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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
65dded18-adec-4e79-9d92-c4b690980214 | storydall-e-adapting-pretrained-text-to-image | 2209.06192 | null | https://arxiv.org/abs/2209.06192v1 | https://arxiv.org/pdf/2209.06192v1.pdf | StoryDALL-E: Adapting Pretrained Text-to-Image Transformers for Story Continuation | Recent advances in text-to-image synthesis have led to large pretrained transformers with excellent capabilities to generate visualizations from a given text. However, these models are ill-suited for specialized tasks like story visualization, which requires an agent to produce a sequence of images given a corresponding sequence of captions, forming a narrative. Moreover, we find that the story visualization task fails to accommodate generalization to unseen plots and characters in new narratives. Hence, we first propose the task of story continuation, where the generated visual story is conditioned on a source image, allowing for better generalization to narratives with new characters. Then, we enhance or 'retro-fit' the pretrained text-to-image synthesis models with task-specific modules for (a) sequential image generation and (b) copying relevant elements from an initial frame. Then, we explore full-model finetuning, as well as prompt-based tuning for parameter-efficient adaptation, of the pre-trained model. We evaluate our approach StoryDALL-E on two existing datasets, PororoSV and FlintstonesSV, and introduce a new dataset DiDeMoSV collected from a video-captioning dataset. We also develop a model StoryGANc based on Generative Adversarial Networks (GAN) for story continuation, and compare it with the StoryDALL-E model to demonstrate the advantages of our approach. We show that our retro-fitting approach outperforms GAN-based models for story continuation and facilitates copying of visual elements from the source image, thereby improving continuity in the generated visual story. Finally, our analysis suggests that pretrained transformers struggle to comprehend narratives containing several characters. Overall, our work demonstrates that pretrained text-to-image synthesis models can be adapted for complex and low-resource tasks like story continuation. | ['Mohit Bansal', 'Darryl Hannan', 'Adyasha Maharana'] | 2022-09-13 | null | null | null | null | ['story-continuation', 'story-visualization'] | ['computer-vision', 'computer-vision'] | [ 6.43804491e-01 4.53760743e-01 2.30747491e-01 -1.25467792e-01
-5.94174922e-01 -8.74235988e-01 1.16436136e+00 -5.22129297e-01
2.71666348e-02 7.88553178e-01 5.21758795e-01 -2.39862263e-01
3.67788255e-01 -8.70754957e-01 -1.09613216e+00 -6.37553632e-01
3.63142610e-01 4.07402366e-01 1.29892929e-02 -3.03120166e-01
-2.11395212e-02 3.48599046e-01 -1.48595154e+00 8.02704215e-01
6.75116181e-01 4.74551201e-01 4.11250859e-01 1.02455103e+00
-1.32655442e-01 1.09425354e+00 -9.49483573e-01 -5.01453936e-01
8.48172307e-02 -1.02777255e+00 -6.54987752e-01 2.68969417e-01
5.41598201e-01 -5.68578482e-01 -4.10852402e-01 4.32292938e-01
3.93752426e-01 1.66759282e-01 8.59811485e-01 -1.59812188e+00
-1.13273263e+00 7.81138003e-01 -4.73773718e-01 -2.67348021e-01
6.52890325e-01 8.72183740e-01 6.16641402e-01 -8.14316869e-01
1.27405322e+00 1.28788853e+00 5.41702032e-01 1.06822371e+00
-1.48455822e+00 -5.82724988e-01 9.80217941e-03 -1.59755293e-02
-7.95263112e-01 -4.27107364e-01 8.68805170e-01 -4.93515819e-01
9.78889585e-01 4.06661600e-01 9.81538773e-01 1.83028710e+00
7.88344257e-03 8.47005904e-01 1.21345019e+00 -2.89743274e-01
2.23028570e-01 1.93927750e-01 -7.36125827e-01 4.18448508e-01
-3.22793603e-01 3.31339926e-01 -7.05719948e-01 3.12192589e-01
1.10479558e+00 -2.62050837e-01 -2.43422642e-01 -3.95858228e-01
-1.35446775e+00 8.72665644e-01 4.60821748e-01 2.55829573e-01
-4.36135441e-01 5.49492955e-01 3.23876619e-01 3.19825500e-01
3.27591687e-01 7.73284316e-01 1.14642076e-01 -2.26198912e-01
-1.08137774e+00 6.52096927e-01 6.47877038e-01 1.08852768e+00
3.40468884e-01 6.10333025e-01 -6.64202571e-01 4.66570139e-01
-2.05181882e-01 4.80906963e-01 3.31189424e-01 -1.04849887e+00
5.17987907e-01 3.98190081e-01 -5.16458377e-02 -6.83577836e-01
-1.14985906e-01 -2.27078021e-01 -8.96884024e-01 5.92826068e-01
4.41289902e-01 -3.57240796e-01 -1.18193531e+00 1.98656082e+00
1.01384856e-01 1.94336995e-01 2.65752286e-01 8.07694197e-01
1.06932116e+00 1.12416995e+00 8.36650133e-02 -3.84667283e-03
1.23831904e+00 -1.19029844e+00 -6.66113377e-01 -3.74883682e-01
3.52140188e-01 -5.52904487e-01 1.64115059e+00 2.01145366e-01
-1.47665632e+00 -6.95747793e-01 -8.92748296e-01 -2.00521290e-01
-4.15593505e-01 -1.21330060e-01 5.43935239e-01 4.27036285e-01
-1.27427328e+00 4.48861271e-01 -4.16648120e-01 -3.68456036e-01
5.73213100e-01 -2.04518214e-01 -3.61950129e-01 7.84646161e-03
-9.59332705e-01 7.91233182e-01 5.52414298e-01 -2.04962656e-01
-1.35379124e+00 -9.82818842e-01 -9.88750279e-01 8.58933926e-02
1.87619448e-01 -1.11065865e+00 1.17570186e+00 -1.29158044e+00
-1.68241143e+00 8.31229150e-01 1.08124092e-01 -5.34148514e-01
8.86030376e-01 6.19585142e-02 -1.81562752e-01 2.06884846e-01
-1.78841129e-02 1.28692555e+00 1.15106797e+00 -1.67616057e+00
-1.47918552e-01 3.39571834e-01 2.93878227e-01 2.72706032e-01
-1.68033764e-01 -6.76074475e-02 -2.56539643e-01 -1.08336008e+00
-6.34627879e-01 -7.18546391e-01 -3.05132531e-02 -1.60682350e-02
-5.00358284e-01 2.08011970e-01 1.09754956e+00 -7.77062595e-01
9.12034273e-01 -2.01088309e+00 3.35119635e-01 -1.50388464e-01
2.26498187e-01 6.56403825e-02 -4.97218013e-01 7.01705635e-01
-3.38689178e-01 3.02646488e-01 -4.56478477e-01 -6.43273592e-01
3.66526865e-03 1.59486294e-01 -6.57153845e-01 -1.49333149e-01
5.12444675e-01 1.34190679e+00 -7.88995504e-01 -4.10029650e-01
2.46191353e-01 6.31453991e-01 -7.22277343e-01 6.07983470e-01
-6.57702148e-01 8.21166337e-01 6.59917668e-02 2.17049718e-01
3.02848935e-01 -2.75652200e-01 -6.65081665e-02 8.24175403e-02
4.45163026e-02 7.47595876e-02 -7.21591175e-01 1.95155871e+00
-5.86350799e-01 1.08293450e+00 -3.43502104e-01 -6.38915956e-01
8.37490797e-01 3.28080982e-01 1.38860762e-01 -7.10602880e-01
7.51691312e-02 -2.32351109e-01 -2.46551514e-01 -6.65349901e-01
7.35519707e-01 -2.26966918e-01 -2.50427127e-01 7.63050020e-01
1.02779828e-01 -6.64974272e-01 3.61494154e-01 4.74968612e-01
1.06510639e+00 7.44408488e-01 1.33252656e-02 3.14310402e-01
3.82150747e-02 2.07902908e-01 -1.71510562e-01 8.04352403e-01
4.54888642e-01 1.07663453e+00 7.27315545e-01 -3.08511227e-01
-1.52790165e+00 -1.19531262e+00 4.05283958e-01 9.26622450e-01
-1.93357825e-01 -3.63535404e-01 -1.07688367e+00 -6.47574425e-01
-2.53963053e-01 1.34595478e+00 -9.33490157e-01 -1.49908558e-01
-7.73104846e-01 -3.93034041e-01 6.00471139e-01 6.05903924e-01
4.59425032e-01 -1.75448763e+00 -1.09784830e+00 2.15822294e-01
-1.03374921e-01 -1.03536546e+00 -5.48341751e-01 -1.14674736e-02
-5.51220834e-01 -6.91361129e-01 -1.03527582e+00 -6.61616981e-01
6.46036208e-01 2.71981508e-02 1.26164174e+00 2.45379889e-03
-1.18850224e-01 4.20565486e-01 -3.37085098e-01 -2.54788727e-01
-1.01744115e+00 -3.38145047e-02 -4.91567820e-01 -6.40210658e-02
-5.68197191e-01 -8.21284711e-01 -4.13269728e-01 2.77734976e-02
-1.28623533e+00 1.07688761e+00 3.72803658e-01 9.73189950e-01
2.92208523e-01 -4.02350724e-01 5.31358778e-01 -9.51457202e-01
7.57421911e-01 -3.92597169e-01 -1.98826686e-01 2.41520107e-01
-3.04577112e-01 1.22968867e-01 8.73311043e-01 -9.08169806e-01
-1.35069442e+00 1.00600896e-02 -1.20995129e-02 -7.41735220e-01
-1.14542924e-01 2.95997381e-01 -1.17900394e-01 3.24911386e-01
7.80401707e-01 3.51215422e-01 -1.00883722e-01 -1.41542941e-01
9.86335874e-01 7.31212422e-02 1.05240870e+00 -6.10266387e-01
1.09794986e+00 4.29005265e-01 -1.53298557e-01 -4.24733311e-01
-3.31520975e-01 4.93743896e-01 -3.76704454e-01 -4.73649472e-01
1.03700423e+00 -8.25119734e-01 -4.79558110e-01 4.05810058e-01
-1.43174624e+00 -1.00923157e+00 -9.34640706e-01 -1.42767921e-01
-1.16473174e+00 -1.45719141e-01 -5.13962328e-01 -5.78093708e-01
-2.22953498e-01 -1.02994239e+00 1.00244474e+00 3.64993215e-01
-5.40763199e-01 -1.05397511e+00 4.70583886e-02 1.76941752e-01
4.57739681e-01 9.88553703e-01 9.24515724e-01 -2.69552976e-01
-5.99902749e-01 1.32313222e-01 -2.71504164e-01 -2.22476237e-02
-1.17564954e-01 7.24853277e-02 -1.02864563e+00 -1.37878299e-01
-2.75819689e-01 -6.27890408e-01 7.81351864e-01 2.29630068e-01
1.06501639e+00 -6.65071249e-01 -1.07718892e-01 9.70779300e-01
1.20236588e+00 3.07429075e-01 9.44128394e-01 3.58274728e-01
8.79887462e-01 5.67131877e-01 2.80136257e-01 3.21409911e-01
3.06722999e-01 6.03188932e-01 4.85107929e-01 -4.71602470e-01
-6.59888148e-01 -9.37245965e-01 5.23080766e-01 2.85898209e-01
-1.19220145e-01 -6.94740772e-01 -6.43788040e-01 5.95966160e-01
-1.80498278e+00 -1.23407912e+00 1.16989389e-01 1.58465326e+00
9.38514888e-01 1.52595537e-02 1.90263644e-01 -8.01610500e-02
4.55934674e-01 4.52335238e-01 -7.75807023e-01 -6.78252280e-01
-4.08354104e-01 3.35500330e-01 -4.38031601e-03 2.66805947e-01
-6.54193997e-01 1.21195149e+00 6.32530022e+00 8.01574051e-01
-1.12334740e+00 2.02868327e-01 7.17181742e-01 -3.26429665e-01
-7.71856785e-01 1.08021043e-01 -1.65044516e-01 3.45829397e-01
7.87734210e-01 -2.66782314e-01 7.18892753e-01 7.42503583e-01
2.32264072e-01 7.49578327e-02 -1.28720593e+00 9.09690261e-01
3.27639103e-01 -1.85713172e+00 5.20485759e-01 -2.07533985e-01
1.03715289e+00 -7.00994432e-01 3.39002132e-01 3.85065883e-01
5.37508726e-01 -1.37104607e+00 1.24847615e+00 5.19395828e-01
1.33028877e+00 -6.19860709e-01 7.84810930e-02 2.07589358e-01
-1.01993656e+00 9.72959995e-02 1.25347391e-01 -1.51839666e-02
4.88448083e-01 -1.69537529e-01 -1.06890512e+00 3.33674997e-01
4.62343246e-01 6.56325221e-01 -7.07558215e-01 4.58390027e-01
-6.34452343e-01 4.41483557e-01 4.00184207e-02 7.22970292e-02
3.35410178e-01 8.21032003e-02 6.11297667e-01 1.28280401e+00
4.70778525e-01 2.43784133e-02 -2.75554031e-01 1.53209233e+00
-2.19255999e-01 -1.80248335e-01 -8.72594893e-01 -2.90760994e-01
2.00101823e-01 1.10833359e+00 -6.88758969e-01 -5.28963268e-01
-1.81594323e-02 1.29878747e+00 2.81272143e-01 5.99648356e-01
-1.14226234e+00 -1.06750697e-01 2.05674052e-01 3.12442034e-01
2.19203785e-01 -8.06780308e-02 -5.14916062e-01 -1.08078802e+00
-1.66961834e-01 -1.20169485e+00 8.44510272e-02 -1.61276948e+00
-7.92781115e-01 8.67106557e-01 2.48605102e-01 -1.07899249e+00
-8.00050735e-01 -1.31134570e-01 -9.92677629e-01 7.56328046e-01
-1.01674986e+00 -1.70674276e+00 -4.07165855e-01 5.96846163e-01
8.77888739e-01 -2.21883982e-01 6.82650328e-01 -2.17413068e-01
-3.26814175e-01 5.28227210e-01 -1.82747513e-01 -5.21546565e-02
6.38259828e-01 -1.29087901e+00 9.45644557e-01 1.04646206e+00
2.39241213e-01 1.23810023e-01 1.01618981e+00 -7.34955609e-01
-1.15302444e+00 -1.13339186e+00 4.94497180e-01 -5.79941034e-01
4.68250424e-01 -7.30271518e-01 -8.44941258e-01 8.05308282e-01
8.00570071e-01 -4.51995879e-01 1.62837803e-01 -4.41343337e-01
-3.72762829e-01 3.51070076e-01 -1.05679309e+00 1.05525184e+00
1.35169923e+00 -3.46801341e-01 -5.24715006e-01 1.67343095e-01
1.01130164e+00 -6.50618255e-01 -5.87552071e-01 -4.97382730e-02
6.30834758e-01 -1.07107735e+00 1.02064919e+00 -6.84083462e-01
1.29366076e+00 -6.97532520e-02 2.28540912e-01 -1.59042287e+00
-1.49989605e-01 -1.04860687e+00 -1.17997140e-01 1.43256879e+00
3.44865859e-01 -2.46433809e-01 6.73220575e-01 5.37970364e-01
-2.16038406e-01 -4.84121442e-01 -5.31663477e-01 -5.98443270e-01
1.03626154e-01 -3.46841186e-01 8.64057899e-01 9.33929801e-01
-2.07648501e-01 3.44901711e-01 -7.60866702e-01 -3.90173554e-01
2.72831470e-01 1.24521323e-01 1.31791735e+00 -5.28498232e-01
-5.71429193e-01 -5.28403699e-01 2.32799441e-01 -8.13475549e-01
1.63470238e-01 -7.48442352e-01 1.40197165e-02 -1.73989403e+00
2.91651845e-01 -1.09749258e-01 3.81504536e-01 6.90171719e-01
-2.10743591e-01 3.77607226e-01 8.56386185e-01 1.16487175e-01
-1.55547678e-01 6.22579515e-01 1.85831809e+00 -1.65123612e-01
-3.33611041e-01 -5.09590328e-01 -6.36035919e-01 4.75042641e-01
5.44909120e-01 -4.50854927e-01 -6.92012131e-01 -3.94155890e-01
3.02897632e-01 4.39101964e-01 8.21909070e-01 -8.90702844e-01
-1.13103926e-01 -3.78488094e-01 7.06660807e-01 -4.27323610e-01
5.76850593e-01 -4.77626324e-01 7.68488109e-01 3.06317657e-01
-5.57070673e-01 2.37108052e-01 3.84603113e-01 2.58470356e-01
1.09915659e-01 -1.08946126e-03 6.22888207e-01 -1.94603339e-01
-4.53175187e-01 -2.34234054e-02 -3.63457292e-01 7.92610422e-02
1.03239989e+00 -5.12880325e-01 -5.61917186e-01 -8.92914593e-01
-6.76589429e-01 1.68382060e-02 7.61777103e-01 6.55975401e-01
7.71088898e-01 -1.47686887e+00 -9.41642463e-01 1.93430647e-01
-4.87227514e-02 1.38856679e-01 4.05900180e-01 2.71480829e-01
-5.09617984e-01 -1.68969538e-02 -6.76162660e-01 -5.44701397e-01
-1.09852374e+00 8.07611644e-01 2.36744992e-02 -4.46017295e-01
-8.55244100e-01 6.89665556e-01 7.57565558e-01 -7.93112665e-02
-1.22172825e-01 -3.13671470e-01 -5.39090335e-02 2.41811504e-03
4.58242744e-01 1.42160878e-02 -5.46962798e-01 -4.64641780e-01
2.15271577e-01 3.21669698e-01 1.82450220e-01 -6.43391192e-01
1.48362648e+00 2.30406821e-02 3.42578113e-01 3.51311624e-01
8.41705739e-01 -2.19768494e-01 -1.91094315e+00 7.57027119e-02
-5.06428301e-01 -2.64199674e-01 -3.92330676e-01 -1.06965923e+00
-1.08574617e+00 9.41675365e-01 1.86091468e-01 1.33571908e-01
1.35813415e+00 -5.99662252e-02 7.38178074e-01 -6.63128681e-03
6.24950081e-02 -6.87063575e-01 6.55705214e-01 1.66786849e-01
1.65003026e+00 -7.89821208e-01 -2.49048054e-01 -5.43525331e-02
-1.12920773e+00 1.08931553e+00 6.57733798e-01 -1.85396641e-01
-1.17904872e-01 4.69688743e-01 -4.48165573e-02 -1.53583467e-01
-9.83562231e-01 -9.32226796e-03 3.30498308e-01 9.36762512e-01
5.01332022e-02 -7.24758208e-02 2.32917011e-01 3.57490420e-01
-7.98962951e-01 -4.08831565e-03 7.10167229e-01 6.67576075e-01
1.80572003e-01 -9.16270852e-01 -5.51376700e-01 1.08064614e-01
2.49848384e-02 -2.17915773e-01 -5.70190489e-01 1.11500180e+00
1.18670270e-01 6.53688848e-01 1.65960521e-01 -3.30296844e-01
2.77874917e-01 1.21269681e-01 7.38414109e-01 -5.93208969e-01
-9.01501894e-01 6.17614314e-02 2.04724982e-01 -4.65218544e-01
-3.06943387e-01 -6.87929749e-01 -1.01061118e+00 -4.12374198e-01
3.01811099e-01 -3.46555620e-01 5.83295226e-01 7.67561376e-01
3.06991696e-01 1.05004418e+00 3.46991003e-01 -1.15940845e+00
-1.05981380e-02 -8.99234474e-01 -1.42007604e-01 7.67585039e-01
2.14733094e-01 -3.30751926e-01 -9.25266147e-02 7.25881517e-01] | [11.203125, 0.5130406022071838] |
0465f2a0-b4ba-41ff-878d-56edd7b1c7fd | mdqe-mining-discriminative-query-embeddings | 2303.14395 | null | https://arxiv.org/abs/2303.14395v1 | https://arxiv.org/pdf/2303.14395v1.pdf | MDQE: Mining Discriminative Query Embeddings to Segment Occluded Instances on Challenging Videos | While impressive progress has been achieved, video instance segmentation (VIS) methods with per-clip input often fail on challenging videos with occluded objects and crowded scenes. This is mainly because instance queries in these methods cannot encode well the discriminative embeddings of instances, making the query-based segmenter difficult to distinguish those `hard' instances. To address these issues, we propose to mine discriminative query embeddings (MDQE) to segment occluded instances on challenging videos. First, we initialize the positional embeddings and content features of object queries by considering their spatial contextual information and the inter-frame object motion. Second, we propose an inter-instance mask repulsion loss to distance each instance from its nearby non-target instances. The proposed MDQE is the first VIS method with per-clip input that achieves state-of-the-art results on challenging videos and competitive performance on simple videos. In specific, MDQE with ResNet50 achieves 33.0\% and 44.5\% mask AP on OVIS and YouTube-VIS 2021, respectively. Code of MDQE can be found at \url{https://github.com/MinghanLi/MDQE_CVPR2023}. | ['Lei Zhang', 'Wangmeng Xiang', 'Shuai Li', 'Minghan Li'] | 2023-03-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Li_MDQE_Mining_Discriminative_Query_Embeddings_To_Segment_Occluded_Instances_on_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Li_MDQE_Mining_Discriminative_Query_Embeddings_To_Segment_Occluded_Instances_on_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-instance-segmentation'] | ['computer-vision'] | [-5.91557734e-02 -1.21258937e-01 -4.97054011e-01 -2.95960307e-01
-1.03095436e+00 -5.92036545e-01 2.21600667e-01 -1.35634005e-01
-5.20504415e-01 5.18397629e-01 1.28296930e-02 1.82532683e-01
-6.28860444e-02 -4.57520515e-01 -9.61998820e-01 -4.82874364e-01
-2.82354027e-01 3.43649119e-01 8.12089145e-01 9.01558772e-02
5.12559786e-02 1.75871372e-01 -1.46383250e+00 2.97153354e-01
8.33038867e-01 1.20579839e+00 4.45260525e-01 6.54211521e-01
-4.18777317e-02 6.70394361e-01 -6.19259059e-01 -3.90132040e-01
3.86384338e-01 -4.05132115e-01 -8.83070230e-01 1.04030684e-01
8.60704005e-01 -4.93649304e-01 -7.97359765e-01 9.75637674e-01
4.69948262e-01 3.92319262e-01 5.05723298e-01 -1.40630400e+00
-6.49937987e-01 4.45625454e-01 -6.74600661e-01 8.64488363e-01
3.21360946e-01 2.92035878e-01 1.07213855e+00 -1.22742724e+00
8.18526149e-01 1.03862333e+00 4.32798415e-01 5.57181358e-01
-1.06193995e+00 -6.89685345e-01 3.63947898e-01 6.57655001e-01
-1.73962617e+00 -3.99824739e-01 5.23725808e-01 -5.08800983e-01
6.67495072e-01 3.12727302e-01 6.20820701e-01 1.00701833e+00
-2.19037607e-01 1.23808205e+00 5.41760683e-01 3.12754661e-01
1.16664127e-01 -2.27963664e-02 6.10016100e-02 6.78910077e-01
-2.51718573e-02 -8.37977603e-02 -4.92740154e-01 2.32181013e-01
8.00289035e-01 1.93811774e-01 -6.16990328e-01 -1.76240221e-01
-1.32416916e+00 7.53960073e-01 6.32422864e-01 1.79561719e-01
-3.34180236e-01 4.19733375e-01 4.00095284e-01 1.48708269e-01
4.55673933e-01 3.25431317e-01 -4.77814734e-01 -4.03268665e-01
-1.16776896e+00 2.39393681e-01 4.82503146e-01 1.23673904e+00
6.94451630e-01 -2.16779001e-02 -4.70903993e-01 9.35617566e-01
3.38278748e-02 5.14690876e-01 1.54355779e-01 -1.13559616e+00
6.76549196e-01 4.28227037e-01 8.54346529e-02 -9.91183937e-01
1.00777432e-01 -3.61938119e-01 -5.48993230e-01 -3.37341666e-01
3.18801463e-01 -8.90688747e-02 -1.12488782e+00 1.27767730e+00
4.12034631e-01 7.75923610e-01 -2.40853861e-01 1.22481441e+00
1.25908351e+00 1.00716412e+00 1.47580430e-01 -3.97408567e-02
1.18477547e+00 -1.32160199e+00 -5.45275807e-01 -2.17147976e-01
5.58777928e-01 -7.17618704e-01 1.02023780e+00 1.84593022e-01
-1.08552718e+00 -8.00937057e-01 -8.57695937e-01 -1.62807018e-01
-2.40379021e-01 1.76067859e-01 1.54541090e-01 1.15979701e-01
-7.79311597e-01 3.34153742e-01 -8.29497814e-01 -5.90549558e-02
9.57864642e-01 3.73181939e-01 -3.46062273e-01 -4.33572114e-01
-1.03273749e+00 2.88119227e-01 3.25370103e-01 1.00442074e-01
-1.32648778e+00 -9.05519187e-01 -9.60640252e-01 -1.09197907e-01
8.93168151e-01 -1.95912912e-01 9.84185934e-01 -1.01639390e+00
-8.51499856e-01 7.64001429e-01 -3.26929539e-01 -4.34931099e-01
6.68078959e-01 -5.62740862e-01 -2.65428156e-01 5.82471132e-01
3.42790753e-01 1.17664981e+00 7.92832911e-01 -1.15765047e+00
-6.91248953e-01 -2.00162083e-01 2.21233293e-01 2.14238241e-01
-1.53850392e-01 -1.08571984e-01 -1.23986399e+00 -7.58523047e-01
-3.84427346e-02 -9.10131693e-01 -2.27487370e-01 2.12302074e-01
-5.35013855e-01 -4.13533062e-01 1.13785148e+00 -6.30907893e-01
1.37736607e+00 -2.43205976e+00 1.71748176e-01 -2.18027011e-01
2.33098000e-01 5.71700096e-01 -4.98434991e-01 2.11207762e-01
2.26021141e-01 1.91730499e-01 -1.65742651e-01 -3.21519941e-01
-3.53898294e-02 3.61516923e-01 -1.86474055e-01 5.57482362e-01
3.49137098e-01 1.08304739e+00 -9.43204463e-01 -6.92608893e-01
2.85348833e-01 6.50578201e-01 -7.17889071e-01 3.14160526e-01
-3.23546827e-01 4.47306722e-01 -5.29520810e-01 8.29666436e-01
7.19265282e-01 -1.98360875e-01 -3.01512837e-01 -3.32054079e-01
6.86820596e-02 -5.37950806e-02 -1.28083897e+00 1.75406551e+00
2.80496869e-02 9.17282045e-01 4.99462560e-02 -1.07459593e+00
4.60791647e-01 1.72678664e-01 6.75086915e-01 -7.40874231e-01
1.15136569e-02 2.23453403e-01 -3.59573871e-01 -7.65649557e-01
5.80686986e-01 5.38602114e-01 9.73823667e-02 -4.05754484e-02
1.25408515e-01 2.79465597e-02 5.38188338e-01 3.43009621e-01
1.10864913e+00 2.24203169e-01 -1.45709336e-01 -7.84662887e-02
4.41439480e-01 -8.34624022e-02 8.48238051e-01 6.62925005e-01
-5.74072480e-01 1.01519930e+00 5.49183607e-01 -4.31267291e-01
-8.57247770e-01 -1.23230076e+00 -2.50530422e-01 1.26867390e+00
8.24056864e-01 -5.06558120e-01 -7.02726066e-01 -9.42049265e-01
-6.30468205e-02 2.77069688e-01 -5.74524879e-01 1.39221430e-01
-7.16633797e-01 -2.94252545e-01 3.41034681e-01 6.58336878e-01
6.23102307e-01 -1.02006972e+00 -4.63850707e-01 2.89424568e-01
-3.51499736e-01 -1.46982479e+00 -8.22616994e-01 -1.51577622e-01
-5.60567796e-01 -1.34314859e+00 -1.02348971e+00 -9.25211251e-01
5.23777604e-01 4.80645061e-01 1.22849774e+00 1.12467960e-01
-4.82341737e-01 3.56453270e-01 -7.26089418e-01 4.61785076e-03
1.55769080e-01 1.82388023e-01 -2.02351362e-01 -3.60704549e-02
4.94914740e-01 -2.81993926e-01 -1.17866302e+00 6.98178053e-01
-1.13037002e+00 -2.13782936e-01 3.53729546e-01 6.60745919e-01
1.02547264e+00 -2.49959394e-01 3.86375517e-01 -5.56835175e-01
3.68207060e-02 -7.53858328e-01 -4.56994355e-01 7.73344636e-02
-1.20092675e-01 -4.50863540e-01 3.63061488e-01 -4.99504685e-01
-3.50606859e-01 1.17439538e-01 -2.67732114e-01 -9.25984561e-01
-2.43519723e-01 1.50969237e-01 -7.72008523e-02 2.72213548e-01
3.81573379e-01 3.85923833e-01 -3.20998669e-01 -4.26727235e-01
2.47970253e-01 5.37143886e-01 4.69993770e-01 -4.13024098e-01
6.41723871e-01 4.10115451e-01 -5.05784690e-01 -9.70724523e-01
-9.66724694e-01 -8.24932992e-01 -5.18948913e-01 -2.77623683e-01
1.34896517e+00 -1.13876474e+00 -3.21120948e-01 1.05533965e-01
-1.00209212e+00 -5.84163368e-01 -3.69654745e-01 5.30637801e-01
-5.02122521e-01 2.94738531e-01 -6.25700891e-01 -3.98019969e-01
-4.18983400e-02 -1.51412034e+00 1.19712961e+00 2.73547292e-01
4.72275391e-02 -7.07164168e-01 -1.50679454e-01 5.18691957e-01
1.65525094e-01 2.38469809e-01 2.32403561e-01 -5.66918373e-01
-1.00545931e+00 -1.62915930e-01 -4.38709587e-01 5.93285143e-01
-1.90341426e-03 4.09723446e-02 -7.38473952e-01 -5.22178233e-01
-4.62115854e-01 -3.16041797e-01 1.04014981e+00 6.35140479e-01
1.51117039e+00 -2.83741415e-01 -4.02626604e-01 8.65234375e-01
1.51290798e+00 1.54685915e-01 6.51011705e-01 2.94061638e-02
1.06663346e+00 2.01854631e-01 9.45174754e-01 4.91132766e-01
2.50686616e-01 8.80440831e-01 6.24321938e-01 -1.03698045e-01
-2.31288478e-01 -7.42865652e-02 4.07322317e-01 5.24548233e-01
-7.77140707e-02 -4.19911563e-01 -7.45834529e-01 9.18076575e-01
-1.79403555e+00 -1.06237113e+00 -2.11247820e-02 1.97275543e+00
5.43936908e-01 7.12744072e-02 3.16048771e-01 -1.06362157e-01
7.69188821e-01 5.97991109e-01 -6.39603078e-01 5.99522628e-02
-8.70998427e-02 1.36610409e-02 5.54659665e-01 4.11903471e-01
-1.42612493e+00 1.02642071e+00 4.51465702e+00 1.03325641e+00
-9.17919278e-01 4.64294136e-01 7.97850251e-01 -5.02370596e-01
-6.05192818e-02 -2.82314092e-01 -9.51957643e-01 7.83795595e-01
6.37170732e-01 1.84403688e-01 2.20932364e-01 9.17671680e-01
2.86260128e-01 -1.57252043e-01 -1.08206236e+00 1.19909298e+00
1.18576676e-01 -1.49840426e+00 -2.48676017e-02 6.45608990e-04
9.23018634e-01 3.77930403e-01 1.79893836e-01 4.98926759e-01
-1.81074291e-01 -1.05032015e+00 7.46936440e-01 1.94465727e-01
8.90590310e-01 -6.69114172e-01 7.84656644e-01 -6.13266677e-02
-1.51710236e+00 -1.35236368e-01 -4.62464660e-01 4.32413965e-01
2.81346828e-01 3.66114825e-01 -4.58548278e-01 3.39655668e-01
1.03884912e+00 1.04657936e+00 -6.28103197e-01 1.34291029e+00
-5.18324487e-02 7.60706306e-01 -3.21033090e-01 2.03494802e-01
6.42219305e-01 -9.55491364e-02 7.30343282e-01 1.42244959e+00
2.28895739e-01 1.65653661e-01 5.11390984e-01 6.64945006e-01
-2.10484937e-01 9.20044184e-02 -4.77430135e-01 6.50559692e-03
5.90526819e-01 9.42708075e-01 -7.70392060e-01 -5.05176306e-01
-3.92938942e-01 1.20914602e+00 9.68830064e-02 6.72964096e-01
-1.40272689e+00 -3.69530439e-01 1.00801587e+00 3.67486537e-01
8.90993357e-01 -2.55182147e-01 3.12534600e-01 -1.17248535e+00
3.71837884e-01 -7.66612589e-01 4.42205697e-01 -4.57315326e-01
-1.01632261e+00 7.07277656e-01 7.58298114e-02 -1.43404341e+00
2.08710745e-01 -3.35444719e-01 -3.88292581e-01 3.70348543e-01
-1.44804299e+00 -8.44569743e-01 -5.25927126e-01 7.13649571e-01
1.28810680e+00 8.76074135e-02 1.71732724e-01 7.33777404e-01
-7.10713506e-01 7.96126187e-01 2.01897800e-01 5.13208985e-01
5.66131234e-01 -1.04832375e+00 1.18824519e-01 7.48222947e-01
3.41966718e-01 1.41661748e-01 4.97625202e-01 -4.63631511e-01
-1.33796465e+00 -1.46734703e+00 3.68113548e-01 -5.39837122e-01
4.17846650e-01 -3.61509144e-01 -9.66738760e-01 6.10750020e-01
1.39404848e-01 6.89762175e-01 4.45853114e-01 -4.08899218e-01
-3.36937219e-01 -1.02558106e-01 -9.52753544e-01 5.00576437e-01
1.34675479e+00 -4.66782838e-01 -3.79465461e-01 5.96049249e-01
9.18410718e-01 -4.57212746e-01 -9.44166064e-01 4.57509458e-01
2.32174620e-01 -1.00390100e+00 1.18478036e+00 -5.16924024e-01
6.36377931e-01 -4.92622226e-01 -3.70702237e-01 -8.70402396e-01
-7.77444094e-02 -5.02449095e-01 -3.02703649e-01 1.08178115e+00
3.10410172e-01 -2.44448602e-01 9.62729037e-01 3.40170950e-01
-2.75527477e-01 -1.17865133e+00 -1.07872188e+00 -9.66937006e-01
-9.38574076e-02 -5.52818239e-01 3.51565123e-01 7.61822104e-01
-5.90091705e-01 2.54844446e-02 -2.43643463e-01 2.64315754e-01
4.99583602e-01 9.08142924e-02 7.36724019e-01 -7.39070475e-01
-2.27618337e-01 -3.49492788e-01 -7.22415388e-01 -1.65611625e+00
3.51347588e-02 -7.08409429e-01 7.06686303e-02 -1.57016444e+00
1.38757065e-01 -5.37899911e-01 -3.47219467e-01 2.37826750e-01
-3.33104074e-01 8.67501974e-01 4.68366474e-01 3.37852508e-01
-1.31957436e+00 6.64220810e-01 1.41421759e+00 -3.21301043e-01
-2.95526832e-01 -1.72429606e-01 -1.27089128e-01 4.83639240e-01
6.93471491e-01 -5.74372649e-01 -5.39004624e-01 -6.89706504e-01
-1.43058211e-01 2.07712641e-03 5.74244618e-01 -1.19643569e+00
4.56209667e-02 -1.54961154e-01 4.02424008e-01 -8.88227582e-01
6.32773161e-01 -6.26647353e-01 2.87299119e-02 3.71694863e-01
-3.11503679e-01 1.11732058e-01 6.90710694e-02 8.62007856e-01
-5.12782156e-01 -1.27081856e-01 6.91426456e-01 -7.16573894e-02
-1.16807306e+00 8.24463964e-01 -3.83936279e-02 6.16617978e-01
1.32504189e+00 -5.57297051e-01 -2.21403077e-01 -2.45087311e-01
-8.08263183e-01 5.55605829e-01 3.54906082e-01 6.34367883e-01
8.83604348e-01 -1.18851423e+00 -8.44672680e-01 -2.69164853e-02
1.37956977e-01 3.17478448e-01 6.59248412e-01 1.10993159e+00
-6.88914359e-01 2.83699661e-01 5.56407422e-02 -9.13068891e-01
-1.31849265e+00 5.71551919e-01 2.30234429e-01 3.05209160e-01
-7.04634964e-01 1.31713045e+00 4.74958807e-01 7.89358094e-02
5.31401992e-01 -2.62762010e-01 -1.33073911e-01 2.19934851e-01
5.82255602e-01 4.14426625e-01 -3.98499429e-01 -9.13976550e-01
-4.21265572e-01 6.81415439e-01 -2.22341225e-01 2.24030599e-01
1.23195648e+00 -9.01881531e-02 3.24579090e-01 1.05257727e-01
1.74608064e+00 -1.80231616e-01 -1.72499537e+00 -2.17973620e-01
-2.60070086e-01 -9.07470703e-01 -2.53526926e-01 -3.33256364e-01
-1.52745998e+00 8.39215875e-01 7.26095021e-01 8.50890800e-02
1.00114179e+00 3.67984354e-01 1.20540440e+00 7.18123913e-02
1.74954101e-01 -1.07809973e+00 2.93527037e-01 3.07941705e-01
7.66801059e-01 -1.43681753e+00 -4.15116027e-02 -4.76630658e-01
-7.25564539e-01 6.74459994e-01 9.00371194e-01 -3.06344867e-01
6.68174326e-01 -1.89775079e-01 9.25874263e-02 -2.84467608e-01
-5.33888996e-01 -3.28002512e-01 4.18517411e-01 4.21938568e-01
1.63644314e-01 -4.88161109e-02 -2.55446255e-01 2.65079856e-01
1.67323753e-01 -7.44264424e-02 3.36453736e-01 7.72749841e-01
-3.60594183e-01 -6.09380603e-01 -1.23766661e-01 5.88517606e-01
-6.68170333e-01 -2.68250052e-02 -6.08889200e-02 8.27151299e-01
4.01510537e-01 8.29585493e-01 2.72741437e-01 -4.44610596e-01
3.65588844e-01 -3.69842887e-01 2.01094091e-01 -5.52674472e-01
-3.69250238e-01 1.21058904e-01 -5.72601706e-02 -1.00844431e+00
-4.09518123e-01 -4.15511906e-01 -1.19101763e+00 -1.52626693e-01
-1.51702330e-01 2.07598642e-01 2.50921130e-01 6.10074937e-01
5.57844162e-01 4.06430066e-01 5.24436772e-01 -1.06978333e+00
-1.92384094e-01 -7.64855862e-01 -4.25173789e-01 6.32363260e-01
2.63688117e-01 -7.66007125e-01 -4.39430505e-01 1.24996476e-01] | [9.2901029586792, 0.06659847497940063] |
75d6c0c6-4a6d-41f1-a967-cd98240b7661 | a-fully-automated-and-explainable-algorithm | 2307.03757 | null | https://arxiv.org/abs/2307.03757v1 | https://arxiv.org/pdf/2307.03757v1.pdf | A Fully Automated and Explainable Algorithm for the Prediction of Malignant Transformation in Oral Epithelial Dysplasia | Oral epithelial dysplasia (OED) is a premalignant histopathological diagnosis given to lesions of the oral cavity. Its grading suffers from significant inter-/intra- observer variability, and does not reliably predict malignancy progression, potentially leading to suboptimal treatment decisions. To address this, we developed a novel artificial intelligence algorithm that can assign an Oral Malignant Transformation (OMT) risk score, based on histological patterns in the in Haematoxylin and Eosin stained whole slide images, to quantify the risk of OED progression. The algorithm is based on the detection and segmentation of nuclei within (and around) the epithelium using an in-house segmentation model. We then employed a shallow neural network fed with interpretable morphological/spatial features, emulating histological markers. We conducted internal cross-validation on our development cohort (Sheffield; n = 193 cases) followed by independent validation on two external cohorts (Birmingham and Belfast; n = 92 cases). The proposed OMTscore yields an AUROC = 0.74 in predicting whether an OED progresses to malignancy or not. Survival analyses showed the prognostic value of our OMTscore for predicting malignancy transformation, when compared to the manually-assigned WHO and binary grades. Analysis of the correctly predicted cases elucidated the presence of peri-epithelial and epithelium-infiltrating lymphocytes in the most predictive patches of cases that transformed (p < 0.0001). This is the first study to propose a completely automated algorithm for predicting OED transformation based on interpretable nuclear features, whilst being validated on external datasets. The algorithm shows better-than-human-level performance for prediction of OED malignant transformation and offers a promising solution to the challenges of grading OED in routine clinical practice. | ['Nasir M Rajpoot', 'Syed Ali Khurram', 'Hisham Mehanna', 'Paul Nankivell', 'Jill Brooks', 'Jacqueline James', 'Stephanie G Craig', 'Kris D McCombe', 'Shan E Ahmed Raza', 'Fayyaz Minhas', 'Mostafa Jahanifar', 'Hanya Mahmood', 'Raja Muhammad Saad Bashir', 'Adam J Shephard'] | 2023-07-06 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 3.75115305e-01 5.83862960e-01 -2.80092180e-01 -9.84163806e-02
-1.02965057e+00 -3.27733010e-01 4.76376384e-01 6.78787351e-01
-6.73720837e-01 4.00424063e-01 2.84046471e-01 -7.56727159e-01
-3.49271655e-01 -7.71734416e-01 -5.40965982e-02 -1.11210287e+00
-5.83946109e-02 9.66505647e-01 1.94872156e-01 4.13210131e-02
-3.26909055e-03 4.10416663e-01 -1.43329668e+00 3.75806987e-01
7.66622782e-01 7.15406001e-01 2.35117599e-01 1.03532839e+00
1.16343491e-01 5.49838543e-01 -3.30470353e-01 -4.06428516e-01
-1.39281034e-01 -1.77469417e-01 -8.09797823e-01 -7.26910159e-02
1.68104008e-01 -7.47915953e-02 8.06282908e-02 7.32625186e-01
4.75599915e-01 -6.01417720e-01 1.29217374e+00 -7.72844553e-01
-3.65661979e-01 5.98774441e-02 -3.26060593e-01 1.65615946e-01
3.37918490e-01 1.02272943e-01 9.72606599e-01 -4.89074320e-01
8.07470441e-01 6.01190090e-01 1.11237049e+00 4.89540488e-01
-1.08217299e+00 -1.00782089e-01 -5.42219162e-01 2.49552667e-01
-1.22834575e+00 -2.70136684e-01 1.44080408e-02 -6.22271180e-01
9.51185822e-01 6.53744400e-01 1.09072924e+00 3.87640387e-01
5.39426565e-01 5.02849519e-01 1.37839282e+00 -7.63487518e-01
6.55166209e-01 2.60992557e-01 1.77086964e-01 7.70057201e-01
5.73325098e-01 1.43610954e-01 1.33102953e-01 -2.96386838e-01
2.42213801e-01 -2.06071138e-02 -1.66117191e-01 1.21505514e-01
-4.99887824e-01 8.23703289e-01 4.17398065e-01 5.41138887e-01
-2.46697113e-01 -2.30081886e-01 4.88800585e-01 1.57180771e-01
3.02381665e-01 -1.33934570e-03 -3.81741732e-01 2.23022446e-01
-7.86029696e-01 -4.35004115e-01 9.17348504e-01 -1.27206817e-01
3.20047349e-01 -6.66544020e-01 4.95220959e-01 1.25415576e+00
7.89565086e-01 1.64684355e-01 1.30812657e+00 -6.43124759e-01
-4.72723663e-01 1.10237312e+00 -3.74511451e-01 -5.93259573e-01
-8.89810503e-01 1.39440885e-02 -8.06174994e-01 5.38524747e-01
7.45293736e-01 3.55701029e-01 -1.31515348e+00 1.18060195e+00
3.79568428e-01 -3.67016435e-01 2.10976288e-01 5.46999753e-01
5.98839283e-01 1.54222608e-01 3.36270005e-01 -8.58325213e-02
1.70482302e+00 -5.55469036e-01 -4.49636221e-01 -7.23875538e-02
1.26724052e+00 -6.05146885e-01 9.48466361e-01 5.25569856e-01
-7.08375812e-01 1.19154014e-01 -1.02345037e+00 -1.97472662e-01
-7.21072376e-01 2.41314084e-03 7.02576935e-01 1.11828780e+00
-1.53185201e+00 1.36117907e-02 -1.10046494e+00 -9.70279217e-01
4.01234597e-01 8.19722652e-01 -7.92124450e-01 4.90748353e-05
-6.75422132e-01 1.03154922e+00 3.36596400e-01 3.12495441e-03
-2.57583529e-01 -4.72802103e-01 -8.11584830e-01 -4.12639588e-01
-2.79330760e-01 -1.05145776e+00 1.32610106e+00 -9.00216103e-01
-1.55167413e+00 1.50269425e+00 -3.04915309e-01 -9.39401388e-02
4.66641903e-01 5.76533735e-01 -3.71818870e-01 2.17601925e-01
1.23399243e-01 3.92193109e-01 1.33700967e-01 -1.08812225e+00
-1.14537191e+00 -8.22154522e-01 -4.15556043e-01 3.18126641e-02
-2.97160834e-01 -4.65904891e-01 -4.37413305e-01 -3.68079156e-01
1.66220069e-01 -1.16186130e+00 -5.89804649e-01 9.05251801e-02
-2.31310412e-01 -5.88390768e-01 3.82036030e-01 -9.26646948e-01
1.08379376e+00 -2.07751036e+00 -4.33830321e-01 6.66553855e-01
2.21384779e-01 2.58162916e-01 1.90736100e-01 1.97662693e-03
4.68163826e-02 4.81362700e-01 -1.32987052e-01 -1.97548866e-01
-2.09169369e-02 4.26273853e-01 7.63315380e-01 7.34181225e-01
-1.50870308e-01 7.89820373e-01 -7.40576327e-01 -6.87499225e-01
2.95055389e-01 5.77720523e-01 -4.21454906e-01 -1.66467264e-01
6.87594339e-02 -2.78835475e-01 -1.41887993e-01 8.33278477e-01
3.18280369e-01 -4.52989489e-01 4.77145463e-01 -4.03603800e-02
1.77318737e-01 -1.89325549e-02 -8.15838873e-01 9.91721511e-01
-5.90986431e-01 6.06731474e-01 1.70706123e-01 -3.48732680e-01
5.71121097e-01 4.89708841e-01 1.71208337e-01 -7.61700094e-01
4.52263445e-01 8.53370547e-01 1.05874643e-01 -7.60854721e-01
4.96358331e-03 -5.51555514e-01 1.07207961e-01 2.44882643e-01
-4.86176787e-03 -2.06382468e-01 3.32514673e-01 -2.13280737e-01
1.19199347e+00 -4.22232062e-01 9.87105370e-01 -4.05994415e-01
7.61089444e-01 1.70277387e-01 1.72082379e-01 3.39992672e-01
-3.53775114e-01 5.52964807e-01 4.66964692e-01 -5.82429409e-01
-8.37937176e-01 -8.73655558e-01 -7.14461863e-01 6.07865691e-01
-1.77970022e-01 6.00185283e-02 -7.22854555e-01 -5.62250555e-01
1.07461708e-02 4.68641728e-01 -1.10797048e+00 9.34758708e-02
-3.41431141e-01 -1.38525236e+00 5.94408274e-01 2.17255384e-01
-5.60824536e-02 -5.94452322e-01 -2.43528023e-01 4.32881534e-01
1.39081135e-01 -5.07225215e-01 3.61583829e-02 6.96742833e-01
-7.58959949e-01 -1.43520367e+00 -8.15849483e-01 -1.37540483e+00
9.85035062e-01 -2.55875409e-01 6.41964853e-01 4.88418967e-01
-5.16487837e-01 1.66316748e-01 1.76652987e-02 -5.53001761e-01
-1.04184771e+00 9.87497941e-02 -1.50260717e-01 -2.55685300e-01
7.97591507e-01 -4.03256923e-01 -8.17290127e-01 3.46649677e-01
-8.67881954e-01 -5.69939390e-02 9.50816989e-01 8.30806315e-01
9.02178526e-01 -6.07090537e-03 4.51031685e-01 -9.07627285e-01
4.40571398e-01 -5.58471262e-01 -1.48283258e-01 8.92978311e-02
-8.81030202e-01 -2.86523312e-01 3.96759301e-01 -1.05329417e-01
-8.43405187e-01 9.03708562e-02 -4.77952868e-01 8.15065086e-01
-5.36693990e-01 6.06522381e-01 2.86612123e-01 -1.04082581e-02
6.69906020e-01 -1.47126034e-01 4.99295652e-01 -1.23445451e-01
-1.59280419e-01 1.12898767e+00 5.68521857e-01 1.80512834e-02
4.42498505e-01 8.34710181e-01 1.73478737e-01 -1.03077114e+00
-3.75314206e-01 -9.73223150e-01 -4.89891291e-01 2.51192432e-02
9.33437347e-01 -8.51910949e-01 -5.54245293e-01 6.53776288e-01
-3.99224728e-01 -6.45458460e-01 -1.47894949e-01 4.98365581e-01
-4.44705874e-01 2.89584309e-01 -8.51688802e-01 -5.86301923e-01
-5.31450212e-01 -9.57095563e-01 1.09021628e+00 4.42402840e-01
-8.98303509e-01 -1.45788610e+00 6.80687428e-01 6.34423733e-01
2.96988189e-01 3.52807879e-01 1.19596791e+00 -9.47678387e-01
1.26046509e-01 -5.73040783e-01 -6.21897765e-02 -4.54107076e-02
3.21849704e-01 3.35352570e-01 -1.02596891e+00 -1.46353515e-02
-1.26209930e-01 8.52104202e-02 9.05242503e-01 4.58416224e-01
4.23566192e-01 -3.20352823e-01 -8.45609546e-01 3.78315896e-01
1.59778464e+00 5.61235845e-01 6.12619340e-01 1.03995800e+00
1.42717332e-01 7.67702878e-01 1.24406666e-01 -9.99297425e-02
6.32326365e-01 3.59831840e-01 3.60563219e-01 -4.64332283e-01
-3.49506974e-01 3.00520416e-02 2.38415048e-01 5.33473432e-01
-2.86432952e-01 -3.45064610e-01 -1.18882823e+00 8.62503290e-01
-1.01287079e+00 -4.03450042e-01 -4.29907262e-01 1.76603281e+00
9.71327364e-01 2.19794810e-01 -1.04101384e-02 6.20533228e-01
5.57876110e-01 -4.64175582e-01 -1.72532499e-01 -7.40916789e-01
-1.74567774e-02 2.86497027e-01 4.55986530e-01 6.51898921e-01
-7.29097486e-01 1.68731540e-01 6.56531239e+00 6.30542338e-01
-1.27580523e+00 -4.45729792e-02 1.01374006e+00 3.09394389e-01
-1.38056740e-01 -5.10974288e-01 -6.26307309e-01 3.45970541e-01
1.14946234e+00 -9.65489261e-03 -2.29496166e-01 6.40866756e-01
2.45094985e-01 -4.23120528e-01 -9.48516250e-01 2.23544002e-01
1.09151259e-01 -1.12453854e+00 -2.84715593e-01 6.88638806e-01
6.41034424e-01 1.81713805e-01 -7.79657364e-02 8.50430951e-02
4.45808619e-01 -1.02180278e+00 2.96070546e-01 3.07777345e-01
9.77917314e-01 -2.78086454e-01 1.50968516e+00 1.31913751e-01
-7.59411156e-01 -4.19329032e-02 8.76859352e-02 8.39576721e-02
5.19008897e-02 4.32559639e-01 -1.94582415e+00 -5.49289882e-02
7.26606846e-01 1.85749903e-01 -8.54871154e-01 1.15297925e+00
-7.99904391e-02 7.32393920e-01 -9.07067537e-01 -3.96724969e-01
1.84050933e-01 1.19802035e-01 1.07989736e-01 1.21094215e+00
4.51648384e-01 1.18888542e-01 -5.42169869e-01 2.30693948e-02
4.42407846e-01 4.21075344e-01 -7.70817623e-02 3.10536295e-01
1.64757222e-01 1.26149416e+00 -1.10719967e+00 -1.21704988e-01
-3.92887890e-01 5.51361442e-01 -8.73214453e-02 4.68045957e-02
-1.02450207e-01 -1.52332559e-01 1.61627635e-01 3.84330958e-01
-2.52968408e-02 5.25244892e-01 -5.77610552e-01 -4.06879187e-01
-4.47243392e-01 -5.26139915e-01 9.76486444e-01 -4.42013651e-01
-1.25876307e+00 5.04793406e-01 -4.30733144e-01 -8.14300776e-01
-2.80505717e-01 -1.06838727e+00 -8.08098912e-01 5.46556950e-01
-1.52656829e+00 -1.39798009e+00 -4.57687378e-01 2.49285046e-02
7.62404427e-02 6.27739802e-02 1.29302311e+00 2.01143464e-03
-4.04648930e-01 7.43361652e-01 2.35351875e-01 -7.99274147e-02
5.39879918e-01 -1.84208167e+00 -3.45677763e-01 -4.48149927e-02
-5.54105341e-01 4.99069422e-01 5.58708370e-01 -4.42229062e-01
-7.85197496e-01 -9.60194290e-01 1.15072918e+00 -3.41374546e-01
8.28356028e-01 3.84820193e-01 -7.34371483e-01 5.80088496e-01
-3.70917208e-02 -2.08593890e-01 1.72078311e+00 -2.09113434e-01
-5.94823994e-02 1.54020742e-01 -1.46210432e+00 7.00849295e-01
2.92219013e-01 -2.81723559e-01 -5.73103607e-01 3.05318028e-01
3.82344052e-02 -1.73990339e-01 -1.44906294e+00 3.74280930e-01
9.51050520e-01 -9.47554767e-01 8.29669893e-01 1.15671881e-01
1.52975813e-01 -4.03676450e-01 -1.31414188e-02 -9.16565895e-01
-3.18459094e-01 -1.76498175e-01 3.53938401e-01 8.92601132e-01
6.65826976e-01 -7.98934639e-01 1.14227772e+00 5.47438264e-01
-3.65093827e-01 -1.03792405e+00 -1.36050224e+00 -1.64626479e-01
2.40123779e-01 -7.74426684e-02 3.36807221e-01 6.14737153e-01
1.96554169e-01 -1.24702476e-01 7.84245312e-01 1.95951164e-01
3.26946944e-01 -6.47945702e-01 4.66863215e-01 -1.28533864e+00
-2.85625011e-02 -9.45149004e-01 -9.92406011e-01 -5.62150702e-02
-2.97701627e-01 -1.14276648e+00 -1.97081968e-01 -1.88760567e+00
3.11941206e-01 -5.43722391e-01 -3.61895740e-01 6.31863415e-01
-1.29762813e-01 7.13980079e-01 -4.72591281e-01 3.86329353e-01
5.80947362e-02 -2.29059041e-01 8.86036634e-01 -2.57353097e-01
-3.50510359e-01 1.60627857e-01 -8.68976176e-01 1.07479215e+00
7.65269041e-01 -3.98082256e-01 1.53516128e-03 2.63684154e-01
1.94712579e-01 -1.91516533e-01 4.18137878e-01 -8.36239696e-01
2.29284927e-01 -1.17644303e-01 4.23014492e-01 -5.07207692e-01
-6.88472064e-03 -8.91924500e-01 6.02116704e-01 1.10588217e+00
-1.00963295e-01 -5.12817025e-01 1.15439825e-01 4.88024592e-01
-1.68130666e-01 -4.90665942e-01 8.21519732e-01 -1.09519005e-01
-6.04720950e-01 -1.21636435e-01 -9.17796075e-01 -6.89609408e-01
1.18857002e+00 -1.00095832e+00 -5.09510279e-01 3.48998122e-02
-9.51556206e-01 -5.15195392e-02 9.02633905e-01 -1.54217780e-01
5.39751016e-02 -8.81728888e-01 -6.90983534e-01 4.04184423e-02
3.37484509e-01 1.72556490e-01 1.91274583e-01 1.18680382e+00
-1.26709378e+00 3.68898004e-01 1.54813468e-01 -7.07094729e-01
-1.47951913e+00 8.59011989e-03 7.71105409e-01 -7.13309228e-01
-3.21040362e-01 8.32572758e-01 -1.51990782e-02 -5.24334908e-01
3.62080429e-03 -4.39710706e-01 -5.38745999e-01 3.35784703e-01
5.32577217e-01 4.41152453e-01 4.40114945e-01 -7.00320542e-01
-1.85224980e-01 6.90293312e-01 -3.78328174e-01 6.16786517e-02
1.29916883e+00 -9.16704386e-02 -3.47743303e-01 2.58734971e-01
1.49652541e+00 8.99930894e-02 -4.66635853e-01 1.98399812e-01
2.10185766e-01 1.44424096e-01 2.05790460e-01 -1.12752187e+00
-7.30991066e-01 2.28070825e-01 1.14574051e+00 4.20620590e-01
1.10288858e+00 1.76714629e-01 6.79850101e-01 -3.63184661e-02
1.44505888e-01 -9.38163400e-01 -2.89756298e-01 -8.52080621e-03
5.00981510e-01 -1.24523056e+00 -1.27292769e-02 -4.38679785e-01
5.71728777e-03 1.34177649e+00 2.05755532e-01 1.35512948e-01
7.97288597e-01 3.50193888e-01 6.24358535e-01 -2.81944782e-01
-6.55420780e-01 1.46769077e-01 2.08436072e-01 9.23935175e-01
4.75063652e-01 2.96365768e-01 -6.98166609e-01 6.29651845e-01
-5.25760829e-01 3.36245626e-01 5.39130270e-01 7.42509484e-01
-4.68571752e-01 -1.11584520e+00 -3.51570368e-01 8.98734868e-01
-7.20535994e-01 3.68704796e-01 -4.02056962e-01 1.29063511e+00
3.38949889e-01 7.24142730e-01 9.57626477e-02 -4.56192084e-02
8.32566470e-02 -4.76713218e-02 -2.28628945e-02 -3.13077301e-01
-7.22582936e-01 2.43320584e-01 3.75661671e-01 1.62800938e-01
-5.67475617e-01 -9.49219108e-01 -1.52179718e+00 3.09302695e-02
-4.28338975e-01 1.26355931e-01 6.67359054e-01 8.17548752e-01
-1.34980723e-01 3.59940261e-01 2.61431217e-01 -4.95556384e-01
-1.12649627e-01 -1.03757179e+00 -6.87158048e-01 2.54326731e-01
4.82024699e-01 -2.41162643e-01 -1.02391839e+00 3.64765346e-01] | [15.197936058044434, -3.1076037883758545] |
381246e3-581b-4c5e-a914-71fb579c6281 | reconstructing-a-large-scale-3d-face-dataset | 2010.08391 | null | https://arxiv.org/abs/2010.08391v2 | https://arxiv.org/pdf/2010.08391v2.pdf | Reconstructing A Large Scale 3D Face Dataset for Deep 3D Face Identification | Deep learning methods have brought many breakthroughs to computer vision, especially in 2D face recognition. However, the bottleneck of deep learning based 3D face recognition is that it is difficult to collect millions of 3D faces, whether for industry or academia. In view of this situation, there are many methods to generate more 3D faces from existing 3D faces through 3D face data augmentation, which are used to train deep 3D face recognition models. However, to the best of our knowledge, there is no method to generate 3D faces from 2D face images for training deep 3D face recognition models. This letter focuses on the role of reconstructed 3D facial surfaces in 3D face identification and proposes a framework of 2D-aided deep 3D face identification. In particular, we propose to reconstruct millions of 3D face scans from a large scale 2D face database (i.e.VGGFace2), using a deep learning based 3D face reconstruction method (i.e.ExpNet). Then, we adopt a two-phase training approach: In the first phase, we use millions of face images to pre-train the deep convolutional neural network (DCNN), and in the second phase, we use normal component images (NCI) of reconstructed 3D face scans to train the DCNN. Extensive experimental results illustrate that the proposed approach can greatly improve the rank-1 score of 3D face identification on the FRGC v2.0, the Bosphorus, and the BU-3DFE 3D face databases, compared to the model trained by 2D face images. Finally, our proposed approach achieves state-of-the-art rank-1 scores on the FRGC v2.0 (97.6%), Bosphorus (98.4%), and BU-3DFE (98.8%) databases. The experimental results show that the reconstructed 3D facial surfaces are useful and our 2D-aided deep 3D face identification framework is meaningful, facing the scarcity of 3D faces. | ['Huibin Li', 'Zihui Zhang', 'Cuican Yu'] | 2020-10-16 | null | null | null | null | ['face-reconstruction'] | ['computer-vision'] | [-3.17571610e-01 -6.25044107e-02 1.88652039e-01 -5.27586460e-01
-4.58969414e-01 -3.68910760e-01 3.89196873e-01 -1.06090593e+00
3.52178607e-03 1.71425417e-01 -2.51159370e-01 -3.42837095e-01
1.61384940e-01 -9.06984627e-01 -6.75128579e-01 -8.55526447e-01
4.80783135e-02 6.31526411e-01 -5.51192701e-01 -9.27947164e-02
7.36315688e-03 1.32532442e+00 -1.98226511e+00 1.24313317e-01
2.43601650e-01 1.51694214e+00 -3.63262057e-01 1.94698945e-01
-4.03190315e-01 -9.17684063e-02 -6.25430584e-01 -6.14711761e-01
7.06025541e-01 -2.44779781e-01 -3.59085739e-01 2.37956449e-01
7.83155501e-01 -8.07065964e-01 -3.00050229e-01 9.91744697e-01
9.26911950e-01 -3.88084054e-01 8.96206200e-01 -1.36435366e+00
-7.02291310e-01 -3.51942964e-02 -8.59390318e-01 -1.67633384e-01
2.01222405e-01 1.04901560e-01 9.86657217e-02 -1.56555295e+00
2.76388615e-01 1.84102333e+00 8.97625983e-01 1.18580854e+00
-7.66877294e-01 -1.25272644e+00 -2.06076831e-01 -2.39287734e-01
-1.67552114e+00 -9.22290981e-01 9.76010442e-01 -3.93096268e-01
8.29460025e-01 -9.69668701e-02 4.97490495e-01 1.12324834e+00
-2.99087614e-01 5.64514935e-01 1.00412273e+00 -3.04155380e-01
-1.31591856e-01 -1.03126399e-01 -7.91740119e-02 1.02871418e+00
3.13120544e-01 3.20117384e-01 -4.80230957e-01 -5.98653927e-02
1.16053414e+00 -2.17414983e-02 -2.20452454e-02 9.13530663e-02
-4.37889069e-01 6.02752447e-01 2.19175220e-01 4.28104550e-02
-3.96158099e-01 -2.90782541e-01 2.74293572e-01 3.90392691e-01
7.68100917e-01 -2.38619223e-01 -3.08955878e-01 3.69822472e-01
-5.71223497e-01 6.30877241e-02 7.52402961e-01 9.39426124e-01
9.85367119e-01 3.20574194e-01 1.44873023e-01 1.13025188e+00
5.72604656e-01 9.21603024e-01 3.07775825e-01 -5.93460798e-01
1.88243017e-01 7.94837415e-01 -2.38701850e-01 -9.24014270e-01
-1.03669107e-01 -3.64393257e-02 -1.16020560e+00 4.18755502e-01
1.69270232e-01 -2.96039134e-02 -1.16033792e+00 1.56433952e+00
7.20752001e-01 2.03828394e-01 9.80788097e-02 9.77204442e-01
1.48252106e+00 5.13862312e-01 -3.18979353e-01 -1.71722710e-01
1.20432186e+00 -4.68315542e-01 -3.11332315e-01 2.12872267e-01
2.56578445e-01 -8.18505406e-01 6.89988315e-01 1.19275369e-01
-9.38094556e-01 -9.69207108e-01 -7.73284853e-01 1.18657507e-01
-2.61420667e-01 4.03650403e-01 4.11195308e-01 1.25711405e+00
-1.22719383e+00 7.77909830e-02 -4.14250910e-01 -1.82424873e-01
1.08374166e+00 8.40603352e-01 -9.84628022e-01 -3.39677423e-01
-9.50000703e-01 5.82857549e-01 2.64309980e-02 3.42309326e-01
-1.31938767e+00 -5.77547848e-01 -7.93154895e-01 -1.48050025e-01
1.01541365e-02 -4.41621810e-01 9.07319844e-01 -7.17542350e-01
-1.45247138e+00 1.64561856e+00 -2.46385396e-01 1.50335506e-01
2.84795851e-01 4.04839218e-02 -4.07501042e-01 5.56082930e-03
-5.02187237e-02 7.47173727e-01 1.11740255e+00 -1.36440670e+00
-3.03091884e-01 -1.01684558e+00 -1.54588863e-01 -7.74818212e-02
-5.19018471e-01 1.20574497e-01 -7.45502353e-01 -2.46812478e-01
1.70147792e-01 -7.40957022e-01 3.51374924e-01 3.03302616e-01
-3.14750731e-01 -6.31014884e-01 1.17886794e+00 -5.56301594e-01
3.75295490e-01 -2.26956129e+00 -2.06055567e-01 2.99582869e-01
1.86087817e-01 7.48062313e-01 -5.83355427e-01 -2.03179419e-01
-2.37857550e-01 3.42900485e-01 -2.17164215e-02 -7.34109163e-01
-2.30552517e-02 5.06982282e-02 -8.19912255e-02 5.60264230e-01
5.64598203e-01 7.98377454e-01 -2.76148081e-01 -2.59775192e-01
2.13915408e-01 9.38166142e-01 -4.59882617e-01 5.21822631e-01
1.56559721e-01 3.95352960e-01 -5.21083593e-01 1.12792826e+00
1.48357153e+00 9.30166692e-02 -2.38568988e-02 -4.30796981e-01
1.22962385e-01 -1.23177305e-01 -9.97647882e-01 1.30176842e+00
-2.43359581e-01 1.04290619e-01 3.60292882e-01 -1.07583153e+00
1.56473446e+00 4.46954757e-01 5.17760873e-01 -6.74389362e-01
3.28247249e-01 4.45041806e-01 -3.16871852e-01 -3.65031868e-01
-2.92768151e-01 -3.00539553e-01 2.36058384e-01 6.08143032e-01
3.46521407e-01 -8.80378261e-02 -3.09462726e-01 -4.70335633e-01
4.87841100e-01 -9.76878032e-02 -1.32046506e-01 -1.40934959e-01
8.84759009e-01 -4.80360895e-01 5.50395966e-01 2.73385018e-01
-1.49309754e-01 5.44280052e-01 4.73406523e-01 -8.46694529e-01
-8.86927545e-01 -9.41334486e-01 -2.01207206e-01 4.20071125e-01
-7.25825727e-02 9.07557607e-02 -9.61594880e-01 -9.24284935e-01
4.29384172e-01 -1.88671872e-01 -5.80445051e-01 -2.00561553e-01
-7.29661226e-01 -7.52684832e-01 1.08941960e+00 2.33282834e-01
8.63467097e-01 -8.83272827e-01 1.38382226e-01 -2.66941190e-01
3.36807370e-01 -9.42519307e-01 -4.34705883e-01 -4.19034362e-01
-8.89292955e-01 -1.34107947e+00 -7.95133650e-01 -1.25162160e+00
1.05385029e+00 4.02266532e-01 9.15012240e-01 4.68755603e-01
-2.67605335e-01 2.86407351e-01 -1.56178489e-01 -4.73300099e-01
-5.40843487e-01 -3.16542804e-01 8.07635128e-01 2.95599043e-01
8.90496612e-01 -4.07048076e-01 -4.19083983e-01 6.98463142e-01
-6.44588768e-01 -2.83399701e-01 6.21538699e-01 9.12709177e-01
6.68154895e-01 2.10621431e-01 8.88979018e-01 -6.35127842e-01
4.29450929e-01 -2.59594709e-01 -8.17847192e-01 1.60796463e-01
-4.36727226e-01 -3.04569334e-01 5.93547940e-01 -3.91660839e-01
-9.83082354e-01 2.50238270e-01 -7.74376035e-01 -1.03741312e+00
-4.58474755e-01 4.15747650e-02 -7.45954990e-01 -6.00723386e-01
4.62473601e-01 2.61332899e-01 6.12003148e-01 -8.02709818e-01
-3.12633254e-02 1.17760181e+00 3.51071239e-01 -5.44821501e-01
1.02542102e+00 3.33151877e-01 1.25498995e-01 -9.75383759e-01
-5.51042736e-01 -1.10058516e-01 -6.54011786e-01 -3.51715267e-01
6.36825383e-01 -1.16420615e+00 -1.13594031e+00 1.15561187e+00
-1.30881488e+00 9.34333056e-02 1.88603327e-01 2.87629515e-01
-2.24262312e-01 2.37114951e-01 -5.16302645e-01 -1.07368767e+00
-7.07215965e-01 -1.40033889e+00 1.53099668e+00 2.61797875e-01
3.87113333e-01 -4.54306662e-01 -3.61094475e-01 5.81241310e-01
2.69113243e-01 2.08217815e-01 8.73917699e-01 -4.26332980e-01
-4.34055835e-01 -4.29328293e-01 -4.61337805e-01 7.45385528e-01
5.70979297e-01 -1.62938878e-01 -1.50354385e+00 -3.24930310e-01
2.17586175e-01 -4.84891504e-01 6.21962905e-01 1.74534127e-01
1.31312788e+00 -2.65493542e-01 -1.46867380e-01 9.39652026e-01
1.00667560e+00 3.75061065e-01 7.05328882e-01 -4.50544447e-01
7.41421878e-01 8.13248873e-01 2.84201503e-01 3.98061156e-01
2.09363803e-01 5.02420902e-01 4.61073250e-01 -7.57408217e-02
-3.94364834e-01 -3.47285062e-01 3.56829017e-01 7.66955078e-01
-2.47460812e-01 1.08958498e-01 -8.24487805e-01 1.32945016e-01
-9.86250460e-01 -7.31564045e-01 2.51488179e-01 2.15332413e+00
6.56853914e-01 -1.95591047e-01 5.26695848e-02 2.72142142e-01
1.02435219e+00 -1.47449806e-01 -8.48853350e-01 -9.36542917e-03
-1.37008771e-01 5.27576685e-01 -4.83509116e-02 1.89487234e-01
-1.05966401e+00 9.53248262e-01 5.21439028e+00 7.94266045e-01
-1.38857400e+00 -8.97438601e-02 1.03919911e+00 3.15552403e-04
3.58287394e-02 -6.85251772e-01 -1.34384823e+00 4.26650673e-01
5.38983345e-01 1.97855964e-01 4.35507357e-01 8.86747837e-01
-3.85433957e-02 6.49755895e-01 -1.25199795e+00 1.63096285e+00
3.77370685e-01 -1.31341612e+00 5.10598421e-01 3.34608138e-01
5.76632679e-01 -3.79105538e-01 2.26484984e-01 3.27632844e-01
-2.83627480e-01 -1.49454498e+00 3.62753630e-01 2.80382574e-01
1.59076309e+00 -9.63625252e-01 7.57528722e-01 3.46173406e-01
-1.18714094e+00 9.06117484e-02 -5.49199641e-01 3.43873024e-01
-2.33089417e-01 6.54568851e-01 -7.64165699e-01 3.59206676e-01
9.75849748e-01 6.34205282e-01 -1.41236424e-01 6.32740617e-01
1.37787998e-01 2.47695357e-01 -3.38696748e-01 6.57942370e-02
-1.17046967e-01 -1.10347062e-01 2.44477302e-01 6.29201651e-01
6.64411366e-01 3.69672924e-01 -9.81592089e-02 9.45630074e-01
-8.10287714e-01 -1.23177692e-01 -1.00033033e+00 -1.37735724e-01
5.29477239e-01 1.26267374e+00 -2.61651754e-01 1.56309269e-02
-3.70130092e-01 7.17453122e-01 8.65869895e-02 6.01385869e-02
-4.27358359e-01 -1.28830582e-01 1.03703940e+00 -1.95171386e-02
2.00118721e-01 -1.26718143e-02 1.36943132e-01 -9.69688118e-01
1.28670484e-01 -1.11341941e+00 1.38701275e-01 -4.73363042e-01
-1.78583574e+00 8.48739028e-01 -3.75384867e-01 -9.57382619e-01
-8.90783444e-02 -1.18253541e+00 -4.59674120e-01 1.26690483e+00
-1.59276569e+00 -1.29923713e+00 -5.64083934e-01 8.07982445e-01
2.83088058e-01 -7.53410995e-01 1.01837873e+00 7.34781682e-01
-5.96036971e-01 1.04253113e+00 -2.58783162e-01 5.81926525e-01
5.59464455e-01 -4.30650234e-01 8.06937218e-01 2.57468909e-01
-1.86036248e-02 5.21463692e-01 -2.91190505e-01 -5.44871688e-01
-2.14249969e+00 -1.21176064e+00 5.70236742e-01 -3.85410130e-01
-2.42338315e-01 -5.62792420e-01 -7.45092452e-01 2.45594174e-01
-5.48436642e-01 2.80866981e-01 7.65112102e-01 -1.36156946e-01
-6.02342129e-01 -4.69872564e-01 -1.70150840e+00 2.31114388e-01
1.47254205e+00 -7.30619907e-01 -2.37518206e-01 2.46375397e-01
3.88226300e-01 -2.37283453e-01 -9.77395892e-01 7.60377407e-01
8.68082702e-01 -8.39433372e-01 1.24117720e+00 -4.79525805e-01
5.47212809e-02 -2.93089479e-01 -3.15947831e-01 -9.16626096e-01
1.64983626e-02 -2.91662633e-01 -1.75656900e-01 1.24981701e+00
8.37176591e-02 -7.35279322e-01 1.14090049e+00 3.15947503e-01
-1.20840617e-01 -6.92071676e-01 -1.22270560e+00 -7.48075068e-01
1.38354450e-01 -1.36376232e-01 1.26285553e+00 9.61336076e-01
-1.00495446e+00 1.68313056e-01 -1.87765986e-01 2.44350493e-01
7.91538000e-01 1.68134734e-01 1.02912271e+00 -1.65268874e+00
3.42825145e-01 -4.64482576e-01 -5.12745082e-01 -1.08126938e+00
6.62528038e-01 -9.69227672e-01 -2.66695738e-01 -9.56158459e-01
1.35059550e-01 -7.62062848e-01 -5.40236793e-02 6.03560805e-01
1.81251034e-01 6.52560949e-01 -8.74231905e-02 1.99374676e-01
1.72174901e-01 7.55778670e-01 1.52122128e+00 -3.28434378e-01
5.71485534e-02 -4.38795611e-02 -7.19954908e-01 4.85968113e-01
6.01377487e-01 -1.55592084e-01 -2.00960040e-01 -6.51868880e-01
-5.06212056e-01 7.07591511e-03 3.97659749e-01 -7.42419064e-01
2.03290023e-02 1.42051533e-01 1.03063500e+00 -7.54924476e-01
6.50329828e-01 -8.00915420e-01 1.61581829e-01 3.03459287e-01
3.58346134e-01 -1.98852390e-01 4.77149755e-01 8.45876187e-02
-2.40336761e-01 -9.43482760e-03 9.43868995e-01 -2.73768932e-01
-5.55427909e-01 1.07060540e+00 6.64263740e-02 -3.65143210e-01
8.96010399e-01 -4.58791435e-01 -2.44616598e-01 -1.85305681e-02
-3.22136939e-01 -8.10340941e-02 4.11267638e-01 5.15365541e-01
1.10830343e+00 -1.78310788e+00 -9.81851459e-01 1.16204035e+00
1.27354302e-02 2.69651651e-01 2.97816992e-01 2.24329889e-01
-4.77390856e-01 3.05531293e-01 -5.51636994e-01 -7.34926641e-01
-1.37959790e+00 2.72107154e-01 6.13757133e-01 4.29811746e-01
-2.22975314e-01 1.16319823e+00 3.85457039e-01 -9.28771377e-01
2.81897724e-01 4.42231685e-01 -2.99718827e-01 8.31147432e-02
6.75610960e-01 -1.35484776e-02 2.94743091e-01 -1.11555278e+00
-5.57086825e-01 1.06883204e+00 -2.73795836e-02 3.89462054e-01
1.46163762e+00 3.77123773e-01 -4.81238455e-01 -2.51628190e-01
1.60390306e+00 -3.11005503e-01 -1.02918029e+00 7.57725835e-02
-5.10318518e-01 -7.05699921e-01 -5.90033159e-02 -5.01087904e-01
-1.68042254e+00 1.14209008e+00 9.72567081e-01 -9.13925245e-02
1.28929341e+00 -3.17101814e-02 8.24378490e-01 5.20509422e-01
5.93813241e-01 -5.48204303e-01 1.18299589e-01 5.72147787e-01
9.47760522e-01 -1.37611234e+00 -4.07840043e-01 -4.47629720e-01
2.06219181e-02 1.25128829e+00 9.59344149e-01 1.40182421e-01
9.13050413e-01 1.92122653e-01 1.32429004e-01 -3.26933205e-01
-3.43822628e-01 6.96602017e-02 2.61339515e-01 7.11961627e-01
2.95553416e-01 -1.35894850e-01 4.05333042e-01 6.26106620e-01
-4.18364853e-02 -5.37448041e-02 -6.81295171e-02 5.39407611e-01
-1.20618343e-01 -1.22023189e+00 -6.29615605e-01 5.10197520e-01
-2.92323083e-01 1.15237191e-01 -6.05410099e-01 6.62051857e-01
2.38145307e-01 8.14910948e-01 1.74128965e-01 -7.58370996e-01
3.81082237e-01 2.13149965e-01 7.32373953e-01 -5.74004352e-01
-2.22823769e-01 -1.24985181e-01 -2.01934054e-01 -3.37196946e-01
-3.62311572e-01 -2.09630176e-01 -9.36675787e-01 -7.30412662e-01
-3.75150323e-01 5.79774268e-02 9.07421410e-01 5.78817964e-01
6.46956325e-01 -2.23069698e-01 1.22256351e+00 -1.16792917e+00
-5.12249947e-01 -8.77865613e-01 -6.87940478e-01 2.87177682e-01
3.50300163e-01 -9.92779016e-01 -4.69366729e-01 -5.70436977e-02] | [13.265971183776855, 0.48118194937705994] |
7c3189f0-7fed-4710-9ec9-ea83c0b45ad5 | 3d-facial-imperfection-regeneration-deep | 2303.14381 | null | https://arxiv.org/abs/2303.14381v1 | https://arxiv.org/pdf/2303.14381v1.pdf | 3D Facial Imperfection Regeneration: Deep learning approach and 3D printing prototypes | This study explores the potential of a fully convolutional mesh autoencoder model for regenerating 3D nature faces with the presence of imperfect areas. We utilize deep learning approaches in graph processing and analysis to investigate the capabilities model in recreating a filling part for facial scars. Our approach in dataset creation is able to generate a facial scar rationally in a virtual space that corresponds to the unique circumstances. Especially, we propose a new method which is named 3D Facial Imperfection Regeneration(3D-FaIR) for reproducing a complete face reconstruction based on the remaining features of the patient face. To further enhance the applicable capacity of the present research, we develop an improved outlier technique to separate the wounds of patients and provide appropriate wound cover models. Also, a Cir3D-FaIR dataset of imperfect faces and open codes was released at https://github.com/SIMOGroup/3DFaIR. Our findings demonstrate the potential of the proposed approach to help patients recover more quickly and safely through convenient techniques. We hope that this research can contribute to the development of new products and innovative solutions for facial scar regeneration. | ['H. Nguyen-Xuan', 'Li-Wei Chou', 'Thanh Q. Nguyen', 'Duong Q. Nguyen', 'Thinh D. Le', 'Phuong D. Nguyen'] | 2023-03-25 | null | null | null | null | ['face-reconstruction'] | ['computer-vision'] | [-1.01981036e-01 4.07773882e-01 -2.18906021e-03 4.24282029e-02
-1.11035906e-01 9.86846834e-02 -1.40572395e-02 -4.88532335e-01
2.18923539e-01 6.75484598e-01 4.97204661e-01 -2.18066871e-01
-1.08664386e-01 -1.19154382e+00 -6.77768290e-01 -6.14051104e-01
-5.40701253e-03 2.50850260e-01 -5.45010924e-01 -4.78652596e-01
1.03650734e-01 1.16051602e+00 -1.93184674e+00 4.77047324e-01
6.49570048e-01 1.00993872e+00 9.11136121e-02 2.56043553e-01
1.12431450e-02 1.84414580e-01 -3.33214104e-01 -5.41031182e-01
6.22280836e-01 -3.04610372e-01 -6.12327099e-01 1.98195770e-01
2.55922556e-01 -7.67770171e-01 -4.85257894e-01 4.80527818e-01
1.04418898e+00 -1.72365308e-01 7.73212552e-01 -1.02440405e+00
-8.85491550e-01 7.30954409e-02 -5.28719425e-01 -3.79409760e-01
5.67642093e-01 8.54888335e-02 1.86829135e-01 -1.19279850e+00
1.06854224e+00 9.79282975e-01 9.80803907e-01 8.81585598e-01
-7.91703641e-01 -8.30938935e-01 -5.98535359e-01 1.15880512e-01
-1.46471930e+00 -5.39554954e-01 9.55810130e-01 -2.96802163e-01
5.60149372e-01 2.14692786e-01 1.17853534e+00 1.25309038e+00
5.10319531e-01 3.52694273e-01 1.22657895e+00 -3.55412364e-01
1.47923186e-01 -1.75624222e-01 -5.98919511e-01 1.04610527e+00
2.44826511e-01 6.38644755e-01 -4.83022571e-01 -5.56852594e-02
1.35777617e+00 7.98329711e-02 -3.95640016e-01 1.14106834e-01
-4.57458764e-01 7.49976993e-01 5.78078806e-01 3.53469163e-01
-7.94985592e-01 -4.30406928e-02 7.06226081e-02 2.18812451e-01
7.07045555e-01 1.58447772e-01 -1.44371137e-01 4.18929875e-01
-8.19249392e-01 7.36601651e-02 7.54439175e-01 8.53411138e-01
6.30252004e-01 4.01250809e-01 3.15209031e-02 8.96715283e-01
2.96921432e-02 1.46293417e-01 3.52747321e-01 -1.02455080e+00
-3.08910370e-01 6.58479214e-01 -3.24779272e-01 -1.00269389e+00
-3.48152906e-01 -3.38509440e-01 -9.67490017e-01 6.62184656e-01
-1.05373025e-01 -2.90377378e-01 -1.34342492e+00 1.13986528e+00
5.94399691e-01 4.30597901e-01 -1.10187225e-01 9.64808106e-01
1.28757405e+00 1.60308853e-01 -1.44080937e-01 -1.09416492e-01
1.02311671e+00 -4.75228220e-01 -7.67825782e-01 3.13398123e-01
5.38322985e-01 -6.99698389e-01 8.53311181e-01 2.82874882e-01
-1.14409673e+00 -1.83262751e-01 -8.01998556e-01 -1.92608975e-03
-2.21230283e-01 1.61743671e-01 9.35614526e-01 3.08478892e-01
-1.47338736e+00 9.13690209e-01 -2.47356489e-01 -5.33897042e-01
1.02766848e+00 3.76980931e-01 -1.01011992e+00 -2.56026894e-01
-8.40809524e-01 9.71063614e-01 -5.43845259e-02 3.35111648e-01
-9.32154536e-01 -8.67216527e-01 -7.97736943e-01 -2.76431352e-01
-1.98982134e-01 -1.08026230e+00 6.95066035e-01 -8.91151011e-01
-1.39283419e+00 1.10451245e+00 1.00779407e-01 -8.46937075e-02
5.18095613e-01 2.38886029e-01 -1.86056420e-01 1.78820834e-01
-1.62214831e-01 4.68146712e-01 7.64632761e-01 -1.65864348e+00
6.06511496e-02 -6.80279434e-01 -3.42509300e-01 2.54203323e-02
-2.07771271e-01 -1.21359127e-02 -9.32745039e-02 -7.28190064e-01
2.56918937e-01 -6.81120098e-01 -1.94012821e-01 6.40802383e-01
-2.58904606e-01 1.19611144e-01 6.56215727e-01 -1.23866284e+00
8.76910388e-01 -1.83729029e+00 1.22854069e-01 3.22220385e-01
3.12001348e-01 3.20224851e-01 -4.70632434e-01 7.85311580e-01
-3.08602244e-01 5.68258762e-01 -2.06302166e-01 -4.86497581e-01
-6.87832355e-01 3.70887280e-01 3.06433022e-01 5.20512760e-01
1.99531704e-01 8.71865034e-01 -2.41695151e-01 -5.27509511e-01
-5.80845773e-02 8.81436527e-01 -5.95748603e-01 1.78588971e-01
1.49137452e-01 3.90940160e-01 -1.86086431e-01 1.33224106e+00
1.11250710e+00 4.03599024e-01 -2.71085590e-01 -1.85449988e-01
1.27237231e-01 -5.99422872e-01 -9.11174417e-01 1.76836538e+00
-5.08264959e-01 1.23587273e-01 5.28210640e-01 -5.22957742e-01
9.90862548e-01 5.03547251e-01 6.76787496e-01 -6.27604127e-01
6.72246993e-01 2.79887497e-01 -4.23924178e-01 -9.20886397e-01
1.50837719e-01 -7.24440873e-01 7.87960410e-01 2.84900755e-01
3.03392828e-01 -1.33371219e-01 -3.60276818e-01 -2.81448781e-01
8.12784910e-01 1.95397958e-01 1.04046337e-01 -2.59290524e-02
3.67990546e-02 -3.86583805e-01 3.75145435e-01 -1.03620598e-02
-2.93854028e-01 1.00448847e+00 3.43207896e-01 -6.05806351e-01
-9.78760839e-01 -9.29066002e-01 -2.87512630e-01 2.83550709e-01
-2.04164490e-01 1.02268003e-01 -8.04131210e-01 -3.84336710e-01
4.07580286e-01 4.13569868e-01 -1.08446717e+00 -3.00577849e-01
-2.56990969e-01 -3.88897717e-01 7.08636343e-01 3.64323348e-01
2.34865457e-01 -1.13209927e+00 -1.12118544e-02 6.07498549e-03
-1.58260971e-01 -4.32162732e-01 -6.41976893e-02 -2.97462791e-01
-1.01858866e+00 -1.25365567e+00 -1.01091218e+00 -1.24455547e+00
9.83599246e-01 9.18075964e-02 5.89282870e-01 8.08460057e-01
-8.04102778e-01 2.36391217e-01 -4.50862229e-01 -6.02146804e-01
-5.21848202e-01 -4.97874498e-01 2.43897513e-01 1.02788962e-01
-2.34971002e-01 -1.07756186e+00 -8.77886474e-01 5.90452254e-02
-9.80216920e-01 9.56470668e-02 6.38585925e-01 8.20049167e-01
5.90905905e-01 7.45883062e-02 7.32717514e-01 -6.07715189e-01
1.09648991e+00 -6.24821126e-01 2.40343779e-01 -1.24223657e-01
-6.59289122e-01 -2.14276731e-01 2.52037585e-01 -1.93468690e-01
-1.03328228e+00 -1.06639870e-01 -6.31634533e-01 -9.08353746e-01
-2.41754860e-01 5.59481144e-01 1.03023991e-01 -5.65200746e-01
9.33714330e-01 5.50026931e-02 8.49671781e-01 -5.27011991e-01
1.98081836e-01 8.67260158e-01 2.47873157e-01 -3.88753444e-01
7.10510194e-01 6.01898074e-01 1.10501379e-01 -7.76603222e-01
-3.86802596e-03 -2.66850859e-01 -4.48848277e-01 -9.06418502e-01
4.66005653e-01 -7.78215528e-01 -5.96653879e-01 6.42192483e-01
-1.15421283e+00 -2.62924761e-01 -3.10169458e-01 2.77403563e-01
-6.57099068e-01 1.90714672e-01 -6.66569471e-01 -8.52682471e-01
-7.40520895e-01 -7.07313716e-01 1.13992608e+00 3.27489942e-01
2.13209137e-01 -7.95492470e-01 1.34722769e-01 7.41188347e-01
4.84450817e-01 9.03975785e-01 8.95546019e-01 1.89576983e-01
-2.13269323e-01 -5.23469210e-01 -4.84049842e-02 6.31820858e-01
2.64308482e-01 2.81729817e-01 -9.32557166e-01 -2.41363555e-01
-2.46441159e-02 -1.72099635e-01 6.90590739e-01 3.00821900e-01
1.13265002e+00 -6.81174219e-01 -2.51510262e-01 7.25069880e-01
1.68032777e+00 3.73335066e-03 1.25581241e+00 -2.25446541e-02
3.72145683e-01 8.12807739e-01 3.46545607e-01 6.91245735e-01
1.05164774e-01 1.39493659e-01 8.35774362e-01 -5.16318917e-01
-8.57462764e-01 -4.71403688e-01 -8.26124325e-02 5.46095073e-01
-9.47547197e-01 -5.55309914e-02 -7.40712941e-01 6.84300840e-01
-1.10845137e+00 -9.47333813e-01 5.27941026e-02 1.84204912e+00
6.87554300e-01 -6.53085947e-01 -2.87903845e-01 3.09412926e-01
6.48468137e-01 -2.74059296e-01 -1.92881510e-01 -5.93981564e-01
-1.31704256e-01 1.02209198e+00 2.01923996e-01 3.19298953e-01
-4.67255473e-01 8.59590113e-01 6.44900370e+00 6.55826807e-01
-1.25713193e+00 -1.10729478e-01 4.49186265e-01 -2.61243619e-02
-5.62279224e-01 -5.78806177e-02 -8.20739344e-02 1.28226563e-01
6.22170568e-01 -1.60673469e-01 6.27859592e-01 7.29140341e-01
5.50867736e-01 -3.11935265e-02 -4.74017054e-01 7.89225757e-01
3.29514980e-01 -1.56998169e+00 3.64523113e-01 3.45988572e-01
5.65532625e-01 -3.87287736e-01 1.44855632e-02 -3.94468643e-02
-2.83240229e-01 -1.30450356e+00 3.57976198e-01 1.03049278e+00
1.42885518e+00 -8.15824389e-01 8.72688115e-01 2.02345744e-01
-6.28428519e-01 -2.31268346e-01 -1.47306189e-01 -1.93652228e-01
-4.04735841e-02 4.65160370e-01 -1.39940155e+00 7.19344556e-01
5.77145994e-01 3.82519215e-01 -3.93639088e-01 1.25204182e+00
6.07779175e-02 3.06495637e-01 -1.24474570e-01 2.11187631e-01
-5.92878938e-01 -2.65190154e-01 4.28928286e-01 5.89375436e-01
7.26884425e-01 2.57312417e-01 -5.03986776e-01 8.65789294e-01
-2.89725780e-01 4.93445992e-01 -1.08139360e+00 5.97971790e-02
2.05244645e-01 9.82524812e-01 -2.54990101e-01 4.68390077e-01
-1.16179451e-01 9.72649634e-01 2.21956000e-01 3.82980317e-01
-4.87832785e-01 -9.24887881e-03 5.81663311e-01 8.60471189e-01
-1.03963971e-01 -2.50776224e-02 -5.84137261e-01 -7.23641455e-01
8.64922702e-02 -6.60673320e-01 1.05738811e-01 -1.29548216e+00
-1.22391927e+00 5.71293712e-01 -4.48381662e-01 -1.16101301e+00
2.22442418e-01 -5.08897185e-01 -7.87751079e-01 8.17473531e-01
-1.41967833e+00 -1.80740786e+00 -6.76556289e-01 7.67147481e-01
2.56640315e-01 -2.16449752e-01 1.28324044e+00 1.65500611e-01
-5.02910078e-01 5.49452662e-01 -2.22205877e-01 3.36533599e-02
4.89031732e-01 -4.62968767e-01 -2.30221942e-01 6.40974820e-01
-2.55819112e-01 3.56970549e-01 4.89880681e-01 -1.04366016e+00
-1.42318380e+00 -1.14436245e+00 6.66194975e-01 1.48025811e-01
-1.97395198e-02 2.72889482e-03 -5.47087729e-01 4.96599257e-01
1.70758605e-01 -1.44009948e-01 8.89500737e-01 -9.66627598e-02
-5.76804169e-02 -1.87345594e-02 -1.79602253e+00 7.92371452e-01
1.38309538e+00 -3.27802539e-01 -1.58720538e-01 5.03675044e-01
4.36449856e-01 -4.79612947e-01 -1.23238218e+00 7.34664381e-01
6.91099346e-01 -9.44847941e-01 6.29616082e-01 -5.88784158e-01
8.64690840e-01 1.29229408e-02 1.60640255e-02 -1.39359248e+00
-2.75741279e-01 -5.35091460e-01 -2.70760376e-02 8.06499779e-01
2.78734088e-01 -6.99126184e-01 1.01533008e+00 5.18585861e-01
-4.83671099e-01 -1.32419503e+00 -1.25812042e+00 -4.74966168e-01
2.53525525e-01 -3.43022734e-01 7.38298953e-01 8.13657701e-01
-2.62865782e-01 -6.34513497e-01 -2.76403219e-01 4.96499799e-02
3.00564557e-01 -2.06385761e-01 5.88309526e-01 -1.33650327e+00
3.36319655e-01 -1.80877715e-01 -6.36341989e-01 1.85816437e-01
2.18978912e-01 -1.16357303e+00 -3.74256581e-01 -1.91691732e+00
-9.34384093e-02 -4.89046246e-01 3.53897028e-02 7.86929965e-01
3.77937645e-01 3.56492311e-01 -1.42956212e-01 1.13348342e-01
7.94803202e-01 8.52493823e-01 2.04565287e+00 5.34114316e-02
-1.56418234e-01 1.63453836e-02 -1.16132784e+00 3.16257566e-01
1.01096797e+00 -2.17002258e-01 -1.98790193e-01 -2.93348730e-01
-2.77030349e-01 4.43631232e-01 4.51554567e-01 -9.81630445e-01
1.21802129e-01 -3.20339292e-01 5.29474080e-01 -1.14541039e-01
3.60247135e-01 -8.43125582e-01 6.33964658e-01 1.65860951e-01
2.88327575e-01 -4.04956967e-01 4.64856207e-01 4.69764769e-01
-4.90624495e-02 -9.22441483e-03 8.04556668e-01 -2.53301740e-01
-3.11312288e-01 7.62141049e-01 -1.69474721e-01 -6.94304049e-01
1.39462972e+00 -8.19264352e-01 -1.33842379e-01 -4.78345633e-01
-9.18302894e-01 -2.17457056e-01 8.25505376e-01 5.11266112e-01
1.32400680e+00 -1.53672755e+00 -1.07089067e+00 6.83835447e-01
4.12904024e-02 -4.18242067e-02 7.59345829e-01 6.37379050e-01
-1.15685642e+00 -1.88284576e-01 -8.83119404e-01 -6.95232907e-03
-1.21944845e+00 3.78165036e-01 6.41287923e-01 4.52930152e-01
-7.60065675e-01 8.56096923e-01 -3.34081739e-01 -3.98278147e-01
1.20281719e-01 1.99101657e-01 -2.39882112e-01 -2.02034209e-02
1.69916749e-01 3.70158136e-01 3.87405068e-01 -6.82633817e-01
2.23489776e-02 6.14882231e-01 2.00079963e-01 2.34149903e-01
1.57820117e+00 2.71313280e-01 -3.60278100e-01 -4.48799103e-01
1.04004049e+00 7.95779303e-02 -8.67647350e-01 4.59812403e-01
-7.56819010e-01 -4.85774666e-01 2.64770448e-01 -8.80481303e-01
-1.72733057e+00 2.97127098e-01 8.02214503e-01 -4.18059975e-01
1.52291119e+00 -1.66849047e-01 8.24367166e-01 -3.79315615e-01
6.37786984e-01 -9.66367066e-01 -4.98420931e-02 -3.83890234e-02
1.83223951e+00 -7.76046336e-01 -5.54212332e-02 -6.30151749e-01
-4.23896313e-01 1.19545317e+00 4.56902981e-01 -2.45224103e-01
1.00130343e+00 3.86349440e-01 1.62487134e-01 -4.43961769e-01
-4.27859932e-01 -2.26546392e-01 -1.01493515e-01 1.13484228e+00
2.50613540e-01 1.50427774e-01 -6.62883639e-01 5.93219042e-01
-3.31409276e-01 5.45253336e-01 7.23087966e-01 7.87966073e-01
-1.69852953e-02 -1.09734082e+00 -1.46058485e-01 7.54628360e-01
-7.78554901e-02 7.63221309e-02 -5.86132169e-01 8.85711551e-01
4.56176013e-01 8.21155250e-01 -3.23820293e-01 -7.67384231e-01
5.64146936e-01 1.53158814e-01 5.16099215e-01 -4.65382457e-01
-4.57157403e-01 -2.77446210e-01 1.70922354e-01 -6.72584593e-01
-1.31787181e-01 -3.44517857e-01 -1.09881592e+00 -6.08954906e-01
-3.48853856e-01 -4.30700988e-01 1.09933293e+00 3.71248186e-01
7.98279345e-01 3.58556032e-01 7.74517536e-01 -9.70615208e-01
-1.61979407e-01 -8.98107529e-01 -1.00121701e+00 1.37346759e-01
1.10630713e-01 -9.00399268e-01 -5.20233333e-01 -1.02113716e-01] | [12.803813934326172, -0.024428170174360275] |
070d61d6-4789-4959-8521-10e6cbb9fedb | global-optimality-of-elman-type-rnn-in-the | 2303.06726 | null | https://arxiv.org/abs/2303.06726v1 | https://arxiv.org/pdf/2303.06726v1.pdf | Global Optimality of Elman-type RNN in the Mean-Field Regime | We analyze Elman-type Recurrent Reural Networks (RNNs) and their training in the mean-field regime. Specifically, we show convergence of gradient descent training dynamics of the RNN to the corresponding mean-field formulation in the large width limit. We also show that the fixed points of the limiting infinite-width dynamics are globally optimal, under some assumptions on the initialization of the weights. Our results establish optimality for feature-learning with wide RNNs in the mean-field regime | ['Sayan Mukherjee', 'Jianfeng Lu', 'Andrea Agazzi'] | 2023-03-12 | null | null | null | null | ['type'] | ['speech'] | [-1.64005741e-01 2.26733550e-01 -1.51033342e-01 -5.73904738e-02
-3.19398463e-01 -4.86229241e-01 3.79084200e-01 -4.75989282e-01
-6.39284849e-01 7.11781025e-01 5.79281271e-01 -3.31900328e-01
-6.89864814e-01 -6.18348181e-01 -8.35608602e-01 -9.86831486e-01
-1.23031046e-02 3.19529980e-01 -9.76107642e-02 -6.08804405e-01
1.45765841e-01 7.45934665e-01 -1.03961086e+00 -3.55850309e-01
4.16014135e-01 7.63212442e-01 2.56043553e-01 1.00337708e+00
-8.03199038e-02 8.01453590e-01 -4.59717065e-01 3.51089865e-01
3.91695499e-01 -2.40338877e-01 -7.63765752e-01 -1.58169895e-01
5.08012474e-01 -1.14572719e-02 -9.81623769e-01 1.04194736e+00
5.92999220e-01 8.38612258e-01 6.78630471e-01 -4.30955738e-01
-8.72151256e-01 8.89438331e-01 -6.34781718e-01 7.94785321e-01
-5.59262812e-01 -2.10031301e-01 1.07531989e+00 -8.94268453e-01
7.30986059e-01 9.76537049e-01 8.84638727e-01 1.01711047e+00
-1.19212449e+00 -5.16996324e-01 4.87828493e-01 -4.31375444e-01
-1.23034704e+00 -5.66396296e-01 6.37960136e-01 -3.07299644e-01
1.38017786e+00 -4.53399748e-01 5.25917828e-01 9.01179910e-01
1.22829504e-01 5.17959893e-01 4.56497222e-01 -5.84054112e-01
1.51538119e-01 -3.14451367e-01 5.04385233e-01 8.18043113e-01
2.58441657e-01 2.37639651e-01 -4.21354353e-01 -2.73543801e-02
1.30642641e+00 1.14985861e-01 -3.91607076e-01 -1.83393389e-01
-9.05743659e-01 1.03750408e+00 6.00532234e-01 6.38546705e-01
-3.78148317e-01 6.38070881e-01 1.39505982e-01 5.06731689e-01
6.10980272e-01 3.59566122e-01 -2.94188887e-01 1.19693264e-01
-9.18761134e-01 2.62333155e-01 5.28162122e-01 1.07293069e+00
4.98465329e-01 5.96193314e-01 3.89745310e-02 8.72088015e-01
1.07000180e-01 8.76239181e-01 7.23568797e-01 -8.94857824e-01
5.88616371e-01 6.69107214e-02 1.79853350e-01 -2.36997694e-01
-7.88019598e-01 -9.06644285e-01 -9.96274292e-01 -1.47995234e-01
5.91468632e-01 -8.44125450e-01 -8.16178739e-01 2.01247358e+00
-2.20722660e-01 1.26906022e-01 3.38335603e-01 7.56722510e-01
3.92654717e-01 8.31961215e-01 -3.06154490e-01 -4.81902778e-01
5.63416898e-01 -7.02213883e-01 -7.00744092e-01 -2.36585528e-01
7.73616195e-01 -2.20152542e-01 7.94044197e-01 -1.92598235e-02
-1.13766813e+00 -2.59839982e-01 -1.00638521e+00 -4.72560078e-02
-5.81527688e-02 3.65963914e-02 2.61744231e-01 1.66934252e-01
-1.31205082e+00 8.95258605e-01 -5.96443653e-01 -1.35036051e-01
1.57876000e-01 5.52024305e-01 -1.58958688e-01 5.26918769e-01
-1.22059953e+00 6.66895568e-01 2.99265146e-01 5.11634648e-01
-8.39519143e-01 -8.46130013e-01 -5.19465804e-01 4.41958345e-02
-2.20923349e-02 -4.90257889e-01 1.32762349e+00 -7.73881316e-01
-1.49080932e+00 3.47838819e-01 -2.53558308e-01 -6.48609042e-01
1.93567529e-01 -4.29356843e-01 -1.17419101e-01 -1.16847008e-01
-3.55042905e-01 4.74253178e-01 8.35478663e-01 -7.52673686e-01
-3.85909408e-01 -3.73330712e-01 -1.94506392e-01 2.88106769e-01
-4.18681830e-01 -5.32461226e-01 3.41546714e-01 -5.52237928e-01
-3.56171727e-02 -8.26170802e-01 -3.15663576e-01 -4.70404178e-01
-3.10987025e-01 -3.01935196e-01 7.22268462e-01 -2.39818200e-01
1.31016707e+00 -2.29168510e+00 3.13315958e-01 5.60606122e-01
3.34109306e-01 8.31976440e-03 -1.51843056e-01 1.99409932e-01
-1.73620418e-01 -9.05084834e-02 1.30541697e-01 -1.28514811e-01
-1.37677878e-01 1.82665139e-01 -8.79365385e-01 8.07547033e-01
-2.60410309e-01 8.35859358e-01 -5.82448423e-01 -1.14900403e-01
-9.20354873e-02 6.56307280e-01 -5.36045730e-01 -7.79450522e-04
5.72572872e-02 1.99711233e-01 -6.57157838e-01 -2.65168875e-01
2.71966070e-01 -3.19677055e-01 2.77610403e-02 9.97331440e-02
-3.60757500e-01 -1.29461333e-01 -8.84306312e-01 1.43010628e+00
-8.37906063e-01 1.12308145e+00 2.35433325e-01 -9.12432432e-01
8.69186103e-01 2.87159979e-01 4.18289810e-01 -3.27022523e-01
1.74635336e-01 2.84701973e-01 4.54092473e-02 4.55567031e-04
4.64459211e-01 -5.25720716e-01 2.38668829e-01 6.50321424e-01
4.06011671e-01 2.51966447e-01 1.35445490e-01 2.46968895e-01
7.18743384e-01 -3.59685183e-01 -1.26883328e-01 -7.24653780e-01
6.83311969e-02 -7.37936378e-01 1.25552461e-01 1.15524995e+00
2.27076367e-01 4.78745431e-01 4.14758235e-01 -4.31822270e-01
-1.39016688e+00 -9.82244074e-01 -3.90902579e-01 1.42107570e+00
-2.00408086e-01 -9.31066424e-02 -8.73237371e-01 -2.81657994e-01
-7.77977183e-02 4.15356040e-01 -8.76794219e-01 -1.37797251e-01
-7.72799253e-01 -6.94967806e-01 7.09022582e-01 7.78678536e-01
4.45780218e-01 -1.16431499e+00 -7.52795994e-01 3.08695406e-01
1.45319268e-01 -9.26214993e-01 -8.80549014e-01 6.30180717e-01
-1.22991371e+00 -6.32766247e-01 -1.29712510e+00 -8.42614532e-01
6.15736246e-01 -2.17577815e-01 5.90633690e-01 -1.32097170e-01
-5.81155904e-03 3.62880677e-01 2.23623037e-01 3.37688290e-02
1.63015779e-02 6.19752347e-01 2.38834381e-01 -3.03487450e-01
3.88697274e-02 -6.51688874e-01 -6.38663530e-01 1.93580449e-01
-7.02847242e-01 -4.21676636e-01 3.15643519e-01 9.10938084e-01
4.71909523e-01 -2.01192334e-01 6.09042525e-01 -6.80121839e-01
1.01225162e+00 -1.59977004e-01 -8.48835349e-01 3.14127803e-01
-3.92462254e-01 8.89122844e-01 7.36814737e-01 -6.92927599e-01
-1.10045660e+00 -1.83418155e-01 -1.09724790e-01 -3.75108659e-01
3.39720756e-01 1.96818337e-01 5.04868865e-01 -1.44264460e-01
9.74142849e-01 1.88673869e-01 -1.13259844e-01 -4.99059588e-01
4.26462501e-01 4.95976388e-01 5.12445450e-01 -7.73831010e-01
5.59058666e-01 5.23857534e-01 1.07676364e-01 -1.18514764e+00
-1.18215525e+00 -1.83928668e-01 -7.04160452e-01 -9.58720669e-02
4.29383636e-01 -5.60455382e-01 -9.59876955e-01 3.80210102e-01
-1.07344544e+00 -9.15206313e-01 -8.99734557e-01 5.26869357e-01
-9.03119922e-01 -8.75472426e-02 -1.02735877e+00 -9.86433566e-01
-6.12363875e-01 -7.22270131e-01 6.84812129e-01 6.36261284e-01
1.54787466e-01 -1.60422289e+00 4.72217858e-01 -6.45777941e-01
7.77005255e-01 -3.34730208e-01 9.29865122e-01 -5.47731996e-01
4.57344726e-02 1.51633427e-01 -1.00116171e-01 1.25907034e-01
-2.72561461e-01 -6.62119240e-02 -8.38773608e-01 -2.42417023e-01
4.43416946e-02 -1.79139376e-01 1.28315389e+00 1.21009517e+00
7.92501152e-01 -2.82196939e-01 -3.24189037e-01 8.65012348e-01
1.41161227e+00 -1.04442112e-01 2.45132715e-01 1.96357638e-01
5.16575336e-01 3.32188278e-01 -3.53461742e-01 4.94145006e-01
-3.57910335e-01 2.75463998e-01 -1.02018856e-01 5.14559522e-02
-4.52152826e-03 -4.36718792e-01 3.73376995e-01 7.91685164e-01
-3.24328959e-01 -7.59995878e-02 -6.40896499e-01 5.15965104e-01
-1.96461844e+00 -1.21635211e+00 4.03330714e-01 2.00573301e+00
8.88838053e-01 2.14212552e-01 2.95638740e-01 -2.23702595e-01
8.85500431e-01 4.34933662e-01 -8.59022200e-01 -4.87563074e-01
-4.76655066e-01 2.33473241e-01 1.09263039e+00 9.92829740e-01
-9.15966928e-01 9.42457557e-01 8.57309628e+00 5.99358559e-01
-1.11175811e+00 1.32918909e-01 2.24207714e-01 -4.17776972e-01
-2.93892711e-01 -3.36088687e-01 -1.37688124e+00 -2.58521596e-03
1.08592737e+00 -2.24935010e-01 6.50204718e-01 6.83969736e-01
2.19722703e-01 2.90973753e-01 -6.41385376e-01 8.44051301e-01
-4.44811732e-01 -1.63631451e+00 -1.94660038e-01 3.97830576e-01
9.24943268e-01 6.40296996e-01 4.35423195e-01 7.87331685e-02
6.35326445e-01 -1.13668048e+00 5.84006667e-01 7.06364393e-01
9.13159788e-01 -9.29972351e-01 4.63752508e-01 3.16641748e-01
-1.07385051e+00 -5.06480694e-01 -1.07162070e+00 6.61421288e-03
2.73591518e-01 6.39047742e-01 -2.15646356e-01 -1.99283093e-01
1.08286507e-01 7.98159122e-01 -9.35699493e-02 7.34503448e-01
2.96089143e-01 5.38479865e-01 -8.42329562e-01 -2.19386935e-01
5.69493175e-01 -2.56943166e-01 4.44660187e-01 1.28315711e+00
3.07695538e-01 3.57889161e-02 -2.37476677e-01 9.10269022e-01
-3.85268927e-01 1.61786433e-02 -8.17556381e-01 -6.75886348e-02
3.00033480e-01 8.85543525e-01 -6.72188759e-01 1.84470937e-01
-1.06218956e-01 3.88116241e-01 7.44123340e-01 9.98294115e-01
-3.87054592e-01 -8.29055429e-01 5.68089604e-01 1.68212116e-01
5.88652551e-01 -3.17667753e-01 -4.44176123e-02 -1.11298621e+00
-1.43396288e-01 -4.75876406e-02 3.28101963e-01 -4.95310813e-01
-1.09674549e+00 7.77942061e-01 1.32107481e-01 -3.67997527e-01
-4.17478919e-01 -7.49528825e-01 -7.50885487e-01 6.95199907e-01
-1.17231238e+00 -5.94887316e-01 5.13560534e-01 8.67976308e-01
2.79566526e-01 -3.69725078e-01 5.10709524e-01 -6.55489862e-02
-5.06681144e-01 7.59085000e-01 6.82730019e-01 4.27396357e-01
-6.23113662e-02 -1.18231177e+00 3.93125892e-01 5.25172234e-01
3.60530585e-01 1.03498352e+00 8.55135977e-01 -3.33272308e-01
-1.08947194e+00 -9.32728648e-01 4.82965976e-01 8.37598592e-02
9.72166657e-01 -3.77957672e-01 -7.93669105e-01 1.06269765e+00
2.18143925e-01 3.77803445e-01 1.78495780e-01 4.51718420e-01
-2.87290871e-01 1.25366688e-01 -7.32494771e-01 3.97749096e-01
1.11274624e+00 -7.35874593e-01 -5.98238528e-01 2.09306583e-01
4.41340476e-01 -3.51068676e-01 -6.71827674e-01 5.51927052e-02
6.49334013e-01 -4.78208601e-01 6.01238191e-01 -1.03352797e+00
-1.97786883e-01 1.66546375e-01 -9.30872858e-02 -1.64291632e+00
-2.99489677e-01 -1.11467612e+00 -2.37083972e-01 5.71115911e-01
6.91050828e-01 -7.33991504e-01 7.81859756e-01 2.65525188e-02
7.69092366e-02 -7.38118172e-01 -1.09120595e+00 -7.23401487e-01
6.51027799e-01 -7.30183870e-02 -8.40901658e-02 1.80767179e-01
2.19088659e-01 6.26239061e-01 -2.52844512e-01 -2.35028505e-01
5.13836622e-01 -1.63870499e-01 -1.31243914e-01 -1.18830717e+00
-2.44926989e-01 -6.99671566e-01 -1.15453497e-01 -1.64878094e+00
8.26700509e-01 -7.94877768e-01 2.19476640e-01 -1.33449626e+00
1.19036138e-01 -1.45661607e-01 -3.34340811e-01 1.39916733e-01
3.55771214e-01 -1.98207974e-01 6.05288800e-03 2.42168009e-01
-4.29377466e-01 4.72359240e-01 1.09604824e+00 1.67961836e-01
-4.65445012e-01 2.52675205e-01 -4.07436788e-01 7.80286908e-01
7.59179771e-01 -4.62671697e-01 -7.60034740e-01 -3.24032336e-01
6.01250648e-01 -1.11672707e-01 3.35660367e-03 -7.93112874e-01
4.89206493e-01 -1.67267472e-02 3.55601490e-01 -4.65071082e-01
8.39195624e-02 -2.84829825e-01 -4.28104252e-01 2.73779005e-01
-1.10091734e+00 2.34435156e-01 -3.68580371e-02 5.45509934e-01
9.19259861e-02 -6.34325206e-01 1.17747688e+00 1.09119611e-02
-1.54067844e-01 3.76508534e-01 -5.43684661e-01 8.01213562e-01
4.30575281e-01 -2.37586759e-02 -4.65943605e-01 -8.19921911e-01
-1.10019970e+00 1.20537408e-01 -3.69634442e-02 -1.65369868e-01
6.75956190e-01 -1.08090460e+00 -4.30527568e-01 2.81648129e-01
-4.88217801e-01 -9.60233286e-02 2.13024348e-01 7.55395293e-01
-2.24712878e-01 5.78068912e-01 7.04343384e-03 -2.61209428e-01
-4.70100701e-01 1.84894755e-01 1.21657825e+00 -1.95806906e-01
-1.01369774e+00 1.00914752e+00 3.71261746e-01 -3.42426121e-01
4.78038251e-01 -3.59065890e-01 -2.39442900e-01 -5.25461622e-02
7.19892383e-01 3.08927774e-01 -1.02024369e-01 -4.91672933e-01
9.96174067e-02 8.99145246e-01 -3.08899283e-01 -5.97838402e-01
1.44727147e+00 -8.82126763e-02 2.47799158e-01 6.19218290e-01
1.40321302e+00 -1.02907948e-01 -1.67276525e+00 -3.79136860e-01
7.15494677e-02 2.95827389e-01 3.68762791e-01 -1.10188916e-01
-1.38317180e+00 1.02314579e+00 5.79553783e-01 1.18981563e-01
5.86819828e-01 2.73206979e-01 6.34707153e-01 8.51162672e-01
-7.48533905e-02 -1.43306875e+00 -1.98558569e-01 1.22895896e+00
6.11878574e-01 -3.21096063e-01 -2.44718090e-01 5.05578876e-01
-4.03229237e-01 1.53511405e+00 3.17832679e-01 -7.88050950e-01
1.18774390e+00 6.40440643e-01 -1.06180519e-01 -2.29784518e-01
-1.08992326e+00 -1.36424556e-01 1.38949305e-01 6.30210519e-01
6.51625633e-01 -2.73037732e-01 2.30375275e-01 1.33472800e-01
-3.08699608e-01 -2.07788065e-01 4.67513889e-01 5.51740944e-01
-9.15586531e-01 -3.98786843e-01 1.94557846e-01 4.27337259e-01
-4.94314551e-01 -1.95237041e-01 -4.39935513e-02 8.23201954e-01
-6.98203027e-01 5.76972246e-01 4.54442799e-01 -7.65555874e-02
3.46161127e-01 3.47015619e-01 8.58998835e-01 -3.67723674e-01
-5.70484221e-01 3.39089096e-01 -3.81655216e-01 -1.62692517e-01
-2.77179539e-01 -4.60157543e-01 -1.64993727e+00 -4.46840465e-01
-6.82552934e-01 4.29727733e-01 4.03431982e-01 1.24241757e+00
1.85217232e-01 2.97812581e-01 4.93766606e-01 -5.26562452e-01
-9.93290424e-01 -1.20350230e+00 -9.95772004e-01 -6.90475926e-02
9.83536780e-01 -4.05969560e-01 -7.29900956e-01 -4.52792615e-01] | [7.883634567260742, 3.5065433979034424] |
2d799a83-b7ed-44eb-88eb-7dd13e57df99 | not-all-poisons-are-created-equal-robust | 2210.09671 | null | https://arxiv.org/abs/2210.09671v1 | https://arxiv.org/pdf/2210.09671v1.pdf | Not All Poisons are Created Equal: Robust Training against Data Poisoning | Data poisoning causes misclassification of test time target examples by injecting maliciously crafted samples in the training data. Existing defenses are often effective only against a specific type of targeted attack, significantly degrade the generalization performance, or are prohibitive for standard deep learning pipelines. In this work, we propose an efficient defense mechanism that significantly reduces the success rate of various data poisoning attacks, and provides theoretical guarantees for the performance of the model. Targeted attacks work by adding bounded perturbations to a randomly selected subset of training data to match the targets' gradient or representation. We show that: (i) under bounded perturbations, only a number of poisons can be optimized to have a gradient that is close enough to that of the target and make the attack successful; (ii) such effective poisons move away from their original class and get isolated in the gradient space; (iii) dropping examples in low-density gradient regions during training can successfully eliminate the effective poisons, and guarantees similar training dynamics to that of training on full data. Our extensive experiments show that our method significantly decreases the success rate of state-of-the-art targeted attacks, including Gradient Matching and Bullseye Polytope, and easily scales to large datasets. | ['Baharan Mirzasoleiman', 'Tian Yu Liu', 'Yu Yang'] | 2022-10-18 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [-4.15958352e-02 -2.46432930e-01 -3.10647070e-01 -1.11436017e-01
-8.80721152e-01 -9.57104921e-01 3.09487224e-01 2.82387108e-01
-6.00887060e-01 5.56027830e-01 -4.05959040e-01 -3.82172316e-01
1.34592401e-02 -9.94345307e-01 -1.12916410e+00 -1.00238907e+00
-3.46107036e-01 6.79805100e-01 6.89208150e-01 -7.46568991e-03
3.13083977e-01 8.62158000e-01 -9.76329744e-01 2.47389063e-01
6.36861742e-01 6.37724876e-01 -5.49723208e-01 7.89490998e-01
9.31328312e-02 6.54531240e-01 -1.27372038e+00 -5.73835015e-01
6.45967603e-01 -2.07858115e-01 -4.71637696e-01 -2.65956968e-01
7.84757316e-01 -6.90472007e-01 -7.08936512e-01 1.52874088e+00
4.22969252e-01 -3.25651675e-01 4.67257649e-01 -1.54861486e+00
-4.64649171e-01 6.06070042e-01 -7.45368481e-01 3.75455558e-01
-2.10591629e-01 4.32423919e-01 6.08345926e-01 -4.63812083e-01
2.54934102e-01 1.16810262e+00 6.43927574e-01 1.15887654e+00
-1.08373773e+00 -1.14584243e+00 2.54543066e-01 -2.00202048e-01
-1.12286508e+00 -1.43803880e-01 4.83086675e-01 -8.87380168e-03
6.36053622e-01 4.64170098e-01 5.52405179e-01 1.27617657e+00
9.44780484e-02 8.44300091e-01 7.48117685e-01 4.10510972e-02
3.74805421e-01 1.62085686e-02 3.00022960e-01 8.77174079e-01
8.37873638e-01 3.34235698e-01 -4.87677306e-01 -7.45476365e-01
4.09725100e-01 3.34483415e-01 -4.40379679e-01 -4.30879861e-01
-5.99586248e-01 1.24057019e+00 5.64498067e-01 -9.22882184e-03
1.70965251e-02 3.24360371e-01 4.94317085e-01 2.88974226e-01
3.96446019e-01 6.16244376e-01 -5.92369199e-01 9.83707011e-02
-4.97318834e-01 6.48748040e-01 9.19422448e-01 5.61004758e-01
5.77828646e-01 1.48358583e-01 -1.27294913e-01 3.81832689e-01
-4.65048030e-02 9.31658089e-01 3.85431409e-01 -4.05031383e-01
6.72098815e-01 7.44611621e-01 4.91820090e-02 -9.10360932e-01
-8.97374898e-02 -5.35742164e-01 -4.71469462e-01 3.29021424e-01
7.55256295e-01 -4.10590529e-01 -1.02842259e+00 1.91762817e+00
6.99137926e-01 3.23398441e-01 -1.05382621e-01 6.72008336e-01
2.64664978e-01 5.39536178e-01 1.67237315e-02 1.52836904e-01
9.28722501e-01 -6.42245412e-01 -4.85862754e-02 -4.33318108e-01
9.91434872e-01 -1.66958511e-01 1.29729509e+00 2.79118448e-01
-8.67689908e-01 2.34319523e-01 -1.24140406e+00 5.75418234e-01
-5.12512207e-01 -6.66242898e-01 4.83102202e-01 1.08562601e+00
-5.03570139e-01 7.85125017e-01 -8.72215390e-01 2.52319098e-01
9.97073889e-01 6.76453590e-01 -1.87245682e-01 -4.06871021e-01
-1.08525681e+00 5.23429990e-01 5.85040264e-02 -3.36330980e-01
-1.65147448e+00 -1.18714011e+00 -6.05864108e-01 1.42479926e-01
3.07396203e-01 -4.07818705e-01 9.84029233e-01 -4.11901832e-01
-8.56087089e-01 6.00932419e-01 2.61011392e-01 -8.19281638e-01
7.26550162e-01 -4.15324360e-01 3.96881141e-02 2.39487737e-01
-4.61779870e-02 3.00772697e-01 9.83375192e-01 -1.14057076e+00
-7.67795205e-01 -6.26492918e-01 8.11938345e-02 9.35172364e-02
-1.15100944e+00 8.54001194e-02 -1.85511768e-01 -3.53573918e-01
-2.82169908e-01 -7.71765411e-01 -4.38890874e-01 8.28912854e-02
-5.93700349e-01 -3.98378372e-02 1.36621630e+00 9.05782450e-03
1.09680128e+00 -2.10896897e+00 -2.35347614e-01 2.47105762e-01
6.55363858e-01 8.60709310e-01 -2.38971353e-01 7.75520429e-02
1.79582015e-01 4.72957194e-01 -2.92868197e-01 -1.59845471e-01
-5.52559346e-02 1.39763415e-01 -8.33712578e-01 1.04931808e+00
-1.55539764e-02 6.73457026e-01 -1.05366266e+00 -8.86639729e-02
-1.22646324e-01 2.34033644e-01 -8.56118679e-01 3.07147890e-01
-2.29097202e-01 1.46850888e-02 -5.95426142e-01 5.33837378e-01
8.63062739e-01 8.60318393e-02 -3.01798761e-01 2.19109058e-01
4.12941873e-01 4.23923790e-01 -7.98125088e-01 6.39364302e-01
-4.51600738e-02 3.99537653e-01 1.58641458e-01 -7.38800108e-01
7.49445617e-01 -1.49527714e-01 4.22248393e-01 -2.54697293e-01
1.19409814e-01 2.45462373e-01 1.61302254e-01 -7.26467445e-02
-2.62043417e-01 -4.38966276e-03 -1.04289740e-01 6.48070872e-01
-3.82353395e-01 -8.25203508e-02 -5.27956486e-02 2.91284770e-01
1.71478653e+00 -5.73994458e-01 -3.20156038e-01 -1.27190515e-01
1.29831761e-01 2.30764732e-01 6.83029473e-01 1.28001177e+00
-4.89719927e-01 9.67077166e-02 7.77316034e-01 -6.53978705e-01
-1.08409715e+00 -1.02599144e+00 -1.04858674e-01 9.51610267e-01
2.56637007e-01 -2.71070659e-01 -1.17903626e+00 -1.44415522e+00
4.30606723e-01 4.58012044e-01 -7.19240367e-01 -8.18881869e-01
-7.47087717e-01 -1.03875089e+00 1.20142353e+00 3.52494121e-01
6.72102571e-01 -8.35679293e-01 -5.53569078e-01 -1.50823817e-01
4.96973127e-01 -7.86578536e-01 -6.01762116e-01 4.39655751e-01
-9.52812731e-01 -1.58267629e+00 -3.14274669e-01 -5.84072351e-01
1.02000964e+00 4.20135677e-01 8.18192363e-01 5.63071251e-01
-3.28888237e-01 -8.90742391e-02 -1.38425127e-01 -6.57697082e-01
-4.91062194e-01 -7.27772415e-02 1.17482334e-01 -2.64141023e-01
6.25511050e-01 -2.08277687e-01 -4.72349703e-01 4.06501174e-01
-9.82162595e-01 -7.93361664e-01 3.45909595e-01 9.43396926e-01
2.38380447e-01 4.49194968e-01 4.65827793e-01 -1.22288036e+00
6.56410933e-01 -4.56899792e-01 -9.38911796e-01 4.21805009e-02
-3.41420203e-01 1.11782765e-02 1.06623793e+00 -1.16606104e+00
-2.67993510e-01 4.86417674e-02 -7.09465668e-02 -1.03278816e+00
2.91626062e-02 -1.59037456e-01 -3.13495308e-01 -4.85179126e-01
1.18135893e+00 1.01566896e-01 -5.82873896e-02 -3.89456570e-01
-4.44320322e-04 3.91584456e-01 3.31599593e-01 -7.00594962e-01
1.40602970e+00 5.78123510e-01 1.68431967e-01 -4.92268384e-01
-1.00557458e+00 -1.93402484e-01 -2.48573914e-01 1.73935816e-01
2.11595163e-01 -5.14075220e-01 -8.66005898e-01 7.41829097e-01
-7.96128631e-01 -6.82582974e-01 -2.47712567e-01 6.95744902e-03
1.52329892e-01 4.22934741e-01 -9.14831698e-01 -8.07608724e-01
-4.47701246e-01 -1.18039310e+00 7.64475524e-01 3.15576822e-01
1.76878437e-01 -9.37801003e-01 -1.05852425e-01 -1.11450488e-02
2.79137790e-01 3.38202655e-01 9.88185406e-01 -1.32788539e+00
-3.74350190e-01 -8.61366391e-01 1.11108311e-01 3.28779161e-01
-1.12876020e-01 3.52292024e-02 -7.87170827e-01 -6.75161362e-01
3.54227185e-01 -5.31789362e-01 9.33076620e-01 3.91685814e-02
1.36868179e+00 -8.15293312e-01 -5.87067246e-01 8.28243494e-01
1.15717447e+00 2.17297718e-01 4.32030439e-01 3.87865961e-01
7.96466827e-01 3.43052685e-01 5.35182893e-01 3.96667600e-01
-4.13622826e-01 3.23176831e-01 9.54063773e-01 -3.44317853e-02
2.23247334e-01 -3.43012303e-01 5.21831155e-01 -1.64046228e-01
9.14291203e-01 -1.65823862e-01 -8.39866042e-01 5.72568059e-01
-1.60791600e+00 -9.14389849e-01 -1.29388288e-01 2.62570500e+00
1.00084984e+00 6.03889167e-01 3.95859212e-01 2.38032788e-01
8.17743361e-01 -6.67422712e-02 -1.11642003e+00 -3.44201863e-01
8.20559636e-02 2.52048671e-01 1.07168663e+00 5.13870955e-01
-1.30558956e+00 1.14151275e+00 6.72415733e+00 1.02581894e+00
-9.82281804e-01 1.46089658e-01 6.65139079e-01 -5.48742235e-01
-1.09796807e-01 -1.61539689e-01 -1.17777085e+00 4.38525587e-01
9.48513508e-01 -1.24933667e-01 5.62821209e-01 1.27934003e+00
-3.98466796e-01 3.60766351e-01 -1.12807965e+00 4.40910399e-01
-2.18186136e-02 -1.24998140e+00 1.26810297e-02 4.03056800e-01
5.05335212e-01 8.38734284e-02 4.94960189e-01 3.78484875e-01
9.26255703e-01 -1.01193488e+00 3.01556319e-01 -4.61991966e-01
4.02413845e-01 -1.25355959e+00 4.47319120e-01 6.56067729e-01
-5.02035677e-01 -5.80708325e-01 -7.68631399e-01 3.47112387e-01
-4.83844429e-01 6.32292211e-01 -9.64681506e-01 -3.51530701e-01
7.29068041e-01 4.64827567e-02 -7.43229985e-01 1.16587889e+00
-3.33507806e-01 8.52426171e-01 -6.57036424e-01 -4.32138503e-01
3.89236361e-01 2.95141071e-01 7.57936418e-01 9.28448021e-01
-2.48056725e-01 -2.54134387e-01 2.65873820e-01 6.49420798e-01
-4.23885554e-01 -2.20204279e-01 -8.98177505e-01 -1.71573699e-01
8.43280733e-01 1.02144349e+00 -5.34922004e-01 -3.16589177e-01
-2.18318757e-02 6.81428850e-01 6.11689091e-01 6.78427294e-02
-1.14136028e+00 -4.80536044e-01 1.12583041e+00 6.57271296e-02
2.28141636e-01 1.68863103e-01 -3.67339045e-01 -1.02780259e+00
-1.07521445e-01 -1.23377538e+00 6.73714638e-01 3.06973024e-03
-1.50843191e+00 5.63482821e-01 -1.54943019e-01 -9.46444035e-01
8.22656304e-02 -8.12625766e-01 -7.27089882e-01 6.80259466e-01
-8.90003800e-01 -5.32633066e-01 4.95957769e-02 8.04768741e-01
2.42955670e-01 -3.32123429e-01 7.27136672e-01 2.50494182e-01
-9.88717020e-01 1.06245410e+00 2.01916527e-02 4.84155059e-01
6.08790100e-01 -1.22544992e+00 6.30933523e-01 1.27016628e+00
1.25244468e-01 7.40088701e-01 7.71699011e-01 -9.48207974e-01
-1.39828503e+00 -1.22460032e+00 2.21467584e-01 -6.99255943e-01
8.18411589e-01 -5.51817477e-01 -1.07324052e+00 4.91781265e-01
-5.89263082e-01 4.02650148e-01 5.22822738e-01 -1.74170900e-02
-8.81317616e-01 -1.03889570e-01 -1.78328228e+00 1.02859533e+00
7.72690833e-01 -4.15919453e-01 -1.06601045e-01 9.12690580e-01
7.31583118e-01 -5.21950722e-01 -4.48409677e-01 3.91589791e-01
1.27151489e-01 -8.01728249e-01 1.08302248e+00 -1.25194383e+00
1.39414832e-01 -1.29016399e-01 1.18993491e-01 -1.11761785e+00
-1.01829760e-01 -6.26383722e-01 -5.59059739e-01 7.83196270e-01
3.24653476e-01 -7.03752995e-01 1.35358083e+00 7.03730643e-01
-6.91998936e-03 -1.09171438e+00 -1.00415313e+00 -8.02047789e-01
5.40926874e-01 -3.85389291e-02 6.49587810e-01 7.40427196e-01
-9.74936560e-02 2.02813912e-02 -2.88345963e-01 6.17213011e-01
1.02317643e+00 -2.76760668e-01 9.17189538e-01 -1.04410970e+00
-4.23020512e-01 -5.71990609e-01 -4.05282199e-01 -8.42672944e-01
1.34471819e-01 -5.05494058e-01 -4.62501030e-03 -8.72312188e-01
2.38827795e-01 -6.22546315e-01 -2.39860505e-01 8.55719686e-01
-5.12654185e-01 4.14685875e-01 3.26544046e-02 2.49246299e-01
-3.97089392e-01 7.85430372e-02 9.91356432e-01 -2.91498274e-01
-1.31922066e-01 2.68993765e-01 -7.78426349e-01 6.80847168e-01
9.11388457e-01 -1.21540475e+00 -3.75118643e-01 -1.69627652e-01
-4.76332344e-02 -4.07309294e-01 2.41300866e-01 -8.52609873e-01
1.25446171e-01 -1.74750641e-01 6.49514079e-01 -3.12447816e-01
1.60604492e-01 -7.12500155e-01 -5.38402379e-01 1.26604879e+00
-4.87780482e-01 7.53904730e-02 2.12477535e-01 6.62871301e-01
5.16515791e-01 -5.87543011e-01 1.29549205e+00 1.13710016e-01
1.55658036e-01 7.08296061e-01 -1.74773306e-01 3.12234968e-01
1.23911166e+00 7.01343715e-02 -9.11017835e-01 -1.57125797e-02
-9.94675606e-02 3.23276937e-01 7.56527603e-01 5.75491115e-02
5.93557417e-01 -1.08924663e+00 -4.79265690e-01 3.05950254e-01
-1.75647378e-01 1.46352902e-01 -1.92518055e-01 3.39198649e-01
-5.36320508e-01 3.04588443e-03 3.54116783e-02 -4.46273386e-01
-1.38154745e+00 1.01343811e+00 7.78707743e-01 -3.99220556e-01
-7.57920444e-01 1.16099119e+00 3.48910958e-01 -2.73951590e-01
6.63355350e-01 2.25858733e-01 2.18306884e-01 -5.21593750e-01
9.15147662e-01 2.25027680e-01 -1.25704706e-01 -2.34341204e-01
-4.55097497e-01 -2.39806399e-02 -6.11263752e-01 8.68283212e-02
1.23928726e+00 8.26812029e-01 1.31450417e-02 -2.20102921e-01
1.36654508e+00 3.09255391e-01 -1.59546149e+00 4.07719389e-02
-1.63511679e-01 -9.02144194e-01 -8.09439048e-02 -6.16604030e-01
-1.22863281e+00 9.14371669e-01 4.83519584e-01 5.24498284e-01
9.61392939e-01 -2.16355085e-01 9.43348110e-01 6.23539746e-01
2.36156031e-01 -5.95740259e-01 4.48894531e-01 2.70383179e-01
3.43874693e-01 -7.10850000e-01 -5.58996014e-02 -3.81505698e-01
-3.61958385e-01 8.68073761e-01 1.07314658e+00 -6.18788779e-01
3.72534335e-01 7.26464570e-01 -1.48143843e-01 -2.42752597e-01
-8.23944747e-01 2.92858779e-01 -9.30017233e-02 7.45411277e-01
-3.45372975e-01 -2.86158562e-01 1.34779871e-01 3.46436501e-01
-1.04868226e-02 -7.22059309e-01 4.33188558e-01 9.34551299e-01
-7.93692589e-01 -1.05035377e+00 -6.05923295e-01 5.25529623e-01
-8.69590580e-01 -3.55077796e-02 -8.00978899e-01 8.40126097e-01
-2.96479106e-01 8.73512626e-01 -2.10102126e-02 -5.93860328e-01
2.50441968e-01 -7.65396878e-02 4.32336777e-01 -6.78835928e-01
-8.69880795e-01 -1.55621082e-01 -2.49554098e-01 -5.24638116e-01
6.21699452e-01 -3.23318124e-01 -1.20329905e+00 -6.52009904e-01
-4.72830653e-01 3.28093261e-01 5.29272139e-01 8.30039859e-01
1.51142165e-01 1.46224141e-01 1.01540792e+00 -4.58394378e-01
-1.38670182e+00 -6.13464832e-01 -4.11940813e-01 4.40147668e-01
4.39258546e-01 -4.95476335e-01 -1.06435537e+00 -6.06963873e-01] | [5.789267539978027, 7.627106666564941] |
016c05fa-7bb2-464a-923a-6d791bceb248 | addressing-the-challenges-of-open-world | 2303.14930 | null | https://arxiv.org/abs/2303.14930v1 | https://arxiv.org/pdf/2303.14930v1.pdf | Addressing the Challenges of Open-World Object Detection | We address the challenging problem of open world object detection (OWOD), where object detectors must identify objects from known classes while also identifying and continually learning to detect novel objects. Prior work has resulted in detectors that have a relatively low ability to detect novel objects, and a high likelihood of classifying a novel object as one of the known classes. We approach the problem by identifying the three main challenges that OWOD presents and introduce OW-RCNN, an open world object detector that addresses each of these three challenges. OW-RCNN establishes a new state of the art using the open-world evaluation protocol on MS-COCO, showing a drastically increased ability to detect novel objects (16-21% absolute increase in U-Recall), to avoid their misclassification as one of the known classes (up to 52% reduction in A-OSE), and to incrementally learn to detect them while maintaining performance on previously known classes (1-6% absolute increase in mAP). | ['Niko Sünderhauf', 'Dimity Miller', 'Feras Dayoub', 'David Pershouse'] | 2023-03-27 | null | null | null | null | ['open-world-object-detection'] | ['computer-vision'] | [ 1.75117910e-01 1.43065199e-01 3.71353142e-02 -1.04004517e-01
-9.60972428e-01 -8.68759453e-01 6.90136254e-01 5.21716297e-01
-6.74951077e-01 5.94635546e-01 -2.79379040e-01 3.28494497e-02
-7.86916092e-02 -8.33510220e-01 -7.71267831e-01 -2.74352908e-01
-4.93711323e-01 7.14093387e-01 1.18164968e+00 1.87547401e-01
1.24578148e-01 7.68949151e-01 -2.09147954e+00 4.34468627e-01
5.89353621e-01 1.15652370e+00 6.43240213e-02 1.03481328e+00
-1.92398280e-01 6.16436899e-01 -5.96963346e-01 -1.32217124e-01
4.24481213e-01 1.92312356e-02 -1.06423020e+00 -1.32986426e-01
1.05517757e+00 -3.14087987e-01 -1.14045285e-01 1.00353181e+00
4.49938595e-01 7.73692206e-02 6.61666572e-01 -1.31123149e+00
-6.06911838e-01 4.50870007e-01 -5.26710689e-01 8.98922265e-01
2.03546137e-01 1.94201887e-01 1.04734850e+00 -1.08450437e+00
7.54331648e-01 1.23447359e+00 8.94111395e-01 4.30434257e-01
-1.41991842e+00 -8.07628632e-01 2.23501265e-01 3.62944573e-01
-1.78340483e+00 -3.85718703e-01 -2.43487582e-01 -3.73303294e-01
1.20641255e+00 4.66510177e-01 1.62282199e-01 5.48264503e-01
7.72533752e-03 8.54035318e-01 8.26920450e-01 -3.72098833e-01
3.78441960e-01 4.00534064e-01 5.76577067e-01 4.40539628e-01
6.65431678e-01 1.59118667e-01 -3.34962755e-01 -1.39183670e-01
6.34511188e-02 -3.41982841e-02 1.38897032e-01 -4.75361705e-01
-1.19459975e+00 5.01107275e-01 7.98552275e-01 1.82999417e-01
-1.20719470e-01 1.05081961e-01 3.87631267e-01 2.01759294e-01
3.43046576e-01 7.64280379e-01 -6.54283822e-01 -7.05658123e-02
-4.59939927e-01 4.58908617e-01 7.02457130e-01 1.11165631e+00
8.56450796e-01 -2.24086225e-01 -3.75241011e-01 5.55304706e-01
-7.95855448e-02 5.17173171e-01 6.14676058e-01 -4.57528651e-01
2.35841393e-01 8.81641567e-01 1.28956854e-01 -5.96325636e-01
-5.21180570e-01 -5.25240541e-01 -5.76591864e-03 3.61086875e-01
2.00002536e-01 1.06460743e-01 -1.24227834e+00 1.55770659e+00
5.50397694e-01 3.17968220e-01 1.25288889e-01 2.64363945e-01
8.89772952e-01 5.61390281e-01 1.47467226e-01 -9.37829241e-02
1.32381439e+00 -7.27450013e-01 -2.02062383e-01 -4.72188711e-01
7.47514665e-01 -5.57538807e-01 8.09310913e-01 9.36845765e-02
-5.86458623e-01 -5.19821405e-01 -1.24457419e+00 2.80416220e-01
-8.52210641e-01 -3.42729211e-01 4.77374673e-01 6.37321234e-01
-6.89695418e-01 3.50054383e-01 -6.75863862e-01 -7.32425153e-01
8.14469397e-01 4.41897959e-01 -3.44583213e-01 -2.61681587e-01
-7.23125100e-01 9.51602757e-01 1.16408861e+00 -5.72830737e-01
-1.11371374e+00 -9.51585948e-01 -5.50567389e-01 1.92090645e-01
8.36721659e-01 -1.02690689e-01 1.49138546e+00 -6.22221172e-01
-3.14847708e-01 7.84010530e-01 1.21910974e-01 -6.87171459e-01
5.09330511e-01 -3.81393313e-01 -9.24987078e-01 8.94284472e-02
4.31950212e-01 8.45242918e-01 5.62002420e-01 -1.21308255e+00
-1.45316279e+00 -2.93518901e-01 1.55138180e-01 3.52465212e-02
-4.69732434e-01 1.63870677e-01 -3.38806547e-02 -5.09957731e-01
2.92196631e-01 -9.99442160e-01 -5.48600666e-02 1.85155526e-01
-1.06687166e-01 -5.33687532e-01 1.26924098e+00 -6.93505304e-03
1.13660145e+00 -2.03657365e+00 -5.45760691e-01 -1.78356748e-02
6.60501838e-01 6.04869664e-01 -2.36832812e-01 -1.14994645e-02
-1.54001117e-01 4.81851585e-02 -1.25412524e-01 2.66018152e-01
-2.10982203e-01 1.93083405e-01 -4.42141116e-01 9.82541293e-02
4.73775417e-01 8.85347128e-01 -1.22825694e+00 -2.34305516e-01
-4.34955135e-02 -1.10387839e-01 -2.86234766e-01 -8.65307916e-03
-2.87045926e-01 -3.88013154e-01 -1.28858626e-01 6.67496562e-01
8.46602619e-01 -2.48072669e-01 -5.05842827e-02 1.69661090e-01
-1.52239323e-01 5.31716365e-03 -1.78770459e+00 7.99140632e-01
-4.19091508e-02 6.60745859e-01 -3.37644815e-01 -5.53920269e-01
7.84941494e-01 1.12792075e-01 3.69992912e-01 -5.55815399e-01
-4.60593998e-02 4.39248860e-01 2.50919521e-01 -3.51715595e-01
4.46102947e-01 4.60471995e-02 2.30790868e-01 6.27619147e-01
4.44376111e-01 2.36249924e-01 4.59714592e-01 3.91560107e-01
1.45350099e+00 -5.16483963e-01 3.38716656e-01 -4.33733523e-01
2.49430805e-01 -1.16683904e-03 5.71878135e-01 1.43053555e+00
-3.72119635e-01 5.74676156e-01 1.52941510e-01 -8.13971400e-01
-7.89237380e-01 -1.48287618e+00 -4.35097694e-01 1.42565119e+00
1.32490769e-01 -4.04253423e-01 -3.54643106e-01 -1.01899195e+00
4.08384740e-01 5.83846450e-01 -6.04146183e-01 -4.71845627e-01
-2.65403032e-01 -7.65661955e-01 5.96931934e-01 7.27915943e-01
2.67178148e-01 -1.04940975e+00 -9.36610699e-01 3.06731224e-01
6.90518767e-02 -1.05178273e+00 -1.19047314e-01 6.39633358e-01
-7.12532640e-01 -1.28694165e+00 -4.62056071e-01 -8.73913825e-01
6.46705389e-01 4.42160070e-01 1.24278271e+00 6.64089471e-02
-8.84325922e-01 2.44441763e-01 -4.23806757e-01 -9.76298690e-01
-4.78075475e-01 1.03698708e-01 4.69124973e-01 3.10726985e-02
7.23775566e-01 4.64333929e-02 -2.18193069e-01 5.86192071e-01
-9.44421589e-01 -5.77070773e-01 4.86230165e-01 5.23150623e-01
4.64686275e-01 1.52764067e-01 7.98779011e-01 -1.04812849e+00
1.06615312e-01 -6.57761395e-01 -6.91796899e-01 4.06001091e-01
-9.72047448e-01 -6.12720400e-02 1.74949095e-01 -7.90545940e-01
-9.53304291e-01 2.11168572e-01 2.31962427e-01 -1.27215087e-01
-2.42035866e-01 7.69874901e-02 1.50819957e-01 -2.45889321e-01
1.32864058e+00 4.32139970e-02 -3.39700192e-01 -5.21826684e-01
3.85579765e-01 7.66752601e-01 7.80759275e-01 -4.36120212e-01
1.07895410e+00 5.69686949e-01 -3.83494198e-01 -5.46112001e-01
-1.09142530e+00 -9.31955099e-01 -9.07503366e-01 1.89449731e-02
5.08052230e-01 -1.00404310e+00 -4.00106728e-01 3.57560813e-01
-8.21446300e-01 1.41276285e-01 -6.33562028e-01 2.65304774e-01
-1.76316172e-01 2.00656280e-01 -2.50272393e-01 -6.46318913e-01
-3.29978108e-01 -8.76807690e-01 7.39386439e-01 3.91796649e-01
-2.51514822e-01 -4.83893663e-01 -7.29309618e-02 1.09766446e-01
3.35145116e-01 1.31115898e-01 7.10752666e-01 -1.12305522e+00
-7.58626938e-01 -5.89572668e-01 -4.74627227e-01 2.14525715e-01
1.31881312e-01 -2.07796842e-01 -1.17258191e+00 -6.33508921e-01
-4.35284793e-01 -4.58529770e-01 1.05080020e+00 -5.19378372e-02
8.46831679e-01 1.22467242e-02 -7.76853800e-01 1.28897250e-01
1.31880283e+00 3.39954287e-01 4.02114093e-01 5.35059333e-01
3.49226445e-01 1.25625759e-01 5.56456983e-01 3.58312309e-01
3.18938605e-02 6.50827765e-01 2.99349338e-01 -2.91920193e-02
-2.57492036e-01 -6.51720986e-02 1.18106388e-01 -1.07153744e-01
3.12578619e-01 -1.83828741e-01 -1.26747954e+00 9.14517581e-01
-1.89707458e+00 -8.55909705e-01 -2.41793931e-01 2.26671267e+00
6.45792842e-01 6.69040978e-01 4.18626904e-01 7.17145354e-02
7.96559930e-01 -2.46845737e-01 -6.62237704e-01 -1.59218967e-01
-8.67558867e-02 1.33562759e-01 5.67351341e-01 4.38801683e-02
-1.36324096e+00 8.51353765e-01 7.68844748e+00 6.13420188e-01
-8.89658749e-01 1.48249358e-01 2.52041101e-01 -2.93679416e-01
4.06531215e-01 1.35514736e-02 -1.17059100e+00 1.78594545e-01
8.21131766e-01 -4.94409323e-01 7.14336559e-02 1.09542584e+00
-5.21067798e-01 -3.65359873e-01 -1.46514654e+00 7.95121908e-01
2.08114922e-01 -1.39797819e+00 1.97147548e-01 -2.14122579e-01
8.37746501e-01 5.50526679e-01 -2.45667040e-01 6.41181648e-01
4.30985570e-01 -6.62733138e-01 8.20838451e-01 6.16551824e-02
6.91220701e-01 -6.10259235e-01 6.40989542e-01 4.31414247e-01
-1.32971311e+00 -5.94487667e-01 -5.23874044e-01 -2.61949837e-01
-2.80156165e-01 2.14005545e-01 -1.34246826e+00 1.86150700e-01
1.21810222e+00 4.67376262e-01 -1.21396542e+00 1.55480742e+00
1.46248639e-01 5.67061782e-01 -6.43912494e-01 6.05138578e-02
1.86249778e-01 7.51332104e-01 7.24990368e-01 1.11710978e+00
-9.26959515e-02 6.65505007e-02 5.56406438e-01 6.48114145e-01
-1.69530258e-01 -1.97090685e-01 -4.87790078e-01 8.65307823e-02
4.90055472e-01 1.01879001e+00 -1.18405175e+00 -6.27951026e-01
-3.12553674e-01 6.20017469e-01 4.35431391e-01 -9.80448350e-02
-7.03945637e-01 -7.43373811e-01 6.52540267e-01 1.92344055e-01
5.94885886e-01 2.41604507e-01 -1.70008957e-01 -9.30424333e-01
1.32687777e-01 -6.32079065e-01 9.62917507e-01 -4.51253772e-01
-1.28447175e+00 5.46537340e-01 1.59073155e-02 -1.23129106e+00
1.98511615e-01 -6.79647207e-01 -4.32012290e-01 4.71685112e-01
-1.23311388e+00 -9.84879613e-01 -3.06357712e-01 1.74502984e-01
6.56532109e-01 -2.09713265e-01 7.27613270e-01 5.35343826e-01
-2.69683301e-01 7.40728378e-01 3.82195830e-01 1.16047412e-01
6.63372338e-01 -1.31457639e+00 7.40361452e-01 9.50649798e-01
3.81784737e-01 2.86912769e-01 5.75585723e-01 -7.00874865e-01
-9.80839849e-01 -1.40475857e+00 8.75918746e-01 -1.16014278e+00
6.99230015e-01 -6.18786573e-01 -1.18289638e+00 9.16287661e-01
-6.47593319e-01 7.16697633e-01 2.71232247e-01 4.48631734e-01
-8.31168294e-01 -1.98019296e-01 -1.28275418e+00 3.50088358e-01
1.19417298e+00 -4.18421090e-01 -8.84782016e-01 4.66166407e-01
9.33982372e-01 -3.72814417e-01 -5.55502295e-01 5.87886095e-01
5.07541358e-01 -6.72819376e-01 1.20805418e+00 -9.53598857e-01
-8.29809979e-02 -5.56484044e-01 -8.55241939e-02 -9.46481764e-01
-6.26209855e-01 -2.27655366e-01 -4.05079037e-01 1.21907556e+00
5.89067698e-01 -6.19646549e-01 6.79882526e-01 4.14962113e-01
2.17862669e-02 -3.88523936e-01 -1.02403581e+00 -1.48469663e+00
-5.03599346e-02 -4.94473338e-01 4.64146733e-01 9.50293303e-01
-2.76261926e-01 3.69986176e-01 -1.66165642e-02 3.04561764e-01
4.64389503e-01 -1.31321788e-01 7.30255425e-01 -1.56021452e+00
3.09127793e-02 -3.32065314e-01 -1.04359567e+00 -6.57519162e-01
-3.95471007e-01 -1.03802013e+00 2.87981212e-01 -1.27929270e+00
5.80446601e-01 -5.73925197e-01 -7.63265967e-01 5.91448128e-01
-3.19666088e-01 5.54889441e-01 3.64085495e-01 1.07495815e-01
-1.21382785e+00 6.34229779e-02 5.22247612e-01 -2.54293531e-01
-2.73234278e-01 2.33564004e-02 -7.52070248e-01 6.73005521e-01
1.23471938e-01 -9.57115173e-01 -2.86429767e-02 -3.30825448e-01
9.40480456e-02 -7.16479540e-01 4.16480094e-01 -1.62416387e+00
1.73509896e-01 -2.78249034e-03 5.36975443e-01 -8.58122766e-01
-1.22099310e-01 -8.23952496e-01 1.04780914e-02 9.04962778e-01
-3.22119892e-01 -1.46686971e-01 5.57822168e-01 9.47700143e-01
1.57527059e-01 -2.34570861e-01 1.04270411e+00 -1.03631288e-01
-1.55675447e+00 1.23504058e-01 -4.17914093e-01 2.78830916e-01
1.57727814e+00 -3.84451032e-01 -5.54938376e-01 3.99884403e-01
-6.98121667e-01 3.96583110e-01 1.15165256e-01 1.03096735e+00
2.23746791e-01 -1.14694345e+00 -6.90149128e-01 3.01533222e-01
1.03576922e+00 -1.66609678e-02 -5.74855432e-02 2.26897120e-01
-5.34448326e-01 3.22321713e-01 -1.32249951e-01 -8.19907069e-01
-1.37968898e+00 7.14919865e-01 3.86007875e-01 -3.60750668e-02
-7.46291637e-01 1.09852207e+00 1.06811598e-01 -6.47133946e-01
4.45866048e-01 -1.67545229e-01 6.11475669e-02 2.22637177e-01
8.52842391e-01 7.82177448e-01 3.93008679e-01 -3.75093490e-01
-7.39494085e-01 1.11485884e-01 -7.97709286e-01 5.59078276e-01
1.35565448e+00 1.54676750e-01 1.03600353e-01 4.90688622e-01
1.10077631e+00 -4.56717521e-01 -1.03641772e+00 -7.19256282e-01
6.53631568e-01 -5.46864986e-01 4.58825473e-03 -1.09374082e+00
-2.60113627e-01 3.49638283e-01 1.42722976e+00 3.22153628e-01
6.66797578e-01 5.04069388e-01 6.70715928e-01 9.36994493e-01
4.66377497e-01 -1.16282034e+00 2.28271306e-01 8.14407349e-01
5.71653008e-01 -1.50170529e+00 1.97053000e-01 -3.22390854e-01
-2.48006016e-01 9.11147058e-01 1.15172994e+00 8.36723074e-02
7.36834764e-01 2.28116930e-01 1.28981546e-01 -3.34031820e-01
-7.47246385e-01 -4.70793724e-01 3.16243023e-01 5.81537962e-01
-1.44333586e-01 1.68998986e-01 1.95118144e-01 1.00928791e-01
2.60408700e-01 -1.07164957e-01 3.97775084e-01 1.29253161e+00
-9.72395003e-01 -5.38754463e-01 -4.52758789e-01 9.21868742e-01
-2.41224900e-01 5.81499673e-02 -3.98242116e-01 6.41791582e-01
3.54393154e-01 6.60423458e-01 5.25422573e-01 -3.10171813e-01
6.02188885e-01 2.08544746e-01 2.94970777e-02 -1.09526968e+00
-4.24225152e-01 -4.31408614e-01 -4.82574031e-02 -4.39938515e-01
-5.71215153e-02 -6.55094922e-01 -1.19216716e+00 1.60991605e-02
-7.44328022e-01 -1.55201063e-01 2.53496170e-01 9.57205713e-01
4.14346784e-01 4.58731830e-01 4.64007735e-01 -4.74951714e-01
-8.25273514e-01 -6.74260557e-01 -5.01969099e-01 5.55771589e-01
5.65729082e-01 -8.93380225e-01 -2.69468963e-01 -7.73466676e-02] | [9.444011688232422, 1.5577353239059448] |
bc24486f-b32d-40a5-8f5b-6e3138a99371 | distributed-submodular-cover-succinctly | null | null | http://papers.nips.cc/paper/5752-distributed-submodular-cover-succinctly-summarizing-massive-data | http://papers.nips.cc/paper/5752-distributed-submodular-cover-succinctly-summarizing-massive-data.pdf | Distributed Submodular Cover: Succinctly Summarizing Massive Data | How can one find a subset, ideally as small as possible, that well represents a massive dataset? I.e., its corresponding utility, measured according to a suitable utility function, should be comparable to that of the whole dataset. In this paper, we formalize this challenge as a submodular cover problem. Here, the utility is assumed to exhibit submodularity, a natural diminishing returns condition preva- lent in many data summarization applications. The classical greedy algorithm is known to provide solutions with logarithmic approximation guarantees compared to the optimum solution. However, this sequential, centralized approach is imprac- tical for truly large-scale problems. In this work, we develop the first distributed algorithm – DISCOVER – for submodular set cover that is easily implementable using MapReduce-style computations. We theoretically analyze our approach, and present approximation guarantees for the solutions returned by DISCOVER. We also study a natural trade-off between the communication cost and the num- ber of rounds required to obtain such a solution. In our extensive experiments, we demonstrate the effectiveness of our approach on several applications, includ- ing active set selection, exemplar based clustering, and vertex cover on tens of millions of data points using Spark. | ['Ashwinkumar Badanidiyuru', 'Baharan Mirzasoleiman', 'Andreas Krause', 'Amin Karbasi'] | 2015-12-01 | null | null | null | neurips-2015-12 | ['data-summarization'] | ['miscellaneous'] | [ 4.85122316e-02 2.84255922e-01 -1.71269253e-01 -2.04254106e-01
-1.03482616e+00 -8.79406393e-01 -2.80301332e-01 7.65591800e-01
-6.10830113e-02 9.95378077e-01 -4.83938940e-02 7.80477673e-02
-5.57640254e-01 -1.22015262e+00 -1.00997949e+00 -7.80596495e-01
-3.54284823e-01 9.03266609e-01 1.16337694e-01 -1.95243239e-01
5.15474156e-02 2.87408859e-01 -1.72700691e+00 7.20578507e-02
9.01970208e-01 9.68447506e-01 1.14231370e-01 3.39885950e-01
1.86465308e-02 3.38279873e-01 -6.09195948e-01 -2.38104269e-01
4.82115835e-01 -2.54080921e-01 -7.54403412e-01 4.39340055e-01
1.27081886e-01 -1.35632411e-01 -1.93005487e-01 9.01114345e-01
4.79826719e-01 -6.19271994e-02 2.74878383e-01 -1.51367652e+00
-9.12126005e-02 9.89049077e-01 -7.90732741e-01 1.07057549e-01
3.59321952e-01 -1.08648352e-01 1.19737053e+00 -3.48907173e-01
8.53722990e-01 9.99461353e-01 4.78140503e-01 5.79164959e-02
-1.47602165e+00 -2.92809695e-01 3.02617755e-02 9.84627102e-03
-1.62552774e+00 -2.72220790e-01 5.39411902e-01 3.63245606e-02
6.52252078e-01 9.26761746e-01 7.00689554e-01 1.65763199e-01
4.41419473e-03 8.88362527e-01 7.69776821e-01 -1.95549384e-01
5.17812312e-01 1.71597123e-01 2.84572631e-01 2.85357058e-01
1.01392341e+00 -3.36967200e-01 -5.79311013e-01 -7.80998766e-01
7.95530602e-02 8.65399316e-02 -4.03650671e-01 -5.53939998e-01
-9.63286519e-01 6.88358665e-01 1.84714466e-01 3.67099009e-02
-6.17947876e-01 3.81547004e-01 4.25384223e-01 6.44286275e-01
5.73154211e-01 3.88811767e-01 -2.46815205e-01 1.07832916e-01
-9.81373072e-01 4.32097167e-01 1.19780934e+00 1.08246434e+00
8.45096469e-01 -4.12676901e-01 -5.35955131e-02 3.23611528e-01
-2.59826273e-01 1.89459503e-01 -2.83077270e-01 -8.41058314e-01
4.55374479e-01 8.20519745e-01 2.41758376e-01 -1.06820345e+00
-4.16621268e-01 -3.55920196e-01 -8.86267960e-01 -1.83665052e-01
3.06202143e-01 -1.90885112e-01 -2.03219891e-01 1.57246780e+00
9.02780294e-01 -1.33624032e-01 -1.43133020e-02 7.27111161e-01
5.69198489e-01 8.07013094e-01 -5.22370994e-01 -1.02434993e+00
1.00403047e+00 -3.85982752e-01 -5.87042630e-01 2.57630080e-01
7.94504821e-01 -4.88896459e-01 7.35421300e-01 4.84210402e-01
-1.31068218e+00 2.66757041e-01 -1.01758599e+00 2.63674766e-01
2.25576282e-01 -4.99407262e-01 7.09938109e-01 6.77598834e-01
-1.09336495e+00 3.59090239e-01 -5.57247996e-01 -3.63342762e-01
3.60251874e-01 5.48703372e-01 -1.79276049e-01 -3.74376148e-01
-6.04069650e-01 3.29773813e-01 5.98773837e-01 -3.51714224e-01
-7.28164792e-01 -7.35504925e-01 -3.57906103e-01 3.11088622e-01
9.13183510e-01 -8.16380739e-01 7.91586459e-01 -6.68176115e-01
-6.52548432e-01 7.59200513e-01 -2.33849823e-01 -5.48313498e-01
5.17136455e-01 4.28925365e-01 2.11179882e-01 1.97289407e-01
-1.02643467e-01 2.78203916e-02 2.22234607e-01 -1.33812249e+00
-6.00741208e-01 -6.84836209e-01 4.41454858e-01 3.03798884e-01
-7.13940382e-01 -6.49667457e-02 -3.30936968e-01 -2.59832919e-01
2.42811307e-01 -7.88909376e-01 -7.57210970e-01 -3.43421027e-02
-4.66787428e-01 -4.52363759e-01 4.03421968e-01 -7.08026662e-02
1.38796842e+00 -2.08417010e+00 1.18930653e-01 6.36567891e-01
4.84001756e-01 -3.59964639e-01 8.61754566e-02 9.04086649e-01
4.74005967e-01 7.92080835e-02 -5.22000194e-01 -1.43469989e-01
-3.91612872e-02 3.80982816e-01 -2.50135958e-01 8.89901400e-01
-3.81510973e-01 7.21478343e-01 -6.73756957e-01 -2.11821899e-01
-4.43419188e-01 -1.17591321e-01 -5.56682408e-01 -9.65362489e-02
-4.89884257e-01 -2.63190448e-01 -4.11966711e-01 6.76291823e-01
1.09770286e+00 -3.61003369e-01 5.72881043e-01 3.05967540e-01
1.93645675e-02 -1.73816606e-02 -1.59521759e+00 1.29911637e+00
-1.78906694e-01 2.02496588e-01 5.55099547e-01 -1.20210254e+00
9.41715360e-01 1.30167738e-01 9.96660173e-01 -2.55205452e-01
-1.04389060e-02 5.16233504e-01 -3.56764585e-01 -7.18323290e-02
5.51669121e-01 1.43715337e-01 -1.75380379e-01 1.07326353e+00
-5.45191348e-01 -6.84781969e-02 5.28392017e-01 5.15882492e-01
1.42917109e+00 -7.33401120e-01 4.82666135e-01 -7.07711041e-01
1.18386969e-01 4.46761906e-01 6.02401078e-01 8.10704648e-01
3.31340641e-01 5.22017777e-01 1.04367030e+00 -5.41849673e-01
-8.84676278e-01 -7.13228285e-01 -1.06492698e-01 7.84718513e-01
5.97423613e-01 -7.48986304e-01 -8.51026833e-01 -4.30535495e-01
3.58581841e-01 5.00641763e-01 -4.27699924e-01 7.62544349e-02
-3.04869503e-01 -1.08212042e+00 -7.97696412e-03 3.60069126e-02
-7.92717189e-02 -7.16603577e-01 -8.07800412e-01 3.56273323e-01
2.01208461e-02 -8.56691539e-01 -4.64153051e-01 1.38673991e-01
-1.01074767e+00 -1.19251442e+00 -2.07991645e-01 -5.45277655e-01
8.45450163e-01 6.62575722e-01 1.16893566e+00 3.01711440e-01
-2.85812616e-01 4.21662569e-01 -2.82595903e-01 -5.53886294e-01
-1.23379059e-01 -2.00890144e-03 6.38635755e-02 7.67981773e-03
2.49729395e-01 -6.85640812e-01 -5.45358121e-01 2.30225489e-01
-1.20267236e+00 -2.52968669e-01 2.52852559e-01 3.50711077e-01
1.27899325e+00 6.49218857e-01 6.81923032e-01 -1.02244329e+00
7.66400874e-01 -8.71558964e-01 -8.24057579e-01 5.26461720e-01
-7.31799662e-01 -1.11466773e-01 5.83308756e-01 -1.87466517e-01
-3.31580222e-01 4.43806350e-01 1.33844033e-01 -9.52218994e-02
4.64544445e-01 6.00440741e-01 -2.63894677e-01 3.55964005e-02
5.31925678e-01 3.54609668e-01 1.21720232e-01 -2.80019999e-01
3.41902018e-01 6.87406301e-01 3.20324063e-01 -6.78288221e-01
7.31218755e-01 9.84913766e-01 3.40519071e-01 -7.10556030e-01
-6.33455634e-01 -6.76017046e-01 -9.72033888e-02 -1.28464609e-01
-7.36012459e-02 -7.33889878e-01 -1.07711422e+00 5.60082532e-02
-1.03078306e+00 1.44459471e-01 -6.54291868e-01 -3.36525925e-02
-6.46450639e-01 2.59991318e-01 6.00134283e-02 -1.02187800e+00
-7.23063767e-01 -9.23358619e-01 7.05331385e-01 -1.32332593e-01
2.68183295e-02 -4.80748862e-01 5.00437915e-02 8.35647509e-02
1.69489697e-01 8.59447896e-01 7.11588144e-01 -6.58257902e-01
-9.63411152e-01 -4.05306578e-01 6.31612353e-03 -1.41862631e-01
-9.63501856e-02 -2.74769187e-01 -5.17862201e-01 -7.04987288e-01
8.43224078e-02 -2.62001395e-01 9.65888739e-01 5.40307224e-01
1.57175088e+00 -7.99971938e-01 -3.88521671e-01 6.40158713e-01
1.58599865e+00 -2.88805962e-01 4.42819774e-01 1.17083460e-01
3.31419170e-01 6.81153834e-01 7.68159747e-01 1.07482052e+00
5.79591095e-01 5.21312594e-01 6.23176396e-01 2.09583431e-01
4.42456126e-01 7.81254619e-02 1.56523043e-03 5.63621402e-01
5.13338596e-02 -5.66698253e-01 -6.90140665e-01 7.03701913e-01
-2.05544400e+00 -8.28302979e-01 -5.18148959e-01 2.66223931e+00
8.27159524e-01 -5.43540344e-03 6.72300458e-01 4.37676340e-01
6.72873497e-01 -1.64041400e-01 -6.52911842e-01 -4.07986820e-01
-4.05846030e-01 -1.25767961e-01 8.87127340e-01 2.35835239e-01
-7.11122811e-01 2.99311072e-01 5.85088968e+00 8.66144776e-01
-7.44955957e-01 1.97251379e-01 8.81571412e-01 -7.28832960e-01
-9.83789146e-01 -1.98362209e-03 -5.95241487e-01 3.38042527e-01
9.03165162e-01 -1.03258169e+00 6.00440085e-01 1.00398469e+00
-3.76852020e-03 -2.41380632e-01 -1.14856255e+00 1.01303530e+00
-1.55304626e-01 -1.70484161e+00 -2.60616273e-01 5.07814407e-01
1.09056842e+00 -1.62215546e-01 -1.27361655e-01 -2.59631127e-01
2.19425425e-01 -8.38746786e-01 6.64466441e-01 5.47558032e-02
9.16401207e-01 -1.33212960e+00 6.03877962e-01 6.79960430e-01
-1.12317491e+00 -2.85958469e-01 -6.53829634e-01 -4.17642482e-02
2.01057196e-01 1.12828684e+00 -6.15676284e-01 6.30539715e-01
7.58714080e-01 1.80731997e-01 -1.68497767e-02 1.30111396e+00
5.10271311e-01 4.20666516e-01 -1.07741463e+00 -2.84868121e-01
1.66887008e-02 -1.65241286e-01 8.48825097e-01 9.09274161e-01
3.28131646e-01 3.85630071e-01 3.86926502e-01 4.54504490e-01
-5.21702588e-01 4.17042851e-01 -6.47786558e-01 4.13587950e-02
8.60942185e-01 1.16023850e+00 -8.31817508e-01 -1.68582290e-01
-1.86217539e-02 5.40909946e-01 3.12197268e-01 -4.22520265e-02
-6.54861450e-01 -3.35353196e-01 6.50762498e-01 6.06414557e-01
2.37135097e-01 -6.54010922e-02 -7.30897844e-01 -7.77030826e-01
5.57243109e-01 -8.39673758e-01 8.08410525e-01 1.97496310e-01
-1.06913590e+00 3.98306876e-01 9.20723900e-02 -9.86485898e-01
8.39499012e-02 2.43123565e-02 -4.86277938e-01 5.41781962e-01
-1.13775182e+00 -5.85456133e-01 -3.06856781e-01 4.57183838e-01
1.04101121e-01 2.80664623e-01 4.55439717e-01 6.28384724e-02
-4.00096238e-01 6.93669260e-01 5.75271666e-01 -7.79859066e-01
1.19084738e-01 -1.17103982e+00 1.51857331e-01 7.58796096e-01
-5.98707907e-02 3.92739117e-01 9.20881927e-01 -5.04745185e-01
-2.09743834e+00 -1.08319390e+00 8.43476057e-01 -1.80243686e-01
3.98594290e-01 -3.63030165e-01 -7.92241395e-01 3.88885915e-01
-2.08690807e-01 1.64334834e-01 5.63201427e-01 9.94751900e-02
7.39820600e-02 -5.63836813e-01 -1.54338109e+00 1.63039848e-01
1.28588426e+00 1.78726628e-01 1.41474549e-02 7.67183125e-01
8.78001392e-01 -3.84246767e-01 -8.63974214e-01 2.76513398e-01
3.17725480e-01 -8.47081363e-01 9.22603905e-01 -3.82624000e-01
1.77854016e-01 -1.12542927e-01 -3.41371000e-01 -9.74719942e-01
-9.42116380e-02 -1.21029282e+00 -6.53843164e-01 1.02559578e+00
1.89967364e-01 -7.04908848e-01 9.28954661e-01 4.56578344e-01
1.93468302e-01 -1.26053834e+00 -1.18693352e+00 -1.05509639e+00
-1.42155111e-01 -2.08734751e-01 1.08668005e+00 7.70768404e-01
2.84290090e-02 2.93695759e-02 -1.97280422e-01 1.38556823e-01
1.06656516e+00 6.37416065e-01 9.70846772e-01 -1.39834023e+00
-3.21390718e-01 -1.38783902e-01 -2.87929505e-01 -8.25502455e-01
-1.71097517e-01 -9.86919224e-01 -2.65926450e-01 -1.71752214e+00
5.59675574e-01 -7.98000693e-01 -1.16800621e-01 3.11539978e-01
1.06571186e-02 2.01565586e-02 1.63765386e-01 3.41341734e-01
-9.28733587e-01 2.50405490e-01 1.03868246e+00 -3.92109789e-02
-6.05583727e-01 2.57433414e-01 -1.12605870e+00 2.68356442e-01
6.85627997e-01 -8.34259510e-01 -4.11271542e-01 -2.56697208e-01
5.94451785e-01 3.08831751e-01 -1.21931344e-01 -5.69880962e-01
3.43371302e-01 -5.49367726e-01 -3.12264860e-01 -8.78986955e-01
1.26574829e-01 -9.02836084e-01 5.69268882e-01 5.60878754e-01
-1.85003519e-01 -2.66827773e-02 -1.65588856e-01 7.97336340e-01
-6.58451580e-03 -1.48588926e-01 5.63585758e-01 3.02849747e-02
-1.93322241e-01 6.10507190e-01 8.91829357e-02 1.41372591e-01
1.67893553e+00 -3.16489398e-01 -3.10095549e-01 -5.11767387e-01
-1.85613096e-01 7.43331492e-01 7.38589942e-01 -1.48364887e-01
4.99596804e-01 -1.00959480e+00 -1.25919294e+00 -2.90077001e-01
2.81657696e-01 3.34253520e-01 2.51645595e-01 1.00178754e+00
-6.03252590e-01 2.75310963e-01 2.61684388e-01 -4.68187630e-01
-1.29452717e+00 8.87805104e-01 -1.55401066e-01 -2.47742087e-01
-6.57934010e-01 7.31170952e-01 4.35277708e-02 -1.89868569e-01
9.98834223e-02 2.76013743e-02 3.78668517e-01 2.43780494e-01
6.62900865e-01 7.89039254e-01 2.81276882e-01 -2.59618372e-01
-4.26698565e-01 -2.52561290e-02 2.06266344e-01 2.17325881e-01
1.75708890e+00 -1.73641622e-01 -6.74691081e-01 9.06718671e-02
1.12729633e+00 1.48139875e-02 -8.16781223e-01 -2.74292946e-01
-7.88602680e-02 -6.33526266e-01 -1.25384871e-02 -2.94444293e-01
-1.04886460e+00 -1.64486337e-02 3.38844806e-02 8.15865934e-01
1.43574643e+00 4.27395135e-01 8.65307927e-01 6.09758019e-01
8.95310581e-01 -1.05331361e+00 -5.21190047e-01 -6.84038922e-02
9.26112771e-01 -9.50592458e-01 3.99971157e-01 -5.71722567e-01
-4.09708709e-01 1.02986753e+00 3.55166614e-01 -1.48913041e-01
5.49138188e-01 4.41038191e-01 -6.36863887e-01 -3.70187163e-01
-1.19818532e+00 -1.50311366e-01 -5.73185757e-02 1.62642419e-01
-1.84802532e-01 5.61331034e-01 -8.25574636e-01 5.78307509e-01
-2.68203020e-01 -1.02599099e-01 7.47019649e-01 9.56013620e-01
-8.55080068e-01 -8.24101329e-01 -7.85042524e-01 9.01988149e-01
-4.06175703e-01 1.85199112e-01 -5.13622880e-01 4.26641315e-01
-2.65913993e-01 1.11369944e+00 2.98026532e-01 -2.85592020e-01
2.27815613e-01 -5.53085983e-01 4.31241840e-01 -5.49527049e-01
-5.39672494e-01 -2.70541962e-02 4.70922627e-02 -4.92522448e-01
7.41336346e-02 -6.19580984e-01 -1.32639372e+00 -8.30006123e-01
-3.19218904e-01 2.77240336e-01 5.96415222e-01 6.56498790e-01
3.58478576e-01 2.63056289e-02 1.05848074e+00 -2.47121051e-01
-1.01635480e+00 -3.47986430e-01 -8.70616019e-01 1.15853027e-01
-7.49007473e-03 -1.57877848e-01 -4.00094479e-01 -4.65822458e-01] | [6.620212554931641, 4.928375244140625] |
5e764ee8-3268-474f-b1fd-d4d1274a514d | decoupling-recognition-from-detection-single | 2207.07253 | null | https://arxiv.org/abs/2207.07253v4 | https://arxiv.org/pdf/2207.07253v4.pdf | Single Shot Self-Reliant Scene Text Spotter by Decoupled yet Collaborative Detection and Recognition | Typical text spotters follow the two-stage spotting paradigm which detects the boundary for a text instance first and then performs text recognition within the detected regions. Despite the remarkable progress of such spotting paradigm, an important limitation is that the performance of text recognition depends heavily on the precision of text detection, resulting in the potential error propagation from detection to recognition. In this work, we propose the single shot Self-Reliant Scene Text Spotter v2 (SRSTS v2), which circumvents this limitation by decoupling recognition from detection while optimizing two tasks collaboratively. Specifically, our SRSTS v2 samples representative feature points around each potential text instance, and conducts both text detection and recognition in parallel guided by these sampled points. Thus, the text recognition is no longer dependent on detection, thereby alleviating the error propagation from detection to recognition. Moreover, the sampling module is learned under the supervision from both detection and recognition, which allows for the collaborative optimization and mutual enhancement between two tasks. Benefiting from such sampling-driven concurrent spotting framework, our approach is able to recognize the text instances correctly even if the precise text boundaries are challenging to detect. Extensive experiments on four benchmarks demonstrate that our method compares favorably to state-of-the-art spotters. | ['Wenjie Pei', 'Chengquan Zhang', 'Guangming Lu', 'Pengyuan Lyu', 'Jingjing Wu'] | 2022-07-15 | null | null | null | null | ['text-spotting'] | ['computer-vision'] | [ 6.34270787e-01 -4.33166325e-01 -8.17476362e-02 -9.79012698e-02
-9.04128373e-01 -2.55938917e-01 7.32123017e-01 1.34071797e-01
-3.47313493e-01 1.81955442e-01 -1.24476478e-01 8.98127034e-02
8.17150623e-02 -7.30539322e-01 -4.92057174e-01 -8.92310262e-01
5.61471522e-01 6.13201082e-01 6.22967005e-01 1.69562772e-01
5.52052796e-01 4.09154683e-01 -1.61605906e+00 3.17198515e-01
1.06716323e+00 1.01696789e+00 5.44129193e-01 7.13823318e-01
-6.32048488e-01 3.95653993e-01 -6.75658166e-01 -1.21671776e-03
1.57925054e-01 -2.76606649e-01 -2.53973275e-01 5.59789300e-01
5.91146946e-01 -3.03711057e-01 -2.39475742e-01 8.48970532e-01
4.14808929e-01 1.64079279e-01 4.67547178e-01 -8.91219854e-01
-2.66158462e-01 5.33063710e-01 -1.03551590e+00 -1.39876515e-01
4.00748760e-01 2.35063896e-01 1.23491240e+00 -1.37176800e+00
4.13593680e-01 8.23196590e-01 5.69385946e-01 4.03933614e-01
-1.14054751e+00 -4.54107434e-01 4.70669359e-01 -9.74188838e-03
-1.66143799e+00 -6.65263116e-01 7.72710919e-01 -4.63114977e-01
7.35645354e-01 5.40305912e-01 4.68307555e-01 8.34570765e-01
-2.45668098e-01 1.36584306e+00 6.22124910e-01 -5.57492137e-01
2.38005310e-01 1.63077906e-01 4.01407421e-01 7.80074954e-01
1.46098688e-01 -2.67022163e-01 -7.45253146e-01 7.37546682e-02
5.87758243e-01 5.16048431e-01 -4.95696723e-01 -3.61248821e-01
-1.48070455e+00 2.50366986e-01 1.44278079e-01 4.98049557e-01
-1.97593778e-01 -2.36807436e-01 3.92123938e-01 1.08785935e-01
5.19844711e-01 8.46872032e-02 -2.91746333e-02 -8.70212093e-02
-1.65622485e+00 -2.18240604e-01 6.09997451e-01 9.36725318e-01
8.57882142e-01 -1.86775267e-01 -6.00286186e-01 9.47874069e-01
2.91735888e-01 7.28714526e-01 4.72743958e-01 2.84872293e-01
8.56840491e-01 1.22966444e+00 5.07906452e-02 -9.12158668e-01
-1.46658421e-01 -3.66517365e-01 -9.18156087e-01 -6.05243444e-03
4.17063564e-01 2.11950373e-02 -1.05218649e+00 1.01143372e+00
4.18739200e-01 1.66578412e-01 -2.00654611e-01 1.03978407e+00
5.54944217e-01 7.79399753e-01 -3.95620823e-01 -7.05626085e-02
1.26419699e+00 -1.19753730e+00 -4.58866328e-01 -4.52227324e-01
7.30936229e-01 -8.47313881e-01 1.13813961e+00 3.19970638e-01
-6.25140965e-01 -4.24895048e-01 -1.08102095e+00 4.56979387e-02
-1.64955303e-01 8.17056060e-01 8.20946470e-02 5.05304098e-01
-7.26586103e-01 2.45406538e-01 -8.97596002e-01 -5.50918996e-01
3.89904022e-01 3.03831398e-01 -1.01248153e-01 7.26295337e-02
-6.04875147e-01 3.41502637e-01 2.56746113e-01 3.75840306e-01
-7.29812503e-01 -3.11139971e-01 -5.73780715e-01 3.02456766e-01
7.61914253e-01 -2.23131582e-01 9.57381904e-01 -9.38043594e-01
-1.56402373e+00 6.93717957e-01 -6.30092978e-01 -1.08492784e-01
9.33198810e-01 -4.13317740e-01 -1.91127121e-01 1.07933823e-02
2.32836440e-01 1.58244178e-01 1.27250814e+00 -1.02837443e+00
-8.69792163e-01 -4.95093435e-01 -5.21054029e-01 4.21453029e-01
-6.37587607e-01 -3.05829365e-02 -1.11220944e+00 -5.00703216e-01
3.42811048e-01 -5.07281601e-01 7.01725259e-02 2.05464333e-01
-8.49541605e-01 -3.61770213e-01 1.30055499e+00 -3.64460677e-01
1.33716595e+00 -2.33916974e+00 9.75781009e-02 1.19843185e-01
3.47555161e-01 2.48851731e-01 -4.36908193e-02 5.40627897e-01
2.85936594e-01 -2.14383498e-01 -1.70265675e-01 -9.67630744e-01
3.10018156e-02 -1.76283181e-01 -6.65833235e-01 6.35651112e-01
6.18503578e-02 8.27063501e-01 -6.92345202e-01 -6.50524199e-01
4.90074396e-01 2.27586567e-01 -6.05744086e-02 3.36022139e-01
-4.18079019e-01 1.02346107e-01 -6.80142105e-01 9.15691674e-01
6.96277320e-01 -4.70160574e-01 4.37241867e-02 8.70174468e-02
-3.65278959e-01 1.10375300e-01 -1.41972411e+00 1.52433717e+00
-2.54542440e-01 7.04758584e-01 2.91854858e-01 -1.02867329e+00
1.08675444e+00 8.17495063e-02 3.15081477e-01 -5.59797227e-01
-2.94062421e-02 1.64385423e-01 -5.68561316e-01 -3.05846691e-01
6.58309400e-01 2.75709033e-01 1.19291522e-01 7.38166034e-01
-4.18933213e-01 9.38688293e-02 1.03235826e-01 2.60152936e-01
9.99200463e-01 1.65486351e-01 6.36206418e-02 7.88568854e-02
7.14500904e-01 1.40077472e-01 4.75865334e-01 9.58686769e-01
-9.88220349e-02 7.18658626e-01 2.77039349e-01 -2.65097588e-01
-7.68302917e-01 -7.70088673e-01 -1.46978214e-01 1.25520003e+00
6.05749130e-01 -4.94366467e-01 -7.87515819e-01 -9.82884288e-01
-2.76440498e-03 3.45296055e-01 -6.35687053e-01 2.56875306e-01
-6.31509006e-01 -6.60147905e-01 4.94055361e-01 4.48825389e-01
5.66364944e-01 -9.19272304e-01 -6.37648344e-01 -1.30906887e-02
-1.47272237e-02 -9.18524742e-01 -8.79339576e-01 1.67296112e-01
-7.74553001e-01 -1.00883603e+00 -9.77707088e-01 -7.85358548e-01
8.43682528e-01 7.85189390e-01 5.92048407e-01 3.24176043e-01
-5.43507278e-01 2.76793748e-01 -4.47020411e-01 1.07695386e-02
-1.71692193e-01 2.94019759e-01 -2.03349784e-01 6.62945330e-01
2.31949270e-01 -9.57625918e-03 -4.75763261e-01 6.55484438e-01
-8.77036989e-01 3.89607996e-01 6.95475280e-01 9.54486668e-01
5.94740093e-01 -9.96340998e-03 1.53656185e-01 -7.62232780e-01
2.93477207e-01 -2.06376225e-01 -8.85542333e-01 5.81445515e-01
-6.72096372e-01 -7.05292225e-02 7.61365473e-01 -3.36971432e-01
-9.59869981e-01 3.10115010e-01 3.19474518e-01 -4.37726736e-01
-2.77772069e-01 1.93912968e-01 -1.53775886e-01 1.13063850e-01
4.99150008e-01 1.04553568e+00 -4.29521315e-02 -5.04495859e-01
7.66177699e-02 9.85137105e-01 1.69913828e-01 -2.71960378e-01
9.94459927e-01 6.33163273e-01 -4.04417455e-01 -1.24377060e+00
-5.16865313e-01 -9.59796190e-01 -6.97273254e-01 -2.49259025e-01
6.05969965e-01 -7.01749802e-01 -5.63159287e-01 8.06370854e-01
-1.08031654e+00 -3.03034782e-01 -1.05708316e-01 1.35455415e-01
-1.86742440e-01 6.85370445e-01 -4.26035255e-01 -1.09788561e+00
-6.01086795e-01 -1.21415067e+00 1.75299764e+00 1.82432368e-01
1.82638466e-01 -7.92580903e-01 -1.09887771e-01 2.58722574e-01
3.27520877e-01 -3.20556521e-01 3.84942085e-01 -6.68153107e-01
-1.01346755e+00 -6.79032505e-01 -5.19610107e-01 -2.01543406e-01
2.22705677e-01 1.17881671e-01 -1.06843245e+00 -4.06197816e-01
-1.39169931e-01 -1.94573373e-01 1.08910847e+00 1.14302784e-01
9.24538851e-01 5.35257608e-02 -7.34550178e-01 5.52085102e-01
1.37414217e+00 1.96189098e-02 3.87254149e-01 2.63351738e-01
9.10088897e-01 4.19806808e-01 6.96243703e-01 5.36371768e-01
1.10485721e-02 8.81882966e-01 2.41690397e-01 -3.04064006e-01
-1.10176420e-02 -2.41402194e-01 4.85432506e-01 4.73503381e-01
5.36641657e-01 -4.90232736e-01 -9.01647031e-01 3.73355627e-01
-2.16741204e+00 -7.94371426e-01 -1.93748653e-01 2.28707290e+00
6.11513555e-01 -1.56418066e-02 5.63878231e-02 3.12257022e-01
1.26112115e+00 4.70609695e-01 -7.88399756e-01 2.63363957e-01
-9.03802142e-02 -1.83596268e-01 3.16528946e-01 4.01061833e-01
-1.13219655e+00 1.12099016e+00 5.47303820e+00 1.10269105e+00
-1.39906752e+00 -2.19718575e-01 4.39350128e-01 -1.97465092e-01
1.29429922e-01 -1.14430105e-02 -1.13890219e+00 4.79900330e-01
1.86069444e-01 -2.67712194e-02 3.95867497e-01 8.26448500e-01
2.08467156e-01 -2.56458431e-01 -1.09976041e+00 1.13655376e+00
2.94133902e-01 -1.21012342e+00 1.11883827e-01 -1.38527393e-01
4.85491723e-01 -1.80174280e-02 -1.00724913e-01 1.07345648e-01
-9.83207822e-02 -6.78722441e-01 7.92091429e-01 4.30070937e-01
1.00884974e+00 -2.91976780e-01 3.52488607e-01 6.89910948e-01
-1.51492131e+00 5.36705777e-02 -1.71464279e-01 2.42630124e-01
3.16969045e-02 1.06655407e+00 -1.23717844e+00 3.95015836e-01
2.95645952e-01 8.68383884e-01 -4.98253554e-01 9.52836812e-01
-2.03428879e-01 5.85366666e-01 -4.81728166e-01 -4.20454383e-01
9.18619260e-02 -1.50359273e-01 7.68318594e-01 1.41581726e+00
2.83655465e-01 -3.04516166e-01 5.08934736e-01 1.08220673e+00
-1.23147294e-02 3.72864962e-01 -3.63654226e-01 -6.41509145e-02
5.48498869e-01 1.35842478e+00 -1.18241513e+00 -5.37214220e-01
-2.81709969e-01 1.43073773e+00 3.32613498e-01 2.94268727e-01
-7.23434150e-01 -7.98231661e-01 1.94459304e-01 3.81882377e-02
4.97083515e-01 -9.82522964e-02 -7.13981986e-01 -1.54138196e+00
5.41850269e-01 -6.25798464e-01 1.43464461e-01 -4.49131310e-01
-9.78220761e-01 3.99966300e-01 -6.61186635e-01 -1.28730822e+00
3.18594575e-01 -4.67088431e-01 -8.73969376e-01 8.86987925e-01
-1.29055905e+00 -1.06064188e+00 -6.27324939e-01 5.44349968e-01
1.03090608e+00 6.32736981e-02 4.58513051e-01 7.82646015e-02
-1.22882950e+00 9.64394510e-01 3.01056027e-01 3.36343616e-01
8.49900186e-01 -1.16719794e+00 7.06206560e-01 1.03692985e+00
2.19278246e-01 3.72960716e-01 1.80687487e-01 -8.37691963e-01
-1.80069768e+00 -1.07816195e+00 7.25054383e-01 -3.15162688e-01
5.70186019e-01 -8.09122801e-01 -1.02577233e+00 3.04226756e-01
-2.31655791e-01 -2.02814206e-01 1.76322937e-01 1.02269389e-01
-3.03935260e-01 -3.15945506e-01 -8.49066556e-01 7.81480849e-01
9.49458122e-01 -5.01696646e-01 -4.52143580e-01 5.12897551e-01
3.30885887e-01 -3.28325480e-01 -1.97888583e-01 -4.30352846e-03
3.78726482e-01 -9.93425608e-01 6.66611552e-01 3.01039100e-01
2.37049639e-01 -4.68141198e-01 5.49453497e-02 -8.08759809e-01
-1.02504138e-02 -6.76940441e-01 -1.13420025e-01 1.34348190e+00
4.72194165e-01 -6.83509231e-01 1.01813340e+00 3.91019672e-01
-5.03899753e-02 -6.03354454e-01 -1.04657686e+00 -6.95376277e-01
-4.51806486e-01 -3.43380034e-01 5.00212133e-01 8.16496491e-01
1.68212891e-01 4.64322329e-01 -3.78339887e-01 2.32893169e-01
5.82814336e-01 6.42102599e-01 1.06169152e+00 -1.15535676e+00
-4.37233120e-01 -6.68477356e-01 -1.92796350e-01 -1.85384369e+00
-1.85546041e-01 -8.17659199e-01 6.07392073e-01 -1.31167495e+00
3.99691552e-01 -6.08478844e-01 6.60558790e-02 5.27363479e-01
-4.94300216e-01 2.77602933e-02 1.93346024e-01 7.49737084e-01
-9.52012002e-01 5.92611253e-01 1.08286083e+00 -3.57120574e-01
-4.30273086e-01 -8.56648013e-03 -4.02999967e-01 4.76888180e-01
3.55378568e-01 -2.97278702e-01 -1.53926715e-01 -6.22085035e-01
-1.06867395e-01 1.10928573e-01 1.49739891e-01 -9.20281410e-01
7.54176319e-01 -1.70701459e-01 3.85599166e-01 -9.37573254e-01
1.25991136e-01 -7.38318086e-01 -4.49063897e-01 2.69391537e-01
-2.17544377e-01 -4.29589897e-01 -1.33147824e-03 7.66369045e-01
1.20491153e-02 -2.14342088e-01 6.61264598e-01 2.30747789e-01
-6.37323380e-01 3.03122759e-01 -2.83458054e-01 -1.74665928e-01
1.08811951e+00 -6.47828341e-01 -2.36649498e-01 1.39010638e-01
-4.35991526e-01 4.05423939e-01 7.81198144e-01 2.05675066e-01
7.10021973e-01 -8.36896658e-01 -6.97138190e-01 6.21543705e-01
3.11631083e-01 3.89140666e-01 9.39715132e-02 1.10207665e+00
-2.89269716e-01 4.33340400e-01 5.98780036e-01 -1.16445863e+00
-1.25882077e+00 4.38573390e-01 2.01331973e-01 -4.27084148e-01
-1.11320949e+00 7.10781753e-01 4.07973886e-01 -1.23400532e-01
5.67750454e-01 -1.34157047e-01 2.19152179e-02 2.83589475e-02
7.93495595e-01 4.50249106e-01 1.44271359e-01 -3.53127092e-01
-4.06891644e-01 9.00698543e-01 -5.24495304e-01 -7.51154795e-02
9.89828646e-01 -2.75337130e-01 7.79456049e-02 5.53441584e-01
7.52104878e-01 1.71009436e-01 -1.31702173e+00 -4.26227033e-01
1.14349015e-01 -5.97024024e-01 7.46828690e-02 -7.48215795e-01
-9.39297438e-01 9.07039046e-01 3.68072391e-01 2.66189009e-01
1.15588009e+00 -3.00876886e-01 9.32838678e-01 5.58566570e-01
3.40346068e-01 -1.12253416e+00 9.50564444e-02 5.97136736e-01
5.73540568e-01 -1.20216370e+00 -7.32824653e-02 -4.41734254e-01
-4.86900598e-01 1.19613135e+00 6.42728269e-01 -1.49656504e-01
2.23053515e-01 4.57963914e-01 -6.20094761e-02 -1.40787259e-01
-6.94436133e-01 -5.15294373e-02 2.21720383e-01 9.05897245e-02
2.80165315e-01 5.80532439e-02 3.46733723e-03 2.24310726e-01
4.99550551e-01 -1.49544984e-01 1.08551897e-01 8.95031154e-01
-7.98112094e-01 -8.58481467e-01 -7.21125126e-01 5.24304509e-01
-3.16164270e-02 -2.90965229e-01 -8.18459690e-01 3.87671858e-01
-3.04843932e-01 8.23243022e-01 3.27895097e-02 -3.90371442e-01
3.27018887e-01 1.53399825e-01 -2.94968318e-02 -8.19865525e-01
-5.54193020e-01 4.05974090e-01 -3.75961930e-01 -4.66146618e-01
1.85456917e-01 -5.95514655e-01 -1.37911618e+00 -1.00793406e-01
-9.76847172e-01 1.58852473e-01 6.13868117e-01 1.02347982e+00
6.55214608e-01 3.10766309e-01 9.87252116e-01 -9.92778957e-01
-6.63433075e-01 -9.64926064e-01 -6.63726568e-01 1.96846873e-01
4.61303353e-01 -4.59529132e-01 -5.89717448e-01 -7.02400655e-02] | [11.997896194458008, 2.242135763168335] |
3eb371ed-3f57-4325-a3e1-8f124c86339b | very-deep-convolutional-neural-networks-for-1 | 1610.00277 | null | http://arxiv.org/abs/1610.00277v1 | http://arxiv.org/pdf/1610.00277v1.pdf | Very Deep Convolutional Neural Networks for Robust Speech Recognition | This paper describes the extension and optimization of our previous work on
very deep convolutional neural networks (CNNs) for effective recognition of
noisy speech in the Aurora 4 task. The appropriate number of convolutional
layers, the sizes of the filters, pooling operations and input feature maps are
all modified: the filter and pooling sizes are reduced and dimensions of input
feature maps are extended to allow adding more convolutional layers.
Furthermore appropriate input padding and input feature map selection
strategies are developed. In addition, an adaptation framework using joint
training of very deep CNN with auxiliary features i-vector and fMLLR features
is developed. These modifications give substantial word error rate reductions
over the standard CNN used as baseline. Finally the very deep CNN is combined
with an LSTM-RNN acoustic model and it is shown that state-level weighted log
likelihood score combination in a joint acoustic model decoding scheme is very
effective. On the Aurora 4 task, the very deep CNN achieves a WER of 8.81%,
further 7.99% with auxiliary feature joint training, and 7.09% with LSTM-RNN
joint decoding. | ['Philip C. Woodland', 'Yanmin Qian'] | 2016-10-02 | null | null | null | null | ['robust-speech-recognition'] | ['speech'] | [ 1.54055983e-01 7.38214925e-02 2.36397326e-01 -6.81568503e-01
-1.36431623e+00 -3.24308127e-01 4.58638191e-01 -4.36309487e-01
-9.28684413e-01 4.45698202e-01 3.78086239e-01 -4.19491619e-01
2.87916716e-02 -2.53198832e-01 -5.78160882e-01 -8.32828879e-01
3.90294045e-02 -1.39524162e-01 3.07731777e-01 -3.25099289e-01
-8.95586237e-02 5.89252830e-01 -1.58715487e+00 3.65052432e-01
1.52856499e-01 1.39518523e+00 4.61373031e-01 1.58523953e+00
2.22556174e-01 4.66081023e-01 -1.12079489e+00 -1.59792438e-01
1.48048237e-01 3.81119587e-02 -7.94095457e-01 -1.99448943e-01
7.62651265e-01 -2.61230439e-01 -6.39025211e-01 9.35440719e-01
1.10721660e+00 6.46773398e-01 2.50326186e-01 -5.23043692e-01
-3.02837044e-01 6.68567836e-01 4.11962681e-02 4.33985412e-01
-1.16300799e-01 2.04288021e-01 9.69830573e-01 -1.14943790e+00
-1.54536664e-01 1.39353287e+00 7.40774095e-01 6.79521561e-01
-7.68541217e-01 -4.61775273e-01 -1.30598053e-01 6.34324774e-02
-1.10923052e+00 -8.32398772e-01 2.70651430e-01 1.50401235e-01
2.00787759e+00 5.81384957e-01 1.89340666e-01 1.20010257e+00
1.84838735e-02 5.41327596e-01 7.20300674e-01 -6.44811153e-01
3.86603661e-02 1.59698039e-01 2.61367112e-01 3.55533004e-01
-2.86935091e-01 8.98522362e-02 -6.09083831e-01 -9.21650901e-02
4.73677278e-01 -5.31973779e-01 -8.32300708e-02 5.49489677e-01
-8.75235319e-01 6.07582271e-01 2.63085216e-01 4.75623846e-01
-6.46070763e-02 5.99505246e-01 8.08198690e-01 5.31883895e-01
5.13516605e-01 4.22168940e-01 -9.29848552e-01 -3.84198934e-01
-8.69443476e-01 1.51410192e-01 8.69422376e-01 9.12659764e-01
3.97676855e-01 7.97998786e-01 -2.50682771e-01 1.67900896e+00
3.81845921e-01 7.12437153e-01 7.23862290e-01 -1.01462877e+00
7.46373534e-01 -2.42541462e-01 -1.05568916e-01 -3.42560768e-01
-5.04421890e-01 -8.33399773e-01 -6.14642203e-01 1.27451420e-01
9.44515765e-02 -4.80123758e-01 -1.42250371e+00 1.60084486e+00
-4.03730303e-01 -1.79857947e-02 4.30750281e-01 5.75876832e-01
9.07444000e-01 9.45571244e-01 2.10521724e-02 1.83304176e-01
1.36759973e+00 -1.12648833e+00 -1.01115000e+00 -5.78564167e-01
7.26093113e-01 -1.09103477e+00 9.31447268e-01 3.57947439e-01
-1.36666393e+00 -9.65023279e-01 -1.43348563e+00 -3.87774676e-01
-7.66529560e-01 5.10671139e-01 3.30401570e-01 9.94238496e-01
-1.50883424e+00 6.50044203e-01 -7.10645318e-01 -7.65872002e-02
-3.37782018e-02 1.05011475e+00 -2.42010653e-01 1.06863238e-01
-1.51684737e+00 1.03766263e+00 6.54631615e-01 5.87285995e-01
-7.78617978e-01 -2.04946548e-01 -1.24268425e+00 2.38154933e-01
-1.08782962e-01 -1.50024995e-01 1.81509328e+00 -6.91673934e-01
-2.14163280e+00 4.73842502e-01 -1.00442708e-01 -8.98538888e-01
-5.46348989e-02 -3.42956603e-01 -8.15867305e-01 -1.52863368e-01
-5.01330793e-01 7.29741275e-01 7.88031340e-01 -4.97875273e-01
-7.36768961e-01 -2.34864671e-02 -3.09349537e-01 2.76142776e-01
-4.99673545e-01 5.48778057e-01 -6.30084515e-01 -6.02235556e-01
2.75225490e-02 -8.32900643e-01 -3.96172702e-01 -5.27468503e-01
-3.74328554e-01 -2.20086992e-01 9.61958647e-01 -9.88597453e-01
1.18266702e+00 -2.32559061e+00 -1.05743580e-01 3.09482127e-01
-3.01609397e-01 9.00693655e-01 -3.46393526e-01 7.93295726e-02
-1.38250321e-01 1.69637501e-01 -1.63675174e-01 -8.62865090e-01
-2.42483113e-02 3.76217157e-01 -8.96019116e-02 2.29838654e-01
3.56639147e-01 7.05782473e-01 -4.26569432e-01 9.19022970e-03
5.36742210e-01 8.93737793e-01 -2.92616099e-01 2.00069040e-01
1.34804040e-01 -2.13550076e-01 -3.32159642e-03 3.21493119e-01
7.58891582e-01 3.77230793e-01 -1.92422390e-01 -3.82886715e-02
-2.26032987e-01 7.52856731e-01 -1.03691030e+00 1.65507340e+00
-9.34123635e-01 1.09040689e+00 2.51501679e-01 -8.28675330e-01
1.21078563e+00 7.25009024e-01 -3.62939477e-01 -6.35948896e-01
3.16574544e-01 6.84816182e-01 2.34569341e-01 -3.98329586e-01
7.23108411e-01 8.34225118e-02 -3.10249716e-01 -2.37728596e-01
7.24438012e-01 -4.84147877e-01 -3.29826444e-01 -2.75013745e-01
8.81651402e-01 -1.43857300e-01 -1.43821523e-01 -2.08953351e-01
5.80066741e-01 -5.10027349e-01 3.32535952e-01 8.74001145e-01
-1.50884300e-01 9.89600360e-01 2.79168159e-01 -3.12243074e-01
-1.23178554e+00 -7.14603603e-01 -3.89743358e-01 1.16835964e+00
-5.48388362e-01 -4.36992586e-01 -7.48799801e-01 -1.79149613e-01
-3.67662162e-01 7.47481704e-01 -5.54792166e-01 -2.51233995e-01
-9.96323526e-01 -9.48330641e-01 1.15673149e+00 7.83742189e-01
5.30905902e-01 -1.11553121e+00 -4.33905125e-02 4.58565533e-01
2.89260671e-02 -1.43499506e+00 -1.83252633e-01 1.09407258e+00
-7.26467788e-01 -2.59585649e-01 -8.43158603e-01 -1.21220648e+00
-1.28555270e-02 -2.11366177e-01 6.49900079e-01 -8.22522268e-02
-1.29194170e-01 -2.33010650e-01 -1.69122741e-01 -4.62309122e-01
-6.60131454e-01 1.79545552e-01 7.83427954e-02 -3.06408316e-01
3.02072883e-01 -3.03107500e-01 -4.35483456e-01 1.44858524e-01
-4.63395715e-01 -3.63915801e-01 6.01470828e-01 1.06047833e+00
2.74042040e-01 -5.04256427e-01 6.04911447e-01 -3.64427179e-01
4.03000891e-01 -4.11199359e-03 -3.69918793e-01 -1.94518358e-01
6.44321379e-04 3.88149172e-02 5.85814536e-01 -2.52315342e-01
-1.05076146e+00 7.36054704e-02 -1.28676891e+00 -1.52602017e-01
-2.69064218e-01 7.96180367e-02 -3.39684427e-01 -4.22768816e-02
7.11187124e-01 4.73903604e-02 -1.25464082e-01 -8.05397868e-01
2.22626522e-01 1.44243157e+00 6.45702004e-01 1.21940970e-01
2.91370720e-01 -6.93081692e-02 -5.09663165e-01 -1.11788726e+00
-6.32972777e-01 -4.82278466e-01 -7.37915933e-01 1.80892497e-01
9.46773052e-01 -1.08151555e+00 -3.77746165e-01 8.81235301e-01
-1.17969155e+00 -4.27982837e-01 -2.41796866e-01 9.35210168e-01
-5.39663196e-01 1.17626689e-01 -1.02187085e+00 -8.32338452e-01
-5.03967762e-01 -1.54990518e+00 1.06896782e+00 5.86213777e-03
8.26750416e-03 -9.17725861e-01 -1.66979194e-01 2.16834709e-01
8.48036528e-01 -2.05579653e-01 5.25981247e-01 -1.02018142e+00
4.87474585e-03 -5.16211331e-01 -1.33677512e-01 1.33090472e+00
-1.41465098e-01 6.46078438e-02 -1.79594350e+00 -2.58531749e-01
-5.94228804e-02 -3.81176591e-01 1.33627057e+00 6.34029627e-01
1.26705062e+00 -1.65838093e-01 -8.91332775e-02 7.78897285e-01
1.10316896e+00 4.67163116e-01 8.33722353e-01 1.78363070e-01
4.96074945e-01 4.03489947e-01 2.31998593e-01 3.68808396e-02
-2.35771224e-01 8.38322341e-01 3.20183933e-01 -2.25493371e-01
-6.45308435e-01 4.01309013e-01 6.81630135e-01 1.45490766e+00
6.30464628e-02 -4.31948036e-01 -6.38521552e-01 6.66492283e-01
-1.43506706e+00 -6.02330923e-01 -9.26291943e-02 2.09409285e+00
6.47664964e-01 4.03988421e-01 -3.64540994e-01 4.81563061e-01
7.06465364e-01 4.36196595e-01 4.46244748e-03 -1.35097063e+00
-3.66395891e-01 7.66245246e-01 7.94757128e-01 9.11353767e-01
-1.36527705e+00 1.11369359e+00 6.89271164e+00 1.31999218e+00
-9.74903762e-01 3.97928417e-01 6.57092810e-01 -2.28248402e-01
2.42298007e-01 -4.50333178e-01 -1.19029450e+00 -2.11715866e-02
1.90790033e+00 7.70416558e-01 6.31654784e-02 6.89981818e-01
2.52711743e-01 1.49734840e-01 -6.22923613e-01 9.66921985e-01
8.60620737e-02 -1.00146723e+00 -1.39881089e-01 5.12761176e-02
4.96350437e-01 5.78652978e-01 3.40100676e-01 6.19219601e-01
2.57832080e-01 -1.18845916e+00 6.02193832e-01 2.13079125e-01
9.23414588e-01 -1.18238509e+00 1.18691123e+00 -1.63903907e-01
-1.13931155e+00 -3.26734126e-01 -8.00415993e-01 -6.39449712e-03
5.03823236e-02 3.70774388e-01 -9.63691354e-01 1.24457613e-01
8.04673791e-01 6.78797886e-02 -5.28371572e-01 9.87968981e-01
-4.50787134e-03 1.00758433e+00 -6.22019291e-01 -3.38412702e-01
7.81047761e-01 6.66079402e-01 4.46501732e-01 1.96459508e+00
4.76184547e-01 -4.71480846e-01 -4.70019549e-01 2.47262970e-01
-1.26402229e-01 7.87847862e-02 -3.50370526e-01 2.37450182e-01
3.16844523e-01 1.16713619e+00 -1.31785616e-01 -4.29671943e-01
-4.40317005e-01 9.10048068e-01 1.88856393e-01 4.79082435e-01
-5.11227250e-01 -1.18691361e+00 9.01612818e-01 -5.04085124e-01
6.09161377e-01 -3.20219994e-01 -2.61852086e-01 -6.36258602e-01
-1.09776385e-01 -5.88240266e-01 2.33043414e-02 -7.42471397e-01
-5.66163123e-01 1.02856708e+00 -3.38201880e-01 -8.41244519e-01
-5.08406401e-01 -1.06170011e+00 -8.04324150e-01 1.33124328e+00
-1.60202014e+00 -9.97077703e-01 1.83665782e-01 4.31094855e-01
9.83902991e-01 -4.74144936e-01 1.14316130e+00 4.48574603e-01
-5.73771060e-01 1.06597531e+00 4.02888685e-01 3.02657455e-01
5.48578858e-01 -1.41873252e+00 1.20377457e+00 7.53896654e-01
7.94975385e-02 4.72860634e-01 4.48322326e-01 -1.08143695e-01
-1.08107078e+00 -1.11137402e+00 1.16158879e+00 -4.62424904e-02
4.94448513e-01 -9.87551332e-01 -8.22377443e-01 5.78706384e-01
4.11388427e-01 1.84612468e-01 4.33513612e-01 5.93402497e-02
-2.56917149e-01 -1.24674283e-01 -9.12960589e-01 4.51013505e-01
5.26894450e-01 -8.73606384e-01 -4.25635904e-01 4.50142086e-01
1.22209430e+00 -5.18469691e-01 -7.29703069e-01 5.48340380e-01
5.19576371e-01 -5.83337784e-01 1.00484252e+00 -4.02456790e-01
-2.53130525e-01 1.53879924e-02 -5.49057424e-01 -1.41174376e+00
-1.37696341e-01 -7.83155859e-01 -1.06220003e-02 9.44613695e-01
1.16191339e+00 -5.67319453e-01 2.91826904e-01 1.71593368e-01
-9.49127197e-01 -5.16376257e-01 -1.35884786e+00 -7.36628532e-01
1.14621826e-01 -1.14752996e+00 2.35648155e-01 1.17323399e-01
-2.23777771e-01 3.41595709e-02 -5.11109054e-01 3.68328035e-01
-8.34048837e-02 -9.28716183e-01 3.34304333e-01 -8.01428676e-01
-4.92345899e-01 -2.99754202e-01 -5.34896553e-01 -1.18722844e+00
-1.28607020e-01 -5.82107246e-01 4.24254924e-01 -1.32742405e+00
-4.49087679e-01 -1.03843965e-01 -7.78499186e-01 7.26769626e-01
1.14681281e-01 5.94313383e-01 -1.30029758e-02 -2.79232323e-01
-1.91134632e-01 3.43583167e-01 8.94452631e-01 7.27182105e-02
-2.23540500e-01 4.25267518e-01 -1.42425194e-01 6.36762142e-01
1.19091797e+00 -2.39058822e-01 -3.40614021e-02 -3.64391387e-01
-4.96025160e-02 -8.72247815e-02 2.17553601e-01 -1.30976105e+00
1.84788123e-01 7.58720219e-01 4.87690687e-01 -5.82929790e-01
1.02412546e+00 -3.09819251e-01 -5.91617882e-01 3.04512203e-01
-5.39568186e-01 -1.19049527e-01 7.72214651e-01 3.01685154e-01
-5.67662239e-01 -4.57964897e-01 9.80407655e-01 -8.92438516e-02
-7.09902763e-01 -1.56844780e-01 -9.38320458e-01 -2.81372607e-01
1.66679755e-01 -8.23099762e-02 -6.81952834e-02 -3.66023630e-01
-1.24959087e+00 -1.69929624e-01 -6.73418522e-01 6.78280175e-01
6.67115927e-01 -1.17645693e+00 -7.64481544e-01 3.92198056e-01
-1.65301561e-02 -1.33249238e-01 1.68678194e-01 7.68039942e-01
-4.89359975e-01 6.13211334e-01 1.10822931e-01 -4.09296483e-01
-1.46143961e+00 -2.17937171e-01 8.66611660e-01 1.60622653e-02
-4.62310582e-01 1.43781900e+00 -1.27460942e-01 -7.22296774e-01
7.32525051e-01 -7.17196226e-01 -2.54550219e-01 -1.77115336e-01
8.10357630e-01 2.89963722e-01 6.95493937e-01 -6.94097459e-01
-2.91283876e-01 3.86379272e-01 -1.96366400e-01 -5.52721262e-01
1.39348018e+00 -1.32938579e-01 2.72753179e-01 3.88395160e-01
1.66861343e+00 -1.34744301e-01 -1.11276913e+00 -2.63078570e-01
-4.10274975e-02 1.45696953e-01 4.99806762e-01 -1.00658870e+00
-1.01122618e+00 1.46093273e+00 8.73731732e-01 -2.25881720e-03
9.08120453e-01 -1.13420337e-01 9.44834948e-01 6.91109896e-01
-3.91107470e-01 -1.38192165e+00 -2.14960650e-01 1.06120276e+00
9.42846298e-01 -8.85438859e-01 -5.66631377e-01 -3.50632258e-02
-2.61783391e-01 1.41215396e+00 4.48269546e-01 -1.81599081e-01
8.22084844e-01 9.64939654e-01 2.73416907e-01 -2.23107189e-02
-6.90414429e-01 -2.85982400e-01 2.70019859e-01 5.18029392e-01
4.76106793e-01 3.99307869e-02 1.06350005e-01 5.81509769e-01
-5.17861843e-01 -7.23325908e-01 3.31964105e-01 6.26262188e-01
-8.11023474e-01 -8.73996019e-01 -3.09861124e-01 2.66985089e-01
-1.01808572e+00 -6.09382093e-01 -1.83125958e-01 4.99991119e-01
-3.24177928e-02 1.13130677e+00 2.38479808e-01 -6.00607574e-01
4.62105721e-01 3.19480509e-01 8.50686245e-03 -6.58806086e-01
-1.11406898e+00 4.45847958e-01 5.18721104e-01 -3.04127723e-01
7.14702606e-02 -2.30555132e-01 -1.20762599e+00 3.44336003e-01
-8.46197486e-01 1.87875450e-01 1.43330419e+00 9.41669345e-01
1.83715060e-01 8.38967800e-01 5.76054990e-01 -1.01623189e+00
-5.51091135e-01 -1.58683658e+00 -5.85119128e-01 -4.08756196e-01
8.80296946e-01 -1.21588238e-01 -6.31326556e-01 -2.62620091e-01] | [14.594536781311035, 6.177769660949707] |
a596c215-31a4-4293-8070-db6fe3212b47 | integration-of-physics-based-and-data-driven | 2206.05508 | null | https://arxiv.org/abs/2206.05508v2 | https://arxiv.org/pdf/2206.05508v2.pdf | Integration of Physics-Based and Data-Driven Models for Hyperspectral Image Unmixing | Spectral unmixing is one of the most important quantitative analysis tasks in hyperspectral data processing. Conventional physics-based models are characterized by clear interpretation. However they may not be suitable for analyzing scenes with unknown complex physical characteristics. Data-driven methods have developed rapidly in recent years, in particular deep learning methods because they possess superior capability in modeling complex and nonlinear systems. Simply transferring these methods as black-boxes to conduct unmixing may lead to low physical interpretability and generalization ability. This article reviews hyperspectral unmixing works that integrate advantages of both physics-based models and data-driven methods by means of deep neural network structures design, prior design and loss design. Most of these methods derive from a common mathematical optimization framework, and combine good interpretability with high accuracy. | ['Susanto Rahardja', 'Cédric Richard', 'Xiuheng Wang', 'Min Zhao', 'Jie Chen'] | 2022-06-11 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 3.54341835e-01 -6.83561862e-01 -1.12715609e-01 -3.15829158e-01
-2.89370656e-01 -2.08135724e-01 5.29738486e-01 5.21967933e-02
-2.00007960e-01 8.01378191e-01 -3.45553644e-02 -2.21836254e-01
-8.24278831e-01 -8.54554892e-01 -2.81935424e-01 -1.16523492e+00
2.85686493e-01 3.19965929e-01 -4.15652156e-01 -4.15956557e-01
-1.97656706e-01 7.43805051e-01 -1.45430350e+00 -1.47296816e-01
1.48955250e+00 1.10004508e+00 2.38221347e-01 3.46196383e-01
-2.78474987e-01 4.57404763e-01 -2.08881423e-01 1.14828087e-01
3.65187228e-01 -3.08682024e-01 -2.48992965e-01 1.22986697e-01
2.12948740e-01 -3.98157090e-01 -6.89987719e-01 1.30306482e+00
5.99976540e-01 4.31715071e-01 7.16695011e-01 -1.33946335e+00
-9.54497159e-01 4.16260391e-01 -9.24831629e-01 -1.11467376e-01
-2.17197061e-01 2.45821625e-01 8.19530368e-01 -6.65707290e-01
-8.09619650e-02 1.17677379e+00 8.46178651e-01 1.12910382e-01
-1.26736784e+00 -6.86086535e-01 1.31437391e-01 2.66146779e-01
-1.23860788e+00 -2.83425838e-01 1.17269266e+00 -6.07533216e-01
6.04758799e-01 4.17176872e-01 7.80430198e-01 6.23114169e-01
-3.67839495e-03 7.09666848e-01 1.25057805e+00 -3.10650438e-01
9.34972614e-03 -4.33679707e-02 2.04287857e-01 4.78975892e-01
5.88265002e-01 5.50462365e-01 -3.28734607e-01 1.39229763e-02
6.49518967e-01 3.33814293e-01 -5.68893373e-01 -3.44678193e-01
-9.76225555e-01 7.90695369e-01 8.33954990e-01 2.54571587e-02
-5.71734369e-01 -1.56577099e-02 3.54248807e-02 1.61444068e-01
5.75172007e-01 5.93590558e-01 -2.44183630e-01 5.21397173e-01
-1.10407472e+00 3.15290064e-01 4.19228077e-01 6.92337990e-01
8.85851145e-01 3.90085042e-01 -8.84185657e-02 9.80798066e-01
5.77193618e-01 8.84307265e-01 5.07940054e-01 -7.24552572e-01
2.18870953e-01 6.37088120e-01 1.00917913e-01 -9.51925635e-01
-7.14101672e-01 -7.08790183e-01 -1.33796883e+00 5.58969736e-01
-4.26161028e-02 -3.07191640e-01 -1.00375998e+00 1.55375946e+00
2.44285110e-02 -2.19130501e-01 2.30635047e-01 1.06502998e+00
1.03846467e+00 1.12636685e+00 4.01589386e-02 -2.18388349e-01
1.07344079e+00 -1.08729577e+00 -9.78251994e-01 -3.21964055e-01
1.11212239e-01 -6.59114778e-01 9.33291554e-01 4.42981333e-01
-8.68928432e-01 -5.14015198e-01 -1.40304673e+00 -4.71948236e-02
-6.79869354e-01 2.74710178e-01 8.19189072e-01 6.20070815e-01
-6.79092288e-01 8.19612443e-01 -6.60134256e-01 -1.55982941e-01
5.47301948e-01 3.62390339e-01 -1.27055168e-01 1.18412934e-01
-1.02633238e+00 8.18237305e-01 6.52651370e-01 5.91326594e-01
-8.29648197e-01 -7.39806116e-01 -7.34248400e-01 2.83484548e-01
2.80459136e-01 -7.72095859e-01 1.08811319e+00 -1.10863268e+00
-1.65264833e+00 3.91475767e-01 -1.49784550e-01 -6.96358457e-02
2.87734658e-01 -2.87263691e-01 -7.42925763e-01 4.66659553e-02
-3.38771999e-01 3.84108126e-01 8.75928938e-01 -1.44206333e+00
-2.92673618e-01 -1.96704328e-01 -9.26423520e-02 2.81815678e-01
-5.86278260e-01 -2.70866007e-01 1.66330501e-01 -6.81913733e-01
4.36854988e-01 -6.90694392e-01 -3.04639310e-01 3.03095698e-01
-5.61774492e-01 3.88198376e-01 1.07114398e+00 -7.32536674e-01
9.76216495e-01 -1.84684312e+00 9.49335843e-02 1.09733328e-01
3.68329197e-01 7.40026951e-01 -1.79187343e-01 4.68122303e-01
-3.51478189e-01 9.24332663e-02 -6.55553281e-01 -1.15548752e-01
1.14643700e-01 -7.24006668e-02 -1.98348656e-01 4.99797344e-01
3.77640486e-01 8.86900723e-01 -7.84703195e-01 3.73776369e-02
6.34256959e-01 3.70270193e-01 -2.60810733e-01 4.20578450e-01
-4.10941929e-01 4.60369796e-01 -5.18937707e-01 7.73563027e-01
1.10464871e+00 -4.36089277e-01 1.75753552e-02 -6.62557602e-01
-1.83331922e-01 -1.36427209e-01 -1.20264769e+00 1.43360770e+00
-2.87768692e-01 6.53103888e-01 4.75363761e-01 -1.31190109e+00
8.38292956e-01 3.36162657e-01 6.66979730e-01 -4.39758539e-01
1.62693113e-01 2.74587601e-01 1.23672292e-01 -7.78711379e-01
3.99023622e-01 -7.26046503e-01 5.96783221e-01 1.71709731e-01
-1.23502247e-01 -4.46723938e-01 2.42752042e-02 -3.74573797e-01
2.51608849e-01 1.59747958e-01 2.92839229e-01 -4.17202145e-01
7.26894736e-01 2.53662586e-01 5.94749093e-01 5.19688964e-01
4.72331652e-03 5.57319760e-01 -4.42629904e-02 -4.18133706e-01
-1.05945933e+00 -8.49328339e-01 -3.07207435e-01 5.58979154e-01
3.48736644e-01 1.95827067e-01 -3.35750282e-01 8.97166058e-02
1.55301334e-03 4.63716924e-01 -1.77932441e-01 -3.83500606e-01
-1.01750411e-05 -1.52708936e+00 3.73106927e-01 4.14944947e-01
9.00488555e-01 -8.65745068e-01 -5.23839183e-02 3.84523183e-01
-6.95043728e-02 -8.25316310e-01 9.46099758e-02 2.22030759e-01
-9.77647007e-01 -9.95323479e-01 -7.97138572e-01 -4.24679726e-01
6.31179154e-01 6.91679537e-01 7.80701339e-01 -1.59387290e-01
-3.11094344e-01 -7.58050606e-02 -1.22691192e-01 -8.36170435e-01
-3.68069619e-01 1.96076669e-02 1.72341317e-01 1.62320033e-01
3.02078098e-01 -6.24151111e-01 -5.49037635e-01 7.31911808e-02
-1.39521468e+00 1.34851292e-01 7.27025330e-01 9.07753885e-01
4.36138004e-01 6.83357596e-01 3.93162847e-01 -5.17700791e-01
7.12763250e-01 -3.79441381e-01 -8.74014914e-01 4.10409093e-01
-8.52801561e-01 5.85409850e-02 7.12745309e-01 -1.97573259e-01
-1.38073766e+00 -1.23935565e-01 3.39640044e-02 -3.99093896e-01
-2.84847260e-01 1.00978827e+00 -5.58976471e-01 -3.09682786e-01
6.36977494e-01 1.81124151e-01 3.97925228e-01 -5.97646415e-01
2.18424812e-01 7.24571824e-01 2.95394599e-01 -6.06207907e-01
1.10063863e+00 7.03631699e-01 3.39153022e-01 -1.06790745e+00
-9.43859220e-01 -5.20135999e-01 -4.01614904e-01 -1.09580331e-01
7.21066356e-01 -1.00718200e+00 -6.17051303e-01 1.01541388e+00
-9.26982462e-01 -3.29788148e-01 -1.46082476e-01 7.70395517e-01
-2.52452284e-01 5.41247129e-01 -3.67606640e-01 -8.04070950e-01
-3.58523637e-01 -1.15256190e+00 7.02515066e-01 6.28465354e-01
3.15529555e-01 -1.14215088e+00 -1.50121316e-01 3.34941030e-01
6.79520249e-01 4.35511440e-01 1.05007994e+00 -1.06004447e-01
-4.80230778e-01 -1.39403552e-01 -6.39622748e-01 6.23207271e-01
3.99664462e-01 2.77904838e-01 -1.28759134e+00 -2.32106552e-01
2.54199892e-01 -2.65022218e-01 1.01019561e+00 8.84704113e-01
1.42733610e+00 -1.11723043e-01 -2.07359537e-01 1.10049403e+00
1.93669307e+00 3.73369902e-01 5.23925900e-01 3.31887603e-01
8.40024829e-01 7.36785233e-01 2.11663693e-01 2.83876240e-01
-1.46536559e-01 2.25058541e-01 6.76837504e-01 -7.00164616e-01
2.10292697e-01 -4.04586680e-02 7.78963864e-02 6.52342498e-01
-1.08777888e-01 -5.54068565e-01 -8.54006827e-01 1.46145537e-01
-1.94351757e+00 -1.04316533e+00 -4.88901109e-01 1.91270006e+00
6.19786322e-01 -1.73425451e-01 -1.75257578e-01 1.41398922e-01
7.91579366e-01 3.75751168e-01 -8.07435095e-01 9.07605588e-02
-6.28579557e-01 4.15966468e-04 7.55428970e-01 4.57684487e-01
-1.17317641e+00 6.41749322e-01 6.83194923e+00 7.45221376e-01
-1.27241874e+00 5.40426821e-02 4.88671303e-01 7.25262314e-02
-4.43326533e-01 -2.48458935e-04 -4.28389639e-01 1.95583299e-01
4.21083868e-01 3.89233790e-02 6.51804388e-01 3.88021439e-01
7.29435682e-01 -5.90371378e-02 -7.68029928e-01 1.13889384e+00
-2.27195740e-01 -1.34547961e+00 2.28087947e-01 1.32279441e-01
8.54113042e-01 3.45481075e-02 1.15412794e-01 -6.17860537e-03
8.34283456e-02 -1.32644093e+00 4.32999611e-01 8.80087376e-01
4.88167346e-01 -3.62716764e-01 7.31714010e-01 3.03294361e-01
-9.31694329e-01 -1.98359489e-01 -5.30282140e-01 -2.59825349e-01
2.78167486e-01 1.14010465e+00 3.16637419e-02 8.55062962e-01
4.69729215e-01 8.84194076e-01 -7.48671368e-02 1.31514609e+00
-2.08456650e-01 5.81176639e-01 -4.67741668e-01 5.84116951e-02
3.46900105e-01 -9.64120805e-01 6.84938431e-01 8.36687028e-01
4.68177974e-01 1.85362175e-01 1.38673306e-01 1.32981443e+00
1.64720386e-01 -1.76167376e-02 -5.48764646e-01 -4.57012028e-01
2.41229475e-01 1.32412696e+00 -3.70381236e-01 -2.73210824e-01
-4.26118910e-01 4.30965900e-01 -2.08790600e-01 6.00350320e-01
-8.60693812e-01 -3.48000616e-01 7.31922686e-01 -5.84300561e-03
-1.71867400e-01 -4.66184884e-01 -3.96903217e-01 -1.27012539e+00
-6.53710887e-02 -9.05008435e-01 1.61541760e-01 -1.13416386e+00
-1.60818827e+00 3.17665726e-01 9.12045836e-02 -1.34662354e+00
5.22393048e-01 -1.24723613e+00 -7.64255404e-01 1.18790710e+00
-2.12209225e+00 -1.05056524e+00 -9.68526602e-01 3.27945322e-01
7.84760714e-02 -2.11452633e-01 7.32760251e-01 4.15839583e-01
-9.97461140e-01 -8.65983143e-02 7.40684628e-01 -2.09727079e-01
5.14047563e-01 -1.24091136e+00 -2.54899085e-01 9.22751129e-01
-4.24550653e-01 5.03521502e-01 6.23279512e-01 -2.64881134e-01
-1.40250432e+00 -1.09391963e+00 1.54203340e-01 2.37673506e-01
6.29167914e-01 1.21941745e-01 -1.05364418e+00 1.13489076e-01
2.01700568e-01 -2.08770201e-01 8.57292056e-01 1.80486031e-03
-3.16796303e-02 -6.25870883e-01 -1.01187360e+00 5.66831827e-01
6.59113169e-01 -3.07860702e-01 -4.37727243e-01 6.18409812e-01
4.38093752e-01 -1.30128190e-01 -7.66208231e-01 6.20042920e-01
3.50825399e-01 -9.36055601e-01 1.09132028e+00 -5.71992874e-01
5.84095657e-01 -5.74284732e-01 -4.72183079e-02 -1.44958186e+00
-7.01930583e-01 -6.73518538e-01 2.45313793e-02 1.11154842e+00
3.72193605e-01 -8.76643896e-01 6.47864401e-01 4.58953291e-01
-1.25790283e-01 -3.13662767e-01 -3.17941099e-01 -9.44541812e-01
5.86595051e-02 -1.81697682e-01 1.05070150e+00 1.02447176e+00
-3.65722269e-01 2.02961892e-01 -3.76371592e-01 4.29030597e-01
7.35848248e-01 2.76031822e-01 4.80010033e-01 -1.53245056e+00
-1.54396445e-01 -9.40293312e-01 7.45062828e-02 -7.27581918e-01
1.26160428e-01 -7.32950389e-01 1.10695921e-01 -1.81794858e+00
1.10385500e-01 -5.36449969e-01 -5.61294198e-01 4.77505654e-01
-2.85131842e-01 1.13129313e-03 -2.02924252e-01 4.54159886e-01
3.18666905e-01 9.90899205e-01 1.28268623e+00 -5.96702516e-01
-3.77652377e-01 -1.02790274e-01 -8.01440001e-01 7.42400765e-01
1.10840178e+00 -2.58076251e-01 -5.08407950e-01 -7.13522494e-01
3.26068491e-01 -1.77693292e-01 4.97901440e-01 -1.12530708e+00
-2.38830596e-02 -6.42294288e-01 3.78519148e-01 -3.99871528e-01
4.49608684e-01 -1.10023165e+00 4.29637671e-01 4.51311529e-01
-1.49141803e-01 -5.18967330e-01 4.78218257e-01 4.66541350e-01
-3.80875856e-01 -4.44660038e-01 9.72331583e-01 -1.38751283e-01
-8.22213650e-01 6.40135229e-01 -3.04718167e-01 -3.92636120e-01
7.50896513e-01 -3.45110118e-01 -3.34413052e-01 -3.12126100e-01
-6.12275779e-01 2.24984035e-01 2.83311099e-01 1.13331050e-01
3.96974623e-01 -1.24143350e+00 -7.09576547e-01 8.60059485e-02
9.03985426e-02 1.08617567e-01 4.00681823e-01 7.80572653e-01
-9.74146008e-01 2.78954327e-01 -4.00947303e-01 -5.09349942e-01
-7.54989862e-01 5.79188585e-01 8.19583178e-01 -3.67598608e-02
-2.85932809e-01 7.09202468e-01 3.19767952e-01 -7.10045516e-01
-1.53177932e-01 -3.18027496e-01 -7.91721940e-02 4.33870219e-02
3.25954914e-01 5.24094701e-01 3.29257622e-02 -5.17831266e-01
-6.23820573e-02 5.66345513e-01 5.27159214e-01 6.01753332e-02
1.51332307e+00 -5.45411035e-02 -4.15722430e-01 2.14509651e-01
1.06280684e+00 -2.92795062e-01 -1.39798164e+00 -3.93421412e-01
-4.26037610e-01 -2.95336366e-01 7.40963519e-01 -8.00789416e-01
-1.22560585e+00 1.19093955e+00 6.19948149e-01 4.73173112e-01
1.43903303e+00 -6.33725107e-01 6.44619346e-01 5.31158626e-01
-2.62411267e-01 -1.12345171e+00 -2.39207357e-01 3.13157678e-01
8.41769636e-01 -1.65059388e+00 4.03401852e-01 -5.49766958e-01
-9.36090052e-02 1.39436054e+00 6.31975591e-01 2.10175291e-01
9.37603712e-01 -3.54498774e-02 1.02411039e-01 -3.97494555e-01
-3.80496196e-02 -2.70273775e-01 4.76322353e-01 7.42486894e-01
3.09165657e-01 1.81253761e-01 -9.21523944e-02 4.71033126e-01
9.59053710e-02 -7.07094073e-02 2.53693342e-01 8.51566672e-01
-5.44137776e-01 -9.89543259e-01 -6.38868868e-01 6.33183718e-01
-4.95181642e-02 -2.29155868e-01 -1.14053220e-01 6.37041867e-01
1.67805895e-01 1.11376369e+00 -2.56447732e-01 -7.56929368e-02
1.10134616e-01 -1.02550715e-01 2.74772495e-01 -5.20017028e-01
-4.36175428e-02 9.54610705e-02 -2.08438799e-01 -3.97013704e-04
-8.90110075e-01 -4.40569907e-01 -9.03376341e-01 -4.87734050e-01
-6.79740310e-01 4.27163504e-02 7.21951902e-01 1.17421401e+00
1.36522233e-01 6.81367815e-01 5.14360428e-01 -1.00082946e+00
-7.34833121e-01 -1.04656541e+00 -1.01015854e+00 1.67480752e-01
5.31970918e-01 -6.63613737e-01 -2.66125947e-01 6.59890249e-02] | [10.12356185913086, -2.0513389110565186] |
bafd08f0-105a-462f-9d0f-267a404896e0 | feature-concatenation-multi-view-subspace | 1901.10657 | null | https://arxiv.org/abs/1901.10657v6 | https://arxiv.org/pdf/1901.10657v6.pdf | Feature Concatenation Multi-view Subspace Clustering | Multi-view clustering is a learning paradigm based on multi-view data. Since statistic properties of different views are diverse, even incompatible, few approaches implement multi-view clustering based on the concatenated features straightforward. However, feature concatenation is a natural way to combine multi-view data. To this end, this paper proposes a novel multi-view subspace clustering approach dubbed Feature Concatenation Multi-view Subspace Clustering (FCMSC), which boosts the clustering performance by exploring the consensus information of multi-view data. Specifically, multi-view data are concatenated into a joint representation firstly, then, $l_{2,1}$-norm is integrated into the objective function to deal with the sample-specific and cluster-specific corruptions of multiple views. Moreover, a graph regularized FCMSC is also proposed in this paper to explore both the consensus information and complementary information of multi-view data for clustering. It is noteworthy that the obtained coefficient matrix is not derived by simply applying the Low-Rank Representation (LRR) to concatenated features directly. Finally, an effective algorithm based on the Augmented Lagrangian Multiplier (ALM) is designed to optimize the objective functions. Comprehensive experiments on six real-world datasets illustrate the superiority of the proposed methods over several state-of-the-art approaches for multi-view clustering. | ['Shanmin Pang', 'Qinghai Zheng', 'Jun Wang', 'Jihua Zhu', 'Zhongyu Li', 'Yaochen Li'] | 2019-01-30 | null | null | null | null | ['multi-view-subspace-clustering'] | ['computer-vision'] | [-3.12518060e-01 -6.86112761e-01 4.80457321e-02 -2.99483567e-01
-7.98753738e-01 -4.76644218e-01 3.54931355e-01 -2.81334430e-01
-1.20981224e-01 3.06515068e-01 3.51099342e-01 4.25129622e-01
-5.10624409e-01 -4.63013768e-01 -1.54266298e-01 -1.27654684e+00
1.56882852e-01 4.22037542e-02 -2.46178538e-01 4.72515449e-02
1.49879947e-01 1.85192287e-01 -1.33459735e+00 3.33592057e-01
8.72555494e-01 5.82637489e-01 2.82819420e-01 2.02368677e-01
2.05022871e-01 3.60922426e-01 -1.55779481e-01 -7.18424171e-02
1.91167653e-01 -4.90311354e-01 -3.44536096e-01 6.87648237e-01
1.21170998e-01 1.27212003e-01 -1.20148294e-01 1.23345649e+00
6.48647904e-01 2.26366714e-01 4.94190276e-01 -1.26613295e+00
-4.27572012e-01 3.41332644e-01 -1.07090962e+00 -5.12633249e-02
2.90516317e-01 -5.81787564e-02 9.84494090e-01 -1.29671717e+00
6.17639422e-01 1.33563435e+00 2.43452147e-01 -1.19631074e-01
-1.34787095e+00 -4.59342927e-01 3.72253597e-01 4.89264280e-01
-1.72841036e+00 -2.35983089e-01 1.24791431e+00 -4.50302988e-01
3.18401337e-01 3.84486735e-01 5.77466965e-01 7.06596553e-01
-1.00882025e-02 8.11470866e-01 1.44073570e+00 -1.20651871e-01
-5.88624887e-02 1.32137373e-01 8.80064368e-02 6.90213442e-01
3.23051810e-01 -1.84568390e-01 -3.52508903e-01 -3.26614141e-01
1.82233110e-01 3.63560140e-01 -4.62977976e-01 -8.94749224e-01
-1.45948076e+00 9.16214883e-01 3.08535933e-01 4.24644500e-01
-3.28251034e-01 -4.32078898e-01 5.06521165e-01 -6.91833207e-03
3.53101403e-01 -1.29060775e-01 -1.73135623e-01 2.50574768e-01
-7.57185459e-01 -1.90648451e-01 4.51438427e-01 9.54506695e-01
9.25216615e-01 1.30623460e-01 1.55730978e-01 9.91009116e-01
6.08367801e-01 6.14880860e-01 3.62427473e-01 -8.44882667e-01
6.26356721e-01 1.02256346e+00 -3.13075364e-01 -1.43199515e+00
-4.58719939e-01 -5.43392777e-01 -1.47314429e+00 -7.92261139e-02
-2.33114101e-02 -1.59163892e-01 -3.11871588e-01 1.52658153e+00
5.54488540e-01 2.31659338e-01 2.96274126e-01 9.91357327e-01
9.90336359e-01 5.44330835e-01 -4.51934844e-01 -9.19817507e-01
1.20388806e+00 -8.59648466e-01 -8.07110906e-01 2.43755668e-01
2.73039520e-01 -8.58641863e-01 5.47548890e-01 6.03617072e-01
-8.51492107e-01 -6.64533317e-01 -1.05475497e+00 3.93145591e-01
-2.32371300e-01 3.57764989e-01 3.82955879e-01 4.06850427e-01
-6.53873205e-01 1.49331450e-01 -6.21762395e-01 -1.45832300e-01
6.67434856e-02 3.12327683e-01 -7.42087364e-01 -5.52725434e-01
-7.45452106e-01 4.39970106e-01 5.72380066e-01 3.08961987e-01
-4.47313666e-01 -2.69585311e-01 -7.45647788e-01 -1.82642907e-01
7.31490135e-01 -5.73497891e-01 1.84718639e-01 -5.02130330e-01
-1.04878151e+00 5.07319987e-01 -4.02078122e-01 4.87367213e-01
1.53141633e-01 -1.80842099e-03 -7.56991923e-01 3.26521277e-01
2.55787313e-01 -1.06440514e-01 7.99124002e-01 -1.92431223e+00
-3.30896765e-01 -6.97671711e-01 -4.20157105e-01 5.49964011e-01
-2.51644373e-01 -1.88987687e-01 -8.70932817e-01 -5.19836605e-01
7.61608541e-01 -9.67682898e-01 -3.66847008e-01 -5.76298296e-01
-6.35519862e-01 3.74420732e-02 1.14259183e+00 -6.31150126e-01
1.43702114e+00 -2.33232737e+00 8.85342419e-01 5.50306380e-01
2.64919668e-01 2.10498068e-02 -3.28378230e-02 7.50923514e-01
-1.89416125e-01 -9.22972187e-02 -3.95420432e-01 -1.51856691e-01
-2.71322876e-01 1.24451481e-01 2.90366769e-01 8.65876019e-01
-3.38754892e-01 4.19393659e-01 -8.41082990e-01 -7.86006570e-01
3.85958731e-01 3.95609587e-01 -4.55313504e-01 1.50451437e-01
5.53467572e-01 7.92701840e-01 -5.30938864e-01 6.48095012e-01
1.06227112e+00 -3.29512954e-01 4.75998312e-01 -7.56398618e-01
-1.78373948e-01 -7.83931792e-01 -1.82424867e+00 1.87232018e+00
-1.64657086e-01 -2.09948137e-01 4.43273872e-01 -1.10136747e+00
6.54708028e-01 3.77293855e-01 1.01296854e+00 -9.27071273e-03
5.78605272e-02 1.38685182e-01 2.38095075e-02 -5.23274601e-01
9.31444466e-02 -1.22263975e-01 -8.01924034e-04 3.76780868e-01
6.30445555e-02 1.34006202e-01 2.98887283e-01 2.98611581e-01
5.58506906e-01 6.30043959e-03 5.07847011e-01 -2.71861374e-01
1.32168162e+00 -3.02898914e-01 8.81980002e-01 9.99137908e-02
7.28826458e-03 6.73651695e-01 2.88676858e-01 -5.70730790e-02
-7.35069156e-01 -1.02988279e+00 -3.65007780e-02 5.95518649e-01
3.66207749e-01 -6.31209373e-01 -5.17376244e-01 -7.80592978e-01
-1.19158834e-01 2.21588567e-01 -3.71817023e-01 -1.30687937e-01
-3.96220297e-01 -1.06385767e+00 -5.48742488e-02 2.17978016e-01
6.52367949e-01 -3.26850176e-01 -7.30525207e-05 1.58223182e-01
-5.22240937e-01 -1.05722630e+00 -5.41750073e-01 -1.44129947e-01
-9.54026401e-01 -1.28618884e+00 -5.75837433e-01 -6.85309649e-01
8.79058361e-01 9.44751859e-01 6.67524517e-01 5.12698814e-02
-1.48056358e-01 7.11828113e-01 -6.08548641e-01 2.11163431e-01
4.84572724e-02 -2.86998272e-01 3.95779282e-01 7.40897417e-01
2.54261643e-01 -7.47607172e-01 -5.68149209e-01 4.62335616e-01
-9.31977034e-01 5.50209433e-02 6.54001057e-01 1.12847102e+00
9.97829020e-01 4.73883897e-01 3.56809229e-01 -7.85974920e-01
3.09014887e-01 -5.12093186e-01 -3.98237139e-01 4.59434658e-01
-6.73140407e-01 -1.84429079e-01 9.16222453e-01 -7.74785429e-02
-1.10684669e+00 2.00861901e-01 3.69379938e-01 -8.93455684e-01
-5.27953692e-02 6.95764303e-01 -7.23112047e-01 1.61103848e-02
-2.85970178e-02 7.20477283e-01 1.41593859e-01 -4.60358977e-01
6.43134415e-01 5.89959323e-01 1.92207128e-01 -3.68073672e-01
1.03878701e+00 8.58906388e-01 3.35928112e-01 -9.04742777e-01
-7.42266238e-01 -9.58329558e-01 -1.01706684e+00 -3.34696621e-01
1.02049613e+00 -1.22353995e+00 -7.86281049e-01 4.00085062e-01
-7.17787147e-01 7.55272567e-01 2.20158964e-01 9.24777687e-01
-4.23587441e-01 9.06038702e-01 -3.14425170e-01 -6.76201165e-01
-1.13172568e-01 -1.26884246e+00 7.91742921e-01 -1.80685297e-02
4.83061254e-01 -9.47296083e-01 1.74905956e-02 6.90878153e-01
-2.07815617e-01 4.08151209e-01 7.96634078e-01 -5.60884655e-01
-3.83804888e-01 3.19578908e-02 -1.81599185e-01 4.44705516e-01
4.96511579e-01 3.56149524e-02 -7.20767915e-01 -7.67616153e-01
4.46365118e-01 3.70676965e-02 6.86770678e-01 3.81504148e-01
9.86086488e-01 -1.84481114e-01 -3.69108289e-01 7.08081543e-01
1.83634627e+00 1.69983536e-01 1.93458125e-01 8.72628316e-02
1.31676280e+00 5.46341240e-01 6.87023461e-01 7.34334528e-01
5.25494754e-01 5.43000042e-01 3.03052992e-01 6.16673827e-02
3.10577333e-01 5.60743287e-02 3.04723948e-01 1.76933324e+00
-2.78487086e-01 9.14476626e-03 -6.85216606e-01 2.33137831e-01
-2.06707764e+00 -1.19618011e+00 -7.01163530e-01 2.19003296e+00
6.93651140e-02 -3.79813105e-01 9.04499367e-02 2.04545453e-01
1.00087643e+00 4.11155611e-01 -4.71901983e-01 2.08394304e-01
-4.16476071e-01 -5.20829022e-01 2.48930454e-01 2.15817705e-01
-1.30160379e+00 4.01871473e-01 5.10783243e+00 8.39102209e-01
-7.50050366e-01 1.17634363e-01 2.86450505e-01 -8.29231739e-02
-3.13966036e-01 1.37644619e-01 -5.13265252e-01 5.18099070e-01
3.51589203e-01 -1.26436323e-01 6.34305239e-01 6.22536659e-01
4.12872016e-01 -1.68006748e-01 -6.75596654e-01 1.47354352e+00
5.51942825e-01 -8.38901341e-01 1.11790307e-01 3.01326394e-01
9.58847225e-01 -2.35054672e-01 -2.72902567e-02 -1.38901109e-02
5.80705367e-02 -4.15973097e-01 2.29751557e-01 6.02528274e-01
6.84514880e-01 -1.10783768e+00 7.30964661e-01 4.25031394e-01
-1.68660688e+00 -1.97957098e-01 -2.95417517e-01 3.07960868e-01
2.70060629e-01 7.45009065e-01 -1.11163624e-01 1.60441124e+00
6.25161529e-01 1.36997449e+00 -8.39286625e-01 7.43480802e-01
1.55062526e-01 2.73061484e-01 -7.53556713e-02 5.33108056e-01
2.42021963e-01 -1.05200410e+00 7.47455180e-01 7.88645685e-01
3.48426968e-01 4.03031319e-01 6.81432545e-01 4.56787616e-01
3.53799075e-01 4.17360544e-01 -7.88021684e-01 2.04757467e-01
2.84848869e-01 1.67734277e+00 -6.32001698e-01 -3.26174974e-01
-8.70421469e-01 9.78310704e-01 1.61026835e-01 4.86323565e-01
-5.50747454e-01 -8.68884102e-03 3.68636668e-01 -4.79902864e-01
4.25390601e-01 -3.66515428e-01 -1.78655908e-01 -1.70640934e+00
1.24937847e-01 -9.90680575e-01 6.34532690e-01 -5.09941280e-01
-1.55965674e+00 2.88060904e-01 5.82099073e-02 -2.01019835e+00
5.59813343e-02 -2.63837039e-01 -5.37965715e-01 7.23511100e-01
-1.23132229e+00 -1.28295708e+00 -3.18042517e-01 1.07363570e+00
4.04497117e-01 -4.06879842e-01 5.93424559e-01 4.35144186e-01
-9.18201983e-01 2.86123425e-01 6.55720711e-01 6.40557706e-02
6.73700929e-01 -1.13197172e+00 -5.62304676e-01 9.75141764e-01
9.17106345e-02 6.98844910e-01 2.58954048e-01 -5.45782626e-01
-1.96199083e+00 -9.43948805e-01 1.77312419e-01 -1.09513015e-01
4.42117751e-01 -1.18979312e-01 -8.54682565e-01 4.90782261e-01
3.07186395e-01 9.84168351e-02 1.31921089e+00 7.61179104e-02
-3.55974793e-01 -3.14464331e-01 -9.66972768e-01 4.62393314e-01
8.09819043e-01 -5.50279021e-01 -3.51919234e-01 1.99597389e-01
3.98508579e-01 -5.02963774e-02 -1.33409405e+00 5.02768815e-01
3.49195868e-01 -1.30400586e+00 1.10230064e+00 -4.28949147e-01
2.36290857e-01 -9.11335707e-01 -6.34289563e-01 -1.47047770e+00
-7.37552345e-01 -1.58332393e-01 -1.29082352e-01 1.59518671e+00
-6.76463395e-02 -4.86631364e-01 4.02508676e-01 6.24521226e-02
-9.69524980e-02 -6.85658395e-01 -9.55541015e-01 -7.56167054e-01
-2.31581643e-01 5.35809696e-02 3.01107168e-01 1.28275251e+00
4.92245145e-02 4.93652731e-01 -5.39284348e-01 4.52585161e-01
1.13582766e+00 6.45451784e-01 6.82236373e-01 -1.21138012e+00
-3.21080923e-01 -1.30612656e-01 -3.43573898e-01 -4.31762636e-01
9.28937644e-02 -1.02161551e+00 -2.49548748e-01 -1.54015207e+00
7.50982583e-01 -1.37795419e-01 -5.23149252e-01 -1.38344616e-01
-5.57448626e-01 -9.84471962e-02 5.71425200e-01 4.70429510e-01
-7.85815060e-01 7.46977687e-01 1.46248519e+00 -3.14843096e-02
-1.37291238e-01 -1.04132146e-01 -6.60060763e-01 8.49722862e-01
4.55107987e-01 -1.70736462e-01 -4.54393268e-01 -9.51158702e-02
-1.39027163e-01 4.43332672e-01 7.28608146e-02 -1.05553544e+00
3.99074286e-01 -3.12849581e-01 4.67543542e-01 -9.61644650e-01
4.60576326e-01 -1.19792330e+00 3.78079414e-01 3.28821242e-01
1.98559329e-01 1.22042626e-01 -3.18596989e-01 1.05389845e+00
-6.65736437e-01 8.84476975e-02 7.20713317e-01 -4.67323720e-01
-6.63457274e-01 4.11304206e-01 -1.04591824e-01 4.14043553e-02
1.07420111e+00 -2.09186509e-01 -5.23928404e-02 -1.78961217e-01
-8.03352356e-01 5.31007946e-01 4.89039510e-01 2.21872523e-01
6.66817665e-01 -1.86086345e+00 -7.47406244e-01 2.19528601e-01
3.12267601e-01 -2.67340634e-02 8.19358647e-01 1.18784106e+00
-2.88010575e-02 2.35167772e-01 -1.07264087e-01 -8.05042982e-01
-1.54675257e+00 9.98069823e-01 -1.75259307e-01 -4.14166361e-01
-3.34004015e-01 4.50760633e-01 2.77992368e-01 -5.18861055e-01
-3.03537190e-01 4.55695778e-01 -5.89003026e-01 6.19725347e-01
2.80533463e-01 7.73474276e-01 -3.01491082e-01 -1.28629279e+00
-5.11132061e-01 1.14476132e+00 1.67973489e-01 -1.07482813e-01
1.38971269e+00 -5.87940931e-01 -5.17145634e-01 7.16525435e-01
1.67421412e+00 1.20516345e-01 -1.03859913e+00 -3.73588741e-01
-2.51313269e-01 -5.46000600e-01 3.20391706e-03 -2.40491495e-01
-1.43280303e+00 7.79641509e-01 3.82977217e-01 -3.59283225e-03
1.37337458e+00 -2.98683941e-01 3.83424729e-01 2.28187323e-01
4.44691062e-01 -1.15462565e+00 1.51834875e-01 1.56386808e-01
7.84748614e-01 -1.37903285e+00 5.45345783e-01 -7.54706919e-01
-9.62801218e-01 1.03985691e+00 7.20626414e-01 -3.25856879e-02
8.91533613e-01 -3.99210542e-01 -3.47300172e-02 -4.02698517e-01
-4.83916312e-01 -2.81910241e-01 4.00510192e-01 2.84744978e-01
1.58368170e-01 6.72274306e-02 -5.26074529e-01 4.57253546e-01
3.12898934e-01 -4.29267675e-01 4.55668837e-01 8.03027749e-01
-1.25786170e-01 -1.13482440e+00 -7.27705538e-01 4.98824090e-01
-2.60981709e-01 7.89472163e-02 -2.10057497e-01 5.06496608e-01
2.62202591e-01 1.32516897e+00 -4.25682604e-01 -7.38760293e-01
2.43769869e-01 -2.17076335e-02 8.18025693e-02 -5.58911145e-01
-4.16896075e-01 7.37062693e-01 -3.23539734e-01 -5.44903815e-01
-9.71290529e-01 -1.01587558e+00 -1.03023589e+00 8.92139971e-04
-4.72732008e-01 2.52322644e-01 4.28456008e-01 7.29385912e-01
3.63640100e-01 3.66108119e-01 1.34992170e+00 -7.94767618e-01
-3.10506552e-01 -5.48013270e-01 -1.17255270e+00 6.37457907e-01
5.20353913e-02 -7.59910583e-01 -5.91565549e-01 5.11807501e-02] | [8.2442045211792, 4.623725414276123] |
1c74e2bb-46e0-4798-bbdd-f0dbfaacd973 | variational-autoencoders-for-semi-supervised | 1603.02514 | null | http://arxiv.org/abs/1603.02514v3 | http://arxiv.org/pdf/1603.02514v3.pdf | Variational Autoencoders for Semi-supervised Text Classification | Although semi-supervised variational autoencoder (SemiVAE) works in image
classification task, it fails in text classification task if using vanilla LSTM
as its decoder. From a perspective of reinforcement learning, it is verified
that the decoder's capability to distinguish between different categorical
labels is essential. Therefore, Semi-supervised Sequential Variational
Autoencoder (SSVAE) is proposed, which increases the capability by feeding
label into its decoder RNN at each time-step. Two specific decoder structures
are investigated and both of them are verified to be effective. Besides, in
order to reduce the computational complexity in training, a novel optimization
method is proposed, which estimates the gradient of the unlabeled objective
function by sampling, along with two variance reduction techniques.
Experimental results on Large Movie Review Dataset (IMDB) and AG's News corpus
show that the proposed approach significantly improves the classification
accuracy compared with pure-supervised classifiers, and achieves competitive
performance against previous advanced methods. State-of-the-art results can be
obtained by integrating other pretraining-based methods. | ['Ying Tan', 'Weidi Xu', 'Haoze Sun', 'Chao Deng'] | 2016-03-08 | null | null | null | null | ['semi-supervised-text-classification-1'] | ['natural-language-processing'] | [ 1.26078740e-01 2.19911024e-01 -5.05207717e-01 -3.47336054e-01
-8.18906903e-01 -2.54482597e-01 7.24234521e-01 -1.01762548e-01
-7.01864183e-01 7.83244073e-01 -8.51821005e-02 -2.81240761e-01
2.78152406e-01 -5.16439021e-01 -6.07740879e-01 -9.70312119e-01
6.12837613e-01 5.16296089e-01 5.52469902e-02 1.22329049e-01
2.20365569e-01 -2.23070607e-01 -1.61969054e+00 1.81574374e-01
9.33785319e-01 1.11520982e+00 5.28796971e-01 2.81271994e-01
-2.37583578e-01 1.19076407e+00 -4.48004723e-01 -4.99350041e-01
-1.58253506e-01 -8.51294816e-01 -6.37245715e-01 5.82635164e-01
-1.14653772e-02 -5.22054493e-01 -7.60800540e-02 1.32632005e+00
3.10059428e-01 2.72033751e-01 1.30304933e+00 -8.80629182e-01
-7.85063386e-01 8.10864329e-01 -3.49271536e-01 -3.98133732e-02
-1.44772992e-01 -3.01884264e-01 1.29362869e+00 -9.46116984e-01
4.83976901e-01 9.82805252e-01 2.84367085e-01 8.00205588e-01
-9.95911777e-01 -4.37595338e-01 2.38269240e-01 4.46820170e-01
-1.20059288e+00 -2.06589624e-01 1.04007518e+00 -4.78543997e-01
8.10763240e-01 -2.28734657e-01 4.56431210e-01 1.43220282e+00
2.27879107e-01 1.26418412e+00 1.38271439e+00 -5.42630970e-01
5.69152176e-01 6.69476032e-01 2.14830711e-01 8.00550282e-01
-3.87960970e-01 -4.33514826e-02 -3.11678320e-01 2.10308880e-01
4.62877363e-01 -1.33414581e-01 -3.16082984e-01 -5.65153182e-01
-9.19161201e-01 1.20546186e+00 1.62981614e-01 5.57479918e-01
-3.69827479e-01 4.13942412e-02 5.85986555e-01 2.63942748e-01
7.72763550e-01 -7.23027512e-02 -4.25086260e-01 -9.23318267e-02
-1.10526299e+00 -2.80891567e-01 5.61634958e-01 8.06183696e-01
3.92010689e-01 5.18315911e-01 -2.01429412e-01 1.06819212e+00
7.00515330e-01 3.89547467e-01 7.88876295e-01 -8.19816828e-01
2.05001950e-01 6.51862681e-01 -1.05894923e-01 -4.28730100e-01
-2.90939748e-01 -5.23700416e-01 -1.00743473e+00 1.26617625e-01
1.93131536e-01 -1.53003365e-01 -8.69218409e-01 1.47178686e+00
3.62970620e-01 1.95752397e-01 4.23835248e-01 9.37702239e-01
9.97465014e-01 9.12458420e-01 1.76730100e-02 -7.92631030e-01
1.29584455e+00 -1.14054215e+00 -1.30486417e+00 4.78549190e-02
6.38838947e-01 -2.82366186e-01 8.99885774e-01 6.01015270e-01
-9.31568623e-01 -8.05211723e-01 -1.19637227e+00 -4.46324870e-02
-3.36991012e-01 7.15230227e-01 3.65738124e-01 8.76784921e-01
-7.40443051e-01 5.67242026e-01 -8.33483100e-01 6.39096871e-02
3.02854359e-01 2.20057324e-01 3.48964892e-02 2.76058882e-01
-1.30744123e+00 9.46704388e-01 4.89542574e-01 1.95135981e-01
-1.27630496e+00 -1.78165883e-01 -1.01352370e+00 1.14853442e-01
2.43021473e-01 -2.40193099e-01 1.32013059e+00 -1.35990095e+00
-2.18257284e+00 6.06244743e-01 -1.27506539e-01 -4.91642505e-01
5.81437647e-01 8.32216442e-02 -3.04190397e-01 4.47445244e-01
-1.20193429e-01 7.16246247e-01 1.38302886e+00 -1.28600156e+00
-4.99485344e-01 -3.58254701e-01 -2.21933678e-01 3.86442006e-01
-7.83229649e-01 -2.31074676e-01 -3.68435919e-01 -4.79435295e-01
-2.04700664e-01 -7.47797310e-01 -3.77046205e-02 -2.82661945e-01
-2.93480724e-01 -6.98065758e-01 7.52153397e-01 -7.58736968e-01
1.17968798e+00 -2.14780140e+00 4.73755270e-01 2.44218800e-02
7.91846141e-02 2.79177666e-01 1.95639133e-01 1.01013981e-01
1.69059843e-01 -1.60752133e-01 -5.06209314e-01 -7.52914369e-01
5.86986281e-02 3.91582161e-01 -7.85623267e-02 6.05972230e-01
1.62857488e-01 7.39091396e-01 -7.20071435e-01 -9.07448471e-01
4.44428533e-01 6.16833985e-01 -4.68134016e-01 1.33312613e-01
-3.83489758e-01 4.53618318e-01 -4.29651111e-01 4.49901551e-01
5.09533882e-01 -4.94072914e-01 4.78968173e-01 2.91292798e-02
1.18555367e-01 9.63771790e-02 -1.04375148e+00 1.65317786e+00
-5.05209148e-01 6.40086710e-01 3.15451324e-02 -1.58139741e+00
9.58108127e-01 6.26648068e-01 3.44247580e-01 -6.47994936e-01
7.88023472e-01 2.35713180e-02 -1.65969744e-01 -5.30073404e-01
3.98929507e-01 -1.02633797e-01 2.09692255e-01 1.87281474e-01
5.73534429e-01 1.04606934e-01 2.66641021e-01 8.33785087e-02
4.64972675e-01 4.41259325e-01 3.08869302e-01 -3.53160173e-01
1.04255593e+00 -3.95235866e-02 4.22217399e-01 4.79568422e-01
-3.82547140e-01 3.64508271e-01 4.69589800e-01 7.29829222e-02
-8.79149377e-01 -8.59738290e-01 -4.70408618e-01 9.57598031e-01
1.46494061e-01 -1.16680801e-01 -1.09446585e+00 -1.01288116e+00
-3.12730610e-01 9.85571504e-01 -8.13361645e-01 -4.54948694e-02
-1.23012237e-01 -8.30092311e-01 2.78477281e-01 4.96492714e-01
7.87621439e-01 -9.24665153e-01 -4.67536807e-01 2.19578847e-01
-2.73590773e-01 -1.11764717e+00 -1.21318139e-02 3.53605866e-01
-9.32426929e-01 -9.51829910e-01 -1.04266548e+00 -1.04415703e+00
6.18111908e-01 -1.02905005e-01 6.75917387e-01 -2.34677672e-01
3.03373396e-01 3.30904394e-01 -6.38658524e-01 -7.68298581e-02
-6.85338199e-01 1.00478716e-01 -2.77714133e-02 3.64167988e-01
4.77838963e-01 -1.29794553e-01 -3.96961600e-01 2.35697493e-01
-5.93633175e-01 -2.55233884e-01 4.75883961e-01 1.41227388e+00
7.96760976e-01 1.64786428e-01 7.54119873e-01 -1.02603519e+00
5.50997078e-01 -4.09334570e-01 -5.60306609e-01 1.98189169e-01
-9.69511449e-01 1.17965080e-01 9.69230592e-01 -5.10932803e-01
-1.53298604e+00 2.64237896e-02 -4.15736824e-01 -5.24245143e-01
-1.10996239e-01 4.91862267e-01 7.88544491e-02 3.71325225e-01
5.28375804e-01 7.26017475e-01 3.00224483e-01 -4.27643687e-01
3.45144361e-01 1.16363120e+00 1.17233366e-01 -5.70021048e-02
1.88602641e-01 3.57406825e-01 -3.42440993e-01 -8.36331785e-01
-9.83390391e-01 -3.84853512e-01 -5.73050916e-01 -5.34342527e-01
1.25650120e+00 -1.08475661e+00 -8.07600856e-01 4.39720482e-01
-9.79920864e-01 -3.30752522e-01 -5.25494330e-02 7.56745100e-01
-5.15103459e-01 4.54755098e-01 -9.14831877e-01 -1.05543208e+00
-2.73234636e-01 -1.42305124e+00 7.75250673e-01 1.03331409e-01
2.05333546e-01 -1.20129180e+00 -1.33405700e-01 5.54899693e-01
1.14513272e-02 -3.03596467e-01 7.91336417e-01 -8.91121566e-01
-1.41063213e-01 -2.17568334e-02 2.15471331e-02 1.08387899e+00
-3.22878301e-01 -1.08814046e-01 -1.14731109e+00 -2.75539070e-01
3.76243979e-01 -7.57463217e-01 9.72547650e-01 6.37754083e-01
1.13990045e+00 -4.82009612e-02 -6.30914867e-02 2.33113289e-01
1.34463274e+00 3.83605748e-01 4.99784529e-01 1.35089487e-01
6.11857891e-01 4.52772945e-01 5.63092530e-01 3.21528345e-01
3.42347264e-01 3.71096104e-01 4.84299272e-01 2.32245177e-01
-1.72293745e-03 -1.64007440e-01 6.28255785e-01 1.55783176e+00
5.61887622e-02 -4.08923835e-01 -3.01535010e-01 2.86524475e-01
-1.90772390e+00 -7.27694273e-01 -2.91842222e-01 1.80108929e+00
6.51617646e-01 2.66855061e-01 -1.50862679e-01 5.13735056e-01
5.88550985e-01 3.01635891e-01 -6.26875341e-01 -5.57743192e-01
3.45438975e-03 -3.29818167e-02 2.79525965e-01 5.36630392e-01
-1.27195084e+00 9.96193051e-01 6.18948650e+00 1.33536458e+00
-1.02126443e+00 5.67396939e-01 5.72173297e-01 2.90348649e-01
-1.01197474e-01 -3.39419574e-01 -7.90872216e-01 5.02422690e-01
1.11108351e+00 5.54177880e-01 2.55323589e-01 1.14613760e+00
-9.38766636e-03 -2.02102378e-01 -8.62817049e-01 1.03653252e+00
4.95431632e-01 -9.84034717e-01 -2.82664150e-02 -1.67600021e-01
7.85918415e-01 -2.27745175e-01 1.30092680e-01 7.37196028e-01
5.83080240e-02 -6.60397291e-01 5.77121377e-01 2.74887472e-01
6.58224702e-01 -9.29194331e-01 9.39972639e-01 7.90656865e-01
-9.01840627e-01 -1.14719555e-01 -6.45328522e-01 -7.03333840e-02
4.72796932e-02 4.81431365e-01 -5.52781105e-01 4.37916845e-01
4.35486972e-01 1.00397289e+00 -3.82689863e-01 2.41264850e-01
-5.28339148e-01 1.08483088e+00 1.54052809e-01 -7.74560332e-01
3.08388501e-01 -5.20473778e-01 2.96683490e-01 9.78434980e-01
-6.99116811e-02 -8.57526213e-02 7.63982683e-02 8.04791093e-01
-1.30753458e-01 4.43032503e-01 -4.32183355e-01 -9.20747593e-02
1.22340180e-01 1.18155777e+00 -7.68222630e-01 -6.28718436e-01
-4.35345709e-01 1.24719644e+00 3.48873913e-01 3.80814523e-01
-9.54341590e-01 -1.89526260e-01 -9.80718285e-02 -5.75949848e-01
5.22775948e-01 3.17986533e-02 -1.89008579e-01 -1.24922085e+00
-2.03910649e-01 -8.14990699e-01 1.99657738e-01 -7.44492412e-01
-9.18174207e-01 6.37755871e-01 -2.65174031e-01 -1.38177860e+00
-5.00038922e-01 -8.55440497e-01 -5.32166287e-02 3.06631148e-01
-1.56894183e+00 -9.42453861e-01 1.97016791e-01 6.81226134e-01
1.00715017e+00 -5.31958580e-01 8.56976926e-01 3.85634512e-01
-8.69992554e-01 7.22316384e-01 6.90675676e-01 1.78359389e-01
2.48095229e-01 -1.29661584e+00 -3.45880598e-01 7.19494045e-01
3.67693931e-01 7.41921440e-02 5.79673946e-01 -2.74082959e-01
-1.22627437e+00 -8.52047145e-01 7.36970544e-01 -1.78469926e-01
5.27217805e-01 -4.54181284e-01 -7.82925189e-01 4.53141630e-01
5.80811501e-01 -3.67413372e-01 7.13097930e-01 -1.80584386e-01
-1.18461624e-01 1.75227419e-01 -1.00317097e+00 4.32176679e-01
5.21135271e-01 -6.25226378e-01 -6.98496997e-01 4.10801709e-01
5.39568543e-01 -2.63620019e-01 -9.56949890e-01 2.55613655e-01
2.91413099e-01 -8.95029247e-01 5.82417190e-01 -2.65000582e-01
9.37089503e-01 -9.30675864e-02 -1.37912691e-01 -1.32026899e+00
-1.55014828e-01 -1.15179785e-01 -5.20380735e-01 1.14918423e+00
4.69142258e-01 -4.79385227e-01 7.47634828e-01 -7.86868017e-03
-9.35257375e-02 -9.48175907e-01 -9.18348134e-01 -7.15614557e-01
1.10330909e-01 -5.69636762e-01 -2.53774226e-01 1.01236582e+00
1.45144731e-01 7.78260112e-01 -6.32625818e-01 -1.46309674e-01
6.80409372e-01 -5.55363595e-02 3.55556160e-01 -1.08272839e+00
-4.41042632e-01 -3.38024825e-01 -1.56197056e-01 -1.48323655e+00
6.86489463e-01 -1.06748128e+00 8.91689211e-02 -1.54629183e+00
1.21290147e-01 2.00654551e-01 -3.61774921e-01 1.91740438e-01
-5.18220998e-02 2.43731320e-01 -5.78557178e-02 5.41392304e-02
-6.25594437e-01 1.04521334e+00 1.38804960e+00 -3.22285116e-01
1.44828381e-02 7.83019699e-03 -2.29015827e-01 7.72796333e-01
7.30417907e-01 -4.73784745e-01 -6.01308048e-01 -8.89976546e-02
8.17332044e-03 2.78887182e-01 8.74580219e-02 -8.42181265e-01
4.09639962e-02 3.38819295e-01 3.72250110e-01 -7.91991830e-01
4.50603545e-01 -8.51507783e-01 -4.82752532e-01 5.71837962e-01
-6.61109626e-01 -3.88585806e-01 -2.27810457e-01 8.44375253e-01
-4.54490840e-01 -8.33440661e-01 9.18727398e-01 5.43635292e-03
-6.89486802e-01 -1.06283896e-01 -9.19489205e-01 3.89330611e-02
9.87269998e-01 1.18957171e-02 1.31115988e-01 -2.88163155e-01
-7.43740201e-01 2.52314031e-01 -3.20255533e-02 2.53755748e-01
6.87155843e-01 -1.10274684e+00 -5.76150835e-01 8.11528191e-02
-1.59639549e-02 -3.67979318e-01 4.40732002e-01 8.67243528e-01
-3.55161488e-01 3.23423296e-01 8.20776299e-02 -7.76534975e-01
-1.05848491e+00 7.54774570e-01 2.29826838e-01 -3.16311359e-01
-5.74698329e-01 8.51452827e-01 -1.24973759e-01 -5.22472382e-01
6.06888771e-01 -2.50075459e-01 -8.33355188e-01 4.66797948e-01
3.68522346e-01 2.15019390e-01 -4.24485952e-02 -6.68196201e-01
-5.15315011e-02 3.65882128e-01 -1.25764206e-01 -3.28959852e-01
9.78748679e-01 -2.61317253e-01 2.15840191e-01 8.66364837e-01
1.48565841e+00 -4.71554637e-01 -1.40044832e+00 -2.79057652e-01
-2.20675543e-01 9.36024636e-02 5.41882217e-01 -5.49329400e-01
-1.17490757e+00 1.23808205e+00 7.86213279e-01 2.79337704e-01
9.73651171e-01 -2.85528898e-01 5.41297257e-01 5.64787805e-01
1.30849063e-01 -1.68993855e+00 1.99962631e-01 3.58276933e-01
4.73659813e-01 -1.59150779e+00 -1.05816722e-01 -1.37831688e-01
-1.14493859e+00 1.02841043e+00 5.04142940e-01 -2.68272460e-01
9.45928156e-01 -8.06133002e-02 1.27514482e-01 6.32862598e-02
-8.10558498e-01 -2.37065583e-01 2.92539865e-01 3.45352739e-01
4.17387336e-01 7.96818826e-03 -7.50578523e-01 5.99749744e-01
1.96878493e-01 -2.08374411e-01 4.79265660e-01 6.96223199e-01
-4.70496982e-01 -9.23808753e-01 -1.79573651e-02 4.73645836e-01
-6.29454851e-01 -6.18222021e-02 -8.28175992e-02 7.50198662e-01
-6.57114461e-02 1.09847546e+00 7.87592158e-02 -3.04703057e-01
-2.15804011e-01 4.32314485e-01 3.49486947e-01 -3.15149456e-01
-4.86302763e-01 4.91863132e-01 9.79958568e-04 -2.02051625e-01
-8.72330487e-01 -6.76405549e-01 -1.16810751e+00 3.22429448e-01
-7.31546044e-01 3.42346013e-01 9.51760888e-01 1.31825280e+00
-2.05346122e-01 8.35449636e-01 7.95853734e-01 -6.46417081e-01
-9.47827101e-01 -1.16759825e+00 -7.91637957e-01 2.64983088e-01
1.47974297e-01 -8.47473264e-01 -5.28207242e-01 1.92658022e-01] | [9.50025463104248, 2.9934909343719482] |
f66b8c94-98a0-4c89-9473-125e678a0576 | unsupervised-deformable-ultrasound-image | 2306.13329 | null | https://arxiv.org/abs/2306.13329v1 | https://arxiv.org/pdf/2306.13329v1.pdf | Unsupervised Deformable Ultrasound Image Registration and Its Application for Vessel Segmentation | This paper presents a deep-learning model for deformable registration of ultrasound images at online rates, which we call U-RAFT. As its name suggests, U-RAFT is based on RAFT, a convolutional neural network for estimating optical flow. U-RAFT, however, can be trained in an unsupervised manner and can generate synthetic images for training vessel segmentation models. We propose and compare the registration quality of different loss functions for training U-RAFT. We also show how our approach, together with a robot performing force-controlled scans, can be used to generate synthetic deformed images to significantly expand the size of a femoral vessel segmentation training dataset without the need for additional manual labeling. We validate our approach on both a silicone human tissue phantom as well as on in-vivo porcine images. We show that U-RAFT generates synthetic ultrasound images with 98% and 81% structural similarity index measure (SSIM) to the real ultrasound images for the phantom and porcine datasets, respectively. We also demonstrate that synthetic deformed images from U-RAFT can be used as a data augmentation technique for vessel segmentation models to improve intersection-over-union (IoU) segmentation performance | ['Howie Choset', 'John Galeotti', 'Ananya Bal', 'Andrew L. Orekhov', 'FNU Abhimanyu'] | 2023-06-23 | null | null | null | null | ['optical-flow-estimation', 'image-registration'] | ['computer-vision', 'computer-vision'] | [ 2.45464802e-01 5.76733589e-01 2.62690037e-01 -3.05318892e-01
-6.89639151e-01 -5.87037683e-01 -6.54754192e-02 -5.44697978e-02
-2.16026664e-01 4.84640092e-01 1.09734572e-01 -3.15171868e-01
1.00476280e-01 -7.72695780e-01 -8.85067344e-01 -4.22628105e-01
-4.74406391e-01 5.25725842e-01 4.34974700e-01 -1.86918467e-01
-3.11398923e-01 6.36028707e-01 -6.57290995e-01 1.18883774e-01
9.49376047e-01 9.30799127e-01 -1.32980645e-01 8.53567481e-01
2.39151835e-01 7.13387012e-01 -3.95731837e-01 -3.79058957e-01
8.02388012e-01 -5.63480437e-01 -9.77984905e-01 1.24689907e-01
7.19734967e-01 -8.46124470e-01 -4.84789163e-01 5.21281302e-01
8.30652714e-01 9.29253548e-02 6.35911047e-01 -8.74725997e-01
-2.78128028e-01 6.64671540e-01 -4.89335388e-01 1.41674802e-01
7.75153413e-02 2.88782209e-01 2.84509391e-01 -6.02869034e-01
1.04627919e+00 1.08517218e+00 1.03947651e+00 5.98472834e-01
-1.47525465e+00 -4.41487551e-01 -6.82081997e-01 -5.37835658e-01
-9.60324705e-01 -3.57752353e-01 4.50075626e-01 -7.23236620e-01
3.18772256e-01 2.45677486e-01 7.33251929e-01 6.59697771e-01
3.45998257e-01 5.70447803e-01 9.96419370e-01 -1.78893268e-01
1.00933686e-01 -2.93204308e-01 -1.09685771e-01 1.10377765e+00
1.64645955e-01 2.48686776e-01 6.42953021e-03 -5.60230613e-02
1.84985662e+00 -6.44560158e-01 -5.27550280e-01 -5.59089065e-01
-1.45057714e+00 7.47943819e-01 6.52634859e-01 9.34061855e-02
-3.61135721e-01 4.32838380e-01 5.62725306e-01 2.07580820e-01
3.47589433e-01 6.51311338e-01 -1.14803888e-01 8.00763536e-03
-7.77865410e-01 5.00043519e-02 9.02653277e-01 8.70055854e-01
4.31020528e-01 3.43101770e-01 -4.10442382e-01 8.49007189e-01
2.50597924e-01 2.21069947e-01 5.01903832e-01 -1.58582127e+00
1.77941099e-03 1.38064891e-01 2.22756356e-01 -9.31700230e-01
-5.70175350e-01 -4.46698487e-01 -8.21588993e-01 4.51328754e-01
8.87465656e-01 -5.99703372e-01 -1.22020686e+00 1.49628246e+00
5.44720888e-01 6.97921276e-01 1.37684122e-01 1.32323062e+00
1.13229644e+00 3.36872697e-01 4.71402667e-02 -1.02138072e-01
1.03507972e+00 -9.20658290e-01 -4.53472555e-01 1.98986754e-01
9.11394775e-01 -7.65983045e-01 8.21019530e-01 4.76161875e-02
-1.42016625e+00 -5.12382746e-01 -8.18072140e-01 1.38659924e-01
5.24806261e-01 -6.41610175e-02 7.40822375e-01 7.24588811e-01
-1.12664509e+00 9.08815861e-01 -1.29841220e+00 -1.30935028e-01
7.46347129e-01 3.36269140e-01 -4.60512191e-01 6.88711479e-02
-9.28066492e-01 7.97479033e-01 6.17873557e-02 1.92647740e-01
-1.00248051e+00 -1.19182062e+00 -1.10966551e+00 -2.28652954e-01
3.82507592e-02 -8.36358845e-01 1.18074381e+00 -5.73428214e-01
-1.75004745e+00 8.38387132e-01 2.57150084e-01 -7.07021415e-01
9.36992526e-01 -1.73460823e-02 6.62699714e-02 6.99107587e-01
1.74994648e-01 9.15882826e-01 5.30851781e-01 -1.32328212e+00
7.76388869e-02 2.36730367e-01 -1.19272597e-01 -1.24838918e-01
2.77958274e-01 -5.05113974e-02 -3.43939453e-01 -9.89910603e-01
2.31428668e-01 -1.21812272e+00 -6.95280552e-01 7.72483885e-01
-3.83312643e-01 4.55639482e-01 5.17808735e-01 -7.64253378e-01
6.21587217e-01 -1.66757607e+00 -1.72536612e-01 4.54731375e-01
3.95631164e-01 3.94343734e-01 -4.58996177e-01 -1.58068091e-01
-2.01381326e-01 1.61712140e-01 -6.66663408e-01 5.31952418e-02
-5.39315224e-01 3.87630165e-01 -2.20934618e-02 5.00351250e-01
8.24810639e-02 1.04207540e+00 -1.04219782e+00 -7.55541384e-01
3.01391900e-01 3.99754256e-01 -7.71362662e-01 2.23482832e-01
1.02730326e-01 1.25241959e+00 -3.06439638e-01 2.63941318e-01
8.16575170e-01 -1.64013192e-01 -4.89558186e-03 -4.90056098e-01
1.20763332e-01 -3.38718742e-01 -1.03258336e+00 1.95665669e+00
-5.53724647e-01 5.56721807e-01 1.37098774e-01 -9.58375752e-01
8.64417493e-01 5.99148810e-01 1.06593204e+00 -5.18720508e-01
4.85613257e-01 4.69728023e-01 3.01998883e-01 -7.04008937e-01
1.86991263e-02 -3.11171353e-01 2.41837054e-01 5.37072778e-01
2.04758987e-01 -4.89207715e-01 3.81408364e-01 1.31208807e-01
8.52504432e-01 3.60179245e-01 -2.94395357e-01 -3.82162660e-01
3.39610130e-01 4.09008749e-02 5.70335865e-01 7.16077030e-01
-2.57215708e-01 1.01839828e+00 4.58062381e-01 -6.08849704e-01
-1.16183138e+00 -1.11217380e+00 -5.13989747e-01 3.18791956e-01
4.31977093e-01 1.95143640e-01 -9.10306394e-01 -5.75323343e-01
3.62641960e-02 2.98622072e-01 -5.23162305e-01 6.61662593e-02
-1.08474994e+00 -7.55812287e-01 7.27850795e-01 5.98071098e-01
4.17272508e-01 -1.04903316e+00 -8.84009123e-01 5.91304243e-01
-2.57760495e-01 -1.50268054e+00 -5.88950515e-01 -2.67217785e-01
-1.05853760e+00 -1.18424356e+00 -1.13782978e+00 -1.05754852e+00
8.95230472e-01 -1.99686632e-01 1.08022964e+00 1.11942276e-01
-5.74858189e-01 3.28804672e-01 -2.91296273e-01 -1.94384113e-01
-1.20942080e+00 -4.42615747e-01 -1.77311882e-01 -3.41951512e-02
-6.07230484e-01 -7.24147618e-01 -1.11928403e+00 6.05189919e-01
-7.81244338e-01 1.35497749e-01 3.23285848e-01 7.96709061e-01
6.11232996e-01 -5.57785273e-01 5.24827838e-01 -1.09399462e+00
5.04749358e-01 -8.90351832e-02 -5.67952633e-01 -8.80018175e-02
-1.39763340e-01 -8.03342611e-02 2.61283189e-01 -5.08910418e-01
-9.77754354e-01 1.58081323e-01 -2.95664877e-01 -7.37872839e-01
3.85784432e-02 4.92475808e-01 7.37591863e-01 -7.05392659e-01
1.01448023e+00 2.00189461e-04 6.26096606e-01 -1.48299620e-01
3.57196808e-01 3.37029666e-01 7.68313885e-01 -6.13952637e-01
6.92323148e-01 6.56773031e-01 3.33295554e-01 -7.26705909e-01
-3.97054315e-01 -1.06536902e-01 -8.72691810e-01 -5.42812943e-01
8.28417122e-01 -6.43042803e-01 -7.15688288e-01 6.75252736e-01
-1.07835865e+00 -8.63535345e-01 -6.71583951e-01 7.36651242e-01
-8.58615279e-01 6.43042684e-01 -1.22376525e+00 -2.87832171e-01
-5.25904357e-01 -1.45225251e+00 9.97029901e-01 2.70002127e-01
-2.25538522e-01 -1.25222647e+00 -2.16861796e-02 3.17126185e-01
7.52551496e-01 1.11448014e+00 6.56108439e-01 -2.29864419e-01
-4.86414671e-01 -8.29213262e-02 -2.17863411e-01 4.87591088e-01
1.99639976e-01 -1.56437397e-01 -6.12094939e-01 -3.30533952e-01
-1.98506415e-01 -4.19715792e-01 5.65549970e-01 9.17725146e-01
1.04035318e+00 -1.88598961e-01 -1.99081406e-01 7.87073731e-01
1.19158626e+00 1.60265967e-01 7.81587720e-01 -4.64245304e-02
7.30480671e-01 4.90869343e-01 4.79146868e-01 2.60669291e-01
-1.42008520e-03 6.37211502e-01 4.55641031e-01 -8.05874109e-01
-6.12101316e-01 7.01536760e-02 -2.53585249e-01 6.15292728e-01
-2.84213126e-01 1.01759806e-01 -8.63507628e-01 5.84341943e-01
-1.46618783e+00 -5.28142512e-01 -3.33820045e-01 2.10127282e+00
1.10116374e+00 -1.54131532e-01 1.76188331e-02 -3.52444023e-01
6.71645463e-01 -3.18128496e-01 -4.53164935e-01 -2.72373229e-01
2.39916995e-01 6.92673862e-01 5.08796394e-01 6.12087011e-01
-1.34742415e+00 6.44105554e-01 6.84472084e+00 2.57741421e-01
-1.46342969e+00 4.12971340e-02 8.71073663e-01 4.25236464e-01
-7.47485757e-02 -2.17166737e-01 1.20277114e-01 2.00848937e-01
7.07454801e-01 -2.27085943e-03 -1.12786539e-01 5.21324456e-01
4.28659588e-01 1.22296959e-02 -9.37042892e-01 7.18946576e-01
-3.36716592e-01 -1.72093236e+00 1.38101336e-02 -1.88611642e-01
7.78406620e-01 -2.62144748e-02 -2.24923730e-01 -2.07828861e-02
4.44961250e-01 -1.07023549e+00 4.57309961e-01 5.04672647e-01
1.11856949e+00 -3.30343962e-01 8.09092343e-01 2.18919329e-02
-9.29334521e-01 5.02925098e-01 -1.50249392e-01 5.92711866e-01
6.68378949e-01 3.48732263e-01 -1.17918801e+00 6.39254391e-01
4.63290572e-01 6.36746228e-01 -3.45071673e-01 1.53958738e+00
2.14391753e-01 6.76908433e-01 -3.49915802e-01 6.34630322e-01
1.05400957e-01 -3.56309772e-01 6.79945290e-01 1.10255313e+00
4.12701130e-01 3.06637082e-02 2.60057956e-01 9.48588908e-01
-4.34925631e-02 9.40104127e-02 -2.37313882e-01 2.41152793e-01
1.99681044e-01 1.20134640e+00 -8.35465014e-01 -3.69232029e-01
-1.09865494e-01 7.37908483e-01 -2.86491454e-01 3.10664177e-01
-9.01508689e-01 -3.18316549e-01 3.12292039e-01 5.00960469e-01
7.29577392e-02 -1.88049480e-01 -1.35055304e-01 -8.81092787e-01
-3.15475374e-01 -4.84158605e-01 1.78641826e-01 -7.84105420e-01
-1.15963030e+00 7.59100914e-01 -2.37174390e-04 -1.63821745e+00
-1.82139203e-01 -4.33601916e-01 -5.01707077e-01 8.44659209e-01
-1.46517372e+00 -9.64034438e-01 -7.35542893e-01 3.36041868e-01
2.01794967e-01 1.01378083e-01 8.17653596e-01 3.52395386e-01
-1.74466893e-01 5.73066711e-01 -1.60054862e-01 7.52626717e-01
6.14020109e-01 -1.21596754e+00 5.31391025e-01 6.72251999e-01
-3.18847388e-01 2.97601402e-01 6.41250789e-01 -5.44506311e-01
-8.78093302e-01 -1.44871783e+00 -1.86251312e-01 -1.96621791e-02
5.19123554e-01 2.27177843e-01 -9.15045083e-01 7.29962885e-01
1.92986414e-01 7.84072638e-01 6.94773078e-01 -8.53668928e-01
2.41749778e-01 1.39000833e-01 -1.51397908e+00 5.95980823e-01
8.78075361e-01 2.28090137e-01 -1.25630558e-01 5.33302486e-01
8.21374178e-01 -1.34121394e+00 -1.69960546e+00 7.04295516e-01
5.27145445e-01 -7.06096411e-01 9.63678956e-01 -3.62693280e-01
6.22468114e-01 -2.75521278e-01 5.48065186e-01 -1.31203449e+00
3.41286473e-02 -1.03881431e+00 1.98397413e-01 5.20676136e-01
2.71108031e-01 -6.65086210e-01 1.00983131e+00 6.29409254e-01
-5.76555669e-01 -7.64685929e-01 -1.17005181e+00 -5.62312901e-01
3.70165169e-01 -1.90524876e-01 1.53737739e-01 1.00402677e+00
-3.29917520e-01 -1.11021213e-01 -1.92749947e-01 1.49097651e-01
8.39766741e-01 1.00464605e-01 6.94644988e-01 -1.06661761e+00
-2.20725149e-01 -5.38889095e-02 -4.93266523e-01 -1.17037582e+00
-1.70096472e-01 -1.00593638e+00 1.90315083e-01 -1.61907148e+00
-3.31521183e-01 -1.07483065e+00 2.64864773e-01 5.50570428e-01
1.57518625e-01 7.81235099e-01 1.04566133e-02 3.14983130e-01
1.41499624e-01 2.65557230e-01 2.49680305e+00 -6.92738444e-02
-3.76047999e-01 1.00125015e-01 -2.50220478e-01 5.43567061e-01
5.74157357e-01 -2.93742687e-01 -2.81039387e-01 -3.02688986e-01
-4.81250703e-01 7.95783758e-01 4.05340940e-01 -1.16787159e+00
-1.05926126e-01 1.33187860e-01 2.30498075e-01 6.21894784e-02
4.39022742e-02 -6.88852966e-01 1.42887264e-01 8.10766637e-01
-3.43505770e-01 -4.64763552e-01 3.66825402e-01 3.01591635e-01
-7.15777501e-02 -1.70135200e-01 1.15615201e+00 -4.09744591e-01
-2.98696697e-01 4.97039139e-01 -3.85399312e-01 7.32755810e-02
1.04022908e+00 -4.33095276e-01 -2.56864041e-01 -3.96945477e-01
-1.23385620e+00 2.51520872e-01 1.62268370e-01 9.31167752e-02
7.27512479e-01 -1.18805850e+00 -1.07612801e+00 3.08187574e-01
-2.01300994e-01 4.94039714e-01 3.39652926e-01 1.17763364e+00
-1.55146694e+00 -4.55747321e-02 -3.34605873e-01 -9.65870857e-01
-8.75738621e-01 9.71596912e-02 1.05068207e+00 -1.68586716e-01
-9.19316709e-01 7.52674282e-01 2.54463732e-01 -6.87801182e-01
-1.66190699e-01 -6.21138096e-01 -1.00023903e-01 -4.98474270e-01
-4.63612750e-03 9.51549932e-02 9.97324884e-02 -7.78591752e-01
-2.07427014e-02 6.48309469e-01 1.46071076e-01 1.12270325e-01
1.27657723e+00 1.14759967e-01 -2.84558255e-02 -2.75047123e-01
1.04820907e+00 -1.48399860e-01 -1.54494512e+00 -1.39992461e-01
-4.78974670e-01 -5.00604212e-01 1.98310474e-03 -7.23788500e-01
-1.70548534e+00 5.93713045e-01 8.61814857e-01 -5.12454361e-02
8.35534096e-01 -1.84569180e-01 1.29285967e+00 -1.15606472e-01
3.65967333e-01 -4.41831499e-01 2.76382506e-01 1.00597955e-01
1.01145625e+00 -1.06647468e+00 -2.28958800e-01 -1.17005265e+00
-6.44767582e-01 1.38758218e+00 6.26124799e-01 -4.30942565e-01
7.27974951e-01 6.46847010e-01 5.90950668e-01 -9.69070662e-03
3.72159518e-02 9.21757221e-02 1.25884086e-01 7.56680608e-01
5.29182255e-01 -1.68843180e-01 -3.10283631e-01 -6.98130578e-02
-2.44428769e-01 2.88811415e-01 9.90229189e-01 9.28876102e-01
-1.19522899e-01 -1.01948190e+00 -2.02257559e-01 5.21540105e-01
-4.00594056e-01 1.38505325e-01 3.97021323e-01 8.45655859e-01
-1.93273664e-01 4.42400515e-01 1.61165118e-01 -2.56493073e-02
5.57742059e-01 -5.55610001e-01 7.18719780e-01 -4.81469959e-01
-6.55628085e-01 2.62193978e-01 1.08009256e-01 -6.72697961e-01
-6.26360595e-01 -2.53076911e-01 -1.47630382e+00 6.65663853e-02
-9.35266763e-02 -1.74517639e-03 3.39231402e-01 6.45935953e-01
1.95466012e-01 7.42587566e-01 5.83752275e-01 -1.12367725e+00
-2.01134339e-01 -9.38690960e-01 -4.59461302e-01 6.68134630e-01
2.43039399e-01 -6.08361125e-01 -1.81180045e-01 4.85748231e-01] | [14.13929271697998, -2.403764247894287] |
6747d81c-4b69-463d-b861-25196046c5ec | elc-ois-ellipsoidal-clustering-for-open-world | 2303.04351 | null | https://arxiv.org/abs/2303.04351v1 | https://arxiv.org/pdf/2303.04351v1.pdf | ElC-OIS: Ellipsoidal Clustering for Open-World Instance Segmentation on LiDAR Data | Open-world Instance Segmentation (OIS) is a challenging task that aims to accurately segment every object instance appearing in the current observation, regardless of whether these instances have been labeled in the training set. This is important for safety-critical applications such as robust autonomous navigation. In this paper, we present a flexible and effective OIS framework for LiDAR point cloud that can accurately segment both known and unknown instances (i.e., seen and unseen instance categories during training). It first identifies points belonging to known classes and removes the background by leveraging close-set panoptic segmentation networks. Then, we propose a novel ellipsoidal clustering method that is more adapted to the characteristic of LiDAR scans and allows precise segmentation of unknown instances. Furthermore, a diffuse searching method is proposed to handle the common over-segmentation problem presented in the known instances. With the combination of these techniques, we are able to achieve accurate segmentation for both known and unknown instances. We evaluated our method on the SemanticKITTI open-world LiDAR instance segmentation dataset. The experimental results suggest that it outperforms current state-of-the-art methods, especially with a 10.0% improvement in association quality. The source code of our method will be publicly available at https://github.com/nubot-nudt/ElC-OIS. | ['Xieyuanli Chen', 'Zhiqiang Zheng', 'Huimin Lu', 'Qinghua Yu', 'Kaihong Huang', 'Wenbang Deng'] | 2023-03-08 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 2.14018404e-01 -1.55691030e-02 -2.91291270e-02 -3.86348993e-01
-8.50055993e-01 -5.94868898e-01 3.72032881e-01 9.13331136e-02
-3.48938942e-01 6.34580970e-01 -8.73854041e-01 -1.73693538e-01
-2.66577452e-01 -1.01741731e+00 -7.63296008e-01 -6.65827155e-01
9.81920213e-02 1.11538756e+00 7.62121320e-01 1.56506926e-01
2.33626917e-01 8.51040423e-01 -1.98796868e+00 -1.62836015e-01
1.31571960e+00 1.16458929e+00 4.77095157e-01 2.32281759e-01
-3.49267036e-01 1.71841774e-02 -4.96631593e-01 -9.76627991e-02
6.44648552e-01 1.08909875e-01 -6.65578723e-01 1.41751900e-01
8.33910346e-01 -1.89485028e-02 2.96983093e-01 1.14819443e+00
1.57675847e-01 2.30209917e-01 7.47419059e-01 -1.31457829e+00
2.04710243e-03 4.58842255e-02 -7.26373971e-01 2.01459333e-01
-4.86577719e-01 -1.96906887e-02 8.37664306e-01 -7.73725152e-01
5.87545276e-01 9.94995356e-01 5.57479024e-01 2.64233977e-01
-1.21995986e+00 -1.01948953e+00 2.12204263e-01 4.06440943e-01
-1.63936782e+00 -1.83229253e-01 6.08413875e-01 -4.34920281e-01
3.86235863e-01 4.16179657e-01 6.14207923e-01 4.98033822e-01
-1.64907023e-01 6.70005798e-01 1.06593657e+00 -8.42653289e-02
4.06167924e-01 2.12010950e-01 4.48382735e-01 3.92119825e-01
4.26887780e-01 2.92449407e-02 -3.13314311e-02 -5.30570261e-02
5.07699907e-01 2.13678107e-01 -9.81579497e-02 -6.83240235e-01
-9.33218181e-01 8.43131602e-01 6.33709908e-01 6.31907135e-02
-4.16161805e-01 -1.89256877e-01 9.87256691e-02 -2.94733018e-01
6.81653857e-01 2.01415733e-01 -6.03525102e-01 2.69721359e-01
-1.08081210e+00 9.70175117e-02 7.02297151e-01 9.85578358e-01
1.17877102e+00 -1.60921350e-01 1.78685069e-01 8.83946478e-01
3.49091977e-01 8.72105598e-01 4.19721007e-02 -1.16783392e+00
4.04053599e-01 8.26482773e-01 3.52315903e-02 -8.27594280e-01
-2.85970420e-01 -4.24216032e-01 -4.99331325e-01 3.24670553e-01
4.09988672e-01 1.65540248e-01 -1.24990571e+00 1.17642033e+00
9.35269773e-01 6.16269827e-01 1.12825446e-01 7.64569461e-01
8.34270477e-01 6.57008708e-01 -1.48535371e-02 -7.76127772e-03
1.39683056e+00 -7.64268696e-01 -3.75190824e-01 -3.68954152e-01
6.84857070e-02 -5.80605030e-01 7.21780300e-01 2.88099021e-01
-4.72920954e-01 -6.16950691e-01 -9.32140589e-01 2.58137107e-01
-6.57277346e-01 2.13479578e-01 4.83299553e-01 5.43679535e-01
-6.86729789e-01 5.07773817e-01 -8.79391015e-01 -4.30672705e-01
5.04191518e-01 4.30003166e-01 -2.00290024e-01 -8.63905773e-02
-7.95049608e-01 5.95774472e-01 7.16761231e-01 2.59536684e-01
-6.29615605e-01 -6.86694205e-01 -7.33503401e-01 -1.48173422e-01
9.17639375e-01 -4.19915527e-01 1.07482469e+00 -5.73878169e-01
-9.66066957e-01 8.89188290e-01 -3.08244765e-01 -7.00124145e-01
4.20054525e-01 -2.17059910e-01 -3.23550075e-01 3.39457721e-01
3.48050177e-01 8.90768766e-01 9.21163023e-01 -1.57823586e+00
-1.03686857e+00 -6.64802253e-01 -1.15009055e-01 1.63022265e-01
1.34773642e-01 -3.71326804e-01 -6.61064923e-01 -1.55773774e-01
5.67863464e-01 -1.13735259e+00 -3.87507886e-01 5.97579591e-02
-3.73102784e-01 -2.20135123e-01 1.25535464e+00 -3.18080306e-01
7.94281840e-01 -2.10874081e+00 -1.89159721e-01 4.06554341e-01
1.46079853e-01 5.38843274e-01 1.96316093e-01 1.36940882e-01
1.30246818e-01 1.02321386e-01 -7.85489798e-01 -2.81151593e-01
-1.87927172e-01 5.80333292e-01 -3.16597611e-01 3.39005709e-01
9.84762758e-02 5.55129826e-01 -7.03108907e-01 -6.70715690e-01
7.04020381e-01 3.02585185e-01 -1.32258654e-01 6.40208572e-02
-4.52526897e-01 6.76909864e-01 -3.88213724e-01 8.20717990e-01
1.22020316e+00 7.88952857e-02 -1.83847681e-01 -9.54468921e-02
-3.02319467e-01 -2.79227883e-01 -1.56701148e+00 1.28095770e+00
-3.07104170e-01 2.90602624e-01 1.67608976e-01 -8.15869331e-01
9.92722988e-01 -8.50538537e-02 5.79674065e-01 -5.34978867e-01
2.94797514e-02 5.45648396e-01 -2.30880782e-01 -2.69714147e-01
6.34589076e-01 -8.55845883e-02 1.78575262e-01 -4.85364124e-02
-6.31957352e-02 -4.97775942e-01 3.57125849e-01 9.56710279e-02
4.78140980e-01 1.04966953e-01 2.45950371e-01 -2.58580506e-01
6.64182127e-01 3.60796899e-01 7.99408972e-01 7.46147156e-01
-2.29886547e-01 7.47034490e-01 -7.88264796e-02 -3.28268170e-01
-5.42673945e-01 -1.12660205e+00 -6.81430280e-01 4.82560396e-01
5.54492474e-01 -1.62868146e-02 -7.80690789e-01 -6.49051607e-01
1.75399899e-01 8.11738968e-01 -3.55726898e-01 2.93593019e-01
-3.38862240e-01 -7.54921794e-01 2.49273524e-01 2.76040405e-01
6.49104178e-01 -1.03346157e+00 -7.24549174e-01 7.26357475e-02
-3.00639421e-01 -1.33886206e+00 2.93771476e-02 1.99483737e-01
-9.68346655e-01 -1.49692655e+00 -4.13469732e-01 -5.01839817e-01
5.90336442e-01 5.21099687e-01 9.23169851e-01 1.33758768e-01
-4.66092527e-01 4.22727197e-01 -3.26600134e-01 -6.64590836e-01
-1.50308594e-01 1.38242915e-01 -1.27221316e-01 2.87145793e-01
5.12920618e-01 -4.12219673e-01 -5.46079338e-01 6.42569125e-01
-7.78889358e-01 -2.40232408e-01 4.19521034e-01 3.81759256e-01
1.24730444e+00 2.79876858e-01 2.37492770e-01 -9.92984116e-01
-8.04516748e-02 -6.03350043e-01 -1.09028232e+00 1.47313505e-01
-5.11090934e-01 -2.57999927e-01 2.79138952e-01 -1.53023079e-01
-9.21256185e-01 4.43414122e-01 -2.04753354e-01 -6.35534346e-01
-8.21817100e-01 2.89996833e-01 -7.79555663e-02 -2.42476031e-01
4.74042952e-01 2.64155474e-02 -1.05558448e-01 -5.93113244e-01
2.58276343e-01 7.97401428e-01 7.36906290e-01 -5.88057220e-01
9.86977458e-01 9.00096238e-01 1.05125360e-01 -1.15816188e+00
-8.94004226e-01 -1.13475037e+00 -9.86122608e-01 -3.24114919e-01
9.56071258e-01 -9.17704403e-01 -3.61072540e-01 5.01926124e-01
-8.78551185e-01 -1.81702763e-01 -4.74395931e-01 3.15172285e-01
-3.61223519e-01 5.34529448e-01 -3.93642150e-02 -1.03144228e+00
-2.29032785e-01 -1.21944237e+00 1.37500978e+00 4.63346243e-01
2.42002249e-01 -7.84511864e-01 -1.25063092e-01 5.43847084e-01
-9.80832726e-02 4.77907985e-01 4.15451616e-01 -8.12675416e-01
-9.91458893e-01 -1.38397828e-01 -2.47220650e-01 4.47401881e-01
2.12529916e-02 3.55814368e-01 -9.29482460e-01 -7.86933675e-02
8.80768821e-02 8.25150907e-02 9.33397174e-01 4.35833216e-01
1.18729544e+00 8.95333067e-02 -5.32296836e-01 5.53040087e-01
1.50526476e+00 2.54577398e-01 6.44936621e-01 3.59256208e-01
6.45350993e-01 6.81304514e-01 1.22993171e+00 1.47983417e-01
4.27019149e-01 6.38659179e-01 8.17196071e-01 -1.53138548e-01
8.40802193e-02 5.10146581e-02 -1.98849246e-01 2.19353139e-01
4.82340157e-03 -1.66158527e-01 -1.12932491e+00 7.75682330e-01
-2.03728104e+00 -8.60902369e-01 -8.03703964e-01 2.39126134e+00
3.90773773e-01 1.86172634e-01 3.67952511e-02 1.98726267e-01
9.68120933e-01 -4.40743230e-02 -6.08640432e-01 -3.31051350e-02
1.63390219e-01 4.31272507e-01 7.63130128e-01 5.03841043e-01
-1.32467031e+00 1.05532527e+00 4.64749241e+00 1.08099437e+00
-8.75607312e-01 1.63494602e-01 3.20981383e-01 2.19122157e-01
1.72365502e-01 -6.91000447e-02 -1.11138511e+00 5.26412368e-01
7.82731950e-01 2.46434569e-01 1.15001485e-01 8.71411920e-01
9.83492881e-02 -5.70720792e-01 -6.31456196e-01 7.84230232e-01
-9.03232917e-02 -1.08949506e+00 -1.56387880e-01 1.86304182e-01
5.99821270e-01 3.54508400e-01 -1.12032801e-01 1.49788991e-01
-1.58369523e-02 -5.64721167e-01 7.46416330e-01 3.23827654e-01
5.91663063e-01 -7.47262180e-01 7.08247364e-01 7.69195795e-01
-1.49970722e+00 -8.32443163e-02 -4.54724699e-01 2.44947881e-01
1.02612108e-01 7.23122716e-01 -8.70578647e-01 9.85862017e-01
9.12152767e-01 6.27588034e-01 -6.19570971e-01 1.45350432e+00
-3.65058690e-01 6.63667262e-01 -8.13037694e-01 4.66316551e-01
3.54351103e-01 -7.38359392e-01 7.10263968e-01 8.74782205e-01
3.05527002e-01 1.92865834e-01 4.36196506e-01 9.10145640e-01
2.54836798e-01 4.62453626e-02 -5.97440958e-01 3.62204581e-01
5.24459004e-01 1.47272587e+00 -1.18022883e+00 -3.71848643e-01
-9.17675793e-02 6.08976841e-01 9.63001475e-02 1.88696519e-01
-9.13404465e-01 -1.93841174e-01 5.69646895e-01 1.66609138e-01
6.33069038e-01 -2.45786920e-01 -3.06132704e-01 -9.21986759e-01
1.51248500e-01 -3.69142890e-01 4.68141973e-01 -6.25231862e-01
-1.19903910e+00 5.64264834e-01 3.43870252e-01 -1.44346440e+00
1.35844499e-02 -5.61006188e-01 -5.41666985e-01 6.90933764e-01
-1.74395001e+00 -1.25684214e+00 -6.19792104e-01 4.69529092e-01
6.41470790e-01 2.11493760e-01 5.37156999e-01 3.71805727e-01
-4.71816391e-01 -8.02486315e-02 2.59410113e-01 -1.38298526e-01
4.68753040e-01 -1.27865911e+00 3.01794678e-01 9.34979022e-01
3.18815440e-01 3.32253039e-01 6.12782598e-01 -7.82403111e-01
-5.67603886e-01 -1.46374178e+00 4.53157365e-01 -4.55533475e-01
5.50332904e-01 -3.01497966e-01 -1.13370371e+00 5.76731861e-01
-3.12687188e-01 2.27837250e-01 3.97980034e-01 -1.75054684e-01
-5.13307229e-02 -2.90538609e-01 -1.40259981e+00 2.20061243e-01
8.98066819e-01 -1.86364040e-01 -5.43993533e-01 4.27097201e-01
5.97708225e-01 -6.86055899e-01 -6.54739916e-01 8.04412603e-01
2.28935748e-01 -1.05165422e+00 1.14557135e+00 -2.35948712e-02
-1.69840250e-02 -8.26515555e-01 -1.31348059e-01 -1.01298356e+00
1.52847156e-01 -8.85764230e-03 -8.71476308e-02 1.19719446e+00
3.16148877e-01 -9.46150422e-01 7.09421039e-01 4.30612236e-01
-3.73710364e-01 -4.80336219e-01 -1.19404769e+00 -1.07797933e+00
-1.31080404e-01 -5.41035831e-01 5.17041087e-01 7.40555286e-01
-9.06912267e-01 8.13336372e-02 6.32392168e-02 7.57446945e-01
9.40826416e-01 5.56714833e-01 9.27554548e-01 -1.83924222e+00
2.55333155e-01 -3.95287666e-03 -4.85947043e-01 -8.00143123e-01
1.76258430e-01 -8.57735097e-01 1.81715339e-01 -1.61811185e+00
-5.27505241e-02 -9.17110562e-01 -3.64614539e-02 4.44354057e-01
-2.11152732e-01 5.05780458e-01 2.93013662e-01 4.28018987e-01
-5.96572399e-01 3.94652635e-01 8.98165643e-01 -1.59313068e-01
-2.87598550e-01 5.13631225e-01 -2.89981127e-01 7.50174880e-01
1.12219048e+00 -7.04058349e-01 -2.37249479e-01 -1.36164695e-01
-4.71822590e-01 -3.06513429e-01 7.02612638e-01 -1.45714259e+00
1.45512640e-01 -1.96855336e-01 2.78290473e-02 -1.29519105e+00
5.45522749e-01 -1.34704661e+00 4.39916760e-01 3.24813128e-01
3.98182750e-01 -4.25142795e-01 3.83127838e-01 8.52766335e-01
-2.29217038e-01 -5.50145984e-01 9.87283051e-01 -1.94178671e-01
-1.06276560e+00 4.87094223e-01 -1.07423472e-03 -7.95119256e-02
1.38401544e+00 -4.89516884e-01 -1.97028041e-01 1.43982485e-01
-7.41119921e-01 5.88744402e-01 5.67102015e-01 3.09476852e-01
4.40634996e-01 -7.96098113e-01 -4.58967596e-01 1.31722286e-01
3.01392943e-01 6.86504602e-01 3.71502370e-01 8.02524924e-01
-5.74402392e-01 3.12057495e-01 1.04399242e-01 -1.16299498e+00
-1.43574274e+00 4.10001546e-01 4.25399512e-01 7.15828827e-03
-6.34900868e-01 5.25141120e-01 2.04264820e-01 -8.69807839e-01
2.21815407e-01 -4.49950278e-01 -3.23994130e-01 2.17643142e-01
1.91703960e-01 4.21807706e-01 2.05352619e-01 -8.20951879e-01
-4.43912297e-01 8.26237738e-01 1.85994133e-01 2.58649051e-01
1.20266056e+00 -1.33942351e-01 -1.93083748e-01 6.50031269e-01
8.31702530e-01 -1.24764912e-01 -1.23470855e+00 -1.98683739e-01
1.04820937e-01 -5.35936058e-01 -1.33104071e-01 -6.41334534e-01
-1.20413005e+00 9.75118399e-01 8.94189954e-01 2.35935569e-01
8.47162127e-01 9.52993035e-02 7.54269421e-01 3.33675772e-01
6.02254808e-01 -1.09948289e+00 -4.49191958e-01 4.82730657e-01
6.56080365e-01 -1.50470781e+00 5.60471676e-02 -9.25536573e-01
-4.80387211e-01 1.02337611e+00 7.05740988e-01 -5.43807745e-02
5.98772347e-01 -5.56756705e-02 1.41326189e-01 -4.71157849e-01
-5.42058013e-02 -6.77997351e-01 3.28120500e-01 6.15175009e-01
-3.78172070e-01 2.48671249e-01 -8.71900320e-02 1.44056350e-01
-4.05368470e-02 -2.36253411e-01 2.78454214e-01 8.08976948e-01
-7.58823931e-01 -9.26926494e-01 -7.85066843e-01 5.41271865e-01
-9.91973430e-02 4.45061661e-02 -3.49070519e-01 9.55539405e-01
5.59656560e-01 9.41477954e-01 1.62356332e-01 -7.49808624e-02
4.60568905e-01 6.43846169e-02 8.01172182e-02 -7.61955917e-01
-3.38834405e-01 1.01291696e-02 -1.50172889e-01 -4.38743442e-01
-7.17906058e-01 -8.17808986e-01 -1.47083843e+00 6.89143464e-02
-6.81991518e-01 2.80618489e-01 8.26084137e-01 7.39734471e-01
2.82486022e-01 3.50587219e-01 5.09541512e-01 -1.12537992e+00
-2.55640328e-01 -6.27772391e-01 -6.14890933e-01 1.05809003e-01
1.78529173e-01 -9.98539209e-01 -5.42146802e-01 -1.42933577e-01] | [8.087016105651855, -2.7031478881835938] |
68caa69a-4873-4c9c-868b-0dd8e940934a | memory-constrained-policy-optimization-1 | 2204.09315 | null | https://arxiv.org/abs/2204.09315v2 | https://arxiv.org/pdf/2204.09315v2.pdf | Learning to Constrain Policy Optimization with Virtual Trust Region | We introduce a constrained optimization method for policy gradient reinforcement learning, which uses a virtual trust region to regulate each policy update. In addition to using the proximity of one single old policy as the normal trust region, we propose forming a second trust region through another virtual policy representing a wide range of past policies. We then enforce the new policy to stay closer to the virtual policy, which is beneficial if the old policy performs poorly. More importantly, we propose a mechanism to automatically build the virtual policy from a memory of past policies, providing a new capability for dynamically learning appropriate virtual trust regions during the optimization process. Our proposed method, dubbed Memory-Constrained Policy Optimization (MCPO), is examined in diverse environments, including robotic locomotion control, navigation with sparse rewards and Atari games, consistently demonstrating competitive performance against recent on-policy constrained policy gradient methods. | ['Svetha Venkatesh', 'Sunil Gupta', 'Kien Do', 'Dung Nguyen', 'Majid Abdolshah', 'Thommen Karimpanal George', 'Hung Le'] | 2022-04-20 | memory-constrained-policy-optimization | https://openreview.net/forum?id=7yuU9VeIpde | https://openreview.net/pdf?id=7yuU9VeIpde | null | ['policy-gradient-methods'] | ['methodology'] | [-4.08447623e-01 1.42525241e-03 -7.04285264e-01 -7.14038610e-02
-2.18348950e-01 -5.06807625e-01 6.00102365e-01 5.37903085e-02
-1.00643229e+00 1.42767346e+00 2.30531409e-01 -2.70649850e-01
-6.86834380e-02 -7.69878805e-01 -9.09921944e-01 -7.33694136e-01
-2.77628571e-01 2.77228624e-01 3.62111926e-01 -5.88937700e-01
5.50917387e-01 4.76480842e-01 -1.31439424e+00 -2.52141416e-01
9.78082538e-01 8.27283740e-01 5.54074645e-01 5.17114699e-01
1.33530125e-01 8.50466788e-01 -4.06318098e-01 1.92743897e-01
2.34594718e-01 -1.67549193e-01 -6.73007905e-01 -1.39723241e-01
-9.53808054e-02 -5.87707758e-01 -3.37609410e-01 9.18369353e-01
4.72485214e-01 7.71487057e-01 2.36951202e-01 -1.03381312e+00
-4.93017495e-01 6.46202624e-01 -1.48561209e-01 2.28118181e-01
4.80856597e-01 3.60050738e-01 8.09416413e-01 -5.19775331e-01
7.57992208e-01 1.30526292e+00 2.88286597e-01 6.87422276e-01
-9.60070789e-01 -5.40984333e-01 9.42741334e-01 1.68423682e-01
-7.25833595e-01 -1.84204847e-01 6.32045567e-01 -7.61255547e-02
9.21904624e-01 -1.12289719e-01 1.19461238e+00 1.22141302e+00
5.10059595e-01 7.42665827e-01 1.37280464e+00 -2.75098085e-01
9.15779710e-01 1.59956198e-02 -4.47652698e-01 9.40980852e-01
9.85153317e-02 7.96853364e-01 -4.80776668e-01 -6.10772967e-01
8.64046872e-01 -2.12100789e-01 -3.61450493e-01 -9.65791583e-01
-1.05583036e+00 8.09797049e-01 4.45417672e-01 -1.83027521e-01
-5.90717316e-01 5.03322184e-01 4.55250144e-01 5.70237160e-01
-4.19534184e-02 8.10562551e-01 -5.77399433e-01 -3.56580526e-01
-3.89784068e-01 7.13414609e-01 9.12143648e-01 5.59800625e-01
9.00545239e-01 5.72122097e-01 -3.30937564e-01 7.06084490e-01
4.05300677e-01 4.45790499e-01 6.94704413e-01 -1.35453045e+00
1.86015710e-01 3.35765928e-01 7.68551648e-01 -6.31951153e-01
-9.36544016e-02 -2.46449977e-01 -2.22069919e-01 9.60784197e-01
2.78847396e-01 -3.89536083e-01 -9.37048376e-01 1.92414045e+00
7.14592338e-01 1.52261183e-01 8.35795850e-02 8.79886687e-01
-1.87524438e-01 6.26004696e-01 -4.95416671e-02 -3.47766638e-01
5.63465953e-01 -9.32979167e-01 -5.75834155e-01 -2.38889560e-01
4.75936919e-01 -2.15485841e-01 1.16925931e+00 4.86661941e-01
-8.70028734e-01 -3.26049626e-01 -1.25812519e+00 5.49520910e-01
-3.65844399e-01 -3.03790450e-01 8.42395544e-01 2.63456076e-01
-1.18531322e+00 1.10654008e+00 -7.68788159e-01 -2.83478945e-01
9.55412015e-02 4.12662894e-01 -6.30737245e-02 1.90055802e-01
-1.33353102e+00 1.17618752e+00 6.50081635e-01 -1.00225925e-01
-1.51498985e+00 -3.79037112e-01 -7.70388961e-01 -2.87439764e-01
7.06177652e-01 -4.57998753e-01 1.33632231e+00 -7.99621224e-01
-2.39767289e+00 2.24819049e-01 1.52097076e-01 -5.51143885e-01
6.71014726e-01 -4.56336260e-01 -9.63653401e-02 -1.81013837e-01
1.44664273e-02 5.48237085e-01 1.27360356e+00 -1.48699975e+00
-6.41613603e-01 -1.66431531e-01 3.85632277e-01 6.88990235e-01
-3.61208618e-01 -4.12813216e-01 -1.32254660e-01 -5.22013724e-01
-2.82166004e-01 -1.14268088e+00 -8.03160906e-01 5.84868677e-02
1.08927637e-01 -1.12412319e-01 7.65072644e-01 -1.61357820e-01
1.22569013e+00 -1.82935739e+00 5.03232181e-01 4.93873149e-01
-8.01272690e-02 3.09823990e-01 -7.46965632e-02 3.32988620e-01
3.41926932e-01 -2.87655860e-01 -4.10921723e-02 1.74059365e-02
9.81396511e-02 8.16506147e-01 -5.38051665e-01 3.62567782e-01
-3.50855529e-01 7.39846110e-01 -1.54131794e+00 -2.30074063e-01
5.89862093e-02 -1.79401204e-01 -7.90494859e-01 3.93645704e-01
-5.68546891e-01 5.60411394e-01 -8.48151445e-01 5.47986209e-01
2.56099671e-01 6.84505105e-02 4.67614114e-01 4.76077855e-01
-4.50642526e-01 -3.86563572e-03 -1.22690582e+00 1.64374840e+00
-4.15019900e-01 -1.17979974e-01 3.63981366e-01 -8.58206034e-01
1.02165866e+00 2.08665177e-01 7.05951691e-01 -6.50169194e-01
5.54963611e-02 1.56239927e-01 -3.09282243e-01 -1.50033891e-01
6.81777716e-01 1.87670767e-01 -7.69430995e-02 5.87663889e-01
1.39454901e-01 -3.26948375e-01 2.03514904e-01 1.93538383e-01
9.80046272e-01 8.10507834e-01 4.10528272e-01 -3.71154130e-01
4.34611470e-01 -5.46785966e-02 9.06992733e-01 1.19685590e+00
-6.76689625e-01 -1.75603628e-01 4.17428702e-01 -4.12604541e-01
-9.36073124e-01 -1.08043635e+00 2.45433405e-01 1.47736382e+00
2.40059838e-01 -3.90625805e-01 -3.95162761e-01 -9.66616511e-01
5.36734521e-01 5.00533760e-01 -8.13358903e-01 -3.59602571e-01
-8.65943015e-01 -4.50628847e-01 2.73446470e-01 4.05090153e-01
6.39601946e-01 -1.19736314e+00 -9.12564993e-01 5.84063530e-01
3.01935613e-01 -6.54710472e-01 -6.91727996e-01 4.87625033e-01
-8.35099936e-01 -1.02835274e+00 -6.17535830e-01 -5.47529340e-01
4.41791981e-01 -8.32240060e-02 8.53447556e-01 9.44768190e-02
4.00371552e-01 5.90483367e-01 -3.76200527e-01 -2.43746623e-01
-4.49814022e-01 -2.49490157e-01 6.16471291e-01 -4.05986518e-01
-4.51205045e-01 -5.99314272e-01 -5.19106746e-01 3.06841731e-01
-4.48656976e-01 -3.81759644e-01 3.87605578e-01 1.29887140e+00
6.63553834e-01 -4.83126611e-01 5.34666657e-01 -4.33001935e-01
1.08866894e+00 -5.32485306e-01 -1.04932606e+00 2.36282840e-01
-1.01888049e+00 5.63465297e-01 1.05044711e+00 -8.87896001e-01
-1.18855906e+00 2.42167842e-02 1.72645509e-01 -4.21124041e-01
2.49793202e-01 3.03559184e-01 2.21547991e-01 -3.74603808e-01
7.40405798e-01 1.91785961e-01 8.89023989e-02 -1.81840703e-01
6.50223553e-01 8.89847428e-02 4.01866823e-01 -1.25363004e+00
5.38823307e-01 4.26622868e-01 -1.11579120e-01 -2.51410544e-01
-5.57064056e-01 -5.44100180e-02 -4.05056357e-01 -4.69876140e-01
4.83715862e-01 -5.50100684e-01 -9.37325478e-01 4.65702504e-01
-6.86354578e-01 -9.00824189e-01 -4.86056834e-01 6.31195128e-01
-9.50300097e-01 4.47514951e-01 -5.73820770e-01 -6.32342160e-01
-2.37777159e-01 -1.18945539e+00 2.96315789e-01 3.51001412e-01
-1.06500082e-01 -9.89478886e-01 5.80319107e-01 -4.81906444e-01
3.34127426e-01 3.02421421e-01 7.31183052e-01 -1.45530790e-01
-3.18608284e-01 3.62386823e-01 3.44756871e-01 2.31765881e-01
1.64065063e-01 -1.50153980e-01 -4.26675051e-01 -8.81773472e-01
-1.67888865e-01 -7.41162896e-01 6.54134154e-01 2.53825843e-01
7.80846775e-01 -4.04949665e-01 -4.02780622e-01 6.55678213e-01
1.28448832e+00 5.36656499e-01 2.51489073e-01 7.27945268e-01
3.49293858e-01 -4.94934469e-02 1.06682611e+00 9.15961146e-01
2.01434910e-01 5.46333551e-01 6.36351407e-01 2.25456193e-01
3.96114528e-01 -5.28435349e-01 7.64766753e-01 4.15233612e-01
-1.53067365e-01 2.54808873e-01 -3.97251040e-01 3.23175400e-01
-2.31055832e+00 -8.35205615e-01 8.14616024e-01 2.33596015e+00
9.77726161e-01 3.60666573e-01 2.30829954e-01 -4.03994352e-01
3.72203559e-01 2.91874379e-01 -1.21549106e+00 -8.13400269e-01
1.66046292e-01 1.57816976e-01 6.20238900e-01 6.64841533e-01
-8.92783582e-01 1.38177371e+00 7.64490652e+00 5.15430391e-01
-1.28274131e+00 -2.03137040e-01 1.42290384e-01 -1.48079455e-01
-1.46810770e-01 2.30230242e-01 -6.92101240e-01 5.26891768e-01
5.32924592e-01 -2.97839016e-01 1.01671243e+00 1.19223595e+00
4.79253680e-01 -4.87581283e-01 -8.26668739e-01 4.49830174e-01
-4.70152617e-01 -1.17578900e+00 5.12067676e-02 4.56003919e-02
9.08573389e-01 1.81581955e-02 1.70099795e-01 6.30921662e-01
1.22798371e+00 -6.63782895e-01 8.22059095e-01 5.28789282e-01
3.52472156e-01 -9.82793033e-01 1.09189652e-01 3.11006814e-01
-1.10875237e+00 -6.45232081e-01 -4.70732570e-01 -2.73107648e-01
-6.29774332e-02 9.22107771e-02 -7.19017625e-01 3.66562009e-01
6.82231009e-01 5.40460110e-01 -1.12306394e-01 8.82486463e-01
-4.92765397e-01 3.84244353e-01 -1.61410183e-01 -4.93353367e-01
6.35695815e-01 -2.99458265e-01 8.23976040e-01 6.27279401e-01
4.01787274e-02 -7.26121217e-02 7.53669202e-01 5.77477694e-01
2.18760759e-01 6.96605220e-02 -8.59210253e-01 2.42868692e-01
6.24139309e-01 9.03628170e-01 -4.43024963e-01 -4.12777960e-01
6.01356030e-02 1.04208040e+00 8.95238638e-01 6.04240596e-01
-8.32358539e-01 -1.60802633e-01 1.14660311e+00 -5.28026760e-01
4.55500841e-01 -5.33974528e-01 2.45679080e-01 -1.11362135e+00
-1.09288618e-01 -7.81234741e-01 2.85502851e-01 -3.97201806e-01
-7.93759108e-01 2.67910242e-01 1.63123623e-01 -1.14530075e+00
-5.47223210e-01 -4.70116019e-01 -5.61522782e-01 4.97990847e-01
-1.56714559e+00 -5.12727976e-01 1.73624843e-01 6.75563872e-01
2.43607625e-01 -3.13165098e-01 8.10987890e-01 -3.09998900e-01
-3.30723912e-01 3.55158150e-01 5.47886550e-01 -3.56031656e-01
8.51289451e-01 -1.41772926e+00 6.20635152e-02 5.89856684e-01
-3.88438433e-01 6.69336379e-01 8.08198094e-01 -8.29934597e-01
-1.44640124e+00 -8.02166760e-01 -1.56166092e-01 5.53384721e-02
7.28451014e-01 -7.36574158e-02 -8.37563813e-01 4.45055455e-01
7.38634318e-02 -6.76578954e-02 1.28791735e-01 -6.56715706e-02
-1.26444390e-02 -9.47612077e-02 -1.10826647e+00 9.42587376e-01
1.02366436e+00 -1.87591508e-01 -3.58791351e-01 8.63229707e-02
6.46935821e-01 -6.73201978e-01 -7.45617330e-01 3.21809739e-01
7.32837439e-01 -7.09681392e-01 8.72135520e-01 -8.16165030e-01
-1.71496123e-01 -3.94640177e-01 -6.13663681e-02 -2.06328225e+00
-5.00028670e-01 -8.81118238e-01 -5.34450591e-01 5.55491149e-01
2.95109730e-02 -8.39366138e-01 9.88787651e-01 2.84067184e-01
-1.29839212e-01 -1.17553246e+00 -1.00135183e+00 -9.98683512e-01
2.14390159e-01 -3.75918113e-02 6.51246428e-01 8.58573735e-01
2.66624123e-01 -2.49045476e-01 -7.18536317e-01 5.50845824e-02
4.58730340e-01 -8.52468237e-02 7.11939335e-01 -7.11278498e-01
-5.67255080e-01 -3.21738482e-01 4.12173457e-02 -1.22179246e+00
4.01434243e-01 -6.18529975e-01 2.03919873e-01 -1.35904598e+00
-3.98848474e-01 -5.44429183e-01 -7.55599380e-01 4.69937891e-01
-2.77074277e-01 -5.65028548e-01 1.56244516e-01 2.45896932e-02
-6.70968771e-01 1.15599322e+00 1.77220201e+00 6.10209769e-03
-6.73701465e-01 6.54826984e-02 -4.19277638e-01 6.54964507e-01
1.04034019e+00 -4.05338496e-01 -6.18490696e-01 -1.52606249e-01
-1.12515418e-02 1.48723558e-01 9.03466996e-03 -1.03565192e+00
1.53188899e-01 -8.19973767e-01 3.81634146e-01 -9.01575983e-02
4.31077957e-01 -4.14573759e-01 -4.57097709e-01 8.20557714e-01
-3.41302663e-01 5.17994821e-01 1.32439077e-01 9.82246578e-01
1.67422935e-01 -1.37771666e-01 7.99918413e-01 -5.57482541e-01
-1.27774990e+00 3.05583090e-01 -8.35088313e-01 3.46282534e-02
1.06432855e+00 -8.48519728e-02 9.30974334e-02 -4.54499841e-01
-8.89327407e-01 8.89869332e-01 7.07438409e-01 3.16653550e-01
9.54373538e-01 -1.27620220e+00 -1.69277787e-01 -3.46672051e-02
-1.02170721e-01 -3.65624040e-01 -2.78495580e-01 2.15101153e-01
-1.11778356e-01 7.17305616e-02 -6.32651567e-01 -2.31786191e-01
-8.60633075e-01 6.16498530e-01 5.86152613e-01 -6.35658741e-01
-4.47758436e-01 6.49462223e-01 -4.29578394e-01 -8.57028902e-01
3.12192321e-01 -4.65973377e-01 -2.59788781e-01 -3.09355617e-01
2.54601240e-01 4.72273350e-01 -2.68221974e-01 -1.58429947e-02
-1.30248472e-01 2.51754552e-01 -3.17520276e-02 -5.62099218e-01
1.08009672e+00 1.70414674e-03 7.96534959e-03 4.05553252e-01
5.54004192e-01 -1.70875311e-01 -1.84017694e+00 -2.31609672e-01
1.01679273e-01 -5.78104675e-01 -1.11919716e-01 -7.89804518e-01
-6.40313625e-01 7.92408958e-02 6.36198342e-01 -3.64323825e-01
5.85365057e-01 -6.49148703e-01 4.24107671e-01 1.04065967e+00
8.36183250e-01 -1.82572126e+00 5.09873927e-01 1.10515261e+00
7.89342046e-01 -1.16228473e+00 2.18259454e-01 3.80515605e-01
-8.84828866e-01 1.06207693e+00 1.04687512e+00 -4.35267508e-01
5.94554126e-01 2.07361594e-01 4.59408052e-02 1.67459980e-01
-9.52276289e-01 -2.11157382e-01 -2.85327882e-01 6.78984106e-01
-1.43188998e-01 8.80621672e-02 -3.09323847e-01 -1.35809004e-01
3.78142036e-02 1.71279252e-01 4.28303480e-01 1.51603019e+00
-8.50113511e-01 -1.37554717e+00 -3.96524340e-01 2.40128160e-01
-4.74707820e-02 2.70946711e-01 -9.97740701e-02 7.22292721e-01
-1.82920396e-01 8.02702427e-01 -2.85160094e-01 -3.20445567e-01
2.66019762e-01 -1.78888813e-01 6.26535356e-01 -4.09470707e-01
-5.37556589e-01 -2.71059513e-01 4.08240482e-02 -9.44106162e-01
8.06973726e-02 -4.15805846e-01 -1.52754533e+00 -2.89014056e-02
-1.17080398e-01 4.60606456e-01 4.50008214e-01 9.16855931e-01
1.36913717e-01 2.59853631e-01 8.65671217e-01 -9.85268235e-01
-1.12134981e+00 -5.78208208e-01 -4.43363011e-01 2.12219611e-01
4.95071679e-01 -1.22251749e+00 1.44271795e-02 -5.92406034e-01] | [4.102975368499756, 2.109081983566284] |
cb538d90-56be-457c-a57a-fa413cfc7e19 | polar-transformer-networks | 1709.01889 | null | http://arxiv.org/abs/1709.01889v3 | http://arxiv.org/pdf/1709.01889v3.pdf | Polar Transformer Networks | Convolutional neural networks (CNNs) are inherently equivariant to
translation. Efforts to embed other forms of equivariance have concentrated
solely on rotation. We expand the notion of equivariance in CNNs through the
Polar Transformer Network (PTN). PTN combines ideas from the Spatial
Transformer Network (STN) and canonical coordinate representations. The result
is a network invariant to translation and equivariant to both rotation and
scale. PTN is trained end-to-end and composed of three distinct stages: a polar
origin predictor, the newly introduced polar transformer module and a
classifier. PTN achieves state-of-the-art on rotated MNIST and the newly
introduced SIM2MNIST dataset, an MNIST variation obtained by adding clutter and
perturbing digits with translation, rotation and scaling. The ideas of PTN are
extensible to 3D which we demonstrate through the Cylindrical Transformer
Network. | ['Christine Allen-Blanchette', 'Xiaowei Zhou', 'Kostas Daniilidis', 'Carlos Esteves'] | 2017-09-06 | polar-transformer-networks-1 | https://openreview.net/forum?id=HktRlUlAZ | https://openreview.net/pdf?id=HktRlUlAZ | iclr-2018-1 | ['rotated-mnist'] | ['computer-vision'] | [ 2.39836127e-02 1.97950497e-01 8.72689039e-02 -3.10151398e-01
-3.19053888e-01 -9.94093597e-01 1.19207060e+00 -6.89461946e-01
-4.64386493e-01 3.38155240e-01 4.07984942e-01 -4.72827047e-01
1.26313224e-01 -5.50413251e-01 -1.08196759e+00 -5.42963147e-01
-5.08739613e-02 4.20331150e-01 -4.54074852e-02 -7.13217974e-01
-1.99467406e-01 9.62921619e-01 -7.93853164e-01 2.56430954e-01
-1.54087871e-01 8.03837121e-01 -3.94159943e-01 7.98366487e-01
4.64001089e-01 7.39418089e-01 -1.87632129e-01 -4.45063174e-01
7.90613592e-01 -1.18063450e-01 -7.06723988e-01 -3.36978287e-01
9.31939662e-01 -5.73420338e-02 -8.59143615e-01 9.31350946e-01
6.79453433e-01 -1.49642827e-03 5.15368104e-01 -1.01106822e+00
-9.75359857e-01 7.66109169e-01 -3.37944925e-01 1.55521229e-01
9.27112345e-03 -9.99384522e-02 1.06905353e+00 -1.18417621e+00
9.56545651e-01 1.60521734e+00 1.26380777e+00 4.96672869e-01
-1.33883882e+00 -3.84701759e-01 6.10866919e-02 3.26424278e-03
-9.74972963e-01 -4.65749949e-01 8.01267326e-01 -2.75484443e-01
1.16687894e+00 2.57420421e-01 6.35242701e-01 1.45007932e+00
5.14284730e-01 4.38853264e-01 1.00486624e+00 -4.12875354e-01
3.58944088e-02 -3.18123937e-01 -3.64616871e-01 5.58296859e-01
-6.71616644e-02 2.36089930e-01 -3.07932734e-01 2.72156328e-01
1.34280264e+00 3.28288302e-02 3.03403642e-02 -9.44518566e-01
-1.66558230e+00 6.40917778e-01 9.05512094e-01 -1.18574448e-01
-5.55990860e-02 5.65497100e-01 4.79254186e-01 5.88786542e-01
2.17705488e-01 7.15651691e-01 -8.28603446e-01 4.78547402e-02
-6.28732264e-01 3.86454374e-01 5.68085670e-01 1.24382293e+00
3.34582031e-01 5.15527070e-01 2.85151541e-01 5.62874138e-01
3.50349635e-01 5.22832274e-01 7.25060403e-01 -9.06583846e-01
4.87839788e-01 2.41093203e-01 -3.58694166e-01 -8.12301576e-01
-7.73830414e-01 -7.99367487e-01 -8.99262726e-01 4.95506257e-01
3.79330039e-01 -1.21798679e-01 -1.09108782e+00 2.03763938e+00
1.27640322e-01 5.01876473e-02 2.03024000e-01 8.54696214e-01
7.78104603e-01 8.94755721e-02 -3.56439739e-01 5.15879750e-01
1.45928395e+00 -8.59734118e-01 -3.69271636e-01 -2.50362270e-02
4.41846311e-01 -1.00255167e+00 8.71503353e-01 1.56069800e-01
-1.31220424e+00 -4.45235580e-01 -1.49083793e+00 -5.51034987e-01
-7.23100722e-01 1.53707415e-01 3.43667626e-01 4.88255382e-01
-1.53627551e+00 7.29499757e-01 -1.17796576e+00 -3.99369895e-01
2.72949547e-01 5.78619123e-01 -8.21659684e-01 2.65404224e-01
-1.22057104e+00 1.04952383e+00 5.18072285e-02 9.02864039e-02
-8.39202881e-01 -9.08422887e-01 -1.21063507e+00 -2.01416939e-01
-4.48940963e-01 -1.03605199e+00 1.54733837e+00 -1.06831360e+00
-1.70825493e+00 7.97188342e-01 1.54829733e-02 -7.49414444e-01
1.08188403e+00 -2.19011500e-01 -3.08459461e-01 -4.64245379e-02
2.35903431e-02 8.20610583e-01 9.23975646e-01 -8.95869315e-01
-1.88966438e-01 -4.38573807e-01 1.93036661e-01 2.24891037e-01
-3.30731622e-03 2.48053856e-03 -1.60937041e-01 -1.16970146e+00
7.81884730e-01 -1.38570726e+00 -1.92321435e-01 8.42764080e-02
-5.29428542e-01 1.49357200e-01 1.15507662e+00 -5.56941092e-01
1.65870994e-01 -2.00071526e+00 3.69492173e-01 1.73877895e-01
3.27315032e-01 2.37344727e-02 -3.24931264e-01 1.46028027e-01
-1.11082745e+00 -5.93693880e-03 1.76664904e-01 -4.66910422e-01
2.81313121e-01 1.98396012e-01 -5.48840880e-01 9.91852462e-01
5.18534005e-01 8.95764530e-01 -7.97822714e-01 3.97568233e-02
2.55315214e-01 9.45514798e-01 -8.20978999e-01 -3.13598037e-01
7.72795603e-02 2.25104526e-01 1.99082091e-01 6.66594088e-01
7.75177360e-01 6.72063008e-02 2.66636536e-02 -5.21717489e-01
-2.46734068e-01 5.38217127e-01 -1.20076108e+00 1.77353525e+00
-3.45396101e-01 9.70917165e-01 -1.35336623e-01 -9.11402225e-01
9.14071262e-01 3.92906070e-01 4.36701447e-01 -6.21269107e-01
3.44800264e-01 9.78695974e-02 2.85423920e-03 1.11909278e-01
6.95622921e-01 -8.50980878e-02 -5.61345257e-02 9.82209817e-02
5.64152896e-01 1.23059921e-01 -1.02616496e-01 1.81746915e-01
1.09682965e+00 6.85585916e-01 2.05501154e-01 -5.88418245e-01
1.78106800e-01 -2.52045602e-01 3.76844913e-01 4.17009801e-01
-4.39370237e-02 1.00105238e+00 6.70570433e-01 -9.73924994e-01
-1.58915651e+00 -1.56141555e+00 -3.57431978e-01 1.02350271e+00
-2.09149480e-01 -3.47230554e-01 -6.38417900e-01 -5.20806313e-01
-8.94353688e-02 2.15037435e-01 -9.84167457e-01 -7.07740560e-02
-9.15619731e-01 -5.43281555e-01 9.96802211e-01 8.48870277e-01
7.68320262e-01 -5.68293452e-01 -4.88530189e-01 -4.92613241e-02
-5.02980035e-03 -1.35449290e+00 -4.19186831e-01 5.63311219e-01
-8.77364337e-01 -8.27173412e-01 -7.83515275e-01 -6.38031185e-01
6.42391324e-01 6.01426810e-02 1.11303544e+00 -9.99500275e-01
-7.02630505e-02 3.13544899e-01 -2.43098348e-01 -5.57799101e-01
-2.37795040e-01 2.18411356e-01 5.30972302e-01 -2.45839998e-01
-8.09470043e-02 -1.17868209e+00 -5.18364727e-01 5.14986038e-01
-7.91690707e-01 1.02338023e-01 5.72522938e-01 6.79624736e-01
6.26645267e-01 -6.68130457e-01 -1.74420048e-02 -6.22981548e-01
8.97418410e-02 -1.86357692e-01 -5.66282988e-01 -3.26657742e-01
-1.94974020e-01 3.67525607e-01 8.67643595e-01 -5.53598642e-01
-6.58417225e-01 3.30564767e-01 -1.87546328e-01 -7.80392647e-01
8.97033364e-02 1.43616751e-01 -2.11431384e-01 -2.80959606e-01
1.04709113e+00 -1.29069611e-01 1.87710486e-02 -4.53962892e-01
8.09704781e-01 -1.45935968e-01 1.10693729e+00 -4.78704154e-01
1.35396433e+00 7.78509080e-01 6.04830205e-01 -6.55302465e-01
-3.28197747e-01 6.88811578e-03 -1.12610018e+00 7.24603087e-02
7.55005717e-01 -1.20512462e+00 -4.87162769e-01 3.41162205e-01
-1.21318591e+00 -2.32823119e-01 -6.64923489e-01 5.84727824e-01
-7.38826156e-01 4.95582186e-02 -5.02892852e-01 1.07516982e-01
-1.71248436e-01 -1.26505053e+00 1.00517511e+00 -9.00942534e-02
-2.85272092e-01 -1.07591295e+00 1.68140173e-01 -4.00483787e-01
6.89820707e-01 6.76845908e-01 6.46019578e-01 -7.10580647e-01
-3.42390269e-01 -3.13000500e-01 -2.07561687e-01 3.63774180e-01
-1.49515122e-01 -3.29003334e-02 -1.08269870e+00 -3.79487097e-01
-8.57368261e-02 -1.23304494e-01 8.61742318e-01 3.69560331e-01
6.01898968e-01 -5.34498036e-01 2.24031016e-01 1.53257740e+00
1.32043588e+00 -1.71993107e-01 7.33815253e-01 8.79997015e-01
1.00724113e+00 1.11537009e-01 -1.83215931e-01 3.27031352e-02
2.74548143e-01 8.63157690e-01 7.99265802e-01 -3.77127171e-01
-4.30893630e-01 -2.35988557e-01 7.78849721e-01 6.62642241e-01
-5.78473866e-01 2.91749150e-01 -7.77547181e-01 3.56147796e-01
-1.64196444e+00 -9.74080145e-01 1.00211821e-01 1.69513106e+00
5.81069231e-01 5.72214611e-02 7.30103522e-04 2.33378038e-01
2.78070152e-01 4.08513486e-01 -5.07643819e-01 -7.08166957e-01
-5.32344699e-01 5.24360418e-01 1.03753161e+00 3.90727758e-01
-1.45524204e+00 1.10117888e+00 6.54936361e+00 1.72072753e-01
-1.59538782e+00 -2.31002998e-02 2.43414551e-01 -9.25427228e-02
-3.44472602e-02 -1.41974777e-01 -6.89230144e-01 -3.07501107e-01
6.73438370e-01 1.56100273e-01 3.97046894e-01 1.19127452e+00
1.09885260e-01 1.01573884e+00 -1.41759348e+00 7.55738258e-01
1.57304913e-01 -1.37129140e+00 2.32179046e-01 -8.34907219e-02
7.48953044e-01 7.67323017e-01 6.62348688e-01 2.19007716e-01
6.58804119e-01 -1.09568155e+00 1.25221634e+00 2.17932448e-01
1.19722641e+00 -6.82411849e-01 5.43933034e-01 -3.84517938e-01
-1.17605436e+00 2.53908873e-01 -5.53819716e-01 -1.22923148e-03
-2.13773549e-01 7.41947368e-02 -1.05367970e+00 5.21925867e-01
8.24060440e-01 1.10822046e+00 -6.53096437e-01 4.08221304e-01
-1.31602943e-01 1.13615341e-01 -4.31517810e-01 4.85886753e-01
5.77237904e-01 -1.39935315e-01 8.81883383e-01 1.43203163e+00
4.39437330e-01 -3.36364120e-01 -3.88859034e-01 6.87760293e-01
-3.26989740e-01 -3.11896265e-01 -7.51286030e-01 4.52048957e-01
9.47393775e-02 1.38680935e+00 -6.47990406e-01 1.00843348e-02
-2.84542829e-01 1.00946832e+00 1.29293889e-01 5.38306236e-01
-8.65463197e-01 -3.97644222e-01 1.04362953e+00 1.30234793e-01
6.14386380e-01 -3.79858613e-01 -3.35088521e-01 -1.37248075e+00
6.72807321e-02 -1.09240282e+00 9.69068408e-02 -8.86953115e-01
-8.47044408e-01 7.37379909e-01 -5.39015569e-02 -1.50642252e+00
-2.20707431e-01 -1.44619906e+00 -5.37453473e-01 9.23798621e-01
-1.38290048e+00 -1.54719293e+00 -2.40609124e-01 9.26823378e-01
2.33665049e-01 -4.90786761e-01 9.39869165e-01 1.67362586e-01
-1.78371027e-01 9.38676000e-01 1.49406523e-01 7.35592842e-01
1.00027955e+00 -1.60157681e+00 1.01568520e+00 1.02504408e+00
2.53398657e-01 9.72245514e-01 8.45666230e-01 1.62346605e-02
-1.67504549e+00 -1.55049157e+00 7.41618991e-01 -7.91873991e-01
7.52317309e-01 -4.81882662e-01 -1.30383030e-01 1.48035264e+00
2.16380090e-01 5.10104358e-01 2.01260269e-01 -1.08228914e-01
-1.26309741e+00 -2.14370787e-01 -9.38480616e-01 9.80259657e-01
1.14820921e+00 -7.01612651e-01 -6.36547208e-01 2.74952829e-01
1.00358772e+00 -1.13402092e+00 -8.49410415e-01 4.38585132e-01
9.52597618e-01 -7.92433262e-01 1.41746783e+00 -9.51056063e-01
5.89740217e-01 -3.95485014e-01 -5.52105486e-01 -1.56154978e+00
-5.43016493e-01 -9.20968235e-01 2.14329541e-01 6.15759313e-01
6.14202738e-01 -7.54325151e-01 5.17179549e-01 6.19921349e-02
-5.82181990e-01 -4.67942476e-01 -1.25369012e+00 -8.66755009e-01
4.43519592e-01 -4.69237477e-01 5.84701180e-01 1.36670482e+00
-1.14086509e-01 3.75800908e-01 -3.35293710e-01 2.61875480e-01
4.19759929e-01 -6.59507990e-01 8.98724437e-01 -9.86760914e-01
-1.96262375e-01 -5.49900889e-01 -1.03487945e+00 -1.02587795e+00
3.46315503e-02 -1.15726674e+00 -3.46082538e-01 -9.48944151e-01
-2.29360953e-01 5.12014516e-02 -3.58109623e-01 4.91247356e-01
5.51141381e-01 6.34047687e-01 2.56031603e-01 7.75066018e-02
-3.26682955e-01 3.03167194e-01 1.18477690e+00 -1.01521507e-01
4.73660417e-02 -2.13078111e-01 -6.48089170e-01 1.04145062e+00
8.29091609e-01 -2.50917137e-01 -2.52548844e-01 -6.54760361e-01
5.64920545e-01 -4.91421223e-01 6.36507332e-01 -1.07402158e+00
8.97964165e-02 3.78849059e-01 8.78595531e-01 -4.66910511e-01
3.76149923e-01 -7.93392777e-01 6.68388829e-02 4.13078278e-01
-3.66539836e-01 1.16151214e+00 3.17671031e-01 3.12362343e-01
-6.48499355e-02 5.67704737e-01 1.01820254e+00 -3.26352678e-02
-2.46830434e-01 5.20772874e-01 4.09204140e-03 -3.46289091e-02
5.44381678e-01 -8.21540505e-02 -5.19900501e-01 -2.70423621e-01
-9.33028936e-01 -3.45567733e-01 5.84831238e-01 5.42175114e-01
4.45324302e-01 -1.70813692e+00 -6.78770304e-01 3.36211890e-01
-1.45700365e-01 2.33832762e-01 -3.02286357e-01 8.66484106e-01
-1.12099743e+00 5.20024955e-01 -5.53494632e-01 -7.15268135e-01
-1.20338643e+00 3.46267879e-01 7.55307198e-01 -2.99984813e-01
-7.87177265e-01 7.89364159e-01 1.13506146e-01 -9.19767439e-01
8.32809508e-02 -7.31754184e-01 -1.51499119e-02 -2.01564223e-01
2.34780163e-01 2.06642032e-01 3.38004321e-01 -9.20716107e-01
-3.90038460e-01 6.62691414e-01 2.62093991e-02 -5.34608960e-01
1.49943531e+00 3.45142186e-01 -2.27019593e-01 1.94288060e-01
1.48858345e+00 3.31917070e-02 -1.29167223e+00 -4.53144640e-01
-3.77913177e-01 1.45140797e-01 -2.45855004e-01 -4.56644386e-01
-1.02746761e+00 1.01461029e+00 5.45297444e-01 -1.58937156e-01
7.31209338e-01 -2.36121923e-01 2.71313816e-01 8.18335950e-01
-2.32704759e-01 -7.40535319e-01 6.30412623e-02 1.21565199e+00
1.19540620e+00 -6.36081576e-01 -1.38290077e-02 2.10508525e-01
-2.35026047e-01 1.46992671e+00 2.97067434e-01 -6.85176551e-01
9.77846444e-01 5.08148193e-01 3.99934053e-01 8.86259787e-03
-4.43273306e-01 3.74057233e-01 3.35746586e-01 6.70546234e-01
4.99284804e-01 1.49418227e-02 1.75761357e-01 2.19729468e-01
-1.09302056e+00 -5.76456964e-01 4.46297586e-01 6.73384309e-01
1.80805326e-01 -9.63799596e-01 -4.48295981e-01 -2.27301955e-01
-6.45406663e-01 -1.69629455e-01 -5.54198742e-01 1.27253342e+00
7.11110532e-02 5.46906702e-02 2.66927063e-01 -5.52307010e-01
6.29330993e-01 6.11437205e-03 4.82590318e-01 -1.10211469e-01
-4.25995111e-01 1.72035247e-02 -1.22415721e-01 -8.47336471e-01
-4.19028193e-01 -8.77467036e-01 -8.73704612e-01 -4.19791788e-01
2.95117468e-01 -4.35074598e-01 9.62508559e-01 7.36382067e-01
2.29602411e-01 9.15808916e-01 5.37878633e-01 -1.25872791e+00
-7.36172259e-01 -1.12736917e+00 -2.35173047e-01 2.62963980e-01
7.47996628e-01 -3.35228354e-01 -2.73909837e-01 2.98384905e-01] | [8.918286323547363, 2.3460941314697266] |
faf54ff8-fd88-4a4e-8a81-4b4e18a205c4 | contrastive-attention-networks-for | 2306.07998 | null | https://arxiv.org/abs/2306.07998v1 | https://arxiv.org/pdf/2306.07998v1.pdf | Contrastive Attention Networks for Attribution of Early Modern Print | In this paper, we develop machine learning techniques to identify unknown printers in early modern (c.~1500--1800) English printed books. Specifically, we focus on matching uniquely damaged character type-imprints in anonymously printed books to works with known printers in order to provide evidence of their origins. Until now, this work has been limited to manual investigations by analytical bibliographers. We present a Contrastive Attention-based Metric Learning approach to identify similar damage across character image pairs, which is sensitive to very subtle differences in glyph shapes, yet robust to various confounding sources of noise associated with digitized historical books. To overcome the scarce amount of supervised data, we design a random data synthesis procedure that aims to simulate bends, fractures, and inking variations induced by the early printing process. Our method successfully improves downstream damaged type-imprint matching among printed works from this period, as validated by in-domain human experts. The results of our approach on two important philosophical works from the Early Modern period demonstrate potential to extend the extant historical research about the origins and content of these books. | ['Taylor Berg-Kirkpatrick', "Max G'Sell", 'Christopher N. Warren', 'Samuel V. Lemley', 'Elizaveta Pertseva', 'Kishore PV Reddy', 'Kartik Goyal', 'Nikolai Vogler'] | 2023-06-12 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 2.41305858e-01 -1.78811595e-01 2.37905189e-01 -1.13447897e-01
-1.04985011e+00 -9.10438776e-01 1.02810311e+00 3.09092969e-01
-2.96790838e-01 5.25186241e-01 3.26816380e-01 4.87455055e-02
-2.99396813e-01 -7.64652133e-01 -9.07080829e-01 -1.71426550e-01
2.87176192e-01 7.57471204e-01 5.45678586e-02 -1.96497262e-01
1.25633776e+00 6.60901666e-01 -1.33034348e+00 4.09632653e-01
7.08683729e-01 4.65933651e-01 3.87972564e-01 7.34610677e-01
-4.80140716e-01 2.86127120e-01 -1.08949900e+00 -1.07481670e+00
3.05028558e-01 -2.59024203e-01 -6.40088439e-01 4.52922843e-02
1.02676654e+00 -2.50584841e-01 -2.67329216e-01 1.00767314e+00
6.45822585e-01 -3.92669737e-01 1.12855101e+00 -6.45933747e-01
-1.33718967e+00 8.40972900e-01 -5.49806476e-01 1.63278937e-01
5.49948573e-01 -4.95484546e-02 1.02313912e+00 -1.00470674e+00
9.69160140e-01 1.19677091e+00 9.66469109e-01 3.01407695e-01
-1.26500475e+00 -4.29026991e-01 -3.18069726e-01 2.53250122e-01
-1.21820772e+00 -3.85152638e-01 1.20524061e+00 -9.20635700e-01
1.88066974e-01 1.96091831e-01 4.89706278e-01 1.26576400e+00
2.23795503e-01 6.68025970e-01 1.31295955e+00 -6.19400382e-01
2.03889832e-01 2.63600945e-01 -1.32198110e-01 6.19808853e-01
1.20348586e-02 -1.29976928e-01 -8.04404080e-01 -1.13985464e-01
8.55336308e-01 -4.51728404e-01 -2.74985075e-01 7.27051422e-02
-1.43970335e+00 5.94239593e-01 1.09813206e-01 4.35019463e-01
4.19076979e-02 -2.11249605e-01 3.06034714e-01 4.68738705e-01
4.56179291e-01 1.18937981e+00 -1.56212166e-01 -3.20309103e-01
-1.12336433e+00 1.45809159e-01 9.44828570e-01 1.10278070e+00
4.64063108e-01 -2.64210284e-01 1.63085788e-01 9.83405173e-01
6.79106936e-02 4.64248061e-01 4.37008113e-01 -7.27964580e-01
4.78019744e-01 4.64895606e-01 1.74558982e-02 -1.43107212e+00
-1.54708594e-01 8.90312903e-03 -4.26594675e-01 3.03983420e-01
6.10206902e-01 1.48399219e-01 -6.63698494e-01 8.80839646e-01
-1.24946870e-01 -4.79714423e-01 -4.11521465e-01 7.97164559e-01
6.69188023e-01 2.63380080e-01 -3.54717076e-01 1.91905752e-01
1.09531796e+00 -5.21315753e-01 -5.70404410e-01 6.25495315e-02
1.27202377e-01 -1.16234505e+00 1.56024444e+00 5.04123569e-01
-1.07943976e+00 -4.10565674e-01 -1.28379786e+00 -2.34713256e-01
-5.65485060e-01 1.91593617e-01 6.70674518e-02 8.00783157e-01
-7.11437345e-01 1.19117785e+00 -4.40521330e-01 -4.30437565e-01
6.01819038e-01 -9.21237394e-02 -2.80911148e-01 1.41075611e-01
-9.09116983e-01 1.18101168e+00 -1.38879139e-02 1.11793756e-01
-5.23063123e-01 -1.00473940e+00 -3.13350469e-01 -5.45899332e-01
1.83516890e-01 -1.82887569e-01 9.89251673e-01 -8.13974261e-01
-1.45699918e+00 1.22817552e+00 2.50382632e-01 1.61569919e-02
1.18418169e+00 -4.36086625e-01 -4.70219284e-01 8.67392793e-02
6.79933205e-02 1.72975942e-01 8.60821843e-01 -1.81769645e+00
-3.28794688e-01 -3.26684117e-01 -4.77902949e-01 -3.94809153e-03
-3.86749029e-01 1.51818588e-01 -4.40948546e-01 -1.21040356e+00
1.98973313e-01 -6.03958189e-01 3.61105800e-01 5.52112877e-01
-4.70749348e-01 2.59831101e-01 3.14605027e-01 -1.27997220e+00
8.77047539e-01 -1.94028366e+00 2.49170959e-01 4.59524751e-01
3.37060587e-03 -1.88365549e-01 6.71968758e-02 6.59221590e-01
3.89406204e-01 3.54059428e-01 -5.92828572e-01 -3.36729795e-01
2.07073763e-01 7.03479648e-02 -4.13902789e-01 8.76775682e-01
-3.14193554e-02 8.02050769e-01 -7.44377792e-01 -5.98184407e-01
2.95888353e-02 4.42023017e-03 1.81123868e-01 1.03925250e-01
-2.35019118e-01 6.07218705e-02 4.22786996e-02 1.08471930e+00
5.42058945e-01 3.93147588e-01 -1.41005591e-01 -2.37621337e-01
-2.68041581e-01 4.11713310e-02 -9.38901186e-01 1.73114347e+00
-2.45013475e-01 1.07105112e+00 -6.82110041e-02 -3.90439361e-01
1.12098455e+00 -2.25399137e-01 -6.04961589e-02 -8.95999491e-01
4.15960997e-02 4.18723851e-01 -7.63348266e-02 -5.81378937e-01
9.16518331e-01 -6.44683987e-02 -6.92178607e-02 6.89879060e-01
-2.14162812e-01 -5.34211755e-01 -7.80821145e-02 1.19512536e-01
7.76038766e-01 1.43775910e-01 -2.33574376e-01 -5.10119021e-01
3.11992824e-01 2.80162152e-02 1.85019180e-01 8.66281271e-01
1.51347175e-01 1.27118480e+00 3.91518831e-01 -3.52437377e-01
-1.79418170e+00 -1.21503150e+00 -3.60803515e-01 8.06802571e-01
1.82942957e-01 -2.04400524e-01 -7.70373106e-01 -2.87476957e-01
2.59692430e-01 5.94875455e-01 -8.08809042e-01 9.88587458e-03
-6.03656113e-01 -4.81294513e-01 9.21570480e-01 5.17585695e-01
1.75551370e-01 -1.09607887e+00 -4.90353584e-01 1.75802082e-01
1.94229782e-01 -6.72316670e-01 -4.06246632e-01 -2.00000748e-01
-6.34057760e-01 -1.17100382e+00 -1.10105205e+00 -7.50081420e-01
6.56851888e-01 -4.38551307e-01 1.09249222e+00 5.85671812e-02
-6.10644639e-01 5.07915676e-01 -3.70044261e-01 -5.05158901e-01
-8.14850628e-01 1.33202374e-01 -8.29531066e-03 -1.32194504e-01
-3.02385222e-02 -2.71508723e-01 -2.10210741e-01 4.45451140e-01
-7.97654867e-01 -2.77284741e-01 5.17557502e-01 6.93527818e-01
3.10510725e-01 2.54483838e-02 2.43721142e-01 -8.93992603e-01
8.48974824e-01 -2.83543408e-01 -4.76579905e-01 5.88600099e-01
-5.35447598e-01 3.49456742e-02 5.41905642e-01 -6.82849288e-01
-1.02724016e+00 -3.44110250e-01 1.44702256e-01 3.61597650e-02
-4.91208173e-02 4.31344926e-01 -3.01161855e-01 -3.00433367e-01
7.50506580e-01 2.40290076e-01 -8.89540613e-02 -8.76171708e-01
4.38257992e-01 9.25042808e-01 1.27072477e+00 -9.96418655e-01
1.08674407e+00 3.50915462e-01 -3.31272513e-01 -1.08064640e+00
-6.96081594e-02 1.17599405e-01 -1.12553191e+00 -3.73097181e-01
4.97758031e-01 -5.95805943e-01 -2.49181941e-01 9.31621313e-01
-9.71248448e-01 -3.23061854e-01 -1.27550393e-01 2.42656991e-02
-3.82083923e-01 7.73423910e-01 -8.09821248e-01 -6.86655223e-01
-2.20961228e-01 -7.39902437e-01 9.65801358e-01 -2.06963252e-03
-4.96073216e-01 -9.16753411e-01 1.81892633e-01 3.40480030e-01
1.68440014e-01 3.58995289e-01 1.24683964e+00 -4.90368754e-01
-3.05204481e-01 -3.18716735e-01 -4.58485819e-02 1.20755389e-01
3.45029145e-01 5.66602707e-01 -9.14844036e-01 2.87236869e-02
-1.45574182e-01 -2.01736521e-02 5.34340680e-01 -3.58208805e-01
8.80621612e-01 -2.50366032e-01 -4.26316932e-02 4.42020357e-01
1.28394079e+00 1.59184441e-01 8.57036889e-01 9.82617974e-01
9.03110504e-01 7.52289236e-01 2.38464043e-01 4.69067276e-01
1.05838001e-01 4.22529101e-01 5.01888804e-02 6.60418943e-02
-2.53252000e-01 -2.76210874e-01 1.34848312e-01 8.07462871e-01
-3.87340784e-02 -1.99529156e-01 -1.19287241e+00 9.36151564e-01
-1.38070643e+00 -6.86908245e-01 -3.51686954e-01 2.19640899e+00
1.04393685e+00 3.07733625e-01 4.92543317e-02 2.67435759e-01
1.04729903e+00 -5.98485656e-02 -3.42156529e-01 -4.40970987e-01
-5.73510706e-01 1.05717734e-01 6.07428193e-01 1.33399233e-01
-9.14452255e-01 6.75160527e-01 7.02046061e+00 5.88598251e-01
-6.67487383e-01 -3.68156612e-01 3.96567643e-01 -2.72680353e-02
-5.63592792e-01 -3.17186385e-01 -2.52314985e-01 9.40910161e-01
4.87231821e-01 -9.62112918e-02 4.59656298e-01 4.67643023e-01
-2.15979263e-01 -9.21758190e-02 -1.28646421e+00 9.47566152e-01
4.52580541e-01 -1.39239156e+00 -2.72197053e-02 -3.60809900e-02
6.60638213e-01 -3.19404632e-01 1.31364182e-01 -4.75405514e-01
5.99476248e-02 -9.49338615e-01 1.33121359e+00 9.69973326e-01
8.74944448e-01 -8.27910542e-01 6.03136539e-01 -4.16223519e-02
-5.70497274e-01 7.54115358e-03 -3.14746022e-01 8.19105282e-02
1.13027915e-02 6.31837428e-01 -6.97660506e-01 3.31753731e-01
6.62040591e-01 5.91197133e-01 -1.08597744e+00 1.17233121e+00
-1.99759603e-01 4.10666168e-01 -2.63891459e-01 -4.54541892e-02
-1.86133847e-01 -1.09100625e-01 5.18543184e-01 1.42551899e+00
2.90882736e-01 -1.92026243e-01 -4.39080417e-01 1.22556818e+00
-3.26669961e-01 3.86925973e-02 -3.68978858e-01 -2.46987417e-01
6.44367278e-01 1.02149093e+00 -9.58557487e-01 5.71825765e-02
-1.01891197e-01 1.38817322e+00 2.81170785e-01 -9.48899388e-02
-4.85241801e-01 -5.51753461e-01 1.20818421e-01 3.81850928e-01
1.59264624e-01 -2.99543440e-01 -9.70915079e-01 -8.10388267e-01
4.52410132e-01 -8.70895565e-01 5.25210090e-02 -7.83123076e-01
-2.01145172e+00 1.81800127e-01 -2.00713441e-01 -8.83064032e-01
4.03937787e-01 -6.82683766e-01 -7.69210458e-01 8.78735363e-01
-8.76284480e-01 -1.21061921e+00 -1.02254882e-01 2.51086205e-01
5.63736856e-01 -4.33180064e-01 5.25682032e-01 1.19943164e-01
-3.72694075e-01 8.23350012e-01 8.73966932e-01 4.45904434e-01
1.02100313e+00 -1.59151864e+00 8.81046176e-01 7.47098804e-01
2.14764446e-01 6.24201179e-01 6.61849976e-01 -1.07492626e+00
-1.52701640e+00 -5.73744416e-01 7.03329563e-01 -8.81540060e-01
9.40356314e-01 -6.94374025e-01 -1.16260374e+00 3.07402253e-01
6.66465983e-02 -6.84532404e-01 5.30778468e-01 -1.01700716e-01
-4.64966774e-01 1.70948267e-01 -1.16915929e+00 6.35829270e-01
9.01092231e-01 -9.92614686e-01 -1.23424554e+00 2.71606356e-01
-7.02508762e-02 -3.45296472e-01 -1.22114575e+00 -1.39569119e-01
9.27043021e-01 -5.05580604e-01 9.39882398e-01 -7.68051371e-02
1.03599417e+00 -1.79882184e-01 4.58011813e-02 -1.10161602e+00
-3.04311782e-01 -7.43477046e-01 1.96701542e-01 1.72605538e+00
5.49470067e-01 -1.60850257e-01 6.14505291e-01 9.60639596e-01
-2.84884393e-01 -2.12008134e-01 -8.32923234e-01 -6.29742265e-01
4.94430333e-01 -8.35517570e-02 3.89299423e-01 1.03888142e+00
-4.95443568e-02 -2.27911189e-01 -2.40206540e-01 -1.16872832e-01
7.64485061e-01 5.11190221e-02 5.92633605e-01 -1.46464038e+00
-1.33806586e-01 -7.89705098e-01 -3.86663556e-01 -4.39300776e-01
6.83354884e-02 -6.84474230e-01 3.56253564e-01 -1.35897374e+00
1.16218843e-01 -4.87552851e-01 2.03862697e-01 1.48877114e-01
-1.72383085e-01 3.64980370e-01 2.15788975e-01 6.42325521e-01
1.30198613e-01 2.56999880e-01 9.16010857e-01 -3.92096400e-01
5.09897619e-02 -3.62885356e-01 -5.98615050e-01 6.71329558e-01
5.94467282e-01 -4.63967770e-01 1.58076420e-01 -6.83631957e-01
3.75750214e-01 -5.69983304e-01 5.18390596e-01 -1.08354425e+00
3.66905808e-01 2.07647964e-01 8.48522365e-01 -5.71514070e-01
-4.15867940e-02 -8.04940760e-01 2.05433622e-01 1.80559114e-01
-5.60628891e-01 2.63781846e-01 1.20180413e-01 6.17702901e-01
3.10516655e-01 -6.65751755e-01 5.68923593e-01 -3.96160394e-01
-6.78695500e-01 -3.45518380e-01 -3.99651021e-01 6.58851564e-02
7.23006308e-01 -5.27864337e-01 -5.61606109e-01 5.45384968e-03
-3.09601128e-01 -1.51063055e-01 1.10366476e+00 4.65418875e-01
5.79427838e-01 -1.11830771e+00 -8.38044465e-01 -5.41900424e-03
9.23898444e-02 -2.81437069e-01 -1.69377178e-01 1.64352655e-01
-1.04647291e+00 -3.76914740e-01 -4.94825214e-01 -2.51326740e-01
-1.04644132e+00 4.91567701e-01 -6.14317991e-02 5.96133530e-01
-9.97002304e-01 9.13834393e-01 -6.23647511e-01 -3.41979563e-01
3.65642726e-01 2.45695282e-02 6.15770230e-03 4.16820973e-01
3.89406979e-01 8.12206089e-01 2.98504591e-01 -6.69408143e-01
-2.29451224e-01 8.28987300e-01 -1.99395910e-01 -4.76762682e-01
1.30141377e+00 -1.93484694e-01 -1.93660357e-03 8.13367128e-01
1.07523477e+00 2.66957074e-01 -1.14531541e+00 -1.30488843e-01
4.08329964e-01 -7.62265921e-01 -1.60550416e-01 -1.18058574e+00
-8.20131600e-01 7.42327988e-01 5.40826082e-01 -8.41091573e-02
6.41941130e-01 -3.80247757e-02 7.28955805e-01 3.80175889e-01
2.66519070e-01 -1.61333191e+00 2.13222764e-02 1.48332655e-01
1.22340918e+00 -1.04840064e+00 4.28485334e-01 1.14628188e-01
-5.46929121e-01 1.43577790e+00 3.22108209e-01 -2.61060633e-02
5.32512181e-02 4.52943593e-01 5.10599501e-02 -3.27656597e-01
8.62028897e-02 3.37921828e-01 4.13058430e-01 7.78283358e-01
2.94970751e-01 -3.22981268e-01 -2.91401297e-01 3.52844715e-01
-6.25014246e-01 -3.85867447e-01 8.27273369e-01 1.35498345e+00
-2.44786769e-01 -9.71741259e-01 -8.57761919e-01 4.38176006e-01
-3.67992848e-01 -2.15507582e-01 -1.20815504e+00 8.92312348e-01
6.15734048e-02 5.38670838e-01 3.81038010e-01 -3.72682184e-01
4.81979370e-01 3.95974517e-02 7.23726273e-01 -2.21130818e-01
-8.98588836e-01 1.10809470e-03 -2.16966718e-02 2.93913037e-02
-3.04794371e-01 -9.12754416e-01 -7.23567545e-01 -5.40443301e-01
-2.82993287e-01 -1.97032347e-01 8.39732349e-01 7.35059619e-01
-3.79320718e-02 3.14122438e-01 3.77735347e-01 -9.79442358e-01
-5.02815545e-01 -8.23309958e-01 -8.34956288e-01 6.47806883e-01
1.14138931e-01 -5.19692779e-01 -4.45856243e-01 4.30548161e-01] | [10.188698768615723, 10.238249778747559] |
75cee51c-bff7-43a5-ad20-57ca7a42b47f | scaling-up-learning-with-gait-prop | 2102.11598 | null | https://arxiv.org/abs/2102.11598v3 | https://arxiv.org/pdf/2102.11598v3.pdf | Gradient-adjusted Incremental Target Propagation Provides Effective Credit Assignment in Deep Neural Networks | Many of the recent advances in the field of artificial intelligence have been fueled by the highly successful backpropagation of error (BP) algorithm, which efficiently solves the credit assignment problem in artificial neural networks. However, it is unlikely that BP is implemented in its usual form within biological neural networks, because of its reliance on non-local information in propagating error gradients. Since biological neural networks are capable of highly efficient learning and responses from BP trained models can be related to neural responses, it seems reasonable that a biologically viable approximation of BP underlies synaptic plasticity in the brain. Gradient-adjusted incremental target propagation (GAIT-prop or GP for short) has recently been derived directly from BP and has been shown to successfully train networks in a more biologically plausible manner. However, so far, GP has only been shown to work on relatively low-dimensional problems, such as handwritten-digit recognition. This work addresses some of the scaling issues in GP and shows it to perform effective multi-layer credit assignment in deeper networks and on the much more challenging ImageNet dataset. | ['Marcel van Gerven', 'Luca Ambrogioni', 'Nasir Ahmad', 'Sander Dalm'] | 2021-02-23 | null | null | null | null | ['handwritten-digit-recognition'] | ['computer-vision'] | [ 5.29337466e-01 2.74550736e-01 2.95340508e-01 -3.13793123e-01
4.84409370e-02 -3.60254645e-02 5.40626049e-01 1.88614398e-01
-7.63025939e-01 1.13565373e+00 -1.38716400e-01 -1.92498609e-01
-2.34866351e-01 -7.99603820e-01 -9.04284954e-01 -8.21313560e-01
-1.47238597e-01 6.15156054e-01 6.16842985e-01 -1.45799324e-01
6.84340298e-01 7.67754436e-01 -1.33704889e+00 2.23118871e-01
4.34957802e-01 9.73577201e-01 1.68269768e-01 7.18409956e-01
1.77068606e-01 9.48726356e-01 -5.12990654e-01 -3.52081656e-01
-4.84670885e-02 -5.78454733e-01 -8.23084414e-01 -5.01499176e-01
3.03486645e-01 -4.84064706e-02 -3.05854827e-01 6.80665016e-01
5.96182048e-01 1.51169643e-01 8.07225049e-01 -8.46210897e-01
-8.05272222e-01 2.29916736e-01 -2.31002435e-01 7.09218204e-01
-1.96836159e-01 1.15657307e-01 7.82522559e-01 -9.66201425e-01
5.01236618e-01 1.04376018e+00 9.74724293e-01 6.71913803e-01
-1.56453729e+00 -3.60877395e-01 -1.11732572e-01 3.24949361e-02
-9.84955609e-01 -2.74172068e-01 4.85814720e-01 -1.63632363e-01
1.62563527e+00 -9.25012454e-02 9.47761595e-01 1.07194793e+00
6.74238920e-01 6.10253990e-01 1.28441334e+00 -4.50049490e-01
4.07951832e-01 -8.63736123e-02 -8.71244296e-02 8.14108551e-01
1.81374162e-01 4.47896570e-01 -6.64910674e-01 -8.22122246e-02
1.01576436e+00 -3.51241887e-01 -2.35272154e-01 1.46591902e-01
-1.08814704e+00 8.48127484e-01 8.47440183e-01 3.99165660e-01
-5.73525131e-01 5.78645170e-01 1.00483082e-01 2.20598504e-01
2.41284654e-01 8.39295447e-01 -5.15222490e-01 -1.42541807e-02
-1.17647576e+00 3.36744934e-01 8.95061731e-01 2.04486653e-01
5.26854515e-01 3.92627686e-01 1.79172948e-01 8.71324420e-01
3.64877909e-01 -1.60000136e-03 7.20412850e-01 -1.22623444e+00
2.80323729e-04 3.13359320e-01 -1.30425230e-01 -1.28461051e+00
-5.58705211e-01 -4.27448392e-01 -8.21230054e-01 8.23500097e-01
7.55156457e-01 -4.38781269e-02 -9.61067319e-01 1.65694344e+00
-2.01180145e-01 1.73331469e-01 2.22421121e-02 7.86941469e-01
4.52506810e-01 6.82635307e-01 1.82782605e-01 2.98082009e-02
7.25472808e-01 -8.91257346e-01 -1.25554979e-01 -9.03096378e-01
1.86435103e-01 -2.06833258e-01 5.22404909e-01 4.97578353e-01
-1.34880185e+00 -4.17547882e-01 -1.27568150e+00 1.84322316e-02
-5.43235540e-01 -4.03546900e-01 8.07918668e-01 4.85923171e-01
-1.43309712e+00 1.23697805e+00 -9.33415592e-01 -5.98783314e-01
8.51706207e-01 8.37577760e-01 -2.72104770e-01 -1.45101491e-02
-1.17019427e+00 1.47784162e+00 4.81597334e-01 2.23905504e-01
-6.59776807e-01 -5.33512115e-01 -3.08115453e-01 8.57044607e-02
-4.44778979e-01 -9.69862938e-01 9.95340466e-01 -1.13680482e+00
-1.60739958e+00 1.01390898e+00 -1.18523292e-01 -9.64460373e-01
1.14927895e-01 3.21304470e-01 4.65787165e-02 2.90863812e-01
-5.12870312e-01 1.54108810e+00 7.54253089e-01 -7.74381101e-01
-4.59422380e-01 -3.86261731e-01 -1.08537890e-01 2.06482753e-01
-2.82240689e-01 1.40930608e-01 2.83921748e-01 -6.56793177e-01
3.71610612e-01 -8.31138194e-01 -2.90659308e-01 4.65737700e-01
2.65185744e-01 -1.16092183e-01 1.86765105e-01 -7.00616658e-01
5.81167936e-01 -1.91578126e+00 1.36845618e-01 1.04294665e-01
1.37913480e-01 2.97263145e-01 -3.57701294e-02 2.31177121e-01
-1.14144534e-01 5.77563047e-02 -6.10085726e-01 8.26617796e-03
-1.47594303e-01 5.17511189e-01 -2.37198234e-01 2.78803527e-01
7.23151863e-01 1.08132195e+00 -7.62979686e-01 -3.69761646e-01
-2.51773149e-01 6.32298172e-01 -3.52341920e-01 -3.60843509e-01
-1.72170311e-01 1.43674016e-01 -4.57292004e-03 4.25639689e-01
2.22195834e-01 -2.16215923e-01 -2.15259969e-01 2.21349210e-01
-4.04966101e-02 3.18993360e-01 -5.97753882e-01 1.57567835e+00
4.79382165e-02 1.03639662e+00 -1.35607988e-01 -1.38490176e+00
1.23894918e+00 4.54977304e-01 1.86538994e-01 -8.15060318e-01
7.53517598e-02 5.93281329e-01 4.85646278e-01 -3.24282907e-02
3.70367095e-02 -6.57258213e-01 6.01625919e-01 4.88919258e-01
2.80801177e-01 -1.85223788e-01 1.77469671e-01 -2.74075150e-01
1.39637589e+00 1.92974791e-01 -8.55421796e-02 -3.01439524e-01
2.62907743e-01 3.25023383e-01 2.88819790e-01 7.81633377e-01
-4.12384957e-01 8.72095644e-01 1.48683161e-01 -5.43166041e-01
-1.11045825e+00 -8.95367920e-01 -4.19650823e-01 8.73935699e-01
-2.61789143e-01 4.77607340e-01 -6.54418528e-01 1.87753234e-02
-1.77476145e-02 4.89038229e-01 -5.65495729e-01 -5.43533742e-01
-5.60065806e-01 -1.30404854e+00 9.95679557e-01 7.66786098e-01
5.13603985e-01 -1.74478805e+00 -8.69076014e-01 9.23717201e-01
5.20883739e-01 -5.50366879e-01 4.14055526e-01 8.98557007e-01
-1.39032006e+00 -6.14518762e-01 -1.13478720e+00 -1.21217620e+00
7.34646320e-01 -3.04744154e-01 1.03115773e+00 7.89800510e-02
-4.46537346e-01 6.83643371e-02 2.37112373e-01 -5.41699350e-01
-3.70296985e-01 -5.58364578e-02 -9.46761817e-02 -4.58350956e-01
6.89408898e-01 -9.21359062e-01 -6.37299180e-01 1.81443006e-01
-6.49944901e-01 -6.11620322e-02 7.57711112e-01 1.09793401e+00
5.76897442e-01 -1.27024669e-02 8.99370432e-01 -5.94508290e-01
9.22791958e-01 -2.91969419e-01 -1.59647152e-01 -3.37655395e-02
-7.87510753e-01 1.81918442e-01 3.80642176e-01 -3.92204165e-01
-8.70569944e-01 -1.25497356e-01 -3.93046409e-01 -5.73177217e-03
-2.55191773e-01 8.27785611e-01 7.59301007e-01 -6.66690528e-01
1.03303766e+00 3.70215684e-01 -1.96907036e-02 -1.31410155e-02
-1.55405089e-01 1.48941413e-01 6.20505095e-01 -5.55695772e-01
2.06173822e-01 3.32102895e-01 2.40941226e-01 -7.08042920e-01
-4.06683922e-01 1.24240771e-01 -4.29883152e-01 -1.15937032e-01
7.73409128e-01 -4.55193818e-01 -5.05530059e-01 8.16397488e-01
-1.19791269e+00 -5.70833683e-01 -4.16792452e-01 4.61359948e-01
-7.88183093e-01 -1.86306193e-01 -1.14522564e+00 -5.96412480e-01
-2.25601166e-01 -6.99393272e-01 1.23940177e-01 4.33571100e-01
-3.00424516e-01 -1.26612961e+00 3.06497753e-01 -1.79347932e-01
7.70746946e-01 2.40629390e-01 1.07383108e+00 -5.94237983e-01
-2.22556859e-01 -3.38964134e-01 -3.04957300e-01 4.14037704e-01
-3.74303877e-01 1.07939228e-01 -1.19904029e+00 -2.91018691e-02
1.85889557e-01 -7.43743658e-01 1.31595707e+00 6.19800866e-01
8.19864929e-01 -8.17013383e-02 -3.50729942e-01 3.19918096e-01
1.43035758e+00 2.15994716e-01 7.97221065e-01 5.13288617e-01
7.42471367e-02 5.83582997e-01 -9.19310525e-02 -2.24666715e-01
4.50553410e-02 6.53596520e-02 4.13377613e-01 -2.43173428e-02
-1.29915655e-01 -4.74098735e-02 9.18430537e-02 6.25835955e-01
-4.04226899e-01 -6.49519265e-02 -1.19875431e+00 4.42012817e-01
-1.69769943e+00 -1.18745875e+00 -5.62379323e-02 1.80958378e+00
1.07084692e+00 7.89659441e-01 -3.28718603e-01 4.37135875e-01
5.32686353e-01 -1.66474029e-01 -8.59779477e-01 -9.08579051e-01
-4.02210295e-01 7.70073950e-01 4.44023728e-01 3.11954588e-01
-8.32103550e-01 5.32813013e-01 7.40641832e+00 3.32044721e-01
-1.08817184e+00 1.29747521e-02 1.00891423e+00 -3.64509933e-02
2.36057922e-01 -3.04659069e-01 -7.85722196e-01 3.71338367e-01
1.46055925e+00 5.66029809e-02 5.77245593e-01 6.03622079e-01
-9.57683623e-02 -3.32102984e-01 -1.23213673e+00 7.60720670e-01
-9.25914571e-02 -1.52592301e+00 -2.62273252e-01 -3.43260402e-03
7.70365894e-01 4.42159623e-01 1.83448941e-01 3.53488803e-01
1.57690242e-01 -1.33352709e+00 6.49124742e-01 6.95585370e-01
1.63368523e-01 -6.63092077e-01 5.10413885e-01 5.85457385e-01
-3.67894918e-01 -2.37389609e-01 -1.01299179e+00 -4.29693073e-01
-1.01008050e-01 5.44861853e-01 -5.99364281e-01 -4.58575606e-01
6.53361619e-01 6.78201437e-01 -5.33474445e-01 1.51245952e+00
-3.14041287e-01 6.13996744e-01 -5.73120952e-01 -1.80613056e-01
4.41722631e-01 1.35925367e-01 1.72106519e-01 1.29084647e+00
2.66263455e-01 1.00081153e-01 -5.04862368e-01 1.14550471e+00
-1.06697083e-01 -3.46691459e-01 -3.39525312e-01 -3.65736634e-01
3.14337254e-01 9.51421440e-01 -1.03818417e+00 -4.47763875e-02
-2.08253279e-01 9.83588398e-01 5.88029444e-01 3.55715334e-01
-6.20657504e-01 -3.57762575e-01 3.75888079e-01 2.21742559e-02
5.05971372e-01 -2.47806892e-01 -5.64969361e-01 -4.90419924e-01
-1.97120875e-01 -4.63427275e-01 -2.21545156e-02 -1.06549096e+00
-1.28184605e+00 5.76192975e-01 -6.08951926e-01 -2.39330918e-01
-2.52517700e-01 -1.17612517e+00 -6.95632696e-01 1.23891366e+00
-1.50688660e+00 -5.83899319e-01 1.74230054e-01 4.67112660e-01
2.55720854e-01 1.24462292e-01 1.21394372e+00 8.52766782e-02
-3.69663835e-01 2.66910583e-01 2.37978816e-01 -1.40589699e-01
3.88148129e-01 -1.24650478e+00 5.94627798e-01 5.84373653e-01
2.49176845e-01 5.81952035e-01 6.97232246e-01 -4.06893671e-01
-8.78240883e-01 -7.05021441e-01 1.08598292e+00 -2.57686883e-01
6.81176245e-01 -2.11802553e-02 -1.22693253e+00 3.71420771e-01
1.34837747e-01 1.77984923e-01 3.73601735e-01 -3.35125178e-01
-1.57814220e-01 2.56404653e-02 -1.26731682e+00 3.74937028e-01
9.05367374e-01 -3.06453645e-01 -1.03661239e+00 2.39109203e-01
1.71840504e-01 -1.52357742e-01 -7.87248075e-01 1.43728778e-01
4.86269176e-01 -9.69975173e-01 1.07093191e+00 -5.55398881e-01
6.97527170e-01 -1.14974923e-01 1.10468738e-01 -1.35269892e+00
-6.56437278e-01 -1.10083945e-01 -3.30172181e-01 8.78201783e-01
7.44657159e-01 -7.99263954e-01 9.60829854e-01 9.43843961e-01
-2.28530094e-01 -1.22438157e+00 -1.08196330e+00 -6.53733075e-01
4.49057311e-01 -5.22576692e-03 -1.00771584e-01 6.32458746e-01
5.13370149e-02 2.21169084e-01 8.98390338e-02 -2.13731527e-01
4.37748641e-01 -3.20790321e-01 -2.16965258e-01 -1.54219639e+00
-6.21028721e-01 -8.94978225e-01 -7.50559926e-01 -7.88215339e-01
-6.54440075e-02 -8.20775449e-01 3.37816387e-01 -1.90520871e+00
-1.02377392e-01 -4.16004211e-01 -6.33417428e-01 5.78380466e-01
1.51522964e-01 8.18765998e-01 -7.90769756e-02 4.16859299e-01
2.69652978e-02 9.24402848e-02 1.05877960e+00 -1.36858985e-01
8.11155438e-02 -2.03806207e-01 -5.37863731e-01 8.18603396e-01
1.01617134e+00 -8.07871699e-01 -2.02651620e-01 -5.22500098e-01
5.34924030e-01 -3.07548404e-01 5.11385858e-01 -1.43850040e+00
4.88455415e-01 5.84947206e-02 9.93931592e-01 -4.41044904e-02
5.42164862e-01 -5.23459554e-01 1.25263631e-02 8.63258064e-01
-5.16476214e-01 -4.44166251e-02 3.64495277e-01 5.41423976e-01
-2.25137338e-01 -5.86833179e-01 1.02964377e+00 -6.76896214e-01
-7.30332077e-01 1.70755446e-01 -8.35457683e-01 2.27407534e-02
7.47729421e-01 -8.93694222e-01 -4.47905838e-01 1.56020224e-01
-9.80265319e-01 -1.39511228e-01 2.43168786e-01 -8.24513473e-03
8.39819252e-01 -1.04567885e+00 -6.49829030e-01 1.01574078e-01
-3.75211984e-01 -2.91329831e-01 -3.07243377e-01 9.40249324e-01
-8.52140188e-01 5.95176399e-01 -7.79819369e-01 -4.01221007e-01
-5.55591106e-01 3.81910443e-01 7.07174480e-01 -8.43484420e-03
-4.90258873e-01 1.52853382e+00 -2.59288639e-01 -9.25089195e-02
2.99738079e-01 -1.89453453e-01 -8.84583369e-02 -2.15126872e-01
4.93828893e-01 8.51038396e-02 1.94039375e-01 -2.63300866e-01
-3.65510225e-01 2.13731691e-01 -1.52018249e-01 -2.30583802e-01
1.81940329e+00 2.74656802e-01 -1.73443392e-01 5.26701510e-01
7.62440801e-01 -9.18359578e-01 -1.67349744e+00 1.07694715e-01
1.93678334e-01 9.73151848e-02 1.26162872e-01 -1.16691470e+00
-9.21377122e-01 1.51036131e+00 4.56519961e-01 6.42133355e-02
9.19829905e-01 -3.78627092e-01 7.19654441e-01 7.06336677e-01
5.38361371e-01 -1.05203938e+00 1.87742516e-01 5.44459283e-01
8.36983502e-01 -9.66823101e-01 -8.45190510e-02 3.92625630e-01
-1.05502978e-01 1.46497524e+00 5.78507602e-01 -6.83430970e-01
6.59009933e-01 3.10517788e-01 -2.76196331e-01 -1.43424153e-01
-8.65777552e-01 3.26359451e-01 -3.31083015e-02 8.37967873e-01
5.51048219e-01 -4.11014259e-01 -3.13978374e-01 1.46854252e-01
1.11995995e-01 4.17866617e-01 4.40519512e-01 1.09474266e+00
-1.04899180e+00 -8.08584988e-01 -1.67261437e-01 6.35632873e-01
-4.63904440e-01 -3.85677010e-01 -4.77519929e-01 5.76354504e-01
6.90692142e-02 4.50644225e-01 4.75270338e-02 -1.17642209e-02
-6.88727647e-02 4.32612628e-01 1.01738560e+00 -5.64623415e-01
-8.52777123e-01 -4.62021828e-01 -1.79404870e-01 -2.43346646e-01
-5.30018389e-01 -4.21871781e-01 -1.52791595e+00 -3.95713776e-01
-1.47026420e-01 -9.46867168e-02 8.50685537e-01 9.87758517e-01
2.59729296e-01 4.78398651e-01 -1.17239319e-01 -1.35333169e+00
-5.16959667e-01 -8.02897692e-01 -5.85744679e-01 5.11677600e-02
2.43740492e-02 -5.93373954e-01 -4.51925874e-01 8.03676099e-02] | [8.201138496398926, 2.886044979095459] |
f0aabdf7-e8dc-43eb-b9c6-fbfb9b555526 | graphganfed-a-federated-generative-framework | 2304.05498 | null | https://arxiv.org/abs/2304.05498v1 | https://arxiv.org/pdf/2304.05498v1.pdf | GraphGANFed: A Federated Generative Framework for Graph-Structured Molecules Towards Efficient Drug Discovery | Recent advances in deep learning have accelerated its use in various applications, such as cellular image analysis and molecular discovery. In molecular discovery, a generative adversarial network (GAN), which comprises a discriminator to distinguish generated molecules from existing molecules and a generator to generate new molecules, is one of the premier technologies due to its ability to learn from a large molecular data set efficiently and generate novel molecules that preserve similar properties. However, different pharmaceutical companies may be unwilling or unable to share their local data sets due to the geo-distributed and sensitive nature of molecular data sets, making it impossible to train GANs in a centralized manner. In this paper, we propose a Graph convolutional network in Generative Adversarial Networks via Federated learning (GraphGANFed) framework, which integrates graph convolutional neural Network (GCN), GAN, and federated learning (FL) as a whole system to generate novel molecules without sharing local data sets. In GraphGANFed, the discriminator is implemented as a GCN to better capture features from molecules represented as molecular graphs, and FL is used to train both the discriminator and generator in a distributive manner to preserve data privacy. Extensive simulations are conducted based on the three bench-mark data sets to demonstrate the feasibility and effectiveness of GraphGANFed. The molecules generated by GraphGANFed can achieve high novelty (=100) and diversity (> 0.9). The simulation results also indicate that 1) a lower complexity discriminator model can better avoid mode collapse for a smaller data set, 2) there is a tradeoff among different evaluation metrics, and 3) having the right dropout ratio of the generator and discriminator can avoid mode collapse. | ['Xiang Sun', 'Wuji Liu', 'Jingjing Yao', 'Daniel Manu'] | 2023-04-11 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 2.48352721e-01 -6.63894974e-03 -1.96100518e-01 -7.14369342e-02
-5.53200245e-01 -8.22445393e-01 3.97130907e-01 1.04799289e-02
-1.14184767e-01 1.17106164e+00 -1.12477034e-01 -3.78888458e-01
1.36448890e-01 -1.22758806e+00 -8.54414940e-01 -1.15929317e+00
-3.49109024e-02 3.77914831e-02 -1.21350616e-01 -1.67671159e-01
-1.23272233e-01 6.65636241e-01 -8.81279945e-01 2.16828018e-01
1.16751325e+00 8.13069284e-01 -1.54699385e-01 4.61823940e-01
-8.59097838e-02 6.26408577e-01 -8.52445066e-01 -5.81090331e-01
5.65705657e-01 -8.76835346e-01 -2.79826850e-01 -4.71965611e-01
-2.80355220e-03 -3.01343381e-01 -6.04275584e-01 1.12672126e+00
9.05354798e-01 -1.11447930e-01 4.87169832e-01 -1.34894013e+00
-9.09062684e-01 4.82751906e-01 -2.36373901e-01 -6.46318048e-02
1.33380458e-01 6.70305729e-01 5.83941460e-01 -2.54289180e-01
6.77043617e-01 1.10621643e+00 4.13741916e-01 8.03595722e-01
-1.15481102e+00 -1.21192861e+00 -1.82398394e-01 -4.57722723e-01
-1.28204799e+00 -3.30459654e-01 6.05426908e-01 -3.53785068e-01
5.17213941e-01 3.19542974e-01 6.21673048e-01 1.13495326e+00
8.01510632e-01 3.39464337e-01 9.44876969e-01 1.45152450e-01
4.12068337e-01 -3.87618467e-02 -6.33160889e-01 6.62926614e-01
3.75976324e-01 4.30441886e-01 -2.34807014e-01 -4.41202849e-01
1.04737997e+00 3.80662322e-01 -3.90025526e-01 -2.77742505e-01
-8.43811929e-01 1.01352441e+00 8.47191215e-01 5.88950701e-02
-3.55020672e-01 2.20533922e-01 4.35648233e-01 5.45078218e-01
9.75858718e-02 5.11624515e-01 -8.52513406e-03 3.71252805e-01
-3.87411028e-01 1.69017628e-01 7.09794998e-01 9.48861539e-01
6.67917550e-01 4.11769450e-01 -4.88546103e-01 2.48424515e-01
5.80118448e-02 5.37230134e-01 7.83892691e-01 -2.96715707e-01
3.13987851e-01 8.40712309e-01 -3.56855243e-01 -1.06836748e+00
-7.91971907e-02 -6.10124230e-01 -1.28665841e+00 7.44104087e-02
1.23658642e-01 -4.77093726e-01 -9.99936283e-01 1.89108074e+00
2.99249738e-01 2.65219897e-01 3.38540763e-01 6.56388462e-01
9.72426355e-01 7.96877265e-01 1.25499204e-01 -1.33410469e-01
1.04386306e+00 -6.00489855e-01 -5.05977511e-01 2.15836063e-01
6.41604602e-01 -4.99537408e-01 6.57437503e-01 -1.58315077e-02
-7.95369804e-01 -3.13179165e-01 -1.26124346e+00 1.76606119e-01
-7.16692507e-01 -2.41507873e-01 8.59348178e-01 8.22428286e-01
-7.50959694e-01 5.69770992e-01 -4.53810155e-01 -5.55902086e-02
1.05637252e+00 7.81174481e-01 -4.37908798e-01 -1.26561984e-01
-1.40034294e+00 1.01692110e-01 5.60437024e-01 -1.41573697e-01
-1.31751847e+00 -6.68787241e-01 -6.29926383e-01 6.98748082e-02
1.03747575e-02 -1.18225515e+00 4.58331883e-01 -1.04026318e+00
-1.50540948e+00 4.95690078e-01 5.95018506e-01 -6.24541402e-01
5.24612486e-01 5.51120877e-01 -5.98928452e-01 -7.57129043e-02
-1.33903861e-01 5.21066248e-01 5.94482064e-01 -9.09851074e-01
-4.39735085e-01 -3.51129949e-01 8.49087015e-02 -1.59294661e-02
-1.63636789e-01 -4.42375779e-01 1.17650010e-01 -8.32749069e-01
-3.75221461e-01 -8.17702413e-01 -3.11660439e-01 -1.93555266e-01
-6.36815190e-01 2.05942929e-01 8.71213496e-01 -3.35786432e-01
1.01738465e+00 -2.05273819e+00 1.92426797e-02 4.13038194e-01
5.58695555e-01 5.82190335e-01 -3.01409841e-01 6.02166355e-01
-1.69127882e-01 4.70914215e-01 -2.35505328e-01 2.27205962e-01
-4.13299203e-01 5.95338829e-02 -2.39565536e-01 4.21178848e-01
1.34568304e-01 1.27896321e+00 -1.10250270e+00 -1.93447582e-02
-7.76699483e-02 4.95054185e-01 -3.79610419e-01 3.50523740e-01
-3.57314557e-01 6.48958862e-01 -8.20688725e-01 7.72515476e-01
8.05042386e-01 -2.56429762e-01 1.42298087e-01 -1.20787978e-01
1.84320450e-01 -2.71569580e-01 -8.33057940e-01 1.46975017e+00
-1.30473003e-01 1.15904860e-01 -1.04400456e-01 -6.66414320e-01
1.21046257e+00 3.82262826e-01 3.59989703e-01 -8.33084881e-01
2.84557730e-01 2.69533217e-01 1.27895877e-01 -1.61951706e-01
-5.35246246e-02 -3.18661481e-01 -1.61169365e-01 3.73041064e-01
8.72106850e-02 2.57504582e-01 -1.68641210e-01 1.74050748e-01
1.41077054e+00 -3.62456679e-01 1.00913532e-01 -5.93363605e-02
6.37682319e-01 -3.22417945e-01 7.67654002e-01 5.05200028e-01
-3.32783088e-02 3.56114239e-01 5.38438737e-01 -5.19214094e-01
-9.32469010e-01 -1.06857407e+00 2.50438094e-01 5.29223621e-01
7.61791915e-02 -1.27094269e-01 -6.04064465e-01 -9.21046436e-01
3.37871127e-02 2.26537108e-01 -4.68593240e-01 -7.16466308e-01
-1.65186986e-01 -7.84335077e-01 8.67935658e-01 1.78589433e-01
7.84616053e-01 -1.01525128e+00 -4.77641970e-02 2.58526593e-01
4.24327254e-01 -4.48607266e-01 -6.08839631e-01 1.98507279e-01
-6.18399858e-01 -1.13482416e+00 -6.07087731e-01 -7.86118746e-01
8.38112354e-01 1.83040008e-01 5.93452930e-01 2.12800279e-02
-2.41333410e-01 -2.83181012e-01 -2.25664571e-01 -4.58091438e-01
-7.39454687e-01 1.10215373e-01 -8.15539211e-02 2.46679172e-01
7.25222453e-02 -7.66475618e-01 -8.52215052e-01 9.68963951e-02
-1.31814241e+00 -2.48530477e-01 7.39281118e-01 1.11192000e+00
6.99480534e-01 2.70333022e-01 9.81438100e-01 -1.44264424e+00
8.50537658e-01 -7.36097813e-01 -7.55006969e-01 1.98352143e-01
-6.20160222e-01 -2.73931213e-02 1.24602067e+00 -4.85824436e-01
-6.47938907e-01 4.11621854e-02 -7.28429854e-02 -4.60288167e-01
2.48620048e-01 4.01056975e-01 -7.02214003e-01 -5.09128332e-01
6.39402211e-01 3.87342572e-01 3.14731866e-01 -4.77388389e-02
3.42563868e-01 5.67614198e-01 3.63410205e-01 -2.70676136e-01
8.83603156e-01 3.21270883e-01 2.82352388e-01 -1.98244169e-01
-6.53287619e-02 1.47681564e-01 1.52835427e-02 1.88044757e-01
6.43057525e-01 -1.09732473e+00 -8.42407584e-01 5.33328950e-01
-8.12998712e-01 -4.73948568e-02 -3.02220136e-01 2.76866645e-01
-2.45495915e-01 -9.22400691e-03 -5.34271300e-01 -4.50360179e-01
-8.67154837e-01 -1.07882535e+00 5.78648388e-01 7.26853251e-01
2.15385407e-01 -9.89234567e-01 -5.30418241e-03 9.20440555e-02
6.73845470e-01 1.22263718e+00 1.19794822e+00 -8.57036233e-01
-9.49682117e-01 -4.76501197e-01 4.05063219e-02 2.57357448e-01
7.13300228e-01 -1.66205242e-01 -7.85535991e-01 -7.24748015e-01
-9.21397731e-02 -1.54640287e-01 6.24813735e-01 1.26951322e-01
1.05626106e+00 -8.51148486e-01 -5.30306935e-01 1.09569192e+00
1.62179649e+00 8.93251836e-01 9.85394061e-01 3.69358584e-02
8.67700815e-01 4.46397103e-02 8.49949569e-02 3.09938014e-01
-7.86753297e-02 1.46483853e-01 6.69575691e-01 -5.16335428e-01
-2.03448180e-02 -7.74657667e-01 3.15868318e-01 5.05381346e-01
2.33299792e-01 -7.24268734e-01 -3.37124795e-01 2.10467160e-01
-1.55585027e+00 -9.87832010e-01 3.13119590e-01 2.41463947e+00
8.91502023e-01 -1.05535910e-01 1.79654106e-01 -3.07550073e-01
8.08288038e-01 1.30482130e-02 -1.03387880e+00 -3.98123860e-01
-3.62933964e-01 5.67776680e-01 7.67787039e-01 7.23342691e-03
-7.48539925e-01 7.66856611e-01 5.63604021e+00 1.08831072e+00
-1.57705772e+00 -5.54292798e-02 1.12225413e+00 -4.30202633e-02
-6.65713489e-01 3.23502310e-02 -4.76223975e-01 9.80123937e-01
8.10012043e-01 -5.06514430e-01 5.01483142e-01 6.92810595e-01
-1.81254856e-02 5.83476961e-01 -9.42853153e-01 8.88266981e-01
-2.85453051e-01 -1.82473779e+00 6.24687731e-01 3.41536283e-01
9.66577053e-01 -2.09786952e-01 3.57479274e-01 5.99088743e-02
4.81213540e-01 -1.52490973e+00 1.10152803e-01 4.42567885e-01
1.04840124e+00 -1.27099442e+00 7.86166728e-01 2.47164249e-01
-1.02501321e+00 -1.52713908e-02 -4.13061529e-01 2.90325403e-01
-2.90652037e-01 5.24521887e-01 -8.88135493e-01 9.33908582e-01
1.84520617e-01 4.72258121e-01 -3.98282290e-01 8.29333663e-01
1.76382251e-02 3.60893667e-01 -1.79260131e-02 -1.02397770e-01
2.29024887e-01 -4.44946378e-01 4.25673813e-01 7.32882142e-01
3.20205212e-01 -5.66150015e-03 4.97319512e-02 1.04952061e+00
-6.49411321e-01 1.94739461e-01 -9.28055048e-01 -5.09627640e-01
6.96052730e-01 1.19320965e+00 -3.91745269e-01 2.86763404e-02
-7.03565106e-02 8.15683365e-01 1.00221179e-01 3.50975811e-01
-8.65318298e-01 -8.56524169e-01 7.06976414e-01 2.15269968e-01
1.19904049e-01 2.16116726e-01 -6.56515593e-03 -7.83779740e-01
-3.71705711e-01 -1.16250169e+00 4.64631975e-01 -3.35420221e-01
-1.48796535e+00 7.74297655e-01 -7.21658826e-01 -1.25311661e+00
-1.33277150e-02 -2.92690724e-01 -8.07014763e-01 1.14393437e+00
-1.23362219e+00 -1.20606923e+00 -2.85701334e-01 8.09141278e-01
-1.87223047e-01 -5.60678244e-01 8.77861083e-01 3.33116204e-01
-7.24999905e-01 1.03583002e+00 4.44463760e-01 2.40798011e-01
5.13796806e-01 -8.00599754e-01 2.06045419e-01 7.83873856e-01
-2.11838126e-01 8.82175207e-01 1.85887158e-01 -8.20126891e-01
-1.71403134e+00 -1.58086276e+00 2.37783447e-01 -1.20532103e-01
1.53725684e-01 -4.86802459e-01 -7.16696680e-01 5.43302715e-01
-4.23715077e-02 2.08738342e-01 1.04169941e+00 -6.35650933e-01
-2.96074212e-01 -4.34623331e-01 -1.72246325e+00 5.63137829e-01
8.14237058e-01 -4.35942173e-01 3.40087950e-01 5.27605057e-01
9.62879896e-01 -5.39817810e-01 -1.14060938e+00 3.13169539e-01
4.34833616e-01 -8.04707944e-01 8.17503691e-01 -6.27468050e-01
3.58628720e-01 -5.91496468e-01 7.74897216e-03 -1.38903427e+00
-3.33534539e-01 -1.02028477e+00 -9.20349434e-02 1.50017571e+00
3.18802059e-01 -1.17835665e+00 9.26699400e-01 3.87875348e-01
9.85612795e-02 -1.01147962e+00 -9.83645737e-01 -9.10702586e-01
1.91802040e-01 6.43657446e-01 1.43103659e+00 9.52697754e-01
-3.88653338e-01 4.35987979e-01 -4.08455402e-01 6.04076032e-03
3.25430065e-01 9.73539650e-02 9.19685900e-01 -7.67581999e-01
-3.05851817e-01 -3.45294267e-01 -6.81138754e-01 -6.31556213e-01
-1.00235954e-01 -1.22170198e+00 -4.29623455e-01 -1.13973069e+00
5.12656383e-02 -7.40195215e-01 -5.86353123e-01 4.72251952e-01
-8.01450163e-02 4.21066396e-02 -4.24122587e-02 8.60513896e-02
-1.51808292e-01 6.40037000e-01 1.51183367e+00 -3.91396493e-01
-3.10040385e-01 -4.58780527e-02 -1.17544436e+00 -1.03737665e-02
8.36595953e-01 -4.99163419e-01 -6.65751576e-01 -5.37154712e-02
1.20636886e-02 1.26686811e-01 2.84433752e-01 -9.91257370e-01
2.02282116e-01 -2.46140584e-01 7.20960617e-01 -4.43061627e-03
-8.73223618e-02 -7.81218410e-01 9.88212645e-01 7.92053699e-01
5.08812675e-03 -6.81592748e-02 9.98682380e-02 8.02375555e-01
-1.14813790e-01 3.41881961e-01 7.43734062e-01 -1.92141265e-01
-1.50948137e-01 1.03839457e+00 1.03974581e-01 -1.02357738e-01
1.36929476e+00 -2.99945831e-01 -6.16917431e-01 -3.01841885e-01
-1.52864218e-01 2.30856225e-01 7.55820513e-01 2.28947550e-01
5.93020201e-01 -1.57027447e+00 -6.37239277e-01 5.71054637e-01
1.56996012e-01 1.17012970e-01 3.55610013e-01 2.63452232e-01
-6.76511228e-01 2.62283593e-01 -4.17494476e-01 -1.97655067e-01
-9.08933938e-01 7.78964758e-01 4.92886841e-01 -3.48070174e-01
-1.46230310e-01 8.79460096e-01 4.31624591e-01 -2.50876486e-01
-1.51495546e-01 4.59031947e-03 1.75062805e-01 -4.12239403e-01
3.39970678e-01 2.46639416e-01 2.43651331e-03 -3.69345516e-01
-3.10304761e-01 2.24715158e-01 -2.77413964e-01 6.06421530e-01
1.22599781e+00 3.98630202e-01 -1.51254296e-01 -2.51859695e-01
1.32614613e+00 5.89705370e-02 -1.24614382e+00 8.21127295e-02
-7.59945154e-01 -4.74887639e-01 -1.06160842e-01 -9.12408769e-01
-1.44495857e+00 4.55219865e-01 7.13162124e-01 9.19241086e-02
1.20762932e+00 -4.14617926e-01 1.15231264e+00 -1.11321956e-01
4.26938564e-01 -5.15840530e-01 1.68135092e-02 -5.04664406e-02
5.36116898e-01 -8.14000070e-01 -3.11571583e-02 -3.15845758e-01
-4.54940200e-01 9.37110186e-01 6.30550086e-01 -1.54713750e-01
3.03667516e-01 6.00756854e-02 -1.42528504e-01 -3.00521493e-01
-5.78106821e-01 3.07869166e-01 -8.24268386e-02 9.01386619e-01
1.49594620e-01 2.09005371e-01 -1.80767953e-01 7.17741132e-01
-1.32018521e-01 1.00542076e-01 3.72584105e-01 9.73412156e-01
-4.48521338e-02 -1.37798643e+00 8.59543011e-02 6.78051949e-01
-5.36937237e-01 -7.84917027e-02 -6.06855214e-01 5.31425297e-01
4.87386256e-01 8.28543603e-01 -2.48976812e-01 -6.70110822e-01
1.70298785e-01 -3.23215187e-01 2.49751449e-01 -4.58602965e-01
-8.39635253e-01 -2.05418274e-01 -4.23686296e-01 -4.91885126e-01
-8.18938203e-03 2.70403177e-03 -1.29555154e+00 -7.11737514e-01
-3.72228980e-01 3.55291426e-01 3.43105733e-01 4.55874294e-01
9.32802618e-01 7.73478031e-01 1.16069436e+00 -3.01723205e-03
-3.32080126e-01 -3.73368829e-01 -7.31716394e-01 3.08211744e-01
2.97863424e-01 -2.35071510e-01 -1.99677184e-01 -1.20909624e-01] | [5.802748680114746, 7.233469009399414] |
1d07ec95-cb17-4d3b-9378-224399118f72 | hybrid-neural-diffeomorphic-flow-for-shape | 2307.01957 | null | https://arxiv.org/abs/2307.01957v1 | https://arxiv.org/pdf/2307.01957v1.pdf | Hybrid Neural Diffeomorphic Flow for Shape Representation and Generation via Triplane | Deep Implicit Functions (DIFs) have gained popularity in 3D computer vision due to their compactness and continuous representation capabilities. However, addressing dense correspondences and semantic relationships across DIF-encoded shapes remains a critical challenge, limiting their applications in texture transfer and shape analysis. Moreover, recent endeavors in 3D shape generation using DIFs often neglect correspondence and topology preservation. This paper presents HNDF (Hybrid Neural Diffeomorphic Flow), a method that implicitly learns the underlying representation and decomposes intricate dense correspondences into explicitly axis-aligned triplane features. To avoid suboptimal representations trapped in local minima, we propose hybrid supervision that captures both local and global correspondences. Unlike conventional approaches that directly generate new 3D shapes, we further explore the idea of shape generation with deformed template shape via diffeomorphic flows, where the deformation is encoded by the generated triplane features. Leveraging a pre-existing 2D diffusion model, we produce high-quality and diverse 3D diffeomorphic flows through generated triplanes features, ensuring topological consistency with the template shape. Extensive experiments on medical image organ segmentation datasets evaluate the effectiveness of HNDF in 3D shape representation and generation. | ['Xiaohui Xie', 'Shanlin Sun', 'Kun Han'] | 2023-07-04 | null | null | null | null | ['3d-shape-generation', '3d-shape-representation'] | ['computer-vision', 'computer-vision'] | [ 2.39666952e-05 2.75140256e-01 -3.07051279e-02 -4.27341282e-01
-4.94374216e-01 -8.53889048e-01 7.40318358e-01 -2.68922076e-02
2.23550558e-01 6.32308364e-01 5.13409734e-01 -1.39296100e-01
-3.26241106e-01 -1.04258418e+00 -6.96031928e-01 -7.70224273e-01
-6.58843592e-02 7.48377323e-01 -6.20829239e-02 -2.67076313e-01
1.19010918e-01 9.57229018e-01 -8.88365924e-01 5.51700443e-02
1.09278476e+00 7.99455225e-01 -1.67492598e-01 4.52715218e-01
-3.66388738e-01 2.18336672e-01 -3.68085876e-02 -3.24934125e-01
6.68788552e-01 -5.18280983e-01 -9.82842207e-01 2.71108180e-01
6.63982511e-01 -2.97908396e-01 -3.59321505e-01 8.46273422e-01
5.09041548e-01 8.71728919e-03 9.37811196e-01 -1.06510115e+00
-1.22004223e+00 4.96040136e-02 -3.82302076e-01 8.40763897e-02
1.74128160e-01 2.18431190e-01 7.60840535e-01 -1.02001512e+00
1.11166215e+00 1.28693712e+00 8.69910300e-01 6.43315017e-01
-1.58598816e+00 -2.04463005e-01 -1.11179382e-01 -4.18470025e-01
-1.15274274e+00 -1.77129149e-01 9.99532938e-01 -6.27458692e-01
7.40113199e-01 2.91124851e-01 8.70804787e-01 9.36844707e-01
3.45592260e-01 6.39795125e-01 1.09172285e+00 -2.10681245e-01
-2.20244117e-02 -1.50914535e-01 -2.42109045e-01 1.04948473e+00
2.27099508e-01 1.89839125e-01 -2.00527504e-01 -2.92520046e-01
1.67621064e+00 1.48410043e-02 -4.61375266e-01 -9.12778139e-01
-1.53771007e+00 9.09142077e-01 7.77621031e-01 4.04268980e-01
-3.46742481e-01 8.99392143e-02 7.35234693e-02 2.64885396e-01
7.93535590e-01 5.41407764e-01 -2.10120991e-01 2.02596001e-02
-6.22729599e-01 3.54987413e-01 7.28810549e-01 9.82484341e-01
6.95797026e-01 9.35879350e-02 -3.59146446e-01 6.52847052e-01
2.15589151e-01 5.31489968e-01 1.78835630e-01 -1.06479311e+00
1.03178002e-01 8.70224655e-01 -1.84009731e-01 -1.31043160e+00
-2.47100696e-01 -3.09743226e-01 -9.99346495e-01 6.87457919e-02
4.56991553e-01 3.23485941e-01 -1.06458056e+00 1.71364152e+00
7.24375188e-01 2.82188833e-01 -1.99577987e-01 1.05111313e+00
7.36748576e-01 2.73868978e-01 -2.09088475e-01 2.81221211e-01
9.73046899e-01 -6.13227546e-01 -4.71910775e-01 4.36204761e-01
5.24850011e-01 -6.17405713e-01 8.18095565e-01 -2.65028358e-01
-1.22792602e+00 -1.71085119e-01 -7.23979354e-01 -4.09758270e-01
-1.87978804e-01 -1.83646873e-01 7.73135841e-01 4.64577287e-01
-1.27508640e+00 6.97966874e-01 -1.11571658e+00 -1.92859650e-01
8.42934728e-01 5.11689126e-01 -4.59547311e-01 -4.69012521e-02
-8.56939375e-01 7.73824096e-01 -1.59246102e-01 8.07901919e-02
-5.31084776e-01 -1.11491585e+00 -1.08624268e+00 -2.21325919e-01
-2.12489262e-01 -1.36544287e+00 7.24634588e-01 -5.37381947e-01
-1.50401998e+00 1.10500562e+00 -1.90450437e-02 -8.80445093e-02
6.64745688e-01 3.06253314e-01 2.64261395e-01 3.66027892e-01
1.57825634e-01 9.72910404e-01 8.34479749e-01 -1.15688467e+00
9.37354267e-02 -5.46485007e-01 4.75643128e-02 3.91069263e-01
-2.42579326e-01 -5.79888642e-01 -2.40926929e-02 -9.79245484e-01
4.77208078e-01 -8.80385876e-01 -2.63199985e-01 7.31158853e-01
-3.73235852e-01 -9.85401794e-02 9.45303142e-01 -4.68596041e-01
6.61154449e-01 -1.80693936e+00 5.54955244e-01 4.72846955e-01
5.33312857e-01 -6.05943128e-02 -2.45780215e-01 -4.34650993e-03
2.13693023e-01 2.48819053e-01 -6.81219161e-01 -2.32153401e-01
-2.25661732e-02 4.19348329e-01 -2.00051501e-01 4.91862535e-01
6.19224489e-01 1.47290087e+00 -1.07685733e+00 -4.42232072e-01
2.08298281e-01 9.39487159e-01 -8.79862368e-01 7.51756281e-02
-9.25239772e-02 9.59341049e-01 -9.29887772e-01 7.17360318e-01
7.87815928e-01 -2.73648798e-01 -4.94180620e-02 -4.32580173e-01
-3.51573266e-02 1.98228434e-01 -6.28680289e-01 2.08978629e+00
-3.92111599e-01 2.19130576e-01 2.15192541e-01 -1.09318590e+00
1.19566405e+00 2.19693750e-01 9.21707153e-01 -5.58293700e-01
1.39786974e-01 5.61647534e-01 -3.06682199e-01 -2.38594413e-01
1.41206965e-01 -4.24338579e-01 9.38479602e-02 5.13199329e-01
1.67200819e-01 -7.05349505e-01 -2.38724247e-01 1.13785170e-01
9.81505036e-01 5.08800149e-01 -1.01705119e-01 -6.88670695e-01
5.39271951e-01 6.81746230e-02 4.10709441e-01 2.24665001e-01
-5.79719841e-02 1.03959572e+00 3.16538215e-01 -7.98583567e-01
-1.33185935e+00 -1.43062747e+00 -5.24728358e-01 1.36341438e-01
2.39346907e-01 8.31725150e-02 -7.27606535e-01 -9.36492383e-01
3.47981870e-01 -9.80644673e-03 -8.66315484e-01 -2.01665074e-01
-9.89581347e-01 -6.33858204e-01 5.56572855e-01 5.18024862e-01
3.80116671e-01 -9.87907648e-01 -4.39087212e-01 3.33430946e-01
-2.28293225e-01 -8.51455986e-01 -1.01908088e+00 -3.62821043e-01
-1.29250526e+00 -1.06871474e+00 -1.22494519e+00 -9.30342972e-01
1.19803345e+00 2.88441092e-01 1.00083196e+00 2.28440732e-01
-6.76863134e-01 5.97755373e-01 -1.09228864e-01 1.47101238e-01
-4.94267434e-01 3.26769426e-02 -7.73157701e-02 1.06948793e-01
-2.98179351e-02 -8.35261583e-01 -7.43412733e-01 2.92424411e-01
-8.46911728e-01 1.55939087e-01 3.12845290e-01 1.19655681e+00
8.70029151e-01 -5.63123524e-01 6.48883343e-01 -7.17806041e-01
4.96279091e-01 -4.28587496e-01 -2.63297558e-01 1.67187780e-01
-3.63440156e-01 3.18354338e-01 3.16797137e-01 -3.85853767e-01
-1.03682351e+00 -2.48021837e-02 -9.78989527e-02 -8.69601786e-01
4.00551260e-02 1.43059492e-01 2.01547474e-01 -4.95090544e-01
6.67227447e-01 4.34936672e-01 6.63164675e-01 -4.13842857e-01
4.51112866e-01 8.75689164e-02 4.12375540e-01 -1.03686357e+00
9.78914738e-01 8.87414813e-01 1.81293055e-01 -6.98597968e-01
-6.62053704e-01 -1.17028445e-01 -9.63934422e-01 -3.79842333e-02
9.01870012e-01 -4.56092268e-01 -4.40283626e-01 4.81936157e-01
-1.13787448e+00 -2.87367523e-01 -7.36556053e-01 2.33031288e-01
-9.74878609e-01 3.88670117e-01 -7.26256132e-01 -2.89216012e-01
-5.57160139e-01 -1.14742815e+00 1.48120832e+00 -5.85558750e-02
-2.63403416e-01 -1.46120679e+00 1.87071219e-01 2.08983094e-01
7.44388461e-01 8.96717668e-01 1.22289145e+00 3.45482491e-03
-7.28243709e-01 1.35527223e-01 -2.67290324e-01 4.10935245e-02
7.64047921e-01 -3.28863829e-01 -5.96291721e-01 -3.62313181e-01
-4.34696637e-02 -2.58114934e-01 6.63501680e-01 5.08393407e-01
1.08623350e+00 -3.53249937e-01 -3.61698002e-01 9.71627235e-01
1.15345693e+00 -1.12159364e-01 3.15563709e-01 -3.45823318e-02
1.02292168e+00 7.89545178e-01 -2.07935814e-02 1.04153953e-01
5.38011909e-01 6.10426724e-01 2.02178955e-01 -4.63195205e-01
-6.85118675e-01 -4.78961080e-01 -1.72654092e-02 1.04749370e+00
-2.95403510e-01 3.21108341e-01 -7.47519314e-01 6.68900907e-01
-1.54328883e+00 -5.74022293e-01 -9.07133371e-02 2.01706433e+00
1.03360641e+00 -1.62148997e-01 -9.33428034e-02 -1.96148857e-01
6.36826515e-01 -3.37595642e-02 -7.59316623e-01 -1.70111433e-01
-2.31415227e-01 4.55455750e-01 3.86551805e-02 5.80837011e-01
-9.72940683e-01 8.04281056e-01 5.70676947e+00 5.40197730e-01
-1.30502307e+00 2.11003004e-03 7.14400768e-01 1.54138789e-01
-1.00636303e+00 -2.08626777e-01 -3.57087374e-01 1.38601452e-01
1.18677460e-01 -1.54147968e-01 2.15623483e-01 2.99147993e-01
1.04900755e-01 5.32052398e-01 -1.04405189e+00 7.60982156e-01
-6.61767200e-02 -1.79548216e+00 5.84320188e-01 3.33473891e-01
1.11357415e+00 -2.13251233e-01 1.62383959e-01 -1.92281276e-01
2.53491819e-01 -1.02148795e+00 7.23414361e-01 6.98820770e-01
1.25310159e+00 -4.95272785e-01 2.66266644e-01 4.01773192e-02
-1.32956553e+00 3.69326472e-01 -2.87680835e-01 4.07333910e-01
2.58122206e-01 5.96463859e-01 -8.40102017e-01 7.52532899e-01
4.67794538e-01 1.06289089e+00 -8.60317871e-02 8.40246260e-01
1.44974217e-01 1.05101101e-01 -4.62432206e-01 3.15773249e-01
3.69750053e-01 -5.65422833e-01 8.58964503e-01 9.48181391e-01
5.32198846e-01 2.99723625e-01 1.25503421e-01 1.58876252e+00
-1.72671214e-01 7.12193176e-02 -1.03014553e+00 -1.81480846e-03
2.51112133e-01 1.15340173e+00 -9.58583653e-01 -1.40635073e-01
-1.26087919e-01 8.42937291e-01 2.75947660e-01 3.24170411e-01
-4.38810170e-01 -3.29932198e-02 8.45714211e-01 3.81638974e-01
2.07542256e-01 -4.85260725e-01 -5.46328068e-01 -1.37017667e+00
1.77883431e-01 -4.62925404e-01 8.46963152e-02 -3.70232612e-01
-1.69141960e+00 6.50847614e-01 -9.85307097e-02 -1.26704168e+00
5.11633083e-02 -4.99448001e-01 -6.03722692e-01 8.29063594e-01
-1.54645753e+00 -1.53433037e+00 -3.55369806e-01 6.83454335e-01
4.26322311e-01 2.08849683e-01 9.15662527e-01 1.69919953e-01
1.21755905e-01 4.56394494e-01 -2.19926536e-01 1.31182909e-01
4.17030245e-01 -1.34400845e+00 6.29569650e-01 2.62022197e-01
3.18162367e-02 6.33931041e-01 5.21339662e-02 -8.42284024e-01
-1.60636210e+00 -1.12763762e+00 6.61849380e-01 -5.93364656e-01
3.85435909e-01 -3.24496865e-01 -1.05801666e+00 4.71181393e-01
-1.60660923e-01 5.15948951e-01 2.92946875e-01 -5.48027217e-01
-2.88026661e-01 2.49882415e-01 -1.50759375e+00 4.85668868e-01
1.66293359e+00 -4.08029765e-01 -5.21496534e-01 2.29236811e-01
6.00197196e-01 -6.70964658e-01 -1.25985742e+00 5.66206396e-01
4.97374952e-01 -6.17232978e-01 1.16425741e+00 -6.08848393e-01
3.52715880e-01 -1.96462214e-01 1.44941285e-01 -1.37486470e+00
-3.55495989e-01 -7.17118144e-01 -8.22165087e-02 7.81185806e-01
8.85042325e-02 -6.32681191e-01 8.51472497e-01 5.41823328e-01
-4.56798196e-01 -9.94906187e-01 -1.09419751e+00 -6.97181404e-01
6.94194853e-01 1.74458846e-01 7.38340318e-01 1.40697551e+00
-2.73619682e-01 -2.36487195e-01 4.24325876e-02 -3.54206830e-01
7.76402414e-01 4.22550499e-01 3.94132078e-01 -1.38313997e+00
8.50340948e-02 -6.54072881e-01 -5.36077261e-01 -1.26882219e+00
2.34523162e-01 -1.65495229e+00 -1.23781204e-01 -1.48630381e+00
-9.64515284e-03 -1.07951760e+00 9.16813537e-02 5.52219272e-01
2.10530266e-01 5.40240943e-01 -4.76697721e-02 2.55519807e-01
2.91616227e-02 1.05780065e+00 2.32387948e+00 -1.90596327e-01
-1.68939710e-01 -3.70653987e-01 -5.41899681e-01 4.42187488e-01
4.88461614e-01 -2.86758959e-01 -4.61860657e-01 -6.88721359e-01
-1.83183193e-01 1.56693697e-01 5.54368198e-01 -4.29951787e-01
8.69783908e-02 -1.10168271e-01 3.89227033e-01 -1.89870253e-01
2.72894442e-01 -5.74606955e-01 1.58913508e-01 3.68726999e-01
-1.67899057e-01 -5.98484650e-02 4.23694178e-02 4.44023550e-01
-2.18191296e-01 3.35776001e-01 8.25085878e-01 -2.81216204e-01
1.54236909e-02 8.91020000e-01 2.68024039e-02 2.34224960e-01
8.99477243e-01 -5.57887793e-01 -1.05018780e-01 -2.95177773e-02
-9.50048268e-01 7.46522844e-02 6.33153081e-01 5.00040650e-01
9.79514182e-01 -1.80014026e+00 -8.10152888e-01 6.82359576e-01
-1.44693300e-01 5.98511875e-01 1.96078658e-01 1.03852046e+00
-7.02124596e-01 2.32340559e-01 -4.73373175e-01 -9.53009367e-01
-6.28597856e-01 -2.45405696e-02 6.82489991e-01 -1.38651803e-01
-1.22631073e+00 7.54637778e-01 5.79020500e-01 -9.41438854e-01
-2.72930741e-01 -4.60018724e-01 1.33598343e-01 -3.43848228e-01
3.50066982e-02 2.02042922e-01 9.99215469e-02 -7.38325119e-01
-1.96136802e-01 1.06114662e+00 1.29098445e-01 1.48215130e-01
1.58880222e+00 3.00801732e-02 -2.24253938e-01 -3.07595786e-02
1.36620545e+00 -1.88798100e-01 -1.60528374e+00 -2.41127670e-01
-1.17881685e-01 -6.91358745e-01 -2.09831880e-04 -4.75891232e-01
-1.44722342e+00 8.26022446e-01 2.60095060e-01 5.64624183e-02
7.70077646e-01 8.72825086e-02 1.22412789e+00 2.09595151e-02
5.75562954e-01 -4.57574397e-01 2.50506282e-01 5.17500699e-01
1.22748101e+00 -9.22681808e-01 -3.21509391e-01 -6.78124964e-01
-2.70523638e-01 1.32195342e+00 5.00919044e-01 -5.66450119e-01
9.06817198e-01 2.79813468e-01 1.73842553e-02 -4.85794097e-01
-4.65842605e-01 2.61521995e-01 5.89230001e-01 7.17817307e-01
4.17505234e-01 -3.91450562e-02 -2.29618147e-01 8.44065323e-02
-2.54751354e-01 -7.61891454e-02 4.10733730e-01 8.86824429e-01
2.20877882e-02 -1.31621063e+00 -1.88393742e-01 4.61542219e-01
-1.75153852e-01 1.37155384e-01 -3.13921005e-01 6.29689097e-01
-6.12979047e-02 1.10554598e-01 2.41808549e-01 4.88340557e-02
3.29813123e-01 -1.32825091e-01 9.76785541e-01 -6.16025627e-01
-5.20008743e-01 1.33217528e-01 -2.94695795e-01 -5.53666115e-01
-4.64661181e-01 -6.87883496e-01 -1.52409601e+00 -2.92176157e-01
8.89743567e-02 -5.50759071e-03 4.50824976e-01 6.98272765e-01
6.59598708e-01 2.24526539e-01 7.74643600e-01 -1.17405152e+00
-6.34051144e-01 -5.17693698e-01 -4.19368267e-01 8.09301198e-01
4.67976063e-01 -1.10551608e+00 -3.20770800e-01 1.56971440e-01] | [8.770262718200684, -3.6708285808563232] |
1b0d0afb-1f9d-4161-9a90-7c5bffbf7edd | naive-regularizers-for-low-resource-neural | null | null | https://aclanthology.org/R19-1013 | https://aclanthology.org/R19-1013.pdf | Naive Regularizers for Low-Resource Neural Machine Translation | Neural machine translation models have little inductive bias, which can be a disadvantage in low-resource scenarios. Neural models have to be trained on large amounts of data and have been shown to perform poorly when only limited data is available. We show that using naive regularization methods, based on sentence length, punctuation and word frequencies, to penalize translations that are very different from the input sentences, consistently improves the translation quality across multiple low-resource languages. We experiment with 12 language pairs, varying the training data size between 17k to 230k sentence pairs. Our best regularizer achieves an average increase of 1.5 BLEU score and 1.0 TER score across all the language pairs. For example, we achieve a BLEU score of 26.70 on the IWSLT15 English{--}Vietnamese translation task simply by using relative differences in punctuation as a regularizer. | ['Anders S{\\o}gaard', 'Ana Valeria Gonzalez', 'Marcel Bollmann', 'Meriem Beloucif'] | 2019-09-01 | null | null | null | ranlp-2019-9 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 2.47303471e-01 -7.37603903e-02 -6.01773858e-01 -4.78188843e-01
-1.48405659e+00 -7.07465470e-01 5.93287170e-01 1.15182042e-01
-1.01064515e+00 1.30483866e+00 4.21582639e-01 -7.92612731e-01
5.36808908e-01 -4.77305174e-01 -9.83003855e-01 -4.09075767e-01
2.83935636e-01 6.00987911e-01 -3.83430213e-01 -4.10749316e-01
1.49590954e-01 4.80402000e-02 -6.62053823e-01 3.64200681e-01
1.16682327e+00 2.57949799e-01 3.02881062e-01 5.06238520e-01
-2.52185073e-02 1.44288749e-01 -6.03839815e-01 -6.70066416e-01
5.39840698e-01 -7.15930700e-01 -7.69583106e-01 -4.49315548e-01
5.48624694e-01 -1.46451011e-01 -7.35996813e-02 1.09996593e+00
7.04276681e-01 -1.36977643e-01 4.91868079e-01 -4.75237876e-01
-8.47806275e-01 1.12004364e+00 -4.76187974e-01 4.12094325e-01
1.07592754e-01 9.64367464e-02 1.21489799e+00 -1.30234814e+00
7.34275579e-01 1.07867110e+00 6.55866683e-01 7.40169942e-01
-1.54095244e+00 -6.52481496e-01 -1.67433783e-01 -1.39880627e-01
-1.21225941e+00 -6.23998284e-01 2.90549189e-01 -4.20319773e-02
1.63003457e+00 2.37447530e-01 1.98304638e-01 1.17431605e+00
5.76903701e-01 4.97535884e-01 1.23975575e+00 -7.95052171e-01
-1.20912433e-01 2.08381340e-01 -8.91451612e-02 4.20286477e-01
2.99647570e-01 -7.09033757e-02 -6.10881925e-01 -6.51518628e-02
4.71810490e-01 -4.74546403e-01 -9.66388881e-02 4.24498677e-01
-1.55031705e+00 9.62120473e-01 2.13153079e-01 4.69618678e-01
-3.07548463e-01 2.07338318e-01 5.89251518e-01 8.44475806e-01
7.13082910e-01 8.81220818e-01 -8.18512321e-01 -4.25102055e-01
-9.50712740e-01 5.91932656e-03 6.38595521e-01 9.62417245e-01
7.22246826e-01 2.31354520e-01 -2.37447605e-01 1.14910460e+00
-1.30272493e-01 8.47075403e-01 7.58218825e-01 -6.60839081e-01
1.09990096e+00 6.83564469e-02 1.24838976e-02 -5.95302939e-01
7.24641420e-03 -4.25309509e-01 -8.12729299e-01 -3.81283581e-01
4.72842991e-01 -5.79128206e-01 -1.03513503e+00 2.04784822e+00
-3.27548563e-01 -4.52678353e-01 3.77921969e-01 7.50644267e-01
3.85984123e-01 9.21031713e-01 -9.30989832e-02 -4.82142210e-01
9.69961703e-01 -1.03538942e+00 -6.70047104e-01 -6.46490633e-01
1.10768628e+00 -1.25961947e+00 1.69600689e+00 4.26183715e-02
-1.41559708e+00 -2.98777699e-01 -9.66628850e-01 -1.27549082e-01
-2.98339576e-01 2.86995173e-01 3.65665257e-01 4.29678977e-01
-1.11615002e+00 7.44530380e-01 -7.13173568e-01 -4.76517439e-01
-8.42905045e-03 5.16093671e-01 -3.22580367e-01 -5.37266694e-02
-1.50738597e+00 1.30412841e+00 3.01824212e-01 8.28395486e-02
-5.14410496e-01 -5.30604959e-01 -6.93167448e-01 -1.38790697e-01
-1.93436787e-01 -5.13378322e-01 1.20860338e+00 -1.36955750e+00
-1.52579045e+00 9.46710646e-01 -3.14070374e-01 -6.00790679e-01
5.38746297e-01 -4.02945161e-01 -2.70963937e-01 -3.74215454e-01
2.27955252e-01 5.65242946e-01 3.73216093e-01 -7.84931123e-01
-4.71567839e-01 -6.97200671e-02 -3.38347524e-01 4.42138791e-01
-4.12860185e-01 5.20014763e-01 -3.76388252e-01 -8.14487696e-01
-9.41090807e-02 -1.03926003e+00 -4.26231205e-01 -7.17833042e-01
-1.87337160e-01 1.01526622e-02 3.69893834e-02 -1.01676929e+00
1.29921675e+00 -1.72852898e+00 1.88608527e-01 4.11720686e-02
-5.55433512e-01 3.09648335e-01 -4.82847929e-01 2.87533581e-01
1.35114774e-01 4.42145705e-01 -3.08410615e-01 -1.97036281e-01
-2.67477870e-01 2.56031215e-01 -1.62511662e-01 2.79498845e-01
5.18736422e-01 9.26862597e-01 -9.23096955e-01 -2.85534233e-01
-3.27730179e-01 1.29779115e-01 -4.29059386e-01 3.73582430e-02
1.00082159e-02 1.53279200e-01 -3.59955877e-02 4.01031166e-01
4.54205155e-01 9.25778002e-02 3.37693334e-01 3.60391498e-01
-7.00512752e-02 9.94696736e-01 -4.77862775e-01 1.76374781e+00
-8.44884396e-01 1.03063560e+00 -1.71308130e-01 -5.78441262e-01
8.97131860e-01 4.49809700e-01 -2.79399287e-03 -8.53880048e-01
-2.43502599e-03 7.80502737e-01 5.11831403e-01 -1.58045083e-01
5.42239130e-01 -3.95086616e-01 -1.76033616e-01 5.29212058e-01
5.24661690e-02 -1.59383416e-01 4.31567758e-01 -8.53579566e-02
1.04573059e+00 -9.92271304e-02 1.40567422e-01 -5.52716851e-01
2.37486228e-01 2.20199883e-01 9.58618104e-01 6.90536320e-01
7.05999322e-03 7.56805182e-01 4.20820266e-01 -3.33842635e-01
-1.74936962e+00 -9.76222157e-01 -1.16816528e-01 1.06577396e+00
-4.42116886e-01 -4.43786085e-01 -9.10898685e-01 -6.26353621e-01
-2.56581843e-01 8.57989848e-01 -2.60122120e-01 -1.87104389e-01
-1.23690081e+00 -1.25079489e+00 9.12486672e-01 4.74238157e-01
3.00683260e-01 -9.72468555e-01 -1.43292192e-02 3.90694290e-01
-4.51647639e-01 -1.00448573e+00 -7.95336068e-01 5.35626709e-01
-1.08062434e+00 -1.70808107e-01 -8.58367980e-01 -1.03364456e+00
7.32185364e-01 -1.45207465e-01 1.51093578e+00 1.74787380e-02
2.98540235e-01 -5.55401266e-01 -1.71618104e-01 -3.15717876e-01
-6.94923580e-01 5.34492254e-01 5.34858286e-01 -5.53836286e-01
5.71302891e-01 -4.08960968e-01 -1.67680949e-01 1.35116816e-01
-4.40193266e-01 1.15497217e-01 8.77777755e-01 1.25051808e+00
4.51313704e-01 -6.91599727e-01 6.73654795e-01 -1.00007939e+00
9.10380542e-01 -3.27756077e-01 -2.88754106e-01 3.34305495e-01
-9.71015334e-01 1.78051203e-01 9.75939691e-01 -5.59562325e-01
-8.18088114e-01 -4.73157130e-02 -8.46374780e-02 1.90462247e-01
1.83238164e-01 6.94329500e-01 7.29222521e-02 2.40754440e-01
9.83577192e-01 2.50829697e-01 -1.56178415e-01 -4.17623848e-01
1.55980900e-01 9.67898667e-01 1.83425650e-01 -6.75181031e-01
7.05062509e-01 -2.72848010e-01 -3.98084611e-01 -5.85443020e-01
-5.80594361e-01 -1.24239773e-01 -6.49191439e-01 3.22899133e-01
5.88273346e-01 -1.15894771e+00 6.30929917e-02 2.36061200e-01
-1.35088789e+00 -5.89958966e-01 -2.65934709e-02 9.45424199e-01
-4.33738202e-01 1.58814400e-01 -1.00894594e+00 -3.27676654e-01
-7.66791344e-01 -1.11202955e+00 5.51631570e-01 -7.21400082e-02
-4.35977042e-01 -8.31068099e-01 9.05910209e-02 2.00657398e-01
4.79762614e-01 -2.03300819e-01 1.14487815e+00 -7.03953087e-01
-1.50125667e-01 -3.24173309e-02 -8.77311304e-02 6.70364618e-01
3.26067477e-01 -1.69496179e-01 -4.87536073e-01 -4.61358130e-01
-2.78903931e-01 -5.07656753e-01 6.65490031e-01 2.99370348e-01
5.39955974e-01 -5.59153140e-01 5.01670949e-02 3.65735412e-01
1.49271345e+00 3.97140859e-03 5.06068528e-01 3.20013195e-01
6.03633523e-01 3.38545352e-01 5.44483006e-01 -4.38293554e-02
4.74458411e-02 5.99899352e-01 -3.13451767e-01 -2.01910228e-01
-1.30838081e-01 -2.44244531e-01 7.86901832e-01 1.52613401e+00
-1.92140564e-01 -2.22640738e-01 -1.16733158e+00 7.01770186e-01
-1.67086661e+00 -6.72913849e-01 -1.87001675e-01 2.31578112e+00
1.56579149e+00 4.45740879e-01 -2.21743435e-01 -3.73689264e-01
7.05327451e-01 -2.37915721e-02 -2.54476249e-01 -1.07128680e+00
-3.59371871e-01 3.34104508e-01 9.64694202e-01 9.36958253e-01
-5.93434334e-01 1.50180531e+00 7.16767597e+00 8.07354033e-01
-1.29745650e+00 1.18472934e-01 6.93437457e-01 -3.15044105e-01
-4.99990940e-01 1.31759653e-02 -8.74197245e-01 5.93653738e-01
1.51430941e+00 -4.09818023e-01 7.43736088e-01 4.87817675e-01
5.38245976e-01 9.30633023e-03 -1.18704152e+00 7.60978758e-01
1.64814964e-01 -1.20727897e+00 1.52187988e-01 -5.27185872e-02
1.08144140e+00 5.50739884e-01 2.92662755e-02 3.51401448e-01
4.86826956e-01 -1.19339037e+00 5.68147421e-01 9.21717733e-02
9.08964217e-01 -8.35017383e-01 9.14252698e-01 4.41116512e-01
-6.11841321e-01 3.83100003e-01 -6.80706084e-01 -2.63126403e-01
4.16023396e-02 5.37137628e-01 -1.06341505e+00 1.83546916e-01
6.01440072e-01 4.81475145e-01 -4.38316315e-01 6.17971361e-01
-3.47142488e-01 1.02575064e+00 -4.43937987e-01 -3.04672807e-01
5.14976263e-01 -2.97762930e-01 3.43233734e-01 1.62224388e+00
3.18130165e-01 -1.93956688e-01 -1.24708682e-01 4.11512166e-01
-5.58962405e-01 5.26723802e-01 -7.09069133e-01 -8.50565881e-02
4.11867768e-01 7.59121478e-01 -3.90686244e-01 -3.80048722e-01
-4.07789499e-01 1.24470329e+00 7.02526331e-01 4.94810373e-01
-6.79591715e-01 -5.64246058e-01 8.25076282e-01 -7.23430663e-02
-1.51720177e-02 -4.50211257e-01 -7.23006248e-01 -1.32180417e+00
3.06419939e-01 -1.19245136e+00 -1.14011154e-01 -4.93794858e-01
-1.18073463e+00 7.31885791e-01 -3.37448210e-01 -9.39221203e-01
-4.75388348e-01 -5.70971668e-01 -5.14671326e-01 1.44074500e+00
-1.36907649e+00 -8.79917741e-01 5.70807934e-01 1.00048549e-01
8.74133110e-01 -3.53516310e-01 1.03273833e+00 5.41098475e-01
-5.29772043e-01 1.04296100e+00 6.61619306e-01 3.48825872e-01
1.02612579e+00 -1.18311703e+00 8.93648267e-01 1.01759064e+00
3.30042303e-01 9.88972127e-01 8.13775957e-01 -6.59532785e-01
-1.24296403e+00 -1.04238045e+00 1.95190465e+00 -5.35209179e-01
5.76376081e-01 -5.30090570e-01 -8.05789709e-01 7.08788216e-01
5.93053520e-01 -3.90963584e-01 7.26867437e-01 3.72933328e-01
-2.95044482e-01 -6.93174452e-03 -9.21928167e-01 1.00653613e+00
9.14189041e-01 -6.73397422e-01 -6.41052186e-01 6.38054430e-01
7.48401105e-01 -3.37547123e-01 -7.60959208e-01 3.71513367e-01
4.13111448e-01 -2.34878838e-01 6.46052182e-01 -1.15234661e+00
6.53534770e-01 -1.43165723e-03 -2.65936196e-01 -1.70522392e+00
-4.00299966e-01 -7.62717366e-01 4.80670750e-01 1.18791413e+00
1.37480116e+00 -4.04365510e-01 6.71680093e-01 4.88086015e-01
-6.33268058e-02 -8.44544530e-01 -9.61042643e-01 -1.02844536e+00
8.15982223e-01 -2.15217739e-01 2.93336660e-01 9.11476970e-01
1.93822727e-01 7.51096964e-01 -5.51060438e-01 -2.97991574e-01
1.53989315e-01 -2.35466808e-01 4.53997284e-01 -6.62705123e-01
-2.18032390e-01 -3.68030161e-01 8.80187601e-02 -8.92712653e-01
3.75118881e-01 -1.16623008e+00 2.83435285e-01 -1.43092477e+00
3.12066406e-01 -3.07694376e-01 -2.76718020e-01 5.38037002e-01
-4.17994231e-01 5.90164244e-01 2.71701068e-02 2.35324308e-01
3.68468203e-02 2.87706405e-01 1.01939201e+00 -1.81889534e-01
-1.87116876e-01 -2.30162814e-01 -5.88180482e-01 5.16468763e-01
1.19492888e+00 -7.17547834e-01 3.28654200e-02 -1.14042628e+00
5.01098096e-01 -2.06773520e-01 -3.94105524e-01 -5.03842652e-01
-1.25333101e-01 -3.88810128e-01 3.44849527e-01 -2.01816469e-01
1.28601283e-01 -4.19167399e-01 -1.98047429e-01 5.29254377e-01
-7.09040642e-01 7.02953935e-01 2.66809613e-01 1.21924810e-01
-1.88217327e-01 -3.06562155e-01 8.04213941e-01 -3.76983732e-01
-1.37029186e-01 -7.77520686e-02 -6.10709369e-01 3.78141612e-01
4.48576003e-01 8.17773864e-02 4.20427136e-02 -1.91338733e-01
-2.55507201e-01 3.53521071e-02 6.21813655e-01 4.67100948e-01
2.74368316e-01 -1.57806432e+00 -1.24981964e+00 1.20031886e-01
-2.07332112e-02 -4.79123265e-01 -5.02284408e-01 5.92323840e-01
-6.69180155e-01 5.04346371e-01 -2.29098707e-01 -2.66690284e-01
-1.30190468e+00 5.16342074e-02 1.39841765e-01 -4.29846734e-01
-2.90343553e-01 8.72799695e-01 -5.29593468e-01 -6.96615577e-01
9.98312607e-02 -4.12553489e-01 3.15686345e-01 -1.59065947e-01
1.91327482e-01 1.93574801e-01 3.82917941e-01 -5.85480988e-01
-3.85404617e-01 4.73108083e-01 -4.63046849e-01 -5.69749475e-01
1.16784298e+00 -3.70466299e-02 -1.42445982e-01 4.25778270e-01
1.41146016e+00 2.17744589e-01 -8.06277454e-01 -3.64099115e-01
2.37473279e-01 -4.11727130e-01 -1.56585589e-01 -9.37092245e-01
-5.78949630e-01 8.78158331e-01 3.35383266e-01 -3.09295416e-01
8.65288854e-01 -3.98030341e-01 1.00213385e+00 8.20533574e-01
3.11773807e-01 -1.54970539e+00 -2.86960930e-01 1.11189115e+00
7.90609896e-01 -1.40694201e+00 -1.15066938e-01 -9.23114549e-03
-6.09625578e-01 8.91496420e-01 4.34076846e-01 -1.80085525e-01
1.41148984e-01 3.15824360e-01 5.08675694e-01 5.25766850e-01
-8.23781908e-01 1.28951877e-01 6.78863823e-02 2.46404693e-01
1.08421898e+00 2.95530409e-01 -1.17917919e+00 1.74038142e-01
-4.87333089e-01 -4.07329351e-01 5.38082838e-01 6.75860763e-01
-3.81905377e-01 -1.52005613e+00 -2.13504031e-01 5.17236352e-01
-1.02885938e+00 -8.58413160e-01 -5.69057703e-01 3.83522451e-01
-1.15957230e-01 9.00612116e-01 3.71250659e-02 -2.36056313e-01
1.69783235e-01 3.08284432e-01 5.03403306e-01 -7.66276479e-01
-1.00852489e+00 -7.64465034e-02 4.20575410e-01 -2.56641984e-01
-1.18523598e-01 -7.04089105e-01 -1.15308630e+00 -4.75811809e-01
-2.80250579e-01 2.90436685e-01 8.47676575e-01 7.65057564e-01
3.40974092e-01 5.02516376e-03 7.21790910e-01 -4.03153300e-01
-9.25992846e-01 -1.30501342e+00 -8.93424898e-02 2.66165465e-01
2.95198381e-01 9.72800925e-02 -3.80344898e-01 1.83863193e-01] | [11.561464309692383, 10.255462646484375] |
80b253ad-5bc4-47a1-b8c7-9da4f8f423b9 | unsupervised-multiple-object-tracking-with-a | 2202.09315 | null | https://arxiv.org/abs/2202.09315v2 | https://arxiv.org/pdf/2202.09315v2.pdf | Unsupervised Multiple-Object Tracking with a Dynamical Variational Autoencoder | In this paper, we present an unsupervised probabilistic model and associated estimation algorithm for multi-object tracking (MOT) based on a dynamical variational autoencoder (DVAE), called DVAE-UMOT. The DVAE is a latent-variable deep generative model that can be seen as an extension of the variational autoencoder for the modeling of temporal sequences. It is included in DVAE-UMOT to model the objects' dynamics, after being pre-trained on an unlabeled synthetic dataset of single-object trajectories. Then the distributions and parameters of DVAE-UMOT are estimated on each multi-object sequence to track using the principles of variational inference: Definition of an approximate posterior distribution of the latent variables and maximization of the corresponding evidence lower bound of the data likehood function. DVAE-UMOT is shown experimentally to compete well with and even surpass the performance of two state-of-the-art probabilistic MOT models. Code and data are publicly available. | ['Xavier Alameda-Pineda', 'Laurent Girin', 'Xiaoyu Lin'] | 2022-02-18 | null | null | null | null | ['temporal-sequences'] | ['reasoning'] | [-3.82428318e-01 8.91941041e-02 3.11443973e-02 -1.17953405e-01
-4.68335658e-01 -3.51776809e-01 9.55410302e-01 -4.18094814e-01
-2.84149736e-01 6.45905674e-01 -6.05658032e-02 2.15461269e-01
-2.34914988e-01 -6.24449849e-01 -9.53520119e-01 -9.37833846e-01
9.09628421e-02 1.22537184e+00 5.05555868e-01 3.54151458e-01
-2.83092946e-01 4.10370409e-01 -1.76751244e+00 3.49776559e-02
3.47085983e-01 6.61267757e-01 4.97557819e-01 9.39950049e-01
-8.27936009e-02 8.07089806e-01 -4.52812731e-01 -4.46796149e-01
-1.44435138e-01 -4.35082048e-01 -4.89229828e-01 2.01437324e-01
7.45998025e-02 -3.72525543e-01 -6.79965079e-01 1.17167616e+00
7.24687725e-02 3.91634405e-01 1.38566470e+00 -1.60745704e+00
-6.76014602e-01 5.55137038e-01 -9.28031653e-02 3.84180784e-01
-2.30218425e-01 8.72173309e-02 8.76579881e-01 -6.64383650e-01
9.68926430e-01 1.44457734e+00 7.07586050e-01 6.80807710e-01
-1.38261628e+00 -8.68149251e-02 -2.33089030e-01 2.11495250e-01
-1.44927454e+00 -5.03645912e-02 5.13704419e-01 -9.84204173e-01
6.33433342e-01 -8.91528055e-02 7.97297597e-01 1.48184955e+00
5.37981570e-01 1.13119149e+00 5.67312479e-01 -2.01103359e-01
4.78063345e-01 2.33496919e-01 -2.76574120e-02 6.89301252e-01
5.59762530e-02 5.05709767e-01 -3.63375425e-01 -3.70503366e-01
8.82174194e-01 4.81697097e-02 2.11228356e-01 -6.58006430e-01
-9.87319589e-01 1.02648139e+00 9.21419486e-02 3.53286922e-01
-6.82494164e-01 6.15579724e-01 1.03129618e-01 -2.53785074e-01
3.24357748e-01 -4.12818819e-01 -9.57978144e-02 -1.85118005e-01
-1.00213110e+00 6.15216732e-01 7.22101092e-01 9.95256960e-01
6.57708526e-01 3.65031838e-01 -3.75598431e-01 4.33911473e-01
1.01136410e+00 7.09757328e-01 3.21998358e-01 -1.21362102e+00
-3.07358474e-01 2.56840527e-01 4.09156173e-01 -4.39341486e-01
2.92572901e-02 -2.23699987e-01 -4.23949540e-01 3.19356859e-01
4.47717220e-01 -2.01087296e-01 -1.04195726e+00 1.85611260e+00
5.11838913e-01 5.66455007e-01 6.97466210e-02 7.21840084e-01
6.18227303e-01 9.65258181e-01 2.95870006e-01 -1.91891104e-01
1.14359713e+00 -6.23263359e-01 -9.50461209e-01 2.48578250e-01
4.97109517e-02 -3.34835082e-01 1.95588484e-01 1.96534798e-01
-8.60268712e-01 -8.71682048e-01 -7.95849323e-01 2.92875856e-01
-2.69160450e-01 3.99737060e-02 4.37233806e-01 5.85388243e-01
-1.04618943e+00 8.02525699e-01 -1.57767200e+00 -3.15494269e-01
3.42702061e-01 1.85853556e-01 2.47902144e-02 2.17165530e-01
-9.00327444e-01 8.66630256e-01 6.66800141e-01 -1.65836886e-02
-1.78586090e+00 -5.16465902e-01 -8.56298745e-01 -6.70020506e-02
-9.63840187e-02 -9.25652266e-01 1.37372780e+00 -6.19428277e-01
-1.47508132e+00 7.00444877e-01 -2.66493082e-01 -7.20056713e-01
5.95099390e-01 -3.08958054e-01 -3.07159156e-01 1.09667577e-01
-2.43148319e-02 9.38047767e-01 1.25505435e+00 -1.43526173e+00
-3.78699750e-01 -1.91088654e-02 -4.82646912e-01 -2.37734392e-01
4.51963805e-02 -1.04288831e-01 -4.70198780e-01 -5.40762901e-01
-3.04690957e-01 -1.14746308e+00 -1.76749483e-01 -3.06561962e-02
-2.32579008e-01 -5.15366137e-01 1.25708568e+00 -5.57096779e-01
8.71806800e-01 -1.97575545e+00 5.80115020e-01 -2.70797938e-01
2.68867582e-01 1.52059197e-01 1.70264930e-01 3.90156537e-01
2.83272296e-01 -3.61893743e-01 -4.00642782e-01 -5.25665104e-01
3.64063561e-01 7.29577422e-01 -2.22730309e-01 6.89793348e-01
2.05513656e-01 9.82968986e-01 -1.03138077e+00 -6.77129567e-01
5.81806242e-01 7.20063269e-01 -4.80652332e-01 3.37417960e-01
-9.88258004e-01 4.87673044e-01 -5.87010264e-01 2.09349602e-01
5.89976311e-01 -4.11534369e-01 2.68557787e-01 6.74438616e-03
-1.54733717e-01 -4.42057371e-01 -1.13713181e+00 1.44053555e+00
2.56847143e-01 9.22298312e-01 -2.86576182e-01 -6.76456034e-01
6.38587475e-01 5.28303742e-01 8.80504131e-01 1.31351054e-01
5.22629023e-01 -8.96186084e-02 -1.22865736e-01 -5.55768073e-01
5.81074715e-01 -3.08150709e-01 1.04450516e-01 4.39621627e-01
8.30438256e-01 9.95134115e-02 4.03569877e-01 3.88426155e-01
7.36266851e-01 7.26387084e-01 -1.01070749e-02 -3.10408086e-01
3.38328958e-01 -9.10179466e-02 4.93462920e-01 7.92066991e-01
-3.54802638e-01 3.79413038e-01 2.43703291e-01 -2.02313930e-01
-1.37685788e+00 -1.61551797e+00 -2.86850750e-01 7.16497123e-01
3.86326686e-02 -3.17036718e-01 -9.33943748e-01 -3.23306918e-01
1.70684397e-01 9.88024175e-01 -7.64630377e-01 2.60174442e-02
-2.25119278e-01 -8.72993231e-01 4.26470160e-01 6.04742348e-01
-3.27705964e-02 -1.23571837e+00 -7.16685295e-01 4.79410261e-01
-1.51271895e-01 -1.29424191e+00 -3.86539027e-02 -4.03054543e-02
-7.11699128e-01 -9.09946144e-01 -7.23056912e-01 -3.26634139e-01
2.27959424e-01 -4.34612632e-01 1.07098317e+00 -4.12949950e-01
-2.07938269e-01 7.32831240e-01 -9.13809761e-02 -7.40566373e-01
-9.12772298e-01 -4.12535608e-01 3.61921608e-01 2.08772019e-01
5.04810631e-01 -6.23771966e-01 -2.49544069e-01 2.39797682e-01
-9.41310227e-01 -2.72814214e-01 3.16791594e-01 7.31121480e-01
8.96692991e-01 1.05438223e-02 -4.40809168e-02 -2.45192826e-01
7.47182965e-02 -6.83125794e-01 -9.91591930e-01 5.01162000e-02
-2.25508586e-01 3.92610043e-01 1.02043577e-01 -9.05672789e-01
-1.08190966e+00 1.59922808e-01 -1.53231785e-01 -1.36344743e+00
-2.37994865e-01 1.46392554e-01 1.74690396e-01 3.02306771e-01
3.32877278e-01 4.32096690e-01 -1.35759741e-01 -5.32339513e-01
5.06587386e-01 3.96958083e-01 6.77537262e-01 -4.73140895e-01
7.33410299e-01 8.55690062e-01 4.16827016e-02 -9.05653656e-01
-7.61469662e-01 -4.02756959e-01 -7.28129864e-01 -7.24826992e-01
1.45957541e+00 -8.06370497e-01 -9.61054027e-01 5.37191629e-01
-1.21758008e+00 -4.60909635e-01 -4.71973389e-01 8.03283393e-01
-9.69455421e-01 2.23666340e-01 -7.36575425e-01 -1.36344552e+00
1.22061171e-01 -9.94960964e-01 1.19585776e+00 4.23449159e-01
-7.67355561e-02 -1.33755708e+00 6.48403585e-01 1.17584720e-01
1.40953669e-03 4.43677455e-01 5.36846578e-01 -6.10042632e-01
-9.19748962e-01 -1.20937482e-01 2.50245839e-01 3.11548620e-01
-4.31820035e-01 5.18438220e-01 -1.01543474e+00 -1.80373728e-01
8.53147078e-03 -1.89710408e-01 8.92256618e-01 1.15174437e+00
7.16772199e-01 -3.43921920e-03 -6.89122736e-01 3.16107035e-01
1.52894080e+00 1.79961815e-01 3.92454296e-01 -5.23832217e-02
5.17074049e-01 2.02222839e-01 4.34015036e-01 5.91560662e-01
2.74276853e-01 6.55274212e-01 6.93295896e-01 7.68041909e-01
8.43871906e-02 -4.49982703e-01 5.57364583e-01 9.16394770e-01
-8.20755810e-02 -6.79010689e-01 -8.62253964e-01 8.67065132e-01
-2.12282181e+00 -1.31421983e+00 -4.15280282e-01 1.74928606e+00
1.23813748e-01 7.45297894e-02 4.96719211e-01 -3.74198020e-01
9.16678846e-01 2.35254884e-01 -5.77315688e-01 5.93627580e-02
1.58705592e-01 -7.66259581e-02 1.12189978e-01 5.12881517e-01
-1.18967187e+00 8.73600662e-01 6.83996964e+00 6.68485940e-01
-5.03627181e-01 5.38174272e-01 -1.80635303e-01 1.08024523e-01
-1.30387396e-01 -2.41127703e-03 -1.12816525e+00 7.27047682e-01
1.48650014e+00 3.43449637e-02 1.70580760e-01 9.61299956e-01
-5.75136468e-02 -3.45465355e-02 -1.07154047e+00 7.43174314e-01
-6.49298951e-02 -1.34285545e+00 8.96544196e-04 3.02560717e-01
7.78336167e-01 4.18508887e-01 7.55984411e-02 5.66371202e-01
8.45383763e-01 -6.78387046e-01 1.07915068e+00 1.11515689e+00
2.34541550e-01 -5.37891924e-01 5.89781225e-01 7.04075456e-01
-1.20517242e+00 9.30492207e-02 -5.77222645e-01 3.56572479e-01
5.58372676e-01 2.94333130e-01 -5.63438058e-01 5.39162755e-01
7.24456012e-01 6.61039710e-01 -9.79560912e-02 9.97195363e-01
-1.06785506e-01 7.58820176e-01 -5.57548821e-01 -2.72224426e-01
3.18954349e-01 -2.77434528e-01 1.19381857e+00 1.24754882e+00
3.30814511e-01 -1.10436350e-01 1.74251825e-01 1.43576241e+00
1.37565002e-01 -5.86196244e-01 -4.24175084e-01 -5.76381803e-01
4.18168157e-01 1.01270986e+00 -5.79147100e-01 -5.15165329e-01
-1.53878648e-02 8.22893679e-01 1.04945965e-01 3.21595043e-01
-1.14458847e+00 3.88140947e-01 6.67648137e-01 -2.83875257e-01
9.60653007e-01 -1.25402302e-01 4.33751434e-01 -1.08367538e+00
-4.02549982e-01 -1.99512869e-01 3.87350261e-01 -1.03763139e+00
-1.31209242e+00 6.34599268e-01 4.56762671e-01 -1.29430950e+00
-8.35820436e-01 -7.46468425e-01 -4.27207947e-01 7.28930056e-01
-8.27926993e-01 -1.12339246e+00 6.04938418e-02 5.21027446e-01
4.82095212e-01 -1.77349731e-01 7.75865495e-01 6.38525113e-02
-4.28923070e-01 4.07804735e-03 5.38748562e-01 1.40008777e-01
-1.27250358e-01 -1.36978924e+00 2.80523270e-01 8.15120220e-01
5.41717172e-01 3.24491173e-01 1.32850599e+00 -8.16074193e-01
-1.15901315e+00 -1.20270622e+00 4.78036046e-01 -1.09175110e+00
7.88568616e-01 -3.28719288e-01 -8.46905172e-01 1.21746683e+00
3.20220709e-01 5.07398993e-02 4.54318047e-01 -2.57139266e-01
1.61033422e-01 5.18960953e-01 -8.89794230e-01 3.15560073e-01
7.09194124e-01 -5.41869640e-01 -8.15424740e-01 3.70828807e-01
4.95145380e-01 -4.63101119e-01 -1.16561329e+00 1.42713010e-01
6.85792089e-01 -8.16130757e-01 1.01376569e+00 -7.67245352e-01
2.88982898e-01 -4.42354620e-01 -2.88700074e-01 -1.02351952e+00
-5.96624196e-01 -6.51451468e-01 -1.01905227e+00 1.04477811e+00
-1.24577703e-02 2.92399321e-02 7.82379746e-01 5.56910411e-02
4.85556480e-03 -3.74552935e-01 -1.02268744e+00 -9.40890372e-01
-7.19746277e-02 -6.54656053e-01 1.09672241e-01 6.50470614e-01
-6.54157341e-01 1.99519768e-01 -5.12804985e-01 4.28932160e-01
1.31125104e+00 1.44431926e-02 7.07960963e-01 -1.50955606e+00
-5.72473466e-01 -1.64605826e-01 -6.46634042e-01 -1.01623213e+00
3.73575836e-01 -6.67253673e-01 2.82283098e-01 -1.35651374e+00
4.45097029e-01 2.06284866e-01 -2.51610875e-01 -1.29978463e-01
-4.25280258e-02 -1.51451170e-01 2.27820158e-01 2.90365964e-01
-7.56884515e-01 1.10233843e+00 9.22230124e-01 -1.15315810e-01
-4.88367081e-02 1.62032828e-01 2.43599027e-01 7.85656989e-01
2.05587402e-01 -1.04097784e+00 -1.97460487e-01 -4.83428985e-02
-2.54808784e-01 3.71045649e-01 7.23011374e-01 -1.06466317e+00
1.32421508e-01 -7.43063241e-02 5.08976936e-01 -1.30340159e+00
7.61198401e-01 -6.19998634e-01 7.53679276e-01 5.01319528e-01
-7.40237981e-02 -1.07235081e-01 1.74963057e-01 1.16865158e+00
-1.00988653e-02 -4.06200677e-01 8.45995963e-01 -1.27308518e-02
-7.75332391e-01 4.18829381e-01 -9.35317695e-01 -1.90320924e-01
1.02680349e+00 -8.44452009e-02 1.51878774e-01 -3.03795606e-01
-1.17194295e+00 5.08345403e-02 1.66096121e-01 3.65397513e-01
4.48842019e-01 -1.49732184e+00 -6.84199810e-01 -1.52176842e-01
-1.35088220e-01 -2.52827227e-01 5.98612368e-01 6.44527733e-01
-2.70201027e-01 5.48280060e-01 -2.54506111e-01 -1.20540130e+00
-8.85742784e-01 8.32481623e-01 5.20633280e-01 -4.72988307e-01
-7.18181551e-01 8.54524553e-01 1.17666595e-01 -4.30541188e-01
7.56763145e-02 -1.46158934e-01 -1.60591975e-01 -1.55858085e-01
3.19359303e-01 3.60448897e-01 -6.32949412e-01 -9.97532785e-01
-2.87217885e-01 1.43791601e-01 2.56588608e-01 -6.36024475e-01
1.35978901e+00 -4.44891118e-03 2.49030456e-01 8.90641212e-01
8.76727700e-01 -5.26680827e-01 -1.84768188e+00 -1.68580100e-01
-6.43002316e-02 -3.85357998e-03 9.88813341e-02 -3.27300519e-01
-6.88580453e-01 8.48336577e-01 9.14878368e-01 2.29228169e-01
3.89530092e-01 3.19178998e-01 6.06473625e-01 6.32922277e-02
2.05848038e-01 -9.01781321e-01 1.05306067e-01 4.23643261e-01
5.34780562e-01 -9.20213997e-01 -1.88931823e-01 5.97449057e-02
-6.67386115e-01 8.90742958e-01 2.94403106e-01 -3.79305691e-01
1.06651795e+00 2.91108400e-01 -2.55002856e-01 -4.20192420e-01
-9.15384233e-01 -3.69732797e-01 5.42621851e-01 7.33254611e-01
3.16767320e-02 1.73526257e-01 2.12615460e-01 3.18666965e-01
9.98825766e-03 2.15996206e-01 9.04789940e-02 7.81351388e-01
-3.76805037e-01 -8.14191878e-01 -4.49767530e-01 1.23826951e-01
-4.40855622e-01 3.24818313e-01 -7.02144532e-03 9.98952985e-01
2.38444358e-01 6.23242795e-01 3.26544106e-01 -2.65700340e-01
-1.43438846e-01 2.57282406e-01 8.00413787e-01 -1.79807320e-01
-6.53607771e-02 3.74442816e-01 -3.18646759e-01 -2.22391710e-01
-7.12106884e-01 -1.15217125e+00 -1.06515467e+00 -1.42665341e-01
-4.03392166e-01 1.81526452e-01 5.78329742e-01 1.24909818e+00
-3.91718978e-03 7.82258093e-01 -1.54997120e-02 -1.11259615e+00
-6.24323010e-01 -1.26109219e+00 -6.74941480e-01 2.97441304e-01
3.02822143e-01 -1.08838415e+00 -1.72763765e-01 5.08181870e-01] | [7.360686302185059, -0.05284833535552025] |
dc3d2dab-2bf0-4908-90e6-21f0a1ee9846 | an-effective-discourse-parser-that-uses-rich | null | null | https://aclanthology.org/N09-1064/ | https://aclanthology.org/N09-1064.pdf | An effective Discourse Parser that uses Rich Linguistic Information | This paper presents a first-order logic learning approach to determine rhetorical relations between discourse segments. Beyond linguistic cues and lexical information, our approach exploits compositional semantics and segment discourse structure data. We report a statistically significant improvement in classifying relations over attribute-value learning paradigms such as Decision Trees, RIPPER and Naive Bayes. For discourse parsing, our modified shift-reduce parsing model that uses our relation classifier significantly
outperforms a right-branching majority-class baseline. | ['Barbara Di Eugenio', 'Rajen Subba'] | 2009-05-31 | null | null | null | proceedings-of-human-language-technologies | ['discourse-parsing'] | ['natural-language-processing'] | [ 5.00743687e-01 1.03167593e+00 -1.19805777e+00 -7.06124663e-01
-1.13941991e+00 -9.40444589e-01 1.02308881e+00 9.73042011e-01
-1.70625627e-01 1.05473936e+00 8.73224914e-01 -1.15824378e+00
-3.43378335e-02 -1.09407270e+00 -5.68119109e-01 -4.42684330e-02
-1.98583037e-01 4.79680270e-01 6.73829973e-01 -5.74288428e-01
6.28678858e-01 7.44116306e-02 -1.04907703e+00 1.22899425e+00
4.66884077e-01 9.14772272e-01 -4.86994356e-01 8.28131139e-01
-7.68870711e-01 2.32323861e+00 -6.66575611e-01 -4.09669310e-01
-3.31200033e-01 -5.94323456e-01 -1.70544863e+00 -2.93023020e-01
3.40455383e-01 9.04484279e-03 -4.70972881e-02 5.68413675e-01
7.32106417e-02 9.49660167e-02 8.14716101e-01 -7.52212763e-01
-3.66996676e-01 1.59268260e+00 -2.31131718e-01 5.96060991e-01
1.13059878e+00 -4.64531928e-01 1.73086119e+00 -2.96639293e-01
1.21384692e+00 1.71886146e+00 6.32100463e-01 2.23539248e-01
-1.42631567e+00 -2.21499860e-01 3.80032986e-01 6.52859688e-01
-6.26894891e-01 -4.50292915e-01 1.06767142e+00 -8.55810940e-01
1.63576877e+00 4.94342059e-01 1.24129355e-01 1.02903879e+00
2.37472624e-01 7.94087291e-01 1.54147243e+00 -1.03508341e+00
2.37742692e-01 -1.90366641e-01 7.51742005e-01 4.76742715e-01
-1.88224748e-01 -3.74749959e-01 -7.44165421e-01 -5.33306062e-01
2.83398051e-02 -9.64962423e-01 1.87089130e-01 3.06940854e-01
-1.00311613e+00 1.09252024e+00 1.73443094e-01 5.21016598e-01
-2.79063404e-01 7.10923923e-03 8.71835947e-01 5.92388034e-01
6.37105882e-01 7.71504104e-01 -5.36867797e-01 -1.60682708e-01
-4.14787114e-01 4.33524579e-01 1.40007055e+00 7.87466288e-01
5.71137726e-01 -4.34235960e-01 -4.42437947e-01 4.17448491e-01
2.02402174e-01 -8.59402493e-02 2.14250565e-01 -1.36141515e+00
7.15371192e-01 8.42925370e-01 -3.08298290e-01 -1.16606987e+00
-4.32399571e-01 4.86533612e-01 1.56334132e-01 -7.28626177e-02
3.75784755e-01 -8.07640627e-02 -2.25538418e-01 1.42999101e+00
4.75144088e-01 -6.88733995e-01 4.80792999e-01 4.69737589e-01
1.23663127e+00 6.13693416e-01 5.43595135e-01 -7.87461340e-01
1.61095059e+00 -6.88995063e-01 -1.07565618e+00 -8.99814442e-02
1.05836689e+00 -8.86960745e-01 9.22577083e-01 1.99749827e-01
-1.06236136e+00 3.61202396e-02 -1.10022485e+00 -5.46557724e-01
-7.69976228e-02 -5.31261027e-01 9.63925302e-01 3.42550278e-01
-5.37365437e-01 4.43341076e-01 -5.58893263e-01 8.06100592e-02
4.63817030e-01 -7.31164441e-02 -1.70228511e-01 5.22704720e-01
-1.35913467e+00 1.17249286e+00 9.97779369e-01 -7.90809691e-01
-1.52255222e-01 -2.13135466e-01 -1.16401708e+00 -3.03570956e-01
8.73883367e-01 7.34292017e-03 1.48859334e+00 -6.07488453e-01
-1.66693664e+00 1.42996478e+00 -5.54627478e-01 -8.18322122e-01
1.19705804e-01 -2.12900326e-01 -4.77391705e-02 2.72903055e-01
4.05195147e-01 1.88202471e-01 1.56548992e-01 -1.06304777e+00
-6.57594025e-01 -6.80371299e-02 6.39551759e-01 2.96503574e-01
1.93506330e-01 6.99464381e-01 6.10765517e-01 -3.99210155e-01
3.67407262e-01 -6.79728448e-01 1.12646684e-01 -6.14454687e-01
-7.33970284e-01 -1.11692643e+00 6.24528110e-01 -6.95531368e-01
1.62854850e+00 -1.54226685e+00 7.02393726e-02 -1.44629478e-02
7.89678842e-02 -2.53943115e-01 6.14004314e-01 4.50734854e-01
-9.41118896e-02 3.61481547e-01 -7.98545405e-02 4.54092860e-01
-1.09005217e-02 4.27642494e-01 -5.70417583e-01 6.73457310e-02
5.12798369e-01 1.02123952e+00 -7.79379904e-01 -1.21533418e+00
1.52684227e-01 -2.43766457e-01 -4.13222879e-01 1.50512964e-01
-9.94710505e-01 1.63584068e-01 -4.81910169e-01 5.00575125e-01
-1.09544836e-01 -2.56162941e-01 1.10264206e+00 -2.82972381e-02
-1.64601132e-01 1.61990345e+00 -5.90204179e-01 1.27452385e+00
-2.02982649e-01 1.04583490e+00 -2.65468270e-01 -1.25369871e+00
1.03687787e+00 3.77784938e-01 -2.63490099e-02 -6.28505290e-01
3.52409065e-01 1.81798153e-02 1.82727695e-01 -8.01227748e-01
2.14709327e-01 -4.75614369e-01 -6.64064586e-01 4.93286580e-01
-1.06307462e-01 7.72474660e-03 2.49748141e-01 1.76919267e-01
1.22083294e+00 4.18253243e-01 1.15407491e+00 -6.14256740e-01
9.00670528e-01 7.51736820e-01 8.09437931e-01 3.81202608e-01
-4.56883870e-02 -3.35098580e-02 1.36452222e+00 -6.64714992e-01
-6.66739941e-01 -8.20009172e-01 -4.50243086e-01 1.88342512e+00
1.05112270e-02 -7.02439487e-01 -4.55172658e-01 -8.59742641e-01
-2.40685582e-01 1.28845811e+00 -6.74825072e-01 4.57493871e-01
-1.43062842e+00 -5.08084238e-01 6.27397656e-01 4.74250168e-01
1.68962970e-01 -1.17840791e+00 -6.94920182e-01 3.03175688e-01
-4.96455312e-01 -1.17259395e+00 3.89486551e-01 5.97591758e-01
-6.11542881e-01 -1.49625862e+00 7.34097302e-01 -9.90958810e-01
-1.13019124e-01 -5.76194465e-01 1.40452957e+00 -9.61927846e-02
3.32222015e-01 -2.23091602e-01 -6.13419116e-01 -5.44938624e-01
-1.01337516e+00 2.46782050e-01 -5.10269105e-01 -7.70291924e-01
8.48344028e-01 -2.11801782e-01 7.99772069e-02 -3.49847168e-01
-2.87041932e-01 -7.92442039e-02 -1.04038484e-01 6.14706695e-01
5.18102288e-01 -1.97378784e-01 4.49430555e-01 -1.54583097e+00
1.13081789e+00 -4.72630590e-01 -8.05141479e-02 5.17366230e-01
-6.31371200e-01 3.92427951e-01 1.53629184e-01 -2.76668131e-01
-1.23863649e+00 -2.95764714e-01 -3.25658359e-02 6.13795996e-01
-2.00917438e-01 7.07743943e-01 -1.31605238e-01 4.77645010e-01
1.20583272e+00 -7.21908212e-01 -5.41902613e-03 -3.33011970e-02
5.01304448e-01 5.62946677e-01 4.59777594e-01 -1.07081759e+00
1.88414529e-01 2.01914519e-01 2.01361641e-01 -5.53823113e-01
-1.33858120e+00 -2.39972204e-01 -7.79015064e-01 -5.78742959e-02
1.33895075e+00 -6.56630397e-01 -9.03706253e-01 -3.20962042e-01
-1.57635760e+00 -4.24039215e-01 -4.30189550e-01 3.49158317e-01
-7.88921714e-01 1.70717061e-01 -1.01819217e+00 -1.01234126e+00
-1.56840846e-01 -6.42290771e-01 6.12394810e-01 9.19556916e-02
-1.15707135e+00 -1.21558750e+00 3.92604172e-02 5.16822398e-01
-1.47617072e-01 8.49266946e-01 1.46954656e+00 -1.32253993e+00
-3.03375542e-01 3.40218306e-01 -1.92368142e-02 -2.43568540e-01
1.84624955e-01 -4.25617844e-02 -8.87085438e-01 5.68633020e-01
1.06159456e-01 -8.07431281e-01 6.39627099e-01 3.10712725e-01
4.06992614e-01 -8.84130836e-01 -3.79281282e-01 -2.05442175e-01
1.10190260e+00 5.40705502e-01 3.15427125e-01 9.03324783e-01
5.12596130e-01 1.38375390e+00 7.01247334e-01 -5.81651293e-02
1.02964008e+00 4.63739514e-01 -1.06716849e-01 4.49470937e-01
-3.16152334e-01 -2.28148043e-01 2.05674022e-02 5.64491749e-01
-2.12517893e-03 1.40961140e-01 -1.44475079e+00 4.49588239e-01
-2.05323982e+00 -1.11963356e+00 -4.59255636e-01 1.26352382e+00
1.80992985e+00 6.04613662e-01 2.32647628e-01 3.67013842e-01
6.68047190e-01 5.26074350e-01 -7.94689134e-02 -1.04275465e+00
-5.86359262e-01 1.84094742e-01 5.03651388e-02 1.07473302e+00
-1.40706205e+00 1.27967072e+00 7.23220158e+00 6.29450321e-01
-7.12782919e-01 2.33754173e-01 5.28251231e-01 3.39410394e-01
-4.93419647e-01 4.38547671e-01 -6.16752803e-01 -8.17368776e-02
1.05657411e+00 -1.80056408e-01 7.26801828e-02 9.11009312e-01
-3.47681880e-01 -4.67574239e-01 -1.38658810e+00 3.44071507e-01
1.65481642e-01 -1.95623267e+00 -2.54275292e-01 -2.41784304e-01
4.18646902e-01 -3.92092168e-01 -6.16203308e-01 3.02451491e-01
7.31410980e-01 -9.97844040e-01 7.70557225e-01 1.71299800e-01
3.90265703e-01 -4.26645607e-01 6.15427792e-01 2.81016320e-01
-9.40093637e-01 -6.85473606e-02 5.35251461e-02 -9.46261048e-01
2.75841862e-01 2.22527564e-01 -1.10510921e+00 2.85498291e-01
2.79897720e-01 3.47589135e-01 -3.19949329e-01 -1.73326209e-02
-9.01870906e-01 1.07737231e+00 -1.15482993e-01 -4.36838537e-01
7.21246526e-02 2.36480281e-01 6.77263081e-01 1.52833784e+00
-8.03974330e-01 6.32885158e-01 5.22773087e-01 4.94440734e-01
8.82394016e-02 2.49532983e-01 -6.36681020e-01 2.24202693e-01
7.64729559e-01 6.49319649e-01 -9.15979326e-01 -6.19450688e-01
-2.63753384e-01 -3.60558815e-02 4.72381562e-01 -1.14465654e-01
-3.56131434e-01 -1.29524484e-01 2.15498835e-01 6.97919503e-02
1.77499443e-01 -9.29380581e-02 -9.93942916e-01 -6.42538667e-01
6.27896711e-02 -9.25166607e-01 7.52753556e-01 -1.87848926e-01
-1.27097273e+00 5.77033579e-01 6.40950680e-01 -4.84332830e-01
-9.08719957e-01 -7.38089859e-01 -6.76277280e-01 5.42806804e-01
-1.28592515e+00 -1.10539937e+00 4.60737407e-01 -6.01186715e-02
5.86126268e-01 -2.40411714e-01 1.22347689e+00 -5.44131100e-01
-9.21430364e-02 7.41897821e-02 -7.00570643e-01 6.30188406e-01
4.76767927e-01 -1.52667904e+00 1.99688166e-01 4.42948222e-01
1.14050604e-01 4.97234732e-01 8.26527953e-01 -6.96702957e-01
-9.91850913e-01 -3.68630707e-01 1.65167534e+00 -7.64707804e-01
1.09743154e+00 -8.13532695e-02 -1.14877784e+00 1.01857507e+00
7.76340306e-01 -4.47460413e-01 1.13068867e+00 1.00297081e+00
-9.41966176e-01 4.36105609e-01 -8.97897899e-01 2.90257126e-01
9.52293038e-01 -9.42459285e-01 -1.89880371e+00 3.93557996e-01
1.03974652e+00 -6.70960844e-01 -1.13596606e+00 5.98409355e-01
3.30107719e-01 -5.75360477e-01 7.63240635e-01 -1.19483984e+00
9.97098505e-01 -6.03434108e-02 -5.04754364e-01 -4.99208957e-01
-1.43215224e-01 -7.25458562e-01 -3.90416890e-01 1.38941753e+00
8.47729087e-01 -2.34884009e-01 6.49883091e-01 9.00247633e-01
1.62006468e-01 -7.01793432e-01 -9.19628322e-01 -4.16051418e-01
7.02884018e-01 -5.38608611e-01 1.85640961e-01 1.21386242e+00
1.04218876e+00 1.39455283e+00 3.56274575e-01 -4.89882082e-01
4.09449786e-01 6.94954991e-01 3.22820365e-01 -1.22427559e+00
-2.99080640e-01 -5.89944899e-01 -3.90634537e-01 -6.51135802e-01
9.79229391e-01 -1.14754450e+00 1.81840938e-02 -1.75570679e+00
-3.23610343e-02 -4.61396873e-01 1.97683901e-01 5.32638550e-01
-2.34603822e-01 -2.10604206e-01 7.88722560e-02 3.07811737e-01
-8.23782563e-01 2.09787954e-02 7.52430797e-01 -4.08958137e-01
-4.23248827e-01 -3.43356967e-01 -6.90992594e-01 1.09904611e+00
8.63583267e-01 -5.86832345e-01 -4.26176906e-01 -1.38397321e-01
5.55120230e-01 2.67829120e-01 -6.86589852e-02 -2.32492298e-01
3.37286383e-01 -6.31364167e-01 9.92635638e-02 -5.51706195e-01
-1.85621865e-02 -1.21469632e-01 -3.73668849e-01 2.28942201e-01
-1.17452180e+00 -7.41510317e-02 1.56659484e-01 1.22570284e-01
-3.21500152e-01 -4.71769035e-01 2.29910582e-01 -1.61068738e-01
-7.07650900e-01 -9.60997522e-01 -7.49676764e-01 7.11139619e-01
1.12699509e+00 6.44532964e-02 -9.02621448e-01 -1.56886145e-01
-8.32403779e-01 1.05880633e-01 3.84836681e-02 1.91433325e-01
3.03885549e-01 -1.07878995e+00 -8.07644069e-01 -5.63262999e-01
1.08485967e-02 1.68108180e-01 -5.79510212e-01 5.08658648e-01
-5.19199014e-01 5.63397467e-01 -6.15698397e-02 -4.80279744e-01
-1.53301239e+00 4.71935034e-01 -1.48740649e-01 -5.29159725e-01
-5.71344852e-01 9.47866619e-01 -3.38615119e-01 -4.24418956e-01
2.08589822e-01 -2.78744519e-01 -7.19359577e-01 5.63965261e-01
4.66665834e-01 2.84474462e-01 -1.74837023e-01 -7.67353892e-01
-7.33822644e-01 1.93175912e-01 -1.62879359e-02 -4.65876251e-01
1.09689069e+00 -1.01581596e-01 -6.18568242e-01 1.20662665e+00
8.78244340e-01 7.89690390e-02 -6.24369740e-01 -5.14146864e-01
1.09938812e+00 3.01798917e-02 -1.27549276e-01 -7.11661220e-01
3.22868347e-01 4.97597307e-01 -4.22607750e-01 9.01298344e-01
6.56877935e-01 8.35123301e-01 2.14185342e-01 7.31550276e-01
2.87901610e-01 -1.34860384e+00 -1.70549825e-01 8.28866184e-01
9.41775024e-01 -1.28996110e+00 5.24426758e-01 -9.00755286e-01
-7.59968817e-01 1.14362848e+00 5.08029222e-01 -2.39668593e-01
5.44601798e-01 8.07799816e-01 2.55013883e-01 -4.01494414e-01
-1.20040548e+00 -3.27206582e-01 1.86415717e-01 6.10249639e-01
1.23859107e+00 3.70938510e-01 -9.46477473e-01 5.02348542e-01
-7.68672466e-01 -5.20315945e-01 2.68437088e-01 1.18731570e+00
-6.20508134e-01 -1.41116560e+00 -4.08791512e-01 3.46514821e-01
-8.44886422e-01 -1.69371113e-01 -9.28578198e-01 8.88009369e-01
-1.71095043e-01 1.33680952e+00 2.73360819e-01 -3.58904213e-01
2.36586288e-01 4.41286743e-01 6.48017108e-01 -8.29852462e-01
-8.96481395e-01 -2.30153389e-02 1.44269395e+00 -4.39026445e-01
-1.19051814e+00 -1.03125858e+00 -1.83995438e+00 -4.13423628e-01
-1.91640854e-01 1.58766046e-01 2.23506242e-01 1.33773792e+00
-2.20267296e-01 5.71061790e-01 3.52066904e-01 -1.35502845e-01
-4.20297176e-01 -9.80144024e-01 3.05867732e-01 3.64263833e-01
1.14171132e-01 -5.88392437e-01 -6.93649501e-02 4.40445513e-01] | [10.746925354003906, 9.337820053100586] |
b3db8e10-b2b9-45e8-a6f2-f0eb7dec1b5d | communication-efficient-federated-bilevel | 2302.06701 | null | https://arxiv.org/abs/2302.06701v1 | https://arxiv.org/pdf/2302.06701v1.pdf | Communication-Efficient Federated Bilevel Optimization with Local and Global Lower Level Problems | Bilevel Optimization has witnessed notable progress recently with new emerging efficient algorithms, yet it is underexplored in the Federated Learning setting. It is unclear how the challenges of Federated Learning affect the convergence of bilevel algorithms. In this work, we study Federated Bilevel Optimization problems. We first propose the FedBiO algorithm that solves the hyper-gradient estimation problem efficiently, then we propose FedBiOAcc to accelerate FedBiO. FedBiO has communication complexity $O(\epsilon^{-1.5})$ with linear speed up, while FedBiOAcc achieves communication complexity $O(\epsilon^{-1})$, sample complexity $O(\epsilon^{-1.5})$ and also the linear speed up. We also study Federated Bilevel Optimization problems with local lower level problems, and prove that FedBiO and FedBiOAcc converges at the same rate with some modification. | ['Heng Huang', 'Feihu Huang', 'Junyi Li'] | 2023-02-13 | null | null | null | null | ['bilevel-optimization'] | ['methodology'] | [-6.02899075e-01 -1.88158117e-02 -3.17083180e-01 -1.13131151e-01
-1.04668152e+00 -6.05516016e-01 1.63218945e-01 2.91833222e-01
-5.50268292e-01 8.96237552e-01 1.26169715e-02 -5.96159339e-01
-5.67745805e-01 -8.56168270e-01 -1.11538756e+00 -8.26817334e-01
-7.28244126e-01 5.94636500e-01 -3.34467322e-01 3.15872729e-02
-3.43796723e-02 3.78909148e-02 -1.03222752e+00 7.03871325e-02
9.31444466e-01 1.01261520e+00 -2.69405276e-01 1.19458675e+00
-1.36959881e-01 8.32006156e-01 -3.41689110e-01 -6.55186236e-01
6.57058358e-01 -6.49863243e-01 -7.83344090e-01 -3.57204527e-01
6.51272476e-01 -1.24541018e-03 -5.52423477e-01 1.35220373e+00
5.50350368e-01 2.00648516e-01 -4.76026302e-03 -1.16367912e+00
-5.13852835e-01 8.75346601e-01 -3.91802222e-01 2.89786428e-01
-4.32011597e-02 2.58209676e-01 1.22995532e+00 -9.92391825e-01
4.91582811e-01 1.07987618e+00 7.47886717e-01 2.39476740e-01
-1.08943009e+00 -7.28777707e-01 4.10746783e-01 3.30186188e-01
-1.01079226e+00 -2.21745357e-01 2.05560580e-01 -4.62774821e-02
5.81526220e-01 4.68988746e-01 8.91127110e-01 4.48942631e-01
9.18623954e-02 8.19282949e-01 1.31642163e+00 -3.15963835e-01
4.03385371e-01 -3.45945247e-02 4.80573386e-01 1.41276848e+00
6.56535864e-01 5.03027081e-01 -7.69781768e-01 -4.77718592e-01
2.73311377e-01 2.28346750e-01 -5.45451403e-01 -3.08697224e-01
-1.12937486e+00 1.15730202e+00 5.94856679e-01 -3.62116732e-02
-1.23544268e-01 6.97048247e-01 4.93948191e-01 8.29230845e-01
4.80795562e-01 2.68267810e-01 -6.98886514e-01 -2.49982789e-01
-6.70138717e-01 -1.41155273e-02 1.50335836e+00 9.50186372e-01
1.04549980e+00 4.53892425e-02 1.49942875e-01 7.12348759e-01
4.27431166e-01 9.39117908e-01 1.28332600e-01 -9.44306076e-01
5.55286288e-01 3.37394804e-01 -1.81912065e-01 -8.60437274e-01
-2.68609226e-01 -4.62979823e-01 -8.88048768e-01 6.72891960e-02
6.63544476e-01 -5.88611186e-01 -4.90843266e-01 1.47294128e+00
7.57778704e-01 2.89373457e-01 -5.81960492e-02 1.06227398e+00
4.18981880e-01 7.61914253e-01 -4.20708716e-01 -4.66455549e-01
1.33234990e+00 -1.51938641e+00 -2.99799651e-01 1.13953920e-02
9.03905153e-01 -7.36986756e-01 8.99735272e-01 5.44776976e-01
-1.30413318e+00 2.62747258e-01 -9.13947821e-01 1.85819373e-01
-1.83240518e-01 -3.55960101e-01 1.05192363e+00 8.75840008e-01
-9.55693722e-01 5.81780374e-01 -8.03359747e-01 2.04002067e-01
4.68893230e-01 5.24731398e-01 -1.38435498e-01 -4.84894425e-01
-6.73014164e-01 5.41847467e-01 -6.99004531e-02 6.88265637e-02
-1.36608207e+00 -1.04663789e+00 -7.66499639e-01 -1.45157501e-01
5.98661900e-01 -8.46993804e-01 1.10720384e+00 -4.58616227e-01
-1.41851592e+00 4.18813437e-01 -1.04268298e-01 -7.14345932e-01
6.40330613e-01 -1.25993833e-01 6.80634975e-02 -1.10096075e-01
-2.41836742e-01 -3.20779741e-01 4.38462824e-01 -9.87188280e-01
-9.96791542e-01 -7.26607978e-01 1.75491080e-01 -2.04293113e-02
-5.39977372e-01 -2.69117355e-01 -4.83843416e-01 -2.57811069e-01
-9.83213261e-02 -1.08731425e+00 -6.50309086e-01 2.45971054e-01
2.05981098e-02 -1.32202283e-01 4.95805472e-01 -3.70499611e-01
9.33252096e-01 -1.86430514e+00 1.25152647e-01 4.29146528e-01
6.82023048e-01 2.01030597e-01 -4.95079719e-02 2.79479831e-01
5.70313692e-01 1.26615226e-01 -2.18595117e-01 -3.89814287e-01
2.27868035e-01 3.58504713e-01 -4.72490825e-02 1.08564949e+00
-9.42866981e-01 7.91219890e-01 -1.11146879e+00 -3.98205549e-01
-8.89524445e-02 4.61774439e-01 -8.16746950e-01 9.31334943e-02
-3.42372149e-01 -2.30607446e-02 -5.24216592e-01 8.23603213e-01
7.56901205e-01 -5.80170155e-01 3.34657073e-01 -1.67571038e-01
-1.97370112e-01 -1.63862295e-02 -1.29539800e+00 1.53596342e+00
-9.26152647e-01 5.94923139e-01 8.36002529e-01 -1.04308796e+00
6.68767691e-01 9.17213485e-02 6.77315533e-01 -5.77632606e-01
2.48369515e-01 5.59857368e-01 -4.19211000e-01 -2.69842833e-01
1.88604921e-01 -1.93805218e-01 7.56098256e-02 9.27319229e-01
-1.23179443e-01 2.17556432e-01 1.99223503e-01 1.29094288e-01
1.35334337e+00 -4.32672441e-01 -4.45369259e-02 -3.84108484e-01
3.61095160e-01 2.58634180e-01 6.08443022e-01 9.19362485e-01
-1.61135897e-01 4.98502590e-02 3.89500499e-01 -7.66651571e-01
-7.88870871e-01 -8.54874134e-01 9.21800733e-02 1.37244976e+00
6.01676404e-01 -7.61735916e-01 -5.66116989e-01 -1.17705178e+00
3.21921110e-01 2.17153370e-01 -4.96866018e-01 1.67752896e-02
-8.47014725e-01 -1.15110159e+00 5.13708770e-01 8.55927318e-02
4.58938926e-01 -5.02568483e-01 -4.86678988e-01 1.22313432e-01
2.79505819e-01 -5.09831369e-01 -9.96696413e-01 3.62090349e-01
-8.47387075e-01 -1.38685787e+00 -5.91765821e-01 -6.58211231e-01
8.26115310e-01 4.47325081e-01 1.00010383e+00 2.37571552e-01
-4.55690473e-01 3.95524323e-01 -2.51350284e-01 -2.76778400e-01
-1.31200835e-01 8.98688510e-02 -1.13142855e-01 -5.81777319e-02
1.67508513e-01 -2.33385742e-01 -8.60739172e-01 1.60978973e-01
-5.35632968e-01 -4.11988735e-01 2.41333678e-01 1.24428689e+00
8.20478439e-01 -2.43906349e-01 1.60506174e-01 -1.10613871e+00
6.67898655e-01 -7.44768620e-01 -1.05362833e+00 2.12809488e-01
-1.22870922e+00 2.19738483e-01 1.08664191e+00 -3.89224499e-01
-6.08755350e-01 -1.01587966e-01 3.45525630e-02 -5.85415304e-01
6.53440595e-01 5.24931490e-01 4.23390806e-01 -4.99664068e-01
6.14141703e-01 2.06412762e-01 1.67366102e-01 -6.97054505e-01
5.92596531e-01 5.11902392e-01 4.35846925e-01 -9.69587684e-01
6.08981431e-01 7.15652943e-01 4.96414565e-02 -4.43266064e-01
-8.86717796e-01 -2.26681694e-01 2.23959595e-01 1.70671150e-01
2.32630864e-01 -9.20740664e-01 -1.47460973e+00 -2.49751434e-01
-7.80352592e-01 -4.57218528e-01 -4.01248187e-01 6.43442333e-01
-3.99123251e-01 3.99484843e-01 -6.87844634e-01 -7.15488374e-01
-1.00213802e+00 -1.16860235e+00 3.21401954e-01 9.89614353e-02
1.78373590e-01 -1.20224559e+00 2.57498920e-01 6.23726606e-01
4.80304867e-01 1.70621648e-01 7.62420475e-01 -5.57090580e-01
-9.28697228e-01 -2.06444785e-01 -2.93439925e-01 3.07816993e-02
-1.89226955e-01 -5.72627485e-01 -2.98296273e-01 -8.27814579e-01
-1.46702781e-01 -3.38417530e-01 8.87030661e-01 -3.12743783e-02
1.05486357e+00 -1.06608808e+00 -2.46162787e-01 1.37881756e+00
1.54894519e+00 -1.36806384e-01 -5.99181093e-02 3.59883994e-01
4.66123134e-01 -5.25186583e-02 4.36482072e-01 8.12617123e-01
6.32148266e-01 4.63732406e-02 8.28566313e-01 -4.47118022e-02
-2.53586531e-01 -9.14200097e-02 4.11082715e-01 1.01334095e+00
3.03405553e-01 2.68152133e-02 -6.98374808e-01 8.55631113e-01
-1.83503807e+00 -5.12356699e-01 -2.23724529e-01 2.13137150e+00
1.11134148e+00 -4.78211880e-01 -4.92426194e-02 -1.16972759e-01
4.14759606e-01 1.13958322e-01 -8.55747998e-01 -6.87528610e-01
-4.33244295e-02 4.80408102e-01 8.46865714e-01 6.68982685e-01
-5.66681325e-01 7.63527811e-01 5.83530807e+00 6.28533304e-01
-1.08598542e+00 5.26675105e-01 5.14496505e-01 -6.76438034e-01
-2.93959111e-01 2.74113327e-01 -7.41116941e-01 6.20063722e-01
8.65676165e-01 -6.82046473e-01 1.05749750e+00 1.20246553e+00
-2.51136243e-01 1.12772666e-01 -1.20073795e+00 1.11002088e+00
1.95104733e-01 -1.62097609e+00 -2.86893070e-01 3.32101613e-01
1.14798665e+00 4.53930259e-01 4.91732061e-02 1.90837234e-01
1.05657434e+00 -1.06520033e+00 2.06763074e-01 -1.19881347e-01
6.59867585e-01 -1.00376928e+00 5.44212639e-01 4.15206432e-01
-1.26180530e+00 -5.85118532e-01 -4.14773524e-01 1.25333548e-01
1.20289497e-01 9.75731969e-01 -6.21313274e-01 3.98489714e-01
1.03691757e+00 3.65244061e-01 2.85781436e-02 1.31598401e+00
4.63396311e-02 4.89854634e-01 -6.81790352e-01 -3.14746529e-01
3.33374411e-01 -3.76671225e-01 5.85712731e-01 1.29548967e+00
4.97316688e-01 1.67696446e-01 5.30369997e-01 4.02204067e-01
-5.85199952e-01 3.80059063e-01 -1.69441402e-01 1.09457724e-01
5.55082977e-01 1.21671224e+00 1.93832163e-02 -3.29117447e-01
-2.12920293e-01 7.08440721e-01 5.43065131e-01 4.76362482e-02
-7.33007431e-01 -3.35370600e-01 1.04867744e+00 -2.15541095e-01
3.87482017e-01 -3.22973728e-01 -4.90559995e-01 -1.25466716e+00
-2.38720849e-01 -1.18037033e+00 1.03228700e+00 1.13579787e-01
-1.37166238e+00 5.27553260e-01 -3.60668898e-01 -4.59425539e-01
2.30871335e-01 -3.81240636e-01 -3.25952560e-01 4.70834166e-01
-1.59135425e+00 -7.42630184e-01 -4.58688319e-01 8.00127029e-01
3.68725955e-01 -3.14401805e-01 6.73410237e-01 4.72876251e-01
-6.56700909e-01 8.89140129e-01 5.70110679e-01 -1.48979142e-01
4.42455739e-01 -1.02422392e+00 -7.59970844e-02 7.10631311e-01
1.66813061e-01 6.09669864e-01 6.04333043e-01 -3.38141561e-01
-2.43783307e+00 -1.23323858e+00 7.56403804e-01 -9.01046619e-02
6.30693078e-01 -1.88644767e-01 -3.65073383e-01 5.78619123e-01
1.97861195e-01 8.13180745e-01 6.31216645e-01 8.82726461e-02
-6.04689538e-01 -4.16917711e-01 -1.36069083e+00 5.74644804e-01
1.21230245e+00 -3.03463072e-01 -2.20763702e-02 6.54105127e-01
7.11145282e-01 -6.19129121e-01 -1.19678485e+00 7.50239268e-02
3.59941006e-01 -7.73185790e-01 1.00440550e+00 -4.77107376e-01
-1.44950971e-01 -1.35716572e-01 -3.50187033e-01 -1.38922811e+00
-1.26739189e-01 -1.22873509e+00 -1.07146096e+00 3.90804052e-01
4.53817159e-01 -1.03623462e+00 1.22874022e+00 2.06083968e-01
1.03013806e-01 -1.12720680e+00 -1.33891559e+00 -7.23643124e-01
3.36776495e-01 -2.44287625e-01 6.67149007e-01 7.72771120e-01
2.33393639e-01 -4.27266434e-02 -5.28723776e-01 1.46866024e-01
1.15408385e+00 5.41422427e-01 9.70572114e-01 -8.16775203e-01
-8.21162224e-01 -2.49825940e-01 2.08656505e-01 -1.33197999e+00
8.71572294e-04 -1.23063993e+00 -1.22477643e-01 -1.15328825e+00
3.19737077e-01 -1.04306209e+00 -3.84371519e-01 6.30412161e-01
-6.90663755e-02 2.77559668e-01 4.25107747e-01 2.43533775e-03
-8.10656965e-01 5.90032279e-01 1.16743314e+00 -3.79518896e-01
-5.67023382e-02 -2.28155553e-01 -7.41744220e-01 4.13006037e-01
6.14380300e-01 -5.84514856e-01 -1.90527469e-01 -8.42937350e-01
4.54715997e-01 2.74468541e-01 1.35096937e-01 -5.70217252e-01
7.09553063e-01 -1.88845128e-01 -1.46972075e-01 -2.13036180e-01
1.82015404e-01 -6.62201345e-01 2.58070119e-02 9.44561660e-01
-2.02271134e-01 3.26611310e-01 -9.88673344e-02 7.69551218e-01
9.49482992e-02 -1.95614219e-01 8.79433215e-01 -3.69026721e-01
-1.14942729e-01 7.81786084e-01 -7.05385283e-02 3.85172039e-01
1.23245895e+00 9.89223346e-02 -5.62909365e-01 -2.45276123e-01
-5.24544299e-01 7.75543988e-01 2.33722672e-01 -6.00409061e-02
5.21421432e-01 -9.52046275e-01 -7.39264548e-01 2.31926158e-01
-3.86203408e-01 1.89833358e-01 -4.86616343e-02 1.14686608e+00
-6.73085153e-01 2.21284211e-01 5.94774723e-01 -4.85745311e-01
-1.42975461e+00 5.26973069e-01 6.54337168e-01 -4.07714605e-01
-7.53193498e-01 1.14831281e+00 -3.52092087e-01 -6.99168622e-01
5.73896766e-01 1.83604375e-01 8.02289128e-01 -2.06487060e-01
5.73433876e-01 1.04098105e+00 -4.48060073e-02 5.67730740e-02
-3.32840830e-01 3.86881709e-01 -1.25123337e-01 1.79319054e-01
1.60180283e+00 -3.55388373e-02 -4.95845258e-01 -1.05228052e-01
1.61371970e+00 1.77577101e-02 -1.24027216e+00 -4.23000455e-01
-1.26863942e-02 -6.48995399e-01 3.68121624e-01 -7.65450597e-01
-1.47429097e+00 5.73864877e-01 6.50140584e-01 -1.40237987e-01
7.99971998e-01 -6.89158440e-02 1.25080895e+00 4.71230567e-01
7.94130206e-01 -1.04349101e+00 6.77911788e-02 6.75824642e-01
4.03632969e-01 -1.12301326e+00 2.40740508e-01 -1.89161122e-01
-7.93024600e-02 1.09251833e+00 3.24105680e-01 -2.67775655e-01
9.63175178e-01 4.98795211e-01 -2.21563876e-01 -2.39856213e-01
-1.18078601e+00 1.89699933e-01 -2.81879455e-01 1.71718538e-01
2.28666127e-01 1.11866277e-02 -6.89273000e-01 4.59054768e-01
-2.96124488e-01 1.03314273e-01 2.77224422e-01 9.24001336e-01
-4.22870278e-01 -1.13152444e+00 -5.78921258e-01 3.98606420e-01
-5.80206335e-01 -1.07432887e-01 -7.80841261e-02 3.56362730e-01
5.89698404e-02 9.49396014e-01 -1.92743659e-01 -1.63344517e-01
-2.20155418e-02 -1.51909828e-01 4.68697369e-01 -2.01327026e-01
-1.01286173e+00 -3.04350048e-01 1.00414768e-01 -6.45953596e-01
2.18597054e-01 -1.99681267e-01 -1.50636649e+00 -9.35071349e-01
-2.60555118e-01 7.61613190e-01 9.71919835e-01 6.01120174e-01
5.86555064e-01 2.93929093e-02 9.72285271e-01 -3.76581341e-01
-1.19217944e+00 -3.15920204e-01 -1.76482975e-01 -2.18037050e-03
6.07521713e-01 3.83610241e-02 -6.56031787e-01 -3.32804859e-01] | [6.214059829711914, 5.049936294555664] |
b5ca2866-cc37-4777-ae9d-bd0a63a36f36 | towards-stroke-patients-upper-limb-automatic | 2212.05062 | null | https://arxiv.org/abs/2212.05062v1 | https://arxiv.org/pdf/2212.05062v1.pdf | Towards Stroke Patients' Upper-limb Automatic Motor Assessment Using Smartwatches | Assessing the physical condition in rehabilitation scenarios is a challenging problem, since it involves Human Activity Recognition (HAR) and kinematic analysis methods. In addition, the difficulties increase in unconstrained rehabilitation scenarios, which are much closer to the real use cases. In particular, our aim is to design an upper-limb assessment pipeline for stroke patients using smartwatches. We focus on the HAR task, as it is the first part of the assessing pipeline. Our main target is to automatically detect and recognize four key movements inspired by the Fugl-Meyer assessment scale, which are performed in both constrained and unconstrained scenarios. In addition to the application protocol and dataset, we propose two detection and classification baseline methods. We believe that the proposed framework, dataset and baseline results will serve to foster this research field. | ['Miguel A. Ferrer', 'Josep Lladós', 'Cristina Carmona-Duarte', 'Alicia Fornés', 'Jialuo Chen', 'Asma Bensalah'] | 2022-12-09 | null | null | null | null | ['human-activity-recognition', 'human-activity-recognition'] | ['computer-vision', 'time-series'] | [ 1.71914622e-01 -5.22587076e-02 -5.26613832e-01 6.59580603e-02
-6.84657753e-01 -4.70682889e-01 5.26942551e-01 -1.32428110e-01
-8.03673565e-01 6.91697836e-01 1.00953710e+00 -1.60957500e-01
-3.20142567e-01 -4.43209976e-01 2.91128606e-02 -3.67308855e-01
-4.49084550e-01 3.75616014e-01 4.52239215e-01 -2.79679358e-01
2.05080822e-01 7.84517705e-01 -1.68933356e+00 5.14880598e-01
6.79849625e-01 6.28394783e-01 1.97177500e-01 6.16733909e-01
4.80449826e-01 7.41763532e-01 -4.14885074e-01 9.65714082e-02
7.83920512e-02 -5.22376835e-01 -9.94524240e-01 -1.35602849e-02
-2.58609951e-02 -5.08151352e-01 -3.47882032e-01 5.03834546e-01
9.72043574e-01 5.11734664e-01 6.82808936e-01 -1.18982899e+00
2.82133073e-01 3.51512849e-01 2.72843633e-02 3.48212868e-01
7.46633768e-01 4.26420420e-01 6.44999921e-01 -5.29639840e-01
6.04760349e-01 7.05180824e-01 6.39643729e-01 5.10758936e-01
-5.59358716e-01 -1.94008097e-01 6.09579198e-02 7.65528321e-01
-1.06884670e+00 -3.42857897e-01 7.55516410e-01 -6.62612259e-01
7.96136022e-01 4.46783960e-01 1.04863310e+00 1.44170761e+00
5.87952994e-02 1.20090413e+00 9.16776717e-01 -3.57258320e-01
5.51293612e-01 -5.74874520e-01 1.82357028e-01 6.96601048e-02
2.97075927e-01 -2.49408390e-02 -5.44765115e-01 1.80937409e-01
6.80561662e-01 2.02590093e-01 -6.47267282e-01 -5.11307418e-01
-1.54101348e+00 3.17095757e-01 2.75373966e-01 5.36910892e-01
-9.92160082e-01 6.12178519e-02 4.68204051e-01 -3.93136963e-02
-1.57894880e-01 3.51690352e-02 -2.52516687e-01 -9.46529508e-01
-9.10266161e-01 4.44878250e-01 7.18411803e-01 6.69843674e-01
-9.59243476e-02 -3.76012295e-01 -6.26218438e-01 4.51760411e-01
1.31692976e-01 2.37046257e-02 6.83969796e-01 -7.04491735e-01
4.89517033e-01 7.45343089e-01 2.67538369e-01 -5.14229178e-01
-8.70115042e-01 -3.94304305e-01 -4.78472710e-01 5.34352064e-01
6.67868793e-01 -1.31629124e-01 -6.46525145e-01 1.32746637e+00
4.27711666e-01 -1.48958698e-01 -2.07890674e-01 1.33779693e+00
5.27462423e-01 -3.05610210e-01 3.49142134e-01 -2.41638571e-01
1.65452600e+00 -8.60676527e-01 -7.56879807e-01 -2.88445562e-01
9.94714856e-01 -5.20424664e-01 1.48750579e+00 3.53759468e-01
-9.59073484e-01 -3.21812868e-01 -8.17331612e-01 1.75131321e-01
-1.30156934e-01 4.77735549e-01 6.39792442e-01 5.81093669e-01
-5.66497624e-01 6.52785599e-01 -1.22627413e+00 -8.00628781e-01
2.75354773e-01 1.37457818e-01 -5.13655305e-01 3.34936559e-01
-1.03713715e+00 1.32363820e+00 4.63978142e-01 4.06076789e-01
-5.46367407e-01 -3.75825554e-01 -5.75257182e-01 -2.14375556e-01
2.27055371e-01 -5.49292386e-01 1.16795373e+00 -3.13454598e-01
-1.72651768e+00 7.83366680e-01 1.14163890e-01 -4.53869626e-02
1.25429487e+00 -6.92202508e-01 -2.59508699e-01 9.32424143e-02
-1.65159568e-01 -2.29123414e-01 1.87435746e-01 -6.60415292e-01
-5.33226430e-01 -4.58141863e-01 -1.16206147e-01 5.60735583e-01
-4.32698697e-01 1.91520616e-01 -3.60114396e-01 -7.53814697e-01
1.00661762e-01 -9.10119653e-01 -5.52475043e-02 -1.91427842e-02
-1.32913828e-01 -1.61954314e-02 6.57335401e-01 -1.16119456e+00
1.38875687e+00 -1.72772157e+00 3.18200737e-01 1.69654384e-01
2.67016212e-03 3.11685115e-01 3.49242836e-01 6.09127820e-01
-8.88686851e-02 -3.98030728e-01 -1.93364039e-01 1.79587185e-01
-8.64622965e-02 1.79641739e-01 1.20846733e-01 5.00600398e-01
-1.59975216e-01 8.36333752e-01 -9.23684716e-01 -3.63556981e-01
5.43410122e-01 1.77979693e-01 -4.49292928e-01 3.15740019e-01
2.95074254e-01 8.98248017e-01 -3.59382689e-01 6.01674855e-01
2.16452569e-01 2.88786888e-01 2.43833467e-01 -4.50023115e-01
-1.80883065e-01 2.05902457e-01 -1.44166183e+00 1.89003372e+00
-1.63278759e-01 2.63018996e-01 -1.71321854e-01 -1.06810451e+00
5.35865724e-01 4.70727503e-01 1.09493840e+00 -3.81634772e-01
4.22246993e-01 2.20770940e-01 1.19023696e-01 -1.10881698e+00
5.46372943e-02 4.88592684e-02 3.52556049e-03 3.15581203e-01
-1.03273697e-01 2.03507721e-01 3.54655266e-01 -1.75177991e-01
1.54012561e+00 6.63148522e-01 7.97068179e-01 -1.32525042e-01
5.29834628e-01 6.71531335e-02 5.22748768e-01 4.63962704e-01
-8.01234365e-01 6.47852182e-01 3.43906581e-01 -3.88715684e-01
-6.37645066e-01 -1.21320605e+00 2.73115128e-01 8.03924561e-01
-1.61822687e-03 -3.78304601e-01 -9.02838588e-01 -6.29275918e-01
-1.52187958e-01 4.23186928e-01 -4.30281729e-01 -3.54898512e-01
-9.52831030e-01 -5.58815897e-01 4.85358417e-01 1.02441633e+00
4.32516843e-01 -1.43853855e+00 -1.24563873e+00 4.05329973e-01
-5.36187112e-01 -1.15512991e+00 -2.58885890e-01 -5.39831705e-02
-9.63368475e-01 -1.51875782e+00 -9.53969836e-01 -7.20093131e-01
1.85158849e-01 -1.01142712e-02 7.34866440e-01 -1.67601421e-01
-2.58934617e-01 3.96562845e-01 -9.16566491e-01 -2.12928221e-01
-8.83667469e-02 3.46738845e-01 2.32428536e-01 -1.43896997e-01
3.28061372e-01 -8.45674932e-01 -1.09758031e+00 4.75987583e-01
-6.91375613e-01 1.11222073e-01 8.45860541e-01 4.77702945e-01
3.65318596e-01 -3.93295050e-01 2.90864259e-01 -2.29635403e-01
9.50275779e-01 -2.01181665e-01 1.15935750e-01 3.02924395e-01
-2.90588498e-01 -2.54781157e-01 3.25208336e-01 -7.45141625e-01
-9.17178571e-01 4.28289980e-01 -3.45760524e-01 9.36355740e-02
-6.26585543e-01 6.80668771e-01 -4.55475241e-01 4.15394828e-02
1.03171802e+00 -1.96041632e-03 9.57029406e-03 -7.48829782e-01
2.14558542e-01 8.66331577e-01 9.17106926e-01 -5.02293944e-01
4.89950150e-01 3.98064822e-01 -2.72799209e-02 -7.46911824e-01
-3.54910702e-01 -8.90482664e-01 -1.22688830e+00 -9.43539083e-01
7.59503424e-01 -5.03803909e-01 -1.03624797e+00 5.95645249e-01
-9.26887035e-01 -5.31412244e-01 -3.82629782e-01 1.17069697e+00
-1.07511294e+00 5.95488846e-01 -4.42361951e-01 -9.68172491e-01
-3.98789555e-01 -1.04272652e+00 1.03042030e+00 7.41109103e-02
-6.84789777e-01 -4.87962991e-01 2.98862368e-01 4.71791953e-01
3.60553920e-01 8.30473006e-01 5.09357691e-01 -4.98498082e-01
1.08842582e-01 -5.42542160e-01 2.07814351e-01 1.03066772e-01
3.58536541e-01 -4.98986244e-01 -4.77074444e-01 -2.47849047e-01
-1.13059059e-02 -1.07293367e-01 1.92646027e-01 2.29248047e-01
8.98636222e-01 -5.10843620e-02 -1.51261732e-01 1.44813314e-01
8.58868003e-01 3.50347646e-02 1.05064225e+00 6.10903978e-01
5.44919431e-01 9.08212781e-01 8.23005557e-01 4.11909372e-01
2.07399637e-01 1.21622491e+00 2.74761409e-01 -7.74765611e-02
-4.09805775e-01 -4.96644713e-02 4.76134568e-01 7.33076632e-01
-9.64327455e-01 2.13537067e-01 -1.31212783e+00 3.69242609e-01
-2.35265255e+00 -1.28250194e+00 -4.82692868e-01 2.22478557e+00
6.07868433e-01 3.84921674e-03 6.46767557e-01 6.34472370e-01
3.31201255e-01 -1.97680697e-01 -3.72605860e-01 1.78348735e-01
2.95931697e-01 3.28931183e-01 3.63646448e-01 1.73310742e-01
-1.02037394e+00 8.00557554e-01 6.31045866e+00 4.51828092e-01
-8.64226937e-01 2.47509003e-01 -3.12063962e-01 -2.63726324e-01
5.43832123e-01 -9.81372744e-02 -1.66475177e-01 5.31121969e-01
7.69477189e-01 5.04103750e-02 2.22531602e-01 7.82193601e-01
9.01447713e-01 -2.94104546e-01 -1.13092363e+00 7.11900830e-01
-1.27384216e-01 -8.26053083e-01 -3.04949790e-01 3.67502607e-02
5.54363504e-02 3.43201421e-02 -7.79187739e-01 3.46814036e-01
-1.23472095e-01 -6.32658303e-01 6.66683376e-01 9.50259626e-01
4.86077428e-01 -4.02577788e-01 8.35211456e-01 5.99111915e-01
-1.41009247e+00 -4.11076359e-02 4.27596658e-01 -4.43634182e-01
6.22999966e-01 2.27608114e-01 -4.74070489e-01 5.77156067e-01
7.41424978e-01 5.34375131e-01 -3.20312023e-01 1.54356921e+00
-5.77648520e-01 6.07962430e-01 -2.56528735e-01 -1.65205225e-01
-2.50129491e-01 -1.14262208e-01 5.91838181e-01 1.14089918e+00
2.81284630e-01 2.35540435e-01 4.00127500e-01 2.68855453e-01
4.99060124e-01 2.77714938e-01 -5.08996010e-01 1.86245635e-01
-3.50254774e-02 1.03737104e+00 -8.14786673e-01 -7.80205848e-03
-1.63658515e-01 1.13900173e+00 1.45058423e-01 2.96425790e-01
-5.92592180e-01 -3.67540717e-01 7.48158813e-01 4.34036583e-01
-5.03845274e-01 -5.44512033e-01 -4.66289192e-01 -1.35682249e+00
6.50916815e-01 -9.40874100e-01 5.71904123e-01 -6.70366466e-01
-8.13295603e-01 -8.79716501e-03 3.31248462e-01 -1.74333060e+00
-6.04885340e-01 -7.29979217e-01 -8.15108478e-01 7.64040172e-01
-9.89773929e-01 -9.99878705e-01 -9.88010466e-01 7.47707427e-01
5.82828879e-01 1.41781077e-01 9.10988033e-01 4.47870761e-01
-4.65596735e-01 3.09452593e-01 -1.90910816e-01 2.25273356e-01
4.41965997e-01 -8.22501957e-01 -2.91498639e-02 1.03750992e+00
-3.58796299e-01 6.05373204e-01 8.77784252e-01 -7.56059408e-01
-1.31304801e+00 -4.54338551e-01 6.62651241e-01 -5.74664652e-01
7.06884384e-01 -1.64431676e-01 -6.13994479e-01 4.29844767e-01
-4.35865581e-01 -1.46685511e-01 5.47725022e-01 -1.59871474e-01
4.41093892e-01 2.18934938e-01 -1.11836684e+00 8.67277503e-01
1.71163416e+00 -2.64138430e-01 -8.61392736e-01 5.34416556e-01
-1.72040656e-01 -5.10500073e-01 -1.05534363e+00 7.57949471e-01
1.19605792e+00 -4.54322517e-01 8.86354446e-01 -9.38570797e-01
1.77491903e-01 -3.47053379e-01 9.77774411e-02 -1.12949979e+00
-2.62889534e-01 -4.83376026e-01 -4.65988338e-01 7.59357095e-01
-1.02201052e-01 -3.03757161e-01 1.21508920e+00 5.14484167e-01
-1.14749767e-01 -6.79301381e-01 -1.11836529e+00 -1.02361369e+00
-2.16320455e-01 -7.62627780e-01 2.62697726e-01 6.16409302e-01
6.67816043e-01 7.22656539e-03 -5.31949759e-01 -1.40237110e-02
4.47435588e-01 -1.89223468e-01 9.00446355e-01 -1.32113528e+00
-1.54805854e-01 -6.13579214e-01 -9.39029753e-01 -6.31054580e-01
-3.20408016e-01 -6.30382657e-01 1.84912786e-01 -1.98438334e+00
-3.60004269e-02 9.99325961e-02 -2.54130989e-01 5.53674698e-01
-3.06460764e-02 7.89048001e-02 9.97550339e-02 5.45369148e-01
-5.89765236e-02 6.01140201e-01 1.04563677e+00 3.51735353e-02
-5.08065403e-01 4.78471786e-01 -1.75524771e-01 6.76128507e-01
1.00216663e+00 -2.57711381e-01 -3.80450904e-01 2.92600431e-02
-2.61871427e-01 -1.39021710e-01 4.22748566e-01 -1.39143455e+00
2.58266985e-01 -3.96537185e-01 1.67427838e-01 -6.20962322e-01
1.02216505e-01 -8.65018845e-01 2.62056410e-01 8.81429374e-01
-3.75672542e-02 1.51646258e-02 -2.38441125e-01 1.83673739e-01
-1.94750372e-02 -1.21282436e-01 4.09800768e-01 2.46753227e-02
-8.74865949e-01 9.08837765e-02 -6.09606206e-01 -2.72419881e-02
1.35818934e+00 -7.16621697e-01 -1.27363876e-01 -2.75798351e-01
-1.21429169e+00 2.28340119e-01 1.05086990e-01 6.13641083e-01
4.81171101e-01 -1.46173573e+00 -5.61353743e-01 -1.85644537e-01
5.50810575e-01 -3.07745159e-01 2.38036722e-01 1.40825236e+00
-7.01020718e-01 2.65062243e-01 -7.51183927e-01 -2.75143147e-01
-1.24271178e+00 6.54745400e-02 2.41056472e-01 -4.15650576e-01
-1.09160256e+00 -8.11151937e-02 -5.80855489e-01 -3.99014980e-01
6.53458536e-01 -1.89332724e-01 -7.72486985e-01 1.54450694e-02
5.63226640e-01 7.53308594e-01 4.43154454e-01 -5.45441210e-01
-5.86870313e-01 2.27701411e-01 3.07723552e-01 -3.42568606e-01
1.40352488e+00 1.54225320e-01 3.40361774e-01 2.16502279e-01
6.91942990e-01 -2.65402198e-01 -1.14966106e+00 3.31139192e-03
5.33974409e-01 -3.94354284e-01 -4.55568768e-02 -8.93260241e-01
-6.84606254e-01 7.32612133e-01 1.18245959e+00 -2.87420332e-01
1.15614140e+00 -1.13517232e-01 7.36818552e-01 3.21974218e-01
7.45878994e-01 -1.42148054e+00 -2.84683798e-02 1.81824148e-01
1.20980096e+00 -9.15855527e-01 6.78706020e-02 -3.58464360e-01
-5.38217068e-01 1.09117973e+00 5.27953982e-01 1.43475965e-01
5.10286629e-01 1.42489761e-01 1.01228729e-01 -8.85556713e-02
5.69695085e-02 -6.68599665e-01 4.97343808e-01 8.29886138e-01
6.20691419e-01 2.02071726e-01 -1.31046283e+00 8.40707898e-01
-1.57341525e-01 6.86994314e-01 3.16902995e-01 1.46581984e+00
-3.90662223e-01 -1.14002347e+00 -2.64612466e-01 5.27426302e-01
-2.18854398e-01 6.15253747e-01 -4.19851363e-01 8.88847470e-01
-5.97489141e-02 9.55113113e-01 -6.40927374e-01 -6.61834657e-01
1.26042819e+00 7.44345710e-02 4.34671223e-01 -6.35293722e-01
-8.18356872e-01 -3.27701271e-01 2.23364234e-01 -1.01109505e+00
-3.72374028e-01 -9.66266870e-01 -1.31923020e+00 -5.29222600e-02
1.12235174e-02 -1.08041957e-01 6.09414041e-01 1.07487285e+00
1.50316074e-01 7.13341415e-01 8.17388222e-02 -1.17605841e+00
-5.95724285e-01 -1.33601511e+00 -3.59237850e-01 6.98138714e-01
-1.65031195e-01 -1.06681263e+00 -9.51871201e-02 5.50513389e-03] | [7.07755184173584, 0.2852584719657898] |
e7a8a28b-0bcf-4095-a974-ff154e1b4518 | strong-but-simple-baseline-with-dual | 2012.05010 | null | https://arxiv.org/abs/2012.05010v2 | https://arxiv.org/pdf/2012.05010v2.pdf | Strong but Simple Baseline with Dual-Granularity Triplet Loss for Visible-Thermal Person Re-Identification | In this letter, we propose a conceptually simple and effective dual-granularity triplet loss for visible-thermal person re-identification (VT-ReID). In general, ReID models are always trained with the sample-based triplet loss and identification loss from the fine granularity level. It is possible when a center-based loss is introduced to encourage the intra-class compactness and inter-class discrimination from the coarse granularity level. Our proposed dual-granularity triplet loss well organizes the sample-based triplet loss and center-based triplet loss in a hierarchical fine to coarse granularity manner, just with some simple configurations of typical operations, such as pooling and batch normalization. Experiments on RegDB and SYSU-MM01 datasets show that with only the global features our dual-granularity triplet loss can improve the VT-ReID performance by a significant margin. It can be a strong VT-ReID baseline to boost future research with high quality. | ['Xichuan Zhou', 'Dong Li', 'Xiaoheng Tan', 'Yanxia Chai', 'Haijun Liu'] | 2020-12-09 | null | null | null | null | ['cross-view-person-re-identification'] | ['computer-vision'] | [-2.26425260e-01 -3.47249746e-01 -2.03072131e-01 -6.54953599e-01
-9.19640958e-01 -4.23894733e-01 7.19269156e-01 2.74909079e-01
-6.73775673e-01 7.56156325e-01 3.16833228e-01 -1.03662789e-01
5.20582460e-02 -7.04697967e-01 -7.01567769e-01 -7.31553376e-01
1.51695192e-01 3.83051157e-01 1.32799670e-01 -6.90837651e-02
-2.07754672e-02 3.91700953e-01 -1.49468315e+00 2.66217142e-01
1.01458812e+00 8.46649408e-01 -3.30238849e-01 2.91618168e-01
8.07462707e-02 1.52150854e-01 -5.44089139e-01 -8.58739436e-01
7.11388171e-01 -1.57429963e-01 -6.31281734e-01 -3.56737614e-01
1.01789248e+00 -4.88596886e-01 -2.56376207e-01 9.76940811e-01
9.44963753e-01 2.24077910e-01 6.92370534e-01 -1.28430521e+00
-6.13226473e-01 5.79912782e-01 -9.04246867e-01 2.36111015e-01
3.28287065e-01 2.85762370e-01 8.59394193e-01 -7.18960881e-01
2.79080927e-01 1.65244293e+00 9.51072872e-01 5.19932985e-01
-1.46821058e+00 -1.07997763e+00 5.33055186e-01 2.73222297e-01
-1.60804784e+00 -2.44606301e-01 4.38914448e-01 -2.83130914e-01
6.44786596e-01 4.61744964e-01 2.51050293e-01 1.31676221e+00
-1.55242786e-01 7.18429506e-01 1.18992960e+00 -3.28891695e-01
-6.04201704e-02 2.25159392e-01 6.98377788e-01 4.12699103e-01
5.64204931e-01 2.72503734e-01 -3.84530902e-01 -3.09780866e-01
6.66761637e-01 6.54270276e-02 -1.19751163e-01 -1.78343907e-01
-1.07754290e+00 6.51860714e-01 5.78090847e-01 -8.61285329e-02
9.26130340e-02 -4.87093888e-02 6.03933036e-01 1.75232261e-01
5.98971248e-01 2.24749014e-01 -3.48590702e-01 4.44737449e-02
-1.01409709e+00 3.75843823e-01 4.91357833e-01 1.21872604e+00
8.27915490e-01 -2.53152221e-01 -8.69949400e-01 9.47078049e-01
-5.35621941e-02 5.08577764e-01 3.25360119e-01 -5.75337112e-01
6.37301683e-01 4.68762696e-01 2.33807668e-01 -6.88088596e-01
-4.55497414e-01 -5.99680245e-01 -9.16126668e-01 -6.98672011e-02
5.71007371e-01 -1.15798905e-01 -9.04698193e-01 2.10423207e+00
4.11115408e-01 3.60162646e-01 -4.94624287e-01 7.73500025e-01
6.74532771e-01 3.46403062e-01 3.94019753e-01 3.58763859e-02
1.70724750e+00 -9.30030525e-01 -4.59824651e-01 2.33720839e-02
7.29318798e-01 -4.35117841e-01 1.18604743e+00 -1.11927487e-01
-1.04837215e+00 -7.53488243e-01 -1.01980889e+00 -2.44636893e-01
-6.17348313e-01 3.62005293e-01 5.12722969e-01 9.58708048e-01
-1.17752850e+00 4.40406144e-01 -5.17468512e-01 -7.18128502e-01
1.89940616e-01 4.06949610e-01 -3.97996306e-01 -1.74299721e-02
-1.15693986e+00 7.17639089e-01 2.28847787e-01 4.27891985e-02
-2.93223172e-01 -1.07511330e+00 -7.18382001e-01 2.40870014e-01
1.77446187e-01 -1.01922643e+00 9.09142077e-01 -3.16418469e-01
-1.34580982e+00 9.20971751e-01 -4.18064833e-01 -5.51497340e-01
9.43030357e-01 -1.91246733e-01 -2.46035844e-01 -2.61236459e-01
2.48619139e-01 7.23528981e-01 5.45203626e-01 -1.30882013e+00
-5.12292206e-01 -6.08834028e-01 6.01939904e-03 1.39842927e-01
-5.82481980e-01 9.93439928e-02 -4.60920274e-01 -9.78402376e-01
-2.36543342e-01 -9.24718738e-01 9.34107825e-02 -1.44927979e-01
-6.31571352e-01 -4.90321100e-01 5.04059553e-01 -5.98388493e-01
1.20174110e+00 -2.04510593e+00 -3.16825688e-01 6.70239180e-02
1.52180105e-01 4.00860757e-01 -2.92255104e-01 3.13517392e-01
-9.75022987e-02 3.11504275e-01 3.38260345e-02 -8.64429235e-01
1.60389513e-01 -3.81792843e-01 -2.14607894e-01 3.95270020e-01
-8.30795541e-02 8.12872827e-01 -7.13909507e-01 -4.35789973e-01
8.74379650e-02 4.61287737e-01 -5.33770740e-01 1.03238389e-01
3.04496765e-01 3.31697971e-01 -3.53348613e-01 5.81514239e-01
1.08159614e+00 -8.75122696e-02 -1.87124431e-01 -4.68838841e-01
-8.71477574e-02 1.92541435e-01 -9.61240292e-01 1.28920829e+00
-3.11671793e-01 4.45921123e-01 -5.57038710e-02 -6.66776657e-01
8.65997791e-01 -9.46446583e-02 9.98499915e-02 -8.09469879e-01
4.06713262e-02 -4.86476012e-02 -3.10030550e-01 6.66730851e-02
8.57637286e-01 -2.81534344e-02 -2.77940363e-01 3.55830699e-01
-2.40322173e-01 8.57448280e-01 2.75855716e-02 1.55372709e-01
8.09824765e-01 -6.19309619e-02 4.37231623e-02 -4.16636527e-01
4.94611830e-01 -4.72710848e-01 7.63861418e-01 1.06373131e+00
-4.14357901e-01 7.21094191e-01 1.42665327e-01 -2.35383898e-01
-1.11244774e+00 -1.23554564e+00 -4.44622874e-01 1.17261553e+00
3.60292166e-01 -3.36687118e-01 -8.04964483e-01 -7.97341645e-01
3.74765873e-01 4.52100366e-01 -8.31101120e-01 -1.60975397e-01
-5.66770673e-01 -1.03205264e+00 7.33319938e-01 7.03168035e-01
9.76073623e-01 -3.80046457e-01 3.03383917e-01 -6.61684275e-02
-2.69381762e-01 -1.20950603e+00 -1.10852897e+00 -1.24103867e-01
-6.17795527e-01 -6.84607208e-01 -1.05798149e+00 -5.43387771e-01
6.63375616e-01 5.85166216e-01 7.88146615e-01 5.12517877e-02
-3.32088530e-01 9.73840356e-02 -3.47435772e-01 -2.52145588e-01
2.17205226e-01 3.25830996e-01 3.02329421e-01 1.74665213e-01
5.75274110e-01 -4.73436385e-01 -8.41154695e-01 6.13504827e-01
-3.75419408e-01 -1.32505804e-01 3.86173964e-01 1.06482387e+00
1.87394649e-01 -3.95313978e-01 5.88292718e-01 -6.66160405e-01
5.37323058e-01 -1.93931401e-01 -4.44665194e-01 4.80577290e-01
-6.18978322e-01 1.31454812e-02 5.25433779e-01 -4.97563452e-01
-1.12347412e+00 -4.22096878e-01 4.85327952e-02 -1.90961272e-01
-5.24751768e-02 -1.42845705e-01 -4.68266249e-01 -8.90723392e-02
5.87295294e-01 1.80959880e-01 -3.85098487e-01 -7.45217621e-01
2.89228231e-01 7.70513415e-01 5.31032145e-01 -9.03837323e-01
8.61762285e-01 4.03585315e-01 -1.38783649e-01 -4.97337550e-01
-7.43186414e-01 -5.42492747e-01 -5.40709794e-01 2.28132427e-01
9.24782276e-01 -1.23110390e+00 -1.00914073e+00 8.14887106e-01
-9.31761861e-01 -3.01254600e-01 -1.22954167e-01 3.17455024e-01
3.92288305e-02 6.49146855e-01 -8.12182844e-01 -8.36768746e-01
-5.23191571e-01 -1.06288469e+00 1.28212786e+00 3.68711591e-01
-3.77272256e-02 -7.50098050e-01 -1.97055712e-01 5.38861752e-01
2.81293392e-01 5.85563555e-02 6.60506785e-01 -6.41733885e-01
-5.83419859e-01 -2.49962166e-01 -7.51012027e-01 1.82195380e-01
2.23071530e-01 -2.55971730e-01 -1.18133986e+00 -7.69776523e-01
-5.39047778e-01 -2.80877531e-01 1.20820689e+00 2.31022328e-01
1.34968877e+00 -2.65381396e-01 -5.51337481e-01 9.61967528e-01
1.19419897e+00 -6.91778138e-02 5.62223673e-01 5.28449953e-01
9.63487387e-01 4.75129575e-01 4.08907831e-01 3.08320940e-01
8.78406405e-01 1.10201001e+00 -2.54154831e-01 -1.36390939e-01
-2.48893753e-01 -4.43042815e-01 2.74276525e-01 2.96643615e-01
-1.26282632e-01 -1.60844505e-01 -6.04278326e-01 4.26246017e-01
-1.81844544e+00 -1.13166630e+00 2.19983563e-01 2.74594975e+00
8.91581535e-01 -2.78802924e-02 7.16968656e-01 -1.38726801e-01
1.16935611e+00 1.26431212e-01 -4.64575738e-01 -3.30179155e-01
-2.23763064e-01 -2.71980613e-01 8.38083565e-01 4.95631695e-01
-1.20173097e+00 8.98508787e-01 6.24531126e+00 1.15388799e+00
-8.93438220e-01 1.25976264e-01 7.95638680e-01 -2.09174827e-01
-1.28721997e-01 -3.00780505e-01 -1.59307003e+00 9.28685844e-01
7.98078239e-01 -1.53555334e-01 1.22812614e-01 5.71455121e-01
1.07494086e-01 9.00719222e-03 -1.37843955e+00 1.34074175e+00
-9.14642960e-02 -1.06150055e+00 2.79470325e-01 3.38525981e-01
6.60563171e-01 -2.06690237e-01 2.63279766e-01 4.23174709e-01
3.11508417e-01 -9.05298591e-01 6.65555894e-01 1.94332510e-01
8.52787495e-01 -8.59755576e-01 7.05187500e-01 5.69195375e-02
-1.47191000e+00 -6.19643554e-02 -5.15983999e-01 1.95970938e-01
1.11251436e-01 4.70635086e-01 -4.26826179e-01 8.56149554e-01
8.06531310e-01 5.52748084e-01 -8.14522266e-01 1.40001941e+00
2.05745324e-01 4.31703925e-01 -4.09501374e-01 1.13699093e-01
-4.41722050e-02 -2.86706120e-01 5.21361172e-01 1.56135631e+00
2.77387410e-01 -2.40824614e-02 1.95791423e-01 8.27054143e-01
-2.20911041e-01 -1.32435292e-01 -1.21443868e-01 5.51787496e-01
8.33304226e-01 1.12650633e+00 -2.90806711e-01 -3.90886307e-01
-2.54245490e-01 1.13470066e+00 5.06123662e-01 6.28280461e-01
-9.76527691e-01 -4.13019776e-01 9.99323249e-01 2.33349442e-01
2.65650600e-01 9.86878853e-03 -6.09447598e-01 -1.31792176e+00
1.86651722e-01 -5.20915449e-01 5.99725485e-01 -2.45961741e-01
-1.77486968e+00 5.07986426e-01 2.48282000e-01 -1.39046609e+00
9.03065652e-02 -4.68338877e-01 -6.99892640e-01 1.08398747e+00
-1.63162518e+00 -1.45057702e+00 -4.29368556e-01 5.27668655e-01
1.09577045e-01 2.42369883e-02 5.71489334e-01 5.76881349e-01
-1.05241072e+00 1.87281644e+00 2.50727803e-01 3.71130496e-01
1.16138613e+00 -1.24946535e+00 4.56794530e-01 8.73061240e-01
-3.64051521e-01 9.30566251e-01 6.52779341e-01 -5.31068146e-01
-8.89467478e-01 -1.21864951e+00 8.29886854e-01 -5.72878480e-01
2.66843170e-01 -7.14626491e-01 -7.81882048e-01 7.84067214e-01
-1.12229191e-01 3.76523994e-02 5.07396877e-01 4.89718050e-01
-8.97197187e-01 -6.18373275e-01 -1.56344616e+00 6.47134244e-01
1.39932048e+00 -7.03157306e-01 -3.42559308e-01 2.18070567e-01
6.36194110e-01 -2.64670312e-01 -8.69054258e-01 3.60494077e-01
7.13714421e-01 -8.46523941e-01 1.25778782e+00 -4.25960809e-01
-2.37489656e-01 -2.39052236e-01 1.01976939e-01 -1.14727283e+00
-5.34628928e-01 -7.14148045e-01 2.62448788e-01 1.72402430e+00
2.21088558e-01 -1.13134420e+00 8.25446725e-01 1.05587780e+00
8.67522359e-02 -4.02732730e-01 -9.72200274e-01 -1.31029201e+00
3.96916568e-01 -5.60379252e-02 1.04668069e+00 8.41077387e-01
-9.10853595e-02 2.17036195e-02 -5.66725969e-01 2.53653407e-01
1.11455226e+00 -4.09169793e-02 1.04493260e+00 -1.24006128e+00
-1.58885270e-01 -5.35293818e-01 -4.66489464e-01 -1.22383308e+00
1.46962672e-01 -7.85782278e-01 -2.87720770e-01 -1.21076405e+00
8.09757710e-01 -7.19216108e-01 -6.40352070e-01 4.54126716e-01
-5.47301531e-01 2.61986762e-01 3.68224263e-01 2.55656570e-01
-4.57538843e-01 7.64539897e-01 1.07859445e+00 -3.53868395e-01
-2.47578815e-01 -9.83268321e-02 -8.55015993e-01 3.21141958e-01
6.21520162e-01 -2.94707030e-01 -1.81889802e-01 -3.20993125e-01
-4.97414708e-01 -4.50922489e-01 6.23698354e-01 -9.19391096e-01
3.51159513e-01 -1.50826089e-02 5.06070137e-01 -6.27115369e-01
5.30623972e-01 -3.50043833e-01 -9.21405926e-02 3.71394269e-02
-2.95885593e-01 -1.00832634e-01 2.67722845e-01 6.39247417e-01
-4.24307212e-02 4.35240090e-01 8.19706917e-01 2.28522867e-01
-4.42971468e-01 5.68219543e-01 3.61169845e-01 3.34640890e-02
8.02986920e-01 -5.29409885e-01 -7.59829938e-01 -1.78149909e-01
-3.59521240e-01 5.69807053e-01 6.06867731e-01 5.48244178e-01
1.29852504e-01 -1.55789435e+00 -9.29795384e-01 1.97948188e-01
3.98662657e-01 -4.09095675e-01 5.09724438e-01 7.52107382e-01
8.42753127e-02 5.49181938e-01 -2.24102378e-01 -5.23955703e-01
-1.42821932e+00 5.06887257e-01 4.71766531e-01 -3.94490182e-01
-6.08126342e-01 1.13884389e+00 8.01695764e-01 -5.01191497e-01
4.40054238e-01 -2.61563867e-01 1.35110274e-01 -1.18871741e-01
9.26091373e-01 6.43417060e-01 -6.28786460e-02 -7.76483715e-01
-5.73800921e-01 6.96779847e-01 -4.53308433e-01 6.61902279e-02
8.60950232e-01 -1.82249248e-01 2.09002886e-02 4.10450809e-02
1.39400256e+00 -1.13378972e-01 -1.36286080e+00 -2.52265483e-01
-1.77710578e-01 -6.76311433e-01 -2.93462276e-01 -8.13790560e-01
-8.60344410e-01 7.63544440e-01 7.85016596e-01 -7.00393394e-02
9.35252190e-01 -9.33378488e-02 9.42809820e-01 1.23207405e-01
7.36405492e-01 -9.90661681e-01 -2.26902112e-01 2.89320260e-01
7.11522937e-01 -1.31577325e+00 -5.36024896e-03 -5.58734357e-01
-3.23099017e-01 6.19012654e-01 8.48889351e-01 7.51657337e-02
5.07510543e-01 1.35306910e-01 -2.21589819e-01 3.99523050e-01
-3.32222074e-01 -1.89383477e-01 5.42465150e-01 5.96046507e-01
3.32193106e-01 2.97436267e-01 -4.74059165e-01 8.10254395e-01
-1.42226383e-01 -1.57303169e-01 6.38285056e-02 3.41241360e-01
-2.27887332e-01 -1.48959351e+00 -4.56177264e-01 5.75919747e-01
-3.31477106e-01 -1.24814130e-01 -2.38229334e-01 7.49740601e-01
2.03334257e-01 9.83019710e-01 1.46556780e-01 -6.39692187e-01
3.65679145e-01 -8.26067179e-02 5.93370080e-01 -3.66590828e-01
-7.93747544e-01 -2.53098041e-01 2.18842313e-01 -5.00963390e-01
-1.34362608e-01 -7.72829473e-01 -9.00376439e-01 -8.67341816e-01
-2.48903900e-01 -2.37462819e-02 2.44256452e-01 7.97043741e-01
7.08235443e-01 -5.49564622e-02 6.53302491e-01 -8.92464221e-01
-8.06908011e-01 -1.06609821e+00 -7.02646315e-01 6.07741058e-01
4.49745476e-01 -7.38412559e-01 -4.63723719e-01 -4.08438116e-01] | [14.72497272491455, 0.9865638017654419] |
5204a529-86b4-485c-b6a9-70a550f8732e | empirical-asset-pricing-via-ensemble-gaussian | 2212.01048 | null | https://arxiv.org/abs/2212.01048v1 | https://arxiv.org/pdf/2212.01048v1.pdf | Empirical Asset Pricing via Ensemble Gaussian Process Regression | We introduce an ensemble learning method based on Gaussian Process Regression (GPR) for predicting conditional expected stock returns given stock-level and macro-economic information. Our ensemble learning approach significantly reduces the computational complexity inherent in GPR inference and lends itself to general online learning tasks. We conduct an empirical analysis on a large cross-section of US stocks from 1962 to 2016. We find that our method dominates existing machine learning models statistically and economically in terms of out-of-sample $R$-squared and Sharpe ratio of prediction-sorted portfolios. Exploiting the Bayesian nature of GPR, we introduce the mean-variance optimal portfolio with respect to the predictive uncertainty distribution of the expected stock returns. It appeals to an uncertainty averse investor and significantly dominates the equal- and value-weighted prediction-sorted portfolios, which outperform the S&P 500. | ['Puneet Pasricha', 'Damir Filipović'] | 2022-12-02 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-6.86206460e-01 -6.45916238e-02 -2.98439503e-01 -3.08769971e-01
-1.07828128e+00 -9.36455548e-01 7.77931511e-01 2.43099779e-02
-3.39841604e-01 9.42950070e-01 2.22871974e-01 -9.86639023e-01
-5.38528204e-01 -1.04889357e+00 -4.86973226e-01 -5.44492900e-01
-4.31984603e-01 3.65146756e-01 -2.01501966e-01 3.10738266e-01
7.42317498e-01 1.55778304e-01 -9.37305093e-01 -3.21396142e-01
7.26123452e-01 1.79545867e+00 -1.45855159e-01 1.66130751e-01
-1.86568592e-02 1.07582998e+00 -1.07833192e-01 -1.15015042e+00
7.74754643e-01 2.25567818e-01 6.76424280e-02 -3.91646683e-01
-2.01123774e-01 -3.80397737e-01 -1.03121705e-01 8.74630213e-01
2.55237013e-01 4.15471941e-01 1.23275566e+00 -9.14638221e-01
-5.18350184e-01 8.33158612e-01 -6.48624778e-01 6.60910726e-01
-2.45644465e-01 2.05798209e-01 1.67311323e+00 -8.68613780e-01
-5.60827032e-02 9.27334726e-01 8.80327821e-01 -1.07740089e-01
-1.21751630e+00 -9.42243099e-01 1.70340791e-01 -3.40915084e-01
-7.89113879e-01 -1.52165323e-01 3.26059282e-01 -9.38993692e-01
8.03902984e-01 -1.75875083e-01 6.84249580e-01 7.40963042e-01
9.51733232e-01 3.18799287e-01 1.02396894e+00 -1.84093658e-02
4.57779467e-01 -4.99073938e-02 2.56578773e-02 -1.08604588e-01
7.76087821e-01 6.81318939e-01 -3.81800860e-01 -3.42018753e-01
8.99932504e-01 1.61923856e-01 2.08950594e-01 6.63214549e-02
-8.37906957e-01 1.03668904e+00 -2.60004580e-01 -2.65804648e-01
-8.62477899e-01 3.23869377e-01 7.82570392e-02 4.70230818e-01
9.11442995e-01 4.02875215e-01 -1.12863934e+00 -3.83242190e-01
-9.96919334e-01 4.91150528e-01 1.11983097e+00 5.48776150e-01
1.68719932e-01 4.72732723e-01 -3.02921623e-01 2.01705486e-01
5.68712115e-01 8.47530961e-01 2.25315943e-01 -1.05182970e+00
8.26404035e-01 2.36588508e-01 5.05491138e-01 -7.58056581e-01
-1.68248489e-01 -7.36630201e-01 -4.00682539e-01 4.74486500e-01
4.68113959e-01 -8.32776546e-01 -2.88188726e-01 1.38208938e+00
-3.67941886e-01 2.36561269e-01 1.79824367e-01 1.18934155e-01
9.49894486e-04 7.50170231e-01 2.27095664e-01 -5.05822420e-01
1.21604395e+00 -6.69907033e-01 -2.88633347e-01 -1.32234141e-01
1.28219813e-01 -2.31723934e-01 1.14882573e-01 5.40192068e-01
-1.24317873e+00 -3.07631791e-01 -6.60029054e-01 9.27722931e-01
1.39497928e-02 -3.61363471e-01 6.97859347e-01 7.03207135e-01
-6.43095255e-01 1.05823004e+00 -7.93801904e-01 1.01663029e+00
6.10909462e-01 2.01597780e-01 2.75120527e-01 6.95974052e-01
-9.38080549e-01 1.00904036e+00 4.13933456e-01 -2.10381716e-01
-7.96925962e-01 -1.26473844e+00 -4.07574207e-01 4.44792181e-01
3.46575588e-01 -6.26548648e-01 1.59544218e+00 -7.72096753e-01
-1.52021551e+00 -1.20399989e-01 3.34915102e-01 -1.03472865e+00
7.14313686e-01 -2.72235900e-01 -4.28767323e-01 -2.31823251e-01
-7.63658583e-02 -2.72558659e-01 7.86688268e-01 -7.36179173e-01
-1.11785090e+00 -3.88640106e-01 -4.61649537e-01 -5.18887788e-02
8.43553469e-02 6.63956404e-02 4.77669775e-01 -1.21135962e+00
1.59019586e-02 -6.03207886e-01 -4.55530226e-01 -8.38131368e-01
-1.51432082e-01 -1.75763533e-01 -1.62756830e-01 -9.17484045e-01
1.26137447e+00 -1.76846611e+00 -5.55830956e-01 6.63695872e-01
-9.23888236e-02 -6.35897696e-01 5.62474012e-01 3.23434412e-01
-3.91203314e-01 2.55325109e-01 -1.27154216e-01 -2.29914263e-01
5.59432566e-01 -3.48708868e-01 -9.00066078e-01 5.06815612e-01
1.60084873e-01 1.04954076e+00 -4.82300729e-01 2.01572001e-01
-4.68634181e-02 -2.11861596e-01 -2.60337740e-01 8.33906233e-02
-6.35153055e-02 1.50225461e-01 -3.98901075e-01 5.73693335e-01
5.64406812e-01 -9.63175669e-02 -1.03980727e-01 5.46629786e-01
-4.48435724e-01 4.70027804e-01 -1.17255509e+00 4.91319537e-01
-4.45779800e-01 2.59002656e-01 -4.20113981e-01 -9.13302124e-01
9.85181153e-01 2.18246013e-01 1.56968653e-01 -3.58977765e-01
7.57258907e-02 3.63407522e-01 3.98158021e-02 2.42315501e-01
2.64727712e-01 -8.34105551e-01 -2.86887735e-01 6.99084342e-01
-2.27610674e-02 -9.30645615e-02 -7.89546315e-03 -2.79813528e-01
7.28458226e-01 2.06623182e-01 4.97533560e-01 -6.93876326e-01
-1.25558704e-01 -5.59236407e-01 7.91443050e-01 9.26303864e-01
2.24182189e-01 -7.65108911e-04 9.15562212e-01 -2.06067532e-01
-8.57043087e-01 -1.49352169e+00 -1.94574371e-01 1.08284080e+00
-7.20418990e-01 2.94603437e-01 -5.38145937e-02 -4.98845071e-01
9.84563828e-01 1.32319164e+00 -6.45395577e-01 3.18248540e-01
1.44154146e-01 -1.17776763e+00 -6.53664097e-02 8.18044901e-01
2.04968721e-01 -8.44837189e-01 -5.60558736e-01 3.86140913e-01
6.55284405e-01 -4.93697464e-01 -2.02396750e-01 3.06400865e-01
-1.02765357e+00 -7.88072884e-01 -7.02049553e-01 5.69744669e-02
9.46997181e-02 -5.10216892e-01 1.43621826e+00 -1.04692864e+00
5.51370025e-01 3.17331731e-01 1.33803561e-01 -1.49747109e+00
-2.00373344e-02 -3.44899088e-01 7.72282183e-02 1.38809890e-01
4.67060030e-01 -6.59051359e-01 -5.65554619e-01 -1.33589387e-01
-1.61334574e-01 -5.35466075e-01 5.17910361e-01 6.20170832e-01
4.20417130e-01 5.47041953e-01 1.12296855e+00 -3.36468250e-01
6.61185503e-01 -7.88865268e-01 -1.61906207e+00 3.01586807e-01
-1.12362659e+00 1.99664518e-01 -6.16405271e-02 -3.48297060e-01
-1.53570056e+00 -3.49243611e-01 3.23572338e-01 -4.97534946e-02
4.76362705e-01 9.68701124e-01 1.79733723e-01 2.51769930e-01
-1.12620816e-01 2.28339911e-01 7.56516075e-03 -6.46837831e-01
-2.74739772e-01 2.91694134e-01 6.22188985e-01 -6.20409071e-01
9.47858691e-01 3.28688145e-01 2.49426767e-01 -2.73324847e-02
-7.86579669e-01 -1.05319299e-01 -2.27621555e-01 -1.23932213e-02
6.01281524e-01 -1.30510104e+00 -1.03856504e+00 3.72623920e-01
-6.19077265e-01 -3.26336801e-01 -5.33674419e-01 1.06464100e+00
-7.85797834e-01 -1.41228765e-01 -8.67420852e-01 -1.81053865e+00
-5.91729403e-01 -7.40399003e-01 5.34633338e-01 2.67461598e-01
-1.43513709e-01 -1.39509106e+00 2.39531532e-01 1.03660636e-01
5.45736432e-01 4.68570471e-01 9.15106714e-01 -1.23987949e+00
-5.32208204e-01 -5.03672004e-01 -1.56128094e-01 5.35783350e-01
-1.86377615e-01 1.02718264e-01 -5.67998886e-01 1.34386821e-02
3.68637919e-01 9.99280810e-02 9.85126317e-01 1.08496094e+00
9.97332931e-01 -4.09206450e-01 8.62783715e-02 3.55323583e-01
1.53861845e+00 6.30508780e-01 5.41541338e-01 7.11460650e-01
-1.70664653e-01 8.34291697e-01 3.98707688e-01 1.08242035e+00
2.47451380e-01 -1.69885084e-01 1.43157586e-01 7.76917398e-01
1.05919039e+00 -5.18326163e-01 5.48172712e-01 4.26325738e-01
-5.83718538e-01 1.52422920e-01 -9.29571152e-01 4.12297308e-01
-1.65978968e+00 -1.41003466e+00 1.82142764e-01 2.32331371e+00
5.95225394e-01 6.94317758e-01 5.34509778e-01 -1.64673567e-01
4.77766514e-01 9.31793824e-02 -4.91759241e-01 -5.28086610e-02
-1.86707363e-01 4.47919250e-01 1.20582747e+00 3.09645265e-01
-1.16244876e+00 2.45587796e-01 7.12928247e+00 8.74625206e-01
-4.18977290e-01 -1.79065615e-01 1.14320171e+00 -1.51649877e-01
-6.13766432e-01 3.29914019e-02 -1.24326527e+00 8.93097997e-01
1.32185578e+00 -9.39295292e-01 3.95290367e-02 1.11474025e+00
1.44354492e-01 -1.13471024e-01 -1.04076326e+00 6.55135334e-01
-6.19442821e-01 -1.63532352e+00 -3.55900228e-01 5.39504051e-01
1.01592624e+00 7.61765689e-02 6.14773035e-01 6.98232472e-01
9.35516953e-01 -1.27688491e+00 1.07826197e+00 1.31305218e+00
3.98001611e-01 -1.36833620e+00 1.18821812e+00 4.09687877e-01
-9.47250903e-01 -6.39480293e-01 -2.22420454e-01 -4.20393199e-01
2.51764715e-01 9.38957334e-01 -2.22464398e-01 4.70791012e-01
7.33415723e-01 5.12528181e-01 -3.72980647e-02 9.07960296e-01
-5.23927482e-03 8.29634190e-01 -3.83178800e-01 -3.34106833e-02
3.73313785e-01 -7.51540005e-01 1.24071978e-01 1.02258027e+00
9.17665005e-01 2.28128895e-01 -4.91820753e-01 9.61525202e-01
-1.71175599e-01 -1.08300015e-01 -3.91760349e-01 -1.61214888e-01
4.27186936e-01 6.52326584e-01 -2.59400278e-01 -2.16399908e-01
-7.85370648e-01 -7.21992552e-02 -3.31669599e-01 2.76656061e-01
-5.29670835e-01 -4.02001947e-01 5.22288620e-01 -4.09030542e-02
8.14653814e-01 -2.08979532e-01 -8.39782655e-01 -1.02257729e+00
-1.18215725e-01 -6.01127565e-01 3.75642031e-01 -3.36432904e-01
-1.80688071e+00 -2.08365902e-01 2.00278655e-01 -9.92335081e-01
-4.91429389e-01 -8.93970311e-01 -1.09456527e+00 1.29121602e+00
-1.62284589e+00 -1.47822827e-01 7.08893716e-01 2.88668685e-02
2.60496855e-01 -9.27576363e-01 3.99540424e-01 -4.88964409e-01
-5.74316680e-01 1.75326645e-01 9.27626252e-01 3.79945755e-01
1.58151329e-01 -1.59305143e+00 7.73716509e-01 6.76210463e-01
-7.33964071e-02 5.89817345e-01 5.60926855e-01 -9.79604959e-01
-9.37563002e-01 -9.22571838e-01 5.85083604e-01 -7.31743276e-01
1.42644155e+00 3.10654253e-01 -5.68711400e-01 1.10606015e+00
1.50003508e-01 -2.54068613e-01 9.33383584e-01 1.39284417e-01
-2.52756774e-01 -4.16190445e-01 -1.03412294e+00 4.21030223e-01
2.23086596e-01 -2.73441643e-01 -1.02215910e+00 -4.80165333e-02
5.12932062e-01 2.31841952e-01 -1.44624138e+00 3.89735520e-01
7.82442331e-01 -1.00775301e+00 8.32245588e-01 -8.21070969e-01
3.74381006e-01 3.74544591e-01 -2.75123775e-01 -1.42362857e+00
-6.58438623e-01 -8.79944563e-01 -2.75911868e-01 1.23450673e+00
8.39101791e-01 -1.23017943e+00 7.99831092e-01 9.54345703e-01
2.05434591e-01 -8.41616094e-01 -1.32560432e+00 -1.05360329e+00
7.88591266e-01 -7.39541590e-01 8.84474039e-01 5.44158757e-01
-1.12451941e-01 -1.75688699e-01 -1.26015693e-01 -6.21219389e-02
1.27943146e+00 4.21391994e-01 3.91349614e-01 -1.90161526e+00
-5.55203378e-01 -8.88943851e-01 9.09155607e-02 -8.00944269e-01
4.42533404e-01 -5.70483029e-01 -3.49327773e-01 -8.75733197e-01
2.30735406e-01 -2.79410809e-01 -5.69297791e-01 -6.30469173e-02
-5.74020483e-02 -3.74563754e-01 1.57490104e-01 7.49546895e-03
-2.58055001e-01 5.33460438e-01 6.85597181e-01 1.61900237e-01
-5.31704202e-02 7.46963561e-01 -1.02164841e+00 1.08608544e+00
1.15046775e+00 -2.59278536e-01 -3.54274958e-02 2.87419498e-01
6.30686462e-01 5.39944947e-01 3.53797168e-01 -4.72657621e-01
-9.93783548e-02 -7.48175204e-01 7.70907938e-01 -9.06298161e-01
-1.17471106e-01 -4.04840022e-01 3.21365654e-01 5.80054283e-01
-2.76931882e-01 5.09802103e-01 1.79259539e-01 8.86655211e-01
-9.88648385e-02 -4.22381133e-01 5.84344268e-01 -3.63646597e-01
-1.70433223e-01 3.32327515e-01 -6.96158230e-01 1.95846140e-01
1.30531645e+00 6.80109486e-02 -1.27474293e-01 -8.43076825e-01
-3.90637279e-01 3.62663627e-01 7.55438581e-02 -5.52476905e-02
2.90802330e-01 -1.09589779e+00 -1.11024821e+00 -7.47162700e-02
-2.85271496e-01 -3.08020830e-01 3.54651967e-03 5.96897781e-01
-2.30874196e-01 8.08014810e-01 1.35383949e-01 2.11805820e-01
-9.31639969e-02 2.43527412e-01 4.62834686e-01 -5.91658115e-01
-1.51859477e-01 1.07362425e+00 2.51474977e-01 1.92725763e-01
7.22817928e-02 -3.60017687e-01 -7.45664984e-02 4.82237786e-01
6.98640943e-01 8.11377347e-01 -2.64787138e-01 -2.18635738e-01
-4.06212695e-02 3.08701307e-01 2.47565314e-01 -5.62609315e-01
1.72059071e+00 3.66856605e-02 -5.03636636e-02 8.26771796e-01
7.64549494e-01 2.15064585e-02 -1.89555991e+00 -8.66860822e-02
8.41165900e-01 -4.20085967e-01 1.25476509e-01 -6.55809164e-01
-1.09427261e+00 7.23827004e-01 -2.52273679e-02 2.49186695e-01
6.12620592e-01 -1.98472098e-01 3.94601822e-01 4.15332735e-01
4.26646054e-01 -1.29205799e+00 -8.36165249e-02 3.57874900e-01
7.74613380e-01 -1.20750940e+00 2.73057908e-01 3.67872506e-01
-9.35202956e-01 1.08469558e+00 -1.38771668e-01 -6.17020428e-01
1.37394619e+00 3.50614727e-01 -4.99325037e-01 1.49852484e-01
-1.12416017e+00 3.59843820e-01 5.41616797e-01 4.88844752e-01
1.27322868e-01 3.58228356e-01 -4.75085229e-02 1.44854844e+00
-6.66061044e-01 -4.35533226e-01 5.80981314e-01 6.25967562e-01
-6.72600389e-01 -5.75564146e-01 -4.81206894e-01 1.08294535e+00
-1.28342295e+00 -3.45940977e-01 2.64761418e-01 7.43955731e-01
-2.83618420e-01 6.28356874e-01 7.44612396e-01 1.74153280e-02
2.43978620e-01 5.10219336e-01 1.29625514e-01 -4.33966398e-01
-3.99864852e-01 3.00944448e-01 2.86357594e-03 -4.90317605e-02
-5.52512370e-02 -1.36178577e+00 -6.51278198e-01 -5.84232271e-01
-2.14409769e-01 6.40069172e-02 3.72276664e-01 9.64534640e-01
-8.19862485e-02 1.25125289e-01 9.95544016e-01 -8.79768491e-01
-1.63380265e+00 -9.77400064e-01 -1.36429095e+00 -3.18903476e-01
3.68236333e-01 -6.55576646e-01 -9.78776157e-01 -2.47504875e-01] | [4.842289447784424, 4.071173667907715] |
27c2f982-2e2d-4584-a8d1-6f369db329e8 | gini-regularized-optimal-transport-with-an | 1712.02512 | null | http://arxiv.org/abs/1712.02512v1 | http://arxiv.org/pdf/1712.02512v1.pdf | Gini-regularized Optimal Transport with an Application to Spatio-Temporal Forecasting | Rapidly growing product lines and services require a finer-granularity
forecast that considers geographic locales. However the open question remains,
how to assess the quality of a spatio-temporal forecast? In this manuscript we
introduce a metric to evaluate spatio-temporal forecasts. This metric is based
on an Opti- mal Transport (OT) problem. The metric we propose is a constrained
OT objec- tive function using the Gini impurity function as a regularizer. We
demonstrate through computer experiments both the qualitative and the
quantitative charac- teristics of the Gini regularized OT problem. Moreover, we
show that the Gini regularized OT problem converges to the classical OT
problem, when the Gini regularized problem is considered as a function of
{\lambda}, the regularization parame-ter. The convergence to the classical OT
solution is faster than the state-of-the-art Entropic-regularized OT[Cuturi,
2013] and results in a numerically more stable algorithm. | ['Leo Razoumov', 'Lin Su', 'Yuyang Wang', 'Lucas Roberts'] | 2017-12-07 | null | null | null | null | ['spatio-temporal-forecasting'] | ['time-series'] | [-3.73132795e-01 -6.13769926e-02 1.28193215e-01 -2.19761655e-01
-6.82168067e-01 -3.50606501e-01 6.90033674e-01 2.26524234e-01
-3.03726673e-01 1.00649226e+00 4.49103564e-02 -3.56731176e-01
-4.97028738e-01 -9.81025517e-01 -4.10243750e-01 -1.06246781e+00
-3.29265058e-01 7.86618114e-01 2.28294834e-01 -5.45561671e-01
2.96143293e-01 6.59578383e-01 -1.44483769e+00 1.91346408e-04
1.53422475e+00 1.39700472e+00 -2.68274009e-01 4.05483931e-01
2.58329451e-01 3.42939109e-01 -1.91743031e-01 -2.77757734e-01
3.02828103e-01 -3.84930253e-01 -9.49780941e-01 6.93258941e-02
4.60965447e-02 4.25666928e-01 3.94004852e-01 1.06478953e+00
4.55227643e-02 7.28183925e-01 1.04601550e+00 -1.21860158e+00
-3.71937394e-01 3.02538276e-01 -4.31184888e-01 3.29199761e-01
-6.57572523e-02 5.77608822e-03 9.37078834e-01 -7.69669294e-01
7.11014986e-01 9.88700390e-01 7.08993852e-01 3.50071900e-02
-1.50408518e+00 -5.73374741e-02 1.91545337e-01 -1.54663175e-01
-1.63686442e+00 3.36741418e-01 2.00909689e-01 -7.73663342e-01
7.24038720e-01 6.04556203e-01 4.62371588e-01 2.57035345e-01
5.79418004e-01 2.21691906e-01 1.29598796e+00 -9.51273218e-02
4.34682250e-01 2.68764138e-01 4.02295701e-02 4.35279399e-01
2.10301861e-01 1.99190363e-01 -2.14273587e-01 -2.51726389e-01
2.65229076e-01 -3.71926874e-01 -2.85279900e-01 -1.01211898e-01
-1.02976024e+00 9.65894997e-01 5.14264166e-01 6.16442621e-01
-6.61429167e-01 4.82350588e-02 2.98047394e-01 2.54076928e-01
1.36903667e+00 4.85760272e-01 -3.60403508e-01 -2.71655589e-01
-1.36902606e+00 6.32498324e-01 8.50607634e-01 5.36127329e-01
7.81561375e-01 3.20400387e-01 -9.02647972e-02 3.36134493e-01
2.11780772e-01 7.59805322e-01 -2.65193079e-02 -9.13192391e-01
4.32331681e-01 2.52055615e-01 4.26965863e-01 -1.01603925e+00
-5.76520383e-01 -4.37744915e-01 -9.50708091e-01 3.09302151e-01
6.95178390e-01 -3.14050674e-01 -5.13264537e-01 1.24388576e+00
5.52418828e-01 1.70001939e-01 -1.53355330e-01 1.03100491e+00
1.40078425e-01 1.14852345e+00 -1.31631568e-01 -4.61262792e-01
8.45673680e-01 -7.89974988e-01 -5.82832992e-01 5.65359950e-01
9.67941225e-01 -7.55440354e-01 6.95718467e-01 4.73031849e-01
-1.00240064e+00 -1.18017465e-01 -6.65217459e-01 4.32832688e-01
-7.04006493e-01 -4.63582128e-02 4.77467805e-01 5.70344627e-01
-1.29778576e+00 1.10973430e+00 -7.37979710e-01 -4.20227647e-01
-2.11709976e-01 1.86741337e-01 -1.20733336e-01 3.94382328e-01
-1.08295870e+00 1.10863400e+00 1.99702397e-01 2.15644762e-01
-5.43921411e-01 -8.92810225e-01 -7.30885148e-01 7.62908757e-02
1.45734847e-01 -2.83036232e-01 9.36290681e-01 -7.80607462e-01
-1.54905665e+00 5.22709787e-01 -6.76254109e-02 -3.48132163e-01
6.36482596e-01 3.84628385e-01 -5.30906498e-01 -6.33661523e-02
3.09604835e-02 7.84076750e-02 5.70583463e-01 -1.24317825e+00
-5.80806732e-01 -1.25785068e-01 -1.38147339e-01 -4.52948287e-02
4.08454895e-01 1.69580579e-01 2.17930660e-01 -6.51488125e-01
-1.93416663e-02 -1.09023476e+00 -3.10535342e-01 -3.18348676e-01
-3.10351342e-01 -1.33119702e-01 6.40589118e-01 -7.72002280e-01
1.26247573e+00 -1.86433959e+00 3.47190708e-01 6.32994413e-01
-1.82382464e-02 5.19711114e-02 1.95319027e-01 6.14812851e-01
-7.64307380e-02 2.90436059e-01 -7.27946639e-01 -2.65974075e-01
1.45235971e-01 1.02199726e-01 -2.11490288e-01 8.71955872e-01
-2.09407808e-04 5.53204954e-01 -8.80524993e-01 -9.13622752e-02
2.48113990e-01 4.56545591e-01 -4.40961391e-01 -3.09931040e-01
-2.76858062e-01 6.57953501e-01 -4.44488823e-01 3.07854712e-01
1.08524263e+00 -1.22399271e-01 -2.01296926e-01 2.07855240e-01
-8.40476811e-01 1.73388332e-01 -1.04052567e+00 8.86656642e-01
-5.85274160e-01 5.76241255e-01 3.18618596e-01 -1.05296302e+00
1.03250706e+00 2.58255720e-01 7.19225466e-01 -6.46151662e-01
3.41330469e-02 6.12687707e-01 -2.89688289e-01 -9.84363332e-02
7.69683242e-01 -3.85897934e-01 2.32477449e-02 4.88916993e-01
-3.06488544e-01 -3.58867764e-01 3.16762507e-01 8.19553584e-02
5.10500729e-01 7.88820833e-02 -1.72236711e-02 -1.22105753e+00
5.58734715e-01 1.43599227e-01 3.48173827e-01 3.17280799e-01
1.19290343e-02 6.19312525e-01 8.03549647e-01 -6.86538458e-01
-1.03841269e+00 -8.00283790e-01 -5.59857070e-01 7.76087284e-01
1.03371948e-01 -4.15146947e-01 -9.84969079e-01 -5.35734892e-01
2.04791725e-01 1.00415921e+00 -7.35084176e-01 1.97329164e-01
-3.30367774e-01 -1.08812940e+00 1.73118457e-01 3.39463651e-02
3.50746661e-01 -6.64092422e-01 -4.01264340e-01 1.24469511e-01
-2.39894643e-01 -8.56169522e-01 -5.29037714e-01 7.75175691e-02
-8.83631110e-01 -5.24161398e-01 -8.54519486e-01 1.30192518e-01
4.80301142e-01 -2.57961631e-01 1.02420306e+00 -2.01488346e-01
2.32759520e-01 6.52142242e-02 -2.40468517e-01 -2.65631527e-02
-3.80489737e-01 1.56739280e-01 2.53800571e-01 4.75370437e-01
-3.90307568e-02 -4.18396235e-01 -3.88167799e-01 7.43313611e-01
-8.53625655e-01 -2.31393412e-01 -2.78515548e-01 4.96949852e-01
6.55068159e-01 1.16896279e-01 1.96684077e-01 -6.11331522e-01
6.24780118e-01 -5.97438574e-01 -1.27125669e+00 2.21876532e-01
-8.90236974e-01 3.65892023e-01 6.07093275e-01 1.23478081e-02
-1.04264188e+00 -3.09726059e-01 1.54739141e-03 -5.93813509e-02
1.72046587e-01 8.40757668e-01 4.15859878e-01 -4.30965424e-01
5.58353961e-01 -1.17870025e-01 -3.18618953e-01 -4.97496188e-01
2.30233610e-01 3.71830076e-01 -3.86411175e-02 -7.50312626e-01
6.06985629e-01 6.07604504e-01 4.19985592e-01 -7.73707151e-01
-6.55588090e-01 -4.52934951e-01 -6.46779001e-01 -5.80874860e-01
1.04108119e+00 -3.67380530e-01 -9.73222017e-01 2.58082777e-01
-1.27195883e+00 -5.56272864e-01 -6.07464612e-01 6.68299258e-01
-7.48171091e-01 -2.24540811e-02 -4.94056016e-01 -1.23553360e+00
4.54929024e-02 -1.23697221e+00 1.17444754e+00 -1.16898052e-01
-1.13042071e-02 -1.54179204e+00 4.15360898e-01 8.28546733e-02
7.73176670e-01 6.96470320e-01 6.44772649e-01 -7.13384077e-02
-5.16399980e-01 -7.71064684e-03 -3.87263447e-01 1.87717974e-01
-1.39886498e-01 3.11302841e-01 -7.80440152e-01 -7.49469846e-02
7.15420544e-02 3.15637827e-01 8.60782802e-01 7.53428936e-01
8.20853591e-01 -4.35532838e-01 -2.34170053e-02 7.77168512e-01
1.59970677e+00 1.96403041e-02 5.23997545e-01 2.45770916e-01
4.81663138e-01 8.50645781e-01 8.67732406e-01 5.71159720e-01
2.73799062e-01 9.29480970e-01 2.74618030e-01 -1.07784513e-02
4.84287798e-01 -2.20848923e-03 1.68050647e-01 1.00636852e+00
-1.00294852e+00 -2.32234970e-01 -1.26336527e+00 4.39099669e-01
-2.28943682e+00 -9.34125841e-01 -5.74546337e-01 2.31767344e+00
4.09933627e-01 -1.67968497e-01 2.93337584e-01 -1.14717156e-01
4.98721302e-01 1.41920790e-01 7.68903568e-02 -9.31991875e-01
-1.61266744e-01 -7.15814680e-02 1.09907532e+00 1.14464879e+00
-1.05498374e+00 7.42818534e-01 6.64491081e+00 9.23552871e-01
-1.40365744e+00 3.22372526e-01 6.38062537e-01 1.73617914e-01
-4.00882810e-01 -6.86436817e-02 -7.39237368e-01 5.56833267e-01
1.42772305e+00 -4.85446841e-01 4.81627434e-01 5.44209123e-01
7.54566967e-01 -4.66855347e-01 -5.44263482e-01 6.06287301e-01
-1.44705996e-01 -1.44084811e+00 -2.96563596e-01 6.83994591e-01
1.18530619e+00 2.11307541e-01 2.36829087e-01 2.70671137e-02
3.76914293e-01 -1.05186343e+00 7.10682809e-01 1.03077471e+00
7.60287642e-01 -9.54689503e-01 8.83765280e-01 3.53579491e-01
-1.50146246e+00 3.83860618e-01 -1.82688802e-01 -4.62768562e-02
5.89336693e-01 7.68454790e-01 -1.88571930e-01 1.07674277e+00
6.89659417e-01 7.11650252e-01 -7.62607008e-02 1.04666114e+00
3.69226992e-01 6.16397381e-01 -6.26902759e-01 9.12295729e-02
9.24922407e-01 -1.10397959e+00 6.91870868e-01 1.29233658e+00
6.43168628e-01 1.11771956e-01 5.47910593e-02 1.06419015e+00
3.66728425e-01 2.60714829e-01 -5.94670475e-01 -1.05834957e-02
-3.62185776e-01 1.04470396e+00 -8.62542570e-01 -4.58316132e-02
2.12602802e-02 7.05840349e-01 1.62355937e-02 7.91891456e-01
-8.09390545e-01 -8.67652521e-02 7.05613077e-01 4.40149069e-01
2.89070308e-01 -3.39591861e-01 -4.15283799e-01 -1.16778505e+00
-3.47975269e-02 -5.39760590e-01 2.41305545e-01 -6.49310052e-01
-1.11914158e+00 8.41318905e-01 4.68758821e-01 -1.24287617e+00
-2.56748974e-01 -7.36740828e-01 -5.59210837e-01 1.18864751e+00
-1.33321500e+00 -7.31252313e-01 1.01997107e-01 3.42224211e-01
1.82966907e-02 1.67512015e-01 5.79457939e-01 2.26602361e-01
-4.43464816e-01 2.89128311e-02 7.15026855e-01 -5.57284892e-01
-1.60610682e-04 -1.26684046e+00 2.39815637e-01 7.86803007e-01
-3.79145920e-01 3.18128198e-01 1.10892439e+00 -6.53177500e-01
-7.49406755e-01 -1.16735995e+00 1.27998161e+00 -3.80941421e-01
9.42051530e-01 -2.83574462e-01 -1.00559449e+00 5.98814309e-01
1.48517594e-01 2.43275851e-01 1.16257608e-01 6.45808950e-02
1.84672531e-02 -2.17783585e-01 -1.36260450e+00 2.36140326e-01
4.26905900e-01 -4.04812783e-01 1.25966081e-02 6.27362430e-01
6.28679037e-01 -2.41316542e-01 -1.15652263e+00 3.52667123e-01
1.47961438e-01 -1.15203130e+00 5.71795702e-01 -4.25806761e-01
-6.63786521e-03 -3.53331566e-01 -1.46984115e-01 -1.58484316e+00
-1.58023685e-01 -9.21078920e-01 3.42371523e-01 9.05341744e-01
8.50554824e-01 -1.01999080e+00 4.00241613e-01 6.87738121e-01
-5.43788224e-02 -9.63730931e-01 -1.41499364e+00 -1.25326645e+00
5.41793227e-01 -6.48480594e-01 5.49329400e-01 9.16505873e-01
9.63573232e-02 -2.01260448e-01 -3.25743824e-01 2.36984923e-01
6.29728198e-01 -4.23354730e-02 7.96999857e-02 -1.44599223e+00
1.34314775e-01 -5.41246951e-01 -4.01216775e-01 -5.63697338e-01
1.63111567e-01 -8.08248341e-01 2.68189330e-02 -1.40614879e+00
-4.58376050e-01 -7.88449466e-01 1.24583365e-02 -1.03663333e-01
3.89383703e-01 2.76226997e-02 2.08056852e-01 -1.00253401e-02
-4.59747225e-01 8.19375634e-01 1.26950824e+00 -1.60336550e-02
-2.68403888e-01 2.13443622e-01 9.35885832e-02 6.08623564e-01
7.80752003e-01 -5.76915503e-01 -1.06752828e-01 3.34971286e-02
5.68627298e-01 3.78219932e-02 2.38405183e-01 -7.53147900e-01
5.69422096e-02 -5.53073406e-01 -5.28337479e-01 -7.24853158e-01
3.73088926e-01 -5.95233738e-01 3.23386729e-01 1.47679642e-01
-6.82336688e-02 2.55928546e-01 3.19991112e-01 3.02883416e-01
-4.75179076e-01 -2.33175203e-01 9.87684786e-01 1.39396250e-01
-7.79798329e-02 4.46698964e-01 -6.21434689e-01 3.03591907e-01
9.51582730e-01 -4.67616953e-02 -3.68992761e-02 -5.82515895e-01
-7.41299689e-01 6.13137335e-02 4.82667267e-01 -2.52931356e-01
1.31186873e-01 -1.35797155e+00 -7.79048026e-01 -2.13889167e-01
-1.03736252e-01 -4.81142431e-01 1.47923112e-01 1.56016660e+00
-1.03223491e+00 7.35542238e-01 2.25914285e-01 -5.62790334e-01
-8.02690744e-01 4.37503278e-01 8.63537848e-01 -5.26656568e-01
-3.86807084e-01 4.78757322e-01 2.47406751e-01 -7.23305345e-01
-1.54248446e-01 -9.08252358e-01 2.13316143e-01 1.48039728e-01
1.69904441e-01 8.24949503e-01 2.72796750e-01 -1.07101095e+00
-3.77728283e-01 8.04226279e-01 6.02529824e-01 -5.99169791e-01
1.29275298e+00 -2.73281485e-01 -3.22018623e-01 7.24297166e-01
1.09541392e+00 -1.52766421e-01 -1.19016325e+00 1.24360196e-01
1.25213340e-01 -4.37101662e-01 3.22146535e-01 -6.06192946e-01
-1.13928974e+00 8.34886491e-01 3.11158210e-01 8.09486568e-01
7.65433133e-01 -2.97710061e-01 5.58608651e-01 5.23604527e-02
4.94048774e-01 -1.36953902e+00 -6.94076896e-01 8.54097843e-01
1.07189512e+00 -1.11141264e+00 2.17212215e-02 -6.18741095e-01
-8.90072763e-01 8.88614655e-01 -9.87569103e-04 -2.07591236e-01
1.27568805e+00 1.77189633e-01 -7.90881068e-02 -2.97751635e-01
-5.64375043e-01 -5.32222271e-01 6.36350095e-01 4.18744951e-01
3.94228667e-01 3.94569039e-01 -6.21353209e-01 2.92694457e-02
-2.21595287e-01 -2.19011992e-01 4.07299906e-01 5.02408206e-01
-2.11418539e-01 -6.44858539e-01 -4.04126972e-01 1.01230487e-01
-2.80655026e-01 -1.81532845e-01 -1.13721788e-01 8.40540946e-01
1.56293795e-01 9.90609586e-01 1.35986745e-01 -1.04611635e-01
2.22429425e-01 -1.00665018e-01 1.41658798e-01 -3.16236675e-01
-7.36095130e-01 -5.56002185e-02 8.83321017e-02 -7.08731294e-01
-5.11499107e-01 -8.67113888e-01 -1.16685462e+00 -9.38489318e-01
-3.83365393e-01 6.29504502e-01 8.44578087e-01 1.26960731e+00
2.35782608e-01 4.11122829e-01 7.13721931e-01 -1.06889689e+00
-1.83153003e-01 -7.45682657e-01 -1.12742603e+00 6.79638982e-02
4.74652141e-01 -7.50247180e-01 -7.63508916e-01 -3.22602391e-01] | [6.427553176879883, 3.8280978202819824] |
ac676919-d1f3-455e-b8ae-21a6b89b8f6d | mixup-of-feature-maps-in-a-hidden-layer-for | 1906.09739 | null | https://arxiv.org/abs/1906.09739v1 | https://arxiv.org/pdf/1906.09739v1.pdf | Mixup of Feature Maps in a Hidden Layer for Training of Convolutional Neural Network | The deep Convolutional Neural Network (CNN) became very popular as a fundamental technique for image classification and objects recognition. To improve the recognition accuracy for the more complex tasks, deeper networks have being introduced. However, the recognition accuracy of the trained deep CNN drastically decreases for the samples which are obtained from the outside regions of the training samples. To improve the generalization ability for such samples, Krizhevsky et al. proposed to generate additional samples through transformations from the existing samples and to make the training samples richer. This method is known as data augmentation. Hongyi Zhang et al. introduced data augmentation method called mixup which achieves state-of-the-art performance in various datasets. Mixup generates new samples by mixing two different training samples. Mixing of the two images is implemented with simple image morphing. In this paper, we propose to apply mixup to the feature maps in a hidden layer. To implement the mixup in the hidden layer we use the Siamese network or the triplet network architecture to mix feature maps. From the experimental comparison, it is observed that the mixup of the feature maps obtained from the first convolution layer is more effective than the original image mixup. | ['Takio Kurita', 'Hideki Oki'] | 2019-06-24 | null | null | null | null | ['image-morphing'] | ['computer-vision'] | [ 2.23792166e-01 -1.29428580e-01 1.72965080e-02 -3.90823066e-01
-3.93296778e-02 2.83268951e-02 5.74189246e-01 -1.39876842e-01
-7.20156610e-01 5.55097997e-01 -6.14877157e-02 2.52219498e-01
2.62824774e-01 -1.10474575e+00 -7.63797879e-01 -7.61900842e-01
3.53509694e-01 2.04988346e-01 1.61754847e-01 -1.39871895e-01
1.98973015e-01 8.13087940e-01 -1.66855335e+00 5.44951320e-01
1.02160680e+00 1.17017543e+00 1.79007873e-01 3.45879138e-01
-5.62843561e-01 5.18991947e-01 -6.34508908e-01 -1.65279478e-01
7.25166857e-01 -5.47117949e-01 -4.52766955e-01 1.96381688e-01
4.03511882e-01 -1.92299470e-01 -4.69710439e-01 1.20033562e+00
2.34865949e-01 1.39423177e-01 4.25630212e-01 -1.43504584e+00
-8.57450724e-01 5.33742964e-01 -6.00484014e-01 6.87474832e-02
-3.06831002e-01 -3.65377851e-02 3.56817901e-01 -9.39951837e-01
3.71846080e-01 1.06628251e+00 5.02239168e-01 5.80085933e-01
-9.48441446e-01 -9.26831424e-01 -1.64683446e-01 3.65178108e-01
-1.47225595e+00 -5.85515946e-02 9.01925802e-01 -2.54805386e-01
3.64645988e-01 2.71320820e-01 7.23762631e-01 6.26884401e-01
-9.76566002e-02 9.11264598e-01 1.13737476e+00 -4.54386801e-01
-9.79881510e-02 2.36535117e-01 1.86409473e-01 7.16162682e-01
3.36010337e-01 5.16604260e-03 -1.99603826e-01 1.35614023e-01
9.50876534e-01 5.90020239e-01 -2.29282126e-01 -1.88995495e-01
-1.18823755e+00 7.53506184e-01 1.10946846e+00 6.27817094e-01
-5.33257604e-01 -6.44879192e-02 2.75359750e-01 1.86627984e-01
9.98657718e-02 6.00760162e-01 -3.31494093e-01 2.80818850e-01
-8.95816982e-01 1.05001576e-01 3.47278506e-01 7.74440885e-01
1.24224925e+00 2.20848337e-01 -4.79125641e-02 8.82955134e-01
2.86119077e-02 2.10772842e-01 1.02009952e+00 -5.47013938e-01
4.73377705e-01 1.20504236e+00 -2.81919271e-01 -9.81232822e-01
-3.91128540e-01 -6.83301449e-01 -1.31336832e+00 3.58745933e-01
5.38260102e-01 -1.96993679e-01 -1.32339954e+00 1.54945767e+00
2.44061649e-01 1.73865601e-01 3.74925673e-01 8.76768112e-01
9.15069759e-01 1.06131327e+00 -2.25307509e-01 2.56783694e-01
1.18051767e+00 -1.34055781e+00 -6.71337605e-01 -1.72840521e-01
6.61542118e-01 -7.91000903e-01 1.00385654e+00 3.18694502e-01
-6.73549533e-01 -1.13424182e+00 -1.26943195e+00 4.57167439e-02
-7.84242928e-01 6.36902511e-01 4.26032156e-01 3.48990530e-01
-5.67457795e-01 7.16820955e-01 -7.08629429e-01 -2.03797638e-01
7.32780278e-01 5.21828771e-01 -7.89153516e-01 -2.07831278e-01
-9.86079216e-01 5.99968612e-01 7.46653020e-01 2.71571279e-01
-5.58898032e-01 -4.65694964e-01 -8.84601474e-01 8.56646150e-02
1.36380762e-01 -8.58262479e-02 6.81322277e-01 -1.56258476e+00
-1.41196406e+00 5.46264291e-01 1.68911532e-01 -3.94048005e-01
3.32550555e-01 -6.72845319e-02 -3.51243347e-01 -2.16337964e-01
-1.54781207e-01 8.64350319e-01 8.33848298e-01 -1.02706182e+00
-5.37138224e-01 -2.17542723e-01 -3.33472162e-01 6.33549169e-02
-5.36277831e-01 -2.41558298e-01 -2.45538309e-01 -8.85040045e-01
3.36383760e-01 -9.73697782e-01 -2.28503600e-01 9.79246385e-03
-5.04318655e-01 5.04105464e-02 1.25441444e+00 -5.56609452e-01
7.12421358e-01 -2.37962532e+00 5.69094270e-02 4.01088446e-01
1.71549901e-01 7.51567066e-01 -4.20132935e-01 1.85198605e-01
-3.94262016e-01 1.25485122e-01 -4.40001994e-01 -1.56856433e-01
-4.58011806e-01 3.20594192e-01 1.38383597e-01 2.38897577e-01
4.57025945e-01 8.05951357e-01 -5.45126319e-01 -1.94671422e-01
1.32534996e-01 3.86374056e-01 -6.13483727e-01 3.84076804e-01
5.36763715e-03 5.01726031e-01 -2.25074351e-01 3.45255196e-01
9.91614997e-01 -9.97731164e-02 -3.79082143e-01 -4.27849770e-01
-5.66146970e-02 -2.63725489e-01 -1.23884606e+00 1.44863236e+00
-1.49479583e-01 6.75134122e-01 -3.16097111e-01 -1.15991461e+00
1.41161752e+00 1.19587138e-01 2.86751658e-01 -4.36235130e-01
3.57901067e-01 2.93959290e-01 4.86040086e-01 -5.82581580e-01
4.16138083e-01 -6.90626949e-02 2.05903962e-01 2.40943938e-01
1.06989130e-01 5.67366965e-02 1.82568222e-01 -2.20074862e-01
5.37025988e-01 -3.22009698e-02 3.36431004e-02 -3.39625552e-02
8.20329607e-01 -1.12636119e-01 6.97268069e-01 3.75611961e-01
7.53575191e-02 7.15338290e-01 2.89682776e-01 -7.23098338e-01
-1.22646856e+00 -6.23531163e-01 -3.00778709e-02 4.38901842e-01
-3.81266959e-02 -1.23013012e-01 -9.08033192e-01 -8.27097595e-01
-3.91658209e-02 5.85648455e-02 -8.03245485e-01 -5.29288650e-01
-7.07836211e-01 -8.91702950e-01 5.50854325e-01 5.91593683e-01
1.60723627e+00 -1.16919482e+00 -2.25725368e-01 5.52204661e-02
3.52552272e-02 -1.02176702e+00 -2.80807763e-01 6.49224371e-02
-8.69441450e-01 -9.13620830e-01 -1.03370845e+00 -9.93508637e-01
1.11709070e+00 -1.07800856e-01 2.07126215e-01 2.91402459e-01
-2.43078858e-01 -5.71176589e-01 -5.26181579e-01 -2.63001949e-01
-4.76625860e-01 3.43131870e-01 7.69540900e-03 6.32981062e-01
3.31732213e-01 -3.49373013e-01 -3.84459943e-01 3.45972747e-01
-1.16814470e+00 4.24262464e-01 9.26853836e-01 1.03330505e+00
3.37833136e-01 1.27183706e-01 5.86548209e-01 -7.94919312e-01
3.52470577e-01 -2.05550194e-01 -4.25791025e-01 -4.27572504e-02
-2.74739355e-01 4.98596817e-01 9.51679766e-01 -6.52096808e-01
-7.43573666e-01 1.98270306e-01 -3.57784182e-01 -4.86288041e-01
-2.95536399e-01 4.60633367e-01 -2.47411370e-01 -3.09514701e-01
5.83544075e-01 3.24476033e-01 3.31611782e-01 -7.10840166e-01
1.13932088e-01 8.24793339e-01 4.83760536e-01 -2.80002415e-01
8.43872786e-01 2.23016575e-01 4.62257024e-03 -7.72067606e-01
-5.25053680e-01 -4.30138707e-02 -5.42360961e-01 -6.27817139e-02
9.56482649e-01 -4.36589718e-01 -3.08211893e-01 8.99897516e-01
-1.02593434e+00 -3.34035695e-01 -3.33867252e-01 7.88821220e-01
2.61576977e-02 1.88823909e-01 -5.15565038e-01 -2.93845594e-01
-2.03135252e-01 -1.48230946e+00 4.78888661e-01 6.90368235e-01
2.89138258e-01 -6.31414175e-01 -2.21273616e-01 -1.88982990e-02
5.35036027e-01 3.12796235e-01 8.14075291e-01 -9.38762844e-01
-4.91206527e-01 -5.11891365e-01 -4.14450645e-01 8.87191713e-01
4.86138433e-01 2.16175064e-01 -9.14434731e-01 -1.00525670e-01
-4.72962931e-02 -3.61691713e-02 9.00598645e-01 -6.96004033e-02
1.33391666e+00 -4.00826663e-01 -1.00340977e-01 8.31342697e-01
1.38370764e+00 3.96329135e-01 8.78232181e-01 5.45862794e-01
8.07790518e-01 6.17633522e-01 3.00152063e-01 6.56188428e-02
-1.82231128e-01 4.19549912e-01 3.96919817e-01 -3.74585301e-01
-8.39448944e-02 -1.24647684e-01 4.26634140e-02 7.37742782e-01
-1.45600662e-01 8.15809071e-02 -7.77443230e-01 3.32361042e-01
-1.67490602e+00 -8.71254146e-01 -4.14112695e-02 2.15869904e+00
7.24468410e-01 1.14713706e-01 -1.01001926e-01 4.40787524e-01
9.68812704e-01 -2.53432114e-02 -3.51442873e-01 -2.32947931e-01
-1.95666164e-01 3.41881454e-01 5.18278301e-01 3.33807915e-01
-1.02101135e+00 7.56544232e-01 5.22085619e+00 8.16534579e-01
-1.53563035e+00 -1.00185066e-01 6.62389159e-01 1.93284690e-01
1.20265573e-01 -2.53320754e-01 -7.38298953e-01 5.56765735e-01
4.67951894e-01 2.19328552e-01 3.25703770e-01 7.97930300e-01
-1.38172626e-01 4.44402173e-03 -8.65143716e-01 1.15240395e+00
1.77577987e-01 -1.35931087e+00 3.76023799e-01 -1.65873468e-02
9.62465525e-01 -1.27989188e-01 7.88396150e-02 3.96646470e-01
-1.68699369e-01 -9.65617180e-01 2.98251152e-01 6.20309293e-01
4.92923051e-01 -7.81227708e-01 1.17891753e+00 3.81129563e-01
-9.95242298e-01 -1.10553473e-01 -4.75241631e-01 -2.13876180e-02
-2.34927446e-01 5.76709688e-01 -7.47082174e-01 4.70585167e-01
5.60796499e-01 5.27997255e-01 -8.78088474e-01 1.24086201e+00
-1.26642287e-01 3.06043625e-01 -4.27492023e-01 -6.53496161e-02
5.19154191e-01 -3.97048384e-01 2.06609204e-01 8.29664290e-01
4.55417782e-01 -1.71770915e-01 -1.07857950e-01 9.76997793e-01
-4.32359844e-01 2.20195219e-01 -5.62343836e-01 -2.25288838e-01
2.05093473e-01 1.40002394e+00 -5.91865420e-01 -6.84629440e-01
-1.18668467e-01 1.07664919e+00 1.33765206e-01 3.04625094e-01
-5.72685838e-01 -8.81668389e-01 4.40027654e-01 3.09096626e-03
1.47954106e-01 -2.37996712e-01 -3.01657528e-01 -1.00343442e+00
-2.79750898e-02 -7.98042476e-01 4.16648090e-02 -6.43525183e-01
-8.82640064e-01 9.69324768e-01 -1.45960793e-01 -1.24569643e+00
4.09616269e-02 -8.08147728e-01 -7.46523559e-01 1.12002683e+00
-1.30956721e+00 -8.99116993e-01 -9.35302734e-01 5.62854052e-01
3.32697421e-01 -4.41594094e-01 6.00666642e-01 5.48369467e-01
-8.15345883e-01 6.36461854e-01 7.07613258e-03 6.91745579e-01
5.34027696e-01 -9.18530226e-01 2.88638771e-01 8.61656308e-01
3.65374386e-02 4.93987978e-01 3.33897769e-01 -4.80205327e-01
-8.04602981e-01 -1.12455976e+00 5.37824690e-01 3.61724406e-01
4.25057672e-02 -2.91371316e-01 -1.18579793e+00 6.67523384e-01
3.36372674e-01 2.06604049e-01 6.01382136e-01 -5.57289183e-01
-1.91359207e-01 -4.94235367e-01 -1.37417459e+00 5.98973930e-01
4.11950976e-01 -1.96169451e-01 -5.21232486e-01 8.47682764e-04
5.36944330e-01 -2.20431477e-01 -7.41532445e-01 5.94108820e-01
4.82969791e-01 -7.96817362e-01 5.58799803e-01 -4.88609105e-01
3.88169616e-01 -6.35778964e-01 -6.57821074e-02 -1.41018915e+00
-2.31788471e-01 -1.11828044e-01 3.40164810e-01 1.15541446e+00
4.31431353e-01 -9.63744879e-01 1.08784974e+00 3.83423269e-01
-7.36156255e-02 -8.27120662e-01 -5.54621398e-01 -6.98405445e-01
1.61657110e-02 7.67826661e-02 9.26332116e-01 1.10024714e+00
-4.10492957e-01 8.62460285e-02 -1.93027198e-01 7.64590651e-02
3.95523369e-01 -1.88979968e-01 8.78449261e-01 -1.18877161e+00
1.89972837e-02 -3.30381572e-01 -7.44079649e-01 -9.59987998e-01
-1.49775162e-01 -9.47967768e-01 -2.22461492e-01 -1.46774435e+00
-6.15555383e-02 -7.97142386e-01 -3.08607161e-01 7.11760342e-01
-1.28788754e-01 6.86552942e-01 3.96197259e-01 1.70192823e-01
2.80862778e-01 7.39339888e-01 1.59789729e+00 -2.08728969e-01
-3.05395603e-01 -2.87713092e-02 -3.99127990e-01 7.04620063e-01
1.24391747e+00 -3.87111723e-01 -8.30597132e-02 -4.20383185e-01
-5.09077013e-02 -3.71104032e-01 1.19567692e-01 -1.54198968e+00
1.54126436e-01 4.63979058e-02 7.55062699e-01 -6.10496402e-01
3.25732619e-01 -9.66775060e-01 1.95171595e-01 7.03247726e-01
-3.35969627e-01 9.72223803e-02 3.17610741e-01 -4.96989787e-02
-4.86414135e-01 -5.69930136e-01 1.08009017e+00 -1.29337341e-01
-5.92749119e-01 4.59938139e-01 -2.01409891e-01 -4.06935006e-01
1.12580264e+00 -2.11479947e-01 -2.92480260e-01 -1.11567065e-01
-7.21876621e-01 -6.39287233e-02 2.73293912e-01 4.43119377e-01
7.60562778e-01 -1.81775069e+00 -5.37467420e-01 7.66918361e-01
-4.24958095e-02 2.86470771e-01 1.41410023e-01 8.00728321e-01
-8.70891511e-01 9.96697345e-04 -7.60900497e-01 -4.16879624e-01
-1.14195907e+00 5.59917092e-01 5.64724803e-01 -1.36322230e-01
-3.04233253e-01 7.16505110e-01 1.31244823e-01 -6.56270564e-01
2.00185683e-02 -6.46420121e-01 -4.16840047e-01 -1.36278063e-01
6.30713165e-01 2.81993598e-01 -5.26208505e-02 -6.02189958e-01
-7.18949735e-03 6.29408479e-01 -2.77458936e-01 2.16608564e-03
1.27708602e+00 5.06833553e-01 -2.91815311e-01 1.92596048e-01
1.65459156e+00 -2.02691182e-01 -9.82660234e-01 -3.62790108e-01
-2.58600116e-01 -4.60240960e-01 -3.84666733e-02 -5.74272871e-01
-1.57013428e+00 1.09468770e+00 7.29063272e-01 1.97588682e-01
1.13418472e+00 -4.96847808e-01 7.83273220e-01 4.81874675e-01
-1.49903238e-01 -8.80396783e-01 1.60743073e-01 4.01270419e-01
1.01585817e+00 -1.21375191e+00 -2.57767707e-01 -1.72232166e-01
-4.24305528e-01 1.38269460e+00 1.05394578e+00 -5.50395608e-01
5.77856004e-01 1.71782106e-01 4.04423326e-02 1.04887202e-01
-1.29464984e-01 -1.84901729e-01 2.52782911e-01 2.90921062e-01
1.78104147e-01 -2.16621712e-01 -2.48524576e-01 3.98494959e-01
-1.96382612e-01 3.29278298e-02 5.16919315e-01 9.06677425e-01
-5.08529484e-01 -1.29869401e+00 -6.55737996e-01 7.22885072e-01
-2.68631995e-01 -1.13700554e-02 -3.47734332e-01 8.53408992e-01
5.41712046e-01 4.15006846e-01 3.24829876e-01 -7.16452181e-01
2.89175451e-01 1.70205638e-01 3.18986446e-01 -4.75071222e-01
-5.96824288e-01 -2.69294888e-01 -3.82252157e-01 -1.86333597e-01
-5.07198989e-01 -3.85550082e-01 -1.42215228e+00 -1.47036344e-01
-3.52088451e-01 3.43886048e-01 7.74337590e-01 8.49553108e-01
2.33747318e-01 7.54345357e-01 7.53891051e-01 -8.50252569e-01
-2.47361377e-01 -1.27192545e+00 -4.53528762e-01 4.11490113e-01
4.29090858e-01 -5.23696244e-01 -3.06434274e-01 -1.19837159e-02] | [9.318387031555176, 2.184021234512329] |
a9e499fa-c9ab-4b9a-8cde-f99848e296fe | point-cloud-registration-driven-robust | 2209.06395 | null | https://arxiv.org/abs/2209.06395v2 | https://arxiv.org/pdf/2209.06395v2.pdf | Point Cloud Registration-Driven Robust Feature Matching for 3D Siamese Object Tracking | Learning robust feature matching between the template and search area is crucial for 3D Siamese tracking. The core of Siamese feature matching is how to assign high feature similarity on the corresponding points between the template and search area for precise object localization. In this paper, we propose a novel point cloud registration-driven Siamese tracking framework, with the intuition that spatially aligned corresponding points (via 3D registration) tend to achieve consistent feature representations. Specifically, our method consists of two modules, including a tracking-specific nonlocal registration module and a registration-aided Sinkhorn template-feature aggregation module. The registration module targets at the precise spatial alignment between the template and search area. The tracking-specific spatial distance constraint is proposed to refine the cross-attention weights in the nonlocal module for discriminative feature learning. Then, we use the weighted SVD to compute the rigid transformation between the template and search area, and align them to achieve the desired spatially aligned corresponding points. For the feature aggregation model, we formulate the feature matching between the transformed template and search area as an optimal transport problem and utilize the Sinkhorn optimization to search for the outlier-robust matching solution. Also, a registration-aided spatial distance map is built to improve the matching robustness in indistinguishable regions (e.g., smooth surface). Finally, guided by the obtained feature matching map, we aggregate the target information from the template into the search area to construct the target-specific feature, which is then fed into a CenterPoint-like detection head for object localization. Extensive experiments on KITTI, NuScenes and Waymo datasets verify the effectiveness of our proposed method. | ['Jian Yang', 'Jin Xie', 'Guangyu Li', 'Le Hui', 'Kaihao Lan', 'Haobo Jiang'] | 2022-09-14 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [-2.50743181e-01 -6.00757539e-01 -2.12327346e-01 -2.48906657e-01
-1.10953176e+00 -4.81694013e-01 3.82371068e-01 -7.07554072e-03
-4.11483705e-01 6.68270588e-02 -2.89716776e-02 4.10851985e-01
-4.92524505e-01 -4.35086787e-01 -6.75875187e-01 -8.53549778e-01
1.13464125e-01 5.82050383e-01 5.42704403e-01 7.25183785e-02
4.66627359e-01 8.62604678e-01 -1.26159072e+00 -4.15679514e-01
9.28625762e-01 1.18451548e+00 2.92905062e-01 5.67351060e-04
3.69152762e-02 -2.21092761e-01 -3.38277519e-01 5.32472953e-02
6.50074899e-01 -1.47422045e-01 -2.86272705e-01 3.18666309e-01
8.36875379e-01 3.94398123e-02 -4.17872965e-01 1.34148645e+00
3.43093634e-01 5.22146225e-01 6.43000782e-01 -1.42161214e+00
-4.15547758e-01 -1.56868473e-01 -9.57764387e-01 2.55943120e-01
1.79300725e-01 4.40154016e-01 8.81807923e-01 -1.23740900e+00
6.94069266e-01 1.10371292e+00 6.47534907e-01 3.06839496e-01
-9.27790523e-01 -8.58956099e-01 2.79868603e-01 4.47877795e-02
-1.81009197e+00 -5.08365393e-01 1.03186738e+00 -4.20611382e-01
5.06995738e-01 2.21614704e-01 7.63716936e-01 5.15038848e-01
3.77933681e-01 4.99413848e-01 6.23715580e-01 5.01343329e-03
2.51188856e-02 -1.76691219e-01 9.32092965e-02 8.50935638e-01
3.88198674e-01 4.37344253e-01 -5.10801554e-01 -1.54884040e-01
8.48631620e-01 4.54566181e-01 -1.95513368e-01 -9.77852762e-01
-1.50368977e+00 7.12472677e-01 1.02633715e+00 3.79503220e-01
-4.88176256e-01 -1.84916109e-02 4.34976928e-02 -1.90002006e-02
2.13250175e-01 2.24037930e-01 -2.33848691e-01 2.65336275e-01
-1.07992268e+00 4.20619607e-01 2.22919315e-01 1.32099462e+00
9.72170651e-01 -1.11564830e-01 -4.09073651e-01 4.76858735e-01
8.72508705e-01 8.83027077e-01 3.78606200e-01 -8.80139172e-01
7.75057614e-01 9.57369626e-01 2.20768750e-01 -1.52349579e+00
-3.13272536e-01 -6.70661092e-01 -6.21727645e-01 1.75112486e-01
6.93487823e-01 2.47484341e-01 -8.35642040e-01 1.59425426e+00
9.87274826e-01 5.59324205e-01 -1.78153366e-01 1.48993659e+00
4.95626986e-01 4.13916856e-01 -1.48670614e-01 -4.90867086e-02
1.31263864e+00 -7.74242818e-01 -3.33257794e-01 -2.08266705e-01
6.73620224e-01 -8.33352804e-01 7.71036327e-01 -3.48806322e-01
-9.26040053e-01 -5.40201485e-01 -1.03153956e+00 3.80979781e-03
-1.49984032e-01 1.91437706e-01 2.32895911e-01 1.04847975e-01
-5.05610526e-01 4.89676535e-01 -8.67231309e-01 -4.17791128e-01
4.10015494e-01 4.74651873e-01 -3.40874881e-01 1.03823490e-01
-8.30813706e-01 7.41207361e-01 1.85743541e-01 3.17560017e-01
-5.78926086e-01 -9.18212712e-01 -1.04430139e+00 -1.56519637e-01
2.85655349e-01 -7.05588102e-01 6.92749023e-01 -4.99772638e-01
-1.19790637e+00 8.95327210e-01 -3.63708198e-01 -1.03684422e-02
4.02500600e-01 5.38516864e-02 -3.24600995e-01 -1.52594090e-01
5.41758955e-01 4.80314761e-01 1.01429701e+00 -1.11633158e+00
-8.76114845e-01 -8.39575469e-01 -5.45357406e-01 5.23267508e-01
1.06372483e-01 -2.47203801e-02 -9.30392087e-01 -6.99686050e-01
8.73491347e-01 -1.01549852e+00 -1.49798959e-01 3.64483088e-01
-2.91116536e-01 -4.55562592e-01 9.20884132e-01 -4.56272870e-01
1.01864648e+00 -2.41023707e+00 3.39677423e-01 8.19764555e-01
2.69175738e-01 -1.90276816e-01 -2.64971793e-01 -2.14207515e-01
6.90391585e-02 -4.76704240e-01 -1.32522926e-01 -2.22108111e-01
5.12183364e-03 -1.53584465e-01 6.45812973e-02 1.09858179e+00
1.67618811e-01 1.06271303e+00 -9.25198078e-01 -7.07638085e-01
2.38043949e-01 3.35211754e-01 -4.53344643e-01 1.30757555e-01
9.96687189e-02 5.08228064e-01 -1.08612251e+00 9.31243598e-01
8.78818154e-01 6.50331303e-02 -6.23724937e-01 -6.96002662e-01
-4.17320192e-01 -1.12900279e-01 -1.58121109e+00 2.08783197e+00
8.03107917e-02 1.62051857e-01 1.78745240e-01 -5.89396060e-01
1.16915882e+00 -2.28171185e-01 8.83171558e-01 -6.90426052e-01
3.24618757e-01 3.76067221e-01 3.71123143e-02 -1.40106440e-01
4.49263096e-01 1.41830027e-01 -1.63098395e-01 2.01309025e-01
-7.55056366e-02 -1.67111769e-01 -2.07778543e-01 -1.02547176e-01
6.18253529e-01 2.36548588e-01 1.18474409e-01 -3.40788066e-01
8.71486485e-01 2.65203357e-01 8.37180853e-01 3.02774876e-01
-5.97075462e-01 5.63692451e-01 -3.18048567e-01 -3.92793745e-01
-8.19618762e-01 -1.19616640e+00 -2.11412773e-01 5.06873846e-01
9.93809938e-01 -2.85399035e-02 -5.76036632e-01 -9.16541934e-01
3.39276105e-01 1.25687435e-01 -5.00132263e-01 -4.24199849e-01
-7.66811192e-01 -4.32987213e-02 6.13051876e-02 4.76714253e-01
5.36738276e-01 -7.79539466e-01 -4.79229093e-01 5.45766652e-02
8.40888694e-02 -7.04343736e-01 -1.42052162e+00 -8.78137872e-02
-9.18429494e-01 -1.04873514e+00 -7.96932697e-01 -8.89610052e-01
1.03174257e+00 5.32835782e-01 3.78313601e-01 2.19721928e-01
-3.62124443e-02 2.52280802e-01 -2.71805073e-03 4.80503030e-02
2.60671556e-01 -5.82510866e-02 4.25311327e-01 2.98492193e-01
5.36277235e-01 -1.67027801e-01 -6.99353218e-01 8.31623852e-01
-4.11788613e-01 -4.03188795e-01 6.76602483e-01 6.11021161e-01
1.20932949e+00 -1.12302393e-01 -1.04686739e-02 1.83208659e-02
1.90031081e-01 -3.68025243e-01 -9.87065434e-01 1.35741040e-01
-3.54408741e-01 1.35652632e-01 1.51610896e-01 -7.25276887e-01
-5.11437893e-01 5.55336058e-01 2.61483520e-01 -1.00992918e+00
6.20392747e-02 2.96099722e-01 -5.35321772e-01 -6.47665739e-01
4.59740967e-01 3.09476227e-01 9.92937982e-02 -3.30707073e-01
2.16828555e-01 4.09910500e-01 6.03199899e-01 -5.44124305e-01
1.56860042e+00 5.01986802e-01 8.73569623e-02 -3.90493512e-01
-7.33317256e-01 -8.16704392e-01 -9.40267980e-01 -2.87297755e-01
8.72775078e-01 -6.93889856e-01 -7.34703481e-01 2.56238699e-01
-9.39878523e-01 1.71418399e-01 -5.11586428e-01 8.48203778e-01
-5.98614275e-01 3.34907293e-01 -9.19705853e-02 -4.36288744e-01
-3.42448771e-01 -1.20429444e+00 1.46661210e+00 6.86004639e-01
-9.75166261e-02 -7.89427757e-01 2.55960166e-01 1.50807083e-01
2.11383477e-01 1.30099282e-01 3.17346036e-01 -6.20172501e-01
-9.79226172e-01 -5.63047945e-01 -4.43341851e-01 -3.56281906e-01
1.54172048e-01 -2.30099529e-01 -5.77235997e-01 -5.47572911e-01
1.71569124e-01 2.38800958e-01 5.39671004e-01 5.22761941e-01
6.86295450e-01 3.62902172e-02 -6.78035557e-01 1.01715446e+00
1.22756016e+00 1.95729107e-01 2.18577951e-01 4.93816048e-01
7.42149174e-01 2.86084801e-01 1.29294705e+00 1.64519146e-01
3.77083361e-01 9.85150099e-01 4.19858903e-01 8.84080902e-02
-1.93421945e-01 -4.83150959e-01 2.70420194e-01 9.35759306e-01
1.30914256e-01 3.94777626e-01 -6.99564874e-01 5.59245408e-01
-1.93896508e+00 -7.91187763e-01 -5.91053143e-02 2.56785440e+00
4.29198742e-01 -4.26528323e-03 6.41935840e-02 -3.53808790e-01
8.36493015e-01 7.16163367e-02 -8.48436892e-01 5.03797233e-01
1.00393012e-01 -1.94086358e-01 5.13882160e-01 5.57734907e-01
-1.28640223e+00 9.71730351e-01 4.43978024e+00 1.04734325e+00
-1.07253003e+00 5.51394187e-02 -1.85838014e-01 -7.88136646e-02
-2.97658574e-02 5.26099056e-02 -1.21693170e+00 6.01342499e-01
2.73557633e-01 -3.64851654e-01 2.67734945e-01 8.49202812e-01
3.25265765e-01 1.72507271e-01 -1.03566921e+00 1.16900468e+00
1.42243505e-01 -1.15567374e+00 -2.66579743e-02 1.73394188e-01
5.01180828e-01 1.96934957e-02 1.64714694e-01 1.48269832e-01
-1.98828861e-01 -5.65561056e-01 8.52753997e-01 8.35122705e-01
5.25692165e-01 -6.72426581e-01 4.43061709e-01 2.35281572e-01
-1.91410923e+00 1.97455660e-01 -5.53123295e-01 5.55140138e-01
6.04320969e-03 1.76004380e-01 -2.28931740e-01 7.06054330e-01
7.75068760e-01 9.77262855e-01 -5.65335274e-01 1.59110391e+00
8.07927698e-02 -2.41502240e-01 -6.52663469e-01 6.46977499e-02
2.31816113e-01 -5.41268051e-01 1.01513243e+00 6.75269246e-01
6.11329496e-01 -1.05789997e-01 6.00540936e-01 1.13305998e+00
2.07838431e-01 2.45271817e-01 -4.67006832e-01 4.56321359e-01
7.38751471e-01 1.34927404e+00 -6.19323611e-01 -3.63240391e-02
-2.54174143e-01 8.82727444e-01 3.51778090e-01 4.04994845e-01
-9.94912922e-01 -5.27806103e-01 8.43680918e-01 1.46023795e-01
3.08130443e-01 -3.43981147e-01 -2.34105587e-01 -1.20888150e+00
2.74107844e-01 -5.09551644e-01 3.86235714e-01 -5.64403892e-01
-1.30453706e+00 3.00702214e-01 -1.09837346e-01 -1.91671276e+00
1.71325281e-01 -3.07231903e-01 -9.41334903e-01 1.10775030e+00
-1.42635524e+00 -1.36195135e+00 -5.05603969e-01 1.02327144e+00
3.58656019e-01 -8.73540938e-02 2.42078811e-01 4.79189932e-01
-5.93155503e-01 7.61402249e-01 -6.27326742e-02 1.82191476e-01
7.09819376e-01 -8.02394629e-01 5.48118763e-02 8.69439840e-01
1.08349644e-01 7.91424513e-01 3.10214877e-01 -1.02552986e+00
-1.67603135e+00 -1.36737502e+00 7.79417396e-01 -5.40910780e-01
7.41059303e-01 -1.72252923e-01 -1.03641927e+00 6.03431404e-01
-5.68058252e-01 3.78423244e-01 -9.37453844e-03 -2.64694065e-01
-2.14401320e-01 -2.83819467e-01 -1.22481239e+00 6.28386140e-01
1.20754230e+00 -3.33145082e-01 -7.61189818e-01 3.01515698e-01
5.98535240e-01 -8.39515865e-01 -9.49488938e-01 4.37455207e-01
4.46758002e-01 -1.65209621e-01 1.14133549e+00 -3.83529782e-01
-3.71325910e-01 -1.05391264e+00 -9.07024592e-02 -1.20634902e+00
-5.58487654e-01 -5.18254101e-01 -5.51006980e-02 1.28896344e+00
1.35710046e-01 -5.08184671e-01 1.13068795e+00 5.31023860e-01
-3.09734195e-01 -5.70577443e-01 -1.45537162e+00 -1.04941249e+00
-5.66414706e-02 -2.81754851e-01 6.65380895e-01 8.21561933e-01
-3.31616402e-01 4.40096520e-02 4.88307439e-02 6.16942704e-01
1.14829814e+00 5.26618481e-01 9.05644476e-01 -1.34656894e+00
3.27735007e-01 -5.21060586e-01 -6.82573855e-01 -1.34596419e+00
2.83479989e-01 -1.26810241e+00 2.10192278e-01 -1.11881602e+00
1.35086432e-01 -8.12615573e-01 -3.25899690e-01 3.02489161e-01
-1.99185982e-01 1.37904987e-01 2.32835293e-01 6.20420039e-01
-7.41979003e-01 7.92925417e-01 1.46107662e+00 -2.87580401e-01
-4.57521558e-01 1.79148436e-01 -3.78351480e-01 5.22983372e-01
4.24353242e-01 -6.55221105e-01 1.48301393e-01 -2.67718941e-01
-3.25214267e-01 -2.23791435e-01 5.89198768e-01 -1.12346578e+00
8.54719698e-01 -3.19734573e-01 4.84255075e-01 -8.75214338e-01
4.29900765e-01 -1.41830611e+00 1.76555261e-01 4.37083870e-01
-7.85315409e-02 -9.55332909e-03 -1.03373371e-01 6.09037936e-01
-5.58698811e-02 -1.12315841e-01 9.37173426e-01 3.75943005e-01
-7.32448399e-01 1.01024652e+00 3.54691148e-01 1.02386802e-01
1.44200456e+00 -6.31195188e-01 1.28345285e-02 1.09623358e-01
-5.91758251e-01 7.57936239e-01 7.46784031e-01 5.26620686e-01
9.22154367e-01 -1.85457516e+00 -5.79407036e-01 5.98820150e-01
3.40938032e-01 1.36687934e-01 1.41693458e-01 1.38394618e+00
-1.11138098e-01 2.83998817e-01 -5.65213263e-02 -1.17186522e+00
-1.18573594e+00 6.12677872e-01 7.04714358e-01 5.28283045e-02
-6.43124878e-01 6.59016907e-01 7.89099038e-02 -5.37786841e-01
3.33042920e-01 -4.90215302e-01 1.55150546e-02 -1.26256451e-01
3.78850251e-01 2.84687907e-01 -1.70564100e-01 -1.24291861e+00
-7.33042717e-01 1.53728104e+00 9.20183957e-02 2.42107034e-01
9.64340448e-01 -2.57078439e-01 1.28891855e-01 3.88438217e-02
1.31316245e+00 1.29890680e-01 -1.48451412e+00 -5.61756313e-01
2.22301304e-01 -8.25967431e-01 1.21489856e-02 -2.02004015e-01
-1.44610584e+00 6.00056350e-01 9.03007746e-01 -3.63043338e-01
6.35993838e-01 2.83808887e-01 9.18205917e-01 9.90344360e-02
3.74072194e-01 -7.96532512e-01 -1.40559241e-01 5.63051343e-01
8.80405307e-01 -1.22992218e+00 2.03386359e-02 -3.69163543e-01
-3.96993905e-01 8.61647189e-01 8.27789605e-01 -4.74860996e-01
6.35291934e-01 1.11702289e-02 -3.46261961e-03 -5.66742420e-01
-4.48747352e-02 -3.69073361e-01 9.04094934e-01 6.93350017e-01
-1.54247880e-01 -2.57188708e-01 -5.85446469e-02 5.02155483e-01
-5.03036827e-02 -2.35564634e-01 -4.07208562e-01 7.07115173e-01
-6.03158176e-01 -7.08297551e-01 -7.39438713e-01 3.35648596e-01
1.73712760e-01 2.56563991e-01 -2.02358633e-01 8.49254489e-01
1.54055938e-01 6.26053393e-01 2.64899582e-01 -4.82163996e-01
7.69378245e-01 -1.66980311e-01 3.55858058e-01 -4.75429088e-01
-6.60490036e-01 3.71701837e-01 -5.41393459e-01 -9.96766448e-01
-2.88230836e-01 -9.39491332e-01 -1.64790046e+00 -9.36086699e-02
-6.58714235e-01 3.42085302e-01 5.17482162e-01 9.02647793e-01
4.40927565e-01 5.33143841e-02 7.40558088e-01 -1.25714183e+00
-5.83853424e-01 -5.24051428e-01 -6.51292384e-01 6.06727839e-01
3.15563917e-01 -9.76080179e-01 -4.94996697e-01 -3.56391668e-01] | [6.649143218994141, -2.393994092941284] |
2ace200a-6950-44e3-96eb-3279850487a5 | spatial-context-aware-deep-neural-network-for | 2111.12296 | null | https://arxiv.org/abs/2111.12296v2 | https://arxiv.org/pdf/2111.12296v2.pdf | Spatial-context-aware deep neural network for multi-class image classification | Multi-label image classification is a fundamental but challenging task in computer vision. Over the past few decades, solutions exploring relationships between semantic labels have made great progress. However, the underlying spatial-contextual information of labels is under-exploited. To tackle this problem, a spatial-context-aware deep neural network is proposed to predict labels taking into account both semantic and spatial information. This proposed framework is evaluated on Microsoft COCO and PASCAL VOC, two widely used benchmark datasets for image multi-labelling. The results show that the proposed approach is superior to the state-of-the-art solutions on dealing with the multi-label image classification problem. | ['Jiang Liu', 'Yitian Zhao', 'Jianfeng Ren', 'Qian Zhang', 'Jialu Zhang'] | 2021-11-24 | null | null | null | null | ['multi-label-image-classification'] | ['computer-vision'] | [ 4.51647282e-01 -3.80788028e-01 -2.42415965e-01 -7.45253623e-01
-7.55008042e-01 -3.64149898e-01 5.80910921e-01 4.44347948e-01
-7.50358462e-01 6.54023051e-01 -2.98868358e-01 -4.90040258e-02
-2.50114679e-01 -4.51030850e-01 -4.34667706e-01 -7.12024570e-01
3.91673476e-01 2.33961239e-01 3.54027450e-01 1.24938995e-01
5.19675016e-01 4.88834888e-01 -1.96947682e+00 3.17758352e-01
5.84909141e-01 1.21887898e+00 3.39557886e-01 1.74666360e-01
-2.63291836e-01 8.35210204e-01 -2.54328370e-01 -8.14155638e-02
-1.02403499e-01 -1.62411049e-01 -1.13271570e+00 2.41652489e-01
6.93217099e-01 2.44826719e-01 2.67730772e-01 1.26645684e+00
3.44582379e-01 1.96036652e-01 6.70749784e-01 -1.07060301e+00
-6.56608820e-01 -2.89517790e-02 -8.05014133e-01 2.33169645e-01
-2.89634615e-02 -4.34926867e-01 1.01016152e+00 -8.16884756e-01
4.10431653e-01 1.33850634e+00 6.59806669e-01 2.44435281e-01
-8.89959276e-01 -6.35421813e-01 3.03525627e-01 5.14614701e-01
-1.48318112e+00 4.51818369e-02 8.35409820e-01 -4.01569635e-01
6.38274133e-01 2.09722668e-01 1.20418906e-01 8.12280834e-01
2.57102668e-01 7.17887759e-01 1.61296356e+00 -6.67106986e-01
-4.52484079e-02 3.45474422e-01 1.87329441e-01 5.28507054e-01
-2.14869499e-01 -1.41529381e-01 -2.95364171e-01 2.22411472e-02
5.00998855e-01 2.45622754e-01 8.79574269e-02 -3.07728916e-01
-9.99052227e-01 8.45850646e-01 5.98722816e-01 7.03507483e-01
-2.28097945e-01 1.75176188e-01 6.47473872e-01 -1.94550067e-01
8.46439958e-01 2.46085659e-01 -5.09436429e-01 1.49900556e-01
-8.18567157e-01 2.23873913e-01 1.77957073e-01 7.22937524e-01
8.77808094e-01 -4.35957015e-01 -9.78919417e-02 1.22888160e+00
5.14533103e-01 1.20589152e-01 4.62791681e-01 -7.48251379e-01
3.47627789e-01 8.03868473e-01 1.28351664e-02 -1.20722270e+00
-8.56581092e-01 -5.46382189e-01 -8.19691837e-01 2.40414396e-01
3.78660649e-01 2.56760389e-01 -8.31385255e-01 1.47633946e+00
5.91584563e-01 3.26568365e-01 -2.01884449e-01 8.04919541e-01
9.57999945e-01 4.24713820e-01 6.77672625e-01 -3.79966535e-02
1.54212439e+00 -1.33979261e+00 -8.23607385e-01 -5.12738526e-01
8.58744860e-01 -8.53843987e-01 7.16687143e-01 1.30376607e-01
-7.44660199e-01 -9.60444629e-01 -8.89125705e-01 -1.52654471e-02
-7.65808880e-01 3.74959111e-01 3.75929058e-01 5.11057913e-01
-8.86566281e-01 3.76434207e-01 -2.25720882e-01 -5.11606276e-01
5.63838243e-01 1.96863532e-01 -6.22978330e-01 -3.51534694e-01
-1.19897521e+00 1.06337690e+00 8.72336090e-01 1.82606235e-01
-6.35771573e-01 -3.42951804e-01 -8.12733471e-01 -2.03644812e-01
5.16955197e-01 -2.09861577e-01 1.09399390e+00 -1.02250862e+00
-9.11124825e-01 1.25808907e+00 -7.72102028e-02 1.01065740e-01
3.54605675e-01 -1.72142424e-02 -5.29241741e-01 4.47186828e-02
4.62807745e-01 8.67349625e-01 3.87305975e-01 -1.52136874e+00
-1.06494391e+00 -6.93139076e-01 1.23025648e-01 4.64665383e-01
-3.33575040e-01 3.26543182e-01 -3.45510274e-01 -5.34892023e-01
1.22125022e-01 -1.20307231e+00 -1.98471904e-01 -2.11108685e-01
-5.00462174e-01 -5.95466197e-01 9.65022624e-01 -3.59677196e-01
8.40617180e-01 -2.06766820e+00 1.04942527e-02 -7.96869993e-02
-1.94188625e-01 3.76584142e-01 -1.54819146e-01 2.56563872e-01
-1.56152964e-01 5.84656373e-03 -9.29205343e-02 -5.68966687e-01
-2.20339745e-01 1.83001265e-01 2.13116869e-01 6.17187381e-01
-9.94517878e-02 7.59113014e-01 -8.52581918e-01 -8.17671120e-01
5.43862879e-01 5.88525593e-01 1.53993005e-02 4.65320237e-02
-2.79395282e-01 9.80726659e-01 -4.86803234e-01 7.12866068e-01
7.58453906e-01 -4.25231993e-01 1.89185306e-01 -2.83737332e-01
-1.19412668e-01 -3.82401645e-01 -1.14464343e+00 2.05778575e+00
-6.32358611e-01 4.35531855e-01 -2.72191048e-01 -1.19853318e+00
8.26551199e-01 3.15604061e-01 5.42784572e-01 -9.23151970e-01
3.07236552e-01 3.32614422e-01 -5.59176445e-01 -7.76052713e-01
4.70706463e-01 -2.43616387e-01 -2.40294710e-02 2.38783374e-01
1.14007138e-01 2.01051220e-01 7.20763206e-02 -2.55419075e-01
6.49205506e-01 3.74403596e-01 5.84476352e-01 -3.21280926e-01
8.68921697e-01 2.10736897e-02 5.51586688e-01 3.93358797e-01
-4.63004589e-01 5.05473733e-01 1.20972887e-01 -6.92819476e-01
-8.46568406e-01 -4.20845777e-01 -3.78371060e-01 1.47053719e+00
5.56688011e-01 -4.66997698e-02 -7.43371964e-01 -9.59449470e-01
-1.03274383e-01 5.49814165e-01 -9.05130565e-01 2.32880652e-01
-3.33685488e-01 -7.27048457e-01 5.21396577e-01 5.67773163e-01
8.91266584e-01 -1.09789932e+00 -5.86593747e-01 2.38381729e-01
-2.60449111e-01 -1.42667091e+00 -2.16864675e-01 2.15339631e-01
-5.81909776e-01 -1.16730380e+00 -8.55418682e-01 -1.15571153e+00
5.07174432e-01 5.63143969e-01 9.12025094e-01 1.20779879e-01
-4.69342738e-01 9.34114307e-02 -4.30942059e-01 -3.03556055e-01
-1.17390871e-01 1.18435092e-01 -4.21241611e-01 3.28945100e-01
5.56078613e-01 -2.93261200e-01 -5.99591315e-01 7.12681770e-01
-9.63478446e-01 1.23034336e-01 5.24349809e-01 7.51690626e-01
8.38786960e-01 3.53492469e-01 7.65485585e-01 -1.10647678e+00
2.69677371e-01 -5.40678978e-01 -5.36251664e-01 5.55002153e-01
-5.03380418e-01 -1.26036093e-01 3.35175931e-01 -2.64540493e-01
-1.21900344e+00 4.18733865e-01 -1.84157252e-01 -8.44024196e-02
-8.76178265e-01 3.71681094e-01 -2.32734188e-01 -3.17210138e-01
3.15085799e-01 8.95540416e-02 -3.58593136e-01 -6.10746861e-01
4.64330494e-01 9.48453486e-01 3.21036547e-01 -4.30100769e-01
-7.89884329e-02 6.56125069e-01 3.32967460e-01 -5.35417020e-01
-1.53690529e+00 -1.05486107e+00 -1.01298892e+00 -4.44759369e-01
1.29473913e+00 -8.52877080e-01 -5.76342285e-01 6.97009444e-01
-1.34214294e+00 1.19208649e-01 2.46795371e-01 2.06805050e-01
-6.07336700e-01 3.03134978e-01 -3.60597849e-01 -5.59693396e-01
-4.32613194e-02 -1.58369935e+00 1.36951745e+00 3.63917261e-01
8.62431377e-02 -1.32432389e+00 1.74973518e-01 7.84170330e-01
3.77980292e-01 4.05500412e-01 8.89281690e-01 -7.63873219e-01
-2.81988949e-01 -2.52064675e-01 -8.80299866e-01 2.43029997e-01
9.31502357e-02 -5.62151372e-01 -1.26470065e+00 -4.84849960e-02
3.09755299e-02 -5.52607417e-01 7.48139441e-01 2.61425465e-01
1.40935731e+00 1.98355630e-01 -6.33470356e-01 2.90562302e-01
1.73318565e+00 2.09068328e-01 2.03316376e-01 6.40553892e-01
1.04812527e+00 9.33553934e-01 9.03557301e-01 2.75446415e-01
7.65488565e-01 1.01483190e+00 7.95046926e-01 -1.70006171e-01
-2.38361731e-02 2.29397565e-01 -6.27807498e-01 3.99097770e-01
1.01179697e-01 -1.45105630e-01 -1.12304962e+00 4.96434301e-01
-2.18617821e+00 -5.83420753e-01 -3.44127357e-01 1.50476789e+00
5.38673699e-01 -6.78389668e-02 -1.99900478e-01 1.68810502e-01
9.64077473e-01 3.26779693e-01 -4.46592540e-01 -3.19052219e-01
-4.49510552e-02 -1.96741685e-01 6.93867743e-01 1.29468799e-01
-1.90234566e+00 9.20412719e-01 5.62875175e+00 1.12175524e+00
-1.05207455e+00 6.93772316e-01 6.96502686e-01 4.16759998e-01
1.35168225e-01 -3.08219403e-01 -8.73542249e-01 3.70215446e-01
7.77541161e-01 5.66045463e-01 -5.66490293e-02 8.76004517e-01
-2.01310396e-01 -3.84010464e-01 -8.36632729e-01 9.80486095e-01
3.85849088e-01 -1.03714919e+00 -2.76527643e-01 8.29916373e-02
9.53644395e-01 -8.96765888e-02 1.77219853e-01 9.41870064e-02
1.77922487e-01 -1.10391355e+00 7.15894639e-01 4.26141739e-01
7.58425057e-01 -8.36741805e-01 1.08692956e+00 5.05902290e-01
-1.42070115e+00 -3.08861971e-01 -2.64878601e-01 2.10976228e-02
2.45922208e-01 6.88925445e-01 -3.56417924e-01 7.06506312e-01
7.63293743e-01 9.33207214e-01 -8.77964854e-01 9.92276251e-01
-2.70130396e-01 1.90462813e-01 1.21261016e-01 2.18508631e-01
7.65957594e-01 8.95762071e-03 -1.79015189e-01 1.13611221e+00
1.81231439e-01 -1.15954861e-01 3.81662786e-01 4.58797961e-01
4.81663682e-02 4.65867609e-01 -6.92013741e-01 4.24446583e-01
1.60853326e-01 1.49169064e+00 -1.13749218e+00 -1.70366108e-01
-6.78860843e-01 8.79781425e-01 6.64475560e-01 1.65981099e-01
-9.18017507e-01 -2.47964058e-02 2.24753693e-01 -3.64118755e-01
4.59001698e-02 -6.64048642e-02 -4.53342497e-01 -5.37441134e-01
-2.13964283e-01 -5.26172221e-01 4.90732193e-01 -7.54301429e-01
-1.09587836e+00 5.66453755e-01 2.59025786e-02 -1.06978858e+00
1.73092350e-01 -6.39052749e-01 -1.76281929e-01 9.10788953e-01
-1.81820309e+00 -1.71523142e+00 -4.44469184e-01 3.70151520e-01
8.32480907e-01 -9.52998251e-02 1.03313541e+00 5.90235531e-01
-5.01175761e-01 2.86112577e-01 3.10155362e-01 -1.57089099e-01
7.24135160e-01 -1.09151649e+00 -9.82562662e-04 3.92757505e-01
2.86455214e-01 1.31127387e-01 3.60755503e-01 -3.96924436e-01
-6.42243624e-01 -1.36725533e+00 1.01033413e+00 -3.06297034e-01
4.21105981e-01 4.57364060e-02 -8.41229081e-01 4.89577085e-01
4.42756474e-01 4.33836788e-01 8.75916123e-01 1.03690475e-01
-4.97105002e-01 1.98717415e-02 -1.20965374e+00 1.92522332e-01
8.09183657e-01 -5.85985601e-01 -2.67675489e-01 5.09743154e-01
5.01904011e-01 -3.32298845e-01 -8.80173981e-01 6.66558504e-01
3.70792836e-01 -9.40421760e-01 1.06611764e+00 -6.16995059e-02
3.28852326e-01 -2.24723175e-01 -2.76498497e-01 -1.46347320e+00
-2.72585243e-01 4.94654566e-01 5.66256881e-01 1.17632568e+00
1.31937861e-01 -3.38866591e-01 5.21990538e-01 1.89912170e-01
7.73028797e-03 -8.72929573e-01 -1.03869486e+00 -4.71639693e-01
6.59616515e-02 -3.20847839e-01 3.41967195e-01 1.28763378e+00
-3.47085506e-01 3.51286292e-01 -4.06069189e-01 4.99285161e-02
6.61375821e-01 -1.96061656e-02 2.17319712e-01 -1.50007439e+00
3.27510744e-01 -4.27054375e-01 -6.40732050e-01 -4.86893326e-01
6.54251993e-01 -8.53467882e-01 7.15563968e-02 -1.73058796e+00
3.92576694e-01 -8.09203744e-01 -6.80089414e-01 4.88947093e-01
-3.49539995e-01 7.95045137e-01 2.38650143e-01 2.41413757e-01
-1.19471753e+00 3.43013644e-01 1.31676221e+00 -3.16229433e-01
4.14084643e-01 -1.83172032e-01 -4.49269444e-01 8.23696434e-01
8.96967888e-01 -6.53384626e-01 -3.23924512e-01 -4.54948634e-01
3.38296443e-01 -4.45956215e-02 2.53406137e-01 -1.23547912e+00
3.61017495e-01 -2.36968443e-01 1.20629281e-01 -7.80791879e-01
3.82485926e-01 -1.14137614e+00 2.30212241e-01 9.09959674e-02
-6.41898274e-01 5.08110300e-02 2.99244165e-01 7.28537560e-01
-5.02644539e-01 -4.92764264e-01 9.57125127e-01 -2.66157955e-01
-1.16619277e+00 1.41461447e-01 -5.61640784e-02 -1.47710249e-01
1.52123094e+00 -6.82043582e-02 -2.47832552e-01 2.35497266e-01
-7.15217948e-01 1.59956142e-01 2.94294178e-01 6.02126062e-01
4.90265638e-01 -1.47291350e+00 -3.47422481e-01 -7.24497586e-02
6.32673144e-01 8.30478817e-02 7.39969790e-01 6.51423931e-01
-4.46440041e-01 6.46930933e-01 -3.08564961e-01 -8.07859123e-01
-1.41833055e+00 7.35435963e-01 3.78941059e-01 -4.79288608e-01
-3.47261339e-01 9.32901978e-01 2.55707860e-01 -7.29586840e-01
3.29672724e-01 1.37180060e-01 -9.22386527e-01 3.50844115e-01
4.93316591e-01 3.45375985e-01 1.59649789e-01 -1.30151618e+00
-4.06554192e-01 1.13599813e+00 -1.58369735e-01 2.01028705e-01
1.06554890e+00 -5.12178004e-01 -2.53616750e-01 7.32140303e-01
1.48598695e+00 -7.42254436e-01 -9.72720087e-01 -5.76965451e-01
4.14318085e-01 -4.49030370e-01 2.01208681e-01 -8.57664526e-01
-1.12456441e+00 1.11723375e+00 9.02566254e-01 1.72215089e-01
9.93711889e-01 -1.21350691e-01 5.72495997e-01 6.26987144e-02
5.49431503e-01 -1.28068006e+00 2.45600387e-01 4.80859518e-01
5.76532841e-01 -1.89974546e+00 -7.09033161e-02 -5.42147934e-01
-6.83059096e-01 9.66911197e-01 7.31524467e-01 1.66683823e-01
9.20178592e-01 -2.88308740e-01 3.55330199e-01 -3.36549431e-01
-2.16574535e-01 -3.30372721e-01 3.73491913e-01 2.32935280e-01
5.42633355e-01 7.76847973e-02 -3.75986785e-01 1.01034038e-01
6.16649032e-01 -4.50037234e-02 -1.64049286e-02 9.82951760e-01
-5.27820110e-01 -1.09151554e+00 -5.83081126e-01 1.77210093e-01
-8.22251856e-01 1.77193791e-01 8.15629065e-02 6.95874274e-01
6.24796510e-01 1.18805206e+00 -8.86454880e-02 -2.57566839e-01
1.83315188e-01 7.21408874e-02 3.19695324e-01 -5.73871493e-01
-6.13070190e-01 -1.35088190e-01 -4.34863754e-02 -5.27975202e-01
-1.18601453e+00 -3.90575707e-01 -1.05765676e+00 1.39530927e-01
-5.22599399e-01 -2.52466928e-02 1.16657221e+00 1.24202597e+00
1.31137326e-01 7.97227561e-01 5.09830594e-01 -1.08324754e+00
-1.37295529e-01 -1.15647542e+00 -7.56245971e-01 6.09802902e-01
4.88005467e-02 -1.10847127e+00 -8.73570517e-02 -1.20787472e-01] | [9.799433708190918, 3.9811649322509766] |
36f4a2ee-0af6-4e96-9653-9bdd7ed94917 | strategies-in-transfer-learning-for-low | 2306.12040 | null | https://arxiv.org/abs/2306.12040v1 | https://arxiv.org/pdf/2306.12040v1.pdf | Strategies in Transfer Learning for Low-Resource Speech Synthesis: Phone Mapping, Features Input, and Source Language Selection | We compare using a PHOIBLE-based phone mapping method and using phonological features input in transfer learning for TTS in low-resource languages. We use diverse source languages (English, Finnish, Hindi, Japanese, and Russian) and target languages (Bulgarian, Georgian, Kazakh, Swahili, Urdu, and Uzbek) to test the language-independence of the methods and enhance the findings' applicability. We use Character Error Rates from automatic speech recognition and predicted Mean Opinion Scores for evaluation. Results show that both phone mapping and features input improve the output quality and the latter performs better, but these effects also depend on the specific language combination. We also compare the recently-proposed Angular Similarity of Phone Frequencies (ASPF) with a family tree-based distance measure as a criterion to select source languages in transfer learning. ASPF proves effective if label-based phone input is used, while the language distance does not have expected effects. | ['Esther Klabbers', 'Jelske Dijkstra', 'Matt Coler', 'Phat Do'] | 2023-06-21 | null | null | null | null | ['transfer-learning', 'speech-synthesis'] | ['miscellaneous', 'speech'] | [ 7.60086924e-02 -2.87714392e-01 -1.82082936e-01 -3.16053659e-01
-1.12195683e+00 -8.15972269e-01 8.18760037e-01 -4.79607321e-02
-8.56651723e-01 1.13946772e+00 4.14449841e-01 -7.54497647e-01
-1.21766627e-01 -5.90377748e-01 -4.31869596e-01 -5.80080628e-01
2.39212587e-01 6.76847816e-01 4.36585486e-01 -3.03456813e-01
3.92146975e-01 5.74942172e-01 -1.57018602e+00 2.33218387e-01
1.31644762e+00 3.83882970e-01 6.45331025e-01 5.19572318e-01
-1.57138810e-01 3.88138682e-01 -8.56039047e-01 -4.04185951e-01
8.27157646e-02 -8.02630246e-01 -9.66120660e-01 -4.10768002e-01
3.30471963e-01 1.47048011e-01 2.10633054e-02 7.47996807e-01
7.94655979e-01 2.02861711e-01 1.07478893e+00 -6.78533614e-01
-7.34269500e-01 8.06535840e-01 1.73813850e-01 2.94599950e-01
4.24886823e-01 -9.44517106e-02 8.00151944e-01 -9.36280370e-01
6.38634264e-01 1.45602071e+00 5.95178246e-01 2.77281970e-01
-1.05861843e+00 -5.90585232e-01 -2.34292626e-01 2.76878238e-01
-1.27715981e+00 -6.02613091e-01 3.80047530e-01 -4.20154124e-01
1.14479804e+00 2.23178223e-01 4.03441370e-01 9.63956416e-01
2.61156391e-02 3.93057793e-01 1.66739082e+00 -1.04553103e+00
1.20729044e-01 7.22282350e-01 -1.92371696e-01 5.72313905e-01
5.57710603e-02 2.83792496e-01 -7.79142737e-01 1.38816670e-01
5.81523299e-01 -8.98215294e-01 -3.39640051e-01 2.12062851e-01
-1.37543142e+00 8.72063577e-01 -1.10689044e-01 9.40227509e-01
-6.60374165e-02 -4.32166040e-01 3.90526295e-01 7.94980347e-01
4.34102923e-01 3.55425477e-01 -6.68420434e-01 -5.43751359e-01
-7.77726412e-01 -1.11732207e-01 7.19013155e-01 5.61015606e-01
4.98589277e-01 4.02674049e-01 -1.76623002e-01 1.47604072e+00
2.20962152e-01 7.98602700e-01 7.73529828e-01 -6.43202245e-01
7.21705437e-01 -6.69056997e-02 -4.62734215e-02 -4.37531352e-01
-5.44561818e-02 -1.11120917e-01 -1.21564269e-01 2.39953518e-01
8.14144015e-01 -2.06597805e-01 -9.27969098e-01 1.81330812e+00
9.61873159e-02 -2.08071053e-01 2.97512978e-01 6.40721500e-01
5.99596918e-01 8.38851392e-01 7.72571042e-02 -5.76010227e-01
1.11477101e+00 -7.69991994e-01 -6.13030910e-01 1.71074152e-01
9.89209294e-01 -1.02854180e+00 1.60746086e+00 4.59512562e-01
-1.03437805e+00 -8.40133131e-01 -7.80621946e-01 2.45119005e-01
-6.08746946e-01 2.79669195e-01 5.56634879e-03 1.24635327e+00
-1.16576409e+00 6.31879747e-01 -5.12055814e-01 -4.61430252e-01
-3.71378332e-01 5.17192245e-01 -4.17369068e-01 1.47982314e-01
-1.38642156e+00 1.29550219e+00 6.79245710e-01 -3.24962974e-01
-2.37998694e-01 -3.36837023e-01 -6.39736652e-01 -9.43795890e-02
-3.69518340e-01 -1.04517400e-01 9.61300731e-01 -1.05218434e+00
-2.13438106e+00 7.93987095e-01 -1.39875163e-03 5.51197678e-02
4.28085715e-01 -6.59855977e-02 -6.35720968e-01 7.06684440e-02
9.94302519e-03 5.63399076e-01 3.93905401e-01 -8.81163478e-01
-7.44992316e-01 -2.58754998e-01 -4.46785688e-01 5.59054852e-01
-4.33178842e-01 4.43658769e-01 -7.74490163e-02 -6.73020124e-01
-2.47295082e-01 -9.11987066e-01 4.44174558e-01 -7.90398300e-01
1.34064853e-01 -3.79225314e-01 4.18564618e-01 -1.27136099e+00
1.16502714e+00 -2.07423544e+00 4.79956642e-02 4.66369659e-01
-7.71837115e-01 3.39287460e-01 -2.06211224e-01 5.15205562e-01
2.50763327e-01 1.60693571e-01 -1.20238654e-01 2.14831978e-01
7.24584982e-03 2.55832523e-01 8.40381905e-02 1.58445030e-01
1.15810700e-01 5.80384731e-01 -6.98694468e-01 -5.92659593e-01
4.22017932e-01 4.84039068e-01 -2.88972706e-01 -1.44793317e-01
3.04772943e-01 6.19314969e-01 8.97285342e-02 3.64755452e-01
1.87614605e-01 5.85619628e-01 1.77414611e-01 1.39491752e-01
-3.89050186e-01 1.00047886e+00 -9.64765787e-01 1.30118930e+00
-9.35980558e-01 7.03121006e-01 -2.70822108e-01 -8.59672368e-01
1.19151068e+00 5.07148325e-01 -2.55355209e-01 -9.37515259e-01
-1.12339163e-04 9.20407236e-01 7.46305168e-01 -4.07765716e-01
4.32760984e-01 -3.66766781e-01 1.48120716e-01 1.28941253e-01
2.11451799e-01 -2.26512104e-01 2.06933275e-01 -4.02161419e-01
6.20148957e-01 2.48208344e-01 2.77240992e-01 -6.63473248e-01
6.39393806e-01 -2.65738010e-01 3.71364743e-01 6.37101650e-01
-4.93290201e-02 5.13549626e-01 2.60693580e-01 3.03473055e-01
-7.85315394e-01 -1.38861799e+00 -4.27259564e-01 1.48307121e+00
-3.40929389e-01 -2.08047051e-02 -6.37123525e-01 -7.56839991e-01
-1.47683114e-01 1.19772601e+00 -9.61432308e-02 -8.37316737e-03
-8.22424293e-01 -3.74221951e-01 6.60645604e-01 4.37798768e-01
1.39761046e-01 -1.45250702e+00 -1.14878163e-01 1.96015581e-01
-3.43399316e-01 -8.63736451e-01 -3.94680172e-01 2.89607823e-01
-9.29809034e-01 -3.04079175e-01 -1.12046099e+00 -1.06624472e+00
1.65712953e-01 -3.10569137e-01 8.01968634e-01 -3.20057571e-01
4.47873503e-01 3.52459401e-01 -5.24269760e-01 -2.32443437e-01
-9.01795268e-01 3.34822834e-01 4.10634875e-01 -3.80839020e-01
2.54316598e-01 -4.25262302e-01 5.75939082e-02 3.50798309e-01
-4.08940583e-01 -4.22617137e-01 5.51523745e-01 7.25153387e-01
2.12306798e-01 -2.51842856e-01 9.37733889e-01 -7.90080249e-01
4.84411657e-01 -1.44920826e-01 -2.76873827e-01 4.68371987e-01
-6.45298600e-01 4.58222181e-02 5.36166131e-01 -5.51213205e-01
-1.20521069e+00 -3.15293103e-01 -5.20580888e-01 1.18869871e-01
-2.75423795e-01 6.54460549e-01 -3.39910388e-01 -7.56181553e-02
7.30306625e-01 2.93736279e-01 -1.78936675e-01 -5.32937050e-01
-4.81444709e-02 1.08816314e+00 3.40422004e-01 -5.46712995e-01
5.30586183e-01 -2.91238666e-01 -4.91806656e-01 -1.32312024e+00
-8.80407989e-02 -1.73827231e-01 -8.10467541e-01 -3.37282002e-01
7.04906464e-01 -7.49159276e-01 -3.08382481e-01 4.63952869e-01
-1.21870244e+00 -5.72156131e-01 -1.29153520e-01 1.30349278e+00
-5.98467171e-01 2.27210060e-01 -6.70161545e-01 -8.56650352e-01
-5.66945449e-02 -1.02348661e+00 7.18751669e-01 -9.47321393e-03
-4.19735491e-01 -1.24966264e+00 2.20628396e-01 3.11019063e-01
2.99154937e-01 -2.76683778e-01 1.37328911e+00 -9.79727864e-01
-9.33860317e-02 3.39084715e-01 1.07895054e-01 6.70248032e-01
4.65026379e-01 -3.86549048e-02 -8.68373990e-01 -1.98440075e-01
-3.03732991e-01 -2.66250938e-01 5.45846939e-01 4.18715328e-01
2.30429590e-01 -7.65311345e-02 1.44204572e-01 5.91142066e-02
1.21576917e+00 7.89075315e-01 6.41990721e-01 3.28412950e-01
5.13245523e-01 8.25056016e-01 6.16899669e-01 -1.14955693e-01
6.46239519e-02 8.60182643e-01 -6.61454260e-01 2.12867871e-01
-3.74763846e-01 -2.75975347e-01 1.09214377e+00 1.71290874e+00
-1.88971862e-01 -4.51823592e-01 -1.05690503e+00 6.70810163e-01
-1.19045627e+00 -7.06507146e-01 3.74669419e-03 2.68164968e+00
9.85434353e-01 2.02268049e-01 3.48228842e-01 4.29467291e-01
7.32077122e-01 -3.37530047e-01 1.94278076e-01 -8.68387282e-01
-3.02066654e-01 6.99559033e-01 3.51959974e-01 1.03629982e+00
-4.97471809e-01 1.19822752e+00 6.08782911e+00 1.14176238e+00
-1.30394816e+00 1.77876383e-01 3.34818095e-01 2.89700717e-01
-4.06299859e-01 -1.45237893e-01 -7.62506247e-01 5.24692297e-01
1.47442627e+00 8.48689079e-02 3.73522013e-01 2.51006544e-01
1.39649168e-01 -2.19951555e-01 -8.62441003e-01 7.15565622e-01
-1.17602888e-02 -8.36006582e-01 -4.80889678e-02 -4.73957732e-02
5.29132605e-01 1.51616529e-01 7.14364648e-02 3.87007445e-01
-6.64116303e-03 -7.89594591e-01 8.85207713e-01 2.57209301e-01
1.04759419e+00 -8.33314717e-01 7.12953568e-01 7.57961273e-02
-9.73552287e-01 3.37317795e-01 -2.67749429e-01 -6.63596988e-02
1.38796195e-02 9.15190503e-02 -1.14019179e+00 3.85279179e-01
5.14649868e-01 3.45729850e-02 -5.60285032e-01 7.10058689e-01
-2.99748749e-01 1.44424236e+00 -5.12392044e-01 -3.76606941e-01
1.60698622e-01 -3.55619252e-01 4.32255596e-01 1.47660792e+00
7.33697832e-01 -3.23964596e-01 -2.03258768e-01 3.50949109e-01
4.52856988e-01 8.46375167e-01 -7.58576751e-01 -1.60818681e-01
7.36832678e-01 6.27145231e-01 -1.00360954e+00 -1.75919205e-01
-4.95049059e-01 1.02660084e+00 1.57631442e-01 4.06760961e-01
-3.52065653e-01 -7.35352755e-01 3.77183527e-01 8.74982625e-02
3.04307848e-01 -2.20776990e-01 -3.81729513e-01 -7.03773081e-01
-9.00463611e-02 -9.80727136e-01 1.29872650e-01 -5.79948366e-01
-1.00050926e+00 6.40956879e-01 1.62576854e-01 -1.11026156e+00
-5.72364330e-01 -9.32602108e-01 -4.58099097e-01 1.39724565e+00
-1.12562096e+00 -7.77239621e-01 5.52831233e-01 3.96303535e-01
4.27169561e-01 -5.58248580e-01 8.61108482e-01 4.10146147e-01
-1.21809073e-01 9.17073727e-01 4.49862242e-01 1.68587998e-01
8.67707610e-01 -1.18454194e+00 3.50743771e-01 5.50442576e-01
5.05615234e-01 4.58963573e-01 4.73490238e-01 -6.78688705e-01
-5.09212315e-01 -6.06368244e-01 1.57310152e+00 -3.65616947e-01
4.20369714e-01 -3.15218508e-01 -7.42672026e-01 2.57321954e-01
3.53396475e-01 -6.12852931e-01 7.45695353e-01 3.87403637e-01
-1.81074038e-01 6.70409820e-04 -9.95923817e-01 5.87052822e-01
9.85800087e-01 -8.02480996e-01 -7.54361987e-01 1.67885318e-01
6.36349201e-01 4.13295254e-02 -9.10886765e-01 2.06449360e-01
6.74874723e-01 -9.81747210e-01 5.78834236e-01 -3.79572093e-01
-1.56395603e-02 -8.12829733e-02 -3.47844690e-01 -1.62317634e+00
-3.00304711e-01 -3.10280055e-01 8.59912395e-01 1.56674862e+00
1.00762570e+00 -9.48114693e-01 4.59407330e-01 -1.02051027e-01
-3.45236778e-01 -2.47980282e-01 -1.24049234e+00 -1.22830665e+00
4.02741641e-01 -2.65539497e-01 1.61690518e-01 9.04590487e-01
1.27647400e-01 6.51513398e-01 -1.30707353e-01 -1.20658219e-01
-1.55643672e-02 -3.38041425e-01 2.53996342e-01 -1.30388951e+00
-4.39244896e-01 -4.30625826e-01 -4.08609360e-01 -4.21704888e-01
3.58675182e-01 -1.07214546e+00 1.07940964e-01 -1.35592604e+00
-4.71292049e-01 -7.03045726e-01 -3.81282389e-01 3.47431540e-01
-1.69574067e-01 1.09887384e-01 5.17302692e-01 5.13826795e-02
1.94357768e-01 4.70544368e-01 9.47158277e-01 2.67471850e-01
-5.67559719e-01 2.94653475e-01 -9.96226072e-02 4.50861782e-01
9.76028860e-01 -5.29296458e-01 -4.44396704e-01 -3.93320292e-01
-1.62425771e-01 2.18996003e-01 -2.37091273e-01 -1.08717287e+00
-2.38872319e-01 -1.88708276e-01 2.11575732e-01 -1.35484278e-01
2.71459818e-01 -4.88516778e-01 -4.76340428e-02 6.71551347e-01
-3.88078570e-01 2.91515231e-01 3.53497267e-01 -8.83710980e-02
-2.60785431e-01 -5.55392623e-01 8.89064133e-01 1.22083969e-01
-5.59206843e-01 -2.86419839e-01 -7.42807269e-01 2.30950609e-01
4.87799138e-01 -6.11932814e-01 4.68569212e-02 -4.30583030e-01
-3.78039241e-01 -3.76131803e-01 2.28559881e-01 5.32207370e-01
2.16535866e-01 -1.40936518e+00 -9.96168077e-01 2.71852612e-01
2.49583181e-03 -9.58642364e-01 -1.64863288e-01 8.90079439e-01
-6.16811216e-01 4.87597853e-01 -5.08687019e-01 -4.12307590e-01
-1.38691998e+00 8.73027816e-02 2.68505770e-04 -9.11787972e-02
7.72695690e-02 9.55409884e-01 1.10212646e-01 -9.95299339e-01
2.59333640e-01 -1.68603137e-01 -2.71104991e-01 2.05811679e-01
-5.30963838e-02 5.61957121e-01 3.63566935e-01 -8.22111845e-01
-4.61184055e-01 4.69635963e-01 8.07630122e-02 -9.47884560e-01
6.44116282e-01 -1.70490086e-01 2.03413412e-01 1.07000995e+00
1.07923400e+00 6.59707189e-01 -3.64360869e-01 -1.43146306e-01
2.32683480e-01 -2.08991006e-01 -1.15302131e-01 -1.06798518e+00
-4.50659424e-01 9.99867916e-01 7.47224391e-01 -9.34933424e-02
9.14629340e-01 -6.15837872e-02 4.91005599e-01 2.24441320e-01
4.08425748e-01 -1.51240218e+00 -3.50962609e-01 8.17087710e-01
6.74607992e-01 -8.17890644e-01 -5.75185299e-01 -4.47632849e-01
-8.62902522e-01 9.39401627e-01 3.07918906e-01 3.49214911e-01
5.85225165e-01 1.19600348e-01 4.88077551e-01 6.74578071e-01
-4.80812192e-01 -4.75183487e-01 3.74981999e-01 9.19708788e-01
9.87309694e-01 1.34622857e-01 -9.13019240e-01 2.19174057e-01
-7.91788161e-01 -3.22224826e-01 1.94262072e-01 7.14348733e-01
-5.24757087e-01 -1.53021586e+00 -5.46867907e-01 4.87594813e-01
-4.13011521e-01 -4.09399182e-01 -4.44437593e-01 9.89782691e-01
2.55864859e-01 9.25740421e-01 1.37409866e-01 -6.72348738e-01
4.92871292e-02 5.69783330e-01 7.01626182e-01 -7.04595089e-01
-8.85016203e-01 1.81262642e-01 5.15245080e-01 2.00456515e-01
-4.54843491e-01 -8.61831963e-01 -1.08393979e+00 -1.79811623e-02
-5.83506525e-01 5.72425127e-01 5.75542092e-01 1.13975763e+00
-9.99210253e-02 1.41742295e-02 6.14909112e-01 -4.94522423e-01
-3.76269966e-01 -1.36439776e+00 -6.56205595e-01 -3.57887568e-03
-5.72543703e-02 -5.77541709e-01 -3.80077809e-01 -6.34709094e-03] | [14.297798156738281, 6.962036609649658] |
b5bcca5d-f565-436a-affd-c2f7a29c8f9c | sdcl-self-distillation-contrastive-learning | 2210.17168 | null | https://arxiv.org/abs/2210.17168v4 | https://arxiv.org/pdf/2210.17168v4.pdf | SDCL: Self-Distillation Contrastive Learning for Chinese Spell Checking | Due to the ambiguity of homophones, Chinese Spell Checking (CSC) has widespread applications. Existing systems typically utilize BERT for text encoding. However, CSC requires the model to account for both phonetic and graphemic information. To adapt BERT to the CSC task, we propose a token-level self-distillation contrastive learning method. We employ BERT to encode both the corrupted and corresponding correct sentence. Then, we use contrastive learning loss to regularize corrupted tokens' hidden states to be closer to counterparts in the correct sentence. On three CSC datasets, we confirmed our method provides a significant improvement above baselines. | ['Yu Sun', 'Xipeng Qiu', 'Hang Yan', 'Xiaotian Zhang'] | 2022-10-31 | null | null | null | null | ['chinese-spell-checking'] | ['natural-language-processing'] | [ 3.60847622e-01 -2.01669782e-01 -2.79172450e-01 -4.83742923e-01
-1.20444214e+00 -4.78046119e-01 3.85876179e-01 2.04693750e-01
-6.70483053e-01 6.15159035e-01 6.41229987e-01 -4.72896725e-01
7.79162586e-01 -3.73827070e-01 -9.05107796e-01 -4.39207047e-01
2.82776803e-01 9.79487300e-02 2.90621370e-01 -6.86158910e-02
1.87255487e-01 2.66591627e-02 -1.11947072e+00 8.31412554e-01
1.19637620e+00 6.54603839e-01 6.09231651e-01 7.19851077e-01
-4.09399003e-01 9.85306740e-01 -8.77595663e-01 -5.82151055e-01
4.96650748e-02 -6.70503020e-01 -8.16148400e-01 -3.59295070e-01
4.57122058e-01 -2.54571289e-01 -2.98962533e-01 1.24517691e+00
4.49946135e-01 2.45503202e-01 5.44875383e-01 -8.47599685e-01
-1.06889534e+00 1.25545931e+00 -2.30093017e-01 3.94961983e-01
3.91563773e-01 2.09333580e-02 1.25929379e+00 -7.82368600e-01
3.14361870e-01 1.35776794e+00 4.54690069e-01 8.97083282e-01
-1.17538762e+00 -8.58694434e-01 3.76529276e-01 5.13755560e-01
-1.28065348e+00 -6.37426615e-01 9.23405707e-01 2.25836225e-02
1.40420580e+00 3.54076236e-01 4.49826956e-01 1.11292934e+00
-5.38070686e-02 1.40049744e+00 1.08511376e+00 -8.37218881e-01
1.89567566e-01 -9.03397948e-02 4.10169996e-02 2.54727125e-01
-1.48287103e-01 -1.82923421e-01 -5.78783214e-01 2.22683921e-01
3.59835595e-01 -3.53974819e-01 -2.14751586e-01 3.70403171e-01
-8.14549625e-01 5.43725371e-01 2.43263006e-01 3.68706971e-01
-7.85161853e-02 4.12934273e-01 5.26367068e-01 1.98545158e-01
2.39875495e-01 4.46160793e-01 -6.22944176e-01 -5.64291120e-01
-7.99516320e-01 4.96378029e-03 6.81370795e-01 9.42747235e-01
4.64653850e-01 2.78912485e-01 -2.23268688e-01 9.05523121e-01
2.92682111e-01 4.90550935e-01 8.82938802e-01 -8.24319661e-01
6.56329930e-01 2.47395709e-01 -2.27385834e-01 -5.40701151e-01
2.33642295e-01 -3.06064546e-01 -6.89223588e-01 -5.89293122e-01
1.29509911e-01 -4.19561528e-02 -8.00381243e-01 1.88090372e+00
-1.40461758e-01 5.99684000e-01 1.34868383e-01 6.20449901e-01
6.59462094e-01 9.50747490e-01 1.37434736e-01 -2.14849934e-01
9.42672968e-01 -1.03643405e+00 -1.09704983e+00 -3.75742346e-01
8.95418704e-01 -6.05781198e-01 1.78183973e+00 3.64233941e-01
-1.31129217e+00 -5.95850766e-01 -9.67212915e-01 -4.13403392e-01
-1.30937859e-01 -3.95895094e-02 1.36179671e-01 7.11904109e-01
-9.77503181e-01 5.05825579e-01 -9.13066447e-01 4.22429889e-01
9.95929390e-02 1.68316767e-01 3.24752589e-04 -7.05519989e-02
-1.62642753e+00 7.28993535e-01 6.65500045e-01 6.58486858e-02
-5.03416598e-01 -5.45761108e-01 -1.23131967e+00 2.65882045e-01
1.38879806e-01 -1.04597740e-01 1.60587442e+00 -1.18269789e+00
-1.79753935e+00 5.77039480e-01 -5.98911822e-01 -4.99641985e-01
1.67319551e-01 -4.79068339e-01 -6.11609459e-01 5.98012395e-02
-1.54053152e-01 6.04281306e-01 5.98796546e-01 -1.07756746e+00
-4.43671614e-01 9.33982283e-02 -2.39444971e-01 3.81046593e-01
-3.40380043e-01 3.09526235e-01 -6.41997635e-01 -9.34310257e-01
-8.73163640e-02 -6.49227500e-01 8.45266953e-02 -4.21207488e-01
-4.76333201e-01 -3.46132308e-01 6.41103327e-01 -9.76003885e-01
1.60943758e+00 -2.40730214e+00 9.30352733e-02 -1.20333515e-01
-3.38073403e-01 6.36421442e-01 -2.78827310e-01 2.91874230e-01
2.00948596e-01 3.23308259e-01 -4.46822643e-01 -7.49035776e-01
1.85542330e-01 4.74822134e-01 -4.23826307e-01 2.31417298e-01
4.30410564e-01 1.02395380e+00 -8.97689581e-01 -5.67311645e-01
6.77406043e-02 4.63132352e-01 -8.80128264e-01 3.16557050e-01
-3.90395671e-01 2.90237844e-01 1.69192225e-01 2.70937592e-01
8.16531897e-01 -1.12893432e-02 3.43238950e-01 1.79417655e-01
3.48987021e-02 1.19615269e+00 -9.76064026e-01 1.68419230e+00
-5.51804900e-01 5.77880979e-01 -1.58456303e-02 -7.70192206e-01
7.48689532e-01 2.49503195e-01 -1.13730140e-01 -1.11026049e+00
-6.16698042e-02 3.26548874e-01 1.52059108e-01 -2.15808228e-01
7.41680622e-01 -3.02407742e-01 -2.91284770e-01 1.95841327e-01
-2.64474481e-01 -3.88654083e-01 -2.24824116e-01 2.59140283e-01
8.72293234e-01 6.37355670e-02 1.69051796e-01 -6.46182224e-02
6.27269506e-01 -4.79500175e-01 8.96422088e-01 7.36944556e-01
-3.26288313e-01 6.92198277e-01 5.81056774e-01 2.34496444e-01
-7.43252099e-01 -1.15962899e+00 1.21623673e-01 1.06624305e+00
7.00091943e-02 -6.04772389e-01 -9.08418894e-01 -7.08449960e-01
-5.56078553e-02 1.15258503e+00 -1.46859989e-01 -4.78305697e-01
-1.12576604e+00 -1.03739850e-01 1.02284896e+00 7.93416679e-01
5.07489026e-01 -1.18880284e+00 -1.44284382e-01 3.39814961e-01
-3.79201442e-01 -1.14294744e+00 -9.95190799e-01 2.16378167e-01
-4.94983941e-01 -5.71005583e-01 -4.45582539e-01 -1.24425232e+00
1.76106766e-01 -6.61658851e-05 8.99222612e-01 3.71536613e-01
2.16536894e-01 -2.05123797e-01 -5.08490801e-01 -2.48628676e-01
-7.84058988e-01 6.07881323e-02 -1.29714280e-01 -2.16750830e-01
4.25505191e-01 -6.39269426e-02 -2.22199336e-01 -2.21733302e-01
-9.90626454e-01 6.50820658e-02 2.16344759e-01 8.61652434e-01
6.19573176e-01 -2.08849870e-02 5.46847343e-01 -1.00940442e+00
6.38374090e-01 -2.40425095e-01 -4.77710217e-01 2.27283388e-01
-4.20398414e-01 2.24882290e-01 9.46449280e-01 -4.40585434e-01
-1.33325040e+00 -3.70063260e-02 -6.24161959e-01 -1.49805948e-01
-1.32075295e-01 6.72450244e-01 -6.52076244e-01 5.29419184e-01
2.18180105e-01 3.45863938e-01 -3.66170734e-01 -6.59628510e-01
4.84472781e-01 1.02041042e+00 8.17218900e-01 -7.76483893e-01
2.47589871e-01 -1.27278611e-01 -6.12352252e-01 -7.44395256e-01
-7.59941339e-01 -3.42165321e-01 -5.58620512e-01 2.35370129e-01
7.76629090e-01 -9.37559307e-01 -5.20358443e-01 7.64169335e-01
-1.42544127e+00 -7.15995550e-01 -2.80746967e-01 5.09003639e-01
-3.00671726e-01 8.56344581e-01 -1.32875693e+00 -8.80064070e-01
-2.46022657e-01 -1.08692181e+00 9.55383539e-01 -8.45401268e-03
-1.66243270e-01 -9.45958674e-01 3.76609415e-02 8.31774920e-02
3.89252931e-01 -4.30182129e-01 1.29674208e+00 -6.49654269e-01
-3.29839319e-01 3.84604260e-02 -1.17875412e-01 8.35116804e-01
2.88337231e-01 -2.69346517e-02 -1.13715768e+00 -1.20907523e-01
-4.44711819e-02 -1.78857908e-01 1.05095077e+00 2.35260412e-01
1.44844532e+00 -6.74618423e-01 1.20813884e-01 6.72344148e-01
1.17371058e+00 3.19117129e-01 7.95704663e-01 8.72988105e-02
8.66483092e-01 3.31569850e-01 2.69301653e-01 3.19709569e-01
6.04952514e-01 6.22055173e-01 2.87289560e-01 2.39122465e-01
-3.95965517e-01 -6.73390687e-01 6.27724171e-01 1.74190843e+00
5.56349516e-01 -4.66629475e-01 -9.69271541e-01 5.25056958e-01
-1.57055688e+00 -8.77515495e-01 -1.48504287e-01 2.07376575e+00
1.50528169e+00 3.33510041e-01 -2.23303169e-01 1.84520200e-01
8.18402588e-01 1.75971285e-01 -2.69691080e-01 -8.41911912e-01
-5.59286594e-01 4.70052660e-01 1.82718322e-01 8.20888698e-01
-9.79869068e-01 1.50850809e+00 6.52153778e+00 9.28027213e-01
-1.23025429e+00 1.45502150e-01 3.71911794e-01 5.94554245e-02
-8.38387430e-01 1.07567199e-01 -7.88028896e-01 9.42344904e-01
1.26372778e+00 3.44693251e-02 5.11297166e-01 4.28858459e-01
1.06138878e-01 9.42969173e-02 -1.00342429e+00 9.53195095e-01
1.02635689e-01 -1.27376723e+00 1.86331630e-01 -4.13539976e-01
8.86460304e-01 -1.58351630e-01 7.21606761e-02 6.65715277e-01
2.90295184e-01 -8.28125119e-01 1.07222641e+00 1.91056997e-01
8.80141199e-01 -7.03248382e-01 5.99012673e-01 4.24261957e-01
-1.17136860e+00 4.80411835e-02 -2.37874314e-01 -3.75041902e-01
2.80508310e-01 1.08094208e-01 -7.01953709e-01 1.53819770e-01
3.03150952e-01 7.81695902e-01 -6.06735587e-01 1.11948097e+00
-6.49600804e-01 1.29135048e+00 1.28949240e-01 -2.82572210e-01
5.39848693e-02 9.38787609e-02 2.60208040e-01 1.75529337e+00
1.01130664e-01 -1.02221668e-01 1.34217724e-01 6.46950364e-01
-3.41963887e-01 2.07574666e-01 -1.99916691e-01 -1.33839026e-01
9.52946782e-01 3.06179285e-01 -2.98086315e-01 -5.60511768e-01
-4.21177596e-01 1.50390089e+00 6.30910099e-01 3.29268068e-01
-7.96707034e-01 -5.79108000e-01 6.81045413e-01 -2.79343933e-01
4.76702422e-01 -3.79048914e-01 -3.95166904e-01 -1.32813060e+00
2.44977288e-02 -9.92812276e-01 2.88652748e-01 -7.23282516e-01
-1.14478171e+00 4.03043061e-01 -1.80146381e-01 -8.22332978e-01
-3.99704427e-01 -4.75491554e-01 -6.86119556e-01 9.61555362e-01
-1.81179833e+00 -8.91398787e-01 2.86880702e-01 4.16927814e-01
8.23870957e-01 1.09044395e-01 7.17075408e-01 4.87939596e-01
-6.22658730e-01 9.90514100e-01 8.03268030e-02 3.86905998e-01
8.33223224e-01 -1.51093090e+00 8.37324619e-01 1.22757721e+00
1.90053999e-01 6.09824300e-01 4.13747609e-01 -8.06186855e-01
-1.15959167e+00 -1.04737556e+00 1.52315235e+00 -1.19339682e-01
5.06537855e-01 -5.81613004e-01 -1.40393353e+00 7.31319726e-01
4.32237506e-01 -5.01371734e-02 6.38569176e-01 -1.78147227e-01
-6.53352857e-01 1.07911332e-02 -7.62084544e-01 7.67298877e-01
8.36487710e-01 -1.14224100e+00 -8.10625792e-01 -1.54051051e-01
1.32859218e+00 -5.91313839e-01 -2.35830277e-01 2.41182297e-02
1.73053563e-01 -5.66353500e-01 4.69671845e-01 -7.76158392e-01
3.40176284e-01 -3.92769724e-02 -4.46292698e-01 -1.46138644e+00
-3.42501014e-01 -7.11268902e-01 -2.53059030e-01 1.39641893e+00
4.94173884e-01 -3.70279104e-01 5.49638271e-01 6.00676537e-01
-7.07062721e-01 -2.31007174e-01 -1.10865748e+00 -9.82620597e-01
5.88477373e-01 -7.45489180e-01 8.27995777e-01 9.77530241e-01
3.20355892e-01 2.23922163e-01 -3.94045800e-01 4.37173918e-02
1.47115186e-01 -3.40805173e-01 1.37644038e-01 -8.52836788e-01
-6.17018104e-01 -5.98196089e-01 -3.59563380e-02 -1.42305923e+00
8.16989481e-01 -1.15832770e+00 6.16978467e-01 -1.36015129e+00
5.13326973e-02 -4.92918283e-01 -4.98010695e-01 3.80255878e-01
-7.02908456e-01 5.36670536e-02 4.27950233e-01 -1.19113818e-01
-6.49001479e-01 6.87141657e-01 9.84908760e-01 -1.62938952e-01
-3.18290114e-01 -1.48559064e-01 -5.32646775e-01 4.39892352e-01
1.04540718e+00 -3.77821177e-01 -7.43632913e-02 -1.00118184e+00
3.05013537e-01 -3.01210359e-02 1.66779093e-03 -8.19549441e-01
1.98343351e-01 -1.64130911e-01 1.42926378e-02 -4.57895696e-01
4.38807338e-01 -5.36822498e-01 -3.20684612e-01 6.35511220e-01
-8.90905499e-01 2.74352849e-01 4.32389289e-01 3.31648946e-01
-3.57601106e-01 -4.40032482e-01 6.65643096e-01 6.30209148e-02
-7.79381096e-01 -7.17115551e-02 -6.69292748e-01 5.39921820e-01
3.42132211e-01 -6.06254712e-02 -3.25980157e-01 -5.12647867e-01
-1.70343205e-01 4.62013949e-03 3.54888976e-01 4.02311563e-01
7.16823757e-01 -1.31783748e+00 -6.59302175e-01 5.35153091e-01
-7.36143440e-02 4.73736180e-03 5.00540659e-02 4.70366597e-01
-5.18264174e-01 3.63373846e-01 2.07332805e-01 -1.98820904e-01
-1.17552137e+00 4.33637887e-01 2.72487879e-01 1.38535321e-01
-4.13931638e-01 1.07174468e+00 -3.70204933e-02 -4.88005877e-01
5.14282644e-01 -6.53964520e-01 1.01728417e-01 -3.73702586e-01
5.17016947e-01 2.94636428e-01 7.41604939e-02 -6.26417518e-01
-4.45475817e-01 -2.42692381e-02 -1.69803530e-01 -4.17345971e-01
9.04396415e-01 -2.31555611e-01 -2.38410737e-02 5.94777048e-01
1.47622752e+00 5.22942901e-01 -1.37750614e+00 -4.04651701e-01
3.68712991e-01 -4.84741330e-01 -3.77665344e-03 -7.40005016e-01
-7.52955675e-01 1.15530121e+00 1.07448876e-01 -1.13290079e-01
9.21160400e-01 -4.86141406e-02 1.13545644e+00 3.51645947e-01
-4.49950062e-02 -1.38351071e+00 1.32042378e-01 1.09660220e+00
5.37657380e-01 -1.15370762e+00 -6.89325273e-01 -4.16778564e-01
-9.54889715e-01 7.02022374e-01 6.95252240e-01 -1.36031210e-01
5.84035218e-01 5.57675481e-01 1.76239714e-01 5.61293960e-01
-8.96682620e-01 -1.47852618e-02 2.10908219e-01 5.73956430e-01
8.59974921e-01 2.80101031e-01 -2.46078774e-01 7.81733036e-01
-4.33518231e-01 -3.10741544e-01 5.78499079e-01 8.86802197e-01
-4.04025376e-01 -1.24459386e+00 9.79924388e-03 3.13817769e-01
-6.45983517e-01 -7.41795659e-01 -3.97139341e-01 4.65405136e-01
-3.15969922e-02 9.98618245e-01 4.11384344e-01 -2.64499307e-01
1.94159076e-01 1.70886636e-01 4.96149719e-01 -9.37495053e-01
-7.77725041e-01 3.56782913e-01 6.22291714e-02 -3.22795719e-01
-3.01587209e-02 -7.53493547e-01 -1.74765956e+00 -2.71768104e-02
-4.60422844e-01 2.23313436e-01 4.79065299e-01 9.37318683e-01
2.08462447e-01 4.35836375e-01 5.15685916e-01 -2.69197792e-01
-7.07641244e-01 -7.76738942e-01 -5.95917702e-01 4.65727657e-01
5.29425085e-01 -7.64720095e-03 -3.24447155e-01 2.37048157e-02] | [10.939059257507324, 10.649880409240723] |
b31c3b1b-c3e0-48ed-b7c4-cf968bd6c719 | continuous-offline-handwriting-recognition | 2112.13328 | null | https://arxiv.org/abs/2112.13328v1 | https://arxiv.org/pdf/2112.13328v1.pdf | Continuous Offline Handwriting Recognition using Deep Learning Models | Handwritten text recognition is an open problem of great interest in the area of automatic document image analysis. The transcription of handwritten content present in digitized documents is significant in analyzing historical archives or digitizing information from handwritten documents, forms, and communications. In the last years, great advances have been made in this area due to applying deep learning techniques to its resolution. This Thesis addresses the offline continuous handwritten text recognition (HTR) problem, consisting of developing algorithms and models capable of transcribing the text present in an image without the need for the text to be segmented into characters. For this purpose, we have proposed a new recognition model based on integrating two types of deep learning architectures: convolutional neural networks (CNN) and sequence-to-sequence (seq2seq) models, respectively. The convolutional component of the model is oriented to identify relevant features present in characters, and the seq2seq component builds the transcription of the text by modeling the sequential nature of the text. For the design of this new model, an extensive analysis of the capabilities of different convolutional architectures in the simplified problem of isolated character recognition has been carried out in order to identify the most suitable ones to be integrated into the continuous model. Additionally, extensive experimentation of the proposed model for the continuous problem has been carried out to determine its robustness to changes in parameterization. The generalization capacity of the model has also been validated by evaluating it on three handwritten text databases using different languages: IAM in English, RIMES in French, and Osborne in Spanish, respectively. The new proposed model provides competitive results with those obtained with other well-established methodologies. | ['Jorge Sueiras'] | 2021-12-26 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 4.48313653e-01 -5.13673842e-01 2.27000162e-01 -3.15408230e-01
-2.67618388e-01 -5.90994418e-01 6.92241669e-01 6.34492338e-02
-5.73292613e-01 6.54122531e-01 -5.15730307e-02 -2.84819335e-01
-2.54316092e-01 -6.72517061e-01 -4.67507631e-01 -7.17038155e-01
1.38905153e-01 5.71061432e-01 4.62381281e-02 -2.61022776e-01
7.32236147e-01 1.09290302e+00 -1.46197236e+00 5.52988231e-01
5.72845578e-01 8.51798892e-01 3.32097650e-01 1.07401276e+00
-4.18917298e-01 1.03349900e+00 -7.96490848e-01 -1.77014574e-01
1.01726390e-01 -6.78500116e-01 -8.67796361e-01 4.68417108e-01
2.08197922e-01 -3.41692686e-01 -3.82528573e-01 7.79711902e-01
4.29641366e-01 1.12045415e-01 8.20427120e-01 -3.78302276e-01
-5.02707064e-01 6.57983899e-01 -1.49273843e-01 1.12685531e-01
2.94178843e-01 -2.83857077e-01 6.31508231e-01 -9.14425731e-01
7.45110393e-01 8.95697236e-01 6.27951801e-01 2.29799017e-01
-8.54822278e-01 -5.51174134e-02 -2.23910362e-01 1.90026879e-01
-1.34212101e+00 -2.15233520e-01 6.41675830e-01 -6.41182780e-01
1.16721618e+00 3.49079937e-01 3.26877028e-01 9.65814531e-01
3.26524556e-01 8.84265721e-01 9.67604637e-01 -9.64334071e-01
2.60434449e-01 2.26756260e-01 4.86456990e-01 5.51319420e-01
-7.92711275e-04 -1.93082511e-01 -2.95094460e-01 3.25636715e-01
7.18717813e-01 -3.37491512e-01 -7.29762614e-02 2.45001540e-02
-8.10218453e-01 7.77376652e-01 1.28770424e-02 9.63195801e-01
-2.74993092e-01 -1.39342383e-01 7.34610379e-01 2.02661753e-01
6.32091239e-02 2.94394553e-01 -6.63524196e-02 -4.12967116e-01
-1.32019150e+00 1.78072974e-01 1.17387795e+00 8.21966410e-01
4.11170781e-01 2.96223670e-01 -1.88541234e-01 9.24499333e-01
1.57980770e-02 2.70901620e-01 9.23458278e-01 -2.25005865e-01
5.75349152e-01 6.31128371e-01 -5.76416925e-02 -1.01203215e+00
-2.92043000e-01 -2.61413306e-01 -9.70583498e-01 1.46443978e-01
4.46841836e-01 -1.49119556e-01 -1.14576387e+00 7.47839093e-01
-3.20537150e-01 -6.63869560e-01 1.99460939e-01 7.53812075e-01
5.32424271e-01 1.10020387e+00 -4.55780506e-01 -3.76771726e-02
9.79975283e-01 -8.98272812e-01 -8.43126893e-01 2.76379287e-02
3.30946982e-01 -9.59439337e-01 6.04330122e-01 8.57574463e-01
-7.47887135e-01 -8.08806062e-01 -1.23297346e+00 -1.77593067e-01
-5.22801697e-01 9.07867730e-01 2.89392937e-02 7.71101952e-01
-7.35619426e-01 6.46252036e-01 -7.77066290e-01 -4.50381637e-01
1.02668703e-01 4.60716516e-01 -2.49526277e-01 3.95666286e-02
-7.90259361e-01 8.64139557e-01 6.63691938e-01 5.96697092e-01
-7.75355041e-01 5.50338104e-02 -3.88241738e-01 2.35709310e-01
5.69718815e-02 1.83976784e-01 1.03036880e+00 -1.57780361e+00
-1.72652173e+00 6.99103773e-01 -5.82414716e-02 -3.58658791e-01
1.01044273e+00 -1.00549974e-01 -5.30461252e-01 2.09839419e-01
-3.75272572e-01 1.36096701e-01 8.30769122e-01 -9.30022836e-01
-5.26300550e-01 -4.29028660e-01 -5.55807233e-01 -4.01998684e-02
-3.43370199e-01 1.83992058e-01 -5.29271424e-01 -9.31420505e-01
4.96422127e-02 -6.76369429e-01 1.85871676e-01 -4.67813849e-01
-3.08565080e-01 4.96877357e-02 9.45680320e-01 -1.12817574e+00
1.02105629e+00 -2.13859510e+00 3.75316173e-01 4.60780621e-01
-4.68598336e-01 5.74307501e-01 -6.58346862e-02 9.26532269e-01
-7.95045719e-02 -1.28956825e-01 -4.18741375e-01 -1.51037529e-01
-1.84980094e-01 2.97833085e-01 -4.89775360e-01 2.88322270e-01
2.26509839e-01 4.38362807e-01 -3.27864766e-01 -3.79552692e-01
1.87759146e-01 4.63731855e-01 1.10186622e-01 1.66236371e-01
-3.20477396e-01 1.22169375e-01 -3.57783169e-01 5.96394300e-01
4.77584571e-01 2.14869574e-01 2.97676414e-01 1.10674739e-01
-5.32066047e-01 -2.06996053e-01 -1.22046363e+00 1.22541845e+00
-2.64424473e-01 9.92592514e-01 -1.35164276e-01 -1.32177031e+00
1.51662743e+00 3.31566364e-01 2.30483353e-01 -6.52857482e-01
4.32427943e-01 5.23553550e-01 -7.49485865e-02 -7.08230674e-01
1.02295744e+00 1.09608613e-01 2.57239580e-01 3.47819239e-01
1.68834537e-01 1.40229180e-01 7.25384831e-01 -3.78882140e-01
6.12816155e-01 1.73555121e-01 -4.80070198e-03 -1.02542415e-01
9.80648339e-01 1.64209023e-01 -6.42890334e-02 6.94383740e-01
3.43880653e-01 6.35294974e-01 5.89803100e-01 -7.28346586e-01
-1.31793690e+00 -4.08628821e-01 -1.26161709e-01 6.17264271e-01
-4.20026362e-01 1.79449528e-01 -8.30196738e-01 -2.32497871e-01
-2.44275436e-01 5.12829363e-01 -6.06930971e-01 3.52412254e-01
-7.17011154e-01 -7.38857925e-01 9.18125987e-01 5.25232553e-01
7.56169260e-01 -1.29549205e+00 -6.35512352e-01 4.78658617e-01
4.85748835e-02 -8.73899221e-01 5.18673323e-02 5.37074625e-01
-1.03938293e+00 -8.49631727e-01 -1.14014852e+00 -1.04646957e+00
5.00966489e-01 -3.67598146e-01 3.84910643e-01 -1.69729888e-01
-5.74469626e-01 1.95984960e-01 -8.66523027e-01 -2.42488280e-01
-8.56725156e-01 5.60889363e-01 -4.40777004e-01 3.49442601e-01
2.59428680e-01 4.27236818e-02 2.60405064e-01 2.26377189e-01
-1.28053021e+00 -2.67768353e-01 8.33858311e-01 9.82079566e-01
3.01233619e-01 1.96353033e-01 3.35057855e-01 -7.97102749e-01
8.65458965e-01 -3.56641822e-02 -8.52439523e-01 4.87463415e-01
-3.64613891e-01 1.47537878e-02 9.46725070e-01 -3.21303070e-01
-1.06097555e+00 3.39377373e-01 -4.28556442e-01 -4.30931337e-02
-3.86914641e-01 8.51466656e-01 7.33928233e-02 -1.29788890e-01
6.03264511e-01 9.08361852e-01 2.10519880e-01 -6.63694859e-01
-1.66431218e-01 1.12863350e+00 6.37738049e-01 -4.88146514e-01
2.41369456e-01 8.46766979e-02 -1.18924305e-01 -1.48492670e+00
-2.39723250e-01 -3.27484637e-01 -1.00869453e+00 -3.35804582e-01
7.00720310e-01 -3.69776756e-01 -4.57174242e-01 1.22578502e+00
-1.25641513e+00 -2.16599673e-01 -1.13354824e-01 2.98714876e-01
-5.32629073e-01 6.96989000e-01 -8.21409285e-01 -8.39047670e-01
-3.27469587e-01 -1.31188715e+00 6.86973393e-01 2.16611937e-01
1.03297317e-03 -1.03853202e+00 1.53815672e-01 1.56422883e-01
1.97468802e-01 5.00851125e-02 1.06561005e+00 -9.36115742e-01
-4.22510952e-01 -7.65480876e-01 -5.01262955e-02 8.55418444e-01
1.32966653e-01 4.90022480e-01 -9.20414150e-01 -2.44821504e-01
-3.25854346e-02 -3.07336748e-01 8.77323985e-01 1.20542422e-01
1.02882719e+00 -8.02409127e-02 1.66859224e-01 3.76280844e-01
1.62519205e+00 7.09961593e-01 1.04873586e+00 5.68139970e-01
3.35245401e-01 4.70877111e-01 3.62549394e-01 6.31431937e-01
-4.70497310e-01 4.57842439e-01 1.34452665e-02 6.88531920e-02
2.65481994e-02 -2.62515545e-02 4.42792714e-01 8.40478122e-01
-1.83299020e-01 -6.43795133e-01 -1.05572248e+00 2.92509437e-01
-1.59270620e+00 -9.01363254e-01 -2.25210845e-01 1.87922072e+00
4.73070472e-01 5.33000492e-02 -5.03290705e-02 7.53670394e-01
7.79181063e-01 -6.97409734e-02 -1.45324677e-01 -9.62298155e-01
-4.04337943e-01 3.53066325e-01 4.22737360e-01 4.63050753e-01
-1.17573225e+00 7.86653280e-01 6.42385197e+00 8.64716887e-01
-1.66753709e+00 -6.07606709e-01 4.52119231e-01 5.16119659e-01
3.21978271e-01 -3.30507398e-01 -9.16318774e-01 4.14593399e-01
8.69293869e-01 2.04819605e-01 3.77048135e-01 6.51486397e-01
2.66529441e-01 -3.08598369e-01 -9.87897336e-01 7.56837130e-01
4.85273868e-01 -1.42516720e+00 5.16804636e-01 -5.64672649e-02
6.49210930e-01 -3.79325390e-01 -8.16070884e-02 -1.00769930e-01
-2.31837854e-01 -1.14970708e+00 9.39063549e-01 8.76962781e-01
5.68049014e-01 -8.27997744e-01 1.21434426e+00 3.56803715e-01
-7.44047403e-01 -3.48513335e-01 -3.05642933e-01 -6.89668432e-02
-7.02123791e-02 2.46196136e-01 -1.03277218e+00 8.95279109e-01
2.26915106e-01 5.87324679e-01 -6.44194603e-01 1.02388728e+00
-5.63074201e-02 5.93518436e-01 3.11438572e-02 -5.30142844e-01
6.38637304e-01 -2.98788309e-01 2.58340687e-01 1.60887551e+00
3.76541495e-01 -1.45128414e-01 -1.49890900e-01 8.58935952e-01
2.05128878e-01 4.66647923e-01 -3.31705838e-01 -5.90081334e-01
-7.50133321e-02 9.80845869e-01 -1.05627549e+00 -3.18145901e-01
-9.06800851e-02 1.18607271e+00 -3.34782377e-02 2.81201273e-01
-4.38743681e-01 -9.90374386e-01 -1.89663351e-01 -1.02298260e-01
5.84039152e-01 -4.48399007e-01 -3.20261896e-01 -8.23958814e-01
3.88448834e-02 -1.11664772e+00 1.24993183e-01 -7.15392649e-01
-8.87878597e-01 8.66979659e-01 -3.13345969e-01 -9.37909961e-01
-4.37496275e-01 -1.10778058e+00 -3.84618431e-01 1.23586977e+00
-9.99280989e-01 -1.00640965e+00 -3.06004673e-01 4.25163954e-01
8.70704114e-01 -6.16367996e-01 8.07087958e-01 2.64764100e-01
-6.97081566e-01 5.45198500e-01 6.88917398e-01 6.40681505e-01
3.92899662e-01 -8.93894076e-01 5.26260063e-02 1.02698779e+00
1.38585567e-01 2.36564502e-01 4.72436309e-01 -6.56477690e-01
-1.30410743e+00 -8.10093403e-01 9.96808708e-01 4.64907028e-02
4.31289732e-01 -2.77124465e-01 -8.88073981e-01 5.36834538e-01
8.23964849e-02 -5.82745552e-01 3.04909557e-01 -5.84885657e-01
-1.50305614e-01 -9.78699774e-02 -8.01756620e-01 9.80108753e-02
-1.02565087e-01 -4.30116147e-01 -7.07598209e-01 2.53776275e-02
-8.85129571e-02 -3.70786250e-01 -7.11654842e-01 -1.36278689e-01
6.87975466e-01 -9.10809577e-01 3.61527383e-01 -4.45877254e-01
8.28423083e-01 -1.10641651e-01 -9.57591906e-02 -8.31274748e-01
-8.98303762e-02 -2.50237107e-01 3.73047858e-01 1.23268449e+00
3.55314732e-01 -3.15223038e-01 8.37853432e-01 1.60660520e-01
-2.31484339e-01 -4.09377694e-01 -7.68231452e-01 -8.16735268e-01
1.88208178e-01 -1.06726438e-01 1.47539005e-01 7.33388543e-01
-3.10950220e-01 5.03748506e-02 -3.61050367e-01 -3.98798063e-02
9.36651826e-02 8.34365040e-02 6.15680814e-01 -1.07669997e+00
-1.58244073e-01 -6.03768468e-01 -6.42352104e-01 -9.46229696e-01
-1.52067877e-02 -6.97560966e-01 1.08212039e-01 -1.33049250e+00
-6.63630813e-02 -9.23875496e-02 2.03647986e-01 4.05592859e-01
4.74500507e-01 -6.15664907e-02 3.02766383e-01 2.81225443e-01
1.87848672e-01 2.88569838e-01 8.58428597e-01 -3.72910023e-01
-1.83211565e-01 8.95673931e-02 1.11786075e-01 3.45698804e-01
6.28620803e-01 -8.42218623e-02 -1.33838449e-02 -4.93531078e-01
3.50312069e-02 1.09101668e-01 3.38225104e-02 -1.11051106e+00
3.27055305e-01 2.66797096e-01 6.88125014e-01 -8.72094035e-01
2.44233496e-02 -8.32989216e-01 1.54634684e-01 6.10796928e-01
-5.75813830e-01 1.21049069e-01 3.57276320e-01 1.82812989e-01
-6.73034787e-01 -9.53134298e-01 7.57058561e-01 -2.81253964e-01
-9.54218805e-01 -2.82559872e-01 -1.04447865e+00 -3.96156222e-01
8.59853208e-01 -7.10082293e-01 5.78452833e-02 -1.72550753e-01
-8.17606747e-01 -3.43451977e-01 2.23964825e-01 3.66271824e-01
6.21254265e-01 -7.44675875e-01 -6.76688850e-01 3.45234245e-01
-1.16175279e-01 -3.03699315e-01 1.75849408e-01 5.94995618e-01
-1.43699813e+00 7.64600337e-01 -6.94160163e-01 -4.19397503e-01
-1.38599658e+00 3.75858456e-01 5.00588834e-01 -2.30808631e-01
-3.19041222e-01 3.98835778e-01 -6.59269273e-01 -1.28679514e-01
5.90811849e-01 -3.83418620e-01 -4.50520396e-01 3.67526948e-01
4.84558105e-01 6.22696400e-01 6.26870394e-01 -6.99114561e-01
-4.79707532e-02 6.52999341e-01 -1.32790387e-01 -1.85075611e-01
1.43298280e+00 3.14540297e-01 -4.62039888e-01 6.21859312e-01
1.09613669e+00 -1.74912691e-01 -7.79246151e-01 2.38804609e-01
2.87340105e-01 -1.55211776e-01 -9.31537822e-02 -1.01867485e+00
-8.50510180e-01 1.12244010e+00 6.05033576e-01 2.31146395e-01
1.12002456e+00 -8.67352009e-01 2.96193600e-01 8.64961565e-01
1.27151996e-01 -1.46238685e+00 -5.01505062e-02 9.48845267e-01
9.94726598e-01 -7.53577650e-01 -4.79969867e-02 -9.42093059e-02
-5.37363827e-01 2.15591359e+00 1.51873708e-01 -7.39949346e-02
4.29386050e-01 3.83493483e-01 2.96679772e-02 7.25861415e-02
-2.25861222e-01 1.72839135e-01 1.46389052e-01 3.08540910e-01
6.84987068e-01 -1.93555325e-01 -5.60456812e-01 4.14505064e-01
1.49567770e-02 3.64887357e-01 9.91161466e-01 1.20847285e+00
-3.81323785e-01 -1.19491160e+00 -7.37391770e-01 2.39651188e-01
-4.98742133e-01 1.69184238e-01 -9.47913945e-01 9.59871054e-01
1.77835990e-02 6.23566329e-01 8.92611444e-02 -1.93547547e-01
3.56826752e-01 3.72630745e-01 4.91039753e-01 -2.42799520e-01
-9.29658771e-01 8.86386335e-02 4.48601414e-03 1.34305075e-01
-3.48441482e-01 -7.01780438e-01 -9.43045735e-01 -8.61850455e-02
-2.28927657e-01 1.26820162e-01 1.08012462e+00 8.89324486e-01
-1.07248351e-01 5.95365942e-01 6.02747440e-01 -8.72989535e-01
-8.13589692e-01 -1.11841440e+00 -7.81385124e-01 5.32582141e-02
2.09679112e-01 -8.54975283e-02 -5.90688549e-02 4.74160045e-01] | [11.813530921936035, 2.627074956893921] |
2aec6988-e224-4640-bbf9-c48bced70268 | cascaded-information-enhancement-and-cross | 2302.08670 | null | https://arxiv.org/abs/2302.08670v1 | https://arxiv.org/pdf/2302.08670v1.pdf | Cascaded information enhancement and cross-modal attention feature fusion for multispectral pedestrian detection | Multispectral pedestrian detection is a technology designed to detect and locate pedestrians in Color and Thermal images, which has been widely used in automatic driving, video surveillance, etc. So far most available multispectral pedestrian detection algorithms only achieved limited success in pedestrian detection because of the lacking take into account the confusion of pedestrian information and background noise in Color and Thermal images. Here we propose a multispectral pedestrian detection algorithm, which mainly consists of a cascaded information enhancement module and a cross-modal attention feature fusion module. On the one hand, the cascaded information enhancement module adopts the channel and spatial attention mechanism to perform attention weighting on the features fused by the cascaded feature fusion block. Moreover, it multiplies the single-modal features with the attention weight element by element to enhance the pedestrian features in the single-modal and thus suppress the interference from the background. On the other hand, the cross-modal attention feature fusion module mines the features of both Color and Thermal modalities to complement each other, then the global features are constructed by adding the cross-modal complemented features element by element, which are attentionally weighted to achieve the effective fusion of the two modal features. Finally, the fused features are input into the detection head to detect and locate pedestrians. Extensive experiments have been performed on two improved versions of annotations (sanitized annotations and paired annotations) of the public dataset KAIST. The experimental results show that our method demonstrates a lower pedestrian miss rate and more accurate pedestrian detection boxes compared to the comparison method. Additionally, the ablation experiment also proved the effectiveness of each module designed in this paper. | ['Kaizheng Wang', 'Kaixiong Xu', 'Yang Yang'] | 2023-02-17 | null | null | null | null | ['pedestrian-detection'] | ['computer-vision'] | [ 1.77903362e-02 -7.61095583e-01 3.48215133e-01 -8.52511972e-02
-6.24500155e-01 -2.43021145e-01 5.25438964e-01 4.42904793e-02
-8.28053355e-01 5.45568705e-01 1.47799775e-01 -2.33241194e-03
2.25207523e-01 -7.61234462e-01 -3.50964189e-01 -1.22691107e+00
3.56911510e-01 -3.81250501e-01 7.17489541e-01 -1.86539739e-01
6.10426925e-02 2.04226211e-01 -1.93455970e+00 2.41645083e-01
9.61116314e-01 9.23382401e-01 3.12603742e-01 6.44194901e-01
8.57663453e-02 3.96677464e-01 -4.52355057e-01 -3.74761850e-01
2.58554608e-01 -9.35973376e-02 -1.17960297e-01 1.97374657e-01
2.67698824e-01 -5.42999208e-01 -2.65084535e-01 1.19243252e+00
8.47272813e-01 1.15137801e-01 4.14742321e-01 -1.33567178e+00
-4.86055702e-01 1.14937015e-01 -1.17879581e+00 5.66788971e-01
2.11709023e-01 6.59322023e-01 5.11476576e-01 -1.04695785e+00
-1.53924286e-01 1.28447390e+00 5.34358799e-01 2.34270111e-01
-7.51794279e-01 -6.98718846e-01 1.36114463e-01 5.51076591e-01
-1.46184587e+00 -2.82458156e-01 7.49075770e-01 -3.77216339e-01
6.43588543e-01 3.25594425e-01 8.83294880e-01 6.55524135e-01
2.22308226e-02 8.59910309e-01 9.97362435e-01 -4.17564213e-01
-2.98114985e-01 4.77014899e-01 2.50034809e-01 8.40272725e-01
4.11423206e-01 4.17705029e-01 -2.09031388e-01 5.67530170e-02
3.68896961e-01 1.12537846e-01 -7.98176974e-02 5.40042967e-02
-1.11210454e+00 5.95960259e-01 6.89476907e-01 4.29919332e-01
-4.49858874e-01 -1.09300651e-01 3.64490390e-01 -2.77346641e-01
2.35545278e-01 -3.53217900e-01 -1.08565800e-01 3.96585047e-01
-6.29418314e-01 3.02504580e-02 7.83225000e-02 7.05307901e-01
1.04521620e+00 -1.52988985e-01 -5.36546826e-01 7.43500292e-01
4.81464982e-01 7.59102046e-01 3.99769932e-01 -3.53630632e-01
5.55114150e-01 7.83149838e-01 3.30896020e-01 -1.07758331e+00
-4.61099476e-01 -5.13197958e-01 -8.77678931e-01 1.63332701e-01
2.83830911e-01 -4.70312834e-01 -7.39103854e-01 1.51647019e+00
5.42945027e-01 1.71826482e-01 -8.45189318e-02 1.31211090e+00
1.04897094e+00 5.89760065e-01 6.20903134e-01 -8.07592869e-02
1.80377805e+00 -9.51199710e-01 -5.61141729e-01 -2.48893768e-01
2.70754874e-01 -1.02372301e+00 6.23371422e-01 -2.46219840e-02
-7.01211870e-01 -1.17178261e+00 -1.12576556e+00 1.11339495e-01
-6.95837200e-01 6.17844999e-01 4.20480907e-01 1.12867773e+00
-8.06774080e-01 -2.15237468e-01 -4.25771594e-01 -4.46865201e-01
3.30932558e-01 2.04503641e-01 -2.12745875e-01 -1.08825758e-01
-1.24366009e+00 8.90768111e-01 6.02734327e-01 3.94884020e-01
-6.73881710e-01 -4.42464471e-01 -6.75652623e-01 -1.69008207e-02
3.38929564e-01 -7.88996994e-01 7.80941546e-01 -8.48903418e-01
-1.00529683e+00 4.79311645e-01 -2.63824403e-01 -2.48553708e-01
5.03002763e-01 -7.42865130e-02 -5.81187844e-01 5.15393689e-02
3.75295222e-01 7.74474025e-01 6.18049562e-01 -1.07469511e+00
-1.48353350e+00 -3.68816406e-01 -2.58802865e-02 4.55883801e-01
-5.43396771e-01 2.17550933e-01 -6.07561886e-01 -2.62460470e-01
-1.11175701e-01 -5.12941897e-01 -1.81710020e-01 -2.60520220e-01
-4.30727273e-01 -1.48557648e-01 1.16722620e+00 -7.66195238e-01
1.02091670e+00 -2.27744722e+00 -1.00602962e-01 2.09707752e-01
9.73504633e-02 4.36027676e-01 -1.64374948e-01 2.19020881e-02
5.35657778e-02 -1.69427678e-01 7.99247697e-02 -3.16057146e-01
-2.17222124e-01 -2.71917313e-01 3.46459001e-01 4.70391691e-01
2.75568247e-01 6.79888308e-01 -9.59956944e-01 -8.25540125e-01
8.49035442e-01 6.71672344e-01 -1.05082236e-01 -2.16560569e-02
4.43346113e-01 4.32142943e-01 -4.71523106e-01 7.26656258e-01
9.60766912e-01 2.23558858e-01 -4.14343596e-01 -5.83576143e-01
-6.43103004e-01 -4.98126000e-01 -1.44355488e+00 1.15128660e+00
-1.20379113e-01 2.72008777e-01 4.67950217e-02 -5.96206725e-01
7.87906885e-01 3.12520802e-01 4.63273287e-01 -7.48721004e-01
3.46144915e-01 1.08050980e-01 -3.25970799e-02 -6.07463598e-01
7.95129061e-01 7.24107474e-02 -2.02656534e-04 2.06395164e-02
-1.60804585e-01 4.12177056e-01 3.62716615e-01 7.24004507e-02
5.79942584e-01 -1.44726574e-01 8.82666484e-02 -1.04340827e-02
1.22201133e+00 -7.77473813e-03 5.46978712e-01 6.48580194e-01
-6.40387118e-01 1.95280448e-01 -1.78720653e-01 -4.16070074e-01
-8.91240478e-01 -8.22101951e-01 1.40167147e-01 1.18877327e+00
5.25579989e-01 -1.83048546e-01 -7.10608959e-01 -6.15667701e-01
-6.27527758e-02 3.34905356e-01 -5.21475255e-01 -2.18240023e-01
-2.72596657e-01 -1.31813300e+00 5.24210095e-01 6.52949274e-01
1.31677222e+00 -6.72947884e-01 -7.95293570e-01 6.05783835e-02
-4.37455505e-01 -9.82146442e-01 -5.39650559e-01 -1.65155783e-01
-5.84315248e-02 -1.19151843e+00 -8.66051137e-01 -7.51535535e-01
5.18790364e-01 8.91392291e-01 3.18449438e-01 2.54273504e-01
-5.49670279e-01 3.95997435e-01 -2.13159069e-01 -4.78594333e-01
8.64023343e-02 -1.67681754e-01 -9.32902694e-02 5.50477564e-01
5.54451525e-01 -3.24919820e-02 -7.86404848e-01 3.71816695e-01
-6.17977321e-01 1.21419385e-01 7.95981824e-01 8.72923017e-01
1.56521484e-01 5.39552271e-01 3.05885732e-01 -1.74306124e-01
3.66266578e-01 -2.31446549e-01 -6.78670287e-01 1.70223728e-01
4.42207232e-03 -4.04336691e-01 3.41905594e-01 -2.48118833e-01
-1.55312502e+00 3.77289683e-01 -1.28768310e-01 -5.25391602e-04
-4.16749716e-01 2.00512618e-01 -5.26188433e-01 -2.08476052e-01
4.08127964e-01 4.59548771e-01 -3.46938670e-01 -2.50420034e-01
3.44294131e-01 7.66401291e-01 6.06080353e-01 -2.93091029e-01
9.08422828e-01 4.91856545e-01 -3.28276902e-02 -9.35227454e-01
-4.38160568e-01 -6.99309945e-01 -7.48031020e-01 -6.98325813e-01
1.25918698e+00 -1.03971338e+00 -8.05901825e-01 7.61404216e-01
-1.18256092e+00 4.50360030e-01 2.42477253e-01 6.12140179e-01
6.89504892e-02 6.52375579e-01 -4.96047795e-01 -1.34534383e+00
-3.84659439e-01 -1.18209994e+00 1.12348604e+00 9.10210013e-01
5.79574227e-01 -5.04695415e-01 -4.77621228e-01 4.31190342e-01
3.71656328e-01 2.57580820e-02 6.19577587e-01 -1.78719819e-01
-6.99904621e-01 -3.57359886e-01 -8.09802890e-01 2.59370804e-01
8.65232795e-02 5.10614626e-02 -1.15144885e+00 -1.63116395e-01
-3.97690445e-01 2.47983038e-01 1.11630857e+00 5.22417843e-01
5.65935969e-01 2.15109691e-01 -5.26270032e-01 9.83172506e-02
1.47574091e+00 3.04961652e-01 5.13425171e-01 3.39270502e-01
6.86000824e-01 5.59843659e-01 7.00334609e-01 3.60544771e-01
4.46503639e-01 5.48318863e-01 4.65925694e-01 -4.87308323e-01
-2.09263682e-01 1.08406231e-01 2.68721431e-01 2.48222187e-01
-3.21287781e-01 9.02138501e-02 -6.71286523e-01 5.98103642e-01
-1.92991972e+00 -1.23531401e+00 -4.42791849e-01 2.26012778e+00
2.61174530e-01 1.00554220e-01 4.85500842e-01 3.83283257e-01
1.23668468e+00 -1.07769653e-01 -2.07409456e-01 2.03795105e-01
-4.81956989e-01 -2.74588674e-01 6.49853885e-01 2.31811002e-01
-1.59110391e+00 6.55500591e-01 4.93702698e+00 8.94937277e-01
-8.08420002e-01 3.26993585e-01 5.74671984e-01 1.37088642e-01
3.41605723e-01 -2.20527649e-01 -9.46754992e-01 6.99699283e-01
4.08561647e-01 1.65428787e-01 3.43200155e-02 5.40896893e-01
3.45294237e-01 -5.55192828e-01 -4.61147606e-01 1.01803195e+00
-4.00532000e-02 -7.43902564e-01 -1.29676580e-01 -4.51226160e-02
4.38626975e-01 -3.20575893e-01 1.70279950e-01 2.82937527e-01
1.04494365e-02 -4.58690375e-01 7.58306682e-01 6.60508811e-01
2.36482129e-01 -1.03126895e+00 1.19891894e+00 3.92551601e-01
-1.90576017e+00 -4.53884900e-01 -3.00481021e-01 7.76878139e-03
2.63126671e-01 5.14988482e-01 -4.31732088e-01 1.01321816e+00
7.98142076e-01 4.85393226e-01 -9.39426959e-01 1.53699708e+00
-1.56442031e-01 2.05812678e-01 -3.27588290e-01 -2.06399426e-01
3.00546885e-01 -6.44455329e-02 5.89349270e-01 1.43362010e+00
2.95973986e-01 4.76848334e-02 4.66222763e-01 6.89851046e-01
5.43815553e-01 2.03143626e-01 -2.11856008e-01 5.85483789e-01
3.91712934e-01 1.60179019e+00 -5.89274347e-01 -5.34203231e-01
-8.25900614e-01 8.17075193e-01 -1.78066745e-01 3.49585325e-01
-1.19065106e+00 -4.53941435e-01 3.76761675e-01 -1.87750727e-01
3.36538076e-01 -6.00661151e-02 -2.61458039e-01 -8.00085008e-01
-3.35507430e-02 -3.20187956e-01 5.42791784e-01 -8.25868785e-01
-1.21440792e+00 3.65397513e-01 -6.80317953e-02 -1.11946547e+00
3.91177535e-01 -6.31831706e-01 -6.74986482e-01 1.32307875e+00
-1.61932397e+00 -1.52732885e+00 -7.23006904e-01 7.26484001e-01
4.10005182e-01 -4.97459210e-02 3.72132123e-01 7.82613575e-01
-1.17153239e+00 4.51125920e-01 -2.46785462e-01 2.74906278e-01
6.03961825e-01 -9.22469318e-01 -1.54955015e-01 1.36630177e+00
-6.30684018e-01 3.06586504e-01 4.78288352e-01 -7.57175922e-01
-9.87136245e-01 -1.28107035e+00 5.93107224e-01 -2.36511969e-05
2.40165338e-01 1.24052063e-01 -5.30275285e-01 2.34942377e-01
2.77699202e-01 1.17225088e-02 5.08431852e-01 -2.80966848e-01
-1.05694495e-01 -3.08856666e-01 -1.13148963e+00 4.23868984e-01
5.96476614e-01 -2.24339664e-01 -4.28720266e-01 1.57395899e-02
3.89697194e-01 -9.52707306e-02 -4.12415415e-01 3.92956108e-01
5.73633730e-01 -1.08576655e+00 1.23657131e+00 -5.20167388e-02
-1.17358332e-02 -1.01411295e+00 -1.98537111e-01 -9.82442260e-01
-6.64587080e-01 1.44501567e-01 3.70330483e-01 1.61423624e+00
1.75563857e-01 -5.82254708e-01 3.54410946e-01 5.04884660e-01
-1.79443389e-01 -1.56347021e-01 -8.82493079e-01 -4.07463372e-01
-4.71839577e-01 -1.37093946e-01 6.76059961e-01 4.07780468e-01
-2.40211636e-01 4.25513357e-01 -4.17497873e-01 5.24022460e-01
9.57338989e-01 -7.41252378e-02 6.48072600e-01 -1.07171834e+00
2.62215156e-02 -4.59804088e-01 -3.69128019e-01 -7.78481781e-01
-3.63090217e-01 -4.23231483e-01 1.75373927e-01 -1.51760101e+00
6.62115574e-01 -2.84023106e-01 -4.36972678e-01 2.59938061e-01
-7.68466651e-01 5.97259104e-01 2.56540000e-01 -1.65933222e-01
-7.40541041e-01 4.44841653e-01 1.03148293e+00 -4.18648273e-01
-1.02947921e-01 -2.37938669e-02 -7.15626240e-01 7.05954671e-01
6.00921452e-01 -1.16168633e-02 2.85855285e-03 -3.63731444e-01
-4.81647283e-01 -3.37392956e-01 9.39713478e-01 -1.50105596e+00
6.09117389e-01 6.36776350e-03 1.08312714e+00 -9.76305544e-01
5.85596800e-01 -8.81747961e-01 -7.94159323e-02 7.27823019e-01
2.17347383e-01 -4.47029993e-02 3.45192611e-01 6.58290029e-01
7.44931912e-03 4.00954336e-02 8.79147530e-01 -2.27501199e-01
-1.23192692e+00 1.87980741e-01 -6.01941645e-01 -5.32441199e-01
1.28253841e+00 -3.18680912e-01 -4.98421192e-01 4.60441262e-02
-3.78940284e-01 3.59401524e-01 -3.11009847e-02 3.84709448e-01
5.84405482e-01 -1.42968678e+00 -9.34068561e-01 4.02935803e-01
1.45499215e-01 -4.76004660e-01 7.21810758e-01 1.03132522e+00
-4.14824747e-02 4.46979254e-01 -4.65636402e-01 -7.81651735e-01
-1.64181232e+00 7.65469372e-01 5.13351798e-01 -8.07191711e-03
-4.56932038e-02 7.47039199e-01 2.23574296e-01 1.23480186e-01
4.28498499e-02 -5.78778144e-03 -6.25771821e-01 1.78267017e-01
8.29521596e-01 6.98609352e-01 -8.82446244e-02 -1.21911776e+00
-5.56388617e-01 6.51786625e-01 1.56233177e-01 -1.30002066e-01
8.52924883e-01 -5.13728082e-01 -3.38784456e-02 -8.81072730e-02
7.89667666e-01 -2.97042932e-02 -9.96745050e-01 -3.45002413e-01
-3.73823941e-01 -6.08403385e-01 1.23446047e-01 -8.39049637e-01
-1.06096756e+00 8.68659079e-01 1.01253998e+00 1.85597599e-01
1.37473249e+00 -3.28590274e-01 6.74867809e-01 -1.48488760e-01
2.63847053e-01 -9.84013319e-01 -1.24525102e-02 1.73604876e-01
3.25630099e-01 -1.32477093e+00 -1.49602816e-01 -6.37175024e-01
-4.47199911e-01 8.96920145e-01 9.59020972e-01 2.02149525e-01
5.19005358e-01 1.55828565e-01 -6.94324747e-02 2.05636963e-01
-1.94666132e-01 -9.51401711e-01 3.58894855e-01 8.38115454e-01
1.59840792e-01 9.06563848e-02 -4.44667190e-01 7.36932337e-01
2.46620059e-01 -2.00932592e-01 5.34778042e-03 6.88595891e-01
-8.76302183e-01 -8.52835894e-01 -9.74743128e-01 2.93283522e-01
-9.57969204e-02 -1.19661115e-01 -1.23816021e-01 5.94790459e-01
9.15266216e-01 1.60338259e+00 -9.86638069e-02 -7.31733739e-01
4.99429703e-01 -3.68456589e-03 3.06748778e-01 -1.03902198e-01
-8.03835869e-01 1.86198294e-01 3.52583230e-02 -2.06678882e-01
-6.15534246e-01 -7.24050164e-01 -8.66156042e-01 -3.66852045e-01
-6.36633456e-01 6.85827509e-02 5.10053396e-01 9.95549738e-01
1.08307607e-01 7.11164176e-01 5.52981436e-01 -1.05318964e+00
-1.46143336e-03 -1.02872503e+00 -4.00367230e-01 3.61627460e-01
2.21683666e-01 -8.40428770e-01 -7.92301074e-02 -1.35312155e-01] | [9.823103904724121, -1.396195650100708] |
f6d2b75d-4a1c-47ab-9c94-79f61e9ad5d4 | 3d-object-proposals-for-accurate-object-class | null | null | http://papers.nips.cc/paper/5644-3d-object-proposals-for-accurate-object-class-detection | http://papers.nips.cc/paper/5644-3d-object-proposals-for-accurate-object-class-detection.pdf | 3D Object Proposals for Accurate Object Class Detection | The goal of this paper is to generate high-quality 3D object proposals in the context of autonomous driving. Our method exploits stereo imagery to place proposals in the form of 3D bounding boxes. We formulate the problem as minimizing an energy function encoding object size priors, ground plane as well as several depth informed features that reason about free space, point cloud densities and distance to the ground. Our experiments show significant performance gains over existing RGB and RGB-D object proposal methods on the challenging KITTI benchmark. Combined with convolutional neural net (CNN) scoring, our approach outperforms all existing results on all three KITTI object classes. | ['Andrew G. Berneshawi', 'Yukun Zhu', 'Sanja Fidler', 'Raquel Urtasun', 'Xiaozhi Chen', 'Huimin Ma', 'Kaustav Kundu'] | 2015-12-01 | null | null | null | neurips-2015-12 | ['vehicle-pose-estimation'] | ['computer-vision'] | [-1.35932192e-01 4.00937855e-01 -9.11635011e-02 -7.61670887e-01
-7.95628130e-01 -4.76793557e-01 8.31497431e-01 -8.66176113e-02
-6.73732519e-01 4.72716302e-01 4.34159413e-02 -9.72784013e-02
1.67263880e-01 -6.89833879e-01 -9.80493844e-01 -1.91306621e-01
2.48924103e-02 1.02117753e+00 7.58783102e-01 -3.97603959e-01
6.65366411e-01 8.42976630e-01 -1.76950073e+00 -1.96096733e-01
4.71975952e-01 1.34745002e+00 1.82104573e-01 4.89784807e-01
-1.41219771e-03 3.37783635e-01 -3.39144319e-01 -3.00142497e-01
7.74893165e-01 5.43109477e-01 -5.12702823e-01 5.41319773e-02
9.62115586e-01 -5.45635760e-01 -5.14282346e-01 8.50724161e-01
2.38856539e-01 2.91956604e-01 8.74181092e-01 -1.47338617e+00
-1.72601387e-01 3.56868021e-02 -4.82699692e-01 1.79113612e-01
1.14447944e-01 5.48782825e-01 1.01592934e+00 -1.15402973e+00
6.83118701e-01 1.55006158e+00 7.91368544e-01 3.98043424e-01
-1.02853286e+00 -6.04259074e-01 2.63834357e-01 8.80268663e-02
-1.40017843e+00 -1.12421580e-01 7.18759954e-01 -5.50238907e-01
1.40671527e+00 -1.38530895e-01 7.47905791e-01 9.79179025e-01
3.35550159e-01 7.63310373e-01 8.87559414e-01 1.36495471e-01
4.70019639e-01 -1.17547296e-01 2.55044810e-02 7.27291167e-01
4.85523403e-01 5.08398235e-01 -6.82483971e-01 -3.14106017e-01
7.39415348e-01 2.64819339e-02 3.29976380e-01 -9.73371327e-01
-1.30601406e+00 1.05587316e+00 9.59878683e-01 -7.57300377e-01
-2.12077737e-01 7.27101624e-01 1.05387993e-01 -5.86023867e-01
5.68290293e-01 4.58172709e-01 -3.53052706e-01 1.05787188e-01
-6.02562428e-01 1.08394563e+00 4.14265156e-01 1.35744286e+00
1.09246182e+00 -6.60765022e-02 -1.35311186e-01 3.07835042e-01
7.82332599e-01 8.80255520e-01 -6.34767041e-02 -1.33261549e+00
7.16826499e-01 6.14367545e-01 5.08190274e-01 -6.46569490e-01
-3.52225274e-01 -2.32660055e-01 -7.55148083e-02 7.57801592e-01
2.39763558e-02 3.81167859e-01 -1.41597438e+00 1.16240442e+00
5.17565846e-01 8.10299814e-03 4.23522629e-02 1.21007848e+00
1.07873118e+00 5.12213528e-01 3.81069095e-03 8.57181728e-01
1.02644753e+00 -9.91057277e-01 -3.75500843e-02 -9.43844318e-01
3.36603314e-01 -4.83911604e-01 6.08117163e-01 7.75153562e-02
-9.96861398e-01 -5.59138179e-01 -9.50326562e-01 -6.72149658e-01
-3.27661067e-01 1.26661837e-01 7.41589785e-01 2.16093987e-01
-1.11932087e+00 2.64669210e-01 -8.77241075e-01 -2.70897239e-01
7.60115623e-01 4.20880795e-01 -4.42644596e-01 -1.62274495e-01
-6.21872604e-01 1.28490484e+00 6.13629222e-01 4.81172428e-02
-1.16148496e+00 -7.28630126e-01 -1.13231492e+00 -4.02903229e-01
-2.43097730e-02 -7.72679985e-01 1.26491678e+00 -1.29716381e-01
-1.22696555e+00 1.34491837e+00 6.49515614e-02 -7.98454762e-01
5.48323691e-01 -2.33264983e-01 4.40499097e-01 -2.71873437e-02
3.39452177e-01 1.78077435e+00 6.22453988e-01 -1.35904682e+00
-1.00802886e+00 -6.45720065e-01 1.10805422e-01 4.03940737e-01
6.46327913e-01 -5.24905801e-01 -4.79337364e-01 -1.70417428e-01
5.81835389e-01 -1.02735710e+00 -5.85859001e-01 3.68987739e-01
-5.31197011e-01 -5.05147159e-01 8.00259292e-01 -2.16620609e-01
1.85214598e-02 -1.99159062e+00 1.44530162e-01 4.06343602e-02
2.37332463e-01 -2.52143353e-01 5.83702028e-02 -1.92286998e-01
4.82117176e-01 -3.52142602e-02 -3.00804764e-01 -5.89009106e-01
4.70279992e-01 3.92729312e-01 -5.21034122e-01 7.49250233e-01
6.57538593e-01 1.06823003e+00 -7.13370562e-01 -4.13045973e-01
5.39246142e-01 5.17505229e-01 -6.79449916e-01 2.33022273e-01
-5.83762288e-01 1.64540529e-01 -5.29257536e-01 8.99844885e-01
8.90164793e-01 1.53775290e-01 -7.54635930e-01 -3.33468430e-02
-2.67643392e-01 4.97241378e-01 -6.65515125e-01 1.95458567e+00
2.64039868e-03 6.96552992e-01 -1.23434603e-01 -4.21032429e-01
1.27039099e+00 -5.33999085e-01 8.39021131e-02 -5.88100016e-01
1.23303637e-01 2.61255980e-01 -4.00303155e-01 3.68207209e-02
8.97043645e-01 -2.52635889e-02 -3.51981759e-01 -1.33396015e-01
-5.37251495e-02 -1.11562920e+00 -3.78105581e-01 2.20558763e-01
1.11903071e+00 4.57137644e-01 1.07052185e-01 -2.23440558e-01
-1.47704959e-01 6.93532884e-01 3.13069791e-01 7.82141626e-01
-4.12412643e-01 9.85815227e-01 2.25502715e-01 -7.89933920e-01
-1.51176381e+00 -1.31956446e+00 -2.07248181e-01 5.55108786e-01
7.62214303e-01 -8.25522318e-02 -1.54706046e-01 -6.34090066e-01
6.98164880e-01 6.40451968e-01 -9.27126348e-01 -1.24036804e-01
-5.45521140e-01 -1.58900827e-01 2.59967178e-01 7.31032491e-01
3.15382212e-01 -8.96834970e-01 -1.11808097e+00 9.19758230e-02
-1.23896729e-02 -1.24540639e+00 -1.14672936e-01 4.71085995e-01
-1.08285844e+00 -8.61941397e-01 -3.51438940e-01 -3.40905249e-01
6.48725331e-01 5.81879616e-01 1.24425197e+00 -2.86827922e-01
-6.46840990e-01 4.24164653e-01 -2.63857096e-01 -9.28092003e-01
8.93163010e-02 1.62002385e-01 -6.16460778e-02 -6.08069241e-01
5.57155788e-01 -3.18152867e-02 -6.44211650e-01 3.46506834e-01
-4.83410656e-01 9.32591781e-02 6.43807173e-01 3.44914079e-01
1.03683925e+00 -6.58005238e-01 -3.36127102e-01 -9.74538773e-02
-2.92486280e-01 -4.37796026e-01 -1.12225652e+00 -4.10278618e-01
-1.62797496e-01 6.04825802e-02 -2.72439957e-01 3.96838710e-02
-5.79116642e-01 5.96542001e-01 -3.74076329e-02 -7.57831693e-01
-3.60295564e-01 -2.28473902e-01 2.80958079e-02 -4.43168640e-01
7.48865962e-01 -2.35304218e-02 -3.70647274e-02 -3.01445723e-01
5.00396848e-01 2.97041148e-01 5.95419347e-01 -4.89106029e-01
1.01319444e+00 1.13177907e+00 2.30480447e-01 -4.60914016e-01
-1.17392421e+00 -7.21725583e-01 -8.38862956e-01 -2.52519965e-01
1.01603472e+00 -1.28397512e+00 -6.14330113e-01 1.05321966e-01
-1.29726517e+00 -5.05507648e-01 -3.34016383e-01 4.39718157e-01
-9.46132600e-01 -1.18259087e-01 -1.57273561e-02 -7.66485572e-01
-1.19638734e-01 -1.26564682e+00 1.99932122e+00 1.38423115e-01
3.74533348e-02 -5.12306988e-01 7.92917386e-02 4.04957920e-01
2.04000235e-01 4.71961170e-01 2.23536715e-01 -1.19466133e-01
-1.40271878e+00 -5.04463851e-01 -5.62348366e-01 -1.12079091e-01
-5.26050985e-01 -1.00144193e-01 -1.06296635e+00 -9.90813151e-02
-3.35056663e-01 -6.04606748e-01 1.37324739e+00 5.56807518e-01
1.08919704e+00 9.91031826e-02 -5.85264444e-01 1.03712726e+00
1.42857862e+00 -1.90833002e-01 4.40735072e-01 6.13417149e-01
6.17673218e-01 5.42276680e-01 1.00955570e+00 4.95487094e-01
7.46118367e-01 7.68269122e-01 1.40202522e+00 -3.40235867e-02
9.12258178e-02 -3.90012503e-01 1.19241796e-01 -4.05086964e-01
1.38123497e-01 -1.82255864e-01 -1.21041048e+00 7.61899412e-01
-1.85091603e+00 -5.62801659e-01 -2.24911422e-01 1.86855865e+00
1.69226959e-01 5.86320102e-01 1.48907602e-01 -2.25722536e-01
3.28039557e-01 5.22711687e-02 -9.14666772e-01 -1.45669758e-01
-2.40245298e-01 3.26040499e-02 1.20935643e+00 5.54870248e-01
-1.35871696e+00 1.11401320e+00 6.77826023e+00 3.22656423e-01
-8.50709081e-01 1.32041484e-01 5.36905408e-01 -4.53260839e-01
-2.27620915e-01 8.87374058e-02 -1.55246532e+00 2.20925286e-02
5.78784108e-01 1.85988560e-01 -6.82510361e-02 1.32859945e+00
2.08647735e-02 -3.13299865e-01 -1.08119690e+00 9.77481544e-01
1.35562375e-01 -1.51680136e+00 -5.98695092e-02 4.50856537e-01
7.19487369e-01 1.04200053e+00 1.90492496e-01 2.59216279e-01
7.19743013e-01 -1.03153741e+00 1.67261302e+00 3.86035860e-01
5.65380752e-01 -6.63560331e-01 7.59184957e-01 1.67531148e-01
-9.42234933e-01 -1.21026359e-01 -8.41632366e-01 -1.26753360e-01
4.60448898e-02 3.86674553e-01 -1.20718479e+00 4.04921509e-02
1.07720435e+00 7.60190904e-01 -5.56794822e-01 1.44200647e+00
-1.34191215e-01 2.40133077e-01 -7.71855533e-01 -8.46524760e-02
8.16422939e-01 9.69774947e-02 7.36183524e-01 8.09152961e-01
2.46496081e-01 2.01985717e-01 3.39185029e-01 1.34202909e+00
1.28438205e-01 -3.21299314e-01 -9.11680281e-01 5.57175159e-01
2.80007809e-01 1.13816369e+00 -7.95243979e-01 -1.49825752e-01
2.12134942e-02 5.90533912e-01 5.31119883e-01 2.47244220e-02
-8.13480675e-01 -8.61468911e-02 1.11446857e+00 3.31005275e-01
6.67699695e-01 -5.69201827e-01 -5.11822462e-01 -7.42785037e-01
9.82749388e-02 8.77101049e-02 -1.26374394e-01 -1.19187272e+00
-1.26569796e+00 4.50329572e-01 2.71072119e-01 -1.21848500e+00
1.41117781e-01 -9.70167041e-01 -4.03932810e-01 9.14443552e-01
-1.77839077e+00 -1.28407705e+00 -6.81310356e-01 1.28143549e-01
5.08792996e-01 4.87139411e-02 1.45135164e-01 -2.68540949e-01
-4.11363170e-02 -5.91508523e-02 -2.38079339e-01 -2.22878620e-01
3.91347617e-01 -1.34732628e+00 1.02369809e+00 4.06180292e-01
-4.08431664e-02 8.99825171e-02 7.25947499e-01 -8.06650937e-01
-1.27466083e+00 -1.44828808e+00 5.44108748e-01 -1.17731035e+00
4.07234132e-01 -7.03571498e-01 -4.47167039e-01 6.88882947e-01
-3.34347308e-01 5.09583116e-01 -9.27450880e-02 -3.35151076e-01
-4.13320303e-01 -3.64472941e-02 -1.21613681e+00 7.99815282e-02
1.21261883e+00 -2.39109069e-01 -5.55579603e-01 5.25164366e-01
8.55779231e-01 -1.04305696e+00 -4.02515471e-01 4.86675173e-01
2.61476219e-01 -9.09544468e-01 1.30602515e+00 -5.59449553e-01
4.35125440e-01 -5.23538888e-01 -5.32128394e-01 -9.61703777e-01
-2.98013002e-01 -1.93813205e-01 7.43378252e-02 3.77479017e-01
3.92353415e-01 -2.33882636e-01 1.30281365e+00 6.21713698e-01
-7.01775312e-01 -4.63498443e-01 -1.37769067e+00 -7.37233996e-01
9.86133981e-03 -7.14775026e-01 6.24277055e-01 1.45118743e-01
-6.00482762e-01 -3.46211456e-02 7.51423910e-02 4.94237423e-01
1.14472413e+00 -7.21406490e-02 1.07599127e+00 -1.41288829e+00
4.45180565e-01 -6.10294998e-01 -9.82528985e-01 -1.14850187e+00
2.81733125e-01 -9.59119380e-01 7.74686515e-01 -1.77861583e+00
-1.06842905e-01 -8.05106163e-01 1.70073286e-01 4.66035575e-01
4.09508608e-02 5.69054127e-01 2.39627319e-03 1.94216430e-01
-8.62354934e-01 8.57958615e-01 9.70848143e-01 -3.22719395e-01
1.10990614e-01 -9.62034240e-02 -1.64631993e-01 6.73867106e-01
6.30362034e-01 -5.26012361e-01 -9.89621952e-02 -7.69147456e-01
2.62041450e-01 -2.60141045e-01 1.02983999e+00 -1.27248085e+00
2.55033895e-02 -2.27090150e-01 5.09906888e-01 -1.53206229e+00
1.00697875e+00 -7.43597865e-01 -2.78610289e-01 3.30383182e-01
-1.42261550e-01 -3.02967466e-02 4.45544839e-01 7.48979986e-01
9.01350901e-02 4.26912010e-02 7.95739353e-01 -3.39272410e-01
-9.31668162e-01 6.57040000e-01 -5.35004400e-02 -9.83455405e-03
1.07981586e+00 -5.65632284e-01 -2.88206041e-01 3.56931388e-02
-3.77296150e-01 4.22150970e-01 8.43878448e-01 6.51062369e-01
1.05184519e+00 -1.32034135e+00 -6.79631948e-01 2.34121785e-01
6.60006583e-01 8.24840844e-01 -1.62730306e-01 4.04360563e-01
-9.37941372e-01 8.59171987e-01 -3.46878201e-01 -1.16623914e+00
-6.96715474e-01 1.46954969e-01 5.65978646e-01 4.03898805e-01
-8.18673491e-01 1.51333880e+00 3.02839875e-01 -8.20312023e-01
3.77816647e-01 -8.41679990e-01 2.67104387e-01 -4.71124679e-01
1.27743155e-01 3.20324861e-02 1.23579755e-01 -8.84101927e-01
-4.98995066e-01 6.10273242e-01 9.07028839e-02 -1.84704155e-01
1.29133236e+00 -8.59623998e-02 1.80321857e-01 2.79248983e-01
9.57731068e-01 -7.16013372e-01 -1.94834220e+00 -1.63147360e-01
-5.74429743e-02 -7.87713766e-01 3.62804860e-01 -5.38453400e-01
-6.38497829e-01 9.24445450e-01 7.75543392e-01 -4.46742117e-01
1.19067557e-01 5.60438514e-01 5.47814488e-01 5.26574969e-01
7.64774382e-01 -8.15523207e-01 1.96023092e-01 6.83058858e-01
1.00826299e+00 -1.84930885e+00 1.39284581e-01 -2.41702348e-01
-5.06315291e-01 9.15546477e-01 9.77133036e-01 -5.68717778e-01
5.33374608e-01 1.56157807e-01 -7.63614029e-02 -6.14259541e-01
-6.86435521e-01 -4.31324899e-01 5.28548598e-01 7.64710009e-01
-2.88008928e-01 1.30698636e-01 5.07062793e-01 1.98550224e-01
-5.21778643e-01 -3.99758577e-01 3.55204433e-01 9.67551708e-01
-9.77015316e-01 -3.94846469e-01 -6.56896830e-01 5.15475035e-01
1.87344253e-01 2.42084563e-01 -6.31690800e-01 8.72599363e-01
3.29797506e-01 4.71229374e-01 6.03626966e-01 -3.05243790e-01
4.99496937e-01 -3.78705710e-01 5.23315251e-01 -8.98887873e-01
-1.76531971e-01 -2.65821606e-01 -2.02135935e-01 -8.24780762e-01
-1.99403718e-01 -7.21173227e-01 -1.26296628e+00 -4.60259207e-02
-3.45398128e-01 -2.93680131e-01 1.28419030e+00 8.64671409e-01
2.36712247e-01 1.69389695e-02 3.07392180e-01 -1.75503278e+00
-2.60063678e-01 -8.12472761e-01 -2.94228405e-01 5.50337397e-02
4.15787101e-01 -1.06262302e+00 -3.87001902e-01 -3.27537954e-01] | [7.775477886199951, -2.6327199935913086] |
71039fed-706e-479e-8b6a-fb7a866495e0 | beyond-deep-residual-learning-for-image | 1611.06345 | null | http://arxiv.org/abs/1611.06345v4 | http://arxiv.org/pdf/1611.06345v4.pdf | Beyond Deep Residual Learning for Image Restoration: Persistent Homology-Guided Manifold Simplification | The latest deep learning approaches perform better than the state-of-the-art
signal processing approaches in various image restoration tasks. However, if an
image contains many patterns and structures, the performance of these CNNs is
still inferior. To address this issue, here we propose a novel feature space
deep residual learning algorithm that outperforms the existing residual
learning. The main idea is originated from the observation that the performance
of a learning algorithm can be improved if the input and/or label manifolds can
be made topologically simpler by an analytic mapping to a feature space. Our
extensive numerical studies using denoising experiments and NTIRE single-image
super-resolution (SISR) competition demonstrate that the proposed feature space
residual learning outperforms the existing state-of-the-art approaches.
Moreover, our algorithm was ranked third in NTIRE competition with 5-10 times
faster computational time compared to the top ranked teams. The source code is
available on page : https://github.com/iorism/CNN.git | ['Woong Bae', 'Jong Chul Ye', 'Jaejun Yoo'] | 2016-11-19 | null | null | null | null | ['color-image-denoising'] | ['computer-vision'] | [-8.53460953e-02 -2.21216559e-01 2.02053964e-01 -5.38537279e-02
-9.15789127e-01 -1.55918866e-01 5.58755815e-01 -4.41816002e-01
-1.40856847e-01 6.83192015e-01 5.40306509e-01 6.12962507e-02
-3.40048134e-01 -5.14886796e-01 -8.21115196e-01 -6.82467282e-01
-1.49174288e-01 2.91743316e-02 1.89963743e-01 -3.46901655e-01
3.80709589e-01 5.18446028e-01 -1.37027895e+00 3.39281291e-01
8.84322286e-01 9.29707289e-01 1.34493515e-01 4.87455577e-01
1.49059847e-01 7.92884707e-01 -9.76383835e-02 1.00099988e-01
5.11279285e-01 -2.76236594e-01 -8.01820636e-01 -1.77968726e-01
7.65483677e-01 -5.17399758e-02 -8.90929580e-01 1.28607857e+00
5.69245398e-01 2.12806553e-01 4.00690734e-01 -8.25344682e-01
-1.00255334e+00 4.00745511e-01 -6.53779089e-01 5.48920035e-01
9.43867192e-02 -2.32527956e-01 7.78757811e-01 -1.36033797e+00
7.66089261e-01 1.02180016e+00 9.04154480e-01 3.75534236e-01
-1.26260924e+00 -7.06380248e-01 -7.46928202e-03 4.62094456e-01
-1.34390366e+00 -4.33096647e-01 1.10671544e+00 -5.27049243e-01
6.17570877e-01 -1.55412499e-02 2.57825196e-01 9.33639169e-01
4.32911783e-01 6.35313630e-01 1.30480957e+00 -7.88850933e-02
5.44772185e-02 -2.68345743e-01 1.67697385e-01 9.62411106e-01
8.27479586e-02 2.23218352e-01 -6.18247747e-01 1.10861823e-01
1.10833240e+00 -9.96680651e-03 -5.59869885e-01 -3.05723399e-01
-1.35177207e+00 8.15018356e-01 7.41930842e-01 6.41023040e-01
-4.14150000e-01 1.19275302e-01 3.90695065e-01 6.20683014e-01
8.54227185e-01 3.85034621e-01 -3.36409599e-01 1.60844907e-01
-1.01315939e+00 3.38918865e-02 3.33361447e-01 3.36978972e-01
7.77374506e-01 3.54131013e-01 1.76955625e-01 8.96191716e-01
-5.17242216e-02 -2.56186742e-02 4.65448111e-01 -1.37174976e+00
1.00733615e-01 3.76337200e-01 5.39600849e-02 -1.33065271e+00
-5.49540460e-01 -8.91715467e-01 -1.51058316e+00 4.70296562e-01
3.86971772e-01 -1.47687283e-03 -8.87184501e-01 1.44908643e+00
1.18850552e-01 7.26441681e-01 -4.86037619e-02 1.17414939e+00
1.04643810e+00 7.45413601e-01 -5.47767997e-01 -2.02783346e-01
1.05859184e+00 -1.03807247e+00 -6.37965679e-01 -6.49475232e-02
2.12110639e-01 -9.86108422e-01 8.30590308e-01 5.94251633e-01
-1.06105971e+00 -9.06953990e-01 -1.07913113e+00 -1.85399279e-01
-1.81238111e-02 2.51057267e-01 6.70019329e-01 1.10467993e-01
-1.37746251e+00 1.23819935e+00 -8.35245967e-01 -2.44481638e-01
6.41377807e-01 2.37548217e-01 -7.26682067e-01 -1.32246196e-01
-9.81994033e-01 6.91130280e-01 -2.08761826e-01 1.89581409e-01
-8.32552671e-01 -9.88137603e-01 -5.80897689e-01 -3.15737903e-01
1.65673882e-01 -6.51201665e-01 7.91307390e-01 -8.86171222e-01
-1.41811883e+00 8.18106651e-01 -9.10897255e-02 -4.84425068e-01
5.34177363e-01 -3.31196934e-01 -3.06510001e-01 4.25548941e-01
3.65019739e-02 2.63640761e-01 1.30591905e+00 -1.27865410e+00
-4.17992353e-01 -2.53171593e-01 -1.51884139e-01 -6.68548048e-02
-7.47137144e-02 -3.61784883e-02 3.79086994e-02 -7.98488259e-01
5.15070021e-01 -7.31711030e-01 -3.87784451e-01 2.47660261e-02
-1.29976556e-01 -1.56272039e-01 9.90578651e-01 -9.11146939e-01
6.00956440e-01 -2.30053949e+00 5.40231109e-01 -2.45165646e-01
4.39259261e-01 2.79305458e-01 -2.21389458e-01 1.88118055e-01
-6.03835106e-01 1.32945135e-01 -1.62568510e-01 -3.31518710e-01
-2.96264201e-01 -2.69457519e-01 -4.37215954e-01 1.04117990e+00
-4.17221896e-02 7.69668043e-01 -7.10678041e-01 -2.70224631e-01
4.81203020e-01 1.01560211e+00 -2.32968509e-01 -1.69425756e-02
3.57516825e-01 9.29463983e-01 -3.48966211e-01 4.22811180e-01
9.19487715e-01 -2.87652314e-01 -2.68701404e-01 -7.16890812e-01
-2.35184759e-01 1.11127734e-01 -1.31856990e+00 1.74534333e+00
-3.35447282e-01 9.65227306e-01 3.06728423e-01 -1.49647486e+00
8.17527115e-01 2.30636567e-01 8.18358421e-01 -7.88487673e-01
3.94132175e-03 3.43729377e-01 -2.75014430e-01 -2.34719679e-01
3.32827955e-01 -1.57658920e-01 4.00973797e-01 1.51062638e-01
7.45311305e-02 1.50126874e-01 -1.44295007e-01 2.11987838e-01
1.32966185e+00 1.74792632e-01 -1.01911999e-01 -5.45898736e-01
6.71927392e-01 -1.86685383e-01 7.81077743e-01 5.20338237e-01
-2.59820938e-01 8.65979135e-01 2.07422242e-01 -8.31978083e-01
-1.04944694e+00 -1.10165977e+00 -3.21553975e-01 7.43754923e-01
9.32451636e-02 -2.17058584e-01 -6.73309565e-01 -3.41036379e-01
-3.45045507e-01 -6.23383885e-03 -6.48421407e-01 4.70927544e-02
-8.84081602e-01 -6.40987515e-01 3.02021593e-01 3.54138672e-01
8.72865081e-01 -1.05657136e+00 -2.34115049e-01 1.21935330e-01
-4.68213648e-01 -1.08836460e+00 -4.66323793e-01 -1.55841276e-01
-1.18239808e+00 -1.02558219e+00 -9.11319792e-01 -1.13133514e+00
6.11028433e-01 4.40879852e-01 9.77142036e-01 2.33364016e-01
-4.25448686e-01 2.60446608e-01 -3.63438308e-01 3.87477398e-01
-1.00506909e-01 -9.28987488e-02 2.69941360e-01 1.24887913e-01
-2.17020065e-01 -9.68844295e-01 -9.45298731e-01 2.01679885e-01
-7.65057683e-01 1.30705342e-01 4.52453434e-01 9.19254065e-01
7.35767841e-01 6.19755447e-01 7.96193063e-01 -5.08422732e-01
4.37127680e-01 -3.39879185e-01 -3.77281606e-01 -1.40432760e-01
-4.28814560e-01 1.49536639e-01 7.50527799e-01 -3.16849828e-01
-9.46197510e-01 1.36795491e-01 -7.65941814e-02 -4.90845174e-01
-8.33640769e-02 3.57554585e-01 3.11474681e-01 -5.12052238e-01
6.82764888e-01 1.99532002e-01 -6.34194864e-03 -8.55003834e-01
2.38611355e-01 3.23544830e-01 7.17311680e-01 -3.19698602e-01
9.85673547e-01 8.46920729e-01 2.17125162e-01 -8.91490221e-01
-8.91973257e-01 -4.61362541e-01 -7.12213755e-01 -3.04291457e-01
7.63712227e-01 -9.61110473e-01 -3.36457968e-01 7.28035033e-01
-9.91978407e-01 -2.92830497e-01 -2.33539075e-01 4.03031588e-01
-6.38240397e-01 4.87651736e-01 -9.65208948e-01 -2.29196638e-01
-4.11390662e-01 -1.18430805e+00 8.64274025e-01 1.31756335e-01
3.44451785e-01 -1.10082805e+00 8.23611021e-02 5.26732206e-01
6.47050798e-01 4.31630880e-01 6.61239982e-01 -1.01811029e-01
-6.16702318e-01 -1.52188055e-02 -4.58393961e-01 6.11348093e-01
1.40999079e-01 -3.41891825e-01 -7.37283826e-01 -5.89259326e-01
3.23883504e-01 -1.55819714e-01 1.40045345e+00 7.59415925e-01
1.14357698e+00 -5.72761409e-02 -1.29280031e-01 9.94188547e-01
1.66393209e+00 -4.08430755e-01 9.19681907e-01 4.48403478e-01
7.51764715e-01 4.74674314e-01 2.92909920e-01 5.68603575e-02
1.23634674e-01 6.17487848e-01 4.43979055e-01 -2.77983010e-01
-6.10187888e-01 -2.82931253e-02 4.37993765e-01 1.00791228e+00
-3.72192889e-01 4.51207429e-01 -6.47757947e-01 6.55597448e-01
-1.89189076e+00 -1.03186011e+00 -3.68779242e-01 1.98267686e+00
5.95551193e-01 -1.23377211e-01 -8.54792148e-02 2.71367699e-01
7.51869559e-01 3.37699175e-01 -3.78437102e-01 1.20101072e-01
-3.14301282e-01 4.32964653e-01 3.23314846e-01 6.41170979e-01
-1.31091309e+00 9.60791111e-01 5.54665470e+00 8.84872675e-01
-1.28040373e+00 3.54451030e-01 6.50228977e-01 9.62153897e-02
1.19037561e-01 -1.40291408e-01 -2.63998985e-01 9.05708000e-02
7.31937885e-01 -1.63409457e-01 5.96231103e-01 4.49494213e-01
2.29811773e-01 2.33036578e-01 -8.44808161e-01 1.28353870e+00
1.06072098e-01 -1.84879541e+00 -1.64965823e-01 5.92445508e-02
8.72887731e-01 4.95753229e-01 2.12361351e-01 1.01551466e-01
-4.24409918e-02 -1.16873562e+00 4.01703477e-01 8.06008220e-01
6.40596092e-01 -7.48418272e-01 7.48253286e-01 7.23050069e-03
-1.30793488e+00 -6.97995499e-02 -5.78643680e-01 -1.07821032e-01
1.80803731e-01 8.15518439e-01 1.30226299e-01 8.13441157e-01
1.20156407e+00 1.29173827e+00 -6.85314655e-01 1.18975997e+00
2.69897617e-02 5.67010581e-01 -1.03629418e-02 6.70226395e-01
9.05986801e-02 -3.34509134e-01 8.98069978e-01 1.08766401e+00
3.97657216e-01 1.92306265e-01 1.54831782e-01 7.16833472e-01
-2.95524597e-01 9.75279808e-02 -5.51806688e-01 3.38718772e-01
-5.55069707e-02 1.62869287e+00 -7.61614144e-01 -6.03175461e-02
-3.92968178e-01 1.09055007e+00 3.04800570e-01 5.65594852e-01
-5.57803929e-01 -3.08549285e-01 6.43856525e-01 3.94773602e-01
3.50711256e-01 -3.96709681e-01 -3.64643872e-01 -1.39020872e+00
1.37378991e-01 -8.46752167e-01 1.25561103e-01 -8.69742930e-01
-1.40353322e+00 8.47941756e-01 -4.18734252e-01 -1.38884711e+00
2.90602416e-01 -7.02276468e-01 -5.73330045e-01 5.76780796e-01
-1.63471282e+00 -1.00880647e+00 -3.37965906e-01 8.09385598e-01
5.82132578e-01 -4.42777038e-01 5.98922789e-01 4.29929286e-01
-3.61426592e-01 1.42013520e-01 3.04011226e-01 2.69484997e-01
6.47798419e-01 -1.07311356e+00 2.03831896e-01 1.10719359e+00
6.00154400e-02 4.65613097e-01 7.55160868e-01 -5.26216745e-01
-1.47808778e+00 -1.16292429e+00 4.27234024e-01 -1.90425292e-01
9.44845557e-01 -9.07131881e-02 -1.15273762e+00 6.61451042e-01
4.48282450e-01 4.91296262e-01 1.80264473e-01 -9.14871246e-02
-4.76624906e-01 -1.54004723e-01 -1.09803128e+00 4.35314655e-01
1.15056741e+00 -5.54570019e-01 -5.29965222e-01 4.53989536e-01
6.56910896e-01 -3.26897234e-01 -9.89869952e-01 5.37756741e-01
2.58001596e-01 -1.03736579e+00 1.30331194e+00 -3.69144261e-01
5.65284848e-01 -5.59421897e-01 -3.51020932e-01 -1.47853410e+00
-8.24625731e-01 -7.58518517e-01 -1.80522770e-01 8.33093584e-01
4.73383181e-02 -7.16831684e-01 5.61918855e-01 -1.97225481e-01
-3.22539240e-01 -7.16978788e-01 -1.31393194e+00 -8.99532259e-01
4.18270379e-01 -2.83982247e-01 7.91452527e-02 1.03489196e+00
-3.19173664e-01 2.82666028e-01 -6.11464143e-01 3.75246853e-01
1.15230775e+00 1.61958978e-01 2.75674760e-01 -1.48882377e+00
-6.18636236e-02 -4.03600603e-01 -5.60547650e-01 -8.46497893e-01
3.23068768e-01 -1.18399847e+00 -1.86161786e-01 -1.81322455e+00
1.60338446e-01 -2.22958520e-01 -5.58516860e-01 3.59689504e-01
2.26314381e-01 7.19965100e-01 1.63318694e-01 3.49079370e-01
-5.99137187e-01 5.31239867e-01 1.35497797e+00 -1.25498578e-01
-1.64920479e-01 -2.90180326e-01 -7.84242332e-01 8.26597333e-01
1.02261913e+00 -2.49624342e-01 -4.13578324e-04 -5.29585004e-01
2.40335897e-01 -1.32144308e-02 8.27918053e-01 -1.19870341e+00
3.91347349e-01 2.23366082e-01 4.30619627e-01 -4.17012960e-01
2.04477519e-01 -5.63575208e-01 1.84119657e-01 4.24858958e-01
-1.83822721e-01 -1.27422929e-01 1.61248475e-01 3.95635426e-01
-2.73495078e-01 1.50318131e-01 1.24314928e+00 -1.99436575e-01
-6.85483158e-01 3.37892771e-01 -2.75869638e-01 1.15500726e-01
4.47351307e-01 -8.01365674e-02 -4.85615045e-01 -4.17378187e-01
-8.37945700e-01 -2.44783983e-01 3.18930715e-01 5.40090978e-01
1.01081872e+00 -1.47910249e+00 -1.08983171e+00 1.85238436e-01
-4.41075593e-01 -1.63704321e-01 5.83600760e-01 1.16713738e+00
-5.45420945e-01 1.41074732e-01 -3.84643048e-01 -6.60641432e-01
-1.09919715e+00 3.64990503e-01 7.15438128e-01 -1.47344694e-01
-1.43149602e+00 7.57471502e-01 1.48004398e-01 -5.38417339e-01
1.75654158e-01 -6.33072630e-02 -1.93383738e-01 -3.03870708e-01
7.69961834e-01 7.51777112e-01 8.55562687e-02 -8.17906976e-01
-3.61306489e-01 1.11382914e+00 -1.53383002e-01 9.42289010e-02
1.88354027e+00 -1.26248926e-01 -3.75225455e-01 4.65666614e-02
1.34094000e+00 -3.86331193e-02 -1.36428607e+00 -4.77051646e-01
-4.52876687e-02 -4.56855446e-01 6.58670247e-01 -5.45023501e-01
-1.59719419e+00 7.01718867e-01 9.36921060e-01 2.20107567e-02
1.41373944e+00 6.89298660e-02 9.84570563e-01 1.19088821e-01
4.44196165e-01 -1.07282317e+00 4.82796341e-01 4.53437626e-01
1.40982664e+00 -1.26217973e+00 8.41443017e-02 -4.30283099e-01
-3.22963804e-01 1.09609926e+00 3.31231773e-01 -8.30381691e-01
1.14982009e+00 1.98761463e-01 -3.99985211e-03 -3.35291207e-01
-4.87868071e-01 3.64321284e-02 4.36127156e-01 5.52769125e-01
5.75433195e-01 -9.06537771e-02 -1.94134787e-01 4.13429201e-01
-2.21303314e-01 1.01483114e-01 7.15526581e-01 3.52689147e-01
-4.35464293e-01 -9.24421668e-01 -2.42873296e-01 3.14966500e-01
-7.16566861e-01 -1.23192571e-01 4.27916870e-02 5.25665343e-01
-1.85398739e-02 1.02408659e+00 -9.92480963e-02 -4.70318794e-01
2.81767011e-01 -4.11929429e-01 5.44883907e-01 -2.96709329e-01
-2.93425381e-01 1.28611192e-01 -2.13942051e-01 -1.04055083e+00
-5.75219512e-01 -6.94299400e-01 -1.36414003e+00 -3.22440356e-01
1.28284752e-01 3.20823006e-02 4.70327973e-01 8.01443875e-01
4.28340495e-01 6.41599774e-01 6.69794142e-01 -1.19631171e+00
-1.44249484e-01 -8.66590381e-01 -5.97044885e-01 3.35603446e-01
6.96078479e-01 -8.56794417e-01 -7.59977996e-01 1.06030554e-01] | [11.165032386779785, -2.2095212936401367] |
69f62ef7-dfaf-49da-ac10-af221a56493d | counterfactual-supervision-based-information | 2208.07798 | null | https://arxiv.org/abs/2208.07798v3 | https://arxiv.org/pdf/2208.07798v3.pdf | Counterfactual Supervision-based Information Bottleneck for Out-of-Distribution Generalization | Learning invariant (causal) features for out-of-distribution (OOD) generalization has attracted extensive attention recently, and among the proposals invariant risk minimization (IRM) is a notable solution. In spite of its theoretical promise for linear regression, the challenges of using IRM in linear classification problems remain. By introducing the information bottleneck (IB) principle into the learning of IRM, IB-IRM approach has demonstrated its power to solve these challenges. In this paper, we further improve IB-IRM from two aspects. First, we show that the key assumption of support overlap of invariant features used in IB-IRM is strong for the guarantee of OOD generalization and it is still possible to achieve the optimal solution without this assumption. Second, we illustrate two failure modes that IB-IRM (and IRM) could fail for learning the invariant features, and to address such failures, we propose a \textit{Counterfactual Supervision-based Information Bottleneck (CSIB)} learning algorithm that provably recovers the invariant features. By requiring counterfactual inference, CSIB works even when accessing data from a single environment. Empirical experiments on several datasets verify our theoretical results. | ['Kui Jia', 'Bin Deng'] | 2022-08-16 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 3.25886935e-01 4.45435584e-01 -6.13642752e-01 -3.54293376e-01
-1.00155902e+00 -2.84529507e-01 5.50758302e-01 -3.30756754e-02
-2.41629481e-01 9.94253874e-01 5.32746091e-02 -7.00560093e-01
-8.28859329e-01 -5.94888628e-01 -1.05891371e+00 -6.64875984e-01
-4.06760544e-01 2.23694444e-01 -6.19785301e-02 8.58810544e-02
3.42440963e-01 3.48013937e-01 -1.67527688e+00 1.83546126e-01
9.22068954e-01 1.02590168e+00 5.88682890e-02 2.59932756e-01
-1.19154975e-02 7.87476718e-01 -4.18763846e-01 -5.07318318e-01
7.31532931e-01 -2.05790654e-01 -9.88283396e-01 -1.33565545e-01
4.34296370e-01 -3.73067886e-01 -3.05820972e-01 9.96008277e-01
5.66279471e-01 1.25291109e-01 9.86330688e-01 -2.00147152e+00
-6.18744493e-01 6.97848380e-01 -6.91315353e-01 2.08158419e-01
1.20286405e-01 -3.12061667e-01 1.26295090e+00 -7.47550726e-01
3.48796934e-01 1.50582671e+00 5.58474183e-01 6.75984681e-01
-1.38358724e+00 -6.87316239e-01 3.22101653e-01 5.60340881e-02
-1.25044489e+00 -2.33261362e-01 5.38943946e-01 -4.33139920e-01
9.01071131e-01 4.15468514e-01 1.32226527e-01 1.07419121e+00
2.91335523e-01 1.29497373e+00 1.37349868e+00 -7.29405582e-01
3.94261122e-01 4.20576841e-01 4.58966076e-01 4.50720996e-01
5.76904655e-01 4.79040235e-01 -8.88676226e-01 -5.19513190e-01
4.69360679e-01 1.22656234e-01 -2.96219438e-01 -6.16076231e-01
-7.63401449e-01 1.09160674e+00 2.86221832e-01 -4.82943840e-02
1.92349609e-02 -4.58968617e-02 3.71926367e-01 8.46726239e-01
7.56573617e-01 3.35923880e-01 -7.40571082e-01 3.03105503e-01
-7.01398253e-01 4.16456968e-01 6.91717386e-01 1.06417394e+00
6.93089664e-01 -2.74794757e-01 -2.56593615e-01 5.14960647e-01
2.14934051e-01 4.99130636e-01 3.78317773e-01 -8.80131900e-01
6.71865940e-01 3.73937905e-01 2.57056713e-01 -8.68094087e-01
-4.67071593e-01 -3.52342278e-01 -8.97927403e-01 2.78154850e-01
4.23916936e-01 3.79899479e-02 -5.21049142e-01 1.86687768e+00
3.46829474e-01 1.02898732e-01 1.76072359e-01 6.34783804e-01
-4.81601171e-02 3.99450481e-01 -2.98225790e-01 -6.53175414e-01
7.51539111e-01 -4.62445021e-01 -5.82255185e-01 -5.07986881e-02
8.58635366e-01 -2.31121376e-01 1.04147506e+00 4.81772780e-01
-7.27592349e-01 -2.23110214e-01 -1.13162363e+00 2.81230062e-01
-2.42818087e-01 -7.53201544e-01 9.18902159e-01 7.78797269e-01
-8.96962702e-01 6.30669355e-01 -5.38875997e-01 -3.09877116e-02
5.12452722e-01 5.13268709e-01 -4.56460118e-01 -3.43081146e-01
-1.29940426e+00 7.66782284e-01 3.49923491e-01 -9.65550691e-02
-8.92400324e-01 -1.05488729e+00 -8.63698184e-01 -1.17641389e-01
6.38152361e-01 -5.33054292e-01 1.07439280e+00 -1.01844215e+00
-1.19092703e+00 5.83311737e-01 -1.39355734e-01 -7.04461992e-01
9.41306055e-01 -4.34658647e-01 -1.94389094e-02 -6.76733628e-02
2.84400254e-01 2.81248927e-01 1.08390188e+00 -1.21977246e+00
-7.39549220e-01 -6.62575841e-01 5.99319823e-02 -1.73203722e-02
-4.10065323e-01 2.24410165e-02 2.70355731e-01 -5.40580034e-01
3.92400511e-02 -8.23096573e-01 -1.75001979e-01 -2.21117258e-01
-5.42802811e-01 -6.40307188e-01 6.07497036e-01 -1.87409818e-01
1.11953127e+00 -2.28377867e+00 -1.08706780e-01 2.54454404e-01
1.05060069e-02 -1.18218005e-01 2.83918586e-02 3.18657100e-01
-1.82839870e-01 4.16735590e-01 -2.93843806e-01 -4.31693345e-01
3.44815403e-01 2.15891466e-01 -7.25705743e-01 6.87153518e-01
4.40462828e-02 7.12420702e-01 -7.17267513e-01 -4.95041430e-01
3.28524597e-02 2.56037917e-02 -6.92881942e-01 1.33501321e-01
-1.04384370e-01 2.27236643e-01 -4.68389124e-01 4.63508368e-01
1.03166115e+00 -1.79800332e-01 2.65622228e-01 3.50516707e-01
-5.03089540e-02 2.70700932e-01 -1.50736737e+00 1.29503763e+00
-4.71323222e-01 3.28685552e-01 -1.21285051e-01 -1.54844701e+00
6.23328030e-01 3.18997085e-01 4.64471489e-01 -6.02498353e-01
-2.30880424e-01 3.24633777e-01 -3.38770360e-01 -3.53655905e-01
2.85156146e-02 -4.66785967e-01 -3.11353087e-01 6.10006571e-01
6.39534369e-02 2.89857358e-01 -3.72129470e-01 1.23099782e-01
9.56135809e-01 -4.26230058e-02 4.83711332e-01 -6.10806346e-01
2.93169916e-01 -3.00129294e-01 8.35697412e-01 1.43133128e+00
-1.71190888e-01 2.92368501e-01 5.88548601e-01 -5.28222978e-01
-6.67726815e-01 -1.23980212e+00 -6.29460514e-01 1.06792462e+00
4.30605114e-02 -1.86746076e-01 -3.49940151e-01 -1.22606468e+00
3.86682928e-01 8.75981927e-01 -7.72972047e-01 -2.19541758e-01
-1.46542639e-01 -9.67498660e-01 2.12712944e-01 3.02979827e-01
4.00278181e-01 -6.41372323e-01 -4.28852797e-01 2.71047112e-02
-9.31712762e-02 -4.57037002e-01 -1.73291281e-01 5.18584728e-01
-7.50483334e-01 -1.01988709e+00 -4.60770845e-01 -2.39807203e-01
5.79530954e-01 5.35201788e-01 9.09572542e-01 -1.58262864e-01
-2.95848101e-01 4.27061856e-01 -2.68794715e-01 -7.65518367e-01
-3.06374997e-01 9.53933895e-02 4.29594308e-01 -3.33198644e-02
3.73748779e-01 -7.11441517e-01 -4.30723578e-01 3.81593883e-01
-9.13913846e-01 -1.56059623e-01 6.30587280e-01 9.84326005e-01
3.63297760e-01 3.69422317e-01 1.01700342e+00 -1.13592958e+00
3.59762639e-01 -6.50664389e-01 -7.88943470e-01 3.76602918e-01
-9.97333288e-01 3.59752446e-01 5.28301358e-01 -2.97258645e-01
-1.12341690e+00 -2.23702192e-01 3.81893545e-01 -4.15559173e-01
2.55910214e-02 4.05775011e-01 -4.23386633e-01 2.84772426e-01
6.31656885e-01 8.79792422e-02 4.47093211e-02 -6.40483141e-01
2.83804744e-01 9.62476611e-01 1.35438830e-01 -8.17070067e-01
8.18533957e-01 5.90500176e-01 2.83986628e-01 -6.04607582e-01
-1.53730333e+00 -3.65624934e-01 -7.21794069e-01 4.53025587e-02
5.07466257e-01 -7.88904250e-01 -8.89920235e-01 8.15177187e-02
-7.56613374e-01 -2.48629287e-01 -3.52413356e-01 5.98370194e-01
-8.29262972e-01 1.82459876e-01 -4.62957397e-02 -1.33886147e+00
2.40162686e-02 -7.58732498e-01 9.00049865e-01 -2.34574959e-01
9.01839361e-02 -9.66713548e-01 -1.01559713e-01 2.98490345e-01
1.45139709e-01 2.27284551e-01 1.01608431e+00 -7.95559406e-01
-5.56467712e-01 -1.98339984e-01 -1.09493412e-01 3.81386429e-01
1.01046465e-01 -5.47793925e-01 -1.15796995e+00 -6.41818643e-01
2.48423725e-01 -6.33346736e-01 1.06772447e+00 5.94954371e-01
1.23854470e+00 -6.93779767e-01 -3.70781064e-01 4.62865353e-01
1.48193860e+00 -1.96342409e-01 2.55587220e-01 5.94787598e-01
1.57271862e-01 7.93029904e-01 9.12091494e-01 8.10126424e-01
1.81834698e-01 6.70386076e-01 4.07973826e-01 2.09445313e-01
3.78078669e-01 -4.81734216e-01 6.84783638e-01 1.27361953e-01
2.86509037e-01 4.68353555e-02 -4.82727766e-01 4.04300004e-01
-2.08816552e+00 -1.16127610e+00 1.26328677e-01 2.74783039e+00
8.86741042e-01 2.39216477e-01 2.05593213e-01 2.69392401e-01
6.18267655e-01 -6.23570345e-02 -7.80210137e-01 -4.49794859e-01
-1.49749592e-01 -1.90287024e-01 6.12840235e-01 4.87969339e-01
-1.11950898e+00 4.30803269e-01 6.60627174e+00 9.07426834e-01
-7.24818826e-01 2.37433314e-01 6.61940634e-01 -4.33716923e-02
-4.16927993e-01 1.54688656e-01 -1.09744489e+00 2.95767695e-01
8.32506120e-01 -2.95412391e-01 4.72685516e-01 1.01032829e+00
5.46905585e-02 -1.49097323e-04 -1.43902445e+00 7.73548722e-01
-9.31153521e-02 -1.01744080e+00 1.80225208e-01 3.15017581e-01
7.89148271e-01 -1.94938749e-01 2.59256572e-01 5.20532489e-01
5.16192853e-01 -9.93849218e-01 7.63651133e-01 2.91368037e-01
7.41410553e-01 -1.01485252e+00 7.31410265e-01 8.48526955e-01
-7.00645030e-01 -4.54449773e-01 -8.51555884e-01 -3.35248530e-01
-3.43894035e-01 9.85092819e-01 -6.75724506e-01 1.01354349e+00
6.39686286e-01 6.36201322e-01 -3.77903134e-01 5.80494285e-01
-5.21680266e-02 6.10550702e-01 -3.66863340e-01 3.56619358e-01
-3.48996818e-02 2.00846434e-01 5.88488340e-01 9.04891253e-01
9.92909595e-02 -3.36564422e-01 3.10977221e-01 8.13933194e-01
2.93359533e-02 1.16755115e-02 -9.85084534e-01 4.69068974e-01
4.56849694e-01 6.82527125e-01 -2.04043239e-01 -6.49586022e-02
-5.32987714e-01 7.52747715e-01 5.11474729e-01 1.63671315e-01
-7.17594743e-01 -1.15951240e-01 5.91153622e-01 -8.80744308e-02
4.91268225e-02 2.80363351e-01 -6.45966604e-02 -1.31482792e+00
2.81413019e-01 -7.60934412e-01 9.50793564e-01 -3.15972678e-02
-1.75003564e+00 -3.75443622e-02 3.51636618e-01 -1.20887613e+00
-1.37501642e-01 -5.73255002e-01 -2.04998240e-01 6.69638753e-01
-1.75521517e+00 -1.06025088e+00 3.66089702e-01 7.52900422e-01
4.98210162e-01 2.37345658e-02 7.29793727e-01 -1.33215496e-03
-6.08851135e-01 9.48703885e-01 4.09727395e-01 -2.57768124e-01
7.80057490e-01 -1.61949193e+00 -1.11429922e-01 6.39928818e-01
1.26919046e-01 6.68245316e-01 6.78794920e-01 -5.15714467e-01
-1.41721690e+00 -1.12425005e+00 7.52584457e-01 -4.52856392e-01
4.99723256e-01 -5.36831737e-01 -6.58902764e-01 9.36932802e-01
-2.22591668e-01 6.73771501e-02 7.20831513e-01 4.77650821e-01
-6.74440742e-01 -3.16435784e-01 -1.09177303e+00 4.19667631e-01
1.21591687e+00 -2.63198912e-01 -8.25040877e-01 5.69431245e-01
1.00157809e+00 1.37549832e-01 -7.36126125e-01 3.78463924e-01
3.97891700e-01 -1.17370069e+00 9.10335839e-01 -1.02650177e+00
2.02099472e-01 -7.48596564e-02 -5.18455923e-01 -1.05476832e+00
-6.31759986e-02 -1.01442707e+00 -7.58003220e-02 1.38740480e+00
3.99531364e-01 -9.84309733e-01 3.19215953e-01 8.40663373e-01
7.13512152e-02 -6.10325813e-01 -1.35609782e+00 -1.44881582e+00
4.34534907e-01 -8.01230013e-01 4.25638378e-01 8.99791300e-01
2.84306854e-01 6.96240738e-02 -7.33613729e-01 3.54581922e-01
9.57715631e-01 3.86517137e-01 8.57250512e-01 -1.32104957e+00
-5.19735157e-01 -1.29254982e-01 -3.65911365e-01 -9.27477241e-01
6.92651689e-01 -1.26089728e+00 -1.63809761e-01 -9.80439007e-01
7.02480197e-01 -7.13945508e-01 -6.07174814e-01 4.95386243e-01
-3.25048536e-01 -3.53714377e-01 2.81380326e-01 2.74610788e-01
-7.11384416e-01 4.59632188e-01 8.30102146e-01 -3.06653343e-02
-2.10875899e-01 4.76886660e-01 -9.08863664e-01 6.56855285e-01
6.27396047e-01 -7.69602537e-01 -4.64915454e-01 5.42063117e-02
3.84633273e-01 6.82892697e-03 4.82821316e-01 -4.69640046e-01
-7.08081126e-02 -3.43695819e-01 2.38961279e-01 -5.12109041e-01
-8.79086107e-02 -8.45761240e-01 -2.18874559e-01 6.05834305e-01
-7.45078325e-01 -3.67830753e-01 -5.00135235e-02 1.07465816e+00
4.01755758e-02 -3.84497315e-01 6.99181557e-01 -6.33145720e-02
-2.45089218e-01 1.66871548e-01 -8.03594217e-02 3.98453504e-01
8.73095632e-01 2.66317159e-01 -3.59072208e-01 -3.08436900e-01
-2.92111218e-01 2.99449682e-01 1.77103192e-01 3.51573467e-01
4.95332927e-01 -1.12887430e+00 -8.27749014e-01 1.53874755e-01
2.03731284e-01 1.78791016e-01 1.52941689e-01 9.42038774e-01
4.44577843e-01 5.44561565e-01 2.85608411e-01 -4.75603461e-01
-8.52063239e-01 1.12000692e+00 2.18044728e-01 -5.34079075e-01
-7.19813824e-01 5.69052458e-01 5.63114285e-01 -4.35165942e-01
4.16068077e-01 1.14726618e-01 3.22372228e-01 -1.26448855e-01
7.74362504e-01 5.26722252e-01 1.11082075e-02 -8.24554882e-04
-3.26091111e-01 1.61266893e-01 -3.29609066e-01 -2.27497861e-01
1.54147601e+00 -3.76179039e-01 -3.27785909e-02 7.46893406e-01
1.28043485e+00 7.85251632e-02 -1.39073849e+00 -3.44374150e-01
5.25075614e-01 -6.47824943e-01 -6.71885163e-02 -5.45179784e-01
-5.24169028e-01 8.12540591e-01 3.84476155e-01 4.60214525e-01
1.06221628e+00 6.58773631e-02 2.17414245e-01 5.01565456e-01
7.10676551e-01 -8.89336467e-01 -2.87090130e-02 1.97883353e-01
8.19150567e-01 -1.41194546e+00 1.10297345e-01 -2.26420313e-01
-4.98412222e-01 9.28873241e-01 4.47505027e-01 1.74218994e-02
8.82051766e-01 7.53909051e-02 -3.30877036e-01 6.18020594e-02
-8.97885740e-01 -2.75625698e-02 1.90759480e-01 5.58492422e-01
9.09180492e-02 -4.70385924e-02 -2.82523096e-01 7.16051638e-01
-6.38764128e-02 -1.99538320e-01 3.34790856e-01 8.64446104e-01
-3.64833444e-01 -9.68190014e-01 -3.57585818e-01 5.64043403e-01
-7.29236543e-01 -5.13447486e-02 -2.61629969e-01 1.11757135e+00
-1.36598170e-01 1.01063681e+00 -1.29698679e-01 -1.94608510e-01
1.68162480e-01 2.89087474e-01 3.35590214e-01 -3.25477540e-01
-1.03981085e-02 -1.54091924e-01 -3.37746412e-01 -6.03318155e-01
-4.95157421e-01 -8.62144530e-01 -8.79152298e-01 -1.66129276e-01
-6.35822535e-01 2.40794316e-01 4.68349755e-01 1.03727090e+00
3.10880303e-01 2.56654155e-02 1.44399679e+00 -4.28695261e-01
-1.46074915e+00 -8.20730031e-01 -9.90346313e-01 3.99124712e-01
6.00879908e-01 -9.00592744e-01 -1.11132610e+00 -2.68006861e-01] | [8.68509292602539, 4.1954827308654785] |
50222f16-eb9d-4b19-ab7f-3c223e945008 | complex-word-identification-based-on | null | null | https://aclanthology.org/W18-0521 | https://aclanthology.org/W18-0521.pdf | Complex Word Identification Based on Frequency in a Learner Corpus | We introduce the TMU systems for the Complex Word Identification (CWI) Shared Task 2018. TMU systems use random forest classifiers and regressors whose features are the number of characters, the number of words, and the frequency of target words in various corpora. Our simple systems performed best on 5 tracks out of 12 tracks. Our ablation analysis revealed the usefulness of a learner corpus for CWI task. | ['Tomoyuki Kajiwara', 'Mamoru Komachi'] | 2018-06-01 | null | null | null | ws-2018-6 | ['complex-word-identification'] | ['natural-language-processing'] | [ 2.62460351e-01 -8.66836235e-02 -5.30178845e-01 -3.02138887e-02
-7.56861150e-01 -7.05517113e-01 9.92858350e-01 2.00236395e-01
-1.04970253e+00 5.84021628e-01 4.48823601e-01 -8.23201835e-01
2.30769455e-01 -4.69434738e-01 -5.10472357e-01 -1.72797307e-01
-6.47813827e-02 5.13149798e-01 1.72825679e-01 -1.50398076e-01
1.42669126e-01 -8.39083642e-02 -1.01850355e+00 2.30321839e-01
5.62377453e-01 8.46721113e-01 1.35748491e-01 7.75386512e-01
-4.77290690e-01 5.92651010e-01 -8.59345615e-01 -5.02100706e-01
6.76643569e-04 -1.06877603e-01 -8.02470446e-01 -6.98070645e-01
6.42786860e-01 -2.54101064e-02 -5.47769606e-01 7.80756533e-01
5.65145910e-01 1.91721711e-02 1.00371146e+00 -9.56367493e-01
-5.93480110e-01 1.38779342e+00 -3.39049488e-01 6.70534968e-01
4.23421711e-01 7.41299316e-02 1.56198776e+00 -1.12348151e+00
8.24847341e-01 1.27244461e+00 6.68606460e-01 6.15795016e-01
-1.15645742e+00 -1.29521275e+00 1.23131648e-01 3.44240725e-01
-1.34234107e+00 -6.05111301e-01 -4.89793122e-02 -5.77855647e-01
1.61442459e+00 7.41591305e-02 5.23747683e-01 1.56277323e+00
1.58019096e-01 1.13321745e+00 8.91979218e-01 -8.37635696e-01
-1.57886431e-01 -9.06828642e-02 1.16482925e+00 5.99225402e-01
5.72979927e-01 1.87954724e-01 -1.01210713e+00 -1.49880707e-01
1.01505958e-01 -5.64825833e-01 -1.96432412e-01 5.74089527e-01
-1.21443939e+00 1.02428579e+00 -1.59640431e-01 3.40903997e-01
3.59807819e-01 3.09242364e-02 5.32560050e-01 2.68881828e-01
6.92268968e-01 8.57476771e-01 -7.71941602e-01 -5.11919260e-01
-5.49499989e-01 1.18873477e-01 6.32398784e-01 1.07181704e+00
4.97140825e-01 1.49497064e-03 -5.36083579e-01 8.87261152e-01
1.13824241e-01 5.51940382e-01 9.35660779e-01 -5.90067245e-02
5.08997500e-01 3.96724135e-01 -1.48305923e-01 -1.30337954e-01
-5.50175488e-01 -6.56853259e-01 -2.25372970e-01 -1.65172994e-01
4.41032499e-01 -3.45908761e-01 -1.07199717e+00 1.50539982e+00
-3.15856665e-01 6.81956783e-02 -2.50091344e-01 1.92273818e-02
1.03960454e+00 4.83199328e-01 4.64733392e-01 -1.69794150e-02
1.15781629e+00 -8.66624057e-01 -7.36367285e-01 -5.17229617e-01
1.05726051e+00 -7.78791308e-01 1.22990453e+00 2.39503250e-01
-6.58930600e-01 -4.28481311e-01 -1.11194372e+00 1.59582391e-01
-7.10858881e-01 3.78396869e-01 6.10468268e-01 7.22472668e-01
-6.21629357e-01 3.22572023e-01 -3.83152246e-01 -1.93306208e-01
3.44242007e-01 -7.73197832e-03 -3.40054393e-01 -1.81737058e-02
-1.36425257e+00 1.30228877e+00 5.19142032e-01 -3.23659956e-01
-7.27975786e-01 -7.26186752e-01 -1.11418033e+00 8.76498222e-03
5.31672388e-02 -2.20968857e-01 1.26298833e+00 -3.70475590e-01
-1.07139134e+00 9.87509191e-01 -2.51449227e-01 -5.98327935e-01
1.96076721e-01 -7.24842012e-01 -5.74543774e-01 -5.25324643e-01
2.42190704e-01 3.30587536e-01 7.26780713e-01 -3.78388911e-01
-8.76432121e-01 5.62412217e-02 -4.69126493e-01 -7.25383312e-02
-5.76103091e-01 1.65012941e-01 -2.47642651e-01 -7.07342803e-01
-3.44944149e-01 -8.90666783e-01 2.12516129e-01 -6.96941257e-01
-6.83815658e-01 -9.11555707e-01 4.82308924e-01 -6.99746132e-01
1.60185766e+00 -2.26262236e+00 -3.80859151e-02 1.51922673e-01
1.40694112e-01 3.86632890e-01 -6.21316552e-01 4.54526134e-02
-1.47540644e-01 5.10160327e-01 2.72416472e-01 -3.90445530e-01
-6.92346618e-02 -9.54942331e-02 -3.10308397e-01 3.07714880e-01
2.35385150e-01 9.30513322e-01 -7.52776980e-01 -2.48152062e-01
-1.43267497e-01 -1.65199950e-01 -2.86945105e-01 1.36416227e-01
-1.70556039e-01 -1.56348348e-01 5.74319810e-03 2.67074734e-01
1.71619669e-01 1.13701925e-01 -1.01142071e-01 1.01584345e-01
-3.48461419e-01 1.01004028e+00 -7.56337404e-01 1.23856699e+00
-5.60592890e-01 9.18135643e-01 -4.43160444e-01 -2.85190403e-01
7.18587339e-01 2.35561579e-01 -7.39649162e-02 -4.19742852e-01
2.13957086e-01 1.04148120e-01 5.15731692e-01 4.77425233e-02
1.99078068e-01 3.65385205e-01 -3.18096697e-01 5.65282464e-01
3.96185428e-01 1.35844052e-01 3.54585022e-01 3.21365625e-01
1.30382931e+00 -1.46211326e-01 5.84380746e-01 -3.40730727e-01
7.08907768e-02 -2.49194726e-02 4.17585462e-01 1.12185383e+00
-9.86100435e-02 3.94952089e-01 1.51228055e-01 -3.68262112e-01
-7.86864042e-01 -9.22346532e-01 -3.53150576e-01 1.58572388e+00
-3.96371186e-01 -9.80043173e-01 -3.35517615e-01 -5.72299242e-01
9.13892463e-02 1.02565062e+00 -7.37339318e-01 -3.38912219e-01
-4.26632762e-01 -4.89771038e-01 8.35838974e-01 5.59097290e-01
-4.95451726e-02 -1.13737452e+00 -1.74015671e-01 1.84503227e-01
-2.13221580e-01 -1.17959106e+00 -9.44332659e-01 4.99652058e-01
-2.04678133e-01 -9.86073077e-01 -3.98646623e-01 -8.12842786e-01
-4.26322781e-03 2.77832955e-01 1.12192976e+00 -3.81677337e-02
-4.63592708e-01 1.66223869e-02 -7.09306657e-01 -8.11899424e-01
-4.13194239e-01 8.40730429e-01 2.67276913e-01 -4.87698615e-01
5.06848991e-01 2.70382278e-02 1.89053491e-01 -1.17523521e-01
-3.02611560e-01 7.79266953e-02 3.12010288e-01 1.06753492e+00
-1.97231770e-03 -2.89503872e-01 1.98417991e-01 -1.11964953e+00
7.07146645e-01 -2.70355880e-01 -2.81077027e-01 5.50774336e-01
-7.49729991e-01 1.43787384e-01 1.02522701e-01 -9.57738101e-01
-6.70941591e-01 -1.20203376e-01 -3.19565237e-01 1.09234482e-01
1.55380383e-01 5.81407249e-01 -1.87721655e-01 2.39933252e-01
8.40207934e-01 1.99667767e-01 -4.83741194e-01 -4.63323504e-01
3.87528062e-01 6.33609116e-01 4.28893566e-01 -3.49138737e-01
7.81334758e-01 -4.22280520e-01 -6.34231925e-01 -8.91823173e-01
-9.95420158e-01 -6.93146944e-01 -6.90730214e-01 -7.06506222e-02
6.18459105e-01 -1.10889864e+00 -4.52329874e-01 6.57372534e-01
-1.11655152e+00 -4.54694986e-01 1.41966268e-01 5.34697175e-01
1.71517432e-01 1.58851326e-01 -5.70834100e-01 -8.31568062e-01
-8.51831555e-01 -8.12550604e-01 5.51000893e-01 3.20398100e-02
-1.08190405e+00 -6.78960145e-01 1.30768910e-01 2.96128951e-02
5.56123197e-01 -3.08391571e-01 1.28916371e+00 -9.76459384e-01
1.83194697e-01 -2.51833290e-01 2.83605815e-03 8.03686380e-02
9.47059691e-02 2.59747654e-01 -8.57601464e-01 -3.43225330e-01
-7.05081224e-01 -3.81412357e-01 1.17819250e+00 3.16962510e-01
8.05090725e-01 -2.81551629e-01 -6.28957868e-01 5.58211446e-01
9.61430073e-01 -3.14059220e-02 6.92326948e-02 4.69770104e-01
6.70355022e-01 1.83192432e-01 2.61743188e-01 3.58342230e-01
6.55059442e-02 6.92761838e-01 -1.58267364e-01 3.44450623e-01
-2.83978730e-01 -4.70338315e-01 6.72396719e-01 6.70039773e-01
1.30909204e-01 -1.85300902e-01 -1.35337949e+00 6.73364639e-01
-1.26149571e+00 -7.51320004e-01 -2.32758313e-01 2.07137394e+00
9.09220040e-01 8.19235086e-01 2.80469209e-01 8.53439346e-02
7.32998431e-01 1.01271026e-01 -4.30175722e-01 -4.29027557e-01
-3.88080329e-01 7.18556583e-01 9.07731950e-01 6.51433408e-01
-1.10720980e+00 1.37470186e+00 7.53865290e+00 1.03545463e+00
-8.44717801e-01 2.76148587e-01 2.90817320e-01 -2.42666215e-01
-3.61271620e-01 -1.53262541e-01 -1.41069448e+00 4.85219479e-01
1.16756368e+00 -3.89011621e-01 4.22677577e-01 7.22298980e-01
-4.79554236e-01 -1.81672633e-01 -1.02064645e+00 7.49173045e-01
-7.79643655e-02 -1.04005158e+00 1.12178594e-01 -8.36222097e-02
3.32611769e-01 6.62512839e-01 2.39392400e-01 7.88755298e-01
7.84431577e-01 -1.05982280e+00 1.02258861e+00 -2.30760816e-02
1.27527642e+00 -5.68974555e-01 5.40944755e-01 2.34235466e-01
-9.68564868e-01 -2.89192144e-02 -1.07648991e-01 -2.34242097e-01
-3.55077088e-01 3.59754801e-01 -1.00111711e+00 -1.41677231e-01
6.90777838e-01 5.40929198e-01 -9.02306139e-01 7.93211699e-01
-5.70412099e-01 1.53414357e+00 -4.67604190e-01 -5.42416990e-01
1.09498955e-01 2.51570374e-01 5.82546115e-01 1.67558563e+00
-7.58230835e-02 -1.98559299e-01 -7.69161806e-02 2.20190629e-01
-1.22148760e-01 3.80832195e-01 -6.00421369e-01 -3.00288171e-01
8.04944098e-01 1.06039298e+00 -2.93419152e-01 -2.38382578e-01
-5.38254976e-01 6.83888674e-01 8.36494982e-01 9.77640152e-02
-3.39784622e-01 -4.81840551e-01 8.27939451e-01 -1.64786786e-01
2.51981378e-01 -4.58944738e-01 -6.78082705e-01 -1.11694169e+00
-3.53314012e-01 -7.12088645e-01 7.53787816e-01 -2.00941294e-01
-1.23411620e+00 5.59053481e-01 -1.29548624e-01 -7.89169967e-01
-3.25611383e-01 -8.66747022e-01 -9.04011130e-01 8.87915254e-01
-1.30012631e+00 -9.02129829e-01 1.96520194e-01 3.80315959e-01
6.82909310e-01 -7.99113572e-01 1.15587318e+00 -1.29517578e-02
-9.77112353e-01 1.23873711e+00 2.44402215e-01 6.22500956e-01
9.02081013e-01 -1.35937488e+00 1.26243913e+00 6.45928204e-01
5.08362770e-01 6.78383589e-01 5.46234787e-01 -6.68610334e-01
-8.92280281e-01 -8.58819842e-01 1.24470925e+00 -5.85794687e-01
9.34111834e-01 -7.22913861e-01 -4.46451515e-01 8.51072073e-01
2.32314408e-01 -3.24819267e-01 9.63020504e-01 9.14624870e-01
-7.78799772e-01 4.45481360e-01 -5.15192211e-01 6.65362716e-01
1.07863414e+00 -6.90010428e-01 -4.57478225e-01 2.61021852e-01
8.08022857e-01 -2.55468726e-01 -4.12378490e-01 1.64295644e-01
8.27280283e-01 -1.32896388e-02 7.21327782e-01 -1.00144160e+00
2.79229581e-01 3.91420960e-01 1.17555827e-01 -1.67875504e+00
-6.88587010e-01 -5.13551295e-01 -1.24417864e-01 1.08837950e+00
1.15041018e+00 -4.90839899e-01 2.82312870e-01 4.41257626e-01
7.37238750e-02 -4.96532530e-01 -1.10494936e+00 -7.70968556e-01
3.60933036e-01 -6.00269854e-01 1.57206401e-01 8.31928968e-01
4.03618366e-01 1.01945281e+00 -2.44429007e-01 -4.27755773e-01
2.34155744e-01 -2.74906307e-01 4.56446856e-01 -1.28627014e+00
-2.05183312e-01 -6.45886660e-01 -1.17583841e-01 -5.57245493e-01
5.72450936e-01 -1.13316607e+00 -5.18273301e-02 -1.06874609e+00
3.94770831e-01 -1.55092001e-01 -2.96772480e-01 7.67148435e-01
-7.27282763e-01 -2.15474945e-02 2.90309221e-01 3.25697005e-01
-3.14129442e-01 2.78943837e-01 6.63897634e-01 -5.79121947e-01
7.44098984e-03 5.16317189e-02 -4.93135035e-01 4.20443833e-01
9.22115028e-01 -4.57550853e-01 3.34975906e-02 -6.40697360e-01
3.51409838e-02 -6.17452204e-01 -5.29052496e-01 -9.40506339e-01
1.60219014e-01 2.14750320e-01 3.22895885e-01 -5.39867938e-01
1.00737058e-01 -9.29164961e-02 -4.50762123e-01 5.70124447e-01
-6.38383329e-01 1.81206167e-01 3.29709589e-01 1.25511780e-01
1.46433502e-01 -4.88498926e-01 6.65957868e-01 5.24355248e-02
-5.40306270e-01 2.04950050e-01 -9.49469805e-01 3.72308671e-01
6.17867053e-01 1.51441813e-01 -3.79319042e-01 -3.36536467e-01
-5.91475904e-01 2.34089315e-01 -1.07479654e-01 7.67835081e-01
3.49021375e-01 -1.09790206e+00 -9.95603025e-01 1.24195926e-01
4.60817516e-01 -3.79152417e-01 -4.88073409e-01 -1.97150651e-03
-1.71876073e-01 5.15786409e-01 -2.01037005e-01 -2.77497191e-02
-1.39870894e+00 1.16269879e-01 1.97077408e-01 -4.93509382e-01
-3.88145506e-01 1.14917397e+00 -2.02881526e-02 -3.00544053e-01
5.12624681e-01 -3.42255756e-02 -3.72197062e-01 3.69242281e-01
7.37323225e-01 3.72627795e-01 1.30262539e-01 -5.13240516e-01
-3.33396912e-01 9.66463685e-02 -7.75128603e-01 -2.50176340e-01
1.10792696e+00 1.90705001e-01 2.55175203e-01 7.04592228e-01
1.20247805e+00 -4.83088791e-02 -5.12726486e-01 -6.63936555e-01
5.20034194e-01 9.73871127e-02 9.80836600e-02 -8.48560512e-01
-4.33413893e-01 6.13881469e-01 4.87835675e-01 -3.03378850e-01
4.05474305e-01 -3.42553668e-02 9.51397479e-01 2.79512495e-01
2.02928215e-01 -1.10473311e+00 -2.50682682e-01 1.24802959e+00
5.26800692e-01 -1.13036239e+00 -6.32124320e-02 -2.78011382e-01
-3.99490297e-01 9.54791307e-01 7.42249012e-01 2.79116780e-01
8.08688283e-01 5.54752529e-01 1.41794503e-01 6.15180470e-02
-9.48354185e-01 -5.86240232e-01 4.21791881e-01 4.41928595e-01
7.75014579e-01 4.03526634e-01 -6.96054161e-01 8.10209811e-01
-6.04690135e-01 -8.09023499e-01 3.31674099e-01 5.67358732e-01
-4.76455778e-01 -8.49550486e-01 -6.63685650e-02 7.86894739e-01
-4.50119078e-01 -6.28361642e-01 -6.28109097e-01 6.33778393e-01
-6.59925044e-02 9.28740978e-01 1.19528040e-01 -6.11856103e-01
3.17923427e-01 6.88779533e-01 3.91690403e-01 -9.37010527e-01
-6.76447809e-01 -4.55061525e-01 3.49085927e-01 -1.49852142e-01
2.94125199e-01 -8.35919082e-01 -7.77069330e-01 4.74288538e-02
-6.06797099e-01 -7.02385083e-02 6.22704387e-01 1.14779162e+00
-1.57394812e-01 1.90994024e-01 3.40534270e-01 -2.94528484e-01
-5.30585706e-01 -1.78034139e+00 -3.63635838e-01 3.76587570e-01
1.38906658e-01 -6.67884529e-01 -2.31212199e-01 -2.47313455e-01] | [10.499053001403809, 10.491033554077148] |
5325e099-2b2a-49a0-b30b-271c5e5105ea | multimodal-emotion-recognition-model-using | 1911.12918 | null | https://arxiv.org/abs/1911.12918v2 | https://arxiv.org/pdf/1911.12918v2.pdf | Multimodal Affective States Recognition Based on Multiscale CNNs and Biologically Inspired Decision Fusion Model | There has been an encouraging progress in the affective states recognition models based on the single-modality signals as electroencephalogram (EEG) signals or peripheral physiological signals in recent years. However, multimodal physiological signals-based affective states recognition methods have not been thoroughly exploited yet. Here we propose Multiscale Convolutional Neural Networks (Multiscale CNNs) and a biologically inspired decision fusion model for multimodal affective states recognition. Firstly, the raw signals are pre-processed with baseline signals. Then, the High Scale CNN and Low Scale CNN in Multiscale CNNs are utilized to predict the probability of affective states output for EEG and each peripheral physiological signal respectively. Finally, the fusion model calculates the reliability of each single-modality signals by the Euclidean distance between various class labels and the classification probability from Multiscale CNNs, and the decision is made by the more reliable modality information while other modalities information is retained. We use this model to classify four affective states from the arousal valence plane in the DEAP and AMIGOS dataset. The results show that the fusion model improves the accuracy of affective states recognition significantly compared with the result on single-modality signals, and the recognition accuracy of the fusion result achieve 98.52% and 99.89% in the DEAP and AMIGOS dataset respectively. | ['Yuxuan Zhao', 'Xinyan Cao', 'Jinlong Lin', 'Xixin Cao', 'Dunshan Yu'] | 2019-11-29 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 2.14237764e-01 -3.17613214e-01 3.52956206e-01 -5.83519936e-01
-4.76866215e-01 -1.49940342e-01 3.27495813e-01 1.97791755e-01
-4.82249856e-01 1.02332807e+00 1.83509201e-01 4.56948608e-01
-1.05055466e-01 -6.87749088e-01 -3.10105771e-01 -1.00011969e+00
-7.58377789e-03 -3.52587909e-01 -3.62691849e-01 -1.19615734e-01
1.04931183e-01 3.02224725e-01 -1.78393424e+00 6.54258609e-01
7.67779112e-01 2.00861430e+00 -4.00174856e-01 6.70479894e-01
1.53825013e-03 5.77978730e-01 -7.00355947e-01 -1.32325860e-02
-2.61405408e-01 -4.62026596e-01 -4.89570260e-01 -3.88099909e-01
-3.31268728e-01 2.39734158e-01 -1.04405463e-01 1.13174987e+00
9.05295372e-01 1.45727515e-01 7.56347060e-01 -1.50504434e+00
-5.27415454e-01 2.22666994e-01 -4.29048419e-01 2.48216897e-01
5.26812375e-01 -3.55911553e-02 4.84998018e-01 -8.85968745e-01
-1.39386743e-01 9.85426307e-01 6.18332744e-01 3.56597126e-01
-9.04020131e-01 -9.07540917e-01 7.08478764e-02 6.89479291e-01
-1.54640496e+00 -3.71284872e-01 7.99310803e-01 -3.12221259e-01
1.16732717e+00 1.24511540e-01 1.10491049e+00 1.21070063e+00
8.42343390e-01 5.96049547e-01 1.50742257e+00 4.71845455e-03
4.42695975e-01 1.79862708e-01 2.61596650e-01 4.92727667e-01
-1.68045983e-01 -2.53165990e-01 -8.51306021e-01 -1.71208560e-01
4.29409802e-01 1.19326122e-01 -1.26760051e-01 4.22909319e-01
-1.10889280e+00 3.38932425e-01 4.89167154e-01 5.70396662e-01
-1.03010297e+00 -9.01308060e-02 5.31524658e-01 3.00504446e-01
4.46082532e-01 2.77490288e-01 -4.78217572e-01 -2.49206260e-01
-8.24482620e-01 -3.25721174e-01 5.42508125e-01 2.49212995e-01
7.62184680e-01 1.83570221e-01 -8.69578496e-02 9.24677193e-01
2.56022632e-01 5.22520721e-01 7.11227834e-01 -7.19492376e-01
-1.65598541e-01 8.22616577e-01 -3.40617485e-02 -1.42889237e+00
-7.93196142e-01 -4.68036495e-02 -1.38074589e+00 9.72482860e-02
-2.00214326e-01 -4.68005478e-01 -6.79883063e-01 1.93865192e+00
1.11997370e-02 3.73806208e-01 4.58315700e-01 9.81892884e-01
1.11119616e+00 1.07356691e+00 4.38572019e-01 -5.48428118e-01
1.46422470e+00 -2.23296225e-01 -1.17137074e+00 -4.43284847e-02
-1.56221062e-01 -3.52748930e-01 7.96112776e-01 6.39608502e-01
-1.20141804e+00 -8.16420674e-01 -1.36070645e+00 3.38869095e-01
-6.54093206e-01 2.43170053e-01 5.75688720e-01 4.82974231e-01
-1.02947080e+00 4.24729824e-01 -7.84926534e-01 -3.61826599e-01
3.90609622e-01 6.67050421e-01 -4.14471507e-01 5.68668604e-01
-1.80229449e+00 1.01815379e+00 4.29940850e-01 4.70622748e-01
-6.36279941e-01 -3.06959957e-01 -7.27534592e-01 2.87474602e-01
-5.10017037e-01 -4.90552902e-01 6.94570124e-01 -1.35200167e+00
-1.67759240e+00 5.33585310e-01 -1.38535410e-01 -3.18482891e-02
-4.99634296e-01 1.44556344e-01 -8.23633850e-01 4.02259320e-01
-3.10726881e-01 8.28022003e-01 5.52766919e-01 -8.58821332e-01
-3.91233057e-01 -5.87111235e-01 -3.89063567e-01 4.33704853e-01
-6.83462858e-01 -2.84316819e-02 2.21790716e-01 -1.43445330e-02
2.44530588e-01 -7.07164824e-01 9.98809487e-02 -4.73650366e-01
-1.62401527e-01 -3.72790933e-01 4.78485346e-01 -5.36953628e-01
1.01107597e+00 -2.19969916e+00 2.83194989e-01 3.48166704e-01
-2.66999695e-02 2.26739682e-02 -4.51144874e-02 9.62615535e-02
-3.55558008e-01 -2.67350767e-02 -6.79465234e-02 -9.60856974e-02
-4.03688662e-02 3.32361599e-03 -3.66559848e-02 2.54909635e-01
5.32409191e-01 8.28856945e-01 -5.26483059e-01 -5.57055652e-01
3.56444955e-01 5.91042578e-01 1.01197869e-01 3.91851038e-01
5.02610922e-01 5.43369234e-01 -3.48964036e-01 7.87779748e-01
6.32904172e-01 -1.01317421e-01 1.38199767e-02 -5.64128101e-01
-1.86923638e-01 -5.52870370e-02 -1.02548528e+00 1.75357378e+00
-4.77692485e-01 4.82298642e-01 2.52344385e-02 -1.08806145e+00
1.29844117e+00 7.65998483e-01 6.54578388e-01 -7.23425388e-01
7.61467338e-01 3.17740068e-02 -9.44518298e-02 -7.10205913e-01
1.71506330e-01 -5.51485479e-01 -4.49026853e-01 1.66165277e-01
4.85823244e-01 -1.41472772e-01 -3.28987867e-01 -1.29287168e-01
9.13058043e-01 -9.41601321e-02 2.61415809e-01 -2.63862193e-01
7.43466854e-01 -4.07009423e-01 6.66997254e-01 2.44632781e-01
-5.26998103e-01 3.78911465e-01 4.26188499e-01 -5.07210851e-01
-4.54757363e-01 -1.11276436e+00 -1.47705629e-01 1.00026715e+00
4.75329459e-01 -8.67379680e-02 -6.14851356e-01 -1.18143067e-01
-3.30219448e-01 3.71333778e-01 -9.04315710e-01 -7.63553321e-01
3.65884721e-01 -1.24116194e+00 7.11862087e-01 7.09677994e-01
8.03335130e-01 -1.48873508e+00 -4.85997409e-01 2.35392883e-01
-4.19075787e-01 -9.00967658e-01 4.29023564e-01 5.09319365e-01
-6.96323931e-01 -6.03249192e-01 -4.54236686e-01 -7.90311396e-01
2.52277732e-01 -5.68371236e-01 6.86544955e-01 -4.00850475e-01
-2.07205325e-01 3.33375961e-01 -8.43299106e-02 -5.46850145e-01
1.50819480e-01 -2.23850146e-01 4.28170294e-01 5.91808498e-01
7.41920471e-01 -1.15836871e+00 -7.57629156e-01 -8.91081914e-02
-6.64435387e-01 9.20078605e-02 6.59474492e-01 5.11777341e-01
6.42168462e-01 -2.07797885e-01 1.22654831e+00 1.60607964e-01
1.05038846e+00 -5.65483451e-01 2.46221244e-01 1.83533445e-01
-3.29614997e-01 -3.59489501e-01 7.16706872e-01 -5.07469475e-01
-1.23475313e+00 7.64546618e-02 -2.61755913e-01 -4.41245317e-01
-6.22359157e-01 7.10792899e-01 -2.64749806e-02 8.99134129e-02
5.48788428e-01 4.33107764e-01 -3.50836366e-01 1.93211153e-01
-8.04965664e-03 1.04081035e+00 6.85761452e-01 -4.44091886e-01
-3.48811984e-01 1.65476039e-01 -1.52222142e-01 -5.60330033e-01
-5.95698655e-01 -1.28988460e-01 -4.67361420e-01 -6.25842273e-01
1.31748855e+00 -9.66734111e-01 -1.19771385e+00 6.86392725e-01
-1.19189978e+00 2.75066793e-01 1.05294518e-01 7.68278241e-01
-5.93246579e-01 1.19008191e-01 -1.05176473e+00 -1.26770329e+00
-7.93239772e-01 -1.05220687e+00 8.53734195e-01 7.59476542e-01
-4.54117596e-01 -6.24064863e-01 2.06698760e-01 -6.76533729e-02
4.45818216e-01 5.44181168e-01 9.86517489e-01 -5.95276833e-01
4.21394825e-01 -3.95880282e-01 -1.23076119e-01 4.84120071e-01
1.33633107e-01 9.33034644e-02 -1.42543983e+00 3.74440670e-01
2.00258985e-01 -7.20693171e-01 5.68598330e-01 4.43530262e-01
1.12625587e+00 1.15431339e-01 1.07685803e-02 2.03063667e-01
1.23238528e+00 8.63372803e-01 8.19201112e-01 -2.32352808e-01
1.51224300e-01 5.28290093e-01 2.84845144e-01 5.86910069e-01
3.50772142e-01 -1.34410724e-01 4.12499070e-01 -1.90980241e-01
5.21903157e-01 3.02764952e-01 5.61748505e-01 1.41634476e+00
-1.48950487e-01 7.65868276e-02 -7.58062899e-01 5.23305714e-01
-1.97793388e+00 -9.83196497e-01 1.75407995e-02 1.81169593e+00
9.57786083e-01 -1.27948120e-01 -2.35990703e-01 2.43925154e-01
8.14868569e-01 -9.73678753e-02 -7.73243964e-01 -8.14662039e-01
-4.24754053e-01 5.22734344e-01 -3.88005525e-01 -2.41717119e-02
-1.11449838e+00 4.78865147e-01 6.14168310e+00 6.58088088e-01
-1.32333708e+00 9.20362994e-02 9.84525442e-01 -3.98974061e-01
1.97099105e-01 -6.23812258e-01 -1.10216476e-01 5.29924810e-01
1.37293649e+00 -5.19610585e-05 4.68948901e-01 5.22013485e-01
2.80340612e-01 -5.05315304e-01 -9.29607332e-01 1.45214796e+00
9.02835354e-02 -6.32366955e-01 -1.17345020e-01 -4.84351903e-01
6.10032201e-01 -1.62379280e-01 1.44258529e-01 4.91272181e-01
-1.83965564e-01 -1.13604581e+00 4.15540099e-01 1.38555348e+00
6.43603265e-01 -1.06750691e+00 1.08768654e+00 2.10363001e-01
-1.27975047e+00 -1.21910021e-01 -3.94099474e-01 -4.02909070e-01
-4.70077619e-02 6.19341791e-01 -1.23045713e-01 4.27161902e-01
1.05973697e+00 7.74555027e-01 -4.45711136e-01 4.78364974e-01
5.40965684e-02 6.44007981e-01 -3.56380105e-01 -4.47580576e-01
-2.52793524e-02 -3.05658430e-01 -3.77096655e-03 1.31818724e+00
4.30504233e-01 5.25587797e-01 -1.02183007e-01 9.74439323e-01
2.99781114e-02 2.34514579e-01 -3.50756347e-01 -1.52740166e-01
2.23498389e-01 1.86732841e+00 -5.94501376e-01 -7.22348571e-01
-1.84804007e-01 9.28007603e-01 -7.95921013e-02 3.44718874e-01
-9.44422960e-01 -7.39601731e-01 6.53750181e-01 -9.36456978e-01
-5.07294774e-01 1.96990833e-01 -7.94295430e-01 -1.25628376e+00
-2.51664817e-01 -4.53208745e-01 2.88026363e-01 -1.45344007e+00
-1.40626645e+00 8.34191561e-01 -2.23839045e-01 -9.80158985e-01
6.88867941e-02 -5.16758502e-01 -9.44617152e-01 1.08016920e+00
-1.18816245e+00 -6.43830299e-01 -6.39493585e-01 9.17326331e-01
1.44813150e-01 1.43987294e-02 1.22634888e+00 2.23424062e-01
-7.30574727e-01 1.46600142e-01 -1.58737123e-01 -2.25363877e-02
6.68451965e-01 -9.47153986e-01 -5.44464827e-01 3.96552384e-01
-6.32521629e-01 6.72179103e-01 3.12941611e-01 -2.63407648e-01
-1.10195458e+00 -6.99867606e-01 5.51961601e-01 7.71455914e-02
4.46083605e-01 -1.51144907e-01 -8.68346095e-01 5.20569868e-02
7.86093473e-01 -1.37972683e-01 1.09643042e+00 -2.46681944e-02
5.07742502e-02 -2.71616369e-01 -1.30407250e+00 5.04759312e-01
3.69977564e-01 -7.03314185e-01 -8.43310475e-01 -1.04438975e-01
2.62517184e-01 1.84341043e-01 -1.45694304e+00 7.65286446e-01
9.46695805e-01 -8.95900071e-01 5.74303031e-01 -5.49977243e-01
6.20607555e-01 -3.60112995e-01 -3.80364239e-01 -1.41968131e+00
-3.46567780e-01 1.58894941e-01 1.12402119e-01 1.10212350e+00
3.93836915e-01 -6.22644305e-01 2.19816625e-01 7.74140894e-01
-1.96266826e-02 -9.64820266e-01 -1.08675444e+00 -2.78243925e-02
-2.56027341e-01 -4.90854710e-01 4.55774337e-01 9.09077227e-01
7.06532657e-01 5.74660301e-01 -1.67755097e-01 5.80718927e-02
7.62758255e-02 2.14544237e-01 1.40881557e-02 -1.25042391e+00
3.28872323e-01 -4.06740278e-01 -6.31782115e-01 -2.46181607e-01
1.39118478e-01 -7.91467011e-01 2.62370169e-01 -1.62655878e+00
3.19217414e-01 2.21833527e-01 -1.31998432e+00 7.23263800e-01
-1.85309187e-01 6.04122341e-01 -2.66292375e-02 -1.55678362e-01
-8.21076393e-01 1.02458858e+00 8.05209756e-01 -1.34569407e-01
-2.94475168e-01 -4.13681686e-01 -5.43654263e-01 7.95595050e-01
1.03934228e+00 -1.61107987e-01 -1.18279442e-01 3.40672940e-01
3.78129810e-01 4.75021034e-01 2.89764911e-01 -1.49460065e+00
2.76434779e-01 7.11692823e-03 1.23561311e+00 -4.94958729e-01
9.46753561e-01 -6.34312153e-01 8.35574940e-02 2.68051922e-01
-4.12814856e-01 1.25962645e-02 4.12569791e-01 3.39065135e-01
-3.48742753e-01 1.35582313e-01 8.44046354e-01 -3.70092355e-02
-6.19350612e-01 1.72051415e-01 -9.58853364e-01 -5.33468306e-01
9.41504240e-01 -2.69660234e-01 -1.20710403e-01 -5.16066611e-01
-1.29031467e+00 -3.48390453e-02 -2.12458342e-01 3.47946763e-01
8.92779231e-01 -1.80373681e+00 -3.56922686e-01 3.04148108e-01
1.89497881e-02 -6.18805408e-01 6.46362841e-01 1.36631036e+00
1.67937770e-01 1.91221219e-02 -9.60705936e-01 -6.33631706e-01
-1.12667775e+00 2.48940453e-01 5.10484576e-01 1.71449602e-01
3.34135562e-01 6.66453540e-01 8.77786055e-02 -2.09371507e-01
-5.30812405e-02 -3.59597892e-01 -8.61053109e-01 5.91831744e-01
5.03428757e-01 2.30590880e-01 1.49112970e-01 -7.00294256e-01
-5.96830189e-01 5.74505091e-01 5.21112144e-01 -3.92950565e-01
1.45784771e+00 -1.78639084e-01 -5.24415910e-01 1.06751680e+00
1.21610713e+00 -5.95243156e-01 -8.05799603e-01 1.25871766e-02
-4.59519982e-01 2.81937510e-01 7.51049072e-02 -1.21783149e+00
-1.09594560e+00 1.35598886e+00 1.06784618e+00 3.29971999e-01
1.75389624e+00 -3.49611342e-01 7.17902064e-01 4.26989675e-01
1.89780384e-01 -1.33212137e+00 1.66953087e-01 3.49626869e-01
8.15451562e-01 -9.74617124e-01 -2.75943279e-01 3.17355186e-01
-9.54939365e-01 1.25105417e+00 7.14428663e-01 -3.23703527e-01
7.86086738e-01 2.70818204e-01 2.10445262e-02 -1.72002822e-01
-9.76364851e-01 -3.64763848e-02 3.25585276e-01 3.51768404e-01
3.59376013e-01 1.70841143e-01 -2.76210308e-01 1.69545400e+00
2.81243660e-02 1.86740831e-01 1.91643253e-01 7.21032619e-01
-6.37531340e-01 -2.22771242e-01 -5.31426907e-01 6.22162521e-01
-5.75198352e-01 -1.78238302e-01 -6.52874351e-01 1.07445158e-01
3.43298346e-01 1.28736413e+00 1.60102025e-01 -1.12470210e+00
3.03318143e-01 7.91800857e-01 3.09722364e-01 -2.84867376e-01
-9.42943275e-01 1.23656809e-01 -1.85525090e-01 -5.78815341e-01
-7.65108764e-01 -3.80091488e-01 -1.47981596e+00 -6.38947040e-02
-2.12687358e-01 2.05769658e-01 7.74634242e-01 1.02328992e+00
6.39713705e-01 7.64092028e-01 7.10292757e-01 -1.16892588e+00
1.61522284e-01 -1.32226205e+00 -6.86433434e-01 4.55907375e-01
1.97148751e-02 -6.93193495e-01 -3.44749689e-01 1.44785335e-02] | [13.177459716796875, 3.442643165588379] |
987695fd-fc34-4567-a420-4fb202ca7eaa | knowledge-driven-event-embedding-for-stock | null | null | https://aclanthology.org/C16-1201 | https://aclanthology.org/C16-1201.pdf | Knowledge-Driven Event Embedding for Stock Prediction | Representing structured events as vectors in continuous space offers a new way for defining dense features for natural language processing (NLP) applications. Prior work has proposed effective methods to learn event representations that can capture syntactic and semantic information over text corpus, demonstrating their effectiveness for downstream tasks such as event-driven stock prediction. On the other hand, events extracted from raw texts do not contain background knowledge on entities and relations that they are mentioned. To address this issue, this paper proposes to leverage extra information from knowledge graph, which provides ground truth such as attributes and properties of entities and encodes valuable relations between entities. Specifically, we propose a joint model to combine knowledge graph information into the objective function of an event embedding learning model. Experiments on event similarity and stock market prediction show that our model is more capable of obtaining better event embeddings and making more accurate prediction on stock market volatilities. | ['Yue Zhang', 'Ting Liu', 'Junwen Duan', 'Xiao Ding'] | 2016-12-01 | knowledge-driven-event-embedding-for-stock-1 | https://aclanthology.org/C16-1201 | https://aclanthology.org/C16-1201.pdf | coling-2016-12 | ['stock-market-prediction', 'stock-prediction'] | ['time-series', 'time-series'] | [-3.67306679e-01 1.90933011e-02 -5.64279914e-01 -5.37703872e-01
-2.06183895e-01 -7.17804313e-01 9.06877160e-01 8.74118209e-01
-4.76301551e-01 6.77261114e-01 8.83366942e-01 -2.13059619e-01
-9.15689915e-02 -1.57645619e+00 -5.59152246e-01 -2.90152192e-01
-4.97879803e-01 2.42504373e-01 2.79840350e-01 -1.94361433e-01
1.54759526e-01 5.59881210e-01 -1.15479696e+00 1.16672821e-01
4.91453260e-01 1.17655575e+00 -1.16741009e-01 1.03254408e-01
-6.89251423e-01 1.34072161e+00 -4.80028629e-01 -8.07132900e-01
2.25711197e-01 -1.73501343e-01 -6.10829294e-01 -2.49238357e-01
-1.94572508e-01 -3.08575958e-01 -9.08926725e-01 9.76621807e-01
1.88659653e-02 2.92272002e-01 8.94220531e-01 -1.40265000e+00
-1.08927989e+00 1.10232842e+00 -2.87551016e-01 8.56775880e-01
3.70776989e-02 -1.74887449e-01 1.76992571e+00 -8.74605358e-01
6.74663782e-01 1.01219487e+00 4.98924255e-01 -6.51342422e-02
-1.01196897e+00 -7.06115484e-01 3.72689068e-01 3.43883067e-01
-1.05954850e+00 6.53636381e-02 8.60254467e-01 -4.75414872e-01
1.34997928e+00 5.71116209e-02 7.97510684e-01 1.08632803e+00
2.97792882e-01 8.37213874e-01 4.06875432e-01 -7.87824467e-02
2.32994422e-01 2.37959877e-01 5.88303924e-01 3.33258539e-01
7.86265433e-01 1.56336769e-01 -7.42636740e-01 -2.22173393e-01
8.10773671e-01 5.95730960e-01 -1.59382477e-01 -3.20452690e-01
-1.30885386e+00 1.34573627e+00 6.34905875e-01 6.38035417e-01
-8.06949198e-01 2.54480958e-01 6.06030643e-01 3.66984725e-01
7.97899067e-01 6.13971770e-01 -7.45792329e-01 -9.42054242e-02
-5.79706192e-01 2.91261852e-01 9.02170956e-01 6.66991413e-01
7.10929632e-01 1.90867990e-01 -1.31981432e-01 3.57221663e-01
3.23115051e-01 3.11042637e-01 7.65967429e-01 -4.85045649e-02
6.20265424e-01 9.84991133e-01 1.83085483e-02 -1.32325780e+00
-4.57263678e-01 -3.88179660e-01 -3.50419253e-01 -2.42523044e-01
4.83975150e-02 -2.03500137e-01 -6.20647073e-01 1.59328985e+00
2.10731283e-01 7.35661447e-01 3.38918060e-01 7.66627729e-01
9.09180403e-01 1.02381289e+00 3.05172533e-01 -1.22305371e-01
1.53746176e+00 -5.50454676e-01 -6.70573831e-01 -1.29333898e-01
5.64848006e-01 -3.13693047e-01 5.73909461e-01 -2.91959077e-01
-7.25734293e-01 -1.94442898e-01 -8.90949965e-01 3.66026498e-02
-9.27510977e-01 -3.07407290e-01 1.17025578e+00 2.89565176e-01
-3.58776927e-01 6.10576570e-01 -1.04901516e+00 -1.33846626e-01
4.23840910e-01 5.78807257e-02 -2.31568441e-01 2.94778168e-01
-1.79375887e+00 9.82589722e-01 8.55293870e-01 -2.60897040e-01
-4.66397882e-01 -1.03049779e+00 -1.34910631e+00 6.00090921e-01
3.40961426e-01 -5.03043950e-01 9.96884167e-01 -4.38321590e-01
-1.10949731e+00 4.03050154e-01 1.20089747e-01 -1.12944400e+00
-2.17506126e-01 -7.24083409e-02 -8.23577106e-01 4.99941856e-02
2.30801981e-02 1.90851107e-01 5.26125550e-01 -4.61332768e-01
-7.46270359e-01 -3.23478907e-01 2.36197561e-01 -1.95783060e-02
-7.41639674e-01 -2.92758867e-02 2.11682301e-02 -1.03215218e+00
-2.08104834e-01 -2.74026036e-01 -1.82729721e-01 -4.00082111e-01
-1.82439417e-01 -6.49292946e-01 5.62198222e-01 -5.65013170e-01
1.45490503e+00 -2.05130029e+00 -1.27716362e-01 3.13011289e-01
3.11578125e-01 1.16143204e-01 -1.08755855e-02 8.45234811e-01
-3.77043821e-02 1.14999851e-02 2.35323627e-02 7.62029216e-02
3.75411302e-01 3.38900298e-01 -1.08753550e+00 2.03139484e-01
6.75079942e-01 1.33643913e+00 -1.02412903e+00 -1.45605281e-01
1.99692905e-01 5.59298158e-01 -3.43871146e-01 5.87767363e-02
-2.45425925e-01 -8.84084702e-02 -8.65501761e-01 3.24508578e-01
1.76816553e-01 -4.53421235e-01 2.02959608e-02 -2.31313005e-01
3.14289965e-02 8.44863176e-01 -1.18644512e+00 1.24801636e+00
-4.48428154e-01 5.56160331e-01 -7.71569312e-01 -1.13175929e+00
9.34661150e-01 3.62234890e-01 5.81995964e-01 -6.71703577e-01
2.47260138e-01 -1.43093452e-01 -3.23524505e-01 -1.87283888e-01
6.98124349e-01 -4.89671826e-01 -2.71287948e-01 5.17221034e-01
1.90987170e-01 2.65288979e-01 2.95894831e-01 4.89122123e-01
1.11011505e+00 -4.29377198e-01 6.77503943e-01 -1.74525026e-02
1.96775824e-01 1.12795942e-01 7.22367346e-01 3.83832246e-01
-4.83748093e-02 5.08984104e-02 6.13237917e-01 -5.20332515e-01
-8.02513659e-01 -1.36774528e+00 -1.64832458e-01 7.93399274e-01
-2.06238348e-02 -6.78369105e-01 5.85374609e-02 -7.44079173e-01
5.88858366e-01 1.05857992e+00 -7.77665198e-01 -3.11509997e-01
-3.11604470e-01 -8.55893731e-01 3.33539307e-01 1.09874988e+00
1.16679855e-01 -1.00178587e+00 -3.62967432e-01 4.44745451e-01
1.19734354e-01 -1.15780437e+00 -4.98841286e-01 3.64197135e-01
-7.56315589e-01 -1.28368235e+00 -3.37444425e-01 -6.92147970e-01
2.88824558e-01 -1.62107840e-01 1.32453334e+00 -4.87241745e-01
-2.63116688e-01 3.59868169e-01 -4.74489301e-01 -6.33794367e-01
2.63748057e-02 -5.22080101e-02 -4.86709960e-02 2.51305312e-01
8.46787095e-01 -5.82646906e-01 -5.70220709e-01 -1.51927963e-01
-1.13420331e+00 -3.52611750e-01 3.27612013e-01 6.44929111e-01
4.61162210e-01 3.20267975e-01 9.88823295e-01 -9.39691365e-01
7.72453487e-01 -9.55426812e-01 -7.19662905e-01 2.50105023e-01
-7.71507263e-01 2.91264445e-01 5.48583508e-01 -4.12095636e-01
-1.11664236e+00 -2.85171211e-01 1.74487710e-01 -2.84330219e-01
2.25143284e-01 9.80714083e-01 9.42056626e-02 4.99906719e-01
3.91288966e-01 2.89764702e-01 -4.30091679e-01 -3.70350540e-01
5.72247088e-01 2.12916419e-01 2.98928857e-01 -2.44706303e-01
9.68422711e-01 4.79509950e-01 -5.57043590e-02 -5.88942111e-01
-9.95306075e-01 -7.25309193e-01 -3.24934602e-01 3.61152589e-01
9.08093512e-01 -1.03149271e+00 -5.59005380e-01 6.65486082e-02
-7.51347005e-01 1.96324438e-01 -6.27818167e-01 1.03309953e+00
-1.71248853e-01 6.04608320e-02 -9.40507710e-01 -6.47066176e-01
-2.30813995e-01 -4.37460721e-01 8.66963625e-01 1.32070571e-01
-1.99898198e-01 -1.53857338e+00 4.34082717e-01 -2.11252451e-01
3.49334896e-01 3.33385736e-01 1.00613272e+00 -1.24586403e+00
-6.75351381e-01 -4.58980918e-01 -3.65857512e-01 1.95562497e-01
3.35121721e-01 -2.55596519e-01 -6.35036349e-01 3.97407077e-02
-1.08748071e-01 -5.37970178e-02 1.28102946e+00 2.50946492e-01
6.21846139e-01 -4.49272841e-01 -2.99743384e-01 3.48217666e-01
1.44863880e+00 7.59528577e-02 5.37655771e-01 2.31710047e-01
5.78181922e-01 4.08325553e-01 3.64801973e-01 7.62953341e-01
5.64546227e-01 3.26166034e-01 2.04890072e-01 3.16558033e-01
2.07194805e-01 -6.85728848e-01 4.68365580e-01 6.89315856e-01
2.88636655e-01 -2.50239015e-01 -7.12366700e-01 6.34732306e-01
-1.75296223e+00 -1.29368794e+00 9.76779908e-02 1.80462885e+00
9.93054450e-01 1.71768948e-01 6.14056252e-02 3.35470922e-02
4.79532599e-01 5.98862708e-01 -4.75458056e-01 4.72227670e-02
-1.70209512e-01 3.22324663e-01 6.57643139e-01 1.47783384e-01
-1.14337552e+00 1.06179762e+00 5.59723282e+00 4.53105807e-01
-9.93878722e-01 -7.08188396e-03 7.29789212e-02 -3.28152925e-02
-8.08473945e-01 -5.85149862e-02 -9.62754309e-01 5.26353419e-01
1.08447027e+00 -8.28988254e-01 3.01825386e-02 6.54230177e-01
-6.58105090e-02 4.21686560e-01 -1.24964666e+00 7.98048198e-01
-7.15786666e-02 -1.84223866e+00 4.13196564e-01 1.36227040e-02
5.41323125e-01 1.86399147e-02 -6.04317747e-02 6.67243302e-01
5.81435144e-01 -7.38474905e-01 5.90263963e-01 5.89762747e-01
1.54258534e-01 -8.81906569e-01 7.36163199e-01 1.33908749e-01
-1.52497935e+00 -1.60009116e-02 -4.74532485e-01 -2.55149812e-01
3.70376229e-01 7.58440554e-01 -1.16054356e+00 7.56613791e-01
2.83533126e-01 1.27092564e+00 -4.21634406e-01 7.68746316e-01
-2.86579937e-01 8.08586836e-01 -2.63165385e-01 -3.05813491e-01
4.61034864e-01 -9.98873934e-02 3.05575192e-01 1.14877880e+00
2.17182815e-01 2.28764191e-01 1.82538018e-01 9.78250682e-01
-3.36554319e-01 3.29482168e-01 -8.66795480e-01 -7.89269865e-01
3.46270919e-01 9.22570825e-01 -8.70254874e-01 -4.22337443e-01
-9.01836932e-01 7.00605929e-01 5.47378778e-01 3.99508268e-01
-9.19490635e-01 -5.04538596e-01 1.05298686e+00 -6.74868897e-02
7.08878696e-01 -2.42030010e-01 -1.61187395e-01 -1.69756389e+00
-1.41045749e-01 -7.59144500e-02 8.33325207e-01 -4.61260229e-01
-1.83322954e+00 2.99068570e-01 -3.84538993e-02 -9.40524876e-01
-3.75480294e-01 -7.55184710e-01 -7.45602846e-01 6.76137984e-01
-1.77765000e+00 -8.77521098e-01 4.04887706e-01 6.38144732e-01
3.40411901e-01 -4.23881441e-01 6.70514405e-01 2.54115969e-01
-5.36946893e-01 2.46955335e-01 2.75233872e-02 6.98779166e-01
3.82637113e-01 -1.40567553e+00 6.14748120e-01 7.50736058e-01
9.70549047e-01 7.06718624e-01 4.40879554e-01 -9.05161321e-01
-1.37386751e+00 -1.51872110e+00 1.27824354e+00 -5.80226302e-01
1.38672364e+00 -8.93036053e-02 -9.36252654e-01 1.09873414e+00
1.96045879e-02 3.39879505e-02 1.01578403e+00 3.25475991e-01
-7.38942564e-01 -2.56656408e-02 -8.45536172e-01 4.13352281e-01
8.05375040e-01 -7.11617827e-01 -1.21864879e+00 1.82668671e-01
1.10381472e+00 5.97185269e-02 -1.02386510e+00 1.52910858e-01
6.06554300e-02 -3.06332827e-01 1.16811132e+00 -1.26359594e+00
4.14093882e-01 -1.24615341e-01 -1.64084017e-01 -1.41464925e+00
-3.09380054e-01 -3.99588086e-02 -5.29111207e-01 1.35203445e+00
5.83886206e-01 -9.54301596e-01 5.49475551e-01 6.26218259e-01
1.12882681e-01 -5.71594775e-01 -7.24404037e-01 -9.11325276e-01
3.40317488e-02 -5.53835928e-01 9.40661073e-01 1.35133135e+00
3.81054431e-01 5.40766180e-01 -1.53780639e-01 4.33900237e-01
3.13070714e-01 5.72176754e-01 2.47938812e-01 -1.55610275e+00
-1.36508688e-01 -4.89886791e-01 -8.34261298e-01 -7.38349378e-01
6.26668692e-01 -1.39395344e+00 -6.36584878e-01 -1.64473200e+00
1.16077229e-01 -6.67624772e-02 -7.02924907e-01 4.48214889e-01
-1.89973578e-01 -1.57743081e-01 1.02845997e-01 -4.73745726e-02
-4.01657790e-01 8.62376571e-01 7.15540826e-01 -1.78216472e-01
-8.96695554e-02 -9.07344148e-02 -6.47101998e-01 5.41393042e-01
8.68161261e-01 -4.95666325e-01 -5.13239801e-01 -2.14621723e-01
5.37671983e-01 7.59606585e-02 5.29927850e-01 -3.61718059e-01
2.31638998e-01 -2.77581424e-01 2.77123243e-01 -5.95362246e-01
1.69325829e-01 -9.73190486e-01 -8.05113316e-02 2.43287265e-01
-5.21107435e-01 1.59804955e-01 -3.21227312e-02 1.09909213e+00
-7.30691552e-01 -1.67993322e-01 1.73629850e-01 6.93171993e-02
-1.01759267e+00 6.27312601e-01 -1.25868067e-01 3.94736946e-01
1.16813040e+00 2.87048727e-01 -2.55865246e-01 -3.43339026e-01
-6.89057827e-01 3.43440086e-01 -1.25220530e-02 6.71074927e-01
6.19993746e-01 -1.57193995e+00 -6.31426036e-01 1.91966623e-01
2.17648983e-01 -4.67882425e-01 -1.41955957e-01 3.89362991e-01
-2.36124024e-01 7.10702956e-01 2.61091907e-03 1.41915262e-01
-6.58959329e-01 7.42832899e-01 3.33096460e-02 -5.43809772e-01
-8.25313330e-01 7.62355983e-01 2.69628525e-01 -1.18841343e-01
2.19763413e-01 -9.14992571e-01 -3.61143142e-01 5.21780550e-01
7.24580765e-01 1.71734929e-01 -6.51408955e-02 -3.61723661e-01
-3.56246144e-01 -3.50737721e-02 -1.10269092e-01 5.01497537e-02
1.80208361e+00 7.47167766e-02 1.20080732e-01 7.25366473e-01
1.27502871e+00 -3.58282886e-02 -9.82935667e-01 -6.83198929e-01
5.01320422e-01 -4.73189443e-01 2.30204821e-01 -4.76976663e-01
-1.23517168e+00 7.63960779e-01 2.49038101e-03 5.82583368e-01
6.16460979e-01 3.13298762e-01 1.24773562e+00 5.31312585e-01
3.22936743e-01 -8.64919305e-01 -7.11415485e-02 5.67148864e-01
6.23562992e-01 -1.16676581e+00 -2.39056528e-01 -3.15846354e-01
-8.49327326e-01 1.17450416e+00 1.45710990e-01 -2.98385352e-01
1.19185793e+00 2.34235108e-01 -3.45107764e-01 -5.42144001e-01
-8.36253941e-01 -5.62240541e-01 4.74339396e-01 2.22778141e-01
3.57576102e-01 2.27619693e-01 -1.84041679e-01 1.07359278e+00
-2.18808994e-01 -1.97062403e-01 2.85813719e-01 7.71915436e-01
-4.54812646e-01 -9.36484337e-01 1.25456527e-01 7.19417334e-01
-5.59101939e-01 -3.55938703e-01 -2.70774394e-01 6.97815359e-01
-3.13607544e-01 5.53867996e-01 5.43723643e-01 -1.34369895e-01
4.32510972e-01 4.44746763e-01 1.00504451e-01 -9.37554300e-01
-5.97131550e-01 -4.25412625e-01 9.04558599e-02 -4.51517552e-01
-2.14320466e-01 -7.27246284e-01 -1.52519226e+00 -2.29524672e-01
-1.24762647e-01 3.44895154e-01 1.87387034e-01 7.76880443e-01
4.56522822e-01 7.49938607e-01 6.22808039e-01 -2.35671997e-01
-7.21436739e-01 -6.60755277e-01 -1.02626753e+00 7.19388902e-01
2.39149675e-01 -8.24238956e-01 -3.84170890e-01 4.90429401e-02] | [8.728378295898438, 8.268895149230957] |
345500b0-6a99-4cbd-965b-318f3c4e5485 | non-autoregressive-text-generation-with-pre | 2102.08220 | null | https://arxiv.org/abs/2102.08220v1 | https://arxiv.org/pdf/2102.08220v1.pdf | Non-Autoregressive Text Generation with Pre-trained Language Models | Non-autoregressive generation (NAG) has recently attracted great attention due to its fast inference speed. However, the generation quality of existing NAG models still lags behind their autoregressive counterparts. In this work, we show that BERT can be employed as the backbone of a NAG model to greatly improve performance. Additionally, we devise mechanisms to alleviate the two common problems of vanilla NAG models: the inflexibility of prefixed output length and the conditional independence of individual token predictions. Lastly, to further increase the speed advantage of the proposed model, we propose a new decoding strategy, ratio-first, for applications where the output lengths can be approximately estimated beforehand. For a comprehensive evaluation, we test the proposed model on three text generation tasks, including text summarization, sentence compression and machine translation. Experimental results show that our model significantly outperforms existing non-autoregressive baselines and achieves competitive performance with many strong autoregressive models. In addition, we also conduct extensive analysis experiments to reveal the effect of each proposed component. | ['Nigel Collier', 'Piji Li', 'Simon Baker', 'David Vandyke', 'Yan Wang', 'Deng Cai', 'Yixuan Su'] | 2021-02-16 | null | https://aclanthology.org/2021.eacl-main.18 | https://aclanthology.org/2021.eacl-main.18.pdf | eacl-2021-2 | ['sentence-compression'] | ['natural-language-processing'] | [ 3.19907099e-01 4.56270464e-02 -2.60289788e-01 -2.74596781e-01
-1.17079914e+00 -3.08662236e-01 6.88713312e-01 -1.71430126e-01
-6.65747747e-02 8.70904565e-01 7.62742102e-01 -6.56783164e-01
3.00243348e-01 -6.17287636e-01 -8.30648541e-01 -6.05921268e-01
2.79390007e-01 3.93145114e-01 -9.26966965e-02 -2.58257896e-01
3.41957301e-01 -8.16588923e-02 -1.01873052e+00 1.60639882e-01
1.39907420e+00 6.91640377e-01 4.43932980e-01 8.41561913e-01
7.12793991e-02 9.49908197e-01 -7.00439453e-01 -5.40560544e-01
-3.15610762e-03 -5.33139884e-01 -5.44027030e-01 -4.32302132e-02
4.37855534e-02 -7.34079003e-01 -4.17244405e-01 7.92304456e-01
6.17498994e-01 1.22808218e-01 8.77881050e-01 -8.52519453e-01
-7.55351365e-01 1.17211413e+00 -7.37114489e-01 1.68731749e-01
3.07350367e-01 4.42361198e-02 1.04350758e+00 -7.81628311e-01
1.66599825e-01 1.47632432e+00 5.08408487e-01 2.97435939e-01
-9.08662081e-01 -6.50339246e-01 4.30518150e-01 4.32181358e-02
-1.03308451e+00 -7.57634699e-01 7.48170137e-01 -6.76040947e-02
1.14864540e+00 1.81587622e-01 1.67888746e-01 1.14105201e+00
3.03292543e-01 1.18085921e+00 7.56878614e-01 -5.61664402e-01
4.85812426e-02 -1.66922018e-01 1.59373313e-01 5.45207560e-01
1.02672435e-01 -2.94725627e-01 -4.18360293e-01 -1.79610565e-01
5.39563000e-01 -2.11346895e-01 -2.99452633e-01 2.76964575e-01
-9.55728650e-01 7.40472615e-01 3.25912014e-02 1.05537869e-01
-4.59473908e-01 3.73908222e-01 3.85836661e-01 5.47472239e-02
7.79377759e-01 3.41324300e-01 -3.62265587e-01 -6.36297822e-01
-1.02574837e+00 3.16189408e-01 6.26722813e-01 1.33809078e+00
3.60475004e-01 5.13828516e-01 -5.17195761e-01 1.04879224e+00
2.83467710e-01 4.21450853e-01 7.86025465e-01 -6.79566622e-01
9.74113107e-01 2.09228262e-01 5.10787638e-03 -7.62722552e-01
-1.05458558e-01 -5.38507044e-01 -1.18206346e+00 -7.84478068e-01
1.08003989e-01 -3.40288252e-01 -7.79660940e-01 1.60288692e+00
-7.91969150e-02 2.57373750e-01 2.03639850e-01 5.80833077e-01
4.49095041e-01 1.08016098e+00 -1.15483284e-01 -5.96663475e-01
1.15684378e+00 -1.34256065e+00 -9.14164126e-01 -2.48786226e-01
6.65140986e-01 -8.71436536e-01 1.03979456e+00 3.00394684e-01
-1.37658072e+00 -4.08077955e-01 -8.73144209e-01 -3.16502213e-01
3.30990434e-01 4.34219956e-01 7.80565798e-01 6.06682658e-01
-8.69556487e-01 5.56503952e-01 -1.07263839e+00 1.19608149e-01
1.25321805e-01 1.30453929e-01 2.65235066e-01 -1.00068472e-01
-1.20270610e+00 5.85248053e-01 3.68538052e-01 3.37612480e-01
-5.40638447e-01 -6.59081399e-01 -8.27336431e-01 3.68077606e-01
2.39219069e-01 -8.29520643e-01 1.60500479e+00 -5.48389435e-01
-1.96910465e+00 -4.76730913e-02 -7.34936297e-01 -8.39609087e-01
5.19698501e-01 -8.36005509e-01 -1.64747447e-01 -4.12715506e-03
-1.21509567e-01 2.25897729e-01 7.49294698e-01 -9.09556568e-01
-4.11149651e-01 -1.35421321e-01 -5.77338748e-02 2.63805062e-01
-4.81522173e-01 -1.90473944e-02 -6.25519931e-01 -1.03339338e+00
-1.88585863e-01 -9.61087883e-01 -3.80014062e-01 -9.41684008e-01
-7.89975524e-01 -3.28503281e-01 3.58598292e-01 -8.08864534e-01
1.76002431e+00 -1.86166704e+00 1.91183597e-01 -8.58877078e-02
-2.62580991e-01 2.73435771e-01 -1.30413994e-01 6.32025719e-01
1.42547816e-01 2.07036331e-01 -2.79880404e-01 -5.70271015e-01
-1.66532192e-02 1.87645674e-01 -7.99640238e-01 -2.54596416e-02
3.03195179e-01 1.09701109e+00 -7.94010937e-01 -4.22673136e-01
-7.38144070e-02 3.68040502e-01 -6.02554381e-01 3.42961580e-01
-2.81900227e-01 2.13118836e-01 -4.55254614e-01 4.11279410e-01
6.04439199e-01 -2.44060218e-01 3.09600025e-01 1.19744875e-01
-2.43494045e-02 7.37322152e-01 -7.76654541e-01 1.47740304e+00
-6.74930334e-01 2.68284768e-01 -4.10353124e-01 -9.21133220e-01
8.49080741e-01 2.87919194e-01 7.16468766e-02 -4.40524548e-01
5.19656204e-02 5.84899746e-02 6.62989244e-02 -2.06908658e-01
1.15342450e+00 1.42477170e-01 -6.33915216e-02 5.64069867e-01
-1.96559608e-01 -1.89015508e-01 3.37276280e-01 3.90114337e-01
8.79723549e-01 5.25011495e-02 4.62197155e-01 -2.75461841e-02
3.74081224e-01 -4.00545597e-01 6.36969388e-01 9.97437835e-01
2.21397504e-01 7.82152712e-01 7.41766155e-01 7.24191889e-02
-9.83436644e-01 -8.18409443e-01 1.38763323e-01 1.18817735e+00
-1.86722815e-01 -7.62912929e-01 -8.72662246e-01 -4.80875492e-01
-4.33897883e-01 1.09801924e+00 -1.99885800e-01 -1.52067184e-01
-9.18166518e-01 -1.07131279e+00 6.40105009e-01 8.76304984e-01
3.41041058e-01 -7.98990190e-01 -1.28655016e-01 3.47059369e-01
-5.94755054e-01 -1.16900098e+00 -6.77237332e-01 -2.00033426e-01
-1.12953210e+00 -4.20916229e-01 -7.89411426e-01 -6.45384431e-01
4.71161336e-01 4.83999461e-01 9.56266999e-01 -7.66882580e-03
4.68772620e-01 -6.95762783e-02 -5.56947947e-01 -5.42602003e-01
-6.30507290e-01 5.76796949e-01 -3.45744332e-03 -2.97329903e-01
-4.17046770e-02 -6.88882709e-01 -5.12472928e-01 8.67920816e-02
-8.78321826e-01 3.61224055e-01 8.66788566e-01 9.26548660e-01
4.97305751e-01 -1.06320545e-01 8.61916542e-01 -1.25812423e+00
1.02706647e+00 -5.62782526e-01 -3.60952199e-01 3.51165324e-01
-6.67181492e-01 2.57714599e-01 9.66346204e-01 -4.25714970e-01
-1.40481448e+00 -2.45444387e-01 -3.92740607e-01 -8.26035291e-02
2.74196476e-01 8.85604143e-01 -1.21276155e-01 7.01062739e-01
1.64670661e-01 5.44447064e-01 -1.81500316e-01 -5.60467005e-01
4.78864938e-01 8.62725258e-01 6.36139750e-01 -5.39679289e-01
6.81139112e-01 -3.89541648e-02 -2.99536318e-01 -6.80738509e-01
-7.49051154e-01 -2.73305863e-01 -2.18320325e-01 3.59709442e-01
2.13814005e-01 -1.22004044e+00 -2.76169181e-01 5.49686849e-01
-1.42712522e+00 -1.94596723e-01 9.40920636e-02 5.39508879e-01
-7.15804935e-01 8.20024133e-01 -9.85802829e-01 -1.11580122e+00
-8.86039972e-01 -1.12669086e+00 1.18122327e+00 1.41786829e-01
-2.10823536e-01 -8.19412768e-01 -5.52225523e-02 4.08020318e-01
4.26932096e-01 -3.25077385e-01 1.15242505e+00 -6.69265330e-01
-4.66268599e-01 -2.60407746e-01 -1.61344066e-01 2.28207409e-01
-4.91177011e-03 8.66884217e-02 -7.86901236e-01 -1.82500258e-01
-8.19038376e-02 -1.53268352e-01 1.11085606e+00 5.01574695e-01
1.26671576e+00 -5.98913968e-01 -9.39192697e-02 4.63405788e-01
9.02338266e-01 5.59963621e-02 9.87540483e-01 8.73196051e-02
6.51573837e-01 1.63676694e-01 6.62222981e-01 6.96412623e-01
6.63475990e-01 5.92182398e-01 1.51134267e-01 1.27594918e-01
8.37405398e-02 -6.52508199e-01 7.03621268e-01 1.70922112e+00
-3.60488534e-01 -7.23617256e-01 -5.70775986e-01 4.26934689e-01
-1.98598123e+00 -8.58116210e-01 -2.60972828e-01 2.15947008e+00
8.51821780e-01 1.79626688e-01 3.72320786e-02 1.14100836e-02
5.91891170e-01 3.07184577e-01 -3.11314344e-01 -6.86520875e-01
-1.45698979e-01 1.88221976e-01 2.74822712e-01 4.73569840e-01
-9.88805830e-01 1.20528984e+00 6.78183794e+00 9.85645056e-01
-7.85778761e-01 -1.23027578e-01 7.62175977e-01 -2.43702321e-03
-3.87505978e-01 1.35986321e-02 -1.11380744e+00 7.16474354e-01
1.23036385e+00 -5.21493196e-01 2.95736879e-01 8.83974075e-01
5.35679400e-01 1.02048412e-01 -9.38236058e-01 6.61115170e-01
1.67527288e-01 -1.11119044e+00 3.50235343e-01 -3.72776389e-02
8.14756215e-01 -8.09200183e-02 1.55645549e-01 6.98880494e-01
3.86405438e-01 -8.83372128e-01 4.90220010e-01 3.43750179e-01
6.41066372e-01 -8.31552029e-01 7.29301333e-01 5.79588473e-01
-1.04123843e+00 -9.27853435e-02 -6.84027672e-01 -2.53049016e-01
5.13218105e-01 6.02754116e-01 -9.46448863e-01 8.38234246e-01
2.24020872e-02 6.28932834e-01 -2.81371832e-01 8.63967240e-01
-4.70454544e-01 1.07968915e+00 -1.44559503e-01 -2.65911967e-02
2.22833410e-01 -2.25769863e-01 5.22741139e-01 1.42016864e+00
5.71518779e-01 9.92127582e-02 -1.15123175e-01 5.14414966e-01
-3.21372360e-01 2.58933425e-01 -4.68663543e-01 -1.51269957e-01
5.44586241e-01 9.35003638e-01 -1.68564096e-01 -5.61097264e-01
-3.02650869e-01 9.29207504e-01 4.30635929e-01 5.04860759e-01
-9.61229026e-01 -3.49339008e-01 2.81524539e-01 -9.77603942e-02
4.28825587e-01 -3.73751551e-01 -1.59958959e-01 -1.37905860e+00
2.07097858e-01 -9.62571979e-01 1.77715853e-01 -7.29732454e-01
-9.87972796e-01 4.87165064e-01 -4.49054688e-02 -1.09632814e+00
-1.09903526e+00 -4.36793566e-02 -8.13286543e-01 8.51321340e-01
-1.79302442e+00 -1.05660331e+00 1.34685844e-01 7.08957464e-02
1.16184092e+00 -3.91202532e-02 6.75040543e-01 1.75203562e-01
-8.42072487e-01 9.29133415e-01 5.26715219e-01 -8.18401277e-02
7.11508453e-01 -1.09047914e+00 8.01379323e-01 1.24305058e+00
4.93047312e-02 8.90977502e-01 7.33229876e-01 -6.34866416e-01
-1.49959826e+00 -1.29898810e+00 1.03273034e+00 -6.05541281e-02
6.30601227e-01 -2.68900275e-01 -1.04669750e+00 8.44876289e-01
2.50271499e-01 -7.43097782e-01 4.96870726e-01 2.43504897e-01
-1.81224421e-01 2.27135010e-02 -3.52019072e-01 6.80157542e-01
8.22101474e-01 -1.47475556e-01 -5.39534509e-01 4.69385028e-01
1.05002344e+00 -5.69003999e-01 -6.96787000e-01 3.74945194e-01
6.96461380e-01 -8.29370439e-01 6.69561148e-01 -5.64070880e-01
9.78931904e-01 9.15947482e-02 1.71175115e-02 -1.55391192e+00
-2.67660171e-01 -1.04494941e+00 -6.19997919e-01 1.51346266e+00
5.02972364e-01 -4.89987522e-01 5.89289188e-01 5.31332374e-01
-4.77069139e-01 -8.00357997e-01 -5.74348211e-01 -8.20792019e-01
2.69346178e-01 -4.86357987e-01 7.22815812e-01 3.75563949e-01
8.91580731e-02 7.11455286e-01 -9.84530151e-01 1.72779858e-02
1.71281502e-01 1.81008935e-01 1.02174747e+00 -6.13086104e-01
-7.07088590e-01 -3.16984892e-01 3.98560017e-02 -1.91938269e+00
2.28016347e-01 -5.72802961e-01 3.37227583e-01 -1.47169185e+00
4.79759783e-01 -1.85372293e-01 -1.18962169e-01 1.29374087e-01
-7.52669990e-01 -1.63376555e-01 1.82732821e-01 2.55981296e-01
-5.99996865e-01 9.44927871e-01 1.06574035e+00 1.24549918e-01
-3.71635497e-01 2.78779566e-01 -1.01167452e+00 6.63931787e-01
9.04727399e-01 -2.59834856e-01 -5.44964671e-01 -7.84464359e-01
1.51588783e-01 3.51731122e-01 -2.70372719e-01 -6.20011806e-01
1.78653196e-01 7.29671866e-02 4.78934459e-02 -8.66365373e-01
2.70026207e-01 -1.87313929e-01 -7.56458193e-02 7.98415765e-02
-5.06622553e-01 3.53724658e-01 -5.21094128e-02 8.22028577e-01
-2.83443928e-01 -3.10468554e-01 4.28202659e-01 1.97097138e-01
-5.32043874e-02 1.44854695e-01 -6.09784842e-01 7.48156086e-02
4.27674770e-01 1.69262290e-01 -4.89631146e-01 -7.09241390e-01
1.92333721e-02 3.97710413e-01 2.02363759e-01 4.35768992e-01
6.57700121e-01 -9.91472900e-01 -8.20181847e-01 -5.59570007e-02
-1.50045663e-01 1.55518249e-01 3.45695227e-01 1.01151538e+00
-3.51523906e-01 7.42964208e-01 3.39674354e-01 -2.51939535e-01
-1.15000844e+00 5.08656800e-01 -3.13229322e-01 -8.97087455e-01
-5.93160212e-01 4.59479332e-01 3.36497933e-01 -2.93259230e-02
1.48023412e-01 -4.25889611e-01 -1.65943608e-01 -3.92196506e-01
7.53735721e-01 5.19289434e-01 3.21378745e-02 -3.93762499e-01
1.09930702e-01 1.27529383e-01 -5.94998121e-01 -3.74563076e-02
1.31664371e+00 -3.65488738e-01 -2.46996582e-02 4.67387110e-01
8.34968448e-01 1.04547694e-01 -9.87056673e-01 -2.71067470e-01
-2.28649303e-01 -2.40471497e-01 -3.49587090e-02 -5.48826873e-01
-7.21211970e-01 1.11307383e+00 -1.49394348e-01 8.18806589e-02
1.18372762e+00 -3.65960121e-01 1.24029136e+00 5.02379954e-01
3.60140540e-02 -9.16973472e-01 -1.95198283e-01 8.00234318e-01
7.68132210e-01 -9.38163698e-01 1.64348066e-01 -5.61749041e-01
-8.26974213e-01 1.04768622e+00 4.18953598e-01 3.14408094e-02
8.93796533e-02 2.33360961e-01 -2.22122729e-01 4.54711169e-01
-1.21979201e+00 1.50812060e-01 1.69321373e-01 4.24919277e-01
8.21171880e-01 5.75128756e-02 -6.16199911e-01 8.65253210e-01
-5.40552199e-01 -5.72783463e-02 7.48426199e-01 5.81996739e-01
-3.69705945e-01 -1.23880029e+00 -1.70206428e-01 5.74505806e-01
-7.74429321e-01 -6.00913405e-01 -2.18131557e-01 5.25507212e-01
-8.55950713e-01 1.12266481e+00 5.94784087e-03 -2.28467107e-01
1.06659636e-01 2.89294515e-02 2.88025677e-01 -4.04720873e-01
-3.90553176e-01 4.67486054e-01 2.62547761e-01 -1.41999602e-01
-1.16041750e-01 -5.11480868e-01 -9.47833776e-01 -3.65317136e-01
-5.76342046e-01 2.89675146e-01 5.27782977e-01 8.94378722e-01
6.05398953e-01 5.92437446e-01 7.49345422e-01 -6.55469716e-01
-1.00391483e+00 -1.41552448e+00 -2.85953820e-01 3.56628597e-02
5.09284325e-02 -3.05914134e-01 -1.40230715e-01 1.87602401e-01] | [11.986348152160645, 9.18514347076416] |
3466cdc9-b0e2-4e1f-82b7-58ab5324ace9 | randgan-randomized-generative-adversarial | 2010.06418 | null | https://arxiv.org/abs/2010.06418v1 | https://arxiv.org/pdf/2010.06418v1.pdf | RANDGAN: Randomized Generative Adversarial Network for Detection of COVID-19 in Chest X-ray | COVID-19 spread across the globe at an immense rate has left healthcare systems incapacitated to diagnose and test patients at the needed rate. Studies have shown promising results for detection of COVID-19 from viral bacterial pneumonia in chest X-rays. Automation of COVID-19 testing using medical images can speed up the testing process of patients where health care systems lack sufficient numbers of the reverse-transcription polymerase chain reaction (RT-PCR) tests. Supervised deep learning models such as convolutional neural networks (CNN) need enough labeled data for all classes to correctly learn the task of detection. Gathering labeled data is a cumbersome task and requires time and resources which could further strain health care systems and radiologists at the early stages of a pandemic such as COVID-19. In this study, we propose a randomized generative adversarial network (RANDGAN) that detects images of an unknown class (COVID-19) from known and labelled classes (Normal and Viral Pneumonia) without the need for labels and training data from the unknown class of images (COVID-19). We used the largest publicly available COVID-19 chest X-ray dataset, COVIDx, which is comprised of Normal, Pneumonia, and COVID-19 images from multiple public databases. In this work, we use transfer learning to segment the lungs in the COVIDx dataset. Next, we show why segmentation of the region of interest (lungs) is vital to correctly learn the task of classification, specifically in datasets that contain images from different resources as it is the case for the COVIDx dataset. Finally, we show improved results in detection of COVID-19 cases using our generative model (RANDGAN) compared to conventional generative adversarial networks (GANs) for anomaly detection in medical images, improving the area under the ROC curve from 0.71 to 0.77. | ['Farzad Khalvati', 'Patrik Rogalla', 'Saman Motamed'] | 2020-10-06 | null | null | null | null | ['covid-19-image-segmentation'] | ['computer-vision'] | [ 5.11651993e-01 -1.25129461e-01 3.47770810e-01 -7.09297359e-02
-6.69421494e-01 -8.93968225e-01 2.23047242e-01 -2.32373141e-02
-5.05868256e-01 8.18697512e-01 -2.92758197e-01 -8.63812268e-01
-5.64561691e-03 -9.07683253e-01 -8.75617266e-01 -8.91214907e-01
8.89889970e-02 1.21342742e+00 -1.67987663e-02 2.86429495e-01
-4.14904147e-01 6.31218493e-01 -9.70419586e-01 4.69725430e-01
6.77255213e-01 6.44524455e-01 1.00415707e-01 1.49257791e+00
1.60571307e-01 6.32587492e-01 -8.24096739e-01 -1.68707997e-01
3.46144825e-01 -7.90775061e-01 -6.86734557e-01 -2.65421987e-01
3.32279533e-01 -6.65995061e-01 -5.86010655e-03 6.47230506e-01
5.69634676e-01 -4.20432448e-01 1.13385916e+00 -1.36204779e+00
-3.83728117e-01 -8.33433867e-02 -3.04167122e-01 4.81585264e-01
-1.74090981e-01 6.19769394e-01 4.45025742e-01 -5.57913005e-01
5.65982401e-01 9.22283113e-01 1.09486997e+00 8.09510171e-01
-8.50141168e-01 -7.15320706e-01 -6.44294143e-01 -3.14038694e-02
-1.18589258e+00 3.56125653e-01 7.45754242e-02 -9.97303128e-01
9.29832160e-01 4.18332934e-01 5.76268911e-01 1.51100731e+00
4.57408130e-01 2.41231546e-01 1.11544287e+00 -1.48151792e-03
6.22638837e-02 -8.58743787e-02 -1.86038166e-01 8.53490770e-01
6.13329709e-01 7.93168321e-02 3.37389171e-01 -3.84892613e-01
9.48917449e-01 7.35758483e-01 -2.42500916e-01 -1.82595123e-02
-1.15467751e+00 1.16174817e+00 2.56116182e-01 2.19787776e-01
-5.86583674e-01 2.57082712e-02 4.21131760e-01 1.72971189e-01
1.62806481e-01 4.97930855e-01 -5.36222041e-01 1.77892640e-01
-7.92832792e-01 -4.46542911e-02 4.29495752e-01 3.53796244e-01
2.50449926e-01 9.09715891e-02 -3.48062992e-01 6.02191627e-01
2.02779248e-01 1.14367390e+00 5.94314635e-01 -4.39329803e-01
2.89295018e-01 6.01720035e-01 4.87607718e-02 -5.40506363e-01
-5.58512866e-01 -4.16359067e-01 -1.11051929e+00 3.77013981e-02
6.06919467e-01 -6.39936268e-01 -1.45474350e+00 1.54225898e+00
3.08720529e-01 3.50019068e-01 2.16109306e-01 6.93261445e-01
6.27087295e-01 6.54941142e-01 -1.14439251e-02 1.53141081e-01
1.50086129e+00 -5.95277011e-01 -2.32141569e-01 -2.08538454e-02
6.55796945e-01 -5.40307045e-01 8.85386705e-01 1.81397229e-01
-4.97324497e-01 -3.90218973e-01 -9.21929479e-01 4.69777256e-01
-4.84141588e-01 -2.41359130e-01 3.03999353e-02 8.87787998e-01
-8.99886847e-01 7.66506195e-02 -9.65771079e-01 -5.37754297e-01
7.53278971e-01 4.76596415e-01 -1.89377561e-01 -3.51975381e-01
-1.00630844e+00 6.77009523e-01 1.80916801e-01 -1.83488578e-01
-1.38993907e+00 -6.68257117e-01 -5.41905344e-01 -2.05334663e-01
2.43225053e-01 -1.06516099e+00 1.03516507e+00 -6.81957126e-01
-6.30254686e-01 9.48409617e-01 2.85179704e-01 -5.06163120e-01
7.50655711e-01 -1.42372265e-01 -3.79968137e-01 4.14453328e-01
4.25057746e-02 6.10217810e-01 1.02041757e+00 -1.11360919e+00
-4.97327626e-01 -3.51261526e-01 -3.15114796e-01 -4.22471374e-01
2.78268129e-01 4.50262390e-02 5.58376312e-02 -7.70036817e-01
-4.43030745e-01 -1.35024524e+00 -1.69922858e-02 -2.63696969e-01
-5.00174344e-01 -1.80936344e-02 1.24319804e+00 -9.88406062e-01
4.62473959e-01 -1.95555270e+00 -4.29639310e-01 2.49743029e-01
5.07283628e-01 8.10007572e-01 -5.73278964e-02 6.33711442e-02
-9.28612351e-02 4.52477515e-01 -6.02147520e-01 2.59489417e-01
-3.49404782e-01 6.74459279e-01 -9.58038121e-02 4.96255815e-01
6.29581690e-01 1.17326725e+00 -8.06270838e-01 -4.11930859e-01
1.03950679e-01 7.22643971e-01 -6.00613117e-01 6.62139475e-01
-2.39474446e-01 9.02044713e-01 -3.54162604e-01 7.14498341e-01
4.81696397e-01 -6.39952898e-01 -8.90188292e-02 2.18479365e-01
6.06261313e-01 -3.20681632e-01 -5.27389348e-01 6.68014586e-01
-3.27052325e-01 4.39297229e-01 -8.66366327e-02 -7.81322002e-01
5.13912797e-01 6.48258567e-01 5.40226460e-01 -2.42444783e-01
2.39385694e-01 2.24615678e-01 3.79102588e-01 -7.30791688e-01
-1.97901025e-01 -4.04671550e-01 1.46628231e-01 7.54286647e-01
-6.23838231e-02 -7.70018026e-02 -1.15525864e-01 -7.83145502e-02
1.41247165e+00 -4.48391557e-01 7.73257092e-02 1.35564417e-01
3.54963183e-01 2.97286630e-01 3.09089452e-01 1.02431846e+00
-2.15220571e-01 8.78679872e-01 3.34548831e-01 -3.43828142e-01
-1.34657860e+00 -1.37218177e+00 -2.25758776e-01 6.73704982e-01
-6.95762098e-01 4.07360345e-01 -9.31192040e-01 -1.06308365e+00
-8.79799351e-02 4.19109553e-01 -8.01106095e-01 4.03348655e-02
-6.96738422e-01 -1.05443490e+00 1.07358074e+00 6.78787708e-01
2.92417377e-01 -1.30789185e+00 -9.20261145e-01 4.46186885e-02
-1.77224010e-01 -9.51246619e-01 -2.49121308e-01 3.23551178e-01
-4.39343661e-01 -1.61390209e+00 -9.52614427e-01 -6.11986816e-01
9.41230595e-01 -2.01577500e-01 1.00870883e+00 2.23395213e-01
-8.40937555e-01 5.71715951e-01 -2.92949140e-01 -9.43415701e-01
-1.03143144e+00 -9.30379853e-02 -9.23002064e-02 -2.07729235e-01
3.26798946e-01 -3.11275106e-02 -7.78866708e-01 4.79151219e-01
-1.12368083e+00 -2.14510202e-01 6.06138468e-01 9.69115555e-01
6.89249456e-01 -4.58435016e-03 6.61097944e-01 -1.15368974e+00
4.14517343e-01 -9.21728373e-01 -4.07738507e-01 9.40112174e-02
-5.29357195e-01 -3.11952055e-01 7.55728900e-01 -3.77418786e-01
-5.94409227e-01 -1.53445080e-01 -3.32581818e-01 -6.73351109e-01
-3.55927408e-01 6.93922071e-03 4.72738713e-01 2.52458453e-01
6.76672459e-01 6.14130795e-02 1.94663763e-01 -2.30914682e-01
-1.94288149e-01 7.69562900e-01 7.56017923e-01 -7.85448104e-02
8.24616432e-01 4.98125106e-01 1.43007696e-01 -6.00749552e-01
-7.18432367e-01 -6.22114003e-01 -4.33136404e-01 -7.47210011e-02
1.62150836e+00 -7.29864240e-01 -4.17160749e-01 6.79383397e-01
-1.02143335e+00 -3.04736853e-01 -4.08081383e-01 5.61667085e-01
-3.14922243e-01 -1.86707731e-02 -6.98866248e-01 -4.10175979e-01
-6.93710268e-01 -1.33238745e+00 9.97610450e-01 -3.81181613e-02
-2.79555589e-01 -1.04888284e+00 3.97867262e-01 4.86431867e-01
5.24987221e-01 7.47807682e-01 1.20828104e+00 -1.28955364e+00
-4.13399100e-01 -3.23818475e-01 -2.31225997e-01 7.39332259e-01
3.06678861e-01 4.40103151e-02 -9.38424945e-01 -4.10717070e-01
2.51084685e-01 -3.89848262e-01 7.38345742e-01 5.40395260e-01
1.18163383e+00 -4.53084797e-01 -2.94824570e-01 7.18980014e-01
1.35881603e+00 7.97237158e-01 5.15555501e-01 -5.26864082e-03
9.94628429e-01 2.83868790e-01 2.09058851e-01 2.14991137e-01
-5.47561757e-02 5.18245399e-02 5.23147106e-01 -5.48902154e-01
-5.20137548e-02 1.49843469e-02 4.70524840e-02 4.18436289e-01
-9.07886624e-02 -7.28665709e-01 -1.37100005e+00 5.66497505e-01
-1.20195711e+00 -8.68974149e-01 -2.58624166e-01 1.84703398e+00
6.71990395e-01 -2.32654080e-01 1.29142970e-01 1.58181638e-01
7.94870019e-01 -4.58021224e-01 -6.84471965e-01 -6.32563293e-01
1.38004720e-01 7.23203540e-01 3.86788577e-01 1.62068784e-01
-1.13429797e+00 1.34139180e-01 6.27821398e+00 2.48625249e-01
-1.38494742e+00 3.91023815e-01 9.41500723e-01 1.13863371e-01
9.92160067e-02 -6.82771206e-01 -4.81888473e-01 6.08271301e-01
1.14681292e+00 4.97327894e-01 1.70265600e-01 7.10952520e-01
-3.83630767e-02 1.78553984e-01 -9.74656224e-01 7.85492599e-01
1.33770511e-01 -1.12717521e+00 -1.49760783e-01 2.28371739e-01
9.27527070e-01 5.56395590e-01 9.60817859e-02 2.13402867e-01
4.89112169e-01 -1.47161257e+00 -5.04128858e-02 3.28958690e-01
1.17486036e+00 -7.19012082e-01 1.36734843e+00 2.01574698e-01
-5.30143082e-01 1.22763857e-01 -6.48180544e-02 6.10608876e-01
-6.95050433e-02 4.07508075e-01 -1.95437455e+00 6.07526302e-02
7.78287709e-01 -3.83433998e-02 -6.43578351e-01 6.96564376e-01
-9.99540016e-02 1.14571142e+00 -2.90691853e-01 1.79433793e-01
3.91444296e-01 6.47584721e-02 4.46654439e-01 1.32942235e+00
4.82992023e-01 6.28135800e-02 -4.04448994e-02 6.96713567e-01
-2.08612457e-01 -1.90564439e-01 -7.93667853e-01 -2.36283764e-01
-1.01943789e-02 9.71372068e-01 -9.11664605e-01 -4.79231745e-01
-2.45709926e-01 8.66181850e-01 -3.85050118e-01 1.28792316e-01
-1.09813035e+00 -2.10284237e-02 3.93240303e-01 4.01866853e-01
5.73334694e-01 1.43567845e-01 -1.10905711e-02 -5.72278678e-01
-3.66755247e-01 -1.14161420e+00 6.53010726e-01 -7.82745361e-01
-1.45398891e+00 6.90892279e-01 -5.89475073e-02 -9.96056855e-01
-6.83854520e-01 -8.49179149e-01 -5.02618730e-01 9.51664209e-01
-1.22073328e+00 -9.98074055e-01 -5.06715000e-01 6.88481390e-01
3.84435296e-01 -2.81802922e-01 9.92042363e-01 1.24449976e-01
-3.36297601e-01 3.73137623e-01 2.72522092e-01 5.90421617e-01
5.83301902e-01 -1.15065038e+00 3.07101130e-01 8.75396430e-01
-2.50151426e-01 3.71488690e-01 2.27410331e-01 -8.72387946e-01
-8.34659934e-01 -1.71145713e+00 5.47868252e-01 -7.19215035e-01
2.76086271e-01 -1.91396236e-01 -9.34461474e-01 7.84939945e-01
8.63283053e-02 9.81072932e-02 9.25031960e-01 -6.81963325e-01
-2.57647604e-01 3.86208683e-01 -1.69102263e+00 1.49529874e-01
5.30972242e-01 -5.21699190e-01 -4.57551569e-01 5.52550673e-01
7.29396701e-01 -2.57214338e-01 -7.27946937e-01 6.17659330e-01
3.31615180e-01 -7.53398180e-01 1.04233754e+00 -7.89500237e-01
4.32193965e-01 -2.13273704e-01 -2.08404548e-02 -1.07339108e+00
8.04731771e-02 -1.57504112e-01 1.63418859e-01 6.80311143e-01
2.83969790e-01 -9.73698854e-01 6.71930552e-01 7.26787597e-02
9.84718651e-02 -7.60695338e-01 -7.18116879e-01 -6.85984552e-01
-1.30792083e-02 -3.30456167e-01 5.91822565e-01 1.03388453e+00
-1.16549790e+00 9.00825337e-02 -1.50700539e-01 3.05611104e-01
1.88995063e-01 -1.70481086e-01 7.23050654e-01 -1.10466301e+00
-5.31819701e-01 1.20752202e-02 -4.51812088e-01 -8.47381130e-02
-4.39306200e-01 -8.41767788e-01 -3.62599380e-02 -1.58785069e+00
3.42458546e-01 -6.11780345e-01 -3.86792779e-01 7.29170084e-01
-2.87791371e-01 7.16518998e-01 -1.80586893e-02 2.32322648e-01
3.49528454e-02 -3.19223136e-01 1.35915375e+00 -2.69169062e-01
1.51346505e-01 1.51197806e-01 -3.69079441e-01 6.92590773e-01
1.02858794e+00 -9.52831447e-01 -5.36884844e-01 -1.34530067e-01
1.14648752e-01 4.60751690e-02 6.96024239e-01 -9.15603101e-01
-4.71716255e-01 -5.27675673e-02 6.89809322e-01 -8.80342364e-01
6.29924908e-02 -8.81907642e-01 4.88728106e-01 1.17833030e+00
1.23385616e-01 5.14388084e-01 1.99550062e-01 5.20688474e-01
1.01565942e-01 -2.64161855e-01 9.26250160e-01 -2.61539578e-01
-1.22823544e-01 3.17452848e-01 -6.90531671e-01 4.68897671e-01
1.27704036e+00 -9.34339464e-02 -5.32371938e-01 -1.48480713e-01
-5.35122097e-01 -1.07807502e-01 4.22455817e-01 2.14404136e-01
5.58995545e-01 -7.10304558e-01 -1.06779921e+00 4.80004579e-01
3.22313644e-02 4.38825846e-01 1.94469586e-01 9.00159419e-01
-1.25057900e+00 3.22120696e-01 -2.74780512e-01 -1.10256171e+00
-1.38535786e+00 7.74670660e-01 5.03012121e-01 -6.35419309e-01
-3.61343145e-01 8.55401754e-01 5.08929312e-01 -3.75417799e-01
-1.99552342e-01 -4.71625596e-01 2.91107930e-02 -2.66221136e-01
3.27305943e-01 2.00074881e-01 2.33078048e-01 -6.32977009e-01
-4.18487072e-01 2.88683087e-01 -8.38719532e-02 3.00821424e-01
1.17872548e+00 5.25970459e-01 -3.28252874e-02 1.47719726e-01
1.32664478e+00 -7.68207237e-02 -8.49788845e-01 2.67201543e-01
-5.09955645e-01 -1.36518970e-01 -3.83972645e-01 -1.07857943e+00
-9.59356427e-01 8.86207223e-01 1.15563881e+00 9.75536779e-02
1.12392831e+00 2.15284556e-01 9.35526073e-01 2.65488476e-01
-1.43979371e-01 -5.16199291e-01 2.55245149e-01 3.01375449e-01
5.42137623e-01 -1.19520032e+00 -3.28300923e-01 -4.56705282e-04
-5.91383040e-01 6.39474690e-01 3.99256557e-01 4.25740890e-02
3.35754067e-01 4.25684720e-01 3.69360864e-01 -5.88070750e-01
-4.30108458e-01 1.93348512e-01 1.37087420e-01 9.99263406e-01
8.65552872e-02 4.53294516e-01 2.26514041e-01 3.01908195e-01
-1.15757473e-01 -1.93871692e-01 3.15317839e-01 1.02946699e+00
-1.26283437e-01 -9.22885120e-01 -8.21587384e-01 1.01476383e+00
-9.78786945e-01 -1.74544320e-01 -4.83979315e-01 8.89178574e-01
6.10418379e-01 7.29743898e-01 3.28147054e-01 -9.66525674e-02
-7.60620087e-02 3.21157247e-01 3.08511734e-01 -7.14458704e-01
-7.23017454e-01 -1.59189358e-01 -1.25609383e-01 -7.65723363e-02
-3.36046547e-01 -5.11382341e-01 -1.34551668e+00 -9.36948806e-02
-4.99671558e-03 -1.37869984e-01 6.80520177e-01 8.12175274e-01
1.86529830e-01 8.69082928e-01 4.58808780e-01 -1.93163961e-01
-3.87979507e-01 -7.80180097e-01 -2.07367048e-01 5.47671020e-01
6.45237207e-01 -3.23036939e-01 -3.51088941e-01 3.65530670e-01] | [15.498934745788574, -1.792251467704773] |
5a728658-5fb0-455e-9043-1668f99cab40 | generating-clarifying-questions-for-query | 2201.09974 | null | https://arxiv.org/abs/2201.09974v1 | https://arxiv.org/pdf/2201.09974v1.pdf | Generating Clarifying Questions for Query Refinement in Source Code Search | In source code search, a common information-seeking strategy involves providing a short initial query with a broad meaning, and then iteratively refining the query using terms gleaned from the results of subsequent searches. This strategy requires programmers to spend time reading search results that are irrelevant to their development needs. In contrast, when programmers seek information from other humans, they typically refine queries by asking and answering clarifying questions. Clarifying questions have been shown to benefit general-purpose search engines, but have not been examined in the context of code search. We present a method for generating natural-sounding clarifying questions using information extracted from function names and comments. Our method outperformed a keyword-based method for single-turn refinement in synthetic studies, and was associated with shorter search duration in human studies. | ['Collin McMillan', 'Zachary Eberhart'] | 2022-01-24 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [ 3.61235470e-01 2.43932068e-01 -1.71405509e-01 -2.27728888e-01
-1.11773252e+00 -1.07611120e+00 3.56661081e-01 7.66984463e-01
-3.03217947e-01 5.44567466e-01 2.87125677e-01 -8.51670682e-01
-1.27561033e-01 -3.48027498e-01 -3.86104465e-01 4.62799698e-01
4.85154480e-01 9.43462253e-02 4.71917301e-01 -3.04864764e-01
1.15483761e+00 -4.10835398e-03 -1.49168003e+00 1.34871140e-01
1.34761381e+00 1.92130566e-01 8.63801599e-01 6.62398458e-01
-6.52350426e-01 8.08152676e-01 -9.08962965e-01 -5.41881382e-01
-1.70746684e-01 -6.74806297e-01 -1.38395667e+00 -3.09647709e-01
3.15027744e-01 -8.11927244e-02 3.02580416e-01 1.05655777e+00
2.17082471e-01 -1.05825163e-01 1.30532056e-01 -1.10465133e+00
-8.34348857e-01 7.64470637e-01 -3.32724959e-01 4.09821898e-01
1.07257104e+00 1.46264900e-02 1.19048238e+00 -9.20210183e-01
6.65643394e-01 8.65479112e-01 6.87345743e-01 5.43800771e-01
-1.32275891e+00 -5.66548586e-01 1.26881963e-02 -1.31352067e-01
-1.16425407e+00 -3.46929014e-01 6.21822715e-01 -7.38651514e-01
1.25972056e+00 4.94848967e-01 5.69099307e-01 5.74488759e-01
2.88237661e-01 2.69721687e-01 4.97365326e-01 -6.83290064e-01
2.95233931e-02 3.68009120e-01 4.57410127e-01 8.05155993e-01
5.55473208e-01 -2.92720944e-01 -3.80638748e-01 -7.81819582e-01
5.70763111e-01 -3.00384581e-01 -3.68626922e-01 -2.31748838e-02
-1.15742517e+00 7.99872041e-01 -2.40997940e-01 5.48536897e-01
-2.30414942e-01 2.43610591e-01 2.65024841e-01 4.90297824e-01
8.97823200e-02 1.47422910e+00 -8.98651958e-01 -5.80752552e-01
-1.02016032e+00 4.07891124e-01 1.35927022e+00 1.31471205e+00
1.03209043e+00 -2.48486340e-01 -2.44139180e-01 8.06397140e-01
3.42112511e-01 2.55860537e-01 5.52635968e-01 -1.30704045e+00
3.45185995e-01 9.81550097e-01 4.82948929e-01 -1.03688657e+00
8.91399086e-02 -4.66944665e-01 1.16036050e-01 1.78936392e-01
4.02015865e-01 -6.03936650e-02 -3.06281269e-01 1.53435934e+00
-1.20534956e-01 -7.56019235e-01 3.88123933e-03 4.29464370e-01
7.89545000e-01 3.13096732e-01 1.60426348e-02 -2.71112084e-01
1.43529570e+00 -9.15759325e-01 -6.19076967e-01 -4.30237502e-01
5.40055871e-01 -9.86452878e-01 1.44663608e+00 3.09865594e-01
-1.09834206e+00 -5.53445101e-01 -8.39206696e-01 -3.61917138e-01
-1.20370649e-01 1.44318491e-01 3.50173265e-01 7.13779628e-01
-1.19262815e+00 1.77359626e-01 -3.27099651e-01 -2.92738318e-01
-1.22198954e-01 -6.88599050e-02 1.46370441e-01 6.23067692e-02
-1.03988600e+00 7.06978977e-01 -4.82443050e-02 -3.51130217e-01
-7.34071374e-01 -1.07611012e+00 -9.83057857e-01 3.71663690e-01
7.30676591e-01 -7.54929423e-01 1.90633023e+00 -7.21287012e-01
-1.01049185e+00 6.37823164e-01 -5.23992956e-01 6.58784807e-02
-1.63911611e-01 -3.40541512e-01 8.15802719e-03 2.50739157e-02
6.66980505e-01 5.26431620e-01 7.01822519e-01 -1.15531123e+00
-4.65646744e-01 -1.35944746e-02 6.94124520e-01 4.94232588e-02
6.97391182e-02 5.58769584e-01 -6.27490759e-01 -6.96617246e-01
-3.30911428e-02 -6.62936747e-01 -4.21832353e-01 -5.00630476e-02
-2.79480398e-01 -2.87038356e-01 3.99892122e-01 -9.73464787e-01
1.97336531e+00 -1.79875410e+00 -1.12539902e-01 -8.19014478e-03
5.39248168e-01 4.76368284e-03 -1.33660600e-01 8.02614570e-01
-1.68747246e-01 8.05050254e-01 -2.02200383e-01 4.32364553e-01
1.79180443e-01 -4.70250309e-01 -1.90014049e-01 -2.81352073e-01
-2.81525198e-02 9.62595403e-01 -1.28639615e+00 -6.23989761e-01
-3.79481137e-01 -1.43757448e-01 -8.97982121e-01 2.71712273e-01
-5.67745805e-01 -3.58764380e-02 -4.74202573e-01 6.82554781e-01
1.52354902e-02 -6.26704454e-01 4.01604082e-03 2.76825935e-01
-2.58938938e-01 6.24224484e-01 -7.24037766e-01 1.76776528e+00
-9.12418783e-01 8.96906435e-01 6.58974946e-02 -4.22576338e-01
9.10720527e-01 2.59432942e-01 -2.59708483e-02 -5.07887721e-01
-5.13621032e-01 2.31886372e-01 -9.04885754e-02 -8.94677520e-01
5.82485020e-01 1.91176936e-01 -2.55884439e-01 1.00933886e+00
-3.29547346e-01 -7.67408311e-01 5.56212485e-01 4.49987739e-01
1.41871464e+00 6.69207945e-02 6.72546923e-01 -3.92548531e-01
5.50365746e-01 4.86856192e-01 1.98979616e-01 1.00126183e+00
3.09193701e-01 2.76976377e-01 6.15543127e-01 -1.51728526e-01
-6.78297281e-01 -5.87157190e-01 2.13736713e-01 1.21457469e+00
7.80833587e-02 -1.21741688e+00 -9.51703906e-01 -8.67059767e-01
-2.84377277e-01 1.20195806e+00 -3.60722750e-01 1.62682272e-02
-4.45022017e-01 2.22524703e-01 5.40347993e-01 1.30929172e-01
1.72586396e-01 -1.14616513e+00 -1.31471825e+00 2.33674586e-01
-3.94808263e-01 -4.68074739e-01 -1.23673284e+00 1.13963033e-03
-8.02665114e-01 -1.31514490e+00 -7.89005399e-01 -8.99435818e-01
8.71245503e-01 1.47797272e-01 1.68244517e+00 9.39020693e-01
-2.05473512e-01 7.85464704e-01 -4.38676476e-01 -2.83300519e-01
-6.80304885e-01 2.47314304e-01 -8.58373463e-01 -1.03868604e+00
3.54171425e-01 -2.25987509e-01 -3.10777754e-01 3.30573499e-01
-1.10168099e+00 1.01519786e-01 6.05914950e-01 8.51886094e-01
-2.68338993e-02 -2.61797369e-01 6.01078749e-01 -1.07423532e+00
1.47001159e+00 -5.47341228e-01 -7.48910129e-01 6.07187033e-01
-1.13340735e+00 5.39588928e-01 3.63415718e-01 -2.97719032e-01
-1.10650241e+00 -1.10300377e-01 -9.39221606e-02 2.92423695e-01
-1.27880216e-01 1.12032700e+00 2.63210922e-01 -1.63922664e-02
1.10145688e+00 2.54767805e-01 -2.27058008e-01 -5.41054487e-01
6.15699030e-02 4.26188439e-01 1.93982333e-01 -9.79659557e-01
8.12519550e-01 -5.48624456e-01 -6.66060567e-01 -6.87833846e-01
-2.79639870e-01 -4.86357361e-01 -1.68234125e-01 1.26408692e-02
5.54272830e-01 -2.80258745e-01 -2.64483452e-01 -5.95835187e-02
-1.54499817e+00 -1.56341314e-01 -2.16711998e-01 -5.05924225e-02
-2.61694938e-01 7.23786056e-01 -2.55682796e-01 -5.80223501e-01
-2.73885846e-01 -1.43429279e+00 9.67448413e-01 3.87927264e-01
-1.07804942e+00 -7.95971394e-01 3.58340949e-01 2.97734380e-01
6.74469829e-01 -2.66278982e-01 1.58572817e+00 -8.20033133e-01
-7.03975856e-01 2.92485859e-02 -1.08481735e-01 -1.93595201e-01
5.44233859e-01 3.34854349e-02 -3.04082483e-01 1.98820725e-01
-5.61096286e-03 -1.00234665e-01 4.99810092e-02 -8.48006383e-02
1.05968380e+00 -7.10323095e-01 -4.17482495e-01 3.44659947e-02
1.35514116e+00 5.35295486e-01 3.35309893e-01 2.65991718e-01
1.52815208e-01 8.41531992e-01 5.86913407e-01 3.19521338e-01
2.81161904e-01 4.38713074e-01 2.94775758e-02 1.91111743e-01
8.50639269e-02 -3.64433616e-01 8.01279850e-04 5.44283986e-01
4.60175812e-01 -3.26247662e-02 -1.37380612e+00 9.56491232e-01
-1.46276188e+00 -6.53468013e-01 -7.20357448e-02 1.86000872e+00
1.35171878e+00 -2.62172110e-02 -3.04986369e-02 -2.15049461e-01
5.35618067e-01 -2.17546642e-01 -5.65964818e-01 -3.76187027e-01
6.29444420e-01 3.90563220e-01 -1.30565330e-01 7.60778964e-01
-3.44932601e-02 6.44258916e-01 7.31464195e+00 3.98272961e-01
-6.45335913e-01 -1.76120594e-01 4.05556053e-01 4.13694948e-01
-1.17646646e+00 6.15498006e-01 -5.70466697e-01 3.16416234e-01
7.45844901e-01 -9.29869235e-01 5.04440904e-01 1.13453138e+00
9.43721831e-02 -5.54985881e-01 -1.36194003e+00 7.38162637e-01
-5.40889539e-02 -1.39904857e+00 -5.29883988e-02 -4.06439066e-01
5.23557723e-01 -5.21874011e-01 -3.50016505e-01 3.60293657e-01
3.38739067e-01 -9.70940053e-01 6.42345965e-01 5.46705544e-01
6.86012208e-01 -3.06449711e-01 3.57132763e-01 7.06639826e-01
-1.09086466e+00 3.92333046e-02 -2.64469627e-03 -3.30283016e-01
-1.38653815e-01 2.63054878e-01 -1.07574189e+00 1.46749988e-01
6.29231751e-01 9.48506072e-02 -9.63778913e-01 1.24771643e+00
-2.01761305e-01 5.55984080e-01 3.34331654e-02 -7.56676853e-01
-1.79250747e-01 1.03544764e-01 5.50519884e-01 1.48250175e+00
5.14999092e-01 1.27209231e-01 -3.93190682e-02 1.47411609e+00
1.66032895e-01 2.13974670e-01 -5.06658673e-01 -5.17187059e-01
6.58364296e-01 9.87126172e-01 -6.46588206e-01 -3.67684990e-01
-6.30793035e-01 6.72046542e-01 -1.06953494e-01 6.07492030e-01
-4.45224196e-01 -1.12604666e+00 2.03029782e-01 3.70616794e-01
1.67373970e-01 -2.12262571e-01 -2.59085655e-01 -9.94090021e-01
2.48076186e-01 -1.44819844e+00 1.20237865e-01 -1.16524470e+00
-7.02293932e-01 6.05858684e-01 4.73866910e-01 -8.75087917e-01
-7.06154287e-01 -4.04993109e-02 -5.78734875e-01 1.33890843e+00
-1.12843835e+00 -2.91232646e-01 -3.18175942e-01 1.99461766e-02
1.08529317e+00 1.88032493e-01 7.31314719e-01 -4.32060584e-02
3.95719521e-02 2.82455295e-01 -6.81656718e-01 -1.00843288e-01
5.68292558e-01 -1.33162987e+00 6.56119585e-01 8.54917109e-01
1.36856422e-01 1.64943898e+00 8.68769705e-01 -9.97674882e-01
-1.31078792e+00 -3.60474169e-01 1.44081569e+00 -5.60510099e-01
6.46375179e-01 -3.36099565e-02 -1.04994822e+00 2.77723402e-01
7.37935841e-01 -6.88389421e-01 7.74827957e-01 5.02755679e-02
-3.29707921e-01 2.79503584e-01 -8.55140567e-01 6.27598584e-01
9.16600049e-01 -8.25318158e-01 -1.10898840e+00 1.19521484e-01
1.02534652e+00 -3.34374517e-01 -4.95753944e-01 -4.30268720e-02
3.76447320e-01 -6.81011438e-01 5.72080016e-01 -4.55933392e-01
5.60340941e-01 -4.72547621e-01 1.06584728e-01 -1.13579559e+00
-2.25702435e-01 -1.18715858e+00 3.73472959e-01 1.01038873e+00
1.11519694e+00 -4.16837245e-01 4.85331923e-01 1.18162382e+00
-1.77633747e-01 -5.83868146e-01 -1.88253686e-01 -3.62907588e-01
-1.22995116e-01 -2.89991498e-01 6.07631385e-01 8.28952730e-01
6.88769221e-01 6.54894292e-01 2.22578853e-01 -1.54593125e-01
1.75084785e-01 2.73566574e-01 7.42711127e-01 -1.28892541e+00
-4.42764461e-01 -4.50054109e-01 5.28858900e-01 -1.44265151e+00
5.13132736e-02 -7.18204319e-01 4.43576366e-01 -1.61184621e+00
3.26532871e-01 -2.69483566e-01 3.74182403e-01 4.63033915e-01
-4.83444124e-01 -4.99032855e-01 -1.00975901e-01 2.52555102e-01
-4.53031152e-01 -2.37600226e-02 1.01131690e+00 -1.06730670e-01
-3.12303901e-01 1.88278228e-01 -1.44249499e+00 5.64884961e-01
5.28938651e-01 -6.47468030e-01 -8.52751315e-01 -3.80537033e-01
1.13708282e+00 6.31777108e-01 1.12571113e-01 -6.56648099e-01
6.59289181e-01 -4.53078538e-01 -3.10816407e-01 -3.80618215e-01
-3.70659918e-01 -7.84026682e-01 4.27068949e-01 5.92854381e-01
-8.44882250e-01 8.07229280e-01 3.65977317e-01 2.57227004e-01
-2.73949355e-01 -1.34763300e+00 6.16053566e-02 -4.99093145e-01
-6.12665176e-01 -1.96421683e-01 -9.01700795e-01 5.09041250e-01
5.49812675e-01 -4.67485964e-01 -3.58233005e-01 -5.41647077e-01
-2.25750759e-01 1.21050790e-01 8.78111839e-01 3.45863491e-01
6.84849858e-01 -8.42081487e-01 -4.13106173e-01 -5.92528246e-02
2.62462437e-01 -2.75720567e-01 -4.72574443e-01 3.10541421e-01
-6.29226804e-01 7.10650265e-01 1.46992639e-01 -3.52282286e-01
-1.24214196e+00 5.16597927e-01 1.01979990e-02 -1.82738826e-01
-1.54625401e-01 7.77089179e-01 2.42236540e-01 -3.57318074e-01
2.10417077e-01 -8.87072384e-01 -2.07770288e-01 -1.75542593e-01
5.65501750e-01 2.09415078e-01 4.99446057e-02 9.79627892e-02
-3.95683467e-01 5.08648634e-01 1.00146011e-02 -5.26613891e-01
8.98849368e-01 -1.71743840e-01 -2.59061635e-01 4.16538060e-01
8.11897159e-01 6.91085756e-01 -5.90883970e-01 -1.74206600e-01
6.53018355e-01 -5.57280838e-01 -3.68498832e-01 -1.14158082e+00
-4.50239450e-01 3.62499416e-01 -2.15032369e-01 7.65108228e-01
1.07775438e+00 3.05521488e-01 4.44897920e-01 9.38829362e-01
4.18484002e-01 -7.10583329e-01 4.70802367e-01 5.93278587e-01
1.15132523e+00 -1.02677429e+00 -2.27742627e-01 -5.62755287e-01
-3.99520427e-01 1.30038893e+00 8.33456159e-01 5.07147908e-01
3.76354635e-01 3.76404852e-01 5.89551851e-02 -5.72745085e-01
-9.54045117e-01 1.88100666e-01 3.49271238e-01 5.27870238e-01
8.53052080e-01 -7.86418974e-01 -5.97735345e-01 7.37175703e-01
-2.21304506e-01 1.72469720e-01 7.47182965e-01 1.31644785e+00
-5.21615267e-01 -1.35877132e+00 -3.35054070e-01 5.80582500e-01
-7.09713995e-01 -8.31504762e-01 -8.28745723e-01 6.11556768e-01
-3.53089184e-01 1.22099292e+00 -4.27215397e-01 -1.37778550e-01
2.55886227e-01 1.08841233e-01 1.96338087e-01 -1.42167962e+00
-1.06827044e+00 -2.33016819e-01 4.96488452e-01 -4.91246790e-01
-1.56308994e-01 -6.77719533e-01 -9.91802692e-01 2.40190849e-01
-7.26688027e-01 1.00344193e+00 5.81665576e-01 8.71018708e-01
5.75965941e-01 2.94503748e-01 1.30133078e-01 -1.44702271e-01
-4.50177461e-01 -8.22666109e-01 2.53465354e-01 1.17934952e-02
4.85983878e-01 -2.78674603e-01 -3.15954238e-01 5.40039778e-01] | [7.617844581604004, 7.998866558074951] |
ec7bf9cb-7a5a-49f2-b37c-726cf09c438d | understanding-the-impact-of-edge-cases-from | 2204.12402 | null | https://arxiv.org/abs/2204.12402v1 | https://arxiv.org/pdf/2204.12402v1.pdf | Understanding the Impact of Edge Cases from Occluded Pedestrians for ML Systems | Machine learning (ML)-enabled approaches are considered a substantial support technique of detection and classification of obstacles of traffic participants in self-driving vehicles. Major breakthroughs have been demonstrated the past few years, even covering complete end-to-end data processing chain from sensory inputs through perception and planning to vehicle control of acceleration, breaking and steering. YOLO (you-only-look-once) is a state-of-the-art perception neural network (NN) architecture providing object detection and classification through bounding box estimations on camera images. As the NN is trained on well annotated images, in this paper we study the variations of confidence levels from the NN when tested on hand-crafted occlusion added to a test set. We compare regular pedestrian detection to upper and lower body detection. Our findings show that the two NN using only partial information perform similarly well like the NN for the full body when the full body NN's performance is 0.75 or better. Furthermore and as expected, the network, which is only trained on the lower half body is least prone to disturbances from occlusions of the upper half and vice versa. | ['Stig Ursing', 'Christian Berger', 'Jens Henriksson'] | 2022-04-26 | null | null | null | null | ['body-detection'] | ['computer-vision'] | [-9.39485580e-02 3.48217815e-01 -3.65445465e-01 -4.79903579e-01
-3.30721796e-01 -1.30259171e-01 4.78775084e-01 -4.63964455e-02
-7.01019168e-01 5.03168344e-01 -4.86097746e-02 -4.34291989e-01
2.23020330e-01 -5.77589571e-01 -9.16936100e-01 -6.29075170e-01
-8.10331199e-03 1.99824199e-01 7.28458345e-01 -4.33202714e-01
-8.84883329e-02 5.44680357e-01 -2.07211900e+00 3.49221975e-01
4.05869007e-01 1.24085677e+00 9.23512354e-02 1.03085124e+00
4.00455713e-01 6.05758607e-01 -3.73244196e-01 -5.39264202e-01
3.72828633e-01 1.03056543e-01 -1.77689835e-01 4.83929664e-02
9.85067964e-01 -2.84314245e-01 -3.51134658e-01 8.46274436e-01
6.75186872e-01 8.79043862e-02 6.57197654e-01 -1.54961097e+00
-8.07377398e-02 -9.20363814e-02 -2.76053160e-01 3.91762972e-01
2.40401745e-01 4.99603152e-01 2.44419575e-01 -9.95576560e-01
3.78858209e-01 1.35087240e+00 9.98616695e-01 6.14469290e-01
-1.02167237e+00 -4.99533415e-01 1.05679505e-01 5.30689836e-01
-1.28396225e+00 -8.29051316e-01 6.03057861e-01 -6.10114574e-01
1.14331269e+00 4.08554077e-02 2.41641909e-01 1.03109109e+00
4.89288121e-01 5.31781673e-01 8.08009863e-01 -4.30235744e-01
2.67157942e-01 5.19723475e-01 1.41262278e-01 6.65071070e-01
4.57382679e-01 5.66080570e-01 -4.41892028e-01 5.08071840e-01
2.64462322e-01 -4.63454098e-01 3.10297340e-01 -5.38634062e-01
-7.39985585e-01 7.07655072e-01 4.02898759e-01 -2.42151514e-01
-3.27699333e-01 1.78864256e-01 7.38141477e-01 1.74658783e-02
8.72587189e-02 -3.27601194e-01 -4.89387959e-01 -1.20700523e-01
-8.41014087e-01 2.88511097e-01 6.70168042e-01 1.00409257e+00
5.18441260e-01 2.34945372e-01 -2.07732052e-01 6.66613400e-01
3.25528264e-01 5.78708172e-01 2.35864475e-01 -8.64052534e-01
8.16831410e-01 3.95085990e-01 2.44006917e-01 -9.01196182e-01
-9.01039362e-01 -4.52347130e-01 -7.09774077e-01 9.82629359e-01
9.77670550e-01 -3.09718013e-01 -1.00093925e+00 1.42720628e+00
5.09820640e-01 -2.38606110e-01 2.20024616e-01 9.98838305e-01
9.77213502e-01 3.64652395e-01 3.19491297e-01 2.51757950e-01
1.49769187e+00 -9.52843785e-01 -5.33605576e-01 -7.72835970e-01
4.91835892e-01 -5.41116059e-01 7.49816477e-01 7.37624049e-01
-8.79765272e-01 -1.46831751e+00 -1.46462762e+00 -6.45807609e-02
-7.83627450e-01 5.41978002e-01 2.83835292e-01 1.24738407e+00
-8.05363536e-01 5.98229945e-01 -7.29368806e-01 -3.77780378e-01
4.59142238e-01 4.24257964e-01 -6.09577715e-01 3.08427233e-02
-1.05917454e+00 1.37845933e+00 2.21357465e-01 4.22602087e-01
-9.32387948e-01 -3.68785352e-01 -9.72077370e-01 -2.59612381e-01
4.44367319e-01 -4.83910918e-01 1.07697487e+00 -8.01368594e-01
-1.35088634e+00 8.42812598e-01 -2.85222471e-01 -8.81271482e-01
9.33475792e-01 -4.47771788e-01 -5.20879924e-01 -1.15539826e-01
2.86390055e-02 8.54028821e-01 8.97592008e-01 -1.34835148e+00
-9.07107174e-01 -3.47847492e-01 -1.85653001e-01 -9.01224241e-02
3.48165065e-01 -1.32812947e-01 -2.36462116e-01 -2.25182083e-02
4.91763018e-02 -6.90904498e-01 -1.89293534e-01 3.55436295e-01
-4.50747281e-01 -2.96238452e-01 7.91744053e-01 -8.02666008e-01
9.39928412e-01 -1.97916007e+00 -3.62980872e-01 2.17664391e-01
5.98037504e-02 5.28597474e-01 1.83189481e-01 1.55705675e-01
-2.14049175e-01 -3.51907730e-01 2.70144910e-01 -4.53608692e-01
2.00418904e-01 3.60689551e-01 1.99195758e-01 8.09365690e-01
2.18773499e-01 5.17603517e-01 -5.56217015e-01 -5.12725413e-01
7.23237455e-01 5.85428655e-01 -2.63102889e-01 1.27761528e-01
2.53964573e-01 5.55508621e-02 7.12837209e-04 7.28549600e-01
7.47361004e-01 6.91352427e-01 -3.16768050e-01 -3.42553258e-01
-3.30655932e-01 -1.90038785e-01 -1.43274617e+00 1.06537223e+00
-4.23030883e-01 1.12321484e+00 1.37965366e-01 -9.57311213e-01
1.00426245e+00 2.42826208e-01 -1.95626736e-01 -1.12770367e+00
2.21323743e-01 1.09919071e-01 3.84378970e-01 -9.69717920e-01
3.25126976e-01 2.54345797e-02 1.20889157e-01 -5.65341353e-01
-8.12972859e-02 3.49536270e-01 4.24128711e-01 -2.26062149e-01
8.34121644e-01 3.05272073e-01 2.82992870e-01 -9.96148065e-02
8.73241961e-01 -3.72370929e-02 4.41320896e-01 9.41412866e-01
-8.76842678e-01 5.31120956e-01 4.48346525e-01 -6.35875463e-01
-1.33554864e+00 -1.17094672e+00 -2.60852695e-01 1.20123184e+00
1.00405797e-01 -5.13086654e-02 -9.93726850e-01 -5.12309074e-01
2.60295302e-01 7.71253407e-01 -7.13591754e-01 -2.42151380e-01
-6.84024572e-01 -3.38329941e-01 5.19883394e-01 1.05845284e+00
5.20760179e-01 -9.57274318e-01 -1.12180543e+00 3.03084403e-01
2.90738583e-01 -1.35380507e+00 1.43796295e-01 2.67258078e-01
-4.89086777e-01 -1.10335147e+00 -4.81891185e-01 -4.13803309e-01
3.07774365e-01 1.06226802e-02 1.02023292e+00 -1.50868833e-01
-4.03279603e-01 3.59966874e-01 -4.81542721e-02 -8.83743286e-01
-5.26223421e-01 -4.69553113e-01 4.29968625e-01 -1.32747153e-02
4.27021325e-01 -1.21530674e-01 -7.60052025e-01 6.40496552e-01
-3.04536708e-02 -1.28400832e-01 5.98676264e-01 5.31713068e-01
2.37849861e-01 4.04050201e-02 3.86474937e-01 -2.55614042e-01
2.25396454e-01 -1.15498997e-01 -7.04328358e-01 -1.34339511e-01
-3.88768852e-01 -1.92194372e-01 5.93754709e-01 -2.98389226e-01
-9.13106859e-01 4.23859447e-01 -4.87793863e-01 -3.23989630e-01
-9.24777508e-01 -1.72685787e-01 -3.02794248e-01 -6.03637770e-02
9.00995910e-01 -1.07849747e-01 1.76299989e-01 -3.15666705e-01
2.53055632e-01 6.33147657e-01 9.74517226e-01 -2.17973754e-01
5.26570737e-01 6.16321862e-01 4.18968946e-01 -1.18187118e+00
-6.00626707e-01 -6.19199932e-01 -1.10807943e+00 -8.80080044e-01
1.11834824e+00 -7.48823822e-01 -1.10514450e+00 3.03826958e-01
-1.13941872e+00 -2.85178125e-01 -1.12152368e-01 5.74388802e-01
-6.12069070e-01 1.57413512e-01 7.76210101e-03 -1.36282980e+00
-6.61490411e-02 -1.10964715e+00 1.04812539e+00 3.73924762e-01
-2.62985807e-02 -6.72974825e-01 -2.47680217e-01 4.69943523e-01
1.90062851e-01 3.74237865e-01 4.33417767e-01 -4.00349587e-01
7.70542165e-03 -6.79059029e-01 -3.68037403e-01 7.77182043e-01
-5.43466568e-01 8.12215880e-02 -1.31316710e+00 -2.02275321e-01
-3.14566076e-01 -5.54009937e-02 9.84110892e-01 7.30709434e-01
7.14832842e-01 1.03786699e-01 -4.34657604e-01 1.63800180e-01
1.22331154e+00 2.49387011e-01 7.55027175e-01 5.92899621e-01
5.71756423e-01 9.90493357e-01 7.39960253e-01 1.50292352e-01
4.19371456e-01 7.72575021e-01 8.47670734e-01 -2.59341717e-01
-5.05772650e-01 -4.89477962e-02 5.94409585e-01 -1.53083995e-01
-2.73236185e-01 -1.29282206e-01 -8.79678190e-01 3.52240115e-01
-1.69251800e+00 -9.67936039e-01 -6.20675027e-01 2.34800959e+00
1.71574056e-02 9.71404374e-01 6.47085011e-01 6.42988443e-01
6.08643711e-01 -4.05218273e-01 -4.05347198e-01 -8.13932598e-01
1.83754712e-01 -2.45184571e-01 8.98059428e-01 5.06011903e-01
-1.47334814e+00 6.02052629e-01 6.20387983e+00 7.50860512e-01
-8.23136091e-01 1.26113445e-01 5.88683665e-01 2.06644312e-02
8.51565003e-01 -4.22515213e-01 -1.41302729e+00 4.91979986e-01
1.15915596e+00 8.11236680e-01 -6.48107529e-02 1.33495748e+00
7.01857507e-01 -7.24795640e-01 -1.25916100e+00 8.43620598e-01
9.25672203e-02 -1.03143632e+00 -5.24506748e-01 -3.29651982e-02
1.89589590e-01 3.12980339e-02 -9.49723348e-02 5.58962286e-01
-2.72565663e-01 -1.01892030e+00 1.22432220e+00 6.47463083e-01
5.45395613e-01 -9.40425158e-01 1.17709529e+00 6.39770031e-01
-1.29357898e+00 -3.64935338e-01 -6.15980148e-01 -3.55481684e-01
1.09662071e-01 3.60415101e-01 -8.63650918e-01 2.99541831e-01
7.33682811e-01 1.86004579e-01 -8.85122895e-01 1.13964760e+00
-5.86025938e-02 6.40541553e-01 -3.37407410e-01 -1.70456972e-02
3.46271902e-01 1.84734866e-01 6.23805940e-01 1.65041196e+00
-1.59518376e-01 -4.00194854e-01 3.18377972e-01 3.45385790e-01
6.71293020e-01 -2.20222414e-01 -6.83471322e-01 8.25190127e-01
2.64835879e-02 1.11623812e+00 -6.39240980e-01 -5.31902552e-01
-5.67990482e-01 4.50315684e-01 1.40116110e-01 2.76134044e-01
-9.88182366e-01 -3.16992134e-01 6.83651328e-01 5.07561207e-01
4.25502121e-01 -1.47584245e-01 -3.62908274e-01 -3.18598896e-01
1.16687372e-01 -3.80400538e-01 2.32187197e-01 -8.99240971e-01
-8.76943111e-01 5.29305100e-01 3.14616114e-01 -1.24798393e+00
-2.25463539e-01 -1.42012334e+00 -6.14780664e-01 7.90806115e-01
-1.28739226e+00 -1.04517186e+00 -4.31352109e-01 2.28754237e-01
7.25117922e-01 -2.46709198e-01 4.08720195e-01 5.24389446e-01
-7.12757409e-01 8.26712787e-01 -2.42290914e-01 2.02728644e-01
3.68550003e-01 -1.05721426e+00 9.07357261e-02 9.94484305e-01
-3.75155956e-01 2.21650526e-02 1.23944569e+00 -5.44756174e-01
-1.11888754e+00 -9.50250626e-01 7.79314578e-01 -6.32412016e-01
1.75950229e-01 -4.12642211e-01 -5.07402062e-01 2.74576604e-01
6.33161664e-02 2.30036855e-01 1.93477169e-01 -8.63201544e-02
-1.14552140e-01 -5.76606512e-01 -1.22040749e+00 4.16445255e-01
7.54743397e-01 -1.10568553e-01 -6.20554447e-01 1.77855432e-01
-2.21013045e-03 -5.69382489e-01 -4.11619395e-01 4.30338472e-01
9.95526254e-01 -1.34153402e+00 1.30553484e+00 -5.98141730e-01
1.99637264e-01 -4.87344265e-01 -3.02089632e-01 -7.16038585e-01
-9.23106149e-02 -1.41091226e-02 -1.01196386e-01 9.39448357e-01
3.93224627e-01 -4.27132308e-01 1.10048687e+00 3.25979829e-01
-4.87936407e-01 -7.34471619e-01 -1.47214007e+00 -9.97904956e-01
-1.16127007e-01 -1.11892581e+00 -1.77165821e-01 5.22216782e-02
-1.98576093e-01 1.52344927e-01 -4.18945551e-01 3.85324746e-01
7.32114792e-01 -5.83951712e-01 1.11381769e+00 -1.05915225e+00
-9.41226408e-02 -4.49899465e-01 -9.50806677e-01 -8.33609343e-01
-2.71596551e-01 -3.48101556e-01 3.19333404e-01 -1.37912261e+00
-2.49683470e-01 5.08792400e-02 5.27879968e-02 2.46386632e-01
1.10355891e-01 4.40323770e-01 1.05215065e-01 -4.36150283e-01
-4.39082205e-01 -2.67771967e-02 1.10930097e+00 6.97665513e-02
2.52470076e-02 5.24184465e-01 -5.41526079e-02 1.08777773e+00
7.74700463e-01 -3.29166532e-01 -1.02772348e-01 1.45135850e-01
1.45155713e-01 -3.96966711e-02 9.19270456e-01 -1.67608213e+00
3.26833129e-01 2.32338473e-01 8.56725693e-01 -1.13814294e+00
6.73748076e-01 -9.16891158e-01 -1.80526108e-01 7.17612624e-01
-1.08124845e-01 1.22758579e-02 5.99631786e-01 4.60403055e-01
1.89832956e-01 -4.27853823e-01 9.48957562e-01 -5.98136485e-02
-1.09924626e+00 -2.59589285e-01 -8.44096899e-01 -3.04172039e-01
1.04977918e+00 -9.15519655e-01 -8.91960561e-02 -3.48468333e-01
-1.04338717e+00 1.04328625e-01 -1.98337778e-01 5.13881385e-01
5.86856604e-01 -1.07248569e+00 -6.16992295e-01 4.45100486e-01
6.96657747e-02 -2.55304575e-01 1.78372607e-01 8.51860940e-01
-3.76459748e-01 8.48727584e-01 -4.60738540e-01 -8.19044173e-01
-1.58804417e+00 7.17349827e-01 6.40588880e-01 1.87031835e-01
-4.98215914e-01 7.18196213e-01 3.21334675e-02 -2.53589690e-01
6.47491992e-01 -4.77839798e-01 -4.02871221e-01 1.12338262e-02
4.83278751e-01 9.83111143e-01 2.20818013e-01 -1.06785202e+00
-4.89936858e-01 6.80447936e-01 2.39801228e-01 4.76066060e-02
7.03850210e-01 -2.02827409e-01 6.83545649e-01 3.32129478e-01
8.30641091e-01 -2.72633374e-01 -1.63082337e+00 3.10149819e-01
-1.66364491e-01 -1.89430013e-01 1.11172274e-01 -8.52998853e-01
-6.58540964e-01 1.25955296e+00 1.24250138e+00 1.90630451e-01
5.78057706e-01 -1.28400281e-01 1.39437109e-01 3.41147810e-01
7.71998093e-02 -1.44920301e+00 -9.70829800e-02 3.06510270e-01
8.69534671e-01 -1.56370938e+00 8.36872235e-02 -4.57395792e-01
-6.53364003e-01 1.23908830e+00 9.78161454e-01 -1.37911782e-01
5.42278528e-01 2.24024117e-01 4.80617285e-02 9.70783178e-03
-5.20794034e-01 -6.25870109e-01 6.96169734e-01 1.21633887e+00
7.66065791e-02 1.66569233e-01 4.39261086e-02 4.71002847e-01
-2.26845682e-01 -2.64330953e-01 2.28895664e-01 7.76917160e-01
-8.65665436e-01 -5.78294873e-01 -7.93978631e-01 4.96767938e-01
-1.11961618e-01 2.71007091e-01 -2.52879248e-03 1.31335402e+00
9.69871879e-01 1.19124007e+00 1.95265859e-01 -3.39870006e-01
8.24524999e-01 6.87825307e-02 3.19126755e-01 -2.68563360e-01
-3.28454137e-01 -4.24193926e-02 5.88761568e-01 -7.98250973e-01
-1.54373229e-01 -8.52659404e-01 -1.14888585e+00 -1.60321087e-01
-2.29454800e-01 -3.41943532e-01 1.03040993e+00 1.04612541e+00
-4.49281149e-02 6.56125724e-01 1.13678716e-01 -1.49019337e+00
-4.66841519e-01 -1.02443576e+00 -2.59065211e-01 5.09637557e-02
6.18963182e-01 -1.04936588e+00 -1.97054788e-01 1.11456923e-01] | [8.082756996154785, -0.7932628393173218] |
b012cd99-ce73-470b-bbb7-b37144dd9265 | transfering-hierarchical-structure-with-dual | 2201.11981 | null | https://arxiv.org/abs/2201.11981v2 | https://arxiv.org/pdf/2201.11981v2.pdf | Transfering Hierarchical Structure with Dual Meta Imitation Learning | Hierarchical Imitation Learning (HIL) is an effective way for robots to learn sub-skills from long-horizon unsegmented demonstrations. However, the learned hierarchical structure lacks the mechanism to transfer across multi-tasks or to new tasks, which makes them have to learn from scratch when facing a new situation. Transferring and reorganizing modular sub-skills require fast adaptation ability of the whole hierarchical structure. In this work, we propose Dual Meta Imitation Learning (DMIL), a hierarchical meta imitation learning method where the high-level network and sub-skills are iteratively meta-learned with model-agnostic meta-learning. DMIL uses the likelihood of state-action pairs from each sub-skill as the supervision for the high-level network adaptation, and use the adapted high-level network to determine different data set for each sub-skill adaptation. We theoretically prove the convergence of the iterative training process of DMIL and establish the connection between DMIL and Expectation-Maximization algorithm. Empirically, we achieve state-of-the-art few-shot imitation learning performance on the Meta-world \cite{metaworld} benchmark and competitive results on long-horizon tasks of Kitchen environments. | ['Feng Chen', 'Yizhou Jiang', 'Chongkai Gao'] | 2022-01-28 | null | null | null | null | ['few-shot-imitation-learning'] | ['methodology'] | [-4.59090434e-02 1.33197710e-01 -7.87836388e-02 -1.21683612e-01
-3.85638624e-01 -1.81863248e-01 5.38371444e-01 -4.21344280e-01
-6.01245344e-01 9.39098120e-01 -5.71388640e-02 1.81498230e-01
-2.38963634e-01 -5.84131777e-01 -1.17241490e+00 -7.59116173e-01
-3.06245595e-01 5.49506426e-01 6.40694022e-01 -3.35756928e-01
1.10720076e-01 1.35473594e-01 -1.56106746e+00 3.59560788e-01
1.00578618e+00 5.08236349e-01 1.05319524e+00 7.80004799e-01
8.12661275e-02 1.12307250e+00 -4.54977632e-01 8.03115070e-02
3.24097782e-01 -6.29453421e-01 -9.84229624e-01 8.50340202e-02
-9.86563861e-02 -4.89077032e-01 -6.94051862e-01 1.08254623e+00
3.41386259e-01 5.83425283e-01 7.53955781e-01 -1.30953872e+00
-2.77149737e-01 6.95200503e-01 -2.80616075e-01 2.02879816e-01
8.98597017e-02 5.49031556e-01 5.31715512e-01 -6.30387127e-01
6.31604254e-01 1.38854373e+00 4.64085579e-01 8.50693166e-01
-8.44687521e-01 -5.93979418e-01 2.84436345e-01 4.99930352e-01
-8.54482949e-01 -1.17428461e-02 4.01828140e-01 -3.40656072e-01
1.31095457e+00 -6.65630877e-01 9.00396347e-01 1.33455706e+00
7.38516867e-01 1.02075899e+00 1.25690055e+00 -1.48308516e-01
3.58398765e-01 -2.45671749e-01 -3.30672443e-01 1.06659853e+00
-8.53715390e-02 5.50668657e-01 -5.80227256e-01 7.75235444e-02
1.10581517e+00 2.17506975e-01 1.33894831e-01 -6.65559232e-01
-1.57671130e+00 4.60766256e-01 6.37163043e-01 1.62441343e-01
-7.14345276e-01 6.00577414e-01 5.95965564e-01 8.86973381e-01
-9.93245766e-02 5.61004758e-01 -6.63373172e-01 -2.45972827e-01
-3.44447911e-01 -7.77625814e-02 7.28807092e-01 1.19909668e+00
9.23142493e-01 2.40113392e-01 -1.14905853e-02 7.62784064e-01
1.14753552e-01 2.44358405e-01 1.04033852e+00 -1.39496684e+00
5.52259088e-01 5.03006637e-01 3.11431382e-02 -1.46195918e-01
-4.50819463e-01 -2.84122616e-01 -7.48258591e-01 4.94420439e-01
5.88409379e-02 -3.47392857e-01 -1.04269218e+00 1.85264456e+00
3.86794269e-01 4.45332408e-01 4.13722545e-01 5.68132579e-01
4.38107908e-01 8.31661820e-01 -9.25527811e-02 -2.89562672e-01
7.98226595e-01 -1.74601960e+00 -2.37589464e-01 -3.54017824e-01
5.41168451e-01 -9.77494717e-02 9.83404458e-01 3.54344696e-01
-1.11537397e+00 -1.02069485e+00 -1.20728481e+00 2.82089025e-01
-1.83061868e-01 -1.61392286e-01 2.68567353e-01 -2.84324020e-01
-8.16722333e-01 1.03989756e+00 -9.38428521e-01 -3.65810782e-01
1.99516580e-01 3.86329859e-01 -3.93724263e-01 -1.14098452e-02
-1.09233296e+00 1.22232926e+00 1.00993359e+00 -1.37405917e-01
-2.01182580e+00 -3.08517843e-01 -8.13959062e-01 2.76408754e-02
6.22913301e-01 -1.00145733e+00 1.61320674e+00 -1.03261936e+00
-2.10367632e+00 3.85782361e-01 2.55591422e-01 -4.89333302e-01
4.76276457e-01 -1.39452264e-01 6.76906062e-03 1.69889539e-01
2.46793717e-01 9.31720734e-01 1.35502374e+00 -1.21124291e+00
-1.14148223e+00 -8.77865851e-02 1.88785776e-01 6.41285956e-01
-6.26000017e-02 -6.01428092e-01 -1.71289101e-01 -4.00123090e-01
-2.09585261e-02 -1.09702444e+00 -3.51923376e-01 -4.74410951e-01
9.93015319e-02 -4.39506799e-01 6.79481626e-01 -3.96758616e-01
6.04289949e-01 -2.06700397e+00 8.87851357e-01 -2.90284783e-01
5.88429421e-02 1.83726534e-01 -3.96138549e-01 5.87523758e-01
2.91680872e-01 -5.62626660e-01 -1.62636220e-01 -2.31647313e-01
3.65472808e-02 5.63983977e-01 -9.37822834e-02 1.42410919e-01
-2.44862422e-01 1.14526176e+00 -1.52347612e+00 -5.05282104e-01
3.09341848e-01 -5.37160002e-02 -2.74425119e-01 5.40267766e-01
-3.62437040e-01 7.38979340e-01 -4.59172845e-01 3.89572591e-01
-4.36536893e-02 -1.88265324e-01 1.58865675e-01 1.72579065e-01
5.54791912e-02 1.97755173e-02 -7.84421146e-01 2.25937033e+00
-8.15801799e-01 3.40175211e-01 -5.24030114e-03 -1.11562204e+00
7.62322605e-01 2.71081716e-01 3.37459177e-01 -5.89783370e-01
9.53902528e-02 2.08244741e-01 2.84395516e-01 -7.65710652e-01
2.59932935e-01 -1.58896029e-01 -2.77451783e-01 4.44036812e-01
7.28434265e-01 -5.24411082e-01 8.99674222e-02 -9.69101265e-02
1.18008399e+00 6.69182777e-01 5.63156664e-01 2.25423370e-02
3.96865755e-01 1.79588541e-01 5.48353016e-01 1.07044470e+00
-3.91812652e-01 -4.99168634e-02 7.78589398e-02 -4.39968616e-01
-1.20746291e+00 -1.27751720e+00 5.85399270e-01 1.45736575e+00
3.31002921e-01 1.92715768e-02 -6.87016129e-01 -8.83807600e-01
3.85361426e-02 5.73639691e-01 -7.40017176e-01 -5.84828258e-01
-6.30596042e-01 -2.07841098e-01 3.22138846e-01 5.51961243e-01
9.53399479e-01 -1.81595170e+00 -1.17085123e+00 6.02602005e-01
2.28644386e-02 -8.90196085e-01 -3.61846268e-01 4.43740577e-01
-1.22444296e+00 -1.03885627e+00 -7.95838654e-01 -1.22117734e+00
6.45210385e-01 3.84212673e-01 7.43771255e-01 -1.50546551e-01
-5.73126562e-02 5.92019260e-01 -3.34564775e-01 -2.35268578e-01
-7.73259878e-01 1.95585310e-01 4.51545835e-01 -4.45414275e-01
-2.10809946e-01 -1.00670445e+00 -3.89461964e-01 5.27310133e-01
-5.72163343e-01 3.39007854e-01 1.15583014e+00 1.32747674e+00
5.75624526e-01 8.58511254e-02 8.45349371e-01 -2.53035426e-01
6.68275297e-01 -4.44384396e-01 -5.76929986e-01 3.55719298e-01
-4.72856462e-01 2.72112161e-01 8.31180692e-01 -9.41159129e-01
-1.14495814e+00 9.02136564e-02 1.97994977e-01 -7.23585665e-01
-4.84957285e-02 4.00724858e-01 1.75817415e-01 4.58408669e-02
5.05628884e-01 7.25392759e-01 1.53224662e-01 -3.45240414e-01
5.03639400e-01 4.58843142e-01 7.35114694e-01 -6.95195138e-01
5.75753808e-01 3.56416583e-01 -5.93081862e-02 -6.24151587e-01
-8.55281949e-01 -4.44259912e-01 -1.01280677e+00 -3.42336923e-01
8.22026193e-01 -9.24437404e-01 -7.77360797e-01 9.08336759e-01
-9.87814724e-01 -1.19861531e+00 -5.29573441e-01 5.37365079e-01
-1.34363759e+00 3.19203675e-01 -8.55296314e-01 -4.58309799e-01
-2.55462050e-01 -1.27538931e+00 7.25954413e-01 2.52493024e-01
2.36207709e-01 -1.01648760e+00 3.45883936e-01 -1.76127516e-02
1.56537876e-01 -1.17992023e-02 7.18073249e-01 -2.88571477e-01
-3.89996201e-01 9.53623131e-02 1.52084321e-01 5.16114652e-01
2.07797319e-01 -5.56838810e-01 -5.44793963e-01 -8.06067646e-01
1.90803587e-01 -1.01626909e+00 8.29598367e-01 2.26929948e-01
9.90622222e-01 -3.14165682e-01 -3.30807090e-01 4.02124584e-01
1.08762813e+00 1.52661145e-01 5.01234770e-01 6.22416794e-01
6.49926126e-01 4.84616309e-01 8.89973521e-01 3.87659013e-01
4.49012280e-01 5.16813695e-01 6.36875272e-01 6.02663815e-01
-1.50074875e-02 -5.97883344e-01 8.62497568e-01 1.18882060e+00
-1.12304702e-01 3.71977359e-01 -5.17694414e-01 5.23166239e-01
-2.31101632e+00 -1.04823554e+00 5.74649036e-01 1.96702278e+00
7.75926054e-01 3.57081681e-01 1.85821399e-01 -5.57650924e-01
6.44847631e-01 -1.11859655e-02 -1.13399565e+00 -2.77858227e-01
3.32843155e-01 -2.01912329e-01 2.30154082e-01 3.39232773e-01
-8.83980334e-01 1.26642847e+00 6.18266916e+00 8.88093233e-01
-7.66865551e-01 3.95769328e-01 -8.02903622e-02 -7.23550990e-02
5.37030101e-01 -4.58845086e-02 -5.72061300e-01 4.58601505e-01
6.93015397e-01 -2.99536377e-01 9.33789432e-01 1.12925553e+00
-4.83121201e-02 -2.44087726e-01 -1.33255005e+00 7.65510917e-01
-1.42197743e-01 -9.98719275e-01 -1.76252816e-02 -5.68148755e-02
1.14957726e+00 5.91667414e-01 -5.60049862e-02 1.08979118e+00
7.94185936e-01 -7.53560781e-01 5.96268058e-01 5.29522240e-01
5.16988933e-01 -5.72528005e-01 5.08690238e-01 1.08437979e+00
-1.25829339e+00 -6.47851944e-01 -8.21071625e-01 -2.11009249e-01
1.10477000e-01 -2.96243370e-01 -8.04648757e-01 5.21414220e-01
6.97467506e-01 8.32191348e-01 -2.72157699e-01 9.07337427e-01
-6.57149315e-01 1.69451579e-01 -2.34210696e-02 -2.13951170e-01
6.39960051e-01 -8.20424333e-02 6.77331924e-01 8.73556733e-01
3.17866415e-01 6.19486310e-02 4.35595512e-01 6.19855881e-01
-4.04854864e-02 -4.71775055e-01 -7.77656138e-01 9.86176580e-02
4.32985127e-01 1.15491819e+00 -3.20048600e-01 -7.18201697e-01
-1.81418985e-01 1.31823146e+00 8.99202526e-01 3.14553291e-01
-7.82279849e-01 -1.93128109e-01 3.51893276e-01 -5.61252594e-01
3.87339920e-01 -4.07743990e-01 3.41407895e-01 -1.09225774e+00
-2.72123396e-01 -9.84010875e-01 3.74361068e-01 -1.05279386e+00
-1.21015084e+00 3.00026357e-01 2.63831586e-01 -1.56673408e+00
-6.88829482e-01 -4.42181736e-01 -7.66459882e-01 4.01338488e-01
-1.33743978e+00 -9.75928724e-01 -3.94092649e-01 6.87879145e-01
1.19622529e+00 -5.98026514e-01 6.22047544e-01 -3.82464588e-01
-3.82904202e-01 3.92884433e-01 3.91394436e-01 -1.58887058e-01
4.38286662e-01 -1.28032172e+00 4.18495566e-01 4.27291751e-01
-8.21332186e-02 2.76907504e-01 6.42127395e-01 -7.47817814e-01
-1.35383201e+00 -1.01600873e+00 1.93265323e-02 -2.84123808e-01
8.54406416e-01 -1.50613442e-01 -8.51610363e-01 8.40887308e-01
2.69149095e-01 -3.31055582e-01 -7.32971355e-02 -2.31309801e-01
-9.68132392e-02 6.32054135e-02 -9.95458841e-01 7.76704788e-01
1.37918460e+00 -4.42105800e-01 -1.43071926e+00 2.66854674e-01
9.34536338e-01 -4.34859574e-01 -9.46638286e-01 4.81498659e-01
6.65716529e-01 -8.36404145e-01 8.84713471e-01 -8.58102798e-01
5.63620448e-01 -1.27657890e-01 1.49915561e-01 -1.99604547e+00
-5.66245317e-01 -5.91888607e-01 -5.30970037e-01 4.36015487e-01
2.05499321e-01 -6.96794748e-01 5.69557071e-01 -3.58527452e-01
-6.35405660e-01 -7.94640183e-01 -1.04725897e+00 -1.45614123e+00
1.53281957e-01 9.88297164e-02 3.20519120e-01 7.32601941e-01
6.30821347e-01 4.73789990e-01 -5.57006180e-01 3.91323529e-02
7.31730998e-01 1.92060575e-01 1.10973740e+00 -8.60695064e-01
-7.85538375e-01 -4.01250988e-01 -1.35419875e-01 -1.33399534e+00
5.15146196e-01 -8.39460313e-01 4.64012235e-01 -1.75350595e+00
3.42274398e-01 -1.27000585e-01 -5.03872633e-01 5.34347534e-01
-1.40007302e-01 -4.47789013e-01 2.97799110e-01 5.85271716e-01
-1.10619283e+00 1.03393245e+00 1.86829317e+00 -2.66564012e-01
-4.80790824e-01 2.56478786e-02 5.14765643e-02 8.36609900e-01
9.84752297e-01 -6.45627797e-01 -7.08398879e-01 -4.34483379e-01
-1.25887096e-01 1.94023386e-01 2.34407261e-01 -1.30197191e+00
4.78608459e-01 -3.64528030e-01 3.46322685e-01 -3.42465132e-01
5.00909626e-01 -6.63870215e-01 -1.60464540e-01 1.11978483e+00
-2.67048061e-01 2.23743752e-01 1.15050890e-01 8.90512228e-01
-2.13929806e-02 -5.82897484e-01 8.16266477e-01 -7.63841093e-01
-1.30378985e+00 3.15483987e-01 -6.24110401e-01 9.54180285e-02
1.19675851e+00 -3.28553766e-01 -3.12248498e-01 -1.80144846e-01
-9.49263990e-01 8.21269989e-01 5.31192064e-01 6.89501345e-01
7.63975024e-01 -1.20328903e+00 -4.68784124e-01 -1.31693184e-01
8.29171687e-02 1.60712212e-01 3.45288783e-01 7.69863963e-01
-1.41949207e-01 -4.45542745e-02 -8.70621979e-01 -6.69863999e-01
-8.07372749e-01 8.74570191e-01 6.24846280e-01 -4.45071608e-01
-8.59807789e-01 7.31692076e-01 3.36272210e-01 -7.63341248e-01
2.98738033e-01 -2.72780329e-01 -1.54550374e-02 -3.34741265e-01
1.86740935e-01 6.98639870e-01 -5.95898747e-01 -1.98878855e-01
9.24270451e-02 3.78461093e-01 -1.25751421e-01 -3.78019214e-01
1.26489794e+00 -6.08494356e-02 -2.70649930e-03 9.49291468e-01
9.75513339e-01 -1.04009032e+00 -1.98270547e+00 -3.44468802e-01
-4.05955408e-03 -1.58329811e-02 -4.94445801e-01 -6.54560089e-01
-4.16457862e-01 9.14346278e-01 3.37144792e-01 -2.89818704e-01
6.13901854e-01 7.16466606e-02 7.68739343e-01 1.21678460e+00
1.08477736e+00 -1.67447615e+00 1.08086801e+00 9.76061285e-01
9.83307123e-01 -1.19005609e+00 -8.27090293e-02 2.83762097e-01
-7.54195035e-01 8.97486627e-01 1.08090925e+00 -3.75256091e-01
3.53880525e-01 -1.39199540e-01 -2.95431525e-01 -1.43045425e-01
-9.94052052e-01 -1.00000232e-01 -3.32876435e-03 6.99011207e-01
-4.94214594e-01 1.05553325e-02 1.37837842e-01 2.34426245e-01
1.14002861e-02 9.69670266e-02 4.17704642e-01 1.27618384e+00
-1.05552828e+00 -8.03339779e-01 3.04675903e-02 4.46468234e-01
2.82372087e-01 3.50091934e-01 -9.60275903e-02 8.39262784e-01
1.30736055e-02 6.15393996e-01 -2.30600551e-01 -5.06507993e-01
3.16895127e-01 1.57013386e-01 1.08681667e+00 -9.26116288e-01
-4.81752843e-01 -2.13739321e-01 -2.06420034e-01 -6.35848224e-01
-3.41318250e-01 -5.47971129e-01 -1.54631352e+00 -4.98714037e-02
-1.55292237e-02 2.06009582e-01 3.78004014e-01 1.31755376e+00
2.83868492e-01 8.27511787e-01 7.14807272e-01 -1.29983294e+00
-1.16671264e+00 -1.34768283e+00 -4.33874816e-01 2.75645614e-01
2.89481103e-01 -9.84570086e-01 -2.63787925e-01 -7.65966102e-02] | [4.353198051452637, 1.1291435956954956] |
3346b7c2-98c7-4ac7-9227-d444ec567830 | d2a-a-dataset-built-for-ai-based | 2102.07995 | null | https://arxiv.org/abs/2102.07995v1 | https://arxiv.org/pdf/2102.07995v1.pdf | D2A: A Dataset Built for AI-Based Vulnerability Detection Methods Using Differential Analysis | Static analysis tools are widely used for vulnerability detection as they understand programs with complex behavior and millions of lines of code. Despite their popularity, static analysis tools are known to generate an excess of false positives. The recent ability of Machine Learning models to understand programming languages opens new possibilities when applied to static analysis. However, existing datasets to train models for vulnerability identification suffer from multiple limitations such as limited bug context, limited size, and synthetic and unrealistic source code. We propose D2A, a differential analysis based approach to label issues reported by static analysis tools. The D2A dataset is built by analyzing version pairs from multiple open source projects. From each project, we select bug fixing commits and we run static analysis on the versions before and after such commits. If some issues detected in a before-commit version disappear in the corresponding after-commit version, they are very likely to be real bugs that got fixed by the commit. We use D2A to generate a large labeled dataset to train models for vulnerability identification. We show that the dataset can be used to build a classifier to identify possible false alarms among the issues reported by static analysis, hence helping developers prioritize and investigate potential true positives first. | ['Zhong Su', 'Alessandro Morari', 'Jim Laredo', 'Bo Yang', 'Edward Epstein', 'Luca Buratti', 'Burn Lewis', 'Saurabh Pujar', 'Yunhui Zheng'] | 2021-02-16 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [-1.05842628e-01 -7.01267570e-02 -2.94800580e-01 -2.25087389e-01
-1.04832017e+00 -9.87077832e-01 -6.28674924e-02 7.05988705e-01
3.03373247e-01 4.34567630e-01 -2.02559501e-01 -7.58622408e-01
1.02569237e-01 -8.37305188e-01 -8.30392003e-01 4.63980203e-03
-3.71319890e-01 -1.12892069e-01 7.91576862e-01 -9.91945788e-02
6.95914090e-01 -7.82517195e-02 -1.57439208e+00 7.05043077e-01
9.42067862e-01 2.85306811e-01 1.43004522e-01 7.78550386e-01
-3.54092091e-01 6.95015132e-01 -1.04065752e+00 -5.23586094e-01
1.89626917e-01 -4.41234916e-01 -9.38062072e-01 -5.82971513e-01
4.64901388e-01 -1.85901105e-01 3.89707536e-01 1.39154649e+00
6.35567978e-02 -6.38906002e-01 2.01393757e-02 -1.45590067e+00
-3.91998976e-01 9.36909497e-01 -8.79393697e-01 5.37513614e-01
8.34397912e-01 1.28801286e-01 1.33120966e+00 -6.59387887e-01
6.80854440e-01 1.01066768e+00 1.11370361e+00 6.77626729e-01
-1.44261217e+00 -3.34071606e-01 -7.82900751e-02 8.88051316e-02
-9.75841403e-01 -1.62419662e-01 6.59103036e-01 -1.06753278e+00
1.44564378e+00 3.90909970e-01 4.10039961e-01 1.02440572e+00
5.09429693e-01 2.99721748e-01 9.05053079e-01 -6.14039779e-01
2.95356303e-01 2.95553952e-01 7.74016082e-01 6.91811919e-01
5.93356550e-01 -1.22433446e-01 -1.28764451e-01 -1.16677821e+00
-3.13228071e-02 -1.16179161e-01 -1.88285768e-01 1.24525905e-01
-9.33484972e-01 8.96645367e-01 -4.14417088e-02 4.17130500e-01
1.21472385e-02 2.22703740e-01 7.83831120e-01 7.55524755e-01
1.75348639e-01 8.81710470e-01 -9.20924067e-01 -3.13219577e-01
-7.45971918e-01 3.53951156e-01 8.95774662e-01 6.72254503e-01
1.01279938e+00 -2.40966260e-01 2.97248811e-01 5.90101600e-01
4.36530828e-01 -5.27171930e-03 5.83256125e-01 -5.07746577e-01
6.54896140e-01 1.33170724e+00 -1.19058162e-01 -1.13771975e+00
-1.22687854e-01 1.73204951e-02 3.14638019e-01 5.19281924e-01
1.92472175e-01 8.91152620e-02 -4.04387295e-01 1.56950271e+00
2.46040784e-02 -7.90383741e-02 -1.41353801e-01 5.63352145e-02
3.20935488e-01 2.25938693e-01 -1.58718064e-01 1.47949606e-02
9.94441152e-01 -4.44703013e-01 -1.38561994e-01 -3.37128252e-01
1.32366395e+00 -7.41764188e-01 1.20958519e+00 3.73224765e-01
-6.08090639e-01 -2.24914208e-01 -1.04152536e+00 6.07926130e-01
-3.75348508e-01 -5.66606186e-02 4.80215371e-01 1.11073613e+00
-1.12514293e+00 6.10262275e-01 -1.01356280e+00 -1.97488263e-01
2.17891797e-01 8.61342475e-02 -3.20275515e-01 2.18227342e-01
-8.07857037e-01 6.28694832e-01 2.93387383e-01 -4.58013535e-01
-1.06379473e+00 -9.84497249e-01 -7.44009793e-01 7.10783750e-02
4.23880547e-01 -3.71196494e-02 1.19193959e+00 -1.02277303e+00
-3.26871783e-01 7.74700999e-01 -4.08283696e-02 -1.85238466e-01
1.32203713e-01 -2.24686548e-01 -4.18636382e-01 -2.92158931e-01
6.15353584e-01 -2.93301374e-01 6.76551878e-01 -9.21340942e-01
-6.17993474e-01 -3.06733698e-01 4.79193866e-01 -1.03205431e+00
-4.52827603e-01 7.19244838e-01 1.63721815e-01 -4.03263867e-01
-2.39153087e-01 -6.70709670e-01 -1.83714464e-01 -5.72403014e-01
-6.05636537e-01 1.94345657e-02 5.93963981e-01 -9.62501287e-01
1.91544878e+00 -2.00530243e+00 -8.76093134e-02 1.30955383e-01
3.42895836e-01 1.58383757e-01 -2.86150753e-01 4.94320869e-01
-5.21550894e-01 7.43491113e-01 -3.80316526e-01 2.24889472e-01
-2.60266095e-01 -2.73238301e-01 -4.87769008e-01 1.00704752e-01
4.23866481e-01 4.83526468e-01 -9.88183916e-01 -1.80391416e-01
-4.63365734e-01 -1.63001642e-01 -8.32984805e-01 2.78374165e-01
-4.57560718e-01 -1.57905638e-01 -4.53784287e-01 9.30943131e-01
4.75332499e-01 -1.37581686e-02 8.12314525e-02 4.00447577e-01
-2.89903998e-01 6.48061216e-01 -8.69480491e-01 1.04037225e+00
-4.60551441e-01 7.57364392e-01 -3.50088537e-01 -7.49456584e-01
9.81359184e-01 3.21207136e-01 -8.83812904e-02 -2.76539296e-01
-2.19122425e-01 5.58597684e-01 2.25204900e-01 -9.00393903e-01
1.13110408e-01 3.54840577e-01 -4.25262809e-01 8.68306577e-01
-4.74736840e-02 2.79710680e-01 2.95015842e-01 1.09939374e-01
1.94321930e+00 -2.16077697e-02 4.11655337e-01 -2.58934647e-01
6.44101918e-01 1.94983125e-01 8.61999452e-01 6.79918349e-01
-1.20553508e-01 3.61208677e-01 1.38720751e+00 -7.92963266e-01
-9.74719942e-01 -6.96320355e-01 1.45388961e-01 9.53656316e-01
-3.45391095e-01 -8.92621756e-01 -9.34991181e-01 -1.37123072e+00
-6.15341291e-02 7.09943712e-01 -7.08342433e-01 -4.68283117e-01
-7.10758686e-01 -7.38034844e-01 7.04397082e-01 5.43552220e-01
-2.76669949e-01 -8.92006814e-01 -1.02281606e+00 2.89567769e-01
-3.40988562e-02 -4.07710463e-01 -2.01400161e-01 1.46049604e-01
-5.72825909e-01 -1.74540031e+00 -3.29612158e-02 -5.59429109e-01
6.15436018e-01 -4.24745074e-03 1.33214211e+00 8.02892029e-01
-8.02129805e-01 2.48894542e-01 -4.99242574e-01 -2.69440144e-01
-1.14483631e+00 -1.31006330e-01 -1.22785039e-01 -3.71531725e-01
4.18341935e-01 -4.32942420e-01 6.16670474e-02 3.33810031e-01
-7.11604595e-01 -6.57953620e-01 1.99445367e-01 6.13494456e-01
1.56441137e-01 3.55651677e-02 7.23249912e-01 -1.01150370e+00
6.60585940e-01 -9.00307834e-01 -8.75044346e-01 5.62883556e-01
-4.55801606e-01 2.40905792e-04 6.12086177e-01 -5.08664668e-01
-9.10263062e-01 -8.28860253e-02 -2.42682293e-01 8.50610435e-02
-1.60466135e-01 9.64231789e-01 -1.13694713e-01 -2.88517147e-01
1.03137577e+00 -1.98744029e-01 -1.89893648e-01 -6.24724805e-01
-2.09414661e-01 4.79610771e-01 1.14437237e-01 -5.98602533e-01
8.46444666e-01 -2.19381928e-01 -5.36705196e-01 -5.29983401e-01
-3.70629191e-01 -2.56040514e-01 -5.40114939e-01 -2.17868939e-01
3.92743886e-01 -4.35709625e-01 -1.10751979e-01 4.11434293e-01
-1.38448191e+00 -1.81362465e-01 2.35812873e-01 -8.30530524e-02
-9.26868469e-02 6.00863338e-01 -2.47359380e-01 -7.56732583e-01
-1.09898731e-01 -1.49675643e+00 6.86638594e-01 6.63546994e-02
-7.60628402e-01 -9.80298162e-01 7.44642317e-01 -7.16461539e-02
3.48794848e-01 6.00950658e-01 1.61512148e+00 -7.63531089e-01
-4.89885002e-01 -5.17868042e-01 1.43528089e-01 4.40803319e-01
2.63694286e-01 7.46972084e-01 -8.58766198e-01 -2.39391655e-01
2.62334459e-02 -1.46349430e-01 4.16605562e-01 -1.49859145e-01
7.88695514e-01 -3.70487034e-01 -6.83666766e-01 2.46105716e-01
1.70744181e+00 2.40434378e-01 4.96060789e-01 8.51561129e-01
5.29857457e-01 7.23804414e-01 6.10334456e-01 3.10000420e-01
1.16768852e-01 4.65570718e-01 6.46064818e-01 5.12485862e-01
3.46921295e-01 2.07695827e-01 8.28848004e-01 4.95552421e-01
1.89608976e-01 5.58275133e-02 -1.64202189e+00 9.85561788e-01
-1.47974718e+00 -7.41190970e-01 -6.33118033e-01 2.47256184e+00
9.92877960e-01 3.57618958e-01 3.06773931e-01 3.31424087e-01
8.22901070e-01 -2.17179865e-01 -1.32259384e-01 -6.50991559e-01
2.69265383e-01 7.03159347e-02 -7.38142431e-02 2.92927235e-01
-9.28117156e-01 4.07525659e-01 6.28109837e+00 2.45675236e-01
-9.73624766e-01 1.51979834e-01 3.24173152e-01 1.63605615e-01
-6.64571822e-01 5.43743253e-01 -9.37037110e-01 5.54536045e-01
1.34895730e+00 -3.77096593e-01 -5.21356463e-02 1.35906065e+00
-2.77046859e-01 -2.31588721e-01 -1.45871866e+00 1.90975830e-01
2.42848173e-02 -1.29623008e+00 -2.14528590e-01 -5.47403172e-02
7.34301388e-01 1.70873129e-03 -2.61443317e-01 4.03829575e-01
4.21203002e-02 -6.63764834e-01 7.80143380e-01 2.91147530e-01
3.22174579e-01 -6.55167341e-01 8.42320323e-01 3.41153085e-01
-1.03673172e+00 -4.64866370e-01 -4.31383371e-01 -4.28351499e-02
-2.58028626e-01 7.34516144e-01 -7.77574778e-01 2.73818433e-01
1.04489672e+00 3.92275214e-01 -1.34998739e+00 1.21735728e+00
-5.55641390e-02 8.75004530e-01 5.35708442e-02 5.21575399e-02
-2.33372912e-01 3.05519849e-01 6.69634640e-01 1.16529369e+00
3.56881589e-01 -6.96423411e-01 2.96730679e-02 1.30818665e+00
4.21487868e-01 -1.28144383e-01 -7.76556194e-01 -3.41029793e-01
4.82958615e-01 1.02102625e+00 -8.41759682e-01 5.95310591e-02
-9.17566478e-01 3.98097277e-01 3.91172945e-01 -1.58214495e-02
-7.62338459e-01 -8.62288535e-01 7.91835725e-01 2.18478009e-01
1.36377990e-01 6.00911118e-02 -1.87844098e-01 -1.14523590e+00
4.57801044e-01 -1.23915815e+00 3.50100428e-01 -2.96689451e-01
-1.13557410e+00 8.48078847e-01 6.04199769e-04 -1.10727072e+00
-2.23299295e-01 -4.94563907e-01 -1.43656993e+00 8.36357713e-01
-1.01410401e+00 -8.09121490e-01 -1.88338175e-01 -1.37047008e-01
3.90823245e-01 -3.99342120e-01 8.02572846e-01 2.55957663e-01
-7.27864027e-01 7.51783371e-01 -1.85426176e-01 2.18397066e-01
6.18879020e-01 -1.37335658e+00 1.06975520e+00 1.28942370e+00
-1.16268359e-01 1.16210985e+00 5.66224337e-01 -1.25373626e+00
-1.21321154e+00 -1.11939275e+00 9.61352289e-01 -1.01816940e+00
1.15538061e+00 -4.32791442e-01 -1.58549190e+00 8.50561798e-01
-2.07398385e-01 -8.35312530e-02 8.74330103e-01 7.77754709e-02
-7.66910136e-01 3.40664893e-01 -1.07405436e+00 1.91812798e-01
4.37338591e-01 -6.34196401e-01 -6.91045225e-01 2.10249156e-01
7.35604107e-01 8.81153420e-02 -7.80561388e-01 8.05730596e-02
3.55747432e-01 -1.35730207e+00 6.19911432e-01 -6.13023698e-01
8.06094825e-01 -1.63121641e-01 4.42292122e-03 -1.01892364e+00
-7.11582974e-02 -6.70761585e-01 -9.18537602e-02 1.74440193e+00
7.99366891e-01 -8.28949630e-01 4.52764779e-01 4.55220163e-01
-3.36555801e-02 -7.57924020e-01 -7.26860285e-01 -8.74772191e-01
8.44711810e-02 -4.27693874e-01 7.98666894e-01 1.14052868e+00
3.06180567e-01 -1.77855030e-01 2.06512362e-01 2.83221781e-01
4.96912062e-01 1.41492814e-01 6.22423172e-01 -1.47545302e+00
-7.17794895e-01 -4.62939918e-01 -6.37592077e-01 5.40009178e-02
2.53251463e-01 -7.12849081e-01 1.30834699e-01 -8.90826464e-01
4.59291935e-01 -5.27490258e-01 -6.83248267e-02 7.97577262e-01
-6.55111432e-01 -1.27275782e-02 -4.34716493e-01 1.29763037e-01
-1.39031574e-01 -3.83066118e-01 2.18699072e-02 -6.47410974e-02
-2.67377436e-01 2.65563279e-01 -5.68921745e-01 9.72790360e-01
7.34894931e-01 -1.01236439e+00 -6.28738776e-02 -2.87515968e-01
1.03492033e+00 6.97570890e-02 4.46372271e-01 -9.06479716e-01
-4.24094051e-01 -1.41278327e-01 -2.62388676e-01 -2.88850695e-01
-6.69805765e-01 -3.13894928e-01 1.49401754e-01 7.95635700e-01
-4.03684080e-01 4.11810756e-01 3.53265673e-01 4.67023671e-01
-9.17689204e-02 -1.17841935e+00 5.67978323e-01 -3.43774915e-01
-7.04691708e-01 -5.55977263e-02 -6.89354539e-01 1.10311151e-01
1.30852866e+00 -1.24440834e-01 -4.94642168e-01 4.81269479e-01
-3.44682097e-01 -1.05243765e-01 1.01273572e+00 7.17581928e-01
4.45192814e-01 -8.73745263e-01 -4.63103652e-01 2.78815687e-01
5.02743542e-01 -5.59538662e-01 1.88179109e-02 6.33127511e-01
-6.05563521e-01 8.84412006e-02 -3.86935949e-01 -4.02237386e-01
-1.60151196e+00 7.07394063e-01 1.25982434e-01 -1.04177199e-01
-5.18550992e-01 9.47714686e-01 -3.12276687e-02 -4.86218691e-01
-2.60748900e-02 -3.92443180e-01 -2.56766289e-01 -3.77238803e-02
8.68097126e-01 5.23119867e-01 2.82616884e-01 -3.48324150e-01
-8.12291741e-01 4.00087416e-01 -2.65445441e-01 2.73436189e-01
1.52159834e+00 4.43300813e-01 -8.49143684e-01 6.17318511e-01
1.22476113e+00 3.65094781e-01 -6.63976729e-01 3.03347886e-01
8.41994166e-01 -7.22547352e-01 -3.76534700e-01 -7.00833857e-01
-7.97193587e-01 9.12868083e-01 3.40599984e-01 6.89222813e-01
7.93715060e-01 2.70021856e-01 3.42512786e-01 2.42582589e-01
6.72539294e-01 -4.24326330e-01 2.11133182e-01 4.03132319e-01
7.41755903e-01 -1.14123797e+00 -2.01961756e-01 -4.38890189e-01
-1.51230425e-01 1.39048588e+00 1.05699074e+00 -1.59787312e-01
2.25730404e-01 6.09098852e-01 -8.20734277e-02 -1.63660750e-01
-9.81752098e-01 5.64764917e-01 -1.12726195e-02 9.57563221e-01
5.60091138e-01 -1.47421420e-01 -2.08231539e-01 9.79201317e-01
1.78310610e-02 -4.65145499e-01 1.31075752e+00 1.30066621e+00
-6.13403261e-01 -1.49598956e+00 -5.99123597e-01 6.56024814e-01
-8.39875638e-01 -1.72782689e-01 -7.86524117e-01 5.42591512e-01
1.78017572e-01 7.50925958e-01 -3.22306603e-01 -6.25212371e-01
1.71266779e-01 2.53599972e-01 3.27701539e-01 -1.23675621e+00
-1.11742866e+00 -4.32623476e-01 1.64471254e-01 -6.76156223e-01
2.53393084e-01 -8.78488064e-01 -9.03601885e-01 1.35052159e-01
-6.55861080e-01 2.08005145e-01 4.69102144e-01 7.85792172e-01
4.91334379e-01 5.55840254e-01 6.40204787e-01 -3.15668494e-01
-6.21735394e-01 -6.67034030e-01 -4.65844944e-02 9.78724509e-02
5.89506686e-01 -7.41431952e-01 -5.85113525e-01 3.50402027e-01] | [7.3006696701049805, 7.754489421844482] |
b3950f8c-5b25-4f18-9159-7f506dade823 | zero-shot-dialogue-relation-extraction-by | 2306.06141 | null | https://arxiv.org/abs/2306.06141v1 | https://arxiv.org/pdf/2306.06141v1.pdf | Zero-Shot Dialogue Relation Extraction by Relating Explainable Triggers and Relation Names | Developing dialogue relation extraction (DRE) systems often requires a large amount of labeled data, which can be costly and time-consuming to annotate. In order to improve scalability and support diverse, unseen relation extraction, this paper proposes a method for leveraging the ability to capture triggers and relate them to previously unseen relation names. Specifically, we introduce a model that enables zero-shot dialogue relation extraction by utilizing trigger-capturing capabilities. Our experiments on a benchmark DialogRE dataset demonstrate that the proposed model achieves significant improvements for both seen and unseen relations. Notably, this is the first attempt at zero-shot dialogue relation extraction using trigger-capturing capabilities, and our results suggest that this approach is effective for inferring previously unseen relation types. Overall, our findings highlight the potential for this method to enhance the scalability and practicality of DRE systems. | ['Yun-Nung Chen', 'Ze-Song Xu'] | 2023-06-09 | null | null | null | null | ['relation-extraction'] | ['natural-language-processing'] | [ 9.19970199e-02 7.93719828e-01 -2.13184819e-01 -2.87692904e-01
-8.97722900e-01 -9.00010288e-01 7.82414198e-01 3.36827368e-01
-1.33853674e-01 1.09950519e+00 4.40691233e-01 -2.80943096e-01
5.68395816e-02 -8.56904328e-01 5.43236434e-02 -6.16040686e-03
-5.53536117e-02 7.05611408e-01 3.90795439e-01 -7.46289909e-01
-2.17656314e-01 3.38708520e-01 -1.01733303e+00 3.54610473e-01
6.36608124e-01 5.74744940e-01 -3.25104535e-01 8.21952760e-01
-2.27207825e-01 1.19827008e+00 -1.02473950e+00 -6.31785214e-01
-3.51482369e-02 -4.32346404e-01 -1.64006162e+00 -4.67601456e-02
-8.82170051e-02 -7.04281032e-01 -5.45936704e-01 5.42587519e-01
3.76746565e-01 2.80805022e-01 3.65240991e-01 -1.16103995e+00
-2.60272712e-01 9.91240561e-01 5.38788503e-03 3.21182609e-01
9.28476632e-01 -1.83271971e-02 1.29075456e+00 -6.27952158e-01
8.39149952e-01 1.14210999e+00 5.25220215e-01 8.44709098e-01
-1.21684086e+00 -4.20275450e-01 8.47948249e-03 4.46516164e-02
-1.12036431e+00 -8.07736039e-01 5.80039859e-01 -4.68929708e-01
1.34152162e+00 6.47170603e-01 3.00225586e-01 1.41657829e+00
-3.24618310e-01 6.83823884e-01 1.03405154e+00 -7.25499392e-01
2.24259235e-02 3.49612176e-01 6.84868872e-01 6.91335618e-01
8.00571218e-02 -2.82896221e-01 -6.99144781e-01 -4.40962315e-01
3.61465991e-01 -4.65485245e-01 -2.70999998e-01 2.12704524e-01
-8.96168470e-01 8.31964672e-01 -1.16701536e-01 4.57840949e-01
-2.79980935e-02 -6.62426770e-01 4.84521329e-01 2.83362299e-01
6.35821640e-01 1.00224435e+00 -6.30893111e-01 -6.15637779e-01
-4.40550774e-01 1.12804838e-01 1.77285886e+00 1.31046367e+00
4.85565215e-01 -5.16616404e-01 -5.35621703e-01 8.67940605e-01
-1.27620190e-01 -1.20831169e-01 2.94333577e-01 -8.25579166e-01
5.74325204e-01 1.01302528e+00 4.31784928e-01 -7.05538988e-01
-5.80243230e-01 2.14999914e-01 -3.16096276e-01 -5.64972281e-01
5.73508203e-01 -6.54780090e-01 -3.83823812e-01 1.58282769e+00
5.79087853e-01 -7.72498697e-02 3.74119937e-01 6.00725114e-01
1.36797762e+00 3.14518631e-01 1.35321498e-01 -5.05033970e-01
1.71012366e+00 -8.07847381e-01 -1.16773760e+00 -1.56387612e-01
9.77191210e-01 -6.23081207e-01 9.18763459e-01 1.16308406e-02
-7.51988530e-01 -1.25745401e-01 -9.86307085e-01 -2.33602285e-01
-2.83994764e-01 6.41543244e-04 1.02128708e+00 6.28846288e-01
-5.72831333e-01 3.16093057e-01 -6.55261815e-01 -6.99617743e-01
-3.98868546e-02 2.53705382e-01 -2.44060755e-01 2.55907029e-01
-1.57575142e+00 1.15985847e+00 7.39624262e-01 -3.49961817e-02
-4.37853366e-01 -3.06176156e-01 -9.69389975e-01 1.20335519e-01
1.10725248e+00 -2.21471369e-01 1.71934140e+00 9.92579088e-02
-1.69047427e+00 5.06394863e-01 -2.99195588e-01 -5.24164319e-01
1.60579830e-01 -4.59306508e-01 -6.35903180e-01 2.18760222e-01
-2.65504606e-03 1.10670097e-01 9.14761499e-02 -1.03602159e+00
-6.03805959e-01 -1.20325431e-01 5.79534769e-01 1.49064705e-01
-6.75826788e-01 3.41679901e-01 -2.93424189e-01 -1.55473158e-01
-1.48191005e-01 -7.79294372e-01 5.05538732e-02 -7.10545123e-01
-8.23508620e-01 -7.34770298e-01 9.26024616e-01 -6.87690139e-01
1.76579273e+00 -1.88406038e+00 -8.15291256e-02 -6.33595735e-02
5.88971794e-01 4.92597967e-01 2.90384650e-01 9.38307941e-01
3.40126812e-01 1.88901529e-01 1.85248330e-01 -1.22444130e-01
-8.00285414e-02 4.72551882e-01 -3.31285894e-01 -3.25893372e-01
5.20365953e-01 1.07169139e+00 -1.13660538e+00 -7.14868188e-01
1.58585802e-01 1.71922490e-01 -1.96679786e-01 6.87723815e-01
-5.01402736e-01 5.25596201e-01 -6.14479780e-01 4.92980063e-01
8.70149757e-04 -4.05107915e-01 6.98201597e-01 -1.53684244e-01
3.53682339e-02 7.86965966e-01 -8.71904254e-01 1.22999501e+00
-3.05546284e-01 6.06147528e-01 -1.07241690e-01 -4.05408591e-01
1.05112600e+00 7.02348828e-01 3.27685565e-01 -1.33649215e-01
3.42239261e-01 -1.64542079e-01 -9.66824666e-02 -8.13080907e-01
8.03698242e-01 -7.76535645e-02 -4.01111633e-01 4.00783956e-01
2.19181821e-01 7.06721693e-02 4.28992242e-01 6.55193448e-01
1.49536121e+00 -1.55390605e-01 7.59560823e-01 3.36473674e-01
5.64704895e-01 4.11349297e-01 5.92203796e-01 6.45136356e-01
-1.60519898e-01 -8.40598121e-02 7.56241918e-01 -1.50888950e-01
-6.55413926e-01 -5.83187640e-01 -9.40545835e-03 1.19102097e+00
-2.04404816e-02 -8.36015046e-01 -7.80046403e-01 -1.04503214e+00
-1.83849558e-01 9.61006939e-01 -3.19163382e-01 -7.27640688e-02
-6.41226768e-01 -5.11642039e-01 1.09151495e+00 4.28736955e-01
5.59519768e-01 -7.94237614e-01 -2.45495334e-01 4.95123893e-01
-7.58507550e-01 -1.60910690e+00 -1.11141093e-01 2.55914390e-01
-4.47227180e-01 -1.33303571e+00 3.89482128e-03 -6.26546681e-01
2.92181134e-01 1.45646721e-01 1.09273231e+00 -9.75449290e-03
-2.05512732e-01 4.22189444e-01 -6.78118408e-01 -2.33721197e-01
-7.57939041e-01 5.37389278e-01 -1.20094299e-01 -3.90343159e-01
9.65679348e-01 -1.92065820e-01 -2.24128231e-01 3.86690021e-01
-3.76485497e-01 -9.23274159e-02 8.63382965e-02 9.19570208e-01
-1.95422441e-01 -2.43039150e-02 7.32469380e-01 -1.30327773e+00
1.19992626e+00 -5.04857540e-01 -2.93388277e-01 5.43458760e-01
-8.08735967e-01 8.17631558e-02 5.45383453e-01 -3.86192709e-01
-1.55912626e+00 5.65157309e-02 -2.49012895e-02 2.48214617e-01
-4.57167149e-01 4.22672749e-01 -1.10241570e-01 2.15832338e-01
8.91429961e-01 -3.03338796e-01 -9.93262157e-02 -6.36389196e-01
6.04423463e-01 1.12322032e+00 5.48868716e-01 -8.32943559e-01
5.93787134e-01 6.48694932e-02 -2.84867406e-01 -6.44078016e-01
-1.36265290e+00 -8.22283804e-01 -8.07779074e-01 -1.33333802e-01
8.71332347e-01 -7.23219156e-01 -9.64954138e-01 -9.73135307e-02
-1.14120615e+00 -2.40949661e-01 -2.50190616e-01 2.46182963e-01
-9.49769393e-02 4.88512456e-01 -1.21805811e+00 -1.13992953e+00
-6.31644368e-01 -6.84003532e-01 8.61238182e-01 3.40971619e-01
-1.01371741e+00 -9.03969944e-01 1.49088949e-01 7.78630137e-01
3.85391712e-02 1.53555125e-01 8.98776770e-01 -1.44315159e+00
-2.76747704e-01 -4.27694589e-01 -1.12595186e-01 -1.47889718e-01
4.47243929e-01 -1.78450823e-01 -1.10934007e+00 -1.71981514e-01
-1.40831634e-01 -8.07451010e-01 1.85657382e-01 -6.01696730e-01
4.09568310e-01 -6.62309408e-01 -3.86992693e-01 -1.38267204e-01
7.99906135e-01 3.69804978e-01 1.56752184e-01 2.62914687e-01
6.32405996e-01 7.26749957e-01 8.96809518e-01 5.94495714e-01
7.80544281e-01 7.97637045e-01 -1.86617360e-01 4.88673970e-02
-1.00008465e-01 -4.27869707e-01 -6.98802173e-02 8.41713548e-01
-2.51038279e-02 -2.41067603e-01 -8.35502744e-01 4.59253311e-01
-1.84923410e+00 -8.65667880e-01 -1.44897684e-01 1.71550751e+00
1.53255963e+00 1.48333147e-01 2.67622262e-01 8.98089632e-02
6.12369001e-01 4.96419445e-02 -2.78742522e-01 -5.51007211e-01
1.56499699e-01 3.84037316e-01 6.69557974e-02 6.72775865e-01
-1.03603125e+00 1.29823101e+00 6.57087517e+00 4.09226567e-01
-6.15831137e-01 1.44483954e-01 2.33862266e-01 1.60535738e-01
1.07070632e-01 2.81339467e-01 -1.25998926e+00 3.05985939e-02
1.10701859e+00 -4.56409752e-01 2.84632564e-01 8.87639463e-01
-7.78818503e-02 -3.83777842e-02 -1.48772681e+00 3.68260175e-01
-1.42224506e-01 -1.20191836e+00 -3.05935919e-01 9.33154952e-03
3.15629989e-01 -4.86757606e-01 -7.27674961e-01 6.92156672e-01
9.31070507e-01 -7.54385591e-01 -1.72170460e-01 2.10457206e-01
6.73262239e-01 -4.65171248e-01 8.34710300e-01 4.79408979e-01
-1.07169986e+00 -1.15063041e-02 -1.36956405e-02 -2.73306221e-01
1.75770566e-01 2.82813996e-01 -1.76483142e+00 5.85315704e-01
2.51739889e-01 2.38198593e-01 -4.80982453e-01 4.24817324e-01
-4.99733776e-01 6.74333513e-01 -2.23006219e-01 -3.49302649e-01
-1.38509765e-01 2.52868861e-01 5.38318574e-01 1.42105734e+00
-1.74709469e-01 6.63072169e-01 5.86311579e-01 5.49713790e-01
-2.40666211e-01 2.60065738e-02 -6.80889666e-01 -2.37030268e-01
9.76719737e-01 1.60199130e+00 -5.99877000e-01 -6.13442600e-01
-5.29304147e-01 8.33340645e-01 5.91555119e-01 2.09149703e-01
-5.35863638e-01 -3.59198809e-01 2.71775365e-01 -3.42640549e-01
3.43007036e-02 -2.58714497e-01 -2.37648904e-01 -1.28263271e+00
-1.68805346e-01 -1.01111817e+00 7.42687404e-01 -2.65495032e-01
-1.30844343e+00 7.58446574e-01 2.73611724e-01 -8.25634778e-01
-7.28522837e-01 -2.25795329e-01 -3.82919341e-01 6.02650046e-01
-1.01531565e+00 -1.05575061e+00 -4.35765952e-01 5.00650823e-01
6.40963137e-01 6.03270233e-02 1.37904370e+00 1.31891638e-01
-8.64615738e-01 7.87168086e-01 -5.18626988e-01 5.99489510e-01
8.10917556e-01 -1.39284587e+00 2.10114107e-01 6.66977167e-01
1.97436154e-01 9.36401904e-01 7.16022849e-01 -8.85387599e-01
-1.31185389e+00 -7.87206113e-01 1.13169909e+00 -8.34609747e-01
9.66236115e-01 -5.09855747e-01 -1.01163363e+00 8.24962616e-01
3.12427908e-01 -4.85529453e-01 1.07782435e+00 6.64062142e-01
-4.12130266e-01 2.09215447e-01 -1.21834385e+00 6.03495479e-01
1.03942323e+00 -7.32410908e-01 -1.09779084e+00 4.84567374e-01
8.03276181e-01 -6.87487483e-01 -1.51895058e+00 4.93150324e-01
3.01825345e-01 -4.91973847e-01 6.54762149e-01 -8.46333921e-01
1.83532119e-01 4.08580787e-02 1.61756411e-01 -9.42411542e-01
-8.91488791e-02 -1.08875561e+00 -7.90083706e-01 1.86674201e+00
7.27611125e-01 -3.96764666e-01 6.63249314e-01 1.42434955e+00
1.35289624e-01 -5.99065125e-01 -7.37874150e-01 -7.07170188e-01
-3.88302505e-01 -1.49911255e-01 6.70254171e-01 1.24025667e+00
8.51820052e-01 1.23801637e+00 -4.83703405e-01 1.20284520e-01
1.03756912e-01 3.50555070e-02 8.77152801e-01 -1.31852734e+00
-3.77388448e-01 1.00620426e-01 -4.17619161e-02 -1.01349533e+00
1.59565002e-01 -5.36032081e-01 1.18458509e-01 -1.43488598e+00
1.10134594e-01 -4.37167645e-01 1.00226521e-01 7.72636652e-01
-5.44431746e-01 -1.64816603e-01 -6.61721230e-02 2.11108074e-01
-6.74207091e-01 3.49477679e-01 1.04580295e+00 5.41019663e-02
-6.14671171e-01 7.06046075e-02 -7.44151771e-01 6.52489662e-01
7.68408418e-01 -2.55355924e-01 -6.19180083e-01 4.41748612e-02
-1.07309245e-01 3.81172299e-01 -1.47023141e-01 -4.40285355e-01
4.35623437e-01 -9.37895328e-02 -8.70537609e-02 -3.38656276e-01
5.52956998e-01 -6.00086570e-01 -3.27905804e-01 -3.34316934e-03
-6.05581164e-01 -3.28128785e-01 1.76335454e-01 5.63503265e-01
-1.85451746e-01 -3.69424373e-01 2.44779125e-01 -2.42778838e-01
-7.43016005e-01 -1.38235599e-01 -4.41261768e-01 4.47461218e-01
1.16134799e+00 2.10977450e-01 -7.57245541e-01 -4.49156702e-01
-7.72087216e-01 3.23879093e-01 -1.02658682e-02 3.82316381e-01
4.44523245e-01 -8.90099764e-01 -3.67972791e-01 -1.30547434e-01
3.78024548e-01 -9.53190923e-02 -3.24242920e-01 4.74783123e-01
5.74442670e-02 4.24871027e-01 2.27858871e-01 -1.50462374e-01
-1.73243690e+00 4.59429741e-01 4.08932082e-02 -6.41312718e-01
-7.95231521e-01 8.43807817e-01 -5.30533910e-01 -3.03874403e-01
4.60479289e-01 -1.51560083e-01 -6.10556960e-01 2.15742633e-01
5.66717565e-01 2.59839743e-01 -3.09654064e-02 -2.78380811e-01
-3.12385678e-01 -1.83471128e-01 -5.66650569e-01 -2.36962095e-01
1.03246701e+00 -2.88513094e-01 -2.59117305e-01 5.72926939e-01
6.17821395e-01 2.19678640e-01 -7.70986736e-01 -5.86232841e-01
7.08413601e-01 -2.00434878e-01 -4.23053861e-01 -9.51508522e-01
-2.50690788e-01 3.60243320e-01 -8.54377896e-02 7.54558921e-01
8.05289447e-01 3.04042548e-01 1.18265176e+00 9.32755828e-01
5.51070809e-01 -1.07882500e+00 -4.99474537e-03 7.71820664e-01
5.68055272e-01 -1.43210101e+00 -1.46485632e-02 -1.06767678e+00
-7.90246785e-01 1.16031349e+00 1.07382190e+00 5.28921902e-01
2.54140437e-01 5.54855406e-01 1.17973037e-01 -4.76751685e-01
-1.04719889e+00 -4.11165416e-01 2.24815592e-01 4.04538155e-01
8.48963320e-01 1.21441156e-01 -4.71671343e-01 5.97526133e-01
-2.81231582e-01 -1.82872325e-01 5.50747931e-01 1.13496959e+00
-3.05644572e-01 -1.52551782e+00 3.78091186e-02 4.97914761e-01
-5.14386594e-01 -1.77239180e-01 -1.19962060e+00 8.44693840e-01
-3.53731185e-01 1.76177537e+00 -4.70593691e-01 -7.34342277e-01
5.33060253e-01 6.15371108e-01 2.49273598e-01 -1.15862858e+00
-9.03082490e-01 -1.53326824e-01 1.38135612e+00 -3.47362995e-01
-2.10156873e-01 -1.75802872e-01 -1.27826726e+00 -3.51427346e-01
-7.78797507e-01 4.64649856e-01 -3.91431116e-02 1.24278069e+00
3.83169621e-01 5.04303753e-01 6.52651370e-01 1.02827705e-01
-6.09547734e-01 -1.28168702e+00 -8.69474933e-02 5.11021376e-01
2.19127871e-02 -8.19970369e-01 -1.23706147e-01 -2.73901727e-02] | [12.442692756652832, 8.072003364562988] |
a7732f94-b2d6-4f38-9cef-a44d802dd350 | multi-scale-cross-form-pyramid-network-for | 1904.11309 | null | https://arxiv.org/abs/1904.11309v3 | https://arxiv.org/pdf/1904.11309v3.pdf | Multi-scale Cross-form Pyramid Network for Stereo Matching | Stereo matching plays an indispensable part in autonomous driving, robotics and 3D scene reconstruction. We propose a novel deep learning architecture, which called CFP-Net, a Cross-Form Pyramid stereo matching network for regressing disparity from a rectified pair of stereo images. The network consists of three modules: Multi-Scale 2D local feature extraction module, Cross-form spatial pyramid module and Multi-Scale 3D Feature Matching and Fusion module. The Multi-Scale 2D local feature extraction module can extract enough multi-scale features. The Cross-form spatial pyramid module aggregates the context information in different scales and locations to form a cost volume. Moreover, it is proved to be more effective than SPP and ASPP in ill-posed regions. The Multi-Scale 3D feature matching and fusion module is proved to regularize the cost volume using two parallel 3D deconvolution structure with two different receptive fields. Our proposed method has been evaluated on the Scene Flow and KITTI datasets. It achieves state-of-the-art performance on the KITTI 2012 and 2015 benchmarks. | ['Yuchao Dai', 'Zhibo Rao', 'Bo Li', 'Mingyi He', 'Zhidong Zhu'] | 2019-04-25 | null | null | null | null | ['3d-feature-matching', 'stereo-matching', '3d-scene-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [-1.44214809e-01 -4.10558701e-01 2.10869327e-01 -4.40365613e-01
-4.61945176e-01 -7.41025433e-02 5.91544569e-01 -2.67341495e-01
-7.26397038e-01 3.69532555e-01 2.13148475e-01 -2.02258050e-01
-8.06827918e-02 -1.03426027e+00 -9.48703885e-01 -4.61027026e-01
1.54739037e-01 2.19561785e-01 7.41323054e-01 -4.92929041e-01
6.22032464e-01 6.93612039e-01 -2.08097792e+00 2.45235860e-01
8.93951356e-01 1.22404146e+00 4.89110023e-01 5.19315183e-01
-1.24495700e-01 8.20618808e-01 6.86195120e-02 2.38612313e-02
7.83742428e-01 -3.63668986e-02 -5.71103036e-01 5.51604778e-02
7.97309160e-01 -5.61865628e-01 -7.46095777e-01 1.21002018e+00
5.87134123e-01 9.93602574e-02 5.68229377e-01 -9.08735275e-01
-1.86458409e-01 -3.08005095e-01 -8.23121965e-01 4.64539915e-01
2.06690684e-01 3.73686373e-01 5.27749538e-01 -1.04782867e+00
6.33236349e-01 1.57048941e+00 6.17262542e-01 8.88364017e-02
-8.58689368e-01 -6.33221686e-01 -1.35144457e-01 5.51588893e-01
-1.12695706e+00 -2.18303964e-01 8.16590726e-01 -7.09070742e-01
1.18672109e+00 -1.79887906e-01 6.24611318e-01 3.44908714e-01
7.41656244e-01 5.72968960e-01 1.38255453e+00 -2.02931371e-02
-1.77271947e-01 -2.44926348e-01 1.94231391e-01 8.87242913e-01
3.76032740e-02 6.67506039e-01 -5.77847242e-01 3.08341771e-01
1.25005007e+00 4.33327705e-01 -2.68276542e-01 -4.17009801e-01
-1.16615617e+00 8.71553183e-01 9.86485600e-01 7.08735138e-02
-4.97110724e-01 -3.20946760e-02 2.16669634e-01 2.10713878e-01
3.93900424e-01 1.69504639e-02 -4.88138825e-01 1.05226405e-01
-6.40796721e-01 3.36142153e-01 2.39315391e-01 8.81953716e-01
1.54688346e+00 6.31732419e-02 -1.28712833e-01 7.91920245e-01
4.00511146e-01 7.50981808e-01 6.85249448e-01 -1.38480914e+00
7.56093144e-01 8.10004950e-01 1.64743215e-02 -1.19488692e+00
-6.67879343e-01 -3.12096328e-01 -9.89029348e-01 6.89524949e-01
1.76693097e-01 1.22931533e-01 -1.02564895e+00 1.34311593e+00
3.39775652e-01 4.19416696e-01 7.51999095e-02 1.33871090e+00
1.28261673e+00 7.99640536e-01 -4.40225393e-01 3.00537944e-01
1.29328859e+00 -1.06309164e+00 -4.05994594e-01 -4.33413386e-01
2.32718721e-01 -7.58825243e-01 5.55897117e-01 -1.26241118e-01
-1.06440902e+00 -1.03882515e+00 -1.15762353e+00 -6.51589811e-01
-5.13399124e-01 -7.29570910e-02 6.26650751e-01 -7.70562813e-02
-1.05477858e+00 6.27095819e-01 -4.40617353e-01 -1.36307359e-01
3.63096863e-01 3.58072996e-01 -8.23764741e-01 -3.48867983e-01
-1.18572319e+00 1.03218651e+00 2.95397550e-01 2.51947701e-01
-6.75105572e-01 -7.78258264e-01 -1.41683590e+00 -7.08849505e-02
-2.00063005e-01 -9.61826026e-01 7.68242300e-01 -5.36052823e-01
-1.42272019e+00 1.26309717e+00 -4.24827158e-01 -3.85978281e-01
3.78660351e-01 -1.04479373e-01 -1.06501780e-01 1.05982833e-01
4.42533106e-01 9.05107796e-01 8.65345716e-01 -8.90895724e-01
-1.14772320e+00 -7.96944678e-01 3.23780999e-02 7.38095045e-01
3.71664047e-01 -3.86911333e-01 -3.38604689e-01 -1.68132707e-01
5.18556058e-01 -4.68078792e-01 -4.65125859e-01 -8.85657147e-02
5.82358874e-02 2.12350339e-02 7.23555684e-01 -7.39937842e-01
5.56777120e-01 -2.11180019e+00 2.84149855e-01 -1.26627982e-01
3.36906940e-01 9.66720879e-02 -4.89030033e-02 -2.16927633e-01
6.27033636e-02 -5.65521121e-01 -2.18632430e-01 -2.74867773e-01
-2.67603159e-01 7.98105299e-02 -1.66095272e-01 5.61789095e-01
2.88672596e-01 8.46903205e-01 -7.78643131e-01 -4.77139652e-01
9.66510117e-01 5.15518844e-01 -7.19065189e-01 3.36097449e-01
4.70038772e-01 5.69677651e-01 -4.73474443e-01 4.45979655e-01
1.30637825e+00 1.70804948e-01 -8.46107781e-01 -3.19656342e-01
-6.15594029e-01 2.51096845e-01 -1.17450309e+00 2.09018731e+00
-4.98292178e-01 7.18468428e-01 2.23132297e-01 -9.78735805e-01
1.21281278e+00 -3.25936556e-01 3.44067067e-01 -1.24266112e+00
3.26314151e-01 4.22801048e-01 -3.04467589e-01 -6.25371456e-01
6.07849121e-01 -5.10038473e-02 -1.63921788e-02 -3.20657939e-01
2.18120396e-01 -6.31009281e-01 -2.17338800e-02 -2.49858677e-01
9.98569906e-01 2.17657164e-01 1.10542104e-01 -4.56502706e-01
1.20398939e+00 -6.68757707e-02 7.23748803e-01 3.63427401e-01
-4.88747239e-01 6.75028086e-01 9.43031684e-02 -8.29182506e-01
-1.07383215e+00 -9.55386102e-01 -2.02843696e-01 4.06109840e-01
9.01841044e-01 1.47451937e-01 -3.14242542e-01 -3.82881522e-01
3.30425054e-01 5.36134653e-02 -5.69109797e-01 -1.26839042e-01
-7.54600823e-01 -2.71372795e-01 -1.17687499e-02 4.49379951e-01
1.24530482e+00 -1.15227830e+00 -9.75249112e-01 1.51589751e-01
-3.22201312e-01 -1.21467340e+00 -5.45979977e-01 2.74165690e-01
-8.92611742e-01 -1.11047876e+00 -6.70693576e-01 -1.10356891e+00
4.03368503e-01 6.84993207e-01 9.84710038e-01 -4.29692864e-01
-3.71240556e-01 -1.27528504e-01 8.45651776e-02 -1.87924415e-01
1.47855327e-01 -2.97372252e-01 -2.23074049e-01 1.05542220e-01
4.18541372e-01 -8.73168290e-01 -7.53076255e-01 4.51459348e-01
-5.87688386e-01 2.88700074e-01 7.42218375e-01 9.93098199e-01
6.22133732e-01 -2.39284649e-01 -2.39530671e-02 -2.49203116e-01
8.34956244e-02 -1.19581409e-01 -9.36661243e-01 -3.71417791e-01
-1.03079163e-01 1.60164446e-01 5.43690979e-01 2.45460764e-01
-1.10574949e+00 1.65679514e-01 -3.90642673e-01 -5.77578127e-01
-2.63887346e-01 2.60525942e-02 -3.72325517e-02 -3.36869478e-01
5.38943529e-01 4.72799301e-01 1.72754209e-02 -3.71370882e-01
2.04782575e-01 5.64103365e-01 8.20127070e-01 -1.64710969e-01
8.73149812e-01 7.57956028e-01 3.92127365e-01 -6.67533517e-01
-8.13391149e-01 -7.88996041e-01 -9.31127965e-01 -2.84444392e-01
1.18476343e+00 -1.40703213e+00 -7.37715423e-01 9.26945210e-01
-1.26646674e+00 -1.29536122e-01 -2.38282800e-01 6.54612422e-01
-7.72337854e-01 2.64031649e-01 -6.81149244e-01 -2.70437598e-01
-3.61169845e-01 -1.46015561e+00 1.38300037e+00 6.40104055e-01
6.56705379e-01 -8.11422288e-01 1.54123053e-01 4.47202891e-01
3.68059158e-01 2.18951046e-01 3.39437306e-01 2.16574371e-01
-9.18953478e-01 1.02670103e-01 -7.03926861e-01 4.26393688e-01
-9.57153291e-02 -4.18686122e-01 -1.06234407e+00 -1.34278730e-01
1.48339808e-01 -2.93085515e-01 1.35712528e+00 6.87660158e-01
8.43002856e-01 2.68121690e-01 -5.66244610e-02 1.39246607e+00
1.53200865e+00 1.58430487e-01 7.57446527e-01 5.27624249e-01
8.80667627e-01 7.70023346e-01 7.04246759e-01 1.68176591e-01
7.12060094e-01 5.35668194e-01 6.25536859e-01 -2.72196978e-01
-1.97533607e-01 -2.26551935e-01 2.82860607e-01 6.17380142e-01
-1.33984208e-01 7.07120121e-01 -7.41532028e-01 4.88098770e-01
-1.95185399e+00 -1.06881881e+00 -1.78258002e-01 2.01947880e+00
3.59188616e-01 3.01566660e-01 -3.02013338e-01 1.21505171e-01
6.60597205e-01 3.05565029e-01 -5.64249277e-01 -3.19014102e-01
-4.49189574e-01 1.84990942e-01 9.35622633e-01 8.00683260e-01
-1.21274316e+00 8.89544666e-01 4.86183214e+00 8.79784465e-01
-1.23860836e+00 2.77830171e-03 6.14059210e-01 5.68418086e-01
-4.12095077e-02 -6.55346289e-02 -8.86948049e-01 3.10097426e-01
3.04924279e-01 4.78702225e-02 2.49053925e-01 7.04902947e-01
1.73847988e-01 -3.47668916e-01 -7.81153500e-01 1.55101156e+00
6.27159625e-02 -1.54354167e+00 -8.34091827e-02 -1.13653475e-02
7.66391873e-01 6.71062350e-01 -1.50328726e-01 2.88677126e-01
5.24456427e-02 -8.52356374e-01 7.39466906e-01 5.46131551e-01
6.18936896e-01 -7.31022358e-01 8.81450653e-01 4.07668710e-01
-1.51759446e+00 -3.07331920e-01 -6.38725698e-01 -4.95476037e-01
5.29988520e-02 5.65320730e-01 -1.21746302e-01 9.19023454e-01
1.06599844e+00 1.44532955e+00 -4.81757402e-01 1.17119801e+00
2.16109846e-02 -2.95081019e-01 -2.88621843e-01 3.43160570e-01
4.80714023e-01 -5.61179340e-01 5.41097283e-01 1.11307597e+00
3.45704079e-01 2.73498874e-02 7.65000656e-02 7.52075732e-01
2.15223253e-01 -1.44787565e-01 -9.24515069e-01 9.82171774e-01
1.70674250e-01 1.32674539e+00 -2.54924834e-01 -4.50112104e-01
-5.60818195e-01 8.08866620e-01 3.98952305e-01 2.71348447e-01
-5.67791045e-01 -5.48707604e-01 8.93162906e-01 -1.61774699e-02
4.88349468e-01 -1.64055005e-01 -5.57380259e-01 -1.39810181e+00
1.11523584e-01 -3.43183458e-01 2.28123546e-01 -1.03647971e+00
-1.07213366e+00 6.05025232e-01 -2.82346666e-01 -1.46957672e+00
-1.97598469e-02 -8.09479296e-01 -6.34739220e-01 1.26957750e+00
-2.35436106e+00 -8.69296312e-01 -8.89702678e-01 1.16982949e+00
6.72581077e-01 -1.03365563e-01 3.70168418e-01 4.79931384e-01
-3.20229411e-01 -6.42287433e-02 -4.11083736e-03 1.33918608e-02
6.20151818e-01 -1.01862490e+00 3.15953761e-01 8.13519239e-01
-6.09779060e-01 8.00205767e-02 3.79522920e-01 -2.60703623e-01
-1.20794475e+00 -1.27169931e+00 8.29602301e-01 -1.66790903e-01
2.65528291e-01 -1.41128719e-01 -6.67094469e-01 4.02519345e-01
5.99324033e-02 4.16456908e-01 -3.38430971e-01 -5.99398911e-01
-3.32099795e-01 -6.60309374e-01 -1.36747813e+00 2.20685035e-01
1.21130669e+00 -6.12968147e-01 -8.01297426e-01 1.78292636e-02
7.23226547e-01 -7.68479645e-01 -6.96570456e-01 7.85860837e-01
4.61921126e-01 -1.62297487e+00 1.19459844e+00 -4.40811478e-02
7.63950467e-01 -5.59767783e-01 -3.30469906e-01 -1.09651864e+00
-5.26223361e-01 -4.13876295e-01 2.87968636e-01 6.93721771e-01
9.56787691e-02 -9.48928833e-01 6.05700314e-01 8.42467248e-02
-6.19764984e-01 -5.24494588e-01 -1.21037459e+00 -4.99430060e-01
-5.39280623e-02 -3.27743739e-01 4.43973660e-01 5.70339680e-01
-2.53504395e-01 5.18093467e-01 -2.36833751e-01 1.63797006e-01
9.38079536e-01 4.01928186e-01 9.58086431e-01 -1.19879234e+00
2.67224899e-03 -6.48116648e-01 -9.81990814e-01 -1.60078800e+00
1.58594489e-01 -7.56158173e-01 2.03079760e-01 -1.50321090e+00
1.21766321e-01 -1.62582085e-01 -2.83935905e-01 5.86192161e-02
-1.35442600e-01 3.05724889e-01 9.46347266e-02 2.00850233e-01
-1.94456458e-01 8.05265307e-01 1.69849527e+00 -2.31396899e-01
-2.74759591e-01 -8.12390000e-02 -1.16711207e-01 7.30922580e-01
6.21807098e-01 1.46536864e-02 -2.05931097e-01 -5.50619483e-01
-2.78710932e-01 2.90732205e-01 5.57902813e-01 -1.24717808e+00
3.76126558e-01 -8.58879760e-02 5.57225764e-01 -1.12151551e+00
5.93071997e-01 -8.99495184e-01 -3.22006494e-01 6.54184103e-01
9.04971585e-02 2.12757930e-01 2.86531508e-01 3.05665433e-01
-7.18880057e-01 3.13576579e-01 1.11061633e+00 -2.77201861e-01
-1.30821288e+00 6.74214840e-01 1.20595312e-02 1.14690550e-01
9.20861900e-01 -3.78922313e-01 -3.33904982e-01 1.34591423e-02
-2.24223390e-01 4.49125558e-01 4.65082884e-01 6.21891558e-01
9.35668230e-01 -1.39892936e+00 -6.66580856e-01 7.62443304e-01
1.65088192e-01 5.36030531e-01 6.18937671e-01 9.11250353e-01
-9.69696164e-01 5.96870124e-01 -8.66707385e-01 -1.13563156e+00
-8.79661679e-01 4.27606553e-01 7.19874084e-01 -2.39229023e-01
-7.34436035e-01 8.87106717e-01 6.90556705e-01 -7.77753949e-01
7.48354271e-02 -4.59298879e-01 -4.78317618e-01 -3.73097003e-01
4.43214476e-01 3.57128322e-01 -3.07749258e-03 -1.08637059e+00
-3.88068765e-01 1.48401165e+00 3.05816770e-01 -5.42487809e-03
1.38429964e+00 -3.08800757e-01 -3.52235079e-01 2.90240556e-01
1.47594559e+00 -5.10961175e-01 -1.80302656e+00 -4.06266749e-01
-4.97649580e-01 -7.38411784e-01 2.61648774e-01 -2.13726461e-01
-1.21385324e+00 1.24992597e+00 7.09256887e-01 -2.92413354e-01
1.24588943e+00 -4.16658044e-01 8.56127918e-01 2.31717721e-01
4.39453334e-01 -8.13268304e-01 -2.02399537e-01 1.01877427e+00
8.76758575e-01 -1.70347428e+00 -1.40543059e-01 -3.12948734e-01
-3.59297901e-01 1.13906872e+00 8.83454204e-01 -7.02041328e-01
9.94449496e-01 7.23052397e-02 4.64160107e-02 -2.07398668e-01
-4.26237136e-01 -4.11946416e-01 4.60521072e-01 5.98981321e-01
-9.40590575e-02 -1.62627816e-01 -1.27422467e-01 1.64027095e-01
-2.27019683e-01 -1.69825461e-02 2.89021850e-01 7.17272818e-01
-6.73188806e-01 -5.58196127e-01 -3.30392689e-01 1.98824003e-01
7.17735365e-02 -4.83985990e-02 2.89680451e-01 6.23453796e-01
5.79471588e-01 8.41968894e-01 3.68137091e-01 -6.75910413e-01
6.81010842e-01 -3.51580173e-01 5.23346066e-01 -2.61296719e-01
-4.23369467e-01 -7.50370622e-02 -2.16100723e-01 -1.19305861e+00
-5.57112992e-01 -5.31399786e-01 -1.25740755e+00 -2.93593615e-01
2.16964945e-01 -1.76328167e-01 6.51609540e-01 9.07810628e-01
3.35292399e-01 3.30256492e-01 9.45432246e-01 -1.48731160e+00
-8.48596469e-02 -9.08947110e-01 -5.79284668e-01 3.77188921e-01
9.02226210e-01 -8.38707507e-01 -3.84271473e-01 -6.25260221e-03] | [8.870176315307617, -2.280872106552124] |
9ab657f6-d383-44f8-9f75-d5a01a0ead94 | context-aware-change-detection-with-semi | 2306.08935 | null | https://arxiv.org/abs/2306.08935v1 | https://arxiv.org/pdf/2306.08935v1.pdf | Context-Aware Change Detection With Semi-Supervised Learning | Change detection using earth observation data plays a vital role in quantifying the impact of disasters in affected areas. While data sources like Sentinel-2 provide rich optical information, they are often hindered by cloud cover, limiting their usage in disaster scenarios. However, leveraging pre-disaster optical data can offer valuable contextual information about the area such as landcover type, vegetation cover, soil types, enabling a better understanding of the disaster's impact. In this study, we develop a model to assess the contribution of pre-disaster Sentinel-2 data in change detection tasks, focusing on disaster-affected areas. The proposed Context-Aware Change Detection Network (CACDN) utilizes a combination of pre-disaster Sentinel-2 data, pre and post-disaster Sentinel-1 data and ancillary Digital Elevation Models (DEM) data. The model is validated on flood and landslide detection and evaluated using three metrics: Area Under the Precision-Recall Curve (AUPRC), Intersection over Union (IoU), and mean IoU. The preliminary results show significant improvement (4\%, AUPRC, 3-7\% IoU, 3-6\% mean IoU) in model's change detection capabilities when incorporated with pre-disaster optical data reflecting the effectiveness of using contextual information for accurate flood and landslide detection. | ['Yifang Ban', 'Andrea Nascetti', 'Ritu Yadav'] | 2023-06-15 | null | null | null | null | ['change-detection'] | ['computer-vision'] | [ 2.14513808e-01 -1.69414714e-01 1.18126422e-01 -8.13147575e-02
-4.29077774e-01 -4.49772894e-01 6.14922106e-01 8.62261176e-01
-6.44398153e-01 8.65304410e-01 5.34014642e-01 -7.20689833e-01
-2.83676505e-01 -1.54656684e+00 -2.33959183e-01 -6.29630625e-01
-5.26922703e-01 1.35674074e-01 3.53963137e-01 -7.34611332e-01
2.33433954e-03 8.75131965e-01 -1.54183817e+00 -1.89190581e-02
7.00322092e-01 7.63314545e-01 3.54968041e-01 5.44974029e-01
-9.37615260e-02 -2.84114890e-02 -2.49410033e-01 5.98694861e-01
4.35287237e-01 7.83105865e-02 -5.65585554e-01 -2.55964458e-01
7.96528608e-02 -7.82762587e-01 1.32138491e-01 9.19799745e-01
5.90483725e-01 -6.77021071e-02 6.10282063e-01 -8.21530521e-01
3.20722908e-02 1.23151250e-01 -7.26460457e-01 7.50182807e-01
4.80089486e-01 1.05111882e-01 5.69576085e-01 -6.72358334e-01
5.48239410e-01 1.21137738e+00 8.53966773e-01 -3.65276933e-01
-1.08472264e+00 -6.11712933e-01 1.01704627e-01 -2.33999893e-01
-1.51734567e+00 -2.80033588e-01 1.69931278e-01 -7.34518468e-01
1.03175652e+00 2.59308279e-01 6.48545384e-01 -2.28805654e-02
4.19191241e-01 2.03857481e-01 1.13057196e+00 -6.39562190e-01
2.21482396e-01 -1.05695277e-01 -6.09220192e-02 1.67910844e-01
6.08210921e-01 4.15879786e-01 -2.67600983e-01 -7.69144446e-02
4.48962301e-01 4.05659050e-01 -2.10851356e-01 3.87241781e-01
-7.78006554e-01 7.30469465e-01 9.48292851e-01 4.89014000e-01
-9.68641341e-01 -2.24642932e-01 2.18393564e-01 3.35345179e-01
9.49234843e-01 3.28627750e-02 -3.24728012e-01 2.67449051e-01
-1.16659248e+00 2.41477177e-01 -1.15944725e-02 3.57366920e-01
1.16176045e+00 6.24472983e-02 -5.92188835e-02 4.46217149e-01
4.69608366e-01 1.16716444e+00 -6.03024252e-02 -4.01621729e-01
4.15934831e-01 9.46412802e-01 2.67803967e-01 -1.35404551e+00
-7.43307531e-01 -3.62348944e-01 -8.99000645e-01 1.92517191e-01
-1.40547499e-01 -3.32822561e-01 -9.59343851e-01 1.39408624e+00
3.10973555e-01 -2.30144769e-01 6.35427982e-02 7.56671429e-01
4.54221547e-01 8.95453393e-01 6.21417820e-01 -3.61297011e-01
1.13979769e+00 3.90954107e-01 -4.91388530e-01 -4.25575167e-01
6.27854228e-01 -4.76539314e-01 5.94621480e-01 -3.64040911e-01
-2.38140240e-01 -4.07647163e-01 -7.24861264e-01 8.33413422e-01
-7.36004651e-01 -1.34076536e-01 4.35550481e-01 4.72731918e-01
-1.12795210e+00 8.41984898e-02 -6.92661583e-01 -1.03776729e+00
3.44133347e-01 1.10315129e-01 -3.91352594e-01 -2.55506724e-01
-1.29078317e+00 8.84090722e-01 5.31808734e-01 2.62458533e-01
-6.39876485e-01 -7.76133776e-01 -6.59877658e-01 -1.82581581e-02
-1.18205585e-02 -2.57958353e-01 3.81256044e-01 -3.96438718e-01
-5.30155778e-01 6.35247290e-01 2.03816846e-01 -4.54137564e-01
2.42250025e-01 -6.87458813e-02 -7.06038296e-01 2.02100262e-01
6.51328146e-01 7.25648463e-01 -4.77686040e-02 -1.21323252e+00
-1.16008687e+00 -8.82304192e-01 -2.15730979e-03 3.01697254e-01
-2.36819789e-01 1.08512297e-01 3.65442038e-01 -4.54428017e-01
4.58086491e-01 -7.18445539e-01 -2.97983229e-01 -7.33434781e-02
2.56822780e-02 2.82281071e-01 1.17505240e+00 -8.00670743e-01
1.43459761e+00 -1.94420421e+00 -8.88439238e-01 3.30844939e-01
-2.97260195e-01 6.37953103e-01 -3.15748341e-02 9.48558152e-01
2.88448840e-01 3.68997604e-01 -8.44368994e-01 3.72259468e-01
-5.15145481e-01 3.10211897e-01 -2.85078496e-01 2.98934430e-01
1.22412719e-01 3.54630768e-01 -8.90101075e-01 -1.60226315e-01
5.85834146e-01 4.02588129e-01 -9.32712778e-02 -7.05659762e-02
-3.13497894e-02 4.50475931e-01 -5.51918626e-01 7.29948401e-01
1.13819265e+00 5.70525646e-01 -6.54018372e-02 3.24801862e-01
-8.75340819e-01 5.70248486e-03 -1.35804486e+00 8.63477945e-01
-3.95641834e-01 6.94838881e-01 9.08387527e-02 -7.19499290e-01
1.27410388e+00 3.88915688e-01 6.05306029e-01 -8.52930367e-01
-3.67341548e-01 7.67157003e-02 -4.13388163e-01 -6.42880023e-01
5.50330520e-01 -3.62349212e-01 -1.80790685e-02 5.47954738e-01
-6.35960460e-01 5.96840866e-04 1.14165835e-01 4.38252464e-02
1.00555038e+00 -3.04674596e-01 3.84767890e-01 -5.76062262e-01
1.65338010e-01 3.04826409e-01 3.96793872e-01 6.09071732e-01
-2.95850873e-01 3.49426448e-01 8.39699432e-02 -7.36786783e-01
-6.54452026e-01 -9.73222375e-01 -4.76899326e-01 8.26282203e-01
-5.47245592e-02 1.74169287e-01 -2.85782635e-01 -8.93380269e-02
1.69179574e-01 7.85987794e-01 -3.14242363e-01 3.40948515e-02
-1.05822437e-01 -1.59212983e+00 6.62335515e-01 3.27225953e-01
1.20841181e+00 -9.38964844e-01 -1.08182728e+00 4.08077925e-01
-3.18403363e-01 -7.89574027e-01 5.33240199e-01 -7.15634599e-02
-1.25521636e+00 -1.01121330e+00 -2.83561051e-01 -2.22288802e-01
6.71786904e-01 6.92779779e-01 8.32457364e-01 8.52439180e-02
-2.75105983e-01 4.90029544e-01 -5.46460390e-01 -5.38715601e-01
-1.61636353e-01 -6.43997192e-02 -1.17987581e-01 -2.56919414e-01
4.41348761e-01 -6.76057816e-01 -8.41959655e-01 1.42113551e-01
-1.16071475e+00 -1.76113188e-01 4.86304581e-01 1.89236738e-02
3.14284533e-01 2.99797893e-01 6.51499510e-01 -4.90972102e-01
4.90165561e-01 -7.65991092e-01 -5.23721755e-01 1.10729516e-01
-8.87792468e-01 -3.60416412e-01 -2.31869251e-01 2.90648162e-01
-1.08755195e+00 5.48835285e-02 1.70508083e-02 5.71693361e-01
-7.02368855e-01 1.27332723e+00 -1.54351750e-02 2.37422004e-01
8.21443558e-01 -2.67937128e-02 -3.25202048e-01 -4.94518787e-01
-2.25735098e-01 1.06220853e+00 5.41967869e-01 4.29199962e-03
7.35412061e-01 8.21317732e-01 -4.07482125e-02 -1.30246437e+00
-4.24859732e-01 -1.03471589e+00 -8.93873334e-01 -5.06493628e-01
4.41794008e-01 -1.11501586e+00 -2.87098233e-02 6.32476151e-01
-9.92034793e-01 -1.41540905e-02 -1.43576274e-02 6.64163291e-01
2.90138811e-01 6.39898255e-02 1.90575883e-01 -1.19225967e+00
-6.66946352e-01 -4.03802246e-01 6.98882937e-01 2.12490141e-01
1.14043683e-01 -9.78441894e-01 5.71668386e-01 -7.34129325e-02
9.24991667e-01 8.96297038e-01 7.33800173e-01 -4.82907183e-02
-3.23721766e-01 -4.99540269e-01 -6.36933923e-01 5.16822785e-02
5.90171933e-01 -8.08909535e-02 -7.72370875e-01 -2.69286424e-01
-4.47509944e-01 4.58121210e-01 1.06844473e+00 6.48502946e-01
1.69058237e-02 -2.50826806e-01 -4.73389059e-01 2.51268119e-01
1.97594309e+00 2.34831244e-01 8.24344099e-01 6.12562656e-01
1.10182531e-01 6.29182160e-01 6.28907263e-01 9.60228264e-01
5.96817970e-01 4.72288400e-01 9.16952312e-01 -2.98305273e-01
-1.95984766e-01 8.04887787e-02 3.95865917e-01 3.91582586e-02
-2.36179620e-01 -1.11031048e-01 -1.76204491e+00 9.77365613e-01
-1.50510573e+00 -1.11695719e+00 -6.05127811e-01 2.39036465e+00
2.90899515e-01 -6.32968768e-02 3.91224138e-02 3.01399171e-01
8.84952307e-01 2.45160371e-01 -1.38833970e-01 -7.05442950e-02
-2.63656020e-01 8.61089081e-02 9.68693256e-01 6.82630837e-01
-1.08098435e+00 8.98054242e-01 5.97475433e+00 -4.28789295e-03
-1.24682891e+00 2.72169784e-02 2.21586645e-01 3.43579620e-01
-3.44972879e-01 2.60012984e-01 -6.97973073e-01 2.20587790e-01
9.43817914e-01 6.34173006e-02 -1.09902710e-01 1.19231388e-01
1.15503001e+00 -1.10186589e+00 3.99338491e-02 7.83274472e-02
-4.92050618e-01 -1.13036323e+00 1.79231107e-01 1.63256600e-01
7.74921358e-01 4.34132278e-01 -3.96009117e-01 -1.92114878e-02
4.18965250e-01 -3.72270465e-01 2.14219630e-01 7.21643090e-01
9.20116663e-01 -5.80673814e-01 1.05756783e+00 3.87620658e-01
-1.38369358e+00 1.05124731e-02 -2.31412858e-01 -6.22993112e-01
1.14689603e-01 8.89085829e-01 -1.03650212e+00 6.30046308e-01
8.38379145e-01 6.18471801e-01 -5.66330552e-01 1.11707532e+00
-1.64878190e-01 7.90774345e-01 -8.29172194e-01 3.95243615e-01
4.18468267e-01 6.70857169e-03 6.24168694e-01 1.22181940e+00
6.16999209e-01 6.02763414e-01 3.19891959e-01 4.40927923e-01
3.50243807e-01 1.50760606e-01 -7.82324672e-01 5.10354102e-01
6.36183262e-01 7.67863631e-01 -6.81534231e-01 -7.54706003e-03
-3.34136449e-02 3.00706238e-01 -5.21066904e-01 3.86109114e-01
-9.55841914e-02 -3.48429948e-01 6.36692822e-01 6.81865156e-01
-1.12079941e-01 -3.77971977e-01 -3.36750269e-01 -3.35037977e-01
-1.57530338e-01 -9.37917531e-02 4.20001775e-01 -7.99325943e-01
-4.53518927e-01 2.88131833e-01 5.20013869e-01 -1.17634773e+00
-2.78618671e-02 -1.48981377e-01 -8.71631265e-01 9.88514900e-01
-1.84601104e+00 -1.24631810e+00 -7.00890124e-01 3.40331614e-01
5.94971031e-02 6.92529185e-03 8.25899661e-01 3.27942073e-01
-4.09868628e-01 -2.79054612e-01 4.14333731e-01 -5.61381467e-02
3.32819313e-01 -7.21844554e-01 1.75119817e-01 1.11049056e+00
-4.69170898e-01 -1.03196874e-02 6.93717837e-01 -1.21564317e+00
-6.06704295e-01 -1.50365877e+00 1.11300385e+00 4.32751715e-01
3.66947979e-01 4.90258694e-01 -6.66060030e-01 3.98730427e-01
-4.23198223e-01 -7.77104795e-02 4.75490272e-01 1.55647099e-01
6.90513253e-02 -5.27699828e-01 -1.59253287e+00 2.07533985e-01
6.56911612e-01 -4.84406859e-01 -1.46370977e-01 2.43468672e-01
3.31148714e-01 3.30449432e-01 -7.34237611e-01 7.00424731e-01
4.48775351e-01 -1.01096499e+00 6.99650288e-01 -6.08897470e-02
1.81105360e-01 -3.63856286e-01 -5.77713549e-01 -1.25429070e+00
-3.68548930e-01 7.76113495e-02 8.60787451e-01 1.20739043e+00
1.89894393e-01 -5.74047208e-01 2.75860369e-01 4.65307772e-01
-1.24819048e-01 1.58526033e-01 -1.06272602e+00 -6.84388876e-01
-9.79920756e-03 -6.81631446e-01 6.79855585e-01 6.24105096e-01
-3.86695772e-01 -2.12452888e-01 5.97662292e-02 9.73102212e-01
2.52442360e-01 -2.37519994e-01 5.82795620e-01 -1.63927448e+00
8.46549869e-01 1.50994202e-02 -6.02143824e-01 -1.46952346e-01
-4.85685557e-01 -5.44510484e-01 -1.95011720e-01 -1.91904128e+00
7.35139251e-02 -8.11437607e-01 -3.00079703e-01 9.73834813e-01
-7.43600130e-02 1.49298295e-01 7.85822340e-05 5.31806171e-01
1.92934975e-01 4.75983292e-01 4.92583901e-01 -1.72492415e-01
-5.47868669e-01 1.33366153e-01 -2.05864593e-01 4.70816731e-01
9.50774014e-01 -8.44416559e-01 9.07493476e-03 -6.63915396e-01
3.62016857e-01 7.28021413e-02 7.45472610e-01 -1.48607147e+00
6.20666854e-02 -4.87803996e-01 3.26035470e-02 -9.96058643e-01
-1.97682321e-01 -9.13978517e-01 4.98325348e-01 7.88718581e-01
1.13418497e-01 8.16377923e-02 6.24081373e-01 4.36134487e-01
-2.42410526e-01 -2.13924170e-01 6.97608292e-01 -9.75474901e-03
-1.01769567e+00 3.28789413e-01 -8.79369080e-01 -4.14577872e-01
7.95541942e-01 -3.41455102e-01 -3.26977521e-01 -3.59129369e-01
-5.98022640e-01 2.57739305e-01 3.51481766e-01 1.55126542e-01
4.49749053e-01 -1.02939498e+00 -1.18242955e+00 3.27607781e-01
3.46005440e-01 -2.12824374e-01 2.53233522e-01 5.87855816e-01
-9.02280927e-01 4.71810520e-01 -5.51954567e-01 -4.94512141e-01
-1.02281547e+00 -5.46136200e-02 2.80686885e-01 -3.03479344e-01
-2.68028945e-01 3.15231621e-01 -2.40495458e-01 -4.85661924e-01
-2.93107361e-01 -2.66088665e-01 -2.26209551e-01 5.66461205e-01
4.69461262e-01 5.96253753e-01 1.60581976e-01 -7.35935152e-01
-5.82277954e-01 4.63621646e-01 3.87451619e-01 -3.58793437e-01
1.60935915e+00 -6.89726114e-01 -1.51466846e-01 3.15538973e-01
6.16241693e-01 -2.55507946e-01 -1.05689192e+00 -4.59544122e-01
2.75293648e-01 -3.58977556e-01 7.73223341e-01 -1.04898429e+00
-8.38484406e-01 8.38867068e-01 1.27706945e+00 1.58402637e-01
1.37735510e+00 -2.86790609e-01 4.61570740e-01 4.54187304e-01
4.22306418e-01 -9.79779124e-01 -6.16077185e-01 5.38027406e-01
1.00784600e+00 -1.42224240e+00 4.50928926e-01 -1.12967074e-01
-1.50412664e-01 9.30169225e-01 7.05904290e-02 1.94040477e-01
1.12021840e+00 -5.32172062e-02 1.12443388e-01 -2.98495919e-01
-1.90164492e-01 -8.35809886e-01 -3.34210455e-01 8.14073205e-01
1.29508466e-01 6.19627237e-01 -1.92659155e-01 -7.96711445e-02
2.59312421e-01 1.14707179e-01 3.85052592e-01 1.13079357e+00
-1.28429115e+00 -7.14412093e-01 -8.79554093e-01 4.85679150e-01
-1.11705586e-01 -2.45199814e-01 -9.00087357e-02 9.21092868e-01
2.33312845e-01 1.20566595e+00 3.03364307e-01 -2.19041437e-01
3.55517060e-01 -3.76680762e-01 -2.11103186e-01 -6.06936514e-01
-5.04099190e-01 -1.49122030e-01 -9.51523706e-02 1.67519751e-03
-7.86006927e-01 -1.01481998e+00 -1.05355370e+00 -6.06495202e-01
-1.38523415e-01 -7.02301413e-02 9.47985649e-01 1.05301678e+00
3.69814545e-01 -1.23535767e-02 8.27602804e-01 -6.96685195e-01
1.63028643e-01 -1.32371533e+00 -6.20965064e-01 -1.61717385e-01
5.00226676e-01 -6.17979467e-01 -2.17676729e-01 -1.19273618e-01] | [9.422934532165527, -1.3801205158233643] |
d6e298e4-0110-4bcd-9cac-f033b800d165 | reinforced-pedestrian-attribute-recognition | 2205.14042 | null | https://arxiv.org/abs/2205.14042v1 | https://arxiv.org/pdf/2205.14042v1.pdf | Reinforced Pedestrian Attribute Recognition with Group Optimization Reward | Pedestrian Attribute Recognition (PAR) is a challenging task in intelligent video surveillance. Two key challenges in PAR include complex alignment relations between images and attributes, and imbalanced data distribution. Existing approaches usually formulate PAR as a recognition task. Different from them, this paper addresses it as a decision-making task via a reinforcement learning framework. Specifically, PAR is formulated as a Markov decision process (MDP) by designing ingenious states, action space, reward function and state transition. To alleviate the inter-attribute imbalance problem, we apply an Attribute Grouping Strategy (AGS) by dividing all attributes into subgroups according to their region and category information. Then we employ an agent to recognize each group of attributes, which is trained with Deep Q-learning algorithm. We also propose a Group Optimization Reward (GOR) function to alleviate the intra-attribute imbalance problem. Experimental results on the three benchmark datasets of PETA, RAP and PA100K illustrate the effectiveness and competitiveness of the proposed approach and demonstrate that the application of reinforcement learning to PAR is a valuable research direction. | ['Shengjia Li', 'Yaodong Wang', 'Zhenfei Hu', 'Zhong Ji'] | 2022-05-21 | null | null | null | null | ['pedestrian-attribute-recognition'] | ['computer-vision'] | [ 2.73795933e-01 -7.54303634e-02 -3.25474560e-01 -7.00590909e-01
-6.82740808e-01 -1.53070658e-01 5.77297688e-01 -7.56218955e-02
-4.36339438e-01 6.43935084e-01 1.29993871e-01 -1.27858192e-01
-3.93083766e-02 -7.46308208e-01 -6.34082139e-01 -9.03922141e-01
1.25447705e-01 4.49820369e-01 1.70835570e-01 3.72624174e-02
2.43017152e-01 2.80377209e-01 -1.58446825e+00 1.56421497e-01
1.18981361e+00 1.21534693e+00 -2.66927592e-02 4.42807585e-01
1.06790820e-02 1.10000229e+00 -6.45130992e-01 -5.18681109e-01
2.72116244e-01 -5.30379653e-01 -7.44055212e-01 6.84243023e-01
1.01739660e-01 -2.52108812e-01 -1.78909093e-01 8.82552266e-01
6.41996026e-01 3.01587969e-01 7.92927921e-01 -1.82314932e+00
-3.30728382e-01 1.56132221e-01 -7.94106483e-01 1.86724663e-01
3.35913211e-01 3.73393744e-01 8.57674479e-01 -3.09593916e-01
8.25638548e-02 1.45997083e+00 3.30392897e-01 5.50200403e-01
-9.22010124e-01 -5.69022655e-01 6.48593783e-01 8.72985244e-01
-1.19108629e+00 -2.52517611e-01 7.90455759e-01 -5.21619618e-01
5.60496092e-01 1.76914752e-01 7.41597056e-01 1.04391193e+00
2.84634948e-01 1.13313174e+00 1.22024298e+00 -1.95488423e-01
5.09695709e-01 -3.53202373e-02 4.64106351e-02 4.32398111e-01
-4.46279161e-02 3.94325405e-02 -2.95904666e-01 -1.53260320e-01
4.66207832e-01 1.91495657e-01 1.35760725e-01 -6.67757511e-01
-1.06448340e+00 7.70232141e-01 2.88357705e-01 -6.18279457e-01
-8.01877856e-01 -1.30774021e-01 5.34235179e-01 1.91822127e-01
2.02421665e-01 1.20016985e-01 -2.54424453e-01 -4.49400656e-02
-2.58177400e-01 2.79230028e-01 5.81758678e-01 9.61201012e-01
6.89883530e-01 -3.51700298e-02 -4.49273705e-01 8.15801919e-01
3.67542744e-01 5.91695189e-01 2.07097426e-01 -9.73605871e-01
4.98867452e-01 6.79950237e-01 5.26263773e-01 -9.26459372e-01
-2.48581678e-01 -1.66685656e-01 -1.01235485e+00 3.57908875e-01
2.37193614e-01 -4.89414871e-01 -8.26269150e-01 1.66091752e+00
6.68190777e-01 3.86272907e-01 3.65434676e-01 9.89414692e-01
5.50054431e-01 7.75323868e-01 6.87115431e-01 -5.10956645e-01
1.49102867e+00 -8.35002124e-01 -8.78339767e-01 -1.94269106e-01
1.73306063e-01 -4.62788671e-01 7.53725231e-01 3.75529021e-01
-8.45288634e-01 -6.24216080e-01 -8.47591579e-01 5.80002248e-01
-6.02586195e-02 2.70320505e-01 4.23405707e-01 4.58222359e-01
-5.64831793e-01 3.80661562e-02 -6.68631256e-01 -3.08785677e-01
7.78936803e-01 5.70785224e-01 -1.79667398e-01 -7.79389590e-03
-1.26690876e+00 7.32964039e-01 3.93812031e-01 1.93115562e-01
-1.01994026e+00 -3.21373314e-01 -8.08953643e-01 -9.70699713e-02
8.22020888e-01 -9.14939225e-01 1.15143657e+00 -1.25650263e+00
-1.74652159e+00 6.61021352e-01 -6.70314953e-02 -5.29776037e-01
5.65278053e-01 -3.79504383e-01 -4.86103326e-01 -1.30048376e-02
5.47772506e-03 5.40342391e-01 1.03397834e+00 -1.08281207e+00
-1.23200536e+00 -5.88937521e-01 -1.85134728e-02 7.57794321e-01
-3.19556966e-02 1.73403278e-01 -1.70218304e-01 -3.85780156e-01
-2.02035934e-01 -7.83559620e-01 -5.40409565e-01 -3.48485738e-01
-4.95591491e-01 -5.25457799e-01 6.23753309e-01 -6.47948265e-01
1.08746696e+00 -2.08186007e+00 1.39745161e-01 9.62540507e-02
1.24704083e-02 2.18944743e-01 -1.53797856e-02 9.80526656e-02
8.07265341e-02 -2.57844180e-01 -1.57116726e-01 -3.49684735e-03
6.88780546e-02 3.35235894e-01 -9.65795815e-02 4.46129322e-01
6.33208215e-01 6.11400604e-01 -7.46789217e-01 -7.10876524e-01
1.62563890e-01 1.84442237e-01 -5.05510151e-01 7.57047176e-01
-2.72114456e-01 5.44577360e-01 -8.60719323e-01 7.04611421e-01
8.06113362e-01 3.93177010e-02 9.89846438e-02 -3.01902026e-01
-1.33095115e-01 -2.86924571e-01 -1.44562137e+00 1.06519115e+00
-6.35919720e-02 2.56351829e-02 -1.22602314e-01 -1.46351731e+00
1.18967938e+00 6.72529936e-02 6.19844437e-01 -7.75596797e-01
9.34045985e-02 -2.05823869e-01 3.96533646e-02 -8.91142666e-01
3.15585852e-01 -3.42852920e-02 -2.95799851e-01 2.28371024e-01
-1.70060962e-01 2.59070426e-01 -6.55745994e-03 -2.21326977e-01
8.60648453e-01 1.82141855e-01 4.57648844e-01 5.35640642e-02
8.81346166e-01 -1.34732574e-03 1.27252412e+00 7.48980880e-01
-6.81882322e-01 2.64712691e-01 9.56289291e-01 -7.72179544e-01
-1.05315685e+00 -8.29346716e-01 1.53348237e-01 1.14920747e+00
4.13302630e-01 8.61553252e-02 -7.55676925e-01 -9.88338590e-01
-1.31290376e-01 3.76559436e-01 -6.29738092e-01 -3.06015491e-01
-4.96724337e-01 -1.19220161e+00 5.69276065e-02 6.50733173e-01
9.31098878e-01 -1.37737477e+00 -7.16245115e-01 2.40786269e-01
-2.31606036e-01 -1.24656451e+00 -2.75953323e-01 -3.52329831e-03
-3.75643939e-01 -1.08608353e+00 -4.96842355e-01 -8.18547487e-01
6.92657590e-01 -5.04799336e-02 1.08175540e+00 -3.00196588e-01
-9.96199772e-02 4.12088037e-01 -5.01506329e-01 -5.83509088e-01
-2.21567124e-01 1.05711482e-02 5.56347668e-02 5.92144370e-01
5.87858617e-01 -1.04673482e-01 -8.87405872e-01 7.59899139e-01
-5.25368392e-01 3.32841240e-02 9.38884616e-01 1.07595205e+00
8.75182688e-01 1.37331158e-01 9.12142813e-01 -7.96842098e-01
4.91410881e-01 -4.75475043e-01 -8.28895450e-01 4.15009767e-01
-3.58679652e-01 -1.27745986e-01 4.83066052e-01 -3.16602260e-01
-1.33423936e+00 4.07993078e-01 -1.10077951e-02 -2.18581155e-01
-4.84901577e-01 7.44461864e-02 -9.61541712e-01 3.34215760e-01
1.42937869e-01 4.35630560e-01 3.66568625e-01 1.03088608e-03
1.97911799e-01 8.45630646e-01 4.88098890e-01 -7.41712034e-01
4.73936766e-01 2.78184801e-01 -3.02472133e-02 -4.43727881e-01
-9.22523201e-01 -4.45239693e-01 -6.40290439e-01 -3.81393850e-01
1.18654084e+00 -1.00371468e+00 -1.31028879e+00 7.97210455e-01
-8.55146945e-01 -2.01893553e-01 -2.72368848e-01 4.90807503e-01
-9.87071633e-01 2.85124630e-01 -1.92960352e-01 -9.90347266e-01
-2.22262964e-01 -1.31824028e+00 7.91294575e-01 6.47062778e-01
3.47734958e-01 -4.42113727e-01 -1.07190102e-01 5.95370770e-01
-4.82683448e-04 4.65145946e-01 7.52682269e-01 -6.37002885e-01
-5.82387805e-01 -6.72556981e-02 -2.33427837e-01 4.46122527e-01
8.30652639e-02 6.56974986e-02 -6.35850072e-01 -1.07784815e-01
-1.74309209e-01 -3.87745708e-01 5.16704917e-01 4.46025252e-01
1.29692960e+00 -5.01145244e-01 -1.85470432e-01 4.57280546e-01
1.13058794e+00 7.59763181e-01 7.46468842e-01 6.27973258e-01
6.45468175e-01 7.82275975e-01 1.36162782e+00 8.07630658e-01
8.16320956e-01 8.00251305e-01 6.58468306e-01 -8.40843171e-02
2.61034787e-01 -9.41013172e-02 4.56182212e-01 4.51848984e-01
-1.01356626e-01 -2.12308601e-01 -8.92634153e-01 9.81730074e-02
-2.31312490e+00 -1.07639945e+00 1.16054356e-01 2.16670251e+00
5.83940923e-01 1.07866026e-01 6.43200934e-01 4.75590229e-02
9.01948154e-01 -6.16398230e-02 -8.52844536e-01 -3.83988798e-01
-2.37983137e-01 -5.60187340e-01 2.54438192e-01 1.77577689e-01
-1.60507584e+00 9.45201457e-01 5.56140614e+00 6.89508796e-01
-7.68649399e-01 -3.58176470e-01 1.28505170e+00 4.92078245e-01
2.86389172e-01 -2.92260528e-01 -9.71076190e-01 6.58152521e-01
6.22191548e-01 3.84975486e-02 2.22703174e-01 7.87679195e-01
3.50432485e-01 -2.05615357e-01 -9.13419127e-01 9.47048426e-01
1.20708626e-02 -7.84699082e-01 1.90869182e-01 -2.12109596e-01
4.69137013e-01 -4.04983371e-01 9.06493664e-02 4.41424370e-01
5.69208562e-01 -8.94371808e-01 4.89406794e-01 7.71224022e-01
4.54008013e-01 -1.20422912e+00 9.57441747e-01 2.54541725e-01
-1.06893098e+00 -5.33969045e-01 -5.49385726e-01 3.07149850e-02
-5.51640755e-03 9.50484872e-02 -5.72156668e-01 6.14711285e-01
6.98635399e-01 1.00000513e+00 -5.34198582e-01 1.21957564e+00
-2.53648758e-01 4.97984409e-01 1.79646522e-01 -8.06737766e-02
1.19109847e-01 -5.54716527e-01 4.57620472e-01 9.28703725e-01
-8.97403210e-02 3.70143503e-01 6.16843045e-01 3.09071541e-01
2.74294555e-01 6.80525526e-02 -2.55516291e-01 4.04393911e-01
5.67461014e-01 1.20808327e+00 -5.05067468e-01 -4.25713062e-01
-3.46804768e-01 8.08007896e-01 2.64842838e-01 3.38753611e-01
-9.81398344e-01 -1.98184609e-01 7.58888662e-01 -2.51355529e-01
5.28786302e-01 1.94627792e-01 -1.20963335e-01 -9.42273080e-01
7.81394541e-03 -1.21927822e+00 6.78352952e-01 -5.59321582e-01
-1.40761077e+00 5.59292197e-01 -6.55661076e-02 -1.58528376e+00
-2.08875000e-01 -2.23984450e-01 -5.66824496e-01 7.37291813e-01
-1.76610589e+00 -1.11590064e+00 -5.41819096e-01 8.23451996e-01
5.64760745e-01 -4.26850736e-01 5.90057015e-01 1.93129808e-01
-1.14462459e+00 5.24162531e-01 -1.32047504e-01 2.20124722e-01
6.04109347e-01 -1.32332933e+00 4.52736728e-02 6.61135435e-01
-5.09859324e-01 -1.54952645e-01 7.54208088e-01 -4.81184125e-01
-1.30031633e+00 -1.28579521e+00 3.13390553e-01 -2.56808192e-01
3.89694780e-01 -2.17527583e-01 -7.22026289e-01 4.79203314e-01
3.43508005e-01 1.24817371e-01 7.88054705e-01 -7.66521841e-02
1.14694489e-02 -5.49556434e-01 -1.11354923e+00 4.23008442e-01
8.54203701e-01 1.17296658e-01 -5.03608525e-01 3.16249222e-01
5.30942798e-01 -2.80118883e-01 -8.88094544e-01 3.96459311e-01
3.36386561e-01 -8.10955048e-01 9.60123718e-01 -9.47874963e-01
3.46027881e-01 -5.00311434e-01 -6.85124770e-02 -1.46858406e+00
-4.14622337e-01 -5.60181975e-01 -9.74631682e-03 1.45695746e+00
-1.06529042e-01 -4.85969603e-01 8.45094562e-01 6.56171203e-01
-2.29176600e-02 -8.53267670e-01 -8.47814023e-01 -5.49360275e-01
-2.99931496e-01 1.70775831e-01 9.01408613e-01 7.20293641e-01
-3.60613376e-01 3.83365601e-01 -6.22325778e-01 4.09098744e-01
8.50007415e-01 1.31843820e-01 9.78980124e-01 -1.10566449e+00
-1.47917688e-01 -2.64886618e-01 -5.75763464e-01 -8.19655895e-01
3.11390698e-01 -1.52257234e-01 3.04830343e-01 -1.33630538e+00
3.03667217e-01 -6.99289680e-01 -5.89318633e-01 4.50820893e-01
-6.61982715e-01 -2.23254263e-01 1.69623032e-01 1.46972150e-01
-1.21914744e+00 9.21898186e-01 1.20418179e+00 -2.09128797e-01
-5.15076816e-02 4.90847349e-01 -6.08349144e-01 6.63697004e-01
8.68049681e-01 -3.24468076e-01 -3.53297055e-01 -1.09468713e-01
-8.79362077e-02 4.24851507e-01 2.93626785e-01 -1.05341125e+00
2.10395843e-01 -6.35075271e-01 3.93276870e-01 -6.81011379e-01
2.13994741e-01 -9.12294507e-01 -3.25488597e-02 5.81798255e-01
-3.00654620e-01 3.05136949e-01 -1.62151262e-01 7.14118004e-01
-3.69734466e-01 2.39688307e-01 7.72234976e-01 8.21545795e-02
-1.18754375e+00 5.94722748e-01 -3.65296632e-01 1.65651783e-01
1.60176563e+00 -2.08142832e-01 -2.50551105e-01 -2.23045349e-01
-7.75887012e-01 7.98515797e-01 -3.90278129e-03 3.67158979e-01
7.19629586e-01 -1.41611433e+00 -9.15688694e-01 1.92494571e-01
1.24490947e-01 6.87633902e-02 5.49095035e-01 6.11833692e-01
-1.97909415e-01 -2.82118674e-02 -7.73472011e-01 -7.54289865e-01
-1.33456588e+00 6.37688994e-01 4.58925128e-01 -3.34834456e-01
-3.77690494e-01 5.41504681e-01 3.19218159e-01 -3.05060267e-01
4.11577225e-01 3.40284526e-01 -9.60325599e-01 3.49515319e-01
4.47674960e-01 4.03141290e-01 -2.23069265e-01 -6.70417488e-01
-3.82168323e-01 5.17489851e-01 -2.27412760e-01 3.95805746e-01
1.41444314e+00 -3.76259774e-01 1.56715289e-01 9.16552171e-02
6.27454638e-01 -7.32555389e-01 -1.90110373e+00 -2.40146443e-01
-1.63730402e-02 -3.98590744e-01 -4.31879401e-01 -7.96703637e-01
-1.01350653e+00 6.79649234e-01 8.51166785e-01 1.34449303e-01
1.37869501e+00 -1.43802628e-01 4.37358946e-01 2.55160540e-01
1.38666332e-02 -1.29334128e+00 3.93403292e-01 3.34569305e-01
7.05679119e-01 -1.51121354e+00 -1.08353771e-01 -4.11842018e-01
-1.30924237e+00 8.77294540e-01 1.10319924e+00 -1.87924087e-01
4.50567693e-01 -5.44047821e-03 2.30178908e-01 1.78162053e-01
-7.87556827e-01 -1.89302683e-01 1.13831619e-02 8.92993689e-01
-9.84753892e-02 2.30611160e-01 -3.26860100e-01 6.84367716e-01
7.23702610e-02 -3.52467336e-02 5.00663996e-01 9.53796625e-01
-4.38740522e-01 -1.13062012e+00 -3.46391767e-01 5.39979041e-01
-3.64457965e-01 3.87993544e-01 6.52936324e-02 4.71605450e-01
3.30708534e-01 9.02278244e-01 2.52929807e-01 -3.99782211e-01
5.27577281e-01 -3.53125066e-01 5.89793213e-02 -2.84449130e-01
-4.62157488e-01 -1.13977209e-01 1.85117647e-02 -6.03128076e-01
-6.30992711e-01 -1.05496609e+00 -8.54619265e-01 -7.80881494e-02
-3.89059111e-02 1.42303199e-01 3.28335822e-01 8.86248827e-01
3.44518036e-01 8.48969877e-01 1.15857470e+00 -4.34111089e-01
-7.80890584e-01 -6.03619456e-01 -4.21156466e-01 6.00668728e-01
1.91469654e-01 -8.67353976e-01 1.45571247e-01 1.32019907e-01] | [14.315349578857422, 1.032454013824463] |
e3f393f8-d2e6-47a1-85a5-8643d4070fee | past-visions-of-artificial-futures-one | 1806.01322 | null | http://arxiv.org/abs/1806.01322v1 | http://arxiv.org/pdf/1806.01322v1.pdf | Past Visions of Artificial Futures: One Hundred and Fifty Years under the Spectre of Evolving Machines | The influence of Artificial Intelligence (AI) and Artificial Life (ALife)
technologies upon society, and their potential to fundamentally shape the
future evolution of humankind, are topics very much at the forefront of current
scientific, governmental and public debate. While these might seem like very
modern concerns, they have a long history that is often disregarded in
contemporary discourse. Insofar as current debates do acknowledge the history
of these ideas, they rarely look back further than the origin of the modern
digital computer age in the 1940s-50s. In this paper we explore the earlier
history of these concepts. We focus in particular on the idea of
self-reproducing and evolving machines, and potential implications for our own
species. We show that discussion of these topics arose in the 1860s, within a
decade of the publication of Darwin's The Origin of Species, and attracted
increasing interest from scientists, novelists and the general public in the
early 1900s. After introducing the relevant work from this period, we
categorise the various visions presented by these authors of the future
implications of evolving machines for humanity. We suggest that current debates
on the co-evolution of society and technology can be enriched by a proper
appreciation of the long history of the ideas involved. | ['Tim Taylor', 'Alan Dorin'] | 2018-06-04 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 3.69583130e-01 2.80448794e-01 9.65661705e-02 -7.41870254e-02
5.75531065e-01 -7.66507745e-01 1.20649374e+00 1.57604188e-01
-7.35137463e-01 7.85883546e-01 3.40918839e-01 -7.52910376e-01
-1.89401016e-01 -6.54518902e-01 -4.16960657e-01 -6.07780218e-01
7.12592602e-02 2.44400263e-01 -4.79248837e-02 -8.28160644e-01
5.68667650e-01 3.84934962e-01 -1.84854484e+00 -3.92299712e-01
1.06230831e+00 4.12222385e-01 2.01367348e-01 8.74522209e-01
-1.37862310e-01 4.10611629e-01 -8.48690867e-01 -6.25395775e-01
-8.42889845e-02 -7.77903795e-01 -6.95702553e-01 -1.36413887e-01
-2.37923667e-01 8.10673758e-02 -2.42830023e-01 7.75262713e-01
3.08441520e-01 -3.58107328e-01 3.12009662e-01 -9.14065599e-01
-9.40063655e-01 5.88423193e-01 -5.18743657e-02 2.05122530e-01
7.41009861e-02 2.86264181e-01 6.03579044e-01 -2.14319661e-01
7.12693095e-01 1.03943908e+00 4.96868551e-01 5.39089322e-01
-7.61630416e-01 -3.66066456e-01 1.11007392e-02 1.85259238e-01
-1.27871633e+00 -6.86243176e-01 5.38184762e-01 -6.23022795e-01
7.22834945e-01 4.55050468e-01 1.52062142e+00 7.07659841e-01
6.52152836e-01 4.76829797e-01 7.85791278e-01 -7.53037095e-01
1.82207853e-01 2.49053866e-01 -1.66245997e-01 4.91088986e-01
7.48219907e-01 1.52396321e-01 -2.25672111e-01 -2.86214918e-01
6.25389934e-01 -3.69038820e-01 -3.01834106e-01 -2.87813246e-02
-1.57166660e+00 7.08706915e-01 -2.64371596e-02 9.75362062e-01
-1.67273536e-01 1.97182298e-01 3.53975594e-01 3.08328211e-01
2.38998950e-01 7.03589141e-01 -3.75025392e-01 -5.40335238e-01
-2.59280026e-01 8.59607086e-02 9.19753790e-01 2.99678922e-01
2.36423969e-01 -1.96165368e-01 7.86708891e-01 4.61222529e-01
3.31186235e-01 2.57035196e-01 5.80786526e-01 -9.34671760e-01
-3.53537977e-01 5.99706054e-01 1.80813178e-01 -9.92848396e-01
-3.07217717e-01 -3.90849322e-01 -5.58435082e-01 7.41957873e-02
4.35273945e-01 -3.45852256e-01 -3.92330468e-01 1.38875961e+00
3.68439138e-01 -2.10910350e-01 1.71461880e-01 6.43172383e-01
4.79471833e-01 6.99494481e-01 7.56407483e-03 -3.55721086e-01
1.21245670e+00 -4.23405915e-01 -4.85549212e-01 -5.97915128e-02
5.30863404e-01 -7.26828694e-01 4.74094272e-01 3.45253408e-01
-1.05333972e+00 -1.60132587e-01 -1.19011557e+00 1.95894137e-01
-5.32521248e-01 -5.72012424e-01 1.03729498e+00 1.31883478e+00
-1.03488100e+00 7.25214839e-01 -8.82871389e-01 -1.17899776e+00
9.27042961e-02 1.66575775e-01 2.08750680e-01 5.11083901e-01
-1.10829973e+00 1.26694846e+00 2.84064054e-01 8.85586664e-02
4.32104915e-02 -3.57160419e-01 -2.26421267e-01 -5.67616403e-01
3.66571844e-01 -9.66345668e-01 1.06136739e+00 -1.34019804e+00
-1.41602397e+00 1.27280700e+00 5.26055954e-02 -5.54612041e-01
5.39468348e-01 1.43463509e-02 -8.72800171e-01 -5.08624837e-02
-4.18361962e-01 1.76057205e-01 1.11151159e-01 -1.08139944e+00
-9.49483156e-01 -4.44529772e-01 -6.20453432e-03 1.09428346e-01
-1.89159378e-01 3.18826973e-01 4.77655008e-02 -6.02729082e-01
-7.30532333e-02 -1.21360254e+00 -1.79417223e-01 1.24680586e-01
2.86368072e-01 -4.74581867e-01 4.58133519e-01 -3.85176659e-01
1.09281802e+00 -2.06774306e+00 1.21979304e-01 -2.36661389e-01
2.67653406e-01 2.75054157e-01 2.39480108e-01 7.67808020e-01
1.89080924e-01 4.15945500e-01 -3.00264001e-01 6.25633538e-01
-8.81538764e-02 3.38309944e-01 3.48647833e-02 7.66173244e-01
-9.83751565e-02 1.05015635e+00 -1.15467453e+00 -2.99474131e-02
8.26729387e-02 5.96950293e-01 -1.63985476e-01 -2.34034896e-01
1.05977848e-01 3.03795040e-01 -3.40080500e-01 4.30785179e-01
-1.39790839e-02 -2.07616895e-01 2.80404925e-01 3.44423175e-01
-6.70793414e-01 4.17529084e-02 -3.54026496e-01 1.46514010e+00
-2.02130228e-02 9.27260339e-01 1.44127822e-02 -9.45145488e-01
8.37954044e-01 4.11599576e-01 4.42093253e-01 -5.78907728e-01
3.84476334e-01 5.96077085e-01 9.55731750e-01 -5.59911489e-01
5.75387359e-01 -5.18983722e-01 7.41670802e-02 6.94115996e-01
-4.65920389e-01 -6.65598750e-01 1.02963053e-01 -9.69300698e-03
8.92690182e-01 3.39969516e-01 6.91224694e-01 -6.20472789e-01
5.20933926e-01 -3.77161689e-02 3.92697334e-01 2.84421206e-01
-4.86792386e-01 1.97765887e-01 1.11624151e-01 -7.59706259e-01
-1.34877849e+00 -8.13160419e-01 -4.02497888e-01 7.87961006e-01
3.67271125e-01 -2.91815013e-01 -6.02323532e-01 -1.14455745e-01
-2.51758192e-03 1.15548444e+00 -7.88351893e-01 -3.52721214e-01
-5.49371779e-01 -9.94615018e-01 3.36473495e-01 -2.77362764e-01
3.44463199e-01 -1.07349682e+00 -1.43153572e+00 2.21832767e-01
2.83089638e-01 -7.26153374e-01 3.95314932e-01 -6.84107542e-02
-7.45493889e-01 -8.21534514e-01 -9.06898022e-01 -5.55554748e-01
4.66351390e-01 -1.19448244e-03 9.27867770e-01 2.73573995e-01
-4.68355477e-01 7.22359836e-01 -3.52869600e-01 -9.02767956e-01
-1.24128175e+00 -1.33137941e-01 1.82438362e-03 -4.36724782e-01
3.81809175e-01 -7.33332634e-01 -5.71934700e-01 2.23641656e-02
-9.09213245e-01 3.01077127e-01 3.76663566e-01 3.78121585e-01
-4.36951756e-01 -7.98866749e-02 8.91343296e-01 -7.84733772e-01
3.94833982e-01 -5.66973209e-01 -6.95387349e-02 5.36062717e-01
-5.75639248e-01 -1.14985317e-01 2.18196988e-01 -2.71028340e-01
-9.09591317e-01 -6.37210906e-01 -1.17710173e-01 6.53387845e-01
-1.49368554e-01 3.19490314e-01 1.53714389e-01 -9.88261476e-02
6.43143713e-01 3.37479413e-01 1.63831413e-01 -3.14434230e-01
3.60483468e-01 9.80696738e-01 5.46203494e-01 -5.58406234e-01
6.42585158e-01 5.40961206e-01 -2.29630962e-01 -1.50801134e+00
-8.91244933e-02 1.45060956e-01 -4.76523578e-01 -4.74316001e-01
7.02244222e-01 -4.60832804e-01 -5.20735145e-01 4.84908700e-01
-1.15648329e+00 -5.12673557e-02 -5.47419310e-01 2.97810972e-01
-4.39047873e-01 2.96625137e-01 -1.33307621e-01 -1.15196419e+00
-3.92171651e-01 -6.12370491e-01 4.14336026e-01 4.13022757e-01
-5.35055518e-01 -1.25119293e+00 2.04685599e-01 3.31015080e-01
5.21456778e-01 6.68765545e-01 1.18182385e+00 -4.06846404e-01
1.35220706e-01 -1.04914740e-01 3.52654427e-01 3.49348150e-02
4.97796386e-01 3.85750979e-01 -6.79271102e-01 9.81417298e-03
2.00725004e-01 1.38863355e-01 3.56042683e-01 -1.38615176e-01
1.91242397e-01 -3.26635331e-01 -3.43675077e-01 1.60745054e-01
1.49931741e+00 9.77906346e-01 6.87023640e-01 7.22641826e-01
2.23755747e-01 9.68146920e-01 2.86215674e-02 5.09328544e-01
5.17212152e-01 9.06445906e-02 9.04026031e-02 7.07765296e-02
4.35510762e-02 2.48643816e-01 2.80860424e-01 1.30209899e+00
-5.05080581e-01 -1.98339403e-01 -1.29688764e+00 7.38866687e-01
-1.66759849e+00 -7.41831005e-01 -1.02689244e-01 2.28128290e+00
5.94292939e-01 1.78691179e-01 4.33632791e-01 4.67785671e-02
9.93370056e-01 -1.06851727e-01 -6.80133462e-01 -7.88977623e-01
-3.95376801e-01 5.50725982e-02 1.26331970e-01 2.08989784e-01
-4.99706447e-01 7.29279041e-01 7.48015499e+00 1.84492812e-01
-1.08358145e+00 -1.61379278e-01 6.05807781e-01 3.15208919e-02
-6.66454554e-01 1.98561296e-01 -2.98812240e-01 5.81484199e-01
1.14224756e+00 -1.03321862e+00 3.05285394e-01 1.94065273e-01
6.56291768e-02 -3.79751593e-01 -7.42996454e-01 5.48465133e-01
1.05730742e-01 -9.61637378e-01 -2.07329303e-01 2.71049291e-01
7.52055407e-01 1.34594500e-01 -9.68029648e-02 -2.43674099e-01
4.59304273e-01 -1.04048896e+00 9.41717625e-01 2.59532988e-01
3.20655823e-01 -9.13766861e-01 3.27962160e-01 3.30398977e-01
-5.26411295e-01 -2.29098350e-01 -2.58968830e-01 -5.96277654e-01
2.47724503e-01 6.04139626e-01 -3.69534552e-01 3.19173634e-01
2.20785171e-01 3.60328287e-01 -3.85509402e-01 9.68851328e-01
3.81036587e-02 5.01713097e-01 -3.92337739e-01 -6.13390088e-01
2.66082346e-01 -2.05282509e-01 7.73650885e-01 1.00049078e+00
3.39149773e-01 4.31573778e-01 -7.56003976e-01 7.28170693e-01
4.98814940e-01 -9.21476446e-03 -4.54342693e-01 -8.43793988e-01
3.19893301e-01 9.52124596e-01 -9.88487780e-01 -2.41860524e-01
-5.31602502e-01 9.11032081e-01 -3.65119249e-01 2.67800856e-02
-8.15421462e-01 -3.91685486e-01 5.93425989e-01 4.31290209e-01
5.09345382e-02 -3.50968003e-01 -3.28007698e-01 -1.07058871e+00
-3.40493798e-01 -8.27736855e-01 -5.31038977e-02 -1.49150848e-01
-7.34009862e-01 3.12899381e-01 -2.52396196e-01 -4.21095878e-01
-8.01564306e-02 -3.19954157e-01 -5.33516467e-01 3.36924464e-01
-8.48443985e-01 -9.07705367e-01 2.77771771e-01 -2.34371006e-01
1.37964994e-01 -1.02755263e-01 8.57522964e-01 -2.62015283e-01
-4.47394103e-01 1.57389268e-01 7.50269651e-01 -3.59745383e-01
2.56960124e-01 -7.79222488e-01 9.98724818e-01 5.15081227e-01
-2.19374865e-01 7.26423860e-01 1.02859640e+00 -5.76124787e-01
-1.62605739e+00 -1.81995556e-01 1.08722174e+00 -3.44101757e-01
9.28192735e-01 -1.02401823e-01 -4.47445512e-01 4.30480003e-01
5.76170921e-01 -8.85037661e-01 7.01655805e-01 -3.32718417e-02
3.10795635e-01 3.00854146e-01 -9.52405512e-01 9.02412593e-01
1.17252755e+00 -8.52224231e-02 -8.79031360e-01 -7.69761810e-03
7.08335042e-01 4.74072218e-01 -8.41483951e-01 3.40560436e-01
1.30819440e+00 -1.01141238e+00 6.29678071e-01 -1.93516046e-01
2.57674694e-01 -9.44420323e-02 2.03902468e-01 -8.73054147e-01
-5.25006831e-01 -1.02193809e+00 4.64686751e-01 1.14460480e+00
2.08675295e-01 -1.14060187e+00 4.75466371e-01 7.30311453e-01
2.13948879e-02 -7.62043178e-01 -9.88106489e-01 -5.56449711e-01
5.25849998e-01 -3.96292433e-02 6.19829178e-01 1.21784949e+00
5.44923723e-01 2.43931457e-01 1.57559723e-01 -6.06773734e-01
4.09056216e-01 -8.82860869e-02 7.54314899e-01 -1.53550625e+00
-1.99025780e-01 -8.22597742e-01 -9.35258925e-01 -3.58537197e-01
-4.91466820e-01 -7.05270290e-01 7.46599883e-02 -1.50457668e+00
1.27667710e-01 -1.04353577e-01 -9.05109793e-02 -1.58378467e-01
-1.11544160e-02 7.84346983e-02 3.75089347e-01 3.98842156e-01
-1.59575582e-01 3.05492491e-01 1.26871705e+00 3.74873281e-01
-1.97255716e-01 -3.72419178e-01 -1.15740848e+00 9.80659604e-01
9.58219051e-01 -3.35184149e-02 -2.02255934e-01 -2.04530627e-01
6.92860842e-01 -2.95889229e-01 -7.00996071e-03 -1.22243190e+00
1.31162345e-01 -4.55988944e-01 2.02737063e-01 1.25897273e-01
1.78587198e-01 -7.71520734e-01 6.80030167e-01 1.06970859e+00
4.70258035e-02 -2.70823594e-02 1.44666374e-01 1.83893755e-01
4.07922179e-01 -1.26553461e-01 8.14995646e-01 -4.86730576e-01
-5.59428394e-01 -4.03798699e-01 -9.32602048e-01 -1.49964571e-01
1.36333466e+00 -5.69322824e-01 -3.97096395e-01 -7.83181041e-02
-5.68208814e-01 4.23888713e-02 1.06401396e+00 5.36082387e-01
1.32005364e-01 -7.14081466e-01 -7.96131849e-01 -1.24291077e-01
6.00543246e-03 -4.14897203e-01 -4.14410010e-02 4.69587296e-01
-9.53652263e-01 4.11125004e-01 -4.54602420e-01 -4.48522493e-02
-1.12422967e+00 5.38769722e-01 2.09179237e-01 4.88213420e-01
-7.55992413e-01 7.46685445e-01 4.54871774e-01 -3.12366411e-02
-1.85329735e-01 1.50510907e-01 -2.28571355e-01 -1.27894163e-01
4.69865441e-01 7.04900980e-01 -7.32258260e-01 -8.88949633e-01
-4.69836682e-01 5.35983801e-01 1.20901793e-01 -4.28188950e-01
1.50686145e+00 -4.09115046e-01 -6.24311447e-01 1.19695425e+00
8.22728395e-01 -6.47133887e-02 -4.29639995e-01 2.35640854e-01
7.52722472e-03 -1.30486935e-01 -4.38744992e-01 -1.00966096e+00
-4.92544234e-01 6.20612204e-01 4.68334734e-01 4.98703599e-01
8.97039175e-01 4.78128940e-02 8.10991526e-01 1.00874521e-01
3.14878970e-01 -1.01685452e+00 -5.38504243e-01 8.15476030e-02
8.89038563e-01 -5.93505383e-01 2.61867285e-01 -2.10262954e-01
-3.14741552e-01 1.23346961e+00 2.41874084e-01 4.06150296e-02
5.90792298e-01 5.97527549e-02 -1.33437574e-01 3.57692689e-02
-7.77064443e-01 -1.42907321e-01 -1.49573117e-01 5.01447797e-01
1.01838732e+00 2.35345408e-01 -1.27073991e+00 1.40700549e-01
-5.60206115e-01 2.15784475e-01 7.35027432e-01 1.31170380e+00
-1.01187301e+00 -1.12979543e+00 -6.31184995e-01 1.22384802e-01
-4.25311267e-01 2.16085061e-01 -9.93273675e-01 6.48552060e-01
6.74073100e-01 7.61411607e-01 1.49589837e-01 -3.36754441e-01
-1.40262723e-01 1.48662686e-01 8.72246742e-01 -1.52285010e-01
-7.44057000e-01 1.18969511e-02 2.55071670e-01 1.42757654e-01
-4.22155499e-01 -1.00534582e+00 -1.10126185e+00 -8.42313111e-01
-3.86066914e-01 5.39115191e-01 1.01896417e+00 9.39909339e-01
2.63833165e-01 2.33674720e-01 1.75184473e-01 -5.05006075e-01
-2.83499137e-02 -6.02473080e-01 -7.46964991e-01 -2.95155048e-01
8.87216628e-02 -5.08795306e-02 -3.98438245e-01 1.79793060e-01] | [5.630260944366455, 4.216994285583496] |
e69e0432-5676-4366-9a1f-844dfdf8ec15 | provably-efficient-causal-model-based | 2202.06545 | null | https://arxiv.org/abs/2202.06545v3 | https://arxiv.org/pdf/2202.06545v3.pdf | Provably Efficient Causal Model-Based Reinforcement Learning for Systematic Generalization | In the sequential decision making setting, an agent aims to achieve systematic generalization over a large, possibly infinite, set of environments. Such environments are modeled as discrete Markov decision processes with both states and actions represented through a feature vector. The underlying structure of the environments allows the transition dynamics to be factored into two components: one that is environment-specific and another that is shared. Consider a set of environments that share the laws of motion as an example. In this setting, the agent can take a finite amount of reward-free interactions from a subset of these environments. The agent then must be able to approximately solve any planning task defined over any environment in the original set, relying on the above interactions only. Can we design a provably efficient algorithm that achieves this ambitious goal of systematic generalization? In this paper, we give a partially positive answer to this question. First, we provide a tractable formulation of systematic generalization by employing a causal viewpoint. Then, under specific structural assumptions, we provide a simple learning algorithm that guarantees any desired planning error up to an unavoidable sub-optimality term, while showcasing a polynomial sample complexity. | ['Marcello Restelli', 'Michael Bronstein', 'Juan Felipe Calderon', 'Emanuele Rossi', 'Riccardo De Santi', 'Mirco Mutti'] | 2022-02-14 | null | null | null | null | ['systematic-generalization'] | ['reasoning'] | [ 3.69893104e-01 5.26687980e-01 -2.33176574e-01 -1.86500654e-01
-5.84367216e-01 -6.53052866e-01 7.85553873e-01 2.33647808e-01
-4.74191129e-01 8.38637054e-01 1.13693036e-01 -3.91039222e-01
-4.57258821e-01 -1.02749836e+00 -8.09391916e-01 -8.84408057e-01
-4.79589760e-01 9.62031364e-01 2.65277565e-01 -1.74064040e-02
2.19808131e-01 5.90474606e-01 -1.32249892e+00 -2.19830111e-01
7.68132925e-01 6.33378923e-01 5.00485659e-01 1.03930652e+00
1.25091925e-01 9.38972533e-01 -1.59915119e-01 1.74731895e-01
4.54935074e-01 -6.13909185e-01 -1.13231146e+00 7.70723164e-01
-3.65110695e-01 -3.59606594e-01 -3.71359140e-01 1.11118829e+00
5.23405634e-02 4.98823613e-01 6.84816241e-01 -1.20109141e+00
-4.04392362e-01 8.07831168e-01 7.48269856e-02 -3.19838431e-03
5.02974570e-01 5.07895231e-01 9.65643883e-01 -2.29406759e-01
7.28372097e-01 1.22295988e+00 -6.31706566e-02 6.48282766e-01
-1.55724454e+00 -5.88763170e-02 5.74110329e-01 4.63131107e-02
-1.07317293e+00 -2.73102701e-01 4.78033483e-01 -5.24724364e-01
8.86852264e-01 1.63878828e-01 6.87264919e-01 8.64808857e-01
3.63346428e-01 7.74180472e-01 1.24211490e+00 -4.99923050e-01
1.01229990e+00 -1.39234369e-04 1.47382781e-01 5.48174620e-01
1.93526924e-01 5.09766579e-01 -2.91909963e-01 -2.43521214e-01
7.06063628e-01 1.97772428e-01 -1.36506528e-01 -7.22951114e-01
-1.08958507e+00 7.37722576e-01 1.83060527e-01 1.09870322e-01
-4.79919165e-01 2.97278404e-01 4.72739339e-02 7.33620346e-01
-1.29207477e-01 6.46557927e-01 -5.21003485e-01 3.92945223e-02
-2.39085466e-01 5.57410240e-01 1.07686090e+00 1.05061197e+00
8.94615054e-01 -1.50625885e-01 6.78435266e-02 -1.19195469e-01
2.47798592e-01 4.76682544e-01 1.10803731e-01 -1.28131080e+00
4.15508956e-01 1.67101622e-01 8.33238006e-01 -5.34223378e-01
-4.99158114e-01 -2.62586564e-01 -4.17186677e-01 3.87192428e-01
4.75391448e-01 -3.34938258e-01 -7.88324893e-01 2.18810749e+00
3.91797692e-01 1.61977336e-02 3.10083270e-01 5.50518572e-01
-3.77403945e-01 6.81706369e-01 -7.89175257e-02 -7.15287268e-01
8.14639211e-01 -8.89786541e-01 -4.74828392e-01 -4.05040234e-01
7.41607130e-01 -1.02394268e-01 8.80851090e-01 3.57502371e-01
-1.05654824e+00 -4.80034091e-02 -1.06594920e+00 4.36777115e-01
1.89798653e-01 -4.97999400e-01 6.99375272e-01 4.00627911e-01
-1.08196545e+00 7.68190503e-01 -1.19038105e+00 -4.42878842e-01
-4.82455920e-03 4.46071118e-01 -1.82842806e-01 -6.99570701e-02
-8.36223364e-01 8.94428790e-01 5.56515574e-01 -6.36993572e-02
-1.65108252e+00 4.24330533e-02 -5.99401414e-01 7.54937157e-02
1.09899211e+00 -8.94897044e-01 1.81499684e+00 -8.36985469e-01
-1.59955072e+00 2.23679721e-01 -3.01033348e-01 -6.05311394e-01
6.02199554e-01 -3.27138901e-02 4.68899906e-02 8.33368450e-02
1.80565119e-01 2.21921634e-02 7.42425442e-01 -1.31236029e+00
-9.00841057e-01 -5.68576336e-01 6.19540811e-01 5.21490812e-01
7.70479590e-02 -8.68957415e-02 7.96922594e-02 3.45420204e-02
2.52912402e-01 -1.40184844e+00 -1.18665922e+00 -3.95120025e-01
-4.26243365e-01 -1.80125386e-01 1.36607543e-01 -2.36336314e-06
7.91271746e-01 -1.84025228e+00 5.27357817e-01 2.35533655e-01
1.57194778e-01 -4.10218060e-01 -1.33857438e-02 7.65795767e-01
3.40431005e-01 2.20180631e-01 -3.34429950e-01 -4.48782116e-01
2.10744202e-01 4.44905370e-01 -4.32386637e-01 5.42894959e-01
-1.98174685e-01 6.37979507e-01 -1.14132249e+00 -6.71690404e-02
8.11464489e-02 -3.59857947e-01 -6.60777509e-01 2.36711562e-01
-7.43857920e-01 5.91969669e-01 -1.13716066e+00 2.02847421e-01
2.89447278e-01 -1.20316990e-01 5.54544449e-01 9.46161151e-01
-2.69164443e-01 3.64408404e-01 -1.53295314e+00 1.57025826e+00
-4.01744634e-01 1.04752868e-01 9.28891972e-02 -8.65519941e-01
4.37151313e-01 2.58776248e-01 4.20086324e-01 -1.02357781e-02
1.25373036e-01 2.33676895e-01 1.14662267e-01 -4.29422826e-01
4.68486875e-01 -4.98263001e-01 -4.63449150e-01 7.78486252e-01
-7.17131793e-02 -1.02530137e-01 -6.59457466e-04 1.66751340e-01
1.44630826e+00 -3.17051522e-02 5.55347919e-01 -2.37739399e-01
2.19161674e-01 1.87468380e-01 6.98917568e-01 1.41990197e+00
-3.35135192e-01 -5.92699684e-02 6.25040114e-01 -4.68371034e-01
-1.01529860e+00 -1.05096340e+00 3.17855507e-01 9.75385487e-01
3.69904816e-01 -2.95197368e-01 -5.54148495e-01 -4.09434944e-01
-2.62170970e-01 9.13587928e-01 -7.79676735e-01 -1.16482817e-01
-4.67915744e-01 -3.73890162e-01 -1.69247389e-01 3.16192597e-01
5.17052054e-01 -1.04187095e+00 -1.12047768e+00 4.24841881e-01
1.33959159e-01 -8.81765544e-01 -3.83380592e-01 6.36557162e-01
-9.78880465e-01 -7.93094039e-01 -1.02170080e-01 -5.21683693e-01
5.78234851e-01 1.95352793e-01 6.17676437e-01 -4.22361493e-01
3.35966945e-01 6.67317808e-01 -1.49405926e-01 -3.58315140e-01
-5.74917436e-01 1.09673487e-02 4.21299696e-01 -7.69960657e-02
8.45004618e-02 -7.31870055e-01 -2.19546229e-01 1.07989106e-02
-8.27114999e-01 1.64572671e-01 5.61163127e-01 5.53958952e-01
5.00994921e-01 6.25127375e-01 2.74411052e-01 -6.10698998e-01
5.15753090e-01 -4.94529456e-01 -8.52194905e-01 2.53288805e-01
-4.20212328e-01 4.75602448e-01 7.36294925e-01 -4.58642870e-01
-1.18155563e+00 3.84793758e-01 4.12366569e-01 -1.09803393e-01
-4.20560837e-01 6.75890923e-01 -6.41780257e-01 3.62928838e-01
6.34505868e-01 5.46719193e-01 -8.76456425e-02 -2.80723929e-01
4.13291097e-01 2.39309609e-01 4.26688254e-01 -9.90888417e-01
8.23588371e-01 4.32592332e-01 3.84239823e-01 -7.41925120e-01
-7.41253138e-01 -5.13825044e-02 -6.32949054e-01 -1.84278294e-01
6.14740074e-01 -5.29797554e-01 -8.47243488e-01 2.49368176e-01
-8.94202590e-01 -8.42677891e-01 -6.19568586e-01 4.18830067e-01
-1.27392018e+00 1.45439014e-01 -1.94928080e-01 -1.30894935e+00
3.91065508e-01 -1.19630516e+00 5.23811758e-01 1.68573886e-01
-5.25509939e-04 -8.42811942e-01 2.79239178e-01 -1.46178678e-01
1.72174066e-01 2.20529646e-01 8.88914466e-01 -9.47804034e-01
-1.09658277e+00 -4.13882174e-02 5.71456671e-01 -8.31161439e-03
1.45729497e-01 -2.85149068e-01 -5.61353981e-01 -3.68419319e-01
3.62785846e-01 -1.06674261e-01 5.76645374e-01 3.66650254e-01
8.01992476e-01 -6.65767729e-01 -5.29969037e-01 1.96985200e-01
1.38299322e+00 7.38559186e-01 1.00266799e-01 5.89804947e-01
2.89842755e-01 6.29276931e-01 5.70929646e-01 6.41945481e-01
5.00050962e-01 4.96064007e-01 6.22352481e-01 7.61252999e-01
6.08041048e-01 -3.59170526e-01 3.98336083e-01 4.19483259e-02
-1.82735920e-01 -1.98171005e-01 -9.50122058e-01 7.22493052e-01
-2.20021367e+00 -1.18091226e+00 3.18580091e-01 2.54672217e+00
6.60941482e-01 2.16505438e-01 1.34207457e-01 -1.24936849e-01
5.75642586e-01 -7.13849664e-02 -9.91032839e-01 -3.99592280e-01
8.91311914e-02 -2.43563071e-01 5.87660253e-01 8.89263332e-01
-9.68014121e-01 8.17447901e-01 6.51271296e+00 2.06443846e-01
-6.90753937e-01 6.62586614e-02 3.63016456e-01 -7.93541968e-02
-5.23797035e-01 4.75376189e-01 -6.76855743e-01 1.53683096e-01
1.26903403e+00 -6.65401936e-01 8.04505348e-01 1.04317582e+00
3.77244174e-01 -2.92002112e-01 -1.57471108e+00 3.71242672e-01
-6.72515512e-01 -1.21106315e+00 -2.02529877e-01 4.44174558e-01
9.73459542e-01 -1.31597430e-01 -6.90963073e-03 2.61120468e-01
1.12861145e+00 -9.69111025e-01 1.02138686e+00 3.93817455e-01
2.26430073e-01 -9.11401868e-01 3.80443186e-01 9.87906039e-01
-9.74919438e-01 -6.85255885e-01 -1.37949377e-01 -5.72273433e-01
3.18127722e-01 2.01106280e-01 -7.34756529e-01 5.85114002e-01
1.10651746e-01 3.08084160e-01 2.22762167e-01 9.97521520e-01
-3.38318259e-01 5.21342516e-01 -4.75888193e-01 -2.02545539e-01
5.39585292e-01 -3.63668412e-01 8.48999083e-01 7.96333671e-01
2.45239213e-01 6.02772653e-01 8.05338085e-01 7.19796062e-01
2.22379729e-01 -2.69460499e-01 -7.76998460e-01 1.10205086e-02
5.77883124e-01 6.95598364e-01 -7.66380191e-01 -4.20816958e-01
-1.48636550e-01 9.75355923e-01 4.46403116e-01 5.50421774e-01
-5.06698132e-01 7.75425285e-02 7.70538867e-01 -1.91399589e-01
1.20454974e-01 -5.04497230e-01 -3.02922338e-01 -1.17115808e+00
-6.15028292e-03 -6.23333395e-01 2.94736713e-01 -3.42915744e-01
-8.01931560e-01 4.22149032e-01 2.48451740e-01 -8.18237305e-01
-5.42816341e-01 -2.34772146e-01 -4.15269196e-01 7.79362679e-01
-1.11558378e+00 -8.17067027e-01 1.62692204e-01 5.23227215e-01
7.12295711e-01 1.11544933e-02 8.51278007e-01 -6.21787369e-01
-4.64378595e-01 -1.53210163e-01 2.43585631e-01 -5.24148166e-01
-1.44589599e-02 -1.52152729e+00 1.60507321e-01 1.11610949e+00
-6.72292635e-02 6.23077393e-01 1.25040960e+00 -6.47913754e-01
-1.63223839e+00 -1.23919678e+00 8.81606877e-01 -5.05334795e-01
8.66271734e-01 -1.06602818e-01 -6.54333234e-01 1.33019638e+00
-7.34760687e-02 -2.62748241e-01 3.21549118e-01 1.57560498e-01
-2.30771061e-02 2.75374919e-01 -1.10030115e+00 9.88832176e-01
1.25694489e+00 -2.24856108e-01 -8.29286516e-01 4.34823036e-01
8.87536347e-01 -3.38840693e-01 -5.41969776e-01 7.31917471e-02
2.57876962e-01 -6.98606193e-01 4.53975171e-01 -1.06953835e+00
4.19686943e-01 -3.64827961e-01 -4.90451336e-01 -1.48849475e+00
-4.86521214e-01 -1.19020450e+00 -1.77596182e-01 6.53427243e-01
3.21729392e-01 -8.49598587e-01 6.36688828e-01 1.07507420e+00
-8.30786750e-02 -6.53797925e-01 -9.34028983e-01 -1.20406926e+00
2.84553349e-01 -6.74528360e-01 6.87523484e-01 3.28268915e-01
3.13212425e-01 3.62093925e-01 -4.21096683e-01 6.37704790e-01
7.72423685e-01 4.13424522e-01 6.87042117e-01 -1.03644478e+00
-6.87647641e-01 -2.29101568e-01 -1.48036838e-01 -1.41622293e+00
2.86163568e-01 -6.25503898e-01 3.83800536e-01 -1.55596030e+00
4.89122480e-01 -7.50545740e-01 -7.10158348e-02 3.97465467e-01
8.93055797e-02 -8.53847504e-01 4.64390576e-01 4.61166590e-01
-7.27595687e-01 5.84376156e-01 1.32175064e+00 8.57588872e-02
-6.11190319e-01 5.65352440e-01 -7.70001888e-01 9.23729181e-01
9.02049780e-01 -4.74702954e-01 -7.08614409e-01 -3.53544384e-01
8.74598548e-02 6.92075908e-01 1.76436558e-01 -7.32579529e-01
4.67992812e-01 -1.07253325e+00 -2.19394952e-01 -2.59711385e-01
2.38330990e-01 -7.73097157e-01 4.05148000e-01 7.00677991e-01
-9.03715730e-01 -1.29383072e-01 -3.25632006e-01 1.03070199e+00
3.92450184e-01 -4.33969408e-01 6.17290914e-01 -3.39298606e-01
-7.37174332e-01 4.49386597e-01 -9.17795360e-01 -2.15275198e-01
1.38988161e+00 3.16028832e-03 1.06227674e-01 -6.60637736e-01
-1.24327695e+00 3.78489405e-01 7.14400291e-01 -9.76411924e-02
3.95972282e-01 -1.02295065e+00 -4.26160187e-01 -2.57739574e-01
-7.24239349e-02 1.63628027e-01 1.33288652e-01 5.91794848e-01
6.57185242e-02 6.40572488e-01 -1.18719473e-01 -2.39845082e-01
-8.77336204e-01 1.02981532e+00 4.54375356e-01 -4.96652156e-01
-7.07934380e-01 6.09054446e-01 3.38075519e-01 -3.18191022e-01
1.81927700e-02 -5.40513515e-01 1.24672405e-01 -3.82093817e-01
5.83539784e-01 2.46775508e-01 -4.52911615e-01 -2.67581761e-01
-1.74118593e-01 -6.76576002e-03 -7.95331895e-02 -7.96586633e-01
1.29843605e+00 -4.82737899e-01 7.99767077e-02 7.03975201e-01
6.44217134e-01 -1.63350537e-01 -1.66841173e+00 -5.93943417e-01
1.99081987e-01 -4.82068986e-01 -2.84920871e-01 -5.62057495e-01
-4.75271642e-01 3.35013509e-01 6.89416975e-02 6.17656112e-01
9.96528029e-01 2.51163363e-01 2.31881157e-01 8.65878463e-01
1.19324100e+00 -8.73084843e-01 3.00506153e-03 7.56593525e-01
6.78064883e-01 -9.31482852e-01 -3.78424138e-01 -1.35855094e-01
-7.51218975e-01 9.95327771e-01 1.26141220e-01 -2.90838629e-01
3.64743501e-01 2.81873465e-01 -6.44088149e-01 3.00825648e-02
-1.28578317e+00 -4.22155857e-01 -4.26307738e-01 7.01944411e-01
-3.71867716e-01 5.57150841e-01 -3.76773141e-02 4.94106889e-01
-2.68348604e-01 4.22082208e-02 9.30371404e-01 1.13838696e+00
-1.12431812e+00 -9.94908214e-01 -3.57178539e-01 3.00722420e-01
-1.67851090e-01 1.96321130e-01 -1.44851789e-01 5.92615664e-01
-2.03236878e-01 1.11881781e+00 -1.13853753e-01 -2.53168970e-01
1.68411031e-01 7.35310791e-03 5.74636161e-01 -8.45048010e-01
1.14230849e-01 -3.95493023e-02 1.17585994e-01 -7.64773548e-01
-3.09582710e-01 -1.20028389e+00 -1.33870697e+00 -2.90573925e-01
-1.82295889e-02 2.08241597e-01 3.20403755e-01 1.39030254e+00
-5.95077360e-03 4.00924742e-01 9.61744905e-01 -8.26753795e-01
-1.17424369e+00 -6.71511471e-01 -7.48543978e-01 -7.23421499e-02
6.31410837e-01 -5.70060015e-01 -5.45225322e-01 2.99703050e-02] | [4.383595943450928, 2.1636734008789062] |
d93cbf71-b886-48d5-b360-d6f983efb6cb | boosting-adversarial-robustness-using-feature | 2306.06462 | null | https://arxiv.org/abs/2306.06462v1 | https://arxiv.org/pdf/2306.06462v1.pdf | Boosting Adversarial Robustness using Feature Level Stochastic Smoothing | Advances in adversarial defenses have led to a significant improvement in the robustness of Deep Neural Networks. However, the robust accuracy of present state-ofthe-art defenses is far from the requirements in critical applications such as robotics and autonomous navigation systems. Further, in practical use cases, network prediction alone might not suffice, and assignment of a confidence value for the prediction can prove crucial. In this work, we propose a generic method for introducing stochasticity in the network predictions, and utilize this for smoothing decision boundaries and rejecting low confidence predictions, thereby boosting the robustness on accepted samples. The proposed Feature Level Stochastic Smoothing based classification also results in a boost in robustness without rejection over existing adversarial training methods. Finally, we combine the proposed method with adversarial detection methods, to achieve the benefits of both approaches. | ['R. Venkatesh Babu', 'Gaurang Sriramanan', 'Samyak Jain', 'Sravanti Addepalli'] | 2023-06-10 | null | null | null | null | ['adversarial-robustness', 'autonomous-navigation'] | ['adversarial', 'computer-vision'] | [ 2.52094746e-01 2.08128452e-01 -1.04706939e-02 -3.48270684e-01
-4.94798779e-01 -7.38835096e-01 7.63928056e-01 -8.16243738e-02
-4.39057589e-01 9.83516455e-01 -1.83715284e-01 -6.03851199e-01
-1.86211437e-01 -8.50815833e-01 -5.95421255e-01 -9.41048205e-01
-1.17726855e-01 -1.37619510e-01 6.63087189e-01 -3.05730194e-01
1.56935424e-01 9.75544870e-01 -1.23878336e+00 1.17670670e-01
9.81388927e-01 1.29137385e+00 -3.85160923e-01 4.52228993e-01
1.41842127e-01 7.35493481e-01 -8.23012173e-01 -7.01668918e-01
5.56575596e-01 -9.60380286e-02 -1.91634327e-01 -4.78364706e-01
3.92426729e-01 -5.32418609e-01 -4.69057620e-01 1.43316495e+00
3.77955735e-01 2.73643762e-01 8.14610541e-01 -1.24339771e+00
-3.78438681e-01 5.68230093e-01 -3.24211270e-01 9.62199830e-03
-1.01628184e-01 6.67260289e-02 5.41898131e-01 -4.55170691e-01
2.64008343e-01 1.20927286e+00 7.46431470e-01 9.85052109e-01
-8.92996192e-01 -1.02691567e+00 3.25203627e-01 1.54271632e-01
-1.05429375e+00 -6.57887101e-01 9.74697411e-01 -1.95693940e-01
5.04395723e-01 1.96195528e-01 1.86520919e-01 1.37288249e+00
5.32678485e-01 5.23501456e-01 9.76910770e-01 -2.70723760e-01
5.74638844e-01 2.45008543e-01 -1.59674764e-01 5.90298831e-01
4.51095760e-01 5.62198699e-01 -1.91441879e-01 -3.00141245e-01
6.42831624e-01 3.78395207e-02 -1.40638679e-01 -3.99338633e-01
-5.03876925e-01 9.63574946e-01 5.22046804e-01 8.60708654e-02
-3.16286415e-01 1.07016757e-01 5.55057228e-01 2.88260996e-01
5.81941247e-01 2.82424450e-01 -2.97102958e-01 2.26310238e-01
-1.03422749e+00 2.61588454e-01 7.16004968e-01 6.11760616e-01
1.36502385e-01 7.46903181e-01 1.04248904e-01 5.95650494e-01
3.22601497e-01 3.61282825e-01 3.06039810e-01 -9.16213334e-01
2.97683477e-01 2.27907583e-01 6.38548732e-02 -1.09070921e+00
-2.73413539e-01 -5.67410946e-01 -9.67082977e-01 1.04415703e+00
4.67961133e-01 -4.82619315e-01 -9.57377315e-01 1.78475046e+00
3.06563973e-01 3.04148853e-01 3.40802610e-01 6.61162019e-01
1.73977092e-01 2.89005846e-01 3.75110321e-02 4.44075242e-02
7.45031953e-01 -5.77144086e-01 -5.09095371e-01 -2.22345293e-01
1.72362179e-01 -4.90998417e-01 4.01759446e-01 6.28496706e-01
-6.12701237e-01 -4.71303403e-01 -1.45633006e+00 6.69964075e-01
-3.27167690e-01 -2.20447287e-01 5.86260319e-01 1.23604441e+00
-5.66364348e-01 8.54752123e-01 -8.19007814e-01 8.63825530e-02
5.15108049e-01 5.99650741e-01 -3.80582422e-01 1.92325205e-01
-1.41468132e+00 1.02423656e+00 4.73857611e-01 2.33383968e-01
-9.97127473e-01 -3.82970870e-01 -6.57507539e-01 -2.82234885e-02
1.73575088e-01 -2.28829861e-01 8.94724905e-01 -8.28422666e-01
-1.50534344e+00 2.97512323e-01 1.98253751e-01 -1.11602247e+00
9.37228680e-01 -3.96090329e-01 -5.49106121e-01 1.31908819e-01
-3.45824629e-01 5.61635613e-01 1.36608505e+00 -1.24572217e+00
-5.89016616e-01 -2.19834760e-01 3.39695692e-01 -1.64854944e-01
-5.99983096e-01 8.04137718e-03 1.80422738e-01 -8.02272618e-01
7.48277381e-02 -8.62728715e-01 -5.96217573e-01 2.04121128e-01
-3.55782390e-01 1.09618932e-01 1.21471608e+00 -5.83705664e-01
7.38022685e-01 -2.30275536e+00 -2.88092375e-01 3.31429571e-01
7.03348890e-02 7.40882814e-01 1.78561538e-01 1.31086916e-01
8.25719163e-03 1.16476864e-02 -2.58808225e-01 -3.42176944e-01
-3.22865173e-02 2.83330321e-01 -9.30160344e-01 6.34983420e-01
5.00729740e-01 3.88789058e-01 -6.80555522e-01 -1.27316996e-01
3.17368329e-01 4.43975657e-01 -5.94991982e-01 6.32510334e-02
-3.40477638e-02 4.83986765e-01 -5.08300781e-01 6.87858820e-01
7.31983721e-01 5.15473902e-01 -1.09497048e-01 4.08109128e-02
3.40110034e-01 2.52339337e-03 -1.14144337e+00 8.34681928e-01
-3.10415298e-01 6.15139246e-01 3.23792577e-01 -1.21049356e+00
1.13196230e+00 2.21847653e-01 1.68015867e-01 -1.38139471e-01
3.00967097e-01 1.57716796e-01 2.99999624e-01 6.76635951e-02
5.77823460e-01 -2.08459318e-01 -4.41734307e-02 6.12158515e-02
-6.50957450e-02 -9.19119790e-02 -3.77633750e-01 6.86922967e-02
1.03561461e+00 1.72072515e-01 8.55406374e-02 -9.85303000e-02
5.87042212e-01 -1.93340018e-01 7.41500616e-01 8.99358511e-01
-5.98202348e-01 4.22012895e-01 4.01766956e-01 -1.73828259e-01
-8.95961165e-01 -1.08058143e+00 -1.75187588e-01 7.35810459e-01
9.56080556e-02 2.18659326e-01 -9.30987835e-01 -1.18488646e+00
2.22824603e-01 7.42031217e-01 -5.36522210e-01 -6.75681353e-01
-4.78571862e-01 -5.74161530e-01 1.07178748e+00 5.95169365e-01
5.06326377e-01 -9.03687358e-01 -3.21701795e-01 1.84992179e-01
3.86269808e-01 -1.18982196e+00 7.20073730e-02 5.02333105e-01
-9.56481397e-01 -9.42107141e-01 -3.71067405e-01 -4.22335446e-01
6.69983447e-01 -3.83081147e-03 4.08671767e-01 1.39189959e-01
2.27126591e-02 -1.63310975e-01 -3.97052646e-01 -6.03989601e-01
-6.73160553e-01 -1.67478830e-01 7.64740109e-01 -5.55734001e-02
-6.09124079e-02 -5.57449102e-01 -4.26877111e-01 2.59059131e-01
-9.33224678e-01 -5.97577095e-01 5.32069564e-01 1.03075957e+00
1.20511921e-02 4.50431585e-01 1.12775064e+00 -8.56814682e-01
6.38337672e-01 -4.27187830e-01 -7.07094491e-01 -3.70269492e-02
-6.26595795e-01 1.77848898e-02 1.17990708e+00 -7.47981608e-01
-9.91045237e-01 -2.56272741e-02 -6.04353428e-01 -6.54749155e-01
-2.52125263e-01 2.28383616e-01 -1.53673157e-01 -7.02193201e-01
8.24597716e-01 -1.86755843e-02 1.64888903e-01 -2.12959319e-01
2.19787434e-01 5.25994062e-01 6.56086981e-01 -5.08452773e-01
1.07298017e+00 4.38663751e-01 2.33905122e-01 -5.15675128e-01
-6.70313358e-01 4.47128620e-03 -6.02873445e-01 -1.82116821e-01
2.51865804e-01 -6.01087093e-01 -5.45295238e-01 6.07195139e-01
-1.02209866e+00 -1.71402439e-01 -1.61625296e-01 3.44204426e-01
-4.40160543e-01 6.82999372e-01 -5.32459557e-01 -1.34717417e+00
-3.52487147e-01 -1.20075607e+00 4.84319240e-01 2.69531608e-01
-5.57862259e-02 -1.01024795e+00 -5.09162188e-01 8.18827078e-02
5.85221231e-01 4.95496482e-01 6.17351234e-01 -1.22644138e+00
-2.89534628e-01 -9.35684085e-01 -3.66882570e-02 9.05253112e-01
-3.47305275e-02 1.33020535e-01 -1.18634915e+00 -2.71833062e-01
3.61091763e-01 -3.63399118e-01 9.39553916e-01 1.74953297e-01
1.01529980e+00 -3.15421075e-01 -1.86957464e-01 4.63513643e-01
1.08980083e+00 5.04317284e-01 6.71078742e-01 2.91660964e-01
4.15839791e-01 6.47360623e-01 8.16389263e-01 3.59650940e-01
-4.49669719e-01 5.19567490e-01 8.74738455e-01 9.78159234e-02
1.19376697e-01 -1.13993049e-01 5.72953820e-01 2.63327897e-01
1.40546143e-01 -2.35548958e-01 -6.71582639e-01 3.56336504e-01
-1.79989529e+00 -1.06886292e+00 1.97486177e-01 2.25386047e+00
6.96321726e-01 8.47078621e-01 -2.41090760e-01 4.19944704e-01
7.62591898e-01 2.01130882e-01 -7.52281249e-01 -4.90806252e-01
-7.54287839e-02 1.07736371e-01 8.50585639e-01 4.95433688e-01
-1.28123450e+00 9.88899171e-01 6.97979832e+00 1.23913980e+00
-1.18292165e+00 -2.97550838e-02 5.66749334e-01 7.19482750e-02
1.47526814e-02 -6.11565672e-02 -9.49757457e-01 3.67802292e-01
9.81648922e-01 6.77773803e-02 1.86662465e-01 1.21394384e+00
1.34472385e-01 2.49595698e-02 -7.36754954e-01 4.11053658e-01
2.93244794e-02 -1.08826911e+00 4.09974121e-02 9.75733772e-02
6.29423738e-01 -2.09459618e-01 3.39575559e-01 3.61780256e-01
5.71972370e-01 -9.22241688e-01 7.33233571e-01 4.24823195e-01
4.42907333e-01 -1.23929846e+00 1.03082943e+00 6.48394406e-01
-7.42206275e-01 -2.43176758e-01 -6.06215239e-01 -8.37795250e-03
2.03217417e-02 6.54525042e-01 -8.14443529e-01 5.54290473e-01
4.07278150e-01 1.16368741e-01 -2.79112488e-01 8.47744823e-01
-4.67938840e-01 7.91800916e-01 -3.56823504e-01 -1.33329585e-01
2.57219017e-01 -1.58594195e-02 8.31043839e-01 1.00210035e+00
3.30840908e-02 -2.06054866e-01 4.15904745e-02 5.62339902e-01
4.26696986e-02 -2.76613712e-01 -7.59005189e-01 1.11443818e-01
6.57220900e-01 1.13196576e+00 -6.73610449e-01 -2.61009574e-01
-9.70990285e-02 8.98802400e-01 2.46244118e-01 2.16656640e-01
-1.05764997e+00 -6.57401085e-01 9.17105019e-01 -2.40891799e-01
2.57403523e-01 -4.06437308e-01 -6.01412356e-01 -9.19220448e-01
-7.88495243e-02 -8.48931134e-01 2.05454659e-02 -2.08969042e-01
-1.34518886e+00 7.01110840e-01 -2.48003662e-01 -1.36206293e+00
-4.94452000e-01 -9.71005499e-01 -9.14279461e-01 8.35568130e-01
-1.15968382e+00 -1.01557970e+00 1.83295980e-01 5.34413338e-01
4.21716303e-01 -6.20256066e-01 8.71716857e-01 1.28641978e-01
-5.01786470e-01 1.01096833e+00 3.16882312e-01 2.38730744e-01
8.14971566e-01 -9.88738060e-01 5.23425758e-01 1.35397208e+00
-9.34307128e-02 8.35676908e-01 1.07799363e+00 -6.94220781e-01
-1.02745509e+00 -1.23700464e+00 1.50167271e-01 -2.41262957e-01
9.34556544e-01 -3.15265507e-01 -9.88797665e-01 1.91566885e-01
-1.65139213e-01 7.61936232e-02 5.25590122e-01 5.76589163e-03
-4.40479070e-01 -1.01635695e-01 -1.74269247e+00 8.33742201e-01
5.95422029e-01 -4.23784375e-01 -4.94865656e-01 1.25623941e-01
7.06954181e-01 -3.80082101e-01 -7.62656450e-01 6.50509834e-01
7.08159089e-01 -9.53211010e-01 1.08829296e+00 -6.99237704e-01
2.25334868e-01 -2.67542571e-01 -2.81852365e-01 -1.22992206e+00
-9.54501927e-02 -4.10390764e-01 -3.92844230e-01 1.28088975e+00
3.27505976e-01 -8.63139570e-01 1.25373065e+00 8.08379412e-01
-2.40597859e-01 -7.18941033e-01 -1.30641484e+00 -1.07595265e+00
1.48841828e-01 -6.93757892e-01 3.23707819e-01 6.77864313e-01
-1.93014637e-01 -5.08890033e-01 -6.47871256e-01 5.73129058e-01
8.80155146e-01 -4.12617505e-01 6.44863486e-01 -1.11553216e+00
-3.51630390e-01 -4.28045094e-01 -8.53627801e-01 -7.07419753e-01
3.96965533e-01 -4.44512755e-01 1.82652116e-01 -1.00877130e+00
-4.64808196e-01 -5.45172215e-01 -5.51881790e-01 5.00583172e-01
-2.49194607e-01 5.41716933e-01 1.23892330e-01 -1.28174480e-02
-1.13756977e-01 5.66230834e-01 5.64861894e-01 -2.08593100e-01
1.38830930e-01 3.67205560e-01 -6.38838410e-01 9.85355616e-01
1.02799392e+00 -6.23072147e-01 -3.52617770e-01 6.55997992e-02
-3.50489765e-01 -1.02059104e-01 3.93559873e-01 -1.37022960e+00
3.28162223e-01 -2.48266920e-01 5.48308492e-01 -2.89457142e-01
5.23139179e-01 -1.03694510e+00 -1.66337609e-01 7.47978091e-01
-1.77755222e-01 -3.89346540e-01 3.32999587e-01 8.64527583e-01
-1.82815045e-01 -5.99136710e-01 1.12670898e+00 3.24526340e-01
-5.50559700e-01 1.70420215e-01 -4.40940380e-01 -3.61066818e-01
9.91723835e-01 -3.01819623e-01 -3.69304359e-01 -3.29610437e-01
-6.60486937e-01 -1.53308630e-01 3.68518561e-01 5.11000633e-01
6.55228078e-01 -1.14455140e+00 -3.91593426e-01 2.78445423e-01
-3.50267589e-01 -2.90635854e-01 1.06504709e-01 4.92027462e-01
-3.17120880e-01 5.76402359e-02 -3.95817667e-01 -1.72527224e-01
-1.19796205e+00 6.72614634e-01 1.91745013e-01 -9.69690382e-02
-4.09063220e-01 8.66854668e-01 4.21695262e-02 -2.69795448e-01
5.07001996e-01 1.10710271e-01 -2.27235898e-01 -2.29564235e-01
5.27412117e-01 2.14784905e-01 7.70017579e-02 -4.76564944e-01
-3.21351260e-01 3.68450955e-02 -2.71992624e-01 -1.55702293e-01
1.05331194e+00 1.65652514e-01 2.87128299e-01 1.19284846e-01
6.75820649e-01 3.52898061e-01 -1.45059824e+00 -5.71782235e-03
-1.98314548e-03 -3.70089591e-01 -3.37124318e-02 -6.29729509e-01
-1.10076773e+00 8.51835012e-01 5.46944916e-01 3.71208042e-01
1.10719061e+00 -5.97915292e-01 7.52622783e-01 5.34838140e-01
6.06089771e-01 -8.04220855e-01 -2.59803236e-01 6.58086240e-01
6.66824937e-01 -1.22739100e+00 5.65341301e-03 -3.74724656e-01
-4.88266826e-01 1.18587625e+00 6.12767100e-01 -5.54000735e-01
5.83484292e-01 4.54645157e-01 5.60094193e-02 4.49450374e-01
-6.11677706e-01 1.56119809e-01 3.20151806e-01 9.83693421e-01
8.54242370e-02 -1.19875453e-01 -2.67344475e-01 7.80249596e-01
-9.86006781e-02 -4.65411186e-01 3.77016634e-01 8.81837964e-01
-6.60879314e-01 -1.13223159e+00 -5.94178915e-01 4.34909523e-01
-9.03993726e-01 -7.67782405e-02 -3.35214525e-01 4.96308297e-01
5.03767189e-03 1.12745261e+00 -3.85388047e-01 -7.91897297e-01
2.31168211e-01 3.16691697e-02 1.09600134e-01 -2.52845913e-01
-5.88995159e-01 -2.26970583e-01 1.52985916e-01 -2.67842650e-01
-7.70279579e-03 -5.68698347e-01 -1.09937084e+00 -3.52784514e-01
-6.26101613e-01 2.01883033e-01 8.50999773e-01 9.94635999e-01
1.19959317e-01 5.07849634e-01 8.12082529e-01 -9.54857171e-01
-1.15868473e+00 -8.87207985e-01 -4.97422010e-01 5.81641570e-02
3.18051994e-01 -8.89494181e-01 -7.29611695e-01 -1.74307197e-01] | [5.511204242706299, 7.826510906219482] |
7ae2ce5a-6750-49e5-ad36-91eedea2d4c5 | power-norm-based-lifelong-learning-for | null | null | https://openreview.net/forum?id=4TMv3kdrhuc | https://openreview.net/pdf?id=4TMv3kdrhuc | Power Norm Based Lifelong Learning for Paraphrase Generations | Seq2seq language generation models are trained with multiple domains in a continue learning manner, where the data from each domain being observed in an online fashion. However, continual learning studies usually suffer a lot from catastrophic forgetting, a persistent challenge for lifelong learning. To handle this problem, existing work has leveraged experience replay or dynamic architecture to consolidate the past knowledge, which however result in incremental memory space or high computational cost. In this work, we propose an innovative framework PNLL that remedies the catastrophic forgetting issues with a power normalization on NLP transformer models. Specifically, PNLLL leverages power norm to achieve a better balance between past experience rehearsal and new knowledge acquisition. Our experiments on, paraphrase generation, show that PNLLL outperforms SOTA models by a considerable margin and remedy forgetting greatly. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 1.26168966e-01 -1.15222774e-01 -3.29204142e-01 -1.71243489e-01
-7.55989313e-01 -4.71981943e-01 6.91238046e-01 -1.31534249e-01
-3.99642497e-01 1.24820387e+00 4.90725487e-01 -2.78297246e-01
1.01028532e-01 -9.31373596e-01 -8.53941917e-01 -4.43424851e-01
3.66771996e-01 3.79256368e-01 2.26299867e-01 -3.44748974e-01
-4.38102223e-02 7.07215071e-02 -1.25454700e+00 2.50284821e-01
1.11022341e+00 5.25638342e-01 3.14439327e-01 4.43730116e-01
-2.99716651e-01 1.12554598e+00 -5.60586512e-01 -7.42954493e-01
1.60467148e-01 -5.36377609e-01 -7.76298761e-01 -4.34349149e-01
4.64912415e-01 -5.84042370e-01 -7.36141980e-01 8.28972578e-01
7.51741827e-01 3.12344790e-01 3.18752259e-01 -1.13663471e+00
-1.23068750e+00 1.02166104e+00 -4.22591269e-01 4.05465603e-01
1.84170395e-01 6.01190805e-01 7.45620668e-01 -1.18366921e+00
5.20339012e-01 1.12017465e+00 9.82617140e-01 7.87978232e-01
-1.09683919e+00 -9.31181908e-01 2.76102155e-01 3.73363495e-01
-1.06940246e+00 -4.60119903e-01 7.13006377e-01 1.86260603e-02
1.36900866e+00 -4.05788898e-01 8.05838466e-01 1.86393094e+00
4.38795656e-01 1.04835105e+00 9.31888044e-01 -2.65440077e-01
3.26811939e-01 -3.59215140e-02 8.31611734e-03 3.07560503e-01
4.01815236e-01 2.70748585e-01 -9.30723488e-01 9.91624147e-02
6.35893285e-01 4.22237128e-01 4.61507663e-02 -2.70971328e-01
-8.95790100e-01 5.34639895e-01 1.63563401e-01 2.96296448e-01
-4.21052456e-01 7.76803121e-02 5.00079393e-01 9.43733215e-01
5.21026611e-01 6.26733899e-01 -6.92306280e-01 -6.07468247e-01
-1.01740360e+00 2.54453570e-01 5.62919021e-01 1.08434331e+00
6.14877582e-01 4.79859650e-01 -4.83205050e-01 9.46344078e-01
-3.14517766e-01 7.03218877e-01 1.12042987e+00 -7.18020320e-01
5.26846707e-01 5.30923426e-01 6.95480555e-02 -4.64186370e-01
-7.32225180e-02 -7.28425682e-01 -9.13643479e-01 -2.42176116e-01
2.09020793e-01 -4.16556686e-01 -1.06884181e+00 2.19929123e+00
-7.31084272e-02 4.70219910e-01 3.22421968e-01 4.12319630e-01
3.58860463e-01 5.67543626e-01 4.45883989e-01 -2.00016558e-01
8.79514635e-01 -1.17505312e+00 -7.50329673e-01 -6.30045891e-01
4.48591202e-01 -4.45950747e-01 1.62770236e+00 5.06458342e-01
-1.42069018e+00 -5.93508959e-01 -1.12579238e+00 -2.51333982e-01
-2.34431371e-01 -1.54883340e-01 6.70024872e-01 2.94871837e-01
-1.07340646e+00 7.88441420e-01 -8.92968655e-01 -3.54136050e-01
4.28166419e-01 -5.98397031e-02 -2.86322534e-02 -3.84145945e-01
-1.66860962e+00 1.05606592e+00 5.15120924e-01 -3.40847164e-01
-1.00848818e+00 -1.25432432e+00 -6.80622518e-01 2.67027795e-01
3.52684498e-01 -1.00185144e+00 1.68575203e+00 -1.02076876e+00
-1.73401725e+00 4.56111997e-01 -3.20097315e-03 -8.90059650e-01
5.96893013e-01 -7.78005123e-01 -6.34192884e-01 -4.15484101e-01
-7.84164816e-02 3.65865886e-01 8.14600050e-01 -7.05557108e-01
-4.11354780e-01 -1.85596555e-01 -1.36220425e-01 4.11596507e-01
-6.80222273e-01 -5.60398877e-01 -7.23737255e-02 -9.70492899e-01
-3.80573273e-01 -6.93300068e-01 -1.79515466e-01 -5.04885674e-01
3.01255912e-01 -1.93774000e-01 6.65049732e-01 -6.52263403e-01
1.46998179e+00 -2.05195928e+00 5.08192740e-02 -5.18216789e-01
-1.07097782e-01 6.68922186e-01 -5.26361883e-01 5.19365430e-01
1.05609491e-01 -2.54657835e-01 -5.94402067e-02 -5.08500636e-01
-1.82232127e-01 2.84083635e-01 -9.05748606e-01 -2.75797099e-01
2.71217883e-01 1.26698089e+00 -1.40769601e+00 -6.41647503e-02
-1.85580686e-01 3.07396919e-01 -6.65843904e-01 3.26392382e-01
-4.47242737e-01 9.89204571e-02 -2.64846161e-02 5.42769551e-01
7.32115090e-01 -2.94384956e-01 2.54435480e-01 3.42621207e-01
1.80230156e-01 4.85575080e-01 -7.34662890e-01 2.27189040e+00
-8.35988224e-01 2.64319956e-01 -3.86954904e-01 -5.73184073e-01
8.03757250e-01 2.60788262e-01 1.43573448e-01 -1.12777674e+00
-2.36191154e-01 2.84844458e-01 -2.44259551e-01 -1.54104993e-01
8.47023547e-01 -5.84027827e-01 -1.11048773e-01 5.78317225e-01
5.32947898e-01 1.02336876e-01 2.10770760e-02 3.11464846e-01
1.38408887e+00 2.27554038e-01 3.07872891e-01 1.81619078e-01
7.65702203e-02 -3.43917981e-02 7.79169023e-01 1.09946036e+00
-2.53469586e-01 4.74163413e-01 -1.57931936e-03 -2.96463042e-01
-1.19856596e+00 -1.50640607e+00 4.69404429e-01 1.29202199e+00
-1.91834509e-01 -2.91719228e-01 -5.56191623e-01 -6.75784528e-01
1.05461806e-01 1.18315482e+00 -3.14893067e-01 -9.76478815e-01
-7.47082889e-01 -7.35378981e-01 6.58218622e-01 8.75787139e-01
6.97970927e-01 -1.25190115e+00 -3.87714922e-01 5.61632991e-01
-1.24773525e-01 -6.52542710e-01 -6.86183453e-01 2.80524194e-01
-1.25629497e+00 -5.17590106e-01 -9.65494096e-01 -7.24851012e-01
2.71317095e-01 5.02281904e-01 1.50720537e+00 -3.05939376e-01
-8.40853900e-02 2.77722955e-01 -3.30341041e-01 -3.76139700e-01
-3.08840185e-01 6.21653736e-01 4.13370878e-01 -4.75520700e-01
6.17800891e-01 -1.15265906e+00 -7.11322010e-01 -1.95722893e-01
-8.82931352e-01 -3.73150185e-02 1.01258671e+00 1.27226794e+00
4.78424430e-01 -2.23578103e-02 1.31683004e+00 -1.06199002e+00
9.52675223e-01 -9.95813608e-01 -2.41957337e-01 4.23483551e-01
-1.13137460e+00 2.63101369e-01 1.06713438e+00 -8.37028146e-01
-1.58030093e+00 -3.45325857e-01 1.08059347e-02 -7.30610371e-01
1.94942102e-01 5.31915188e-01 -1.11528330e-01 4.03926849e-01
8.73603106e-01 7.29164660e-01 -1.63890049e-01 -5.74032366e-01
7.14125276e-01 3.54827821e-01 7.15224147e-01 -6.35232091e-01
7.03827560e-01 -2.27665659e-02 -6.13196671e-01 -2.99531788e-01
-1.28321695e+00 -1.05080806e-01 -4.06322032e-01 3.00792605e-01
1.77205279e-01 -1.28697705e+00 -3.47690620e-02 7.70060003e-01
-1.00119865e+00 -6.92869246e-01 -8.13757181e-01 3.17635894e-01
-3.80118698e-01 1.67973191e-01 -8.08668554e-01 -5.61894596e-01
-7.03040659e-01 -2.12849855e-01 5.06331742e-01 3.97842765e-01
-2.60770351e-01 -9.82272685e-01 5.47440171e-01 5.63820302e-02
8.95124912e-01 -2.98833638e-01 8.99661660e-01 -5.46227217e-01
-4.52608585e-01 1.61145970e-01 -2.35810280e-02 5.79665601e-01
1.60510674e-01 -6.78510725e-01 -9.01677847e-01 -6.18141115e-01
9.89442542e-02 -7.24117219e-01 1.03187859e+00 -1.68885440e-01
9.65813041e-01 -5.86428821e-01 -6.74053431e-02 5.14091134e-01
1.31846607e+00 2.88548678e-01 6.22859538e-01 1.60462797e-01
4.22778547e-01 2.73685567e-02 5.84044278e-01 5.36235750e-01
4.04591978e-01 1.99735507e-01 -7.31604472e-02 3.23668599e-01
-6.64957523e-01 -1.15527987e+00 7.76020288e-01 1.21747196e+00
3.13346893e-01 -3.39906782e-01 -8.36468339e-01 9.04402971e-01
-1.83303690e+00 -1.12672162e+00 7.24878848e-01 2.05303216e+00
1.56347620e+00 2.44142264e-01 -1.88160524e-01 -1.14120930e-01
1.73183650e-01 1.67750210e-01 -1.11276805e+00 -2.30093181e-01
-3.26901734e-01 5.70331454e-01 3.08044314e-01 3.43227923e-01
-5.76968729e-01 1.22793257e+00 6.60445547e+00 9.01313007e-01
-1.11906183e+00 3.36515218e-01 2.86272794e-01 -4.81772989e-01
-3.78688067e-01 2.59671547e-02 -9.48726118e-01 6.91608787e-01
1.30530417e+00 -6.57012701e-01 7.13135898e-01 9.96624947e-01
-2.02416107e-01 1.01348452e-01 -1.12313962e+00 1.03400159e+00
1.22516587e-01 -1.18947673e+00 4.96856689e-01 -4.80628431e-01
1.03183556e+00 2.41350040e-01 5.04699230e-01 1.12344468e+00
7.99997926e-01 -9.04895425e-01 4.82838929e-01 6.76084757e-01
7.69930482e-01 -7.13877439e-01 5.18440425e-01 5.71880817e-01
-7.95342922e-01 -3.16349685e-01 -4.41707969e-01 -3.62582803e-01
3.09325576e-01 6.52522802e-01 -1.10146666e+00 3.97073299e-01
4.42647964e-01 7.49365509e-01 -7.82267749e-01 9.16597784e-01
-3.57095689e-01 9.39856410e-01 3.25165838e-02 1.28461078e-01
7.36671267e-03 1.58610865e-01 3.81768882e-01 1.09813428e+00
7.46375680e-01 1.96195003e-02 -8.47126320e-02 7.29591548e-01
-4.73032445e-01 -1.92041934e-01 -8.06622505e-01 -1.75683364e-01
7.83578098e-01 7.80131161e-01 5.98759539e-02 -4.00186419e-01
-4.35689121e-01 1.35041857e+00 6.64956391e-01 3.30320835e-01
-6.91841125e-01 -4.61304724e-01 5.52005708e-01 2.77746528e-01
2.05870956e-01 -3.46052766e-01 -2.37468719e-01 -1.52736366e+00
1.02324955e-01 -7.92715192e-01 4.62883294e-01 -8.85828972e-01
-1.63742137e+00 4.33621079e-01 -1.87879696e-01 -9.96779442e-01
-4.27363634e-01 -2.28723124e-01 -7.27275908e-01 6.36411250e-01
-1.81721723e+00 -1.08912432e+00 -1.25768960e-01 6.55296206e-01
8.75985265e-01 -3.44074726e-01 7.84760833e-01 4.86967385e-01
-3.30880105e-01 1.02605045e+00 3.13395977e-01 -2.93626219e-01
9.88485098e-01 -1.23305690e+00 6.00350916e-01 8.23781371e-01
2.80729006e-03 8.24935138e-01 6.22160256e-01 -8.85874033e-01
-1.38789356e+00 -1.37504244e+00 1.19753194e+00 -7.08209097e-01
6.81851983e-01 -3.32432866e-01 -1.17638290e+00 9.04905736e-01
4.95446861e-01 -3.37889671e-01 5.70065856e-01 2.09389970e-01
-5.59879065e-01 -2.07310826e-01 -1.02616370e+00 9.09996450e-01
1.47963929e+00 -6.24555230e-01 -1.05769908e+00 4.75545526e-02
1.12810814e+00 -2.68620998e-01 -6.70057654e-01 1.91611275e-01
3.55043530e-01 -6.43777132e-01 8.46201181e-01 -8.47030222e-01
5.32256305e-01 -7.41232978e-03 2.57654816e-01 -1.75367415e+00
-3.47313702e-01 -8.49110246e-01 -7.29022145e-01 1.38093424e+00
3.58332515e-01 -5.10565042e-01 8.03377032e-01 3.36205810e-01
-1.34022117e-01 -4.58161294e-01 -8.19727361e-01 -1.21237266e+00
4.75626826e-01 -1.43715814e-01 8.15063715e-01 9.01924551e-01
1.31727964e-01 9.09586430e-01 -8.47245395e-01 -4.32724178e-01
5.26975155e-01 -7.38618150e-02 6.60041273e-01 -9.60905492e-01
-4.06722844e-01 -6.56199157e-02 3.25536013e-01 -1.27953887e+00
1.27782539e-01 -9.52159941e-01 -1.38562933e-01 -1.09079218e+00
4.16483581e-01 -2.17144117e-01 -7.79051960e-01 7.04472542e-01
-4.16476935e-01 -2.98767507e-01 8.86823982e-02 3.38581465e-02
-8.75474989e-01 1.10111547e+00 1.19952893e+00 -6.00806288e-02
-3.92910719e-01 -1.46793276e-01 -1.00879514e+00 4.65160549e-01
8.85434806e-01 -4.24251825e-01 -9.64402139e-01 -8.44214678e-01
5.32470763e-01 4.19991650e-02 1.68431938e-01 -1.15538108e+00
4.56699282e-01 -2.15978399e-01 3.66606474e-01 -4.65718061e-01
2.34097198e-01 -6.04333401e-01 1.75072789e-01 4.12624329e-01
-4.25802201e-01 3.76043618e-01 2.30797350e-01 8.33252490e-01
-3.20126295e-01 -2.39963997e-02 7.25957274e-01 -3.68858725e-01
-7.63931096e-01 3.74742061e-01 -7.19704777e-02 5.11038601e-01
7.05024362e-01 6.08063750e-02 -4.56192762e-01 -3.29594374e-01
-6.42101526e-01 3.92858803e-01 4.86222953e-01 8.16108882e-01
6.65830553e-01 -1.68228078e+00 -5.38123846e-01 3.83155912e-01
1.52617469e-02 -6.34284019e-02 6.22568607e-01 4.54648405e-01
1.40465513e-01 3.36849093e-01 -3.69214952e-01 4.20585908e-02
-6.72313035e-01 6.64605439e-01 6.79328218e-02 -8.08479071e-01
-5.44993281e-01 7.94840813e-01 -6.41507879e-02 -3.46937507e-01
2.00073376e-01 -2.13636070e-01 9.73483995e-02 8.00154954e-02
7.24273622e-01 4.36195374e-01 8.12732354e-02 2.14913040e-01
3.35820653e-02 4.54081111e-02 -7.22524047e-01 -9.35759470e-02
1.37859464e+00 -3.25794876e-01 4.20296788e-01 5.69747567e-01
7.55930543e-01 -1.52640760e-01 -1.60272896e+00 -7.93000519e-01
1.24208525e-01 -3.04956138e-01 -2.82157272e-01 -1.29300261e+00
-7.56975055e-01 8.79933178e-01 4.76666033e-01 -4.26687986e-01
1.15558195e+00 -3.97986382e-01 1.51011014e+00 5.62257349e-01
6.61165416e-01 -1.40214145e+00 7.14099109e-01 1.20188677e+00
7.65632689e-01 -1.01080978e+00 -2.77049273e-01 4.40983713e-01
-8.31176937e-01 6.50234997e-01 9.96019781e-01 -1.46356523e-01
5.20604134e-01 2.11018518e-01 -2.33441412e-01 3.40622604e-01
-1.32816494e+00 2.58301765e-01 -3.25036913e-01 5.55206239e-01
2.45171353e-01 -1.02070041e-01 -1.69053674e-01 1.25818264e+00
-2.79644907e-01 6.65889859e-01 3.52223098e-01 1.03117132e+00
-4.14768308e-01 -1.25318825e+00 3.24477643e-01 2.84971505e-01
-4.12833869e-01 -6.48194075e-01 -1.36218399e-01 4.14774656e-01
2.33721789e-02 4.68907058e-01 -1.00373000e-01 -3.40803534e-01
6.13339901e-01 5.65970480e-01 5.51217079e-01 -7.23609924e-01
-7.10011303e-01 -5.74924290e-01 -1.98710904e-01 -3.60297292e-01
1.49228036e-01 -5.05458593e-01 -1.03037012e+00 -3.87439638e-01
-1.59276575e-02 4.20061611e-02 -9.40961167e-02 5.37108660e-01
8.16957712e-01 4.59057033e-01 4.51011747e-01 -1.90923512e-01
-1.18643987e+00 -1.11501396e+00 -5.86221933e-01 4.90732819e-01
3.89116555e-01 -5.45252502e-01 -2.94394463e-01 -3.54318246e-02] | [9.921573638916016, 3.544647693634033] |
09e54e3a-55f9-40a4-abdb-7589fb1c7b17 | translatotron-2-robust-direct-speech-to | 2107.08661 | null | https://arxiv.org/abs/2107.08661v5 | https://arxiv.org/pdf/2107.08661v5.pdf | Translatotron 2: High-quality direct speech-to-speech translation with voice preservation | We present Translatotron 2, a neural direct speech-to-speech translation model that can be trained end-to-end. Translatotron 2 consists of a speech encoder, a linguistic decoder, an acoustic synthesizer, and a single attention module that connects them together. Experimental results on three datasets consistently show that Translatotron 2 outperforms the original Translatotron by a large margin on both translation quality (up to +15.5 BLEU) and speech generation quality, and approaches the same of cascade systems. In addition, we propose a simple method for preserving speakers' voices from the source speech to the translation speech in a different language. Unlike existing approaches, the proposed method is able to preserve each speaker's voice on speaker turns without requiring for speaker segmentation. Furthermore, compared to existing approaches, it better preserves speaker's privacy and mitigates potential misuse of voice cloning for creating spoofing audio artifacts. | ['Roi Pomerantz', 'Tal Remez', 'Michelle Tadmor Ramanovich', 'Ye Jia'] | 2021-07-19 | translatotron-2-robust-direct-speech-to-1 | https://openreview.net/forum?id=HTfUrAxjPkR | https://openreview.net/pdf?id=HTfUrAxjPkR | null | ['speech-to-speech-translation', 'voice-cloning'] | ['speech', 'speech'] | [ 2.14878023e-01 4.22264516e-01 -7.33680800e-02 -2.78704941e-01
-1.40705156e+00 -7.25182414e-01 4.81392592e-01 -3.47312123e-01
-2.07706064e-01 5.54353118e-01 4.61721450e-01 -6.99731410e-01
7.49220967e-01 -3.32588464e-01 -9.63499367e-01 -3.70270312e-01
4.61788595e-01 3.15665603e-01 -3.71399224e-02 -3.85656729e-02
-3.28811318e-01 2.09068865e-01 -9.91572499e-01 3.43486041e-01
8.06629777e-01 8.71294320e-01 -2.49053221e-02 7.11436987e-01
5.73485754e-02 5.24493873e-01 -8.84549260e-01 -8.27099085e-01
1.66050464e-01 -8.07445943e-01 -7.33570814e-01 8.93493071e-02
4.46221381e-01 -5.54731548e-01 -5.26381373e-01 1.04075849e+00
8.96170437e-01 -9.20318067e-02 1.46653205e-01 -1.04784226e+00
-1.06963384e+00 1.18904686e+00 -3.14429045e-01 -1.13217436e-01
3.13528597e-01 1.04766838e-01 7.94337809e-01 -9.35153544e-01
3.10209930e-01 1.65645361e+00 7.15699375e-01 7.35883415e-01
-1.25058627e+00 -9.13590431e-01 -3.39139968e-01 -4.36572790e-01
-1.31761086e+00 -1.27778232e+00 5.84581792e-01 -2.04618961e-01
9.62598443e-01 3.70084554e-01 2.55057126e-01 1.53666151e+00
1.21323787e-01 6.42995954e-01 7.11351931e-01 -4.42604631e-01
4.27965932e-02 4.75246280e-01 -1.63356155e-01 5.03954113e-01
-2.40588859e-01 2.82464474e-01 -6.67713404e-01 -2.96365827e-01
3.03750306e-01 -6.08276606e-01 -4.05835420e-01 1.76613599e-01
-1.12809980e+00 5.67404091e-01 3.54988351e-02 1.26647979e-01
-1.70563757e-01 2.18529969e-01 5.58420956e-01 5.25941014e-01
6.08127415e-01 2.00587988e-01 -1.59572184e-01 -5.37696481e-02
-9.49243426e-01 -1.36211649e-01 8.01883280e-01 1.39946699e+00
2.48421818e-01 6.05096698e-01 -1.71549678e-01 8.69806468e-01
4.53520834e-01 9.26895976e-01 7.13101923e-01 -7.32865512e-01
6.79179311e-01 -1.95455477e-01 -1.58781279e-02 -6.27481401e-01
2.22001463e-01 -5.00630140e-01 -6.05801344e-01 -2.50960886e-01
-9.48625132e-02 -5.75893283e-01 -8.07565510e-01 1.85830045e+00
1.31657228e-01 1.50972411e-01 1.27752513e-01 7.57480383e-01
8.30134153e-01 9.36537147e-01 -2.87980705e-01 -2.11121500e-01
1.28221321e+00 -1.33209682e+00 -1.00253940e+00 -2.56245464e-01
3.23105276e-01 -1.21770799e+00 1.22247779e+00 -5.53114451e-02
-1.40847325e+00 -5.71158469e-01 -1.01663625e+00 -1.50461599e-01
-2.27361489e-02 2.73769826e-01 -4.90186326e-02 1.19852316e+00
-1.38238525e+00 2.28491321e-01 -7.22406566e-01 -2.41493553e-01
1.85809761e-01 2.72367269e-01 -2.59977579e-01 5.28477609e-01
-1.23547637e+00 6.27910674e-01 1.97801758e-02 -1.65003330e-01
-1.13851964e+00 -5.58400750e-01 -9.33671534e-01 2.20493317e-01
6.92476481e-02 -7.51577675e-01 1.77477312e+00 -1.09321356e+00
-2.31958842e+00 6.64541066e-01 -4.03279155e-01 -6.55602455e-01
5.57688832e-01 -3.23176146e-01 -9.21923280e-01 1.19063035e-01
1.34296969e-01 6.14579558e-01 1.15958464e+00 -1.10385191e+00
-5.16615808e-01 4.54017287e-03 -4.41951752e-01 6.77559897e-02
-4.13375050e-01 6.20917797e-01 -4.88870621e-01 -8.51384282e-01
-3.08031291e-01 -9.96580958e-01 2.39254117e-01 -1.29935384e-01
-8.90307903e-01 2.04992399e-01 8.64947200e-01 -1.02380860e+00
1.11858904e+00 -2.38731885e+00 -1.42011568e-01 -1.11603513e-01
-1.88590854e-01 7.01968253e-01 -3.13107103e-01 4.85425919e-01
-8.49377364e-02 3.14303339e-01 -2.84549713e-01 -1.02390599e+00
6.22429736e-02 -2.28004605e-01 -7.05455840e-01 4.00140792e-01
1.39540881e-01 8.66428792e-01 -5.65980494e-01 -7.66045675e-02
-6.74520433e-02 7.27544427e-01 -5.01024365e-01 2.37227008e-01
7.34202191e-02 3.87138009e-01 3.62203419e-02 5.59147418e-01
6.48416102e-01 2.79147893e-01 7.05290660e-02 4.20152813e-01
2.04686001e-02 1.17171383e+00 -5.85533082e-01 1.58639693e+00
-6.10388339e-01 8.05902839e-01 5.15469491e-01 -3.89993906e-01
1.08766484e+00 1.00571990e+00 -7.73552433e-02 -5.62447071e-01
3.52387488e-01 3.83468717e-01 -2.22285882e-01 -4.30240393e-01
4.62820739e-01 -2.86508173e-01 -7.47471452e-02 7.60688543e-01
1.88081756e-01 -1.40305266e-01 -4.36263055e-01 8.14923900e-04
6.69224322e-01 -3.18383873e-01 -4.15817238e-02 -1.23628877e-01
4.45557684e-01 -5.73539317e-01 5.66211402e-01 3.71730328e-01
-2.65532374e-01 6.55814469e-01 5.99569827e-02 1.42387688e-01
-1.20797265e+00 -1.13348174e+00 2.43341908e-01 1.03145027e+00
-1.25900060e-01 -3.28345686e-01 -1.18709636e+00 -5.92410803e-01
-2.79370844e-01 9.97088909e-01 -6.39414415e-02 -5.08927107e-01
-6.49563432e-01 -7.81437308e-02 1.45787358e+00 4.08091158e-01
4.31418210e-01 -9.12799835e-01 2.58030239e-02 2.37846822e-01
-6.52606905e-01 -1.19224691e+00 -1.40458810e+00 -1.79620057e-01
-4.58581328e-01 -1.80594936e-01 -6.52271867e-01 -1.01582241e+00
2.45916635e-01 4.09343153e-01 8.82267475e-01 -4.03271139e-01
2.58717924e-01 -1.42112464e-01 -1.89441219e-01 -5.42079270e-01
-1.25008512e+00 2.70332485e-01 3.33630532e-01 3.22556019e-01
2.15583518e-01 -5.60204923e-01 -7.35809207e-02 2.91600704e-01
-6.13497496e-01 -2.73170978e-01 2.38018259e-01 6.51537061e-01
1.15901522e-01 -1.24940209e-01 9.54139709e-01 -3.75022084e-01
8.67723048e-01 -3.40174377e-01 -4.41319257e-01 1.37312651e-01
-4.34234530e-01 -2.59833753e-01 8.23436320e-01 -3.97395492e-01
-1.00580788e+00 -7.05792755e-02 -4.23675954e-01 -5.24482906e-01
-2.47855201e-01 1.34664821e-02 -6.01094782e-01 1.13188632e-01
5.74517906e-01 4.20607299e-01 2.77112126e-01 -6.02866530e-01
5.52268088e-01 1.40791416e+00 7.62563348e-01 -6.94892481e-02
7.16986477e-01 1.94577277e-01 -8.28814387e-01 -8.36582899e-01
-3.34422946e-01 -1.30050540e-01 -3.95143300e-01 2.22660914e-01
6.64538682e-01 -1.12563801e+00 -5.43479502e-01 6.04640365e-01
-1.44887662e+00 -8.53809938e-02 -2.47309357e-01 6.12812817e-01
-4.46432143e-01 3.95145655e-01 -9.00858283e-01 -8.32679451e-01
-7.14075923e-01 -1.33187401e+00 1.14505100e+00 -1.21075355e-01
-3.11656296e-01 -7.17458665e-01 -2.58942004e-02 7.48346031e-01
5.62223792e-01 -2.96890318e-01 6.55695975e-01 -7.50505149e-01
-3.73682231e-01 -1.09261826e-01 1.47133827e-01 7.79912353e-01
3.71840447e-01 -2.15151180e-02 -1.26541436e+00 -4.97042567e-01
1.83925509e-01 -2.34345436e-01 7.92159379e-01 1.89768940e-01
4.94008482e-01 -8.23360443e-01 -4.94901165e-02 6.70122683e-01
6.90891206e-01 5.26081026e-01 5.74530721e-01 -2.00569838e-01
5.95166564e-01 5.35577118e-01 -2.11064480e-02 -3.16285156e-02
2.19766662e-01 7.23263681e-01 1.78786561e-01 -1.53219521e-01
-6.17346168e-01 -6.05159819e-01 1.15027094e+00 1.41653681e+00
5.92752159e-01 -6.43174529e-01 -4.39335197e-01 7.20628619e-01
-1.36604655e+00 -8.78224313e-01 1.73448510e-02 2.29053116e+00
9.62204814e-01 -1.61180347e-01 3.53726417e-01 9.26794708e-02
1.14197052e+00 2.06452653e-01 -3.76222640e-01 -9.29675639e-01
-6.04937226e-02 1.28690928e-01 3.41620415e-01 8.72305930e-01
-8.74279022e-01 1.29820228e+00 7.27806425e+00 7.11240590e-01
-1.36403120e+00 4.49709177e-01 5.02435088e-01 -2.90598959e-01
-4.61757541e-01 -1.84977517e-01 -6.70090377e-01 5.17667949e-01
1.43088591e+00 -1.57372966e-01 6.80580974e-01 5.91745019e-01
4.41724360e-01 7.86950469e-01 -1.01316130e+00 7.51927793e-01
3.03618610e-01 -1.12421834e+00 1.82372048e-01 1.11166954e-01
4.43208128e-01 2.25227363e-02 3.31588656e-01 3.37887406e-02
2.96133071e-01 -1.03665817e+00 1.22407651e+00 -7.04563260e-02
1.09578645e+00 -1.03431058e+00 4.90439147e-01 2.92157263e-01
-9.04290199e-01 1.32784322e-01 -1.14356831e-01 3.52229983e-01
5.28346837e-01 4.15814430e-01 -1.24591148e+00 5.56065917e-01
4.48008329e-01 2.86438227e-01 1.93069298e-02 6.94193721e-01
-4.04750377e-01 1.08587456e+00 -1.47595987e-01 1.68239832e-01
2.19873801e-01 1.05684780e-01 8.47105205e-01 1.33222187e+00
6.70452714e-01 -4.89645690e-01 -8.56496245e-02 8.70551229e-01
-6.17062807e-01 6.82223514e-02 -6.36766076e-01 -3.65911484e-01
9.96013701e-01 7.50434399e-01 -8.12185407e-02 -2.94996619e-01
-1.42207757e-01 1.32536221e+00 -1.03549331e-01 3.93010676e-01
-8.88871968e-01 -6.91613853e-01 8.73058915e-01 -6.65646344e-02
1.91932797e-01 6.93743769e-03 -3.01165044e-01 -9.99495506e-01
3.25655311e-01 -1.30715418e+00 -2.12371111e-01 -6.51221395e-01
-9.33221459e-01 1.12501240e+00 -7.56335795e-01 -1.03295696e+00
-4.45063502e-01 4.68090875e-03 -5.46518743e-01 1.11767721e+00
-1.19274056e+00 -1.33188212e+00 4.57361430e-01 6.49046302e-01
6.72103047e-01 -5.11727452e-01 1.03433418e+00 4.98966575e-01
-6.57806158e-01 1.31933320e+00 8.20382088e-02 3.60281348e-01
9.67311800e-01 -7.52373695e-01 1.22246861e+00 1.16682041e+00
2.22951561e-01 6.26880050e-01 5.69178045e-01 -5.43969393e-01
-1.08060277e+00 -1.31081593e+00 1.55856383e+00 -2.80021667e-01
4.96083200e-01 -6.65052831e-01 -8.06582332e-01 9.18183208e-01
8.76056373e-01 -4.76328492e-01 8.53924990e-01 -3.34956467e-01
-6.66958749e-01 -1.64279029e-01 -1.11699033e+00 6.03806913e-01
8.81818593e-01 -1.06723523e+00 -5.79934239e-01 1.70938820e-01
1.41784728e+00 -3.85211945e-01 -5.49818814e-01 -1.18068323e-01
4.83761132e-01 -6.89821124e-01 8.01941752e-01 -4.97536927e-01
1.52041078e-01 -1.88815832e-01 -2.17763662e-01 -1.47029412e+00
-1.92068398e-01 -1.61664712e+00 7.73966908e-02 1.81140399e+00
7.18878388e-01 -8.25046599e-01 2.56505191e-01 2.73585826e-01
-6.23908579e-01 -1.88182488e-01 -1.18697095e+00 -1.08749986e+00
1.50889322e-01 -3.13830465e-01 1.01677585e+00 9.74288464e-01
-7.82371778e-03 6.67295098e-01 -9.21000719e-01 5.51054955e-01
4.26620811e-01 -1.85720459e-01 7.25610793e-01 -5.44587433e-01
-4.61561978e-01 -4.85589445e-01 1.09222122e-01 -1.29175162e+00
3.16878289e-01 -1.05643976e+00 2.93222010e-01 -9.50385571e-01
-2.71962374e-01 -1.06819116e-01 -1.13430178e-04 4.87773657e-01
2.48138644e-02 1.52124926e-01 1.93494275e-01 -7.53795579e-02
5.89526072e-02 8.58934045e-01 1.07552862e+00 -1.59507021e-01
-2.63584048e-01 2.81636715e-01 -9.83928919e-01 4.83004719e-01
9.19404149e-01 -7.51760185e-01 -4.63609368e-01 -8.06831419e-01
-3.83174241e-01 2.70917207e-01 1.24319769e-01 -8.21031153e-01
1.48934767e-01 2.10636035e-01 -3.69713694e-01 -3.26916367e-01
3.96658719e-01 -6.24037325e-01 1.20672300e-01 3.51803303e-01
-6.50497735e-01 8.89217705e-02 1.54450044e-01 4.01059568e-01
-3.31339687e-01 6.07623346e-02 1.01022148e+00 1.97130650e-01
2.69078672e-01 2.22189471e-01 -8.40880215e-01 -1.24781996e-01
6.46760941e-01 1.78994611e-01 -4.06613201e-01 -7.09826827e-01
-4.30543035e-01 -6.46059290e-02 3.04959774e-01 8.77919793e-01
4.28395778e-01 -1.45715272e+00 -1.00622106e+00 5.64059258e-01
-1.92503080e-01 -4.29663599e-01 -1.35609865e-01 6.28241241e-01
-2.78577715e-01 5.47560513e-01 1.69693276e-01 -2.97484934e-01
-1.41843820e+00 5.14924645e-01 4.08340603e-01 3.30407381e-01
-5.54557145e-01 1.11823630e+00 -5.15220538e-02 -7.09704161e-01
4.98007506e-01 -4.25245166e-01 5.55008292e-01 -2.86243439e-01
7.14771688e-01 2.88463116e-01 2.36271843e-01 -1.08548748e+00
-3.84814858e-01 3.36622670e-02 -5.68210520e-02 -6.91038311e-01
9.24973488e-01 -6.63988709e-01 -1.17906995e-01 3.05803150e-01
1.35428572e+00 4.69265759e-01 -9.82589722e-01 -4.42284614e-01
-3.82969677e-01 -3.56424630e-01 -6.95221964e-03 -9.04868126e-01
-1.12792444e+00 1.18258095e+00 3.34661186e-01 2.59546965e-01
9.23333526e-01 -1.33939132e-01 1.56267154e+00 1.09367378e-01
1.95091590e-01 -9.09446955e-01 -1.70391470e-01 5.74377775e-01
9.69382405e-01 -9.22035158e-01 -7.84933507e-01 -3.82676512e-01
-7.78283656e-01 8.25192809e-01 1.20839313e-01 3.07579488e-01
5.27793050e-01 4.52097863e-01 5.92086554e-01 5.02282619e-01
-7.82777190e-01 2.23440319e-01 2.42847741e-01 7.68896043e-01
5.19101381e-01 2.95884967e-01 2.34687760e-01 5.68611801e-01
-8.37396026e-01 -3.35159898e-01 4.25277621e-01 4.97637957e-01
-3.28540385e-01 -1.22010493e+00 -4.59704429e-01 -2.50634342e-01
-8.93564284e-01 -4.03178692e-01 -8.81918550e-01 1.99847773e-01
-7.40257353e-02 1.61895430e+00 3.83489691e-02 -6.82496607e-01
2.30976343e-01 2.95592964e-01 -2.12465286e-01 -5.62733412e-01
-1.02843082e+00 5.21968603e-01 5.26377976e-01 -4.62505758e-01
1.52706373e-02 -5.73003709e-01 -1.04104757e+00 -6.81766450e-01
-4.89704281e-01 1.44272238e-01 8.64514530e-01 7.87477136e-01
8.69074225e-01 5.17573357e-01 1.02877128e+00 -4.22048360e-01
-7.99692869e-01 -9.74493861e-01 -5.00256658e-01 -4.04827446e-02
8.60502064e-01 1.80689141e-01 -2.93590307e-01 2.21925631e-01] | [14.741936683654785, 6.8988752365112305] |
1afe0f0a-1e98-49e1-893c-abbc1dc1133f | visual-question-answering-from-another-1 | 2212.01639 | null | https://arxiv.org/abs/2212.01639v1 | https://arxiv.org/pdf/2212.01639v1.pdf | Visual Question Answering From Another Perspective: CLEVR Mental Rotation Tests | Different types of mental rotation tests have been used extensively in psychology to understand human visual reasoning and perception. Understanding what an object or visual scene would look like from another viewpoint is a challenging problem that is made even harder if it must be performed from a single image. We explore a controlled setting whereby questions are posed about the properties of a scene if that scene was observed from another viewpoint. To do this we have created a new version of the CLEVR dataset that we call CLEVR Mental Rotation Tests (CLEVR-MRT). Using CLEVR-MRT we examine standard methods, show how they fall short, then explore novel neural architectures that involve inferring volumetric representations of a scene. These volumes can be manipulated via camera-conditioned transformations to answer the question. We examine the efficacy of different model variants through rigorous ablations and demonstrate the efficacy of volumetric representations. | ['Christopher Pal', 'Derek Nowrouzezahrai', 'Sina Honari', 'Florian Golemo', 'Martin Weiss', 'Christopher Beckham'] | 2022-12-03 | visual-question-answering-from-another | https://openreview.net/forum?id=aYbCpFNnHdh | https://openreview.net/pdf?id=aYbCpFNnHdh | null | ['visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning'] | [ 3.81581217e-01 1.87815472e-01 2.99387246e-01 -3.35698217e-01
-2.67810941e-01 -9.58768129e-01 9.03402805e-01 -1.20258048e-01
-5.15586078e-01 2.37516686e-01 1.33707091e-01 -5.99358916e-01
-9.53210443e-02 -6.13960862e-01 -8.57540607e-01 -3.89948696e-01
1.58880979e-01 4.13307309e-01 4.23557572e-02 -7.62694925e-02
5.23328483e-01 5.83882272e-01 -1.34365368e+00 3.29837084e-01
1.79648176e-01 3.59826446e-01 2.86908895e-01 6.73833370e-01
3.86672348e-01 7.37444639e-01 -3.80059153e-01 -2.42311671e-01
5.47313035e-01 -5.51850855e-01 -1.10196972e+00 3.19442928e-01
6.86458170e-01 -3.51421803e-01 -3.23978633e-01 9.51881170e-01
-1.78321544e-02 4.64315832e-01 8.66724670e-01 -1.03986740e+00
-8.94297898e-01 2.26029843e-01 -7.39437401e-01 7.07112789e-01
6.13179982e-01 3.50257188e-01 9.10799742e-01 -7.73052335e-01
8.28669906e-01 1.69801319e+00 1.88137233e-01 4.07562345e-01
-1.79395235e+00 -1.62877604e-01 3.29553068e-01 1.15055762e-01
-1.09670377e+00 -5.10308683e-01 5.42757571e-01 -6.25962198e-01
8.54029357e-01 1.93064854e-01 7.23136842e-01 1.09095383e+00
2.12060466e-01 3.15173894e-01 1.53674603e+00 -2.84837216e-01
2.06027061e-01 -6.08360507e-02 1.64139301e-01 7.31654882e-01
4.92898971e-01 5.50672598e-02 -3.43684375e-01 2.07173362e-01
1.11804938e+00 -2.43020788e-01 -3.63554686e-01 -6.43050790e-01
-1.45969105e+00 7.80355871e-01 6.16030455e-01 5.45742549e-02
-8.75650123e-02 1.82483330e-01 -3.64037789e-02 2.69050300e-01
-3.61033380e-02 9.18398201e-01 -1.83302566e-01 3.71150315e-01
-5.82677364e-01 4.64386940e-01 2.87786514e-01 6.31641507e-01
5.59296727e-01 8.16994831e-02 -2.92004108e-01 4.44186389e-01
4.39283699e-01 2.39942044e-01 2.85938889e-01 -1.21138227e+00
3.69396478e-01 2.84627765e-01 1.50574371e-01 -7.29470372e-01
-6.30927086e-01 -1.91189378e-01 -2.78018951e-01 5.61822236e-01
7.74242163e-01 1.89163864e-01 -1.16304338e+00 2.02989483e+00
2.48796269e-01 -6.40096962e-02 -1.86992288e-01 1.04389572e+00
6.18356943e-01 2.58381039e-01 1.55381069e-01 6.11071326e-02
1.44735944e+00 -5.04911900e-01 -2.60903716e-01 -6.89679921e-01
5.61549664e-01 -3.80780727e-01 1.43172169e+00 4.03916031e-01
-1.16150987e+00 -5.69527030e-01 -1.09739542e+00 -6.66770101e-01
-2.92775333e-01 -2.51913607e-01 7.83554792e-01 8.09192717e-01
-1.20734990e+00 3.58942926e-01 -8.00566077e-01 -3.97484660e-01
6.33405089e-01 9.34859067e-02 -4.73051041e-01 -1.49681747e-01
-6.94504619e-01 1.32093096e+00 -3.62059660e-02 5.41418456e-02
-1.43735516e+00 -4.87602353e-01 -1.05549884e+00 8.32046643e-02
2.31246814e-01 -1.04537189e+00 1.13898182e+00 -7.39823282e-01
-1.00658727e+00 1.45653915e+00 -3.91121000e-01 -1.93877742e-01
3.39753658e-01 -9.11583155e-02 1.83582008e-01 3.62883031e-01
4.43433113e-02 6.52776718e-01 9.96774077e-01 -1.39572453e+00
2.70113319e-01 -6.61190927e-01 5.20465910e-01 2.55612463e-01
4.71543372e-01 -2.94467658e-02 -3.21272165e-02 -4.91520673e-01
3.91240001e-01 -1.17601395e+00 -5.87162636e-02 1.51890844e-01
-3.27649713e-01 1.91106290e-01 4.24576283e-01 -5.44698954e-01
2.33475477e-01 -2.09210777e+00 2.65903860e-01 8.13952684e-02
5.53245485e-01 6.46455511e-02 -3.67320538e-01 1.10038996e-01
-4.08437520e-01 4.26135451e-01 -2.91524589e-01 -6.52525350e-02
-7.25110695e-02 -2.50114072e-02 -5.30538201e-01 8.65043461e-01
3.58352870e-01 1.16949284e+00 -6.75419450e-01 -2.40637556e-01
3.19429040e-01 5.46284258e-01 -8.69313538e-01 -1.48283973e-01
-1.79962069e-01 9.06301737e-01 -2.51899213e-01 2.51300603e-01
6.26662374e-01 -2.67146438e-01 2.42876336e-01 -5.22903427e-02
7.73082599e-02 3.88238311e-01 -7.82807291e-01 1.55508614e+00
-2.37328678e-01 1.00637233e+00 -1.39789224e-01 -1.04969144e+00
5.97651005e-01 -1.12335779e-01 -6.93992302e-02 -8.14021587e-01
2.87593454e-01 -5.29352009e-01 4.37177628e-01 -6.90855026e-01
2.75920570e-01 -6.63407624e-01 1.62492007e-01 5.63310683e-01
3.82019691e-02 -6.76079154e-01 2.04767004e-01 2.35041738e-01
1.08366513e+00 4.76182312e-01 4.02902991e-01 -2.13453382e-01
1.10567391e-01 -2.52308864e-02 2.52230793e-01 8.48701119e-01
-3.68824303e-01 7.94939578e-01 8.12263727e-01 -2.20716864e-01
-1.07661128e+00 -1.59650540e+00 -3.07958841e-01 9.64455724e-01
2.61989146e-01 -5.67204282e-02 -5.45600116e-01 -2.74984777e-01
-8.03872570e-02 9.89105940e-01 -1.09610915e+00 -3.27859253e-01
-2.59947330e-01 -6.68211460e-01 3.58047396e-01 5.34703135e-01
3.12305748e-01 -1.11556447e+00 -1.14777827e+00 -5.62639415e-01
-3.28849666e-02 -1.34679365e+00 1.06972203e-01 -1.20240420e-01
-9.03910935e-01 -1.21735132e+00 -3.46095771e-01 -5.02303183e-01
7.88367629e-01 6.02812290e-01 1.36509240e+00 8.65650643e-03
-4.45119798e-01 6.85725391e-01 -2.93078065e-01 -3.63121003e-01
-1.49430126e-01 -3.80416751e-01 1.55890780e-02 -1.53821304e-01
1.20977663e-01 -5.94797671e-01 -4.41999555e-01 2.55470753e-01
-1.02308691e+00 1.40079498e-01 5.71380258e-01 3.30655187e-01
4.35072452e-01 -3.09557587e-01 2.00960904e-01 -1.00367785e+00
6.13272130e-01 -2.87856549e-01 -5.61782241e-01 1.37645274e-01
-4.34710681e-02 2.60592282e-01 4.09653097e-01 -4.17892814e-01
-1.07410443e+00 -2.13805363e-01 2.79610246e-01 -3.19896698e-01
-1.93543747e-01 2.97923207e-01 -7.16496632e-02 -8.79765451e-02
7.48725533e-01 -5.53143173e-02 -2.40930915e-01 -1.18885137e-01
5.92784584e-01 -2.01232210e-01 4.95130599e-01 -6.66693866e-01
1.10189402e+00 9.56594169e-01 2.59483635e-01 -9.94295835e-01
-1.03262436e+00 -8.96599218e-02 -9.18053389e-01 -1.48309529e-01
1.13974822e+00 -7.73151278e-01 -9.92516339e-01 -9.50779319e-02
-9.77659106e-01 -6.35631204e-01 -2.04753876e-01 5.98802269e-01
-7.62298763e-01 3.43462437e-01 -1.32856935e-01 -6.04617655e-01
4.84925717e-01 -1.23385727e+00 1.03644872e+00 3.65209058e-02
-4.76900518e-01 -9.70554411e-01 1.18768245e-01 4.76782709e-01
1.76460907e-01 2.29119524e-01 1.27573144e+00 -2.03325927e-01
-7.12226808e-01 2.78877113e-02 -5.57407260e-01 -6.46806955e-02
-2.24531770e-01 -1.63433686e-01 -1.24985552e+00 -3.95683557e-01
3.11439872e-01 -5.31351030e-01 1.13746381e+00 5.80321550e-01
1.44627523e+00 1.48863524e-01 -2.59650107e-02 9.93606567e-01
1.35005474e+00 -8.41878504e-02 9.72142339e-01 2.87273109e-01
5.18470407e-01 7.62099504e-01 2.91171163e-01 -1.14324316e-02
3.26504439e-01 5.03595114e-01 4.22357142e-01 -2.60277670e-02
-1.62440255e-01 -2.40679070e-01 2.31680855e-01 1.20351188e-01
-3.51021647e-01 -9.42878276e-02 -8.66641939e-01 3.91167402e-01
-1.13944161e+00 -1.20983887e+00 -5.23827747e-02 2.15176272e+00
4.06676888e-01 2.32874990e-01 -7.96990916e-02 1.68583095e-02
3.16878647e-01 4.70155120e-01 -7.02934325e-01 -5.31033933e-01
-7.49348924e-02 3.64714235e-01 8.33306164e-02 4.88556683e-01
-7.94172883e-01 9.01528180e-01 7.46898556e+00 -9.57772732e-02
-9.22888577e-01 -8.94393474e-02 6.93458200e-01 -1.15952894e-01
-3.29029918e-01 3.86333436e-01 -3.84737223e-01 -3.06244522e-01
7.01447904e-01 -4.72379057e-03 8.05135787e-01 5.78135967e-01
2.61528403e-01 -6.34923398e-01 -1.53874874e+00 8.97282302e-01
4.59366769e-01 -8.75752091e-01 1.42925948e-01 1.42325521e-01
4.31901425e-01 -3.95547189e-02 6.86835885e-01 2.87727326e-01
3.63495320e-01 -1.63923275e+00 5.76191664e-01 6.28658712e-01
7.54366040e-01 -1.23740517e-01 -4.57082875e-02 1.96436360e-01
-7.73389101e-01 9.66630504e-02 -5.59172332e-01 -2.57407069e-01
-2.48060003e-03 2.97275543e-01 -8.06489348e-01 -3.95692177e-02
5.02709985e-01 3.52430701e-01 -1.11001253e+00 9.81220484e-01
-3.14386815e-01 2.96617091e-01 3.27465497e-02 3.79895210e-01
-1.54045403e-01 -2.85959125e-01 4.40223485e-01 6.40013695e-01
3.42520662e-02 1.22555206e-02 -2.51622319e-01 1.45913315e+00
-4.03982289e-02 -2.76424497e-01 -9.03991163e-01 -6.50685048e-04
3.59693193e-03 1.22823906e+00 -9.46292400e-01 -8.84999707e-02
-3.33636850e-01 9.70786512e-01 6.85734034e-01 6.93538189e-01
-8.06840658e-01 1.89259410e-01 6.24085546e-01 2.24485114e-01
1.77309766e-01 -6.65037036e-01 -2.66097307e-01 -1.43094373e+00
-6.10608142e-03 -7.63268650e-01 -6.32437170e-02 -1.54955161e+00
-8.51291358e-01 4.07921761e-01 3.69853765e-01 -9.07014489e-01
-1.66324049e-01 -9.65828538e-01 -5.21354496e-01 8.21060538e-01
-1.38765967e+00 -7.78246105e-01 -6.03966653e-01 6.41182363e-01
4.52831298e-01 2.68596947e-01 6.72289014e-01 -4.17613506e-01
-4.30037618e-01 1.30057901e-01 -3.63561273e-01 1.78981304e-01
6.06033385e-01 -1.16331744e+00 4.42189932e-01 9.36230183e-01
5.11769593e-01 1.03796005e+00 6.80299878e-01 -4.28431362e-01
-1.38521814e+00 -6.87422633e-01 2.65146732e-01 -1.20999408e+00
3.53955269e-01 -5.71380734e-01 -9.21242177e-01 1.27613640e+00
9.51775536e-02 6.15829043e-02 4.78024185e-01 1.07689440e-01
-8.78251851e-01 3.75824004e-01 -9.46162939e-01 8.46542180e-01
1.30339682e+00 -8.80561233e-01 -9.20893729e-01 2.47901917e-01
5.92844486e-01 -1.01440549e-01 -4.19267774e-01 2.99491346e-01
5.33329248e-01 -1.25790811e+00 1.19970334e+00 -1.02858007e+00
6.10436738e-01 -2.31679499e-01 -2.50165045e-01 -1.44399011e+00
-3.92477423e-01 -9.28245410e-02 2.50930488e-01 6.52401209e-01
3.67401838e-01 -6.20154440e-01 2.94499964e-01 6.70520246e-01
2.30913669e-01 -1.85077772e-01 -6.52871251e-01 -6.62797868e-01
3.08360934e-01 -5.05744100e-01 1.66403100e-01 1.05128026e+00
-1.25186697e-01 4.40309912e-01 8.41648281e-02 1.79686114e-01
4.56377923e-01 1.01544656e-01 8.38093042e-01 -1.17440367e+00
-4.26105320e-01 -3.95656377e-01 -5.81029117e-01 -8.74906898e-01
1.95395902e-01 -1.00130129e+00 -2.38341138e-01 -1.51638508e+00
4.69369262e-01 3.78708877e-02 1.44349858e-01 2.09575221e-01
-1.46416828e-01 3.05862725e-01 3.87635112e-01 1.67645350e-01
-4.87916440e-01 3.23105246e-01 1.58313954e+00 1.21479101e-01
3.48915160e-02 -4.12092716e-01 -8.93647134e-01 8.54666710e-01
8.08708787e-01 -1.85180306e-01 -6.63865209e-01 -5.67748427e-01
7.42170691e-01 7.71698505e-02 9.95970905e-01 -8.19038332e-01
-3.50784004e-01 -2.37656817e-01 8.34899187e-01 -1.74801022e-01
6.29855275e-01 -4.06296551e-01 -1.58025727e-01 2.78234184e-01
-6.33341312e-01 4.00728524e-01 2.87000597e-01 5.63729227e-01
1.82860032e-01 -1.51497766e-01 9.75904942e-01 -4.23824966e-01
-6.09110057e-01 8.85545462e-02 -5.27761638e-01 4.37752336e-01
1.10392416e+00 -8.88457224e-02 -6.69298828e-01 -4.70629245e-01
-9.59681213e-01 -2.07399085e-01 7.35229850e-01 2.73239225e-01
8.71917486e-01 -9.33478355e-01 -6.28135622e-01 2.75981158e-01
1.54434502e-01 -1.35218605e-01 2.16370523e-01 1.02079427e+00
-7.02020347e-01 4.60890055e-01 -4.85600263e-01 -4.82181072e-01
-9.83484566e-01 1.06010520e+00 4.92462069e-01 3.98515128e-02
-7.97450602e-01 7.59368181e-01 1.06223249e+00 -3.42371315e-01
-4.76432085e-01 -4.83820170e-01 -2.64209032e-01 -2.20162973e-01
4.77984279e-01 -1.78988934e-01 -1.91892058e-01 -7.95068383e-01
-2.17177644e-01 6.13446295e-01 1.30023912e-01 -5.75154185e-01
1.10816538e+00 -2.72309959e-01 -4.11553867e-02 7.20701277e-01
9.25088465e-01 1.12616882e-01 -1.26183379e+00 2.90837400e-02
-3.93289328e-01 -8.92113686e-01 -2.81971954e-02 -5.92861116e-01
-8.41914713e-01 1.02585208e+00 3.37990046e-01 7.86915347e-02
9.30346191e-01 2.73083717e-01 -8.00731033e-02 4.64391768e-01
2.32492536e-01 -6.06699884e-01 2.77908713e-01 4.12606418e-01
1.11162174e+00 -1.20045960e+00 3.66021305e-01 -2.60915339e-01
-8.82471263e-01 7.01009989e-01 5.48470080e-01 -5.48903704e-01
4.97236311e-01 -2.67144859e-01 -1.34362295e-01 -4.63339925e-01
-7.05134392e-01 -4.26764220e-01 4.08288002e-01 8.08427215e-01
4.53711987e-01 -5.08482158e-02 1.81631133e-01 1.99119095e-02
-5.86801469e-01 -4.51106369e-01 9.80136037e-01 6.60549164e-01
-2.71917909e-01 -4.66405332e-01 -6.19872153e-01 5.72091877e-01
-1.44705206e-01 -1.00193657e-01 -6.71432137e-01 9.45453227e-01
-1.33455932e-01 7.95365632e-01 1.41368374e-01 -1.72800109e-01
2.49031652e-02 1.25281110e-01 1.24357450e+00 -1.00622869e+00
-5.81348836e-02 -1.14966579e-01 -1.09817810e-01 -6.39322042e-01
-6.34104669e-01 -8.65948975e-01 -1.04441726e+00 -2.01988965e-01
1.88401863e-02 -5.01984417e-01 5.76482654e-01 1.15543747e+00
-4.95564304e-02 6.06173873e-01 1.72313854e-01 -9.79293585e-01
-1.91763192e-01 -8.74565065e-01 -5.05706549e-01 7.13008642e-01
4.58203316e-01 -1.01577985e+00 -4.69640702e-01 1.36522278e-01] | [10.535155296325684, 2.08100962638855] |
21f529d8-6994-47e1-be67-bd605a31cb0f | jailbreaking-chatgpt-via-prompt-engineering | 2305.13860 | null | https://arxiv.org/abs/2305.13860v1 | https://arxiv.org/pdf/2305.13860v1.pdf | Jailbreaking ChatGPT via Prompt Engineering: An Empirical Study | Large Language Models (LLMs), like ChatGPT, have demonstrated vast potential but also introduce challenges related to content constraints and potential misuse. Our study investigates three key research questions: (1) the number of different prompt types that can jailbreak LLMs, (2) the effectiveness of jailbreak prompts in circumventing LLM constraints, and (3) the resilience of ChatGPT against these jailbreak prompts. Initially, we develop a classification model to analyze the distribution of existing prompts, identifying ten distinct patterns and three categories of jailbreak prompts. Subsequently, we assess the jailbreak capability of prompts with ChatGPT versions 3.5 and 4.0, utilizing a dataset of 3,120 jailbreak questions across eight prohibited scenarios. Finally, we evaluate the resistance of ChatGPT against jailbreak prompts, finding that the prompts can consistently evade the restrictions in 40 use-case scenarios. The study underscores the importance of prompt structures in jailbreaking LLMs and discusses the challenges of robust jailbreak prompt generation and prevention. | ['Yang Liu', 'Tianwei Zhang', 'Lida Zhao', 'Ying Zhang', 'Yaowen Zheng', 'Yuekang Li', 'Zhengzi Xu', 'Gelei Deng', 'Yi Liu'] | 2023-05-23 | null | null | null | null | ['prompt-engineering'] | ['natural-language-processing'] | [-1.96018979e-01 -1.60652861e-01 -3.26029509e-01 -4.92914207e-02
-9.65204954e-01 -1.16536069e+00 7.46352375e-01 3.69675338e-01
-3.15246373e-01 3.61360729e-01 5.60616314e-01 -1.15725327e+00
-5.68597376e-01 -2.53126889e-01 -2.96515167e-01 8.17794502e-02
3.30745041e-01 3.19726259e-01 3.17094624e-01 -4.74246740e-01
8.11563075e-01 4.04983431e-01 -1.02859724e+00 5.61594725e-01
1.09840488e+00 1.72881827e-01 2.55352318e-01 1.11508155e+00
-6.27581954e-01 1.22559547e+00 -1.39677751e+00 -2.16587290e-01
1.16794975e-02 -5.22217825e-02 -1.00181699e+00 -3.98856670e-01
8.76755834e-01 -6.06343389e-01 -1.90618426e-01 3.83020312e-01
4.24118578e-01 1.35819942e-01 2.72422314e-01 -1.51435184e+00
-7.56474137e-01 7.20907867e-01 -8.38584080e-02 5.65786779e-01
8.10133636e-01 3.98548752e-01 9.13012505e-01 -4.51302409e-01
7.11512983e-01 1.30885518e+00 6.91688120e-01 6.10411227e-01
-1.02982712e+00 -6.90961838e-01 -1.75612241e-01 -1.39875144e-01
-1.16583061e+00 -4.60510135e-01 2.62434244e-01 -6.63330078e-01
1.52495968e+00 8.49479377e-01 2.89321374e-02 1.22759223e+00
4.70042139e-01 4.70332235e-01 9.63712454e-01 -5.91284990e-01
6.22459464e-02 2.67751157e-01 6.32750511e-01 2.53141940e-01
-3.27114314e-02 -3.65432203e-01 -5.66026628e-01 -1.14930534e+00
3.88815254e-01 -2.66156018e-01 1.72748312e-01 8.03109646e-01
-7.86111414e-01 8.70883107e-01 -3.81843001e-01 6.06607497e-01
-4.85335514e-02 -1.67456508e-01 3.31674695e-01 3.72472584e-01
2.39199236e-01 1.09274316e+00 -6.69855595e-01 -6.10192060e-01
-8.70025218e-01 5.54777682e-01 1.29181147e+00 1.07803810e+00
5.01913965e-01 -1.43832490e-01 -4.68252480e-01 1.17630768e+00
1.79757133e-01 6.38249815e-01 1.56014025e-01 -8.74441683e-01
8.11901987e-01 5.75205326e-01 3.11812550e-01 -1.03073263e+00
-4.61617023e-01 -4.22476791e-02 1.07597187e-01 -4.86756116e-01
5.93280256e-01 -4.36878085e-01 -3.61377686e-01 1.49427450e+00
-9.38393548e-02 -2.24757493e-01 -5.40431440e-01 2.16488108e-01
6.34763598e-01 8.33634377e-01 3.79612982e-01 -1.18469164e-01
8.56379509e-01 -5.21264136e-01 -7.26076722e-01 -4.58721459e-01
1.14317727e+00 -1.44130909e+00 1.61655653e+00 5.18721379e-02
-8.11886430e-01 -1.87847182e-01 -4.61167067e-01 -1.38526820e-02
-3.14655423e-01 -1.70824066e-01 5.11340201e-01 1.02870309e+00
-1.12084532e+00 6.08340502e-02 -2.57908881e-01 -6.90183401e-01
-2.95059264e-01 -1.89084828e-01 1.85021907e-01 8.97098333e-03
-1.20744336e+00 1.02100348e+00 -5.80247752e-02 -3.30344200e-01
-5.77423990e-01 -1.25069082e+00 -4.96118605e-01 6.17587566e-02
5.16160369e-01 -3.54635455e-02 1.64086318e+00 -2.33862307e-02
-1.05723536e+00 3.56905460e-01 -4.72793728e-02 -1.17784888e-02
3.64406884e-01 -4.52202737e-01 -6.25895262e-01 -5.30300811e-02
3.89777452e-01 1.14551641e-01 4.90205795e-01 -9.26522851e-01
-4.80536520e-01 3.31174761e-01 1.93337306e-01 -1.78554311e-01
-4.71535683e-01 8.09101105e-01 1.12331193e-02 -4.70557392e-01
-6.62403703e-01 -8.06017697e-01 -2.76381820e-01 -6.49480224e-01
-7.55919814e-01 -5.45066774e-01 9.19873953e-01 -8.25648963e-01
1.97982550e+00 -1.87191498e+00 -6.79765105e-01 2.17282087e-01
1.70285881e-01 6.19476736e-01 -6.47094488e-01 1.44708633e+00
3.76314044e-01 9.73558187e-01 2.71170050e-01 -1.16385579e-01
1.00778341e-01 3.29890400e-01 -6.82882905e-01 -1.10312812e-02
-1.35464326e-01 8.80623460e-01 -7.00379074e-01 -5.04930913e-01
3.02771360e-01 9.42104012e-02 -8.02203119e-01 3.75390321e-01
-5.39921165e-01 1.62152633e-01 -4.68028098e-01 6.58786952e-01
5.22021234e-01 -3.13014537e-01 1.71754479e-01 6.64064646e-01
-6.73966348e-01 9.90246713e-01 -5.99160016e-01 1.11709976e+00
-5.51996768e-01 7.03687847e-01 1.13012582e-01 1.16076566e-01
8.90589297e-01 2.40665257e-01 1.87504202e-01 -7.70348370e-01
-1.92160249e-01 3.10985178e-01 -1.46636114e-01 -9.61216271e-01
7.51995742e-01 4.23121810e-01 -5.09888887e-01 1.16184175e+00
-1.07094370e-01 -2.04781011e-01 2.72314519e-01 8.17567945e-01
1.49489701e+00 -5.79949021e-01 3.81787308e-02 -2.72352964e-01
-4.84218542e-03 2.84332335e-02 1.18589826e-01 1.73596323e+00
-1.43525779e-01 1.68683678e-01 6.63736165e-01 -2.07824752e-01
-8.52539480e-01 -7.30994105e-01 -5.40136243e-04 1.55156648e+00
-3.18958938e-01 -9.74051237e-01 -7.74073124e-01 -9.14898574e-01
5.49820215e-02 1.16312230e+00 -2.22906455e-01 -3.76462154e-02
-7.11776376e-01 -3.33842486e-01 1.15126109e+00 2.01389268e-01
3.06802064e-01 -1.08648562e+00 -7.59991765e-01 2.92601526e-01
-7.77400076e-01 -1.13625956e+00 -8.39164376e-01 -1.14077419e-01
-4.09672409e-01 -1.18253589e+00 -1.11756429e-01 -2.28328615e-01
1.73073530e-01 4.95834827e-01 1.15684426e+00 8.43757451e-01
-2.19592080e-01 8.63929033e-01 -6.11736178e-01 -1.88934773e-01
-9.62608576e-01 4.79878992e-01 -1.49875328e-01 -7.32198775e-01
4.62961942e-01 -3.75797421e-01 2.97453199e-02 3.78881216e-01
-9.75586474e-01 -3.50466728e-01 3.41388345e-01 4.39187676e-01
-4.78443354e-01 -4.79186624e-01 7.94172823e-01 -1.04558694e+00
1.53119564e+00 -7.73577690e-01 -2.12560415e-01 5.87771773e-01
-5.99009156e-01 -2.90970176e-01 6.73697650e-01 -5.20264983e-01
-1.25228047e+00 -9.05113459e-01 -5.67879617e-01 3.36502731e-01
-5.09735584e-01 6.37443721e-01 2.03469649e-01 -6.91372603e-02
9.94969070e-01 -8.16371441e-02 -5.89649200e-01 -8.44970584e-01
3.91874135e-01 1.03966844e+00 1.05349094e-01 -1.01524150e+00
6.86291039e-01 -1.23764336e-01 -8.63621593e-01 -1.37271261e+00
-5.87829053e-01 -7.87492335e-01 -1.22736134e-01 -2.78499454e-01
4.28573221e-01 -1.85597539e-01 -6.71681881e-01 4.06616956e-01
-1.48854709e+00 -6.79310083e-01 1.47557259e-01 5.33354804e-02
-7.65916482e-02 7.29527652e-01 -9.85339940e-01 -9.82381523e-01
-3.25578123e-01 -7.23249257e-01 5.97634852e-01 2.42714465e-01
-1.09232044e+00 -8.85002971e-01 3.15613180e-01 6.24955595e-01
7.96286762e-01 -1.89602703e-01 1.57280672e+00 -1.13959742e+00
-3.78484339e-01 -2.21968472e-01 -1.15140922e-01 7.90800601e-02
1.72215596e-01 4.41406578e-01 -6.20584607e-01 -1.57689899e-01
-3.21535505e-02 -3.23066533e-01 -1.03394426e-01 5.95593974e-02
4.23577309e-01 -1.07518601e+00 -9.60971490e-02 1.95508137e-01
9.82933760e-01 4.73129243e-01 4.48392987e-01 5.71710110e-01
4.07259256e-01 6.90816462e-01 5.03215849e-01 6.33966565e-01
1.90001070e-01 4.22560513e-01 1.34594589e-01 4.13009346e-01
7.51188546e-02 -7.37331092e-01 2.97001511e-01 1.02684879e+00
5.55367172e-01 -6.02182984e-01 -1.21466100e+00 6.02509201e-01
-1.63047600e+00 -6.48874104e-01 -4.32296246e-01 1.68290615e+00
7.45048404e-01 1.43119723e-01 8.49303901e-02 -2.58547813e-01
3.60843986e-01 5.51975071e-01 -1.16029475e-03 -1.00278115e+00
9.31104645e-02 5.02943769e-02 3.29990238e-01 7.69413173e-01
-5.24081230e-01 8.69436741e-01 7.66791344e+00 7.03577459e-01
-9.88544643e-01 2.13752985e-01 2.66091257e-01 -3.73502448e-02
-5.95619261e-01 3.36083263e-01 -1.07774949e+00 7.87083268e-01
1.27632558e+00 -2.97652274e-01 4.27276403e-01 8.10767293e-01
5.11655688e-01 -1.57350034e-01 -7.74186730e-01 3.71840566e-01
4.05599810e-02 -1.56242836e+00 1.50147900e-01 -1.51563570e-01
5.16121149e-01 1.14344157e-01 -1.30712569e-01 6.43510580e-01
5.09168029e-01 -8.24685574e-01 6.09867752e-01 7.06864372e-02
5.87096930e-01 -5.09636164e-01 2.16969818e-01 9.63079572e-01
-5.71858227e-01 -2.29886845e-01 -1.63734496e-01 -4.08968866e-01
1.81326658e-01 2.44170174e-01 -1.24900651e+00 1.90315276e-01
6.69690132e-01 1.56787783e-01 -9.12324369e-01 1.03617167e+00
-2.72433281e-01 1.17233205e+00 -5.08385420e-01 -3.57430011e-01
3.94153804e-01 2.81832337e-01 7.10099280e-01 1.88656354e+00
-3.05855889e-02 -5.32773621e-02 2.07018763e-01 7.35254288e-01
2.58358419e-01 8.16724077e-02 -5.51797032e-01 -4.10586596e-01
1.18857241e+00 1.05205548e+00 -4.16740716e-01 -1.12352863e-01
-3.15350026e-01 4.31825548e-01 1.00489803e-01 6.46354795e-01
-5.95965564e-01 -3.99904341e-01 7.02982664e-01 4.00485456e-01
-3.94357145e-01 -4.80633587e-01 -4.09847945e-01 -9.12070215e-01
-5.09772822e-03 -1.25253892e+00 4.05744821e-01 -8.54669273e-01
-1.17209220e+00 3.92549187e-01 3.46643120e-01 -4.49024320e-01
-3.92114609e-01 -1.37978896e-01 -1.13861859e+00 8.50147009e-01
-1.01771176e+00 -9.74603355e-01 1.05727892e-02 4.52756315e-01
6.24753773e-01 4.06173259e-01 7.26078987e-01 5.48535585e-01
-4.47989821e-01 7.59885073e-01 -6.52651722e-03 -9.17035341e-02
9.52539325e-01 -8.78575802e-01 7.05335021e-01 9.38412249e-01
-3.82807106e-02 1.42856336e+00 8.36706161e-01 -1.00157273e+00
-1.15072405e+00 -8.21347415e-01 1.61231565e+00 -1.06974757e+00
8.58395755e-01 -7.54187286e-01 -1.12442780e+00 9.29173291e-01
3.21530193e-01 -8.43158782e-01 7.78213322e-01 3.21413130e-01
-5.31918943e-01 5.86157262e-01 -9.19835806e-01 6.65357411e-01
7.20874369e-01 -9.81729686e-01 -5.66644549e-01 5.20740330e-01
8.75607133e-01 -2.93836027e-01 -5.47974110e-01 -2.63309985e-01
6.78515911e-01 -8.44787598e-01 6.39439523e-01 -1.06291187e+00
3.47264528e-01 3.37320507e-01 -8.91797617e-02 -1.06317353e+00
-1.94513083e-01 -1.30728066e+00 7.83753470e-02 1.73353946e+00
4.94161010e-01 -7.31486082e-01 3.08425963e-01 1.15265119e+00
-2.10238189e-01 -4.47270155e-01 -9.48557436e-01 -8.77927005e-01
3.30846399e-01 -6.85571194e-01 4.49135184e-01 1.23837519e+00
2.11235151e-01 2.23744333e-01 -6.94552600e-01 -1.24590047e-01
1.30726248e-01 -2.91282296e-01 9.05220926e-01 -8.42138171e-01
-3.81935209e-01 -2.24438250e-01 5.08569837e-01 -1.25630319e+00
-1.23017676e-01 -6.03896141e-01 -1.00375049e-01 -1.54711032e+00
7.66907409e-02 -7.65708625e-01 2.52870321e-01 7.54334807e-01
-1.78720728e-01 -5.35084248e-01 5.61169207e-01 4.44021940e-01
-5.55566490e-01 -1.79552376e-01 7.18311310e-01 1.09408461e-01
-5.02124548e-01 -3.02741025e-02 -9.20371830e-01 6.48993671e-01
9.09191012e-01 -7.31072545e-01 -4.05418426e-01 -6.11461937e-01
5.56632102e-01 2.01196894e-01 2.68588126e-01 -4.73877549e-01
4.20586318e-01 -7.12448180e-01 -4.32408035e-01 -5.09943962e-01
-2.34889597e-01 -3.53135109e-01 -1.29218325e-01 2.41557553e-01
-7.65661597e-01 3.78585279e-01 5.99786758e-01 1.66762501e-01
3.88725966e-01 -5.82966983e-01 3.47138792e-01 5.17009124e-02
-5.29692531e-01 -3.11560214e-01 -1.33941805e+00 6.65416598e-01
5.14668941e-01 9.11326483e-02 -1.02854562e+00 -5.20994067e-01
5.09808585e-02 3.32212031e-01 2.84495890e-01 9.46006954e-01
3.18982601e-01 -6.86317086e-01 -5.28419971e-01 1.48964331e-01
8.06817189e-02 -6.78988457e-01 3.00933182e-01 7.14921713e-01
-2.73632824e-01 7.73348331e-01 7.81759098e-02 -3.70210081e-01
-1.25344574e+00 1.61097944e-01 1.96152598e-01 -4.19855505e-01
-5.11107326e-01 7.11799860e-01 -5.43522537e-02 -7.40604162e-01
2.28540182e-01 -2.70627916e-01 1.78555757e-01 -1.89149737e-01
6.84376299e-01 6.16429269e-01 1.41475588e-01 -1.51181668e-01
-5.12102842e-01 -4.16011922e-02 -5.55526137e-01 -9.41058993e-02
9.37472403e-01 -1.89887196e-01 -1.23175591e-01 2.30349019e-01
9.87672985e-01 5.93343258e-01 -7.60282874e-01 -6.86837547e-03
5.33975244e-01 -7.20085502e-01 -6.57173812e-01 -1.33354890e+00
-1.27824455e-01 4.83835042e-01 -1.79106489e-01 6.37654841e-01
4.61620450e-01 2.57798638e-02 1.09544408e+00 5.36288559e-01
3.14052194e-01 -1.05482221e+00 1.63261428e-01 8.82615626e-01
8.97204757e-01 -6.55444920e-01 -4.67080116e-01 -1.44800648e-01
-3.98009241e-01 9.48570549e-01 7.84300089e-01 2.54928321e-01
4.10643667e-01 5.52037060e-01 5.01066625e-01 -2.79272258e-01
-1.07434404e+00 6.54316485e-01 -2.51866311e-01 5.55984557e-01
5.22486806e-01 -2.48088464e-02 -6.65202200e-01 6.42330229e-01
-2.02035293e-01 -1.12740844e-01 8.89860749e-01 1.28496182e+00
-4.85493690e-01 -1.19621003e+00 -5.42009532e-01 6.38384044e-01
-6.63801849e-01 -4.27877367e-01 -1.07351100e+00 8.67093325e-01
-1.54040828e-01 1.76846409e+00 -3.57402563e-01 -6.07585907e-01
3.91612023e-01 3.02273154e-01 -1.27854273e-01 -9.86441553e-01
-1.36039340e+00 -1.32621258e-01 5.96486986e-01 -4.20524359e-01
3.59720826e-01 -4.21778947e-01 -8.02946746e-01 -8.58108997e-01
-4.82087672e-01 5.31391859e-01 3.44766587e-01 1.03868461e+00
5.83104908e-01 -2.54858658e-02 3.31938684e-01 -3.07188220e-02
-7.44910479e-01 -1.15770340e+00 -1.51172489e-01 2.47173563e-01
2.30948582e-01 -1.25108510e-01 -5.73976696e-01 -3.32406431e-01] | [11.158808708190918, 8.391949653625488] |
0f3fdee5-0005-46c9-99f5-09379f6cdc20 | pandaset-advanced-sensor-suite-dataset-for | 2112.12610 | null | https://arxiv.org/abs/2112.12610v1 | https://arxiv.org/pdf/2112.12610v1.pdf | PandaSet: Advanced Sensor Suite Dataset for Autonomous Driving | The accelerating development of autonomous driving technology has placed greater demands on obtaining large amounts of high-quality data. Representative, labeled, real world data serves as the fuel for training deep learning networks, critical for improving self-driving perception algorithms. In this paper, we introduce PandaSet, the first dataset produced by a complete, high-precision autonomous vehicle sensor kit with a no-cost commercial license. The dataset was collected using one 360{\deg} mechanical spinning LiDAR, one forward-facing, long-range LiDAR, and 6 cameras. The dataset contains more than 100 scenes, each of which is 8 seconds long, and provides 28 types of labels for object classification and 37 types of labels for semantic segmentation. We provide baselines for LiDAR-only 3D object detection, LiDAR-camera fusion 3D object detection and LiDAR point cloud segmentation. For more details about PandaSet and the development kit, see https://scale.com/open-datasets/pandaset. | ['Diange Yang', 'Yunlong Wang', 'Kun Jiang', 'Kai Sun', 'Jian Wu', 'Zesong Li', 'Judy Jiao', 'Xiaolin Chai', 'Zishuo Zhang', 'Steven Hao', 'Zhenlei Shao', 'Pengchuan Xiao'] | 2021-12-23 | null | null | null | null | ['point-cloud-segmentation'] | ['computer-vision'] | [-2.07556188e-02 -3.20151567e-01 -3.86469901e-01 -1.08063316e+00
-9.15324032e-01 -8.79906654e-01 3.79446685e-01 8.45851377e-03
-6.05339348e-01 2.94410378e-01 -6.27168536e-01 -5.15552402e-01
1.86267152e-01 -7.80468524e-01 -1.09498775e+00 -2.99300551e-01
1.87072277e-01 7.72876441e-01 5.12287378e-01 -1.74127426e-02
2.05346629e-01 1.01918459e+00 -2.06769776e+00 -1.91808380e-02
6.31128550e-01 1.36557543e+00 4.86429125e-01 6.66642845e-01
-2.14024752e-01 3.32682520e-01 -1.47792548e-01 -1.79309592e-01
7.39460826e-01 4.01607841e-01 -2.65696794e-01 -2.87346225e-02
1.22086906e+00 -6.04764640e-01 -2.26483762e-01 8.48948061e-01
3.64666075e-01 -1.64397910e-01 5.45464754e-01 -1.94604194e+00
-2.90931672e-01 1.99750811e-01 -6.55856431e-01 1.36382133e-01
-2.31292784e-01 4.06470567e-01 8.11889052e-01 -9.95579839e-01
4.88786966e-01 1.22069609e+00 9.04235065e-01 3.42550248e-01
-1.03323734e+00 -1.24104381e+00 -3.56152743e-01 1.25802055e-01
-1.46076536e+00 -6.35442913e-01 7.14875042e-01 -8.50098431e-01
9.18103874e-01 -2.00750515e-01 6.36916280e-01 9.35095310e-01
2.38006175e-01 4.54027772e-01 1.00311327e+00 1.50933653e-01
1.81241766e-01 1.08860679e-01 4.30747658e-01 6.27384663e-01
5.81702173e-01 4.13728565e-01 -4.18636709e-01 2.44024634e-01
2.18302637e-01 4.43952419e-02 5.89304924e-01 -7.27133214e-01
-8.85006249e-01 8.45430136e-01 4.71914142e-01 -5.35635114e-01
4.97529432e-02 3.63012433e-01 2.93567896e-01 2.87066549e-02
3.29310238e-01 6.28836155e-02 -6.65672362e-01 -1.60999045e-01
-6.66751206e-01 5.14294803e-01 2.99496472e-01 1.46055508e+00
1.39043677e+00 1.05773970e-01 6.66996002e-01 6.76387489e-01
3.96538794e-01 1.17130232e+00 -7.31170699e-02 -1.62319911e+00
3.81213784e-01 6.93678677e-01 1.05121724e-01 -5.72484374e-01
-5.87607443e-01 -3.34628820e-02 -1.11421138e-01 5.73605061e-01
1.11574993e-01 -1.03405349e-01 -1.08339071e+00 1.22134364e+00
4.14839357e-01 -1.27019146e-02 -6.02641590e-02 8.91982555e-01
1.13418758e+00 4.67054427e-01 -1.43068850e-01 5.30119181e-01
1.22531903e+00 -7.18465805e-01 -2.00505570e-01 -6.53178334e-01
6.55668795e-01 -7.20627964e-01 1.02044487e+00 2.51033008e-01
-5.61627626e-01 -9.58159745e-01 -1.31362152e+00 -6.34292364e-01
-8.34396064e-01 1.42375201e-01 5.95232785e-01 5.48714697e-01
-9.61265147e-01 1.20342590e-01 -7.88179338e-01 -4.51825351e-01
8.24799478e-01 1.87693968e-01 -3.14381421e-01 -3.75063062e-01
-7.48073816e-01 9.54850852e-01 4.09488171e-01 -2.09593505e-01
-1.04856026e+00 -6.99861109e-01 -9.17648673e-01 -5.90549052e-01
4.75214064e-01 -2.36476645e-01 1.61008465e+00 -2.17020899e-01
-9.51903403e-01 1.45791829e+00 -9.94118080e-02 -6.65515184e-01
4.06478584e-01 -3.68708611e-01 -3.39363754e-01 -2.84001119e-02
5.52184761e-01 1.62570405e+00 7.04057097e-01 -1.45306790e+00
-1.20214558e+00 -6.78527057e-01 -3.26407671e-01 -5.47006391e-02
3.76015037e-01 -2.08285928e-01 -2.46302024e-01 3.94760013e-01
6.92315325e-02 -1.19519281e+00 -5.71204051e-02 3.32610279e-01
-3.16714376e-01 -2.98484147e-01 1.49322939e+00 4.90602572e-03
7.87091032e-02 -2.43360043e+00 -7.83018053e-01 -3.14373285e-01
2.31577978e-01 2.03231052e-01 -1.71850562e-01 1.40155047e-01
1.22913465e-01 2.87028681e-02 -2.85897225e-01 -5.38598716e-01
1.42366841e-01 6.10948324e-01 -5.31502008e-01 4.96703655e-01
2.99802274e-01 9.55487609e-01 -7.65942335e-01 -5.23616254e-01
6.98593676e-01 2.58469582e-01 -1.27672970e-01 -4.87547852e-02
-3.05942178e-01 2.61503369e-01 -2.60286659e-01 8.87076974e-01
1.18023741e+00 2.30999425e-01 -6.44624412e-01 -7.53750280e-02
-6.08380795e-01 2.20778048e-01 -8.97823513e-01 1.63505697e+00
-1.64823771e-01 1.17268169e+00 3.20235819e-01 -6.03544533e-01
1.24954987e+00 -5.40307045e-01 6.32599294e-01 -8.53969216e-01
2.99595475e-01 4.83373314e-01 -3.40813249e-01 -5.23719907e-01
8.82596195e-01 1.46868691e-01 -2.08282098e-01 1.26602545e-01
-1.57055065e-01 -9.41633403e-01 3.21703225e-01 1.58911154e-01
7.12646484e-01 2.43281618e-01 -2.55634189e-01 -2.05390975e-01
-3.28012295e-02 8.26312184e-01 6.15424693e-01 6.09279394e-01
-5.85031688e-01 5.05792320e-01 5.32272272e-02 -4.45879310e-01
-1.33474565e+00 -1.24701011e+00 -6.30198896e-01 9.08480465e-01
4.34414983e-01 -2.93349233e-02 -5.36957443e-01 -7.32057095e-02
8.03014338e-01 7.36562729e-01 -8.50989297e-02 4.52225916e-02
-4.62423652e-01 -1.78064361e-01 6.47924483e-01 7.66261697e-01
6.60868526e-01 -7.60069191e-01 -1.07928479e+00 -1.04386233e-01
1.30771205e-01 -1.69382823e+00 1.37233241e-02 4.66415018e-01
-7.58683085e-01 -1.18457532e+00 2.56335318e-01 -7.39413440e-01
1.27333581e-01 9.20840204e-01 1.14161897e+00 -4.33392584e-01
-5.23659706e-01 1.71410307e-01 -1.05969422e-01 -1.10376942e+00
-1.49138615e-01 5.65404482e-02 2.77712286e-01 -4.79676306e-01
1.07983899e+00 -3.04981738e-01 -2.64337420e-01 3.84749621e-01
-5.08817434e-01 8.03459063e-02 5.04781604e-01 -4.19603027e-02
8.55697095e-01 9.49005783e-03 3.22964162e-01 -3.86385441e-01
4.18082178e-02 -3.67696971e-01 -1.16791725e+00 -7.61410058e-01
-6.06411576e-01 -4.43674564e-01 1.70856267e-01 1.55402109e-01
-6.34644032e-01 5.28671741e-01 -1.63069174e-01 -5.89577675e-01
-9.02267992e-01 2.40637525e-03 -1.26172532e-03 -6.21311404e-02
7.92854548e-01 -2.04955146e-01 1.81757435e-01 -4.13408995e-01
7.15884387e-01 1.03179467e+00 9.32540119e-01 -4.47438210e-01
9.29397881e-01 7.95158744e-01 1.15773603e-01 -9.61779416e-01
-9.15234029e-01 -7.35859454e-01 -1.06497359e+00 -2.44112954e-01
1.10991430e+00 -1.31142557e+00 -6.06012881e-01 8.57751548e-01
-9.62107778e-01 -5.63847840e-01 -3.97978336e-01 4.43535805e-01
-4.86005485e-01 -1.97636768e-01 -2.15057418e-01 -7.63473988e-01
-3.32519896e-02 -1.35858786e+00 1.40439618e+00 3.63293022e-01
2.36276314e-01 -1.91363260e-01 -1.53098360e-01 8.12058330e-01
1.31113425e-01 3.10026467e-01 3.83337975e-01 -3.73792887e-01
-1.00727630e+00 -3.46690804e-01 -6.50758862e-01 5.71766198e-01
-1.45469829e-01 4.19895917e-01 -1.26561296e+00 -6.26730472e-02
-2.54147112e-01 -6.73099518e-01 9.91075337e-01 3.14775646e-01
1.09440577e+00 5.30191958e-01 -5.23543298e-01 7.88484275e-01
1.27724624e+00 3.05177927e-01 8.45314637e-02 2.90414095e-01
1.03151131e+00 6.06015265e-01 8.59884143e-01 8.50594565e-02
9.06900942e-01 4.09718513e-01 7.58444428e-01 8.13145284e-03
-1.29706040e-01 -2.10901916e-01 2.67396688e-01 3.70478779e-01
4.56625342e-01 4.83929925e-03 -1.35582685e+00 6.40085161e-01
-1.37921262e+00 -6.43621087e-01 -7.47402906e-01 1.87811565e+00
3.46864313e-01 3.88802916e-01 1.69452652e-01 -1.92894917e-02
6.44181311e-01 1.15222663e-01 -1.16092265e+00 -4.19682026e-01
-7.11113960e-02 -1.27368808e-01 1.11012828e+00 3.87043506e-01
-1.26946366e+00 1.06121397e+00 6.11936712e+00 5.39224446e-01
-1.28164184e+00 5.43172769e-02 2.66889453e-01 -1.49867237e-01
1.13024697e-01 -1.27403378e-01 -1.37801754e+00 3.14601243e-01
1.26006019e+00 -2.22346745e-02 7.63099045e-02 1.37058544e+00
3.56954932e-01 -4.15078580e-01 -9.77541029e-01 9.78956759e-01
-1.63163692e-01 -1.24662459e+00 -4.86664951e-01 3.53028864e-01
4.91254270e-01 1.25229049e+00 6.63055331e-02 2.73097485e-01
5.69419682e-01 -6.60366833e-01 1.08659530e+00 1.42986923e-01
1.12336767e+00 -7.67978013e-01 5.33234715e-01 5.88147163e-01
-1.33731675e+00 -2.40340427e-01 -5.36935508e-01 -1.69923469e-01
1.12011187e-01 6.05620027e-01 -8.73392880e-01 1.70975864e-01
1.10320973e+00 9.79591012e-01 -7.68693209e-01 8.16530168e-01
-5.67455515e-02 4.48005587e-01 -7.24329412e-01 2.07028896e-01
3.69782865e-01 -4.21125531e-01 4.34469432e-01 1.01338255e+00
2.47740299e-01 -1.65601835e-01 5.10535538e-01 8.82734835e-01
-1.59492120e-01 -3.96829784e-01 -1.15282798e+00 4.34342809e-02
9.00803626e-01 1.58116949e+00 -7.07206905e-01 -3.27493817e-01
-3.83089215e-01 9.22711343e-02 -4.94880527e-02 2.96086585e-03
-8.54322314e-01 -3.44039828e-01 1.11187649e+00 4.30131227e-01
1.95165679e-01 -8.49086702e-01 -7.37574637e-01 -5.71395814e-01
3.94533621e-03 -3.01662892e-01 -9.05401483e-02 -1.23257685e+00
-1.24472094e+00 1.72873616e-01 2.97720462e-01 -1.18922377e+00
-1.06130630e-01 -9.59549308e-01 -1.63853779e-01 7.48514056e-01
-1.89859509e+00 -1.29009008e+00 -9.79950428e-01 3.61451119e-01
6.50957882e-01 -1.25151739e-01 3.04041237e-01 4.79439586e-01
-4.38894421e-01 -5.48752863e-03 -2.31201630e-02 1.52041376e-01
7.61363924e-01 -1.01257610e+00 7.43217409e-01 6.78605258e-01
-1.58170968e-01 1.80267859e-02 5.50620139e-01 -6.42599583e-01
-1.62228441e+00 -1.65382314e+00 4.73501563e-01 -9.52332437e-01
7.07634628e-01 -6.78303123e-01 -7.22119331e-01 8.36512446e-01
-2.59413272e-02 1.75279707e-01 3.94034773e-01 -2.67202795e-01
-3.61530811e-01 -6.87047005e-01 -1.33542907e+00 9.22283158e-02
1.27966809e+00 -4.36755747e-01 -4.31430697e-01 3.45414191e-01
9.66169059e-01 -6.93672121e-01 -5.02127588e-01 6.78764403e-01
5.37414610e-01 -1.00620711e+00 9.09052610e-01 3.85467522e-02
3.03989768e-01 -5.00211060e-01 -5.26387811e-01 -7.96658099e-01
-1.81688756e-01 1.97512489e-02 2.34334856e-01 9.85195637e-01
3.30324948e-01 -8.50196779e-01 1.05893326e+00 3.83534908e-01
-1.00471568e+00 -3.35121483e-01 -9.54605281e-01 -9.35161293e-01
2.34786928e-01 -1.06974721e+00 6.98706508e-01 7.59484768e-01
-9.25773144e-01 4.23779190e-01 2.10192308e-01 2.12554246e-01
1.01670194e+00 2.29703799e-01 1.29617441e+00 -1.69315481e+00
6.35817587e-01 -4.51404899e-01 -5.79079151e-01 -1.00960195e+00
1.64445713e-01 -8.90236795e-01 3.50949943e-01 -1.62720239e+00
-1.58331841e-01 -8.02599847e-01 3.10327679e-01 6.09180152e-01
4.58160669e-01 5.76105773e-01 1.10961264e-02 3.12279761e-01
-3.41061026e-01 2.99722552e-01 8.28698814e-01 -1.68924093e-01
1.59267172e-01 -1.76406667e-01 -4.30788577e-01 8.30405295e-01
1.15099108e+00 -5.95008910e-01 -4.13421780e-01 -7.60740459e-01
8.31775665e-02 -5.18798172e-01 5.73548019e-01 -1.42362380e+00
2.17981562e-01 -2.80676901e-01 3.57306719e-01 -1.61962259e+00
7.60871232e-01 -9.03998077e-01 -5.43530956e-02 2.90703267e-01
6.90991580e-02 9.45590585e-02 4.80697185e-01 2.15580314e-01
-2.24825703e-02 3.78060788e-02 1.13351715e+00 -1.20520614e-01
-1.41371238e+00 5.89882612e-01 -3.34933996e-01 1.40244775e-02
1.27019358e+00 -5.82048833e-01 -6.41491473e-01 1.50142446e-01
-1.03496611e-01 6.34829700e-01 7.90904284e-01 8.89389575e-01
5.46038389e-01 -1.31787109e+00 -6.72295094e-01 4.83616084e-01
5.49384952e-01 7.33367145e-01 6.76202178e-02 3.95504624e-01
-7.50020802e-01 5.69079936e-01 -3.36491913e-01 -1.31134307e+00
-1.11206102e+00 4.63411629e-01 3.25475574e-01 8.72766018e-01
-7.03580320e-01 8.23959410e-01 -2.47337833e-01 -8.67778361e-01
-5.05960174e-02 -4.47056621e-01 1.73650295e-01 1.64901480e-01
2.00868621e-01 4.70191836e-01 1.88909486e-01 -1.00467896e+00
-4.89197999e-01 7.46324897e-01 2.21643895e-01 -5.12438230e-02
1.08875728e+00 -3.35762054e-01 3.94236557e-02 8.57346773e-01
1.17186105e+00 -3.33174378e-01 -1.50166678e+00 -8.47809464e-02
-1.67444453e-01 -4.91514176e-01 2.00625390e-01 -4.35818940e-01
-1.03588378e+00 1.27398467e+00 8.96385968e-01 1.87322855e-01
5.90864897e-01 2.93816328e-01 9.86471474e-01 6.40301406e-01
7.16713071e-01 -1.16093361e+00 -1.82218045e-01 8.56031656e-01
5.02268255e-01 -1.73949647e+00 -1.00016303e-01 -2.82989889e-01
-4.60529625e-01 7.87648082e-01 1.05978525e+00 -1.10982202e-01
5.57951748e-01 6.01850450e-01 6.06661201e-01 -4.50550586e-01
-4.57880378e-01 -3.92348260e-01 -1.12834662e-01 8.70437324e-01
-7.11368322e-02 3.77757221e-01 4.46618974e-01 -3.02021820e-02
-6.87107503e-01 -4.77490984e-02 5.71060002e-01 1.01268768e+00
-1.04678524e+00 -7.17588663e-01 -3.83448809e-01 5.56959212e-01
2.82279462e-01 2.59523273e-01 -3.29293311e-01 7.99855411e-01
7.58497000e-01 1.10807252e+00 5.16661108e-01 -6.19570553e-01
4.55333114e-01 7.76213780e-02 1.52542919e-01 -5.97554386e-01
1.94842935e-01 -2.58604467e-01 1.11034915e-01 -7.56208599e-01
-1.48733854e-01 -9.64005351e-01 -1.53600252e+00 -4.81367499e-01
3.54329683e-02 -3.42976391e-01 1.50528085e+00 6.34301722e-01
6.80554271e-01 3.21258307e-02 7.28050351e-01 -1.28395820e+00
-2.86224514e-01 -8.36821973e-01 -5.23695469e-01 7.00453017e-03
4.45107669e-01 -9.88297820e-01 -3.90943468e-01 1.01289473e-01] | [7.828461647033691, -2.368272066116333] |
b9d7c0c7-fa00-4ea7-80dc-4b3684b65d81 | transfer-learning-for-material-classification | 1609.06188 | null | http://arxiv.org/abs/1609.06188v1 | http://arxiv.org/pdf/1609.06188v1.pdf | Transfer Learning for Material Classification using Convolutional Networks | Material classification in natural settings is a challenge due to complex
interplay of geometry, reflectance properties, and illumination. Previous work
on material classification relies strongly on hand-engineered features of
visual samples. In this work we use a Convolutional Neural Network (convnet)
that learns descriptive features for the specific task of material recognition.
Specifically, transfer learning from the task of object recognition is
exploited to more effectively train good features for material classification.
The approach of transfer learning using convnets yields significantly higher
recognition rates when compared to previous state-of-the-art approaches. We
then analyze the relative contribution of reflectance and shading information
by a decomposition of the image into its intrinsic components. The use of
convnets for material classification was hindered by the strong demand for
sufficient and diverse training data, even with transfer learning approaches.
Therefore, we present a new data set containing approximately 10k images
divided into 10 material categories. | ['Hendrik P. A. Lensch', 'Patrick Wieschollek'] | 2016-09-20 | null | null | null | null | ['material-classification', 'material-recognition'] | ['computer-vision', 'computer-vision'] | [ 5.50848007e-01 -3.56997728e-01 1.86068892e-01 -3.57000560e-01
-4.56313312e-01 -4.38014388e-01 7.62527466e-01 8.54668915e-02
-1.53230906e-01 7.23475814e-01 -2.48417482e-02 6.46995986e-03
-2.19942406e-01 -1.17814779e+00 -8.41660798e-01 -8.58593822e-01
8.11852589e-02 1.95582047e-01 2.81164646e-01 -3.06811601e-01
3.43057156e-01 9.47381258e-01 -1.76970363e+00 6.38566613e-01
5.12537658e-01 1.54578912e+00 3.13993365e-01 7.49907494e-01
-2.96400487e-01 6.11154318e-01 -4.23043489e-01 1.15902394e-01
3.79100472e-01 -1.60562202e-01 -8.59779298e-01 2.74715334e-01
7.84169316e-01 -2.77910978e-01 -3.15438628e-01 5.99817216e-01
4.43943650e-01 1.53434217e-01 1.23459828e+00 -7.99885571e-01
-7.56188095e-01 9.49289724e-02 -1.48354113e-01 -5.37224673e-02
2.87697643e-01 1.85154155e-01 1.00148141e+00 -9.22600985e-01
5.94814062e-01 7.88701713e-01 7.25671411e-01 5.02662599e-01
-1.38502991e+00 -1.85605794e-01 -5.31019047e-02 2.01505914e-01
-1.07856917e+00 -2.19532952e-01 1.25052965e+00 -8.12384367e-01
1.01170516e+00 1.98840290e-01 9.22047257e-01 1.05757093e+00
-8.63832757e-02 6.22427821e-01 1.19926584e+00 -5.79470456e-01
2.29846567e-01 2.66099840e-01 5.06106429e-02 6.78959966e-01
1.92089945e-01 8.69381055e-02 -5.88007689e-01 9.07540619e-02
8.48991156e-01 1.94608737e-02 -2.39814654e-01 -5.08118451e-01
-1.00576735e+00 5.43578565e-01 7.26306498e-01 3.34763229e-01
-3.93069386e-01 1.16733342e-01 2.31145769e-01 3.94192159e-01
4.85652924e-01 5.00737667e-01 -7.15640783e-01 7.97933787e-02
-7.55458474e-01 9.30649191e-02 8.44549716e-01 6.10166490e-01
1.02319002e+00 3.61439526e-01 -8.07973072e-02 1.07235873e+00
1.38269499e-01 6.50177479e-01 -3.62617821e-02 -6.41470194e-01
2.20819741e-01 7.74579108e-01 -1.68561593e-01 -9.76341844e-01
-3.44741404e-01 -3.18378955e-01 -6.91513658e-01 4.74715114e-01
5.67884088e-01 2.90951192e-01 -8.76916349e-01 1.25284326e+00
1.82891339e-01 -5.40032864e-01 1.25273496e-01 6.34541452e-01
1.11811471e+00 4.78501141e-01 -1.91965118e-01 4.09235597e-01
1.01236963e+00 -7.56252706e-01 3.48694995e-03 7.34985396e-02
2.80497730e-01 -8.36663544e-01 9.78015721e-01 4.70708877e-01
-8.83435726e-01 -7.07412064e-01 -1.02310181e+00 4.36111018e-02
-8.57740462e-01 4.39927429e-01 6.45581722e-01 5.35312891e-01
-7.64130712e-01 8.84380579e-01 -5.69533288e-01 -4.49749529e-01
8.29695642e-01 5.93873322e-01 -3.21026623e-01 -3.58449399e-01
-5.79619944e-01 6.45279408e-01 2.98629671e-01 3.88220511e-02
-7.78843343e-01 -7.99646258e-01 -6.86599910e-01 -2.15828955e-01
1.26296997e-01 -6.09100878e-01 7.51074195e-01 -1.26214612e+00
-1.68064988e+00 1.11921299e+00 4.97862279e-01 1.54380083e-01
5.93641937e-01 -5.67836426e-02 -1.21036611e-01 2.82219559e-01
-4.04330224e-01 4.15927559e-01 1.16264009e+00 -1.65639710e+00
-1.66482478e-01 -3.53528932e-03 8.89584646e-02 -3.65815639e-01
-5.54054499e-01 -1.19126588e-01 -5.88128455e-02 -5.46756983e-01
3.89596671e-02 -8.17270577e-01 1.23205006e-01 2.42539674e-01
-3.27879965e-01 -1.18235938e-01 7.74024963e-01 -4.99568969e-01
3.07006508e-01 -2.10896564e+00 7.05377832e-02 2.19290584e-01
1.74905360e-01 3.41628551e-01 -2.75840640e-01 6.12911940e-01
-6.33456632e-02 -2.77262449e-01 -8.72543454e-02 -8.69952440e-02
-3.87518406e-02 -8.21860880e-02 2.39169169e-02 5.56092441e-01
6.82888091e-01 9.59874988e-01 -6.30169868e-01 -3.00407708e-01
5.48664689e-01 7.03531623e-01 -4.83908474e-01 3.89346987e-01
-2.68147528e-01 5.54404140e-01 -3.21035266e-01 8.98987651e-01
6.38853967e-01 -1.86568946e-01 -1.51765332e-01 -7.15157151e-01
-5.37760742e-02 1.95153624e-01 -6.90711975e-01 1.58564746e+00
-6.70909464e-01 9.65041518e-01 1.56057224e-01 -1.18419611e+00
1.15695882e+00 5.70866577e-02 8.21728170e-01 -6.63248539e-01
4.31930006e-01 4.00053322e-01 4.02423330e-02 -4.75289017e-01
2.83568650e-01 -3.41439284e-02 4.33532387e-01 4.20893341e-01
1.55465096e-01 -3.39799076e-01 -1.26287401e-01 -2.77568191e-01
1.02912068e+00 3.60866368e-01 -2.11584792e-01 -5.28547585e-01
5.59047520e-01 -7.88620710e-02 -2.77339332e-02 5.49314380e-01
1.63634941e-01 6.77856982e-01 1.41972862e-02 -9.46286023e-01
-1.10402691e+00 -1.15880632e+00 -2.64855146e-01 1.06816053e+00
-1.65743217e-01 -1.82509527e-01 -6.18261456e-01 -4.40099239e-01
2.11318374e-01 -9.16817039e-02 -8.86438489e-01 -3.04604352e-01
-7.35925138e-01 -6.89830720e-01 7.66330138e-02 8.33066285e-01
6.00855172e-01 -1.41403925e+00 -6.58669174e-01 5.98809458e-02
3.27973872e-01 -1.20525444e+00 2.65332550e-01 3.98780227e-01
-6.05703294e-01 -1.12331021e+00 -7.97232151e-01 -8.77930284e-01
8.84994686e-01 2.11717561e-01 1.37736881e+00 2.61808574e-01
-9.79216695e-01 7.76502132e-01 -3.53621989e-01 -4.58199233e-01
-4.74243551e-01 3.34746897e-01 -2.97652692e-01 2.20522240e-01
1.34986460e-01 -7.28061259e-01 -6.87604010e-01 1.13561183e-01
-8.62144828e-01 1.34769708e-01 7.95158625e-01 7.84148335e-01
2.65534461e-01 -7.34572951e-03 4.12751079e-01 -6.67496920e-01
3.07157755e-01 -1.52999386e-01 -3.15501630e-01 2.28525847e-01
-2.17723027e-01 1.37314005e-02 6.64080381e-01 -4.77151155e-01
-8.03277314e-01 3.88988927e-02 2.30047386e-02 -2.10503981e-01
-4.83786225e-01 1.62776008e-01 -9.04345587e-02 -5.84520936e-01
6.43278480e-01 8.27507973e-02 -2.20012199e-02 -4.63003635e-01
1.41237825e-01 6.23945117e-01 3.14931750e-01 -1.00165677e+00
8.79778862e-01 3.84173840e-01 4.19814467e-01 -1.37998748e+00
-6.98122203e-01 -2.09873438e-01 -9.76010978e-01 -5.75045586e-01
7.36839354e-01 -5.24563670e-01 -6.69636369e-01 8.26260865e-01
-8.94756734e-01 -7.43816972e-01 -5.35449982e-01 2.29775786e-01
-7.48275816e-01 7.47088194e-02 -6.97365999e-01 -7.09550619e-01
-2.88761318e-01 -1.04398918e+00 1.09787691e+00 -1.82404831e-01
3.40442695e-02 -1.20437884e+00 -9.36100036e-02 3.35930645e-01
7.84313381e-01 4.87645835e-01 1.15531790e+00 -1.37466818e-01
-7.75824070e-01 -2.47737542e-01 -6.02165699e-01 6.61260903e-01
6.14783466e-01 2.65970409e-01 -1.36788833e+00 -2.36063361e-01
-2.91672587e-01 -5.84595501e-01 1.10226452e+00 2.43366748e-01
1.43996155e+00 1.47692963e-01 7.02874810e-02 5.10325611e-01
1.61032581e+00 -1.98000804e-01 5.29261470e-01 5.12948930e-01
1.09436023e+00 9.95316803e-01 6.06618151e-02 2.69532889e-01
1.62013453e-02 5.59490204e-01 4.21268880e-01 -3.91296178e-01
-5.98857880e-01 1.21845379e-01 7.35438168e-02 6.58473372e-01
-5.46335757e-01 -8.11235234e-02 -9.24376249e-01 3.05592984e-01
-1.43305731e+00 -4.77741718e-01 4.53399979e-02 1.91163528e+00
5.30814707e-01 2.14788049e-01 1.98639452e-01 5.48023760e-01
3.24635476e-01 -4.80415188e-02 -4.73174274e-01 -2.21089974e-01
-1.43610820e-01 3.38809639e-01 4.59418833e-01 -3.13820830e-03
-1.13680351e+00 6.24482810e-01 7.03855848e+00 6.33480966e-01
-1.32755613e+00 -5.01152098e-01 4.40474182e-01 2.16691926e-01
-1.84070125e-01 -6.11387372e-01 -4.78947610e-01 1.45674258e-01
5.95415592e-01 6.68753564e-01 6.50419891e-01 6.95514858e-01
-2.35523522e-01 -4.84741479e-02 -1.32632375e+00 1.16727984e+00
2.89662898e-01 -1.17549789e+00 4.80278283e-02 1.56419620e-01
7.76705444e-01 7.14931730e-03 2.65798658e-01 -1.30784974e-01
1.39900953e-01 -1.07003546e+00 7.22497582e-01 6.02709889e-01
7.79164791e-01 -2.46986851e-01 4.79871958e-01 -1.82289377e-01
-1.27746975e+00 -5.75894043e-02 -3.89468551e-01 -1.08333118e-01
-3.67670268e-01 5.98118603e-01 -5.63278615e-01 5.93672037e-01
7.38835812e-01 1.05240071e+00 -6.41658485e-01 7.82500625e-01
2.63951480e-01 5.24055481e-01 -3.80785972e-01 -3.21756564e-02
1.19910374e-01 -2.87665188e-01 1.06961936e-01 1.25823426e+00
3.56151201e-02 -4.45523620e-01 4.54392105e-01 9.29887474e-01
-2.60588378e-01 1.78611964e-01 -6.82067633e-01 -1.61984384e-01
-2.57712036e-01 1.33984089e+00 -1.09087598e+00 -1.57892868e-01
-5.39821267e-01 6.23658895e-01 5.47604620e-01 2.04469919e-01
-2.51527101e-01 -9.43302661e-02 4.48516339e-01 2.96863228e-01
6.85158491e-01 -4.01432753e-01 -2.74219930e-01 -8.63577783e-01
2.40917072e-01 -4.86431211e-01 -1.16950966e-01 -7.74496138e-01
-1.79885793e+00 5.25254250e-01 -8.01196024e-02 -1.13133645e+00
1.42438293e-01 -1.41477370e+00 -5.93984365e-01 7.08591044e-01
-1.79445970e+00 -1.50402379e+00 -7.60885119e-01 4.05283660e-01
3.69744062e-01 -2.98911631e-01 9.76668656e-01 2.75711060e-01
-2.31234297e-01 2.43002281e-01 1.62083030e-01 3.38497579e-01
5.31467378e-01 -1.25380492e+00 1.07631184e-01 1.29405901e-01
-1.11856632e-01 1.06134310e-01 5.24992943e-01 -2.18656674e-01
-1.57221437e+00 -9.56174970e-01 1.29540533e-01 -4.77207512e-01
6.41883612e-01 -5.69247842e-01 -8.03366065e-01 1.81262612e-01
4.51133028e-02 3.58747691e-01 9.13716137e-01 5.13420515e-02
-6.62077904e-01 -3.65886509e-01 -1.00949061e+00 2.32134759e-01
1.02878785e+00 -8.62753034e-01 -3.02103758e-01 3.53532314e-01
2.25275457e-01 2.99657825e-02 -1.05283439e+00 3.33024353e-01
7.78303385e-01 -9.32656169e-01 1.15053988e+00 -5.59760630e-01
7.81667948e-01 1.58370331e-01 -5.78286469e-01 -1.13952875e+00
-3.16505969e-01 -3.47328186e-02 1.81713909e-01 1.24300337e+00
2.50868320e-01 -4.87680465e-01 1.03839839e+00 6.00305021e-01
-1.54441670e-01 -7.57160246e-01 -3.98759544e-01 -7.98854411e-01
2.89422959e-01 -3.23892981e-01 4.46207583e-01 6.26206636e-01
-5.85686684e-01 1.36931822e-01 -1.16260059e-01 -3.71338010e-01
5.48612893e-01 4.47134107e-01 6.39464915e-01 -1.54330993e+00
-2.38785341e-01 -5.93217373e-01 -7.54887640e-01 -5.66417456e-01
2.95000851e-01 -9.08575714e-01 1.56585649e-01 -1.51844108e+00
2.56420821e-01 -6.41738474e-01 -4.17518586e-01 3.10065240e-01
2.58262098e-01 5.50396264e-01 9.65769887e-02 9.99016836e-02
-3.86550397e-01 5.98456919e-01 1.29100752e+00 -6.37923181e-01
-1.05339833e-01 -1.31284162e-01 -2.28295982e-01 5.22754073e-01
9.98205423e-01 -7.13466806e-03 -1.13520376e-01 -4.94295001e-01
2.56361067e-01 -5.27341425e-01 4.77037221e-01 -1.34819913e+00
-1.56140804e-01 -9.90811065e-02 7.15838194e-01 -6.24997795e-01
6.51205420e-01 -1.12932241e+00 5.09515852e-02 3.72789353e-01
-2.88771212e-01 -4.28330660e-01 2.03306198e-01 2.58552253e-01
-6.24172725e-02 -1.86270401e-01 6.25432432e-01 -4.06688809e-01
-6.05317891e-01 2.87768126e-01 -3.97323906e-01 -2.84044385e-01
7.42113352e-01 -4.20703977e-01 -2.19481871e-01 7.67951012e-02
-5.80269277e-01 -5.62220454e-01 7.12191582e-01 1.41793817e-01
6.80167317e-01 -1.34779561e+00 -5.63020051e-01 3.58296216e-01
2.66625673e-01 6.90623969e-02 9.15895328e-02 3.44376087e-01
-7.96426415e-01 2.24459589e-01 -7.15995133e-01 -6.90295994e-01
-1.22283697e+00 4.47492689e-01 4.07117307e-01 -1.79060008e-02
-6.40618443e-01 4.30586457e-01 2.07413211e-01 -4.20331657e-01
2.25363635e-02 -3.03337276e-01 -1.19458757e-01 -9.97513235e-02
2.63582855e-01 3.62026215e-01 4.27550256e-01 -6.47675216e-01
-2.27360308e-01 1.02257884e+00 5.47613613e-02 4.32898045e-01
1.90256310e+00 3.16909164e-01 -3.04737717e-01 5.80337942e-01
1.52676725e+00 -2.11008534e-01 -1.23166013e+00 -3.13512444e-01
-1.47403061e-01 -5.18194437e-01 1.47939667e-01 -6.59162641e-01
-1.33164120e+00 1.00222707e+00 5.73746860e-01 3.92735362e-01
1.05988038e+00 7.51446187e-02 4.96847570e-01 5.56423545e-01
2.83486754e-01 -1.11480796e+00 5.53724945e-01 3.16590428e-01
9.34300840e-01 -1.50217080e+00 2.18710333e-01 -8.17610621e-01
9.46888626e-02 1.56467986e+00 4.21659946e-01 -3.48503113e-01
9.82983232e-01 1.19727634e-01 8.73197019e-02 -5.83621919e-01
-2.67364681e-01 -1.87833130e-01 7.60963857e-01 8.15860331e-01
4.40370947e-01 1.34282053e-01 5.03802598e-01 1.86655045e-01
-1.92899793e-01 -2.59468257e-01 1.22172654e-01 1.23489976e+00
-3.59528393e-01 -1.09375691e+00 -2.82636344e-01 7.14826941e-01
-2.74270684e-01 1.87594727e-01 -5.82397521e-01 5.19984603e-01
5.63063510e-02 6.40151799e-01 1.47155195e-01 -4.51981008e-01
2.55457252e-01 -1.65298924e-01 9.00470734e-01 -5.84733129e-01
-6.08149707e-01 -2.64084339e-01 5.94780929e-02 -3.21213871e-01
-9.75084007e-01 -6.76965356e-01 -5.35575926e-01 -1.89873695e-01
-2.08440647e-01 -4.08152133e-01 9.57836390e-01 8.97508085e-01
-1.85655430e-01 8.50536942e-01 8.17761540e-01 -1.40710163e+00
-1.65412188e-01 -8.63222182e-01 -6.89776540e-01 7.06947625e-01
4.20072883e-01 -8.53289545e-01 -3.37453842e-01 2.55991608e-01] | [10.199088096618652, -0.16858245432376862] |
14571e2c-9c75-40e4-b29d-581e37d72cf2 | why-out-of-distribution-detection-in-cnns | 2110.07043 | null | https://arxiv.org/abs/2110.07043v1 | https://arxiv.org/pdf/2110.07043v1.pdf | Why Out-of-distribution Detection in CNNs Does Not Like Mahalanobis -- and What to Use Instead | Convolutional neural networks applied for real-world classification tasks need to recognize inputs that are far or out-of-distribution (OoD) with respect to the known or training data. To achieve this, many methods estimate class-conditional posterior probabilities and use confidence scores obtained from the posterior distributions. Recent works propose to use multivariate Gaussian distributions as models of posterior distributions at different layers of the CNN (i.e., for low- and upper-level features), which leads to the confidence scores based on the Mahalanobis distance. However, this procedure involves estimating probability density in high dimensional data using the insufficient number of observations (e.g. the dimensionality of features at the last two layers in the ResNet-101 model are 2048 and 1024, with ca. 1000 observations per class used to estimate density). In this work, we want to address this problem. We show that in many OoD studies in high-dimensional data, LOF-based (Local Outlierness-Factor) methods outperform the parametric, Mahalanobis distance-based methods. This motivates us to propose the nonparametric, LOF-based method of generating the confidence scores for CNNs. We performed several feasibility studies involving ResNet-101 and EffcientNet-B3, based on CIFAR-10 and ImageNet (as known data), and CIFAR-100, SVHN, ImageNet2010, Places365, or ImageNet-O (as outliers). We demonstrated that nonparametric LOF-based confidence estimation can improve current Mahalanobis-based SOTA or obtain similar performance in a simpler way. | ['Henryk Maciejewski', 'Tomasz Walkowiak', 'Kamil Szyc'] | 2021-10-13 | null | null | null | null | ['2048'] | ['playing-games'] | [-6.20090902e-01 -9.94482338e-02 2.54212886e-01 -6.42985940e-01
-7.57323027e-01 -2.49015555e-01 4.80491847e-01 6.18039668e-02
-8.98868740e-01 1.00199425e+00 -1.81641072e-01 -2.68748343e-01
-3.25311571e-01 -7.86510646e-01 -9.77152586e-01 -6.32429242e-01
-1.95582956e-01 5.82969368e-01 3.50535780e-01 4.54434574e-01
2.26766393e-01 7.78441668e-01 -1.75289977e+00 1.75581887e-01
8.30211163e-01 1.33285463e+00 -2.23118942e-02 7.91371763e-01
-1.47215098e-01 3.50152701e-01 -1.04665720e+00 -3.27400476e-01
1.97018400e-01 -6.11840561e-02 -2.36838713e-01 -5.49791217e-01
7.12135315e-01 -3.49373221e-01 -1.74887300e-01 1.48495102e+00
4.22176868e-01 8.57149288e-02 1.36072958e+00 -1.50227964e+00
-6.74097240e-01 3.62988949e-01 -2.65771866e-01 3.07364047e-01
-3.51700097e-01 -2.38119647e-01 4.41700459e-01 -1.08574212e+00
9.92836878e-02 1.16490781e+00 8.78647923e-01 3.82816046e-01
-1.17225361e+00 -1.06325221e+00 -1.52439222e-01 2.87263542e-01
-1.81533599e+00 -1.56805053e-01 3.26627254e-01 -6.50539756e-01
9.33941364e-01 -1.39676556e-01 3.10750693e-01 1.27535784e+00
5.20619631e-01 3.18801135e-01 9.84282732e-01 -1.13333121e-01
7.20409036e-01 2.82786727e-01 1.48889512e-01 2.99363196e-01
6.00640357e-01 2.39377603e-01 -2.49720365e-01 -2.92778641e-01
7.98811555e-01 4.78459410e-02 -7.07135946e-02 -7.70128518e-02
-7.21286714e-01 9.15182412e-01 5.61225832e-01 3.17840219e-01
-2.69976199e-01 3.76867652e-01 2.33312815e-01 2.54448086e-01
5.55391252e-01 5.30232042e-02 -6.77123904e-01 -2.15442665e-02
-1.20288551e+00 6.55458719e-02 9.14935827e-01 1.11288261e+00
7.90127218e-01 2.21575141e-01 -1.50180086e-01 7.68064380e-01
6.52045786e-01 4.78783101e-01 7.52384126e-01 -4.87395793e-01
4.49602216e-01 1.52377725e-01 3.08034550e-02 -9.30205643e-01
-2.20904022e-01 -5.09091079e-01 -1.23823059e+00 6.59628987e-01
6.55740499e-01 -1.55263856e-01 -1.39399230e+00 1.64475036e+00
-2.27890059e-01 7.63334155e-01 2.28907138e-01 6.49940491e-01
6.30892813e-01 5.06248534e-01 1.57828201e-02 2.35391349e-01
9.79688406e-01 -4.18704450e-01 -3.68748367e-01 7.52588809e-02
6.44063413e-01 -6.13411248e-01 1.04303205e+00 7.12422073e-01
-5.39523005e-01 -7.56108522e-01 -1.09569478e+00 4.04521286e-01
-6.36759698e-01 4.74099755e-01 1.85852662e-01 9.16676104e-01
-1.05616212e+00 9.70654786e-01 -8.39402139e-01 1.27871307e-02
6.48042500e-01 3.52061123e-01 -5.27879953e-01 -9.17109400e-02
-1.13855827e+00 7.00322390e-01 8.62739146e-01 1.64969012e-01
-9.81129706e-01 -6.78418756e-01 -6.68219864e-01 2.73705870e-01
-2.71658719e-01 -1.39793113e-01 8.76807809e-01 -7.93708384e-01
-1.21737862e+00 3.90983731e-01 1.73017681e-01 -8.07001472e-01
6.11639500e-01 -2.84511447e-01 -4.95989203e-01 -1.07415147e-01
-1.34247869e-01 8.05239856e-01 8.91589224e-01 -9.19531405e-01
-4.92818922e-01 -2.67883807e-01 -5.27389824e-01 -2.55447745e-01
-3.15700799e-01 -1.95172444e-01 -1.64406613e-01 -6.09601140e-01
4.41281229e-01 -7.07394481e-01 -1.07221484e-01 -1.74471632e-01
-4.85768497e-01 -2.91675776e-01 5.49666762e-01 -4.34246778e-01
8.81305993e-01 -2.34933472e+00 -6.37777627e-01 6.91757381e-01
1.35397270e-01 4.35465962e-01 -1.04940221e-01 -2.28987291e-01
-3.03937882e-01 3.70662183e-01 -7.42803002e-03 -3.27458948e-01
9.72076729e-02 2.77572691e-01 -2.76477784e-01 6.01724327e-01
2.12203383e-01 4.94149387e-01 -5.64545870e-01 -3.02318066e-01
2.79965043e-01 7.10395038e-01 -5.38817465e-01 -1.62901112e-03
1.07055984e-01 7.52784610e-02 -1.08333983e-01 1.22519754e-01
1.09217799e+00 1.59868136e-01 -4.67021227e-01 -2.84928977e-01
9.08649042e-02 -1.16590865e-01 -1.48191226e+00 1.21650243e+00
-3.87267649e-01 7.57867515e-01 -5.41989028e-01 -8.66013467e-01
1.29973114e+00 3.63575459e-01 7.58627579e-02 -2.69305110e-01
2.84256309e-01 4.83578324e-01 1.47919893e-01 -9.16742757e-02
2.38139167e-01 1.05409950e-01 2.00739980e-01 -3.46626997e-01
6.30998969e-01 -1.07529238e-01 3.74700651e-02 -1.37569681e-01
8.75315249e-01 -2.73829997e-01 -3.59872133e-02 -4.13093776e-01
2.57929921e-01 -5.04271686e-01 4.69648838e-01 9.80789602e-01
-1.74106404e-01 9.85312283e-01 7.84755766e-01 -1.83668047e-01
-1.21615827e+00 -1.50373352e+00 -6.19989097e-01 3.27097148e-01
-3.48535091e-01 -1.75020799e-01 -8.27849686e-01 -6.31530881e-01
-2.76230648e-02 9.24382627e-01 -5.47258258e-01 -3.35356712e-01
-2.01551449e-02 -6.47855580e-01 9.86295044e-01 7.11205184e-01
5.73458731e-01 -8.60830069e-01 4.12772745e-02 1.82434306e-01
4.24307555e-01 -9.99811471e-01 -1.75466724e-02 4.33496207e-01
-1.06200898e+00 -8.68238389e-01 -9.06217933e-01 -5.48957825e-01
7.34053016e-01 -4.79013413e-01 9.27802801e-01 -4.01368469e-01
-1.09697536e-01 -6.21883124e-02 -2.40045324e-01 -5.72930574e-01
-1.78537816e-01 -1.53419673e-01 4.16448116e-01 -2.03596935e-01
6.04543030e-01 -7.97766447e-01 -4.62832093e-01 5.27906835e-01
-9.64867353e-01 -5.81423938e-01 7.23119736e-01 7.14972854e-01
7.20042229e-01 4.28092182e-01 6.64043427e-01 -6.70681894e-01
6.24002695e-01 -5.86172223e-01 -8.81461620e-01 -1.39801912e-02
-4.95240360e-01 1.58646166e-01 9.63902771e-01 -7.74062395e-01
-6.81211829e-01 -1.47429988e-01 -5.67311704e-01 -9.89383221e-01
-7.63378620e-01 5.37443399e-01 -1.60859153e-01 9.51597467e-02
1.07830834e+00 -8.72072652e-02 -4.56971407e-01 -5.12146473e-01
2.22398546e-02 7.86478758e-01 5.91290176e-01 -5.82727492e-01
6.77314460e-01 2.40429953e-01 -1.65578142e-01 -7.81929851e-01
-6.88126504e-01 -4.46438372e-01 -6.03160322e-01 8.64130780e-02
8.44697595e-01 -9.26017821e-01 -6.30443096e-01 7.11029887e-01
-1.17373431e+00 -1.95969597e-01 -2.42697403e-01 1.21785903e+00
-4.14984435e-01 4.98115178e-03 -5.18644035e-01 -9.02074158e-01
-8.46715569e-02 -1.16891909e+00 6.56653643e-01 3.28417540e-01
1.96305886e-02 -8.95562589e-01 -9.00239795e-02 -4.23927158e-01
4.48506802e-01 3.13288458e-02 7.23955214e-01 -1.17346430e+00
-2.47789040e-01 -5.19772470e-01 -5.93260407e-01 1.06158745e+00
-2.03234375e-01 1.78906038e-01 -1.22993147e+00 -1.51229993e-01
-1.19711757e-02 -2.68835396e-01 1.04535437e+00 6.92647815e-01
1.77069068e+00 -2.76824713e-01 7.99470171e-02 7.53227055e-01
1.41562319e+00 3.77835929e-02 9.92098987e-01 -1.31923221e-02
4.71866995e-01 1.61100283e-01 4.95199978e-01 4.47983950e-01
-1.60922885e-01 2.13502571e-01 5.47519863e-01 2.37530038e-01
2.89218295e-02 -2.25118145e-01 2.43519872e-01 6.22180223e-01
4.73130960e-03 -2.67655522e-01 -9.08287048e-01 3.98634791e-01
-1.66815567e+00 -8.23791981e-01 -2.67639577e-01 2.53285742e+00
7.34764636e-01 6.66179121e-01 -3.20948750e-01 1.78723335e-01
7.57642269e-01 -3.93204391e-01 -4.86613899e-01 -2.74006754e-01
5.25923371e-02 3.41981143e-01 8.14344347e-01 2.86966324e-01
-1.00931466e+00 7.63708770e-01 5.60754728e+00 1.37969780e+00
-1.00696111e+00 2.41818018e-02 8.45121622e-01 2.71653920e-01
4.02861536e-02 -2.79596001e-01 -1.21737957e+00 6.39240026e-01
1.28674614e+00 2.16594547e-01 6.39821142e-02 1.18872666e+00
-2.23871451e-02 -3.14192265e-01 -1.13428080e+00 1.22727358e+00
-2.23047427e-05 -1.08252013e+00 -1.83286682e-01 9.18329358e-02
8.32652092e-01 5.58837056e-01 2.22117633e-01 4.88746285e-01
3.35250348e-01 -1.26753521e+00 7.74840295e-01 8.15762997e-01
9.73254323e-01 -1.06021070e+00 1.28007793e+00 5.47481477e-01
-9.13963139e-01 2.90204465e-01 -1.03092253e+00 2.22760066e-01
-2.84452170e-01 1.28355551e+00 -9.06343222e-01 1.39498249e-01
1.05010545e+00 4.40621465e-01 -5.73627830e-01 1.51808262e+00
-2.19073564e-01 8.00300717e-01 -7.28433967e-01 -1.05770387e-01
2.05692470e-01 -8.80609155e-02 2.00841248e-01 1.03044760e+00
8.40924859e-01 -4.80642438e-01 -2.81989396e-01 1.19220054e+00
-2.07689106e-01 -2.22015887e-01 -3.98795724e-01 1.68131605e-01
4.61822599e-01 9.94223893e-01 -8.05548429e-01 -3.20504874e-01
-2.06803009e-01 6.89472139e-01 3.16581875e-01 5.81806123e-01
-1.05690634e+00 -7.55151272e-01 5.97718954e-01 -1.58815637e-01
3.36050063e-01 -4.82630849e-01 -1.57705307e-01 -1.03507340e+00
-1.70847595e-01 -4.11743432e-01 1.01990446e-01 -7.59184062e-01
-1.76078916e+00 8.95700812e-01 2.57857978e-01 -1.23016047e+00
-1.84353352e-01 -1.08161116e+00 -6.54147983e-01 1.21687198e+00
-1.43755269e+00 -4.19388235e-01 -2.80480146e-01 7.17117429e-01
2.61349142e-01 -3.90581161e-01 8.72078955e-01 4.39890236e-01
-3.28564912e-01 8.10018301e-01 5.89537263e-01 5.07347941e-01
7.72592366e-01 -1.19899845e+00 3.93230110e-01 6.15008652e-01
3.96778107e-01 4.82880294e-01 5.91854990e-01 -5.24567366e-01
-2.40516439e-01 -1.36018908e+00 6.25551045e-01 -2.15273485e-01
5.83067179e-01 -4.28357810e-01 -9.86000836e-01 4.08016235e-01
-5.02483428e-01 4.63123292e-01 5.46158135e-01 1.92099400e-02
-4.57691073e-01 -2.14540884e-01 -1.39767885e+00 3.94717902e-01
5.81094146e-01 -4.67339605e-01 -5.71340978e-01 9.49470699e-02
4.32725668e-01 -2.63082862e-01 -9.27059650e-01 4.61216033e-01
5.10467052e-01 -1.21944034e+00 8.93435776e-01 -4.36839491e-01
-1.28431059e-02 -3.63137960e-01 -5.06878793e-01 -1.49048865e+00
7.27101266e-02 9.03514475e-02 -4.02476825e-02 1.15192628e+00
5.88939309e-01 -9.76577878e-01 9.71650362e-01 5.49239337e-01
-2.84050733e-01 -5.46530068e-01 -1.22352099e+00 -1.18796456e+00
2.96404511e-01 -9.74707663e-01 4.38125342e-01 5.39761662e-01
-7.38840044e-01 -3.57712120e-01 3.90996821e-02 5.50859451e-01
5.96674025e-01 -8.10506940e-01 6.34576142e-01 -1.52159774e+00
-1.80880919e-01 -3.80153865e-01 -1.02091610e+00 -7.36603677e-01
2.50161767e-01 -6.25072718e-01 2.58623421e-01 -1.29446292e+00
-2.79148430e-01 -6.47807539e-01 -5.64438760e-01 2.89052576e-01
2.58896321e-01 4.53624427e-01 -7.49036893e-02 1.52790859e-01
-3.30749243e-01 6.20002747e-01 6.21947825e-01 9.20515582e-02
3.95729877e-02 3.29937428e-01 -6.78106248e-02 1.15058219e+00
1.13641739e+00 -9.15578485e-01 -3.81632537e-01 -1.66354194e-01
2.42170095e-01 -4.19713140e-01 6.52643979e-01 -1.80097795e+00
3.13027859e-01 1.48937702e-01 8.43945086e-01 -9.28740144e-01
2.55664915e-01 -1.07184947e+00 -4.16658409e-02 2.32925236e-01
-1.72711357e-01 -3.97545286e-02 2.24942625e-01 7.24337876e-01
-4.72045064e-01 -7.07647324e-01 9.00647461e-01 1.90140888e-01
-5.70454478e-01 5.02749145e-01 -4.08167988e-01 7.24836439e-02
7.78978407e-01 -2.85109937e-01 -3.44121337e-01 -2.02233717e-01
-7.66442955e-01 -3.43754530e-01 -4.34703715e-02 1.15672536e-01
9.37807560e-01 -1.33378768e+00 -5.42849541e-01 4.86721456e-01
1.56954318e-01 2.34373659e-01 1.70258522e-01 7.46002018e-01
-6.82313144e-01 1.58998728e-01 -3.42505336e-01 -8.47932279e-01
-8.86717260e-01 1.05587952e-01 4.40842450e-01 2.15499587e-02
-1.22751519e-01 1.08095312e+00 3.20979580e-02 -4.99932468e-01
5.87186754e-01 -8.52918446e-01 -1.93812758e-01 -1.85084656e-01
3.70623916e-01 3.07750016e-01 4.48799670e-01 -3.16101938e-01
-2.91220844e-01 3.65540653e-01 8.68520513e-02 -2.17959210e-01
1.08152831e+00 2.94127733e-01 2.34400079e-01 6.26667023e-01
1.51793396e+00 -2.42318347e-01 -1.32829392e+00 8.88632331e-03
9.58559066e-02 -4.75165427e-01 1.02605745e-01 -5.03674984e-01
-9.27115738e-01 1.14423263e+00 9.50297475e-01 1.38993785e-01
7.91572511e-01 -3.66108194e-02 2.18715340e-01 6.34657919e-01
2.55660743e-01 -1.08439302e+00 2.74883993e-02 8.77286136e-01
7.50250757e-01 -1.13799536e+00 -3.92016679e-01 5.47208004e-02
-3.85450304e-01 1.23873425e+00 7.98279703e-01 -5.48868120e-01
1.19685972e+00 2.68860281e-01 -8.53238329e-02 2.15095639e-01
-4.26250905e-01 -3.13167386e-02 5.84336102e-01 7.44656563e-01
3.98916572e-01 1.58216693e-02 1.35018826e-01 7.65845060e-01
-4.19023305e-01 -8.90593603e-02 3.96260321e-01 3.99565458e-01
-5.49771488e-01 -6.14295363e-01 -3.99035126e-01 8.97907019e-01
-4.11546320e-01 -3.22432697e-01 2.21889734e-01 7.13546455e-01
3.01527679e-01 8.02500725e-01 4.08146799e-01 -4.05220479e-01
2.49939427e-01 4.19706814e-02 1.57855958e-01 -5.52919030e-01
1.03817746e-01 -1.10769004e-01 -1.46038547e-01 -3.69197160e-01
8.01594555e-03 -1.80321246e-01 -1.24036515e+00 -2.30073333e-01
-6.67465150e-01 2.43273541e-01 1.11226761e+00 8.09252858e-01
1.71970800e-01 5.66438794e-01 2.84634739e-01 -1.01157844e+00
-7.94215739e-01 -1.26536477e+00 -1.00240207e+00 2.95074821e-01
1.36543810e-01 -7.93787003e-01 -8.34949672e-01 -2.86317945e-01] | [7.57492733001709, 3.6338205337524414] |
d5179d35-b78a-4845-8e18-a4f88ad7420b | denoising-time-series-data-using-asymmetric | null | null | https://doi.org/10.1007/978-3-319-93040-4_23 | https://www.researchgate.net/publication/325808737_Denoising_Time_Series_Data_Using_Asymmetric_Generative_Adversarial_Networks | Denoising Time Series Data Using Asymmetric Generative Adversarial Networks | Denoising data is a preprocessing step for several time series mining algorithms. This step is especially important if the noise in data originates from diverse sources. Consequently, it is commonly used in biomedical applications that use Electroencephalography (EEG) data. In EEG data noise can occur due to ocular, muscular and cardiac activities. In this paper, we explicitly learn to remove noise from time series data without assuming a prior distribution of noise. We propose an online, fully automated, end-to-end system for denoising time series data. Our model for denoising time series is trained using unpaired training corpora and does not need information about the source of the noise or how it is manifested in the time series. We propose a new architecture called AsymmetricGAN that uses a generative adversarial network for denoising time series data. To analyze our approach, we create a synthetic dataset that is easy to visualize and interpret. We also evaluate and show the effectiveness of our approach on an existing EEG dataset. | ['Sunil Gandhi', 'David Hairston', 'Tinoosh Mohsenin', 'Tim Oates'] | 2018-06-17 | null | null | null | advances-in-knowledge-discovery-and-data | ['eeg-denoising'] | ['methodology'] | [ 1.77882969e-01 -1.97471499e-01 7.94336319e-01 -4.03981298e-01
-3.82063746e-01 -5.62568605e-01 3.85618269e-01 2.36758329e-02
-4.91652012e-01 9.19202805e-01 1.39152646e-01 -8.83641243e-02
-1.59397274e-01 -6.01544738e-01 -7.02873349e-01 -7.87503898e-01
-2.58448511e-01 1.43047441e-02 -2.19108686e-01 -2.51769871e-01
-9.14133117e-02 4.18118268e-01 -1.17724764e+00 1.42497301e-01
8.47014427e-01 8.41261685e-01 -3.02446615e-02 5.55066347e-01
2.17700392e-01 6.32022202e-01 -1.33905864e+00 -1.80161566e-01
3.84678334e-01 -8.16214502e-01 -2.58090109e-01 -6.43095225e-02
-1.30537808e-01 -1.44166440e-01 -4.83858883e-01 1.30144882e+00
6.54884934e-01 3.86314131e-02 5.06112397e-01 -1.35921240e+00
-1.03719808e-01 6.71883702e-01 -5.63449025e-01 5.10059476e-01
1.37913719e-01 1.13579802e-01 1.57312602e-01 -1.75373331e-01
5.27949333e-01 8.18942189e-01 7.67620206e-01 5.97022116e-01
-1.52885473e+00 -8.36012542e-01 -7.83005878e-02 2.86569893e-01
-1.18365526e+00 -2.27319568e-01 1.29111063e+00 -3.81264836e-01
5.02849579e-01 2.16677412e-01 7.66364217e-01 1.76510227e+00
5.53472698e-01 4.70189005e-01 1.32643771e+00 -1.64505869e-01
5.97869158e-01 -3.51404458e-01 1.95480779e-01 -8.86371955e-02
-3.36526968e-02 4.94746584e-03 -4.52013373e-01 -1.57174900e-01
4.99911219e-01 2.13479176e-01 -3.91100705e-01 1.48933411e-01
-9.99180794e-01 3.66888285e-01 1.96416870e-01 4.58865553e-01
-7.52696514e-01 3.71411324e-01 5.52877545e-01 8.63979518e-01
5.57404697e-01 3.72868627e-01 -5.06407380e-01 -4.36863154e-01
-1.27108443e+00 2.63664429e-03 7.86109865e-01 6.19056582e-01
2.53858179e-01 2.66401708e-01 8.98592621e-02 5.07346511e-01
-2.14801848e-01 3.89763802e-01 8.31220865e-01 -6.71258271e-01
6.81658238e-02 1.95073098e-01 -1.34820744e-01 -8.40737164e-01
-5.58866680e-01 -5.08874774e-01 -1.30003953e+00 1.59727916e-01
3.57020587e-01 -5.95352948e-01 -1.05185139e+00 2.01088881e+00
-7.74543658e-02 7.00522304e-01 -2.62437109e-02 6.89360321e-01
6.13705516e-01 5.34068048e-01 -8.35629478e-02 -5.66527605e-01
1.03145993e+00 -3.06406796e-01 -1.20936418e+00 5.05900867e-02
1.27436146e-01 -5.47296882e-01 6.95194364e-01 9.33661878e-01
-8.66034806e-01 -2.12274298e-01 -1.05770576e+00 3.01214978e-02
-4.23188210e-01 -7.87904263e-02 2.92915821e-01 5.25115430e-01
-7.90301383e-01 9.06084657e-01 -1.27969754e+00 -1.73892364e-01
4.44231570e-01 4.28198457e-01 -5.17200649e-01 3.51570636e-01
-1.05984652e+00 8.81271780e-01 1.18338287e-01 1.73099279e-01
-8.42042446e-01 -6.85245097e-01 -4.86924708e-01 -6.77046878e-03
6.43817857e-02 -3.46821606e-01 1.02791870e+00 -1.26972604e+00
-1.28035438e+00 3.13923448e-01 -9.29084048e-02 -7.51196206e-01
7.53897607e-01 -7.43073449e-02 -6.58667505e-01 5.55505566e-02
-1.89198002e-01 -9.86586325e-03 1.21134472e+00 -8.86427462e-01
8.03837851e-02 -4.98664588e-01 -3.93822789e-01 -5.26859879e-01
-1.97106272e-01 -1.95542693e-01 3.30730900e-02 -1.25688148e+00
7.49758109e-02 -5.85434556e-01 -8.34996700e-02 -2.53245175e-01
-3.52319121e-01 1.45344734e-01 9.19379354e-01 -9.60889459e-01
1.11664438e+00 -2.23907995e+00 2.06105322e-01 6.34574473e-01
3.04680377e-01 -8.13080817e-02 -1.53439641e-01 5.13075292e-01
-4.08126920e-01 -1.71278231e-02 -5.08329988e-01 -3.67074043e-01
-2.14861378e-01 1.97210878e-01 -3.26281130e-01 6.12637937e-01
1.68291911e-01 5.11730909e-01 -7.10506320e-01 2.01038733e-01
1.04293980e-01 8.06794465e-01 -2.99405158e-01 2.82259881e-01
-1.04162153e-02 1.05769479e+00 -2.79201195e-02 1.58406556e-01
6.60700440e-01 1.91302836e-01 -2.49968860e-02 -4.42867577e-01
-1.47586013e-03 2.30889097e-02 -1.03656435e+00 1.81991899e+00
-2.26996660e-01 7.85778522e-01 4.71892394e-03 -1.19394708e+00
7.73832619e-01 5.02942026e-01 6.42160892e-01 -8.51583004e-01
5.16191483e-01 2.04166472e-01 3.08496773e-01 -6.16827428e-01
-1.56221732e-01 -1.68162167e-01 1.96601257e-01 6.20580554e-01
3.52649719e-01 -1.67766705e-01 2.32072905e-01 -1.98200643e-01
1.61205375e+00 -5.72485588e-02 1.68664455e-01 -1.46699831e-01
2.48907015e-01 -3.70423079e-01 5.40031016e-01 5.45901716e-01
8.80206469e-03 7.38141894e-01 8.23424995e-01 -4.38744903e-01
-1.12702096e+00 -7.67634392e-01 -4.71081808e-02 3.34680468e-01
-2.46162459e-01 -4.27135199e-01 -1.10305560e+00 -4.20817584e-01
-3.69412214e-01 6.63566887e-01 -8.20911944e-01 -3.59310180e-01
-3.62024903e-01 -7.61287093e-01 5.29308140e-01 5.10362089e-01
2.52619028e-01 -1.09259951e+00 -6.86776757e-01 5.08941889e-01
-2.14326382e-01 -8.91500652e-01 -3.34602565e-01 6.35383725e-01
-9.46130455e-01 -1.04468226e+00 -6.57612383e-01 -3.27959746e-01
7.38003433e-01 -2.00706586e-01 8.95891488e-01 -1.58765063e-01
-3.92163604e-01 3.52763623e-01 -2.86953896e-01 -1.02402115e+00
-1.63463309e-01 -2.47041106e-01 1.05809495e-01 1.90547466e-01
5.90623081e-01 -1.19887054e+00 -6.41959429e-01 2.37945672e-02
-1.24724746e+00 -1.57561421e-01 1.73913300e-01 9.18304265e-01
6.06930494e-01 5.39124489e-01 7.75476575e-01 -7.59792089e-01
1.12318313e+00 -5.87872982e-01 -5.49300432e-01 -4.52692062e-02
-1.25216931e-01 4.24268022e-02 9.55133438e-01 -6.65206611e-01
-5.97371578e-01 2.40129903e-02 -3.59751612e-01 -5.63280940e-01
-3.14930737e-01 7.34544516e-01 -5.01795374e-02 7.73961395e-02
7.01763272e-01 1.92435145e-01 8.03392082e-02 -6.37546301e-01
-5.52040339e-02 5.49123824e-01 7.42809236e-01 -2.83988535e-01
7.56986499e-01 5.51443517e-01 5.60940104e-03 -1.01995885e+00
-4.25612599e-01 -1.38695195e-01 -5.37905812e-01 -5.71061186e-02
6.09576702e-01 -7.93075979e-01 -6.86305702e-01 6.98586881e-01
-1.26863110e+00 -3.63098741e-01 -2.90616632e-01 5.62398195e-01
-5.42158604e-01 1.08829856e-01 -5.03213346e-01 -7.17505753e-01
-4.87380743e-01 -9.34954703e-01 6.82117224e-01 1.70290228e-02
-3.10151786e-01 -9.67423916e-01 8.16093236e-02 -3.46545666e-01
4.61899608e-01 6.63868010e-01 7.39106476e-01 -7.57503450e-01
-1.53491810e-01 -3.43591005e-01 3.86969179e-01 6.31039977e-01
3.44496250e-01 -1.52042374e-01 -1.15349817e+00 -3.41509730e-01
7.57544637e-01 2.66764499e-02 6.31614566e-01 5.20616770e-01
1.65249741e+00 -1.78688347e-01 1.60754114e-01 8.44741285e-01
1.24935758e+00 3.90236408e-01 8.78351212e-01 1.06246278e-01
5.47593892e-01 5.69487572e-01 1.22683877e-02 5.67604959e-01
-6.16166405e-02 1.21425986e-01 2.86189020e-01 -1.51428118e-01
2.58573920e-01 6.38587121e-03 3.44278842e-01 9.18581784e-01
-3.24529678e-01 -2.87013680e-01 -7.87387729e-01 4.01681215e-01
-1.62256360e+00 -1.07831502e+00 -1.05780840e-01 2.07312465e+00
8.54973674e-01 -5.01659289e-02 7.84992874e-02 7.35292912e-01
5.02340376e-01 -3.60616982e-01 -7.12835133e-01 -3.97770256e-02
-1.97201148e-01 8.69058073e-01 3.13307285e-01 3.93378139e-02
-9.81649280e-01 3.19649667e-01 6.13884068e+00 4.55582768e-01
-1.47913456e+00 9.50966701e-02 5.29240191e-01 -2.50779510e-01
-1.04355998e-01 -6.34938240e-01 2.65139759e-01 7.79692113e-01
1.06365883e+00 -4.18754309e-01 7.92404413e-01 2.50797004e-01
6.69622838e-01 7.09197894e-02 -1.27163994e+00 1.38912010e+00
-5.99561110e-02 -7.76219368e-01 -3.38723570e-01 -2.33894840e-01
4.71842855e-01 -7.90743083e-02 -9.25222784e-02 -1.14666440e-01
-1.80839822e-01 -1.22623837e+00 7.70770311e-01 7.96022892e-01
4.39964205e-01 -9.33876812e-01 9.65277493e-01 4.63326186e-01
-6.72579885e-01 6.91683218e-02 -5.88357858e-02 -1.49309963e-01
4.92621101e-02 8.59070480e-01 -3.45997930e-01 3.65154684e-01
6.50256038e-01 7.86135435e-01 -5.07616043e-01 1.23004353e+00
-5.39206862e-01 8.02437425e-01 -3.96787882e-01 2.86404163e-01
-1.93757996e-01 -3.20504189e-01 6.05132878e-01 9.63104725e-01
6.46639884e-01 2.68118352e-01 -2.99801052e-01 8.65632534e-01
-1.79050311e-01 -1.94779232e-01 -6.47380590e-01 -2.87330508e-01
1.15133554e-01 1.16921842e+00 -8.54281247e-01 1.02938944e-02
-4.34161335e-01 1.03835678e+00 -1.15032308e-01 5.85177481e-01
-7.92474627e-01 -6.44268692e-01 4.43898648e-01 -9.33283865e-02
9.63674337e-02 -2.38982141e-01 -2.84828216e-01 -1.36860979e+00
2.93788970e-01 -1.12373221e+00 1.06319033e-01 -8.22430849e-01
-1.72263265e+00 8.24645042e-01 -2.14987084e-01 -1.32525718e+00
-2.42736697e-01 -4.02481556e-01 -6.97972894e-01 9.21973050e-01
-1.16975296e+00 -6.90618038e-01 -4.12364393e-01 9.11055982e-01
1.70737430e-01 -9.03552547e-02 8.17531705e-01 4.70289111e-01
-3.81399512e-01 3.56210262e-01 1.12998351e-01 2.56820589e-01
7.61853814e-01 -1.34524214e+00 3.76877636e-01 1.07431650e+00
-2.14204807e-02 7.88492322e-01 9.82165694e-01 -6.56494141e-01
-1.37956893e+00 -1.15109193e+00 3.98487985e-01 -9.32115242e-02
8.14708889e-01 -6.02543831e-01 -1.15076101e+00 6.29401147e-01
5.41438162e-01 3.64473127e-02 8.07186365e-01 -1.78918824e-01
-5.36160078e-03 -3.66432279e-01 -1.35307968e+00 4.93213326e-01
7.27105558e-01 -4.61493939e-01 -8.29086542e-01 1.71250179e-01
2.72523224e-01 -3.25550258e-01 -8.43687534e-01 6.73363358e-02
3.64675611e-01 -7.97416091e-01 5.14817417e-01 -4.27524060e-01
4.00045723e-01 -3.54222268e-01 2.24605054e-01 -1.96367967e+00
5.07979915e-02 -1.03133178e+00 -2.02162907e-01 1.05119324e+00
1.42386988e-01 -7.13042736e-01 4.55079734e-01 5.17838538e-01
-1.34033365e-02 -1.29670411e-01 -7.93005884e-01 -6.34386063e-01
-5.41139133e-02 -6.22095048e-01 7.69039989e-01 9.52330410e-01
-7.94975013e-02 1.74484357e-01 -5.14490604e-01 1.27564147e-01
7.03182340e-01 -2.19361573e-01 6.00803494e-01 -1.27729762e+00
-5.36891371e-02 -5.67390844e-02 -6.30323827e-01 -3.62974733e-01
7.23679662e-02 -5.37532270e-01 9.05788392e-02 -1.12390828e+00
-1.54684171e-01 1.42323554e-01 -5.65638423e-01 5.16451538e-01
1.07193328e-01 3.25825781e-01 -5.18607534e-02 -1.73096970e-01
-2.67854630e-04 5.77646196e-01 1.02022922e+00 -1.24324560e-01
-2.45571628e-01 -1.07361928e-01 -4.02623773e-01 7.94084847e-01
1.00388539e+00 -6.93181992e-01 -5.48728049e-01 -4.45965260e-01
1.41302034e-01 1.06734447e-01 4.15988028e-01 -1.17880380e+00
3.70261669e-01 2.70909369e-01 6.44004524e-01 -4.74609703e-01
1.81478590e-01 -1.27776289e+00 5.99567413e-01 4.15106505e-01
-1.41791627e-01 3.42106074e-01 3.01229358e-01 5.41520059e-01
-4.55807298e-01 1.95728429e-03 6.44519210e-01 -2.23439671e-02
-2.78056085e-01 2.60136575e-01 -4.70000833e-01 -1.55187711e-01
9.40670013e-01 1.45378277e-01 -1.97988585e-01 -6.31868064e-01
-1.06716645e+00 7.91225806e-02 2.14050785e-01 2.40507782e-01
5.51056564e-01 -1.19675350e+00 -5.78770638e-01 5.72905600e-01
-7.01975748e-02 -4.06924248e-01 2.22716033e-01 1.05649996e+00
-4.21188354e-01 -2.49829918e-01 -5.69158733e-01 -4.63470131e-01
-1.18047845e+00 4.52670634e-01 3.59391868e-01 2.67933309e-01
-7.39619732e-01 4.25646454e-01 -2.29131967e-01 -3.95058133e-02
4.05678272e-01 -6.65685534e-01 -2.84937918e-01 2.18314573e-01
7.93199122e-01 3.59547108e-01 3.61041754e-01 -1.61994293e-01
-2.74963230e-01 2.34261632e-01 2.50566900e-01 -2.33453169e-01
1.96322548e+00 1.20827004e-01 -4.70046312e-01 7.34045923e-01
1.11221230e+00 -6.39617071e-02 -1.11384463e+00 2.10751742e-01
-1.10150449e-01 -1.04232274e-01 8.52725357e-02 -7.95642674e-01
-1.31379580e+00 8.91033411e-01 8.64287078e-01 5.26673496e-01
1.73984158e+00 -6.16331160e-01 6.25518799e-01 3.15119594e-01
4.89511609e-01 -1.02363038e+00 -1.32270709e-01 3.78663123e-01
9.98950362e-01 -1.00617075e+00 -2.04216629e-01 4.88159880e-02
-2.88961411e-01 1.28484368e+00 1.27027154e-01 -4.11707908e-01
8.41486692e-01 7.52391338e-01 2.50048101e-01 -1.20324515e-01
-7.07669437e-01 5.82123781e-03 1.28667817e-01 7.60631561e-01
3.55545133e-01 -1.21133476e-01 -5.86688042e-01 9.93914366e-01
-5.27496159e-01 2.63628066e-01 6.84155583e-01 9.28738594e-01
1.78127885e-01 -1.16439688e+00 -4.14856970e-01 5.37988365e-01
-9.11813915e-01 -1.78982224e-02 -3.69928479e-01 4.90568608e-01
1.83854178e-01 1.05358064e+00 1.83254499e-02 -3.93270701e-01
4.28444356e-01 9.54217091e-02 5.17103016e-01 -2.78942972e-01
-8.06457758e-01 4.59942043e-01 -3.49955499e-01 -6.20358407e-01
-6.04087710e-01 -5.56735873e-01 -1.06550801e+00 -3.03817123e-01
-6.68878332e-02 1.04121186e-01 9.24158990e-01 8.31594110e-01
3.30217063e-01 1.09991801e+00 5.24984658e-01 -7.08535016e-01
-2.12615013e-01 -1.10718226e+00 -7.45714545e-01 7.56644130e-01
6.71899676e-01 -3.86187464e-01 -5.31448662e-01 4.80716497e-01] | [13.179211616516113, 3.4496243000030518] |
6233c1ba-668b-4359-9af0-867ae1ab14b8 | question-answer-sentence-graph-for-joint | 2203.03549 | null | https://arxiv.org/abs/2203.03549v2 | https://arxiv.org/pdf/2203.03549v2.pdf | Question-Answer Sentence Graph for Joint Modeling Answer Selection | This research studies graph-based approaches for Answer Sentence Selection (AS2), an essential component for retrieval-based Question Answering (QA) systems. During offline learning, our model constructs a small-scale relevant training graph per question in an unsupervised manner, and integrates with Graph Neural Networks. Graph nodes are question sentence to answer sentence pairs. We train and integrate state-of-the-art (SOTA) models for computing scores between question-question, question-answer, and answer-answer pairs, and use thresholding on relevance scores for creating graph edges. Online inference is then performed to solve the AS2 task on unseen queries. Experiments on two well-known academic benchmarks and a real-world dataset show that our approach consistently outperforms SOTA QA baseline models. | ['Yizhou Sun', 'Alessandro Moschitti', 'Thuy Vu', 'Roshni G. Iyer'] | 2022-02-16 | null | null | null | null | ['answer-selection'] | ['natural-language-processing'] | [ 3.22068334e-01 5.56601286e-01 -8.09606835e-02 -5.11925101e-01
-1.40350628e+00 -6.08518541e-01 2.81872123e-01 9.88378227e-01
-2.90736794e-01 2.52705485e-01 3.02194089e-01 -7.07156122e-01
-2.05057904e-01 -1.24236321e+00 -5.72613835e-01 2.43694812e-01
-2.02763863e-02 1.03483188e+00 8.28317046e-01 -6.62514687e-01
4.53525811e-01 -3.03565282e-02 -1.01680386e+00 6.67905688e-01
1.14050210e+00 9.26627457e-01 -2.93145150e-01 1.31746483e+00
-6.90828085e-01 1.50967503e+00 -7.76246905e-01 -1.07193327e+00
-1.87955439e-01 -6.48364723e-01 -1.58948433e+00 -3.45002800e-01
1.30947506e+00 -5.68677559e-02 -4.99655306e-01 1.08233607e+00
4.71524924e-01 4.34035003e-01 5.30935228e-01 -7.61357188e-01
-1.08426344e+00 5.70689917e-01 5.95126301e-02 4.62575257e-01
1.16949213e+00 -1.16922485e-03 1.79321635e+00 -9.44560528e-01
8.25238287e-01 1.25957680e+00 3.50372136e-01 3.63385290e-01
-1.00880611e+00 2.42369995e-01 -7.13701472e-02 7.02696800e-01
-9.54256177e-01 -2.23272726e-01 7.56473303e-01 -3.44785266e-02
1.31582141e+00 6.21842504e-01 5.31881154e-01 4.84216720e-01
4.07273740e-01 1.01783633e+00 6.35488093e-01 -6.71613812e-01
1.98305815e-01 -3.60098243e-01 1.09725821e+00 1.29734874e+00
-8.21678266e-02 -8.22781026e-01 -6.13127410e-01 -5.57006836e-01
-6.47557713e-03 -5.36291540e-01 -4.31942850e-01 -1.04220808e-01
-7.35329449e-01 1.07256925e+00 6.57454193e-01 5.51269203e-02
-4.06734526e-01 4.34147045e-02 2.20578626e-01 1.12394631e+00
3.99809390e-01 9.17604387e-01 -3.30364108e-01 2.98708141e-01
-4.23600614e-01 2.43786320e-01 1.44066787e+00 8.26624572e-01
9.31721091e-01 -5.66422641e-01 -1.11669791e+00 8.04588020e-01
3.18335354e-01 3.40520978e-01 1.23149492e-01 -9.41678762e-01
9.29864705e-01 1.22249246e+00 -3.91076028e-01 -1.24029267e+00
-3.18979800e-01 -2.86033183e-01 -4.97113466e-01 -6.24576747e-01
3.34996015e-01 1.97865628e-02 -8.76150906e-01 1.19032466e+00
4.38054383e-01 -3.32273543e-01 -1.22439951e-01 7.53017008e-01
1.65141618e+00 4.53684509e-01 3.14302668e-02 -1.67003237e-02
1.59490383e+00 -1.42698681e+00 -6.52491927e-01 -4.61363763e-01
1.01035786e+00 -5.16547740e-01 1.45818973e+00 3.83946523e-02
-1.16391575e+00 -3.61027360e-01 -5.50824225e-01 -6.35151863e-01
-4.05218244e-01 -2.50396043e-01 3.91291142e-01 5.60894012e-01
-1.56898522e+00 2.22195402e-01 -3.72344479e-02 -3.51368308e-01
1.65685788e-01 3.21091503e-01 -1.62558332e-01 -4.90452260e-01
-1.62468696e+00 1.01076317e+00 -1.02608632e-02 5.70719279e-02
-8.08196306e-01 -2.42546469e-01 -1.02914095e+00 3.23652804e-01
6.87943578e-01 -1.34627283e+00 1.50079346e+00 -5.15904546e-01
-1.36847425e+00 1.00312078e+00 -5.07945538e-01 -5.56933463e-01
-1.84412435e-01 -8.19483846e-02 -4.17673588e-01 8.86609018e-01
6.77244365e-02 4.67079312e-01 8.67209971e-01 -8.32139730e-01
-1.42343506e-01 -3.85360152e-01 5.42132556e-01 4.11085010e-01
-3.29738408e-01 1.13955483e-01 -6.93946242e-01 2.85466313e-02
3.28184038e-01 -3.77573788e-01 -1.93073362e-01 -5.06685853e-01
-6.06744826e-01 -1.09145188e+00 3.27822834e-01 -9.76094246e-01
1.66403568e+00 -1.23761857e+00 1.12035789e-01 4.82430339e-01
7.23994434e-01 -3.40624824e-02 -5.88524520e-01 7.93790221e-01
2.43771181e-01 1.04112662e-01 -7.98877999e-02 -1.38732910e-01
-5.28888544e-03 -1.14108309e-01 -1.31092265e-01 -1.72989160e-01
3.23742628e-01 1.51251519e+00 -1.20975411e+00 -9.28712308e-01
-3.22011858e-01 -2.88782209e-01 -5.18481672e-01 5.53026557e-01
-7.82542706e-01 -1.39364138e-01 -6.98039532e-01 7.60428846e-01
1.85673669e-01 -7.78647780e-01 1.72812656e-01 -2.45381612e-02
9.39846754e-01 8.47776592e-01 -5.52560985e-01 1.44101453e+00
-1.87862486e-01 5.38291395e-01 2.75635924e-02 -8.47738445e-01
8.72430503e-01 1.41404539e-01 -1.39865369e-01 -1.03144157e+00
4.45094034e-02 -1.02690002e-02 -2.84910172e-01 -9.19459045e-01
8.09710085e-01 6.16991878e-01 -3.65042947e-02 4.84929770e-01
4.09909159e-01 -7.48848379e-01 6.82600260e-01 1.08339155e+00
1.78465414e+00 -4.47553068e-01 2.24497635e-02 -4.48918417e-02
9.11155939e-01 3.78988266e-01 -3.62198621e-01 1.14191020e+00
-1.59363493e-01 6.37299776e-01 9.14058685e-01 -1.16638087e-01
-2.52618760e-01 -1.07514632e+00 6.28537834e-01 1.45360708e+00
-6.08958527e-02 -7.30039537e-01 -8.94592762e-01 -1.38057697e+00
6.62951265e-03 1.09306026e+00 -6.29015565e-01 -4.08816397e-01
-8.06231916e-01 -6.40610009e-02 2.17449158e-01 1.10492222e-01
1.49327770e-01 -1.15647173e+00 2.58132398e-01 1.36939898e-01
-4.14888561e-01 -1.04260719e+00 -7.98534632e-01 -2.37066939e-01
-9.53254461e-01 -1.48394692e+00 -3.35485667e-01 -1.12597930e+00
7.71476567e-01 5.34778178e-01 2.34898400e+00 7.07977712e-01
5.26788235e-02 1.12675488e+00 -4.34175134e-01 -6.08494133e-02
-4.30043697e-01 2.63289690e-01 -7.30743945e-01 -4.03572693e-02
6.49039984e-01 -5.63526750e-02 -8.51854920e-01 7.18703419e-02
-8.21585715e-01 -5.39677739e-01 9.06091928e-02 8.30103993e-01
5.40178895e-01 -5.75153351e-01 9.92824018e-01 -1.05755866e+00
1.62333143e+00 -5.52849293e-01 -4.75471884e-01 1.12949109e+00
-6.86652720e-01 6.31671473e-02 4.18183357e-01 1.81945041e-01
-8.55646610e-01 -5.01081169e-01 -3.07608962e-01 -2.60866135e-02
1.95485324e-01 9.85151947e-01 1.95277348e-01 -2.57031232e-01
1.04863548e+00 -2.30312701e-02 -1.79240450e-01 -7.17892824e-03
8.49296629e-01 6.02775156e-01 3.80420089e-01 -3.90777260e-01
6.14605010e-01 -9.46089774e-02 -1.47380590e-01 -5.35462201e-01
-1.41622627e+00 -9.35275435e-01 -3.53517383e-01 -4.38303471e-01
1.04905081e+00 -5.29378057e-01 -8.52383733e-01 -1.70406237e-01
-1.28380036e+00 -2.15802133e-01 -4.49007958e-01 3.00964154e-03
-9.99969319e-02 6.65595293e-01 -9.29458499e-01 -7.10373402e-01
-9.67594624e-01 -7.67415166e-01 1.07937837e+00 4.35070366e-01
-2.43145674e-01 -1.40597379e+00 3.54398340e-01 1.26365006e+00
3.75456870e-01 -4.06475663e-01 1.31068242e+00 -1.00964844e+00
-9.01735127e-01 -6.66083097e-01 -3.49808991e-01 2.05898747e-01
-3.59593153e-01 -1.98968977e-01 -7.30406523e-01 -1.14382267e-01
3.43767852e-02 -8.61473143e-01 1.06749213e+00 1.08651340e-01
1.05562639e+00 -2.14528054e-01 5.83821908e-02 -2.18756482e-01
1.02968061e+00 -4.40852672e-01 4.92679656e-01 -5.94980344e-02
7.54114449e-01 7.47856379e-01 4.32940334e-01 -2.74359882e-01
1.07901847e+00 1.50640368e-01 5.98407626e-01 2.78369486e-01
-4.10234451e-01 -4.39692855e-01 3.02024305e-01 1.46012986e+00
6.08409643e-01 -7.84705102e-01 -9.65811849e-01 7.79784858e-01
-1.75207019e+00 -7.08519399e-01 -8.44094217e-01 2.02829218e+00
8.88291776e-01 6.56743869e-02 -1.37568057e-01 -4.06720996e-01
5.87731898e-01 2.50177532e-01 -2.77441770e-01 -5.57851017e-01
-3.06851953e-01 7.76020288e-01 1.44755943e-02 8.89560163e-01
-8.29795122e-01 1.03411210e+00 6.69224119e+00 8.54101658e-01
-3.12563211e-01 1.12484381e-01 7.33471513e-01 4.71251458e-01
-8.98217440e-01 2.39974752e-01 -4.16827947e-01 -3.20528209e-01
1.03857338e+00 -2.31205598e-01 2.82345533e-01 6.24903560e-01
-3.68470341e-01 -2.80103832e-01 -9.88745332e-01 4.71750915e-01
6.72647834e-01 -1.54249871e+00 4.14514422e-01 -7.35234857e-01
6.13450348e-01 1.58762470e-01 -3.73020172e-01 7.00550437e-01
5.71409106e-01 -8.14928651e-01 -3.97272408e-02 6.62168741e-01
2.55315274e-01 -2.26823404e-01 8.59933257e-01 2.70066291e-01
-1.05845463e+00 1.96795203e-02 -3.68922055e-01 1.44395560e-01
1.91781625e-01 5.96602201e-01 -1.11589599e+00 8.80980253e-01
5.83343685e-01 1.57278419e-01 -1.42386544e+00 1.06408441e+00
-9.27001357e-01 1.17439616e+00 -1.04678474e-01 -6.80991471e-01
1.77758157e-01 -2.77133763e-01 5.19288421e-01 1.00381291e+00
-2.16644868e-01 9.14142951e-02 4.51996997e-02 6.07976854e-01
-7.90616393e-01 6.68434441e-01 -5.30850828e-01 -7.78725892e-02
3.46841365e-01 1.43692958e+00 -4.00003314e-01 -7.34636366e-01
-3.12879652e-01 1.01658762e+00 9.11238551e-01 4.26048130e-01
-9.52535868e-02 -7.35366046e-01 -3.96246493e-01 -1.10714652e-01
1.96614657e-02 -1.06740393e-01 -1.11492909e-01 -1.39007449e+00
2.75733322e-01 -1.15020132e+00 1.35619676e+00 -9.88998890e-01
-1.73537004e+00 4.89068985e-01 -2.65331864e-01 -6.18820131e-01
-2.44566128e-01 -4.77560431e-01 -1.07577109e+00 8.10182333e-01
-1.58039224e+00 -8.05048585e-01 -3.85777384e-01 7.37136245e-01
4.27242547e-01 3.55613464e-03 8.41338038e-01 1.08922169e-01
-4.44448084e-01 5.75406730e-01 -3.61239612e-01 3.51435870e-01
6.02166235e-01 -1.60335732e+00 8.32677007e-01 7.04544246e-01
6.97844863e-01 6.56569779e-01 2.18398809e-01 -7.31269956e-01
-1.80126643e+00 -7.99935400e-01 1.67010832e+00 -1.02456760e+00
8.70015323e-01 -1.52192444e-01 -1.18729389e+00 3.92299265e-01
8.37740541e-01 -2.19533667e-01 6.44089341e-01 4.15068805e-01
-3.52364451e-01 1.25100434e-01 -8.01159501e-01 6.74517810e-01
8.22349191e-01 -1.05763960e+00 -1.02379382e+00 1.19519544e+00
1.22743869e+00 -3.75032544e-01 -8.79464984e-01 1.77748278e-01
-2.77613848e-01 -8.51155281e-01 9.29757059e-01 -1.16101527e+00
2.43366271e-01 -1.53426126e-01 2.81510204e-01 -1.09976900e+00
-2.68725067e-01 -8.75752807e-01 -6.59370005e-01 6.72580183e-01
8.95921350e-01 -6.64874017e-01 9.64256704e-01 6.14717007e-01
-2.17858061e-01 -8.83534908e-01 -9.27016914e-01 -3.50250334e-01
-1.24466985e-01 -1.52629077e-01 3.57101202e-01 7.26149857e-01
2.12552458e-01 1.27551448e+00 3.98018748e-01 3.65032069e-02
5.28334498e-01 2.67523348e-01 7.22976685e-01 -1.10176170e+00
-3.70450020e-01 -4.86625880e-01 -5.02336025e-02 -1.37545586e+00
1.98751673e-01 -1.11292899e+00 -8.40777829e-02 -2.45075512e+00
1.03731016e-02 1.87410653e-01 -8.93709734e-02 1.08390369e-01
-1.03084433e+00 -1.51831686e-01 -1.16501719e-01 -2.00677309e-02
-1.48298776e+00 5.40860534e-01 1.52993560e+00 -4.81085718e-01
6.77735209e-02 4.46553193e-02 -5.48444271e-01 2.76214451e-01
6.24626577e-01 -5.50308049e-01 -6.38960779e-01 -7.72351027e-01
9.15187359e-01 4.43557829e-01 1.50888801e-01 -4.43273515e-01
8.03246260e-01 1.50809005e-01 -1.36805907e-01 -6.58977747e-01
1.50818184e-01 -1.73571765e-01 -9.34839904e-01 2.50084072e-01
-7.57320285e-01 4.87444878e-01 -3.86722326e-01 7.24104106e-01
-5.84193707e-01 -6.88436687e-01 3.05571523e-03 -1.90765798e-01
-2.32322499e-01 3.81146938e-01 -1.13492660e-01 8.77825320e-01
2.47959986e-01 1.28216609e-01 -9.47960258e-01 -1.15517879e+00
-5.69229722e-01 9.43266571e-01 -2.28840500e-01 2.93093681e-01
1.05100679e+00 -9.97620821e-01 -1.02312100e+00 -3.69401962e-01
4.64144677e-01 4.30578273e-03 3.17073971e-01 7.86505759e-01
-6.91934645e-01 4.26747620e-01 7.74958134e-01 -4.55513537e-01
-1.26183927e+00 2.95122176e-01 3.98764968e-01 -9.23759520e-01
-2.50130519e-02 1.29334915e+00 -4.89673793e-01 -9.43084538e-01
1.00399189e-01 -1.58415198e-01 -7.30592728e-01 5.31381331e-02
2.45087683e-01 2.08248258e-01 3.98808211e-01 -9.42437872e-02
-1.34990275e-01 2.80343205e-01 -6.07171133e-02 8.79126638e-02
7.48515844e-01 -5.63103072e-02 -7.26501524e-01 2.89677307e-02
1.21444142e+00 -4.89303544e-02 -5.28019816e-02 -5.31365156e-01
5.07107198e-01 -1.50408000e-01 -8.21080357e-02 -6.19105637e-01
-6.35832250e-01 9.63985085e-01 6.18307628e-02 9.01780069e-01
8.79105985e-01 1.84348956e-01 1.04266870e+00 1.38578558e+00
-1.60348136e-02 -1.29858243e+00 6.59705997e-01 9.81991529e-01
1.00993407e+00 -1.37378812e+00 -4.95586544e-02 -5.61331272e-01
-5.03779292e-01 1.16427600e+00 8.18911433e-01 -1.35257825e-01
3.40309203e-01 -8.12414050e-01 2.35346749e-01 -9.77545857e-01
-1.19155145e+00 -4.49554294e-01 9.66568172e-01 2.78623164e-01
5.58222294e-01 -2.48523295e-01 -2.50459015e-01 6.73047081e-02
-1.59443140e-01 -3.86182070e-01 2.49941707e-01 9.28777039e-01
-6.78642333e-01 -1.08428514e+00 -4.73242402e-02 9.66216266e-01
-3.30146998e-01 -6.17003500e-01 -1.07930124e+00 2.97005653e-01
-8.81137908e-01 1.65265107e+00 -2.19741136e-01 -4.34226364e-01
6.58532500e-01 3.94848645e-01 5.06615222e-01 -9.55885947e-01
-1.39701355e+00 -7.71064341e-01 9.08465624e-01 -5.69625676e-01
7.12274015e-02 -1.65390465e-02 -1.15492404e+00 1.31342873e-01
-7.67211139e-01 7.15101004e-01 3.19427162e-01 8.82120132e-01
7.48979926e-01 5.26460886e-01 5.53982139e-01 3.72457415e-01
-8.83341134e-01 -9.64523435e-01 7.26498291e-02 6.67468667e-01
6.52056336e-02 2.25319102e-01 -4.24540132e-01 -3.43988627e-01] | [10.967705726623535, 8.011178970336914] |
6f3e3820-5b7f-4aee-99a5-48faa1fb78f9 | on-contrastive-learning-of-semantic | 2305.03843 | null | https://arxiv.org/abs/2305.03843v1 | https://arxiv.org/pdf/2305.03843v1.pdf | On Contrastive Learning of Semantic Similarity forCode to Code Search | This paper introduces a novel code-to-code search technique that enhances the performance of Large Language Models (LLMs) by including both static and dynamic features as well as utilizing both similar and dissimilar examples during training. We present the first-ever code search method that encodes dynamic runtime information during training without the need to execute either the corpus under search or the search query at inference time and the first code search technique that trains on both positive and negative reference samples. To validate the efficacy of our approach, we perform a set of studies demonstrating the capability of enhanced LLMs to perform cross-language code-to-code search. Our evaluation demonstrates that the effectiveness of our approach is consistent across various model architectures and programming languages. We outperform the state-of-the-art cross-language search tool by up to 44.7\%. Moreover, our ablation studies reveal that even a single positive and negative reference sample in the training process results in substantial performance improvements demonstrating both similar and dissimilar references are important parts of code search. Importantly, we show that enhanced well-crafted, fine-tuned models consistently outperform enhanced larger modern LLMs without fine tuning, even when enhancing the largest available LLMs highlighting the importance for open-sourced models. To ensure the reproducibility and extensibility of our research, we present an open-sourced implementation of our tool and training procedures called Cosco. | ['Gail Kaiser', 'Saikat Chakraborty', 'Anthony Saieva'] | 2023-05-05 | null | null | null | null | ['code-search', 'code-search', 'semantic-textual-similarity', 'semantic-similarity'] | ['computer-code', 'computer-vision', 'natural-language-processing', 'natural-language-processing'] | [-1.17453650e-01 -4.44397748e-01 -6.15238667e-01 -1.32773623e-01
-1.26099825e+00 -8.90054286e-01 8.07053447e-01 1.97791994e-01
-5.03146529e-01 1.93390131e-01 -9.00125876e-02 -9.20859694e-01
-2.77298521e-02 -3.54973465e-01 -8.58911633e-01 1.11451536e-01
-1.16020240e-01 2.79749811e-01 4.51753259e-01 -2.69685745e-01
4.59777504e-01 2.22763821e-01 -1.67470109e+00 3.69328946e-01
1.07364678e+00 5.22773445e-01 5.84159851e-01 8.56375575e-01
-2.64067531e-01 6.02572918e-01 -5.34999430e-01 -4.62103903e-01
2.48137400e-01 -2.62932889e-02 -6.97444618e-01 -7.29057312e-01
6.41474843e-01 -1.30546987e-02 2.31838003e-01 8.74340296e-01
5.58314204e-01 -1.22094736e-01 1.53326958e-01 -1.00149965e+00
-5.08330643e-01 6.52313530e-01 -6.17878616e-01 3.75301778e-01
4.49284643e-01 3.88211399e-01 1.13803148e+00 -9.42280054e-01
6.59044802e-01 7.12434292e-01 9.91376638e-01 5.26707768e-01
-1.47710145e+00 -8.57228696e-01 -1.23732641e-01 -3.74715775e-01
-1.47075236e+00 -7.63006806e-01 3.85456741e-01 -7.08366692e-01
1.83989370e+00 3.63616973e-01 3.49624783e-01 1.02278745e+00
1.40813425e-01 5.65023184e-01 7.95613766e-01 -9.11004424e-01
-8.92883986e-02 5.19513369e-01 4.11179848e-02 1.07002938e+00
1.07577257e-01 4.10589129e-01 -3.75045329e-01 -7.53416896e-01
1.47699371e-01 -3.60421896e-01 -8.76088291e-02 -4.72220629e-01
-1.00978756e+00 7.70107508e-01 -9.39718559e-02 6.66114688e-01
1.82855606e-01 4.67051655e-01 7.30476677e-01 2.44941950e-01
1.36027157e-01 9.24279392e-01 -7.47796118e-01 -6.75770462e-01
-1.62643290e+00 9.67215002e-02 1.02018523e+00 1.12718189e+00
7.53233552e-01 -8.36758967e-03 -1.16051398e-01 9.76084411e-01
2.29095891e-01 5.65111697e-01 8.71528924e-01 -6.69838905e-01
6.26972258e-01 8.22872639e-01 -3.51549894e-01 -8.33291888e-01
-1.04131885e-01 -9.39543903e-01 -3.14250141e-02 1.62142232e-01
-8.41305032e-02 1.14579484e-01 -4.43983078e-01 1.86293066e+00
-1.25717431e-01 -1.66388884e-01 -1.03403501e-01 1.86138824e-01
4.88663316e-01 2.07606941e-01 2.29337782e-01 2.19384387e-01
1.00607944e+00 -1.19682217e+00 1.02920972e-01 -5.99742651e-01
1.25072205e+00 -1.02123368e+00 1.48038721e+00 1.18954696e-01
-1.21360457e+00 -5.90035081e-01 -1.06506789e+00 9.93903130e-02
-3.02616328e-01 4.65324163e-01 6.85585141e-01 7.43748188e-01
-1.35276783e+00 3.70060325e-01 -9.07942832e-01 -4.25134957e-01
7.36571550e-02 4.65524584e-01 -2.90056556e-01 1.49163127e-01
-7.52068579e-01 9.99318719e-01 1.54682532e-01 -5.76970518e-01
-9.31508780e-01 -9.59035397e-01 -7.98156798e-01 1.31251454e-01
2.27415070e-01 -4.86304611e-01 1.43112040e+00 -9.14232910e-01
-8.05702806e-01 1.18813622e+00 -4.18149561e-01 -3.69350731e-01
4.55244571e-01 -2.74592012e-01 -5.91961324e-01 -3.30897063e-01
4.04074073e-01 3.74840468e-01 5.28689861e-01 -1.18710899e+00
-4.53036010e-01 2.60333627e-01 2.16778412e-01 -3.41211140e-01
-4.51878250e-01 3.37593853e-01 -9.16556060e-01 -5.25708199e-01
-6.68851078e-01 -9.59306359e-01 8.70412663e-02 -2.99257874e-01
-1.38791427e-01 6.56818748e-02 3.76501888e-01 -5.71494877e-01
2.01340103e+00 -2.40470576e+00 -1.37268707e-01 3.45247418e-01
6.17730021e-02 4.16215122e-01 -4.87383217e-01 6.58340514e-01
-1.33587301e-01 5.78329802e-01 -2.85566986e-01 -4.16779339e-01
3.19455713e-01 -4.96071801e-02 -2.97423601e-01 2.22335517e-01
6.05491959e-02 1.12734330e+00 -9.47743893e-01 -7.37097383e-01
-7.30701238e-02 9.39751193e-02 -1.02161968e+00 6.80385306e-02
-3.41114134e-01 -1.47068769e-01 -3.33064944e-01 6.01576626e-01
1.60783514e-01 -5.96018851e-01 1.36309803e-01 1.76551253e-01
-3.60870630e-01 4.75565851e-01 -7.32632339e-01 2.06353664e+00
-1.26956010e+00 6.00942373e-01 6.34777918e-02 -3.56413990e-01
5.85234463e-01 4.50634249e-02 2.00507976e-02 -9.64490056e-01
-3.74313086e-01 7.66332448e-01 9.73030180e-02 -4.68526155e-01
4.96374011e-01 3.02189797e-01 -1.38498738e-01 7.33687639e-01
2.12367903e-02 -1.96617857e-01 4.83679712e-01 3.81241769e-01
1.44966412e+00 3.24348480e-01 2.80609667e-01 -5.37675798e-01
7.29410172e-01 2.12748215e-01 2.34544016e-02 1.10929096e+00
1.85292765e-01 1.33373365e-01 2.35889211e-01 -2.59125773e-02
-9.55097377e-01 -5.96652985e-01 -2.81299889e-01 1.64712405e+00
-2.99712092e-01 -9.92489219e-01 -6.29320025e-01 -7.95610547e-01
1.15091592e-01 9.99047458e-01 -6.07380807e-01 -2.14305684e-01
-6.20950222e-01 -4.15968478e-01 1.11476171e+00 4.96477783e-01
2.54836917e-01 -7.95176327e-01 -8.76703680e-01 9.36655775e-02
1.48629680e-01 -7.70482421e-01 -8.26888561e-01 3.63933355e-01
-6.45263016e-01 -1.12443912e+00 -2.33043462e-01 -8.70621324e-01
5.04383862e-01 -1.06068827e-01 1.71454382e+00 7.95687854e-01
-3.93277913e-01 5.22646308e-01 -1.72408834e-01 1.31358072e-01
-1.10354841e+00 6.47106767e-01 -3.22342902e-01 -9.52771485e-01
2.30062574e-01 -5.52597940e-01 -2.51060069e-01 1.83499768e-01
-7.27564991e-01 5.73174469e-02 6.95112765e-01 9.45586860e-01
6.97086081e-02 -2.77985930e-01 2.62190521e-01 -8.09625983e-01
9.10380960e-01 -4.86894816e-01 -9.04776216e-01 6.63507521e-01
-1.12885308e+00 5.94831824e-01 6.32801831e-01 -2.79148310e-01
-7.11877704e-01 -1.48619518e-01 -2.68644452e-01 -2.40747690e-01
1.86105907e-01 8.14091563e-01 4.57401574e-01 -4.31244999e-01
1.05553508e+00 4.80006367e-01 -1.25931054e-01 -5.69411218e-01
3.27707350e-01 5.26847839e-01 4.65676725e-01 -1.09062350e+00
8.46356273e-01 4.30451222e-02 -3.58018488e-01 -1.72056049e-01
-2.41605282e-01 -6.71854317e-01 -3.44356865e-01 3.33843678e-01
3.00191104e-01 -8.51992071e-01 -5.33907115e-01 6.31416216e-02
-1.00117636e+00 -6.37441039e-01 1.89517718e-02 4.40783128e-02
-2.81379610e-01 4.19722438e-01 -4.42238897e-01 -6.51982129e-01
-5.31416893e-01 -1.58660102e+00 1.29643011e+00 -2.37826362e-01
-6.16016865e-01 -1.26064265e+00 3.80555004e-01 5.57268076e-02
8.89271796e-01 -3.33454639e-01 1.33027148e+00 -1.03460276e+00
-5.60330093e-01 -2.83772588e-01 3.45809013e-02 1.41620532e-01
-1.26982972e-01 3.21825236e-01 -8.21925938e-01 -5.04891038e-01
-4.08891290e-01 -4.92197335e-01 7.15175986e-01 -2.75772631e-01
9.67812121e-01 -3.46137062e-02 -6.77336216e-01 7.77674258e-01
1.72437370e+00 -1.15552232e-01 2.98592448e-02 6.99400306e-01
3.14395189e-01 1.60978995e-02 2.22681358e-01 1.80621773e-01
1.55719727e-01 8.48866105e-01 2.77381748e-01 2.01790091e-02
-5.88027015e-02 -3.12738180e-01 4.46624309e-01 9.39637959e-01
4.85269129e-01 1.12967685e-01 -1.35401547e+00 8.15834820e-01
-1.50972521e+00 -8.03689182e-01 1.98538527e-01 2.40022373e+00
1.32940388e+00 3.49007100e-01 -1.65842965e-01 -3.31796497e-01
1.57985941e-01 -9.05599594e-02 -3.22836220e-01 -6.34283960e-01
1.25585690e-01 7.66373694e-01 4.42410439e-01 7.33939171e-01
-8.19951296e-01 8.61860633e-01 6.69579792e+00 1.21595132e+00
-1.21872866e+00 2.54063040e-01 1.08228689e-02 -1.78645238e-01
-6.34977341e-01 2.03627974e-01 -9.41584587e-01 4.54340607e-01
1.17770565e+00 -4.56066966e-01 7.74349689e-01 1.38313794e+00
-1.15601756e-01 -2.06695214e-01 -1.38401234e+00 7.42880642e-01
8.30568001e-02 -1.47296619e+00 -3.76447201e-01 -1.01537868e-01
7.46311724e-01 6.35368466e-01 -2.19999403e-02 8.55152786e-01
5.85351646e-01 -1.04185307e+00 9.89122152e-01 2.20375657e-01
1.12398374e+00 -3.23280513e-01 3.79606485e-01 5.15229642e-01
-1.34652090e+00 -1.41411483e-01 2.07834914e-01 1.75487295e-01
-2.46718973e-01 2.79930800e-01 -8.50976586e-01 3.73002499e-01
4.32803959e-01 5.28647840e-01 -1.28416228e+00 1.02228808e+00
9.08851996e-02 6.99509621e-01 -3.64903003e-01 -2.57014558e-02
2.59732217e-01 4.43524241e-01 2.64212400e-01 1.87910199e+00
3.46625090e-01 -7.35764027e-01 2.37524942e-01 1.30949044e+00
-6.01359783e-03 2.02802882e-01 -4.49254125e-01 -2.21076369e-01
6.31223679e-01 1.06302691e+00 -3.23496491e-01 -3.98692340e-01
-6.66614711e-01 7.57775307e-01 6.05417192e-01 2.36327365e-01
-9.30829227e-01 -4.08160657e-01 4.66517389e-01 -2.87768655e-02
3.65909785e-01 -2.33007550e-01 -2.01833233e-01 -1.09551835e+00
2.46524274e-01 -1.38050115e+00 2.99629271e-01 -5.19637227e-01
-9.52063918e-01 7.01626062e-01 2.77010631e-02 -8.32089484e-01
-5.89683414e-01 -3.14379066e-01 -5.52812696e-01 1.34746981e+00
-1.45031679e+00 -1.04071522e+00 -8.92144721e-03 1.99045032e-01
5.31512260e-01 -3.86051118e-01 9.57447648e-01 5.21911740e-01
-3.13875973e-01 1.08274472e+00 2.30510324e-01 5.16866818e-02
6.82893634e-01 -1.07403743e+00 8.80631447e-01 9.15398121e-01
1.81163132e-01 1.65026033e+00 4.20209557e-01 -7.14189649e-01
-1.46885920e+00 -6.53628767e-01 9.14165676e-01 -7.82120347e-01
1.04517365e+00 -5.25981486e-01 -7.94682503e-01 5.74724317e-01
2.08300769e-01 -7.83127993e-02 5.04949033e-01 3.52144390e-01
-6.79126918e-01 8.70535374e-02 -6.25902653e-01 4.42346007e-01
1.04050493e+00 -1.09720004e+00 -5.36228955e-01 1.50189221e-01
6.37752533e-01 -5.01390338e-01 -7.47357607e-01 4.33162153e-01
7.48970211e-01 -1.05985534e+00 9.05282676e-01 -5.31858325e-01
5.79008400e-01 -1.37384564e-01 -3.70318383e-01 -9.26989794e-01
-1.52754232e-01 -5.25930464e-01 -7.55707175e-02 1.21667182e+00
9.23617959e-01 -5.98993242e-01 3.50116849e-01 6.35865271e-01
-2.13439494e-01 -9.39185321e-01 -6.44513965e-01 -9.27640080e-01
2.31489301e-01 -7.24556565e-01 4.40501928e-01 8.40472341e-01
1.53121084e-01 1.05777718e-01 8.37251395e-02 -7.78350383e-02
1.79692183e-03 4.23198938e-01 8.66575897e-01 -8.30460012e-01
-1.04951382e+00 -8.53805840e-01 7.99290910e-02 -1.06711030e+00
4.48627651e-01 -1.38594782e+00 -1.63302809e-01 -1.08716440e+00
6.06048048e-01 -8.94562960e-01 -1.31961644e-01 6.18030965e-01
-2.02112630e-01 -7.53680691e-02 1.21505342e-01 4.43359882e-01
-7.00567603e-01 7.09614456e-02 4.31908607e-01 -9.89340544e-02
-4.28253375e-02 -1.31472304e-01 -6.64232850e-01 3.81704837e-01
3.86173069e-01 -5.21664917e-01 -5.19021809e-01 -5.95998764e-01
6.33772910e-01 -1.65836453e-01 4.07245189e-01 -1.03124690e+00
3.64804953e-01 1.38428792e-01 -1.86459437e-01 -8.05983767e-02
-1.60860270e-01 -7.28900075e-01 2.38214970e-01 6.70177937e-01
-6.32558823e-01 5.79092145e-01 9.23244298e-01 2.20365390e-01
4.32525426e-02 -7.38936901e-01 4.55001295e-01 -3.61287922e-01
-9.02097642e-01 -1.43070996e-01 -3.47027415e-03 4.89837438e-01
7.34303236e-01 -5.00718020e-02 -4.57514793e-01 8.97107646e-02
-2.34250650e-01 1.51109397e-01 1.01601040e+00 5.32518506e-01
-8.75760242e-02 -8.43102813e-01 -3.16385359e-01 3.37798208e-01
4.79865611e-01 -6.44556284e-01 -3.13818663e-01 8.81719649e-01
-4.57809508e-01 7.15731919e-01 3.29437405e-01 -6.96427405e-01
-1.33595967e+00 6.69305265e-01 5.07260442e-01 -5.90496123e-01
-1.70605347e-01 8.69444907e-01 -1.05283469e-01 -7.93942630e-01
1.27659604e-01 -3.05176765e-01 4.82100904e-01 -5.16711235e-01
1.77579716e-01 -4.59085568e-04 5.35738707e-01 -3.61511827e-01
-6.58302367e-01 5.51394701e-01 2.24387273e-02 -6.60004690e-02
1.06673920e+00 2.34742433e-01 -2.58135825e-01 4.11993265e-01
1.39920795e+00 9.07576442e-01 -7.22286582e-01 -4.53296676e-02
2.11688146e-01 -4.80871677e-01 -9.71129760e-02 -1.29262388e+00
-7.24685967e-01 6.98023140e-01 4.95826811e-01 -4.86341044e-02
8.97078335e-01 1.69160385e-02 4.58066553e-01 3.63701940e-01
6.94811106e-01 -7.62971044e-01 -1.02599815e-01 3.91461641e-01
6.13046885e-01 -1.21666551e+00 2.67030927e-03 -1.00191664e-02
-2.37613827e-01 9.55059528e-01 6.98710024e-01 2.54277557e-01
4.91463274e-01 7.83362746e-01 -6.67346790e-02 -2.81384677e-01
-1.09385371e+00 6.29249513e-02 4.42945093e-01 1.62068710e-01
9.27656710e-01 -3.82057339e-01 -1.58731982e-01 3.21154892e-01
-2.44396582e-01 3.56363654e-02 -6.82557821e-02 1.19435203e+00
-3.94848064e-02 -1.50900447e+00 -3.84621210e-02 5.10776639e-01
-5.33324957e-01 -9.17152464e-01 -1.78135410e-01 1.11043823e+00
-7.82720894e-02 7.41476834e-01 -3.53361189e-01 -4.68548536e-01
2.15571821e-01 4.36646521e-01 3.29523325e-01 -8.73044908e-01
-1.21248710e+00 -3.47590387e-01 1.53150260e-01 -7.46243834e-01
-6.37675598e-02 -4.50658858e-01 -1.05490053e+00 -1.73115715e-01
-7.16320455e-01 3.11531156e-01 7.98274696e-01 7.68254101e-01
8.28838110e-01 3.67963135e-01 1.48471028e-01 -4.20526743e-01
-8.49986136e-01 -6.65685296e-01 1.29578918e-01 3.22624892e-01
3.79251868e-01 -5.47294855e-01 -4.62200135e-01 3.82931381e-02] | [7.625427722930908, 7.989203929901123] |
dbfeda6b-8ded-4c11-88b8-fb544c9c5551 | learning-to-learn-better-for-video-object | 2212.02112 | null | https://arxiv.org/abs/2212.02112v1 | https://arxiv.org/pdf/2212.02112v1.pdf | Learning to Learn Better for Video Object Segmentation | Recently, the joint learning framework (JOINT) integrates matching based transductive reasoning and online inductive learning to achieve accurate and robust semi-supervised video object segmentation (SVOS). However, using the mask embedding as the label to guide the generation of target features in the two branches may result in inadequate target representation and degrade the performance. Besides, how to reasonably fuse the target features in the two different branches rather than simply adding them together to avoid the adverse effect of one dominant branch has not been investigated. In this paper, we propose a novel framework that emphasizes Learning to Learn Better (LLB) target features for SVOS, termed LLB, where we design the discriminative label generation module (DLGM) and the adaptive fusion module to address these issues. Technically, the DLGM takes the background-filtered frame instead of the target mask as input and adopts a lightweight encoder to generate the target features, which serves as the label of the online few-shot learner and the value of the decoder in the transformer to guide the two branches to learn more discriminative target representation. The adaptive fusion module maintains a learnable gate for each branch, which reweighs the element-wise feature representation and allows an adaptive amount of target information in each branch flowing to the fused target feature, thus preventing one branch from being dominant and making the target feature more robust to distractor. Extensive experiments on public benchmarks show that our proposed LLB method achieves state-of-the-art performance. | ['DaCheng Tao', 'Lefei Zhang', 'Jing Zhang', 'Meng Lan'] | 2022-12-05 | null | null | null | null | ['semi-supervised-video-object-segmentation', 'video-object-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.99715897e-01 2.12505013e-01 -5.62887371e-01 -4.33946520e-01
-1.01507926e+00 -3.02270532e-01 4.41927612e-01 -3.94086093e-02
-2.93769777e-01 5.41868985e-01 -3.74939787e-04 3.59777249e-02
3.76730934e-02 -7.77395308e-01 -8.34156871e-01 -1.12092388e+00
2.53664792e-01 2.62642890e-01 6.81976020e-01 1.45772070e-01
-4.06047031e-02 1.34832665e-01 -1.51078486e+00 4.86297667e-01
9.96463895e-01 1.36941898e+00 3.41215193e-01 2.55987376e-01
-3.48797619e-01 1.14835894e+00 -4.55143899e-01 -3.82921219e-01
4.42827940e-01 -5.75013459e-01 -6.33329868e-01 2.41994157e-01
2.93796122e-01 -6.14421666e-01 -4.07772899e-01 1.17448568e+00
4.65634584e-01 1.52608871e-01 5.73776424e-01 -1.28733432e+00
-4.38355833e-01 7.84853458e-01 -6.18781328e-01 2.20892876e-01
1.53049305e-01 4.40139681e-01 1.09658992e+00 -7.74089992e-01
4.74905282e-01 1.29124224e+00 2.94121683e-01 5.43235362e-01
-9.92280781e-01 -7.92915523e-01 4.88901407e-01 3.76445472e-01
-1.38438857e+00 -4.92460191e-01 9.24361885e-01 -2.59391487e-01
4.56146389e-01 9.22930539e-02 7.98452079e-01 8.95754039e-01
1.10353149e-01 1.30510747e+00 1.00030255e+00 -2.96741784e-01
2.83529341e-01 2.56812274e-01 8.93833712e-02 1.06880844e+00
5.78532666e-02 1.74797729e-01 -5.41701853e-01 9.55922306e-02
6.14843071e-01 1.57794029e-01 -3.02190870e-01 -3.19630474e-01
-1.00208175e+00 7.65583456e-01 5.32610476e-01 1.90473661e-01
-4.11611348e-01 1.39556989e-01 2.61894286e-01 1.89346761e-01
1.05627142e-01 -4.96699587e-02 -2.96103626e-01 1.33301198e-01
-9.75424051e-01 -1.92054156e-02 5.66038609e-01 1.04281986e+00
1.20687759e+00 -2.43983138e-02 -7.25955009e-01 7.42830336e-01
6.07887983e-01 3.00799817e-01 4.68160063e-01 -9.76675391e-01
4.60601091e-01 9.10130084e-01 -6.62129465e-03 -7.61070251e-01
6.23088591e-02 -4.67166394e-01 -4.25833166e-01 6.97071850e-02
4.97884959e-01 -1.21441642e-02 -1.12311196e+00 1.78382003e+00
5.56718290e-01 5.76200843e-01 5.52210445e-03 1.02502716e+00
7.87505627e-01 9.19745982e-01 9.87079889e-02 -3.31068069e-01
1.05630708e+00 -1.38911247e+00 -5.07374465e-01 -3.66556972e-01
8.12439382e-01 -6.55214369e-01 6.19791508e-01 2.20496133e-01
-9.43534315e-01 -7.15879083e-01 -1.17062485e+00 -3.83508056e-02
-4.51624580e-02 1.40567631e-01 3.78155380e-01 1.88760534e-01
-7.24178731e-01 4.26964402e-01 -7.24373281e-01 -3.44610289e-02
7.01267838e-01 3.41752023e-01 -7.91779459e-02 -3.89149398e-01
-1.16698647e+00 6.93202376e-01 4.84001994e-01 1.86232887e-02
-1.32903910e+00 -5.57510376e-01 -8.21347117e-01 -1.01871558e-01
7.52414584e-01 -4.33614522e-01 1.00086224e+00 -1.35070956e+00
-1.48671162e+00 6.22802079e-01 -1.44853637e-01 -3.02831084e-01
4.60264355e-01 3.10242046e-02 -2.20080912e-01 2.19723061e-01
2.02488095e-01 1.10259044e+00 1.05964386e+00 -1.28514493e+00
-1.10827804e+00 -2.21602306e-01 2.30306938e-01 3.44606936e-01
-1.85127079e-01 -3.02470028e-01 -7.15509593e-01 -5.44393361e-01
1.39757931e-01 -6.70094371e-01 -7.61359930e-02 1.27031505e-01
-1.87213331e-01 -4.65520501e-01 8.86320591e-01 -4.83546585e-01
1.24100614e+00 -2.34337187e+00 2.39450917e-01 9.75874998e-03
9.92889032e-02 3.23546439e-01 -2.74047434e-01 -6.68049678e-02
1.58859685e-01 -2.49583498e-01 -1.00106098e-01 -2.20382735e-01
-2.45175287e-01 4.41759765e-01 -1.84237048e-01 5.53043962e-01
4.25869167e-01 9.30641472e-01 -1.21784091e+00 -8.96422625e-01
2.83465296e-01 1.42475456e-01 -5.15517473e-01 4.43983853e-01
-2.36582473e-01 2.71883845e-01 -6.71003938e-01 8.77917409e-01
5.86195171e-01 -1.26858771e-01 -9.63630825e-02 -5.34592509e-01
2.46020369e-02 1.27153262e-01 -1.23505592e+00 1.64515662e+00
-1.40301093e-01 1.93939805e-01 -3.33419838e-03 -1.07619774e+00
1.04408205e+00 1.36988103e-01 4.66066509e-01 -7.59493470e-01
3.25613618e-01 2.26337686e-01 -7.01550171e-02 -5.47377050e-01
-1.15946591e-01 -1.39041662e-01 -3.55715826e-02 1.87066495e-01
4.65064377e-01 2.11260051e-01 1.41022474e-01 2.25665942e-01
9.61845815e-01 3.51759583e-01 2.29656249e-01 6.44839695e-03
8.30801606e-01 -4.78023402e-02 1.11310613e+00 7.31646478e-01
-5.32131016e-01 3.51843745e-01 3.66786838e-01 -3.66396934e-01
-6.74680471e-01 -1.12285781e+00 2.52368450e-01 1.12897682e+00
6.01276457e-01 -1.61610141e-01 -6.40247881e-01 -1.25115955e+00
-5.16195260e-02 7.24460840e-01 -5.55703878e-01 -6.90997422e-01
-4.89810437e-01 -4.25665408e-01 3.74634475e-01 6.17423594e-01
7.44848490e-01 -8.95817101e-01 -4.45459217e-01 2.16642663e-01
-3.59893739e-01 -8.64730656e-01 -6.14598989e-01 4.07409191e-01
-6.65282249e-01 -1.00255394e+00 -4.28491861e-01 -7.68174708e-01
7.72272825e-01 3.55763227e-01 6.44206464e-01 1.19333640e-01
3.82624678e-02 1.14807360e-01 -5.58799326e-01 -2.12178946e-01
-4.77648765e-01 -1.71734676e-01 -2.68564790e-01 5.44599175e-01
4.40226793e-01 -2.02656314e-01 -5.74101806e-01 3.95712465e-01
-8.82246912e-01 1.32894933e-01 6.51015520e-01 9.27772105e-01
6.43738389e-01 -1.27579018e-01 6.02403820e-01 -5.64436853e-01
-1.23910002e-01 -6.20572507e-01 -6.37908757e-01 4.10217047e-01
-5.26853621e-01 1.01345994e-01 7.05512583e-01 -5.37210405e-01
-1.06245875e+00 2.55628437e-01 6.49574548e-02 -6.58673465e-01
1.44974217e-01 1.98773235e-01 -4.42595512e-01 -5.59174493e-02
3.63297850e-01 3.24836731e-01 4.35232855e-02 -2.83803999e-01
3.71414453e-01 8.32561672e-01 3.35327625e-01 -4.07351613e-01
7.24725902e-01 4.90276992e-01 -2.54571855e-01 -8.75961855e-02
-1.25964832e+00 -4.65556473e-01 -4.82888818e-01 -6.34587169e-01
8.78895819e-01 -1.12401521e+00 -3.36887658e-01 3.83247286e-01
-9.73482072e-01 -1.96476266e-01 -5.90734661e-01 4.11229104e-01
-6.05974257e-01 1.60540581e-01 -5.31313300e-01 -7.40045846e-01
-1.04292162e-01 -1.66076922e+00 1.09221685e+00 6.46337330e-01
1.20933846e-01 -6.43832564e-01 -3.59513611e-01 4.92205948e-01
5.25846928e-02 -9.29562747e-02 8.41235518e-01 -8.32470238e-01
-9.01643693e-01 -9.69769508e-02 -2.24450141e-01 6.49467051e-01
1.87941939e-01 -8.80512968e-02 -1.06450737e+00 -2.60039300e-01
8.69634449e-02 -4.86993730e-01 1.04024947e+00 1.95762739e-01
8.14345777e-01 -2.34540910e-01 -3.79578680e-01 5.45240760e-01
1.40406597e+00 3.77748609e-01 6.08880818e-01 1.65633246e-01
8.10899436e-01 4.83439893e-01 9.61311340e-01 3.54118407e-01
4.31072056e-01 5.19236207e-01 3.54823291e-01 1.66036576e-01
-4.15802866e-01 -2.87454396e-01 8.57168138e-01 8.40598345e-01
2.39180520e-01 -1.09639592e-01 -4.72496659e-01 5.11619329e-01
-1.89612007e+00 -7.63716638e-01 3.06380272e-01 2.16682053e+00
1.09708989e+00 1.61788881e-01 1.91103760e-02 4.61901277e-02
8.76900673e-01 2.52860397e-01 -5.97750187e-01 -5.73820174e-02
7.58378729e-02 -2.57766992e-01 4.23533738e-01 3.58046144e-01
-9.49625790e-01 9.76307034e-01 5.26945686e+00 1.30192709e+00
-1.26757586e+00 3.51874590e-01 6.89924538e-01 -1.19740292e-01
-2.90724158e-01 3.34170073e-01 -1.02820349e+00 6.91242218e-01
6.13769829e-01 1.22266769e-01 5.42883933e-01 8.13032150e-01
-1.11226644e-02 -2.84617722e-01 -1.14173102e+00 8.44106317e-01
-8.13961681e-03 -1.19072115e+00 2.60796100e-01 -2.18568653e-01
6.84824228e-01 -4.36851522e-03 -2.04770893e-01 5.01608968e-01
2.04402775e-01 -5.20339906e-01 1.07740366e+00 6.89765751e-01
5.06335676e-01 -6.77823424e-01 7.02761769e-01 4.69951838e-01
-1.17933810e+00 -4.62747723e-01 -3.42495382e-01 2.84476340e-01
9.80597734e-02 6.05598032e-01 -7.46512234e-01 5.67991793e-01
5.60609281e-01 7.69393504e-01 -4.80950236e-01 9.81311798e-01
-1.92727864e-01 7.84377337e-01 -2.92252302e-01 2.51744449e-01
4.24465895e-01 -2.07698792e-01 6.79557621e-01 8.82359505e-01
1.61418766e-01 -6.77890405e-02 6.62692130e-01 7.59498060e-01
-8.22926611e-02 7.12046251e-02 -2.69022465e-01 -2.77555306e-02
5.48995793e-01 1.29284155e+00 -8.05908024e-01 -5.92053175e-01
-6.50572002e-01 7.97897637e-01 3.93678308e-01 3.45277220e-01
-1.05486071e+00 -1.08987704e-01 2.06438661e-01 -4.32710275e-02
4.96029735e-01 3.21358770e-01 -1.86149359e-01 -9.74665463e-01
-1.05950385e-01 -8.15613329e-01 5.88444710e-01 -5.80158651e-01
-1.13862264e+00 2.39642426e-01 -2.74062511e-02 -1.36788476e+00
-7.09873140e-02 -2.17575118e-01 -4.99254584e-01 7.16125965e-01
-1.51583803e+00 -1.24486661e+00 -2.86348253e-01 4.99287724e-01
6.87853813e-01 -2.24793613e-01 3.28367025e-01 3.97697449e-01
-7.88286388e-01 6.14910305e-01 -2.48154715e-01 5.04840128e-02
6.48757994e-01 -9.88273799e-01 -4.44524258e-01 8.76852691e-01
-1.99246034e-02 9.99362543e-02 3.93259674e-01 -7.34726489e-01
-1.25998783e+00 -1.29795194e+00 5.19517064e-01 -6.44178689e-02
3.77773941e-01 -1.00949071e-01 -8.64464939e-01 5.54218292e-01
-1.72408283e-01 4.25349861e-01 4.89840329e-01 -4.91560757e-01
-3.70643556e-01 -6.35180295e-01 -1.19355178e+00 4.45023090e-01
8.86057377e-01 -2.80782402e-01 -5.88770807e-01 -1.16473844e-03
8.60410154e-01 -1.88026384e-01 -6.31834626e-01 5.85743546e-01
4.32983100e-01 -8.92636895e-01 6.16811872e-01 -1.93129390e-01
3.27634662e-01 -7.55885899e-01 -2.90734828e-01 -9.98823404e-01
-4.56228405e-01 -3.39190960e-01 -4.04277146e-01 1.49123681e+00
2.29822680e-01 -4.00884926e-01 5.41929603e-01 3.19257468e-01
-3.04783642e-01 -1.28651369e+00 -1.05861890e+00 -5.56063950e-01
-4.40039098e-01 -3.08249831e-01 5.71474791e-01 6.36491776e-01
-3.38249683e-01 2.43965685e-01 -3.14130753e-01 1.36151731e-01
5.53272069e-01 2.36113697e-01 4.70944494e-01 -9.19395864e-01
-2.94224441e-01 -3.10452163e-01 -4.05140936e-01 -1.29523945e+00
1.78399794e-02 -1.09297633e+00 3.36691409e-01 -1.37418652e+00
3.55813891e-01 -6.21131659e-01 -5.97879946e-01 7.40588486e-01
-3.45852524e-01 3.05375546e-01 2.99420893e-01 2.42330700e-01
-9.73200321e-01 6.67983294e-01 1.55410850e+00 -3.84339631e-01
-1.61068946e-01 -9.19311345e-02 -7.31719315e-01 8.31668019e-01
4.12956387e-01 -6.31077349e-01 -4.76074487e-01 -2.73078769e-01
-1.60427809e-01 -7.78000131e-02 3.74099314e-01 -8.79737437e-01
4.58779722e-01 -2.09565207e-01 5.34666955e-01 -5.70173562e-01
1.98623255e-01 -9.86891866e-01 -1.70476079e-01 4.28844750e-01
-2.75983751e-01 -6.30631745e-01 -2.34155133e-01 6.89327776e-01
-1.39265656e-01 -4.30828661e-01 9.58841980e-01 -6.94360211e-02
-8.91435027e-01 3.82988542e-01 -1.04009464e-01 -1.11824580e-01
1.31631303e+00 -4.32123661e-01 -6.18617088e-02 -1.23164795e-01
-5.23343027e-01 6.20349288e-01 3.19352120e-01 4.74243671e-01
6.17152631e-01 -1.41246879e+00 -4.93945211e-01 3.07847291e-01
1.58121824e-01 2.28088275e-01 3.12749565e-01 1.03033650e+00
-1.36131078e-01 1.43621871e-02 -2.80481968e-02 -6.82351053e-01
-7.82451451e-01 6.79300010e-01 4.46910590e-01 -2.43011162e-01
-4.39149588e-01 1.01632380e+00 3.46910745e-01 -3.02087013e-02
4.19117779e-01 -4.26748879e-02 -2.28809074e-01 3.84118944e-01
5.42143583e-01 4.80713665e-01 -1.94593996e-01 -7.44822860e-01
-3.06863517e-01 3.14963818e-01 -3.46432000e-01 2.26041615e-01
1.10177732e+00 -3.62408310e-01 -1.40855357e-01 6.52006269e-01
1.25517392e+00 -2.53569305e-01 -1.66119838e+00 -4.75313842e-01
-1.95448026e-01 -3.58997345e-01 3.20375144e-01 -7.56386459e-01
-1.32495737e+00 5.34484148e-01 6.69770300e-01 -2.25152656e-01
1.32967651e+00 2.23668724e-01 1.04494131e+00 8.70235935e-02
5.22500336e-01 -1.27196705e+00 2.07174629e-01 2.29275733e-01
4.07579243e-01 -1.15316665e+00 -2.55516976e-01 -5.20259559e-01
-6.96487486e-01 9.40029323e-01 9.26118791e-01 -1.60173029e-01
4.67423260e-01 2.93933302e-01 -4.48199660e-02 -1.87788010e-02
-7.72222221e-01 -3.49192500e-01 3.26930791e-01 2.95992047e-01
-1.09902307e-01 -7.88588151e-02 -2.07447335e-01 8.14060211e-01
5.48765004e-01 1.57531962e-01 3.51077430e-02 9.14021552e-01
-6.46427274e-01 -1.10654914e+00 -3.05739462e-01 7.18072236e-01
-2.65963793e-01 9.37795043e-02 -1.82265043e-03 2.93379158e-01
8.04524779e-01 8.26486409e-01 -3.55708674e-02 -6.46642387e-01
6.61414117e-02 1.93651095e-01 5.71271956e-01 -6.18464887e-01
-3.83778960e-01 3.17236930e-01 -1.37314379e-01 -6.78337693e-01
-5.37852705e-01 -6.59996867e-01 -1.52774978e+00 2.68460214e-01
-6.45116627e-01 3.31758596e-02 1.23562261e-01 1.35130894e+00
1.85994327e-01 4.72832024e-01 8.97048652e-01 -7.20894814e-01
-6.42953634e-01 -8.54335368e-01 -4.44753081e-01 3.11525434e-01
2.88578659e-01 -9.63116407e-01 -2.84885705e-01 9.59954411e-02] | [9.31706428527832, 1.3343416452407837] |
37654268-ad1c-45e7-9fd4-dbbd58a63f82 | exploration-of-algorithmic-trading-strategies | 2110.14936 | null | https://arxiv.org/abs/2110.14936v1 | https://arxiv.org/pdf/2110.14936v1.pdf | Exploration of Algorithmic Trading Strategies for the Bitcoin Market | Bitcoin is firmly becoming a mainstream asset in our global society. Its highly volatile nature has traders and speculators flooding into the market to take advantage of its significant price swings in the hope of making money. This work brings an algorithmic trading approach to the Bitcoin market to exploit the variability in its price on a day-to-day basis through the classification of its direction. Building on previous work, in this paper, we utilise both features internal to the Bitcoin network and external features to inform the prediction of various machine learning models. As an empirical test of our models, we evaluate them using a real-world trading strategy on completely unseen data collected throughout the first quarter of 2021. Using only a binary predictor, at the end of our three-month trading period, our models showed an average profit of 86\%, matching the results of the more traditional buy-and-hold strategy. However, after incorporating a risk tolerance score into our trading strategy by utilising the model's prediction confidence scores, our models were 12.5\% more profitable than the simple buy-and-hold strategy. These results indicate the credible potential that machine learning models have in extracting profit from the Bitcoin market and act as a front-runner for further research into real-world Bitcoin trading. | ['Tomas Ward', 'Eoin Brophy', 'Nathan Crone'] | 2021-10-28 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-6.87798321e-01 1.56427249e-01 -6.07727706e-01 -1.10056162e-01
-2.43689984e-01 -9.55777884e-01 9.37244713e-01 -9.50688720e-02
-1.79376453e-01 6.98147595e-01 1.51109695e-01 -8.15025985e-01
-1.79556504e-01 -8.44290197e-01 -2.97747105e-01 -4.44814116e-01
-2.77940869e-01 7.00646698e-01 1.54769579e-02 -3.01633567e-01
5.17930746e-01 2.66093880e-01 -5.69853663e-01 8.73690471e-02
4.07761604e-01 1.72662425e+00 -6.99300349e-01 2.77065873e-01
1.04629025e-01 1.26847041e+00 -5.09852529e-01 -7.64597535e-01
1.14615524e+00 -6.26450777e-01 -3.81217510e-01 -4.06554013e-01
-2.97868103e-01 -3.84381324e-01 -2.56419301e-01 8.34780633e-01
-1.95736717e-02 -4.18832064e-01 2.72798598e-01 -1.24926972e+00
-7.47268051e-02 1.15473092e+00 -2.44487092e-01 4.34417605e-01
-2.27397621e-01 4.90904540e-01 1.84595561e+00 -3.95341426e-01
5.97696424e-01 4.61757660e-01 6.36751413e-01 -2.84461886e-01
-1.31054664e+00 -9.29490566e-01 -4.47517782e-01 6.33744821e-02
-6.08827770e-01 -4.10739571e-01 1.02584660e+00 -5.98130524e-01
1.14042509e+00 3.66836756e-01 1.33105803e+00 5.28634369e-01
6.31271839e-01 3.59884262e-01 1.50548601e+00 -2.35063672e-01
2.07487047e-01 1.52038351e-01 -3.73001218e-01 -1.63079705e-03
4.39753979e-01 9.82316852e-01 -3.39119196e-01 -5.40623605e-01
7.65671849e-01 4.51435708e-02 1.16949379e-01 -3.32034349e-01
-1.38586247e+00 1.14195943e+00 5.74168921e-01 5.21265686e-01
-5.72298825e-01 3.61461550e-01 3.66454244e-01 9.52307522e-01
5.58103144e-01 8.13044131e-01 -8.68029296e-01 -5.54685831e-01
-1.33424711e+00 4.85180408e-01 1.44364011e+00 1.06884092e-01
4.17747825e-01 2.43023083e-01 2.95769930e-01 1.95551276e-01
2.93904811e-01 2.72512555e-01 7.32133150e-01 -1.29768705e+00
3.62689942e-01 4.00563955e-01 1.09542482e-01 -8.50935757e-01
-9.91556495e-02 -6.95892990e-01 -2.84649521e-01 5.21059513e-01
8.50399792e-01 -4.65249896e-01 -3.47362250e-01 1.11872435e+00
-2.42859840e-01 -1.94575176e-01 -1.08910181e-01 7.71354318e-01
-5.04155755e-01 3.38279068e-01 -4.46098953e-01 -5.36804557e-01
8.26672852e-01 -9.90948379e-01 -3.70898128e-01 2.26385266e-01
5.62137604e-01 -8.04224253e-01 2.52613245e-04 8.39928865e-01
-1.04735410e+00 2.25327224e-01 -8.06885898e-01 8.03293884e-01
-2.30626076e-01 -8.16711366e-01 1.08831942e+00 6.42248809e-01
-7.33379483e-01 9.06264901e-01 -6.74819767e-01 2.79491633e-01
2.42497534e-01 2.48812616e-01 1.99722238e-02 8.99037302e-01
-1.21563530e+00 1.03356886e+00 4.63053107e-01 1.16602063e-01
-5.85128665e-01 -5.32690942e-01 -2.10316211e-01 3.38295996e-01
5.67703903e-01 -2.58194268e-01 1.44096017e+00 -1.38898468e+00
-1.52973557e+00 3.40851068e-01 6.25781536e-01 -1.22102726e+00
9.31990564e-01 1.77273855e-01 -5.08699059e-01 -1.14165492e-01
-2.05601886e-01 -4.81981300e-02 6.12457693e-01 -6.61669374e-01
-6.61447048e-01 -1.20751522e-01 -1.84476689e-01 -3.04978192e-01
2.90458560e-01 2.01104462e-01 6.43046975e-01 -1.07421839e+00
2.90151924e-01 -1.00848055e+00 -1.20326579e-01 -5.82220495e-01
-1.50110023e-02 1.06578603e-01 3.41839969e-01 -8.46120954e-01
1.23773503e+00 -1.59574378e+00 -4.21268225e-01 8.02672088e-01
2.97976583e-01 -5.60821332e-02 6.73297584e-01 6.55365050e-01
-2.39047587e-01 4.43189234e-01 4.80130762e-02 5.49366951e-01
2.34714687e-01 -9.14911553e-02 -4.09090370e-01 3.06961566e-01
-9.11565274e-02 1.30809939e+00 -6.11483216e-01 -2.57158317e-02
6.02457859e-02 -3.88363957e-01 -5.86109638e-01 6.68190140e-03
-3.45651597e-01 2.96924144e-01 -1.70725480e-01 8.43831480e-01
4.57103066e-02 -2.02468082e-01 6.10274911e-01 2.49058440e-01
-2.51372904e-01 5.15911758e-01 -6.19202793e-01 8.65280449e-01
-4.45645414e-02 6.89922333e-01 2.42921673e-02 -1.10633266e+00
1.00980473e+00 4.26283956e-01 8.52109194e-01 -1.07106304e+00
2.90908009e-01 1.00159812e+00 8.72535467e-01 3.46356630e-02
9.98351648e-02 -6.76822424e-01 -1.20236702e-01 9.99207199e-01
-2.07329020e-01 -1.99305773e-01 4.94431332e-02 -1.48733467e-01
9.44540143e-01 -1.31654903e-01 3.54762912e-01 -3.88280481e-01
-8.03249702e-02 9.58117619e-02 7.37876892e-01 4.32484031e-01
-4.34540093e-01 7.02917129e-02 9.49514031e-01 -8.75341773e-01
-1.20005894e+00 -8.47579122e-01 -4.02149893e-02 4.54522103e-01
-5.26141405e-01 -1.87128589e-01 -1.39319211e-01 -7.89504528e-01
7.47867405e-01 5.90601921e-01 -4.42018837e-01 1.38455257e-01
-4.89361942e-01 -8.67309749e-01 4.40477371e-01 2.20163703e-01
6.15163922e-01 -1.07716739e+00 -9.89099622e-01 5.45846641e-01
2.43898839e-01 -4.31523710e-01 -2.48374432e-01 5.33883572e-01
-9.20291007e-01 -1.22318077e+00 -5.75437129e-01 -1.14601627e-02
-9.53808576e-02 -4.79084700e-01 1.21927261e+00 -1.86773717e-01
8.43244717e-02 -2.00668335e-01 -1.36075109e-01 -4.86426920e-01
-5.13894022e-01 -7.17633218e-02 -1.42637193e-01 -5.71795478e-02
4.26537186e-01 -8.29592645e-01 -6.93375766e-01 3.80303055e-01
-3.64770770e-01 -3.30976486e-01 5.41926682e-01 9.81653094e-01
-4.80978154e-02 3.44348736e-02 9.72412586e-01 -7.25111783e-01
5.69036603e-01 -7.88336277e-01 -1.18817389e+00 -7.48145729e-02
-1.61611986e+00 7.13739544e-02 1.36069834e-01 -2.08424374e-01
-6.25524521e-01 -5.42976379e-01 5.69986403e-01 -4.50242072e-01
6.98378801e-01 9.52786744e-01 5.29786170e-01 -4.05488238e-02
1.83935210e-01 -1.54559717e-01 3.89054060e-01 -5.95236301e-01
2.66319811e-01 4.78436977e-01 2.28328362e-01 -3.17719907e-01
8.57963204e-01 1.59607172e-01 -1.25195086e-01 3.83132622e-02
-2.85784662e-01 -4.57088947e-02 -2.20748976e-01 -8.35310966e-02
1.95745111e-01 -6.94519520e-01 -1.03322268e+00 6.25211179e-01
-6.10243618e-01 -4.53713775e-01 -5.18527985e-01 7.37100363e-01
-6.38821900e-01 -2.04895213e-01 -7.84467161e-01 -1.02773798e+00
-2.49779090e-01 -6.09075785e-01 -2.64706939e-01 2.31273863e-02
-4.85331267e-01 -1.21380353e+00 4.39879209e-01 6.11617625e-01
9.89208221e-01 5.62916338e-01 6.92441404e-01 -1.52277172e+00
-7.38842785e-01 -2.82424867e-01 -1.49900228e-01 5.98617613e-01
2.01175749e-01 -4.28548940e-02 -5.15183091e-01 -3.11424941e-01
5.21923125e-01 -1.00848265e-01 8.00072193e-01 8.37203041e-02
3.47032160e-01 -9.11582053e-01 1.67351216e-01 4.39205766e-01
1.31591952e+00 5.63046575e-01 2.74417669e-01 8.88834000e-01
1.59553234e-02 4.92854953e-01 4.20257747e-01 5.34664869e-01
2.43590474e-01 4.47029233e-01 5.87106049e-01 2.99631864e-01
4.78954822e-01 -3.22782129e-01 3.50035667e-01 8.27149689e-01
-3.20339322e-01 3.24971169e-01 -1.10362232e+00 5.07989287e-01
-1.74655211e+00 -1.23789525e+00 3.61399919e-01 2.13460350e+00
8.49430978e-01 8.30817938e-01 6.31927669e-01 1.89249828e-01
2.98283815e-01 3.20296496e-01 -6.06904089e-01 -8.20560932e-01
-1.04590252e-01 1.60469517e-01 1.14345455e+00 2.07822323e-01
-6.09303534e-01 5.02576292e-01 6.84880257e+00 2.69959211e-01
-1.53579664e+00 -1.12517051e-01 1.13726377e+00 -3.05973560e-01
-4.50496942e-01 7.01581717e-01 -4.19841707e-02 9.20520604e-01
1.33274126e+00 -7.18621314e-01 1.00582027e+00 6.09887302e-01
2.63006568e-01 -1.47884488e-01 -1.06617486e+00 7.29837060e-01
-3.26716304e-01 -1.58638561e+00 -5.19263685e-01 7.49346256e-01
8.18916321e-01 3.99576575e-01 -4.50578108e-02 1.00379497e-01
6.91831589e-01 -1.14746785e+00 1.06730723e+00 6.86059415e-01
2.55247831e-01 -6.49673223e-01 1.06320608e+00 2.66671449e-01
-5.96978128e-01 -5.05877197e-01 4.25270796e-01 -6.56427920e-01
1.16704755e-01 7.22441733e-01 -9.19969499e-01 5.40230989e-01
5.40303051e-01 5.92369556e-01 -1.77024290e-01 8.42507780e-01
8.43250751e-02 9.62555408e-01 -6.44570649e-01 1.99459702e-01
4.84211415e-01 -5.81374824e-01 4.99545783e-01 6.48544848e-01
3.62122595e-01 -2.20497906e-01 -2.18333945e-01 1.18725276e+00
-2.44539037e-01 -2.03605741e-01 -3.75089198e-01 -6.11774981e-01
4.18354511e-01 9.21121120e-01 -5.30803561e-01 -2.43544295e-01
-4.58875149e-01 3.42556715e-01 -2.92066514e-01 2.14629531e-01
-4.78421152e-01 -3.98227930e-01 4.03380394e-01 3.61248851e-01
5.65387249e-01 -3.04461360e-01 -7.49528885e-01 -1.27161098e+00
1.77514583e-01 -1.12114227e+00 3.16788048e-01 -1.94844961e-01
-1.38257122e+00 -1.76092554e-02 -3.29962283e-01 -9.40479517e-01
-7.94792473e-01 -3.91527742e-01 -9.46325660e-01 9.65774298e-01
-1.48863053e+00 -7.51161277e-01 6.78438663e-01 2.22686127e-01
-5.88924810e-02 -5.92284501e-01 2.92167425e-01 -2.38253877e-01
-7.77796283e-03 1.15687378e-01 5.58922529e-01 5.25222898e-01
5.03083289e-01 -1.26522815e+00 4.36269045e-01 5.14611661e-01
6.00717902e-01 4.66190517e-01 4.73348856e-01 -7.99419105e-01
-1.08806944e+00 -2.89315850e-01 1.02960026e+00 -4.96718138e-01
1.55887580e+00 6.11336865e-02 -3.57304931e-01 1.02750111e+00
4.22756374e-01 2.13907217e-03 5.83273113e-01 5.08237258e-03
-3.94664943e-01 -3.76721442e-01 -1.18531621e+00 -1.24882400e-01
2.90830493e-01 -4.26183283e-01 -8.95459354e-01 -2.73848981e-01
3.62593234e-01 1.68218821e-01 -1.16121864e+00 2.71685392e-01
1.09438837e+00 -1.39876115e+00 4.03166503e-01 -5.71136355e-01
2.20751598e-01 2.80535698e-01 -1.31615505e-01 -1.25616789e+00
-1.10087968e-01 -1.39270115e+00 -3.08073759e-01 9.81090426e-01
5.02994120e-01 -1.40030074e+00 9.64864492e-01 7.78090894e-01
5.52611530e-01 -7.94094443e-01 -1.39748180e+00 -8.58478427e-01
6.12524867e-01 -2.14939713e-01 8.62496138e-01 1.14462590e+00
4.59779978e-01 -2.94749022e-01 -2.86405802e-01 -6.98622108e-01
7.48611450e-01 5.40326416e-01 4.92602289e-01 -1.41341937e+00
-5.47253311e-01 -9.93228912e-01 -4.50905472e-01 -6.00960791e-01
-2.37048507e-01 -1.04040349e+00 -6.90155387e-01 -5.47918081e-01
2.24885046e-02 -7.59159029e-01 -9.14991498e-01 4.48296756e-01
5.95480084e-01 1.58520505e-01 7.83617198e-01 7.35703647e-01
2.35811286e-02 1.85741022e-01 8.21338952e-01 -1.76201731e-01
7.35892635e-03 1.43319234e-01 -6.78512335e-01 6.00894392e-01
1.01354551e+00 -2.95179129e-01 1.51033118e-01 1.81090668e-01
6.04842067e-01 4.33333218e-01 5.14556766e-01 -4.85478520e-01
-1.35593163e-03 -6.63954675e-01 2.16362223e-01 -2.67950773e-01
-5.48469611e-02 -9.92956519e-01 6.10205531e-01 9.64874148e-01
-2.24340320e-01 2.14877948e-01 -3.42384696e-01 2.81400561e-01
-3.97187144e-01 1.21698201e-01 4.21569437e-01 -4.34409916e-01
-6.25489205e-02 -1.09277107e-01 -4.19844799e-02 1.26018643e-01
1.12728143e+00 -7.94512108e-02 -3.06475490e-01 -5.87010562e-01
-8.06715548e-01 4.00397152e-01 6.33185387e-01 3.29929322e-01
1.14243701e-02 -1.25294113e+00 -6.68091655e-01 2.63940454e-01
-4.77052748e-01 -5.94002008e-01 -7.23976254e-01 1.09621525e+00
-7.23731995e-01 8.63276243e-01 -4.07951236e-01 -5.90863377e-02
-5.36383927e-01 3.75336170e-01 7.14911580e-01 -6.72603250e-01
-3.29709768e-01 2.36852810e-01 -5.74474514e-01 3.22956331e-02
-2.24006206e-01 -4.00610298e-01 4.74476129e-01 4.29471970e-01
-7.56026879e-02 5.89787304e-01 -1.17707193e-01 -4.56987560e-01
-1.08089343e-01 -9.68639448e-04 2.21876368e-01 -5.56176841e-01
1.51213169e+00 1.23722538e-01 -4.33747530e-01 8.44261944e-01
8.63687158e-01 2.26488620e-01 -1.12426889e+00 -1.07065193e-01
8.45564663e-01 -6.47813559e-01 -3.25042307e-02 -1.36594772e+00
-1.48245835e+00 4.44314957e-01 1.16919711e-01 8.67104590e-01
4.47740972e-01 -1.27131343e-01 9.17229295e-01 2.09396183e-01
5.62882423e-01 -1.22448945e+00 -3.78175646e-01 1.89361200e-01
6.59073651e-01 -1.14340377e+00 -1.74325407e-02 5.25459170e-01
-4.90695953e-01 1.13831913e+00 -1.95961028e-01 -4.47295696e-01
1.02301407e+00 4.13469188e-02 2.83458233e-01 -7.29987919e-02
-9.70716596e-01 5.41742384e-01 -2.16366410e-01 -1.28347054e-01
1.93466306e-01 5.63937068e-01 -6.22062206e-01 8.62898290e-01
-6.75662816e-01 2.36558348e-01 4.39068615e-01 8.51480246e-01
-3.76276016e-01 -1.33872509e+00 -2.10866675e-01 8.15313220e-01
-9.65896368e-01 -8.98353085e-02 -5.90052366e-01 7.37359047e-01
-6.03800789e-02 7.08838046e-01 1.23204120e-01 -3.55926782e-01
7.11780181e-03 5.00792742e-01 -7.26624653e-02 -1.08127326e-01
-1.22554481e+00 2.51995325e-01 3.61432582e-01 -6.50855482e-01
-2.19098672e-01 -1.00118542e+00 -8.58889759e-01 -7.90626585e-01
-5.36393225e-01 4.48817521e-01 6.31805837e-01 7.18191028e-01
1.21871933e-01 -7.55112693e-02 1.23212469e+00 -5.89736700e-01
-1.36328256e+00 -9.12941396e-01 -9.04038489e-01 2.26697221e-01
4.29532319e-01 -3.15898389e-01 -7.30460405e-01 -2.03132212e-01] | [4.651387691497803, 4.163681983947754] |
5b039777-e807-4d6e-b778-d2d4bbe1fd11 | learning-vector-quantized-shape-code-for | 2012.00985 | null | https://arxiv.org/abs/2012.00985v1 | https://arxiv.org/pdf/2012.00985v1.pdf | Learning Vector Quantized Shape Code for Amodal Blastomere Instance Segmentation | Blastomere instance segmentation is important for analyzing embryos' abnormality. To measure the accurate shapes and sizes of blastomeres, their amodal segmentation is necessary. Amodal instance segmentation aims to recover the complete silhouette of an object even when the object is not fully visible. For each detected object, previous methods directly regress the target mask from input features. However, images of an object under different amounts of occlusion should have the same amodal mask output, which makes it harder to train the regression model. To alleviate the problem, we propose to classify input features into intermediate shape codes and recover complete object shapes from them. First, we pre-train the Vector Quantized Variational Autoencoder (VQ-VAE) model to learn these discrete shape codes from ground truth amodal masks. Then, we incorporate the VQ-VAE model into the amodal instance segmentation pipeline with an additional refinement module. We also detect an occlusion map to integrate occlusion information with a backbone feature. As such, our network faithfully detects bounding boxes of amodal objects. On an internal embryo cell image benchmark, the proposed method outperforms previous state-of-the-art methods. To show generalizability, we show segmentation results on the public KINS natural image benchmark. To examine the learned shape codes and model design choices, we perform ablation studies on a synthetic dataset of simple overlaid shapes. Our method would enable accurate measurement of blastomeres in in vitro fertilization (IVF) clinics, which potentially can increase IVF success rate. | ['Hanspeter Pfister', 'Daniel Needleman', 'Dalit Ben-Yosef', 'James Tompkin', 'Helen Yang', 'Brian Leahy', 'Xingxuan Zhang', 'Donglai Wei', 'Won-Dong Jang'] | 2020-12-02 | null | null | null | null | ['amodal-instance-segmentation'] | ['computer-vision'] | [ 3.00719649e-01 7.35787868e-01 1.57791212e-01 -4.04378027e-01
-6.18120432e-01 -9.08957839e-01 2.75189131e-01 1.67712897e-01
-5.35343550e-02 3.15037102e-01 -2.22499639e-01 -2.76498795e-01
2.28642449e-01 -9.78533745e-01 -1.00684202e+00 -8.10815752e-01
2.88667560e-01 9.92935538e-01 1.00935996e-01 7.34145492e-02
1.67293787e-01 8.15805733e-01 -1.31070709e+00 5.16938329e-01
7.56895840e-01 8.97241533e-01 1.56632692e-01 7.45172679e-01
-2.16147184e-01 -1.20128058e-01 -4.73326057e-01 -6.37242913e-01
5.05383730e-01 -4.10757303e-01 -7.07734227e-01 2.49264151e-01
5.81011295e-01 -2.86031246e-01 4.95831594e-02 8.60321164e-01
3.77932996e-01 -2.45560542e-01 1.25573528e+00 -9.39531684e-01
-4.04506713e-01 3.80095690e-01 -7.07858682e-01 7.06056878e-02
-3.21342498e-02 1.74922407e-01 1.10098255e+00 -9.08459485e-01
8.35721016e-01 1.15695941e+00 5.80177605e-01 6.25625253e-01
-1.60266411e+00 -4.73030210e-01 -3.44878644e-01 -5.92914820e-01
-1.17225015e+00 -2.85656422e-01 4.62811351e-01 -7.29299784e-01
5.37283540e-01 2.11152956e-01 7.92518616e-01 4.47336257e-01
1.86056077e-01 9.89414871e-01 5.14936209e-01 -2.53386468e-01
2.52501786e-01 -2.57647317e-02 -1.84694365e-01 8.47089112e-01
3.32144827e-01 -8.94394815e-02 -9.45373923e-02 -1.43210366e-02
1.16527641e+00 3.33818700e-03 -2.02704653e-01 -5.11658192e-01
-8.87705445e-01 6.93552136e-01 3.63108486e-01 3.51598173e-01
-1.58522367e-01 1.03835717e-01 -4.11668643e-02 -2.90128235e-02
2.57073730e-01 5.73734581e-01 -3.86415869e-01 2.14374810e-01
-1.19119215e+00 1.65984005e-01 6.28772676e-01 7.30099559e-01
1.03391445e+00 -7.39714429e-02 -3.68166417e-01 7.65552342e-01
3.46067250e-01 5.17287672e-01 1.04509860e-01 -1.25203323e+00
2.48615682e-01 9.69615698e-01 -1.99175343e-01 -9.00466263e-01
-4.42738384e-01 -3.59397322e-01 -7.61988819e-01 2.80613065e-01
7.65348673e-01 -9.85377803e-02 -1.54062164e+00 1.31485987e+00
5.48396885e-01 2.91720361e-01 -5.31176031e-02 8.89591396e-01
1.14155793e+00 6.93658888e-01 -4.07782972e-01 1.22565292e-01
1.28735137e+00 -4.04695362e-01 -4.47438776e-01 -2.13780105e-01
6.86879992e-01 -6.06841326e-01 6.05007410e-01 2.35778540e-01
-1.27590013e+00 -3.62352133e-01 -8.87405634e-01 -2.44053170e-01
-1.45058975e-01 4.23256636e-01 5.74998081e-01 6.27018213e-01
-9.41464007e-01 6.71616197e-01 -1.19365215e+00 -1.45546004e-01
9.50246930e-01 5.94651222e-01 -4.33225960e-01 2.52135009e-01
-4.72463012e-01 5.24077952e-01 1.65794179e-01 1.61745086e-01
-1.02448630e+00 -9.13854122e-01 -1.25975049e+00 3.36443961e-01
-1.47775576e-01 -4.39092487e-01 8.83908749e-01 -8.31049144e-01
-1.26218939e+00 1.21501195e+00 -2.35802516e-01 -2.09164008e-01
2.94761091e-01 3.35277498e-01 4.35609341e-01 3.11752021e-01
8.02378803e-02 1.21709716e+00 8.30711246e-01 -1.58693588e+00
-5.94824076e-01 -3.68326038e-01 -1.11507058e-01 -6.62240759e-02
-1.02023505e-01 -4.55397666e-01 -7.87934124e-01 -3.79702479e-01
6.15041256e-01 -7.91912794e-01 -1.22662872e-01 2.12240010e-01
-3.99453521e-01 4.35419194e-02 6.59312069e-01 -4.91371691e-01
7.59330213e-01 -2.18847179e+00 3.48459214e-01 4.15260524e-01
2.32463047e-01 1.29630685e-01 -2.96486497e-01 -8.59104991e-02
-6.83599943e-03 2.94978321e-01 -4.82733220e-01 -5.15126884e-01
-2.37484783e-01 4.04041201e-01 -1.90325633e-01 7.04251289e-01
4.26244974e-01 1.11640549e+00 -5.37901342e-01 -7.55684793e-01
1.64568350e-01 4.59339708e-01 -8.62567306e-01 2.42740750e-01
-1.83854327e-01 5.63198149e-01 -2.46294081e-01 8.66355479e-01
8.26694846e-01 -3.39376003e-01 -3.76397744e-02 -6.14328533e-02
2.13729069e-01 -4.57141191e-01 -1.05880272e+00 1.60435593e+00
-7.91406780e-02 8.83884490e-01 2.34541178e-01 -1.27359509e+00
7.92697966e-01 2.11624965e-01 6.97898149e-01 -3.72890651e-01
2.96758503e-01 1.63413018e-01 1.78225309e-01 -3.41345936e-01
7.08398670e-02 -2.67998219e-01 2.27208912e-01 4.37944159e-02
2.56505311e-01 -5.72281420e-01 3.05535376e-01 1.70845956e-01
8.91099870e-01 1.39494479e-01 -1.38005123e-01 -2.36012980e-01
3.77727449e-01 -1.32378682e-01 9.70494092e-01 6.12686098e-01
2.94996575e-02 1.59624052e+00 9.15854216e-01 -6.48404777e-01
-1.19262087e+00 -1.18770063e+00 -7.01538622e-01 5.75805843e-01
2.75982440e-01 2.48991624e-02 -1.01421881e+00 -6.48487568e-01
2.24423692e-01 3.65119427e-01 -9.67006624e-01 2.70961314e-01
-5.63872397e-01 -7.77568400e-01 6.81710780e-01 5.36668003e-01
-1.07242569e-01 -1.15390885e+00 -7.17457831e-01 -4.83624171e-03
1.61832217e-02 -9.94172633e-01 -3.48211020e-01 1.74979717e-02
-8.59689713e-01 -1.15017581e+00 -1.05329573e+00 -9.45845306e-01
1.23088384e+00 -2.91670620e-01 1.07400894e+00 4.90685374e-01
-5.58270395e-01 2.68874280e-02 -1.81393951e-01 -3.82139236e-01
-5.46212375e-01 -4.75252308e-02 -3.51106465e-01 1.78028300e-01
-6.40289932e-02 -3.77490103e-01 -7.57389724e-01 2.46513769e-01
-1.05049825e+00 1.39405146e-01 5.62639773e-01 1.04935205e+00
9.12125766e-01 -2.99342394e-01 1.10026084e-01 -1.01724076e+00
4.21753759e-03 -3.14382136e-01 -9.26531971e-01 3.40101004e-01
-2.81965613e-01 1.20916583e-01 3.85658026e-01 -2.94704586e-01
-8.33717346e-01 3.15077603e-01 -2.33779982e-01 -6.13660395e-01
-1.58301234e-01 3.73020887e-01 -1.74828187e-01 5.39405011e-02
5.98803818e-01 3.22818965e-01 1.48772181e-03 -1.61329567e-01
5.83160035e-02 4.99263197e-01 5.25437236e-01 -6.08691096e-01
7.56560504e-01 8.50612462e-01 3.27740997e-01 -8.55517685e-01
-4.18351799e-01 -3.41362238e-01 -6.79601490e-01 -9.35089439e-02
9.98890042e-01 -6.12500310e-01 -7.75347114e-01 1.33094534e-01
-1.25389028e+00 -5.05057693e-01 -1.96861506e-01 1.60985574e-01
-5.41976690e-01 2.39231378e-01 -8.90685260e-01 -6.53033197e-01
-2.30950713e-01 -1.50880337e+00 1.70437396e+00 4.49506879e-01
-2.91302018e-02 -7.65158713e-01 6.30583689e-02 5.03242791e-01
-1.49608448e-01 3.82413149e-01 9.21788216e-01 -5.86813390e-01
-6.88643217e-01 -2.12773576e-01 -2.12552175e-01 7.35119730e-02
3.64794545e-02 4.35495079e-01 -1.04218638e+00 -3.97127420e-01
-5.09399295e-01 -2.15792924e-01 1.03920257e+00 9.68849182e-01
1.01015043e+00 9.01839659e-02 -5.76559484e-01 1.09281111e+00
1.23065805e+00 1.66742966e-01 6.39299631e-01 -1.16784580e-01
6.69945121e-01 7.24647224e-01 5.03260911e-01 3.47830445e-01
8.99276044e-03 2.66651273e-01 5.56686819e-01 -4.70164746e-01
-1.90767797e-03 -1.38588762e-02 -9.91684794e-02 2.66966671e-01
-4.85895127e-02 -3.88664871e-01 -1.10841358e+00 7.88504839e-01
-1.48880649e+00 -5.18004835e-01 -6.23722188e-02 1.98375416e+00
7.00847805e-01 9.00544524e-02 -2.73133934e-01 -2.59297267e-02
7.44306624e-01 -2.48163432e-01 -4.87135708e-01 -3.71312112e-01
3.16534285e-03 3.30207437e-01 3.26846242e-01 5.53462148e-01
-9.67283368e-01 1.13242960e+00 5.72988653e+00 8.13223183e-01
-1.14262211e+00 -2.44355336e-01 1.21156418e+00 1.18934162e-01
-4.57466036e-01 1.51669875e-01 -8.78378034e-01 3.68727088e-01
3.03553909e-01 2.44924814e-01 1.59064278e-01 5.76830149e-01
-1.11997724e-01 -5.86624816e-02 -1.44894183e+00 8.74361932e-01
-1.77645981e-01 -1.66154480e+00 1.40215317e-02 1.74473286e-01
8.50546002e-01 -1.40025094e-01 -3.17069404e-02 1.52188897e-01
-9.88317188e-04 -1.35914838e+00 5.06503999e-01 4.08295274e-01
9.56546664e-01 -5.83325982e-01 9.37368810e-01 3.16665709e-01
-1.04723585e+00 1.96304813e-01 -6.41444504e-01 2.74561495e-01
2.46356893e-02 4.98018503e-01 -1.09635305e+00 2.30814576e-01
5.99094093e-01 5.64123690e-01 -4.50383484e-01 1.15743196e+00
8.72038826e-02 6.23031855e-01 -5.98773003e-01 3.47718239e-01
-1.09338120e-01 -4.34294462e-01 4.27389622e-01 1.20693839e+00
4.52673852e-01 2.97790438e-01 -2.60844022e-01 1.41058350e+00
-2.16824234e-01 -2.98194885e-02 -5.04722297e-01 -2.65066326e-01
3.17637891e-01 1.45029962e+00 -1.39715934e+00 -3.56537402e-02
-3.63090187e-01 7.87473679e-01 2.95826972e-01 4.38671649e-01
-6.12179160e-01 -1.51559830e-01 6.00047827e-01 1.22519530e-01
4.63738412e-01 2.87406057e-01 -6.91089690e-01 -9.30803061e-01
-1.56726420e-01 -5.90475261e-01 2.86002547e-01 -5.23192585e-01
-8.49098623e-01 3.55439156e-01 -2.70721048e-01 -1.26338315e+00
6.79287314e-03 -7.42734730e-01 -7.24640489e-01 6.81694090e-01
-1.08803988e+00 -1.15997994e+00 -1.25074878e-01 -1.31394595e-01
5.62363148e-01 -1.12672463e-01 6.62732363e-01 1.38627738e-01
-5.27851343e-01 5.33705652e-01 -3.02047543e-02 6.43342555e-01
3.30825120e-01 -1.53318274e+00 2.87058681e-01 6.13853931e-01
3.95481914e-01 5.79712987e-01 7.07465470e-01 -6.67891920e-01
-1.23064458e+00 -9.81735051e-01 3.35266560e-01 -7.04284728e-01
-1.86828580e-02 -4.05097097e-01 -8.95339787e-01 7.60674536e-01
-6.13707490e-02 4.60869223e-01 5.09481668e-01 -2.83536881e-01
2.66709059e-01 4.19926606e-02 -1.11344886e+00 5.09415805e-01
6.61958814e-01 -1.80010311e-02 -4.96427804e-01 4.97545265e-02
3.48121017e-01 -8.84758055e-01 -9.95572686e-01 7.70792603e-01
6.91609859e-01 -9.48244750e-01 1.01712441e+00 -3.99010003e-01
8.27943385e-01 -3.28076452e-01 2.19631910e-01 -1.09847915e+00
-9.28034335e-02 -3.51531565e-01 1.09066665e-02 1.21144724e+00
7.65085936e-01 -2.04153284e-01 1.44191968e+00 7.41251290e-01
-2.54927009e-01 -1.15713704e+00 -1.01376593e+00 -2.63127238e-01
5.10205507e-01 -2.47986928e-01 5.26038885e-01 7.91471124e-01
-1.80223271e-01 -1.42033014e-03 1.41390309e-01 3.69488925e-01
4.13706869e-01 4.92083728e-01 7.66705334e-01 -1.25364721e+00
-2.91682214e-01 -5.42521894e-01 -5.33537626e-01 -1.08528936e+00
1.57247320e-01 -1.15081310e+00 1.34726778e-01 -1.55241215e+00
3.47618103e-01 -4.87016112e-01 6.85581267e-02 6.08269691e-01
-2.86216319e-01 5.39863229e-01 -2.63405940e-03 -7.56096765e-02
-2.84267962e-01 4.29380000e-01 1.55657446e+00 -4.10184741e-01
-3.53896588e-01 5.23315445e-02 -4.04666215e-01 8.72159600e-01
4.47364420e-01 -6.66631460e-01 -1.72882825e-01 -5.59607208e-01
2.29824930e-01 3.98587614e-01 1.72982484e-01 -6.04821205e-01
2.69784540e-01 -2.83067320e-02 9.37983334e-01 -7.41411567e-01
4.17764962e-01 -7.07945287e-01 -1.02731986e-02 3.57215643e-01
-2.21715182e-01 -5.59801579e-01 2.00467333e-01 3.11003953e-01
-1.29270256e-01 -5.03512084e-01 8.54164362e-01 -3.34406346e-02
8.86050165e-02 4.92710471e-01 -3.17845553e-01 1.33620799e-01
9.39776182e-01 -7.14804113e-01 5.55026531e-02 -2.23965675e-01
-8.02548468e-01 4.76760328e-01 7.60696471e-01 4.40153889e-02
9.31847394e-01 -9.00848508e-01 -8.37149680e-01 5.28285503e-01
-9.59168822e-02 9.01154935e-01 -2.24244557e-02 6.55593693e-01
-1.16848660e+00 1.90938547e-01 2.80979723e-02 -1.17414379e+00
-1.14603770e+00 1.79318339e-01 6.05006993e-01 5.23236431e-02
-4.71492916e-01 1.15938175e+00 7.79929876e-01 -6.54041409e-01
1.30257890e-01 -5.47665954e-01 -3.91989738e-01 -4.71919356e-03
1.20432481e-01 -3.16052735e-02 -1.50571195e-02 -7.69090950e-01
-1.70727506e-01 9.95323658e-01 8.44163522e-02 -1.25893801e-02
1.36450100e+00 1.89100489e-01 -2.29473971e-03 5.52894510e-02
1.12186992e+00 4.07061391e-02 -1.60594010e+00 1.98452875e-01
-1.07965574e-01 -4.46759015e-01 1.05545158e-02 -3.28846872e-01
-1.61746478e+00 1.18648994e+00 4.93394673e-01 5.19735590e-02
8.54304314e-01 2.52435505e-01 6.41389728e-01 2.66120285e-01
5.62320538e-02 -8.65082502e-01 -4.30311337e-02 1.80345833e-01
7.86219895e-01 -1.38979590e+00 2.98848357e-02 -5.66118777e-01
-5.12160122e-01 1.28346777e+00 6.61861062e-01 -2.81711435e-03
5.64737618e-01 4.94972855e-01 1.93146244e-01 -3.63222480e-01
-4.60655332e-01 -1.21681228e-01 5.09459436e-01 3.35113138e-01
4.22881186e-01 -2.12820143e-01 -1.42984390e-01 7.34467328e-01
-1.38670862e-01 -2.61800200e-01 5.05647421e-01 5.94755650e-01
-3.29585761e-01 -8.89549613e-01 -3.11079741e-01 6.13376796e-01
-7.22574055e-01 2.12830566e-02 -3.63540560e-01 5.84695935e-01
3.84291768e-01 4.98921543e-01 4.93670821e-01 1.91107690e-02
1.40634015e-01 -2.03815579e-01 4.15702105e-01 -9.50216115e-01
-3.99615705e-01 3.29149961e-01 -5.06821215e-01 -4.49697137e-01
2.29906524e-03 -6.70326471e-01 -1.94949746e+00 2.00991526e-01
-4.67050314e-01 1.21964820e-01 4.79263246e-01 8.98329616e-01
1.84312612e-01 4.12437081e-01 3.70315999e-01 -1.08905661e+00
1.59788981e-01 -7.25355625e-01 -6.23542726e-01 6.51982546e-01
3.49445134e-01 -4.50725555e-01 -4.86732513e-01 4.29850787e-01] | [14.610157012939453, -2.923241138458252] |
7efa701d-348e-4edf-af31-8a2853ba019f | exemplar-based-generative-facial-editing | 2006.00472 | null | https://arxiv.org/abs/2006.00472v1 | https://arxiv.org/pdf/2006.00472v1.pdf | Exemplar-based Generative Facial Editing | Image synthesis has witnessed substantial progress due to the increasing power of generative model. This paper we propose a novel generative approach for exemplar based facial editing in the form of the region inpainting. Our method first masks the facial editing region to eliminates the pixel constraints of the original image, then exemplar based facial editing can be achieved by learning the corresponding information from the reference image to complete the masked region. In additional, we impose the attribute labels constraint to model disentangled encodings in order to avoid undesired information being transferred from the exemplar to the original image editing region. Experimental results demonstrate our method can produce diverse and personalized face editing results and provide far more user control flexibility than nearly all existing methods. | ['Zuowei Zhou', 'Yi Liu', 'Jingtao Guo', 'Zhenzhen Qian'] | 2020-05-31 | null | null | null | null | ['facial-editing'] | ['computer-vision'] | [ 6.64101243e-01 6.11747384e-01 -6.98608384e-02 -5.38061500e-01
-4.33575809e-01 -4.05567855e-01 5.76166451e-01 -8.62367988e-01
1.15803868e-01 9.98679519e-01 3.05130035e-01 4.42666262e-01
2.91176178e-02 -6.33902013e-01 -8.02795172e-01 -7.54847765e-01
5.56770504e-01 1.78263441e-01 -6.52524769e-01 -8.60701054e-02
4.80771139e-02 7.10124373e-01 -1.41690290e+00 4.17023927e-01
9.78939176e-01 6.52823269e-01 1.34238631e-01 3.30840498e-01
-9.52338502e-02 5.71714282e-01 -7.14912176e-01 -5.00494361e-01
4.57251489e-01 -8.49148989e-01 -2.71731257e-01 7.77972519e-01
6.83547139e-01 -5.65180302e-01 -5.19915104e-01 1.03095937e+00
4.95202839e-01 1.41923517e-01 5.21557987e-01 -1.39099264e+00
-1.32168257e+00 2.33512461e-01 -7.90029705e-01 -5.47327757e-01
3.54817033e-01 6.11423180e-02 5.62570810e-01 -1.09060097e+00
9.96378303e-01 1.49741459e+00 6.88121766e-02 1.11971700e+00
-1.56210959e+00 -1.00573742e+00 2.15916678e-01 -1.74994901e-01
-1.47904265e+00 -8.38303864e-01 1.08015764e+00 -2.25338593e-01
3.03348511e-01 4.24043328e-01 7.97169268e-01 1.15083051e+00
2.00007826e-01 5.02002716e-01 1.29715896e+00 -3.81032735e-01
-1.48280049e-02 1.48097754e-01 -6.48514032e-01 1.10177863e+00
-9.83137861e-02 2.72104472e-01 -6.64979279e-01 -8.29307139e-02
1.32111645e+00 7.30140805e-02 -3.79186898e-01 -3.92516553e-01
-9.87673938e-01 6.90859497e-01 1.99103817e-01 -1.98453113e-01
-3.58818918e-01 2.94279546e-01 -1.54305562e-01 4.42478150e-01
5.23963749e-01 5.18183947e-01 -9.38889384e-02 3.82060617e-01
-1.14469492e+00 3.13273042e-01 3.96706045e-01 1.33157659e+00
9.42791939e-01 3.44945073e-01 -4.31498349e-01 9.14836407e-01
2.83877939e-01 3.81501257e-01 3.87303042e-03 -1.51876581e+00
9.83642936e-02 3.82430226e-01 1.95677400e-01 -7.42482185e-01
2.93280721e-01 -3.03546160e-01 -8.44550490e-01 5.47897518e-01
-9.15959328e-02 -5.51794842e-02 -1.18038857e+00 2.02759194e+00
2.78861076e-01 3.37953925e-01 -3.21190655e-01 7.80932665e-01
6.31503165e-01 5.99250734e-01 -2.10064411e-01 -4.83149558e-01
1.06075287e+00 -7.54255533e-01 -1.12613773e+00 4.66736220e-02
-2.81831082e-02 -7.70641208e-01 1.04787314e+00 3.09686691e-01
-1.33592069e+00 -5.22449195e-01 -9.79278684e-01 -1.93612471e-01
2.06176147e-01 2.48469993e-01 5.33218980e-01 6.52768135e-01
-1.09172106e+00 4.89606619e-01 -5.08191049e-01 5.52465096e-02
8.48594069e-01 4.02372181e-01 -6.15105331e-01 -1.62246808e-01
-8.76440704e-01 6.31482422e-01 1.42578408e-01 1.46765932e-01
-9.66794550e-01 -1.08404565e+00 -8.62152040e-01 2.72841752e-02
3.04809898e-01 -1.00445783e+00 7.85592258e-01 -1.40848374e+00
-1.84828222e+00 1.05552006e+00 -4.16423231e-01 6.26919344e-02
6.26122475e-01 -8.83021057e-02 -2.26701796e-01 -1.75237134e-02
-2.02347469e-02 1.15577233e+00 1.33846390e+00 -1.68828619e+00
-1.34779140e-01 -1.50396362e-01 -1.23864613e-01 1.01621576e-01
-1.81256026e-01 -6.57189339e-02 -5.12240887e-01 -1.16561580e+00
4.08796556e-02 -9.91261423e-01 -1.36560410e-01 5.63237488e-01
-4.48585451e-01 3.39987159e-01 1.02498674e+00 -6.63460135e-01
1.12925446e+00 -2.31238413e+00 5.03206372e-01 1.27947286e-01
4.79925007e-01 3.00326180e-02 -3.96746725e-01 4.87858832e-01
-2.74138063e-01 1.18876934e-01 -4.20910537e-01 -6.64902866e-01
-5.79063818e-02 5.07392704e-01 -5.99093854e-01 3.74499381e-01
4.08771932e-01 8.94321620e-01 -6.67066932e-01 -6.68896496e-01
4.23101969e-02 8.24427724e-01 -8.42501640e-01 2.18012691e-01
-2.88021803e-01 9.74347532e-01 -3.16654712e-01 6.18628502e-01
8.32038939e-01 1.03227906e-01 -9.58715379e-02 -2.91013867e-01
1.38555646e-01 -3.34247619e-01 -8.51709425e-01 1.94485641e+00
-3.84525865e-01 5.29372871e-01 3.04904431e-01 -4.80535477e-01
9.97225344e-01 4.27416712e-01 4.39405471e-01 -4.92139041e-01
1.00177638e-02 -2.39187732e-01 -6.41759112e-02 -2.92579234e-01
4.23560470e-01 -3.18978608e-01 1.41184866e-01 4.54535902e-01
4.65993360e-02 -4.03265953e-01 -4.98377942e-02 1.87795579e-01
4.27661031e-01 6.61162019e-01 2.24998537e-02 -2.03594029e-01
3.82621288e-01 -3.87350410e-01 8.21502030e-01 4.11402196e-01
-3.63968983e-02 8.28387320e-01 3.84438157e-01 -2.64935732e-01
-1.17637086e+00 -1.16837502e+00 -6.34262487e-02 8.50298047e-01
-7.87962452e-02 -3.59458387e-01 -1.14026165e+00 -5.53299129e-01
-3.33565548e-02 8.05955827e-01 -9.10616517e-01 -4.27575260e-01
-7.11095095e-01 -3.90438557e-01 4.42355931e-01 3.80042940e-01
6.46292448e-01 -9.09107625e-01 -6.55394867e-02 -4.94206585e-02
-4.86359000e-02 -7.66489148e-01 -1.06441855e+00 -5.62369406e-01
-7.55862594e-01 -5.69654882e-01 -5.65601230e-01 -7.00190604e-01
1.24364460e+00 5.77543639e-02 7.71727145e-01 5.94617575e-02
-4.15912271e-01 3.48169535e-01 6.38745204e-02 -3.45005244e-01
-5.66278100e-01 -5.21836042e-01 -1.87258765e-01 5.18236697e-01
-2.47064784e-01 -7.15048254e-01 -6.44684255e-01 2.36702546e-01
-9.87751305e-01 4.69625056e-01 4.57354635e-01 1.01057482e+00
9.29133534e-01 -1.69971704e-01 6.86784923e-01 -1.08794260e+00
6.40426934e-01 -1.79137677e-01 -4.38633621e-01 2.77399898e-01
-5.69695711e-01 1.52169794e-01 4.28685874e-01 -5.94843388e-01
-1.55977845e+00 2.43702650e-01 1.77658737e-01 -8.36355627e-01
2.75881793e-02 6.43955469e-02 -6.43926203e-01 -2.19744235e-01
3.75830442e-01 2.64237285e-01 2.53895164e-01 -3.00650716e-01
7.18299866e-01 2.61747837e-01 6.02582812e-01 -8.00315797e-01
7.85791039e-01 6.67346716e-01 2.03015849e-01 -6.10681951e-01
-6.26586378e-01 5.50925255e-01 -6.87191248e-01 -3.58027220e-01
7.64596581e-01 -8.97530496e-01 -4.86025780e-01 9.90353301e-02
-1.19113815e+00 -4.71180864e-02 -6.27238333e-01 1.02404043e-01
-8.46626520e-01 1.31406978e-01 -4.83184457e-01 -6.39488161e-01
-9.71565917e-02 -1.13818359e+00 1.11827552e+00 2.45731041e-01
-4.11693484e-01 -7.02877998e-01 -1.81177199e-01 2.52702624e-01
2.91839123e-01 5.71494162e-01 1.00646472e+00 1.44038811e-01
-1.02558875e+00 -2.35199984e-02 -1.04834475e-01 2.44287327e-01
4.96468127e-01 1.47617057e-01 -8.94310355e-01 -2.75015622e-01
1.19610159e-02 -4.02169824e-02 7.81753838e-01 2.53241032e-01
1.31598127e+00 -4.56368506e-01 -2.72383094e-01 9.12361085e-01
1.13387132e+00 3.29064906e-01 9.93895411e-01 -3.68524104e-01
8.69584441e-01 6.37538850e-01 4.61803138e-01 4.65091139e-01
1.15918210e-02 6.05652392e-01 2.45008186e-01 -2.54960358e-01
-4.37863529e-01 -6.51952088e-01 2.84758329e-01 4.19552058e-01
-1.91705033e-01 -1.34305790e-01 -7.58915469e-02 1.63287997e-01
-1.72053182e+00 -1.21388435e+00 3.67533416e-01 2.14209318e+00
9.40809369e-01 -4.67244864e-01 -2.71706432e-01 -2.51081765e-01
8.00523102e-01 1.97973952e-01 -5.86611390e-01 -4.23100233e-01
-8.44441131e-02 5.53781748e-01 8.41176230e-03 6.84016347e-01
-6.68975413e-01 1.07886839e+00 7.39587641e+00 7.95582294e-01
-8.69596541e-01 1.01541109e-01 7.13062286e-01 -6.02395952e-01
-7.41289139e-01 6.38066977e-02 -6.00965321e-01 4.26929086e-01
2.81135619e-01 -2.26791143e-01 6.83845341e-01 5.34255028e-01
4.33054715e-01 1.64161831e-01 -1.19638252e+00 1.05529559e+00
2.18814626e-01 -1.36140156e+00 6.71954870e-01 2.59058118e-01
1.05263770e+00 -1.03685939e+00 4.62443024e-01 -9.39713046e-02
1.73110694e-01 -1.28632522e+00 7.14273751e-01 1.03593075e+00
1.24403596e+00 -9.86155450e-01 -1.24537632e-01 1.86286747e-01
-8.24502587e-01 2.12358862e-01 -3.28997284e-01 1.37270927e-01
1.51353016e-01 3.91551822e-01 -5.17712653e-01 2.45398715e-01
1.42938659e-01 5.57353675e-01 -1.82662487e-01 3.83202136e-01
-3.97824794e-01 2.79500633e-01 5.36597483e-02 4.67379123e-01
-2.39070445e-01 -7.76660144e-01 6.93517506e-01 7.82989204e-01
4.01901811e-01 2.91955292e-01 -4.79212515e-02 1.53268409e+00
-4.47324246e-01 -8.79780427e-02 -7.45503843e-01 -7.44234920e-02
5.93398511e-01 1.20004225e+00 -3.31440538e-01 -2.61852562e-01
-1.89049721e-01 1.63862133e+00 2.93479979e-01 6.69452786e-01
-9.53501225e-01 -1.78914607e-01 8.51137996e-01 2.28945181e-01
1.77801639e-01 -7.76378959e-02 -2.62017190e-01 -1.05736101e+00
-8.34510550e-02 -8.09661567e-01 -1.10292867e-01 -9.77076232e-01
-1.02042079e+00 4.55915451e-01 -1.77672170e-02 -1.04025662e+00
-1.20162673e-01 -1.95595786e-01 -5.95096886e-01 1.25130081e+00
-1.09169304e+00 -1.53769052e+00 -2.53739774e-01 7.32013226e-01
4.85723615e-01 -1.65288284e-01 8.85034800e-01 2.99612373e-01
-5.55310667e-01 9.09980953e-01 -1.69745207e-01 -8.30831230e-02
9.53280926e-01 -8.44743609e-01 4.72321957e-02 6.91431761e-01
9.70063806e-02 9.36464429e-01 6.10580981e-01 -7.73359954e-01
-1.33494580e+00 -1.25550008e+00 5.23919344e-01 -4.11544681e-01
-7.71459714e-02 -5.56759536e-01 -7.78001904e-01 9.57214594e-01
4.48691428e-01 3.88917290e-02 7.93472409e-01 -3.73524159e-01
-5.43441534e-01 -1.84021130e-01 -1.43134391e+00 9.72982883e-01
1.43588555e+00 -5.61586559e-01 -2.70679086e-01 1.27449319e-01
7.02610373e-01 -4.33598101e-01 -8.25652242e-01 2.99210817e-01
7.12623954e-01 -8.28450620e-01 8.73527348e-01 -6.14129722e-01
6.04597867e-01 -4.33802992e-01 -5.01595289e-02 -1.48110247e+00
-6.12559855e-01 -1.15976310e+00 -3.10148336e-02 1.45277631e+00
1.80510938e-01 -5.15505314e-01 8.36961925e-01 7.75750935e-01
-2.63371229e-01 -7.25726783e-01 -7.26801872e-01 -4.48483735e-01
-4.48993109e-02 1.70247704e-01 8.78700733e-01 1.06243837e+00
-3.57987523e-01 3.49042602e-02 -8.51206720e-01 4.69969064e-02
8.20493519e-01 3.74701113e-01 7.58179784e-01 -9.21800077e-01
-4.67709780e-01 -2.98302442e-01 -1.49090886e-01 -9.19005930e-01
4.28378969e-01 -9.17980552e-01 -2.01182947e-01 -1.30658364e+00
3.07121396e-01 -1.75889447e-01 3.23631652e-02 4.21071649e-01
-1.51123315e-01 4.51859564e-01 2.93207884e-01 1.36641860e-01
-3.85427289e-02 8.45537305e-01 2.08589888e+00 -5.97999319e-02
-6.53891042e-02 -3.06649923e-01 -9.33584094e-01 5.07752299e-01
5.57372391e-01 -3.64622265e-01 -7.55754650e-01 -3.01890045e-01
-5.49848378e-02 2.95169383e-01 4.15925711e-01 -4.94393617e-01
-1.48219585e-01 -5.04295588e-01 8.56694639e-01 -2.33673722e-01
6.79863453e-01 -8.29246223e-01 7.92084038e-01 1.62541181e-01
-4.73562002e-01 -1.28854364e-01 9.89488959e-02 5.28325260e-01
-1.63054749e-01 1.23712374e-02 1.06359887e+00 -1.69308320e-01
-3.33303690e-01 5.74743032e-01 -9.35068056e-02 -1.98466569e-01
1.09879744e+00 -4.19579238e-01 -7.71807581e-02 -5.56326926e-01
-1.00574362e+00 -9.52988788e-02 7.11750925e-01 4.07429725e-01
8.73174548e-01 -1.67051589e+00 -8.16969991e-01 6.96017563e-01
-8.64976123e-02 -2.46259600e-01 2.92882800e-01 4.48767155e-01
-2.48950139e-01 -5.52973375e-02 -5.54391623e-01 -2.12885618e-01
-1.33848917e+00 6.09981954e-01 4.40253437e-01 3.09580415e-01
-5.10140002e-01 7.83847511e-01 6.88369215e-01 -1.11927345e-01
-9.64939743e-02 2.65344858e-01 2.49586388e-01 -1.51074693e-01
4.40821588e-01 2.21908942e-01 -3.99977118e-01 -6.57422006e-01
1.46822646e-01 5.55517197e-01 -2.62503535e-01 -3.55055809e-01
1.12788320e+00 -2.86761791e-01 -3.00811559e-01 2.82785092e-02
1.10777295e+00 3.30965251e-01 -1.72444165e+00 1.79496203e-02
-7.57409692e-01 -8.62597823e-01 -1.08118549e-01 -8.05891573e-01
-1.31505823e+00 7.00434566e-01 4.11442429e-01 -4.06320482e-01
1.34204209e+00 -1.99067578e-01 5.13771415e-01 -1.00261167e-01
3.24614435e-01 -9.85239387e-01 7.24725798e-02 -2.01058341e-03
1.36152065e+00 -7.97057211e-01 5.29012531e-02 -7.98149109e-01
-6.28539920e-01 8.14570963e-01 6.82019651e-01 -3.87338489e-01
4.48318958e-01 4.36440736e-01 -1.69047549e-01 -1.34879306e-01
-6.48863196e-01 2.69503623e-01 3.65351111e-01 7.73950458e-01
4.83327299e-01 -1.37423232e-01 -3.64675522e-01 4.27907318e-01
-8.89569595e-02 9.10081416e-02 3.70167226e-01 7.87274778e-01
-1.64136887e-01 -1.40566826e+00 -3.70090604e-01 2.63666809e-01
-2.91957527e-01 -1.18057095e-01 -5.97015142e-01 7.96757042e-01
3.56514037e-01 5.29769599e-01 3.01431268e-02 -1.35514021e-01
3.35572153e-01 2.31695265e-01 1.09112298e+00 -6.54019177e-01
-3.16722363e-01 3.46101403e-01 -1.14036098e-01 -6.90555871e-01
-3.32733780e-01 -5.14727294e-01 -1.25842535e+00 -2.79026598e-01
-2.42705539e-01 -7.44640902e-02 3.10172945e-01 5.81369638e-01
8.07520390e-01 5.41156888e-01 6.99037552e-01 -8.36193204e-01
-1.90000266e-01 -6.96325600e-01 -5.98939121e-01 5.11507154e-01
2.26129249e-01 -6.77745104e-01 -7.33826235e-02 5.86710036e-01] | [12.524622917175293, -0.255819708108902] |
143c6d6c-7c95-413d-83ac-151824d4b473 | a-web-based-mpox-skin-lesion-detection-system | 2306.14169 | null | https://arxiv.org/abs/2306.14169v1 | https://arxiv.org/pdf/2306.14169v1.pdf | A Web-based Mpox Skin Lesion Detection System Using State-of-the-art Deep Learning Models Considering Racial Diversity | The recent 'Mpox' outbreak, formerly known as 'Monkeypox', has become a significant public health concern and has spread to over 110 countries globally. The challenge of clinically diagnosing mpox early on is due, in part, to its similarity to other types of rashes. Computer-aided screening tools have been proven valuable in cases where Polymerase Chain Reaction (PCR) based diagnosis is not immediately available. Deep learning methods are powerful in learning complex data representations, but their efficacy largely depends on adequate training data. To address this challenge, we present the "Mpox Skin Lesion Dataset Version 2.0 (MSLD v2.0)" as a follow-up to the previously released openly accessible dataset, one of the first datasets containing mpox lesion images. This dataset contains images of patients with mpox and five other non-mpox classes (chickenpox, measles, hand-foot-mouth disease, cowpox, and healthy). We benchmark the performance of several state-of-the-art deep learning models, including VGG16, ResNet50, DenseNet121, MobileNetV2, EfficientNetB3, InceptionV3, and Xception, to classify mpox and other infectious skin diseases. In order to reduce the impact of racial bias, we utilize a color space data augmentation method to increase skin color variability during training. Additionally, by leveraging transfer learning implemented with pre-trained weights generated from the HAM10000 dataset, an extensive collection of pigmented skin lesion images, we achieved the best overall accuracy of $83.59\pm2.11\%$. Finally, the developed models are incorporated within a prototype web application to analyze uploaded skin images by a user and determine whether a subject is a suspected mpox patient. | ['Taufiq Hasan', 'Anzirun Nahar Asma', 'Nawsabah Noor', 'S. M. Sakeef Sani', 'Joydip Paul', 'Tasnim Jahan', 'Md. Tazuddin Ahmed', 'Shams Nafisa Ali'] | 2023-06-25 | null | null | null | null | ['transfer-learning'] | ['miscellaneous'] | [ 4.17862982e-01 -4.48211551e-01 -2.53757518e-02 -1.03477046e-01
-6.03113592e-01 -5.28286636e-01 3.63986641e-01 3.59457582e-01
-5.38775325e-01 8.40371907e-01 -3.74760419e-01 -6.09150708e-01
-9.47737545e-02 -7.99717426e-01 -4.30039823e-01 -6.48678899e-01
-3.84042025e-01 2.86995828e-01 -3.53336730e-03 6.25068769e-02
4.32212763e-02 5.60957313e-01 -9.67413366e-01 5.01824021e-01
1.07502973e+00 9.05813932e-01 -3.09800878e-02 9.26203191e-01
1.25279605e-01 5.50715685e-01 -5.52249789e-01 -2.69163728e-01
2.62468755e-01 -3.83709341e-01 -7.93488443e-01 -1.35040238e-01
3.42842758e-01 -5.45554757e-01 -1.95663080e-01 6.79370403e-01
6.16537094e-01 -1.53474554e-01 7.92998016e-01 -1.14962947e+00
-7.07360566e-01 -1.79510847e-01 -9.20971632e-01 4.20053750e-01
3.75166815e-03 8.01776171e-01 5.96140325e-01 -3.41208160e-01
8.14372718e-01 8.67287755e-01 1.20607936e+00 7.97833145e-01
-9.44181025e-01 -7.02705383e-01 -3.36353958e-01 5.57668686e-01
-1.35059202e+00 1.64293274e-02 1.82269275e-01 -5.70564568e-01
1.20482171e+00 5.33754051e-01 6.43306315e-01 1.36906540e+00
1.77349925e-01 4.18655753e-01 1.76149726e+00 -2.00295016e-01
1.01882517e-01 -1.51092531e-02 5.10816351e-02 9.58363295e-01
1.44988969e-02 -1.20895561e-02 -2.42926516e-02 -5.36736131e-01
7.14416504e-01 1.79485083e-01 -4.19423491e-01 2.93161154e-01
-7.67972291e-01 9.95513678e-01 6.52096689e-01 -4.80583310e-02
-7.12045908e-01 1.90171733e-01 1.95985928e-01 1.57660276e-01
5.28139293e-01 3.48857105e-01 -4.46355790e-01 3.17855477e-02
-4.96854126e-01 -6.34779315e-03 4.03767437e-01 5.21982908e-02
5.41280627e-01 -3.27270597e-01 -2.78922748e-02 1.23677349e+00
3.66660476e-01 6.04205072e-01 1.49563879e-01 -4.21382487e-01
3.97925496e-01 7.04725862e-01 -1.06253237e-01 -7.45985508e-01
-7.34026134e-01 1.02746710e-02 -9.81401622e-01 4.08771783e-01
4.85104531e-01 -4.30658728e-01 -1.59723830e+00 1.48590600e+00
4.13757652e-01 2.21554101e-01 -1.57318577e-01 9.13912117e-01
7.32989907e-01 6.23497128e-01 6.20776594e-01 2.82975018e-01
1.65396667e+00 -6.84446096e-01 -3.60479087e-01 -1.25866398e-01
2.99507380e-01 -6.51960313e-01 8.31450164e-01 4.84879464e-01
-5.06132424e-01 -1.40425831e-01 -8.97351444e-01 2.07395762e-01
-8.60948741e-01 -3.51643562e-02 7.43907869e-01 7.75883317e-01
-1.10199833e+00 5.69428504e-01 -7.59781361e-01 -1.34109890e+00
8.48588467e-01 4.26635325e-01 -5.91074944e-01 -2.98735410e-01
-1.27991498e+00 1.15174627e+00 2.25782111e-01 1.06724419e-01
-7.34826863e-01 -7.67024815e-01 -4.98266369e-01 -3.16052020e-01
3.04879367e-01 -7.49091446e-01 6.98959053e-01 -6.70218825e-01
-1.23967516e+00 1.13888812e+00 4.34843153e-01 -3.31892490e-01
3.86851668e-01 1.09393522e-02 -6.33291125e-01 5.47541440e-01
-3.51587921e-01 7.90434062e-01 6.53321981e-01 -8.81689250e-01
-6.74463451e-01 -5.28823674e-01 9.69325528e-02 -2.82382816e-01
1.80766612e-01 4.03891504e-01 -4.22824562e-01 -3.70193779e-01
-7.66021073e-01 -8.28380704e-01 -3.70837629e-01 3.82565081e-01
-5.40296912e-01 -3.65635604e-01 6.39794230e-01 -1.24572480e+00
8.42679322e-01 -1.81308055e+00 -3.87857050e-01 5.66924214e-01
5.01266420e-01 8.27715933e-01 -7.06910670e-01 5.74511051e-01
-6.12594783e-02 4.44256246e-01 -3.28523070e-01 2.38829866e-01
-1.82437718e-01 2.67045177e-03 3.50496709e-01 5.48514843e-01
8.03238869e-01 9.51806843e-01 -6.90747857e-01 -2.98873186e-01
3.04225117e-01 8.45387340e-01 -1.08085446e-01 1.21505342e-01
-1.27972662e-01 1.01745211e-01 -2.65194893e-01 1.07796729e+00
8.89102459e-01 -6.93359792e-01 2.47891262e-01 -5.79685643e-02
1.58301517e-01 -3.68783623e-01 -6.75508082e-01 1.21325994e+00
-3.65581483e-01 4.94843364e-01 2.34420389e-01 -5.72392285e-01
5.73791146e-01 2.14976326e-01 3.11036229e-01 -6.33709967e-01
3.61680806e-01 2.27938704e-02 -1.11218560e-02 -1.00790811e+00
-6.57116994e-02 9.90628824e-03 5.49505115e-01 3.47096443e-01
-1.97869152e-01 4.81474966e-01 2.58963823e-01 1.22391749e-02
1.42189050e+00 -2.04301119e-01 5.27098298e-01 7.16353580e-02
1.86238021e-01 2.64710099e-01 3.06313366e-01 5.88362515e-01
-5.54073334e-01 8.23900998e-01 5.61206043e-01 -5.24756193e-01
-8.74985635e-01 -1.26904678e+00 -2.22231671e-01 8.13333154e-01
-3.42861325e-01 9.90492553e-02 -8.76972973e-01 -8.18706870e-01
1.51258796e-01 1.20809220e-01 -1.10520768e+00 2.46084824e-01
-1.49666891e-01 -1.46955025e+00 8.07358980e-01 4.98046845e-01
6.28082514e-01 -1.15685105e+00 -6.04818881e-01 1.35912150e-01
1.08138271e-01 -6.10386431e-01 -1.92908779e-01 3.53970453e-02
-8.39922130e-02 -2.03479934e+00 -1.02472162e+00 -5.10822713e-01
6.61927938e-01 -1.33504093e-01 7.17477858e-01 3.09936315e-01
-1.18381512e+00 3.26657742e-01 -4.26317424e-01 -3.07669580e-01
-3.00044596e-01 -7.27349445e-02 -8.59740004e-02 1.18988648e-01
8.04399014e-01 1.16871923e-01 -1.04457533e+00 -1.62558869e-01
-9.40219760e-01 -3.31641026e-02 5.86690366e-01 7.00127184e-01
3.75275820e-01 -2.49962881e-01 6.75087750e-01 -1.01262581e+00
7.36162364e-01 -9.22655463e-01 -2.40058675e-01 5.87188125e-01
-3.00808579e-01 -7.12740242e-01 5.27097821e-01 -5.66759408e-01
-7.76771128e-01 -1.35081202e-01 -3.96475762e-01 -1.99119151e-01
-3.93398881e-01 5.31557560e-01 4.07158911e-01 -1.70924455e-01
6.74201667e-01 -1.31517783e-01 4.27514672e-01 -4.67170566e-01
1.14275262e-01 1.01604593e+00 3.35489213e-01 -5.92493359e-03
4.59612876e-01 4.28699344e-01 -1.67009458e-01 -1.09779680e+00
-1.50783196e-01 -5.59878349e-01 -9.36578810e-02 -1.29394099e-01
1.23467982e+00 -6.22848988e-01 -9.27218497e-01 1.14443791e+00
-1.03316391e+00 -6.58158898e-01 3.45656842e-01 2.71502256e-01
1.10104732e-01 3.49722564e-01 -8.18406582e-01 -6.17430806e-01
-7.73158789e-01 -8.32967103e-01 6.18054807e-01 4.82750714e-01
-3.05506647e-01 -1.11655545e+00 3.33920151e-01 5.38258374e-01
6.10591710e-01 7.40122318e-01 1.30797541e+00 -7.48356760e-01
-2.67065287e-01 -2.39816636e-01 -7.98534095e-01 5.94757617e-01
2.75003612e-01 3.94077957e-01 -8.81904006e-01 -3.08716148e-01
-5.45746267e-01 -5.29195964e-01 9.91902351e-01 3.68853897e-01
8.51828873e-01 -1.36274472e-01 -3.54489714e-01 6.70648336e-01
1.71566737e+00 5.18353581e-01 7.76945949e-01 2.05870196e-01
6.65180624e-01 7.71425664e-01 -2.59714518e-02 3.83496374e-01
3.06318492e-01 4.97874282e-02 3.72955322e-01 -5.36701083e-01
-1.55959696e-01 6.22371584e-02 -1.41646981e-01 1.96016014e-01
-3.11955720e-01 -4.92945671e-01 -1.29720545e+00 5.17010093e-01
-1.24147165e+00 -6.91209376e-01 -3.09521794e-01 1.84775448e+00
6.90181673e-01 -5.27271628e-01 1.63281068e-01 -2.13687554e-01
7.93543518e-01 -9.49622914e-02 -6.43637300e-01 -4.27174300e-01
-3.52054425e-02 7.18815982e-01 4.22878444e-01 1.75942764e-01
-1.24532354e+00 5.75618327e-01 6.30183840e+00 6.68982863e-01
-1.66786361e+00 1.78302899e-01 8.67228746e-01 8.44343454e-02
1.71001747e-01 -6.19412780e-01 -1.79089099e-01 7.01679409e-01
6.80206895e-01 3.89276922e-01 6.55755460e-01 3.55465412e-01
8.83901343e-02 -2.23077402e-01 -6.90755606e-01 7.69974768e-01
9.04138386e-02 -1.46307158e+00 -2.33849332e-01 5.73460087e-02
4.54781473e-01 5.13689637e-01 2.50073880e-01 2.80961007e-01
1.85078815e-01 -1.30237877e+00 -3.25604409e-01 3.52603078e-01
1.29607463e+00 -6.02703214e-01 8.45706284e-01 -1.51356339e-01
-7.30797589e-01 2.23111510e-02 -1.54511020e-01 3.73944581e-01
-1.72366306e-01 2.38456085e-01 -1.19804883e+00 1.52061343e-01
8.81320536e-01 1.64065287e-01 -8.63604963e-01 1.12675750e+00
2.05472279e-02 5.72420239e-01 -4.92584944e-01 -3.04545909e-01
5.03555119e-01 4.39313203e-02 4.85068001e-02 1.26938438e+00
1.33200735e-01 2.05652103e-01 -1.76534146e-01 8.27809632e-01
-5.63254114e-03 2.88744807e-01 -3.84939015e-01 -2.29809567e-01
1.48394808e-01 1.48025143e+00 -7.02953696e-01 -2.59622455e-01
-5.13126433e-01 9.17541981e-01 1.67802379e-01 5.85225105e-01
-9.06633735e-01 -7.17363834e-01 7.29252994e-01 2.34175891e-01
2.02571720e-01 4.60953712e-01 1.51412934e-01 -8.00295949e-01
-4.02177423e-01 -1.09633183e+00 5.80293000e-01 -8.40792596e-01
-1.80733967e+00 5.99788368e-01 -3.32023561e-01 -6.47918522e-01
9.92112905e-02 -9.85950768e-01 -8.88968050e-01 1.19187689e+00
-1.66095340e+00 -1.44963276e+00 -5.31188130e-01 4.83448058e-01
2.42196232e-01 -1.91720694e-01 1.04432511e+00 3.14815372e-01
-8.09301198e-01 4.82915461e-01 1.91675663e-01 1.75106183e-01
5.87438047e-01 -1.18377519e+00 3.48923773e-01 5.93924642e-01
-6.15815222e-01 6.68099642e-01 -4.74480097e-04 -8.14890146e-01
-1.16997886e+00 -1.34306765e+00 6.60047293e-01 -3.26123506e-01
7.46096134e-01 -2.35831410e-01 -9.43581045e-01 3.59410673e-01
4.45251912e-01 -9.30544734e-02 1.02906120e+00 -5.73207391e-04
-5.05442798e-01 4.88406345e-02 -1.80400407e+00 4.22290266e-01
3.04941684e-01 -5.32908380e-01 -3.52191776e-01 5.63962996e-01
1.32107690e-01 -1.15932681e-01 -8.06381762e-01 3.28040570e-01
4.94707435e-01 -6.30660653e-01 8.75513911e-01 -8.24215770e-01
4.95818853e-01 -4.58503142e-03 1.15473799e-01 -1.31314921e+00
-3.25038850e-01 -5.13939679e-01 1.08018786e-01 9.26549911e-01
9.33304653e-02 -9.03822899e-01 8.02744806e-01 2.99974799e-01
3.86473358e-01 -1.26945996e+00 -8.35989833e-01 -4.27959204e-01
-1.84342954e-02 -1.32782608e-01 6.64271832e-01 1.14346492e+00
-2.83152580e-01 1.68203294e-01 -2.21469536e-01 2.64604330e-01
6.13862336e-01 -4.14978385e-01 4.38934982e-01 -9.95326757e-01
-1.56702071e-01 -3.99793833e-01 -3.37819874e-01 -8.49219188e-02
-3.46010029e-01 -8.58563185e-01 -2.33454093e-01 -1.67288756e+00
3.55868220e-01 -4.28368002e-01 -6.97248936e-01 9.25310016e-01
-3.34737152e-01 9.33709681e-01 1.85332835e-01 -2.62901366e-01
-1.42042086e-01 -3.19556266e-01 1.07173193e+00 -3.16896796e-01
-4.17033061e-02 -1.95360944e-01 -5.29543936e-01 6.09066427e-01
1.02591610e+00 -7.03120455e-02 -1.97719380e-01 -4.49548334e-01
-2.73658931e-02 -2.36660406e-01 6.62485838e-01 -8.46761584e-01
-4.90336828e-02 -3.85388374e-01 5.28972268e-01 -5.75377166e-01
2.63310462e-01 -5.07267535e-01 2.32213497e-01 7.94373393e-01
4.87180352e-02 -1.21293530e-01 3.23316425e-01 3.22694927e-01
2.88756758e-01 -6.98549068e-03 1.10401380e+00 -1.62436679e-01
-9.82304871e-01 6.10414267e-01 -6.77744567e-01 -2.11888347e-02
1.37489963e+00 -1.46564394e-01 -9.39897716e-01 1.15384690e-01
-4.41627562e-01 3.13179404e-01 3.56986314e-01 4.41614300e-01
7.55968869e-01 -9.03583467e-01 -9.19973135e-01 2.67015815e-01
2.35393822e-01 -4.46650505e-01 8.84318471e-01 8.74296188e-01
-1.42570651e+00 2.33537883e-01 -5.84753990e-01 -4.29772317e-01
-1.47524309e+00 4.17795151e-01 2.99455434e-01 -2.68925160e-01
-2.10367754e-01 9.56121266e-01 -1.24098606e-01 -4.94930238e-01
1.40901536e-01 -2.20027566e-01 -8.67137909e-02 -3.80382426e-02
7.50290513e-01 5.90942740e-01 2.10237242e-02 -2.26509780e-01
-7.44918227e-01 4.13526386e-01 -3.50927919e-01 5.26421249e-01
1.36368597e+00 3.91635925e-01 -2.19227523e-01 -2.74602950e-01
1.35804558e+00 -4.42029864e-01 -9.70507264e-01 2.53348738e-01
-4.20286983e-01 -3.19527119e-01 -2.68018961e-01 -1.60824871e+00
-1.10797310e+00 1.03753293e+00 1.22884309e+00 1.29034728e-01
1.16831231e+00 -1.62884057e-01 9.17200744e-01 2.48340324e-01
1.54598117e-01 -9.91373062e-01 -2.14939460e-01 2.40113497e-01
4.93587762e-01 -1.48605359e+00 -8.69043246e-02 -1.86470091e-01
-5.09324670e-01 8.45373333e-01 7.38254726e-01 1.09842606e-03
3.72553140e-01 -5.89755643e-03 7.24410713e-01 -3.07182074e-01
-9.07193542e-01 -1.45502672e-01 2.69750506e-02 1.32041740e+00
1.18492156e-01 2.27175951e-01 -1.85519889e-01 2.95051843e-01
5.13935864e-01 2.21746102e-01 2.95706838e-01 8.91167104e-01
-1.16782472e-01 -9.63746786e-01 -3.95939022e-01 9.23118889e-01
-7.10830986e-01 -1.88929543e-01 -6.36789441e-01 1.22388589e+00
5.48920274e-01 7.52184749e-01 2.91755319e-01 -3.50745350e-01
1.91446647e-01 -1.24756292e-01 5.33699751e-01 -3.49432886e-01
-9.50445831e-01 -2.36990750e-01 7.76028931e-02 -3.28184575e-01
-2.04401284e-01 -1.20122008e-01 -1.00532651e+00 -6.11951470e-01
3.92035618e-02 -2.94692218e-01 7.36536384e-01 5.40266395e-01
2.76344448e-01 3.26653481e-01 3.67172599e-01 -2.27788419e-01
-4.22920734e-01 -9.56587315e-01 -7.63853014e-01 6.07962072e-01
2.30028063e-01 -1.53977007e-01 -1.03171498e-01 -6.80507571e-02] | [15.685322761535645, -2.986222982406616] |
febe4afc-126f-4857-a07d-a055fde9ef02 | real-world-video-super-resolution-a-benchmark | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Yang_Real-World_Video_Super-Resolution_A_Benchmark_Dataset_and_a_Decomposition_Based_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Yang_Real-World_Video_Super-Resolution_A_Benchmark_Dataset_and_a_Decomposition_Based_ICCV_2021_paper.pdf | Real-World Video Super-Resolution: A Benchmark Dataset and a Decomposition Based Learning Scheme | Video super-resolution (VSR) aims to improve the spatial resolution of low-resolution (LR) videos. Existing VSR methods are mostly trained and evaluated on synthetic datasets, where the LR videos are uniformly downsampled from their high-resolution (HR) counterparts by some simple operators (e.g., bicubic downsampling). Such simple synthetic degradation models, however, cannot well describe the complex degradation processes in real-world videos, and thus the trained VSR models become ineffective in real-world applications. As an attempt to bridge the gap, we build a real-world video super-resolution (RealVSR) dataset by capturing paired LR-HR video sequences using the multi-camera system of iPhone 11 Pro Max. Since the LR-HR video pairs are captured by two separate cameras, there are inevitably certain misalignment and luminance/color differences between them. To more robustly train the VSR model and recover more details from the LR inputs, we convert the LR-HR videos into YCbCr space and decompose the luminance channel into a Laplacian pyramid, and then apply different loss functions to different components. Experiments validate that VSR models trained on our RealVSR dataset demonstrate better visual quality than those trained on synthetic datasets under real-world settings. They also exhibit good generalization capability in cross-camera tests. The dataset and code can be found at https://github.com/IanYeung/RealVSR. | ['Lei Zhang', 'Hui Zeng', 'Wangmeng Xiang', 'Xi Yang'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['video-super-resolution'] | ['computer-vision'] | [ 4.24831063e-01 -6.31959558e-01 -2.26377591e-01 -5.89400865e-02
-9.51933384e-01 -4.51025754e-01 2.61206657e-01 -9.24971640e-01
-1.61176305e-02 9.28056717e-01 2.22943008e-01 2.79985368e-02
2.95625657e-01 -6.17476940e-01 -8.31210017e-01 -7.30167270e-01
1.21080160e-01 -3.70623767e-01 3.78912359e-01 -3.76576036e-01
2.69738548e-02 4.26868170e-01 -1.62164390e+00 6.05079591e-01
9.11681712e-01 7.66954243e-01 5.12666345e-01 9.18293178e-01
5.00036538e-01 9.17809665e-01 -6.54290557e-01 -1.73846111e-01
5.78867495e-01 -5.40137112e-01 -4.48714048e-01 2.38744229e-01
5.67754149e-01 -6.92567825e-01 -8.94653141e-01 1.39752483e+00
3.97268325e-01 1.31895006e-01 1.58653572e-01 -9.10657942e-01
-1.20881164e+00 2.02017322e-01 -1.07426906e+00 6.25750542e-01
7.49888301e-01 1.73771799e-01 5.76859415e-01 -9.81726468e-01
5.92629373e-01 1.47233617e+00 5.10154128e-01 5.78244090e-01
-1.42661214e+00 -8.97373021e-01 -2.35543817e-01 4.36447948e-01
-1.51296818e+00 -5.66701472e-01 6.38800383e-01 2.06142701e-02
3.61214876e-01 3.34251016e-01 2.01793462e-01 1.48742414e+00
5.13953604e-02 2.98789412e-01 1.47723448e+00 -1.04223557e-01
-1.39103621e-01 9.38972235e-02 -4.13592935e-01 2.31369093e-01
3.07085425e-01 3.46176535e-01 -5.35976291e-01 2.33804047e-01
1.38733232e+00 1.51822135e-01 -9.72455025e-01 1.39254900e-02
-1.29721034e+00 4.31115419e-01 3.16414595e-01 2.77558535e-01
-1.02028608e-01 -2.22919390e-01 5.35661764e-02 5.83603740e-01
3.32377344e-01 1.31131455e-01 -2.01969042e-01 2.05539867e-01
-8.14076900e-01 -6.49334714e-02 8.44700634e-02 1.16227603e+00
6.05808794e-01 2.15300947e-01 -6.43795505e-02 1.20284748e+00
-9.52189639e-02 5.78525603e-01 6.07457161e-01 -1.31744015e+00
6.55793250e-01 -1.21483535e-01 4.04865146e-01 -1.02645206e+00
1.88322797e-01 -1.58314258e-01 -1.21568382e+00 2.00965092e-01
2.35155329e-01 1.27562284e-01 -7.57915854e-01 1.39266360e+00
8.79335627e-02 4.57554460e-01 3.08603942e-01 1.32728899e+00
8.59573483e-01 9.60424423e-01 -2.60532200e-01 -6.63965225e-01
1.17387474e+00 -8.17846894e-01 -8.83542478e-01 -6.12939820e-02
-2.07845658e-01 -9.04212058e-01 1.40040183e+00 5.65967500e-01
-1.20856237e+00 -1.07365596e+00 -1.13910294e+00 -2.04209805e-01
4.45910022e-02 2.29162499e-01 -1.61993271e-03 4.08749104e-01
-1.00113928e+00 4.51270193e-01 -5.74207366e-01 -5.53830601e-02
2.78975517e-01 -2.46623695e-01 -5.15233457e-01 -7.83155143e-01
-1.44094265e+00 6.73840046e-01 1.76822692e-01 6.19790517e-02
-1.08282185e+00 -6.40251696e-01 -8.40360045e-01 -2.66619772e-01
4.17506933e-01 -2.93482155e-01 9.65844512e-01 -1.08655417e+00
-1.45749474e+00 8.15230727e-01 -1.04830220e-01 -1.38792872e-01
5.19171357e-01 -3.04389000e-01 -9.81946707e-01 4.61165249e-01
-1.42182065e-02 1.03467435e-01 1.33713770e+00 -1.71152627e+00
-6.02921963e-01 -1.91128582e-01 2.64359377e-02 2.17657164e-01
-8.16486478e-02 2.63102829e-01 -5.66849887e-01 -9.65678930e-01
9.62520465e-02 -5.71767509e-01 2.47921664e-02 -2.58636922e-01
-2.06925794e-01 5.79334736e-01 1.09173870e+00 -1.06216705e+00
1.14414132e+00 -2.26986170e+00 8.84461552e-02 -4.12805438e-01
1.02452636e-01 4.49193180e-01 -3.28204572e-01 -3.99841480e-02
-4.29452300e-01 6.29456714e-02 -7.31283426e-02 4.53045368e-02
-6.68650806e-01 1.32690612e-02 -5.79372466e-01 6.38613105e-01
6.01488128e-02 5.24639368e-01 -8.73681664e-01 -4.77974534e-01
3.11509192e-01 9.35828626e-01 -2.10428342e-01 4.61508244e-01
1.71568751e-01 8.10924888e-01 -1.34280220e-01 7.29869604e-01
9.95101631e-01 -1.80529535e-01 7.35053793e-02 -6.27681196e-01
-7.77209597e-03 -2.85750538e-01 -1.25307739e+00 1.38796031e+00
-5.54165661e-01 8.12880695e-01 1.22177161e-01 -6.71910048e-01
9.72003400e-01 2.04306036e-01 3.01539153e-01 -1.03102577e+00
-2.67320037e-01 1.25341371e-01 -4.51203197e-01 -7.06332743e-01
6.28285646e-01 -1.81982666e-01 2.52523482e-01 2.48468723e-02
-2.25411177e-01 3.95452566e-02 1.10045522e-01 1.06346577e-01
8.79041016e-01 2.24078566e-01 2.13080063e-01 2.77066201e-01
8.02311122e-01 -3.47587675e-01 7.61200190e-01 3.91287148e-01
-1.39830872e-01 1.33941054e+00 1.10769778e-01 -1.94849104e-01
-1.47070813e+00 -1.34759820e+00 -2.13005245e-01 7.60649264e-01
5.30626118e-01 -1.40085697e-01 -6.22803867e-01 -1.76805809e-01
-4.88123775e-01 5.72240114e-01 -3.42462957e-01 -4.69094403e-02
-7.24908471e-01 -8.58762622e-01 3.83461744e-01 2.97208697e-01
9.42359805e-01 -8.27498794e-01 -4.27159369e-01 -8.72034580e-02
-6.36332035e-01 -1.65616655e+00 -5.67601323e-01 -5.61141491e-01
-6.53789163e-01 -1.22901285e+00 -1.03543806e+00 -6.56895399e-01
3.68601739e-01 8.84980977e-01 1.05119157e+00 -5.20050414e-02
-3.95022839e-01 2.55036920e-01 -6.18104935e-01 4.05224055e-01
-5.19819975e-01 -5.59471428e-01 2.75584966e-01 3.05145383e-01
2.38539372e-02 -4.61571544e-01 -7.48112381e-01 6.08694196e-01
-1.17043495e+00 1.81881100e-01 5.48779547e-01 8.11230898e-01
8.92020166e-01 6.99992895e-01 3.92598629e-01 -4.66840416e-01
3.69769126e-01 -3.53128761e-01 -6.87473416e-01 2.22092465e-01
-2.79254884e-01 -3.93986493e-01 9.91018295e-01 -6.29663646e-01
-1.48924232e+00 -2.98967242e-01 1.60028279e-01 -1.03835046e+00
-1.80289388e-01 -2.59989142e-01 -3.09413522e-01 -9.97622311e-02
5.37050366e-01 4.99238580e-01 -2.63641655e-01 -5.56551993e-01
3.65143746e-01 8.53193045e-01 9.41507757e-01 -2.90448308e-01
1.21884024e+00 6.47913218e-01 -2.32730865e-01 -1.01483703e+00
-7.85562634e-01 -1.37506872e-01 -3.38360041e-01 -1.50587097e-01
9.07300234e-01 -1.43929815e+00 -3.30531627e-01 5.32816172e-01
-8.25497150e-01 -5.11906981e-01 -1.37261659e-01 5.86395442e-01
-6.27270877e-01 5.74545324e-01 -9.59503174e-01 -5.74784398e-01
-2.02168487e-02 -1.09176314e+00 1.00624430e+00 5.57282686e-01
4.36723620e-01 -6.10570312e-01 -8.80154371e-02 4.98539239e-01
5.59726894e-01 1.45701081e-01 5.94158828e-01 3.15143913e-01
-7.44469225e-01 1.51670650e-01 -6.16377175e-01 7.76535153e-01
2.59636283e-01 9.24781784e-02 -8.45126033e-01 -6.83520317e-01
3.09637845e-01 -2.69104600e-01 7.68637776e-01 3.41147423e-01
1.44101477e+00 -3.19986343e-01 7.80008286e-02 1.10119915e+00
1.79904985e+00 1.04976028e-01 1.17926753e+00 3.94897133e-01
7.56822586e-01 3.43191683e-01 1.03296196e+00 2.01007187e-01
2.27365240e-01 9.18236673e-01 2.84325719e-01 -2.59346008e-01
-5.89337647e-01 -3.28119516e-01 8.13836336e-01 5.56426644e-01
-4.99253541e-01 -6.31635189e-02 -3.46856177e-01 1.15295753e-01
-1.26779008e+00 -1.24973738e+00 -1.92539468e-01 2.50381064e+00
8.87949228e-01 -1.94314823e-01 -7.88822547e-02 9.72991716e-03
1.06570196e+00 4.81389731e-01 -4.98038888e-01 1.03050649e-01
-7.04811096e-01 -3.98317277e-02 6.24078751e-01 4.10191625e-01
-9.02929187e-01 8.13042223e-01 5.66951275e+00 8.95127535e-01
-1.25853968e+00 3.02353084e-01 9.09256995e-01 -2.80096054e-01
-3.12068462e-02 -3.06520432e-01 -6.43464446e-01 6.07544661e-01
8.13630104e-01 -8.37794840e-02 9.85810041e-01 4.84449863e-01
5.11434913e-01 5.21103144e-02 -7.53743708e-01 1.52977514e+00
1.93343863e-01 -1.01536727e+00 -2.18163729e-02 -1.95375919e-01
9.76413488e-01 -6.37043267e-02 4.04606313e-01 2.90927112e-01
5.59354164e-02 -1.22258568e+00 4.55268949e-01 4.94082034e-01
1.64621031e+00 -5.35105407e-01 5.08855283e-01 1.11586265e-02
-1.38220417e+00 -2.89710015e-01 -8.16485822e-01 3.86427253e-01
1.83433637e-01 3.03295940e-01 2.46252809e-02 6.95968091e-01
1.32833886e+00 9.12271917e-01 -7.69053757e-01 6.02396011e-01
-2.53315151e-01 3.45584959e-01 2.40442142e-01 1.04238129e+00
-4.39584345e-01 -4.89180923e-01 7.17404604e-01 1.08909750e+00
5.22838116e-01 5.05932570e-01 -2.25987047e-01 8.29283655e-01
-8.28766525e-02 -2.45467529e-01 -5.70645452e-01 2.61210680e-01
4.33560699e-01 1.34454811e+00 -2.81332701e-01 -2.80252427e-01
-7.80420482e-01 1.23766410e+00 -1.36251286e-01 9.68745530e-01
-1.20213497e+00 -3.60940486e-01 7.23378360e-01 2.46347293e-01
3.63226384e-01 -8.78004730e-02 6.37682602e-02 -1.78149843e+00
6.38446063e-02 -1.50802791e+00 3.08176130e-01 -1.41069102e+00
-1.27078998e+00 8.20643127e-01 -1.01395650e-02 -1.78743434e+00
1.04865439e-01 -3.68284822e-01 -3.38451654e-01 8.31054986e-01
-1.76005781e+00 -7.56397188e-01 -7.95670211e-01 9.76918519e-01
8.48682404e-01 -1.72526658e-01 2.71772981e-01 4.32594389e-01
-7.82023668e-01 4.28516239e-01 3.44973803e-01 1.00079037e-01
9.89906907e-01 -8.62761438e-01 1.18516617e-01 1.19629800e+00
-2.07996771e-01 4.51771438e-01 9.30937588e-01 -4.65317786e-01
-1.42995524e+00 -1.34109259e+00 -1.85701363e-02 -2.94537097e-01
4.27028924e-01 -8.05714503e-02 -1.30247355e+00 7.04186380e-01
1.56224385e-01 3.88887316e-01 1.58624113e-01 -6.79192424e-01
-5.40244222e-01 -4.06805992e-01 -1.25465798e+00 6.35221064e-01
1.08718765e+00 -6.76813185e-01 -5.80031931e-01 6.40211180e-02
9.91866887e-01 -4.31706131e-01 -1.18419397e+00 5.33489287e-01
3.81981313e-01 -1.33830869e+00 1.45452523e+00 -2.42927074e-01
8.36187124e-01 -7.37651646e-01 -5.87624431e-01 -1.24814653e+00
-1.60732254e-01 -5.87676287e-01 -1.63953543e-01 1.21209121e+00
-2.88400427e-02 -6.27301037e-01 3.40695649e-01 2.34508157e-01
2.73743898e-01 -4.36221123e-01 -4.99965668e-01 -8.81997883e-01
-3.09907734e-01 -2.95022558e-02 5.22368848e-01 1.02293622e+00
-5.07969737e-01 1.35174587e-01 -7.99354017e-01 5.39274931e-01
1.11123693e+00 1.64963424e-01 7.46137679e-01 -5.93091488e-01
-3.91915679e-01 -3.85873429e-02 -1.46718368e-01 -8.73301268e-01
-1.10098876e-01 -2.57364035e-01 -1.35693744e-01 -1.17804074e+00
3.20474088e-01 -1.59780473e-01 -2.48165950e-01 -1.00938901e-01
-3.09194028e-01 7.77666509e-01 1.95342094e-01 4.83394355e-01
-5.21217346e-01 5.89389563e-01 1.48588419e+00 1.61960527e-01
-2.29249880e-01 -1.16657406e-01 -5.79057992e-01 7.02893972e-01
6.54759943e-01 -6.47611767e-02 -3.02966624e-01 -3.95804644e-01
-9.21497568e-02 6.36047781e-01 4.86578077e-01 -1.00826275e+00
-1.42655849e-01 -2.92950511e-01 7.76361823e-01 -2.00186491e-01
4.84779060e-01 -7.30925858e-01 5.48789203e-01 1.71147078e-01
-3.66461247e-01 1.04321809e-02 -5.50060533e-02 6.59715772e-01
-5.00828326e-01 2.82813162e-01 1.28313756e+00 -2.18783140e-01
-7.95346320e-01 4.75642204e-01 7.85774961e-02 -2.25037262e-02
9.78426993e-01 -2.95389563e-01 -8.22980583e-01 -5.21716237e-01
-4.37074184e-01 -9.69010666e-02 1.06862533e+00 5.19518375e-01
1.07909918e+00 -1.33784795e+00 -8.87576580e-01 3.82905781e-01
-8.44147578e-02 -1.77217517e-02 7.68486917e-01 7.91310668e-01
-6.73272908e-01 1.19863279e-01 -5.47701538e-01 -3.70038331e-01
-1.49857163e+00 1.02970481e+00 5.01851916e-01 4.38574925e-02
-9.33425009e-01 4.23223943e-01 5.31422079e-01 1.05274223e-01
2.11347360e-02 1.01177916e-02 -3.81641388e-01 -3.60832930e-01
1.09102046e+00 6.15698218e-01 -3.52221221e-01 -9.06849980e-01
-8.86134729e-02 8.80761802e-01 7.48832449e-02 1.47094252e-03
1.32751501e+00 -8.14169407e-01 1.17862217e-01 1.54490083e-01
1.32272375e+00 2.43427739e-01 -1.55660427e+00 -4.36499238e-01
-5.11915803e-01 -1.24822545e+00 -7.94360340e-02 -2.54803330e-01
-1.40672004e+00 6.86761320e-01 8.35354269e-01 2.36260116e-01
1.81684983e+00 -8.86791497e-02 7.36228824e-01 -1.21022291e-01
5.34108877e-01 -8.16327095e-01 3.15785229e-01 -2.12826896e-02
1.07701337e+00 -1.37354100e+00 1.13925248e-01 -4.55080211e-01
-9.51551497e-01 1.12947392e+00 6.41713142e-01 -1.80429503e-01
4.56725508e-02 9.63512659e-02 1.38551816e-01 2.96537429e-01
-7.69930363e-01 -8.28329623e-02 -5.53012267e-02 8.20470154e-01
2.42032468e-01 -1.69764355e-01 3.11288703e-02 3.05743009e-01
-2.62391530e-02 5.49365878e-02 1.07769752e+00 1.84806153e-01
-2.84082323e-01 -6.94532752e-01 -8.36866140e-01 1.26014888e-01
-7.80210078e-01 1.80497840e-02 3.00394416e-01 6.71854556e-01
-9.04275626e-02 1.07451916e+00 -1.95473045e-01 -5.56988597e-01
4.24050987e-01 -6.77568376e-01 4.84463006e-01 -2.76988328e-01
2.72867590e-01 1.31788284e-01 -3.03166598e-01 -9.59568262e-01
-5.02198458e-01 -5.03872156e-01 -7.35544920e-01 -4.06228691e-01
1.31745949e-01 -2.47365236e-01 2.55421966e-01 3.33620459e-01
8.27978179e-02 6.00633860e-01 1.03740585e+00 -1.04698586e+00
-5.91911912e-01 -7.77034342e-01 -9.30703640e-01 7.24171698e-01
6.62484527e-01 -4.30782318e-01 -6.36089623e-01 4.52781260e-01] | [11.088287353515625, -2.083588123321533] |
97e3aadf-7ca9-4378-a6d1-74c847c5de15 | causal-intervention-for-subject-deconfounded | 2204.07935 | null | https://arxiv.org/abs/2204.07935v1 | https://arxiv.org/pdf/2204.07935v1.pdf | Causal Intervention for Subject-Deconfounded Facial Action Unit Recognition | Subject-invariant facial action unit (AU) recognition remains challenging for the reason that the data distribution varies among subjects. In this paper, we propose a causal inference framework for subject-invariant facial action unit recognition. To illustrate the causal effect existing in AU recognition task, we formulate the causalities among facial images, subjects, latent AU semantic relations, and estimated AU occurrence probabilities via a structural causal model. By constructing such a causal diagram, we clarify the causal effect among variables and propose a plug-in causal intervention module, CIS, to deconfound the confounder \emph{Subject} in the causal diagram. Extensive experiments conducted on two commonly used AU benchmark datasets, BP4D and DISFA, show the effectiveness of our CIS, and the model with CIS inserted, CISNet, has achieved state-of-the-art performance. | ['Yun Liang', 'Yizhou Wang', 'Tao Wang', 'Diqi Chen', 'Yingjie Chen'] | 2022-04-17 | null | null | null | null | ['facial-action-unit-detection'] | ['computer-vision'] | [ 6.51277065e-01 2.63195187e-01 -5.31759083e-01 -5.63712239e-01
-2.88501054e-01 -1.57915289e-03 6.73576772e-01 -8.09044361e-01
2.57155806e-01 7.21115112e-01 8.44102681e-01 4.70441654e-02
-2.43675008e-01 -5.86059213e-01 -9.45791602e-01 -1.04533851e+00
-1.60189211e-01 -3.11950177e-01 -3.21773261e-01 3.51877868e-01
-1.94515675e-01 1.09285928e-01 -1.76810181e+00 4.46435332e-01
4.88909245e-01 1.21113133e+00 -5.65233171e-01 4.82549220e-02
1.68941960e-01 1.22543585e+00 -5.55936992e-01 -6.31272554e-01
9.68425944e-02 -9.17546511e-01 -5.93884170e-01 2.62673497e-01
5.46841443e-01 -6.99057341e-01 -5.45904338e-01 8.79861891e-01
3.02162319e-01 -2.98950046e-01 1.04760218e+00 -1.79436004e+00
-8.56501162e-01 8.12425852e-01 -1.05110979e+00 1.52334467e-01
4.83502775e-01 4.45718132e-02 1.24601328e+00 -8.27392101e-01
5.25889158e-01 1.82622433e+00 1.65924117e-01 8.67834449e-01
-1.23191512e+00 -1.32150805e+00 5.28176546e-01 6.25836015e-01
-1.33149123e+00 -4.90094453e-01 7.07486272e-01 -5.66704631e-01
7.67167926e-01 1.97569251e-01 3.76350909e-01 1.61464846e+00
2.52411664e-01 1.13461792e+00 9.09978628e-01 -2.55209982e-01
-1.29007548e-01 -6.46418035e-01 6.43222556e-02 9.64602947e-01
-1.34635329e-01 3.54026198e-01 -1.06543612e+00 -3.64250243e-01
8.12857866e-01 -9.16304141e-02 -8.42534825e-02 9.12929848e-02
-7.21907854e-01 7.22435832e-01 3.15061718e-01 2.62697469e-02
-5.06514490e-01 5.93211174e-01 1.55391306e-01 -1.01729639e-01
4.53299433e-01 -4.81858283e-01 -3.20283085e-01 3.24060358e-02
-3.55731517e-01 9.55117717e-02 3.35052341e-01 8.14175248e-01
4.79952008e-01 -1.21671014e-01 -5.85062563e-01 7.56974697e-01
5.39595842e-01 5.77119291e-01 1.54368043e-01 -6.56972587e-01
-1.05773680e-01 8.45322728e-01 -3.34789127e-01 -1.12571788e+00
-7.86963925e-02 2.20657036e-01 -8.39204609e-01 -1.60805181e-01
1.14755340e-01 -1.78647563e-01 -9.85971868e-01 2.11203122e+00
4.28161860e-01 9.74075675e-01 -1.70466140e-01 7.61136651e-01
1.04523396e+00 2.93551475e-01 5.88840425e-01 -6.79919362e-01
1.50732458e+00 -5.58501184e-01 -1.04396701e+00 1.16381519e-01
2.06568465e-01 -4.16248232e-01 7.49743044e-01 2.33184233e-01
-4.29476470e-01 -3.10679078e-01 -5.93338847e-01 7.58915097e-02
2.96334684e-01 3.50281358e-01 9.81399655e-01 4.33901697e-01
-4.68552530e-01 3.53992283e-01 -6.52866006e-01 -2.16081738e-01
9.53382790e-01 3.98375571e-01 -5.82249343e-01 -6.06417246e-02
-1.51743150e+00 2.69346654e-01 -3.63457873e-02 2.41914749e-01
-1.53749609e+00 -1.08211720e+00 -7.76317894e-01 -5.44638652e-03
6.55335546e-01 -6.06702626e-01 1.07299972e+00 -1.14227962e+00
-1.39026880e+00 7.11156070e-01 -5.81978321e-01 1.72933042e-02
1.87317833e-01 -3.36542547e-01 -4.02952850e-01 -7.59820342e-02
-2.69233752e-02 6.29815280e-01 1.20539105e+00 -1.16111398e+00
-7.83978999e-01 -5.66579998e-01 -1.30317286e-02 -1.32222116e-01
-2.10818395e-01 5.11905670e-01 -4.70143050e-01 -6.20536625e-01
-1.77322224e-01 -6.63173556e-01 2.37204716e-01 3.67181003e-02
-2.94836462e-01 -9.10204351e-01 8.73260140e-01 -5.68089783e-01
1.39628935e+00 -2.28279281e+00 3.12399000e-01 6.87313825e-02
3.51313472e-01 -2.32320204e-01 -1.53659135e-01 -1.25175267e-01
-7.03310966e-01 6.09211922e-02 -2.03495637e-01 -2.15245903e-01
-4.00647707e-03 5.09963870e-01 -4.79291528e-01 6.14494979e-01
4.92910951e-01 1.00757647e+00 -7.73079216e-01 -5.66551924e-01
3.23630497e-02 4.57635581e-01 -6.87550485e-01 3.76999140e-01
-2.78701454e-01 2.83838630e-01 -4.50800419e-01 8.75251114e-01
5.43493092e-01 -5.86220697e-02 2.07755968e-01 -5.09078681e-01
1.81916669e-01 -2.76637468e-02 -9.99134004e-01 1.57484794e+00
9.08051357e-02 3.60838383e-01 -1.84650987e-01 -7.93395162e-01
8.19351554e-01 5.13280153e-01 9.47252989e-01 -4.85806018e-01
5.45505285e-01 -3.68469924e-01 3.74322082e-03 -7.53115833e-01
-3.35183948e-01 -3.65156293e-01 6.63361773e-02 3.83172572e-01
2.71777809e-01 7.94736803e-01 -1.04327396e-01 9.38952267e-02
1.25235343e+00 3.13704967e-01 4.25242424e-01 -1.65840536e-01
6.83866918e-01 -5.34766614e-01 9.56248403e-01 4.17966425e-01
-4.67353135e-01 1.82785645e-01 1.37750208e+00 -4.01438236e-01
-1.97999880e-01 -1.11645532e+00 -3.39393646e-01 1.07456064e+00
1.59881134e-02 -5.53181231e-01 -7.46390343e-01 -1.13734436e+00
2.36469537e-01 6.84374392e-01 -1.51354730e+00 -4.02499527e-01
-3.04764122e-01 -1.08747733e+00 6.80167735e-01 5.77304780e-01
5.71176708e-01 -1.06210089e+00 -5.44521213e-02 -2.57355124e-01
-3.89295459e-01 -9.81031597e-01 -4.13182348e-01 -4.53410149e-01
-2.35682279e-01 -1.42303813e+00 -1.03689745e-01 -1.44166991e-01
7.35553861e-01 -3.13659310e-02 7.27427661e-01 -2.81109691e-01
-5.94738364e-01 4.28297609e-01 -2.31936425e-01 -5.45142829e-01
-2.86429990e-02 -6.67879462e-01 2.20167905e-01 8.12244892e-01
8.75680625e-01 -5.19169986e-01 -7.36707270e-01 6.27255619e-01
-8.25338900e-01 -7.23868385e-02 5.11301517e-01 8.63451719e-01
3.96014839e-01 -1.22156322e-01 4.90009815e-01 -7.59482145e-01
1.76139414e-01 -6.90496981e-01 -3.67161155e-01 2.88175732e-01
-4.32923526e-01 -1.57733694e-01 -1.22988738e-01 -4.68053401e-01
-1.52183354e+00 2.52826452e-01 -1.45149501e-02 -8.29219520e-01
-3.40560049e-01 4.40935373e-01 -8.11378300e-01 3.44453275e-01
6.18741572e-01 -6.69759661e-02 1.94680274e-01 -1.48146927e-01
5.19546330e-01 5.42135715e-01 4.27698731e-01 -5.68524539e-01
4.03795123e-01 8.01376641e-01 1.39576197e-01 -5.96545696e-01
-1.00987840e+00 -1.53267831e-01 -7.89873183e-01 -5.27350426e-01
1.09894264e+00 -1.03171742e+00 -1.19078314e+00 4.48425114e-01
-1.36546457e+00 -2.93045491e-03 -3.99236009e-02 3.41481030e-01
-3.82049561e-01 -5.91326952e-02 -3.04966956e-01 -7.63646007e-01
3.14572491e-02 -1.10895956e+00 1.25761771e+00 1.13428675e-01
-3.43800783e-01 -5.83069444e-01 -7.06552491e-02 4.26778108e-01
-4.52910125e-01 3.22026610e-01 8.18583429e-01 1.51451444e-02
-6.22188330e-01 1.87275298e-02 -5.66897571e-01 2.64659256e-01
7.54106224e-01 3.50344062e-01 -1.20372951e+00 2.94234097e-01
-2.66841173e-01 -2.05537006e-01 8.77061486e-01 4.77197617e-01
1.53544056e+00 -5.03872335e-01 -5.17926395e-01 3.88681620e-01
8.00957203e-01 2.14344278e-01 9.82795835e-01 -5.97431004e-01
9.31136906e-01 8.06962490e-01 6.70728505e-01 6.40980601e-01
1.76117197e-01 7.32559323e-01 5.94034135e-01 -7.41407182e-03
-2.22001508e-01 -5.18633604e-01 7.01927841e-01 1.54191166e-01
-3.95857066e-01 -1.23773687e-01 -3.96348476e-01 2.75704414e-01
-2.06629825e+00 -1.25014544e+00 -2.12023750e-01 1.56813121e+00
1.04723585e+00 -5.31428099e-01 2.09813222e-01 -2.61759311e-01
4.44490373e-01 2.35321075e-01 -6.11474276e-01 4.75326106e-02
-1.37193382e-01 -4.67676036e-02 2.91163206e-01 4.19475555e-01
-1.05889583e+00 1.17392302e+00 6.68171167e+00 9.79332268e-01
-8.96188200e-01 2.84572989e-01 8.04637253e-01 -1.80327699e-01
-2.41397589e-01 -5.17778397e-02 -1.07062769e+00 3.54526013e-01
6.33313775e-01 -2.12118357e-01 1.12543724e-01 6.67813540e-01
4.97656882e-01 -3.46935466e-02 -1.35288966e+00 1.02448380e+00
5.14677405e-01 -1.15590227e+00 4.18220609e-01 2.36246958e-01
5.69657207e-01 -5.31009078e-01 -2.33543217e-01 -3.98282856e-02
7.37458110e-01 -1.13256967e+00 4.99105573e-01 5.55047274e-01
1.10587394e+00 -4.39716429e-01 2.87753016e-01 -2.43052587e-01
-1.03079855e+00 -1.16955996e-01 -2.20234036e-01 -1.41592741e-01
-8.33070129e-02 3.65285993e-01 -5.94976246e-01 5.41477203e-01
8.78342390e-01 1.16864598e+00 -2.68848002e-01 1.52319402e-01
-9.86266017e-01 1.00832212e+00 -8.81800056e-02 8.36313516e-03
-2.18160555e-01 3.01802866e-02 3.48189175e-01 9.37585831e-01
-1.65235087e-01 8.01224113e-01 -4.65218037e-01 1.19771016e+00
-1.17244922e-01 2.71493178e-02 -5.77369690e-01 9.98293906e-02
-7.08065927e-02 1.06785989e+00 -2.51298666e-01 1.07938297e-01
-6.78686380e-01 1.00892293e+00 2.22917303e-01 4.07679439e-01
-1.39403117e+00 4.72414523e-01 1.24768496e+00 -2.19861865e-01
2.81203277e-02 1.27895057e-01 -2.25915480e-02 -1.16713834e+00
-1.44307449e-01 -8.80675912e-01 8.67995679e-01 -7.14335203e-01
-1.47174907e+00 1.44184336e-01 4.35736895e-01 -8.03034604e-01
3.61828506e-02 -7.48435795e-01 -3.62879813e-01 7.09185541e-01
-1.10750949e+00 -1.36172283e+00 -3.15975130e-01 1.03795397e+00
4.05286551e-01 -5.44853993e-02 8.07047963e-01 5.00729561e-01
-1.09932733e+00 8.39416504e-01 -8.43397856e-01 3.25799972e-01
7.14264274e-01 -6.14739120e-01 -2.93743134e-01 8.94500911e-01
1.39732257e-01 7.11610794e-01 5.21366537e-01 -1.03171325e+00
-1.33857107e+00 -1.00954926e+00 6.69481456e-01 -6.34214640e-01
1.05652606e+00 -4.89660740e-01 -6.03822529e-01 9.50907528e-01
1.31420389e-01 4.05652672e-02 1.03516793e+00 3.70727211e-01
-7.05000281e-01 -3.81323963e-01 -7.76965141e-01 7.26573646e-01
1.63044071e+00 -2.90404946e-01 -5.81731439e-01 1.93801835e-01
6.21422708e-01 5.40579781e-02 -6.83904171e-01 6.69552326e-01
8.83916140e-01 -7.18296170e-01 8.50142181e-01 -9.66033757e-01
1.00070441e+00 -2.45458037e-01 -1.53192595e-01 -1.05119574e+00
-4.41978693e-01 -6.71771467e-01 -1.77235767e-01 1.54178822e+00
3.08085144e-01 -4.66386944e-01 7.45065212e-01 9.57544148e-01
3.89924139e-01 -4.84544694e-01 -1.20947099e+00 -3.42856586e-01
-2.81632096e-01 -6.15268469e-01 7.10190415e-01 1.01648128e+00
-4.44547413e-03 5.08506835e-01 -7.34286606e-01 3.15599293e-01
8.34918678e-01 3.01813520e-02 7.38567650e-01 -8.88314128e-01
-1.55164152e-01 -4.22689140e-01 -4.92662638e-01 -7.28074610e-01
6.41308606e-01 -5.05573750e-01 5.38503416e-02 -1.04727173e+00
5.14895201e-01 -1.43827960e-01 -3.98733884e-01 1.15068865e+00
-6.27608180e-01 -3.73834595e-02 -2.26798654e-01 1.22016571e-01
-1.98253632e-01 9.92385328e-01 1.10850453e+00 -1.68842316e-01
2.53957547e-02 -3.22432876e-01 -6.60601199e-01 8.45198154e-01
1.82031602e-01 -4.78953809e-01 -5.31755209e-01 -2.37094089e-01
-2.75238100e-02 -9.50332209e-02 6.09685540e-01 -1.02080964e-01
-6.85066637e-03 -4.40745533e-01 2.48872057e-01 -3.54514033e-01
1.21219985e-01 -8.88023496e-01 6.79628402e-02 2.41933703e-01
-3.32635969e-01 -4.89885002e-01 2.44261757e-01 7.05265522e-01
-3.10618460e-01 3.30675513e-01 5.07207215e-01 2.98514009e-01
-7.72991061e-01 7.22466588e-01 -1.74624816e-01 -4.21047211e-01
1.19828367e+00 3.98275703e-02 -3.67030025e-01 -1.95062980e-01
-7.12214410e-01 2.43030548e-01 -5.12621820e-01 8.34800899e-01
6.48099244e-01 -1.62942898e+00 -8.51361811e-01 2.98307031e-01
3.48138869e-01 -4.08243150e-01 5.10457814e-01 1.00008249e+00
2.44116992e-01 2.51166075e-01 -1.14375353e-01 -6.37966931e-01
-1.72073936e+00 4.50452596e-01 3.65131736e-01 3.17493498e-01
-2.67614424e-01 1.29878712e+00 1.13476419e+00 1.06246278e-01
3.53035450e-01 -1.17185041e-01 -4.09815073e-01 4.86604035e-01
7.31891394e-01 1.56623751e-01 -3.48443985e-01 -6.96722686e-01
-5.46979487e-01 3.71741205e-01 1.55350193e-01 1.42845929e-01
1.13306201e+00 1.80925243e-02 -5.94281077e-01 2.82356262e-01
1.01651394e+00 -2.60070831e-01 -1.61830306e+00 -1.73626974e-01
-3.76719624e-01 -6.37243092e-01 -3.98051217e-02 -7.64419436e-01
-1.25320792e+00 5.93426526e-01 5.02084017e-01 -4.46012706e-01
1.38202643e+00 4.01324183e-01 1.33875504e-01 -2.98945874e-01
1.72100320e-01 -4.52288359e-01 1.29261181e-01 -3.19436118e-02
1.28608441e+00 -1.09457314e+00 1.66735098e-01 -1.11422753e+00
-4.13558245e-01 7.36593127e-01 8.74279678e-01 6.28557894e-03
1.01997948e+00 3.60640854e-01 7.61382841e-03 -4.28204626e-01
-9.96373951e-01 -2.35867172e-01 4.71226096e-01 2.80498356e-01
3.60590816e-01 2.75610000e-01 -5.96944749e-01 9.06406462e-01
1.08843908e-01 3.86065006e-01 4.38279994e-02 4.21087384e-01
2.58516908e-01 -9.57625508e-01 -2.12345213e-01 3.00565660e-01
-4.39963788e-01 -8.98642689e-02 -8.17928553e-01 7.76267469e-01
5.89487135e-01 8.65283012e-01 2.42704138e-01 -5.90253949e-01
4.53187913e-01 1.43238947e-01 7.22243249e-01 -3.40537548e-01
-1.26220621e-02 2.31426671e-01 7.88482353e-02 -1.15580153e+00
-8.12563181e-01 -1.20967388e+00 -1.25797033e+00 -2.17402577e-01
-9.90847424e-02 -3.53162616e-01 2.28822634e-01 1.20673645e+00
4.19541031e-01 6.47913694e-01 7.10106671e-01 -1.07422590e-01
6.08082488e-02 -1.05322587e+00 -5.33327281e-01 5.31351686e-01
2.19195150e-02 -1.34597647e+00 -4.96365279e-01 5.78367591e-01] | [13.59408187866211, 1.4924367666244507] |
6da968b2-5757-48e8-84d2-499583b32237 | stable-deep-mri-reconstruction-using | 2210.13834 | null | https://arxiv.org/abs/2210.13834v3 | https://arxiv.org/pdf/2210.13834v3.pdf | Stable Deep MRI Reconstruction using Generative Priors | Data-driven approaches recently achieved remarkable success in magnetic resonance imaging (MRI) reconstruction, but integration into clinical routine remains challenging due to a lack of generalizability and interpretability. In this paper, we address these challenges in a unified framework based on generative image priors. We propose a novel deep neural network based regularizer which is trained in a generative setting on reference magnitude images only. After training, the regularizer encodes higher-level domain statistics which we demonstrate by synthesizing images without data. Embedding the trained model in a classical variational approach yields high-quality reconstructions irrespective of the sub-sampling pattern. In addition, the model shows stable behavior when confronted with out-of-distribution data in the form of contrast variation. Furthermore, a probabilistic interpretation provides a distribution of reconstructions and hence allows uncertainty quantification. To reconstruct parallel MRI, we propose a fast algorithm to jointly estimate the image and the sensitivity maps. The results demonstrate competitive performance, on par with state-of-the-art end-to-end deep learning methods, while preserving the flexibility with respect to sub-sampling patterns and allowing for uncertainty quantification. | ['Thomas Pock', 'Florian Knoll', 'Martin Zach'] | 2022-10-25 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 3.97059947e-01 2.23497346e-01 9.07162502e-02 -4.21592802e-01
-1.33164442e+00 -3.37854505e-01 6.13005817e-01 8.69270191e-02
-6.37423396e-01 8.58180225e-01 2.99202472e-01 -2.17586849e-02
-3.72273207e-01 -5.21848381e-01 -9.32215333e-01 -1.15344179e+00
-1.04642130e-01 6.71970606e-01 1.79320052e-01 1.25599027e-01
-1.74918473e-01 4.38998699e-01 -1.04740393e+00 1.46217987e-01
8.13639700e-01 9.66968834e-01 3.39759678e-01 4.23825145e-01
4.87117648e-01 5.41184068e-01 -2.64172316e-01 -1.77215114e-01
1.54444769e-01 -5.03538370e-01 -7.40754068e-01 1.98669031e-01
8.50793272e-02 -5.58953047e-01 -5.28924584e-01 1.17489493e+00
7.05252171e-01 1.48183227e-01 9.48763967e-01 -6.02202296e-01
-6.33017540e-01 5.19400418e-01 -3.23881567e-01 2.26149917e-01
-9.66203511e-02 2.58810848e-01 8.58545899e-01 -7.90380716e-01
7.86776245e-01 7.47727036e-01 6.23852968e-01 4.77655172e-01
-1.91305518e+00 1.21932469e-01 -6.98147789e-02 4.35213819e-02
-1.24449599e+00 -3.48758280e-01 8.51442337e-01 -6.31866395e-01
5.72498083e-01 1.03657603e-01 3.39354306e-01 1.10017145e+00
5.74049592e-01 6.14488721e-01 1.44740152e+00 -3.25862020e-01
5.11291623e-01 -6.97542801e-02 -2.18954712e-01 4.38154936e-01
8.97341594e-02 1.34478524e-01 -2.40969270e-01 -2.02638164e-01
1.12147593e+00 -7.43587166e-02 -6.33560956e-01 -7.53209233e-01
-1.34572458e+00 8.80308867e-01 5.91276884e-01 3.00689429e-01
-7.05862284e-01 1.88633248e-01 3.48189294e-01 -2.51355410e-01
4.50687110e-01 2.59329587e-01 -8.10736939e-02 2.29407385e-01
-1.37795520e+00 1.89313591e-01 4.99185652e-01 5.36237240e-01
4.08435464e-01 1.86233550e-01 -4.46864218e-01 7.77633607e-01
3.71968031e-01 4.55331713e-01 4.92653370e-01 -1.12952089e+00
1.51311979e-03 -1.30627319e-01 3.94507274e-02 -7.42331028e-01
-6.21823251e-01 -8.51582050e-01 -1.13132513e+00 3.41570377e-01
4.76746291e-01 3.12567018e-02 -1.05426097e+00 1.96742070e+00
2.91064501e-01 -2.32131896e-03 -2.25058287e-01 1.08104753e+00
5.79768181e-01 3.52716804e-01 -7.59265572e-02 -4.66940105e-01
1.35480881e+00 -6.04932725e-01 -7.86528349e-01 -8.02832935e-03
1.62628070e-01 -4.47007656e-01 9.15308416e-01 5.56252122e-01
-1.40303302e+00 -2.42072895e-01 -1.07879853e+00 -3.28044631e-02
2.35383034e-01 4.83452715e-02 1.56680450e-01 5.64308524e-01
-1.27177846e+00 9.56672490e-01 -1.26481283e+00 1.36448875e-01
5.19509315e-01 3.33568871e-01 -2.19409660e-01 -2.87098829e-02
-1.11310363e+00 9.82441723e-01 3.21952879e-01 1.02948233e-01
-1.12771082e+00 -9.76536334e-01 -8.55487585e-01 -1.64550200e-01
2.42290989e-01 -7.97464371e-01 1.18718719e+00 -5.87992430e-01
-1.75091469e+00 8.27409267e-01 9.56865773e-02 -5.72357476e-01
9.84161913e-01 1.27082631e-01 -1.27470598e-01 6.43082261e-01
2.11454034e-01 7.15993226e-01 9.96519506e-01 -1.22524500e+00
1.92440644e-01 -3.39500248e-01 -2.53925413e-01 -1.13720246e-01
2.55620927e-01 -2.80675292e-01 -3.21225643e-01 -7.82258153e-01
2.21557394e-01 -9.34112847e-01 -4.72761393e-01 1.57619596e-01
-3.74042332e-01 4.35021281e-01 -5.80629744e-02 -8.81026208e-01
1.00961864e+00 -1.92615271e+00 4.20657009e-01 2.56931901e-01
4.76733446e-01 -1.12910479e-01 1.72483966e-01 9.43310559e-02
-2.23269016e-01 -2.24572256e-01 -9.36048031e-01 -3.43768686e-01
1.03048928e-01 3.63497525e-01 -9.79177505e-02 9.08123672e-01
3.45506132e-01 8.94971669e-01 -1.03409624e+00 -4.94393080e-01
4.56853151e-01 8.06786656e-01 -7.77815998e-01 1.89378664e-01
-1.31092086e-01 1.09666216e+00 -2.98379123e-01 1.42061323e-01
7.96136081e-01 -5.62989950e-01 4.95082319e-01 -5.10494709e-01
5.63016422e-02 7.82306567e-02 -9.09261882e-01 2.09323931e+00
-6.10541463e-01 3.02011728e-01 4.31944042e-01 -1.21631026e+00
4.18276936e-01 4.11497891e-01 6.82124853e-01 -5.64563334e-01
2.38096073e-01 5.05971849e-01 1.75221384e-01 -3.40112209e-01
1.61129907e-02 -7.73625016e-01 1.26670450e-01 3.23097289e-01
3.89849752e-01 -3.31358045e-01 -1.15490360e-02 1.85067207e-01
9.16735291e-01 3.32258433e-01 2.29948759e-01 -6.46786451e-01
4.07932580e-01 -4.57756460e-01 3.11795235e-01 7.80985057e-01
-1.74848527e-01 1.21752226e+00 5.33699095e-01 -1.82850569e-01
-1.24557126e+00 -1.31819570e+00 -8.62694681e-01 4.93474811e-01
-2.52707064e-01 5.28479693e-04 -9.12680686e-01 -4.07522321e-01
-2.31464908e-01 7.32374251e-01 -7.47661650e-01 -1.47286588e-02
-5.25476873e-01 -1.21717095e+00 1.67790532e-01 4.59294170e-01
2.82556325e-01 -7.99328804e-01 -7.39684761e-01 4.69587296e-01
-3.69269073e-01 -1.27798069e+00 -2.71187752e-01 3.66825700e-01
-9.61335540e-01 -6.86846614e-01 -1.12809622e+00 -2.16407865e-01
6.42119348e-01 -3.47705752e-01 1.37546635e+00 -3.54451537e-01
-3.61276448e-01 4.33989793e-01 1.25304982e-01 7.39254281e-02
-5.91404200e-01 -2.39012107e-01 -5.57903275e-02 -2.70676315e-02
-5.16752899e-01 -1.02785599e+00 -1.07800353e+00 7.83163980e-02
-1.21420741e+00 -1.29190743e-01 6.38448596e-01 1.25803328e+00
9.82415795e-01 -2.02979431e-01 3.95404607e-01 -8.34603250e-01
2.67679334e-01 -5.52001655e-01 -4.87564683e-01 5.07008433e-02
-6.48212731e-01 5.19117892e-01 4.60424662e-01 -1.29841492e-01
-1.10099828e+00 -6.09756038e-02 -4.13327426e-01 -3.63018364e-01
6.27402822e-03 4.78100747e-01 7.41942823e-02 -1.40726522e-01
8.52000058e-01 3.47606957e-01 3.85766208e-01 -3.71552736e-01
3.78130555e-01 1.60903037e-01 7.47499466e-01 -8.26941609e-01
3.40466589e-01 7.83361554e-01 3.73765588e-01 -6.08175099e-01
-7.75667608e-01 -6.75482228e-02 -5.84372938e-01 -2.29156598e-01
9.87689495e-01 -7.38858104e-01 -4.20464277e-01 4.21173692e-01
-1.00370371e+00 -2.35005975e-01 -5.45246303e-01 6.27245545e-01
-9.88954127e-01 6.12101555e-01 -7.24458337e-01 -5.52044988e-01
-3.34868491e-01 -1.69412076e+00 1.17811203e+00 -2.77722925e-01
3.64742875e-02 -1.12215602e+00 4.22542244e-02 1.65988401e-01
6.24076128e-01 6.78141356e-01 8.59680951e-01 -3.56275707e-01
-6.59726441e-01 3.83084118e-02 -2.08925217e-01 5.42920411e-01
-5.03006987e-02 -5.10792971e-01 -9.80067551e-01 -3.36701393e-01
4.65998590e-01 -2.95747519e-01 9.08965111e-01 1.15859950e+00
1.31670201e+00 -1.23553298e-01 3.57215367e-02 8.00198317e-01
1.46927643e+00 -3.21801305e-01 7.03261435e-01 8.37539658e-02
4.63538677e-01 4.39519525e-01 -2.18146309e-01 4.45591658e-01
7.90700614e-02 7.84892321e-01 4.52852637e-01 5.72784990e-02
-2.96675205e-01 2.67529227e-02 -1.22611187e-01 6.08634889e-01
-1.12742938e-01 1.07699141e-01 -7.76260972e-01 4.75994915e-01
-1.70383453e+00 -7.39523411e-01 9.02904570e-03 2.24011135e+00
1.11629081e+00 1.55415043e-01 4.21259403e-02 -3.10237240e-02
4.28580195e-01 2.83362806e-01 -5.76983571e-01 1.34330764e-01
-4.48497161e-02 3.72187108e-01 5.72833180e-01 6.57251954e-01
-9.21819091e-01 1.72180951e-01 6.82048559e+00 1.04219711e+00
-1.27744949e+00 5.71249783e-01 9.60904956e-01 -6.05593733e-02
-6.79308474e-01 -2.49138087e-01 -2.01456159e-01 4.98337954e-01
9.15538132e-01 2.50765324e-01 4.84827787e-01 5.03975153e-01
1.40106589e-01 -1.74583599e-01 -1.09455228e+00 9.27496910e-01
-4.51738611e-02 -1.46811008e+00 -1.57091200e-01 -4.70248237e-02
7.18987167e-01 2.43265465e-01 2.39440933e-01 -8.82350653e-02
-7.73587972e-02 -1.02406073e+00 9.11705911e-01 6.79382682e-01
1.08974051e+00 -4.80252087e-01 5.28481066e-01 2.19762251e-01
-4.94520545e-01 3.81748766e-01 -1.85715020e-01 4.45005894e-01
4.86166984e-01 1.03182566e+00 -5.42106330e-01 6.97411180e-01
4.42631215e-01 3.82548153e-01 1.47042526e-02 8.24972868e-01
-2.33036995e-01 4.85484481e-01 -3.97022367e-01 5.40034950e-01
2.26354837e-01 -3.86498511e-01 7.53278375e-01 1.18227863e+00
2.40389705e-01 8.31125900e-02 -1.45316869e-03 1.29268026e+00
-8.49074051e-02 -1.72510564e-01 -1.07389212e-01 3.02522451e-01
-6.13436364e-02 1.22172666e+00 -8.66264760e-01 -5.96459806e-02
-1.36261642e-01 8.92993033e-01 2.83604503e-01 4.91840005e-01
-7.42364764e-01 1.66134879e-01 3.00245881e-02 4.88789231e-01
3.13846499e-01 -1.93582967e-01 -2.81664222e-01 -1.19547629e+00
2.45146006e-01 -6.33082092e-01 2.35383093e-01 -6.82346463e-01
-1.37273908e+00 7.91628361e-01 4.25678313e-01 -9.51427996e-01
-5.73145330e-01 -7.81443059e-01 -3.65636468e-01 9.62856591e-01
-1.61320448e+00 -9.62184489e-01 1.06683960e-02 3.72704804e-01
1.20662391e-01 2.42971361e-01 7.20687687e-01 3.95485401e-01
-2.32481286e-01 3.82978559e-01 4.52706814e-01 -1.74396843e-01
4.54348236e-01 -1.44381523e+00 -1.26980469e-01 8.46698582e-01
-8.26599821e-02 6.39061153e-01 9.70133781e-01 -3.75479668e-01
-1.14646745e+00 -8.02731037e-01 3.03554565e-01 -2.43896201e-01
7.39633381e-01 -1.97363853e-01 -1.00198054e+00 5.77130020e-01
2.29692653e-01 5.76356173e-01 4.52083051e-01 -1.76090017e-01
-9.61277410e-02 -1.70486665e-03 -1.46430695e+00 2.83000022e-01
6.84551597e-01 -5.17207444e-01 -4.90264863e-01 4.30697769e-01
4.98702854e-01 -7.53983676e-01 -1.04583299e+00 4.45654154e-01
3.76287967e-01 -1.21649301e+00 1.08361900e+00 -2.25091025e-01
5.42634964e-01 -2.17721298e-01 -2.07895473e-01 -1.39175701e+00
-3.49929065e-01 -7.60905623e-01 -5.11216149e-02 6.94480598e-01
2.36499995e-01 -5.72949886e-01 5.45262158e-01 6.31269872e-01
-3.72565717e-01 -9.33202028e-01 -1.37133563e+00 -6.10965252e-01
3.40775907e-01 -5.42859972e-01 9.68275517e-02 5.89383900e-01
-2.35363424e-01 -1.74898524e-02 -4.24520344e-01 1.19997717e-01
1.15466917e+00 -4.93623763e-02 -8.25813487e-02 -8.24485898e-01
-7.74940610e-01 -5.04080653e-01 -3.75934780e-01 -9.32692468e-01
5.98193482e-02 -9.99231994e-01 1.96950957e-01 -1.33151066e+00
4.08676416e-01 -3.23980987e-01 -4.44609433e-01 -3.38182668e-03
-9.18806121e-02 4.05074865e-01 -1.22546680e-01 2.86330223e-01
-3.10047746e-01 7.30722666e-01 1.42014945e+00 3.92584540e-02
1.73129737e-01 -1.58245847e-01 -4.47279871e-01 5.66287518e-01
4.71886754e-01 -3.97505045e-01 -3.96678209e-01 -3.41333002e-01
1.49931833e-01 3.72844279e-01 7.49889076e-01 -9.79721248e-01
-8.50782245e-02 2.45696783e-01 5.18795967e-01 -3.18466395e-01
2.52914518e-01 -6.95094049e-01 3.47871989e-01 3.68183076e-01
-3.30709428e-01 -3.51313651e-01 2.07606852e-01 5.91885984e-01
-1.30367815e-01 -3.79303247e-01 1.23846030e+00 -1.25242829e-01
-3.30637582e-02 3.05539638e-01 -2.48842090e-01 3.19972187e-01
5.25975227e-01 1.66041300e-01 2.02242702e-01 -4.11490411e-01
-1.16165924e+00 -2.08394274e-01 3.95736456e-01 -1.80159166e-01
4.95308012e-01 -1.36787331e+00 -8.76866579e-01 5.21272868e-02
-1.19917691e-01 5.27379252e-02 7.59378850e-01 1.37115705e+00
-5.67705035e-01 3.08520228e-01 -2.62324393e-01 -9.89345610e-01
-3.58540028e-01 4.96700138e-01 7.39253223e-01 -5.78893542e-01
-7.98473418e-01 6.53714597e-01 4.23647434e-01 -3.65083158e-01
-5.72511405e-02 -5.24933636e-01 1.68221354e-01 -1.96636960e-01
3.67314041e-01 6.24167845e-02 3.56146485e-01 -6.13280833e-01
-3.94445181e-01 4.03883994e-01 1.10951826e-01 -4.63725865e-01
1.51586425e+00 -2.86509413e-02 -4.57174564e-03 3.34079891e-01
1.26720107e+00 -1.21676609e-01 -1.68011284e+00 -4.17547345e-01
-3.29854876e-01 -1.92102611e-01 5.59615731e-01 -8.80454302e-01
-1.24243164e+00 9.54023242e-01 6.72902167e-01 -5.55315576e-02
9.35768068e-01 -2.51523014e-02 5.03875792e-01 -1.59692675e-01
3.72087538e-01 -8.53631020e-01 -3.33569758e-02 7.57768610e-03
1.14266860e+00 -1.30055094e+00 2.21484855e-01 -2.72336781e-01
-5.78056276e-01 9.90274131e-01 -2.40439549e-01 -3.66034418e-01
8.36253941e-01 3.51988226e-01 -2.10608125e-01 -2.58896559e-01
-2.75146812e-01 7.41611496e-02 5.77446938e-01 6.96632624e-01
4.92876619e-01 1.94647118e-01 -3.54552746e-01 5.32411456e-01
2.31834754e-01 1.08073363e-02 3.48760158e-01 7.30443954e-01
-1.28033370e-01 -1.02714527e+00 -2.68606722e-01 2.90859103e-01
-7.47575402e-01 -2.57108241e-01 4.74231929e-01 6.89216077e-01
-1.44030929e-01 4.70003515e-01 -1.30218893e-01 3.14463228e-01
2.13369206e-01 -5.08130305e-02 9.43043888e-01 -3.91330570e-01
-2.24757388e-01 4.81102377e-01 -1.82386503e-01 -6.17906630e-01
-4.44706678e-01 -8.69844615e-01 -1.35703826e+00 1.10900901e-01
1.07372263e-02 1.27524421e-01 5.75692654e-01 8.93180907e-01
2.82222867e-01 7.79747307e-01 3.68325055e-01 -1.10345685e+00
-9.68173027e-01 -8.76022637e-01 -7.27738738e-01 5.46310484e-01
4.63923812e-01 -8.03970158e-01 -3.48304927e-01 -6.47645369e-02] | [13.498917579650879, -2.3579514026641846] |
e9783db5-f5ef-45c8-94b5-b3af06032929 | aspect-category-opinion-sentiment-quadruple | null | null | https://aclanthology.org/2021.acl-long.29 | https://aclanthology.org/2021.acl-long.29.pdf | Aspect-Category-Opinion-Sentiment Quadruple Extraction with Implicit Aspects and Opinions | Product reviews contain a large number of implicit aspects and implicit opinions. However, most of the existing studies in aspect-based sentiment analysis ignored this problem. In this work, we introduce a new task, named Aspect-Category-Opinion-Sentiment (ACOS) Quadruple Extraction, with the goal to extract all aspect-category-opinion-sentiment quadruples in a review sentence and provide full support for aspect-based sentiment analysis with implicit aspects and opinions. We furthermore construct two new datasets, Restaurant-ACOS and Laptop-ACOS, for this new task, both of which contain the annotations of not only aspect-category-opinion-sentiment quadruples but also implicit aspects and opinions. The former is an extension of the SemEval Restaurant dataset; the latter is a newly collected and annotated Laptop dataset, twice the size of the SemEval Laptop dataset. We finally benchmark the task with four baseline systems. Experiments demonstrate the feasibility of the new task and its effectiveness in extracting and describing implicit aspects and implicit opinions. The two datasets and source code of four systems are publicly released at \url{https://github.com/NUSTM/ACOS}. | ['Jianfei Yu', 'Rui Xia', 'Hongjie Cai'] | 2021-08-01 | null | null | null | acl-2021-5 | ['aspect-category-opinion-sentiment-quadruple'] | ['natural-language-processing'] | [-3.11481375e-02 9.39663500e-02 -4.36104119e-01 -8.96245062e-01
-9.65649068e-01 -8.29766631e-01 5.07955015e-01 3.97013307e-01
-2.61496395e-01 6.62575305e-01 3.07295293e-01 -3.27033997e-01
5.31074226e-01 -6.40172005e-01 -4.95872796e-01 -5.01592278e-01
4.10436541e-01 1.67286783e-01 -2.42893938e-02 -5.92306435e-01
5.30016065e-01 -2.50910670e-01 -1.57260334e+00 8.16482902e-01
8.23781848e-01 1.23634410e+00 -2.13864714e-01 5.23761272e-01
-3.56885523e-01 6.14550471e-01 -6.79195821e-01 -9.56604600e-01
1.53397888e-01 -9.63754803e-02 -4.37025160e-01 2.02100411e-01
4.05114353e-01 1.98282614e-01 5.78116357e-01 9.68623936e-01
3.20448905e-01 -4.09978889e-02 5.87716043e-01 -1.29584146e+00
-9.90622342e-01 5.30233443e-01 -7.95826554e-01 -1.21826805e-01
3.37796688e-01 -8.67847875e-02 1.57127142e+00 -1.26032567e+00
3.29576254e-01 9.56541359e-01 7.91380823e-01 2.60178983e-01
-5.78236878e-01 -5.47409177e-01 8.45427096e-01 -1.97769091e-01
-9.32058215e-01 -2.67062843e-01 6.71412826e-01 -1.84404105e-01
1.08344531e+00 3.63749713e-01 8.75567198e-01 7.82736063e-01
1.03052355e-01 1.24189639e+00 1.34072471e+00 -2.65400231e-01
3.60078961e-01 8.70942295e-01 9.87664878e-01 3.87988746e-01
6.08800828e-01 -6.04601026e-01 -6.48331702e-01 -2.41065323e-01
3.62120718e-02 1.13972053e-01 -1.75795868e-01 -1.45889610e-01
-9.78282213e-01 8.93274248e-01 -1.91535074e-02 1.53963029e-01
-3.07954341e-01 -3.09929967e-01 6.57360196e-01 4.07429397e-01
1.20120263e+00 5.13747573e-01 -1.41628420e+00 -1.19142093e-01
-4.11160350e-01 3.70095581e-01 1.35228336e+00 1.38828743e+00
1.07469678e+00 -2.46117234e-01 6.36728555e-02 9.42277491e-01
3.92767131e-01 8.14658344e-01 4.74404931e-01 -4.39336658e-01
5.43021202e-01 1.13416266e+00 1.83596626e-01 -1.04767883e+00
-2.71447092e-01 -3.76825690e-01 -5.70849180e-01 -2.65384912e-01
-4.11394835e-01 -3.89629751e-01 -9.48554397e-01 1.20442343e+00
5.81362605e-01 -2.78788298e-01 1.63499907e-01 6.25647128e-01
1.30534720e+00 5.14217675e-01 1.33806512e-01 -1.40658647e-01
1.88278747e+00 -1.47343194e+00 -9.81447279e-01 -5.63625276e-01
8.65055382e-01 -9.89550173e-01 1.34579933e+00 4.19741511e-01
-6.26269102e-01 -3.64595532e-01 -1.27045977e+00 -1.97577387e-01
-1.02333951e+00 4.08655018e-01 1.14233875e+00 5.92757404e-01
-8.47466648e-01 -4.98275533e-02 -4.39717263e-01 -1.04055824e-02
2.83574969e-01 2.12501451e-01 -4.25036699e-01 2.25501433e-02
-1.21141815e+00 5.81953824e-01 -1.58665761e-01 1.74472138e-01
-1.47982925e-01 -7.65729129e-01 -1.48218095e+00 -2.08456084e-01
6.44149661e-01 -7.47301579e-01 1.63671267e+00 -1.53735638e+00
-1.09993923e+00 9.17900443e-01 -6.04287148e-01 1.42371565e-01
-4.06400144e-01 -5.83011031e-01 -7.39985704e-01 -2.18102321e-01
7.14041650e-01 3.36046487e-01 7.70551324e-01 -1.41143572e+00
-9.41666126e-01 -6.63983107e-01 5.60187399e-01 3.49819869e-01
-3.93047154e-01 1.36011336e-02 -6.32866859e-01 -6.54056847e-01
-2.53399253e-01 -1.08612740e+00 -4.65351045e-01 -5.38946688e-01
-5.94269931e-01 -4.05474603e-01 7.17643738e-01 -3.41850877e-01
1.35310423e+00 -1.96169162e+00 -4.05830890e-01 -8.08775350e-02
1.31028950e-01 3.28725785e-01 -4.34260547e-01 2.68176615e-01
-1.30953923e-01 2.41599977e-01 -4.66513336e-01 -6.62346601e-01
1.07740201e-01 5.13477772e-02 -4.08735752e-01 -1.07767291e-01
4.54589397e-01 1.00303018e+00 -9.56284404e-01 -2.32558653e-01
-3.56700957e-01 3.18611652e-01 -4.96163815e-01 1.11987717e-01
-4.61747378e-01 1.30514562e-01 -9.57713604e-01 1.01320457e+00
8.68364573e-01 -1.16142236e-01 -3.23940851e-02 -4.06217217e-01
-2.23307252e-01 7.36981690e-01 -8.58223140e-01 1.30423772e+00
-6.91776514e-01 3.22135627e-01 4.57047336e-02 -5.73523760e-01
9.04639721e-01 3.02639753e-01 4.28765357e-01 -5.89890718e-01
1.48405313e-01 3.44102800e-01 -4.82088119e-01 -4.42858905e-01
1.23109066e+00 -3.25517654e-01 -5.86833000e-01 6.81308448e-01
4.16010525e-03 -7.57017076e-01 7.09777057e-01 2.89099962e-01
7.79855251e-01 1.27910018e-01 6.39830410e-01 -2.73698539e-01
8.52694154e-01 2.53104120e-01 6.10257924e-01 1.98230401e-01
-1.28639996e-01 6.94117546e-01 8.20598304e-01 -5.62489033e-01
-7.77308106e-01 -4.93990928e-01 -1.57471616e-02 1.12535584e+00
8.37738737e-02 -1.20502877e+00 -4.80249196e-01 -1.38375878e+00
-8.73883516e-02 7.95438290e-01 -9.26301777e-01 2.65588850e-01
-2.69025207e-01 -8.88220251e-01 -1.80817649e-01 4.82848048e-01
3.39132577e-01 -1.09395587e+00 3.87870558e-02 -1.94869772e-01
-4.16421413e-01 -1.31205034e+00 -8.49657238e-01 1.24885775e-01
-5.46696842e-01 -1.20095801e+00 -4.66736823e-01 -7.98664570e-01
8.63470554e-01 5.48140228e-01 1.87953460e+00 -1.22234470e-03
2.42811978e-01 5.08845747e-01 -8.29919100e-01 -1.12584043e+00
1.37885466e-01 3.66760105e-01 -1.68122783e-01 6.28598407e-02
1.22472227e+00 -3.97075206e-01 -7.98714459e-01 3.13438296e-01
-1.01137888e+00 -1.68129191e-01 6.69426143e-01 4.42002147e-01
1.14556897e+00 -1.49438992e-01 6.00494742e-01 -1.72242689e+00
8.20501387e-01 -6.54592812e-01 -1.40884057e-01 -7.50236213e-02
-6.76191986e-01 -3.31563979e-01 6.42607450e-01 -8.88276473e-02
-1.01844478e+00 -1.10384658e-01 -4.39618707e-01 3.73683751e-01
-7.12197348e-02 9.33505774e-01 -3.76434952e-01 5.26935577e-01
1.40689045e-01 5.98548204e-02 -3.75269532e-01 -3.22921187e-01
5.06978929e-01 9.19047594e-01 -5.79710715e-02 -3.02225947e-01
4.29559559e-01 4.90888864e-01 -4.80817497e-01 -8.60512316e-01
-1.93927038e+00 -1.08585238e+00 -6.05558574e-01 1.50854856e-01
6.43427670e-01 -1.33224714e+00 -1.77104801e-01 4.93105352e-01
-1.12598252e+00 9.84110087e-02 -5.32861829e-01 1.51643351e-01
-2.96038389e-01 3.08158696e-01 -3.69684011e-01 -7.19754696e-01
-8.74799907e-01 -9.54701245e-01 1.57952404e+00 1.93438455e-01
-4.44447041e-01 -1.05157518e+00 3.36703807e-01 5.69623709e-01
2.84809530e-01 9.13113430e-02 6.59204185e-01 -8.74654531e-01
-1.35841697e-01 -5.12400448e-01 -3.15698013e-02 8.85668814e-01
3.91318381e-01 -1.54778557e-02 -9.08288598e-01 9.13452208e-02
3.74866039e-01 -3.48638356e-01 9.08224761e-01 1.42309070e-01
6.57301843e-01 -4.48709786e-01 -2.89344210e-02 2.01290831e-01
1.25853288e+00 -5.60141914e-02 5.18744707e-01 5.47503471e-01
6.79966569e-01 6.54411077e-01 1.34946394e+00 3.06870818e-01
9.23089206e-01 3.59209806e-01 1.81117862e-01 -1.12388581e-01
2.16531888e-01 8.13721865e-03 6.72683120e-01 1.72833025e+00
-6.83770701e-03 -1.04375668e-01 -4.15110260e-01 9.37638879e-01
-1.86571312e+00 -5.55916846e-01 -5.24914861e-01 1.67457986e+00
9.62618411e-01 8.41842443e-02 -2.52106003e-02 -9.43418667e-02
2.95079529e-01 5.85248232e-01 -5.33985078e-01 -8.26573730e-01
-2.30004564e-01 2.50475556e-01 -7.29361996e-02 3.25625032e-01
-1.52759433e+00 8.18551958e-01 5.41508532e+00 5.44257641e-01
-6.10925376e-01 1.84225768e-01 7.78651357e-01 1.20671697e-01
-7.88881004e-01 9.65698510e-02 -1.23936427e+00 1.71639234e-01
8.58077645e-01 -6.00129440e-02 -4.72967774e-01 1.27357674e+00
2.34035198e-02 -2.41582155e-01 -9.06609297e-01 4.76649374e-01
5.72913826e-01 -6.77713156e-01 2.21018463e-01 -3.50433975e-01
1.11930132e+00 -7.75996298e-02 2.61534750e-01 7.32483387e-01
1.11562219e-02 -6.19973123e-01 5.43205261e-01 1.54978082e-01
7.38420606e-01 -8.58461499e-01 1.31821251e+00 2.43824348e-02
-1.54394054e+00 3.60967457e-01 -2.66923070e-01 -1.01260185e-01
2.99788684e-01 1.20260501e+00 -1.49950847e-01 5.65801382e-01
8.76147449e-01 1.45606995e+00 -6.17686808e-01 4.74164695e-01
-7.56728172e-01 4.44423616e-01 1.17798083e-01 -3.54187518e-01
4.95630741e-01 -5.02791822e-01 5.23010790e-01 1.34037304e+00
1.28352880e-01 1.19105861e-01 -4.34514433e-02 4.05089825e-01
-6.06848598e-02 3.82719308e-01 -6.98138654e-01 -2.41326883e-01
-7.02074245e-02 1.88899708e+00 -5.29923499e-01 -6.34093225e-01
-8.51030946e-01 6.75507784e-01 4.29880954e-02 2.68772870e-01
-6.49710894e-01 -6.70106053e-01 1.03551459e+00 -2.89427135e-02
6.41203821e-01 5.91265559e-02 -4.21112955e-01 -1.64599693e+00
4.30146039e-01 -1.30460250e+00 2.02649295e-01 -6.90423548e-01
-1.60958982e+00 1.00012910e+00 -2.27572888e-01 -1.35468054e+00
-2.74999496e-02 -8.54095936e-01 -4.70774680e-01 6.85586154e-01
-2.01752448e+00 -1.37085438e+00 -1.98015630e-01 4.20276225e-01
1.02332413e+00 -5.87446503e-02 9.21443224e-01 3.65077108e-01
-3.47699672e-01 4.15795386e-01 -4.17093188e-01 -3.93265337e-02
9.01519716e-01 -1.46308601e+00 6.55635118e-01 3.79789770e-01
-1.05876118e-01 1.03105235e+00 6.24817967e-01 -6.40162230e-01
-1.32721400e+00 -1.26030064e+00 1.22130847e+00 -1.12140775e+00
9.82767940e-01 -5.16611934e-01 -6.44219697e-01 9.33956087e-01
4.34215426e-01 -2.66836762e-01 1.26265633e+00 5.48841774e-01
-5.97086072e-01 -9.86379236e-02 -7.63356090e-01 6.06284320e-01
5.49777806e-01 -5.06808400e-01 -7.13926256e-01 3.19824398e-01
1.17598796e+00 -1.21204950e-01 -6.74101412e-01 7.48397112e-01
5.57343304e-01 -9.53844666e-01 5.56604505e-01 -4.57606584e-01
8.98927510e-01 -4.36420590e-01 -2.12529063e-01 -1.50501394e+00
1.98950097e-01 -1.26254365e-01 -1.78677306e-01 1.50750816e+00
1.11363506e+00 -7.31286526e-01 6.87468886e-01 5.66042900e-01
-2.99168348e-01 -1.32580411e+00 -1.97626263e-01 -1.94480211e-01
-1.25929952e-01 -6.33282483e-01 6.98752761e-01 8.16288531e-01
4.99424674e-02 1.13120461e+00 1.55576607e-02 -1.33563131e-01
2.46330142e-01 6.16407037e-01 9.14955199e-01 -6.93934619e-01
-2.07651764e-01 -9.70533118e-02 -1.29132867e-01 -1.11512733e+00
9.08716545e-02 -5.17778575e-01 1.27835244e-01 -1.61845446e+00
4.79327768e-01 -3.33712280e-01 -2.55981177e-01 5.88607311e-01
-5.89490950e-01 5.32279253e-01 -1.27517283e-01 -1.25925019e-01
-1.13382828e+00 7.13369310e-01 1.36072731e+00 -3.70961457e-01
-1.60103589e-01 5.31512380e-01 -1.56914008e+00 1.09145856e+00
7.74250329e-01 -4.51577485e-01 -5.33969641e-01 -2.35425517e-01
9.43939328e-01 -7.53557324e-01 -3.92442197e-01 -2.28901953e-01
-1.76796868e-01 5.44803441e-02 -1.08486831e-01 -1.04974627e+00
3.84385169e-01 -8.20584297e-01 -4.86598879e-01 -1.95240915e-01
-1.10513486e-01 4.40362036e-01 3.06499690e-01 5.08526206e-01
-8.18927169e-01 -1.80860266e-01 1.20736763e-01 -3.78232419e-01
-7.66975641e-01 4.08580661e-01 -4.29375559e-01 2.09240839e-01
8.59413505e-01 2.10324734e-01 -5.40010810e-01 -4.80815887e-01
-4.97085720e-01 3.94261152e-01 2.48600587e-01 6.88053668e-01
5.76774895e-01 -1.17708576e+00 -5.35300374e-01 5.85341640e-02
6.48730755e-01 2.08024323e-01 1.79172471e-01 9.78470922e-01
-9.55066532e-02 5.27187526e-01 3.96203011e-01 -1.98099568e-01
-1.42156017e+00 5.14032245e-01 -1.47391960e-01 -6.66054785e-01
-2.72861104e-02 7.39814162e-01 5.55691183e-01 -1.21206784e+00
-3.54623199e-01 -4.63829905e-01 -7.59792686e-01 3.72647464e-01
6.89966559e-01 -2.06978291e-01 2.03065544e-01 -7.62216687e-01
-3.14231932e-01 8.25933278e-01 -4.15661901e-01 1.23489253e-01
1.56502688e+00 -3.20333004e-01 -7.72770345e-01 8.37656975e-01
1.29194474e+00 5.36062300e-01 -4.77260470e-01 -4.89935651e-02
1.37768416e-02 -2.45368928e-01 -3.06571662e-01 -9.48040307e-01
-1.12961948e+00 6.56111479e-01 -1.90865248e-01 3.77469629e-01
1.23192179e+00 6.52542859e-02 1.02046096e+00 3.75350833e-01
2.35325962e-01 -1.12314248e+00 7.93543160e-02 9.60045636e-01
8.87687445e-01 -1.56643224e+00 3.85203719e-01 -7.65281439e-01
-1.26149178e+00 6.36087120e-01 7.79865801e-01 -7.21405521e-02
1.10864460e+00 2.68101871e-01 6.28150761e-01 -5.98710895e-01
-1.11710775e+00 -3.83216113e-01 5.02321541e-01 4.52514082e-01
8.83689106e-01 1.62982047e-01 -5.72511256e-01 1.35930991e+00
-4.44165170e-01 -2.35605925e-01 4.79705036e-01 1.14840424e+00
-2.31179476e-01 -1.24986672e+00 1.91273749e-01 6.27419829e-01
-7.57350028e-01 -5.44930339e-01 -7.75740325e-01 6.44654095e-01
1.66302502e-01 1.29019570e+00 -2.75601447e-01 -2.64047801e-01
7.34127402e-01 -1.47483855e-01 -2.30377227e-01 -1.07732952e+00
-9.97089446e-01 -6.67455792e-02 6.27889693e-01 -6.07603967e-01
-7.93117464e-01 -4.42576021e-01 -9.90303040e-01 5.00012599e-02
-5.59908748e-01 6.99383199e-01 9.76384282e-01 9.14032102e-01
5.40183127e-01 4.63969737e-01 8.25975537e-01 -6.14672661e-01
-1.37889117e-01 -1.01412237e+00 -7.77204871e-01 4.58885938e-01
3.97773623e-01 -2.53027141e-01 -6.06905043e-01 1.87662914e-02] | [11.4527006149292, 6.667998790740967] |
446c711d-c4e3-4e50-b668-3bdcf44d374e | qr-mix-distributional-value-function | 2009.04197 | null | https://arxiv.org/abs/2009.04197v5 | https://arxiv.org/pdf/2009.04197v5.pdf | QR-MIX: Distributional Value Function Factorisation for Cooperative Multi-Agent Reinforcement Learning | In Cooperative Multi-Agent Reinforcement Learning (MARL) and under the setting of Centralized Training with Decentralized Execution (CTDE), agents observe and interact with their environment locally and independently. With local observation and random sampling, the randomness in rewards and observations leads to randomness in long-term returns. Existing methods such as Value Decomposition Network (VDN) and QMIX estimate the value of long-term returns as a scalar that does not contain the information of randomness. Our proposed model QR-MIX introduces quantile regression, modeling joint state-action values as a distribution, combining QMIX with Implicit Quantile Network (IQN). However, the monotonicity in QMIX limits the expression of joint state-action value distribution and may lead to incorrect estimation results in non-monotonic cases. Therefore, we proposed a flexible loss function to approximate the monotonicity found in QMIX. Our model is not only more tolerant of the randomness of returns, but also more tolerant of the randomness of monotonic constraints. The experimental results demonstrate that QR-MIX outperforms the previous state-of-the-art method QMIX in the StarCraft Multi-Agent Challenge (SMAC) environment. | ['Siyue Hu', 'Shih-wei Liao', 'Jian Hu', 'Haibin Wu', 'Seth Austin Harding'] | 2020-09-09 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-6.79233789e-01 7.83959925e-02 -4.01862383e-01 -2.15785772e-01
-7.12680519e-01 -6.01422131e-01 5.49931824e-01 1.85071036e-01
-7.72503853e-01 1.35388267e+00 1.18941426e-01 7.16949180e-02
-5.89617908e-01 -9.54869330e-01 -7.87061632e-01 -8.76360118e-01
-6.53334141e-01 8.17434311e-01 5.73513508e-02 -3.66426408e-01
-1.30055528e-02 1.01373699e-02 -1.23072445e+00 -5.48163578e-02
9.00724530e-01 1.00917137e+00 9.92496684e-02 5.72994590e-01
1.98996842e-01 1.54871476e+00 -8.94863129e-01 -1.51740760e-01
5.06984532e-01 -3.16384494e-01 -4.27836865e-01 -2.45391250e-01
-2.59588897e-01 -7.82141924e-01 -3.33516568e-01 9.19747412e-01
5.54860950e-01 5.32706320e-01 4.31747347e-01 -1.73750937e+00
-4.27790135e-01 1.15590978e+00 -4.77024794e-01 -1.74439698e-01
-1.37341097e-01 5.03793597e-01 1.09142697e+00 -9.06109286e-04
3.92372996e-01 1.25614715e+00 3.71971816e-01 4.44928735e-01
-1.05223012e+00 -3.39271635e-01 5.02676666e-01 1.21210739e-01
-9.46705639e-01 1.83068931e-01 3.99939865e-01 -1.50844023e-01
9.23639238e-01 -1.48009747e-01 7.08381534e-01 9.45285916e-01
3.46657485e-01 1.13177538e+00 1.15156651e+00 1.22013226e-01
7.87622154e-01 -1.91683948e-01 -3.09735507e-01 4.54699934e-01
1.24396928e-01 6.45074308e-01 -6.65157199e-01 -4.59111363e-01
9.67066288e-01 2.03674793e-01 1.17737360e-01 -5.98554552e-01
-1.26081765e+00 8.29669237e-01 2.86732137e-01 -1.97679937e-01
-7.33090103e-01 8.59529138e-01 6.23751283e-01 8.31480026e-01
2.54814059e-01 4.76083517e-01 -7.58927345e-01 -5.79917610e-01
-5.49595237e-01 7.51078844e-01 9.10068512e-01 7.71355569e-01
7.23336041e-01 3.07321191e-01 -4.57236826e-01 2.98397690e-01
2.35628217e-01 5.73102415e-01 5.74138463e-01 -1.53626561e+00
7.71610141e-01 2.49547914e-01 7.36607313e-01 -2.86411971e-01
-4.51369792e-01 -6.66659892e-01 -5.55996895e-01 5.40270090e-01
5.43205798e-01 -7.45282590e-01 -3.66908401e-01 2.03709149e+00
3.25751305e-01 2.05618776e-02 4.19453949e-01 9.83873010e-01
1.72187641e-01 3.76189411e-01 -1.60838544e-01 -2.71645069e-01
6.33565903e-01 -9.05633032e-01 -6.23729765e-01 6.84170350e-02
4.89255100e-01 -3.95578369e-02 8.42547774e-01 5.76791942e-01
-1.06476879e+00 -2.22378224e-01 -9.39609885e-01 5.68649769e-01
1.23338319e-01 -2.96846598e-01 7.08749115e-01 3.37164551e-01
-9.42241609e-01 9.48330164e-01 -1.17282593e+00 4.12686825e-01
1.60754845e-01 4.88747209e-01 8.18072557e-02 2.26714969e-01
-1.12782335e+00 7.62658596e-01 4.20347482e-01 -1.06738053e-01
-1.57995427e+00 -5.64408302e-01 -5.89878023e-01 2.33542517e-01
1.01541579e+00 -5.80472171e-01 1.61048520e+00 -8.61291051e-01
-2.12623477e+00 -3.91138405e-01 5.30496716e-01 -8.08764040e-01
8.25158119e-01 -3.53211015e-01 1.45880967e-01 1.63215585e-02
1.50797740e-01 3.45393300e-01 5.69566727e-01 -1.03122854e+00
-7.26279616e-01 -2.48643100e-01 5.46580911e-01 5.52591860e-01
5.16028926e-02 -5.67242026e-01 3.61117512e-01 -2.20662951e-01
-4.72940862e-01 -6.87682450e-01 -5.98500252e-01 -3.97624344e-01
-2.23256964e-02 -4.61811662e-01 2.06160024e-01 -2.41419569e-01
8.00985456e-01 -1.94506693e+00 1.82219461e-01 1.83015421e-01
5.29599972e-02 -3.59238893e-01 -4.12674099e-01 9.10342455e-01
6.50178313e-01 -1.61010250e-01 4.42706272e-02 -5.02158344e-01
5.30687213e-01 6.40133619e-01 -3.30292493e-01 6.45736635e-01
-9.36008170e-02 7.70967662e-01 -1.29192555e+00 -1.85686171e-01
5.71930856e-02 -5.03378212e-02 -6.59922063e-01 4.24128532e-01
-8.61800194e-01 3.39567125e-01 -7.27227390e-01 5.44395030e-01
4.35308903e-01 -4.18352149e-02 3.80205184e-01 4.83950734e-01
-2.05678463e-01 8.85613412e-02 -1.34207296e+00 1.55081296e+00
-4.28428978e-01 3.34318466e-02 8.35402012e-02 -8.11236143e-01
7.52793074e-01 4.03109640e-01 7.66852796e-01 -7.06987679e-01
6.10720292e-02 1.44035697e-01 7.77967498e-02 -3.36578965e-01
6.50545716e-01 -6.08499860e-03 -2.70765930e-01 5.70564628e-01
2.94306017e-02 -8.52727238e-03 4.11253303e-01 2.03765437e-01
1.20549774e+00 5.50289810e-01 1.97782099e-01 -8.99397433e-02
4.58030635e-03 -1.64460897e-01 1.01776266e+00 1.15502322e+00
-3.63992244e-01 8.30460563e-02 1.09645808e+00 -3.81176174e-01
-9.59895849e-01 -1.16103959e+00 3.31767201e-01 1.30598938e+00
2.85733998e-01 -2.47026131e-01 -4.19214755e-01 -7.56474733e-01
4.15136456e-01 7.36341476e-01 -7.00099051e-01 1.68772221e-01
-3.33361685e-01 -7.50232637e-01 4.01679665e-01 4.28115577e-01
4.19753909e-01 -1.15363908e+00 -9.13030684e-01 6.12551451e-01
-7.55676553e-02 -7.77551830e-01 -5.57189286e-01 4.34772223e-01
-6.80588424e-01 -1.01714540e+00 -4.47277486e-01 6.56636059e-02
3.12081158e-01 -1.86793625e-01 1.03569210e+00 -3.05876762e-01
3.46574277e-01 6.43865883e-01 -3.99164230e-01 -3.39347035e-01
-2.51218408e-01 4.38475832e-02 1.79274991e-01 -1.27057105e-01
-1.23234220e-01 -6.41006172e-01 -8.10464084e-01 1.90296665e-01
-9.12012219e-01 -4.21801925e-01 5.17206430e-01 1.15798700e+00
4.15237874e-01 9.20701399e-03 1.05636168e+00 -6.18930697e-01
9.73854840e-01 -6.99693501e-01 -1.00972331e+00 2.55393267e-01
-8.11997652e-01 4.71062839e-01 9.93042767e-01 -7.31921256e-01
-1.03122687e+00 -2.74933964e-01 3.96651894e-01 -2.98575789e-01
2.42407650e-01 6.97450042e-01 1.42940059e-01 3.73476118e-01
5.19603491e-01 1.58538043e-01 3.75495940e-01 -8.00288618e-02
3.59530181e-01 2.63072729e-01 2.47519106e-01 -1.08570766e+00
3.43665272e-01 2.60576874e-01 2.07188979e-01 -1.63906738e-01
-5.59130311e-01 -1.05674729e-01 1.15579695e-01 -2.59435952e-01
2.37138480e-01 -1.10188544e+00 -1.45337594e+00 3.86645526e-01
-8.44420135e-01 -9.16422606e-01 -7.86868811e-01 7.11261034e-01
-1.12699342e+00 3.45364541e-01 -7.34436154e-01 -1.18154991e+00
-2.02666447e-01 -1.11839139e+00 6.18643820e-01 2.86469549e-01
3.68925720e-01 -9.68658209e-01 5.31270027e-01 8.86916276e-03
4.83451217e-01 4.27245468e-01 4.91001368e-01 -4.91686463e-01
-7.76022494e-01 3.84223126e-02 3.11792493e-01 3.99156332e-01
-1.10727005e-01 -2.22317263e-01 -4.63399589e-01 -6.26257360e-01
-8.55052620e-02 -8.05621982e-01 5.59937835e-01 5.48923552e-01
7.21250057e-01 -6.41268313e-01 2.87695825e-01 3.47791463e-01
1.59729707e+00 1.20758697e-01 4.04963821e-01 7.28380203e-01
1.48528323e-01 3.33871365e-01 9.68528032e-01 1.31858313e+00
8.25338781e-01 3.80408764e-01 1.26672995e+00 4.13447499e-01
7.05961168e-01 -3.65098894e-01 9.40708697e-01 3.39634776e-01
-2.19159931e-01 -2.49500066e-01 -4.52498466e-01 6.74852073e-01
-2.52570748e+00 -1.10012722e+00 3.22410673e-01 2.28181314e+00
1.01049161e+00 1.12914257e-02 5.77805519e-01 -5.48343301e-01
3.34967285e-01 5.71383983e-02 -1.02806044e+00 -3.37571919e-01
-1.10496350e-01 -1.75512224e-01 8.51140380e-01 4.82201099e-01
-6.63746774e-01 6.89173400e-01 5.73994541e+00 8.42536986e-01
-7.53132045e-01 2.01355636e-01 4.23117161e-01 -6.16384089e-01
-3.08653146e-01 2.88593024e-02 -5.04435182e-01 6.41568303e-01
8.66684437e-01 -1.22325584e-01 9.72327948e-01 1.10838723e+00
2.86571622e-01 -2.27879509e-01 -1.02195406e+00 5.25670409e-01
-5.48911095e-01 -1.02699804e+00 -5.58770180e-01 2.05099016e-01
1.03745914e+00 5.78306556e-01 6.73011616e-02 7.31601000e-01
1.28696787e+00 -8.33916545e-01 1.01218498e+00 6.38674676e-01
2.72229373e-01 -1.09593236e+00 9.92158234e-01 7.42204607e-01
-9.28578258e-01 -5.62393188e-01 -5.28047085e-01 -3.49945992e-01
-5.65895773e-02 2.71604478e-01 -7.78722405e-01 6.09201312e-01
4.25432652e-01 4.42619264e-01 1.00638002e-01 8.37199092e-01
-3.59648257e-01 4.82967645e-01 -3.49175394e-01 -4.23542976e-01
5.43513596e-01 -3.66712898e-01 4.14783686e-01 5.39862752e-01
8.18805918e-02 -2.49617770e-01 6.52531445e-01 8.32824230e-01
1.14182442e-01 -1.41132340e-01 -1.31581232e-01 8.66552368e-02
6.34979486e-01 1.05352926e+00 -1.81405053e-01 -2.12819949e-01
-2.49283463e-01 4.37493622e-01 6.80757701e-01 5.94229102e-01
-8.45996439e-01 -7.63406828e-02 7.88238823e-01 -2.51241893e-01
4.78184760e-01 -4.36295271e-01 4.03223634e-02 -1.07189858e+00
1.54296607e-01 -9.33279395e-01 2.45774239e-01 -1.10272892e-01
-1.50073850e+00 3.74114215e-01 5.40023334e-02 -1.34757388e+00
-8.40799749e-01 -2.48322830e-01 -5.63831866e-01 4.55780596e-01
-1.65587425e+00 -8.53330970e-01 2.72175431e-01 5.80240786e-01
2.64092058e-01 -3.96022826e-01 5.08494318e-01 -1.97010934e-01
-3.96829396e-01 4.48884845e-01 7.77774036e-01 -6.62862062e-02
6.19435132e-01 -1.61951745e+00 -7.30240270e-02 3.17350388e-01
-3.86718154e-01 2.71220595e-01 8.02421212e-01 -5.70055008e-01
-1.39331043e+00 -9.37299728e-01 -7.51338676e-02 -1.99968264e-01
9.03904080e-01 -1.10287778e-02 -3.90960485e-01 5.57829201e-01
3.05160373e-01 1.82723720e-02 3.74705881e-01 3.96185704e-02
-1.63165569e-01 -3.10664296e-01 -1.30414879e+00 3.77661794e-01
5.13834417e-01 -2.40370452e-01 -1.60908848e-01 1.02279432e-01
8.70887876e-01 -3.11338246e-01 -1.09790051e+00 1.20044492e-01
4.78452682e-01 -1.02174938e+00 5.03998995e-01 -6.49941564e-01
3.90673220e-01 -3.24202895e-01 -2.45283946e-01 -1.73747027e+00
6.10420946e-03 -8.17246199e-01 -4.72177595e-01 8.64168763e-01
1.80023924e-01 -9.03551102e-01 8.10263395e-01 6.05464041e-01
-6.17980547e-02 -8.86586070e-01 -1.27914584e+00 -1.16072190e+00
1.91422373e-01 -4.30954956e-02 9.73890960e-01 6.02190673e-01
1.67529836e-01 -2.37912357e-01 -6.16886854e-01 2.02089727e-01
1.09941351e+00 1.97452009e-01 8.19300950e-01 -7.88483381e-01
-9.18640733e-01 -2.22136766e-01 -3.94288488e-02 -9.76783037e-01
1.83026820e-01 -3.49183083e-01 2.48050213e-01 -1.29503703e+00
2.42731050e-01 -5.48087418e-01 -5.65038979e-01 4.84989136e-01
6.38263747e-02 -5.80854952e-01 4.87060755e-01 1.60459444e-01
-1.27589178e+00 1.16808665e+00 1.35435939e+00 -9.14485082e-02
-4.36972469e-01 1.59928590e-01 -4.34463710e-01 3.61373365e-01
1.03838634e+00 -6.47156179e-01 -5.90787172e-01 -2.92507142e-01
5.91578007e-01 7.58564293e-01 2.90853530e-01 -7.79796898e-01
3.32912266e-01 -9.00860727e-01 -3.64207774e-02 -4.90035921e-01
1.78365931e-01 -6.29133523e-01 5.33978119e-02 5.22603035e-01
-4.87984121e-01 1.44573748e-01 -2.36634851e-01 7.21722245e-01
-2.48544216e-01 -3.40555519e-01 2.57828206e-01 -3.70899588e-01
-2.56350398e-01 4.02445048e-01 -4.44120020e-01 2.21265927e-01
8.68130028e-01 1.73693880e-01 -5.81297994e-01 -7.72356033e-01
-4.46826160e-01 1.01671004e+00 2.62806028e-01 -2.09713783e-02
4.95249212e-01 -1.14726293e+00 -7.92254746e-01 -3.59787047e-01
-2.95230985e-01 3.14543664e-01 3.06215346e-01 5.17528772e-01
-1.87252745e-01 -2.49707140e-02 -1.92174286e-01 -2.70661622e-01
-4.32069391e-01 3.51526856e-01 6.10519111e-01 -8.73819888e-01
-1.53377712e-01 3.01298141e-01 -2.09705651e-01 -7.23644495e-01
3.94720852e-01 -3.32667470e-01 5.89634739e-02 -1.33794591e-01
2.45407730e-01 6.66587889e-01 -4.72308815e-01 2.19342470e-01
7.06139207e-02 -7.77537972e-02 -9.87521261e-02 -5.45113742e-01
1.63995838e+00 -1.53386921e-01 1.50182424e-02 4.61616367e-01
4.97071832e-01 -3.63681316e-01 -2.12328458e+00 -4.03800279e-01
-1.02866307e-01 -2.39892647e-01 -1.00474395e-01 -1.02648723e+00
-9.49269116e-01 2.79009670e-01 3.95451099e-01 2.60530531e-01
7.03943491e-01 -2.82067090e-01 4.26525801e-01 7.60744810e-01
9.19926286e-01 -1.74301338e+00 3.56335133e-01 8.02479088e-01
7.27937400e-01 -1.15699434e+00 -3.56990844e-03 5.51267564e-01
-1.07549345e+00 9.68821824e-01 8.64905357e-01 -3.39890808e-01
2.59283543e-01 3.37994099e-01 -2.21682005e-02 1.81666002e-01
-1.39251757e+00 -2.80430615e-01 -5.03339469e-01 7.30871558e-01
-1.15755253e-01 5.06899416e-01 -6.66472465e-02 7.89426863e-01
-1.92434579e-01 8.57036188e-02 8.30549777e-01 1.08809197e+00
-3.45094174e-01 -1.14836204e+00 -2.71157771e-01 2.92032450e-01
-3.63066226e-01 2.10364863e-01 2.01906592e-01 8.06340277e-01
-4.46181335e-02 9.69755828e-01 2.72075012e-02 -6.11032769e-02
2.13649184e-01 -4.48934048e-01 4.09552991e-01 -2.03378439e-01
-9.69722688e-01 3.35644148e-02 -4.81685326e-02 -6.57344580e-01
-3.17230076e-01 -6.36507928e-01 -1.52523351e+00 -4.31550324e-01
-1.89140245e-01 3.39401513e-01 6.51895821e-01 8.22306097e-01
2.03890756e-01 5.76781988e-01 9.49532330e-01 -5.67600608e-01
-1.75566828e+00 -9.22471941e-01 -9.21039701e-01 1.51962638e-01
6.13537967e-01 -7.06295133e-01 -4.54251021e-01 -6.87672496e-01] | [3.8020694255828857, 2.1220955848693848] |
07b57547-0d10-4121-88fa-7d8642024cc1 | contextual-outpainting-with-object-level | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Contextual_Outpainting_With_Object-Level_Contrastive_Learning_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Contextual_Outpainting_With_Object-Level_Contrastive_Learning_CVPR_2022_paper.pdf | Contextual Outpainting With Object-Level Contrastive Learning | We study the problem of contextual outpainting, which aims to hallucinate the missing background contents based on the remaining foreground contents. Existing image outpainting methods focus on completing object shapes or extending existing scenery textures, neglecting the semantically meaningful relationship between the missing and remaining contents. To explore the semantic cues provided by the remaining foreground contents, we propose a novel ConTextual Outpainting GAN (CTO-GAN), leveraging the semantic layout as a bridge to synthesize coherent and diverse background contents. To model the contextual correlation between foreground and background contents, we incorporate an object-level contrastive loss to regularize the learning of cross-modal representations of foreground contents and the corresponding background semantic layout, facilitating accurate semantic reasoning. Furthermore, we improve the realism of the generated background contents via detecting generated context in adversarial training. Extensive experiments demonstrate that the proposed method achieves superior performance compared with existing solutions on the challenging COCO-stuff dataset. | ['Zhiwei Xiong', 'Chang Chen', 'Jiacheng Li'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['image-outpainting'] | ['computer-vision'] | [ 6.46071613e-01 1.25026003e-01 4.66401055e-02 -1.60331488e-01
-7.55886376e-01 -4.73526835e-01 5.39765537e-01 -3.91553640e-01
3.83571059e-01 8.40818703e-01 3.30790609e-01 9.76922363e-02
3.87775838e-01 -8.48277628e-01 -1.28436601e+00 -8.56495261e-01
5.55163145e-01 2.06127167e-01 1.96224675e-01 -1.43015981e-01
6.31472841e-02 3.32184315e-01 -1.35257065e+00 7.23133981e-01
1.06102681e+00 1.05416489e+00 5.63497305e-01 3.83617699e-01
-4.73101199e-01 1.31372476e+00 -1.04640031e+00 -5.37457466e-01
4.04475868e-01 -8.70123684e-01 -4.62188035e-01 7.48090327e-01
7.36750007e-01 -5.52420080e-01 -6.03391707e-01 1.26349294e+00
1.82323962e-01 4.73874770e-02 3.33638519e-01 -1.34714603e+00
-9.40443754e-01 4.14080441e-01 -8.00954342e-01 -4.42500375e-02
4.12053138e-01 2.87435830e-01 5.88192761e-01 -7.67581165e-01
7.92815924e-01 1.48632753e+00 4.64637727e-01 6.35305822e-01
-1.25883842e+00 -6.86478674e-01 5.69920182e-01 9.52212736e-02
-1.21221375e+00 -2.01052040e-01 1.41036880e+00 -1.98174506e-01
1.36656398e-02 3.50878745e-01 5.74008942e-01 1.51475894e+00
-6.95316643e-02 1.12745976e+00 1.24292588e+00 -3.52015138e-01
1.11441746e-01 6.92589134e-02 -6.15315437e-01 7.09552765e-01
2.15785161e-01 3.69000472e-02 -4.13562477e-01 1.60657172e-03
1.04363394e+00 2.30286986e-01 -4.58250105e-01 -3.95678610e-01
-1.04034042e+00 5.53532720e-01 4.09994960e-01 -8.58154520e-02
-2.80762672e-01 3.03706199e-01 7.30965436e-02 -2.39031374e-01
4.85488653e-01 2.82151520e-01 -2.92125165e-01 4.82019752e-01
-1.04738545e+00 3.28335792e-01 2.96027809e-01 1.48032618e+00
7.40210593e-01 4.79955316e-01 -6.10937655e-01 7.57004499e-01
1.18105724e-01 6.57458663e-01 6.82211369e-02 -1.27763116e+00
6.27342224e-01 5.39788663e-01 3.58338475e-01 -1.09779394e+00
3.56248468e-01 -4.02674735e-01 -8.08041334e-01 1.56809893e-02
3.70075524e-01 -3.81622761e-02 -1.06440020e+00 1.86388278e+00
5.38925827e-01 6.12421870e-01 1.06422246e-01 8.44676852e-01
8.68775606e-01 7.12667227e-01 1.35141283e-01 -4.51016836e-02
1.13433421e+00 -1.40156949e+00 -9.86557722e-01 -5.09468198e-01
-8.17964524e-02 -9.08144176e-01 1.18588459e+00 9.83109102e-02
-1.36586714e+00 -6.47423565e-01 -1.00361288e+00 -3.35167944e-01
-1.26416683e-01 1.89868793e-01 5.90448022e-01 3.05961549e-01
-5.88487566e-01 3.65629375e-01 -5.56825399e-01 2.48817444e-01
7.61295199e-01 -2.82799333e-01 -9.50751752e-02 -5.09856224e-01
-1.09772956e+00 5.67566693e-01 5.44352889e-01 2.18554735e-01
-1.25616074e+00 -9.90554035e-01 -1.02939594e+00 3.87587808e-02
6.04906082e-01 -8.30720723e-01 7.60331571e-01 -1.52124059e+00
-1.29065597e+00 7.06518829e-01 -1.51284754e-01 -4.96864915e-02
8.81162286e-01 -2.48479709e-01 -2.53273815e-01 3.55616838e-01
3.31722289e-01 9.10375118e-01 1.15696585e+00 -1.94590330e+00
-4.26385969e-01 -1.50720790e-01 6.23114333e-02 2.94688582e-01
1.46655262e-01 -3.90501559e-01 -7.00737655e-01 -1.36860788e+00
-5.03682718e-03 -5.84407210e-01 -1.93619847e-01 1.99555501e-01
-5.86955070e-01 5.13783574e-01 1.38370073e+00 -1.13666499e+00
7.08788037e-01 -2.23903894e+00 2.93897867e-01 -1.75926864e-01
-1.89787373e-02 2.03180891e-02 -2.89090812e-01 2.04698518e-01
1.37861177e-01 -1.96724012e-01 -5.74461520e-01 -4.65720683e-01
-1.37310594e-01 5.02881587e-01 -8.02050531e-01 7.68082738e-02
5.49604297e-01 1.07275283e+00 -1.15702999e+00 -4.57539350e-01
3.38705301e-01 7.52673745e-01 -4.61295813e-01 5.36165476e-01
-7.72508860e-01 1.00237942e+00 -4.86348897e-01 8.98336768e-01
1.09980750e+00 -8.45235437e-02 7.96560049e-02 -2.57640272e-01
6.03719175e-01 -1.57221586e-01 -8.90328348e-01 1.95436358e+00
-4.87133294e-01 5.09587407e-01 2.70646274e-01 -7.70543039e-01
8.32766712e-01 -1.16329573e-01 2.06655815e-01 -6.48752809e-01
1.02042601e-01 -1.95043143e-02 -4.53324974e-01 -4.93611187e-01
4.87137109e-01 -2.03398421e-01 7.68838376e-02 -2.04787571e-02
-9.64024514e-02 -4.58283961e-01 -2.12803230e-01 2.88773298e-01
6.11351073e-01 6.84767306e-01 -1.56044081e-01 -3.74697265e-03
5.11789918e-01 -2.21526459e-01 6.02449596e-01 5.43243825e-01
-3.30172665e-02 1.15864742e+00 5.56738675e-01 -2.77819037e-01
-1.09434557e+00 -1.18766892e+00 2.29175448e-01 7.19470263e-01
6.96937621e-01 -7.23555833e-02 -1.07492292e+00 -8.47003102e-01
-1.47352502e-01 1.07997406e+00 -9.30225551e-01 -2.49830157e-01
-5.68098605e-01 -4.28051054e-01 4.28152978e-01 6.70206845e-01
8.10671687e-01 -1.18543899e+00 -2.62230545e-01 1.11857362e-01
-7.97093630e-01 -1.52500045e+00 -6.89272404e-01 -1.02519058e-01
-7.37237573e-01 -9.34492350e-01 -8.45066071e-01 -7.16050208e-01
8.56139362e-01 3.62939119e-01 1.21258724e+00 2.24438965e-01
-4.81805176e-01 2.35615969e-01 -2.90745765e-01 -1.44530460e-01
-5.71868360e-01 -6.23828053e-01 -6.59959495e-01 3.52261931e-01
-2.18344808e-01 -6.21278346e-01 -7.51660407e-01 2.23575085e-01
-1.34461427e+00 5.40402770e-01 6.23571217e-01 9.99263763e-01
7.65229404e-01 1.17985852e-01 3.79975885e-01 -1.12675166e+00
1.19907791e-02 -5.22832394e-01 -2.97104627e-01 4.00649697e-01
1.87386535e-02 -5.54960556e-02 6.82404816e-01 -6.59428775e-01
-1.56921005e+00 -1.09188169e-01 9.84842144e-03 -1.02628136e+00
-1.83140934e-01 -2.61464655e-01 -9.68074441e-01 8.43944177e-02
1.20310217e-01 6.14343762e-01 -3.49961847e-01 -2.90247947e-01
8.36051285e-01 4.26128432e-02 8.75149965e-01 -1.15471292e+00
7.42617488e-01 9.34326351e-01 -5.61922565e-02 -5.50124526e-01
-1.20837808e+00 -5.42797782e-02 -3.39481026e-01 -1.05947345e-01
1.13588393e+00 -1.04438818e+00 2.89826002e-02 4.29611444e-01
-1.43733668e+00 -4.94890779e-01 -6.33728385e-01 -1.35053664e-01
-8.28950465e-01 5.42328000e-01 -5.90595841e-01 -5.37631392e-01
-4.13847826e-02 -1.23598897e+00 1.59687650e+00 1.21042602e-01
1.45133078e-01 -9.08765376e-01 -4.74057585e-01 6.95010543e-01
1.67536944e-01 8.72256756e-01 1.01193321e+00 -1.42163467e-02
-1.06456590e+00 1.24025792e-01 -5.47412097e-01 4.57116812e-01
1.68538317e-01 -1.63058192e-01 -9.58250582e-01 -4.24233750e-02
9.31255743e-02 -2.00183928e-01 8.58694196e-01 1.95220381e-01
1.49732816e+00 -5.06911933e-01 -1.96737364e-01 8.05481434e-01
1.46477795e+00 1.88408419e-01 1.00218117e+00 1.27122805e-01
1.21907151e+00 7.66296625e-01 6.77788377e-01 3.45200449e-01
9.99416858e-02 4.59218264e-01 5.96212327e-01 -3.08171988e-01
-7.70039201e-01 -8.60387266e-01 2.65338749e-01 3.03138494e-01
2.38024443e-01 -3.53670925e-01 -2.61698902e-01 4.85273659e-01
-1.74115324e+00 -1.03768682e+00 4.58645634e-02 1.63829386e+00
8.66926849e-01 -3.29727866e-02 -3.28898430e-01 -2.95033693e-01
1.04756379e+00 3.19235504e-01 -7.19722867e-01 1.34777263e-01
-5.01576245e-01 1.77337229e-01 2.35424444e-01 4.57965791e-01
-8.86095941e-01 1.23171127e+00 5.82655525e+00 1.27067268e+00
-8.52416337e-01 2.25037187e-01 9.94405210e-01 -1.79369804e-02
-6.74305797e-01 4.44679298e-02 -5.28159320e-01 8.02051723e-01
7.50456825e-02 3.63464683e-01 4.40576136e-01 9.38754678e-01
8.12360123e-02 -1.27867639e-01 -8.85400414e-01 8.14632297e-01
3.25802863e-01 -1.43896961e+00 5.77294767e-01 -2.70661622e-01
1.36057818e+00 -8.59919071e-01 2.49771759e-01 1.66749865e-01
1.98958039e-01 -8.62558305e-01 1.21782708e+00 6.45846844e-01
8.68392646e-01 -7.40055025e-01 3.14944774e-01 8.76784697e-02
-9.82973099e-01 -7.38886222e-02 -3.27944249e-01 1.90924302e-01
1.19992584e-01 4.96405244e-01 -1.97329819e-01 6.72603428e-01
5.05549312e-01 6.10687971e-01 -5.79718769e-01 4.46922630e-01
-5.70937574e-01 2.99714059e-01 1.60055652e-01 6.37840569e-01
1.79091528e-01 -3.43575388e-01 5.41377664e-01 9.97504890e-01
2.55641639e-01 3.58100459e-02 2.26841748e-01 1.52467036e+00
-2.58005798e-01 -1.84893325e-01 -5.47333658e-01 1.13708235e-01
3.59756440e-01 1.05861974e+00 -8.82555425e-01 -5.97053170e-01
-3.97394001e-01 1.53408861e+00 1.10887781e-01 6.40689731e-01
-1.24271345e+00 -1.36292145e-01 5.11798978e-01 2.30287358e-01
5.97902536e-01 1.23573415e-01 -4.73313004e-01 -1.41080308e+00
2.69207567e-01 -1.04519939e+00 3.89009081e-02 -1.26476800e+00
-1.35184705e+00 4.32996184e-01 -2.65907357e-03 -1.17427313e+00
2.69709080e-01 -1.91605747e-01 -6.88640177e-01 8.02048981e-01
-1.49813938e+00 -1.75648797e+00 -6.18306041e-01 6.02312982e-01
9.51948822e-01 2.54398175e-02 3.36420357e-01 3.23475957e-01
-3.13654929e-01 4.19602573e-01 -1.68186605e-01 6.81833997e-02
7.20133841e-01 -1.06737447e+00 1.47427499e-01 1.14259303e+00
-4.82889190e-02 4.67156380e-01 7.62322783e-01 -8.98519218e-01
-1.08560634e+00 -1.57785213e+00 8.23412910e-02 -4.96115834e-01
2.84607142e-01 -4.32228208e-01 -8.65691364e-01 7.18604922e-01
3.38170946e-01 2.42115915e-01 2.96603888e-01 -7.12800205e-01
-6.57081723e-01 6.36558682e-02 -1.34035277e+00 8.21392894e-01
1.19579661e+00 -4.88972217e-01 -5.93599021e-01 1.37662604e-01
1.12276101e+00 -5.74313462e-01 -2.85174549e-01 5.53279221e-01
2.26783350e-01 -1.03569913e+00 1.27992952e+00 -5.06576002e-01
1.08459997e+00 -5.83104730e-01 -5.11254370e-01 -1.15533876e+00
1.02061071e-01 -6.04973674e-01 -3.51447910e-01 1.40619409e+00
-2.31163323e-01 -1.64883295e-05 9.49356019e-01 3.20367664e-01
-2.52582818e-01 -5.63622892e-01 -3.72607499e-01 -5.77876449e-01
5.52312955e-02 -2.55550593e-01 7.43022978e-01 1.14896786e+00
-8.58384311e-01 1.11664928e-01 -7.33806849e-01 2.37619117e-01
8.21215570e-01 4.33960587e-01 8.38842273e-01 -5.13350427e-01
-6.03704810e-01 -2.54697025e-01 -1.66860744e-01 -1.15487516e+00
3.51234555e-01 -6.72199488e-01 4.81856726e-02 -1.33553386e+00
2.19245255e-01 -3.31805676e-01 -7.79512897e-02 1.61223799e-01
-6.68296218e-01 5.38344979e-01 2.31125623e-01 -6.30019605e-02
-6.08965337e-01 8.63316298e-01 1.96759045e+00 -4.20796990e-01
1.42383873e-01 -4.46423680e-01 -8.51517975e-01 8.19869757e-01
4.22919482e-01 -3.37715983e-01 -6.22017741e-01 -4.79110122e-01
-2.21823454e-01 3.70306969e-01 8.21367860e-01 -8.86188686e-01
-2.19000965e-01 -3.75130534e-01 9.49239612e-01 -7.03730702e-01
4.85068262e-01 -9.65528071e-01 2.95198619e-01 1.01103559e-01
-3.67531419e-01 -3.34342718e-01 2.74738163e-01 1.04627693e+00
-3.40720534e-01 -2.26719584e-02 9.69636619e-01 -2.99119651e-01
-6.51049078e-01 2.98136652e-01 3.52903605e-02 4.35000151e-01
1.19721258e+00 -1.40665680e-01 -2.61933476e-01 -2.71811008e-01
-7.27897108e-01 -3.71282697e-02 7.01069832e-01 4.76503551e-01
7.45712936e-01 -1.50152922e+00 -4.33532476e-01 3.57921511e-01
2.29354538e-02 4.06114787e-01 7.84843862e-01 4.31385249e-01
-7.44118869e-01 -9.50300395e-02 -4.40324962e-01 -4.85404521e-01
-9.95796740e-01 1.03388357e+00 1.57629281e-01 -1.17936060e-01
-8.09103310e-01 9.58428741e-01 1.08409929e+00 -7.75584131e-02
2.14855611e-01 -2.54910082e-01 3.49506080e-01 -2.84193993e-01
5.55166304e-01 2.23198503e-01 -3.30601841e-01 -5.11607766e-01
9.51632932e-02 4.04576242e-01 -1.48510644e-02 -1.32940216e-02
1.04570138e+00 -3.30586582e-01 -1.54186606e-01 2.79490978e-01
1.06332350e+00 3.76471788e-01 -1.86680734e+00 -1.18277408e-01
-2.67884374e-01 -9.29995537e-01 -2.92211235e-01 -7.98182487e-01
-1.32362580e+00 9.43931937e-01 2.25337505e-01 -3.04024875e-01
1.35649514e+00 -3.30374576e-02 1.12218308e+00 -1.93106607e-01
3.65242720e-01 -7.68136203e-01 6.08838260e-01 5.34549318e-02
1.01569247e+00 -9.90067244e-01 -5.03842011e-02 -8.98838699e-01
-8.53241265e-01 8.72899234e-01 8.84509802e-01 -3.69404018e-01
2.03152984e-01 3.30439866e-01 1.26862064e-01 8.28226283e-02
-3.89594316e-01 -1.49652407e-01 1.54428929e-01 7.90854394e-01
-4.27769572e-02 -1.62460171e-02 6.24537952e-02 7.02077329e-01
1.68814167e-01 -3.89133871e-01 4.92547005e-01 7.29678333e-01
-2.01185662e-02 -1.07348442e+00 -5.77030301e-01 -7.04225078e-02
-6.08738303e-01 -3.12822402e-01 -5.44090450e-01 7.48679698e-01
5.86308658e-01 7.01209605e-01 -3.11835073e-02 3.09957489e-02
1.41445294e-01 3.46346609e-02 8.19936395e-01 -4.46576893e-01
-7.58723542e-02 5.19875467e-01 -2.53253311e-01 -6.57823324e-01
-3.85729432e-01 -3.48923743e-01 -9.86095488e-01 2.47645061e-02
-8.92099887e-02 -1.94588900e-01 3.26776624e-01 8.47667813e-01
2.12224647e-01 1.07444835e+00 4.66380894e-01 -1.03799033e+00
-1.10989556e-01 -6.32731676e-01 -6.12225175e-01 8.69449139e-01
2.90454596e-01 -8.45306575e-01 -3.64234805e-01 4.04682904e-01] | [11.411153793334961, -0.9594154953956604] |
f1e722c0-117b-44bd-be14-e7c5dbc68ad1 | when-deep-learning-met-code-search | 1905.03813 | null | https://arxiv.org/abs/1905.03813v4 | https://arxiv.org/pdf/1905.03813v4.pdf | When Deep Learning Met Code Search | There have been multiple recent proposals on using deep neural networks for code search using natural language. Common across these proposals is the idea of $\mathit{embedding}$ code and natural language queries, into real vectors and then using vector distance to approximate semantic correlation between code and the query. Multiple approaches exist for learning these embeddings, including $\mathit{unsupervised}$ techniques, which rely only on a corpus of code examples, and $\mathit{supervised}$ techniques, which use an $\mathit{aligned}$ corpus of paired code and natural language descriptions. The goal of this supervision is to produce embeddings that are more similar for a query and the corresponding desired code snippet. Clearly, there are choices in whether to use supervised techniques at all, and if one does, what sort of network and training to use for supervision. This paper is the first to evaluate these choices systematically. To this end, we assembled implementations of state-of-the-art techniques to run on a common platform, training and evaluation corpora. To explore the design space in network complexity, we also introduced a new design point that is a $\mathit{minimal}$ supervision extension to an existing unsupervised technique. Our evaluation shows that: 1. adding supervision to an existing unsupervised technique can improve performance, though not necessarily by much; 2. simple networks for supervision can be more effective that more sophisticated sequence-based networks for code search; 3. while it is common to use docstrings to carry out supervision, there is a sizeable gap between the effectiveness of docstrings and a more query-appropriate supervision corpus. The evaluation dataset is now available at arXiv:1908.09804 | ['Jose Cambronero', 'Hongyu Li', 'Satish Chandra', 'Koushik Sen', 'Seohyun Kim'] | 2019-05-09 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [ 2.84008328e-02 -7.91822970e-02 -2.30769038e-01 -5.22033334e-01
-6.27644122e-01 -7.22752869e-01 4.92493570e-01 3.87202889e-01
-6.27963185e-01 2.97878146e-01 1.14436276e-01 -7.59233475e-01
-3.63467008e-01 -7.88338721e-01 -6.11917615e-01 -2.77020156e-01
-9.55921710e-02 4.70142365e-01 2.98743576e-01 -1.94730207e-01
4.85803157e-01 3.22527736e-01 -1.65338469e+00 4.51202923e-03
6.73008978e-01 8.29478860e-01 3.27209443e-01 6.92860305e-01
-6.04639471e-01 9.19681430e-01 -4.79052246e-01 -5.91539383e-01
3.38194042e-01 -3.68538529e-01 -1.04937255e+00 -2.92180628e-01
2.66830891e-01 -1.62927255e-01 -2.51978248e-01 1.13670015e+00
2.60021240e-01 1.96032114e-02 6.52740121e-01 -1.01807570e+00
-7.48326540e-01 6.44796252e-01 -1.75952807e-01 3.23438764e-01
3.96226764e-01 1.68734804e-01 1.29730141e+00 -7.54950345e-01
5.99901378e-01 7.67925680e-01 7.78472483e-01 4.60488707e-01
-1.29478753e+00 -6.37690544e-01 -2.95187354e-01 -1.97030187e-01
-1.40099204e+00 -3.90720189e-01 5.39660215e-01 -7.11782038e-01
1.49321365e+00 1.70876622e-01 4.14407521e-01 6.80790722e-01
-2.20414221e-01 5.58267355e-01 7.04511821e-01 -6.96516871e-01
1.01671256e-01 5.58309674e-01 2.93419093e-01 1.02519500e+00
2.48610437e-01 2.05868289e-01 -9.20328870e-02 -3.79256576e-01
5.86568713e-01 1.33912832e-01 -2.27690592e-01 -7.24824488e-01
-7.77695894e-01 1.18476462e+00 3.89678776e-01 6.46207154e-01
1.72272861e-01 3.55356246e-01 6.95262492e-01 4.88705695e-01
2.62818076e-02 1.00975418e+00 -5.27351320e-01 -4.28842932e-01
-1.21552885e+00 3.27838480e-01 9.61051464e-01 1.10077572e+00
1.17669106e+00 2.50191931e-02 2.83302248e-01 9.90395486e-01
4.08479720e-01 9.08466876e-02 6.65669978e-01 -1.00284386e+00
4.67176318e-01 7.42497146e-01 -1.83218464e-01 -1.03332496e+00
-1.13586649e-01 -1.62913203e-01 -2.76943237e-01 2.51177728e-01
5.02358675e-01 -7.82922953e-02 -6.58783555e-01 1.69690681e+00
-3.77632588e-01 -3.48395675e-01 -3.02668735e-02 5.35797179e-01
4.65283483e-01 3.76108617e-01 -1.42059326e-01 2.03968495e-01
1.10146046e+00 -9.40836728e-01 -1.51553467e-01 -2.54902422e-01
1.22127950e+00 -8.76713753e-01 1.22086859e+00 1.11810319e-01
-1.02419138e+00 -4.83436733e-01 -1.21317911e+00 -7.75565654e-02
-7.66299903e-01 -9.02682021e-02 7.05179334e-01 8.94369423e-01
-1.32777989e+00 8.09140265e-01 -5.66687346e-01 -6.82584822e-01
2.11595789e-01 5.45148969e-01 -2.81287134e-01 -1.18917160e-01
-9.81199563e-01 8.31050575e-01 4.61916536e-01 -5.09061933e-01
-6.49600744e-01 -4.12286937e-01 -1.19291556e+00 1.96258858e-01
2.22647175e-01 -2.42305562e-01 1.19945633e+00 -1.29363084e+00
-9.61100221e-01 1.01020646e+00 5.49275391e-02 -6.06254935e-01
-1.18437828e-03 1.65888548e-01 -3.10464948e-01 6.67277500e-02
1.01627842e-01 6.89470828e-01 4.07740265e-01 -1.01902676e+00
-3.20180684e-01 -6.45151585e-02 4.62674886e-01 -1.55031577e-01
-7.46521533e-01 3.29040080e-01 -4.87937778e-01 -6.02035642e-01
-2.21552119e-01 -8.53410184e-01 -9.15487111e-02 1.39505163e-01
-1.08961664e-01 -4.18257564e-01 5.42929053e-01 -3.27077150e-01
1.38343656e+00 -2.25899148e+00 -1.06906131e-01 3.77127916e-01
1.98175088e-01 5.19945681e-01 -1.22388989e-01 7.19288528e-01
-2.38718748e-01 4.73798335e-01 -4.10719037e-01 -1.86897933e-01
2.41694272e-01 3.12507093e-01 -2.00080648e-01 2.43303508e-01
2.78520346e-01 6.31867886e-01 -8.92713010e-01 -6.04769707e-01
-1.45653933e-01 2.69131511e-01 -9.81640220e-01 1.09379388e-01
-2.74850637e-01 -3.26999426e-01 -3.37540239e-01 3.33000124e-01
1.22954629e-01 -3.54577839e-01 9.58113670e-02 1.89862192e-01
-2.33795270e-01 6.84788167e-01 -1.06979012e+00 1.80770230e+00
-4.02232677e-01 8.14133525e-01 -3.18050422e-02 -1.28716493e+00
1.08751607e+00 2.51709074e-01 3.70114386e-01 -5.05537748e-01
1.27205461e-01 6.20385587e-01 5.99559722e-03 -5.72594464e-01
5.73489189e-01 -1.16170943e-02 -1.73057541e-01 7.30928540e-01
3.34453374e-01 -2.66670614e-01 4.31855470e-01 1.59861714e-01
1.59599721e+00 2.45473415e-01 1.18609510e-01 -5.10309339e-01
4.50817347e-01 2.52124965e-01 2.09271654e-01 8.10458422e-01
-1.36252373e-01 5.21641970e-01 5.33350110e-01 -4.81162310e-01
-1.18851137e+00 -8.28514278e-01 -5.20731732e-02 1.12027180e+00
-1.43051654e-01 -7.43586361e-01 -7.13770688e-01 -7.65670478e-01
-2.45958287e-02 6.25403762e-01 -4.17853057e-01 -5.20381927e-02
-5.15855432e-01 -1.48417115e-01 8.97280455e-01 6.23739779e-01
2.93567479e-01 -1.06675959e+00 -7.69365191e-01 1.37863010e-01
2.74451822e-01 -6.00177765e-01 -5.65304160e-01 7.46488273e-01
-7.30073333e-01 -9.77459013e-01 -4.77008522e-01 -9.50620115e-01
6.45005107e-01 1.91167444e-02 1.31430876e+00 6.95444763e-01
-2.94933736e-01 4.65649694e-01 -4.95632946e-01 -2.43285105e-01
-5.83613932e-01 1.22266926e-01 -1.07955329e-01 -7.37591147e-01
8.46081376e-01 -5.99852383e-01 -4.37628657e-01 5.63701205e-02
-1.26425052e+00 -4.87456083e-01 5.12322009e-01 8.43842149e-01
1.38977274e-01 -5.07351011e-03 1.33660436e-01 -8.04914057e-01
7.88585126e-01 -3.69865119e-01 -5.63252032e-01 4.75529060e-02
-9.23165083e-01 5.41274905e-01 8.98889899e-01 -1.47593558e-01
-3.66405845e-01 1.01365641e-01 -4.62247908e-01 -4.79284555e-01
-1.72612473e-01 7.93595254e-01 3.65947306e-01 -3.53002287e-02
1.01892471e+00 2.36964464e-01 1.58453301e-01 -4.40477878e-01
2.32061088e-01 7.67583847e-01 2.85553098e-01 -7.84038961e-01
7.99068034e-01 5.48026338e-02 -5.20057976e-01 -5.73275983e-01
-4.58860159e-01 -5.96757054e-01 -5.24485290e-01 3.34112465e-01
8.70761931e-01 -4.87583786e-01 -3.80667925e-01 -1.29069433e-01
-1.09320819e+00 -3.97956163e-01 -3.35764319e-01 4.05894518e-01
-6.34324372e-01 4.77863848e-01 -5.38782537e-01 -5.36268890e-01
-5.94422221e-02 -1.42260349e+00 6.46905661e-01 1.12311259e-01
-5.69779038e-01 -1.00593364e+00 1.84435755e-01 1.61925197e-01
6.57709658e-01 -1.96892768e-01 1.11133277e+00 -1.11760986e+00
-5.50233662e-01 -3.78606886e-01 -3.74084681e-01 6.33266509e-01
1.32085249e-01 1.27350524e-01 -7.98784137e-01 -3.53272915e-01
1.85801834e-02 -6.03367507e-01 6.41201079e-01 -1.67176679e-01
1.09132457e+00 -2.47383758e-01 -3.03052485e-01 5.36166668e-01
1.84670329e+00 3.79370183e-01 6.41163349e-01 4.36382025e-01
4.25220400e-01 4.94722128e-01 1.36477798e-01 7.59072006e-02
3.16041023e-01 6.47251010e-01 3.16260517e-01 1.71944693e-01
1.57463908e-01 -2.07032785e-01 4.25856918e-01 9.57762539e-01
2.70880789e-01 -8.99738353e-03 -1.25654972e+00 7.96762347e-01
-1.48458660e+00 -9.70069528e-01 2.93020993e-01 2.20922828e+00
8.99272263e-01 3.48562330e-01 1.61069304e-01 1.07955866e-01
3.36832285e-01 4.89489026e-02 -2.01714456e-01 -9.17115748e-01
3.10205847e-01 5.86519957e-01 5.24349689e-01 4.08077061e-01
-8.71114314e-01 6.55411661e-01 5.71719646e+00 7.99934030e-01
-1.11099434e+00 7.26916939e-02 2.39621013e-01 1.46405131e-01
-4.39915627e-01 3.32108557e-01 -6.86078787e-01 3.99839908e-01
1.19480276e+00 -8.93930942e-02 5.99326491e-01 1.02238953e+00
-2.33036265e-01 5.92581443e-02 -1.35022151e+00 8.05206060e-01
1.64493203e-01 -1.39675999e+00 -3.07881206e-01 -2.01781280e-02
3.51672888e-01 2.44420737e-01 -2.70619988e-01 5.54392636e-01
5.88757992e-01 -1.10653150e+00 4.28529888e-01 1.92310587e-01
7.61382103e-01 -5.65100312e-01 6.95203722e-01 2.47499809e-01
-1.23231506e+00 -2.56715059e-01 -3.17916334e-01 -1.60040662e-01
-3.20211262e-01 1.61858171e-01 -5.73613942e-01 3.21692050e-01
7.98857749e-01 5.52678227e-01 -7.23896980e-01 1.04265213e+00
-2.94954563e-03 4.37018901e-01 -2.70017892e-01 -4.75730598e-01
5.29438794e-01 3.80847119e-02 2.61142135e-01 1.53314221e+00
3.88724029e-01 -3.36316794e-01 2.28394061e-01 1.14556658e+00
-4.52569984e-02 1.63752571e-01 -1.02673006e+00 -4.26492393e-01
4.65714723e-01 8.93314838e-01 -6.60209358e-01 -4.23175782e-01
-6.26320302e-01 7.42169559e-01 4.98218983e-01 2.39578024e-01
-6.95499361e-01 -9.48160410e-01 4.95882422e-01 2.01352209e-01
5.63839436e-01 -3.92157644e-01 -1.01134025e-01 -1.07030356e+00
2.54656106e-01 -1.04890943e+00 3.90157342e-01 -7.42103100e-01
-1.11205649e+00 7.11354792e-01 6.52760565e-02 -1.10660374e+00
-4.97138828e-01 -7.82023072e-01 -4.50968117e-01 9.20787990e-01
-1.34102559e+00 -4.20992553e-01 8.97536352e-02 4.21643347e-01
3.93794864e-01 -3.25045258e-01 1.11715102e+00 5.01505017e-01
-1.60954058e-01 7.18407929e-01 1.49372116e-01 6.46653771e-01
5.37800610e-01 -1.27453542e+00 2.09289178e-01 6.87505364e-01
5.26397943e-01 1.16570294e+00 7.30021656e-01 -2.81720847e-01
-1.34890008e+00 -7.68867791e-01 1.08221948e+00 -5.35526812e-01
8.57886434e-01 -4.44752932e-01 -9.81870115e-01 6.10965967e-01
3.63953263e-01 -1.60513911e-02 6.53924704e-01 -4.25747782e-02
-6.89221501e-01 -2.10918803e-02 -1.07493174e+00 4.94195461e-01
7.87682772e-01 -9.09982443e-01 -6.91820860e-01 2.28249237e-01
8.98569763e-01 -1.25998318e-01 -9.44045186e-01 1.27384886e-01
3.91762316e-01 -1.22360826e+00 8.10897946e-01 -3.97371352e-01
7.40391910e-01 -2.66060263e-01 -3.22886586e-01 -9.98075485e-01
-9.39560682e-02 -3.79444093e-01 3.39437872e-01 1.27428114e+00
7.61531055e-01 -6.39550686e-01 8.40868711e-01 6.28239274e-01
-1.67771995e-01 -9.85949397e-01 -5.65815449e-01 -8.24825943e-01
3.50757152e-01 -5.50562203e-01 4.41739380e-01 1.05108798e+00
2.34573781e-01 1.94889650e-01 1.18194252e-01 -2.76236773e-01
1.47128195e-01 -6.25994429e-02 7.09097922e-01 -1.05584693e+00
-6.51682973e-01 -7.41309285e-01 -6.66445971e-01 -1.05141222e+00
1.03914641e-01 -1.16414964e+00 5.77791072e-02 -1.39248407e+00
3.14075947e-02 -6.70378983e-01 -3.23050767e-01 6.38670206e-01
2.50394136e-01 -3.62392738e-02 1.09259561e-01 2.08450228e-01
-3.79279852e-01 1.48577660e-01 4.95445043e-01 -2.13862255e-01
-1.19253263e-01 -3.71887296e-01 -7.24970877e-01 6.12111866e-01
7.30485559e-01 -6.43157184e-01 -5.94488144e-01 -7.02561200e-01
4.37200665e-01 2.15593707e-02 2.18924746e-01 -1.08670592e+00
3.52047384e-01 1.70143723e-01 -8.50309655e-02 -1.23901717e-01
1.10588796e-01 -7.73469865e-01 -1.56658605e-01 4.05381978e-01
-5.21629333e-01 4.73170966e-01 3.68621141e-01 3.81552488e-01
-3.12200636e-01 -9.55663323e-01 7.68262982e-01 -4.34909433e-01
-5.23303151e-01 2.52700765e-02 -3.86706173e-01 3.21985960e-01
5.96425176e-01 -3.95825475e-01 -7.41506666e-02 -3.05965930e-01
-3.14978600e-01 4.56052013e-02 8.51690352e-01 3.55756432e-01
4.63481933e-01 -1.11789298e+00 -1.83407098e-01 2.56975502e-01
3.33755136e-01 -1.99838296e-01 -5.99631906e-01 5.72096348e-01
-6.40558839e-01 5.11960804e-01 4.51581776e-02 -4.40169334e-01
-1.06790340e+00 6.93454266e-01 4.01029646e-01 -3.07880640e-01
-2.21992254e-01 1.01332426e+00 -3.65975976e-01 -7.94527173e-01
3.65242332e-01 -2.78352916e-01 -6.98938146e-02 -1.91686407e-01
3.24819177e-01 -2.14178085e-01 3.54974046e-02 -4.66927737e-01
-3.32001507e-01 5.30030727e-01 -9.82134491e-02 -1.38326408e-02
1.39539909e+00 2.38533795e-01 -3.36261958e-01 4.19853061e-01
1.70746076e+00 3.03936284e-03 -7.44367719e-01 -2.49916017e-01
4.63371962e-01 -4.62048441e-01 -2.92258918e-01 -5.38596272e-01
-1.08593273e+00 9.72426713e-01 5.80946445e-01 5.57631433e-01
8.33681643e-01 -2.44835373e-02 5.90917647e-01 6.71194971e-01
4.08677161e-01 -9.56165135e-01 1.08683631e-01 6.55621409e-01
4.12084222e-01 -9.94349062e-01 -7.59642199e-02 1.61958567e-03
-1.58923328e-01 1.30891192e+00 5.93635559e-01 -3.75810266e-01
6.88758492e-01 3.16743135e-01 1.82787571e-02 -5.46002626e-01
-5.10200977e-01 -1.21422209e-01 -3.48704048e-02 3.59894633e-01
8.47802758e-01 -4.41277146e-01 -2.49057561e-01 1.77998811e-01
-2.45661497e-01 8.33397657e-02 3.78422201e-01 1.26181710e+00
-3.88292044e-01 -1.50086188e+00 -8.23799521e-02 7.43776500e-01
-4.69877362e-01 -3.34952682e-01 -3.52481663e-01 9.11978781e-01
2.61080861e-01 8.38739693e-01 -2.97770873e-02 -4.06476557e-01
2.38673046e-01 4.31413621e-01 3.20314020e-01 -1.01233506e+00
-8.02661717e-01 -3.33675444e-01 4.91414145e-02 -3.94647390e-01
-4.50732678e-01 -5.37431717e-01 -1.30307329e+00 -1.22682005e-01
-4.44565386e-01 3.77837867e-01 6.85149252e-01 1.01263976e+00
3.34307879e-01 1.89708441e-01 3.45078021e-01 -6.45481527e-01
-6.94238961e-01 -6.41379714e-01 -3.26100618e-01 4.92309928e-01
1.74885064e-01 -5.54136753e-01 -5.38273335e-01 -2.28728689e-02] | [7.58123254776001, 7.949474334716797] |
1729fee7-0154-41e4-9f3e-229fab323cfb | discovering-causality-for-efficient | 2306.11846 | null | https://arxiv.org/abs/2306.11846v1 | https://arxiv.org/pdf/2306.11846v1.pdf | Discovering Causality for Efficient Cooperation in Multi-Agent Environments | In cooperative Multi-Agent Reinforcement Learning (MARL) agents are required to learn behaviours as a team to achieve a common goal. However, while learning a task, some agents may end up learning sub-optimal policies, not contributing to the objective of the team. Such agents are called lazy agents due to their non-cooperative behaviours that may arise from failing to understand whether they caused the rewards. As a consequence, we observe that the emergence of cooperative behaviours is not necessarily a byproduct of being able to solve a task as a team. In this paper, we investigate the applications of causality in MARL and how it can be applied in MARL to penalise these lazy agents. We observe that causality estimations can be used to improve the credit assignment to the agents and show how it can be leveraged to improve independent learning in MARL. Furthermore, we investigate how Amortized Causal Discovery can be used to automate causality detection within MARL environments. The results demonstrate that causality relations between individual observations and the team reward can be used to detect and punish lazy agents, making them develop more intelligent behaviours. This results in improvements not only in the overall performances of the team but also in their individual capabilities. In addition, results show that Amortized Causal Discovery can be used efficiently to find causal relations in MARL. | ['Corentin Artaud', 'Varuna De Silva', 'Rafael Pina'] | 2023-06-20 | null | null | null | null | ['causal-discovery', 'multi-agent-reinforcement-learning'] | ['knowledge-base', 'methodology'] | [-1.63217038e-01 2.74676949e-01 1.43232923e-02 -6.01377860e-02
-4.60168600e-01 -5.63131273e-01 5.60189307e-01 4.59896117e-01
-6.08786583e-01 1.10080814e+00 -4.48634773e-02 -7.16421828e-02
-6.43742383e-01 -6.76031768e-01 -7.25681067e-01 -8.88492882e-01
-6.05949223e-01 8.06421280e-01 3.39072675e-01 -2.50492960e-01
1.83267266e-01 3.25308383e-01 -1.42466486e+00 -2.54686587e-02
8.36494327e-01 4.74433035e-01 1.62029088e-01 7.83349633e-01
4.91888642e-01 1.52424872e+00 -8.93610179e-01 -5.98724484e-02
2.37056091e-01 -7.11445630e-01 -7.66573071e-01 3.11496913e-01
-1.95828289e-01 -3.77526045e-01 2.94588298e-01 9.34696436e-01
2.46893957e-01 2.47785226e-01 5.46540678e-01 -1.85683918e+00
2.74839461e-01 9.16771948e-01 -6.64967120e-01 1.84612975e-01
1.93973750e-01 3.59144837e-01 1.26189303e+00 -6.80330545e-02
2.39463896e-01 1.49632394e+00 2.75598377e-01 4.14805621e-01
-1.11823320e+00 -7.80596435e-01 3.45014304e-01 4.04392809e-01
-7.81364202e-01 -3.75154227e-01 4.14131820e-01 -3.93753856e-01
9.65299010e-01 9.59731787e-02 5.89248061e-01 6.14988923e-01
3.73590857e-01 8.13198566e-01 1.16149461e+00 -4.42446858e-01
5.24522841e-01 -1.09708242e-01 -2.25278288e-01 8.83930027e-01
3.79648805e-01 4.74103868e-01 -8.56144667e-01 -3.28132153e-01
6.47030890e-01 -2.72251874e-01 1.95130527e-01 -3.79972696e-01
-1.05402374e+00 9.94942665e-01 3.34640712e-01 -1.95103455e-02
-8.05963278e-01 5.70828319e-01 3.04884195e-01 8.35893929e-01
1.07026413e-01 9.51331556e-01 -3.92500430e-01 -3.45973343e-01
-2.22397551e-01 3.21211755e-01 7.41523087e-01 1.89485192e-01
8.50982249e-01 1.03286393e-01 1.12835705e-01 4.16300714e-01
2.75228173e-01 3.42090696e-01 2.36514524e-01 -1.55572939e+00
8.00235644e-02 8.47431004e-01 2.71104366e-01 -6.15320385e-01
-6.44346535e-01 -3.16966265e-01 -2.26550475e-01 1.09121752e+00
4.68530536e-01 -4.31049854e-01 -1.92275882e-01 1.79944587e+00
3.72802347e-01 8.57821330e-02 4.29721266e-01 8.52382898e-01
-2.20636398e-01 2.97330707e-01 -1.42561734e-01 -4.85984951e-01
1.10696912e+00 -8.33366275e-01 -3.94703090e-01 -3.24473292e-01
1.08021343e+00 -4.71953571e-01 4.18585032e-01 6.92827642e-01
-1.15988016e+00 -9.31772217e-02 -7.93971539e-01 8.01917136e-01
3.71815920e-01 -5.12892246e-01 6.94753170e-01 3.69440049e-01
-7.90898979e-01 7.12811828e-01 -1.01483464e+00 -1.02574885e-01
5.25378168e-01 6.14756227e-01 -1.12908250e-02 2.33980209e-01
-9.59089994e-01 9.58469987e-01 3.13238472e-01 -3.80922705e-01
-1.32775295e+00 -3.83901387e-01 -3.76527518e-01 5.65000661e-02
1.18451929e+00 -4.56225067e-01 1.53079128e+00 -1.30057311e+00
-1.20860577e+00 1.39168158e-01 1.63931414e-01 -7.17484832e-01
6.27639353e-01 -4.49047424e-02 2.04652131e-01 1.23149045e-01
4.18620855e-01 2.90735483e-01 8.36814821e-01 -1.24192822e+00
-1.43045247e+00 -2.74121076e-01 3.59709620e-01 5.49492776e-01
-1.50016919e-01 3.41916382e-02 4.63097215e-01 -2.08199173e-01
-5.20652771e-01 -1.10857093e+00 -4.79517132e-01 -6.04406536e-01
5.91538474e-02 -7.84140050e-01 4.51065540e-01 1.47376861e-03
6.11535966e-01 -1.89927554e+00 1.61871508e-01 3.46280187e-01
4.65537488e-01 -6.57904148e-02 -2.66079247e-01 5.14172494e-01
2.96498716e-01 1.17136955e-01 1.92040384e-01 -1.96677178e-01
-3.34571637e-02 7.19211936e-01 -5.93626350e-02 5.09111166e-01
3.78789008e-01 6.07222915e-01 -1.19061601e+00 -2.36429200e-01
-2.18393113e-02 -2.64867544e-01 -4.22214180e-01 4.79833633e-01
-3.83832544e-01 5.08770585e-01 -6.27465189e-01 1.91894904e-01
-1.50884297e-02 -8.73255208e-02 4.29774672e-01 8.31577420e-01
-1.61233678e-01 3.71635884e-01 -1.23834872e+00 6.60092711e-01
-3.93768191e-01 4.31100547e-01 2.28642970e-01 -1.23418081e+00
7.76736677e-01 4.91193980e-01 5.76997399e-01 -7.80996978e-01
2.36997567e-02 2.82323092e-01 6.31557703e-01 -4.53151911e-01
3.53108943e-01 -1.38516903e-01 5.53792268e-02 1.20601189e+00
-3.09999913e-01 1.02475211e-01 6.10048771e-01 2.72901237e-01
1.46026373e+00 -1.02303326e-01 2.68804610e-01 -2.63004135e-02
3.26455742e-01 4.19897944e-01 8.90569746e-01 1.10809028e+00
-1.28888831e-01 -3.72475147e-01 1.00010157e+00 -3.47166449e-01
-9.26174283e-01 -7.58016467e-01 4.98411179e-01 1.38904798e+00
4.52798642e-02 -1.26975119e-01 -3.00246328e-01 -9.61919248e-01
9.01410878e-02 5.89461982e-01 -5.41127324e-01 -3.25630933e-01
-5.22871554e-01 -6.99917078e-01 3.85103762e-01 5.06646693e-01
2.63143450e-01 -1.29087257e+00 -1.40431154e+00 5.29711425e-01
1.96412981e-01 -7.35801995e-01 -1.71619162e-01 6.62241638e-01
-6.14176154e-01 -1.52624917e+00 -1.43380865e-01 -2.53858835e-01
7.67163694e-01 3.56438667e-01 1.03638411e+00 4.30029005e-01
3.49189863e-02 6.77403986e-01 -5.52721560e-01 -6.49760544e-01
-1.04691339e+00 -1.27868086e-01 4.42329675e-01 -1.55967683e-01
2.75104821e-01 -5.38081527e-01 -3.21678817e-01 5.16182959e-01
-6.14879549e-01 -2.49217212e-01 5.57132721e-01 9.17188466e-01
-1.08628429e-01 7.03879297e-01 9.91553724e-01 -7.06850946e-01
9.67548192e-01 -4.18263584e-01 -9.74656761e-01 1.78501114e-01
-8.56633782e-01 4.36241031e-01 8.10728729e-01 -5.95410585e-01
-8.50600958e-01 -2.28110924e-01 6.41919553e-01 -2.22277656e-01
-8.46377164e-02 4.63111639e-01 4.25005466e-01 1.21256232e-01
6.07292414e-01 -1.33193240e-01 4.51340675e-01 -3.38690318e-02
-1.13474391e-02 3.39324683e-01 -6.80648983e-02 -6.98162436e-01
7.42560208e-01 2.79064447e-01 3.29421222e-01 -2.51668334e-01
-5.70745587e-01 -2.84435302e-01 -1.44368187e-01 -5.82678735e-01
5.20138681e-01 -9.64136899e-01 -1.38096535e+00 1.77310109e-01
-7.71403730e-01 -8.19067478e-01 -3.42743009e-01 5.02894163e-01
-7.78974473e-01 -7.07726320e-03 -2.57739544e-01 -1.21865177e+00
2.44434655e-01 -1.19125032e+00 2.47640491e-01 3.97548199e-01
-3.99695337e-01 -1.13992620e+00 1.25681907e-01 4.07537639e-01
1.40803277e-01 -1.13529310e-01 7.72986352e-01 -9.20821071e-01
-6.27867222e-01 1.59852982e-01 2.76000321e-01 1.99907050e-01
1.53579205e-01 -1.36971727e-01 -5.63334823e-01 -5.02051294e-01
-1.93997964e-01 -6.61470652e-01 5.02425313e-01 2.38278493e-01
1.51363134e-01 -5.69328904e-01 -3.99853885e-02 -3.10590118e-01
1.01501977e+00 5.21520793e-01 4.25814688e-02 7.07548201e-01
3.62696081e-01 1.09754241e+00 9.59789753e-01 7.87683725e-01
4.35604095e-01 6.22482121e-01 7.66267836e-01 8.20350423e-02
1.77505612e-01 7.20995814e-02 7.63936877e-01 3.51250350e-01
-3.17741454e-01 1.15607060e-01 -9.32086468e-01 4.86810654e-01
-2.51530933e+00 -1.09691846e+00 -3.54345411e-01 2.18389606e+00
9.15286005e-01 2.23956257e-01 6.59815371e-01 4.35690163e-03
4.18042958e-01 -3.39377820e-01 -7.35614717e-01 -7.14426160e-01
3.17991003e-02 -2.12457329e-01 6.49832606e-01 6.19275212e-01
-5.84732234e-01 7.74522841e-01 5.66895533e+00 5.19277096e-01
-7.48715520e-01 1.77895993e-01 4.77218688e-01 -2.60274678e-01
3.74283232e-02 3.66199106e-01 -5.38807571e-01 8.31893682e-02
8.00908387e-01 -3.15809071e-01 8.48275244e-01 7.11306453e-01
5.92544138e-01 -6.14473224e-01 -1.23473155e+00 3.61354947e-01
-2.92582452e-01 -9.78223503e-01 -5.03965020e-01 4.07386988e-01
8.47599089e-01 6.09705807e-04 -2.95859128e-01 3.02381128e-01
1.35728538e+00 -1.08168483e+00 7.62134075e-01 3.27103168e-01
-2.86364287e-01 -1.25340009e+00 7.41508842e-01 7.10371733e-01
-8.35060537e-01 -5.28647900e-01 -2.45235488e-01 -8.11302364e-01
-2.77482748e-01 2.41019502e-01 -1.37120485e+00 3.88102680e-01
5.11106908e-01 3.83065313e-01 -5.16319931e-01 8.71236503e-01
-6.84722781e-01 6.15118921e-01 -1.72077954e-01 -3.36366683e-01
4.42670763e-01 -1.05246328e-01 7.27503240e-01 4.24395621e-01
-1.68445751e-01 -1.83497235e-01 6.51106715e-01 8.47242832e-01
1.45777747e-01 -1.79753453e-01 -3.38195026e-01 -1.77210346e-02
4.54730779e-01 1.05932057e+00 -6.96849644e-01 -2.84192711e-01
-2.01942354e-01 4.79060382e-01 6.28815591e-01 2.96649933e-01
-5.48887730e-01 1.54064611e-01 8.24478269e-01 -1.69954047e-01
2.08741590e-01 -1.34388968e-01 -1.84919119e-01 -6.98677659e-01
-1.89464167e-01 -1.27067542e+00 4.20477211e-01 -5.60521841e-01
-1.31975353e+00 1.44524843e-01 -1.73373237e-01 -9.10461783e-01
-6.25684142e-01 -7.12998286e-02 -7.02033162e-01 5.23674607e-01
-1.61948788e+00 -4.61236268e-01 4.62468788e-02 4.35190529e-01
5.01073956e-01 -3.86298269e-01 5.99111617e-01 -4.44883645e-01
-6.88374698e-01 6.62867278e-02 -2.41923681e-03 -1.46826550e-01
7.07729459e-01 -1.58361578e+00 -2.40779802e-01 8.13726664e-01
1.84325501e-01 7.56089166e-02 8.56085658e-01 -6.26942158e-01
-1.22987783e+00 -9.05810058e-01 3.65976751e-01 -2.42903903e-01
9.57808495e-01 -1.43610714e-02 -7.59422362e-01 4.57442284e-01
2.88043588e-01 -2.96017408e-01 4.97335851e-01 3.92055959e-01
-1.03755854e-01 -6.48457184e-02 -8.17602098e-01 6.48554981e-01
5.92688322e-01 -2.75218915e-02 -6.34349406e-01 1.56627253e-01
3.40610355e-01 1.79321259e-01 -7.39071131e-01 8.87040282e-04
1.60160556e-01 -1.10008395e+00 4.12524670e-01 -7.11554825e-01
4.32346791e-01 -2.93358505e-01 3.92634302e-01 -1.74674737e+00
-3.42039526e-01 -7.84829438e-01 2.26363186e-02 1.08240616e+00
4.02130097e-01 -7.78350294e-01 6.80681229e-01 5.56322336e-01
2.55268335e-01 -3.35584372e-01 -1.10858929e+00 -1.10055172e+00
2.07620651e-01 -3.62440467e-01 2.82481581e-01 9.55866992e-01
1.58219472e-01 5.38221121e-01 -4.71177161e-01 2.71436751e-01
1.04561460e+00 -1.26358122e-02 9.76491570e-01 -1.30907691e+00
-6.57406747e-01 -5.60878634e-01 -1.46577984e-01 -3.72127086e-01
2.77264088e-01 -6.48970425e-01 2.20527023e-01 -1.32046223e+00
1.93699881e-01 -9.66560364e-01 -2.12508529e-01 9.24303651e-01
-3.13368976e-01 -3.41016985e-02 5.76989233e-01 4.33620781e-01
-9.93240952e-01 1.35459766e-01 1.00112402e+00 6.41544089e-02
-4.76984560e-01 1.94248021e-01 -5.59561849e-01 7.88389564e-01
1.07048023e+00 -9.68947172e-01 -3.19146216e-01 -1.53076267e-02
8.22184205e-01 2.28936985e-01 5.68125129e-01 -8.46530318e-01
3.95596057e-01 -5.48677742e-01 -1.97709098e-01 -3.25423554e-02
-1.39577866e-01 -8.10553670e-01 -4.04614471e-02 9.14301097e-01
-5.10801017e-01 -1.96467619e-02 -8.62822831e-02 5.66611350e-01
-4.27515320e-02 -5.97414434e-01 5.52593589e-01 -3.66414636e-01
-4.66657221e-01 -4.05170232e-01 -9.95999634e-01 9.46120545e-02
1.44358432e+00 2.12322883e-02 -3.21439147e-01 -5.53566337e-01
-3.08576196e-01 9.31581140e-01 4.99215662e-01 1.69981033e-01
3.80650491e-01 -8.03876102e-01 -9.95021701e-01 -1.70357466e-01
-1.12178899e-01 -4.38160915e-03 -1.33218020e-01 1.19510448e+00
2.02378314e-02 1.16564203e-02 -4.15705651e-01 -5.18339694e-01
-1.53233242e+00 3.48181397e-01 5.31283081e-01 -5.94437420e-01
-4.40604091e-01 6.01494670e-01 1.92790315e-01 -1.60646793e-02
3.94539863e-01 -5.47006307e-03 -1.62747428e-01 3.17665070e-01
5.63013136e-01 4.15746361e-01 -2.13896915e-01 -3.16283517e-02
-4.40978199e-01 5.55726420e-03 -1.08733110e-01 -2.19619110e-01
1.63968217e+00 -1.89172015e-01 -1.61503583e-01 5.17646849e-01
3.39327306e-01 1.38374846e-02 -1.77848673e+00 -1.38343677e-01
2.58351803e-01 -2.35519335e-01 5.15991226e-02 -8.95908356e-01
-9.37314868e-01 4.01091307e-01 1.41484052e-01 8.24740648e-01
9.42109168e-01 1.47880480e-01 2.16062173e-01 5.26419759e-01
5.27988672e-01 -1.30029881e+00 5.23166597e-01 4.70120996e-01
5.21756113e-01 -1.40408719e+00 7.10143223e-02 5.60736768e-02
-9.97217953e-01 1.20300901e+00 5.23932099e-01 -2.28783324e-01
-1.53062999e-01 3.33214223e-01 1.10890195e-01 -3.32868487e-01
-1.45297551e+00 -5.09295762e-01 -4.52254802e-01 7.43452191e-01
-7.39281103e-02 2.73879766e-01 -6.72145709e-02 3.94463986e-02
3.28615382e-02 -1.19990177e-01 1.18919480e+00 1.02408457e+00
-7.23591626e-01 -1.31497908e+00 -7.12756693e-01 5.28344810e-01
-3.79768103e-01 3.97911161e-01 -4.96204555e-01 4.97828066e-01
1.35947585e-01 1.36815262e+00 1.02956757e-01 -1.21656014e-02
2.29465798e-01 -1.96868911e-01 3.56214017e-01 -5.94289064e-01
-8.06900799e-01 5.89935370e-02 3.59483063e-01 -5.33124208e-01
-7.10297108e-01 -7.91050255e-01 -1.55905068e+00 -2.02553421e-01
-2.92154968e-01 3.86970639e-01 2.47726157e-01 1.15992534e+00
1.18763320e-01 7.95799494e-01 9.77595687e-01 -3.42652619e-01
-8.48926008e-01 -5.96476078e-01 -5.66028595e-01 2.51127154e-01
4.28005427e-01 -9.05316234e-01 -6.03524089e-01 -2.12577268e-01] | [3.7839460372924805, 1.985613226890564] |
c44008ef-f298-43bc-827a-5fc2aef36b56 | multi-class-skin-cancer-classification | 2303.07520 | null | https://arxiv.org/abs/2303.07520v1 | https://arxiv.org/pdf/2303.07520v1.pdf | Multi-class Skin Cancer Classification Architecture Based on Deep Convolutional Neural Network | Skin cancer detection is challenging since different types of skin lesions share high similarities. This paper proposes a computer-based deep learning approach that will accurately identify different kinds of skin lesions. Deep learning approaches can detect skin cancer very accurately since the models learn each pixel of an image. Sometimes humans can get confused by the similarities of the skin lesions, which we can minimize by involving the machine. However, not all deep learning approaches can give better predictions. Some deep learning models have limitations, leading the model to a false-positive result. We have introduced several deep learning models to classify skin lesions to distinguish skin cancer from different types of skin lesions. Before classifying the skin lesions, data preprocessing and data augmentation methods are used. Finally, a Convolutional Neural Network (CNN) model and six transfer learning models such as Resnet-50, VGG-16, Densenet, Mobilenet, Inceptionv3, and Xception are applied to the publically available benchmark HAM10000 dataset to classify seven classes of skin lesions and to conduct a comparative analysis. The models will detect skin cancer by differentiating the cancerous cell from the non-cancerous ones. The models performance is measured using performance metrics such as precision, recall, f1 score, and accuracy. We receive accuracy of 90, 88, 88, 87, 82, and 77 percent for inceptionv3, Xception, Densenet, Mobilenet, Resnet, CNN, and VGG16, respectively. Furthermore, we develop five different stacking models such as inceptionv3-inceptionv3, Densenet-mobilenet, inceptionv3-Xception, Resnet50-Vgg16, and stack-six for classifying the skin lesions and found that the stacking models perform poorly. We achieve the highest accuracy of 78 percent among all the stacking models. | ['Alfredo Cuzzocrea', 'Sweta Sneha', 'Hossain Shahriar', 'Mst Shapna Akter'] | 2023-03-13 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [-6.81311935e-02 -1.72357827e-01 -2.37314492e-01 1.09406687e-01
-3.51264834e-01 -2.21002653e-01 3.88875037e-01 -2.28742715e-02
-4.27295953e-01 4.96340662e-01 -2.28028268e-01 -3.88894171e-01
1.06382363e-01 -1.13876152e+00 -1.91092432e-01 -9.13581491e-01
2.24936992e-01 -7.63628483e-02 4.18786794e-01 -7.90731460e-02
-3.30041088e-02 2.94802159e-01 -1.05283880e+00 7.12366641e-01
1.12680817e+00 1.05425012e+00 -1.03456691e-01 1.00251627e+00
-2.03499600e-01 8.34140003e-01 -4.00036961e-01 -3.69124472e-01
1.27096310e-01 -3.31358522e-01 -7.63920784e-01 -1.73971608e-01
3.29593003e-01 -3.03244978e-01 -3.08280349e-01 1.11284840e+00
6.53312802e-01 -5.48101902e-01 9.24844086e-01 -1.05605710e+00
-3.64239395e-01 3.38607468e-02 -9.52567935e-01 -7.39595816e-02
1.11621372e-01 2.21558020e-01 3.91452938e-01 -5.53936601e-01
4.18797076e-01 8.27048898e-01 1.28106225e+00 1.02747297e+00
-4.75056976e-01 -4.37652439e-01 -7.30700418e-02 2.71023095e-01
-1.36097836e+00 2.80477285e-01 1.56228855e-01 -5.02212524e-01
5.70030272e-01 7.47815013e-01 8.02973330e-01 1.17294228e+00
6.60786986e-01 8.65006626e-01 1.28861964e+00 -3.78818810e-01
1.75620243e-01 2.24652320e-01 3.11076224e-01 1.02583778e+00
2.54002839e-01 -7.63928965e-02 4.42255400e-02 7.18321651e-02
8.64639223e-01 4.25470918e-01 -1.59165263e-01 4.59450543e-01
-3.53933722e-01 7.17579424e-01 8.70493889e-01 2.33024985e-01
-3.63519728e-01 5.00448281e-03 5.32872498e-01 1.70731172e-01
3.28930616e-01 3.42776142e-02 -4.51319873e-01 4.45935369e-01
-3.99981916e-01 -2.06938505e-01 5.51445842e-01 5.48657775e-01
4.52897608e-01 -2.39003897e-01 -2.69628227e-01 1.09349871e+00
2.47916684e-01 1.98336244e-01 8.10494423e-01 -1.78088799e-01
-1.01350538e-01 8.92974377e-01 -3.62237275e-01 -6.86504543e-01
-7.10651278e-01 -2.56872952e-01 -1.51964653e+00 1.67633414e-01
2.38163576e-01 -4.24700648e-01 -1.67355990e+00 1.05367553e+00
1.89596951e-01 1.91055298e-01 9.49216783e-02 9.13740873e-01
1.31611264e+00 5.10717630e-01 4.78180975e-01 2.87174165e-01
1.44450712e+00 -1.14215124e+00 -3.19233447e-01 -7.32731223e-02
8.46505702e-01 -5.50469279e-01 7.96016932e-01 5.11982143e-01
-5.75720847e-01 -5.84784746e-01 -7.48186409e-01 8.49383324e-02
-7.41306424e-01 3.23674470e-01 9.21455741e-01 7.22319424e-01
-1.16129422e+00 3.97497743e-01 -7.86681294e-01 -9.90000665e-01
5.60135365e-01 1.41109928e-01 -4.46606070e-01 -2.62628317e-01
-1.11378658e+00 9.12763536e-01 3.20153832e-01 2.56047755e-01
-9.73281920e-01 -5.32450080e-01 -5.47764778e-01 -3.26528162e-01
-5.22552580e-02 -5.85875809e-01 1.02820611e+00 -1.26064336e+00
-1.04965019e+00 9.40899074e-01 1.33266941e-01 -3.38342726e-01
6.37969732e-01 7.93376714e-02 -6.87168002e-01 2.29523048e-01
-1.77293643e-01 5.80652654e-01 1.77598998e-01 -9.54442084e-01
-8.47368658e-01 -3.46786499e-01 9.75706242e-03 -6.19367510e-02
-4.84949976e-01 -1.98280662e-01 -5.82487345e-01 -1.03179447e-01
-3.09557058e-02 -7.94763386e-01 -4.88827795e-01 2.98081338e-01
-9.89992440e-01 -2.60167956e-01 7.38289177e-01 -9.13750887e-01
9.57867801e-01 -2.01203489e+00 -4.80961740e-01 3.49241644e-01
2.36038849e-01 7.94673920e-01 -4.75637168e-01 2.63794690e-01
-1.38036087e-01 5.81616819e-01 1.78653643e-01 2.18213588e-01
-5.08786261e-01 1.39883488e-01 5.56837440e-01 2.49867946e-01
2.54455775e-01 9.10673022e-01 -6.57775879e-01 -5.03784478e-01
4.50311482e-01 5.44025898e-01 -9.68893468e-02 7.54917413e-02
2.94534285e-02 -2.56835967e-01 -3.08580369e-01 1.18451142e+00
8.21856797e-01 -3.82601857e-01 1.45404264e-01 -3.50054920e-01
1.56309173e-01 -3.85341048e-01 -7.28906453e-01 9.46703792e-01
-5.04926383e-01 5.09685278e-01 1.27386013e-02 -7.08534539e-01
7.32186675e-01 3.96913141e-01 3.46792728e-01 -7.21176982e-01
2.18018621e-01 3.67748961e-02 -3.15402560e-02 -8.62761497e-01
2.03706697e-02 6.29813317e-03 3.27788860e-01 -2.24301055e-01
-5.65556921e-02 3.63227308e-01 1.59685001e-01 5.90933720e-04
1.25686061e+00 -5.18363297e-01 3.22662473e-01 -7.91822523e-02
5.14792025e-01 -3.90207991e-02 4.78641093e-01 6.43736124e-01
-3.72326165e-01 7.63310671e-01 6.61527097e-01 -8.08315754e-01
-7.09432304e-01 -1.12260067e+00 -2.07179233e-01 7.66109705e-01
1.77191794e-01 -1.51508346e-01 -1.07355654e+00 -1.11073446e+00
-2.06242651e-02 2.71789879e-01 -9.26532805e-01 -2.76329875e-01
7.91699737e-02 -1.36200643e+00 7.28093088e-01 7.03920901e-01
1.14464498e+00 -8.88861895e-01 -9.95771959e-02 -2.19712839e-01
1.09164327e-01 -6.07110143e-01 -2.51226164e-02 5.39194085e-02
-5.21913290e-01 -1.74161863e+00 -9.36641693e-01 -1.06408525e+00
1.04363823e+00 7.61915520e-02 7.32071340e-01 4.85182345e-01
-9.48624730e-01 4.17573266e-02 -3.86009753e-01 -5.30029833e-01
-2.72018760e-01 2.01301798e-02 -2.27023095e-01 -3.23225334e-02
7.47346520e-01 2.14833722e-01 -7.23707676e-01 1.80718273e-01
-8.02933812e-01 5.34314774e-02 9.35638845e-01 9.03651595e-01
5.38826406e-01 2.83804178e-01 4.65955794e-01 -1.12050509e+00
4.87591207e-01 -6.49332762e-01 -6.08785674e-02 4.53882784e-01
-5.16963661e-01 -6.27023518e-01 6.63031578e-01 -4.80154365e-01
-7.76814222e-01 3.18936557e-01 -5.26383758e-01 -1.60464346e-01
-3.69224101e-01 5.22212744e-01 -3.26936170e-02 -2.31974989e-01
8.92585218e-01 2.60346621e-01 6.49430379e-02 -3.48426551e-01
-5.65961003e-02 9.12127376e-01 2.55296141e-01 2.07675532e-01
4.19678181e-01 3.51074457e-01 -1.26767278e-01 -1.05766153e+00
-6.78362012e-01 -4.50672299e-01 -4.68060851e-01 -3.52463186e-01
1.06137526e+00 -9.13881898e-01 -3.55364531e-01 1.36958897e+00
-8.27780426e-01 -4.72029150e-01 1.55915827e-01 9.06605422e-02
2.76741028e-01 3.85913134e-01 -1.05977952e+00 -7.18333066e-01
-7.48232245e-01 -7.96123207e-01 8.53468478e-01 9.57541347e-01
-1.06042668e-01 -1.50786293e+00 -2.85125703e-01 1.82477191e-01
4.72159415e-01 5.99829197e-01 8.22999001e-01 -6.00150347e-01
-1.47026274e-02 -6.36060357e-01 -4.85569149e-01 6.07945263e-01
4.38559830e-01 6.77086294e-01 -9.54141140e-01 -2.66739726e-01
-6.53799415e-01 -3.17319036e-01 1.28403163e+00 5.46121120e-01
1.59466839e+00 -2.98806310e-01 -8.27284396e-01 8.77799392e-01
1.86269581e+00 4.29580927e-01 1.22326267e+00 2.92445660e-01
6.34325683e-01 3.28900039e-01 2.95710891e-01 -1.96799770e-01
2.98194170e-01 -1.42534465e-01 6.54869258e-01 -9.87502337e-01
-2.11092353e-01 -1.34830967e-01 8.66513029e-02 3.67415488e-01
-2.54099578e-01 -2.99013585e-01 -9.14480209e-01 6.19161308e-01
-1.21000028e+00 -7.05011249e-01 -5.21298766e-01 1.87358034e+00
7.01682687e-01 8.94770250e-02 1.26661509e-01 1.90194994e-01
8.49308550e-01 -1.30603105e-01 -5.75973511e-01 -4.50787663e-01
-1.18059153e-02 3.01064283e-01 5.55231810e-01 1.96500391e-01
-1.38147104e+00 8.39270592e-01 6.34815216e+00 6.98967695e-01
-1.79357457e+00 -1.77675337e-01 1.19855797e+00 3.14522922e-01
1.34860262e-01 -5.82149148e-01 -7.04982877e-01 6.66090548e-01
5.81279755e-01 3.58152241e-01 -8.44490081e-02 1.04623234e+00
-1.91996619e-01 -3.54313791e-01 -8.10853004e-01 8.91143620e-01
2.45803352e-02 -1.21803093e+00 9.34901461e-02 -1.37328029e-01
7.31811523e-01 2.26349477e-02 1.11539125e-01 3.16683322e-01
3.39938849e-01 -1.41421342e+00 -2.46547535e-01 8.02327335e-01
1.04461563e+00 -6.94720447e-01 1.32344174e+00 2.57257342e-01
-8.52239788e-01 -6.24730550e-02 -4.37799931e-01 2.25118846e-01
-4.83559668e-01 7.69062757e-01 -1.10331273e+00 5.43727636e-01
8.57564747e-01 5.53978145e-01 -8.90440106e-01 1.20983529e+00
-1.94406882e-01 6.59784675e-01 -5.16558215e-02 -6.13361835e-01
2.45101407e-01 2.82150537e-01 -2.63404995e-01 1.42350507e+00
2.47849837e-01 -5.58023155e-02 6.02295287e-02 6.03146255e-01
2.25259691e-01 -9.99579057e-02 -3.73931795e-01 2.55780518e-02
1.40533850e-01 1.67024171e+00 -6.15594327e-01 -2.48580411e-01
-3.55573714e-01 1.03174114e+00 -1.64003551e-01 2.75366873e-01
-7.29435265e-01 -7.89750099e-01 5.18743098e-01 2.94416189e-01
-2.36854091e-01 4.23774600e-01 -3.47994715e-01 -9.07653809e-01
-4.33087796e-01 -7.23820746e-01 5.58436871e-01 -7.83423424e-01
-1.69497228e+00 6.55944705e-01 -5.57795584e-01 -8.28595340e-01
2.32830882e-01 -1.09999907e+00 -1.26744020e+00 9.76217449e-01
-1.47395909e+00 -1.13633013e+00 -1.10610223e+00 5.57597160e-01
4.23534334e-01 -3.03510308e-01 9.22104836e-01 1.46007553e-01
-9.52482760e-01 7.82452881e-01 7.66775617e-03 8.30692410e-01
4.95785743e-01 -1.40706313e+00 3.55118126e-01 5.11899412e-01
-5.25395155e-01 2.53663868e-01 -6.81698918e-02 -5.98439872e-01
-1.02812660e+00 -1.62077343e+00 5.05028486e-01 -9.34569314e-02
3.05065185e-01 -1.02263354e-01 -7.46734917e-01 3.04213017e-01
2.14375090e-02 1.92771673e-01 8.83490324e-01 4.44155410e-02
-2.60075569e-01 -3.45417380e-01 -1.58243811e+00 5.56017756e-01
4.45082337e-01 -3.43857080e-01 1.29574507e-01 7.21364796e-01
1.94775820e-01 -3.37925255e-01 -8.46557915e-01 4.17336673e-01
6.57592237e-01 -1.03527749e+00 8.03165317e-01 -5.87395549e-01
7.89341033e-01 1.07561059e-01 1.42422348e-01 -1.40329647e+00
-4.35886025e-01 2.33322591e-01 2.25431830e-01 9.59877789e-01
4.40915942e-01 -7.85237610e-01 1.09737527e+00 3.03460151e-01
1.06861964e-01 -1.23656082e+00 -5.03056586e-01 -5.50993323e-01
3.71920347e-01 1.48095384e-01 3.77830595e-01 1.09811568e+00
-2.69531965e-01 -1.61396042e-02 -1.18096471e-01 2.24145293e-01
5.08550644e-01 -5.70517719e-01 5.15260756e-01 -1.14075494e+00
2.51070559e-01 -3.33816797e-01 -5.92359364e-01 -4.61465865e-01
-5.06335199e-01 -6.45511329e-01 -3.27589154e-01 -2.07314944e+00
4.71016049e-01 -4.67667609e-01 -6.40738249e-01 9.39397335e-01
-3.16191971e-01 4.43837494e-01 -1.56677037e-01 -2.53375620e-01
-3.53223681e-01 -2.67393291e-01 1.44824278e+00 -4.74339724e-01
7.44033745e-03 7.50710219e-02 -7.60329127e-01 9.72571969e-01
1.01491940e+00 1.73629239e-01 -7.86950514e-02 -4.36129689e-01
-3.27853382e-01 -2.80115485e-01 5.46920896e-01 -1.23935294e+00
3.96717936e-01 -2.94340014e-01 1.12883759e+00 -5.24041772e-01
1.94397420e-01 -3.84869009e-01 -3.84873524e-02 1.12005663e+00
-1.76998928e-01 -5.43718696e-01 1.80093169e-01 1.96640030e-01
-9.87509042e-02 -2.23305255e-01 1.16648173e+00 -4.49268162e-01
-1.16749322e+00 5.38591146e-01 -5.20366669e-01 -5.78354716e-01
1.34854031e+00 -3.39517206e-01 -6.82832479e-01 -5.38124982e-03
-8.43971133e-01 2.80506670e-01 1.92988619e-01 3.04287255e-01
7.39131570e-01 -1.30194175e+00 -7.25657463e-01 3.09461594e-01
8.43927339e-02 4.91566211e-02 5.76232016e-01 7.41973817e-01
-1.04862893e+00 2.28318676e-01 -5.06451070e-01 -6.65460944e-01
-1.62381697e+00 1.56418219e-01 1.05840063e+00 -1.36672050e-01
-1.12910978e-01 1.19741321e+00 1.52708426e-01 -6.42424881e-01
2.98678637e-01 -3.80339354e-01 -3.09374452e-01 -2.93422788e-01
6.49544954e-01 4.97203290e-01 4.12482768e-02 -5.89834340e-02
-4.21108425e-01 6.46273732e-01 -5.03636539e-01 8.02054048e-01
8.76172125e-01 5.26348948e-01 -2.28431061e-01 -2.38022760e-01
1.25458848e+00 -4.79429126e-01 -8.34663808e-01 2.89155692e-01
-2.03566372e-01 -1.44786909e-01 2.99163703e-02 -1.47349465e+00
-1.41915298e+00 9.57655966e-01 1.24152863e+00 3.13227296e-01
1.51852953e+00 -2.92461783e-01 8.18194211e-01 1.05222069e-01
1.83962047e-01 -1.21816647e+00 1.68713450e-01 3.55041265e-01
4.84234720e-01 -1.31538785e+00 -1.33357778e-01 -5.55223703e-01
-6.20749593e-01 1.24491370e+00 1.32138789e+00 -3.01699847e-01
7.24776149e-01 3.93900782e-01 3.41255397e-01 -5.87263778e-02
-8.27058852e-01 -2.27975175e-01 9.39102694e-02 1.00976121e+00
4.07562107e-01 2.48698488e-01 -3.83119017e-01 6.78954363e-01
1.87910095e-01 1.01250380e-01 4.39209104e-01 6.73400402e-01
-5.16698658e-01 -7.71385729e-01 -1.73351556e-01 8.99465799e-01
-5.70813477e-01 1.00538088e-02 -1.05539119e+00 1.20215511e+00
4.10811812e-01 8.35136592e-01 1.86794847e-01 -9.87591803e-01
1.96111426e-01 -2.77683437e-01 2.52984375e-01 -3.75605464e-01
-6.53891087e-01 -1.06565930e-01 2.31092069e-02 -1.92756444e-01
6.15814030e-02 -1.02075366e-02 -1.20508325e+00 -3.64996940e-01
-3.64958197e-01 5.57165174e-03 6.37826204e-01 6.36382639e-01
4.84305955e-02 6.26710951e-01 6.36237204e-01 -1.31236389e-01
-3.65740746e-01 -1.09122038e+00 -7.32465684e-01 1.50253117e-01
1.70645565e-01 -1.00192077e-01 -3.11812431e-01 -2.06387937e-01] | [15.680010795593262, -2.9975733757019043] |
b1a66dc2-1caa-491c-9ac7-410da84df7e8 | mere-account-mein-kitna-balance-hai-on | 2010.16411 | null | https://arxiv.org/abs/2010.16411v1 | https://arxiv.org/pdf/2010.16411v1.pdf | Mere account mein kitna balance hai? -- On building voice enabled Banking Services for Multilingual Communities | Tremendous progress in speech and language processing has brought language technologies closer to daily human life. Voice technology has the potential to act as a horizontal enabling layer across all aspects of digitization. It is especially beneficial to rural communities in scenarios like a pandemic. In this work we present our initial exploratory work towards one such direction -- building voice enabled banking services for multilingual societies. Speech interaction for typical banking transactions in multilingual communities involves the presence of filled pauses and is characterized by Code Mixing. Code Mixing is a phenomenon where lexical items from one language are embedded in the utterance of another. Therefore speech systems deployed for banking applications should be able to process such content. In our work we investigate various training strategies for building speech based intent recognition systems. We present our results using a Naive Bayes classifier on approximate acoustic phone units using the Allosaurus library. | ['Alan W Black', 'Sai Krishna Rallabandi', 'Akshat Gupta'] | 2020-10-09 | null | null | null | null | ['intent-recognition'] | ['natural-language-processing'] | [ 6.27254881e-03 -5.67547759e-05 -2.63772532e-02 -3.61981839e-01
-7.92468131e-01 -5.55991113e-01 7.73128390e-01 1.23153172e-01
-3.58983785e-01 5.04879594e-01 9.25407827e-01 -9.07501161e-01
3.03056985e-01 -4.49441731e-01 5.68946153e-02 -4.16760057e-01
-1.56180011e-02 5.50117254e-01 7.38004893e-02 -5.62937021e-01
2.69890845e-01 5.44329464e-01 -1.17000759e+00 6.36001945e-01
4.19503510e-01 2.96630293e-01 5.90188324e-01 1.03372121e+00
-8.06997895e-01 7.60958493e-01 -6.30618513e-01 -3.90460312e-01
-1.24646485e-01 -3.06210488e-01 -8.76471817e-01 2.29485631e-01
-4.02853757e-01 -1.66134268e-01 1.23347063e-02 7.79389858e-01
6.76008999e-01 7.33719170e-02 4.23143029e-01 -6.83029175e-01
-1.51759520e-01 1.07257330e+00 -1.64822251e-01 4.46995169e-01
9.77995932e-01 -1.53243348e-01 7.49974012e-01 -8.85554790e-01
5.38739622e-01 1.34311068e+00 7.57272661e-01 2.99491227e-01
-9.72690940e-01 -3.30217540e-01 -2.29265630e-01 -2.00213343e-01
-1.38506925e+00 -1.17780328e+00 4.82189059e-01 -3.92559648e-01
1.85620153e+00 3.19573820e-01 6.08436346e-01 1.00847125e+00
1.19530328e-01 5.35883069e-01 7.08689749e-01 -8.52032304e-01
9.15362760e-02 5.95388591e-01 -1.59550965e-01 4.67271775e-01
-3.17118704e-01 -2.58580208e-01 -5.81412137e-01 -2.29961395e-01
4.06136572e-01 -2.80447572e-01 -1.53613627e-01 3.01601619e-01
-1.29394758e+00 9.62614179e-01 -1.85242161e-01 8.28764021e-01
-4.08996105e-01 -2.73558740e-02 4.43061918e-01 4.02463406e-01
2.68841177e-01 -5.15149198e-02 -2.34745622e-01 -8.85170102e-01
-8.90097797e-01 -1.87159553e-01 1.34015012e+00 9.19724286e-01
3.29988748e-01 5.78193702e-02 5.92981160e-01 1.34258997e+00
7.76386201e-01 4.06044990e-01 8.09371710e-01 -8.08065832e-01
5.57931304e-01 1.41525909e-01 -9.04928967e-02 -7.62767315e-01
-2.89742768e-01 7.14891404e-02 -2.51327991e-01 -4.71476763e-01
1.52564242e-01 -2.33416468e-01 -5.38505852e-01 1.10507953e+00
3.22225690e-01 -3.78059417e-01 4.15281415e-01 2.81031787e-01
5.45814455e-01 1.01073098e+00 4.37550023e-02 -3.94374430e-01
1.52559817e+00 -6.00996077e-01 -9.49840248e-01 -1.98107243e-01
7.14598536e-01 -1.53702867e+00 1.17884815e+00 2.98462361e-01
-8.43670845e-01 -1.75657734e-01 -8.97486269e-01 8.16871449e-02
-3.88789713e-01 -2.74207234e-01 5.30145824e-01 1.48886323e+00
-1.23602247e+00 -8.68549868e-02 -8.45156670e-01 -1.07208312e+00
-2.68763334e-01 1.62506044e-01 -2.11208627e-01 1.90202028e-01
-1.00809085e+00 8.39473069e-01 4.71128672e-02 -1.22729518e-01
-3.57473582e-01 9.06370059e-02 -8.59142900e-01 -1.77454546e-01
-1.04967371e-01 -1.26093924e-01 1.63428915e+00 -7.15673268e-01
-1.60293651e+00 9.24285948e-01 -3.46784890e-01 -4.53398943e-01
1.29029661e-01 -1.29722673e-02 -9.72886443e-01 2.27685273e-03
-1.50170773e-02 5.51786542e-01 6.30631804e-01 -9.86410022e-01
-8.02694499e-01 -4.52093817e-02 -3.90806854e-01 3.65025252e-01
-3.27342927e-01 6.75505102e-01 -1.76086068e-01 -7.12601244e-01
7.84297734e-02 -9.38094139e-01 1.77710086e-01 -5.88450730e-01
-7.72418231e-02 1.09649457e-01 8.15513551e-01 -9.60988700e-01
1.44825017e+00 -2.10776615e+00 -3.77167284e-01 1.43774942e-01
-3.68562311e-01 -1.02949981e-03 3.46489340e-01 1.11060786e+00
2.48495519e-01 2.15146184e-01 -1.28678262e-01 -4.03958499e-01
7.01963753e-02 5.68469644e-01 -4.70166445e-01 2.55413085e-01
-2.01860443e-01 3.86545092e-01 -6.72981977e-01 -5.30490279e-01
1.37503639e-01 6.77983522e-01 -4.70811546e-01 -6.34984896e-02
2.78186530e-01 -9.23654065e-02 -2.64970995e-02 9.02628601e-01
8.53417143e-02 2.96418071e-01 4.73242700e-01 4.14456993e-01
-4.37622339e-01 8.21150422e-01 -9.60901618e-01 1.53493035e+00
-9.52257395e-01 7.27809131e-01 5.63867331e-01 -6.23408735e-01
8.21324050e-01 7.90666699e-01 1.26189590e-01 -3.30164164e-01
1.81693971e-01 5.37933707e-01 3.04495484e-01 -7.43910611e-01
7.29244292e-01 -5.93325377e-01 -1.55554980e-01 5.46108246e-01
-2.86717433e-02 -3.51529002e-01 -2.81035483e-01 8.72353166e-02
9.66904104e-01 -4.31748241e-01 7.62606263e-01 -3.58024180e-01
4.81200010e-01 6.02476448e-02 8.43291506e-02 4.87957150e-01
-5.50602913e-01 2.49631271e-01 -3.21062258e-03 -8.72341469e-02
-9.16567445e-01 -8.89651716e-01 -2.31406808e-01 1.12753451e+00
-4.32138801e-01 -4.03041005e-01 -6.83271050e-01 -1.99328110e-01
-3.03301930e-01 9.97587264e-01 2.20086217e-01 3.02659482e-01
-5.15351951e-01 -3.29347521e-01 9.67079043e-01 2.30900452e-01
2.55165309e-01 -1.32246697e+00 -3.99220884e-01 7.29952812e-01
-3.62826914e-01 -1.22224653e+00 -7.09995806e-01 3.83666605e-01
-6.71068430e-01 -2.65172422e-01 -4.99567479e-01 -1.06943870e+00
1.25389658e-02 3.62299055e-01 7.93046296e-01 -1.35152221e-01
-1.37202978e-01 5.86060345e-01 -6.31169796e-01 -5.90451002e-01
-1.12123907e+00 1.37960449e-01 3.14859957e-01 -9.92459953e-02
6.62632883e-01 -5.54060102e-01 -2.24979952e-01 2.90679663e-01
-6.40161276e-01 -2.45481834e-01 3.72931540e-01 5.23027122e-01
-2.70498455e-01 -2.70447228e-02 6.48166835e-01 -5.13452232e-01
1.03633130e+00 -7.22170234e-01 1.49126381e-01 6.96604326e-02
-1.78502828e-01 -2.85515308e-01 2.21708149e-01 -2.90191799e-01
-1.22030842e+00 7.65257999e-02 -7.28181541e-01 4.71145928e-01
-3.46522242e-01 7.15600371e-01 -2.12690860e-01 1.78581536e-01
6.03834927e-01 1.66575462e-01 1.11337729e-01 -4.05675262e-01
2.21456021e-01 1.73717546e+00 3.36890161e-01 -4.87644523e-01
1.95946604e-01 4.25124019e-01 -9.06994283e-01 -1.97733080e+00
3.35507482e-01 -1.14561129e+00 -3.30776602e-01 -5.06483555e-01
8.83907497e-01 -7.56863773e-01 -3.63817841e-01 3.66333932e-01
-1.24288857e+00 -1.90145239e-01 -6.05637059e-02 5.55703163e-01
-3.49249154e-01 5.51246166e-01 -7.26467609e-01 -1.28509080e+00
-1.14480205e-01 -1.07947814e+00 1.06033325e+00 9.47755426e-02
-8.51847172e-01 -1.20361292e+00 3.64987582e-01 6.71196222e-01
4.74168658e-01 -5.92147648e-01 7.90730715e-01 -9.69613791e-01
8.16955566e-02 -1.19734772e-01 3.47797334e-01 2.54354686e-01
5.08091152e-01 9.54733938e-02 -1.09116685e+00 -8.96322206e-02
3.13293368e-01 -6.74194843e-02 1.96533427e-01 8.36187676e-02
1.61018223e-02 -2.54283965e-01 -8.86575580e-02 -5.78936636e-02
1.05165327e+00 8.68136287e-01 5.26364982e-01 3.09084654e-01
2.91129261e-01 9.27749038e-01 2.43393883e-01 3.99485797e-01
4.53592569e-01 6.03531957e-01 -3.71910483e-01 3.47186923e-01
-7.55857080e-02 -2.34807014e-01 7.79563069e-01 1.70569706e+00
3.70550454e-01 -3.72423798e-01 -1.48572004e+00 7.77112544e-01
-1.37666774e+00 -1.04553485e+00 -5.63018918e-02 2.03102970e+00
8.14896762e-01 1.73073068e-01 1.80528134e-01 1.25858411e-01
6.92476928e-01 -3.18128392e-02 2.26713881e-01 -9.98264432e-01
1.68407381e-01 2.27367312e-01 2.59496331e-01 9.36770976e-01
-7.64552712e-01 9.65083420e-01 6.79007196e+00 6.16878569e-01
-1.06892765e+00 3.04833829e-01 2.47846901e-01 1.85195684e-01
-3.35639387e-01 1.07607329e-02 -6.79127276e-01 3.41890424e-01
1.43107760e+00 5.01505211e-02 6.66678786e-01 6.91806138e-01
6.38488352e-01 -4.44647282e-01 -7.28516221e-01 9.54678535e-01
1.32065415e-01 -1.02938414e+00 -2.39957646e-01 2.81660169e-01
2.76762873e-01 3.58510733e-01 -2.89859027e-01 1.39285579e-01
3.38074744e-01 -8.98492277e-01 7.47024775e-01 1.16648711e-01
5.55497885e-01 -6.57098114e-01 6.36664689e-01 4.00654942e-01
-1.32957232e+00 2.63858195e-02 -6.65835617e-03 -3.23539078e-02
7.94062912e-01 3.04507196e-01 -1.38356090e+00 1.11412920e-01
4.87810373e-01 6.61749169e-02 -7.82392845e-02 8.42030287e-01
2.58899301e-01 8.25331926e-01 -6.50387526e-01 -2.23986670e-01
1.99030310e-01 -3.23770970e-01 7.39087939e-01 1.63803530e+00
7.18825877e-01 -3.17303129e-02 -2.65250318e-02 2.15293765e-02
2.85800368e-01 4.86850500e-01 -1.13017714e+00 -5.11508524e-01
6.89538717e-01 9.85690653e-01 -1.15916324e+00 -1.36540636e-01
-7.07422018e-01 9.89164114e-01 -3.07928264e-01 -7.87509754e-02
-4.23882723e-01 -3.98333937e-01 7.48634577e-01 1.04847603e-01
1.15034670e-01 -5.99819720e-01 -1.53346196e-01 -8.02252889e-01
-2.34173238e-01 -1.36543071e+00 3.04270741e-02 -6.25015676e-01
-8.32082808e-01 6.67748511e-01 -3.96674186e-01 -8.30421507e-01
-5.36285400e-01 -4.14932817e-01 -4.79412943e-01 8.22744250e-01
-7.29628026e-01 -1.00520420e+00 2.93738633e-01 2.50443578e-01
9.73642170e-01 -6.46705270e-01 1.22723556e+00 5.01096606e-01
-1.79145157e-01 1.63453937e-01 3.99461091e-02 5.98599762e-03
5.67766428e-01 -1.02443492e+00 5.71984828e-01 7.98029423e-01
7.40614235e-01 1.06960905e+00 8.40161979e-01 -7.16361582e-01
-1.38166904e+00 -3.59305888e-01 1.53349686e+00 -3.76449704e-01
9.73738134e-01 -4.13852394e-01 -7.05830574e-01 5.65755904e-01
6.52258873e-01 -9.18150783e-01 1.22566879e+00 1.48763821e-01
6.03026636e-02 9.18433517e-02 -1.01709282e+00 5.93914986e-01
7.50828803e-01 -1.26146162e+00 -8.36744726e-01 1.60023078e-01
6.57028556e-01 1.03058264e-01 -6.51351750e-01 -3.08972687e-01
7.36733794e-01 -9.15116310e-01 5.81556737e-01 -1.28084570e-01
-1.07708938e-01 -1.85004354e-01 -6.54890358e-01 -1.11461222e+00
2.97129869e-01 -1.18801570e+00 3.26474637e-01 1.58325231e+00
6.49761915e-01 -6.99206114e-01 4.11812991e-01 3.02262992e-01
-2.72762209e-01 -1.06990479e-01 -1.06439614e+00 -5.58572710e-01
-1.58535197e-01 -9.63826597e-01 3.51037890e-01 8.27186704e-01
7.23336279e-01 5.64090014e-01 -1.58173099e-01 7.85153285e-02
9.33361128e-02 -6.18819952e-01 4.88350809e-01 -7.39008129e-01
-4.65523571e-01 -5.01097620e-01 -3.00353825e-01 -9.73382592e-01
-2.17857674e-01 -7.80854583e-01 3.10845107e-01 -1.37979090e+00
-2.28173748e-01 -2.86919683e-01 4.34170634e-01 1.28200337e-01
5.22784233e-01 -7.36169368e-02 8.96687284e-02 1.84836760e-01
1.62932411e-01 1.29039720e-01 3.57411474e-01 -2.94703897e-03
-4.79098320e-01 1.94204941e-01 -5.15500784e-01 6.41291797e-01
8.55205417e-01 -3.78887862e-01 -6.14711642e-01 -9.32046399e-02
3.37527156e-01 4.27503616e-01 -3.47708195e-01 -9.75770891e-01
3.57235640e-01 7.65246600e-02 -3.43419939e-01 -4.09114420e-01
2.44609326e-01 -1.04363489e+00 5.09814084e-01 4.38502520e-01
1.39378309e-01 2.55001009e-01 1.76889583e-01 3.70168746e-01
-2.99927980e-01 -4.92261976e-01 5.01641929e-01 -1.53393432e-01
-5.75630963e-01 -4.89666939e-01 -1.45990050e+00 -1.79982588e-01
8.02713573e-01 -3.81966829e-01 -1.09836208e-02 -8.22039485e-01
-7.47815311e-01 -6.34904727e-02 4.84710574e-01 5.74606895e-01
5.03567517e-01 -9.42054272e-01 -2.90501535e-01 3.88980359e-01
-7.56705506e-03 -6.27548039e-01 -2.53377736e-01 6.77736580e-01
-1.10640156e+00 6.99246526e-01 5.59810065e-02 -3.21841449e-01
-1.51798713e+00 9.38541442e-02 3.48830998e-01 1.00058593e-01
-4.00954217e-01 8.17466080e-01 -4.37375963e-01 -6.12355530e-01
3.86840820e-01 -3.22944582e-01 -2.64248639e-01 4.11342680e-01
6.29206717e-01 3.75418335e-01 2.56761640e-01 -1.03673267e+00
-5.29576600e-01 -7.51637444e-02 -9.16289911e-02 -1.04910135e+00
1.06700587e+00 -5.39144099e-01 -1.29555166e-01 9.46050882e-01
1.13518274e+00 5.77369571e-01 -5.21483123e-01 2.03499198e-01
5.14436960e-01 -1.76580086e-01 -1.50465267e-03 -4.81338501e-01
-3.22296172e-01 7.23531246e-01 4.58942801e-01 7.72234559e-01
7.11300433e-01 8.82951021e-02 7.95012295e-01 5.75492918e-01
6.55360401e-01 -1.39036250e+00 -3.25929463e-01 6.46345735e-01
7.79622018e-01 -8.93933654e-01 -3.67155582e-01 -1.79046303e-01
-9.35388744e-01 1.07573056e+00 -2.49664620e-01 5.19382656e-01
1.04498458e+00 6.96865976e-01 5.40811062e-01 -1.73827216e-01
-5.82886100e-01 -3.03560346e-01 -3.14287007e-01 7.61909902e-01
8.84386539e-01 3.56704742e-01 -4.48040485e-01 2.44763732e-01
-5.14825523e-01 -4.14884865e-01 5.12805402e-01 1.22638738e+00
-7.99046934e-01 -1.48604298e+00 -5.52322626e-01 3.08531284e-01
-6.72193527e-01 -4.51541394e-01 -6.13384485e-01 5.64129591e-01
-5.21958321e-02 1.40617239e+00 1.02843478e-01 -4.75858986e-01
1.20703064e-01 7.30450273e-01 5.40607795e-02 -7.84194529e-01
-7.45550454e-01 5.53994596e-01 7.19069779e-01 -2.04637408e-01
-5.54018319e-01 -1.19255435e+00 -1.41507089e+00 -4.81130600e-01
-1.67054072e-01 2.86520094e-01 1.10797262e+00 8.29060435e-01
-4.84551489e-02 1.47041842e-01 4.43259180e-01 -6.89580977e-01
-2.55930215e-01 -1.04267216e+00 -6.84651494e-01 -3.01987439e-01
2.57371277e-01 -8.29044357e-02 -3.90106201e-01 4.33043629e-01] | [14.264182090759277, 6.998344421386719] |
5722964f-ccda-4c9b-aafb-5f439eb06a1d | safe-exploration-for-constrained | 2112.00885 | null | https://arxiv.org/abs/2112.00885v3 | https://arxiv.org/pdf/2112.00885v3.pdf | DOPE: Doubly Optimistic and Pessimistic Exploration for Safe Reinforcement Learning | Safe reinforcement learning is extremely challenging--not only must the agent explore an unknown environment, it must do so while ensuring no safety constraint violations. We formulate this safe reinforcement learning (RL) problem using the framework of a finite-horizon Constrained Markov Decision Process (CMDP) with an unknown transition probability function, where we model the safety requirements as constraints on the expected cumulative costs that must be satisfied during all episodes of learning. We propose a model-based safe RL algorithm that we call Doubly Optimistic and Pessimistic Exploration (DOPE), and show that it achieves an objective regret $\tilde{O}(|\mathcal{S}|\sqrt{|\mathcal{A}| K})$ without violating the safety constraints during learning, where $|\mathcal{S}|$ is the number of states, $|\mathcal{A}|$ is the number of actions, and $K$ is the number of learning episodes. Our key idea is to combine a reward bonus for exploration (optimism) with a conservative constraint (pessimism), in addition to the standard optimistic model-based exploration. DOPE is not only able to improve the objective regret bound, but also shows a significant empirical performance improvement as compared to earlier optimism-pessimism approaches. | ['Jean-Francois Chamberland', 'Srinivas Shakkottai', 'Dileep Kalathil', 'Aria HasanzadeZonuzy', 'Archana Bura'] | 2021-12-01 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 1.26349172e-02 7.12742150e-01 -3.01227003e-01 -9.26995650e-02
-8.69778216e-01 -5.40622711e-01 3.03691000e-01 3.43397111e-01
-8.79422307e-01 1.24175608e+00 -2.54740179e-01 -6.23758912e-01
-4.21197772e-01 -8.74039471e-01 -8.31784785e-01 -7.85395741e-01
-5.75561106e-01 4.93907064e-01 1.69705451e-01 -1.36827901e-01
2.14799449e-01 1.74484059e-01 -1.27199221e+00 -3.71612221e-01
8.38545442e-01 1.28702676e+00 -5.12737632e-02 4.44151133e-01
3.37957114e-01 9.69113111e-01 -4.35022026e-01 -1.66598320e-01
4.43147928e-01 -5.48083425e-01 -9.54745531e-01 8.28807428e-02
-3.96720946e-01 -6.10453665e-01 -1.95098460e-01 1.36527324e+00
1.29714772e-01 3.46659511e-01 2.39106640e-01 -1.41197348e+00
1.52045891e-01 7.62236536e-01 -6.06915236e-01 -2.50120852e-02
1.55658782e-01 6.30757034e-01 8.17783594e-01 2.18041092e-01
4.55348283e-01 1.27895272e+00 -1.02399893e-01 7.26807654e-01
-1.03826022e+00 -6.48550868e-01 6.83528244e-01 1.19903065e-01
-1.03396010e+00 2.52131652e-02 2.29968697e-01 -2.06110045e-01
1.00006771e+00 3.08842778e-01 9.20767009e-01 6.93002880e-01
6.19777620e-01 7.96699703e-01 1.37238884e+00 -3.66566777e-01
1.10067058e+00 -1.86919942e-01 -1.55692324e-01 6.55910492e-01
3.31165165e-01 7.12602913e-01 -3.58060986e-01 1.77848770e-03
6.40962005e-01 -6.36126101e-02 1.52965998e-02 -4.25139368e-01
-9.23314750e-01 1.02566397e+00 -1.09787032e-01 -3.20962936e-01
-4.15999293e-01 4.89429325e-01 3.38670671e-01 4.64037240e-01
-1.54776335e-01 4.79212284e-01 -2.44369403e-01 -5.00848114e-01
-7.28160322e-01 5.26229739e-01 7.24170089e-01 8.21260571e-01
5.99023581e-01 4.98656005e-01 -1.07525900e-01 -5.26827537e-02
2.09269166e-01 5.81810415e-01 -9.64089576e-03 -1.51915371e+00
5.61629593e-01 3.65742922e-01 8.15832376e-01 -1.48871675e-01
-3.24842304e-01 -3.02460223e-01 -3.75361949e-01 9.11175370e-01
4.48152930e-01 -7.31409609e-01 -7.14342594e-01 1.99086905e+00
3.06986123e-01 -3.93431753e-01 3.68909597e-01 5.98710001e-01
-4.40312505e-01 7.64268219e-01 -8.20391849e-02 -9.57450926e-01
9.01243329e-01 -5.79764843e-01 -6.06887221e-01 -4.50120091e-01
4.19118136e-01 -9.21760127e-02 9.94493008e-01 8.14162910e-01
-1.53348875e+00 2.11539909e-01 -1.21249175e+00 8.55909705e-01
1.71393067e-01 -5.85511923e-01 6.90936267e-01 7.99810350e-01
-7.14170516e-01 6.84361815e-01 -1.18748057e+00 3.66737038e-01
1.70076385e-01 5.65295577e-01 -8.84022936e-02 1.05962612e-01
-9.12552357e-01 1.01629853e+00 7.44383991e-01 2.36598067e-02
-1.84021664e+00 -1.41361400e-01 -8.78932297e-01 2.05567643e-01
1.30223680e+00 -3.76790278e-02 1.58193123e+00 -4.88935828e-01
-1.74754751e+00 7.31220841e-02 4.01853979e-01 -8.85567904e-01
7.59510100e-01 -1.26870573e-01 -2.59035118e-02 1.53205022e-01
-1.72324911e-01 4.45068300e-01 5.91101468e-01 -1.11308455e+00
-7.57954836e-01 -3.98473024e-01 5.25945365e-01 3.04561377e-01
-7.95546994e-02 -6.32203370e-02 1.63553894e-01 -1.65309846e-01
-4.61665094e-02 -1.29147255e+00 -8.10373068e-01 -4.05453265e-01
-2.27290750e-01 -5.34052029e-03 3.29487383e-01 -1.74810305e-01
1.29930770e+00 -1.88017440e+00 9.57239792e-02 4.11709875e-01
-2.69900948e-01 -1.18839648e-02 1.25758350e-01 5.07035851e-01
3.54615360e-01 -1.09958902e-01 -4.45874125e-01 2.06893235e-02
2.44742900e-01 5.18734217e-01 -5.48056066e-01 4.01947856e-01
-3.81918371e-01 3.83903384e-01 -9.72408116e-01 -1.32978469e-01
1.42398030e-01 -2.21912965e-01 -5.96336782e-01 1.55895591e-01
-6.88319862e-01 2.10211441e-01 -5.41201234e-01 4.81482744e-01
3.85564417e-01 1.53886914e-01 5.11468709e-01 7.30905354e-01
-3.48534524e-01 4.34151217e-02 -1.76469362e+00 1.09452415e+00
-2.47966990e-01 -1.16656363e-01 2.51086563e-01 -7.27127314e-01
5.46803653e-01 2.41555482e-01 5.73066235e-01 -7.55859613e-01
1.93993658e-01 1.83825660e-02 -1.15641639e-01 4.90841679e-02
1.98958293e-01 -4.45108891e-01 -4.16494012e-01 7.54702151e-01
-2.27924794e-01 -2.89643079e-01 2.52324075e-01 9.42836031e-02
1.22023070e+00 3.44488323e-01 3.26267421e-01 -3.79409343e-01
2.47127533e-01 1.61151942e-02 1.14091861e+00 8.86332929e-01
-4.42399442e-01 -4.62693274e-01 1.29078496e+00 -3.31051528e-01
-6.91305816e-01 -1.00697958e+00 2.78717607e-01 7.92982697e-01
3.60180676e-01 -3.00322413e-01 -9.23747897e-01 -9.32424664e-01
-6.83831349e-02 1.25293219e+00 -6.45015836e-01 -3.16715628e-01
-3.84190142e-01 -5.96410036e-01 2.95981735e-01 4.64639604e-01
5.11535645e-01 -1.08556044e+00 -1.52009881e+00 1.85181126e-01
2.26622626e-01 -5.97807169e-01 -2.97844827e-01 7.41605163e-01
-6.63147390e-01 -8.41137648e-01 -2.86027580e-01 7.95042887e-02
7.09835887e-01 -3.91871214e-01 5.35480082e-01 -4.93298650e-01
1.08152434e-01 3.71478945e-01 -6.68694004e-02 -7.81837344e-01
-2.93589920e-01 -4.27790046e-01 3.29860568e-01 -4.17988837e-01
-7.34276772e-02 -2.40456402e-01 -6.24249101e-01 6.12732284e-02
-9.52970207e-01 -4.26475443e-02 4.52459246e-01 6.38899505e-01
7.96484411e-01 3.95536900e-01 6.05872154e-01 -4.96173948e-01
5.25143802e-01 -5.76453730e-02 -1.37781858e+00 3.07440490e-01
-9.66553986e-01 4.63491082e-01 5.54179370e-01 -3.99262667e-01
-1.09845412e+00 1.17768198e-01 1.80221662e-01 -3.25475425e-01
1.91285223e-01 2.62041122e-01 -1.64714068e-01 4.47369404e-02
4.59436238e-01 4.03706670e-01 1.25539094e-01 -5.07864654e-02
8.69618729e-04 1.44448057e-01 2.05498144e-01 -8.68184268e-01
5.34903169e-01 2.20256239e-01 4.69810218e-01 -8.52236524e-02
-7.67602205e-01 2.77960479e-01 3.70126553e-02 -3.86649102e-01
4.60303485e-01 -7.49037504e-01 -1.75896704e+00 1.14149051e-02
-3.74192894e-01 -8.94203603e-01 -6.82855368e-01 5.56997657e-01
-1.19448137e+00 3.70226741e-01 -3.98283243e-01 -1.61043870e+00
-1.24533273e-01 -1.30949008e+00 2.40364015e-01 3.47466707e-01
1.67264659e-02 -4.81548011e-01 -8.39322060e-02 1.61737248e-01
1.49767593e-01 6.26658082e-01 7.66930699e-01 -2.99018413e-01
-8.41701508e-01 -2.94245454e-03 3.66605341e-01 5.15308380e-01
-3.05299163e-01 -2.62445480e-01 -2.69560009e-01 -7.91656375e-01
2.27182806e-01 -6.72258675e-01 5.46841383e-01 2.56356746e-01
1.07345653e+00 -1.14251935e+00 2.25028321e-02 3.15725654e-02
1.53416312e+00 8.53582382e-01 4.36289728e-01 5.34080684e-01
-9.37838778e-02 4.41931963e-01 1.25131297e+00 1.20638585e+00
3.26625377e-01 3.93259436e-01 1.10651302e+00 6.04320884e-01
7.88717449e-01 -7.97016919e-02 8.41113269e-01 -2.31515601e-01
-1.35689303e-02 2.54041329e-02 -8.28542411e-01 5.96152425e-01
-1.98907423e+00 -8.86178732e-01 5.12963951e-01 2.89247131e+00
1.20264363e+00 6.89772666e-01 3.48977745e-01 9.25097615e-02
1.74256369e-01 2.17258539e-02 -8.48546922e-01 -1.08233154e+00
3.62010777e-01 2.84492493e-01 8.81030262e-01 8.41385007e-01
-5.91031194e-01 8.59736860e-01 5.42817307e+00 7.51178145e-01
-8.88105631e-01 -1.92762375e-01 4.18442339e-01 -7.24936247e-01
-2.15330586e-01 4.48350787e-01 -9.10371184e-01 6.22297943e-01
1.18960631e+00 -2.98762053e-01 7.80147076e-01 1.13439357e+00
1.53220996e-01 -9.30287242e-01 -1.19754314e+00 2.85184383e-01
-5.47203064e-01 -1.02679110e+00 -4.10608828e-01 1.78927079e-01
7.78047919e-01 -4.21252817e-01 1.40010417e-01 4.84962285e-01
1.04608536e+00 -9.00367022e-01 1.11949575e+00 3.34844738e-01
5.70197999e-01 -1.57718325e+00 4.19415087e-01 8.42272580e-01
-7.91047931e-01 -6.81667268e-01 -6.16668575e-02 -3.29527229e-01
2.23220602e-01 3.99523646e-01 -6.39201581e-01 4.10029173e-01
6.14522576e-01 -2.99818039e-01 1.80816546e-01 7.47992218e-01
-5.30828238e-01 2.95691550e-01 -3.64379883e-01 -8.10707435e-02
6.97847724e-01 -2.86729395e-01 4.49236721e-01 6.11572146e-01
7.34493658e-02 2.41351157e-01 4.91017520e-01 6.73881173e-01
3.88154835e-01 -3.81431848e-01 -2.09972128e-01 -4.81737927e-02
4.90888119e-01 7.12719381e-01 -7.52620161e-01 -2.70299643e-01
2.17211366e-01 5.53997636e-01 5.26019819e-02 2.65632749e-01
-8.67929041e-01 -3.90840232e-01 6.03137136e-01 -2.71892637e-01
2.00772449e-01 -3.56972903e-01 -3.58052284e-01 -6.24503136e-01
3.68890129e-02 -8.51406991e-01 5.81671715e-01 -2.72197545e-01
-4.17522490e-01 2.54758865e-01 3.17774177e-01 -9.84047353e-01
-6.43350720e-01 -9.45594721e-03 -2.98342705e-01 5.30852854e-01
-1.39427733e+00 -5.09281456e-01 3.15898210e-01 5.77366292e-01
3.46877933e-01 1.46602942e-02 6.75874710e-01 -4.79896277e-01
-6.96311235e-01 4.01489079e-01 2.72654802e-01 -4.52108383e-01
1.23581797e-01 -1.30509031e+00 -1.45468876e-01 8.40391457e-01
-6.10422969e-01 2.44621634e-01 9.92537916e-01 -7.38476157e-01
-1.36410034e+00 -9.15424109e-01 3.00491303e-01 -5.24443574e-02
4.12451684e-01 -1.17481507e-01 -4.58634585e-01 9.48499918e-01
9.17242691e-02 -2.08319053e-01 2.69567609e-01 -2.51162618e-01
-1.03365503e-01 -1.61667004e-01 -1.41340089e+00 8.15947592e-01
5.72217107e-01 -8.37769732e-03 -3.74544978e-01 2.36910582e-01
7.09066153e-01 -6.17011189e-01 -9.33526874e-01 5.25020242e-01
3.52154285e-01 -9.70919430e-01 4.73506242e-01 -5.03741205e-01
-4.28204536e-02 -1.57189250e-01 -1.35646343e-01 -1.09726381e+00
1.18561402e-01 -1.14065635e+00 -5.26328743e-01 5.36226928e-01
2.28841111e-01 -5.04122972e-01 9.88828361e-01 1.11234772e+00
-1.04751110e-01 -1.08952200e+00 -1.44600749e+00 -1.26140869e+00
3.09655905e-01 -2.22495452e-01 5.58547080e-01 5.05874991e-01
5.57080388e-01 -4.01146501e-01 -5.60414433e-01 9.46951956e-02
9.17212188e-01 7.82065913e-02 2.87112474e-01 -6.14057422e-01
-6.45532668e-01 -2.52683133e-01 3.22590739e-01 -3.41850400e-01
2.09343359e-02 -1.72219545e-01 2.14540824e-01 -1.39111292e+00
1.42567292e-01 -5.01983941e-01 -4.36724514e-01 9.39160109e-01
1.68863595e-01 -7.07703233e-01 5.14189482e-01 -2.22561583e-01
-1.07645094e+00 7.63238311e-01 1.18023634e+00 9.81540158e-02
-4.60758805e-01 2.70459741e-01 -5.01030624e-01 7.06729710e-01
7.64674067e-01 -3.85585904e-01 -6.18545294e-01 1.25902724e-02
5.54578662e-01 1.04314065e+00 6.94354847e-02 -9.22637939e-01
2.13160478e-02 -1.14888251e+00 -1.22461893e-01 -4.53947186e-01
3.84715170e-01 -7.86877036e-01 1.99684203e-01 1.19930661e+00
-7.63619840e-01 4.85590994e-02 1.27658173e-01 7.56959379e-01
1.51391014e-01 -3.71415973e-01 1.04360163e+00 -3.26924115e-01
-4.28456277e-01 1.38704181e-01 -7.50533104e-01 -2.93656588e-02
1.65308797e+00 -6.00629374e-02 -9.75020677e-02 -5.20081997e-01
-6.81498051e-01 8.31580162e-01 4.12778646e-01 -4.42498885e-02
4.64524031e-01 -8.07035208e-01 -2.66654968e-01 -1.55158669e-01
-2.06870288e-01 2.74062574e-01 4.17426854e-01 6.62903070e-01
-7.02396706e-02 4.48025882e-01 -4.36009794e-01 -4.43459675e-02
-8.88982415e-01 7.45557129e-01 4.81318802e-01 -7.39187002e-01
-3.85242760e-01 6.94286942e-01 -1.78917348e-01 6.62763938e-02
7.36869872e-01 -1.81090832e-01 3.20731342e-01 -1.40226722e-01
4.53245103e-01 7.39853740e-01 -1.36425123e-01 1.79708689e-01
-2.62099534e-01 -1.53603405e-01 -1.68296665e-01 -7.88702548e-01
1.13677001e+00 -1.12575829e-01 2.03343868e-01 2.00871632e-01
3.19050252e-01 -3.58709216e-01 -2.03196716e+00 -5.72463535e-02
6.69234693e-02 -3.28360409e-01 5.60701042e-02 -1.19062173e+00
-6.88926220e-01 7.77005076e-01 5.97242415e-01 1.06686279e-01
9.51157391e-01 -4.41946864e-01 5.38761139e-01 5.86634398e-01
8.94337118e-01 -1.76778901e+00 2.68891364e-01 6.46075249e-01
6.03626668e-01 -8.47141862e-01 7.89425597e-02 2.74258971e-01
-1.21340322e+00 7.73587644e-01 7.38383114e-01 -2.24406719e-02
8.57189074e-02 3.14216435e-01 -5.01984775e-01 1.38811901e-01
-9.69343424e-01 -9.71692204e-02 -5.62158108e-01 1.66484728e-01
-4.49141562e-01 2.84106165e-01 -3.93897831e-01 5.99053800e-01
-1.43973371e-02 6.35703281e-02 8.93308818e-01 1.48247159e+00
-9.30564344e-01 -1.09084630e+00 -3.86780679e-01 2.55191684e-01
-4.73196328e-01 3.93056363e-01 2.76852008e-02 6.59483612e-01
-4.98419218e-02 9.44630444e-01 -1.01314627e-01 6.24889368e-03
1.65055066e-01 6.53794855e-02 5.58818400e-01 -4.16226804e-01
-4.09272909e-01 2.09402472e-01 1.45446181e-01 -9.14064169e-01
9.94889587e-02 -7.01018989e-01 -1.82617402e+00 -2.80404568e-01
2.01626495e-02 4.15974468e-01 5.59480369e-01 1.08933258e+00
1.88287757e-02 4.14196432e-01 9.77699459e-01 -2.09292278e-01
-1.41503990e+00 -4.13352907e-01 -7.88165569e-01 -3.40761006e-01
3.91740948e-01 -6.56822920e-01 -2.79349804e-01 -5.97257197e-01] | [4.386439800262451, 2.7627532482147217] |
eac9eb16-25c5-4fb3-a54d-4140f3429e4a | knowledge-preserving-incremental-social-event | 2101.08747 | null | https://arxiv.org/abs/2101.08747v2 | https://arxiv.org/pdf/2101.08747v2.pdf | Knowledge-Preserving Incremental Social Event Detection via Heterogeneous GNNs | Social events provide valuable insights into group social behaviors and public concerns and therefore have many applications in fields such as product recommendation and crisis management. The complexity and streaming nature of social messages make it appealing to address social event detection in an incremental learning setting, where acquiring, preserving, and extending knowledge are major concerns. Most existing methods, including those based on incremental clustering and community detection, learn limited amounts of knowledge as they ignore the rich semantics and structural information contained in social data. Moreover, they cannot memorize previously acquired knowledge. In this paper, we propose a novel Knowledge-Preserving Incremental Heterogeneous Graph Neural Network (KPGNN) for incremental social event detection. To acquire more knowledge, KPGNN models complex social messages into unified social graphs to facilitate data utilization and explores the expressive power of GNNs for knowledge extraction. To continuously adapt to the incoming data, KPGNN adopts contrastive loss terms that cope with a changing number of event classes. It also leverages the inductive learning ability of GNNs to efficiently detect events and extends its knowledge from previously unseen data. To deal with large social streams, KPGNN adopts a mini-batch subgraph sampling strategy for scalable training, and periodically removes obsolete data to maintain a dynamic embedding space. KPGNN requires no feature engineering and has few hyperparameters to tune. Extensive experiment results demonstrate the superiority of KPGNN over various baselines. | ['Philip S. Yu', 'JianXin Li', 'Yingtong Dou', 'Jia Wu', 'Hao Peng', 'Yuwei Cao'] | 2021-01-21 | null | null | null | null | ['twitter-event-detection'] | ['natural-language-processing'] | [ 1.74530089e-01 2.59294331e-01 -6.12272441e-01 -3.11523587e-01
-2.11470023e-01 -3.09946060e-01 2.17249170e-01 7.91110575e-01
-3.11471522e-01 5.10509968e-01 3.45075727e-01 2.00515930e-02
-3.79855752e-01 -1.17316556e+00 -6.31268978e-01 -4.19569999e-01
-6.10065997e-01 2.37546340e-01 4.90621209e-01 -3.60072255e-01
-2.99133956e-01 2.12690130e-01 -1.20578730e+00 2.32907072e-01
7.31322646e-01 9.32572186e-01 -3.23222391e-02 4.55933154e-01
-8.94341469e-02 1.11514604e+00 -3.14863712e-01 -4.78913546e-01
1.65622085e-01 -2.81593114e-01 -6.66417539e-01 9.97081771e-02
-9.13135558e-02 -3.81772012e-01 -9.28249359e-01 8.43214154e-01
5.05859256e-01 2.87145853e-01 1.18706480e-01 -1.51888847e+00
-6.16970003e-01 1.14356625e+00 -4.44970757e-01 4.17230666e-01
4.30092156e-01 -7.90773705e-02 1.13885438e+00 -5.45954168e-01
7.70154178e-01 9.46545780e-01 1.00081813e+00 1.08028829e-01
-9.48215783e-01 -8.63181472e-01 7.19298005e-01 5.23853481e-01
-1.39741182e+00 -1.74140751e-01 9.56550241e-01 -4.76129353e-02
8.57843339e-01 1.75146163e-01 1.01043189e+00 1.06529593e+00
-3.21593225e-01 1.04101741e+00 3.89985800e-01 -8.28236714e-02
2.70750940e-01 2.13736854e-02 1.88394368e-01 7.26386249e-01
4.08946544e-01 -3.44858229e-01 -7.98963606e-01 -4.55297172e-01
3.88268113e-01 6.94604456e-01 -3.06827962e-01 -2.98409313e-01
-1.05566382e+00 9.43752229e-01 6.48466289e-01 2.55260319e-01
-5.61017513e-01 1.75958797e-01 6.86477125e-01 6.63103819e-01
8.07645023e-01 1.58643335e-01 -5.18913329e-01 -3.37874852e-02
-6.43393815e-01 -1.85846955e-01 1.05158794e+00 9.53953505e-01
1.03182220e+00 -3.98908034e-02 9.48052555e-02 7.28908360e-01
-6.77040070e-02 4.18725759e-01 4.24474448e-01 -6.14231765e-01
5.26384115e-01 1.26130736e+00 -5.29360533e-01 -1.70536387e+00
-5.68861485e-01 -6.49939656e-01 -1.02909875e+00 -9.07244802e-01
-1.19344942e-01 -3.23444337e-01 -4.78371263e-01 1.72376180e+00
8.83269310e-01 7.92719603e-01 -1.17509894e-01 3.45345527e-01
8.79312158e-01 6.41499102e-01 -1.72791913e-01 -4.05305177e-01
8.65305483e-01 -6.05740130e-01 -5.42061865e-01 -2.10307702e-01
7.02666938e-01 -1.39082566e-01 7.56202221e-01 1.43462911e-01
-7.45177329e-01 -5.24812303e-02 -8.39193702e-01 1.68313161e-01
-4.16182160e-01 -5.78444123e-01 9.69371617e-01 3.24159712e-01
-9.49524701e-01 7.29901314e-01 -8.13328326e-01 -4.73002553e-01
5.85700691e-01 3.78574848e-01 -6.63788691e-02 -2.27225959e-01
-1.59116614e+00 -2.99352761e-02 8.25896204e-01 -6.08712919e-02
-5.93815982e-01 -7.69050658e-01 -9.69732404e-01 1.79753661e-01
8.90216053e-01 -5.78081310e-01 1.04899085e+00 -9.25932109e-01
-1.25831723e+00 2.93549597e-01 1.12327479e-01 -6.23338282e-01
1.95689753e-01 2.97181625e-02 -7.69532442e-01 5.95950961e-01
-5.69642112e-02 1.08002305e-01 9.55213726e-01 -7.81584978e-01
-6.60268247e-01 -1.33615717e-01 2.74786234e-01 1.47175863e-01
-1.13701594e+00 -2.50771046e-01 -4.68404233e-01 -7.63479650e-01
1.46467716e-01 -8.15205812e-01 -2.64430642e-01 -2.88036615e-01
-4.23143744e-01 -2.95051187e-01 9.50961828e-01 -4.30983841e-01
1.60647583e+00 -2.12275028e+00 -3.79523635e-02 6.70689702e-01
8.05189729e-01 1.82006210e-01 -2.86523700e-01 8.43190134e-01
2.66554773e-01 -1.61868989e-01 -1.13189116e-01 -1.63495719e-01
-2.45393962e-02 2.99341500e-01 -2.34394476e-01 3.41968238e-01
-1.31658956e-01 1.14539957e+00 -1.40543640e+00 -5.14353514e-01
-2.03728184e-01 3.70077789e-01 -6.65782452e-01 -9.93426740e-02
-2.45248571e-01 4.38673422e-02 -7.21544325e-01 6.58106625e-01
4.28077042e-01 -9.52503800e-01 4.45591718e-01 -8.76280665e-02
6.15390003e-01 2.77406145e-02 -1.32430685e+00 1.37410617e+00
-3.46106768e-01 3.68325651e-01 -3.85263972e-02 -1.15940917e+00
5.64725041e-01 1.46460578e-01 7.60209441e-01 -5.38709342e-01
8.04011226e-02 -6.53662160e-02 -4.19327646e-01 -4.17286754e-01
5.71656585e-01 1.16600104e-01 -1.73266456e-01 7.71256983e-01
1.37981474e-01 5.16510129e-01 3.08594316e-01 9.72506344e-01
1.60936379e+00 -7.47601509e-01 3.94950807e-01 2.27504224e-01
3.43825966e-01 -3.57685953e-01 8.16585481e-01 8.29702437e-01
-8.17990154e-02 1.44971147e-01 4.86321837e-01 -4.31823671e-01
-4.79522109e-01 -1.06710410e+00 4.24172699e-01 1.48185670e+00
3.13890129e-01 -7.73784578e-01 -3.81069839e-01 -1.02143478e+00
2.66360313e-01 3.68558288e-01 -4.34899658e-01 -6.16571844e-01
-6.68213964e-01 -9.58453000e-01 2.89145172e-01 3.16994786e-01
4.98039514e-01 -1.01981282e+00 2.03136783e-02 4.65189278e-01
-3.64610970e-01 -1.18022108e+00 -6.25411928e-01 -3.49737167e-01
-7.35803366e-01 -1.46730018e+00 -9.16615948e-02 -8.58380914e-01
6.73464835e-01 6.60931826e-01 1.06947601e+00 1.80617630e-01
-2.20638484e-01 1.02676415e+00 -7.60881364e-01 -4.46590483e-01
-3.19108725e-01 5.38075387e-01 7.27381483e-02 5.05596936e-01
5.07037520e-01 -1.10933018e+00 -7.75785208e-01 1.81127891e-01
-1.21970809e+00 -2.80034721e-01 3.67819756e-01 5.14905989e-01
4.34638113e-01 4.19667870e-01 1.23104811e+00 -1.32225096e+00
7.22247958e-01 -1.05168641e+00 -2.34142497e-01 1.60591319e-01
-7.60622799e-01 -3.39983016e-01 7.58397877e-01 -7.57449865e-01
-8.48771274e-01 -3.12767506e-01 3.16251218e-01 -2.88311750e-01
4.38661367e-01 9.54809546e-01 1.54166669e-01 -6.59067407e-02
5.29176891e-01 2.97089189e-01 -1.71739101e-01 -2.52643615e-01
4.76322144e-01 5.55447578e-01 2.00456336e-01 -1.11240081e-01
9.21652734e-01 8.03947210e-01 -2.77116686e-01 -9.59480464e-01
-9.58216429e-01 -8.48872542e-01 -1.39410928e-01 -2.61388630e-01
1.68792740e-01 -9.63927805e-01 -9.93875504e-01 4.88720387e-01
-6.65151298e-01 -2.92411953e-01 -7.61598587e-01 5.25502086e-01
-1.00333713e-01 7.16417134e-01 -9.81885433e-01 -5.43429255e-01
-5.49209297e-01 -1.72522515e-01 6.13109410e-01 6.63866550e-02
-4.22079638e-02 -1.21714044e+00 -7.48606324e-02 1.42804414e-01
3.82398278e-01 4.67542350e-01 6.47457480e-01 -1.02240455e+00
-6.97533965e-01 -6.13905013e-01 -2.58964986e-01 1.71393469e-01
3.67993683e-01 -5.33597708e-01 -6.42637670e-01 -7.31329083e-01
-1.58496082e-01 -4.86294568e-01 1.06648123e+00 2.96300687e-02
1.17334867e+00 -7.52050698e-01 -4.48722869e-01 4.65985775e-01
1.00582778e+00 -1.49477988e-01 1.43838048e-01 5.67284301e-02
1.01671338e+00 5.03860176e-01 3.09666097e-01 8.55199158e-01
9.06426370e-01 2.21350472e-02 4.05480653e-01 4.19901162e-02
6.97036684e-02 -5.90484083e-01 4.56329882e-01 1.42145753e+00
1.93163753e-01 -3.03743392e-01 -6.38157725e-01 6.05436921e-01
-2.07367277e+00 -1.11199617e+00 3.78093630e-01 1.98284805e+00
9.62244689e-01 2.40246087e-01 3.02535355e-01 2.06995383e-01
9.12028015e-01 5.66965938e-01 -8.51596415e-01 1.98981673e-01
-1.39199868e-01 -1.62416562e-01 5.50746560e-01 2.35988542e-01
-1.05520785e+00 7.58405864e-01 4.91837072e+00 9.43411767e-01
-9.18839157e-01 1.89884245e-01 3.11337143e-01 -4.16477352e-01
-3.23564321e-01 9.60515533e-03 -7.11853564e-01 4.59738880e-01
1.04002261e+00 -5.51714838e-01 4.88469064e-01 8.90239835e-01
-2.51924507e-02 2.98093408e-01 -9.17609990e-01 8.52710247e-01
1.46678954e-01 -1.40143800e+00 5.81363775e-02 -2.02473193e-01
8.80400836e-01 4.80495811e-01 -1.62103608e-01 5.77781737e-01
5.37144125e-01 -3.52282465e-01 1.63041458e-01 5.19049466e-01
3.41703981e-01 -8.97880495e-01 5.48214376e-01 4.35905784e-01
-1.61005437e+00 -4.32159543e-01 -2.75333911e-01 -3.02161183e-02
2.79208690e-01 9.37084198e-01 -1.07769489e+00 6.67563379e-01
6.65934563e-01 1.37919915e+00 -6.32852554e-01 8.10074270e-01
1.09962458e-02 1.05529845e+00 -6.63383901e-01 -2.04842806e-01
2.71254063e-01 3.71300504e-02 7.67369390e-01 1.17520940e+00
1.25077397e-01 1.36317536e-01 6.88744307e-01 2.90732861e-01
-5.86799920e-01 4.19577748e-01 -5.71802199e-01 -3.26701283e-01
6.25702977e-01 1.06090963e+00 -8.22987735e-01 -4.58804637e-01
-4.88478959e-01 8.15479875e-01 5.83288729e-01 5.62156320e-01
-5.86344182e-01 -6.30846083e-01 2.85320193e-01 3.58928502e-01
5.41297793e-01 -7.39365118e-03 5.26501596e-01 -1.43847895e+00
2.41762534e-01 -6.43913209e-01 1.10836232e+00 -1.51755661e-01
-1.59797490e+00 4.39678371e-01 1.38792291e-01 -1.12733066e+00
-1.34237006e-01 7.71516189e-02 -5.54109573e-01 -2.27092445e-01
-1.59720242e+00 -1.32166553e+00 -3.18239450e-01 8.77635360e-01
2.09533066e-01 -2.02292994e-01 2.66310900e-01 4.96993035e-01
-5.47873974e-01 9.01109517e-01 1.86341017e-01 3.43425333e-01
4.82911080e-01 -1.02224600e+00 2.12133035e-01 6.92584872e-01
1.62994251e-01 5.14908493e-01 2.71271229e-01 -9.78572845e-01
-1.49321544e+00 -1.69855821e+00 8.44194829e-01 1.48944920e-02
1.01960969e+00 -3.56836796e-01 -1.13268220e+00 8.38333607e-01
-2.16199830e-01 3.06432128e-01 9.80402708e-01 3.71709555e-01
-5.29785514e-01 -5.99690676e-01 -9.18997884e-01 4.80093241e-01
1.47226369e+00 -7.72413313e-01 -2.09057212e-01 7.41793096e-01
1.30462170e+00 -1.10821471e-01 -1.02817917e+00 3.94837826e-01
8.61811712e-02 -5.71719944e-01 9.39759195e-01 -7.20908225e-01
-1.12869687e-01 -1.12953968e-02 1.50936618e-01 -1.14471042e+00
-2.64665663e-01 -1.08803129e+00 -8.34901929e-01 1.19609714e+00
3.80884141e-01 -1.05221260e+00 9.92907047e-01 4.38445836e-01
1.51169956e-01 -6.61042035e-01 -7.40100205e-01 -8.34925413e-01
-6.21837258e-01 -4.70638692e-01 7.76702523e-01 1.40018809e+00
2.18387797e-01 4.40019995e-01 -5.17768919e-01 1.75423965e-01
5.72250605e-01 3.34484488e-01 7.04294145e-01 -1.41907811e+00
-4.36503500e-01 8.73915479e-03 -6.09695733e-01 -7.80681014e-01
1.38980374e-01 -1.32681847e+00 -5.85411787e-01 -1.36295128e+00
2.51733541e-01 -3.31999600e-01 -8.23318183e-01 5.26379883e-01
-1.68735802e-01 1.56520277e-01 7.38346055e-02 1.58708766e-01
-1.55164874e+00 7.01480925e-01 8.15413415e-01 -2.12424159e-01
-6.86629713e-01 1.41540363e-01 -8.35306168e-01 8.96718979e-01
8.11440349e-01 -6.39124632e-01 -9.27242875e-01 -7.11533949e-02
9.45436060e-01 -1.46765605e-01 3.64862353e-01 -7.60857701e-01
6.21425092e-01 -7.57012740e-02 4.61299019e-03 -5.34330249e-01
5.27378060e-02 -7.18402505e-01 5.92989400e-02 2.99338639e-01
-2.33264551e-01 -6.00564294e-02 -1.76515460e-01 1.47596359e+00
-2.82939881e-01 2.69448400e-01 3.12039971e-01 -5.87548465e-02
-6.85547411e-01 8.60347152e-01 -1.12229861e-01 5.34806490e-01
1.02708554e+00 -2.00678527e-01 -1.57998621e-01 -8.06392670e-01
-7.97206879e-01 7.32284009e-01 2.67421365e-01 4.76883322e-01
8.52299809e-01 -1.37435901e+00 -5.74451685e-01 3.08046266e-02
3.96423191e-01 1.32775620e-01 6.42510772e-01 8.08551550e-01
-1.12229124e-01 -2.20612079e-01 5.85218072e-01 -3.01697999e-01
-1.07959867e+00 8.16304266e-01 -1.51847601e-01 -4.06749904e-01
-9.03122663e-01 7.45150864e-01 -9.11661461e-02 -4.46690559e-01
3.78261864e-01 -1.86139181e-01 -2.24228188e-01 4.40869153e-01
6.89524233e-01 6.24208689e-01 7.32880086e-04 -2.61526823e-01
-1.49705112e-01 2.42260508e-02 -4.26860183e-01 6.01096749e-01
1.67949116e+00 -3.96342248e-01 -1.30822718e-01 4.91345465e-01
1.53045940e+00 -1.03654817e-01 -1.00111079e+00 -1.14788377e+00
2.16547921e-02 -3.50684106e-01 8.98025930e-02 -2.97353238e-01
-1.36608601e+00 2.10633919e-01 1.35390028e-01 5.47485113e-01
1.38195109e+00 2.17961714e-01 1.52912331e+00 9.10578489e-01
2.53529608e-01 -1.38823688e+00 3.91856045e-01 4.84395623e-01
4.83012646e-01 -1.13377345e+00 2.04863995e-01 -5.65412998e-01
-4.51917857e-01 8.45092356e-01 3.81842554e-01 -1.55318715e-03
1.18291867e+00 -2.10227594e-01 -4.69515979e-01 -4.77833539e-01
-7.95322597e-01 -2.71654189e-01 1.17650539e-01 3.67864460e-01
-4.14432675e-01 -6.66662380e-02 -4.34409641e-02 7.19453752e-01
1.20773897e-01 7.19767511e-02 4.96602207e-01 1.14378107e+00
-4.45618838e-01 -7.61524320e-01 3.43078285e-01 9.25489724e-01
-2.55788147e-01 -1.15826949e-01 -2.76107788e-01 5.54674685e-01
6.21670857e-03 8.58725011e-01 -2.47704476e-01 -6.08029068e-01
3.09814721e-01 -2.47616708e-01 -1.04636304e-01 -5.98967612e-01
-5.38866282e-01 -3.23993117e-01 9.83865187e-02 -6.77889705e-01
-3.71248841e-01 -7.03972578e-01 -1.28016698e+00 -5.29103816e-01
-3.74296159e-01 3.29964697e-01 1.36610135e-01 7.12045312e-01
8.81036937e-01 4.15881544e-01 1.16289210e+00 -4.24512446e-01
-4.68849570e-01 -5.54419100e-01 -6.92030668e-01 6.04762435e-01
3.55406731e-01 -4.72166032e-01 -4.92401928e-01 -2.12371856e-01] | [7.246924877166748, 6.043272495269775] |
36ac86f7-f1b0-4fee-acbb-3077f965ef60 | snomed2vec-random-walk-and-poincare | 1907.08650 | null | https://arxiv.org/abs/1907.08650v1 | https://arxiv.org/pdf/1907.08650v1.pdf | Snomed2Vec: Random Walk and Poincaré Embeddings of a Clinical Knowledge Base for Healthcare Analytics | Representation learning methods that transform encoded data (e.g., diagnosis and drug codes) into continuous vector spaces (i.e., vector embeddings) are critical for the application of deep learning in healthcare. Initial work in this area explored the use of variants of the word2vec algorithm to learn embeddings for medical concepts from electronic health records or medical claims datasets. We propose learning embeddings for medical concepts by using graph-based representation learning methods on SNOMED-CT, a widely popular knowledge graph in the healthcare domain with numerous operational and research applications. Current work presents an empirical analysis of various embedding methods, including the evaluation of their performance on multiple tasks of biomedical relevance (node classification, link prediction, and patient state prediction). Our results show that concept embeddings derived from the SNOMED-CT knowledge graph significantly outperform state-of-the-art embeddings, showing 5-6x improvement in ``concept similarity" and 6-20\% improvement in patient diagnosis. | ['Khushbu Agarwal', 'Sutanay Choudhury', 'Tome Eftimov', 'Suzanne Tamang', 'Robert Rallo', 'Raghavendra Addanki'] | 2019-07-19 | null | null | null | null | ['clinical-knowledge'] | ['miscellaneous'] | [ 8.73031020e-02 7.47793913e-01 -5.93818426e-01 -2.48728484e-01
-5.02218425e-01 -1.97657377e-01 3.01752120e-01 1.33334422e+00
-4.87901211e-01 2.87120283e-01 9.26881135e-01 -6.74833953e-01
-5.82342803e-01 -9.15987432e-01 -8.62242207e-02 -4.11155671e-01
-6.55848682e-01 6.92941129e-01 -2.77554750e-01 -2.61494994e-01
-1.71073288e-01 2.13495046e-01 -9.02007878e-01 2.01437235e-01
4.40524101e-01 9.13290441e-01 -2.96607941e-01 5.65546513e-01
-3.34477156e-01 8.43458295e-01 -3.37423801e-01 -6.11887991e-01
-2.25646511e-01 -9.27197039e-02 -9.72013056e-01 -5.55395663e-01
1.76320095e-02 7.47295767e-02 -7.27588654e-01 1.04766965e+00
7.38108933e-01 1.48667634e-01 9.63873982e-01 -9.92055357e-01
-1.47825849e+00 5.84260523e-01 -7.58211091e-02 3.80264759e-01
6.63768172e-01 -4.30693567e-01 1.46018755e+00 -6.40471339e-01
9.66546834e-01 1.01983058e+00 1.04103160e+00 6.33271813e-01
-1.14202511e+00 -2.41331846e-01 -3.67853850e-01 4.04197425e-01
-1.51682746e+00 7.54562318e-02 3.58957171e-01 -7.11837292e-01
1.46409547e+00 2.18520656e-01 7.42105961e-01 1.08972752e+00
7.44817734e-01 3.57176155e-01 3.87441337e-01 -2.93279529e-01
2.71848232e-01 2.65068024e-01 3.39708149e-01 9.28884923e-01
8.43165517e-01 6.64180443e-02 -3.72367501e-02 -7.74798214e-01
5.42921364e-01 4.29344147e-01 -4.00131494e-01 -5.42578220e-01
-1.19572175e+00 1.47155261e+00 8.42580318e-01 5.57686985e-01
-5.43379545e-01 4.00083601e-01 8.65036547e-01 2.32095495e-01
5.08285999e-01 8.08064103e-01 -6.04389012e-01 -9.69065912e-03
-4.53963608e-01 3.24428789e-02 7.29602635e-01 7.27187097e-01
1.11195482e-01 -1.31850228e-01 -3.03267807e-01 7.49543607e-01
4.21924770e-01 2.19141021e-02 8.80928397e-01 -2.65980333e-01
2.69356787e-01 7.82211244e-01 -2.92448401e-01 -1.35502291e+00
-7.05254674e-01 -3.46277624e-01 -7.46124506e-01 -4.33031887e-01
-3.09919089e-01 -2.40481608e-02 -9.07568932e-01 1.32724881e+00
3.55394661e-01 3.36007357e-01 4.44405943e-01 5.27318954e-01
1.33769882e+00 3.77420813e-01 3.70012790e-01 -5.41367242e-03
1.75319648e+00 -7.01167285e-01 -9.54443276e-01 2.67209679e-01
1.27654314e+00 -2.34671399e-01 3.78266692e-01 -1.11152261e-01
-6.41522110e-01 -3.71337742e-01 -8.73007059e-01 -2.08142117e-01
-8.57536793e-01 -5.93898773e-01 1.03295541e+00 5.56648791e-01
-9.26398396e-01 6.55031800e-01 -8.62607479e-01 -5.71276128e-01
6.84028924e-01 3.77610296e-01 -8.02290559e-01 -6.78407550e-01
-1.47528458e+00 1.09143412e+00 4.45408702e-01 -6.56036317e-01
-4.60362673e-01 -1.15241003e+00 -1.56143701e+00 2.65002787e-01
-1.97586119e-01 -9.87042904e-01 7.07829058e-01 -1.20063178e-01
-6.05353117e-01 7.74883628e-01 4.48818535e-01 -6.35631859e-01
-1.94405526e-01 -8.51669684e-02 -1.00689018e+00 3.57803643e-01
-5.48106991e-03 4.77865845e-01 2.01877862e-01 -6.58576548e-01
-2.85338223e-01 -3.36805344e-01 -9.22222063e-02 1.51677225e-02
-8.80562007e-01 -4.10764307e-01 -8.59296769e-02 -7.13264287e-01
-1.39682412e-01 -7.25666821e-01 -6.97263539e-01 2.42073908e-01
-1.69789135e-01 -4.62330699e-01 3.83089602e-01 -8.96162331e-01
1.40755880e+00 -2.18138909e+00 1.47252962e-01 1.96310163e-01
7.48280942e-01 1.83372453e-01 -2.12054059e-01 6.64330184e-01
-5.50780773e-01 2.91174769e-01 -1.94043621e-01 1.48260713e-01
-1.42858282e-01 4.84620810e-01 2.93353349e-01 5.57450652e-01
3.05306874e-02 1.30408335e+00 -1.42057204e+00 -4.40860033e-01
5.50469980e-02 1.06076372e+00 -7.53860235e-01 2.79808436e-02
1.92638382e-01 -1.26705453e-01 -4.74577606e-01 6.20673418e-01
2.47069374e-02 -5.11161327e-01 5.90960324e-01 -3.90733749e-01
5.90945959e-01 1.94981575e-01 -4.73243892e-01 1.94811380e+00
-3.73397827e-01 5.92338920e-01 -6.10307336e-01 -1.07164454e+00
5.87062180e-01 8.21862936e-01 1.17969179e+00 -4.06409055e-01
2.25319818e-01 -1.21821679e-01 1.36907548e-01 -7.83229172e-01
3.41896474e-01 -1.71441346e-01 -1.78345382e-01 1.91542432e-01
3.88510138e-01 2.03664362e-01 -3.45066130e-01 3.19451064e-01
1.61751330e+00 -6.37857497e-01 8.29156876e-01 -3.24069351e-01
1.56880391e-03 5.74966557e-02 2.56010711e-01 1.82159960e-01
-1.87753573e-01 2.36955538e-01 5.14999628e-01 -5.96527576e-01
-8.26278150e-01 -1.17679596e+00 -4.86744136e-01 6.96646214e-01
-3.04931015e-01 -8.03456485e-01 -2.94646263e-01 -8.79708529e-01
7.50830650e-01 6.69243693e-01 -1.21165085e+00 -6.97604597e-01
-8.20259079e-02 -7.95488477e-01 4.60950345e-01 1.05611455e+00
-6.06569529e-01 -6.99619055e-01 -4.86093819e-01 3.67657334e-01
3.09689552e-01 -8.31732333e-01 -4.52539980e-01 2.99307436e-01
-1.12282562e+00 -1.45491242e+00 -5.84055185e-01 -1.09093988e+00
7.97314584e-01 -2.03005567e-01 1.20444405e+00 2.58604418e-02
-9.58445132e-01 8.68613780e-01 -6.98120654e-01 -2.68908203e-01
-2.83568412e-01 -3.02913904e-01 2.34612137e-01 -4.16348249e-01
7.41598010e-01 -2.74284899e-01 -8.61456633e-01 -3.58134747e-01
-1.17054510e+00 -5.83255053e-01 4.75195289e-01 1.18076611e+00
6.94691300e-01 -1.38881370e-01 6.12529516e-01 -1.18192470e+00
1.19300377e+00 -1.02283525e+00 1.75506935e-01 1.64411947e-01
-1.29857898e+00 2.31198892e-01 1.20266117e-01 -2.60211349e-01
-1.04418598e-01 -4.20544118e-01 -2.93021202e-01 -5.36209941e-01
1.92669243e-01 1.18458748e+00 5.06099284e-01 6.66914731e-02
8.77695143e-01 -1.32196113e-01 2.75541190e-03 -3.64452511e-01
9.19707298e-01 8.42898905e-01 2.40415096e-01 9.65946615e-02
2.64285773e-01 1.87889427e-01 -3.05276625e-02 -5.84611714e-01
-3.54897976e-01 -7.94550300e-01 -3.90984476e-01 5.70647359e-01
1.46573234e+00 -8.49510074e-01 -3.71743202e-01 -7.40255773e-01
-9.01650667e-01 3.32261056e-01 -5.77184856e-01 8.16629529e-01
-1.86526909e-01 2.85422355e-01 -8.36243749e-01 -1.98417798e-01
-7.28960037e-01 -7.81915665e-01 7.97651589e-01 -2.88609087e-01
-6.26216531e-01 -1.69909358e+00 5.65911591e-01 1.27260193e-01
3.68484080e-01 6.56885862e-01 1.74308407e+00 -1.21049178e+00
2.97355354e-01 -6.81706488e-01 -2.16557950e-01 1.51789129e-01
5.43821514e-01 -5.71729660e-01 -6.77523375e-01 -5.67842305e-01
-4.93822366e-01 -3.27692218e-02 8.38251710e-01 4.19835895e-01
1.17388105e+00 -3.27112198e-01 -8.22401345e-01 5.82318425e-01
1.70844781e+00 2.48828024e-01 5.67556918e-01 -8.43283981e-02
8.92714918e-01 3.57755363e-01 1.09620184e-01 4.57093209e-01
4.88261998e-01 4.06641424e-01 4.60995972e-01 -8.23704898e-02
-8.61617178e-02 -1.78142548e-01 -9.59297642e-02 1.15749443e+00
3.22663449e-02 -1.01666003e-01 -1.11811650e+00 1.02805674e+00
-1.55782878e+00 -5.91380060e-01 1.36690378e-01 1.81101739e+00
8.49737823e-01 -2.08162025e-01 -2.87260771e-01 2.21505716e-01
3.50558549e-01 4.60903235e-02 -3.50301325e-01 -7.46704400e-01
6.73726678e-01 7.80310452e-01 5.52790821e-01 2.41706058e-01
-8.34587395e-01 5.23858428e-01 6.77640486e+00 1.92784801e-01
-7.45090246e-01 5.09784341e-01 2.86453336e-01 3.12599242e-02
-6.33308351e-01 -6.01758957e-01 -1.73389286e-01 2.26542667e-01
1.29748285e+00 -1.73184544e-01 -4.07849997e-02 9.20141399e-01
-3.57236803e-01 5.57672024e-01 -1.41820621e+00 1.28220582e+00
4.11157757e-01 -1.75645018e+00 1.62148759e-01 2.41646975e-01
7.30671763e-01 1.06482208e-01 1.40880942e-01 4.18797851e-01
3.53526592e-01 -1.45895374e+00 -3.26802909e-01 4.53914076e-01
1.47010827e+00 -6.92892849e-01 1.19970596e+00 -5.90308666e-01
-1.15208256e+00 -1.75629005e-01 -4.85652000e-01 4.05759633e-01
1.48814052e-01 6.64835989e-01 -1.28877294e+00 8.94647360e-01
5.88938534e-01 1.19429195e+00 -4.30059373e-01 8.01252425e-01
2.24369746e-02 4.82208520e-01 4.09197122e-01 9.72475335e-02
1.72158241e-01 1.92464679e-01 9.38078091e-02 1.43457222e+00
3.20734113e-01 9.60405022e-02 8.65034759e-02 4.17596102e-01
-3.19673747e-01 2.26078302e-01 -1.10257518e+00 -5.60246408e-01
2.45366380e-01 1.06331038e+00 -4.18729961e-01 -4.16911542e-01
-6.14433110e-01 9.03772175e-01 1.59156129e-01 1.04224153e-01
-7.91895628e-01 -7.65093446e-01 1.30185056e+00 1.04900338e-01
3.99474174e-01 1.19646952e-01 -5.67893013e-02 -8.74765515e-01
-5.99321842e-01 -5.83492398e-01 7.38854885e-01 -4.07940477e-01
-1.68207860e+00 7.62493312e-01 -2.11607680e-01 -1.09391463e+00
-2.97543526e-01 -9.30670023e-01 -1.79750130e-01 5.52563429e-01
-1.50079286e+00 -8.45439136e-01 5.67819439e-02 6.13924146e-01
-1.76356919e-02 -5.26051223e-01 1.70341897e+00 6.24577940e-01
-1.05259597e-01 8.22913766e-01 4.01551843e-01 4.88924593e-01
4.75582302e-01 -1.28132963e+00 4.81462121e-01 -1.47386000e-01
3.29972744e-01 8.09555411e-01 3.10248405e-01 -5.63068151e-01
-1.46318698e+00 -1.43849242e+00 1.21249747e+00 -8.16782355e-01
7.18767524e-01 -3.52616720e-02 -7.05808699e-01 7.79660463e-01
1.54888347e-01 5.32128990e-01 1.66632593e+00 2.26400465e-01
-6.23381078e-01 1.72481939e-01 -1.55422044e+00 3.51380914e-01
1.01421976e+00 -8.15813541e-01 -8.83141398e-01 4.95886177e-01
1.28538609e+00 -6.82078302e-02 -1.92871499e+00 4.47649390e-01
4.92591053e-01 -1.13139622e-01 1.35471702e+00 -1.50951982e+00
5.55471838e-01 1.58928663e-01 -4.44056511e-01 -1.53498828e+00
-8.64318013e-01 3.92456353e-02 -5.97107768e-01 4.23039407e-01
3.93308133e-01 -7.04822540e-01 3.84931624e-01 3.61284763e-01
-2.16741279e-01 -1.41429782e+00 -9.43525314e-01 -5.43204546e-01
1.23279482e-01 -2.02640131e-01 3.98460746e-01 1.64789474e+00
5.79207540e-01 3.26419801e-01 6.64908141e-02 1.36998489e-01
1.55503258e-01 -2.35567078e-01 -1.10121772e-01 -1.41711271e+00
-2.88980037e-01 -2.22931564e-01 -1.45787883e+00 -2.86086679e-01
-1.02063604e-01 -1.61748123e+00 -5.41080117e-01 -2.32639122e+00
1.41241670e-01 -2.23195031e-01 -1.03112125e+00 5.55109203e-01
-1.74502015e-01 8.31986070e-02 -1.76775977e-01 -2.41025940e-01
-3.27977061e-01 4.05909806e-01 9.23373222e-01 -5.67237020e-01
1.96297765e-02 -5.74916959e-01 -1.02695620e+00 1.80484444e-01
5.06955385e-01 -6.84585929e-01 -6.54973209e-01 -5.90453267e-01
2.86098182e-01 1.90148488e-01 -3.58994938e-02 -8.44855607e-01
-9.98232886e-02 2.61135072e-01 3.16236466e-01 3.13845798e-02
2.52562344e-01 -8.33033741e-01 -1.11028338e-02 9.15601790e-01
-5.64449668e-01 4.45844769e-01 2.05608040e-01 1.24373317e+00
-2.11177617e-01 -3.35370377e-02 2.84419328e-01 -3.24235600e-03
-6.59163296e-01 5.55987954e-01 -1.27080902e-01 1.90848187e-01
1.08648145e+00 7.74623826e-02 -1.57597795e-01 -6.34659380e-02
-1.00551510e+00 5.71239255e-02 -2.35514287e-02 6.96957886e-01
1.10839605e+00 -1.65661478e+00 -5.73175967e-01 -9.53818392e-03
7.52957165e-01 -4.00429636e-01 -4.57087494e-02 8.14345837e-01
-6.27178133e-01 6.48140550e-01 -3.87663618e-02 -2.30444774e-01
-1.18597090e+00 1.11704338e+00 8.09341595e-02 -4.49285209e-01
-8.76052618e-01 1.01919901e+00 -1.47064447e-01 -4.53263491e-01
2.19975621e-01 -7.65021563e-01 -6.32097721e-01 2.80724436e-01
5.68678975e-01 1.77547112e-01 1.42675430e-01 -2.94789821e-01
-7.30279744e-01 3.48762780e-01 -2.65464243e-02 3.02807271e-01
1.72248960e+00 5.58947623e-01 -1.32100046e-01 1.68827772e-01
1.83703458e+00 -2.48727337e-01 -1.63119733e-01 -1.86427504e-01
9.19276625e-02 -2.10917592e-01 3.03427994e-01 -7.26076722e-01
-1.04070187e+00 7.41995573e-01 1.05335987e+00 1.02973141e-01
6.19509161e-01 3.06003094e-01 1.01079786e+00 3.38323146e-01
2.76515782e-01 -7.28927910e-01 9.25496072e-02 2.41101757e-01
5.90462327e-01 -1.10695219e+00 1.88275054e-01 -1.56850025e-01
-6.56197727e-01 1.01924467e+00 -1.31673291e-02 -7.09981844e-02
1.24850667e+00 -1.88458160e-01 3.75120528e-03 -8.50376666e-01
-6.11847520e-01 -2.65354272e-02 5.31451225e-01 9.13553774e-01
8.08448970e-01 5.27272999e-01 -4.69453484e-01 8.19900274e-01
1.37398422e-01 -1.06133901e-01 1.93610996e-01 9.97042716e-01
-1.24569036e-01 -1.02269733e+00 7.93442056e-02 1.07862639e+00
-5.77511489e-01 -5.35587251e-01 -5.12195230e-02 4.49834764e-01
2.93088257e-01 7.10613787e-01 6.72708675e-02 -7.15458751e-01
3.71916950e-01 2.65811622e-01 3.21046472e-01 -1.14748406e+00
-5.60325086e-01 -6.23549283e-01 1.13466553e-01 -5.35618126e-01
-2.10518584e-01 -3.57164025e-01 -1.42034030e+00 -6.35688081e-02
-2.59347290e-01 3.18935484e-01 5.63953817e-01 4.93215114e-01
7.41616666e-01 1.09362209e+00 7.75226653e-02 -7.71864653e-02
-4.27680403e-01 -8.38246584e-01 -6.63747489e-01 7.24900424e-01
3.34305197e-01 -6.46954715e-01 5.32300696e-02 -1.22846216e-01] | [7.946043968200684, 7.092226982116699] |
9e9a89c7-84cd-4613-8de5-485d24df1de7 | monoedge-monocular-3d-object-detection-using | 2301.01802 | null | https://arxiv.org/abs/2301.01802v1 | https://arxiv.org/pdf/2301.01802v1.pdf | MonoEdge: Monocular 3D Object Detection Using Local Perspectives | We propose a novel approach for monocular 3D object detection by leveraging local perspective effects of each object. While the global perspective effect shown as size and position variations has been exploited for monocular 3D detection extensively, the local perspectives has long been overlooked. We design a local perspective module to regress a newly defined variable named keyedge-ratios as the parameterization of the local shape distortion to account for the local perspective, and derive the object depth and yaw angle from it. Theoretically, this module does not rely on the pixel-wise size or position in the image of the objects, therefore independent of the camera intrinsic parameters. By plugging this module in existing monocular 3D object detection frameworks, we incorporate the local perspective distortion with global perspective effect for monocular 3D reasoning, and we demonstrate the effectiveness and superior performance over strong baseline methods in multiple datasets. | ['Huei Peng', 'Panqu Wang', 'Lingting Ge', 'Minghan Zhu'] | 2023-01-04 | null | null | null | null | ['monocular-3d-object-detection'] | ['computer-vision'] | [-1.9017529e-01 -2.4681997e-01 -1.9507676e-01 -2.8234300e-01
-2.3403086e-01 -8.2718563e-01 8.3303660e-01 -2.6675674e-01
-2.8685254e-01 1.1312797e-01 1.2641288e-01 -1.7757913e-01
3.2860044e-01 -4.8564255e-01 -8.2864624e-01 -7.9237783e-01
3.2827774e-01 2.6631802e-01 6.1307502e-01 2.8182185e-01
4.8209450e-01 9.8245072e-01 -1.2955277e+00 -1.2102623e-01
4.5412704e-01 1.0164212e+00 2.4570721e-01 6.9868326e-01
3.8130152e-01 4.9112007e-01 -5.1335561e-01 -7.1626663e-02
7.2799754e-01 -1.4771318e-01 1.4505533e-02 3.5921335e-01
7.9587942e-01 -8.0192375e-01 -4.5787239e-01 9.5005327e-01
6.7037767e-01 -2.5184622e-01 8.0586910e-01 -1.1565257e+00
-3.5503820e-01 2.3858912e-02 -9.4226956e-01 2.0265581e-01
4.9571455e-01 3.4171247e-01 8.2048172e-01 -1.2213831e+00
8.7817472e-01 1.3534416e+00 4.5179090e-01 7.1087979e-02
-1.0406057e+00 -6.2525839e-01 3.4097341e-01 8.1380494e-02
-1.2356526e+00 -3.0604488e-01 9.5274585e-01 -5.3724074e-01
8.8827592e-01 2.6197355e-02 6.4558381e-01 9.2470187e-01
2.3229052e-01 9.8766744e-01 1.1293659e+00 -5.0457096e-01
1.5369293e-02 3.6716479e-01 -7.3390372e-02 5.0193971e-01
5.7752037e-01 4.3843442e-01 -3.3942622e-01 6.0059786e-02
9.4703275e-01 -3.4101043e-02 -2.8048488e-01 -1.1295063e+00
-1.1466813e+00 5.0469404e-01 5.1126409e-01 -1.3199019e-01
5.0959822e-02 1.4273560e-02 -2.6665749e-02 -1.6157955e-01
4.5218086e-01 4.0470999e-01 -3.6333197e-01 6.7506224e-02
-5.5186421e-01 4.2906553e-01 2.8252098e-01 1.3318607e+00
6.1426187e-01 -1.2444510e-01 -2.4302281e-01 5.1733702e-01
3.1384537e-01 8.9684588e-01 -3.7520979e-02 -9.6867120e-01
5.5439371e-01 1.1572311e+00 4.8568696e-01 -1.0439906e+00
-6.3435489e-01 -6.9570011e-01 -2.0124097e-01 3.5813215e-01
6.0980535e-01 4.1460425e-02 -8.4396100e-01 1.5238780e+00
8.3126730e-01 -3.0649373e-01 -5.1163644e-01 1.2983798e+00
7.6007408e-01 2.1207280e-01 -5.9358197e-01 -2.7758902e-02
1.3395523e+00 -8.3463883e-01 -1.8357582e-01 -5.3414673e-01
4.5138463e-01 -1.0187919e+00 7.3838770e-01 2.7715608e-01
-1.2194656e+00 -3.3206475e-01 -1.1848750e+00 -2.5762600e-01
-3.6257841e-02 5.0891322e-01 4.2249572e-01 6.1918879e-01
-5.9139472e-01 -2.0925973e-01 -5.7792455e-01 -3.0394754e-01
2.4181132e-01 1.8089217e-01 -1.7295656e-01 -9.4544627e-02
-5.5080199e-01 1.1709716e+00 2.4522100e-01 -1.5464379e-01
-7.1365935e-01 -7.5212133e-01 -7.2147804e-01 -1.8775569e-02
7.4147230e-01 -9.9906993e-01 1.0631303e+00 -3.6282709e-01
-1.3589362e+00 1.1398607e+00 -2.0833488e-01 -2.6780608e-01
8.3391243e-01 -3.0788282e-01 2.1504515e-01 2.6715612e-01
1.9575201e-02 5.7718551e-01 9.8333842e-01 -1.3804348e+00
-6.4341170e-01 -7.3173106e-01 3.7710848e-01 6.9701022e-01
1.2733449e-01 8.0836654e-02 -6.7018872e-01 -3.1615484e-01
4.6799040e-01 -9.6993220e-01 3.7167605e-02 3.8139537e-01
-3.1660855e-01 1.0924998e-01 9.9028474e-01 -2.1688361e-01
7.1044546e-01 -2.2589574e+00 -1.1025803e-01 -2.6560834e-01
2.0252071e-01 1.7459579e-01 1.9603339e-01 1.9463448e-01
2.5950873e-01 -3.5035285e-01 1.6482867e-01 -2.7024338e-01
-3.3718236e-02 -2.4667835e-01 -2.1602249e-01 8.3809859e-01
2.0662247e-01 1.1147256e+00 -9.2777735e-01 -1.9512194e-01
4.7781450e-01 4.5731419e-01 -7.3686290e-01 1.6573673e-01
-6.5509275e-02 2.5547621e-01 -4.0805379e-01 8.8139606e-01
1.2238511e+00 -1.4565013e-01 -1.9975762e-01 -3.9925861e-01
-3.2454920e-01 2.8603464e-01 -1.2178636e+00 1.2608237e+00
-1.6250537e-01 4.3614149e-01 1.4836177e-01 -3.9584696e-01
9.0413111e-01 -1.3473615e-01 3.0838788e-01 -4.8551321e-01
3.2652119e-01 1.2354892e-01 -1.2957099e-01 -2.0001216e-01
5.0070214e-01 6.2957972e-02 1.1696734e-01 3.2795751e-01
-1.8737641e-01 -5.1436812e-01 -1.7532584e-01 1.4956143e-01
7.4258906e-01 4.8414314e-01 7.3608553e-01 4.7215126e-02
4.0706447e-01 -6.6865616e-02 2.9724562e-01 7.2857344e-01
-1.8089956e-01 1.0277730e+00 5.1803422e-01 -1.7725717e-01
-1.2014141e+00 -1.0760990e+00 -3.4028792e-01 3.6766031e-01
7.0203477e-01 -2.0495105e-01 -4.4078872e-01 -7.2123510e-01
4.1715100e-01 3.9094779e-01 -4.4038746e-01 -2.0728588e-01
-6.5605980e-01 -6.5906632e-01 1.4606483e-01 5.2351314e-01
4.6586800e-01 -4.1678402e-01 -1.0071797e+00 -1.2604660e-01
9.5240206e-02 -1.6456199e+00 -6.9415909e-01 4.8815068e-02
-6.6569066e-01 -1.0644017e+00 -5.3335315e-01 -1.1708277e-01
6.9235086e-01 1.1249350e+00 8.4447217e-01 -3.3816177e-01
-5.0967562e-01 4.5178449e-01 -1.6660994e-01 -5.0010109e-01
1.3853744e-02 -2.0023535e-01 3.0771406e-02 5.4985665e-02
2.0090657e-01 -3.5196152e-01 -1.1704876e+00 7.1103030e-01
-4.9017611e-01 2.5651014e-01 6.4102519e-01 7.2497922e-01
2.7120700e-01 -3.8056582e-01 3.7499782e-02 -4.9109909e-01
-2.0449047e-01 -1.2220636e-01 -1.1366719e+00 -2.8540313e-01
-5.0105304e-01 -4.2085055e-01 1.2615190e-01 -4.3293047e-01
-9.5037919e-01 3.1096604e-01 4.7276661e-01 -6.8523282e-01
-5.7119966e-02 -1.2341042e-01 -5.6849939e-01 -3.5054848e-01
5.1699406e-01 1.8754996e-01 -2.9343423e-01 -2.5773805e-01
5.1421833e-01 5.8782190e-01 3.4497187e-01 -2.6284540e-01
1.0888155e+00 1.1429253e+00 1.8969749e-01 -7.7149045e-01
-9.5005298e-01 -8.5890621e-01 -8.4851098e-01 -2.9937658e-01
8.0624127e-01 -1.1458894e+00 -7.1309632e-01 5.3509003e-01
-1.3736773e+00 1.8499042e-01 -8.3433568e-02 6.4708263e-01
-3.9323997e-01 4.5575002e-01 -1.4541301e-01 -9.6068329e-01
7.0863396e-02 -1.1690291e+00 1.3965633e+00 3.0664103e-02
1.2602466e-01 -5.6928235e-01 -4.4130337e-01 3.0941716e-01
6.9793351e-02 1.7463738e-01 7.6232141e-01 -2.4710885e-01
-1.2961046e+00 -4.9460983e-01 -6.9626451e-01 1.8644299e-01
-2.3530830e-02 -9.0287760e-02 -1.1206468e+00 -1.9601068e-01
3.5957417e-01 2.0264268e-01 6.4406675e-01 5.6866091e-01
5.4090506e-01 1.4741097e-01 -5.1731688e-01 9.4755799e-01
1.3315291e+00 1.7002468e-01 1.5964305e-01 4.5590228e-01
9.1385072e-01 3.6755189e-01 8.3377230e-01 4.8067942e-01
4.1080952e-01 1.0731226e+00 6.8278909e-01 -7.3045202e-02
-4.3026191e-01 -3.1687453e-01 3.7738097e-01 2.7077490e-01
2.2637637e-02 -1.1495680e-01 -6.9836742e-01 3.2228827e-01
-1.5166657e+00 -7.0854145e-01 -1.9800873e-01 2.3189185e+00
2.8051835e-01 2.4709038e-01 1.8003786e-01 -1.3511534e-01
4.6262124e-01 2.3611499e-01 -8.1081063e-01 1.8534559e-01
-4.4285756e-01 -5.1930195e-01 1.0392690e+00 4.7180712e-01
-1.0100845e+00 1.0062035e+00 6.3019977e+00 5.4127008e-01
-1.2440950e+00 -2.7993649e-02 9.7947881e-02 -4.0076351e-01
-1.7549212e-01 2.1843424e-01 -1.5539669e+00 1.8035874e-01
-1.5920426e-01 -5.8456335e-02 1.7275605e-01 8.9134443e-01
3.7381771e-01 -4.1439646e-01 -1.2052974e+00 1.1042370e+00
3.5929546e-01 -7.0894986e-01 4.3126427e-02 3.0435029e-01
9.2845482e-01 4.8261654e-02 1.8369728e-01 -2.7039954e-01
-3.6864426e-02 -4.8425201e-01 1.0787822e+00 1.7680317e-01
4.2908999e-01 -4.3486106e-01 3.1538221e-01 5.7612956e-01
-1.0707642e+00 -2.4001026e-01 -4.5785034e-01 -5.2796364e-02
1.7643033e-01 7.6726270e-01 -1.2389393e+00 3.8186294e-01
4.8573151e-01 7.5312078e-01 -7.3042452e-01 1.2510413e+00
-3.1379458e-01 3.3726420e-02 -5.2716261e-01 2.4206342e-01
3.8175106e-02 -2.0313358e-01 1.2148422e+00 9.8891371e-01
3.3127123e-01 -1.9196807e-01 -1.0408649e-01 1.1124758e+00
9.0896271e-02 -1.1114073e-01 -7.4351192e-01 3.0730239e-01
5.7729506e-01 1.2208540e+00 -5.9618920e-01 -4.3208495e-02
-6.8626505e-01 6.8813038e-01 -2.2231933e-02 5.1541442e-01
-8.4334636e-01 6.4250275e-02 5.8196831e-01 5.3296602e-01
7.4333102e-01 -4.6404406e-01 -5.5728364e-01 -1.4409239e+00
5.9229064e-01 -7.3439860e-01 8.4281594e-02 -1.0758783e+00
-9.3316942e-01 1.6719867e-01 1.9892898e-01 -1.5609795e+00
-4.0286992e-02 -8.1496352e-01 -3.0388311e-01 6.5286499e-01
-1.6296582e+00 -1.2960219e+00 -5.2578592e-01 3.6340162e-01
4.6805140e-01 8.0033220e-02 8.5909054e-02 1.2134664e-01
-1.4080797e-01 4.9065539e-01 -2.2357841e-01 -2.2881967e-01
9.8259425e-01 -1.1426265e+00 3.4111789e-01 9.9943328e-01
-6.6543169e-02 6.9454885e-01 7.1300912e-01 -4.8842821e-01
-1.7640965e+00 -6.9599235e-01 5.7200772e-01 -1.1092703e+00
4.8053759e-01 -8.9937001e-01 -5.4092264e-01 6.1339432e-01
-2.2542907e-01 1.9478239e-01 -2.6586434e-01 -2.9191315e-01
-5.6100488e-01 -1.5703164e-01 -7.8070605e-01 6.1271220e-01
1.2982079e+00 -4.8561004e-01 -3.7081492e-01 5.3052723e-02
6.0760593e-01 -8.3581853e-01 -3.5992983e-01 6.8229413e-01
7.8091967e-01 -1.2754126e+00 1.4090337e+00 4.4402864e-02
1.6574676e-01 -6.8657404e-01 -1.7793128e-01 -1.0105789e+00
-2.3688823e-01 -2.2006153e-01 -1.7518759e-01 9.6860850e-01
1.4935727e-01 -7.1373504e-01 8.0456144e-01 2.7254650e-01
-1.9360065e-01 -5.6321788e-01 -1.0080987e+00 -8.1018293e-01
-2.3895416e-01 -2.6096457e-01 4.9493071e-01 6.8431002e-01
-3.7705505e-01 2.8480205e-01 -3.5718691e-01 4.7201112e-01
5.6553388e-01 4.7603166e-01 1.4130788e+00 -1.1085739e+00
-3.2590565e-01 -4.7799024e-01 -6.3802600e-01 -1.8778926e+00
-3.5676810e-01 -7.5976008e-01 -2.5488132e-01 -1.0733405e+00
5.2278733e-01 -2.0109111e-01 -6.8082221e-02 -1.9713326e-01
-2.2000660e-01 2.0061471e-01 2.9207647e-01 2.6939151e-01
-3.2270566e-01 5.2604383e-01 1.5616775e+00 1.6864981e-01
-1.9378763e-01 8.4100895e-02 -6.5109730e-01 8.8487667e-01
2.8500220e-01 -4.3487689e-01 -2.4097681e-01 -4.0709588e-01
2.8931317e-01 -4.5944139e-02 7.4501628e-01 -7.3121262e-01
2.4557601e-01 -3.0848712e-01 6.8322295e-01 -1.1722928e+00
5.4991412e-01 -9.8359531e-01 -1.6461349e-01 2.2330497e-01
1.2908183e-01 -2.9191315e-01 -7.0241094e-04 5.7244223e-01
1.8311886e-01 9.0650655e-03 9.9231935e-01 -1.5152606e-01
-3.2543707e-01 2.4540013e-01 8.1467159e-02 -9.3399175e-02
1.0923651e+00 -5.5383492e-01 -4.5177090e-01 -2.6307330e-01
-2.4919190e-01 9.1254875e-02 1.0898015e+00 6.0937482e-01
6.2079698e-01 -1.2275248e+00 -4.8737124e-01 5.0855333e-01
3.9533108e-01 5.5634793e-02 -1.6542634e-01 1.1203965e+00
-4.4245446e-01 6.9238955e-01 1.5698466e-01 -1.0547866e+00
-1.3652091e+00 6.9097638e-01 6.0162741e-01 2.0533843e-01
-6.5673518e-01 6.4269030e-01 9.3447149e-01 -3.5790521e-01
1.3239321e-01 -4.7184250e-01 9.6003272e-02 -1.1012325e-01
2.6784170e-01 3.5611030e-01 -8.8694692e-03 -5.7730502e-01
-3.7138060e-01 1.0758624e+00 1.1400045e-02 1.8161756e-01
1.0431174e+00 -6.3680279e-01 2.7493364e-01 5.1724476e-01
9.8031276e-01 4.0217227e-01 -1.6861340e+00 -3.6901313e-01
-3.5320425e-01 -9.3394870e-01 3.6536225e-03 -7.2826821e-01
-7.1115267e-01 9.1596240e-01 5.3406268e-01 -3.6266944e-01
7.7362251e-01 2.2710475e-01 4.1491491e-01 9.9273354e-02
4.8230326e-01 -7.5368506e-01 6.6447414e-02 5.5272335e-01
8.1492859e-01 -1.3689615e+00 3.1614253e-01 -7.6716226e-01
-2.4627885e-01 9.8272663e-01 6.8328786e-01 -1.4774464e-01
3.4573376e-01 1.6122165e-01 8.4902383e-02 -1.4344259e-01
-4.7313395e-01 -2.2518431e-01 5.5643654e-01 5.2139097e-01
-5.3501964e-02 -1.3222414e-01 -1.7314035e-01 2.7483848e-01
-8.3622582e-02 -3.6255956e-01 3.8175583e-01 6.6537613e-01
-4.0716508e-01 -8.5705334e-01 -5.3109711e-01 -5.7177350e-02
-2.0772000e-01 8.6044364e-02 -5.9874934e-01 9.8205096e-01
2.1880160e-01 5.5783802e-01 -7.6057918e-02 -8.9325495e-02
5.6047738e-01 -2.5896990e-01 8.8627225e-01 -5.7539022e-01
-2.2162503e-01 5.5155396e-01 -1.7327762e-01 -4.9185789e-01
-1.9493251e-01 -7.6309580e-01 -7.6345199e-01 9.9178426e-02
-7.5528318e-01 -6.2383020e-01 8.2201540e-01 7.6069283e-01
3.2202065e-01 -3.1891499e-02 7.7508539e-01 -1.1934615e+00
-5.8194816e-01 -7.0541269e-01 -5.3413302e-01 2.2720900e-01
4.4626227e-01 -1.0334462e+00 -8.5538381e-01 -3.2555926e-01] | [7.846647262573242, -2.551821231842041] |
89ddf788-b025-4c60-a38c-01c9226476d6 | part-based-deep-hashing-for-large-scale | 1705.02145 | null | http://arxiv.org/abs/1705.02145v1 | http://arxiv.org/pdf/1705.02145v1.pdf | Part-based Deep Hashing for Large-scale Person Re-identification | Large-scale is a trend in person re-identification (re-id). It is important
that real-time search be performed in a large gallery. While previous methods
mostly focus on discriminative learning, this paper makes the attempt in
integrating deep learning and hashing into one framework to evaluate the
efficiency and accuracy for large-scale person re-id. We integrate spatial
information for discriminative visual representation by partitioning the
pedestrian image into horizontal parts. Specifically, Part-based Deep Hashing
(PDH) is proposed, in which batches of triplet samples are employed as the
input of the deep hashing architecture. Each triplet sample contains two
pedestrian images (or parts) with the same identity and one pedestrian image
(or part) of the different identity. A triplet loss function is employed with a
constraint that the Hamming distance of pedestrian images (or parts) with the
same identity is smaller than ones with the different identity. In the
experiment, we show that the proposed Part-based Deep Hashing method yields
very competitive re-id accuracy on the large-scale Market-1501 and
Market-1501+500K datasets. | ['Liang Zheng', 'Xiangwei Kong', 'Haiyan Fu', 'Qi Tian', 'Fuqing Zhu'] | 2017-05-05 | null | null | null | null | ['large-scale-person-re-identification'] | ['computer-vision'] | [-5.40857315e-01 -6.18778169e-01 2.88219471e-02 -3.52982014e-01
-8.86123180e-01 -3.40885460e-01 6.17028594e-01 3.03626835e-01
-7.61458039e-01 6.96151555e-01 4.20129716e-01 3.43921274e-01
3.32580775e-01 -8.23731065e-01 -7.13500321e-01 -9.48600054e-01
6.44598231e-02 6.42606676e-01 1.47640690e-01 7.41886571e-02
1.20536089e-01 4.66692150e-01 -1.72719657e+00 5.05611561e-02
2.70540029e-01 1.02220917e+00 2.40949262e-02 4.80581522e-01
5.17883301e-01 1.47416249e-01 -7.79302239e-01 -8.62660885e-01
6.49631560e-01 -3.12213808e-01 -7.57978618e-01 -1.04035951e-01
7.55899489e-01 -6.50638103e-01 -8.19865108e-01 1.00512195e+00
1.15625441e+00 2.84899801e-01 6.42310202e-01 -1.50813055e+00
-8.77907872e-01 1.88891321e-01 -6.99722350e-01 3.12523544e-01
4.78992283e-01 1.74689904e-01 8.42045367e-01 -9.80444133e-01
2.93141544e-01 1.39818883e+00 8.65637600e-01 4.70570445e-01
-1.24739468e+00 -7.71579802e-01 -3.62530559e-01 7.08877265e-01
-2.12593365e+00 -4.10109282e-01 4.69628841e-01 -3.67976278e-01
8.24568927e-01 4.93843444e-02 7.87859261e-01 8.77185643e-01
1.01852998e-01 7.60956764e-01 8.22096586e-01 -3.78634393e-01
8.79201666e-02 3.68539512e-01 3.78551751e-01 3.99962693e-01
6.71872854e-01 1.31183878e-01 -4.78826493e-01 -2.28759393e-01
3.44074428e-01 4.31699187e-01 6.59050196e-02 -3.77368063e-01
-1.03442299e+00 1.05634952e+00 7.69763887e-01 1.91638559e-01
-2.23097578e-01 1.77124500e-01 5.67273974e-01 6.41375482e-02
-1.47061363e-01 -7.65980184e-02 2.68018633e-01 9.91757512e-02
-1.01358008e+00 5.49019873e-01 4.48325515e-01 8.69161427e-01
9.44432497e-01 -6.30407095e-01 -3.71866554e-01 8.96381855e-01
1.23123474e-01 6.22503638e-01 8.36476743e-01 -5.12101531e-01
3.17245543e-01 6.46796286e-01 2.77209103e-01 -1.12605023e+00
-9.51514468e-02 -1.01091281e-01 -1.08698237e+00 -1.45940050e-01
3.92204970e-01 1.01315044e-01 -7.57447541e-01 1.67127967e+00
4.03726041e-01 5.19914553e-02 2.02291962e-02 1.08858538e+00
8.56544197e-01 5.69165051e-01 8.06627497e-02 3.34971547e-01
1.80154240e+00 -1.07447696e+00 -2.18158677e-01 7.53210336e-02
4.86750722e-01 -8.39099467e-01 6.16611063e-01 -4.20455672e-02
-9.44631100e-01 -1.16652036e+00 -1.00152004e+00 -3.55206668e-01
-6.92615986e-01 3.22767287e-01 -2.70896535e-02 8.72456491e-01
-1.08807302e+00 2.02714697e-01 -2.03333214e-01 -7.24043250e-01
1.21029027e-01 5.75238824e-01 -8.02707613e-01 -3.77870440e-01
-1.30198526e+00 7.68611491e-01 4.27903116e-01 -6.09279517e-03
-9.84389901e-01 -2.09643692e-01 -9.87811983e-01 1.15696549e-01
-3.90418023e-01 -6.95590436e-01 7.23981738e-01 -5.09815753e-01
-7.68480659e-01 1.24864769e+00 -2.43283555e-01 -7.40164876e-01
4.74625736e-01 -1.20756768e-01 -4.16803390e-01 1.08014077e-01
4.34925407e-01 1.04582047e+00 7.74346054e-01 -1.14737105e+00
-8.27125192e-01 -6.72316730e-01 -1.16402127e-01 1.91778585e-01
-4.19965327e-01 3.57529260e-02 -5.95548570e-01 -5.14854133e-01
-3.29761177e-01 -1.13633251e+00 1.97897017e-01 -1.37265906e-01
-6.14793718e-01 -3.72704446e-01 8.58610153e-01 -7.88659871e-01
1.18552804e+00 -2.23558474e+00 -6.39192834e-02 1.55017436e-01
6.81586638e-02 2.54246563e-01 -1.25387087e-01 5.60742199e-01
-1.61484275e-02 -1.71527058e-01 1.75941542e-01 -8.34218025e-01
2.98309505e-01 5.53067913e-03 -1.71087682e-02 6.85501516e-01
-2.93313205e-01 7.61599839e-01 -5.38586736e-01 -6.59155250e-01
2.49994487e-01 5.95366538e-01 -2.89599568e-01 2.79217094e-01
7.68738091e-01 -5.52575663e-02 -4.61536460e-02 7.10718930e-01
1.02267611e+00 -1.48900211e-01 -2.45266765e-01 -5.10581732e-01
-1.16912991e-01 1.37135312e-02 -1.35628772e+00 1.39476311e+00
-2.96454784e-03 4.27395284e-01 -2.40532994e-01 -7.73752749e-01
8.16528618e-01 1.96406320e-01 3.83803636e-01 -9.36160743e-01
-1.62310168e-01 6.72965273e-02 -2.36171842e-01 7.86095709e-02
8.84077489e-01 6.76491857e-02 -3.48382920e-01 5.60720444e-01
-9.13168397e-03 6.12630486e-01 2.93200791e-01 1.14930116e-01
7.28100002e-01 -3.71986330e-01 2.61123091e-01 -1.37783363e-01
8.26841474e-01 -2.01776460e-01 4.19401556e-01 7.80181825e-01
-7.60549486e-01 8.07693958e-01 -1.03176601e-01 -8.08455646e-01
-1.58444369e+00 -9.71054494e-01 -2.89097369e-01 1.03512836e+00
6.25221550e-01 -4.13922578e-01 -6.76690698e-01 -5.34061491e-01
4.79722023e-01 8.41384232e-02 -8.22290003e-01 -2.05742329e-01
-5.88726938e-01 -8.24601650e-01 7.65020728e-01 5.83101690e-01
1.12089419e+00 -8.56172144e-01 -4.75874841e-01 4.32461202e-02
-4.68962520e-01 -9.77318347e-01 -1.04225183e+00 -9.05527696e-02
-1.60629973e-01 -1.00317299e+00 -1.46825993e+00 -1.21824348e+00
5.75023174e-01 6.49404466e-01 7.58996367e-01 4.93025407e-02
-6.04031563e-01 3.99704456e-01 -2.41104573e-01 3.53685111e-01
8.93891901e-02 2.41579097e-02 4.65294659e-01 2.75175512e-01
8.44264686e-01 -5.96965775e-02 -1.16942263e+00 7.39372790e-01
-5.65512002e-01 -3.57136071e-01 4.26135361e-01 1.01691794e+00
6.10785663e-01 1.60022005e-01 5.60645126e-02 1.83005393e-01
4.19943780e-01 -3.73377800e-01 -2.33002514e-01 4.23129290e-01
-3.67293477e-01 -6.25128224e-02 4.80294406e-01 -3.77113491e-01
-5.43286979e-01 1.43527538e-01 -1.13934822e-01 -2.39391685e-01
-1.87765181e-01 -1.26099378e-01 -3.25079739e-01 -2.39900216e-01
3.79678071e-01 7.06871569e-01 -1.99620947e-01 -5.26685238e-01
2.15108663e-01 8.64562690e-01 6.53892636e-01 -3.89586508e-01
8.44844401e-01 4.84570384e-01 -1.87124968e-01 -6.50313973e-01
-1.26808926e-01 -8.00308406e-01 -8.92020524e-01 4.48078364e-02
1.03575814e+00 -1.26798606e+00 -8.41903090e-01 7.99385250e-01
-9.17841554e-01 1.94269821e-01 -1.72743022e-01 3.59143168e-01
-2.28640083e-02 7.99110830e-01 -7.57274270e-01 -5.83040237e-01
-6.49109423e-01 -1.18329620e+00 1.33771896e+00 5.36090612e-01
-2.85995379e-02 -4.92288023e-01 4.33914244e-01 3.80688518e-01
6.16711825e-02 -7.41500854e-02 4.74542767e-01 -8.24381709e-01
-4.67151046e-01 -7.08126187e-01 -5.86521626e-01 2.88483083e-01
1.40210167e-01 -4.31334347e-01 -8.66187334e-01 -8.68737757e-01
-5.30860901e-01 -4.40381885e-01 8.61406922e-01 7.43903965e-02
8.08857441e-01 -2.81414896e-01 -5.40360928e-01 5.25080025e-01
1.40429664e+00 1.43544555e-01 7.39792943e-01 6.46182775e-01
6.96263075e-01 4.58062202e-01 4.66174990e-01 7.84675539e-01
8.43933165e-01 1.11465752e+00 4.23752517e-02 -3.88663597e-02
-1.74075514e-01 -4.31359053e-01 2.71902233e-01 2.18673289e-01
1.19440794e-01 -1.58174425e-01 -6.85069323e-01 8.22879970e-01
-1.65548170e+00 -1.31452501e+00 3.17629129e-01 2.54251719e+00
5.17895520e-01 -3.02231073e-01 7.56926477e-01 2.17434838e-01
1.19693911e+00 -3.16291489e-02 -4.47858602e-01 -6.87026232e-02
-2.04297364e-01 -2.22321033e-01 5.57719827e-01 2.67741561e-01
-1.46120942e+00 6.89380348e-01 5.49797297e+00 8.94415379e-01
-8.05984020e-01 -4.40789461e-02 6.46491766e-01 7.06821084e-02
2.18728513e-01 -2.52668172e-01 -1.48695314e+00 9.29582596e-01
8.06322455e-01 -3.34791243e-02 1.49750710e-01 8.59391272e-01
-2.54578203e-01 -2.30825469e-01 -1.16257107e+00 1.55097067e+00
2.78140068e-01 -8.79085124e-01 2.18744218e-01 4.04675484e-01
4.41182017e-01 -4.56968993e-01 2.97375798e-01 3.08436453e-01
1.67977229e-01 -9.02907550e-01 7.27578640e-01 1.88468680e-01
8.10444355e-01 -1.05669022e+00 1.03853285e+00 1.47295725e-02
-1.72354877e+00 -1.34656131e-01 -6.78374410e-01 2.28716761e-01
7.02346787e-02 8.57577547e-02 -6.75545275e-01 4.91871029e-01
1.18497157e+00 5.60110986e-01 -8.06527019e-01 1.38109982e+00
3.42267901e-01 -3.20892385e-03 -4.54144806e-01 1.30089611e-01
9.15881470e-02 1.06692419e-03 1.50995523e-01 1.35576284e+00
4.15217996e-01 -1.43789947e-01 2.45882332e-01 5.14406323e-01
-3.21856327e-02 4.03882004e-02 -5.99829018e-01 3.06331217e-01
5.97289920e-01 1.07704628e+00 -4.03347075e-01 -5.66022575e-01
-4.46786284e-01 1.56785607e+00 1.51375398e-01 1.41677529e-01
-7.82122850e-01 -4.17358190e-01 9.19686079e-01 9.61370245e-02
6.22942328e-01 -1.55903369e-01 1.43512383e-01 -9.95116889e-01
-1.23318881e-01 -7.21417665e-01 7.47971892e-01 -4.26904917e-01
-1.56719410e+00 4.84578431e-01 1.17124766e-02 -1.24744833e+00
-1.67708531e-01 -3.25923502e-01 -3.62428010e-01 1.03788686e+00
-1.19171035e+00 -1.49509847e+00 -5.62217772e-01 1.02666700e+00
1.44561082e-01 -3.07680219e-01 8.28910053e-01 7.21472979e-01
-7.74553239e-01 1.29596162e+00 3.55658829e-01 6.65496826e-01
9.66429889e-01 -9.82187033e-01 5.51933587e-01 8.50943387e-01
-2.96877827e-02 8.07691514e-01 4.39316899e-01 -5.98075151e-01
-1.21612835e+00 -9.71294522e-01 1.12246037e+00 -2.59246171e-01
1.36375889e-01 -5.02734244e-01 -5.74205697e-01 4.12894815e-01
1.18193179e-01 1.36537209e-01 9.55857635e-01 -3.57034534e-01
-6.24673128e-01 -3.42083752e-01 -1.48503458e+00 3.78480822e-01
7.60303199e-01 -8.14266860e-01 -5.35090566e-01 1.55114800e-01
3.05029154e-01 3.44007425e-02 -7.73738027e-01 9.09717306e-02
8.21466327e-01 -9.87912595e-01 1.49293387e+00 -2.52063125e-01
-9.96074304e-02 -5.08788109e-01 -4.54454631e-01 -7.74597704e-01
-6.34226978e-01 -6.31121770e-02 1.27100542e-01 1.27425504e+00
-3.30565512e-01 -5.86159468e-01 7.51166999e-01 6.57064438e-01
3.55492294e-01 -2.36506507e-01 -1.16678905e+00 -8.47000420e-01
-6.53307652e-03 5.37613392e-01 8.41434479e-01 5.17560840e-01
-3.29018950e-01 3.22771110e-02 -7.49364018e-01 2.21553594e-01
9.77028728e-01 2.66342133e-01 1.06469095e+00 -1.12813282e+00
-9.11529884e-02 -1.74231485e-01 -8.88257027e-01 -1.02982306e+00
-1.17461339e-01 -6.48243368e-01 -9.32667851e-02 -1.24182487e+00
8.23869944e-01 -2.28773803e-01 -4.50558692e-01 3.62964928e-01
-1.84953123e-01 7.50687242e-01 3.08881909e-01 6.20839775e-01
-7.20704556e-01 7.72381485e-01 6.90650165e-01 -4.95058835e-01
1.79807499e-01 -1.05505750e-01 -4.03734297e-01 1.92708030e-01
4.92029518e-01 -3.03583801e-01 7.70664141e-02 -1.59040377e-01
-5.23799479e-01 -2.95059413e-01 7.89600849e-01 -1.38084018e+00
5.35029113e-01 6.25572801e-01 9.46952164e-01 -9.00941312e-01
5.46301484e-01 -7.46074975e-01 2.78855741e-01 7.71989942e-01
-2.26412997e-01 5.30149698e-01 -2.34964862e-02 5.29894531e-01
-2.93352693e-01 -2.03261912e-01 9.95961130e-01 -1.34252846e-01
-9.88396883e-01 4.24256295e-01 1.58164427e-01 -2.33569846e-01
1.14954245e+00 -5.96018076e-01 -2.40225822e-01 -2.25385070e-01
-4.46847379e-01 1.26217276e-01 8.59777868e-01 2.66433775e-01
7.90415585e-01 -1.80372930e+00 -6.84693873e-01 3.45023692e-01
3.38880062e-01 -6.06447756e-01 4.21966821e-01 3.97226453e-01
-4.49638814e-01 6.86497569e-01 -5.26809156e-01 -4.22268331e-01
-1.65048921e+00 9.08510923e-01 4.14174020e-01 -2.65750945e-01
-5.84667861e-01 9.67584610e-01 2.92212963e-01 -1.97569549e-01
5.16143262e-01 3.80031735e-01 -1.09198101e-01 1.46854326e-01
9.12779808e-01 6.19767725e-01 -1.56199381e-01 -1.28507447e+00
-5.81718266e-01 7.59675145e-01 -4.98508126e-01 -1.13161199e-01
8.57859254e-01 -3.48489523e-01 -7.87683483e-03 -7.92625248e-02
1.59275842e+00 -2.67028302e-01 -1.05954242e+00 -4.57344234e-01
-3.39028299e-01 -7.52491891e-01 -3.79961282e-01 -4.41436529e-01
-8.09643567e-01 8.02741826e-01 1.39897704e+00 -1.50748655e-01
8.73362064e-01 -1.30943909e-01 1.45519805e+00 2.27180690e-01
6.62317276e-01 -1.09156775e+00 2.39857569e-01 1.43145889e-01
5.41097939e-01 -1.37801850e+00 -2.14435421e-02 3.73258173e-01
-6.70083165e-01 7.37158477e-01 5.72141171e-01 -2.11853325e-01
5.22281587e-01 -1.52452305e-01 -3.35264206e-01 1.21137343e-01
-3.00292075e-02 -3.11851472e-01 3.12839448e-01 7.25334823e-01
-3.15207131e-02 9.89174247e-02 -8.83505866e-02 3.71752262e-01
-1.88547701e-01 -3.03948104e-01 -1.01391196e-01 7.79842317e-01
-4.45061356e-01 -1.22058654e+00 -7.61385083e-01 -5.96862007e-03
-1.85984656e-01 -4.85793836e-02 -3.66830468e-01 7.81027079e-01
4.04835761e-01 7.41762459e-01 1.96211368e-01 -6.64146125e-01
5.05796000e-02 -8.62988606e-02 4.08338934e-01 -1.92383692e-01
-6.38330460e-01 -2.40405962e-01 -3.07472616e-01 -3.52878124e-01
-9.14237201e-02 -6.42393112e-01 -8.98122728e-01 -8.36624324e-01
4.34130579e-02 2.81075239e-01 3.26415330e-01 5.11028707e-01
2.94783384e-01 -4.18339193e-01 6.24766529e-01 -1.01477838e+00
-4.39455867e-01 -7.43838489e-01 -7.79286504e-01 7.78940022e-01
5.00450015e-01 -7.16288686e-01 -2.47022077e-01 -3.12259998e-02] | [14.737329483032227, 0.905683696269989] |
a4630821-ceb2-4581-adc3-2da9a1d9940e | cg-nerf-conditional-generative-neural | 2112.03517 | null | https://arxiv.org/abs/2112.03517v1 | https://arxiv.org/pdf/2112.03517v1.pdf | CG-NeRF: Conditional Generative Neural Radiance Fields | While recent NeRF-based generative models achieve the generation of diverse 3D-aware images, these approaches have limitations when generating images that contain user-specified characteristics. In this paper, we propose a novel model, referred to as the conditional generative neural radiance fields (CG-NeRF), which can generate multi-view images reflecting extra input conditions such as images or texts. While preserving the common characteristics of a given input condition, the proposed model generates diverse images in fine detail. We propose: 1) a novel unified architecture which disentangles the shape and appearance from a condition given in various forms and 2) the pose-consistent diversity loss for generating multimodal outputs while maintaining consistency of the view. Experimental results show that the proposed method maintains consistent image quality on various condition types and achieves superior fidelity and diversity compared to existing NeRF-based generative models. | ['Jaegul Choo', 'Soyoung Yang', 'Sanghun Jung', 'Gyumin Shim', 'Kyungmin Jo'] | 2021-12-07 | null | null | null | null | ['3d-aware-image-synthesis'] | ['computer-vision'] | [ 2.87487477e-01 -2.64731526e-01 1.61822811e-01 -3.33430648e-01
-6.60650849e-01 -5.56544065e-01 7.84532309e-01 -6.43088877e-01
3.93257499e-01 7.94178188e-01 2.84630090e-01 4.05317508e-02
-9.00199562e-02 -8.59542847e-01 -7.06397951e-01 -9.84550834e-01
6.25331223e-01 2.40008160e-01 -3.58814932e-02 -2.37714425e-01
2.23431543e-01 7.07522333e-01 -1.97011960e+00 3.90706092e-01
1.17690647e+00 8.09611559e-01 5.52519739e-01 8.01818788e-01
2.21448705e-01 5.09912014e-01 -7.34026670e-01 -3.69119465e-01
2.83408761e-01 -8.46763253e-01 -2.48023286e-01 5.22196531e-01
7.82391310e-01 -5.50087988e-01 -3.20788741e-01 9.32498753e-01
8.80403697e-01 8.47103223e-02 1.02650404e+00 -1.15879643e+00
-1.34597492e+00 1.30951643e-01 -6.88226283e-01 -5.66588081e-02
6.15938187e-01 2.12145373e-01 6.90225959e-01 -8.00212145e-01
7.26746082e-01 1.43104911e+00 3.19000214e-01 7.81286955e-01
-1.36196446e+00 -4.69233513e-01 7.77900666e-02 -9.94841605e-02
-1.51771080e+00 -2.89513141e-01 9.41688538e-01 -2.21033946e-01
6.74951017e-01 4.05491143e-01 4.67509001e-01 1.34918499e+00
4.93702441e-01 6.39192343e-01 1.29344213e+00 -2.93920994e-01
4.90926988e-02 -2.75729666e-03 -5.41720569e-01 4.85454917e-01
2.25419432e-01 2.30867997e-01 -5.99591196e-01 -3.14682722e-02
9.78919744e-01 -8.01405162e-02 -6.04330420e-01 -4.41794395e-01
-1.20249259e+00 4.98622715e-01 2.30955929e-01 9.12878290e-02
-4.26460266e-01 1.23837233e-01 -2.86070228e-01 4.82067503e-02
3.98544788e-01 1.28828391e-01 5.02172709e-02 2.76675045e-01
-7.51900077e-01 3.84659946e-01 3.22802067e-01 1.33632195e+00
7.75177360e-01 5.98211288e-01 -6.30260885e-01 9.85610425e-01
3.83337587e-01 1.04935873e+00 2.45159671e-01 -7.56458998e-01
3.60553563e-01 3.68725419e-01 1.92010060e-01 -1.02790976e+00
-1.22559354e-01 -5.99098980e-01 -1.10665500e+00 2.88679123e-01
-1.67494148e-01 -2.34275207e-01 -1.31189585e+00 1.90818477e+00
2.34448522e-01 -1.11550957e-01 1.16619110e-01 8.52047682e-01
1.15484321e+00 9.28769648e-01 -7.98435435e-02 -1.26146898e-01
8.84361267e-01 -6.67447865e-01 -9.48069811e-01 7.83848688e-02
-2.65806407e-01 -1.04631186e+00 8.37128699e-01 3.38135839e-01
-1.40914547e+00 -9.70796585e-01 -9.91074145e-01 5.14317192e-02
-1.88953444e-01 2.38986731e-01 3.90899211e-01 9.77441728e-01
-1.42069101e+00 9.99979228e-02 -2.04654738e-01 -1.34662002e-01
6.11369498e-02 -3.58505696e-02 -1.66745782e-01 -4.20210928e-01
-1.10349524e+00 6.72202170e-01 2.45025441e-01 -1.62021928e-02
-1.21690667e+00 -6.58000708e-01 -1.01516378e+00 6.12620153e-02
-2.30740253e-02 -1.29703379e+00 8.89869690e-01 -7.83637106e-01
-1.53692067e+00 6.12465322e-01 -2.14868501e-01 3.62969369e-01
3.38350415e-01 2.92420704e-02 -5.72791815e-01 1.28315642e-01
-1.40257284e-01 1.00800216e+00 1.03481221e+00 -2.14691687e+00
-4.25984591e-01 -2.28797734e-01 5.81216179e-02 5.38285315e-01
-5.17623918e-03 -3.69997114e-01 -5.64368486e-01 -7.73475468e-01
1.20946713e-01 -8.01500082e-01 -1.10567190e-01 -2.11914450e-01
-6.13848865e-01 2.80478776e-01 9.97024715e-01 -3.75057310e-01
1.08519924e+00 -1.88791609e+00 2.98950613e-01 2.19133943e-01
6.19849227e-02 5.85560165e-02 -3.23634356e-01 7.26356447e-01
-2.66766567e-02 1.18550546e-01 -9.06643271e-02 -3.19198817e-01
2.71781702e-02 2.68061638e-01 -8.74454677e-02 2.13529155e-01
3.38992864e-01 1.05380225e+00 -6.94097161e-01 -3.38345677e-01
4.10053939e-01 1.02394450e+00 -5.98361731e-01 3.49247545e-01
-2.07621232e-01 6.41165555e-01 -3.43853444e-01 6.70577288e-01
1.20016360e+00 -1.10794596e-01 -3.14086787e-02 -5.50626934e-01
2.11428538e-01 -4.56029296e-01 -1.39686060e+00 1.69572794e+00
-4.95210558e-01 3.02716613e-01 -1.11535624e-01 -3.25371832e-01
1.14672554e+00 2.76250452e-01 2.24871546e-01 -7.37900674e-01
1.37383655e-01 -3.97804119e-02 -2.45050430e-01 -4.43221658e-01
8.99151623e-01 -2.79676437e-01 -2.59448588e-02 3.40317041e-01
2.99099714e-01 -4.51734036e-01 1.82021141e-01 2.46055126e-01
4.04648393e-01 5.24021745e-01 -1.50231784e-02 -1.36818752e-01
4.01718587e-01 -5.02822816e-01 4.40902114e-01 7.54441679e-01
2.83264607e-01 1.39770925e+00 -6.85406942e-03 -1.26123562e-01
-1.19614422e+00 -1.48955297e+00 -8.72303545e-02 5.49594820e-01
6.23586535e-01 -2.22567648e-01 -5.58413267e-01 -3.43172759e-01
-2.53079534e-01 8.18549752e-01 -5.90676248e-01 -8.88750255e-02
-2.14355752e-01 -8.36782634e-01 4.33624059e-01 3.82094115e-01
7.79448867e-01 -8.11010897e-01 -4.13371950e-01 -2.03655720e-01
-5.66715002e-01 -1.04728782e+00 -6.81844294e-01 -1.91207409e-01
-5.83507180e-01 -6.36035860e-01 -1.12659276e+00 -7.70666718e-01
1.09085190e+00 5.13265550e-01 1.29480314e+00 -2.65637487e-01
-3.80817592e-01 6.66299641e-01 -3.75061303e-01 -2.17008844e-01
-6.81260884e-01 -3.62238526e-01 -2.65095919e-01 3.40864927e-01
-2.88695931e-01 -6.23262823e-01 -7.82780230e-01 3.67922276e-01
-1.37318981e+00 4.12603945e-01 6.61964774e-01 8.88438642e-01
8.62402856e-01 2.44033590e-01 6.97207689e-01 -7.24132955e-01
7.75340140e-01 -3.35778266e-01 -2.26035550e-01 4.76970583e-01
-5.74279487e-01 5.03753312e-02 6.89849019e-01 -4.45734143e-01
-1.88906443e+00 -5.42205684e-02 -5.58153726e-02 -5.86055815e-01
-4.32087213e-01 9.69548002e-02 -5.66165984e-01 -1.16386674e-01
6.55648112e-01 7.40785956e-01 -1.22596569e-01 -3.33211154e-01
4.93351668e-01 4.39229369e-01 6.48970425e-01 -7.36846864e-01
9.86826360e-01 3.86057824e-01 1.28325135e-01 -5.26595712e-01
-4.06707406e-01 -5.93261756e-02 -3.26246560e-01 -5.71087420e-01
9.33350146e-01 -1.09001255e+00 -3.27889711e-01 7.59797454e-01
-1.04033208e+00 1.08873434e-01 -1.53932795e-01 3.01815361e-01
-7.22820163e-01 2.97480017e-01 -5.13838172e-01 -8.21060359e-01
-2.36439958e-01 -1.17910302e+00 1.44602036e+00 6.26608133e-01
1.95412442e-01 -8.55438471e-01 7.32042715e-02 2.62057126e-01
6.19030952e-01 5.02235889e-01 9.37879801e-01 3.57788682e-01
-8.65826547e-01 -7.29962112e-03 -2.20939338e-01 3.04205120e-01
4.51189041e-01 8.45422000e-02 -1.04344165e+00 -2.76605606e-01
-1.01661630e-01 -1.90230593e-01 5.64465523e-01 4.60380942e-01
9.47101593e-01 -1.43944785e-01 -6.56480119e-02 7.51635194e-01
1.82100189e+00 3.89087379e-01 1.19619071e+00 -9.79790166e-02
7.43965864e-01 4.12006110e-01 3.22828680e-01 6.03456557e-01
2.10869163e-01 5.00551760e-01 4.97233897e-01 -4.37707871e-01
-6.10128343e-01 -5.65957665e-01 2.35886991e-01 6.70417964e-01
-2.27414265e-01 -8.90083849e-01 -3.06891799e-01 5.02857983e-01
-1.65637934e+00 -1.31546307e+00 -4.31298949e-02 1.92279625e+00
6.48943663e-01 -3.88801575e-01 -2.87190795e-01 -9.98145863e-02
9.46412861e-01 2.53990829e-01 -5.31525671e-01 -5.09338915e-01
-7.05228806e-01 8.83883312e-02 1.67546168e-01 3.28871697e-01
-6.38543725e-01 5.20927131e-01 7.18646193e+00 1.01038587e+00
-1.03745258e+00 -2.14145437e-01 8.25084865e-01 -7.69695342e-02
-1.15734875e+00 -2.72605330e-01 -8.49612176e-01 3.67738366e-01
3.39601159e-01 -7.95876682e-02 3.18908870e-01 5.03345609e-01
2.23062947e-01 -1.31515458e-01 -6.93285465e-01 1.14444077e+00
7.67044842e-01 -1.11037683e+00 8.08139563e-01 1.66599024e-02
1.48424387e+00 -6.04689181e-01 5.47534347e-01 -3.43648940e-02
2.09414914e-01 -1.12772071e+00 8.41302872e-01 7.55353570e-01
1.19041574e+00 -9.49355066e-01 3.81367296e-01 2.18657345e-01
-9.45829928e-01 2.47750372e-01 -2.10518137e-01 4.94900703e-01
4.33850765e-01 4.78033990e-01 -4.77631837e-01 1.22327697e+00
6.06527328e-01 4.61605757e-01 -6.15844727e-01 9.00027990e-01
-1.69216916e-01 1.05805181e-01 6.28635809e-02 2.41273299e-01
5.88693134e-02 -3.49056929e-01 6.99770510e-01 9.98263299e-01
9.23824131e-01 -5.43488972e-02 -1.20255709e-01 1.23488855e+00
1.70945391e-01 -1.22015499e-01 -8.25011671e-01 2.75812149e-01
2.41826355e-01 1.22879207e+00 -4.24187392e-01 -5.12052067e-02
-2.57631004e-01 1.23724639e+00 -1.64994135e-01 7.39828348e-01
-9.35565591e-01 -1.83294594e-01 3.91849428e-01 2.36341879e-02
6.11097872e-01 -9.01006162e-02 -1.95336580e-01 -1.14961410e+00
6.64832667e-02 -9.69033182e-01 1.52574152e-01 -1.58012533e+00
-1.32925880e+00 8.45406532e-01 2.58576274e-01 -1.59924126e+00
-5.01180589e-01 -3.33228946e-01 -4.34274793e-01 1.16251481e+00
-1.59214926e+00 -1.74528873e+00 -6.49345219e-01 8.06425989e-01
5.07943332e-01 -2.39652961e-01 7.03891516e-01 3.42757046e-01
-2.46169195e-01 5.27104616e-01 1.50057852e-01 -4.80720609e-01
8.07198346e-01 -1.09897542e+00 1.11663572e-01 1.02409172e+00
8.36597160e-02 4.90646183e-01 7.02927649e-01 -6.63679898e-01
-1.40713358e+00 -1.06631541e+00 5.98334491e-01 -2.64498711e-01
-3.59777331e-01 -1.14991300e-01 -4.42546159e-01 3.89224321e-01
6.42459273e-01 -4.05338109e-01 6.62406027e-01 -3.45122695e-01
-3.36674184e-01 -1.25510752e-01 -1.40667868e+00 7.51830339e-01
1.09443176e+00 -3.29901516e-01 -1.22404426e-01 -1.42115340e-01
5.61281025e-01 -6.62893832e-01 -7.45707750e-01 6.14950955e-01
4.46605563e-01 -1.30263364e+00 1.07836187e+00 -1.79931179e-01
6.70427382e-01 -6.92139864e-01 -4.19721037e-01 -1.64081454e+00
-7.15159595e-01 -6.96964622e-01 1.53442234e-01 1.36501384e+00
3.51619512e-01 -3.27179343e-01 4.00846571e-01 3.81285936e-01
-2.64508903e-01 -6.10782802e-01 -5.06944835e-01 -6.98589385e-01
-1.62393495e-01 -4.99582067e-02 9.04549003e-01 8.20451736e-01
-6.59218192e-01 3.51003677e-01 -8.42360914e-01 2.99856454e-01
7.50635147e-01 3.93243879e-01 7.24944890e-01 -7.01996446e-01
-2.99058914e-01 -1.86433926e-01 -3.70318592e-01 -1.02950835e+00
-2.38442898e-01 -6.60896719e-01 1.02642901e-01 -1.90786183e+00
5.18465877e-01 -2.63786823e-01 -1.32270738e-01 4.97321524e-02
-3.68305117e-01 5.11728227e-01 1.02051854e-01 3.58014763e-03
-3.83928806e-01 9.32411551e-01 1.93829143e+00 -3.18413228e-02
-8.96799320e-05 -3.39739770e-01 -9.19796586e-01 2.05630317e-01
5.94403148e-01 2.30719219e-03 -1.02576637e+00 -6.75855935e-01
2.31422618e-01 1.38491958e-01 4.21548635e-01 -1.02938306e+00
-8.73879045e-02 -2.82188863e-01 8.68689477e-01 -9.45225656e-01
4.54317361e-01 -6.78116441e-01 8.95268261e-01 -2.21460848e-03
-1.79182455e-01 1.45339876e-01 2.42138475e-01 6.43520951e-01
-3.36674869e-01 3.14161717e-03 8.37595940e-01 -3.93125266e-02
-7.54592717e-01 3.22785467e-01 -1.97179928e-01 -1.00852102e-01
9.71696854e-01 -5.03058434e-01 -5.67771971e-01 -8.54363203e-01
-3.72151881e-01 -1.55793488e-01 7.44192541e-01 5.52654684e-01
9.41526055e-01 -1.80856669e+00 -9.17262018e-01 3.95360082e-01
2.33865693e-01 -7.46285617e-02 9.69083250e-01 2.72798240e-01
-3.61062646e-01 2.26269633e-01 -5.65863669e-01 -6.93637311e-01
-1.13701284e+00 4.52020913e-01 2.59110481e-01 -6.78116158e-02
-3.83757502e-01 6.21631265e-01 8.27635825e-01 -3.97917718e-01
-2.15596721e-01 1.52826786e-01 -2.23320156e-01 -2.29457095e-01
4.06235099e-01 1.01260804e-02 -2.18152869e-02 -8.56612980e-01
5.25439419e-02 8.06197226e-01 1.14579253e-01 -1.82191715e-01
1.12576091e+00 -4.66056347e-01 3.29418391e-01 1.85574144e-01
9.04958367e-01 7.34005645e-02 -1.38702393e+00 -2.91413050e-02
-1.16307294e+00 -7.88741171e-01 2.07042228e-02 -1.11649585e+00
-1.36378253e+00 5.23183286e-01 7.09887445e-01 -7.05668554e-02
1.64021194e+00 -2.14172065e-01 7.07856596e-01 -2.04604581e-01
4.73284721e-01 -9.03598249e-01 2.56009489e-01 1.65393189e-01
1.03295970e+00 -9.04662728e-01 -6.89948425e-02 -5.80884814e-01
-7.96550751e-01 8.38733315e-01 8.28803658e-01 1.58689111e-01
1.44267097e-01 2.83207685e-01 1.54952243e-01 -1.68011889e-01
-8.23860943e-01 -2.05406129e-01 5.89041471e-01 1.22855234e+00
3.50828916e-01 -1.21753670e-01 -1.11164697e-01 2.80538470e-01
3.92440334e-02 -1.83081970e-01 5.19578338e-01 7.23246276e-01
-1.59181714e-01 -1.03833210e+00 -6.21015429e-01 2.10408509e-01
-2.66052634e-01 -7.22448528e-02 -3.63863587e-01 5.57353199e-01
2.70335406e-01 9.96809363e-01 -1.08673312e-01 -5.27647257e-01
3.50430876e-01 -1.43196955e-01 9.28437293e-01 -3.86403590e-01
-3.19227010e-01 2.96694309e-01 5.13394326e-02 -3.60975266e-01
-7.19126880e-01 -3.93043876e-01 -6.89879954e-01 -1.57236621e-01
-3.20004672e-01 -2.66174972e-01 5.75824797e-01 4.64788258e-01
7.00123429e-01 7.30717063e-01 8.51729453e-01 -1.04415059e+00
-2.32844815e-01 -8.17005873e-01 -7.36934841e-01 7.68209815e-01
7.37569258e-02 -7.45058656e-01 -3.14233243e-01 4.97764856e-01] | [11.543147087097168, -0.5757175087928772] |
82742bcd-f009-42eb-816d-8e3c245b9738 | box-of-lies-multimodal-deception-detection-in | null | null | https://aclanthology.org/N19-1175 | https://aclanthology.org/N19-1175.pdf | Box of Lies: Multimodal Deception Detection in Dialogues | Deception often takes place during everyday conversations, yet conversational dialogues remain largely unexplored by current work on automatic deception detection. In this paper, we address the task of detecting multimodal deceptive cues during conversational dialogues. We introduce a multimodal dataset containing deceptive conversations between participants playing the Box of Lies game from The Tonight Show Starring Jimmy Fallon, in which they try to guess whether an object description provided by their opponent is deceptive or not. We conduct annotations of multimodal communication behaviors, including facial and linguistic behaviors, and derive several learning features based on these annotations. Initial classification experiments show promising results, performing well above both a random and a human baseline, and reaching up to 69{\%} accuracy in distinguishing deceptive and truthful behaviors. | ["Ver{\\'o}nica P{\\'e}rez-Rosas", 'Felix Soldner', 'Rada Mihalcea'] | 2019-06-01 | null | null | null | naacl-2019-6 | ['deception-detection'] | ['miscellaneous'] | [ 1.10845000e-01 4.27297205e-02 1.06340267e-01 -7.35876024e-01
-1.00873280e+00 -1.01661050e+00 9.99748170e-01 -1.95448995e-01
-4.49762404e-01 9.37912941e-01 3.17467034e-01 -5.51304184e-02
5.28325379e-01 1.30909145e-01 -1.46936983e-01 -6.82711184e-01
1.58716068e-01 3.49478573e-01 -3.13584656e-01 -3.26173753e-01
5.64162195e-01 4.15239841e-01 -1.09405839e+00 9.97011185e-01
5.44191539e-01 1.13865793e+00 -9.31723356e-01 1.31754363e+00
2.78004318e-01 1.40002573e+00 -1.25217783e+00 -1.34776831e+00
-2.27176979e-01 -6.57695711e-01 -1.15391481e+00 1.89779192e-01
7.98559248e-01 -8.25472474e-01 -5.13536394e-01 1.07563674e+00
2.89833605e-01 2.33563796e-01 6.93759441e-01 -1.63417101e+00
-3.64564627e-01 1.47206679e-01 -3.83212596e-01 3.54726642e-01
1.01233578e+00 4.35931981e-01 7.28825331e-01 -7.25596666e-01
2.51192093e-01 1.41814280e+00 4.66238588e-01 1.25870168e+00
-1.10331798e+00 -8.29263151e-01 -2.52067298e-01 4.12607193e-01
-1.09094906e+00 -1.21919072e+00 7.32828736e-01 -4.09593493e-01
6.55906558e-01 6.22670233e-01 3.46961915e-01 1.95160675e+00
-7.41593316e-02 1.14279842e+00 1.16097486e+00 -7.77767599e-02
-1.01669095e-01 3.28285903e-01 2.82017589e-01 8.97445500e-01
-3.21141601e-01 -5.14078140e-02 -1.11829484e+00 -7.65516281e-01
-2.81582773e-03 -5.24985373e-01 -4.41895097e-01 1.21585369e-01
-8.26285899e-01 1.02032030e+00 -3.66702117e-02 -1.83723625e-02
-8.37545693e-02 -7.03448653e-02 7.81626105e-01 5.54447412e-01
4.73398209e-01 4.92503375e-01 2.02299535e-01 -9.33185101e-01
-6.00779831e-01 3.14251930e-01 1.24193323e+00 4.59282726e-01
1.43665284e-01 -5.09045310e-02 -3.17637771e-02 1.04264939e+00
-1.04230247e-01 4.35723275e-01 3.08175653e-01 -1.14068151e+00
5.13508260e-01 4.47520941e-01 4.70630288e-01 -1.47338486e+00
-3.92464101e-01 4.34603989e-01 -5.64061344e-01 1.32915527e-01
8.44883740e-01 -4.90730822e-01 -3.57272625e-01 1.34111142e+00
1.92882255e-01 -4.21713501e-01 1.99335873e-01 1.25072837e+00
9.24979627e-01 5.02202332e-01 5.06738089e-02 -1.57804281e-01
1.39197695e+00 -7.58630633e-01 -7.32516706e-01 -1.51375413e-01
5.10107458e-01 -6.96001232e-01 8.18845928e-01 7.96189904e-01
-1.15524709e+00 2.05321740e-02 -6.50895953e-01 -1.89920038e-01
-2.25492418e-01 -6.21527471e-02 5.67033768e-01 1.18019116e+00
-6.23280168e-01 2.49722466e-01 -4.65561956e-01 -2.34381378e-01
4.91059750e-01 1.53578162e-01 -6.76253319e-01 -1.52308702e-01
-1.17620409e+00 1.12590861e+00 -9.81952325e-02 4.43975210e-01
-1.27437449e+00 -1.17794842e-01 -1.04793823e+00 -2.70297438e-01
3.89649034e-01 -3.89897525e-01 1.55232811e+00 -1.45001304e+00
-1.74993336e+00 1.44074380e+00 -2.66115725e-01 -3.55114520e-01
8.88605356e-01 -1.42111868e-01 -6.99508667e-01 4.47234392e-01
-3.17875355e-01 5.09735763e-01 1.28757823e+00 -1.39016449e+00
-4.60248500e-01 -5.98861217e-01 4.25857842e-01 4.96139050e-01
-2.75817931e-01 4.54903185e-01 1.78464383e-01 -1.33910224e-01
-3.34316432e-01 -8.56969953e-01 6.18124783e-01 -6.18612431e-02
-7.78339803e-01 -2.21170634e-01 1.06231380e+00 -9.24576163e-01
9.07628477e-01 -2.21220040e+00 1.28772378e-01 4.68723796e-04
5.35404921e-01 4.22512025e-01 2.76056211e-02 5.10405123e-01
1.10193171e-01 1.42330825e-01 4.58039567e-02 -5.96678317e-01
3.33119839e-01 9.79916975e-02 -2.27005199e-01 7.36626267e-01
-7.79283419e-02 9.63266909e-01 -8.35048079e-01 -4.14564699e-01
-7.43324235e-02 9.21495333e-02 6.76336363e-02 4.69357669e-01
2.89855182e-01 5.15825450e-01 -2.02732921e-01 9.18040216e-01
5.90303361e-01 2.81117857e-01 5.01506440e-02 -1.92613900e-01
2.37402692e-01 3.03703994e-01 -2.06982210e-01 9.83762741e-01
-1.19152270e-01 1.14026296e+00 8.68431032e-01 -6.40915513e-01
6.12399518e-01 3.33501071e-01 -2.48631656e-01 -2.88714856e-01
5.68890333e-01 6.69170469e-02 8.77256393e-02 -8.44532907e-01
7.04434335e-01 -6.37442708e-01 -6.07366145e-01 5.18024445e-01
-1.43635541e-01 -2.67361850e-01 -1.05350830e-01 3.67960840e-01
1.13008535e+00 -4.50444639e-01 2.08023474e-01 2.18737602e-01
8.62080097e-01 1.21903017e-01 -8.84828996e-03 7.22978771e-01
-9.84657168e-01 4.48589593e-01 1.15467882e+00 -5.47031105e-01
-4.42927510e-01 -8.83250475e-01 3.18161339e-01 1.41581476e+00
2.29169831e-01 -9.32650343e-02 -8.31192672e-01 -1.24788797e+00
1.35146491e-02 8.98528695e-01 -7.26253211e-01 -5.46596467e-01
-3.55634153e-01 -5.29206932e-01 1.33407974e+00 -4.31695059e-02
6.98804140e-01 -8.29005420e-01 -3.71948957e-01 -5.50066113e-01
-7.24070787e-01 -1.20161080e+00 -7.32098877e-01 -3.99181932e-01
-2.62875289e-01 -1.34623766e+00 -4.83093321e-01 -3.65745395e-01
4.51367557e-01 1.97942987e-01 1.03479242e+00 3.12874645e-01
-2.14637011e-01 6.97114706e-01 -2.81990737e-01 -2.38880992e-01
-7.84218729e-01 -2.53795266e-01 5.60596511e-02 4.15757090e-01
4.17674094e-01 -2.09871400e-02 -3.96552980e-01 4.71017092e-01
-5.98548770e-01 -2.40786985e-01 -8.65654554e-03 9.31690693e-01
-7.15539157e-01 -3.67269516e-01 2.95792848e-01 -4.79608417e-01
1.24857628e+00 -3.35083812e-01 -1.56735443e-02 2.79991496e-02
2.41949722e-01 -6.03078425e-01 5.11004925e-01 -3.48144591e-01
-1.24614525e+00 -2.63464302e-01 9.42340773e-03 -2.51341462e-01
-5.73368013e-01 -1.44875556e-01 -8.95798132e-02 -4.24083501e-01
5.27496159e-01 3.79787654e-01 2.63511002e-01 -8.10880065e-02
9.65864062e-02 1.08604681e+00 5.60159147e-01 -7.00069904e-01
5.51216245e-01 4.35813963e-01 -3.52233648e-01 -1.17792654e+00
-9.66750979e-01 -4.47770834e-01 -5.26095986e-01 -6.28310084e-01
6.47508085e-01 -4.15889144e-01 -1.81115210e+00 8.34432185e-01
-1.42734241e+00 2.50626113e-02 5.27932048e-01 1.19609497e-02
-4.25516754e-01 8.23154449e-01 -9.54553783e-01 -1.34649181e+00
-1.13558047e-01 -8.73844922e-01 1.05611610e+00 -1.05863668e-01
-7.26436138e-01 -9.47182298e-01 -8.02383721e-02 1.38181841e+00
1.49682119e-01 2.74214357e-01 7.96999574e-01 -1.09729099e+00
9.36013907e-02 -4.87693936e-01 -2.27946296e-01 4.37532395e-01
3.61119211e-02 -2.20107257e-01 -1.01401818e+00 -1.86505690e-01
3.49558033e-02 -1.18056226e+00 7.06066251e-01 -3.80378664e-01
1.01763546e+00 -7.33370304e-01 -1.17360383e-01 3.40816081e-02
4.26362276e-01 1.18920572e-01 6.08969808e-01 -1.52165323e-01
5.23627102e-01 1.08364964e+00 2.71970958e-01 4.76279408e-01
4.21900809e-01 6.65161967e-01 7.20816433e-01 3.84684324e-01
3.48192841e-01 -9.49987620e-02 6.75914168e-01 -5.35131730e-02
4.12940159e-02 -6.16820335e-01 -8.67909551e-01 3.96788776e-01
-1.56730878e+00 -1.27634263e+00 -3.08204949e-01 1.93413520e+00
8.51469219e-01 -2.01483697e-01 4.98060852e-01 -5.51592335e-02
6.99510932e-01 3.69503081e-01 -4.12904412e-01 -1.08361256e+00
-2.81540826e-02 -5.10009468e-01 -9.61546600e-02 8.68681788e-01
-1.26657808e+00 1.00496352e+00 6.73388004e+00 5.95968425e-01
-8.69114697e-01 7.98585191e-02 9.64324117e-01 -4.40212458e-01
1.98144868e-01 -6.23895526e-01 -4.27164704e-01 4.46019918e-01
9.80339170e-01 2.06436932e-01 6.06351376e-01 4.69777942e-01
2.00568840e-01 -4.88388926e-01 -1.39211690e+00 1.37050736e+00
1.03418863e+00 -7.57779002e-01 -9.74490121e-02 -9.35018584e-02
2.68140435e-01 -7.09862411e-01 1.83881730e-01 4.08348948e-01
2.63393581e-01 -1.57921064e+00 4.04334098e-01 5.57155132e-01
3.64507318e-01 -8.74409020e-01 9.30335939e-01 7.12980986e-01
-6.75048456e-02 -5.59937805e-02 2.24269316e-01 -2.20174685e-01
2.18276769e-01 -1.38905481e-01 -1.08799720e+00 -2.79959664e-02
4.51446444e-01 3.48511040e-01 -3.72357339e-01 4.06584680e-01
-1.57691807e-01 5.59786618e-01 1.53514162e-01 -3.70049387e-01
3.56695831e-01 3.28183931e-04 8.49171162e-01 1.35026646e+00
-3.67802054e-01 3.75708193e-01 -6.91706315e-02 7.77458310e-01
-4.53818947e-01 -2.96746373e-01 -6.04447484e-01 -3.39461952e-01
-7.24026337e-02 1.39581025e+00 4.93395217e-02 -2.68291295e-01
-7.01163942e-03 1.73870873e+00 2.86180228e-01 1.90422297e-01
-9.64896917e-01 -2.46130124e-01 1.01248729e+00 -3.28940779e-01
-3.33426446e-01 -9.19033960e-02 -7.25753680e-02 -1.24080729e+00
-2.25050692e-02 -1.30355942e+00 4.93059844e-01 -8.21395278e-01
-1.71859229e+00 4.80467409e-01 -1.64000884e-01 -6.61314666e-01
-5.07839441e-01 -6.44337475e-01 -6.15955055e-01 7.30981290e-01
-9.58384097e-01 -1.21860564e+00 -5.07304847e-01 6.49008214e-01
6.23829484e-01 -2.20667422e-01 8.66878748e-01 -1.75105274e-01
-7.27090597e-01 7.94421613e-01 -3.88493121e-01 5.08753538e-01
8.52569938e-01 -9.95493948e-01 -1.75656572e-01 2.94572353e-01
-1.55494586e-01 5.10366321e-01 8.70621204e-01 -3.34193140e-01
-1.37181997e+00 -2.86486536e-01 8.03472161e-01 -1.04603899e+00
8.01231086e-01 -5.89611113e-01 -8.61920655e-01 4.60928380e-01
4.68759745e-01 -2.83317566e-01 9.18475091e-01 6.56305403e-02
-7.40778804e-01 2.75036842e-01 -1.60752094e+00 5.99737465e-01
8.14302087e-01 -8.74721885e-01 -7.79393494e-01 3.65098447e-01
-2.23096788e-01 -4.30142671e-01 -4.02433962e-01 -9.18600410e-02
9.83307838e-01 -1.35451972e+00 6.71522617e-01 -1.31504107e+00
5.64211786e-01 3.17382991e-01 -4.58152592e-02 -1.39238322e+00
4.16081458e-01 -9.14200604e-01 -9.65355784e-02 1.13974535e+00
2.11593971e-01 -4.83602941e-01 1.00744307e+00 1.15771699e+00
1.95190549e-01 -1.96354195e-01 -1.26349437e+00 -2.93059736e-01
6.71926066e-02 -1.69542477e-01 -3.84361520e-02 1.18598616e+00
7.66771972e-01 5.83258748e-01 -9.05718327e-01 -1.01022273e-01
6.12536013e-01 -1.35009438e-01 9.33089912e-01 -8.12786341e-01
1.16202794e-01 -6.18148208e-01 -4.35647726e-01 -1.20733154e+00
8.68262649e-01 -4.62697446e-01 1.44767895e-01 -7.80598223e-01
5.34753561e-01 3.92572075e-01 3.94589514e-01 4.88940299e-01
-3.44104581e-02 6.06207788e-01 1.32924572e-01 1.37675926e-01
-9.89411235e-01 6.64283037e-01 1.06591427e+00 -4.33429450e-01
1.82158828e-01 1.41699046e-01 -6.18276894e-01 1.02232659e+00
7.55416989e-01 -1.77173227e-01 1.51920348e-01 -1.45508721e-01
-2.08532393e-01 4.84948128e-01 8.39001358e-01 -2.91346520e-01
2.15288371e-01 -1.79137498e-01 3.67964149e-01 -2.23807573e-01
1.14885068e+00 -6.28235817e-01 -6.02127075e-01 2.22890005e-01
-6.93724811e-01 -1.07580699e-01 2.43849322e-01 6.05762005e-01
-3.07227761e-01 -3.17565233e-01 8.80676270e-01 -1.90535322e-01
-5.28768778e-01 -2.96820968e-01 -8.96980166e-01 1.53054491e-01
8.86105120e-01 -2.07623363e-01 -8.48769069e-01 -1.40473521e+00
-8.64037097e-01 2.40046203e-01 3.14619809e-01 5.18388093e-01
8.71880949e-01 -9.94038105e-01 -8.80325615e-01 -1.65997848e-01
2.61528879e-01 -1.00976980e+00 3.17767441e-01 1.00115931e+00
-3.40959847e-01 2.14936554e-01 -1.27756476e-01 -3.04009020e-01
-1.98773360e+00 4.15557064e-02 8.16389918e-01 2.38113925e-01
4.04827707e-02 9.76154804e-01 1.78317994e-01 -4.99964625e-01
2.37995192e-01 5.10771751e-01 -1.12351701e-01 9.54028144e-02
8.29047799e-01 8.26852560e-01 5.70860319e-03 -1.24465060e+00
-5.37016928e-01 -2.01468706e-01 -1.93640217e-01 -2.86590219e-01
5.25640905e-01 -3.03734779e-01 -3.21987033e-01 5.24567485e-01
1.43689847e+00 1.35896370e-01 -7.86450863e-01 3.20666224e-01
-8.73447433e-02 -9.54221249e-01 -3.54305148e-01 -1.38328886e+00
-5.28067052e-01 9.11597490e-01 4.48685698e-02 5.71155667e-01
6.23302937e-01 6.59056231e-02 8.38747203e-01 5.00151217e-01
2.39224747e-01 -7.78687894e-01 6.34360075e-01 6.50989473e-01
1.32049787e+00 -1.75089037e+00 -1.45724788e-01 -2.80786812e-01
-1.27407241e+00 1.27982664e+00 7.17494547e-01 1.50721475e-01
-4.87114601e-02 -1.92446411e-01 3.04112166e-01 -3.78455967e-01
-7.75950372e-01 2.21182406e-01 4.66518909e-01 6.97106421e-01
2.42087796e-01 1.10442191e-01 -3.87390666e-02 5.27018011e-01
-3.38102579e-01 -6.41716540e-01 6.23302937e-01 6.27570212e-01
-3.59749407e-01 -7.08589017e-01 -6.31990075e-01 1.53695434e-01
-6.29415810e-01 1.78562179e-02 -1.79731488e+00 8.03793669e-01
-3.22375834e-01 1.62918568e+00 -1.07110873e-01 -5.32015741e-01
3.84466171e-01 4.16822076e-01 4.87335920e-01 -2.26069734e-01
-1.02931809e+00 -4.18721795e-01 9.69809890e-01 -7.37002730e-01
-1.30378425e-01 -8.22778463e-01 -6.18313849e-01 -8.76189828e-01
-8.11181068e-02 2.41350949e-01 4.58155990e-01 9.93749738e-01
-6.23240992e-02 -4.33777779e-01 6.96910262e-01 -8.59323859e-01
-7.65257835e-01 -1.07775843e+00 -3.44145983e-01 7.86708534e-01
8.40405464e-01 -3.36213559e-01 -8.51925135e-01 -2.01915041e-01] | [8.234585762023926, 10.382684707641602] |
3be66e0c-b2de-4ea3-94fe-752fb5d0f244 | ground-plane-based-absolute-scale-estimation | 1903.00912 | null | http://arxiv.org/abs/1903.00912v1 | http://arxiv.org/pdf/1903.00912v1.pdf | Ground Plane based Absolute Scale Estimation for Monocular Visual Odometry | Recovering the absolute metric scale from a monocular camera is a challenging
but highly desirable problem for monocular camera-based systems. By using
different kinds of cues, various approaches have been proposed for scale
estimation, such as camera height, object size etc. In this paper, firstly, we
summarize different kinds of scale estimation approaches. Then, we propose a
robust divide and conquer the absolute scale estimation method based on the
ground plane and camera height by analyzing the advantages and disadvantages of
different approaches. By using the estimated scale, an effective scale
correction strategy has been proposed to reduce the scale drift during the
Monocular Visual Odometry (VO) estimation process. Finally, the effectiveness
and robustness of the proposed method have been verified on both public and
self-collected image sequences. | ['Yuchao Dai', 'Dingfu Zhou', 'Hongdong Li'] | 2019-03-03 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-1.25958771e-01 -5.13658881e-01 -1.18241541e-01 -3.53657871e-01
-1.74465418e-01 -4.19357300e-01 4.21579659e-01 -2.49495089e-01
-4.11559224e-01 6.08336151e-01 6.40743747e-02 2.04994172e-01
-1.68197360e-02 -7.12776542e-01 -5.69250703e-01 -4.98048156e-01
5.65795600e-01 1.58107415e-01 5.92737615e-01 5.10789678e-02
6.52978897e-01 6.23128712e-01 -1.45937133e+00 -6.42737508e-01
7.36734807e-01 1.04241931e+00 4.00761008e-01 6.59480214e-01
2.10062191e-01 4.74742621e-01 -6.18368447e-01 -1.54731959e-01
6.34002745e-01 -3.79139572e-01 -1.87446907e-01 5.73019266e-01
6.10983074e-01 -5.79029083e-01 -3.75032842e-01 1.38117802e+00
6.49312854e-01 -1.97430268e-01 5.01915932e-01 -1.07231975e+00
-2.98935711e-01 -5.47328517e-02 -8.41961205e-01 1.12463452e-01
6.00865006e-01 -9.35901925e-02 5.93014896e-01 -8.48321259e-01
7.25095093e-01 1.06173825e+00 6.45579159e-01 4.53866757e-02
-7.19557166e-01 -5.91556549e-01 -1.51575774e-01 3.90315562e-01
-1.67037940e+00 -3.09139460e-01 8.86872709e-01 -5.28441668e-01
5.11126697e-01 -1.10085554e-01 6.90181911e-01 9.98030379e-02
3.52532715e-01 2.19713539e-01 1.32107472e+00 -5.10194600e-01
-1.63568817e-02 2.46990159e-01 -1.06841631e-01 6.78003848e-01
8.27399850e-01 3.53593826e-02 -2.97842443e-01 1.37738446e-02
1.14300263e+00 3.46425213e-02 -3.83823454e-01 -9.54740465e-01
-1.26070333e+00 5.29830575e-01 4.87463415e-01 9.12313759e-02
-1.36734754e-01 -9.59646627e-02 9.66794342e-02 7.98071027e-02
2.09299102e-01 2.94021159e-01 -3.70785058e-01 -2.53373478e-02
-8.04295182e-01 -5.42979069e-06 3.53796929e-01 1.40817177e+00
8.64595950e-01 4.62102108e-02 4.69915569e-01 6.70089126e-01
3.80658507e-01 9.68348742e-01 5.21253943e-01 -7.32892394e-01
5.14943898e-01 7.81401336e-01 3.62887263e-01 -1.38260400e+00
-5.68341553e-01 -9.40975696e-02 -5.63718021e-01 2.98461795e-01
1.58178508e-01 -2.23049477e-01 -7.07375526e-01 1.17808962e+00
5.32217324e-01 2.37467512e-02 -1.25470415e-01 1.13937521e+00
7.54218221e-01 2.67354548e-01 -5.16248465e-01 -3.25680137e-01
1.32333100e+00 -7.64833212e-01 -7.69486129e-01 -3.25449705e-01
3.10125619e-01 -1.24153507e+00 6.77366376e-01 3.66920173e-01
-9.27046061e-01 -7.06819355e-01 -1.45516598e+00 -1.62194967e-01
-1.35844007e-01 5.84048092e-01 2.01673612e-01 6.29497290e-01
-7.68828511e-01 3.29729050e-01 -7.60489404e-01 -6.73276424e-01
-3.98367912e-01 3.17127436e-01 -3.03979158e-01 6.06376901e-02
-7.67669857e-01 1.12805271e+00 3.54033947e-01 -1.36019243e-02
-1.60729095e-01 -4.49089296e-02 -8.15864027e-01 -2.69811302e-01
3.68156344e-01 -6.34192169e-01 1.02525294e+00 -6.23357713e-01
-1.64038110e+00 9.87560451e-01 -2.58319825e-01 -2.23168373e-01
5.82324922e-01 -3.43221933e-01 -2.97376126e-01 2.82878876e-01
-3.04899830e-03 4.43736702e-01 6.53771996e-01 -9.19107556e-01
-7.10467219e-01 -5.94068885e-01 -1.14014344e-02 7.33177304e-01
1.18965834e-01 2.71815583e-02 -5.57949007e-01 -2.74474025e-01
9.21877861e-01 -1.10820794e+00 2.41509043e-02 2.01866895e-01
-9.75890830e-02 2.82307535e-01 7.30448484e-01 -8.19963992e-01
1.20569944e+00 -2.10673952e+00 2.02636689e-01 -1.30015358e-01
8.11493993e-02 7.46050552e-02 5.04528046e-01 2.91948259e-01
3.70031893e-01 -3.84090155e-01 1.15131535e-01 7.06752539e-02
-4.18589354e-01 -7.58703277e-02 1.08161397e-01 8.30682993e-01
-2.67480850e-01 4.53748435e-01 -5.34098089e-01 -5.97035170e-01
6.72676921e-01 3.31874371e-01 -3.04778755e-01 2.72293031e-01
3.47235233e-01 3.96965027e-01 -9.64819193e-02 8.02177489e-01
1.12493908e+00 5.90966269e-02 1.18993379e-01 -5.08735538e-01
-5.58075368e-01 1.28607869e-01 -1.73856711e+00 1.38427138e+00
-3.48970711e-01 7.06939101e-01 5.95688596e-02 -4.72459376e-01
1.30522525e+00 2.03839019e-01 3.24068725e-01 -4.79952574e-01
4.33206379e-01 5.41343272e-01 -1.32188559e-01 -3.63044947e-01
7.99854696e-01 -7.00446069e-02 2.62047678e-01 -2.50730842e-01
-2.45838925e-01 -7.18746841e-01 2.21075490e-01 -2.66564101e-01
5.10107458e-01 3.14213961e-01 1.09822559e+00 -2.21382827e-01
9.65237796e-01 2.79924572e-02 7.13328779e-01 6.45690486e-02
-5.12781918e-01 7.35152900e-01 1.18324058e-02 -4.45329517e-01
-1.35587549e+00 -8.04585159e-01 -2.35405430e-01 1.29794432e-02
1.00193691e+00 -1.89775825e-01 -5.57517290e-01 -2.24229977e-01
1.44747436e-01 -9.30144414e-02 -1.77790701e-01 9.29762870e-02
-6.49220526e-01 -4.11160767e-01 4.21862975e-02 5.59764326e-01
9.95017469e-01 -4.87307012e-01 -8.10938656e-01 6.14860915e-02
-2.95011550e-01 -1.40479457e+00 -4.87610310e-01 -2.57089138e-01
-1.00833857e+00 -1.19893467e+00 -7.54525483e-01 -7.24389672e-01
6.81059003e-01 1.14759231e+00 6.30373061e-01 -6.29325956e-02
-6.96266256e-03 1.46822199e-01 -1.16450824e-01 -3.77675056e-01
3.22592892e-02 -4.49043289e-02 3.18805069e-01 -1.02289677e-01
3.96928966e-01 -4.52451438e-01 -8.12171578e-01 8.22913170e-01
-3.51781696e-01 1.79251269e-01 6.17693305e-01 4.33780491e-01
5.06169081e-01 -8.81017968e-02 -3.71545349e-04 -3.05961490e-01
5.64280860e-02 3.27417664e-02 -1.32412016e+00 -5.12157455e-02
-7.68798709e-01 -1.78107128e-01 3.40908110e-01 -3.27968031e-01
-1.03229702e+00 1.83568925e-01 2.31896505e-01 -2.58052677e-01
-3.09494473e-02 9.07623768e-02 -2.78801858e-01 -4.51920539e-01
5.11028528e-01 1.19429164e-01 -4.59642261e-02 -4.09880191e-01
2.09195882e-01 9.84697819e-01 5.16994774e-01 -4.83714119e-02
1.05057180e+00 8.69721174e-01 2.75494814e-01 -1.00370562e+00
-6.35252714e-01 -9.10027802e-01 -1.11737573e+00 -2.78932333e-01
1.02941346e+00 -1.32619095e+00 -6.36308551e-01 7.69210696e-01
-1.12526989e+00 2.80607998e-01 1.75739035e-01 1.10308242e+00
-5.45025408e-01 7.25021780e-01 -5.24508476e-01 -6.66584492e-01
-1.77642360e-01 -1.18839991e+00 1.14937079e+00 6.64966702e-01
1.21234521e-01 -6.99159205e-01 1.70434147e-01 2.94393986e-01
2.33054578e-01 1.64356843e-01 2.88104981e-01 2.85808861e-01
-9.60603952e-01 -2.46797130e-01 -5.18879414e-01 2.14745030e-01
3.55282485e-01 3.61876637e-02 -5.67131698e-01 -4.04548645e-01
1.62345126e-01 1.93547443e-01 2.10362539e-01 4.33773696e-01
1.72552645e-01 8.72410387e-02 -1.65707886e-01 9.90797758e-01
1.85664892e+00 4.94695067e-01 6.39734209e-01 7.65734017e-01
7.46453822e-01 2.03657180e-01 1.06234539e+00 5.50135076e-01
3.91929269e-01 1.00287175e+00 3.17626595e-01 1.25891998e-01
-1.21816225e-01 -2.61525154e-01 4.02420700e-01 1.14825583e+00
-1.70285106e-01 3.09532732e-01 -5.13277531e-01 2.58078307e-01
-1.67801774e+00 -8.12199295e-01 -3.85045409e-01 2.41427398e+00
4.72023547e-01 4.82059047e-02 -4.93857339e-02 8.64647478e-02
9.24459934e-01 1.96718559e-01 -4.55552220e-01 -2.63767242e-01
-2.61511475e-01 -4.51310277e-01 9.97628510e-01 5.94700515e-01
-1.07002747e+00 8.66812646e-01 6.22616386e+00 2.01493412e-01
-1.50237417e+00 -1.66481972e-01 -2.43541569e-01 2.61576235e-01
2.51483977e-01 2.98153698e-01 -1.24744260e+00 3.01616102e-01
2.96861410e-01 -3.60535920e-01 4.34798151e-01 1.17681420e+00
1.68531522e-01 -6.12482548e-01 -5.33129811e-01 1.32682574e+00
4.75582153e-01 -7.74276197e-01 -3.75400305e-01 8.68264809e-02
8.61110389e-01 -1.03772329e-02 -4.05411690e-01 -1.95403725e-01
-1.24111153e-01 -1.51486501e-01 7.60969877e-01 2.83322722e-01
7.52654135e-01 -4.75072294e-01 8.72653782e-01 4.71829385e-01
-1.64019299e+00 3.44836377e-02 -7.90871561e-01 -4.83310819e-01
2.31598780e-01 6.03255630e-01 -7.56790578e-01 7.72843361e-01
5.12781382e-01 6.78532302e-01 -8.64427745e-01 1.37642765e+00
-4.83048767e-01 -8.08585510e-02 -3.07570577e-01 9.73985195e-02
-2.58392572e-01 -6.74498498e-01 6.15712523e-01 7.42394447e-01
6.63350105e-01 9.20900777e-02 3.33778225e-02 4.25633788e-01
1.93472534e-01 3.40098619e-01 -7.46687353e-01 4.39768672e-01
5.49419522e-01 1.20535159e+00 -6.34419084e-01 -4.03885394e-01
-7.19372392e-01 7.76697934e-01 -6.23016544e-02 7.65383467e-02
-9.19438720e-01 -5.25411487e-01 3.85124654e-01 2.58831590e-01
2.45485857e-01 -7.36257732e-01 -4.48178738e-01 -1.54901433e+00
3.16026419e-01 -6.47558689e-01 1.01446025e-01 -1.20851171e+00
-4.15220678e-01 4.43232000e-01 8.44844356e-02 -1.89099789e+00
-3.83303547e-03 -6.43857181e-01 -2.68215984e-01 6.33778155e-01
-1.49910402e+00 -8.61001670e-01 -8.46221328e-01 3.15113783e-01
5.80820084e-01 2.55660191e-02 1.95241690e-01 3.56957674e-01
-2.82583654e-01 1.96106985e-01 8.13833773e-02 -2.15978716e-02
9.87312257e-01 -9.85208809e-01 1.36704102e-01 1.14206326e+00
-1.59031332e-01 6.23791993e-01 8.08026552e-01 -7.30192840e-01
-1.36426151e+00 -5.75785697e-01 8.66254866e-01 -2.62025356e-01
4.64851975e-01 -6.74306676e-02 -5.40420830e-01 7.21105158e-01
1.67675093e-01 -9.03392136e-02 -2.06644639e-01 -2.89920926e-01
-1.53969094e-01 -2.83254653e-01 -1.08085179e+00 5.09962618e-01
9.56510484e-01 -3.39852661e-01 -6.42709851e-01 -7.92729631e-02
5.42757988e-01 -8.33553612e-01 -6.76073611e-01 6.16967559e-01
7.03294098e-01 -1.19612837e+00 8.33319306e-01 3.28375995e-01
1.72960684e-02 -8.64773273e-01 -3.14650297e-01 -1.05455363e+00
-3.26776236e-01 -2.97060281e-01 1.87316015e-01 1.17056644e+00
-2.38980502e-01 -7.71047115e-01 8.23308110e-01 2.47352064e-01
8.40422064e-02 -2.73806900e-01 -7.57529378e-01 -8.37512434e-01
-5.49073935e-01 9.96356085e-02 4.31541681e-01 6.97899044e-01
-1.15929194e-01 4.73469734e-01 -6.13491297e-01 5.45974910e-01
6.89713597e-01 2.80552089e-01 1.41540313e+00 -1.30196631e+00
-1.09478123e-01 3.12489234e-02 -9.40813005e-01 -1.40832317e+00
-5.41009426e-01 -1.94240764e-01 -1.02596059e-01 -1.43589020e+00
2.34490842e-01 5.12186550e-02 1.39125749e-01 -4.92782503e-01
-1.50755912e-01 2.16223851e-01 3.70630682e-01 5.19820750e-01
-2.07427830e-01 2.85262853e-01 1.23088932e+00 2.97440767e-01
-1.84688255e-01 1.11278892e-01 -1.95318341e-01 1.06492126e+00
6.93038821e-01 -3.06227863e-01 -3.04816127e-01 -4.46510464e-01
3.35236400e-01 3.25996131e-01 1.26262888e-01 -1.32155955e+00
2.24452436e-01 -3.59472603e-01 2.99720526e-01 -1.12153649e+00
3.74134541e-01 -9.64672267e-01 1.28214553e-01 6.42473459e-01
2.80929178e-01 5.17324328e-01 -1.58656731e-01 2.67689466e-01
-3.53950351e-01 -2.33872697e-01 1.17523444e+00 -1.24733232e-01
-8.53470862e-01 4.56170253e-02 1.25564963e-01 -2.26692051e-01
1.11907530e+00 -5.08144915e-01 -3.31264704e-01 -4.52885181e-01
-3.17844510e-01 -6.99746907e-02 9.69435930e-01 3.13855797e-01
5.87482989e-01 -1.56113660e+00 -3.58851463e-01 2.16330662e-01
3.21143955e-01 -1.95744082e-01 1.34685904e-01 1.08905745e+00
-1.06071043e+00 6.29086792e-01 -3.69052052e-01 -7.94366181e-01
-1.66663206e+00 7.76022851e-01 3.14791828e-01 4.09738906e-02
-3.70407194e-01 2.54486024e-01 2.44671509e-01 -4.71629590e-01
-2.53263898e-02 -5.30750692e-01 -2.08549052e-01 -1.82934269e-01
3.33040833e-01 7.31408656e-01 -4.08738852e-02 -1.04815316e+00
-3.63341361e-01 1.48614585e+00 3.73536229e-01 -2.88443744e-01
8.15260470e-01 -8.92683744e-01 4.06218171e-02 4.11977053e-01
1.07081747e+00 4.15930361e-01 -1.25602484e+00 -3.21837455e-01
8.32906272e-03 -9.86040175e-01 -2.22901478e-01 -1.76526904e-01
-7.55920231e-01 8.67509127e-01 7.42979407e-01 -1.25975460e-01
1.07574916e+00 -3.94846261e-01 6.21242404e-01 2.02316642e-01
7.73951650e-01 -1.22883344e+00 -2.31708914e-01 3.08451593e-01
7.24548519e-01 -1.40275562e+00 6.19964361e-01 -7.93128312e-01
-4.25720215e-01 1.26944613e+00 7.82026410e-01 -1.65467665e-01
5.07199168e-01 5.93931861e-02 3.02921355e-01 1.50210038e-01
-2.36988977e-01 -3.36260408e-01 2.21919551e-01 4.28350836e-01
3.66570503e-01 5.43809719e-02 -8.85497272e-01 -1.86025783e-01
-2.06864923e-01 2.03182399e-02 8.91848087e-01 9.99859631e-01
-9.02932048e-01 -8.96006703e-01 -9.18214262e-01 -1.19931027e-01
-6.65914193e-02 2.02783510e-01 -3.66259933e-01 9.65727985e-01
1.72734365e-01 8.75366509e-01 -1.22865811e-01 -4.38313335e-01
6.46672904e-01 -2.84855515e-01 6.07338607e-01 -3.73615533e-01
8.83779377e-02 2.36243188e-01 -1.20612189e-01 -3.52077931e-01
-5.50757766e-01 -6.21066153e-01 -1.14631522e+00 -2.67557293e-01
-7.53171444e-01 -1.62132949e-01 9.24109697e-01 6.58412337e-01
1.97732270e-01 -1.21106155e-01 9.23906684e-01 -8.50724161e-01
-6.41766727e-01 -8.79643500e-01 -7.04894125e-01 3.93360078e-01
1.32638872e-01 -8.49465668e-01 -5.73518932e-01 -4.90790717e-02] | [7.84186315536499, -2.1521358489990234] |
8d7cff07-248a-4dc7-ad91-f2187b8287c3 | improving-the-results-of-string-kernels-in | 1808.08409 | null | http://arxiv.org/abs/1808.08409v2 | http://arxiv.org/pdf/1808.08409v2.pdf | Improving the results of string kernels in sentiment analysis and Arabic dialect identification by adapting them to your test set | Recently, string kernels have obtained state-of-the-art results in various
text classification tasks such as Arabic dialect identification or native
language identification. In this paper, we apply two simple yet effective
transductive learning approaches to further improve the results of string
kernels. The first approach is based on interpreting the pairwise string kernel
similarities between samples in the training set and samples in the test set as
features. Our second approach is a simple self-training method based on two
learning iterations. In the first iteration, a classifier is trained on the
training set and tested on the test set, as usual. In the second iteration, a
number of test samples (to which the classifier associated higher confidence
scores) are added to the training set for another round of training. However,
the ground-truth labels of the added test samples are not necessary. Instead,
we use the labels predicted by the classifier in the first training iteration.
By adapting string kernels to the test set, we report significantly better
accuracy rates in English polarity classification and Arabic dialect
identification. | ['Andrei M. Butnaru', 'Radu Tudor Ionescu'] | 2018-08-25 | improving-the-results-of-string-kernels-in-1 | https://aclanthology.org/D18-1135 | https://aclanthology.org/D18-1135.pdf | emnlp-2018-10 | ['native-language-identification'] | ['natural-language-processing'] | [ 3.44874352e-01 9.93540883e-02 -1.44188687e-01 -6.90486073e-01
-7.24858046e-01 -8.90672207e-01 6.26497269e-01 3.84861678e-01
-4.95513469e-01 6.74383521e-01 -5.01820147e-02 -3.71353954e-01
1.89618036e-01 -9.08626020e-01 -4.29688096e-01 -9.06301320e-01
1.33486718e-01 7.86535382e-01 4.70895380e-01 -4.35934365e-01
3.05586666e-01 3.08588415e-01 -1.34139025e+00 5.66213489e-01
1.24271441e+00 1.00845051e+00 -3.20151418e-01 7.00429261e-01
-5.24052829e-02 7.71050215e-01 -2.92524129e-01 -8.43060434e-01
7.37471208e-02 -7.88765967e-01 -1.01899981e+00 3.26821879e-02
1.91731960e-01 2.67487187e-02 -1.49507374e-02 1.18764377e+00
2.46743888e-01 4.16326560e-02 8.95794809e-01 -9.80459154e-01
-4.89062101e-01 8.05025399e-01 -2.61926770e-01 -3.84156644e-01
5.94337225e-01 -2.16812536e-01 1.01312971e+00 -1.06593740e+00
2.78902709e-01 8.42687964e-01 7.49616325e-01 4.08120900e-01
-1.22938848e+00 -4.61907029e-01 -8.48736167e-02 2.21526906e-01
-1.03280425e+00 -5.07053971e-01 9.00461376e-01 -3.33220482e-01
5.19495487e-01 4.10962701e-01 6.95990801e-01 5.95128357e-01
-4.12375361e-01 6.56354249e-01 1.36926973e+00 -9.19335425e-01
4.83795196e-01 7.93809712e-01 6.60986960e-01 8.34720492e-01
-6.91182241e-02 -2.20142379e-02 -1.98137209e-01 -3.92111361e-01
2.29159743e-01 -6.29588664e-01 -2.15440005e-01 -4.08690810e-01
-1.01368546e+00 1.03510177e+00 3.59010816e-01 3.88593823e-01
-2.08417445e-01 -6.62236691e-01 5.19382000e-01 4.87053394e-01
6.69508874e-01 4.65220183e-01 -5.10183871e-01 -2.20951848e-02
-7.85588682e-01 -6.60798997e-02 1.14994419e+00 3.38471264e-01
9.45762396e-01 -4.15320367e-01 1.91422775e-01 1.20077872e+00
1.68312073e-01 3.63637239e-01 6.16546512e-01 -1.85387418e-01
3.15620214e-01 9.41150367e-01 -4.57044393e-02 -7.04412401e-01
-1.56117216e-01 4.72292751e-02 -5.32539487e-01 1.51279747e-01
9.17719185e-01 -2.72159189e-01 -8.24042857e-01 1.53090942e+00
4.24718320e-01 -3.79663594e-02 3.96555811e-01 7.29440570e-01
5.52143455e-01 5.95962822e-01 -2.96914846e-01 -1.99732304e-01
1.14055049e+00 -8.82913649e-01 -7.69346580e-02 1.12682935e-02
1.09390163e+00 -1.04493296e+00 9.54044819e-01 3.97396445e-01
-7.76205719e-01 -3.58631611e-01 -1.05994201e+00 3.86095256e-01
-6.13480628e-01 5.62853932e-01 6.21111631e-01 1.01295996e+00
-9.83029902e-01 5.18178165e-01 -2.47739032e-01 -5.41779101e-01
1.55947089e-01 5.24315774e-01 -5.61863065e-01 -1.77461371e-01
-1.30156541e+00 8.89838398e-01 4.15064454e-01 5.81114646e-03
-1.78463891e-01 -1.97264314e-01 -8.07479560e-01 -1.74685076e-01
-1.63556151e-02 2.64953643e-01 1.13156283e+00 -1.67110658e+00
-1.88362992e+00 1.37106907e+00 -3.79527081e-03 -8.76100436e-02
4.30106997e-01 1.62840784e-01 -3.44098806e-01 -9.95223671e-02
-2.24725460e-03 3.41428518e-01 7.72760153e-01 -1.21713495e+00
-6.35399282e-01 -2.48491511e-01 5.96344434e-02 9.87460986e-02
-3.79467487e-01 8.82719681e-02 -3.86314839e-01 -5.18162608e-01
1.92932293e-01 -1.31575000e+00 2.68038630e-01 -4.49374467e-01
-3.43669027e-01 -4.70330149e-01 5.11887312e-01 -6.56153679e-01
9.12398994e-01 -2.32617545e+00 1.57002315e-01 7.87214398e-01
-1.07190520e-01 5.17325819e-01 -1.35269597e-01 2.41244748e-01
-3.77789438e-01 -3.01474750e-01 -3.79954398e-01 -2.98730866e-03
-1.76697105e-01 -5.75404577e-02 -3.23173821e-01 2.98144042e-01
3.91145647e-01 4.98188376e-01 -1.00531244e+00 -3.70194137e-01
-1.06470577e-01 -4.41662110e-02 -3.64735782e-01 2.30048522e-01
2.06461288e-02 2.42321566e-01 -6.46785870e-02 5.36462307e-01
7.25126088e-01 6.32533729e-02 3.66680086e-01 -3.04899961e-01
1.45967335e-01 3.66688818e-01 -1.16021943e+00 8.24169040e-01
-2.46874496e-01 6.42919242e-01 -3.22205037e-01 -1.27045202e+00
9.19995427e-01 2.08433867e-01 1.62227362e-01 -4.28463846e-01
1.03008069e-01 5.47142863e-01 3.60203922e-01 -3.19691360e-01
3.67040783e-01 -1.66852966e-01 -6.96932450e-02 7.18869150e-01
1.53422505e-01 -5.17409109e-02 2.58991778e-01 -3.82865779e-02
6.43469810e-01 1.47221744e-01 2.34250575e-01 -2.21353546e-01
1.01200926e+00 6.45887926e-02 3.58300120e-01 4.58004534e-01
-1.06133543e-01 3.44347954e-01 7.71133721e-01 -5.34251928e-01
-1.01965010e+00 -1.02426922e+00 -4.13871616e-01 1.33094072e+00
-1.98606625e-01 -3.38515431e-01 -1.07595897e+00 -1.20284891e+00
-1.13748228e-02 5.51728189e-01 -9.08971786e-01 -3.21404159e-01
-4.08579677e-01 -1.15996099e+00 5.89620292e-01 5.41315734e-01
5.61931312e-01 -1.01672435e+00 -9.65621695e-02 4.32228707e-02
-1.85733974e-01 -7.39782870e-01 -4.01370615e-01 3.55284184e-01
-6.61547840e-01 -1.14557099e+00 -5.89239359e-01 -1.23940420e+00
1.04077852e+00 -1.32804975e-01 8.12239408e-01 -5.87223470e-02
2.14179933e-01 1.21410806e-02 -6.21946931e-01 -2.23927453e-01
-8.05950344e-01 8.65162238e-02 2.05864385e-01 5.11670887e-01
5.52362144e-01 4.11261059e-02 -8.39320943e-02 4.96868104e-01
-4.90199655e-01 -1.86423779e-01 3.83374423e-01 1.11646318e+00
1.86407402e-01 -1.03709288e-01 7.01042712e-01 -1.01415837e+00
6.51097059e-01 -1.77338451e-01 -5.56522787e-01 5.35476506e-01
-5.50088584e-01 1.56172678e-01 7.10910618e-01 -9.40946758e-01
-8.32976401e-01 2.76769012e-01 -1.43230423e-01 2.95404911e-01
5.31100780e-02 7.77400732e-01 -5.31736109e-03 -2.59065837e-01
8.94905508e-01 3.28971088e-01 2.43593931e-01 -7.53733739e-02
1.37564808e-01 1.05965662e+00 3.07816684e-01 -5.17619848e-01
4.71217215e-01 -4.79012541e-02 -3.75230014e-01 -8.07790577e-01
-6.39742434e-01 -3.12714845e-01 -9.26758528e-01 -3.16100746e-01
6.38774931e-01 -4.71749514e-01 -5.71427405e-01 1.05388165e+00
-6.91130280e-01 -4.71300900e-01 -1.78727716e-01 6.00263536e-01
-4.10868824e-01 3.69604468e-01 -4.66508090e-01 -8.12438607e-01
-1.68215692e-01 -1.08400345e+00 7.49606371e-01 2.64004976e-01
-2.85571367e-01 -1.16565263e+00 2.81834602e-01 1.58497453e-01
3.20123718e-03 -1.64966062e-01 1.33876383e+00 -1.13521624e+00
1.76318049e-01 -8.89642894e-01 -1.02395795e-01 6.79418504e-01
4.23175782e-01 -3.14148329e-02 -1.23709130e+00 -2.29102239e-01
-5.37658818e-02 -6.86785460e-01 6.47741437e-01 -2.50827696e-04
7.62585759e-01 3.67665812e-02 -2.65091658e-01 3.20037633e-01
9.52748299e-01 2.78435022e-01 5.94714284e-01 1.44093737e-01
6.04742885e-01 7.96637237e-01 5.73476076e-01 -6.89432546e-02
3.24428290e-01 5.37467003e-01 -3.28882307e-01 -1.20991901e-01
2.64889389e-01 -8.09736997e-02 5.79314828e-01 7.88239062e-01
-8.36426392e-02 1.58560127e-01 -1.01169086e+00 2.68158019e-01
-1.81484079e+00 -7.81615853e-01 -5.39668389e-02 2.70682597e+00
1.07960832e+00 2.51791269e-01 5.23920536e-01 6.50749147e-01
8.21982622e-01 -4.03303087e-01 -4.36953485e-01 -4.51732606e-01
-1.76069647e-01 1.37738183e-01 1.11554228e-01 6.68217659e-01
-1.22237706e+00 1.12936223e+00 6.17931080e+00 5.98798692e-01
-1.49609494e+00 -3.65548491e-01 9.23238039e-01 4.61847872e-01
1.07571110e-02 7.63914734e-02 -6.55956149e-01 3.90133858e-01
9.29832101e-01 -4.42952104e-02 5.28996587e-01 6.03747249e-01
-3.96043837e-01 -4.75691646e-01 -1.22584045e+00 8.13899159e-01
3.28379393e-01 -9.32604194e-01 -1.45186529e-01 -6.40875101e-02
5.25709033e-01 -2.92334914e-01 -4.33132797e-02 4.00505334e-01
2.01358944e-01 -7.53272235e-01 5.26808739e-01 2.72125691e-01
8.72875452e-01 -8.60849380e-01 1.09678674e+00 3.44324082e-01
-8.42899144e-01 -5.50947674e-02 -3.45541596e-01 -7.78115988e-02
-3.60919476e-01 6.43416762e-01 -1.09325385e+00 3.56194705e-01
5.13391435e-01 5.85644424e-01 -8.18714440e-01 9.05454218e-01
-2.36283228e-01 9.92440641e-01 -3.64681393e-01 -4.61430728e-01
1.20059423e-01 -6.05441868e-01 2.56270081e-01 1.12697124e+00
-3.57428379e-02 -1.01667374e-01 9.04238001e-02 2.48990357e-01
2.84823310e-02 6.13788307e-01 -5.51417530e-01 -6.70773461e-02
9.07101259e-02 1.29518867e+00 -9.02223587e-01 -6.24035716e-01
-4.02086735e-01 1.19011104e+00 7.81808019e-01 1.89987361e-01
-4.88962591e-01 -6.19153023e-01 2.23494425e-01 -2.22573265e-01
8.79873484e-02 1.48204923e-01 -3.05858076e-01 -1.26431191e+00
1.73871443e-02 -9.54535246e-01 4.14126486e-01 -5.41147888e-01
-1.56729043e+00 6.40834093e-01 -1.12010792e-01 -1.27421153e+00
-5.19704759e-01 -1.02436495e+00 -7.04508007e-01 1.16486907e+00
-9.07073557e-01 -1.10248661e+00 -1.86675359e-02 4.33844924e-01
-5.09705320e-02 -4.57326472e-01 1.24882066e+00 3.27690900e-03
-3.93132955e-01 8.80554199e-01 3.44338059e-01 7.46301115e-01
9.50243413e-01 -1.47805834e+00 8.64553005e-02 5.45043528e-01
1.39483586e-01 3.84901583e-01 3.46903056e-01 -5.24588168e-01
-9.47106659e-01 -6.42120659e-01 1.17213333e+00 -4.39257383e-01
7.73865283e-01 -6.22501910e-01 -1.13702559e+00 4.70667243e-01
-2.49165639e-01 -2.70050496e-01 1.17479539e+00 4.55188125e-01
-5.88812232e-01 -2.12959841e-01 -1.16108096e+00 5.22339821e-01
3.87988985e-01 -8.09238911e-01 -6.03844643e-01 3.38202208e-01
4.75522354e-02 -2.60537535e-01 -7.30079591e-01 2.36452281e-01
8.16181600e-01 -6.31921113e-01 5.10707080e-01 -6.44196570e-01
1.90009460e-01 -1.27615109e-01 -1.79016769e-01 -1.61449277e+00
-1.40312865e-01 -2.72503436e-01 3.97011846e-01 1.21557724e+00
7.89825022e-01 -9.51287985e-01 7.57966399e-01 4.43906188e-01
3.19290310e-01 -7.98873305e-01 -6.80093825e-01 -4.90166843e-01
3.65481049e-01 -2.48948827e-01 3.51237714e-01 1.26989567e+00
6.71087623e-01 5.75200677e-01 2.65985541e-02 -6.55100495e-02
5.02573788e-01 2.33495280e-01 6.17154956e-01 -1.36852360e+00
-2.52544403e-01 -2.86396742e-01 -4.87550139e-01 -5.54113805e-01
4.32889074e-01 -1.14391243e+00 1.13250151e-01 -9.14006531e-01
1.30829573e-01 -7.54496932e-01 -1.63034871e-01 6.60239637e-01
-5.94966948e-01 4.64425087e-01 -7.66152069e-02 2.30987072e-01
-1.60741001e-01 1.61236495e-01 8.34865451e-01 -8.92158300e-02
-5.70156515e-01 3.96726310e-01 -4.62844133e-01 6.24341547e-01
9.39564466e-01 -1.93003878e-01 -4.22506452e-01 1.86638951e-01
1.36848584e-01 -2.19193161e-01 1.46225497e-01 -7.96556652e-01
-7.27902204e-02 1.40290454e-01 5.86152136e-01 -7.07273558e-02
1.07575387e-01 -4.16034758e-01 -5.01235902e-01 4.83963877e-01
-3.24045032e-01 1.63831282e-02 1.65210560e-01 1.42785341e-01
-1.20492950e-01 -5.91594100e-01 1.05202579e+00 2.70135403e-01
-3.38527352e-01 -9.20830220e-02 -5.19309223e-01 -8.75943676e-02
8.87414634e-01 -4.11390334e-01 -5.95682338e-02 -3.24636102e-01
-8.38109434e-01 -7.62948990e-02 6.41313672e-01 2.92701513e-01
4.22407895e-01 -1.18268156e+00 -8.72647226e-01 5.25152206e-01
3.87711197e-01 -4.50166225e-01 -3.65763128e-01 8.55730116e-01
-3.30257386e-01 1.57139257e-01 -1.47347763e-01 -5.41498482e-01
-1.31436884e+00 4.10244316e-01 4.33384895e-01 -1.35140553e-01
-1.67564243e-01 7.67362058e-01 1.76145844e-02 -9.15697753e-01
-8.79907161e-02 -1.10509805e-01 -4.08259898e-01 3.17386657e-01
4.44501102e-01 5.65212294e-02 3.02904218e-01 -7.95941591e-01
-3.52065831e-01 5.27891278e-01 -4.64987904e-01 -4.60887432e-01
1.07639122e+00 4.69319850e-01 -4.94183183e-01 7.66885400e-01
1.17719722e+00 3.52540433e-01 -6.81205451e-01 -3.73052329e-01
1.00039706e-01 -2.57221818e-01 -2.17788562e-01 -1.04112267e+00
-5.89505315e-01 5.55658102e-01 7.59999871e-01 4.60045874e-01
1.08667541e+00 2.64422819e-02 1.34050757e-01 4.93301392e-01
2.44860440e-01 -1.31872702e+00 -1.30058065e-01 7.10106671e-01
5.63658416e-01 -1.42134714e+00 -3.85846168e-01 -6.08638346e-01
-7.33345985e-01 1.28301144e+00 3.96327496e-01 -5.90150952e-02
7.44990051e-01 -9.34544485e-03 3.75265926e-01 1.60228983e-01
-4.06967193e-01 -1.25142992e-01 4.53009754e-01 7.21144438e-01
7.92092085e-01 1.75608426e-01 -3.59899282e-01 3.55597675e-01
-3.22057754e-01 -9.71733183e-02 2.68517166e-01 7.68350244e-01
-3.02718967e-01 -1.34500325e+00 -4.88927990e-01 7.46724606e-01
1.35526925e-01 -2.62334943e-01 -9.26367342e-01 4.50329721e-01
-8.59053805e-02 9.44424570e-01 5.29681519e-02 -7.59766638e-01
3.16315293e-01 5.71041465e-01 6.52000368e-01 -4.29833174e-01
-6.51782155e-01 -2.25606814e-01 3.37207913e-01 1.34640679e-01
-3.44179034e-01 -8.48660707e-01 -1.27908647e+00 -1.68650821e-01
-5.85595906e-01 5.40851533e-01 6.05053365e-01 1.00005150e+00
-9.81360674e-02 -1.55476615e-01 9.75242198e-01 -6.56153142e-01
-5.66324115e-01 -1.14023340e+00 -5.30184090e-01 6.18993878e-01
2.10788324e-01 -4.50393617e-01 -4.95319337e-01 2.26385504e-01] | [10.217606544494629, 10.52568531036377] |
2fbaac9a-ec46-4e4a-8a59-17fa4564b82d | local-aggressive-adversarial-attacks-on-3d | 2105.09090 | null | https://arxiv.org/abs/2105.09090v2 | https://arxiv.org/pdf/2105.09090v2.pdf | Local Aggressive Adversarial Attacks on 3D Point Cloud | Deep neural networks are found to be prone to adversarial examples which could deliberately fool the model to make mistakes. Recently, a few of works expand this task from 2D image to 3D point cloud by using global point cloud optimization. However, the perturbations of global point are not effective for misleading the victim model. First, not all points are important in optimization toward misleading. Abundant points account considerable distortion budget but contribute trivially to attack. Second, the multi-label optimization is suboptimal for adversarial attack, since it consumes extra energy in finding multi-label victim model collapse and causes instance transformation to be dissimilar to any particular instance. Third, the independent adversarial and perceptibility losses, caring misclassification and dissimilarity separately, treat the updating of each point equally without a focus. Therefore, once perceptibility loss approaches its budget threshold, all points would be stock in the surface of hypersphere and attack would be locked in local optimality. Therefore, we propose a local aggressive adversarial attacks (L3A) to solve above issues. Technically, we select a bunch of salient points, the high-score subset of point cloud according to gradient, to perturb. Then a flow of aggressive optimization strategies are developed to reinforce the unperceptive generation of adversarial examples toward misleading victim models. Extensive experiments on PointNet, PointNet++ and DGCNN demonstrate the state-of-the-art performance of our method against existing adversarial attack methods. | ['Mingjie Wang', 'Zhiyu Chen', 'Feng Chen', 'Yiming Sun'] | 2021-05-19 | null | null | null | null | ['image-to-3d'] | ['computer-vision'] | [-4.97837700e-02 2.82612324e-01 2.12686941e-01 6.51536733e-02
-6.53047204e-01 -8.76210570e-01 5.49830794e-01 -2.34985456e-01
-1.03752136e-01 5.63069820e-01 -2.27780566e-01 1.19751245e-02
6.47501845e-04 -1.00024414e+00 -1.04493284e+00 -1.01341617e+00
-1.14338644e-01 4.61987913e-01 1.52549595e-01 -4.86479670e-01
3.40038776e-01 1.03998530e+00 -1.02194178e+00 -6.87212348e-02
9.49688077e-01 7.78336287e-01 -2.41812363e-01 4.14697737e-01
1.53090417e-01 5.40567040e-01 -1.25911605e+00 -8.20692182e-01
8.95760238e-01 -9.66259092e-02 -5.87838471e-01 -3.13240886e-01
6.05816424e-01 -5.53975165e-01 -3.84384423e-01 1.77832854e+00
8.61182690e-01 2.73997366e-01 6.61162913e-01 -1.68900275e+00
-8.42838585e-01 3.89668673e-01 -7.78833330e-01 -7.47396203e-04
-2.71081254e-02 5.18014193e-01 6.53253376e-01 -8.92303169e-01
2.51122802e-01 1.54368222e+00 6.29718423e-01 7.42687702e-01
-8.44668806e-01 -9.90475416e-01 3.84242743e-01 9.83166248e-02
-1.35838628e+00 -4.74431776e-02 1.30615330e+00 -4.35894765e-02
6.34101808e-01 8.30913961e-01 6.36877358e-01 1.35782754e+00
3.30220371e-01 5.95410049e-01 6.69760644e-01 1.06923878e-01
2.22431451e-01 1.03156187e-01 -5.31682193e-01 5.63418806e-01
1.94388703e-01 7.25649536e-01 -1.94170550e-01 -4.46210653e-01
7.50677884e-01 9.26842913e-02 -3.02900851e-01 -2.60841370e-01
-9.25863266e-01 1.03732800e+00 1.12469482e+00 -1.93618730e-01
-3.86533409e-01 4.47803199e-01 1.75169080e-01 1.94135383e-01
3.84243727e-01 9.61623907e-01 -2.57951021e-01 2.35621020e-01
-5.55317283e-01 5.43499351e-01 3.89441818e-01 9.51054752e-01
4.99706060e-01 4.97911036e-01 -1.34026945e-01 5.04683554e-01
2.46749431e-01 7.46460140e-01 1.35718018e-01 -1.09202719e+00
3.86888981e-01 6.45923138e-01 -2.93409210e-02 -1.59304726e+00
-3.06874990e-01 -6.71452701e-01 -1.12305212e+00 1.04444122e+00
1.98369190e-01 -3.37351799e-01 -8.83968055e-01 1.79171324e+00
5.91379821e-01 4.27994072e-01 -1.49745550e-02 1.30503738e+00
6.71250999e-01 6.36324465e-01 -7.55863786e-02 6.47946671e-02
8.02611887e-01 -8.46060336e-01 -2.13442788e-01 -7.51133710e-02
1.32197663e-01 -7.86625743e-01 8.98128986e-01 3.09069067e-01
-1.40572059e+00 -3.73992562e-01 -1.09629798e+00 3.32141876e-01
-3.70501965e-01 -5.66179514e-01 3.68887186e-01 5.87376297e-01
-6.76726580e-01 9.47846591e-01 -6.73480868e-01 3.19042534e-01
6.99931324e-01 4.41755354e-01 4.52357717e-02 3.03805918e-01
-1.39620554e+00 9.66947198e-01 6.75883964e-02 7.88098853e-03
-1.37150908e+00 -1.01233661e+00 -5.01936913e-01 -7.14142397e-02
1.46778509e-01 -7.80645669e-01 7.16231108e-01 -1.06104755e+00
-1.31333065e+00 7.56586730e-01 4.92945313e-01 -5.06113887e-01
1.01592958e+00 -1.84399337e-01 -1.68603897e-01 9.81825590e-02
2.27553137e-02 9.91555631e-01 1.22552681e+00 -1.76498425e+00
-2.95005888e-01 -3.20594102e-01 4.85882878e-01 6.15635693e-01
-2.01944649e-01 9.38310623e-02 -1.49886757e-01 -1.07884073e+00
4.07028534e-02 -9.65627253e-01 -4.55897331e-01 3.16904843e-01
-8.59377027e-01 1.08873630e-02 1.07455182e+00 -8.58821571e-02
7.48199403e-01 -2.12670326e+00 2.92398214e-01 2.77275383e-01
5.28535187e-01 3.44772041e-01 -2.81690750e-02 5.10554351e-02
-3.39297920e-01 6.02039278e-01 -2.78450131e-01 -2.60637850e-01
4.40278724e-02 1.74861804e-01 -9.29629683e-01 7.51253009e-01
2.32740104e-01 9.90283847e-01 -9.88291383e-01 -2.76380181e-01
3.34083766e-01 3.05332899e-01 -7.33362913e-01 1.05973780e-01
4.67671156e-02 3.97275060e-01 -5.66988289e-01 8.75613868e-01
1.21013284e+00 3.12550157e-01 -9.06906724e-01 -7.20991418e-02
2.74369210e-01 -2.86697537e-01 -9.96896505e-01 1.22305429e+00
1.91544108e-02 1.29051760e-01 3.69871035e-02 -7.79782236e-01
7.48578608e-01 1.04190461e-01 4.58883911e-01 -1.73774451e-01
2.67561585e-01 3.02897599e-02 -4.16367911e-02 -1.48900777e-01
2.71459430e-01 -3.16159040e-01 -8.39132965e-02 1.84506506e-01
-3.78532797e-01 -4.95570600e-01 -7.37990439e-01 8.71052891e-02
1.02862406e+00 -1.59227282e-01 -1.32875964e-01 -1.33970484e-01
4.51942563e-01 1.55216366e-01 7.45308876e-01 7.78155684e-01
-5.32460868e-01 9.09987628e-01 4.09712613e-01 -4.54352021e-01
-1.00527072e+00 -1.38610733e+00 -1.02866497e-02 6.61066532e-01
7.86043465e-01 1.08815536e-01 -8.36306870e-01 -1.01659060e+00
2.11742759e-01 8.92275035e-01 -6.66549087e-01 -9.21875000e-01
-7.74440408e-01 -8.13850045e-01 1.08477664e+00 2.66945571e-01
6.57394588e-01 -1.22137237e+00 -3.19134325e-01 -8.73930454e-02
4.43698429e-02 -5.61756074e-01 -4.04815644e-01 -4.59101126e-02
-7.38371849e-01 -1.03434086e+00 -6.66453421e-01 -5.18613040e-01
1.00319362e+00 3.04634571e-01 1.03492284e+00 3.79844129e-01
-1.54040292e-01 -7.96411708e-02 -1.84328616e-01 -9.35096979e-01
-4.02718127e-01 -4.44723964e-01 4.75473374e-01 -2.49824390e-01
4.40626442e-02 -7.36360371e-01 -6.52074039e-01 6.09609306e-01
-8.14951003e-01 -4.30288583e-01 2.69401938e-01 7.50923514e-01
5.99293649e-01 3.31423908e-01 3.78264487e-01 -3.27202439e-01
6.41103864e-01 -5.95177472e-01 -4.49182123e-01 -2.28208169e-01
-2.44860485e-01 -3.52559388e-01 1.22842813e+00 -6.58426821e-01
-5.70164561e-01 -2.64387995e-01 -1.59305573e-01 -1.44598186e+00
-1.21274926e-02 -2.13459656e-01 -4.33940202e-01 -7.03465879e-01
1.12751853e+00 8.57213959e-02 -5.24076670e-02 -1.01573788e-01
4.22344625e-01 -1.71026886e-02 6.22566640e-01 -6.12882972e-01
1.72462249e+00 6.98304951e-01 3.41537744e-01 -3.40446264e-01
-6.49027109e-01 7.83188827e-03 -2.59481609e-01 -4.28929061e-01
6.47192717e-01 -6.43799961e-01 -7.43630707e-01 7.18054295e-01
-1.24484611e+00 -1.90104768e-02 -5.35317719e-01 2.52797324e-02
-4.52809900e-01 4.43137795e-01 -2.18936861e-01 -5.66381097e-01
-3.43970716e-01 -1.35182405e+00 9.95139301e-01 1.41206861e-01
1.20781846e-01 -8.38968813e-01 -8.85720178e-02 7.18016699e-02
2.32595846e-01 9.02152658e-01 6.31316185e-01 -7.21134901e-01
-7.26947904e-01 -3.80578458e-01 1.96411368e-02 4.11046445e-01
-4.18792516e-02 1.73935324e-01 -9.34790254e-01 -5.66801250e-01
5.13397813e-01 -3.94434392e-01 4.70766425e-01 2.80202359e-01
1.43942249e+00 -7.64074028e-01 -3.51373494e-01 1.13640177e+00
1.23907125e+00 1.78499460e-01 6.09661520e-01 3.86377722e-01
1.07442486e+00 3.72113943e-01 6.87574267e-01 1.67531490e-01
-2.29503393e-01 5.70672035e-01 1.45679533e+00 -2.98715055e-01
-1.18338717e-02 -2.16737375e-01 3.74468625e-01 2.00043127e-01
-1.15018763e-01 -2.67678112e-01 -6.79875135e-01 3.76026094e-01
-1.56560278e+00 -1.11611104e+00 -2.67687067e-02 2.01723766e+00
6.13795280e-01 3.79198402e-01 -7.32675716e-02 4.48968522e-02
8.66207063e-01 3.34089905e-01 -9.69224930e-01 -3.49758238e-01
-2.30281293e-01 -2.34983593e-01 7.99746990e-01 4.20603126e-01
-1.13898706e+00 1.05410671e+00 6.12078571e+00 1.12898588e+00
-1.11190438e+00 2.29214709e-02 6.88080490e-01 -6.30382955e-01
-4.07993644e-01 -1.75243810e-01 -5.72792053e-01 6.98078334e-01
2.83688605e-01 -3.03881168e-01 7.21323669e-01 1.12983751e+00
1.07144944e-01 5.34893155e-01 -8.97595525e-01 9.80357826e-01
2.93272641e-03 -1.22534299e+00 4.96935010e-01 1.77052662e-01
8.89413476e-01 6.12157807e-02 6.06022775e-01 1.87363684e-01
7.26255774e-01 -1.22632933e+00 8.28709304e-01 2.26674333e-01
4.77105975e-01 -1.30592108e+00 4.91976768e-01 6.00635409e-01
-6.79233372e-01 -5.55855110e-02 -9.17492986e-01 2.65085042e-01
8.64561200e-02 3.76113117e-01 -5.28163910e-01 4.87965405e-01
8.78014922e-01 2.48227686e-01 -4.25346464e-01 8.14475834e-01
-4.21452343e-01 1.76122084e-01 -5.57500362e-01 1.05714530e-01
6.73724055e-01 -3.03387139e-02 1.56673133e+00 5.22157907e-01
1.38051853e-01 1.79598480e-01 1.83506813e-02 1.23482728e+00
-1.83297127e-01 -2.71122456e-01 -9.38203871e-01 6.79536283e-01
6.38147354e-01 1.15172791e+00 -1.99591294e-01 -6.11503907e-02
1.49694309e-01 9.74614143e-01 2.66962737e-01 2.28466257e-01
-1.35172284e+00 -3.85406762e-01 1.20029163e+00 -2.09938735e-01
-1.23972371e-01 2.05331728e-01 -7.20933318e-01 -9.32577968e-01
1.04473673e-01 -1.04555595e+00 1.11843303e-01 -7.91956782e-01
-1.48461294e+00 6.30809128e-01 -2.74040382e-02 -1.71276474e+00
1.12810895e-01 -3.74837160e-01 -1.27163529e+00 8.84024322e-01
-1.23459864e+00 -1.02471077e+00 -4.06010151e-01 9.67135072e-01
4.91074502e-01 -4.08402473e-01 5.38308501e-01 1.52637601e-01
-5.61696410e-01 9.80477333e-01 -1.56730339e-01 9.24814492e-02
6.54651880e-01 -1.30891156e+00 7.29959786e-01 1.11098576e+00
-1.21646017e-01 5.61384499e-01 8.75016332e-01 -7.58367598e-01
-1.04882729e+00 -1.54426658e+00 1.23156816e-01 -9.23306525e-01
5.75690985e-01 -9.04892907e-02 -8.86934042e-01 4.89318013e-01
-1.26002077e-03 -2.74014529e-02 1.73044994e-01 -5.27310133e-01
-4.49749641e-02 9.34837982e-02 -1.68322074e+00 1.14675498e+00
1.18919444e+00 -1.28125548e-01 -6.48826063e-01 6.28723979e-01
1.09860241e+00 -7.26830244e-01 -7.08856642e-01 6.31117642e-01
-2.33961359e-01 -9.52578604e-01 1.58518422e+00 -7.01347232e-01
5.66658497e-01 -4.72346157e-01 -3.86715792e-02 -1.61668599e+00
-5.10461688e-01 -9.91915345e-01 -2.48402506e-01 9.88892853e-01
2.23853618e-01 -6.49245560e-01 9.12180960e-01 4.29674983e-01
-4.88836557e-01 -9.66014087e-01 -1.21798122e+00 -8.34417284e-01
5.76769412e-01 -2.38284618e-01 9.55078304e-01 1.20965147e+00
-4.08053339e-01 -1.97263911e-01 -4.74310964e-01 7.69658625e-01
9.97816682e-01 -2.65137643e-01 8.70157719e-01 -9.03765738e-01
-1.09873526e-01 -7.83644855e-01 -5.20487070e-01 -8.62441659e-01
-3.36238518e-02 -9.19775307e-01 -1.04938030e-01 -9.60938096e-01
-3.42198461e-01 -6.49527848e-01 -2.57766038e-01 4.06428069e-01
-5.36824465e-01 2.66977698e-01 5.04593670e-01 6.26787305e-01
-1.33874506e-01 6.52289808e-01 1.60376716e+00 -5.19178569e-01
-4.16949727e-02 1.69435635e-01 -8.73940766e-01 9.97192800e-01
7.86505878e-01 -7.81619549e-01 -3.80829215e-01 -4.44426984e-01
1.47310242e-01 -2.76632816e-01 7.62602568e-01 -1.08366632e+00
1.93916470e-01 -3.80941212e-01 6.80424869e-01 -5.88614404e-01
6.06485546e-01 -1.07169557e+00 1.36433691e-01 6.76584005e-01
-4.77109812e-02 9.85447690e-02 2.43474647e-01 5.97194016e-01
-2.21709907e-02 -3.32566231e-01 1.26850402e+00 -3.28063786e-01
-2.22075611e-01 7.67254412e-01 7.67579824e-02 1.64831027e-01
1.52540326e+00 -3.10339600e-01 -5.65332651e-01 -2.14464158e-01
-6.25348091e-01 4.39664990e-01 7.72989452e-01 6.03921950e-01
7.24101424e-01 -1.69700873e+00 -7.12283492e-01 1.71065211e-01
-2.54222095e-01 5.85243642e-01 2.11480200e-01 4.01523590e-01
-7.15952516e-01 -2.40723789e-01 -3.63347292e-01 -5.83924294e-01
-1.02409315e+00 1.09857094e+00 7.28576899e-01 -3.66580412e-02
-6.77511692e-01 1.34055305e+00 5.70798218e-01 -6.11078203e-01
3.27990919e-01 1.24493510e-01 -4.57607917e-02 -2.51466244e-01
2.37524956e-01 5.10868728e-01 -1.77133530e-01 -6.17909074e-01
-4.33922768e-01 6.51361525e-01 3.00245713e-02 2.11148709e-01
9.72827375e-01 1.51494339e-01 -8.43362212e-02 -2.12685540e-01
1.28857422e+00 1.82348505e-01 -1.44936633e+00 2.41175190e-01
-8.92166615e-01 -6.41269565e-01 -1.34711906e-01 -6.48637652e-01
-1.38404393e+00 7.60830641e-01 3.47333431e-01 3.62826318e-01
1.09984994e+00 -2.90839404e-01 8.47027421e-01 3.22649986e-01
2.20551550e-01 -8.16257060e-01 2.19226539e-01 3.46664280e-01
1.43341565e+00 -1.05980122e+00 6.31269068e-02 -3.39152515e-01
-7.86815226e-01 7.39287078e-01 1.10170925e+00 -8.33438098e-01
5.15734196e-01 8.89864266e-02 2.43646920e-01 -3.79110932e-01
-4.47173774e-01 4.92243975e-01 1.43042400e-01 7.57550895e-01
-7.90741920e-01 -1.00754932e-01 3.02695841e-01 2.51425624e-01
-6.68220341e-01 -8.32509696e-01 3.51449072e-01 6.87407196e-01
-4.25934434e-01 -6.36540115e-01 -8.58356357e-01 9.70689133e-02
-6.23168647e-01 -7.00867176e-02 -7.22552478e-01 7.62935579e-01
2.66119033e-01 7.21325159e-01 -1.66369230e-02 -6.28204823e-01
5.16739190e-01 -4.96138126e-01 1.30306721e-01 -2.38724172e-01
-1.02040732e+00 -1.29381284e-01 -4.33518231e-01 -7.26453900e-01
1.87530860e-01 -4.06098127e-01 -1.24850476e+00 -7.65422344e-01
-2.77732909e-01 -8.34577009e-02 4.63422626e-01 5.32176375e-01
2.97870189e-01 6.58630788e-01 1.26938760e+00 -1.25148809e+00
-1.00713265e+00 -5.91519713e-01 -1.94691598e-01 5.26122034e-01
3.58168483e-01 -8.00021172e-01 -1.08972251e+00 -4.66116995e-01] | [7.712672710418701, -4.456973075866699] |
2ddf514f-f3df-423e-9916-214635342945 | finbert-lstm-deep-learning-based-stock-price | 2211.07392 | null | https://arxiv.org/abs/2211.07392v1 | https://arxiv.org/pdf/2211.07392v1.pdf | FinBERT-LSTM: Deep Learning based stock price prediction using News Sentiment Analysis | Economy is severely dependent on the stock market. An uptrend usually corresponds to prosperity while a downtrend correlates to recession. Predicting the stock market has thus been a centre of research and experiment for a long time. Being able to predict short term movements in the market enables investors to reap greater returns on their investments. Stock prices are extremely volatile and sensitive to financial market. In this paper we use Deep Learning networks to predict stock prices, assimilating financial, business and technology news articles which present information about the market. First, we create a simple Multilayer Perceptron (MLP) network and then expand into more complex Recurrent Neural Network (RNN) like Long Short Term Memory (LSTM), and finally propose FinBERT-LSTM model, which integrates news article sentiments to predict stock price with greater accuracy by analysing short-term market information. We then train the model on NASDAQ-100 index stock data and New York Times news articles to evaluate the performance of MLP, LSTM, FinBERT-LSTM models using mean absolute error (MAE), mean absolute percentage error (MAPE) and accuracy metrics. | ['Shayan Halder'] | 2022-11-11 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-7.59972334e-01 -3.48489404e-01 -3.66888374e-01 -1.09438539e-01
-1.86350241e-01 -5.42202890e-01 8.90280902e-01 -2.08744127e-02
-4.73096788e-01 1.04845381e+00 5.26766419e-01 -6.19107306e-01
1.14013478e-01 -1.15075111e+00 -5.35878539e-01 -3.21277797e-01
-8.99863020e-02 2.12678149e-01 1.43922687e-01 -7.26236880e-01
9.02535975e-01 4.68084633e-01 -1.20927489e+00 9.38583016e-02
1.26241684e-01 1.58028746e+00 3.54873911e-02 4.83654290e-01
-5.41995287e-01 1.71589673e+00 -5.95473111e-01 -2.35516757e-01
4.31635112e-01 -1.25664085e-01 -7.00768888e-01 -4.14030612e-01
-3.93388927e-01 -2.37101570e-01 6.52570277e-02 1.08220637e+00
2.26862490e-01 -8.19291994e-02 4.12805855e-01 -8.88439596e-01
-9.25992250e-01 9.15891767e-01 -5.33796608e-01 9.83258843e-01
-3.15241180e-02 -1.72711477e-01 1.09520435e+00 -1.03030550e+00
3.70507270e-01 7.34496295e-01 8.15942705e-01 -1.08676746e-01
-6.66300893e-01 -8.30137253e-01 5.95840737e-02 1.04448535e-01
-5.66213846e-01 -1.91991821e-01 6.32613897e-01 -5.72915316e-01
1.59905076e+00 5.63699193e-02 8.49135637e-01 7.65567899e-01
1.03150010e+00 4.76016730e-01 1.25652444e+00 -2.53266811e-01
1.82361037e-01 3.51372123e-01 1.66604072e-01 9.15873125e-02
2.40250364e-01 2.99603105e-01 -3.02518338e-01 1.42261729e-01
6.96712494e-01 4.98306304e-01 1.83257218e-02 6.19978070e-01
-1.16363132e+00 8.70623291e-01 6.46413326e-01 8.64523590e-01
-1.11436701e+00 5.89639731e-02 5.20247936e-01 1.13327324e+00
8.05536211e-01 4.41188782e-01 -8.49897683e-01 -4.20507431e-01
-1.25130403e+00 3.90524268e-01 9.45430636e-01 8.94064680e-02
3.36722076e-01 6.62346840e-01 2.38726959e-01 7.85889149e-01
4.32486147e-01 5.26252925e-01 1.49553216e+00 -2.53004402e-01
3.07713658e-01 5.87421715e-01 2.56586134e-01 -1.24923539e+00
-5.58009326e-01 -8.41708779e-01 -7.09387362e-01 5.89910507e-01
-1.68040600e-02 -4.50149000e-01 -7.04280317e-01 1.08340681e+00
-4.32283521e-01 2.53246516e-01 6.02178156e-01 3.86609256e-01
4.49105859e-01 1.34162128e+00 -2.52598345e-01 -4.95685309e-01
1.07380319e+00 -8.65323186e-01 -7.69260406e-01 -3.07664037e-01
4.59130079e-01 -8.63873780e-01 2.82234251e-01 3.21300566e-01
-1.16121888e+00 -3.30967247e-01 -1.01740110e+00 5.02034307e-01
-1.10390234e+00 -3.23668331e-01 -1.68024059e-02 1.81194946e-01
-1.07391977e+00 9.46513414e-01 -6.85258925e-01 1.75175920e-01
-3.35205644e-02 4.01355326e-01 8.02430511e-02 9.93992865e-01
-1.49955404e+00 1.50189769e+00 7.67670453e-01 2.87445724e-01
-5.20651899e-02 -4.47893411e-01 -5.43272018e-01 4.90869097e-02
-2.01539934e-01 -1.32501379e-01 1.49469841e+00 -1.07142043e+00
-1.78355038e+00 4.75649565e-01 3.18558186e-01 -1.23731673e+00
2.73940206e-01 -4.57094749e-03 -9.25223112e-01 -5.12222707e-01
-2.00392827e-01 1.03322349e-01 4.66675341e-01 -3.83353174e-01
-9.40842927e-01 -2.22619161e-01 -3.07432950e-01 -5.12074307e-02
-1.73144728e-01 3.02766651e-01 7.92071581e-01 -1.00345004e+00
2.39511192e-01 -7.03304708e-01 2.26548370e-02 -1.06083429e+00
-5.26853129e-02 -2.99376816e-01 8.85324359e-01 -9.58568037e-01
1.30165386e+00 -1.68774152e+00 -4.80100572e-01 3.53712678e-01
-1.58831000e-01 3.02556574e-01 3.77294719e-01 5.37319243e-01
-4.83946174e-01 1.47288978e-01 8.83642212e-02 1.02452487e-01
1.80603474e-01 1.71183676e-01 -9.71729934e-01 1.97912633e-01
1.70162633e-01 1.12786639e+00 -3.55014384e-01 1.13067262e-01
9.05375034e-02 1.54391870e-01 4.71046567e-02 -1.84644222e-01
-2.38056391e-01 -2.39258692e-01 -3.07693601e-01 7.43542314e-01
3.63954693e-01 -3.63622010e-01 -5.24269044e-01 2.45081559e-01
-9.41859543e-01 5.23262560e-01 -9.55685437e-01 6.67160928e-01
-4.34311301e-01 9.42865729e-01 -5.26828885e-01 -1.05817986e+00
1.48404515e+00 6.56614780e-01 1.09501444e-01 -1.24594414e+00
3.91155452e-01 1.10768723e+00 -1.01357117e-01 -2.37486914e-01
7.36957967e-01 -8.66015196e-01 1.24583401e-01 6.01687849e-01
-2.52052784e-01 5.16380847e-01 1.55607104e-01 -4.25897419e-01
6.28287673e-01 -2.71986485e-01 4.39585090e-01 -2.83806980e-01
4.68133688e-01 -1.43142670e-01 3.94992530e-01 5.75389676e-02
-1.31284952e-01 1.97589695e-01 6.17396951e-01 -9.46140170e-01
-1.11778402e+00 -5.46593070e-01 -2.39785120e-01 9.41028655e-01
-5.62783778e-01 4.59953249e-01 -1.77721322e-01 -1.17613733e-01
7.32105970e-02 9.51310158e-01 -7.03341484e-01 1.99392334e-01
-4.55812246e-01 -1.14813340e+00 1.89801097e-01 5.47770381e-01
7.83904850e-01 -1.77731884e+00 -1.01504242e+00 6.68727160e-01
3.95183444e-01 -5.17517745e-01 4.40915748e-02 3.37804705e-01
-9.88195717e-01 -4.81361926e-01 -1.17077589e+00 -6.09393179e-01
-1.10218130e-01 -5.62113643e-01 1.18417084e+00 -4.08444971e-01
4.96610135e-01 -3.12925994e-01 -2.75401622e-01 -9.34158444e-01
-4.39773232e-01 1.05795622e-01 2.21582577e-02 -5.26302233e-02
5.02637506e-01 -6.60247326e-01 -4.25849229e-01 -6.46482483e-02
-6.53510988e-01 -2.93845177e-01 8.28467369e-01 6.08171046e-01
3.24092418e-01 2.71714658e-01 1.13216364e+00 -6.45274699e-01
8.54085922e-01 -8.93587708e-01 -9.69687521e-01 -9.30124149e-02
-1.07501841e+00 2.26443902e-01 3.38029265e-01 -8.19464698e-02
-1.05264115e+00 -6.05924606e-01 -3.73079687e-01 -1.43863991e-01
4.60213661e-01 1.39179492e+00 8.68958175e-01 2.92142808e-01
3.29797387e-01 6.17110133e-01 1.34499287e-02 -5.45196950e-01
-3.46482366e-01 7.17499316e-01 5.43195546e-01 2.64076084e-01
4.81764644e-01 1.08841136e-01 -3.14447999e-01 -4.98862207e-01
-8.58564675e-01 -2.47861490e-01 -4.36856389e-01 -3.20100248e-01
6.25502944e-01 -8.96652281e-01 -4.90059704e-01 9.27612424e-01
-9.97204423e-01 -2.25624330e-02 -2.57494509e-01 6.72301948e-01
-2.13792831e-01 -4.71816778e-01 -1.01270008e+00 -1.14312530e+00
-8.17455113e-01 -7.04706788e-01 1.67759880e-01 3.31788868e-01
-2.79240519e-01 -1.46432245e+00 6.54321373e-01 -3.15436572e-01
9.92818117e-01 6.80018365e-01 3.70638549e-01 -1.20202243e+00
-9.84837264e-02 -7.08792865e-01 -1.74204409e-01 7.36656725e-01
5.71961515e-02 1.77551236e-03 -7.84274399e-01 1.77026968e-02
5.76166511e-01 -1.48493886e-01 9.61580276e-01 7.18108714e-01
-2.05873828e-02 -7.30068803e-01 2.60026366e-01 3.00365746e-01
1.67749631e+00 8.15342605e-01 6.95906341e-01 1.34891188e+00
2.20551684e-01 2.21893996e-01 1.47808447e-01 5.24269402e-01
3.64945203e-01 -2.20403284e-01 2.98004746e-01 1.53329059e-01
7.41555810e-01 -1.88297331e-02 8.41999769e-01 1.22872376e+00
-2.18967035e-01 1.72550932e-01 -1.01682222e+00 3.86280566e-01
-1.68922043e+00 -1.26430452e+00 1.09798841e-01 1.72466123e+00
6.79308176e-01 9.31590855e-01 3.83031636e-01 5.86991310e-01
4.30222362e-01 4.86408025e-01 -4.96899426e-01 -7.31724560e-01
-3.59650999e-01 3.55067588e-02 1.01601744e+00 3.05421025e-01
-8.92673016e-01 6.69505775e-01 6.04508734e+00 3.21146131e-01
-1.85619044e+00 -1.05710380e-01 9.25284982e-01 4.45893332e-02
-1.79089144e-01 -3.26529980e-01 -7.82063484e-01 7.23719478e-01
1.65803075e+00 -4.98184115e-01 -8.48427936e-02 9.42018807e-01
3.30897033e-01 5.91570772e-02 -5.02256393e-01 7.27080584e-01
-2.86144525e-01 -1.90777409e+00 -6.87912628e-02 2.42439955e-01
7.43319988e-01 7.11330652e-01 4.22503620e-01 4.69631344e-01
1.21877447e-01 -8.50717425e-01 1.05590677e+00 1.10528028e+00
-1.35888860e-01 -9.98434305e-01 1.38109231e+00 4.56936568e-01
-1.02501869e+00 -3.83232802e-01 -4.65875089e-01 -7.05435157e-01
6.77859411e-02 5.07229209e-01 -5.40759027e-01 1.06669098e-01
7.40065455e-01 1.00883090e+00 -2.07054868e-01 6.96637332e-01
3.79333168e-01 4.22852010e-01 -3.50410104e-01 -4.61072117e-01
9.71617877e-01 -3.41759473e-01 2.19402969e-01 1.06806362e+00
6.78244770e-01 1.54813051e-01 -4.08717632e-01 8.39184046e-01
4.71516186e-03 3.11147094e-01 -5.05670965e-01 -4.29884583e-01
1.52311236e-01 6.63225949e-01 -8.41610312e-01 -5.25873244e-01
-6.81137621e-01 5.10161638e-01 -1.79670364e-01 2.49814793e-01
-3.51009727e-01 -4.01425213e-01 3.36522847e-01 1.33940861e-01
3.16320926e-01 1.26786232e-01 -4.83613133e-01 -1.13913977e+00
4.45953906e-02 -4.48339760e-01 3.07801872e-01 -6.60486460e-01
-1.22513759e+00 9.22523022e-01 -3.34162116e-01 -1.25899076e+00
-7.52245188e-01 -8.17380965e-01 -1.04472673e+00 1.07295620e+00
-2.02567649e+00 -5.26024342e-01 5.78927457e-01 1.49250060e-01
5.06591499e-01 -9.41783309e-01 4.94431734e-01 1.29530132e-01
-6.79456234e-01 -1.74346983e-01 6.03835106e-01 4.67330813e-01
2.67992392e-02 -1.34731078e+00 7.75523901e-01 5.61770201e-01
2.99269129e-02 3.33431780e-01 8.39181662e-01 -6.85978174e-01
-6.75501883e-01 -9.25375879e-01 1.31547868e+00 -1.75922319e-01
1.32782471e+00 4.62158144e-01 -9.94009018e-01 1.12482917e+00
4.25411642e-01 -2.64681846e-01 3.95344138e-01 -2.04357848e-01
-1.96271971e-01 -1.92677364e-01 -9.03247178e-01 1.12578444e-01
-3.78188372e-01 -3.19511354e-01 -1.12245607e+00 -2.60829646e-02
5.61966419e-01 -9.22561586e-02 -1.01154280e+00 2.54178375e-01
7.22453892e-01 -1.38708806e+00 6.20507240e-01 -1.65505573e-01
4.78381634e-01 1.54920384e-01 -7.91185275e-02 -1.43315911e+00
-3.25880647e-01 -3.24198812e-01 -2.48395540e-02 9.40568805e-01
9.91965532e-01 -1.21837258e+00 7.84732282e-01 6.80771351e-01
-4.01337929e-02 -8.36289585e-01 -7.90767729e-01 -7.23572552e-01
4.46626872e-01 -3.53989184e-01 9.04092014e-01 1.11944199e+00
2.47375116e-01 1.18282139e-01 -4.32919860e-01 -2.30128348e-01
7.42592439e-02 2.17972562e-01 6.87442198e-02 -1.33288026e+00
-9.37709026e-03 -1.13907921e+00 -3.98998797e-01 -5.58004498e-01
6.04032800e-02 -5.90213776e-01 -4.83514786e-01 -1.42564368e+00
-3.16350788e-01 1.26462663e-02 -1.06560349e+00 2.60169804e-01
4.45787549e-01 1.56332925e-01 2.26443276e-01 5.87723434e-01
-5.33663556e-02 3.06992054e-01 7.63044119e-01 1.15486067e-02
-3.99118066e-01 4.21995044e-01 -3.67803544e-01 9.08303440e-01
1.03279591e+00 -1.94804937e-01 1.17768310e-01 4.76960614e-02
1.06819832e+00 1.91391155e-01 -1.67777296e-02 -1.03327024e+00
3.85686576e-01 -4.87616397e-02 6.36857390e-01 -1.03619838e+00
-2.50712652e-02 -5.41015089e-01 2.20999658e-01 9.09741163e-01
-2.93017507e-01 8.81608665e-01 3.14862609e-01 1.14660650e-01
-9.52899158e-01 -3.32142472e-01 6.70986593e-01 -4.90335137e-01
-7.19166219e-01 3.38158533e-02 -6.47664487e-01 -4.13145214e-01
8.74163866e-01 -4.41000819e-01 -7.01038018e-02 -3.47003490e-01
-6.51707470e-01 2.39808798e-01 -3.00175846e-02 4.75664198e-01
6.43996477e-01 -1.26912475e+00 -8.58314037e-01 2.84499913e-01
-4.31336999e-01 -5.33181310e-01 -1.17581494e-01 6.58921778e-01
-9.70488369e-01 8.42053235e-01 -4.80241299e-01 2.94883214e-02
-5.88542223e-01 3.83225739e-01 7.22329915e-01 -4.03903872e-01
-5.53644657e-01 1.00264931e+00 -5.16968310e-01 -6.50324449e-02
1.47252753e-01 -9.12569821e-01 -9.18534577e-01 8.61641407e-01
7.97491789e-01 4.13712531e-01 7.86150470e-02 -8.25316906e-01
-5.74219003e-02 5.77365816e-01 -1.43129870e-01 -3.21350694e-01
1.92706847e+00 1.77016743e-02 -4.35659856e-01 1.41096532e+00
1.31099558e+00 -4.31669176e-01 -7.38996983e-01 -2.72020161e-01
9.07554448e-01 1.80842638e-01 5.74637234e-01 -8.06739330e-01
-1.21988571e+00 5.10330737e-01 5.88961542e-01 8.90935183e-01
9.65418577e-01 -5.51350057e-01 1.42086983e+00 5.26959598e-01
-2.10866988e-01 -1.42996764e+00 -2.09403068e-01 8.30832660e-01
8.64068270e-01 -1.30228460e+00 -2.88624078e-01 7.94422984e-01
-6.41051292e-01 1.49200594e+00 -2.05974698e-01 -6.84002519e-01
1.48156786e+00 3.43184948e-01 3.15320134e-01 -2.89251626e-01
-1.17444110e+00 2.09613726e-01 6.26908019e-02 -4.17935699e-01
5.51527739e-01 -1.62473153e-02 -3.48021388e-02 5.31440496e-01
-8.04731905e-01 3.16769719e-01 6.12836778e-01 9.75923717e-01
-1.00442147e+00 -6.92264557e-01 -5.43648481e-01 7.29176223e-01
-1.16221869e+00 -1.90882310e-01 -1.59455165e-01 6.53953373e-01
-2.92508572e-01 4.51949626e-01 6.43473864e-01 -2.71478713e-01
4.09644604e-01 3.72676402e-01 -4.23949927e-01 -2.51701385e-01
-8.69477391e-01 2.01202556e-01 -8.26188102e-02 1.44356778e-02
-5.18257201e-01 -7.83806026e-01 -1.23818600e+00 -6.08118951e-01
-7.26870149e-02 2.44964212e-01 8.22375596e-01 1.25759649e+00
-6.67235106e-02 5.08030593e-01 8.89259577e-01 -7.15862036e-01
-8.24324191e-01 -1.25360084e+00 -1.11500931e+00 -1.01918496e-01
8.75600696e-01 -2.66598791e-01 -4.80456591e-01 2.07320116e-02] | [4.449534893035889, 4.275467395782471] |
930efa30-4e32-4227-82cc-148da3c2d8a5 | positive-unlabeled-classification-under-class-1 | 2107.05045 | null | https://arxiv.org/abs/2107.05045v2 | https://arxiv.org/pdf/2107.05045v2.pdf | Positive-Unlabeled Classification under Class-Prior Shift: A Prior-invariant Approach Based on Density Ratio Estimation | Learning from positive and unlabeled (PU) data is an important problem in various applications. Most of the recent approaches for PU classification assume that the class-prior (the ratio of positive samples) in the training unlabeled dataset is identical to that of the test data, which does not hold in many practical cases. In addition, we usually do not know the class-priors of the training and test data, thus we have no clue on how to train a classifier without them. To address these problems, we propose a novel PU classification method based on density ratio estimation. A notable advantage of our proposed method is that it does not require the class-priors in the training phase; class-prior shift is incorporated only in the test phase. We theoretically justify our proposed method and experimentally demonstrate its effectiveness. | ['Masashi Sugiyama', 'Shota Nakajima'] | 2021-07-11 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 2.51406848e-01 -1.16512559e-01 -6.35722637e-01 -3.36393297e-01
-5.66683173e-01 -4.58961487e-01 3.17668736e-01 3.68990928e-01
-3.66034001e-01 1.19579077e+00 -4.61927295e-01 -4.13576722e-01
1.28661931e-01 -1.08898163e+00 -6.99585438e-01 -8.74893546e-01
6.07237995e-01 4.03083473e-01 5.35693049e-01 2.21051186e-01
3.36371884e-02 1.40507624e-01 -1.43947458e+00 -1.31934792e-01
1.17565858e+00 1.24931335e+00 2.31035858e-01 4.61075962e-01
-1.01702683e-01 7.23911047e-01 -6.51061833e-01 -1.57408595e-01
3.34641963e-01 -7.30823755e-01 -4.70353603e-01 2.36587688e-01
-1.28980100e-01 -1.76413476e-01 -3.62007022e-02 1.43189561e+00
2.68730074e-01 9.49072391e-02 1.34024668e+00 -1.32168603e+00
-4.91662353e-01 2.75304556e-01 -1.02609980e+00 2.91461170e-01
5.71226440e-02 -4.88503963e-01 9.70961511e-01 -6.57979548e-01
2.49439925e-01 7.41798043e-01 3.90437454e-01 4.04236585e-01
-8.54170620e-01 -5.97288609e-01 1.20511437e-02 4.82014567e-01
-1.46974134e+00 -2.66739313e-04 8.67488146e-01 -4.08059269e-01
2.78956722e-02 1.68310478e-02 6.72531486e-01 8.47641289e-01
8.23723674e-02 1.04247141e+00 1.18464828e+00 -7.73247182e-01
5.88771641e-01 6.04530931e-01 5.28660178e-01 3.09239179e-01
5.41869521e-01 -1.02801919e-01 4.93027829e-02 -2.29832292e-01
6.41688347e-01 7.55413994e-02 -4.95691895e-01 -7.03407824e-01
-7.50813365e-01 6.69841290e-01 -6.26400858e-03 2.36894384e-01
-1.71620309e-01 -4.93070334e-01 1.25995830e-01 1.18572310e-01
3.33059609e-01 -3.39253604e-01 -3.54073197e-01 6.22167997e-02
-6.75717294e-01 -1.19263321e-01 9.59302068e-01 1.17687106e+00
8.32906425e-01 -2.27700397e-01 -1.83001887e-02 8.52984488e-01
4.07230735e-01 6.02833211e-01 4.75683838e-01 -1.91928461e-01
3.06449950e-01 3.73009384e-01 3.25644612e-01 -5.95853388e-01
2.82615349e-02 -4.67031717e-01 -6.07374489e-01 3.92040983e-03
7.90345490e-01 -3.12797606e-01 -8.38975489e-01 1.46637082e+00
4.37588513e-01 4.38028932e-01 2.65235484e-01 5.99836349e-01
5.43861151e-01 6.91874087e-01 4.06470001e-02 -5.88310599e-01
8.90466154e-01 -6.88703001e-01 -6.49991691e-01 -1.57378659e-01
4.85604435e-01 -8.10703278e-01 8.87717009e-01 4.22359169e-01
-4.51517999e-01 -4.79106873e-01 -1.51735926e+00 4.04309660e-01
-1.13049015e-01 4.11572367e-01 3.33150744e-01 9.71072912e-01
-3.31105143e-01 3.45354974e-01 -7.33747900e-01 -3.87428224e-01
2.44006500e-01 4.50126469e-01 -2.01364085e-01 -1.99935392e-01
-1.03932869e+00 6.57019317e-01 8.36872578e-01 2.50109881e-02
-5.73678732e-01 -2.01739937e-01 -7.11208820e-01 2.85597354e-01
4.31112051e-01 -1.17824011e-01 1.50142157e+00 -9.47191298e-01
-1.43668461e+00 3.74884903e-01 -2.55331606e-01 -1.25002474e-01
7.10501432e-01 7.27677420e-02 -4.29234296e-01 5.78021482e-02
-1.95476323e-01 2.55509526e-01 7.88215518e-01 -1.28132963e+00
-9.73189056e-01 -2.04995885e-01 -1.44629076e-01 1.64892226e-01
-3.12493354e-01 -4.28004950e-01 -3.65502298e-01 -4.06643271e-01
5.04338264e-01 -6.40668690e-01 1.37501404e-01 3.22157927e-02
-3.47247571e-01 -3.65093619e-01 1.29460096e+00 -2.98131317e-01
9.70188320e-01 -2.25469089e+00 -5.89841306e-01 4.04680252e-01
-8.13998953e-02 3.60716701e-01 2.49078691e-01 1.41689792e-01
2.36299597e-02 -1.86617151e-01 -3.73068959e-01 2.57646274e-02
-9.81890038e-02 3.47774625e-01 -3.62699658e-01 7.32440650e-01
2.98953593e-01 3.24900389e-01 -1.14035821e+00 -8.23116899e-01
2.35699102e-01 2.97232121e-01 -3.50155622e-01 2.47832969e-01
-1.20279469e-01 6.01067424e-01 -5.99416792e-01 4.39232916e-01
1.05556297e+00 -3.27148855e-01 2.22513914e-01 -6.39284924e-02
1.95439994e-01 1.02625377e-01 -1.48713446e+00 7.93649316e-01
-7.51331896e-02 3.76800597e-01 -2.59291410e-01 -1.33012962e+00
1.04853761e+00 4.49990511e-01 3.93645942e-01 -4.75237370e-01
4.52260226e-01 2.69394785e-01 2.00784892e-01 -2.29680315e-01
2.18888372e-01 -5.34506679e-01 2.36600175e-01 6.34487271e-01
1.29995391e-01 9.51853022e-02 4.26396817e-01 -1.75160095e-01
6.32384717e-01 1.07479431e-01 6.22045696e-01 -3.51074725e-01
6.95176005e-01 -3.36234868e-01 1.01483500e+00 6.61058068e-01
-4.13064986e-01 5.74370861e-01 7.19905615e-01 1.45470381e-01
-7.63257742e-01 -1.23341763e+00 -5.46420455e-01 4.83032614e-01
4.94248897e-01 5.38466498e-02 -6.80609882e-01 -1.09867370e+00
-2.20774725e-01 5.85605681e-01 -4.56044912e-01 -4.09965925e-02
-1.34042472e-01 -1.06475854e+00 -9.29076299e-02 6.17587507e-01
4.93957222e-01 -5.49303412e-01 -2.66337991e-01 5.43479547e-02
-1.20881759e-01 -9.96163011e-01 -2.40549147e-01 2.89562732e-01
-9.29975927e-01 -1.47343087e+00 -7.03373909e-01 -1.02896929e+00
1.10652447e+00 4.45946366e-01 6.25691056e-01 -2.95078903e-01
2.18519509e-01 2.59651810e-01 -5.68337917e-01 -7.05622494e-01
-5.05264997e-01 -1.42115280e-01 4.71290685e-02 2.75915951e-01
6.70641005e-01 -3.03891778e-01 -3.64423305e-01 6.05377078e-01
-7.58798420e-01 -2.91952044e-01 4.94885743e-01 1.07771873e+00
7.03964233e-01 4.86772627e-01 8.98522794e-01 -1.10722792e+00
2.06473351e-01 -5.62907517e-01 -8.16358507e-01 6.36341035e-01
-3.43276262e-01 -8.53371620e-02 6.56356215e-01 -7.41868973e-01
-1.19920599e+00 2.65631173e-02 -1.30246565e-01 -2.46823132e-01
-1.20885603e-01 3.26742887e-01 -6.07515693e-01 -4.84129973e-02
5.67147851e-01 8.17986801e-02 -1.16615236e-01 -4.47865158e-01
-8.42103213e-02 9.43205833e-01 4.60647494e-01 -7.81806290e-01
6.57831430e-01 3.91551971e-01 -8.43671039e-02 -1.06737328e+00
-9.24493074e-01 -6.44820869e-01 -7.35314250e-01 -5.92467152e-02
5.20410180e-01 -7.17063606e-01 -5.13988197e-01 3.34528476e-01
-8.06033671e-01 -1.34834826e-01 -3.35690290e-01 1.03018379e+00
-5.08169293e-01 8.01616788e-01 -3.11538517e-01 -1.03802049e+00
-7.77989179e-02 -1.02826726e+00 3.53287131e-01 5.87108612e-01
6.61511496e-02 -1.12558889e+00 -6.49398789e-02 8.33821855e-03
7.95958191e-03 -6.54160306e-02 1.05905581e+00 -1.03004849e+00
-3.01998198e-01 -5.10284901e-01 -3.94399881e-01 6.79526269e-01
4.05795038e-01 -4.34991829e-02 -1.15272355e+00 -2.50710577e-01
2.91270167e-01 -3.61499161e-01 5.36816895e-01 4.19322222e-01
1.09436703e+00 1.44226685e-01 -4.25899297e-01 -5.69856651e-02
1.29422033e+00 5.38177133e-01 5.57572663e-01 -2.79424906e-01
4.53392178e-01 4.24344778e-01 1.08947778e+00 6.27598941e-01
-9.34937671e-02 3.20013165e-01 2.07044467e-01 9.50747579e-02
2.58136928e-01 -3.26779038e-01 8.26406777e-02 6.43467069e-01
2.27368250e-01 -5.24460435e-01 -6.65433705e-01 3.90997320e-01
-1.79076540e+00 -8.50703359e-01 -1.20926410e-01 2.77725172e+00
8.17781627e-01 3.31000090e-01 -4.22969535e-02 7.70436704e-01
1.13068759e+00 -2.43639126e-01 -4.96636301e-01 2.63733983e-01
1.21326990e-01 7.61983618e-02 3.60768110e-01 4.33127165e-01
-1.33447719e+00 2.78570861e-01 6.56072140e+00 1.03688037e+00
-1.22130823e+00 -2.06066035e-02 8.00251365e-01 4.65093851e-01
-6.07243367e-02 5.43524362e-02 -9.32803631e-01 4.28224802e-01
5.41232824e-01 -2.35864490e-01 -2.40015537e-01 1.00789082e+00
-1.45641819e-01 -6.97736263e-01 -1.35334253e+00 9.85508919e-01
7.79287424e-04 -6.48784399e-01 -3.00250113e-01 -1.02024889e-02
6.48223817e-01 -3.86625648e-01 -2.79675983e-02 5.49652755e-01
4.39181551e-02 -4.51690376e-01 7.36650944e-01 1.22392409e-01
6.76044464e-01 -8.18228006e-01 1.15641904e+00 7.16505826e-01
-9.62701738e-01 -1.44018773e-02 -6.38953567e-01 -1.24512337e-01
-1.14656322e-01 8.24820399e-01 -1.20368755e+00 5.19133449e-01
2.40669280e-01 7.37593830e-01 -2.33750954e-01 1.45943558e+00
-2.38074496e-01 9.47846413e-01 -4.26285177e-01 -7.55566806e-02
-2.53648341e-01 -3.89657915e-01 2.64272273e-01 7.66788960e-01
5.30148327e-01 7.03473983e-04 4.07619536e-01 6.56251371e-01
-6.76719751e-03 4.62424994e-01 -4.34428632e-01 -5.72832339e-02
4.27810103e-01 9.67376351e-01 -1.12444389e+00 -4.40078169e-01
-8.04335117e-01 5.51082671e-01 9.95424986e-02 3.62845868e-01
-7.70527124e-01 -2.20125750e-01 -1.29618168e-01 1.80649664e-02
3.33326340e-01 9.92903709e-02 -1.21075116e-01 -1.28284919e+00
8.10320377e-02 -3.14727515e-01 4.76216465e-01 -3.02330583e-01
-1.65952241e+00 6.95554614e-02 7.52088875e-02 -1.77091992e+00
-2.27774531e-01 -7.39211202e-01 -6.68560505e-01 8.46787989e-01
-1.51859391e+00 -8.02336633e-01 -4.62510623e-02 4.63433504e-01
1.80480003e-01 3.48886773e-02 6.52164340e-01 4.31815863e-01
-6.51344597e-01 6.37291372e-01 4.54282016e-01 3.84301513e-01
8.79854679e-01 -1.17888916e+00 -3.53178293e-01 8.30424786e-01
-7.77055398e-02 4.04541969e-01 8.14108551e-01 -5.40798008e-01
-8.82871449e-01 -8.39231789e-01 5.82022727e-01 -6.17633387e-02
6.34219646e-01 -8.73030573e-02 -8.86671662e-01 5.91435730e-01
-1.31675109e-01 5.19333482e-01 1.05883646e+00 5.36630023e-03
-2.21563447e-02 -2.63834476e-01 -1.30028164e+00 1.99079201e-01
2.85516858e-01 -1.66775927e-01 -5.54388702e-01 1.23376742e-01
1.30857795e-01 -3.46664101e-01 -6.90504909e-01 6.37991130e-01
5.73037446e-01 -8.79987419e-01 4.20021951e-01 -8.96762833e-02
-7.36475512e-02 -5.59696972e-01 -1.57638699e-01 -1.09968543e+00
1.06004015e-01 1.67327330e-01 -1.57345131e-01 1.18145680e+00
3.74286473e-01 -9.42319334e-01 1.08234239e+00 4.30015236e-01
5.57777397e-02 -5.76103151e-01 -8.36876810e-01 -7.64819741e-01
-1.06311239e-01 -3.48733902e-01 2.12524086e-01 9.32822108e-01
2.32140020e-01 3.56389642e-01 -2.01722026e-01 2.06459329e-01
6.02137566e-01 7.47159645e-02 5.53590000e-01 -1.60121679e+00
-6.37484848e-01 -4.74184155e-02 -5.29688656e-01 -1.02510560e+00
7.62051195e-02 -5.36059976e-01 1.79207474e-01 -1.39712596e+00
5.23889601e-01 -8.19481015e-01 -5.53211868e-01 3.58415127e-01
-3.55915964e-01 9.87972245e-02 -2.51714110e-01 2.60747641e-01
-3.46066147e-01 7.36495674e-01 1.18856740e+00 -1.32276481e-02
-1.96668997e-01 5.19025803e-01 -3.10627520e-01 7.58897305e-01
8.28817725e-01 -5.01290619e-01 -9.08896863e-01 2.00690821e-01
-1.77509636e-01 2.16692425e-02 -5.56734204e-02 -1.20302176e+00
-3.74886841e-02 -2.41690502e-01 5.57496727e-01 -8.28159809e-01
8.67582709e-02 -1.14180756e+00 -2.05840513e-01 3.29464376e-01
2.22359136e-01 -5.82273781e-01 3.67889926e-03 9.15547252e-01
-3.38338971e-01 -7.97798872e-01 1.03302801e+00 2.15693980e-01
-4.53552634e-01 3.68747830e-01 -2.40285888e-01 -1.18838124e-01
1.46266472e+00 -2.59464860e-01 -3.09307277e-01 -3.26153874e-01
-5.56086481e-01 1.45227179e-01 5.61487973e-01 -5.78561611e-02
3.51378739e-01 -1.24212646e+00 -2.29057923e-01 3.16056013e-01
2.30227008e-01 1.83213681e-01 2.30283186e-01 8.06618094e-01
-5.07783949e-01 3.13475281e-01 -1.63614675e-01 -7.01377690e-01
-1.03002357e+00 7.38327384e-01 1.39922291e-01 -1.42311648e-01
-2.88938433e-01 4.23116028e-01 3.85661125e-01 -2.31498241e-01
2.95681179e-01 -9.40597281e-02 -5.06391168e-01 8.18021670e-02
7.02007473e-01 3.19027543e-01 1.10559876e-03 -4.03677613e-01
1.46748453e-01 2.96855867e-01 -2.38021210e-01 -7.34145045e-02
9.22738194e-01 1.57821432e-01 4.64110225e-02 7.99296379e-01
1.14044487e+00 1.03479870e-01 -1.24490070e+00 -3.79415035e-01
-6.85881823e-02 -4.12618399e-01 -2.35147506e-01 -1.35438725e-01
-7.81819403e-01 1.13703322e+00 6.84990048e-01 1.67040795e-01
1.03539598e+00 -1.65169463e-01 4.46565092e-01 4.03056473e-01
5.44963598e-01 -1.25884569e+00 -9.63213518e-02 2.69150287e-01
1.61578417e-01 -1.54014516e+00 -6.41728491e-02 -8.67920518e-01
-3.23517859e-01 1.10344863e+00 7.12979436e-01 -7.14808237e-03
9.95234311e-01 1.13197684e-01 7.79436231e-02 4.07832742e-01
-4.25117016e-01 -1.64940551e-01 1.91004008e-01 6.04561031e-01
5.63824654e-01 1.44077772e-02 -6.65955424e-01 5.74430287e-01
2.42344379e-01 4.68139462e-02 6.38443828e-01 1.37964094e+00
-6.02198601e-01 -1.39956510e+00 -5.80990374e-01 6.57950699e-01
-3.94634098e-01 3.35301191e-01 6.38810098e-02 6.33445382e-01
1.39660060e-01 9.07115102e-01 4.51098122e-02 -2.26810381e-01
4.11678106e-02 2.56021619e-01 6.30342484e-01 -8.75448644e-01
4.46577042e-01 2.57564694e-01 -2.30632871e-01 2.76253104e-01
-4.17655438e-01 -7.43921816e-01 -1.36245728e+00 -3.33390199e-02
-9.01852429e-01 6.37955964e-01 3.14336658e-01 1.01311171e+00
-1.81601346e-01 1.75183639e-01 6.39641166e-01 -5.81010640e-01
-7.17414379e-01 -1.22507787e+00 -1.08807576e+00 -3.40383649e-02
2.36115642e-02 -9.98402596e-01 -4.70131546e-01 8.31958726e-02] | [9.120147705078125, 3.9715349674224854] |
6733282f-0f70-41d4-94c3-bf8ae389db40 | attention-modulation-for-zero-shot-cross | null | null | https://aclanthology.org/2022.codi-1.11 | https://aclanthology.org/2022.codi-1.11.pdf | Attention Modulation for Zero-Shot Cross-Domain Dialogue State Tracking | Dialog state tracking (DST) is a core step for task-oriented dialogue systems aiming to track the user’s current goal during a dialogue. Recently a special focus has been put on applying existing DST models to new domains, in other words performing zero-shot cross-domain transfer. While recent state-of-the-art models leverage large pre-trained language models, no work has been made on understanding and improving the results of first developed zero-shot models like SUMBT. In this paper, we thus propose to improve SUMBT zero-shot results on MultiWOZ by using attention modulation during inference. This method improves SUMBT zero-shot results significantly on two domains and does not worsen the initial performance with the great advantage of needing no additional training. | ['Sophie Rosset', 'Guillaume Bernard', 'Olivier Galibert', 'Mathilde Veron'] | null | null | null | null | coling-codi-crac-2022-10 | ['dialogue-state-tracking', 'task-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing'] | [-7.35963061e-02 4.39404100e-01 -9.99078006e-02 -4.40279305e-01
-5.12097180e-01 -2.80331850e-01 1.20619500e+00 2.03868672e-01
-7.01838672e-01 8.84747028e-01 5.14413238e-01 -2.14506686e-01
1.12230383e-01 -5.62022567e-01 8.38238671e-02 -1.89934358e-01
2.66695172e-01 9.81846452e-01 7.43159831e-01 -9.15359378e-01
2.75329441e-01 4.19217348e-02 -8.99677575e-01 2.30261073e-01
7.85395920e-01 4.03815269e-01 2.10849449e-01 8.77097249e-01
-5.52223742e-01 7.25260854e-01 -9.34196055e-01 -5.68969548e-01
-1.44487366e-01 -8.69779408e-01 -1.32397878e+00 3.02726571e-02
1.45431861e-01 -4.55745399e-01 -3.97701800e-01 7.74208784e-01
7.52217591e-01 6.86036587e-01 5.25822639e-01 -9.35775101e-01
-3.54673594e-01 5.54641604e-01 -1.22608908e-01 2.36966357e-01
6.55871212e-01 2.44940892e-01 8.54204655e-01 -4.85557497e-01
7.90263236e-01 1.44566393e+00 5.43402910e-01 1.19389617e+00
-1.25259185e+00 -2.25219727e-01 1.41034350e-01 1.94986835e-01
-7.27372468e-01 -5.82608223e-01 7.38073826e-01 -2.44954422e-01
1.29177630e+00 9.45334285e-02 2.99486548e-01 1.21978927e+00
-7.58671016e-02 1.01539588e+00 1.17711771e+00 -7.26735175e-01
2.17101425e-01 4.05576020e-01 3.55874479e-01 5.25863290e-01
-4.63735819e-01 -4.17446345e-01 -5.12730241e-01 -8.26519430e-02
4.65874344e-01 -3.65609795e-01 -4.73008566e-02 -3.61574650e-01
-9.89519596e-01 1.04142904e+00 3.20779569e-02 9.62888122e-01
2.08372977e-02 -2.84002572e-01 7.64423609e-01 5.20514548e-01
7.56956637e-01 6.09894097e-01 -5.00287056e-01 -7.33546734e-01
-6.95679188e-01 3.11026990e-01 1.23930383e+00 7.46153057e-01
4.78076905e-01 -7.01528862e-02 -7.61209130e-01 1.21270502e+00
-7.32461363e-02 6.05646186e-02 8.02657187e-01 -8.23595166e-01
6.57107651e-01 6.35705411e-01 2.16670245e-01 -3.64555508e-01
-5.04312158e-01 2.15123728e-01 -3.84362847e-01 8.09164494e-02
7.50940025e-01 -5.72538733e-01 -8.47088873e-01 1.73710775e+00
1.82078168e-01 -1.02055818e-01 2.99598604e-01 7.08731353e-01
8.40003610e-01 7.10493684e-01 1.30386740e-01 -4.03954759e-02
1.38808298e+00 -1.01430166e+00 -1.06004322e+00 -4.43061829e-01
9.12124276e-01 -5.65796554e-01 1.25043917e+00 2.02014759e-01
-1.11384475e+00 -5.01620293e-01 -9.33143020e-01 -8.70178342e-02
-6.18639708e-01 -4.20673490e-01 5.17450273e-01 9.41594899e-01
-9.03024375e-01 7.13340104e-01 -6.25615537e-01 -7.92347193e-01
1.35833263e-01 2.33665034e-01 -3.00825268e-01 2.21302137e-02
-1.64480996e+00 1.62566638e+00 4.75085318e-01 -4.92814690e-01
-6.12806320e-01 -4.70279723e-01 -9.12351191e-01 1.91747025e-01
6.86105609e-01 -3.77676338e-01 1.88639557e+00 -5.15952706e-01
-2.39830709e+00 7.97523141e-01 -1.87534049e-01 -6.79999292e-01
5.18309414e-01 -2.98565894e-01 -3.09088707e-01 -1.21204421e-01
-1.54170766e-01 6.04113221e-01 5.53981781e-01 -8.74619007e-01
-5.93148887e-01 -2.35813051e-01 3.61416668e-01 2.78030664e-01
-6.58962846e-01 1.52140781e-01 -3.21697086e-01 -2.65377611e-01
-6.10631764e-01 -8.29539537e-01 -3.05397034e-01 -3.20246100e-01
-5.65798432e-02 -7.63100684e-01 1.00669146e+00 -5.73724389e-01
1.21013319e+00 -1.85125458e+00 3.30564588e-01 -6.64746106e-01
-7.44811445e-02 1.11265504e+00 -6.75601512e-02 7.86575675e-01
2.75357008e-01 -2.04410747e-01 -2.39997700e-01 -4.81608361e-01
2.21261412e-01 3.14886510e-01 -1.08603567e-01 -2.22568982e-03
2.98211604e-01 8.93551826e-01 -1.06643558e+00 -4.24922764e-01
6.60402179e-01 4.38144989e-02 -6.17869794e-01 5.09754181e-01
-5.20229816e-01 5.74689329e-01 -2.87018836e-01 -2.19714135e-01
4.11595553e-01 -1.00221686e-01 2.83425242e-01 3.54781091e-01
-2.52210647e-01 8.52369189e-01 -7.71429777e-01 2.14980412e+00
-7.36184537e-01 6.01641893e-01 -8.81533697e-02 -1.03681898e+00
9.23494816e-01 5.80698550e-01 2.40081072e-01 -7.12718487e-01
3.65255475e-01 -2.29146555e-01 2.73984402e-01 -3.83643866e-01
6.44673884e-01 -4.82531101e-01 -4.07006234e-01 5.00227273e-01
5.94452202e-01 -1.60786450e-01 1.95998356e-01 2.54777640e-01
8.60502422e-01 1.15555622e-01 6.05888247e-01 -1.93463698e-01
6.66076243e-01 1.41901553e-01 1.36859745e-01 5.83486974e-01
-3.92532498e-01 3.98753971e-01 6.01833940e-01 9.48017091e-03
-7.62714267e-01 -7.32218146e-01 1.36578664e-01 1.42769194e+00
-1.05149068e-01 -3.68208528e-01 -9.85983968e-01 -8.77256572e-01
-1.57164320e-01 1.27355921e+00 -6.09548450e-01 -4.29986298e-01
-5.85322022e-01 -4.12703782e-01 6.64309740e-01 3.31993341e-01
7.67813742e-01 -1.11587059e+00 -8.02323222e-01 5.07704198e-01
-2.61126876e-01 -1.05760705e+00 -4.90232646e-01 2.51819640e-01
-7.89598227e-01 -6.44189417e-01 -1.17945004e+00 -4.43740278e-01
-3.68965454e-02 1.72359630e-01 9.97136354e-01 -4.56584036e-01
-6.95882738e-02 3.75133157e-01 -4.51725602e-01 -3.06352854e-01
-7.03113377e-01 4.21820253e-01 4.47024498e-03 -1.48379117e-01
6.03823364e-01 -1.81068107e-01 -2.22255066e-01 4.10799146e-01
-5.13738930e-01 -2.03986522e-02 2.33467177e-01 9.99334931e-01
-5.64370394e-01 -5.51983416e-01 8.86642158e-01 -1.07960379e+00
1.14723694e+00 -1.93273142e-01 -3.71538669e-01 2.55413681e-01
-5.05559206e-01 3.67618203e-01 5.89742005e-01 -4.98445153e-01
-1.65099311e+00 -3.66748393e-01 -3.49532306e-01 -2.64433652e-01
-3.15405488e-01 2.08824798e-01 -1.01411350e-01 1.46438837e-01
5.82325161e-01 1.30744800e-01 8.58723372e-02 -6.54256344e-01
6.36870563e-01 9.52844799e-01 3.06450099e-01 -4.49871808e-01
2.78710812e-01 1.04085572e-01 -4.79367614e-01 -9.86109316e-01
-8.82515550e-01 -7.91736424e-01 -8.86992455e-01 -1.17688999e-01
1.07485342e+00 -4.16636914e-01 -8.32817972e-01 4.57516372e-01
-1.08410394e+00 -7.66275465e-01 -2.36452937e-01 1.62766218e-01
-6.41309083e-01 6.23201787e-01 -6.72255754e-01 -1.00864840e+00
-2.93057382e-01 -8.58861566e-01 7.24904418e-01 2.08203390e-01
-7.01690495e-01 -1.39066792e+00 4.02034312e-01 4.43094999e-01
5.81178784e-01 -4.35543984e-01 6.91007435e-01 -1.11168551e+00
-6.31123036e-02 -1.43366754e-01 -4.15120982e-02 2.90703416e-01
2.69116998e-01 -6.54401064e-01 -1.20532000e+00 -2.17347205e-01
1.42091393e-01 -5.02168775e-01 6.23955131e-01 7.84390941e-02
3.32937717e-01 -5.66753410e-02 -1.92126945e-01 -7.33185858e-02
7.53479779e-01 4.17739809e-01 5.81646740e-01 2.12937891e-01
2.37473249e-01 8.20680797e-01 9.90221143e-01 4.56823200e-01
3.39328557e-01 1.13786173e+00 -5.38813621e-02 1.55647188e-01
-2.83930063e-01 -1.88746527e-01 4.79109675e-01 6.01988792e-01
1.36851326e-01 -4.72718239e-01 -9.20882106e-01 6.97574258e-01
-2.03700638e+00 -9.64761317e-01 1.81008518e-01 2.05501485e+00
1.00920212e+00 4.45374936e-01 2.07105532e-01 -2.43714213e-01
5.63591540e-01 5.05261481e-01 -4.27485049e-01 -7.14899242e-01
3.36744249e-01 2.79644221e-01 -2.16243020e-03 7.18810678e-01
-1.01459408e+00 1.44501698e+00 5.95060682e+00 7.19603360e-01
-1.01441360e+00 3.89270008e-01 1.97093248e-01 8.59966129e-02
2.37161651e-01 1.01817869e-01 -9.73487914e-01 2.69092143e-01
1.23969948e+00 -3.69032800e-01 6.20464832e-02 9.37176466e-01
2.21965387e-02 -2.84757495e-01 -8.85113418e-01 5.63193738e-01
2.94775635e-01 -9.33368802e-01 -2.67867625e-01 -8.79314318e-02
3.95605743e-01 -2.19421342e-01 -2.60870934e-01 9.07235205e-01
6.81542933e-01 -6.29375339e-01 4.38765362e-02 3.10236543e-01
6.08976007e-01 -5.47219276e-01 7.70586729e-01 6.79077506e-01
-6.13924444e-01 1.78233445e-01 -3.29704374e-01 -2.90812999e-01
4.02150124e-01 -1.35175571e-01 -1.56120276e+00 4.53437179e-01
3.10851276e-01 4.70638692e-01 -2.81501859e-01 7.66461790e-01
-1.21762350e-01 4.84926343e-01 -3.08467075e-03 -4.48882729e-01
5.33046901e-01 -5.14596887e-02 5.98625064e-01 1.40980983e+00
-2.89582424e-02 2.55066101e-02 3.13820928e-01 6.00765765e-01
4.25167494e-02 -4.44827639e-02 -6.46064639e-01 -6.23300672e-02
9.77119356e-02 1.06823540e+00 -4.18643981e-01 -7.45841861e-01
-6.37381911e-01 1.39988768e+00 3.44936877e-01 -1.51580036e-01
-5.86221278e-01 -4.82585788e-01 8.71895194e-01 -3.70255709e-02
3.49017709e-01 -2.88139075e-01 -1.77267864e-01 -1.22429025e+00
-5.85882843e-01 -6.02528095e-01 4.53571528e-01 -5.49656153e-01
-1.14770973e+00 5.76508939e-01 2.34431088e-01 -9.74055827e-01
-6.93208635e-01 -6.24192119e-01 -7.31019199e-01 1.05794013e+00
-1.33897817e+00 -9.77738857e-01 -5.78966998e-02 6.64247870e-01
1.24708462e+00 -2.31092155e-01 1.26689374e+00 4.26936001e-02
-4.90427792e-01 4.82735872e-01 3.48522700e-02 1.81429431e-01
1.30194652e+00 -1.45230472e+00 7.28257179e-01 5.93959033e-01
-1.58485383e-01 6.30992353e-01 1.03932357e+00 -6.04726076e-01
-9.35426652e-01 -6.58713281e-01 1.07717741e+00 -7.12748826e-01
8.25630307e-01 -6.03742182e-01 -1.21273685e+00 4.93736684e-01
8.73098493e-01 -4.81886148e-01 5.73268294e-01 5.06344140e-01
-9.40616876e-02 3.23546797e-01 -1.09277272e+00 7.55479336e-01
7.34407544e-01 -4.42034364e-01 -1.27113903e+00 1.86701864e-01
6.18301153e-01 -3.49488914e-01 -9.26319242e-01 4.87198941e-02
1.56139702e-01 -9.06874001e-01 7.81520605e-01 -8.88678193e-01
1.74602151e-01 2.49172300e-01 3.19036573e-01 -1.68626225e+00
-2.01803610e-01 -7.06107497e-01 -1.81982189e-01 1.19744492e+00
2.44543731e-01 -5.82203805e-01 8.30799818e-01 7.74317801e-01
-3.92949164e-01 -1.16970666e-01 -1.05968165e+00 -8.42568457e-01
2.42209256e-01 9.42035113e-03 1.60733372e-01 1.11867094e+00
8.35623503e-01 1.07397413e+00 -8.42247844e-01 -4.81594503e-01
1.71773106e-01 8.88816863e-02 1.09015656e+00 -1.34266818e+00
-2.47094646e-01 -4.54698086e-01 -5.03765084e-02 -1.32914901e+00
1.78193763e-01 -5.84771335e-01 1.75084785e-01 -1.51184809e+00
-5.08531891e-02 8.19702521e-02 1.29342107e-02 5.76792955e-01
-4.77777719e-01 -2.50262976e-01 4.58059877e-01 -1.37395218e-01
-8.83266807e-01 9.30820465e-01 1.31661344e+00 -9.25789624e-02
-3.81042391e-01 2.72089839e-01 -3.91254723e-01 5.32949567e-01
9.24262941e-01 -2.54573166e-01 -5.08346260e-01 -3.19983587e-02
-5.64717591e-01 4.50317860e-01 -8.31434131e-02 -9.68559027e-01
3.20496321e-01 -1.15420306e-02 -1.33498579e-01 -3.95209879e-01
8.55767310e-01 -5.02863407e-01 -6.40582085e-01 5.86146593e-01
-4.69994515e-01 -2.98428506e-01 4.51680601e-01 5.10881364e-01
-1.94567204e-01 -6.84491217e-01 9.93237376e-01 -3.49620610e-01
-1.00718713e+00 -1.70078471e-01 -8.15314114e-01 2.22272143e-01
9.72535312e-01 -1.07565433e-01 -3.18551660e-01 -5.78606367e-01
-7.27706730e-01 2.79986739e-01 1.92715183e-01 7.37650931e-01
2.83194244e-01 -8.05056393e-01 -5.17950654e-01 -1.43806428e-01
2.61700153e-01 -3.92000318e-01 2.70567864e-01 6.14102066e-01
-4.31707688e-02 9.55009580e-01 -4.91872251e-01 -2.97877908e-01
-1.58097970e+00 5.09448946e-01 2.24316150e-01 -7.56940544e-01
-7.33713806e-01 7.15598226e-01 -6.24452643e-02 -7.11354792e-01
1.20811716e-01 3.81423123e-02 -6.39823198e-01 2.77322173e-01
5.57784677e-01 2.10978687e-01 4.65818159e-02 -2.96479225e-01
-2.70418078e-02 4.39416133e-02 -5.94928265e-01 -4.18172061e-01
1.13971949e+00 -1.77173018e-01 3.84608328e-01 9.20743763e-01
8.96311462e-01 -3.54149431e-01 -1.24748158e+00 -4.21633542e-01
3.98128539e-01 -4.72712398e-01 -1.21835627e-01 -8.82903457e-01
-3.23625088e-01 1.29327238e+00 5.10545075e-01 4.33183819e-01
6.37653768e-01 -5.24359457e-02 9.51961040e-01 5.90759635e-01
5.67814767e-01 -1.21043634e+00 4.24998105e-01 1.19243026e+00
5.84034681e-01 -1.50526989e+00 -3.76938045e-01 -2.23263428e-01
-1.24387956e+00 9.42639649e-01 7.73437440e-01 1.08409762e-01
3.42894524e-01 -4.79553789e-02 1.56781584e-01 -6.21223338e-02
-8.49395454e-01 -5.93987882e-01 7.16528893e-02 6.28854811e-01
6.07506692e-01 -2.20021188e-01 -3.96874428e-01 4.26004052e-01
1.39178067e-01 1.03080317e-01 4.45563853e-01 1.04807377e+00
-7.60786414e-01 -1.53804576e+00 -1.27433380e-02 2.23075584e-01
-2.82319993e-01 -6.45140707e-02 -7.00379550e-01 8.21709573e-01
-5.64817071e-01 9.89310622e-01 -9.77441594e-02 -2.05895767e-01
6.11608982e-01 7.87060022e-01 3.79693985e-01 -1.03789127e+00
-9.85868156e-01 -1.44634619e-01 5.71927071e-01 -2.04486758e-01
-3.20438057e-01 -6.74452901e-01 -9.79648232e-01 -2.75645316e-01
-4.99999583e-01 3.58747572e-01 3.49687457e-01 1.01968491e+00
2.26298466e-01 7.26334214e-01 2.31368348e-01 -6.71961367e-01
-6.30695879e-01 -1.67027390e+00 -1.88693419e-01 4.73101437e-01
1.08238496e-01 -7.62353659e-01 7.24865124e-02 -2.82538950e-01] | [12.75997257232666, 7.919741630554199] |
62e3b811-2b7f-4a3d-86dc-cce62bf9f464 | feature-selection-and-classification-of | null | null | https://doi.org/10.1109/LGRS.2007.905116 | https://doi.org/10.1109/LGRS.2007.905116 | Feature Selection and Classification of Hyperspectral Images With Support Vector Machines | Hyperspectral images consist of large number of bands which require sophisticated analysis to extract. One approach to reduce computational cost, information representation, and accelerate knowledge discovery is to eliminate bands that do not add value to the classification and analysis method which is being applied. In particular, algorithms that perform band elimination should be designed to take advantage of the structure of the classification method used. This letter introduces an embedded-feature-selection (EFS) algorithm that is tailored to operate with support vector machines (SVMs) to perform band selection and classification simultaneously. We have successfully applied this algorithm to determine a reasonable subset of bands without any user-defined stopping criteria on some sample AVIRIS images; a problem occurs in benchmarking recursive-feature-elimination methods for the SVMs. | ['George Fann', 'Rick Archibald'] | 2007-10-15 | null | null | null | ieee-geoscience-and-remote-sensing-letters-11 | ['classification-of-hyperspectral-images', 'few-shot-image-classification'] | ['computer-vision', 'computer-vision'] | [ 9.15382922e-01 -4.40902114e-01 -6.11449825e-03 -5.62388301e-01
-3.07317317e-01 -6.13509059e-01 3.31647545e-01 9.22847092e-02
-1.99264571e-01 8.95349741e-01 -4.54447120e-01 -4.94485885e-01
-5.73100746e-01 -1.05242705e+00 -6.35252297e-02 -9.34287250e-01
-7.71918222e-02 8.26788843e-02 4.97077927e-02 -1.78095013e-01
5.25508821e-01 9.70551133e-01 -2.10453367e+00 4.37905461e-01
1.15621161e+00 1.18282139e+00 4.30457175e-01 7.02677667e-01
-1.62882462e-01 4.78931516e-01 -4.45233405e-01 2.04735130e-01
5.32913685e-01 -7.26876259e-01 -7.36746728e-01 5.55813432e-01
1.42533645e-01 4.59504016e-02 2.72450626e-01 1.20125556e+00
1.23716414e-01 2.72437662e-01 1.02734399e+00 -9.39492047e-01
8.01660419e-02 2.64980167e-01 -5.64597487e-01 2.93456346e-01
-1.70012578e-01 -5.27021363e-02 7.88991988e-01 -7.69528627e-01
3.10308725e-01 4.52225119e-01 4.98623818e-01 1.02178305e-02
-1.23599100e+00 -3.51938128e-01 -2.41422608e-01 3.15568119e-01
-1.48225212e+00 -2.27108955e-01 1.03939736e+00 -7.53519714e-01
9.32641447e-01 1.00742173e+00 1.03570831e+00 2.07069203e-01
-1.19041257e-01 3.19736063e-01 1.17861092e+00 -7.72074997e-01
4.48227733e-01 4.69665676e-01 8.15815449e-01 5.69058359e-01
5.12488067e-01 9.06216726e-02 -3.47996980e-01 -4.49425042e-01
2.40387380e-01 -1.07917048e-01 -5.22788465e-01 -4.03071791e-01
-5.76098204e-01 1.21415710e+00 1.53393388e-01 2.97581196e-01
-6.69131637e-01 -6.33375466e-01 2.49168739e-01 4.15963560e-01
5.34246147e-01 8.46421480e-01 -5.98512113e-01 4.05384392e-01
-1.21008456e+00 2.19792686e-02 5.55344939e-01 4.39055294e-01
1.22024202e+00 7.35528767e-02 8.95097926e-02 8.66166830e-01
-5.51090986e-02 4.06062067e-01 4.06688541e-01 -3.64352584e-01
1.35802880e-01 1.02537179e+00 7.00733736e-02 -8.69961023e-01
-6.24350667e-01 -7.62550652e-01 -5.40351868e-01 6.28112078e-01
1.36338443e-01 -1.20364279e-02 -9.92105722e-01 6.80409551e-01
4.36493009e-01 -2.59062260e-01 3.77584010e-01 8.59782875e-01
5.95433176e-01 6.63984537e-01 -3.20695728e-01 -4.28644031e-01
1.16870701e+00 -6.86044335e-01 -4.24481064e-01 -1.21911168e-01
7.54947603e-01 -7.13028491e-01 1.02779567e+00 6.26296103e-01
-4.31913286e-01 -4.55266118e-01 -1.37488317e+00 6.10199392e-01
-6.89033329e-01 5.90486407e-01 8.79727244e-01 1.03165591e+00
-3.90517980e-01 5.78671873e-01 -3.41733158e-01 -7.47046024e-02
2.24850804e-01 5.63509226e-01 -2.84936965e-01 3.29008967e-01
-9.17924345e-01 8.83845806e-01 9.28531528e-01 4.02765274e-01
-2.02679023e-01 -5.70928037e-01 -5.85854053e-01 1.80461109e-01
2.77303606e-01 -3.15089345e-01 6.65031195e-01 -1.60502195e+00
-1.28827500e+00 6.20367587e-01 -1.01928391e-01 -4.43024486e-01
2.30746448e-01 8.21322575e-02 -5.41104138e-01 4.86135334e-01
-1.98933125e-01 -1.38854980e-01 1.15102708e+00 -1.16945767e+00
-8.26563239e-01 -3.65346074e-01 -3.16352725e-01 1.67524382e-01
-5.30546308e-01 -7.86589980e-02 2.97096699e-01 -3.44867170e-01
5.64221621e-01 -9.66642380e-01 -2.24394947e-01 -4.39480305e-01
-5.43201506e-01 2.47526735e-01 1.09240961e+00 -6.80757523e-01
9.34644938e-01 -2.11475182e+00 -3.39475460e-02 9.26551759e-01
-1.86882272e-01 4.81363326e-01 4.35925052e-02 4.79197681e-01
-3.68940622e-01 -1.72553167e-01 -4.14503634e-01 2.91550159e-01
-5.20429075e-01 -8.96302536e-02 -4.32632208e-01 7.49780953e-01
3.09551358e-01 -6.89987745e-03 -3.89585108e-01 -3.52657884e-01
4.95459765e-01 1.63147047e-01 -3.83766443e-01 8.73934925e-02
-6.99887350e-02 4.02231589e-02 -4.82668221e-01 8.49944890e-01
9.05221760e-01 -1.80501848e-01 6.70230482e-03 -6.19564235e-01
-5.05304873e-01 -1.65599421e-01 -1.35565770e+00 7.68639147e-01
-1.03928894e-01 5.49223781e-01 3.61506082e-02 -1.31371737e+00
8.96240830e-01 6.46882132e-03 5.99856734e-01 3.54609303e-02
1.01653010e-01 4.11148280e-01 -2.69378647e-02 -5.77090502e-01
4.90212739e-01 -1.60211012e-01 4.92220908e-01 -7.98065215e-02
-3.69083434e-02 -2.58305162e-01 3.44854206e-01 -4.21405256e-01
6.22906923e-01 -2.34971538e-01 6.84362471e-01 -5.78495502e-01
9.15888965e-01 8.11043620e-01 2.26163849e-01 6.67809606e-01
-5.13787121e-02 3.15725684e-01 1.63079873e-01 -6.10640109e-01
-7.00459898e-01 -3.83301467e-01 -5.53423524e-01 9.89816189e-01
7.33232722e-02 1.06070995e-01 -4.33488339e-01 -5.74762583e-01
-4.47448716e-03 7.71870315e-01 -4.68870074e-01 -9.94545668e-02
6.73075914e-02 -1.32145584e+00 8.99725258e-02 -8.13019425e-02
4.23465818e-01 -7.04250753e-01 -8.13118756e-01 2.19405472e-01
2.01304421e-01 -5.35830021e-01 1.75538898e-01 7.91274548e-01
-9.56003904e-01 -1.25888777e+00 -3.42096567e-01 -5.09623289e-01
8.42296302e-01 5.53357959e-01 4.07052338e-01 -6.22548945e-02
-3.52685332e-01 7.35627487e-02 -6.57633007e-01 -6.28412485e-01
-2.39444911e-01 1.95433766e-01 -3.94555211e-01 3.18224043e-01
3.48110884e-01 -2.93622464e-01 -3.77650142e-01 1.62536874e-01
-7.68737912e-01 2.07026184e-01 5.31643450e-01 1.16909349e+00
5.52818775e-01 7.05904841e-01 4.30598676e-01 -8.02261412e-01
3.64071310e-01 -1.49439692e-01 -9.65544581e-01 4.56337124e-01
-8.04799676e-01 -9.14784968e-02 8.18627954e-01 -6.79082498e-02
-1.00272000e+00 6.75751626e-01 1.16717443e-01 -7.52922222e-02
-3.06421965e-01 9.45955157e-01 -9.99631956e-02 -7.49703467e-01
9.45734560e-01 7.55453050e-01 9.58664492e-02 -3.98529917e-01
-1.41919658e-01 1.02352333e+00 1.70391560e-01 -4.07555513e-02
8.15254927e-01 5.29608727e-01 4.06199336e-01 -1.44073784e+00
-7.27137923e-01 -8.27878892e-01 -6.53427362e-01 -3.08213592e-01
5.37934422e-01 -7.66612113e-01 -2.59052485e-01 4.09966558e-01
-8.29895377e-01 5.86107671e-02 -8.58512446e-02 7.42433429e-01
-3.42921734e-01 3.63450199e-01 7.99995586e-02 -1.24365103e+00
-4.09010530e-01 -9.45363879e-01 6.40896499e-01 3.28504860e-01
1.36335805e-01 -6.74530208e-01 -2.30416313e-01 3.95006776e-01
3.04203004e-01 2.53575504e-01 8.97417665e-01 -7.39864290e-01
-3.54697138e-01 -3.95954430e-01 -1.21284261e-01 6.36768639e-01
4.18345541e-01 3.63419354e-01 -1.21186244e+00 -3.16714555e-01
1.44299895e-01 8.19135383e-02 1.08703279e+00 5.03026843e-01
1.22452366e+00 -1.65909782e-01 -2.82585293e-01 9.19554472e-01
1.79977798e+00 5.96264422e-01 4.17398036e-01 6.93549037e-01
3.61412287e-01 7.80140519e-01 7.70687103e-01 4.91577357e-01
-4.93616939e-01 4.10045475e-01 4.56845939e-01 -4.30476516e-01
3.15072209e-01 4.04500484e-01 -2.57330853e-03 8.57965425e-02
-4.12890911e-01 1.24952741e-01 -8.59148622e-01 2.49069616e-01
-1.70125806e+00 -1.16951632e+00 -5.63816071e-01 1.99962378e+00
7.10153997e-01 -6.03757389e-02 1.14002973e-01 6.57621324e-01
7.09270954e-01 -7.16900676e-02 -4.37968731e-01 -1.18152194e-01
-4.00483072e-01 2.55526096e-01 8.15104842e-01 5.21812737e-01
-1.51390338e+00 6.37077749e-01 6.32107830e+00 7.36844361e-01
-1.52991962e+00 -4.05465931e-01 3.65145892e-01 1.95757434e-01
-1.48079410e-01 1.02806622e-02 -6.65526867e-01 2.24601746e-01
6.16622508e-01 -1.80234328e-01 4.43441838e-01 1.11489189e+00
2.14350581e-01 -4.63023573e-01 -4.72397149e-01 8.66973877e-01
-2.05409333e-01 -1.41405940e+00 1.62306547e-01 -2.57591959e-02
5.72169483e-01 -2.22593665e-01 -1.60533592e-01 -2.79623151e-01
-4.09422517e-01 -6.71463311e-01 5.76056898e-01 4.10164297e-01
3.52983177e-01 -8.79793644e-01 6.65047765e-01 3.09273452e-01
-1.02321565e+00 -5.14011443e-01 -6.01004362e-01 -2.38057598e-01
-3.50119531e-01 1.01589012e+00 -1.09491611e+00 8.24207127e-01
4.89956141e-01 2.64785945e-01 -5.58152854e-01 1.32067716e+00
1.89008757e-01 6.31095707e-01 -2.28328079e-01 -2.93904036e-01
3.13085914e-01 -5.58893859e-01 5.29675722e-01 1.17465663e+00
5.52490056e-01 1.45281374e-01 2.33501107e-01 5.33015966e-01
9.20235455e-01 5.92842162e-01 -5.32898903e-01 -3.61802518e-01
3.03195328e-01 1.22301173e+00 -1.15754533e+00 -4.67601359e-01
-5.05824566e-01 8.23733449e-01 -1.85048416e-01 3.14275503e-01
-4.60990608e-01 -7.68227100e-01 6.59327745e-01 5.89347328e-04
5.04906714e-01 -3.22929472e-01 -6.14141703e-01 -9.07296181e-01
-1.00773789e-01 -7.38188684e-01 6.10688686e-01 -5.89593768e-01
-1.15177655e+00 8.17033231e-01 -1.30146012e-01 -1.38527942e+00
4.67964262e-02 -8.11470151e-01 -3.68078172e-01 1.18564010e+00
-1.77163100e+00 -1.02711439e+00 -4.73751605e-01 6.79861426e-01
4.13617432e-01 -3.64206105e-01 8.72852802e-01 -1.29547879e-01
-4.68023568e-01 -3.18820495e-03 3.19661409e-01 -1.73629239e-01
1.49586827e-01 -1.07710719e+00 -5.71949244e-01 1.05854726e+00
-5.74232005e-02 3.50489169e-01 9.52179372e-01 -5.88169634e-01
-1.42717505e+00 -1.16577184e+00 6.10251665e-01 2.32405469e-01
5.31475306e-01 2.26413403e-02 -7.95359969e-01 2.87674934e-01
-1.19872659e-01 -2.41107568e-01 1.17858493e+00 1.46303535e-01
1.43751912e-02 -4.21442062e-01 -1.27905035e+00 2.77118206e-01
4.81285751e-01 -4.09522533e-01 -5.36063492e-01 5.97188711e-01
5.68491630e-02 -6.32809773e-02 -5.10048389e-01 4.85944360e-01
3.86782169e-01 -9.99333203e-01 9.47474062e-01 -5.55911720e-01
4.30587307e-02 -6.19073272e-01 -2.29063690e-01 -1.30463600e+00
-4.50007558e-01 -3.14196557e-01 2.55805582e-01 6.30549431e-01
6.35085940e-01 -9.73725379e-01 8.40495288e-01 3.44545960e-01
-1.74931541e-01 -3.58431756e-01 -7.22020984e-01 -1.05532408e+00
-5.93177915e-01 -1.71097279e-01 6.72081232e-01 1.03652883e+00
1.02011934e-01 7.31562525e-02 -3.76762897e-01 6.93054020e-01
6.99129701e-01 8.33090365e-01 4.87394005e-01 -1.59115207e+00
-2.10033447e-01 -4.79177028e-01 -4.82478708e-01 -1.35200471e-01
1.32485367e-02 -8.33563626e-01 -6.44789636e-02 -1.24110460e+00
-2.65911538e-02 -3.20265114e-01 -2.94969350e-01 7.22927988e-01
-1.77587494e-01 1.59653276e-01 -3.32959853e-02 3.06973487e-01
2.91875899e-01 4.02420402e-01 8.23416531e-01 -3.23267519e-01
-6.39177978e-01 2.74277538e-01 -7.04695046e-01 6.37598574e-01
9.44107473e-01 -3.75320971e-01 -4.99366254e-01 3.57563853e-01
2.90605780e-02 -7.06266984e-02 2.92069376e-01 -1.19857788e+00
7.65447319e-02 -7.37721562e-01 5.97138464e-01 -6.88274384e-01
2.16697320e-01 -1.08162594e+00 4.88491446e-01 7.85215676e-01
-1.36420250e-01 -4.98232186e-01 9.74009112e-02 2.17983142e-01
-3.67220968e-01 -7.88899422e-01 9.50344503e-01 -3.47906277e-02
-9.50451016e-01 -7.90645406e-02 -7.49662757e-01 -9.16469514e-01
1.26698887e+00 -6.73732579e-01 -3.15971524e-02 7.10317194e-02
-8.27583551e-01 -2.22459704e-01 2.74215311e-01 -2.89340675e-01
4.73394543e-01 -8.98570061e-01 -6.57942414e-01 5.10521650e-01
2.94479430e-01 -5.56385636e-01 1.56298265e-01 6.35413945e-01
-8.06625545e-01 4.54883695e-01 -4.15228486e-01 -4.66258287e-01
-1.67472231e+00 5.47412813e-01 3.93968880e-01 1.29437605e-02
-5.60275972e-01 8.41149628e-01 -2.15218991e-01 2.16607526e-02
-2.09747806e-01 -2.71420479e-01 -4.81022000e-01 4.78573471e-01
6.47131443e-01 3.52520078e-01 5.13347328e-01 -4.78867412e-01
-3.45502555e-01 3.34776372e-01 2.45461911e-01 1.67969301e-01
1.65262187e+00 1.63902551e-01 -2.89459467e-01 1.02016859e-01
9.23559546e-01 -4.06490117e-02 -8.59325111e-01 1.29470274e-01
4.17908020e-02 -7.23183036e-01 5.38475394e-01 -7.68235028e-01
-8.73345077e-01 5.35262108e-01 8.30602646e-01 8.12443316e-01
1.60769653e+00 -4.79277074e-01 2.85628606e-02 7.24841654e-01
3.39158088e-01 -1.31505895e+00 -5.94208360e-01 2.79451787e-01
8.05924475e-01 -1.09739590e+00 2.19702676e-01 -9.93292034e-01
-3.83216172e-01 1.71130395e+00 3.20993364e-01 4.86025102e-02
8.44893634e-01 1.15676470e-01 -4.61502895e-02 -2.94277728e-01
-1.76842615e-01 -6.73592210e-01 4.32134837e-01 6.38843536e-01
1.26584247e-01 2.13409007e-01 -7.06758857e-01 5.61081648e-01
-7.20670000e-02 -1.26950787e-02 6.26791358e-01 1.02548540e+00
-1.17332065e+00 -7.53967404e-01 -7.67499268e-01 8.08311284e-01
8.30984339e-02 -1.56333730e-01 -4.65165019e-01 5.67331612e-01
2.18958586e-01 9.06610250e-01 -4.11377758e-01 -3.92795175e-01
1.18041992e-01 3.03992629e-01 2.29690224e-01 -5.19878864e-01
-3.57133806e-01 1.25690848e-01 3.39206517e-01 -1.97937593e-01
-5.21708846e-01 -7.15176463e-01 -9.87018287e-01 2.00168401e-01
-9.34011817e-01 5.42892873e-01 8.55999708e-01 9.89169478e-01
8.17336366e-02 3.39950204e-01 9.18147564e-01 -8.41589808e-01
-6.51881218e-01 -7.53994882e-01 -1.33713710e+00 -4.43563573e-02
3.93149555e-01 -6.53028905e-01 -6.93530917e-01 2.40540549e-01] | [9.710464477539062, -1.7328022718429565] |
58f5223d-f34c-4391-9bd4-1a768ce0b5c5 | a-joint-model-for-definition-extraction-with | 1911.01678 | null | https://arxiv.org/abs/1911.01678v4 | https://arxiv.org/pdf/1911.01678v4.pdf | A Joint Model for Definition Extraction with Syntactic Connection and Semantic Consistency | Definition Extraction (DE) is one of the well-known topics in Information Extraction that aims to identify terms and their corresponding definitions in unstructured texts. This task can be formalized either as a sentence classification task (i.e., containing term-definition pairs or not) or a sequential labeling task (i.e., identifying the boundaries of the terms and definitions). The previous works for DE have only focused on one of the two approaches, failing to model the inter-dependencies between the two tasks. In this work, we propose a novel model for DE that simultaneously performs the two tasks in a single framework to benefit from their inter-dependencies. Our model features deep learning architectures to exploit the global structures of the input sentences as well as the semantic consistencies between the terms and the definitions, thereby improving the quality of the representation vectors for DE. Besides the joint inference between sentence classification and sequential labeling, the proposed model is fundamentally different from the prior work for DE in that the prior work has only employed the local structures of the input sentences (i.e., word-to-word relations), and not yet considered the semantic consistencies between terms and definitions. In order to implement these novel ideas, our model presents a multi-task learning framework that employs graph convolutional neural networks and predicts the dependency paths between the terms and the definitions. We also seek to enforce the consistency between the representations of the terms and definitions both globally (i.e., increasing semantic consistency between the representations of the entire sentences and the terms/definitions) and locally (i.e., promoting the similarity between the representations of the terms and the definitions). | ['Franck Dernoncourt', 'Thien Huu Nguyen', 'Dejing Dou', 'Amir Pouran Ben Veyseh'] | 2019-11-05 | null | null | null | null | ['definition-extraction'] | ['natural-language-processing'] | [ 4.80418533e-01 2.49485090e-01 -2.88842559e-01 -4.04927611e-01
-3.13125372e-01 -7.26887882e-01 7.76807487e-01 5.96356988e-01
-1.86948568e-01 3.26186121e-01 2.99183965e-01 -5.38630068e-01
-3.32277119e-01 -9.66285050e-01 -3.40798676e-01 -6.27985656e-01
1.68126509e-01 3.22172642e-01 1.20822256e-02 -2.33649641e-01
3.01769078e-02 4.67028432e-02 -1.31741667e+00 1.89642146e-01
9.11249280e-01 9.66001928e-01 3.05009931e-01 2.78880835e-01
-5.78967273e-01 7.91448891e-01 -5.46527147e-01 -4.66632605e-01
-1.22156233e-01 -6.54045641e-01 -1.20352137e+00 2.24804521e-01
2.80181468e-01 2.51839787e-01 -2.44373411e-01 1.21350741e+00
-1.69030458e-01 1.51747704e-01 4.68097150e-01 -1.03921306e+00
-5.30616641e-01 6.95356011e-01 -3.55679363e-01 7.80673176e-02
3.17665100e-01 -4.13108319e-01 1.60180938e+00 -5.98695338e-01
4.49345767e-01 9.86115992e-01 4.56568956e-01 3.75710011e-01
-1.12331700e+00 -3.67156982e-01 5.57084620e-01 2.50412405e-01
-1.34225476e+00 -4.29642320e-01 9.32539701e-01 -5.17989635e-01
1.19419885e+00 2.16267928e-01 4.30367559e-01 9.07450438e-01
1.43408686e-01 5.88282824e-01 7.71076322e-01 -6.85477316e-01
1.94010228e-01 7.82058090e-02 9.11153078e-01 6.31643474e-01
9.05202106e-02 -3.58881086e-01 -3.02029759e-01 -4.41763774e-02
4.97350305e-01 -2.26332322e-01 -3.10646147e-01 -3.44201326e-01
-9.93447244e-01 8.53006899e-01 3.62983137e-01 8.00971687e-01
-3.42092097e-01 -5.48861288e-02 6.34530842e-01 4.52816673e-02
4.99307185e-01 5.48727095e-01 -5.76620221e-01 2.40365401e-01
-8.54659557e-01 1.95142929e-03 8.15873027e-01 9.41611111e-01
8.85665059e-01 -3.91870916e-01 -1.82868406e-01 6.49388731e-01
4.36042011e-01 6.59953281e-02 4.22822773e-01 -2.25177839e-01
6.28601253e-01 8.88532281e-01 4.35898686e-03 -1.27237928e+00
-5.28311908e-01 -6.12853706e-01 -7.21444249e-01 -4.84186083e-01
2.99191266e-01 -2.10741460e-01 -7.04681456e-01 2.05219674e+00
3.27412874e-01 1.42162755e-01 4.44881953e-02 6.39558256e-01
1.11376786e+00 4.39822823e-01 2.37208381e-01 -1.47922814e-01
1.61963129e+00 -1.02231348e+00 -9.68569994e-01 -6.60225868e-01
8.41319442e-01 -3.05288911e-01 6.81953907e-01 -1.96743265e-01
-7.98042178e-01 -4.81516659e-01 -1.16471422e+00 -2.45508939e-01
-6.02461934e-01 1.68276578e-01 5.88019192e-01 1.43144280e-01
-8.44853044e-01 3.37844223e-01 -6.38014853e-01 -2.03942627e-01
1.79162830e-01 2.55687445e-01 -2.71062970e-01 6.06593452e-02
-1.71708810e+00 1.06760228e+00 5.47174394e-01 2.06586674e-01
-3.07687670e-01 -4.15313363e-01 -1.22069967e+00 5.27220786e-01
6.75191998e-01 -7.76698232e-01 9.44415271e-01 -1.30329776e+00
-8.53261590e-01 8.60566735e-01 -4.95324790e-01 -3.92752737e-01
-1.67564809e-01 -1.44206449e-01 -1.43423468e-01 -3.36089209e-02
2.51060247e-01 2.20268726e-01 4.85540420e-01 -1.23909557e+00
-5.63718319e-01 -4.91813213e-01 5.74577451e-01 3.17133605e-01
-3.76516879e-01 5.73349558e-02 -5.14066160e-01 -4.48732764e-01
1.63507685e-01 -6.73485518e-01 -8.83920267e-02 -2.83343792e-01
-5.19397676e-01 -7.78801262e-01 4.38402176e-01 -7.88374960e-01
1.51943183e+00 -2.16680956e+00 4.72435862e-01 1.54324040e-01
4.44922566e-01 3.64360631e-01 -7.02854916e-02 6.65251017e-01
-2.80567706e-01 2.34695166e-01 -3.94430339e-01 -5.48889816e-01
2.24040091e-01 4.97070938e-01 -2.92226940e-01 2.61271119e-01
3.63164514e-01 9.11843300e-01 -1.16653788e+00 -3.42167825e-01
6.36461824e-02 3.95196229e-01 1.28493518e-01 2.47608900e-01
-3.95153850e-01 2.26788417e-01 -4.62624699e-01 -2.54110899e-02
5.32877743e-01 -3.22856933e-01 6.12056851e-01 -1.64972410e-01
5.58591485e-02 1.06899107e+00 -9.64015007e-01 1.44533598e+00
-7.49455690e-01 4.27143395e-01 6.42129853e-02 -1.52053320e+00
8.75698149e-01 5.62782168e-01 4.21235561e-01 -6.42001867e-01
-2.18640137e-02 1.91455424e-01 -4.42992523e-02 -4.98341262e-01
2.65148133e-01 -2.86027998e-01 -1.16598725e-01 4.53271925e-01
1.08557105e-01 -2.35327687e-02 4.13585544e-01 2.37517282e-01
8.90065908e-01 2.02375893e-02 6.50823712e-01 -2.65207887e-01
6.55171573e-01 -1.11187167e-01 6.07037187e-01 6.09754205e-01
2.27227017e-01 3.85830075e-01 8.20965767e-01 -2.94447154e-01
-6.58928037e-01 -8.82864058e-01 -1.11018062e-01 8.69072080e-01
3.83045644e-01 -7.75039196e-01 -6.03959799e-01 -1.10244775e+00
-2.48589769e-01 8.66671979e-01 -7.00040579e-01 -1.36524484e-01
-5.86973906e-01 -2.85785466e-01 3.12741727e-01 4.63795930e-01
6.11314595e-01 -8.97898495e-01 -3.88919294e-01 1.28611386e-01
-3.43462825e-01 -1.38091242e+00 -4.06853914e-01 4.60769922e-01
-4.05200124e-01 -1.18564904e+00 -1.22065190e-02 -1.04652250e+00
5.55755258e-01 1.36999294e-01 1.37088156e+00 2.99380988e-01
3.89413625e-01 1.91977054e-01 -4.59276348e-01 -1.86208695e-01
-1.21089227e-01 2.03769118e-01 -2.75623888e-01 2.75785118e-01
5.89603424e-01 -5.54841638e-01 -2.14587599e-01 -5.06960526e-02
-1.28348935e+00 2.79502600e-01 3.68775368e-01 8.71214151e-01
4.18264866e-01 3.41948420e-01 6.24439418e-01 -1.09585512e+00
7.02686906e-01 -6.50804281e-01 -2.81218469e-01 7.54896045e-01
-6.98915303e-01 2.97544092e-01 7.45773971e-01 -3.64215732e-01
-8.70556772e-01 -1.03082724e-01 -2.39577532e-01 -2.04497892e-02
-2.15760514e-01 1.03783655e+00 -4.31196213e-01 5.22798896e-01
1.23590797e-01 3.69687110e-01 -4.50742394e-01 -2.49030069e-01
3.72909129e-01 6.40766621e-01 2.14828119e-01 -3.96896005e-01
6.33835196e-01 1.80770218e-01 -7.04361200e-02 -7.91092575e-01
-1.26956105e+00 -7.26678133e-01 -9.31183517e-01 9.16627645e-02
8.53290141e-01 -5.53870261e-01 -1.50557399e-01 1.85210913e-01
-1.48946965e+00 2.62930579e-02 -1.52306616e-01 2.31449261e-01
-1.21464416e-01 6.47969782e-01 -4.31095600e-01 -7.49701560e-01
-3.44611704e-01 -9.73597646e-01 1.07685149e+00 2.27469534e-01
-4.03300792e-01 -1.51476729e+00 -2.05258150e-02 1.94897160e-01
3.71721268e-01 1.71670198e-01 1.47971487e+00 -9.65957105e-01
-1.55630261e-01 -2.13629380e-01 -4.18696821e-01 3.77104968e-01
5.51507533e-01 -2.21667439e-01 -7.33756125e-01 -5.18302945e-03
4.36655551e-01 -6.42835572e-02 9.22821045e-01 1.55535415e-01
6.52506411e-01 -4.67966586e-01 -3.71079773e-01 1.88919663e-01
1.27598703e+00 1.47793621e-01 4.83508408e-01 2.22035855e-01
7.55913377e-01 8.74309361e-01 3.40877026e-01 1.22566968e-01
5.78246713e-01 8.04038405e-01 2.38820016e-01 -2.94407099e-01
-7.37375468e-02 -1.79396451e-01 1.13688119e-01 8.39488983e-01
5.58532834e-01 -3.82962197e-01 -9.94119704e-01 7.17614472e-01
-2.00262642e+00 -6.05772436e-01 -9.61962864e-02 1.89736009e+00
9.91559029e-01 1.19305901e-01 -2.10485309e-01 9.84966010e-02
8.25274169e-01 5.64429760e-01 -3.00401330e-01 -6.42745078e-01
6.61274493e-02 5.24605572e-01 -1.63407475e-01 5.93723416e-01
-1.32661796e+00 7.87390411e-01 4.59210157e+00 6.94488704e-01
-1.02313256e+00 -1.03452511e-01 3.11399251e-01 2.99274594e-01
-4.99288678e-01 3.50143284e-01 -7.28768826e-01 2.47741282e-01
6.42462909e-01 -1.04186259e-01 1.93858892e-01 4.64495987e-01
-2.98454035e-02 4.00583260e-02 -1.41386259e+00 6.16554201e-01
2.62003273e-01 -1.01729286e+00 2.09751934e-01 -2.17647091e-01
4.42913443e-01 -3.45759213e-01 -3.23896259e-01 3.19595188e-01
-3.78323086e-02 -1.04915249e+00 6.79336667e-01 4.23464656e-01
4.66603041e-01 -4.31124121e-01 1.19766688e+00 4.55654681e-01
-1.46875918e+00 1.56477630e-01 3.00016291e-02 -4.40871447e-01
4.01072577e-02 7.47160614e-01 -4.60214257e-01 1.03602040e+00
7.48933703e-02 9.31876004e-01 -4.63791519e-01 6.05730772e-01
-8.90338182e-01 4.21327233e-01 -8.05080533e-02 -1.37823761e-01
3.64478976e-01 -4.22649860e-01 3.83454382e-01 1.37679780e+00
-2.41081402e-01 -1.66253731e-01 3.92462313e-01 1.02091897e+00
-7.87818730e-02 1.68288514e-01 -4.82566416e-01 -1.35653794e-01
3.03985238e-01 1.12181962e+00 -5.31988263e-01 -3.14826369e-01
-6.74350679e-01 5.74997306e-01 6.87361956e-01 4.36748117e-01
-6.86731160e-01 -6.49251401e-01 6.14432514e-01 -1.82964504e-01
3.99958432e-01 -6.03485525e-01 -5.42775691e-01 -1.26083434e+00
4.56285179e-01 -5.15205324e-01 4.18806791e-01 -3.23478162e-01
-1.38165581e+00 7.66390979e-01 1.30675569e-01 -7.60764956e-01
-1.94447204e-01 -4.95968997e-01 -7.94548929e-01 1.13111174e+00
-1.67696559e+00 -1.40942514e+00 -3.08340099e-02 2.92300373e-01
3.73024046e-01 5.46719670e-01 1.00723290e+00 4.52781878e-02
-8.10163736e-01 4.65063393e-01 -3.03829014e-01 6.32223010e-01
1.08881213e-01 -1.18428981e+00 3.50325823e-01 1.05062187e+00
5.25651515e-01 1.00812209e+00 4.02790725e-01 -4.76156920e-01
-9.38629329e-01 -8.95803928e-01 1.65008914e+00 -9.14675891e-02
8.73197615e-01 -6.75194442e-01 -1.10999858e+00 6.34902120e-01
3.01489294e-01 -3.22845876e-01 7.09739208e-01 5.06934702e-01
-5.84899902e-01 -1.61782857e-02 -4.61673290e-01 5.74393332e-01
9.30781305e-01 -8.77549589e-01 -9.92577732e-01 3.97837847e-01
7.68659830e-01 -2.42455304e-01 -6.01732790e-01 4.76980656e-01
3.47314984e-01 -5.87171733e-01 6.87441826e-01 -8.04427624e-01
5.47157049e-01 -2.20585629e-01 -5.64373024e-02 -1.05309021e+00
-2.02487424e-01 -1.88313037e-01 -1.80316105e-01 1.59732223e+00
7.74931848e-01 -4.72476304e-01 1.78954527e-01 5.73339343e-01
-6.20976686e-02 -1.08763576e+00 -1.04103446e+00 -5.91940165e-01
2.15729028e-01 -4.24226642e-01 4.90561157e-01 9.88891065e-01
2.43531510e-01 1.10868657e+00 -1.25634689e-02 3.29464972e-01
1.06711544e-01 7.02936769e-01 1.87658042e-01 -9.55346286e-01
-2.66413510e-01 -5.34756124e-01 -2.11772487e-01 -1.36180699e+00
6.63910806e-01 -1.12257707e+00 2.45331079e-01 -2.11837649e+00
2.47320548e-01 -4.74174082e-01 -3.46603632e-01 6.14969075e-01
-5.05455256e-01 -4.85453814e-01 -8.09053630e-02 1.24921195e-01
-7.15405881e-01 6.02665603e-01 9.47804868e-01 -3.31785321e-01
-1.51711613e-01 -5.77644259e-03 -7.36334503e-01 5.17272472e-01
6.85453057e-01 -3.94085944e-01 -4.60857362e-01 -7.77641058e-01
3.36936623e-01 -1.67289838e-01 2.51581460e-01 -5.51570594e-01
3.74513626e-01 -1.81748852e-01 -2.15688333e-01 -9.14273635e-02
1.06327586e-01 -7.54365623e-01 -3.50153804e-01 1.81618556e-01
-5.96196532e-01 -1.12628572e-01 -1.43376336e-01 5.17758191e-01
-4.84701276e-01 -4.53628540e-01 5.47574580e-01 -1.83263764e-01
-5.28355360e-01 8.49243999e-02 -1.93689629e-01 1.78686738e-01
8.20149839e-01 1.59188539e-01 -3.20292085e-01 -3.03893030e-01
-7.14860559e-01 4.26757067e-01 1.12412460e-01 5.04935920e-01
6.74157143e-01 -9.69578564e-01 -6.25046790e-01 1.36843190e-01
1.72361851e-01 3.41912746e-01 -1.10324040e-01 7.29655087e-01
5.95383570e-02 4.78410542e-01 3.95814449e-01 -3.92990947e-01
-1.30338073e+00 7.33318210e-01 5.72278976e-01 -9.17493284e-01
-4.42447782e-01 7.26136625e-01 4.80414301e-01 -4.52986926e-01
3.71634305e-01 -4.14604604e-01 -7.10150003e-01 1.18096359e-01
3.75513136e-01 -2.79077858e-01 1.02236420e-01 -7.47070014e-01
-5.49538136e-01 5.95105350e-01 -2.23485321e-01 2.52104491e-01
1.28351235e+00 -2.24054992e-01 -6.03173494e-01 6.03284597e-01
1.30402255e+00 -1.44966826e-01 -6.31999135e-01 -6.75209701e-01
5.92602432e-01 -4.50588427e-02 9.49049518e-02 -4.67808217e-01
-8.90529811e-01 8.71339917e-01 -2.36325726e-01 4.78873104e-01
1.22813416e+00 3.84295970e-01 9.56237614e-01 1.91883177e-01
2.08727211e-01 -7.42001295e-01 -2.23332960e-02 9.36844349e-01
6.90687060e-01 -9.58863378e-01 1.61986006e-03 -8.00483763e-01
-4.25741196e-01 1.32542777e+00 4.26628619e-01 1.71334296e-02
5.74734390e-01 1.09187946e-01 -9.51484963e-02 -5.05343020e-01
-7.42196143e-01 -5.03777504e-01 6.42818987e-01 3.90883207e-01
5.57957709e-01 1.15413265e-02 -5.88521242e-01 9.13192630e-01
1.04926676e-01 -4.15316731e-01 -1.61568113e-02 8.24736357e-01
-1.18757889e-01 -1.26141810e+00 2.28945658e-01 3.99348885e-01
-3.43970239e-01 -4.61845458e-01 -8.76608133e-01 5.16126394e-01
4.51543421e-01 1.44222653e+00 -3.92884389e-02 -3.98736477e-01
2.65196353e-01 2.44586304e-01 3.92213240e-02 -1.01639771e+00
-7.73542404e-01 -3.09978127e-01 2.80052990e-01 -1.25523433e-01
-7.08047032e-01 -2.95732498e-01 -1.35115778e+00 1.39652282e-01
-6.71554267e-01 4.18863535e-01 5.24965048e-01 1.67052853e+00
1.59912020e-01 6.87776685e-01 7.33314812e-01 -1.63063616e-01
-4.63489562e-01 -8.88284445e-01 -5.87755084e-01 7.38776028e-01
2.59314269e-01 -5.39190948e-01 -5.01702070e-01 -2.43497729e-01] | [10.029236793518066, 8.674064636230469] |
5f42555b-d093-4480-9de4-209f8d7bdbcf | temporal-knowledge-base-completion-new | 2005.05035 | null | https://arxiv.org/abs/2005.05035v2 | https://arxiv.org/pdf/2005.05035v2.pdf | Temporal Knowledge Base Completion: New Algorithms and Evaluation Protocols | Temporal knowledge bases associate relational (s,r,o) triples with a set of times (or a single time instant) when the relation is valid. While time-agnostic KB completion (KBC) has witnessed significant research, temporal KB completion (TKBC) is in its early days. In this paper, we consider predicting missing entities (link prediction) and missing time intervals (time prediction) as joint TKBC tasks where entities, relations, and time are all embedded in a uniform, compatible space. We present TIMEPLEX, a novel time-aware KBC method, that also automatically exploits the recurrent nature of some relations and temporal interactions between pairs of relations. TIMEPLEX achieves state-of-the-art performance on both prediction tasks. We also find that existing TKBC models heavily overestimate link prediction performance due to imperfect evaluation mechanisms. In response, we propose improved TKBC evaluation protocols for both link and time prediction tasks, dealing with subtle issues that arise from the partial overlap of time intervals in gold instances and system predictions. | ['Sushant Rathi', 'Prachi Jain', 'Mausam', 'Soumen Chakrabarti'] | 2020-05-02 | null | https://aclanthology.org/2020.emnlp-main.305 | https://aclanthology.org/2020.emnlp-main.305.pdf | emnlp-2020-11 | ['knowledge-base-completion', 'temporal-knowledge-graph-completion', 'knowledge-base-completion', 'temporal-information-extraction'] | ['graphs', 'knowledge-base', 'knowledge-base', 'natural-language-processing'] | [-3.08942646e-01 4.47075933e-01 -8.12554717e-01 -4.34248149e-01
-6.44782305e-01 -5.90454876e-01 8.14081192e-01 6.48680031e-01
-1.29994780e-01 1.12383699e+00 2.39231765e-01 -4.52208072e-01
-6.18442059e-01 -8.95781457e-01 -6.58982038e-01 -3.85037549e-02
-7.13327706e-01 1.09134710e+00 8.38761747e-01 -2.36594245e-01
-2.96709418e-01 3.62127155e-01 -1.20191872e+00 6.50245368e-01
5.90413511e-01 9.61474538e-01 -5.05659580e-01 5.40415764e-01
5.94035126e-02 1.42474031e+00 -4.30638522e-01 -7.56237566e-01
-9.97553840e-02 1.02740392e-01 -1.29113746e+00 -8.52635264e-01
-1.03939749e-01 -3.62946815e-03 -8.07239532e-01 1.96471095e-01
5.58962114e-02 1.65879458e-01 3.45120072e-01 -2.02578449e+00
-3.43148768e-01 1.18186963e+00 -3.00033152e-01 4.39598948e-01
5.22736311e-01 -3.61659765e-01 1.28768587e+00 -4.36593980e-01
1.05353236e+00 7.19554961e-01 1.18382120e+00 1.00144498e-01
-1.31233513e+00 -3.73369694e-01 -3.19961794e-02 9.13391709e-01
-1.79050469e+00 -4.75345552e-01 3.65283608e-01 -2.63068706e-01
1.84457731e+00 5.93874574e-01 7.22735167e-01 8.71111453e-01
8.03424567e-02 6.25125110e-01 6.76949739e-01 -3.40826303e-01
-1.61873084e-02 9.90867317e-02 1.95998028e-01 2.94756889e-01
-4.99534281e-03 8.42745975e-02 -9.85936940e-01 -3.42153400e-01
3.02989297e-02 -3.04878980e-01 -2.25788608e-01 -3.28276679e-02
-1.27408051e+00 1.30733654e-01 1.57135755e-01 1.88855574e-01
-3.34838629e-01 3.30289155e-01 6.06579185e-01 3.84869486e-01
7.75224566e-01 3.23299319e-01 -9.99245286e-01 -5.63257515e-01
-8.13053787e-01 4.58200246e-01 1.32806480e+00 1.41108215e+00
6.80791199e-01 -5.85732818e-01 -3.38653266e-01 5.01652896e-01
-1.11933634e-01 1.32658526e-01 2.21132301e-02 -8.45592201e-01
7.04566538e-01 5.48052371e-01 4.97553289e-01 -1.06338596e+00
-6.57876372e-01 -8.99409503e-02 -4.31527764e-01 -8.67479920e-01
2.85116851e-01 2.06454128e-01 -4.77724642e-01 1.71979046e+00
5.19094050e-01 7.17827678e-01 3.36552143e-01 5.62234700e-01
6.83051348e-01 8.20028424e-01 3.13830763e-01 -6.97851956e-01
1.13895059e+00 -6.31500840e-01 -9.09350336e-01 7.50722289e-02
1.14388084e+00 -5.03064096e-01 2.02364877e-01 -7.94793069e-02
-9.06688988e-01 1.04126029e-01 -8.53294313e-01 -3.17083061e-01
-7.43577361e-01 -2.42718056e-01 1.02073956e+00 2.14759380e-01
-9.64875698e-01 8.38120997e-01 -1.00203001e+00 -5.52089393e-01
-2.98539609e-01 2.98078150e-01 -5.05682647e-01 1.31122261e-01
-1.80819321e+00 1.51284862e+00 5.88770986e-01 1.41683817e-01
-3.26508462e-01 -1.05476058e+00 -5.67421377e-01 -2.12709233e-01
7.60382891e-01 -2.73040324e-01 1.22091019e+00 1.07736461e-01
-6.97374225e-01 5.93666255e-01 -3.49076867e-01 -1.01323354e+00
3.81476343e-01 -2.88808763e-01 -1.30667269e+00 -1.57933071e-01
4.29594070e-01 2.57934839e-01 4.23426330e-02 -8.57598186e-01
-8.96820366e-01 -2.08436355e-01 -5.56691960e-02 1.35659590e-01
8.43915902e-03 -1.37080401e-01 -8.97514880e-01 -2.43841022e-01
1.82063326e-01 -8.54196966e-01 1.99583441e-01 -5.30965984e-01
-7.62120426e-01 -6.31527841e-01 9.13337767e-01 -7.30335891e-01
1.99413109e+00 -1.79642892e+00 1.01970986e-01 3.33001018e-01
1.41094986e-03 -1.06047690e-01 2.14066893e-01 1.03674912e+00
-2.06031114e-01 1.70125768e-01 2.63536334e-01 -4.05950367e-01
-5.14699295e-02 6.33860707e-01 -7.39294469e-01 3.53511006e-01
-1.99048538e-02 9.77556407e-01 -9.02920842e-01 -8.73930871e-01
-2.70481646e-01 1.16462059e-01 5.03980890e-02 -2.34266743e-01
-5.89003563e-01 -1.89790010e-01 -2.10559160e-01 8.67292881e-01
1.12319261e-01 -5.05156279e-01 7.86049485e-01 -6.64899230e-01
-2.22811043e-01 7.50516355e-01 -8.20096850e-01 1.28700364e+00
-2.91452296e-02 5.36586642e-01 -7.92603910e-01 -7.10457325e-01
6.00493908e-01 8.54811490e-01 1.01604486e+00 -9.13205385e-01
-7.40271389e-01 1.55866355e-01 -3.31682712e-01 -3.67232770e-01
1.03863156e+00 1.61916152e-01 -7.83445686e-02 5.02241433e-01
-1.38244238e-02 1.68426543e-01 3.88455331e-01 5.35849392e-01
1.57895374e+00 2.75973409e-01 5.15164852e-01 1.33952826e-01
-5.05833030e-02 4.43482757e-01 8.45327497e-01 4.19271022e-01
-2.06765950e-01 2.26015337e-02 7.65512884e-01 -7.00519204e-01
-1.19935250e+00 -8.51130784e-01 -3.08431610e-02 8.37797761e-01
1.36825874e-01 -1.15036500e+00 3.06481451e-01 -8.15873861e-01
4.50549632e-01 9.93786514e-01 -8.34651232e-01 -2.74925858e-01
-6.41578078e-01 -3.86048734e-01 9.48350370e-01 7.54942119e-01
-1.25697076e-01 -6.79184616e-01 -1.08064100e-01 6.91284955e-01
-7.30158210e-01 -1.38214087e+00 -1.91736698e-01 3.03152621e-01
-6.97008729e-01 -1.22271717e+00 1.49130225e-01 -1.71026528e-01
-4.02257917e-03 6.77690357e-02 1.48969233e+00 -8.27396810e-02
-1.41928047e-01 4.20776576e-01 -5.01367450e-01 -3.76247942e-01
7.93410540e-02 1.14783846e-01 9.82055813e-02 -3.79160017e-01
6.09710634e-01 -7.22701550e-01 -3.64726096e-01 5.35790563e-01
-4.19314921e-01 3.78821194e-01 -4.57224287e-02 3.31787914e-01
6.62610114e-01 3.59568447e-01 5.77440143e-01 -9.51477110e-01
1.67203754e-01 -8.83630812e-01 -4.48233664e-01 9.99417424e-01
-1.17824006e+00 -3.19937174e-03 9.37441215e-02 -3.66909325e-01
-9.54032004e-01 -3.94849986e-01 5.81575692e-01 -2.19908848e-01
5.30924857e-01 1.05759871e+00 3.66992384e-01 5.73628426e-01
7.13559031e-01 1.41359745e-02 -4.64620352e-01 -2.58211643e-01
4.99678046e-01 2.00607568e-01 5.55515707e-01 -1.01449049e+00
4.25526887e-01 5.41106284e-01 1.13084786e-01 -1.96718022e-01
-1.20676076e+00 -8.00548375e-01 -5.69403172e-01 -2.71536529e-01
3.63871604e-01 -1.15114069e+00 -8.21160614e-01 -1.00989379e-01
-1.32430089e+00 -5.40986657e-01 -4.23136264e-01 4.83179241e-01
-5.01334310e-01 7.73965269e-02 -8.28470767e-01 -9.11009312e-01
-2.24037796e-01 -3.00575107e-01 8.51070285e-01 -2.56879479e-01
-6.92322373e-01 -1.09623075e+00 2.51915097e-01 2.00946182e-01
1.98430353e-04 3.45546067e-01 9.95121181e-01 -8.80881310e-01
-7.56022096e-01 -4.34861690e-01 -2.61942387e-01 -5.20386219e-01
8.36662576e-03 2.89004207e-01 -5.19862652e-01 1.31401509e-01
-8.16165805e-01 -2.02508286e-01 4.24379647e-01 -8.49170685e-02
7.95679390e-01 -5.00922203e-01 -1.10484290e+00 1.89064503e-01
1.33970380e+00 4.13599759e-01 8.36087167e-01 3.48028630e-01
4.50721920e-01 6.19299114e-01 1.06746650e+00 6.67851746e-01
1.14807796e+00 1.07955945e+00 5.25636859e-02 2.58136958e-01
9.12537500e-02 -3.09677809e-01 3.95373665e-02 5.55279136e-01
-6.63022161e-01 -4.50871766e-01 -1.24692178e+00 1.05641246e+00
-2.45017242e+00 -1.35250020e+00 -5.61438680e-01 2.13661528e+00
1.39216352e+00 1.81971043e-01 5.67662716e-03 8.98929760e-02
4.70821768e-01 -3.48031186e-02 -4.23828989e-01 1.41666969e-03
-3.18506867e-01 -3.42669040e-02 7.20243037e-01 3.40704113e-01
-8.76663506e-01 1.03020179e+00 6.31626797e+00 8.26275408e-01
-7.98175693e-01 1.23597406e-01 3.69133383e-01 -2.05884442e-01
-2.91388243e-01 4.66685772e-01 -9.89813268e-01 3.87181193e-01
1.32907093e+00 -5.05638003e-01 3.91434222e-01 6.13149941e-01
-6.64158463e-02 -1.16958879e-01 -1.65142739e+00 7.31178880e-01
-4.28679198e-01 -1.62302530e+00 -4.27022368e-01 -1.80557333e-02
6.11977160e-01 6.42523095e-02 -2.31445253e-01 6.65732145e-01
5.96988738e-01 -8.07902396e-01 8.46926749e-01 9.87842560e-01
9.03938055e-01 -4.15800124e-01 6.73576236e-01 6.54720217e-02
-1.61202264e+00 1.57173485e-01 -2.30342541e-02 3.55899222e-02
3.21866810e-01 9.52061355e-01 -1.32488620e+00 1.17314053e+00
8.96540642e-01 1.02776885e+00 -4.39874947e-01 7.54612803e-01
8.93750116e-02 6.69862747e-01 -6.70288801e-01 1.71547979e-01
-1.65161297e-01 2.82660574e-01 3.16965610e-01 1.04232574e+00
4.92103606e-01 4.31828588e-01 -3.86120938e-02 3.51538748e-01
-3.02127766e-04 -1.67305648e-01 -7.64844835e-01 -8.60561505e-02
1.21861410e+00 9.14183140e-01 -4.88294005e-01 -5.13278306e-01
-5.35560191e-01 5.32300055e-01 6.93341851e-01 5.54992974e-01
-1.22269464e+00 5.46025187e-02 4.96473908e-01 1.24368288e-01
1.92783296e-01 -1.52554154e-01 -1.97645515e-01 -9.39408839e-01
2.57589728e-01 -7.92517439e-02 9.04780984e-01 -9.68444109e-01
-1.11915410e+00 2.73153335e-01 3.01421881e-01 -1.27812707e+00
-3.40444565e-01 3.85600477e-02 -1.65966466e-01 6.37114584e-01
-1.12500823e+00 -1.23405778e+00 8.27822760e-02 6.15383804e-01
-2.15072662e-01 1.94918305e-01 8.68403137e-01 4.22183394e-01
-3.19683969e-01 3.20161432e-01 -8.32791775e-02 -1.07450709e-01
1.09114933e+00 -1.41724014e+00 2.78507292e-01 4.43541229e-01
2.47964785e-02 6.87219977e-01 8.92090619e-01 -1.08393514e+00
-1.26448274e+00 -1.36882377e+00 1.75545251e+00 -7.57469058e-01
1.11184454e+00 5.79742044e-02 -8.17180693e-01 1.41204536e+00
-7.17579871e-02 2.79369980e-01 8.17582428e-01 9.43955481e-01
-7.47607470e-01 -6.22555971e-01 -6.76649034e-01 5.08851051e-01
1.09816051e+00 -9.10059988e-01 -4.79677796e-01 6.43203676e-01
1.15421903e+00 -4.21834260e-01 -1.68058944e+00 8.09128940e-01
6.45178318e-01 -6.90387845e-01 9.66645896e-01 -8.41794491e-01
2.22426936e-01 -4.45671767e-01 -2.80726433e-01 -9.02105749e-01
-1.83815598e-01 -6.14646137e-01 -1.02823627e+00 1.39703345e+00
9.69251633e-01 -4.11699742e-01 8.33779812e-01 9.19381499e-01
-1.05599597e-01 -7.71983624e-01 -1.02436697e+00 -8.71829212e-01
-6.37089670e-01 -7.41987348e-01 8.83812368e-01 1.38183618e+00
7.53449798e-01 2.56856740e-01 -6.72723830e-01 3.06023896e-01
3.35276991e-01 4.17170078e-01 5.07410645e-01 -1.34121478e+00
-1.75135776e-01 1.46140173e-01 -1.99558273e-01 -4.95568573e-01
-2.97845621e-02 -5.76325357e-01 -8.83086324e-02 -1.63709772e+00
-2.77790558e-02 -9.45455134e-01 -2.21275449e-01 9.96748209e-01
1.96593821e-01 -6.00792617e-02 -2.56160975e-01 6.04989946e-01
-1.09744859e+00 3.05816293e-01 7.03768015e-01 -1.33570194e-01
-2.08308980e-01 -2.33060922e-02 -1.99352115e-01 4.10631187e-02
4.67198163e-01 -4.18656290e-01 -7.31194019e-01 -1.76629096e-01
1.00240052e+00 9.94387388e-01 2.82787144e-01 -8.65107775e-01
9.45393264e-01 -3.38194281e-01 6.15314096e-02 -1.16659546e+00
6.46571398e-01 -7.85830557e-01 1.10000193e+00 1.66674539e-01
-3.66873682e-01 3.29247303e-02 2.80167788e-01 8.16861093e-01
-3.15557390e-01 3.62693995e-01 -1.90223277e-01 1.77973375e-01
-9.24103796e-01 3.99245858e-01 1.31088614e-01 -1.86958030e-01
1.20444429e+00 4.14013900e-02 -9.61158454e-01 -3.54126632e-01
-1.17505765e+00 5.91742933e-01 1.67759489e-02 2.27619693e-01
4.07555163e-01 -1.38415182e+00 -4.20589834e-01 -8.18434238e-01
5.72889566e-01 5.35418745e-03 2.78172284e-01 1.31386495e+00
-2.17243865e-01 7.77633607e-01 4.13765699e-01 -2.86987185e-01
-1.15242326e+00 6.87190354e-01 5.09374626e-02 -7.71819293e-01
-6.18729889e-01 6.71587586e-01 -6.08931124e-01 -3.00198197e-01
2.09071293e-01 -2.98924357e-01 -4.08656523e-02 2.53527194e-01
1.38801157e-01 4.67108846e-01 3.74131262e-01 -2.84140468e-01
-6.03829920e-01 -1.03257760e-01 -3.08774889e-01 -1.07764803e-01
1.31537044e+00 -8.80944505e-02 -4.60607290e-01 8.22913885e-01
7.85782337e-01 -2.14808241e-01 -6.98744237e-01 -6.01110756e-01
6.34714663e-01 6.19559586e-02 -4.71152514e-01 -1.14541698e+00
-4.73437548e-01 4.36090045e-02 -1.23499215e-01 7.23107100e-01
8.48279893e-01 3.13723654e-01 8.11929047e-01 4.24274385e-01
6.92235053e-01 -1.19856894e+00 -4.62491632e-01 5.86111367e-01
6.64625585e-01 -8.96670163e-01 3.76035273e-01 -7.07692802e-01
-5.28211951e-01 7.61466384e-01 7.76754200e-01 6.62947357e-01
7.55174398e-01 3.13596100e-01 -4.53951299e-01 -4.06299889e-01
-1.76991987e+00 -2.45819122e-01 3.11996669e-01 5.47828496e-01
4.92336690e-01 4.58488792e-01 -4.37112629e-01 5.71938634e-01
-2.22589388e-01 1.43304527e-01 2.22532421e-01 9.57424462e-01
5.68129402e-03 -1.06948996e+00 4.20537889e-02 5.45839548e-01
-2.68139273e-01 -1.73778981e-01 -2.28257492e-01 1.14586294e+00
1.62528479e-03 8.69302332e-01 2.35538170e-01 -6.96055532e-01
3.56277823e-01 1.80541217e-01 2.90055573e-01 -3.04089576e-01
-2.57577837e-01 -2.96754956e-01 1.01028967e+00 -9.37130332e-01
-4.91706520e-01 -7.24915862e-01 -1.35102582e+00 -5.94307482e-01
-3.15342546e-01 5.98745108e-01 2.47766867e-01 9.14519370e-01
5.43997824e-01 4.17711735e-01 1.82947427e-01 2.89189164e-02
1.22242700e-02 -8.86573374e-01 -7.04540372e-01 2.63050497e-01
1.90949421e-02 -8.01765919e-01 1.59883142e-01 2.03853697e-01] | [8.565873146057129, 7.932240962982178] |
d7a914c7-363a-465c-9d13-c572afdfbad4 | is-my-automatic-audio-captioning-system-so | 2211.08983 | null | https://arxiv.org/abs/2211.08983v1 | https://arxiv.org/pdf/2211.08983v1.pdf | Is my automatic audio captioning system so bad? spider-max: a metric to consider several caption candidates | Automatic Audio Captioning (AAC) is the task that aims to describe an audio signal using natural language. AAC systems take as input an audio signal and output a free-form text sentence, called a caption. Evaluating such systems is not trivial, since there are many ways to express the same idea. For this reason, several complementary metrics, such as BLEU, CIDEr, SPICE and SPIDEr, are used to compare a single automatic caption to one or several captions of reference, produced by a human annotator. Nevertheless, an automatic system can produce several caption candidates, either using some randomness in the sentence generation process, or by considering the various competing hypothesized captions during decoding with beam-search, for instance. If we consider an end-user of an AAC system, presenting several captions instead of a single one seems relevant to provide some diversity, similarly to information retrieval systems. In this work, we explore the possibility to consider several predicted captions in the evaluation process instead of one. For this purpose, we propose SPIDEr-max, a metric that takes the maximum SPIDEr value among the scores of several caption candidates. To advocate for our metric, we report experiments on Clotho v2.1 and AudioCaps, with a transformed-based system. On AudioCaps for example, this system reached a SPIDEr-max value (with 5 candidates) close to the SPIDEr human score of reference. | ['Julien Pinquier', 'Thomas Pellegrini', 'Etienne Labbé'] | 2022-11-14 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.96383607e-01 2.45645687e-01 3.46524656e-01 -3.03796947e-01
-1.27753627e+00 -7.81964302e-01 7.81360984e-01 4.84344900e-01
-3.88782203e-01 9.68663573e-01 4.75868255e-01 2.65709236e-02
-6.74988925e-02 -2.79303074e-01 -5.28871834e-01 -4.97970223e-01
1.37790442e-01 6.85443103e-01 3.39428812e-01 -2.34650105e-01
1.85004622e-01 1.20265737e-01 -1.77671182e+00 6.36599839e-01
7.35427558e-01 8.94636095e-01 5.68952203e-01 7.61398613e-01
-3.15123141e-01 4.80551273e-01 -9.79418576e-01 -5.77939153e-01
-1.88940823e-01 -9.07164156e-01 -8.14406574e-01 -1.96312085e-01
2.86560535e-01 3.14477533e-01 3.70443135e-01 1.06701267e+00
5.34215271e-01 -1.04409680e-01 6.84309065e-01 -1.26394582e+00
2.01634932e-04 1.22159576e+00 2.18849048e-01 -2.28224128e-01
1.10562742e+00 1.39339745e-01 1.25406313e+00 -6.58962727e-01
5.22802532e-01 1.16620481e+00 2.61973530e-01 6.38308406e-01
-1.27142203e+00 -2.91665196e-01 -1.12497889e-01 2.55119443e-01
-1.39430046e+00 -4.46631461e-01 7.02774525e-01 -4.88797009e-01
3.28912318e-01 6.96352065e-01 5.94765544e-01 1.31852019e+00
-4.84444872e-02 6.42726958e-01 1.01136971e+00 -6.73527062e-01
5.31283259e-01 4.66863006e-01 1.28196776e-01 -9.48251560e-02
-2.87018158e-02 -4.25971329e-01 -6.04062498e-01 -5.41439772e-01
-5.15702814e-02 -7.65994251e-01 -7.26599157e-01 -2.71680709e-02
-1.29862845e+00 5.30467451e-01 9.49656293e-02 6.01859510e-01
-5.53335369e-01 2.40975991e-02 3.55527073e-01 2.48558313e-01
-5.87184429e-02 7.96136379e-01 -1.63604870e-01 -5.54512382e-01
-9.49206769e-01 2.90593207e-01 9.62172687e-01 7.45471299e-01
3.26065689e-01 -2.66709566e-01 -5.75903058e-01 8.27383637e-01
3.44888180e-01 2.55352885e-01 6.74182355e-01 -5.98373890e-01
4.64157104e-01 2.44530797e-01 4.48926061e-01 -5.91514349e-01
-1.70218602e-01 -2.85972118e-01 -5.03217459e-01 2.41321832e-01
5.86489201e-01 -1.60993025e-01 -6.65050745e-01 1.83925939e+00
-9.67163295e-02 -1.23002805e-01 3.09149027e-01 1.25174081e+00
5.09261966e-01 8.96922946e-01 2.81560391e-01 -5.93842924e-01
1.39366674e+00 -8.28877449e-01 -8.94948423e-01 -2.86438018e-01
3.41964364e-01 -1.16516614e+00 1.32951474e+00 8.08737099e-01
-1.02448606e+00 -5.88109136e-01 -9.62670922e-01 4.01982844e-01
-9.80380699e-02 -4.98162489e-03 -1.54624760e-01 6.87272191e-01
-9.24568713e-01 5.38586855e-01 -2.62012243e-01 -3.15045774e-01
-3.70770812e-01 1.13217801e-01 -1.32199153e-01 2.43804768e-01
-1.36046243e+00 7.01460123e-01 3.82210284e-01 -9.97718200e-02
-6.60271406e-01 -1.18878879e-01 -4.03556377e-01 3.71994853e-01
1.08384080e-01 -5.92262030e-01 1.59270668e+00 -1.23218310e+00
-1.68772602e+00 4.67890829e-01 -9.37997997e-02 -6.83585525e-01
7.05412269e-01 -1.82274461e-01 -5.17883539e-01 2.05390871e-01
-1.17534257e-01 8.42663944e-01 9.18378353e-01 -1.43132496e+00
-3.48230779e-01 1.74356580e-01 4.58473191e-02 1.69225991e-01
-1.24555893e-01 2.49417961e-01 -2.36797556e-01 -4.55086082e-01
-1.76045790e-01 -9.98255789e-01 -5.48948813e-03 -2.98878103e-01
-5.99548876e-01 -3.10224175e-01 1.82169884e-01 -5.12823522e-01
1.41399932e+00 -2.30342031e+00 1.02551043e-01 2.80335456e-01
-3.15889031e-01 2.83490747e-01 -2.65743971e-01 5.38893223e-01
-1.72019929e-01 2.82870203e-01 -3.98532629e-01 -3.59550148e-01
1.57703787e-01 -1.58919036e-01 -3.27142149e-01 -5.16144671e-02
1.29233986e-01 3.12186003e-01 -1.12322342e+00 -6.75890744e-01
1.23024900e-02 5.02941668e-01 -2.97889382e-01 2.82654136e-01
-5.45242965e-01 4.07503933e-01 -3.13451380e-01 1.00413524e-01
3.23733628e-01 6.41859099e-02 2.05665767e-01 -9.34690088e-02
-3.01798787e-02 4.76424634e-01 -1.25300264e+00 1.54537451e+00
-6.12238884e-01 7.50748575e-01 -1.55007064e-01 -4.29663926e-01
1.15942538e+00 8.65311980e-01 2.76790619e-01 -4.01009619e-01
1.37834370e-01 5.93478680e-01 2.02423662e-01 -3.18853587e-01
5.70108235e-01 -4.58869375e-02 -2.50966400e-01 4.33428198e-01
-1.17357552e-01 -2.89304912e-01 4.38322484e-01 -2.17551785e-03
1.14632392e+00 -8.26407745e-02 3.25839341e-01 -2.42576525e-02
8.69270504e-01 -1.45156935e-01 -3.88942920e-02 8.37269664e-01
-5.17890695e-03 1.20580602e+00 6.82565570e-01 2.75600832e-02
-1.11334729e+00 -1.03035820e+00 3.08556445e-02 7.14357674e-01
3.10300151e-03 -7.69824862e-01 -1.03519595e+00 -4.49711174e-01
-5.38648546e-01 1.00348175e+00 -1.48234397e-01 1.31108891e-02
-4.34005052e-01 -1.80875450e-01 5.84577978e-01 -1.42843902e-01
-3.53861973e-02 -1.41631448e+00 -7.56235361e-01 5.32130897e-01
-6.80634201e-01 -9.81852412e-01 -4.56181169e-01 1.83607459e-01
-4.85434622e-01 -6.72043145e-01 -9.64434326e-01 -6.26158237e-01
2.36020043e-01 -2.48813644e-01 1.18352246e+00 -1.68231577e-01
2.27383360e-01 2.66673744e-01 -8.85217011e-01 -3.92655253e-01
-1.12417006e+00 9.14501250e-02 1.11045159e-01 1.37512088e-01
6.37097433e-02 -5.05535066e-01 -3.16642880e-01 3.17119479e-01
-9.71537471e-01 1.04726955e-01 5.87822020e-01 6.25470817e-01
4.60719466e-01 -5.15820086e-01 6.20798349e-01 -2.81835645e-01
1.14679563e+00 -2.46563882e-01 -3.07081699e-01 4.50171500e-01
-2.49181166e-01 2.94656873e-01 1.10734165e+00 -6.75021172e-01
-6.32716894e-01 4.81685996e-01 -4.26422447e-01 -2.94228792e-01
-3.77676487e-01 4.12850797e-01 -2.46581003e-01 3.04824293e-01
1.05184102e+00 2.58108437e-01 -1.60366818e-01 -5.13053060e-01
3.30056429e-01 1.14455914e+00 4.32659805e-01 -4.47425038e-01
5.61573088e-01 -3.92094970e-01 -3.45252484e-01 -7.28124559e-01
-5.21944880e-01 -3.15431744e-01 -1.56667545e-01 -5.17393529e-01
7.06326008e-01 -5.81882000e-01 -5.94958663e-01 6.75102249e-02
-1.68726611e+00 9.92588848e-02 -3.66800338e-01 6.65351927e-01
-8.81994426e-01 2.24047452e-01 -7.25949109e-02 -1.08269525e+00
-4.03751969e-01 -1.38007700e+00 1.06698322e+00 1.54122263e-01
-7.07197070e-01 -2.92165071e-01 1.50624931e-01 1.18825719e-01
2.29635447e-01 1.73856735e-01 6.93887293e-01 -1.00220561e+00
-9.74166542e-02 -1.50313675e-01 2.31080890e-01 4.54107136e-01
9.87945274e-02 -1.44464388e-01 -1.11539006e+00 -2.02476932e-03
-2.20011443e-01 -1.60003066e-01 4.32920098e-01 5.03596663e-02
9.67559278e-01 -4.45136428e-01 -1.95562355e-02 -2.59939224e-01
1.16217542e+00 4.57560182e-01 7.74053931e-01 2.20437989e-01
2.24756729e-02 6.92305326e-01 7.81685650e-01 4.82454389e-01
-2.00275287e-01 1.08623469e+00 4.58714962e-01 2.71255136e-01
-9.71735939e-02 -2.92193800e-01 5.62237501e-01 8.71250093e-01
1.82325751e-01 -8.58146846e-01 -8.52249324e-01 3.99791926e-01
-1.62937868e+00 -9.13731754e-01 -3.46536219e-01 2.45292830e+00
7.78030932e-01 4.67219502e-01 2.71151364e-01 6.20254576e-01
1.14106834e+00 -1.48309954e-02 3.39746960e-02 -7.95373023e-01
-4.91320156e-02 9.14997235e-02 -1.29032284e-01 6.22244835e-01
-7.10742950e-01 3.99407208e-01 5.56624699e+00 7.78658390e-01
-1.12472606e+00 -3.89222726e-02 4.79132622e-01 7.80040249e-02
-3.57614070e-01 -7.50063285e-02 -4.02386427e-01 8.43761563e-01
1.44280434e+00 -5.37193239e-01 4.38395143e-01 7.42345095e-01
4.00182933e-01 -4.75889780e-02 -1.36106122e+00 1.08812559e+00
1.25840589e-01 -8.38104606e-01 2.85667330e-01 -1.38569802e-01
2.82080770e-01 -2.25744247e-01 -1.58726335e-01 -2.55279839e-02
-2.63806731e-01 -8.01225781e-01 1.21941912e+00 6.28366053e-01
5.21943867e-01 -4.98084724e-01 8.22812200e-01 2.05145314e-01
-1.03173029e+00 5.16796783e-02 -4.05166717e-03 1.21517271e-01
6.76903844e-01 4.44157869e-01 -1.25266910e+00 5.95925510e-01
2.59925574e-01 -2.55731791e-01 -4.62008119e-01 1.75799978e+00
-2.88546026e-01 7.39692271e-01 -5.35680473e-01 -6.67066455e-01
2.91251749e-01 1.01853073e-01 1.00145161e+00 1.33869338e+00
8.35615993e-01 -2.04469755e-01 -1.33565053e-01 6.70726776e-01
1.03092864e-01 6.92136943e-01 -3.62280309e-01 6.06712839e-03
4.78585690e-01 1.02385461e+00 -4.28245455e-01 -2.85737067e-01
9.18327942e-02 1.03670311e+00 -3.36359441e-01 1.04134835e-01
-1.01804018e+00 -5.86353004e-01 3.46111506e-01 2.68562883e-01
-1.05361499e-01 6.78984076e-02 1.08518817e-01 -6.34185255e-01
2.53372967e-01 -9.95038092e-01 1.41377538e-01 -1.34293985e+00
-1.04371941e+00 1.23976183e+00 1.11523353e-01 -1.89664173e+00
-7.15636075e-01 -3.46197546e-01 -5.87797105e-01 7.92308271e-01
-1.07124853e+00 -4.24289465e-01 -3.53726149e-01 2.37494498e-01
7.73703337e-01 1.08651511e-01 8.67003620e-01 4.18845594e-01
1.33015150e-02 3.68977100e-01 -9.43470001e-02 -3.03897917e-01
1.02513921e+00 -1.35303390e+00 2.41521075e-01 5.70280433e-01
5.83010554e-01 3.08169544e-01 1.44808471e+00 -2.93505520e-01
-7.65795231e-01 -7.31899738e-01 1.27551103e+00 -2.44189091e-02
6.12720788e-01 -1.71465144e-01 -1.03042996e+00 -3.76259126e-02
4.12757069e-01 -4.21962947e-01 5.90069056e-01 -2.19958454e-01
5.57552688e-02 -1.00478813e-01 -8.42644334e-01 6.58293128e-01
6.52101994e-01 -3.36073577e-01 -8.06991398e-01 3.66022766e-01
8.66376162e-01 -2.18568861e-01 -7.32998431e-01 1.86261490e-01
5.71103632e-01 -9.52720225e-01 4.73528355e-01 -2.24202424e-01
2.09079862e-01 -6.20862603e-01 -2.67549045e-02 -1.59254897e+00
5.48772179e-02 -9.29362059e-01 1.46844342e-01 1.41631234e+00
8.29359651e-01 -8.10377747e-02 4.64430183e-01 4.27789629e-01
-2.96633095e-01 -3.24349105e-01 -1.18079913e+00 -8.83332014e-01
-3.80550355e-01 -6.13444448e-01 7.97859907e-01 3.63380224e-01
3.23331445e-01 5.21529377e-01 -5.25037944e-01 -1.09422756e-02
1.94374457e-01 -1.84055328e-01 6.56155586e-01 -1.49315441e+00
-4.02019858e-01 -4.41067189e-01 -4.33990955e-01 -8.57482910e-01
-3.75187285e-02 -7.33432651e-01 5.87015867e-01 -1.46074402e+00
-2.96399921e-01 -1.41575426e-01 -2.62101650e-01 3.59117031e-01
8.63846242e-02 3.62109303e-01 6.44991219e-01 2.57300250e-02
-5.92242241e-01 4.12697077e-01 9.56360698e-01 -3.16681892e-01
-5.53724647e-01 2.33232796e-01 -4.35520321e-01 5.13125479e-01
9.01325762e-01 -7.82292306e-01 -1.86047137e-01 -8.04711282e-02
2.50610828e-01 3.46570045e-01 2.27009252e-01 -1.54197431e+00
2.70406246e-01 7.65819252e-02 -8.88622552e-02 -1.75886914e-01
5.95680714e-01 -8.33403528e-01 5.24413586e-01 4.24653172e-01
-5.98826170e-01 -2.71235630e-02 -4.05534431e-02 2.57209063e-01
-4.85491037e-01 -9.03703749e-01 5.81505537e-01 2.40358301e-02
-3.04391831e-01 -4.51841474e-01 -7.53942192e-01 -2.59827018e-01
6.32474184e-01 -2.08755001e-01 -1.13898836e-01 -8.33504260e-01
-9.45709348e-01 -1.45329908e-02 1.68594792e-01 5.21453619e-01
4.41183507e-01 -1.37987602e+00 -9.58653390e-01 -3.04439992e-01
4.19854224e-01 -4.17569041e-01 -9.93360654e-02 5.60129821e-01
-3.90323758e-01 4.15498555e-01 -1.59102529e-01 -5.79384208e-01
-1.38576758e+00 3.51126969e-01 1.88488573e-01 -2.29638323e-01
-1.84020087e-01 3.40982169e-01 -1.84409544e-01 2.25163653e-01
2.75232583e-01 -3.47460777e-01 -4.99822766e-01 2.34464243e-01
6.19004905e-01 -1.71934459e-02 1.77683532e-01 -6.94419265e-01
-4.14628059e-01 3.97857428e-01 3.82326365e-01 -6.97966158e-01
7.94865251e-01 -5.80180511e-02 1.39428869e-01 5.72903991e-01
9.37593937e-01 3.04983974e-01 -6.78189516e-01 2.40108952e-01
2.21537277e-01 -2.62482166e-01 -2.91972101e-01 -9.88315225e-01
-2.26781309e-01 6.54805064e-01 7.86427677e-01 8.70288193e-01
1.17582369e+00 4.14600074e-02 5.14678836e-01 4.89619464e-01
4.13712740e-01 -8.61764491e-01 4.82575446e-02 2.85835803e-01
1.25505483e+00 -8.08886766e-01 -4.04789865e-01 -2.49064028e-01
-8.51495147e-01 1.38533568e+00 4.08951014e-01 3.49639058e-01
-6.60625286e-03 -3.79745737e-02 1.65263116e-01 4.00952309e-01
-7.40468085e-01 -3.04844618e-01 4.57490176e-01 3.91764939e-01
5.85889876e-01 1.19825721e-01 -1.01752734e+00 7.45227039e-01
-6.24175072e-01 -3.43969576e-02 7.99121380e-01 5.07580698e-01
-5.65426409e-01 -1.40256310e+00 -7.66724885e-01 1.61589399e-01
-4.17344034e-01 -6.56188354e-02 -7.15398610e-01 3.13267857e-01
-5.85796572e-02 1.34936297e+00 -4.11485620e-02 -5.56718767e-01
4.60453689e-01 4.33476567e-01 1.46329761e-01 -4.58971739e-01
-8.17061961e-01 6.26971349e-02 3.65143627e-01 -1.34965226e-01
-4.22060192e-01 -5.93300283e-01 -9.35065091e-01 3.25055897e-01
-3.74969959e-01 9.18021202e-01 9.85753119e-01 8.72453928e-01
4.91690934e-02 3.72725010e-01 7.15816438e-01 -8.19183111e-01
-3.18132669e-01 -1.17231584e+00 -1.71666726e-01 4.50775921e-01
2.03342393e-01 -3.05020124e-01 -6.41645372e-01 2.08239317e-01] | [15.248779296875, 4.850766181945801] |
d9179acd-067c-4d2d-96ea-56c10d4fc9b7 | let-s-be-explicit-about-that-distant | 2106.03192 | null | https://arxiv.org/abs/2106.03192v1 | https://arxiv.org/pdf/2106.03192v1.pdf | Let's be explicit about that: Distant supervision for implicit discourse relation classification via connective prediction | In implicit discourse relation classification, we want to predict the relation between adjacent sentences in the absence of any overt discourse connectives. This is challenging even for humans, leading to shortage of annotated data, a fact that makes the task even more difficult for supervised machine learning approaches. In the current study, we perform implicit discourse relation classification without relying on any labeled implicit relation. We sidestep the lack of data through explicitation of implicit relations to reduce the task to two sub-problems: language modeling and explicit discourse relation classification, a much easier problem. Our experimental results show that this method can even marginally outperform the state-of-the-art, in spite of being much simpler than alternative models of comparable performance. Moreover, we show that the achieved performance is robust across domains as suggested by the zero-shot experiments on a completely different domain. This indicates that recent advances in language modeling have made language models sufficiently good at capturing inter-sentence relations without the help of explicit discourse markers. | ['Robert Östling', 'Murathan Kurfali'] | 2021-06-06 | null | null | null | null | ['implicit-discourse-relation-classification', 'implicit-relations'] | ['natural-language-processing', 'natural-language-processing'] | [ 3.53308648e-01 1.08369005e+00 -4.23715919e-01 -3.21216077e-01
-4.85674828e-01 -4.66995806e-01 1.32629025e+00 6.42296791e-01
-3.30041945e-01 1.08764970e+00 4.79484439e-01 -5.69258392e-01
-4.85038087e-02 -6.40081346e-01 -2.25158691e-01 -5.09796977e-01
-2.13873666e-02 6.51094913e-01 4.11449254e-01 -6.85494959e-01
2.46515900e-01 -3.93962227e-02 -1.54420328e+00 7.32712805e-01
7.96581030e-01 5.40936768e-01 9.75814238e-02 3.98330569e-01
-3.25562060e-01 1.72871816e+00 -7.06707418e-01 -4.24560040e-01
-4.60100442e-01 -6.14563167e-01 -1.46897292e+00 1.88638568e-01
6.97607920e-02 1.06984258e-01 -1.95914390e-03 6.95335031e-01
7.98913091e-02 1.01029322e-01 8.45612705e-01 -8.27876389e-01
-6.16819501e-01 7.16072023e-01 -1.96481630e-01 1.70155197e-01
6.96407855e-01 -4.00479585e-01 1.31308091e+00 -6.11708999e-01
9.49503362e-01 1.34497344e+00 5.20467222e-01 4.27789479e-01
-1.46623611e+00 2.22311094e-02 4.47326958e-01 4.35068935e-01
-9.22517300e-01 -6.73205793e-01 7.42953181e-01 -7.91145384e-01
1.28148174e+00 5.96925795e-01 1.51278734e-01 9.63267624e-01
-4.06622529e-01 5.32455087e-01 1.18799782e+00 -1.00206935e+00
3.42628472e-02 3.96072030e-01 6.67320371e-01 3.66672486e-01
1.37215957e-01 -2.36678883e-01 -4.06406969e-01 -9.72566083e-02
2.64101624e-01 -6.17221177e-01 -4.38893527e-01 -2.27205247e-01
-1.15837908e+00 8.55187595e-01 2.60085762e-01 9.58363414e-01
-5.01932055e-02 -5.27344704e-01 5.87612212e-01 5.70256829e-01
9.44842339e-01 6.70749962e-01 -3.01591307e-01 -2.48566613e-01
-5.99957168e-01 3.51848125e-01 1.25723743e+00 6.80815935e-01
4.93051469e-01 -3.47783566e-01 -9.23805460e-02 8.19817185e-01
1.01330474e-01 -2.85059899e-01 3.02606583e-01 -9.97723937e-01
6.21819258e-01 1.00247371e+00 1.55240491e-01 -1.20272088e+00
-4.05879647e-01 -3.28866439e-03 -5.82649946e-01 1.35489807e-01
8.76153469e-01 5.68596721e-02 -2.55581766e-01 1.58945918e+00
1.27730787e-01 -1.51004434e-01 1.68392986e-01 7.57756531e-01
9.19525623e-01 5.01014829e-01 1.80886269e-01 -8.68684411e-01
1.44514120e+00 -1.06416452e+00 -1.14812291e+00 -2.12794915e-01
1.21059704e+00 -7.00135946e-01 1.03170681e+00 2.17288032e-01
-9.41102743e-01 -3.68092328e-01 -1.00474823e+00 -2.94883728e-01
-2.51358092e-01 -8.77635702e-02 9.47805583e-01 2.60766029e-01
-7.34233022e-01 3.58325273e-01 -5.19242644e-01 -2.83202618e-01
5.03147393e-02 1.90529898e-01 -5.54948211e-01 3.57457489e-01
-1.37720549e+00 1.50975800e+00 5.78530729e-01 2.98052784e-02
4.40176465e-02 -2.30229154e-01 -1.06303680e+00 -1.07557997e-01
8.11875641e-01 -2.86412328e-01 1.43937349e+00 -1.32805109e+00
-1.43817639e+00 1.31384456e+00 -4.59515333e-01 -7.05993652e-01
5.03232121e-01 -3.28968614e-01 -1.38565332e-01 -1.07980296e-01
-4.88753617e-02 -2.14372072e-02 3.35369289e-01 -1.26530516e+00
-4.54566866e-01 -5.35733514e-02 6.19815767e-01 2.86192864e-01
-2.11513460e-01 2.58024961e-01 1.56723022e-01 -4.27187055e-01
-1.50209144e-02 -6.49011910e-01 -1.08124912e-01 -3.46539021e-01
-3.38498741e-01 -7.42214620e-01 8.01448166e-01 -6.16236448e-01
1.45994163e+00 -1.94559836e+00 4.84388560e-01 -4.17842150e-01
2.93024600e-01 5.58841586e-01 2.14535847e-01 6.01120114e-01
-3.03569704e-01 1.18976526e-01 -2.68346786e-01 -4.95492905e-01
-8.21378976e-02 5.34084022e-01 -4.47354674e-01 3.33684027e-01
3.80144805e-01 7.77343452e-01 -9.89295661e-01 -6.49873495e-01
1.21712707e-01 1.08638264e-01 -2.04906851e-01 3.14340740e-01
-5.46784043e-01 4.98222917e-01 -2.64816195e-01 1.38718724e-01
6.16103448e-02 -3.63283545e-01 8.72253716e-01 1.39450803e-01
-2.37016663e-01 1.27006173e+00 -9.31662023e-01 1.48099327e+00
-3.35856169e-01 1.05226552e+00 8.02027062e-02 -1.63418722e+00
9.90802884e-01 8.02459538e-01 2.03182340e-01 -6.44021988e-01
2.77611226e-01 1.29227817e-01 5.70437610e-01 -7.85803854e-01
5.10818541e-01 -4.83836710e-01 -1.82453394e-02 4.57792729e-01
-2.63886787e-02 9.73749012e-02 4.99868423e-01 2.12379675e-02
8.20744038e-01 1.72306046e-01 6.30777478e-01 -4.88719225e-01
9.62672353e-01 1.66275337e-01 3.77050161e-01 4.27115351e-01
1.08355097e-01 3.19857031e-01 9.69411433e-01 -3.22572976e-01
-8.24444950e-01 -4.80823934e-01 -1.84777707e-01 1.05441082e+00
1.44433692e-01 -6.54109478e-01 -6.90878391e-01 -6.11329436e-01
-3.11526418e-01 9.38710868e-01 -7.22414076e-01 3.56305152e-01
-1.07590938e+00 -6.28923297e-01 3.45269471e-01 1.48391873e-01
1.16156921e-01 -1.08504784e+00 -5.43482423e-01 2.55574495e-01
-4.21252847e-01 -1.30747902e+00 3.96141559e-01 2.08875433e-01
-6.70637965e-01 -1.24682033e+00 -2.78538793e-01 -9.42087173e-01
3.51853997e-01 -1.15883701e-01 1.34480226e+00 4.48015064e-01
1.73443511e-01 -7.82937407e-02 -6.08384788e-01 -2.61495888e-01
-7.25448668e-01 3.16886932e-01 -3.03681910e-01 -3.12916785e-01
5.11678517e-01 -4.15014267e-01 8.56382251e-02 5.28627858e-02
-4.94128495e-01 3.83792430e-01 5.56760207e-02 9.25840199e-01
1.64800677e-02 -2.06299514e-01 8.26701880e-01 -1.41781282e+00
8.23480368e-01 -3.48291129e-01 -1.09324314e-01 3.69620234e-01
-4.18669045e-01 8.24322626e-02 2.65660584e-01 -5.49176097e-01
-1.17560124e+00 -4.79644418e-01 -9.66545939e-02 5.56054175e-01
-2.26601228e-01 8.17667007e-01 -1.40165031e-01 4.96845067e-01
7.43240952e-01 -2.67580837e-01 2.04067707e-01 -4.92847949e-01
3.81567985e-01 7.20061362e-01 2.09967688e-01 -5.13291001e-01
4.61356133e-01 3.65933068e-02 -5.25857881e-02 -1.12140846e+00
-1.31291234e+00 -4.09829974e-01 -9.55832958e-01 -1.17851058e-02
1.06008041e+00 -6.74540460e-01 -6.08327568e-01 -1.26971796e-01
-1.48280275e+00 -5.84131658e-01 -3.63040775e-01 2.67923027e-01
-5.53374350e-01 5.88910460e-01 -8.65633726e-01 -1.19042289e+00
4.51199226e-02 -8.98206413e-01 5.73617697e-01 -1.51526526e-01
-1.02571249e+00 -1.31223774e+00 1.47913322e-01 5.67667663e-01
1.83677197e-01 4.54216510e-01 1.17959702e+00 -9.96140420e-01
-4.42836195e-01 7.02273995e-02 -9.42749530e-02 2.10121840e-01
4.63435650e-01 -1.86310887e-01 -1.02543831e+00 1.60425082e-01
3.58528674e-01 -5.11952519e-01 5.36893845e-01 -1.86342493e-01
5.94482601e-01 -4.29419756e-01 -2.24512324e-01 -1.89963415e-01
9.07836735e-01 -9.45545908e-04 6.40276253e-01 6.53066516e-01
3.61928433e-01 1.21956527e+00 8.75607967e-01 -4.46649678e-02
6.00103855e-01 1.01519191e+00 1.23445652e-01 -1.21376477e-01
-2.92248070e-01 1.78377107e-01 2.23658159e-01 9.20409620e-01
-4.71129298e-01 -1.55518875e-01 -1.03827274e+00 5.24093270e-01
-2.25119400e+00 -1.20022917e+00 -7.03220308e-01 1.95754886e+00
1.40942347e+00 3.56383502e-01 4.76275152e-03 5.12900710e-01
4.54476058e-01 3.10247630e-01 3.25014740e-01 -6.62445307e-01
-2.67111331e-01 1.82037488e-01 -2.68452287e-01 1.08116579e+00
-1.19745934e+00 7.99467146e-01 6.01159048e+00 6.13279223e-01
-1.00887954e+00 1.42598599e-01 5.98440766e-01 3.74497771e-01
-1.18286170e-01 1.91404730e-01 -4.47020531e-01 2.35945702e-01
9.88132477e-01 -1.05041966e-01 1.40702322e-01 5.42183876e-01
5.48644252e-02 -5.38403392e-01 -1.58229899e+00 4.98560309e-01
1.09362125e-01 -1.48170662e+00 -2.73862928e-01 -7.61119798e-02
3.80236834e-01 -4.85412508e-01 -4.25635189e-01 4.47758138e-01
-7.54349381e-02 -1.33312500e+00 5.81305683e-01 2.66106486e-01
4.70941603e-01 -2.09671140e-01 9.13123429e-01 9.59160447e-01
-7.47756898e-01 2.07461908e-01 -5.05026095e-02 -1.08855867e+00
3.53392094e-01 2.66194224e-01 -9.64785337e-01 6.44701838e-01
4.62773666e-02 5.21960020e-01 -3.19694787e-01 3.55199069e-01
-5.00489712e-01 6.13230050e-01 -2.73985509e-02 -4.58671227e-02
4.48728688e-02 -6.17981814e-02 4.95411754e-01 1.28893924e+00
-2.09917635e-01 3.49205792e-01 2.35108912e-01 6.09658539e-01
1.48087531e-01 2.55178779e-01 -6.25400245e-01 7.30573535e-02
2.75248379e-01 9.36144173e-01 -5.70992649e-01 -4.17282492e-01
-5.80364466e-01 5.60033679e-01 7.45252430e-01 1.66023403e-01
-4.03426439e-01 4.47585806e-02 2.59722620e-01 3.94596964e-01
2.94026993e-02 -4.17834908e-01 -3.89193386e-01 -1.09932363e+00
1.21954530e-01 -9.38019097e-01 2.04172879e-01 -3.90295595e-01
-1.29438746e+00 7.52454817e-01 2.14471564e-01 -8.96768749e-01
-8.10854256e-01 -6.81247056e-01 -5.16332507e-01 1.03948367e+00
-1.47835064e+00 -1.14637721e+00 3.61920930e-02 1.05905302e-01
6.35589898e-01 1.10945597e-01 1.32187271e+00 2.25931332e-01
-1.73674718e-01 2.57443905e-01 -4.31370258e-01 3.81140947e-01
7.92926908e-01 -1.24184418e+00 -7.70012476e-03 5.09841084e-01
1.82003841e-01 6.65922403e-01 1.14213789e+00 -3.84674609e-01
-7.64848709e-01 -4.36089754e-01 1.68039012e+00 -8.69714022e-01
1.09841418e+00 -4.77805585e-01 -1.33682048e+00 7.93148994e-01
8.10791612e-01 -2.51597136e-01 8.69074941e-01 8.04430604e-01
-3.34880501e-01 5.01532018e-01 -6.80743635e-01 4.37070757e-01
9.13958549e-01 -6.49259329e-01 -1.44057703e+00 4.50636506e-01
8.12258840e-01 -3.64150375e-01 -7.94646323e-01 5.18222690e-01
-4.27229516e-02 -9.32343900e-01 8.44645739e-01 -7.75527418e-01
8.01896036e-01 -1.58351004e-01 -1.38397694e-01 -7.51206160e-01
-5.77273220e-03 -3.44464242e-01 -4.66655254e-01 1.63347697e+00
7.03548670e-01 -6.27816796e-01 6.04058981e-01 8.84330571e-01
-1.06434129e-01 -7.75154412e-01 -9.10569608e-01 -7.29836941e-01
2.18954906e-01 -1.92421079e-01 -1.11642964e-02 1.42310286e+00
8.65447283e-01 1.16885829e+00 -2.19461516e-01 -3.63066405e-01
6.48355037e-02 3.93195152e-01 6.65082932e-01 -1.37742198e+00
-5.16651332e-01 -5.90209484e-01 -3.05502862e-01 -1.12486649e+00
6.80972815e-01 -8.21115494e-01 2.75004618e-02 -1.50622022e+00
6.91307932e-02 -6.95460975e-01 1.88145936e-01 3.65306824e-01
-3.38516623e-01 5.94751351e-02 2.05758423e-01 2.69230634e-01
-6.57314658e-01 5.28416514e-01 1.09171319e+00 -1.37998790e-01
-2.22849146e-01 -8.87633488e-03 -6.14884853e-01 1.02712917e+00
6.61454499e-01 -1.95566610e-01 -6.16571426e-01 -3.14970136e-01
1.35892361e-01 2.83062279e-01 1.34324595e-01 -5.44732034e-01
3.32664838e-03 -2.12100908e-01 -3.37104917e-01 -1.41593277e-01
5.24286330e-01 -6.20673180e-01 -1.55845150e-01 3.21796447e-01
-7.41367042e-01 -1.73307762e-01 1.00110404e-01 2.53163964e-01
-6.28273189e-01 -5.81377983e-01 1.29922122e-01 -4.44543250e-02
-6.78850234e-01 -5.27671337e-01 -6.66401029e-01 2.76142627e-01
1.01894224e+00 -1.37488663e-01 -6.64703369e-01 -3.31530243e-01
-1.05752790e+00 -5.40944189e-02 3.77322912e-01 3.96499902e-01
6.72237054e-02 -9.76805270e-01 -7.47943819e-01 -3.57407391e-01
6.49381205e-02 2.67977566e-01 -2.60040104e-01 1.08506393e+00
-3.66715699e-01 5.43264329e-01 2.18802541e-01 -5.93535662e-01
-1.77473617e+00 6.21925771e-01 8.93913060e-02 -6.70505226e-01
-7.37059176e-01 6.78134382e-01 2.72337347e-01 -4.05656099e-01
2.80868083e-01 -2.52668053e-01 -7.37069607e-01 3.91207218e-01
5.30051053e-01 -1.61761716e-01 6.07986450e-02 -9.40003335e-01
-2.31038034e-01 3.70310768e-02 -6.24738447e-02 3.79621424e-02
1.36614597e+00 -2.37736896e-01 -5.61213076e-01 9.80515540e-01
6.83053315e-01 1.91654205e-01 -6.34717405e-01 -3.77239317e-01
7.97313392e-01 -2.30181858e-01 -4.15932059e-01 -7.33276546e-01
-7.20050409e-02 8.65508437e-01 -1.87379450e-01 8.82809818e-01
7.30819702e-01 2.82333642e-01 1.96189731e-01 4.67671871e-01
3.30123633e-01 -8.69527817e-01 -3.38315591e-02 9.15883958e-01
9.16267097e-01 -1.33779526e+00 2.74198800e-01 -1.04302931e+00
-5.94277501e-01 1.10058415e+00 6.10717475e-01 -1.63797662e-01
4.42330480e-01 3.43636662e-01 1.82813093e-01 -1.88100725e-01
-1.10235977e+00 -2.06637412e-01 2.92892694e-01 5.63041747e-01
1.29349709e+00 3.60430479e-02 -7.76611805e-01 4.46713656e-01
-1.71639815e-01 -2.40303889e-01 5.12959361e-01 9.42472398e-01
-4.56199914e-01 -1.70589817e+00 -2.38100022e-01 1.36899859e-01
-4.57174182e-01 -4.95754890e-02 -6.85064852e-01 1.23738801e+00
4.59569097e-02 1.06723547e+00 1.77560732e-01 -2.41193529e-02
2.43343934e-01 2.11844578e-01 6.04429245e-01 -1.04332614e+00
-5.91005504e-01 -2.46967778e-01 1.04344165e+00 -1.27556667e-01
-1.02819574e+00 -6.97336316e-01 -1.20297170e+00 -3.60024840e-01
-5.39191782e-01 4.65637237e-01 8.69444758e-03 1.50773025e+00
-8.85366350e-02 6.55900121e-01 -2.34834105e-02 -7.55864680e-01
-4.63171721e-01 -1.31070793e+00 -4.39771153e-02 7.37870872e-01
4.58922535e-01 -8.35072279e-01 -1.82751685e-01 2.53347278e-01] | [10.782835960388184, 9.323203086853027] |
22535b5c-ab23-4c92-b3ed-26c0be996e95 | neural-sheaf-diffusion-a-topological | 2202.04579 | null | https://arxiv.org/abs/2202.04579v4 | https://arxiv.org/pdf/2202.04579v4.pdf | Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs | Cellular sheaves equip graphs with a "geometrical" structure by assigning vector spaces and linear maps to nodes and edges. Graph Neural Networks (GNNs) implicitly assume a graph with a trivial underlying sheaf. This choice is reflected in the structure of the graph Laplacian operator, the properties of the associated diffusion equation, and the characteristics of the convolutional models that discretise this equation. In this paper, we use cellular sheaf theory to show that the underlying geometry of the graph is deeply linked with the performance of GNNs in heterophilic settings and their oversmoothing behaviour. By considering a hierarchy of increasingly general sheaves, we study how the ability of the sheaf diffusion process to achieve linear separation of the classes in the infinite time limit expands. At the same time, we prove that when the sheaf is non-trivial, discretised parametric diffusion processes have greater control than GNNs over their asymptotic behaviour. On the practical side, we study how sheaves can be learned from data. The resulting sheaf diffusion models have many desirable properties that address the limitations of classical graph diffusion equations (and corresponding GNN models) and obtain competitive results in heterophilic settings. Overall, our work provides new connections between GNNs and algebraic topology and would be of interest to both fields. | ['Michael M. Bronstein', 'Pietro Liò', 'Benjamin Paul Chamberlain', 'Francesco Di Giovanni', 'Cristian Bodnar'] | 2022-02-09 | null | null | null | null | ['node-classification-on-non-homophilic'] | ['graphs'] | [-1.22302420e-01 6.20591283e-01 -8.08666497e-02 2.62172848e-01
4.79275823e-01 -8.22032690e-01 8.59805286e-01 6.42668828e-02
-1.13701008e-01 4.31375772e-01 6.67405576e-02 -2.41890088e-01
-4.92836922e-01 -1.08758903e+00 -7.94594884e-01 -1.14681709e+00
-5.70286930e-01 5.99564493e-01 3.09257805e-01 -5.21672487e-01
4.38274108e-02 7.19611168e-01 -9.04707491e-01 -2.79137701e-01
9.54179585e-01 4.80900645e-01 -3.13790023e-01 1.07997131e+00
-7.48320296e-02 7.20604956e-01 1.76173046e-01 -4.77238834e-01
4.96699631e-01 -6.27880931e-01 -6.61323488e-01 2.25737281e-02
5.09148180e-01 3.78538847e-01 -9.02593195e-01 1.34631133e+00
1.33362457e-01 4.14302498e-02 1.00621176e+00 -1.29402578e+00
-1.02331305e+00 8.75372767e-01 -5.05302608e-01 5.10501325e-01
-4.33853865e-01 2.00896397e-01 1.02280891e+00 -5.34707844e-01
8.95553470e-01 1.25912833e+00 1.01716053e+00 4.60397094e-01
-1.60961950e+00 -1.83255881e-01 1.33219138e-01 -2.52928734e-01
-1.52151346e+00 -3.26350093e-01 6.40830994e-01 -8.28477204e-01
2.31927410e-01 1.22237481e-01 9.67257142e-01 1.05423129e+00
5.89442432e-01 3.74777645e-01 9.46997464e-01 -2.23688334e-01
4.56149846e-01 -5.70068806e-02 3.41898859e-01 1.12577927e+00
7.81069458e-01 -1.28813118e-01 -1.30510747e-01 -1.54668733e-01
1.35260737e+00 -6.73894063e-02 -6.16246998e-01 -1.08870840e+00
-1.15599799e+00 1.08908355e+00 7.15116978e-01 7.20117509e-01
-3.30408990e-01 3.77335608e-01 5.07619902e-02 5.68557680e-01
2.86509126e-01 3.95497561e-01 2.14593172e-01 4.74122792e-01
-5.36012411e-01 -8.48460868e-02 1.41789520e+00 8.79590154e-01
7.11352527e-01 2.32948437e-02 3.82050127e-01 2.09463656e-01
3.39662582e-02 3.14016163e-01 7.45510682e-02 -8.58999312e-01
1.27985984e-01 5.20305991e-01 -2.13617072e-01 -1.29674613e+00
-6.38193846e-01 -8.05211842e-01 -1.35087585e+00 -6.59195706e-02
9.82090890e-01 -7.02736303e-02 -5.21278262e-01 1.90009904e+00
1.96492642e-01 3.04762602e-01 3.52019072e-02 7.21396327e-01
-1.45271560e-02 5.05123377e-01 -4.12733182e-02 -6.82282746e-02
7.35101044e-01 -5.33188522e-01 -4.56224501e-01 1.87118590e-01
9.60904539e-01 -6.12534992e-02 1.02899694e+00 6.36779368e-02
-1.16171622e+00 9.90099832e-02 -9.71629560e-01 1.17063895e-01
-2.65531063e-01 -4.71355170e-01 4.46460485e-01 6.85055137e-01
-1.64977372e+00 9.05678153e-01 -8.97915900e-01 -8.44424129e-01
4.92753834e-01 3.11212957e-01 -7.16245994e-02 -5.40627772e-03
-1.12034369e+00 6.37443662e-01 -1.04713403e-02 9.43260938e-02
-7.25953341e-01 -7.21411645e-01 -6.29275858e-01 7.25105330e-02
1.77356347e-01 -8.82935166e-01 5.19202769e-01 -1.19539082e+00
-1.10152602e+00 6.26815557e-01 4.19765502e-01 -6.18519962e-01
1.00879419e+00 5.02297997e-01 -1.90061212e-01 3.64769012e-01
-2.72670031e-01 2.08162934e-01 8.12258363e-01 -9.84689057e-01
-1.37033850e-01 -3.62249821e-01 1.58675343e-01 1.60680190e-02
-3.65859777e-01 -7.44402230e-01 5.05719446e-02 -4.91504669e-01
-4.96094376e-02 -1.21753311e+00 -2.58123040e-01 1.97191164e-01
-3.33427906e-01 -1.65703818e-01 6.53272510e-01 -1.56325072e-01
1.08185995e+00 -2.27508163e+00 6.02781951e-01 6.69160724e-01
9.72655654e-01 -1.34477854e-01 -2.35531241e-01 6.03888214e-01
2.12806106e-01 3.67898017e-01 -4.36260968e-01 1.97717756e-01
3.19816135e-02 2.20783368e-01 -1.75794214e-01 1.26580405e+00
-8.30681697e-02 1.14333439e+00 -9.82494533e-01 -2.13884607e-01
-2.70882338e-01 8.06260347e-01 -5.82118511e-01 -5.04506230e-01
8.45827069e-03 3.41720730e-01 -5.41357994e-01 2.53022388e-02
4.59362984e-01 -6.99562013e-01 2.79788643e-01 5.64882047e-02
3.48441117e-02 -4.73314077e-01 -1.05711937e+00 1.11058819e+00
-1.26104578e-01 8.80632818e-01 4.72060144e-01 -1.08852041e+00
6.75236762e-01 -7.88795725e-02 3.36372912e-01 -1.24921232e-01
2.36578956e-01 1.43211454e-01 4.20414031e-01 -2.84609318e-01
7.27384463e-02 -2.49524742e-01 2.46492311e-01 4.40403134e-01
-5.50561473e-02 1.72997981e-01 1.37900874e-01 5.30817688e-01
1.39024019e+00 -5.41268110e-01 -5.87524064e-02 -1.24909270e+00
3.59075308e-01 -1.40020266e-01 1.12747721e-01 9.44483042e-01
-1.99160278e-01 1.61860958e-01 1.05310607e+00 -4.37596977e-01
-1.40680432e+00 -1.28891933e+00 -1.15839936e-01 8.09007347e-01
3.98521900e-01 1.40251210e-02 -1.01079643e+00 -4.59032387e-01
1.49595559e-01 2.68990934e-01 -1.33502102e+00 -5.20839274e-01
-5.31340182e-01 -8.01015139e-01 7.27024615e-01 6.72596022e-02
4.90741879e-01 -4.35077935e-01 -3.14269334e-01 1.69385578e-02
3.02281141e-01 -8.86400878e-01 -8.61497998e-01 -4.73174155e-02
-1.07261038e+00 -1.28296959e+00 -9.33095098e-01 -8.33634555e-01
7.78131664e-01 8.44071954e-02 9.30996716e-01 1.57568231e-01
6.86437078e-03 7.60803401e-01 5.00918664e-02 8.02926440e-03
-5.94248176e-01 5.02244830e-01 2.23384835e-02 6.53783858e-01
-7.94252679e-02 -7.47603238e-01 -7.11409032e-01 4.51272279e-01
-1.18968594e+00 -1.72327206e-01 2.68460780e-01 4.07417536e-01
2.10367322e-01 1.72229484e-01 2.38304898e-01 -1.02865899e+00
7.58277476e-01 -7.71599591e-01 -5.66454589e-01 3.12953979e-01
-5.67528784e-01 4.91042316e-01 7.12178171e-01 -7.17696190e-01
-6.68525040e-01 -2.82892942e-01 7.31297374e-01 -2.99488604e-01
5.27183712e-01 4.26248193e-01 9.34971198e-02 -5.67017555e-01
8.27814639e-01 -4.24416959e-02 3.46151948e-01 -1.37959152e-01
5.28588951e-01 -2.77009513e-02 3.13801289e-01 -4.20988083e-01
9.06703055e-01 8.71111572e-01 6.49817407e-01 -1.34190667e+00
-3.54309887e-01 -1.30771594e-02 -7.50873864e-01 -2.83032477e-01
8.05224121e-01 -4.15094465e-01 -6.03075743e-01 4.79926288e-01
-7.55929530e-01 -7.70353734e-01 -5.86469769e-01 1.93103030e-01
-6.40979528e-01 2.75651306e-01 -9.70807612e-01 -8.09045732e-01
2.45812789e-01 -6.90692902e-01 2.70296216e-01 1.87590141e-02
8.71150643e-02 -1.88745213e+00 3.00216556e-01 -3.91487330e-01
7.98289895e-01 4.46988136e-01 1.16724384e+00 -5.54801524e-01
-6.97367907e-01 -6.68507889e-02 -1.62491292e-01 -3.24192166e-01
-2.76621461e-01 3.07332594e-02 -4.32165772e-01 -2.74616510e-01
5.89838363e-02 3.85269344e-01 1.14936602e+00 7.48939455e-01
3.68394524e-01 -6.65746033e-01 -3.60667318e-01 8.03135812e-01
1.49237037e+00 -2.42943168e-01 3.95444542e-01 -2.44583376e-02
1.04068804e+00 9.23967838e-01 -6.32595837e-01 -3.28387693e-02
2.98204720e-01 2.43850365e-01 2.97380656e-01 4.36280668e-02
-7.09614977e-02 -2.86368281e-01 2.09679544e-01 1.06192768e+00
-3.09217006e-01 -3.85557443e-01 -1.11997926e+00 3.49694550e-01
-1.83796203e+00 -9.59543705e-01 -4.44699019e-01 2.08502793e+00
5.80776870e-01 -3.75606269e-02 4.56392258e-01 -1.88076973e-01
1.19043756e+00 -2.15187874e-02 -7.65882790e-01 -2.54267991e-01
-6.48492157e-01 -2.16933936e-01 9.92944419e-01 9.39100087e-01
-7.28004694e-01 6.65805995e-01 6.17263460e+00 5.03276944e-01
-1.02264750e+00 2.29909271e-01 6.62248135e-01 3.63925733e-02
-5.19412458e-01 -1.12805441e-01 -4.07489449e-01 2.88083434e-01
9.14080322e-01 -3.91813695e-01 8.10540736e-01 3.52813303e-01
-8.71774852e-02 3.04464936e-01 -1.20223558e+00 6.79129124e-01
-1.65110365e-01 -1.48409808e+00 6.47502020e-02 8.92849207e-01
9.89612222e-01 9.71479416e-02 2.31551692e-01 -1.53333575e-01
8.04126620e-01 -1.16126680e+00 5.95938206e-01 7.97290921e-01
5.65316141e-01 -6.53940260e-01 3.77586216e-01 2.16922745e-01
-1.13876057e+00 -3.53567190e-02 -3.74199390e-01 7.99763948e-02
-1.76610611e-02 5.45739949e-01 -3.55219871e-01 9.72501114e-02
1.69352502e-01 7.33472347e-01 -5.34425259e-01 1.04902542e+00
4.00265485e-01 5.68192780e-01 -4.14120436e-01 -1.31502196e-01
4.03510541e-01 -7.39614010e-01 7.84165025e-01 1.12249815e+00
3.23707342e-01 1.15663849e-01 -1.16467506e-01 9.00352657e-01
-2.32009873e-01 1.94328815e-01 -1.01792812e+00 -3.88692558e-01
8.67250785e-02 1.06400609e+00 -1.35612714e+00 -6.83790073e-02
-2.41649762e-01 6.99531555e-01 5.12623310e-01 7.28490531e-01
-4.87388790e-01 -3.50027680e-01 5.66732049e-01 7.16813505e-01
3.83893460e-01 -4.50912297e-01 -2.49278188e-01 -1.26952374e+00
-3.97262633e-01 -4.69789535e-01 4.01130706e-01 -3.25520784e-01
-1.55463076e+00 5.23714006e-01 -2.17181265e-01 -5.04687607e-01
4.01958168e-01 -5.74926615e-01 -6.92849278e-01 5.19533336e-01
-9.83733773e-01 -8.56465757e-01 -1.26125425e-01 5.74828625e-01
-8.57872218e-02 1.60921872e-01 1.67875782e-01 -8.80899578e-02
-5.29758096e-01 1.67541385e-01 3.66930306e-01 2.07236081e-01
9.05687958e-02 -1.47054529e+00 5.39462030e-01 6.67940557e-01
1.84983626e-01 6.58512890e-01 8.10906649e-01 -7.55931497e-01
-1.70241821e+00 -1.06548023e+00 4.99194473e-01 -6.57921374e-01
1.50403345e+00 -5.11199117e-01 -1.06625414e+00 7.49000132e-01
9.00491048e-03 2.41749778e-01 1.76023945e-01 -3.85650024e-02
-3.31839859e-01 2.33513415e-01 -9.31935430e-01 8.56422424e-01
1.50026011e+00 -5.02031147e-01 1.14771634e-01 2.06248745e-01
4.65355933e-01 2.09810540e-01 -7.81927705e-01 1.11327082e-01
4.55146492e-01 -8.58103812e-01 7.22767353e-01 -7.33920455e-01
1.71336263e-01 -1.70479678e-02 3.84704582e-02 -1.39965892e+00
-4.83781010e-01 -9.36513484e-01 -2.49584630e-01 6.70218945e-01
4.57720965e-01 -9.89592671e-01 7.09954739e-01 4.97613966e-01
4.07961339e-01 -6.49310827e-01 -7.58542240e-01 -8.47489119e-01
7.05238044e-01 1.72420263e-01 2.55062133e-01 1.18572617e+00
-9.72154960e-02 4.22263831e-01 -5.87496161e-02 5.62269315e-02
9.74581301e-01 -4.14045721e-01 3.44705164e-01 -1.78517592e+00
-2.52186507e-01 -1.01109850e+00 -6.50954843e-01 -8.62490296e-01
-6.08060397e-02 -1.28898680e+00 -3.80683988e-01 -1.12389338e+00
-1.11224586e-02 -6.06735229e-01 -9.01368037e-02 -3.90769631e-01
4.42221493e-01 1.98363438e-01 -7.09383842e-03 5.23463070e-01
-4.99575704e-01 2.02759594e-01 1.44915640e+00 -9.98880435e-03
-4.79312509e-01 -1.75921142e-01 -6.66569471e-01 6.55041397e-01
8.87654364e-01 -1.37506172e-01 -4.85639006e-01 -2.13335395e-01
7.12646246e-01 -1.47892520e-01 5.98316789e-01 -6.78501844e-01
6.22542739e-01 -3.38775441e-02 1.18622649e-02 3.55570793e-01
-1.51401863e-01 -5.62457561e-01 3.85312706e-01 6.55025542e-01
-7.11362123e-01 7.54065886e-02 -3.81100506e-01 1.15579128e+00
5.48492491e-01 -3.05569302e-02 1.06756711e+00 -8.27957690e-02
5.35976887e-02 6.36488974e-01 -4.89109010e-01 5.27274966e-01
1.17553449e+00 -3.07244420e-01 -4.30692047e-01 -5.89035213e-01
-1.06145191e+00 1.94797441e-02 9.01307762e-01 -5.11532510e-03
5.42556569e-02 -1.27964354e+00 -5.05767167e-01 1.85516506e-01
-2.03402460e-01 -2.88490981e-01 8.70689303e-02 1.47217882e+00
-6.62064493e-01 3.83986011e-02 -1.17593244e-01 -6.02790952e-01
-6.56610310e-01 5.99120975e-01 8.59048963e-01 -7.86265954e-02
-7.90789247e-01 5.90082943e-01 8.73207211e-01 -6.07895814e-02
1.20989949e-01 -3.00539345e-01 1.39828753e-02 1.38397306e-01
8.31066072e-02 6.49696827e-01 -3.81727040e-01 -6.20206177e-01
6.93048164e-02 6.01189077e-01 1.75122857e-01 -8.43939856e-02
1.27401888e+00 -3.74269933e-01 -2.64957130e-01 6.90662503e-01
1.11733854e+00 1.35980427e-01 -1.57361174e+00 -2.08671585e-01
1.82345062e-01 8.12752638e-03 -2.93019786e-02 -7.43943751e-02
-1.19736135e+00 8.40884626e-01 1.41384780e-01 1.16619420e+00
4.48513031e-01 4.44666505e-01 3.59035224e-01 2.55107492e-01
1.14293188e-01 -9.55117643e-01 -5.99126704e-02 5.85198045e-01
7.13383257e-01 -5.96935451e-01 -5.81843793e-01 -4.77679163e-01
-2.35695809e-01 1.21492302e+00 8.07510987e-02 -6.84240699e-01
1.00879109e+00 2.77952790e-01 -4.25518304e-01 -5.47244906e-01
-5.67683220e-01 -2.54888250e-03 1.14631481e-01 7.11885571e-01
-4.46840748e-02 1.22512281e-01 -3.05811483e-02 1.10985592e-01
-1.63106680e-01 -3.99311662e-01 9.61400807e-01 3.36456537e-01
-6.76082671e-01 -5.01406074e-01 5.15771806e-02 4.64643627e-01
-2.63861239e-01 -5.40135056e-02 -8.08185339e-01 1.05801070e+00
-3.26322168e-01 4.77585167e-01 2.69104719e-01 -9.31612700e-02
-1.37539925e-02 -8.81220028e-02 6.90913737e-01 -1.90098315e-01
-5.24146408e-02 -9.47892442e-02 -2.22432315e-01 -2.93754429e-01
-3.17906857e-01 -7.99733698e-01 -7.86150455e-01 -9.33944285e-01
-1.88271075e-01 1.94536433e-01 2.00557217e-01 8.26947868e-01
4.30403084e-01 1.58793092e-01 4.10551220e-01 -5.46872616e-01
-5.20568490e-01 -5.33654690e-01 -1.03812790e+00 1.26435757e-01
6.50922954e-01 -4.77919042e-01 -8.91892374e-01 -9.72312987e-02] | [6.849144458770752, 5.953017711639404] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.