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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
add37c0e-e70b-4150-be09-02b48593970b | towards-end-to-end-unified-scene-text | 2203.15143 | null | https://arxiv.org/abs/2203.15143v2 | https://arxiv.org/pdf/2203.15143v2.pdf | Towards End-to-End Unified Scene Text Detection and Layout Analysis | Scene text detection and document layout analysis have long been treated as two separate tasks in different image domains. In this paper, we bring them together and introduce the task of unified scene text detection and layout analysis. The first hierarchical scene text dataset is introduced to enable this novel research task. We also propose a novel method that is able to simultaneously detect scene text and form text clusters in a unified way. Comprehensive experiments show that our unified model achieves better performance than multiple well-designed baseline methods. Additionally, this model achieves state-of-the-art results on multiple scene text detection datasets without the need of complex post-processing. Dataset and code: https://github.com/google-research-datasets/hiertext and https://github.com/tensorflow/models/tree/master/official/projects/unified_detector. | ['Michalis Raptis', 'Yasuhisa Fujii', 'Alessandro Bissacco', 'Dmitry Panteleev', 'Siyang Qin', 'Shangbang Long'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Long_Towards_End-to-End_Unified_Scene_Text_Detection_and_Layout_Analysis_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Long_Towards_End-to-End_Unified_Scene_Text_Detection_and_Layout_Analysis_CVPR_2022_paper.pdf | cvpr-2022-1 | ['document-layout-analysis', 'scene-text-detection'] | ['computer-vision', 'computer-vision'] | [-1.92942508e-02 -7.83182800e-01 4.73388396e-02 -2.60200500e-01
-7.62429595e-01 -6.27700388e-01 7.80926049e-01 2.00668827e-01
-1.55240893e-01 -2.38142535e-01 3.66712928e-01 -2.91099906e-01
3.31661731e-01 -4.54963326e-01 -4.80400741e-01 -5.42298615e-01
5.44896185e-01 3.76435697e-01 5.26561141e-01 2.06488177e-01
6.19334102e-01 1.15362801e-01 -1.46245480e+00 7.24170089e-01
8.76797795e-01 5.67739904e-01 7.66041875e-01 1.05474508e+00
-2.43907630e-01 9.46452677e-01 -3.78448576e-01 -3.38735640e-01
1.41635254e-01 -1.88571066e-01 -7.83144832e-01 5.22266448e-01
9.14312303e-01 -5.53759098e-01 -6.95286810e-01 1.23690367e+00
5.89876235e-01 -1.26791149e-01 3.45629215e-01 -1.09513187e+00
-8.47172976e-01 7.73336470e-01 -9.80645478e-01 3.23398262e-01
4.46612418e-01 2.05238298e-01 1.01338494e+00 -1.33024681e+00
6.14611804e-01 1.27353704e+00 3.07856709e-01 -3.80052105e-02
-1.05639565e+00 -4.25970376e-01 4.91611928e-01 2.06049472e-01
-1.39691925e+00 -5.79448700e-01 7.28069246e-01 -6.71680927e-01
6.32924974e-01 3.72360915e-01 1.59008369e-01 1.20804334e+00
-2.57996004e-02 1.61340010e+00 8.10250640e-01 -4.72491384e-01
-1.56275630e-01 3.49234731e-04 4.58532035e-01 9.04575765e-01
3.34076464e-01 -8.30887616e-01 -5.85756958e-01 3.04932922e-01
6.30377889e-01 1.89888105e-01 -2.29814425e-01 -3.82053107e-01
-1.62118173e+00 6.20714605e-01 2.05602974e-01 6.15257204e-01
4.96583655e-02 5.34509495e-02 6.68191850e-01 -1.65554732e-01
4.61515307e-01 1.50302246e-01 -1.65958703e-01 -9.20898616e-02
-1.11306679e+00 2.71861404e-01 4.81091291e-01 1.27506089e+00
3.19344789e-01 8.89812112e-02 -3.82369488e-01 1.01029611e+00
3.18481237e-01 7.70202219e-01 3.99128199e-01 -3.62794012e-01
8.27318966e-01 7.67295778e-01 -1.46035120e-01 -1.10266912e+00
-4.25843626e-01 -2.08865866e-01 -8.27074051e-01 -3.86780143e-01
4.30396289e-01 5.32266051e-02 -9.27160323e-01 9.04853284e-01
3.50718051e-01 -1.32911935e-01 -5.19544840e-01 8.65490675e-01
1.14965022e+00 6.96173429e-01 -2.10781381e-01 2.32877806e-01
1.63136899e+00 -1.58945024e+00 -8.63931596e-01 -2.29894131e-01
8.82566273e-01 -1.40298188e+00 1.52055776e+00 6.40420616e-01
-9.44183886e-01 -5.76911807e-01 -1.12898493e+00 -5.84958076e-01
-6.46341443e-01 8.36671233e-01 6.10976398e-01 4.76078570e-01
-8.55441749e-01 -7.11296359e-03 -9.15970087e-01 -8.61873627e-01
5.06578803e-01 -1.24169849e-01 -9.51708555e-02 -2.46324465e-01
-2.63621390e-01 3.24287832e-01 2.17549682e-01 -6.68401420e-02
-8.35397899e-01 -3.55134815e-01 -5.51490366e-01 -8.51050913e-02
5.70686698e-01 -5.87213695e-01 1.31683338e+00 -5.11349618e-01
-1.09352529e+00 9.79629934e-01 -3.57619852e-01 3.28932256e-02
6.86694026e-01 -6.38599932e-01 -3.57024789e-01 1.73904210e-01
3.90880048e-01 3.37142199e-01 8.32951069e-01 -1.22778881e+00
-6.49404049e-01 -6.11129224e-01 -4.43364084e-01 2.81221688e-01
-7.26438046e-01 4.00126755e-01 -1.25257027e+00 -8.27691376e-01
1.43470213e-01 -7.37808228e-01 1.16924599e-01 -5.55522740e-02
-1.16202605e+00 -1.86432660e-01 1.21399105e+00 -6.87654018e-01
1.34895372e+00 -2.07698989e+00 -1.03263162e-01 -3.31176370e-01
3.72620791e-01 1.15891233e-01 -2.29471236e-01 5.75145721e-01
7.07224756e-02 2.84763694e-01 3.70111726e-02 -8.77394855e-01
2.03105643e-01 -4.94948387e-01 -4.48484957e-01 6.25632048e-01
-2.33331114e-01 8.52784216e-01 -5.24786592e-01 -8.40252042e-01
8.06415379e-01 2.67672539e-01 -2.05465108e-01 -4.38799569e-03
-2.64185131e-01 1.04373276e-01 -6.65322959e-01 8.56274605e-01
9.92849946e-01 -5.55925488e-01 1.99620739e-01 -3.52683187e-01
-3.55672717e-01 2.81958789e-01 -1.41465724e+00 1.92016804e+00
8.48422274e-02 1.11467147e+00 7.62931556e-02 -8.84812713e-01
6.21490777e-01 -1.52355477e-01 3.64728004e-01 -7.45662272e-01
5.68318129e-01 -1.52826458e-01 -4.88402188e-01 -6.38019562e-01
9.45077002e-01 9.36700344e-01 -1.53037623e-01 3.55812430e-01
-1.58326805e-01 -5.89280427e-02 6.31616414e-01 5.47390044e-01
1.05523729e+00 1.71451241e-01 -7.94214010e-02 -2.66516030e-01
3.98866653e-01 1.81655064e-01 6.98792338e-02 9.41515326e-01
-2.17904389e-01 8.56242895e-01 5.05759656e-01 -2.63367414e-01
-1.18058419e+00 -1.03154039e+00 -3.18143576e-01 1.50265551e+00
3.10133696e-01 -9.55193996e-01 -7.13163018e-01 -5.37382424e-01
7.17601478e-02 5.63274920e-01 -4.94166225e-01 6.10056400e-01
-4.18467432e-01 -9.12833452e-01 6.83881402e-01 5.68752825e-01
6.46851182e-01 -6.71787918e-01 -3.34413648e-01 -3.66398185e-01
-3.28985482e-01 -1.66812181e+00 -5.94949901e-01 -2.77762096e-02
-6.35374963e-01 -1.00604796e+00 -5.94885170e-01 -7.67049074e-01
6.70008540e-01 8.71446490e-01 8.06234419e-01 2.21898660e-01
-7.94944823e-01 4.17550802e-01 -5.35247564e-01 -3.40422451e-01
-4.12567817e-02 1.97942406e-01 -2.85978228e-01 5.48616350e-02
2.45017290e-01 -9.51949731e-02 -7.15848088e-01 2.90452778e-01
-9.71822083e-01 6.25420749e-01 3.33745629e-01 3.78885090e-01
3.24800253e-01 -2.43187565e-02 -2.48390332e-01 -8.43458891e-01
6.48108125e-01 -7.99393281e-02 -8.00324678e-01 2.72997826e-01
-3.39063555e-01 -2.79402524e-01 4.92345095e-01 -1.83354482e-01
-1.01719916e+00 2.30304003e-01 2.74419874e-01 -3.09201300e-01
-5.65258980e-01 2.44362175e-01 -1.56594902e-01 3.31695080e-01
4.27770764e-01 3.27098966e-01 -5.88066339e-01 -7.81785011e-01
5.78407407e-01 8.94708991e-01 5.15395939e-01 -5.12736082e-01
7.98617244e-01 8.75661373e-01 -3.99191260e-01 -1.22685945e+00
-8.63567472e-01 -1.05048144e+00 -1.07922745e+00 -9.98656526e-02
1.04804587e+00 -1.10248744e+00 -5.19713938e-01 9.01832044e-01
-1.19595194e+00 -6.44807935e-01 3.47979516e-01 1.03323348e-01
-2.53843099e-01 9.80317116e-01 -6.47880793e-01 -6.19050801e-01
-3.83134216e-01 -1.13433921e+00 1.60573065e+00 6.35955408e-02
1.97823435e-01 -9.72797275e-01 -9.86886993e-02 6.86378956e-01
7.69141540e-02 7.18104392e-02 4.20852095e-01 -4.57903683e-01
-8.44009161e-01 -2.47996375e-01 -7.43272781e-01 4.14317008e-03
1.25213312e-02 5.88512659e-01 -1.10559261e+00 -1.97910070e-01
-4.01732147e-01 1.34606604e-02 1.25539124e+00 3.64461601e-01
1.41315961e+00 6.35098293e-02 -5.20801783e-01 8.59839499e-01
1.42989755e+00 -1.38620973e-01 4.82404888e-01 5.66758215e-01
1.32062936e+00 3.81257743e-01 5.30541241e-01 7.39592731e-01
6.15513682e-01 7.09934533e-01 2.06201881e-01 -3.90670210e-01
-3.81388009e-01 -9.57377180e-02 3.20279151e-01 8.02060068e-01
3.97460848e-01 -7.32448578e-01 -1.30355644e+00 5.70843041e-01
-2.25532532e+00 -8.68803740e-01 -1.02927959e+00 1.69799876e+00
3.21599603e-01 1.13439873e-01 1.61872268e-01 7.52549395e-02
9.03671324e-01 4.76690054e-01 -3.34082305e-01 -7.77649321e-03
-4.42043066e-01 -2.79112220e-01 6.12606704e-01 2.98306823e-01
-1.71241522e+00 1.46067023e+00 5.15147495e+00 1.00769246e+00
-1.05576992e+00 3.22557747e-01 5.37707210e-01 -1.88818201e-01
2.24537730e-01 -6.37154328e-03 -9.49176669e-01 4.35114384e-01
4.64275301e-01 -1.43738180e-01 1.36332437e-01 8.26452553e-01
4.14832383e-01 -4.14934844e-01 -8.06833208e-01 1.11067367e+00
3.16530228e-01 -1.35459363e+00 1.55446887e-01 -1.49492016e-02
7.64878452e-01 2.78578371e-01 1.71992570e-01 -1.69765070e-01
2.14316636e-01 -5.74077606e-01 9.92425203e-01 2.04075679e-01
6.55387163e-01 -1.90005481e-01 3.31295907e-01 9.56291407e-02
-1.56124234e+00 6.27664500e-04 -3.43218386e-01 1.07928410e-01
1.27512505e-02 9.34667230e-01 -7.90815115e-01 6.43804371e-01
8.71825457e-01 1.16392040e+00 -1.32337952e+00 1.16011834e+00
-1.99459597e-01 5.08373559e-01 -1.59527272e-01 -2.20116824e-01
1.39113650e-01 -1.26310870e-01 4.94265407e-01 1.75126755e+00
6.29189834e-02 -4.98845935e-01 4.91063863e-01 8.32369268e-01
-5.39877713e-02 4.68762279e-01 -6.26149654e-01 -1.13715373e-01
2.75862575e-01 1.70705819e+00 -1.41851604e+00 -4.53104317e-01
-5.02618134e-01 1.20370138e+00 1.44688323e-01 3.93576145e-01
-9.30510163e-01 -4.44988281e-01 2.35270008e-01 6.41740933e-02
4.22361821e-01 -5.91646850e-01 -6.70521379e-01 -1.58493233e+00
2.27445826e-01 -6.68745458e-01 4.04456288e-01 -1.11727011e+00
-9.86120045e-01 1.65798828e-01 -3.12007785e-01 -1.08322501e+00
4.52267349e-01 -8.68445814e-01 -6.82332993e-01 3.03862244e-01
-1.18453276e+00 -1.54838586e+00 -7.74685204e-01 7.63613403e-01
1.07821453e+00 -5.26276827e-02 3.80999386e-01 4.40984607e-01
-1.25243342e+00 5.13305902e-01 4.53490555e-01 6.68488503e-01
1.00112867e+00 -1.29745018e+00 7.23274708e-01 1.29915547e+00
3.24420869e-01 4.44867849e-01 5.27467549e-01 -7.27741301e-01
-1.77433324e+00 -1.12812245e+00 4.28930342e-01 -9.23338294e-01
9.27051008e-01 -9.82968867e-01 -6.49233997e-01 8.04593563e-01
6.06655121e-01 -3.78437012e-01 2.60567516e-01 2.32954443e-01
-5.35789490e-01 1.03548899e-01 -3.91771555e-01 8.88102710e-01
1.01824999e+00 -4.74580914e-01 -2.09768042e-01 9.66183305e-01
5.26383162e-01 -5.01634061e-01 -5.80997288e-01 1.74622219e-02
4.21310604e-01 -1.02652931e+00 7.70806611e-01 1.05887987e-01
4.34143335e-01 -5.14042556e-01 -2.79424995e-01 -4.30552840e-01
-2.36666322e-01 -5.45765579e-01 -2.54614484e-02 1.42256308e+00
2.21997365e-01 -4.47449923e-01 5.27779818e-01 -9.46162641e-03
-3.17637742e-01 -2.92503446e-01 -4.69124645e-01 -7.03352332e-01
-2.10090935e-01 -7.10283160e-01 2.16251820e-01 1.04335499e+00
5.91739714e-02 5.73222280e-01 -2.28443041e-01 1.75732046e-01
6.83157086e-01 2.76010722e-01 1.23710763e+00 -9.53359544e-01
1.36598647e-01 -9.40920889e-01 -1.83122888e-01 -1.22951257e+00
-1.76486865e-01 -9.92999911e-01 -5.38840033e-02 -1.94177830e+00
7.27383375e-01 3.43648084e-02 8.45198408e-02 4.90280151e-01
-2.01035842e-01 2.63456106e-01 5.94879746e-01 3.31095129e-01
-1.33099258e+00 4.04316455e-01 1.11191773e+00 -2.66379803e-01
3.11297160e-02 -3.19704771e-01 -6.62322819e-01 7.20447421e-01
8.16060960e-01 -2.63368040e-01 -8.62861872e-02 -8.73045087e-01
1.68406740e-01 -4.39980298e-01 4.63781148e-01 -1.00580800e+00
5.03014028e-01 7.72842988e-02 5.61348379e-01 -1.26694512e+00
1.25622690e-01 -3.11334580e-01 -5.40036380e-01 1.91980436e-01
-1.76901862e-01 1.46407396e-01 2.97645777e-01 3.51345330e-01
9.55394134e-02 -6.56816736e-02 6.73790872e-01 1.35362744e-01
-8.46745729e-01 2.49514021e-02 -4.21717674e-01 -9.71039087e-02
8.59025419e-01 -9.31280851e-03 -9.75732207e-01 -1.61050230e-01
-3.23126376e-01 4.65325177e-01 7.33439982e-01 8.67498696e-01
5.37773490e-01 -8.07784259e-01 -6.91850364e-01 -3.62675413e-02
3.60037237e-01 -2.61734389e-02 3.77744675e-01 1.11522651e+00
-8.10180306e-01 7.38195360e-01 1.56842738e-01 -1.01018560e+00
-1.60089839e+00 7.83240199e-01 -1.75062921e-02 -1.98798954e-01
-8.63987446e-01 5.61417341e-01 5.07388353e-01 -3.36699009e-01
3.18407536e-01 -5.44133902e-01 1.45858556e-01 3.37971225e-02
5.76082706e-01 4.39445078e-01 1.31299853e-01 -5.10361254e-01
-3.33210260e-01 7.22976804e-01 -1.79189786e-01 1.82196796e-01
1.11964428e+00 -4.20989335e-01 -1.55515075e-01 6.15494967e-01
1.00964475e+00 7.23447278e-02 -8.64838421e-01 -2.19432160e-01
1.27553910e-01 -6.17071450e-01 2.17243642e-01 -6.37858093e-01
-8.76580119e-01 1.01920569e+00 6.91536963e-01 4.71510798e-01
1.03780925e+00 1.20804988e-01 4.91214544e-01 4.52456445e-01
-2.45430321e-01 -1.32546842e+00 6.02518022e-01 5.93760967e-01
7.74573207e-01 -1.39397991e+00 4.42954630e-01 -6.66893959e-01
-6.04213059e-01 1.22798848e+00 7.09823072e-01 6.24747714e-03
4.35164899e-01 4.58207548e-01 3.88318971e-02 -4.94930536e-01
-7.30061829e-01 -6.16606832e-01 3.75563383e-01 2.40838259e-01
8.46927106e-01 1.91575512e-01 4.05938104e-02 2.18232200e-01
-1.50212482e-01 -4.97735679e-01 7.11865127e-01 9.48861480e-01
-3.86730999e-01 -8.23942363e-01 -5.99139631e-01 4.52122450e-01
-3.63192022e-01 -2.91373610e-01 -6.99068785e-01 8.22500527e-01
-4.69385773e-01 1.08839369e+00 2.14294568e-01 -2.62529492e-01
2.94260651e-01 9.17128772e-02 3.40914816e-01 -5.72399139e-01
-2.91988581e-01 7.41450250e-01 -4.75072898e-02 -5.29451191e-01
-1.43097132e-01 -8.92620206e-01 -1.18132293e+00 -5.13099611e-01
-3.75903249e-01 -5.45851469e-01 9.50672805e-01 5.27076066e-01
4.93862957e-01 7.50340939e-01 5.95610023e-01 -9.22308564e-01
1.39975235e-01 -1.11179507e+00 -6.16418898e-01 4.77816820e-01
5.92234880e-02 -3.51534516e-01 -1.32188901e-01 4.86400157e-01] | [12.02387523651123, 2.2239794731140137] |
59918c1c-5274-4260-91e9-6e465c55477e | modeling-intelligent-decision-making-command | 1903.08412 | null | http://arxiv.org/abs/1903.08412v1 | http://arxiv.org/pdf/1903.08412v1.pdf | Modeling Intelligent Decision Making Command And Control Agents: An Application to Air Defense | The paper is a half-way between the agent technology and the mathematical
reasoning to model tactical decision making tasks. These models are applied to
air defense (AD) domain for command and control (C2). It also addresses the
issues related to evaluation of agents. The agents are designed and implemented
using the agent-programming paradigm. The agents are deployed in an air combat
simulated environment for performing the tasks of C2 like electronic counter
counter measures, threat assessment, and weapon allocation. The simulated AD
system runs without any human intervention, and represents state-of-the-art
model for C2 autonomy. The use of agents as autonomous decision making entities
is particularly useful in view of futuristic network centric warfare. | ['Sumanta Kumar Das'] | 2019-03-20 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [-1.78500637e-01 5.52187562e-01 -1.64559275e-01 -1.03560507e-01
4.35657531e-01 -1.09101915e+00 1.38540936e+00 3.02833200e-01
-7.75996327e-01 8.84944618e-01 -4.15797438e-03 -1.15265894e+00
-6.69234335e-01 -9.07975078e-01 2.47622296e-01 -3.88521016e-01
-5.90928495e-01 1.04573715e+00 1.57716230e-01 -1.12468290e+00
1.78884000e-01 1.05344391e+00 -1.29251182e+00 1.86499488e-03
3.94400328e-01 1.01448977e+00 -2.40036622e-01 1.11869550e+00
4.01096433e-01 1.46854794e+00 -1.01093614e+00 7.72982463e-02
5.83069980e-01 -1.16700523e-01 -6.72977388e-01 -4.04209942e-01
-4.07925665e-01 -5.64666152e-01 -9.57377702e-02 9.08309102e-01
4.35386598e-01 3.36349189e-01 6.80463672e-01 -1.68161142e+00
4.83088642e-01 2.93182194e-01 2.63055831e-01 1.23465918e-01
5.22600234e-01 3.02199662e-01 1.76335737e-01 -2.30059609e-01
6.92571104e-01 1.29060864e+00 1.65958911e-01 6.83254898e-01
-5.00337839e-01 -3.74727100e-01 2.93413609e-01 1.28578469e-01
-9.75140154e-01 -3.65610681e-02 3.13371986e-01 -7.51016617e-01
1.20053601e+00 6.50832057e-01 7.20158637e-01 5.06729305e-01
8.01990509e-01 2.14726508e-01 1.18051994e+00 -5.60602844e-01
7.86432385e-01 3.95615548e-02 5.43647587e-01 5.84003031e-01
5.25383532e-01 9.31254447e-01 8.48137736e-02 -6.20444775e-01
7.70979822e-01 -3.04384321e-01 1.76565479e-02 -7.80358091e-02
-1.07482421e+00 8.73695910e-01 1.16765045e-01 3.86653364e-01
-1.30189896e+00 -4.40036282e-02 5.43989003e-01 4.02948260e-01
-1.63770989e-01 9.21141982e-01 -2.35014021e-01 -4.08150554e-01
-2.51519263e-01 8.61710727e-01 1.25567997e+00 4.24781322e-01
-1.93485290e-01 7.19560564e-01 2.24727601e-01 -3.33509445e-01
4.90015537e-01 4.88187253e-01 3.94048318e-02 -1.04111195e+00
-1.47942489e-03 5.61579347e-01 5.49716353e-01 -8.57834876e-01
-7.60447204e-01 -3.75547230e-01 -2.22605556e-01 1.48908389e+00
-2.94753373e-01 -8.15799892e-01 -6.60666764e-01 1.00555813e+00
5.33690810e-01 -3.40507060e-01 7.31266022e-01 9.37066615e-01
5.15665174e-01 7.36969769e-01 4.48694229e-01 -2.66482264e-01
1.62575066e+00 -5.98880112e-01 -1.07516146e+00 1.18194968e-01
4.42485988e-01 -1.70399740e-01 -1.07246041e-01 6.97342873e-01
-1.33781052e+00 -9.30565074e-02 -1.32935083e+00 8.18942904e-01
-6.40689015e-01 -5.25406182e-01 8.75782192e-01 8.54012728e-01
-7.66166568e-01 3.19787487e-03 -5.73259056e-01 -6.04843237e-02
-4.22995180e-01 6.39908195e-01 -1.38786845e-02 4.49794680e-01
-1.44142544e+00 1.65504527e+00 6.38734162e-01 -6.78668730e-03
-1.58648789e+00 -2.45389342e-01 -7.59027660e-01 -1.71122715e-01
4.06808347e-01 -5.98447561e-01 1.41171980e+00 -6.75757587e-01
-1.49699187e+00 5.33714354e-01 7.54203856e-01 -9.25891519e-01
4.60246235e-01 1.30487317e-02 -8.66015732e-01 1.88752562e-01
-3.16203833e-01 1.56183302e-01 4.96175975e-01 -1.46773160e+00
-9.74201679e-01 -2.53244579e-01 1.07950807e+00 5.80120623e-01
6.30119801e-01 5.82472682e-01 8.78115237e-01 -1.91969469e-01
-5.31427085e-01 -1.06834066e+00 -6.28180683e-01 -6.01379097e-01
2.42687657e-01 8.74476060e-02 9.89555418e-01 -2.96716541e-01
1.23409092e+00 -1.57069957e+00 3.32911640e-01 3.55689853e-01
2.55930424e-01 6.33272707e-01 1.41435862e-01 1.08804238e+00
-1.53899506e-01 -6.37750234e-03 2.80245632e-01 6.41531289e-01
3.73700142e-01 7.76973292e-02 -5.55711091e-01 4.51023057e-02
-3.16761136e-01 1.57908291e-01 -8.79272759e-01 -1.13810904e-01
3.93862218e-01 -1.54038727e-01 -2.34100502e-02 3.23920399e-01
-4.58555222e-01 2.41061479e-01 -9.51440513e-01 7.10285723e-01
4.78682548e-01 8.94216120e-01 6.85806274e-01 8.58448744e-02
-6.99259281e-01 6.88213184e-02 -9.23641145e-01 1.00671864e+00
-2.59610832e-01 3.44003469e-01 9.57462370e-01 -5.79533219e-01
6.15492761e-01 9.58975554e-01 5.01140833e-01 -4.33097273e-01
6.82672381e-01 1.51620209e-01 7.68648505e-01 -5.72987139e-01
4.36086059e-01 -5.16789675e-01 -4.48832780e-01 7.87801266e-01
-2.69339383e-01 -6.85846746e-01 1.44005492e-01 1.67528346e-01
1.34367216e+00 -1.55579478e-01 6.43122435e-01 -5.06040454e-01
7.99978554e-01 8.76766324e-01 3.16979557e-01 6.33359253e-01
-6.10971987e-01 -7.71649480e-01 3.11393470e-01 -1.08403265e+00
-5.42901337e-01 -8.36529434e-01 3.63674343e-01 7.27844417e-01
2.15187535e-01 -2.91198790e-01 -8.51368308e-01 -7.95343995e-01
-3.58380765e-01 1.46535850e+00 -2.65749574e-01 1.90289050e-01
-7.79853463e-01 -5.45433223e-01 5.18649817e-01 3.33802402e-02
4.77308750e-01 -8.81001949e-01 -1.60668862e+00 4.45474148e-01
3.52810800e-01 -1.12266779e+00 4.07809407e-01 -1.58631191e-01
-3.73231262e-01 -1.35480118e+00 3.13225865e-01 -2.13412493e-01
5.72992027e-01 8.64186659e-02 8.16861033e-01 4.16358262e-01
-2.40902945e-01 1.03713381e+00 -5.48815250e-01 -1.31557763e+00
-8.52619588e-01 -5.63012779e-01 6.75269604e-01 -4.45034266e-01
3.03579569e-01 -9.22879353e-02 -4.64791715e-01 3.33305508e-01
-9.53000963e-01 -4.00249753e-03 2.33926371e-01 3.94806057e-01
-4.88623083e-01 7.73200750e-01 1.79228693e-01 -3.28737676e-01
1.68389988e+00 -2.46536747e-01 -8.00950527e-01 2.45933145e-01
-4.36326802e-01 -4.40928400e-01 5.78522503e-01 2.61713360e-02
-9.19445276e-01 -5.53435922e-01 3.97788644e-01 2.76535302e-01
-3.21316093e-01 5.04488647e-01 -1.72710158e-02 -3.64662498e-01
6.61123097e-01 -3.29222143e-01 5.64222753e-01 1.98828265e-01
-2.49731258e-01 6.75834000e-01 1.63839340e-01 -8.24611664e-01
9.15256798e-01 3.10039133e-01 4.92571533e-01 -5.68424940e-01
-4.34500873e-02 2.36789212e-01 -2.80502528e-01 -8.47578347e-01
9.68725860e-01 -4.74438548e-01 -1.26321781e+00 1.06133103e-01
-1.21712935e+00 -4.59414065e-01 -2.95505792e-01 9.14415598e-01
-5.46180069e-01 -1.97484389e-01 -2.85561800e-01 -1.44764829e+00
-4.66639727e-01 -1.24249983e+00 3.80389392e-01 4.12179269e-02
-2.37370670e-01 -1.13154042e+00 5.36211252e-01 2.90572613e-01
5.31012952e-01 7.29064703e-01 1.07609081e+00 -9.46275651e-01
-3.44789743e-01 -7.32265711e-01 5.39969981e-01 1.36903435e-01
-1.59082428e-01 -6.58735856e-02 -3.06832045e-01 -2.11829990e-01
2.56379485e-01 8.45417287e-03 -5.63219190e-01 1.32418185e-01
1.13168396e-01 -3.46680731e-01 -2.67883360e-01 -4.32324201e-01
1.27704072e+00 1.41572320e+00 6.89084530e-01 7.57370234e-01
-3.62447560e-01 1.12300539e+00 1.60633075e+00 6.41001761e-01
5.46992309e-02 9.75775182e-01 7.64969766e-01 -1.62145346e-02
3.49308908e-01 2.58363396e-01 4.62286919e-01 3.17205280e-01
-9.65521574e-01 -3.95292103e-01 -1.38341975e+00 -1.01882651e-01
-1.97303760e+00 -1.31427014e+00 4.71730679e-02 1.79399359e+00
2.49852836e-01 3.51451367e-01 1.50314957e-01 -8.09365362e-02
5.00376642e-01 2.67845839e-02 -8.14258680e-03 -9.46778595e-01
3.37686867e-01 3.19058038e-02 5.57036340e-01 9.94163990e-01
-8.70717704e-01 9.06711221e-01 7.01231527e+00 5.57036757e-01
-6.34982944e-01 1.74559981e-01 -6.28524721e-02 1.46834373e-01
7.72524551e-02 3.14530075e-01 -4.96898890e-01 8.03487822e-02
1.22654676e+00 -3.14642936e-01 5.18728435e-01 6.95882499e-01
7.46644795e-01 -2.42586449e-01 -5.27213991e-01 3.93426538e-01
-1.33936405e-01 -1.38249910e+00 8.00782815e-02 2.42352128e-01
1.07129484e-01 -5.26454568e-01 -3.80913883e-01 4.60085511e-01
8.51541996e-01 -8.16298604e-01 9.75507498e-01 5.53991675e-01
2.23554879e-01 -9.38256204e-01 1.17506194e+00 5.68558991e-01
-7.50055730e-01 -5.90515673e-01 6.52919039e-02 -5.38011074e-01
7.17015803e-01 -2.34576777e-01 -1.18020093e+00 7.15674818e-01
1.02261961e-01 -5.72822690e-01 1.28967479e-01 4.85429347e-01
-1.14342168e-01 7.49097615e-02 -1.71049945e-02 -4.24747020e-01
7.56669164e-01 -4.43048358e-01 1.10746729e+00 7.07148850e-01
-1.43969283e-01 1.06262267e+00 3.10572654e-01 9.36804637e-02
1.11873102e+00 -3.47578228e-01 -8.26198697e-01 -1.95795789e-01
5.14389277e-01 1.21835685e+00 -3.97270173e-01 -4.61417437e-01
2.42692426e-01 4.04768214e-02 -5.90043783e-01 2.40100220e-01
-9.59341705e-01 -5.23669362e-01 9.71498847e-01 4.12928522e-01
-5.38506866e-01 -8.42371762e-01 1.50779977e-01 -3.22678834e-01
-8.15869987e-01 -1.57956898e+00 8.36819038e-02 -5.88954151e-01
-6.64285541e-01 1.32221401e+00 7.70602405e-01 -1.37762022e+00
-8.24175835e-01 -1.13398075e+00 -8.74790788e-01 3.23408365e-01
-7.78405905e-01 -1.08292878e+00 -4.14670229e-01 3.65188241e-01
3.68458807e-01 -1.17399085e+00 1.22683537e+00 -9.96155962e-02
-2.87520826e-01 -5.75410545e-01 -6.62289798e-01 -2.16985047e-01
-1.43213615e-01 -8.92313480e-01 1.51444986e-01 6.91459894e-01
-9.14703965e-01 4.35872644e-01 1.10877264e+00 -9.64745939e-01
-1.57241917e+00 -2.93996722e-01 1.29537687e-01 -3.96195590e-01
8.07982624e-01 -1.58079103e-01 1.54337794e-01 4.95872408e-01
8.63803029e-01 -8.07799459e-01 7.81055927e-01 -3.14169854e-01
5.06541610e-01 -4.46309708e-02 -1.43625486e+00 9.82072651e-01
6.19745433e-01 -5.59408497e-03 -6.84809566e-01 7.49718845e-01
5.99903226e-01 -2.19277397e-01 -7.23242044e-01 4.57340717e-01
2.69393653e-01 -6.60755873e-01 9.08608139e-01 -1.33892548e+00
-4.00249243e-01 -7.22023010e-01 -2.33236656e-01 -1.22842443e+00
-2.23757803e-01 -1.17202342e+00 -6.33472903e-03 3.26742709e-01
2.98910588e-01 -1.23927701e+00 3.60087335e-01 1.28997624e+00
-3.20517510e-01 -1.97014168e-01 -8.88765275e-01 -7.24100530e-01
-4.50859845e-01 -1.76854774e-01 7.42497146e-01 8.70131731e-01
3.42925727e-01 1.56465590e-01 -5.44902831e-02 3.37162644e-01
6.19390428e-01 -4.42507327e-01 8.26604009e-01 -1.28241503e+00
-2.27847233e-01 -2.23902971e-01 -8.36918652e-01 -2.05327868e-01
1.98823050e-01 -2.28270918e-01 -2.43002445e-01 -1.40561914e+00
-7.54731357e-01 -5.65639615e-01 1.90583542e-01 2.89519072e-01
7.68530369e-01 -7.32443273e-01 5.97139537e-01 -3.18341367e-02
-1.06262982e-01 2.30626285e-01 9.11839306e-01 -4.35960703e-02
-5.34358853e-03 2.80414134e-01 -2.07136825e-01 8.90603483e-01
1.23163962e+00 -3.76173288e-01 -5.92394769e-01 -4.03252803e-02
2.22942814e-01 6.99414134e-01 3.77892524e-01 -1.15337849e+00
6.42970741e-01 -9.08314347e-01 -3.77669573e-01 -4.55716640e-01
5.33035100e-01 -1.51047230e+00 4.50507343e-01 1.19279730e+00
-5.94401322e-02 9.79247630e-01 1.56579301e-01 2.35898450e-01
-4.04931456e-01 -9.06051934e-01 3.76307696e-01 -2.74171382e-01
-7.94207156e-01 -2.45066553e-01 -1.21751618e+00 -7.22435892e-01
1.97210813e+00 -2.46277168e-01 -1.02654707e+00 -6.30701900e-01
-7.13951468e-01 3.15274894e-01 2.08292812e-01 4.73015793e-02
5.94045579e-01 -7.89077759e-01 -7.16672480e-01 -1.18912691e-02
-5.33911169e-01 -7.27589965e-01 3.84321995e-02 4.80828673e-01
-1.42114449e+00 7.62770295e-01 -8.66461456e-01 2.94585317e-01
-1.33720791e+00 5.68365097e-01 7.06607878e-01 -3.31743032e-01
-8.83932412e-02 2.38142878e-01 -3.27224173e-02 -4.49252754e-01
4.41803411e-03 3.37678552e-01 -6.17512345e-01 -2.25586906e-01
9.03649867e-01 5.83892226e-01 -1.40143275e-01 -4.74157751e-01
-5.68986237e-01 9.39047486e-02 -8.34862888e-02 -9.90222275e-01
1.11250103e+00 3.71363282e-01 -3.22507650e-01 -1.15209915e-01
-1.43977702e-01 3.48027162e-02 -4.63695645e-01 6.79603040e-01
-8.08279067e-02 -2.27828830e-01 3.36250216e-01 -1.24124324e+00
-2.81301647e-01 3.88765872e-01 6.94628119e-01 7.94117987e-01
9.05583739e-01 -8.95401537e-01 4.71943803e-02 5.28414965e-01
9.04747665e-01 -1.40113926e+00 -3.55752498e-01 5.76947927e-01
1.33339238e+00 -6.76193118e-01 4.69095618e-01 -4.07098562e-01
-1.21300960e+00 1.34876251e+00 6.62094951e-01 -1.85080573e-01
7.85915077e-01 7.54171491e-01 1.79732323e-01 -9.52565014e-01
-1.09587741e+00 -9.49323848e-02 -1.04835533e-01 1.02696562e+00
-6.40648827e-02 5.06311476e-01 -6.92779839e-01 4.01003063e-01
-5.03916629e-02 -1.34212866e-01 8.80401790e-01 1.68084610e+00
-7.41056204e-01 -1.23044336e+00 -7.16800749e-01 -2.52265543e-01
-1.43518895e-01 1.62819818e-01 -8.42960835e-01 1.34603643e+00
1.57789826e-01 1.18101728e+00 -1.15269937e-01 -5.44586062e-01
6.94952309e-01 -2.85678029e-01 2.20610172e-01 -3.34519476e-01
-1.39530253e+00 -6.86110198e-01 9.85614419e-01 -2.68129349e-01
-3.69104624e-01 -4.40087140e-01 -1.54357433e+00 -5.46344519e-01
2.27021396e-01 8.06953549e-01 1.33295226e+00 5.49448311e-01
1.94210202e-01 7.20222533e-01 6.16297603e-01 -8.01501095e-01
-7.38481879e-01 -7.19563723e-01 -5.15346885e-01 -3.72279882e-01
-7.45694041e-02 -8.68257880e-01 -2.65742719e-01 -5.59776127e-01] | [3.835899829864502, 1.7008708715438843] |
3bdd2031-70c9-44f5-b054-a2a8e3b32472 | mitigating-biased-activation-in-weakly | 2305.15354 | null | https://arxiv.org/abs/2305.15354v1 | https://arxiv.org/pdf/2305.15354v1.pdf | Mitigating Biased Activation in Weakly-supervised Object Localization via Counterfactual Learning | In this paper, we focus on an under-explored issue of biased activation in prior weakly-supervised object localization methods based on Class Activation Mapping (CAM). We analyze the cause of this problem from a causal view and attribute it to the co-occurring background confounders. Following this insight, we propose a novel Counterfactual Co-occurring Learning (CCL) paradigm to synthesize the counterfactual representations via coupling constant foreground and unrealized backgrounds in order to cut off their co-occurring relationship. Specifically, we design a new network structure called Counterfactual-CAM, which embeds the counterfactual representation perturbation mechanism into the vanilla CAM-based model. This mechanism is responsible for decoupling foreground as well as background and synthesizing the counterfactual representations. By training the detection model with these synthesized representations, we compel the model to focus on the constant foreground content while minimizing the influence of distracting co-occurring background. To our best knowledge, it is the first attempt in this direction. Extensive experiments on several benchmarks demonstrate that Counterfactual-CAM successfully mitigates the biased activation problem, achieving improved object localization accuracy. | ['Jun Xiao', 'Yi Yang', 'Ping Liu', 'Lei Chen', 'Yawei Luo', 'Feifei Shao'] | 2023-05-24 | null | null | null | null | ['object-localization', 'weakly-supervised-object-localization'] | ['computer-vision', 'computer-vision'] | [ 7.19301581e-01 4.17014629e-01 -3.84642988e-01 -1.71190515e-01
-4.04750764e-01 -3.08335155e-01 9.55971599e-01 -1.18922509e-01
-3.45028102e-01 8.61969411e-01 3.42697531e-01 -4.98641163e-01
6.16600104e-02 -8.06986868e-01 -1.05256772e+00 -8.46710205e-01
-1.24840662e-01 4.77756461e-04 1.92354739e-01 1.24434695e-01
2.21655965e-01 3.33661586e-01 -1.47696936e+00 6.63707495e-01
9.45756912e-01 6.02996111e-01 1.62509282e-03 1.82256058e-01
-1.21739380e-01 9.81620014e-01 -7.42924690e-01 -4.09596503e-01
3.97546440e-01 -7.16076851e-01 -5.69114566e-01 -4.92162257e-02
5.95042408e-01 -1.79824501e-01 -6.23677447e-02 1.12594354e+00
4.32467878e-01 -1.91207137e-02 6.90730572e-01 -1.33122289e+00
-8.94461989e-01 9.35930729e-01 -1.09247088e+00 6.57823801e-01
-9.48291570e-02 3.63966554e-01 1.05403948e+00 -8.89147341e-01
3.25565726e-01 1.57250226e+00 2.53656715e-01 7.14968204e-01
-1.56575584e+00 -9.29451942e-01 9.99508083e-01 2.71852855e-02
-1.30243134e+00 -2.37710983e-01 1.16679788e+00 -4.18831706e-01
4.30329531e-01 3.05450112e-01 6.73293591e-01 1.47805190e+00
5.62360091e-03 1.00460458e+00 1.34365714e+00 -7.22127974e-01
3.37488413e-01 1.38711855e-01 -3.89847159e-02 6.19779110e-01
5.59184134e-01 2.45018959e-01 -3.93241525e-01 -1.87715188e-01
5.68696260e-01 -1.11035712e-01 -5.61349392e-01 -6.48008764e-01
-1.24188137e+00 8.19593608e-01 7.33683228e-01 2.77879030e-01
-2.44989097e-01 3.82305324e-01 8.65504369e-02 -1.81574702e-01
7.55555153e-01 5.61686039e-01 -2.47331947e-01 6.82660580e-01
-7.85597086e-01 2.81744123e-01 3.20656270e-01 5.69626033e-01
5.31438708e-01 1.48064628e-01 -6.05262458e-01 5.65043271e-01
3.16314459e-01 4.51212913e-01 5.68718672e-01 -5.93477249e-01
4.54757422e-01 6.80034637e-01 1.73493534e-01 -9.56258774e-01
-1.36534765e-01 -6.54722452e-01 -5.79927385e-01 4.45234001e-01
5.44896841e-01 -2.21195981e-01 -1.03074753e+00 2.15595794e+00
5.96349418e-01 7.54185855e-01 -1.97823375e-01 1.09380019e+00
3.60197872e-01 5.45156062e-01 4.67810959e-01 -3.23829025e-01
1.19787967e+00 -8.02596867e-01 -7.09816456e-01 -3.11564296e-01
5.12875021e-01 -2.21754953e-01 1.31146288e+00 1.06157504e-01
-7.21539736e-01 -3.64962518e-01 -1.37468815e+00 3.08031976e-01
-4.30255651e-01 -2.24986556e-03 6.79702342e-01 7.17860818e-01
-6.02571726e-01 3.91451269e-01 -6.20197117e-01 1.82533097e-02
8.46160114e-01 1.30671293e-01 -3.88896652e-02 3.77153866e-02
-1.31265450e+00 1.00119698e+00 4.68478680e-01 2.18045533e-01
-1.11218953e+00 -1.00724268e+00 -7.93251514e-01 -1.35906087e-02
8.95417988e-01 -6.75325394e-01 9.29462254e-01 -1.59709716e+00
-1.25747788e+00 8.96246076e-01 -3.36162895e-02 -6.07074142e-01
6.65638745e-01 -2.00841874e-01 -2.73097754e-01 -1.05067842e-01
1.03643797e-01 6.00457728e-01 1.08474743e+00 -1.71397769e+00
-3.85815442e-01 -4.54880506e-01 2.22413361e-01 2.16336653e-01
-2.29740828e-01 -6.82060868e-02 -9.06735361e-02 -9.38482881e-01
-9.90730673e-02 -7.55050302e-01 -1.84567019e-01 9.26661789e-02
-2.92816758e-01 -4.30888459e-02 7.61214733e-01 -1.75970018e-01
1.08950877e+00 -2.12916589e+00 -1.00950181e-01 -1.00322478e-02
3.82540315e-01 2.18603715e-01 -1.96295395e-01 -3.76410484e-01
-5.19247174e-01 2.04319969e-01 -4.80125159e-01 1.37914252e-02
-1.82078794e-01 1.21375531e-01 -7.55137563e-01 7.75154471e-01
6.47392333e-01 7.84408808e-01 -1.22279704e+00 -4.74661708e-01
4.66240123e-02 3.95897508e-01 -6.69495404e-01 8.22105408e-02
-4.28396642e-01 2.80279428e-01 -3.00926685e-01 4.93832320e-01
8.95429850e-01 3.84297967e-02 3.41391295e-01 -1.86530337e-01
-1.21150255e-01 2.58554071e-01 -1.15793550e+00 1.21068919e+00
-3.98800641e-01 4.09043163e-01 5.57516590e-02 -1.11331451e+00
6.53088391e-01 1.02947168e-01 2.17581794e-01 -6.72436118e-01
2.13794291e-01 6.87301159e-02 3.70521784e-01 -3.00533533e-01
1.05468254e-03 -5.84626734e-01 5.39647155e-02 3.05523932e-01
-1.27322510e-01 3.58244926e-01 -1.13915488e-01 1.94382414e-01
7.35720992e-01 3.86087567e-01 5.19098461e-01 -4.40551549e-01
5.49205422e-01 -1.59014672e-01 6.88776135e-01 1.00811207e+00
-4.19739723e-01 5.17103195e-01 6.86987340e-01 -3.03264469e-01
-4.27463651e-01 -1.30797756e+00 4.23316620e-02 1.10246062e+00
2.38051802e-01 5.24926968e-02 -7.71852314e-01 -1.13767934e+00
9.83503908e-02 1.02871680e+00 -9.68170583e-01 -6.19127572e-01
-8.70083630e-01 -1.30025268e+00 3.84274721e-01 5.64803541e-01
4.57370460e-01 -1.07010794e+00 -7.74264097e-01 5.20027988e-02
-2.38068804e-01 -6.20634437e-01 -3.38365793e-01 1.71064451e-01
-5.54401219e-01 -1.08217776e+00 -7.88107395e-01 -2.23035693e-01
6.39186144e-01 4.17413414e-01 1.02927899e+00 5.13257720e-02
-3.18199068e-01 6.69927569e-03 -1.36461677e-02 -8.23734224e-01
-2.41948754e-01 -3.74420822e-01 -9.18376446e-02 4.60044175e-01
2.47282669e-01 -4.42643523e-01 -6.81673050e-01 8.15616101e-02
-8.85283470e-01 2.09869251e-01 6.05370343e-01 9.33629572e-01
4.09756929e-01 -5.60092069e-02 6.12830043e-01 -1.06415546e+00
3.46124709e-01 -6.65218949e-01 -7.03101814e-01 2.04389334e-01
-5.88755071e-01 9.50866789e-02 4.99795169e-01 -9.04126048e-01
-1.63473034e+00 4.45829183e-02 3.33595634e-01 -5.38363814e-01
-9.11408588e-02 1.05246704e-03 -6.83960497e-01 2.81933278e-01
7.54013002e-01 -5.70659526e-02 -4.54631209e-01 -1.78543553e-01
8.17320466e-01 1.91053003e-01 6.71223581e-01 -6.85193419e-01
7.13491857e-01 1.07233524e+00 -2.09116101e-01 -3.02283615e-01
-1.36364532e+00 -1.97695315e-01 -4.54148978e-01 -2.28809595e-01
8.09364736e-01 -7.46999800e-01 -4.46478426e-01 9.94316563e-02
-1.37335968e+00 -3.21584940e-01 -4.12818074e-01 3.34839404e-01
-3.04462790e-01 2.39626411e-02 2.66242176e-02 -1.18947935e+00
2.32876047e-01 -8.25490952e-01 8.80726576e-01 4.41877283e-02
-7.17557967e-02 -5.88181555e-01 6.33043200e-02 1.47791296e-01
2.15667024e-01 4.83486414e-01 1.00143957e+00 -5.20436287e-01
-6.56974673e-01 1.97847962e-01 -4.53933060e-01 8.73080194e-02
1.36307508e-01 -2.53300995e-01 -1.37224352e+00 -5.87489046e-02
1.56981319e-01 1.59266844e-01 1.35855305e+00 4.29214776e-01
1.33799994e+00 -5.38968027e-01 -5.24410248e-01 4.22369093e-01
1.34581006e+00 2.08579510e-01 5.02227545e-01 2.87373185e-01
7.57873356e-01 7.98785269e-01 5.28778076e-01 1.09427989e-01
1.93544198e-02 6.35777235e-01 8.74793947e-01 -4.50549781e-01
-2.80865043e-01 -3.83344054e-01 2.61151761e-01 -2.62971252e-01
4.00715545e-02 -2.75439203e-01 -7.68674850e-01 8.10089350e-01
-1.91481113e+00 -1.12551403e+00 -1.47428155e-01 2.15354276e+00
8.47305477e-01 4.04478222e-01 5.57870902e-02 5.67473704e-03
1.02521861e+00 4.34049696e-01 -5.13552487e-01 -7.63057917e-02
-1.26341552e-01 2.02268385e-03 4.64133382e-01 4.67626661e-01
-1.25666285e+00 8.93027186e-01 5.66286850e+00 6.21416092e-01
-1.15429235e+00 2.94055402e-01 7.62419045e-01 -4.40824211e-01
-4.03134346e-01 8.73373076e-02 -7.39001751e-01 6.32327795e-01
6.48621261e-01 -5.84639842e-03 2.38679841e-01 8.84993792e-01
3.15342963e-01 -1.00751251e-01 -1.12606144e+00 6.50494456e-01
2.13850901e-01 -1.36454511e+00 4.86004859e-01 -5.54052666e-02
6.15739107e-01 -3.17078561e-01 1.22085303e-01 3.67686778e-01
4.16947007e-01 -9.53501999e-01 1.15373361e+00 2.82529414e-01
5.21288335e-01 -5.37604451e-01 5.94369531e-01 3.18581074e-01
-8.21652472e-01 -3.08969498e-01 -9.73597169e-02 -2.34391227e-01
-4.48486879e-02 8.10315549e-01 -7.09667146e-01 1.74611598e-01
5.43297410e-01 1.96338966e-01 -4.29493815e-01 7.11074233e-01
-4.48595613e-01 7.71791995e-01 -1.59191876e-03 1.37804121e-01
-1.50333876e-02 1.83051482e-01 6.79329634e-01 1.39165735e+00
-1.77747533e-01 -3.16365808e-02 -3.27474158e-03 1.50770962e+00
-1.68163449e-01 -1.47195607e-01 -8.25073719e-01 3.43705446e-01
2.63221532e-01 9.82794523e-01 -1.00909591e+00 -5.10375023e-01
-3.29793662e-01 9.20452297e-01 4.54021960e-01 5.33403039e-01
-1.27524400e+00 3.09909787e-03 5.17768145e-01 3.30921352e-01
1.97529271e-01 3.66401315e-01 -5.41797042e-01 -1.24686694e+00
2.68646479e-02 -9.35455680e-01 5.25146008e-01 -5.46237707e-01
-1.27684057e+00 2.31900781e-01 4.07376677e-01 -7.80532718e-01
1.88739896e-01 -6.67612195e-01 -7.41715074e-01 7.78203428e-01
-1.78206050e+00 -1.09263325e+00 -6.32696822e-02 5.38879275e-01
5.19073069e-01 3.64268184e-01 3.39275122e-01 1.06283732e-01
-8.44736814e-01 3.19381058e-01 -5.51626861e-01 -6.71616867e-02
7.28486836e-01 -1.30607998e+00 -6.94456473e-02 1.15973556e+00
1.93605900e-01 8.65213513e-01 6.93125606e-01 -7.11861849e-01
-8.65826547e-01 -1.22880137e+00 5.92836976e-01 -5.24899483e-01
6.33707464e-01 -7.14077532e-01 -1.01448083e+00 5.59818566e-01
2.12025940e-02 3.18498760e-01 3.48232746e-01 4.64409105e-02
-6.69923127e-01 -1.82365730e-01 -1.09659111e+00 9.19816315e-01
1.20362425e+00 -3.11641097e-01 -9.45819497e-01 1.46002829e-01
8.35117042e-01 -3.71161215e-02 3.15863788e-01 4.43724990e-01
2.52448708e-01 -8.06302905e-01 1.12721241e+00 -6.21867061e-01
4.28460658e-01 -5.80602109e-01 1.43055305e-01 -1.25182402e+00
-3.60102028e-01 -4.37998533e-01 -3.23920697e-01 1.24503374e+00
3.44070435e-01 -6.07926011e-01 7.15147078e-01 2.31936052e-01
5.92779601e-03 -6.22173548e-01 -9.97423410e-01 -7.35699058e-01
9.90404338e-02 -3.90541166e-01 5.98070860e-01 1.09994721e+00
-2.73628742e-01 2.83616364e-01 -2.48183101e-01 4.34956789e-01
6.01826489e-01 2.49497727e-01 6.05068266e-01 -9.75127399e-01
-4.17305470e-01 -4.75733131e-01 7.26615191e-02 -6.68188214e-01
3.62536579e-01 -8.23226035e-01 3.13564479e-01 -1.00113690e+00
4.18044508e-01 -1.56187400e-01 -5.83119094e-01 4.97074902e-01
-6.04055762e-01 2.22796783e-01 1.21349543e-01 3.13648605e-03
-3.54137540e-01 5.48816979e-01 1.21348107e+00 -2.53559351e-01
-1.25456810e-01 -2.83378959e-01 -1.13773978e+00 1.09277225e+00
6.88312232e-01 -7.00012565e-01 -5.33807456e-01 -2.43044853e-01
-1.67271271e-01 -4.85697895e-01 8.58934283e-01 -5.53843319e-01
-2.49488160e-01 -5.37975848e-01 4.52104151e-01 -2.17519045e-01
5.42267188e-02 -6.76285803e-01 -2.48438537e-01 6.03361726e-01
-5.27893007e-01 -3.60230654e-01 2.09102884e-01 9.51283574e-01
1.54423609e-01 1.62767936e-02 1.04684770e+00 -3.32063645e-01
-5.69990814e-01 -2.21190959e-01 -3.38570148e-01 5.85716516e-02
1.11411524e+00 -1.36757404e-01 -4.76984382e-01 2.94142682e-02
-4.36062396e-01 4.64675762e-02 1.28721952e-01 3.93229991e-01
4.41106886e-01 -1.17921817e+00 -5.67731977e-01 1.56673282e-01
9.77602974e-02 -2.81346560e-01 2.94084530e-02 8.26068461e-01
7.26680234e-02 2.58656293e-01 6.48086146e-02 -4.05322403e-01
-9.57416892e-01 9.81027424e-01 5.24047792e-01 -3.50283265e-01
-4.89337027e-01 8.27001095e-01 1.00109434e+00 -7.69704655e-02
1.54159769e-01 -4.59111631e-01 -1.54937193e-01 4.01852503e-02
6.02697909e-01 3.16823512e-01 -1.75954357e-01 -3.09425265e-01
-5.15676856e-01 1.55977942e-02 -8.83902162e-02 -1.78243607e-01
9.81775165e-01 5.27689978e-02 -3.29209194e-02 3.82451952e-01
8.75127017e-01 2.42893174e-01 -1.47042155e+00 1.46031514e-01
2.17600048e-01 -6.41658843e-01 1.52547240e-01 -9.76409674e-01
-9.98537004e-01 8.23566556e-01 7.10062742e-01 -8.12015589e-03
1.16683316e+00 1.64501801e-01 1.12281710e-01 -4.73079756e-02
1.09426953e-01 -8.47264409e-01 2.95593143e-01 9.11029503e-02
9.96885836e-01 -1.20058227e+00 2.42113341e-02 -5.62739074e-01
-5.00573456e-01 4.44899619e-01 1.05074787e+00 -4.60847676e-01
4.49026883e-01 3.28199506e-01 1.73050631e-02 -2.84774989e-01
-6.60765827e-01 -2.83144355e-01 2.53512859e-01 6.52575791e-01
4.74380761e-01 5.01030348e-02 -4.22387749e-01 5.90317547e-01
1.91412732e-01 -1.53403014e-01 3.57830197e-01 7.21077323e-01
-1.54620215e-01 -5.52267551e-01 -6.68424606e-01 1.41522333e-01
-6.00484252e-01 -2.95284748e-01 -6.33752763e-01 1.05152953e+00
7.05712318e-01 7.15710104e-01 1.72351390e-01 1.91975161e-01
4.09228653e-01 8.27689990e-02 4.77991283e-01 -7.59923398e-01
-4.89032716e-01 2.57513583e-01 -1.55127302e-01 -5.21531165e-01
-5.71079850e-01 -3.89210343e-01 -1.08042777e+00 3.77647042e-01
-6.06847525e-01 -1.59545779e-01 2.86891550e-01 8.96253467e-01
1.61151856e-01 8.16506088e-01 3.23486596e-01 -8.23686719e-01
-6.94930673e-01 -7.35817611e-01 -3.69391054e-01 5.23856461e-01
5.99747598e-01 -1.13640809e+00 -6.63351893e-01 6.60534715e-03] | [9.886211395263672, 1.7273269891738892] |
cd0e5cca-76e4-4633-818d-0473044fc2fc | deep-depth-completion-a-survey | 2205.05335 | null | https://arxiv.org/abs/2205.05335v3 | https://arxiv.org/pdf/2205.05335v3.pdf | Deep Depth Completion from Extremely Sparse Data: A Survey | Depth completion aims at predicting dense pixel-wise depth from an extremely sparse map captured from a depth sensor, e.g., LiDARs. It plays an essential role in various applications such as autonomous driving, 3D reconstruction, augmented reality, and robot navigation. Recent successes on the task have been demonstrated and dominated by deep learning based solutions. In this article, for the first time, we provide a comprehensive literature review that helps readers better grasp the research trends and clearly understand the current advances. We investigate the related studies from the design aspects of network architectures, loss functions, benchmark datasets, and learning strategies with a proposal of a novel taxonomy that categorizes existing methods. Besides, we present a quantitative comparison of model performance on three widely used benchmarks, including indoor and outdoor datasets. Finally, we discuss the challenges of prior works and provide readers with some insights for future research directions. | ['Tin Lun Lam', 'Honghai Liu', 'Qing Gao', 'Chenyou Fan', 'Mete Ozay', 'Chenyu Bao', 'Junjie Hu'] | 2022-05-11 | null | null | null | null | ['depth-completion'] | ['computer-vision'] | [ 3.77747148e-01 1.52696073e-02 -4.87749785e-01 -7.52597511e-01
-6.59781754e-01 -1.23425767e-01 4.14430857e-01 -1.99560031e-01
-5.62776685e-01 7.67592430e-01 1.25245541e-01 -2.59206980e-01
-1.75752252e-01 -7.95882940e-01 -6.86955690e-01 -5.48208594e-01
-3.37544501e-01 2.09339797e-01 -1.93447247e-02 -5.06955795e-02
5.10916829e-01 5.08732736e-01 -1.72334278e+00 -4.15855832e-02
8.16201806e-01 1.45426023e+00 4.32264179e-01 3.83018613e-01
-6.70025591e-03 9.22960699e-01 -2.31528729e-01 -3.20033103e-01
3.92687678e-01 2.06996009e-01 -4.41982508e-01 -9.39886495e-02
6.46129310e-01 -7.11335838e-01 -8.25797915e-01 1.04921246e+00
4.31737036e-01 7.13716522e-02 4.40780908e-01 -1.50138879e+00
-3.04919273e-01 2.57207900e-01 -6.62244558e-01 2.10658964e-02
2.86293536e-01 4.48868014e-02 9.21810150e-01 -1.05840349e+00
5.32236695e-01 1.13364804e+00 7.30690360e-01 7.49915242e-01
-5.86424291e-01 -7.42769718e-01 4.61632162e-01 6.54388368e-01
-1.24593520e+00 -4.61501867e-01 8.72586727e-01 -3.15548509e-01
1.08763373e+00 -1.10734023e-01 5.74920416e-01 1.21897197e+00
3.52503777e-01 1.09462357e+00 8.80755544e-01 -1.29812481e-02
3.91570181e-01 -1.48639619e-01 1.69039458e-01 7.02675462e-01
4.42772806e-01 2.44993418e-01 -7.61624455e-01 3.22552145e-01
7.39547968e-01 -4.56960984e-02 -2.92334497e-01 -8.36363733e-01
-9.69506204e-01 8.61673176e-01 7.42227912e-01 -2.51705796e-01
-2.95641690e-01 3.25658739e-01 2.97490150e-01 1.07940651e-01
4.21672851e-01 4.16039266e-02 -5.81279278e-01 -1.55577496e-01
-6.39284432e-01 3.81002367e-01 5.83002210e-01 1.16450870e+00
1.04758298e+00 6.07596971e-02 4.16105807e-01 7.79788196e-01
5.43691814e-01 4.22638953e-01 2.31348813e-01 -1.28580117e+00
7.01836288e-01 3.31265509e-01 -2.33930685e-02 -7.79342473e-01
-5.64520717e-01 -3.23260486e-01 -1.00705516e+00 3.63918722e-01
1.61492795e-01 -1.56186923e-01 -8.28147531e-01 1.46693277e+00
4.50914390e-02 4.84172970e-01 -9.47984867e-03 9.49121594e-01
1.10397291e+00 3.38844627e-01 -3.69746238e-01 1.93534419e-01
8.80193174e-01 -1.07204473e+00 -5.86486578e-01 -9.02510941e-01
2.66041160e-01 -2.44981393e-01 9.37384009e-01 6.34969532e-01
-9.13901567e-01 -6.07114077e-01 -1.41325808e+00 -6.27161443e-01
-2.97685236e-01 2.82291602e-02 1.03316092e+00 4.93925214e-01
-1.13988817e+00 6.14971817e-01 -1.22321653e+00 -4.98647332e-01
5.49666703e-01 3.39709699e-01 -2.59644002e-01 -6.87039375e-01
-9.28157508e-01 8.12932551e-01 -1.99992172e-02 2.42679998e-01
-9.81983185e-01 -5.25693059e-01 -1.27804995e+00 -3.59808803e-01
1.45472571e-01 -9.09353614e-01 1.42076182e+00 8.77495110e-02
-1.44128048e+00 8.74771774e-01 -4.37746733e-01 -7.62161851e-01
5.50064683e-01 -4.40349132e-01 1.79352891e-02 -1.01383574e-01
2.52466768e-01 1.03748679e+00 3.45249236e-01 -1.15093720e+00
-1.20309210e+00 -6.90300822e-01 2.36725926e-01 3.25281590e-01
-9.10886228e-02 -5.13499677e-01 -6.72077239e-01 -4.77551343e-03
6.74407303e-01 -7.98384011e-01 -6.18420780e-01 3.99313122e-01
-4.27367300e-01 1.01510109e-03 6.34711564e-01 -1.21038295e-01
8.95983994e-01 -2.13450384e+00 1.49605080e-01 -1.33715287e-01
5.96160948e-01 -3.66983712e-01 1.09183155e-01 3.37898970e-01
2.91064352e-01 -1.58958048e-01 -4.57736403e-01 -1.06157923e+00
9.15035829e-02 3.70726526e-01 -3.00339520e-01 5.93786299e-01
-1.25268996e-01 8.94673288e-01 -8.60137582e-01 -9.33142528e-02
5.34188747e-01 5.60034573e-01 -3.78429383e-01 1.09586474e-02
-5.28438091e-02 4.00705487e-01 -3.56135875e-01 7.91510701e-01
7.27333426e-01 -1.28139913e-01 -2.80721515e-01 -5.50863184e-02
-2.48783544e-01 7.40232408e-01 -8.42502713e-01 2.18944025e+00
-6.52782440e-01 1.12760985e+00 3.03748667e-01 -9.52868044e-01
1.01099110e+00 -2.26239383e-01 5.33258796e-01 -9.25031364e-01
-4.76702629e-03 4.38680500e-01 -2.57558346e-01 -3.38324517e-01
8.46202493e-01 9.62368175e-02 -8.72244872e-03 -1.47347199e-02
-3.07514220e-01 -3.65532994e-01 -9.98575836e-02 -3.07430718e-02
1.24354172e+00 1.23127684e-01 2.57309765e-01 2.03150570e-01
3.68326485e-01 2.63473783e-02 6.20820343e-01 7.35271931e-01
-5.91482222e-01 6.68543637e-01 1.62157878e-01 -5.63659310e-01
-7.64285266e-01 -1.10656548e+00 -3.98229122e-01 4.67377394e-01
6.13152683e-01 -2.80648619e-01 -2.20109656e-01 -3.78794074e-01
3.47777367e-01 3.42007041e-01 -6.58919990e-01 -1.02005877e-01
-4.44241732e-01 -6.11341059e-01 4.50522780e-01 8.52402389e-01
9.01217818e-01 -8.14128935e-01 -7.43418276e-01 -1.14292979e-01
-2.66696095e-01 -1.46951008e+00 3.19538206e-01 4.06295747e-01
-1.35736358e+00 -1.00970399e+00 -4.64219540e-01 -8.57968628e-01
2.01084033e-01 7.71187127e-01 1.28920031e+00 -2.66657710e-01
-6.33287504e-02 3.56735796e-01 -4.68888693e-02 -4.95543331e-01
4.36398387e-01 2.64094532e-01 7.36787692e-02 -6.25082195e-01
7.73569703e-01 -7.45840847e-01 -8.03809822e-01 1.16648838e-01
-5.45279145e-01 3.44524123e-02 5.66940725e-01 5.37796199e-01
8.59287798e-01 -4.56881477e-03 4.39720780e-01 -6.94515169e-01
4.58436251e-01 -5.53928196e-01 -7.05066442e-01 -4.65535641e-01
-8.61450851e-01 -3.05163473e-01 8.91005099e-02 3.19722980e-01
-8.39862645e-01 1.01605892e-01 -5.13022542e-01 -2.20583633e-01
-2.19183624e-01 6.17901444e-01 -2.47980982e-01 -3.54414225e-01
6.04958296e-01 1.29667446e-01 -9.65917781e-02 -4.98411477e-01
2.28226468e-01 5.82177639e-01 7.06008255e-01 -3.35429788e-01
5.07278264e-01 9.96762693e-01 8.22504535e-02 -9.91121769e-01
-8.37026596e-01 -6.00375235e-01 -5.35857141e-01 -5.68678277e-03
3.91528636e-01 -1.31257379e+00 -5.82092643e-01 7.21834302e-01
-1.24012172e+00 -5.84709466e-01 -7.74090290e-02 5.86775303e-01
-7.86504626e-01 2.83412516e-01 -8.52825165e-01 -6.62462056e-01
-9.42818597e-02 -1.19827032e+00 1.26457465e+00 2.98305660e-01
1.52959794e-01 -1.07205069e+00 1.83167029e-02 3.86147141e-01
2.94827223e-01 2.63770670e-01 6.50493622e-01 -5.25216293e-03
-1.18189776e+00 -9.18585956e-02 -4.28676665e-01 3.77055585e-01
1.01621948e-01 -2.53796607e-01 -1.21653712e+00 -2.18426704e-01
5.26482016e-02 -3.03201407e-01 1.25630081e+00 7.51972973e-01
1.24831796e+00 2.56664336e-01 -6.25905275e-01 1.27042222e+00
1.57194149e+00 2.12098256e-01 8.57879877e-01 6.80753171e-01
8.73885155e-01 7.88287282e-01 7.55813360e-01 4.40314472e-01
9.59032595e-01 2.79776096e-01 9.70816076e-01 2.41927788e-01
-1.30734771e-01 -2.73804307e-01 2.93445028e-02 5.39236426e-01
1.79443657e-01 -2.51680672e-01 -8.31067204e-01 6.29647851e-01
-1.78290665e+00 -4.27048147e-01 -1.24628410e-01 2.00641012e+00
1.85800195e-01 3.87541324e-01 -2.58634657e-01 2.84572065e-01
4.37247902e-01 4.62220430e-01 -1.07969737e+00 1.76516827e-02
-1.71417743e-01 9.11695883e-02 7.58949816e-01 6.26034617e-01
-1.14443600e+00 9.91539061e-01 6.72554970e+00 2.22540781e-01
-1.21892166e+00 -8.88355076e-02 3.72746527e-01 -3.54034305e-01
-2.70554274e-01 -2.60327399e-01 -9.60054338e-01 1.83456466e-01
5.29540837e-01 -7.22349137e-02 3.78033340e-01 1.13116884e+00
3.50115269e-01 -4.64733779e-01 -1.45175326e+00 1.38920605e+00
4.00944352e-02 -1.38229764e+00 -3.01988602e-01 2.52685279e-01
8.98050129e-01 9.45572317e-01 2.74524122e-01 3.00504982e-01
7.83614218e-02 -9.70240593e-01 5.13361096e-01 2.89388597e-01
9.06874597e-01 -6.67859674e-01 8.52258503e-01 4.55672801e-01
-1.23196280e+00 -2.66012937e-01 -5.66205263e-01 -7.76073456e-01
3.68364871e-01 7.70920217e-01 -4.16031748e-01 3.91854525e-01
9.86785650e-01 1.38739538e+00 -1.74833119e-01 1.33371186e+00
-5.34706235e-01 1.68155491e-01 -3.69840741e-01 9.36690196e-02
3.51978272e-01 -1.42931998e-01 2.62670219e-01 8.42366755e-01
3.13549012e-01 -1.90038055e-01 -5.50728152e-03 6.74410105e-01
-1.72219217e-01 -3.69402796e-01 -9.59128022e-01 4.39686865e-01
6.65676713e-01 9.85262394e-01 -4.09400553e-01 3.35593782e-02
-7.07662046e-01 7.77609229e-01 2.92941779e-01 3.71825546e-01
-5.38021505e-01 -5.44540823e-01 1.39812565e+00 1.11974344e-01
-4.91175847e-03 -7.63125956e-01 -9.24446642e-01 -1.05127871e+00
2.22019941e-01 -4.28759873e-01 -1.11010209e-01 -8.62414122e-01
-1.17439735e+00 4.27942365e-01 -8.28212723e-02 -1.08731818e+00
-3.07785451e-01 -9.49080050e-01 -3.87188613e-01 7.97857165e-01
-2.25246119e+00 -5.69064438e-01 -1.07106352e+00 4.23056930e-01
7.13074505e-01 2.73481421e-02 4.64654088e-01 6.60404861e-01
-6.26670778e-01 2.97833622e-01 1.69780403e-01 2.79302639e-03
5.23400128e-01 -1.17632926e+00 1.08992708e+00 7.13201284e-01
-1.37543708e-01 3.92625809e-01 5.01039565e-01 -5.72295785e-01
-1.45597386e+00 -1.05570698e+00 7.60931611e-01 -4.23153102e-01
4.25112605e-01 -3.49017709e-01 -5.99214077e-01 8.49431217e-01
-2.62859076e-01 1.39819682e-01 3.45940351e-01 1.04284145e-01
-2.54557040e-02 -3.11725050e-01 -1.14294124e+00 4.37866032e-01
1.50499225e+00 -4.88757163e-01 -3.80707271e-02 5.23481704e-02
7.08775640e-01 -8.45780194e-01 -3.29767704e-01 7.94219196e-01
8.32752228e-01 -1.55284524e+00 1.04635036e+00 -7.61574134e-02
7.06640244e-01 -1.50068164e-01 -5.25270104e-01 -1.00420916e+00
-2.06070408e-01 -2.19881877e-01 -4.28373784e-01 4.52434987e-01
4.12564129e-02 -7.01988339e-01 1.57040846e+00 4.81614858e-01
-6.30938649e-01 -1.12523139e+00 -9.75594997e-01 -5.77047348e-01
2.14453787e-02 -1.04241490e+00 4.88444567e-01 4.96969819e-01
-3.70013148e-01 4.10985768e-01 -2.92937756e-01 3.34173232e-01
9.05434370e-01 -1.98307484e-01 9.03249741e-01 -1.39595604e+00
2.13622689e-01 -4.53886479e-01 -6.80303037e-01 -2.04597878e+00
2.00130448e-01 -4.68984753e-01 2.14525759e-01 -2.19075680e+00
-9.14414376e-02 -6.35858297e-01 -1.32708967e-01 4.52782772e-02
2.63742983e-01 4.57052141e-01 -1.23386167e-01 1.62686124e-01
-5.41355312e-01 6.80496156e-01 1.15176976e+00 -2.94191450e-01
-1.54626042e-01 2.52526790e-01 -8.46049130e-01 8.94052684e-01
8.25792730e-01 -1.85031906e-01 -7.48009205e-01 -1.07540417e+00
3.38602602e-01 -5.73365614e-02 2.75964379e-01 -1.28372860e+00
3.92320424e-01 -3.34499069e-02 2.40790084e-01 -1.09963596e+00
7.45893419e-01 -8.69895875e-01 -4.35289085e-01 3.46975565e-01
6.20476380e-02 1.00580141e-01 1.78885266e-01 7.32350707e-01
-4.13643956e-01 3.90692838e-02 6.22340977e-01 -1.38310418e-01
-1.42252743e+00 6.73251033e-01 -6.60932437e-02 -1.99434713e-01
8.28494966e-01 -5.43445826e-01 -3.37842047e-01 -7.94968784e-01
-3.40569049e-01 6.35305226e-01 5.14097571e-01 5.26767671e-01
1.04834735e+00 -1.23330450e+00 -4.27384794e-01 2.84801066e-01
2.56850034e-01 6.07033849e-01 2.64997482e-01 5.43949425e-01
-6.23537362e-01 7.35110343e-01 -2.22702026e-01 -8.32619786e-01
-7.46895909e-01 3.06762278e-01 3.60689402e-01 6.07351176e-02
-6.56888127e-01 1.13576758e+00 4.30546522e-01 -4.79187518e-01
8.69655311e-01 -7.00038075e-01 -1.20000042e-01 -3.79533917e-01
4.56206888e-01 5.68272531e-01 2.21380904e-01 -2.87194133e-01
-5.67949533e-01 7.41743743e-01 1.11595258e-01 7.42979674e-03
1.29570818e+00 -5.75739980e-01 1.65500507e-01 6.42735720e-01
1.25063896e+00 -2.56696731e-01 -1.73992574e+00 -3.85769159e-01
-6.40525147e-02 -4.67010319e-01 1.61461070e-01 -5.81491709e-01
-1.23497963e+00 1.29706752e+00 5.49866736e-01 -2.17913494e-01
1.09141827e+00 -8.70577991e-02 9.47933197e-01 6.75585926e-01
1.00873780e+00 -7.89184809e-01 2.18998194e-02 9.20692503e-01
5.70028961e-01 -1.68071365e+00 4.09118086e-02 -4.41702574e-01
-2.15028241e-01 7.43955791e-01 8.29152703e-01 -1.86492532e-01
8.21244061e-01 3.28930795e-01 1.14754833e-01 -1.04733586e-01
-6.56515598e-01 -3.06492567e-01 -7.73181021e-02 9.39795613e-01
3.70934755e-01 -1.72862425e-01 1.38060421e-01 4.90065157e-01
-4.96661991e-01 4.30240482e-02 4.06860977e-01 9.62582052e-01
-6.52658820e-01 -9.52235878e-01 -7.45874792e-02 4.81981337e-01
-3.25668380e-02 -1.20645940e-01 -3.18932444e-01 8.09704006e-01
5.75813726e-02 1.01515746e+00 2.67130941e-01 -6.84034407e-01
3.84748399e-01 -3.78762305e-01 6.88983679e-01 -6.77940547e-01
2.70561039e-01 -5.08461654e-01 -8.89374167e-02 -8.75796854e-01
-2.60625511e-01 -6.00473642e-01 -1.13189113e+00 -5.35285294e-01
1.86700687e-01 -2.33844250e-01 1.12533689e+00 8.57147038e-01
3.11178237e-01 5.29017329e-01 5.30024171e-01 -1.11626637e+00
-2.56836355e-01 -7.77933121e-01 -6.55389726e-01 -3.24291259e-01
7.08118141e-01 -8.56996477e-01 -4.26105231e-01 -4.24642265e-01] | [8.539555549621582, -2.6261703968048096] |
9086e637-892d-47b2-a74f-9359d2da2c1e | 3d-pose-transfer-with-correspondence-learning | 2109.15025 | null | https://arxiv.org/abs/2109.15025v6 | https://arxiv.org/pdf/2109.15025v6.pdf | 3D Pose Transfer with Correspondence Learning and Mesh Refinement | 3D pose transfer is one of the most challenging 3D generation tasks. It aims to transfer the pose of a source mesh to a target mesh and keep the identity (e.g., body shape) of the target mesh. Some previous works require key point annotations to build reliable correspondence between the source and target meshes, while other methods do not consider any shape correspondence between sources and targets, which leads to limited generation quality. In this work, we propose a correspondence-refinement network to achieve the 3D pose transfer for both human and animal meshes. The correspondence between source and target meshes is first established by solving an optimal transport problem. Then, we warp the source mesh according to the dense correspondence and obtain a coarse warped mesh. The warped mesh will be better refined with our proposed Elastic Instance Normalization, which is a conditional normalization layer and can help to generate high-quality meshes. Extensive experimental results show that the proposed architecture can effectively transfer the poses from source to target meshes and produce better results with satisfied visual performance than state-of-the-art methods. | ['Guosheng Lin', 'Fayao Liu', 'Ruibo Li', 'Jiacheng Wei', 'Chaoyue Song'] | 2021-09-30 | null | http://proceedings.neurips.cc/paper/2021/hash/18a411989b47ed75a60ac69d9da05aa5-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/18a411989b47ed75a60ac69d9da05aa5-Paper.pdf | neurips-2021-12 | ['pose-transfer'] | ['computer-vision'] | [ 2.11110964e-01 -1.14412475e-02 1.37566119e-01 -3.14120054e-01
-4.52848256e-01 -2.55847305e-01 3.79593819e-01 -1.25966191e-01
-1.78426117e-01 5.65233231e-01 1.31260464e-02 4.21575993e-01
2.27848187e-01 -1.00087023e+00 -9.37615633e-01 -6.12390280e-01
2.55464524e-01 6.74883246e-01 5.60251594e-01 -3.49818558e-01
-4.29817941e-03 6.03439748e-01 -1.24717557e+00 -1.30309284e-01
7.71008670e-01 9.53127503e-01 2.01588973e-01 1.81756109e-01
1.69643108e-02 -1.04743190e-01 -3.70404363e-01 -3.56788486e-01
3.67312789e-01 -6.02536380e-01 -8.20660949e-01 1.52214214e-01
5.16832352e-01 -2.15339243e-01 -9.02237147e-02 1.25980628e+00
5.56716621e-01 1.37946919e-01 8.26839924e-01 -1.35663152e+00
-6.76992059e-01 2.48811707e-01 -8.47865820e-01 -2.62855768e-01
1.62386030e-01 -1.47476792e-01 4.98328924e-01 -8.23015392e-01
9.64942694e-01 1.47950447e+00 6.38545394e-01 6.75409317e-01
-1.36249018e+00 -1.03005087e+00 1.28449395e-01 -1.92171633e-01
-1.43031466e+00 -2.62083590e-01 9.51781273e-01 -5.41092455e-01
-5.98862395e-02 1.61242068e-01 7.86945462e-01 1.04404092e+00
2.95454592e-01 2.68586189e-01 8.97385180e-01 1.98935587e-02
-1.02733523e-01 -1.59875602e-01 -5.65410614e-01 7.21129298e-01
1.68423995e-01 -1.52786359e-01 -3.95207047e-01 6.18560286e-03
1.52076507e+00 -8.83239284e-02 -4.62893158e-01 -7.09708452e-01
-1.75015926e+00 4.35300589e-01 8.47132981e-01 2.33011156e-01
-4.91715163e-01 2.85632499e-02 3.21018040e-01 -1.21646300e-01
6.99561596e-01 2.26092383e-01 -2.51795471e-01 3.59813273e-01
-7.67196953e-01 4.86795276e-01 3.90215248e-01 1.13418961e+00
9.21200097e-01 -1.43465502e-02 -2.55137116e-01 7.23194361e-01
5.88895559e-01 6.98594093e-01 1.09975748e-01 -9.11515236e-01
4.92882848e-01 6.95632219e-01 1.10207558e-01 -1.51627529e+00
-2.47211054e-01 -3.48105758e-01 -1.36956370e+00 3.36969644e-01
3.31215948e-01 1.20679159e-02 -1.03912270e+00 1.77124178e+00
8.74827385e-01 4.35484231e-01 -1.62743911e-01 1.19037008e+00
9.57154870e-01 7.41206229e-01 1.02999575e-01 -1.24211125e-01
1.21477175e+00 -7.12666154e-01 -6.30353868e-01 -1.18212432e-01
-6.67801797e-02 -9.67278779e-01 6.88148975e-01 -1.12011820e-01
-1.12026596e+00 -7.75267363e-01 -1.01480806e+00 -1.66835725e-01
1.39506802e-01 5.79040684e-02 1.11715429e-01 -7.20368624e-02
-5.98978519e-01 8.14842522e-01 -9.06892002e-01 -3.31625730e-01
4.76685405e-01 2.49100447e-01 -6.97641909e-01 1.73372805e-01
-1.16755259e+00 8.23509872e-01 3.70275319e-01 2.94789433e-01
-7.41204143e-01 -8.93261611e-01 -1.06756020e+00 -3.40710759e-01
9.13246870e-02 -1.04215372e+00 8.46795201e-01 -8.96879017e-01
-1.58701825e+00 1.13382757e+00 7.74716288e-02 2.45864168e-01
7.23945498e-01 1.04397029e-01 -1.85392812e-01 -1.44131362e-01
3.77646863e-01 8.15671861e-01 1.02136183e+00 -1.58125377e+00
-6.31676078e-01 -3.02131414e-01 -1.82168931e-01 3.16196978e-01
-2.65711658e-02 -1.81898817e-01 -9.19763684e-01 -8.54755282e-01
4.90537882e-01 -8.67027044e-01 -1.29188716e-01 5.73200405e-01
-4.01544839e-01 1.11626706e-03 7.86913812e-01 -9.48995054e-01
8.76167774e-01 -1.95776188e+00 8.88151884e-01 3.99269372e-01
2.82353640e-01 4.05321866e-02 -2.34255522e-01 1.90404914e-02
-1.66207239e-01 3.96377444e-02 -4.32855368e-01 -2.79499441e-01
-1.59114718e-01 5.71609773e-02 2.77268235e-02 6.04381502e-01
1.86254695e-01 7.96532750e-01 -8.98104370e-01 -7.97709823e-01
2.07499295e-01 7.61768043e-01 -4.96667445e-01 5.66060066e-01
-1.41638130e-01 1.04331315e+00 -5.70055246e-01 3.65458369e-01
8.73666763e-01 -1.21861286e-01 -1.97395042e-01 -9.15269077e-01
-4.01703119e-02 -2.55347073e-01 -1.37665296e+00 2.05907822e+00
-2.66808003e-01 -1.19270673e-02 4.31687862e-01 -5.56194961e-01
1.15493059e+00 2.78291970e-01 6.60501897e-01 -4.36151147e-01
4.95709717e-01 1.95648268e-01 -1.46372706e-01 -1.85381040e-01
7.77631402e-02 -3.57941240e-01 -7.50461891e-02 1.00541033e-01
-7.87069350e-02 -5.27462125e-01 -9.36783403e-02 -2.04539374e-01
2.32870713e-01 6.05528891e-01 -3.21217775e-02 -3.10897827e-01
6.70651138e-01 -1.28053442e-01 9.50496137e-01 -1.67539135e-01
2.90449560e-01 1.11118889e+00 2.67945081e-01 -3.74399543e-01
-1.41170168e+00 -1.22468519e+00 -1.66456148e-01 5.12009799e-01
6.54945731e-01 -1.73486128e-01 -1.02337146e+00 -5.00959456e-01
-9.16683301e-02 1.66097239e-01 -7.89530218e-01 -2.44724944e-01
-8.63661885e-01 -4.82659757e-01 2.19197765e-01 3.43274772e-01
6.46163166e-01 -9.74939883e-01 -2.53134906e-01 2.99717158e-01
-4.52585548e-01 -8.95131528e-01 -9.88054633e-01 -6.50760591e-01
-7.51600325e-01 -1.07470405e+00 -1.15401793e+00 -1.14576304e+00
1.06837034e+00 -2.43810713e-01 8.70228350e-01 3.50720674e-01
3.63893574e-03 -2.94275641e-01 -2.28664741e-01 -1.97308421e-01
-6.36481583e-01 9.78896990e-02 2.12757349e-01 3.57879251e-01
-5.30466437e-01 -7.96013355e-01 -6.92107856e-01 7.36300111e-01
-9.50066805e-01 6.37262702e-01 4.66114610e-01 5.61422408e-01
1.14431810e+00 -1.29008209e-02 2.46845558e-01 -5.96661329e-01
3.08338046e-01 -2.80131400e-01 -4.75639075e-01 9.19502452e-02
-1.22670986e-01 -1.40301352e-02 4.12966400e-01 -6.93702757e-01
-8.21634114e-01 1.34576559e-01 -3.40980738e-01 -7.21701086e-01
-6.55061007e-02 1.82210654e-01 -5.13311982e-01 -8.35871771e-02
4.77503002e-01 4.71499674e-02 2.59736866e-01 -6.70112967e-01
1.02062188e-01 2.90376961e-01 6.78286791e-01 -7.44673908e-01
1.22395790e+00 3.81055921e-01 1.87920555e-01 -4.01871294e-01
-7.12589800e-01 2.90606283e-02 -1.03918719e+00 -3.10129613e-01
1.18928146e+00 -6.76164091e-01 -5.14952958e-01 6.56788528e-01
-1.42347527e+00 -2.62979865e-01 -1.11708932e-01 3.98606539e-01
-5.78414559e-01 2.81306446e-01 -5.41835070e-01 -6.43678457e-02
-5.88150561e-01 -1.13955724e+00 1.39525390e+00 2.84042954e-01
-1.60543725e-01 -7.93655932e-01 7.78680667e-02 6.28082082e-02
2.85840750e-01 9.51566577e-01 8.31839859e-01 1.21735632e-01
-3.69868189e-01 -9.46119651e-02 -2.20439389e-01 2.62690187e-01
3.93279970e-01 1.51734591e-01 -3.92316192e-01 -4.16075051e-01
-1.96417049e-01 5.28598987e-02 1.66470990e-01 1.42398775e-01
8.36586297e-01 -2.63203204e-01 -4.62636679e-01 8.61784756e-01
1.16200149e+00 -7.81996697e-02 6.21990681e-01 1.11694776e-01
1.15707099e+00 6.40093446e-01 7.80997336e-01 1.51169747e-02
3.54233176e-01 9.08545613e-01 4.49181378e-01 -4.62457389e-01
-3.18429083e-01 -6.46538854e-01 3.27871256e-02 1.13901091e+00
-3.74930739e-01 1.09074987e-01 -8.95628631e-01 3.56710792e-01
-1.82589912e+00 -6.71264172e-01 -1.82702333e-01 2.24344397e+00
1.06017756e+00 4.69644368e-02 -2.84596253e-02 -2.08657086e-01
1.02386665e+00 -3.72437164e-02 -6.06187224e-01 2.87606120e-01
2.76117384e-01 2.24592432e-01 2.58347809e-01 4.22435939e-01
-9.87063050e-01 9.03923571e-01 5.31675673e+00 8.43879461e-01
-1.33812356e+00 9.74674150e-02 4.14053112e-01 1.97404251e-01
-1.45199642e-01 -2.65201151e-01 -5.33872247e-01 5.53405106e-01
2.77819425e-01 -3.75178844e-01 1.81577221e-01 5.85453868e-01
1.81941912e-01 3.84973645e-01 -1.15228117e+00 1.03067029e+00
-3.67191397e-02 -1.22496915e+00 4.77466911e-01 -7.74935260e-02
8.42017055e-01 -4.90978479e-01 -2.80012012e-01 -7.88146257e-02
3.75980549e-02 -1.05038321e+00 1.11945474e+00 9.55622196e-01
1.20116770e+00 -9.03867960e-01 7.73828328e-01 4.22116250e-01
-1.62714398e+00 7.82029569e-01 -5.34327209e-01 3.10886592e-01
4.35034275e-01 2.80485719e-01 -2.95314968e-01 9.22542751e-01
8.52823198e-01 8.75521183e-01 -3.64388674e-01 1.00753164e+00
-1.84213519e-01 -1.35089019e-02 -2.28241578e-01 3.96813691e-01
-2.24307880e-01 -4.68012094e-01 5.79921842e-01 6.95154667e-01
5.14170706e-01 8.70465115e-03 3.83405924e-01 1.16797626e+00
-2.68704444e-01 5.30776024e-01 -3.50561231e-01 3.48668814e-01
4.14896905e-01 1.36887348e+00 -7.86707461e-01 -2.55586356e-01
9.95311067e-02 1.16740847e+00 1.87241539e-01 5.17795533e-02
-9.78390694e-01 -3.38980138e-01 8.03564489e-01 4.74515170e-01
-1.15861781e-02 -1.38804227e-01 -2.74829566e-02 -1.13714814e+00
1.52067756e-02 -7.46661842e-01 -4.82602045e-02 -8.10659409e-01
-1.46326005e+00 7.84330368e-01 1.67611182e-01 -1.60214984e+00
1.94955409e-01 -1.10420644e-01 -6.29309118e-01 1.17226028e+00
-1.06665301e+00 -1.47069263e+00 -6.48299932e-01 4.92466509e-01
2.79591441e-01 3.08948725e-01 6.50525570e-01 5.32938182e-01
-3.49987984e-01 3.96573991e-01 -3.80565524e-01 3.54479015e-01
8.54706049e-01 -7.64704287e-01 6.19290411e-01 6.16963744e-01
-4.11809951e-01 4.98930424e-01 5.87443352e-01 -1.09660399e+00
-1.08070195e+00 -1.54608071e+00 4.39838588e-01 -3.29403192e-01
2.12095201e-01 -2.76931107e-01 -1.12543714e+00 5.66606283e-01
-7.33656958e-02 1.19195096e-01 2.85012349e-02 -4.18677211e-01
-4.85821925e-02 -8.67613852e-02 -1.14593565e+00 6.93211436e-01
1.30239975e+00 -1.06548890e-01 -5.09944201e-01 8.24474245e-02
7.37491727e-01 -1.11540449e+00 -1.21249211e+00 7.17221856e-01
5.48545599e-01 -5.77553928e-01 1.13191926e+00 -4.09808099e-01
6.17894948e-01 -8.25347185e-01 2.67577380e-01 -1.46412063e+00
-4.93158877e-01 -2.95348525e-01 4.11062658e-01 1.50130761e+00
1.23391017e-01 -2.72618324e-01 6.10914171e-01 4.53450680e-01
-7.02302828e-02 -7.19986439e-01 -8.55561674e-01 -5.40865064e-01
2.73149371e-01 -4.08105291e-02 9.30554688e-01 1.06892323e+00
-6.15632176e-01 3.54874849e-01 -5.14843345e-01 3.43135089e-01
8.40345144e-01 2.45232418e-01 9.37195361e-01 -1.39246821e+00
2.96132535e-01 -3.35439205e-01 -5.65530181e-01 -9.85148728e-01
1.63659438e-01 -1.06297874e+00 2.63759851e-01 -1.76011169e+00
2.77355593e-02 -7.10645735e-01 1.15193851e-01 4.30855572e-01
-1.40603155e-01 6.11675620e-01 3.44605111e-02 1.77940667e-01
-1.11499637e-01 8.11362743e-01 2.01337457e+00 -6.77138045e-02
-1.65657282e-01 -1.58784434e-01 -3.73020202e-01 8.58835340e-01
6.33065879e-01 -4.67963189e-01 -1.81107730e-01 -4.78505254e-01
-2.19989326e-02 1.04461037e-01 4.97983813e-01 -1.06101990e+00
-9.94729344e-03 -4.16544288e-01 5.84610522e-01 -3.76337320e-01
2.70007432e-01 -1.11086488e+00 8.21946442e-01 3.37222815e-01
-5.34893535e-02 -1.03362687e-01 1.37088686e-01 3.80315036e-01
-2.63016112e-02 1.56151950e-01 1.22829783e+00 -5.20131886e-02
-2.59347558e-01 9.77999926e-01 4.91498470e-01 1.13562994e-01
1.20880437e+00 -1.36945739e-01 2.43681043e-01 -1.42021952e-02
-7.23909736e-01 2.36294970e-01 9.71437037e-01 7.03900695e-01
7.27605283e-01 -1.94843423e+00 -9.70287800e-01 3.64583939e-01
1.27128931e-02 7.35217333e-01 3.11054111e-01 7.97908843e-01
-7.23880708e-01 -3.42630416e-01 -7.48958349e-01 -6.82089806e-01
-1.40184283e+00 3.29530776e-01 6.07707500e-01 6.70474395e-02
-7.31658101e-01 7.34641671e-01 3.51152927e-01 -6.18599892e-01
4.40321825e-02 -2.50263333e-01 -1.82053983e-01 -2.73729246e-02
3.02077085e-01 2.71484733e-01 -2.13033617e-01 -1.31628311e+00
-3.75427663e-01 1.33665395e+00 3.00663531e-01 1.63444579e-01
1.39992726e+00 -5.51567173e-05 -3.78232300e-01 2.41849169e-01
1.30885851e+00 6.49587736e-02 -1.37428546e+00 -2.50510573e-01
-6.74750924e-01 -6.54894650e-01 -1.68099165e-01 -1.58371300e-01
-1.48632812e+00 7.29791641e-01 4.86209273e-01 -3.04994076e-01
8.49369884e-01 3.00811953e-03 1.15722585e+00 -3.70865434e-01
7.16547847e-01 -7.18976438e-01 -9.27975401e-02 2.87610859e-01
1.44165874e+00 -9.49340582e-01 1.01898558e-01 -7.81170487e-01
-3.13672125e-01 9.77669120e-01 8.84192288e-01 -2.39442527e-01
6.14248872e-01 8.21482465e-02 2.29053516e-02 -2.68733323e-01
-6.70149773e-02 1.01500481e-01 8.08687687e-01 7.14122891e-01
1.85412481e-01 -9.81651619e-02 -1.22508705e-01 4.32575107e-01
-3.37613761e-01 -1.04997344e-01 -9.54617560e-03 6.69292390e-01
-1.52263075e-01 -1.26198518e+00 -5.75331807e-01 3.24343979e-01
-1.36616230e-01 2.42776811e-01 -1.88984439e-01 6.38005853e-01
5.11804640e-01 4.62407082e-01 2.10170850e-01 -4.89590913e-01
8.88020933e-01 -2.23698676e-01 5.09302855e-01 -8.32883060e-01
-5.01624942e-01 1.81399837e-01 -3.92889261e-01 -5.91484547e-01
-5.76557636e-01 -3.55490834e-01 -1.44976401e+00 -5.76300740e-01
-1.12472869e-01 1.10085897e-01 3.56918603e-01 5.51665306e-01
3.13677937e-01 5.54461777e-01 4.44941849e-01 -1.46498823e+00
-2.86754426e-02 -7.87774026e-01 -4.46764886e-01 8.81963372e-01
1.32584482e-01 -9.09199655e-01 -8.91914815e-02 4.26771611e-01] | [7.264889240264893, -1.4791414737701416] |
f66f9b61-8330-44d9-9379-9653094bec25 | impact-analysis-of-the-use-of-speech-and | null | null | https://aclanthology.org/2022.lrec-1.316 | https://aclanthology.org/2022.lrec-1.316.pdf | Impact Analysis of the Use of Speech and Language Models Pretrained by Self-Supersivion for Spoken Language Understanding | Pretrained models through self-supervised learning have been recently introduced for both acoustic and language modeling. Applied to spoken language understanding tasks, these models have shown their great potential by improving the state-of-the-art performances on challenging benchmark datasets. In this paper, we present an error analysis reached by the use of such models on the French MEDIA benchmark dataset, known as being one of the most challenging benchmarks for the slot filling task among all the benchmarks accessible to the entire research community. One year ago, the state-of-art system reached a Concept Error Rate (CER) of 13.6% through the use of a end-to-end neural architecture. Some months later, a cascade approach based on the sequential use of a fine-tuned wav2vec2.0 model and a fine-tuned BERT model reaches a CER of 11.2%. This significant improvement raises questions about the type of errors that remain difficult to treat, but also about those that have been corrected using these models pre-trained through self-supervision learning on a large amount of data. This study brings some answers in order to better understand the limits of such models and open new perspectives to continue improving the performance. | ['Yannick Estève', 'Nathalie Camelin', 'Bassam Jabaian', 'Sahar Ghannay', 'Gaëlle Laperriere', 'Antoine Caubrière', 'Valentin Pelloin', 'Salima Mdhaffar'] | null | null | null | null | lrec-2022-6 | ['spoken-language-understanding', 'slot-filling', 'spoken-language-understanding'] | ['natural-language-processing', 'natural-language-processing', 'speech'] | [ 7.85982087e-02 4.47838634e-01 2.31209770e-02 -6.37837887e-01
-9.94623721e-01 -1.81534782e-01 6.90017700e-01 4.34398204e-01
-9.86833990e-01 6.89067781e-01 2.11438090e-01 -2.33561292e-01
1.61611751e-01 -4.95710850e-01 -8.75452936e-01 -4.13023740e-01
1.22284450e-01 9.49778676e-01 4.43153948e-01 -4.51174825e-01
-1.07992768e-01 8.57506692e-03 -1.96561217e+00 5.21799326e-01
7.40784466e-01 1.00517702e+00 1.76869169e-01 4.64273334e-01
-5.19472957e-01 4.74076301e-01 -5.00260413e-01 -3.99591535e-01
-2.00882614e-01 -1.74883664e-01 -9.36777353e-01 -1.89480081e-01
3.46260756e-01 -4.79680486e-02 1.40288532e-01 7.43581235e-01
4.98092115e-01 1.95685089e-01 2.62144446e-01 -8.31582606e-01
-3.99056613e-01 1.10148597e+00 -4.29701805e-02 1.94570601e-01
2.26859301e-01 -2.10248038e-01 1.17240727e+00 -9.17361796e-01
6.46970272e-01 1.12912691e+00 5.87969601e-01 7.52870500e-01
-1.26856577e+00 -5.60621202e-01 7.83032253e-02 5.66501200e-01
-1.27323818e+00 -6.66916132e-01 4.68042850e-01 -2.58074164e-01
1.33818626e+00 1.89084068e-01 2.95124918e-01 1.18005776e+00
-3.55101615e-01 8.43849659e-01 1.21830022e+00 -9.15883362e-01
2.58328140e-01 7.95072317e-01 2.93852061e-01 4.32533294e-01
-2.16812283e-01 8.74589533e-02 -5.66919029e-01 3.90158780e-02
1.26570985e-01 -5.07318974e-01 -2.08848998e-01 -2.27026448e-01
-8.83815825e-01 1.13577318e+00 3.48180205e-01 9.16030884e-01
-1.00229204e-01 -2.45732501e-01 5.18364966e-01 2.17269823e-01
8.95384252e-01 4.33534831e-01 -9.44920123e-01 -2.70871580e-01
-1.17035270e+00 8.47007707e-02 8.96619976e-01 3.40939015e-01
6.36547506e-01 4.18956578e-02 -8.42238516e-02 1.30038404e+00
2.87021399e-01 2.67460912e-01 7.47237444e-01 -3.17447901e-01
3.70071977e-01 5.33867717e-01 -1.38938800e-01 -3.23164403e-01
-5.91177821e-01 -6.89798594e-01 -5.67522407e-01 4.81687114e-02
6.15116298e-01 -1.43716484e-01 -9.98333395e-01 1.65587187e+00
3.39373529e-01 3.62401515e-01 9.43422988e-02 6.61979675e-01
9.35902774e-01 8.66963744e-01 1.37487710e-01 9.05974011e-04
1.15925944e+00 -1.16218114e+00 -8.39324474e-01 -3.40517491e-01
7.99042225e-01 -8.00911725e-01 1.15402377e+00 5.10832191e-01
-9.42576349e-01 -7.57024407e-01 -9.65699792e-01 5.95556088e-02
-7.09321916e-01 1.51707783e-01 2.81817615e-01 8.05685580e-01
-1.10599625e+00 5.98016202e-01 -7.28969693e-01 -6.33264899e-01
1.81204066e-01 2.27473587e-01 -2.95646906e-01 -3.79351812e-04
-1.43486631e+00 1.18111479e+00 3.44477922e-01 -1.88059099e-02
-7.50215352e-01 -8.18147361e-01 -6.81730568e-01 1.31331950e-01
5.29455721e-01 -2.13418990e-01 1.29692876e+00 -7.61961401e-01
-1.71832180e+00 1.06434178e+00 -1.89541176e-01 -9.50048506e-01
5.45067728e-01 -5.55320501e-01 -6.11851037e-01 -3.22089046e-01
-1.53385535e-01 5.51173210e-01 5.11138022e-01 -1.15846431e+00
-6.45183980e-01 -2.24803150e-01 -2.00474679e-01 -1.20367937e-01
-4.13309127e-01 7.31531978e-02 -3.55300874e-01 -3.81645352e-01
-1.52679414e-01 -8.33626926e-01 -2.82306969e-01 -4.53986615e-01
-1.97929606e-01 -5.61815381e-01 4.49899197e-01 -7.74812162e-01
1.28171492e+00 -2.15395069e+00 1.83485463e-01 -1.22948429e-02
-1.42944232e-01 7.77843893e-01 -1.83373317e-01 5.86541951e-01
-6.51587322e-02 1.12529472e-01 -3.74360442e-01 -6.86083615e-01
1.67409237e-02 1.42469153e-01 -3.25077295e-01 2.47065686e-02
3.65771770e-01 7.06852078e-01 -6.85864687e-01 7.93294311e-02
3.23292911e-01 7.48309195e-01 -5.27891815e-01 4.47670788e-01
-4.26611155e-01 4.37169611e-01 -3.20940055e-02 2.64529907e-03
5.44174552e-01 3.61556038e-02 -2.36384990e-03 1.90909877e-02
-2.30436355e-01 6.38514400e-01 -1.29040647e+00 1.67362440e+00
-6.54511392e-01 5.30657768e-01 -2.90595274e-02 -1.14771938e+00
1.02935362e+00 2.99415648e-01 2.36194462e-01 -7.70904779e-01
2.33779341e-01 4.45451021e-01 -1.17583163e-02 -2.95950383e-01
4.05526549e-01 -1.97322756e-01 9.94043201e-02 2.64949858e-01
5.45641065e-01 -1.44963950e-01 1.57371372e-01 -8.90544802e-03
7.64227271e-01 -1.89025596e-01 2.13001981e-01 -1.78923622e-01
8.80631745e-01 -1.36778161e-01 2.61175036e-01 7.47571230e-01
8.25785007e-03 6.63931012e-01 1.50808334e-01 -4.65031117e-01
-1.04974413e+00 -7.15634882e-01 -3.58952433e-01 1.25690746e+00
-4.60727811e-01 -4.58288342e-01 -9.28445637e-01 -5.13799727e-01
-1.84259191e-02 8.95515203e-01 -7.15186059e-01 -7.32640103e-02
-3.67257476e-01 -8.22558284e-01 5.67298889e-01 4.77188230e-01
3.20136219e-01 -1.13602924e+00 -2.89037734e-01 5.18151939e-01
-1.69528052e-01 -1.47578084e+00 1.99873820e-01 3.78158867e-01
-7.13734388e-01 -7.65702128e-01 -7.53350616e-01 -6.13514900e-01
1.16000965e-01 -2.55114168e-01 1.20614278e+00 1.51906237e-01
-1.39359891e-01 2.86311448e-01 -6.31474495e-01 -6.69765055e-01
-6.41557276e-01 4.23132628e-01 7.18201324e-02 2.65700787e-01
4.66846466e-01 -1.68474659e-01 4.53545004e-02 1.79913759e-01
-7.51634181e-01 -2.28470504e-01 1.97815970e-01 9.90357697e-01
5.31185985e-01 -2.59319365e-01 5.86990952e-01 -1.11919618e+00
4.95984912e-01 -2.93892294e-01 -5.68288982e-01 2.18027920e-01
-7.98977256e-01 2.46249333e-01 3.85549694e-01 -2.37736240e-01
-1.10337234e+00 -1.07748702e-01 -8.71955693e-01 -7.30156004e-02
-4.62436855e-01 5.19691288e-01 5.36614060e-02 2.02494562e-01
5.89139819e-01 2.10483178e-01 -1.20585553e-01 -7.78990805e-01
5.40590703e-01 8.63273740e-01 2.63433456e-01 -2.49694914e-01
4.23367977e-01 2.50624537e-01 -6.11570716e-01 -1.11179316e+00
-1.18903410e+00 -6.39762521e-01 -6.21819854e-01 3.08265984e-02
8.68931234e-01 -9.32151377e-01 -2.12813422e-01 5.25048077e-01
-1.08383083e+00 -5.30898273e-01 -4.01429385e-01 4.90588397e-01
-2.64292121e-01 1.56055331e-01 -4.83749539e-01 -9.92583156e-01
-3.54405999e-01 -9.04653192e-01 8.33278120e-01 1.46767050e-01
-1.44696191e-01 -1.10818601e+00 2.59716064e-01 5.97741067e-01
9.22677636e-01 -3.53448778e-01 6.73000753e-01 -1.07828462e+00
-1.27062663e-01 -1.96586818e-01 1.05689187e-02 7.07652211e-01
-2.28348598e-01 -1.34583235e-01 -1.59985578e+00 -2.80457884e-01
-8.54969025e-02 -4.51248467e-01 1.29089999e+00 2.45363429e-01
1.04210222e+00 1.48692191e-01 -2.27598220e-01 2.96181351e-01
1.13052917e+00 -1.02414107e-02 5.34950614e-01 4.44692254e-01
3.80418569e-01 6.69770539e-01 3.48916471e-01 1.90191641e-01
5.04245400e-01 8.83095562e-01 2.90248930e-01 -9.31884050e-02
-2.67948270e-01 -2.03660905e-01 2.06654727e-01 1.11745429e+00
1.26187131e-01 -1.58288613e-01 -1.04466546e+00 6.54628217e-01
-1.66286051e+00 -5.89321315e-01 -1.98933214e-01 2.36047459e+00
8.14039052e-01 4.54284012e-01 9.26800221e-02 4.77352768e-01
4.21658546e-01 2.16056541e-01 -1.58185765e-01 -7.26672173e-01
-2.51744151e-01 5.44256091e-01 1.72883138e-01 7.65223384e-01
-1.24687874e+00 1.27551150e+00 6.10038567e+00 9.29893017e-01
-1.26553798e+00 4.35544938e-01 6.09354854e-01 2.76465297e-01
-1.64216116e-01 -5.30857816e-02 -1.25242841e+00 1.95149601e-01
1.60621929e+00 1.04936667e-01 2.33445078e-01 7.75414288e-01
3.31026688e-02 -2.74684519e-01 -8.46117973e-01 7.28497684e-01
2.83006877e-01 -1.23233461e+00 -1.05535895e-01 -2.80906618e-01
5.67339897e-01 4.94843900e-01 -6.65411875e-02 7.38439023e-01
9.08496901e-02 -9.39891577e-01 5.06949306e-01 4.04520124e-01
5.72306216e-01 -5.84838927e-01 1.10318804e+00 4.21743631e-01
-7.97609270e-01 3.49847265e-02 -4.51839954e-01 -2.09677964e-01
3.52593839e-01 8.89938712e-01 -1.10864341e+00 3.47739488e-01
7.96189904e-01 3.87146413e-01 -5.84068894e-01 1.05302882e+00
-2.73850292e-01 1.07801914e+00 -5.10717869e-01 -2.57817179e-01
4.31778878e-01 1.58688590e-01 3.65412265e-01 1.46549189e+00
9.27137360e-02 -2.66661555e-01 -1.27637848e-01 7.40684271e-01
4.42353860e-02 5.14289439e-01 -2.56961733e-01 -1.38872847e-01
2.59623975e-01 1.17423749e+00 -4.27526027e-01 -3.99833977e-01
-3.97958964e-01 7.29399383e-01 7.32913494e-01 1.17372237e-01
-6.12912774e-01 -1.92967594e-01 4.68388021e-01 1.89999640e-01
4.57237065e-01 -3.80008966e-02 -1.92233011e-01 -1.12054873e+00
-1.93471164e-01 -6.01672411e-01 2.98523962e-01 -5.01756132e-01
-1.18075824e+00 1.00382531e+00 -1.67940229e-01 -7.13588655e-01
-4.41189706e-01 -7.30759144e-01 -3.95936936e-01 7.67243803e-01
-1.83161390e+00 -8.54894042e-01 -7.71605596e-02 2.91333556e-01
7.43611217e-01 -3.82062674e-01 1.35395908e+00 4.92669463e-01
-5.15452683e-01 7.11601019e-01 2.65092194e-01 9.95091423e-02
7.53449082e-01 -1.14356422e+00 4.14339632e-01 5.02404749e-01
7.16738224e-01 1.66059107e-01 8.20078492e-01 -2.21670717e-01
-7.38875628e-01 -7.38498151e-01 1.52878571e+00 -5.26106894e-01
7.39286304e-01 -5.95400810e-01 -1.35604703e+00 5.00879467e-01
3.81218195e-01 9.13371239e-03 6.25152707e-01 4.61829364e-01
-5.03497839e-01 -2.18528464e-01 -8.33839655e-01 8.03409293e-02
6.31515622e-01 -5.10381520e-01 -8.02347779e-01 2.84257978e-01
7.72000313e-01 -4.70908880e-01 -6.32888198e-01 4.41507250e-01
4.01776046e-01 -1.22850370e+00 7.12028623e-01 -6.62119925e-01
2.74824053e-01 1.92583904e-01 -2.23469809e-01 -1.46689630e+00
3.93098742e-02 -2.20527351e-01 1.09964617e-01 1.49242961e+00
8.20288837e-01 -5.72560966e-01 5.87767363e-01 3.12134176e-01
-8.44485462e-02 -8.43910575e-01 -1.18992949e+00 -6.94950402e-01
3.69870454e-01 -7.98855126e-01 3.22773606e-01 7.94531405e-01
-1.09330133e-01 4.93468434e-01 -3.53352159e-01 -1.58914715e-01
3.25103998e-01 -2.25832120e-01 7.63623655e-01 -1.44834828e+00
-1.67182684e-01 -4.57369119e-01 -2.57617652e-01 -8.24073434e-01
3.60499144e-01 -9.31136847e-01 5.04946560e-02 -1.31597078e+00
-7.25846812e-02 -4.55444992e-01 -4.99134004e-01 4.63347703e-01
-1.13492846e-01 2.11768642e-01 2.98344791e-01 -2.89473176e-01
-5.30048788e-01 6.29100621e-01 6.47234678e-01 -9.89161804e-02
-2.75923908e-01 1.35706395e-01 -3.75141948e-01 6.58225775e-01
7.40372539e-01 -5.25582612e-01 2.99954712e-02 -5.27269006e-01
1.50428712e-01 -3.46649110e-01 5.19634858e-02 -1.14092672e+00
1.28816813e-01 4.13784206e-01 -3.43324314e-03 -4.22547221e-01
5.38808644e-01 -7.54058480e-01 -2.99012363e-01 4.03416902e-01
-4.90613103e-01 -3.04338306e-01 6.86603248e-01 3.99762005e-01
-5.46541035e-01 -3.19217056e-01 7.96865761e-01 -6.51687458e-02
-8.49005938e-01 -2.29321808e-01 -3.87936354e-01 1.13170490e-01
7.53322065e-01 2.85507798e-01 -5.73499361e-03 -4.97181058e-01
-1.09984505e+00 9.50152576e-02 -2.76330382e-01 8.04214954e-01
3.41107637e-01 -8.65828097e-01 -9.26959276e-01 4.89572942e-01
2.89803475e-01 -3.14181715e-01 3.77281755e-01 7.40527511e-01
-9.23375189e-02 6.25072420e-01 -3.53456428e-03 -7.92117238e-01
-1.27090883e+00 2.34204054e-01 4.37480748e-01 -5.60625792e-01
-3.91806841e-01 1.22844112e+00 -3.22669357e-01 -8.50346327e-01
5.02536058e-01 -3.72704387e-01 -5.06239355e-01 2.87791014e-01
6.21181548e-01 2.19435900e-01 4.91086870e-01 -6.83324158e-01
-4.62061286e-01 5.47732234e-01 -1.18054412e-01 -2.41710514e-01
1.52744198e+00 -1.29102305e-01 1.92933127e-01 8.83094668e-01
9.97463644e-01 -1.41538130e-02 -8.63544643e-01 -5.99347234e-01
3.05136085e-01 6.64094463e-02 3.11845422e-01 -1.16634285e+00
-1.02104366e+00 1.43897319e+00 8.35567474e-01 2.62819201e-01
7.48856544e-01 1.55770272e-01 6.88511610e-01 2.57307142e-01
2.24856451e-01 -1.20158911e+00 -2.22341284e-01 9.58278954e-01
8.97617221e-01 -1.56206858e+00 -4.01041299e-01 -2.88696676e-01
-6.67236328e-01 1.01420164e+00 3.24682802e-01 -1.08542226e-01
9.32561934e-01 2.21861497e-01 3.73654038e-01 2.22327501e-01
-8.50950837e-01 -5.90675771e-01 5.16915202e-01 4.09765005e-01
6.88384056e-01 1.35585546e-01 -5.38977027e-01 8.68425906e-01
-3.26908350e-01 5.56883514e-02 4.14096951e-01 4.61028963e-01
-5.39554656e-01 -1.41063857e+00 -1.16952300e-01 2.84099102e-01
-4.11380738e-01 -1.29890695e-01 -3.00546736e-01 6.61344409e-01
6.71321899e-02 1.22283614e+00 3.23148742e-02 -3.86179417e-01
6.32612646e-01 6.64329767e-01 2.63178885e-01 -7.93993354e-01
-7.27214396e-01 8.40155780e-02 2.18042716e-01 -4.76404458e-01
-5.66672325e-01 -5.65668344e-01 -1.23215890e+00 1.12590723e-01
-4.57443625e-01 3.42218280e-01 9.36823606e-01 1.24085712e+00
1.62317157e-01 6.76391244e-01 3.51071835e-01 -7.22973824e-01
-6.40137196e-01 -1.36093652e+00 -3.89251709e-01 3.51825774e-01
2.70547688e-01 -6.75579607e-01 -4.88592863e-01 -2.21977428e-01] | [13.839347839355469, 6.807491779327393] |
d4ddcfbb-e734-41d9-a56d-c040ff47b26b | explainable-end-to-end-deep-learning-for | null | null | https://doi.org/10.1117/1.JMI.7.4.044503 | https://www.spiedigitallibrary.org/journalArticle/Download?fullDOI=10.1117%2F1.JMI.7.4.044503 | Explainable end-to-end deep learning for diabetic retinopathy detection across multiple datasets | Purpose: Diabetic retinopathy (DR) is characterized by retinal lesions affecting people having diabetes for several years. It is one of the leading causes of visual impairment worldwide. To diagnose this disease, ophthalmologists need to manually analyze retinal fundus images. Computer-aided diagnosis systems can help alleviate this burden by automatically detecting DR on retinal images, thus saving physicians’ precious time and reducing costs. The objective of this study is to develop a deep learning algorithm capable of detecting DR on retinal fundus images. Nine public datasets and more than 90,000 images are used to assess the efficiency of the proposed technique. In addition, an explainability algorithm is developed to visually show the DR signs detected by the deep model.
Approach: The proposed deep learning algorithm fine-tunes a pretrained deep convolutional neural network for DR detection. The model is trained on a subset of EyePACS dataset using a cosine annealing strategy for decaying the learning rate with warm up, thus improving the training accuracy. Tests are conducted on the nine datasets. An explainability algorithm based on gradient-weighted class activation mapping is developed to visually show the signs selected by the model to classify the retina images as DR.
Result: The proposed network leads to higher classification rates with an area under curve (AUC) of 0.986, sensitivity = 0.958, and specificity = 0.971 for EyePACS. For MESSIDOR, MESSIDOR-2, DIARETDB0, DIARETDB1, STARE, IDRID, E-ophtha, and UoA-DR, the AUC is 0.963, 0.979, 0.986, 0.988, 0.964, 0.957, 0.984, and 0.990, respectively.
Conclusions: The obtained results achieve state-of-the-art performance and outperform past published works relying on training using only publicly available datasets. The proposed approach can robustly classify fundus images and detect DR. An explainability model was developed and showed that our model was able to efficiently identify different signs of DR and detect this health issue. | ['Moulay A. Akhloufi', 'Mohamed Chetoui'] | 2020-08-20 | null | null | null | null | ['diabetic-retinopathy-detection', 'diabetic-retinopathy-grading'] | ['medical', 'medical'] | [-2.33641088e-01 8.76266044e-03 1.25467598e-01 -4.49925423e-01
-2.16390193e-01 -2.73045093e-01 1.60244673e-01 -1.15384020e-01
-3.90604019e-01 8.07208300e-01 -5.94980456e-02 -4.07664299e-01
-3.80189717e-01 -7.15199769e-01 -2.55287588e-01 -6.93144143e-01
-3.05932853e-02 2.29002714e-01 -4.25201990e-02 1.40717342e-01
4.48697329e-01 8.43641937e-01 -2.04117227e+00 4.66348559e-01
1.45572770e+00 9.67700362e-01 6.77608699e-02 9.31192756e-01
2.63048679e-01 9.68760252e-01 -5.00296295e-01 -1.56261876e-01
3.54984671e-01 -6.42365456e-01 -6.09053016e-01 2.67043591e-01
6.86725199e-01 -5.25086939e-01 -7.03911558e-02 1.00976539e+00
8.21526527e-01 -2.53872633e-01 7.40006387e-01 -4.85500306e-01
-6.32209599e-01 -4.12806347e-02 -8.23710203e-01 3.67272049e-01
-1.88008994e-01 3.77724379e-01 3.37495148e-01 -4.22560185e-01
5.74337542e-01 9.19919908e-01 4.38905865e-01 5.97202063e-01
-1.01223600e+00 -2.94638157e-01 -5.59445620e-01 4.80650038e-01
-1.20233548e+00 -3.48617136e-01 8.41309354e-02 -7.91490257e-01
9.67406631e-01 3.36503237e-01 9.56418633e-01 2.33057305e-01
3.64433080e-01 2.48441860e-01 1.68726075e+00 -4.91402686e-01
2.03740492e-01 2.81510174e-01 4.59835887e-01 8.67321253e-01
5.81628859e-01 2.68330604e-01 2.11669251e-01 8.57101986e-04
7.22323477e-01 -2.35656381e-01 -2.83341169e-01 9.46364477e-02
-5.93195438e-01 8.07796359e-01 4.82313961e-01 1.56384721e-01
-6.26736343e-01 -3.25680077e-01 2.08630502e-01 3.26075196e-01
2.82017052e-01 5.21113992e-01 -1.65494308e-01 1.22693665e-01
-6.26162350e-01 -1.58678710e-01 2.50337422e-01 3.26235980e-01
5.23867607e-01 -1.41981214e-01 -2.66865790e-01 8.86021376e-01
2.60707051e-01 3.47839773e-01 4.76260304e-01 -8.89055312e-01
-2.81159520e-01 1.02258098e+00 2.93822289e-01 -5.47787249e-01
-6.89959466e-01 -6.00588620e-01 -9.13020492e-01 8.25461984e-01
3.45790505e-01 -3.38941753e-01 -1.45125186e+00 9.11087573e-01
5.26592545e-02 -3.62321883e-02 3.78098041e-01 1.23284924e+00
8.43229115e-01 3.31226766e-01 5.78317093e-03 -1.65407404e-01
1.51286602e+00 -7.62923658e-01 -5.00634670e-01 7.88422376e-02
5.73402762e-01 -8.49494040e-01 7.95886636e-01 7.75774181e-01
-1.07636046e+00 -6.56576931e-01 -8.95765603e-01 -1.22154079e-01
-3.25643420e-02 1.23290813e+00 6.40847206e-01 6.70143306e-01
-1.49077642e+00 5.74974418e-02 -5.35534918e-01 -8.15332115e-01
6.34666324e-01 5.09479284e-01 -1.48400605e-01 -3.44461590e-01
-5.50651670e-01 1.06692886e+00 2.21871227e-01 3.47275585e-01
-4.80838329e-01 -2.95291930e-01 -2.91038066e-01 -2.78822958e-01
-3.69872063e-01 -9.77714002e-01 7.66574800e-01 -1.20356977e+00
-1.29740524e+00 1.42873275e+00 -2.79146075e-01 -7.84365356e-01
4.59017873e-01 -1.38303533e-01 -6.97091341e-01 4.56900924e-01
-2.07647145e-01 7.54994929e-01 4.62729901e-01 -8.57055545e-01
-1.02917278e+00 -4.91796255e-01 -9.76042822e-02 -4.99313362e-02
9.30458754e-02 3.62862825e-01 -2.03847930e-01 -7.04126656e-02
-9.32165310e-02 -7.50106871e-01 -1.82188451e-02 2.40571499e-02
-3.93040597e-01 -2.96034604e-01 3.75760436e-01 -7.98148274e-01
9.62141693e-01 -1.91992986e+00 -3.76817465e-01 2.82730579e-01
6.02051973e-01 9.92746651e-01 3.51574160e-02 -1.31293714e-01
-2.24485338e-01 2.21588612e-02 -2.18350440e-02 3.74981999e-01
-6.74919009e-01 -9.12161246e-02 3.61333877e-01 6.13258004e-01
4.29409355e-01 5.15226424e-01 -4.33221549e-01 -1.23051614e-01
4.58130181e-01 6.65861785e-01 -2.90149689e-01 2.61858582e-01
1.99605208e-02 3.30235749e-01 -2.21357122e-01 5.66871285e-01
6.96920335e-01 -5.17677069e-01 5.28944843e-02 -4.28064883e-01
-4.05923843e-01 -3.20945621e-01 -9.84663546e-01 7.70763397e-01
-4.74138930e-02 1.03355706e+00 -6.39277339e-01 -8.79496396e-01
1.34528792e+00 1.05388187e-01 1.63327962e-01 -7.43022442e-01
3.43903720e-01 4.39140052e-01 5.70507228e-01 -1.24026334e+00
-1.45719618e-01 1.73435599e-01 1.03430974e+00 5.05148396e-02
-3.98102909e-01 6.71226442e-01 2.77168065e-01 -2.89453328e-01
8.29909980e-01 -1.91153064e-01 5.45345306e-01 -5.40375225e-02
8.77054691e-01 2.71065563e-01 2.13698596e-01 4.55617756e-01
-3.18038553e-01 6.10404491e-01 6.60549164e-01 -9.81583595e-01
-9.73936856e-01 -5.34334958e-01 -6.50760412e-01 -8.57276749e-03
-6.02545477e-02 4.59948063e-01 -7.83052027e-01 -2.21037462e-01
-8.10498223e-02 5.87861180e-01 -7.49812365e-01 -1.85790304e-02
-2.10287273e-01 -1.00469208e+00 2.56687254e-01 1.21302061e-01
9.85200882e-01 -8.05847168e-01 -7.63903022e-01 -1.20528661e-01
1.75316125e-01 -6.64447725e-01 1.77879050e-01 -6.55709803e-01
-9.67351854e-01 -1.53115380e+00 -1.04790366e+00 -9.10576463e-01
8.78696382e-01 2.24974841e-01 8.73780727e-01 2.18724996e-01
-9.75631595e-01 4.65280749e-02 -1.34727955e-01 -2.86461622e-01
-3.93176526e-01 -4.73536968e-01 -1.76190928e-01 5.71158588e-01
9.15850878e-01 -7.51340166e-02 -1.20258772e+00 2.72412479e-01
-4.81368214e-01 -6.54065534e-02 1.10342455e+00 6.53133214e-01
6.79596186e-01 -4.95053045e-02 5.16378880e-01 -7.16053426e-01
4.52568650e-01 -1.99952170e-01 -1.00980568e+00 3.07664752e-01
-9.82159376e-01 -2.34513447e-01 9.45476517e-02 -1.91434309e-01
-9.18241084e-01 1.00833423e-01 3.22114170e-01 -1.50149688e-01
-4.08413291e-01 2.68919289e-01 3.67463619e-01 -3.48962963e-01
1.17414665e+00 4.21117321e-02 4.01170790e-01 -4.89670694e-01
1.51916826e-02 1.22698534e+00 4.22696203e-01 1.92967862e-01
8.40434507e-02 4.63918895e-01 2.33678743e-02 -9.28120077e-01
-6.97964311e-01 -5.21605849e-01 -1.51739523e-01 -4.88486081e-01
1.25732172e+00 -9.66997266e-01 -8.38418365e-01 7.90292382e-01
-1.01356351e+00 2.46087629e-02 1.66769341e-01 9.83950436e-01
-1.68570891e-01 2.50344038e-01 -2.34138414e-01 -9.14829850e-01
-5.56874931e-01 -1.15824556e+00 5.09220958e-01 7.93373168e-01
8.95405859e-02 -7.42672026e-01 4.70195152e-02 7.13825941e-01
4.25557375e-01 4.77020442e-01 1.03155577e+00 -3.99649769e-01
-5.86569905e-01 -1.66471928e-01 -8.38231683e-01 7.99426258e-01
3.12575936e-01 5.10833144e-01 -1.05722880e+00 -1.56648964e-01
-3.37945104e-01 -2.05303952e-01 9.88079131e-01 1.05374682e+00
1.01766574e+00 -2.30755091e-01 -2.18549371e-01 6.55574203e-01
1.77435768e+00 6.35684609e-01 1.25578868e+00 6.66688442e-01
2.67573774e-01 6.07304990e-01 4.20175016e-01 2.04703733e-01
3.07271540e-01 3.13639820e-01 4.68356609e-01 -7.16641068e-01
-5.09462893e-01 5.88707209e-01 -3.61818597e-02 -5.03567643e-02
-7.81663537e-01 -1.51457340e-01 -1.06387067e+00 6.76272929e-01
-1.48844087e+00 -7.59044290e-01 -7.83374250e-01 2.23147631e+00
7.57238865e-01 -1.40114948e-01 3.40158463e-01 -6.76615983e-02
9.56973374e-01 -5.83705187e-01 -6.43015623e-01 -5.91891110e-01
-3.52277696e-01 3.81184459e-01 4.63419378e-01 3.52256566e-01
-9.16177094e-01 7.03978658e-01 4.95268917e+00 -1.27969146e-01
-1.44186056e+00 -3.79511327e-01 9.90943551e-01 -7.04122111e-02
4.40693617e-01 -3.39619488e-01 -6.29730165e-01 4.58363861e-01
8.90561759e-01 1.05362922e-01 2.82216519e-01 5.14186502e-01
7.55523682e-01 -3.65276217e-01 -7.33471870e-01 9.58962262e-01
9.99362841e-02 -1.43800437e+00 1.96072772e-01 1.05909497e-01
9.57880378e-01 1.80434227e-01 1.63089678e-01 -2.67476976e-01
8.78895372e-02 -1.43144059e+00 -3.08260620e-01 1.21711814e+00
1.03266120e+00 -7.74854541e-01 1.28676713e+00 -2.73761392e-01
-3.17471594e-01 -2.09153146e-01 -5.95386505e-01 1.25529438e-01
-4.19071227e-01 7.30461776e-01 -1.04771030e+00 1.57684937e-01
7.71928251e-01 7.04684734e-01 -7.41615832e-01 1.80125129e+00
-2.67827362e-01 6.48724198e-01 1.50976971e-01 -8.25728104e-02
2.34944239e-01 -3.44023019e-01 3.48016262e-01 8.37101102e-01
4.73196507e-01 3.64412755e-01 -6.02012634e-01 8.42249453e-01
2.51161546e-01 2.10941449e-01 -2.85387963e-01 1.11075036e-01
9.91928354e-02 1.07122970e+00 -2.80790329e-01 -2.70309567e-01
-3.98036569e-01 5.16739845e-01 -1.13641270e-01 5.53239346e-01
-5.73910594e-01 -6.00961566e-01 6.08478129e-01 3.23267668e-01
1.88002601e-01 3.97857875e-01 -4.13349718e-01 -6.87198818e-01
-1.74975321e-02 -1.00982475e+00 4.68789756e-01 -1.26385570e+00
-1.09010243e+00 8.15012276e-01 -6.29384160e-01 -1.31805718e+00
9.96124670e-02 -8.28061640e-01 -4.46808636e-01 1.38731682e+00
-1.93598819e+00 -8.20523202e-01 -8.79920959e-01 5.52530587e-01
9.05092880e-02 -8.15168977e-01 5.27085602e-01 1.48200423e-01
-7.32876003e-01 3.90124142e-01 1.45714492e-01 4.07122485e-02
8.23130786e-01 -1.12249970e+00 -3.03089857e-01 8.27125967e-01
-5.05621970e-01 6.48257256e-01 4.07630533e-01 -5.19109905e-01
-7.01856196e-01 -1.29985464e+00 9.39865232e-01 4.53050472e-02
2.53984779e-01 7.35514820e-01 -8.05457056e-01 2.84352452e-01
2.25691065e-01 1.00754552e-01 6.06063783e-01 -2.34162048e-01
-1.72946274e-01 -4.98361677e-01 -1.40687728e+00 3.84684443e-01
5.21511495e-01 -3.49459723e-02 -4.52067614e-01 6.18855774e-01
-2.64087897e-02 -3.72064561e-01 -8.75237763e-01 2.58151501e-01
6.36958838e-01 -1.40535259e+00 6.40074730e-01 -7.28564560e-01
6.49058640e-01 -5.21393418e-01 3.33850026e-01 -8.38335931e-01
-1.05893292e-01 -5.71937621e-01 1.08341619e-01 6.62870884e-01
4.84246254e-01 -1.22686887e+00 4.88076031e-01 3.61601651e-01
-3.10119800e-02 -6.96677566e-01 -5.27637661e-01 -4.48890805e-01
-2.02327073e-01 4.22239959e-01 2.93448061e-01 8.06241572e-01
-7.64503598e-01 -5.04611284e-02 -5.88633977e-02 5.72696269e-01
6.60998106e-01 -2.95564588e-02 5.53594768e-01 -1.48957205e+00
1.52976369e-03 -5.07938504e-01 -9.39586461e-01 -3.69070917e-01
-5.31347036e-01 -6.02574229e-01 -5.92495620e-01 -2.00059652e+00
2.60768294e-01 -3.94297004e-01 -4.20977682e-01 6.35340989e-01
-9.84884128e-02 3.74507546e-01 -5.48044264e-01 4.15174514e-01
1.53639823e-01 -1.30738512e-01 1.33178174e+00 2.89118886e-02
-5.09994686e-01 2.37898111e-01 -8.71329665e-01 6.29931211e-01
1.25858676e+00 -8.40160474e-02 -5.26547372e-01 -5.34024835e-01
1.81924999e-01 -1.99713618e-01 7.13713527e-01 -1.18646157e+00
5.34975715e-02 2.98445113e-02 4.71073896e-01 -4.25046504e-01
-1.69404343e-01 -3.23712975e-01 8.45923275e-02 6.95251882e-01
-2.01127365e-01 -4.50791746e-01 2.55903155e-01 3.32427472e-01
-2.36889482e-01 2.87691709e-02 1.36664760e+00 7.81051889e-02
-7.42244601e-01 1.65982898e-02 -4.19639468e-01 -2.60709971e-01
1.21359599e+00 -6.13588989e-01 -8.38323414e-01 -8.72523859e-02
-7.56317079e-01 2.50823051e-02 3.91009152e-01 8.81915390e-02
6.90502584e-01 -7.65822828e-01 -1.15938973e+00 1.28708601e-01
3.00193340e-01 -2.31894344e-01 3.89729142e-01 1.15029967e+00
-1.16934800e+00 4.26460952e-01 -6.68102920e-01 -8.03578079e-01
-1.70121956e+00 1.22935921e-01 1.03788161e+00 2.93960243e-01
-7.45617509e-01 8.19342554e-01 -1.62444815e-01 2.69696951e-01
1.49420753e-01 -5.68470478e-01 -8.02883744e-01 -5.42011671e-02
7.07893431e-01 5.84282875e-01 1.96171686e-01 -2.18179792e-01
-1.34638891e-01 1.05313778e+00 -2.84779578e-01 4.72357810e-01
1.34401417e+00 6.48672655e-02 -5.02071619e-01 -2.28869066e-01
7.48067439e-01 -1.71114162e-01 -9.07029808e-01 9.92546901e-02
-1.77032351e-01 -3.73787135e-01 4.76199210e-01 -1.53129137e+00
-1.24216747e+00 8.96206677e-01 1.61518860e+00 1.61725387e-01
1.56197441e+00 -2.34724045e-01 4.10154641e-01 3.37810814e-01
-1.99661329e-02 -7.74695396e-01 -3.91847312e-01 -1.07288256e-01
7.96905518e-01 -1.23832464e+00 7.12497532e-02 -1.17488496e-01
-7.55478084e-01 1.51394868e+00 4.63960707e-01 -6.29911646e-02
2.90466011e-01 -3.90866399e-01 5.22706330e-01 -3.84741157e-01
-4.41828787e-01 -5.58438182e-01 5.88781774e-01 6.87907696e-01
4.33322817e-01 3.27602141e-02 -8.29931855e-01 1.41912282e-01
6.66529238e-02 5.12078285e-01 9.04707134e-01 3.10837269e-01
-7.75165200e-01 -7.97321141e-01 -3.97342779e-02 7.67249465e-01
-3.53998780e-01 -1.40979188e-02 -6.51796937e-01 9.33087349e-01
3.28591377e-01 1.06093800e+00 1.59007609e-01 -7.73775652e-02
1.97967768e-01 -2.15410277e-01 3.11591387e-01 -3.06357473e-01
-3.38205338e-01 1.02756443e-02 5.60067356e-01 -4.10318792e-01
-8.00579131e-01 -3.63674343e-01 -1.09974122e+00 -7.75066167e-02
2.79340874e-02 -1.57737523e-01 5.45041323e-01 6.27727807e-01
7.86015928e-01 4.08939898e-01 6.49063945e-01 1.67388931e-01
-2.65540928e-01 -9.15460348e-01 -5.76809764e-01 -6.05956130e-02
6.45384192e-01 -5.24657369e-01 -3.96884650e-01 3.44307423e-01] | [15.857301712036133, -4.003580570220947] |
29082770-ac92-43c7-aa93-b7300fdc8fab | hand-drawn-symbol-recognition-of-surgical | 2006.16546 | null | https://arxiv.org/abs/2006.16546v1 | https://arxiv.org/pdf/2006.16546v1.pdf | Hand-drawn Symbol Recognition of Surgical Flowsheet Graphs with Deep Image Segmentation | Perioperative data are essential to investigating the causes of adverse surgical outcomes. In some low to middle income countries, these data are computationally inaccessible due to a lack of digitization of surgical flowsheets. In this paper, we present a deep image segmentation approach using a U-Net architecture that can detect hand-drawn symbols on a flowsheet graph. The segmentation mask outputs are post-processed with techniques unique to each symbol to convert into numeric values. The U-Net method can detect, at the appropriate time intervals, the symbols for heart rate and blood pressure with over 99 percent accuracy. Over 95 percent of the predictions fall within an absolute error of five when compared to the actual value. The deep learning model outperformed template matching even with a small size of annotated images available for the training set. | ['Marcel Durieux', 'Donald Brown', 'William Adorno III', 'Angela Yi'] | 2020-06-30 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 3.69772822e-01 4.91053730e-01 -4.46465045e-01 -1.57416865e-01
-6.32058084e-01 -5.31160176e-01 7.01780170e-02 7.17645168e-01
-6.22499526e-01 5.98206520e-01 4.05595042e-02 -9.24056351e-01
-1.13441281e-01 -8.80468190e-01 -5.82216322e-01 -1.56280935e-01
-1.16151057e-01 3.82395685e-01 -6.66726902e-02 1.10370718e-01
3.88201267e-01 5.77535272e-01 -8.75162184e-01 2.77147561e-01
6.61827385e-01 1.28049219e+00 -1.29268304e-01 9.30030763e-01
-7.99692273e-02 6.09552741e-01 -7.39579380e-01 -2.04564363e-01
8.28241825e-01 -5.46650529e-01 -3.72478902e-01 -2.26756006e-01
7.94951856e-01 -6.19250774e-01 -5.48283219e-01 8.94200087e-01
3.29253554e-01 -2.10535616e-01 5.32642484e-01 -5.84355295e-01
3.26517150e-02 3.94252986e-01 -1.05548650e-01 4.63591784e-01
1.92679226e-01 1.64515808e-01 3.96526545e-01 -3.88343662e-01
6.52017415e-01 5.45444846e-01 1.04752886e+00 2.67275035e-01
-9.66412723e-01 -6.94527507e-01 -5.46465993e-01 -5.07604957e-01
-1.14860773e+00 -1.23123437e-01 3.79945695e-01 -6.89020395e-01
7.48235524e-01 4.37224597e-01 1.02204466e+00 2.88324773e-01
1.04390061e+00 4.22661155e-02 8.51396024e-01 -5.69000900e-01
-7.45154396e-02 -1.52167007e-01 -1.01865835e-01 1.27789974e+00
5.57031810e-01 4.78343755e-01 -1.99715957e-01 7.19169378e-02
1.65362811e+00 3.41969728e-01 -1.06424272e-01 -2.56836176e-01
-1.29947960e+00 7.10549653e-01 5.49071133e-01 1.80984259e-01
-2.36637965e-01 8.19746181e-02 5.46871722e-01 1.75631270e-01
-1.43147826e-01 7.93827474e-01 -1.60786897e-01 -2.70942658e-01
-1.14366043e+00 -3.73779356e-01 8.00626278e-01 9.46445346e-01
2.11457983e-01 4.56092879e-03 -2.36810297e-01 3.37549537e-01
-1.46863118e-01 -7.23941531e-03 7.03045547e-01 -1.05992353e+00
2.61324137e-01 6.58061147e-01 -1.72305689e-03 -8.93784046e-01
-8.52145314e-01 -5.95751226e-01 -9.70247746e-01 4.74789470e-01
9.63908613e-01 -6.68494284e-01 -1.44035029e+00 9.06808734e-01
-1.24024525e-01 -2.14335434e-02 -1.00090973e-01 6.17613375e-01
6.65388048e-01 3.22863422e-02 1.21199340e-01 -1.60999913e-02
1.32452404e+00 -5.98084033e-01 -5.16341031e-01 -3.52540135e-01
8.33458722e-01 -6.76864922e-01 5.50056338e-01 3.47159147e-01
-9.88741398e-01 -4.68822420e-01 -1.18110549e+00 -1.23328501e-02
-2.47451946e-01 1.85261145e-01 6.88284039e-01 8.73683035e-01
-8.21405947e-01 1.08637273e+00 -9.96997416e-01 -1.48011357e-01
6.81924164e-01 6.91651523e-01 -4.74092394e-01 2.19649464e-01
-9.45314348e-01 1.20812678e+00 4.03804332e-01 2.37592295e-01
-4.70849991e-01 -7.81583130e-01 -1.20593965e+00 1.61590621e-01
-1.41320024e-02 -5.08513749e-01 1.26156878e+00 -8.50493014e-01
-1.04192841e+00 1.20279467e+00 3.14975500e-01 -9.71766531e-01
1.05781484e+00 3.73267792e-02 -1.71592355e-01 2.73749381e-01
-6.49380684e-02 6.29914463e-01 3.04092348e-01 -6.74704969e-01
-6.70065582e-01 -2.68141538e-01 -2.36552447e-01 6.52035698e-02
-3.60889100e-02 -3.46426010e-01 -4.02934253e-01 -6.38885140e-01
2.99943358e-01 -8.57214332e-01 -6.89517140e-01 4.72163618e-01
-2.54757106e-01 6.06193304e-01 2.59587616e-01 -9.34328675e-01
1.16624069e+00 -2.20300412e+00 -7.67987728e-01 5.34496725e-01
4.26234841e-01 2.97036499e-01 4.83407766e-01 -2.84995679e-02
-1.01503797e-01 6.42829314e-02 -2.24679291e-01 2.20582262e-01
-4.61683601e-01 2.62600839e-01 5.24091087e-02 6.19077027e-01
-3.35095018e-01 7.52557516e-01 -9.54092681e-01 -7.61152208e-01
5.38194895e-01 9.44960043e-02 -3.18149865e-01 -1.05675921e-01
3.80498052e-01 2.24823460e-01 2.79264878e-02 5.98832309e-01
2.34929711e-01 1.12096660e-01 4.89281297e-01 6.95011243e-02
-2.05281764e-01 2.34936923e-02 -8.07727218e-01 1.71255362e+00
-3.74055982e-01 8.53767693e-01 -9.26736519e-02 -6.21356666e-01
1.29279137e+00 3.40519339e-01 8.01344275e-01 -8.17545414e-01
5.19735992e-01 5.13622284e-01 6.03740633e-01 -1.97277978e-01
4.10468966e-01 -5.83754361e-01 -8.97777304e-02 3.22744697e-02
-2.51678795e-01 -2.25016594e-01 2.79326797e-01 -7.62799084e-02
9.50010061e-01 -4.63302702e-01 6.22443438e-01 -4.01988328e-01
-3.69493850e-02 4.44135547e-01 6.87649369e-01 1.11081862e+00
-5.29328763e-01 7.53807306e-01 8.50028038e-01 -9.44446564e-01
-1.00650990e+00 -9.75612044e-01 -5.40529668e-01 1.77367046e-01
1.47878006e-01 -1.02647550e-01 -6.71052337e-01 -7.00327158e-01
2.74037838e-01 4.34150517e-01 -8.97224545e-01 -3.90140414e-01
-5.74002802e-01 -2.48002350e-01 4.79876101e-01 1.04760396e+00
1.19181186e-01 -1.07121766e+00 -1.36812913e+00 5.14535069e-01
5.22024095e-01 -9.08728004e-01 -5.30020773e-01 5.99967480e-01
-1.18275046e+00 -1.29906380e+00 -8.39653432e-01 -1.15630305e+00
1.12279975e+00 -4.45638210e-01 8.16052616e-01 2.61083484e-01
-8.52427423e-01 -5.35883307e-02 1.95706829e-01 -8.22285771e-01
-6.01552188e-01 3.50101222e-03 -3.30682129e-01 -6.32646620e-01
3.73642951e-01 2.04983279e-02 -9.79053080e-01 1.36876315e-01
-4.55496132e-01 1.01747468e-01 7.93683529e-01 6.96372926e-01
5.41114867e-01 -4.63610530e-01 7.42823631e-02 -9.95539546e-01
5.70842326e-01 -1.23740599e-01 -6.44116879e-01 -1.90806855e-02
-7.56835938e-01 -2.92142760e-02 5.14475346e-01 -2.07869694e-01
-4.26447630e-01 4.12710667e-01 2.70384073e-01 -5.45666754e-01
-4.97019552e-02 4.92313445e-01 8.17935288e-01 -2.04016760e-01
7.45472252e-01 -2.71489263e-01 6.59342825e-01 -6.33519366e-02
-1.69969365e-01 4.00267959e-01 1.03532302e+00 -1.41080588e-01
2.72785693e-01 2.93131292e-01 3.38865101e-01 -3.93421024e-01
-4.93041217e-01 -3.87995213e-01 -8.64331067e-01 -3.31568331e-01
7.67547488e-01 -4.88525212e-01 -4.21619952e-01 1.21039920e-01
-8.59864295e-01 -4.50944096e-01 -4.95947927e-01 8.50037098e-01
-4.60742295e-01 2.09121287e-01 -8.39206398e-01 -3.29333216e-01
-5.56080282e-01 -1.26481307e+00 5.43787539e-01 3.57131869e-01
-6.09500885e-01 -1.01962566e+00 -1.51499435e-01 -3.07605136e-02
1.98930174e-01 7.55696177e-01 8.73251736e-01 -8.31825018e-01
-2.06824034e-01 -1.09158218e+00 -2.16158316e-01 1.81739181e-01
3.04239154e-01 9.40917730e-02 -3.12446445e-01 5.39954342e-02
-2.59237200e-01 3.21292102e-01 5.99425316e-01 8.21913123e-01
1.37551939e+00 -1.17671201e-02 -5.73847771e-01 7.58567929e-01
1.28025496e+00 8.15305650e-01 7.44761586e-01 5.33303082e-01
2.81277299e-01 4.79611270e-02 4.12487060e-01 4.36495036e-01
-4.93507199e-02 5.44931926e-02 3.10340047e-01 -6.89386189e-01
-2.63289884e-02 -8.03144872e-02 -4.49169278e-01 9.79755670e-02
-5.04450239e-02 4.21716630e-01 -1.34247744e+00 6.42177045e-01
-1.38301647e+00 -5.92285037e-01 -1.25814881e-02 2.46066523e+00
8.68599653e-01 7.36448944e-01 -2.03218818e-01 -8.75545666e-02
7.42736876e-01 -2.30079100e-01 -5.50496876e-01 -8.37181628e-01
5.48089027e-01 5.76186001e-01 1.38077652e+00 3.72541338e-01
-1.01532733e+00 4.67118472e-01 7.17400694e+00 -8.71867761e-02
-1.19045675e+00 -4.43442941e-01 1.00595725e+00 2.85400054e-03
3.55479240e-01 -1.86130151e-01 -4.40379471e-01 3.40091288e-01
9.47215199e-01 -2.30802298e-01 -1.67676434e-01 7.78764784e-01
2.09540769e-01 -4.12278891e-01 -1.11454272e+00 1.02180159e+00
2.57647485e-02 -1.88686359e+00 -4.72793221e-01 1.18532620e-01
5.35591006e-01 -3.30521703e-01 -6.68347776e-02 -1.49872927e-02
1.89891458e-01 -1.24053562e+00 4.44663584e-01 5.49782753e-01
1.49901497e+00 -5.68255067e-01 9.35819149e-01 -5.70438281e-02
-7.65209794e-01 -1.09140292e-01 -1.67828694e-01 -1.59470737e-01
-3.31534967e-02 2.17068717e-01 -1.42398596e+00 1.06473947e-02
3.00307572e-01 2.16829374e-01 -2.64346004e-01 1.53047860e+00
7.58000240e-02 2.69454449e-01 -4.25774813e-01 2.14779899e-01
3.55522990e-01 -9.63400975e-02 2.59922165e-02 1.14940965e+00
3.96063060e-01 1.74545348e-01 6.40355796e-02 4.07937348e-01
-1.86656371e-01 4.69030365e-02 -6.18393838e-01 1.44049093e-01
2.84428090e-01 9.88116086e-01 -1.20694685e+00 -4.64327782e-01
-5.24965525e-01 6.64077580e-01 -2.31476828e-01 -8.22769850e-02
-5.02080917e-01 -8.77127349e-01 3.81226689e-01 4.07263458e-01
-1.14059113e-01 -2.26418376e-01 -1.10778534e+00 -4.81463522e-01
-1.63582146e-01 -4.74707782e-01 7.01042891e-01 -2.65380025e-01
-4.89936680e-01 3.16372067e-01 -3.97573471e-01 -1.36045277e+00
-3.82390529e-01 -1.03457832e+00 -7.31012702e-01 9.18626070e-01
-1.09645343e+00 -3.99295598e-01 -6.02546573e-01 1.40293151e-01
1.92618743e-01 4.67521660e-02 1.01319623e+00 7.18374401e-02
-3.42497081e-01 6.06128573e-01 1.34268194e-01 9.77059126e-01
6.66033685e-01 -1.35020435e+00 3.51836830e-01 7.24275708e-01
-2.29413256e-01 7.87257552e-01 4.56599653e-01 -8.42963278e-01
-9.55767453e-01 -8.67835939e-01 6.50823712e-01 -2.60314822e-01
3.77199441e-01 -6.56622695e-03 -4.49421883e-01 9.95911360e-01
-1.20475858e-01 3.86521369e-01 9.27016735e-01 -2.27028444e-01
2.78223097e-01 -6.59752563e-02 -1.29586923e+00 4.30250227e-01
6.23062313e-01 -2.45212376e-01 -6.29643679e-01 3.16488147e-01
-1.19893653e-02 -1.39344668e+00 -1.20633769e+00 3.96422595e-01
9.43166196e-01 -7.74218857e-01 7.85901904e-01 -6.25962377e-01
6.19149446e-01 2.43646652e-01 4.46030676e-01 -9.73958492e-01
2.38866769e-02 -5.86950898e-01 2.91517168e-01 2.97617823e-01
6.55484080e-01 -5.18201232e-01 1.17096233e+00 1.25807667e+00
-4.83083546e-01 -8.09728444e-01 -9.42925096e-01 -5.05799294e-01
1.65361822e-01 -1.11285195e-01 2.04957053e-01 8.06343973e-01
3.75557423e-01 -5.18980980e-01 8.24037120e-02 -1.04734130e-01
2.95632333e-01 1.63761094e-01 3.71160448e-01 -1.14009094e+00
8.59602466e-02 -7.03916848e-01 -8.69648516e-01 -4.00839686e-01
-3.26264262e-01 -8.24442387e-01 1.55595914e-01 -1.88269782e+00
-2.12583363e-01 -4.70915079e-01 -4.91989613e-01 6.61239505e-01
3.40788998e-02 2.85927683e-01 4.33482043e-02 -4.13360298e-02
2.33887821e-01 -3.92711490e-01 1.24286854e+00 -6.33120388e-02
-5.54869950e-01 8.66161361e-02 -6.34155273e-01 6.98110342e-01
1.01220548e+00 -4.69767630e-01 -4.77163531e-02 -2.86685497e-01
-2.77448654e-01 6.16945207e-01 1.51979759e-01 -1.25021958e+00
2.93770134e-01 -8.18565562e-02 1.03835094e+00 -3.59156191e-01
-1.76625058e-01 -8.64100695e-01 6.01456612e-02 1.17764819e+00
-4.44818199e-01 1.10569477e-01 6.27286315e-01 1.77233383e-01
-1.04936451e-01 -2.75729835e-01 9.78079438e-01 -4.85643744e-01
-5.00975370e-01 3.04785848e-01 -4.30812776e-01 -6.76318556e-02
1.19737935e+00 -7.32853174e-01 -7.11645856e-02 -1.48545176e-01
-8.82437050e-01 2.97083501e-02 3.97834092e-01 1.08579338e-01
6.51329994e-01 -8.29933703e-01 -3.08308929e-01 2.87396908e-01
-1.22142121e-01 2.76584327e-01 1.95824549e-01 1.02125907e+00
-1.43188047e+00 5.40500343e-01 -5.90443909e-01 -1.95149735e-01
-1.09555328e+00 4.80406433e-01 8.34790885e-01 -1.83773890e-01
-1.09955740e+00 7.12345541e-01 -6.49517681e-03 1.32675812e-01
2.88205147e-01 -7.16277838e-01 4.12107892e-02 -6.17662445e-02
4.39401805e-01 1.17802083e-01 1.82536453e-01 -1.57218263e-01
-1.68015987e-01 2.44676903e-01 -5.61416000e-02 1.83354154e-01
8.87746930e-01 4.73986804e-01 2.83337712e-01 8.41218531e-02
1.02214599e+00 -2.57005990e-01 -1.42882931e+00 1.42121285e-01
-9.98325497e-02 -4.28505540e-01 2.56179214e-01 -9.24815953e-01
-1.21625185e+00 7.52025783e-01 7.80031443e-01 -1.23799413e-01
7.85314500e-01 -4.21734303e-01 8.01647186e-01 -7.65973004e-03
-3.06861065e-02 -1.06097782e+00 -4.98881578e-01 1.13445535e-01
4.04498756e-01 -1.05765045e+00 1.66147985e-02 -2.82647103e-01
-4.94405657e-01 1.63419008e+00 6.30703092e-01 -2.04896778e-01
5.33719718e-01 5.96923292e-01 4.87937063e-01 -1.94862902e-01
1.13051280e-01 2.20421895e-01 1.94766879e-01 4.26698685e-01
4.13599283e-01 2.59287655e-01 -4.77410525e-01 3.10628861e-01
-4.16109711e-01 1.95595667e-01 8.05021346e-01 1.24468482e+00
-4.84035879e-01 -6.42200649e-01 -3.66483986e-01 1.25149691e+00
-6.90745831e-01 -2.91339129e-01 -7.91729167e-02 9.54496980e-01
-1.67169824e-01 4.78593439e-01 5.81207275e-01 1.76016986e-02
5.50178945e-01 1.94466278e-01 3.22386295e-01 -6.49413228e-01
-7.38403678e-01 6.75180256e-02 1.24231480e-01 -4.86143112e-01
2.61967510e-01 -7.08574355e-01 -1.66071665e+00 3.87700349e-02
1.66325778e-01 -1.26953036e-01 6.59997106e-01 6.25987113e-01
8.72834548e-02 7.36513078e-01 7.64293075e-02 -5.36438882e-01
-3.35749418e-01 -7.21821249e-01 -6.17204607e-01 1.39804214e-01
5.04666209e-01 -4.21009541e-01 -1.52832836e-01 1.04779363e-01] | [14.406479835510254, -2.63227915763855] |
ac269534-eba3-4fb2-ada5-6621e93175cb | a-multi-stream-deep-neural-network-with-late | null | null | https://doi.org/10.1016/j.eswa.2022.117030 | https://reader.elsevier.com/reader/sd/pii/S0957417422004468?token=0DF82822FFD4CF570219D85B4D22640194174214B62C2EF2F9D312C3B9A94571E27EDA274DE53CE91F489F49976617E2&originRegion=eu-west-1&originCreation=20220420222220 | A multi-stream deep neural network with late fuzzy fusion for real-world anomaly detection | Abnormal event detection in video is alternatively known as outlier detection, where machine learning can be highly effective. While testing an unknown video, the objective of such methods is to verify the video’s category, e.g. normal or abnormal. This paper exploits visual information from normal as well as abnormal videos to train a deep multiple instance learning classifier for classification of videos. Existing multiple instance learning classifiers presume that the training videos contain only short-duration anomalous events. This assumption may not be valid for all real-world anomalies. Also, multiple occurrences of anomalies in training videos cannot be ruled out. This paper shows that by injecting temporal information in feature extraction, anomaly detection performance can be improved. To accomplish this, two spatio-temporal deep feature extractors have been applied in parallel on the training videos. These streams are then used to train a modified multiple instance learning-based classifier. Finally, a fuzzy aggregation is employed to fuse the anomaly scores. Additionally, two lightweight deep-learning classifiers have been used to substantiate the model’s efficacy for classifying fire and accident events. To understand the reliability and performance of the proposed method, extensive experiments have been carried out using UCF-Crime video dataset containing 13 anomaly categories. The dataset has been restructured into five broad categories based on the severity of actions to study the robustness of the proposed method. The paper provides sufficient empirical evidence which proves that by incorporating temporal features in the pipeline, anomaly detection accuracy can be significantly improved. Moreover, the model helps to detect long-duration anomalies in videos, which was not possible using existing methods. The proposed end-to-end multi-stream architecture performs abnormal event detection with accuracy as high as 84.48%, which is better than the performance of existing video anomaly detection methods. Moreover, the class-wise detection accuracy has improved by 6%–14% across various broad categories. | ['Ig-JaeKim', 'Heeseung Choi', 'Debi Prosad Dogra', 'Nitin Sharma', 'Kamalakar Vijay Thakare'] | 2022-03-27 | null | null | null | expert-systems-with-applications-2022-3 | ['anomaly-detection-in-surveillance-videos', 'abnormal-event-detection-in-video', 'anomaly-detection-in-surveillance-videos', 'abnormal-event-detection-in-video'] | ['computer-vision', 'computer-vision', 'methodology', 'methodology'] | [ 2.88349062e-01 -2.58192599e-01 4.45111170e-02 -1.96179345e-01
-5.38678110e-01 -3.20229977e-01 4.79845107e-01 4.37543362e-01
-3.38081032e-01 4.21867341e-01 -1.08537741e-01 -1.39933363e-01
-2.45641038e-01 -6.08848214e-01 -7.26397574e-01 -7.41242111e-01
-4.87525403e-01 -6.95142224e-02 3.14275861e-01 7.14364871e-02
3.42560977e-01 6.55145466e-01 -1.97282338e+00 8.32392216e-01
6.93536639e-01 1.27481532e+00 -3.04182231e-01 6.29296899e-01
-6.59911633e-02 8.19989800e-01 -8.71041656e-01 8.84517282e-02
4.18273509e-01 -3.45978320e-01 -4.17150319e-01 4.30433154e-01
4.01969820e-01 -7.85186112e-01 -1.97727934e-01 7.28352606e-01
2.18986556e-01 1.79931313e-01 5.03701329e-01 -1.51766980e+00
9.79016051e-02 7.16429278e-02 -6.14418924e-01 7.73049414e-01
5.85904479e-01 2.34072149e-01 7.23871469e-01 -7.33980596e-01
2.57951260e-01 8.00553441e-01 3.66744667e-01 9.36237499e-02
-7.02089012e-01 -6.99808836e-01 2.06686899e-01 6.44324064e-01
-1.11766720e+00 -2.18127340e-01 7.53989875e-01 -5.02402723e-01
1.07900500e+00 2.08498999e-01 7.27520347e-01 1.00921535e+00
4.61611927e-01 7.56657422e-01 8.10302734e-01 -2.44999006e-01
1.68397576e-01 -1.49738729e-01 -8.09173137e-02 5.63702285e-01
2.69482762e-01 1.21253081e-01 -4.43394452e-01 -7.45422998e-03
2.74346530e-01 4.96815115e-01 -8.85184556e-02 9.78432074e-02
-1.10478747e+00 6.28375947e-01 2.15927616e-01 4.81006593e-01
-7.85218239e-01 -2.72996962e-01 9.25342858e-01 4.90343899e-01
4.72640991e-01 2.53897130e-01 -2.69922018e-01 -2.75168389e-01
-1.02620602e+00 5.23032621e-02 3.52771103e-01 4.81893480e-01
3.54094893e-01 5.04806519e-01 -2.65724629e-01 5.04387379e-01
8.83732662e-02 1.80253282e-01 5.27090549e-01 -7.27771282e-01
3.44098330e-01 9.40141439e-01 -2.12120800e-03 -1.29392588e+00
-4.28076416e-01 -3.88495952e-01 -7.20828593e-01 4.22839284e-01
3.90249014e-01 -8.99918154e-02 -9.29017246e-01 1.18630636e+00
2.87061572e-01 7.89097369e-01 6.38105422e-02 9.28920031e-01
3.95187169e-01 7.87510693e-01 1.70110270e-01 -4.35733289e-01
1.11798811e+00 -4.63532060e-01 -8.50141406e-01 1.25251606e-01
6.77861750e-01 -6.43639624e-01 7.35205829e-01 7.46526182e-01
-7.58756578e-01 -6.02659881e-01 -1.09343410e+00 7.27332890e-01
-3.20240617e-01 9.99164581e-02 3.66777539e-01 5.44194281e-01
-4.96151954e-01 4.35391814e-01 -9.53702867e-01 -3.33980471e-01
5.77181995e-01 3.30427110e-01 -4.95510280e-01 -1.77061260e-01
-1.03807044e+00 5.75209618e-01 7.28259802e-01 2.12658152e-01
-9.51137483e-01 -4.63619888e-01 -6.62134647e-01 -2.14409605e-02
3.67744416e-01 -8.45498964e-02 7.43729830e-01 -1.36951351e+00
-8.11339915e-01 3.72590333e-01 -1.96920037e-01 -7.20820367e-01
5.32193422e-01 -2.61327118e-01 -9.77419555e-01 5.83375812e-01
5.90481013e-02 4.96515110e-02 1.02596128e+00 -9.22818124e-01
-1.20395768e+00 -3.58174533e-01 2.08441481e-01 5.64049557e-02
-5.20210385e-01 2.16923222e-01 -9.95283574e-02 -7.95194924e-01
1.06253696e-03 -7.34683752e-01 2.38348186e-01 -3.06453556e-01
-8.85193199e-02 -7.45539367e-02 1.36079478e+00 -8.56174350e-01
1.51522243e+00 -2.32827830e+00 -2.89091706e-01 4.57166314e-01
-1.91119052e-02 4.06730175e-01 2.67144945e-02 2.49659389e-01
-2.51975179e-01 -4.94748168e-02 -2.32383907e-01 1.53353170e-01
-4.20068443e-01 2.74156451e-01 -1.93330675e-01 5.54836035e-01
5.81682861e-01 2.76773214e-01 -7.24571764e-01 -2.99801260e-01
4.33644384e-01 1.99472293e-01 -4.43471193e-01 2.24128991e-01
4.18667272e-02 4.80530322e-01 -2.35993236e-01 9.48643625e-01
3.93867701e-01 3.07995558e-01 -2.51721054e-01 -2.01154560e-01
-1.05975367e-01 -3.79477471e-01 -1.25600636e+00 1.18994105e+00
-1.79059803e-01 6.51987195e-01 -3.80026639e-01 -1.36806369e+00
7.26373851e-01 7.00421214e-01 9.07403529e-01 -7.73118496e-01
-7.34334067e-02 1.41294256e-01 2.85872251e-01 -9.88977313e-01
1.62715361e-01 8.79661292e-02 1.20422058e-01 1.96479559e-01
4.60215990e-04 5.99570930e-01 5.34432888e-01 3.85706984e-02
1.46319973e+00 9.38414484e-02 7.11372048e-02 1.94017112e-01
7.34743655e-01 1.27651617e-01 6.25422657e-01 6.60457313e-01
-3.06593984e-01 4.18614298e-01 4.56493646e-01 -6.21043563e-01
-9.38264489e-01 -9.81907904e-01 -4.85347807e-02 7.99847782e-01
-1.86288003e-02 -2.14659438e-01 -5.36049604e-01 -8.69294524e-01
-1.99461982e-01 5.75504065e-01 -4.75677967e-01 -3.24844569e-01
-4.93757457e-01 -9.17346597e-01 6.45088911e-01 5.33459425e-01
6.29676402e-01 -1.06279898e+00 -9.29806232e-01 1.85178146e-01
-1.69081539e-01 -1.17658830e+00 1.30778566e-01 -1.40278026e-01
-7.95580745e-01 -1.45066488e+00 -1.42200708e-01 -2.85247833e-01
5.99278867e-01 1.55912504e-01 4.66072917e-01 3.18849325e-01
-3.50172907e-01 5.07800937e-01 -7.36594200e-01 -4.90053207e-01
-3.87196124e-01 -3.92159492e-01 2.51091927e-01 5.32446325e-01
5.08253336e-01 -3.38763893e-01 -5.63262880e-01 3.42810571e-01
-1.29621446e+00 -6.17000222e-01 6.96775675e-01 6.94982946e-01
4.40090775e-01 6.19414270e-01 8.63583505e-01 -4.80363429e-01
3.71840507e-01 -7.90147722e-01 -3.28986228e-01 -5.91991618e-02
-3.12950820e-01 -3.50325018e-01 8.21200907e-01 -4.43249524e-01
-1.04855454e+00 -1.26207545e-02 -1.07197575e-01 -6.41788363e-01
-6.40782416e-01 6.45637989e-01 -5.60531672e-03 3.02805126e-01
4.23054487e-01 1.41157642e-01 8.30975100e-02 -1.28180664e-02
-4.16717857e-01 6.95497870e-01 4.21588719e-01 -1.08035348e-01
6.27584994e-01 5.78444123e-01 4.83557321e-02 -9.84385967e-01
-5.79922020e-01 -7.66450942e-01 -5.92074037e-01 -7.66244471e-01
9.63955402e-01 -9.40257668e-01 -4.41750735e-01 6.95555270e-01
-8.24443698e-01 9.81563106e-02 -3.89363393e-02 7.07073390e-01
-1.99997962e-01 4.80374545e-01 -4.12801743e-01 -9.49689090e-01
-5.53927459e-02 -1.00015783e+00 7.98207104e-01 5.22170737e-02
-3.16001564e-01 -7.90956557e-01 -3.72747511e-01 4.57033098e-01
1.90749273e-01 7.11566567e-01 7.69576788e-01 -1.21619606e+00
-4.37366128e-01 -7.20338523e-01 6.70544505e-02 4.80402291e-01
2.81848371e-01 3.85670662e-01 -8.65841508e-01 -3.54224950e-01
-6.57321811e-02 1.42282978e-01 7.13398755e-01 2.38130584e-01
1.18450105e+00 -2.83400774e-01 -1.48611769e-01 2.77232587e-01
1.30317652e+00 6.86005712e-01 8.39177668e-01 7.25048423e-01
6.05291843e-01 4.47003633e-01 9.95356500e-01 7.26659238e-01
-1.14804395e-01 5.15444756e-01 7.72244155e-01 9.57396701e-02
2.90533066e-01 3.61834437e-01 8.24231684e-01 3.95622402e-01
-2.53623426e-01 -1.96902946e-01 -8.39528620e-01 5.04497528e-01
-1.79757082e+00 -1.58539760e+00 -2.63895720e-01 2.17885876e+00
6.09976240e-02 4.16858643e-01 3.29765111e-01 8.76936674e-01
7.20698297e-01 -1.00443629e-03 -4.47879374e-01 -6.07905149e-01
-3.28511335e-02 3.17895226e-02 8.45416412e-02 -3.65349539e-02
-1.44059765e+00 3.24304760e-01 4.70800924e+00 5.49766183e-01
-1.30443776e+00 -8.64656717e-02 4.12088633e-01 -4.77413923e-01
3.33732396e-01 -2.24597976e-01 -3.06141526e-01 7.35837281e-01
1.10965145e+00 9.51892510e-02 5.53353913e-02 6.96668506e-01
6.80049181e-01 -2.90003747e-01 -8.07336569e-01 6.67605639e-01
3.04380298e-01 -1.04496205e+00 1.12685792e-01 -9.34087336e-02
4.74522948e-01 -2.15938956e-01 -2.04437092e-01 3.78434837e-01
-5.63786328e-01 -7.68328309e-01 5.65285683e-01 6.13562465e-01
4.00124609e-01 -1.11703289e+00 1.20425832e+00 3.64738703e-01
-1.12237453e+00 -7.25629687e-01 7.10115954e-02 -2.00066283e-01
2.32026324e-01 4.87718701e-01 -8.64443660e-01 8.01671445e-01
9.99365568e-01 7.65084743e-01 -7.13346601e-01 1.24739087e+00
8.54704306e-02 7.98712850e-01 -3.80545408e-01 4.39481258e-01
3.78891319e-01 -2.11862139e-02 7.51039565e-01 1.18219101e+00
6.58914983e-01 3.92824747e-02 3.09147090e-01 1.78852975e-01
4.28320736e-01 1.01989947e-01 -8.76348674e-01 3.91065292e-02
1.66536689e-01 1.13225496e+00 -8.33777606e-01 -2.71647066e-01
-6.80424213e-01 8.84150207e-01 -2.91349113e-01 2.50270009e-01
-1.16261864e+00 -3.70318472e-01 4.77304131e-01 2.86722600e-01
3.22326124e-01 -1.80749409e-02 1.85390357e-02 -1.00701725e+00
2.42264286e-01 -9.47385669e-01 7.44835734e-01 -5.53110540e-01
-9.62234974e-01 4.94582862e-01 1.13421619e-01 -1.74001348e+00
-4.35477644e-01 -5.81721663e-01 -8.54627669e-01 3.31878901e-01
-1.06744492e+00 -1.08813083e+00 -4.33657557e-01 7.05654085e-01
7.77879357e-01 -5.00987768e-01 5.31927586e-01 5.51682293e-01
-9.00510132e-01 4.90388572e-01 -1.06878102e-01 3.63002181e-01
5.42781949e-01 -9.98740554e-01 -3.24041545e-01 1.57392812e+00
2.26029996e-02 1.92152619e-01 6.70891106e-01 -7.33915269e-01
-1.17235506e+00 -1.39579582e+00 2.93353647e-01 -1.95512757e-01
6.90715015e-01 2.19094738e-01 -1.20838583e+00 5.95778048e-01
1.44971102e-01 1.99761480e-01 6.32202804e-01 -4.02979702e-01
-3.25560085e-02 -2.09607407e-01 -1.26690781e+00 2.78230786e-01
6.18980825e-01 -2.53473401e-01 -6.06056333e-01 1.04620188e-01
1.98771402e-01 -2.57462889e-01 -8.38664711e-01 7.37251103e-01
5.67324102e-01 -1.21788788e+00 6.60640657e-01 -7.63356030e-01
5.08135200e-01 -5.59632778e-01 -2.70833224e-01 -1.11633110e+00
2.67592240e-02 -9.23136529e-03 -5.24960697e-01 1.24335313e+00
1.70319527e-01 -6.45635366e-01 4.32828218e-01 3.71259183e-01
-4.67580318e-01 -7.65267968e-01 -9.69340384e-01 -7.58311868e-01
-5.72391391e-01 -7.98883498e-01 3.96016896e-01 9.56161678e-01
-1.45578563e-01 -1.34471059e-01 -3.12291920e-01 6.18108809e-01
3.97787392e-01 -2.21442416e-01 6.01556063e-01 -1.13754380e+00
-6.22976664e-03 -3.41146559e-01 -9.89651501e-01 -1.12656146e-01
1.61657616e-01 -6.24901712e-01 -2.95187235e-01 -1.31771898e+00
-6.34775758e-02 -2.47279368e-03 -6.77032232e-01 5.90419769e-01
-7.51586929e-02 3.60370457e-01 -7.48194754e-02 4.67881635e-02
-4.88237768e-01 1.34020790e-01 6.45117700e-01 -1.65751681e-01
-1.24961630e-01 2.73413539e-01 -9.22201015e-03 8.33249152e-01
9.93055046e-01 -2.31955454e-01 -3.83883774e-01 -7.96165019e-02
-4.38789092e-02 -7.06281736e-02 6.45260215e-01 -1.38044453e+00
6.68223873e-02 -4.55098189e-02 6.27549887e-01 -6.47522151e-01
1.35968000e-01 -1.19721997e+00 -4.71571274e-03 4.17376161e-01
-3.81792523e-02 3.99378091e-01 4.06732947e-01 8.24356079e-01
-6.86888993e-01 8.96328967e-03 6.24018669e-01 1.22119106e-01
-1.24263191e+00 1.95313990e-01 -7.02550888e-01 -3.03490967e-01
1.59129083e+00 -6.90087914e-01 -1.07828394e-01 -1.72621757e-01
-7.91215122e-01 4.82974090e-02 1.58907250e-01 5.24215996e-01
9.07417715e-01 -1.29517782e+00 -6.79706097e-01 3.22410405e-01
3.78798634e-01 -3.20080906e-01 5.83359122e-01 1.18984675e+00
-5.58897734e-01 8.46604928e-02 -4.30054873e-01 -8.95855784e-01
-1.43965673e+00 6.02306962e-01 2.21564680e-01 6.71555027e-02
-6.14653707e-01 2.10037559e-01 -4.68335837e-01 2.73242801e-01
2.14156210e-01 -2.07867295e-01 -4.49691415e-01 3.74166816e-01
6.40631378e-01 5.44855177e-01 3.05760443e-01 -7.49746919e-01
-3.94828022e-01 1.91439733e-01 -1.05554514e-01 2.91975588e-01
1.22366786e+00 6.71548694e-02 8.77996683e-02 4.58310902e-01
9.72390652e-01 -9.33749005e-02 -1.09273505e+00 2.42185071e-01
1.98127180e-01 -5.35023451e-01 -6.92472905e-02 -5.73745489e-01
-1.15908527e+00 8.15607786e-01 9.30306911e-01 3.99868369e-01
1.58066607e+00 -4.33670312e-01 6.98676467e-01 2.03193337e-01
1.37139738e-01 -1.18926692e+00 2.00623482e-01 1.85595393e-01
7.44564950e-01 -1.48069966e+00 -1.16710626e-01 1.45172417e-01
-7.06302643e-01 1.17550492e+00 8.12274516e-01 -1.78477570e-01
4.31088805e-01 1.81565583e-01 -8.04452300e-02 -1.29097864e-01
-6.09923780e-01 -1.32928401e-01 3.70172262e-01 3.02552313e-01
2.05866575e-01 -2.00779036e-01 -1.36216238e-01 3.76818717e-01
3.20925444e-01 -5.47970831e-02 6.26182020e-01 1.03687668e+00
-4.05853510e-01 -5.91411531e-01 -5.02901196e-01 9.27732110e-01
-9.31536615e-01 3.68822247e-01 -6.55066920e-03 9.27339613e-01
4.08817828e-01 1.21261537e+00 3.07901263e-01 -5.21444380e-01
4.27416563e-01 2.87540823e-01 4.63079065e-02 -3.41622382e-01
-5.76227248e-01 -5.30594448e-03 5.60417436e-02 -7.46890247e-01
-6.06203139e-01 -7.52393961e-01 -1.32213163e+00 -1.03271201e-01
-1.69414684e-01 -7.22414581e-04 3.24891627e-01 1.32891202e+00
3.54619920e-01 7.60614693e-01 6.60916030e-01 -5.70637047e-01
-1.02152348e-01 -8.50601256e-01 -2.83292741e-01 8.03002000e-01
4.60609108e-01 -8.16234946e-01 -5.91497123e-01 1.64038181e-01] | [7.826423168182373, 1.5839239358901978] |
5e1c19c7-181e-465b-b010-664daaab7bab | bangla-grammatical-error-detection-using-t5 | 2303.10612 | null | https://arxiv.org/abs/2303.10612v1 | https://arxiv.org/pdf/2303.10612v1.pdf | Bangla Grammatical Error Detection Using T5 Transformer Model | This paper presents a method for detecting grammatical errors in Bangla using a Text-to-Text Transfer Transformer (T5) Language Model, using the small variant of BanglaT5, fine-tuned on a corpus of 9385 sentences where errors were bracketed by the dedicated demarcation symbol. The T5 model was primarily designed for translation and is not specifically designed for this task, so extensive post-processing was necessary to adapt it to the task of error detection. Our experiments show that the T5 model can achieve low Levenshtein Distance in detecting grammatical errors in Bangla, but post-processing is essential to achieve optimal performance. The final average Levenshtein Distance after post-processing the output of the fine-tuned model was 1.0394 on a test set of 5000 sentences. This paper also presents a detailed analysis of the errors detected by the model and discusses the challenges of adapting a translation model for grammar. Our approach can be extended to other languages, demonstrating the potential of T5 models for detecting grammatical errors in a wide range of languages. | ['Khondker Salman Sayeed', 'H. A. Z. Sameen Shahgir'] | 2023-03-19 | null | null | null | null | ['grammatical-error-detection'] | ['natural-language-processing'] | [ 9.44248587e-02 1.25391394e-01 6.24459505e-01 -6.65061176e-01
-1.23765981e+00 -4.00112391e-01 1.65356070e-01 3.70705336e-01
-4.85066384e-01 6.61281049e-01 -1.01281172e-02 -7.65332997e-01
6.37752339e-02 -7.49157965e-01 -6.98288262e-01 -1.07853726e-01
8.98584053e-02 7.44404912e-01 3.37625474e-01 -6.76832795e-01
5.17830253e-01 2.63169408e-01 -1.03824568e+00 6.71617687e-01
1.33090687e+00 2.07808018e-01 5.44141769e-01 9.48758483e-01
-1.46623299e-01 5.17035067e-01 -1.17000008e+00 -8.58353436e-01
-2.75600459e-02 -7.87901342e-01 -1.22215855e+00 -1.07170478e-01
3.60375494e-01 9.72657837e-03 1.82621598e-01 1.17551231e+00
6.59242570e-01 -3.73589188e-01 3.90264273e-01 -7.34972358e-01
-7.46667981e-01 8.30042899e-01 2.06710279e-01 4.55072045e-01
4.71711636e-01 -3.21358860e-01 9.31548119e-01 -1.23425639e+00
4.97299463e-01 1.29405785e+00 1.00601685e+00 4.53351706e-01
-9.38416362e-01 -3.82839888e-01 -4.98093694e-01 7.08695650e-02
-1.27970982e+00 -3.03022414e-01 2.30051383e-01 -2.51942098e-01
1.81797695e+00 3.68456721e-01 3.09538990e-01 6.45597696e-01
8.16640437e-01 5.49734175e-01 1.30139387e+00 -1.26486909e+00
-2.13332489e-01 2.97976762e-01 2.35053413e-02 8.77325833e-01
1.87340304e-01 -1.78950533e-01 -4.50027913e-01 1.40007928e-01
3.32585931e-01 -6.20453000e-01 1.11676104e-01 4.84252512e-01
-1.14461613e+00 9.13404405e-01 -2.02725127e-01 6.40714169e-01
3.88808474e-02 -2.18645021e-01 7.47841954e-01 1.06556666e+00
7.56317616e-01 4.32907045e-01 -6.38047040e-01 -3.33848447e-01
-1.05043447e+00 2.39404410e-01 6.01010740e-01 1.48008215e+00
6.30707264e-01 9.02793631e-02 -1.87235147e-01 1.10048568e+00
3.96194637e-01 7.38652766e-01 4.75364834e-01 -3.74540389e-01
8.98176432e-01 5.92958510e-01 -8.34002048e-02 -7.65570223e-01
-4.26145703e-01 -2.50646859e-01 -1.92735538e-01 -6.34024590e-02
6.52669489e-01 -9.37968791e-02 -9.02408957e-01 1.19730508e+00
-5.87096773e-02 -1.02621627e+00 1.04938261e-01 4.28888589e-01
6.69515789e-01 8.19301546e-01 -1.10830866e-01 -7.59411678e-02
1.22880483e+00 -7.85864949e-01 -7.99284697e-01 -2.17276514e-01
1.63340747e+00 -1.53394842e+00 1.27662444e+00 3.16051215e-01
-1.29942989e+00 -5.42319357e-01 -9.18236434e-01 -3.65383834e-01
-3.97522032e-01 3.84230554e-01 -6.68200627e-02 1.01947355e+00
-1.26329792e+00 5.77949762e-01 -5.90396821e-01 -9.31474745e-01
-2.50537306e-01 4.47169811e-01 -3.22961569e-01 -7.75025412e-02
-1.28996742e+00 1.45425332e+00 3.03714186e-01 -3.74609903e-02
-2.00049445e-01 7.67330825e-02 -7.23357439e-01 -2.99368769e-01
-3.12262714e-01 -1.24456093e-01 1.37097812e+00 -7.45207071e-01
-1.36238158e+00 1.31235683e+00 -3.10177386e-01 -3.01415056e-01
5.99302828e-01 -1.48888096e-01 -8.47488463e-01 -2.17189804e-01
2.61839956e-01 1.48941025e-01 4.38753039e-01 -7.13459969e-01
-9.89302635e-01 -4.51003015e-01 -3.01757544e-01 5.86132519e-02
-1.65034041e-01 1.05926085e+00 -2.71370769e-01 -1.05968940e+00
1.52282134e-01 -8.19880366e-01 3.73712391e-01 -6.27409756e-01
1.84819400e-01 -5.78126907e-01 7.06324399e-01 -1.35066807e+00
1.68509412e+00 -1.81003857e+00 -9.09160674e-02 2.72536010e-01
-4.64187086e-01 6.02251232e-01 -2.84033746e-01 1.00019741e+00
9.00837332e-02 1.47231847e-01 -2.93225288e-01 -2.82333046e-01
3.41512486e-02 4.69536960e-01 -8.28470737e-02 3.25916618e-01
3.48380804e-01 8.57922316e-01 -9.53059316e-01 -5.41809916e-01
8.17087516e-02 8.22202712e-02 -3.73977244e-01 8.63931552e-02
1.76073506e-01 3.85152176e-02 -2.40486674e-02 6.79547369e-01
5.66550970e-01 4.99942243e-01 9.25636441e-02 4.28063422e-01
-2.22208247e-01 9.52492654e-01 -8.15385103e-01 1.32246351e+00
-6.04912877e-01 6.83070540e-01 -1.89915821e-01 -8.67877126e-01
1.44174004e+00 4.03295994e-01 -1.66465551e-01 -8.16295505e-01
1.50138751e-01 9.23802376e-01 2.69307733e-01 -5.90806723e-01
7.67533898e-01 -2.00148314e-01 -4.24062848e-01 4.18113083e-01
3.38225394e-01 -4.21898663e-01 4.21621591e-01 -1.31366299e-02
1.06786871e+00 4.43218872e-02 1.94649726e-01 -5.95061123e-01
6.94945872e-01 5.02418220e-01 4.17133301e-01 7.80383825e-01
-1.90441653e-01 6.14280462e-01 1.07871257e-01 -4.97801274e-01
-1.46477234e+00 -4.81161028e-01 -2.27797523e-01 1.15201533e+00
-4.78053391e-01 -6.74420118e-01 -1.06598902e+00 -1.09299016e+00
-2.45604873e-01 1.06127536e+00 -2.77978659e-01 -9.00750533e-02
-9.48446989e-01 -7.92890310e-01 1.03459728e+00 1.67505667e-01
2.97066599e-01 -1.37239575e+00 -2.24171862e-01 7.00789928e-01
-5.97959578e-01 -8.99631679e-01 -5.25041640e-01 2.42912397e-01
-9.49464142e-01 -6.93731308e-01 -3.61354440e-01 -1.36095262e+00
6.51041269e-01 -1.14523374e-01 1.16548491e+00 3.64698648e-01
5.08124121e-02 -8.07451308e-02 -8.95466089e-01 -6.23705566e-01
-1.42703176e+00 1.20148975e-02 -1.42280743e-01 -6.37657404e-01
1.00562418e+00 1.02851868e-01 2.65556455e-01 4.59981084e-01
-5.86815774e-01 -2.62129277e-01 4.48978037e-01 9.57085967e-01
-9.03554820e-03 1.11343265e-01 5.16582251e-01 -9.96126950e-01
8.09439957e-01 4.17820700e-02 -4.90967721e-01 3.62051278e-01
-7.45914817e-01 -2.14411225e-02 6.44848347e-01 -8.36907141e-03
-7.73132205e-01 -3.06761891e-01 -8.15692723e-01 5.24285614e-01
1.04851820e-01 4.41785544e-01 -1.96918145e-01 -3.01430881e-01
7.32301533e-01 2.18161345e-01 -1.78519115e-01 -6.25705004e-01
-2.07245961e-01 1.34048319e+00 2.25439057e-01 -4.83230174e-01
4.08220470e-01 -5.33877432e-01 -4.63431537e-01 -7.13959098e-01
-4.04343724e-01 -3.20904642e-01 -8.18011343e-01 -1.06681705e-01
3.93731385e-01 -7.10424423e-01 1.48704752e-01 8.64791870e-01
-1.49792540e+00 -1.05558336e-01 -7.43897632e-02 5.19307494e-01
-5.31148851e-01 4.94875699e-01 -9.30413365e-01 -4.94853199e-01
-5.76585114e-01 -1.13219666e+00 1.14460969e+00 -5.36163032e-01
-4.51416582e-01 -1.09409988e+00 4.66042720e-02 4.17485654e-01
3.33568454e-01 -3.40593308e-01 1.27395892e+00 -6.02527976e-01
2.02108011e-01 -1.33236453e-01 -8.06487203e-02 7.32344866e-01
1.88503265e-01 1.14616670e-03 -6.02260649e-01 -3.96142721e-01
1.21230278e-02 -4.69971262e-02 4.42517519e-01 -3.56266238e-02
5.08044660e-01 -3.48723710e-01 2.21643433e-01 1.40603527e-01
1.19266438e+00 2.36223444e-01 8.20764780e-01 6.56076789e-01
5.05129099e-01 6.44564271e-01 9.98932183e-01 1.22513801e-01
5.61037958e-01 7.40052938e-01 1.17376797e-01 7.34985527e-03
-2.85608768e-01 -2.26040229e-01 8.90618563e-01 1.59032071e+00
3.20958085e-02 -3.65924597e-01 -1.19434631e+00 8.42813730e-01
-1.48035645e+00 -6.32206798e-01 -7.40931928e-01 2.04049802e+00
1.07705021e+00 3.10206741e-01 9.47284177e-02 4.76096541e-01
9.01414931e-01 -3.95365596e-01 2.98466265e-01 -1.49265993e+00
-2.25685462e-01 5.03744125e-01 4.85710979e-01 9.35020387e-01
-8.38587403e-01 1.55101681e+00 7.27415323e+00 8.40004265e-01
-8.81840110e-01 4.00771171e-01 -7.69894198e-02 5.96688569e-01
-1.04470417e-01 5.09457812e-02 -1.06861687e+00 6.07597470e-01
1.41058242e+00 -1.57864317e-01 1.42788276e-01 3.89978111e-01
3.77604902e-01 4.15130518e-02 -7.94863284e-01 4.82102066e-01
3.19488645e-01 -7.31566072e-01 1.02767751e-01 -1.59725159e-01
4.57475841e-01 1.08371198e-01 -3.61802161e-01 4.78988647e-01
3.62818003e-01 -7.00137198e-01 9.14164662e-01 -4.74883430e-02
8.85934889e-01 -8.60542655e-01 1.13404012e+00 2.78726339e-01
-4.99340922e-01 1.93904847e-01 -7.97147214e-01 -3.01439464e-01
-1.05139047e-01 6.22512400e-01 -1.17699575e+00 5.67386866e-01
8.93176973e-01 4.93650705e-01 -7.47547805e-01 8.10010552e-01
-3.34903449e-01 9.74786222e-01 -1.70970127e-01 -2.89475888e-01
2.98388273e-01 -1.98646203e-01 5.63792825e-01 1.98871219e+00
9.03135955e-01 -2.99850941e-01 -3.12159121e-01 1.90270737e-01
9.86548588e-02 7.32159674e-01 -4.48566467e-01 1.62859246e-01
3.24953765e-01 7.35537529e-01 -5.11647224e-01 -3.97193760e-01
-4.41601664e-01 1.36195076e+00 4.79356080e-01 -1.30763009e-01
-5.29620767e-01 -1.03301954e+00 1.86972380e-01 9.38790664e-02
3.22358847e-01 -2.60441601e-01 -3.50737184e-01 -9.63768780e-01
2.64592409e-01 -1.44563127e+00 3.03398401e-01 -7.65714467e-01
-9.53021109e-01 8.76316309e-01 -2.65203983e-01 -1.05252087e+00
-1.73505351e-01 -9.42609131e-01 -3.09750825e-01 1.32106435e+00
-1.07612216e+00 -1.25362659e+00 2.37385005e-01 3.79921973e-01
8.10269713e-01 -2.41631791e-01 1.07089388e+00 4.57414657e-01
-2.97692955e-01 1.00151813e+00 3.47551048e-01 2.34768391e-01
1.03089130e+00 -1.41546202e+00 8.78334761e-01 1.25586867e+00
9.62153822e-02 3.80663455e-01 8.36186051e-01 -7.41014838e-01
-1.01979387e+00 -1.42805493e+00 2.29913259e+00 -6.60026252e-01
4.92518634e-01 -5.60981333e-01 -1.03798854e+00 7.86854923e-01
2.49842688e-01 -3.52305591e-01 3.71653318e-01 9.82590169e-02
3.66408355e-03 3.64243984e-02 -1.37293375e+00 2.22150475e-01
1.00418174e+00 -6.80296600e-01 -1.18325019e+00 5.95153809e-01
3.55140328e-01 -5.85729003e-01 -1.05363917e+00 2.80521870e-01
1.72233820e-01 -7.42174625e-01 1.66327432e-01 -5.22424877e-01
4.01297837e-01 -2.94750094e-01 -1.22880369e-01 -1.62462485e+00
-4.57639843e-01 -6.79100394e-01 4.45754558e-01 1.31958067e+00
7.20644057e-01 -5.97793519e-01 2.05889046e-01 4.62575443e-02
-6.33876443e-01 -3.56801718e-01 -1.11042595e+00 -1.04026902e+00
5.77127039e-01 -5.73682249e-01 4.70736414e-01 8.18893433e-01
2.81811655e-01 2.46647090e-01 -4.27769631e-01 2.22823456e-01
2.32711598e-01 -2.94533819e-01 4.97751951e-01 -8.97843540e-01
-4.28553410e-02 -5.29749356e-02 -6.29047573e-01 -8.40212107e-01
-1.88761443e-01 -1.14040375e+00 1.89662114e-01 -1.40468156e+00
-1.15822226e-01 -5.53562760e-01 -1.44407572e-02 6.03443265e-01
-2.42131904e-01 5.71399868e-01 4.46994528e-02 1.44028604e-01
-7.00252354e-02 2.86917716e-01 9.29592431e-01 -2.30330639e-02
-3.26373614e-02 -7.66379014e-02 -4.51644480e-01 4.02052253e-01
9.53109860e-01 -9.13774729e-01 1.74649358e-01 -1.05252659e+00
4.08561081e-01 -2.88419664e-01 -2.91370451e-01 -7.99701393e-01
-3.26287374e-02 2.77600378e-01 1.56609327e-01 -7.53667057e-01
-4.59982425e-01 -6.54132783e-01 -1.07766651e-01 5.81051230e-01
-1.44020706e-01 8.78449440e-01 3.09804797e-01 -2.05779448e-01
-3.35240901e-01 -7.25031912e-01 8.00137103e-01 -7.91618153e-02
-6.56562030e-01 -4.25757796e-01 -9.86307561e-01 1.69160977e-01
7.60125816e-01 -3.87317806e-01 -2.10345760e-01 -1.74494833e-02
-6.19051397e-01 -3.86464223e-02 4.89788830e-01 6.30888283e-01
5.11521101e-01 -1.06537426e+00 -1.22330379e+00 5.48539042e-01
5.42757452e-01 -5.60144842e-01 -4.48831826e-01 7.00037539e-01
-1.09113336e+00 5.25314569e-01 -3.43829244e-01 -6.54057384e-01
-1.75121105e+00 2.04937264e-01 3.10143530e-01 -2.76272565e-01
-6.07622623e-01 7.61795402e-01 -6.23790622e-01 -1.03009284e+00
-1.33081809e-01 -5.50605178e-01 5.55949891e-03 -3.73496175e-01
2.93507069e-01 4.93620902e-01 9.80794132e-01 -1.07396913e+00
-4.12151456e-01 5.98347247e-01 -3.38529408e-01 -7.91459009e-02
1.17816472e+00 -5.60739040e-01 -4.97249931e-01 3.29962432e-01
9.79766846e-01 3.62425089e-01 -1.13397188e-01 -2.91458309e-01
4.24087465e-01 -4.00664240e-01 -3.77567224e-02 -1.04901600e+00
-3.45648527e-01 9.64142382e-01 5.07039011e-01 6.48268461e-02
1.07706475e+00 -1.75222710e-01 8.80338728e-01 4.86738026e-01
6.04476511e-01 -1.60139632e+00 -4.82458591e-01 1.31965065e+00
8.69619668e-01 -1.10124922e+00 -4.91533041e-01 -5.20502746e-01
-3.86766970e-01 1.35482812e+00 2.80245513e-01 5.90656362e-02
2.42362887e-01 2.32969761e-01 5.32276988e-01 -3.96681018e-02
-4.71065521e-01 5.39874434e-02 1.11401096e-01 4.88464355e-01
8.56293142e-01 1.16003051e-01 -1.09282637e+00 8.67446139e-02
-5.23188829e-01 -1.80884182e-01 7.27084875e-01 1.17150831e+00
-6.19131505e-01 -1.72022867e+00 -5.85505128e-01 2.59690404e-01
-8.52436185e-01 -4.07964677e-01 -4.67805713e-01 7.86411941e-01
2.46766880e-01 1.24331474e+00 -2.81119831e-02 -4.59024072e-01
7.08442748e-01 3.32283944e-01 7.58487642e-01 -8.40428650e-01
-1.07518923e+00 -1.14858739e-01 4.94138688e-01 -2.43908852e-01
-2.46183768e-01 -7.98536301e-01 -1.03702855e+00 -5.46517015e-01
-6.41769469e-01 4.56601173e-01 7.82638788e-01 9.95962679e-01
2.06785142e-01 1.38234615e-01 6.04322374e-01 6.07674271e-02
-6.41797483e-01 -1.47057867e+00 -4.65781003e-01 3.00150603e-01
3.14454474e-02 -3.84775624e-02 -2.46869817e-01 9.44782719e-02] | [11.066154479980469, 10.62102222442627] |
7e9304a9-034a-41fc-b91b-182eb32e9f79 | facial-synthesizing-dynamic-talking-face-with | 2108.07938 | null | https://arxiv.org/abs/2108.07938v1 | https://arxiv.org/pdf/2108.07938v1.pdf | FACIAL: Synthesizing Dynamic Talking Face with Implicit Attribute Learning | In this paper, we propose a talking face generation method that takes an audio signal as input and a short target video clip as reference, and synthesizes a photo-realistic video of the target face with natural lip motions, head poses, and eye blinks that are in-sync with the input audio signal. We note that the synthetic face attributes include not only explicit ones such as lip motions that have high correlations with speech, but also implicit ones such as head poses and eye blinks that have only weak correlation with the input audio. To model such complicated relationships among different face attributes with input audio, we propose a FACe Implicit Attribute Learning Generative Adversarial Network (FACIAL-GAN), which integrates the phonetics-aware, context-aware, and identity-aware information to synthesize the 3D face animation with realistic motions of lips, head poses, and eye blinks. Then, our Rendering-to-Video network takes the rendered face images and the attention map of eye blinks as input to generate the photo-realistic output video frames. Experimental results and user studies show our method can generate realistic talking face videos with not only synchronized lip motions, but also natural head movements and eye blinks, with better qualities than the results of state-of-the-art methods. | ['Xiaohu Guo', 'Madhukar Budagavi', 'Saifeng Ni', 'Ming Zeng', 'Yifei HUANG', 'Yifan Zhao', 'Chenxu Zhang'] | 2021-08-18 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zhang_FACIAL_Synthesizing_Dynamic_Talking_Face_With_Implicit_Attribute_Learning_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zhang_FACIAL_Synthesizing_Dynamic_Talking_Face_With_Implicit_Attribute_Learning_ICCV_2021_paper.pdf | iccv-2021-1 | ['3d-face-animation', 'talking-face-generation'] | ['computer-vision', 'computer-vision'] | [ 9.89992693e-02 9.72638354e-02 1.41999438e-01 -4.40438569e-01
-6.92953348e-01 -4.58678842e-01 4.72927153e-01 -9.63806510e-01
3.16115618e-01 6.31457984e-01 4.07675177e-01 3.72362822e-01
4.91408288e-01 -4.50933784e-01 -9.59569752e-01 -8.45610976e-01
2.62518615e-01 1.61738053e-01 -1.27637088e-01 -4.18857932e-02
-7.17166737e-02 3.51853013e-01 -2.06271458e+00 4.62027520e-01
5.45553029e-01 1.14934814e+00 -6.61755353e-02 8.58530223e-01
-1.69029161e-02 8.66252124e-01 -8.75544310e-01 -3.58148783e-01
-1.64928734e-02 -6.87904477e-01 -2.10489973e-01 4.18103844e-01
7.28961587e-01 -5.39474368e-01 -5.45493960e-01 9.98162985e-01
9.14354801e-01 1.31813744e-02 5.65904379e-01 -1.79490864e+00
-8.39897573e-01 9.82801914e-02 -5.54022551e-01 -2.37277895e-01
8.81880641e-01 7.35034764e-01 2.06095770e-01 -9.42619681e-01
5.43594062e-01 1.74555862e+00 4.86691117e-01 1.02273369e+00
-1.03043127e+00 -1.26438761e+00 2.30342045e-01 5.02983212e-01
-1.60247588e+00 -1.06066823e+00 9.74883616e-01 -3.10875565e-01
2.72724092e-01 3.71057838e-01 9.36304152e-01 1.67071927e+00
-1.43843710e-01 5.94142377e-01 7.51984358e-01 -1.91277385e-01
-5.69293387e-02 9.56145227e-02 -7.11476028e-01 4.25846130e-01
-4.92727488e-01 2.92212039e-01 -5.20210266e-01 1.70692485e-02
9.15065110e-01 -1.85177222e-01 -6.58752918e-01 1.24032954e-02
-1.17601490e+00 5.14699697e-01 -3.81393284e-02 -1.40314624e-01
-3.67355824e-01 1.29630804e-01 9.54965875e-02 1.86723825e-02
3.49251211e-01 -1.09335676e-01 -1.05198056e-01 2.96068303e-02
-7.48999953e-01 1.53875992e-01 6.60842538e-01 1.32900310e+00
6.30311728e-01 5.71367800e-01 -4.66119349e-01 7.63565540e-01
3.67992699e-01 9.52039301e-01 4.67828363e-01 -1.30013704e+00
1.06574230e-01 9.81982574e-02 1.67183980e-01 -1.00759089e+00
5.82506619e-02 1.49210140e-01 -8.78628671e-01 1.98900685e-01
1.45808458e-01 -3.54498804e-01 -7.90595770e-01 2.32506657e+00
5.19783020e-01 8.75583410e-01 9.32612550e-03 8.86515737e-01
1.25783312e+00 7.46517599e-01 9.02875606e-03 -8.05565774e-01
1.19406664e+00 -8.40512574e-01 -1.46071148e+00 1.41641907e-02
-2.60701448e-01 -1.04045522e+00 1.22025228e+00 2.38757059e-01
-1.25584888e+00 -1.01605082e+00 -6.85171604e-01 -1.71600226e-02
2.14015231e-01 2.28177503e-01 2.15840548e-01 6.40492737e-01
-1.10612428e+00 1.28014103e-01 -3.77868891e-01 -3.78368348e-02
2.25643143e-01 2.29631573e-01 -3.92873198e-01 9.58195478e-02
-1.18061161e+00 4.01296139e-01 -9.90625843e-02 6.47014752e-02
-1.13391781e+00 -7.99193263e-01 -9.82226670e-01 1.47033471e-03
5.07942438e-01 -8.16289008e-01 1.30566454e+00 -1.60370362e+00
-2.05114698e+00 6.83129787e-01 -5.89160204e-01 2.63555855e-01
4.95705247e-01 -1.27590403e-01 -7.93063641e-01 2.21859872e-01
-2.72443205e-01 8.89582574e-01 1.36571896e+00 -1.46511722e+00
-1.72707498e-01 -3.01886559e-01 -7.87936077e-02 3.58627856e-01
-2.34065056e-01 2.02950895e-01 -7.72192299e-01 -6.75506592e-01
-2.96117812e-01 -9.62357938e-01 5.83046079e-01 3.55969906e-01
-6.23117626e-01 1.19169824e-01 1.14290190e+00 -6.70719028e-01
9.70768213e-01 -2.39271474e+00 1.63927466e-01 -2.65129894e-01
-5.85150011e-02 2.22568691e-01 -3.96914095e-01 2.74477527e-02
-3.30487996e-01 4.46510408e-03 2.29663759e-01 -5.64899981e-01
-2.90415078e-01 1.45680308e-01 -3.79567474e-01 3.57471228e-01
1.97302461e-01 7.43245661e-01 -9.18157816e-01 -6.64324105e-01
1.48291081e-01 1.10396135e+00 -6.10917389e-01 6.57835007e-01
-3.15414846e-01 9.34171855e-01 -1.73174292e-01 6.61093712e-01
7.29595244e-01 2.50451535e-01 -1.41525909e-01 -5.87090135e-01
1.63458005e-01 4.69106883e-02 -1.18201542e+00 1.51218176e+00
-4.62306380e-01 6.87847316e-01 2.71651715e-01 -1.98904619e-01
9.86352563e-01 7.40573823e-01 3.68400574e-01 -3.50597948e-01
2.38963470e-01 -2.00953826e-01 -7.18328059e-02 -9.82953370e-01
-1.43043712e-01 -6.62576407e-03 4.10135150e-01 3.03548872e-01
1.64221153e-01 -4.34396118e-01 -3.43322039e-01 -3.84200849e-02
4.77755547e-01 3.68413121e-01 -1.26329035e-01 1.00407936e-01
9.43425000e-01 -8.89206469e-01 6.14433765e-01 -1.08060606e-01
-1.18374135e-02 9.54116642e-01 4.56873089e-01 -1.61470756e-01
-9.41954732e-01 -1.14576232e+00 2.59960473e-01 7.89094687e-01
2.13528603e-01 -3.24197918e-01 -1.33725774e+00 -2.06526190e-01
-3.89748394e-01 6.37353122e-01 -7.55704582e-01 -4.18437362e-01
-4.83233362e-01 -1.25002995e-01 5.97368181e-01 3.70379359e-01
5.64365447e-01 -1.50156319e+00 1.17416857e-02 -1.02466062e-01
-5.56169033e-01 -1.22580945e+00 -1.17484558e+00 -9.23482597e-01
-3.10360014e-01 -1.08088183e+00 -8.37868035e-01 -7.45075762e-01
7.31658280e-01 -5.87180853e-02 8.48325193e-01 -3.96409445e-02
-2.28062168e-01 3.38836908e-01 -3.36319618e-02 -4.26360428e-01
-6.19756162e-01 -6.10959232e-01 3.81691813e-01 8.46408427e-01
9.53557938e-02 -7.53982663e-01 -8.14798892e-01 5.38565993e-01
-7.19127238e-01 2.37803444e-01 1.47873133e-01 7.27137268e-01
5.89505434e-01 -2.60466695e-01 6.26274467e-01 -5.65357149e-01
4.10800993e-01 -3.25513393e-01 -2.06748784e-01 1.15176804e-01
8.66578594e-02 -4.99036223e-01 7.58520126e-01 -1.20156038e+00
-1.29496837e+00 1.24488227e-01 -3.29483271e-01 -1.14605343e+00
-4.36982155e-01 -3.73893917e-01 -1.04727387e+00 1.75357968e-01
4.59729910e-01 3.76256317e-01 3.61890137e-01 -1.96800843e-01
4.24842894e-01 8.80873501e-01 1.08382523e+00 -4.99000907e-01
8.44552815e-01 4.51691270e-01 -1.92933336e-01 -1.00931776e+00
-5.65062582e-01 1.80655941e-01 -3.99690002e-01 -5.09607673e-01
8.82427394e-01 -1.13422728e+00 -1.21987808e+00 1.02710092e+00
-1.35119319e+00 -2.22436085e-01 -1.80104747e-01 4.05596435e-01
-1.11114359e+00 2.37962335e-01 -5.35201430e-01 -7.45895982e-01
-2.03177959e-01 -1.47761559e+00 1.15026319e+00 4.96305436e-01
-1.48810402e-01 -5.22390306e-01 -2.35422313e-01 3.33739877e-01
2.52504170e-01 3.93100083e-01 6.81451380e-01 -2.00481743e-01
-6.68367147e-01 2.38820240e-02 1.25019282e-01 4.04841512e-01
5.74441552e-01 4.95173007e-01 -1.32203507e+00 -1.96539596e-01
-9.00243036e-03 -1.65880412e-01 -3.96380052e-02 4.42736626e-01
1.33287311e+00 -8.17013562e-01 -1.83685690e-01 1.07566059e+00
7.67192066e-01 4.17465687e-01 9.16567981e-01 -5.79666615e-01
8.14397573e-01 7.67093480e-01 6.10486269e-01 4.12622541e-01
2.68566132e-01 9.34516609e-01 5.76729417e-01 -1.88427240e-01
-4.76085454e-01 -5.90988994e-01 5.11810422e-01 6.88066721e-01
-1.44935459e-01 -2.97850817e-01 -2.33833805e-01 3.40624690e-01
-1.38117063e+00 -1.33279884e+00 1.00148298e-01 2.22444654e+00
1.06447566e+00 -2.99271286e-01 1.40095174e-01 9.65960026e-02
1.14845884e+00 -7.07893493e-03 -9.28536773e-01 -1.04644835e-01
-1.02564514e-01 1.08369432e-01 -3.76223803e-01 6.21018171e-01
-6.69954836e-01 9.26525593e-01 5.68967152e+00 7.41306245e-01
-1.50928736e+00 8.09359327e-02 7.18568683e-01 -5.10974586e-01
-5.90649068e-01 -4.56110090e-01 -6.39933884e-01 9.24270034e-01
8.18194807e-01 -2.21082404e-01 7.63777733e-01 6.58284485e-01
4.39129412e-01 4.41084564e-01 -1.26028061e+00 1.30389380e+00
5.10109127e-01 -1.12334454e+00 3.51700008e-01 -1.78293616e-01
4.66181904e-01 -6.69402838e-01 3.87174964e-01 7.23027810e-02
-5.77852353e-02 -1.27184999e+00 8.29221189e-01 6.62920415e-01
1.68939030e+00 -8.03855181e-01 2.21874028e-01 3.07416499e-01
-1.20380223e+00 7.70596787e-02 3.43447924e-02 3.36583465e-01
2.80846804e-01 -1.46948531e-01 -4.56708580e-01 1.14765204e-01
8.22451472e-01 5.16321838e-01 -8.92814770e-02 7.36533642e-01
-4.71864313e-01 3.10462177e-01 -1.49089932e-01 3.61225277e-01
-4.22918230e-01 -5.84276766e-02 6.60241246e-01 8.18171203e-01
5.73279858e-01 4.36060578e-01 -1.94871411e-01 1.01985562e+00
-2.26671040e-01 1.65641338e-01 -8.40264440e-01 1.36692747e-01
9.14343357e-01 1.15812647e+00 1.24431141e-01 -2.35253230e-01
-2.72810578e-01 9.24471617e-01 -3.71072382e-01 5.99468112e-01
-1.10820758e+00 -1.63120151e-01 1.29703081e+00 4.33791906e-01
-4.75589447e-02 2.49350116e-01 3.07067931e-01 -9.10670340e-01
-1.05136055e-02 -1.22510409e+00 -1.17092967e-01 -1.45957243e+00
-8.20411921e-01 9.72135961e-01 -1.93416104e-01 -1.30997765e+00
-6.56976700e-01 -8.47070217e-02 -9.66243923e-01 1.18095732e+00
-1.32453990e+00 -1.36182952e+00 -6.98289633e-01 1.19973660e+00
7.76943386e-01 -2.96538264e-01 8.70109677e-01 3.36835653e-01
-4.34654742e-01 1.07581139e+00 -4.94246274e-01 2.19991922e-01
1.12279665e+00 -5.58581650e-01 3.87755126e-01 3.97066742e-01
-2.02570427e-02 3.30767959e-01 7.70989597e-01 -4.83706802e-01
-1.31798673e+00 -1.13152766e+00 5.44318318e-01 -4.38990325e-01
1.59275711e-01 -4.39484894e-01 -9.48955595e-01 7.48642802e-01
3.87095183e-01 2.11959690e-01 6.44894540e-01 -6.75081789e-01
-2.61734009e-01 -2.52422094e-01 -1.22517860e+00 7.79718459e-01
1.14846635e+00 -6.62552774e-01 -2.72067070e-01 3.23016584e-01
9.70306456e-01 -6.67431891e-01 -6.19301915e-01 5.34100831e-01
6.85183883e-01 -1.04830194e+00 1.14219737e+00 -5.18626630e-01
3.05400133e-01 -3.59040886e-01 9.79251340e-02 -1.32924366e+00
-5.64860702e-02 -1.25045347e+00 -3.91003350e-03 1.77466989e+00
-1.12433801e-03 -3.57068986e-01 5.38941264e-01 3.90203118e-01
-1.06722035e-01 -5.17656326e-01 -7.15963304e-01 -4.39751923e-01
-2.44438544e-01 -1.35023788e-01 1.02303338e+00 8.98208678e-01
-3.37037295e-01 3.91540051e-01 -7.29801059e-01 2.51630217e-01
5.09429514e-01 -6.50600120e-02 1.15593815e+00 -1.18873823e+00
-1.10080548e-01 -1.05568096e-01 -2.18414128e-01 -8.17610443e-01
5.61544657e-01 -2.25219563e-01 3.89660187e-02 -8.78666401e-01
-6.83151111e-02 -9.30034220e-02 2.60335118e-01 2.15939626e-01
-1.85545936e-01 3.74135762e-01 1.52242780e-01 1.61532294e-02
-3.90110202e-02 8.06911767e-01 1.69903016e+00 -1.40532956e-01
-1.75927818e-01 1.51336342e-01 -4.62327033e-01 1.01549232e+00
3.69220018e-01 -1.08554520e-01 -7.73628652e-01 -3.69171560e-01
-2.87329674e-01 7.42486596e-01 3.59053791e-01 -9.38668072e-01
1.50290325e-01 -3.02333951e-01 5.48555434e-01 -3.26221436e-01
1.03540039e+00 -7.97374189e-01 5.88673294e-01 1.07055835e-01
-3.27762961e-01 6.99982122e-02 2.59652704e-01 4.69144106e-01
-2.90167630e-01 2.79132843e-01 8.50255311e-01 -8.37578103e-02
-2.26515517e-01 8.78193140e-01 -1.04414396e-01 1.80649266e-01
1.17266822e+00 -3.25419098e-01 -2.61961490e-01 -1.24910331e+00
-9.63609219e-01 -1.12888133e-02 6.03818834e-01 8.17829788e-01
7.00343192e-01 -1.52813399e+00 -8.48424852e-01 7.89014816e-01
-1.15740106e-01 4.47753668e-02 5.66536188e-01 4.69765067e-01
-1.72042593e-01 -9.64712873e-02 -5.01129627e-01 -6.90714180e-01
-1.60731113e+00 8.50383878e-01 4.11278456e-01 7.45500326e-01
-4.40277994e-01 8.05654645e-01 7.33609319e-01 -9.06297565e-03
5.18049836e-01 1.47281006e-01 -3.54468316e-01 -4.77304636e-03
7.43305445e-01 8.12636912e-02 -4.04093742e-01 -1.08341336e+00
-7.34523833e-02 8.86710882e-01 3.81637931e-01 -1.64263113e-03
8.23299885e-01 -2.84081548e-01 2.64597714e-01 3.46511543e-01
1.13196754e+00 4.24565941e-01 -1.65960884e+00 -7.34890103e-02
-1.04837203e+00 -5.92159390e-01 -3.92182499e-01 -6.71906412e-01
-1.48546684e+00 1.10404265e+00 4.85389292e-01 -2.51629561e-01
1.40429676e+00 -5.14264926e-02 7.35448420e-01 -2.24215895e-01
2.48994738e-01 -6.94589794e-01 4.75650996e-01 3.85510065e-02
1.17002475e+00 -9.39866900e-01 -5.28411984e-01 -7.64596999e-01
-7.84272313e-01 8.79242837e-01 1.02559376e+00 1.80092379e-01
5.13066232e-01 5.14848769e-01 4.44654614e-01 3.31461638e-01
-9.11354840e-01 -2.30601672e-02 1.53137058e-01 1.11490500e+00
1.72409400e-01 -2.51467019e-01 4.08417165e-01 5.92468619e-01
-3.52550924e-01 6.79491311e-02 3.77883047e-01 -7.41432905e-02
-2.73237955e-02 -8.00597966e-01 -5.53060114e-01 -1.81948408e-01
-4.69693840e-01 -1.15015984e-01 -3.15804631e-01 7.79627740e-01
3.27476263e-01 9.36937392e-01 4.16059405e-01 -5.22098064e-01
2.88539767e-01 3.29578221e-01 5.25797367e-01 -4.73711282e-01
-2.91895419e-01 2.77314693e-01 -1.80833399e-01 -6.90517783e-01
-2.87157327e-01 -4.60839778e-01 -1.17933547e+00 -4.10640359e-01
-1.20176211e-01 -8.54696333e-02 4.82461840e-01 6.68620765e-01
4.97504890e-01 5.84371865e-01 9.18553770e-01 -1.20340908e+00
-2.42216423e-01 -1.23727882e+00 -3.41777116e-01 7.73892105e-01
5.97301900e-01 -6.78747952e-01 -6.21048689e-01 5.60649812e-01] | [13.200947761535645, -0.4139115810394287] |
466a2086-556a-4468-8091-fa0b65c3ab21 | faster-ltn-a-neuro-symbolic-end-to-end-object | 2107.01877 | null | https://arxiv.org/abs/2107.01877v1 | https://arxiv.org/pdf/2107.01877v1.pdf | Faster-LTN: a neuro-symbolic, end-to-end object detection architecture | The detection of semantic relationships between objects represented in an image is one of the fundamental challenges in image interpretation. Neural-Symbolic techniques, such as Logic Tensor Networks (LTNs), allow the combination of semantic knowledge representation and reasoning with the ability to efficiently learn from examples typical of neural networks. We here propose Faster-LTN, an object detector composed of a convolutional backbone and an LTN. To the best of our knowledge, this is the first attempt to combine both frameworks in an end-to-end training setting. This architecture is trained by optimizing a grounded theory which combines labelled examples with prior knowledge, in the form of logical axioms. Experimental comparisons show competitive performance with respect to the traditional Faster R-CNN architecture. | ['Fabrizio Lamberti', 'Lia Morra', 'Filomeno Davide Miro', 'Francesco Manigrasso'] | 2021-07-05 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 3.45683992e-01 5.93748987e-01 -8.08064565e-02 -6.41296983e-01
-2.01980919e-01 -2.66214550e-01 7.91648567e-01 1.37019873e-01
-3.50581497e-01 3.73705328e-02 3.93343233e-02 -4.69198227e-01
-3.59790146e-01 -7.32003212e-01 -9.98080790e-01 -5.40953279e-02
-2.36287743e-01 6.06104791e-01 1.75137654e-01 -2.78189898e-01
1.69179931e-01 4.08957213e-01 -1.40917349e+00 8.82281661e-01
4.70989674e-01 1.54506814e+00 -1.65193707e-01 4.43958938e-01
-2.43559286e-01 1.79000962e+00 -1.22964874e-01 -6.72029793e-01
8.43309611e-02 -1.81741968e-01 -1.20189512e+00 5.85235655e-02
7.99473226e-01 -6.34537756e-01 -4.19432491e-01 1.11459792e+00
-2.44121239e-01 -2.97866315e-02 2.82587379e-01 -1.40348458e+00
-1.15742600e+00 8.70259702e-01 -5.33973314e-02 2.40019500e-01
3.51377755e-01 1.33523405e-01 1.32468367e+00 -9.23188627e-01
6.28593624e-01 1.68414164e+00 5.99952757e-01 4.07117128e-01
-1.33300066e+00 -1.40735880e-01 2.84637958e-01 4.69152063e-01
-1.19331884e+00 -4.81522530e-01 7.96889663e-01 -4.27476645e-01
1.27005875e+00 -6.47031143e-02 5.54902256e-01 9.71112490e-01
-2.67098993e-01 1.11379981e+00 8.83350790e-01 -5.06401360e-01
4.29644436e-01 -2.47899909e-02 1.83283016e-01 1.23466766e+00
2.55721360e-01 -1.75760522e-01 -5.92387855e-01 6.31688535e-02
8.06283534e-01 -6.79341778e-02 3.00032627e-02 -7.15063035e-01
-1.41982567e+00 6.56062782e-01 1.12541401e+00 1.61134034e-01
-3.46372038e-01 9.21131790e-01 4.98884201e-01 1.68330297e-01
2.25408569e-01 5.90674758e-01 -3.42362970e-01 4.13455665e-01
-6.01531386e-01 2.27807984e-01 6.37300253e-01 9.82847035e-01
4.71550137e-01 -5.02201356e-02 1.08880429e-02 2.14548871e-01
3.81468326e-01 1.77659139e-01 2.01321185e-01 -1.26644063e+00
3.38612765e-01 9.95521605e-01 -2.80463159e-01 -9.08987224e-01
-2.30573818e-01 -2.46688113e-01 -5.54555178e-01 3.27944577e-01
3.40681285e-01 2.50933707e-01 -6.42683685e-01 1.41813064e+00
6.23065531e-02 4.57053155e-01 2.65428603e-01 1.05410492e+00
9.44379508e-01 2.40337610e-01 5.14461845e-02 4.70170945e-01
1.54858482e+00 -7.08905995e-01 -4.04437780e-01 -2.74712503e-01
6.22941852e-01 -2.86712617e-01 9.30482507e-01 4.23217267e-01
-1.03827775e+00 -3.13786954e-01 -1.06365430e+00 -6.17395282e-01
-6.56926394e-01 1.39879003e-01 1.10785270e+00 1.13307029e-01
-1.18223596e+00 6.03965104e-01 -1.00848722e+00 -5.10051072e-01
1.05286205e+00 4.08960104e-01 -3.33569467e-01 -1.71500772e-01
-8.74278069e-01 8.89825642e-01 7.02593863e-01 4.02209431e-01
-8.56944263e-01 -5.90700150e-01 -1.14454901e+00 9.78743136e-02
7.09454238e-01 -8.97744954e-01 1.50672030e+00 -1.17994952e+00
-1.23360121e+00 1.03672290e+00 1.52403906e-01 -8.30098629e-01
2.10236818e-01 -2.79921860e-01 -1.72466964e-01 3.61924738e-01
4.98605631e-02 9.56606269e-01 5.90498686e-01 -1.02650511e+00
-3.84124517e-01 -3.20954680e-01 7.35319853e-01 -7.81353116e-02
-9.02943835e-02 1.85878292e-01 -1.79397762e-01 -9.52952802e-02
2.04629213e-01 -6.88785791e-01 -2.30058759e-01 4.17047709e-01
-5.70033491e-01 -4.32633996e-01 7.09563911e-01 -2.63646245e-01
4.27471191e-01 -1.97678423e+00 9.08595771e-02 1.77344263e-01
5.01583397e-01 2.26733252e-01 -6.67436346e-02 1.15544401e-01
-2.34070182e-01 -4.97453064e-02 -4.71781045e-02 -1.31019264e-01
3.05518627e-01 4.61001277e-01 -5.67862391e-01 3.55159461e-01
8.21033478e-01 1.42373657e+00 -1.21947002e+00 -5.40183246e-01
2.16035828e-01 4.02014971e-01 -5.04772067e-01 -8.05892274e-02
-8.57253432e-01 -9.60638970e-02 -5.53984940e-01 4.84738827e-01
1.66908413e-01 -5.86754441e-01 3.17227781e-01 -3.38776439e-01
6.00657873e-02 6.34142518e-01 -8.72436345e-01 1.78921318e+00
-2.98169643e-01 9.10679936e-01 -3.57717544e-01 -1.29499912e+00
7.02946186e-01 4.44420129e-01 1.44164786e-01 -7.67754793e-01
3.73550624e-01 2.04114065e-01 2.49085464e-02 -6.94908023e-01
1.23824298e-01 -1.57368273e-01 1.64427489e-01 2.53383666e-01
3.51955086e-01 -6.53725937e-02 1.60351753e-01 4.97889370e-01
1.13838911e+00 4.83154237e-01 2.15400219e-01 -1.04932360e-01
4.02031034e-01 2.55906045e-01 1.78559899e-01 7.30599046e-01
9.88940671e-02 3.01597506e-01 9.30004120e-01 -1.00250959e+00
-1.00173414e+00 -8.79787922e-01 8.09757337e-02 1.06249607e+00
-5.74222803e-02 -4.52838928e-01 -5.07099211e-01 -6.28223658e-01
1.30467296e-01 6.58631265e-01 -5.36014676e-01 -5.00998050e-02
-5.22979736e-01 -1.00889079e-01 6.93785250e-01 8.26643050e-01
5.41215718e-01 -1.28143752e+00 -8.97509217e-01 -4.01475467e-02
1.60298049e-01 -1.79198611e+00 3.63355219e-01 3.33792388e-01
-9.27266419e-01 -1.21702361e+00 1.25508994e-01 -7.14437604e-01
8.81318688e-01 -1.62386149e-02 1.36044848e+00 3.10939193e-01
-3.40885699e-01 3.54960948e-01 -1.04750015e-01 -4.33616549e-01
-1.74307808e-01 -1.63918570e-01 -2.95494348e-01 7.98763260e-02
4.11901712e-01 -3.78892183e-01 -3.68952602e-01 -1.19136736e-01
-9.67216849e-01 3.97038251e-01 6.78596318e-01 6.81171238e-01
3.86166513e-01 7.77506158e-02 6.90903887e-02 -8.28911722e-01
3.55881035e-01 -2.72912890e-01 -7.67553806e-01 3.77637595e-01
-2.86279380e-01 4.45747674e-01 6.64858401e-01 -1.44839093e-01
-8.92617345e-01 1.74000442e-01 2.02031359e-01 -6.40621185e-01
-3.16085398e-01 7.12850273e-01 -6.93835318e-02 -1.18080802e-01
8.17559183e-01 -1.46633282e-01 -9.40695554e-02 -2.71577090e-01
8.22490513e-01 2.73733586e-01 9.92643833e-01 -7.53524959e-01
6.61103725e-01 7.18243659e-01 4.87207681e-01 -4.33130413e-01
-1.41691947e+00 -2.73377299e-01 -9.38570380e-01 3.97331528e-02
9.55988228e-01 -8.40026438e-01 -8.96761656e-01 -3.25376876e-02
-1.36728454e+00 -5.07916510e-01 -4.40545887e-01 4.29407716e-01
-7.59952545e-01 5.56031242e-02 -5.83897412e-01 -5.97308576e-01
-7.62234405e-02 -9.69626307e-01 1.01097202e+00 -1.33267403e-01
-1.52718544e-01 -1.05234945e+00 -4.99528438e-01 4.95572358e-01
2.27947056e-01 3.97713840e-01 9.32785749e-01 -6.86324775e-01
-1.00913930e+00 -1.73138946e-01 -6.30297601e-01 3.95095617e-01
-2.93432295e-01 2.98244543e-02 -1.02671623e+00 2.10793406e-01
-1.74000710e-01 -6.93145633e-01 9.87352252e-01 2.00656503e-01
1.23695362e+00 -4.37322497e-01 -1.39267430e-01 4.56627995e-01
1.42337883e+00 -1.91786006e-01 6.79145217e-01 3.31730902e-01
9.20575202e-01 4.02475268e-01 1.15670666e-01 -3.11446805e-02
6.13619626e-01 4.59066004e-01 7.76449382e-01 -9.76559371e-02
-2.50073731e-01 -1.40202865e-01 1.68008924e-01 1.89021621e-02
-9.40525904e-02 2.60384113e-01 -1.32280397e+00 4.69202220e-01
-2.22739863e+00 -1.13188541e+00 -1.88298985e-01 1.73366272e+00
6.62664890e-01 3.61768544e-01 -2.01995775e-01 -3.76329124e-02
1.87984824e-01 -4.63700257e-02 -3.85754704e-01 -5.55020690e-01
6.93719015e-02 2.54324973e-01 3.61136198e-01 3.88412982e-01
-1.22969675e+00 1.01017320e+00 6.23956680e+00 1.70104399e-01
-9.59393978e-01 -1.95420738e-02 5.07020712e-01 -1.02024660e-01
-5.78289367e-02 3.39525998e-01 -3.90663087e-01 -2.12310433e-01
8.70714724e-01 2.47408912e-01 5.73026597e-01 1.01785660e+00
-1.25330687e-01 3.60890515e-02 -1.72846723e+00 9.29514587e-01
1.89019114e-01 -1.60587394e+00 7.86328036e-03 -1.26939625e-01
4.58519489e-01 2.08186463e-01 5.62025383e-02 2.53798515e-01
6.59197986e-01 -1.26294947e+00 1.03239870e+00 5.58636010e-01
5.50956547e-01 -3.89435768e-01 7.02303648e-01 8.13238993e-02
-8.79186749e-01 -1.54142544e-01 -3.98232371e-01 -2.99468786e-01
-1.53792664e-01 4.19255883e-01 -1.01813591e+00 3.27544212e-01
5.71214199e-01 9.78141308e-01 -6.94090605e-01 7.49255300e-01
-5.31202435e-01 4.03828204e-01 -3.13288331e-01 -5.83578367e-03
4.82668728e-01 3.06266695e-01 2.64854431e-01 1.11866999e+00
-2.31155112e-01 1.55938938e-01 2.07388639e-01 1.58096027e+00
-1.91448107e-01 -3.61661762e-01 -8.87270749e-01 -3.07058066e-01
-9.74728689e-02 1.25428796e+00 -7.38064468e-01 -5.22672534e-01
-5.71503639e-01 6.84823751e-01 6.31841421e-01 2.93620288e-01
-5.70392191e-01 -1.09704673e-01 4.12629575e-01 -1.18811056e-01
4.42399740e-01 -3.61916989e-01 -2.97302157e-01 -1.16323018e+00
1.95275053e-01 -6.27557755e-01 3.63227665e-01 -1.21443641e+00
-1.20643497e+00 3.37952673e-01 3.19179624e-01 -7.11459517e-01
-4.01114486e-02 -1.29692841e+00 -3.53040516e-01 6.31850600e-01
-1.53874230e+00 -1.44808757e+00 -3.23787987e-01 5.94658673e-01
2.13200212e-01 -1.71957567e-01 8.48154366e-01 -1.78300768e-01
-4.43495780e-01 4.91169244e-02 -4.30609941e-01 6.36663020e-01
-1.09948426e-01 -1.44748199e+00 4.63024855e-01 7.94748783e-01
4.11978006e-01 8.33478153e-01 4.72363472e-01 -3.23996097e-01
-1.56342983e+00 -1.04823220e+00 8.11901510e-01 -6.27960682e-01
8.71731699e-01 -4.86787587e-01 -8.95668924e-01 1.06135595e+00
6.47601113e-02 5.52045584e-01 4.46869731e-01 3.79573852e-01
-1.07923377e+00 -1.69912040e-01 -8.23156655e-01 7.21871257e-01
1.06342268e+00 -9.11425352e-01 -7.82925248e-01 5.79578757e-01
7.10375786e-01 -4.22304899e-01 -8.08241844e-01 2.45352477e-01
4.41692173e-01 -8.54317367e-01 1.15923560e+00 -9.00992155e-01
8.36045325e-01 -4.41045374e-01 -2.61151373e-01 -7.24765599e-01
-2.53383577e-01 -3.96786064e-01 -2.58620650e-01 8.29705715e-01
1.73801273e-01 -1.36398554e-01 6.84343696e-01 7.44569302e-01
-1.18283086e-01 -6.30459309e-01 -7.38713861e-01 -7.05550551e-01
-3.04692179e-01 -9.05166090e-01 6.47226632e-01 8.74469161e-01
3.52950335e-01 7.02570498e-01 2.00388774e-01 3.30749005e-01
7.06887305e-01 2.14681979e-02 4.35229123e-01 -1.29941857e+00
-1.48691773e-01 -7.17673242e-01 -1.03591466e+00 -8.23058546e-01
5.87916613e-01 -1.23286319e+00 -1.20593011e-02 -1.77060461e+00
9.00774822e-02 -1.51768342e-01 -4.74155754e-01 1.02637172e+00
3.31993639e-01 1.24093346e-01 7.80638978e-02 5.77544272e-02
-1.15095937e+00 2.97223419e-01 1.08167636e+00 -3.42524737e-01
2.23258525e-01 -6.93067074e-01 -7.10982978e-01 9.06937361e-01
5.49615383e-01 -3.77643734e-01 -6.12687588e-01 -7.70381868e-01
7.33483851e-01 -1.04151748e-01 1.25273359e+00 -1.05915952e+00
4.74565059e-01 -2.07974046e-01 3.25906426e-01 -3.44930589e-01
1.21591218e-01 -9.28708017e-01 -1.91029802e-01 2.77443588e-01
-7.86156714e-01 -5.44890203e-02 5.91659024e-02 3.84698629e-01
-1.65688336e-01 -7.23222569e-02 4.38371152e-01 -4.14047092e-01
-1.15242016e+00 -6.55322568e-03 8.65533501e-02 -2.15052515e-01
6.80587590e-01 -1.19235173e-01 -6.05414152e-01 -3.74140143e-02
-4.09260333e-01 2.95824260e-01 8.85692760e-02 3.58792245e-01
9.54705715e-01 -1.30995262e+00 -4.04177934e-01 -1.10765263e-01
3.19817871e-01 5.02989054e-01 -3.37065935e-01 9.92180765e-01
-7.02103376e-01 7.59405792e-01 -2.73639709e-01 -6.26742899e-01
-8.04333508e-01 8.55083406e-01 7.19790041e-01 5.69579862e-02
-8.77793789e-01 9.23014760e-01 1.97073638e-01 -4.09512669e-01
3.21786135e-01 -8.22385430e-01 -7.63991475e-02 -4.40809697e-01
5.10045469e-01 -6.85375780e-02 1.03289992e-01 -5.86124182e-01
-4.95422453e-01 8.86594653e-02 -1.11432575e-01 -6.17410690e-02
1.33025038e+00 3.94644201e-01 -5.92609465e-01 4.58962262e-01
1.08363461e+00 -7.99710035e-01 -9.81048584e-01 -6.52462780e-01
4.57919061e-01 -1.92996562e-01 3.16302359e-01 -7.44353414e-01
-9.87917960e-01 8.58753741e-01 1.88838333e-01 4.07256372e-02
8.92010927e-01 3.37586343e-01 4.31263655e-01 1.02117574e+00
2.09899649e-01 -7.57786632e-01 5.43951511e-01 6.02743566e-01
7.82269120e-01 -1.29772615e+00 -1.39773982e-02 -3.89147997e-01
-4.72918510e-01 1.41298115e+00 5.49211144e-01 -4.03411150e-01
4.22052294e-01 1.66474760e-01 -2.39814162e-01 -8.97317708e-01
-1.04480338e+00 -5.17878056e-01 4.77132320e-01 5.92455983e-01
4.77388978e-01 1.92479175e-02 4.67691213e-01 1.73848048e-01
-4.80383262e-02 5.10121405e-01 3.28949124e-01 9.05718565e-01
-2.43176520e-01 -6.84799552e-01 -9.51011777e-02 3.20162982e-01
-3.25969934e-01 -2.25939587e-01 -3.46817315e-01 6.29752398e-01
3.42503071e-01 7.82337368e-01 2.81310618e-01 -5.18391132e-02
3.72464448e-01 2.18885094e-01 7.35917389e-01 -7.57219672e-01
-5.47235250e-01 -3.74105126e-01 1.39093801e-01 -8.15043747e-01
-5.76555133e-01 -4.36893165e-01 -1.61057293e+00 -1.49155647e-01
-4.42615896e-02 -3.94205928e-01 6.48825169e-01 1.23369062e+00
1.45044506e-01 5.86719573e-01 -1.01483107e-01 -6.05651736e-01
-5.07293224e-01 -5.11489868e-01 -3.26073170e-01 6.88684344e-01
4.94728357e-01 -5.15261471e-01 8.76930654e-02 3.34056348e-01] | [10.5514497756958, 2.211491107940674] |
4791d68c-eee6-4009-9bc3-c96315bad49d | learning-sequence-descriptor-based-on | 2305.11467 | null | https://arxiv.org/abs/2305.11467v1 | https://arxiv.org/pdf/2305.11467v1.pdf | Learning Sequence Descriptor based on Spatiotemporal Attention for Visual Place Recognition | Sequence-based visual place recognition (sVPR) aims to match frame sequences with frames stored in a reference map for localization. Existing methods include sequence matching and sequence descriptor-based retrieval. The former is based on the assumption of constant velocity, which is difficult to hold in real scenarios and does not get rid of the intrinsic single frame descriptor mismatch. The latter solves this problem by extracting a descriptor for the whole sequence, but current sequence descriptors are only constructed by feature aggregation of multi-frames, with no temporal information interaction. In this paper, we propose a sequential descriptor extraction method to fuse spatiotemporal information effectively and generate discriminative descriptors. Specifically, similar features on the same frame focu on each other and learn space structure, and the same local regions of different frames learn local feature changes over time. And we use sliding windows to control the temporal self-attention range and adpot relative position encoding to construct the positional relationships between different features, which allows our descriptor to capture the inherent dynamics in the frame sequence and local feature motion. | ['Chen Ye', 'Wenjie Mu', 'Gengxuan Tian', 'Yingfeng Cai', 'Junqiao Zhao', 'Fenglin Zhang'] | 2023-05-19 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 4.19749096e-02 -8.80234778e-01 -4.35907006e-01 -2.41625518e-01
-7.10869491e-01 -5.93830168e-01 7.79209614e-01 1.58317491e-01
-5.55248797e-01 5.80194414e-01 2.97441691e-01 6.61073267e-01
-2.21316323e-01 -5.61965704e-01 -5.14855385e-01 -1.00037956e+00
-1.36585161e-01 -2.35285893e-01 7.33495891e-01 -1.71818405e-01
6.59065187e-01 7.42228627e-01 -1.74354088e+00 1.90660700e-01
4.12979424e-01 1.09528518e+00 7.21073091e-01 5.41235924e-01
-1.02814473e-01 9.83097434e-01 -5.27561247e-01 3.95219535e-01
9.94698852e-02 -5.33162653e-01 -4.42266226e-01 2.16088109e-02
5.08928299e-01 -3.65219176e-01 -1.02487135e+00 1.05163443e+00
5.23955226e-01 7.22819626e-01 4.05802757e-01 -1.34104633e+00
-5.97506106e-01 -2.14883670e-01 -2.58819252e-01 6.47259533e-01
8.37491751e-01 7.05865249e-02 9.10623372e-01 -7.02207685e-01
1.03618526e+00 1.06070876e+00 4.75690961e-01 3.99681717e-01
-9.14426029e-01 -4.38742429e-01 2.77319521e-01 8.65987241e-01
-1.86475503e+00 -6.99036002e-01 9.80773211e-01 -3.70197088e-01
9.56568897e-01 3.34860802e-01 1.01807201e+00 9.53941107e-01
3.60345572e-01 7.61603177e-01 6.12696648e-01 -2.09395170e-01
1.53284743e-01 -5.07828116e-01 -1.49892405e-01 5.21941185e-01
-3.14366192e-01 3.25268120e-01 -1.05651152e+00 -6.33192062e-02
9.56007421e-01 5.41771412e-01 -5.87014735e-01 -5.67879438e-01
-1.65602160e+00 6.11429095e-01 5.15535772e-01 7.52299845e-01
-3.63503993e-01 2.47934163e-01 3.14511389e-01 2.33577281e-01
5.27695417e-02 6.89427853e-02 5.29727479e-03 -4.29058552e-01
-1.00852191e+00 3.43954295e-01 4.19415742e-01 1.11649179e+00
1.17090309e+00 -1.62513763e-01 -3.85021806e-01 4.27991897e-01
2.37942442e-01 6.73225403e-01 9.25576150e-01 -1.05452645e+00
2.10113227e-01 2.28366882e-01 2.23599523e-01 -1.57952952e+00
-2.01294377e-01 3.33206803e-01 -4.77859229e-01 -1.89468995e-01
3.82644355e-01 4.22058403e-01 -7.22297013e-01 1.72778785e+00
2.44974032e-01 7.04478741e-01 -3.10455352e-01 1.09660470e+00
6.18633151e-01 8.02260518e-01 -1.09264560e-01 -3.43820900e-01
1.14290607e+00 -7.47398317e-01 -8.66100192e-01 -7.18636066e-02
5.59224010e-01 -6.63039029e-01 5.49153686e-01 -2.01019540e-01
-7.57497311e-01 -8.07556868e-01 -1.16430008e+00 -1.10319637e-01
-3.51835221e-01 -2.64486700e-01 2.48221084e-01 2.34206989e-02
-1.15485835e+00 6.47865236e-01 -9.77639973e-01 -4.85343605e-01
-5.66318184e-02 3.75956714e-01 -6.43886983e-01 -1.26292586e-01
-1.20454347e+00 8.20381224e-01 5.01214683e-01 3.33020203e-02
-6.92720115e-01 -3.28720808e-01 -1.06742632e+00 -1.09120250e-01
3.85483876e-02 -3.51475239e-01 8.84750783e-01 -9.59569395e-01
-1.38346457e+00 5.61866403e-01 -7.67905831e-01 -2.18153507e-01
2.06764698e-01 2.15394702e-02 -4.65831548e-01 3.80344450e-01
1.87689811e-01 5.38474202e-01 8.12108755e-01 -7.32631028e-01
-8.28573346e-01 -2.91429937e-01 -5.98062761e-02 5.08850873e-01
-4.56550717e-02 1.08421639e-01 -5.52773476e-01 -5.02418160e-01
3.82685900e-01 -7.51068830e-01 -4.18464988e-02 1.92872390e-01
2.41056725e-01 -1.43588856e-01 1.03265274e+00 -3.93214405e-01
1.47623229e+00 -2.34133625e+00 1.19354844e-01 1.81365266e-01
9.59145427e-02 1.56979218e-01 -1.74061567e-01 5.36238968e-01
1.52061671e-01 -3.95352185e-01 2.38544986e-01 7.93495551e-02
-2.23202497e-01 4.36781019e-01 -2.44518504e-01 8.55981231e-01
-1.18325697e-02 9.42296326e-01 -1.29468000e+00 -7.61262476e-01
5.03598571e-01 5.60590804e-01 -3.70003343e-01 2.83534020e-01
2.25052804e-01 6.06398880e-01 -8.22197616e-01 4.59918410e-01
4.33617890e-01 -5.47094867e-02 -4.91164774e-02 -3.52447450e-01
-5.37050903e-01 1.38727501e-01 -1.30130506e+00 2.25594640e+00
-6.72726473e-03 9.37327266e-01 -4.11387622e-01 -1.04977238e+00
1.13266599e+00 2.41172403e-01 7.96555579e-01 -1.18645322e+00
-1.47903431e-02 1.59911737e-01 -2.65098572e-01 -7.02862203e-01
6.56556666e-01 2.78747499e-01 7.39084706e-02 1.06966630e-01
1.50349945e-01 2.61719227e-01 1.55074596e-01 -1.78805083e-01
1.17010069e+00 2.93679655e-01 6.53106570e-01 -4.88063879e-02
8.24082613e-01 -2.70963073e-01 8.03708792e-01 6.88788414e-01
-6.41391695e-01 7.97228456e-01 -5.48589788e-03 -7.32395291e-01
-1.08128691e+00 -9.58337665e-01 -3.21884491e-02 8.66473913e-01
9.08977449e-01 -6.42988324e-01 -3.57429981e-01 -3.53962064e-01
-1.38205349e-01 1.08615212e-01 -6.26761258e-01 -3.33149403e-01
-9.14321065e-01 -8.48623142e-02 2.81606942e-01 2.90235341e-01
4.88860428e-01 -1.06852388e+00 -1.08123493e+00 4.52542245e-01
-5.19123435e-01 -9.63126898e-01 -9.39069569e-01 -1.68990269e-01
-4.54723537e-01 -9.09155667e-01 -8.04646313e-01 -8.38397563e-01
4.07793403e-01 6.52263820e-01 6.68411553e-01 -1.81570612e-02
-2.15922877e-01 4.09489244e-01 -5.38612843e-01 2.90265858e-01
1.63913950e-01 -8.65599811e-02 7.99972191e-02 1.82776928e-01
5.25822878e-01 -5.88398218e-01 -8.19714308e-01 5.01698911e-01
-7.34423280e-01 -9.74649042e-02 1.93649977e-01 1.06448615e+00
8.84275556e-01 -3.89230102e-01 4.74690571e-02 1.33769631e-01
9.05647129e-02 -3.78994435e-01 -4.93573457e-01 4.59955931e-01
-5.47863543e-02 7.45113716e-02 2.91141868e-01 -6.69508696e-01
-6.34385943e-01 2.66148508e-01 6.82057664e-02 -8.71807933e-01
-2.32544348e-01 2.73425281e-01 -1.38777256e-01 -3.66466820e-01
3.92620564e-01 8.07360232e-01 -4.36740145e-02 -9.31804329e-02
2.25975335e-01 3.09890926e-01 5.72289586e-01 -4.37435865e-01
5.31354725e-01 6.50902927e-01 -2.79265139e-02 -8.02584469e-01
-4.53370243e-01 -7.76962399e-01 -1.02850199e+00 -3.84236544e-01
9.60633039e-01 -7.89574921e-01 -7.79617429e-01 3.08930397e-01
-1.40853381e+00 -1.74161990e-03 -1.44084513e-01 7.56016731e-01
-8.00012708e-01 6.37051404e-01 -2.95327336e-01 -5.65247655e-01
1.23730823e-01 -1.16036272e+00 1.20831096e+00 4.59188730e-01
-1.03584565e-01 -9.04047012e-01 4.44832951e-01 -5.59967160e-01
2.40401030e-01 4.16705906e-01 2.15554610e-01 -4.81677800e-01
-9.27236974e-01 -1.70123413e-01 -1.50346845e-01 -1.62516013e-01
2.51367211e-01 -1.04575828e-02 -6.84908569e-01 -3.38814378e-01
1.19996794e-01 2.40777671e-01 5.75801730e-01 4.38035250e-01
6.94828093e-01 -1.56803876e-01 -5.48105061e-01 7.67144084e-01
1.36307657e+00 7.45244205e-01 7.38847613e-01 4.93262798e-01
6.96983695e-01 3.33004773e-01 8.30489159e-01 5.38688242e-01
3.94084603e-01 1.12839055e+00 1.92356836e-02 3.07672530e-01
-5.90466522e-03 -3.65312010e-01 4.04470682e-01 8.55004787e-01
-2.37831473e-01 -1.88729130e-02 -7.82401800e-01 5.85114121e-01
-2.17656541e+00 -1.58570313e+00 2.86590010e-01 2.33126950e+00
4.61973399e-01 -3.41671258e-01 -6.44224510e-02 -1.53191939e-01
9.10503328e-01 5.89950740e-01 -5.28599143e-01 2.20989481e-01
-3.57354552e-01 -1.06466815e-01 4.97784078e-01 4.36350495e-01
-1.08907664e+00 8.85063410e-01 6.41229534e+00 8.62302482e-01
-1.31248844e+00 1.02455556e-01 3.36528607e-02 -5.56188188e-02
-6.37297183e-02 2.31644467e-01 -6.36434436e-01 6.66498780e-01
6.68520629e-01 -5.68322361e-01 4.37718153e-01 6.13775611e-01
1.81441560e-01 -1.98664010e-01 -1.12300158e+00 1.50861287e+00
2.33820498e-01 -1.34408033e+00 -6.82913512e-02 4.27482091e-02
5.08706152e-01 5.49833942e-03 -2.19967261e-01 -8.41752738e-02
-3.16042185e-01 -7.87081838e-01 9.16297853e-01 1.07956338e+00
5.52935779e-01 -6.52390540e-01 4.14263189e-01 2.27082670e-01
-1.83380425e+00 -6.62949309e-02 -5.48915207e-01 -8.71413872e-02
2.86352664e-01 8.20224434e-02 -2.63609409e-01 5.70093274e-01
6.65441275e-01 1.30981827e+00 -5.09490848e-01 1.30684102e+00
1.73588723e-01 -9.93327722e-02 -2.66805321e-01 -6.35463297e-02
2.59149939e-01 -2.36916140e-01 6.58621907e-01 1.16828859e+00
6.54895663e-01 2.45239928e-01 4.06417280e-01 5.19616961e-01
5.47643602e-01 -8.42618942e-03 -8.70159864e-01 2.47994527e-01
8.16533804e-01 8.32494617e-01 -5.70064664e-01 -3.51939529e-01
-6.37745976e-01 1.15484548e+00 2.32960060e-01 5.52993536e-01
-8.09249282e-01 -5.61235130e-01 9.66254771e-01 -1.49944335e-01
3.57094467e-01 -7.19812036e-01 3.40756178e-01 -1.46631038e+00
6.41620308e-02 -3.99060071e-01 2.96459138e-01 -7.14032829e-01
-9.70328391e-01 4.51047570e-01 1.13772139e-01 -1.88318288e+00
-5.79316199e-01 -1.52371094e-01 -3.77260119e-01 7.27016807e-01
-1.58753741e+00 -8.56307328e-01 -4.79703814e-01 1.12507796e+00
4.95907336e-01 3.02336942e-02 7.36637890e-01 3.43668193e-01
-2.51365036e-01 4.88231689e-01 3.13764840e-01 1.19717933e-01
8.19093645e-01 -7.49440491e-01 2.72644937e-01 7.92318761e-01
2.54053861e-01 6.84948981e-01 4.88446236e-01 -5.41752517e-01
-1.53890073e+00 -6.14265442e-01 9.88314867e-01 -4.15390968e-01
7.47934461e-01 -2.92040586e-01 -1.09066546e+00 3.52041483e-01
-1.38947457e-01 5.36375523e-01 4.14804190e-01 -6.30234063e-01
-2.73315072e-01 -1.29308805e-01 -8.12371314e-01 3.69700521e-01
1.35675907e+00 -1.03018355e+00 -6.24090910e-01 -4.08598334e-02
5.05619168e-01 -6.46149635e-01 -8.39070857e-01 2.24432461e-02
9.06737566e-01 -1.08848953e+00 1.17021251e+00 -2.72311389e-01
-7.95130283e-02 -7.69226789e-01 -5.81317961e-01 -8.78654838e-01
-5.73121369e-01 -5.86452425e-01 -7.19212890e-02 1.14284539e+00
-3.87907594e-01 -5.14396548e-01 4.44086313e-01 5.45338988e-01
1.08319804e-01 -2.93874353e-01 -1.37731099e+00 -8.69969070e-01
-5.38828254e-01 -1.71477750e-01 7.21992433e-01 1.01077926e+00
5.96289039e-02 -1.64486468e-01 -4.69162881e-01 1.13805987e-01
4.58009630e-01 2.09735513e-01 5.71028411e-01 -9.68850434e-01
2.07401011e-02 -3.65146905e-01 -1.31878746e+00 -1.32492578e+00
3.78841668e-01 -5.91451228e-01 2.75028825e-01 -1.07856631e+00
6.09440841e-02 -1.23498507e-01 -5.80292881e-01 2.40814358e-01
-8.74031186e-02 2.21300218e-02 3.80053669e-01 7.17557728e-01
-8.14121127e-01 8.37432146e-01 1.24347723e+00 -6.80922791e-02
-1.27926290e-01 -3.77462119e-01 1.53308406e-01 3.59743953e-01
2.99869061e-01 -4.03108329e-01 -4.87690687e-01 -4.24569905e-01
-3.62347990e-01 2.82184392e-01 5.54448783e-01 -1.19464779e+00
6.49560452e-01 -5.23588002e-01 5.64544320e-01 -5.27164102e-01
4.32486087e-01 -8.89682233e-01 3.92033041e-01 2.79814869e-01
-2.76634812e-01 3.97104889e-01 -1.59276962e-01 6.98457599e-01
-6.10880852e-01 -6.80532902e-02 6.35695219e-01 -1.07226513e-01
-1.52522171e+00 5.80470800e-01 -2.00445995e-01 -1.26769394e-01
1.11443663e+00 -6.60710692e-01 -1.38252839e-01 -3.93661946e-01
-5.76596200e-01 5.45210466e-02 7.86081731e-01 6.33900404e-01
7.75638640e-01 -1.89942026e+00 -3.08779031e-01 3.05396914e-01
3.62195551e-01 -3.25622410e-01 5.06998837e-01 7.77320921e-01
-4.44255501e-01 4.05959308e-01 -5.69807649e-01 -9.25047815e-01
-1.07564259e+00 7.19996154e-01 3.64655882e-01 1.53115973e-01
-8.89230847e-01 4.60596114e-01 3.66400927e-01 1.72017843e-01
5.74741922e-02 -8.85576606e-02 -3.83157909e-01 2.79436022e-01
9.33099926e-01 2.47077763e-01 -1.59559190e-01 -1.21831834e+00
-5.65997481e-01 1.09966469e+00 1.19372874e-01 -1.12981595e-01
9.68820155e-01 -6.79886580e-01 -5.89646865e-03 7.02883065e-01
1.71747339e+00 -3.64728898e-01 -1.44250345e+00 -5.91553092e-01
1.33830339e-01 -1.11768472e+00 -4.58193943e-02 5.08042090e-02
-8.56865883e-01 6.54837370e-01 8.24864268e-01 -3.01021188e-02
1.13216507e+00 -4.33547162e-02 7.83525944e-01 4.06734884e-01
6.59643888e-01 -1.00662482e+00 -1.59205217e-02 7.05512106e-01
6.73455715e-01 -9.36979949e-01 -9.58833843e-02 -8.82550050e-03
-4.84741062e-01 1.35392988e+00 3.91833931e-01 -2.71383822e-01
5.00071704e-01 -8.07450488e-02 -1.02279127e-01 1.75992828e-02
-6.28088176e-01 -5.07614315e-01 3.92087698e-01 8.72158408e-01
2.35654846e-01 -2.63774812e-01 -2.68464088e-01 1.26776351e-02
1.74632668e-01 -1.93845257e-02 3.18154171e-02 1.19451189e+00
-5.29406965e-01 -1.03679526e+00 -2.83885539e-01 -3.71275581e-02
8.93230960e-02 1.82816982e-01 7.87786469e-02 7.20477939e-01
1.24316663e-01 6.50913596e-01 3.57060492e-01 -4.88250256e-01
3.42782289e-01 -2.68036872e-02 5.48995614e-01 -1.11255653e-01
-1.63439319e-01 3.51687908e-01 -4.39465463e-01 -1.02609551e+00
-8.53251040e-01 -1.04099023e+00 -1.17565906e+00 -3.14588666e-01
-5.22187278e-02 6.02474287e-02 3.97689015e-01 9.85003889e-01
4.51206982e-01 2.20876098e-01 7.94571638e-01 -1.33243310e+00
-2.57793684e-02 -5.24803758e-01 -7.09457517e-01 4.75023448e-01
8.46456289e-01 -9.69489634e-01 -2.05861956e-01 2.82828093e-01] | [7.950123310089111, -1.3689895868301392] |
0ea58a89-6c47-4a88-bf48-5dbfaf738b50 | learning-montezumas-revenge-from-a-single | 1812.03381 | null | http://arxiv.org/abs/1812.03381v1 | http://arxiv.org/pdf/1812.03381v1.pdf | Learning Montezuma's Revenge from a Single Demonstration | We propose a new method for learning from a single demonstration to solve
hard exploration tasks like the Atari game Montezuma's Revenge. Instead of
imitating human demonstrations, as proposed in other recent works, our approach
is to maximize rewards directly. Our agent is trained using off-the-shelf
reinforcement learning, but starts every episode by resetting to a state from a
demonstration. By starting from such demonstration states, the agent requires
much less exploration to learn a game compared to when it starts from the
beginning of the game at every episode. We analyze reinforcement learning for
tasks with sparse rewards in a simple toy environment, where we show that the
run-time of standard RL methods scales exponentially in the number of states
between rewards. Our method reduces this to quadratic scaling, opening up many
tasks that were previously infeasible. We then apply our method to Montezuma's
Revenge, for which we present a trained agent achieving a high-score of 74,500,
better than any previously published result. | ['Tim Salimans', 'Richard Chen'] | 2018-12-08 | null | null | null | null | ['montezumas-revenge'] | ['playing-games'] | [ 1.04929313e-01 5.56922376e-01 4.13611624e-03 8.16667378e-02
-9.54964399e-01 -7.97185779e-01 6.13163650e-01 -6.31718934e-02
-9.90767479e-01 1.27524543e+00 -1.50957704e-01 -4.12631601e-01
-3.56040942e-03 -5.85701764e-01 -9.95908737e-01 -6.87571824e-01
-4.48076844e-01 7.39872932e-01 1.43001035e-01 -3.60990733e-01
3.70415419e-01 1.41903833e-01 -1.26608825e+00 -2.13597596e-01
9.02848840e-01 5.64479828e-01 5.79577029e-01 1.12002635e+00
4.22952503e-01 1.02392101e+00 -7.53156602e-01 2.71154158e-02
4.70123500e-01 -7.76887238e-01 -8.61697435e-01 9.53795910e-02
-3.78674716e-02 -8.29641223e-01 -4.87635821e-01 1.06671643e+00
3.72616678e-01 4.02215034e-01 3.35587919e-01 -1.27614379e+00
-2.14465410e-01 8.31877351e-01 -5.42418957e-01 -3.11933178e-02
3.33355457e-01 5.28837502e-01 1.04360330e+00 -1.15935318e-01
7.83766448e-01 1.02570903e+00 2.04928339e-01 7.36708403e-01
-1.18637896e+00 -6.00406528e-01 2.62131304e-01 9.30163413e-02
-5.48250735e-01 -1.38066962e-01 3.68412197e-01 -2.63987839e-01
1.03303146e+00 -2.64811605e-01 8.99069846e-01 1.19083452e+00
-1.91836894e-01 1.02263522e+00 1.47936964e+00 -4.14897949e-01
6.86608374e-01 -4.12263513e-01 -3.42439145e-01 9.18694019e-01
4.55958843e-02 6.77566886e-01 -3.40457171e-01 -9.40016210e-02
9.54186499e-01 -2.80012041e-02 -9.96221695e-03 -7.27688134e-01
-1.49494147e+00 1.13416362e+00 4.87098008e-01 -9.05413702e-02
-5.53760469e-01 6.38648510e-01 3.16866189e-01 7.44347274e-01
-3.08056474e-01 9.15097535e-01 -3.23897332e-01 -6.32910967e-01
-6.46487057e-01 5.78076184e-01 9.93248999e-01 5.88209093e-01
5.79048097e-01 9.56277549e-02 7.72793517e-02 3.16693187e-01
-1.14612348e-01 6.29660785e-01 4.46491569e-01 -1.70689535e+00
5.88354528e-01 5.22206724e-02 7.27583230e-01 3.24771600e-03
-3.79952461e-01 -3.42637062e-01 -2.30891585e-01 1.02439487e+00
7.57526517e-01 -8.36106479e-01 -6.70306563e-01 1.93438101e+00
2.35912636e-01 1.14210449e-01 4.67193663e-01 8.15322638e-01
3.49109620e-02 5.84586859e-01 -3.42292637e-01 -3.06626409e-01
9.05956030e-01 -1.47585869e+00 -2.64307648e-01 -3.53250742e-01
7.49022841e-01 -2.56478548e-01 1.07852912e+00 7.98337758e-01
-1.24673545e+00 -1.87330589e-01 -1.05835915e+00 3.33812147e-01
2.41158530e-02 -1.28170952e-01 8.14313591e-01 3.09528738e-01
-9.19443607e-01 1.02029836e+00 -1.07352805e+00 -4.09416929e-02
2.28357688e-01 4.70332026e-01 -2.35155404e-01 1.31982595e-01
-9.74985301e-01 1.10905230e+00 5.25757730e-01 -6.28367141e-02
-1.78413308e+00 -2.74477880e-02 -8.21573496e-01 1.94746003e-01
1.03954709e+00 -3.97342175e-01 1.96665168e+00 -8.19369018e-01
-2.11352539e+00 3.61221761e-01 1.66340426e-01 -7.51034439e-01
6.95655704e-01 -4.13513929e-01 4.40442175e-01 1.16234824e-01
2.25110024e-01 8.40941072e-01 6.43400788e-01 -1.05242348e+00
-6.25072956e-01 3.03893845e-04 7.31327355e-01 5.08731842e-01
1.98934600e-01 -1.82492688e-01 1.38819501e-01 -1.06097586e-01
-3.74217868e-01 -1.26628315e+00 -6.64907873e-01 -3.34979296e-01
-1.05784781e-01 -3.10753465e-01 2.99195081e-01 -1.24366887e-01
5.44694304e-01 -1.84989691e+00 4.55640793e-01 -7.38917813e-02
3.48655432e-01 1.49005473e-01 -4.61840212e-01 6.27638817e-01
1.59607112e-01 -2.71213830e-01 -2.78087765e-01 -2.68454462e-01
3.83576393e-01 4.18289870e-01 -2.95056343e-01 3.59757364e-01
-1.93021342e-01 1.15033185e+00 -1.45255411e+00 -2.36884534e-01
3.71284299e-02 -2.48678267e-01 -7.76234746e-01 5.28483689e-01
-5.92770696e-01 4.96408075e-01 -5.09484708e-01 1.67825446e-01
9.45232585e-02 -3.71001393e-01 3.73693198e-01 9.14103627e-01
-1.10107496e-01 4.04652178e-01 -1.03846717e+00 2.01559234e+00
-4.54578489e-01 4.52268600e-01 1.02377430e-01 -1.11776006e+00
4.59504932e-01 2.43985862e-01 3.86877447e-01 -6.53520942e-01
7.80741945e-02 2.56814569e-01 3.22951466e-01 -4.07984853e-01
5.27217925e-01 -2.17073977e-01 -3.85708123e-01 8.03109884e-01
8.05303454e-03 -4.93204534e-01 5.10924876e-01 4.23783153e-01
1.54140306e+00 5.96108794e-01 3.31092775e-01 2.00581163e-01
-7.21699521e-02 2.18503058e-01 2.82317907e-01 1.24210262e+00
-2.81800836e-01 7.09125251e-02 9.73563790e-01 -2.58899033e-01
-1.17526484e+00 -1.23828697e+00 6.37404025e-01 1.11791205e+00
9.94491205e-02 -4.04708713e-01 -8.20772767e-01 -9.02756989e-01
-2.46626288e-02 6.03567243e-01 -8.07090223e-01 1.02100633e-01
-7.01593518e-01 -1.63389310e-01 3.43323350e-01 2.44449854e-01
5.04851639e-01 -1.75220466e+00 -1.42425001e+00 2.76453912e-01
1.25308242e-02 -6.87360525e-01 -4.53985214e-01 6.48564756e-01
-8.23992729e-01 -1.14736676e+00 -8.70624065e-01 -5.59208810e-01
4.87663954e-01 -1.93408132e-02 1.08071792e+00 3.74713242e-02
-2.03118175e-01 3.93088043e-01 -1.64142549e-01 -1.67826816e-01
-4.09634948e-01 2.17489600e-01 1.40550494e-01 -1.01608574e+00
-1.76435679e-01 -6.25255466e-01 -5.99953055e-01 -1.01250052e-01
-6.14830494e-01 8.86773542e-02 7.92707503e-01 1.29706299e+00
1.37413248e-01 -1.77946344e-01 7.02064276e-01 -6.70503795e-01
9.16965187e-01 -4.95804578e-01 -1.24324691e+00 -1.39139652e-01
-4.32631493e-01 4.38310146e-01 8.39481711e-01 -7.22149789e-01
-6.94597542e-01 2.50344455e-01 2.80010909e-01 -2.92100221e-01
-2.03157421e-02 2.42884055e-01 5.26177883e-01 6.65316582e-02
7.73032963e-01 2.58668453e-01 3.65462661e-01 -1.01490960e-01
4.60969120e-01 1.94149047e-01 5.11752486e-01 -8.71894181e-01
8.89371097e-01 -3.19717010e-03 -8.44972022e-03 -3.46852511e-01
-7.23027050e-01 7.38605708e-02 -2.43325755e-01 7.77659100e-03
5.51858664e-01 -8.93050313e-01 -1.71529841e+00 2.49312326e-01
-8.42077076e-01 -1.17457986e+00 -6.79535568e-01 5.39156199e-01
-1.15667486e+00 3.09068859e-01 -7.22467124e-01 -9.52592850e-01
1.45223677e-01 -1.24837470e+00 7.54944563e-01 4.33564484e-01
-3.02200858e-03 -6.15314066e-01 5.54157794e-01 1.60249829e-01
3.07518661e-01 8.11058283e-02 3.68309051e-01 -4.38638747e-01
-7.41803885e-01 2.61502504e-01 1.30776599e-01 8.53382796e-02
-2.21991062e-01 -6.22390211e-01 -4.36015218e-01 -5.79219997e-01
-3.15356292e-02 -1.22058713e+00 7.33208716e-01 2.08207279e-01
7.53111899e-01 -3.98509532e-01 -5.24518564e-02 2.39601895e-01
1.16921258e+00 4.22099054e-01 5.71215153e-01 8.30794752e-01
9.86355916e-02 2.91921824e-01 9.69205618e-01 6.26138866e-01
1.98605344e-01 5.06624460e-01 8.14643502e-01 1.01523951e-01
4.46918428e-01 -4.99575555e-01 7.89011419e-01 6.75846487e-02
-1.97618380e-01 4.43139933e-02 -4.16430265e-01 5.30487657e-01
-2.21427464e+00 -1.11563051e+00 7.04930723e-01 2.17758632e+00
9.86216664e-01 2.72649676e-01 6.40129030e-01 -2.23838463e-01
2.35928997e-01 1.12743817e-01 -9.44634318e-01 -4.92933124e-01
3.82949144e-01 4.12749499e-01 4.70071882e-01 8.13309550e-01
-7.38636374e-01 1.27857983e+00 6.81925631e+00 6.02467716e-01
-7.86326170e-01 -3.80120687e-02 2.93905139e-01 -6.85006559e-01
1.04217105e-01 2.91862965e-01 -3.98959130e-01 3.22204411e-01
8.35039496e-01 -7.78730661e-02 1.20593894e+00 1.08149922e+00
1.69585377e-01 -6.13220453e-01 -1.29649317e+00 7.02489614e-01
-2.69545615e-01 -1.04593933e+00 -8.21222842e-01 3.18682641e-01
8.02763522e-01 1.77178979e-01 3.92101668e-02 8.63142312e-01
1.29333019e+00 -1.22334611e+00 4.97460991e-01 -1.33667275e-01
3.81738782e-01 -8.23012114e-01 4.02932614e-01 7.99011827e-01
-5.91158688e-01 -1.74600095e-01 -3.15960735e-01 -6.79362297e-01
1.24759085e-01 -3.47460024e-02 -1.21361601e+00 1.45798713e-01
2.02039346e-01 1.73338994e-01 -7.74616227e-02 8.49101126e-01
-6.59358680e-01 4.67724264e-01 -4.18548286e-01 -4.92313951e-01
6.99170530e-01 -3.43518168e-01 3.40352058e-01 6.64895773e-01
3.14033389e-01 2.25488067e-01 5.62372565e-01 8.71595740e-01
-8.16068426e-02 -3.34241420e-01 -7.33503461e-01 -1.57988101e-01
2.50274390e-01 1.15196562e+00 -5.98485172e-01 -5.17260849e-01
-7.52661191e-03 1.17359912e+00 8.82009149e-01 2.93524176e-01
-9.59526896e-01 -5.53526521e-01 3.08685988e-01 -3.40086848e-01
5.98653078e-01 -5.34689963e-01 3.69018078e-01 -1.23814499e+00
6.94888318e-03 -1.10954833e+00 1.31775975e-01 -7.74533272e-01
-6.90055966e-01 4.74181026e-01 -3.59500572e-02 -9.65747356e-01
-1.00311041e+00 -3.75584155e-01 -5.91605306e-01 5.79021633e-01
-1.45377874e+00 -4.14367169e-01 1.20994955e-01 4.57608253e-01
6.97285533e-01 -1.72130018e-01 7.81003654e-01 -4.28539544e-01
-2.06058919e-01 2.64949054e-01 1.80656344e-01 -8.57798010e-02
4.52564299e-01 -1.65898860e+00 4.64106172e-01 6.33725762e-01
1.70856819e-01 3.54689628e-01 8.18834364e-01 -4.99575347e-01
-1.32393122e+00 -1.60538092e-01 2.39424571e-01 -1.59108117e-02
9.59026515e-01 -4.22870994e-01 -3.92656595e-01 8.68253350e-01
4.98851031e-01 -1.57390267e-01 4.13021371e-02 1.91933841e-01
-1.98858038e-01 3.64754945e-01 -9.24472511e-01 8.51898134e-01
1.00908136e+00 -2.67036587e-01 -8.75726402e-01 4.20352072e-01
5.85345387e-01 -7.89358556e-01 -4.22667682e-01 -1.93352699e-01
6.08366609e-01 -9.48654413e-01 7.30705082e-01 -8.09957445e-01
7.38784492e-01 -2.76218742e-01 2.17204005e-01 -1.78624666e+00
-9.45842788e-02 -1.19391656e+00 -3.03499609e-01 2.22475544e-01
3.89992028e-01 -6.99601829e-01 1.01599026e+00 1.97402924e-01
9.29078013e-02 -8.07954133e-01 -8.39798629e-01 -1.03120017e+00
3.53036076e-01 1.62842169e-01 2.42489845e-01 4.88951236e-01
5.96928954e-01 5.11471689e-01 -6.69192612e-01 -7.79638067e-02
7.61771739e-01 3.70260417e-01 1.05997741e+00 -7.87151396e-01
-1.02149057e+00 -3.08371872e-01 3.09311599e-01 -1.51321399e+00
3.60197634e-01 -7.09918916e-01 4.35355067e-01 -1.26848662e+00
4.29230183e-01 -4.14399952e-01 -1.54319540e-01 6.54031336e-01
-1.01276711e-01 -1.19944595e-01 7.06959307e-01 4.40184958e-02
-1.15590179e+00 6.36262476e-01 1.68297863e+00 3.18152048e-02
-3.33333910e-01 -1.05978893e-02 -6.36535048e-01 6.42525077e-01
8.54591191e-01 -5.32040417e-01 -5.01470983e-01 -1.59208283e-01
2.06271291e-01 7.07773268e-01 2.95097768e-01 -8.53029132e-01
2.02248916e-01 -3.77135575e-01 1.54465407e-01 -2.05583170e-01
4.72167611e-01 -5.31614959e-01 -2.35492453e-01 1.02597213e+00
-6.38406277e-01 2.34364241e-01 -5.82351908e-02 5.24590909e-01
1.66947007e-01 -7.27660537e-01 5.95046639e-01 -4.71411377e-01
-4.04501349e-01 -2.76627559e-02 -6.67504132e-01 2.69565612e-01
1.22938704e+00 6.28587753e-02 -3.90667498e-01 -8.45286369e-01
-1.01031303e+00 5.95414579e-01 6.18176758e-01 -8.46309364e-02
4.66825187e-01 -9.88558412e-01 -4.76661682e-01 -2.25181803e-01
-3.67391914e-01 9.95570645e-02 3.07403114e-02 4.72760558e-01
-2.69511193e-01 2.01750278e-01 -4.79428321e-01 -2.53243089e-01
-1.10642767e+00 6.58561766e-01 2.10180536e-01 -9.30376709e-01
-7.98412025e-01 4.32873994e-01 9.24786702e-02 -4.29938585e-01
3.30483317e-01 -2.72751927e-01 -3.43926280e-04 -3.61114442e-01
2.62018621e-01 2.84174293e-01 -5.77268422e-01 2.38070279e-01
8.90938938e-02 1.25693411e-01 -2.86852866e-01 -8.43572259e-01
1.48057461e+00 2.59813219e-01 3.19389999e-01 3.38716596e-01
8.28030527e-01 -1.74806803e-01 -1.86625063e+00 2.08783578e-02
-1.72714263e-01 -4.60135847e-01 -3.66021276e-01 -9.35032368e-01
-6.38241410e-01 6.55745625e-01 4.47512448e-01 1.91569164e-01
7.78352678e-01 -1.06189974e-01 7.18107760e-01 1.23998880e+00
8.39342654e-01 -1.14004362e+00 7.05627680e-01 8.30470443e-01
6.85157239e-01 -1.40545225e+00 -6.10817447e-02 4.16074365e-01
-9.80061471e-01 9.88004029e-01 5.22437811e-01 -6.83768749e-01
-2.05573484e-01 3.01441789e-01 -4.28032190e-01 -7.23787546e-02
-9.55157757e-01 -3.05660903e-01 -5.20358086e-01 5.71229398e-01
-9.36891511e-02 2.14882940e-01 -2.13085860e-02 2.11365342e-01
-5.24590075e-01 2.77159423e-01 9.80711579e-01 1.25576854e+00
-7.87183583e-01 -1.34045947e+00 -2.31334791e-01 2.92398721e-01
-3.09857726e-01 -1.69145577e-02 -1.94239587e-01 8.74650061e-01
-3.30247819e-01 7.30254591e-01 -8.10588673e-02 2.30275206e-02
4.66591567e-02 -3.27376015e-02 1.00799060e+00 -6.80521846e-01
-5.31756759e-01 -3.93453576e-02 1.32885113e-01 -7.94270217e-01
-1.38772845e-01 -7.14891076e-01 -1.54613817e+00 -3.11931342e-01
-9.29047689e-02 4.06513304e-01 5.26656091e-01 9.13851321e-01
-1.07459109e-02 4.32020038e-01 7.54767954e-01 -1.20701647e+00
-1.13697183e+00 -7.60155737e-01 -5.45177877e-01 1.85937524e-01
5.85383892e-01 -8.27529252e-01 -5.03711343e-01 -4.53143209e-01] | [4.0217061042785645, 1.6628053188323975] |
b410e9f1-4417-49ce-8fbe-9519352dfe93 | complex-gated-recurrent-neural-networks | 1806.08267 | null | http://arxiv.org/abs/1806.08267v2 | http://arxiv.org/pdf/1806.08267v2.pdf | Complex Gated Recurrent Neural Networks | Complex numbers have long been favoured for digital signal processing, yet
complex representations rarely appear in deep learning architectures. RNNs,
widely used to process time series and sequence information, could greatly
benefit from complex representations. We present a novel complex gated
recurrent cell, which is a hybrid cell combining complex-valued and
norm-preserving state transitions with a gating mechanism. The resulting RNN
exhibits excellent stability and convergence properties and performs
competitively on the synthetic memory and adding task, as well as on the
real-world tasks of human motion prediction. | ['Moritz Wolter', 'Angela Yao'] | 2018-06-21 | complex-gated-recurrent-neural-networks-1 | http://papers.nips.cc/paper/8253-complex-gated-recurrent-neural-networks | http://papers.nips.cc/paper/8253-complex-gated-recurrent-neural-networks.pdf | neurips-2018-12 | ['music-transcription'] | ['music'] | [ 3.84102404e-01 -4.36370701e-01 -3.87579352e-02 -5.25391735e-02
-4.02967155e-01 -2.15644553e-01 8.45726252e-01 -2.67667979e-01
-6.79204822e-01 9.14803028e-01 3.49704087e-01 -3.09888124e-01
3.49665105e-01 -6.50511563e-01 -3.92973304e-01 -9.13363695e-01
-4.04986978e-01 2.08999470e-01 2.46428058e-01 -3.71799588e-01
9.74292457e-02 4.64905977e-01 -1.45034730e+00 2.96144456e-01
3.12760234e-01 9.30195093e-01 2.16419294e-01 8.91320467e-01
3.67524534e-01 1.29380107e+00 -5.09288073e-01 1.17946252e-01
3.23030911e-02 -6.59432948e-01 -3.98243278e-01 -3.54593337e-01
-1.13271564e-01 1.58270210e-01 -9.45361078e-01 6.69854939e-01
8.52453113e-01 6.31323159e-01 6.48234487e-01 -7.98514247e-01
-4.55933809e-01 3.59743208e-01 -2.52832681e-01 7.05438197e-01
3.32781732e-01 2.82889366e-01 9.81998384e-01 -9.12333071e-01
6.74293280e-01 1.22195268e+00 7.33219504e-01 6.96733475e-01
-1.25235772e+00 -5.75612366e-01 1.16956368e-01 3.57886076e-01
-1.15889204e+00 -5.13415933e-01 6.63398385e-01 -2.45145991e-01
1.60548317e+00 5.21352887e-02 1.07074034e+00 1.50767934e+00
5.93307853e-01 1.19135880e+00 7.76810110e-01 -7.41911530e-02
2.62613147e-01 -8.61227691e-01 -1.75917104e-01 6.49855494e-01
6.48778751e-02 3.10096592e-01 -6.08961761e-01 1.13531789e-02
9.88275290e-01 1.55552194e-01 -4.73317534e-01 -4.66941595e-02
-1.78415620e+00 7.22021401e-01 4.78117913e-01 3.06562364e-01
-6.78941131e-01 9.00738478e-01 6.87805653e-01 4.21564788e-01
1.94542214e-01 4.24886823e-01 -6.29312918e-02 -5.77132523e-01
-7.99147367e-01 2.56539166e-01 7.56984591e-01 5.53117871e-01
1.93405643e-01 9.40799773e-01 -2.78241128e-01 5.90287149e-01
7.87581354e-02 6.39828265e-01 1.10179687e+00 -8.77890289e-01
2.44850054e-01 2.81148911e-01 -6.67770803e-02 -1.04334223e+00
-7.88292468e-01 -5.06954432e-01 -1.50313699e+00 8.13783929e-02
1.86857283e-01 6.61424398e-02 -1.10934079e+00 1.94916570e+00
-3.28809768e-01 6.41636848e-01 2.36820504e-01 8.87003958e-01
7.10963130e-01 1.11420739e+00 9.60184559e-02 -4.91637498e-01
8.61164629e-01 -6.33718789e-01 -8.68883729e-01 -2.75415540e-01
5.79670846e-01 -3.69725078e-01 8.66847456e-01 4.16553736e-01
-1.25736582e+00 -6.68055356e-01 -1.22596812e+00 -3.97892334e-02
-1.87529638e-01 -1.62726313e-01 8.38843644e-01 3.18079054e-01
-1.14603817e+00 8.25452268e-01 -1.41572273e+00 -6.42813696e-03
2.65953451e-01 3.20489109e-01 -8.95199925e-02 1.34947523e-01
-1.48345530e+00 9.47185576e-01 3.37075233e-01 4.92807567e-01
-8.38610590e-01 -3.17158490e-01 -9.67254877e-01 -1.10048577e-01
-1.62339300e-01 -8.24592113e-01 1.22729385e+00 -7.96916127e-01
-1.87359118e+00 6.12115562e-01 2.72261002e-03 -1.00177431e+00
3.64259213e-01 -2.52841622e-01 -6.67986095e-01 -3.67585127e-03
-2.99524397e-01 7.62636542e-01 9.68729019e-01 -2.84385115e-01
-2.42702916e-01 -4.52174284e-02 -5.55844188e-01 2.40595862e-01
8.33120272e-02 -2.82340318e-01 1.45372804e-02 -1.28004241e+00
2.60333061e-01 -1.18891859e+00 -7.51438439e-01 -6.30031079e-02
-1.37130320e-02 -3.02383266e-02 7.19601631e-01 -5.35868704e-01
1.53629613e+00 -2.17689037e+00 4.06287223e-01 1.45355001e-01
-8.68131146e-02 5.28359711e-01 -3.18795025e-01 3.31399202e-01
-1.22196496e-01 -2.12820768e-01 -4.12572324e-01 7.47693237e-03
-2.22446434e-02 4.45627362e-01 -5.80190420e-01 5.11428297e-01
4.03559178e-01 1.32594597e+00 -1.11807907e+00 -1.14616543e-01
2.03225594e-02 5.30928254e-01 -5.14616907e-01 4.26855944e-02
-2.29029343e-01 3.24257821e-01 -1.85097337e-01 4.86507863e-01
-1.27650574e-01 -3.68129909e-01 1.45706713e-01 9.72125977e-02
3.30952369e-02 4.25937057e-01 -9.13320303e-01 1.72604764e+00
-2.97827810e-01 1.06455922e+00 -3.94667625e-01 -9.90818083e-01
9.87902582e-01 4.16864723e-01 3.24191689e-01 -1.07928169e+00
9.58800782e-03 3.21730554e-01 1.82810172e-01 -1.77724123e-01
6.85746193e-01 -2.89537787e-01 -3.31959009e-01 5.28484106e-01
-1.00974284e-01 -3.34355950e-01 7.93084577e-02 -1.36736691e-01
1.27274549e+00 1.37698516e-01 5.57301521e-01 -9.51341912e-02
2.69220650e-01 -3.40001166e-01 8.00874650e-01 4.79011118e-01
-3.14729214e-01 7.22655594e-01 3.10779691e-01 -7.21636713e-01
-9.00323451e-01 -1.17266428e+00 4.66847271e-01 1.08932126e+00
-6.44473583e-02 -6.96283802e-02 -5.92916049e-02 7.38101825e-02
-3.12034339e-01 5.43833375e-01 -4.73952472e-01 -5.58177412e-01
-1.14998591e+00 -8.77384126e-01 7.88922846e-01 9.14815366e-01
5.04785717e-01 -1.66987228e+00 -1.01795793e+00 7.68293917e-01
-2.23638862e-01 -8.71481121e-01 -4.60187078e-01 5.52659035e-01
-1.10974693e+00 -5.44083297e-01 -8.86769593e-01 -1.09915459e+00
1.74852952e-01 1.94801542e-03 1.05504644e+00 -2.35843241e-01
-3.30413282e-01 6.81430250e-02 -2.23467752e-01 -1.35250926e-01
-3.63048345e-01 7.11327791e-02 1.21178187e-01 -5.15601113e-02
-1.46340132e-01 -6.73726499e-01 -6.77886009e-01 1.89259067e-01
-1.01749074e+00 1.04668625e-01 5.48756540e-01 1.33856845e+00
6.94700599e-01 -3.90953988e-01 9.35415149e-01 -5.45242190e-01
7.65851021e-01 -2.98551917e-01 -2.46656820e-01 -3.25394608e-02
-1.02825515e-01 2.88478911e-01 7.38944650e-01 -7.33723760e-01
-8.40020955e-01 1.49786010e-01 -2.34446988e-01 -5.05697250e-01
3.06287467e-01 7.09016562e-01 3.31489921e-01 1.06223710e-01
7.34148204e-01 5.72546184e-01 6.89317435e-02 9.82175693e-02
2.62164801e-01 1.34671498e-02 7.15765595e-01 -2.39600018e-01
2.55497366e-01 4.30546820e-01 1.50664717e-01 -1.03769410e+00
-2.76108861e-01 -1.39589369e-01 -4.20444041e-01 -4.20355126e-02
7.37252533e-01 -9.77837503e-01 -5.14335275e-01 7.43297279e-01
-1.25638700e+00 -6.53136849e-01 -5.27659833e-01 7.61958718e-01
-1.11944366e+00 1.69734672e-01 -1.00455844e+00 -6.82642639e-01
-3.21234435e-01 -8.02761912e-01 6.23880327e-01 2.23701820e-01
-3.93014371e-01 -1.00430644e+00 2.82520443e-01 -5.12847662e-01
5.76619864e-01 5.25682688e-01 9.12600636e-01 -3.26138228e-01
-5.51812351e-01 -2.80562103e-01 1.42133892e-01 2.75921345e-01
-1.63597509e-01 -9.75596979e-02 -7.66285062e-01 -4.93332535e-01
-1.78725526e-01 -4.25722092e-01 1.49444413e+00 5.21256328e-01
1.04012346e+00 -3.61956470e-03 -2.01627046e-01 5.55116415e-01
1.00257218e+00 5.27250171e-01 1.01450360e+00 7.73775131e-02
6.20713174e-01 8.22882503e-02 2.27191120e-01 5.72092831e-01
1.33848980e-01 3.38614643e-01 2.67101109e-01 -6.68031350e-02
-1.12650283e-01 -2.89350092e-01 6.91658556e-01 1.66559565e+00
-4.25742745e-01 -3.93009394e-01 -9.45129275e-01 4.67729688e-01
-1.99252701e+00 -1.38468385e+00 3.05167019e-01 1.85693836e+00
6.20312631e-01 4.20340836e-01 1.97857455e-03 4.39344704e-01
4.98647928e-01 6.79528832e-01 -8.32167387e-01 -4.58246648e-01
-5.20268798e-01 4.50262308e-01 3.08194041e-01 5.21726012e-02
-1.10595965e+00 1.03963768e+00 7.84295368e+00 6.59134805e-01
-1.38201165e+00 -2.77744979e-01 5.38669765e-01 -9.09654722e-02
-9.40963402e-02 -6.07290268e-01 -3.89147490e-01 1.97396412e-01
1.34873939e+00 -1.68298453e-01 3.84841919e-01 4.30161804e-01
2.88066298e-01 3.26518327e-01 -1.14065123e+00 1.43811738e+00
-2.07005993e-01 -1.45332348e+00 2.83861190e-01 -2.62474984e-01
8.67995083e-01 2.00555608e-01 4.46821392e-01 5.65154195e-01
2.88624465e-01 -1.50840700e+00 7.13956475e-01 8.89303803e-01
6.95239186e-01 -6.59129441e-01 6.09113932e-01 1.95617989e-01
-1.47170854e+00 -3.28858286e-01 -2.21231788e-01 -3.85962665e-01
5.23208559e-01 2.31097475e-01 -5.23572743e-01 9.38185602e-02
3.74368787e-01 1.24164045e+00 -2.36485913e-01 1.04994369e+00
3.85913402e-02 8.36918473e-01 -1.83488786e-01 -3.73314828e-01
4.06483233e-01 -4.41144668e-02 4.42937881e-01 1.45692217e+00
4.98737216e-01 2.67923325e-01 5.07847741e-02 3.75941098e-01
-3.04962005e-02 -3.47754836e-01 -8.77141297e-01 -5.54599226e-01
2.64129996e-01 7.34675705e-01 -8.94502819e-01 -3.61711621e-01
-3.70214283e-01 9.80887055e-01 1.43229946e-01 6.09905779e-01
-7.95188069e-01 -4.82095033e-01 8.19472134e-01 -2.15251803e-01
4.98772442e-01 -8.02544832e-01 -7.30199888e-02 -1.12529576e+00
-1.07258148e-01 -8.53862405e-01 4.64237750e-01 -7.01747537e-01
-1.06366539e+00 8.65559161e-01 -4.74225938e-01 -1.73886991e+00
-8.73823345e-01 -7.27995098e-01 -5.89987636e-01 4.62388635e-01
-1.32544601e+00 -6.34958267e-01 8.25348720e-02 6.55324340e-01
5.10904431e-01 -2.64844477e-01 8.57717574e-01 2.91565180e-01
-5.09078562e-01 3.45280230e-01 2.09906265e-01 1.94977641e-01
2.04856738e-01 -8.48858595e-01 1.03154218e+00 7.59412110e-01
2.00457752e-01 5.25475681e-01 4.78797704e-01 -4.18204129e-01
-1.54834569e+00 -1.24750340e+00 6.05447888e-01 1.75598748e-02
7.66154528e-01 -1.78762138e-01 -1.01726627e+00 5.48403621e-01
1.92186814e-02 3.58110070e-01 4.63097274e-01 -4.44267005e-01
-3.64561498e-01 8.95560756e-02 -6.26446426e-01 8.48122418e-01
1.43061030e+00 -8.33778679e-01 -6.34788036e-01 -1.82379670e-02
6.81372106e-01 -4.08559144e-01 -5.93037724e-01 4.90876436e-01
6.83475196e-01 -6.51597142e-01 1.10006726e+00 -7.80959547e-01
1.62145317e-01 -2.55959809e-01 -3.61581862e-01 -1.54586565e+00
-6.84407949e-01 -9.41247940e-01 -6.24513865e-01 3.13592821e-01
5.02505422e-01 -7.33947754e-01 8.60536993e-01 3.74124362e-03
-3.09624642e-01 -8.48738670e-01 -1.22200072e+00 -9.00177717e-01
-6.78769872e-02 -4.81423467e-01 1.87573060e-01 6.82750762e-01
-8.08018609e-04 3.46027553e-01 -5.35409987e-01 -3.65207344e-01
1.30421951e-01 8.96286741e-02 1.77613467e-01 -1.13711214e+00
-1.70582414e-01 -8.66282463e-01 -8.30092847e-01 -1.48965240e+00
1.72400832e-01 -8.57641220e-01 2.69926131e-01 -1.41367507e+00
-4.15129334e-01 -4.85179983e-02 -6.28350198e-01 9.51700509e-02
4.88049462e-02 4.04869765e-01 1.73604831e-01 1.50897250e-01
-7.38085270e-01 1.01410902e+00 1.35392654e+00 -1.12727128e-01
-3.20284665e-01 1.28111467e-01 -1.26039118e-01 6.67310655e-01
7.60267794e-01 -6.55001029e-02 -3.91629457e-01 -1.87344268e-01
4.92825419e-01 5.08126497e-01 2.13532984e-01 -1.29402506e+00
3.36244315e-01 -6.98578879e-02 6.05462432e-01 -7.31466651e-01
6.71715140e-01 -2.87847102e-01 2.37080619e-01 9.74965394e-01
-5.81063926e-01 7.34800160e-01 5.69570698e-02 8.64558220e-01
-5.13298094e-01 3.44676912e-01 8.05685878e-01 -3.00660610e-01
-1.15864396e+00 3.99308413e-01 -8.95793259e-01 1.85123309e-01
7.19338655e-01 -4.49295878e-01 -1.42269671e-01 -6.06459260e-01
-9.04884398e-01 3.63236368e-02 -3.68024409e-02 5.57362199e-01
9.89126801e-01 -1.58713996e+00 -7.33036697e-01 4.19387639e-01
-2.56475270e-01 -1.37170255e-01 1.37340846e-02 7.21444070e-01
-6.99034810e-01 4.48906541e-01 -5.86129367e-01 -6.21932328e-01
-7.44228840e-01 4.41279501e-01 3.79348308e-01 -4.43979561e-01
-7.69061565e-01 7.25481391e-01 3.65809165e-02 -1.26836747e-01
3.27265680e-01 -8.61015975e-01 -3.17464978e-01 2.14121446e-01
5.89132905e-01 5.85189164e-01 -3.58270518e-02 -6.51210546e-01
-3.06745410e-01 3.38206649e-01 1.60672531e-01 -2.70565808e-01
1.32734704e+00 1.24530561e-01 1.16479054e-01 9.72406566e-01
1.19041872e+00 -8.25621843e-01 -1.26145506e+00 -1.95200369e-01
2.84070313e-01 3.22021507e-02 -2.11912736e-01 -2.85199642e-01
-9.70927536e-01 1.07726133e+00 5.86426020e-01 2.41001621e-01
1.21092629e+00 -5.42583406e-01 1.03345704e+00 7.76868701e-01
4.41255838e-01 -1.05676723e+00 5.76163411e-01 1.14250338e+00
9.89803374e-01 -9.28068757e-01 -2.18685642e-01 1.94172651e-01
-6.99379623e-01 1.21296537e+00 3.57045203e-01 -5.05293846e-01
7.05082774e-01 1.97290152e-01 -9.94629599e-03 1.31639704e-01
-1.38526595e+00 -2.14409754e-01 1.98303536e-01 7.09608674e-01
6.43571615e-01 2.22669408e-01 -1.05228156e-01 3.17942411e-01
-1.42106444e-01 4.19629551e-02 4.24716920e-01 1.10192311e+00
-4.57872808e-01 -5.23230314e-01 -6.80438755e-03 4.36106771e-01
-1.71151116e-01 -5.61348796e-02 2.15256102e-02 4.66596633e-01
-5.29579163e-01 4.65946764e-01 2.08244592e-01 -5.26423097e-01
2.33617663e-01 3.49647291e-02 4.01955426e-01 -3.16235244e-01
-6.50135696e-01 2.46020071e-02 1.30571857e-01 -7.32904136e-01
-4.42895591e-01 -5.95990300e-01 -1.59885502e+00 -2.25333750e-01
1.40223876e-01 -2.28981659e-01 3.56391937e-01 6.26328588e-01
2.88369656e-01 8.01100791e-01 4.10378963e-01 -1.38632262e+00
-4.66040224e-01 -7.99047709e-01 -5.15382767e-01 2.95362294e-01
6.64610744e-01 -5.20147979e-01 -1.23144537e-01 8.66828337e-02] | [7.488323211669922, 3.413301467895508] |
465f991e-d9df-4611-9fa6-a0301eb48971 | multi-accdoa-localizing-and-detecting | 2110.07124 | null | https://arxiv.org/abs/2110.07124v2 | https://arxiv.org/pdf/2110.07124v2.pdf | Multi-ACCDOA: Localizing and Detecting Overlapping Sounds from the Same Class with Auxiliary Duplicating Permutation Invariant Training | Sound event localization and detection (SELD) involves identifying the direction-of-arrival (DOA) and the event class. The SELD methods with a class-wise output format make the model predict activities of all sound event classes and corresponding locations. The class-wise methods can output activity-coupled Cartesian DOA (ACCDOA) vectors, which enable us to solve a SELD task with a single target using a single network. However, there is still a challenge in detecting the same event class from multiple locations. To overcome this problem while maintaining the advantages of the class-wise format, we extended ACCDOA to a multi one and proposed auxiliary duplicating permutation invariant training (ADPIT). The multi- ACCDOA format (a class- and track-wise output format) enables the model to solve the cases with overlaps from the same class. The class-wise ADPIT scheme enables each track of the multi-ACCDOA format to learn with the same target as the single-ACCDOA format. In evaluations with the DCASE 2021 Task 3 dataset, the model trained with the multi-ACCDOA format and with the class-wise ADPIT detects overlapping events from the same class while maintaining its performance in the other cases. Also, the proposed method performed comparably to state-of-the-art SELD methods with fewer parameters. | ['Yuki Mitsufuji', 'Emiru Tsunoo', 'Naoya Takahashi', 'Shusuke Takahashi', 'Yuichiro Koyama', 'Kazuki Shimada'] | 2021-10-14 | null | null | null | null | ['sound-event-localization-and-detection'] | ['audio'] | [ 1.26305178e-01 -3.19964290e-01 4.37569380e-01 6.89824671e-03
-1.17544472e+00 -6.16145015e-01 4.13255006e-01 3.46765220e-02
-3.83481801e-01 6.58379316e-01 2.49510370e-02 -1.86665341e-01
-6.30591750e-01 -5.75938106e-01 -9.05921936e-01 -9.11025047e-01
-3.43672872e-01 3.11609864e-01 7.34380305e-01 3.85026187e-01
8.91501084e-02 5.86089730e-01 -1.75678504e+00 7.66350508e-01
3.34358364e-01 1.26905215e+00 5.29335856e-01 7.72647679e-01
1.24578536e-01 6.50517523e-01 -8.57955098e-01 3.47156763e-01
2.06005558e-01 -5.70921123e-01 -4.73280877e-01 -6.83509350e-01
7.32362747e-01 1.50532159e-03 -2.62580365e-01 5.56220114e-01
7.79470146e-01 2.63623446e-01 6.02720499e-01 -1.52961147e+00
1.14884205e-01 4.23255682e-01 -4.23247397e-01 6.14894152e-01
3.54967594e-01 -2.83476889e-01 1.02911162e+00 -1.03686488e+00
3.61837000e-01 9.11180317e-01 9.51090693e-01 -3.77246290e-02
-8.27810347e-01 -8.32352757e-01 1.71405345e-01 4.39363420e-01
-1.53222680e+00 -9.47294459e-02 8.42619896e-01 -3.22861999e-01
1.13988864e+00 4.86157417e-01 5.15943944e-01 1.08097363e+00
2.51098365e-01 6.84954703e-01 9.01255310e-01 -4.35760707e-01
3.74880642e-01 -3.41354400e-01 1.01885110e-01 4.06181306e-01
-8.95942375e-02 3.52984183e-02 -9.84491587e-01 -1.29145533e-01
6.70755684e-01 1.43720340e-02 -2.18436390e-01 2.89594824e-03
-1.51045084e+00 5.85786343e-01 1.44629195e-01 5.38856387e-01
-5.63550472e-01 2.02914268e-01 4.37719941e-01 1.10305950e-01
3.33270639e-01 5.60845733e-01 -7.34014392e-01 -3.56918275e-01
-1.19197583e+00 1.99048519e-01 7.94323683e-01 6.97015107e-01
4.80348438e-01 3.11265022e-01 -5.71429193e-01 8.11107337e-01
-4.56754584e-03 5.95236897e-01 4.10615563e-01 -6.72241747e-01
6.16187274e-01 2.40813494e-01 3.87754962e-02 -9.86283958e-01
-8.83744180e-01 -1.00666368e+00 -5.24916291e-01 1.26782283e-01
6.40589058e-01 -2.88109839e-01 -1.05674970e+00 1.88533020e+00
1.90510795e-01 7.97946334e-01 -4.49001566e-02 7.49087334e-01
5.99856555e-01 1.09543765e+00 -2.63988495e-01 -3.66290808e-01
1.46614051e+00 -8.66202533e-01 -6.70803905e-01 -9.60034654e-02
5.52489579e-01 -8.54138851e-01 7.22436130e-01 6.21483624e-01
-7.38673210e-01 -7.01941192e-01 -9.72951949e-01 4.44630325e-01
-3.98522705e-01 6.79539561e-01 5.15238106e-01 3.32690597e-01
-9.01721179e-01 2.46067092e-01 -8.70871782e-01 -2.65718311e-01
6.08531237e-02 2.38399327e-01 -4.63071972e-01 2.85979271e-01
-1.15309250e+00 5.44702828e-01 4.33913529e-01 3.54523212e-02
-1.04098451e+00 -1.06695735e+00 -6.36923671e-01 3.35207939e-01
2.87778348e-01 -1.41930729e-01 1.14632797e+00 -4.22350794e-01
-1.21062505e+00 1.18381016e-01 -2.72028834e-01 -5.94204545e-01
2.14661554e-01 -2.70713270e-01 -6.84345663e-01 3.28762740e-01
4.61488575e-01 4.91559625e-01 7.82048643e-01 -8.74601424e-01
-1.14248729e+00 -5.99844903e-02 -2.73177117e-01 -7.21256342e-03
-1.37722388e-01 1.05214126e-01 -4.32862863e-02 -7.93506801e-01
5.06151676e-01 -7.92186499e-01 3.97998244e-01 -1.78787410e-01
-4.73657161e-01 -4.39198971e-01 1.11034083e+00 -3.74924928e-01
1.38745117e+00 -2.39802051e+00 -1.51768893e-01 4.32131588e-02
-4.85825241e-02 1.67952582e-01 -2.02538729e-01 6.59043431e-01
-3.93606693e-01 -3.36351782e-01 4.62024137e-02 -4.32551295e-01
-1.70782924e-01 6.78641647e-02 -4.94748980e-01 2.77563751e-01
2.85193264e-01 2.58738816e-01 -8.19396675e-01 -2.03860015e-01
1.37946069e-01 3.00966710e-01 -6.21912420e-01 1.47124782e-01
1.04485601e-02 3.12801868e-01 -3.42521608e-01 4.49763894e-01
6.82623565e-01 1.45315319e-01 -8.62168819e-02 -6.30138397e-01
-5.07670343e-01 5.97840667e-01 -1.85046017e+00 1.43740177e+00
-5.95898032e-01 8.29186976e-01 -2.35268250e-01 -1.04925871e+00
8.19047093e-01 7.41384983e-01 6.44045591e-01 -6.23239815e-01
-2.48538464e-01 4.52961445e-01 2.86304116e-01 -3.93349618e-01
3.11936855e-01 1.05437882e-01 -2.32652575e-01 5.06711125e-01
4.14495081e-01 2.59292662e-01 1.19961761e-01 3.12189311e-02
1.38429832e+00 1.10696428e-01 7.01136142e-02 -9.68370885e-02
1.38464361e-01 -3.43471885e-01 6.31518304e-01 1.28683829e+00
3.23717520e-02 1.09157574e+00 4.21671689e-01 -5.06191134e-01
-5.50249934e-01 -1.28455913e+00 -3.26228708e-01 1.01840699e+00
-6.47130385e-02 -3.61731589e-01 -3.59027147e-01 -7.80087769e-01
-1.41370729e-01 9.02549684e-01 -5.21893442e-01 6.79911301e-02
-7.48415947e-01 -6.43696606e-01 1.01325214e+00 7.11028159e-01
4.67804402e-01 -1.09502316e+00 -6.99393809e-01 4.60271358e-01
-4.24282968e-01 -1.13720930e+00 -5.07802546e-01 9.38180625e-01
-2.88249850e-01 -7.75320470e-01 -5.74487209e-01 -8.09322000e-01
3.17040354e-01 2.03843951e-01 6.53576314e-01 -4.66268927e-01
-2.60036618e-01 3.25543851e-01 -4.80056047e-01 -6.62257850e-01
7.26983994e-02 -3.69235985e-02 1.51726380e-01 4.59096760e-01
1.28480092e-01 -7.99211621e-01 -3.89369726e-01 4.17702407e-01
-7.26629257e-01 -1.92161992e-01 3.38955998e-01 7.20868170e-01
6.45688295e-01 2.00423837e-01 1.17675710e+00 -3.44722867e-01
3.45512480e-01 -7.22072065e-01 -2.53815979e-01 -2.28614407e-03
-2.74903268e-01 -1.86026990e-01 7.46911883e-01 -7.66389549e-01
-6.66847229e-01 3.40767175e-01 -2.61050463e-01 -5.42453170e-01
-3.73127908e-01 4.33909684e-01 -3.46264429e-03 3.91912043e-01
6.19785786e-01 4.05127615e-01 -4.34062868e-01 -5.85224032e-01
-9.57873091e-02 5.93197823e-01 5.49753547e-01 -3.59171629e-01
4.37621891e-01 2.39361599e-01 6.96389899e-02 -6.85111165e-01
-8.84221375e-01 -6.48336351e-01 -4.70578581e-01 -1.65987566e-01
1.03515291e+00 -9.43855464e-01 -2.30733097e-01 8.60014558e-01
-1.39854062e+00 1.13882549e-01 -3.65176201e-01 8.83967221e-01
-2.35721961e-01 -6.48999121e-03 -2.11612329e-01 -8.39380205e-01
-5.57973273e-02 -9.41367805e-01 1.23199213e+00 3.09772432e-01
-2.57790953e-01 -4.34471279e-01 1.30680591e-01 -2.63635010e-01
1.83628663e-01 1.20478809e-01 8.19951117e-01 -1.25717962e+00
-5.03128111e-01 -3.26189816e-01 7.41384923e-02 1.14609949e-01
7.81961605e-02 -3.34563673e-01 -1.19120681e+00 4.47533699e-03
-4.25373344e-03 1.87983021e-01 6.94309354e-01 5.28822064e-01
9.87407804e-01 -9.14781243e-02 -3.64452541e-01 4.25628006e-01
9.32471335e-01 7.32707858e-01 4.39514488e-01 2.88652450e-01
5.18499374e-01 1.45244319e-02 7.07291424e-01 4.76047665e-01
3.78121644e-01 1.13464332e+00 5.01693130e-01 -1.07276239e-01
-4.76314217e-01 -2.73311794e-01 2.50352651e-01 6.52449548e-01
2.59404778e-01 -5.99076331e-01 -7.13557422e-01 8.14680934e-01
-1.70178413e+00 -1.29568839e+00 -4.21016335e-01 2.19593215e+00
7.10219204e-01 -2.45787529e-03 3.06307003e-02 6.35588527e-01
6.76565230e-01 1.18527316e-01 -1.53895512e-01 -2.92744726e-01
-1.37278467e-01 4.15835023e-01 8.43702555e-02 1.52459741e-01
-1.22145009e+00 4.27667648e-01 5.72334385e+00 1.03741121e+00
-1.33823621e+00 3.50369632e-01 3.88951302e-02 -2.09346205e-01
1.40441194e-01 -2.02352077e-01 -1.16166627e+00 7.91610360e-01
9.98015821e-01 4.00446862e-01 1.47535115e-01 5.05929768e-01
3.61191690e-01 -3.90909374e-01 -1.29580760e+00 9.94779944e-01
1.48333341e-01 -1.21014059e+00 -2.41933167e-01 -1.55527040e-01
5.68734586e-01 -1.29198311e-02 -1.76694542e-01 5.01593530e-01
-4.76044983e-01 -4.53163147e-01 1.17008758e+00 4.41921741e-01
5.87250769e-01 -5.67413211e-01 6.89977527e-01 5.20375073e-01
-1.53557122e+00 -3.61083478e-01 -8.72549489e-02 -1.81343272e-01
3.69529694e-01 6.63949788e-01 -9.70395744e-01 7.77458131e-01
1.03445888e+00 5.13700366e-01 -4.27091271e-01 1.51319492e+00
-6.27254844e-02 1.15909362e+00 -8.58436346e-01 6.44473433e-02
2.41216287e-01 2.55249470e-01 1.00999427e+00 1.26951742e+00
8.87191892e-01 -2.59216040e-01 1.19993247e-01 5.76370955e-01
1.26877189e-01 -2.33953431e-01 -4.34699297e-01 2.71999210e-01
9.16932821e-01 9.17377591e-01 -8.58221531e-01 -8.72715265e-02
-2.66823143e-01 7.81049669e-01 -1.41177503e-02 3.06414485e-01
-1.05256689e+00 -7.89361477e-01 4.02619243e-01 1.80847496e-01
6.94219470e-01 -4.76153195e-02 -2.27843925e-01 -5.51530778e-01
2.67461538e-01 -6.22448027e-01 5.54579198e-01 -9.75342214e-01
-9.81708646e-01 7.77497888e-01 3.21663707e-01 -1.72845984e+00
-2.09714979e-01 -3.64413768e-01 -9.21914041e-01 7.96520054e-01
-1.27515817e+00 -1.14661455e+00 -7.88125396e-02 5.35385847e-01
6.99599326e-01 -2.27891937e-01 9.04585123e-01 7.65645385e-01
-5.26749074e-01 8.29579294e-01 -3.87841314e-02 -1.00393035e-02
8.10561001e-01 -1.16552341e+00 -1.29803464e-01 1.06151330e+00
5.82987964e-01 2.25470275e-01 5.24181664e-01 -5.41424692e-01
-1.07982981e+00 -1.19522333e+00 1.15270209e+00 -2.58855790e-01
3.79522145e-01 -3.76469672e-01 -7.32555509e-01 5.87085307e-01
-8.95315707e-02 2.62202233e-01 2.88119346e-01 7.95557275e-02
-1.57926515e-01 -3.95151556e-01 -9.39342499e-01 3.45211513e-02
8.37057531e-01 -6.51774585e-01 -6.59733415e-01 3.62647384e-01
6.56623602e-01 -5.14120221e-01 -5.49352765e-01 3.40557635e-01
4.40553069e-01 -7.26769209e-01 9.44368303e-01 -1.68490335e-01
1.44296542e-01 -8.94179344e-01 -1.72134891e-01 -1.19222212e+00
-2.41758689e-01 -3.68929595e-01 -2.87150025e-01 1.38057268e+00
4.44079459e-01 -7.13390291e-01 1.12727083e-01 -2.92378396e-01
-7.43523359e-01 -5.64614654e-01 -1.60569549e+00 -1.03171396e+00
-3.88834864e-01 -7.81344533e-01 3.94921958e-01 6.94113493e-01
-4.13784593e-01 2.10201561e-01 -5.75987637e-01 8.16246331e-01
1.13103479e-01 8.20042267e-02 2.75400698e-01 -1.26304626e+00
-4.77190822e-01 -4.34520952e-02 -4.89540964e-01 -9.86274004e-01
-2.73928881e-01 -9.73471880e-01 3.47272724e-01 -1.52446866e+00
-1.62113562e-01 -5.04386008e-01 -6.82404459e-01 9.30508196e-01
-2.70674489e-02 4.03491408e-01 9.30684358e-02 3.73106971e-02
-4.31418538e-01 1.37302816e-01 7.79132426e-01 5.39919436e-02
-3.94872844e-01 4.32915807e-01 -2.59227484e-01 6.28055632e-01
5.79970717e-01 -1.14951169e+00 -4.83222812e-01 -4.16007310e-01
1.54109731e-01 1.49762452e-01 6.46557748e-01 -1.66729081e+00
5.27557254e-01 3.41010153e-01 4.29778337e-01 -1.23937178e+00
6.22878730e-01 -8.36863399e-01 3.67239654e-01 1.92544192e-01
-2.27677390e-01 4.96970601e-02 4.88310874e-01 6.16917431e-01
-3.83892626e-01 -3.13959390e-01 5.15388668e-01 3.48303109e-01
-6.28672361e-01 -1.51680976e-01 -6.21966124e-01 4.88135405e-02
1.05853665e+00 -2.74066150e-01 -3.15882176e-01 -3.11685026e-01
-7.83091962e-01 -1.66685745e-01 -6.19584560e-01 5.84938407e-01
3.37527663e-01 -1.28071856e+00 -5.90021193e-01 3.66107494e-01
-1.15477324e-01 -1.40774712e-01 4.58688825e-01 1.04484499e+00
-9.60896388e-02 4.84280914e-01 -1.32389218e-01 -9.13388491e-01
-1.03538692e+00 5.32004908e-02 5.56780219e-01 -3.32836688e-01
-6.42435670e-01 9.30367708e-01 3.04780543e-01 -3.70923907e-01
4.31015193e-01 -4.61208403e-01 -3.41607034e-01 3.33876967e-01
4.92269963e-01 4.79914039e-01 4.38753486e-01 -4.37165141e-01
-6.66192889e-01 5.13677835e-01 1.91022113e-01 -2.17013165e-01
1.41166162e+00 1.43866345e-01 3.05906296e-01 4.31645960e-01
1.17975509e+00 3.30426693e-01 -1.06051540e+00 9.25193354e-02
-2.74082750e-01 -1.31081983e-01 1.54240981e-01 -1.24192286e+00
-1.01620293e+00 9.90171790e-01 9.03522611e-01 2.77400494e-01
1.30092144e+00 3.64108896e-03 5.81943989e-01 1.83314875e-01
4.39352393e-01 -7.58743346e-01 1.09631969e-02 7.35159755e-01
8.32239807e-01 -5.51015854e-01 -3.99864167e-01 -2.00491905e-01
-4.83746469e-01 1.05456078e+00 4.71847206e-01 8.15268680e-02
7.98042834e-01 4.88416046e-01 2.21853908e-02 -1.11975476e-01
-5.80131531e-01 -3.78237776e-02 3.02445978e-01 6.27826333e-01
1.61286641e-03 -7.21143335e-02 -3.47938240e-01 1.03875136e+00
-8.00141990e-02 -1.61150441e-01 3.21507663e-01 9.71072316e-01
-1.49520427e-01 -1.01825309e+00 -4.28727686e-01 5.40799379e-01
-5.54255307e-01 -1.72691137e-01 2.67326444e-01 7.15895474e-01
4.42415148e-01 1.05192113e+00 3.48398536e-01 -5.43552697e-01
4.86178011e-01 4.14071023e-01 2.98487600e-02 -6.35897279e-01
-9.10782635e-01 3.34715694e-01 1.98023945e-01 -5.06811798e-01
-2.02540550e-02 -9.73409832e-01 -1.39558136e+00 6.11192644e-01
-4.90710765e-01 2.68636614e-01 5.69194138e-01 9.86342549e-01
7.50073135e-01 1.06261194e+00 7.56815791e-01 -7.87948847e-01
-2.83684671e-01 -1.11797988e+00 -6.68739855e-01 2.46205889e-02
4.62054908e-01 -8.82159710e-01 -5.20189762e-01 -1.61755213e-03] | [15.192118644714355, 5.249176025390625] |
857610b6-eb7b-41c1-ad17-4206c602824c | correlative-information-maximization-a | 2306.04810 | null | https://arxiv.org/abs/2306.04810v2 | https://arxiv.org/pdf/2306.04810v2.pdf | Correlative Information Maximization: A Biologically Plausible Approach to Supervised Deep Neural Networks without Weight Symmetry | The backpropagation algorithm has experienced remarkable success in training large-scale artificial neural networks, however, its biological-plausibility is disputed, and it remains an open question whether the brain employs supervised learning mechanisms akin to it. Here, we propose correlative information maximization between layer activations as an alternative normative approach to describe the signal propagation in biological neural networks in both forward and backward directions. This new framework addresses many concerns about the biological-plausibility of conventional artificial neural networks and the backpropagation algorithm. The coordinate descent-based optimization of the corresponding objective, combined with the mean square error loss function for fitting labeled supervision data, gives rise to a neural network structure that emulates a more biologically realistic network of multi-compartment pyramidal neurons with dendritic processing and lateral inhibitory neurons. Furthermore, our approach provides a natural resolution to the weight symmetry problem between forward and backward signal propagation paths, a significant critique against the plausibility of the conventional backpropagation algorithm. This is achieved by leveraging two alternative, yet equivalent forms of the correlative mutual information objective. These alternatives intrinsically lead to forward and backward prediction networks without weight symmetry issues, providing a compelling solution to this long-standing challenge. | ['Alper T Erdogan', 'Cengiz Pehlevan', 'Bariscan Bozkurt'] | 2023-06-07 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 5.23698211e-01 4.77035671e-01 8.73850733e-02 -5.53084314e-01
1.50315925e-01 -1.09948039e-01 8.17193210e-01 2.44466104e-02
-8.58368158e-01 8.66519034e-01 -4.83580772e-03 -6.44238412e-01
-5.90027094e-01 -5.85196078e-01 -5.99966288e-01 -1.04152870e+00
-1.04827240e-01 2.71185040e-01 1.63864657e-01 -2.71228075e-01
4.70791429e-01 6.21002972e-01 -1.16923630e+00 4.89280559e-02
6.23355389e-01 1.12800467e+00 1.48024157e-01 5.23791611e-01
-8.32448900e-02 1.10280204e+00 -1.02110445e-01 -6.22316360e-01
1.39081432e-03 -4.90288407e-01 -1.08040547e+00 -2.59270161e-01
7.49813542e-02 1.95374146e-01 -1.94795787e-01 9.24087346e-01
2.45769516e-01 -6.53817654e-02 8.17568064e-01 -1.04166067e+00
-6.67588234e-01 6.08389020e-01 -2.98532277e-01 5.13175786e-01
-2.64922976e-01 7.10441172e-02 1.15324616e+00 -8.90940428e-01
4.73334134e-01 9.38569248e-01 1.03412700e+00 5.98990679e-01
-1.66026330e+00 -3.24232340e-01 1.69033423e-01 3.77908647e-02
-1.10724211e+00 -3.49967092e-01 6.20050609e-01 -6.10381305e-01
1.07792354e+00 1.24361031e-01 8.96900713e-01 9.60553944e-01
3.54954273e-01 5.31084299e-01 1.02816474e+00 -5.80353022e-01
1.79996178e-01 2.43320584e-01 3.19146067e-01 6.67846560e-01
2.29862049e-01 3.87651414e-01 -6.35219514e-01 -3.77710089e-02
7.21207201e-01 -7.02997968e-02 -4.86865431e-01 -1.48197800e-01
-1.27998078e+00 6.36364639e-01 6.67907774e-01 5.34852147e-01
-3.91125321e-01 1.59858480e-01 1.25251919e-01 3.22825283e-01
3.18688065e-01 7.03521967e-01 -5.68901241e-01 3.71506661e-01
-1.08242893e+00 1.58802807e-01 8.10624778e-01 2.33599275e-01
7.17700422e-01 1.76291585e-01 3.22012961e-01 7.26394117e-01
9.61503744e-01 7.10296929e-02 6.00763500e-01 -1.14135921e+00
2.09359631e-01 4.39486861e-01 -3.55618060e-01 -1.22461998e+00
-6.80836618e-01 -8.16476047e-01 -9.93744195e-01 6.38054490e-01
7.65723348e-01 -2.46682435e-01 -3.28977048e-01 1.96039391e+00
5.47813289e-02 -1.35747880e-01 6.86677545e-02 8.63897681e-01
2.02499956e-01 4.13801342e-01 3.65951508e-02 -2.53568530e-01
9.06179965e-01 -8.06174636e-01 -5.26320696e-01 -4.92052466e-01
5.83589196e-01 -2.90280640e-01 5.60584009e-01 2.97155261e-01
-1.29125750e+00 -1.63537189e-01 -1.28186047e+00 5.13549661e-03
-4.25561368e-01 -3.41874421e-01 8.68212342e-01 4.87101525e-01
-1.22118068e+00 1.19342971e+00 -9.46676791e-01 -2.36861825e-01
5.27100503e-01 5.36537945e-01 -4.57905293e-01 4.37770933e-01
-1.00949585e+00 1.27759969e+00 3.97894949e-01 4.62536961e-01
-1.97420299e-01 -6.90214396e-01 -5.20910680e-01 1.11381486e-01
-2.97937244e-01 -9.44653630e-01 9.83579874e-01 -1.14440715e+00
-1.42405772e+00 1.05003929e+00 -1.89412460e-01 -7.40055144e-01
3.45990390e-01 2.00537771e-01 -1.78551137e-01 -8.03698599e-02
-2.97861069e-01 9.47847068e-01 4.29862052e-01 -1.03630674e+00
-3.35338026e-01 -6.48805797e-01 -3.86857361e-01 1.96529537e-01
-3.91129583e-01 7.38405297e-03 2.44509622e-01 -6.62108839e-01
5.17611563e-01 -7.35002697e-01 -2.06528023e-01 5.04347384e-01
-2.98568726e-01 1.14683077e-01 1.33304879e-01 -2.14239240e-01
1.02695858e+00 -2.01515198e+00 1.32638097e-01 3.84365499e-01
3.29274207e-01 3.62539478e-02 -1.58988368e-02 2.15634450e-01
-4.67762232e-01 8.80857632e-02 -7.67062366e-01 -2.50263959e-01
-1.45604178e-01 2.87840515e-01 -2.93936521e-01 6.42599225e-01
3.86616498e-01 8.79352033e-01 -7.63505638e-01 -3.24101210e-01
-4.11284417e-01 7.50611007e-01 -3.51952344e-01 -1.34386420e-01
2.11301193e-01 3.46395761e-01 -3.18577103e-02 1.20025940e-01
3.72983515e-01 -4.44943041e-01 2.08596483e-01 6.45158738e-02
-3.65765601e-01 4.92640972e-01 -8.11277628e-01 1.49346757e+00
4.43969145e-02 1.06996953e+00 1.32621005e-01 -1.42770040e+00
9.11948621e-01 2.48305216e-01 3.07190657e-01 -6.31716371e-01
1.42427310e-01 4.50241804e-01 3.79645020e-01 -1.76773384e-01
-6.04354441e-02 -6.44115031e-01 7.65401304e-01 5.37995338e-01
4.30890083e-01 6.69903457e-02 6.21816628e-02 -1.15448304e-01
1.07022774e+00 1.87919453e-01 9.84502118e-03 -5.78793049e-01
5.53359032e-01 -6.26548752e-02 4.77941811e-01 5.29376030e-01
-2.63954759e-01 6.85669065e-01 4.28453088e-01 -4.12157029e-01
-9.27246094e-01 -1.01786351e+00 -4.72327650e-01 7.60069549e-01
-2.93489635e-01 6.60985485e-02 -7.62168050e-01 -1.43437982e-01
-1.15443751e-01 4.19590503e-01 -7.74906814e-01 -3.55504751e-01
-5.00568271e-01 -1.19357908e+00 9.63727355e-01 2.77154356e-01
3.55332315e-01 -1.00799811e+00 -7.49033511e-01 2.88103193e-01
7.71325156e-02 -7.08126545e-01 1.94063827e-01 8.13895941e-01
-1.26717627e+00 -1.03893077e+00 -9.18250620e-01 -8.58570874e-01
7.79265821e-01 -1.83387950e-01 9.29530025e-01 1.06667235e-01
-2.02626660e-01 3.37849371e-02 2.65474349e-01 -4.11167622e-01
-3.81425619e-01 -8.12887177e-02 1.47216961e-01 -5.69534898e-02
3.30419660e-01 -9.34419453e-01 -6.42318189e-01 2.37175599e-01
-7.75944233e-01 4.80216295e-02 5.14372408e-01 9.70584750e-01
4.68652129e-01 -2.30848283e-01 6.95112169e-01 -8.93718600e-01
7.51550198e-01 -5.47152936e-01 -3.22826266e-01 1.64880887e-01
-1.10894525e+00 2.77018875e-01 4.58575279e-01 -1.63830131e-01
-1.08049512e+00 -2.49891520e-01 -4.39724445e-01 1.28617525e-01
-1.16783865e-01 6.37544274e-01 3.95225167e-01 -3.39302003e-01
8.93372297e-01 4.02280271e-01 2.95490146e-01 -3.17434192e-01
1.89613149e-01 1.30570754e-01 6.60021067e-01 -2.44767800e-01
4.91796434e-01 6.38280690e-01 1.69705406e-01 -7.11888850e-01
-6.72088981e-01 -1.34731963e-01 -7.48399079e-01 -1.99914843e-01
8.33133101e-01 -3.93897951e-01 -8.05779040e-01 5.21045923e-01
-1.38476121e+00 -4.01238263e-01 -4.11017001e-01 7.16074049e-01
-9.00413394e-01 1.89291954e-01 -8.41269970e-01 -8.11348200e-01
-2.93606550e-01 -7.84659445e-01 2.38339081e-01 -4.45374846e-02
-3.91140044e-01 -1.52336502e+00 2.26003096e-01 5.93764856e-02
6.71214700e-01 1.30721360e-01 1.07211232e+00 -7.61123598e-01
4.38543074e-02 -1.68445446e-02 -2.53601611e-01 4.87966448e-01
-2.42464617e-01 1.56827539e-01 -1.17818630e+00 1.26496300e-01
5.21037996e-01 -2.26064369e-01 1.06575036e+00 4.07386333e-01
8.13013256e-01 -4.14117098e-01 -2.07055166e-01 7.12005377e-01
1.33019793e+00 8.45458508e-02 4.26680148e-01 3.76772702e-01
2.48191670e-01 1.03716457e+00 -2.41105467e-01 7.34497607e-02
1.88847765e-01 7.62290210e-02 3.96158963e-01 -2.09657792e-02
-2.76502483e-02 -1.05457515e-01 2.04585999e-01 1.10180378e+00
-2.55544692e-01 8.47012177e-02 -1.05605638e+00 3.36534470e-01
-1.72974694e+00 -1.24369371e+00 -3.38037223e-01 2.11527848e+00
9.62840676e-01 5.39925754e-01 -1.81949988e-01 6.16946459e-01
4.87835020e-01 -4.77980450e-02 -4.85210180e-01 -4.78365749e-01
-5.15077651e-01 -3.50159034e-03 4.24271792e-01 4.15780038e-01
-7.16976523e-01 3.24405253e-01 7.96915293e+00 4.61021245e-01
-1.25354826e+00 -6.18371665e-02 7.50462234e-01 8.01421031e-02
-1.74385861e-01 -6.15876429e-02 -6.33272946e-01 3.05452317e-01
1.20229602e+00 -5.65717518e-02 4.80203718e-01 3.70539755e-01
2.00869560e-01 9.11621191e-03 -1.27354252e+00 7.66466737e-01
-8.77096429e-02 -1.66318691e+00 -2.03410372e-01 8.09471384e-02
5.18789411e-01 5.56619108e-01 1.62579030e-01 -1.31402254e-01
2.03421459e-01 -9.99227464e-01 8.35923254e-01 7.09295154e-01
3.30711275e-01 -3.01886946e-01 4.01094645e-01 7.03857422e-01
-5.36859095e-01 -2.11380944e-01 -3.32976550e-01 -3.71197999e-01
1.67371139e-01 7.56659627e-01 -4.10572708e-01 -1.77951783e-01
4.93780464e-01 5.05076408e-01 -3.30924511e-01 1.13762736e+00
-1.54668182e-01 6.26062453e-01 -4.42738116e-01 6.66576717e-03
3.22523803e-01 -2.77943552e-01 4.80071157e-01 1.31506217e+00
2.73951441e-02 -1.87263772e-01 -7.28570223e-01 1.30121326e+00
-1.20345913e-02 -1.01585351e-02 -6.88828290e-01 -1.28722757e-01
1.79411024e-01 1.06055641e+00 -8.10065806e-01 9.02860761e-02
-2.95061141e-01 7.51622975e-01 6.81030273e-01 5.02166510e-01
-4.48368102e-01 -4.09374505e-01 4.46024299e-01 1.48731217e-01
-1.44225881e-01 -1.62720114e-01 -7.68402696e-01 -8.55412245e-01
6.92775026e-02 -2.95623511e-01 6.83725253e-02 -6.87927365e-01
-1.28651321e+00 5.34441471e-01 -4.01299417e-01 -6.19619966e-01
-6.77731782e-02 -1.05469716e+00 -8.32562566e-01 9.52483594e-01
-1.77951014e+00 -6.10792398e-01 2.78196692e-01 4.35397655e-01
8.35368708e-02 1.90486591e-02 9.48511779e-01 3.78183424e-01
-5.28217852e-01 4.12746906e-01 3.51810217e-01 5.82968146e-02
2.53844351e-01 -1.33952904e+00 1.55031741e-01 5.10123610e-01
1.29502520e-01 8.80681038e-01 7.28909671e-01 -1.07247926e-01
-9.62436736e-01 -8.70488405e-01 1.19442225e+00 -2.06118271e-01
9.58444118e-01 -9.41651389e-02 -1.17114592e+00 5.77396154e-01
1.30210608e-01 3.83461677e-02 9.64539111e-01 9.36512426e-02
-4.13806349e-01 7.95113444e-02 -1.02463484e+00 4.94229734e-01
9.16885436e-01 -3.95964354e-01 -8.34806204e-01 2.74761289e-01
3.93223315e-01 3.68062228e-01 -9.04228628e-01 2.38062397e-01
6.64174139e-01 -1.04388523e+00 8.03508878e-01 -7.79450476e-01
6.00755692e-01 -2.84548607e-02 -7.74561539e-02 -1.25220025e+00
-4.54725891e-01 -5.02809465e-01 2.32764613e-02 8.23506296e-01
1.00632811e+00 -9.46222484e-01 8.34296048e-01 7.76283920e-01
-1.89218089e-01 -1.23409498e+00 -9.51693952e-01 -6.52259290e-01
6.05300784e-01 -5.65308750e-01 -1.56732827e-01 1.04394817e+00
5.89541852e-01 4.37091619e-01 1.60783470e-01 -3.36995959e-01
6.68374240e-01 -4.41565156e-01 -3.45741212e-02 -1.50635898e+00
-4.02925730e-01 -1.07652295e+00 -4.07091886e-01 -1.04212439e+00
2.18554243e-01 -1.24468064e+00 4.37072106e-02 -1.46142268e+00
-7.40464181e-02 -4.29601699e-01 -5.26413977e-01 3.78451705e-01
3.25839579e-01 2.72542268e-01 4.66302708e-02 5.55052817e-01
-1.40668556e-01 2.37706378e-01 1.02592254e+00 -8.16851184e-02
4.13660184e-02 1.17490422e-02 -6.41921997e-01 1.25372732e+00
8.26931357e-01 -6.31626070e-01 -3.11388940e-01 -7.68735707e-01
8.24975014e-01 -2.43791953e-01 3.89271647e-01 -1.02214921e+00
7.36986637e-01 6.95587844e-02 4.07025009e-01 -2.85776090e-02
4.10480559e-01 -7.61831462e-01 -7.21831396e-02 6.26335979e-01
-8.27799141e-01 1.53125199e-02 -3.37662287e-02 4.79074925e-01
-2.60100722e-01 -5.00421047e-01 1.02681577e+00 -2.45625168e-01
-2.55227238e-01 8.26964378e-02 -8.25501978e-01 -2.68163402e-02
6.95942998e-01 -5.50531626e-01 -3.28103572e-01 1.48757190e-01
-1.11225986e+00 7.17154220e-02 1.76697493e-01 -3.05710863e-02
7.13150501e-01 -1.03565264e+00 -5.95092297e-01 2.98308700e-01
-1.54467851e-01 -3.14010650e-01 -7.85364807e-02 1.33697081e+00
-6.18560135e-01 4.88405645e-01 -4.39593256e-01 -3.58265281e-01
-7.30309844e-01 3.01085919e-01 8.59976351e-01 -1.24527112e-01
-4.03655440e-01 1.25070560e+00 7.26297125e-02 -4.82392102e-01
1.91816539e-01 -1.25112951e-01 -3.66338164e-01 -3.61636132e-02
4.70024973e-01 3.41393501e-01 6.48664981e-02 -5.85537910e-01
-2.87858725e-01 2.38176167e-01 4.38918583e-02 -2.84663111e-01
1.42466795e+00 -1.44292295e-01 -4.02610451e-01 8.40391159e-01
1.27467394e+00 -5.44365883e-01 -1.12990654e+00 -2.73883283e-01
4.53461647e-01 1.56387702e-01 1.34479091e-01 -8.77692342e-01
-9.93269801e-01 1.28342569e+00 3.28677952e-01 3.88007462e-01
8.04370105e-01 -3.62620860e-01 3.90707225e-01 7.44842529e-01
-1.91836674e-02 -1.07915735e+00 -2.60141909e-01 6.71907723e-01
7.63248205e-01 -1.16807342e+00 -1.24310389e-01 3.20521221e-02
-1.54194325e-01 1.33462334e+00 4.02302802e-01 -2.05978721e-01
1.01893151e+00 3.94047141e-01 -2.00512987e-02 -3.09371114e-01
-1.04887104e+00 1.96208268e-01 6.56552985e-02 6.02992177e-01
8.28827202e-01 -2.69645602e-01 -4.96245444e-01 3.59124959e-01
-1.95485085e-01 -2.96179373e-02 1.99831545e-01 8.26846361e-01
-6.97797120e-01 -7.39094198e-01 -1.73199013e-01 2.70888597e-01
-6.87690318e-01 -2.21881270e-01 -3.70102078e-01 5.12685418e-01
1.26348408e-02 7.26314127e-01 1.32051155e-01 -1.31679982e-01
-6.54470250e-02 2.88093686e-01 5.40541768e-01 -3.91365021e-01
-5.74011564e-01 -2.43935436e-01 -2.83150136e-01 -2.04736918e-01
-5.69326639e-01 -4.09721613e-01 -1.51718342e+00 -1.94544613e-01
-4.15974796e-01 1.19975157e-01 1.07864439e+00 1.13360012e+00
1.89831704e-01 4.22102511e-01 3.08390886e-01 -9.42658365e-01
-7.44999290e-01 -7.54666448e-01 -5.77789545e-01 2.97496438e-01
3.32334489e-01 -4.36872333e-01 -5.53755343e-01 7.87210241e-02] | [8.082579612731934, 3.3808465003967285] |
8e9a0dd0-31b6-4e7c-bd88-f831cd809970 | a-self-paced-bci-system-with-low-latency-for | 2204.05450 | null | https://arxiv.org/abs/2204.05450v1 | https://arxiv.org/pdf/2204.05450v1.pdf | A self-paced BCI system with low latency for motor imagery onset detection based on time series prediction paradigm | In a self-paced motor-imagery brain-computer interface (MI-BCI), the onsets of the MI commands presented in a continuous electroencephalogram (EEG) signal are unknown. To detect these onsets, most self-paced approaches apply a window function on the continuous EEG signal and split it into long segments for further analysis. As a result, the system has a high latency. To reduce the system latency, we propose an algorithm based on the time series prediction concept and use the data of the previously received time samples to predict the upcoming time samples. Our predictor is an encoder-decoder (ED) network built with long short-term memory (LSTM) units. The onsets of the MI commands are detected shortly by comparing the incoming signal with the predicted signal. The proposed method is validated on dataset IVc from BCI competition III. The simulation results show that the proposed algorithm improves the average F1-score achieved by the winner of the competition by 26.7% for latencies shorter than one second. | ['Elnaz Banan Sadeghian', 'Navid Ayoobi'] | 2022-04-12 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 6.11206770e-01 -3.00772756e-01 -7.25568756e-02 -2.75179267e-01
-5.59964657e-01 -3.36504616e-02 4.01412815e-01 1.13904593e-03
-7.13099062e-01 8.05571914e-01 1.36488199e-01 -1.09861881e-01
-1.70340911e-01 -3.06514531e-01 -7.17059672e-01 -4.22261298e-01
-5.77629149e-01 2.28898004e-02 1.89872637e-01 6.64524958e-02
5.28991461e-01 3.97666730e-02 -1.62884617e+00 6.46391213e-01
8.90534163e-01 1.30243707e+00 6.99607432e-01 4.72542107e-01
4.69248779e-02 6.34342849e-01 -8.56271803e-01 7.28183270e-01
-1.38679102e-01 -6.86172545e-01 -6.70381725e-01 -3.04932892e-01
-4.39946294e-01 -2.37547323e-01 -6.12210631e-01 9.92925107e-01
5.19291759e-01 -9.09798294e-02 4.55765426e-01 -1.27299929e+00
1.03380799e-01 5.29081583e-01 -3.06112468e-01 7.24361479e-01
5.38900614e-01 -1.33882731e-01 1.63931683e-01 -7.45494187e-01
5.04355907e-01 4.82542604e-01 3.81879330e-01 4.64186400e-01
-9.19263899e-01 -9.42850351e-01 2.11901248e-01 1.06877387e+00
-1.65623355e+00 -3.42804462e-01 7.57679462e-01 -4.68863368e-01
1.14901495e+00 3.05189900e-02 7.53376245e-01 1.25788760e+00
9.36948955e-01 8.93500268e-01 1.08842909e+00 -3.79919559e-01
4.52140659e-01 -9.55880210e-02 4.79317546e-01 1.16598401e-02
-4.16802973e-01 1.75695345e-01 -9.40484226e-01 2.47888103e-01
4.71318215e-01 1.14027776e-01 -7.41620183e-01 4.89343226e-01
-1.35431612e+00 2.84896910e-01 3.01397443e-01 5.09218872e-01
-9.12801325e-01 -1.78393483e-01 4.81900424e-01 6.48062527e-01
3.78218383e-01 1.17529131e-01 -3.45351458e-01 -7.32070804e-01
-1.25566232e+00 -3.09216917e-01 6.51046395e-01 8.94172370e-01
3.40019524e-01 -7.04289589e-04 -2.73918420e-01 5.36516786e-01
-2.45577812e-01 1.94766507e-01 1.09545541e+00 -2.54489958e-01
4.85882521e-01 2.88601935e-01 1.69162657e-02 -6.93380952e-01
-4.00522470e-01 -4.97471899e-01 -7.62914240e-01 1.88198492e-01
8.82639140e-02 -3.94162655e-01 -8.53430688e-01 1.42535138e+00
-3.43701452e-01 6.25686765e-01 -1.41880676e-01 9.55861449e-01
3.85336131e-01 9.15814579e-01 -1.58049241e-02 -6.93007648e-01
1.11435735e+00 -4.60383803e-01 -1.03732109e+00 -4.91819859e-01
1.62738353e-01 -4.07956898e-01 6.09345615e-01 8.23664546e-01
-8.46348882e-01 -7.65121162e-01 -1.29238677e+00 6.34592891e-01
-7.19018504e-02 2.35480309e-01 6.55699745e-02 1.59187813e-03
-7.63404012e-01 9.56754148e-01 -1.08952177e+00 -1.54835492e-01
1.36907399e-01 6.73338413e-01 -1.89275339e-01 4.42797691e-01
-1.46854842e+00 9.10447776e-01 5.47948420e-01 3.22944999e-01
-1.02710700e+00 -5.39217174e-01 -4.02147263e-01 4.17866409e-02
-8.83451775e-02 -1.12081543e-02 1.08560634e+00 -1.23597884e+00
-1.47538149e+00 5.09670854e-01 -6.21018052e-01 -7.14712918e-01
3.26973885e-01 -1.97350413e-01 -8.99216712e-01 9.65859145e-02
-9.23010111e-02 3.86908561e-01 1.09336913e+00 -5.23988068e-01
-9.52646196e-01 -4.43177849e-01 -5.62658250e-01 1.34202883e-01
-3.31322879e-01 2.65635550e-02 -1.27402887e-01 -4.71131891e-01
3.72552536e-02 -7.97500014e-01 3.50843251e-01 -8.00448418e-01
-1.78541586e-01 -1.79090962e-01 6.84770525e-01 -8.70195925e-01
1.51661968e+00 -2.39333653e+00 1.39420414e-02 7.89312348e-02
-4.66964096e-02 1.52821511e-01 -1.52703911e-01 2.71752506e-01
-5.20439625e-01 -5.46173930e-01 -3.90936509e-02 2.22143680e-01
-4.41947758e-01 -1.33664891e-01 -3.94675702e-01 4.42236811e-01
-2.63063367e-02 5.92769027e-01 -7.82114089e-01 -3.02809980e-02
4.31127921e-02 3.93125564e-01 2.01753557e-01 5.26582599e-01
1.81917399e-01 6.50980532e-01 1.39518201e-01 4.54699211e-02
3.23884994e-01 1.34179309e-01 -1.34364337e-01 -2.03504130e-01
-4.57464337e-01 7.05570936e-01 -9.14559484e-01 1.99394059e+00
-4.04890686e-01 1.08299148e+00 -3.42910916e-01 -1.04870903e+00
8.31360996e-01 7.31842756e-01 5.49394011e-01 -1.06731749e+00
2.71368027e-01 2.32805416e-01 3.92214477e-01 -5.78611255e-01
-1.94609985e-01 1.31892189e-01 4.39394832e-01 2.71311522e-01
3.14087197e-02 4.19638783e-01 1.38479352e-01 -1.84305549e-01
1.26088870e+00 1.24675982e-01 1.87919110e-01 -1.75777957e-01
5.42547166e-01 -1.31086349e-01 5.81889689e-01 5.70709586e-01
-1.19485222e-01 2.89661855e-01 2.69082606e-01 -4.31318164e-01
-5.97837210e-01 -8.71291399e-01 -1.79158837e-01 7.49747217e-01
2.07274169e-01 -3.66779387e-01 -8.40200961e-01 -1.49129763e-01
-5.01870334e-01 9.24779475e-01 -4.48117167e-01 -5.54694414e-01
-5.18867373e-01 -4.81428832e-01 9.49624553e-03 5.46554267e-01
2.83248961e-01 -1.53744256e+00 -1.09939146e+00 7.79600203e-01
-4.45232511e-01 -8.96988213e-01 -3.74532640e-01 7.31848538e-01
-9.18388367e-01 -7.44362175e-01 -7.14694738e-01 -1.05133653e+00
4.18000877e-01 1.30478395e-02 4.21796918e-01 -4.63726580e-01
-1.32858559e-01 -3.24261226e-02 -3.33494991e-01 -6.69112861e-01
1.77310854e-02 -7.12576807e-02 2.59858400e-01 1.22066364e-01
8.16796184e-01 -8.72913122e-01 -7.09600568e-01 2.55175173e-01
-5.01994669e-01 4.13407534e-01 7.13118196e-01 7.94809580e-01
6.90900385e-01 2.41637990e-01 1.09566879e+00 -2.74716526e-01
8.10573220e-01 -5.18585682e-01 -4.13271517e-01 1.25999048e-01
-7.42048323e-01 -4.72312234e-02 8.65559936e-01 -9.07019138e-01
-8.70624602e-01 8.19313675e-02 7.15416223e-02 -2.59265095e-01
-2.55201101e-01 6.29342496e-01 7.53685012e-02 9.73395035e-02
5.39615929e-01 8.95806730e-01 -2.36298531e-01 -3.35101962e-01
-3.53237092e-01 1.21332991e+00 8.53764057e-01 1.40357763e-02
1.15357555e-01 1.23564256e-02 -5.59928238e-01 -9.30178583e-01
-3.28483373e-01 -4.13579434e-01 -5.55661380e-01 -5.61137795e-01
5.42690158e-01 -1.11393321e+00 -6.75189674e-01 6.03300750e-01
-1.43445051e+00 -2.70870477e-01 1.91438958e-01 1.16677558e+00
-6.70212507e-01 1.30021155e-01 -5.54463744e-01 -8.75946105e-01
-7.10347891e-01 -9.90416169e-01 5.23345292e-01 3.21079791e-01
-6.55867398e-01 -2.20013380e-01 -1.59247383e-01 -3.01379919e-01
3.74686450e-01 -6.03050478e-02 7.74379373e-01 -6.19114578e-01
-2.21172608e-02 -5.30965626e-01 1.90312043e-01 2.09599435e-01
2.25453809e-01 -6.14957511e-01 -1.04725087e+00 -2.29133353e-01
5.86185336e-01 2.80022234e-01 5.20118117e-01 6.38760984e-01
1.34563196e+00 -1.20179795e-01 -5.08747220e-01 5.56121349e-01
1.28164327e+00 1.09331656e+00 9.00896728e-01 3.16522568e-01
2.37633303e-01 2.47545391e-01 4.36904937e-01 5.90729713e-01
7.62535185e-02 4.04658765e-01 1.43780559e-01 2.46532336e-01
2.63473123e-01 -1.83323190e-01 7.67862260e-01 1.03991604e+00
-6.77117109e-02 1.38034388e-01 -7.18529463e-01 6.72850847e-01
-1.78775167e+00 -9.53937888e-01 -2.54886687e-01 2.69080329e+00
9.35004473e-01 5.29331148e-01 -1.12665705e-01 6.88364923e-01
7.90990710e-01 -3.02381486e-01 -8.43538344e-01 -5.04702151e-01
1.77641675e-01 6.49625301e-01 3.40202391e-01 2.35000357e-01
-7.82717884e-01 5.26028991e-01 5.80187702e+00 8.54771137e-01
-1.49493003e+00 1.06386542e-01 3.58404487e-01 -5.26230156e-01
4.52400684e-01 -3.66596550e-01 -4.89757389e-01 9.23832417e-01
1.62096667e+00 -5.92453122e-01 8.45458627e-01 6.20352387e-01
7.68989086e-01 -6.41762614e-01 -1.47410786e+00 1.53100240e+00
-2.22059600e-02 -9.07746851e-01 -4.50032353e-01 -2.18063548e-01
2.44153038e-01 2.51598597e-01 -2.95217216e-01 3.47078800e-01
-7.91386962e-01 -1.05401242e+00 7.91182339e-01 8.64966631e-01
9.96161580e-01 -9.10561323e-01 6.57642126e-01 7.33217359e-01
-1.18521035e+00 -1.78303733e-01 -2.83149093e-01 -4.34517384e-01
1.13475427e-01 5.64539194e-01 -6.61556721e-01 2.19107285e-01
4.95582402e-01 8.22887063e-01 -2.25069627e-01 1.33157694e+00
-4.18586075e-01 1.06533730e+00 -1.99982718e-01 -1.37251332e-01
-4.59497608e-03 -1.66909490e-02 6.50922835e-01 1.05065751e+00
5.31610966e-01 2.41481960e-01 -4.05761868e-01 6.49237752e-01
3.03647101e-01 -1.11518309e-01 -5.01976073e-01 1.10449314e-01
3.98464352e-01 9.35779989e-01 -4.68852967e-01 -3.24469298e-01
-1.45574212e-01 1.45278239e+00 7.18406066e-02 4.77323472e-01
-8.14481795e-01 -9.52827811e-01 1.18660569e-01 -1.55091569e-01
2.88477074e-02 -1.36659265e-01 -4.65396464e-01 -9.73491311e-01
4.20266092e-01 -6.80442691e-01 1.19593874e-01 -8.91934574e-01
-8.14446151e-01 1.02834475e+00 -2.42464319e-01 -1.61427009e+00
-5.21853268e-01 -2.30388716e-01 -8.25444341e-01 1.08189893e+00
-1.27053511e+00 -5.27696870e-02 -2.82626301e-01 7.85083830e-01
7.40986586e-01 8.00908078e-03 1.00383770e+00 4.27311242e-01
-4.14853066e-01 2.95420498e-01 1.12710617e-01 -5.21044955e-02
5.96528590e-01 -8.74797344e-01 1.74833491e-01 8.78569424e-01
-2.02809367e-02 3.88330579e-01 7.55309463e-01 -7.49574065e-01
-1.30872083e+00 -7.90099800e-01 1.03195548e+00 2.38766983e-01
4.33637142e-01 -4.02661800e-01 -1.03624380e+00 4.61059004e-01
4.15218204e-01 -4.80719179e-01 3.48693907e-01 -4.02356356e-01
4.13373083e-01 -3.92830551e-01 -7.69461274e-01 3.71709496e-01
8.72221231e-01 -4.96987194e-01 -1.04666853e+00 9.90810022e-02
1.71072766e-01 -1.38414487e-01 -7.01848745e-01 2.24478528e-01
6.76601887e-01 -6.53863370e-01 4.56539363e-01 -2.91220248e-01
2.95193225e-01 -1.09029897e-01 2.95799911e-01 -1.75978434e+00
-3.15500945e-01 -8.09971631e-01 -9.42667276e-02 5.16892731e-01
3.30149591e-01 -4.11652625e-01 5.20497799e-01 5.79616427e-01
-1.69511110e-01 -6.62642360e-01 -1.07161415e+00 -9.44832683e-01
-5.61336935e-01 -6.36914968e-01 3.18678916e-01 2.82078028e-01
9.60344374e-01 4.53996867e-01 -5.12085915e-01 3.65986861e-02
3.84876877e-01 -8.31234574e-05 9.29049700e-02 -1.10250175e+00
1.00180700e-01 -1.77832276e-01 -4.98961836e-01 -1.13132131e+00
8.58811960e-02 -8.88102889e-01 5.15864968e-01 -1.42514706e+00
-6.46770820e-02 1.71653926e-01 -7.90704310e-01 3.29178512e-01
-1.04993284e-01 -3.25037628e-01 -1.11545183e-01 1.35103106e-01
-3.03608298e-01 5.60449839e-01 7.06270635e-01 -2.52182662e-01
-5.91349304e-01 1.54405013e-01 1.30057991e-01 6.52824998e-01
9.90801036e-01 -7.11616635e-01 -4.83072847e-01 -2.40771532e-01
-3.94326419e-01 6.49846911e-01 -1.33874640e-01 -1.85421288e+00
7.69927680e-01 1.34279668e-01 7.68612027e-01 -1.02487481e+00
3.72859597e-01 -7.60579586e-01 2.01218605e-01 8.43398631e-01
-4.76953655e-01 2.29077488e-02 3.07397813e-01 5.60199797e-01
-3.50810081e-01 -2.22231984e-01 4.45795238e-01 2.20522672e-01
-8.44016314e-01 3.08822691e-01 -1.01064563e+00 -1.99123353e-01
1.06925476e+00 -4.33360130e-01 1.08354382e-01 -3.55185896e-01
-7.99405932e-01 1.33696020e-01 -3.40026289e-01 4.91579682e-01
9.10564899e-01 -1.21688414e+00 -5.38446248e-01 6.10870361e-01
1.00158691e-01 -5.70958793e-01 2.22582906e-01 1.10245430e+00
7.24117132e-03 5.88376999e-01 -7.61454225e-01 -5.81573725e-01
-1.15596902e+00 4.04537559e-01 2.38336235e-01 2.45116472e-01
-9.61154163e-01 7.93156326e-01 -4.33522724e-02 7.09579468e-01
6.98292851e-01 -5.17569065e-01 -5.74361861e-01 1.26844570e-01
1.02581978e+00 2.23923415e-01 3.59907210e-01 -3.30656797e-01
-6.02291703e-01 -5.73715288e-03 -5.48735186e-02 -3.97301793e-01
1.61413229e+00 2.37181000e-02 -5.52303856e-03 8.81510735e-01
1.18052006e+00 -5.96557140e-01 -1.15413499e+00 -1.43120229e-01
3.26159924e-01 -1.70608521e-01 2.51889527e-01 -9.67114747e-01
-9.23172057e-01 8.70306194e-01 1.08654845e+00 -2.12970093e-01
1.46295726e+00 -6.63183808e-01 1.07878351e+00 2.97826529e-01
6.67085767e-01 -1.36206031e+00 -4.11742419e-01 3.80442232e-01
7.99154580e-01 -7.49139011e-01 -4.66993660e-01 3.33503962e-01
-6.24795139e-01 1.48890018e+00 5.00620127e-01 -2.08750978e-01
7.24819183e-01 2.50533491e-01 -1.89055189e-01 1.18427239e-01
-8.77319038e-01 3.79128614e-03 2.07751945e-01 6.04207873e-01
2.84148067e-01 2.43062690e-01 -9.35851455e-01 9.56750214e-01
-8.94774497e-02 6.45548761e-01 6.71364546e-01 7.86169231e-01
-4.87402529e-01 -6.63426816e-01 -3.56602430e-01 7.89543152e-01
-3.31850380e-01 -2.26841137e-01 3.17673385e-02 2.52908885e-01
-5.01208566e-02 1.33112276e+00 2.12233916e-01 -8.36495519e-01
4.07285869e-01 4.64894921e-01 4.50133175e-01 -3.15467626e-01
-3.77024114e-01 1.82778209e-01 -1.45104691e-01 -6.74046397e-01
-9.30273831e-02 -4.56041455e-01 -1.57299542e+00 1.60957873e-01
-8.18354040e-02 1.88894629e-01 9.99562025e-01 1.15256369e+00
5.37278235e-01 7.81097412e-01 8.07111502e-01 -8.17507744e-01
-2.62181878e-01 -1.31741905e+00 -6.84327781e-01 1.89909950e-01
3.89673799e-01 -4.73848522e-01 -2.84675211e-01 1.73453495e-01] | [13.087825775146484, 3.459249973297119] |
0432ee6b-cea8-4727-a652-8a14e96643d9 | revisiting-personalized-federated-learning | 2302.01677 | null | https://arxiv.org/abs/2302.01677v2 | https://arxiv.org/pdf/2302.01677v2.pdf | Revisiting Personalized Federated Learning: Robustness Against Backdoor Attacks | In this work, besides improving prediction accuracy, we study whether personalization could bring robustness benefits to backdoor attacks. We conduct the first study of backdoor attacks in the pFL framework, testing 4 widely used backdoor attacks against 6 pFL methods on benchmark datasets FEMNIST and CIFAR-10, a total of 600 experiments. The study shows that pFL methods with partial model-sharing can significantly boost robustness against backdoor attacks. In contrast, pFL methods with full model-sharing do not show robustness. To analyze the reasons for varying robustness performances, we provide comprehensive ablation studies on different pFL methods. Based on our findings, we further propose a lightweight defense method, Simple-Tuning, which empirically improves defense performance against backdoor attacks. We believe that our work could provide both guidance for pFL application in terms of its robustness and offer valuable insights to design more robust FL methods in the future. We open-source our code to establish the first benchmark for black-box backdoor attacks in pFL: https://github.com/alibaba/FederatedScope/tree/backdoor-bench. | ['Minhao Cheng', 'Bolin Ding', 'Yaliang Li', 'Daoyuan Chen', 'Liuyi Yao', 'Zeyu Qin'] | 2023-02-03 | null | null | null | null | ['personalized-federated-learning'] | ['methodology'] | [-3.57256174e-01 -5.60624003e-01 -7.26750314e-01 -7.62975737e-02
-7.35351443e-01 -1.37736154e+00 3.55428547e-01 -9.16260108e-02
-1.84856400e-01 6.82470262e-01 7.01133683e-02 -1.02306366e+00
-1.30006969e-01 -8.82048845e-01 -8.58606160e-01 -4.35744077e-01
-3.79885703e-01 -1.34424075e-01 2.21905947e-01 -3.79954457e-01
3.96547765e-01 7.32751727e-01 -9.51235175e-01 4.44291979e-01
6.96227908e-01 5.94670713e-01 -4.90417629e-01 5.11877537e-01
4.69771743e-01 6.02522790e-01 -6.19362652e-01 -8.45573545e-01
6.82915628e-01 -8.84945020e-02 -9.01632965e-01 -6.19857907e-01
6.13721430e-01 -5.53435683e-01 -7.80657589e-01 9.81708169e-01
4.83669043e-01 7.59115368e-02 1.24440445e-02 -1.47248638e+00
-6.85474873e-01 1.07824326e+00 -5.71864724e-01 5.14598787e-01
3.03183138e-01 4.55982298e-01 1.13258183e+00 -5.68337440e-01
3.99315566e-01 1.30878377e+00 8.05366099e-01 5.51762879e-01
-1.43513036e+00 -1.35047019e+00 1.09595113e-01 3.11523110e-01
-1.29884851e+00 -5.25736749e-01 5.16790867e-01 1.14772832e-02
8.74145448e-01 7.94173479e-01 8.58808830e-02 1.50731802e+00
1.62446469e-01 6.74023628e-01 1.06509006e+00 -1.44659445e-01
2.48597730e-02 2.95522928e-01 6.63931549e-01 6.46191716e-01
7.36771345e-01 7.21288741e-01 -4.09503788e-01 -1.08825731e+00
1.62542939e-01 1.44674201e-02 -4.95334506e-01 -1.41048983e-01
-8.00379217e-01 1.10845077e+00 6.07775092e-01 1.46793619e-01
1.50509164e-01 2.61551499e-01 5.69148600e-01 5.56469262e-01
1.14653885e-01 7.13910460e-01 -6.14733875e-01 1.38678476e-01
-8.63817036e-01 5.77927887e-01 9.43116605e-01 5.45243144e-01
6.43595576e-01 -1.32770926e-01 -3.41992944e-01 3.65564704e-01
2.24730629e-03 4.92933959e-01 1.97995886e-01 -1.12362897e+00
4.42858845e-01 6.51461352e-03 -5.91741800e-02 -1.13981390e+00
-1.44663200e-01 -7.14197099e-01 -2.56750196e-01 -3.43057275e-01
3.23931515e-01 -4.09779727e-01 -2.25049138e-01 1.96064281e+00
3.94875199e-01 4.31065589e-01 -1.15671016e-01 5.48700452e-01
3.59817058e-01 3.77416790e-01 -1.62246168e-01 7.81230778e-02
1.33848202e+00 -1.05850780e+00 -2.74946898e-01 1.32011384e-01
1.07679749e+00 -7.15274155e-01 1.21969807e+00 3.08997542e-01
-6.38049901e-01 -7.34899938e-02 -9.17330623e-01 3.66889447e-01
-3.66623878e-01 -2.72120595e-01 1.11348855e+00 1.29390740e+00
-5.87787271e-01 6.46568000e-01 -9.16574299e-01 -3.44251722e-01
4.95189160e-01 2.35085770e-01 -3.56337935e-01 6.96569756e-02
-1.55283093e+00 4.25163239e-01 3.18010539e-01 -4.73474026e-01
-1.14412928e+00 -1.02562022e+00 -3.22403908e-01 1.58944547e-01
6.04747057e-01 -6.72479212e-01 1.10610199e+00 -2.70352632e-01
-1.09057605e+00 4.21593577e-01 1.12672746e-01 -1.14680755e+00
3.07074457e-01 -3.76783013e-01 -4.81363416e-01 -3.98689061e-02
-2.24584237e-01 2.53361344e-01 6.91722572e-01 -1.12313795e+00
-3.71687859e-01 -3.28815967e-01 6.26977444e-01 -5.11596620e-01
-9.77804780e-01 2.19462425e-01 -2.16091573e-01 -8.15450788e-01
-5.83200455e-01 -1.22428143e+00 -1.25779822e-01 -3.14021707e-01
-8.80758524e-01 1.42688796e-01 9.93615985e-01 -2.88851619e-01
1.72024560e+00 -2.11665869e+00 -4.97835249e-01 3.65165234e-01
2.27220640e-01 5.52829087e-01 -3.30465794e-01 8.85844350e-01
-4.57320586e-02 5.60091197e-01 4.21166234e-02 -1.34172037e-01
5.78568168e-02 -6.94953427e-02 -1.08597374e+00 6.63706303e-01
-5.66283524e-01 7.73332834e-01 -5.17863929e-01 -4.64087501e-02
-1.05528004e-01 4.08253431e-01 -1.13576603e+00 -1.21566735e-01
-9.82187912e-02 2.30948731e-01 -6.58779740e-01 7.40499437e-01
1.00729144e+00 -3.50247830e-01 1.07928291e-01 -2.29077250e-01
1.22518830e-01 3.13809156e-01 -8.97852004e-01 1.10334384e+00
-3.35013330e-01 3.01050901e-01 3.74188609e-02 -6.02682114e-01
4.61813033e-01 8.32772106e-02 4.59359363e-02 -2.33287677e-01
1.69273302e-01 1.07568868e-01 -2.16288082e-02 8.51712525e-02
3.67013872e-01 2.67988861e-01 -6.08121306e-02 7.50882447e-01
-3.38314384e-01 3.99839699e-01 9.67342257e-02 4.70276415e-01
1.24061084e+00 -4.78882551e-01 1.01904362e-01 -4.45545614e-01
6.93919301e-01 -1.96593165e-01 4.75264907e-01 1.31568837e+00
-3.76656324e-01 4.10758443e-02 4.27229017e-01 -4.86273885e-01
-2.80905694e-01 -1.10858405e+00 -1.94919661e-01 1.28596401e+00
9.60965678e-02 -1.13136053e+00 -8.83238494e-01 -1.05276406e+00
4.42892939e-01 8.28169763e-01 -5.99636376e-01 -3.95534039e-01
-4.18520123e-01 -8.66032481e-01 1.49245393e+00 2.15859562e-01
6.65605545e-01 -4.35586184e-01 -3.33587110e-01 -2.37246186e-01
-2.06920549e-01 -1.00410402e+00 -7.81673908e-01 -6.81314766e-02
-8.44758868e-01 -1.20897746e+00 -1.82232454e-01 1.30529320e-02
2.09129065e-01 6.97021604e-01 7.70970523e-01 4.71192032e-01
-8.43905360e-02 1.60094708e-01 -2.48915121e-01 -2.07375199e-01
-2.57747591e-01 4.09956992e-01 4.22520190e-01 -1.98162496e-01
2.43303210e-01 -6.99466228e-01 -7.54008234e-01 6.34262145e-01
-7.98560977e-01 -5.69748282e-01 -3.87536385e-03 7.48632193e-01
2.74792731e-01 7.75436386e-02 2.69017071e-01 -1.16632092e+00
8.33415270e-01 -6.65635407e-01 -8.49094212e-01 3.06716442e-01
-6.18279159e-01 2.29203235e-02 8.48757029e-01 -4.55951452e-01
-5.77176154e-01 -4.28239971e-01 -3.44451845e-01 -7.82006502e-01
2.02343270e-01 2.94497967e-01 8.46957564e-02 -5.93284369e-01
1.20058584e+00 4.26654518e-02 -3.04809600e-01 -5.83689272e-01
4.86405492e-01 4.19132203e-01 4.11221772e-01 -1.03678691e+00
1.16262782e+00 6.07499301e-01 -1.23496793e-01 -5.54116786e-01
-7.15474308e-01 -6.81750923e-02 3.29925865e-01 4.56505805e-01
2.33257592e-01 -7.72029281e-01 -9.20720160e-01 2.17700332e-01
-7.74475694e-01 -3.82096589e-01 1.20817579e-01 2.12876514e-01
-1.68995813e-01 8.63428891e-01 -1.04496717e+00 -4.73019838e-01
-7.02881753e-01 -1.16934073e+00 4.96589363e-01 1.41753376e-01
-8.48297924e-02 -1.11978579e+00 2.55434573e-01 5.86237609e-01
5.00324428e-01 8.28124061e-02 7.41387308e-01 -1.05195343e+00
-5.50175071e-01 -1.90004319e-01 -1.72496289e-02 2.65475065e-01
9.51792486e-03 4.99026179e-02 -1.25991750e+00 -9.00384188e-01
-1.27969652e-01 -4.05298233e-01 1.04530263e+00 1.11373395e-01
1.76964939e+00 -6.42458379e-01 -5.95921338e-01 1.19095647e+00
1.29551101e+00 -2.33103633e-01 4.28055614e-01 6.03432715e-01
4.43788111e-01 2.50152439e-01 5.99878252e-01 6.22484565e-01
1.28427267e-01 6.22071147e-01 4.94770080e-01 7.59323537e-02
2.94104189e-01 -4.85022545e-01 6.37822211e-01 -3.88589017e-02
2.21711665e-01 -5.13517499e-01 -7.93334663e-01 1.63411692e-01
-1.37602687e+00 -1.05985844e+00 1.15510173e-01 2.33793020e+00
6.63417459e-01 1.73115969e-01 3.40403765e-01 9.31491628e-02
5.49211204e-01 2.11798981e-01 -4.06834483e-01 -5.23042023e-01
1.60874352e-02 1.64880902e-01 9.07555997e-01 4.82612342e-01
-1.11480594e+00 1.17105925e+00 6.39697027e+00 1.23054683e+00
-1.24891889e+00 1.57012701e-01 6.78649604e-01 -3.20000291e-01
-4.76999521e-01 4.25938487e-01 -1.09071612e+00 4.42978740e-01
1.25186646e+00 -5.67255378e-01 8.54496479e-01 8.69690359e-01
-1.09134370e-03 5.28275490e-01 -1.07927990e+00 5.85471392e-01
-3.68842334e-01 -1.65779865e+00 1.02788836e-01 6.03897035e-01
6.54420018e-01 3.30507725e-01 4.89585698e-01 4.40496266e-01
4.60605323e-01 -9.15275872e-01 2.83992827e-01 -2.27566063e-02
5.82819402e-01 -1.01904464e+00 3.03126991e-01 4.69046146e-01
-1.01155925e+00 -4.93325919e-01 -4.52066332e-01 1.33251399e-01
-1.85333356e-01 4.05811012e-01 -5.54459333e-01 8.13580871e-01
7.45185435e-01 3.27397764e-01 -7.24775016e-01 8.70758891e-01
-1.30108278e-02 1.20152628e+00 -5.85152626e-01 3.36304247e-01
1.23144306e-01 4.40316558e-01 7.92357445e-01 1.23842299e+00
1.23355381e-01 -1.53430313e-01 2.94582784e-01 8.87924671e-01
-2.69309968e-01 8.34059864e-02 -6.77425027e-01 -2.63304234e-01
1.08102405e+00 1.03448629e+00 -3.83489192e-01 -9.23746601e-02
-1.76553115e-01 5.75793982e-01 1.77624494e-01 1.95614174e-01
-1.22055042e+00 -4.20162082e-01 1.22053409e+00 2.14821473e-01
1.04527362e-01 -1.88628044e-02 -1.41825601e-01 -1.43805230e+00
-4.08833802e-01 -1.39920437e+00 9.63413537e-01 -9.72088203e-02
-1.51765108e+00 6.24177396e-01 1.52111501e-01 -7.78067827e-01
2.48028599e-02 -3.04694414e-01 -8.43381643e-01 5.55817306e-01
-1.28886914e+00 -1.01331878e+00 1.73373401e-01 1.03762066e+00
4.19427864e-02 -2.69316912e-01 8.28832030e-01 3.16341251e-01
-8.63681734e-01 1.69639313e+00 2.53851503e-01 6.66090921e-02
9.10885036e-01 -5.59324563e-01 7.44213104e-01 1.09833598e+00
3.90316993e-01 1.47051644e+00 8.13009441e-01 -5.65038621e-01
-1.56625831e+00 -1.02619243e+00 3.78737539e-01 -6.20673478e-01
8.98765206e-01 -3.73683989e-01 -8.71228218e-01 8.98120522e-01
-6.03644699e-02 2.61290669e-01 1.01066053e+00 3.20156276e-01
-9.01004791e-01 -1.58409819e-01 -1.39828992e+00 7.45366454e-01
9.55085337e-01 -5.93853533e-01 -8.57930109e-02 4.14633930e-01
1.05879116e+00 -3.04217100e-01 -9.15261984e-01 3.75776112e-01
5.82422078e-01 -1.26731539e+00 1.27930176e+00 -6.50492668e-01
-1.56547517e-01 5.52890152e-02 -5.50450087e-01 -9.09030437e-01
-3.00824612e-01 -8.63039374e-01 -4.54285502e-01 1.17496061e+00
3.83249670e-01 -1.30330145e+00 8.68973732e-01 1.99133545e-01
4.38581973e-01 -7.87749708e-01 -7.14427412e-01 -1.14978135e+00
6.02672756e-01 -6.46559775e-01 1.17933559e+00 1.03674567e+00
-1.77484140e-01 -2.55742908e-01 -5.48327744e-01 4.14057702e-01
8.98940146e-01 1.47474572e-01 1.09794152e+00 -7.86395669e-01
-8.19367945e-01 -2.57437915e-01 8.27886239e-02 -7.85484195e-01
4.47588265e-01 -1.10428250e+00 -9.14664447e-01 -4.96937722e-01
2.25528046e-01 -4.45158362e-01 -6.36328995e-01 8.27218473e-01
-4.48362045e-02 4.48436975e-01 5.22442043e-01 3.00925493e-01
-3.92906636e-01 3.82894218e-01 8.01049590e-01 3.81532535e-02
-1.53036848e-01 2.47031271e-01 -1.23634565e+00 3.32719713e-01
1.07347715e+00 -6.74016416e-01 -6.25019908e-01 -1.47177339e-01
-1.15618385e-01 -1.24780402e-01 3.59379172e-01 -8.09884250e-01
1.98903922e-02 -2.70877987e-01 -2.33436227e-02 -2.72589564e-01
1.81278400e-02 -5.60067415e-01 1.45849273e-01 8.81945312e-01
-3.19111288e-01 8.38195011e-02 5.93325734e-01 5.01292467e-01
2.60587066e-01 1.62666857e-01 7.19713807e-01 1.08758092e-01
-3.42102110e-01 5.25358915e-01 -1.47903651e-01 1.32389084e-01
9.78359282e-01 -4.67671268e-03 -8.02345991e-01 -2.96539098e-01
-2.72903353e-01 4.25487369e-01 7.21083939e-01 4.00058836e-01
3.44301879e-01 -1.03496861e+00 -6.71509981e-01 4.37829375e-01
1.17819376e-01 -1.09643900e+00 2.46431023e-01 7.58025289e-01
-4.79974240e-01 7.35533714e-01 -6.64670691e-02 -2.07795426e-01
-1.63596570e+00 1.02017331e+00 5.66894829e-01 -4.84381586e-01
-5.11563838e-01 9.76481378e-01 2.87699759e-01 -5.71607828e-01
3.29136580e-01 2.15658277e-01 3.15023839e-01 -4.97893393e-01
7.97149241e-01 5.94240010e-01 3.30882445e-02 -3.05328965e-01
-7.15960860e-01 -2.53587328e-02 -4.50655967e-01 1.97029896e-02
9.08816993e-01 8.03120807e-02 -9.94179249e-02 -3.20731342e-01
1.30614913e+00 6.06290996e-01 -8.25021207e-01 -1.10408984e-01
-1.92009643e-01 -9.41982448e-01 1.10699296e-01 -8.08913887e-01
-1.23272359e+00 5.09862244e-01 6.11471295e-01 9.99697074e-02
1.17565989e+00 -2.77490079e-01 9.94442165e-01 7.49320149e-01
7.35582888e-01 -4.34745938e-01 -2.29366556e-01 4.88970459e-01
5.66707015e-01 -6.51378751e-01 1.07396446e-01 -4.60421652e-01
-3.17948133e-01 7.96416700e-01 5.59377789e-01 -3.52123231e-01
7.94560373e-01 3.03031892e-01 -2.50024140e-01 4.36203405e-02
-1.15240121e+00 3.64914805e-01 8.18235986e-03 2.81467378e-01
3.49208146e-01 4.64667343e-02 -3.83357793e-01 7.85794318e-01
-4.97790158e-01 -7.59139359e-02 4.33837742e-01 7.37874448e-01
-1.48874298e-01 -1.59744060e+00 -7.47244120e-01 2.78510183e-01
-1.04643416e+00 -1.80087328e-01 -4.91042018e-01 7.55543292e-01
-2.14264587e-01 1.13616002e+00 -5.33591032e-01 -9.55719948e-01
1.07688792e-01 -1.99202582e-01 3.46720368e-01 -2.20935911e-01
-1.08640718e+00 -3.71430606e-01 1.68795586e-01 -9.06077981e-01
4.60015237e-01 -3.86628151e-01 -9.32315230e-01 -1.19706023e+00
-5.32500744e-01 5.81183195e-01 3.64832938e-01 4.09913957e-01
7.22111583e-01 -3.33388969e-02 9.55545723e-01 -4.12012309e-01
-1.13180315e+00 -3.93711567e-01 -4.31250542e-01 1.17423803e-01
2.08427399e-01 -6.17971003e-01 -6.56098068e-01 -6.78291976e-01] | [5.787728786468506, 7.384841442108154] |
91c5460a-9e57-4095-8aa8-86c86bca11e9 | panic-3d-stylized-single-view-3d | 2303.14587 | null | https://arxiv.org/abs/2303.14587v1 | https://arxiv.org/pdf/2303.14587v1.pdf | PAniC-3D: Stylized Single-view 3D Reconstruction from Portraits of Anime Characters | We propose PAniC-3D, a system to reconstruct stylized 3D character heads directly from illustrated (p)ortraits of (ani)me (c)haracters. Our anime-style domain poses unique challenges to single-view reconstruction; compared to natural images of human heads, character portrait illustrations have hair and accessories with more complex and diverse geometry, and are shaded with non-photorealistic contour lines. In addition, there is a lack of both 3D model and portrait illustration data suitable to train and evaluate this ambiguous stylized reconstruction task. Facing these challenges, our proposed PAniC-3D architecture crosses the illustration-to-3D domain gap with a line-filling model, and represents sophisticated geometries with a volumetric radiance field. We train our system with two large new datasets (11.2k Vroid 3D models, 1k Vtuber portrait illustrations), and evaluate on a novel AnimeRecon benchmark of illustration-to-3D pairs. PAniC-3D significantly outperforms baseline methods, and provides data to establish the task of stylized reconstruction from portrait illustrations. | ['Matthias Zwicker', 'Xiao Yang', 'Janus Kristjansson', 'Sizhe An', 'Guoxian Song', 'Yiheng Zhu', 'Heng Wang', 'Yichun Shi', 'Kevin Zhang', 'Shuhong Chen'] | 2023-03-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Chen_PAniC-3D_Stylized_Single-View_3D_Reconstruction_From_Portraits_of_Anime_Characters_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_PAniC-3D_Stylized_Single-View_3D_Reconstruction_From_Portraits_of_Anime_Characters_CVPR_2023_paper.pdf | cvpr-2023-1 | ['single-view-3d-reconstruction'] | ['computer-vision'] | [ 1.18940890e-01 4.65263017e-02 1.54307798e-01 -1.90270960e-01
-4.83466119e-01 -1.07893109e+00 7.39077687e-01 -6.93805218e-01
5.26096165e-01 5.09151280e-01 1.57895237e-02 -3.30193341e-01
5.40401757e-01 -6.74331903e-01 -9.11357462e-01 -3.24679255e-01
3.27189237e-01 9.78296578e-01 -5.42455018e-02 -3.39955360e-01
-6.27140654e-03 1.06743920e+00 -1.23982680e+00 1.16761096e-01
7.24747121e-01 8.14390123e-01 -2.33040657e-02 8.93675745e-01
-3.68671298e-01 4.04627591e-01 -9.28886592e-01 -9.43805516e-01
3.87573093e-01 -3.79656464e-01 -1.87370226e-01 5.55519164e-01
1.24005580e+00 -5.99080920e-01 -3.82283032e-01 4.28833753e-01
5.30774474e-01 -3.19868296e-01 1.26292467e+00 -1.16510570e+00
-1.07775664e+00 2.26722464e-01 -1.28011537e+00 -3.11878085e-01
7.15251088e-01 3.82528394e-01 6.85639739e-01 -1.18984020e+00
1.27020621e+00 1.79865479e+00 8.38797152e-01 7.09867239e-01
-1.45844913e+00 -6.45986855e-01 1.61672816e-01 -6.39978826e-01
-1.36507165e+00 -2.78017014e-01 9.14066792e-01 -2.00058013e-01
5.89382768e-01 5.59522450e-01 1.11895669e+00 1.79115117e+00
1.49423152e-01 9.95679140e-01 1.36983788e+00 -3.76556697e-03
-1.04837321e-01 3.14496756e-01 -3.58045489e-01 9.44122910e-01
-5.43944985e-02 -1.40641592e-02 -4.64149356e-01 -1.18964359e-01
1.45776653e+00 -3.16445798e-01 -4.32754248e-01 -5.87310374e-01
-1.16945457e+00 4.67345685e-01 4.71138835e-01 -2.99060553e-01
6.52140900e-02 3.06465656e-01 1.39130667e-01 -1.31848022e-01
4.91001725e-01 5.31043053e-01 4.40627448e-02 9.18479078e-03
-9.32655334e-01 4.53386426e-01 7.84700930e-01 1.61065936e+00
2.96467364e-01 8.43518734e-01 2.97588054e-02 8.70934308e-01
1.37136867e-02 1.10842490e+00 -2.48118162e-01 -1.14223516e+00
5.01787663e-01 3.59887570e-01 -5.54846972e-02 -8.70684266e-01
-1.98851615e-01 -3.06442857e-01 -1.07434773e+00 4.34101105e-01
4.68912870e-01 5.77060878e-02 -1.03545654e+00 1.08568168e+00
4.90298003e-01 3.70940380e-02 -1.63694575e-01 1.07872605e+00
1.49743199e+00 7.24177837e-01 -4.10660177e-01 2.76016504e-01
1.38695979e+00 -8.21285069e-01 -6.67227030e-01 2.07564384e-02
3.42845991e-02 -8.55476975e-01 1.52748060e+00 5.27521670e-01
-1.52305651e+00 -4.47766453e-01 -9.29226875e-01 -5.73047340e-01
-1.10053927e-01 1.06361933e-01 5.80214918e-01 7.01324999e-01
-9.95459199e-01 4.80102479e-01 8.92796144e-02 -2.95023769e-02
7.65080392e-01 -2.58499235e-01 -4.56220508e-01 -2.48126015e-01
-6.69677734e-01 7.78301060e-01 -3.12948138e-01 5.32055162e-02
-8.71890903e-01 -1.14193523e+00 -1.02241659e+00 -2.10365102e-01
8.65601972e-02 -8.73821735e-01 1.08548176e+00 -5.35740435e-01
-1.74366057e+00 1.30422843e+00 1.44450650e-01 5.30508645e-02
1.15649569e+00 1.35743037e-01 -4.05347526e-01 2.68767625e-01
-6.62179962e-02 8.69307756e-01 1.17376530e+00 -2.13144207e+00
1.05027445e-01 -2.62734950e-01 -1.48070902e-01 2.25833729e-01
4.07300919e-01 -4.51403797e-01 -8.51844907e-01 -9.20112014e-01
4.03745994e-02 -7.87068486e-01 2.12829828e-01 5.76337576e-01
-1.09851253e+00 3.01793516e-01 1.06088531e+00 -6.79984987e-01
6.71958089e-01 -1.91121089e+00 1.82361633e-01 1.38958007e-01
4.72759753e-01 2.03769580e-02 -2.55236864e-01 2.24862859e-01
-5.33834025e-02 4.01353151e-01 -2.85293847e-01 -9.56246436e-01
1.93836078e-01 1.58428371e-01 -7.28879392e-01 4.27085876e-01
2.11424053e-01 1.21415758e+00 -6.63691223e-01 -7.47379899e-01
3.53211552e-01 7.05928147e-01 -2.66377419e-01 8.75461400e-02
-4.04589474e-01 6.15060508e-01 -2.80362815e-01 1.26261616e+00
1.11031651e+00 -1.49672776e-01 -8.54487251e-03 -2.47957066e-01
1.85015827e-01 -4.27527189e-01 -8.95556450e-01 1.83301198e+00
-5.17751932e-01 7.68612683e-01 2.01339364e-01 -2.56213043e-02
1.11402714e+00 -9.13923606e-02 1.63128495e-01 -8.03910255e-01
9.39072296e-02 4.07111272e-02 -7.25560784e-01 -1.51027501e-01
9.01143551e-01 -1.24348089e-01 -3.34093451e-01 3.44021469e-01
-2.91848898e-01 -1.27798820e+00 -3.06538999e-01 4.13257390e-01
5.89975655e-01 5.91943264e-01 -1.65008958e-02 -6.20589294e-02
3.83956730e-02 -6.13421649e-02 2.45286450e-01 4.04888988e-01
3.68741602e-01 1.36212933e+00 8.51196229e-01 -4.94212717e-01
-1.43292153e+00 -1.61658049e+00 -5.20266473e-01 2.60570735e-01
2.55278379e-01 -2.63602585e-01 -7.25812733e-01 -6.51921749e-01
4.42027181e-01 9.66074526e-01 -8.15982103e-01 4.50197250e-01
-7.68476605e-01 -7.81313553e-02 8.33036900e-01 3.80877316e-01
4.97423202e-01 -7.57564545e-01 -3.44860464e-01 -3.48788768e-01
1.48246169e-01 -1.25792229e+00 -7.74631798e-01 -2.06583440e-01
-7.01115906e-01 -9.03599620e-01 -1.13348520e+00 -4.82041776e-01
5.94586670e-01 3.78378063e-01 2.03420568e+00 2.45645978e-02
-5.70782125e-01 6.66499794e-01 -2.58899350e-02 -4.11504775e-01
-5.78495622e-01 -3.11130285e-01 -2.16948107e-01 -4.07628387e-01
-5.80084860e-01 -5.90798199e-01 -6.99692547e-01 6.45118296e-01
-6.76711023e-01 4.99314010e-01 3.13017637e-01 6.03543878e-01
9.67046857e-01 -6.45450413e-01 1.61980256e-01 -9.29628253e-01
4.21334356e-01 -2.17366219e-01 -5.30553043e-01 2.55320191e-01
3.17253103e-03 -3.34430307e-01 7.88208783e-01 -4.73641217e-01
-1.19674683e+00 -1.64896935e-01 9.50465426e-02 -1.14976346e+00
-1.15569524e-01 -4.78263348e-01 -5.44353664e-01 1.26646802e-01
7.01645672e-01 5.48265949e-02 -9.86269638e-02 -4.62191373e-01
8.66310775e-01 1.97424904e-01 8.70841980e-01 -7.65020549e-01
1.13739347e+00 6.68007493e-01 3.36622685e-01 -1.06846488e+00
-2.83342302e-01 3.57812285e-01 -7.55071223e-01 -4.67961282e-01
8.99524510e-01 -8.90359402e-01 -7.22250342e-01 4.15226191e-01
-1.22383142e+00 -6.50471568e-01 -6.85341537e-01 -1.92503914e-01
-6.68352306e-01 4.03744787e-01 -6.59200013e-01 -6.57764256e-01
-3.22189450e-01 -1.13647664e+00 1.97553992e+00 1.13647141e-01
-4.43694800e-01 -8.48817885e-01 -1.97877765e-01 5.06923378e-01
-8.10090825e-02 8.57855856e-01 1.09533429e+00 4.39408511e-01
-9.35262501e-01 -7.80485524e-03 -4.74460989e-01 -1.40460804e-01
-7.73902461e-02 5.04241467e-01 -1.35099769e+00 2.88712848e-02
-5.21341383e-01 -4.70298648e-01 6.16991162e-01 4.05657232e-01
1.36769545e+00 1.57110821e-02 -3.20872784e-01 1.28419137e+00
1.10018039e+00 -1.63815930e-01 8.14495564e-01 -3.66655797e-01
1.18844199e+00 5.51087916e-01 4.88615781e-01 3.66849303e-01
3.93377036e-01 8.45638990e-01 5.32503426e-01 -4.06234533e-01
-6.88386202e-01 -8.77787948e-01 4.11099792e-02 6.33456767e-01
-1.75928250e-01 -7.71200001e-01 -5.48891127e-01 6.09338470e-02
-1.08215296e+00 -7.92756557e-01 -5.52838564e-01 1.79758596e+00
5.23458719e-01 -2.01443993e-02 1.32961392e-01 -2.06381842e-01
3.44653249e-01 4.20183718e-01 -1.02417338e+00 -8.43197346e-01
-6.38513088e-01 8.34692940e-02 2.90428281e-01 3.33921134e-01
-5.97325861e-01 1.17267597e+00 6.83268309e+00 1.00158167e+00
-1.01767719e+00 -4.37345892e-01 1.05711710e+00 -1.64024100e-01
-1.16132247e+00 -4.60304052e-01 -6.36008203e-01 2.56754279e-01
3.31953347e-01 2.35847056e-01 3.81224126e-01 7.01088846e-01
-7.86228180e-02 -8.77838023e-03 -1.29128253e+00 1.52760041e+00
4.33876157e-01 -1.56662261e+00 4.39704686e-01 2.05222443e-01
7.85945773e-01 -5.38848281e-01 7.50927448e-01 2.04526097e-01
3.84887040e-01 -1.59075916e+00 1.19489694e+00 6.37055755e-01
1.79842699e+00 -6.56140745e-01 -2.04111308e-01 1.18325315e-01
-1.31731713e+00 4.80082244e-01 -2.43202075e-01 6.42450690e-01
2.76133090e-01 2.28240505e-01 -9.56172109e-01 6.07859194e-01
5.65292776e-01 8.98752749e-01 -6.01755977e-01 4.79904771e-01
-1.95683137e-01 5.84008135e-02 -1.15934618e-01 -4.71075810e-03
1.10233195e-01 -6.90801382e-01 6.28119051e-01 1.15883446e+00
4.07459289e-01 6.23380356e-02 -4.09833044e-01 1.46326482e+00
-5.02816260e-01 -1.78900182e-01 -1.10020959e+00 1.09778479e-01
3.93694699e-01 1.25364244e+00 -8.60768199e-01 -1.89559132e-01
7.70509541e-02 1.44161749e+00 2.36548796e-01 4.67420548e-01
-1.16127121e+00 9.61954221e-02 4.92961854e-01 4.22995925e-01
2.88917333e-01 -2.26845145e-01 -5.55467665e-01 -1.25246859e+00
-1.26367360e-01 -8.81714344e-01 -3.94443065e-01 -1.57858384e+00
-1.51148593e+00 7.41369069e-01 9.90704298e-02 -1.43137836e+00
-8.48781615e-02 -5.84662497e-01 -6.29473925e-01 9.20117438e-01
-1.25024211e+00 -1.64562237e+00 -5.33931196e-01 4.35668439e-01
7.55990028e-01 -4.87751625e-02 9.18066084e-01 -1.07186660e-01
-4.76310670e-01 6.26997113e-01 1.35968506e-01 -1.96532264e-01
7.20988572e-01 -1.64633119e+00 1.12528265e+00 1.75427929e-01
2.54988372e-01 1.13962322e-01 6.00875974e-01 -7.58296490e-01
-1.86565721e+00 -9.49559033e-01 -5.49988113e-02 -1.13151944e+00
9.11170617e-02 -9.30435836e-01 -7.66073704e-01 5.26853621e-01
2.76958823e-01 7.68032447e-02 4.28328574e-01 -3.59071940e-01
-6.68013871e-01 1.67770579e-01 -1.37420166e+00 1.02322638e+00
1.53181434e+00 -3.59221935e-01 -1.76202312e-01 1.72173515e-01
5.84857345e-01 -9.45199072e-01 -9.82092202e-01 -2.85912417e-02
1.01673448e+00 -1.08396947e+00 1.48431945e+00 -2.66787797e-01
7.46462166e-01 -1.53014317e-01 -1.05165206e-01 -1.51213872e+00
1.38527334e-01 -8.39550495e-01 -1.38249442e-01 1.10652208e+00
1.98161542e-01 -5.80703281e-02 8.81995320e-01 5.78549743e-01
-2.46224612e-01 -8.49506319e-01 -6.45455062e-01 -5.36560833e-01
3.46382111e-01 -1.73109606e-01 1.01659524e+00 9.00966942e-01
-7.03332424e-01 4.50018406e-01 -6.08232975e-01 -3.84958893e-01
8.96032035e-01 6.12085879e-01 1.34264529e+00 -1.16175222e+00
-1.13088056e-01 -4.94839996e-01 7.04936981e-02 -1.42950416e+00
2.62370467e-01 -8.56001794e-01 -3.68426800e-01 -1.49992347e+00
-8.45936090e-02 -6.88832402e-01 9.45462048e-01 -1.42406598e-01
9.95411500e-02 4.04190660e-01 5.24950445e-01 3.55066434e-02
-1.11947253e-01 7.58667231e-01 2.24412155e+00 -2.95036644e-01
-2.03150325e-02 -1.78308189e-01 -4.11925465e-01 8.52811396e-01
2.25537255e-01 1.47847831e-01 -5.52144587e-01 -6.38377547e-01
-6.22473937e-03 5.26905596e-01 5.57919264e-01 -4.32370335e-01
-3.83410126e-01 1.83030054e-01 1.16393483e+00 -1.27327287e+00
1.28136897e+00 -8.66435707e-01 4.13333237e-01 -1.28980502e-01
1.31511226e-01 3.37989092e-01 3.30136538e-01 7.77288258e-01
4.64776367e-01 1.95024624e-01 7.92622387e-01 -3.62057090e-01
-3.27105701e-01 4.37999278e-01 -1.44881546e-01 2.74472684e-01
8.73697579e-01 -8.53330970e-01 -4.97849792e-01 -5.27945518e-01
-3.93070430e-01 1.24855243e-01 1.16818213e+00 2.91320384e-01
1.19024491e+00 -1.51391292e+00 -6.36717141e-01 3.14256698e-01
-4.65705767e-02 5.22532165e-01 5.35784483e-01 1.39998555e-01
-1.06138396e+00 5.56000173e-02 -2.83588856e-01 -8.45757425e-01
-1.36006224e+00 3.69693696e-01 2.97232419e-01 5.94827048e-02
-1.20951068e+00 8.70015025e-01 6.18758440e-01 -7.37076461e-01
2.45972097e-01 -5.51987529e-01 7.78223351e-02 5.29339490e-03
1.36359394e-01 2.57783532e-01 -3.73913705e-01 -8.66166592e-01
5.74633591e-02 1.03408039e+00 1.99400708e-01 -8.89710858e-02
1.15034688e+00 8.98562744e-02 4.08630461e-01 3.96070063e-01
1.00872755e+00 3.48539799e-01 -1.54362178e+00 4.29804511e-02
-7.76998222e-01 -7.76225328e-01 -3.90581131e-01 -7.82566369e-01
-1.16762781e+00 1.09280455e+00 6.70566037e-02 -1.41093671e-01
5.86671472e-01 -5.02036139e-02 8.53337526e-01 2.23311275e-01
4.18980092e-01 -6.29164636e-01 3.91479969e-01 1.84746936e-01
1.49801326e+00 -9.47541058e-01 3.13776881e-01 -6.92059934e-01
-1.17869282e+00 1.21196508e+00 7.23133266e-01 -2.07586825e-01
2.80894458e-01 3.90234858e-01 9.28470865e-03 -4.96518999e-01
-3.54041159e-01 3.92347306e-01 3.52645397e-01 7.64866590e-01
9.06696171e-03 1.77956969e-01 8.14980090e-01 3.27550590e-01
-7.57722378e-01 -7.07466662e-01 6.56978607e-01 4.52546984e-01
3.10411692e-01 -6.97635710e-01 -7.16156960e-01 3.47319245e-01
1.79633334e-01 1.84009224e-01 -8.41735303e-01 1.33150780e+00
-1.42748818e-01 4.29491371e-01 3.85900855e-01 -1.16399616e-01
6.61754310e-01 -4.07349080e-01 9.30260897e-01 -3.33141208e-01
-5.60433388e-01 3.38100016e-01 2.56117135e-01 -4.71278101e-01
1.01943471e-01 -4.90532398e-01 -9.20028925e-01 -9.07957315e-01
-8.61379039e-03 -3.62496912e-01 6.52727306e-01 1.16955169e-01
2.09759504e-01 6.83151126e-01 8.01272869e-01 -1.41104984e+00
-8.85945261e-02 -5.98832190e-01 -9.85338628e-01 4.51308370e-01
1.85538352e-01 -7.02028930e-01 -2.07231537e-01 -1.45477086e-01] | [12.323234558105469, -0.5785751342773438] |
61250486-7771-4aca-a25d-f05bad3410fc | two-in-one-a-model-hijacking-attack-against | 2305.07406 | null | https://arxiv.org/abs/2305.07406v1 | https://arxiv.org/pdf/2305.07406v1.pdf | Two-in-One: A Model Hijacking Attack Against Text Generation Models | Machine learning has progressed significantly in various applications ranging from face recognition to text generation. However, its success has been accompanied by different attacks. Recently a new attack has been proposed which raises both accountability and parasitic computing risks, namely the model hijacking attack. Nevertheless, this attack has only focused on image classification tasks. In this work, we broaden the scope of this attack to include text generation and classification models, hence showing its broader applicability. More concretely, we propose a new model hijacking attack, Ditto, that can hijack different text classification tasks into multiple generation ones, e.g., language translation, text summarization, and language modeling. We use a range of text benchmark datasets such as SST-2, TweetEval, AGnews, QNLI, and IMDB to evaluate the performance of our attacks. Our results show that by using Ditto, an adversary can successfully hijack text generation models without jeopardizing their utility. | ['Ahmed Salem', 'Yang Zhang', 'Michael Backes', 'Wai Man Si'] | 2023-05-12 | null | null | null | null | ['face-recognition', 'text-summarization'] | ['computer-vision', 'natural-language-processing'] | [ 6.02406085e-01 -8.67871046e-02 4.87999544e-02 -4.64854799e-02
-6.43503010e-01 -9.32067394e-01 1.07704413e+00 -9.94124785e-02
-2.67131776e-01 5.92296958e-01 -2.49593914e-01 -7.53788531e-01
3.66317213e-01 -7.82016456e-01 -5.30106425e-01 -7.01649010e-01
2.89140910e-01 1.70270130e-01 1.76817775e-01 -6.57151104e-04
3.96270663e-01 5.41890621e-01 -9.42369998e-01 3.78042787e-01
5.79166114e-01 7.97796488e-01 -5.74604988e-01 7.32068300e-01
2.76768319e-02 7.82608867e-01 -1.00912976e+00 -1.09889746e+00
1.66127041e-01 -3.25863510e-01 -9.37556744e-01 -3.10785115e-01
3.55292588e-01 -5.23105800e-01 -4.84813392e-01 1.03796661e+00
7.80171156e-01 -4.29823488e-01 5.60862243e-01 -1.67528558e+00
-2.99406648e-01 9.38919425e-01 -7.14864612e-01 -1.66101575e-01
3.04591507e-01 5.49368486e-02 6.89668298e-01 -6.03376627e-01
3.50553364e-01 1.43367636e+00 2.68581033e-01 9.14938152e-01
-1.10426474e+00 -1.36076522e+00 -3.34147692e-01 5.89276515e-02
-1.34314787e+00 -4.18938339e-01 6.46581590e-01 -4.90888506e-01
8.28064978e-01 5.44853091e-01 -5.68912104e-02 1.60015929e+00
5.65961063e-01 9.58684742e-01 1.22033954e+00 -4.33304757e-01
2.54328310e-01 2.06740886e-01 8.28248188e-02 5.50916910e-01
5.59710145e-01 1.52067617e-01 -5.41345239e-01 -7.61456370e-01
2.05731332e-01 -3.91953975e-01 -1.49579436e-01 2.94053741e-02
-8.84163022e-01 1.07282662e+00 -2.44312420e-01 6.55160770e-02
2.45549187e-01 2.87277400e-01 6.65134609e-01 2.11807474e-01
4.39598203e-01 2.87661761e-01 -2.31626600e-01 6.84158802e-02
-8.94235730e-01 8.14318657e-02 9.83711362e-01 5.80118716e-01
4.24244344e-01 2.23862812e-01 -1.28575966e-01 2.66116023e-01
1.72295868e-01 8.26091349e-01 6.47366464e-01 -2.95500070e-01
6.49941146e-01 2.05185667e-01 -3.74014467e-01 -1.12132192e+00
-2.20879823e-01 -1.22833312e-01 -1.13204312e+00 -1.93107158e-01
7.12527633e-02 -5.36985338e-01 -7.99582720e-01 1.68217826e+00
2.28192046e-01 3.56004506e-01 8.44609439e-02 2.62711614e-01
6.27652168e-01 5.63227117e-01 1.20977961e-01 -3.22150290e-02
1.24261379e+00 -7.95682669e-01 -3.04030657e-01 -6.89124912e-02
9.49849784e-01 -9.49066341e-01 6.32728994e-01 6.67503774e-01
-8.31482828e-01 -1.64307103e-01 -9.50530767e-01 2.23586157e-01
-4.92317736e-01 -3.86437848e-02 6.09386504e-01 1.36750960e+00
-6.67413712e-01 2.64821559e-01 -8.08655262e-01 -4.16413814e-01
5.56280792e-01 5.48923254e-01 -5.04228830e-01 1.40832156e-01
-1.30414295e+00 6.30536497e-01 5.44636369e-01 -3.33828747e-01
-8.23496759e-01 -3.33337575e-01 -6.06998384e-01 6.17395900e-02
3.61024797e-01 -8.39703143e-01 1.17104089e+00 -5.75287938e-01
-1.65922606e+00 8.25191498e-01 1.22940041e-01 -8.84071410e-01
5.88665903e-01 -1.21463068e-01 -4.33286607e-01 1.07635789e-01
-2.31960282e-01 4.93050188e-01 1.23255789e+00 -9.02686596e-01
-3.01333606e-01 -3.36625665e-01 -1.11769121e-02 -3.07336539e-01
-8.74261081e-01 5.18831670e-01 -1.17723364e-02 -8.13968897e-01
-5.52454293e-01 -1.16525817e+00 5.23527153e-02 -4.90172714e-01
-1.09165800e+00 -1.28494605e-01 1.18382072e+00 -3.49498123e-01
1.42242396e+00 -2.00640702e+00 -4.27751318e-02 2.32174709e-01
1.31470755e-01 7.94762552e-01 -1.88053340e-01 7.69835353e-01
-8.05393793e-03 8.44838977e-01 -2.40812272e-01 -3.87125343e-01
-2.48609763e-02 4.92658839e-02 -9.39889550e-01 2.96365947e-01
1.34141773e-01 9.71382618e-01 -2.89166212e-01 -4.13693726e-01
-2.66531743e-02 2.88037866e-01 -6.25929117e-01 -1.08671285e-01
-2.33029336e-01 6.56622946e-02 -6.43721640e-01 4.33973700e-01
8.21919024e-01 -2.08198935e-01 1.31521061e-01 1.42759845e-01
3.22438389e-01 4.21867743e-02 -7.70360053e-01 1.00749445e+00
-3.18975002e-01 5.92051983e-01 -2.74180651e-01 -7.62403786e-01
5.70869625e-01 3.82028192e-01 1.56985372e-01 -2.95560569e-01
4.49536741e-01 -3.96075211e-02 1.31031955e-02 -3.33252370e-01
4.80614930e-01 8.21283236e-02 -3.35452080e-01 8.92139554e-01
-1.17021553e-01 -9.95441452e-02 -1.37339355e-02 4.95694429e-01
9.15522873e-01 -2.61990577e-01 3.43613058e-01 1.08971335e-01
6.62556171e-01 -2.58335292e-01 -1.00836433e-01 1.05236709e+00
1.68674905e-02 3.13399315e-01 6.22244775e-01 -1.47520155e-01
-5.61135292e-01 -8.29487324e-01 5.71698435e-02 9.96105194e-01
-8.50067735e-02 -8.43022048e-01 -1.00476432e+00 -1.23170316e+00
-3.42771634e-02 6.84872568e-01 -4.61419940e-01 -6.23684049e-01
-4.74067718e-01 -1.06586230e+00 1.54727030e+00 2.18058214e-01
8.10433984e-01 -8.55350614e-01 -3.91973913e-01 -4.70829345e-02
-3.38727862e-01 -1.37607837e+00 -5.69792628e-01 -3.51338863e-01
-5.77476799e-01 -9.30795252e-01 -3.42079163e-01 -3.94161195e-01
4.77712810e-01 1.62104309e-01 7.65519798e-01 1.16439745e-01
-4.12786275e-01 2.02942044e-01 -1.46478727e-01 -6.93845689e-01
-9.48767781e-01 4.79545891e-01 -3.45316492e-02 2.86917508e-01
1.69748470e-01 -2.55168825e-01 -1.98184341e-01 3.68920088e-01
-1.54893875e+00 7.16770589e-02 5.29413760e-01 7.48855531e-01
-1.47777840e-01 2.32876688e-01 6.48535311e-01 -1.19718575e+00
7.41727293e-01 -4.33322072e-01 -6.32892370e-01 4.71585602e-01
-5.81839561e-01 -1.43121500e-02 8.13177049e-01 -7.54688680e-01
-9.34433818e-01 -1.28907189e-01 -2.85258323e-01 -1.70506120e-01
-2.68294156e-01 4.81178701e-01 -2.08674833e-01 -2.55383372e-01
6.43714011e-01 5.80033123e-01 -5.15887290e-02 -2.20696926e-01
2.60408372e-01 8.63379061e-01 2.14931220e-01 -4.51183736e-01
1.33281887e+00 4.77769732e-01 2.28634164e-01 -1.00126231e+00
-4.87263709e-01 1.00250244e-02 4.34619784e-02 7.78918415e-02
5.89567900e-01 -5.99567890e-01 -8.87927651e-01 1.13056648e+00
-1.28921092e+00 -6.35908395e-02 5.09209871e-01 1.59512922e-01
-1.77070215e-01 8.84975791e-01 -7.13730037e-01 -6.11022234e-01
-9.43991363e-01 -1.16335571e+00 8.92290890e-01 -4.78897579e-02
-1.99125931e-01 -8.53136063e-01 -1.27970859e-01 5.98552823e-01
4.72027510e-01 3.42299432e-01 1.03922153e+00 -1.23278701e+00
-3.81648242e-01 -4.46914971e-01 -1.05073705e-01 3.56757879e-01
-6.01894862e-04 2.44724631e-01 -1.04703343e+00 -5.43835223e-01
-7.00612068e-02 -4.68191028e-01 7.97649682e-01 -2.62782753e-01
1.27569735e+00 -7.62262166e-01 -3.37875009e-01 5.16173065e-01
1.08996069e+00 1.93946555e-01 8.62626135e-01 2.01306120e-01
6.44232333e-01 4.67570424e-01 1.93856701e-01 4.53375846e-01
-1.81966333e-03 7.80359805e-01 3.83612394e-01 2.96317413e-02
4.18208897e-01 -2.13229612e-01 6.10991538e-01 3.27314645e-01
4.29811031e-01 -7.35199332e-01 -7.56607890e-01 -1.02855489e-01
-1.55701232e+00 -9.82707262e-01 2.65800059e-02 2.07946491e+00
8.75872612e-01 2.95752674e-01 6.38185516e-02 9.92468223e-02
5.28403878e-01 1.46636471e-01 -3.83037776e-01 -6.40376985e-01
5.92771657e-02 3.90427262e-01 4.84350085e-01 1.67715162e-01
-1.15158689e+00 1.31081772e+00 5.58244848e+00 1.26107645e+00
-1.46590054e+00 -6.59685731e-02 7.69589901e-01 7.92546105e-03
2.94770133e-02 -5.01384512e-02 -9.85706985e-01 5.46561658e-01
8.72290969e-01 -6.12854064e-01 3.66149336e-01 8.33887458e-01
-8.02972987e-02 2.20299691e-01 -8.76348913e-01 8.62349153e-01
3.46039206e-01 -1.31122994e+00 6.41268432e-01 2.57837236e-01
4.81584787e-01 -1.85665712e-01 2.96583444e-01 3.45151752e-01
3.69325966e-01 -1.16761482e+00 3.73243749e-01 -2.03048557e-01
7.48358786e-01 -9.84066427e-01 6.85597241e-01 5.33554852e-01
-5.64963102e-01 1.20655693e-01 -1.32011324e-01 2.35562995e-01
-8.00189227e-02 2.88068622e-01 -1.07393754e+00 8.19999993e-01
1.68767646e-01 7.93204755e-02 -8.36491108e-01 5.43720901e-01
-1.94409773e-01 9.37427759e-01 -3.16181421e-01 2.62537114e-02
4.80642468e-02 2.81368464e-01 4.69828159e-01 1.11109531e+00
2.60955215e-01 -1.90142870e-01 8.27968791e-02 5.45485556e-01
-3.70521516e-01 1.17863476e-01 -8.68224621e-01 -4.05394733e-01
3.64744335e-01 1.30736816e+00 -5.13224483e-01 -2.79391795e-01
-2.00356677e-01 1.08791733e+00 -1.95910379e-01 1.23940006e-01
-1.12275445e+00 -4.97962445e-01 4.90904689e-01 -8.69667232e-02
-5.53484857e-02 -2.27462828e-01 -1.21249557e-01 -1.38126934e+00
-4.01595682e-01 -1.27254784e+00 5.42811036e-01 -4.01081026e-01
-1.00953567e+00 8.01828504e-01 1.85565740e-01 -6.81479931e-01
-5.64799964e-01 -4.33067977e-01 -6.78319871e-01 6.71761215e-01
-1.24932730e+00 -1.26755393e+00 1.26399949e-01 8.15853655e-01
2.23567501e-01 -4.61749047e-01 7.60927618e-01 2.21423477e-01
-8.98195982e-01 1.28906476e+00 -1.15288816e-01 5.05533755e-01
7.32959330e-01 -6.51276708e-01 7.07569122e-01 1.01527536e+00
2.81956613e-01 6.83292568e-01 5.22791266e-01 -5.95650077e-01
-1.36382401e+00 -1.27279389e+00 8.50126624e-01 -3.20498824e-01
7.72553027e-01 -6.59408212e-01 -6.23443365e-01 8.32402468e-01
1.97565094e-01 -4.70856786e-01 7.39775360e-01 -6.21145070e-01
-6.65839553e-01 1.80808619e-01 -1.17768049e+00 9.59667563e-01
5.17611146e-01 -5.70269942e-01 6.98559955e-02 4.61528331e-01
7.59793699e-01 -2.14631006e-01 -4.11917865e-01 4.15491521e-01
6.45752132e-01 -7.12126315e-01 8.93845320e-01 -8.39213014e-01
4.01358843e-01 1.60464868e-01 -7.73114040e-02 -8.97422314e-01
2.55207896e-01 -1.10008836e+00 -1.93681493e-01 1.56664252e+00
2.83349246e-01 -1.11334288e+00 8.30485046e-01 3.70122403e-01
5.04626930e-01 -4.30048406e-01 -9.28253770e-01 -8.27907145e-01
4.64409679e-01 -4.17827457e-01 8.08780849e-01 9.69053328e-01
-2.03763083e-01 4.92313981e-01 -7.54696012e-01 1.85458779e-01
7.51674891e-01 -1.74877524e-01 1.15084267e+00 -1.03292358e+00
-2.82331973e-01 -4.34510320e-01 -2.10330337e-01 -7.06743240e-01
3.35832983e-01 -1.03733146e+00 -5.14535904e-01 -7.06424892e-01
1.88882083e-01 7.57682770e-02 1.64205581e-01 6.18260622e-01
2.97522377e-02 5.18701971e-01 4.96874750e-01 1.30141214e-01
-1.91810459e-01 2.96876162e-01 8.06342065e-01 -2.38916323e-01
1.55100018e-01 3.05331618e-01 -7.02919602e-01 6.69302344e-01
1.33536351e+00 -7.80303061e-01 -4.08726186e-01 -1.87673464e-01
2.64069110e-01 8.20434988e-02 3.97694141e-01 -7.94691384e-01
1.93568245e-01 -2.19047129e-01 -3.96667793e-02 -3.25292587e-01
1.19677065e-02 -5.81318438e-01 5.48048504e-02 6.91756487e-01
-4.59154993e-01 2.77416222e-02 3.42019767e-01 3.89012158e-01
9.70490649e-02 -2.41253957e-01 7.91569412e-01 1.51878864e-01
-7.35984147e-02 4.44223672e-01 -5.69135904e-01 3.10625858e-03
1.05915260e+00 7.27441087e-02 -8.37623298e-01 -4.30144280e-01
-4.29541469e-01 2.18483582e-02 2.78807819e-01 5.20933151e-01
6.11093104e-01 -8.98919642e-01 -9.06263471e-01 4.92630154e-01
9.87695679e-02 -5.47466576e-01 6.12193160e-02 6.25934362e-01
-3.43408644e-01 6.74287558e-01 6.07654713e-02 -3.78161907e-01
-1.86394787e+00 5.90553701e-01 2.63471067e-01 -6.70250118e-01
-2.18819857e-01 4.02807653e-01 4.58387703e-01 -2.95349151e-01
1.92368105e-01 -9.00567025e-02 -1.38567509e-02 -1.14437446e-01
7.43575215e-01 2.71589339e-01 1.28642187e-01 -4.19694573e-01
-3.21255773e-01 2.38249302e-01 -4.04324174e-01 -8.48562568e-02
5.42765677e-01 8.56729597e-02 -3.24931622e-01 -3.17720830e-01
1.20897722e+00 2.69560993e-01 -4.22041804e-01 -2.32603759e-01
1.16005115e-01 -2.73831040e-01 -2.07077309e-01 -8.42756927e-01
-9.64503944e-01 1.08747315e+00 2.15624779e-01 5.97646356e-01
1.06217265e+00 -2.49253049e-01 1.11626327e+00 4.94976640e-01
4.56297666e-01 -5.88014245e-01 3.51272583e-01 5.61528444e-01
6.41761482e-01 -9.59193230e-01 -2.06598714e-01 -4.59160566e-01
-5.90368748e-01 1.03554034e+00 5.87983549e-01 3.77930373e-01
4.98596489e-01 4.76314634e-01 -1.69387758e-01 1.63847789e-01
-8.96290123e-01 3.90679389e-01 4.03377004e-02 3.83821338e-01
1.31898457e-02 7.51418918e-02 -1.60653651e-01 4.25591409e-01
-3.36443573e-01 8.55446458e-02 9.60578501e-01 8.87519121e-01
-1.09828658e-01 -1.41095924e+00 -5.54427981e-01 3.12516510e-01
-1.14108026e+00 -3.11837375e-01 -8.25746059e-01 6.94321573e-01
-2.96495914e-01 1.20682049e+00 -4.15382564e-01 -7.81200528e-01
-1.75632060e-01 1.69686511e-01 2.69040763e-01 -5.22596717e-01
-8.60281229e-01 -1.44748524e-01 1.58114269e-01 -1.39977217e-01
-3.46205868e-02 -2.19103605e-01 -9.27660286e-01 -9.54207242e-01
-5.37263393e-01 1.99700356e-01 6.66899323e-01 7.96219110e-01
4.62115377e-01 2.04223003e-02 8.58544409e-01 -2.39045218e-01
-1.02679765e+00 -7.78038502e-01 -2.97514170e-01 8.33929181e-02
-2.50324439e-02 -1.37266383e-01 -5.13918042e-01 -3.97714712e-02] | [5.918694972991943, 7.799533367156982] |
884dc1a9-45f8-41c7-966d-0b3cf79998f5 | margin-based-parallel-corpus-mining-with | 1811.01136 | null | https://arxiv.org/abs/1811.01136v2 | https://arxiv.org/pdf/1811.01136v2.pdf | Margin-based Parallel Corpus Mining with Multilingual Sentence Embeddings | Machine translation is highly sensitive to the size and quality of the training data, which has led to an increasing interest in collecting and filtering large parallel corpora. In this paper, we propose a new method for this task based on multilingual sentence embeddings. In contrast to previous approaches, which rely on nearest neighbor retrieval with a hard threshold over cosine similarity, our proposed method accounts for the scale inconsistencies of this measure, considering the margin between a given sentence pair and its closest candidates instead. Our experiments show large improvements over existing methods. We outperform the best published results on the BUCC mining task and the UN reconstruction task by more than 10 F1 and 30 precision points, respectively. Filtering the English-German ParaCrawl corpus with our approach, we obtain 31.2 BLEU points on newstest2014, an improvement of more than one point over the best official filtered version. | ['Mikel Artetxe', 'Holger Schwenk'] | 2018-11-03 | margin-based-parallel-corpus-mining-with-1 | https://aclanthology.org/P19-1309 | https://aclanthology.org/P19-1309.pdf | acl-2019-7 | ['parallel-corpus-mining', 'cross-lingual-bitext-mining'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.17007607e-01 -6.33126870e-02 -3.41978043e-01 -3.54729265e-01
-1.41326439e+00 -7.90196776e-01 1.12626970e+00 8.77957404e-01
-1.05318475e+00 1.06233776e+00 4.79942411e-01 -5.20850718e-01
-1.08588412e-01 -5.10764182e-01 -8.70167613e-01 -6.46517098e-01
2.72234589e-01 8.14037979e-01 3.13916445e-01 -4.42765921e-01
4.32680279e-01 1.62489846e-01 -1.26838696e+00 2.82905549e-01
1.13780570e+00 6.86464131e-01 2.43161172e-01 3.97598267e-01
-1.58293560e-01 -2.66770143e-02 -5.40504932e-01 -9.05666053e-01
1.69083104e-01 -3.49234700e-01 -9.08021927e-01 -3.65006387e-01
6.20371282e-01 2.96002001e-01 -1.32825668e-03 1.23541415e+00
6.24533296e-01 4.04999629e-02 5.14935076e-01 -5.40006459e-01
-6.74856186e-01 8.61608624e-01 -3.26549888e-01 3.76318991e-01
6.02485240e-01 -4.28137094e-01 1.33186495e+00 -1.13012445e+00
9.55688119e-01 9.35291767e-01 5.79694092e-01 3.56840819e-01
-1.27465332e+00 -3.93806100e-01 -9.87067521e-02 5.51764011e-01
-1.23882663e+00 -4.18412894e-01 2.91399032e-01 -2.57482499e-01
1.17407823e+00 3.33728582e-01 4.24147695e-01 1.12578523e+00
3.64916772e-01 5.11415184e-01 1.24849403e+00 -1.08390284e+00
1.21667936e-01 3.19301218e-01 2.36737683e-01 2.69230306e-01
3.62645864e-01 -1.96248218e-01 -4.05642122e-01 -2.77544260e-01
1.46694511e-01 -1.83573499e-01 -3.68821830e-01 -3.35297436e-01
-1.61094236e+00 8.71921301e-01 3.94292511e-02 7.42192149e-01
-2.19686568e-01 -2.95022488e-01 4.70226258e-01 6.11824393e-01
7.05868840e-01 7.34227538e-01 -7.59181321e-01 -5.45227230e-01
-9.30881977e-01 2.51326293e-01 7.46157765e-01 8.38921726e-01
5.90013027e-01 -6.71871305e-01 -5.84177207e-03 1.07190883e+00
5.36783179e-03 5.61895192e-01 7.54990816e-01 -5.08825481e-01
7.93345094e-01 4.97020066e-01 8.26814920e-02 -6.93686485e-01
-3.78682651e-02 -4.96782273e-01 -5.31946957e-01 -2.92593807e-01
6.70252442e-01 -8.66294205e-02 -5.16005993e-01 1.61416197e+00
2.77155221e-01 -3.27597916e-01 1.30892783e-01 7.81566739e-01
2.76195824e-01 6.54146075e-01 -5.25688708e-01 -5.62613904e-01
1.33123243e+00 -1.05756569e+00 -7.77134418e-01 -8.95907283e-02
8.05433095e-01 -1.49361122e+00 1.08086014e+00 4.73528922e-01
-9.60428298e-01 -3.44844848e-01 -1.18417621e+00 3.51042151e-02
-4.32663649e-01 5.84240910e-03 3.47039938e-01 6.95927322e-01
-8.73547912e-01 9.39103723e-01 -4.66926396e-01 -6.37847185e-01
-2.18120366e-01 3.35356891e-01 -5.79058766e-01 -4.18865412e-01
-1.34383249e+00 1.20960951e+00 2.90064216e-01 -9.89900678e-02
2.83116177e-02 -4.40010875e-01 -5.66697478e-01 -1.67669967e-01
2.88667858e-01 -1.37752101e-01 9.68572140e-01 -7.67116308e-01
-1.43745828e+00 7.12371409e-01 -3.26233447e-01 -6.26706004e-01
5.16621470e-01 -3.73303860e-01 -6.67619407e-01 -9.98465940e-02
1.01599820e-01 3.12372684e-01 4.03552294e-01 -6.60556018e-01
-5.31946480e-01 -3.57849479e-01 -5.63584231e-02 7.61702657e-02
-5.23396671e-01 2.29164571e-01 -4.80055898e-01 -6.07131302e-01
8.30657706e-02 -1.02286649e+00 -2.29493544e-01 -5.70944905e-01
5.14286458e-02 -5.41298091e-01 2.81744421e-01 -6.96507156e-01
1.29675341e+00 -1.89268959e+00 3.34418833e-01 1.42020322e-02
-2.58198977e-01 3.96294147e-01 -3.70532960e-01 8.43446553e-01
-5.51750977e-03 3.99546251e-02 -2.18314886e-01 -2.49357566e-01
1.57748550e-01 1.71154425e-01 -2.73870051e-01 5.87358773e-01
1.02489330e-01 5.97046912e-01 -9.53868210e-01 -4.23913151e-01
2.76616053e-03 3.62022519e-01 -4.18502420e-01 -1.97947577e-01
3.11285667e-02 6.83647841e-02 -5.45357689e-02 3.01604897e-01
6.20239019e-01 1.60842806e-01 4.15833801e-01 1.30044639e-01
-2.86358148e-01 8.72514367e-01 -1.04995346e+00 1.96350300e+00
-5.64904571e-01 6.79631054e-01 -3.66778344e-01 -8.11661601e-01
9.75831807e-01 4.71470267e-01 2.86497474e-01 -8.12068105e-01
7.43028522e-02 6.70157373e-01 2.45533764e-01 -1.49914727e-01
6.77760780e-01 -4.73844856e-02 -6.01533614e-02 3.62030983e-01
3.22439730e-01 -7.73582533e-02 7.16768682e-01 1.86160997e-01
9.90655363e-01 4.50833365e-02 2.76947290e-01 -7.16180265e-01
7.32535720e-01 2.94532459e-02 5.93744338e-01 4.61124450e-01
-1.58990383e-01 7.19575882e-01 4.38651681e-01 -5.49664676e-01
-1.22463858e+00 -9.12905514e-01 -5.02503872e-01 7.50068486e-01
-9.39448625e-02 -8.68335485e-01 -8.18342745e-01 -9.80216146e-01
-2.11433396e-01 7.91723251e-01 -5.62531054e-01 9.52891558e-02
-7.95106888e-01 -9.94018734e-01 2.41231561e-01 1.10144027e-01
-2.84306332e-02 -6.92807734e-01 1.68787152e-01 3.27121466e-01
-5.67566931e-01 -1.11876667e+00 -7.45972514e-01 1.22505262e-01
-9.66723502e-01 -1.06583619e+00 -7.34623194e-01 -8.30194473e-01
3.58050793e-01 4.58095632e-02 1.17522168e+00 -3.08178693e-01
8.76985267e-02 -2.71363676e-01 -6.69191957e-01 -2.53935277e-01
-5.48708439e-01 4.62967008e-01 4.01502848e-01 -2.24973053e-01
8.49861920e-01 -3.29432249e-01 -3.26565951e-01 1.49614379e-01
-6.37928545e-01 -3.11258197e-01 6.99211538e-01 1.13787889e+00
5.68544209e-01 -4.75218266e-01 5.78893781e-01 -7.81410456e-01
7.52315402e-01 -3.03717971e-01 -5.95530152e-01 4.40611720e-01
-1.19614506e+00 3.87432307e-01 5.95979631e-01 -4.33805645e-01
-7.08832383e-01 -2.94607785e-02 -2.24492013e-01 2.33931035e-01
2.33778693e-02 3.14722508e-01 5.57830967e-02 9.46204141e-02
7.02553928e-01 3.26690197e-01 -1.88418955e-01 -6.79329097e-01
4.08418626e-01 8.82202864e-01 2.05956146e-01 -5.10832012e-01
6.50895655e-01 -3.48342881e-02 -3.16567421e-01 -5.83769441e-01
-6.26229644e-01 -6.25625074e-01 -8.55843425e-01 1.22868910e-01
6.04817092e-01 -6.96688473e-01 -1.73249781e-01 -9.65840816e-02
-1.38154697e+00 3.15883368e-01 -1.39587581e-01 1.06123209e+00
-3.88946146e-01 6.55831814e-01 -7.12961614e-01 -3.90033841e-01
-4.53453034e-01 -1.27158654e+00 8.56674135e-01 -3.02255392e-01
-3.84472460e-01 -9.20893133e-01 6.42292857e-01 5.64007461e-01
3.59100729e-01 -1.77669585e-01 8.88247430e-01 -1.01409221e+00
-1.56453922e-01 -4.02968079e-01 -3.31316292e-02 4.62544262e-01
1.02184683e-01 -3.43194544e-01 -6.38060808e-01 -4.46323514e-01
5.49105406e-02 -1.98585078e-01 7.29143322e-01 -9.38127935e-02
3.22461039e-01 -1.64446309e-01 -2.07932979e-01 4.16831821e-02
1.50526857e+00 -6.44399524e-02 5.86678326e-01 6.10275269e-01
1.78582177e-01 4.74997371e-01 7.50187576e-01 -7.12481653e-03
1.08019464e-01 8.65944743e-01 6.30052686e-02 1.46710873e-01
-9.29981545e-02 -1.76086128e-01 5.49283445e-01 1.56721115e+00
-1.48663536e-01 -1.15385145e-01 -7.13405728e-01 7.62637734e-01
-1.77347767e+00 -7.64400840e-01 -2.93792993e-01 2.48278236e+00
1.18427014e+00 3.67340863e-01 -2.07643852e-01 1.82679385e-01
6.11705124e-01 5.80168106e-02 1.31341115e-01 -8.77353132e-01
-3.04793715e-01 5.40197551e-01 6.80603564e-01 7.26390719e-01
-8.57905984e-01 9.85697389e-01 6.35950279e+00 1.09445310e+00
-8.56420338e-01 3.34192455e-01 8.56209099e-02 -1.33853495e-01
-3.73110741e-01 1.47816002e-01 -9.03195322e-01 4.36685741e-01
1.20634449e+00 -2.03006700e-01 4.04995650e-01 4.84720856e-01
5.55710644e-02 -3.69040780e-02 -1.06685483e+00 7.06592560e-01
4.54445750e-01 -1.23676705e+00 -7.00618252e-02 1.97549328e-01
9.51410353e-01 4.17061508e-01 -3.37900698e-01 2.55192757e-01
2.36153543e-01 -6.69041395e-01 5.08026659e-01 1.82575777e-01
5.41108847e-01 -8.80125403e-01 1.13052905e+00 2.75409073e-01
-7.17484295e-01 2.83463985e-01 -5.27498424e-01 -1.35783896e-01
1.29532903e-01 8.94024968e-01 -8.09576809e-01 8.22942853e-01
5.74305654e-01 5.43762803e-01 -5.47489226e-01 1.07316971e+00
-2.96874404e-01 6.72236204e-01 -3.26159954e-01 -3.84025395e-01
2.96032369e-01 -4.39059228e-01 5.31764150e-01 1.28668928e+00
5.07457495e-01 -5.64273596e-01 -1.04874291e-01 1.27916574e-01
-1.97062552e-01 8.26844096e-01 -4.34127778e-01 4.43129390e-02
2.70783663e-01 1.04149401e+00 -3.58195573e-01 -4.19058502e-01
-5.61511159e-01 1.23778749e+00 5.31146526e-01 -7.08239386e-03
-6.36980355e-01 -6.69511080e-01 6.12182796e-01 -1.44195497e-01
4.60709304e-01 -3.32767457e-01 -2.97514088e-02 -1.34449685e+00
5.53805411e-01 -1.07492995e+00 1.38213858e-01 -8.97147059e-02
-1.29238141e+00 9.81114149e-01 -3.24174404e-01 -1.30102611e+00
-3.61185312e-01 -6.11188889e-01 -2.34480798e-01 8.98174047e-01
-1.52845013e+00 -8.09538960e-01 6.17757201e-01 2.12412700e-01
7.47913599e-01 -1.86121091e-01 1.24552655e+00 5.55764079e-01
-2.14805424e-01 6.97590411e-01 8.53385329e-01 -6.28692657e-02
1.35870147e+00 -1.22689152e+00 5.08617699e-01 9.59129572e-01
6.82343960e-01 7.81879187e-01 9.10032988e-01 -4.37562227e-01
-1.29266775e+00 -7.03955710e-01 2.14048553e+00 -6.03029251e-01
1.03502977e+00 -6.71809018e-01 -6.86317205e-01 3.81748468e-01
6.87574387e-01 -3.42989057e-01 6.83188975e-01 4.57944751e-01
-6.55619860e-01 -1.61071122e-01 -9.19471383e-01 5.43729663e-01
9.19311702e-01 -3.77152145e-01 -1.08522534e+00 6.71130598e-01
7.18855977e-01 -1.02363691e-01 -1.05994904e+00 3.68985564e-01
5.75619578e-01 -7.75585532e-01 6.74820602e-01 -5.12319148e-01
3.98242772e-01 -1.14276610e-01 -3.81613493e-01 -1.66533530e+00
-2.89898038e-01 -4.39098179e-01 2.87214339e-01 1.13867223e+00
8.58602285e-01 -5.66401124e-01 4.67930734e-01 -4.69106399e-02
-1.29349887e-01 -7.11338878e-01 -1.23856592e+00 -1.03413379e+00
5.93640983e-01 -3.45829755e-01 5.67632437e-01 9.64324236e-01
4.06250954e-01 5.77045500e-01 -2.89464116e-01 -1.40490517e-01
5.25559187e-01 1.65682539e-01 4.45751160e-01 -1.31115329e+00
-3.42582822e-01 -5.06518424e-01 -2.75516927e-01 -1.01298964e+00
1.30144879e-01 -1.07308459e+00 -7.17206672e-02 -1.39441144e+00
2.58865833e-01 -9.66434255e-02 -3.71150762e-01 -5.66508844e-02
-1.57696143e-01 4.71247822e-01 4.94210385e-02 1.26368582e-01
-5.99493504e-01 4.96570379e-01 1.01836383e+00 -1.24109268e-01
2.17702627e-01 -2.13931367e-01 -4.18542027e-01 4.38733399e-01
7.89836347e-01 -6.27000272e-01 8.33878443e-02 -8.16174269e-01
3.80310953e-01 -2.93764621e-01 -3.28880638e-01 -8.74287784e-01
1.81954548e-01 -2.08546948e-02 9.93436724e-02 -5.33458531e-01
2.55483419e-01 -7.14023113e-01 -2.22524986e-01 4.61863756e-01
-3.91879052e-01 4.80435580e-01 4.28114422e-02 5.50211549e-01
-4.51694101e-01 -5.08966625e-01 4.07736450e-01 1.11016400e-01
-1.90050602e-01 -1.41474158e-01 -4.85866964e-01 3.41404751e-02
6.99918866e-01 3.24686855e-01 -2.54282117e-01 -6.18960224e-02
-4.94124383e-01 3.18870023e-02 5.40319562e-01 8.02433848e-01
2.79305667e-01 -1.46004426e+00 -1.02522624e+00 1.41210303e-01
4.29372728e-01 -6.62006259e-01 -2.06302881e-01 1.05147898e+00
-3.45992655e-01 7.24828660e-01 1.19677000e-01 -2.91010737e-01
-1.44130886e+00 5.23316503e-01 -1.50844738e-01 -5.37051678e-01
-3.46366316e-01 4.09709752e-01 -5.66097796e-01 -7.10100234e-01
6.03465661e-02 -2.83090055e-01 -8.23140442e-02 8.00379813e-02
6.46409392e-01 3.19253117e-01 6.83916867e-01 -7.18097627e-01
-4.92707342e-01 6.08195126e-01 -3.41882586e-01 -3.47214818e-01
1.37581956e+00 -8.88733789e-02 -3.59825492e-01 5.63166440e-01
1.32260072e+00 5.86021125e-01 -4.45641160e-01 -4.77250516e-01
4.54970747e-01 -6.04607344e-01 -1.79787546e-01 -8.41138005e-01
-5.14926970e-01 7.96940863e-01 5.83516181e-01 7.20732063e-02
6.74869776e-01 2.05073338e-02 9.42686796e-01 5.59507012e-01
4.88637418e-01 -1.34794855e+00 -4.02382284e-01 7.88297951e-01
7.52297044e-01 -1.41193342e+00 2.53326651e-02 -1.44215465e-01
-2.39941314e-01 1.26235223e+00 -2.56441757e-02 -2.59612173e-01
3.40855211e-01 7.80210048e-02 2.05533400e-01 2.95331299e-01
-8.20553541e-01 3.04201189e-02 3.64103496e-01 2.95491219e-01
7.84460843e-01 5.77002764e-02 -1.38499999e+00 4.34281379e-01
-2.34863684e-01 -2.08060876e-01 3.68699223e-01 6.74436390e-01
-4.68187392e-01 -2.08447242e+00 -3.42989326e-01 2.22869664e-01
-6.86465323e-01 -4.41984355e-01 -4.61900771e-01 5.99582195e-01
7.99112674e-03 1.05911291e+00 -8.72730762e-02 -3.75458866e-01
4.26512986e-01 3.53943229e-01 6.45070255e-01 -4.40747023e-01
-6.77052081e-01 1.33885592e-01 3.19266975e-01 -3.56136113e-01
-4.14078265e-01 -8.73246610e-01 -7.88852453e-01 -2.88466066e-01
-6.84700429e-01 6.71576440e-01 9.18463647e-01 9.76127625e-01
2.37051889e-01 -4.92873676e-02 6.90114737e-01 -4.70567524e-01
-9.00644779e-01 -1.38319027e+00 -1.55574217e-01 5.69413424e-01
-3.71132717e-02 -3.80011827e-01 -4.38701600e-01 -2.28414878e-01] | [11.373188972473145, 10.23267650604248] |
f0a477bd-1493-4767-a30f-565d7f99a201 | conformal-language-modeling | 2306.10193 | null | https://arxiv.org/abs/2306.10193v1 | https://arxiv.org/pdf/2306.10193v1.pdf | Conformal Language Modeling | We propose a novel approach to conformal prediction for generative language models (LMs). Standard conformal prediction produces prediction sets -- in place of single predictions -- that have rigorous, statistical performance guarantees. LM responses are typically sampled from the model's predicted distribution over the large, combinatorial output space of natural language. Translating this process to conformal prediction, we calibrate a stopping rule for sampling different outputs from the LM that get added to a growing set of candidates until we are confident that the output set is sufficient. Since some samples may be low-quality, we also simultaneously calibrate and apply a rejection rule for removing candidates from the output set to reduce noise. Similar to conformal prediction, we prove that the sampled set returned by our procedure contains at least one acceptable answer with high probability, while still being empirically precise (i.e., small) on average. Furthermore, within this set of candidate responses, we show that we can also accurately identify subsets of individual components -- such as phrases or sentences -- that are each independently correct (e.g., that are not "hallucinations"), again with statistical guarantees. We demonstrate the promise of our approach on multiple tasks in open-domain question answering, text summarization, and radiology report generation using different LM variants. | ['Regina Barzilay', 'Tommi S. Jaakkola', 'Jae Ho Sohn', 'Adam Yala', 'Tal Schuster', 'Adam Fisch', 'Victor Quach'] | 2023-06-16 | null | null | null | null | ['conformal-prediction', 'question-answering', 'text-summarization', 'open-domain-question-answering', 'conformal-prediction'] | ['computer-vision', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'reasoning'] | [ 6.36611760e-01 9.55592096e-01 -1.34102866e-01 -4.18722600e-01
-1.84886742e+00 -7.50988305e-01 3.43470007e-01 3.66436541e-01
-4.46303003e-02 9.20487642e-01 6.87124074e-01 -1.81750521e-01
-1.24667943e-01 -8.06405842e-01 -1.00914693e+00 -6.08467340e-01
1.73201159e-01 1.08459294e+00 1.01935789e-01 1.31816730e-01
1.53494298e-01 -3.99911664e-02 -1.08987522e+00 5.42470396e-01
1.13089013e+00 5.63351095e-01 1.09570585e-01 1.01457763e+00
-1.03490613e-01 8.73215199e-01 -7.14886785e-01 -6.21893406e-01
-2.12367594e-01 -9.00134981e-01 -9.18331385e-01 -2.94845030e-02
3.36858600e-01 -6.14665821e-03 1.18567623e-01 9.76001263e-01
4.88831967e-01 -6.19535595e-02 9.42885220e-01 -9.13706481e-01
-6.47061348e-01 8.72825861e-01 -2.43902117e-01 -1.98443994e-01
6.41438246e-01 -1.36482298e-01 1.16467750e+00 -9.92747605e-01
6.94015443e-01 1.03249753e+00 7.87640274e-01 8.58456314e-01
-1.70853055e+00 -4.80924815e-01 2.30068527e-02 -4.54825610e-01
-1.15051901e+00 -6.04334056e-01 5.16693592e-01 -2.18556970e-01
6.49796367e-01 7.47123539e-01 4.55843300e-01 1.19693410e+00
5.90038240e-01 7.89184451e-01 7.37696528e-01 -4.87354845e-01
5.61452806e-01 4.89228040e-01 1.15109093e-01 5.57747483e-01
2.43052796e-01 -3.43149364e-01 -7.18798339e-01 -7.94207692e-01
3.15791070e-01 -3.33718866e-01 -4.81530875e-01 -7.13536292e-02
-1.16449285e+00 9.94684279e-01 -1.20639637e-01 2.33937278e-02
-3.76328349e-01 1.49463624e-01 -3.84251284e-03 1.17931455e-01
6.90424442e-01 4.82129484e-01 -3.01881671e-01 1.64836466e-01
-1.17718673e+00 3.91592890e-01 1.09197414e+00 1.15295017e+00
4.20770854e-01 -2.78124779e-01 -6.82373643e-01 9.39334750e-01
3.03907525e-02 8.37327480e-01 3.63794655e-01 -8.78942847e-01
3.57039690e-01 1.56700313e-01 2.87278414e-01 -7.95567930e-01
-4.28479701e-01 -6.46213830e-01 -8.86989415e-01 -2.41449535e-01
2.04477817e-01 -1.33805275e-01 -5.12246132e-01 2.23161364e+00
1.11467965e-01 -2.68244222e-02 1.37820899e-01 6.17558897e-01
6.45489156e-01 7.52426565e-01 -2.25968286e-02 -6.65693164e-01
1.23818111e+00 -6.29660428e-01 -4.25549120e-01 -1.46250337e-01
6.36034667e-01 -7.36142814e-01 1.22810543e+00 5.22205114e-01
-1.44252133e+00 -2.49587595e-01 -8.19745660e-01 2.60511011e-01
6.80667341e-01 1.72282130e-01 2.22366989e-01 4.81252223e-01
-1.23970175e+00 5.31422317e-01 -4.38247859e-01 -5.77045269e-02
2.54009217e-01 1.37384906e-01 -2.40559772e-01 3.39172594e-02
-8.57487500e-01 7.65711606e-01 -1.81435719e-02 -4.48676318e-01
-7.22127855e-01 -6.63730025e-01 -4.48458910e-01 2.24673644e-01
2.78810143e-01 -1.11732769e+00 1.58094764e+00 -1.06246924e+00
-1.18443584e+00 5.89327753e-01 -7.52892435e-01 -6.07825756e-01
5.68705738e-01 -2.00356226e-02 -1.81078389e-01 3.32319736e-01
2.48304844e-01 6.82220459e-01 9.23006654e-01 -1.34491372e+00
-4.41289335e-01 1.25467044e-03 -1.79498553e-01 2.75180519e-01
-3.01215291e-01 -2.40108281e-01 -2.69987911e-01 -5.64486086e-01
3.87002379e-01 -1.03357315e+00 -5.28749943e-01 -3.58588934e-01
-8.51410031e-01 -3.91353071e-01 -1.32619739e-01 -5.69188058e-01
1.42916846e+00 -1.75737536e+00 1.05676755e-01 4.42968786e-01
4.49041307e-01 -3.96333963e-01 -1.92916662e-01 3.54369938e-01
-6.44795876e-03 4.30273980e-01 -5.13014197e-01 -6.17745638e-01
-6.76689222e-02 1.33455694e-01 -8.47601831e-01 2.88464487e-01
2.47391686e-01 1.01063144e+00 -8.06711376e-01 -6.65090024e-01
-2.97179043e-01 -1.10578939e-01 -9.93693650e-01 1.71052203e-01
-7.85506368e-01 3.96757096e-01 -5.13566256e-01 2.04736486e-01
4.73948479e-01 -5.68837523e-01 1.69916138e-01 2.78817445e-01
5.15182853e-01 4.66468930e-01 -1.18234956e+00 1.25183618e+00
-5.14489472e-01 4.55632418e-01 -3.64324123e-01 -7.88137436e-01
7.90717959e-01 4.60152149e-01 3.22417289e-01 -1.88160092e-01
-2.76093513e-01 2.60881037e-01 -1.60697997e-01 -2.82044470e-01
4.79449183e-01 -4.65185314e-01 -2.54169166e-01 8.03848803e-01
9.11081880e-02 -3.32062870e-01 1.30999386e-02 5.68078518e-01
1.20174956e+00 -1.92827284e-01 3.04060221e-01 -2.60782868e-01
4.07521218e-01 1.16212860e-01 5.80400884e-01 1.33311975e+00
3.34099919e-01 1.00306559e+00 9.60545123e-01 1.64154932e-01
-1.28893435e+00 -1.31458354e+00 -2.49774814e-01 7.34515131e-01
-2.02864856e-01 -6.23152137e-01 -9.63987708e-01 -5.75086415e-01
-2.91607946e-01 1.15827584e+00 -5.50859153e-01 -2.05411553e-01
-4.26206350e-01 -5.89395821e-01 4.21478868e-01 3.87063593e-01
-4.98820901e-01 -1.12220919e+00 -3.23462039e-01 4.45433408e-01
-4.56348002e-01 -7.07263410e-01 -7.22329319e-01 1.91629931e-01
-8.57973278e-01 -9.32147324e-01 -5.78953505e-01 -7.18443394e-01
1.00142539e+00 -1.05039172e-01 1.53988159e+00 -4.66236174e-02
2.60543704e-01 5.52810311e-01 -2.78625160e-01 -3.10804158e-01
-1.05579352e+00 -3.81658524e-02 -7.47679099e-02 -8.09471458e-02
-4.14700955e-02 -3.65696967e-01 -4.90420103e-01 -7.07975999e-02
-7.46595502e-01 3.17263663e-01 5.95839858e-01 1.02868819e+00
9.62318897e-01 -3.39148045e-01 9.60391223e-01 -1.18443000e+00
9.86440361e-01 -6.89227283e-01 -2.21057072e-01 4.30582881e-01
-5.51831365e-01 3.69914383e-01 8.88478458e-01 -7.56291926e-01
-7.25037098e-01 -2.20563829e-01 -2.38261268e-01 -2.31903464e-01
-4.51430455e-02 5.63740849e-01 4.48131934e-02 4.90608990e-01
1.06549835e+00 4.69341397e-01 6.16268814e-02 -6.48734421e-02
4.02939141e-01 4.87404525e-01 6.92358136e-01 -6.21963024e-01
6.61499739e-01 2.52203375e-01 -1.28344968e-01 -5.11371791e-01
-1.02569890e+00 -2.14979827e-01 2.89852303e-02 -1.01450376e-01
5.78769743e-01 -6.70542121e-01 -4.95867789e-01 -2.41570041e-01
-1.39862609e+00 2.58740522e-02 -6.61780119e-01 3.20934653e-01
-7.68938541e-01 3.11349809e-01 -4.36213553e-01 -1.16512167e+00
-7.29177237e-01 -9.00784791e-01 1.20785546e+00 6.33465126e-02
-9.04692292e-01 -7.65430987e-01 2.18031183e-01 3.34843621e-02
1.55543625e-01 -1.27964020e-01 1.34991360e+00 -1.07794690e+00
-4.19910997e-01 -4.36615855e-01 4.62319374e-01 2.78002232e-01
-3.20580930e-01 -1.25721827e-01 -8.50523889e-01 -2.59710271e-02
2.85951734e-01 -2.79271007e-01 8.64879072e-01 6.62142694e-01
1.30436265e+00 -7.18945205e-01 -5.32082081e-01 1.00814775e-01
1.05092931e+00 -1.31612435e-01 5.53584039e-01 -4.17761415e-01
2.50800569e-02 5.41513443e-01 2.76221693e-01 5.00206232e-01
1.56967252e-01 3.69305015e-01 -1.37983397e-01 1.09547667e-01
1.73212901e-01 -6.75960541e-01 4.27697033e-01 9.08878207e-01
5.74544609e-01 -6.04721963e-01 -5.58722496e-01 4.84893918e-01
-1.88808692e+00 -1.04682052e+00 -3.26142579e-01 2.53633738e+00
1.22602510e+00 3.84123534e-01 -3.46880383e-03 -3.33161280e-02
5.00398815e-01 -2.25641266e-01 -5.09423256e-01 -4.17525291e-01
-3.52562398e-01 5.89044034e-01 1.00977480e-01 7.50235021e-01
-6.85862303e-01 5.22661269e-01 6.62549305e+00 1.02252913e+00
-5.59888124e-01 2.00977251e-01 1.06884038e+00 -2.52915889e-01
-1.37854266e+00 1.77753106e-01 -7.51838028e-01 4.05702919e-01
1.27447534e+00 -4.20894384e-01 4.79543172e-02 8.09364974e-01
1.59490377e-01 -3.23140144e-01 -1.31596684e+00 5.44431388e-01
2.69830644e-01 -1.71863663e+00 8.66698846e-02 -1.86866537e-01
9.71632063e-01 -3.95688295e-01 1.15689188e-01 2.38873199e-01
3.40241134e-01 -1.14281666e+00 9.59283590e-01 8.04299176e-01
7.90717006e-01 -7.47515082e-01 6.17082775e-01 8.39952052e-01
-6.34893537e-01 5.36389416e-03 -4.73070413e-01 5.60098946e-01
2.08265632e-01 1.00814152e+00 -1.11494076e+00 3.70904356e-01
1.05688117e-01 5.04549071e-02 -1.81721792e-01 9.58884358e-01
-2.89066166e-01 1.09198117e+00 -3.76919568e-01 -5.01163661e-01
-2.44590998e-01 -1.07347384e-01 6.51200891e-01 1.13219833e+00
5.43079436e-01 2.21580476e-01 1.51279464e-01 1.11244369e+00
-1.71803087e-01 3.37031901e-01 -5.58312058e-01 1.80865079e-01
6.99178874e-01 1.05966938e+00 -4.78609830e-01 -4.81999338e-01
-1.37414202e-01 9.09109294e-01 3.22597593e-01 2.40577340e-01
-8.15423608e-01 9.38338265e-02 1.16928875e-01 1.52702600e-01
6.26471778e-03 2.80672818e-01 -7.36766696e-01 -8.90977204e-01
2.84120739e-01 -9.54396009e-01 4.69562292e-01 -9.84579563e-01
-1.38041449e+00 6.40384853e-01 -1.72923565e-01 -1.36341596e+00
-6.29376352e-01 -6.01937920e-02 -6.86100006e-01 1.06360734e+00
-8.03090751e-01 -6.86842203e-01 2.38094389e-01 1.81858972e-01
5.95440090e-01 -1.76900774e-02 1.28206503e+00 -3.79789323e-01
-7.88889453e-03 6.85552180e-01 6.36039823e-02 -3.57314020e-01
5.89482725e-01 -1.29646277e+00 1.67270482e-01 6.61669552e-01
3.38528901e-01 5.68435848e-01 8.18407059e-01 -7.75221705e-01
-1.16681814e+00 -1.03446567e+00 1.32977486e+00 -7.80713856e-01
3.50091070e-01 -3.86160105e-01 -9.03598249e-01 5.84402978e-01
-5.09394407e-02 -5.01289845e-01 8.30298662e-01 1.47913322e-01
-6.71732128e-02 1.82182133e-01 -1.03722680e+00 8.38127255e-01
7.37040579e-01 -3.76312792e-01 -6.68129086e-01 8.46152306e-01
8.81635010e-01 -4.09177065e-01 -5.22561491e-01 3.63873571e-01
2.87832618e-01 -7.18145072e-01 7.04077542e-01 -7.68171966e-01
8.08701992e-01 -5.00624580e-03 -1.88184887e-01 -1.29806352e+00
-2.26519421e-01 -8.67720664e-01 -2.56436884e-01 9.55134153e-01
9.09201980e-01 -4.55896497e-01 8.28110158e-01 6.22510493e-01
-2.40739003e-01 -1.13626313e+00 -1.01938128e+00 -3.87191534e-01
4.18707341e-01 -7.81101525e-01 3.34849954e-01 2.49087647e-01
2.20110551e-01 6.73548937e-01 -4.65562165e-01 1.50593311e-01
5.87349176e-01 3.35450053e-01 7.00694203e-01 -1.15191674e+00
-7.86713362e-01 -3.05000663e-01 1.38260707e-01 -1.17469239e+00
2.18056098e-01 -1.04941010e+00 2.84228921e-01 -1.69299543e+00
7.67155886e-01 -4.52676266e-01 1.23966597e-01 2.30930984e-01
-4.28039283e-01 -7.83235729e-02 -2.35238168e-02 3.51627290e-01
-4.86133516e-01 3.58299434e-01 1.14490569e+00 6.89437322e-04
-1.44217744e-01 4.50973809e-01 -1.14977789e+00 5.75716972e-01
5.91923416e-01 -7.29707479e-01 -5.35709143e-01 1.34443343e-01
4.45166767e-01 6.86161697e-01 2.58803666e-01 -6.94684029e-01
1.91829592e-01 -5.65725826e-02 4.25571054e-01 -7.91518271e-01
9.65120643e-02 -2.61857539e-01 2.37239540e-01 4.93349642e-01
-1.03955090e+00 -9.33475941e-02 -1.73953652e-01 7.93346107e-01
2.48253837e-01 -6.19257867e-01 6.42243147e-01 3.51109654e-02
2.35583469e-01 1.63844883e-01 -4.85249281e-01 5.36008418e-01
7.28810728e-01 1.48783252e-01 -1.99257672e-01 -9.47061837e-01
-7.47592032e-01 2.61843324e-01 1.83066785e-01 7.60108382e-02
7.44724095e-01 -1.13760853e+00 -1.18084514e+00 9.32841524e-02
7.59643540e-02 9.23672915e-02 1.37594029e-01 9.57234323e-01
4.13153805e-02 3.96618426e-01 7.56414413e-01 -6.46836936e-01
-9.38636065e-01 3.68760675e-01 2.93459952e-01 -4.04606313e-01
-8.26025605e-01 9.44776714e-01 3.87496710e-01 -3.63946289e-01
9.76880491e-02 -3.21136326e-01 6.86018169e-02 -2.44532049e-01
4.20361519e-01 -2.14014828e-01 2.45157909e-02 -8.75350833e-02
-1.99221194e-01 4.68768477e-02 9.89391878e-02 -6.40533328e-01
1.06576478e+00 2.61370819e-02 -1.16926499e-01 4.76042897e-01
1.01708615e+00 5.11089981e-01 -9.08756793e-01 -2.91915953e-01
-2.20679492e-03 -1.20474018e-01 -5.02231717e-01 -8.34191203e-01
-3.12717587e-01 4.64924008e-01 1.31464913e-01 3.24789733e-01
9.47018206e-01 6.16437912e-01 5.85798800e-01 2.84738243e-01
3.17430347e-01 -9.85180438e-01 2.27074936e-01 1.98302373e-01
1.10993397e+00 -8.02116871e-01 -2.07930714e-01 -2.07884088e-01
-8.73554647e-01 1.01381946e+00 1.64560288e-01 -1.02479480e-01
3.07530373e-01 5.48235297e-01 -3.04129273e-01 1.92606494e-01
-1.42038834e+00 3.14688027e-01 5.44403911e-01 4.53050852e-01
5.18213809e-01 3.18365961e-01 -4.16922063e-01 1.16986370e+00
-7.04249203e-01 -2.92713642e-01 6.94980919e-01 3.87399048e-01
-8.17830265e-01 -9.35430110e-01 -4.07184601e-01 1.02920353e+00
-3.14283848e-01 -3.58696282e-01 -3.27482909e-01 1.64165139e-01
-3.24943364e-01 1.23262894e+00 -6.34460226e-02 -3.38703662e-01
1.84369490e-01 3.67259264e-01 4.79444295e-01 -7.41268694e-01
-3.78539622e-01 3.21251512e-01 3.44268292e-01 -1.87210038e-01
4.82291505e-02 -8.43730032e-01 -1.37400138e+00 -1.28634468e-01
-4.42784131e-01 4.12139475e-01 2.44570494e-01 1.07444358e+00
1.99656337e-01 3.36141348e-01 7.49393642e-01 -1.38273522e-01
-1.23799324e+00 -9.98101592e-01 -2.21114784e-01 2.59692490e-01
2.06216499e-01 3.13453525e-02 -3.13681841e-01 6.81780875e-02] | [11.992600440979004, 9.056231498718262] |
01ab4758-c134-41e5-8526-7f509c049df5 | connecting-vision-and-language-with-video | 2302.11217 | null | https://arxiv.org/abs/2302.11217v2 | https://arxiv.org/pdf/2302.11217v2.pdf | Connecting Vision and Language with Video Localized Narratives | We propose Video Localized Narratives, a new form of multimodal video annotations connecting vision and language. In the original Localized Narratives, annotators speak and move their mouse simultaneously on an image, thus grounding each word with a mouse trace segment. However, this is challenging on a video. Our new protocol empowers annotators to tell the story of a video with Localized Narratives, capturing even complex events involving multiple actors interacting with each other and with several passive objects. We annotated 20k videos of the OVIS, UVO, and Oops datasets, totalling 1.7M words. Based on this data, we also construct new benchmarks for the video narrative grounding and video question answering tasks, and provide reference results from strong baseline models. Our annotations are available at https://google.github.io/video-localized-narratives/. | ['Vittorio Ferrari', 'Radu Soricut', 'Jordi Pont-Tuset', 'Soravit Changpinyo', 'Paul Voigtlaender'] | 2023-02-22 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Voigtlaender_Connecting_Vision_and_Language_With_Video_Localized_Narratives_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Voigtlaender_Connecting_Vision_and_Language_With_Video_Localized_Narratives_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-narrative-grounding', 'video-question-answering'] | ['computer-vision', 'computer-vision'] | [ 5.26593253e-02 -8.94089118e-02 -4.75973904e-01 -7.55986497e-02
-8.49494874e-01 -1.06414127e+00 7.08117008e-01 6.27069548e-02
-5.28206348e-01 5.17782211e-01 6.30488336e-01 6.19015433e-02
3.61569405e-01 -9.26714912e-02 -8.91576111e-01 -3.59605312e-01
-2.02622637e-01 1.99012533e-02 3.57070416e-01 1.82476759e-01
-2.19973922e-02 -1.41312897e-01 -1.25297916e+00 1.04604018e+00
7.47194290e-02 9.92991984e-01 3.74588400e-01 1.13707173e+00
-1.66711748e-01 1.81912315e+00 -5.14126420e-01 -8.62784445e-01
-7.30533823e-02 -3.32538754e-01 -1.09790027e+00 2.91357100e-01
9.00165796e-01 -7.44392037e-01 -1.03666747e+00 8.16556036e-01
1.32123068e-01 1.86970234e-01 1.16270386e-01 -1.64090562e+00
-7.63759017e-01 8.21862936e-01 -5.80771327e-01 4.29316938e-01
1.17776179e+00 4.07671660e-01 1.13819480e+00 -9.98662472e-01
1.12767148e+00 1.05460835e+00 4.14077669e-01 6.66542470e-01
-7.57189870e-01 -7.07268000e-01 3.46049994e-01 5.18428087e-01
-1.41934276e+00 -6.90386117e-01 5.42265892e-01 -6.92484558e-01
6.23040020e-01 3.97714198e-01 6.34875357e-01 1.80295396e+00
-3.86657745e-01 1.15177071e+00 4.32745755e-01 4.66608442e-02
-1.82184771e-01 -1.24889854e-02 -1.45431384e-02 8.78656566e-01
-3.05303663e-01 -4.30018276e-01 -1.18915868e+00 -6.11495599e-02
6.18608117e-01 2.02933043e-01 -6.89630866e-01 -8.94689262e-02
-1.79839277e+00 5.23625493e-01 7.51010850e-02 1.60589084e-01
-2.16362551e-01 6.74112082e-01 5.72706282e-01 1.02630295e-02
2.21208215e-01 1.50302291e-01 -2.76466999e-02 -6.64503574e-01
-8.43249440e-01 2.49915153e-01 7.33636379e-01 1.26762497e+00
3.98909390e-01 -4.28383231e-01 -4.91427451e-01 3.25647771e-01
2.00626180e-01 5.71509898e-01 -1.95422675e-03 -1.55274928e+00
8.66171122e-01 4.81492877e-01 4.32705045e-01 -1.16241980e+00
-1.49700686e-01 3.71061832e-01 -5.06210923e-01 -6.36880696e-01
5.18071115e-01 -1.71982184e-01 -6.55005157e-01 1.62210929e+00
2.31228396e-01 5.69740772e-01 -2.58954287e-01 1.12365854e+00
1.40080941e+00 8.95230949e-01 3.13451886e-01 -1.93654969e-01
1.74353945e+00 -1.31137168e+00 -1.12389755e+00 -1.15036570e-01
5.89538574e-01 -6.12138093e-01 1.37342465e+00 3.91633391e-01
-1.15328503e+00 -3.81264836e-01 -5.08207321e-01 -3.60213280e-01
-7.97310472e-02 1.17997728e-01 4.11587715e-01 -1.07631102e-01
-9.18214619e-01 3.91237391e-03 -9.10600185e-01 -3.66744876e-01
5.90124071e-01 -2.20347837e-01 -6.39048934e-01 -2.51243472e-01
-9.66070414e-01 3.47330481e-01 1.72700360e-01 2.35244464e-02
-1.32421744e+00 -5.32624245e-01 -9.39828455e-01 -3.33367676e-01
7.81684458e-01 -3.39762360e-01 1.57184207e+00 -9.52646017e-01
-1.03047776e+00 1.04481602e+00 -4.30146486e-01 -3.77771586e-01
5.81260860e-01 -6.14514649e-01 -3.29183191e-01 9.16513085e-01
2.24357381e-01 8.57442558e-01 6.62470937e-01 -1.18013954e+00
-5.84253788e-01 5.88993207e-02 7.45346785e-01 1.15087489e-02
-4.80866879e-01 4.21727926e-01 -1.15357101e+00 -5.95901728e-01
-4.73134786e-01 -8.34661961e-01 2.39423335e-01 8.44011828e-02
-4.86354053e-01 -1.22689404e-01 9.45016086e-01 -8.75681460e-01
1.31956041e+00 -2.47731733e+00 3.73248100e-01 -5.22881389e-01
6.21797681e-01 -2.16743618e-01 -3.43718976e-01 4.73551959e-01
5.26459292e-02 2.38068849e-01 7.20937103e-02 -8.00897181e-01
-1.34860039e-01 3.36200088e-01 -4.40064967e-01 3.65355134e-01
-4.34064418e-02 1.12645876e+00 -1.16389871e+00 -9.03428674e-01
-3.35239284e-02 3.35826963e-01 -5.49968898e-01 3.64775360e-01
-4.21543002e-01 6.52829230e-01 -3.35763723e-01 8.07247996e-01
2.31980860e-01 -7.19896674e-01 2.29566962e-01 -3.72830480e-01
-2.23249570e-02 -6.09725639e-02 -9.04218316e-01 2.25411749e+00
-1.21425889e-01 1.23728168e+00 2.07824893e-02 -7.06301689e-01
-9.67348926e-03 7.74532974e-01 7.22694218e-01 -3.82544935e-01
8.44981670e-02 -3.13577533e-01 -6.87164366e-01 -1.16552579e+00
6.63047552e-01 4.07400757e-01 -3.43854815e-01 3.59420270e-01
3.31277311e-01 3.59789997e-01 4.77695554e-01 7.50531256e-01
1.36494851e+00 1.91619083e-01 8.05581436e-02 4.50568110e-01
3.41038227e-01 1.66036591e-01 3.41630697e-01 7.95776188e-01
-3.80962282e-01 7.32098997e-01 7.56202698e-01 -5.08201778e-01
-7.41391599e-01 -1.09043503e+00 3.63098919e-01 1.34226680e+00
6.00007355e-01 -1.12087929e+00 -6.18991792e-01 -6.54706180e-01
-3.60733151e-01 3.54526013e-01 -5.73350310e-01 3.00835371e-01
-5.66824615e-01 5.44503592e-02 7.23358870e-01 4.45495963e-01
5.04112422e-01 -1.03990412e+00 -6.77473664e-01 -1.26003161e-01
-1.13566887e+00 -2.05760026e+00 -9.30933297e-01 -4.43927407e-01
-2.10758597e-01 -1.39737940e+00 -7.03008115e-01 -7.20010519e-01
3.81268084e-01 5.29735386e-01 1.27550507e+00 1.60486117e-01
-1.33406624e-01 1.01160610e+00 -6.56427622e-01 2.93582417e-02
-1.30724177e-01 -1.68869436e-01 -1.48729637e-01 3.40176493e-01
8.45425278e-02 -2.02954650e-01 -5.72651505e-01 4.00542438e-01
-9.70057845e-01 4.15095329e-01 -9.80105773e-02 3.07531983e-01
5.68852484e-01 -3.71738076e-01 -6.85544387e-02 -5.02758980e-01
5.27021214e-02 -8.79930139e-01 -1.98014796e-01 2.23275498e-01
4.83406425e-01 -5.01189947e-01 2.28189692e-01 -8.65541995e-01
-7.19090283e-01 1.36614099e-01 3.24777335e-01 -1.04274595e+00
-3.93250138e-01 4.08868730e-01 -1.58982128e-01 3.06506902e-01
4.30327624e-01 -8.28667656e-02 -4.13732737e-01 -2.10642621e-01
6.06902122e-01 4.09411818e-01 9.17471528e-01 -4.72583860e-01
6.97096705e-01 8.43421519e-01 -4.60692048e-01 -8.35831583e-01
-1.11954069e+00 -7.32757568e-01 -3.22306573e-01 -7.65641987e-01
1.30227971e+00 -1.24291849e+00 -1.16424274e+00 3.30701321e-01
-1.56161714e+00 -6.02611959e-01 -1.22303776e-01 3.79659116e-01
-6.04096532e-01 5.06189167e-01 -7.94661760e-01 -5.96784770e-01
1.10372536e-01 -1.01607335e+00 1.10733974e+00 1.37201726e-01
-4.70902622e-01 -7.01238155e-01 -2.70698041e-01 8.87163699e-01
-2.30795413e-01 4.84781802e-01 1.74809489e-02 -3.96223038e-01
-7.59784222e-01 -2.71235466e-01 -2.42042765e-01 2.73382142e-02
-1.19397767e-01 6.60897717e-02 -7.76796877e-01 -3.18081416e-02
-2.26979792e-01 -7.04925656e-01 7.61827469e-01 1.11328242e-02
1.39464259e+00 -5.55292487e-01 -4.47874218e-01 5.15635252e-01
1.09536695e+00 8.51956829e-02 4.08931345e-01 1.46053523e-01
9.37513590e-01 4.00957137e-01 5.85545599e-01 4.33413357e-01
7.44512618e-01 6.84682906e-01 6.28115177e-01 2.34112158e-01
-1.45196557e-01 -3.28311473e-01 7.64997125e-01 6.11082554e-01
-3.75342011e-01 -8.38604271e-01 -1.13260400e+00 7.17634439e-01
-2.21218872e+00 -1.46390736e+00 -3.49491745e-01 1.60613239e+00
5.87302566e-01 -2.11661592e-01 1.71408147e-01 -2.91765660e-01
7.85133779e-01 5.30992925e-01 -2.43255049e-01 3.02372366e-01
-2.00234950e-01 -2.99467921e-01 2.64922082e-01 3.92671376e-01
-1.24289286e+00 8.43567610e-01 5.69513559e+00 6.07717931e-01
-9.00593817e-01 5.08127332e-01 3.81003469e-01 -8.48996043e-01
-1.35635510e-01 -5.31080663e-02 -5.75377643e-01 7.00300515e-01
9.89808559e-01 6.94766119e-02 6.46133304e-01 5.25634825e-01
3.21949780e-01 -3.23252678e-01 -1.24683690e+00 1.38023531e+00
4.13231552e-01 -1.62189066e+00 1.80615202e-01 -1.61242753e-01
7.32845008e-01 4.91172783e-02 -8.76223147e-02 1.36198595e-01
-1.80085242e-01 -9.77727592e-01 1.24966979e+00 6.66543901e-01
9.82143700e-01 -2.06568703e-01 5.03653049e-01 2.32938826e-01
-1.44367206e+00 -1.34055629e-01 2.74467260e-01 -7.55628794e-02
6.46543086e-01 -6.01612851e-02 -2.92676151e-01 3.28364134e-01
1.05185413e+00 1.00340569e+00 -5.24690688e-01 6.17102325e-01
-2.86493748e-01 6.86405241e-01 -2.10504115e-01 2.20870465e-01
2.38541529e-01 1.58443138e-01 6.25004232e-01 1.32193005e+00
2.06490904e-01 3.96595955e-01 2.69188762e-01 7.06254125e-01
-5.59945464e-01 -2.36515567e-01 -6.21616662e-01 -5.66653430e-01
4.92396384e-01 1.27248335e+00 -6.92173064e-01 -4.34516460e-01
-8.22003961e-01 1.23633587e+00 1.98811650e-01 6.16028607e-01
-1.38346016e+00 2.79455949e-02 5.78149319e-01 2.33977243e-01
2.42828503e-01 -5.60957849e-01 5.40535510e-01 -1.56382740e+00
2.79588789e-01 -8.51330459e-01 5.90392947e-01 -1.38189769e+00
-9.23257589e-01 4.40099597e-01 2.31766492e-01 -1.16293573e+00
-2.94024888e-02 -4.24617052e-01 -4.27676857e-01 6.96975878e-03
-1.01807129e+00 -1.20209110e+00 -9.62646306e-01 1.02265346e+00
7.95014918e-01 1.25425860e-01 5.08383274e-01 5.85346401e-01
-4.34336513e-01 3.10030669e-01 -4.70866323e-01 5.61508536e-01
8.99936557e-01 -8.47954750e-01 1.10057153e-01 7.51174748e-01
6.78643465e-01 3.37645352e-01 6.09403253e-01 -5.15383244e-01
-1.62706923e+00 -9.41150725e-01 7.35383987e-01 -1.05460942e+00
1.07810664e+00 -5.70101321e-01 -6.71601653e-01 1.12970352e+00
6.45344079e-01 2.55211741e-01 7.51721025e-01 -2.08215430e-01
-2.83313096e-01 3.75804424e-01 -4.63795632e-01 6.80441678e-01
1.48533368e+00 -9.22514319e-01 -4.83361065e-01 6.12014771e-01
1.04764378e+00 -7.42775559e-01 -7.30444908e-01 1.40834544e-02
5.56030631e-01 -7.04986513e-01 9.26458955e-01 -7.89480686e-01
8.99852335e-01 -3.40800703e-01 -3.07646841e-01 -5.84910691e-01
1.65765420e-01 -9.65341151e-01 -6.29265606e-01 1.37156880e+00
1.65998101e-01 1.38055906e-01 5.67384124e-01 4.97152925e-01
1.89819578e-02 -3.57003421e-01 -7.88501263e-01 -5.69357753e-01
-5.76510131e-01 -8.36336792e-01 1.90349907e-01 1.11869335e+00
2.22479016e-01 3.05110365e-01 -7.55262613e-01 3.22650552e-01
4.42980260e-01 -1.89123675e-01 9.27712977e-01 -6.23034358e-01
-2.23961577e-01 -1.96941182e-01 -3.63619089e-01 -1.17446482e+00
3.38705748e-01 -6.83728099e-01 -8.73000398e-02 -1.41034579e+00
5.89902818e-01 4.42571670e-01 -9.61535871e-02 5.83100796e-01
-1.12161241e-01 7.08211064e-01 5.19915760e-01 3.80505264e-01
-1.49453545e+00 2.02912688e-01 1.20774245e+00 -2.06768215e-01
1.66651711e-01 -6.64451241e-01 -2.88635284e-01 8.65196466e-01
6.21945739e-01 -4.84901994e-01 -3.36085647e-01 -7.54903972e-01
6.35185897e-01 4.78869021e-01 1.02208662e+00 -8.22775126e-01
4.29880947e-01 -2.84003824e-01 3.50913182e-02 -5.34041822e-01
5.82186103e-01 -6.83906376e-01 2.37112060e-01 -2.58972775e-02
-4.90171134e-01 8.68370160e-02 2.67647326e-01 8.20505142e-01
-4.11951602e-01 1.79615791e-03 2.13551208e-01 -1.64574832e-01
-1.03323281e+00 5.58500290e-01 -4.33537900e-01 4.46097612e-01
1.18505371e+00 -1.10304318e-01 -5.62792480e-01 -8.59601319e-01
-1.07480192e+00 5.36174119e-01 4.76329327e-01 6.29119813e-01
6.99730694e-01 -1.45434272e+00 -5.98117054e-01 -5.52130222e-01
4.18733209e-01 -3.36498022e-02 5.86737633e-01 9.99658525e-01
-5.86785257e-01 1.99786827e-01 6.23589344e-02 -7.03371227e-01
-1.40107203e+00 6.52754068e-01 3.25003341e-02 3.66019845e-01
-7.45934844e-01 9.33081985e-01 3.53144467e-01 3.48059446e-01
4.69744354e-01 -2.49726519e-01 -2.45674148e-01 3.62515628e-01
8.77395213e-01 1.61789298e-01 -6.85629845e-01 -7.43067622e-01
-4.63773698e-01 4.38999563e-01 3.23408246e-01 -2.85175622e-01
1.08379972e+00 -5.96282661e-01 -5.31554557e-02 7.77592003e-01
1.23121083e+00 1.35047391e-01 -1.37481904e+00 -3.21546316e-01
-3.07012927e-02 -4.79535103e-01 -3.00764263e-01 -4.66912895e-01
-9.60856795e-01 7.01448619e-01 1.00475915e-01 1.86566666e-01
9.12665546e-01 5.37895441e-01 9.17709112e-01 3.32532734e-01
2.30971724e-01 -7.86572099e-01 7.61250138e-01 4.35291111e-01
1.05359578e+00 -1.36409211e+00 -2.45254889e-01 -1.61131144e-01
-9.75755692e-01 1.03864515e+00 5.23569405e-01 1.93453297e-01
2.57845551e-01 3.64543766e-01 -5.57082370e-02 -3.17948908e-01
-1.04751086e+00 -1.15814254e-01 3.05112809e-01 2.90184975e-01
3.08687717e-01 -1.05571046e-01 9.59689245e-02 8.30451369e-01
9.83606949e-02 1.51025817e-01 7.63755739e-01 7.98644483e-01
-3.52396742e-02 -4.44832474e-01 -3.14350605e-01 -1.12284340e-01
-5.12499034e-01 7.67771015e-03 -5.06666362e-01 8.52246165e-01
2.36716211e-01 1.20585442e+00 3.20074499e-01 -3.97623509e-01
8.84378403e-02 1.15482472e-01 3.37571234e-01 -4.16971654e-01
-2.03220502e-01 -1.89085230e-01 3.49649876e-01 -1.19938540e+00
-7.82592952e-01 -7.93060839e-01 -1.28525484e+00 -3.81449044e-01
2.60110825e-01 1.35876834e-01 3.17847401e-01 8.16174328e-01
2.10079759e-01 3.77301008e-01 1.78857654e-01 -8.41661453e-01
2.73392767e-01 -8.67612481e-01 -1.09797470e-01 6.41287863e-01
5.44514537e-01 -4.78720784e-01 -4.29943353e-01 9.74911749e-01] | [10.328194618225098, 0.8022266030311584] |
c9fdb885-fafc-4561-ad23-2e8721df274b | missing-entries-matrix-approximation-and | 1302.6768 | null | http://arxiv.org/abs/1302.6768v2 | http://arxiv.org/pdf/1302.6768v2.pdf | Missing Entries Matrix Approximation and Completion | We describe several algorithms for matrix completion and matrix approximation
when only some of its entries are known. The approximation constraint can be
any whose approximated solution is known for the full matrix. For low rank
approximations, similar algorithms appears recently in the literature under
different names. In this work, we introduce new theorems for matrix
approximation and show that these algorithms can be extended to handle
different constraints such as nuclear norm, spectral norm, orthogonality
constraints and more that are different than low rank approximations. As the
algorithms can be viewed from an optimization point of view, we discuss their
convergence to global solution for the convex case. We also discuss the optimal
step size and show that it is fixed in each iteration. In addition, the derived
matrix completion flow is robust and does not require any parameters. This
matrix completion flow is applicable to different spectral minimizations and
can be applied to physics, mathematics and electrical engineering problems such
as data reconstruction of images and data coming from PDEs such as Helmholtz
equation used for electromagnetic waves. | ['Gil Shabat', 'Yaniv Shmueli', 'Amir Averbuch'] | 2013-02-27 | null | null | null | null | ['electrical-engineering'] | ['miscellaneous'] | [ 3.74663621e-01 1.45189971e-01 9.65396464e-02 3.69024426e-02
-5.11677742e-01 -5.46422303e-01 9.58991274e-02 1.53645441e-01
-4.56734717e-01 7.12454319e-01 2.28444040e-01 -7.80730769e-02
-5.31044364e-01 -3.88331622e-01 -7.36389995e-01 -9.17065084e-01
-1.61816344e-01 4.79640156e-01 -1.58553228e-01 -3.65856409e-01
1.67752132e-01 5.83140433e-01 -1.18472075e+00 -5.42322844e-02
8.26225519e-01 6.74765229e-01 1.62901044e-01 7.80120373e-01
1.14513434e-01 4.74352777e-01 -1.67333484e-01 -7.31119514e-02
6.79507196e-01 -2.10319251e-01 -7.88108885e-01 4.54832226e-01
4.68960434e-01 -4.30033207e-02 -4.14174616e-01 1.27549231e+00
4.96550083e-01 3.89637053e-01 6.39964640e-01 -1.11304808e+00
-3.71709347e-01 3.56144071e-01 -7.56295800e-01 4.68244255e-02
4.87253457e-01 -5.69684863e-01 7.28090048e-01 -1.28467703e+00
6.90602481e-01 1.23612273e+00 9.07347977e-01 2.81026185e-01
-1.57033014e+00 -1.44084796e-01 -1.27305701e-01 2.13133246e-01
-1.44974208e+00 -3.72970313e-01 6.15534842e-01 -7.37539172e-01
3.01783979e-01 5.34147680e-01 4.25223708e-01 6.64041638e-01
-1.39793232e-01 5.96814573e-01 1.16572618e+00 -5.62226772e-01
1.65162355e-01 1.33900449e-01 4.68029052e-01 6.72045529e-01
5.66730142e-01 -2.55985707e-01 -1.85720325e-01 -4.49279636e-01
6.96950436e-01 -5.90364933e-02 -8.35809946e-01 -5.99926412e-01
-1.40502274e+00 1.06490088e+00 5.72189689e-02 4.19593215e-01
-4.45018262e-01 -1.42895048e-02 2.45404527e-01 3.94360036e-01
2.86357433e-01 3.30257922e-01 -1.75647922e-02 2.30802238e-01
-7.95862436e-01 3.57743800e-01 1.21331882e+00 1.12873721e+00
9.34209704e-01 2.51486808e-01 -6.44810721e-02 9.08644497e-01
1.06870547e-01 7.85658419e-01 5.18333390e-02 -1.15654171e+00
3.45097721e-01 2.39993278e-02 2.22143114e-01 -1.23500097e+00
-7.12340295e-01 -4.08886909e-01 -1.22251403e+00 2.39705164e-02
5.06794631e-01 -1.65447637e-01 -4.33291346e-01 1.61805105e+00
4.57170069e-01 4.51866657e-01 6.59338990e-03 1.16107833e+00
6.06858492e-01 7.90833294e-01 -6.86346233e-01 -7.89292812e-01
1.16904008e+00 -5.72545528e-01 -9.81817603e-01 1.14079565e-01
4.84703153e-01 -1.08692610e+00 6.43788934e-01 6.71876073e-01
-1.12399387e+00 -2.23418429e-01 -9.45709765e-01 -5.25574237e-02
6.53126910e-02 2.44943112e-01 4.61178601e-01 6.54841661e-01
-1.03437769e+00 7.36848831e-01 -6.83519065e-01 -3.96669149e-01
-2.53578335e-01 4.90128785e-01 -4.23795551e-01 -1.40455723e-01
-8.66140425e-01 7.05056965e-01 1.88980043e-01 4.29096580e-01
-5.14670014e-01 -7.52040744e-01 -6.82165325e-01 -1.56471878e-01
2.81566083e-01 -8.45473111e-01 8.02320063e-01 -8.21902275e-01
-1.25705564e+00 6.59763396e-01 -2.55290866e-01 -3.68137449e-01
5.78659654e-01 -8.62496048e-02 -6.11685663e-02 1.65698737e-01
6.25433475e-02 -2.67008752e-01 1.09423018e+00 -9.90536392e-01
-1.44505098e-01 -2.44536340e-01 1.48348048e-01 2.09070385e-01
-2.73141742e-01 -1.51980668e-01 -2.90234119e-01 -7.48387218e-01
4.59213257e-01 -1.19305527e+00 -7.50658333e-01 -1.36661837e-02
-4.19561595e-01 3.12036812e-01 5.97831786e-01 -6.62547648e-01
1.16552556e+00 -2.08992791e+00 9.27042365e-01 5.34207642e-01
1.81599855e-01 -6.34700758e-03 -2.07333807e-02 8.94269645e-01
-3.02985221e-01 -1.01674549e-01 -6.56277955e-01 -2.78630912e-01
-1.24375716e-01 2.04817683e-01 -2.22224519e-01 1.09629261e+00
-2.51700908e-01 1.55299976e-01 -7.29144573e-01 -2.52742320e-01
1.99452221e-01 5.23032486e-01 -6.74079120e-01 -5.93288392e-02
2.75125712e-01 5.53155661e-01 -4.18921351e-01 1.02207363e-01
1.09034514e+00 -8.27524960e-02 1.01482674e-01 -6.42417192e-01
-2.56995767e-01 -4.66929317e-01 -2.03334594e+00 1.65542495e+00
-6.02612913e-01 6.55939877e-01 1.09078085e+00 -1.49203658e+00
5.96247435e-01 4.31420326e-01 7.96174943e-01 2.77088466e-03
-1.05110994e-02 3.70281190e-01 -2.84095667e-02 -5.12506962e-01
5.44721842e-01 -3.64124089e-01 4.01490539e-01 2.90918142e-01
-2.93813974e-01 6.19502850e-02 6.89420402e-01 4.50354964e-01
1.10197902e+00 -3.57394248e-01 4.04839128e-01 -7.96552956e-01
1.08736503e+00 -1.20939471e-01 5.62182844e-01 6.91758394e-01
4.33298379e-01 7.02547610e-01 3.31570387e-01 -2.53219157e-01
-1.15536845e+00 -7.22853601e-01 -5.73799074e-01 6.70206726e-01
1.31247595e-01 -5.21047771e-01 -5.52764595e-01 6.46446571e-02
-8.69247839e-02 1.79137979e-02 -4.72758740e-01 2.15297863e-01
-7.21917152e-01 -8.85109007e-01 1.17430024e-01 8.75051394e-02
1.61198616e-01 -3.29948455e-01 -9.84795988e-02 2.64104307e-01
-2.27717176e-01 -1.33308923e+00 -5.31211376e-01 -9.38528329e-02
-1.08562624e+00 -1.23998070e+00 -1.02210021e+00 -6.86771512e-01
9.66823280e-01 3.08656007e-01 7.69457221e-01 5.87279396e-03
-2.94748873e-01 8.93337429e-01 -2.47917101e-01 -8.76175910e-02
-3.95959735e-01 -2.23505616e-01 4.53462094e-01 5.53654909e-01
-3.58634293e-01 -7.52860010e-01 -4.32787359e-01 4.24037725e-01
-1.16641939e+00 -8.62150788e-02 4.07870173e-01 9.54827785e-01
7.57247150e-01 -5.74350320e-02 2.97572017e-01 -1.23756993e+00
7.77393878e-01 -4.11229074e-01 -7.20885336e-01 2.56895758e-02
-3.83510619e-01 3.46860439e-01 8.68803263e-01 -5.12077212e-01
-7.45530546e-01 4.67883706e-01 2.14557927e-02 -6.52853131e-01
3.06139141e-01 6.67507052e-01 -7.49965664e-03 -4.31935698e-01
7.56178617e-01 2.28996769e-01 1.30518138e-01 -7.35067785e-01
4.55951214e-01 2.97854871e-01 4.45096374e-01 -7.30363250e-01
1.15677917e+00 8.95035028e-01 6.92116678e-01 -1.52998710e+00
-6.81467235e-01 -9.13599730e-01 -5.46974838e-01 -8.33608210e-02
5.83879590e-01 -7.24868238e-01 -7.75061488e-01 6.05819002e-02
-1.24075282e+00 9.98392403e-02 -2.77971596e-01 7.99859285e-01
-7.85906494e-01 8.62580717e-01 -6.94553137e-01 -9.03870225e-01
-6.88798949e-02 -1.13778281e+00 6.25853658e-01 -1.70078620e-01
1.48128256e-01 -1.36954272e+00 3.08618814e-01 -4.87879291e-02
2.54821330e-01 3.43500316e-01 5.82558572e-01 -2.60678291e-01
-4.64550376e-01 -1.87767409e-02 -8.84156616e-04 4.26485360e-01
1.63799301e-02 -2.99202621e-01 -5.73287666e-01 -6.80179417e-01
3.87504339e-01 9.04563442e-02 7.69144535e-01 5.96885204e-01
8.07203770e-01 -7.25812137e-01 -2.93226659e-01 8.47396314e-01
1.66715384e+00 -4.12351251e-01 4.65179741e-01 -4.44675051e-02
9.01413202e-01 7.59153306e-01 3.80814552e-01 6.15595520e-01
-2.72797972e-01 8.27759981e-01 2.66540080e-01 -1.17456391e-01
2.15949222e-01 3.47474664e-01 2.62075156e-01 1.12082326e+00
-2.00666070e-01 2.04624876e-01 -6.10485494e-01 4.84005094e-01
-2.17897248e+00 -1.05560136e+00 -8.27060223e-01 2.54529715e+00
4.10059810e-01 -5.02543569e-01 1.80786625e-01 3.50416750e-01
9.17498171e-01 -1.98106393e-02 -3.09796691e-01 -3.80111009e-01
-2.93668389e-01 1.40924826e-01 8.60550821e-01 1.07940078e+00
-9.02863145e-01 3.93198490e-01 7.21981287e+00 7.32257009e-01
-7.34895766e-01 1.72755271e-01 6.80499431e-03 -2.73859035e-03
-4.91692871e-01 1.87048957e-01 -3.67187947e-01 7.33802170e-02
4.94847268e-01 -2.77823687e-01 6.49679184e-01 6.46051764e-01
3.80517274e-01 3.89827080e-02 -1.14932346e+00 1.42194438e+00
5.84393367e-02 -1.21508718e+00 -9.64498371e-02 3.23330313e-01
8.86256516e-01 -2.34741062e-01 -4.95066568e-02 -2.55093247e-01
-1.72848076e-01 -9.61717904e-01 2.51123279e-01 6.12176895e-01
4.56349790e-01 -7.32320011e-01 4.71226096e-01 4.27357227e-01
-1.25698996e+00 1.49022669e-01 -6.21950984e-01 -1.97930843e-01
4.12566334e-01 9.74453747e-01 -4.44384813e-01 8.47896755e-01
1.87717855e-01 8.79754841e-01 -3.24852988e-02 1.20597458e+00
3.57249260e-01 4.66644138e-01 -7.02953815e-01 1.27429485e-01
1.29967421e-01 -1.04599535e+00 1.14352632e+00 1.23905706e+00
5.17760754e-01 4.05324340e-01 4.46454555e-01 5.66219449e-01
3.23830456e-01 5.78986526e-01 -6.67320013e-01 2.45719239e-01
2.09163688e-02 1.41881275e+00 -6.57554209e-01 -1.39158204e-01
-6.41437829e-01 9.90412176e-01 6.31817952e-02 7.47761846e-01
-4.46719527e-01 -3.40956479e-01 6.20711088e-01 3.54238093e-01
1.39605641e-01 -3.70209575e-01 1.97666548e-02 -1.41146564e+00
7.84247369e-02 -6.60502493e-01 4.31121379e-01 -3.13426286e-01
-1.19011104e+00 3.94882739e-01 2.07890749e-01 -1.22895455e+00
-3.20293531e-02 -9.73563910e-01 -3.50056708e-01 7.50999510e-01
-1.06553066e+00 -5.78165352e-01 -1.93079472e-01 8.02816689e-01
1.27403766e-01 -4.85378839e-02 4.55215842e-01 6.19671524e-01
-5.16310751e-01 8.67290944e-02 6.22466922e-01 -1.02476098e-01
6.05793774e-01 -1.30885124e+00 -4.04312044e-01 1.02854717e+00
1.68252990e-01 7.13231921e-01 1.29878914e+00 -3.02837402e-01
-1.89775789e+00 -7.02430367e-01 4.01882887e-01 -6.77281767e-02
8.08971345e-01 -2.05412805e-01 -8.82825851e-01 6.39702320e-01
1.87515676e-01 8.27457160e-02 5.84472775e-01 1.48934321e-02
-2.47890279e-02 -9.44248140e-02 -9.63147044e-01 3.44141573e-01
8.79811585e-01 -2.24135563e-01 -1.37864396e-01 8.06702793e-01
2.18590572e-01 -6.59529984e-01 -9.74223137e-01 2.55824596e-01
3.03707987e-01 -7.96376228e-01 1.10043716e+00 -5.77112913e-01
-1.15382589e-01 -6.48272097e-01 -3.93204570e-01 -1.21849775e+00
-4.83618528e-01 -1.14262617e+00 -7.17355162e-02 8.76647294e-01
2.20953494e-01 -5.63310921e-01 5.69528818e-01 3.61758590e-01
-2.34266385e-01 -6.87817752e-01 -1.08370256e+00 -9.28692162e-01
-1.50208876e-01 -4.22235399e-01 -3.05826887e-02 1.02394009e+00
1.02095552e-01 4.59971100e-01 -9.36535716e-01 3.97637993e-01
1.05704999e+00 7.61602670e-02 7.30020463e-01 -1.31359923e+00
-6.10373795e-01 -1.99186757e-01 -4.39661443e-01 -1.19173491e+00
2.12169439e-01 -1.19118726e+00 -1.96673721e-01 -1.54725945e+00
1.23501852e-01 -5.19310892e-01 1.81752846e-01 -9.53344107e-02
1.08664870e-01 1.39036074e-01 2.47468829e-01 1.24546371e-01
-1.77763656e-01 4.03042167e-01 1.16875803e+00 -1.09084286e-01
-2.19918966e-01 1.29613176e-01 -3.88450950e-01 7.19797015e-01
3.93153340e-01 -3.85681182e-01 -4.66969281e-01 -2.39905521e-01
6.05401576e-01 2.56611466e-01 6.80756271e-02 -9.79411721e-01
3.25214297e-01 -8.34332332e-02 -1.73722073e-01 -2.55878180e-01
4.63693261e-01 -1.01255763e+00 5.11831105e-01 4.60031092e-01
-1.66076645e-01 1.05300069e-01 -6.29337737e-03 8.25206101e-01
-1.66603774e-01 -8.13496172e-01 9.40286160e-01 -1.20361768e-01
-3.35497379e-01 5.14664412e-01 -5.29121161e-01 2.44949326e-01
8.35655570e-01 -2.57488698e-01 1.92060441e-01 -6.55048609e-01
-1.12264228e+00 1.83832005e-01 3.22574973e-01 -1.38655603e-01
5.88891566e-01 -1.53677237e+00 -1.01372671e+00 -6.53391480e-02
-7.78969452e-02 -2.04697192e-01 3.87569815e-01 1.40827644e+00
-7.75325418e-01 2.72370547e-01 1.04042023e-01 -7.56317139e-01
-1.22529054e+00 7.80413210e-01 2.89602309e-01 -1.25534952e-01
-6.09318972e-01 3.24678957e-01 3.89089376e-01 -2.74887353e-01
1.26066580e-01 -1.66473433e-01 -6.06899448e-02 5.41853793e-02
7.03185558e-01 8.11729014e-01 -3.66711915e-02 -9.67994034e-01
-2.83637553e-01 1.06328630e+00 2.97027409e-01 -2.55178839e-01
1.20986545e+00 -3.50806952e-01 -4.97043431e-01 4.89228427e-01
1.39748919e+00 4.58947361e-01 -6.97973609e-01 -3.75810623e-01
-1.71838537e-01 -3.49239320e-01 -4.28796560e-02 1.72672257e-01
-1.13573790e+00 6.83202386e-01 3.61667842e-01 2.34712124e-01
1.14963484e+00 -3.18863779e-01 3.99164468e-01 6.92442834e-01
2.93254316e-01 -9.71430659e-01 -2.39464834e-01 5.24298251e-01
1.14793241e+00 -9.40320075e-01 3.05902064e-01 -7.99721360e-01
-1.56257913e-01 1.36936295e+00 1.27702355e-01 -5.54134905e-01
8.46518815e-01 2.42874682e-01 -2.88140982e-01 1.16652310e-01
-3.18852395e-01 -3.41960460e-01 4.93777603e-01 6.21861100e-01
5.22144616e-01 -1.20844521e-01 -1.04729891e+00 -2.37056483e-02
3.49791050e-02 -4.42999989e-01 1.03786159e+00 5.00710905e-01
-3.66166919e-01 -1.31115544e+00 -8.15518677e-01 3.83646458e-01
-4.51488376e-01 -1.19158007e-01 -7.01845735e-02 6.14441812e-01
-2.81740278e-01 7.60278881e-01 -3.76288205e-01 4.99581806e-02
3.47163558e-01 -1.66723892e-01 7.29253769e-01 -5.22444606e-01
-2.30521441e-01 3.59891444e-01 1.02504008e-01 -5.96765041e-01
-7.24142671e-01 -7.47695863e-01 -1.17622077e+00 -3.99715304e-01
-3.15048814e-01 6.49104714e-01 5.28435111e-01 6.89154088e-01
8.14790130e-02 1.56963468e-01 5.70905983e-01 -8.99810433e-01
-4.98130530e-01 -8.42100680e-01 -9.85862792e-01 4.38582271e-01
5.93136072e-01 -4.79212672e-01 -4.49114203e-01 4.46613096e-02] | [7.083764553070068, 4.443610668182373] |
9bc92634-ed36-4da7-ba4f-4ad264d3a67e | decipherment-with-a-million-random-restarts | null | null | https://aclanthology.org/D13-1087 | https://aclanthology.org/D13-1087.pdf | Decipherment with a Million Random Restarts | null | ['Taylor Berg-Kirkpatrick', 'Dan Klein'] | 2013-10-01 | null | null | null | emnlp-2013-10 | ['decipherment'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.320093154907227, 3.69858980178833] |
bffae6d8-0846-41da-8034-4e4f24147a84 | global-sensitivity-analysis-of-a-symmetric | 2107.04647 | null | https://arxiv.org/abs/2107.04647v2 | https://arxiv.org/pdf/2107.04647v2.pdf | Global sensitivity analysis of asymmetric energy harvesters | Parametric variability is inevitable in actual energy harvesters. It can significantly affect crucial aspects of the system performance, especially in harvesting systems that present geometric parameters, material properties, or excitation conditions that are susceptible to small perturbations. This work aims to develop an investigation to identify the most critical parameters in the dynamic behavior of asymmetric bistable energy harvesters with nonlinear piezoelectric coupling, considering the variability of their physical and excitation properties. For this purpose, a global sensitivity analysis based on orthogonal variance decomposition, employing Sobol indices, is performed to quantify the effect of the harvester parameters on the variance of the recovered power. This technique quantifies the variance concerning each parameter individually and collectively regarding the total variation of the model. The results indicate that the frequency and amplitude of excitation, asymmetric terms and electrical proprieties of the piezoelectric coupling are the most critical parameters that affect the mean power harvested. It is also shown that the order of importance of the parameters can change according to the stability of the harvester's dynamic response. In this way, a better understanding of the system under analysis is obtained since the study allows the identification of vital parameters that rule the change of dynamic behavior and therefore constitutes a powerful tool in the robust design, optimization, and response prediction of nonlinear harvesters. | ['Paulo Sérgio Varoto', 'Samuel da Silva', 'Americo Cunha Jr', 'João Pedro Norenberg'] | 2021-07-09 | null | null | null | null | ['robust-design'] | ['miscellaneous'] | [ 1.88119218e-01 -2.38799676e-01 9.13771987e-02 4.30825651e-01
2.06936244e-02 -8.95381510e-01 3.69376987e-01 1.86428994e-01
-2.49926329e-01 7.41978645e-01 3.52925472e-02 2.23561347e-01
-7.04586327e-01 -7.02419579e-01 -4.36636031e-01 -1.54348230e+00
-3.79286528e-01 5.61359003e-02 1.01583824e-01 -4.89372253e-01
2.16303587e-01 8.35352957e-01 -1.70488799e+00 -7.18022525e-01
7.86411941e-01 1.05039382e+00 2.43649274e-01 4.07569557e-01
5.13326526e-01 -6.28010556e-02 -6.95948422e-01 2.52128065e-01
-5.32325916e-02 -4.29393381e-01 -1.13681868e-01 -3.45957547e-01
-4.46243614e-01 -7.48589858e-02 3.54889393e-01 7.12831259e-01
6.34824991e-01 1.22968920e-01 1.34775364e+00 -8.76734197e-01
-2.12981880e-01 7.20362484e-01 -9.06294435e-02 2.34169383e-02
5.36949933e-02 2.01234609e-01 6.17192924e-01 -3.86221170e-01
3.58307540e-01 6.36181414e-01 2.58818239e-01 2.58930624e-01
-1.45752966e+00 -5.53498149e-01 -5.96536756e-01 2.35682443e-01
-1.17754650e+00 -4.11644191e-01 1.11430764e+00 -5.96501768e-01
3.41679752e-01 1.61122128e-01 7.99346268e-01 5.63522398e-01
7.80033886e-01 -2.36887023e-01 1.06300735e+00 -7.19201028e-01
3.75023067e-01 -4.14627828e-02 2.66679004e-02 1.92980587e-01
6.76653028e-01 5.63364744e-01 -2.91996062e-01 -4.70896289e-02
4.16973203e-01 -7.97407210e-01 -4.58644718e-01 -5.19284308e-01
-6.55011952e-01 4.06483054e-01 2.27928042e-01 9.89088356e-01
-5.01864493e-01 1.90394849e-01 4.36693393e-02 3.84618118e-02
-2.70411205e-02 5.79795003e-01 -2.95069683e-02 -1.26463354e-01
-6.42988324e-01 -5.48168607e-02 8.95550907e-01 1.26180351e-01
4.08975840e-01 2.20513299e-01 -1.45262018e-01 3.02090675e-01
-2.70927157e-02 1.23160350e+00 2.47498289e-01 -6.56841695e-01
-1.04919314e-01 4.18371618e-01 2.35182300e-01 -1.22419488e+00
-7.68314898e-01 -5.44264913e-01 -7.83101797e-01 1.18427776e-01
4.78489757e-01 -5.15023172e-01 -4.47941154e-01 1.70856726e+00
2.67901510e-01 -7.01392472e-01 1.26107678e-01 7.86066115e-01
5.96097410e-01 5.73977649e-01 1.57189563e-01 -5.69063723e-01
1.27720237e+00 3.12982231e-01 -8.53759646e-01 1.35006294e-01
1.76103368e-01 -5.92959821e-01 5.42993069e-01 -9.56024081e-02
-9.95136559e-01 -3.48763913e-02 -1.03104806e+00 5.93503535e-01
-4.71761644e-01 4.83256608e-01 -9.32084024e-02 5.28594077e-01
-6.86115265e-01 5.92320979e-01 -6.78221583e-01 -3.71864706e-01
-4.35789287e-01 4.52183843e-01 -2.32798071e-03 6.04320049e-01
-1.06718421e+00 1.26430607e+00 4.76663351e-01 7.33461857e-01
-3.96332830e-01 -7.04156518e-01 -2.15748131e-01 5.46024978e-01
6.48578107e-02 -4.54709858e-01 5.63407421e-01 -3.57304633e-01
-1.84754348e+00 2.99888462e-01 1.79075524e-01 -8.52329433e-02
4.46697474e-01 2.91748226e-01 -1.18962362e-01 2.66427189e-01
-1.09643415e-01 1.45643633e-02 7.50333428e-01 -1.29074001e+00
1.18022136e-01 -4.77197886e-01 -5.43739796e-01 -5.89987673e-02
-4.35810804e-01 -2.58180469e-01 5.01917839e-01 -2.56364852e-01
4.88165542e-02 -9.70157564e-01 1.17909476e-01 -5.47205508e-01
-3.49731773e-01 -7.97390267e-02 6.09367132e-01 -3.41558337e-01
8.31257761e-01 -2.18945599e+00 7.74935544e-01 6.12242103e-01
-2.65452415e-01 -1.32764667e-01 1.80334896e-01 1.08795476e+00
3.88692468e-01 -1.01079322e-01 -3.40974897e-01 6.66658401e-01
-7.71336183e-02 -2.57501721e-01 -8.75348672e-02 5.26436567e-01
7.63388127e-02 5.82430959e-01 -5.72202504e-01 -2.64027536e-01
1.58876002e-01 6.77854240e-01 -1.29650563e-01 2.45523408e-01
4.73016389e-02 4.96098280e-01 -8.12862992e-01 4.09812361e-01
3.46261829e-01 4.50465798e-01 2.36476943e-01 -7.32361436e-01
-7.35663116e-01 -5.24207771e-01 -8.40977013e-01 4.61258531e-01
-3.70463014e-01 4.69052821e-01 4.82479900e-01 -1.07134986e+00
1.24824572e+00 2.10925594e-01 8.97286057e-01 -7.61686087e-01
5.49280107e-01 4.74122524e-01 1.15969233e-01 -6.07376873e-01
4.97262001e-01 -4.13460284e-01 -1.85969487e-01 2.48930112e-01
2.74218316e-03 -3.86810035e-01 3.11875850e-01 -4.21495199e-01
6.12132788e-01 -2.77938962e-01 1.98440671e-01 -8.51616323e-01
6.03459299e-01 -1.94265947e-01 2.45146081e-01 2.59397417e-01
-1.17674153e-02 -2.25782767e-01 8.07094634e-01 2.39258289e-01
-6.68869972e-01 -7.77914822e-01 -4.35983241e-01 6.26785040e-01
6.22070551e-01 4.18400764e-01 -5.18500626e-01 5.31781316e-01
2.13217124e-01 6.73362076e-01 -5.40336788e-01 -6.19544804e-01
-5.78423560e-01 -1.08094776e+00 2.85199225e-01 5.78700118e-02
3.26296002e-01 -1.02631354e+00 -1.24072480e+00 2.80431002e-01
-8.08882415e-02 -9.36931551e-01 9.27323624e-02 5.19309044e-01
-1.04168487e+00 -9.47218478e-01 -5.87264836e-01 -3.36276740e-01
5.50265491e-01 -2.97312140e-01 7.14701176e-01 -1.82772279e-01
-2.82039225e-01 5.58864832e-01 -3.31938624e-01 -2.99872190e-01
-5.41830778e-01 3.32762152e-01 3.46148238e-02 -2.09776908e-01
-3.80415171e-01 -8.51413846e-01 -5.82571745e-01 6.59409523e-01
-9.58407462e-01 -6.38739526e-01 5.97612917e-01 4.31108922e-01
4.16628748e-01 4.69972938e-01 6.91294909e-01 -3.45274284e-02
1.02370417e+00 -3.49186689e-01 -8.75441611e-01 3.44930559e-01
-5.91741264e-01 4.78389144e-01 7.88811445e-01 -4.27877516e-01
-8.56772661e-01 -5.10583669e-02 5.08416831e-01 2.92715609e-01
-2.19070408e-02 5.77507675e-01 -4.12249535e-01 -2.92046070e-01
4.05133486e-01 3.93605828e-01 8.02487284e-02 -2.95275867e-01
1.01763383e-01 1.47491544e-01 3.82221639e-01 -6.04289711e-01
1.00416887e+00 1.58723667e-01 8.98140609e-01 -1.33309031e+00
3.17032754e-01 8.79568234e-02 -2.36659199e-01 -6.77288175e-01
6.32251501e-01 -9.68427882e-02 -1.11384988e+00 7.84747601e-01
-6.56409085e-01 -4.34854925e-01 -1.21428408e-01 3.78131986e-01
-4.26546305e-01 -4.75169159e-04 9.21377167e-03 -1.14888096e+00
-6.45240605e-01 -1.05741024e+00 6.77651525e-01 5.56587696e-01
6.75642118e-02 -9.29108441e-01 1.69723824e-01 -1.23227619e-01
9.17005301e-01 7.68601954e-01 1.22933805e+00 1.91621080e-01
-3.96562040e-01 -2.03188986e-01 3.76132756e-01 -1.03473380e-01
2.20803380e-01 4.04335022e-01 -3.84740114e-01 -3.28399956e-01
-4.11503315e-02 2.14157492e-01 8.00626397e-01 9.37761664e-01
7.69797936e-02 -1.29143700e-01 -4.16615218e-01 3.03400695e-01
1.86638641e+00 2.85710245e-01 3.24017406e-01 1.58695281e-01
6.19045980e-02 8.15343082e-01 3.08003247e-01 4.91602033e-01
-1.28980801e-01 8.46879542e-01 7.03917980e-01 3.70334908e-02
2.98508406e-01 1.77482158e-01 6.54177427e-01 3.79665405e-01
-4.57663506e-01 -3.02290469e-01 -5.11330009e-01 2.64014602e-01
-1.12337077e+00 -9.38651502e-01 -1.85282361e-02 2.41672993e+00
3.17489624e-01 -3.87179047e-01 3.28411430e-01 2.56767869e-01
7.29714572e-01 5.13170734e-02 -3.50614071e-01 -4.78042841e-01
-4.79893297e-01 1.03605101e-02 8.81681561e-01 5.19139946e-01
-4.59266543e-01 2.72915095e-01 6.30110931e+00 2.98439413e-01
-1.43515611e+00 -5.14722884e-01 -9.49166790e-02 -9.29029137e-02
-5.52875757e-01 -1.24518953e-01 -4.88263994e-01 6.63458705e-01
7.33013630e-01 -5.57331860e-01 5.11002421e-01 8.47202539e-02
5.87558270e-01 -7.26144314e-01 -4.89152580e-01 1.81026489e-01
-2.64474303e-01 -6.27780676e-01 -2.90707350e-01 3.79497558e-01
3.34146112e-01 -3.87639374e-01 1.43576220e-01 -3.16818118e-01
-4.76120770e-01 -4.54667211e-01 5.89032769e-01 1.00129199e+00
4.05009657e-01 -6.75042868e-01 8.41404974e-01 3.26257408e-01
-1.16543972e+00 -3.07476878e-01 -7.66812786e-02 1.73635304e-01
4.09552217e-01 9.21391308e-01 -5.81814229e-01 3.72127414e-01
4.22505707e-01 -1.07258640e-01 -2.70436347e-01 7.16950595e-01
-9.61873457e-02 6.81488276e-01 -7.81147957e-01 -5.83825469e-01
-4.24381047e-01 -5.69918871e-01 9.45577979e-01 5.08373380e-01
6.81279719e-01 2.53742397e-01 -6.69063151e-01 1.18616605e+00
3.29077721e-01 1.67875752e-01 -5.19460082e-01 -3.68760765e-01
6.79286540e-01 1.07365739e+00 -9.55673039e-01 5.37254214e-01
4.74433988e-01 9.80797112e-02 -4.14134920e-01 4.55418766e-01
-6.57956541e-01 -3.36873561e-01 3.41739148e-01 2.53575206e-01
5.97848654e-01 -4.28735256e-01 -1.62413847e-02 -6.53836608e-01
2.07689211e-01 -2.76579168e-02 -8.27864651e-03 -3.06699306e-01
-7.25206316e-01 1.71886280e-01 4.54812020e-01 -1.01524711e+00
-3.35199744e-01 -3.29617172e-01 -8.26569200e-01 5.55861890e-01
-1.26865482e+00 -9.74470317e-01 -1.53639644e-01 2.09135935e-01
-5.13935447e-01 1.67116120e-01 6.10123813e-01 -1.38311625e-01
-7.67080307e-01 2.78695256e-01 7.74215639e-01 -4.33235198e-01
2.88455278e-01 -6.87704742e-01 -7.78384745e-01 7.11442530e-01
-7.52532125e-01 4.27045435e-01 1.37148476e+00 -4.89933252e-01
-1.65310383e+00 -9.82602462e-02 5.28366268e-01 3.10585082e-01
5.31575143e-01 -2.96445061e-02 -4.20952827e-01 -3.20372015e-01
6.10321201e-02 -5.49544036e-01 6.08215392e-01 -2.86390841e-01
4.93237436e-01 -6.14070415e-01 -1.09533238e+00 3.63535941e-01
3.84360880e-01 -1.52781323e-01 -1.11005217e-01 -1.77681357e-01
-3.86263281e-02 2.11854383e-01 -1.24658513e+00 7.10526228e-01
1.17484796e+00 -8.37067902e-01 6.97473466e-01 7.46925082e-03
1.89582884e-01 -1.38078928e-01 9.88169461e-02 -1.31313074e+00
-2.56040365e-01 -4.21494365e-01 -2.51792162e-03 1.50564146e+00
3.32959861e-01 -8.85380924e-01 3.28052938e-01 6.42468393e-01
4.18164700e-01 -5.37677050e-01 -1.11252975e+00 -9.12403643e-01
3.08026820e-01 5.23551702e-01 2.84506768e-01 8.55927393e-02
1.83846787e-01 2.28897780e-02 3.50871496e-02 3.01040858e-01
7.17367411e-01 3.19318771e-01 2.41667122e-01 -1.36124897e+00
1.30417630e-01 -4.76299584e-01 -3.48195374e-01 -3.69660258e-02
-1.05962701e-01 -3.28980327e-01 2.30331451e-01 -1.11293781e+00
-2.02862218e-01 -2.76966274e-01 -3.45767021e-01 -3.69290076e-02
-5.34058958e-02 -1.22582197e-01 1.94338650e-01 3.41615587e-01
4.32243973e-01 9.72652376e-01 1.10594058e+00 2.03701742e-02
-7.69513667e-01 1.75383314e-01 -2.92352408e-01 7.19503015e-02
8.03094387e-01 -4.92394984e-01 -1.30397633e-01 -5.87590002e-02
2.46179968e-01 2.36509278e-01 3.79902422e-01 -1.34760582e+00
3.12325746e-01 -2.32451245e-01 -7.47495592e-02 -3.44502032e-01
1.20994903e-01 -1.09032476e+00 6.29523218e-01 6.95681572e-01
-1.34380504e-01 -3.65692616e-01 2.15340391e-01 3.97668064e-01
3.67658176e-02 -3.47546458e-01 5.37492692e-01 4.48169887e-01
-1.28575295e-01 -3.97798955e-01 -7.32716799e-01 -2.39908457e-01
9.60503519e-01 -3.13255996e-01 -2.76301116e-01 -4.03454542e-01
-5.50137579e-01 1.11189678e-01 3.92783970e-01 -1.13834046e-01
-3.23014781e-02 -8.19393933e-01 -5.28297365e-01 1.59698769e-01
-2.95566797e-01 -5.74893177e-01 4.08259958e-01 1.15874887e+00
-3.97092998e-01 3.09910178e-01 -6.56018376e-01 -4.72499460e-01
-1.20741475e+00 1.38085246e-01 5.17437935e-01 -2.13464409e-01
1.87980831e-01 1.82379574e-01 -2.02796727e-01 4.14108008e-01
-2.26911187e-01 -3.41073811e-01 -6.41817510e-01 6.65330648e-01
-2.89440542e-01 6.98912024e-01 -1.32352943e-02 -7.54457533e-01
-4.71426308e-01 1.35095584e+00 8.99262011e-01 -6.69252947e-02
1.34922051e+00 -1.57707602e-01 -3.66533756e-01 3.45673025e-01
8.18353355e-01 1.60331413e-01 -9.81962681e-01 4.11981672e-01
-3.43432486e-01 9.51774493e-02 3.12305719e-01 -7.23261416e-01
-1.09767354e+00 4.24323320e-01 7.08702385e-01 7.50551403e-01
1.51366079e+00 -3.92422266e-02 3.29731971e-01 3.64172786e-01
3.54527920e-01 -1.22144306e+00 -3.16963553e-01 1.15092993e-01
1.03012121e+00 -5.80264628e-01 1.74708545e-01 -4.64111030e-01
-6.98985755e-02 1.42800415e+00 2.58722398e-02 2.74925008e-02
8.02797854e-01 4.58141595e-01 -1.38646513e-01 5.77506758e-02
-2.28602901e-01 -1.72693506e-01 1.44889921e-01 4.46604133e-01
3.55582267e-01 4.68282342e-01 -1.12483788e+00 5.98952711e-01
-2.19359830e-01 -1.24285789e-02 3.46644759e-01 6.95795774e-01
-6.42038345e-01 -9.22677040e-01 -5.78644216e-01 1.84970442e-02
-9.80925262e-02 4.91417915e-01 -7.88374066e-01 7.90742278e-01
1.26230896e-01 8.99978697e-01 -3.02770436e-01 -2.26728991e-01
6.98673785e-01 5.46140745e-02 4.52468902e-01 2.29527354e-01
-4.70547974e-01 -7.07954019e-02 -1.18794732e-01 -8.81023854e-02
-5.29126823e-01 -5.90555072e-01 -1.11441207e+00 -2.95491934e-01
-8.77860188e-01 5.24970651e-01 9.26161170e-01 9.16480064e-01
2.98182756e-01 2.87463278e-01 7.70116150e-01 -9.95838165e-01
-5.81959128e-01 -8.52572978e-01 -8.61083627e-01 -4.57903184e-02
1.91035852e-01 -9.45839047e-01 -8.57409477e-01 -2.45891258e-01] | [5.9905829429626465, 3.1897833347320557] |
896b9d67-c025-40e4-a1f1-92ed6ecfa39c | g-map-general-memory-augmented-pre-trained | 2212.03613 | null | https://arxiv.org/abs/2212.03613v2 | https://arxiv.org/pdf/2212.03613v2.pdf | G-MAP: General Memory-Augmented Pre-trained Language Model for Domain Tasks | Recently, domain-specific PLMs have been proposed to boost the task performance of specific domains (e.g., biomedical and computer science) by continuing to pre-train general PLMs with domain-specific corpora. However, this Domain-Adaptive Pre-Training (DAPT; Gururangan et al. (2020)) tends to forget the previous general knowledge acquired by general PLMs, which leads to a catastrophic forgetting phenomenon and sub-optimal performance. To alleviate this problem, we propose a new framework of General Memory Augmented Pre-trained Language Model (G-MAP), which augments the domain-specific PLM by a memory representation built from the frozen general PLM without losing any general knowledge. Specifically, we propose a new memory-augmented layer, and based on it, different augmented strategies are explored to build the memory representation and then adaptively fuse it into the domain-specific PLM. We demonstrate the effectiveness of G-MAP on various domains (biomedical and computer science publications, news, and reviews) and different kinds (text classification, QA, NER) of tasks, and the extensive results show that the proposed G-MAP can achieve SOTA results on all tasks. | ['Qun Liu', 'Xin Jiang', 'Guangyong Chen', 'Lifeng Shang', 'Jiaxin Shi', 'Wei zhang', 'Yichun Yin', 'Zhongwei Wan'] | 2022-12-07 | null | null | null | null | ['general-knowledge'] | ['miscellaneous'] | [ 1.33784056e-01 1.41411990e-01 -1.40546694e-01 -2.85345733e-01
-5.21542728e-01 2.41487729e-03 5.96499145e-01 1.60062209e-01
-6.23470247e-01 1.05605304e+00 2.10755065e-01 -6.35090247e-02
4.93003279e-02 -7.59712815e-01 -7.19695926e-01 -6.80138767e-01
2.67042160e-01 6.13894880e-01 5.18656313e-01 -2.16171697e-01
2.97089338e-01 2.93070465e-01 -1.06399941e+00 4.58471566e-01
1.22637236e+00 5.65115452e-01 8.87449861e-01 1.78608820e-01
-5.03790557e-01 7.20584869e-01 -5.31926870e-01 -3.45886022e-01
-4.20290411e-01 -3.75024199e-01 -8.99934232e-01 -5.41935973e-02
-8.57840106e-02 -1.19925544e-01 -5.31439722e-01 9.93294537e-01
5.67969799e-01 4.69038606e-01 5.30680776e-01 -9.25653040e-01
-1.05261791e+00 7.32907176e-01 -6.34343386e-01 4.02070105e-01
-1.40630826e-01 -2.50907335e-02 3.00516099e-01 -1.23995602e+00
5.51304102e-01 1.52260482e+00 6.81378722e-01 8.94548237e-01
-8.79376292e-01 -7.93560386e-01 4.50719088e-01 2.50614643e-01
-1.44356465e+00 -3.89349490e-01 6.55856013e-01 -9.31740403e-02
1.03354347e+00 -3.23195100e-01 1.67384416e-01 1.24591887e+00
7.48219013e-01 1.03730464e+00 9.26544309e-01 -3.93403292e-01
1.35197833e-01 2.97984570e-01 5.55569232e-01 7.04065979e-01
4.00318384e-01 -2.53417820e-01 -7.06499994e-01 -1.88515007e-01
6.48375988e-01 1.84442461e-01 -3.71808976e-01 -5.96140809e-02
-1.05546916e+00 5.46111941e-01 1.01460017e-01 6.12027228e-01
-6.18037045e-01 -4.47116613e-01 4.54059482e-01 3.87805343e-01
5.90528667e-01 2.51710236e-01 -8.37324619e-01 2.52111733e-01
-1.02115941e+00 -4.11832072e-02 4.62772816e-01 9.97066736e-01
8.20879936e-01 5.60307391e-02 -2.22603023e-01 1.15524232e+00
1.86295882e-01 4.10476118e-01 1.30725873e+00 -3.46177578e-01
4.43523914e-01 8.38448048e-01 -6.23785555e-02 -8.33208919e-01
-6.93398654e-01 -7.29237020e-01 -1.28902376e+00 -5.09789765e-01
-1.79686144e-01 -1.60761029e-01 -1.30953205e+00 2.03968978e+00
-6.00732118e-02 3.73963445e-01 3.19704890e-01 5.71771502e-01
1.09154820e+00 1.00987661e+00 7.23812044e-01 -5.53457141e-01
1.15149844e+00 -1.20793629e+00 -9.99423921e-01 -7.24907696e-01
6.27962589e-01 -3.35571408e-01 1.13711500e+00 5.43639660e-01
-1.05039072e+00 -9.38179255e-01 -1.10741079e+00 -6.96238279e-02
-5.53679645e-01 1.75378948e-01 1.66573107e-01 2.82510847e-01
-9.60166335e-01 7.25670457e-01 -8.02302182e-01 -5.19702137e-01
3.38172644e-01 1.77821591e-01 -3.89181972e-01 -5.15857279e-01
-1.68840730e+00 1.26288486e+00 1.15633249e+00 -4.06696759e-02
-1.05267119e+00 -6.46492481e-01 -6.53949559e-01 5.36558963e-02
2.77923763e-01 -8.30018342e-01 1.02862549e+00 -8.07771146e-01
-1.38513398e+00 7.81892896e-01 -1.89775243e-01 -5.23721039e-01
9.93003920e-02 -4.48980421e-01 -8.74324262e-01 1.03011094e-02
1.38971219e-02 6.42405272e-01 9.26843345e-01 -1.10268104e+00
-5.47401845e-01 -2.63689429e-01 -2.92174071e-01 2.92386532e-01
-7.93318272e-01 -2.52470165e-01 -4.38730150e-01 -8.24149728e-01
1.53561030e-02 -6.36850476e-01 -2.41936237e-01 -4.97601181e-01
1.11942291e-01 -2.99098551e-01 6.84183896e-01 -1.12133992e+00
1.58782995e+00 -2.27959895e+00 2.72520959e-01 -3.10119033e-01
1.40977398e-01 7.15323329e-01 -4.85148996e-01 1.97510287e-01
-3.58782783e-02 -8.02466422e-02 -5.13002694e-01 -4.86345440e-01
-3.52713138e-01 5.91594279e-01 -2.78980941e-01 9.36826468e-02
2.19760671e-01 8.90830994e-01 -1.14666820e+00 -5.76686144e-01
-4.09270823e-02 4.55071509e-01 -2.63573498e-01 1.78849623e-01
-3.52717042e-01 4.35709059e-01 -5.04997671e-01 3.33788693e-01
9.52808321e-01 -4.62740421e-01 2.40060270e-01 -5.80110848e-02
2.20998585e-01 3.12521726e-01 -8.94946635e-01 2.13788986e+00
-4.57516402e-01 6.83423728e-02 -1.71651244e-01 -1.08861506e+00
1.04665947e+00 3.85268092e-01 3.86171043e-02 -7.40681112e-01
1.38442278e-01 1.59377992e-01 -1.54957563e-01 -2.87488639e-01
5.59620976e-01 -4.61894333e-01 -1.53696500e-02 2.81447262e-01
6.46966279e-01 4.16663617e-01 -6.48136064e-02 2.43297085e-01
1.07206798e+00 1.18855163e-01 4.21989948e-01 -5.12919962e-01
9.73873138e-01 2.78414208e-02 8.26948941e-01 6.36217833e-01
-1.40810132e-01 4.35976177e-01 -1.40740812e-01 -3.91068071e-01
-6.70283437e-01 -9.90930319e-01 -1.68869495e-02 1.28852272e+00
1.14145353e-01 -3.11318636e-01 -6.41303897e-01 -8.21241796e-01
-2.47349769e-01 1.13773561e+00 -3.98095816e-01 -9.88176823e-01
-7.24964976e-01 -1.18146920e+00 4.99128610e-01 4.56246674e-01
1.12968469e+00 -1.33309877e+00 -8.90253112e-02 4.86954987e-01
-3.52699310e-01 -9.10254180e-01 -4.87842679e-01 2.54561394e-01
-1.31835091e+00 -5.61515868e-01 -9.79141295e-01 -1.00475240e+00
7.13607669e-01 2.38073692e-01 1.21785390e+00 -1.45060673e-01
3.32666874e-01 3.49560648e-01 -3.71916473e-01 -2.45159030e-01
-5.97053766e-01 4.29390699e-01 2.54489481e-01 -9.44539383e-02
4.84565616e-01 -5.41635752e-01 -2.37113193e-01 9.36867204e-03
-1.11368871e+00 9.99593139e-02 1.11298513e+00 1.18135977e+00
6.03916943e-01 4.18668836e-01 1.22772467e+00 -1.10717952e+00
8.15742493e-01 -7.90129602e-01 4.35792878e-02 4.94966328e-01
-8.36001456e-01 3.52654248e-01 8.11074495e-01 -9.01129127e-01
-1.65734935e+00 -3.61195982e-01 -1.64808139e-01 -4.88078505e-01
-4.82809357e-03 8.76495481e-01 -5.44201493e-01 1.59742668e-01
6.62336409e-01 8.84773612e-01 -2.20205158e-01 -9.00499046e-01
2.56172478e-01 6.20594323e-01 6.77865505e-01 -6.55296624e-01
4.26267594e-01 -6.79816976e-02 -3.76956135e-01 -6.80050135e-01
-1.10224962e+00 -1.88662544e-01 -6.18491530e-01 3.04742038e-01
4.39376682e-01 -1.07818067e+00 2.25329190e-01 7.00346649e-01
-1.34879565e+00 -1.98722929e-01 -1.30437985e-01 4.74976212e-01
-2.96941102e-01 5.51461875e-01 -7.69158840e-01 -4.79799777e-01
-6.16204202e-01 -5.91627419e-01 6.15845025e-01 4.07021672e-01
-9.52202380e-02 -1.11219358e+00 1.57979116e-01 -6.92182556e-02
3.70993525e-01 -2.51737654e-01 1.25082898e+00 -7.44609952e-01
-4.45097461e-02 1.25266731e-01 -1.10857151e-01 6.18942499e-01
1.39057696e-01 -8.84312630e-01 -8.55972290e-01 -5.64112782e-01
3.61235142e-01 -1.93555966e-01 1.13740337e+00 1.44838467e-01
1.02432370e+00 -3.92857730e-01 -6.75535500e-01 2.71789432e-01
1.27746058e+00 3.46314222e-01 6.41388416e-01 3.41643125e-01
3.46745044e-01 2.22816482e-01 7.34891415e-01 1.87660530e-01
3.16861391e-01 1.56170204e-01 -1.50110394e-01 1.06555976e-01
-2.60346472e-01 -4.72975850e-01 4.69473094e-01 1.35180843e+00
2.02499643e-01 -3.73148561e-01 -7.97830820e-01 8.57415318e-01
-1.94064903e+00 -4.74883497e-01 3.02535444e-01 2.09254909e+00
1.19947302e+00 3.27775598e-01 -3.92645150e-01 -1.57454550e-01
8.93407226e-01 -5.80748729e-02 -7.46936500e-01 -1.68910593e-01
-3.28711271e-01 2.42225870e-01 2.37586349e-01 2.78101265e-01
-8.99320483e-01 1.12841702e+00 6.04575062e+00 1.15763593e+00
-9.90163743e-01 5.99072754e-01 4.41270590e-01 2.37395018e-01
-4.18309011e-02 -2.76607513e-01 -9.96662080e-01 6.82616234e-01
1.36190248e+00 -5.71221650e-01 1.23467758e-01 8.19849610e-01
-2.64210135e-01 1.58250138e-01 -8.71069431e-01 1.08795416e+00
2.40639552e-01 -1.03094840e+00 7.04075933e-01 -4.34427410e-01
5.74336827e-01 -5.52960187e-02 3.55865657e-02 9.79090929e-01
2.20561013e-01 -7.85909832e-01 2.75137633e-01 7.50972450e-01
5.81217945e-01 -7.62459815e-01 8.74797761e-01 9.05464590e-01
-7.99416184e-01 -5.50092719e-02 -8.60380888e-01 1.51448265e-01
1.28111988e-01 7.73754120e-01 -5.79409480e-01 8.87596488e-01
6.45032406e-01 7.51584649e-01 -6.18682742e-01 8.45919132e-01
-1.73828661e-01 6.79487646e-01 -2.40691602e-02 2.78172284e-01
3.82495858e-02 3.04623008e-01 4.30752665e-01 1.43640423e+00
4.71936196e-01 3.76881212e-01 4.97194827e-02 6.59967542e-01
-2.64785826e-01 1.15636371e-01 -3.39834362e-01 -2.19487518e-01
4.84649509e-01 1.05218625e+00 -4.95458871e-01 -7.27862597e-01
-3.69658053e-01 1.19599390e+00 4.69252288e-01 2.90055066e-01
-6.32517159e-01 -3.94940346e-01 3.36008698e-01 1.53860033e-01
3.65452021e-02 -2.17215344e-01 -1.93215787e-01 -1.27963960e+00
-1.81342259e-01 -7.21310198e-01 6.78906798e-01 -7.55952179e-01
-1.56955886e+00 8.64577472e-01 -3.57750989e-02 -9.73252356e-01
-1.03718573e-02 -4.84064937e-01 -3.04055929e-01 9.68512356e-01
-1.64590359e+00 -1.03451490e+00 -6.64327443e-02 7.47069418e-01
6.61864758e-01 -4.65074897e-01 8.27794552e-01 6.13009095e-01
-5.11705697e-01 6.39913976e-01 9.68825370e-02 -8.28774571e-02
9.58682477e-01 -7.17054844e-01 2.15516999e-01 7.66486406e-01
-3.06928158e-01 9.19753015e-01 5.59518695e-01 -1.00171804e+00
-1.17929339e+00 -1.47352695e+00 1.25260103e+00 -4.93026733e-01
3.26727182e-01 -3.98181900e-02 -1.59379387e+00 8.44895422e-01
3.40485632e-01 -3.02263081e-01 4.55902189e-01 -6.34429976e-02
-8.47392157e-02 -3.05851530e-02 -1.30162108e+00 3.66737187e-01
1.08582616e+00 -2.35064834e-01 -1.37313187e+00 1.56729624e-01
1.05824006e+00 -2.72911340e-01 -7.29371965e-01 6.07647002e-01
2.12471304e-03 -2.72760451e-01 8.31703126e-01 -7.28627324e-01
1.79162711e-01 -3.74534398e-01 1.27938479e-01 -1.53843987e+00
-7.77685583e-01 -1.62634090e-01 -4.04147804e-01 1.33472764e+00
1.76445350e-01 -6.95956707e-01 4.52281266e-01 3.05872858e-01
-5.26818216e-01 -3.48114014e-01 -8.72262418e-01 -9.23521876e-01
4.35934424e-01 -7.63341412e-02 5.33421695e-01 1.15402818e+00
-3.14484350e-02 8.87766480e-01 -6.20583355e-01 2.05898628e-01
3.55052382e-01 -1.75343066e-01 1.96358040e-01 -1.34565759e+00
-1.59720659e-01 -1.58148214e-01 9.22883116e-03 -1.39455175e+00
2.57876456e-01 -9.22655582e-01 1.07647471e-01 -1.49994910e+00
3.60186160e-01 -2.95396984e-01 -9.68386173e-01 7.86247849e-01
-5.06282210e-01 -2.90607572e-01 -9.02728587e-02 2.00476542e-01
-9.15172756e-01 7.84784138e-01 1.30105519e+00 -3.05156499e-01
-3.59457046e-01 -3.58461350e-01 -9.23045039e-01 6.53202653e-01
7.12040126e-01 -5.66644490e-01 -6.12050891e-01 -5.58833122e-01
-1.11407219e-02 1.38470139e-02 4.13186513e-02 -1.16374481e+00
4.16861415e-01 -1.43435746e-01 5.66007912e-01 -6.46067798e-01
1.68968350e-01 -6.30037487e-01 9.21078175e-02 5.77337027e-01
-1.58299550e-01 -4.49799970e-02 6.31038547e-01 5.67695439e-01
-4.05333340e-01 -3.93967420e-01 1.01679265e+00 -5.04162312e-01
-1.05610561e+00 3.29101831e-01 -2.97622889e-01 6.08918741e-02
6.31123543e-01 2.16401309e-01 -4.19757187e-01 4.89828773e-02
-1.01471269e+00 3.89548600e-01 7.07522109e-02 6.85939074e-01
8.09601128e-01 -1.48624814e+00 -7.73116767e-01 2.21901625e-01
8.01954977e-03 2.12274073e-03 7.26429880e-01 5.81692815e-01
4.87741567e-02 5.30862510e-01 -4.18448955e-01 -3.01036328e-01
-7.30387032e-01 1.02193213e+00 1.72085032e-01 -7.37521708e-01
-6.72404051e-01 7.22970128e-01 6.54068232e-01 -6.25445902e-01
6.11882433e-02 -2.47648820e-01 -4.59758699e-01 -1.31581396e-01
8.09147656e-01 2.09417954e-01 2.96902835e-01 -2.84705251e-01
-4.60813403e-01 2.32856378e-01 -5.68973064e-01 2.08774567e-01
1.38836992e+00 -3.44924212e-01 -1.16956644e-01 5.23967922e-01
7.83754706e-01 -3.58712494e-01 -8.49942446e-01 -7.81585813e-01
2.01551944e-01 8.22517648e-02 -2.17610486e-02 -1.05879509e+00
-9.92925584e-01 1.07026982e+00 5.42691052e-01 -4.00851220e-01
1.46069813e+00 -2.03465641e-01 1.12884319e+00 4.86714840e-01
6.48570240e-01 -1.21711135e+00 1.74151048e-01 8.41919780e-01
7.88655818e-01 -8.63103390e-01 -1.82211027e-01 8.15838762e-03
-6.46359444e-01 9.91469085e-01 8.67496789e-01 -4.49708477e-02
7.71657467e-01 6.02186285e-02 -2.73474902e-01 2.03122810e-01
-9.31359589e-01 1.28542766e-01 2.23126903e-01 5.76499879e-01
2.31034234e-01 -1.31718233e-01 -7.17768192e-01 1.57429242e+00
2.54363239e-01 5.67411304e-01 2.37300307e-01 1.07949603e+00
-7.47761846e-01 -1.18824661e+00 -1.58612326e-01 5.10478735e-01
-2.97310144e-01 -3.95086199e-01 -7.60460496e-02 5.45222223e-01
3.79931301e-01 6.39725268e-01 -3.45285565e-01 -5.51918507e-01
3.08174759e-01 7.15495110e-01 3.69697303e-01 -7.84314334e-01
-3.60382915e-01 -1.06396817e-01 -9.48226973e-02 -6.30235597e-02
-3.11825544e-01 -2.91346073e-01 -1.30041492e+00 -2.43705496e-01
-6.82785064e-02 4.06707734e-01 2.59609497e-03 1.04440081e+00
6.06434584e-01 7.20370650e-01 -6.08987138e-02 -2.73967057e-01
-3.77051771e-01 -1.36379886e+00 -7.00320959e-01 2.97929138e-01
1.49687961e-01 -9.52005506e-01 -2.22259298e-01 1.40515313e-01] | [10.429265022277832, 8.136680603027344] |
2a3ed11a-c345-4e95-abfc-a9205bf39a90 | lamad-a-linguistic-attentional-model-for | null | null | https://aclanthology.org/2021.findings-emnlp.317 | https://aclanthology.org/2021.findings-emnlp.317.pdf | LAMAD: A Linguistic Attentional Model for Arabic Text Diacritization | In Arabic Language, diacritics are used to specify meanings as well as pronunciations. However, diacritics are often omitted from written texts, which increases the number of possible meanings and pronunciations. This leads to an ambiguous text and makes the computational process on undiacritized text more difficult. In this paper, we propose a Linguistic Attentional Model for Arabic text Diacritization (LAMAD). In LAMAD, a new linguistic feature representation is presented, which utilizes both word and character contextual features. Then, a linguistic attention mechanism is proposed to capture the important linguistic features. In addition, we explore the impact of the linguistic features extracted from the text on Arabic text diacritization (ATD) by introducing them to the linguistic attention mechanism. The extensive experimental results on three datasets with different sizes illustrate that LAMAD outperforms the existing state-of-the-art models. | ['Jianliang Gao', 'Raeed Al-Sabri'] | null | null | null | null | findings-emnlp-2021-11 | ['arabic-text-diacritization'] | ['natural-language-processing'] | [-1.37159443e-02 -4.71537620e-01 -1.64224952e-01 -3.70215476e-01
-1.32846653e-01 -5.24507701e-01 5.27946055e-01 3.96160036e-01
-5.99199891e-01 4.86818433e-01 4.72329617e-01 -3.00115049e-01
1.63274363e-01 -8.37483823e-01 -2.32978195e-01 -7.26053596e-01
4.49721247e-01 7.44344145e-02 -1.00925229e-01 -4.82710987e-01
5.68808019e-01 4.42165464e-01 -1.29107165e+00 4.22040731e-01
1.16662002e+00 8.39306474e-01 3.20276201e-01 1.44150332e-01
-6.99350536e-01 6.20198667e-01 -9.65767086e-01 -3.94862205e-01
-1.26574978e-01 -3.71697217e-01 -6.65584922e-01 2.37232730e-01
-4.52205762e-02 -2.63541460e-01 -4.74745296e-02 1.14431429e+00
2.69668907e-01 3.22141230e-01 8.11158717e-01 -7.69954622e-01
-1.26620400e+00 9.95377421e-01 -7.56696343e-01 3.95464808e-01
3.48554969e-01 -1.37021035e-01 1.18168199e+00 -1.39343274e+00
1.08090073e-01 1.75958335e+00 2.68363118e-01 1.86665639e-01
-7.26066113e-01 -5.57474911e-01 5.94412208e-01 3.33837986e-01
-1.53279567e+00 -1.97923779e-01 7.81379282e-01 -2.28139982e-01
1.09698021e+00 1.59960702e-01 3.52664858e-01 5.70287645e-01
3.14022422e-01 1.10248077e+00 8.33203197e-01 -9.11717534e-01
-1.08917356e-01 -1.36403829e-01 5.19330621e-01 4.43720728e-01
2.23169729e-01 -4.82225209e-01 -2.48873219e-01 2.80352443e-01
4.24973756e-01 1.40675968e-02 -2.38666907e-01 4.89533365e-01
-1.16251075e+00 8.90830040e-01 2.38792196e-01 6.30148828e-01
-3.33600074e-01 -2.22925231e-01 4.77490604e-01 9.03993994e-02
3.00578982e-01 2.70835549e-01 -3.35381240e-01 -3.24769318e-02
-4.00178879e-01 -1.82710141e-01 3.59813035e-01 8.44960749e-01
7.21720099e-01 1.54406995e-01 -2.64875323e-01 9.03883100e-01
5.54402947e-01 7.48855352e-01 9.58883166e-01 -4.00441021e-01
5.79606473e-01 9.84292030e-01 7.92610198e-02 -1.19146430e+00
-2.61368543e-01 1.88892305e-01 -7.17013419e-01 -2.13608921e-01
3.20799738e-01 -3.38570207e-01 -8.74745607e-01 1.26099432e+00
7.54915550e-02 -4.11777377e-01 2.55670965e-01 8.53649557e-01
7.82964647e-01 1.14264250e+00 9.27513093e-02 -3.05245310e-01
1.53924704e+00 -1.02627528e+00 -1.26898146e+00 -4.34690595e-01
4.47269201e-01 -1.17673051e+00 1.68064833e+00 3.25084805e-01
-8.44871044e-01 -5.44400394e-01 -1.08595431e+00 -3.27332675e-01
-6.05540514e-01 6.49689853e-01 3.98210526e-01 6.03947699e-01
-3.86996061e-01 3.92576084e-02 -7.85879910e-01 -9.63003114e-02
1.13255896e-01 2.49864519e-01 -9.43571702e-03 2.79832501e-02
-1.47676265e+00 9.40985501e-01 7.99838126e-01 5.26235223e-01
-1.49958059e-01 1.73936665e-01 -1.00181091e+00 2.11856291e-01
2.81731963e-01 1.37728751e-01 9.29928243e-01 -1.35773861e+00
-1.59351003e+00 5.00406206e-01 -3.27337265e-01 -2.06084140e-02
-1.70459524e-02 -4.64681864e-01 -5.73893547e-01 1.11757718e-01
-1.50688784e-02 4.27447289e-01 9.54912305e-01 -8.66356730e-01
-6.38788939e-01 -1.69052467e-01 1.47219643e-01 5.97416878e-01
-6.27106905e-01 3.74773383e-01 -5.71285665e-01 -1.16395915e+00
1.31978348e-01 -7.14348137e-01 1.72953948e-01 -3.40837628e-01
-3.74908119e-01 -6.55152977e-01 8.80136073e-01 -7.88220167e-01
1.78373778e+00 -2.38207722e+00 1.97369188e-01 1.53543502e-01
-1.16768800e-01 4.20463979e-01 -9.21630785e-02 3.95532250e-01
1.60265207e-01 3.26183528e-01 -2.91873902e-01 -9.09105837e-02
8.68866816e-02 5.21670818e-01 -2.23227069e-01 2.12755390e-02
6.54508233e-01 8.53179455e-01 -8.73430610e-01 -5.76389790e-01
1.83181688e-01 3.81242543e-01 -2.50156939e-01 -6.18443042e-02
-1.39983520e-01 4.96990718e-02 -5.96933782e-01 6.90750599e-01
7.26087213e-01 1.02379462e-02 2.04567567e-01 -1.31042436e-01
-2.64667541e-01 5.58050871e-01 -9.10937548e-01 1.13819182e+00
-4.75116581e-01 4.23415810e-01 -3.55730087e-01 -7.32375205e-01
8.65237236e-01 1.41180426e-01 -8.79338533e-02 -5.06703496e-01
7.18787193e-01 1.40650481e-01 3.87417734e-01 -4.69029814e-01
7.33336747e-01 7.45247642e-04 -1.75859183e-01 6.30867958e-01
-3.75473380e-01 7.12418482e-02 6.35465384e-01 9.92057994e-02
2.46072769e-01 -6.73669949e-02 5.10454416e-01 -4.49233353e-01
9.96853590e-01 -2.10641742e-01 4.78391290e-01 2.54461795e-01
-7.05207288e-02 4.05479193e-01 5.77710330e-01 -3.79371524e-01
-5.62654197e-01 -7.46977210e-01 -1.56217843e-01 1.27953172e+00
1.73685744e-01 -5.71954250e-01 -8.80903184e-01 -8.62370431e-01
-1.68703839e-01 9.39843774e-01 -7.17677236e-01 -2.42622897e-01
-7.58954942e-01 -9.19419467e-01 4.52895701e-01 5.72306752e-01
4.65412378e-01 -1.55941606e+00 -3.03520977e-01 2.44677424e-01
-4.60510373e-01 -8.66545975e-01 -7.19122767e-01 -7.19189644e-02
-3.96893561e-01 -7.74165690e-01 -5.48634708e-01 -1.14151371e+00
8.99937093e-01 3.32381636e-01 6.66424870e-01 2.60973334e-01
1.74375609e-01 -1.89295918e-01 -8.42938840e-01 -6.46727562e-01
-3.71715516e-01 1.43192723e-01 -3.37708257e-02 2.21019700e-01
9.89689350e-01 4.49453630e-02 -1.63997024e-01 2.16685962e-02
-1.15024710e+00 -8.18008780e-02 6.43631577e-01 9.09018815e-01
4.89072829e-01 2.04602703e-02 5.73566377e-01 -8.03787172e-01
1.02635705e+00 -3.46555680e-01 -4.16494071e-01 3.89088124e-01
-4.00482863e-01 2.81924218e-01 1.05724800e+00 -7.15181470e-01
-1.20634818e+00 -2.54326075e-01 -2.85488337e-01 4.46289107e-02
2.61066724e-02 1.21079516e+00 -4.99387980e-01 3.65794122e-01
2.49181762e-01 4.43563670e-01 1.11788198e-01 -4.26579982e-01
2.05812007e-01 1.06340849e+00 8.26453790e-02 -5.06572306e-01
4.69564080e-01 1.34005055e-01 -3.81705821e-01 -1.12149262e+00
-9.24656987e-01 -2.96825785e-02 -6.87604487e-01 1.82477385e-01
7.70867050e-01 -6.16890609e-01 -3.97573680e-01 7.85502493e-01
-1.24065256e+00 6.63654581e-02 4.22100769e-03 4.57230300e-01
-2.45423652e-02 8.36040199e-01 -7.86920369e-01 -6.28902674e-01
-2.95580178e-01 -1.34806907e+00 8.05081546e-01 3.33489895e-01
-2.24442944e-01 -9.23566759e-01 -5.30815840e-01 -5.96312471e-02
2.71848381e-01 -2.22551897e-01 1.30992126e+00 -6.25914335e-01
-8.60727727e-02 1.40494341e-02 -4.59155113e-01 3.73570830e-01
7.30637252e-01 2.78161913e-01 -6.64007366e-01 -2.76992675e-02
-1.28513038e-01 -2.54393399e-01 6.57902896e-01 1.81203127e-01
9.62923765e-01 -5.77052593e-01 1.06341735e-01 1.32476151e-01
9.15291965e-01 3.62028122e-01 6.67436063e-01 4.10529494e-01
7.15485752e-01 4.72241908e-01 1.00819778e+00 5.04421353e-01
4.90629077e-01 3.01906765e-01 2.98196316e-01 1.29946113e-01
4.00970988e-02 8.25587139e-02 6.51948094e-01 1.39443874e+00
2.23006248e-01 -5.61241567e-01 -8.95794332e-01 5.58813393e-01
-1.66783333e+00 -4.08822477e-01 -2.46682242e-01 1.79112971e+00
1.19423628e+00 1.14989109e-01 -3.45553219e-01 2.40905911e-01
8.87556016e-01 1.63654119e-01 -2.42089555e-01 -7.68669426e-01
-3.72651517e-01 6.99328780e-02 1.08728237e-01 6.98082566e-01
-1.12493503e+00 1.44009531e+00 5.49722862e+00 1.05303490e+00
-1.35308957e+00 -2.17852190e-01 4.46502656e-01 3.55235934e-01
-4.03577894e-01 -1.44066572e-01 -9.79698241e-01 6.18053436e-01
4.04805779e-01 -3.86720121e-01 3.89348418e-01 4.21329170e-01
2.51031280e-01 -1.00886874e-01 -6.09947920e-01 8.39329064e-01
3.98854405e-01 -7.25266337e-01 6.48671567e-01 -3.83373857e-01
4.90891248e-01 -3.26598465e-01 7.89184645e-02 1.73506007e-01
5.88110760e-02 -8.71928453e-01 7.44404137e-01 2.75975585e-01
6.90633595e-01 -1.11194825e+00 9.12542522e-01 1.82846114e-01
-1.17064786e+00 -4.16611359e-02 -6.26167357e-01 -2.26486936e-01
-6.10826463e-02 8.26334730e-02 -5.21903276e-01 4.32950437e-01
4.68078166e-01 9.00623739e-01 -8.41110528e-01 5.26634097e-01
-7.41084635e-01 7.44308472e-01 -3.33227128e-01 -6.34270191e-01
7.06967473e-01 -3.65193725e-01 3.56016606e-01 1.24506700e+00
2.51203448e-01 1.85658827e-01 2.35818714e-01 6.26770437e-01
-3.24023925e-02 7.24315882e-01 -1.07282758e-01 -5.30728281e-01
5.91903985e-01 1.06485426e+00 -9.06324685e-01 -4.22751963e-01
-5.77214599e-01 1.00407577e+00 3.41265410e-01 3.38834018e-01
-7.61642933e-01 -9.48764801e-01 4.38904017e-01 -2.82127202e-01
4.04897988e-01 -3.63336116e-01 -1.04003340e-01 -1.23261893e+00
4.75342236e-02 -1.03089345e+00 2.95214027e-01 -5.95356047e-01
-1.48580515e+00 6.63709462e-01 -1.76602215e-01 -1.14673805e+00
-7.17371404e-02 -7.03028023e-01 -4.63467211e-01 1.10085058e+00
-1.77511406e+00 -1.15942883e+00 3.87163609e-02 3.67871463e-01
8.43960524e-01 -1.64997473e-01 8.43662202e-01 3.18827212e-01
-7.61387765e-01 7.19750285e-01 2.10855186e-01 5.25401592e-01
8.02927375e-01 -1.02680075e+00 3.53609860e-01 1.04008365e+00
-1.99092831e-03 9.27003264e-01 2.95909792e-01 -5.53876817e-01
-1.02419412e+00 -1.04597127e+00 1.39464259e+00 -1.48618907e-01
8.42984140e-01 -1.43620729e-01 -1.09963751e+00 7.77392387e-01
5.76956868e-01 -3.99188519e-01 7.35900819e-01 -1.24921434e-01
6.31693527e-02 1.35496527e-01 -6.73783302e-01 1.03671861e+00
4.00932074e-01 -2.23044068e-01 -9.28782284e-01 6.17978685e-02
7.93720305e-01 -2.43680298e-01 -3.50192606e-01 -7.46057257e-02
5.08201659e-01 -4.57837194e-01 5.76620519e-01 -4.73897010e-01
3.98820728e-01 -5.03060877e-01 -8.67816210e-02 -1.22307312e+00
-4.03050274e-01 -4.85731900e-01 -2.86878180e-03 1.26140332e+00
4.04834837e-01 -6.92857623e-01 8.56647715e-02 2.38618359e-01
-1.37421876e-01 -5.07147968e-01 -6.30187929e-01 -6.20839477e-01
3.74304593e-01 -1.19003013e-01 8.71100068e-01 1.11008549e+00
1.54251754e-01 7.44648397e-01 -4.83283624e-02 4.56287991e-03
-1.34469911e-01 1.82618439e-01 5.37716113e-02 -8.88579607e-01
2.34060958e-01 -6.60922289e-01 -1.78577259e-01 -1.35176659e+00
4.19748187e-01 -8.23808253e-01 7.14268163e-02 -1.30451369e+00
-3.06347370e-01 -4.24683928e-01 -3.00213695e-01 7.36646235e-01
-6.43058777e-01 8.85935202e-02 3.13176572e-01 4.36580442e-02
-3.10088903e-01 7.78783023e-01 1.26282132e+00 -3.30614060e-01
-3.61518532e-01 -3.03363264e-01 -6.97764933e-01 9.69164729e-01
1.05846715e+00 -2.97516108e-01 -2.14401141e-01 -7.47296691e-01
2.07443461e-01 -4.26100701e-01 -3.84963781e-01 -5.61643422e-01
-5.19873425e-02 -4.27033901e-01 3.54242772e-01 -7.88466275e-01
1.18753098e-01 -7.08729684e-01 -8.39132726e-01 3.02923530e-01
-1.91028699e-01 5.65833390e-01 4.02241051e-01 3.85873914e-02
-3.75249833e-01 -6.33378386e-01 6.19353294e-01 -6.41534850e-02
-7.38728762e-01 3.39038484e-02 -1.04342687e+00 1.25113949e-01
7.47434258e-01 -2.42121577e-01 -7.42763504e-02 -5.33916019e-02
-5.30689001e-01 2.84581780e-01 2.62794405e-01 7.24059105e-01
8.26599419e-01 -1.35873628e+00 -8.22020411e-01 5.64253569e-01
1.93031892e-01 -6.57131150e-02 -1.61853790e-01 6.43206716e-01
-8.12957704e-01 3.10517132e-01 -1.85497656e-01 -1.79678246e-01
-1.29888868e+00 6.41583204e-01 6.38106912e-02 1.11529520e-02
-2.57025123e-01 6.67367399e-01 2.14347601e-01 -4.08663929e-01
1.74483448e-01 -5.71539342e-01 -7.54900098e-01 3.17409307e-01
6.32054985e-01 1.26954690e-01 -3.03640626e-02 -1.08248365e+00
-3.65302682e-01 4.58878696e-01 -4.54216897e-01 2.30198279e-02
8.64875615e-01 -5.47981679e-01 -5.26901186e-01 5.70445418e-01
7.61434257e-01 5.26071191e-01 -9.31859255e-01 -2.56602496e-01
2.59301811e-01 -4.55119282e-01 -5.35062514e-02 -8.32535088e-01
-7.57432580e-01 1.27354181e+00 -1.73301492e-02 1.73155844e-01
1.07060289e+00 -4.07848686e-01 9.87065852e-01 6.65201128e-01
-8.52679089e-02 -1.33112478e+00 5.36440387e-02 1.30419695e+00
9.66582417e-01 -1.10552275e+00 -1.59674495e-01 -4.51217800e-01
-9.34476793e-01 1.31733394e+00 8.49081218e-01 -1.03149310e-01
6.49886787e-01 1.45515978e-01 4.11539555e-01 2.11542845e-01
-4.32064950e-01 -1.99237540e-01 2.48978123e-01 4.77536052e-01
8.78533363e-01 -5.95795028e-02 -8.84485126e-01 9.76376951e-01
-2.73826092e-01 -5.42594194e-01 7.85622120e-01 1.04682386e+00
-5.51552296e-01 -1.32862139e+00 -4.86993253e-01 2.51661837e-01
-6.69773579e-01 -6.43896401e-01 -6.63935900e-01 7.03557909e-01
1.76535726e-01 1.06391513e+00 2.61364728e-01 4.55818884e-02
-1.83305722e-02 9.47300121e-02 3.28013569e-01 -7.46094227e-01
-5.39825618e-01 4.30718720e-01 -1.44944146e-01 1.84605658e-01
-2.55172372e-01 -6.34921074e-01 -1.84119630e+00 -2.65971392e-01
-7.66241372e-01 2.63567179e-01 2.78648376e-01 1.05378926e+00
2.33846173e-01 5.68800390e-01 5.98258138e-01 -5.37001848e-01
-4.82303619e-01 -1.12990510e+00 -5.00875533e-01 2.04629511e-01
2.38960758e-01 -6.53449833e-01 -1.69483259e-01 6.54710233e-02] | [10.15091323852539, 10.131697654724121] |
db8568c1-2191-4932-b12e-b8cb772a0c86 | a-study-of-few-shot-audio-classification | 2012.01573 | null | https://arxiv.org/abs/2012.01573v1 | https://arxiv.org/pdf/2012.01573v1.pdf | A Study of Few-Shot Audio Classification | Advances in deep learning have resulted in state-of-the-art performance for many audio classification tasks but, unlike humans, these systems traditionally require large amounts of data to make accurate predictions. Not every person or organization has access to those resources, and the organizations that do, like our field at large, do not reflect the demographics of our country. Enabling people to use machine learning without significant resource hurdles is important, because machine learning is an increasingly useful tool for solving problems, and can solve a broader set of problems when put in the hands of a broader set of people. Few-shot learning is a type of machine learning designed to enable the model to generalize to new classes with very few examples. In this research, we address two audio classification tasks (speaker identification and activity classification) with the Prototypical Network few-shot learning algorithm, and assess performance of various encoder architectures. Our encoders include recurrent neural networks, as well as one- and two-dimensional convolutional neural networks. We evaluate our model for speaker identification on the VoxCeleb dataset and ICSI Meeting Corpus, obtaining 5-shot 5-way accuracies of 93.5% and 54.0%, respectively. We also evaluate for activity classification from audio using few-shot subsets of the Kinetics~600 dataset and AudioSet, both drawn from Youtube videos, obtaining 51.5% and 35.2% accuracy, respectively. | ['Lauren Phillips', 'Brian Hutchinson', 'Chris Careaga', 'Piper Wolters'] | 2020-12-02 | null | null | null | null | ['few-shot-audio-classification'] | ['audio'] | [ 2.22151667e-01 -1.87481537e-01 -1.90330356e-01 -4.48481381e-01
-1.05936456e+00 -2.71382481e-01 1.46556929e-01 -2.84438636e-02
-4.32178468e-01 6.64550781e-01 3.50961238e-01 1.27642276e-02
7.81936273e-02 -6.04609430e-01 -3.48394394e-01 -3.59958202e-01
-1.50498658e-01 3.26643854e-01 3.99007387e-02 -1.53343737e-01
-3.97756435e-02 1.79141417e-01 -1.67313135e+00 5.31863332e-01
5.39035797e-01 1.22285318e+00 -1.66929603e-01 8.31800580e-01
1.21926032e-02 8.24019134e-01 -7.99579620e-01 -2.68802255e-01
-9.95968506e-02 -2.66181201e-01 -8.16036463e-01 4.39473614e-02
4.22011465e-01 -3.26643050e-01 -4.03303087e-01 6.64525151e-01
8.79877985e-01 4.38586146e-01 5.84996700e-01 -1.20514965e+00
-5.50516665e-01 7.29866087e-01 -3.90813291e-01 6.18717194e-01
7.44513690e-01 7.66804144e-02 9.89321172e-01 -8.19001079e-01
2.93247133e-01 1.12205839e+00 7.07300961e-01 7.97189653e-01
-9.29266214e-01 -1.02174115e+00 -1.92251410e-02 4.48697060e-01
-1.41028321e+00 -9.79071140e-01 7.04000115e-01 -5.43596268e-01
1.18032217e+00 1.53947994e-02 5.67351162e-01 1.49682415e+00
-4.08477187e-02 6.42803729e-01 4.40127403e-01 -4.18850213e-01
3.96089613e-01 2.11870134e-01 2.65372962e-01 4.14191276e-01
-2.22593248e-01 -1.12172559e-01 -9.31064069e-01 -1.15586765e-01
2.41043225e-01 2.07773715e-01 -3.47762376e-01 1.85356483e-01
-9.92942214e-01 7.98045874e-01 4.66760434e-02 5.27069390e-01
-2.06409767e-01 8.91014747e-03 7.13580012e-01 4.40345228e-01
5.47658563e-01 3.22342724e-01 -3.30061018e-01 -8.03191662e-01
-9.78321850e-01 6.56369934e-03 9.62569237e-01 7.47232497e-01
4.05369520e-01 5.71233153e-01 -5.96398190e-02 1.30284441e+00
-4.80238758e-02 1.76359996e-01 1.05322325e+00 -8.80335689e-01
3.27383548e-01 3.72490078e-01 -1.65489957e-01 -6.27185225e-01
-2.32131928e-01 -4.88881826e-01 -8.34322214e-01 -2.35002533e-01
2.26454869e-01 -4.45511222e-01 -8.52350354e-01 1.73340094e+00
-5.45156002e-02 4.30084705e-01 6.02953583e-02 6.88754857e-01
8.67802560e-01 9.80731666e-01 -1.11716345e-01 -3.05816174e-01
1.19637704e+00 -7.79948115e-01 -8.19086552e-01 -2.81973660e-01
6.62916183e-01 -4.70103174e-01 1.10099137e+00 5.76580465e-01
-1.00432098e+00 -6.44875348e-01 -1.03220487e+00 2.37825379e-01
-4.18891430e-01 -2.98264802e-01 4.95418578e-01 8.64864945e-01
-8.31831276e-01 7.00834513e-01 -6.47387922e-01 -5.27933717e-01
6.80752277e-01 2.71990627e-01 -3.34106833e-01 -4.98071173e-03
-1.40302598e+00 6.48782074e-01 7.15891048e-02 -1.81144953e-01
-1.09597588e+00 -7.79605269e-01 -7.97206342e-01 3.97876859e-01
3.77335817e-01 -1.36303991e-01 1.56620610e+00 -9.47013140e-01
-1.49154961e+00 5.70173383e-01 -1.48862317e-01 -8.12546730e-01
1.27269819e-01 -2.82883018e-01 -8.84934843e-01 5.25353756e-03
-3.47907655e-02 4.89078015e-01 8.65959227e-01 -3.73655558e-01
-6.15767300e-01 -2.97744095e-01 -2.07142860e-01 -8.53235181e-03
-8.64781797e-01 3.27179104e-01 3.60244922e-02 -5.31444192e-01
-4.27010477e-01 -6.76782966e-01 1.43878430e-01 -3.23468506e-01
-1.30919933e-01 -3.76232982e-01 9.39729691e-01 -8.09724808e-01
1.37969804e+00 -2.35010552e+00 -1.63997728e-02 -1.80921644e-01
7.28171319e-02 3.44999731e-01 -2.29247976e-02 3.42550248e-01
-2.14962572e-01 1.09500244e-01 7.88583793e-03 -1.42619133e-01
2.92517859e-02 -4.08843979e-02 -3.35026354e-01 2.13213161e-01
6.93045706e-02 5.45604050e-01 -6.66861653e-01 -1.90009013e-01
3.79657894e-02 4.48415160e-01 -5.79436541e-01 1.83019444e-01
1.19723529e-01 -3.16345273e-03 -1.28534893e-02 7.99480021e-01
-7.05169886e-02 -9.22007784e-02 -5.67531064e-02 2.41970509e-01
3.57455425e-02 3.49327415e-01 -1.11241782e+00 1.81384897e+00
-6.40843749e-01 9.57344830e-01 -2.37877905e-01 -1.19684243e+00
9.00385916e-01 8.52023423e-01 5.09111762e-01 -4.55379844e-01
1.89189762e-01 6.11750074e-02 3.20006818e-01 -4.65148181e-01
2.33610034e-01 -3.02421540e-01 -3.74256223e-01 6.69667900e-01
6.45118475e-01 3.32248449e-01 7.69407004e-02 2.11588532e-01
1.31157124e+00 -5.92607617e-01 2.93157607e-01 1.47386074e-01
1.60759151e-01 -4.04677123e-01 8.09572637e-01 7.07945168e-01
-5.01814485e-01 4.91120994e-01 3.28802109e-01 -6.22372627e-01
-8.52989018e-01 -8.56081903e-01 -1.10158354e-01 1.59067714e+00
-5.40914237e-01 -4.20450091e-01 -7.38717914e-01 -2.19144732e-01
-1.19140200e-01 7.65719950e-01 -4.58173752e-01 -4.48670834e-01
-2.72232503e-01 -4.45693225e-01 7.84434736e-01 6.47018671e-01
4.16249245e-01 -1.02574766e+00 -6.38714373e-01 4.56383079e-01
-1.26544133e-01 -1.12574279e+00 -4.27070588e-01 2.44631365e-01
-5.57987332e-01 -8.25045466e-01 -7.32497275e-01 -8.01369369e-01
-1.42557800e-01 1.72235847e-01 8.75315607e-01 -3.57635111e-01
-5.06180346e-01 5.04843950e-01 -4.47799265e-01 -7.69482195e-01
-2.63313293e-01 2.66874373e-01 6.52424753e-01 2.28933796e-01
7.05051422e-01 -7.00528085e-01 -1.90491393e-01 3.10797125e-01
-3.98273587e-01 -3.35659981e-01 2.68774331e-01 8.91097784e-01
7.06509277e-02 8.30995217e-02 1.07718301e+00 -5.75142205e-01
7.65371978e-01 -4.44257736e-01 1.88328996e-02 2.52810389e-01
-2.17232451e-01 -3.56288612e-01 4.63818967e-01 -7.21436858e-01
-9.19932127e-01 -4.38326858e-02 -2.30929241e-01 -7.08724737e-01
-3.67202818e-01 3.75172377e-01 -8.53987187e-02 2.55405545e-01
8.67978513e-01 1.33878142e-01 -1.10081375e-01 -5.77060878e-01
1.24936193e-01 1.34092259e+00 5.07725775e-01 -1.82369798e-01
3.25385928e-01 1.07415840e-01 -7.36262321e-01 -1.29065990e+00
-9.49586749e-01 -6.26081347e-01 -5.47063231e-01 -2.99598753e-01
8.59448552e-01 -1.00717866e+00 -5.01463890e-01 4.53886449e-01
-1.04641795e+00 -5.52396029e-02 -4.26579356e-01 5.35257697e-01
-4.34106350e-01 -1.64683178e-01 -5.48538625e-01 -1.05957639e+00
-4.24621522e-01 -9.26283598e-01 7.20480561e-01 2.44090259e-01
-7.57553518e-01 -7.57618427e-01 1.95958167e-01 3.93241137e-01
4.09181178e-01 -3.23089436e-02 6.58178151e-01 -1.11660707e+00
1.63163655e-02 -5.26493371e-01 2.32197523e-01 4.67369705e-01
2.03272492e-01 -1.45966828e-01 -1.56886041e+00 -3.51850986e-01
-4.95126881e-02 -7.09593058e-01 8.81453693e-01 4.47148174e-01
1.41790617e+00 -3.01180482e-01 -2.28186548e-01 3.52356523e-01
7.81247914e-01 5.85111260e-01 3.61721158e-01 -1.52938068e-01
4.41835850e-01 4.77326185e-01 2.75272220e-01 6.92970514e-01
1.01006918e-01 7.91040480e-01 1.44870663e-02 4.19181019e-01
4.95993122e-02 -1.83595911e-01 5.13494790e-01 1.02688336e+00
-1.48789838e-01 -1.24003505e-03 -1.17694962e+00 5.60756922e-01
-1.57339644e+00 -1.45725405e+00 6.13098145e-01 2.17409730e+00
7.07598269e-01 1.66122884e-01 5.53631008e-01 5.63386977e-01
7.38224447e-01 1.69988081e-01 -6.09234512e-01 -5.75560093e-01
2.87351340e-01 3.96382600e-01 -2.07920894e-01 2.76347268e-02
-1.10510850e+00 7.34970093e-01 6.45654869e+00 7.26553202e-01
-1.20580113e+00 3.13971728e-01 6.90286815e-01 -7.68810749e-01
3.42550784e-01 -4.76731986e-01 -9.16722238e-01 6.23614550e-01
1.74053729e+00 -4.63608593e-01 5.08519053e-01 1.06516981e+00
1.62457153e-01 2.72803783e-01 -1.33813906e+00 1.54900968e+00
2.84521192e-01 -1.42826045e+00 -2.28340343e-01 -1.00529408e-02
5.52098572e-01 -5.09017557e-02 2.28627130e-01 8.01920950e-01
1.86959766e-02 -1.14475799e+00 5.50312519e-01 3.95323157e-01
9.19142842e-01 -1.00985146e+00 6.91396356e-01 3.58976007e-01
-1.12461519e+00 -5.69841921e-01 -3.15185457e-01 -3.34221959e-01
1.90947875e-01 3.66628855e-01 -9.27599728e-01 -1.42761357e-02
1.01909745e+00 7.74170101e-01 -2.59188235e-01 9.70810175e-01
2.77020305e-01 8.11710179e-01 -1.35273755e-01 -2.91196078e-01
7.76438266e-02 3.02391797e-01 4.41720128e-01 1.13344550e+00
4.33010787e-01 7.35348240e-02 -9.10193026e-02 4.63724583e-01
-4.38553065e-01 4.98105697e-02 -8.71725202e-01 -4.21854496e-01
7.95193076e-01 1.02848363e+00 -1.77251115e-01 -3.49929005e-01
-5.71013391e-01 6.75448775e-01 3.09059739e-01 2.64231414e-01
-6.91376925e-01 -8.73019636e-01 8.73760223e-01 1.92791507e-01
1.33087561e-01 1.04228660e-01 4.31469642e-02 -1.12933445e+00
-2.48584077e-01 -1.13675487e+00 4.51959044e-01 -5.33481956e-01
-1.19961572e+00 5.49444675e-01 -1.99144855e-01 -1.29822826e+00
-7.91695237e-01 -4.39642310e-01 -9.53407347e-01 6.07650220e-01
-8.96680057e-01 -7.45587945e-01 -1.18506052e-01 5.24331510e-01
9.65719879e-01 -7.64206409e-01 1.10040867e+00 5.09634078e-01
-7.80343413e-01 7.54731774e-01 9.37326811e-04 4.21067268e-01
8.11544240e-01 -8.17703187e-01 2.94763356e-01 4.31236714e-01
5.73685884e-01 5.49678862e-01 4.68958169e-01 -7.37067908e-02
-1.25125241e+00 -1.02097869e+00 8.05757046e-01 -4.25207138e-01
6.75345480e-01 -5.74843585e-01 -1.01189756e+00 7.06730843e-01
1.25260517e-01 -3.40301357e-02 1.31262016e+00 6.54112875e-01
-3.01350147e-01 -2.87291199e-01 -9.91778374e-01 2.91319638e-01
9.44545269e-01 -9.25388157e-01 -7.64002144e-01 1.62736565e-01
7.55563140e-01 -1.89750716e-02 -8.21178138e-01 8.38293731e-02
8.82712543e-01 -8.63221943e-01 8.09235275e-01 -1.03072095e+00
3.23204786e-01 3.22873205e-01 -4.95548129e-01 -1.54528451e+00
-2.29901940e-01 -6.25436187e-01 -4.98602390e-01 1.33008063e+00
4.44502890e-01 -3.89066726e-01 7.31210649e-01 5.88090301e-01
-2.51029909e-01 -7.67187595e-01 -1.19063735e+00 -8.94655824e-01
-3.79628353e-02 -8.81640434e-01 5.56392133e-01 1.09725165e+00
3.60459387e-01 7.11430013e-01 -6.24296188e-01 -2.46608540e-01
4.44080025e-01 -1.57854259e-01 7.05215693e-01 -1.42407787e+00
-4.72314268e-01 -2.53907263e-01 -7.87868500e-01 -4.95766848e-01
2.95329750e-01 -6.42748415e-01 -1.68065965e-01 -1.17921221e+00
1.92053720e-01 9.46643725e-02 -4.86702770e-01 6.31656885e-01
3.34556788e-01 2.11170077e-01 -3.56397522e-03 3.36354785e-02
-6.24582767e-01 5.69780886e-01 5.78689277e-01 -4.54429358e-01
-4.19152796e-01 2.05713838e-01 -8.28600287e-01 9.04780328e-01
8.39948535e-01 -2.36362025e-01 -4.17355955e-01 -1.87543675e-01
-8.75429884e-02 5.38799949e-02 1.37474149e-01 -1.47866535e+00
1.77855939e-01 8.88494179e-02 6.22662067e-01 -3.03070217e-01
7.41772115e-01 -4.46090013e-01 -1.04178995e-01 3.48562747e-01
-5.61081767e-01 -3.65447849e-01 1.59017593e-01 5.75862527e-01
-3.87683749e-01 -2.27303803e-01 8.47751856e-01 -8.49586651e-02
-9.03489769e-01 3.63756120e-01 -5.41842818e-01 1.87550396e-01
9.98932838e-01 -2.68933237e-01 -2.11479127e-01 -7.98823118e-01
-1.01043534e+00 3.73911262e-02 -5.61593063e-02 6.94153488e-01
6.80062294e-01 -1.55095220e+00 -5.24929523e-01 2.91629672e-01
2.67508864e-01 -3.39082569e-01 4.52871084e-01 6.23203814e-01
6.98152557e-02 4.67226803e-01 -3.23595017e-01 -4.95015651e-01
-1.37719274e+00 3.18903714e-01 2.54005104e-01 1.67073831e-01
-5.29242098e-01 1.10360658e+00 -5.82343154e-02 -2.15860710e-01
6.59124017e-01 -3.13244969e-01 -2.87026882e-01 4.34791774e-01
1.09116983e+00 5.57105839e-01 4.97470051e-02 -4.41812247e-01
-3.75326574e-01 1.62046656e-01 -3.13691229e-01 -2.36960590e-01
1.54931545e+00 1.98482707e-01 5.39303780e-01 1.05685902e+00
1.28857327e+00 -5.40705621e-01 -9.59681809e-01 -2.48275563e-01
-8.83051232e-02 -3.74771416e-01 1.83488727e-01 -6.01435304e-01
-8.41905832e-01 1.48420382e+00 8.65090013e-01 4.25232649e-01
8.22365403e-01 6.79238886e-02 8.91565323e-01 5.53405762e-01
3.60688061e-01 -1.44034755e+00 4.55647916e-01 5.04389584e-01
6.98210478e-01 -1.32755554e+00 -3.01749110e-01 1.99739903e-01
-7.42714942e-01 9.36921418e-01 5.71337223e-01 2.60480881e-01
6.74929261e-01 1.74475178e-01 -1.58688873e-01 5.70992157e-02
-1.22094166e+00 6.01212904e-02 1.56415895e-01 6.87060714e-01
7.00315893e-01 3.39728594e-02 3.80219311e-01 1.04867399e+00
-2.18573555e-01 1.00095816e-01 4.51592565e-01 8.41010690e-01
-7.73296893e-01 -7.65909493e-01 -2.58067459e-01 8.14081132e-01
-5.07329643e-01 5.04699871e-02 -3.61749023e-01 2.81548500e-01
-9.93351191e-02 1.05059242e+00 4.87215042e-01 -7.49675691e-01
2.50769347e-01 6.61390603e-01 2.34968171e-01 -9.81924832e-01
-3.24888319e-01 -3.25005651e-02 2.84818739e-01 -2.92855650e-01
-1.50239378e-01 -7.56336987e-01 -8.99466038e-01 -4.00088727e-01
-2.38858685e-01 1.24355808e-01 5.57820380e-01 8.55094671e-01
3.77042741e-01 7.24717796e-01 7.22844720e-01 -9.10732269e-01
-7.31527925e-01 -1.33231676e+00 -8.68852735e-01 1.86859220e-01
1.81688666e-01 -7.38597870e-01 -3.40813339e-01 3.29289623e-02] | [14.544051170349121, 5.680128574371338] |
a5372334-9c6c-466e-bc34-aa906122bc87 | interpretable-neural-architecture-search-and | 2305.11917 | null | https://arxiv.org/abs/2305.11917v1 | https://arxiv.org/pdf/2305.11917v1.pdf | Interpretable neural architecture search and transfer learning for understanding sequence dependent enzymatic reactions | Finely-tuned enzymatic pathways control cellular processes, and their dysregulation can lead to disease. Creating predictive and interpretable models for these pathways is challenging because of the complexity of the pathways and of the cellular and genomic contexts. Here we introduce Elektrum, a deep learning framework which addresses these challenges with data-driven and biophysically interpretable models for determining the kinetics of biochemical systems. First, it uses in vitro kinetic assays to rapidly hypothesize an ensemble of high-quality Kinetically Interpretable Neural Networks (KINNs) that predict reaction rates. It then employs a novel transfer learning step, where the KINNs are inserted as intermediary layers into deeper convolutional neural networks, fine-tuning the predictions for reaction-dependent in vivo outcomes. Elektrum makes effective use of the limited, but clean in vitro data and the noisy, yet plentiful in vivo data that captures cellular context. We apply Elektrum to predict CRISPR-Cas9 off-target editing probabilities and demonstrate that Elektrum achieves state-of-the-art performance, regularizes neural network architectures, and maintains physical interpretability. | ['Olga Troyanskaya', 'Michael Shelley', 'Adam R. Lamson', 'Zijun Zhang'] | 2023-05-18 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 2.75307268e-01 -7.09337816e-02 -9.35122743e-02 -7.09196227e-03
-6.57174706e-01 -1.16895974e+00 4.38305825e-01 3.90783817e-01
-1.80414408e-01 1.10208189e+00 2.10316658e-01 -8.03904533e-01
-2.97027647e-01 -3.44909370e-01 -1.31064546e+00 -8.72202218e-01
-1.43681124e-01 5.09341776e-01 -7.58413076e-02 -1.29238158e-01
7.08433762e-02 7.81931698e-01 -8.53753924e-01 6.31408989e-01
8.00882816e-01 5.77306867e-01 -5.02788536e-02 1.19491887e+00
1.75299972e-01 4.31393236e-01 -2.74544686e-01 -4.91922684e-02
-3.85264456e-02 -6.57438934e-01 -4.90049392e-01 -8.57687950e-01
-3.20379063e-02 1.11641683e-01 -3.25802326e-01 2.98477322e-01
6.68509185e-01 -2.36496046e-01 8.59316587e-01 -8.34319770e-01
-1.19139159e+00 1.91215560e-01 1.12780087e-01 1.76926628e-01
1.35238603e-01 6.47726119e-01 8.87823582e-01 -7.50984490e-01
8.02189350e-01 1.04228783e+00 8.91290963e-01 6.49398088e-01
-1.81028676e+00 -3.87938678e-01 -1.49946287e-01 -1.97031931e-03
-1.07573140e+00 -4.37246323e-01 1.02527030e-01 -5.79611063e-01
1.35968471e+00 2.70514369e-01 7.97311127e-01 1.55949402e+00
7.93431163e-01 1.06717370e-01 6.88997030e-01 -1.84827417e-01
4.24526393e-01 -8.76970470e-01 -1.81112841e-01 9.40851808e-01
2.18015745e-01 1.14900537e-01 -5.60516477e-01 -4.31970179e-01
8.04857016e-01 9.58239064e-02 -2.76312351e-01 2.87729204e-02
-1.55086696e+00 1.97118536e-01 8.26534033e-02 -4.36563529e-02
-3.66625547e-01 4.46084321e-01 3.62676382e-01 5.99017479e-02
-3.89946885e-02 1.13728189e+00 -1.29289973e+00 -8.16422552e-02
-4.26394999e-01 1.76117808e-01 7.78580844e-01 3.57720762e-01
4.64214891e-01 -2.66944319e-01 -1.09072402e-01 4.96582866e-01
3.07788789e-01 6.92719743e-02 3.81569803e-01 -1.01974380e+00
-1.93455234e-01 6.74378276e-01 3.69179696e-01 -4.29045290e-01
-6.91800773e-01 -8.65606889e-02 -8.19823682e-01 8.48340988e-02
7.86653697e-01 -4.58804905e-01 -1.16585767e+00 1.94176757e+00
2.45811924e-01 3.24891567e-01 -2.80782171e-02 7.81811953e-01
4.26466703e-01 9.89875078e-01 3.85523021e-01 -1.74087733e-01
1.05362093e+00 -5.43093860e-01 -5.59008539e-01 9.49688628e-02
6.13815486e-01 -3.83973151e-01 1.09108341e+00 5.38683236e-01
-1.05517995e+00 7.16223717e-02 -1.14443195e+00 -6.87081099e-01
-8.93476605e-01 1.34563729e-01 8.13702941e-01 3.09156507e-01
-9.51068282e-01 8.33421469e-01 -1.21072233e+00 -5.01609683e-01
4.90690798e-01 6.09450161e-01 -4.04041499e-01 3.39033484e-01
-1.06635058e+00 8.31194460e-01 5.01346409e-01 2.27922082e-01
-1.35981643e+00 -1.33285809e+00 -3.99210632e-01 6.03956580e-01
-1.14101529e-01 -9.71871793e-01 9.93086398e-01 -2.65733719e-01
-1.73417842e+00 6.03512764e-01 -1.46298215e-01 -3.51976454e-02
2.39952475e-01 1.60070881e-01 -1.58982709e-01 -2.40421265e-01
-4.55558300e-01 6.79682910e-01 -1.05357654e-01 -7.08284557e-01
2.81885993e-02 -3.50598693e-01 -7.26558045e-02 -3.05688202e-01
1.79970190e-01 -1.13128208e-01 -1.90309435e-01 -4.69587922e-01
-1.51180163e-01 -8.93954217e-01 -2.36015186e-01 3.58349621e-01
-6.92634225e-01 4.15928587e-02 5.13882101e-01 -5.31555235e-01
8.17366838e-01 -1.68873429e+00 4.78765756e-01 1.43339857e-02
4.90839094e-01 3.79306316e-01 -3.74019980e-01 8.29944551e-01
-4.59974200e-01 5.10176003e-01 -4.11592312e-02 2.83497363e-01
-2.45262105e-02 1.77802406e-02 -2.47677881e-02 3.71522605e-01
6.47642791e-01 1.29405582e+00 -8.17479968e-01 1.40179381e-01
7.24553913e-02 7.76480436e-01 -5.41018426e-01 2.43946776e-01
-8.89302492e-01 8.74146104e-01 -5.53096950e-01 7.30537832e-01
1.71686828e-01 -5.86045682e-01 7.92125344e-01 -2.40048215e-01
-1.64163187e-01 2.33240858e-01 -5.10240376e-01 1.78319061e+00
-2.75920302e-01 5.72254539e-01 -2.25096703e-01 -9.92352307e-01
5.34528911e-01 3.06773365e-01 3.55489075e-01 -3.47614080e-01
8.88653249e-02 7.65355006e-02 8.96717012e-02 -4.19535160e-01
-2.85696179e-01 4.89474908e-02 -6.68798164e-02 3.27987701e-01
2.23250791e-01 1.77309871e-01 1.09320745e-01 6.85819834e-02
1.35251856e+00 5.28833032e-01 4.77067769e-01 -3.17410439e-01
2.75623292e-01 6.79972675e-03 8.37352872e-01 6.07681274e-01
-2.65083015e-01 4.66985464e-01 1.15366244e+00 -7.59909272e-01
-1.62707186e+00 -1.16557872e+00 1.26160935e-01 1.16880798e+00
-2.98054367e-01 -3.25576812e-01 -9.39845920e-01 -2.63615489e-01
4.05950360e-02 2.37839937e-01 -1.02495623e+00 -4.25563604e-01
-2.73178697e-01 -1.02308655e+00 1.26115143e+00 5.50822318e-01
-3.07914257e-01 -7.12393343e-01 2.36985013e-02 4.09694254e-01
1.25055566e-01 -6.83617890e-01 -4.16747957e-01 8.05739224e-01
-4.67346340e-01 -1.61755621e+00 -1.25093073e-01 -4.56672519e-01
5.05703449e-01 -2.07891062e-01 7.58045375e-01 4.57150936e-02
-7.19834089e-01 -2.90664852e-01 2.76212364e-01 -8.23626578e-01
-7.52214491e-01 -2.49333549e-02 2.19104871e-01 -4.68983501e-01
2.54766196e-01 -7.28425741e-01 -6.18906319e-01 1.60950258e-01
-1.07066977e+00 1.90409675e-01 5.95202208e-01 1.23492241e+00
1.06185997e+00 -5.78900695e-01 9.47254539e-01 -8.60342562e-01
6.50328696e-01 -5.12463152e-01 -7.99461544e-01 6.77402675e-01
-5.59144914e-01 4.89689767e-01 1.20298970e+00 -3.73202860e-01
-1.10079432e+00 2.36203805e-01 -1.20688744e-01 4.68658060e-02
-3.07894766e-01 5.67401588e-01 -3.73994440e-01 -6.55021193e-03
9.10170376e-01 2.22872511e-01 -1.50294593e-02 -2.70811439e-01
5.23438096e-01 2.29152903e-01 4.97582942e-01 -8.62710416e-01
1.30137518e-01 4.02381897e-01 4.89916921e-01 -5.41564047e-01
-7.05974758e-01 -3.71525474e-02 -6.49033844e-01 1.11777723e-01
7.74942219e-01 -7.00666964e-01 -1.47976744e+00 6.88120663e-01
-1.12812078e+00 -7.96255767e-01 2.28873134e-01 2.12655455e-01
-8.77403259e-01 1.23589255e-01 -1.14521468e+00 -4.28586453e-01
-4.43861693e-01 -1.18451321e+00 1.12893116e+00 1.32562026e-01
-2.56221920e-01 -9.05908048e-01 3.98445398e-01 1.01989679e-01
4.66260314e-02 6.35171235e-01 1.55775118e+00 -6.06388032e-01
-7.03403056e-01 -6.73553497e-02 -2.72468656e-01 -1.82385534e-01
1.48786038e-01 5.22059321e-01 -1.00288069e+00 2.38780137e-02
-9.72478569e-01 -5.49538672e-01 7.81884074e-01 5.38250923e-01
1.59100771e+00 -3.17904532e-01 -5.47693551e-01 7.96867371e-01
1.23591089e+00 2.07982942e-01 6.35095775e-01 1.01029895e-01
6.50252461e-01 8.47830325e-02 6.21825308e-02 2.07553223e-01
8.30394849e-02 3.26622367e-01 4.10406947e-01 -1.45893902e-01
1.82013944e-01 -3.88781309e-01 2.74356484e-01 3.29283446e-01
-2.65692741e-01 -7.08570361e-01 -8.55765343e-01 1.58428937e-01
-1.93595946e+00 -8.80643070e-01 -1.01507135e-01 1.87984860e+00
1.16647756e+00 -1.88397378e-01 -2.62859426e-02 -3.39230925e-01
4.41060215e-01 -3.69092554e-01 -1.15956008e+00 -7.11271346e-01
-3.66203040e-01 2.39637136e-01 3.87211710e-01 3.99907142e-01
-7.62834907e-01 7.72484720e-01 7.31515503e+00 5.35720110e-01
-1.16915381e+00 -1.30129561e-01 9.82452273e-01 -2.74856210e-01
-2.34920636e-01 -1.34241194e-01 -3.47104639e-01 4.13761348e-01
1.45772314e+00 -1.11377895e-01 7.44473517e-01 1.98930323e-01
9.45828795e-01 3.13915879e-01 -1.62827051e+00 4.02362347e-01
-7.64118075e-01 -2.35677814e+00 -1.24269044e-02 1.66451707e-01
6.37049198e-01 3.50588262e-01 -7.19094127e-02 -8.77822004e-03
7.87956774e-01 -1.52016628e+00 4.01583374e-01 9.48136628e-01
9.16476607e-01 -5.62624514e-01 3.78917277e-01 4.41881269e-01
-5.75713158e-01 -1.05848908e-01 -4.55556244e-01 -1.39573496e-02
-2.62919098e-01 7.63301790e-01 -1.00298965e+00 1.33228809e-01
4.58712220e-01 5.11364102e-01 -1.58624668e-02 6.81983769e-01
-6.86214566e-02 8.04271698e-01 -3.60199958e-01 -1.93060741e-01
3.09300460e-02 -8.93608332e-02 -3.01708914e-02 1.42045593e+00
2.60423303e-01 3.79110962e-01 -1.59758344e-01 1.32722414e+00
-5.19011378e-01 -2.44732559e-01 -3.77349794e-01 -7.05258310e-01
5.46321750e-01 1.23972845e+00 -4.47621018e-01 -1.42859474e-01
1.39754370e-01 8.01709235e-01 7.15040267e-01 5.42490125e-01
-9.17153597e-01 -1.88602239e-01 1.29889500e+00 -2.03716919e-01
-1.45114660e-01 -3.95348579e-01 -2.54238069e-01 -9.93437052e-01
-2.55535781e-01 -7.25794554e-01 3.68880779e-01 -8.62196207e-01
-1.02073455e+00 -3.77092332e-01 -7.44953513e-01 -3.79958183e-01
7.92496279e-02 -1.15227425e+00 -6.46865547e-01 1.04420054e+00
-1.13330054e+00 -1.09786153e+00 1.09359771e-02 -3.65249813e-02
4.29973811e-01 4.43104833e-01 1.10202730e+00 1.59138694e-01
-1.10913372e+00 4.23309565e-01 7.91256011e-01 -1.36544704e-01
8.11992645e-01 -1.47374046e+00 7.25970089e-01 3.69064182e-01
-4.17411447e-01 1.15973151e+00 5.84571481e-01 -6.17169201e-01
-1.77321339e+00 -1.42917156e+00 7.17022240e-01 -9.38440084e-01
8.30833852e-01 -6.58669055e-01 -8.89978349e-01 8.51130545e-01
-2.33610451e-01 1.54866502e-01 1.15208721e+00 7.82983527e-02
-6.08693719e-01 1.64240092e-01 -1.09808791e+00 7.59388566e-01
1.11010623e+00 -6.06713891e-01 -3.05064976e-01 6.67733788e-01
7.81676710e-01 -5.09195864e-01 -1.27912247e+00 -1.13453224e-01
1.06026745e+00 -6.36396706e-01 1.13202226e+00 -1.62113297e+00
6.85724974e-01 -3.68343771e-01 2.88057476e-01 -1.62869203e+00
-5.89069009e-01 -7.81571448e-01 -2.43428096e-01 6.60976768e-01
8.86864424e-01 -5.96795678e-01 7.85543978e-01 5.40490627e-01
-3.62980932e-01 -1.06639969e+00 -5.97593606e-01 -4.37667072e-01
2.65719593e-01 6.29754290e-02 6.62125528e-01 1.18185949e+00
5.38675845e-01 1.65974870e-01 -4.08670157e-02 2.83127666e-01
2.77000368e-01 -2.74824500e-01 4.99131590e-01 -1.05430460e+00
-3.45510751e-01 -2.04645291e-01 -2.00053781e-01 -6.33789539e-01
7.64294993e-03 -6.83322072e-01 1.70490965e-01 -1.43719256e+00
2.20734358e-01 -6.64297789e-02 -8.35738599e-01 7.19199717e-01
-1.64385706e-01 1.35928970e-02 -3.67746443e-01 -3.64557281e-02
-5.69157422e-01 3.41227293e-01 1.07370675e+00 -3.02806854e-01
-6.83637038e-02 -6.39711320e-01 -8.91879857e-01 4.75044072e-01
8.47390056e-01 -2.75010377e-01 -2.52126396e-01 -3.69099230e-01
4.70317006e-01 9.31236073e-02 4.41019148e-01 -7.19807148e-01
6.18364774e-02 -6.32474184e-01 1.10632181e+00 -8.98273960e-02
1.50256976e-01 -2.96784282e-01 5.58517039e-01 4.31903124e-01
-9.49265540e-01 -5.99358454e-02 5.00927150e-01 9.75259423e-01
5.51418900e-01 6.66796625e-01 6.71087205e-01 -2.49413297e-01
-1.35209724e-01 5.47865748e-01 -8.83004248e-01 -1.83604956e-01
7.57687628e-01 -1.57974750e-01 -6.94010675e-01 2.88172532e-03
-9.06092405e-01 1.38893381e-01 6.53159678e-01 2.26330906e-01
1.85251370e-01 -7.24761367e-01 -4.66369003e-01 3.26044373e-02
2.42869869e-01 -1.12105474e-01 3.07983577e-01 5.64871728e-01
-9.96093631e-01 8.17133605e-01 -3.72397125e-01 -3.20046991e-01
-6.55059338e-01 5.54723263e-01 7.12988675e-01 -1.34317249e-01
5.55397309e-02 6.83117390e-01 3.59678447e-01 -8.75316739e-01
-4.08265263e-01 -7.41248369e-01 1.64865792e-01 -6.52961373e-01
3.35648060e-01 1.44919425e-01 1.22271135e-01 1.23086236e-01
-1.44705281e-01 5.84768914e-02 -2.93457191e-02 6.48554027e-01
1.48753667e+00 1.03155293e-01 -4.21057969e-01 2.67989069e-01
1.02602875e+00 -3.57949585e-01 -1.65704918e+00 3.96494716e-01
-1.38493761e-01 1.08197562e-01 -1.39776051e-01 -1.58747864e+00
-3.81811738e-01 7.83444345e-01 2.60329634e-01 -2.04932570e-01
8.53439808e-01 -2.98038483e-01 8.13004315e-01 1.07239795e+00
1.28870934e-01 -8.77719820e-01 -2.92221427e-01 4.99260008e-01
4.14016783e-01 -7.76053488e-01 -2.44085401e-01 -3.52351546e-01
2.33450785e-01 1.25739777e+00 5.89816034e-01 2.24587664e-01
2.18792841e-01 4.18381512e-01 5.22503108e-02 -8.04116353e-02
-1.28610671e+00 5.80366790e-01 -4.37065214e-02 7.00616896e-01
8.07605803e-01 7.89273251e-03 -3.10558379e-01 7.89884686e-01
3.51294845e-01 5.43666124e-01 6.89612865e-01 4.81925398e-01
-3.20864111e-01 -1.08346140e+00 -1.01474874e-01 4.94974107e-01
-6.22971356e-01 -1.58085421e-01 -8.19541991e-01 4.79752690e-01
7.51286000e-02 5.90572536e-01 -3.11967462e-01 -8.40282009e-04
2.31759399e-01 3.16375107e-01 5.05845487e-01 -3.26403946e-01
-3.69620234e-01 -1.71131417e-01 1.09659679e-01 -8.22563589e-01
1.68837428e-01 -4.23399180e-01 -1.72829771e+00 -4.31068003e-01
-6.06254637e-02 1.39920443e-01 7.41565883e-01 1.00492513e+00
1.32231307e+00 8.50863993e-01 1.69870764e-01 -8.20828915e-01
-2.12977186e-01 -5.85636914e-01 -2.59576082e-01 7.66434520e-02
4.68845159e-01 -4.54200566e-01 1.02329709e-01 5.44019163e-01] | [4.947540760040283, 5.625450134277344] |
13f3c5eb-ecea-4acc-ad64-3c6131678176 | sart-similarity-analogies-and-relatedness-for | 1904.00365 | null | http://arxiv.org/abs/1904.00365v1 | http://arxiv.org/pdf/1904.00365v1.pdf | SART - Similarity, Analogies, and Relatedness for Tatar Language: New Benchmark Datasets for Word Embeddings Evaluation | There is a huge imbalance between languages currently spoken and
corresponding resources to study them. Most of the attention naturally goes to
the "big" languages: those which have the largest presence in terms of media
and number of speakers. Other less represented languages sometimes do not even
have a good quality corpus to study them. In this paper, we tackle this
imbalance by presenting a new set of evaluation resources for Tatar, a language
of the Turkic language family which is mainly spoken in Tatarstan Republic,
Russia.
We present three datasets: Similarity and Relatedness datasets that consist
of human scored word pairs and can be used to evaluate semantic models; and
Analogies dataset that comprises analogy questions and allows to explore
semantic, syntactic, and morphological aspects of language modeling. All three
datasets build upon existing datasets for the English language and follow the
same structure. However, they are not mere translations. They take into account
specifics of the Tatar language and expand beyond the original datasets. We
evaluate state-of-the-art word embedding models for two languages using our
proposed datasets for Tatar and the original datasets for English and report
our findings on performance comparison. | ['Adín Ramírez Rivera', 'Albina Khusainova', 'Adil Khan'] | 2019-03-31 | null | null | null | null | ['embeddings-evaluation'] | ['natural-language-processing'] | [-4.97660756e-01 -2.31616378e-01 -3.07485640e-01 -2.86889523e-01
-4.69098955e-01 -5.81551075e-01 8.21788907e-01 3.11877638e-01
-9.93642926e-01 5.38453460e-01 6.07510865e-01 -2.46209636e-01
-1.13388292e-01 -8.63772273e-01 -2.26557657e-01 -2.99580485e-01
2.64671803e-01 8.89974833e-01 1.17625229e-01 -8.90246868e-01
1.72075897e-01 1.34873375e-01 -1.31129146e+00 2.31858805e-01
1.00711930e+00 4.14691985e-01 2.43393496e-01 8.65280703e-02
-6.35453999e-01 4.73582178e-01 -3.81031811e-01 -6.71914935e-01
-6.24995977e-02 -3.29650998e-01 -9.02603269e-01 -3.10751408e-01
3.59390199e-01 1.18762583e-01 -3.69750798e-01 9.06143546e-01
5.82912803e-01 -8.38620961e-03 8.30832720e-01 -1.09249532e+00
-1.17581987e+00 1.15599692e+00 -2.18024582e-01 2.79054582e-01
4.45053160e-01 -3.68362904e-01 1.32101238e+00 -1.05232811e+00
8.54005516e-01 1.51284039e+00 6.45847082e-01 6.62134051e-01
-9.73634958e-01 -6.28948748e-01 -2.31950334e-03 4.47979033e-01
-1.64689302e+00 -4.03887659e-01 6.94234431e-01 -4.46787417e-01
1.15203965e+00 1.21908702e-01 4.59480077e-01 1.32680058e+00
-1.72671005e-01 4.40281034e-01 1.08187747e+00 -6.83209002e-01
1.60018563e-01 7.64520824e-01 3.49752843e-01 5.11043966e-01
1.54054165e-01 -3.17401707e-01 -5.35866559e-01 1.39814252e-02
2.58325696e-01 -3.15219462e-01 -3.58432263e-01 -2.68738747e-01
-1.22248423e+00 1.13085198e+00 2.57911813e-02 8.61102045e-01
-8.50310251e-02 -2.32681438e-01 5.95635295e-01 3.97553742e-01
3.94858450e-01 3.84718806e-01 -6.34754598e-01 -8.16484354e-03
-6.70000970e-01 3.49816531e-01 9.72652376e-01 8.22365642e-01
5.14206409e-01 -7.86802694e-02 2.45744884e-01 1.29799056e+00
4.37760115e-01 5.85023940e-01 1.09080553e+00 -3.95790696e-01
5.24698615e-01 7.92616546e-01 -3.35852772e-01 -9.13241029e-01
-2.14247912e-01 -2.66470551e-01 -4.64785695e-01 -3.29810172e-01
3.88722986e-01 1.44036025e-01 -5.32343984e-01 1.75937808e+00
7.53431441e-03 -1.57863051e-01 5.65499544e-01 7.66939461e-01
1.27435410e+00 6.83826506e-01 9.47924405e-02 6.78043962e-02
1.74449635e+00 -9.78237629e-01 -7.28293836e-01 -1.75984189e-01
9.01177466e-01 -9.38667357e-01 1.59487283e+00 1.57532409e-01
-9.29601133e-01 -3.52797568e-01 -9.01547551e-01 -2.48399884e-01
-1.18491900e+00 1.27803981e-01 2.56178826e-01 8.33434105e-01
-9.81357992e-01 3.98699671e-01 -2.28555918e-01 -9.01733577e-01
-1.43160209e-01 -1.45242870e-01 -6.45458460e-01 -1.45247519e-01
-1.62566447e+00 1.43567157e+00 3.96403491e-01 -2.84044296e-01
-2.66787380e-01 -6.91644847e-01 -1.00175130e+00 -1.56549141e-01
2.28936970e-01 -3.09338808e-01 8.86985600e-01 -8.24764073e-01
-1.14846325e+00 1.24414146e+00 3.52974236e-02 -2.86365598e-01
2.61864036e-01 -7.08620623e-02 -7.27307856e-01 -2.43059143e-01
2.61060119e-01 5.02432466e-01 7.43732899e-02 -9.40118134e-01
-2.87791789e-01 -4.30738688e-01 4.02401648e-02 1.45517319e-01
-7.46984065e-01 4.71781820e-01 -3.59718591e-01 -8.64060760e-01
-3.02931339e-01 -8.38802695e-01 2.17564888e-02 -2.37160131e-01
3.06698978e-02 -4.37736899e-01 5.06215870e-01 -8.11903059e-01
1.52545190e+00 -2.03318405e+00 2.86162525e-01 -6.39913902e-02
-1.78026721e-01 2.13855296e-01 -4.30538654e-01 1.02977073e+00
-4.22979556e-02 4.28510308e-01 -1.67874545e-01 -2.79086709e-01
2.74933368e-01 5.78462660e-01 -4.16785210e-01 1.36040747e-01
-1.37087135e-02 9.25669551e-01 -8.03331077e-01 -7.17488945e-01
3.21288966e-02 4.42763746e-01 -4.08702195e-01 -1.79141745e-01
-1.42809395e-02 8.40390921e-02 -2.72204965e-01 5.79908729e-01
4.62360620e-01 2.24261701e-01 3.20693582e-01 -2.67764062e-01
-2.93816954e-01 5.48015952e-01 -9.17552233e-01 1.69853568e+00
-7.04898775e-01 4.87109810e-01 -3.53464544e-01 -8.75739336e-01
9.04514134e-01 2.01215863e-01 1.36971310e-01 -7.97517061e-01
1.83931813e-01 5.09304047e-01 2.76703447e-01 -6.40739083e-01
7.41263747e-01 -1.55802459e-01 -3.77822578e-01 2.12665707e-01
7.48036429e-02 -2.89532363e-01 4.00728315e-01 1.02180364e-02
8.67799520e-01 -2.38837544e-02 5.18133879e-01 -6.16320431e-01
7.68109262e-01 6.50861161e-03 4.70115453e-01 1.72506660e-01
-9.91279110e-02 3.51972729e-01 3.14140946e-01 -3.00715834e-01
-1.09977770e+00 -1.12094378e+00 -4.05502409e-01 1.19434631e+00
3.96897905e-02 -7.28477299e-01 -5.95844448e-01 -3.95769387e-01
-8.77492651e-02 9.61554348e-01 -6.68616176e-01 7.73203447e-02
-4.84783202e-01 -5.75339317e-01 8.46374929e-01 3.40318948e-01
5.23976505e-01 -1.14451635e+00 -1.83752477e-01 2.21620649e-02
-2.05849126e-01 -1.29122210e+00 -3.31627101e-01 -1.82905182e-01
-5.36139011e-01 -9.07402635e-01 -6.80803895e-01 -1.00242162e+00
3.19884978e-02 -6.07596785e-02 1.40932190e+00 -1.25494257e-01
-1.47320092e-01 4.29737657e-01 -6.09977484e-01 -7.00529039e-01
-5.11810899e-01 1.00679733e-01 1.73122406e-01 -3.08084518e-01
9.60373878e-01 -3.92060250e-01 -6.17734157e-02 2.04770133e-01
-8.91024232e-01 -2.86836714e-01 2.90703297e-01 6.29937887e-01
2.43407086e-01 -3.19574565e-01 6.60648406e-01 -8.49960923e-01
7.42233813e-01 -7.55861104e-01 -2.56842762e-01 5.13247967e-01
-3.28702778e-01 2.17098057e-01 4.96533394e-01 -3.76374513e-01
-6.43181026e-01 -4.80041385e-01 -3.03026587e-01 -8.71129110e-02
-1.83474794e-02 7.02395499e-01 -1.65786043e-01 1.55402064e-01
4.02772516e-01 3.41626406e-01 -1.66398287e-01 -7.73413360e-01
5.83751380e-01 9.25557554e-01 2.96825439e-01 -6.39792144e-01
4.77004945e-01 1.90676525e-01 -4.18209374e-01 -1.42336190e+00
-5.00791252e-01 -5.03951311e-01 -5.91993511e-01 1.25999704e-01
9.29953992e-01 -8.07315588e-01 -2.76437581e-01 2.71079242e-01
-1.22750461e+00 1.49729520e-01 -1.90680996e-01 6.80745661e-01
-8.83080214e-02 4.91359472e-01 -4.94903624e-01 -5.99239051e-01
-3.41409355e-01 -8.81858766e-01 9.13463712e-01 -1.76770687e-01
-4.89487618e-01 -1.45665514e+00 4.27558899e-01 2.26123378e-01
5.93641996e-01 -7.03048110e-02 1.39886451e+00 -9.86406565e-01
2.10402206e-01 2.18477249e-01 1.82994753e-02 4.94711578e-01
1.47262216e-01 -1.35169134e-01 -8.80937099e-01 -6.40057698e-02
-1.65612921e-01 -5.24496019e-01 8.15764070e-01 -6.75472170e-02
5.93222022e-01 -1.33568704e-01 -4.85417955e-02 8.33045989e-02
1.43529487e+00 8.21331963e-02 4.77187127e-01 4.41708684e-01
4.24178511e-01 1.05558360e+00 3.09208810e-01 2.28225425e-01
7.51823664e-01 9.89210784e-01 -2.55388096e-02 1.98556423e-01
-1.29281446e-01 -3.24706048e-01 5.93981266e-01 1.79947567e+00
1.06608629e-01 -3.03467870e-01 -1.18983996e+00 9.24697340e-01
-1.65451038e+00 -6.15666151e-01 -2.42735118e-01 2.09089518e+00
9.91236687e-01 -2.73651391e-01 3.21719274e-02 2.15567257e-02
4.94331241e-01 2.22902596e-02 1.46679103e-01 -6.52516246e-01
-4.42599624e-01 4.59299624e-01 1.46596566e-01 5.69652796e-01
-8.28190207e-01 1.24687207e+00 6.07237339e+00 1.06416643e+00
-9.04937148e-01 3.65655988e-01 4.63163592e-02 1.40717685e-01
-7.09914505e-01 -5.65488413e-02 -7.60181606e-01 4.13742334e-01
1.27423644e+00 -2.43495271e-01 2.47970641e-01 6.52821302e-01
5.79332337e-02 2.25209698e-01 -1.10514688e+00 1.05781746e+00
4.70032781e-01 -1.09832919e+00 3.78031820e-01 -6.12452552e-02
4.13212150e-01 2.89951235e-01 -1.33808851e-01 4.77645338e-01
3.53761166e-02 -1.20032668e+00 6.85808480e-01 2.72256583e-01
6.18047714e-01 -6.98766351e-01 9.04822052e-01 3.22196096e-01
-1.01748395e+00 3.02851588e-01 -6.79312885e-01 1.35781085e-02
8.33676830e-02 1.45164788e-01 -3.94574314e-01 4.72319901e-01
8.35760474e-01 6.68690979e-01 -7.43436575e-01 5.57867289e-01
-2.37016231e-01 6.16591394e-01 -1.76753640e-01 -5.48361480e-01
4.55280483e-01 -5.02453864e-01 5.30471623e-01 1.38694084e+00
4.51469839e-01 -2.03004882e-01 1.63696468e-01 7.87015080e-01
-2.60848813e-02 9.07310486e-01 -9.70608234e-01 -2.60245025e-01
6.71544731e-01 1.06176424e+00 -4.22666639e-01 -2.67989874e-01
-8.48458171e-01 8.78302813e-01 4.18207258e-01 2.39008144e-02
-7.13491261e-01 -5.31311631e-01 8.39218557e-01 1.36851206e-01
2.76746824e-02 -1.90548256e-01 8.62120092e-02 -1.28935826e+00
3.34738358e-03 -8.99825096e-01 3.32692176e-01 -5.79605222e-01
-1.84406996e+00 8.92789602e-01 1.53336301e-01 -9.48223829e-01
-2.44054779e-01 -1.00564790e+00 -2.28087321e-01 9.28020954e-01
-1.66786861e+00 -1.43142021e+00 -5.39097078e-02 5.94810128e-01
7.62274384e-01 -6.05145276e-01 1.23160374e+00 5.56088507e-01
-3.63209605e-01 6.57983243e-01 1.99572682e-01 2.88426310e-01
7.13611782e-01 -1.17387068e+00 3.75758082e-01 3.91744524e-01
5.09618163e-01 7.10339487e-01 5.31841397e-01 -3.89583111e-01
-1.26508522e+00 -7.65214264e-01 1.64354253e+00 -7.37218678e-01
1.18769956e+00 -5.56288898e-01 -1.14294755e+00 4.62045282e-01
6.81944609e-01 -2.88738728e-01 9.55986857e-01 2.05060706e-01
-6.53412819e-01 1.05978593e-01 -1.01483929e+00 7.54991353e-01
1.05396843e+00 -6.09563947e-01 -1.19488502e+00 2.65854716e-01
6.56197727e-01 1.74033284e-01 -7.14420617e-01 4.19664010e-02
5.02699077e-01 -7.26612449e-01 8.41320574e-01 -8.32573414e-01
5.29408455e-01 -5.75887524e-02 -5.66018879e-01 -1.37812924e+00
-8.41483921e-02 -1.61878139e-01 5.09226978e-01 1.65859914e+00
5.36290348e-01 -7.84426153e-01 3.60444576e-01 1.17008992e-01
-3.37474644e-02 -5.81819713e-01 -1.13104093e+00 -1.02825379e+00
6.72020555e-01 -5.35732627e-01 5.95324099e-01 1.30868554e+00
6.95424676e-02 7.36197174e-01 -3.79799791e-02 -2.13488445e-01
2.70322621e-01 -7.10635558e-02 4.55753356e-01 -1.20068014e+00
-2.05776542e-01 -5.70998728e-01 -6.13564134e-01 -2.97418267e-01
5.88771105e-01 -1.21418989e+00 -3.74851704e-01 -1.44685054e+00
1.59564823e-01 -3.24236125e-01 -8.48854706e-02 4.65804935e-01
2.52519518e-01 3.65056068e-01 7.60637671e-02 1.76278222e-02
-1.04579903e-01 6.96270883e-01 6.69406772e-01 -1.09061755e-01
-6.55024648e-02 -6.54111505e-01 -6.55099690e-01 9.81188059e-01
7.69025981e-01 -2.76305199e-01 -5.23348033e-01 -7.04607308e-01
5.06978989e-01 -4.93211687e-01 2.13853881e-01 -7.11424708e-01
-1.23853594e-01 -1.20496444e-01 -1.09181166e-01 -3.34193617e-01
3.69239211e-01 -8.54901671e-01 -1.11023128e-01 3.63968432e-01
-3.31278563e-01 5.62210560e-01 2.40128234e-01 1.16529435e-01
-5.12233317e-01 -5.50332069e-01 7.23703206e-01 -2.69312769e-01
-9.89873409e-01 2.56662536e-02 -4.10564452e-01 4.97395813e-01
9.38037694e-01 -1.12299524e-01 -2.77341932e-01 -3.53356928e-01
-2.95340300e-01 1.44844517e-01 3.71248960e-01 9.58465993e-01
5.78208148e-01 -1.47786283e+00 -1.05982459e+00 9.18566156e-03
6.27735555e-01 -8.85693908e-01 -3.50215659e-02 6.84568942e-01
-5.96506417e-01 4.70211118e-01 -3.00562590e-01 1.25004482e-02
-1.26960492e+00 5.62527657e-01 1.29681587e-01 -1.60400257e-01
-4.30711836e-01 5.68369865e-01 1.25856742e-01 -1.05516040e+00
-1.23199418e-01 -3.28509748e-01 -7.97388911e-01 4.64948624e-01
3.47076476e-01 2.74217695e-01 -2.60554980e-02 -1.13594735e+00
-6.76845789e-01 8.22692096e-01 7.78711513e-02 -4.36929375e-01
1.31547213e+00 -1.67892247e-01 -4.77600336e-01 9.07443762e-01
1.33519948e+00 4.95716065e-01 1.59805626e-01 -4.26138192e-01
2.25335196e-01 -3.38014632e-01 -1.23685293e-01 -7.67847598e-01
-6.78498864e-01 1.20742238e+00 4.54793900e-01 -2.92381831e-03
6.67220533e-01 2.33204290e-01 8.37998092e-01 3.81558836e-01
3.29235584e-01 -1.25062048e+00 -1.17884077e-01 1.03459692e+00
1.07932138e+00 -1.00766969e+00 -3.96515816e-01 -3.79081935e-01
-7.94193804e-01 1.03427434e+00 4.34639245e-01 -3.17197256e-02
6.87301755e-01 -9.06217471e-02 2.94853181e-01 -5.77117316e-02
-6.63977504e-01 -3.28348577e-01 4.27680761e-01 6.60819829e-01
9.75122213e-01 -1.67439058e-02 -1.00002193e+00 1.10901725e+00
-6.16710901e-01 -3.70209515e-01 4.12728339e-01 5.45490444e-01
-2.77223766e-01 -1.45115542e+00 -1.03760183e-01 1.23427600e-01
-4.04091209e-01 -5.91782808e-01 -7.40096390e-01 1.20598781e+00
3.56234699e-01 8.03411722e-01 1.13104694e-01 -3.01134646e-01
4.04885143e-01 2.67095536e-01 4.07458574e-01 -6.03894889e-01
-5.27356327e-01 -3.45066339e-01 3.35581601e-01 -1.69959262e-01
-3.50475609e-01 -3.97658944e-01 -1.20134878e+00 -4.80676979e-01
-1.06691577e-01 4.70125586e-01 7.83319235e-01 8.11309695e-01
7.08718970e-02 1.52333170e-01 2.59536237e-01 -4.71281320e-01
-6.14028156e-01 -1.23445714e+00 -7.33306468e-01 5.09391367e-01
-3.11107546e-01 -5.09791315e-01 -2.08473936e-01 -2.93538630e-01] | [10.713164329528809, 9.702906608581543] |
07066ea9-c00c-4e2b-8a7f-319b8db514b1 | transformers-for-ct-reconstruction-from | 2305.06965 | null | https://arxiv.org/abs/2305.06965v1 | https://arxiv.org/pdf/2305.06965v1.pdf | Transformers for CT Reconstruction From Monoplanar and Biplanar Radiographs | Computed Tomography (CT) scans provide detailed and accurate information of internal structures in the body. They are constructed by sending x-rays through the body from different directions and combining this information into a three-dimensional volume. Such volumes can then be used to diagnose a wide range of conditions and allow for volumetric measurements of organs. In this work, we tackle the problem of reconstructing CT images from biplanar x-rays only. X-rays are widely available and even if the CT reconstructed from these radiographs is not a replacement of a complete CT in the diagnostic setting, it might serve to spare the patients from radiation where a CT is only acquired for rough measurements such as determining organ size. We propose a novel method based on the transformer architecture, by framing the underlying task as a language translation problem. Radiographs and CT images are first embedded into latent quantized codebook vectors using two different autoencoder networks. We then train a GPT model, to reconstruct the codebook vectors of the CT image, conditioned on the codebook vectors of the x-rays and show that this approach leads to realistic looking images. To encourage further research in this direction, we make our code publicly available on GitHub: XXX. | ['Daniel Truhn', 'Johannes Stegmaier', 'Christiane Kuhl', 'Sven Nebelung', 'Tianyu Han', 'Gustav Müller-Franzes', 'Firas Khader'] | 2023-05-11 | null | null | null | null | ['computed-tomography-ct'] | ['methodology'] | [ 2.43666589e-01 3.45513999e-01 -8.34415182e-02 -4.11355197e-01
-7.23445237e-01 -3.60258549e-01 5.08255541e-01 1.55960992e-01
-2.88893133e-01 4.54166532e-01 3.30266118e-01 -3.08532268e-01
2.42904555e-02 -1.21647513e+00 -9.15714741e-01 -7.62234509e-01
1.39190787e-02 1.04297817e+00 -6.36225790e-02 -1.39579559e-02
-3.38733792e-01 6.38835013e-01 -1.31240213e+00 3.92557979e-01
1.39976770e-01 9.30151224e-01 2.93546796e-01 6.80571377e-01
-8.78426060e-03 7.55147517e-01 -3.85010451e-01 -3.33819240e-01
2.89262682e-01 -5.88545620e-01 -9.66295302e-01 4.93042648e-01
3.03626414e-02 -7.74857819e-01 -3.59539241e-01 7.34533250e-01
2.34868422e-01 -9.38031301e-02 6.58914626e-01 -6.84443593e-01
-7.01938093e-01 7.06034899e-01 -2.14063689e-01 -2.48303309e-01
5.44529855e-01 -1.08291253e-01 7.89748311e-01 -5.56935310e-01
8.92713785e-01 1.15224445e+00 5.05777955e-01 7.15704858e-01
-9.92043495e-01 -8.42500776e-02 -4.29275125e-01 -8.28853548e-02
-9.55301404e-01 -1.36602158e-02 7.14227974e-01 -4.51803327e-01
3.84860843e-01 5.84770620e-01 1.10013378e+00 1.27074969e+00
4.34299439e-01 5.52517474e-01 1.20275831e+00 -4.66741800e-01
3.03426445e-01 7.04078451e-02 -3.47128630e-01 8.48973989e-01
-6.58841729e-02 1.87725857e-01 -1.90975312e-02 -1.37855828e-01
1.07046628e+00 5.82163036e-01 -5.46609521e-01 -4.19591606e-01
-1.33107984e+00 1.08924747e+00 7.36200988e-01 6.10653222e-01
-6.94141924e-01 1.38359576e-01 2.76449621e-01 2.90452868e-01
1.65431932e-01 1.24471299e-01 -1.90817155e-02 3.38925868e-01
-8.61047864e-01 2.51669616e-01 7.31903732e-01 6.70828104e-01
2.60827273e-01 -1.66480988e-01 2.44914591e-01 6.17857873e-01
3.52571785e-01 5.98434508e-01 8.57770801e-01 -9.02659416e-01
2.63999045e-01 4.48056698e-01 -3.15234840e-01 -7.05164611e-01
-4.41378653e-01 -1.60372779e-01 -9.22252953e-01 1.26531953e-02
3.20382416e-01 1.62422553e-01 -1.03079569e+00 1.27675641e+00
4.80009466e-01 -2.11573616e-01 -2.21965298e-01 1.26655042e+00
8.76240253e-01 8.10427308e-01 -1.18154697e-01 -8.64204019e-03
1.61987484e+00 -4.83667433e-01 -4.96066689e-01 1.15179360e-01
3.40273201e-01 -5.17729521e-01 7.26553202e-01 5.59253871e-01
-1.28928947e+00 -3.99259597e-01 -1.02577841e+00 -2.28374794e-01
-1.18548639e-01 5.17348982e-02 3.10697228e-01 4.79055434e-01
-9.97693837e-01 7.79389501e-01 -1.33920681e+00 -3.45974952e-01
2.10837170e-01 2.72568703e-01 -5.10965347e-01 -2.87454456e-01
-1.00482821e+00 8.32688332e-01 1.62550956e-01 4.45235595e-02
-6.99207067e-01 -3.82922769e-01 -9.26983893e-01 1.01002112e-01
3.63880008e-01 -8.79294276e-01 1.14062750e+00 -6.04081571e-01
-1.44744670e+00 9.18839693e-01 2.66200155e-01 -3.29357058e-01
5.96688032e-01 -8.56948271e-03 3.76168755e-03 5.54499447e-01
-1.20285325e-01 6.18817329e-01 8.12264740e-01 -1.32841527e+00
8.00317712e-03 -7.15224147e-01 1.66888356e-01 8.46397951e-02
1.03342481e-01 -1.95281729e-01 -4.91417319e-01 -5.49280107e-01
5.76168537e-01 -1.11547709e+00 -2.84684747e-01 1.69500157e-01
-5.12747228e-01 3.99721473e-01 5.00132322e-01 -1.03443885e+00
5.36240399e-01 -1.81921637e+00 4.94284391e-01 1.99416295e-01
3.12242985e-01 -2.48258799e-01 2.61453271e-01 4.96558547e-01
-2.88579643e-01 -2.58104373e-02 -6.69128478e-01 -3.22874725e-01
-1.16633110e-01 6.06265604e-01 -2.02831194e-01 5.86150885e-01
-3.32732163e-02 9.77169752e-01 -7.55373299e-01 -6.35235608e-01
4.43536162e-01 8.06472480e-01 -3.80282044e-01 2.24777937e-01
-5.70053682e-02 8.84361565e-01 -6.55798018e-01 5.16464114e-01
4.24144775e-01 -4.35617119e-01 2.95464694e-01 -1.55455410e-01
2.39263594e-01 9.35600996e-02 -7.19249487e-01 1.77501631e+00
-4.86761659e-01 2.41024524e-01 9.90061164e-02 -1.15823138e+00
5.84985316e-01 6.23943746e-01 8.19188654e-01 -8.29186797e-01
4.33148146e-01 -3.94066647e-02 -1.44528314e-01 -5.25071144e-01
9.08649862e-02 -9.76175547e-01 -1.43263891e-01 6.25660777e-01
2.87408624e-02 -6.97176695e-01 -1.93339318e-01 -7.28723034e-02
1.00919867e+00 4.66675609e-02 2.97072649e-01 2.42601842e-01
4.21645224e-01 -2.08039805e-01 -6.50369152e-02 3.28374237e-01
3.04178447e-01 9.18276012e-01 1.96645662e-01 -7.75760889e-01
-1.34946966e+00 -1.36892676e+00 -5.10330856e-01 4.61440593e-01
-3.49573702e-01 -1.12140022e-01 -8.76519501e-01 -4.60915565e-01
-1.69380724e-01 5.18976927e-01 -8.32871974e-01 7.40500987e-02
-6.23165548e-01 -4.90600020e-01 2.22126797e-01 5.15979111e-01
1.19017214e-01 -1.12073112e+00 -1.05749321e+00 1.85244918e-01
-3.62085611e-01 -1.05126154e+00 -6.19714111e-02 1.37473553e-01
-1.26169050e+00 -1.05597973e+00 -1.11799240e+00 -5.74191153e-01
6.96502209e-01 -2.59838939e-01 1.11279321e+00 4.04804975e-01
-4.75288421e-01 6.22645199e-01 -4.42772239e-01 -3.46792340e-01
-8.27648938e-01 -3.90453190e-01 2.06381781e-03 -6.56374171e-02
-2.76795864e-01 -6.27560973e-01 -4.90845799e-01 -3.57841328e-02
-1.34072804e+00 1.19500905e-01 5.38015008e-01 9.39108253e-01
8.89215946e-01 -1.13866769e-01 -2.49653429e-01 -1.06281829e+00
2.74829537e-01 -5.13695121e-01 -4.49851036e-01 4.74406034e-02
1.25933224e-02 2.46087104e-01 8.69561791e-01 -1.48007095e-01
-8.01425397e-01 1.13391038e-03 -6.85209036e-01 -6.78916752e-01
-2.23409474e-01 4.40925896e-01 1.97913364e-01 -5.07751387e-03
5.16132653e-01 4.47797298e-01 -1.83899887e-02 -5.11220694e-01
1.99405447e-01 3.85581732e-01 7.05516934e-01 -3.43176723e-01
5.48752129e-01 8.07407022e-01 1.37087703e-01 -9.33970451e-01
-5.03869236e-01 -2.17141896e-01 -9.05722916e-01 -2.45660201e-01
1.08393705e+00 -6.29428685e-01 -5.94119251e-01 -6.07586466e-02
-8.93673480e-01 -5.60179017e-02 -5.13347924e-01 8.03910315e-01
-8.48164856e-01 4.94619042e-01 -9.29154813e-01 -3.67727995e-01
-1.91138178e-01 -1.41151488e+00 1.47163475e+00 -2.36091971e-01
8.82307347e-03 -1.14326119e+00 1.96275011e-01 3.78281683e-01
8.46362188e-02 5.49537599e-01 1.10980856e+00 -2.58325845e-01
-5.73358119e-01 -3.47371578e-01 2.63850361e-01 3.49390507e-01
2.22493690e-02 -4.91543382e-01 -7.93768585e-01 -2.59885877e-01
7.17370033e-01 -4.31305319e-01 6.39495194e-01 6.29805148e-01
1.47909904e+00 -2.46977657e-01 -2.47976974e-01 7.20383942e-01
1.56826866e+00 8.79312009e-02 6.63369358e-01 -7.86143467e-02
6.82088971e-01 5.64479649e-01 1.28777295e-01 5.13214886e-01
2.87500739e-01 6.53325737e-01 5.37193358e-01 -1.08484440e-01
-3.85016538e-02 -1.97682321e-01 -1.15856670e-01 1.10285664e+00
-3.06293875e-01 -2.33061984e-01 -9.80610013e-01 2.95826316e-01
-1.01427627e+00 -8.52876902e-01 -6.03611916e-02 2.12300205e+00
5.62479794e-01 -9.94030535e-02 -1.07861280e-01 3.49108398e-01
3.84438604e-01 -1.52279601e-01 -2.87289709e-01 -3.09570402e-01
3.49006772e-01 5.56005001e-01 3.16575080e-01 4.43549067e-01
-7.96266377e-01 2.06247449e-01 6.74638271e+00 8.65903348e-02
-1.53953588e+00 3.39507729e-01 5.54827750e-01 -7.40811462e-03
-4.09504324e-01 -1.46608576e-01 -1.00890724e-02 4.13164049e-01
1.06062961e+00 2.97788799e-01 2.22079158e-01 6.72298312e-01
-1.12593502e-01 -2.25507244e-01 -1.34264433e+00 1.07632446e+00
9.82391760e-02 -1.19572592e+00 1.13346674e-01 2.53836095e-01
2.99819320e-01 1.87860131e-01 7.07697570e-02 1.05846137e-01
1.89661354e-01 -9.96820986e-01 7.29772508e-01 6.09739184e-01
8.41969371e-01 -3.86863351e-01 6.87027752e-01 5.58060884e-01
-7.24043012e-01 4.93352339e-02 -3.92894208e-01 2.51620322e-01
4.10784900e-01 5.41622341e-01 -9.95746076e-01 7.53901005e-01
5.95297217e-01 2.53970057e-01 -2.46305317e-01 7.21769333e-01
-5.10778278e-02 5.83387256e-01 -4.49795753e-01 2.96362877e-01
1.90452725e-01 -4.09476101e-01 3.79486799e-01 6.62924945e-01
6.01855040e-01 4.80413377e-01 3.55519936e-03 9.14224565e-01
-1.05853625e-01 8.90969858e-02 -7.94836521e-01 8.86459276e-02
-1.39702916e-01 9.87520218e-01 -9.80148196e-01 -5.02331555e-01
-4.23026174e-01 1.07863212e+00 6.15050197e-02 -3.41953039e-02
-8.14862847e-01 2.59011477e-01 -2.35813990e-01 4.25510377e-01
2.57807851e-01 -9.53406990e-02 -3.59404869e-02 -1.19782901e+00
1.31278053e-01 -6.78012371e-01 2.52309680e-01 -1.08073902e+00
-9.40868795e-01 8.63289773e-01 1.63359359e-01 -1.36424673e+00
-6.24344170e-01 -7.24597514e-01 -2.18415692e-01 5.62693298e-01
-1.06377721e+00 -1.16029692e+00 -1.66777194e-01 5.05440891e-01
3.72691244e-01 3.38028044e-01 1.02075648e+00 2.19645962e-01
1.92390099e-01 4.23339009e-02 3.06973040e-01 3.23877662e-01
1.99997365e-01 -1.31508136e+00 2.52119809e-01 2.72116154e-01
3.49397808e-01 4.79222000e-01 6.15412056e-01 -6.07336938e-01
-1.58809435e+00 -8.81698549e-01 6.66925371e-01 -5.97974598e-01
1.11406505e-01 -2.93854803e-01 -9.06002998e-01 1.06695032e+00
-1.96004976e-02 3.02808732e-01 6.98240697e-01 -3.40975195e-01
1.13051504e-01 3.33428919e-01 -1.16874433e+00 1.37547210e-01
6.01456285e-01 -4.23429519e-01 -7.26097047e-01 5.57728410e-01
4.55534875e-01 -7.79793918e-01 -1.29166639e+00 -4.53057960e-02
6.16166294e-01 -1.08951735e+00 1.21838093e+00 -2.95275658e-01
8.92889977e-01 2.27363959e-01 -2.04901457e-01 -1.34235835e+00
-3.08295842e-02 2.14707151e-01 1.80593133e-01 3.34451526e-01
3.78139466e-02 -4.38201010e-01 8.43828082e-01 3.05624485e-01
-5.48500456e-02 -8.41676235e-01 -9.98603880e-01 -4.22086358e-01
2.34759688e-01 -6.48682892e-01 6.52672231e-01 9.08785343e-01
-1.89924061e-01 9.76131391e-03 -2.59875417e-01 4.96521294e-02
6.17163420e-01 2.78213352e-01 3.71579260e-01 -1.17186308e+00
-6.95206821e-01 7.63279647e-02 -5.57146788e-01 -8.17781627e-01
-1.54328182e-01 -1.11218512e+00 -7.51455575e-02 -1.47318697e+00
2.48963535e-01 -3.43222022e-01 2.62020677e-01 3.73937905e-01
3.51456881e-01 4.30578917e-01 4.22550678e-01 2.24501148e-01
-6.97707236e-02 4.70685899e-01 1.70756447e+00 -1.28459170e-01
1.56776845e-01 6.35480061e-02 -2.30376944e-01 8.55698228e-01
3.10042351e-01 -6.58475637e-01 -2.81680107e-01 -5.69026828e-01
1.00064069e-01 1.02260602e+00 5.82604825e-01 -1.11498928e+00
-2.61386856e-02 1.06247798e-01 7.93776989e-01 -7.65558541e-01
6.22628391e-01 -1.04106832e+00 4.25163835e-01 7.52168000e-01
-2.24981964e-01 1.55700639e-01 -1.46648422e-01 4.53133881e-01
-3.44306320e-01 -3.72081965e-01 6.54213130e-01 -7.80141950e-01
5.23903733e-03 4.07290310e-01 -3.03901225e-01 -4.07910079e-01
8.63382936e-01 -1.41081765e-01 4.86412525e-01 -6.02098286e-01
-1.24349141e+00 -2.27775037e-01 5.97076356e-01 1.02137104e-01
8.99578869e-01 -1.39575124e+00 -6.93376005e-01 5.02445638e-01
-1.47133589e-01 2.86072284e-01 2.87901998e-01 7.39256024e-01
-9.15438056e-01 6.98683739e-01 -4.93188769e-01 -9.27055657e-01
-1.09498787e+00 7.34831810e-01 3.01389992e-01 -4.10514921e-01
-1.18791151e+00 5.48764944e-01 4.58381325e-01 -7.17246294e-01
-2.47784510e-01 -8.49887311e-01 -1.16069816e-01 -2.73934513e-01
4.68142599e-01 -1.17662013e-01 3.41649801e-01 -9.08227146e-01
-4.30551060e-02 7.50355482e-01 2.75036782e-01 -4.20461923e-01
1.62718379e+00 -1.41286582e-01 -4.81921062e-02 4.99257416e-01
1.51120508e+00 -1.73869506e-01 -6.54138446e-01 -1.30179673e-01
-4.90687937e-01 -3.99470180e-01 1.73525319e-01 -5.56815684e-01
-1.26167238e+00 1.15131176e+00 5.89293599e-01 1.72037020e-01
1.32635057e+00 3.60941231e-01 9.04947579e-01 1.74013242e-01
6.14472687e-01 -6.26310408e-01 -3.02338880e-02 4.58926968e-02
1.09072149e+00 -9.40401077e-01 2.68347442e-01 -3.44525367e-01
-4.84069973e-01 1.33121824e+00 -1.74229190e-01 -1.71917826e-01
5.91381490e-01 2.40353450e-01 1.05705969e-02 -4.77400213e-01
-5.48446178e-01 1.41609445e-01 2.56444871e-01 5.02226353e-01
4.00008917e-01 3.38031918e-01 -1.39984608e-01 2.72273839e-01
-6.21677220e-01 1.02888457e-01 6.28865600e-01 1.12778366e+00
-1.10103413e-01 -1.25296175e+00 -9.45212483e-01 5.02398133e-01
-5.65231085e-01 2.34003901e-01 -1.64159670e-01 8.99278760e-01
6.02216311e-02 1.57886282e-01 6.45434931e-02 -1.11476712e-01
2.02120349e-01 8.31678063e-02 9.82903481e-01 -6.17196620e-01
-3.06092769e-01 1.21224053e-01 -3.48644286e-01 -6.63736880e-01
-3.99766535e-01 -7.92581022e-01 -1.44001210e+00 -3.53077687e-02
-1.43213391e-01 -1.72401499e-02 1.06304324e+00 9.76111114e-01
-4.38136429e-01 6.19983733e-01 5.13014257e-01 -9.25029933e-01
-6.58360004e-01 -6.72199726e-01 -7.42564201e-01 7.48800099e-01
5.36069274e-01 -5.81780493e-01 2.00498123e-02 4.55115020e-01] | [13.376460075378418, -2.6028034687042236] |
e54a7ecc-5e67-474f-88e7-c349e886d02d | acinoset-a-3d-pose-estimation-dataset-and | 2103.13282 | null | https://arxiv.org/abs/2103.13282v1 | https://arxiv.org/pdf/2103.13282v1.pdf | AcinoSet: A 3D Pose Estimation Dataset and Baseline Models for Cheetahs in the Wild | Animals are capable of extreme agility, yet understanding their complex dynamics, which have ecological, biomechanical and evolutionary implications, remains challenging. Being able to study this incredible agility will be critical for the development of next-generation autonomous legged robots. In particular, the cheetah (acinonyx jubatus) is supremely fast and maneuverable, yet quantifying its whole-body 3D kinematic data during locomotion in the wild remains a challenge, even with new deep learning-based methods. In this work we present an extensive dataset of free-running cheetahs in the wild, called AcinoSet, that contains 119,490 frames of multi-view synchronized high-speed video footage, camera calibration files and 7,588 human-annotated frames. We utilize markerless animal pose estimation to provide 2D keypoints. Then, we use three methods that serve as strong baselines for 3D pose estimation tool development: traditional sparse bundle adjustment, an Extended Kalman Filter, and a trajectory optimization-based method we call Full Trajectory Estimation. The resulting 3D trajectories, human-checked 3D ground truth, and an interactive tool to inspect the data is also provided. We believe this dataset will be useful for a diverse range of fields such as ecology, neuroscience, robotics, biomechanics as well as computer vision. | ['Amir Patel', 'Mackenzie W. Mathis', 'Alexander Mathis', 'Fred Nicolls', 'Ricardo Jericevich', 'Naoya Muramatsu', 'Liam Clark', 'Daniel Joska'] | 2021-03-24 | null | null | null | null | ['animal-pose-estimation'] | ['computer-vision'] | [-1.49577469e-01 -3.68233681e-01 1.50269508e-01 -6.69580624e-02
-2.56770909e-01 -7.16072023e-01 2.18835399e-01 -1.70543045e-01
-6.76243663e-01 7.94153810e-01 -1.71247333e-01 1.28343239e-01
-5.97634055e-02 -3.75106245e-01 -8.83275092e-01 -4.86421674e-01
-8.68719637e-01 5.78928888e-01 4.72734004e-01 -4.58555996e-01
2.06317842e-01 6.77458823e-01 -1.41730630e+00 -5.01405358e-01
5.68638742e-01 6.13545835e-01 3.57623219e-01 7.21675634e-01
6.71484709e-01 1.67888060e-01 -1.09789357e-01 -3.66057962e-01
2.47724220e-01 -4.79103744e-01 -3.74052316e-01 -1.07570425e-01
5.02010643e-01 -5.15116751e-01 -1.36456236e-01 6.56880498e-01
4.80825543e-01 2.02372357e-01 3.91830862e-01 -1.15977669e+00
2.92884558e-01 2.00088322e-01 -4.96240646e-01 2.02810422e-01
3.92187178e-01 4.83165979e-01 7.36074328e-01 -4.81451780e-01
1.06025445e+00 1.09230185e+00 1.15549839e+00 5.03950357e-01
-1.24838972e+00 -6.19875491e-01 -4.18980181e-01 3.44076276e-01
-1.03455102e+00 -3.76770228e-01 5.39996684e-01 -9.19937968e-01
7.83721149e-01 -1.90893322e-01 1.35452914e+00 1.30472112e+00
6.50876641e-01 3.93815309e-01 8.83996665e-01 2.52619952e-01
2.06033766e-01 -7.78963625e-01 -3.80772948e-01 9.98589516e-01
5.28516114e-01 1.54085502e-01 -8.19977522e-01 -3.01678181e-01
9.03976679e-01 1.61497250e-01 -4.00947601e-01 -7.34329283e-01
-1.67937803e+00 5.66854119e-01 4.47518885e-01 -5.25508583e-01
-3.10814530e-01 4.62280899e-01 2.49424845e-01 1.16346493e-01
5.70753776e-02 5.17884851e-01 -5.67648232e-01 -7.51105070e-01
-5.70357740e-01 8.29589725e-01 8.76750767e-01 8.93180668e-01
5.79929888e-01 3.39465141e-02 8.58157516e-01 6.09243929e-01
5.38230598e-01 9.42686558e-01 3.71174514e-01 -1.38248909e+00
3.46069962e-01 5.49376369e-01 7.30198920e-02 -1.20169759e+00
-9.84041095e-01 -2.41124004e-01 -4.14770037e-01 4.16816235e-01
3.69497299e-01 -3.17763656e-01 -4.62805778e-01 1.70564973e+00
6.69435084e-01 -1.75890431e-01 -2.73741126e-01 1.20523429e+00
4.95822668e-01 2.70504922e-01 -2.90793687e-01 7.74197504e-02
1.24492359e+00 -6.02997839e-01 -1.76870868e-01 -4.85512286e-01
5.37131429e-01 -4.76427734e-01 7.35791743e-01 2.25319475e-01
-7.82076776e-01 -7.21262321e-02 -1.21556902e+00 1.35876954e-01
-1.21756174e-01 -2.73301769e-02 5.13491750e-01 1.20330170e-01
-4.65084136e-01 8.00787091e-01 -1.55817699e+00 -9.28521276e-01
2.99067497e-01 1.75692454e-01 -9.21443880e-01 3.52299750e-01
-8.18081260e-01 1.20683718e+00 8.48842338e-02 2.06221998e-01
-1.13408494e+00 -4.87082273e-01 -1.03646481e+00 -5.36659896e-01
3.40633661e-01 -8.18757772e-01 1.24061775e+00 -8.45655054e-02
-1.48836863e+00 1.19926167e+00 3.90114099e-01 -6.49883568e-01
8.06647539e-01 -7.38803089e-01 9.53679010e-02 2.35342875e-01
3.60561728e-01 7.03773022e-01 5.15164256e-01 -8.69147539e-01
-5.62574506e-01 -7.68717945e-01 -3.67276371e-01 3.34963381e-01
2.56042928e-01 -2.99652725e-01 -1.78027362e-01 -6.41703010e-01
5.30505776e-01 -1.45672250e+00 -3.13086569e-01 6.78215861e-01
-1.09140076e-01 4.85793114e-01 6.86634362e-01 -7.15463877e-01
5.99291742e-01 -1.91452718e+00 6.93913460e-01 -2.99402267e-01
1.65852547e-01 -5.45695052e-02 2.51154780e-01 7.32204139e-01
5.04889607e-01 -1.69034421e-01 -5.49868286e-01 -1.55250560e-02
-2.89668828e-01 2.04146072e-01 1.90741420e-01 1.13092244e+00
-1.62188664e-01 6.77320600e-01 -1.03191769e+00 -5.10157466e-01
2.13089213e-01 1.37685955e-01 -7.32405603e-01 2.52365559e-01
4.04763743e-02 6.62253857e-01 -2.40301728e-01 9.69492793e-01
2.49204710e-01 1.79313645e-01 2.33381689e-01 -1.91481739e-01
-5.28661430e-01 -2.13223368e-01 -8.56655478e-01 2.05209851e+00
-1.04643784e-01 7.26618767e-01 5.95831454e-01 -6.15336716e-01
6.35109186e-01 -2.74544179e-01 7.66620278e-01 -2.54393458e-01
2.88614303e-01 4.77872074e-01 1.31965593e-01 -7.62400270e-01
5.12957633e-01 -1.08059995e-01 -1.77883700e-01 -1.11215301e-02
1.41315803e-01 -5.43111742e-01 1.62403345e-01 -6.65821582e-02
1.22499526e+00 7.57869720e-01 4.88032997e-01 -3.57485026e-01
4.25948165e-02 5.60724616e-01 7.21094370e-01 1.92054585e-01
-5.46148837e-01 6.86064541e-01 1.79658592e-01 -4.66920018e-01
-1.17120516e+00 -1.23624265e+00 -2.36511737e-01 6.36101663e-01
4.62794781e-01 -5.05280077e-01 -6.86323881e-01 4.25522029e-02
5.71668684e-01 -1.43885612e-01 -5.06370187e-01 -3.04390073e-01
-9.47032094e-01 -5.11879444e-01 5.97142696e-01 3.31789166e-01
6.64146841e-01 -8.41509104e-01 -1.49901187e+00 4.47981626e-01
-3.00112516e-01 -1.18953538e+00 -9.01173614e-03 1.17000945e-01
-9.17014897e-01 -1.31719041e+00 -6.08563900e-01 -4.18661237e-01
1.81095496e-01 4.48498130e-01 6.66150510e-01 2.83618104e-02
-4.53712940e-01 2.47988239e-01 -5.61010122e-01 -1.90090612e-01
-1.19900137e-01 -3.02462019e-02 5.90231478e-01 -6.85375214e-01
-2.10647225e-01 -8.24240446e-01 -5.67187130e-01 8.00030231e-01
-3.59325051e-01 1.45397678e-01 4.15010214e-01 8.09379160e-01
7.32253194e-01 -6.52705967e-01 5.56299277e-02 -3.27837408e-01
2.31431201e-02 -5.89298606e-01 -7.45060623e-01 -2.14530021e-01
1.90787628e-01 -3.47911447e-01 5.91089308e-01 -2.44957000e-01
-5.33773184e-01 4.17532653e-01 -4.16107655e-01 -4.11276817e-02
-1.30129710e-01 6.66900694e-01 1.38923183e-01 -3.41010332e-01
8.30184698e-01 -3.35095115e-02 5.01931369e-01 -5.52092314e-01
1.84933677e-01 5.02635777e-01 9.51339185e-01 -4.16094840e-01
7.42021620e-01 7.56206512e-01 3.92984301e-01 -1.33562803e+00
-3.27942729e-01 -4.00305450e-01 -8.34668398e-01 -6.30667865e-01
9.01281357e-01 -9.63320792e-01 -8.89172316e-01 8.32397103e-01
-8.38392615e-01 -6.88755155e-01 2.02362835e-01 8.27419996e-01
-1.12400150e+00 5.15872538e-01 -4.48438495e-01 -4.10736561e-01
-2.90056378e-01 -1.16438341e+00 1.28936207e+00 1.98383003e-01
-4.34354722e-01 -5.08452475e-01 5.05100191e-01 5.33397496e-01
1.03005178e-01 1.06703639e+00 2.77750164e-01 -1.49620697e-01
-3.91215384e-01 -2.68952817e-01 4.17663515e-01 -1.39791027e-01
-7.38021359e-02 3.75114322e-01 -2.83308953e-01 -5.51947594e-01
-1.80813700e-01 -5.35890818e-01 5.09486318e-01 3.76663625e-01
4.39378887e-01 -2.59496663e-02 -4.23172206e-01 1.03190362e+00
1.08828199e+00 -1.08866431e-01 2.08019897e-01 7.10524082e-01
6.57404542e-01 7.22363710e-01 9.43864048e-01 5.81802189e-01
6.49675667e-01 1.03370917e+00 9.38962936e-01 3.22865427e-01
5.45473620e-02 -4.34359699e-01 5.85308790e-01 7.29197145e-01
-5.57660937e-01 5.52562252e-02 -1.13130939e+00 3.90199363e-01
-1.89397037e+00 -1.02333462e+00 -3.72772276e-01 2.27781343e+00
3.19294393e-01 -3.46896239e-02 3.68473917e-01 -1.53422222e-01
4.25316244e-01 -2.36247018e-01 -8.85204256e-01 2.31692091e-01
-8.87162536e-02 -3.65492672e-01 8.01453888e-01 6.96152747e-02
-1.08136988e+00 8.29549491e-01 6.22893381e+00 7.35203549e-02
-1.10190356e+00 -3.19910824e-01 -3.15972477e-01 -1.11250296e-01
3.95438820e-01 4.34302576e-02 -7.20311642e-01 4.59309280e-01
7.85408080e-01 -1.16326019e-01 4.07312125e-01 1.18525422e+00
2.91823983e-01 -4.63274777e-01 -9.27199781e-01 9.94044602e-01
1.23776041e-01 -1.18272769e+00 -5.82922816e-01 3.26298088e-01
5.14225364e-01 5.52458584e-01 -5.00283539e-01 -1.85817048e-01
3.02779693e-02 -5.29514372e-01 1.22008610e+00 5.70854425e-01
5.70334136e-01 -5.24387360e-01 6.71635449e-01 7.63136983e-01
-1.28267848e+00 7.48086497e-02 -2.97162831e-01 -3.68682981e-01
5.64323127e-01 1.77820712e-01 -4.63428408e-01 3.91601950e-01
1.07541144e+00 1.06973040e+00 -7.14508355e-01 1.42355812e+00
-1.68540180e-01 3.53946656e-01 -7.07809567e-01 -1.76815540e-01
-2.07634032e-01 -4.17845637e-01 1.00740147e+00 7.87141800e-01
3.95271033e-01 4.34813201e-02 4.38403003e-02 6.03406489e-01
1.79420322e-01 -1.36494637e-01 -7.50221670e-01 3.93293612e-02
4.02347863e-01 1.41763401e+00 -1.00052774e+00 -6.34947866e-02
-8.39761868e-02 8.58811855e-01 1.42900810e-01 -2.09551170e-01
-9.59146559e-01 -3.20786595e-01 9.86741543e-01 1.82459235e-01
-7.92064518e-02 -1.01218605e+00 7.05572963e-02 -1.50831807e+00
4.63845534e-03 -6.40492022e-01 1.17002361e-01 -6.91055417e-01
-1.08263278e+00 1.44231811e-01 3.59528631e-01 -1.72396231e+00
-3.32183450e-01 -6.23093128e-01 -3.01176906e-01 1.34662569e-01
-9.17448938e-01 -1.11859047e+00 -6.09012067e-01 1.50184661e-01
4.64732081e-01 1.70909554e-01 2.94637591e-01 2.02242970e-01
-3.96136820e-01 -6.82052691e-03 1.81649536e-01 -2.22622544e-01
6.54789746e-01 -1.00845087e+00 5.63153923e-01 7.79853761e-01
-1.21263251e-01 6.33653641e-01 1.12662184e+00 -9.71223712e-01
-2.01139140e+00 -8.33700836e-01 1.18525863e-01 -5.87161660e-01
8.28862011e-01 -3.70485187e-01 -6.29836857e-01 7.41398096e-01
-5.26733875e-01 -3.96719202e-02 6.05103731e-01 -3.29063296e-01
4.32424456e-01 3.38751823e-03 -1.05338442e+00 6.64436936e-01
1.35305297e+00 1.09035254e-01 -6.89006388e-01 1.98987797e-01
3.43639672e-01 -7.92378783e-01 -9.18288350e-01 5.04862785e-01
1.30963182e+00 -9.98655438e-01 9.47904050e-01 -4.07967776e-01
4.72012162e-01 -5.72998583e-01 -1.04531623e-01 -1.35704601e+00
-1.54172361e-01 -3.63155991e-01 8.61257538e-02 6.95593417e-01
1.01331612e-02 -4.35018390e-01 8.41209352e-01 7.97293335e-02
-5.55582821e-01 -5.19698977e-01 -1.16968536e+00 -1.03905237e+00
1.46736000e-02 -9.29145142e-02 1.76127911e-01 6.16241813e-01
4.71080244e-02 4.87147495e-02 -6.37575030e-01 1.11194469e-01
1.04512596e+00 1.15617432e-01 1.49237978e+00 -1.39411104e+00
6.64521828e-02 -1.43313184e-01 -1.06349611e+00 -1.11052692e+00
-6.88998252e-02 -4.14486349e-01 4.08590317e-01 -1.29485810e+00
-2.11540267e-01 -1.30033299e-01 6.54967785e-01 3.64064872e-01
2.08915189e-01 3.62734735e-01 6.31809561e-03 3.21812123e-01
-3.96433145e-01 7.62521267e-01 1.08743536e+00 4.07235146e-01
5.04185632e-02 -2.97449250e-02 8.47706795e-02 1.03392243e+00
6.76874220e-01 -4.19362366e-01 9.13762227e-02 -2.46413291e-01
1.38625130e-01 3.21546346e-01 6.45135164e-01 -1.57903385e+00
2.25723848e-01 -3.77683312e-01 1.76610738e-01 -5.84699333e-01
5.38810015e-01 -9.37407136e-01 6.32290483e-01 9.35750902e-01
1.02117375e-01 1.90701708e-01 7.45732486e-02 7.28363216e-01
1.04888208e-01 4.22910489e-02 9.23241735e-01 -4.49031562e-01
-8.88762236e-01 3.06312174e-01 -7.70954788e-01 2.09088594e-01
1.21139896e+00 -4.85127032e-01 -2.88590103e-01 1.86522137e-02
-3.04779887e-01 4.20175254e-01 1.20109403e+00 2.68216431e-01
6.64422691e-01 -1.06665087e+00 -5.50197184e-01 3.30826551e-01
2.88479775e-01 7.40477219e-02 4.97482508e-01 1.06479728e+00
-1.56829512e+00 -2.52003253e-01 -9.06756878e-01 -9.13131952e-01
-1.20263588e+00 -1.20641023e-01 2.46084616e-01 5.24329424e-01
-8.40314865e-01 7.74868011e-01 -2.80479819e-01 -6.37821436e-01
-2.58296490e-01 -3.99629474e-01 -5.81102408e-02 2.50669103e-02
1.03298694e-01 7.11198032e-01 -1.01016521e-01 -1.11926961e+00
-5.88722408e-01 1.04288864e+00 6.56145394e-01 -2.72423010e-02
1.63881731e+00 -3.44713628e-01 -8.35223496e-02 5.48215091e-01
9.70647633e-01 -1.30193368e-01 -1.51137280e+00 3.63252997e-01
-1.11711130e-01 -6.21074319e-01 -2.65718877e-01 -1.58261493e-01
-6.60805047e-01 8.87065470e-01 4.90659297e-01 -6.49181157e-02
5.41016400e-01 -1.41046606e-02 9.04921472e-01 6.49745286e-01
9.99524534e-01 -9.97580767e-01 -5.43247685e-02 4.54683572e-01
1.06849909e+00 -1.28218591e+00 4.30136561e-01 -1.27719328e-01
-5.42517781e-01 1.15051270e+00 6.98986709e-01 -3.70978683e-01
4.16246295e-01 1.98623270e-01 1.12636685e-02 -1.87137350e-01
-6.45038247e-01 -6.64313138e-02 -1.22854665e-01 6.15556777e-01
9.22934934e-02 1.39414147e-02 -3.50801080e-01 3.49190563e-01
-7.76367128e-01 -1.28153265e-01 7.08279014e-01 1.46269321e+00
-8.10772598e-01 -6.84640169e-01 -3.79601091e-01 3.08776230e-01
-1.64681315e-01 4.23413575e-01 -3.43992829e-01 8.36708844e-01
-3.32862996e-02 4.42432344e-01 -2.10979149e-01 -6.56338036e-01
6.00498021e-01 -4.53157932e-01 4.94291544e-01 -5.37284136e-01
-2.40704700e-01 -4.77235466e-02 1.22542970e-01 -6.80107057e-01
-5.74977040e-01 -9.59475756e-01 -1.28448331e+00 -5.54223299e-01
-2.24612668e-01 -2.98951287e-02 9.09626782e-01 7.26367712e-01
2.23340690e-01 -1.01515889e-01 2.19689280e-01 -1.54221892e+00
-2.95245647e-01 -8.54555368e-01 -5.62337279e-01 1.91106483e-01
3.73013794e-01 -1.29664040e+00 -5.32721281e-01 4.01624799e-01] | [7.555330753326416, -1.0780836343765259] |
3ad28049-7300-450e-a854-86df11592298 | type-supervised-sequence-labeling-based-on | 2210.10240 | null | https://arxiv.org/abs/2210.10240v2 | https://arxiv.org/pdf/2210.10240v2.pdf | Type-supervised sequence labeling based on the heterogeneous star graph for named entity recognition | Named entity recognition is a fundamental task in natural language processing, identifying the span and category of entities in unstructured texts. The traditional sequence labeling methodology ignores the nested entities, i.e. entities included in other entity mentions. Many approaches attempt to address this scenario, most of which rely on complex structures or have high computation complexity. The representation learning of the heterogeneous star graph containing text nodes and type nodes is investigated in this paper. In addition, we revise the graph attention mechanism into a hybrid form to address its unreasonableness in specific topologies. The model performs the type-supervised sequence labeling after updating nodes in the graph. The annotation scheme is an extension of the single-layer sequence labeling and is able to cope with the vast majority of nested entities. Extensive experiments on public NER datasets reveal the effectiveness of our model in extracting both flat and nested entities. The method achieved state-of-the-art performance on both flat and nested datasets. The significant improvement in accuracy reflects the superiority of the multi-layer labeling strategy. | ['Hong Qi', 'Yu Jiang', 'Luguang Liang', 'Haotian Tang', 'Changjiang Zhou', 'Xueru Wen'] | 2022-10-19 | null | null | null | null | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [-3.54626067e-02 6.58060730e-01 -2.68626839e-01 -2.69104868e-01
-2.46096909e-01 -8.36643755e-01 4.21335906e-01 5.89899600e-01
-6.13329768e-01 6.73539400e-01 2.71413505e-01 -3.43516022e-01
-2.08550543e-02 -9.08496797e-01 -5.44225276e-01 -5.19958794e-01
-1.26572296e-01 6.36045814e-01 5.53311646e-01 -2.25081459e-01
1.72987044e-01 5.41890740e-01 -1.13759959e+00 1.42622426e-01
1.00794041e+00 5.37599266e-01 1.77041844e-01 2.42664441e-01
-9.08841372e-01 9.14498568e-01 -6.33179784e-01 -8.83837581e-01
-6.47978857e-02 -5.03002927e-02 -1.14600396e+00 -2.66004745e-02
3.73668820e-01 3.04077923e-01 -3.09225202e-01 1.08872151e+00
3.87137353e-01 -2.05655191e-02 5.50114632e-01 -1.16452551e+00
-5.68913996e-01 1.10518467e+00 -5.11149228e-01 3.49676222e-01
2.75251418e-01 -4.52133298e-01 1.51172817e+00 -8.18705261e-01
8.42652678e-01 9.55631554e-01 8.54693651e-01 3.16398263e-01
-8.28986347e-01 -3.47610980e-01 4.13898289e-01 2.42927074e-01
-1.59623325e+00 -1.63334653e-01 6.40102267e-01 -4.55420792e-01
1.08768785e+00 1.50841579e-01 9.41838548e-02 6.68816984e-01
-1.59132600e-01 6.60845160e-01 7.03989804e-01 -4.79184091e-01
-7.97977205e-03 9.79587212e-02 7.38258958e-01 8.73591959e-01
7.91673362e-01 -5.68699062e-01 -1.64335489e-01 -1.61571607e-01
4.48385924e-01 -2.47413889e-01 -2.60053486e-01 -4.15366918e-01
-1.07087684e+00 6.08122945e-01 4.96124655e-01 7.62645125e-01
-4.01836663e-01 -2.42911577e-01 5.39089143e-01 -7.18152598e-02
3.57464969e-01 5.03465116e-01 -7.95693099e-01 3.00970107e-01
-7.44709313e-01 -3.14660013e-01 1.14777720e+00 1.18136394e+00
7.77654648e-01 -1.10455729e-01 -1.51586697e-01 6.89381719e-01
1.78672865e-01 -2.02318300e-02 4.77892995e-01 -2.77091563e-01
6.71984851e-01 1.16433334e+00 2.23621428e-02 -1.19332933e+00
-9.30606663e-01 -6.53187513e-01 -7.30834961e-01 -4.49620754e-01
4.60016340e-01 -1.99918255e-01 -9.10685778e-01 1.60762882e+00
5.53008378e-01 1.92193896e-01 1.17667004e-01 5.83452165e-01
1.16522086e+00 5.03566504e-01 3.89808118e-01 -1.01550020e-01
1.72948134e+00 -9.67693925e-01 -9.27071929e-01 -1.11061096e-01
9.55779731e-01 -4.68745053e-01 5.55432558e-01 -2.23425150e-01
-5.84509552e-01 -3.44158113e-01 -6.97691441e-01 -1.64541587e-01
-9.12382603e-01 3.42785090e-01 6.89924717e-01 7.18664169e-01
-8.57713103e-01 3.87558192e-01 -5.71924210e-01 -5.64328432e-01
1.28931999e-02 3.90233576e-01 -5.74333608e-01 2.13905483e-01
-1.44295359e+00 7.67775536e-01 9.97510552e-01 2.28635117e-01
-1.86588556e-01 -5.07641137e-01 -1.12733924e+00 4.45070684e-01
8.54113042e-01 -4.50837612e-01 1.00294447e+00 -7.70377815e-01
-8.18768978e-01 8.95995080e-01 -2.21862689e-01 -4.75407690e-01
2.25391537e-01 3.61455940e-02 -6.69645548e-01 1.21037468e-01
1.34238392e-01 2.30041325e-01 3.52480531e-01 -1.23103261e+00
-6.48140669e-01 -3.10112655e-01 2.69629240e-01 1.37849748e-01
-4.92775917e-01 -4.89684520e-03 -5.36998510e-01 -6.86917543e-01
2.03307629e-01 -8.22877347e-01 -2.30022296e-01 -6.03803158e-01
-6.41264856e-01 -7.26290166e-01 6.37949347e-01 -7.76529431e-01
1.72382545e+00 -1.77686405e+00 -4.20847982e-02 1.43530533e-01
3.90604883e-01 4.58194345e-01 5.84738404e-02 8.31684291e-01
-2.05258653e-01 4.97237116e-01 -1.93591550e-01 -5.18655032e-02
1.16035461e-01 1.50422245e-01 -1.91649556e-01 2.85289109e-01
1.92288548e-01 9.03649926e-01 -8.81584466e-01 -8.62615883e-01
-2.63078153e-01 2.51352221e-01 -2.84089684e-01 7.80682191e-02
-1.53613701e-01 1.51109174e-01 -5.68740845e-01 5.37635863e-01
5.54921806e-01 -5.96408546e-01 8.54486108e-01 -5.64308703e-01
-2.00895965e-01 4.90971446e-01 -1.38354766e+00 1.40437627e+00
-1.27905011e-01 1.47068471e-01 -6.61405548e-02 -8.87301803e-01
8.97678196e-01 4.22011107e-01 4.40671384e-01 -2.12092414e-01
-7.56201446e-02 2.63389528e-01 5.60219847e-02 -7.19236076e-01
7.39354730e-01 1.32224038e-01 -2.98583090e-01 8.65790397e-02
2.98082322e-01 6.76551700e-01 6.24719799e-01 5.22953451e-01
1.06090426e+00 1.62550896e-01 7.46803820e-01 -3.38152170e-01
7.89512694e-01 5.98884746e-03 8.67493689e-01 6.67595029e-01
-5.50505593e-02 1.65536612e-01 7.39183486e-01 -4.45133746e-01
-8.99507642e-01 -6.30017936e-01 1.24261349e-01 1.19054794e+00
2.29413494e-01 -6.47000253e-01 -6.17153704e-01 -1.23229599e+00
-1.78625718e-01 6.59230173e-01 -4.72642511e-01 1.39309272e-01
-8.64061058e-01 -7.77249575e-01 7.64110863e-01 5.74388504e-01
4.89377856e-01 -1.09494364e+00 -2.82198757e-01 3.03615302e-01
-5.27936935e-01 -1.48335338e+00 -3.76635373e-01 1.01134866e-01
-5.42117536e-01 -1.22343779e+00 -4.04102445e-01 -1.24099267e+00
7.38654077e-01 -1.25165030e-01 1.18904936e+00 3.22246373e-01
1.07284218e-01 1.97178736e-01 -5.83776891e-01 -1.29114762e-01
-1.70897290e-01 6.76432967e-01 -1.51401713e-01 2.52938181e-01
4.29217815e-01 -2.67680496e-01 -2.01611951e-01 1.77713782e-01
-9.02662575e-01 -2.22675413e-01 6.68744326e-01 6.71729326e-01
2.77936548e-01 2.02478409e-01 7.01501191e-01 -1.45741284e+00
2.48560667e-01 -6.21133149e-01 -4.93237674e-01 5.95512450e-01
-6.91205680e-01 3.80137920e-01 7.13326395e-01 -2.64027745e-01
-1.23015690e+00 2.25266591e-01 -2.79309273e-01 6.17348626e-02
-2.68334597e-01 8.60325694e-01 -5.27994514e-01 1.19042411e-01
1.76898301e-01 6.45368323e-02 -7.77728260e-01 -6.99302435e-01
2.77634919e-01 6.09205365e-01 3.91065449e-01 -4.74330932e-01
7.32408583e-01 2.09971040e-01 -7.34268036e-03 -8.55341017e-01
-8.65197420e-01 -7.53273964e-01 -1.00950181e+00 1.51433907e-02
8.72379422e-01 -7.13836014e-01 -6.25074923e-01 2.36669809e-01
-1.34617794e+00 6.08699583e-02 -3.05500440e-02 2.09681138e-01
1.87912405e-01 9.40394223e-01 -8.31498086e-01 -8.08334470e-01
-2.60924786e-01 -6.80798769e-01 9.80304837e-01 2.15652555e-01
1.20560741e-02 -1.15257537e+00 1.04615130e-01 1.98360994e-01
-1.01575367e-02 1.27221510e-01 1.18076873e+00 -1.50571072e+00
-3.77260983e-01 -2.22112596e-01 -3.53801042e-01 -3.56193155e-01
5.64961024e-02 -1.21289544e-01 -6.84638917e-01 -1.13235749e-01
-4.43034351e-01 9.55068842e-02 8.20280254e-01 -1.76293314e-01
6.89606607e-01 -4.53030676e-01 -5.87201059e-01 2.88471103e-01
1.53021479e+00 1.16792604e-01 4.88678157e-01 4.07956451e-01
1.06755936e+00 8.30644906e-01 4.08760041e-01 2.34047666e-01
7.26956189e-01 5.87346733e-01 5.02665937e-01 -6.12033941e-02
7.60265142e-02 -4.95425642e-01 7.11962115e-03 1.00680006e+00
1.42991960e-01 -7.32869804e-01 -1.27497184e+00 6.89915895e-01
-1.73971319e+00 -9.45284426e-01 -6.06522322e-01 1.95609534e+00
6.96659863e-01 1.19316831e-01 -5.46977557e-02 7.89343119e-02
1.23747063e+00 2.51102626e-01 -1.49420232e-01 -2.06352666e-01
-2.63352811e-01 -1.48611799e-01 7.21196473e-01 3.53380740e-01
-1.23910260e+00 1.17739820e+00 5.69179583e+00 6.71644449e-01
-7.62095988e-01 5.10232821e-02 1.61761045e-01 8.13748002e-01
-3.01359743e-01 8.19015130e-02 -1.25879097e+00 4.69023883e-01
9.36311126e-01 -9.30486061e-03 -1.16908975e-01 7.19169259e-01
-1.02330759e-01 2.25815073e-01 -7.55271554e-01 5.16460359e-01
1.29256666e-01 -1.23014963e+00 1.66046038e-01 3.15125771e-02
5.00163615e-01 -5.97647168e-02 -6.78761721e-01 5.28354049e-01
4.08352584e-01 -6.14068329e-01 6.72435522e-01 3.19750607e-01
6.62992418e-01 -5.28156400e-01 1.14836705e+00 5.11566401e-01
-1.63169980e+00 -8.47697482e-02 -1.44896254e-01 2.02803805e-01
2.74117947e-01 4.79672313e-01 -8.90043139e-01 9.94813323e-01
5.54914296e-01 4.60530818e-01 -7.24281967e-01 1.13783193e+00
-4.59336549e-01 7.06174314e-01 -3.30430835e-01 -1.35534644e-01
2.54320592e-01 -2.33485997e-01 5.81132412e-01 1.68210816e+00
7.69133354e-03 1.53760910e-01 2.71508813e-01 4.08165365e-01
-4.21038896e-01 4.04205114e-01 -5.36556721e-01 -2.70874232e-01
6.43790781e-01 1.50520480e+00 -1.11802638e+00 -4.98391658e-01
-5.62569201e-01 7.67632306e-01 8.11792910e-01 2.48920843e-01
-7.90869653e-01 -6.95603013e-01 6.03551418e-02 -3.88027728e-02
7.77698636e-01 -3.52081001e-01 -1.58777505e-01 -1.29733396e+00
5.61618432e-02 -5.80965340e-01 9.33999896e-01 -5.01459777e-01
-1.27616143e+00 8.29113960e-01 -6.55274838e-02 -8.58715296e-01
4.34513427e-02 -5.48665822e-01 -4.62497264e-01 4.78976905e-01
-1.51976752e+00 -1.22145808e+00 -1.60398260e-01 4.24549490e-01
3.59794617e-01 1.64882004e-01 6.84839964e-01 5.92123210e-01
-8.93662930e-01 6.18306458e-01 -9.81139466e-02 7.97166646e-01
4.98156726e-01 -1.49228036e+00 3.69562626e-01 1.21765089e+00
3.77861142e-01 7.20479131e-01 5.07278860e-01 -9.18094337e-01
-1.14481103e+00 -1.10727334e+00 1.71552122e+00 -2.89990634e-01
9.54293251e-01 -4.31567311e-01 -1.17318618e+00 9.90185857e-01
2.43300870e-01 -7.98978284e-02 7.03671098e-01 1.77758977e-01
-6.13474309e-01 1.71347648e-01 -9.92432594e-01 4.66811776e-01
1.26402938e+00 -3.45912069e-01 -9.81796324e-01 1.30413994e-01
9.71480727e-01 -2.45413676e-01 -9.12128270e-01 4.23464179e-01
1.21168271e-01 -5.18118262e-01 6.18236721e-01 -7.75537252e-01
5.69715211e-03 -4.76376295e-01 1.91093370e-01 -9.16182935e-01
-4.16939288e-01 -5.37211776e-01 -3.53526354e-01 1.88274741e+00
6.72078729e-01 -7.51686037e-01 7.78099298e-01 4.13450986e-01
-1.69726968e-01 -3.27423245e-01 -7.46296644e-01 -7.15843976e-01
-2.87274033e-01 5.13707027e-02 5.93285501e-01 1.40771246e+00
1.89191222e-01 7.83921480e-01 -9.15819407e-02 5.61765969e-01
5.45952618e-01 2.58923054e-01 2.60518461e-01 -1.46263099e+00
5.24661988e-02 -2.06936315e-01 -4.32213694e-01 -8.94705176e-01
4.90885049e-01 -1.14708233e+00 -1.60988774e-02 -1.85354614e+00
2.47821189e-03 -5.37344992e-01 -3.25609833e-01 7.53355503e-01
-2.97433555e-01 -8.49442109e-02 6.80093393e-02 3.59746158e-01
-8.33789349e-01 2.00896442e-01 6.46468163e-01 -1.79595396e-01
-1.22391693e-01 -2.46286333e-01 -6.35939240e-01 7.50365138e-01
8.50091815e-01 -6.39792860e-01 -1.04174577e-01 -4.44060087e-01
4.96260285e-01 -1.65333092e-01 5.83070796e-04 -6.79139495e-01
5.99322796e-01 3.99937965e-02 -6.87204972e-02 -6.83681905e-01
-1.04423560e-01 -1.09672701e+00 1.31772876e-01 3.23855758e-01
-3.29028457e-01 2.04005882e-01 -1.44358054e-01 7.45255113e-01
-1.69656202e-01 -5.44342339e-01 5.11605024e-01 -2.27830216e-01
-1.13349831e+00 2.18847468e-01 -2.22550794e-01 3.06008250e-01
9.05506670e-01 1.26964338e-02 -5.23414493e-01 4.68361452e-02
-8.14428806e-01 2.55505323e-01 2.85616815e-01 3.85057986e-01
1.51475593e-01 -1.02856398e+00 -5.12099504e-01 -1.98700875e-01
1.77288249e-01 -3.69216055e-02 5.80954105e-02 9.09942389e-01
-6.00756586e-01 5.11457741e-01 1.10967092e-01 -1.94365576e-01
-1.26525176e+00 8.85503411e-01 3.33115846e-01 -7.63206244e-01
-5.05993247e-01 5.41143119e-01 1.07377477e-01 -5.51993251e-01
1.71083838e-01 -1.01427518e-01 -9.62468028e-01 4.04703617e-01
2.19385460e-01 3.91001672e-01 2.35197917e-01 -1.07579482e+00
-5.34749150e-01 4.14681345e-01 -1.05929568e-01 4.00313526e-01
1.19905984e+00 -3.32911402e-01 -4.57601637e-01 3.84738296e-01
9.48651850e-01 3.06211054e-01 -5.78089476e-01 -4.41224903e-01
9.34196532e-01 1.78223044e-01 -3.22419852e-01 -5.83618224e-01
-9.78179991e-01 4.72316712e-01 6.01404719e-03 5.39993107e-01
7.01926887e-01 1.49858028e-01 7.06297517e-01 6.19949579e-01
4.40255225e-01 -9.22085285e-01 -4.81319398e-01 7.72606134e-01
3.58925521e-01 -1.06474245e+00 -1.92092583e-02 -9.11310852e-01
-5.97831249e-01 1.05291426e+00 6.62627816e-01 1.61540106e-01
4.89300817e-01 2.18383536e-01 -1.00466318e-01 -3.73917103e-01
-4.64561939e-01 -5.52899837e-01 3.54752302e-01 4.21360284e-01
5.15412927e-01 -1.28144160e-01 -7.60976732e-01 6.19335890e-01
1.70930773e-01 -4.44390178e-01 5.33551335e-01 9.43451583e-01
-4.43466544e-01 -9.92248774e-01 -4.98141050e-02 2.61431873e-01
-8.60443234e-01 -4.08302367e-01 -5.84036827e-01 9.21803474e-01
1.80069536e-01 1.07220244e+00 2.21361872e-02 -1.38598070e-01
4.21748996e-01 4.22005236e-01 -9.52829514e-03 -6.49103642e-01
-9.90553737e-01 -1.72282442e-01 5.06959915e-01 -1.90534815e-01
-5.45358062e-01 -4.85650957e-01 -1.62636054e+00 6.91889673e-02
-6.56410754e-01 6.26452863e-01 4.73102570e-01 8.77339482e-01
4.50484186e-01 4.74742115e-01 4.07454431e-01 -3.44334930e-01
-3.49257112e-01 -9.02593791e-01 -6.38811588e-01 5.77569902e-01
3.05917859e-02 -5.49585760e-01 -2.70281225e-01 6.67382851e-02] | [9.594117164611816, 9.343138694763184] |
cfd531bf-beb6-4f6e-94ee-c9e719a6b158 | value-added-chemical-discovery-using | 1911.07630 | null | https://arxiv.org/abs/1911.07630v1 | https://arxiv.org/pdf/1911.07630v1.pdf | Value-Added Chemical Discovery Using Reinforcement Learning | Computer-assisted synthesis planning aims to help chemists find better reaction pathways faster. Finding viable and short pathways from sugar molecules to value-added chemicals can be modeled as a retrosynthesis planning problem with a catalyst allowed. This is a crucial step in efficient biomass conversion. The traditional computational chemistry approach to identifying possible reaction pathways involves computing the reaction energies of hundreds of intermediates, which is a critical bottleneck in silico reaction discovery. Deep reinforcement learning has shown in other domains that a well-trained agent with little or no prior human knowledge can surpass human performance. While some effort has been made to adapt machine learning techniques to the retrosynthesis planning problem, value-added chemical discovery presents unique challenges. Specifically, the reaction can occur in several different sites in a molecule, a subtle case that has never been treated in previous works. With a more versatile formulation of the problem as a Markov decision process, we address the problem using deep reinforcement learning techniques and present promising preliminary results. | ['Rajeev Surendran Assary', 'Hieu Doan', 'Sandeep Madireddy', 'Prasanna Balaprakash', 'Peihong Jiang'] | 2019-11-10 | null | null | null | null | ['retrosynthesis'] | ['medical'] | [ 3.97379994e-01 2.76411504e-01 -4.18694645e-01 1.30406231e-01
-4.00683224e-01 -1.11231554e+00 6.92269683e-01 4.94741589e-01
-4.33528125e-01 1.27996755e+00 -5.78754283e-02 -8.33037674e-01
2.17239425e-01 -8.62283945e-01 -7.46369958e-01 -7.71197915e-01
-1.91203475e-01 6.12689078e-01 5.77942468e-02 -2.92463392e-01
5.02653241e-01 8.35819066e-01 -8.46655905e-01 -2.34304997e-03
5.90987384e-01 5.17921031e-01 3.30369145e-01 8.95140529e-01
-2.19469503e-01 9.90970254e-01 -3.90265197e-01 -3.74555737e-01
3.24046314e-01 -7.41636693e-01 -1.23221028e+00 -2.56063521e-01
-3.36509615e-01 -1.60690784e-01 -1.73809510e-02 8.04041624e-01
2.79465765e-01 4.22470123e-01 8.89072359e-01 -1.00278556e+00
-9.63828713e-02 7.40476608e-01 -1.74614251e-01 8.93849507e-02
4.69557881e-01 2.97414124e-01 1.04539061e+00 -5.86146176e-01
7.80687094e-01 9.61398423e-01 3.29291940e-01 5.14514804e-01
-1.29356885e+00 -5.41596234e-01 5.16217239e-02 2.02335194e-01
-1.08479321e+00 -2.85418689e-01 4.96670365e-01 -4.38853562e-01
1.52443910e+00 -1.09643288e-01 9.55573320e-01 1.01376474e+00
5.46822011e-01 5.06544948e-01 9.08004522e-01 -4.32023972e-01
7.01026380e-01 -2.98204899e-01 -5.98488152e-01 7.49466896e-01
1.77464902e-01 3.91858906e-01 -3.67898464e-01 1.19820110e-01
5.93413651e-01 -8.81621987e-02 1.00520730e-01 -2.75881618e-01
-1.43843257e+00 1.04938018e+00 4.65458900e-01 2.13703841e-01
-5.32588303e-01 4.44334179e-01 2.75493085e-01 1.57651417e-02
-1.91398665e-01 1.36153936e+00 -8.48355234e-01 -1.41563430e-01
-8.52477849e-01 2.03281134e-01 1.24056864e+00 7.32004166e-01
6.68965578e-01 1.39439970e-01 4.16194141e-01 -1.43230304e-01
1.87670916e-01 -2.21037522e-01 -4.61228602e-02 -1.11114049e+00
6.71937317e-02 3.18395466e-01 3.76869410e-01 -3.91543210e-01
-5.54959536e-01 -2.26574950e-03 -5.31385422e-01 4.54317302e-01
9.33609545e-01 -5.76537192e-01 -8.73159409e-01 1.41546190e+00
4.72644448e-01 -4.88126308e-01 3.21115464e-01 7.42209136e-01
3.16875488e-01 1.27887547e+00 5.27323365e-01 -7.22774565e-01
9.86392319e-01 -1.03355277e+00 -5.84920645e-01 1.85747638e-01
5.62284350e-01 -8.66928279e-01 1.99058920e-01 8.85315120e-01
-1.38398886e+00 -5.45438379e-02 -1.28330541e+00 -8.47972408e-02
-7.47327328e-01 -4.60419655e-01 1.25174451e+00 6.19424582e-01
-5.98756850e-01 1.25812340e+00 -7.82058060e-01 -4.05420691e-01
2.38554180e-01 5.87408304e-01 -3.36058110e-01 -4.52203117e-03
-1.16381073e+00 1.27582049e+00 8.14139783e-01 6.55366033e-02
-1.44921029e+00 -7.04894006e-01 -5.33001482e-01 2.76781749e-02
1.09957969e+00 -6.71572089e-01 1.58813727e+00 -1.01058996e+00
-2.03019619e+00 3.25485319e-01 -1.25951111e-01 -2.69004703e-01
6.66761816e-01 3.49535048e-01 -1.32960722e-01 6.10714853e-02
-2.43371665e-01 9.02014732e-01 5.98166108e-01 -7.50285506e-01
-5.61922312e-01 -2.13122368e-02 2.09286839e-01 1.71113148e-01
6.39075041e-01 9.78082642e-02 3.05102915e-01 -2.85273552e-01
-6.80299774e-02 -8.27053905e-01 -7.48157084e-01 -7.88415298e-02
-3.72095406e-01 -5.16476870e-01 5.84791079e-02 -4.91443038e-01
4.14879858e-01 -1.31953537e+00 4.14125472e-01 1.85944423e-01
-3.19754216e-03 -2.23085582e-02 1.29748911e-01 1.03652000e+00
-2.56005675e-01 5.16621828e-01 4.99579832e-02 6.48230731e-01
-9.91643369e-02 -4.77044843e-03 -5.23012802e-02 2.09742352e-01
3.85037780e-01 7.62538612e-01 -1.37524545e+00 -1.55134276e-01
-1.83145832e-02 1.01650141e-01 -4.37605947e-01 -9.39483382e-03
-7.31503904e-01 5.04093051e-01 -5.18737853e-01 8.74238312e-01
6.66568577e-02 -1.37012467e-01 7.15651751e-01 -2.14269683e-02
-6.86514497e-01 4.46116418e-01 -1.02409649e+00 1.87108922e+00
-2.54545093e-01 2.12851420e-01 -2.48611525e-01 -9.59642172e-01
6.57763898e-01 4.92050201e-01 6.02226913e-01 -5.13111413e-01
7.06291124e-02 2.40403861e-01 4.75905508e-01 -1.94617540e-01
4.97529894e-01 -8.00399125e-01 1.12476267e-01 3.61389250e-01
8.40868428e-02 -4.32373583e-01 5.81794918e-01 2.44038198e-02
1.08263862e+00 5.49060643e-01 8.74827981e-01 -3.27254742e-01
1.83701649e-01 6.11519396e-01 6.56106293e-01 5.79622984e-01
-3.15253615e-01 1.26952166e-02 6.97358727e-01 -7.25833058e-01
-1.39216471e+00 -8.20020676e-01 3.82090151e-01 1.06524622e+00
-4.42995615e-02 -4.40214068e-01 -5.54654300e-01 -5.41671038e-01
-3.51188123e-01 6.97494924e-01 -2.61503130e-01 -9.58830025e-03
-2.98071682e-01 -7.00407386e-01 1.63156152e-01 5.31629980e-01
9.56063420e-02 -1.00313091e+00 -6.18228197e-01 7.51556098e-01
1.59713566e-01 -7.20866084e-01 -3.00304562e-01 9.48112667e-01
-6.19918525e-01 -1.28165698e+00 -6.30491674e-01 -6.07781947e-01
3.59634042e-01 4.32565175e-02 8.52718711e-01 -2.41696119e-01
-7.95572877e-01 -9.12210718e-02 -1.32803038e-01 -7.60394514e-01
-8.22682500e-01 1.21427909e-01 -1.67507246e-01 -7.91203141e-01
-4.34463359e-02 -2.93098599e-01 -7.52686024e-01 8.06694701e-02
-4.20737863e-01 2.18291089e-01 5.62558353e-01 7.98245668e-01
7.09477544e-01 4.42193657e-01 5.32917321e-01 -4.91329044e-01
3.85286629e-01 -4.44164544e-01 -8.82336497e-01 2.46249348e-01
-6.11379862e-01 5.51772535e-01 8.58707845e-01 -3.74015719e-01
-1.04116344e+00 6.75281703e-01 -9.03721247e-03 3.03069204e-01
-2.06881255e-01 6.20806217e-01 -4.05513883e-01 4.29728590e-02
4.29029733e-01 6.51570261e-02 -8.57339948e-02 -2.41366569e-02
4.32744712e-01 -3.44705790e-01 -9.05006677e-02 -9.62782562e-01
4.48994219e-01 1.15855433e-01 8.09910893e-01 -7.93459952e-01
-4.43253428e-01 -1.99093089e-01 -7.58866251e-01 -8.16341210e-03
9.98179972e-01 -6.97300553e-01 -1.31284535e+00 -3.09081256e-01
-1.26158237e+00 -6.93799198e-01 -3.53764027e-01 4.33440834e-01
-8.15924406e-01 2.11684033e-01 -3.59470248e-01 -5.92452228e-01
-7.25409016e-02 -1.34927845e+00 5.19821584e-01 2.31997699e-01
-4.55152065e-01 -9.60063338e-01 2.01167002e-01 1.31382212e-01
-3.79086100e-02 4.89329785e-01 1.12617147e+00 -5.92509449e-01
-9.33193564e-01 -1.36927754e-01 2.37748682e-01 -2.54284441e-01
3.31607848e-01 4.37821299e-01 -7.11217523e-01 -1.12676844e-02
-5.89728355e-01 -5.07821620e-01 8.22156131e-01 3.53343606e-01
1.07329369e+00 -1.11279733e-01 -3.66116017e-01 1.44043460e-01
1.32413566e+00 8.99013698e-01 4.92238611e-01 4.08905685e-01
4.58308041e-01 7.35280275e-01 5.85041583e-01 3.68752718e-01
8.20406079e-02 1.95951954e-01 6.40439034e-01 1.08937845e-02
2.37822831e-01 -4.33466911e-01 2.45906159e-01 1.30814418e-01
-7.02383935e-01 -4.67551947e-01 -9.36320066e-01 9.79375318e-02
-1.47351718e+00 -1.30466509e+00 9.81665850e-02 2.03726101e+00
9.48681772e-01 3.05854268e-02 4.50397164e-01 3.94849867e-01
3.88954759e-01 -1.77056089e-01 -7.82278657e-01 -9.52484310e-01
2.47435346e-01 4.95242566e-01 7.23093808e-01 6.08469963e-01
-9.19254601e-01 1.27483487e+00 7.43340683e+00 4.93049622e-01
-1.17231464e+00 -2.85219342e-01 4.89523202e-01 -9.42517519e-02
-5.09335622e-02 5.80981076e-01 -6.39159441e-01 5.03387339e-02
9.45758283e-01 -3.37488145e-01 9.34081852e-01 6.89567983e-01
5.07133782e-01 -3.43598783e-01 -1.62908137e+00 4.94724363e-01
-7.42521465e-01 -1.65548003e+00 -5.63391969e-02 2.10112706e-01
6.87854588e-01 -4.95120287e-01 -3.13528448e-01 1.07311279e-01
7.41895556e-01 -1.25803077e+00 7.38754928e-01 2.64625937e-01
4.84082133e-01 -1.07013047e+00 1.43457264e-01 2.88211823e-01
-1.08415270e+00 -1.37499824e-01 -3.35624546e-01 -3.31314117e-01
-2.49974262e-02 3.22850406e-01 -1.41448152e+00 6.48428738e-01
-6.23446750e-03 5.39133966e-01 1.47618309e-01 9.43053007e-01
-4.77292001e-01 3.75781983e-01 -2.94660300e-01 -5.33582091e-01
5.70192039e-01 -3.49717826e-01 4.22109812e-02 1.08356023e+00
3.15228164e-01 3.98651332e-01 2.38767877e-01 1.01099133e+00
-1.00436293e-01 9.79707092e-02 -5.78295052e-01 -7.14591980e-01
2.12275356e-01 1.13301075e+00 -1.25727892e+00 -1.62849665e-01
-2.73483366e-01 8.11113119e-01 6.23703338e-02 3.28352869e-01
-6.22421324e-01 -3.50727439e-01 5.64740837e-01 -1.20623164e-01
3.14911127e-01 -4.90512490e-01 1.35987729e-01 -7.46894896e-01
-6.69613481e-01 -8.59435558e-01 3.45246375e-01 -5.98395884e-01
-7.18781292e-01 -1.31987974e-01 -3.12449545e-01 -4.79390502e-01
-2.11360022e-01 -8.26827526e-01 -7.61091113e-01 7.46967912e-01
-1.50588191e+00 -7.55211592e-01 2.42748857e-01 3.30760747e-01
9.01980937e-01 -4.67117243e-02 9.54697728e-01 -3.45429718e-01
-8.64057958e-01 -4.99319518e-03 1.36135295e-01 -5.36754906e-01
5.72671235e-01 -1.51484644e+00 1.53781176e-01 7.78201997e-01
-1.67938188e-01 4.82957363e-01 8.76908302e-01 -6.42763138e-01
-1.98589170e+00 -8.58075142e-01 6.87864780e-01 -1.44687548e-01
7.03637838e-01 -2.60356784e-01 -2.66744077e-01 4.36661631e-01
4.15401965e-01 -3.32180917e-01 7.46559858e-01 -2.71547548e-02
-9.41712260e-02 2.75650740e-01 -9.24213350e-01 7.33409703e-01
8.59630644e-01 -1.41653389e-01 -2.62506932e-01 6.38739347e-01
5.69790900e-01 -2.00854629e-01 -9.71962512e-01 -1.84250921e-01
5.99602163e-01 -4.09643441e-01 1.00251973e+00 -9.45234776e-01
7.46342480e-01 -2.81242102e-01 8.18539336e-02 -1.30738986e+00
-3.27485472e-01 -1.26160324e+00 -1.05733603e-01 5.91214001e-01
6.76560104e-01 -1.92916125e-01 8.49654377e-01 8.30508649e-01
-9.95639712e-02 -5.30765116e-01 -5.18805683e-01 -8.88506055e-01
5.90405047e-01 3.81193832e-02 6.47005558e-01 9.60611880e-01
4.33098823e-01 3.97922367e-01 -1.66361794e-01 2.70057023e-02
4.88660216e-01 1.60992146e-01 4.15035576e-01 -1.15519309e+00
-2.15506658e-01 -6.57096446e-01 9.98511240e-02 -6.12514019e-01
1.73794970e-01 -9.92924809e-01 1.82949305e-01 -1.62114584e+00
1.87118426e-01 -2.00603038e-01 -1.55825362e-01 5.14233232e-01
1.81213185e-01 -4.45032418e-01 8.94784257e-02 -1.85512975e-01
-4.12386775e-01 3.66693825e-01 1.49685192e+00 -4.33126420e-01
-2.23433897e-01 -9.29609165e-02 -8.10615301e-01 6.00985765e-01
9.72642720e-01 -5.96837342e-01 -4.32359666e-01 1.97689250e-01
8.24285924e-01 5.40841877e-01 -1.02150273e-02 -5.91263294e-01
2.92208880e-01 -8.63358438e-01 5.43673277e-01 -4.29848462e-01
3.36484939e-01 -6.90613210e-01 2.94329703e-01 9.97929513e-01
-4.40133870e-01 -1.59509346e-01 1.23580620e-01 6.11688673e-01
3.12621117e-01 -4.59986657e-01 6.15439177e-01 -8.67094517e-01
-8.01990151e-01 3.38283569e-01 -1.19070446e+00 -4.52019453e-01
1.30152106e+00 -2.37449571e-01 9.34056414e-04 -2.38193460e-02
-1.02289796e+00 1.15674138e-01 4.11841154e-01 -5.32344058e-02
3.40663075e-01 -6.64228857e-01 -1.45186424e-01 -5.35857260e-01
-2.07081586e-01 1.24683134e-01 -3.39441031e-01 5.13174176e-01
-9.89175558e-01 6.75317585e-01 -4.60945606e-01 3.80319208e-02
-9.81549978e-01 9.86720443e-01 5.30204296e-01 -2.69187182e-01
-1.02618903e-01 8.33826423e-01 -4.25926521e-02 -1.90541372e-02
9.42505300e-02 -5.60884118e-01 1.05045192e-01 2.53619075e-01
2.94095337e-01 5.52361548e-01 -3.94825116e-02 2.12138072e-01
-3.22859138e-01 2.01804459e-01 -2.58412868e-01 -2.75426861e-02
1.22502470e+00 3.43915135e-01 7.32366815e-02 1.17052130e-01
5.78157365e-01 -4.18698430e-01 -1.35144615e+00 3.02700251e-01
1.96248129e-01 1.10988520e-01 3.23581845e-01 -1.10486054e+00
-5.31518996e-01 1.03449368e+00 7.43231028e-02 -3.60934958e-02
5.24899960e-01 -3.81676793e-01 5.53697169e-01 1.13071620e+00
3.47288966e-01 -1.22525549e+00 4.36731011e-01 4.41861093e-01
5.55532694e-01 -1.53593826e+00 3.09899598e-01 -2.72567779e-01
-4.72698659e-01 1.60108495e+00 2.62730151e-01 2.20179453e-01
2.41055712e-01 1.12151690e-01 -6.27882361e-01 1.03119209e-01
-8.00368965e-01 -2.47427747e-01 -3.97565752e-01 4.57742512e-01
6.81274891e-01 1.90425385e-02 -3.53292018e-01 1.73677489e-01
6.91387653e-02 1.04850776e-01 8.28946710e-01 1.32979810e+00
-5.13255894e-01 -1.40756869e+00 -2.28797734e-01 -2.16507465e-01
-5.18892407e-01 -1.29812375e-01 -7.84042358e-01 8.47711802e-01
1.47263438e-01 1.00897825e+00 -3.33863735e-01 1.68248758e-01
2.29305048e-02 4.27534819e-01 1.19247591e+00 -5.36649942e-01
-5.71639717e-01 1.87875517e-02 2.99301356e-01 -5.98388255e-01
-3.60467345e-01 -6.75586581e-01 -1.63975132e+00 -5.20023167e-01
-2.85308987e-01 4.11572248e-01 1.06851673e+00 1.06390524e+00
-3.57040856e-03 6.94331348e-01 5.71901143e-01 -1.09148741e+00
-3.22307467e-01 -2.06745222e-01 -4.11797822e-01 -3.47929657e-01
2.61933982e-01 -4.87025231e-01 1.59697309e-01 2.76925564e-01] | [4.49893045425415, 6.09076452255249] |
ca1f8804-3cd0-4151-a645-bb8d5b94e5b1 | cross-lingual-transfer-for-unsupervised | null | null | https://aclanthology.org/K15-1012 | https://aclanthology.org/K15-1012.pdf | Cross-lingual Transfer for Unsupervised Dependency Parsing Without Parallel Data | null | ['Steven Bird', 'Long Duong', 'Trevor Cohn', 'Paul Cook'] | 2015-07-01 | null | null | null | conll-2015-7 | ['unsupervised-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.350193977355957, 3.6745247840881348] |
8ed4c691-6372-4693-9ad2-0db3b6711c3f | m2fpa-a-multi-yaw-multi-pitch-high-quality-1 | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Li_M2FPA_A_Multi-Yaw_Multi-Pitch_High-Quality_Dataset_and_Benchmark_for_Facial_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Li_M2FPA_A_Multi-Yaw_Multi-Pitch_High-Quality_Dataset_and_Benchmark_for_Facial_ICCV_2019_paper.pdf | M2FPA: A Multi-Yaw Multi-Pitch High-Quality Dataset and Benchmark for Facial Pose Analysis | Facial images in surveillance or mobile scenarios often have large view-point variations in terms of pitch and yaw angles. These jointly occurred angle variations make face recognition challenging. Current public face databases mainly consider the case of yaw variations. In this paper, a new large-scale Multi-yaw Multi-pitch high-quality database is proposed for Facial Pose Analysis (M2FPA), including face frontalization, face rotation, facial pose estimation and pose-invariant face recognition. It contains 397,544 images of 229 subjects with yaw, pitch, attribute, illumination and accessory. M2FPA is the most comprehensive multi-view face database for facial pose analysis. Further, we provide an effective benchmark for face frontalization and pose-invariant face recognition on M2FPA with several state-of-the-art methods, including DR-GAN, TP-GAN and CAPG-GAN. We believe that the new database and benchmark can significantly push forward the advance of facial pose analysis in real-world applications. Moreover, a simple yet effective parsing guided discriminator is introduced to capture the local consistency during GAN optimization. Extensive quantitative and qualitative results on M2FPA and Multi-PIE demonstrate the superiority of our face frontalization method. Baseline results for both face synthesis and face recognition from state-of-the-art methods demonstrate the challenge offered by this new database.
| ['Pei-Pei Li', ' Zhenan Sun', ' Ran He', ' Yibo Hu', ' Xiang Wu'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['robust-face-recognition'] | ['computer-vision'] | [-1.84619024e-01 -5.80075011e-02 -1.43539719e-02 -7.01396108e-01
-8.65897655e-01 -4.70419347e-01 3.69956702e-01 -1.30673349e+00
3.52902651e-01 3.68341953e-01 1.90235913e-01 3.60695660e-01
1.94461837e-01 -5.49128592e-01 -6.20567381e-01 -1.17760694e+00
1.92330718e-01 3.39442611e-01 -5.83916128e-01 -2.40963563e-01
-2.83357620e-01 1.15621746e+00 -1.75741029e+00 8.02557841e-02
2.27118894e-01 1.26338160e+00 -3.52016419e-01 2.28450194e-01
5.11482298e-01 -2.89102614e-01 -4.96618956e-01 -9.90160167e-01
4.99937087e-01 -2.22962677e-01 -2.35241517e-01 5.11454523e-01
1.12324917e+00 -5.30499458e-01 -2.29675174e-01 7.62897849e-01
1.04558218e+00 -2.42521405e-01 4.88878816e-01 -1.41601777e+00
-3.16693813e-01 -1.66577220e-01 -9.69799697e-01 -2.61812925e-01
5.39020419e-01 3.85902703e-01 3.08990419e-01 -1.47208965e+00
6.88811660e-01 1.77745676e+00 7.79914021e-01 8.67654383e-01
-8.18231702e-01 -1.07714510e+00 1.76480204e-01 -2.99971774e-02
-1.56130004e+00 -1.02122402e+00 8.61570418e-01 -3.04818302e-01
5.39477289e-01 2.19933763e-01 6.35949969e-01 1.29999006e+00
1.72344834e-01 2.99144268e-01 1.05891633e+00 -1.71230182e-01
-3.04781586e-01 -3.48415047e-01 -5.27806878e-01 1.15229928e+00
9.24585387e-02 1.48804560e-01 -7.53670335e-01 -1.67126730e-02
9.07643735e-01 -8.60196650e-02 -3.26645553e-01 -1.50478944e-01
-8.23531389e-01 5.79651952e-01 2.52082869e-02 -2.67857283e-01
-2.60129750e-01 -5.20190857e-02 2.62223512e-01 1.25248194e-01
6.67186737e-01 1.30601767e-02 -4.86683071e-01 9.75733548e-02
-9.08977211e-01 2.93238938e-01 4.58316773e-01 1.07296622e+00
5.61085641e-01 4.23900783e-01 -2.79072762e-01 9.42615330e-01
7.66492486e-01 1.27400088e+00 -5.85937761e-02 -9.16050553e-01
3.40613604e-01 2.47006416e-01 -2.95116067e-01 -1.25783753e+00
-4.69999790e-01 -5.05336486e-02 -7.93389440e-01 -5.80794504e-03
2.64528126e-01 -2.13296846e-01 -9.68534291e-01 1.88015592e+00
7.72286355e-01 -8.46926570e-02 -1.50477827e-01 7.68626809e-01
1.21162844e+00 4.39619720e-01 -3.56802523e-01 -6.85239971e-01
1.77695334e+00 -8.19606423e-01 -9.59204018e-01 -1.09931558e-01
-7.70231634e-02 -1.10268366e+00 7.26258337e-01 4.20261592e-01
-1.11588681e+00 -5.35698414e-01 -7.01885045e-01 1.36401877e-01
5.14042228e-02 6.01749778e-01 4.77686226e-01 1.31133389e+00
-1.04700351e+00 -6.57642111e-02 -5.36110103e-01 -3.24420333e-01
7.97722697e-01 5.10465205e-01 -8.08079362e-01 -3.03372473e-01
-7.38794029e-01 5.87550461e-01 -4.31925893e-01 5.41035712e-01
-1.11834562e+00 -8.17346215e-01 -9.29541290e-01 -3.49543184e-01
4.08896536e-01 -5.79562724e-01 9.28793132e-01 -6.93470597e-01
-1.94813633e+00 1.27229035e+00 -5.32211304e-01 4.15864617e-01
4.02827978e-01 -2.29467019e-01 -7.42683172e-01 2.23515734e-01
-9.56842974e-02 7.11450040e-01 1.41719604e+00 -1.14424133e+00
1.10954583e-01 -1.09184706e+00 -4.42987949e-01 1.20736539e-01
-3.16814572e-01 4.45408881e-01 -8.51620197e-01 -7.00716615e-01
3.05991955e-02 -1.15961194e+00 4.97723907e-01 2.89374530e-01
-4.04875189e-01 -8.27788487e-02 1.23474026e+00 -7.79663444e-01
8.00293386e-01 -2.03932714e+00 1.26065508e-01 1.54538333e-01
4.27883193e-02 2.88869411e-01 -3.50345105e-01 9.46718603e-02
-2.27725729e-01 -6.25936091e-02 2.21475780e-01 -5.70751607e-01
-1.62568480e-01 3.20359804e-02 -1.87887564e-01 8.46565664e-01
2.18832165e-01 8.92222226e-01 -1.87545940e-01 -4.36141908e-01
1.25999421e-01 9.01878178e-01 -9.07206118e-01 2.63760120e-01
1.90597430e-01 6.25484049e-01 -3.71692985e-01 1.45050752e+00
1.44624746e+00 2.10046098e-01 2.14615270e-01 -7.91669011e-01
2.08163336e-01 -4.53457624e-01 -9.68421757e-01 1.50526416e+00
-3.69525939e-01 4.84899789e-01 5.14386833e-01 -2.64277905e-01
1.06291866e+00 4.34773535e-01 6.23155832e-01 -4.23731625e-01
3.80412668e-01 1.20914541e-01 -3.17925483e-01 -5.04683793e-01
9.09077078e-02 -1.60182476e-01 1.81732729e-01 -1.18082225e-01
4.25751776e-01 -2.32800320e-01 -1.34466752e-01 -5.72651863e-01
2.46845081e-01 3.86625864e-02 5.40659167e-02 -3.35169733e-01
7.89255679e-01 -1.06211686e+00 7.47794807e-01 -1.60984993e-01
-3.08473110e-01 9.02776361e-01 5.32609701e-01 -5.15890837e-01
-7.60138988e-01 -1.00366521e+00 -4.70865965e-01 8.34976435e-01
-2.17737645e-01 -4.57541615e-01 -1.14145982e+00 -7.02460766e-01
6.13399548e-03 -9.96749774e-02 -6.73908830e-01 2.98808329e-02
-7.41650760e-01 -9.83088017e-01 6.46405101e-01 4.08445001e-01
8.53695571e-01 -8.33029985e-01 -5.00163175e-02 -4.68668491e-01
-6.97066933e-02 -1.40174258e+00 -9.30326164e-01 -7.52937853e-01
-5.21995842e-01 -1.24820840e+00 -7.65514195e-01 -5.67230940e-01
9.22544181e-01 3.29342149e-02 1.10768747e+00 -2.22058505e-01
-4.38634694e-01 5.17078161e-01 -9.31921601e-02 -3.08484197e-01
-2.30850074e-02 -3.59697074e-01 4.79270339e-01 4.54432487e-01
7.84884319e-02 -3.54070663e-01 -8.29495370e-01 1.02357984e+00
-4.55228478e-01 -2.27386579e-01 3.46142024e-01 8.58676910e-01
4.98333722e-01 -2.47236073e-01 5.62152863e-01 -4.37212616e-01
1.45386919e-01 -8.78606960e-02 -7.37574697e-01 3.13895971e-01
-4.31114614e-01 -4.83790666e-01 2.33872995e-01 -7.61671811e-02
-1.26281524e+00 3.34160291e-02 -4.70122069e-01 -6.45873785e-01
-8.20966586e-02 -6.39290735e-02 -1.04330182e+00 -5.07348359e-01
3.23403835e-01 2.69124173e-02 4.45908487e-01 -2.43991494e-01
2.69576162e-01 4.87096667e-01 4.66500700e-01 -5.42841196e-01
8.69027376e-01 4.77323681e-01 3.91635984e-01 -1.12071288e+00
-4.45221454e-01 5.83771691e-02 -5.52877188e-01 -5.68471313e-01
8.44303966e-01 -1.17258596e+00 -1.19042897e+00 1.09370840e+00
-1.09388411e+00 1.78471386e-01 1.81440443e-01 2.60871559e-01
-4.21842426e-01 1.55611485e-01 -3.20529699e-01 -7.59869695e-01
-4.38720554e-01 -1.59703028e+00 1.69665849e+00 4.10366833e-01
3.22881848e-01 -6.21715069e-01 -2.19343036e-01 7.64957905e-01
3.23700726e-01 3.89020622e-01 4.02829766e-01 1.26695529e-01
-3.24805349e-01 -7.03094676e-02 5.79769574e-02 3.77455115e-01
4.06346500e-01 4.93361205e-01 -1.31215882e+00 -7.06615746e-01
-4.61099856e-02 -4.51997429e-01 5.11656821e-01 5.38086176e-01
1.29152954e+00 -4.47398007e-01 -1.49218440e-01 1.27410495e+00
9.96952415e-01 2.93797910e-01 6.91378474e-01 -3.42775375e-01
8.92034352e-01 6.49716496e-01 6.39294326e-01 6.14139497e-01
2.88401335e-01 1.16109872e+00 5.47866821e-01 -1.04042895e-01
-2.01671705e-01 -1.34580076e-01 6.51735842e-01 6.27874672e-01
-3.56028825e-01 -1.29231274e-01 -7.13380754e-01 -2.06276663e-02
-1.13722217e+00 -9.43875670e-01 5.45318544e-01 2.01058912e+00
6.30284190e-01 -7.50479639e-01 2.08193421e-01 -5.61836585e-02
7.12259948e-01 2.86444306e-01 -4.07357961e-01 -1.90559365e-02
-1.98251292e-01 2.49348268e-01 1.38118654e-01 2.89151251e-01
-1.14458346e+00 8.26342106e-01 6.42444897e+00 9.71759379e-01
-1.48360419e+00 -3.04567721e-03 1.08250630e+00 -3.10004354e-01
-8.22027400e-02 -7.67638743e-01 -1.45797718e+00 2.87545472e-01
6.85908020e-01 3.56914371e-01 4.19862062e-01 9.15699244e-01
-1.35077043e-02 2.89272457e-01 -1.01964438e+00 1.59397078e+00
6.62837386e-01 -1.16996670e+00 6.85381442e-02 2.38011360e-01
8.31821382e-01 -4.46904838e-01 6.87263429e-01 1.15589716e-01
-5.32892227e-01 -1.39470422e+00 4.38672692e-01 5.18502355e-01
1.51378226e+00 -1.01943707e+00 5.86062014e-01 -3.79594803e-01
-1.31114841e+00 -3.77600710e-03 -1.40824929e-01 6.89588249e-01
2.60380376e-02 3.68976295e-01 -6.17273748e-01 7.00191081e-01
7.52861202e-01 7.85194218e-01 -6.47702217e-01 2.59136558e-01
-2.55939737e-02 2.48131573e-01 -3.72285485e-01 4.27191794e-01
-3.20636004e-01 -2.54977822e-01 6.35354996e-01 8.20700645e-01
6.03909254e-01 6.80478737e-02 -1.47051573e-01 6.37335956e-01
-3.49982440e-01 1.12501178e-02 -6.33813560e-01 5.19157462e-02
5.35574138e-01 1.60663819e+00 -3.93529207e-01 1.53768331e-01
-3.56246084e-01 5.65039456e-01 -3.35427254e-01 2.38477960e-01
-1.00098228e+00 3.60411078e-01 1.20698321e+00 3.44446570e-01
3.86216432e-01 -9.40581113e-02 1.72446519e-01 -1.21819210e+00
2.92162746e-01 -1.38704336e+00 1.54501900e-01 -5.52301526e-01
-1.06906438e+00 9.18528378e-01 2.10958168e-01 -1.05744874e+00
-3.63035232e-01 -9.68291581e-01 -4.54756260e-01 7.50663877e-01
-1.40657294e+00 -1.75010145e+00 -7.38829017e-01 9.50702012e-01
5.40114880e-01 -8.45245123e-01 9.70834196e-01 4.85319346e-01
-9.51683879e-01 1.38422215e+00 -3.07129443e-01 7.94321969e-02
9.58382368e-01 -3.64851832e-01 2.83491731e-01 6.98774755e-01
-1.49732545e-01 5.81684887e-01 3.78879040e-01 -4.10948157e-01
-2.11036706e+00 -1.19433558e+00 2.43288442e-01 -5.69025040e-01
-1.74031541e-01 -5.93255997e-01 -3.53251398e-01 5.75136065e-01
5.38964570e-02 5.76212704e-01 6.28537238e-01 -5.01761734e-02
-5.22366405e-01 -7.64190614e-01 -1.59718359e+00 4.45543766e-01
1.14723289e+00 -4.78953898e-01 1.22411534e-01 4.13891524e-01
2.02084512e-01 -6.94503129e-01 -1.05610394e+00 8.56865883e-01
9.21143651e-01 -1.02020288e+00 1.08254969e+00 -2.13463098e-01
1.78811684e-01 -1.63311496e-01 -3.49109501e-01 -1.21566236e+00
4.39448804e-02 -1.03313994e+00 1.84040703e-02 1.56158328e+00
-5.19698970e-02 -7.69231379e-01 9.11519825e-01 2.29018331e-01
-4.30054329e-02 -9.12129521e-01 -1.18895471e+00 -6.49608433e-01
-1.01589493e-01 -7.76723120e-03 9.79898810e-01 7.31028259e-01
-6.53946757e-01 -8.88145640e-02 -5.44281840e-01 2.41933480e-01
7.10123181e-01 -2.96162954e-03 1.03550506e+00 -9.61665988e-01
1.53337911e-01 -2.33928770e-01 -5.41440189e-01 -6.90886438e-01
4.23239410e-01 -4.95222181e-01 -2.61729032e-01 -6.33588254e-01
1.08464614e-01 2.22423170e-02 4.55106944e-01 5.25755644e-01
1.31600201e-01 6.42673969e-01 1.21450990e-01 -8.46660286e-02
5.19222990e-02 8.94431770e-01 1.64534032e+00 9.56717692e-03
2.63913929e-01 -6.73200712e-02 -5.22025585e-01 7.68936396e-01
4.19184178e-01 -2.39169281e-02 -4.27000225e-01 -1.86913148e-01
-7.52931461e-02 3.20326895e-01 3.33790153e-01 -7.13932514e-01
-1.35249525e-01 -2.16594547e-01 8.63916874e-01 -5.72830975e-01
9.55032587e-01 -7.42642820e-01 3.29017818e-01 2.64725298e-01
3.42471063e-01 3.62913013e-01 4.15120423e-01 2.22987503e-01
-3.53229135e-01 5.29540002e-01 1.21659005e+00 1.31609932e-01
-2.99474210e-01 9.84454274e-01 1.86449453e-01 8.77452791e-02
1.22586536e+00 -2.51755357e-01 -4.76765901e-01 -3.73067081e-01
-5.79501510e-01 -1.74056530e-01 3.82318646e-01 5.29738843e-01
8.61642420e-01 -1.68637395e+00 -9.84833300e-01 1.03673458e+00
1.13880828e-01 -4.73208874e-02 5.35261035e-01 1.01523924e+00
-4.95806038e-01 4.52751696e-01 -6.94182694e-01 -7.02253640e-01
-1.86547649e+00 1.17368713e-01 5.46998560e-01 2.50252783e-01
-1.75856829e-01 1.09421945e+00 4.65169668e-01 -5.21196306e-01
-4.09467220e-02 5.77854589e-02 -1.17547862e-01 3.02704602e-01
5.67454636e-01 3.34176749e-01 2.93986708e-01 -1.19965649e+00
-7.19730496e-01 1.25781465e+00 5.05399294e-02 1.50874272e-01
1.32873333e+00 8.97680372e-02 -2.16030210e-01 -3.49937111e-01
1.31695819e+00 2.04736277e-01 -1.40191579e+00 2.72619665e-01
-9.11239505e-01 -9.23250377e-01 -1.67535022e-01 -5.89672625e-01
-1.91230571e+00 8.18791211e-01 8.08656096e-01 -6.37301028e-01
1.49230611e+00 -1.69647828e-01 3.96145999e-01 1.28494740e-01
4.31861848e-01 -9.18886781e-01 3.82905871e-01 3.85761023e-01
1.58871722e+00 -1.24750316e+00 1.50655620e-02 -8.95711243e-01
-4.84259188e-01 1.31123006e+00 1.00264740e+00 2.62135059e-01
8.69961441e-01 4.41412657e-01 8.69977549e-02 -2.63080180e-01
-5.01796901e-01 4.30700123e-01 6.27515376e-01 7.51049340e-01
4.53245521e-01 -5.21547012e-02 2.06059605e-01 4.89769697e-01
-5.01874387e-01 -3.73132825e-01 -1.63974054e-02 3.63220096e-01
2.30109751e-01 -8.86424482e-01 -7.64369249e-01 1.88026763e-02
-5.47015071e-01 2.07170561e-01 -2.24958703e-01 9.79261220e-01
1.26128197e-01 7.14471102e-01 3.12649421e-02 -3.88735622e-01
3.74815524e-01 5.57613447e-02 1.09821165e+00 -2.85783440e-01
-4.40246433e-01 3.66396517e-01 1.53340986e-02 -9.00419176e-01
-3.05570960e-01 -7.69871235e-01 -5.96893609e-01 -6.52387440e-01
-1.17687792e-01 -4.11036402e-01 7.58260846e-01 5.18692434e-01
5.19231141e-01 2.30251774e-01 1.10835564e+00 -1.17229307e+00
-5.66036522e-01 -9.70829785e-01 -6.39750302e-01 1.70831606e-01
3.34671885e-01 -1.09425759e+00 -2.90339023e-01 -7.42083509e-03] | [13.18496036529541, 0.3313773274421692] |
2f236501-e9a4-4e8d-90da-2eb4a628e29d | lite-mdetr-a-lightweight-multi-modal-detector | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Lou_Lite-MDETR_A_Lightweight_Multi-Modal_Detector_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Lou_Lite-MDETR_A_Lightweight_Multi-Modal_Detector_CVPR_2022_paper.pdf | Lite-MDETR: A Lightweight Multi-Modal Detector | Recent multi-modal detectors based on transformers and modality encoders have successfully achieved impressive results on end-to-end visual object detection conditioned on a raw text query. However, they require a large model size and an enormous amount of computations to achieve high performance, which makes it difficult to deploy mobile applications that are limited by tight hardware resources. In this paper, we present a Lightweight modulated detector, Lite-MDETR, to facilitate efficient end-to-end multi-modal understanding on mobile devices. The key primitive is that Dictionary-Lookup-Transformormations (DLT) is proposed to replace Linear Transformation (LT) in multi-modal detectors where each weight in Linear Transformation (LT) is approximately factorized into a smaller dictionary, index, and coefficient. This way, the enormous linear projection with weights is converted into lite linear projection with dictionaries, a few lookups and scalings with indices and coefficients. DLT can be directly applied to pre-trained detectors, removing the need to perform expensive training from scratch. To tackle the challenging training of DLT due to the non-differentiable index, we convert the index and coefficient into a sparse matrix, train this sparse matrix during the fine-tuning phase, and recover it back to index and coefficient during the inference phase. Extensive experiments on several tasks such as phrase grounding, referring expression comprehension and segmentation show that our Lite-MDETR achieves similar detection accuracy to the prior multi-modal detectors with ~ 4.1xmodel size reduction. | ['Hongxia Jin', 'Yilin Shen', 'Ting Hua', 'Burak Uzkent', 'Yen-Chang Hsu', 'Qian Lou'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['phrase-grounding'] | ['natural-language-processing'] | [ 3.31250697e-01 -6.85625300e-02 -4.68124688e-01 -2.33567506e-01
-1.09606171e+00 -5.57043731e-01 3.23202729e-01 3.05445362e-02
-6.57934725e-01 5.51105514e-02 1.05111562e-01 -4.15711790e-01
3.86930764e-01 -5.30341625e-01 -9.15300369e-01 -4.85871762e-01
5.49088955e-01 6.04387045e-01 2.59864300e-01 -7.03890026e-02
-2.40768753e-02 2.44298987e-02 -1.49220657e+00 4.83712941e-01
6.99353874e-01 1.14531338e+00 5.87136686e-01 5.67836523e-01
-1.69729143e-01 5.98943591e-01 -3.41021538e-01 -5.24120092e-01
7.61758983e-02 -2.40913019e-01 -4.95592356e-01 4.92965151e-03
5.20885110e-01 -6.85749054e-01 -4.58729208e-01 1.21796095e+00
6.52690947e-01 5.89389959e-03 5.80320358e-01 -1.02995884e+00
-6.67634428e-01 6.65273488e-01 -9.04388249e-01 1.01819545e-01
5.52850962e-01 -3.40749025e-02 1.22042704e+00 -1.40632451e+00
5.88530958e-01 1.51167965e+00 6.04658365e-01 5.66503823e-01
-1.07061243e+00 -7.45443583e-01 3.40134621e-01 1.59367517e-01
-1.57156730e+00 -4.65414822e-01 4.77592111e-01 -2.29034767e-01
1.20779836e+00 2.64192581e-01 5.21863222e-01 9.91190076e-01
-2.82956183e-01 1.28496754e+00 6.19860411e-01 -4.59407538e-01
-5.95482476e-02 3.07219118e-01 -1.60766635e-02 1.17261624e+00
-1.42931310e-03 -2.26776019e-01 -6.48158431e-01 4.70675528e-03
7.05205858e-01 1.52113840e-01 -1.18418977e-01 -1.26063675e-01
-1.24697316e+00 8.56610179e-01 4.98904467e-01 6.79528117e-02
-3.15400273e-01 1.72338441e-01 4.72214341e-01 1.00226030e-01
-9.12529677e-02 -1.20516635e-01 -3.13479811e-01 -3.91129330e-02
-9.33076620e-01 -1.94410086e-02 5.18610477e-01 1.15173686e+00
7.64974236e-01 -3.79594304e-02 -1.85300946e-01 1.02372897e+00
5.11970043e-01 9.93401825e-01 6.84316874e-01 -4.93564039e-01
8.68141353e-01 7.86431253e-01 -3.15919578e-01 -9.43427980e-01
-4.27763194e-01 -2.14832038e-01 -9.95749891e-01 -3.83977294e-01
1.60665020e-01 -3.53461355e-02 -1.08106422e+00 1.74903619e+00
3.29362690e-01 -3.96394171e-02 7.15200417e-03 9.81531560e-01
9.40502584e-01 7.99041927e-01 3.21388096e-01 -1.05560303e-01
1.79507005e+00 -8.59409213e-01 -5.09246051e-01 -6.59183800e-01
8.94356251e-01 -7.91065156e-01 1.35641241e+00 2.00403392e-01
-9.92505670e-01 -5.16304433e-01 -9.77406740e-01 -5.61042428e-01
-1.76973879e-01 5.30351639e-01 7.79621959e-01 4.09402400e-01
-6.43657148e-01 -7.32702389e-02 -7.59696126e-01 -2.57685006e-01
4.53414530e-01 4.35350448e-01 -2.97934949e-01 -1.99231416e-01
-1.06272900e+00 7.33103693e-01 5.23419619e-01 -1.20340288e-01
-8.63792241e-01 -6.19930506e-01 -9.87087071e-01 1.11291289e-01
6.24580204e-01 -8.47419202e-01 1.27809286e+00 -4.07656938e-01
-1.29006171e+00 9.91351187e-01 -4.91082489e-01 -3.68733525e-01
1.11368567e-01 -2.36820638e-01 -3.15230966e-01 2.00769305e-01
3.94197613e-01 8.87404919e-01 1.14284468e+00 -7.39709198e-01
-1.02328813e+00 -3.45696628e-01 2.24766731e-01 5.49327493e-01
-6.90412760e-01 6.03553541e-02 -1.15794218e+00 -6.27598226e-01
4.66069549e-01 -9.27186668e-01 9.66541395e-02 -7.67963147e-03
-4.65303808e-01 -3.01787823e-01 8.74621809e-01 -5.96880198e-01
1.40713608e+00 -2.49506259e+00 3.55456829e-01 5.33292666e-02
3.61616462e-01 1.66875109e-01 -2.05576897e-01 -4.38659026e-06
1.58386424e-01 -1.95959017e-01 -8.23294148e-02 -4.97023553e-01
2.38319203e-01 1.63467571e-01 -4.21294570e-01 2.47218132e-01
-9.86996386e-03 1.08916903e+00 -7.74373412e-01 -7.89140046e-01
3.44744921e-01 3.66927177e-01 -7.85185397e-01 6.59697130e-02
-3.20201248e-01 1.04789529e-02 -4.50922757e-01 1.12777674e+00
5.12341857e-01 -5.24145961e-01 3.75075415e-02 -6.91449523e-01
2.00999990e-01 2.50670493e-01 -1.18363810e+00 1.96428645e+00
-7.53747404e-01 6.28677189e-01 8.92343521e-02 -9.36364651e-01
5.15766442e-01 1.63063541e-01 3.18251908e-01 -9.03235257e-01
3.96563798e-01 2.38260806e-01 -2.43899673e-01 -3.09942842e-01
5.11142731e-01 -1.24491237e-01 -4.17536527e-01 3.90507519e-01
2.72091001e-01 5.60855009e-02 2.68615812e-01 4.61194217e-01
9.06312108e-01 -1.67941779e-01 1.69962585e-01 3.49364281e-01
6.06821239e-01 -2.55776662e-02 5.44045031e-01 5.77376902e-01
1.63014874e-01 3.63182068e-01 2.47576296e-01 -1.30272210e-01
-8.20723474e-01 -9.45087612e-01 -2.06904933e-02 1.63898015e+00
2.53675878e-01 -6.67830288e-01 -6.08350694e-01 -5.53819358e-01
-7.90831298e-02 5.15887320e-01 -1.85105100e-01 -2.08962113e-01
-5.36858439e-01 -9.58115518e-01 7.37115920e-01 8.23391259e-01
6.65319264e-01 -7.54756868e-01 -6.60251021e-01 2.81053841e-01
-3.92013907e-01 -1.44000244e+00 -7.48628795e-01 3.92480224e-01
-7.65598059e-01 -7.75046825e-01 -5.75758517e-01 -1.12113094e+00
7.89386153e-01 4.11522001e-01 8.36426258e-01 -6.64941296e-02
-2.01575324e-01 1.97826251e-01 -1.67615056e-01 -2.89569706e-01
-1.85751133e-02 8.73639509e-02 3.80601957e-02 -1.04034685e-01
5.68378329e-01 -4.16951388e-01 -5.02721548e-01 3.75945210e-01
-8.21731508e-01 2.75015026e-01 9.55382109e-01 1.11499059e+00
1.06162691e+00 -1.87557355e-01 5.01585864e-02 -7.04705834e-01
4.26955879e-01 -1.03198968e-01 -6.38923168e-01 3.07796746e-01
-4.02448803e-01 8.40975195e-02 3.99965852e-01 -9.38604534e-01
-7.73709953e-01 3.87552053e-01 -2.93451905e-01 -7.69183278e-01
2.48885691e-01 6.91563308e-01 -3.33988965e-01 -4.89112847e-02
5.48143208e-01 4.06342536e-01 -4.07455057e-01 -3.88617098e-01
7.30502009e-01 6.98128521e-01 7.62481272e-01 -2.69016743e-01
8.56125355e-01 4.34627652e-01 -2.45714620e-01 -6.74154222e-01
-1.05873442e+00 -6.10414445e-01 -4.10086751e-01 2.13511899e-01
8.15164983e-01 -1.35927606e+00 -7.70666957e-01 2.40060270e-01
-1.05607271e+00 -1.05714969e-01 -2.25439087e-01 5.03952801e-01
-3.10828209e-01 3.06917846e-01 -6.46728218e-01 -5.66250682e-01
-5.84302843e-01 -1.06999791e+00 1.39484572e+00 2.39116400e-01
-1.19938143e-01 -7.69082844e-01 -3.02734286e-01 4.52647597e-01
1.64396629e-01 -3.36831689e-01 1.00571680e+00 -2.82548159e-01
-6.21276557e-01 -4.84116435e-01 -4.86044377e-01 4.82187420e-02
-2.68839180e-01 -4.68208075e-01 -8.91961813e-01 -3.63731146e-01
-4.87011246e-04 -4.79069173e-01 9.90749776e-01 2.12313265e-01
1.11684108e+00 -1.19422652e-01 -6.09093487e-01 9.89818811e-01
1.31661332e+00 5.98886982e-02 2.55038500e-01 1.95660189e-01
1.18894470e+00 -1.80056408e-01 7.85212278e-01 4.08030003e-01
6.89097762e-01 6.98014200e-01 2.83024937e-01 -1.29645228e-01
-2.65061140e-01 -6.03786469e-01 4.37279135e-01 9.24931765e-01
2.56783366e-01 -1.49919301e-01 -7.51437008e-01 4.94447261e-01
-1.81815553e+00 -8.07252228e-01 1.10461451e-01 1.94176209e+00
7.42050350e-01 2.91873068e-01 6.85041696e-02 6.23276234e-02
5.16006529e-01 9.49876606e-02 -8.44099045e-01 -1.60691857e-01
-2.62481123e-01 1.18354196e-02 5.43132782e-01 4.04778510e-01
-1.09672940e+00 1.41669583e+00 5.89796734e+00 1.08280396e+00
-1.26909244e+00 4.15728807e-01 2.40588665e-01 -1.43934205e-01
-1.89618647e-01 3.29392264e-03 -1.28090537e+00 1.80876955e-01
7.93365777e-01 5.16768433e-02 4.90133703e-01 1.12042940e+00
-1.81053668e-01 -1.08439587e-01 -1.13467443e+00 1.60518134e+00
1.78287253e-01 -1.16151249e+00 1.98304042e-01 -8.06011707e-02
3.33754748e-01 1.07164942e-01 2.94083685e-01 6.75764680e-01
-7.83618689e-02 -7.08079875e-01 8.34973633e-01 2.60248221e-02
1.08322990e+00 -4.71268386e-01 3.56516272e-01 5.27937531e-01
-1.48571062e+00 -3.83336842e-01 -5.55865407e-01 2.49696374e-01
3.16263884e-01 4.13479090e-01 -9.07494366e-01 2.15350818e-02
5.96038342e-01 6.13702357e-01 -3.79440904e-01 5.50261855e-01
-2.70133078e-01 3.99352580e-01 -6.55205548e-01 -5.98013960e-03
1.23093508e-01 5.20403348e-02 4.55844700e-01 1.25436234e+00
2.75251120e-01 9.36168879e-02 2.68275291e-01 6.53638721e-01
-4.01333869e-01 1.07664853e-01 -1.34512722e-01 -1.55628026e-01
6.07341409e-01 1.27672756e+00 -5.83558202e-01 -6.65740669e-01
-7.56576657e-01 1.25642300e+00 2.81920671e-01 1.72041744e-01
-1.12882292e+00 -2.17779666e-01 3.36600184e-01 -7.53621943e-03
5.54495513e-01 -1.35684475e-01 -2.49820441e-01 -1.36279333e+00
2.89812565e-01 -1.04366970e+00 5.26500940e-01 -6.16178632e-01
-1.01595116e+00 5.17131150e-01 -1.36025816e-01 -1.27485609e+00
-3.16148192e-01 -6.17571235e-01 -1.25653565e-01 7.95028865e-01
-1.38952911e+00 -1.45631397e+00 -1.94684207e-01 9.68516231e-01
6.50139213e-01 -2.04347804e-01 9.22815681e-01 6.48675799e-01
-7.83645809e-01 1.09058499e+00 -1.59239128e-01 1.25300705e-01
5.76889992e-01 -1.11053920e+00 2.57365465e-01 8.12550485e-01
3.96425396e-01 7.36955047e-01 3.42053980e-01 -5.21528780e-01
-1.88657355e+00 -9.69758272e-01 7.94366002e-01 -3.90440643e-01
7.27051079e-01 -4.61477429e-01 -8.62304688e-01 7.69657731e-01
-2.58507013e-01 1.17306121e-01 5.97576380e-01 2.06005290e-01
-5.91712475e-01 -8.29708651e-02 -8.36640298e-01 5.86451411e-01
1.05490172e+00 -1.09387112e+00 -6.24283910e-01 5.92431843e-01
6.36101842e-01 -9.15526032e-01 -7.04014838e-01 2.07865477e-01
6.56562150e-01 -4.34770763e-01 1.06686878e+00 -2.65497744e-01
4.73964959e-02 -3.86577219e-01 -3.83142710e-01 -7.42250085e-01
-3.08296204e-01 -4.37480539e-01 -5.80454111e-01 1.00185156e+00
5.45816779e-01 -2.66530365e-01 7.38750875e-01 5.85042834e-01
-6.84345216e-02 -7.24097073e-01 -9.91691411e-01 -4.21904057e-01
-3.15491229e-01 -6.01153553e-01 3.75929177e-01 7.16688395e-01
-3.04524717e-03 9.07939851e-01 -4.57401752e-01 2.85431951e-01
4.96133745e-01 2.68955350e-01 6.93910599e-01 -9.51084614e-01
-4.96804267e-01 -2.10794434e-01 -3.12744200e-01 -1.88751423e+00
-2.23422255e-02 -8.98973167e-01 -6.45784149e-03 -1.43705380e+00
4.61093575e-01 -3.78953069e-01 -1.95974946e-01 7.36807346e-01
-2.14861855e-01 3.65612119e-01 3.90143424e-01 3.25200886e-01
-9.07719433e-01 3.96596760e-01 1.10277343e+00 -4.45411623e-01
-3.69875908e-01 -2.21288040e-01 -7.42771626e-01 9.33947980e-01
2.73382992e-01 -5.29461920e-01 -3.97107869e-01 -7.21741736e-01
3.55442196e-01 1.22489989e-01 2.90249437e-01 -9.43207383e-01
5.18453479e-01 2.54558951e-01 4.03603166e-01 -9.62605774e-01
6.67237520e-01 -7.44736314e-01 -2.72061735e-01 2.95068443e-01
-1.66674510e-01 3.63773331e-02 2.54443556e-01 5.07685721e-01
-2.92906195e-01 -1.94139138e-01 5.19818664e-01 5.64933522e-03
-1.02510083e+00 3.40303868e-01 -1.88532203e-01 -1.56605691e-02
6.90305352e-01 -2.93143690e-01 -1.72709171e-02 -2.43884712e-01
-9.67607260e-01 2.62769371e-01 1.44036800e-01 4.56612617e-01
7.93601513e-01 -1.45013928e+00 -3.04834753e-01 1.79766029e-01
1.35171011e-01 1.81132451e-01 3.02383184e-01 9.26347733e-01
-1.09111384e-01 5.22515953e-01 1.22155681e-01 -8.59234393e-01
-1.31165636e+00 6.66047812e-01 1.23556271e-01 -3.16940397e-01
-6.99999392e-01 1.11443269e+00 4.02298391e-01 -1.04473926e-01
5.12522876e-01 -5.26204407e-01 5.26500158e-02 3.40903223e-01
6.70630038e-01 -2.97470782e-02 -5.11840768e-02 -8.28024387e-01
-4.34902757e-01 8.37713599e-01 -3.19659472e-01 -1.94454372e-01
1.00250888e+00 -2.37354904e-01 2.03561991e-01 2.16838107e-01
1.32195842e+00 -1.40218109e-01 -9.33263719e-01 -5.37143171e-01
-1.73447773e-01 -2.95745134e-01 3.81628126e-01 -4.83781159e-01
-9.84858513e-01 9.44125533e-01 7.82395899e-01 -2.97734767e-01
1.35523200e+00 2.06435427e-01 1.12637162e+00 7.86547542e-01
3.91538769e-01 -1.12120485e+00 2.24593773e-01 5.80499768e-01
6.26887918e-01 -1.35052764e+00 -3.81330103e-02 -4.83353406e-01
-6.62470341e-01 7.70774305e-01 5.99100351e-01 1.64577425e-01
5.39333284e-01 2.60083318e-01 -1.22099996e-01 -2.24318534e-01
-4.93625075e-01 -4.29920912e-01 3.98610175e-01 3.65451604e-01
1.81487933e-01 3.20323184e-02 2.94705555e-02 5.72592080e-01
-1.43059731e-01 -9.36972797e-02 -1.10694215e-01 8.39729071e-01
-5.02006352e-01 -1.04029453e+00 -4.03302342e-01 5.02269983e-01
-3.15976381e-01 -3.54327321e-01 3.02531961e-02 6.73566103e-01
2.86061198e-01 7.52358377e-01 1.53198726e-02 -6.13657117e-01
5.45440078e-01 -6.34466410e-02 3.80806983e-01 -7.95116544e-01
-3.40415061e-01 5.42806208e-01 -1.16009906e-01 -6.34505332e-01
-1.70469448e-01 -6.04277253e-01 -1.58630073e+00 -8.56086910e-02
-5.75924516e-01 -3.14992070e-01 6.49145246e-01 9.82047677e-01
2.86876500e-01 5.95390677e-01 1.97748750e-01 -7.53310680e-01
-5.10480523e-01 -9.26224053e-01 -2.96181709e-01 4.16613698e-01
1.82614669e-01 -6.88074172e-01 -8.67328048e-02 7.72116631e-02] | [9.990898132324219, 0.9806615710258484] |
5865e954-6ccb-4784-8a06-540a526bf69b | online-conversation-disentanglement-with | 2010.11080 | null | https://arxiv.org/abs/2010.11080v1 | https://arxiv.org/pdf/2010.11080v1.pdf | Online Conversation Disentanglement with Pointer Networks | Huge amounts of textual conversations occur online every day, where multiple conversations take place concurrently. Interleaved conversations lead to difficulties in not only following the ongoing discussions but also extracting relevant information from simultaneous messages. Conversation disentanglement aims to separate intermingled messages into detached conversations. However, existing disentanglement methods rely mostly on handcrafted features that are dataset specific, which hinders generalization and adaptability. In this work, we propose an end-to-end online framework for conversation disentanglement that avoids time-consuming domain-specific feature engineering. We design a novel way to embed the whole utterance that comprises timestamp, speaker, and message text, and proposes a custom attention mechanism that models disentanglement as a pointing problem while effectively capturing inter-utterance interactions in an end-to-end fashion. We also introduce a joint-learning objective to better capture contextual information. Our experiments on the Ubuntu IRC dataset show that our method achieves state-of-the-art performance in both link and conversation prediction tasks. | ['Shafiq Joty', 'Tao Yu'] | 2020-10-21 | null | https://aclanthology.org/2020.emnlp-main.512 | https://aclanthology.org/2020.emnlp-main.512.pdf | emnlp-2020-11 | ['conversation-disentanglement'] | ['natural-language-processing'] | [ 2.37929717e-01 -1.21915471e-02 -2.68171638e-01 -6.52902901e-01
-1.10776532e+00 -6.27009690e-01 1.04000604e+00 1.10832088e-01
-3.50603461e-01 7.05511570e-01 8.99666429e-01 -3.26393753e-01
-7.10597485e-02 -2.81079710e-01 -2.71040529e-01 -3.90506208e-01
2.59720236e-02 5.36903501e-01 -1.88295886e-01 -2.50465631e-01
4.82227877e-02 6.93362802e-02 -1.34217501e+00 5.39721787e-01
7.68337131e-01 7.95616746e-01 2.68883705e-02 1.03684878e+00
-2.94057310e-01 1.09160936e+00 -7.35058308e-01 -5.14787197e-01
-9.22429338e-02 -3.34900141e-01 -9.11668241e-01 3.23744029e-01
2.35920891e-01 -5.15726566e-01 -5.79748631e-01 4.11993921e-01
4.61015671e-01 2.12738842e-01 3.17823112e-01 -1.53161180e+00
-3.10546041e-01 7.72648692e-01 -1.66935652e-01 2.50027776e-01
7.41236031e-01 1.01128705e-01 1.40356541e+00 -7.56648540e-01
1.79618075e-01 1.29006469e+00 4.00533319e-01 4.38203782e-01
-1.39450049e+00 -7.99388230e-01 4.08003539e-01 2.52899855e-01
-9.37278032e-01 -8.65178108e-01 1.00469697e+00 -3.19763333e-01
1.06633055e+00 7.72528529e-01 4.49352205e-01 1.48075294e+00
-1.69736534e-01 1.17871916e+00 7.55136251e-01 -2.40613565e-01
-2.63020903e-01 7.29730651e-02 3.47820640e-01 2.84672201e-01
-2.88317353e-01 -3.65097553e-01 -8.70949030e-01 -4.50298667e-01
2.43526921e-01 2.17201859e-01 -5.44570148e-01 -1.00341454e-01
-1.46034718e+00 7.53576577e-01 -1.13973226e-02 1.18786976e-01
-1.12220533e-01 -3.33092749e-01 6.85424864e-01 5.96602619e-01
6.36362016e-01 3.59512627e-01 -5.28512478e-01 -8.31116974e-01
-6.00973547e-01 2.19463512e-01 1.31255019e+00 9.66359854e-01
7.41814494e-01 -5.18144011e-01 -2.26568431e-01 9.73915219e-01
1.82405621e-01 1.96466684e-01 5.22904456e-01 -7.54621327e-01
9.61475790e-01 7.58621812e-01 2.53325135e-01 -1.10264015e+00
-4.70244497e-01 -2.59911753e-02 -6.73332095e-01 -5.43337464e-01
4.86030281e-01 -3.88917953e-01 -1.27542749e-01 1.67146623e+00
3.72403473e-01 5.37793696e-01 2.00347006e-02 7.28341997e-01
9.44274902e-01 5.06361246e-01 -3.44098866e-01 -5.16774535e-01
1.41855478e+00 -1.52958524e+00 -9.96812284e-01 -3.66239220e-01
6.61464036e-01 -8.21095169e-01 1.25165689e+00 2.07468823e-01
-9.44832444e-01 -2.57501364e-01 -9.90521550e-01 -4.33517694e-01
-1.31034413e-02 -1.10534400e-01 5.46088338e-01 3.75771999e-01
-5.40767252e-01 2.59733438e-01 -6.98093653e-01 -1.06666148e-01
-3.27131599e-02 3.24755758e-01 -4.87029523e-01 1.78897247e-01
-1.28110516e+00 7.59316981e-01 1.09404258e-01 2.09812164e-01
-4.31523770e-01 -5.62897265e-01 -8.44159484e-01 2.21178234e-01
6.48164511e-01 -5.22383392e-01 1.59088802e+00 -8.37500811e-01
-2.08239341e+00 6.91452503e-01 -5.29897392e-01 -2.39651650e-01
6.11750722e-01 -5.34621894e-01 -6.87196553e-01 -2.17245892e-01
-1.96721449e-01 -1.12940326e-01 6.53009534e-01 -9.49321270e-01
-6.06960952e-01 -1.62906647e-01 3.32806140e-01 3.99095505e-01
-6.36785090e-01 8.85750279e-02 -6.33034587e-01 -3.58134866e-01
-1.53096274e-01 -1.04034269e+00 3.17714334e-01 -2.69137144e-01
-5.80746651e-01 -3.47224325e-01 8.77362549e-01 -5.58070600e-01
1.37338257e+00 -1.97668600e+00 3.04337054e-01 -2.55121827e-01
6.15050852e-01 2.56798416e-01 -1.93500593e-01 7.99432278e-01
1.02559410e-01 -1.04939714e-01 1.89096779e-01 -1.06063652e+00
1.38328657e-01 1.06568500e-01 -2.62017012e-01 3.11461389e-01
-1.16375498e-01 7.20194817e-01 -9.77713108e-01 -4.17330682e-01
2.23922804e-01 3.77125829e-01 -4.69894886e-01 7.21856415e-01
-2.05478415e-01 7.42303193e-01 -3.74999255e-01 1.56840220e-01
5.30826211e-01 -3.74409527e-01 4.64635223e-01 -8.78121238e-03
7.08180368e-02 1.11590731e+00 -7.72118330e-01 1.84638262e+00
-1.03090167e+00 1.09118259e+00 2.52902001e-01 -7.17092991e-01
8.63232136e-01 6.09253883e-01 2.61260420e-01 -3.05864185e-01
2.59605378e-01 -1.48502037e-01 -5.12462035e-02 -7.19323933e-01
5.83169997e-01 1.09237246e-01 -3.71808022e-01 8.54633570e-01
-1.94660038e-01 2.57880449e-01 -2.42345734e-03 3.25127304e-01
1.02226722e+00 -3.56949151e-01 5.73819578e-01 1.60884440e-01
6.60198987e-01 -5.79816401e-01 5.33391774e-01 4.35562015e-01
-2.49911681e-01 5.46989322e-01 7.98751116e-01 -3.71077210e-01
-5.53436458e-01 -8.87812436e-01 2.75564343e-01 1.28874779e+00
1.80182949e-01 -6.75167918e-01 -4.11443025e-01 -8.24082553e-01
-2.21862346e-02 8.01606059e-01 -5.54040790e-01 -5.91899231e-02
-6.66444480e-01 -5.41909397e-01 4.46198374e-01 4.24262166e-01
1.83786035e-01 -7.25841045e-01 -3.73407267e-02 2.89712101e-01
-9.73684609e-01 -1.49274194e+00 -1.01852775e+00 -3.12573642e-01
-4.66514766e-01 -1.08320534e+00 -1.17918663e-01 -4.58395332e-01
2.60286301e-01 7.26614118e-01 1.16836238e+00 -1.12738472e-03
1.18886314e-01 5.86557984e-02 -2.94833839e-01 -8.81077573e-02
-2.95404941e-01 3.66117716e-01 -8.51682797e-02 4.46550399e-01
6.11116111e-01 -8.50769162e-01 -4.90727395e-01 4.74462748e-01
-5.24062395e-01 3.27521205e-01 3.01270604e-01 9.59036946e-01
-2.28779733e-01 -5.30535519e-01 7.40410984e-01 -9.79454756e-01
1.05616784e+00 -5.88579655e-01 -1.01625197e-01 2.93938905e-01
-2.70150125e-01 6.70633018e-02 5.93841136e-01 -5.72442770e-01
-1.21599388e+00 -4.01822329e-01 -5.77712711e-03 -1.09200515e-01
-1.19820543e-01 3.74515742e-01 -5.44849813e-01 5.68467677e-01
2.65952528e-01 3.20362240e-01 1.70265198e-01 -4.79777426e-01
4.52208251e-01 1.24013042e+00 2.90628314e-01 -5.36657214e-01
5.22325456e-01 3.27098429e-01 -7.53112912e-01 -8.42534065e-01
-9.46783364e-01 -7.75496721e-01 -4.13709462e-01 -1.17874973e-01
4.85887647e-01 -1.09565890e+00 -1.49688566e+00 2.49105111e-01
-1.32128131e+00 -2.41979390e-01 4.98809755e-01 5.33823609e-01
-5.12838423e-01 4.58362073e-01 -6.54161870e-01 -8.48660469e-01
-4.58943933e-01 -1.06204700e+00 1.07325423e+00 -1.35422889e-02
-8.32924187e-01 -1.02831173e+00 2.33563960e-01 1.05595291e+00
1.89465791e-01 9.75602716e-02 4.72994834e-01 -1.01655066e+00
-5.39989471e-01 -3.98331404e-01 -3.19021553e-01 6.26194384e-03
5.91843963e-01 -1.12654932e-01 -1.09335339e+00 -3.44977051e-01
3.55830863e-02 -3.49333227e-01 6.27240837e-01 -3.89092743e-01
1.06292295e+00 -8.54332805e-01 -2.90888786e-01 3.46993506e-01
7.20212042e-01 -1.37595430e-01 2.91904390e-01 1.20195568e-01
9.13612068e-01 7.20700383e-01 4.81742114e-01 5.98331690e-01
9.09520447e-01 9.43455100e-01 3.24409127e-01 2.27611601e-01
1.70908421e-01 -1.62668154e-01 3.04888815e-01 1.19186342e+00
3.01711321e-01 -5.97858846e-01 -5.93852401e-01 6.33637071e-01
-2.31310749e+00 -9.59640920e-01 -9.92069840e-02 2.00728369e+00
1.22376096e+00 3.03388476e-01 2.66912222e-01 -1.13500785e-02
5.36874950e-01 5.45584202e-01 -3.50416988e-01 -2.82630801e-01
3.14879507e-01 -4.78439331e-01 -1.01206988e-01 9.89044309e-01
-1.00050676e+00 4.75425094e-01 5.27314281e+00 5.47275245e-01
-1.13380313e+00 5.37456386e-02 1.84635043e-01 -6.00737691e-01
-4.54631478e-01 -2.57798042e-02 -7.14449704e-01 7.49106050e-01
1.00049424e+00 -3.36132348e-01 5.98477483e-01 5.62675893e-01
3.76461864e-01 2.48646095e-01 -1.63496149e+00 1.24348307e+00
1.55712098e-01 -1.25393867e+00 -2.74389058e-01 -1.98884979e-01
3.87645006e-01 -7.22147822e-02 -3.06939960e-01 3.60004067e-01
2.09646910e-01 -7.71695256e-01 4.06235665e-01 3.81173193e-01
3.51950169e-01 -7.33088374e-01 5.54800630e-01 6.48579657e-01
-1.03904235e+00 9.44672897e-02 2.40348563e-01 -4.85987872e-01
2.81919628e-01 5.77352047e-01 -1.21919370e+00 6.11841381e-01
1.92126632e-01 7.59345770e-01 -3.01607605e-02 4.97646719e-01
-2.54161686e-01 4.67687398e-01 -1.33781508e-01 -2.34789476e-01
2.16999883e-03 6.75664321e-02 7.03084052e-01 1.37584817e+00
-1.50925800e-01 -4.56869192e-02 2.97477961e-01 4.24529165e-01
-4.36032921e-01 -2.15894766e-02 -4.38288838e-01 -9.12863985e-02
9.14711416e-01 1.25208759e+00 -1.25436381e-01 -3.22617441e-01
-7.86521077e-01 9.14304495e-01 6.82139099e-01 3.22565109e-01
-7.53744364e-01 -3.03254843e-01 1.23496914e+00 -2.60989785e-01
9.78898630e-02 -2.93848187e-01 -2.74933934e-01 -1.61273670e+00
2.67631769e-01 -1.07926679e+00 2.06419691e-01 -2.88312972e-01
-1.43073750e+00 5.70863605e-01 -2.09560543e-01 -1.15781772e+00
-5.38155019e-01 -1.00231916e-01 -8.45642447e-01 7.86262095e-01
-1.38979673e+00 -1.08145392e+00 -4.72446203e-01 4.69730556e-01
9.18771446e-01 1.08070068e-01 1.07135201e+00 3.14821631e-01
-8.52463067e-01 9.62412179e-01 1.62146732e-01 3.93311203e-01
1.08630908e+00 -1.10455549e+00 5.14542997e-01 4.71311897e-01
8.45293701e-02 8.19010615e-01 8.76406252e-01 -9.49234962e-02
-1.40867281e+00 -7.54006445e-01 1.52009869e+00 -7.39038348e-01
8.60145152e-01 -9.79205728e-01 -9.47418690e-01 9.61374402e-01
4.54395950e-01 -2.66930044e-01 1.17664218e+00 7.68765986e-01
-5.88118553e-01 -2.51879513e-01 -6.56214893e-01 9.78771269e-01
9.04662132e-01 -1.07255387e+00 -7.13428438e-01 2.27774635e-01
1.01520336e+00 -4.99788612e-01 -6.78322911e-01 -7.30736181e-02
8.95725012e-01 -9.00250554e-01 8.57626259e-01 -6.33004904e-01
4.89418954e-01 1.50296375e-01 5.56531660e-02 -1.23546588e+00
7.57034421e-02 -1.37051606e+00 -5.97532988e-01 1.40813780e+00
5.24400175e-01 -7.63366103e-01 6.97010458e-01 9.86422420e-01
9.41405147e-02 -6.37889206e-01 -6.95600748e-01 -5.72611988e-01
-4.01955217e-01 -3.71598154e-01 8.58662307e-01 1.18801594e+00
7.46236742e-01 1.01166105e+00 -8.47022533e-01 -1.09639708e-02
2.69499332e-01 4.89333719e-01 1.34658122e+00 -1.21759176e+00
-4.63200539e-01 -3.68359327e-01 -2.29806341e-02 -1.60302293e+00
4.92713511e-01 -5.04654825e-01 1.17448963e-01 -1.27683449e+00
2.79477060e-01 -1.75917298e-01 1.47398124e-02 1.09352559e-01
-5.09299815e-01 -2.93208838e-01 1.59112960e-01 2.49426439e-01
-8.62630904e-01 7.72688746e-01 1.09355450e+00 -1.74618825e-01
-3.48889738e-01 2.43248388e-01 -7.94794679e-01 5.91056466e-01
7.76965320e-01 -4.24260616e-01 -6.82466030e-01 -5.41908324e-01
-2.98468582e-03 5.11617005e-01 2.92594414e-02 -2.93774068e-01
2.86238045e-01 -2.57027000e-01 -4.13986385e-01 -5.03631175e-01
7.92964816e-01 -8.63651335e-01 -3.07982594e-01 -1.48788765e-01
-8.06393027e-01 -1.47109851e-01 9.20944959e-02 9.97072816e-01
-2.74421662e-01 2.16795295e-01 1.52406946e-01 2.62389123e-01
-1.72102824e-01 1.48879081e-01 -3.84211779e-01 2.04172209e-01
8.17300320e-01 1.27707243e-01 -4.14772063e-01 -8.63359869e-01
-7.36711204e-01 5.82666755e-01 1.96427450e-01 8.78458261e-01
3.18761587e-01 -1.22706962e+00 -8.48230481e-01 9.21704397e-02
2.67384231e-01 -1.75255742e-02 3.12322795e-01 8.75293970e-01
-5.49160643e-03 4.29190725e-01 2.60761321e-01 -4.73488718e-01
-1.82525086e+00 4.01187509e-01 -2.16002241e-02 -5.23398578e-01
-4.82361764e-01 8.92865598e-01 2.12780029e-01 -7.41043508e-01
5.25417209e-01 -4.05901015e-01 -8.87428299e-02 2.73480684e-01
8.13674629e-01 3.49499255e-01 9.33322534e-02 -3.83100748e-01
-3.86080205e-01 -2.70933975e-02 -5.42495072e-01 -7.22966492e-02
1.13378119e+00 -7.09442198e-01 1.21287983e-02 7.79954851e-01
1.74622047e+00 1.45276919e-01 -1.45481205e+00 -7.73605347e-01
-5.64260688e-03 -6.68219686e-01 -1.58231065e-01 -6.63141012e-01
-6.21038139e-01 8.08093071e-01 -3.11247200e-01 5.78224003e-01
8.74208272e-01 -5.52279502e-02 1.21342468e+00 5.31654239e-01
-2.85451077e-02 -8.20109248e-01 3.65015149e-01 5.66528261e-01
1.04773688e+00 -1.60807860e+00 -1.52478740e-01 -5.78434348e-01
-6.85610890e-01 1.09958673e+00 4.77312952e-01 3.05872142e-01
6.71724856e-01 2.83214092e-01 1.08472139e-01 1.17112890e-01
-1.29805636e+00 8.14383551e-02 2.52543181e-01 1.74034148e-01
6.74201667e-01 2.13015154e-01 -2.15512797e-01 7.10621655e-01
-2.45296896e-01 -1.73812076e-01 4.87096071e-01 8.80717754e-01
-8.87702331e-02 -1.28023803e+00 -5.71550317e-02 1.02237448e-01
-2.95920014e-01 -1.04014866e-01 -6.74429715e-01 6.34925425e-01
-5.03691018e-01 1.50418663e+00 2.39918098e-01 -5.52275538e-01
2.21046269e-01 2.20777154e-01 1.66878819e-01 -5.45761704e-01
-6.39725685e-01 -9.78968963e-02 8.82962108e-01 -4.19829667e-01
-3.34076792e-01 -6.08418226e-01 -9.65729415e-01 -8.44996035e-01
-5.23694217e-01 2.62556553e-01 3.81518304e-01 1.26781678e+00
5.75042903e-01 6.57375813e-01 1.11162710e+00 -7.83248484e-01
-5.01333356e-01 -1.28494048e+00 -8.68798569e-02 5.49282312e-01
9.78986263e-01 -4.54487175e-01 -4.36685681e-01 -6.64829090e-02] | [12.59268856048584, 7.781098365783691] |
48898cf5-89e4-4b02-a92b-69cfafaa92a4 | spectral-efficiency-analysis-of-uplink | 2212.02164 | null | https://arxiv.org/abs/2212.02164v2 | https://arxiv.org/pdf/2212.02164v2.pdf | Spectral Efficiency Analysis of Uplink-Downlink Decoupled Access in C-V2X Networks | The uplink (UL)/downlink (DL) decoupled access has been emerging as a novel access architecture to improve the performance gains in cellular networks. In this paper, we investigate the UL/DL decoupled access performance in cellular vehicle-to-everything (C-V2X). We propose a unified analytical framework for the UL/DL decoupled access in C-V2X from the perspective of spectral efficiency (SE). By modeling the UL/DL decoupled access C-V2X as a Cox process and leveraging the stochastic geometry, we obtain the joint association probability, the UL/DL distance distributions to serving base stations and the SE for the UL/DL decoupled access in C-V2X networks with different association cases. We conduct extensive Monte Carlo simulations to verify the accuracy of the proposed unified analytical framework, and the results show a better system average SE of UL/DL decoupled access in C-V2X. | ['Shen', 'Xuemin', 'Haibo Zhou', 'Tianqi Zhang', 'Yunting Xu', 'Kai Yu', 'Luofang Jiao'] | 2022-12-05 | null | null | null | null | ['spectral-efficiency-analysis-of-uplink'] | ['computer-code'] | [-8.87311041e-01 1.05137154e-01 -3.57021064e-01 7.55374581e-02
-4.84424442e-01 -6.78396761e-01 3.39375794e-01 -2.90075630e-01
-4.16633822e-02 1.47442591e+00 -2.74471603e-02 -1.38200104e+00
-1.92050412e-01 -6.91872180e-01 -2.38271326e-01 -1.11443841e+00
-2.63474107e-01 4.63894576e-01 -1.27735317e-01 -1.78546309e-01
-4.96102333e-01 8.68031800e-01 -4.65883583e-01 -9.21655297e-01
8.15877199e-01 1.41455221e+00 1.74978375e-01 7.04364002e-01
-1.09025091e-01 2.78798610e-01 -5.50740063e-01 -4.56722379e-01
1.98614717e-01 -3.37355793e-01 -2.69615084e-01 -6.46307021e-02
-6.67057335e-01 -5.24983644e-01 -1.14727259e+00 4.67020124e-01
8.05966318e-01 -2.25114837e-01 9.46935892e-01 -1.84492159e+00
-3.96961182e-01 2.39902541e-01 -5.00519097e-01 5.34282327e-01
-3.48002553e-01 2.80249178e-01 6.49055302e-01 -8.90047431e-01
2.98116386e-01 1.41539240e+00 6.23791218e-01 2.67089754e-01
-8.51730347e-01 -8.97770464e-01 -2.43817165e-01 2.15611830e-01
-2.27764511e+00 -6.12100601e-01 -6.02715760e-02 -1.48720875e-01
4.11939025e-01 4.88295674e-01 5.53641438e-01 4.38046753e-01
2.83790320e-01 1.02252746e+00 6.54540122e-01 -1.10677630e-01
3.21036994e-01 4.35735762e-01 3.03640574e-01 5.80100238e-01
6.37630224e-01 -1.55701816e-01 4.54233021e-01 -4.94793892e-01
9.19878483e-01 -2.80472994e-01 -4.76538241e-01 -2.24362805e-01
-6.12012982e-01 5.67950904e-01 -5.03056683e-02 -2.92858720e-01
-5.50050020e-01 6.58640623e-01 -1.29993483e-01 3.30042481e-01
-4.64784577e-02 -5.19013822e-01 -1.68000266e-01 1.36477575e-01
-9.10678267e-01 1.45137265e-01 6.55409038e-01 2.06823206e+00
4.90627401e-02 2.43441775e-01 -9.97207642e-01 3.71232271e-01
9.89530087e-01 1.33567703e+00 -7.97478795e-01 -1.03000724e+00
4.43334252e-01 -3.68719757e-01 5.34407794e-01 -2.76792943e-01
-2.71257579e-01 -1.68886209e+00 -1.20215070e+00 -2.28629619e-01
2.70810038e-01 -8.86810601e-01 -2.44674236e-01 1.56837571e+00
-1.37416735e-01 -4.52872626e-02 6.61794305e-01 5.39862931e-01
4.28129792e-01 8.52014482e-01 -6.20948710e-02 -9.95850921e-01
1.17751610e+00 -6.45737350e-01 -1.38030493e+00 4.43459511e-01
5.92197418e-01 -4.25759822e-01 -6.63007572e-02 -8.83745104e-02
-1.51796746e+00 -2.27500916e-01 -1.18538892e+00 4.81756836e-01
4.31404948e-01 3.76096636e-01 -5.38715310e-02 1.09225023e+00
-1.05878782e+00 -5.75756609e-01 -3.29492241e-01 -6.61600351e-01
7.31034279e-01 2.50951827e-01 2.65688956e-01 -3.76771539e-01
-1.26211047e+00 3.19140345e-01 -2.96143629e-02 -2.59113669e-01
-1.06562543e+00 -7.05837071e-01 -2.07361490e-01 4.27952856e-01
5.80099165e-01 -7.32509077e-01 1.33095145e+00 4.83490795e-01
-8.70062172e-01 -6.17360659e-02 -4.94972378e-01 -4.51323271e-01
4.97748017e-01 4.73074734e-01 -5.43520153e-01 1.19128928e-01
1.11847296e-01 -8.27703997e-02 -2.15187609e-01 -1.60096371e+00
-1.23163843e+00 -5.45830488e-01 -9.11654830e-02 -1.99986666e-01
8.74919072e-02 -2.53158987e-01 -9.95850563e-01 -5.88056862e-01
-1.38165399e-01 -1.09119713e+00 -2.59714246e-01 -3.15690003e-02
-6.74582660e-01 -3.68526906e-01 9.15268242e-01 -6.36396885e-01
1.74006557e+00 -2.22437739e+00 -7.91367665e-02 5.38281918e-01
4.01499093e-01 1.50564119e-01 3.26615572e-01 7.05454946e-01
7.90786028e-01 7.29440078e-02 4.71835643e-01 -6.36717975e-01
4.95188795e-02 3.72847974e-01 -3.44159454e-01 4.32904601e-01
-7.97881663e-01 8.82991254e-01 -9.25863087e-01 -3.50465149e-01
-3.69173810e-02 2.14579135e-01 -1.88873678e-01 -7.76850572e-03
2.89922833e-01 4.92917806e-01 -8.55722547e-01 7.59231746e-01
1.08003521e+00 -1.92104667e-01 2.43600950e-01 -2.19382271e-01
-2.56873637e-01 -7.69279897e-01 -6.80425763e-01 8.09047401e-01
-4.41269636e-01 6.88038230e-01 3.90636325e-01 -5.80842972e-01
5.46800077e-01 7.97582507e-01 1.64474636e-01 -4.49523270e-01
5.44851542e-01 5.05891502e-01 -1.05295040e-01 -1.04114011e-01
-2.24159677e-02 -6.44833595e-02 5.52127399e-02 3.92643996e-02
1.46373242e-01 6.83441639e-01 -3.52601781e-02 3.62191767e-01
6.95111394e-01 -6.22336447e-01 7.33907938e-01 -7.32415318e-01
7.61948407e-01 -7.57441282e-01 4.88678008e-01 1.00510025e+00
-3.77799600e-01 1.05028793e-01 4.41884726e-01 3.86070043e-01
-1.08460963e+00 -1.72310472e+00 -3.67537409e-01 3.04632317e-02
8.07479680e-01 -2.25300834e-01 -3.55796218e-01 -1.36427253e-01
2.28917494e-01 1.35824108e+00 1.01393141e-01 -3.19346637e-02
2.97714233e-01 -6.27832174e-01 1.05854344e+00 4.79099959e-01
8.25371504e-01 4.94832247e-01 2.97862232e-01 5.88938177e-01
-7.78844804e-02 -1.50918102e+00 -3.84996831e-01 -1.03883348e-01
-2.42664993e-01 -3.89385760e-01 -1.12222469e+00 -2.68214554e-01
9.23885182e-02 7.45712340e-01 5.00236690e-01 -1.76291510e-01
3.40870470e-01 7.47966886e-01 8.80291872e-03 -4.29056495e-01
-1.52714655e-01 3.87214348e-02 5.94923854e-01 4.10986334e-01
-2.16306821e-02 -3.17209035e-01 -7.43597388e-01 7.91994810e-01
8.09752792e-02 -3.30862343e-01 3.29290628e-01 1.21233627e-01
5.03254592e-01 2.40615681e-01 8.73493552e-01 2.45054439e-01
6.26252294e-01 -1.18979979e+00 -5.33301234e-01 4.56543356e-01
-5.69417596e-01 -2.87005335e-01 2.19578385e-01 2.54618019e-01
-9.28247869e-01 -6.76858068e-01 -4.07793939e-01 -3.84635687e-01
3.91639858e-01 6.17051780e-01 -8.58155072e-01 -7.53685310e-02
-3.00045460e-02 2.33216196e-01 1.29095837e-01 -2.57193416e-01
3.43207598e-01 1.13173139e+00 3.58534008e-01 -4.64875966e-01
9.36221957e-01 6.85054898e-01 4.05206949e-01 -1.41118586e+00
-2.55449325e-01 -3.67527843e-01 1.58207174e-02 -3.76265675e-01
7.63823032e-01 -1.47917485e+00 -9.57483888e-01 3.27501595e-01
-1.28756404e+00 -1.46042630e-01 2.51335073e-02 8.46036434e-01
-4.68552887e-01 3.55824292e-01 -4.48277354e-01 -1.50170720e+00
8.66636112e-02 -1.19517410e+00 5.74828267e-01 3.17796022e-01
3.23525041e-01 -8.40121150e-01 -5.32816768e-01 -2.88137682e-02
5.62263191e-01 -1.60317957e-01 1.03623819e+00 -2.28776515e-01
-8.89349043e-01 -1.55997071e-02 -5.76932132e-01 4.56376411e-02
-3.56794544e-03 -4.08770740e-01 -4.36359763e-01 -7.15267360e-01
-2.20495120e-01 7.10262179e-01 -3.48537751e-02 1.16021419e+00
9.46150959e-01 -1.90716788e-01 -1.25935876e+00 6.98220015e-01
1.47728240e+00 7.05094695e-01 1.15351987e+00 -2.34271154e-01
3.26645583e-01 -9.24414396e-02 6.30224347e-01 8.38422120e-01
8.12282264e-01 9.66874957e-01 4.83531177e-01 1.11840576e-01
-2.67900396e-02 1.50924951e-01 -1.19633943e-01 5.54446161e-01
-9.72417668e-02 -1.53119826e+00 -7.44693637e-01 6.44221961e-01
-1.72731102e+00 -8.99129450e-01 -9.24772859e-01 1.88686776e+00
2.91396752e-02 -1.00191481e-01 1.23072602e-01 -5.43909483e-02
6.41763628e-01 -5.77233136e-01 -2.40822762e-01 2.77349561e-01
-6.80551887e-01 -3.07645291e-01 1.29714417e+00 6.88571155e-01
-3.54092658e-01 3.81455362e-01 5.44334412e+00 1.61698258e+00
-4.93135154e-01 6.04657412e-01 8.35516989e-01 -8.91849771e-03
-5.55594742e-01 -1.57863483e-01 -1.14937568e+00 5.51966250e-01
6.36686325e-01 -5.86534619e-01 -1.10825099e-01 2.81234622e-01
8.53494763e-01 3.81763689e-02 -5.63158453e-01 1.43404746e+00
-5.38824081e-01 -1.67002881e+00 2.19973475e-01 9.28320646e-01
4.12460089e-01 8.60532597e-02 -6.36681616e-02 3.35694492e-01
3.86916213e-02 -6.54232264e-01 6.61509275e-01 1.19685233e+00
1.15271175e+00 -1.00275385e+00 1.03561640e+00 2.25476637e-01
-1.23789561e+00 -7.07358241e-01 3.11344773e-01 3.09796870e-01
1.21527910e+00 5.59620619e-01 -1.08607866e-01 9.78049517e-01
3.22663933e-01 2.51965113e-02 2.09178120e-01 1.40425467e+00
2.39001662e-01 7.28286266e-01 -2.21864566e-01 2.37262417e-02
5.10819912e-01 -5.73144853e-01 1.14718592e+00 8.13831985e-01
1.16026449e+00 4.94067907e-01 -5.86655401e-02 5.34652650e-01
-3.07893995e-02 1.55193999e-01 -3.92121434e-01 4.43936557e-01
1.22108603e+00 9.51740324e-01 -3.58061433e-01 -4.10082132e-01
-6.49619102e-01 5.90572298e-01 -6.13701403e-01 1.05888307e+00
-1.05371523e+00 -6.17339015e-01 1.25783527e+00 4.70244527e-01
4.46961403e-01 -7.49146640e-01 -5.08502543e-01 -5.97133100e-01
-2.89855272e-01 -1.45405501e-01 -1.81358263e-01 -7.68902421e-01
-8.80929887e-01 1.40717372e-01 1.04328059e-01 -1.24351060e+00
4.07629877e-01 -2.58571893e-01 -4.52297240e-01 9.70652759e-01
-1.42924094e+00 -9.06751156e-01 -4.15191054e-01 3.39048445e-01
1.28547937e-01 -9.17223155e-01 2.97644377e-01 1.01007414e+00
-4.84589189e-01 9.82983708e-01 1.08915889e+00 -1.37029856e-01
6.05611131e-02 -7.72767901e-01 -2.96350658e-01 7.21844733e-01
-9.58853304e-01 3.92836511e-01 7.06526101e-01 -3.37415278e-01
-1.25772846e+00 -1.30268669e+00 4.62216675e-01 5.65028191e-03
1.16262898e-01 -3.19971442e-01 -2.63880551e-01 5.10625362e-01
3.14614892e-01 -3.38128544e-02 1.06577766e+00 -6.21455133e-01
4.21985358e-01 -1.99608758e-01 -1.19176650e+00 1.15900362e+00
1.07837558e+00 -2.54209697e-01 3.75458330e-01 6.11075610e-02
7.40076661e-01 1.92601174e-01 -8.59917104e-01 7.08623305e-02
5.77272058e-01 -4.09515917e-01 9.62975264e-01 -7.77620897e-02
-8.96448255e-01 -6.31234169e-01 -1.20845187e+00 -9.82699513e-01
-5.93183577e-01 -1.02459669e+00 -5.25754213e-01 1.39415812e+00
2.96138555e-01 -6.40170157e-01 3.64153177e-01 -1.37529969e-01
-1.79712594e-01 -6.98315859e-01 -1.19941723e+00 -1.33855939e+00
9.21876431e-02 -5.22326171e-01 7.35042453e-01 1.36366948e-01
-4.09364641e-01 6.90413654e-01 -3.79484147e-01 9.16697145e-01
9.62865710e-01 -7.66720116e-01 9.02641058e-01 -1.25598180e+00
-6.36146143e-02 -5.19982457e-01 -3.36773634e-01 -1.87397921e+00
4.07574438e-02 -9.53014195e-01 -4.48831350e-01 -1.64763629e+00
-1.31544724e-01 -9.15957749e-01 2.58845221e-02 -3.65992308e-01
2.33579338e-01 6.48167655e-02 1.14958093e-01 3.17962289e-01
-2.53340542e-01 1.20654571e+00 9.37235713e-01 6.58664033e-02
1.62379190e-01 6.19000673e-01 -7.53649890e-01 8.43186453e-02
6.66664124e-01 2.60565132e-01 -8.31089675e-01 -1.68696597e-01
-2.65214294e-01 6.95565343e-01 2.92344540e-01 -1.20959520e+00
-3.09956074e-03 -1.53686469e-02 9.08939540e-02 -9.33177531e-01
3.57184887e-01 -1.08712292e+00 5.40841341e-01 6.04882360e-01
4.68165547e-01 -8.30338955e-01 7.95732290e-02 1.06406248e+00
3.90520632e-01 3.91867124e-02 7.91224658e-01 5.67300260e-01
-3.60130221e-01 1.04096115e+00 -1.22164214e+00 -2.58722961e-01
1.76414549e+00 -2.96002090e-01 1.93708539e-02 -1.19761348e+00
-1.13448858e+00 1.15223086e+00 -1.24533042e-01 -1.92405596e-01
4.01772559e-01 -1.72946107e+00 -7.40862727e-01 -1.57952383e-01
2.28050515e-01 -8.55569601e-01 3.46828461e-01 1.33346593e+00
-2.55012482e-01 1.31599832e+00 2.87802666e-02 -8.33602905e-01
-1.19616234e+00 4.88318950e-01 6.39225543e-01 3.55569005e-01
5.89566082e-02 3.49818498e-01 3.94530922e-01 -1.17270015e-02
1.65901676e-01 2.53316164e-01 -8.88487548e-02 -1.99586570e-01
1.79957345e-01 1.03021097e+00 -2.73443490e-01 -8.07678044e-01
-2.34673113e-01 2.27403015e-01 3.42765599e-01 -3.12106937e-01
4.91385758e-01 -1.43665946e+00 7.23428428e-02 -2.57814862e-02
1.14351678e+00 3.74210268e-01 -8.62623513e-01 -4.22412664e-01
-5.72328627e-01 -3.50553930e-01 2.75403678e-01 -2.19413161e-01
-9.91277695e-01 5.76409817e-01 8.70053172e-01 3.64322782e-01
4.80467856e-01 4.11824465e-01 8.59508812e-01 4.12539802e-02
1.02427661e+00 -6.78282380e-01 -5.31882286e-01 6.06349945e-01
6.31573260e-01 -4.41035777e-01 -4.59194541e-01 -6.00744367e-01
-1.68198466e-01 8.56909513e-01 3.31053376e-01 7.26434469e-01
1.28567433e+00 1.76366437e-02 -3.54545593e-01 -2.36288130e-01
-6.34974122e-01 -5.92108786e-01 -2.91596562e-01 7.03657031e-01
7.23592639e-02 4.53963101e-01 -5.76878428e-01 8.28893900e-01
2.44682878e-01 -1.76245674e-01 8.47626209e-01 4.26717728e-01
-3.58955353e-01 -1.07062030e+00 -1.51897117e-01 7.63810635e-01
-3.61021638e-01 1.15986109e-01 5.99276423e-01 7.64008045e-01
-1.13838218e-01 1.27531874e+00 3.40167321e-02 -4.71273027e-02
2.59918123e-01 -2.67749518e-01 1.77733451e-01 1.08501524e-01
9.49172676e-01 2.91295469e-01 2.18008310e-01 2.00926736e-01
3.68176311e-01 -5.07502675e-01 -1.50427997e+00 -7.39623547e-01
-2.42327720e-01 7.05627859e-01 2.68355668e-01 1.24430954e+00
5.39103270e-01 6.92689419e-01 9.52816129e-01 -6.33220300e-02
-2.66711980e-01 -4.32749689e-01 -1.25478184e+00 -6.11632824e-01
6.70916617e-01 -8.13946307e-01 -1.77467793e-01 -5.65398395e-01] | [6.154084205627441, 1.424092173576355] |
cea09e51-fe8b-42e3-82f3-ecaf4af5bf18 | a-vision-for-semantically-enriched-data | 2303.01378 | null | https://arxiv.org/abs/2303.01378v1 | https://arxiv.org/pdf/2303.01378v1.pdf | A Vision for Semantically Enriched Data Science | The recent efforts in automation of machine learning or data science has achieved success in various tasks such as hyper-parameter optimization or model selection. However, key areas such as utilizing domain knowledge and data semantics are areas where we have seen little automation. Data Scientists have long leveraged common sense reasoning and domain knowledge to understand and enrich data for building predictive models. In this paper we discuss important shortcomings of current data science and machine learning solutions. We then envision how leveraging "semantic" understanding and reasoning on data in combination with novel tools for data science automation can help with consistent and explainable data augmentation and transformation. Additionally, we discuss how semantics can assist data scientists in a new manner by helping with challenges related to trust, bias, and explainability in machine learning. Semantic annotation can also help better explore and organize large data sources. | ['Horst Samulowitz', 'Sainyam Galhotra', 'Kavitha Srinivas', 'Udayan Khurana'] | 2023-03-02 | null | null | null | null | ['common-sense-reasoning'] | ['reasoning'] | [-6.45216033e-02 5.89261293e-01 -6.91028714e-01 -9.17330444e-01
-3.30521852e-01 -5.81820250e-01 5.94803214e-01 9.73168492e-01
-1.88217610e-01 7.46141076e-01 6.12675786e-01 -3.46012056e-01
-4.28700387e-01 -7.66264021e-01 -6.24797106e-01 6.71033710e-02
3.10181588e-01 8.15890729e-01 -9.33125168e-02 -1.27169907e-01
4.03550386e-01 3.14723581e-01 -1.60472560e+00 3.33833098e-01
9.74037230e-01 7.08915412e-01 -1.18087701e-01 5.45958802e-02
-7.46223569e-01 7.49928892e-01 -3.15274328e-01 -4.59360778e-01
1.47542000e-01 1.86816692e-01 -1.07375801e+00 -3.28889459e-01
2.03669354e-01 1.54605940e-01 1.57473966e-01 8.77808988e-01
8.25035498e-02 4.16224860e-02 1.47882864e-01 -1.62704992e+00
-8.58376503e-01 8.25143576e-01 -7.85905048e-02 -7.75297210e-02
3.01318228e-01 1.41163141e-01 1.06731534e+00 -6.25852048e-01
8.45410347e-01 1.20647359e+00 7.35512257e-01 7.32372105e-01
-1.30215514e+00 -6.01203144e-01 2.20520392e-01 4.51168180e-01
-1.10912430e+00 -4.71797615e-01 6.29877687e-01 -5.79640031e-01
1.02493978e+00 3.59730691e-01 8.62943828e-01 6.45685732e-01
-4.12783146e-01 5.93553483e-01 9.78797257e-01 -4.79702532e-01
3.25895995e-01 7.92962432e-01 5.71706414e-01 7.47515559e-01
7.60307610e-01 1.82461347e-02 -7.56790698e-01 -1.29869595e-01
4.09525216e-01 1.81127921e-01 1.98923856e-01 -4.79040504e-01
-1.19723558e+00 9.36060488e-01 4.13950741e-01 2.35576600e-01
-3.10234129e-01 1.39921010e-01 2.96787560e-01 2.51667589e-01
4.46392596e-01 1.38104701e+00 -1.12640047e+00 -2.27523148e-01
-4.66322899e-01 2.97829121e-01 1.14854515e+00 1.27138257e+00
1.08485055e+00 -1.65583611e-01 3.74195546e-01 6.30141199e-01
5.53254366e-01 2.32687756e-01 4.48619992e-01 -1.12274122e+00
2.98026711e-01 1.31808031e+00 2.28804350e-01 -8.32692564e-01
-7.81120896e-01 -2.52312303e-01 -3.01435918e-01 9.17943418e-02
3.77479196e-01 2.66157859e-03 -8.66845250e-01 1.56107211e+00
5.70969522e-01 -7.47542381e-02 5.15064716e-01 7.25749493e-01
8.33285391e-01 -2.92715561e-02 5.66312313e-01 1.92035586e-01
1.38878298e+00 -4.40410316e-01 -8.35371912e-01 -5.54104269e-01
1.02959633e+00 -5.11751533e-01 1.35168111e+00 2.14860380e-01
-8.32018435e-01 -9.12733376e-02 -9.83115256e-01 -6.19635880e-01
-9.18113649e-01 -5.19804835e-01 1.24480128e+00 5.46948433e-01
-6.29240334e-01 6.14780426e-01 -7.67271698e-01 -8.53087127e-01
9.09245133e-01 2.64839709e-01 -5.26803911e-01 -9.75415483e-02
-1.23337007e+00 1.21941435e+00 7.08883643e-01 -5.96770287e-01
-5.88032961e-01 -1.44951463e+00 -9.20620382e-01 -5.59331514e-02
6.91299915e-01 -1.06224430e+00 1.05973291e+00 -7.32761323e-01
-7.35389113e-01 8.63508403e-01 -2.40738228e-01 -6.55987144e-01
-2.84020416e-03 -4.17829603e-01 -5.22066414e-01 -5.86801410e-01
-8.08885396e-02 6.74966335e-01 1.63845181e-01 -1.10634458e+00
-6.68034077e-01 -9.15964246e-01 -1.81495592e-01 2.95473207e-02
-5.17286181e-01 9.14851055e-02 -4.46178690e-02 -3.47826272e-01
2.04544097e-01 -6.32071137e-01 -3.10534149e-01 1.47059336e-01
-2.08939880e-01 -2.99351454e-01 9.88612175e-01 -5.60130715e-01
1.20749927e+00 -1.78108776e+00 -1.54087633e-01 4.58395600e-01
4.90395874e-01 6.62378073e-02 2.15881735e-01 4.00949806e-01
9.86061767e-02 8.06392848e-01 2.84575403e-01 1.04667917e-01
2.05582768e-01 5.76907992e-01 -2.73713350e-01 -2.57492423e-01
1.84111431e-01 9.20094371e-01 -9.83572781e-01 -3.06521773e-01
2.43445978e-01 2.59623051e-01 -6.87976182e-01 2.27405325e-01
-7.68460870e-01 3.42775702e-01 -7.03363955e-01 5.03564775e-01
1.45195603e-01 -6.52646005e-01 3.82770240e-01 -2.64820665e-01
-1.04464859e-01 5.99765778e-01 -1.08237767e+00 1.72319126e+00
-4.96048510e-01 4.86914992e-01 -1.49543196e-01 -6.39942348e-01
1.10284829e+00 1.37975857e-01 3.77910286e-01 -5.09932816e-01
-2.63680249e-01 1.35797635e-01 -5.06814383e-02 -8.51690769e-01
4.60383147e-01 -1.86862245e-01 2.42930561e-01 6.79876745e-01
-1.43183842e-01 -3.16795558e-01 -1.22423209e-01 2.43979603e-01
9.66425657e-01 1.20538630e-01 5.55418909e-01 -3.25984061e-01
-1.19125672e-01 8.35790932e-01 6.76992178e-01 4.85610455e-01
1.96455672e-01 2.06251591e-02 2.37052917e-01 -8.54217410e-01
-1.19152212e+00 -6.21136963e-01 1.76302437e-02 1.33155382e+00
-1.60687581e-01 -8.52634907e-01 -4.19256985e-01 -6.76644444e-01
4.29720759e-01 1.21163905e+00 -5.79341888e-01 -3.45506728e-01
1.01463012e-02 -5.15255749e-01 3.46612513e-01 8.54482710e-01
2.34572992e-01 -7.15108752e-01 -5.15632212e-01 1.12633482e-01
1.11410372e-01 -1.05284381e+00 2.26083905e-01 2.05739617e-01
-1.06800163e+00 -1.22278011e+00 5.21268487e-01 -3.15119475e-01
6.40130818e-01 1.41480371e-01 1.41667020e+00 2.30822355e-01
-3.48759502e-01 4.62391526e-01 -3.23443830e-01 -1.24447179e+00
-6.59976423e-01 1.43505305e-01 5.03354780e-02 -6.90636396e-01
1.15690351e+00 -3.86220723e-01 -3.56006652e-01 3.87699872e-01
-9.17941272e-01 4.35155183e-01 2.77009219e-01 5.84239364e-01
5.13803542e-01 -2.34569639e-01 7.97921896e-01 -1.51993012e+00
6.38468504e-01 -7.43182063e-01 -5.05883992e-01 6.29156530e-01
-1.74596989e+00 4.20568585e-01 3.00029457e-01 9.44006369e-02
-1.23229647e+00 -1.47995487e-01 2.26537138e-01 -1.59541696e-01
-3.75295728e-01 7.23653376e-01 -3.59469354e-01 1.62588298e-01
9.84516799e-01 -5.90060830e-01 3.12359720e-01 -6.69919014e-01
7.68336475e-01 6.54537380e-01 1.85111895e-01 -6.95984423e-01
5.35566568e-01 5.16675413e-01 -2.25221347e-02 -3.20394397e-01
-1.05972695e+00 -2.41365358e-01 -5.05942822e-01 4.19918031e-01
7.68733561e-01 -7.94139087e-01 -6.82825387e-01 -2.43837684e-01
-7.45132685e-01 -2.30544358e-01 -9.00423050e-01 4.47470307e-01
-3.76040071e-01 -3.06211263e-01 1.76052660e-01 -4.93048400e-01
-1.58470646e-01 -6.66829944e-01 5.47151864e-01 5.01593292e-01
-8.22464526e-01 -1.37415504e+00 -8.84958506e-02 7.69721091e-01
6.16343260e-01 -1.89841278e-02 1.23829746e+00 -1.52168787e+00
-6.06432915e-01 -2.06216320e-01 -1.90530598e-01 1.39406517e-01
2.48618171e-01 6.21897168e-03 -1.07794452e+00 3.28996748e-01
-3.26004654e-01 -3.19316387e-01 1.96701050e-01 -3.91009171e-03
1.50186253e+00 -5.64952195e-01 -5.08859992e-01 5.24016798e-01
1.12723815e+00 8.46211761e-02 1.91524684e-01 4.74370807e-01
7.77899206e-01 1.06961823e+00 8.75140190e-01 5.20483673e-01
8.19849253e-01 2.78784662e-01 1.87656328e-01 -1.29891336e-01
6.53494522e-02 -5.39918244e-01 -4.94490027e-01 1.91564694e-01
1.23190634e-01 2.98559695e-01 -1.49206793e+00 4.84556973e-01
-2.27393270e+00 -7.92041004e-01 -4.07379540e-03 1.99259031e+00
1.04325426e+00 -1.33878127e-01 -1.89534336e-01 -1.82332143e-01
3.64591777e-01 -3.94271493e-01 -9.74924028e-01 -4.69956189e-01
-6.05240315e-02 -1.46176443e-02 6.05444372e-01 4.65255290e-01
-6.42452121e-01 1.22358084e+00 6.51262283e+00 2.18113542e-01
-6.59739614e-01 -3.37805715e-03 4.40790683e-01 -1.62154641e-02
-1.01352406e+00 4.86696929e-01 -8.03121626e-01 4.66391584e-03
8.05448890e-01 -5.56550145e-01 6.39632404e-01 1.11602879e+00
3.87271643e-01 -1.08206645e-02 -1.65743089e+00 7.51824021e-01
-9.20116156e-02 -1.88343287e+00 2.90655326e-02 -9.02940184e-02
6.76758707e-01 1.04703240e-01 -2.05924347e-01 5.09505756e-02
9.77767289e-01 -1.37084341e+00 1.60809994e-01 8.75178874e-01
4.77971435e-01 -4.71959502e-01 5.39827049e-01 2.00430080e-01
-5.62007368e-01 -2.49186754e-01 -2.90200233e-01 -2.86435425e-01
-4.17177975e-01 5.83370924e-01 -1.58551300e+00 4.15018767e-01
7.69774020e-01 1.02547061e+00 -7.54564226e-01 8.48184168e-01
-1.76462829e-01 4.62209910e-01 -3.35452914e-01 2.72455271e-02
-3.44360858e-01 9.57389770e-04 2.70637214e-01 8.69811833e-01
-1.86912920e-02 -2.27049962e-02 1.08048327e-01 9.97527182e-01
-2.27666408e-01 -2.75467690e-02 -8.70763183e-01 -5.51230669e-01
9.82810736e-01 1.05623794e+00 -1.07775725e-01 -4.52818513e-01
-5.92655361e-01 5.07888235e-02 2.93126762e-01 1.95373908e-01
-7.03596324e-02 -5.01880199e-02 1.13057446e+00 1.47532851e-01
-4.75967944e-01 -9.61766765e-02 -1.07506680e+00 -1.03906500e+00
-2.34251440e-01 -1.05833733e+00 9.34687197e-01 -9.67700362e-01
-1.49394727e+00 -2.79751886e-02 1.04759991e-01 -6.66533947e-01
-3.92988592e-01 -4.89069313e-01 -1.71749234e-01 6.93290949e-01
-1.36377120e+00 -1.40242529e+00 -5.35454929e-01 5.01762211e-01
2.33586252e-01 -3.31923217e-01 1.08412051e+00 -1.45381421e-01
-2.72621483e-01 2.01909646e-01 -9.22880322e-02 -2.05633044e-01
9.57729936e-01 -1.33731294e+00 4.45147991e-01 2.66906857e-01
1.49881959e-01 1.08704054e+00 8.25018585e-01 -9.37113047e-01
-1.56202984e+00 -1.06037366e+00 7.33139753e-01 -9.85238791e-01
9.06603217e-01 -1.48590654e-01 -1.17833984e+00 1.05572331e+00
-3.79120409e-02 -3.60064283e-02 1.23357666e+00 8.60300958e-01
-7.16503263e-01 -3.47755104e-01 -1.46581888e+00 5.35209358e-01
1.16029382e+00 -5.07810652e-01 -8.39098096e-01 4.10189599e-01
9.59540844e-01 -6.68701157e-02 -1.22095847e+00 3.28975379e-01
4.98205870e-01 -3.91616374e-01 7.50549972e-01 -1.69178796e+00
1.22337319e-01 -4.01324749e-01 -2.72275925e-01 -1.32753277e+00
-3.14130895e-02 -7.93425798e-01 -1.00851849e-01 1.22634494e+00
7.73378193e-01 -6.92791998e-01 8.42194498e-01 1.92708564e+00
-5.37010580e-02 -3.13457727e-01 -1.81936815e-01 -4.06486183e-01
-1.01030707e-01 -8.47822547e-01 1.37466931e+00 1.39781952e+00
2.94828206e-01 2.90063113e-01 -6.52455725e-03 9.68136936e-02
5.08475184e-01 1.26439124e-01 1.03115380e+00 -1.74154449e+00
1.45688161e-01 -1.79284334e-01 -5.05156398e-01 -2.29792088e-01
-6.56473711e-02 -1.14229381e+00 -5.75293243e-01 -1.77111566e+00
4.26479220e-01 -9.37512040e-01 -2.35677838e-01 9.63707626e-01
-1.48663417e-01 -3.58997315e-01 3.80872153e-02 1.01345249e-01
-4.72630858e-01 7.19185099e-02 9.12453115e-01 1.53642192e-01
-2.02844933e-01 -3.87304753e-01 -1.44077194e+00 9.95727003e-01
9.58042741e-01 -5.19888520e-01 -5.57428241e-01 -7.06683695e-01
7.57774711e-01 -4.73416775e-01 5.03218114e-01 -6.45630956e-01
4.36082214e-01 -7.04985917e-01 4.76286799e-01 -6.03630506e-02
7.76367355e-03 -1.24151254e+00 3.65395278e-01 1.30570486e-01
-7.60054886e-01 1.27213802e-02 2.12613821e-01 5.45981109e-01
1.90291516e-02 -1.73085764e-01 3.98735881e-01 -2.58836865e-01
-9.31098282e-01 2.33869664e-02 2.42151588e-01 3.89884084e-01
1.17182255e+00 -2.53713652e-02 -8.15407932e-01 -1.74348876e-01
-9.88937914e-01 6.42796159e-01 7.55221069e-01 7.22051144e-01
4.11896616e-01 -9.73510265e-01 -4.95587617e-01 3.24612677e-01
5.13004959e-01 4.57348339e-02 -8.93624425e-02 4.63680059e-01
-1.10355340e-01 4.15581018e-01 -2.18723476e-01 -3.19543153e-01
-1.25061345e+00 5.43008506e-01 1.53551981e-01 1.68169081e-01
-2.98116535e-01 5.14158189e-01 -2.83565491e-01 -5.17906010e-01
2.96067506e-01 -2.80043811e-01 -4.55199629e-02 -4.74890247e-02
5.32741308e-01 4.13919151e-01 5.31704538e-02 1.67599648e-01
-4.82249141e-01 1.06613591e-01 -2.27625459e-01 1.44587979e-01
1.61221039e+00 -1.35415524e-01 -1.66657686e-01 5.85018933e-01
6.12748563e-01 -1.16145834e-01 -9.27251577e-01 -3.68395269e-01
7.27477729e-01 -6.66353762e-01 2.73484178e-02 -1.45046473e+00
-5.84955573e-01 9.36514854e-01 3.21997762e-01 3.57928246e-01
7.99746037e-01 4.99244064e-01 2.88287640e-01 7.43843853e-01
2.94375062e-01 -1.33254552e+00 -1.93393826e-01 2.21635014e-01
8.17372322e-01 -1.76847649e+00 1.86604261e-01 -3.58914912e-01
-9.03638303e-01 1.08584940e+00 9.02962506e-01 5.40195584e-01
7.39489973e-01 3.96820605e-01 1.62002981e-01 -5.93765795e-01
-1.16056561e+00 -2.31806219e-01 3.83212596e-01 8.90612602e-01
6.21626139e-01 1.57862946e-01 5.32438755e-02 8.59551370e-01
-3.27163160e-01 3.84489536e-01 2.42926270e-01 9.29233015e-01
-5.28951645e-01 -1.39590287e+00 -1.77044287e-01 9.52891827e-01
-2.52020478e-01 -1.69390351e-01 -7.39657760e-01 9.12144840e-01
1.32192940e-01 9.99987960e-01 1.09170020e-01 -4.05558407e-01
2.86934406e-01 3.89879644e-01 9.79827493e-02 -1.03449368e+00
-5.24254918e-01 -5.46558857e-01 5.15514553e-01 -7.19749391e-01
-3.16342771e-01 -5.19060910e-01 -1.63423312e+00 -3.56626719e-01
-1.18008427e-01 3.22808862e-01 1.17493737e+00 1.05816984e+00
9.78038371e-01 3.22542012e-01 1.91818904e-02 3.28233808e-01
-4.46998000e-01 -6.26866400e-01 -4.18400951e-02 6.20633781e-01
4.37316077e-04 -3.63041997e-01 3.43637057e-02 1.87428296e-01] | [8.97663402557373, 7.307485103607178] |
5aa6fa9e-04c1-4798-945d-673309eb1a82 | clustering-piecewise-stationary-processes | 1906.10921 | null | https://arxiv.org/abs/1906.10921v1 | https://arxiv.org/pdf/1906.10921v1.pdf | Clustering piecewise stationary processes | The problem of time-series clustering is considered in the case where each data-point is a sample generated by a piecewise stationary ergodic process. Stationary processes are perhaps the most general class of processes considered in non-parametric statistics and allow for arbitrary long-range dependence between variables. Piecewise stationary processes studied here for the first time in the context of clustering, relax the last remaining assumption in this model: stationarity. A natural formulation is proposed for this problem and a notion of consistency is introduced which requires the samples to be placed in the same cluster if and only if the piecewise stationary distributions that generate them have the same set of stationary distributions. Simple, computationally efficient algorithms are proposed and are shown to be consistent without any additional assumptions beyond piecewise stationarity. | ['Azadeh Khaleghi', 'Daniil Ryabko'] | 2019-06-26 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 1.25088662e-01 -2.82368720e-01 -1.78917661e-01 -2.72195399e-01
-5.39052546e-01 -6.97523117e-01 7.95130432e-01 4.41267043e-01
-2.54409194e-01 7.35843778e-01 -9.66077112e-03 -1.70183063e-01
-4.91477460e-01 -5.19592643e-01 -4.27524090e-01 -1.22388840e+00
-4.92283911e-01 1.03348815e+00 1.40093014e-01 2.41225004e-01
7.20910132e-02 5.93602180e-01 -1.38836443e+00 -5.77092230e-01
7.35575914e-01 5.38554132e-01 5.80591112e-02 9.05176401e-01
-1.34124517e-01 3.28485817e-01 -6.32325530e-01 -6.23267032e-02
1.50105849e-01 -4.06524926e-01 -5.54524243e-01 3.07215840e-01
-3.00220907e-01 2.35323116e-01 6.85267448e-02 1.06989264e+00
1.56401724e-01 4.82140809e-01 1.14770103e+00 -1.68726647e+00
-2.96289146e-01 4.18813527e-01 -7.01282144e-01 3.34672272e-01
7.48145655e-02 -2.28160173e-01 6.35970771e-01 -4.37744945e-01
1.68696105e-01 1.27937484e+00 7.09898770e-01 2.05015227e-01
-1.70822310e+00 -2.00116247e-01 5.54640368e-02 -1.73452064e-01
-1.35294235e+00 -2.46239170e-01 5.30146956e-01 -4.42553729e-01
3.84170562e-01 4.61438358e-01 5.00255167e-01 8.82788122e-01
4.29406583e-01 3.96308988e-01 1.11016428e+00 -3.69925201e-01
6.98273540e-01 -3.85560751e-01 6.02848411e-01 -1.34671759e-02
5.75327516e-01 -2.78817564e-01 9.62533504e-02 -7.24316537e-01
8.33714068e-01 2.92644739e-01 1.26954466e-01 -5.60575008e-01
-1.20891392e+00 7.92762995e-01 -4.25926447e-01 6.26125693e-01
-5.57757080e-01 1.78457215e-01 2.96921909e-01 2.10025385e-01
5.79992831e-01 7.05157593e-02 -3.54528695e-01 -2.63729006e-01
-9.06487286e-01 1.61106870e-01 1.13839221e+00 1.20176470e+00
4.41764265e-01 -2.95966230e-02 -2.03444496e-01 4.48150337e-01
-4.19352502e-02 7.63842881e-01 2.30444580e-01 -9.15953994e-01
1.71796367e-01 -1.47469118e-01 6.46856844e-01 -5.54933548e-01
-2.75433004e-01 -3.00804347e-01 -8.80019009e-01 -1.65414542e-01
6.22652471e-01 -8.55656192e-02 -7.37149954e-01 1.70659411e+00
2.79454678e-01 5.24299145e-01 5.35090193e-02 1.33863941e-01
-1.76616564e-01 8.77940297e-01 2.84395099e-01 -8.56503904e-01
9.92676914e-01 -4.17910457e-01 -7.74029434e-01 6.09739959e-01
-1.17092028e-01 -7.60492027e-01 6.86861098e-01 2.84488022e-01
-1.20438659e+00 -4.34595525e-01 -4.14483190e-01 3.90710652e-01
7.77582601e-02 -4.20756519e-01 4.80141193e-01 5.04511595e-01
-1.13960123e+00 6.88413322e-01 -1.16189289e+00 -5.95400214e-01
-2.40135238e-01 3.90875667e-01 2.92312857e-02 1.76224947e-01
-5.97189426e-01 3.31877410e-01 1.78995103e-01 5.24273235e-03
-7.76152074e-01 -5.02131045e-01 -3.25261623e-01 2.16872871e-01
5.11136353e-02 -5.61226249e-01 1.21236992e+00 -1.14387357e+00
-1.23217738e+00 6.36469543e-01 -7.88370192e-01 -4.31196243e-01
7.61585772e-01 1.50512680e-01 -6.78745091e-01 7.73461461e-02
4.14200991e-01 -1.61331385e-01 1.08248210e+00 -1.42039371e+00
-5.56090355e-01 -2.92134017e-01 -5.48084617e-01 -1.58803854e-02
1.20226867e-01 3.36327672e-01 -4.51147825e-01 -6.15715444e-01
1.67766854e-01 -1.24742889e+00 -3.83530259e-01 -5.87497532e-01
-4.61754590e-01 -5.83397150e-01 7.95066059e-01 -4.54409361e-01
9.78217602e-01 -2.21314120e+00 -1.37614021e-02 5.62504768e-01
-4.48208414e-02 -2.40817592e-01 1.82082251e-01 8.04334879e-01
-3.67914915e-01 2.36280099e-01 -6.34222388e-01 -7.07552254e-01
1.78131517e-02 4.99854386e-01 -3.83597016e-01 9.32247519e-01
-1.32359713e-02 3.27186346e-01 -1.08343327e+00 -4.06121105e-01
3.14256996e-01 2.09292993e-01 9.29185376e-02 9.69949737e-02
1.57360416e-02 7.92960167e-01 -2.89575547e-01 2.29085654e-01
7.40269065e-01 -1.03344209e-01 3.02442200e-02 5.99160075e-01
-3.75222951e-01 -2.19807252e-01 -1.43770599e+00 8.75253260e-01
-1.86934635e-01 4.00878966e-01 1.17839083e-01 -1.07779682e+00
8.19962680e-01 6.74092114e-01 1.11430323e+00 1.73709035e-01
1.06419958e-01 7.98343793e-02 3.29360031e-02 -3.31228644e-01
3.95380199e-01 -6.68837368e-01 -3.35049868e-01 4.31248635e-01
-1.89893991e-01 -6.35278150e-02 2.85016298e-01 -1.68455750e-01
9.58010554e-01 -2.19561651e-01 4.20321524e-01 -7.39926457e-01
3.13679814e-01 2.22109333e-02 6.95030391e-01 9.77890193e-01
-2.28906065e-01 4.70280021e-01 4.78877276e-01 6.64718598e-02
-1.31212413e+00 -1.54010677e+00 -3.39520544e-01 9.40149367e-01
1.24957800e-01 1.61500424e-01 -4.49580610e-01 1.13470219e-01
-1.69686731e-02 6.65244520e-01 -7.33160913e-01 3.11151505e-01
-4.01218325e-01 -8.37658048e-01 2.23313138e-01 3.41145456e-01
1.45856008e-01 -6.53332174e-01 -1.82531148e-01 5.06212533e-01
-8.33866075e-02 -8.15866470e-01 -3.90141279e-01 1.90077752e-01
-1.27008688e+00 -8.62445474e-01 -8.84585977e-01 -8.38349283e-01
6.96455002e-01 2.98639297e-01 8.25505376e-01 -2.74890333e-01
1.27290905e-01 7.18922377e-01 -1.94194913e-02 -5.43870926e-01
-5.75552881e-01 -2.19955668e-01 1.45084754e-01 1.83112666e-01
3.23593855e-01 -6.00720108e-01 -1.72518730e-01 2.99205363e-01
-9.83847499e-01 -5.41774809e-01 3.96163343e-03 5.47495782e-01
8.23852956e-01 7.25115955e-01 4.76528853e-01 -6.78139329e-01
5.54361582e-01 -8.58471274e-01 -5.24069369e-01 2.17050165e-01
-8.84735063e-02 -9.92476419e-02 7.47859061e-01 -5.95086575e-01
-9.62567985e-01 1.05755655e-02 4.05921280e-01 -5.15300393e-01
-5.83557129e-01 2.94891417e-01 -3.52993280e-01 4.58902866e-01
6.89105093e-02 3.34477872e-01 2.19943915e-02 -4.16607648e-01
1.32046551e-01 2.29112059e-01 8.15198004e-01 -7.44459748e-01
8.27734053e-01 1.00087059e+00 5.38768232e-01 -1.16592669e+00
-2.36991882e-01 -1.04593647e+00 -8.70155394e-01 6.57250211e-02
8.22665274e-01 -7.46597767e-01 -6.64799869e-01 5.38678885e-01
-9.17479634e-01 -3.36160213e-01 -5.10203242e-01 5.91769516e-01
-9.16937232e-01 5.10233760e-01 -4.23311859e-01 -1.27825582e+00
2.30851024e-01 -5.44421494e-01 7.87326157e-01 -6.20697550e-02
-3.60992283e-01 -1.71145606e+00 3.09904665e-01 -3.54459703e-01
1.59173712e-01 5.40028393e-01 9.44335461e-01 -8.82937372e-01
-4.22821343e-02 -3.94018501e-01 3.38283628e-01 1.82313949e-01
6.23338938e-01 5.78012705e-01 -5.38724780e-01 -2.80325443e-01
4.58428085e-01 8.20210516e-01 3.94007981e-01 9.75579977e-01
1.01777840e+00 -2.28753999e-01 -5.27608931e-01 1.37400106e-01
1.47917652e+00 3.96719962e-01 3.44627976e-01 8.13224763e-02
5.77047348e-01 7.12853253e-01 2.32202992e-01 5.77881813e-01
1.77012771e-01 2.60265082e-01 2.33180765e-02 -1.38726989e-02
6.70088649e-01 -4.57920022e-02 2.23831251e-01 9.75494564e-01
-3.25989664e-01 -3.65872055e-01 -9.48288381e-01 9.47368979e-01
-1.99079728e+00 -1.62284863e+00 -8.73769820e-01 2.74748278e+00
6.15522146e-01 -2.13431474e-02 6.06735706e-01 3.03986669e-01
1.45489168e+00 -2.34906986e-01 -3.17912966e-01 -2.94608265e-01
-2.09438547e-01 2.51348436e-01 8.43136072e-01 6.86811030e-01
-1.19413257e+00 4.24622387e-01 7.43538380e+00 5.90750992e-01
-8.97518754e-01 9.28157121e-02 3.55064392e-01 4.97318432e-02
-2.50702709e-01 1.52107522e-01 -5.06731987e-01 9.02443171e-01
1.19152832e+00 -5.79502881e-01 -1.67894028e-02 6.15068972e-01
7.91273296e-01 -3.46051186e-01 -1.18484092e+00 6.79744005e-01
-4.27554607e-01 -5.66109061e-01 -2.40362942e-01 2.66666919e-01
1.01894045e+00 -3.16186190e-01 2.20911875e-01 -1.42530575e-01
9.20975924e-01 -9.55901086e-01 6.31580889e-01 8.61926377e-01
3.14533263e-01 -9.85260844e-01 4.60843116e-01 3.54917169e-01
-1.36501026e+00 3.48181836e-02 -3.48459423e-01 -2.07046881e-01
5.75130641e-01 9.79441643e-01 -5.08118808e-01 6.02355897e-01
4.37716752e-01 7.36254632e-01 -1.36033237e-01 1.42480040e+00
1.75213531e-01 1.05371130e+00 -5.14472485e-01 1.50296301e-01
2.13964850e-01 -6.25357091e-01 7.63377905e-01 1.15158677e+00
6.05015218e-01 -4.07066010e-02 3.03820729e-01 3.21908832e-01
4.81822044e-01 -3.29048485e-02 -6.03498220e-01 2.39142910e-01
7.09053278e-01 8.98825109e-01 -1.22247207e+00 -6.08590961e-01
-3.32368672e-01 8.72154593e-01 -3.50156963e-01 5.44493496e-01
-7.12794900e-01 5.89308143e-02 5.15027940e-01 2.64613360e-01
2.81390250e-01 -7.63153732e-01 -4.50193197e-01 -8.40396225e-01
-1.81766748e-01 -3.70075516e-02 5.24873734e-01 -3.97053093e-01
-1.89943469e+00 -5.10657625e-03 5.15853465e-01 -1.14278305e+00
-3.40308845e-01 -1.13085553e-01 -1.01095998e+00 1.16450167e+00
-9.74744499e-01 -7.96753705e-01 8.94051343e-02 1.08140349e+00
4.39684093e-01 2.13308975e-01 6.51742995e-01 -2.98536181e-01
-5.13575852e-01 -2.42556825e-01 9.88916636e-01 -2.60827541e-01
4.11704093e-01 -1.54642224e+00 2.69200772e-01 8.67137969e-01
-1.67464450e-01 8.02037954e-01 1.23433495e+00 -9.14176881e-01
-1.02742743e+00 -1.27140152e+00 9.53058183e-01 -4.69379544e-01
1.06325316e+00 -2.42477059e-01 -1.15894389e+00 7.48986244e-01
6.62573725e-02 -3.63869816e-01 8.82499456e-01 1.27959341e-01
2.20365986e-01 2.98395008e-01 -1.09477854e+00 4.61683720e-01
6.39233291e-01 -2.85300612e-01 -7.15632439e-01 2.74712414e-01
2.77741134e-01 2.59157032e-01 -9.11463022e-01 2.17743125e-02
1.22551240e-01 -4.78809983e-01 7.42008865e-01 -4.31711048e-01
-2.18112975e-01 -4.86355603e-01 -7.77399167e-02 -1.09350491e+00
-5.32025754e-01 -9.82153058e-01 1.98265284e-01 1.47810984e+00
4.22615893e-02 -8.43593240e-01 4.28740054e-01 7.02310145e-01
1.08191915e-01 6.65026307e-02 -1.36319757e+00 -1.48057044e+00
2.14592010e-01 -3.91035199e-01 6.02949142e-01 1.14222181e+00
-1.48547173e-01 1.14209816e-01 -2.64325380e-01 3.06563973e-01
8.57588589e-01 5.51428273e-02 6.18340790e-01 -1.91573489e+00
-2.97974050e-01 -3.68726909e-01 -1.64458469e-01 -7.61996388e-01
4.11940366e-01 -4.46047902e-01 4.86629158e-01 -1.40676630e+00
2.18618438e-01 -7.81254232e-01 1.61835272e-02 -2.29337350e-01
-1.55625850e-01 -3.44257116e-01 -1.52970672e-01 8.80082488e-01
-3.53782743e-01 4.12795782e-01 6.99891269e-01 4.38357323e-01
-4.16743517e-01 7.15443611e-01 -2.88953602e-01 6.57278121e-01
9.44585741e-01 -4.99292552e-01 -5.74809790e-01 1.47956580e-01
-3.56541067e-01 3.69320899e-01 4.41483766e-01 -8.48145962e-01
1.62291467e-01 -4.88674909e-01 1.79085076e-01 -5.85917592e-01
3.11451733e-01 -1.06663346e+00 8.03447068e-01 3.26489747e-01
-1.78154036e-01 4.79738235e-01 1.32068796e-02 1.07654881e+00
-2.31361259e-02 -3.50292027e-01 7.96281576e-01 5.37290648e-02
-3.72284591e-01 3.92503530e-01 -8.72508645e-01 -5.63435964e-02
1.42741525e+00 -2.28437707e-01 2.61557966e-01 -6.47394419e-01
-1.20252228e+00 8.43155459e-02 6.29366457e-01 -1.95578672e-02
-9.90420207e-02 -1.11979020e+00 -4.36857164e-01 -2.58929938e-01
-2.57999688e-01 4.50833552e-02 1.58247575e-01 9.38919902e-01
-6.26742020e-02 3.09044093e-01 -5.35724573e-02 -9.05257046e-01
-1.22365034e+00 9.28585708e-01 9.21211615e-02 -2.25054279e-01
-5.27655900e-01 1.74927622e-01 1.31216705e-01 8.93034935e-02
-2.25306153e-01 -3.83284301e-01 8.69974941e-02 1.07368723e-01
2.50248075e-01 7.95857906e-01 -1.89318031e-01 -8.37282419e-01
-1.29224151e-01 3.42603207e-01 3.28487068e-01 -6.99838042e-01
1.25964844e+00 -4.17741030e-01 -4.72051263e-01 1.47378588e+00
9.37675714e-01 7.94713497e-02 -1.39615285e+00 -2.25369766e-01
3.39585513e-01 -2.70916760e-01 -7.79180884e-01 1.71170961e-02
-4.19755757e-01 6.01333618e-01 1.87806264e-01 8.75369668e-01
9.18651044e-01 1.80690154e-01 1.30403802e-01 -2.21439511e-01
5.39072216e-01 -1.04841924e+00 -3.33133608e-01 4.82360125e-01
5.58990180e-01 -7.03508973e-01 -2.61467993e-01 -6.28178775e-01
-5.39404511e-01 1.04410088e+00 -1.07022204e-01 -6.29195809e-01
9.03962195e-01 1.67509317e-01 -5.17383397e-01 1.47268146e-01
-4.65215176e-01 -2.45874286e-01 -1.00218900e-01 9.81699765e-01
4.24138099e-01 4.44665879e-01 -4.37722474e-01 3.01866353e-01
-1.71822131e-01 -2.39059597e-01 5.94540715e-01 1.03942740e+00
-3.87463897e-01 -9.25828159e-01 -7.64286280e-01 4.45517689e-01
-6.57417059e-01 5.60179949e-02 3.81881371e-02 8.47676575e-01
-2.05883145e-01 1.15513384e+00 5.62795281e-01 4.43707794e-01
8.89903158e-02 2.76774585e-01 3.10343951e-01 -7.38115907e-01
-3.03849429e-01 5.43149769e-01 -3.21546376e-01 1.89677820e-01
-6.67033136e-01 -1.36381423e+00 -1.30620313e+00 -8.06435227e-01
-6.30578399e-02 5.42683780e-01 4.81914192e-01 9.21734631e-01
-7.57489353e-02 2.77104527e-01 8.98956716e-01 -6.12276673e-01
-3.91745180e-01 -7.91562498e-01 -9.91216063e-01 3.41318399e-01
3.66809547e-01 -5.09257257e-01 -5.14950573e-01 5.37062585e-01] | [7.064774513244629, 3.892709970474243] |
f71667f4-5e34-4630-99cb-dcaa6c59f235 | semi-automatic-construction-of-word-formation | null | null | https://aclanthology.org/L18-1291 | https://aclanthology.org/L18-1291.pdf | Semi-Automatic Construction of Word-Formation Networks (for Polish and Spanish) | null | ["Zden{\\v{e}}k {\\v{Z}}abokrtsk{\\'y}", "Magda {\\v{S}}ev{\\v{c}}{\\'\\i}kov{\\'a}", 'Mateusz Lango'] | 2018-05-01 | semi-automatic-construction-of-word-formation-1 | https://aclanthology.org/L18-1291 | https://aclanthology.org/L18-1291.pdf | lrec-2018-5 | ['sequential-pattern-mining'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.27558708190918, 3.759993076324463] |
987cca09-3024-4333-8c2a-993d2b1fcd27 | twist-two-way-inter-label-self-training-for | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Chu_TWIST_Two-Way_Inter-Label_Self-Training_for_Semi-Supervised_3D_Instance_Segmentation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Chu_TWIST_Two-Way_Inter-Label_Self-Training_for_Semi-Supervised_3D_Instance_Segmentation_CVPR_2022_paper.pdf | TWIST: Two-Way Inter-Label Self-Training for Semi-Supervised 3D Instance Segmentation | We explore the way to alleviate the label-hungry problem in a semi-supervised setting for 3D instance segmentation. To leverage the unlabeled data to boost model performance, we present a novel Two-Way Inter-label Self-Training framework named TWIST. It exploits inherent correlations between semantic understanding and instance information of a scene. Specifically, we consider two kinds of pseudo labels for semantic- and instance-level supervision. Our key design is to provide object-level information for denoising pseudo labels and make use of their correlation for two-way mutual enhancement, thereby iteratively promoting the pseudo-label qualities. TWIST attains leading performance on both ScanNet and S3DIS, compared to recent 3D pre-training approaches, and can cooperate with them to further enhance performance, e.g., +4.4% AP50 on 1%-label ScanNet data-efficient benchmark. Code is available at https://github.com/dvlab-research/TWIST. | ['Jiaya Jia', 'Chi-Wing Fu', 'Xiaojuan Qi', 'Xiao Tan', 'Zhengzhe Liu', 'Xiaoqing Ye', 'Ruihang Chu'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['3d-instance-segmentation-1'] | ['computer-vision'] | [ 3.81183088e-01 5.11381865e-01 -5.82379460e-01 -8.98170650e-01
-9.73696291e-01 -6.13721848e-01 3.85591358e-01 6.95380270e-02
-2.85815120e-01 1.66884959e-01 -5.94838038e-02 -2.67529666e-01
1.90953031e-01 -4.94999796e-01 -7.84738898e-01 -5.52911460e-01
1.12678774e-01 6.10740483e-01 3.49828869e-01 1.16994306e-01
1.51278786e-02 2.70080715e-01 -1.44116592e+00 2.52531976e-01
8.29436481e-01 1.13274229e+00 3.51686984e-01 2.62341857e-01
-2.29733050e-01 8.99228036e-01 4.30149138e-02 -1.80109903e-01
4.26805496e-01 -1.39495790e-01 -1.20316923e+00 5.84724128e-01
4.90036368e-01 -2.56450534e-01 -1.85032219e-01 1.08620417e+00
4.50634599e-01 9.32146162e-02 5.96664071e-01 -9.53290880e-01
-3.57809156e-01 5.53566754e-01 -9.41786110e-01 7.14572659e-03
-7.39676654e-02 2.46241763e-01 9.99286294e-01 -9.08925712e-01
7.05663145e-01 1.29023194e+00 5.50244808e-01 6.60909474e-01
-1.38020205e+00 -7.23847270e-01 3.42350304e-01 -1.11407889e-02
-1.15293109e+00 -3.30891252e-01 9.67698216e-01 -3.27742845e-01
6.03010893e-01 -4.01519574e-02 4.22180146e-01 1.03014493e+00
-4.43903118e-01 1.32791102e+00 1.48248613e+00 -1.95425078e-01
7.66250491e-02 8.85907486e-02 3.64764750e-01 8.95997643e-01
-2.75811374e-01 1.06332645e-01 -5.56769848e-01 1.73766136e-01
6.70251787e-01 -1.44981071e-01 -4.54904847e-02 -4.78165239e-01
-1.13405776e+00 6.02355063e-01 7.84872890e-01 -8.01209826e-03
-1.55992389e-01 1.11181088e-01 2.47479171e-01 1.23159334e-01
1.09354556e+00 3.93255413e-01 -7.78816998e-01 5.00824414e-02
-9.29985344e-01 -9.29141641e-02 5.16418278e-01 1.12714827e+00
1.05284119e+00 -1.63453504e-01 -2.21277416e-01 1.08476377e+00
6.10008061e-01 5.95424950e-01 -1.18585989e-01 -1.35987079e+00
3.38136435e-01 6.08451366e-01 -2.55107701e-01 -4.37552810e-01
-4.28475201e-01 -7.95524359e-01 -6.42102420e-01 1.20495364e-01
3.43350112e-01 1.90586999e-01 -1.59512866e+00 1.71990168e+00
8.52757096e-01 5.78966320e-01 -2.37645328e-01 9.45925772e-01
1.02487576e+00 3.54962081e-01 2.69873708e-01 2.06810266e-01
1.35780549e+00 -1.47161162e+00 -3.83215934e-01 -4.41060871e-01
8.50139558e-01 -6.77179396e-01 1.17521632e+00 2.55187631e-01
-9.78480518e-01 -7.31589794e-01 -8.19862843e-01 -1.83948562e-01
-1.66719675e-03 -6.44831881e-02 5.85295916e-01 4.14894640e-01
-1.10447645e+00 5.78682482e-01 -1.00323343e+00 -8.55835062e-03
9.55958784e-01 1.55330464e-01 -2.02650189e-01 -4.50056463e-01
-9.30873156e-01 5.03917098e-01 4.63844657e-01 -1.56653360e-01
-1.26952112e+00 -1.00935745e+00 -8.36886764e-01 -4.44592386e-01
6.88404679e-01 -5.07813454e-01 1.34721005e+00 -7.94585943e-01
-1.23107076e+00 1.56507504e+00 -9.04730856e-02 -1.51424736e-01
5.50380945e-01 -1.34082168e-01 -2.28229575e-02 2.39870191e-01
4.44450110e-01 1.36019063e+00 8.14380825e-01 -1.76126146e+00
-6.54909372e-01 -4.73312289e-01 1.38466239e-01 3.17245126e-01
-3.98122109e-02 -3.55174094e-01 -8.98714960e-01 -6.08031690e-01
5.24857998e-01 -1.16928697e+00 -4.91424412e-01 4.89269048e-02
-6.23017728e-01 -2.78175801e-01 7.37156451e-01 -5.14773548e-01
5.48012495e-01 -2.15527248e+00 8.88613425e-03 1.89066708e-01
5.22036731e-01 2.26898000e-01 -3.04581165e-01 -1.98517993e-01
-3.85326110e-02 1.37787938e-01 -5.94459295e-01 -8.87591720e-01
-7.57184327e-02 4.64120179e-01 1.24013461e-01 4.32436585e-01
6.27405465e-01 1.19512784e+00 -1.13349795e+00 -6.48405433e-01
2.82617658e-01 4.82952356e-01 -4.71514672e-01 2.24526823e-01
-5.27882874e-01 1.05551505e+00 -8.09824228e-01 9.36805427e-01
8.40270817e-01 -7.06531346e-01 7.29274517e-03 -3.05046856e-01
2.27991566e-01 4.31544870e-01 -9.84823585e-01 2.17299128e+00
-5.59549809e-01 3.00419211e-01 2.92405069e-01 -9.20707226e-01
8.36979926e-01 6.43232837e-02 6.24947488e-01 -8.95827949e-01
2.91828930e-01 2.46278882e-01 -5.70009708e-01 -3.35807949e-01
1.10985219e-01 -3.44366580e-02 2.31329307e-01 5.73220134e-01
2.71340489e-01 -3.73293459e-01 3.90656330e-02 3.75113338e-01
1.09130371e+00 4.63701069e-01 -1.79035991e-01 -3.77140641e-01
2.66597629e-01 6.86362684e-02 6.54301047e-01 5.91941059e-01
-1.90478176e-01 6.80282414e-01 1.92542151e-01 -2.10564345e-01
-9.18666363e-01 -1.07917607e+00 -4.01896983e-01 1.02957463e+00
5.79241991e-01 -1.43757477e-01 -6.65538490e-01 -1.17872167e+00
1.06804825e-01 7.67557621e-01 -6.15663171e-01 -2.27397308e-01
-3.38074178e-01 -6.30012035e-01 3.83033991e-01 5.77714682e-01
6.28355801e-01 -8.02118242e-01 -1.49798632e-01 3.68885212e-02
-3.93849134e-01 -1.33501828e+00 -4.34528857e-01 6.75919116e-01
-1.09957552e+00 -8.63508582e-01 -5.66519320e-01 -7.69593835e-01
8.26318502e-01 5.66858232e-01 1.45744538e+00 2.05063596e-01
-1.14064454e-03 3.98744762e-01 -6.18508875e-01 -1.79220602e-01
-4.30398583e-01 2.90728927e-01 -3.36481601e-01 -8.12233165e-02
1.31087974e-01 -6.93171561e-01 -6.17385626e-01 6.61333442e-01
-8.11057568e-01 3.80285025e-01 6.58534229e-01 7.20981896e-01
1.00304043e+00 -3.10072809e-01 3.98169696e-01 -1.13895082e+00
-2.20704645e-01 -5.37235200e-01 -3.74337167e-01 7.80404881e-02
-7.68759608e-01 1.88675418e-01 1.91523451e-02 -2.65481204e-01
-1.21544266e+00 3.22888732e-01 -3.22907865e-01 -4.73643810e-01
-4.09409940e-01 2.27605492e-01 -2.90572882e-01 -3.26405227e-01
4.74565804e-01 -5.12590855e-02 -7.00675324e-02 -6.67971194e-01
6.88981950e-01 5.21683455e-01 4.95429724e-01 -8.93622160e-01
8.63105774e-01 6.69429302e-01 -4.56408598e-02 -3.05900961e-01
-1.65896058e+00 -1.00241351e+00 -7.08808601e-01 -2.87302703e-01
9.76466656e-01 -1.27221859e+00 -2.95704037e-01 6.51425540e-01
-9.58691418e-01 -8.38416159e-01 -4.82322782e-01 2.76727110e-01
-5.90425014e-01 3.66988361e-01 -6.85953319e-01 -4.80783701e-01
-1.00466385e-01 -1.23697603e+00 1.54680073e+00 1.16140001e-01
1.98269442e-01 -9.31777656e-01 -1.32445142e-01 8.66333783e-01
3.60681862e-02 1.85070172e-01 5.45007169e-01 -5.99288762e-01
-6.53061271e-01 2.00961396e-01 -6.67457879e-01 7.21909940e-01
8.62604827e-02 -5.44716597e-01 -1.35017705e+00 -1.76201135e-01
1.42524496e-01 -8.13724995e-01 1.15870142e+00 2.60360360e-01
1.25856090e+00 2.23826692e-01 -3.89806986e-01 8.92987370e-01
1.07470191e+00 -3.18727404e-01 2.08033174e-01 1.44714296e-01
1.18288577e+00 7.21594453e-01 9.96772766e-01 2.42404193e-01
6.85380042e-01 6.55063331e-01 7.58614480e-01 -5.25293887e-01
-6.08513415e-01 -2.80722857e-01 -5.86494729e-02 8.46039057e-01
1.32484987e-01 -2.17093810e-01 -1.08572173e+00 5.75046182e-01
-1.79837310e+00 -3.28600198e-01 -4.52983975e-01 1.72964311e+00
1.14966238e+00 2.64070004e-01 -6.04973361e-02 -1.02930471e-01
6.09866381e-01 3.73866528e-01 -8.81610096e-01 3.96143734e-01
-1.01860635e-01 1.05635606e-01 7.81383514e-01 4.57858264e-01
-1.44988978e+00 1.11628592e+00 4.97585821e+00 1.12423933e+00
-7.43963957e-01 4.86086547e-01 1.20605242e+00 5.90484142e-02
-4.24914181e-01 -5.94632067e-02 -7.58633018e-01 2.62632608e-01
6.73160613e-01 4.80291188e-01 1.40465781e-01 8.92862499e-01
1.11267976e-01 -4.87357713e-02 -1.13080359e+00 8.45344067e-01
-1.13187350e-01 -1.15006399e+00 -2.21243784e-01 6.05975837e-02
1.14174247e+00 5.06625354e-01 5.28576598e-02 1.24137878e-01
5.01757205e-01 -6.88803792e-01 8.66718829e-01 1.51788816e-01
7.77104557e-01 -5.75290740e-01 6.30592406e-01 3.27930272e-01
-1.22460389e+00 2.18746230e-01 -7.09801987e-02 3.02516639e-01
4.41431046e-01 1.09095824e+00 -7.23505735e-01 7.96319246e-01
8.02973390e-01 1.13986051e+00 -5.97122669e-01 7.36231744e-01
-5.81687570e-01 9.38330829e-01 -5.21260202e-01 5.87842107e-01
5.99273860e-01 -1.10869072e-01 3.55565071e-01 1.03077328e+00
-9.67333391e-02 2.47190475e-01 5.33372939e-01 8.71094048e-01
-2.48203561e-01 -1.26640499e-01 -1.37501657e-01 2.84315825e-01
3.70909661e-01 1.33897388e+00 -1.08632636e+00 -4.33387369e-01
-3.03683162e-01 1.00265074e+00 3.20381194e-01 1.82682857e-01
-1.01906300e+00 3.70322227e-01 4.17063743e-01 1.76580414e-01
3.04148912e-01 -1.54569715e-01 -5.55845320e-01 -1.08359957e+00
1.51603529e-02 -6.71230018e-01 2.52106756e-01 -6.56119227e-01
-1.31834066e+00 5.40962219e-01 -2.48086005e-02 -1.12300825e+00
2.66047359e-01 -3.84112388e-01 -2.21728370e-01 5.61637878e-01
-1.89075160e+00 -1.56337154e+00 -5.34024715e-01 4.20211345e-01
7.75775135e-01 3.49318326e-01 3.82687688e-01 6.50255501e-01
-3.76430422e-01 4.02078241e-01 -9.57152918e-02 -5.95846139e-02
7.84036458e-01 -1.46664369e+00 7.48811126e-01 6.03331923e-01
4.27586198e-01 -7.54809603e-02 3.38073134e-01 -6.07102454e-01
-8.85184944e-01 -1.42233098e+00 5.54647803e-01 -6.75231457e-01
4.41595912e-01 -3.42633754e-01 -9.42884982e-01 5.32339156e-01
-2.78224647e-01 3.47740620e-01 4.56140280e-01 2.51801163e-02
-6.17118537e-01 8.32441300e-02 -1.13868594e+00 2.83736199e-01
1.68717027e+00 -4.82250512e-01 -3.08011860e-01 6.15517259e-01
1.26685679e+00 -6.40122175e-01 -9.19275522e-01 8.08872521e-01
1.28593907e-01 -7.60109127e-01 1.24611652e+00 -2.44635582e-01
4.94565934e-01 -4.06542510e-01 -1.62718922e-01 -1.12734938e+00
-2.63549328e-01 -3.48487079e-01 2.76625212e-02 1.35606658e+00
5.42994559e-01 -3.40729415e-01 1.06243706e+00 4.08488870e-01
-5.57124794e-01 -8.77548218e-01 -7.57501781e-01 -7.67296851e-01
-1.61415026e-01 -7.72199035e-01 4.27571952e-01 1.17017531e+00
-6.82021081e-01 3.56328100e-01 -3.26926202e-01 2.43509546e-01
1.04835737e+00 -1.04166940e-02 5.71833014e-01 -1.11021054e+00
-2.49911964e-01 -2.42369995e-01 -1.89981535e-01 -1.56577587e+00
3.92879933e-01 -1.31283331e+00 2.49954715e-01 -1.30107844e+00
1.95730612e-01 -1.04523325e+00 -5.56994259e-01 8.13569546e-01
-3.06027859e-01 7.29186833e-01 1.14537537e-01 3.01334381e-01
-1.04938233e+00 7.38252580e-01 1.53779149e+00 -3.95491093e-01
-1.55773312e-01 4.43240702e-02 -6.05271399e-01 6.90150976e-01
9.33593512e-01 -8.39255989e-01 -6.54803336e-01 -6.55971706e-01
-9.43914428e-02 -2.64735520e-01 4.91418153e-01 -8.58813167e-01
-4.42154221e-02 -5.49879633e-02 -3.03957406e-02 -7.83271194e-01
2.39514604e-01 -7.41251528e-01 -5.37679419e-02 1.85705334e-01
-4.45986241e-01 -5.59692621e-01 7.14503750e-02 6.60247862e-01
-2.49050796e-01 -9.35383439e-02 9.68702674e-01 -1.52651504e-01
-9.97576594e-01 6.31776989e-01 3.56630862e-01 3.49698097e-01
8.67734313e-01 -7.57255107e-02 -2.21832126e-01 -1.08088285e-01
-7.59927154e-01 8.13157439e-01 5.85253119e-01 3.05490226e-01
3.82214636e-01 -1.24197102e+00 -6.44822896e-01 -1.04800221e-02
4.08458710e-01 6.82835400e-01 4.40106362e-01 9.63052273e-01
-2.37669989e-01 -9.86170489e-03 2.20361929e-02 -1.06555700e+00
-1.01511610e+00 3.07238400e-01 1.53181985e-01 -4.10361469e-01
-6.58518910e-01 1.36902344e+00 3.66398692e-01 -7.99379945e-01
2.88826764e-01 -9.97464582e-02 2.06916317e-01 -1.39334574e-01
2.64562219e-01 1.14139423e-01 7.29952306e-02 -5.80606818e-01
-4.91083682e-01 7.86746919e-01 -2.06980020e-01 -9.46714310e-04
1.44499600e+00 -3.13828677e-01 7.46971890e-02 3.97331893e-01
1.47130573e+00 -3.98141772e-01 -1.83600247e+00 -6.77613854e-01
8.61733928e-02 -4.52371508e-01 3.86504084e-01 -9.67056632e-01
-1.57121944e+00 8.01501334e-01 6.06303275e-01 -8.54093954e-02
1.10105729e+00 6.00572586e-01 8.25399220e-01 2.31165051e-01
3.38224232e-01 -9.43897963e-01 3.49252433e-01 3.36899102e-01
3.84076834e-01 -1.76788187e+00 -4.36463580e-02 -9.05244887e-01
-7.43609786e-01 5.46937466e-01 6.29626453e-01 2.93113850e-02
7.90585935e-01 2.07794875e-01 3.60976666e-01 -4.89197373e-01
-5.83959818e-01 -5.49911737e-01 4.91260141e-01 6.79756463e-01
2.51552880e-01 1.12416605e-02 2.13845521e-01 3.33016634e-01
2.23553196e-01 -1.77270800e-01 2.97808219e-02 7.87261963e-01
-3.60826612e-01 -1.28672349e+00 -2.89970869e-03 4.15839881e-01
-1.91593915e-01 -1.30858615e-01 -2.07768217e-01 6.09824955e-01
2.86924541e-01 9.50535715e-01 -1.45614371e-01 -5.59869528e-01
2.71854043e-01 -1.61411062e-01 3.64162624e-01 -7.83415437e-01
-3.42940181e-01 2.42411390e-01 3.33469391e-01 -9.22813177e-01
-9.73581553e-01 -6.25571430e-01 -1.30871737e+00 -8.61544833e-02
-4.10880119e-01 -1.52744114e-01 6.05714798e-01 9.16524529e-01
3.42110664e-01 5.64158678e-01 8.10210884e-01 -1.01815104e+00
-3.30066442e-01 -7.49213278e-01 -5.11149228e-01 5.91473460e-01
1.05485819e-01 -8.73047233e-01 -4.12822723e-01 1.42195180e-01] | [9.6158447265625, 0.7287827134132385] |
f9eb71d5-c695-484b-a8ca-ce413f659baf | detect-localize-repair-a-unified-framework | 2211.14875 | null | https://arxiv.org/abs/2211.14875v3 | https://arxiv.org/pdf/2211.14875v3.pdf | Detect-Localize-Repair: A Unified Framework for Learning to Debug with CodeT5 | Automated software debugging is a crucial task for improving the productivity of software developers. Many neural-based techniques have been proven effective for debugging-related tasks such as bug localization and program repair (or bug fixing). However, these techniques often focus only on either one of them or approach them in a stage-wise manner, ignoring the mutual benefits between them. In this work, we propose a novel unified \emph{Detect-Localize-Repair} framework based on a pretrained programming language model CodeT5 to seamlessly address these tasks, named CodeT5-DLR. Specifically, we propose three objectives to adapt the generic CodeT5 for debugging: a bug detection objective to determine whether a given code snippet is buggy or not, a bug localization objective to identify the buggy lines, and a program repair objective to translate the buggy code to its fixed version. We evaluate it on each of these tasks and their combined setting on two newly collected line-level debugging datasets in Java and Python. Extensive results show that our model significantly outperforms existing baselines from both NLP and software engineering domains. | ['Steven Hoi', 'Yue Wang', 'Nghi D. Q. Bui'] | 2022-11-27 | null | null | null | null | ['program-repair', 'program-repair'] | ['computer-code', 'reasoning'] | [-1.94198146e-01 -1.60181627e-01 -2.32735455e-01 -4.54189122e-01
-9.53210354e-01 -4.84067172e-01 -2.50571549e-01 4.18539286e-01
1.82556361e-01 3.83536071e-01 -1.90539286e-01 -5.70477605e-01
1.20519519e-01 -4.11761999e-01 -1.07705462e+00 -3.32579650e-02
-2.16720700e-01 -1.70349702e-01 2.50764787e-02 2.33574256e-01
5.03956974e-01 -2.77951241e-01 -1.29897392e+00 1.90627888e-01
1.38523257e+00 3.57103139e-01 3.59714925e-01 7.55834103e-01
-1.59482852e-01 1.21822393e+00 -8.64070535e-01 -4.72049534e-01
-3.38389933e-01 -3.54591906e-01 -8.11683595e-01 -2.41429806e-01
5.02110898e-01 -2.96345592e-01 3.03522229e-01 1.51200151e+00
1.95802912e-01 -2.52043962e-01 -3.82921123e-03 -1.37497330e+00
-1.08956647e+00 9.34150219e-01 -1.08444786e+00 1.57854632e-01
4.64883775e-01 1.01661056e-01 1.03240180e+00 -7.83147216e-01
2.83233911e-01 1.10686946e+00 1.05994189e+00 5.53190410e-01
-1.11922503e+00 -4.37449902e-01 9.97846499e-02 2.63810121e-02
-1.15512013e+00 -9.87525210e-02 6.89530313e-01 -8.06987166e-01
1.40542769e+00 -1.79649726e-01 9.21615865e-03 9.64666426e-01
4.54134792e-01 4.99779224e-01 5.66008568e-01 -5.53579926e-01
8.59050974e-02 -8.69831741e-02 6.48464024e-01 1.32498515e+00
2.47021973e-01 -2.24093720e-01 -4.79457527e-01 -4.07774866e-01
2.43262500e-01 -6.87967315e-02 -5.19711673e-01 2.20624000e-01
-9.41415012e-01 8.36144209e-01 3.53309214e-01 4.14033741e-01
3.12606804e-02 8.11162233e-01 5.74497879e-01 4.58406687e-01
4.53202158e-01 7.76266575e-01 -8.01776528e-01 -5.50453544e-01
-8.30272853e-01 3.71655434e-01 8.52129817e-01 9.97239769e-01
9.17672336e-01 1.59945965e-01 -3.08345944e-01 7.82864034e-01
5.09427488e-01 2.15952452e-02 5.98214090e-01 -7.26234496e-01
7.15491951e-01 1.10706234e+00 9.08375010e-02 -1.09579742e+00
-2.46284589e-01 -4.48702157e-01 -1.54410824e-01 1.84344366e-01
3.91590409e-02 -2.82372206e-01 -5.09871721e-01 1.92189097e+00
-5.03783859e-02 4.19183195e-01 -3.12719136e-01 5.64534664e-01
7.68151641e-01 5.24009049e-01 -2.07810193e-01 1.42541125e-01
1.16868222e+00 -1.46201777e+00 -4.27400976e-01 -7.04056203e-01
1.13376141e+00 -6.36866868e-01 1.39505494e+00 2.79359460e-01
-9.34080958e-01 -5.55011630e-01 -1.21411192e+00 -2.41348907e-01
2.23284811e-02 7.95261562e-01 5.94170749e-01 5.88797510e-01
-1.33166420e+00 8.62665951e-01 -1.02875173e+00 3.98350656e-02
1.40226856e-01 -4.79300916e-02 -2.36451030e-01 -1.21008880e-01
-6.09447658e-01 6.02246702e-01 2.52970457e-01 1.07997745e-01
-1.00686264e+00 -7.30587602e-01 -1.15585458e+00 4.22415257e-01
3.91351134e-01 -5.45323551e-01 1.68145204e+00 -8.59926701e-01
-1.06419182e+00 7.10422993e-01 -1.26547426e-01 -2.99448133e-01
1.25344768e-01 -4.99967337e-01 -7.08467960e-02 -5.29343009e-01
5.06799102e-01 2.82425106e-01 7.87018836e-01 -8.99141848e-01
-6.51001513e-01 -7.09613562e-02 2.67266065e-01 -5.49704134e-01
-3.30312014e-01 3.94106567e-01 -5.46718717e-01 -5.08913636e-01
-4.87331569e-01 -5.67496777e-01 -4.48035933e-02 -2.04137117e-01
-3.99694324e-01 -4.26784277e-01 5.12265861e-01 -1.16858244e+00
1.68949854e+00 -2.23995733e+00 3.75917077e-01 -3.55910063e-01
6.10133886e-01 7.97474459e-02 -4.20468092e-01 2.30774596e-01
-1.86733395e-01 3.93637061e-01 -3.93343836e-01 -5.27577341e-01
1.40489683e-01 -1.74458608e-01 -2.25595489e-01 2.51983285e-01
6.35073662e-01 8.52379024e-01 -1.17121303e+00 -2.28271261e-01
-5.51070690e-01 1.87989324e-01 -9.20621216e-01 4.29045677e-01
-6.54083788e-01 1.33545458e-01 -3.12626153e-01 8.64038110e-01
5.78359962e-01 -3.63816649e-01 -6.67652339e-02 4.44425136e-01
-1.12754650e-01 4.00999248e-01 -5.40931046e-01 1.83331239e+00
-8.53166580e-01 6.81203544e-01 -9.67755243e-02 -6.26946867e-01
9.80664432e-01 8.59515294e-02 -3.00707936e-01 -3.65850359e-01
-1.43990621e-01 4.28745002e-01 -1.62071865e-02 -9.66160893e-01
4.24412102e-01 4.10123914e-01 -4.40203369e-01 6.68739796e-01
3.61878008e-01 2.75371343e-01 2.34213635e-01 4.88672815e-02
1.82754087e+00 5.32347202e-01 3.39480847e-01 -6.60258085e-02
2.63763875e-01 -1.54171869e-01 9.00432825e-01 8.32160234e-01
-2.85785887e-02 5.95415175e-01 1.27389145e+00 -2.85390615e-01
-6.47029936e-01 -5.75442910e-01 3.93705130e-01 1.10025358e+00
-2.11689025e-01 -8.12451124e-01 -1.04148936e+00 -1.20083892e+00
4.29086275e-02 8.21294308e-01 -8.04271221e-01 -5.44004798e-01
-6.87366426e-01 -3.88911694e-01 7.85477996e-01 7.22123206e-01
2.62579799e-01 -9.74855840e-01 -7.57389843e-01 2.87116230e-01
-5.23916297e-02 -5.61302185e-01 -9.31760013e-01 3.85335892e-01
-4.86322820e-01 -1.19193995e+00 -2.93839663e-01 -1.00926626e+00
6.79107308e-01 6.82251677e-02 1.52188432e+00 8.04560423e-01
-3.23698997e-01 5.53322434e-02 -4.18452442e-01 -5.51398322e-02
-6.03313208e-01 -3.71304639e-02 -4.24188256e-01 -4.94790733e-01
2.35498399e-01 -5.39073169e-01 4.15335447e-02 2.27290969e-02
-5.97293615e-01 -3.47752810e-01 3.95971864e-01 8.66287410e-01
1.85560018e-01 1.12723075e-01 6.25600219e-01 -9.38494265e-01
9.67786252e-01 -7.72562265e-01 -9.48665202e-01 4.96291518e-01
-3.91914010e-01 1.40236378e-01 5.74290633e-01 -3.10955644e-01
-8.77817452e-01 -1.73032209e-01 -3.46238196e-01 -5.09791195e-01
1.32632807e-01 1.38925493e+00 -3.08165431e-01 -1.77379444e-01
7.19263673e-01 7.16274083e-02 -4.84370619e-01 -6.60886526e-01
-8.91744792e-02 4.64277357e-01 7.56520510e-01 -8.87945950e-01
5.70643187e-01 -4.53821659e-01 -5.88986278e-01 -1.27551347e-01
-7.00352907e-01 -1.53656542e-01 -1.65115327e-01 7.26303144e-04
6.39058292e-01 -8.49367142e-01 -4.90824342e-01 5.61477482e-01
-1.72979522e+00 -2.87733823e-01 1.93345055e-01 -6.75069988e-02
-1.66248947e-01 5.16936064e-01 -6.69618428e-01 -5.23338318e-01
-3.79307568e-01 -1.63787937e+00 1.27194595e+00 3.10975403e-01
-2.71066636e-01 -9.55564499e-01 4.07395750e-01 2.50467602e-02
4.40845639e-01 4.08050597e-01 1.49383545e+00 -3.94847602e-01
-5.45177400e-01 -4.15412158e-01 -2.97607005e-01 6.03987336e-01
2.85039663e-01 3.43859106e-01 -7.51694560e-01 -2.04900816e-01
2.62699902e-01 -3.79127771e-01 7.33558238e-01 2.03807354e-02
1.26504993e+00 -2.68715888e-01 -3.32144737e-01 6.25309467e-01
1.53147209e+00 2.32323837e-02 3.89985949e-01 2.48574927e-01
8.34219396e-01 3.53915185e-01 4.31526512e-01 3.06013674e-01
5.51010787e-01 5.11851788e-01 6.41933084e-01 4.09290969e-01
1.16447367e-01 -2.69689381e-01 7.80375004e-01 9.12872374e-01
4.49637622e-01 -2.49036193e-01 -1.32351017e+00 7.59086430e-01
-1.99902868e+00 -3.45270067e-01 -3.29511851e-01 1.99890757e+00
1.09739971e+00 8.23166128e-03 -1.85849562e-01 -2.34720945e-01
6.34006858e-01 -1.51201651e-01 -4.33185756e-01 -5.57024360e-01
5.50996065e-01 8.17306936e-02 -2.60743916e-01 3.36246252e-01
-9.92933154e-01 7.19705105e-01 5.23713923e+00 5.07952392e-01
-1.12076938e+00 4.20441210e-01 3.09719682e-01 2.68941075e-01
-3.06009620e-01 2.21499354e-01 -7.64148891e-01 6.60820425e-01
9.29548979e-01 -3.99981976e-01 5.34890652e-01 1.43977559e+00
-1.11279681e-01 -2.06533700e-01 -1.65400982e+00 6.40620828e-01
2.48757094e-01 -1.13333774e+00 -3.51125509e-01 -4.22647327e-01
6.10505104e-01 -1.42475098e-01 -1.05941303e-01 7.77609885e-01
1.72168627e-01 -1.07935393e+00 9.32280004e-01 4.82211590e-01
4.98746812e-01 -6.08043015e-01 9.03192103e-01 3.66739631e-01
-1.00654638e+00 -6.77810097e-03 -3.06362927e-01 -6.03960752e-02
-2.54780680e-01 6.98490441e-01 -7.02236533e-01 4.71180946e-01
9.36229289e-01 9.65128660e-01 -1.05871141e+00 1.29635727e+00
-6.72313809e-01 7.29787767e-01 3.53413552e-01 6.91123828e-02
9.55386087e-02 2.52968371e-01 4.25671130e-01 1.31841469e+00
5.41061342e-01 -9.21514988e-01 1.53636768e-01 1.73983622e+00
-3.25045347e-01 -3.69669080e-01 -4.31208223e-01 -3.58208716e-01
6.16886497e-01 1.18320131e+00 -4.84336257e-01 8.22973996e-02
-6.36990488e-01 8.14313471e-01 8.59018922e-01 1.27473801e-01
-1.23380029e+00 -9.57664847e-01 7.43288040e-01 -2.69982338e-01
2.88117319e-01 -1.02641784e-01 -2.59896696e-01 -1.21993124e+00
6.45460665e-01 -1.06601858e+00 5.60918674e-02 -8.78434956e-01
-1.06655788e+00 6.88271761e-01 -3.19305003e-01 -9.76885676e-01
-2.32201278e-01 -3.93305272e-01 -1.16174722e+00 9.23609614e-01
-1.50098896e+00 -8.86367321e-01 -4.19297457e-01 -1.54464483e-01
5.22612453e-01 -2.56506260e-02 6.18999124e-01 5.21231413e-01
-9.59801793e-01 8.75406027e-01 -3.17994624e-01 9.60719213e-02
6.61147654e-01 -1.59036648e+00 9.80348170e-01 1.53430068e+00
-2.36517593e-01 1.14163613e+00 5.53650200e-01 -9.31544423e-01
-1.45473468e+00 -1.58370936e+00 9.76559103e-01 -5.92330635e-01
8.92859459e-01 -4.05470610e-01 -1.43880796e+00 8.91085505e-01
1.54400662e-01 -1.08131833e-01 1.80933997e-01 1.49758145e-01
-6.42149508e-01 3.28443825e-01 -9.10782993e-01 2.10909873e-01
6.95237935e-01 -5.80975831e-01 -4.13128883e-01 3.62799853e-01
1.35608399e+00 -7.47452199e-01 -6.20066226e-01 6.97288513e-02
6.34621233e-02 -1.03208935e+00 4.63451058e-01 -3.77169728e-01
1.18534458e+00 -6.23477697e-01 7.48309642e-02 -1.47237754e+00
-2.64124483e-01 -6.04127049e-01 -1.65452942e-01 1.61727560e+00
5.10619402e-01 -5.17508507e-01 2.64721483e-01 2.64326572e-01
-6.77722216e-01 -7.03171313e-01 -7.06392825e-01 -7.63588846e-01
-1.47183448e-01 -3.28291595e-01 7.76256740e-01 1.16016090e+00
7.48283789e-02 3.30260813e-01 -3.65481079e-01 4.73717719e-01
2.00540826e-01 1.71987355e-01 6.16636753e-01 -1.00531912e+00
-9.65096176e-01 -4.99823302e-01 -2.27016956e-01 -7.39946127e-01
5.79125583e-01 -8.45409513e-01 5.63498378e-01 -1.28099990e+00
3.92282426e-01 -1.97899491e-01 5.95595241e-02 1.01728725e+00
-7.37555325e-01 -4.09265131e-01 -3.57102931e-01 -1.80512667e-01
-6.17566645e-01 1.83217034e-01 5.12137949e-01 -3.17366242e-01
-1.55363500e-01 -2.97859628e-02 -9.21979129e-01 7.30381489e-01
4.12196398e-01 -8.74155223e-01 -1.44463941e-01 -1.21087372e+00
8.81046951e-01 4.11013633e-01 4.84286934e-01 -9.80370581e-01
1.53461665e-01 7.04963207e-02 -4.14348572e-01 -1.20330200e-01
-5.04590333e-01 -2.99125016e-01 -2.07226172e-01 4.19971645e-01
-2.82482654e-01 6.35780454e-01 3.94832283e-01 6.45560384e-01
-3.91049147e-01 -8.80030692e-01 5.45772970e-01 5.44610806e-03
-6.94730759e-01 -2.82228179e-02 -1.66625440e-01 1.03036731e-01
8.47960591e-01 2.69907862e-01 -7.84553051e-01 1.64315253e-01
-1.39995798e-01 4.70638931e-01 6.24050438e-01 8.53398621e-01
5.12075186e-01 -1.03343189e+00 -7.13218689e-01 3.56015377e-03
5.42600870e-01 -4.65781279e-02 8.66940096e-02 8.22841287e-01
-6.10174775e-01 1.16753839e-01 -9.37474612e-03 -4.97692943e-01
-1.21471381e+00 5.08028865e-01 4.84360427e-01 -4.60307032e-01
-3.96053493e-01 1.35289204e+00 1.28529176e-01 -7.57807076e-01
4.23020422e-01 -9.63120162e-01 3.08918860e-02 -5.56096017e-01
6.09144390e-01 1.88293725e-01 3.36030811e-01 1.32524267e-01
-4.88930702e-01 2.63210833e-01 -1.98102087e-01 7.11575687e-01
1.41960776e+00 3.72144282e-01 -7.82131910e-01 4.36247021e-01
1.01691043e+00 1.14250459e-01 -1.06633615e+00 9.45433751e-02
5.20873606e-01 -3.84871066e-01 -3.18078585e-02 -1.07059550e+00
-1.10527968e+00 9.66067553e-01 2.21567139e-01 7.99869597e-02
1.01016045e+00 6.44367933e-02 5.66472888e-01 3.54619741e-01
6.09941483e-01 -3.32843065e-01 3.11807603e-01 5.96472383e-01
9.54774678e-01 -1.09959340e+00 -4.21907991e-01 -2.99734056e-01
-1.53229162e-01 1.24569297e+00 1.18667173e+00 -1.28026634e-01
-1.93943195e-02 6.36698902e-01 -3.79989415e-01 -2.59409040e-01
-9.91372466e-01 3.29267889e-01 2.65027061e-02 5.84627569e-01
8.41503561e-01 -9.06365067e-02 -7.75853321e-02 1.17520821e+00
1.17867544e-01 1.66445300e-01 8.95226061e-01 1.35426068e+00
-2.91826099e-01 -1.15795577e+00 -2.65778571e-01 3.83490592e-01
-4.22940493e-01 -3.82421523e-01 -4.88333046e-01 5.82480013e-01
2.05607355e-01 7.62507915e-01 -3.27253968e-01 -6.53283715e-01
2.80909747e-01 -7.76566658e-03 5.40668547e-01 -1.16589189e+00
-1.02078474e+00 -2.47979596e-01 4.84514125e-02 -5.96999168e-01
1.53917089e-01 -4.70463127e-01 -1.08454335e+00 -8.36802553e-03
-4.45972741e-01 9.66645107e-02 5.12250960e-01 9.24875557e-01
5.71246386e-01 1.06886554e+00 3.07268083e-01 -5.81309199e-01
-6.13571346e-01 -8.39649975e-01 -2.57699490e-01 -2.39380792e-01
5.57485282e-01 -6.73986137e-01 -2.11437732e-01 1.80978999e-01] | [7.599259853363037, 7.7202863693237305] |
502e8940-2f21-4827-989f-474e58d26d07 | endurance-aware-mapping-of-spiking-neural | 2103.05707 | null | https://arxiv.org/abs/2103.05707v1 | https://arxiv.org/pdf/2103.05707v1.pdf | Endurance-Aware Mapping of Spiking Neural Networks to Neuromorphic Hardware | Neuromorphic computing systems are embracing memristors to implement high density and low power synaptic storage as crossbar arrays in hardware. These systems are energy efficient in executing Spiking Neural Networks (SNNs). We observe that long bitlines and wordlines in a memristive crossbar are a major source of parasitic voltage drops, which create current asymmetry. Through circuit simulations, we show the significant endurance variation that results from this asymmetry. Therefore, if the critical memristors (ones with lower endurance) are overutilized, they may lead to a reduction of the crossbar's lifetime. We propose eSpine, a novel technique to improve lifetime by incorporating the endurance variation within each crossbar in mapping machine learning workloads, ensuring that synapses with higher activation are always implemented on memristors with higher endurance, and vice versa. eSpine works in two steps. First, it uses the Kernighan-Lin Graph Partitioning algorithm to partition a workload into clusters of neurons and synapses, where each cluster can fit in a crossbar. Second, it uses an instance of Particle Swarm Optimization (PSO) to map clusters to tiles, where the placement of synapses of a cluster to memristors of a crossbar is performed by analyzing their activation within the workload. We evaluate eSpine for a state-of-the-art neuromorphic hardware model with phase-change memory (PCM)-based memristors. Using 10 SNN workloads, we demonstrate a significant improvement in the effective lifetime. | ['Francky Catthoor', 'Nagarajan Kandasamy', 'Nikil Dutt', 'Jeffrey Krichmar', 'Anup Das', 'Shihao Song', 'Twisha Titirsha'] | 2021-03-09 | null | null | null | null | ['graph-partitioning'] | ['graphs'] | [-6.90783486e-02 -4.00793314e-01 -8.17577690e-02 2.41550818e-01
4.29651469e-01 -3.48900080e-01 3.69507894e-02 1.32736906e-01
-5.30747831e-01 8.39748442e-01 -4.96497661e-01 -2.96954691e-01
-1.52454212e-01 -1.05516231e+00 -1.08111703e+00 -1.07621467e+00
5.44857383e-02 4.14928705e-01 1.11121559e+00 -1.50733471e-01
5.19809663e-01 6.51872158e-01 -1.79647791e+00 2.61024237e-01
7.23000944e-01 9.55743372e-01 2.22992152e-01 1.93687215e-01
-1.85108304e-01 3.46190751e-01 -9.63096797e-01 -1.37787731e-02
3.36498991e-02 -1.43734530e-01 -1.21620417e-01 -5.43065548e-01
1.61468133e-01 1.65336162e-01 -7.40539730e-01 7.93298900e-01
3.04563105e-01 -1.26145869e-01 5.50728798e-01 -1.45794308e+00
-3.97312194e-01 1.08214951e+00 -3.80240679e-01 7.52964735e-01
-5.09145677e-01 1.04403168e-01 1.89806774e-01 -5.78234494e-01
4.46365654e-01 6.68763518e-01 4.08586055e-01 7.35404551e-01
-1.27315581e+00 -1.02086532e+00 -3.16036522e-01 1.86919436e-01
-1.38062453e+00 -3.96859378e-01 6.09766960e-01 -4.41083722e-02
1.82856417e+00 4.84108567e-01 1.40083122e+00 7.90503919e-01
1.45989084e+00 7.75854141e-02 1.07681775e+00 -7.80106783e-02
8.43074799e-01 -1.99757382e-01 5.92719615e-01 4.38479990e-01
8.61000717e-01 -1.11440055e-01 -1.02313566e+00 -2.18475491e-01
6.57488227e-01 1.27462804e-01 -1.77202776e-01 4.47023660e-02
-5.49776971e-01 1.17719866e-01 6.56912804e-01 3.14009100e-01
-2.28228390e-01 7.98948765e-01 4.48110878e-01 4.98495027e-02
-1.56697959e-01 4.88297820e-01 1.89354032e-01 -1.65656760e-01
-1.04921854e+00 -1.01287186e-01 7.26795912e-01 5.07707000e-01
4.39362526e-01 3.25794548e-01 8.73449966e-02 7.92732239e-01
3.50702107e-01 9.84012425e-01 6.56696081e-01 -4.99799669e-01
-1.28502801e-01 1.02378428e+00 -4.87766087e-01 -6.87361658e-01
-4.52071935e-01 -1.90716743e-01 -8.13728988e-01 3.12764794e-01
-2.82791704e-01 3.51318419e-01 -1.22189355e+00 1.56220472e+00
-2.83303469e-01 2.67232120e-01 -1.99464634e-01 7.87988901e-01
4.46906686e-01 1.04343128e+00 1.26498774e-01 -2.87568867e-01
1.19709766e+00 -5.96876681e-01 -7.04242349e-01 -2.47731969e-01
-3.29634547e-02 -2.25803450e-01 6.04102075e-01 8.47168788e-02
-1.31002009e+00 -3.44723165e-01 -1.74935627e+00 2.15586990e-01
-5.24514437e-01 -3.74769151e-01 2.86136389e-01 7.63181388e-01
-1.54681742e+00 8.18613648e-01 -1.44025123e+00 -3.43803346e-01
2.99396783e-01 8.89601350e-01 2.12400675e-01 5.04454911e-01
-8.97821903e-01 6.90068007e-01 1.09669790e-01 -2.74459273e-01
-8.30929101e-01 -7.72928476e-01 -1.34032652e-01 2.77160317e-01
-3.24229956e-01 -5.66910386e-01 4.63674575e-01 -7.17961073e-01
-1.34609425e+00 6.17758095e-01 2.34342534e-02 -6.54447854e-01
-3.04511368e-01 5.44604957e-01 -5.02821922e-01 -7.13464692e-02
-5.45168340e-01 7.63775170e-01 4.79016125e-01 -1.07713223e+00
-2.66585704e-02 -5.78150868e-01 -5.01413345e-01 -3.42336237e-01
-9.64914799e-01 -2.07818970e-01 -1.76791415e-01 -2.37340972e-01
3.94841373e-01 -1.11234200e+00 2.13325799e-01 -1.55864686e-01
-4.52671379e-01 -1.91027805e-01 1.03148484e+00 8.53148103e-02
1.44375646e+00 -2.03600240e+00 2.64025599e-01 5.76906979e-01
3.12429518e-01 3.54671240e-01 2.05894053e-01 4.00150299e-01
4.21266705e-01 6.98818415e-02 -2.33233213e-01 -2.17938749e-03
-4.29936022e-01 4.43446279e-01 -5.51260769e-01 4.27696735e-01
1.42715022e-01 7.42798328e-01 -3.21802020e-01 -9.88132134e-02
-2.35408351e-01 4.69313562e-01 -2.35560462e-01 -2.52481133e-01
4.19823341e-02 -2.68867642e-01 -4.40074578e-02 7.07566321e-01
9.80009973e-01 -3.00792724e-01 4.50621307e-01 -1.29266471e-01
-5.15297711e-01 2.32863426e-01 -6.85473025e-01 1.12609673e+00
2.54365802e-02 7.89362609e-01 -2.74268270e-01 -5.30555487e-01
1.34077752e+00 -3.62920821e-01 5.44832170e-01 -1.20039058e+00
1.53221652e-01 6.47968650e-01 3.95951688e-01 -2.63125598e-02
5.26234388e-01 3.21232468e-01 2.50884354e-01 5.93593836e-01
2.09527686e-02 1.90350324e-01 1.87016591e-01 -1.31557370e-02
1.76704419e+00 -5.12472510e-01 -7.40207553e-01 -1.18800020e+00
-4.03701067e-02 -4.86071520e-02 4.74314868e-01 5.13827801e-01
-2.20969349e-01 7.34956488e-02 5.04827321e-01 -3.71379673e-01
-1.41148293e+00 -1.58107936e+00 -4.72046643e-01 6.21774614e-01
4.90344912e-01 -2.87785947e-01 -8.92628670e-01 3.76160473e-01
1.12721451e-01 3.98458302e-01 -5.31195939e-01 -7.42525876e-01
-6.63650513e-01 -1.09574795e+00 7.70774186e-01 6.09130979e-01
3.01061243e-01 -1.08845413e+00 -1.60843885e+00 3.84324193e-01
7.02813804e-01 -5.79106092e-01 -2.53949344e-01 1.27725720e+00
-1.30808854e+00 -4.45693254e-01 1.44391671e-01 -6.58505678e-01
1.02756965e+00 1.08097181e-01 1.05158401e+00 3.60889524e-01
-5.69556773e-01 -2.42634758e-01 1.39605775e-01 -2.56937861e-01
-1.18008845e-01 1.01279117e-01 4.36665833e-01 -5.28571427e-01
3.22201222e-01 -9.92713690e-01 -9.62005138e-01 3.82770807e-01
-9.82301533e-01 -4.06128839e-02 3.13773900e-01 7.23165274e-01
1.08139563e+00 7.39596635e-02 5.26968420e-01 -5.55852592e-01
3.37008029e-01 -5.98803222e-01 -6.47639215e-01 1.86992176e-02
-9.68127310e-01 3.17558438e-01 7.53762066e-01 -6.49170697e-01
-2.94327617e-01 -2.51065463e-01 1.49336353e-01 -4.85856026e-01
7.58331597e-01 1.23261914e-01 6.86591817e-03 -4.49377209e-01
7.51171410e-01 3.65143120e-01 4.10593450e-02 9.07571837e-02
-5.97200096e-01 4.54257071e-01 2.06885248e-01 -4.35284644e-01
1.85784817e-01 5.92157900e-01 3.59828830e-01 -6.79170668e-01
5.64580679e-01 3.38493258e-01 7.25054592e-02 -5.53893328e-01
3.61648560e-01 -4.30040270e-01 -8.95952046e-01 6.54739022e-01
-8.93010318e-01 -6.53889060e-01 -4.77579497e-02 -1.50312468e-01
-1.54042408e-01 -4.33505177e-01 -1.02844679e+00 -8.86337161e-01
-7.53839731e-01 -1.16719770e+00 6.05622590e-01 9.73012865e-01
-1.93645153e-02 -5.78624606e-01 1.43474624e-01 -2.03091040e-01
8.00106287e-01 -2.91042298e-01 1.32432926e+00 -4.49081421e-01
-6.84725225e-01 4.16448534e-01 -3.72569859e-02 -2.06983253e-01
-2.64176577e-01 4.66455817e-01 -7.57334650e-01 -5.41265845e-01
-1.72800586e-01 1.20921105e-01 1.05526030e+00 4.88893926e-01
1.21813154e+00 -2.60089841e-02 -8.98976922e-01 5.33449292e-01
1.68764412e+00 7.61424005e-01 1.03320146e+00 2.25299865e-01
5.25714874e-01 9.78788435e-02 -2.84073621e-01 2.36188427e-01
8.84561613e-02 6.34083271e-01 6.35645151e-01 2.07132891e-01
-1.00537084e-01 -6.46776985e-03 5.61799705e-01 1.42053664e+00
3.13364446e-01 -5.80101728e-01 -1.08445179e+00 5.47502458e-01
-1.78634560e+00 -6.02128923e-01 -1.28468633e-01 2.35370898e+00
1.11058009e+00 2.91368484e-01 -8.82475823e-02 6.12289868e-02
7.60786414e-01 -2.99267173e-01 -1.13080478e+00 -7.53898442e-01
-4.61054832e-01 5.16578197e-01 1.05080211e+00 -1.40632734e-01
-1.00633815e-01 6.11557007e-01 5.82293034e+00 7.25321412e-01
-1.67762780e+00 2.15783536e-01 6.30470634e-01 -7.21307755e-01
-5.81791639e-01 -1.39669970e-01 -9.34875667e-01 1.28939414e+00
1.53300476e+00 6.20999001e-02 7.46500790e-01 2.56640017e-01
-8.35534781e-02 -4.96693075e-01 -8.41405392e-01 9.81968284e-01
-8.76068249e-02 -1.74135745e+00 3.69648226e-02 1.72322750e-01
7.65133739e-01 4.15037908e-02 3.05833578e-01 -1.66417286e-01
-3.96738172e-01 -9.47107196e-01 7.96674132e-01 7.23299265e-01
6.38332427e-01 -1.02320945e+00 4.47502375e-01 -9.72826406e-02
-7.87062883e-01 -3.66710752e-01 -9.53024685e-01 1.97826609e-01
-1.66378006e-01 8.91241431e-01 -2.23148227e-01 -3.90005857e-01
1.05508792e+00 -6.19903728e-02 -3.84874552e-01 1.25071394e+00
5.77589512e-01 4.65664178e-01 -4.84068215e-01 -7.33705699e-01
-3.25640887e-01 -2.71816373e-01 4.09311682e-01 8.90858948e-01
4.65621650e-01 -5.95953725e-02 -4.05959159e-01 1.32547069e+00
-5.52151144e-01 -2.59570360e-01 -4.38133687e-01 -8.97225440e-02
1.33457029e+00 1.11114299e+00 -1.40611076e+00 -4.35330868e-02
2.77112126e-01 7.12395966e-01 2.34882027e-01 -7.35715125e-03
-8.49505723e-01 -6.08243763e-01 9.15324628e-01 7.10140944e-01
-7.99621418e-02 -3.83985907e-01 -1.04655802e+00 -1.81679621e-01
-4.90661450e-02 -3.36989760e-01 -3.57836932e-01 -6.02368832e-01
-7.74401248e-01 7.08203375e-01 -5.63635409e-01 -4.55151379e-01
5.87820709e-01 -5.70687413e-01 -9.77948308e-01 5.82165241e-01
-1.04558885e+00 -4.00454640e-01 -3.22895318e-01 1.28157109e-01
1.96651965e-01 -4.14650254e-02 4.76590902e-01 3.68910909e-01
-9.97058570e-01 8.06617141e-01 2.60337412e-01 -5.31979859e-01
4.48204607e-01 -6.16181076e-01 3.90853792e-01 7.78701842e-01
-2.19272658e-01 9.01505232e-01 5.12242496e-01 -1.06745553e+00
-2.31239605e+00 -1.03861749e+00 4.49461371e-01 -4.93204482e-02
6.66434407e-01 -8.28522146e-01 -1.33165538e+00 6.71452880e-02
2.73525774e-01 -1.72273472e-01 6.65519893e-01 -5.60770512e-01
-7.38880336e-02 -5.37542284e-01 -1.19171774e+00 7.01388240e-01
1.19546342e+00 -3.96822125e-01 1.09943680e-01 5.82368299e-02
4.66814369e-01 -2.50597864e-01 -9.12070572e-01 4.38997686e-01
6.02863967e-01 -8.43742192e-01 4.59256798e-01 -3.05771828e-02
3.51527840e-01 -4.25462335e-01 3.68802324e-02 -1.11666691e+00
-7.97380134e-02 -2.62558907e-01 -4.06775564e-01 9.75488424e-01
5.12710750e-01 -8.77776623e-01 8.52222085e-01 3.00204039e-01
-5.56007683e-01 -1.22180903e+00 -1.36130285e+00 -9.54712629e-01
9.09939408e-02 3.57066840e-01 4.83527273e-01 2.50171632e-01
1.81384832e-01 -5.86816519e-02 3.08821917e-01 -2.89742108e-02
5.37115335e-01 -2.85747856e-01 -1.01253964e-01 -1.11245000e+00
-1.67588726e-01 -6.18501008e-01 -6.33995652e-01 -3.61076832e-01
2.81522900e-01 -7.53811061e-01 2.62141228e-01 -1.16163731e+00
5.42632520e-01 -9.40777302e-01 -6.41650498e-01 6.28768444e-01
2.46753260e-01 4.87607330e-01 -5.13818264e-02 4.46676850e-01
-3.56208235e-01 2.01823279e-01 6.72497988e-01 -1.71703622e-01
-2.60812521e-01 -5.97169101e-01 -2.47023758e-02 4.07341011e-02
7.63245165e-01 -6.58746600e-01 -5.79175115e-01 -6.60806596e-01
6.04027867e-01 -1.46872923e-01 2.08810687e-01 -1.72461379e+00
9.64713871e-01 1.24889225e-01 4.48995382e-01 -4.43316132e-01
4.31878507e-01 -5.77742398e-01 9.45235074e-01 1.04956675e+00
4.11791764e-02 6.37080789e-01 7.60061622e-01 3.33224863e-01
3.84169370e-01 -1.67553753e-01 8.94630969e-01 5.82616210e-01
-3.83422196e-01 2.96225220e-01 -8.56216729e-01 -3.14142793e-01
1.09546316e+00 -6.52408838e-01 -1.13844550e+00 5.61660945e-01
-1.66708007e-01 -3.65819931e-02 9.22771752e-01 2.28457853e-01
1.03066230e+00 -1.11973596e+00 -1.26905859e-01 4.98955697e-01
-1.39960632e-01 -6.88133061e-01 4.96717542e-01 8.15886974e-01
-7.03423798e-01 2.71117628e-01 -1.17861497e+00 -7.32073486e-01
-9.25378680e-01 3.36548150e-01 4.76718336e-01 1.92322731e-01
-2.62096375e-01 1.01987445e+00 -3.69184822e-01 4.89299804e-01
-5.02993241e-02 -1.13226607e-01 8.00571665e-02 -1.97782934e-01
2.23260105e-01 4.28238362e-01 5.10246992e-01 -1.07279114e-01
-7.33176351e-01 5.81771314e-01 -1.09180100e-02 -1.04532719e-01
1.31275570e+00 2.15892911e-01 -8.19050908e-01 7.23477066e-01
8.39734793e-01 -2.63859749e-01 -1.03096831e+00 3.66459310e-01
3.21487859e-02 -1.21021457e-02 1.15811452e-01 -7.00356841e-01
-1.39201188e+00 5.64954340e-01 9.59220231e-01 1.24105692e-01
1.31472933e+00 -1.83335721e-01 1.07221282e+00 1.97494984e-01
7.34797895e-01 -1.38122153e+00 3.24853301e-01 5.81335723e-01
2.54957736e-01 -1.88335583e-01 -2.97539979e-01 -2.07160711e-02
-3.68447416e-02 1.19111180e+00 1.17339981e+00 -4.12851423e-01
6.57885075e-01 1.23891401e+00 -3.93740475e-01 -5.28718650e-01
-1.14076138e+00 5.29868603e-01 -2.68844664e-01 2.92934507e-01
-2.46890401e-03 3.97082210e-01 -4.95080292e-01 4.11568105e-01
3.68395559e-02 -1.94719985e-01 7.16074646e-01 1.09846425e+00
-9.48540092e-01 -1.01442754e+00 -1.27829211e-02 1.00234640e+00
-1.94039177e-02 -1.67764187e-01 -2.21917436e-01 8.67344216e-02
7.91632161e-02 6.07829273e-01 9.16624784e-01 -8.72860253e-01
6.86782226e-02 8.16062167e-02 7.19269931e-01 -3.25959325e-01
-9.89842176e-01 -2.73149878e-01 -3.97187382e-01 -4.53442693e-01
4.98894811e-01 -1.86865851e-01 -2.07852435e+00 -7.54632711e-01
-1.42254904e-01 -8.67573824e-03 1.37135017e+00 3.22002500e-01
7.59966135e-01 1.11663520e+00 5.29780686e-01 -4.25286710e-01
-1.93505809e-01 -4.04445887e-01 -8.96032572e-01 -2.61923939e-01
-2.19006673e-01 -8.90149474e-01 -1.52908206e-01 -5.51194131e-01] | [8.219243049621582, 2.5034701824188232] |
d66e91a2-7323-4390-ab35-16033edc5734 | visualizing-ensemble-predictions-of-music | 2112.07627 | null | https://arxiv.org/abs/2112.07627v2 | https://arxiv.org/pdf/2112.07627v2.pdf | Visualizing Ensemble Predictions of Music Mood | Music mood classification has been a challenging problem in comparison with other music classification problems (e.g., genre, composer, or period). One solution for addressing this challenge is to use an ensemble of machine learning models. In this paper, we show that visualization techniques can effectively convey the popular prediction as well as uncertainty at different music sections along the temporal axis while enabling the analysis of individual ML models in conjunction with their application to different musical data. In addition to the traditional visual designs, such as stacked line graph, ThemeRiver, and pixel-based visualization, we introduce a new variant of ThemeRiver, called "dual-flux ThemeRiver", which allows viewers to observe and measure the most popular prediction more easily than stacked line graph and ThemeRiver. Together with pixel-based visualization, dual-flux ThemeRiver plots can also assist in model-development workflows, in addition to annotating music using ensemble model predictions. | ['Min Chen', 'Zelin Ye'] | 2021-12-14 | null | null | null | null | ['music-classification'] | ['music'] | [-7.85838664e-02 -3.52348745e-01 1.34891301e-01 -1.90342274e-02
-5.24052918e-01 -9.59908485e-01 4.06392097e-01 5.46137154e-01
1.97696567e-01 3.42418581e-01 2.61226296e-01 -2.55041301e-01
-4.34148401e-01 -6.09750152e-01 -1.40100881e-01 -4.44409490e-01
-2.22444683e-01 1.57319412e-01 -1.96003765e-01 -2.22453266e-01
6.40249074e-01 3.39234531e-01 -1.86161375e+00 7.00325668e-01
6.88776910e-01 1.00109994e+00 4.32109199e-02 6.41257107e-01
-3.86184543e-01 3.35975498e-01 -9.58938062e-01 -2.26905242e-01
-9.65203494e-02 -5.54122269e-01 -1.42576993e-01 -4.87479508e-01
5.84389150e-01 4.58768100e-01 4.32434946e-01 7.47958899e-01
5.53324282e-01 -3.07993125e-02 4.95597512e-01 -1.41462600e+00
-1.49114907e-01 8.80522609e-01 -9.65287924e-01 -2.05211323e-02
6.79862142e-01 4.78509329e-02 1.15282917e+00 -6.66534483e-01
5.77390015e-01 1.28375101e+00 8.95277917e-01 -1.01055820e-02
-1.81056142e+00 -1.06647491e+00 1.26954680e-02 3.76726031e-01
-1.17822766e+00 3.74112800e-02 1.14798641e+00 -9.34226453e-01
6.29145980e-01 1.06862867e+00 1.17237198e+00 5.81018329e-01
6.34602308e-02 7.08840370e-01 1.46397173e+00 -4.83954847e-01
7.44863227e-02 4.49541621e-02 6.81276172e-02 3.16874176e-01
-2.00706407e-01 -2.01472178e-01 -1.20960915e+00 -4.86146182e-01
5.90089679e-01 -8.86955038e-02 -3.29252571e-01 -3.11190903e-01
-1.50893199e+00 4.54966784e-01 3.01191330e-01 7.18757510e-02
-8.61988962e-02 1.94277659e-01 4.35349405e-01 1.60001040e-01
6.35480881e-01 9.29349005e-01 -3.54212672e-01 -4.89627898e-01
-1.43710542e+00 5.65394998e-01 6.55787170e-01 3.35693628e-01
6.72880352e-01 1.44830391e-01 -1.31748572e-01 8.76188517e-01
3.34230542e-01 2.18661383e-01 3.43873441e-01 -9.73625362e-01
-4.24010307e-03 8.18427444e-01 -5.26069924e-02 -9.91400957e-01
-6.26844347e-01 -5.89917481e-01 -4.99684036e-01 9.86177564e-01
5.15972495e-01 2.30389252e-01 -6.43240452e-01 1.42018437e+00
3.14601421e-01 1.62410438e-01 -4.17372167e-01 7.12365568e-01
9.47657287e-01 4.91740227e-01 7.19131827e-02 -1.43014237e-01
1.46888423e+00 -5.77648401e-01 -6.72681630e-01 2.13316619e-01
5.97350299e-01 -1.01823997e+00 1.57647550e+00 8.30926657e-01
-9.35537040e-01 -4.88373429e-01 -1.20309937e+00 -7.04461783e-02
-3.88798028e-01 8.80583674e-02 3.84989530e-01 4.89449531e-01
-7.07001030e-01 1.10250902e+00 -7.49957979e-01 -1.34079888e-01
3.67979228e-01 -5.05266450e-02 -2.68270195e-01 7.63241470e-01
-6.13609254e-01 7.77847469e-01 2.45846123e-01 -2.73994178e-01
-1.60037160e-01 -1.32996404e+00 -6.50976539e-01 1.09475181e-01
-2.04567716e-01 -3.53259444e-01 1.34029818e+00 -9.23559904e-01
-1.44518661e+00 7.42682040e-01 8.31167772e-02 3.42964120e-02
5.12207806e-01 -4.17340100e-01 -4.55297768e-01 -3.49863619e-01
-2.91649282e-01 4.33091044e-01 5.06832361e-01 -1.28036368e+00
-5.13591766e-01 -3.97214442e-01 -2.66275913e-01 1.45837232e-01
-2.07540050e-01 1.41791314e-01 -1.08554631e-01 -1.04867053e+00
4.23163533e-01 -6.85597897e-01 3.00833061e-02 1.97547987e-01
-5.85675418e-01 -4.00971472e-02 1.31495190e+00 -8.72865319e-01
1.93776572e+00 -2.25335598e+00 2.51500696e-01 3.68898839e-01
6.99302331e-02 -2.67971933e-01 3.41158897e-01 6.53910995e-01
-4.17773664e-01 1.93199590e-01 -2.20249623e-01 -3.27914357e-01
1.77359730e-01 -2.62843728e-01 -2.73480564e-01 4.50564511e-02
-1.94413096e-01 5.97075760e-01 -7.48505175e-01 -4.74934578e-01
3.13161731e-01 5.02625644e-01 -2.95508057e-01 -3.48120987e-01
-4.98848796e-01 7.78228581e-01 2.68732458e-01 5.60015380e-01
5.66070020e-01 -1.17900141e-01 5.38948119e-01 -3.68772328e-01
-5.35433352e-01 4.82160002e-01 -1.40324044e+00 1.86541164e+00
-3.43368560e-01 1.03792346e+00 -4.56729792e-02 -2.29729027e-01
1.26803780e+00 1.04060031e-01 3.36826265e-01 -4.77339745e-01
-4.34293389e-01 6.69798553e-02 -6.43371716e-02 -2.09361166e-01
6.92462862e-01 -1.63628682e-01 -8.76959786e-02 9.25396860e-01
-3.75847965e-01 1.21753573e-01 2.39179015e-01 4.56086807e-02
5.85543036e-01 8.41986775e-01 3.35330606e-01 -2.41133541e-01
-1.99656142e-03 2.50019170e-02 4.15313989e-01 3.60868573e-01
4.59699124e-01 7.01227903e-01 6.78859591e-01 -4.23286706e-01
-9.23418939e-01 -1.09140646e+00 -2.34017968e-01 1.22488058e+00
-2.33987719e-01 -1.35714436e+00 -3.41532201e-01 -2.48587444e-01
2.18590096e-01 8.59822750e-01 -5.69090962e-01 2.55376101e-01
-2.97239751e-01 -4.55068737e-01 4.72089022e-01 3.33199054e-01
-1.55783668e-01 -1.04062402e+00 -1.09295523e+00 1.17183244e-02
-8.42516199e-02 5.56865968e-02 -1.57863915e-01 3.39097947e-01
-9.45382595e-01 -8.84087741e-01 -5.11968136e-01 -1.78829014e-01
-7.46082887e-02 -1.49585620e-01 1.54949939e+00 -2.64724523e-01
-5.04787087e-01 2.20807552e-01 -2.87116021e-01 -8.18083644e-01
-2.01278448e-01 -1.49084002e-01 -2.26064011e-01 -2.50866205e-01
3.38933524e-03 -1.21704566e+00 -8.93251717e-01 1.31506965e-01
-6.39632940e-01 7.67408013e-01 -5.71239851e-02 4.51495051e-01
8.82340789e-01 -3.09197873e-01 4.65904474e-01 -7.22152412e-01
8.22278619e-01 -2.89871961e-01 -3.40702504e-01 4.17272478e-01
-7.74409831e-01 -1.46733969e-01 2.61984497e-01 -5.16520560e-01
-7.15493441e-01 -5.86687811e-02 1.90842777e-01 -5.95549107e-01
7.03725591e-02 8.28334212e-01 2.58931890e-02 2.32657462e-01
8.83161247e-01 -1.26251057e-01 8.08340684e-02 -9.80951369e-01
5.10107398e-01 5.67999601e-01 4.22853559e-01 -2.78036416e-01
3.66860896e-01 2.45638877e-01 -2.84405444e-02 -4.80215073e-01
-2.91881353e-01 -1.70851365e-01 -6.53386056e-01 -8.73963237e-01
6.40974998e-01 -5.90797722e-01 -9.89268005e-01 1.34104013e-01
-7.02145278e-01 -4.27983880e-01 -4.75546300e-01 3.57568741e-01
-6.82204664e-01 1.33285612e-01 -1.35535955e-01 -1.11183655e+00
-3.24229240e-01 -6.66099191e-01 9.93550003e-01 3.90294820e-01
-1.11974764e+00 -8.76234293e-01 4.42979395e-01 -2.71398900e-03
9.85885784e-02 7.67863333e-01 1.28923023e+00 -2.33292952e-01
-2.02030599e-01 -1.13943703e-01 2.46908262e-01 -1.91497251e-01
-1.05570639e-02 3.88974696e-01 -1.31580234e+00 1.08106017e-01
-5.90040565e-01 -1.07904151e-01 5.78661859e-01 3.36471319e-01
1.38831174e+00 9.42823291e-03 -2.33020484e-01 5.46020091e-01
1.07281530e+00 1.98965102e-01 5.56572556e-01 5.73017716e-01
4.97337759e-01 5.74663818e-01 6.10755980e-01 6.12868845e-01
5.37763648e-02 1.07222915e+00 3.85201991e-01 -1.62627280e-01
-2.63806954e-02 -4.84239161e-01 5.91078922e-02 4.68853354e-01
-1.21517234e-01 2.20251426e-01 -9.39267755e-01 -2.31571689e-01
-2.02497745e+00 -1.09691083e+00 -4.41706210e-01 2.56887889e+00
5.68713963e-01 -2.20572487e-01 4.94765908e-01 7.76278138e-01
4.57865834e-01 3.22035968e-01 -4.92292434e-01 -6.01818383e-01
-3.03095996e-01 2.04750583e-01 -1.29995674e-01 1.20802134e-01
-9.02988911e-01 4.00395095e-01 6.60750818e+00 7.74416864e-01
-1.49333155e+00 -1.80066396e-02 4.43922311e-01 -7.28610754e-01
-5.96399248e-01 2.08796144e-01 -9.57319438e-02 4.53482866e-01
5.55079758e-01 -2.11458489e-01 3.57643872e-01 6.14226282e-01
4.81627077e-01 -1.83066592e-01 -1.10894811e+00 1.42835712e+00
-1.49781018e-01 -1.65735364e+00 -9.11910608e-02 -1.17175886e-03
3.18357378e-01 -5.24368227e-01 1.46547422e-01 4.66414727e-02
-6.61967918e-02 -1.04556513e+00 1.08161080e+00 8.03065658e-01
1.00388622e+00 -7.83196211e-01 -1.19571447e-01 8.48762020e-02
-1.24102485e+00 1.02373296e-02 2.68774807e-01 -2.89033473e-01
3.62733215e-01 6.01039648e-01 -5.51363587e-01 7.05963790e-01
1.13007557e+00 9.44220901e-01 -7.73481488e-01 1.34221447e+00
6.31977469e-02 5.27328968e-01 -3.59873414e-01 1.69918075e-01
-6.87384427e-01 -3.89644295e-01 7.22188771e-01 1.45743084e+00
8.50586712e-01 -4.49093461e-01 -8.53732154e-02 9.78105307e-01
3.84372115e-01 4.23068017e-01 -1.80247813e-01 -1.29616037e-01
4.90974396e-01 1.53405130e+00 -8.15559804e-01 -1.13261141e-01
7.84472898e-02 7.41675138e-01 1.18084207e-01 2.77876288e-01
-4.97730821e-01 -2.30848804e-01 6.89151168e-01 4.27811265e-01
-1.18102379e-01 -2.52341211e-01 -9.88633692e-01 -6.35178685e-01
-1.53792083e-01 -1.10473335e+00 5.99799931e-01 -1.44271517e+00
-1.02997136e+00 4.12903696e-01 -1.62481144e-01 -1.68028820e+00
2.44011283e-02 -3.83846819e-01 -1.00177741e+00 1.13266075e+00
-6.57950044e-01 -1.25601411e+00 -3.01273942e-01 2.42182583e-01
1.44763693e-01 2.30049603e-02 1.29904151e+00 1.29679367e-01
-3.87116224e-01 3.01790982e-01 3.05871397e-01 -4.37331140e-01
8.71400893e-01 -1.68103778e+00 3.85945030e-02 5.41186705e-02
8.27644408e-01 3.74913037e-01 1.00434411e+00 -5.85355818e-01
-8.78348947e-01 -4.66952741e-01 3.60413194e-01 -4.16950881e-01
5.95782220e-01 -3.45629722e-01 -8.90704036e-01 2.98975050e-01
2.94314533e-01 -5.16202927e-01 1.32116461e+00 9.72466350e-01
-6.09727144e-01 -1.19574226e-01 -6.66321635e-01 7.43588865e-01
7.32390881e-01 -4.26507920e-01 -1.67956516e-01 4.38210703e-02
1.09345764e-01 -4.19292599e-01 -1.15796447e+00 4.45096612e-01
1.24506378e+00 -1.28627956e+00 8.08735847e-01 -5.09401977e-01
3.99438471e-01 -7.21850991e-01 1.01860622e-02 -1.38230479e+00
-3.36014032e-01 -5.67673266e-01 -2.80629426e-01 1.31655550e+00
6.44218802e-01 -5.27861230e-02 7.32453048e-01 2.17464581e-01
-1.14625312e-01 -7.45793283e-01 -7.87274957e-01 -2.84139335e-01
-2.41365850e-01 -8.99982810e-01 5.66663206e-01 9.64883566e-01
6.18400991e-01 3.31476107e-02 -4.63908970e-01 -2.23471731e-01
2.83710569e-01 6.17801011e-01 9.30952668e-01 -1.76125312e+00
-4.94128644e-01 -9.16516602e-01 -6.16734564e-01 -2.96810597e-01
-4.20823783e-01 -1.10259318e+00 -4.70785052e-01 -1.68292832e+00
1.86866835e-01 -6.98887289e-01 -5.66708744e-01 5.30975997e-01
2.93156295e-03 6.34402454e-01 5.93287170e-01 4.44502652e-01
-3.04952115e-01 5.48460260e-02 1.07718956e+00 9.90390964e-03
-5.56610346e-01 -1.10999539e-01 -5.53787053e-01 7.18296170e-01
8.01254094e-01 -4.04901356e-01 -3.07480663e-01 -2.43265331e-02
7.18623519e-01 -9.86529216e-02 5.46528876e-01 -1.15653849e+00
-1.15243569e-01 1.92574076e-02 7.38506317e-01 -1.04142499e+00
5.88382721e-01 -2.75788754e-01 1.01250362e+00 2.27539882e-01
-4.33651924e-01 5.17749310e-01 5.84475100e-01 4.58681464e-01
-1.73053935e-01 1.57337189e-01 5.15927792e-01 6.74376264e-02
-3.42575073e-01 -4.77330476e-01 -1.58843398e-01 -4.82386112e-01
7.53852427e-01 -4.22842056e-01 -3.73365343e-01 -5.92615187e-01
-1.19338298e+00 4.69294637e-02 5.84285617e-01 5.70156574e-01
3.93150598e-01 -1.35076845e+00 -5.20720780e-01 8.71347114e-02
2.93483138e-01 -4.12882656e-01 5.30737758e-01 9.05494750e-01
-4.62759465e-01 -2.28405103e-01 -3.96050602e-01 -9.68033612e-01
-1.81243610e+00 3.19297224e-01 2.36247703e-01 -9.37518403e-02
-8.46245408e-01 6.86670542e-01 1.45228490e-01 -3.51014316e-01
2.35161781e-01 1.60627197e-02 -4.86036330e-01 5.48144042e-01
6.63278580e-01 5.49137056e-01 -2.38863635e-03 -3.85088772e-01
-2.47555852e-01 5.53786159e-01 4.24646944e-01 -4.70415443e-01
1.17850268e+00 3.93849704e-03 -2.12241024e-01 1.41388404e+00
3.69734764e-01 4.38388884e-01 -1.19073379e+00 2.65328109e-01
2.78658867e-01 -5.07366836e-01 -2.85033017e-01 -1.43954480e+00
-6.93597734e-01 1.02236557e+00 1.07115614e+00 4.99851704e-01
1.10469759e+00 -4.53564823e-02 6.96959393e-03 -2.37964317e-01
9.54258069e-02 -1.03946960e+00 -3.97997722e-02 -7.65744247e-04
1.17936933e+00 -8.17594230e-01 2.28364095e-01 -1.41058579e-01
-5.69592178e-01 1.36072767e+00 4.06915128e-01 3.69370759e-01
6.47853315e-01 4.90726650e-01 5.75388908e-01 -2.59742379e-01
-7.98942506e-01 8.97908360e-02 6.14715755e-01 3.28498960e-01
1.18570280e+00 2.95411348e-01 -3.25698286e-01 6.84585989e-01
-8.48573029e-01 -1.38782099e-01 2.38865577e-02 8.94161463e-01
-3.74983728e-01 -1.35929501e+00 -7.16171980e-01 7.41120040e-01
-3.46838534e-01 -6.92039430e-02 -5.20675719e-01 5.16396344e-01
4.05522108e-01 8.06238413e-01 3.84132028e-01 -7.19119132e-01
2.84094900e-01 4.92224514e-01 4.46401924e-01 -4.29657459e-01
-9.83432233e-01 4.38390374e-01 1.99762896e-01 -3.84433061e-01
-1.63103968e-01 -5.70708752e-01 -1.05887485e+00 -2.83477336e-01
-9.43112299e-02 1.64202601e-01 1.04173851e+00 3.99227440e-01
4.19676244e-01 7.87051260e-01 1.42512515e-01 -1.09873831e+00
1.59724027e-01 -1.28611028e+00 -8.03767443e-01 3.50161821e-01
-8.40741843e-02 -6.05308712e-01 -7.35849738e-02 7.54838735e-02] | [15.963419914245605, 5.3220391273498535] |
cc18bcf8-5b33-481a-bed1-ec14c3278b7f | hybrid-energy-based-model-in-the-feature | 2305.16966 | null | https://arxiv.org/abs/2305.16966v3 | https://arxiv.org/pdf/2305.16966v3.pdf | Hybrid Energy Based Model in the Feature Space for Out-of-Distribution Detection | Out-of-distribution (OOD) detection is a critical requirement for the deployment of deep neural networks. This paper introduces the HEAT model, a new post-hoc OOD detection method estimating the density of in-distribution (ID) samples using hybrid energy-based models (EBM) in the feature space of a pre-trained backbone. HEAT complements prior density estimators of the ID density, e.g. parametric models like the Gaussian Mixture Model (GMM), to provide an accurate yet robust density estimation. A second contribution is to leverage the EBM framework to provide a unified density estimation and to compose several energy terms. Extensive experiments demonstrate the significance of the two contributions. HEAT sets new state-of-the-art OOD detection results on the CIFAR-10 / CIFAR-100 benchmark as well as on the large-scale Imagenet benchmark. The code is available at: https://github.com/MarcLafon/heatood. | ['Nicolas Thome', 'Clément Rambour', 'Elias Ramzi', 'Marc Lafon'] | 2023-05-26 | null | null | null | null | ['density-estimation'] | ['methodology'] | [-4.81538624e-01 -4.11519855e-02 -2.14710325e-01 -4.09934402e-01
-9.02282655e-01 -3.51413012e-01 6.88259900e-01 -2.63669282e-01
-2.90732384e-01 4.15108681e-01 -6.07524998e-02 -1.61361143e-01
3.34229916e-01 -5.81070185e-01 -8.95156384e-01 -8.06702852e-01
-2.31148660e-01 4.80396479e-01 1.53743058e-01 4.02574539e-01
-9.55772772e-02 5.28596282e-01 -1.59568131e+00 -2.17578143e-01
7.10817158e-01 1.31169558e+00 3.23838204e-01 6.41302288e-01
-1.55921131e-01 7.28819311e-01 -8.36779714e-01 -2.99843252e-01
1.00165874e-01 -2.24791944e-01 -2.75572568e-01 -2.03766063e-01
6.89666748e-01 -7.64827728e-01 -4.95056689e-01 1.32782519e+00
6.00557268e-01 1.88187271e-01 1.31996918e+00 -1.58201313e+00
-6.37122631e-01 1.81192696e-01 -7.39822626e-01 5.82383811e-01
-4.28103596e-01 -2.90351719e-01 6.71991765e-01 -1.34105766e+00
4.44289982e-01 1.39022613e+00 5.82521021e-01 5.12914300e-01
-1.05365753e+00 -7.90025234e-01 5.07397167e-02 2.27377877e-01
-1.58762944e+00 -4.18959290e-01 6.88975751e-01 -3.44273537e-01
1.20683110e+00 -3.99116814e-01 7.29240894e-01 1.22083807e+00
4.01058137e-01 1.28598487e+00 7.86252201e-01 -4.71186221e-01
6.72187746e-01 1.55653223e-01 2.30605289e-01 6.75663710e-01
5.41243315e-01 -2.89564696e-03 -5.50904989e-01 -2.18587443e-01
7.69408286e-01 -3.66560549e-01 -1.01590335e-01 -4.83616829e-01
-4.42164779e-01 1.01906168e+00 4.40807462e-01 1.31288230e-01
-2.78269023e-01 6.54530942e-01 1.85294688e-01 -2.19958410e-01
9.64584053e-01 -2.35912889e-01 -1.72171310e-01 -3.08154464e-01
-1.12072361e+00 2.25304395e-01 9.26624358e-01 1.13349295e+00
8.70395422e-01 2.14864045e-01 -2.16673948e-02 1.06918502e+00
8.73337805e-01 5.49659669e-01 2.01852992e-01 -1.06847692e+00
5.90968356e-02 1.17343239e-01 -8.75137225e-02 -4.25249934e-01
-1.14980685e-02 -6.01890147e-01 -7.89935768e-01 4.34810519e-01
2.51879126e-01 -2.36924797e-01 -1.32355750e+00 1.74938297e+00
5.12270868e-01 4.52309072e-01 -2.04061329e-01 7.77025998e-01
6.84741855e-01 7.02479899e-01 1.46703973e-01 3.49032283e-01
1.45989645e+00 -9.67974484e-01 -6.69504642e-01 -3.70510608e-01
5.21939874e-01 -5.38195968e-01 7.37252593e-01 4.98047978e-01
-8.20365846e-01 -2.87311524e-01 -1.25354099e+00 -1.92433164e-01
-4.39044774e-01 2.25619361e-01 5.09865046e-01 8.55598688e-01
-1.12372386e+00 3.20157856e-01 -1.26348436e+00 -1.70850456e-01
1.06589723e+00 3.49231027e-02 4.77256067e-02 -1.86202571e-01
-8.56390655e-01 7.20257759e-01 1.24157690e-01 -2.33130783e-01
-1.47161961e+00 -1.00407541e+00 -8.85716319e-01 4.07776758e-02
-1.41032174e-01 -5.22131562e-01 1.31764340e+00 -6.94539249e-01
-1.21927047e+00 8.22645783e-01 -1.36790872e-01 -5.95715225e-01
5.33250391e-01 -5.91364264e-01 -8.34539458e-02 3.72497857e-01
-6.99138343e-02 1.16370833e+00 1.06641996e+00 -1.26283050e+00
-4.61175352e-01 -1.58285871e-01 -4.27766263e-01 -5.20603508e-02
-3.05853188e-01 2.79568438e-03 -6.41734123e-01 -7.27562666e-01
-3.20142746e-01 -7.99780905e-01 4.05168444e-01 1.19528130e-01
-5.86113691e-01 -5.43462217e-01 8.76331091e-01 -6.18425786e-01
9.19582367e-01 -2.25279832e+00 -2.59657681e-01 5.81341721e-02
5.01624227e-01 -1.77888498e-02 2.21980270e-03 1.02306657e-01
1.85740404e-02 6.05987906e-02 -1.42543375e-01 -9.22330558e-01
5.49459040e-01 6.14139400e-02 -2.44009048e-01 6.98040783e-01
4.65699375e-01 6.23322129e-01 -5.33042133e-01 -4.90399033e-01
3.44954371e-01 1.22911823e+00 -4.42131579e-01 3.16724926e-01
-1.01332232e-01 -1.79766104e-01 -1.42617270e-01 8.42223048e-01
1.01482463e+00 -6.14041872e-02 -2.84201592e-01 -3.90783727e-01
1.28399199e-02 3.01534116e-01 -9.39479768e-01 1.51948583e+00
-1.72085211e-01 9.11671519e-01 1.23078860e-01 -8.04134846e-01
7.37892389e-01 1.44974872e-01 4.37288493e-01 -4.21993464e-01
5.40681601e-01 2.03243926e-01 -1.45103097e-01 5.57811223e-02
3.97879839e-01 1.61852121e-01 2.01936543e-01 1.88069999e-01
8.69921386e-01 2.02663261e-02 3.30709577e-01 4.01927590e-01
1.00454366e+00 2.42698818e-01 1.59644544e-01 -6.61511302e-01
-3.15795302e-01 -2.54479110e-01 3.23446542e-01 7.90521443e-01
-4.57854629e-01 8.06351364e-01 5.19268155e-01 2.05063698e-05
-1.03970182e+00 -1.40408492e+00 -6.74968362e-01 7.76318550e-01
-1.99896917e-02 -3.05195600e-01 -1.19807756e+00 -6.79922223e-01
8.73778686e-02 8.76602769e-01 -5.63179135e-01 -3.15265149e-01
-8.31154808e-02 -1.21603179e+00 6.52790129e-01 6.06224835e-01
4.99055773e-01 -5.80754161e-01 -5.80699503e-01 1.51807070e-01
-1.04129955e-01 -1.15100229e+00 -3.71270925e-01 7.17912734e-01
-6.38444722e-01 -7.05450535e-01 -9.73036289e-01 -6.59974337e-01
5.37100136e-01 -6.49027303e-02 1.30221272e+00 -4.07793403e-01
-6.33257747e-01 5.94011307e-01 -2.30015278e-01 -8.35148573e-01
-3.60072881e-01 -9.47062001e-02 7.76110813e-02 -3.58752489e-01
8.53883684e-01 -6.53930664e-01 -7.98640072e-01 1.77544624e-01
-6.43641829e-01 -2.37968907e-01 5.69380045e-01 5.39143980e-01
7.11127043e-01 1.25997007e-01 5.96135974e-01 -3.64413440e-01
3.48186374e-01 -7.62671649e-01 -6.50339961e-01 -9.95716825e-02
-5.55554986e-01 1.69860548e-03 5.75557351e-02 -5.40884495e-01
-1.16916656e+00 -1.72728926e-01 -4.45189059e-01 -8.42811406e-01
-4.33507890e-01 1.05490781e-01 -2.57326007e-01 -6.87773526e-02
5.02329528e-01 -8.15057021e-04 -1.83518976e-01 -8.33153665e-01
2.57946312e-01 6.18251026e-01 5.97948074e-01 -6.48285091e-01
5.48522353e-01 4.49629873e-01 -2.51263529e-01 -9.04688358e-01
-7.49624074e-01 -6.44200742e-01 -4.23060030e-01 -2.65407115e-01
9.02269959e-01 -1.25494266e+00 -1.87617481e-01 9.41013873e-01
-9.38964546e-01 -7.07168102e-01 -1.59022987e-01 6.25762343e-01
-3.86212558e-01 1.88608348e-01 -7.24483728e-01 -1.07270777e+00
-5.19752741e-01 -8.71833622e-01 1.32646430e+00 4.71635550e-01
8.78789425e-02 -1.11185586e+00 2.54201472e-01 -8.66776630e-02
2.77473032e-01 -2.03030761e-02 7.73954630e-01 -6.90238476e-01
-4.37862784e-01 -1.61620811e-01 -4.19799119e-01 6.94771707e-01
-1.48987398e-01 2.18830019e-01 -1.38236260e+00 -2.36320585e-01
-1.75137386e-01 -4.60226148e-01 1.13018095e+00 9.88481343e-01
1.11479759e+00 1.36847183e-01 -3.94278526e-01 7.24337280e-01
1.52513111e+00 -5.03106369e-03 8.19712758e-01 1.53879151e-01
5.73333681e-01 1.12729199e-01 4.33849573e-01 5.71481824e-01
5.25676727e-01 4.67095822e-01 5.56202531e-01 -9.02206898e-02
-6.38880968e-01 -2.18915850e-01 4.58970636e-01 4.83912140e-01
3.26878667e-01 -6.36389315e-01 -8.86930406e-01 6.45789742e-01
-1.38869679e+00 -6.92012131e-01 1.26008438e-02 1.90572393e+00
6.81723773e-01 1.08120821e-01 3.08138400e-01 -6.02746420e-02
7.21324921e-01 -1.93370320e-02 -7.18113482e-01 -8.13259855e-02
1.77817374e-01 3.45595717e-01 6.42792821e-01 2.65154421e-01
-1.44556534e+00 6.39850080e-01 6.08316946e+00 1.29954410e+00
-7.23888934e-01 4.31117505e-01 9.25553501e-01 -4.86202240e-02
-7.73214102e-02 -4.81427312e-01 -1.36524487e+00 6.19694293e-01
1.13960409e+00 8.34735259e-02 2.96782434e-01 1.18865788e+00
-1.75030947e-01 -4.81073499e-01 -9.28687990e-01 1.13005638e+00
1.43201604e-01 -1.03232014e+00 -3.55353594e-01 4.62232471e-01
5.81418574e-01 7.88702965e-01 1.26992539e-01 5.35913408e-01
4.74989504e-01 -7.68290520e-01 1.07412922e+00 2.72686720e-01
8.03510249e-01 -7.86910713e-01 6.02846742e-01 2.03006789e-01
-1.29803407e+00 3.55074793e-01 -7.85210431e-01 4.43922162e-01
7.19391853e-02 1.17693555e+00 -6.57372594e-01 1.39839603e-02
1.14096773e+00 6.72456145e-01 -3.23417097e-01 1.24030650e+00
-1.43684864e-01 9.34534609e-01 -8.64941061e-01 1.49834305e-01
4.61759418e-02 -1.20916933e-01 5.79781294e-01 1.30760717e+00
5.88807225e-01 -4.99538213e-01 -2.62725234e-01 1.33757198e+00
-5.60788631e-01 -3.16193491e-01 -2.38559559e-01 -2.14640662e-01
7.63644516e-01 1.49047363e+00 -8.17143440e-01 -3.01737249e-01
-3.99028569e-01 8.30344141e-01 4.67320502e-01 4.05345678e-01
-1.17005408e+00 -4.25705552e-01 8.54851186e-01 -1.80335045e-01
4.92901772e-01 -1.35043964e-01 1.50887176e-01 -1.25411999e+00
-2.05517530e-01 -3.67058516e-01 2.29363784e-01 -6.88284814e-01
-1.62908173e+00 4.22291249e-01 4.15755689e-01 -9.29387629e-01
-2.69050356e-02 -8.25714588e-01 -7.67825663e-01 9.11229908e-01
-1.96645856e+00 -9.01486397e-01 -4.55697745e-01 1.83240443e-01
4.65326220e-01 1.68830454e-02 6.59522653e-01 7.44548500e-01
-7.65842319e-01 6.41705096e-01 2.26131558e-01 2.50497878e-01
6.49273932e-01 -1.56262159e+00 7.03959167e-01 8.31705868e-01
1.43685564e-01 3.25433254e-01 6.79851174e-01 -5.07182002e-01
-1.17746449e+00 -1.19144332e+00 2.33294830e-01 -3.60696435e-01
4.72089171e-01 -7.20556080e-01 -8.39847028e-01 5.60898781e-01
2.97757357e-01 4.14399624e-01 5.72445154e-01 -2.70395458e-01
-3.48118782e-01 1.99020863e-01 -1.17910492e+00 4.10933271e-02
8.54463398e-01 -3.84924203e-01 -4.47992802e-01 1.61293224e-01
4.13158447e-01 -3.86761725e-01 -9.70125496e-01 2.13287279e-01
5.94709694e-01 -1.10569525e+00 1.13305104e+00 1.13976493e-01
1.15269631e-01 -1.37912601e-01 -3.65679473e-01 -1.29354477e+00
-2.04700321e-01 -2.27643013e-01 -8.91792953e-01 1.60773611e+00
2.31635496e-02 -5.72780907e-01 8.18131626e-01 3.63719136e-01
-3.51350993e-01 -8.29241991e-01 -1.06509352e+00 -1.07151246e+00
3.02871674e-01 -6.98692739e-01 2.90266693e-01 4.64857340e-01
-4.96458739e-01 -7.25961253e-02 -1.15347683e-01 2.53754050e-01
1.17059088e+00 -6.10176861e-01 5.43941617e-01 -1.18995190e+00
-2.73796916e-01 -3.85353833e-01 -5.12452602e-01 -1.16327834e+00
2.26165131e-01 -7.54804492e-01 2.84392506e-01 -1.56876874e+00
2.48439014e-01 -3.06223541e-01 -3.82121861e-01 2.60883093e-01
1.51333287e-01 4.90325332e-01 -1.73289441e-02 1.13953076e-01
-6.16020262e-01 9.68335986e-01 4.72099960e-01 2.39182711e-02
3.24702598e-02 -2.72843748e-01 -3.83466035e-01 8.28197241e-01
9.49265540e-01 -8.86780977e-01 -4.75827187e-01 -2.79413104e-01
-3.51771303e-02 -7.29394853e-01 6.84075177e-01 -1.24964440e+00
-1.09619163e-01 3.71237904e-01 6.01379991e-01 -7.77366877e-01
4.84589547e-01 -6.07511938e-01 -1.76057070e-01 1.69229984e-01
7.59735331e-02 -3.20349783e-01 4.76790637e-01 6.71826422e-01
-1.00741103e-01 -3.44650179e-01 1.05848682e+00 2.38023788e-01
-8.51891816e-01 4.64548588e-01 -3.82071584e-01 5.56020260e-01
9.18578923e-01 -6.64478689e-02 -7.96200097e-01 -3.60825956e-01
-2.96877950e-01 -1.26374839e-02 2.99188435e-01 2.70595729e-01
4.96517420e-01 -1.38913620e+00 -4.33864921e-01 1.48988232e-01
1.04240030e-01 6.79139495e-02 2.56401211e-01 1.06596625e+00
-5.39633453e-01 5.17451875e-02 3.62764634e-02 -7.54382610e-01
-7.73526609e-01 3.03258225e-02 7.35101521e-01 1.44802481e-02
-6.10943854e-01 1.18850577e+00 5.97390056e-01 -1.62835494e-01
5.01914680e-01 -3.69177073e-01 1.11023515e-01 3.90176922e-02
7.03409195e-01 3.05826515e-01 1.40994132e-01 -3.96625876e-01
-5.52288234e-01 1.34645343e-01 -3.05481125e-02 -1.78025842e-01
1.36875200e+00 -2.62687266e-01 2.30121732e-01 4.21632439e-01
1.48904312e+00 -2.98675448e-01 -1.71746099e+00 -5.62581979e-03
-3.04906011e-01 -3.46251875e-01 8.60169232e-01 -7.98130929e-01
-1.08846056e+00 1.13684893e+00 9.96295452e-01 7.21004382e-02
9.37837064e-01 3.10199261e-01 7.41387963e-01 1.04318172e-01
1.47086829e-01 -1.12959659e+00 1.30981863e-01 3.07097882e-01
4.89317566e-01 -1.03726232e+00 -1.69145808e-01 -2.37958342e-01
-3.71150672e-01 9.69787896e-01 6.45976961e-01 -3.93075168e-01
1.17422867e+00 8.69099081e-01 -1.75436601e-01 -1.58308208e-01
-7.08292544e-01 -1.95811257e-01 5.17963588e-01 8.29354525e-01
4.87320244e-01 -2.37014130e-01 4.30083096e-01 5.81063628e-01
5.69016906e-03 -8.20892081e-02 3.74630898e-01 8.68315160e-01
-5.27228415e-01 -7.53762901e-01 -1.59683913e-01 7.08436131e-01
-5.78572989e-01 -3.22964728e-01 7.11532915e-03 7.47793376e-01
-4.89916876e-02 7.22115457e-01 4.83198166e-01 -1.60143912e-01
-1.51427284e-01 2.89900720e-01 6.08356476e-01 -3.77122402e-01
1.34274170e-01 5.39866745e-01 1.39403995e-03 -4.70026970e-01
-3.62045795e-01 -6.47859633e-01 -1.34735620e+00 -2.58575112e-01
-4.13901299e-01 -1.39760822e-01 1.10234487e+00 7.12356746e-01
5.38990438e-01 6.35177553e-01 1.84397355e-01 -1.22186542e+00
-4.65074211e-01 -1.07574964e+00 -9.98066008e-01 7.92596042e-02
3.49385381e-01 -1.05959487e+00 -7.46005058e-01 6.06073327e-02] | [7.5902910232543945, 3.6289761066436768] |
76c0ea7f-b5fb-4652-b6dc-7b9c32a15039 | ldp-net-an-unsupervised-pansharpening-network | 2111.12483 | null | https://arxiv.org/abs/2111.12483v1 | https://arxiv.org/pdf/2111.12483v1.pdf | LDP-Net: An Unsupervised Pansharpening Network Based on Learnable Degradation Processes | Pansharpening in remote sensing image aims at acquiring a high-resolution multispectral (HRMS) image directly by fusing a low-resolution multispectral (LRMS) image with a panchromatic (PAN) image. The main concern is how to effectively combine the rich spectral information of LRMS image with the abundant spatial information of PAN image. Recently, many methods based on deep learning have been proposed for the pansharpening task. However, these methods usually has two main drawbacks: 1) requiring HRMS for supervised learning; and 2) simply ignoring the latent relation between the MS and PAN image and fusing them directly. To solve these problems, we propose a novel unsupervised network based on learnable degradation processes, dubbed as LDP-Net. A reblurring block and a graying block are designed to learn the corresponding degradation processes, respectively. In addition, a novel hybrid loss function is proposed to constrain both spatial and spectral consistency between the pansharpened image and the PAN and LRMS images at different resolutions. Experiments on Worldview2 and Worldview3 images demonstrate that our proposed LDP-Net can fuse PAN and LRMS images effectively without the help of HRMS samples, achieving promising performance in terms of both qualitative visual effects and quantitative metrics. | ['Yi Zhang', 'Leyuan Fang', 'Jiliu Zhou', 'Mingzheng Hou', 'Zhongzhou Zhang', 'Zhimin Shao', 'Jiahui Ni'] | 2021-11-24 | null | null | null | null | ['pansharpening'] | ['computer-vision'] | [ 7.41898417e-01 -5.74754000e-01 6.12198301e-02 -2.47132644e-01
-7.89036095e-01 -3.53276432e-01 3.53765696e-01 -2.38180041e-01
-4.18268383e-01 6.02689505e-01 -1.17510751e-01 -1.14290059e-01
-3.76365036e-01 -1.12708318e+00 -4.62970555e-01 -1.18268359e+00
4.70978826e-01 -2.87820309e-01 2.50177085e-01 -2.45510235e-01
-2.69564748e-01 5.61316311e-01 -1.47104907e+00 -2.33416613e-02
1.53578854e+00 1.07651246e+00 8.09811950e-01 4.16478604e-01
2.53354788e-01 5.28921962e-01 -1.63840726e-01 9.81597453e-02
6.63868248e-01 -4.48355854e-01 -2.82250166e-01 5.13882816e-01
5.67898691e-01 -4.66529697e-01 -3.86070430e-01 1.60163975e+00
3.30013812e-01 2.68414378e-01 2.55636513e-01 -9.29717362e-01
-7.54365265e-01 3.47187012e-01 -1.24440730e+00 1.52593061e-01
-2.65916616e-01 2.70885974e-01 8.00041616e-01 -4.08616483e-01
8.40160251e-02 9.86968517e-01 4.68003094e-01 -1.21197440e-01
-1.21815324e+00 -6.60274446e-01 7.45084882e-02 2.64239311e-01
-1.21825027e+00 -1.53638929e-01 9.82638955e-01 -2.63621032e-01
4.04778749e-01 3.11296016e-01 7.95953810e-01 4.79703158e-01
1.64412364e-01 5.42227983e-01 1.58392143e+00 -2.52801061e-01
-1.91249400e-01 -1.08768128e-01 1.44530311e-02 4.07404989e-01
1.66358635e-01 4.75383759e-01 9.55191255e-02 1.15740381e-01
8.98125291e-01 3.15092057e-01 -7.32026398e-01 -2.68529028e-01
-9.62607980e-01 5.87296665e-01 8.30414772e-01 2.87103206e-01
-6.39146566e-01 -3.09166133e-01 -4.80402447e-02 3.06410700e-01
5.28040707e-01 1.73621282e-01 -3.42496932e-01 5.61441660e-01
-1.03470957e+00 -5.86484037e-02 3.12760435e-02 4.82729286e-01
1.22333574e+00 7.39153400e-02 3.56907368e-01 1.20114350e+00
3.34346890e-01 9.06218171e-01 4.35579896e-01 -7.64600456e-01
4.34789807e-01 5.68343878e-01 2.02747315e-01 -1.26161385e+00
-3.15674871e-01 -3.60011995e-01 -1.23331678e+00 2.84472018e-01
-1.72642902e-01 -1.34400889e-01 -9.83024001e-01 1.56033421e+00
2.12639913e-01 6.13083169e-02 2.33727712e-02 1.14625931e+00
4.72806513e-01 1.18877256e+00 -5.80209643e-02 -6.08484149e-01
9.41987693e-01 -1.13792253e+00 -6.81569040e-01 -5.74641287e-01
-1.16606757e-01 -6.52822077e-01 1.06330872e+00 2.97504365e-01
-8.13063502e-01 -6.11719191e-01 -1.26978266e+00 7.03850538e-02
-3.33468914e-01 3.43505263e-01 5.32437742e-01 4.38851207e-01
-7.87853420e-01 5.44490755e-01 -7.95674920e-01 -1.72293261e-01
2.44642809e-01 -2.94783022e-02 -1.87083378e-01 -2.40091324e-01
-1.16634595e+00 8.08990479e-01 6.68392658e-01 5.27011037e-01
-6.52162731e-01 -4.87496734e-01 -7.70463645e-01 1.03239760e-01
4.54710841e-01 -1.84476197e-01 7.31147289e-01 -1.20315778e+00
-1.46833408e+00 6.23874664e-01 2.63024896e-01 -6.25923276e-02
3.92639041e-01 -3.24769467e-01 -1.00788355e+00 3.13865393e-01
8.41559619e-02 3.46038938e-01 1.03288031e+00 -1.36830771e+00
-8.03017318e-01 -4.45198774e-01 -7.06315637e-02 4.93755102e-01
-2.95196891e-01 -1.56244487e-01 -3.16341132e-01 -7.22894669e-01
5.27465999e-01 -5.61526060e-01 -2.03292161e-01 4.92812023e-02
-2.69976169e-01 3.96340698e-01 1.08834267e+00 -1.00605166e+00
7.88551450e-01 -2.18822575e+00 1.14396736e-01 5.01241796e-02
-4.23920117e-02 7.05270529e-01 -5.42665958e-01 1.60676256e-01
-3.58249485e-01 -2.88759887e-01 -7.62768447e-01 2.29723766e-01
-4.70171839e-01 1.46408930e-01 -5.46585679e-01 7.42350936e-01
-2.03546453e-02 6.65649831e-01 -8.97227347e-01 -3.10757518e-01
6.12057447e-01 5.72079539e-01 7.27523193e-02 2.23161697e-01
-2.36797780e-01 5.08038163e-01 -2.48987988e-01 5.47047913e-01
1.40457368e+00 -9.74620506e-03 -1.86380111e-02 -6.47492170e-01
-4.65452224e-01 -2.65272468e-01 -1.16700995e+00 1.42932832e+00
-5.78813434e-01 2.86372721e-01 4.74601954e-01 -8.95754158e-01
9.88262713e-01 5.16183972e-02 3.93051088e-01 -9.87490594e-01
-6.75975755e-02 1.86242342e-01 -2.84317911e-01 -3.71086955e-01
4.74789321e-01 -4.66840416e-01 3.77744526e-01 4.64933485e-01
-2.33071640e-01 -4.60625112e-01 -2.71936040e-02 -3.08215320e-01
2.45815977e-01 1.84892327e-01 1.47849008e-01 -1.22493900e-01
7.75130689e-01 -1.11545287e-01 8.09435487e-01 4.66669172e-01
-2.07150385e-01 4.17582154e-01 5.76074235e-02 -1.96457207e-01
-9.94571030e-01 -1.07832122e+00 -2.06266701e-01 7.31850326e-01
5.64322889e-01 2.06943631e-01 -4.05594915e-01 -3.32583278e-01
-2.65221536e-01 5.34723938e-01 -2.96864659e-01 -1.70996502e-01
-3.70205224e-01 -1.28653550e+00 1.91986084e-01 2.12968916e-01
1.27765810e+00 -8.34548056e-01 -5.75907886e-01 1.52252138e-01
-3.37844461e-01 -1.04837739e+00 -3.66511881e-01 -6.90693036e-02
-9.41468716e-01 -1.13284826e+00 -7.83235490e-01 -4.43776816e-01
4.25724417e-01 9.45267141e-01 4.33512688e-01 -3.90641719e-01
-1.54917002e-01 -2.37710811e-02 -2.96273381e-01 2.99650896e-02
-2.39224449e-01 -2.81589925e-01 -2.16549382e-01 4.13900822e-01
-7.60200201e-03 -8.40533495e-01 -5.82324803e-01 2.88833052e-01
-1.48007345e+00 4.72136885e-01 1.06611729e+00 9.14973378e-01
7.89395630e-01 8.89655769e-01 2.39429533e-01 -5.05613029e-01
1.28283128e-01 -1.48384616e-01 -9.80041504e-01 5.41828454e-01
-6.55958354e-01 -4.05434400e-01 5.74394703e-01 -2.36871988e-01
-1.60769439e+00 7.49461949e-02 -8.44853446e-02 -2.90723503e-01
-1.12128921e-01 6.78566933e-01 -5.62519968e-01 -2.64328033e-01
4.76149559e-01 5.26331484e-01 -7.58316815e-02 -6.29054248e-01
4.96588141e-01 8.90352011e-01 9.73962009e-01 -6.93247989e-02
1.04117072e+00 8.11819732e-01 -1.07330695e-01 -1.19893348e+00
-1.04330826e+00 -5.77163160e-01 -5.21801591e-01 -9.59664509e-02
8.40889335e-01 -1.16814172e+00 -1.42716780e-01 1.15204132e+00
-6.64316595e-01 -4.27759558e-01 -2.42244303e-02 4.56122071e-01
-4.60483909e-01 9.52087760e-01 -5.69764137e-01 -6.82177722e-01
-4.01359499e-01 -8.74405384e-01 7.96117604e-01 6.29941702e-01
7.08728313e-01 -6.92398190e-01 3.70077416e-02 4.43413019e-01
4.74954993e-01 2.86528558e-01 8.33031893e-01 3.62078607e-01
-5.16662896e-01 -1.16174012e-01 -8.57050002e-01 8.37404609e-01
6.16817772e-01 1.69224478e-02 -8.93157125e-01 -4.10375983e-01
4.55935597e-01 -2.63684809e-01 1.16254508e+00 4.77110684e-01
1.21561813e+00 -2.70787448e-01 -1.20873541e-01 1.02981973e+00
1.87690520e+00 1.86894074e-01 7.88037241e-01 5.07706046e-01
8.45228970e-01 5.17197132e-01 7.59351432e-01 1.29824206e-01
-2.62992587e-02 4.32537317e-01 6.01454377e-01 -3.98825765e-01
2.18380373e-02 -2.58353442e-01 2.88672239e-01 5.35951316e-01
-5.11012832e-03 2.45940126e-02 -6.58047497e-01 4.47340846e-01
-1.86557102e+00 -8.82454753e-01 -2.36143440e-01 2.23379803e+00
8.33416045e-01 -3.84716421e-01 -2.92132258e-01 5.20906448e-02
9.29552734e-01 8.25065553e-01 -8.24178517e-01 4.83068466e-01
-7.42903829e-01 -1.53786018e-02 6.92627013e-01 5.04681051e-01
-1.48924339e+00 7.03204691e-01 5.06879044e+00 7.74309158e-01
-1.45930290e+00 7.31116533e-02 4.49728489e-01 2.93245077e-01
-2.04453886e-01 -2.72241812e-02 -2.34560579e-01 3.97515744e-01
4.38309252e-01 -1.60442054e-01 6.88427985e-01 5.40825307e-01
3.29507679e-01 -4.02858436e-01 -2.93478042e-01 1.15261602e+00
-2.13600858e-03 -8.41865301e-01 9.78368744e-02 -1.35322288e-01
9.81751561e-01 2.53071636e-01 1.66529849e-01 -1.55639499e-01
3.13083887e-01 -6.91183567e-01 4.60575879e-01 7.17576504e-01
8.21943402e-01 -6.86583161e-01 5.64880967e-01 4.48590249e-01
-1.18605769e+00 -3.37641180e-01 -6.06070697e-01 2.00256795e-01
1.67736113e-01 9.55736160e-01 2.03046277e-01 1.06015348e+00
7.86277533e-01 1.12374628e+00 -4.65013653e-01 1.10186207e+00
-6.52362525e-01 2.11275756e-01 -3.29504699e-01 8.65557969e-01
2.73174763e-01 -8.87953341e-01 5.75922728e-01 8.45190287e-01
3.84680808e-01 4.02981848e-01 1.56574875e-01 1.12829328e+00
3.11169386e-01 -5.76585121e-02 -2.38673896e-01 -2.92863041e-01
6.15011603e-02 1.55031109e+00 -4.96974289e-01 -2.74831980e-01
-2.98317462e-01 1.02482855e+00 -2.00234532e-01 6.14379346e-01
-6.12312257e-01 -2.97130674e-01 4.82021689e-01 -3.45044315e-01
3.08276951e-01 -1.87816322e-01 -1.76849663e-01 -1.50056303e+00
-1.73116833e-01 -9.04919922e-01 3.75515372e-01 -1.25828397e+00
-1.26033700e+00 4.93712664e-01 -9.11591500e-02 -1.42798913e+00
4.39014137e-01 -4.22931015e-01 -6.25384986e-01 1.33678865e+00
-2.01993632e+00 -1.31387961e+00 -8.62384737e-01 5.61795771e-01
1.27952665e-01 2.16731399e-01 4.74672318e-01 2.24143222e-01
-8.12669516e-01 -1.66699871e-01 4.72447306e-01 -1.92242548e-01
6.11876369e-01 -1.00495481e+00 -9.75437760e-02 1.32366705e+00
-2.99276590e-01 5.15296496e-02 5.89349627e-01 -5.68293452e-01
-9.51568127e-01 -1.45671546e+00 3.69311690e-01 3.87227237e-01
5.46371460e-01 3.44130635e-01 -1.17960703e+00 5.03642917e-01
4.16184179e-02 -2.79260240e-02 2.59614259e-01 -4.79040116e-01
-2.95646787e-01 -6.16188645e-01 -1.00067794e+00 4.10111219e-01
5.07390797e-01 -6.16842389e-01 -5.44326901e-01 3.57138306e-01
5.26754498e-01 -2.89518267e-01 -6.84838057e-01 7.49878645e-01
3.53528947e-01 -1.18753743e+00 8.79493535e-01 9.30872709e-02
6.83546543e-01 -6.53488576e-01 -2.82580346e-01 -1.59521866e+00
-4.32521969e-01 -1.53738841e-01 2.94295788e-01 9.99027669e-01
-6.68852031e-02 -7.32542634e-01 2.86316872e-01 1.89064339e-01
-1.45854037e-02 -2.46054277e-01 -4.34647948e-01 -7.92397857e-01
1.44843594e-03 -1.30746767e-01 6.15307629e-01 1.07581103e+00
-6.40100956e-01 2.25592047e-01 -6.20710909e-01 8.09136271e-01
9.69164431e-01 6.48673117e-01 4.54375237e-01 -1.15435207e+00
-4.03696686e-01 -5.10743976e-01 1.14915058e-01 -9.28753555e-01
-2.61200350e-02 -4.84736592e-01 1.91394508e-01 -1.60786736e+00
3.79710376e-01 -3.72507602e-01 -4.06549573e-01 5.10370433e-01
-3.51372480e-01 2.38625675e-01 1.05757453e-01 4.18086648e-01
-1.18307080e-02 8.64384115e-01 1.38262868e+00 -4.24475700e-01
-2.40857139e-01 -8.25707763e-02 -4.92551059e-01 7.18620896e-01
7.35782802e-01 -1.52189583e-01 -2.86672622e-01 -5.69990277e-01
1.22145154e-01 2.65415967e-01 6.45506144e-01 -1.01054621e+00
7.76231438e-02 -5.29328406e-01 2.68727064e-01 -8.59825015e-01
2.36514315e-01 -8.35125625e-01 4.15605605e-01 2.78702915e-01
9.96819660e-02 -5.84209561e-01 2.17764929e-01 5.88949323e-01
-4.48460162e-01 -1.88830476e-02 1.33108020e+00 -1.61092639e-01
-1.02747047e+00 5.94170272e-01 3.84649783e-02 -5.91705322e-01
8.28387022e-01 -1.51624037e-02 -5.89511514e-01 -3.44577104e-01
-4.63125378e-01 3.30620378e-01 6.74213469e-01 2.79514313e-01
5.28040171e-01 -1.08954847e+00 -6.06175780e-01 3.90013158e-01
7.49190077e-02 9.39640328e-02 9.54759002e-01 8.82018268e-01
-6.20804906e-01 2.34175339e-01 -6.16591573e-01 -4.29891706e-01
-1.01425588e+00 6.00406170e-01 8.31788301e-01 -1.92530736e-01
-7.20424235e-01 5.23767114e-01 4.26728547e-01 -6.46527708e-01
-2.18215957e-01 -1.22199610e-01 -2.47617394e-01 6.13166206e-02
6.49155855e-01 4.34436858e-01 -3.09265573e-02 -8.89091253e-01
-8.55214894e-02 6.54971242e-01 1.95180923e-01 -1.87870394e-02
1.58414531e+00 -4.64731544e-01 -5.58899701e-01 3.12125623e-01
1.09149456e+00 -1.38638899e-01 -1.64495170e+00 -7.02432752e-01
-4.50268209e-01 -7.66413629e-01 6.52626276e-01 -8.86395752e-01
-1.50299573e+00 1.04301715e+00 8.33356559e-01 5.15683964e-02
1.76540220e+00 -4.36886877e-01 8.48034918e-01 2.63387591e-01
1.00926422e-01 -1.08495390e+00 -1.37236089e-01 3.21056962e-01
8.33051682e-01 -1.28457141e+00 2.61247903e-01 -5.03630221e-01
-4.24189478e-01 1.21654928e+00 5.28560996e-01 -1.26768341e-02
5.07671475e-01 -2.20272794e-01 3.60941023e-01 -7.27804899e-02
-5.83451129e-02 -2.21092194e-01 1.75214186e-01 5.49503386e-01
-1.73751384e-01 1.32446811e-01 -1.36089996e-01 2.73598701e-01
1.34598985e-01 -2.58002356e-02 5.52967370e-01 5.02862930e-01
-5.57114124e-01 -7.88407683e-01 -5.22246182e-01 3.80885184e-01
-1.28310367e-01 -3.27585377e-02 7.84966797e-02 4.48208123e-01
2.09381551e-01 9.32008803e-01 -3.33926268e-02 -2.98150778e-01
1.96776375e-01 -4.69119847e-01 2.18315512e-01 -3.43627155e-01
1.64455891e-01 4.13766325e-01 -4.35517073e-01 -3.90393078e-01
-9.33594227e-01 -5.66777766e-01 -6.01743758e-01 -1.31165519e-01
-4.98591900e-01 1.07887378e-02 3.87917459e-01 8.68757248e-01
-1.76038325e-01 2.52426565e-01 9.24241900e-01 -9.44855571e-01
-5.53285182e-01 -9.54348385e-01 -1.39704657e+00 1.69285595e-01
5.61638236e-01 -4.12271053e-01 -4.18748766e-01 -6.78065270e-02] | [10.177998542785645, -1.9597004652023315] |
b38436da-2cff-49fb-b59d-193a3bb336e9 | an-experimental-review-of-speaker-diarization | 2305.18074 | null | https://arxiv.org/abs/2305.18074v1 | https://arxiv.org/pdf/2305.18074v1.pdf | An Experimental Review of Speaker Diarization methods with application to Two-Speaker Conversational Telephone Speech recordings | We performed an experimental review of current diarization systems for the conversational telephone speech (CTS) domain. In detail, we considered a total of eight different algorithms belonging to clustering-based, end-to-end neural diarization (EEND), and speech separation guided diarization (SSGD) paradigms. We studied the inference-time computational requirements and diarization accuracy on four CTS datasets with different characteristics and languages. We found that, among all methods considered, EEND-vector clustering (EEND-VC) offers the best trade-off in terms of computing requirements and performance. More in general, EEND models have been found to be lighter and faster in inference compared to clustering-based methods. However, they also require a large amount of diarization-oriented annotated data. In particular EEND-VC performance in our experiments degraded when the dataset size was reduced, whereas self-attentive EEND (SA-EEND) was less affected. We also found that SA-EEND gives less consistent results among all the datasets compared to EEND-VC, with its performance degrading on long conversations with high speech sparsity. Clustering-based diarization systems, and in particular VBx, instead have more consistent performance compared to SA-EEND but are outperformed by EEND-VC. The gap with respect to this latter is reduced when overlap-aware clustering methods are considered. SSGD is the most computationally demanding method, but it could be convenient if speech recognition has to be performed. Its performance is close to SA-EEND but degrades significantly when the training and inference data characteristics are less matched. | ['Stefano Squartini', 'Alessio Brutti', 'Enrico Zovato', 'Giovanni Morrone', 'Samuele Cornell', 'Luca Serafini'] | 2023-05-29 | null | null | null | null | ['speech-separation', 'speaker-diarization'] | ['speech', 'speech'] | [ 2.92317290e-02 4.58589159e-02 1.17327526e-01 -4.96527880e-01
-9.32785511e-01 -4.84001487e-01 6.85511768e-01 -1.91320300e-01
-3.89578760e-01 4.67768699e-01 3.45411092e-01 -6.44236386e-01
-3.69078487e-01 -2.88674265e-01 1.13915056e-02 -8.93820226e-01
1.03759490e-01 1.09308434e+00 1.30325958e-01 -1.62732348e-01
-1.57759666e-01 4.71579283e-01 -1.65915322e+00 2.15657219e-01
8.72756839e-01 6.44142032e-01 3.36926043e-01 7.49651492e-01
-4.21676964e-01 3.45283061e-01 -9.18555737e-01 -2.29221553e-01
7.08241239e-02 -5.00600278e-01 -9.33870971e-01 2.00200021e-01
2.47624040e-01 1.22205243e-01 -4.06736165e-01 7.00164616e-01
1.03062534e+00 4.39215750e-01 7.11916506e-01 -1.35056806e+00
-8.07640105e-02 7.86910474e-01 -1.97840840e-01 2.85634518e-01
2.34948844e-01 -1.51807874e-01 7.81338990e-01 -8.61747980e-01
6.21162295e-01 1.57745600e+00 6.99765861e-01 9.83788967e-01
-1.44022286e+00 -5.79991937e-01 2.41114631e-01 1.03924321e-02
-1.45604479e+00 -8.25588405e-01 6.36557162e-01 -2.28815526e-01
1.33538198e+00 7.24596977e-01 3.03849280e-01 1.15575814e+00
-5.31202137e-01 9.07689273e-01 7.68309295e-01 -6.22295260e-01
5.37233472e-01 3.46182764e-01 3.64184231e-01 2.39381373e-01
-1.95997264e-02 5.00316843e-02 -4.41886753e-01 -3.05096865e-01
1.89191893e-01 -4.49813575e-01 -3.86415094e-01 -8.56077150e-02
-9.83711660e-01 7.44293511e-01 -2.66709298e-01 8.00628483e-01
-3.69903594e-01 -4.49158996e-01 6.51465237e-01 4.95126724e-01
6.05591834e-01 3.18697572e-01 -4.51911151e-01 -5.94056845e-01
-1.36914241e+00 2.77117252e-01 1.23489010e+00 1.04835987e+00
3.63361776e-01 4.84334528e-01 -4.04961944e-01 1.50815570e+00
-3.19125652e-02 3.18514854e-01 6.77194357e-01 -1.07282102e+00
7.67068267e-01 6.70878142e-02 -3.18380713e-01 -8.05991232e-01
-6.10779703e-01 -4.59056377e-01 -1.08231974e+00 -1.06668763e-03
4.74727690e-01 -3.76059949e-01 -8.85832071e-01 1.69607854e+00
4.61293869e-02 -2.79250354e-01 3.83774728e-01 8.85090530e-01
1.02207756e+00 7.50507593e-01 -2.44355410e-01 -5.31896472e-01
1.11145937e+00 -9.17980373e-01 -9.71310556e-01 -2.76318580e-01
6.62835062e-01 -8.12746704e-01 1.03275061e+00 5.62354624e-01
-1.16303277e+00 -6.65921509e-01 -9.37172413e-01 3.54557574e-01
-3.22348714e-01 -5.59219718e-02 4.28262591e-01 1.04620194e+00
-1.41128290e+00 3.30789089e-01 -6.86051369e-01 -7.57200301e-01
-4.97751050e-02 4.58796442e-01 -4.42297645e-02 1.01735085e-01
-1.06383038e+00 8.12572002e-01 3.32980096e-01 -5.68566397e-02
-5.64266920e-01 -3.66300255e-01 -5.87197244e-01 1.70916110e-01
2.94454157e-01 -4.01395053e-01 1.47998476e+00 -7.99021125e-01
-1.84980142e+00 7.15166092e-01 -4.27174449e-01 -5.05794287e-01
4.48867530e-01 1.99708968e-01 -6.01899624e-01 -4.96309139e-02
-1.45265460e-01 6.65165365e-01 6.62415624e-01 -1.59035063e+00
-5.85479617e-01 -2.04642251e-01 -2.65752196e-01 3.14150512e-01
-3.52465063e-01 1.63168926e-02 -8.63532364e-01 -5.89117885e-01
2.54455507e-01 -1.13080239e+00 -6.22464269e-02 -6.36357427e-01
-6.30713761e-01 -4.59205538e-01 1.03449333e+00 -6.23551071e-01
1.76315725e+00 -2.12453413e+00 3.62659305e-01 3.49767894e-01
3.23563606e-01 7.38656819e-01 -2.71960106e-02 5.83683968e-01
-1.70151815e-01 1.16203718e-01 -2.24259660e-01 -8.72996449e-01
3.44037861e-01 5.42823017e-01 1.66356519e-01 1.94067836e-01
-3.35299611e-01 2.82039851e-01 -6.90230489e-01 -5.05770743e-01
3.87123734e-01 4.87677842e-01 -7.95438111e-01 3.62067342e-01
7.89229795e-02 2.28389129e-01 5.14889322e-02 4.20095921e-01
8.02141726e-01 2.49305293e-01 4.54393357e-01 -1.34294048e-01
-1.75614983e-01 5.47635496e-01 -1.22561014e+00 1.44525361e+00
-8.09294820e-01 1.01008701e+00 5.45918465e-01 -1.28785098e+00
9.28116679e-01 7.41949379e-01 2.69474566e-01 -6.27182186e-01
7.76860341e-02 2.74655879e-01 2.87899166e-01 -4.98846769e-01
5.60739756e-01 -6.02725372e-02 9.41808447e-02 2.54285991e-01
5.31872846e-02 -3.19352388e-01 2.74509966e-01 3.80899042e-01
9.29167986e-01 -5.66645920e-01 4.63893339e-02 -3.37040305e-01
4.80660796e-01 -1.21792525e-01 6.91112697e-01 6.96291924e-01
-2.19417691e-01 7.47500718e-01 3.77523541e-01 1.92719787e-01
-8.93023670e-01 -1.03798068e+00 -1.92196712e-01 8.61841440e-01
-1.78869367e-01 -4.76179510e-01 -1.19033790e+00 -5.20434618e-01
-2.28580043e-01 1.11239493e+00 -2.48384420e-02 -1.23146418e-02
-5.38351893e-01 -7.28302479e-01 8.57886493e-01 2.77787983e-01
4.79774892e-01 -7.65260994e-01 1.25383452e-01 2.49925852e-01
-5.30346990e-01 -1.18726552e+00 -5.26933432e-01 6.60915256e-01
-6.52283669e-01 -4.04269040e-01 -8.63567948e-01 -8.92913163e-01
3.33460271e-01 4.38288748e-01 1.16049182e+00 -1.01386920e-01
1.67219087e-01 4.43842143e-01 -5.41711748e-01 -2.71950543e-01
-9.08415020e-01 3.42452139e-01 3.66609991e-01 -1.03437036e-01
6.73582911e-01 -6.69356287e-01 -2.70084172e-01 5.77376604e-01
-6.14809573e-01 -3.36196214e-01 3.36668164e-01 9.74222302e-01
2.54633546e-01 2.95338422e-01 6.06274605e-01 -8.02780151e-01
9.82613385e-01 -3.36058557e-01 -7.81223252e-02 -2.08654776e-02
-7.55599082e-01 -3.99445891e-02 5.46443462e-01 -4.19504225e-01
-1.10559082e+00 -3.84501845e-01 -5.32940447e-01 -6.43461287e-01
-4.74359661e-01 3.91397178e-01 -4.53355432e-01 4.28632736e-01
6.27050638e-01 1.64081186e-01 1.19798832e-01 -7.16190696e-01
5.28274812e-02 1.36081541e+00 3.38820666e-01 -4.54365075e-01
4.00809020e-01 6.31400719e-02 -8.02882016e-01 -1.49493992e+00
-3.20496619e-01 -6.36226475e-01 -4.64055121e-01 -1.79841846e-01
8.47157001e-01 -5.94204903e-01 -5.31028509e-01 5.12189567e-01
-1.14541852e+00 -3.42055649e-01 -1.71290055e-01 6.47606730e-01
-3.87188792e-01 6.19153678e-01 -5.12662709e-01 -1.15714371e+00
-3.13206494e-01 -1.31215763e+00 7.86618710e-01 -2.53808767e-01
-6.98138714e-01 -1.05369282e+00 -7.89238587e-02 6.51578069e-01
4.52804655e-01 -3.28415275e-01 9.59283590e-01 -1.08125949e+00
1.25070840e-01 1.16579182e-01 1.89685505e-02 6.46176636e-01
2.55352944e-01 6.11168481e-02 -1.18608630e+00 -3.62517387e-01
5.00019602e-02 1.07090250e-01 5.32911301e-01 6.00697398e-01
1.08649254e+00 -2.87566036e-01 -3.02625179e-01 3.46087962e-01
8.79444599e-01 7.61871457e-01 5.06219566e-01 1.25820324e-01
4.08504099e-01 9.46382880e-01 5.79427660e-01 1.96512163e-01
3.05629969e-01 1.16309845e+00 -6.04981184e-02 -4.39758986e-01
-4.90069270e-01 2.36079350e-01 4.12616491e-01 1.57825458e+00
5.21599092e-02 -9.18752611e-01 -1.03654945e+00 5.61766505e-01
-1.82159758e+00 -1.15197849e+00 -2.45223135e-01 2.15692329e+00
7.17431724e-01 2.89301306e-01 6.28206372e-01 6.09993100e-01
9.24974084e-01 2.02514250e-02 -2.78667241e-01 -8.99645209e-01
-2.04419091e-01 4.59813513e-02 7.76811764e-02 6.66872144e-01
-7.66806245e-01 8.09961438e-01 6.44700241e+00 1.22852075e+00
-8.42108369e-01 2.61795670e-01 3.65320176e-01 -2.19649330e-01
-2.92645931e-01 -3.42564851e-01 -8.54923189e-01 6.37698591e-01
1.32016051e+00 9.87993777e-02 4.70450848e-01 7.34737337e-01
5.51323593e-01 -2.94268318e-02 -1.00648928e+00 1.28061008e+00
1.99754819e-01 -1.09398329e+00 -8.86978507e-02 6.40077740e-02
4.42202121e-01 2.33943071e-02 -2.76973277e-01 4.69499677e-01
3.23541045e-01 -7.01927483e-01 5.13629317e-01 8.77679288e-02
6.42739892e-01 -9.35776055e-01 8.05851638e-01 3.27872813e-01
-9.66020584e-01 8.00849795e-02 -2.52318591e-01 2.03277633e-01
2.42304549e-01 7.35472083e-01 -8.64539981e-01 6.43420279e-01
7.73366332e-01 1.47240013e-01 5.70207573e-02 7.55418539e-01
3.20764184e-01 1.08911622e+00 -3.72039974e-01 -1.09252632e-01
4.96681362e-01 -3.45528066e-01 8.00369382e-01 1.72247148e+00
1.76273316e-01 -9.10393670e-02 2.01116323e-01 5.00184953e-01
2.77511835e-01 -1.70893788e-01 -4.73720700e-01 9.15230960e-02
9.23549831e-01 9.11796629e-01 -6.64005041e-01 -5.74972808e-01
-2.37424657e-01 9.93962109e-01 9.11827907e-02 5.17280459e-01
-9.06288922e-01 -5.85163832e-01 1.09135008e+00 -7.96712656e-03
4.69646245e-01 -2.68773407e-01 -1.49275973e-01 -7.45543480e-01
-1.38211399e-01 -1.19452667e+00 2.20076054e-01 -4.81489480e-01
-1.15392005e+00 1.18602669e+00 1.60388440e-01 -1.25525665e+00
-4.89203095e-01 -1.92263231e-01 -5.16191781e-01 7.73099542e-01
-1.04438508e+00 -5.08329928e-01 -2.32235461e-01 6.13549531e-01
1.07343972e+00 -4.56002980e-01 9.08791840e-01 7.59480596e-01
-9.37489331e-01 9.81580794e-01 4.91341412e-01 -7.07669780e-02
6.37624025e-01 -1.17153478e+00 3.08029324e-01 6.97999299e-01
1.54226959e-01 5.12376726e-01 7.96887219e-01 -2.71496415e-01
-8.90059531e-01 -9.65425968e-01 1.13939643e+00 -2.11455837e-01
1.48330778e-01 -7.14978755e-01 -9.85175014e-01 2.52388477e-01
3.76745164e-01 -7.80190647e-01 6.39577806e-01 4.21369255e-01
1.71445496e-02 -8.15413222e-02 -9.65385020e-01 6.00225985e-01
1.25393224e+00 -6.43886566e-01 -5.67676783e-01 2.84437180e-01
8.52108002e-01 -1.70324147e-01 -7.75770545e-01 -6.70625046e-02
1.96412414e-01 -1.19948959e+00 9.13472056e-01 -1.91165254e-01
-2.91544676e-01 -1.93350062e-01 -3.51585716e-01 -1.44091821e+00
-4.08768326e-01 -7.53091216e-01 1.53188989e-01 1.78845513e+00
6.44911706e-01 -6.15950882e-01 7.49969661e-01 4.25314754e-01
-5.25086999e-01 -4.48226839e-01 -1.08103788e+00 -1.26850843e+00
9.84006971e-02 -8.03334117e-01 4.74425346e-01 1.03071070e+00
1.06349504e-02 5.51680386e-01 -2.82229900e-01 8.93933848e-02
3.38977069e-01 -1.39050722e-01 7.80983925e-01 -1.01631188e+00
-1.93012834e-01 -6.33042276e-01 -2.42724344e-01 -1.24702835e+00
2.57929265e-01 -9.96734440e-01 1.57598779e-01 -1.54263508e+00
-3.59127343e-01 -6.15758419e-01 2.93922871e-01 2.30535239e-01
2.21190620e-02 -1.33474842e-01 1.34650230e-01 1.35198742e-01
-2.86195725e-01 4.81653363e-01 7.48276114e-01 -3.05268973e-01
-6.90234125e-01 3.41089040e-01 -3.19296002e-01 5.84386289e-01
8.37740898e-01 -3.93366009e-01 -6.90243959e-01 -3.51672828e-01
-5.91107607e-01 5.24777919e-02 -6.77218288e-02 -1.09719479e+00
4.07747656e-01 3.05935383e-01 -2.07459092e-01 -8.40825260e-01
6.80867195e-01 -6.53756142e-01 1.62730992e-01 2.74996102e-01
-2.30724901e-01 -9.27359983e-02 2.56026030e-01 3.04502040e-01
-4.94088173e-01 -8.58734623e-02 8.02119613e-01 1.96847558e-01
-5.72268009e-01 -1.14576057e-01 -1.16200233e+00 2.25080132e-01
5.96683681e-01 -7.37517595e-01 1.08396202e-01 -7.69037187e-01
-8.85961413e-01 6.03169873e-02 -3.35508622e-02 4.31964248e-01
4.94318902e-01 -1.09960353e+00 -6.38138473e-01 2.06708074e-01
-1.09072059e-01 -1.19608097e-01 3.84904176e-01 1.03318977e+00
-2.68766433e-01 6.93016827e-01 3.44673902e-01 -8.81929517e-01
-1.63472164e+00 5.01602113e-01 2.73132563e-01 -1.46581486e-01
-4.45299357e-01 9.94599283e-01 1.18576279e-02 -9.12439227e-01
8.61425698e-01 -4.30094063e-01 -9.57313105e-02 3.23884428e-01
1.00140586e-01 7.70745933e-01 5.27897656e-01 -5.83466947e-01
-6.07518077e-01 2.92972118e-01 -1.23422861e-01 -2.58225918e-01
1.09593022e+00 -4.41554695e-01 2.56745607e-01 4.30647254e-01
1.21067631e+00 1.57182738e-02 -7.69479871e-01 -2.10056052e-01
8.73472616e-02 -2.42526874e-01 3.28013599e-01 -7.52463222e-01
-1.20835662e+00 7.46396363e-01 5.19178092e-01 5.82655132e-01
1.10327470e+00 6.16621878e-03 8.32257092e-01 3.61483216e-01
1.63806781e-01 -1.40530264e+00 -1.91378772e-01 5.52928030e-01
8.35126698e-01 -1.00733888e+00 -4.05578434e-01 -4.14938599e-01
-8.59478951e-01 8.35954964e-01 5.01666963e-01 5.19973040e-01
8.22052181e-01 3.85707796e-01 3.15171152e-01 -1.47944152e-01
-7.44808972e-01 -4.41248178e-01 8.91871899e-02 8.34950507e-01
3.88380736e-01 1.20208152e-02 -1.56239212e-01 5.02360821e-01
-4.83062744e-01 -6.01272225e-01 1.06259145e-01 6.58455789e-01
-1.70101076e-01 -1.16070175e+00 -5.36784530e-01 3.91278118e-01
-4.05197069e-02 -1.60113856e-01 -5.96551716e-01 8.41452599e-01
1.05721183e-01 1.71735418e+00 2.69803911e-01 -7.93142200e-01
5.59042752e-01 1.86027139e-01 -1.89956799e-02 -5.90396166e-01
-9.05171216e-01 3.72452617e-01 8.99705172e-01 -3.59168410e-01
-5.81850111e-01 -7.85697162e-01 -1.09753513e+00 -8.40758204e-01
-7.24014401e-01 6.25844359e-01 7.65790522e-01 9.16245580e-01
5.84090889e-01 7.38576114e-01 7.24999428e-01 -8.56267154e-01
-3.01697433e-01 -1.07022405e+00 -6.96151197e-01 1.83883697e-01
1.82017982e-01 -6.05656326e-01 -8.60002995e-01 -2.59727538e-01] | [14.613312721252441, 6.255340576171875] |
b4da51e5-6e46-4875-b500-d87c274fbee5 | interpretable-network-structure-for-modeling | null | null | https://openreview.net/forum?id=BkgUB1SYPS | https://openreview.net/pdf?id=BkgUB1SYPS | Interpretable Network Structure for Modeling Contextual Dependency | Neural language models have achieved great success in many NLP tasks, to a large extent, due to the ability to capture contextual dependencies among terms in a text. While many efforts have been devoted to empirically explain the connection between the network hyperparameters and the ability to represent the contextual dependency, the theoretical analysis is relatively insufficient. Inspired by the recent research on the use of tensor space to explain the neural network architecture, we explore the interpretable mechanism for neural language models. Specifically, we define the concept of separation rank in the language modeling process, in order to theoretically measure the degree of contextual dependencies in a sentence. Then, we show that the lower bound of such a separation rank can reveal the quantitative relation between the network structure (e.g. depth/width) and the modeling ability for the contextual dependency. Especially, increasing the depth of the neural network can be more effective to improve the ability of modeling contextual dependency. Therefore, it is important to design an adaptive network to compute the adaptive depth in a task. Inspired by Adaptive Computation Time (ACT), we design an adaptive recurrent network based on the separation rank to model contextual dependency. Experiments on various NLP tasks have verified the proposed theoretical analysis. We also test our adaptive recurrent neural network in the sentence classification task, and the experiments show that it can achieve better results than the traditional bidirectional LSTM. | ['Ming Zhou.', 'Yuexian Hou', 'Nan Duan', 'Yehua Zhang', 'Xiaoliu Mao', 'Peng Zhang', 'Xindian Ma'] | 2019-09-25 | null | null | null | null | ['sentence-classification'] | ['natural-language-processing'] | [-5.48271127e-02 -2.91182578e-01 -3.44380915e-01 -4.55680400e-01
1.16653256e-01 -1.86111853e-01 3.56811970e-01 1.29939755e-02
-3.02999973e-01 2.76303798e-01 4.84765172e-01 -6.21619880e-01
-3.43217790e-01 -7.61870086e-01 -4.19232637e-01 -8.29974294e-01
7.21532777e-02 1.93730555e-02 9.68088135e-02 -3.43782872e-01
4.11407262e-01 3.54896665e-01 -1.03440106e+00 3.98575604e-01
9.85781014e-01 9.71792638e-01 5.55403531e-01 2.09585026e-01
-4.26300168e-01 8.48525882e-01 -4.63652045e-01 -1.08900383e-01
-1.86159387e-01 -6.01367295e-01 -8.55277538e-01 -2.16886044e-01
-3.45570266e-01 3.93510871e-02 -3.43134791e-01 1.18738198e+00
2.28069365e-01 8.64470154e-02 3.96541297e-01 -9.34952140e-01
-7.14663208e-01 1.22781682e+00 -1.77788943e-01 5.26667535e-01
-9.85786915e-02 -3.56208980e-01 1.18532884e+00 -7.20738471e-01
7.42834583e-02 1.31380975e+00 2.74071693e-01 2.28077680e-01
-9.96091545e-01 -6.05124056e-01 5.33461988e-01 5.32937944e-01
-1.24018097e+00 -1.03603173e-02 1.05675077e+00 -2.60139495e-01
7.79516339e-01 1.45764127e-01 6.22148752e-01 8.99544537e-01
3.64340305e-01 6.97255492e-01 1.15541089e+00 -6.33828700e-01
-4.88002934e-02 5.44972792e-02 6.59556031e-01 8.16011727e-01
1.19706757e-01 -1.61930069e-01 -3.91353637e-01 1.49322286e-01
8.92383754e-01 4.76709791e-02 -2.83892751e-01 2.88841695e-01
-1.05941844e+00 8.18556488e-01 8.04529786e-01 8.76361489e-01
-1.19733490e-01 2.42328972e-01 5.64886808e-01 2.90502280e-01
4.55103934e-01 4.61678267e-01 -5.07277131e-01 8.51722881e-02
-4.87547666e-01 -4.13560152e-01 4.69490618e-01 4.37209219e-01
5.85013986e-01 9.90397036e-02 -1.21559449e-01 9.59980965e-01
3.37683022e-01 1.23784505e-01 8.30164015e-01 -5.09490013e-01
6.31340921e-01 9.49683726e-01 -3.72475743e-01 -1.38692212e+00
-6.81152701e-01 -5.33954918e-01 -1.22109473e+00 -4.57677126e-01
2.79784888e-01 -3.85150798e-02 -5.53415537e-01 1.78748715e+00
-4.63899784e-02 1.01339228e-01 -5.88349737e-02 1.03923118e+00
5.59996128e-01 8.98389280e-01 8.58767033e-02 -3.06398809e-01
1.59150863e+00 -8.04729283e-01 -9.42030430e-01 -2.44260639e-01
1.01167798e+00 -7.17122197e-01 1.32050633e+00 1.84982330e-01
-6.53451681e-01 -6.29931688e-01 -1.02938926e+00 1.50988623e-02
-3.15510213e-01 4.36433256e-01 8.16727936e-01 5.23454845e-01
-6.62004471e-01 4.44242090e-01 -8.28103483e-01 -1.49184346e-01
-2.42464498e-01 3.23753536e-01 8.12664703e-02 2.36293212e-01
-1.77946949e+00 9.27359343e-01 7.27545679e-01 8.76718581e-01
-2.55262464e-01 -1.22262858e-01 -5.13841391e-01 1.67056873e-01
2.57640243e-01 -5.90100229e-01 7.44577348e-01 -1.03487778e+00
-1.37807095e+00 2.28229567e-01 -1.81720316e-01 -3.81539166e-01
-7.62148425e-02 -1.99670717e-02 -4.66563404e-01 8.91087502e-02
-3.95065933e-01 2.98490942e-01 5.73800862e-01 -8.54294479e-01
-4.66115475e-01 -5.17047048e-01 4.60758805e-01 2.52566278e-01
-9.61182296e-01 1.10458001e-01 -4.02528733e-01 -7.25575984e-01
4.44013685e-01 -1.05166543e+00 -3.31043273e-01 -5.75652361e-01
-5.13048828e-01 -5.57038009e-01 6.62774026e-01 -5.57579994e-01
1.64883840e+00 -2.07850599e+00 4.41546232e-01 2.63858914e-01
1.20446034e-01 1.60752207e-01 -1.20937124e-01 2.83855021e-01
-1.29727051e-01 4.39606637e-01 -1.27205953e-01 4.13132980e-02
-2.69669861e-01 4.20726240e-01 -5.60580313e-01 2.08622500e-01
-1.78762730e-02 8.48541498e-01 -5.70053458e-01 -4.90013301e-01
-8.16092044e-02 5.29410660e-01 -2.89804220e-01 2.43698016e-01
-1.20551147e-01 2.41953388e-01 -8.44916403e-01 1.57280609e-01
3.54817152e-01 -3.58682424e-01 4.01461959e-01 -4.60593551e-01
-1.06850125e-01 4.96628702e-01 -9.90256846e-01 1.24830437e+00
-7.32141733e-01 7.48265624e-01 -1.75232857e-01 -1.23404884e+00
1.31891859e+00 3.49194735e-01 6.17412925e-02 -7.30974793e-01
2.79653788e-01 2.60482728e-01 4.21143740e-01 -7.27942169e-01
4.11801696e-01 -1.82580888e-01 9.66569856e-02 3.88314277e-01
-4.11753744e-01 3.25670719e-01 1.15124084e-01 1.25279999e-03
6.96812272e-01 -2.84200758e-01 -1.06023148e-01 -3.77639771e-01
9.27837431e-01 -3.82157445e-01 3.84707361e-01 3.07338923e-01
7.38957673e-02 3.70014012e-02 7.35691071e-01 -5.38255215e-01
-8.04311812e-01 -2.84661353e-01 -2.21173882e-01 1.31747067e+00
2.90316254e-01 -3.18262160e-01 -6.74822807e-01 -4.68547851e-01
-5.15275359e-01 5.74093878e-01 -6.43845558e-01 -4.15162802e-01
-8.76242340e-01 -1.04299819e+00 6.37279153e-01 7.07857013e-01
7.35435903e-01 -1.18040967e+00 -3.18524182e-01 9.43351611e-02
-5.21938920e-01 -1.25346375e+00 -3.31201375e-01 1.69323996e-01
-1.19538939e+00 -8.19263875e-01 -2.19863996e-01 -9.69742060e-01
7.64628291e-01 2.47619152e-01 7.10006833e-01 6.54600561e-01
2.98707992e-01 -1.05424851e-01 -5.21170974e-01 2.44744457e-02
-2.38113791e-01 4.89522457e-01 6.10749684e-02 3.90840508e-02
2.06548318e-01 -5.71512163e-01 -4.51874703e-01 5.08277237e-01
-1.22037876e+00 2.11248919e-01 6.54658437e-01 7.70819902e-01
2.87225395e-01 3.04791927e-01 5.54402411e-01 -6.22827888e-01
1.15995789e+00 -1.88457251e-01 -4.40857142e-01 3.92703086e-01
-7.31090486e-01 5.96227765e-01 9.05630350e-01 -5.46561062e-01
-1.18829882e+00 -4.81282532e-01 -2.93966204e-01 -7.86825120e-02
3.07669103e-01 1.18568885e+00 -1.13195784e-01 1.20179646e-01
2.38876030e-01 2.50499934e-01 -3.45792413e-01 -4.12739277e-01
2.90688753e-01 5.44109404e-01 -1.33771086e-02 -7.57815301e-01
4.11741227e-01 2.16164142e-01 1.77792460e-01 -7.96440065e-01
-9.65553105e-01 -6.47845045e-02 -7.04963863e-01 -1.17405437e-01
7.77177036e-01 -3.53254914e-01 -9.47801828e-01 1.93899125e-01
-1.51324022e+00 -6.83583915e-02 4.27127600e-01 7.36293733e-01
-6.88816234e-02 4.37962055e-01 -8.77912581e-01 -8.01566005e-01
-3.98162544e-01 -1.26587808e+00 4.53632146e-01 7.57008418e-02
1.07040174e-01 -1.35013378e+00 -1.72182932e-01 1.94514751e-01
3.98591310e-01 -2.01474607e-01 1.49066401e+00 -6.36087716e-01
-5.57523191e-01 -2.19112523e-02 -4.29752141e-01 4.39425737e-01
-1.59726471e-01 1.05284430e-01 -5.72528124e-01 6.72408640e-02
3.57234120e-01 2.07012556e-02 9.12778974e-01 4.17475671e-01
1.25352001e+00 -3.58297974e-01 -2.62454957e-01 4.25279796e-01
1.16105938e+00 3.92574221e-01 7.12288201e-01 3.80151898e-01
7.69371033e-01 7.14310706e-01 4.12063807e-01 1.19300343e-01
2.82861203e-01 6.35571539e-01 3.57616186e-01 -6.97996765e-02
2.07786947e-01 -2.62109011e-01 4.22768921e-01 1.56020772e+00
-2.78871089e-01 -1.45255148e-01 -8.98592293e-01 1.62703276e-01
-1.93169737e+00 -6.11556113e-01 -2.84218848e-01 1.97413933e+00
7.22375393e-01 4.47141707e-01 -2.27655798e-01 1.72781751e-01
8.42341602e-01 2.55691797e-01 -2.15889588e-01 -6.97504938e-01
-1.79480582e-01 -3.58071327e-01 1.59527212e-01 6.32278979e-01
-4.58455503e-01 9.82105672e-01 5.70671225e+00 8.91523361e-01
-1.46923721e+00 -7.65432743e-03 8.09459805e-01 3.21253508e-01
-5.07573247e-01 1.65354535e-01 -8.84622395e-01 4.45500761e-01
9.26361144e-01 -4.12244163e-02 4.15589780e-01 5.43464124e-01
3.62707168e-01 2.15574577e-01 -9.94912803e-01 5.99445105e-01
-7.89647624e-02 -1.10715878e+00 2.90018082e-01 8.51205643e-03
3.31881613e-01 -1.86045349e-01 -4.25187536e-02 3.33786130e-01
-1.94497049e-01 -9.63300169e-01 3.50548089e-01 6.01788640e-01
3.20420712e-01 -7.55769551e-01 9.83926117e-01 7.60544598e-01
-1.35262012e+00 -3.97921652e-01 -6.90237224e-01 -3.21794301e-01
3.05898767e-02 7.94897497e-01 -7.05972254e-01 3.77260119e-01
4.84843135e-01 4.24917668e-01 -6.19410396e-01 6.09114170e-01
-6.18071258e-01 6.89881742e-01 -2.92929500e-01 -5.71154833e-01
2.40544066e-01 -5.31447530e-01 2.81409055e-01 1.13214219e+00
1.44061491e-01 2.40761474e-01 -4.57024314e-02 8.09676230e-01
-6.34178817e-02 3.08153927e-01 -3.15218687e-01 -1.87806591e-01
5.71918905e-01 1.14799488e+00 -8.46240163e-01 -1.88155621e-01
-8.24501440e-02 5.78745127e-01 6.00632906e-01 5.00210047e-01
-8.60815585e-01 -2.92912155e-01 2.03216374e-01 -1.16115510e-01
2.51074508e-02 -7.20349491e-01 -6.15758240e-01 -1.07230842e+00
1.97209373e-01 -5.98317266e-01 2.20284611e-01 -6.24908864e-01
-1.00791454e+00 1.01841879e+00 -3.07100601e-02 -9.56946850e-01
-7.98628330e-02 -6.96289718e-01 -6.51706338e-01 9.13770139e-01
-1.41969383e+00 -8.93054962e-01 -1.03816185e-02 4.46346313e-01
4.25824225e-01 1.20319642e-01 7.67707705e-01 2.21497595e-01
-1.15665567e+00 4.26901817e-01 -1.83922559e-01 2.60776609e-01
2.04268768e-01 -7.24471211e-01 5.79134841e-03 8.46584499e-01
2.08104163e-01 1.16802847e+00 6.76012516e-01 -3.54041070e-01
-1.37505567e+00 -7.12245941e-01 1.03793418e+00 -1.71564236e-01
9.70254838e-01 -3.01011205e-01 -1.16252053e+00 5.22224307e-01
1.79066151e-01 -2.10314006e-01 4.65914637e-01 5.04433095e-01
-3.95745665e-01 -4.72426474e-01 -4.07025427e-01 7.48785555e-01
9.01133478e-01 -7.32206225e-01 -7.97498584e-01 1.81340992e-01
1.16522515e+00 4.15487029e-02 -8.11904490e-01 5.99330604e-01
5.31379104e-01 -9.52767909e-01 7.49126673e-01 -4.12349015e-01
6.42686963e-01 -2.66602963e-01 -1.69736482e-02 -1.30036151e+00
-4.66690451e-01 8.66845250e-02 1.90339107e-02 1.25231636e+00
6.00791454e-01 -7.63831556e-01 5.05156338e-01 6.41833782e-01
1.90568101e-02 -1.29173183e+00 -9.25577760e-01 -5.59828281e-01
6.56763837e-03 -5.60343504e-01 5.20052016e-01 8.92851532e-01
1.09887831e-01 7.36937940e-01 -2.76968747e-01 2.25880206e-01
-4.38183621e-02 2.90080726e-01 2.23744273e-01 -1.10525346e+00
-1.96013659e-01 -7.48210192e-01 -6.13275990e-02 -1.56555235e+00
3.89165878e-01 -7.75815845e-01 -2.43165374e-01 -1.59970677e+00
1.33617729e-01 -5.60034633e-01 -7.22940147e-01 4.29699212e-01
-2.99004704e-01 -2.76497066e-01 1.00355834e-01 6.25112295e-01
-3.14975530e-01 5.99835157e-01 1.35330391e+00 -2.39911117e-02
-3.22980404e-01 -7.48338504e-03 -6.81053638e-01 7.14867890e-01
7.97119558e-01 -4.19864029e-01 -3.76908422e-01 -7.34679937e-01
8.05662811e-01 2.18616799e-01 7.87813962e-02 -5.93220949e-01
4.00250673e-01 -2.87750274e-01 1.35439456e-01 -4.62828577e-01
1.79535449e-01 -1.02046812e+00 -3.11608255e-01 6.04606390e-01
-6.66663289e-01 2.44951814e-01 -1.09197266e-01 3.54508460e-01
-3.42716962e-01 -3.78495932e-01 3.65365952e-01 1.23371318e-01
-4.95426029e-01 9.74343196e-02 -2.63805240e-01 -3.42370838e-01
3.66398156e-01 8.33388567e-02 -3.96877438e-01 -1.50681466e-01
-5.19102156e-01 1.84001297e-01 -1.56908825e-01 6.52367353e-01
6.36311233e-01 -1.26876390e+00 -4.34313655e-01 1.92955390e-01
-2.22691521e-01 -2.98538327e-01 1.41426533e-01 9.21060920e-01
-3.52887839e-01 6.79693818e-01 1.26709700e-01 -4.52963412e-01
-1.11876094e+00 5.27613163e-01 5.04367709e-01 -5.71436882e-01
-2.62578964e-01 6.62066579e-01 2.44878337e-01 -2.06925482e-01
2.37423256e-01 -7.07894266e-01 -6.07518911e-01 3.06598209e-02
3.71000022e-01 1.63032174e-01 5.42349834e-03 -4.01548594e-01
-2.53300399e-01 6.64280593e-01 -1.43829351e-02 5.05538248e-02
1.05868351e+00 -2.29888409e-01 -6.74151659e-01 7.09686935e-01
1.18058634e+00 -5.41249253e-02 -6.07671201e-01 -3.55796397e-01
2.63062000e-01 -1.64219826e-01 1.51985079e-01 -4.64832097e-01
-1.06873322e+00 1.23820972e+00 3.08608025e-01 8.70619893e-01
1.22706664e+00 -1.06894404e-01 7.61717737e-01 7.02580929e-01
1.25856891e-01 -1.02435875e+00 2.31752038e-01 9.71103311e-01
8.94063115e-01 -8.29434276e-01 -8.40919837e-02 -4.52751815e-01
-5.35170436e-01 1.44680297e+00 6.92247808e-01 -8.30221269e-03
6.43478632e-01 1.79005593e-01 6.38507074e-03 -1.68588117e-01
-6.65876508e-01 1.21658802e-01 3.85806322e-01 -1.77895010e-01
8.17965448e-01 1.46537974e-01 -6.58226371e-01 7.97417700e-01
-3.24212611e-01 -3.09393078e-01 2.56424218e-01 1.68580055e-01
-6.01004720e-01 -1.20430696e+00 -4.13624108e-01 7.14825988e-02
-3.64797413e-01 -2.96086580e-01 -2.53619939e-01 4.35869545e-01
1.05700918e-01 1.04784310e+00 -3.56842065e-03 -5.68403840e-01
3.14779788e-01 4.85840347e-03 2.13258058e-01 -3.20592195e-01
-4.94435430e-01 6.08802214e-02 3.42193320e-02 -1.48373038e-01
-4.79635090e-01 3.56228761e-02 -1.55462897e+00 -1.13534600e-01
-7.59300053e-01 4.35583085e-01 7.86734283e-01 1.36623585e+00
1.58873737e-01 6.77602291e-01 8.88868630e-01 -3.47494483e-01
-3.63115579e-01 -1.09320176e+00 -4.40178335e-01 2.31919646e-01
2.32768077e-02 -5.54643691e-01 -4.12160933e-01 -3.99626791e-01] | [10.508101463317871, 9.128493309020996] |
a6e93f80-f23d-4241-8599-196358b6e6dd | self-supervised-domain-adaptation-for-1 | 2107.09372 | null | https://arxiv.org/abs/2107.09372v1 | https://arxiv.org/pdf/2107.09372v1.pdf | Self-Supervised Domain Adaptation for Diabetic Retinopathy Grading using Vessel Image Reconstruction | This paper investigates the problem of domain adaptation for diabetic retinopathy (DR) grading. We learn invariant target-domain features by defining a novel self-supervised task based on retinal vessel image reconstructions, inspired by medical domain knowledge. Then, a benchmark of current state-of-the-art unsupervised domain adaptation methods on the DR problem is provided. It can be shown that our approach outperforms existing domain adaption strategies. Furthermore, when utilizing entire training data in the target domain, we are able to compete with several state-of-the-art approaches in final classification accuracy just by applying standard network architectures and using image-level labels. | ['Daniel Sonntag', 'Alexander Prange', 'Ngoc T. T. Than', 'Truong T. N. Mai', 'Duy M. H. Nguyen'] | 2021-07-20 | null | null | null | null | ['diabetic-retinopathy-grading'] | ['medical'] | [ 3.97911191e-01 4.58985060e-01 -3.94846916e-01 -7.78890967e-01
-8.04199219e-01 -3.42806429e-01 4.65075493e-01 -1.56350806e-03
-5.94907463e-01 1.11472201e+00 2.84353197e-01 -1.86744153e-01
-4.35052842e-01 -5.93312979e-01 -2.71949738e-01 -7.03772962e-01
1.33065850e-01 6.70259237e-01 3.29852939e-01 -1.51015937e-01
3.03229094e-01 8.15895200e-01 -1.50725162e+00 3.91548187e-01
1.12836719e+00 9.20903802e-01 -2.95699924e-01 7.75703728e-01
-1.51461869e-01 8.60274434e-01 -5.32786369e-01 -1.60496801e-01
4.35544223e-01 -4.88680840e-01 -1.13080442e+00 3.28175068e-01
1.20953047e+00 -3.95054549e-01 -4.55876559e-01 9.90251958e-01
8.55508387e-01 6.01336248e-02 9.66884971e-01 -3.10349703e-01
-1.00476623e+00 -3.34339589e-02 -1.02043882e-01 5.32130718e-01
-6.44108951e-02 1.68575406e-01 3.40285897e-01 -4.78642374e-01
1.20987904e+00 8.54674935e-01 5.16064286e-01 1.15866172e+00
-1.63969469e+00 -2.31470287e-01 -5.23575470e-02 3.43962431e-01
-8.79095435e-01 -5.17125487e-01 6.31136596e-01 -9.69409406e-01
7.53299952e-01 -2.50145495e-01 4.46778417e-01 1.11368656e+00
2.89894007e-02 3.36172432e-01 1.61373782e+00 -5.81472874e-01
3.92075568e-01 3.18241537e-01 4.13045734e-01 5.79337955e-01
1.02342710e-01 6.39752984e-01 -2.41029084e-01 8.23229253e-02
1.11588585e+00 -6.68875873e-01 -2.34277681e-01 -1.08343816e+00
-8.73944283e-01 8.95965457e-01 4.43700820e-01 5.34772761e-02
-2.56550908e-01 -4.28232998e-01 5.60577512e-01 5.70186615e-01
5.55707216e-01 5.08318543e-01 -6.05545282e-01 3.35889637e-01
-6.95743918e-01 -6.71036541e-05 7.75585890e-01 7.82988608e-01
5.88459790e-01 -1.01943396e-01 -4.76320624e-01 1.07909667e+00
4.98632751e-02 5.05077541e-02 7.20111907e-01 -1.31027055e+00
-1.57677621e-01 7.69505024e-01 3.80084142e-02 1.80693999e-01
-7.35703290e-01 -7.79248238e-01 -8.91233742e-01 1.06713498e+00
8.34623098e-01 -1.89874113e-01 -1.49854016e+00 1.29754162e+00
4.07005921e-02 1.93638414e-01 5.65970182e-01 8.52589965e-01
1.25388277e+00 -3.83296490e-01 1.91055715e-01 -2.19196573e-01
1.05238342e+00 -9.31705534e-01 -2.02744037e-01 -6.44793063e-02
4.36575502e-01 -5.16053677e-01 7.92040110e-01 5.57975650e-01
-8.87117624e-01 -7.12304413e-01 -7.71441460e-01 -2.26344630e-01
-3.28827828e-01 4.26927269e-01 4.37715709e-01 6.50680661e-01
-1.30937183e+00 4.42131579e-01 -5.05127370e-01 -9.85280752e-01
9.51112449e-01 4.14377630e-01 -6.39610350e-01 -2.83071131e-01
-6.47113979e-01 1.43648982e+00 1.86301216e-01 -4.64360058e-01
-8.12311292e-01 -8.83097708e-01 -5.17011344e-01 -6.17670476e-01
-2.08037332e-01 -1.27702284e+00 1.31879210e+00 -1.08463919e+00
-1.88435578e+00 1.89692509e+00 -1.00717217e-01 -9.02334809e-01
6.62217021e-01 -7.19918460e-02 -7.52247930e-01 6.66586459e-01
-9.41029191e-02 5.75466514e-01 8.61164927e-01 -1.11433530e+00
-9.38079596e-01 -4.22797561e-01 -1.45162597e-01 -8.12728629e-02
-2.12978289e-01 -3.27385291e-02 6.05244711e-02 -3.79393995e-01
-2.60940105e-01 -6.25202894e-01 -5.09739637e-01 5.26575446e-01
-9.11059827e-02 -4.02795672e-01 2.47689158e-01 -5.29479444e-01
7.83864021e-01 -2.11602116e+00 1.44487590e-01 1.81253374e-01
6.27010703e-01 7.01623023e-01 -3.22789550e-01 -2.03635529e-01
-4.12655592e-01 -2.93505162e-01 -2.40624547e-01 -2.54230574e-04
-4.33071434e-01 3.48893672e-01 -8.21649097e-03 6.09296799e-01
3.68782789e-01 5.57613969e-01 -7.41052210e-01 -4.79874223e-01
3.83884519e-01 2.58427799e-01 -3.70177597e-01 2.91757405e-01
1.65069960e-02 1.01086533e+00 -4.17909026e-01 4.30784822e-01
5.34983873e-01 -4.50428605e-01 -9.29739624e-02 -2.59971619e-01
-1.76151231e-01 9.07837003e-02 -6.72879100e-01 1.86593688e+00
-2.63262600e-01 7.40809023e-01 -3.41925502e-01 -1.57307506e+00
1.14595270e+00 3.16515923e-01 5.34048080e-01 -6.13381386e-01
-1.54978130e-02 4.49967444e-01 4.35875766e-02 -7.55873024e-01
-2.60129422e-01 -1.84359089e-01 6.21422350e-01 -2.53714234e-01
6.34905636e-01 1.92090407e-01 3.17328751e-01 -3.01896214e-01
1.23281586e+00 3.78949046e-01 8.60240102e-01 -2.28465542e-01
8.75705123e-01 4.34874773e-01 4.14994329e-01 7.55644202e-01
-6.04084074e-01 7.95962214e-01 5.04819989e-01 -8.35429668e-01
-1.11306965e+00 -1.34630108e+00 -8.51421237e-01 8.26928616e-01
-1.27790153e-01 4.35776830e-01 -5.36223650e-01 -1.04022300e+00
8.10449347e-02 5.48969150e-01 -1.08533955e+00 -1.00657076e-01
-4.08762783e-01 -6.26442671e-01 4.28258449e-01 5.75019121e-01
5.85480630e-01 -6.43075824e-01 -3.75193566e-01 2.38605812e-01
4.10975099e-01 -1.01257944e+00 1.09201103e-01 7.41801932e-02
-1.33698583e+00 -1.30026722e+00 -1.28029430e+00 -1.19333076e+00
7.27206111e-01 -2.05369622e-01 1.43303931e+00 -4.35090125e-01
-6.27607942e-01 6.67109847e-01 -1.76653087e-01 -3.74031216e-01
-7.43217945e-01 1.46456454e-02 -2.12669462e-01 1.39934868e-01
9.67389822e-01 -7.48676896e-01 -7.68612623e-01 2.16382563e-01
-3.68128359e-01 -4.43269193e-01 9.77964044e-01 1.06745279e+00
8.94519687e-01 -4.34170723e-01 7.75568426e-01 -1.23298395e+00
3.89240384e-01 2.00474150e-02 -6.14345849e-01 3.02984178e-01
-9.30856764e-01 8.16474631e-02 2.62091696e-01 -5.33277154e-01
-1.21416163e+00 3.22807729e-01 1.98831260e-01 -3.00874978e-01
-1.05764163e+00 -8.53573717e-03 3.68405282e-01 -7.65421391e-01
1.69117105e+00 2.50267863e-01 4.77716506e-01 -6.92223668e-01
6.45383120e-01 7.94316292e-01 1.04014897e+00 -1.23803601e-01
6.92578852e-01 6.67850375e-01 2.61291713e-01 -6.91323280e-01
-1.12834835e+00 -8.84476781e-01 -1.09886360e+00 -3.13928612e-02
1.15790200e+00 -9.81378436e-01 -7.57747442e-02 5.73035479e-01
-8.42749298e-01 -3.32420498e-01 -8.29414308e-01 7.54956484e-01
-1.15612566e+00 2.58068323e-01 -2.15753555e-01 -1.98531911e-01
8.73420611e-02 -6.19026780e-01 6.17455363e-01 3.69960546e-01
9.43410695e-02 -1.28076887e+00 5.58858812e-01 4.30652678e-01
5.01279831e-01 3.05482805e-01 1.09845710e+00 -9.26982820e-01
-3.89522493e-01 1.60224941e-02 -6.43625081e-01 9.31095660e-01
2.61847556e-01 -3.53749067e-01 -1.00231397e+00 -2.23058954e-01
-2.46820360e-01 -2.84064710e-01 1.20335078e+00 7.83950329e-01
9.01921093e-01 4.28310037e-02 -4.21639085e-01 9.26380992e-01
1.55498910e+00 4.59372401e-02 9.78520870e-01 7.51888514e-01
6.47950396e-02 6.48264170e-01 3.89647037e-01 7.92076737e-02
3.13073695e-01 4.30667281e-01 2.11883217e-01 -8.16742659e-01
-9.34374094e-01 2.33493835e-01 -3.77002537e-01 -2.26583883e-01
-6.17635429e-01 2.47374564e-01 -9.65443730e-01 9.37788427e-01
-1.68825769e+00 -6.86413646e-01 -5.68002164e-02 2.05357790e+00
1.03045750e+00 -7.78420269e-02 3.84453118e-01 -4.61252868e-01
6.44659102e-01 -5.35026312e-01 -8.72492313e-01 -1.95611954e-01
-3.27695251e-01 6.51249528e-01 7.84153163e-01 1.33481085e-01
-1.50451207e+00 7.67833114e-01 7.52897120e+00 2.16931552e-01
-8.94014716e-01 8.08051378e-02 4.17873234e-01 1.28000975e-01
4.26037133e-01 -2.18988821e-01 -6.23102367e-01 -9.03529376e-02
8.50433230e-01 -4.79329787e-02 1.10958099e-01 8.79331946e-01
8.11550617e-02 1.07043102e-01 -1.33452260e+00 8.62493098e-01
3.07465762e-01 -1.38525748e+00 1.33498870e-02 -2.12901793e-02
9.00099397e-01 3.01032484e-01 1.85715631e-01 7.71613568e-02
4.33234721e-01 -1.04429650e+00 -4.69947755e-01 1.05917931e+00
1.13877487e+00 -2.26827905e-01 7.97740698e-01 -3.54463756e-01
-3.70369822e-01 -2.99834758e-01 -6.13087893e-01 3.20739269e-01
-1.44374222e-01 5.24441600e-01 -8.26064706e-01 4.88366723e-01
7.46961415e-01 1.16414535e+00 -8.64224195e-01 1.79295993e+00
-3.03657293e-01 6.14273310e-01 1.64043203e-01 5.93040168e-01
2.08448879e-02 -1.52911276e-01 7.85142899e-01 1.17734909e+00
1.15524665e-01 3.06225102e-02 -7.50292316e-02 5.76192200e-01
1.10247910e-01 2.67221510e-01 -7.55532861e-01 3.71852934e-01
-2.04185292e-01 9.19660807e-01 9.05133039e-02 -3.92669320e-01
-6.86125875e-01 7.32218027e-01 3.64287704e-01 6.48849010e-01
1.30154669e-01 -4.31164503e-01 8.09846103e-01 1.86123699e-01
4.60275531e-01 1.62815258e-01 -6.26505196e-01 -1.03160596e+00
-2.31017172e-01 -7.27188706e-01 6.65052593e-01 -7.90130913e-01
-1.99616814e+00 5.81553936e-01 -3.23558986e-01 -1.83365679e+00
-1.36304811e-01 -1.07749724e+00 -2.08223850e-01 1.05965161e+00
-2.35399175e+00 -1.31992769e+00 -6.01074576e-01 8.94418597e-01
1.68046683e-01 -1.26435089e+00 1.20981765e+00 1.72967449e-01
-1.01701118e-01 5.97269177e-01 4.35694098e-01 2.72500843e-01
1.54436803e+00 -1.53969920e+00 -7.87000060e-02 6.40980959e-01
-3.00226212e-01 3.17375988e-01 4.30646777e-01 -2.82735586e-01
-3.07734549e-01 -1.42088556e+00 7.37006247e-01 -6.37009144e-01
6.47907555e-01 5.23687422e-01 -9.32881653e-01 6.10095859e-01
1.49666607e-01 5.18508852e-01 9.27872121e-01 1.42604023e-01
-5.85846364e-01 -4.02715206e-01 -1.58072352e+00 3.73254955e-01
1.20828319e+00 -3.58859628e-01 -1.09992802e+00 7.19744802e-01
3.24635237e-01 -5.23618639e-01 -1.36798561e+00 4.36949670e-01
2.82586753e-01 -9.46858585e-01 1.07349133e+00 -1.43145776e+00
6.18545890e-01 -2.45134756e-01 1.40817270e-01 -1.27662432e+00
-4.91989851e-01 -6.09913707e-01 -6.75938725e-02 8.29454541e-01
3.56695950e-01 -9.42342877e-01 8.70686173e-01 1.69605464e-01
-2.66153514e-01 -2.97820926e-01 -1.13771343e+00 -8.53074551e-01
6.64971232e-01 3.41449648e-01 -2.40038857e-02 7.78836787e-01
-3.86656523e-01 2.34576076e-01 -2.26054415e-02 2.84161687e-01
9.90862966e-01 1.13418907e-01 6.31948233e-01 -1.92329550e+00
-1.04487054e-01 -5.22446811e-01 -1.09082246e+00 -8.25737953e-01
1.24560311e-01 -9.20411110e-01 -2.70548373e-01 -1.79556894e+00
-8.31224844e-02 -2.57821620e-01 -6.53080344e-01 5.59439182e-01
2.26191670e-01 4.70450163e-01 -4.95064259e-01 2.21303746e-01
-2.81610191e-01 -6.82201162e-02 1.40563130e+00 -2.37295955e-01
-4.38357383e-01 3.33942860e-01 -8.07086408e-01 6.76288068e-01
9.75952446e-01 -2.53361464e-01 -2.77167261e-01 -1.71261638e-01
-3.75775814e-01 -1.27734169e-01 6.49347901e-01 -1.08748078e+00
2.63657570e-01 -8.54833275e-02 4.08210337e-01 -1.38320342e-01
-2.06290390e-02 -4.80984330e-01 -6.46568239e-01 1.90307260e-01
-6.13514781e-01 -8.84088576e-01 1.66990072e-01 6.62038922e-01
-2.82091677e-01 -1.35490656e-01 1.55419540e+00 -1.82448283e-01
-1.02761793e+00 1.82587191e-01 -3.94810021e-01 3.14276129e-01
1.01952279e+00 -5.29038489e-01 -7.94553280e-01 1.97953284e-01
-1.50151289e+00 -8.85440595e-03 5.25713444e-01 1.87581167e-01
4.77232605e-01 -8.98552656e-01 -1.30966425e+00 2.37505019e-01
7.87575364e-01 -1.51678652e-01 3.04171473e-01 9.78152990e-01
-4.72348988e-01 5.65993428e-01 -7.75328815e-01 -7.28551030e-01
-1.32509029e+00 5.16399682e-01 9.83545780e-01 -2.10647494e-01
-8.59935701e-01 7.23565519e-01 1.88800663e-01 -3.95586669e-01
2.25684270e-01 -1.21486470e-01 -8.27034533e-01 -1.43099919e-01
5.11859000e-01 3.09456766e-01 1.88582391e-01 -6.29839748e-02
-5.83146922e-02 1.04329276e+00 -2.72721231e-01 3.07996035e-01
1.34031463e+00 -1.03088826e-01 -9.32722241e-02 -1.11681074e-02
7.32964218e-01 -5.58620572e-01 -1.31414318e+00 -7.98051238e-01
8.59365240e-03 -3.37686002e-01 2.14006320e-01 -1.62466335e+00
-8.47206473e-01 8.29191804e-01 1.69209766e+00 -3.11879784e-01
1.58286977e+00 1.07732952e-01 1.48050532e-01 6.31053567e-01
1.81539401e-01 -1.14997423e+00 -1.73891306e-01 1.84042931e-01
8.08336198e-01 -1.47540641e+00 4.35860716e-02 -4.69604284e-01
-6.78307652e-01 1.35245991e+00 5.66533387e-01 -4.45164889e-01
6.72173023e-01 -1.83163717e-01 8.21561694e-01 1.09574415e-01
-4.76326048e-01 -6.08863056e-01 7.43710399e-01 1.78764606e+00
2.22381949e-01 -2.60861993e-01 -6.12324595e-01 5.43405935e-02
1.47660732e-01 5.57579041e-01 7.75345087e-01 5.18209457e-01
-5.09151101e-01 -1.54679120e+00 1.50474131e-01 8.12329590e-01
-4.29304779e-01 3.52796763e-02 -3.89042288e-01 8.89872015e-01
1.63115442e-01 5.42939603e-01 -4.48165759e-02 2.07765266e-01
5.41514456e-01 1.69073611e-01 5.87451398e-01 -9.39514816e-01
-2.92575091e-01 -1.98375911e-01 3.90113235e-01 -5.33527851e-01
-9.58366930e-01 -5.43655694e-01 -8.94342184e-01 6.11208916e-01
2.44453579e-01 -3.88598204e-01 3.24283987e-01 7.30882049e-01
4.56692398e-01 5.76057374e-01 4.75913882e-01 -2.70314455e-01
-7.83211708e-01 -9.42888975e-01 -7.11879015e-01 3.57776314e-01
8.49205136e-01 -7.54397690e-01 -2.22444668e-01 6.31806195e-01] | [15.774700164794922, -3.937795877456665] |
192fca71-e4ce-457c-a81c-a519da8759ca | facetoponet-facial-expression-recognition | 2209.06322 | null | https://arxiv.org/abs/2209.06322v1 | https://arxiv.org/pdf/2209.06322v1.pdf | FaceTopoNet: Facial Expression Recognition using Face Topology Learning | Prior work has shown that the order in which different components of the face are learned using a sequential learner can play an important role in the performance of facial expression recognition systems. We propose FaceTopoNet, an end-to-end deep model for facial expression recognition, which is capable of learning an effective tree topology of the face. Our model then traverses the learned tree to generate a sequence, which is then used to form an embedding to feed a sequential learner. The devised model adopts one stream for learning structure and one stream for learning texture. The structure stream focuses on the positions of the facial landmarks, while the main focus of the texture stream is on the patches around the landmarks to learn textural information. We then fuse the outputs of the two streams by utilizing an effective attention-based fusion strategy. We perform extensive experiments on four large-scale in-the-wild facial expression datasets - namely AffectNet, FER2013, ExpW, and RAF-DB - and one lab-controlled dataset (CK+) to evaluate our approach. FaceTopoNet achieves state-of-the-art performance on three of the five datasets and obtains competitive results on the other two datasets. We also perform rigorous ablation and sensitivity experiments to evaluate the impact of different components and parameters in our model. Lastly, we perform robustness experiments and demonstrate that FaceTopoNet is more robust against occlusions in comparison to other leading methods in the area. | ['Ali Etemad', 'Alireza Sepas-Moghaddam', 'Mojtaba Kolahdouzi'] | 2022-09-13 | null | null | null | null | ['facial-expression-recognition'] | ['computer-vision'] | [-3.58640142e-02 -1.99322417e-01 -1.25096008e-01 -7.41785944e-01
-4.01349723e-01 -6.59735277e-02 5.48045516e-01 -3.43275279e-01
-2.72759885e-01 2.17064157e-01 1.70864448e-01 2.22218230e-01
2.29978651e-01 -5.18342853e-01 -6.37592793e-01 -8.91945601e-01
-2.60126859e-01 1.45989373e-01 -1.43318981e-01 -2.90394455e-01
-6.19705357e-02 6.96330905e-01 -1.49697649e+00 4.73005563e-01
3.55775654e-01 1.63897920e+00 -3.42427671e-01 1.83779523e-01
-2.19468489e-01 1.22025943e+00 -4.07947749e-01 -8.05060685e-01
1.57775596e-01 -3.45051169e-01 -5.55055082e-01 2.69086242e-01
6.85253620e-01 -3.03741604e-01 -2.84846723e-01 8.53643298e-01
7.16414213e-01 -1.06520817e-01 6.23570085e-01 -1.30744338e+00
-2.09421843e-01 7.06599131e-02 -9.88715172e-01 -9.88873392e-02
9.53043178e-02 6.44990951e-02 9.00924921e-01 -1.23360145e+00
6.75365388e-01 1.39515305e+00 6.70940161e-01 5.60111582e-01
-1.00839388e+00 -1.07457018e+00 2.62049466e-01 1.18242361e-01
-1.31932437e+00 -9.60770965e-01 1.06465518e+00 -2.08007395e-01
5.35940051e-01 -1.65599883e-01 7.35451519e-01 1.20651472e+00
1.81458980e-01 1.01731646e+00 1.02929854e+00 -2.35177904e-01
1.43711120e-01 -1.29002795e-01 -2.20955536e-01 1.35466158e+00
-5.66516280e-01 2.77899504e-02 -1.04398835e+00 -3.51824164e-01
7.34682024e-01 -1.20653354e-01 -2.07464635e-01 -3.69549960e-01
-2.99474657e-01 8.54004383e-01 6.39632404e-01 5.89849465e-02
-5.12992978e-01 3.21469903e-01 5.19718587e-01 3.32986802e-01
6.98776424e-01 3.42821819e-03 -3.61052811e-01 -1.27771229e-01
-8.20802152e-01 1.64253086e-01 7.41365731e-01 5.74064255e-01
9.16015625e-01 8.48973393e-02 -2.29037911e-01 1.32655776e+00
5.25702894e-01 2.95332044e-01 1.86229274e-01 -9.50910509e-01
1.86316624e-01 6.25052512e-01 -3.76830697e-01 -1.07450700e+00
-1.98857799e-01 -2.52311766e-01 -6.75861716e-01 3.15392107e-01
2.78480709e-01 -2.57556349e-01 -9.11845386e-01 2.01829410e+00
4.17703450e-01 4.29276079e-01 -3.17701787e-01 9.55257237e-01
7.13997602e-01 5.59930265e-01 3.21155220e-01 4.16673459e-02
1.33542216e+00 -1.11736691e+00 -4.39539820e-01 -1.57543540e-01
5.33897042e-01 -6.41966641e-01 9.46074069e-01 4.13099915e-01
-9.00253356e-01 -3.51053059e-01 -8.96519661e-01 -5.15689515e-02
-7.45410994e-02 3.64895999e-01 7.25529730e-01 2.46766299e-01
-1.00351501e+00 4.01833653e-01 -7.06056297e-01 -7.45480955e-02
1.10797346e+00 3.84209871e-01 -5.84696293e-01 -9.64191705e-02
-8.09364378e-01 6.47069275e-01 -3.02834213e-01 3.10047925e-01
-1.04144037e+00 -8.29162836e-01 -9.68194783e-01 4.95189801e-02
1.41725197e-01 -5.41289151e-01 1.17799449e+00 -1.59710288e+00
-1.80387509e+00 1.15554667e+00 -5.47131658e-01 -7.59678930e-02
4.07817215e-01 -3.17237675e-01 -2.73316741e-01 4.23350424e-01
-3.14253429e-03 8.94207299e-01 1.24812412e+00 -1.09510481e+00
-4.24587131e-01 -7.34202564e-01 -1.05943225e-01 1.23185301e-02
-4.33462650e-01 2.33881310e-01 -9.35000718e-01 -4.82881337e-01
-1.18620932e-01 -9.51431334e-01 2.55503841e-02 6.35688305e-01
-5.43492138e-02 -3.49101782e-01 1.00308180e+00 -4.45200980e-01
1.05997121e+00 -2.50470996e+00 2.84656525e-01 4.29387331e-01
2.07383096e-01 8.19072053e-02 -3.37622494e-01 1.68370418e-02
-1.60552338e-01 -1.02908149e-01 -1.07813485e-01 -8.51703823e-01
-1.18610054e-01 1.61985159e-01 -2.14047283e-01 5.40059447e-01
4.90010291e-01 8.33684027e-01 -5.49458265e-01 -4.92149860e-01
-3.44294570e-02 7.15291560e-01 -5.43821156e-01 2.88925499e-01
-2.61580616e-01 2.80241966e-01 -5.57149708e-01 9.86423135e-01
5.59453547e-01 -1.48776293e-01 7.44023323e-02 -3.63799930e-01
2.26821482e-01 -2.09449738e-01 -5.91583490e-01 1.71730709e+00
-6.74092293e-01 6.57353938e-01 3.15483272e-01 -7.39992559e-01
1.09791577e+00 2.60082424e-01 5.23768663e-01 -9.47395205e-01
3.78257871e-01 1.70188427e-01 -2.09448084e-01 -5.00863731e-01
9.71775725e-02 -1.95826426e-01 1.34047717e-01 3.41481417e-01
5.05920529e-01 1.69001549e-01 5.73796406e-03 1.32418066e-01
8.33730102e-01 1.57477781e-01 -9.54618901e-02 -2.33091518e-01
4.69262660e-01 -6.78250909e-01 5.40682793e-01 9.44629610e-02
-3.63394231e-01 4.17580485e-01 8.95901561e-01 -6.52992249e-01
-6.87140465e-01 -7.91756272e-01 -9.30446014e-02 1.47997725e+00
-1.61623836e-01 -5.55766046e-01 -7.92120159e-01 -9.05455291e-01
9.75780860e-02 2.29157016e-01 -1.13825190e+00 -1.72908038e-01
-2.98084319e-01 -6.43686950e-01 6.72840178e-01 5.17342031e-01
7.09113955e-01 -1.21521711e+00 -4.85726953e-01 -3.58330384e-02
-4.01454940e-02 -1.15695143e+00 -4.30516183e-01 3.75922566e-04
-3.48927170e-01 -1.06839633e+00 -4.86103088e-01 -6.58290684e-01
7.82413721e-01 -2.39837512e-01 9.50067937e-01 2.56543726e-01
-2.31366903e-01 3.64373267e-01 -2.65015602e-01 -4.91435021e-01
8.37362856e-02 -1.12327456e-01 -1.87711403e-01 8.47639263e-01
4.14293289e-01 -6.03129447e-01 -5.39110541e-01 3.74494284e-01
-6.85581684e-01 -1.74195662e-01 4.56147283e-01 8.04471791e-01
5.69372058e-01 -2.05236718e-01 2.86569357e-01 -8.06884527e-01
4.59102541e-01 -4.85266387e-01 -4.21520621e-01 9.83615816e-02
-1.19275451e-02 1.82353631e-02 5.22513032e-01 -3.61767501e-01
-9.26732242e-01 2.18755543e-01 -3.13640058e-01 -8.03069353e-01
1.05977476e-01 4.15996522e-01 -4.09189701e-01 -4.35358286e-01
4.14210528e-01 1.68230815e-03 3.29579800e-01 -3.87348950e-01
7.96976089e-02 3.94450098e-01 4.38212454e-01 -8.17168057e-01
4.13703203e-01 6.72482073e-01 1.00120097e-01 -1.04121590e+00
-9.95970070e-01 -1.10633478e-01 -4.62358832e-01 -3.56662273e-01
6.19523823e-01 -9.79066014e-01 -7.15701520e-01 9.08614755e-01
-9.76595461e-01 -5.38064837e-01 5.64327165e-02 -1.83656111e-01
-5.24004877e-01 -1.68943465e-01 -7.89380848e-01 -7.14214206e-01
-4.45419252e-01 -1.20678139e+00 1.48144376e+00 3.24795693e-01
-2.35486969e-01 -9.28458989e-01 -6.19737245e-02 2.39523172e-01
4.16931093e-01 4.01401877e-01 8.40471745e-01 -2.19706357e-01
-1.67587787e-01 2.68739406e-02 -2.07280487e-01 3.35710794e-01
-5.85002974e-02 3.88518095e-01 -1.26342928e+00 -1.98684379e-01
-1.53578967e-01 -9.00322437e-01 1.03580964e+00 1.39767632e-01
1.56208420e+00 -2.59055227e-01 -1.29468679e-01 1.09701025e+00
1.16500711e+00 -4.77863289e-02 8.39903593e-01 1.54328853e-01
6.72968686e-01 7.29802191e-01 4.75703895e-01 5.33355057e-01
2.29351953e-01 8.72747958e-01 3.59329373e-01 -3.82459521e-01
5.31959124e-02 -3.22837412e-01 5.23382068e-01 3.20782870e-01
-3.20062153e-02 9.93328318e-02 -6.47540033e-01 2.44082436e-01
-1.67221570e+00 -7.65968502e-01 5.84707320e-01 1.59205830e+00
8.42940569e-01 -1.49436504e-01 1.88237101e-01 -1.23636216e-01
9.80515629e-02 4.22463357e-01 -6.10480070e-01 -4.07077730e-01
-2.19299480e-01 6.52370930e-01 -8.15030262e-02 3.35692555e-01
-1.12693286e+00 1.33523417e+00 5.72252512e+00 8.28928590e-01
-1.85855091e+00 -8.18379223e-02 1.09271455e+00 -3.35657656e-01
1.07763924e-01 -3.56411576e-01 -5.68297386e-01 4.34296668e-01
6.94215834e-01 1.85688660e-01 1.69888213e-01 9.91021276e-01
1.52043477e-01 -2.59685926e-02 -1.02187777e+00 1.09949791e+00
1.84250265e-01 -1.16907561e+00 1.34420633e-01 -3.60289216e-03
4.84378815e-01 6.34069145e-02 2.89725602e-01 2.77842820e-01
8.35781991e-02 -1.32042301e+00 6.91070557e-01 5.03963053e-01
9.31116879e-01 -9.93049085e-01 5.98077238e-01 -1.72929928e-01
-1.03684485e+00 -9.38633531e-02 -2.37887263e-01 2.67732084e-01
-5.89510314e-02 3.54704887e-01 -5.54472208e-01 1.56438798e-01
8.59766126e-01 8.32986474e-01 -5.25126517e-01 5.77043176e-01
-5.36746860e-01 7.73485601e-01 -3.66994232e-01 1.02136135e-01
3.21848899e-01 -1.61412820e-01 1.65613174e-01 1.16123521e+00
7.73055851e-02 1.87231883e-01 -7.14348257e-02 7.38196373e-01
-4.47592020e-01 1.71187669e-01 -4.30651695e-01 8.36578310e-02
2.75966406e-01 1.51198781e+00 -1.64898843e-01 -1.77036002e-01
-3.99725527e-01 1.06360853e+00 6.66934609e-01 5.45930982e-01
-8.70119095e-01 -7.47231618e-02 1.03685641e+00 2.35588998e-01
4.85616356e-01 3.10791433e-02 1.60685971e-01 -9.01825249e-01
5.82855269e-02 -1.16719985e+00 3.03527474e-01 -6.94724202e-01
-1.28357160e+00 8.15261066e-01 -3.48212063e-01 -5.77593207e-01
-1.27317041e-01 -6.41031742e-01 -6.47357106e-01 7.52874374e-01
-1.55217242e+00 -1.39048648e+00 -4.50947016e-01 8.85579646e-01
2.79864162e-01 -4.08109903e-01 8.76347303e-01 3.76544982e-01
-8.85477483e-01 9.58696306e-01 -4.68273431e-01 5.06541550e-01
7.05556929e-01 -8.08636963e-01 -2.29175115e-04 4.19179291e-01
3.00171375e-01 5.16744256e-01 1.86903477e-01 -1.42562285e-01
-1.34729993e+00 -1.10897827e+00 5.36643863e-01 -1.67584643e-01
5.61799228e-01 -5.80201924e-01 -8.96246374e-01 5.13658047e-01
-3.59741449e-02 3.91276002e-01 7.54038393e-01 2.31421024e-01
-7.81925797e-01 -4.82363015e-01 -9.90477443e-01 5.09621739e-01
9.95262504e-01 -5.94898582e-01 5.53463325e-02 -1.20548591e-01
2.82347322e-01 -3.67320895e-01 -7.31992304e-01 5.59164941e-01
9.61990416e-01 -1.12024307e+00 4.56157863e-01 -6.89897656e-01
8.12541544e-01 -3.27469409e-02 -9.10924971e-02 -1.34792304e+00
-2.44819194e-01 -7.03651905e-01 -4.90692966e-02 1.17273736e+00
2.94410616e-01 -3.25517476e-01 9.97334719e-01 4.38474566e-01
1.87202960e-01 -1.42432356e+00 -9.02011514e-01 -1.76843405e-01
-1.41893059e-01 -2.74552912e-01 6.31690323e-01 8.72382462e-01
-3.17007571e-01 5.83549201e-01 -3.85571659e-01 -2.40989998e-01
4.54523951e-01 1.35895878e-01 1.03146732e+00 -8.83076310e-01
-5.75269796e-02 -5.91750145e-01 -4.72625941e-01 -9.95251417e-01
6.65246964e-01 -7.29013383e-01 -2.03364417e-01 -7.62424707e-01
1.73542038e-01 -4.20850128e-01 -2.34867558e-01 8.89541566e-01
-4.81790602e-02 3.05977941e-01 1.87022835e-01 -7.51826689e-02
-6.07862711e-01 1.06610084e+00 1.18661451e+00 -8.66323337e-02
-6.50011301e-02 -2.34424680e-01 -7.70600080e-01 8.29078853e-01
4.51005667e-01 -1.68766290e-01 -3.62349510e-01 -5.39125204e-01
-1.85400382e-01 -2.00439170e-01 3.10125113e-01 -7.51051307e-01
1.38354272e-01 1.08824015e-01 5.73602498e-01 -1.04145613e-03
5.53093553e-01 -8.51929843e-01 -2.42580011e-01 -5.41919321e-02
-2.37260520e-01 -1.25503361e-01 4.60503548e-01 1.48336232e-01
-3.62344086e-01 3.22663665e-01 1.15969861e+00 2.22194105e-01
-7.97594905e-01 9.44985449e-01 -1.20537914e-03 -1.92176420e-02
8.34148943e-01 -6.33513480e-02 1.72344334e-02 -6.21591330e-01
-6.09062314e-01 3.21858883e-01 3.52734685e-01 5.83079636e-01
6.19848430e-01 -1.39357507e+00 -7.91968107e-01 4.99789387e-01
1.62812799e-01 -1.55343711e-01 1.33439451e-02 8.09256017e-01
-4.99338746e-01 -1.33196563e-01 -3.95082414e-01 -7.50033498e-01
-1.47896981e+00 -8.65028277e-02 7.86573172e-01 -1.14081308e-01
-2.14694366e-01 1.22537804e+00 4.35556144e-01 -2.51154453e-01
4.84180540e-01 1.84118107e-01 4.78200391e-02 1.38006508e-01
7.32316852e-01 -1.46824792e-01 1.70051195e-02 -9.68677342e-01
-4.19751137e-01 7.07049429e-01 -2.11464912e-01 -1.18695535e-01
1.48036802e+00 2.68945992e-01 -3.69900823e-01 1.40574083e-01
1.67803407e+00 6.67468412e-03 -1.44377947e+00 -3.10742289e-01
-1.37910485e-01 -4.09710318e-01 1.08465940e-01 -7.51213610e-01
-1.62556362e+00 1.01400161e+00 4.99005169e-01 -7.14117110e-01
1.56282020e+00 -5.33144996e-02 7.17626214e-01 8.92884284e-02
3.72074127e-01 -9.21246290e-01 4.95204419e-01 5.43956816e-01
1.00837076e+00 -1.01445603e+00 -2.30665326e-01 -3.93488050e-01
-7.26624608e-01 1.04183984e+00 8.96558106e-01 4.70904000e-02
9.17401373e-01 5.09502888e-01 3.47162098e-01 -5.39903224e-01
-1.02201366e+00 -1.03003606e-02 1.81952968e-01 2.02878967e-01
5.28119087e-01 -2.40263626e-01 2.24396437e-01 5.54251671e-01
-2.97197282e-01 2.46104449e-01 -3.08497667e-01 7.59325802e-01
-1.61163375e-01 -9.94879901e-01 1.78787531e-03 2.66084969e-01
-6.12307847e-01 1.05953902e-01 -8.58782291e-01 6.92251086e-01
1.42342895e-01 5.31633794e-01 1.86986133e-01 -4.43573833e-01
4.03919160e-01 1.56117976e-01 7.35473931e-01 -3.68246049e-01
-4.86015022e-01 1.65713698e-01 1.19716920e-01 -1.03739190e+00
-3.21174443e-01 -5.58394969e-01 -1.21120656e+00 -3.27328235e-01
7.43603483e-02 6.18874058e-02 4.56649363e-01 7.59659052e-01
6.10580683e-01 2.07349017e-01 8.81056070e-01 -7.93292522e-01
-2.46937156e-01 -9.38641071e-01 -5.66612542e-01 6.39943719e-01
2.91477263e-01 -9.06441331e-01 -2.76699126e-01 -1.14194214e-01] | [13.595128059387207, 1.6415380239486694] |
4879973c-eceb-4556-9b07-1f09eaa8c320 | lexicon-learning-for-few-shot-sequence | null | null | https://aclanthology.org/2021.acl-long.382 | https://aclanthology.org/2021.acl-long.382.pdf | Lexicon Learning for Few Shot Sequence Modeling | Sequence-to-sequence transduction is the core problem in language processing applications as diverse as semantic parsing, machine translation, and instruction following. The neural network models that provide the dominant solution to these problems are brittle, especially in low-resource settings: they fail to generalize correctly or systematically from small datasets. Past work has shown that many failures of systematic generalization arise from neural models{'} inability to disentangle lexical phenomena from syntactic ones. To address this, we augment neural decoders with a lexical translation mechanism that generalizes existing copy mechanisms to incorporate learned, decontextualized, token-level translation rules. We describe how to initialize this mechanism using a variety of lexicon learning algorithms, and show that it improves systematic generalization on a diverse set of sequence modeling tasks drawn from cognitive science, formal semantics, and machine translation. | ['Jacob Andreas', 'Ekin Akyurek'] | 2021-08-01 | null | null | null | acl-2021-5 | ['systematic-generalization'] | ['reasoning'] | [ 5.48404455e-01 2.15007231e-01 -5.31161666e-01 -4.67637420e-01
-7.70038426e-01 -7.59031892e-01 6.09811604e-01 1.38690680e-01
-5.30582786e-01 9.78843689e-01 4.47344363e-01 -1.15881324e+00
2.47538134e-01 -7.72861838e-01 -1.05554175e+00 -6.28316030e-02
3.62614036e-01 6.55075908e-01 1.51262641e-01 -5.32325625e-01
2.17963174e-01 1.16256937e-01 -1.16812229e+00 4.03872162e-01
8.47956061e-01 2.16314539e-01 2.94403642e-01 3.71576548e-01
-4.95586485e-01 7.32686341e-01 -3.33738089e-01 -3.52575690e-01
-8.62341598e-02 -6.96192443e-01 -1.03540516e+00 -4.38252509e-01
2.74387777e-01 -2.02261761e-01 -1.32737949e-01 1.06969810e+00
1.83852404e-01 1.56025519e-03 6.39909625e-01 -7.87868321e-01
-1.05608571e+00 1.05386031e+00 -2.06228569e-01 4.80504096e-01
2.98906058e-01 2.21811473e-01 1.11395478e+00 -7.52362967e-01
8.07718396e-01 1.56532490e+00 7.14182556e-01 1.06054306e+00
-1.50435662e+00 -6.12531185e-01 1.88477159e-01 -5.52888699e-02
-8.57380927e-01 -5.22944033e-01 4.33308780e-01 -3.54918838e-01
1.62317371e+00 -1.65884838e-01 2.69472539e-01 1.43947852e+00
3.24730873e-01 7.84624636e-01 1.11906278e+00 -7.32136726e-01
3.72922570e-02 -1.59449071e-01 3.65289629e-01 7.88980782e-01
6.07663393e-01 1.19959891e-01 -6.55402958e-01 -1.19007453e-01
6.19398117e-01 -4.52119380e-01 -6.99149743e-02 2.54089888e-02
-1.02017629e+00 9.66013610e-01 -1.10121951e-01 4.14574534e-01
-4.32553478e-02 3.23224396e-01 5.07263899e-01 6.76882684e-01
3.96525323e-01 8.62287641e-01 -9.82863665e-01 -2.03410551e-01
-4.91056204e-01 1.90170959e-01 1.01976287e+00 9.42294359e-01
7.52863586e-01 2.70992517e-01 3.31355244e-01 7.50457287e-01
2.09955037e-01 4.78123754e-01 9.64101493e-01 -8.45406830e-01
5.13642013e-01 4.13562953e-01 -2.65304714e-01 -5.65002084e-01
-4.29443091e-01 -2.00506598e-01 -1.23745836e-01 -3.42850745e-01
4.69667107e-01 -2.75666088e-01 -9.59399521e-01 2.46018839e+00
-1.42490014e-01 -1.42590314e-01 1.81363374e-01 5.38325965e-01
7.33067811e-01 5.58625579e-01 7.01904953e-01 -7.71627054e-02
1.28411424e+00 -5.63889444e-01 -4.24571455e-01 -1.02378345e+00
1.09071863e+00 -5.88576496e-01 1.45860112e+00 1.09560117e-01
-1.31853235e+00 -3.56370300e-01 -1.05201900e+00 -2.86326289e-01
-4.26162541e-01 -3.16396952e-01 8.17189872e-01 5.71607113e-01
-1.23271656e+00 5.37603199e-01 -8.80823433e-01 -7.15203464e-01
2.37600669e-01 3.09920639e-01 -2.90958136e-01 -7.07040876e-02
-1.38708997e+00 1.27801800e+00 6.56499684e-01 -3.39852303e-01
-7.32481658e-01 -4.79544967e-01 -9.85955477e-01 7.95078129e-02
3.88778597e-01 -1.08540833e+00 1.66305828e+00 -1.31455326e+00
-1.47137117e+00 1.08615780e+00 -3.83623898e-01 -4.79409277e-01
-1.47164181e-01 -2.60331541e-01 -1.32598370e-01 -2.47446969e-01
1.64429486e-01 7.26809621e-01 3.75189453e-01 -9.37193453e-01
-3.15156043e-01 -4.96830434e-01 -1.72985911e-01 2.88055241e-01
-1.85443476e-01 1.20347366e-01 9.27360952e-02 -4.73071367e-01
-9.93283029e-05 -7.74421394e-01 -4.47187498e-02 -7.59664714e-01
-1.52707309e-01 -2.42366061e-01 1.74845055e-01 -6.38905644e-01
1.07016575e+00 -1.79162669e+00 3.09267551e-01 -2.23699391e-01
-3.97175923e-02 2.51681000e-01 -3.93800378e-01 4.69357342e-01
-7.90683627e-02 4.39124197e-01 -2.03755513e-01 -1.52755097e-01
1.07416175e-01 5.43619692e-01 -5.18561482e-01 -4.04051244e-02
4.01394784e-01 1.31102049e+00 -9.38569307e-01 -1.78345397e-01
-2.49810770e-01 4.04855888e-03 -6.79390967e-01 -1.18783928e-01
-7.71553397e-01 1.89358786e-01 -4.51023251e-01 3.63683462e-01
-8.55183750e-02 -3.93642157e-01 6.43420875e-01 4.85731393e-01
1.11289218e-01 1.15705740e+00 -4.24062908e-01 2.09385514e+00
-4.08423007e-01 5.92648029e-01 -1.48378029e-01 -1.04047418e+00
6.40535891e-01 2.78412253e-01 -6.83344901e-02 -7.13334441e-01
2.19777659e-01 4.55369532e-01 3.97769153e-01 -5.43176830e-01
4.86461401e-01 -6.77300036e-01 -2.21816510e-01 8.84681165e-01
2.77636528e-01 -1.07209682e-02 5.34580722e-02 1.68517202e-01
1.18592024e+00 4.38379288e-01 3.84251326e-01 -3.80629748e-01
1.10861890e-01 5.33171713e-01 6.51673913e-01 8.34513187e-01
5.39478473e-02 2.55241394e-01 4.28866565e-01 -4.36166406e-01
-1.17841887e+00 -1.06241357e+00 1.36404246e-01 1.72498906e+00
-1.48336291e-01 -2.51725107e-01 -9.95204389e-01 -4.34575945e-01
-1.69198379e-01 1.08128905e+00 -2.74499714e-01 -5.26122749e-01
-9.28672075e-01 -9.24041629e-01 9.57398176e-01 6.52340114e-01
8.24248698e-03 -1.32566559e+00 -4.67866719e-01 5.58780789e-01
-2.52849251e-01 -1.18277323e+00 -1.89231962e-01 4.47599947e-01
-1.04842269e+00 -8.75800431e-01 4.42074463e-02 -1.05851877e+00
3.56753200e-01 -1.74321476e-02 1.45379925e+00 3.45599800e-01
-5.09951636e-03 1.56983450e-01 -4.91580516e-02 -4.98150498e-01
-1.18056238e+00 5.82007170e-01 2.10773036e-01 -7.81696856e-01
1.05033910e+00 -4.76272911e-01 8.00013915e-02 -3.52620035e-02
-1.08495295e+00 5.57827167e-02 6.04600787e-01 1.02322614e+00
1.78897277e-01 -5.92972159e-01 1.05361080e+00 -1.12987769e+00
1.17913187e+00 -6.85793817e-01 -4.12399113e-01 2.87925780e-01
-6.96195722e-01 5.68321586e-01 7.49022782e-01 -4.33906734e-01
-9.17753339e-01 -2.37334639e-01 -1.60692990e-01 7.51413032e-02
-4.56029534e-01 6.64032698e-01 -1.96364224e-01 2.21703216e-01
8.98646832e-01 4.52669948e-01 1.86107427e-01 -4.46404219e-01
6.50672019e-01 3.21921587e-01 4.72708106e-01 -1.18090653e+00
5.09023845e-01 -6.95627183e-02 -1.61200404e-01 -7.43641496e-01
-7.15385079e-01 -1.31549807e-02 -4.55909222e-01 5.49466670e-01
9.35148537e-01 -7.39647746e-01 -5.09575784e-01 1.31790370e-01
-1.32885635e+00 -7.63895392e-01 -1.68637067e-01 4.82938826e-01
-7.66324997e-01 2.40820408e-01 -1.14697528e+00 -1.74547940e-01
-2.67840356e-01 -1.09355640e+00 8.23592961e-01 9.17388573e-02
-6.12383127e-01 -1.28843331e+00 1.57979846e-01 8.36162940e-02
5.89986265e-01 -3.47001970e-01 1.79485035e+00 -1.13144243e+00
-4.79842007e-01 3.50259960e-01 2.47057397e-02 2.09732085e-01
5.25605902e-02 -3.00889075e-01 -7.62441754e-01 -9.66290664e-03
3.54356803e-02 -6.26496792e-01 8.88797402e-01 1.49756178e-01
7.15272784e-01 -2.24217013e-01 -3.24465960e-01 4.41344321e-01
1.20005774e+00 2.27285460e-01 2.63846993e-01 3.86711180e-01
4.52697128e-01 5.55655062e-01 -1.83490574e-01 -4.92780745e-01
6.06679797e-01 3.33113313e-01 -8.80514458e-02 1.84125215e-01
-2.51146674e-01 -6.45421147e-01 5.75971723e-01 1.07640016e+00
2.32936218e-01 -2.16044426e-01 -1.11012423e+00 4.67833847e-01
-1.80291152e+00 -8.08453262e-01 2.65700966e-01 1.93575358e+00
1.29405022e+00 3.54768604e-01 -1.52690053e-01 -4.00258392e-01
4.26175714e-01 -3.44959274e-02 -5.75432777e-01 -9.19880092e-01
-1.88681990e-01 5.08866727e-01 4.48249817e-01 7.74448693e-01
-5.63978672e-01 1.71578801e+00 7.43192196e+00 5.33177793e-01
-1.05571306e+00 2.55308867e-01 4.21376377e-01 2.06809491e-01
-7.65795588e-01 1.61977671e-02 -8.51241529e-01 1.61751360e-02
1.40401089e+00 -1.54190883e-01 7.32339978e-01 4.99957502e-01
-6.46528080e-02 1.79162905e-01 -1.53995025e+00 4.17404801e-01
2.07644865e-01 -1.36895561e+00 3.90492678e-01 -2.24572390e-01
5.59666932e-01 3.50517571e-01 1.07616931e-02 5.19265115e-01
8.43178689e-01 -1.29780614e+00 5.48817515e-01 2.81387474e-02
6.88849270e-01 -3.95894408e-01 1.99438274e-01 5.81942976e-01
-5.74179053e-01 -8.64376605e-04 -3.47339928e-01 -4.40032333e-01
1.60769016e-01 -7.53536969e-02 -1.01188445e+00 3.17611508e-02
-1.84273105e-02 3.96345675e-01 -4.42263842e-01 3.69177997e-01
-5.55914998e-01 8.85426223e-01 -1.86081707e-01 -3.60252142e-01
3.44356477e-01 -2.04932410e-03 5.08528709e-01 1.34304023e+00
1.26163289e-01 1.47082910e-01 -1.10950237e-02 9.63583171e-01
-3.15057606e-01 1.23727672e-01 -9.90453064e-01 -2.56944031e-01
5.54884076e-01 5.26535392e-01 -7.21989512e-01 -4.05257851e-01
-6.66130722e-01 7.43139625e-01 5.46445370e-01 6.57938182e-01
-3.84358257e-01 -6.07527932e-03 8.50574374e-01 -2.38751061e-02
7.62822330e-02 -5.59500813e-01 -5.95870495e-01 -1.31247723e+00
-2.01613560e-01 -1.15569139e+00 1.94963858e-01 -8.03391755e-01
-1.20962799e+00 5.03239512e-01 1.33275911e-01 -2.23473147e-01
-9.71086383e-01 -1.06015968e+00 -3.15019071e-01 1.09242415e+00
-1.35632253e+00 -9.32968020e-01 5.96176565e-01 3.69324178e-01
8.71881127e-01 -3.04692864e-01 1.01871371e+00 4.69960235e-02
-4.08750802e-01 5.46053410e-01 -1.43509895e-01 2.02985734e-01
4.88303512e-01 -1.08082795e+00 9.68851089e-01 9.94034171e-01
3.20255518e-01 1.12756145e+00 7.29434967e-01 -6.58223748e-01
-1.57696629e+00 -8.49598825e-01 1.20295072e+00 -9.17175829e-01
8.42584133e-01 -4.87110376e-01 -1.10902917e+00 1.24586773e+00
3.09468955e-01 -4.61775780e-01 7.27153897e-01 3.38538557e-01
-6.86138332e-01 5.02702534e-01 -6.05795026e-01 7.95098960e-01
1.26135004e+00 -9.57673192e-01 -1.22416759e+00 1.36882991e-01
1.05131769e+00 -3.07496130e-01 -4.32597429e-01 1.49323627e-01
4.48833644e-01 -4.44624662e-01 6.50441647e-01 -1.43734837e+00
6.63395703e-01 1.49069607e-01 -3.22441101e-01 -1.49184060e+00
-1.75388202e-01 -6.26933396e-01 1.80187747e-01 7.24596024e-01
9.21136200e-01 -8.21935534e-01 6.63155973e-01 5.50310850e-01
-4.31111097e-01 -5.72211683e-01 -7.86224365e-01 -7.14002609e-01
8.56177568e-01 -6.25520408e-01 7.02430844e-01 1.04966199e+00
2.73932874e-01 1.07033110e+00 1.49846405e-01 -2.90440679e-01
3.48339856e-01 6.53967187e-02 4.74279404e-01 -1.22364986e+00
-4.87642914e-01 -8.23695600e-01 -1.15094595e-01 -1.28695273e+00
6.96005762e-01 -1.33910584e+00 2.73934603e-01 -1.43658388e+00
3.89935732e-01 -3.64596993e-01 -9.72283259e-02 7.12072015e-01
-2.14506403e-01 -1.17966853e-01 -1.13015734e-01 9.19498354e-02
-3.45263541e-01 2.25621626e-01 9.84206319e-01 7.73208365e-02
-3.25022936e-02 -4.84589070e-01 -1.10262406e+00 8.34582746e-01
9.20403898e-01 -6.43818915e-01 -5.38281739e-01 -1.13253844e+00
6.50083244e-01 1.09264836e-01 2.34110169e-02 -3.83359998e-01
3.83380055e-03 -3.82789195e-01 5.65533638e-02 1.56808823e-01
2.08117701e-02 -2.70669580e-01 -2.79812366e-01 4.97556627e-01
-7.64488101e-01 4.86163408e-01 5.54979026e-01 3.36456537e-01
9.63019505e-02 -2.68574089e-01 4.32737619e-01 -5.61724246e-01
-9.08846617e-01 -6.85237795e-02 -6.75070167e-01 5.60471654e-01
4.18268055e-01 -5.43068424e-02 -6.89019859e-01 -1.69901848e-01
-6.66062474e-01 -5.43030575e-02 7.00941801e-01 7.09308088e-01
3.71450514e-01 -1.07811224e+00 -5.73110521e-01 2.98847437e-01
-6.59964904e-02 -1.49921089e-01 -4.40731585e-01 5.41496217e-01
-4.30164009e-01 8.36834133e-01 -2.32323200e-01 -2.94853806e-01
-7.16538131e-01 5.72869062e-01 4.25381243e-01 -1.12285160e-01
-3.97276640e-01 7.20054150e-01 4.17148829e-01 -7.06913650e-01
-6.72440305e-02 -7.48865426e-01 1.46606967e-01 -4.92398769e-01
2.90335268e-01 -1.03012025e-01 1.47787675e-01 -6.06100559e-01
-2.56600231e-01 3.49289685e-01 -9.03306678e-02 -3.23173463e-01
1.09745657e+00 -1.71274602e-01 -7.80917704e-02 6.36601329e-01
9.60106552e-01 -2.12681428e-01 -8.68554890e-01 -5.38168252e-01
3.97834837e-01 2.76977330e-01 -4.13502634e-01 -8.26418996e-01
-3.22026938e-01 1.10519660e+00 -5.65796345e-02 6.45518452e-02
6.65383816e-01 8.74693096e-02 9.96358156e-01 7.33590841e-01
4.65905339e-01 -9.73074973e-01 -1.15589008e-01 1.12766707e+00
2.34309644e-01 -1.09653711e+00 -4.73851621e-01 -4.07173187e-02
-2.91017890e-01 1.03166914e+00 6.58574700e-01 1.21433757e-01
3.24732363e-01 3.81074548e-01 4.19004075e-02 -1.69641733e-01
-1.16884613e+00 -1.07735716e-01 -4.40279953e-02 4.09318745e-01
8.23047996e-01 1.96732163e-01 -5.00322461e-01 6.83564067e-01
-2.96527714e-01 1.08554900e-01 5.71843088e-01 9.49249804e-01
-8.02480400e-01 -1.33859885e+00 -1.26152635e-01 2.39172727e-01
-6.83725893e-01 -6.35965943e-01 -5.07795751e-01 7.90508270e-01
-1.95271805e-01 7.42580414e-01 7.64581859e-02 -2.87239254e-01
9.74695571e-03 8.34186435e-01 7.01116025e-01 -1.08140969e+00
-4.99020338e-01 -1.21501245e-01 3.62183332e-01 -4.67230022e-01
-1.72570437e-01 -6.17374420e-01 -1.54620922e+00 -2.83086181e-01
2.21610982e-02 1.55666664e-01 5.90410471e-01 1.34330344e+00
3.49636316e-01 3.54014277e-01 -9.54468027e-02 -2.52289683e-01
-8.10121894e-01 -9.23154712e-01 -5.66647435e-03 4.70207393e-01
1.96964666e-01 -3.20287257e-01 -1.36502907e-01 1.28767684e-01] | [10.691649436950684, 9.120824813842773] |
34b3854c-9fb8-4328-9004-cc166e891efa | video-coding-for-machine-compact-visual | 2110.09241 | null | https://arxiv.org/abs/2110.09241v1 | https://arxiv.org/pdf/2110.09241v1.pdf | Video Coding for Machine: Compact Visual Representation Compression for Intelligent Collaborative Analytics | Video Coding for Machines (VCM) is committed to bridging to an extent separate research tracks of video/image compression and feature compression, and attempts to optimize compactness and efficiency jointly from a unified perspective of high accuracy machine vision and full fidelity human vision. In this paper, we summarize VCM methodology and philosophy based on existing academia and industrial efforts. The development of VCM follows a general rate-distortion optimization, and the categorization of key modules or techniques is established. From previous works, it is demonstrated that, although existing works attempt to reveal the nature of scalable representation in bits when dealing with machine and human vision tasks, there remains a rare study in the generality of low bit rate representation, and accordingly how to support a variety of visual analytic tasks. Therefore, we investigate a novel visual information compression for the analytics taxonomy problem to strengthen the capability of compact visual representations extracted from multiple tasks for visual analytics. A new perspective of task relationships versus compression is revisited. By keeping in mind the transferability among different machine vision tasks (e.g. high-level semantic and mid-level geometry-related), we aim to support multiple tasks jointly at low bit rates. In particular, to narrow the dimensionality gap between neural network generated features extracted from pixels and a variety of machine vision features/labels (e.g. scene class, segmentation labels), a codebook hyperprior is designed to compress the neural network-generated features. As demonstrated in our experiments, this new hyperprior model is expected to improve feature compression efficiency by estimating the signal entropy more accurately, which enables further investigation of the granularity of abstracting compact features among different tasks. | ['Jiaying Liu', 'Ling-Yu Duan', 'Yueyu Hu', 'Haofeng Huang', 'Wenhan Yang'] | 2021-10-18 | null | null | null | null | ['feature-compression'] | ['computer-vision'] | [ 7.3097157e-01 1.2976253e-01 -2.5129807e-01 -1.7295535e-01
-4.5177025e-01 -1.6750449e-01 7.1152371e-01 1.3926560e-01
-3.6451322e-01 2.2155853e-01 1.7127620e-01 -7.6364294e-02
-4.5087793e-01 -5.8708978e-01 -5.1690286e-01 -5.0420529e-01
-2.0089133e-01 2.4457317e-02 -1.1345281e-01 1.0241328e-01
5.5445617e-01 8.3488470e-01 -2.1625957e+00 4.0010306e-01
6.4348072e-01 1.4751990e+00 5.7301152e-01 7.1299034e-01
9.3352780e-02 8.6575484e-01 -3.7012386e-01 -5.1894462e-01
3.5268191e-01 -2.0831282e-01 -8.9858109e-01 4.2130491e-01
6.4185107e-01 -3.2733721e-01 -4.2188534e-01 1.0380816e+00
1.4728235e-01 5.8655702e-02 9.1921675e-01 -1.5339712e+00
-5.7440674e-01 2.8850719e-01 -3.9845127e-01 3.2991612e-01
1.2927912e-01 2.7912024e-01 1.0062479e+00 -9.0024650e-01
5.8435780e-01 1.2015123e+00 5.6240195e-01 2.2740053e-01
-9.9440080e-01 -1.8194424e-01 -4.4306248e-02 5.0733244e-01
-1.4654329e+00 -5.2636564e-01 7.2938222e-01 -6.1774826e-01
1.1380955e+00 3.6637706e-01 6.9133270e-01 7.3696667e-01
2.3285125e-01 5.8759934e-01 7.8877717e-01 -5.9718448e-01
2.9011002e-01 1.7531732e-01 -3.6847319e-02 7.7178520e-01
4.5150638e-01 4.2013645e-02 -7.1291763e-01 1.5680653e-01
7.8311259e-01 -1.2649056e-01 -3.1595108e-01 -5.9764898e-01
-1.1351399e+00 1.0145856e+00 2.2529273e-01 3.8555747e-01
-4.0683511e-01 1.8518047e-01 5.7229948e-01 2.1182434e-01
1.8242413e-01 4.3822703e-01 -2.9066211e-01 -2.9959854e-01
-1.3895600e+00 5.9339702e-03 6.6557342e-01 1.1378725e+00
6.6968691e-01 3.3841935e-01 -2.1228707e-01 7.0457208e-01
3.7619820e-01 2.9126087e-01 4.8831016e-01 -1.2857909e+00
5.0329423e-01 4.8705563e-01 -3.2596844e-01 -1.4088469e+00
-3.4372309e-01 -3.2932770e-01 -9.8969805e-01 2.1875125e-01
1.3705182e-01 2.9265076e-01 -7.0356590e-01 1.4696990e+00
-4.7640063e-02 -1.6278552e-01 7.2527699e-02 9.2851686e-01
6.9983488e-01 5.3882504e-01 -2.4349961e-02 -3.9795098e-01
1.4857652e+00 -8.0587238e-01 -5.3672570e-01 -9.2663668e-02
5.9340751e-01 -6.3563174e-01 9.9006861e-01 5.4019314e-01
-1.3410834e+00 -9.4231212e-01 -1.2936913e+00 -4.2016304e-01
-4.0402964e-01 1.5183799e-01 6.5515608e-01 6.8956757e-01
-1.0703408e+00 6.4350933e-01 -7.4785548e-01 -2.6946130e-01
7.0170528e-01 2.1120894e-01 -2.3172621e-01 -1.8437660e-01
-9.1223663e-01 9.0740818e-01 4.6986765e-01 -1.4362495e-01
-5.8376539e-01 -7.5999993e-01 -9.9092156e-01 2.9745686e-01
6.0760364e-02 -8.4938473e-01 8.8040215e-01 -1.0430517e+00
-1.0754689e+00 9.1670865e-01 -3.2727964e-02 -8.7510806e-01
3.4621966e-01 3.5691462e-02 -2.1880352e-01 6.5124607e-01
-1.7159531e-01 1.0486780e+00 1.2469705e+00 -9.7232318e-01
-8.0296534e-01 -3.3964860e-01 -5.1474642e-02 2.6318067e-01
-6.2496310e-01 -1.6514325e-01 -4.4897553e-01 -7.0993751e-01
-5.3341880e-02 -7.3919809e-01 -5.4567240e-02 2.6830119e-01
6.8714499e-02 5.5027720e-02 8.7476003e-01 -5.5398846e-01
1.2541322e+00 -2.3363845e+00 3.8725039e-01 1.3571480e-01
4.6159339e-01 1.9599782e-01 -8.1705235e-02 2.6651821e-01
1.1199388e-01 5.8406234e-02 -2.2025207e-01 -3.7820774e-01
-5.9229564e-02 1.5710063e-01 -2.6329601e-01 5.3924030e-01
1.9119008e-01 9.2404830e-01 -5.2592957e-01 -8.1290275e-01
5.4240477e-01 5.8107382e-01 -6.9027311e-01 -7.4344195e-02
6.0467362e-02 -4.3857493e-02 -3.7203214e-01 8.2287812e-01
7.0931321e-01 -4.8741081e-01 -1.8380273e-02 -5.9910655e-01
-4.9071617e-02 -2.2796910e-01 -1.1075159e+00 1.7502745e+00
-4.2606348e-01 1.0583223e+00 1.4295197e-01 -1.4059392e+00
7.9446548e-01 1.9170937e-01 7.7379203e-01 -6.7094427e-01
2.6922747e-01 1.1415008e-01 -3.1049248e-02 -4.7793195e-01
9.6497846e-01 1.3368474e-01 1.7817949e-01 -7.3341973e-02
2.8719077e-01 -3.5144439e-01 2.8291473e-01 6.3370064e-02
8.2857722e-01 -8.1927143e-02 6.8319166e-01 -1.1822354e-01
4.7561154e-01 1.9905382e-01 7.0742577e-02 6.8194556e-01
-3.5626635e-01 6.8014073e-01 1.1946411e-01 -2.8733993e-01
-1.3585658e+00 -7.6360476e-01 -5.2936774e-01 1.0046822e+00
1.4335832e-01 -6.6457945e-01 -7.9406404e-01 -1.8878405e-01
3.3116333e-02 4.5891166e-01 -3.6869687e-01 -2.0486104e-01
-3.2751569e-01 -5.2275711e-01 4.8229763e-01 4.0315196e-01
4.7888127e-01 -1.0144757e+00 -1.4371872e+00 4.8364303e-03
-1.1564261e-01 -1.3985184e+00 -7.9530701e-02 1.8490396e-01
-1.0627946e+00 -1.0188676e+00 -5.5383307e-01 -6.2919515e-01
3.1955999e-01 5.3812212e-01 1.1249666e+00 1.0509024e-01
-6.2121695e-01 8.8021064e-01 -5.6402141e-01 -3.5653022e-01
-4.3097740e-01 -1.3869186e-02 2.7353358e-03 -2.1855293e-02
2.8205055e-01 -5.9791785e-01 -6.0989320e-01 6.5687925e-02
-1.1892359e+00 2.9812154e-01 7.5755447e-01 5.9997910e-01
6.5166688e-01 5.8447447e-02 1.0244595e-01 -2.4892196e-01
3.7941664e-01 -4.2251387e-01 -4.6772459e-01 1.8201017e-01
-7.3438728e-01 1.6509832e-01 6.0573500e-01 -3.3865947e-01
-6.6299772e-01 4.4058621e-02 1.1344666e-01 -8.4261280e-01
-2.4043395e-01 5.1242161e-01 -8.4227845e-02 -2.7048498e-01
5.6239963e-01 5.4477149e-01 2.1633726e-01 -1.9751993e-01
6.6913325e-01 6.9165039e-01 5.2002716e-01 -4.2687461e-01
6.0724360e-01 5.4314739e-01 4.0818009e-01 -1.1665564e+00
-3.4119579e-01 -4.3360460e-01 -7.8791898e-01 -3.3945304e-01
9.7729862e-01 -8.4093505e-01 -5.1814407e-01 1.5794756e-01
-1.0910580e+00 1.6650511e-01 -5.6075644e-01 5.7286006e-01
-1.1580827e+00 6.2181330e-01 -4.6363246e-01 -6.2857425e-01
-2.7129751e-01 -1.3878344e+00 1.1581138e+00 -1.3609547e-02
-2.3372555e-01 -8.0228496e-01 -4.8830366e-01 4.4270140e-01
4.2046645e-01 2.2243972e-01 8.6250943e-01 -3.4559453e-01
-7.6743501e-01 1.8492162e-02 -4.2256773e-01 5.4503655e-01
-1.8050753e-01 -7.2238125e-02 -1.0352848e+00 -4.0947217e-01
2.6737407e-01 -4.1613594e-01 9.3260163e-01 6.4057648e-01
1.4317496e+00 -3.0444801e-01 -1.5870963e-01 9.8820072e-01
1.5511178e+00 1.4232819e-01 7.7725679e-01 4.0530038e-01
5.9505796e-01 7.9564881e-01 7.2575468e-01 6.6784352e-01
8.0394767e-02 6.7126918e-01 6.8226403e-01 3.0487742e-02
-3.8438088e-01 -3.0550698e-02 2.7263439e-01 8.3530098e-01
-3.4304252e-01 -4.5244720e-02 -6.1426330e-01 3.0626428e-01
-1.5334811e+00 -9.7379011e-01 1.4274438e-01 2.1135621e+00
4.0449074e-01 2.9644707e-02 -6.0100555e-02 5.3787547e-01
4.9655783e-01 3.4160468e-01 -4.6519020e-01 -6.5562439e-01
-9.5880836e-02 1.7089557e-02 6.3023013e-01 1.7971966e-01
-1.1541561e+00 5.3929615e-01 6.1303167e+00 1.2399993e+00
-1.1245824e+00 -3.0225556e-02 6.8286240e-01 -2.9868290e-01
-9.4692506e-02 -6.2534422e-02 -6.6872174e-01 3.4514016e-01
1.0349721e+00 -2.5245178e-01 5.4362333e-01 1.0645930e+00
3.6822595e-02 4.9536567e-02 -1.3015777e+00 1.5547328e+00
3.2811043e-01 -1.5749632e+00 5.3319782e-01 4.1070271e-01
4.8206842e-01 -9.2073582e-02 2.8612497e-01 7.6758966e-02
-3.7407064e-01 -1.2218949e+00 1.0395678e+00 5.7164872e-01
1.0680838e+00 -5.9419459e-01 4.2265639e-01 2.4643461e-01
-1.4275188e+00 -4.3339679e-01 -5.0919592e-01 -5.1791470e-02
1.4203741e-01 4.6095261e-01 -5.0545210e-01 7.3706126e-01
7.4667525e-01 8.0739510e-01 -6.2273949e-01 7.9588991e-01
5.6624007e-01 1.5699445e-01 -3.5198078e-02 1.7871620e-01
2.5979632e-01 -1.6926502e-01 5.0291920e-01 1.4117978e+00
4.8570070e-01 -5.7076082e-02 1.3183672e-02 8.5770333e-01
1.1736991e-01 1.7840099e-01 -8.5212183e-01 -8.8490449e-02
4.6246696e-01 1.2648063e+00 -7.8685361e-01 -2.1335247e-01
-6.6389835e-01 9.9093115e-01 1.4598022e-01 2.3665182e-01
-8.5475582e-01 -1.5808009e-01 5.0302154e-01 2.3993912e-01
5.1058847e-01 -3.3790311e-01 -5.2399677e-01 -9.9682230e-01
8.2667328e-02 -8.2742614e-01 2.6831499e-01 -8.2621986e-01
-8.5938221e-01 4.4674182e-01 2.8810784e-01 -1.5587572e+00
-3.9679453e-01 -7.0591426e-01 -9.0957314e-02 5.0767291e-01
-1.4741025e+00 -1.0965463e+00 -3.9320898e-01 5.5405277e-01
8.3880818e-01 -4.0361214e-01 5.2942562e-01 2.5457320e-01
-2.4033056e-01 5.9500903e-01 5.3555705e-02 -1.9421025e-01
1.4547184e-01 -7.3939461e-01 7.0671096e-02 8.1423652e-01
3.5842961e-01 3.4148189e-01 5.4183489e-01 -2.2862178e-01
-1.6945205e+00 -1.1032029e+00 5.5443722e-01 -2.1218118e-01
3.3850282e-01 -4.4839121e-02 -7.3888934e-01 3.3943599e-01
-4.8180282e-02 4.0095713e-02 5.0797683e-01 -3.9960515e-01
-2.3928767e-01 7.5643383e-02 -1.1457050e+00 3.6815709e-01
1.1256360e+00 -6.6807145e-01 -4.4091138e-01 5.0199520e-02
6.5152627e-01 -9.3842499e-02 -1.0712008e+00 4.4077879e-01
4.3828413e-01 -1.1837995e+00 1.2831988e+00 -3.5982218e-01
7.1959770e-01 -9.0403073e-02 -8.2503009e-01 -7.8937322e-01
-3.4196284e-01 -3.0905429e-01 -5.9412950e-01 9.8633242e-01
-1.3864829e-01 -1.5437542e-01 6.1374003e-01 3.4270853e-01
-3.2583195e-01 -9.5780170e-01 -1.1685985e+00 -7.7715367e-01
-1.9612189e-01 -7.1777368e-01 2.8517357e-01 7.5090581e-01
-1.0743393e-01 1.0325958e-01 -3.9659175e-01 -6.0070787e-02
6.0600853e-01 3.9147515e-02 5.4551274e-01 -1.3449391e+00
-2.1950199e-01 -9.3514681e-01 -9.4766766e-01 -1.1442744e+00
-7.2547629e-02 -1.0719208e+00 -3.2736295e-01 -1.2169434e+00
3.0816367e-01 -1.6056323e-01 -8.3941117e-02 -1.4932360e-02
3.9855632e-01 3.0132198e-01 5.7262158e-01 4.1537032e-01
-5.1266390e-01 5.9677696e-01 1.1589997e+00 -1.9838893e-01
5.9634154e-03 -2.7680835e-01 -6.9119120e-01 6.5387577e-01
5.4854810e-01 -1.3216822e-01 -6.5308183e-01 -3.2905850e-01
-4.7110029e-02 1.5784766e-01 6.2337124e-01 -1.2946194e+00
1.9693935e-01 -1.7160070e-01 3.7792000e-01 -5.2397186e-01
5.0220299e-01 -9.2849511e-01 2.0672480e-02 4.7336957e-01
-2.8216067e-01 -6.4439699e-02 5.9517253e-02 6.1878777e-01
-4.1932088e-01 -3.6700237e-01 8.6741346e-01 -7.4369334e-02
-1.0218606e+00 3.4653732e-01 -2.3691483e-01 1.1759829e-02
1.0169141e+00 -9.3531686e-01 -2.4517914e-02 -2.2634932e-01
-5.6270945e-01 -2.9024965e-01 4.7380906e-01 5.2028829e-01
9.3844670e-01 -1.2703655e+00 -6.3657469e-01 3.5595542e-01
1.7263275e-01 -2.5931633e-01 3.3157989e-01 6.9032425e-01
-5.4843462e-01 8.0528355e-01 -4.8333067e-01 -8.7398875e-01
-1.2889274e+00 8.4596992e-01 1.4394993e-01 -6.2147360e-02
-7.5649559e-01 4.9947685e-01 1.6485204e-01 4.6732932e-01
4.6003741e-01 -5.6541663e-01 -3.1196302e-01 1.7895886e-01
5.0441653e-01 6.6831499e-01 9.4540320e-02 -8.8492376e-01
-1.2982252e-01 8.2258415e-01 2.3942854e-01 1.5209244e-01
1.1496860e+00 -2.7826038e-01 7.8758009e-02 2.7478015e-01
1.4809200e+00 -3.4080538e-01 -1.4446118e+00 -3.4338433e-02
-3.0643983e-02 -5.3819245e-01 2.6225498e-01 -8.5236214e-02
-1.1718913e+00 9.0820682e-01 8.5053790e-01 2.7495545e-01
1.4515589e+00 7.3460884e-02 4.8743126e-01 1.9994801e-01
4.2476314e-01 -1.1603638e+00 1.5807082e-01 2.8227395e-01
1.0849872e+00 -1.0999963e+00 2.5050738e-01 -4.4951436e-01
-5.8283758e-01 1.3008256e+00 2.5752226e-01 1.1333257e-01
6.3773400e-01 2.8586322e-01 -5.7345694e-01 -3.2399753e-01
-4.9903113e-01 -1.7790106e-01 5.4911357e-01 8.0954343e-01
2.1051963e-01 -4.1838493e-02 -2.6709700e-01 3.5612306e-01
-3.7225819e-01 -5.3440582e-02 4.0226546e-01 8.1327081e-01
-6.4276093e-01 -6.6005886e-01 -4.6673882e-01 7.0017397e-01
-3.3623624e-01 -1.4155057e-01 7.2885863e-02 8.0042320e-01
3.1226209e-01 8.7479740e-01 3.8695368e-01 -4.8847058e-01
1.2485704e-01 -1.4533037e-01 5.8342570e-01 -1.8854997e-01
-2.0080850e-01 -2.0379201e-01 -1.8309624e-01 -6.2012410e-01
-6.4636433e-01 -6.6859198e-01 -9.3076116e-01 -3.2873476e-01
-3.9519418e-02 -1.6742654e-01 8.1765234e-01 8.1344175e-01
4.4621521e-01 5.0696689e-01 4.2135686e-01 -1.1973892e+00
-5.5808175e-01 -7.6468945e-01 -6.5904081e-01 5.2812290e-01
2.9086089e-01 -8.5002345e-01 -3.3192152e-01 4.0693089e-01] | [11.267890930175781, -1.5425772666931152] |
d6533c01-a698-490b-bcc0-6ccb6911d367 | nonnegative-opls-for-supervised-design-of | 2112.12280 | null | https://arxiv.org/abs/2112.12280v1 | https://arxiv.org/pdf/2112.12280v1.pdf | Nonnegative OPLS for Supervised Design of Filter Banks: Application to Image and Audio Feature Extraction | Audio or visual data analysis tasks usually have to deal with high-dimensional and nonnegative signals. However, most data analysis methods suffer from overfitting and numerical problems when data have more than a few dimensions needing a dimensionality reduction preprocessing. Moreover, interpretability about how and why filters work for audio or visual applications is a desired property, especially when energy or spectral signals are involved. In these cases, due to the nature of these signals, the nonnegativity of the filter weights is a desired property to better understand its working. Because of these two necessities, we propose different methods to reduce the dimensionality of data while the nonnegativity and interpretability of the solution are assured. In particular, we propose a generalized methodology to design filter banks in a supervised way for applications dealing with nonnegative data, and we explore different ways of solving the proposed objective function consisting of a nonnegative version of the orthonormalized partial least-squares method. We analyze the discriminative power of the features obtained with the proposed methods for two different and widely studied applications: texture and music genre classification. Furthermore, we compare the filter banks achieved by our methods with other state-of-the-art methods specifically designed for feature extraction. | ['Vanessa Gómez-Verdejo', 'Jerónimo Arenas García', 'Sergio Muñoz-Romero'] | 2021-12-22 | null | null | null | null | ['genre-classification'] | ['computer-vision'] | [ 1.88954592e-01 -3.61593187e-01 7.56951123e-02 -1.93973064e-01
-1.29087105e-01 -5.58721364e-01 1.47710800e-01 -9.39613767e-03
-4.04866010e-01 6.03838980e-01 1.28902972e-01 -2.13445015e-02
-7.31180370e-01 -6.10477567e-01 -1.73259959e-01 -9.35973823e-01
1.37941852e-01 1.31648496e-01 -1.97347224e-01 -2.81466782e-01
3.13648880e-01 7.09746301e-01 -2.03992462e+00 1.31314397e-01
8.32112968e-01 1.22214317e+00 8.16751719e-02 4.14049327e-01
-1.31591544e-01 2.19006881e-01 -3.05748880e-01 1.55788697e-02
2.82789171e-01 -3.19081455e-01 -3.69820654e-01 5.68202138e-01
5.74810430e-02 1.02078684e-01 2.13828728e-01 1.19166005e+00
4.41256285e-01 2.91267276e-01 8.64193618e-01 -1.24235010e+00
-4.69228029e-01 1.36143267e-01 -5.75210452e-01 6.66356385e-02
2.88060963e-01 -1.99997857e-01 1.01127160e+00 -1.02521110e+00
2.89381683e-01 1.01342535e+00 4.10003781e-01 2.81020284e-01
-1.30253255e+00 -2.57188916e-01 -6.80013224e-02 3.87524575e-01
-1.30043817e+00 -5.32321215e-01 1.13453460e+00 -6.83981359e-01
3.20997119e-01 7.31747508e-01 7.38278568e-01 7.26449311e-01
-6.76434264e-02 3.00287604e-01 9.18378532e-01 -4.41723317e-01
2.01225087e-01 3.95814925e-01 1.87156200e-01 3.51477265e-01
4.11397755e-01 -3.36902380e-01 -3.68165523e-01 -1.80881068e-01
4.64196712e-01 -1.07125780e-02 -4.54149425e-01 -6.56256437e-01
-1.16433787e+00 7.40958989e-01 -1.83766395e-01 7.32015848e-01
-5.54612994e-01 -3.53198797e-01 3.77967328e-01 2.98019439e-01
3.55704933e-01 4.66280967e-01 -2.14408383e-01 -2.40251217e-02
-7.28836954e-01 -1.90327436e-01 7.05104470e-01 4.56889600e-01
3.93834412e-01 1.48509666e-01 6.69890866e-02 8.27685177e-01
2.49819070e-01 3.84869844e-01 6.91134810e-01 -5.22000313e-01
3.26123774e-01 6.07522368e-01 2.44101346e-01 -1.52711499e+00
-6.69211268e-01 -5.30809999e-01 -1.22449291e+00 1.19928122e-01
5.60590506e-01 1.00715175e-01 -1.40450299e-01 1.43804646e+00
3.26636434e-01 -1.95044175e-01 1.10769980e-01 1.12133610e+00
6.27034903e-01 6.28588021e-01 -2.48804480e-01 -7.31406331e-01
1.39442098e+00 -4.01019901e-01 -9.67830122e-01 1.18457429e-01
1.55476406e-01 -1.01737154e+00 1.29759479e+00 9.65916812e-01
-8.71405602e-01 -5.50683081e-01 -8.48758280e-01 7.21832961e-02
-9.20709521e-02 8.15632582e-01 3.88834119e-01 6.39796495e-01
-5.29769480e-01 5.55489182e-01 -4.70681489e-01 -1.81475773e-01
-2.43523344e-01 2.63025612e-01 -4.12544161e-01 4.43699270e-01
-8.68230343e-01 6.87703669e-01 2.09720343e-01 7.03975499e-01
-8.45006183e-02 -2.51618445e-01 -4.70735610e-01 3.17192644e-01
1.83100775e-01 -2.89354891e-01 5.47193229e-01 -1.15351045e+00
-1.32004499e+00 5.77792227e-01 -1.57933474e-01 5.91564178e-02
5.55996954e-01 -1.23715140e-02 -6.08050406e-01 1.55253127e-01
-2.61559308e-01 -1.47470579e-01 1.20181859e+00 -9.00106013e-01
-2.01555997e-01 -4.41682607e-01 -2.69162685e-01 -2.23135632e-02
-1.15564561e+00 -2.60500789e-01 -8.71242359e-02 -8.37578773e-01
6.60373867e-01 -7.45156825e-01 -5.93736395e-02 2.10189931e-02
-3.38328242e-01 -1.09503224e-01 8.14904451e-01 -6.09196246e-01
1.48256874e+00 -2.61222959e+00 4.07804847e-01 6.14020705e-01
3.48120853e-02 2.64456701e-02 -2.00019717e-01 3.74547720e-01
-4.01139557e-01 -1.38693094e-01 6.47174008e-03 -7.48089142e-03
-1.24495871e-01 -6.47733510e-02 -1.05703592e-01 7.30668306e-01
1.47504896e-01 -1.24066696e-02 -4.87020820e-01 -4.25501347e-01
2.97521740e-01 6.22803152e-01 -5.74920893e-01 9.24171358e-02
1.40641451e-01 5.55803061e-01 -5.17259359e-01 3.07726175e-01
4.88427609e-01 -3.77019607e-02 3.26228887e-01 -7.28985131e-01
-3.13944221e-01 -1.25840783e-01 -1.88909447e+00 1.27172422e+00
-4.87956405e-01 5.76295197e-01 2.17235357e-01 -1.38763297e+00
1.11321759e+00 4.40380603e-01 8.31077099e-01 -5.15593350e-01
2.28439793e-01 4.05629069e-01 5.44281751e-02 -9.28338110e-01
3.65143269e-01 -3.31350178e-01 3.51792574e-01 2.41905048e-01
-2.18765229e-01 -7.56719708e-02 4.51725930e-01 -3.39644700e-01
5.45452178e-01 -4.25881773e-01 4.00506705e-01 -4.04290169e-01
1.11449933e+00 -3.83258879e-01 5.63716590e-01 1.45043537e-01
1.80800736e-01 4.93616849e-01 7.22584784e-01 -4.51259315e-01
-9.99126613e-01 -4.10744965e-01 -3.79685700e-01 6.90544069e-01
-1.30915165e-01 -3.97578180e-02 -4.58692253e-01 -2.04215571e-01
-1.24139063e-01 4.33846712e-01 -4.68072653e-01 -9.06763598e-02
-3.92350078e-01 -7.12968171e-01 -4.82385904e-02 1.42500564e-01
1.13681838e-01 -8.76485407e-01 -5.81267536e-01 2.05995977e-01
-1.85682118e-01 -9.20574605e-01 -2.43991792e-01 1.77362457e-01
-9.04994845e-01 -1.07653832e+00 -7.30306983e-01 -6.65406764e-01
7.44142771e-01 3.56027216e-01 7.38997459e-01 1.48898493e-02
-1.69041142e-01 2.49986038e-01 -4.04701352e-01 -2.02717125e-01
-1.65466979e-01 -1.06478788e-01 3.90512854e-01 7.52655387e-01
3.82845588e-02 -6.16814852e-01 -4.69277561e-01 5.24654806e-01
-1.11331987e+00 -1.75980687e-01 5.31442583e-01 8.53628695e-01
5.36967516e-01 4.14797753e-01 7.59877563e-01 -4.13878947e-01
9.05893743e-01 -1.72689766e-01 -7.65877783e-01 3.56462121e-01
-3.00554067e-01 3.21285903e-01 1.01991844e+00 -6.34773433e-01
-8.19897711e-01 3.38982254e-01 -2.42818762e-02 -2.85224229e-01
2.72265747e-02 4.62583572e-01 -3.95265311e-01 -2.24819764e-01
7.45324910e-01 2.14624420e-01 2.99859524e-01 -5.30010879e-01
1.76118046e-01 8.19831133e-01 1.44741341e-01 -4.12311852e-01
6.64121389e-01 5.10357738e-01 3.07031214e-01 -1.29408991e+00
-4.68834251e-01 -6.43123686e-01 -4.38355565e-01 -4.32591826e-01
5.81552029e-01 -3.28191370e-01 -1.19996619e+00 2.76910603e-01
-1.13688970e+00 5.51658392e-01 -4.44190025e-01 8.14961612e-01
-5.76346040e-01 6.57868028e-01 -2.05938071e-01 -1.21176052e+00
-2.58452952e-01 -1.09865105e+00 6.82638288e-01 -1.36634842e-01
-1.74418986e-01 -7.39855826e-01 -6.05053082e-02 4.71378863e-02
3.05921584e-01 1.88582838e-01 1.07456434e+00 -4.00622308e-01
2.69143637e-02 -2.51401812e-01 -5.75889014e-02 5.73001444e-01
3.58223945e-01 6.07293956e-02 -7.78444886e-01 -3.00392032e-01
5.58512628e-01 2.15439256e-02 5.37459731e-01 3.31908822e-01
1.35270321e+00 -4.41424906e-01 2.84264505e-01 2.75964022e-01
1.28006136e+00 1.48408294e-01 2.84654409e-01 1.48790836e-01
2.18688145e-01 9.57320273e-01 7.70001113e-01 9.53557730e-01
-2.69854635e-01 7.33693838e-01 5.75206995e-01 -1.34296536e-01
4.22462851e-01 2.65786409e-01 1.80263400e-01 9.79884744e-01
-2.56581306e-01 -1.07133836e-01 -4.92372483e-01 3.90193462e-01
-1.86111259e+00 -9.89753902e-01 -5.09713054e-01 2.41129327e+00
6.32649481e-01 -1.06399506e-01 2.92843282e-01 1.01256192e+00
7.03568220e-01 -2.93042473e-02 -3.45996648e-01 -3.82902294e-01
-1.92851707e-01 -6.29512891e-02 1.38265580e-01 1.84054658e-01
-1.01767480e+00 8.12201574e-02 5.31333733e+00 9.42374229e-01
-1.40510190e+00 -2.72576511e-01 1.66575402e-01 -8.83288980e-02
-4.72824603e-01 -2.31543049e-01 -3.27338874e-01 2.98169494e-01
4.28830087e-01 7.47382268e-02 5.33054829e-01 7.01494932e-01
6.06040120e-01 -2.07173582e-02 -8.37122500e-01 1.36014450e+00
-3.31241675e-02 -8.12011719e-01 5.74423233e-04 1.11171044e-02
3.84871751e-01 -8.59448850e-01 2.42918089e-01 -4.77858901e-01
-9.39671278e-01 -7.86533594e-01 8.96826863e-01 8.18116963e-01
3.98575127e-01 -7.37242639e-01 5.35064280e-01 4.80039477e-01
-1.06559360e+00 -2.70970941e-01 -5.17936587e-01 -1.15991943e-01
1.33259028e-01 1.21964145e+00 -4.86549884e-01 4.14664596e-01
3.45717758e-01 5.91523170e-01 -3.89363915e-01 1.22255623e+00
-5.46875000e-02 2.39091456e-01 -4.81321841e-01 -3.99140567e-01
-1.33095443e-01 -6.52564645e-01 6.70537055e-01 7.82087743e-01
7.58082092e-01 2.08188996e-01 -3.01313043e-01 7.62820303e-01
3.86891961e-01 7.17848301e-01 -5.19174814e-01 -3.57920766e-01
5.44343516e-02 1.40215588e+00 -8.14507544e-01 1.87133625e-02
-4.70571905e-01 6.39314711e-01 -1.86185479e-01 3.29223037e-01
-6.01781785e-01 -3.60921979e-01 5.37530720e-01 1.70008793e-01
1.23846136e-01 -3.74403447e-01 -3.43787074e-01 -1.37484944e+00
3.53182256e-01 -9.96009529e-01 2.33649895e-01 -4.36998725e-01
-1.19167006e+00 5.74926972e-01 -3.34951729e-01 -1.83437860e+00
-9.64180157e-02 -7.02457011e-01 -2.36096218e-01 5.64132869e-01
-1.18854833e+00 -5.78770041e-01 -2.22205088e-01 7.70735264e-01
3.33527565e-01 -1.09112792e-01 7.88810313e-01 6.35270476e-01
-5.11583865e-01 2.27680221e-01 2.55327553e-01 -2.02416897e-01
5.61573386e-01 -7.65259683e-01 -4.74896908e-01 6.43172562e-01
3.40348125e-01 5.24433970e-01 9.87748325e-01 -2.05749705e-01
-1.59038174e+00 -2.82195210e-01 9.14769351e-01 2.73433536e-01
6.52913332e-01 -2.83081114e-01 -7.99914420e-01 -2.45995764e-02
-1.64468318e-01 3.44571797e-03 8.04709077e-01 1.86020255e-01
6.66568875e-02 -2.93847442e-01 -1.04217005e+00 2.36542314e-01
6.16951883e-01 -2.18807101e-01 -4.12339449e-01 3.20145726e-01
5.24646193e-02 1.62106320e-01 -7.63391495e-01 3.50466847e-01
8.30918849e-01 -1.13722789e+00 8.82577300e-01 -6.51347399e-01
1.58782482e-01 -4.91264343e-01 -1.07076719e-01 -1.12578893e+00
-2.86855966e-01 -2.42151320e-01 2.16434494e-01 1.13235295e+00
2.03408822e-01 -6.03140950e-01 6.35381520e-01 3.61903548e-01
3.65039080e-01 -4.62816149e-01 -9.90571439e-01 -7.99840808e-01
-5.28879404e-01 -4.01871979e-01 3.33568782e-01 1.09076273e+00
1.62558362e-01 3.18541229e-01 -6.41531646e-01 3.78240682e-02
5.21163464e-01 3.07465851e-01 5.29942513e-01 -1.56409907e+00
-3.77085239e-01 -3.11004758e-01 -5.81081688e-01 -6.46560073e-01
5.49168326e-02 -3.87133479e-01 -2.62269855e-01 -1.12516940e+00
-2.00615868e-01 -3.58457208e-01 -8.54401812e-02 1.77344278e-01
3.63348067e-01 2.21977249e-01 2.60513127e-01 9.68029052e-02
-4.20125462e-02 6.41396701e-01 1.21629798e+00 -1.42574579e-01
-4.77984339e-01 2.93931574e-01 -3.57290566e-01 8.81766915e-01
4.57639128e-01 -2.91494310e-01 -5.11342287e-01 -2.00283751e-01
7.75290787e-01 2.25220427e-01 2.67407805e-01 -1.05419636e+00
9.60657448e-02 -2.22583041e-01 2.46960297e-01 -4.15517122e-01
3.80946279e-01 -1.43815708e+00 4.78786260e-01 3.49968702e-01
-1.91193253e-01 -6.43687993e-02 -1.47392839e-01 3.36426467e-01
-5.15104830e-01 -5.53520620e-01 6.98896825e-01 2.32838243e-01
-4.49050128e-01 -2.66530812e-01 -4.33922052e-01 -3.35220516e-01
8.48414958e-01 -3.07596415e-01 1.40432879e-01 -5.73840737e-01
-1.17625940e+00 -1.47080451e-01 6.33052960e-02 2.69052476e-01
6.46741629e-01 -1.35005999e+00 -5.37655711e-01 5.08712530e-01
-6.07904159e-02 -4.61752146e-01 3.53669226e-01 1.31951094e+00
-3.12900692e-01 3.49165022e-01 -3.56667459e-01 -4.31647778e-01
-1.33600688e+00 7.09710240e-01 7.46537074e-02 4.48626615e-02
-2.41505295e-01 2.25702867e-01 1.78213697e-02 1.71726998e-02
2.38397643e-01 -4.56786722e-01 -7.05648661e-01 7.31543660e-01
3.47505599e-01 7.02311933e-01 3.87070060e-01 -6.90575957e-01
-3.51045668e-01 8.55722845e-01 6.30047143e-01 -1.32814229e-01
1.54736853e+00 -2.05460321e-02 -3.32245618e-01 4.35843706e-01
1.16428304e+00 3.54089677e-01 -6.90556705e-01 2.52965502e-02
-1.99910313e-01 -4.47552234e-01 -2.59409938e-02 -1.19748123e-01
-1.18674862e+00 9.88959551e-01 6.03632748e-01 7.53093779e-01
1.54027283e+00 -3.78966898e-01 2.59070814e-01 4.36952889e-01
2.31394172e-01 -1.09631598e+00 -1.92266107e-02 1.89589828e-01
1.27452433e+00 -9.39787507e-01 8.39982182e-03 -5.41415572e-01
-3.92725945e-01 1.57800806e+00 1.11157879e-01 6.48609269e-03
8.09881032e-01 -6.81690499e-02 -1.51135266e-01 6.57695532e-02
-1.91521421e-01 -2.47924611e-01 7.47060835e-01 1.49728134e-01
5.29171824e-01 -1.02270648e-01 -1.12062669e+00 8.28080237e-01
-2.26423889e-01 -1.33009136e-01 5.10672927e-01 4.55805510e-01
-4.31499928e-01 -1.11861002e+00 -7.90393710e-01 2.34960333e-01
-2.77541429e-01 1.98572934e-01 -4.28115666e-01 5.22443175e-01
1.07026454e-02 1.08715856e+00 -2.26654097e-01 -2.60119200e-01
5.78819454e-01 2.89536100e-02 2.63458222e-01 -2.71173149e-01
-3.26968372e-01 5.08204341e-01 -1.28496721e-01 -2.84846842e-01
-6.48266852e-01 -6.54287398e-01 -1.07050645e+00 -5.28800767e-03
-5.99417210e-01 4.21271265e-01 8.57442379e-01 7.47319639e-01
1.09930806e-01 3.48786801e-01 8.37975264e-01 -6.35945678e-01
-6.67168617e-01 -8.25701773e-01 -9.89441633e-01 5.67864895e-01
3.60260457e-01 -7.19188869e-01 -5.87328732e-01 -1.20548323e-01] | [12.387856483459473, 0.6351650953292847] |
b1bd20fe-c928-4eb2-9676-3328f65d5a86 | few-shot-object-detection-via-variational | 2301.13411 | null | https://arxiv.org/abs/2301.13411v1 | https://arxiv.org/pdf/2301.13411v1.pdf | Few-Shot Object Detection via Variational Feature Aggregation | As few-shot object detectors are often trained with abundant base samples and fine-tuned on few-shot novel examples,the learned models are usually biased to base classes and sensitive to the variance of novel examples. To address this issue, we propose a meta-learning framework with two novel feature aggregation schemes. More precisely, we first present a Class-Agnostic Aggregation (CAA) method, where the query and support features can be aggregated regardless of their categories. The interactions between different classes encourage class-agnostic representations and reduce confusion between base and novel classes. Based on the CAA, we then propose a Variational Feature Aggregation (VFA) method, which encodes support examples into class-level support features for robust feature aggregation. We use a variational autoencoder to estimate class distributions and sample variational features from distributions that are more robust to the variance of support examples. Besides, we decouple classification and regression tasks so that VFA is performed on the classification branch without affecting object localization. Extensive experiments on PASCAL VOC and COCO demonstrate that our method significantly outperforms a strong baseline (up to 16\%) and previous state-of-the-art methods (4\% in average). Code will be available at: \url{https://github.com/csuhan/VFA} | ['Gui-Song Xia', 'Ke Yan', 'Jian Ding', 'Yuqiang Ren', 'Jiaming Han'] | 2023-01-31 | null | null | null | null | ['few-shot-object-detection'] | ['computer-vision'] | [-7.88176805e-02 -2.01428428e-01 -3.10623616e-01 -5.19176483e-01
-1.15540934e+00 -3.91799808e-01 7.60716558e-01 1.93432078e-01
-2.50832498e-01 5.81791699e-01 -4.84403037e-02 3.93279374e-01
-1.06257841e-01 -8.99052143e-01 -8.10076475e-01 -8.17782521e-01
2.13278025e-01 3.33435625e-01 6.00722730e-01 -6.70327693e-02
6.15227185e-02 2.47084215e-01 -2.04678631e+00 6.20302558e-01
8.07050526e-01 1.13885891e+00 2.32254982e-01 5.79259813e-01
-1.68265581e-01 5.72611094e-01 -7.36066639e-01 -1.96693003e-01
2.64551222e-01 -3.76265407e-01 -4.38941389e-01 1.56151697e-01
5.08586764e-01 -2.87502408e-01 -2.32622117e-01 1.07535470e+00
4.76704180e-01 4.93639231e-01 1.01701987e+00 -1.40075660e+00
-8.38499248e-01 5.59080005e-01 -5.94635963e-01 4.27288532e-01
-8.50695819e-02 3.63210022e-01 1.25323355e+00 -1.45667696e+00
4.81247753e-01 1.15885937e+00 4.52920198e-01 7.05527306e-01
-1.34657049e+00 -7.22078919e-01 4.58778441e-01 3.95215869e-01
-1.53153574e+00 -4.79935527e-01 8.68976176e-01 -6.04218900e-01
7.67522097e-01 1.34832397e-01 5.05127072e-01 1.24247742e+00
-2.98932046e-02 1.25237417e+00 8.82080913e-01 -2.53501207e-01
5.15151978e-01 4.48719025e-01 5.21676958e-01 5.49132764e-01
2.50236183e-01 -6.25431985e-02 -4.58449095e-01 -1.19168080e-01
3.44947457e-01 4.67873037e-01 -1.28395334e-01 -6.26897395e-01
-9.07184958e-01 1.13556504e+00 6.86698854e-01 2.40175039e-01
-3.32708836e-01 1.66808829e-01 3.68330270e-01 1.74470711e-02
5.26910961e-01 2.23432094e-01 -3.81880879e-01 1.60044104e-01
-8.00783813e-01 4.18090880e-01 5.05095422e-01 1.06278193e+00
1.10381782e+00 -8.60793144e-03 -7.96690822e-01 1.14019668e+00
3.15159887e-01 4.18809205e-01 5.51541448e-01 -8.75353932e-01
4.39288951e-02 6.53787374e-01 1.46516168e-03 -5.92916667e-01
-2.45279912e-02 -6.26082957e-01 -4.28844541e-01 1.92596465e-01
1.14291437e-01 1.66999884e-02 -1.05481839e+00 1.68900084e+00
5.11052728e-01 2.75840074e-01 -7.18067121e-03 6.55853271e-01
9.73296106e-01 7.49225438e-01 5.01003154e-02 -1.66608095e-01
1.22725344e+00 -1.04395080e+00 -5.49705207e-01 -1.48077399e-01
5.71171761e-01 -4.29265559e-01 1.21951139e+00 9.44545642e-02
-7.88674474e-01 -6.87228739e-01 -1.11614716e+00 7.72695467e-02
-5.30630708e-01 8.91790017e-02 4.36660737e-01 4.56705064e-01
-5.59924483e-01 6.28861308e-01 -8.32489192e-01 -1.92823395e-01
8.93848360e-01 -6.67527866e-06 7.03884810e-02 -5.48367463e-02
-9.10238326e-01 5.55497468e-01 4.27574724e-01 -2.91987509e-01
-1.20391464e+00 -9.71724093e-01 -1.00170040e+00 1.63547680e-01
5.16900659e-01 -5.73546767e-01 1.38803685e+00 -9.06027615e-01
-1.24335599e+00 5.64551294e-01 -3.18577141e-01 -4.42107975e-01
3.27203810e-01 -2.54866481e-01 -2.26684317e-01 6.97548836e-02
3.66181821e-01 6.64857388e-01 1.15802491e+00 -1.27573264e+00
-9.50260878e-01 -2.32127786e-01 -8.88753459e-02 -7.00829104e-02
-4.32888448e-01 -2.90576667e-01 -3.01484287e-01 -5.76770902e-01
-1.36567995e-01 -6.61751091e-01 8.32570791e-02 2.02673033e-01
-1.52069822e-01 -6.79992735e-01 9.53993082e-01 -3.66296060e-02
1.24647141e+00 -2.40144086e+00 1.13356588e-02 -1.12211160e-01
2.62908190e-01 2.33724043e-01 -2.53366202e-01 1.68172747e-01
2.11220205e-01 -1.82843119e-01 -2.57101834e-01 -3.88885975e-01
2.57868588e-01 2.77114928e-01 -4.53528553e-01 3.56555015e-01
4.58405316e-01 8.50337267e-01 -1.04877079e+00 -4.98158485e-01
2.93796450e-01 4.42602426e-01 -6.63639247e-01 2.41570935e-01
-4.10683453e-01 1.05071127e-01 -4.58784908e-01 9.93032932e-01
6.07220471e-01 -3.67249936e-01 -2.68518418e-01 -2.75911808e-01
1.45925032e-02 2.01312527e-02 -1.18265748e+00 1.37338638e+00
-3.76039863e-01 3.56680900e-01 -2.58204311e-01 -9.62248623e-01
8.75853479e-01 -7.16675073e-02 3.18095714e-01 -3.75916541e-01
2.24794358e-01 2.51976252e-01 -6.60544932e-02 -1.97871462e-01
3.17600936e-01 7.84074143e-03 -2.23632175e-02 1.97640687e-01
5.39549649e-01 -4.36901450e-02 3.96447122e-01 2.56158590e-01
8.55813682e-01 9.10306871e-02 3.50021392e-01 -2.11082548e-01
3.05925548e-01 -1.31478190e-01 8.58117223e-01 1.12796259e+00
-5.95331967e-01 6.41435087e-01 3.25666040e-01 -3.67676735e-01
-8.01943243e-01 -1.37600601e+00 -3.85360390e-01 1.38845408e+00
1.31066605e-01 -4.75387335e-01 -5.41654825e-01 -9.83479679e-01
2.32595682e-01 9.56119776e-01 -8.86501968e-01 -4.53932673e-01
-1.94441721e-01 -7.68387973e-01 1.04480376e-02 8.29800308e-01
2.74850458e-01 -8.79217327e-01 -5.80063105e-01 1.52814150e-01
1.94241107e-01 -8.78530979e-01 -4.09699947e-01 3.28609437e-01
-6.42849743e-01 -1.04225123e+00 -7.76173472e-01 -5.58374465e-01
3.93969119e-01 3.50576490e-01 9.53224719e-01 -1.32798433e-01
-5.65881968e-01 4.57882434e-01 -5.79269707e-01 -7.25225151e-01
-9.30886120e-02 -8.86442289e-02 1.47889629e-01 2.63106883e-01
6.59970641e-01 -5.01360059e-01 -6.19874895e-01 1.21662803e-01
-7.61929333e-01 -3.51533532e-01 4.68742698e-01 1.14613783e+00
8.19889665e-01 -3.08416784e-01 6.98660195e-01 -8.49795103e-01
2.94478595e-01 -7.26705074e-01 -4.67914373e-01 2.41143465e-01
-4.51460868e-01 -6.48076162e-02 5.36838353e-01 -5.88541687e-01
-8.73034656e-01 7.25729391e-02 1.69977233e-01 -9.77591515e-01
-1.53250009e-01 9.45339054e-02 -1.30347639e-01 1.65254295e-01
9.83486474e-01 3.40295374e-01 -1.98074400e-01 -3.16245288e-01
4.89426285e-01 6.17724955e-01 1.53717324e-01 -5.69815993e-01
7.00250685e-01 6.08822525e-01 -4.25508857e-01 -9.34329093e-01
-1.23333812e+00 -6.72428071e-01 -6.10289216e-01 -1.98717803e-01
6.78386688e-01 -1.04568529e+00 -1.45631567e-01 3.63615751e-01
-9.29486394e-01 -2.86840498e-01 -7.87028015e-01 4.86194313e-01
-6.43683016e-01 3.26395780e-02 -3.30047011e-01 -8.58270288e-01
-2.15540797e-01 -1.14274633e+00 1.19951665e+00 4.44247991e-01
-5.73469289e-02 -6.86314762e-01 8.05356503e-02 2.25063898e-02
1.93053991e-01 1.48052499e-01 6.45663440e-01 -9.40863490e-01
-7.16420472e-01 -3.04039180e-01 -2.29997560e-01 4.44225758e-01
1.76939115e-01 1.54352352e-01 -1.37300825e+00 -3.30139220e-01
-2.21614257e-01 -6.02699876e-01 1.37017143e+00 5.06293476e-01
1.26183760e+00 -9.91867632e-02 -3.88449013e-01 5.43822289e-01
1.26124895e+00 -2.45508086e-02 2.20497742e-01 9.50734392e-02
5.70420802e-01 3.18503052e-01 7.56616354e-01 7.96011031e-01
2.43547544e-01 6.14276230e-01 3.18291545e-01 4.46346521e-01
-2.86039233e-01 -2.57237047e-01 3.04343462e-01 4.28997785e-01
1.68447673e-01 -2.30204873e-02 -5.59731066e-01 8.12581360e-01
-1.96644354e+00 -1.12899435e+00 2.20020726e-01 2.17110419e+00
7.63864815e-01 1.38558790e-01 3.28836530e-01 8.32516886e-03
7.19624758e-01 3.05385798e-01 -7.37569153e-01 -1.03201099e-01
5.44427335e-02 1.23256132e-01 7.01453462e-02 3.30814928e-01
-1.26416898e+00 9.16541278e-01 5.49503422e+00 1.10918176e+00
-8.90927672e-01 3.66125315e-01 5.10408759e-01 -4.97249186e-01
-2.56927788e-01 -1.82489380e-01 -1.17319477e+00 5.23070455e-01
6.45778060e-01 -3.01400483e-01 9.28405598e-02 1.40961647e+00
-3.60002071e-01 7.86285698e-02 -1.22537208e+00 9.89782512e-01
1.30179703e-01 -1.40386081e+00 1.22466587e-01 -2.66892225e-01
7.93444812e-01 2.32199967e-01 2.59871542e-01 9.84951079e-01
3.46565962e-01 -5.30343592e-01 7.79380500e-01 5.36164761e-01
3.76738489e-01 -7.37983882e-01 5.62188268e-01 3.44281971e-01
-1.24277997e+00 -4.03722763e-01 -7.15860069e-01 1.82573661e-01
-1.60242453e-01 7.26776719e-01 -5.79552889e-01 3.20160180e-01
9.99187529e-01 9.65147316e-01 -6.31941020e-01 9.70476508e-01
-2.07285672e-01 5.25023878e-01 -1.98346764e-01 -3.62892687e-01
1.27917260e-01 1.93384528e-01 7.05540776e-01 1.16053259e+00
1.70731008e-01 5.16607165e-02 4.02147293e-01 1.19938564e+00
-4.92219999e-02 -7.02626631e-02 -4.42711532e-01 9.34223086e-03
6.20549917e-01 1.33303773e+00 -6.60068631e-01 -6.25374854e-01
-5.13448775e-01 9.26136136e-01 5.35913110e-01 3.73970926e-01
-8.98310959e-01 -5.46313345e-01 7.41218746e-01 -6.55998960e-02
7.46213734e-01 2.13611096e-01 -6.91519901e-02 -1.38809276e+00
-5.83312055e-03 -5.78121245e-01 6.43954098e-01 -4.58672523e-01
-1.67978704e+00 5.33709168e-01 1.87202632e-01 -1.31504858e+00
-1.39071599e-01 -6.60650849e-01 -7.30204999e-01 6.17140114e-01
-1.48990846e+00 -9.98402715e-01 -2.33787879e-01 5.96540451e-01
9.47906494e-01 -2.76187629e-01 6.45989180e-01 1.45911068e-01
-5.59059978e-01 8.36851239e-01 2.67003298e-01 7.11423233e-02
6.36163771e-01 -1.15744865e+00 -4.25495617e-02 6.91585243e-01
3.06562364e-01 5.41338265e-01 5.59283793e-01 -5.20090997e-01
-1.10480773e+00 -1.40058815e+00 4.47113037e-01 -7.07115233e-01
6.77901447e-01 -4.36671942e-01 -1.08663559e+00 6.34552002e-01
-1.55430153e-01 6.98812962e-01 7.32821882e-01 1.65250659e-01
-6.82343304e-01 -3.67773116e-01 -1.07457590e+00 4.05740529e-01
8.73012185e-01 -5.33982277e-01 -6.65135562e-01 3.60823005e-01
8.52957547e-01 -8.09575990e-02 -5.20484328e-01 5.61862886e-01
4.27571803e-01 -9.72355008e-01 8.77599120e-01 -6.28269196e-01
3.40601832e-01 -3.73811096e-01 -5.75450242e-01 -1.24904621e+00
-6.12743378e-01 -8.44938084e-02 -6.64700866e-01 1.20190203e+00
3.93823415e-01 -5.69403291e-01 5.12474120e-01 2.80220777e-01
-2.19281241e-01 -9.67349708e-01 -9.20919776e-01 -1.06193888e+00
4.72106040e-03 -3.68264526e-01 4.29073930e-01 8.87840331e-01
-3.08231145e-01 2.29955241e-01 -1.19760558e-01 1.30174488e-01
7.55917132e-01 4.98789400e-01 7.56751060e-01 -1.18503559e+00
-5.03495336e-01 -5.34940720e-01 -5.04944444e-01 -6.41218126e-01
1.26797855e-01 -9.85833764e-01 2.04444066e-01 -1.23227382e+00
4.86869484e-01 -2.45173275e-01 -6.71844721e-01 5.91051042e-01
-3.91365290e-01 1.88813820e-01 3.26405823e-01 2.38877505e-01
-7.96690226e-01 9.61203814e-01 8.51881862e-01 -3.10361505e-01
-2.71513551e-01 1.15534350e-01 -6.62972510e-01 7.85256982e-01
6.18852615e-01 -6.23006284e-01 -3.58605266e-01 -1.48785546e-01
-2.36445501e-01 -5.28440773e-01 4.68607008e-01 -1.08300519e+00
3.45482081e-02 -1.61393315e-01 6.22314632e-01 -7.13397205e-01
4.63922203e-01 -4.76215720e-01 -3.82575542e-01 4.40936118e-01
-3.36240649e-01 -5.78778684e-01 8.60655308e-02 8.37399602e-01
-9.81618389e-02 -3.44076276e-01 1.13931739e+00 -1.37436852e-01
-8.24300945e-01 5.63650429e-01 -6.62251189e-02 1.86382741e-01
1.27287316e+00 -1.01686619e-01 -2.83465713e-01 -9.84481201e-02
-6.97373450e-01 3.31375062e-01 1.85073167e-01 4.97642398e-01
7.00671613e-01 -1.42959273e+00 -7.37163603e-01 2.18963742e-01
6.34060323e-01 4.76545505e-02 4.28708971e-01 6.92694724e-01
7.95746073e-02 3.37892324e-02 6.67880028e-02 -9.76297796e-01
-9.54793155e-01 8.11497986e-01 2.73306251e-01 1.23885766e-01
-5.13867438e-01 1.16345167e+00 5.00117004e-01 -3.72819632e-01
3.09971720e-01 -1.69506103e-01 -1.31255433e-01 3.91322821e-01
8.96249175e-01 3.59303504e-01 -1.94130510e-01 -4.04776841e-01
-4.24021333e-01 3.79508108e-01 -4.51293737e-01 1.85568079e-01
1.37288749e+00 8.43053758e-02 2.93697089e-01 8.99228275e-01
1.22697890e+00 -2.43277594e-01 -1.57956922e+00 -5.46523929e-01
-1.79891229e-01 -5.58617711e-01 -4.62453999e-02 -4.98455465e-01
-9.81261432e-01 9.46110845e-01 7.72290826e-01 9.49766263e-02
9.18484986e-01 4.09753770e-01 4.16772217e-01 4.11592275e-01
1.50593460e-01 -1.20196283e+00 4.20576364e-01 5.09465992e-01
8.26432168e-01 -1.45147073e+00 -6.70571700e-02 -2.17317984e-01
-7.18505859e-01 8.56330454e-01 9.43774939e-01 -4.28284526e-01
7.87877619e-01 -4.00154516e-02 -1.05842024e-01 -7.43613765e-02
-8.60078514e-01 -4.30946648e-01 5.19306540e-01 5.45182467e-01
1.68608114e-01 1.89430360e-02 -5.31486459e-02 1.08410060e+00
2.24848628e-01 -7.27175847e-02 2.29225799e-01 1.04372776e+00
-8.20347071e-01 -6.46284103e-01 -1.66085526e-01 6.12632096e-01
-1.29138157e-01 -4.71689254e-02 -1.28294662e-01 6.31919146e-01
2.58579910e-01 6.75397992e-01 1.46343768e-01 -3.37649733e-01
3.22170287e-01 2.62098819e-01 4.14458185e-01 -1.01750302e+00
-2.90229887e-01 8.96158293e-02 -4.27413344e-01 -6.78330004e-01
-2.66804010e-01 -7.20514238e-01 -1.08067358e+00 1.96474984e-01
-5.64911902e-01 5.31848613e-03 3.29536349e-01 8.30898821e-01
4.12150741e-01 7.32003868e-01 7.26477623e-01 -9.80825245e-01
-9.63209450e-01 -9.27734613e-01 -6.84088588e-01 3.35909575e-01
4.32763189e-01 -1.00599658e+00 -5.85259259e-01 -2.16228947e-01] | [9.794238090515137, 2.381481170654297] |
46dde1a3-1e28-439e-a70c-95616043dc6a | hallucinet-ing-spatiotemporal-representations | 1912.04430 | null | https://arxiv.org/abs/1912.04430v3 | https://arxiv.org/pdf/1912.04430v3.pdf | HalluciNet-ing Spatiotemporal Representations Using a 2D-CNN | Spatiotemporal representations learned using 3D convolutional neural networks (CNN) are currently used in state-of-the-art approaches for action related tasks. However, 3D-CNN are notorious for being memory and compute resource intensive as compared with more simple 2D-CNN architectures. We propose to hallucinate spatiotemporal representations from a 3D-CNN teacher with a 2D-CNN student. By requiring the 2D-CNN to predict the future and intuit upcoming activity, it is encouraged to gain a deeper understanding of actions and how they evolve. The hallucination task is treated as an auxiliary task, which can be used with any other action related task in a multitask learning setting. Thorough experimental evaluation shows that the hallucination task indeed helps improve performance on action recognition, action quality assessment, and dynamic scene recognition tasks. From a practical standpoint, being able to hallucinate spatiotemporal representations without an actual 3D-CNN can enable deployment in resource-constrained scenarios, such as with limited computing power and/or lower bandwidth. Codebase is available here: https://github.com/ParitoshParmar/HalluciNet. | ['Paritosh Parmar', 'Brendan Morris'] | 2019-12-10 | null | null | null | null | ['action-quality-assessment', 'action-recognition-in-still-images', 'action-anticipation', 'scene-recognition', 'fine-grained-action-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 3.30999196e-02 1.38330698e-01 -4.04606313e-01 7.40299076e-02
-3.26358318e-01 -2.74992406e-01 5.99012196e-01 -1.05431885e-01
-2.86845088e-01 4.50396597e-01 4.95888025e-01 -1.25870019e-01
2.04006523e-01 -6.63759410e-01 -4.92500275e-01 -5.66699326e-01
-4.19393592e-02 2.32150704e-01 3.05991203e-01 -9.56306309e-02
3.92367356e-02 6.29180908e-01 -1.56208146e+00 5.08404076e-01
3.68267119e-01 1.04566908e+00 2.49088913e-01 5.54718792e-01
3.34477812e-01 1.43469083e+00 -6.18268728e-01 1.91940993e-01
1.36280507e-01 -3.52697641e-01 -7.39871204e-01 3.60167652e-01
3.50562125e-01 -5.89931786e-01 -8.55557799e-01 5.38275540e-01
5.68676114e-01 4.23921883e-01 2.94722915e-01 -1.34771287e+00
-7.16150522e-01 5.09349220e-02 -2.17985302e-01 5.89683592e-01
3.50866616e-01 6.59242094e-01 4.93271142e-01 -6.45961285e-01
5.26443779e-01 1.12154603e+00 4.04165566e-01 6.05710089e-01
-1.02179873e+00 -4.41616029e-01 1.69510409e-01 4.51128393e-01
-1.17425168e+00 -5.77379167e-01 6.43465996e-01 -3.14626455e-01
1.38054943e+00 -3.78795676e-02 9.25725937e-01 1.73763597e+00
1.90439507e-01 1.21855903e+00 8.95005167e-01 1.46669373e-01
3.51190865e-01 -9.13098827e-02 -3.93176287e-01 5.99176705e-01
-2.47545972e-01 1.72624111e-01 -8.94685447e-01 2.95076609e-01
9.83913064e-01 2.56267399e-01 -2.33087942e-01 -3.70771736e-01
-1.29339778e+00 5.99815845e-01 5.71454406e-01 3.68969291e-01
-7.19029248e-01 5.98233044e-01 4.71390933e-01 2.46346101e-01
6.73842728e-01 6.60276115e-01 -3.59443456e-01 -7.52666533e-01
-7.90430725e-01 2.39921793e-01 2.30190858e-01 8.02210331e-01
4.49004471e-01 4.75142419e-01 -4.22644913e-01 4.45629120e-01
-1.95142984e-01 2.51918882e-01 7.83556998e-01 -1.31793618e+00
1.85549513e-01 6.54974937e-01 7.89107308e-02 -6.73726201e-01
-3.55738401e-01 -3.76725048e-01 -6.10836506e-01 2.88005203e-01
2.20521778e-01 -3.20041776e-02 -1.06752396e+00 1.54388058e+00
1.89428404e-01 7.41906464e-01 1.17435515e-01 1.02367699e+00
7.33500600e-01 5.40192485e-01 2.39828810e-01 1.51006907e-01
1.26897228e+00 -1.10922158e+00 -5.48224628e-01 -3.92924935e-01
8.59580338e-01 -2.42114365e-01 1.09969723e+00 4.11547601e-01
-1.11108863e+00 -8.04871500e-01 -8.09799671e-01 -2.16281503e-01
-5.38442016e-01 2.03202859e-01 8.46063852e-01 2.29608715e-01
-1.11495459e+00 8.17603171e-01 -1.26587284e+00 -6.84158981e-01
9.24348533e-01 1.29417270e-01 -4.03538108e-01 5.03066462e-03
-9.59997833e-01 1.43326044e+00 3.50070059e-01 -1.72781527e-01
-1.51835096e+00 -6.79949462e-01 -8.25327992e-01 1.48219898e-01
2.87710726e-01 -5.26856422e-01 1.39904070e+00 -7.85242140e-01
-1.27870214e+00 8.17038059e-01 1.16795348e-02 -6.74384236e-01
5.20231247e-01 -3.84978563e-01 -3.26769650e-01 3.35474402e-01
1.27345890e-01 7.84830987e-01 7.43932247e-01 -6.51967347e-01
-2.94070184e-01 -4.49874938e-01 2.14435816e-01 5.87332845e-01
-4.41420197e-01 -3.32295656e-01 -3.45405251e-01 -4.36589032e-01
-8.72463658e-02 -1.00745463e+00 -2.29185432e-01 5.27885973e-01
-2.27625072e-01 -2.85270661e-01 1.28627658e+00 -4.16033238e-01
6.99226081e-01 -2.29531932e+00 6.69622049e-02 -4.64545429e-01
2.48350263e-01 5.77433527e-01 -2.50349700e-01 4.66998309e-01
-3.23210478e-01 -1.21998712e-01 2.69709200e-01 -4.77152765e-01
-3.13411474e-01 2.76897371e-01 -2.25187480e-01 5.64094424e-01
5.32822073e-01 1.09093046e+00 -1.18246198e+00 -1.53295234e-01
5.33256471e-01 7.04567432e-01 -2.26437777e-01 1.88687250e-01
-4.14880544e-01 7.89815009e-01 -5.93654156e-01 7.86595881e-01
1.09270826e-01 -5.32348394e-01 5.11869378e-02 -8.07578340e-02
-6.10288717e-02 3.10104430e-01 -6.72181308e-01 2.20659876e+00
-7.17608154e-01 1.02399457e+00 -5.28966963e-01 -1.18776023e+00
8.10678065e-01 7.12438226e-01 7.27293611e-01 -1.03224766e+00
1.15349665e-01 -2.43062854e-01 -2.72278726e-01 -8.32136393e-01
5.46010613e-01 -9.23410952e-02 6.88927993e-02 5.88114679e-01
3.46268982e-01 1.85508300e-02 -2.44661327e-02 4.09897864e-02
1.49349761e+00 6.26879990e-01 2.23045737e-01 1.15295589e-01
-1.10154063e-01 1.83489606e-01 3.30026329e-01 4.44783092e-01
-5.15632153e-01 4.68791068e-01 4.04901236e-01 -7.29638338e-01
-8.95567775e-01 -1.03734779e+00 2.67049462e-01 1.06504095e+00
-1.48441996e-02 -2.73004323e-01 -2.41071984e-01 -5.89087546e-01
-5.28951623e-02 8.05064559e-01 -8.40794444e-01 -5.06247282e-01
-4.38470185e-01 -2.20038295e-01 6.48626626e-01 8.96605909e-01
6.26829624e-01 -1.22561359e+00 -1.30587399e+00 2.10082680e-01
-8.55717286e-02 -1.16386437e+00 -1.97018951e-01 2.92083740e-01
-9.71563339e-01 -9.60922420e-01 -8.77817333e-01 -3.45867932e-01
3.81790638e-01 5.26515663e-01 9.13684249e-01 -8.42694640e-02
-3.15375358e-01 6.40088975e-01 -4.76327300e-01 -4.08422559e-01
-8.64768326e-02 -1.82364926e-01 1.85276404e-01 -1.97881222e-01
5.80103040e-01 -1.02698672e+00 -8.31470072e-01 1.64662614e-01
-7.80482233e-01 1.32229105e-01 4.65697795e-01 4.18066889e-01
3.71820569e-01 -3.53661738e-02 3.86999547e-01 -3.49782318e-01
5.60543001e-01 -6.10177577e-01 -2.51554906e-01 -3.92071940e-02
-2.55116343e-01 -5.55415899e-02 4.70617771e-01 -9.55819130e-01
-8.72943282e-01 8.22410285e-02 2.13674173e-01 -1.15977585e+00
-4.93387520e-01 3.18649858e-01 1.62946373e-01 2.91369766e-01
8.74380767e-01 5.19115150e-01 3.63074914e-02 -4.41411585e-01
2.80111581e-01 4.74248439e-01 5.30059040e-01 -2.44021952e-01
4.15776610e-01 6.48710489e-01 -3.39390151e-02 -8.22460949e-01
-1.14192569e+00 -4.81283396e-01 -6.69965684e-01 -5.82924068e-01
1.09353697e+00 -1.26791251e+00 -8.85631204e-01 3.08487177e-01
-1.16269100e+00 -9.04022932e-01 -5.21001697e-01 5.04855275e-01
-9.74687099e-01 -6.13861345e-02 -2.41034344e-01 -6.05851650e-01
1.18752033e-01 -1.11159074e+00 1.17298388e+00 2.54946887e-01
-4.75392014e-01 -1.12787163e+00 9.06553641e-02 5.91250539e-01
5.57418227e-01 4.50692981e-01 5.56111634e-01 -6.41846836e-01
-5.66093922e-01 -1.76997453e-01 -4.65463661e-02 2.08047986e-01
1.51188225e-01 -4.77264792e-01 -1.16682541e+00 -1.24688320e-01
-8.81621838e-02 -8.13069582e-01 7.09801018e-01 2.29659438e-01
1.60278738e+00 -2.23603964e-01 -1.51101664e-01 4.07388985e-01
1.15506959e+00 2.13307306e-01 7.60201991e-01 1.57239869e-01
4.59881127e-01 3.36003631e-01 6.11707032e-01 7.09514380e-01
3.34483743e-01 7.41608500e-01 7.75313079e-01 -5.50999753e-02
-4.73045796e-01 -4.31807131e-01 4.41629648e-01 2.15005279e-01
-2.52977401e-01 -3.01236451e-01 -1.07083488e+00 6.93000853e-01
-2.12813783e+00 -1.20274365e+00 2.77694792e-01 1.89321160e+00
5.24468422e-01 -5.78345507e-02 2.11408958e-01 -8.59411284e-02
2.43631899e-01 5.23108125e-01 -1.12471116e+00 -2.34350383e-01
2.16327786e-01 2.32311308e-01 1.77778065e-01 -6.72282353e-02
-1.07368040e+00 1.11823368e+00 5.72298574e+00 5.44799030e-01
-1.40475154e+00 2.85951197e-01 4.37329412e-01 -4.77349788e-01
2.07690313e-01 -2.22664103e-01 -3.52740884e-01 3.58475059e-01
1.13193381e+00 -1.95749402e-01 2.79803932e-01 1.01526439e+00
5.62408149e-01 -3.28709006e-01 -1.34627748e+00 1.12361479e+00
1.01955585e-01 -1.51494896e+00 2.89188810e-02 1.79081541e-02
5.30483365e-01 3.39672744e-01 3.45974386e-01 5.15371263e-01
2.78159261e-01 -1.32314360e+00 5.40054917e-01 6.57133043e-01
7.15014756e-01 -3.87540013e-01 2.82660007e-01 5.80660820e-01
-1.05457389e+00 -1.08595178e-01 -2.69177735e-01 -2.99696952e-01
1.11024596e-01 1.40298441e-01 -7.34635293e-01 2.45351195e-01
7.37378895e-01 1.11605465e+00 -5.57462573e-01 8.89090300e-01
-2.24744067e-01 4.07310516e-01 -8.55474398e-02 4.29773480e-02
4.88742352e-01 1.66725129e-01 4.47997332e-01 8.37641358e-01
5.53098619e-01 3.33511353e-01 2.12410912e-01 8.02714467e-01
5.26763313e-03 -4.08828408e-01 -1.01557052e+00 -3.88657600e-01
2.47024029e-01 9.20294464e-01 -4.42306638e-01 -4.70948011e-01
-3.07402581e-01 1.17425036e+00 3.27534288e-01 3.95245492e-01
-9.06014264e-01 9.65617448e-02 9.23207879e-01 2.10431457e-01
3.23786646e-01 -4.77025509e-01 -7.47485235e-02 -1.25898361e+00
-2.22383812e-01 -7.81667233e-01 1.46358356e-01 -1.48539376e+00
-6.63691938e-01 4.49082404e-01 7.40970718e-03 -1.51532912e+00
-2.78909862e-01 -5.48729062e-01 -7.56714642e-01 4.99444634e-01
-1.40913510e+00 -1.11748242e+00 -5.88332593e-01 8.85186791e-01
8.93533826e-01 -1.14168704e-01 1.02230585e+00 5.58972545e-03
-5.82999349e-01 7.28259534e-02 -4.90565538e-01 6.99661747e-02
4.08555776e-01 -9.41534817e-01 1.81509361e-01 6.95499182e-01
2.48914659e-01 2.02607051e-01 7.36670077e-01 -4.89760280e-01
-1.47617006e+00 -1.38461542e+00 5.54691792e-01 -6.00751936e-01
7.57360339e-01 -1.56795993e-01 -7.41402924e-01 8.18547130e-01
3.17071855e-01 3.42380434e-01 7.36617506e-01 2.54189819e-02
-2.84526229e-01 1.95254728e-01 -8.48666966e-01 8.32050622e-01
1.29731870e+00 -7.72731304e-01 -5.22013485e-01 7.74168253e-01
5.01885474e-01 -3.79192889e-01 -9.82248425e-01 1.05623841e-01
4.03100759e-01 -1.00858128e+00 8.33884239e-01 -9.80528653e-01
7.12474346e-01 -8.36327821e-02 -7.56967664e-02 -1.40377223e+00
-8.96913186e-02 -5.71071744e-01 -5.71311891e-01 3.46780360e-01
1.98393151e-01 -2.90598243e-01 1.11398053e+00 4.17067319e-01
-3.18794221e-01 -8.34900200e-01 -8.81838799e-01 -1.08975303e+00
-1.96718499e-01 -5.89279294e-01 2.35381499e-01 9.41520989e-01
1.01644583e-01 2.29513094e-01 -6.02112114e-01 1.97826698e-01
2.28766188e-01 -7.60801360e-02 7.85475791e-01 -8.75366449e-01
-2.44923085e-01 -1.97709113e-01 -6.24414325e-01 -1.28989434e+00
2.29120225e-01 -7.74213791e-01 -9.49217156e-02 -1.61132026e+00
2.63409801e-02 -4.46611643e-02 -3.06586385e-01 1.03541803e+00
3.26660395e-01 3.03364605e-01 3.28033447e-01 4.69639674e-02
-8.67901444e-01 9.44801152e-01 1.63105714e+00 -1.40251338e-01
-1.86224908e-01 1.08115472e-01 -4.63296741e-01 5.58620870e-01
1.08496809e+00 -3.30454588e-01 -8.18828404e-01 -5.13856888e-01
-1.65384084e-01 4.19823468e-01 8.32617104e-01 -1.41836131e+00
2.57219195e-01 -2.41817012e-01 4.24280494e-01 -4.55201626e-01
1.08253467e+00 -6.30683005e-01 1.49585620e-01 5.00149488e-01
-4.38772082e-01 2.46570241e-02 3.23199570e-01 5.00783563e-01
-2.12708399e-01 2.13557750e-01 5.92482150e-01 -5.52556276e-01
-9.64445710e-01 5.90356529e-01 -7.93778002e-01 -1.67931497e-01
1.25299048e+00 -6.37283325e-01 -5.08658946e-01 -6.55780017e-01
-1.12443936e+00 9.24878716e-02 3.66119772e-01 6.80177212e-01
8.60674560e-01 -1.39959133e+00 -3.74083161e-01 2.74103358e-02
2.59894788e-01 -2.02980071e-01 4.20687169e-01 9.82282877e-01
-2.01305971e-01 5.00804484e-01 -6.90960348e-01 -5.57785571e-01
-1.15590835e+00 4.90958244e-01 3.84631038e-01 -1.28796279e-01
-7.68379867e-01 6.67673945e-01 -2.21377276e-02 -1.33338973e-01
3.75443667e-01 -2.26926759e-01 -1.20857641e-01 6.18539713e-02
4.66607213e-01 3.63883764e-01 -1.48133665e-01 -4.50012684e-01
-2.48146027e-01 1.78637225e-02 1.18136480e-01 4.49719541e-02
1.64468050e+00 1.49800941e-01 4.19666827e-01 5.50804853e-01
1.15188575e+00 -9.34759676e-01 -1.62529528e+00 -2.49589235e-01
-2.10451469e-01 -4.61134493e-01 3.64721864e-01 -7.28557467e-01
-1.07371354e+00 1.27256179e+00 6.08393073e-01 1.67780891e-01
1.16316593e+00 1.27232581e-01 7.16503203e-01 4.67379272e-01
3.19789886e-01 -9.42649722e-01 8.96852553e-01 4.94967341e-01
9.96515572e-01 -1.17431223e+00 -3.70884873e-02 2.49878734e-01
-9.98424172e-01 1.09333062e+00 8.93163800e-01 -1.21435836e-01
6.22040510e-01 -2.28381157e-03 -4.92159789e-03 -4.81336623e-01
-1.18170846e+00 -4.64444309e-01 6.77106008e-02 8.13942194e-01
2.32234955e-01 -4.22863215e-02 6.08828902e-01 5.73128909e-02
1.60073057e-01 2.36570805e-01 6.56413496e-01 1.05318284e+00
-2.72707999e-01 -7.63345838e-01 1.95315063e-01 3.11165631e-01
-1.30938396e-01 1.45909414e-01 -1.19303927e-01 7.14631557e-01
1.99989155e-01 8.06975245e-01 3.26398134e-01 -2.93516755e-01
3.74165326e-01 9.00356770e-02 4.68449831e-01 -9.48618174e-01
-4.51572329e-01 -2.83465505e-01 2.03554168e-01 -1.21994805e+00
-6.48548484e-01 -6.33665860e-01 -9.01837051e-01 -2.63124913e-01
1.60584807e-01 -4.45033997e-01 5.44904113e-01 8.30429494e-01
6.56567693e-01 6.91585302e-01 3.29575986e-01 -1.07484221e+00
-4.67918485e-01 -1.00519729e+00 -4.40670639e-01 2.00561970e-01
3.39044183e-01 -8.68871748e-01 -4.16434510e-03 9.43417400e-02] | [8.337058067321777, 0.6803867220878601] |
27f19269-743d-4798-97ee-ded0c4ac2846 | domain-aware-triplet-loss-in-domain | 2303.01233 | null | https://arxiv.org/abs/2303.01233v1 | https://arxiv.org/pdf/2303.01233v1.pdf | Domain-aware Triplet loss in Domain Generalization | Despite much progress being made in the field of object recognition with the advances of deep learning, there are still several factors negatively affecting the performance of deep learning models. Domain shift is one of these factors and is caused by discrepancies in the distributions of the testing and training data. In this paper, we focus on the problem of compact feature clustering in domain generalization to help optimize the embedding space from multi-domain data. We design a domainaware triplet loss for domain generalization to help the model to not only cluster similar semantic features, but also to disperse features arising from the domain. Unlike previous methods focusing on distribution alignment, our algorithm is designed to disperse domain information in the embedding space. The basic idea is motivated based on the assumption that embedding features can be clustered based on domain information, which is mathematically and empirically supported in this paper. In addition, during our exploration of feature clustering in domain generalization, we note that factors affecting the convergence of metric learning loss in domain generalization are more important than the pre-defined domains. To solve this issue, we utilize two methods to normalize the embedding space, reducing the internal covariate shift of the embedding features. The ablation study demonstrates the effectiveness of our algorithm. Moreover, the experiments on the benchmark datasets, including PACS, VLCS and Office-Home, show that our method outperforms related methods focusing on domain discrepancy. In particular, our results on RegnetY-16 are significantly better than state-of-the-art methods on the benchmark datasets. Our code will be released at https://github.com/workerbcd/DCT | ['Brian Lovell', 'Kaiyu Guo'] | 2023-03-01 | null | null | null | null | ['metric-learning', 'object-recognition', 'metric-learning'] | ['computer-vision', 'computer-vision', 'methodology'] | [-3.29732858e-02 -2.29569808e-01 -2.96071619e-01 -6.72663629e-01
-5.59449375e-01 -5.07596433e-01 2.63126612e-01 -9.79196653e-02
-3.72786790e-01 6.91170394e-01 2.17935681e-01 -5.47041669e-02
-5.38459778e-01 -6.39129162e-01 -5.22423029e-01 -9.05052662e-01
1.42541528e-01 3.95559788e-01 1.51669040e-01 -7.31673837e-02
1.94908574e-01 5.30897498e-01 -1.49973726e+00 2.71945655e-01
1.09523296e+00 1.14984596e+00 2.98384517e-01 -3.97760458e-02
-8.09460133e-02 3.22929680e-01 -6.01410508e-01 -1.85271025e-01
4.57378358e-01 -2.77019799e-01 -6.83192968e-01 1.38186634e-01
2.98841774e-01 -3.83503765e-01 -4.82986748e-01 1.17855620e+00
6.38338387e-01 1.06140770e-01 9.12254810e-01 -1.57061052e+00
-9.08681154e-01 3.28178942e-01 -4.57458168e-01 1.43673629e-01
-2.04577699e-01 -1.22642837e-01 1.03325081e+00 -9.27047372e-01
4.92260665e-01 1.11590314e+00 3.64954114e-01 7.17594743e-01
-1.01662803e+00 -9.97278631e-01 2.65108287e-01 6.21254802e-01
-1.58094823e+00 -1.87212557e-01 1.06558526e+00 -5.54621041e-01
5.90599656e-01 -4.37500142e-02 2.12719306e-01 1.18216324e+00
6.93696216e-02 6.78541243e-01 7.65105069e-01 -3.78370076e-01
3.80930603e-01 5.86857796e-01 1.15506083e-01 2.61080146e-01
2.86096424e-01 1.70593455e-01 -3.64240378e-01 -1.79686114e-01
5.98634899e-01 1.24617517e-01 -3.33212912e-01 -9.12102938e-01
-9.08913136e-01 1.03163052e+00 5.11360228e-01 3.78006399e-01
-1.06125616e-01 -2.38432795e-01 3.00836205e-01 4.35511887e-01
5.47713935e-01 4.50655639e-01 -5.39282441e-01 7.76451230e-02
-7.24209785e-01 1.95477739e-01 5.17421901e-01 1.02571344e+00
7.97481656e-01 -3.91389728e-01 2.52976455e-02 9.96670723e-01
3.85471463e-01 1.47647560e-01 6.53432190e-01 -8.50745201e-01
6.81070864e-01 7.47281969e-01 -1.27652064e-01 -8.07325900e-01
-3.56322885e-01 -5.39885640e-01 -7.64735103e-01 1.40739515e-01
4.94782865e-01 1.31016178e-02 -8.60256016e-01 1.81906557e+00
3.71310502e-01 1.58397660e-01 1.84628054e-01 8.72563183e-01
7.54390299e-01 2.77416468e-01 8.91775172e-03 1.75701723e-01
1.07864201e+00 -7.12436974e-01 -5.36319911e-01 -1.39457032e-01
9.45086241e-01 -5.47721684e-01 1.11996865e+00 2.77193516e-01
-5.22223532e-01 -6.19825721e-01 -1.21822715e+00 -1.72489267e-02
-5.23081243e-01 1.95592642e-01 4.23916817e-01 6.72876298e-01
-7.45917201e-01 6.17244661e-01 -5.55747807e-01 -4.00702178e-01
7.83571601e-01 4.22757953e-01 -4.26607549e-01 -1.60756811e-01
-1.26362634e+00 7.78386176e-01 6.22524023e-01 -3.00999343e-01
-5.28613389e-01 -1.05652392e+00 -7.12398708e-01 7.74046034e-02
9.52447206e-02 -5.21776080e-01 9.40790296e-01 -9.40502226e-01
-1.13871121e+00 8.89450908e-01 -4.65758890e-02 -3.92874658e-01
4.60959941e-01 -8.80321264e-02 -4.40424114e-01 9.56291333e-02
1.57017291e-01 7.99928010e-01 7.12735116e-01 -1.08767164e+00
-8.18127692e-01 -5.87767005e-01 -1.24169797e-01 2.90381283e-01
-8.19705844e-01 -2.19477266e-01 -3.11575353e-01 -6.58455729e-01
2.34892964e-01 -8.92958224e-01 3.35538499e-02 1.66698992e-01
-1.14181556e-01 -5.58648288e-01 1.03830755e+00 -3.42186153e-01
1.13485789e+00 -2.74058414e+00 6.21086322e-02 2.23021656e-01
2.61687607e-01 2.52589174e-02 -1.17820896e-01 7.08572865e-02
-4.11651045e-01 6.42235056e-02 -2.73520648e-01 -2.45731652e-01
1.16924442e-01 1.96395323e-01 -5.82902372e-01 6.54669404e-01
1.67801276e-01 5.09853840e-01 -6.17847860e-01 -4.13326174e-01
3.27677913e-02 3.41476202e-01 -7.89677978e-01 2.11277120e-02
4.00192942e-03 4.69741881e-01 -5.53236842e-01 5.82476854e-01
1.08257508e+00 -2.76531488e-01 2.29174662e-02 -2.18845874e-01
2.69782484e-01 3.32560956e-01 -1.22188830e+00 1.71940947e+00
-1.74530894e-01 5.12886345e-01 -2.44567946e-01 -1.44473696e+00
9.60230350e-01 5.85782528e-02 7.42478549e-01 -6.69851542e-01
4.29793522e-02 3.41938198e-01 2.07532674e-01 -2.44611412e-01
2.82159775e-01 -1.40120283e-01 -1.50732640e-02 3.05774868e-01
2.03123599e-01 6.65986463e-02 -9.38997567e-02 2.99294461e-02
7.65148878e-01 -2.17142120e-01 7.92420730e-02 -4.09379393e-01
4.05238748e-01 -1.25256121e-01 8.09283197e-01 4.38490063e-01
-4.58360761e-01 8.38164210e-01 5.21237552e-01 -1.51244864e-01
-9.68411386e-01 -1.28766549e+00 -6.63545132e-01 8.70044947e-01
4.69934493e-01 -2.33228683e-01 -5.40634215e-01 -1.01176965e+00
2.53730118e-01 7.95133591e-01 -5.51614225e-01 -5.29447496e-01
-4.56993401e-01 -7.01732337e-01 4.03516501e-01 6.92426741e-01
6.40214205e-01 -7.83099473e-01 -2.50122607e-01 8.03598315e-02
-7.84893557e-02 -1.14558244e+00 -3.98625135e-01 2.04188198e-01
-9.48397696e-01 -1.01492381e+00 -7.04604506e-01 -8.50841224e-01
7.23860621e-01 2.49195874e-01 8.44171166e-01 -1.03629760e-01
-2.38471583e-01 3.56363058e-01 -4.68081892e-01 -5.74622214e-01
-1.79080695e-01 3.23743403e-01 2.36536846e-01 3.70768085e-03
8.19804549e-01 -6.15416884e-01 -6.55089974e-01 5.86657286e-01
-1.09519386e+00 -3.64772975e-01 4.28112894e-01 9.00704086e-01
3.99345696e-01 4.18350399e-01 7.07923949e-01 -6.21361256e-01
5.28477252e-01 -7.25115061e-01 -4.38807398e-01 2.63414085e-01
-7.38388598e-01 9.15765837e-02 6.54324770e-01 -5.47537506e-01
-8.92734587e-01 -3.83833870e-02 1.13058195e-01 -5.32342255e-01
-2.43575096e-01 1.32244498e-01 -5.83578050e-01 3.67625058e-02
6.85075879e-01 1.93482667e-01 1.36043802e-01 -5.06285310e-01
1.84567049e-01 8.95414829e-01 -3.38402167e-02 -7.49479890e-01
8.76554608e-01 6.72523320e-01 -1.87180787e-01 -7.77822495e-01
-6.41912341e-01 -4.60198909e-01 -6.60784543e-01 2.20895186e-01
6.67876124e-01 -9.22225356e-01 -1.61408871e-01 3.85948688e-01
-9.74838138e-01 -8.38676989e-02 -3.57109845e-01 6.60936654e-01
-5.36715090e-01 4.12035465e-01 -8.09997320e-02 -3.52688313e-01
4.57521118e-02 -1.26405287e+00 8.41029823e-01 2.43709713e-01
-1.67187348e-01 -1.12102985e+00 -1.15967989e-01 1.14251383e-01
2.30831638e-01 -1.01255342e-01 1.17476189e+00 -9.18615699e-01
-4.16289717e-01 1.67432889e-01 -3.27805370e-01 7.09175587e-01
3.96053314e-01 -3.99294168e-01 -1.00306880e+00 -4.47042972e-01
1.71527892e-01 -2.09700540e-01 8.86835873e-01 4.70801204e-01
1.61635315e+00 1.52561301e-02 -5.73663116e-01 8.86295855e-01
1.24418187e+00 3.72474968e-01 6.60545588e-01 6.66394830e-01
4.40394878e-01 7.98564970e-01 8.41605604e-01 5.16247153e-01
2.74824142e-01 7.07921565e-01 3.73697668e-01 -1.61945075e-01
-1.21683888e-01 -1.49485320e-01 9.67152789e-02 1.74771741e-01
5.14500558e-01 -5.47099560e-02 -8.52664471e-01 6.91328108e-01
-1.75462842e+00 -7.10805833e-01 2.92680830e-01 2.12108278e+00
7.22979009e-01 -2.24074218e-02 8.80655423e-02 -4.74177767e-03
7.45908678e-01 -1.51845079e-03 -8.92116189e-01 -1.68459788e-01
-2.04094842e-01 1.09431878e-01 4.70977008e-01 2.22943887e-01
-1.23649502e+00 7.94825256e-01 5.34839725e+00 8.84923041e-01
-1.16654825e+00 9.54773929e-03 6.20422244e-01 -3.30149770e-01
-3.26713771e-01 -2.38594294e-01 -1.06939316e+00 7.51490533e-01
4.76901621e-01 -2.58045167e-01 2.43157640e-01 1.12375498e+00
-1.28429383e-01 2.91618884e-01 -1.37922001e+00 1.08972275e+00
-4.83266115e-02 -1.13390207e+00 -3.46121788e-02 3.26930225e-01
6.29789948e-01 8.80531445e-02 5.45961440e-01 4.50920463e-01
-5.83497994e-02 -9.82808411e-01 5.76395273e-01 -1.42537942e-02
7.68764973e-01 -7.06246853e-01 6.07485473e-01 2.22007066e-01
-8.31912100e-01 -4.22793627e-01 -7.55165398e-01 2.17929780e-01
-2.45149061e-01 8.12445998e-01 -8.97177339e-01 4.51480955e-01
1.01070070e+00 8.08756411e-01 -5.18448770e-01 9.81196523e-01
-9.96679068e-02 4.46805537e-01 -3.45643759e-01 2.21537843e-01
1.79072127e-01 -2.02492163e-01 3.75916958e-01 8.50189328e-01
5.29741704e-01 -6.45505637e-02 -1.09190471e-01 1.01809692e+00
-7.30771124e-02 -1.38334306e-02 -7.44196415e-01 1.65806770e-01
6.73390806e-01 8.65190387e-01 -3.68438780e-01 -8.28342512e-02
-5.25753736e-01 8.82049084e-01 3.13280016e-01 5.03437102e-01
-1.07275200e+00 -4.97245997e-01 1.19539404e+00 2.43263572e-01
5.53161025e-01 -1.86877996e-01 -6.44457698e-01 -1.08394468e+00
4.24831361e-01 -8.36523771e-01 5.60533106e-01 -1.43542022e-01
-1.56551027e+00 4.24620867e-01 1.66431352e-01 -1.50263739e+00
2.23944113e-02 -8.45565557e-01 -3.80255103e-01 8.91233444e-01
-1.69948804e+00 -7.53196120e-01 -1.48911178e-01 7.59538412e-01
3.73901755e-01 -4.82082754e-01 5.44240475e-01 6.00796759e-01
-4.24448788e-01 1.02924311e+00 5.48048735e-01 1.67058185e-01
1.07017314e+00 -9.97504115e-01 1.97422020e-02 5.86960196e-01
-4.33686189e-02 6.26013517e-01 4.70273972e-01 -3.85301203e-01
-8.69217157e-01 -1.15894020e+00 6.26939774e-01 -3.71987045e-01
5.84703028e-01 -5.36346674e-01 -1.05832553e+00 5.43389797e-01
-1.24986619e-01 8.68789405e-02 9.93790209e-01 1.82469875e-01
-4.52237666e-01 -4.42134857e-01 -1.36291409e+00 4.39773500e-01
1.27158701e+00 -3.85531932e-01 -6.58420861e-01 3.23128611e-01
6.51790321e-01 -2.16240138e-01 -8.14610004e-01 6.51814044e-01
4.03591305e-01 -8.89643848e-01 1.03609633e+00 -6.92234516e-01
3.24426889e-01 -2.85154432e-01 -4.34780091e-01 -1.49080515e+00
-4.76957530e-01 2.80345958e-02 8.88412148e-02 1.28238392e+00
3.62071276e-01 -9.58067238e-01 9.82974589e-01 5.77999711e-01
-3.34878415e-02 -7.77052343e-01 -1.22208035e+00 -1.13470876e+00
6.76735640e-01 -2.31112584e-01 7.67794430e-01 1.00866354e+00
-2.33781546e-01 9.96580422e-02 9.04535949e-02 3.13401282e-01
6.50934815e-01 1.67987585e-01 5.00467002e-01 -1.37204397e+00
-1.36510476e-01 -5.37255883e-01 -5.38992345e-01 -1.16786206e+00
4.44840044e-01 -9.78112340e-01 -3.15717906e-01 -1.26691926e+00
2.01815575e-01 -7.70622075e-01 -6.08189404e-01 3.76198262e-01
6.25641271e-02 -1.94508970e-01 1.48112550e-01 2.82853603e-01
-4.74256843e-01 9.01918948e-01 1.16946495e+00 -1.86141133e-01
-2.05338001e-01 -2.17073653e-02 -1.08439171e+00 4.59691316e-01
1.19113171e+00 -6.85889959e-01 -6.55898750e-01 -4.67288971e-01
-1.93876222e-01 -6.76658452e-01 3.50004017e-01 -9.49917257e-01
1.45204797e-01 -2.92763934e-02 5.34879506e-01 -5.82560956e-01
2.57808238e-01 -1.11405838e+00 -3.26546222e-01 3.35147917e-01
-1.90771654e-01 -1.96413979e-01 2.70536959e-01 4.32403415e-01
-3.60539854e-01 -2.28721142e-01 1.03342843e+00 1.93402886e-01
-9.21809673e-01 4.10000533e-01 3.81374583e-02 1.74905300e-01
1.18186212e+00 -3.92897666e-01 -4.51613754e-01 -1.25718728e-01
-5.42997301e-01 3.69317472e-01 5.16231716e-01 8.03491533e-01
5.97261071e-01 -1.57531369e+00 -6.13589287e-01 4.19889897e-01
3.29590380e-01 7.65540376e-02 3.27190638e-01 6.74095035e-01
-1.89273179e-01 4.59325165e-01 -1.73329532e-01 -8.22198153e-01
-1.05691195e+00 5.39802969e-01 4.31479394e-01 -7.54138827e-02
-5.03361166e-01 8.39763105e-01 7.54608035e-01 -6.95463538e-01
5.57455301e-01 -3.77047747e-01 -3.97271477e-02 7.96151459e-02
5.02697289e-01 2.90777266e-01 1.69418544e-01 -1.74454644e-01
-6.72821999e-01 6.76602781e-01 -4.62150782e-01 1.06246591e-01
1.34110928e+00 -5.44169284e-02 1.38548613e-01 3.69631462e-02
1.69762230e+00 -3.30303937e-01 -1.37858534e+00 -4.10959989e-01
-1.72944851e-02 -5.30614674e-01 -5.54048829e-02 -6.93433166e-01
-9.32448983e-01 9.08891797e-01 9.90306318e-01 2.62014717e-02
1.29299176e+00 3.44662189e-01 6.50037885e-01 1.93171605e-01
1.69000447e-01 -1.43412685e+00 2.71358311e-01 4.94984478e-01
9.01171505e-01 -1.36845744e+00 -1.25328302e-01 -3.17230791e-01
-7.11091101e-01 1.00680220e+00 9.13014829e-01 -9.99018624e-02
9.27924991e-01 3.21097672e-02 9.95705649e-02 -4.03228812e-02
-3.93688440e-01 4.04193215e-02 2.66486228e-01 7.61202216e-01
2.64346480e-01 -7.96662793e-02 -2.73671031e-01 7.23364115e-01
-3.17088038e-01 -1.70971498e-01 1.80952802e-01 5.97287476e-01
-4.30091172e-01 -1.29361379e+00 -3.37201804e-01 2.65338629e-01
-2.67257214e-01 5.42438552e-02 -3.45151663e-01 8.96996439e-01
3.25058281e-01 8.39919806e-01 2.16075867e-01 -4.04412955e-01
3.51565927e-01 -3.47376838e-02 4.62156147e-01 -6.03398144e-01
3.32892872e-02 -3.22102100e-01 -3.53274971e-01 -4.93828803e-01
-5.47072738e-02 -6.32946610e-01 -1.36799979e+00 -8.49484205e-02
-1.72411412e-01 1.23494707e-01 6.64251506e-01 7.26631701e-01
6.53740704e-01 4.18472528e-01 8.58247280e-01 -3.42742294e-01
-1.04458404e+00 -7.78340638e-01 -8.48201275e-01 6.01676345e-01
2.50944942e-01 -1.08521938e+00 -7.01257229e-01 -3.67572397e-01] | [10.201225280761719, 3.1018683910369873] |
71febecf-a6c5-45c2-8abb-431ffef2b22e | learning-to-reach-goals-without-reinforcement-1 | 1912.06088 | null | https://arxiv.org/abs/1912.06088v4 | https://arxiv.org/pdf/1912.06088v4.pdf | Learning to Reach Goals via Iterated Supervised Learning | Current reinforcement learning (RL) algorithms can be brittle and difficult to use, especially when learning goal-reaching behaviors from sparse rewards. Although supervised imitation learning provides a simple and stable alternative, it requires access to demonstrations from a human supervisor. In this paper, we study RL algorithms that use imitation learning to acquire goal reaching policies from scratch, without the need for expert demonstrations or a value function. In lieu of demonstrations, we leverage the property that any trajectory is a successful demonstration for reaching the final state in that same trajectory. We propose a simple algorithm in which an agent continually relabels and imitates the trajectories it generates to progressively learn goal-reaching behaviors from scratch. Each iteration, the agent collects new trajectories using the latest policy, and maximizes the likelihood of the actions along these trajectories under the goal that was actually reached, so as to improve the policy. We formally show that this iterated supervised learning procedure optimizes a bound on the RL objective, derive performance bounds of the learned policy, and empirically demonstrate improved goal-reaching performance and robustness over current RL algorithms in several benchmark tasks. | ['Dibya Ghosh', 'Coline Devin', 'Sergey Levine', 'Benjamin Eysenbach', 'Ashwin Reddy', 'Abhishek Gupta', 'Justin Fu'] | 2019-12-12 | null | https://openreview.net/forum?id=rALA0Xo6yNJ | https://openreview.net/pdf?id=rALA0Xo6yNJ | iclr-2021-1 | ['multi-goal-reinforcement-learning'] | ['methodology'] | [ 1.13091469e-01 2.03752071e-01 -5.86648941e-01 -3.44660990e-02
-8.93574715e-01 -7.86088824e-01 6.45650804e-01 7.79547319e-02
-6.96298599e-01 1.23546410e+00 3.55356224e-02 -3.78213733e-01
-2.29656056e-01 -3.85392517e-01 -9.02032971e-01 -8.93838763e-01
-4.12401408e-01 4.41801578e-01 5.51901050e-02 -1.99192762e-02
3.92112732e-01 6.79472089e-01 -1.13451374e+00 -4.75075871e-01
9.09271061e-01 4.65021670e-01 4.02607352e-01 9.58241880e-01
3.19474429e-01 1.19637549e+00 -5.41678727e-01 4.90877569e-01
3.92128199e-01 -8.16626728e-01 -7.48251617e-01 3.47907990e-01
-1.42433926e-01 -8.23072731e-01 -6.28752232e-01 9.83996570e-01
1.65186152e-01 5.65775037e-01 7.37011313e-01 -1.42416704e+00
-1.87953860e-01 6.43830478e-01 -3.95593852e-01 -8.93478543e-02
4.54592496e-01 8.32690537e-01 8.29010963e-01 -3.01297694e-01
6.64283216e-01 1.12597191e+00 2.27552816e-01 7.12224901e-01
-1.10583949e+00 -5.44652760e-01 1.90492734e-01 -3.20311226e-02
-7.83303618e-01 -2.48315364e-01 4.16837662e-01 -2.54405588e-01
1.00801814e+00 -4.55323160e-01 6.93468392e-01 1.09398222e+00
2.31866673e-01 1.02881432e+00 1.18768775e+00 -3.52544755e-01
5.09520411e-01 -2.92932242e-01 -2.66819775e-01 1.06787324e+00
-1.19390912e-01 6.80413246e-01 -3.38292927e-01 -8.99712816e-02
1.05379283e+00 1.69275284e-01 -1.31950155e-01 -7.33932793e-01
-1.43608987e+00 7.80662894e-01 4.86555785e-01 1.24199793e-01
-7.83248246e-01 7.28067815e-01 1.33078828e-01 8.80222499e-01
-3.66415024e-01 6.74787521e-01 -3.14033031e-01 -5.99730432e-01
-6.29761100e-01 3.22575510e-01 7.59051979e-01 7.97705233e-01
8.62466276e-01 3.73705328e-01 -1.10374652e-01 3.16765845e-01
5.79837821e-02 8.56007040e-01 3.48370612e-01 -1.58430362e+00
4.02043819e-01 4.77966458e-01 6.69004321e-01 -2.32328206e-01
-2.61221230e-01 -8.31404328e-03 -2.18458369e-01 1.01300895e+00
5.66579103e-01 -5.32242417e-01 -8.01610649e-01 1.79690063e+00
3.26986551e-01 2.91948795e-01 3.73758346e-01 7.03720629e-01
-1.54794902e-01 7.33322263e-01 -2.96261162e-02 -4.77340072e-01
1.75310448e-01 -1.22485662e+00 -2.67342299e-01 -1.73243180e-01
8.70595515e-01 -1.74804926e-01 1.08730912e+00 3.95574778e-01
-1.08292592e+00 -3.74800891e-01 -9.93533313e-01 5.14758945e-01
2.40112334e-01 1.76070616e-01 1.86380759e-01 2.96732783e-03
-8.78264964e-01 1.31177819e+00 -1.20168602e+00 -2.25231573e-01
2.38455623e-01 4.29091215e-01 -1.76535577e-01 1.53876767e-01
-6.15332603e-01 1.05763316e+00 4.99784529e-01 -3.04413378e-01
-1.96941292e+00 -3.04969639e-01 -7.62989104e-01 -4.61348929e-02
8.74173760e-01 -3.69786471e-01 1.70488930e+00 -9.36221242e-01
-2.00110579e+00 2.81292677e-01 1.46957468e-02 -6.74082339e-01
5.83305478e-01 -2.95365274e-01 2.04460859e-01 2.71127164e-01
2.03958958e-01 6.56038880e-01 1.29442203e+00 -1.01255012e+00
-8.63197744e-01 -5.72519116e-02 3.34114075e-01 5.26695609e-01
7.81854894e-03 -4.15582448e-01 -9.59252007e-03 -7.26368427e-02
-3.89100343e-01 -1.34045863e+00 -4.53045219e-01 -6.75993739e-03
-2.37752885e-01 -5.29372394e-01 6.19311333e-01 -2.56248593e-01
7.95047343e-01 -2.12497401e+00 6.03505194e-01 2.35397235e-01
7.43665174e-02 2.11825892e-01 -2.96580613e-01 7.43922889e-01
4.41092700e-01 -3.95871490e-01 -2.26745531e-01 -4.98722121e-02
8.99705067e-02 4.12787348e-01 -5.46355069e-01 6.32210612e-01
-1.47973210e-01 1.00670791e+00 -1.48799634e+00 -2.71959245e-01
3.35520059e-01 -8.91697556e-02 -3.90412658e-01 5.81740916e-01
-5.96903205e-01 9.18219090e-01 -7.76806831e-01 2.28865594e-01
-2.61721492e-01 -2.89937347e-01 2.73451418e-01 6.69523716e-01
-1.89034194e-01 3.06616068e-01 -9.57968593e-01 1.72968841e+00
-5.46561837e-01 4.46989805e-01 1.52526289e-01 -1.01479030e+00
8.16451073e-01 1.60318390e-01 5.50049126e-01 -4.17280197e-01
4.42405008e-02 3.29810411e-01 3.18468995e-02 -5.54838359e-01
1.28007814e-01 -7.53730908e-02 -2.49362677e-01 9.98209774e-01
9.74693149e-02 -5.73240757e-01 2.84843445e-01 2.43441105e-01
1.30261207e+00 7.47008741e-01 5.42668045e-01 2.01491505e-01
3.51354063e-01 3.97213191e-01 2.03587875e-01 1.17030525e+00
-3.38721544e-01 -2.18685821e-01 5.19066155e-01 -1.00981914e-01
-1.11177778e+00 -1.31839883e+00 8.54978204e-01 1.14370751e+00
2.59387255e-01 -1.71191692e-02 -5.25711238e-01 -9.75433528e-01
1.18471503e-01 8.56258750e-01 -4.66691434e-01 -2.16559619e-01
-9.11299586e-01 3.50867003e-01 3.62286001e-01 4.97716695e-01
3.11196804e-01 -1.65649569e+00 -1.19858778e+00 4.39762145e-01
2.12526038e-01 -6.21908724e-01 -6.81775749e-01 2.33423024e-01
-1.01557481e+00 -1.31456459e+00 -7.59118855e-01 -7.99938083e-01
7.54922032e-01 1.85389593e-01 5.98615825e-01 3.62548716e-02
-3.20757814e-02 9.09627259e-01 -1.63501054e-01 -2.14191183e-01
-7.87991345e-01 -2.30060071e-02 3.12310547e-01 -5.30307889e-01
-1.44730493e-01 -5.41516423e-01 -4.32505935e-01 1.53971121e-01
-4.00317848e-01 -9.68269035e-02 6.94437623e-01 9.25116837e-01
5.76260090e-01 -5.57206310e-02 7.80677021e-01 -1.76840991e-01
9.83562052e-01 -3.93471569e-01 -8.85578275e-01 1.62011385e-01
-6.06535256e-01 5.91366827e-01 1.02855372e+00 -7.49040604e-01
-7.93421924e-01 3.18302214e-01 2.18493700e-01 -5.99092782e-01
-2.41184399e-01 2.84340054e-01 4.69755083e-01 1.82882756e-01
7.81682909e-01 7.27320790e-01 5.32153130e-01 -1.21635914e-01
6.01343155e-01 3.34904134e-01 5.14205694e-01 -8.14672470e-01
8.46129596e-01 2.62249410e-01 -1.39503842e-02 -6.60214901e-01
-6.70268357e-01 -3.14827859e-01 -5.00928819e-01 -5.42294383e-01
4.47600126e-01 -6.05306804e-01 -1.09392107e+00 2.23889515e-01
-6.35860860e-01 -1.21263397e+00 -6.48697674e-01 6.52012706e-01
-1.24213195e+00 3.64783913e-01 -5.14298081e-01 -9.78186429e-01
-8.23388770e-02 -1.16987884e+00 6.73124850e-01 4.25860345e-01
-2.61572301e-01 -8.23237121e-01 2.49671936e-01 -3.92020553e-01
1.64711013e-01 3.37197930e-01 5.98777235e-01 -2.87965447e-01
-6.78938448e-01 7.08782906e-03 2.89635450e-01 3.12688291e-01
2.95069158e-01 -3.54458570e-01 -1.36746049e-01 -7.62620807e-01
-1.41037881e-01 -1.01158488e+00 5.44363439e-01 3.76883894e-01
6.88797653e-01 -4.53098506e-01 -3.23809832e-01 1.87003449e-01
1.20112693e+00 4.67012346e-01 2.02664420e-01 4.25695658e-01
3.71322423e-01 3.92503738e-01 9.91312921e-01 4.98493493e-01
1.07298046e-03 2.64102340e-01 4.67812717e-01 4.02704239e-01
9.40433890e-02 -7.02812552e-01 8.82692397e-01 4.20913011e-01
-7.62896538e-02 2.30643615e-01 -4.61113483e-01 6.21116519e-01
-2.05998111e+00 -1.00740445e+00 5.39744020e-01 2.27163601e+00
8.70198846e-01 1.95066229e-01 5.25260270e-01 -1.47812888e-01
2.43632391e-01 1.41093479e-02 -1.36198902e+00 -1.93526864e-01
4.77430403e-01 1.68741301e-01 4.96962547e-01 7.34522283e-01
-6.75165474e-01 1.23235619e+00 6.57153559e+00 5.33757567e-01
-1.04690516e+00 -2.21472308e-01 1.35944616e-02 -3.65677357e-01
2.27687880e-01 1.17346719e-01 -6.05485976e-01 2.10771292e-01
8.00591350e-01 -4.70764101e-01 1.20302308e+00 1.07784784e+00
5.00263691e-01 -3.23953569e-01 -1.36235547e+00 6.02289855e-01
-4.07833308e-01 -1.04028809e+00 -3.78597319e-01 2.76594125e-02
8.22974682e-01 1.69846848e-01 2.08436996e-02 6.97999954e-01
1.11195624e+00 -1.00720632e+00 4.56377506e-01 3.32175285e-01
5.82667172e-01 -8.35538805e-01 -4.08999622e-02 9.93787825e-01
-8.45348716e-01 -4.33224857e-01 -1.80713326e-01 -3.05773020e-01
1.71365395e-01 -3.16094071e-01 -1.16186774e+00 1.79947942e-01
1.22225337e-01 7.64177322e-01 -7.73310894e-03 8.27364266e-01
-9.20498371e-01 6.43866956e-01 -4.11694109e-01 -7.23790526e-01
8.84325862e-01 -3.50290060e-01 6.76272213e-01 7.47409821e-01
2.89189011e-01 -3.27811837e-02 7.00586736e-01 7.53123164e-01
8.32769349e-02 -2.10999116e-01 -9.77457225e-01 -2.99314290e-01
4.57029432e-01 9.54388499e-01 -5.49267948e-01 -5.17756581e-01
1.44457817e-02 9.73304749e-01 7.96620071e-01 5.00151038e-01
-7.73712814e-01 -2.46555910e-01 6.20332837e-01 -3.20563346e-01
3.97341758e-01 -6.65718675e-01 3.53624910e-01 -8.29837322e-01
-1.94844484e-01 -9.86377537e-01 1.77877218e-01 -7.52773464e-01
-8.14229131e-01 1.19928010e-01 -2.69638989e-02 -1.31782091e+00
-1.00722945e+00 -2.54109770e-01 -5.77547729e-01 5.03342628e-01
-1.41037261e+00 -5.20574629e-01 6.27896041e-02 7.64471471e-01
6.82900488e-01 -2.96505630e-01 5.81006944e-01 -5.52508295e-01
-2.12451458e-01 2.08485007e-01 3.86032015e-01 -1.53350774e-02
4.76194590e-01 -1.41854918e+00 6.35860208e-03 5.92523694e-01
1.26101226e-01 4.89770621e-01 7.84874439e-01 -7.54311144e-01
-1.57983315e+00 -7.71568298e-01 5.94864674e-02 -5.35418466e-02
1.03501749e+00 2.20113620e-01 -5.47341049e-01 9.76305008e-01
1.50112078e-01 -3.52677673e-01 -9.50088575e-02 -3.29779059e-01
-2.07224376e-02 1.98400885e-01 -9.14419353e-01 1.03206027e+00
8.33359659e-01 -2.83586979e-01 -7.61375546e-01 3.19719046e-01
5.73346555e-01 -4.54972208e-01 -6.27383232e-01 6.65034447e-03
4.26855534e-01 -6.22713447e-01 7.70953238e-01 -8.85189474e-01
3.55532467e-01 -3.39308053e-01 3.32879603e-01 -1.75083268e+00
-8.15703049e-02 -1.07888877e+00 -4.24081087e-01 3.09598148e-01
3.33753675e-01 -5.75055122e-01 8.01547050e-01 1.41491205e-01
-6.22531064e-02 -8.78330886e-01 -8.95568371e-01 -1.24907804e+00
2.25936040e-01 -9.12586600e-02 2.39062548e-01 4.46034461e-01
4.51292187e-01 3.70722443e-01 -6.68452263e-01 3.46840955e-02
8.50777805e-01 4.11295563e-01 1.23772371e+00 -5.68342090e-01
-5.08141756e-01 -3.50994736e-01 2.47094870e-01 -1.72016633e+00
5.08606374e-01 -8.43397260e-01 4.88703966e-01 -1.57926083e+00
-7.22841024e-02 -3.87334317e-01 -2.03455105e-01 6.83343351e-01
-7.55695701e-02 -5.61039507e-01 4.30800378e-01 4.41080630e-01
-9.27149951e-01 6.59115970e-01 1.67554605e+00 -5.23177721e-02
-7.36150444e-01 1.44279242e-01 -2.45485142e-01 6.92931116e-01
1.07050288e+00 -6.93047345e-01 -6.58059120e-01 -1.50152475e-01
-1.55262142e-01 6.77525103e-01 2.79866099e-01 -8.73624682e-01
3.55623692e-01 -5.46598554e-01 2.50514708e-02 -3.75933886e-01
2.57277995e-01 -6.27014458e-01 -2.82565862e-01 1.09620905e+00
-7.81005740e-01 -3.54315713e-02 -1.82124034e-01 1.00234413e+00
2.38477245e-01 -5.33432007e-01 7.81045318e-01 -2.69081831e-01
-7.17538774e-01 2.63313442e-01 -8.32792938e-01 5.11310510e-02
1.38543212e+00 -1.06756045e-02 1.11373570e-02 -7.74397314e-01
-7.90847659e-01 7.40572929e-01 6.07469559e-01 2.20921963e-01
7.07583010e-01 -1.08987737e+00 -4.91020888e-01 -2.09947571e-01
-1.77055016e-01 -2.63509929e-01 -2.22313926e-01 6.79674983e-01
-1.01356559e-01 2.56741673e-01 -3.00555915e-01 -4.44725037e-01
-1.06392527e+00 7.46107757e-01 3.79287720e-01 -5.62654197e-01
-1.03316343e+00 2.85296768e-01 -2.14443773e-01 -4.73601103e-01
4.62522358e-01 -3.93121719e-01 -1.26995891e-01 -5.82101524e-01
3.39299142e-01 4.89319563e-01 -7.59906828e-01 -5.67068122e-02
7.67676309e-02 3.33809882e-01 2.79508028e-02 -6.12625182e-01
1.17861998e+00 -5.66655919e-02 3.03489834e-01 5.65908968e-01
9.48876381e-01 -3.02195340e-01 -2.13159490e+00 -3.28417093e-01
8.52977782e-02 -2.47371271e-01 -4.76224422e-01 -7.37234950e-01
-4.98638332e-01 4.77860510e-01 1.58648595e-01 7.74068311e-02
6.53708935e-01 -6.26529232e-02 7.89773643e-01 1.04938471e+00
8.80628467e-01 -1.33929980e+00 7.83535004e-01 6.59237385e-01
7.64558494e-01 -1.12082446e+00 -3.36147845e-02 2.77638823e-01
-9.01085734e-01 1.07312763e+00 5.25474489e-01 -6.58281207e-01
2.58284479e-01 1.61235407e-01 -2.28588909e-01 7.35354871e-02
-7.38887787e-01 -3.35774094e-01 -2.95207232e-01 7.63207853e-01
-2.73390710e-01 1.61191300e-01 -1.14741579e-01 -3.03503811e-01
-6.50640056e-02 2.32584238e-01 6.42163277e-01 1.28739870e+00
-1.01686966e+00 -1.04432237e+00 -1.87647030e-01 4.19311255e-01
-6.99016377e-02 4.19981182e-01 -2.72212267e-01 7.68763721e-01
-5.83579719e-01 8.59253049e-01 -2.96236426e-01 -9.17077065e-02
1.56838819e-01 -6.41303137e-02 9.22759593e-01 -7.37995803e-01
-3.60904127e-01 -4.42441814e-02 2.67768167e-02 -7.87962794e-01
-3.07771891e-01 -7.14941025e-01 -1.83817494e+00 -1.35521993e-01
2.22268701e-02 3.01940292e-01 5.40996611e-01 1.16759384e+00
1.80230990e-01 4.25435066e-01 9.20211792e-01 -8.75104606e-01
-1.30970156e+00 -8.52672458e-01 -4.52973425e-01 3.20054382e-01
6.21284664e-01 -7.39629388e-01 -4.17896479e-01 -4.18140031e-02] | [4.239220142364502, 1.5589128732681274] |
f92bb73f-8ad2-4cb0-9161-fa4cddcdc2e7 | improve-sinhala-speech-recognition-through | null | null | https://aclanthology.org/2021.icon-main.26 | https://aclanthology.org/2021.icon-main.26.pdf | Improve Sinhala Speech Recognition Through e2e LF-MMI Model | Automatic speech recognition (ASR) has experienced several paradigm shifts over the years from template-based approaches and statistical modeling to the popular GMM-HMM approach and then to deep learning hybrid model DNN-HMM. The latest shift is to end-to-end (e2e) DNN architecture. We present a study to build an e2e ASR system using state-of-the-art deep learning models to verify the applicability of e2e ASR models for the highly inflected and yet low-resource Sinhala language. We experimented on e2e Lattice-Free Maximum Mutual Information (e2e LF-MMI) model with the baseline statistical models with 40 hours of training data to evaluate. We used the same corpus for creating language models and lexicon in our previous study, which resulted in the best accuracy for the Sinhala language. We were able to achieve a Word-error-rate (WER) of 28.55% for Sinhala, only slightly worse than the existing best hybrid model. Our model, however, is more context-independent and faster for Sinhala speech recognition and so more suitable for general purpose speech-to-text translation. | ['Ruwan Weerasinghe', 'Thilini Nadungodage', 'Randil Pushpananda', 'Buddhi Gamage'] | null | null | null | null | icon-2021-12 | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 1.01767238e-02 2.54746705e-01 1.82120547e-01 -4.17942852e-01
-1.35156369e+00 -3.20490927e-01 6.63227499e-01 -4.03353542e-01
-7.29170024e-01 6.46169901e-01 5.45402110e-01 -9.29793894e-01
4.46837544e-01 -3.78313690e-01 -4.18493181e-01 -4.41503823e-01
3.41648251e-01 9.23733711e-01 1.66384473e-01 -5.83831549e-01
-1.54204980e-01 4.10855561e-01 -8.54758978e-01 6.06331110e-01
3.70134532e-01 3.57257247e-01 5.99321723e-01 9.49391544e-01
-1.15100250e-01 6.66306436e-01 -8.62866521e-01 -4.37410474e-01
6.80768639e-02 -2.41552621e-01 -8.55556846e-01 -2.48709053e-01
2.82013834e-01 -1.56912237e-01 -6.01006091e-01 6.01867914e-01
1.05452418e+00 2.02104032e-01 5.36423802e-01 -4.26595330e-01
-8.16873014e-01 8.70531499e-01 -1.44671798e-02 3.93520623e-01
3.03602755e-01 -1.15786204e-02 8.25127721e-01 -1.19860888e+00
6.52461648e-01 1.45247614e+00 5.28908432e-01 1.04296744e+00
-1.13690376e+00 -6.25333786e-01 -1.65632889e-01 3.46766740e-01
-1.66434979e+00 -1.15006101e+00 3.93624753e-01 3.83887030e-02
2.02200818e+00 2.72362977e-01 3.40590417e-01 1.45791590e+00
-9.02422220e-02 1.04357970e+00 1.34047329e+00 -8.79918039e-01
2.01017201e-01 2.52248406e-01 -2.31989026e-01 2.90179908e-01
-4.18657362e-01 7.29585364e-02 -6.35320008e-01 2.43140355e-01
6.30321681e-01 -5.97309053e-01 1.00594752e-01 6.69755757e-01
-1.32399976e+00 7.84141302e-01 -2.84595042e-01 8.17756414e-01
-4.05620545e-01 -1.19211681e-01 4.79196012e-01 5.39390862e-01
5.64220607e-01 1.28292650e-01 -1.03095531e+00 -5.43073833e-01
-1.14103854e+00 -9.06692892e-02 9.12961841e-01 8.59486938e-01
1.56847268e-01 7.40286291e-01 -1.74499959e-01 1.51132071e+00
4.07072783e-01 9.09913242e-01 8.35708201e-01 -5.17043471e-01
7.93160677e-01 -1.80035502e-01 -3.48938435e-01 -2.24174157e-01
-9.71165001e-02 -4.13604051e-01 -6.76759601e-01 -2.36780092e-01
2.78194636e-01 -3.30111265e-01 -1.25616550e+00 1.65094495e+00
-2.22559825e-01 -2.10998014e-01 4.77808088e-01 4.62539256e-01
8.57794344e-01 1.16178763e+00 -1.17091797e-02 -2.73778915e-01
1.08260989e+00 -8.85711074e-01 -9.00642753e-01 -4.48237658e-01
1.03018475e+00 -1.16445506e+00 1.18188381e+00 3.05419922e-01
-1.27437913e+00 -4.31108207e-01 -8.53094935e-01 -1.03806235e-01
-4.54050332e-01 3.62131119e-01 -2.59083640e-02 1.20821214e+00
-1.63722301e+00 1.10367276e-01 -8.95879507e-01 -6.80543125e-01
-2.12683856e-01 3.70620608e-01 -4.10187840e-01 7.20447376e-02
-1.52861500e+00 1.38000965e+00 3.79796863e-01 -3.21146264e-03
-6.94337606e-01 -1.60904303e-01 -7.10212052e-01 -1.78680435e-01
4.68811840e-02 -1.65559649e-01 1.62224936e+00 -6.61566854e-01
-2.23631954e+00 9.91915941e-01 -6.00945294e-01 -5.18812895e-01
1.75453663e-01 -1.59601271e-01 -7.99343884e-01 -5.76395988e-01
-4.80000943e-01 5.91222584e-01 2.53802299e-01 -9.17120934e-01
-4.98756826e-01 -3.98998529e-01 -6.40216172e-01 1.48082957e-01
-8.59525055e-02 8.31373394e-01 -1.19552374e-01 -6.99248254e-01
3.73493619e-02 -1.02877963e+00 -5.76929450e-02 -1.12332845e+00
-1.59767941e-01 -4.83502179e-01 5.45545220e-01 -1.44902015e+00
1.47500789e+00 -1.72107589e+00 -1.08886749e-01 -3.81989256e-02
-6.17058516e-01 8.44931066e-01 -3.59975010e-01 7.46251047e-01
-1.72172010e-01 1.71348199e-01 -1.14681765e-01 -6.94182873e-01
2.53950637e-02 4.56770986e-01 -3.17470253e-01 2.09205538e-01
8.67482871e-02 9.70069110e-01 -5.86694479e-01 -2.15540901e-01
3.87466252e-01 7.13689506e-01 -7.99427405e-02 2.99700230e-01
1.08225346e-01 1.87884524e-01 1.90370038e-01 5.39042532e-01
3.03498566e-01 7.22013295e-01 2.12139934e-01 4.59869832e-01
-2.24921271e-01 1.27634239e+00 -7.62249708e-01 1.50060427e+00
-1.07303476e+00 8.53741646e-01 1.49133503e-01 -9.31428015e-01
1.33758688e+00 8.95864308e-01 1.78561863e-02 -9.63700294e-01
1.51340440e-02 8.85712087e-01 3.61870021e-01 -9.40648392e-02
4.69984800e-01 -2.57917136e-01 7.72108659e-02 1.68657735e-01
2.78696328e-01 -3.81994039e-01 -2.81027436e-01 -2.89062828e-01
1.08774364e+00 -8.39405656e-02 4.64051396e-01 -2.65308201e-01
5.43746233e-01 -3.03489685e-01 4.13780063e-01 7.15299010e-01
-1.40874878e-01 8.28449667e-01 -1.60007328e-01 -2.28520975e-01
-1.37914658e+00 -9.80644882e-01 -6.11523204e-02 1.09205043e+00
-9.21760857e-01 -3.07951629e-01 -1.06826770e+00 -4.55724686e-01
-7.75707960e-01 1.60347509e+00 1.77339315e-01 1.70320317e-01
-1.27213287e+00 -6.91515923e-01 1.08877742e+00 3.29316139e-01
3.30633223e-01 -1.43071353e+00 3.76983523e-01 7.40157485e-01
-3.83315474e-01 -1.36573541e+00 -4.20405090e-01 2.34896615e-01
-7.27769315e-01 -1.08703496e-02 -9.44312155e-01 -1.00499725e+00
-1.73121631e-01 -1.68620721e-01 1.18889570e+00 -4.32711333e-01
2.65003771e-01 5.94085455e-02 -3.94783139e-01 -5.98612726e-01
-1.36483598e+00 3.19107026e-01 2.51974940e-01 -3.78867745e-01
8.64337862e-01 -4.78517830e-01 2.07840335e-02 2.87772149e-01
-5.88361144e-01 -1.52747616e-01 8.14829946e-01 6.98461115e-01
3.45407248e-01 -5.45782804e-01 8.03596318e-01 -5.55461407e-01
6.20583534e-01 -2.25485265e-01 -3.20577145e-01 2.48506024e-01
-5.06830513e-01 -4.53339070e-02 6.58495665e-01 -3.84667754e-01
-1.12853444e+00 9.27289277e-02 -1.25456309e+00 -2.57467246e-03
-5.04189014e-01 3.84350061e-01 -4.72978145e-01 4.88052696e-01
5.68477154e-01 7.43786275e-01 -3.24308395e-01 -7.17185318e-01
4.13907439e-01 1.51839948e+00 3.54562998e-01 -2.62718171e-01
3.49438459e-01 -2.30195835e-01 -5.49351931e-01 -1.45834005e+00
-3.80195260e-01 -5.82621574e-01 -7.04830706e-01 5.65259717e-02
8.68721128e-01 -8.66822660e-01 -2.25524172e-01 8.79048645e-01
-1.76197016e+00 -6.48921788e-01 -1.77167857e-03 6.52715921e-01
-6.32594824e-01 2.07841024e-01 -8.61853898e-01 -1.04340768e+00
-5.98933697e-01 -1.17537189e+00 1.19839489e+00 -5.47421336e-01
-3.35789710e-01 -1.19878352e+00 4.16620314e-01 5.19564509e-01
8.42602730e-01 -7.07275033e-01 8.54803562e-01 -1.26574504e+00
-7.38408491e-02 -9.05124944e-06 5.36503494e-02 8.68291974e-01
2.43845563e-02 -3.04686993e-01 -1.11652827e+00 -2.62082875e-01
1.54621199e-01 -1.29441887e-01 5.61325669e-01 3.34866941e-01
5.93816459e-01 -5.45054078e-01 1.14662126e-01 1.85419068e-01
9.88558650e-01 6.09374881e-01 1.07624185e+00 3.16166908e-01
5.93288124e-01 4.41251069e-01 2.03691840e-01 -2.21774653e-01
5.28372288e-01 1.17084062e+00 -3.41759294e-01 -1.43586442e-01
-5.06370485e-01 -1.88805595e-01 1.30266023e+00 1.94136429e+00
1.88938349e-01 -6.46846056e-01 -1.30786467e+00 7.96648383e-01
-1.51174355e+00 -7.50687540e-01 -1.00913249e-01 2.22555709e+00
1.15543807e+00 -3.03397042e-04 9.03613046e-02 -8.19411278e-02
7.12987483e-01 1.93122789e-01 2.38934040e-01 -1.17191756e+00
-3.57935905e-01 5.04373193e-01 4.82557803e-01 8.29129279e-01
-7.03060865e-01 1.62708223e+00 6.55145979e+00 1.47606337e+00
-1.21517479e+00 6.60610318e-01 5.65484524e-01 8.51671621e-02
-4.75864597e-02 -1.48685783e-01 -1.31872678e+00 2.44394422e-01
2.12081218e+00 1.94468156e-01 6.21074855e-01 7.35110223e-01
4.70882535e-01 3.92007738e-01 -1.00013137e+00 1.05140519e+00
9.18868780e-02 -1.08152628e+00 3.31797041e-02 3.79749872e-02
6.21240377e-01 8.73525143e-01 -2.46430054e-01 6.22870982e-01
4.63828176e-01 -1.14195931e+00 7.61780262e-01 6.25206307e-02
1.04025829e+00 -6.61680996e-01 9.66088653e-01 5.65150619e-01
-8.09971154e-01 3.66120309e-01 -4.18715805e-01 1.50020614e-01
3.88769925e-01 2.48789117e-01 -1.54723537e+00 1.38151407e-01
1.58924207e-01 6.56768456e-02 -2.00307637e-01 4.73117948e-01
-1.01970091e-01 1.38457584e+00 -5.54012835e-01 -4.36848044e-01
4.04731631e-01 -5.87973297e-02 9.40213442e-01 1.94749749e+00
5.74637830e-01 -1.64456412e-01 -9.53084156e-02 3.85903835e-01
-5.63215045e-03 5.65290689e-01 -5.39288402e-01 1.62104182e-02
4.90725517e-01 7.33827472e-01 -4.03672904e-01 -4.11834955e-01
-4.87951368e-01 1.10602903e+00 3.37296158e-01 2.02563435e-01
-3.53434533e-01 -2.10189641e-01 4.78716850e-01 2.00310096e-01
1.70916319e-01 -7.93875515e-01 -2.24774778e-01 -9.09717560e-01
9.24335495e-02 -1.37922132e+00 -1.57091931e-01 -5.91683328e-01
-1.14028132e+00 1.14845109e+00 -1.23845279e-01 -6.09877586e-01
-6.53317809e-01 -8.46412420e-01 -4.70428348e-01 1.30527782e+00
-1.27752101e+00 -1.41647875e+00 6.49663389e-01 3.25932175e-01
1.17662513e+00 -8.03304613e-01 1.31795895e+00 5.34727871e-01
-2.92847931e-01 8.06783617e-01 5.25652289e-01 2.91486293e-01
4.93167281e-01 -1.25766134e+00 1.27285683e+00 9.86065865e-01
6.08347297e-01 4.90011305e-01 4.82052147e-01 -6.40980482e-01
-1.06508720e+00 -1.06336796e+00 1.73856497e+00 -6.16462529e-01
7.08836079e-01 -8.50702167e-01 -9.68055546e-01 6.56992316e-01
5.23184180e-01 -5.18706501e-01 5.01548469e-01 1.93059340e-01
-1.02917068e-01 8.24063346e-02 -8.36764276e-01 8.56523037e-01
1.00695610e+00 -8.68691683e-01 -6.99975848e-01 3.92052621e-01
9.09734547e-01 -1.83277741e-01 -6.42570376e-01 3.36661905e-01
3.63692641e-01 -5.81379771e-01 6.08939767e-01 -5.52755117e-01
-1.45979002e-01 1.78447608e-02 -6.36002183e-01 -1.69650924e+00
-1.69988960e-01 -9.58323538e-01 1.58513278e-01 1.35917974e+00
9.73542273e-01 -7.26238728e-01 4.80581701e-01 2.53097087e-01
-5.43815672e-01 -5.39373696e-01 -1.47044730e+00 -1.23014510e+00
4.38039750e-01 -8.81506741e-01 3.48506182e-01 7.08671510e-01
-2.12767199e-01 5.76232195e-01 -3.24345827e-01 1.32161388e-02
2.87845153e-02 -8.53204906e-01 4.74042296e-01 -5.61170042e-01
-3.76069129e-01 -5.22659063e-01 -3.84113103e-01 -1.16865456e+00
4.58727121e-01 -1.12230682e+00 3.12459528e-01 -1.51225638e+00
-9.80684236e-02 -3.27462167e-01 -1.29067227e-01 5.53594410e-01
1.36344388e-01 7.75963068e-02 7.97793195e-02 -9.24835503e-02
-6.01278022e-02 6.19351208e-01 6.68466032e-01 1.35773689e-01
-3.49573463e-01 1.83566734e-01 -7.47815371e-02 3.81366998e-01
9.74215508e-01 -6.14910960e-01 2.13057492e-02 -6.48031533e-01
-1.74437791e-01 1.62973717e-01 -2.32304811e-01 -8.37626934e-01
7.41746426e-02 1.14025466e-01 -4.79392745e-02 -6.37128770e-01
4.38246936e-01 -3.41748774e-01 2.86978018e-03 1.75420389e-01
-2.91583747e-01 2.57609665e-01 3.20858449e-01 -2.58448213e-01
-3.09943229e-01 -3.87873232e-01 8.53558779e-01 -1.25446096e-01
-5.95508575e-01 -8.59484226e-02 -1.13091159e+00 -2.14875042e-02
2.77874351e-01 -1.63898870e-01 -7.49934688e-02 -6.70482934e-01
-7.62648344e-01 -5.62534630e-01 5.02602831e-02 7.86840439e-01
6.01184785e-01 -1.02732110e+00 -1.30794585e+00 2.90525198e-01
-1.22527763e-01 -4.81993109e-01 -1.74570039e-01 5.48257887e-01
-4.13584858e-01 1.02363658e+00 1.86894000e-01 -3.46599728e-01
-1.35350108e+00 5.86311929e-02 4.46298212e-01 -3.69992465e-01
-3.60809892e-01 8.39798152e-01 -2.87392259e-01 -1.06411767e+00
1.72606915e-01 -4.20524478e-02 -3.44499201e-02 -3.47361773e-01
3.65403980e-01 5.02019286e-01 7.36498415e-01 -1.08213913e+00
-4.82802391e-01 1.04214467e-01 -2.45201260e-01 -8.46687257e-01
1.26207995e+00 -2.08032832e-01 6.11697659e-02 6.29235208e-01
1.24019766e+00 4.09431547e-01 -2.98992813e-01 -3.52385432e-01
2.62704521e-01 7.94879124e-02 4.41638201e-01 -1.10331941e+00
-5.67590594e-01 1.10682869e+00 6.42353475e-01 1.45392856e-02
6.74125493e-01 -8.88536498e-02 1.30419040e+00 7.81545937e-01
2.29697138e-01 -1.47602141e+00 -5.10753453e-01 1.36960304e+00
9.61913228e-01 -1.06024086e+00 -7.28217423e-01 -5.73949714e-04
-7.27186203e-01 1.06559730e+00 2.70993561e-01 1.57561630e-01
7.52836883e-01 4.91630584e-01 5.26877284e-01 2.20436469e-01
-8.07731152e-01 -2.49818504e-01 3.22903514e-01 6.79054856e-01
9.72336411e-01 4.07358199e-01 -3.12622815e-01 4.28833604e-01
-6.79631889e-01 -3.06542546e-01 3.59395653e-01 4.82178360e-01
-3.21321934e-01 -1.57139599e+00 -2.23419964e-01 1.65971406e-02
-8.22264254e-01 -6.43878222e-01 -5.68849802e-01 7.34284163e-01
-3.80520642e-01 1.20162189e+00 -2.55819913e-02 -4.41130728e-01
4.52051133e-01 6.15203261e-01 3.05800796e-01 -9.11967039e-01
-6.14346087e-01 3.38678896e-01 6.80281639e-01 -2.96633124e-01
-6.57622963e-02 -7.56614327e-01 -1.01656497e+00 -1.99917674e-01
-4.09701377e-01 -2.54721828e-02 1.28256440e+00 1.15544260e+00
1.85411945e-01 2.19262376e-01 6.29241288e-01 -6.77948952e-01
-5.79485118e-01 -1.58804739e+00 -3.69352758e-01 -1.28564224e-01
9.47796330e-02 7.60114565e-02 -2.47860134e-01 -1.19010657e-01] | [14.34303092956543, 7.047705173492432] |
9ef8060a-d857-445e-bd69-9388fc829aec | good-for-misconceived-reasons-an-empirical | 2105.14462 | null | https://arxiv.org/abs/2105.14462v1 | https://arxiv.org/pdf/2105.14462v1.pdf | Good for Misconceived Reasons: An Empirical Revisiting on the Need for Visual Context in Multimodal Machine Translation | A neural multimodal machine translation (MMT) system is one that aims to perform better translation by extending conventional text-only translation models with multimodal information. Many recent studies report improvements when equipping their models with the multimodal module, despite the controversy of whether such improvements indeed come from the multimodal part. We revisit the contribution of multimodal information in MMT by devising two interpretable MMT models. To our surprise, although our models replicate similar gains as recently developed multimodal-integrated systems achieved, our models learn to ignore the multimodal information. Upon further investigation, we discover that the improvements achieved by the multimodal models over text-only counterparts are in fact results of the regularization effect. We report empirical findings that highlight the importance of MMT models' interpretability, and discuss how our findings will benefit future research. | ['Ben Kao', 'Xiang Li', 'Wei Bi', 'Lingpeng Kong', 'Zhiyong Wu'] | 2021-05-30 | null | https://aclanthology.org/2021.acl-long.480 | https://aclanthology.org/2021.acl-long.480.pdf | acl-2021-5 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 4.19262201e-01 4.53858346e-01 -5.79316974e-01 -3.12559456e-01
-1.19844639e+00 -6.87792897e-01 9.86178696e-01 -1.39949664e-01
-2.76611209e-01 7.92705059e-01 7.17011690e-01 -6.76721454e-01
3.39893438e-02 -2.46931046e-01 -8.71477008e-01 -2.80617118e-01
3.19231719e-01 8.98296058e-01 -6.31275356e-01 -6.42638206e-01
1.89558212e-02 -1.00902811e-01 -9.03566062e-01 7.77306795e-01
1.09427774e+00 3.75859201e-01 -1.31490320e-01 4.88510311e-01
-1.94484055e-01 6.74652636e-01 -3.72436762e-01 -1.14374506e+00
1.41327441e-01 -5.65977693e-01 -6.85440481e-01 1.12920761e-01
4.74649787e-01 -3.55603904e-01 -2.94560075e-01 7.75246441e-01
4.40472007e-01 -2.38570794e-01 7.15832949e-01 -1.33688629e+00
-1.37067389e+00 8.65934610e-01 -3.42888683e-01 -4.70134914e-01
5.69608569e-01 1.88865140e-01 1.35307086e+00 -1.22900140e+00
8.91462922e-01 1.53445590e+00 6.98148787e-01 5.70287406e-01
-1.45611215e+00 -1.19522423e-01 1.63443968e-01 1.48301469e-02
-1.01590872e+00 -6.94609106e-01 4.89910632e-01 -2.45074943e-01
1.00553000e+00 5.12645423e-01 2.63471007e-01 1.57279599e+00
3.67784381e-01 1.13687241e+00 1.20367384e+00 -7.37518370e-01
-3.09849471e-01 4.69565988e-01 -7.91209936e-02 5.42144120e-01
2.21100673e-01 -8.68213326e-02 -8.02188694e-01 -1.08594894e-01
4.50725406e-01 -2.83595711e-01 -2.11285189e-01 -3.07698131e-01
-1.58740866e+00 9.50074375e-01 1.67133193e-02 2.77229130e-01
-2.24175945e-01 1.20838165e-01 3.45077187e-01 7.61062324e-01
5.74156880e-01 5.00569820e-01 -5.59466779e-01 -3.85366917e-01
-7.77867377e-01 -1.17138535e-01 7.91685164e-01 1.08817613e+00
8.25377405e-01 -1.23790100e-01 -2.30953351e-01 9.83282864e-01
4.20917243e-01 8.34862053e-01 2.58937657e-01 -1.00050759e+00
1.03629637e+00 6.53985023e-01 8.30198005e-02 -7.42940128e-01
-2.72441536e-01 -3.43130261e-01 -4.36252415e-01 -4.06804055e-01
3.37184966e-01 -5.41438401e-01 -6.17146015e-01 1.82955313e+00
-3.63299578e-01 -7.00718462e-01 2.47776911e-01 1.03837287e+00
6.88870609e-01 4.72211033e-01 2.58036971e-01 -1.22890972e-01
1.19565976e+00 -1.06552923e+00 -1.05047035e+00 -4.46217328e-01
1.05954802e+00 -1.15218139e+00 1.10963202e+00 -6.66365102e-02
-1.07876492e+00 -1.11929707e-01 -6.68403745e-01 -3.54509830e-01
-4.70807701e-01 5.01350105e-01 9.27322030e-01 8.17067802e-01
-1.24707401e+00 3.23058486e-01 -6.46108568e-01 -7.62756467e-01
6.44795001e-02 5.24094999e-01 -4.15314823e-01 6.05613589e-02
-1.26771605e+00 1.48124123e+00 -1.37132883e-01 1.60935789e-01
-5.06587863e-01 -2.01703563e-01 -9.30088162e-01 -2.12288365e-01
2.06046730e-01 -1.10417199e+00 1.40436292e+00 -1.54765427e+00
-1.61261094e+00 6.86971962e-01 -6.74636304e-01 -1.14561357e-01
4.35594499e-01 -7.08628893e-02 -1.67591438e-01 5.16575128e-02
-3.19138579e-02 9.20759439e-01 5.62379062e-01 -1.53135443e+00
-2.69826859e-01 -2.20440313e-01 1.89955756e-02 4.88790601e-01
-6.13737285e-01 1.56417638e-01 -4.46779490e-01 -4.41253096e-01
-1.42732367e-01 -1.19755137e+00 4.73101251e-03 -2.61405379e-01
-5.60892642e-01 -3.07413042e-02 4.86739218e-01 -8.04843068e-01
1.13745856e+00 -1.76778769e+00 7.96720743e-01 -2.28532627e-01
1.40254200e-01 -8.65314305e-02 -8.01617861e-01 1.09419882e+00
-2.35716961e-02 3.62984300e-01 -2.71231502e-01 -1.01846254e+00
6.30203366e-01 3.78143817e-01 -2.96905041e-01 1.67001292e-01
5.35554647e-01 1.58061707e+00 -3.31696719e-01 -4.51923043e-01
-3.05667762e-02 5.40233314e-01 -3.40489775e-01 -1.72030210e-01
-3.41453493e-01 4.68871653e-01 -2.81607985e-01 8.66159022e-01
5.78339159e-01 -2.08322078e-01 4.48704123e-01 -1.87082648e-01
-9.81512889e-02 2.53754824e-01 -2.49044165e-01 1.79450321e+00
-2.35256478e-01 1.02632511e+00 1.84333339e-01 -6.35865748e-01
3.82435888e-01 5.11800170e-01 2.74126351e-01 -8.85054052e-01
2.32607678e-01 3.57332587e-01 1.58572286e-01 -7.08090723e-01
8.69626045e-01 -4.63677466e-01 -2.49806587e-02 6.89274251e-01
9.75831673e-02 4.83649634e-02 -1.66901909e-02 4.27410543e-01
6.10724628e-01 4.79511172e-01 -3.11021149e-01 1.94170922e-01
6.54020384e-02 4.43143666e-01 1.62888765e-01 8.09629738e-01
-1.02073900e-01 5.52453041e-01 5.43420494e-01 -9.93425176e-02
-1.14705920e+00 -6.82559252e-01 -8.23796690e-02 1.32502544e+00
-1.36907801e-01 -5.05183876e-01 -6.21156573e-01 -7.49756634e-01
-6.46003634e-02 7.55456030e-01 -6.48865640e-01 -7.32998997e-02
-5.62789679e-01 -1.13455021e+00 8.27848732e-01 3.74625564e-01
-2.67238840e-02 -4.12577510e-01 -7.53500091e-04 -1.53739721e-01
-7.95895636e-01 -1.14880276e+00 -4.31889832e-01 1.08323701e-01
-1.15098333e+00 -3.52841586e-01 -8.67050529e-01 -6.22367561e-01
5.20506501e-01 2.91173130e-01 1.17686236e+00 -5.03387228e-02
5.21800280e-01 8.80315483e-01 -4.96772617e-01 -2.78694063e-01
-6.72025263e-01 4.80153769e-01 -9.60094258e-02 -2.07557127e-01
5.31230271e-01 -2.35808983e-01 -1.55943364e-01 3.25160414e-01
-9.59403276e-01 4.71751064e-01 9.79157209e-01 7.89248228e-01
-1.29362822e-01 -8.16479743e-01 7.04964399e-01 -6.49159014e-01
1.01271915e+00 -5.46666980e-01 1.04906477e-01 5.21976471e-01
-6.12428427e-01 1.11574955e-01 1.85875177e-01 -5.15767515e-01
-1.18301773e+00 -2.41140187e-01 1.65034845e-01 -1.35981053e-01
2.72034612e-02 9.83761013e-01 -9.86854266e-03 -2.88472533e-01
3.25122982e-01 1.34138778e-01 2.98813343e-01 -3.90182644e-01
6.90977633e-01 6.40635729e-01 -6.25494719e-02 -9.01510477e-01
7.94530809e-01 1.32491693e-01 -9.14617702e-02 -4.92297202e-01
-2.85015583e-01 1.08940586e-01 -5.58768928e-01 -2.28954077e-01
7.24177182e-01 -1.05449784e+00 -4.67338324e-01 -2.01123774e-01
-1.33566499e+00 -1.37311772e-01 2.43363455e-01 7.17745721e-01
-5.34542084e-01 6.75280809e-01 -9.67254460e-01 -9.70522642e-01
-8.18149373e-02 -1.34799254e+00 1.38253868e+00 -8.30519348e-02
-4.73889053e-01 -1.40847099e+00 1.85977817e-01 9.37921643e-01
6.73389137e-01 -2.96904504e-01 1.34481585e+00 -6.34043455e-01
-7.17572629e-01 -1.35340795e-01 -3.78791213e-01 -1.07849352e-01
2.64115036e-02 -7.79886469e-02 -9.03396904e-01 -1.48815468e-01
-2.37253323e-01 -3.68436843e-01 7.66009331e-01 1.11571707e-01
-3.24512273e-02 -3.23459238e-01 -1.44012019e-01 2.46599108e-01
1.20347178e+00 -1.75302356e-01 5.06104469e-01 3.82505357e-01
7.55881131e-01 8.44075859e-01 3.71489495e-01 4.08482887e-02
9.75815177e-01 7.95568347e-01 4.17593241e-01 -4.02405798e-01
-2.29093552e-01 -3.31735551e-01 8.38322163e-01 1.31451714e+00
-2.01598883e-01 -4.38921630e-01 -8.53831589e-01 4.35427636e-01
-2.24318981e+00 -6.13649309e-01 -3.25603336e-01 1.91091061e+00
8.32896829e-01 -4.04393166e-01 8.49170536e-02 -5.34157991e-01
3.05570483e-01 -1.98743746e-01 -1.09046169e-01 -7.58211553e-01
-7.21554160e-01 -1.06859915e-01 3.68538499e-01 7.09951103e-01
-7.29576170e-01 1.09859574e+00 7.72165203e+00 5.05847633e-01
-9.63736475e-01 1.90774798e-01 4.55450475e-01 -1.57142967e-01
-8.79201591e-01 2.34753057e-01 -2.68437028e-01 5.72437197e-02
1.17268860e+00 1.41519010e-01 6.20146751e-01 7.88496062e-02
3.46284866e-01 -5.63146658e-02 -1.50939918e+00 5.32944560e-01
3.17444950e-01 -9.98195350e-01 5.12371063e-01 3.99527937e-01
7.65049100e-01 3.36540602e-02 5.32846689e-01 4.92572129e-01
6.46600053e-02 -1.08765137e+00 8.35609078e-01 5.87400734e-01
4.29425210e-01 -4.37450230e-01 8.11783016e-01 2.70037115e-01
-5.07010221e-01 -1.63185559e-02 -1.39155462e-01 -2.59035528e-01
1.70987919e-01 -3.27156813e-05 -6.95214689e-01 1.00096834e+00
-3.96163389e-02 6.76049650e-01 -8.18589091e-01 5.54094017e-01
1.67644173e-02 5.60056448e-01 1.00125149e-01 -2.36131884e-02
3.78355116e-01 -5.59226990e-01 7.39309907e-01 1.38645113e+00
3.01286936e-01 -4.69670355e-01 -2.17502639e-01 9.18634832e-01
-1.78639486e-01 3.75917375e-01 -8.15515280e-01 -6.60260081e-01
-5.50141335e-02 1.06325626e+00 -2.17275128e-01 -2.12186024e-01
-8.57188225e-01 1.28151643e+00 2.80949533e-01 8.47825885e-01
-8.32046628e-01 1.34290591e-01 4.52156723e-01 -3.19987625e-01
-4.09086347e-02 -6.06659889e-01 -8.58393967e-01 -1.69421411e+00
1.08416542e-01 -1.28318155e+00 7.84080997e-02 -1.03299606e+00
-1.34224916e+00 4.28712159e-01 -1.07203685e-01 -1.12379479e+00
-3.17568332e-01 -7.21727550e-01 -6.97839186e-02 9.11476135e-01
-1.37661636e+00 -1.72837031e+00 4.48980898e-01 2.24932253e-01
4.31612074e-01 -7.30179846e-02 1.02294385e+00 3.91054839e-01
-6.57100379e-01 8.39056432e-01 1.95284262e-01 1.62321832e-02
1.19435489e+00 -9.76917386e-01 5.80636300e-02 6.01326942e-01
1.98544577e-01 1.18180346e+00 7.74112105e-01 -7.21207321e-01
-2.26578903e+00 -5.42827427e-01 1.43965638e+00 -1.21961260e+00
7.11730480e-01 -3.18258822e-01 -4.97395366e-01 1.22315586e+00
1.21915293e+00 -9.77692842e-01 9.39460397e-01 3.71398777e-01
-4.04939890e-01 5.04244804e-01 -7.38079190e-01 9.85214114e-01
8.29048276e-01 -7.21093893e-01 -6.96766257e-01 2.22678602e-01
7.81307876e-01 -5.36285564e-02 -1.01576853e+00 3.74305367e-01
7.52192855e-01 -6.26530349e-01 7.21781969e-01 -9.42508221e-01
8.45807493e-01 -9.15714204e-02 -5.42207837e-01 -1.28347206e+00
7.08033098e-03 -6.28246784e-01 1.40705764e-01 9.82606590e-01
1.08000278e+00 -7.46366441e-01 4.24724311e-01 9.18467641e-01
-2.09500313e-01 -3.97826225e-01 -1.09985995e+00 -5.40928006e-01
7.07078159e-01 -4.11826342e-01 3.26494277e-01 1.22905588e+00
5.88820934e-01 7.65700579e-01 -7.81588674e-01 -6.98531866e-02
4.67173636e-01 -2.61327106e-04 7.77444541e-01 -9.05611932e-01
-3.27691555e-01 -6.59421504e-01 9.35066417e-02 -1.11979055e+00
3.00001889e-01 -1.06896985e+00 -2.63490319e-01 -1.64208663e+00
7.04149485e-01 1.47235051e-01 3.03899586e-01 7.15568841e-01
-1.46629378e-01 4.97620374e-01 5.58599889e-01 4.36294883e-01
-7.80150652e-01 8.46375227e-01 1.52701461e+00 -2.89194435e-01
-1.17950412e-02 -3.31160605e-01 -1.05748641e+00 2.39105046e-01
5.49539745e-01 -2.16228113e-01 -2.38035411e-01 -1.23724246e+00
6.61467016e-01 2.31833488e-01 2.17352360e-01 -9.44648981e-02
2.53822178e-01 -2.38761365e-01 2.67053008e-01 -3.14165920e-01
7.00443685e-01 -9.16831434e-01 1.87171310e-01 -1.34157408e-02
-5.84168613e-01 3.71331811e-01 5.16120374e-01 2.55324006e-01
-2.33096465e-01 9.37016122e-03 -2.62967907e-02 1.34103596e-01
-4.12093475e-02 -3.20741624e-01 -1.05090797e+00 -5.17905891e-01
2.93981642e-01 -9.83766615e-02 -7.64114499e-01 -1.04138219e+00
-6.62969649e-01 2.03760356e-01 6.38386130e-01 6.66405857e-01
4.04767513e-01 -1.55357885e+00 -8.21937680e-01 -1.99034274e-01
2.99134612e-01 -9.48165774e-01 -1.58390731e-01 1.61826897e+00
2.85999347e-02 8.61392438e-01 2.95379548e-03 -4.74035770e-01
-1.06554353e+00 2.47754917e-01 3.52024436e-01 9.02817026e-03
-5.94656654e-02 1.49593100e-01 -3.80388163e-02 -9.18060303e-01
6.12414889e-02 -6.80783093e-02 2.40937606e-01 3.65296453e-02
-7.82178249e-03 2.42445543e-01 -6.06008843e-02 -8.66206944e-01
-2.82820255e-01 5.71033716e-01 -7.43873194e-02 -7.71840930e-01
1.08917975e+00 -6.03977501e-01 -5.23147523e-01 5.34833133e-01
9.43955421e-01 2.11090833e-01 -4.85332012e-01 -2.11647540e-01
6.24057241e-02 -1.52931646e-01 -3.39074820e-01 -1.43212616e+00
-6.26903236e-01 9.15732861e-01 3.77789378e-01 -1.48865581e-02
9.47199583e-01 -2.83885356e-02 6.49966896e-01 5.29302180e-01
2.27195561e-01 -9.40236270e-01 -1.82754099e-01 7.45972335e-01
8.01720202e-01 -1.47764552e+00 -1.57435596e-01 -2.10741758e-01
-1.00679243e+00 1.23535669e+00 3.67125005e-01 6.29639983e-01
-7.02553391e-02 -1.07789375e-02 4.07729298e-01 -1.92561839e-02
-1.12223983e+00 -3.12648773e-01 4.92207408e-01 5.89648664e-01
9.91058886e-01 1.20710708e-01 -7.41258383e-01 6.79626882e-01
1.18031492e-02 -2.66481042e-01 3.43440801e-01 8.56130898e-01
-1.35199472e-01 -1.57871044e+00 -5.51825166e-01 -5.55336066e-02
-2.66228199e-01 -4.33838487e-01 -1.30361402e+00 8.80795181e-01
-2.43799537e-01 1.40551698e+00 -3.47536027e-01 -6.94117188e-01
2.02159345e-01 5.51325202e-01 5.91352165e-01 -2.11019918e-01
-7.90071070e-01 3.57613623e-01 5.87687850e-01 -3.42335254e-01
-4.87591296e-01 -7.06496477e-01 -8.07494402e-01 -4.45983976e-01
-2.30207324e-01 5.60319237e-02 9.26175475e-01 1.08065474e+00
6.11253023e-01 1.72508240e-01 2.82707959e-01 -7.46266782e-01
-4.62642848e-01 -1.15722632e+00 2.19905764e-01 2.44151622e-01
3.57402980e-01 -2.24812046e-01 -3.69225770e-01 1.03422001e-01] | [11.454014778137207, 1.4890210628509521] |
426caf2c-f1a1-40d2-9f57-72a267e4ec55 | data-driven-intelligent-computational-design | 2301.12382 | null | https://arxiv.org/abs/2301.12382v2 | https://arxiv.org/pdf/2301.12382v2.pdf | Data-driven intelligent computational design for products: Method, techniques, and applications | Data-driven intelligent computational design (DICD) is a research hotspot emerged under the context of fast-developing artificial intelligence. It emphasizes on utilizing deep learning algorithms to extract and represent the design features hidden in historical or fabricated design process data, and then learn the combination and mapping patterns of these design features for the purposes of design solution retrieval, generation, optimization, evaluation, etc. Due to its capability of automatically and efficiently generating design solutions and thus supporting human-in-the-loop intelligent and innovative design activities, DICD has drawn the attentions from both academic and industrial fields. However, as an emerging research subject, there are still many unexplored issues that limit the development and application of DICD, such as specific dataset building, engineering design related feature engineering, systematic methods and techniques for DICD implementation in the entire product design process, etc. In this regard, a systematic and operable road map for DICD implementation from full-process perspective is established, including a general workflow for DICD project planning, an overall framework for DICD project implementation, the computing mechanisms for DICD implementation, key enabling technologies for detailed DICD implementation, and three application scenarios of DICD. The road map reveals the common mechanisms and calculation principles of existing DICD researches, and thus it can provide systematic guidance for the possible DICD applications that have not been explored. | ['Yuhao Liu', 'Tianshuo Zang', 'Pingyu Jiang', 'Maolin Yang'] | 2023-01-29 | null | null | null | null | ['feature-engineering'] | ['methodology'] | [-1.27291277e-01 -4.94598508e-01 -1.14982851e-01 -3.71586531e-01
1.16395364e-02 -2.88379490e-01 3.76008838e-01 9.36024860e-02
4.29934949e-01 1.80520847e-01 2.26262450e-01 -3.23655307e-01
-1.08201742e+00 -1.11587584e+00 -1.20419012e-02 -6.12096667e-01
1.64648280e-01 7.73981929e-01 -6.77529931e-01 -2.27083489e-01
5.43736815e-01 1.05719769e+00 -1.92845440e+00 5.30481100e-01
3.83112103e-01 1.16218460e+00 3.57624620e-01 8.58424883e-03
-4.20124024e-01 5.54517210e-01 -3.41111004e-01 -7.33478144e-02
1.99213699e-01 -2.01453865e-01 -4.78404224e-01 1.68722898e-01
-3.44581723e-01 -1.86755821e-01 -3.13278019e-01 4.00495619e-01
9.51366484e-01 -1.52251363e-01 6.46796644e-01 -1.06682050e+00
-9.20763493e-01 4.18649763e-01 -2.75377750e-01 -4.27018583e-01
3.83936226e-01 4.98756081e-01 7.93681920e-01 -1.47075820e+00
6.87434018e-01 9.93637979e-01 3.33344460e-01 2.57736087e-01
-8.16197932e-01 -4.40653920e-01 -2.68890053e-01 2.06448123e-01
-1.33365464e+00 -1.10902168e-01 1.33821344e+00 -8.18451464e-01
9.87845600e-01 1.57445222e-02 9.02423978e-01 7.41220534e-01
4.37038928e-01 9.57866728e-01 6.25109494e-01 -5.16526997e-01
7.59031236e-01 1.00957481e-02 2.19504654e-01 3.33388805e-01
2.42057666e-01 2.83184469e-01 -6.54851556e-01 2.43032321e-01
7.57255137e-01 1.15623333e-01 2.47668669e-01 -6.05371654e-01
-9.78613615e-01 5.48826218e-01 3.60835314e-01 4.61549580e-01
-8.41564536e-01 -2.00226128e-01 4.01907831e-01 3.56213093e-01
1.41879946e-01 7.80388892e-01 -6.95290625e-01 -1.59820750e-01
-7.91889548e-01 6.65158391e-01 5.64009547e-01 1.15738440e+00
8.00257206e-01 1.59802869e-01 -1.74757808e-01 5.87360978e-01
4.51333672e-01 2.70048320e-01 1.39936954e-01 -8.69186521e-01
3.82190198e-01 1.08697510e+00 -5.05395085e-02 -1.18295133e+00
-6.86270356e-01 -5.00570714e-01 -1.00201213e+00 3.40919107e-01
-2.43771672e-01 -2.02500939e-01 -6.71678960e-01 9.52237666e-01
3.13041806e-01 -6.20978475e-01 -2.26320520e-01 9.66016769e-01
1.02348638e+00 6.60992563e-01 -1.13689162e-01 -7.49969780e-02
1.15252519e+00 -4.11486745e-01 -8.32084417e-01 2.88585909e-02
7.15814650e-01 -9.97502327e-01 8.20575774e-01 5.68342865e-01
-9.09853518e-01 -8.76538932e-01 -1.28705215e+00 7.34969899e-02
-5.50357461e-01 3.62700254e-01 1.08150136e+00 5.44323027e-01
-4.98008937e-01 6.26259804e-01 -4.11886871e-01 -1.57622769e-01
8.12137723e-01 4.11255091e-01 2.33845506e-02 -5.40696084e-01
-9.37944174e-01 9.33042645e-01 4.26935375e-01 5.28986156e-01
-8.70889068e-01 -1.17367518e+00 -5.52357078e-01 -1.40493602e-01
2.26285636e-01 -9.29130793e-01 9.48843002e-01 -3.13457042e-01
-1.58673382e+00 5.10373235e-01 3.05015475e-01 1.02147363e-01
1.64127741e-02 -1.27294540e-01 -6.66477919e-01 -4.94425684e-01
-1.59871802e-01 2.12189466e-01 4.65069205e-01 -9.91366088e-01
-4.72472370e-01 -5.49719870e-01 -3.42677653e-01 9.16703269e-02
-2.94996351e-01 2.00716797e-02 -2.89184064e-01 -4.36754376e-01
2.86400497e-01 -5.58902383e-01 -4.25165266e-01 3.22607219e-01
-3.42254251e-01 -2.72002459e-01 1.13422513e+00 -4.44718748e-01
1.53281021e+00 -2.13130832e+00 2.28779435e-01 4.20044363e-01
1.36092141e-01 2.19388813e-01 3.46347280e-02 9.23475981e-01
-5.06700762e-02 -2.24598974e-01 3.16531882e-02 2.52353787e-01
3.63811046e-01 -1.38023928e-01 4.03215177e-02 2.41269648e-01
2.36665592e-01 9.53497171e-01 -5.89861035e-01 6.00804798e-02
8.76588345e-01 4.20217991e-01 -3.77321959e-01 1.95176870e-01
-4.23583806e-01 2.35195518e-01 -9.79624867e-01 1.01715755e+00
7.17123806e-01 -3.51591385e-04 2.07373738e-01 -6.49838626e-01
-5.71735322e-01 4.43176888e-02 -1.46137619e+00 1.86222422e+00
-7.75898159e-01 6.66201115e-01 -1.77828774e-01 -1.03317058e+00
1.63549769e+00 1.41168475e-01 7.95202911e-01 -1.05756187e+00
3.89511168e-01 3.67820770e-01 -2.97054630e-02 -7.37115920e-01
4.15760607e-01 2.15816572e-01 -1.76545292e-01 7.00068057e-01
-1.30370364e-01 -3.82073402e-01 5.85026965e-02 -3.32557678e-01
9.06261623e-01 3.65260512e-01 2.11026054e-02 -3.37816298e-01
5.34086525e-01 6.05271049e-02 2.89038241e-01 8.49423651e-03
3.64045590e-01 4.81659889e-01 8.66348073e-02 -1.11811769e+00
-1.15853190e+00 -6.85676575e-01 -1.05253913e-01 3.18218768e-01
2.10332513e-01 -5.21916449e-01 -4.02096897e-01 -8.35132599e-02
2.60116637e-01 7.89111376e-01 -2.96902269e-01 -2.82949984e-01
-6.48424566e-01 -6.23178363e-01 -8.46797526e-02 4.54329997e-01
4.18559730e-01 -1.27624238e+00 -6.11454368e-01 6.19204640e-01
6.14341557e-01 -6.03981674e-01 2.30908766e-01 2.01222166e-01
-8.28575730e-01 -1.03095829e+00 -4.11344796e-01 -1.01025224e+00
5.71857393e-01 1.45163864e-01 1.20380032e+00 6.22863770e-02
-8.86212468e-01 -8.70643184e-02 -9.61806774e-02 -5.22116184e-01
-4.87907737e-01 1.56926751e-01 -4.97053340e-02 -2.66445905e-01
5.08848250e-01 -6.26209974e-01 -8.13615799e-01 5.37026644e-01
-5.75550020e-01 4.29697722e-01 8.01149428e-01 4.99155849e-01
6.66884601e-01 6.98776484e-01 6.80564582e-01 -2.65137285e-01
6.95865273e-01 -4.01619941e-01 -5.86094260e-01 2.25642949e-01
-8.47245395e-01 4.24703173e-02 4.75149304e-01 2.95408159e-01
-1.13099682e+00 6.68937787e-02 -2.98431993e-01 -2.68611908e-01
-3.02958786e-01 8.03080738e-01 -9.14912820e-01 2.94291496e-01
3.97810131e-01 1.80130843e-02 1.41023651e-01 -9.95651901e-01
5.30741334e-01 9.14408445e-01 -9.87834390e-03 -7.66118824e-01
7.21273363e-01 5.25776073e-02 2.45324120e-01 -7.27523565e-01
-3.18058342e-01 -1.27437338e-02 -6.86927259e-01 -4.85527486e-01
5.37853658e-01 -5.50093949e-01 -6.83825374e-01 4.04942662e-01
-1.24673176e+00 2.71183819e-01 -5.47685981e-01 3.38697016e-01
-4.35851276e-01 -3.58651847e-01 -3.33341897e-01 -6.07310534e-01
-5.82870841e-01 -1.51003695e+00 9.38345730e-01 1.93767071e-01
-5.90317011e-01 -6.76947653e-01 -1.48474872e-01 1.59383774e-01
5.00262856e-01 2.74710625e-01 1.60926366e+00 -5.82760237e-02
-8.06874990e-01 -3.34454685e-01 -9.77253318e-02 2.13561043e-01
5.24505675e-01 2.79266596e-01 -5.97744942e-01 1.81380495e-01
-5.30521832e-02 1.68593556e-01 -8.91513973e-02 4.96005446e-01
1.25303245e+00 9.89596173e-02 -4.53328222e-01 3.17025483e-01
1.55775523e+00 7.84269631e-01 7.98785150e-01 3.63445610e-01
5.98195076e-01 9.57335293e-01 8.07145596e-01 8.41940343e-01
-5.23471609e-02 8.71004283e-01 4.21664566e-01 -8.00218508e-02
-3.02626401e-01 -2.32885689e-01 -1.23539910e-01 8.47326398e-01
3.03768925e-03 -1.75952092e-01 -1.00811720e+00 4.21685219e-01
-1.79454267e+00 -7.62473047e-01 -2.97412068e-01 1.78640938e+00
2.86580324e-01 2.09032539e-02 -2.13941947e-01 5.25318503e-01
7.24185765e-01 -1.24490045e-01 -7.10965693e-01 -6.08250022e-01
-4.28393036e-02 6.45997643e-01 -4.55917120e-02 -4.77027446e-01
-7.85137534e-01 4.63364571e-01 5.48049974e+00 8.05198252e-01
-9.91506398e-01 -3.25920433e-01 4.99754310e-01 -5.95531054e-02
-5.44281185e-01 -5.48503436e-02 -6.22898221e-01 5.28028965e-01
6.36424422e-01 -3.21267456e-01 1.85837284e-01 1.23894870e+00
6.28291130e-01 3.24154913e-01 -1.31270409e+00 1.16353095e+00
-4.36436713e-01 -2.16404605e+00 1.47274703e-01 4.28817630e-01
9.99664724e-01 -6.18891835e-01 8.16861168e-02 -2.15987936e-02
-5.62154464e-02 -8.42504501e-01 7.19119191e-01 5.59836507e-01
4.84862715e-01 -1.21449208e+00 6.91569686e-01 -1.31490499e-01
-1.26675510e+00 -2.86023527e-01 -9.30957943e-02 -3.04412633e-01
4.01600420e-01 1.11467743e+00 -4.76229012e-01 9.75393295e-01
5.01813650e-01 9.55467343e-01 -2.34171882e-01 1.09003639e+00
8.95236433e-02 4.85916957e-02 5.44962659e-02 -2.42001221e-01
1.97230801e-01 -2.36644745e-01 6.47608042e-02 5.49658656e-01
7.06019521e-01 -2.05950469e-01 -8.97199810e-02 1.21159554e+00
1.36249825e-01 7.89499134e-02 -6.42264187e-01 -1.76554397e-01
6.04800642e-01 1.00487626e+00 -5.83369493e-01 2.62838840e-01
-1.92757204e-01 3.51131052e-01 -4.07140821e-01 1.25350848e-01
-7.36809969e-01 -5.28712988e-01 9.53718960e-01 4.28525358e-01
4.04581934e-01 -4.30979550e-01 -1.00725937e+00 -3.31604987e-01
1.40577316e-01 -7.87312150e-01 -1.22193014e-02 -7.56219923e-01
-1.28874755e+00 2.38448605e-01 -1.78469747e-01 -1.56450725e+00
-9.54713859e-03 -8.10117066e-01 -6.42275631e-01 6.68310642e-01
-9.89239693e-01 -1.07593763e+00 -3.32139403e-01 1.08007267e-01
7.95437396e-01 -6.70366466e-01 7.25195050e-01 7.39034355e-01
-8.52508545e-01 1.27000198e-01 3.63339633e-01 -3.02551776e-01
-2.48950440e-02 -5.35235226e-01 5.20533562e-01 3.67179275e-01
-2.60216862e-01 4.42105383e-01 4.49796557e-01 -5.71037948e-01
-2.33643770e+00 -1.03771245e+00 5.56699753e-01 -1.32743537e-01
4.31165576e-01 -1.66116089e-01 -5.14501631e-01 -1.36083007e-01
-3.66708525e-02 -4.89969790e-01 6.12783670e-01 2.61146948e-02
3.50984395e-01 -4.57873791e-01 -1.03531337e+00 5.47298133e-01
1.24701881e+00 -2.37099797e-01 -2.89091527e-01 4.97304559e-01
5.49814701e-01 -5.10347448e-02 -1.09875715e+00 4.25355285e-01
6.58699036e-01 -8.33168805e-01 9.27453101e-01 -2.65183479e-01
7.74400711e-01 -5.19790292e-01 -1.63445264e-01 -8.17549527e-01
-4.72664863e-01 -4.71816301e-01 -3.18339258e-01 1.45657015e+00
2.20228553e-01 2.29571592e-02 9.88453209e-01 7.52612591e-01
-7.36535132e-01 -1.19653523e+00 -7.46131063e-01 -4.30386186e-01
-2.94614494e-01 -8.69268239e-01 1.35901582e+00 5.02713859e-01
-4.22599286e-01 2.23728701e-01 -1.46319538e-01 -1.05106309e-01
4.62746352e-01 5.07735312e-01 6.77917361e-01 -1.31892681e+00
1.72999308e-01 -6.24871612e-01 -6.22974277e-01 -7.66198218e-01
-2.26480722e-01 -7.78459728e-01 -1.93860695e-01 -1.91423500e+00
-3.41409802e-01 -7.73609877e-01 -1.91160947e-01 6.36244118e-02
7.06529319e-01 -3.17558259e-01 -1.07594877e-01 1.42884448e-01
-3.80227976e-02 8.68756175e-01 1.48646629e+00 -3.98986369e-01
-4.10284251e-01 7.06886798e-02 -8.20579171e-01 4.57707405e-01
4.80441451e-01 -1.83667809e-01 -5.61264455e-01 -4.88616347e-01
4.12601173e-01 -2.32207581e-01 9.48294178e-02 -9.32841003e-01
2.94083357e-01 -2.33964786e-01 5.62876701e-01 -1.24474835e+00
-8.06912407e-02 -1.14883566e+00 8.57285738e-01 4.79702324e-01
8.75338074e-03 -5.94754480e-02 1.09264024e-01 -3.09316870e-02
-1.63384706e-01 -1.99050382e-01 4.20795083e-01 -2.64954735e-02
-1.16205251e+00 3.59956890e-01 -4.93094623e-01 -7.07836986e-01
1.15279770e+00 -5.62314451e-01 1.13124080e-01 3.81851196e-01
-4.72048730e-01 1.27709791e-01 2.51894295e-01 6.93418622e-01
8.76597524e-01 -1.70127046e+00 -4.58766609e-01 6.60891652e-01
9.78540927e-02 2.14480877e-01 7.28589475e-01 1.15155190e-01
-7.96063423e-01 5.44644833e-01 -4.16480303e-01 -4.53335315e-01
-5.64058900e-01 7.29051888e-01 1.49478719e-01 6.20823412e-04
-6.75857127e-01 3.53875637e-01 -1.36874795e-01 -2.66750813e-01
1.58647150e-01 -1.08361535e-01 7.79279927e-03 2.29359984e-01
5.19133270e-01 6.56323731e-01 6.50408328e-01 -2.71935731e-01
-5.88877380e-01 9.67141092e-01 5.41102663e-02 3.75642151e-01
1.71052945e+00 2.22195758e-05 3.07223061e-04 -4.83237989e-02
1.15348709e+00 -8.10421348e-01 -8.62312436e-01 -8.83775130e-02
3.41240883e-01 -4.91161466e-01 5.16091406e-01 -8.02507579e-01
-1.43249536e+00 7.49317229e-01 8.71459126e-01 -1.69240654e-01
1.14674485e+00 4.16126847e-02 8.74785244e-01 3.17667454e-01
6.04347050e-01 -1.33321154e+00 1.25239372e-01 1.89998761e-01
1.21676016e+00 -7.48592436e-01 4.97593760e-01 -2.74369448e-01
-2.38844216e-01 1.51174414e+00 4.95262653e-01 3.12260002e-01
1.05476141e+00 4.76830512e-01 -2.88507253e-01 -8.10278535e-01
-4.81618464e-01 1.45194814e-01 2.87483096e-01 6.98527217e-01
3.97987366e-01 -1.44080445e-01 -1.94539770e-01 8.67455542e-01
-1.13581851e-01 3.21801573e-01 -1.54863521e-01 1.37716126e+00
-2.22895458e-01 -1.39372647e+00 -2.03152642e-01 4.72114503e-01
3.24145406e-01 2.16942221e-01 -1.67410687e-01 7.27882504e-01
5.13768137e-01 6.55782402e-01 9.38445553e-02 -7.12558389e-01
9.87046182e-01 -2.46082887e-01 3.69518489e-01 -5.02928078e-01
-5.69100916e-01 -2.39172429e-01 8.15191716e-02 -3.13335836e-01
-1.57566100e-01 -5.02473474e-01 -1.00350976e+00 -4.82985407e-01
-3.02037418e-01 -1.48468345e-01 1.26101243e+00 9.36536610e-01
9.29660976e-01 8.00403893e-01 1.11065316e+00 -9.64459717e-01
1.95751086e-01 -4.41792876e-01 -6.14158571e-01 5.35529945e-03
-4.08318102e-01 -8.91721845e-01 1.99949428e-01 -2.36947328e-01] | [5.936655521392822, 3.169497489929199] |
742cb093-ae6e-4ae4-9365-e4121e54b31e | neural-architecture-search-for-joint-human | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zeng_Neural_Architecture_Search_for_Joint_Human_Parsing_and_Pose_Estimation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zeng_Neural_Architecture_Search_for_Joint_Human_Parsing_and_Pose_Estimation_ICCV_2021_paper.pdf | Neural Architecture Search for Joint Human Parsing and Pose Estimation | Human parsing and pose estimation are crucial for the understanding of human behaviors. Since these tasks are closely related, employing one unified model to perform two tasks simultaneously allows them to benefit from each other. However, since human parsing is a pixel-wise classification process while pose estimation is usually a regression task, it is non-trivial to extract discriminative features for both tasks while modeling their correlation in the joint learning fashion. Recent studies have shown that Neural Architecture Search (NAS) has the ability to allocate efficient feature connections for specific tasks automatically. With the spirit of NAS, we propose to search for an efficient network architecture (NPPNet) to tackle two tasks at the same time. On the one hand, to extract task-specific features for the two tasks and lay the foundation for the further searching of feature interaction, we propose to search their encoder-decoder architectures, respectively. On the other hand, to ensure two tasks fully communicate with each other, we propose to embed NAS units in both multi-scale feature interaction and high-level feature fusion to establish optimal connections between two tasks. Experimental results on both parsing and pose estimation benchmark datasets have demonstrated that the searched model achieves state-of-the-art performances on both tasks. | ['Wu Liu', 'Chi Su', 'Junjie Zhang', 'Qian Bao', 'Yuhang Huang', 'Dan Zeng'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['human-parsing'] | ['computer-vision'] | [ 2.98912168e-01 1.61258608e-01 -3.03846262e-02 -6.14803135e-01
-7.80986607e-01 -8.19423497e-02 1.65052578e-01 -1.13754533e-01
-6.72127903e-01 4.16514754e-01 -7.32935518e-02 1.15800552e-01
-9.43094939e-02 -5.31181276e-01 -8.18288565e-01 -6.44724607e-01
-1.03519395e-01 3.79893988e-01 4.21983421e-01 3.12990993e-02
-4.60486077e-02 2.38851443e-01 -1.45363557e+00 1.62930146e-01
7.78985918e-01 1.34605801e+00 7.80819774e-01 6.05774045e-01
-1.13480315e-02 6.08101606e-01 -3.30906987e-01 -6.35468781e-01
1.85831204e-01 -2.09420338e-01 -6.98737681e-01 1.44856393e-01
3.37513387e-01 -2.80232310e-01 -2.53593713e-01 9.75433111e-01
6.31134510e-01 4.86928411e-02 4.90394235e-01 -1.08585405e+00
-2.32299894e-01 3.31035346e-01 -8.44174027e-01 7.93730393e-02
1.52371719e-01 1.51352644e-01 9.40281689e-01 -7.21088290e-01
3.21889728e-01 1.33686996e+00 3.86400610e-01 5.72599471e-01
-9.54321504e-01 -6.24560833e-01 4.94520783e-01 1.82602182e-01
-1.16416609e+00 -3.09514403e-01 1.11615431e+00 -2.97318995e-01
7.32440650e-01 -2.93076523e-02 6.65341794e-01 1.20618784e+00
3.38915259e-01 1.19801402e+00 6.67031527e-01 -8.14885423e-02
-3.12192552e-03 -7.57760853e-02 1.09367929e-01 9.21442926e-01
2.78500944e-01 -1.06565863e-01 -8.38838696e-01 2.55738109e-01
9.10747647e-01 -8.33026022e-02 -2.73793489e-01 -5.22920132e-01
-1.19987881e+00 6.48341238e-01 7.28378296e-01 2.66344815e-01
-5.14266253e-01 3.98345828e-01 4.22812521e-01 1.95535850e-02
1.27192140e-01 4.20090526e-01 -7.12730587e-01 -3.88905592e-02
-5.76025546e-01 1.89864457e-01 5.75282812e-01 8.84798884e-01
6.97435975e-01 -3.92031163e-01 -3.95802975e-01 7.24347353e-01
5.19810319e-01 2.51606792e-01 4.74192381e-01 -7.02604830e-01
1.00984955e+00 7.05427945e-01 8.96257833e-02 -1.18295765e+00
-7.72811770e-01 -5.10532260e-01 -9.26325560e-01 -2.72170901e-01
6.77783370e-01 -1.82426915e-01 -6.51648164e-01 2.06564617e+00
4.09488827e-01 1.06843233e-01 -1.58779055e-01 1.03169644e+00
6.15879238e-01 4.49779242e-01 3.92082661e-01 -2.78007123e-03
1.87066984e+00 -1.15490997e+00 -6.54290795e-01 -8.18423748e-01
5.29498100e-01 -5.18632591e-01 8.68135691e-01 2.60509253e-01
-1.03010511e+00 -9.57188725e-01 -1.23376703e+00 -2.88968593e-01
-9.50264335e-02 6.47833347e-01 7.52325296e-01 4.18115884e-01
-6.86300755e-01 5.22979259e-01 -1.20820355e+00 -2.02554375e-01
4.58024055e-01 6.33191049e-01 -3.55694145e-01 2.20249668e-01
-1.21945667e+00 8.55196476e-01 4.37817514e-01 5.64207375e-01
-5.00939071e-01 -1.05975725e-01 -9.83233809e-01 1.89573199e-01
5.43802738e-01 -9.11290467e-01 1.16054177e+00 -7.15757906e-01
-1.41984916e+00 5.58998764e-01 -1.87132493e-01 -3.43061507e-01
4.71224397e-01 -6.54129148e-01 -3.55803110e-02 1.06271632e-01
1.87251702e-01 1.03363752e+00 8.27361107e-01 -9.92227435e-01
-8.70281935e-01 -7.27318108e-01 -9.00333971e-02 4.68290746e-01
-7.50074923e-01 -1.90786660e-01 -1.07974243e+00 -4.52577680e-01
3.13303739e-01 -8.37725818e-01 -2.65824765e-01 1.65081456e-01
-4.99820411e-01 -4.47069734e-01 7.12634921e-01 -7.92079687e-01
1.16343677e+00 -2.05945182e+00 6.29097939e-01 -2.06095800e-02
3.05720747e-01 2.09140122e-01 -1.41837567e-01 -9.87049076e-04
1.42718747e-01 -2.17447490e-01 -2.23015815e-01 -7.66151011e-01
-5.46989739e-02 2.16398776e-01 1.35198206e-01 4.69239682e-01
5.42748809e-01 1.05824530e+00 -6.57054842e-01 -8.04799616e-01
1.65852144e-01 4.23412114e-01 -5.88016570e-01 4.34230864e-01
1.04984327e-03 6.62408113e-01 -9.96993482e-01 4.62981433e-01
4.44982409e-01 -3.74872804e-01 2.73687243e-01 -4.54794526e-01
2.95750976e-01 9.26619116e-03 -1.23522913e+00 2.05687022e+00
-4.92695957e-01 2.84722447e-01 2.26445138e-01 -1.14763236e+00
8.73718619e-01 2.97540814e-01 4.34194416e-01 -7.47741461e-01
4.44660544e-01 2.28055432e-01 6.34573102e-02 -4.97796267e-01
2.77933151e-01 2.62013435e-01 -3.71229023e-01 1.03683680e-01
1.30580530e-01 2.39211306e-01 9.04396027e-02 -2.67686307e-01
1.01427662e+00 3.89374495e-01 3.22617084e-01 -9.55283791e-02
7.89765418e-01 -4.41231966e-01 7.43425310e-01 5.11903524e-01
-4.37570065e-01 4.69553173e-01 3.74818325e-01 -6.30214810e-01
-5.66983700e-01 -8.57909203e-01 1.11464195e-01 1.17457712e+00
5.01517892e-01 -2.36951783e-01 -9.01930034e-01 -9.23653483e-01
-3.07229728e-01 1.99053630e-01 -6.40801907e-01 -2.28189960e-01
-9.81136620e-01 -6.97497249e-01 3.53632927e-01 9.11320984e-01
8.71322930e-01 -1.10685420e+00 -1.25413895e+00 3.92835587e-01
-3.42113346e-01 -1.39620256e+00 -5.88170409e-01 6.20599627e-01
-6.16235673e-01 -1.23949969e+00 -8.01688194e-01 -9.48421419e-01
7.61583090e-01 1.52559921e-01 9.93882477e-01 1.96320176e-01
-3.60103011e-01 1.32627040e-01 -1.87979445e-01 -1.66861400e-01
2.09102169e-01 4.60917264e-01 -2.10327670e-01 1.07236505e-01
2.55309433e-01 -5.14820457e-01 -7.79759765e-01 4.84352112e-01
-7.92691946e-01 2.25631252e-01 9.87465262e-01 8.79392087e-01
5.60504496e-01 -1.46657214e-01 3.86575222e-01 -4.55091417e-01
4.73166227e-01 -6.02028146e-02 -5.53191543e-01 4.30164695e-01
-1.02633886e-01 3.97625685e-01 4.57126677e-01 -3.36688161e-01
-1.01005924e+00 7.06278622e-01 -1.73018485e-01 -2.83458918e-01
-2.06273273e-01 3.76482308e-01 -6.01303339e-01 2.70771980e-01
2.70023227e-01 2.91720420e-01 -1.81529313e-01 -4.92815822e-01
2.61572182e-01 4.64267999e-01 5.44608235e-01 -6.82587206e-01
7.30663240e-01 1.54757947e-01 1.11641333e-01 -5.81010401e-01
-9.36466694e-01 -3.98107141e-01 -1.06368780e+00 -1.25286609e-01
1.49710083e+00 -9.71332848e-01 -8.39344680e-01 4.15079415e-01
-1.48151278e+00 -9.37183201e-03 1.41454116e-01 3.93954158e-01
-6.12546742e-01 3.85922998e-01 -2.91396976e-01 -7.31222868e-01
-3.24017435e-01 -1.66995239e+00 1.57168579e+00 4.30266798e-01
-2.41739631e-01 -7.24661231e-01 -1.94566369e-01 4.29251015e-01
5.25845625e-02 6.78509623e-02 6.58939779e-01 -6.88817739e-01
-5.51990747e-01 -3.42241585e-01 -3.95447373e-01 3.10453605e-02
2.13451236e-02 -3.59407216e-01 -9.72750008e-01 -2.55166084e-01
2.19229415e-01 -3.80268246e-01 9.26035702e-01 4.17210549e-01
1.44039118e+00 4.35166806e-02 -6.09067321e-01 6.13793373e-01
9.35943186e-01 1.31642163e-01 4.53395754e-01 2.56497145e-01
7.87087739e-01 8.69623840e-01 6.13769889e-01 3.46818864e-01
6.68215096e-01 1.07749736e+00 4.83276248e-01 -2.07014773e-02
-7.11184293e-02 -1.85617059e-01 3.38650912e-01 6.69989824e-01
-1.67357847e-02 3.80017497e-02 -6.97980225e-01 1.62846074e-01
-2.07885790e+00 -5.06042361e-01 1.48815103e-02 2.06484389e+00
6.91589892e-01 2.56163806e-01 1.15667135e-02 -1.33750364e-01
7.35807240e-01 2.56236553e-01 -5.85267723e-01 2.59562805e-02
3.52554739e-01 -4.67161201e-02 4.12936628e-01 -1.47188250e-02
-1.50893068e+00 8.04692328e-01 4.65007687e+00 8.37590754e-01
-8.76197159e-01 -4.44150865e-02 8.01974058e-01 -9.76649374e-02
2.60664314e-01 -2.72697210e-01 -1.05987620e+00 3.95772219e-01
6.44060194e-01 4.81336653e-01 2.93674737e-01 9.14412618e-01
-9.39021260e-02 -9.16643590e-02 -1.38511574e+00 1.11629140e+00
-1.98559806e-01 -6.88833237e-01 -1.79532245e-01 5.09976270e-03
2.98477799e-01 -3.63338798e-01 -2.58508734e-02 4.63894099e-01
1.08706672e-02 -8.78682911e-01 8.26832950e-01 4.41520900e-01
3.59067470e-01 -7.11964667e-01 7.23202705e-01 6.12474918e-01
-1.76808035e+00 -1.70359135e-01 -3.57097000e-01 1.36075929e-01
3.13607842e-01 4.09754425e-01 -3.83090377e-01 6.48634970e-01
7.52064347e-01 5.97945452e-01 -7.65677273e-01 9.01637137e-01
-5.35996795e-01 4.15493036e-03 -1.65765181e-01 -1.29875526e-01
1.77498117e-01 3.07691116e-02 3.40859026e-01 9.27730203e-01
2.80591279e-01 -2.63686210e-01 4.78022754e-01 7.77984381e-01
-1.89541019e-02 2.60799434e-02 -2.94940203e-01 7.04549849e-02
4.21378762e-01 1.34641755e+00 -9.40503061e-01 -1.05382074e-02
-3.78003150e-01 1.19035065e+00 7.50012636e-01 -3.42403427e-02
-1.09855759e+00 -3.21276128e-01 6.39586449e-01 -1.51046708e-01
4.95949626e-01 -4.41437125e-01 -3.67318571e-01 -9.94123697e-01
3.92575890e-01 -6.30626917e-01 2.34534428e-01 -2.60946572e-01
-1.08523309e+00 6.95788622e-01 -6.03632927e-02 -1.21359515e+00
-3.16798568e-01 -8.09614062e-01 -4.22937155e-01 8.15874696e-01
-1.45459926e+00 -1.34541237e+00 -3.44570100e-01 5.79880476e-01
7.93761849e-01 1.69260412e-01 6.56786621e-01 2.40377218e-01
-9.68996346e-01 6.44440591e-01 -5.20016193e-01 2.95994520e-01
4.64043170e-01 -1.04321051e+00 4.64249074e-01 8.53568077e-01
2.12700948e-01 6.03449404e-01 3.50224376e-01 -6.42036498e-01
-1.62264979e+00 -9.26029742e-01 7.45458364e-01 -2.72368461e-01
2.35279977e-01 -5.82190633e-01 -7.88538635e-01 3.41335565e-01
-5.92349693e-02 7.92295337e-02 4.29030448e-01 1.77947938e-01
-1.77192017e-02 -5.12714125e-02 -6.32539690e-01 5.34100890e-01
1.15611446e+00 -3.40840399e-01 -5.15593052e-01 2.39226162e-01
6.47178352e-01 -4.77573812e-01 -7.06011772e-01 4.40730155e-01
6.31575942e-01 -9.44982409e-01 1.04111612e+00 -3.57923239e-01
5.61633825e-01 -3.48041385e-01 -2.44861901e-01 -1.00656533e+00
-1.95195332e-01 -2.82287300e-01 -1.84485510e-01 1.08315241e+00
3.68102849e-01 -3.60673785e-01 8.01491439e-01 7.78765261e-01
-3.66003029e-02 -1.17821074e+00 -1.08658111e+00 -7.63743818e-01
-2.66468167e-01 -4.28093463e-01 4.42362845e-01 2.86391377e-01
-3.62918139e-01 6.08237565e-01 -6.31071270e-01 3.36891741e-01
4.64130133e-01 7.26132765e-02 8.79067719e-01 -1.34264266e+00
-4.61386800e-01 -4.95978415e-01 -2.24326536e-01 -1.53476918e+00
2.62765586e-01 -5.55080354e-01 3.85466099e-01 -1.52810705e+00
2.85985857e-01 -2.77463228e-01 -3.11972529e-01 5.28790832e-01
-5.47083020e-01 -4.74705994e-02 2.24832878e-01 2.01343447e-01
-7.45352149e-01 8.17632914e-01 1.49351072e+00 -3.31753306e-02
-1.85657218e-01 2.00205296e-01 -8.40228796e-01 8.42520535e-01
4.69895095e-01 -3.35878819e-01 -5.01315951e-01 -7.24243760e-01
1.06365867e-01 9.61664692e-02 3.88206363e-01 -1.27483177e+00
4.08307821e-01 9.88933742e-02 6.47882342e-01 -6.17324591e-01
6.32937670e-01 -1.00885093e+00 -3.56855690e-01 6.01783097e-01
-3.56797248e-01 5.00937067e-02 4.22003530e-02 8.27733457e-01
-2.83000290e-01 -2.22242519e-01 6.43398643e-01 -1.23235531e-01
-9.31382596e-01 3.92172217e-01 5.83078340e-02 -1.26633793e-01
1.12747407e+00 -2.33253166e-01 4.11288477e-02 -1.30825177e-01
-6.76772356e-01 4.42737907e-01 -7.56848976e-02 5.42858541e-01
6.59899592e-01 -1.08259714e+00 -3.62372994e-01 2.29724735e-01
2.91526411e-02 3.34412307e-01 2.88199514e-01 9.38866258e-01
-1.39988050e-01 6.09660327e-01 -3.30451280e-01 -5.94410479e-01
-1.08681285e+00 4.69055355e-01 2.83135206e-01 -8.14287126e-01
-2.55187899e-01 1.25320292e+00 5.85604429e-01 -3.70695204e-01
5.01719594e-01 -2.85115868e-01 -4.68579173e-01 1.55136019e-01
4.40513790e-01 5.38744181e-02 8.26637670e-02 -5.53736687e-01
-4.44556266e-01 6.69735789e-01 -9.35756862e-02 2.97417045e-01
1.39786339e+00 -2.48160437e-01 1.55755877e-01 1.23159349e-01
1.25784647e+00 -4.92887229e-01 -1.63342083e+00 -2.36941576e-01
1.14145719e-01 -2.59884983e-01 -2.31863149e-02 -5.23037136e-01
-1.43824589e+00 1.13774157e+00 6.59505963e-01 -2.81381272e-02
1.22396028e+00 3.90099771e-02 8.84887457e-01 4.24750715e-01
4.45837826e-01 -1.18588996e+00 3.64031434e-01 3.32124114e-01
7.94531345e-01 -1.25062537e+00 -1.70279220e-02 -7.63412833e-01
-7.07687557e-01 1.24640858e+00 1.14215481e+00 -1.93769392e-02
3.90030652e-01 2.34160349e-01 -2.84525126e-01 -3.15663815e-01
-5.86404622e-01 -2.06663296e-01 7.27114379e-01 4.40239519e-01
5.87149978e-01 -5.32992221e-02 -2.58866459e-01 1.12054181e+00
-3.69472951e-02 -2.18780145e-01 -3.73133361e-01 9.22720790e-01
-4.95589733e-01 -1.24184680e+00 -2.61957675e-01 4.16008443e-01
-2.02257141e-01 2.93820769e-01 -2.17721030e-01 7.85207212e-01
9.43180248e-02 7.03424335e-01 -7.19785988e-02 -5.78607082e-01
3.67430538e-01 1.13200493e-01 4.01086420e-01 -5.26150286e-01
-6.09092712e-01 1.91708669e-01 -1.48782749e-02 -8.25842261e-01
-3.62834960e-01 -4.93049264e-01 -1.18508041e+00 3.82206440e-01
-3.72584701e-01 -3.06988835e-01 6.04400098e-01 1.32894576e+00
2.66074389e-01 9.97911930e-01 2.91827142e-01 -1.07457113e+00
-7.72096097e-01 -9.56048250e-01 -4.50515389e-01 2.14710191e-01
-2.52881080e-01 -1.12092352e+00 2.64412999e-01 -1.70301437e-01] | [7.510929584503174, -0.38988006114959717] |
8421872a-aa79-4f26-849a-1c3142296441 | towards-nonlinear-motion-aware-and-occlusion | 2303.18125 | null | https://arxiv.org/abs/2303.18125v2 | https://arxiv.org/pdf/2303.18125v2.pdf | Towards Nonlinear-Motion-Aware and Occlusion-Robust Rolling Shutter Correction | This paper addresses the problem of rolling shutter correction in complex nonlinear and dynamic scenes with extreme occlusion. Existing methods suffer from two main drawbacks. Firstly, they face challenges in estimating the accurate correction field due to the uniform velocity assumption, leading to significant image correction errors under complex motion. Secondly, the drastic occlusion in dynamic scenes prevents current solutions from achieving better image quality because of the inherent difficulties in aligning and aggregating multiple frames. To tackle these challenges, we model the curvilinear trajectory of pixels analytically and propose a geometry-based Quadratic Rolling Shutter (QRS) motion solver, which precisely estimates the high-order correction field of individual pixel. Besides, to reconstruct high-quality occlusion frames in dynamic scenes, we present a 3D video architecture that effectively Aligns and Aggregates multi-frame context, namely, RSA^2-Net. We evaluate our method across a broad range of cameras and video sequences, demonstrating its significant superiority. Specifically, our method surpasses the state-of-the-arts by +4.98, +0.77, and +4.33 of PSNR on Carla-RS, Fastec-RS, and BS-RSC datasets, respectively. | ['Xuelong Li', 'Bin Zhao', 'Dong Wang', 'Zhigang Wang', 'Yizhen Lao', 'Delin Qu'] | 2023-03-31 | null | null | null | null | ['unrolling'] | ['computer-vision'] | [ 3.44822466e-01 -7.47558355e-01 5.40760793e-02 -1.84138328e-01
-7.82283545e-01 -3.22512358e-01 1.57620400e-01 -4.98361766e-01
-4.46170986e-01 6.08059585e-01 1.01386584e-01 -6.17617704e-02
1.29976481e-01 -3.03195059e-01 -7.88329244e-01 -6.41823471e-01
2.29683191e-01 -1.70864299e-01 5.16367733e-01 -1.66968629e-01
2.65302747e-01 3.50087792e-01 -1.28557169e+00 1.02294706e-01
9.90208268e-01 9.87995207e-01 3.44732136e-01 7.69196332e-01
1.51427254e-01 9.34153557e-01 -2.57481784e-01 -3.29166770e-01
4.84703779e-01 -3.68717104e-01 -3.40992630e-01 2.99031019e-01
8.69547844e-01 -8.23224783e-01 -5.04561901e-01 1.04056573e+00
3.88377935e-01 1.92516863e-01 6.21223375e-02 -1.00007200e+00
-2.31549501e-01 -2.27078810e-01 -1.01597536e+00 3.25813740e-01
4.97699022e-01 2.38315046e-01 5.97518682e-01 -1.03548133e+00
8.43845189e-01 1.07912600e+00 8.21273446e-01 3.42761904e-01
-1.08895195e+00 -5.71630478e-01 1.52892873e-01 3.53448719e-01
-1.36153042e+00 -7.11920440e-01 6.55203819e-01 -4.63766813e-01
7.78469384e-01 4.27329272e-01 6.99911714e-01 6.91674054e-01
3.57805729e-01 5.99824250e-01 1.01219773e+00 -1.00398622e-01
8.89152754e-03 -5.45392394e-01 -1.05859138e-01 4.37016726e-01
7.78877661e-02 2.34778702e-01 -7.32904851e-01 7.25497603e-02
1.10219634e+00 2.48588286e-02 -5.88341951e-01 -3.65817875e-01
-1.30996728e+00 3.62723559e-01 1.78418726e-01 -1.06816702e-01
-3.60048205e-01 3.10459912e-01 2.53191918e-01 -2.16655713e-02
5.75861633e-01 8.73933434e-02 -4.31586176e-01 -3.21645498e-01
-1.28477633e+00 3.99476707e-01 4.44632828e-01 1.03234100e+00
6.08293712e-01 3.23239207e-01 -3.50572020e-02 7.89168656e-01
2.15081170e-01 5.63416779e-01 6.98602945e-02 -1.42116094e+00
7.05608070e-01 4.75487523e-02 4.02904809e-01 -1.25755835e+00
-3.27494740e-01 -4.67409015e-01 -9.53440011e-01 1.61012858e-01
4.33144301e-01 -1.67061773e-03 -6.37668312e-01 1.39397275e+00
5.55537224e-01 5.86681902e-01 -1.76540300e-01 1.24720824e+00
6.10477030e-01 8.63447964e-01 -2.59385735e-01 -7.15312600e-01
1.13624525e+00 -1.22133338e+00 -1.05791056e+00 -3.08780253e-01
1.79708973e-01 -1.07695055e+00 8.04307163e-01 5.49167335e-01
-1.42897773e+00 -4.96825844e-01 -1.00275505e+00 -3.22346419e-01
5.88581204e-01 5.31388260e-02 3.35993707e-01 2.55708277e-01
-1.03738427e+00 5.04337668e-01 -9.74362731e-01 7.43172783e-03
3.64164293e-01 1.78375140e-01 -1.51514322e-01 -4.80915427e-01
-7.17501938e-01 6.88410521e-01 -9.06585827e-02 2.33947933e-01
-6.89827383e-01 -1.06623149e+00 -7.55537629e-01 -5.71532287e-02
6.85721159e-01 -7.06232607e-01 1.12477779e+00 -9.52917159e-01
-1.55986011e+00 4.99294668e-01 -5.61192513e-01 -3.12939793e-01
9.44699645e-01 -6.62409306e-01 -3.10756326e-01 2.25166887e-01
1.04914494e-01 4.67859894e-01 8.32369983e-01 -1.14260530e+00
-6.00313067e-01 -1.72518432e-01 1.29002715e-02 4.31545377e-01
-4.55924012e-02 1.77511796e-01 -9.95059192e-01 -1.03976393e+00
3.75686526e-01 -1.01938069e+00 -4.04619128e-01 3.42571139e-01
-3.10332686e-01 6.07284904e-01 8.87414753e-01 -9.53220606e-01
1.36483073e+00 -2.25241327e+00 1.50696710e-01 -2.88648188e-01
2.38923192e-01 3.79007310e-01 9.21937004e-02 8.19490179e-02
-6.33368408e-03 -4.24969107e-01 -3.27770352e-01 -4.72140521e-01
-4.89032000e-01 2.34575197e-01 -3.05070788e-01 7.76634097e-01
-4.30436432e-02 6.52939141e-01 -8.84281099e-01 -3.33054721e-01
5.16861916e-01 7.30380356e-01 -8.86563838e-01 7.96926022e-02
8.11634213e-02 7.62571871e-01 -3.21160883e-01 5.82622051e-01
1.03676903e+00 -1.95341751e-01 1.08356938e-01 -4.37501669e-01
-4.89933670e-01 5.11404872e-02 -1.62017536e+00 1.85083866e+00
-3.06282371e-01 9.52492177e-01 3.46666455e-01 -6.06245637e-01
4.94255006e-01 9.73735303e-02 8.18817317e-01 -8.85334730e-01
-1.07045211e-01 2.89377451e-01 -4.11047190e-01 -5.27707994e-01
7.62301505e-01 1.78296231e-02 4.20879960e-01 1.15077898e-01
-5.11959672e-01 -2.32782125e-01 2.27579892e-01 1.03787027e-01
9.72581863e-01 3.91534984e-01 1.57312050e-01 -1.64342985e-01
6.24428034e-01 -2.05723315e-01 1.12356937e+00 4.56545919e-01
-3.12179953e-01 1.07666218e+00 3.01617026e-01 -7.09021032e-01
-1.00369895e+00 -9.81499434e-01 -1.19222738e-01 6.68347538e-01
6.16418123e-01 -5.05657017e-01 -6.28556848e-01 -6.22148961e-02
-2.04372853e-01 4.19134080e-01 -8.00515339e-02 2.42026523e-01
-1.02597785e+00 -8.18309426e-01 9.24027897e-03 4.39600289e-01
6.47529840e-01 -3.94791037e-01 -7.11563706e-01 4.67347831e-01
-6.83685005e-01 -1.57748592e+00 -8.30436289e-01 -4.56860095e-01
-1.01147199e+00 -1.24004400e+00 -6.40119731e-01 -4.52991337e-01
6.03730381e-01 9.25390065e-01 1.07517374e+00 1.97640955e-01
-3.42053860e-01 8.62595066e-02 -1.23003609e-01 1.85489550e-01
-1.22505888e-01 -5.34003794e-01 2.01681759e-02 1.16216257e-01
-3.85230392e-01 -3.25135738e-01 -9.19218719e-01 7.77828455e-01
-9.74646568e-01 5.27308822e-01 2.03053623e-01 7.74240494e-01
7.97055304e-01 -6.23608045e-02 4.06561717e-02 -5.95672250e-01
-5.18440045e-02 -1.50737390e-01 -9.63065028e-01 6.97260723e-02
-4.25524980e-01 -3.33718538e-01 4.67902690e-01 -2.37206742e-01
-1.22391164e+00 3.15843999e-01 -9.38178673e-02 -5.61576009e-01
4.28771138e-01 1.31961349e-02 3.27986921e-03 -3.55073422e-01
3.40737075e-01 1.80834696e-01 -1.26891434e-01 -2.88080066e-01
2.48683959e-01 1.74320072e-01 8.13169718e-01 -3.01464856e-01
8.68841887e-01 1.00180244e+00 1.69515252e-01 -8.51707220e-01
-8.34012806e-01 -5.32246232e-01 -4.20940220e-01 -4.21099186e-01
9.40241814e-01 -1.35070598e+00 -7.06252277e-01 9.42098975e-01
-1.30842423e+00 -3.09431106e-01 -1.17912620e-01 6.14378989e-01
-6.01451933e-01 7.23516405e-01 -9.22342777e-01 -4.18715179e-01
-3.30069005e-01 -1.52969348e+00 8.78413856e-01 1.40539497e-01
6.48532957e-02 -6.39207721e-01 -1.27237424e-01 5.17180920e-01
5.73880553e-01 3.83920997e-01 2.73465633e-01 6.25434279e-01
-1.17560387e+00 4.24536318e-02 -3.95108521e-01 4.04625058e-01
-8.07348348e-04 3.11407954e-01 -7.29365587e-01 -4.62642580e-01
2.82563150e-01 1.76138505e-01 5.71581483e-01 7.54317522e-01
1.07283115e+00 -2.00215086e-01 -5.20951860e-02 1.28529656e+00
1.58259118e+00 1.89745218e-01 8.43857050e-01 3.66205037e-01
9.60634887e-01 2.70309687e-01 9.92896736e-01 7.38667190e-01
5.12132168e-01 1.13633645e+00 5.34966826e-01 -4.86874208e-02
-3.22747827e-01 1.02190167e-01 4.02354598e-01 9.92969871e-01
-2.53822565e-01 -2.14634001e-01 -6.30762517e-01 3.75208020e-01
-1.86402166e+00 -9.34133410e-01 -7.58885026e-01 2.17732978e+00
7.42478609e-01 -1.63614750e-01 -1.83144465e-01 -4.12857905e-02
7.25627542e-01 4.55083936e-01 -5.47483146e-01 6.26890212e-02
-3.81552935e-01 -1.52152091e-01 8.09734583e-01 6.35037303e-01
-9.71842885e-01 8.20507526e-01 6.16748714e+00 7.94931233e-01
-1.19955039e+00 2.60315329e-01 6.78809941e-01 -4.63341117e-01
-6.69689290e-03 1.42767832e-01 -6.79835021e-01 6.07033730e-01
5.63640356e-01 8.61345232e-03 7.28457153e-01 4.35930163e-01
6.10965073e-01 -4.14860159e-01 -6.62697732e-01 1.28586411e+00
1.68347791e-01 -1.50682676e+00 -3.59456539e-01 -4.52764295e-02
1.14801836e+00 9.61559564e-02 7.51212239e-02 -3.28367621e-01
-5.62834851e-02 -6.35621548e-01 1.03396094e+00 4.82480079e-01
8.76896739e-01 -5.84160566e-01 3.92513245e-01 1.56048551e-01
-1.18408430e+00 7.84934983e-02 -3.36364418e-01 -1.23614594e-01
6.90231681e-01 9.06667769e-01 -1.03220671e-01 6.22478247e-01
8.36747468e-01 1.07790482e+00 -2.44282290e-01 1.14332247e+00
-1.95787683e-01 3.08238298e-01 -3.26959282e-01 6.95980489e-01
2.90117133e-02 -5.64517915e-01 5.76446772e-01 9.64125097e-01
5.67021549e-01 3.99166405e-01 -5.19899949e-02 5.34289181e-01
1.02627642e-01 -1.36623383e-01 -9.50095579e-02 4.51658398e-01
1.46689743e-01 1.10853493e+00 -5.77734828e-01 -3.02988797e-01
-5.67451715e-01 1.20243621e+00 -8.32483321e-02 4.89989161e-01
-1.22312081e+00 1.49545118e-01 9.37785149e-01 3.05147886e-01
2.86073834e-01 -6.31389678e-01 -4.00014520e-01 -1.52862644e+00
3.29306662e-01 -8.19129050e-01 8.15280993e-03 -9.54814792e-01
-7.23500848e-01 4.21617448e-01 -2.84049779e-01 -1.57930326e+00
3.38497721e-02 -2.55004138e-01 -3.66546512e-01 6.89030290e-01
-1.83013511e+00 -7.60735929e-01 -7.56895602e-01 6.87694848e-01
8.10119331e-01 3.12520355e-01 1.29504994e-01 9.53272521e-01
-6.33342385e-01 2.19839811e-01 4.42279905e-01 -2.66404837e-01
9.30631578e-01 -7.08360195e-01 5.14464736e-01 1.20678341e+00
-1.85099378e-01 2.99071521e-01 7.63765812e-01 -5.46739161e-01
-1.80713499e+00 -1.11892331e+00 6.53272629e-01 -2.96178758e-01
4.60847706e-01 -1.38408005e-01 -9.34150279e-01 5.29735029e-01
-2.49257740e-02 5.95489740e-01 1.92676380e-01 -6.00927055e-01
-7.85245374e-02 -3.01718265e-01 -9.03550565e-01 5.78944921e-01
1.24740696e+00 -2.31874436e-01 1.43540144e-01 3.98672193e-01
6.78605258e-01 -1.06381524e+00 -7.46133447e-01 3.08364779e-01
5.99932194e-01 -1.39898336e+00 1.29309082e+00 1.03834055e-01
7.52116978e-01 -5.59224963e-01 -2.29496643e-01 -9.95403409e-01
-1.63743734e-01 -1.04360843e+00 -2.12030992e-01 1.03709483e+00
-2.47951314e-01 -4.48001444e-01 7.05639601e-01 6.22543037e-01
-2.59080619e-01 -6.96181297e-01 -1.14110136e+00 -5.09777904e-01
-5.05259156e-01 -6.40589058e-01 3.42839777e-01 9.72423196e-01
-7.38570094e-01 -1.48349598e-01 -8.49289060e-01 2.57611483e-01
8.17012489e-01 -1.08734993e-02 1.05922198e+00 -5.86381853e-01
-4.17714983e-01 -6.50239596e-03 -2.95460612e-01 -1.53288436e+00
-1.71908230e-01 -1.79189339e-01 1.59231395e-01 -1.37523830e+00
6.03436902e-02 -5.34036875e-01 4.52485383e-02 -1.23282820e-01
-3.74584764e-01 5.01614928e-01 4.77779061e-01 4.09798801e-01
-6.70301974e-01 4.08479035e-01 1.52753770e+00 1.76723361e-01
-1.45581692e-01 -1.22229256e-01 -3.10455084e-01 8.24157894e-01
3.42071474e-01 -3.65691215e-01 -2.60867000e-01 -1.05264771e+00
3.63647044e-01 5.07652760e-01 5.22425830e-01 -1.04137611e+00
3.00901681e-01 -3.02146852e-01 2.98330009e-01 -8.29793394e-01
5.59355915e-01 -8.10353756e-01 6.36679053e-01 3.74201775e-01
5.89811020e-02 3.79729301e-01 1.25713468e-01 6.19251072e-01
-2.39660770e-01 1.03979252e-01 1.04786003e+00 1.44226044e-01
-7.92396486e-01 4.48615402e-01 1.71505641e-02 1.86704010e-01
1.04257405e+00 -4.49270010e-01 -4.04354721e-01 -4.19148892e-01
-2.53133327e-01 1.56064793e-01 8.42267811e-01 3.41314048e-01
7.41981685e-01 -1.23849130e+00 -6.72200739e-01 4.24054712e-01
-2.31456459e-01 3.45853090e-01 7.42222011e-01 1.30632079e+00
-1.21529543e+00 5.49769253e-02 4.78403531e-02 -8.15795422e-01
-1.31469190e+00 3.18663359e-01 1.41785607e-01 -1.22095965e-01
-9.44174409e-01 7.78883994e-01 5.06857932e-01 1.75278544e-01
-1.00304298e-01 -4.17962223e-01 2.80574650e-01 -1.41751811e-01
6.18500888e-01 7.64863551e-01 2.84047537e-02 -8.79285097e-01
-3.28584999e-01 9.79887664e-01 1.11772947e-01 4.97043356e-02
1.40442359e+00 -4.84544843e-01 -5.05814813e-02 -2.33625937e-02
1.21249402e+00 1.45393595e-01 -2.03415394e+00 -2.77747422e-01
-4.53594983e-01 -1.17940557e+00 1.59114018e-01 -3.97843540e-01
-1.50924075e+00 6.81948960e-01 4.97756213e-01 -3.78311485e-01
1.30074489e+00 -6.15462959e-01 1.21396375e+00 -1.38112247e-01
3.45016927e-01 -1.07170796e+00 -7.21276924e-02 4.86704558e-01
7.14212358e-01 -1.13794434e+00 5.66723406e-01 -8.09972703e-01
-5.94662488e-01 1.13170433e+00 3.80438566e-01 -1.68336764e-01
2.49821812e-01 3.17951620e-01 1.16618022e-01 1.47021011e-01
-5.33463180e-01 4.37002540e-01 1.96929395e-01 2.50729293e-01
2.85337389e-01 -2.02301756e-01 -3.01158875e-01 -1.79505184e-01
1.45669386e-01 1.77930415e-01 8.07207465e-01 9.11073983e-01
-1.17725007e-01 -6.99583769e-01 -6.33346379e-01 9.87172499e-02
-5.82351446e-01 -1.18381582e-01 4.96449202e-01 6.62445188e-01
-1.05583146e-01 9.74394560e-01 -8.93456861e-03 2.53620022e-03
4.41858351e-01 -7.20153213e-01 3.53846610e-01 -1.80334881e-01
-3.05938870e-01 4.46161956e-01 -1.41375363e-01 -1.13304627e+00
-7.21315861e-01 -8.01549256e-01 -1.14621174e+00 -5.30551672e-01
-3.71372193e-01 -3.80304009e-01 7.75779307e-01 6.63823962e-01
4.56985563e-01 5.80414832e-01 6.97057545e-01 -1.20674813e+00
-3.76121938e-01 -3.64976376e-01 -4.72443670e-01 4.32652414e-01
5.06800234e-01 -5.96546113e-01 -4.09108281e-01 2.95033157e-01] | [10.780856132507324, -1.8073160648345947] |
e89898f6-e83b-4940-82f9-b14536a66625 | transfer-learning-for-underrepresented-music | 2306.00281 | null | https://arxiv.org/abs/2306.00281v1 | https://arxiv.org/pdf/2306.00281v1.pdf | Transfer Learning for Underrepresented Music Generation | This paper investigates a combinational creativity approach to transfer learning to improve the performance of deep neural network-based models for music generation on out-of-distribution (OOD) genres. We identify Iranian folk music as an example of such an OOD genre for MusicVAE, a large generative music model. We find that a combinational creativity transfer learning approach can efficiently adapt MusicVAE to an Iranian folk music dataset, indicating potential for generating underrepresented music genres in the future. | ['Matthew Guzdial', 'Anahita Doosti'] | 2023-06-01 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 1.12856366e-01 1.44507438e-01 5.25746532e-02 1.66175336e-01
-8.77449632e-01 -8.44695389e-01 5.20535827e-01 -7.24189222e-01
2.75892951e-02 7.37534285e-01 5.59154451e-01 3.43524367e-02
-4.93507326e-01 -1.01388729e+00 -8.46085846e-01 -3.81557792e-01
9.30731650e-03 1.08895743e+00 -6.31925464e-01 -4.43265885e-01
3.40289563e-01 2.34039858e-01 -1.20469034e+00 6.69917226e-01
4.97819543e-01 4.51477170e-01 6.62202016e-02 7.33963370e-01
6.61606938e-02 3.92442077e-01 -1.40262854e+00 -3.36175352e-01
4.93332654e-01 -9.51509655e-01 -7.30526090e-01 -2.83706725e-01
6.58466816e-01 6.78893551e-02 -1.55289635e-01 7.62903512e-01
1.04524505e+00 3.41973871e-01 1.17254889e+00 -8.62070739e-01
-1.30165720e+00 1.57071006e+00 2.84310896e-02 -1.34986416e-02
1.50905401e-01 3.15033853e-01 1.05864751e+00 -6.96645439e-01
7.72930980e-01 1.15645671e+00 9.93017673e-01 6.72957182e-01
-1.41318464e+00 -1.28017068e+00 -8.93242419e-01 -3.57313901e-02
-1.31280589e+00 -2.48991996e-01 1.45043945e+00 -3.72022390e-01
1.01906288e+00 5.28486222e-02 1.26616418e+00 1.33890331e+00
8.20965990e-02 8.37189615e-01 8.18129480e-01 -4.79215890e-01
3.15859884e-01 -1.23969525e-01 -8.89504135e-01 9.07133147e-02
1.51022822e-01 5.38593531e-01 -1.15040421e+00 -7.93627948e-02
1.04191673e+00 -9.11130905e-01 2.83997357e-01 1.23204656e-01
-1.36715233e+00 1.17547691e+00 4.61021692e-01 5.40239513e-01
-2.70842910e-01 4.33904767e-01 5.63655376e-01 4.86366630e-01
3.28132510e-01 1.47559094e+00 -5.60106099e-01 -5.22498667e-01
-1.24015915e+00 6.81557894e-01 5.83777726e-01 8.21580768e-01
3.51158857e-01 1.10798240e+00 -2.81327933e-01 1.15355837e+00
-1.32591262e-01 5.03918409e-01 9.36188936e-01 -1.21968377e+00
2.32217103e-01 2.58710802e-01 -4.03725475e-01 -5.19305289e-01
-5.84781431e-02 -6.64900839e-01 -9.12879288e-01 1.96149543e-01
1.28822833e-01 -2.82733411e-01 -7.11565197e-01 1.65854132e+00
-2.91537046e-01 1.51210204e-01 4.02397901e-01 4.22520906e-01
9.78282869e-01 7.31453896e-01 -1.57145262e-01 3.31095457e-02
7.62068212e-01 -8.81727517e-01 -3.73026371e-01 2.63216585e-01
2.08185136e-01 -9.27486062e-01 1.42745221e+00 7.75697589e-01
-1.28649974e+00 -1.11842370e+00 -9.66780305e-01 2.64148086e-01
-2.11539894e-01 3.26225489e-01 9.29085612e-01 7.12134421e-01
-8.21283758e-01 8.44090223e-01 -6.68526366e-02 -8.70014131e-02
6.66373253e-01 3.92589509e-01 2.09520698e-01 4.59774196e-01
-1.18302357e+00 4.60542023e-01 1.04011297e+00 -3.32461268e-01
-1.10081780e+00 -9.72788393e-01 -3.87216657e-01 1.19107313e-01
-4.42590803e-01 -7.59093344e-01 1.30359602e+00 -1.42175901e+00
-1.86852169e+00 7.54680455e-01 7.92251229e-01 -4.16070998e-01
9.45657790e-02 9.62835699e-02 -4.62201536e-01 -1.90931797e-01
4.67637070e-02 1.15135205e+00 9.69820559e-01 -9.88774061e-01
-2.52624631e-01 9.03008729e-02 -5.08407831e-01 2.74487048e-01
-5.40675342e-01 -2.91873991e-01 2.32049882e-01 -1.52319872e+00
-2.73545891e-01 -1.14564919e+00 4.04083550e-01 -7.39598811e-01
-2.55233496e-01 -5.27251840e-01 3.95712733e-01 -4.14715827e-01
1.00578403e+00 -2.00678372e+00 4.07818317e-01 1.81177840e-01
-4.47293609e-01 1.19915456e-01 -5.07601082e-01 3.85583133e-01
-9.40291584e-02 1.63901761e-01 1.44810766e-01 1.15759008e-01
1.80541247e-01 1.09660752e-01 -4.62499440e-01 -5.21670759e-01
2.58081317e-01 1.25165212e+00 -7.67867684e-01 -1.19472072e-01
-3.18646520e-01 5.62878907e-01 -7.67107487e-01 -5.15380800e-02
-5.73275566e-01 7.56575882e-01 6.79997578e-02 5.67207396e-01
1.31300151e-01 4.76183683e-01 -8.53970721e-02 -9.40957945e-03
6.27477840e-02 2.08580762e-01 -9.22786355e-01 2.10707569e+00
-3.36471707e-01 7.18301594e-01 -7.13855028e-01 -5.23197889e-01
1.41537297e+00 4.86647397e-01 3.08566868e-01 -5.10596871e-01
2.43539274e-01 3.88851881e-01 6.43543303e-01 -8.39349404e-02
6.73843265e-01 -7.12402701e-01 -4.05166775e-01 6.17373884e-01
4.59498167e-01 -6.32815599e-01 1.13277934e-01 -4.68101144e-01
7.78966308e-01 4.73603845e-01 2.50558890e-02 -3.02336395e-01
-3.57198250e-03 2.48876646e-01 3.90206635e-01 7.55079389e-01
2.63707608e-01 6.64576232e-01 1.55357286e-01 -5.59418678e-01
-1.34559500e+00 -1.21763718e+00 6.64152950e-02 1.24283612e+00
-9.43534732e-01 -2.77645290e-01 -8.07817996e-01 -2.22652033e-01
-1.05903573e-01 1.20641005e+00 -7.45392799e-01 -4.73016560e-01
-5.07004440e-01 -5.21264970e-01 1.56160629e+00 6.26036108e-01
4.76964623e-01 -2.18197107e+00 -3.09164703e-01 5.22271395e-01
-1.77041322e-01 -2.38077030e-01 -5.93473911e-01 1.28743768e-01
-1.07944453e+00 -5.54571271e-01 -1.21483052e+00 -1.25738907e+00
-2.73791313e-01 -5.74415982e-01 1.30418110e+00 -7.51989841e-01
-2.08989635e-01 2.55759150e-01 -3.58853191e-01 -1.04854715e+00
-1.08785594e+00 4.42432880e-01 8.73349700e-03 -4.79050934e-01
4.63134170e-01 -7.92874753e-01 -1.14973553e-01 -1.90821782e-01
-8.59094501e-01 1.47401631e-01 6.19973242e-01 8.28799546e-01
6.97816074e-01 5.19038498e-01 1.15489089e+00 -5.58212936e-01
1.23900759e+00 -1.19211666e-01 -2.03338981e-01 -4.28164303e-02
-4.77585137e-01 -1.26916379e-01 4.98698413e-01 -1.16635299e+00
-1.04240310e+00 -1.18159251e-02 3.49045619e-02 -6.73712730e-01
-2.52342105e-01 5.75958014e-01 1.90418512e-02 2.30181798e-01
1.06149042e+00 6.10636115e-01 -2.88766176e-01 -4.95575398e-01
4.72600818e-01 7.20087290e-01 6.75494194e-01 -9.16036367e-01
8.18510175e-01 -1.65302858e-01 -4.18846942e-02 -8.71734619e-01
-3.85751516e-01 2.31599167e-01 -5.26104093e-01 -2.58551210e-01
8.38188171e-01 -9.23443675e-01 -3.30375463e-01 4.52259839e-01
-8.46251130e-01 -7.13061452e-01 -1.13611674e+00 5.40448010e-01
-1.29809713e+00 -5.05907357e-01 -7.45065570e-01 -5.22557378e-01
-8.87838125e-01 -6.62861228e-01 9.90743697e-01 1.88092396e-01
-8.39913607e-01 -7.07459688e-01 9.54116225e-01 2.98606843e-01
3.88334692e-01 6.51661009e-02 1.33948052e+00 -5.60008287e-01
-2.16838017e-01 -1.36247287e-02 3.98931384e-01 3.57614607e-01
9.04965848e-02 -1.94104567e-01 -9.39986825e-01 -5.83767146e-02
-3.84777993e-01 -7.04780281e-01 6.47783577e-01 4.88986701e-01
7.45223582e-01 -2.55022585e-01 3.90259087e-01 7.78689921e-01
1.02003562e+00 4.52304035e-01 7.92016864e-01 4.02497083e-01
5.55886328e-01 2.50533372e-02 1.94164798e-01 4.17266250e-01
-1.04570232e-01 7.26912141e-01 -2.46258602e-01 1.22897573e-01
-8.82703483e-01 -6.41528368e-01 5.93657911e-01 1.33646309e+00
-5.19680560e-01 -2.73571424e-02 -7.69478500e-01 5.51739216e-01
-1.25075877e+00 -1.33167470e+00 3.75974089e-01 1.90605080e+00
1.27307594e+00 -8.03760290e-02 6.17785633e-01 4.68937427e-01
3.85087401e-01 -4.16884720e-01 -4.98520702e-01 -8.79067600e-01
-5.07032275e-01 1.02172112e+00 -3.07180941e-01 -2.68962979e-01
-9.61309016e-01 1.42766929e+00 7.04548979e+00 1.01662159e+00
-9.10410404e-01 9.44395512e-02 7.53623545e-02 -2.36477867e-01
-2.85639763e-01 -3.87663126e-01 -5.34469664e-01 3.74552816e-01
1.14851522e+00 -1.21952824e-01 1.02874005e+00 6.56772196e-01
-1.10215526e-02 6.64559543e-01 -9.44986284e-01 1.06255019e+00
4.10831183e-01 -1.50398505e+00 6.80720806e-01 3.48072857e-01
1.47567034e+00 -2.26619422e-01 4.57486421e-01 8.16713512e-01
1.45566329e-01 -1.28133345e+00 8.74155223e-01 4.80316460e-01
1.07237470e+00 -1.31919813e+00 3.56819808e-01 -4.16266844e-02
-8.91475618e-01 2.34462935e-02 -5.95266759e-01 -5.65103367e-02
-1.61083475e-01 -1.07554324e-01 -1.24340045e+00 2.20367685e-01
6.07675731e-01 6.85011864e-01 -7.52394915e-01 1.09932232e+00
-3.19800004e-02 1.20113456e+00 -1.10172993e-03 1.22141540e-02
1.15106806e-01 -1.02467045e-01 6.68445706e-01 1.04620159e+00
1.00491822e+00 -3.06805491e-01 -1.87721193e-01 1.31215775e+00
-2.57126600e-01 2.40965351e-01 -8.46642673e-01 -7.09751666e-01
1.23512566e-01 7.31867969e-01 -6.96572185e-01 -3.06397408e-01
2.36550912e-01 1.02425885e+00 -1.16647795e-01 1.71031475e-01
-6.49381638e-01 -5.01534998e-01 3.13921213e-01 -4.85776067e-02
5.27796388e-01 -4.18938696e-02 -5.67396820e-01 -8.54201436e-01
-8.48338544e-01 -1.47786236e+00 3.07582885e-01 -1.36997783e+00
-1.57420027e+00 6.43619061e-01 -1.82674155e-01 -1.50220776e+00
-4.81349766e-01 -3.93372625e-01 -7.65447915e-01 7.54941344e-01
-6.31491959e-01 -1.74995136e+00 2.13141918e-01 4.91812557e-01
8.66435528e-01 -1.26633036e+00 1.48798764e+00 -6.24389425e-02
1.09893940e-01 8.20927024e-01 2.17404038e-01 2.45155364e-01
7.27106929e-01 -1.30933452e+00 3.65139633e-01 -6.09306172e-02
1.06155968e+00 5.23331016e-02 3.73802125e-01 -7.32692242e-01
-9.80327845e-01 -1.20835698e+00 6.20777965e-01 -4.56265569e-01
2.47486785e-01 -4.50758152e-02 -4.62709755e-01 5.75099707e-01
7.41906106e-01 -1.22079909e+00 1.28801453e+00 4.67872262e-01
-3.45084846e-01 9.18968543e-02 -7.93695092e-01 7.85938144e-01
1.05094588e+00 -3.12461436e-01 -1.16543746e+00 1.29317984e-01
4.81936157e-01 -6.51873946e-02 -1.05044770e+00 3.16604972e-01
7.88592279e-01 -5.79868257e-01 9.93768632e-01 -7.61026204e-01
7.18186617e-01 -2.24366352e-01 -1.87035352e-01 -1.92019224e+00
-7.16354370e-01 -8.89251113e-01 -1.13272198e-01 1.47536480e+00
3.45112860e-01 7.29588717e-02 8.69103432e-01 -5.91052294e-01
-3.16405773e-01 7.40442723e-02 -9.23048973e-01 -1.15184987e+00
7.76710868e-01 -4.39946651e-01 8.04878771e-01 1.08280265e+00
-1.43314674e-01 8.98934543e-01 -3.99766862e-01 -5.67476392e-01
4.40495789e-01 2.38322318e-01 9.50696051e-01 -1.48556936e+00
-9.01491284e-01 -9.40773070e-01 -3.30963850e-01 -1.24606760e-02
3.09536278e-01 -1.63172185e+00 -8.06161165e-02 -1.14194894e+00
8.48150104e-02 -3.44975382e-01 -4.17022496e-01 5.86137354e-01
5.47374249e-01 1.07266068e+00 4.85171109e-01 2.12023452e-01
-5.06355651e-02 6.38147295e-01 1.53301501e+00 -4.66003388e-01
-8.28044415e-01 1.68637633e-01 -8.91127825e-01 6.86418533e-01
1.24467874e+00 -7.05725372e-01 -5.70063829e-01 -3.72033179e-01
4.97351497e-01 -9.58824307e-02 -6.24462068e-02 -1.43987858e+00
-4.05191362e-01 -9.87908244e-02 1.01073003e+00 -5.38470030e-01
4.27623868e-01 -1.16648324e-01 7.24319518e-01 5.29465437e-01
-6.35496914e-01 8.97169393e-03 7.17228293e-01 2.00770542e-01
-1.54749840e-01 -2.24599332e-01 5.75498462e-01 -3.49401236e-01
-2.53171384e-01 -1.68177962e-01 -6.22320950e-01 5.00497699e-01
5.40249884e-01 -4.94005680e-01 6.14202349e-03 -5.00923038e-01
-9.25510764e-01 -8.20573151e-01 2.19585717e-01 6.04523778e-01
4.70820427e-01 -1.91136730e+00 -1.28267205e+00 4.90602911e-01
8.44341293e-02 -5.08338809e-01 9.88173038e-02 9.06930789e-02
-4.76533830e-01 3.40254486e-01 -8.16382229e-01 -3.73762287e-02
-9.32542026e-01 3.09035957e-01 7.68269077e-02 -1.66474104e-01
-5.62108576e-01 9.36706543e-01 -1.39666811e-01 -7.12442517e-01
1.37578472e-01 1.60359472e-01 -1.64695978e-01 1.08647004e-01
1.75522849e-01 4.48135257e-01 -2.46321931e-01 -6.12998903e-01
1.75404415e-01 5.51053703e-01 4.78174955e-01 -5.34034252e-01
1.43979859e+00 6.76282108e-01 1.60344362e-01 7.61547744e-01
5.89512646e-01 2.21463144e-01 -6.88020706e-01 4.19072390e-01
-2.09722310e-01 3.98986638e-02 -1.88164130e-01 -1.51068413e+00
-8.00243020e-01 9.13609087e-01 5.57671845e-01 -1.85963690e-01
1.23478949e+00 -1.91999115e-02 8.76956105e-01 6.52995110e-01
2.15948120e-01 -1.23926032e+00 4.93276358e-01 7.84561992e-01
1.42030776e+00 -3.42432141e-01 -2.41200179e-01 5.62644601e-01
-8.73451948e-01 1.31017709e+00 4.75159556e-01 -5.46438694e-01
4.17204708e-01 1.44356489e-01 5.16154170e-02 1.35377020e-01
-6.63126945e-01 1.19120568e-01 6.51317060e-01 9.95061338e-01
4.92275119e-01 3.43808442e-01 -1.06068768e-01 9.65363204e-01
-1.26302743e+00 2.31620714e-01 4.70991641e-01 2.71620363e-01
-2.11958230e-01 -1.17242014e+00 -4.54870909e-01 3.92080575e-01
-2.05409169e-01 -4.59885865e-01 -9.06072438e-01 6.93753600e-01
7.93062091e-01 5.64771175e-01 3.18886697e-01 -5.33939719e-01
2.06466094e-01 6.70781553e-01 1.03069890e+00 -7.78649032e-01
-1.20003617e+00 6.59387112e-01 -9.34936181e-02 3.10316294e-01
-2.82118589e-01 -5.21156907e-01 -8.92553985e-01 -8.58779773e-02
-9.40427855e-02 2.25383472e-02 4.01786983e-01 3.34530145e-01
4.59668249e-01 8.58520567e-01 1.21142857e-01 -1.22328722e+00
-4.28438097e-01 -1.48480201e+00 -1.00577223e+00 3.99544656e-01
-6.60147369e-01 -3.15623134e-01 3.60215902e-02 2.66885936e-01] | [16.040964126586914, 5.508218765258789] |
710f60b1-02f1-4dd4-8f6b-f38ff982d8c2 | generalized-nonconvex-approach-for-low-tubal | null | null | https://ieeexplore.ieee.org/abstract/document/9340243 | https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9340243 | Generalized Nonconvex Approach for Low-Tubal-Rank Tensor Recovery | The tensor-tensor product-induced tensor nuclear norm (t-TNN) (Lu et al., 2020) minimization for low-tubal-rank tensor recovery attracts broad attention recently. However, minimizing the t-TNN faces some drawbacks. For example, the obtained solution could be suboptimal to the original problem due to its loose approximation. In this article, we extract a unified nonconvex surrogate of the tensor tubal rank as a tighter regularizer, which involves many popular nonconvex penalty functions. An iterative reweighted t-TNN algorithm is proposed to solve the resulting generalized nonconvex tubal rank minimization for tensor recovery. It converges to a critical point globally with rigorous proofs based on the Kurdyka-Łojasiwicz property. Furthermore, we provide the theoretical guarantees for exact and robust recovery by developing the tensor null space property. Extensive experiments demonstrate that our approach markedly enhances recovery performance compared with several state-of-the-art convex and nonconvex methods. | ['and Xinling Liu', 'Jianwen Huang', 'TingWen Huang', 'Jianjun Wang', 'Feng Zhang', 'Hailin Wang'] | 2022-08-04 | null | null | null | ieee-transactions-on-neural-networks-and-11 | ['image-inpainting', 'low-rank-matrix-completion'] | ['computer-vision', 'methodology'] | [-2.90503204e-01 -2.55593687e-01 -2.56858796e-01 1.40783668e-01
-9.67333019e-01 -3.58480185e-01 -6.32339045e-02 -4.77556586e-01
-2.24911943e-02 6.53380752e-01 6.13426030e-01 4.73433658e-02
-6.92585409e-01 -1.07897520e-01 -6.67000771e-01 -1.00740123e+00
-3.89869571e-01 2.22215861e-01 -1.59961343e-01 -2.95798630e-01
1.25456110e-01 2.54439265e-01 -9.16527271e-01 4.03548330e-02
1.13738728e+00 9.95738983e-01 2.02348549e-02 2.10668147e-01
4.79849130e-01 7.19806135e-01 8.01729113e-02 -2.11985335e-01
5.31845808e-01 3.54001708e-02 -9.33712602e-01 3.49161774e-01
5.07798612e-01 -4.06307727e-01 -3.47268522e-01 1.46378136e+00
2.88053155e-01 3.82452190e-01 3.58473808e-01 -9.80365992e-01
-7.55644858e-01 5.12803137e-01 -1.19432998e+00 1.08727880e-01
2.16763332e-01 -2.91404873e-01 1.26011968e+00 -1.71047258e+00
6.10920489e-01 9.90130186e-01 8.95743668e-01 -8.56399089e-02
-1.13562548e+00 -2.86426991e-01 -2.71489322e-02 6.15787320e-02
-1.61239302e+00 -2.32827350e-01 7.58758008e-01 -5.38508713e-01
2.99320787e-01 5.14306068e-01 4.90882069e-01 4.58590329e-01
-2.36634299e-01 1.07332945e+00 1.11291039e+00 -5.06946910e-03
-2.51551270e-01 -4.20044631e-01 1.51397452e-01 9.44757998e-01
5.90278029e-01 -7.94990733e-02 -4.21023697e-01 -4.70607281e-01
8.97027552e-01 7.27346074e-03 -6.40079141e-01 -3.06450009e-01
-1.84490144e+00 6.29392028e-01 6.05806768e-01 4.72909778e-01
-4.56640631e-01 2.43412912e-01 4.95147556e-01 1.14604039e-02
7.74833500e-01 2.65676230e-01 -9.13440958e-02 3.03822085e-02
-1.06988847e+00 3.62410337e-01 4.54330832e-01 9.90115285e-01
7.01882124e-01 4.45342779e-01 -2.05799580e-01 9.01763320e-01
4.26784962e-01 6.94889486e-01 3.70288827e-02 -1.20825970e+00
7.37908900e-01 2.30543166e-01 1.54902369e-01 -1.53206646e+00
-2.56822824e-01 -8.10127735e-01 -1.22930062e+00 -3.70359451e-01
5.79341471e-01 1.76535502e-01 -4.40277010e-01 1.31195009e+00
4.63301152e-01 5.64153910e-01 -2.35644013e-01 1.46812928e+00
4.78712767e-01 6.20836675e-01 -4.81687486e-01 -4.33840960e-01
1.10831940e+00 -8.99271429e-01 -9.42177296e-01 2.58617222e-01
7.23407924e-01 -9.63793814e-01 8.13674629e-01 4.78017837e-01
-1.28193474e+00 1.66265909e-02 -7.79047191e-01 -2.33370513e-01
4.92893994e-01 4.11627829e-01 5.84305882e-01 2.49051198e-01
-8.56346428e-01 5.60531020e-01 -7.59102702e-01 1.70806035e-01
1.16414495e-01 1.98867217e-01 -4.39812064e-01 -4.97308850e-01
-8.26735437e-01 5.35519481e-01 7.93394446e-02 8.48138392e-01
-7.78978527e-01 -8.45234096e-01 -5.43849707e-01 -3.31441998e-01
5.97058713e-01 -5.93719304e-01 7.25636542e-01 -3.38595390e-01
-1.29562175e+00 6.91735685e-01 -1.34739667e-01 2.04626158e-01
5.29095471e-01 -4.04235929e-01 -1.54147008e-02 4.09123033e-01
3.46619427e-01 -3.38366717e-01 9.97768998e-01 -1.41769230e+00
-6.37269095e-02 -3.65186065e-01 -8.81866962e-02 1.75104767e-01
-4.60294724e-01 5.07023409e-02 -4.43524152e-01 -1.22385967e+00
7.79192924e-01 -1.11249673e+00 -4.04840142e-01 -1.60945669e-01
-7.63533175e-01 1.27604259e-02 5.36430061e-01 -1.02145159e+00
1.26717293e+00 -2.19747138e+00 7.08645105e-01 5.80438614e-01
7.04009831e-01 3.20680849e-02 -1.49620280e-01 2.67264694e-01
-1.51769549e-01 5.17730266e-02 -4.93770629e-01 -3.23210835e-01
-4.64835130e-02 1.58567414e-01 -3.18787992e-01 1.21624970e+00
-2.38467962e-01 6.50210261e-01 -1.17639768e+00 -3.85139972e-01
-1.21852912e-01 4.62525100e-01 -7.78856575e-01 3.57974432e-02
4.10878122e-01 6.17910266e-01 -6.73586965e-01 9.72508550e-01
9.99798656e-01 -2.64028698e-01 2.19406262e-02 -8.54024470e-01
-3.45181197e-01 -2.17361584e-01 -1.25692821e+00 1.67752218e+00
-5.40288426e-02 3.80079269e-01 9.24177170e-01 -1.21737278e+00
4.98210132e-01 3.77122492e-01 9.67387974e-01 -3.14807519e-02
1.40213460e-01 9.13815618e-01 -2.92129636e-01 -5.23312032e-01
7.83897102e-01 -4.48567033e-01 7.61039183e-02 1.64424270e-01
-3.79976749e-01 4.10747230e-02 3.17533821e-01 2.73717403e-01
9.34486091e-01 2.05206461e-02 -1.19688340e-01 -8.63346756e-01
6.67502463e-01 -1.67904764e-01 8.43714654e-01 2.29309022e-01
-1.60795629e-01 7.28575647e-01 3.90515596e-01 -1.42247796e-01
-1.15108764e+00 -7.34088480e-01 -3.61431211e-01 9.17665541e-01
2.31813475e-01 -2.59842336e-01 -5.45591354e-01 -3.21910799e-01
8.98528546e-02 -7.51843629e-03 -4.21506196e-01 2.39863217e-01
-9.35215652e-01 -9.88993287e-01 3.69416505e-01 2.52250761e-01
4.28822607e-01 -9.00483429e-02 4.21121925e-01 2.18353700e-02
-7.19752967e-01 -1.22642386e+00 -1.15912890e+00 -5.85455418e-01
-1.23473585e+00 -1.05591261e+00 -1.22838771e+00 -7.09982872e-01
1.03741515e+00 8.02696943e-01 7.75508881e-01 3.27802181e-01
7.09803179e-02 4.24452096e-01 -3.01747590e-01 5.65335155e-01
1.02267817e-01 -2.67720699e-01 5.00118375e-01 6.27872705e-01
-5.60872376e-01 -5.38127184e-01 -7.08823502e-01 6.54514074e-01
-1.02835846e+00 -5.99233545e-02 4.69858825e-01 9.48471546e-01
8.99299800e-01 1.16371557e-01 2.45469972e-01 -5.67373097e-01
7.58654594e-01 -5.82773089e-01 -4.31291372e-01 1.46710113e-01
-5.10881186e-01 4.60620783e-02 4.38962609e-01 -4.14803386e-01
-7.43455291e-01 -2.25583255e-01 4.32379872e-01 -9.09166098e-01
9.86267388e-01 1.05727541e+00 3.92998531e-02 -5.70423245e-01
2.49531105e-01 2.42913753e-01 3.69066931e-02 -6.68088377e-01
4.59533811e-01 9.71966088e-02 6.42309964e-01 -1.21563411e+00
1.41081977e+00 8.65313411e-01 3.37437153e-01 -9.20688570e-01
-1.27214229e+00 -8.54625165e-01 -4.61678207e-01 -3.15167457e-01
4.67685491e-01 -8.95146608e-01 -8.33816290e-01 3.33472937e-01
-1.02157557e+00 -4.52872142e-02 -2.15912730e-01 9.32326257e-01
-5.44304252e-01 9.63593781e-01 -7.99859464e-01 -6.90147042e-01
-1.85974121e-01 -1.17212367e+00 9.02899325e-01 -4.77209747e-01
4.14507598e-01 -1.11471736e+00 1.05010778e-01 7.80698776e-01
3.74685735e-01 5.19588888e-01 3.81057203e-01 1.70872316e-01
-7.24278271e-01 -2.16634706e-01 -6.59181058e-01 6.12449765e-01
-1.87568113e-01 -2.32179135e-01 -2.99806803e-01 -6.15588844e-01
3.91788423e-01 -6.69664666e-02 6.94908440e-01 4.14539516e-01
1.07729125e+00 -7.68108428e-01 3.57662216e-02 1.06155503e+00
1.51448715e+00 -7.31850624e-01 5.26601732e-01 2.21691594e-01
1.36694610e+00 4.32579815e-01 7.01573610e-01 5.52042425e-01
4.28031027e-01 6.80932999e-01 5.80519855e-01 -1.14364177e-01
-3.10409013e-02 1.91183448e-01 5.94656885e-01 1.79801428e+00
-8.76580834e-01 3.58406872e-01 -7.26107657e-01 8.43305588e-01
-2.19725513e+00 -9.00573730e-01 -8.19754064e-01 2.39820170e+00
8.68986607e-01 -5.38643837e-01 1.04404621e-01 1.79906070e-01
9.04525399e-01 1.59444839e-01 -2.71095961e-01 2.02051416e-01
-3.61813724e-01 -2.66357839e-01 8.74929905e-01 7.06189930e-01
-8.70016813e-01 4.83020991e-01 6.28815842e+00 9.02033389e-01
-7.72438705e-01 3.81262004e-01 1.17062286e-01 -1.00476202e-02
-4.77727264e-01 4.91489433e-02 -3.41864496e-01 2.98653603e-01
3.05255175e-01 -4.28512037e-01 7.81571448e-01 6.61804676e-01
5.72012603e-01 2.82238990e-01 -6.36362135e-01 1.16374040e+00
1.75543085e-01 -1.16568756e+00 -1.88357562e-01 4.62032795e-01
1.08650088e+00 4.17652130e-02 1.52641550e-01 -6.53069913e-02
2.43861042e-03 -7.79401779e-01 7.85024345e-01 6.64278865e-01
8.71577144e-01 -6.65179968e-01 6.73842311e-01 -8.90646502e-02
-1.49338365e+00 9.31505635e-02 -4.43920463e-01 1.29035965e-01
4.57835585e-01 1.36107111e+00 -2.93384373e-01 9.32966828e-01
4.94978487e-01 1.30845916e+00 -2.50936206e-02 1.20320940e+00
9.99184027e-02 7.30120480e-01 -4.33106244e-01 4.35620397e-01
4.40534532e-01 -9.25995648e-01 1.26905775e+00 1.00988829e+00
5.38095772e-01 3.79122645e-01 6.03984594e-01 6.49945617e-01
-2.10108057e-01 3.97198290e-01 -3.17210466e-01 1.23193786e-01
5.41413110e-03 1.39082646e+00 -2.96179056e-01 -1.58975333e-01
-3.12678069e-01 6.97131872e-01 1.29641339e-01 6.16663516e-01
-6.35554492e-01 -3.48998643e-02 5.64161003e-01 3.50414187e-01
8.78536478e-02 -6.13837838e-01 -3.52621734e-01 -1.68537354e+00
5.04178286e-01 -8.42670321e-01 1.98708475e-01 -5.52253664e-01
-1.63397229e+00 3.92097503e-01 1.24819003e-01 -1.68386412e+00
4.60781515e-01 -6.69986486e-01 -2.73766905e-01 6.53046191e-01
-1.55712557e+00 -1.27795959e+00 -2.33077735e-01 7.64168739e-01
-2.49155946e-02 4.79690321e-02 6.96866512e-02 8.71733844e-01
-7.80358732e-01 3.75606716e-01 4.94367361e-01 1.35572553e-01
4.55131710e-01 -1.17706585e+00 -5.77235997e-01 1.06635225e+00
-5.84267139e-01 8.95391822e-01 6.45983517e-01 -6.15855396e-01
-1.96376538e+00 -1.20152092e+00 5.08604944e-01 2.68111564e-02
1.28791416e+00 1.85259208e-01 -1.00863695e+00 6.67417288e-01
-1.66685939e-01 5.20079494e-01 4.80861247e-01 5.99231161e-02
-4.74277943e-01 -2.25685373e-01 -9.41820145e-01 4.21400368e-01
9.76218224e-01 -4.82396036e-01 -2.71714747e-01 8.31911504e-01
4.42853808e-01 -4.71201569e-01 -1.42935407e+00 6.58828199e-01
4.01435047e-01 -6.91604376e-01 1.15459728e+00 -4.35178280e-01
2.92372555e-01 -7.53552318e-01 -5.12671173e-01 -9.87754524e-01
-5.55546701e-01 -1.28407431e+00 -3.90059352e-01 8.30103874e-01
7.06681833e-02 -5.35328567e-01 5.38640022e-01 3.25061440e-01
-6.58032656e-01 -8.95019531e-01 -1.28191447e+00 -1.16142261e+00
1.14946335e-03 -3.32643986e-01 5.24381511e-02 1.29524672e+00
1.20291136e-01 8.78942013e-03 -9.36153233e-01 3.32197130e-01
1.31414890e+00 -7.77215138e-02 4.83886302e-01 -8.99814546e-01
-1.33106977e-01 -3.15845340e-01 -3.07639360e-01 -1.32863784e+00
1.09465770e-01 -1.33536589e+00 2.37393484e-01 -1.40401292e+00
4.34335679e-01 -7.80971766e-01 -3.80811572e-01 2.22024009e-01
-2.13377342e-01 5.23785770e-01 1.85335279e-01 9.28891420e-01
-6.27166748e-01 8.20448816e-01 2.01587176e+00 -1.24309294e-01
7.12960586e-02 -2.07182348e-01 -5.94931841e-01 7.35064328e-01
3.72570187e-01 -3.09416354e-01 -2.19578654e-01 -5.49310207e-01
6.30983770e-01 2.47634843e-01 2.52956420e-01 -6.73110247e-01
5.00340089e-02 -3.57711077e-01 -4.50883180e-01 -5.57586610e-01
2.63448775e-01 -6.53886139e-01 4.05239947e-02 1.81796804e-01
-4.79424931e-02 -3.56685445e-02 -3.34809452e-01 7.54886210e-01
-2.69534767e-01 -2.06568733e-01 7.25687385e-01 4.73112278e-02
-3.63446735e-02 7.34409451e-01 5.80617003e-02 3.15987468e-01
5.08440018e-01 -2.15598866e-01 -3.05413753e-01 -2.89787918e-01
-6.65174544e-01 1.85915917e-01 2.89366335e-01 -6.44594282e-02
7.72066832e-01 -1.87267756e+00 -1.17856848e+00 -3.70059937e-01
-5.34085855e-02 1.32190824e-01 3.38093698e-01 1.85885727e+00
-6.39181376e-01 2.12356508e-01 3.47712427e-01 -7.04563797e-01
-8.43062639e-01 3.93262863e-01 2.40268618e-01 -5.32633126e-01
-8.26138914e-01 7.22010493e-01 2.56307214e-01 -4.68085825e-01
-1.22056313e-01 -4.63261664e-01 1.88392356e-01 -1.49293408e-01
4.29382056e-01 1.01395094e+00 2.79919896e-02 -1.08810866e+00
-3.81801844e-01 9.31254625e-01 2.84390271e-01 -9.01131630e-02
1.53324211e+00 -2.25013778e-01 -9.53135788e-01 1.52043179e-01
1.58087623e+00 5.04521072e-01 -9.68183219e-01 -3.95202726e-01
-2.09292304e-02 -8.36334407e-01 4.02017266e-01 -4.42422740e-02
-1.44479477e+00 2.11554155e-01 -3.86433564e-02 1.43823832e-01
1.00262415e+00 -2.30668038e-01 1.24845409e+00 3.55384946e-01
3.93246830e-01 -1.02525830e+00 1.94633469e-01 6.31180048e-01
1.39910209e+00 -9.07556951e-01 6.79623544e-01 -9.01632786e-01
-3.02381933e-01 1.15207887e+00 6.37815967e-02 -4.89169568e-01
7.23855436e-01 -2.08515987e-01 -3.17999214e-01 -3.71925622e-01
-1.42247573e-01 -1.35456383e-01 6.97001755e-01 2.00600728e-01
5.38036108e-01 3.11133653e-01 -7.51540065e-01 7.70022720e-02
-9.19019431e-02 -2.35444918e-01 6.64847672e-01 5.46461344e-01
-1.91330150e-01 -8.54317665e-01 -8.59801352e-01 3.70584726e-01
-5.61267734e-01 -1.23618506e-01 3.37499499e-01 4.12576109e-01
-1.89433992e-01 7.77085066e-01 -5.66572547e-01 -3.59426141e-01
2.25016758e-01 -6.40304208e-01 5.56403160e-01 -5.05159914e-01
-3.41566086e-01 5.81090212e-01 1.52739599e-01 -6.81865990e-01
-7.40612864e-01 -9.08548653e-01 -1.12691951e+00 -5.06136060e-01
-5.42845309e-01 2.62549937e-01 4.88253564e-01 8.24826241e-01
1.51700288e-01 6.63821846e-02 9.63514805e-01 -1.06005764e+00
-6.35570765e-01 -8.32415700e-01 -9.77106392e-01 5.82077980e-01
4.03110981e-01 -8.24853182e-01 -7.35054910e-01 3.03344317e-02] | [7.40500020980835, 4.469921588897705] |
1565700d-f9d2-491a-9cdd-06b916498faa | inference-and-dynamic-decision-making-for | 2209.01092 | null | https://arxiv.org/abs/2209.01092v1 | https://arxiv.org/pdf/2209.01092v1.pdf | Inference and dynamic decision-making for deteriorating systems with probabilistic dependencies through Bayesian networks and deep reinforcement learning | In the context of modern environmental and societal concerns, there is an increasing demand for methods able to identify management strategies for civil engineering systems, minimizing structural failure risks while optimally planning inspection and maintenance (I&M) processes. Most available methods simplify the I&M decision problem to the component level due to the computational complexity associated with global optimization methodologies under joint system-level state descriptions. In this paper, we propose an efficient algorithmic framework for inference and decision-making under uncertainty for engineering systems exposed to deteriorating environments, providing optimal management strategies directly at the system level. In our approach, the decision problem is formulated as a factored partially observable Markov decision process, whose dynamics are encoded in Bayesian network conditional structures. The methodology can handle environments under equal or general, unequal deterioration correlations among components, through Gaussian hierarchical structures and dynamic Bayesian networks. In terms of policy optimization, we adopt a deep decentralized multi-agent actor-critic (DDMAC) reinforcement learning approach, in which the policies are approximated by actor neural networks guided by a critic network. By including deterioration dependence in the simulated environment, and by formulating the cost model at the system level, DDMAC policies intrinsically consider the underlying system-effects. This is demonstrated through numerical experiments conducted for both a 9-out-of-10 system and a steel frame under fatigue deterioration. Results demonstrate that DDMAC policies offer substantial benefits when compared to state-of-the-art heuristic approaches. The inherent consideration of system-effects by DDMAC strategies is also interpreted based on the learned policies. | ['Philippe Rigo', 'Konstantinos G. Papakonstantinou', 'Charalampos P. Andriotis', 'Pablo G. Morato'] | 2022-09-02 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [-2.69032884e-02 4.29920524e-01 -1.28737697e-02 1.50930002e-01
-4.06331599e-01 -8.47890601e-02 2.11526558e-01 3.57978046e-01
-1.66447297e-01 8.95424306e-01 -1.74791873e-01 -2.11292580e-01
-9.94452000e-01 -7.72464573e-01 -4.62208599e-01 -1.14003634e+00
-2.07303450e-01 8.91292274e-01 -1.41433533e-02 -2.80761093e-01
9.59451497e-02 5.41393578e-01 -1.49287093e+00 -3.11712891e-01
6.59159422e-01 9.09683228e-01 3.50539476e-01 6.59927428e-01
4.75788325e-01 8.70960534e-01 -5.99755049e-01 -8.60418901e-02
2.02744901e-01 -1.72085151e-01 -6.52226388e-01 4.16968614e-01
-6.15999579e-01 -3.59864146e-01 -1.08121477e-01 8.71861637e-01
5.44374645e-01 4.48768675e-01 9.60723400e-01 -1.22700298e+00
-3.50078642e-01 5.96713543e-01 -3.56289536e-01 -1.43103078e-01
2.27469448e-02 5.16762733e-01 7.09178865e-01 -3.06585878e-01
-1.04526039e-02 1.53699231e+00 2.39999101e-01 2.63689697e-01
-1.30072117e+00 3.13758552e-01 7.10110188e-01 4.72347379e-01
-1.16204989e+00 1.48682877e-01 7.47045517e-01 -7.39001095e-01
1.06330216e+00 -1.78521514e-01 6.88424766e-01 8.62109184e-01
9.15541708e-01 4.58614230e-01 1.02281165e+00 -3.59579921e-01
9.06366587e-01 -2.82965928e-01 7.49142095e-02 2.64297813e-01
3.44658762e-01 4.37224954e-01 1.13371521e-01 -1.30102023e-01
4.62250262e-01 -8.72263536e-02 1.72283933e-01 -3.69838059e-01
-5.24284005e-01 6.32702947e-01 1.30417258e-01 4.41464782e-03
-9.76940334e-01 4.89217997e-01 3.26526135e-01 3.17144394e-01
3.73035908e-01 3.77014935e-01 -5.52579641e-01 2.58747429e-01
-5.43190837e-01 3.67230535e-01 9.58561659e-01 6.07082725e-01
5.89602947e-01 5.12382388e-01 -2.82939732e-01 5.63823402e-01
8.93352866e-01 8.02555978e-01 -1.86846107e-01 -1.23919165e+00
-7.69058103e-03 2.31190249e-01 5.37535608e-01 -8.38834107e-01
-6.69798017e-01 -5.93826711e-01 -7.23195910e-01 7.09326684e-01
1.06062084e-01 -3.99463385e-01 -6.61047280e-01 1.60234630e+00
5.19360840e-01 -6.38879463e-02 1.42876208e-01 7.31538594e-01
-4.27065849e-01 7.65059412e-01 1.06812954e-01 -6.75366282e-01
1.39150941e+00 -5.26726902e-01 -1.09117019e+00 -1.32374495e-01
1.51512533e-01 -3.01580161e-01 4.06571567e-01 8.83462429e-01
-1.24362421e+00 -4.53376025e-01 -1.00402868e+00 9.70048010e-01
-2.21970469e-01 1.80377454e-01 -1.91858321e-01 5.09995222e-01
-1.14210892e+00 6.79642916e-01 -1.06297612e+00 -1.99482087e-02
-1.08399495e-01 3.04757208e-01 4.01453376e-01 3.34962606e-02
-1.48042977e+00 1.47407877e+00 4.26632106e-01 8.28049958e-01
-1.94621420e+00 -3.70786667e-01 -6.13126278e-01 2.99536169e-01
1.09714508e+00 -7.80122399e-01 1.39467359e+00 -4.65022683e-01
-1.96120334e+00 -2.26676747e-01 4.69561577e-01 -3.42946172e-01
5.67782581e-01 -4.17931497e-01 -3.33275884e-01 3.65584195e-01
-3.12658548e-01 -2.32855469e-01 1.14946783e+00 -1.52856874e+00
-4.48147178e-01 -1.87631547e-01 3.69596750e-01 2.10632816e-01
-5.54437153e-02 -1.19136088e-01 1.26322821e-01 -1.93890959e-01
-4.61417794e-01 -1.07312238e+00 -9.23634171e-01 -4.42949265e-01
-4.63030130e-01 -2.72442967e-01 6.45734131e-01 -9.59316909e-01
1.25215650e+00 -1.67857826e+00 7.12119877e-01 4.51445758e-01
-3.05186957e-01 6.47351593e-02 2.31643870e-01 9.89499509e-01
3.89060341e-02 -1.61065534e-01 -5.51873803e-01 -4.44356710e-01
3.91805857e-01 5.84448755e-01 8.71557668e-02 5.57807505e-01
2.82570481e-01 1.47231802e-01 -9.38811958e-01 -3.76801759e-01
4.44072157e-01 1.77797854e-01 -3.63618970e-01 3.62392306e-01
-5.75269759e-01 3.24405819e-01 -8.88811290e-01 6.47885621e-01
4.35855001e-01 1.92861244e-01 5.12148738e-01 -1.01443447e-01
-2.22155213e-01 -6.61133885e-01 -1.53872824e+00 1.24697638e+00
-9.45407391e-01 -1.63467795e-01 6.82154179e-01 -1.43514955e+00
8.15020263e-01 6.65489137e-01 6.56905115e-01 -3.40203106e-01
2.71919549e-01 8.81787241e-02 -2.10862439e-02 -7.51825392e-01
3.68859261e-01 -3.09731036e-01 -2.55793422e-01 3.27623814e-01
7.39308372e-02 -5.04818916e-01 1.16735548e-01 -1.33326441e-01
1.29467928e+00 2.84569949e-01 6.38334602e-02 -5.91856897e-01
7.05540121e-01 -3.59977931e-01 7.17452884e-01 6.98138893e-01
-7.71338418e-02 -1.64480716e-01 8.01249683e-01 -1.16740614e-01
-8.67359579e-01 -9.22677815e-01 -1.70288421e-02 5.26156723e-01
1.71330601e-01 3.11888725e-01 -8.11413765e-01 -2.65831530e-01
1.16651893e-01 9.89551663e-01 -6.08699799e-01 -5.29048085e-01
-5.61675251e-01 -8.94714773e-01 -8.15638155e-02 2.70086229e-01
1.69680826e-02 -7.99841225e-01 -9.17529583e-01 9.25555229e-01
4.33369190e-01 -7.81284750e-01 2.76431143e-01 2.93693900e-01
-7.37118244e-01 -1.11616898e+00 -5.76362848e-01 -7.41007254e-02
5.77340662e-01 -5.04007459e-01 8.94249499e-01 -2.37927027e-02
-3.53713781e-01 9.57451522e-01 -2.03219026e-01 -3.14743280e-01
-8.89556110e-01 -2.32944246e-02 3.77477974e-01 2.95621425e-01
-5.73486209e-01 -4.81594145e-01 -6.12719834e-01 5.60874820e-01
-1.02620697e+00 -5.36530137e-01 6.70847952e-01 9.23528075e-01
4.80146140e-01 6.54252172e-01 8.22151840e-01 -5.15052676e-01
6.73603594e-01 -6.13440931e-01 -9.77815688e-01 6.28274262e-01
-1.05637038e+00 4.03943062e-01 7.45555162e-01 -1.87107727e-01
-1.64156234e+00 -1.53420582e-01 -1.53722856e-02 -2.49409735e-01
-2.47483820e-01 8.32531929e-01 -4.13555324e-01 3.96766663e-01
4.35169600e-02 -1.11063667e-01 2.31944725e-01 -4.67406303e-01
1.57249346e-01 4.75892037e-01 2.10060269e-01 -1.26898921e+00
6.99171305e-01 1.69200912e-01 5.28207123e-01 -5.39828718e-01
-7.83586264e-01 2.49916539e-02 -4.88521397e-01 -9.93300021e-01
9.78366315e-01 -6.54226661e-01 -1.11635256e+00 7.15645373e-01
-9.63729262e-01 -4.72184062e-01 -4.24445271e-01 4.41611499e-01
-1.03867209e+00 2.53786266e-01 -6.09004498e-01 -1.66984475e+00
1.84303717e-04 -1.40464866e+00 8.66775990e-01 -2.91985013e-02
3.09686184e-01 -1.16971457e+00 4.34258729e-01 1.72796339e-01
4.85982597e-01 5.85043907e-01 1.11430740e+00 -1.67461589e-01
-4.89463657e-01 -2.53089011e-01 6.20665312e-01 6.56003714e-01
-5.97740486e-02 2.58097589e-01 -6.47130489e-01 -5.22583842e-01
2.78733790e-01 -6.32963702e-02 2.83193976e-01 6.00251853e-01
9.28089678e-01 -3.29350978e-01 -8.50891396e-02 -2.63309658e-01
1.69299710e+00 4.80678350e-01 4.17816460e-01 4.24082756e-01
3.17542136e-01 1.25702357e+00 9.48289514e-01 1.05126619e+00
3.28869075e-01 6.65155888e-01 1.28195524e+00 2.86428154e-01
4.28575367e-01 1.95173576e-01 5.86481154e-01 6.16911829e-01
-2.10613623e-01 -4.91095573e-01 -9.38816011e-01 5.40882647e-01
-2.09468031e+00 -9.08841610e-01 1.89676985e-01 2.14902234e+00
4.26275969e-01 2.77112961e-01 -8.46821219e-02 1.44937620e-01
9.34865832e-01 -1.60086915e-01 -7.46754885e-01 -6.30185008e-01
1.39558703e-01 -3.26421082e-01 5.10233402e-01 5.69470227e-01
-6.76065922e-01 1.43567294e-01 5.90203667e+00 7.49256849e-01
-4.86889511e-01 1.35819835e-03 6.66633248e-01 4.11513746e-02
-2.64035631e-02 1.98854223e-01 -6.47677541e-01 4.01696086e-01
1.57895482e+00 8.22765678e-02 4.46727812e-01 7.21634030e-01
8.66922677e-01 -3.17211837e-01 -8.60906959e-01 8.84344876e-02
-5.33982515e-01 -6.83802664e-01 -5.08767307e-01 1.42671719e-01
6.52800500e-01 -4.04002309e-01 -1.31171510e-01 3.17623407e-01
5.41242361e-01 -4.87379551e-01 1.16255808e+00 1.08407593e+00
9.61614400e-02 -1.01160276e+00 9.89915609e-01 5.08022785e-01
-9.47001874e-01 -7.70860434e-01 -1.81414541e-02 -3.86859328e-02
9.57601070e-01 6.93562806e-01 -3.75592202e-01 1.08370674e+00
5.71724057e-01 4.20846611e-01 -7.82944784e-02 8.22908044e-01
-2.28075489e-01 6.80171371e-01 -2.66893595e-01 -1.21134922e-01
1.46688491e-01 -3.79649103e-01 6.00110948e-01 6.95453763e-01
3.32114726e-01 -1.25011846e-01 4.67418343e-01 7.72454858e-01
8.22940409e-01 -4.44498211e-01 -2.20615119e-01 1.71115130e-01
3.31526041e-01 1.06203711e+00 -7.93056965e-01 -2.13018209e-01
-1.22842845e-02 1.55503064e-01 -3.78521197e-02 4.20307785e-01
-1.04040706e+00 -1.90920651e-01 5.83159566e-01 2.30987649e-02
3.19235057e-01 -2.81187743e-01 1.10232301e-01 -5.85218787e-01
-1.75430968e-01 -6.83127582e-01 2.15340689e-01 -6.69233084e-01
-1.40686405e+00 4.44004089e-01 7.23443151e-01 -1.21369231e+00
-5.44635892e-01 -8.49952936e-01 -5.28854966e-01 5.27504444e-01
-1.51678884e+00 -8.03528845e-01 3.67582560e-01 2.61775762e-01
4.03918892e-01 9.00812168e-03 4.49902087e-01 1.62539497e-01
-1.19732976e+00 -1.01162598e-01 6.19200468e-01 -6.68837249e-01
1.61487758e-01 -1.38199329e+00 -3.53846908e-01 8.96213472e-01
-9.37206626e-01 1.65286452e-01 1.28871822e+00 -6.54744506e-01
-1.57410347e+00 -8.64519000e-01 -1.23221911e-01 2.02137064e-02
9.99422431e-01 1.88121051e-02 -8.55999947e-01 1.28246978e-01
5.03784120e-01 -3.20253819e-01 1.70167327e-01 -1.27409384e-01
5.13283610e-01 -1.72895670e-01 -1.23339570e+00 3.12221020e-01
4.10100639e-01 -2.34766588e-01 -3.08668226e-01 3.56919616e-01
7.56369293e-01 -1.40709862e-01 -1.48138154e+00 4.93697286e-01
1.91746533e-01 -3.51115108e-01 8.56812894e-01 -4.01714832e-01
1.42950445e-01 -4.68738586e-01 -1.08674576e-03 -1.62773263e+00
-4.00995672e-01 -7.54238248e-01 -6.83081448e-01 1.23485577e+00
3.52967381e-01 -5.82501411e-01 1.68635935e-01 6.68922842e-01
-6.32950544e-01 -8.61662447e-01 -1.28338468e+00 -1.00453782e+00
-6.47329353e-03 -3.16263169e-01 4.77535427e-01 2.96976626e-01
-3.48661959e-01 -1.42236769e-01 -5.18598199e-01 7.54429400e-01
9.27843869e-01 -1.79387569e-01 8.75583887e-02 -9.44893897e-01
-8.20428312e-01 -2.40104720e-01 -5.72380200e-02 -3.14102262e-01
3.54104072e-01 -2.15071645e-02 5.85258067e-01 -1.53484654e+00
-3.16060573e-01 -2.65187234e-01 -6.39929116e-01 1.76179171e-01
3.97748426e-02 -7.52578139e-01 1.60735130e-01 -3.95286381e-02
-5.07808566e-01 1.16587138e+00 1.10288692e+00 -2.61750340e-01
1.99698448e-01 3.28686297e-01 -2.10470930e-01 5.57465017e-01
6.79268122e-01 -6.05582237e-01 -4.77550298e-01 -3.49091858e-01
4.66203541e-01 8.07298601e-01 4.06587124e-01 -1.14266539e+00
1.08563185e-01 -6.25073910e-01 -3.92019391e-01 -5.87570965e-01
4.65269327e-01 -1.25520062e+00 6.58437133e-01 1.02553320e+00
-2.49037117e-01 2.97273904e-01 9.89040881e-02 1.13235986e+00
-2.71447301e-01 -6.65231347e-01 7.17094660e-01 3.22107449e-02
-5.38988352e-01 2.63450928e-02 -1.03152239e+00 -4.61893678e-01
1.23477697e+00 2.25867093e-01 -1.30843207e-01 -1.59075066e-01
-1.25219297e+00 7.00776398e-01 3.67687978e-02 1.02590658e-02
3.11253369e-01 -8.42279851e-01 -6.83933198e-01 -4.89587218e-01
-2.83081740e-01 -5.51759228e-02 7.85974503e-01 7.80798793e-01
-3.46227407e-01 1.61071718e-01 -1.39572471e-01 -5.28334796e-01
-5.40365636e-01 7.36077368e-01 6.60008907e-01 -6.68738484e-01
-1.10443801e-01 3.10085863e-01 -1.78296223e-01 -2.26912931e-01
-1.35335177e-01 -3.85365844e-01 -2.63219774e-01 2.85816669e-01
1.84083968e-01 9.88674819e-01 2.77973950e-01 -2.95038939e-01
-9.90468785e-02 4.45661545e-01 3.37732077e-01 -3.23382467e-01
1.47455037e+00 -3.49897474e-01 -4.49169092e-02 5.82125843e-01
3.72446895e-01 -7.37012208e-01 -1.74957418e+00 2.73217969e-02
1.60960764e-01 1.11849092e-01 3.85607600e-01 -7.98537910e-01
-1.04293883e+00 6.76134884e-01 7.57950485e-01 7.18646705e-01
1.16886246e+00 -2.45753288e-01 1.72238141e-01 4.66723323e-01
5.79489648e-01 -1.64820492e+00 2.05807596e-01 4.94440913e-01
1.01443768e+00 -6.66790009e-01 1.79432422e-01 9.01176333e-02
-5.40156305e-01 1.04415059e+00 6.88614011e-01 -3.86852413e-01
1.02121794e+00 4.39490765e-01 -2.88354099e-01 -1.78732410e-01
-1.11239600e+00 5.14057837e-02 -2.12351799e-01 4.03796017e-01
-2.62776792e-01 1.31016016e-01 -2.02943936e-01 5.55354059e-01
6.76675200e-01 -2.94950217e-01 7.91915178e-01 1.37192428e+00
-5.17179608e-01 -1.06274772e+00 -7.90078163e-01 1.38108850e-01
-3.36264461e-01 4.87796217e-01 4.08923775e-01 7.00969577e-01
-1.41845061e-03 1.05184412e+00 -3.03520143e-01 4.09626327e-02
7.25738943e-01 -2.38978311e-01 2.08397374e-01 -5.45427918e-01
-6.86145246e-01 1.62465796e-01 1.60051167e-01 -5.16978621e-01
-4.34132934e-01 -1.00172138e+00 -1.12955976e+00 2.24473223e-01
-6.60322189e-01 4.14561570e-01 8.03772390e-01 1.12254477e+00
3.04614305e-01 1.31601226e+00 8.83943319e-01 -1.11816621e+00
-1.49809921e+00 -9.33167696e-01 -9.62876379e-01 -3.62562805e-01
1.18965589e-01 -1.21092558e+00 -4.43558931e-01 -1.23011217e-01] | [4.679698944091797, 2.3638508319854736] |
d6172700-3864-4861-904f-88c70ef74a4b | exploiting-network-structures-to-improve | 2107.05885 | null | https://arxiv.org/abs/2107.05885v1 | https://arxiv.org/pdf/2107.05885v1.pdf | Exploiting Network Structures to Improve Semantic Representation for the Financial Domain | This paper presents the participation of the MiniTrue team in the FinSim-3 shared task on learning semantic similarities for the financial domain in English language. Our approach combines contextual embeddings learned by transformer-based language models with network structures embeddings extracted on external knowledge sources, to create more meaningful representations of financial domain entities and terms. For this, two BERT based language models and a knowledge graph embedding model are used. Besides, we propose a voting function to joint three basic models for the final inference. Experimental results show that the model with the knowledge graph embeddings has achieved a superior result than these models with only contextual embeddings. Nevertheless, we also observe that our voting function brings an extra benefit to the final system. | ['Shi-jie We', 'Chao Feng'] | 2021-07-13 | null | https://aclanthology.org/2021.finnlp-1.10 | https://aclanthology.org/2021.finnlp-1.10.pdf | finnlp-2021-8 | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [-7.57871509e-01 5.39276898e-01 -3.59015465e-01 -3.99488032e-01
-1.19360924e-01 -3.91811758e-01 1.01446903e+00 2.69661605e-01
-5.49360991e-01 5.17554283e-01 5.27612746e-01 -3.21590990e-01
-2.17125818e-01 -1.16933477e+00 -5.22491097e-01 4.46833558e-02
-7.71508813e-02 6.18786335e-01 4.78767127e-01 -4.35548156e-01
1.38300776e-01 2.14849994e-01 -1.10627210e+00 3.43886256e-01
6.02428794e-01 1.01267636e+00 -4.34913449e-02 1.83359861e-01
-7.05777049e-01 1.47738159e+00 -2.40911156e-01 -1.17445540e+00
1.00926094e-01 7.14636687e-03 -1.14975703e+00 -6.07519209e-01
-1.00111021e-02 -2.81766038e-02 -6.34584904e-01 1.07610071e+00
2.09270835e-01 3.78385872e-01 8.36271584e-01 -1.43566096e+00
-1.39140761e+00 1.23899376e+00 1.69100419e-01 1.15833201e-01
2.86219746e-01 -3.99946064e-01 1.47370434e+00 -1.24666560e+00
8.42493594e-01 1.52294648e+00 7.98144221e-01 1.86615646e-01
-8.75750840e-01 -4.49368864e-01 2.74475873e-01 8.69717121e-01
-1.42339385e+00 -1.03036195e-01 8.85729611e-01 -3.26559424e-01
1.74256623e+00 -2.32978821e-01 6.34589970e-01 1.23366499e+00
-5.61741106e-02 5.46802878e-01 7.11359918e-01 -4.28047657e-01
1.13766566e-01 7.18968034e-01 5.26059747e-01 8.84684861e-01
3.15499455e-01 6.58977255e-02 -4.95155185e-01 -2.66649574e-01
5.32379270e-01 -9.02737118e-03 1.17075339e-01 -4.24085230e-01
-9.45928633e-01 1.17858374e+00 7.10149288e-01 8.12006354e-01
-2.99067914e-01 2.32996121e-01 4.48820293e-01 4.81211931e-01
6.48523986e-01 8.04502547e-01 -6.91686690e-01 1.02079637e-01
-7.16094732e-01 1.36186272e-01 1.26271725e+00 1.02590120e+00
7.12315559e-01 3.89195681e-02 1.20096788e-01 7.67807603e-01
7.12564826e-01 2.06647128e-01 6.56442583e-01 -7.78371930e-01
4.36886311e-01 9.05197144e-01 -5.78083806e-02 -1.41035998e+00
-1.94032431e-01 -3.53922367e-01 -3.13118935e-01 -2.21994787e-01
5.99836223e-02 -8.53089765e-02 -5.13861954e-01 1.51427996e+00
1.00649213e-02 4.65407908e-01 5.64774513e-01 6.38483703e-01
1.20751309e+00 5.39412081e-01 2.34733820e-01 3.72100145e-01
1.59831941e+00 -1.33393323e+00 -9.28055465e-01 -1.19083554e-01
7.73767054e-01 -6.78630769e-01 5.80321491e-01 7.19907600e-03
-8.92490864e-01 -6.30959749e-01 -1.02633202e+00 -3.68232787e-01
-1.28144836e+00 -2.01067284e-01 6.32297814e-01 4.00204301e-01
-1.10190463e+00 6.08497500e-01 -5.60450613e-01 -5.98160028e-01
-1.01260757e-02 -1.52396888e-01 -3.96376818e-01 1.47355278e-03
-1.97336113e+00 1.60222995e+00 9.83141422e-01 -1.63683772e-01
-4.64925349e-01 -7.36041605e-01 -1.29763258e+00 2.68861592e-01
3.75773549e-01 -1.03594434e+00 8.38049650e-01 -5.91407299e-01
-1.32183921e+00 7.53579140e-01 -8.45075399e-02 -8.17532957e-01
2.36101076e-01 -2.96533763e-01 -7.52709746e-01 1.51379436e-01
-1.78567541e-03 2.99169898e-01 4.43532467e-01 -9.02673841e-01
-3.29254955e-01 -1.71149001e-01 4.84099478e-01 -1.61645208e-02
-7.94166684e-01 1.43412739e-01 -4.44248468e-01 -7.22290039e-01
-3.69583100e-01 -3.78542364e-01 -1.80404231e-01 -1.83334365e-01
-1.29901618e-01 -7.47113168e-01 6.20159447e-01 -9.91662562e-01
1.33526230e+00 -1.91518199e+00 2.31059164e-01 3.41700822e-01
3.10697407e-01 2.84880459e-01 -1.64376438e-01 7.20177531e-01
-1.44003987e-01 2.94001728e-01 1.01811849e-01 -3.80573332e-01
3.16264063e-01 4.09769863e-01 -5.25088727e-01 -2.95991778e-01
3.53968501e-01 1.23369992e+00 -1.02656496e+00 -5.78839779e-01
2.92710185e-01 5.13441443e-01 -5.47724903e-01 2.42457002e-01
-1.30684540e-01 -4.80187654e-01 -5.05329967e-01 4.38555211e-01
3.80090714e-01 -2.72554368e-01 5.97599328e-01 -5.82545340e-01
3.00790995e-01 4.17050600e-01 -1.18727124e+00 1.79271472e+00
-6.95384681e-01 2.19128937e-01 -3.35575700e-01 -1.31087315e+00
1.13478827e+00 3.76538783e-01 1.29159704e-01 -5.55941164e-01
2.71657705e-01 6.17750473e-02 -1.83880508e-01 -6.76457644e-01
7.34246492e-01 -3.33936453e-01 -8.97150785e-02 2.28722155e-01
7.13829935e-01 -7.35037588e-03 3.74811105e-02 5.94478130e-01
9.74824488e-01 2.56959260e-01 3.57466906e-01 -3.34724635e-01
4.55683529e-01 -1.78685009e-01 5.20707071e-01 3.08942169e-01
5.63739042e-04 1.57291234e-01 5.51894844e-01 -5.03626645e-01
-8.57226312e-01 -1.12357056e+00 -4.30767015e-02 7.65003145e-01
5.16625717e-02 -9.14772332e-01 -2.93777466e-01 -1.02877033e+00
3.00069839e-01 1.06288254e+00 -7.33135402e-01 -2.85127342e-01
-2.99000025e-01 -4.09685761e-01 6.78924322e-01 9.60186183e-01
3.89777094e-01 -8.99016440e-01 3.85508798e-02 3.05273354e-01
-2.35991463e-01 -1.49221182e+00 -8.22341964e-02 3.44697237e-02
-7.04608858e-01 -1.33056307e+00 -3.58104676e-01 -9.08788562e-01
2.90797532e-01 -3.42884064e-01 1.27759600e+00 -3.14470120e-02
-5.11929803e-02 4.81910288e-01 -5.78794599e-01 -1.04317963e-01
-1.35328218e-01 2.78782435e-02 -7.14176521e-02 -1.34335726e-01
7.68008053e-01 -5.78743100e-01 -1.18632689e-01 2.82035246e-02
-7.91665614e-01 -2.66459316e-01 1.75837234e-01 7.86409378e-01
2.15695471e-01 -2.69765891e-02 7.65775442e-01 -9.53816891e-01
8.01954448e-01 -8.75615776e-01 -3.75804961e-01 5.51310122e-01
-8.26429129e-01 5.05881608e-01 7.19067216e-01 -6.81980997e-02
-1.14486015e+00 -5.02505779e-01 -9.47324559e-02 -5.37105501e-01
2.95351565e-01 9.79847014e-01 -7.74693564e-02 1.44390255e-01
5.35091102e-01 -8.65032002e-02 -1.89128101e-01 -7.90512383e-01
8.85379493e-01 5.10890126e-01 1.64301991e-01 -7.14973986e-01
7.58852541e-01 1.07635893e-01 -2.49710292e-01 -6.14828229e-01
-8.21828604e-01 -3.99811327e-01 -5.43316662e-01 2.76926726e-01
9.62542474e-01 -1.06405580e+00 -3.17355782e-01 1.77212268e-01
-1.32579660e+00 8.81777778e-02 -4.16900784e-01 6.67384267e-01
-1.67013139e-01 5.15814424e-01 -7.37482667e-01 -5.03165364e-01
-2.71858007e-01 -6.28210366e-01 4.00852323e-01 5.81981763e-02
-1.79385975e-01 -1.96524692e+00 2.05336928e-01 3.04367274e-01
6.56252384e-01 2.07173765e-01 1.12485969e+00 -1.18705595e+00
-3.10602993e-01 -5.24167903e-02 -4.76522952e-01 6.56517982e-01
1.33396626e-01 -3.43251340e-02 -9.57067013e-01 1.43786937e-01
-2.97404230e-01 -2.80203134e-01 1.03806889e+00 -2.84952581e-01
9.64209676e-01 -1.69051722e-01 -3.71513426e-01 4.52355504e-01
1.61178720e+00 -1.13246195e-01 5.42942166e-01 3.70820045e-01
7.07839310e-01 5.11077821e-01 4.21159446e-01 2.55684406e-01
1.01691484e+00 5.66670716e-01 2.59880334e-01 1.95189819e-01
-1.70232311e-01 -6.70144856e-01 4.75216240e-01 1.30051768e+00
-2.47513562e-01 -6.01155460e-02 -9.41692650e-01 9.47125316e-01
-2.05591559e+00 -9.61425006e-01 4.89798412e-02 1.62841439e+00
7.73612738e-01 1.34090260e-01 -1.84348732e-01 -2.76838064e-01
5.37288606e-01 1.61243409e-01 -2.04400718e-01 -6.86114669e-01
-2.81334341e-01 6.15912557e-01 2.78951347e-01 7.53640413e-01
-8.65683258e-01 1.19265151e+00 6.67513371e+00 8.46912444e-01
-6.05171323e-01 4.75225210e-01 -1.25065148e-01 2.35918120e-01
-7.70931721e-01 3.67796719e-01 -7.40512967e-01 3.69828522e-01
1.22458017e+00 -5.86245477e-01 2.09805936e-01 8.99666667e-01
-3.51102501e-01 4.77868199e-01 -1.20275128e+00 6.88221216e-01
1.98370293e-01 -1.55579865e+00 3.98221046e-01 -2.07875997e-01
5.16112983e-01 7.58218616e-02 -3.57859761e-01 8.19508553e-01
8.26227725e-01 -8.73473287e-01 5.64442396e-01 1.09518266e+00
1.98927969e-01 -7.67334282e-01 1.17964184e+00 8.92999023e-02
-1.44941521e+00 -5.66276349e-02 -5.73822558e-01 -1.28334373e-01
1.88632324e-01 7.79066920e-01 -8.06278169e-01 1.42570329e+00
5.33386588e-01 1.13800347e+00 -6.37147009e-01 5.05129337e-01
-6.76771879e-01 4.58447993e-01 1.03694377e-02 -2.02225074e-01
1.85845718e-01 -9.94895771e-02 3.14257681e-01 1.33854973e+00
2.81853139e-01 -1.94573060e-01 8.31100270e-02 1.27881801e+00
-3.20263207e-01 2.04997241e-01 -9.03690517e-01 -3.25832218e-01
5.77284753e-01 1.24091029e+00 -1.21060334e-01 -7.86421657e-01
-7.61025846e-01 8.38678181e-01 7.95577705e-01 3.02507877e-01
-8.41828227e-01 -7.99985945e-01 6.88339174e-01 -8.31418633e-02
4.61950719e-01 -2.39103287e-01 1.14755869e-01 -1.56324184e+00
-6.74360842e-02 -2.16522112e-01 6.75584614e-01 -9.40522552e-01
-1.86910665e+00 5.68304479e-01 2.00991407e-01 -6.45284534e-01
-4.47941214e-01 -8.54060531e-01 -6.38694167e-01 9.39478695e-01
-1.98806250e+00 -1.37621474e+00 1.59482583e-01 7.96469688e-01
4.16899705e-03 -5.52941859e-01 1.16589952e+00 4.01705354e-01
-2.35043868e-01 5.96989512e-01 -1.46273538e-01 5.60092211e-01
6.58049703e-01 -1.36660147e+00 4.25336003e-01 5.97479403e-01
4.24354166e-01 9.68191862e-01 3.28554422e-01 -5.38280666e-01
-1.16614211e+00 -1.19985878e+00 1.68705559e+00 -7.11758077e-01
1.18952072e+00 -7.00625852e-02 -9.70822096e-01 9.96613204e-01
6.38913453e-01 9.39992592e-02 8.65294576e-01 4.48454857e-01
-8.08268666e-01 7.60052651e-02 -1.17019403e+00 3.16251487e-01
1.13133264e+00 -7.91944981e-01 -1.66295135e+00 7.32913241e-02
1.23424447e+00 1.69356287e-01 -1.47404337e+00 2.14754969e-01
2.51704633e-01 -4.63932723e-01 1.16737103e+00 -1.09679747e+00
6.71900153e-01 -3.12694848e-01 -2.88288444e-01 -1.51167905e+00
-3.56750280e-01 1.04286417e-01 -4.24160421e-01 1.59574187e+00
6.77604377e-01 -1.03809524e+00 2.69232243e-01 5.59025466e-01
7.85606280e-02 -4.18921113e-01 -9.03021574e-01 -1.05243909e+00
3.86873841e-01 -5.01376390e-01 7.41482079e-01 1.57387102e+00
4.57389563e-01 4.48646426e-01 -1.58666551e-01 -1.20633084e-03
3.81967276e-01 1.06631987e-01 3.34572434e-01 -1.30142498e+00
-2.32805014e-01 -1.60531700e-01 -7.00878322e-01 -6.56068265e-01
7.66143858e-01 -1.48335195e+00 -7.36897528e-01 -1.84135389e+00
6.46858737e-02 -1.85015559e-01 -9.35316622e-01 7.10904062e-01
2.12789010e-02 -2.35892981e-01 3.67757976e-01 -2.37575635e-01
-6.24182463e-01 8.35510552e-01 8.18095565e-01 -2.58732885e-01
4.12041962e-01 -6.42069936e-01 -9.44414794e-01 7.36462355e-01
6.73562765e-01 -4.40083563e-01 -4.02648091e-01 -6.46941304e-01
6.19283378e-01 -2.65615284e-01 6.25332952e-01 -5.96832156e-01
3.47052932e-01 5.83186410e-02 -4.12644297e-02 -1.45900816e-01
3.37751091e-01 -1.10164392e+00 -4.35681343e-02 3.98211271e-01
-2.68770009e-01 1.47277564e-01 1.15195557e-01 5.09797096e-01
-7.82190025e-01 -4.58855867e-01 2.83059508e-01 -2.68521160e-01
-1.04188824e+00 5.97515926e-02 -1.03427015e-01 3.40259969e-01
7.40157306e-01 1.38152465e-01 -2.77929604e-01 -9.36142579e-02
-9.95586216e-01 4.32115942e-01 6.50439458e-03 8.66794288e-01
6.42159581e-01 -1.74437487e+00 -6.49306953e-01 5.39406352e-02
2.58961409e-01 -4.80645508e-01 -1.21813603e-01 5.77507198e-01
-3.15121025e-01 5.50829828e-01 -1.02411188e-01 1.14267111e-01
-9.20564473e-01 5.70681810e-01 4.29466903e-01 -6.85569763e-01
-4.20223027e-01 7.89621353e-01 -2.75290608e-01 -1.01972902e+00
-5.62623106e-02 -6.43417120e-01 -6.17208123e-01 3.24295789e-01
2.87046492e-01 4.25615609e-01 8.59406292e-02 -3.52428645e-01
-5.10591865e-01 5.56710184e-01 -3.57683897e-02 9.67043266e-03
1.54500282e+00 1.29961058e-01 -2.63388664e-01 4.37028587e-01
1.32751608e+00 -2.89376024e-02 -3.70742947e-01 -9.12952602e-01
3.25769961e-01 -3.19542736e-01 8.53352398e-02 -7.63399065e-01
-1.14496505e+00 9.02020752e-01 1.11191779e-01 3.52009356e-01
4.80973035e-01 7.34837949e-02 7.43366182e-01 6.18548632e-01
4.31961447e-01 -1.27890670e+00 -7.90244564e-02 8.99829566e-01
9.20060039e-01 -9.88191187e-01 -2.57209271e-01 -3.10519308e-01
-6.80898130e-01 1.36842453e+00 5.67240834e-01 -3.88308257e-01
1.15339887e+00 5.59850074e-02 -1.80464432e-01 -4.06923831e-01
-9.34150815e-01 -2.64623791e-01 3.17108363e-01 6.02792323e-01
3.90558153e-01 5.98388575e-02 -4.22646850e-01 1.38174820e+00
-3.53863835e-01 7.41429254e-02 2.01219797e-01 6.37980402e-01
-3.15343082e-01 -1.42024803e+00 1.64961189e-01 5.34471311e-02
-3.06419641e-01 -5.16729116e-01 -5.46415389e-01 9.23466444e-01
2.18853280e-01 8.07242811e-01 5.86888799e-03 -5.95698893e-01
6.55313313e-01 7.00892866e-01 2.50843734e-01 -7.19018340e-01
-6.95322335e-01 -7.20165849e-01 4.62222725e-01 -6.12499774e-01
-5.46018362e-01 -2.55710751e-01 -1.17343760e+00 -3.68797451e-01
-2.58575857e-01 4.10050064e-01 4.78561729e-01 1.02672958e+00
5.42948425e-01 8.41630936e-01 4.09001082e-01 -8.01038221e-02
-5.28577447e-01 -1.05880785e+00 -7.15099692e-01 4.18588102e-01
-3.11949581e-01 -7.69500971e-01 -4.55175370e-01 -1.88114457e-02] | [9.419897079467773, 8.268871307373047] |
9b159b48-cad0-4af4-a95d-0199c718832c | set-prediction-without-imposing-structure-as-1 | 2010.04109 | null | https://arxiv.org/abs/2010.04109v2 | https://arxiv.org/pdf/2010.04109v2.pdf | Set Prediction without Imposing Structure as Conditional Density Estimation | Set prediction is about learning to predict a collection of unordered variables with unknown interrelations. Training such models with set losses imposes the structure of a metric space over sets. We focus on stochastic and underdefined cases, where an incorrectly chosen loss function leads to implausible predictions. Example tasks include conditional point-cloud reconstruction and predicting future states of molecules. In this paper, we propose an alternative to training via set losses by viewing learning as conditional density estimation. Our learning framework fits deep energy-based models and approximates the intractable likelihood with gradient-guided sampling. Furthermore, we propose a stochastically augmented prediction algorithm that enables multiple predictions, reflecting the possible variations in the target set. We empirically demonstrate on a variety of datasets the capability to learn multi-modal densities and produce different plausible predictions. Our approach is competitive with previous set prediction models on standard benchmarks. More importantly, it extends the family of addressable tasks beyond those that have unambiguous predictions. | ['Cees G. M. Snoek', 'Gertjan J. Burghouts', 'David W. Zhang'] | 2020-10-08 | set-prediction-without-imposing-structure-as | https://openreview.net/forum?id=04ArenGOz3 | https://openreview.net/pdf?id=04ArenGOz3 | iclr-2021-1 | ['point-cloud-reconstruction'] | ['computer-vision'] | [ 6.12563550e-01 2.60087132e-01 -4.50945258e-01 -7.86630988e-01
-1.29314733e+00 -5.29994071e-01 7.73745060e-01 4.41480689e-02
-1.01673238e-01 1.36753118e+00 5.63315861e-02 -3.28147173e-01
-1.65789336e-01 -9.22260225e-01 -1.32567906e+00 -8.41368496e-01
-1.54259196e-02 1.09872770e+00 -1.02724731e-01 2.59202197e-02
2.77286768e-01 5.44865608e-01 -1.61629272e+00 3.83053482e-01
7.55918324e-01 9.71751332e-01 1.20270826e-01 3.36864382e-01
-6.22665361e-02 8.11536670e-01 -1.52182117e-01 -4.61178482e-01
2.44955085e-02 -2.50594944e-01 -6.35610461e-01 -1.40483961e-01
6.69587076e-01 -5.68509735e-02 1.02835104e-01 9.57756937e-01
1.96963087e-01 3.89993906e-01 1.30749977e+00 -1.21056414e+00
-6.45744562e-01 5.73392272e-01 -2.66568035e-01 -2.27581158e-01
2.80558676e-01 2.80912101e-01 1.23984516e+00 -1.04442203e+00
5.11645198e-01 1.28346860e+00 9.15167868e-01 6.84304118e-01
-1.75672269e+00 -6.25219226e-01 8.30306038e-02 1.81950897e-01
-1.33117878e+00 -4.37209040e-01 5.97351074e-01 -5.62946439e-01
1.27243292e+00 4.22618598e-01 4.03708845e-01 1.35262787e+00
3.16439509e-01 6.15594447e-01 8.75578642e-01 -3.19099993e-01
6.88692927e-01 2.44592354e-01 -8.15747678e-03 5.07213593e-01
3.20663452e-01 3.50250721e-01 -7.53099859e-01 -5.91896296e-01
2.60666519e-01 1.99090108e-01 -3.39342058e-01 -7.64100611e-01
-1.03937507e+00 9.37002480e-01 1.01307452e-01 -3.35855305e-01
-1.98733032e-01 5.37360370e-01 1.29118487e-01 3.14888917e-02
7.61130750e-01 6.05568111e-01 -9.12405610e-01 -3.24824033e-03
-8.75129282e-01 5.26712239e-01 1.01063013e+00 8.96405458e-01
1.01462591e+00 -2.58701503e-01 -1.22457884e-01 4.10988688e-01
3.53987724e-01 4.00327832e-01 -3.11048459e-02 -9.56452131e-01
2.58535624e-01 9.08154622e-02 6.00817919e-01 -4.00402844e-01
-2.37208068e-01 -3.35127175e-01 -8.27605247e-01 4.06269580e-01
4.89526331e-01 -7.81656429e-02 -9.57608759e-01 1.91818607e+00
2.53119737e-01 4.95742559e-01 -1.75206691e-01 6.94155574e-01
3.67387474e-01 7.15040445e-01 2.37878174e-01 -4.54764932e-01
5.41372180e-01 -4.93358046e-01 -4.51949835e-01 1.71204247e-02
5.94792545e-01 -2.74569869e-01 1.20349562e+00 7.45884418e-01
-9.63154912e-01 -2.06463605e-01 -8.53116035e-01 -1.23913668e-01
-3.99474949e-01 -3.86269361e-01 9.26296353e-01 4.99076098e-01
-1.07064295e+00 1.27911246e+00 -1.07193565e+00 1.23960823e-01
7.21961379e-01 7.01221585e-01 -1.79450348e-01 7.15471208e-02
-9.42099035e-01 8.90502036e-01 3.68372321e-01 -3.60177606e-01
-8.29975486e-01 -1.17492640e+00 -7.90979266e-01 7.90593475e-02
1.26482263e-01 -8.62404048e-01 1.19837868e+00 -8.94326568e-01
-1.43255389e+00 5.49050987e-01 -3.03101927e-01 -7.41914570e-01
5.78645945e-01 -1.86555073e-01 -1.29506662e-01 -4.52674747e-01
-2.10453689e-01 7.72580922e-01 7.61647820e-01 -1.17594242e+00
-1.31401718e-01 -2.75814027e-01 -2.10544929e-01 1.58063054e-01
4.75327447e-02 -3.73463660e-01 2.66346186e-01 -3.12387317e-01
-1.15303800e-01 -8.49980772e-01 -5.13966918e-01 2.96430826e-01
-7.33606815e-01 -2.46663764e-01 5.50383627e-01 -6.30518869e-02
8.90670121e-01 -1.73480165e+00 2.97418475e-01 2.45445296e-01
2.40219772e-01 -5.33595264e-01 1.47968054e-01 4.17672753e-01
-1.51760995e-01 3.44293684e-01 -7.25700498e-01 -6.78645372e-01
4.02271569e-01 3.04895163e-01 -7.12941289e-01 6.42873108e-01
3.93261880e-01 8.02340508e-01 -7.77656734e-01 -3.87159497e-01
3.02182764e-01 5.44379711e-01 -8.59880209e-01 -3.80121591e-03
-1.02745295e+00 4.53819960e-01 -3.93832058e-01 5.19744396e-01
7.39450276e-01 -6.48056030e-01 8.58500227e-02 -3.54506150e-02
6.19521327e-02 5.06281793e-01 -9.02791142e-01 1.65694571e+00
-3.18212420e-01 2.81232357e-01 -4.89368945e-01 -9.01203156e-01
8.16047728e-01 -1.19814001e-01 5.21166086e-01 1.12299457e-01
-4.16457325e-01 1.17003746e-01 -2.49765962e-01 5.36136366e-02
4.61419553e-01 -7.21914887e-01 -6.42725229e-02 3.58590662e-01
3.55264008e-01 -5.39977908e-01 -3.32739979e-01 -5.13268672e-02
9.39398468e-01 6.37239873e-01 4.29286927e-01 -3.97347569e-01
1.43377930e-01 -8.49539563e-02 4.76967216e-01 1.01938963e+00
7.88260326e-02 7.18221426e-01 3.67875904e-01 -5.63196659e-01
-1.19377029e+00 -1.45126617e+00 -6.37757897e-01 9.24072981e-01
4.94292900e-02 -3.51851046e-01 -2.26560503e-01 -6.64768159e-01
2.81046093e-01 1.26596677e+00 -5.27663171e-01 -3.09754372e-01
-2.17496023e-01 -1.06603873e+00 3.23701464e-02 2.84981102e-01
-2.14614674e-01 -9.13855255e-01 3.37364548e-03 2.23286539e-01
8.58399123e-02 -5.09295702e-01 -3.32513630e-01 6.30142152e-01
-9.51650918e-01 -9.07563508e-01 -9.81422812e-02 -4.28136140e-01
2.88598716e-01 -5.29998720e-01 1.62185860e+00 -5.44043362e-01
-2.37161011e-01 1.69329017e-01 1.46248847e-01 -6.79753780e-01
-4.17838484e-01 -1.96963456e-03 4.06571478e-01 -1.87250540e-01
7.07057476e-01 -8.00083339e-01 -5.02755702e-01 -1.06914304e-01
-5.52646577e-01 5.89244217e-02 2.20753342e-01 9.71313775e-01
1.32404876e+00 -2.43797839e-01 7.18136549e-01 -1.10227764e+00
2.85367012e-01 -8.69804323e-01 -5.58335721e-01 3.84181440e-01
-5.73039889e-01 4.89280969e-01 7.96637952e-01 -3.70973676e-01
-1.04753864e+00 4.17288721e-01 -6.84498549e-02 -5.04413128e-01
-1.73317268e-01 1.26506716e-01 -2.26573244e-01 3.70045099e-03
8.30572188e-01 2.99980372e-01 -7.05049112e-02 -5.22049248e-01
3.86141151e-01 2.27828264e-01 2.04139471e-01 -8.59015048e-01
4.61283296e-01 5.45294583e-01 3.90109003e-01 -3.91827255e-01
-1.10159338e+00 -4.19128360e-03 -5.14646173e-01 2.30953451e-02
5.79905152e-01 -8.51330101e-01 -7.39431024e-01 1.28789574e-01
-1.03303432e+00 -5.43431878e-01 -7.43778348e-01 4.87497330e-01
-1.15463829e+00 2.55236149e-01 -3.98314595e-01 -1.08005381e+00
-3.25782567e-01 -1.01296985e+00 1.41645563e+00 -1.68979779e-01
-2.80638158e-01 -1.29789007e+00 2.48029932e-01 -1.56765625e-01
9.08978283e-02 4.79319602e-01 9.24423218e-01 -8.85139048e-01
-8.90114665e-01 1.39310911e-01 2.71438271e-01 1.41509250e-01
-7.91589613e-04 1.10775396e-01 -1.13694167e+00 -8.74528736e-02
7.91409332e-03 -7.29003251e-01 1.32176220e+00 6.79405034e-01
1.96615267e+00 -4.86368299e-01 -6.36551201e-01 8.48300159e-01
1.50845671e+00 -1.60360172e-01 4.46281701e-01 -1.16667479e-01
4.86869127e-01 3.86921495e-01 5.51740289e-01 8.12380791e-01
1.05960324e-01 3.81504953e-01 6.52179062e-01 2.63897002e-01
3.34902525e-01 -6.69248223e-01 2.64172643e-01 1.52929962e-01
2.19687477e-01 -5.51824391e-01 -7.03328729e-01 2.04375714e-01
-1.82214105e+00 -1.29052818e+00 2.81450078e-02 2.41162610e+00
1.18952620e+00 2.25660950e-01 5.11637144e-03 -3.45656842e-01
5.35217762e-01 1.50526479e-01 -1.36370373e+00 -1.94419637e-01
-1.53307110e-01 6.58654034e-01 6.14715040e-01 7.46452272e-01
-1.17635787e+00 7.52906859e-01 7.17182112e+00 1.14689088e+00
-8.53231728e-01 3.14161405e-02 1.22096622e+00 -5.01756310e-01
-1.07259607e+00 -3.85922678e-02 -9.63971376e-01 6.14596725e-01
1.05943775e+00 -1.49427250e-01 4.77338672e-01 8.96361530e-01
-7.84324948e-03 1.70049071e-01 -1.68803012e+00 9.32889938e-01
-2.85928726e-01 -1.77901316e+00 1.14035495e-01 2.35963479e-01
7.26305366e-01 3.26335877e-01 3.94396305e-01 3.88325185e-01
8.78789544e-01 -1.61410141e+00 6.52414143e-01 1.24870026e+00
1.04470396e+00 -8.64084125e-01 2.85946786e-01 5.52652061e-01
-5.62677979e-01 2.15766191e-01 -5.02158999e-01 -3.41365993e-01
1.37462720e-01 9.01940882e-01 -9.57191586e-01 2.30440050e-02
3.50842834e-01 1.08792436e+00 7.36049637e-02 7.70651042e-01
2.04270244e-01 8.08777034e-01 -8.06833148e-01 -3.30997348e-01
-1.37067422e-01 -6.27282441e-01 4.89874154e-01 1.04908073e+00
3.94651324e-01 -1.04517631e-01 1.37649819e-01 1.42577469e+00
-2.71489203e-01 -1.53136075e-01 -8.70764434e-01 2.38232613e-01
5.71090043e-01 7.04809606e-01 -6.65176585e-02 -1.50386587e-01
-1.70184508e-01 7.60900915e-01 7.18482554e-01 4.06006604e-01
-8.94230723e-01 1.56675845e-01 1.16030216e+00 1.21797100e-01
3.38169903e-01 -6.99018613e-02 -5.05912066e-01 -1.22232580e+00
-9.44910869e-02 -4.44482058e-01 2.49016643e-01 -6.49111569e-01
-1.90426576e+00 1.00538462e-01 4.79203872e-02 -1.00080752e+00
-3.78093988e-01 -9.05382276e-01 -6.06123447e-01 9.51903403e-01
-1.45910347e+00 -7.42055178e-01 4.42050159e-01 1.77460074e-01
4.06456828e-01 -3.85981821e-03 1.06540370e+00 -1.81284353e-01
-3.59732658e-01 5.31386197e-01 5.56137741e-01 -6.87188029e-01
5.48371255e-01 -1.73458076e+00 3.36566150e-01 3.10618162e-01
1.91647261e-01 4.66313630e-01 9.85971391e-01 -5.27267754e-01
-1.35980546e+00 -1.42547226e+00 7.27183223e-01 -9.88701940e-01
6.00535631e-01 -5.55993259e-01 -9.40398216e-01 8.15750360e-01
-2.80730277e-01 2.66500235e-01 1.09104168e+00 2.53816009e-01
-3.24013621e-01 1.36483490e-01 -1.50626314e+00 3.31534773e-01
1.30283570e+00 -5.47443509e-01 -2.24319175e-01 9.90654528e-01
8.27242374e-01 -3.26074958e-01 -9.57071543e-01 5.18552184e-01
4.58014607e-01 -9.72935379e-01 1.23190653e+00 -1.16529918e+00
6.59499943e-01 1.89514905e-02 -6.03239596e-01 -1.22671402e+00
-3.06733340e-01 -6.70366704e-01 -6.53499901e-01 7.42414951e-01
6.99118555e-01 -5.65587699e-01 1.11913621e+00 9.91584957e-01
-1.09706320e-01 -1.23918188e+00 -9.94391859e-01 -5.47147512e-01
6.38281465e-01 -7.64562905e-01 8.44701290e-01 7.27664828e-01
-2.85643549e-03 4.84465994e-02 -4.31567609e-01 6.62600026e-02
1.00619471e+00 3.04220259e-01 3.13448995e-01 -1.24603426e+00
-7.88898408e-01 -2.50484139e-01 -2.98877060e-01 -1.11175942e+00
5.93694329e-01 -1.01208127e+00 7.15848804e-02 -1.15283501e+00
4.70197141e-01 -7.49431968e-01 -3.70389938e-01 2.94236720e-01
-5.74891716e-02 4.44042385e-02 -1.22609496e-01 1.54665381e-01
-8.43140781e-01 8.69907022e-01 9.46264148e-01 -1.11415766e-01
3.95803899e-02 3.88179123e-01 -6.48828030e-01 7.10317373e-01
9.81239378e-01 -4.67670232e-01 -4.88982409e-01 -2.97179688e-02
4.47690398e-01 1.73376817e-02 5.15916228e-01 -8.19643676e-01
-6.67382404e-02 -6.39752030e-01 6.26144588e-01 -7.05142856e-01
7.22904205e-01 -5.39713085e-01 3.56133938e-01 9.36280116e-02
-6.32037580e-01 -6.04540348e-01 1.06985360e-01 9.25484598e-01
1.55742660e-01 -6.90523535e-02 7.12662756e-01 -2.19647259e-01
-4.76282299e-01 7.34727204e-01 -2.27262061e-02 1.56457677e-01
8.81762445e-01 -1.02272719e-01 1.06269464e-01 -4.17247087e-01
-9.78768528e-01 9.34380740e-02 7.13575542e-01 -8.35433155e-02
7.16414273e-01 -1.15769672e+00 -6.67931497e-01 3.69551629e-02
1.45332783e-01 2.67213255e-01 6.11889511e-02 1.82772428e-01
-2.34375045e-01 3.79577488e-01 3.19393069e-01 -6.98970079e-01
-6.99466109e-01 5.07194161e-01 8.17044258e-01 -3.41088384e-01
-2.85923153e-01 1.13687015e+00 3.12704057e-01 -9.51840401e-01
7.60680288e-02 -4.73560512e-01 2.93586344e-01 -4.31862891e-01
3.22882503e-01 1.33517355e-01 -1.55222982e-01 -2.44073942e-01
-3.01280022e-01 9.27942023e-02 -8.56998041e-02 8.67435187e-02
1.49769187e+00 1.47352204e-01 1.06077187e-01 7.90286899e-01
1.23075521e+00 -2.52107978e-01 -1.93493366e+00 -1.33427441e-01
-1.31716639e-01 -2.62382239e-01 -6.40915707e-02 -1.07108712e+00
-3.53308916e-01 8.42927754e-01 4.41662252e-01 -2.05681145e-01
6.92798376e-01 3.49751413e-01 4.64275390e-01 7.44485795e-01
6.90843046e-01 -7.38840818e-01 -2.28420883e-01 4.28462893e-01
7.70150006e-01 -1.64966846e+00 1.40388623e-01 -3.00183326e-01
-4.69208449e-01 1.08481967e+00 5.88503659e-01 -1.23646460e-01
9.61230397e-01 4.21766639e-01 -7.38854051e-01 -7.70607516e-02
-1.23174095e+00 1.02329902e-01 3.91357630e-01 7.85386741e-01
4.88866329e-01 3.77596021e-01 1.58577546e-01 6.45177245e-01
-4.56811637e-01 9.69183892e-02 3.93943518e-01 5.16697288e-01
-6.15733743e-01 -1.07262492e+00 -7.95180127e-02 1.08548176e+00
-3.93912345e-01 -3.85246873e-01 -3.02273661e-01 5.75586140e-01
9.74114835e-02 5.39358437e-01 3.34815353e-01 -2.25591108e-01
-1.68550164e-01 2.64096320e-01 7.22403467e-01 -8.98552299e-01
-1.42177446e-02 -4.76649672e-01 8.00439790e-02 -5.20695329e-01
-2.83244789e-01 -1.00502396e+00 -1.16715932e+00 -2.56806433e-01
-3.04178655e-01 8.90924688e-03 4.32357103e-01 8.40476632e-01
5.70180655e-01 1.49901614e-01 4.55667913e-01 -1.01002777e+00
-1.25213563e+00 -7.89601028e-01 -7.30308473e-01 3.85184646e-01
6.02888167e-01 -8.21725190e-01 -5.30024529e-01 -1.29323155e-01] | [7.087588787078857, 4.087517738342285] |
3ed49fe0-257c-4c82-a97f-6f31cb00a383 | xricl-cross-lingual-retrieval-augmented-in | 2210.13693 | null | https://arxiv.org/abs/2210.13693v1 | https://arxiv.org/pdf/2210.13693v1.pdf | XRICL: Cross-lingual Retrieval-Augmented In-Context Learning for Cross-lingual Text-to-SQL Semantic Parsing | In-context learning using large language models has recently shown surprising results for semantic parsing tasks such as Text-to-SQL translation. Prompting GPT-3 or Codex using several examples of question-SQL pairs can produce excellent results, comparable to state-of-the-art finetuning-based models. However, existing work primarily focuses on English datasets, and it is unknown whether large language models can serve as competitive semantic parsers for other languages. To bridge this gap, our work focuses on cross-lingual Text-to-SQL semantic parsing for translating non-English utterances into SQL queries based on an English schema. We consider a zero-shot transfer learning setting with the assumption that we do not have any labeled examples in the target language (but have annotated examples in English). This work introduces the XRICL framework, which learns to retrieve relevant English exemplars for a given query to construct prompts. We also include global translation exemplars for a target language to facilitate the translation process for large language models. To systematically evaluate our model, we construct two new benchmark datasets, XSpider and XKaggle-dbqa, which include questions in Chinese, Vietnamese, Farsi, and Hindi. Our experiments show that XRICL effectively leverages large pre-trained language models to outperform existing baselines. Data and code are publicly available at https://github.com/Impavidity/XRICL. | ['Jimmy Lin', 'He Bai', 'Rui Zhang', 'Peng Shi'] | 2022-10-25 | null | null | null | null | ['text-to-sql', 'semantic-parsing'] | ['computer-code', 'natural-language-processing'] | [-2.54926784e-03 1.18939921e-01 -4.22465682e-01 -7.81201720e-01
-1.84296823e+00 -7.99636066e-01 3.64201933e-01 -5.08427136e-02
-4.09714222e-01 6.47046804e-01 3.75859618e-01 -7.87171781e-01
3.26480925e-01 -9.39067721e-01 -1.15210748e+00 4.49525006e-02
6.73748672e-01 9.17078078e-01 2.10648239e-01 -5.29180586e-01
-1.12001337e-01 -2.07862958e-01 -8.91535521e-01 9.14309800e-01
9.75631535e-01 6.22901082e-01 3.48238438e-01 5.54520607e-01
-7.73622751e-01 9.97872412e-01 -5.63755333e-01 -7.89817214e-01
1.82820354e-02 -4.10336077e-01 -1.30575752e+00 -4.49936420e-01
4.79289979e-01 -2.16806233e-01 -2.90601254e-02 7.33991861e-01
5.21734297e-01 2.04656683e-02 1.64444104e-01 -1.01058471e+00
-1.14054644e+00 8.53342712e-01 -5.89141473e-02 5.44020906e-02
6.58236802e-01 2.07293898e-01 1.42453170e+00 -9.46349680e-01
9.03245330e-01 1.45416200e+00 5.32860041e-01 8.08403969e-01
-9.99811947e-01 -6.01731241e-01 -1.94013402e-01 2.02194884e-01
-1.02790511e+00 -4.18648779e-01 4.72135037e-01 3.44847068e-02
1.48924959e+00 2.23768860e-01 -1.50855467e-01 1.52948630e+00
-5.70509247e-02 1.00638223e+00 1.08078706e+00 -6.98438048e-01
7.94679299e-02 2.26231739e-01 2.87088990e-01 7.51265526e-01
-2.68942058e-01 -6.79961517e-02 -5.98476708e-01 -2.71114558e-01
1.79195717e-01 -1.93081781e-01 9.96211246e-02 -2.89910343e-02
-1.03852701e+00 1.20208120e+00 1.92371845e-01 1.34987801e-01
-4.82246764e-02 2.06771895e-01 5.71758747e-01 5.87506354e-01
4.05485094e-01 6.82564199e-01 -1.02518165e+00 -3.72338235e-01
-4.42745417e-01 2.77284145e-01 1.08596361e+00 1.51031899e+00
1.02964926e+00 -2.75993884e-01 -2.14333370e-01 1.16093385e+00
-4.61385772e-02 7.25300193e-01 5.99606931e-01 -1.13229191e+00
1.06576431e+00 6.06332958e-01 7.41297286e-03 -4.37300280e-02
-7.68416971e-02 -2.00801231e-02 1.57752737e-01 -5.93005419e-01
4.32364821e-01 -3.27730894e-01 -6.97877944e-01 1.85627580e+00
2.85277814e-01 -2.89832026e-01 7.21642375e-01 6.10059917e-01
1.07810104e+00 8.96585763e-01 4.97299343e-01 3.17065418e-01
1.64976954e+00 -1.19614565e+00 -4.89666522e-01 -6.42847419e-01
1.17152667e+00 -9.54152107e-01 2.05850029e+00 -2.07260802e-01
-1.06329060e+00 -6.58342123e-01 -3.83230269e-01 -5.68608880e-01
-6.63514316e-01 1.98008120e-01 4.83882964e-01 4.12815571e-01
-8.98832798e-01 1.25596344e-01 -8.66910279e-01 -9.97003615e-01
1.33135960e-01 -2.28506356e-01 -8.10757801e-02 -5.58790207e-01
-1.44983387e+00 8.43249202e-01 3.31435293e-01 -5.51581383e-01
-8.46753061e-01 -9.79933381e-01 -1.12258136e+00 4.62327376e-02
5.68659186e-01 -7.53299296e-01 1.77325177e+00 -7.68236756e-01
-1.26635027e+00 1.02604306e+00 -5.33486187e-01 -3.73259813e-01
1.12793662e-01 -5.00323772e-01 -3.59250903e-01 6.96252286e-02
6.80218399e-01 7.84817994e-01 3.44355434e-01 -9.09437597e-01
-5.05804777e-01 -4.21730012e-01 2.59323537e-01 1.59582570e-01
-2.01510891e-01 6.89187586e-01 -6.47766650e-01 -4.44158942e-01
-2.92867810e-01 -7.88077593e-01 -5.28818369e-03 -4.90442127e-01
-3.54808331e-01 -5.22167206e-01 5.42283177e-01 -8.15647066e-01
9.55441475e-01 -1.92202199e+00 -3.46545696e-01 -4.02479261e-01
-4.58746850e-01 1.15322836e-01 -5.48009276e-01 7.89163172e-01
1.35132402e-01 2.12535903e-01 -2.63832629e-01 -1.66904449e-01
2.15445593e-01 4.70183671e-01 -7.90832341e-01 -3.81382108e-01
5.76719761e-01 1.52206969e+00 -8.80969107e-01 -6.22270703e-01
-1.16935089e-01 1.23880289e-01 -7.11845577e-01 6.72543585e-01
-7.07257390e-01 4.95648906e-02 -7.47404516e-01 8.95753205e-01
6.00310229e-02 -4.04733777e-01 8.31877738e-02 1.18663898e-02
2.82251924e-01 9.23128188e-01 -3.20938379e-01 2.09142542e+00
-1.02206516e+00 1.48265257e-01 -2.65650541e-01 -9.06406283e-01
8.67977440e-01 3.96079123e-01 1.23043738e-01 -1.09191751e+00
-2.87223071e-01 6.03712380e-01 -4.43045348e-01 -8.47846985e-01
4.62641954e-01 -2.53154576e-01 -7.17038572e-01 5.46044886e-01
3.02472353e-01 -2.95930713e-01 1.74849629e-01 3.59567255e-01
1.33606565e+00 5.09162903e-01 7.26362765e-02 -3.00015304e-02
2.78405070e-01 5.96287310e-01 3.92611712e-01 8.17324638e-01
5.96678965e-02 4.53870416e-01 4.17479932e-01 -1.94270357e-01
-9.00906444e-01 -1.18296719e+00 5.12977950e-02 1.80845404e+00
-6.93001822e-02 -3.34761590e-01 -9.27066565e-01 -1.13252187e+00
1.78362634e-02 1.33948016e+00 -3.28348964e-01 4.83221561e-03
-8.23073268e-01 -4.26989198e-01 7.92168438e-01 6.87700152e-01
3.44161779e-01 -1.42389250e+00 -5.60240865e-01 4.01520729e-01
-6.28561020e-01 -1.63302910e+00 -5.77171743e-01 2.76730865e-01
-6.78989291e-01 -9.67878938e-01 -3.22579473e-01 -9.74304736e-01
3.47804278e-01 -4.13209051e-02 1.77737129e+00 -8.78529027e-02
-2.32167661e-01 7.23655343e-01 -6.54277861e-01 -4.08988833e-01
-8.45694363e-01 5.19049942e-01 -6.41786933e-01 -6.41458035e-01
1.11967969e+00 -7.25180516e-03 -2.40518436e-01 3.23452860e-01
-8.36169600e-01 8.71758237e-02 3.14994991e-01 8.02972317e-01
5.13058782e-01 -7.06801832e-01 8.40419292e-01 -1.41321743e+00
6.36570990e-01 -7.57481337e-01 -4.87477571e-01 8.01382840e-01
-3.45741749e-01 2.63498217e-01 9.67755139e-01 -5.40421084e-02
-1.45269489e+00 -1.78357571e-01 -4.73560482e-01 -1.21359162e-01
-3.26688200e-01 6.57351613e-01 -1.65292531e-01 3.76021057e-01
8.70325565e-01 1.54042900e-01 -3.42411399e-01 -5.84986269e-01
6.92416430e-01 8.23234677e-01 5.18693209e-01 -1.29574656e+00
5.51386058e-01 1.29654750e-01 -7.77844965e-01 -4.35878009e-01
-1.12006474e+00 -5.68747103e-01 -3.20411325e-01 4.02907073e-01
1.07889211e+00 -9.87759769e-01 -2.24295214e-01 -1.13195218e-01
-1.35705268e+00 -7.33684480e-01 -4.36128139e-01 1.47646606e-01
-7.34787643e-01 5.60269831e-03 -1.02194822e+00 -3.24426055e-01
-5.55478573e-01 -1.21147275e+00 1.45853722e+00 -2.09999178e-02
-3.87742668e-01 -1.16495192e+00 6.56761900e-02 8.69588315e-01
4.71592754e-01 -3.64552975e-01 1.48336160e+00 -1.15968013e+00
-6.80723548e-01 2.60048937e-02 -1.94654062e-01 2.49932393e-01
-3.67446020e-02 -4.97587264e-01 -7.61371672e-01 -4.20341827e-02
6.12416246e-04 -1.23841631e+00 5.35395443e-01 -1.20036602e-01
1.02947211e+00 -3.24453294e-01 -4.86912839e-02 5.20176649e-01
1.58893394e+00 1.71473116e-01 4.42797631e-01 4.34769750e-01
5.52404821e-01 8.29258442e-01 9.22243059e-01 3.49437334e-02
9.19475317e-01 5.22880495e-01 -6.21591806e-02 -5.93101121e-02
-3.71868879e-01 -7.64624059e-01 5.02471626e-01 9.68486011e-01
8.74489963e-01 -2.07556233e-01 -1.25258756e+00 8.27608645e-01
-1.58338678e+00 -3.96892637e-01 1.91624805e-01 1.83812904e+00
1.23103392e+00 -2.45349839e-01 -4.00327474e-01 -1.00065351e+00
3.37452233e-01 -1.63911954e-02 -6.40777946e-01 -5.71608424e-01
-7.96990097e-02 7.79817581e-01 2.89831668e-01 6.36142790e-01
-8.19629073e-01 1.70811152e+00 5.55414724e+00 7.56918490e-01
-8.89107466e-01 5.99792480e-01 6.20244801e-01 6.13651797e-02
-6.95387840e-01 3.81988287e-01 -1.26433539e+00 3.11610520e-01
1.38952446e+00 -1.35382727e-01 5.07282257e-01 9.97806668e-01
-7.02766031e-02 2.00957641e-01 -1.18265522e+00 5.63872635e-01
1.37083232e-01 -1.24365389e+00 1.19210884e-01 -5.19514918e-01
5.55710912e-01 5.30748963e-01 -1.22235157e-01 1.11507201e+00
8.15203428e-01 -8.34444404e-01 3.65955532e-01 -2.11129785e-02
8.78022909e-01 -3.49361181e-01 5.15135467e-01 2.76087344e-01
-8.97208393e-01 4.36051525e-02 -5.52297533e-01 3.70771825e-01
2.39917129e-01 -1.33800671e-01 -9.57034767e-01 6.48903608e-01
8.23817909e-01 4.81028438e-01 -7.32958138e-01 2.38767967e-01
-3.96710604e-01 9.61636603e-01 -1.35777578e-01 -1.84701920e-01
4.41369861e-01 -3.69516090e-02 7.33392537e-02 1.25260699e+00
4.50428724e-01 1.29569203e-01 3.04314613e-01 1.22908032e+00
-3.95473629e-01 5.41499615e-01 -5.43105543e-01 -1.87718689e-01
5.79766035e-01 1.04432344e+00 -2.25439519e-01 -7.17517376e-01
-1.06343305e+00 9.76113677e-01 5.88694155e-01 6.62818670e-01
-7.91901767e-01 -4.94412005e-01 5.58443248e-01 7.88051486e-02
1.47792488e-01 7.40643144e-02 -2.40406618e-01 -1.44044077e+00
2.26165637e-01 -1.38236594e+00 7.65867829e-01 -1.08592057e+00
-1.49597502e+00 5.62464595e-01 -3.54752690e-02 -6.22698545e-01
-5.23479462e-01 -5.82006633e-01 -5.95500469e-01 9.49154615e-01
-1.80402839e+00 -1.44337058e+00 -1.03726551e-01 6.25198781e-01
1.11914933e+00 -3.39700431e-01 1.03521621e+00 2.90001303e-01
-3.92293155e-01 6.71716630e-01 -6.56773821e-02 3.81521672e-01
1.05319345e+00 -1.15092540e+00 9.64385688e-01 7.94988573e-01
2.78479785e-01 8.14225197e-01 3.83533537e-01 -7.67503321e-01
-1.66075385e+00 -1.40080261e+00 1.38791287e+00 -1.05215752e+00
9.26347494e-01 -5.02208829e-01 -1.23511589e+00 1.14953578e+00
3.79508644e-01 1.73745789e-02 7.76416004e-01 2.38629788e-01
-5.62874854e-01 9.51720998e-02 -9.01299179e-01 4.57815140e-01
9.65383232e-01 -9.10721779e-01 -9.27241981e-01 6.99130237e-01
1.38318181e+00 -4.40759718e-01 -9.42819118e-01 3.67982298e-01
8.50079581e-02 -4.95736390e-01 9.19458687e-01 -1.24766099e+00
6.94810510e-01 1.81550443e-01 -4.98934150e-01 -1.01019788e+00
3.03971320e-01 -4.39418554e-01 2.11743191e-01 1.29857910e+00
8.28985929e-01 -6.99230015e-01 7.37398386e-01 8.09536397e-01
-4.00441498e-01 -8.46283138e-01 -8.52918506e-01 -9.94031489e-01
6.50395274e-01 -6.52768970e-01 8.20332527e-01 8.91766369e-01
-6.74382225e-02 7.75482416e-01 2.12911040e-01 -8.60929042e-02
3.69779348e-01 4.19436574e-01 8.79750788e-01 -5.52463472e-01
-5.28629780e-01 6.11146986e-02 3.64424467e-01 -1.14994204e+00
7.30480552e-01 -1.23568833e+00 1.27356797e-01 -1.57729220e+00
2.90451288e-01 -7.10353732e-01 3.63084711e-02 9.13404703e-01
-4.81263638e-01 -1.10637933e-01 -1.97466630e-02 4.99186553e-02
-7.46783257e-01 6.14544213e-01 7.07268715e-01 -7.11349025e-03
4.67131250e-02 -2.22748041e-01 -9.09281611e-01 2.43732721e-01
7.80492485e-01 -6.23905599e-01 -4.95495111e-01 -1.02138460e+00
9.41764042e-02 3.53375256e-01 1.01397566e-01 -5.41732252e-01
-1.09215798e-02 -1.37248904e-01 -1.59339234e-01 -1.48094922e-01
1.53306693e-01 -4.97087985e-01 -3.30124021e-01 1.79788128e-01
-8.18344295e-01 3.82573009e-01 2.86473662e-01 1.39156237e-01
-4.00085598e-01 -5.29664576e-01 5.48236251e-01 -5.20683706e-01
-8.23729217e-01 1.87417567e-01 1.60142649e-02 9.92643535e-01
5.43455184e-01 2.84363270e-01 -7.44525433e-01 -3.87797475e-01
-2.13835999e-01 4.09435421e-01 4.82452184e-01 9.05999899e-01
4.26586807e-01 -1.13636804e+00 -7.81376839e-01 1.20745674e-01
7.37998664e-01 -2.17910618e-01 5.80264516e-02 4.39086825e-01
-4.84189004e-01 8.57226312e-01 1.37390748e-01 -4.23910916e-01
-9.35969114e-01 6.60677493e-01 6.53573796e-02 -3.90758276e-01
-2.93403924e-01 8.51016223e-01 2.94891834e-01 -1.35385084e+00
-7.64569873e-03 -3.00811499e-01 3.11946869e-01 -5.28134525e-01
1.99891850e-01 -5.13574369e-02 2.01560810e-01 -3.17057997e-01
-2.49348208e-01 3.38164240e-01 -2.29189783e-01 -1.98827460e-01
1.17604661e+00 -9.94216353e-02 -1.63721099e-01 2.51012385e-01
1.60372651e+00 -5.77020682e-02 -7.01599717e-01 -7.32083380e-01
5.19460678e-01 -2.14479536e-01 -5.37935495e-01 -1.09216166e+00
-6.32128716e-01 1.05682230e+00 1.82104707e-01 -2.99907535e-01
9.39131379e-01 5.62075436e-01 1.44416738e+00 7.35142350e-01
5.23533762e-01 -1.15811813e+00 2.77563840e-01 8.04972351e-01
6.64142251e-01 -1.57248139e+00 -7.57969439e-01 -3.03733945e-01
-8.64674389e-01 9.13361549e-01 9.93024707e-01 3.11175615e-01
2.03118503e-01 1.92154467e-01 5.10347188e-01 -2.04669148e-01
-1.04857576e+00 -2.61274159e-01 -2.54655927e-02 4.37748820e-01
8.64879072e-01 2.18101516e-02 -4.16707024e-02 8.44939649e-01
-4.65022445e-01 -4.76130582e-02 1.74389899e-01 1.14738095e+00
-3.45824003e-01 -1.55993545e+00 -6.58641160e-02 3.79878432e-01
-7.28676438e-01 -6.77289546e-01 -2.57234156e-01 7.73508489e-01
-3.92239660e-01 9.94528294e-01 -2.55791862e-02 1.28416466e-02
5.54223239e-01 6.87258840e-01 2.38615349e-01 -1.19987261e+00
-7.91121960e-01 -1.75912127e-01 4.12209779e-01 -8.19760859e-01
1.31852254e-01 -5.01352966e-01 -1.44607306e+00 9.39191580e-02
5.41364104e-02 4.37217325e-01 6.38723373e-01 8.15976143e-01
5.86777270e-01 1.91963896e-01 2.58299440e-01 3.44994068e-01
-8.55880558e-01 -8.63356888e-01 6.77559748e-02 6.24447525e-01
-1.36670396e-01 1.34324692e-02 -1.59668252e-02 1.55884787e-01] | [10.8541841506958, 8.8656644821167] |
709b76ad-e707-4fa9-9768-f9e22fe907a8 | fast-dynamic-vision-detection-and-tracking | 2103.05903 | null | https://arxiv.org/abs/2103.05903v2 | https://arxiv.org/pdf/2103.05903v2.pdf | FAST-Dynamic-Vision: Detection and Tracking Dynamic Objects with Event and Depth Sensing | The development of aerial autonomy has enabled aerial robots to fly agilely in complex environments. However, dodging fast-moving objects in flight remains a challenge, limiting the further application of unmanned aerial vehicles (UAVs). The bottleneck of solving this problem is the accurate perception of rapid dynamic objects. Recently, event cameras have shown great potential in solving this problem. This paper presents a complete perception system including ego-motion compensation, object detection, and trajectory prediction for fast-moving dynamic objects with low latency and high precision. Firstly, we propose an accurate ego-motion compensation algorithm by considering both rotational and translational motion for more robust object detection. Then, for dynamic object detection, an event camera-based efficient regression algorithm is designed. Finally, we propose an optimizationbased approach that asynchronously fuses event and depth cameras for trajectory prediction. Extensive real-world experiments and benchmarks are performed to validate our framework. Moreover, our code will be released to benefit related researches. | ['Fei Gao', 'Chao Xu', 'Qianli Dong', 'Zhiwei Zhang', 'Dong Wang', 'Siyuan Wu', 'Haojia Li', 'Botao He'] | 2021-03-10 | null | null | null | null | ['robust-object-detection'] | ['computer-vision'] | [ 7.58859888e-02 -7.00291812e-01 7.55807161e-02 -2.43590206e-01
3.78132053e-03 -8.14299524e-01 2.28473514e-01 -1.84692711e-01
-4.11831528e-01 4.61979061e-01 -5.26104093e-01 -1.94764026e-02
-1.76327199e-01 -6.58358574e-01 -3.54690552e-01 -7.41934001e-01
-1.50121987e-01 8.16036761e-02 9.03331518e-01 -2.21257940e-01
3.45829315e-02 6.54265106e-01 -1.62375915e+00 -2.38683820e-02
6.51889086e-01 1.06952465e+00 6.93338633e-01 9.18825269e-01
7.34435976e-01 4.73929018e-01 -4.52890068e-01 1.30294427e-01
6.56122565e-01 -1.57111272e-01 -3.25905800e-01 3.89629304e-01
2.57119477e-01 -7.17877865e-01 -1.71374738e-01 9.46755946e-01
4.45552349e-01 4.42988157e-01 1.91458255e-01 -1.44303524e+00
1.79600045e-01 -5.34404367e-02 -5.20769000e-01 2.78598875e-01
3.72471601e-01 4.01013136e-01 5.27303576e-01 -7.13115871e-01
4.74660158e-01 9.28415537e-01 4.58254665e-01 2.00346693e-01
-5.70536673e-01 -3.42398196e-01 3.81245881e-01 3.96759272e-01
-1.40132201e+00 -2.49208093e-01 3.30314457e-01 -5.36855936e-01
9.15370703e-01 2.51662731e-01 1.03716516e+00 5.71530163e-01
5.59879184e-01 5.56449175e-01 3.96999151e-01 -6.85717724e-03
1.56049863e-01 -2.20913410e-01 -3.58057737e-01 7.11603463e-01
5.57544529e-01 2.67472118e-01 -1.40906319e-01 1.29317999e-01
6.19913101e-01 4.82613236e-01 -6.64220870e-01 -6.35453880e-01
-1.66930187e+00 2.82376230e-01 4.52436566e-01 -1.56771287e-01
-5.01269639e-01 1.26878247e-01 2.21655875e-01 1.21029794e-01
1.46140615e-02 1.93057373e-01 -4.47464108e-01 -1.77599162e-01
-7.87610829e-01 3.82426143e-01 5.63263118e-01 1.52211404e+00
3.65809083e-01 2.55830020e-01 2.00838819e-01 2.65842229e-01
3.55190068e-01 8.41035008e-01 4.19458717e-01 -9.07044411e-01
4.44319338e-01 8.28367591e-01 5.97117841e-01 -1.26774490e+00
-5.59515536e-01 -2.51349092e-01 -7.29699433e-01 3.66050065e-01
1.71221681e-02 -3.80264938e-01 -5.02572060e-01 1.07704902e+00
7.28410721e-01 1.96559772e-01 2.05253065e-01 1.43576717e+00
3.35247993e-01 9.12113547e-01 -4.05679613e-01 -3.22862834e-01
1.19162548e+00 -1.05913675e+00 -5.12753665e-01 -3.08118641e-01
4.97803539e-01 -8.04226041e-01 5.52716136e-01 5.03126562e-01
-9.12606239e-01 -7.20359385e-01 -1.29384303e+00 1.97945818e-01
-7.70497508e-03 6.85253918e-01 4.19896245e-01 2.64381707e-01
-6.40706360e-01 3.64600956e-01 -1.19334996e+00 -5.60764611e-01
2.20207293e-02 6.13917291e-01 -2.27960318e-01 -2.68653035e-02
-6.05981767e-01 6.37703896e-01 5.56139112e-01 2.19024077e-01
-1.06343174e+00 -2.87611276e-01 -6.20199859e-01 -1.98442176e-01
6.15118325e-01 -9.58080947e-01 1.35673153e+00 -7.48713076e-01
-1.55929577e+00 4.80870485e-01 2.99849417e-02 -5.98864615e-01
5.06693363e-01 -4.56387818e-01 -3.04560542e-01 1.91106677e-01
1.85188223e-02 5.95343649e-01 1.08646131e+00 -8.88346016e-01
-1.41650975e+00 -4.23065960e-01 2.85071194e-01 6.37841880e-01
-2.41952926e-01 -2.23569959e-01 -3.59908760e-01 -3.74142379e-01
4.01852101e-01 -1.18423045e+00 -2.54991263e-01 2.65689939e-01
1.11461561e-02 1.61357224e-01 1.13118303e+00 -1.79149151e-01
1.05093682e+00 -2.16374159e+00 4.85001087e-01 -3.18029106e-01
2.07968485e-02 2.35794157e-01 1.89166442e-01 3.46877128e-01
3.41390491e-01 -6.90815508e-01 -8.92115235e-02 -8.23981315e-02
-3.64932418e-01 -1.83587968e-02 -6.89157784e-01 6.48723960e-01
-9.96101797e-02 3.35334003e-01 -8.64868641e-01 -2.92760491e-01
4.80859965e-01 1.23028584e-01 -4.52147752e-01 2.17250034e-01
-1.73126101e-01 3.16888630e-01 -5.89008570e-01 1.14484203e+00
7.98160851e-01 1.16015747e-01 3.42682116e-02 -2.44046986e-01
-7.70879924e-01 -3.51968914e-01 -1.39015567e+00 1.44408369e+00
-2.32517108e-01 6.83464408e-01 3.81836027e-01 -5.05632341e-01
8.21314752e-01 1.13242775e-01 5.78910887e-01 -5.57173491e-02
4.88074839e-01 1.40410155e-01 -1.29564852e-01 -5.24397552e-01
9.94782090e-01 1.25489518e-01 7.88732693e-02 -2.24301875e-01
-2.66197950e-01 -4.72740352e-01 1.92655548e-01 -1.46564856e-01
1.06662393e+00 4.31911111e-01 4.79636908e-01 3.12521122e-02
5.75800955e-01 7.59242356e-01 9.33980882e-01 1.28781334e-01
-4.73135620e-01 3.14856976e-01 -2.93755591e-01 -7.23631442e-01
-5.67121327e-01 -7.69872189e-01 1.52697787e-01 5.61372101e-01
9.43229854e-01 -4.59355384e-01 -5.14903128e-01 -3.42431962e-01
5.74145131e-02 2.95870662e-01 1.65983170e-01 -1.12786829e-01
-4.39262956e-01 -7.97480881e-01 -3.64954025e-02 5.74583471e-01
6.50669098e-01 -5.32466471e-01 -1.74206758e+00 3.91650707e-01
-3.32087159e-01 -1.48313189e+00 -2.84129918e-01 -2.18834251e-01
-1.08163953e+00 -1.17076087e+00 -5.87193489e-01 -6.11836553e-01
6.34887934e-01 1.40796316e+00 4.01708841e-01 9.09173638e-02
-6.93407416e-01 6.67252421e-01 -5.22494912e-01 -5.35880208e-01
2.74195701e-01 -2.45264038e-01 6.46308720e-01 -9.63483229e-02
-1.20087661e-01 -2.22734287e-01 -7.59586692e-01 8.46448958e-01
-6.74606085e-01 5.56743629e-02 5.94048142e-01 3.87543857e-01
8.42042565e-01 4.87559438e-01 -1.85997203e-01 4.36941115e-03
4.20552269e-02 -2.95440972e-01 -1.43620121e+00 1.53343931e-01
-2.56658494e-01 -7.75596797e-01 8.23949277e-01 -4.91880238e-01
-8.06636572e-01 6.58929765e-01 5.78472674e-01 -5.97023606e-01
-1.60938218e-01 2.31818527e-01 -1.17193483e-01 -3.21072161e-01
4.08900678e-01 1.50921658e-01 -3.15970689e-01 8.51739794e-02
-1.42602259e-02 6.15164459e-01 5.24808407e-01 8.04705825e-03
9.86529350e-01 8.72794628e-01 2.60553181e-01 -1.19978273e+00
-4.54486519e-01 -6.57082319e-01 -8.19122136e-01 -5.76970816e-01
7.86870956e-01 -1.42728317e+00 -8.84299338e-01 4.52482313e-01
-1.22397995e+00 -2.22429767e-01 -1.13280818e-01 1.08317125e+00
-6.92937136e-01 4.23391521e-01 -6.71977177e-02 -8.61570179e-01
-1.87757656e-01 -1.18203270e+00 1.05844867e+00 4.76402968e-01
1.76887497e-01 -4.15558249e-01 1.53235048e-01 2.17806160e-01
6.85492605e-02 2.73417801e-01 -2.44612396e-02 9.06183049e-02
-1.40659535e+00 -6.41482294e-01 -1.15983345e-01 -1.57672063e-01
1.16112605e-01 3.63548309e-01 -4.46570605e-01 -6.66419923e-01
7.75602460e-02 1.63146004e-01 6.14749789e-01 4.27560627e-01
6.24932587e-01 3.49411219e-02 -7.37053096e-01 9.85926986e-01
1.45047784e+00 5.11959732e-01 9.11367685e-02 3.31745893e-01
5.91101229e-01 4.13513720e-01 1.76481760e+00 9.65417087e-01
4.35736567e-01 9.06590819e-01 1.01864433e+00 4.03374285e-01
3.62404436e-01 1.55046225e-01 7.21091628e-01 7.61288226e-01
-5.46655953e-01 -6.94802999e-01 -6.91629946e-01 5.16155183e-01
-2.04681015e+00 -9.35187101e-01 -4.80651617e-01 2.25592232e+00
-1.37488872e-01 -5.70253171e-02 -1.95198003e-02 -5.53032979e-02
5.94237208e-01 6.92335516e-02 -6.58296406e-01 2.50230074e-01
1.33556321e-01 -7.62600958e-01 8.07856381e-01 2.71189570e-01
-1.37714648e+00 1.19810247e+00 5.16575766e+00 3.38222533e-01
-1.10428739e+00 -2.06594437e-01 -4.85076249e-01 -1.81883529e-01
4.94427830e-01 2.25990444e-01 -1.28180492e+00 3.17511499e-01
5.37786305e-01 -1.67039469e-01 4.09574091e-01 1.25660932e+00
2.06343740e-01 -4.15906996e-01 -6.37202382e-01 1.09489763e+00
7.90232196e-02 -9.34622347e-01 -1.67814106e-01 -8.49049985e-02
5.97269177e-01 -1.00646175e-01 -1.57254279e-01 -3.61956172e-02
4.87691425e-02 -1.16408803e-02 8.14753652e-01 6.24023736e-01
4.08005953e-01 -7.53454208e-01 7.41973281e-01 6.88039839e-01
-1.64554596e+00 -4.75094080e-01 -1.04974902e+00 -2.88202465e-01
3.48892391e-01 4.22706723e-01 -9.18612123e-01 6.79939389e-01
8.16290855e-01 1.02345109e+00 -3.15214694e-01 1.38427997e+00
-1.59475610e-01 2.32091937e-02 -3.69810075e-01 -9.03104022e-02
1.30302295e-01 -3.34830374e-01 1.11400056e+00 6.48322821e-01
9.26321745e-01 6.68617606e-01 5.31470656e-01 3.87792528e-01
3.84920388e-01 -1.32804886e-01 -7.88019001e-01 -4.06828448e-02
2.79751450e-01 1.57506645e+00 -8.48611593e-01 -1.31772831e-01
-3.79111111e-01 1.25726676e+00 -2.23751545e-01 -2.43978892e-02
-1.14656591e+00 -5.87194145e-01 7.79434264e-01 1.32335126e-01
3.26615393e-01 -9.61195111e-01 4.37357098e-01 -1.31113505e+00
3.89817618e-02 -4.02405679e-01 1.61966980e-01 -1.00932693e+00
-4.81722772e-01 6.98461533e-01 -6.79922327e-02 -2.06987643e+00
-3.48333828e-02 -7.22675145e-01 -4.52802747e-01 2.80986071e-01
-1.41684687e+00 -1.07039499e+00 -1.07687962e+00 5.54823577e-01
9.09044623e-01 -2.97141433e-01 5.83242118e-01 2.70857587e-02
-6.77325726e-01 -2.31184125e-01 -8.40749815e-02 -4.68680203e-01
7.86752999e-01 -8.45929146e-01 1.17546126e-01 1.30806804e+00
-1.50117859e-01 1.91275328e-01 6.67232633e-01 -9.85105813e-01
-1.93159294e+00 -1.19867730e+00 2.47365683e-01 -3.89116704e-01
3.55308503e-01 1.33941444e-02 -4.11560059e-01 7.67562628e-01
-1.17944948e-01 1.90787427e-02 2.53402621e-01 -6.90381825e-01
5.45400202e-01 -4.77349967e-01 -8.50732386e-01 6.09949827e-01
1.03950739e+00 2.34391794e-01 -2.06673786e-01 5.03303349e-01
8.82278919e-01 -8.39255095e-01 -5.49957991e-01 6.81746185e-01
5.81125736e-01 -1.03649676e+00 9.67434704e-01 -1.60423905e-01
-6.27440214e-02 -9.97622788e-01 -3.36075485e-01 -1.06170058e+00
-4.12305623e-01 -7.32180059e-01 -1.62576705e-01 7.84417570e-01
-2.46361747e-01 -3.95230532e-01 5.55282533e-01 2.01672003e-01
-5.60747147e-01 -4.32383388e-01 -6.44575715e-01 -9.40794587e-01
-1.04142463e+00 -1.57695755e-01 4.02834147e-01 5.73900282e-01
-1.32027626e-01 5.90853132e-02 -5.13234556e-01 9.73624885e-01
5.80746233e-01 6.00075722e-01 1.11571169e+00 -1.16663444e+00
3.25814821e-02 8.87691975e-02 -7.99289167e-01 -1.24941802e+00
-2.68349439e-01 -4.07276601e-01 3.91023010e-01 -1.41983283e+00
-3.41935515e-01 8.30126181e-02 3.10561389e-01 5.00767119e-02
9.88974199e-02 -1.21375330e-01 1.90514624e-01 3.49997222e-01
-7.94658899e-01 6.68900907e-01 1.13635516e+00 9.89963710e-02
-3.65758836e-01 4.89199728e-01 1.49789557e-01 1.03253484e+00
8.66363823e-01 -4.05821890e-01 -4.85395104e-01 -6.23038054e-01
1.01555087e-01 4.69805479e-01 6.10605717e-01 -1.69530261e+00
7.67115414e-01 -4.30735707e-01 4.74799067e-01 -1.11960113e+00
7.26944745e-01 -1.44643331e+00 -1.61384344e-02 7.99650371e-01
4.87245888e-01 5.13955116e-01 3.79337281e-01 8.37335885e-01
-2.33296692e-01 -8.35889429e-02 7.87804008e-01 -2.84541445e-03
-1.07202506e+00 5.66293836e-01 -7.57923961e-01 -3.33664954e-01
1.81265426e+00 -3.86352271e-01 -2.09922269e-01 -1.58301800e-01
-4.37430769e-01 5.36838293e-01 7.52736032e-01 3.96843910e-01
1.14447272e+00 -1.00269234e+00 -2.89376885e-01 3.78706485e-01
1.44823954e-01 1.92311749e-01 5.36712408e-01 1.09102583e+00
-8.68857622e-01 5.44609606e-01 -4.61511284e-01 -7.79162884e-01
-1.54909277e+00 6.42430007e-01 1.14832453e-01 3.90370935e-01
-4.51770246e-01 7.04399467e-01 4.62580249e-02 -6.31624311e-02
1.71414874e-02 -6.84733927e-01 -7.56669184e-03 -4.28267680e-02
5.30099571e-01 8.03792596e-01 -2.92393208e-01 -6.16624951e-01
-5.65251589e-01 9.20107484e-01 2.04020485e-01 -1.42076509e-02
1.07652962e+00 -5.93466222e-01 5.76524027e-02 2.51625687e-01
4.03057456e-01 -5.24494424e-02 -1.52938080e+00 1.96281239e-01
-3.04525077e-01 -6.83409393e-01 1.34956151e-01 -2.52586275e-01
-7.76021600e-01 9.66170430e-01 6.80541635e-01 4.66644317e-02
1.33888364e+00 -7.50385940e-01 8.88163984e-01 8.44749570e-01
8.84125531e-01 -9.94590878e-01 -8.47883150e-03 5.59949934e-01
6.99330986e-01 -1.27233660e+00 4.18072373e-01 -6.11555517e-01
-7.92219520e-01 1.33469510e+00 9.19869840e-01 -1.65468633e-01
2.96252936e-01 1.63008139e-01 -1.73041284e-01 8.09633061e-02
-8.90408874e-01 -2.78482258e-01 1.09585315e-01 4.59464163e-01
-2.47825503e-01 7.39637241e-02 9.32978317e-02 4.60679561e-01
8.73590931e-02 5.38013466e-02 6.87956393e-01 1.20494866e+00
-8.71310711e-01 -7.42439866e-01 -5.12984276e-01 -8.94193277e-02
4.68051657e-02 2.39604220e-01 -2.66169816e-01 8.56719971e-01
2.53761023e-01 9.30715978e-01 -6.58907443e-02 -7.01871455e-01
6.26335800e-01 -4.93166387e-01 4.16330755e-01 -6.04875684e-01
-2.09555134e-01 7.77531415e-02 -3.90131325e-01 -7.68510997e-01
-3.63513380e-01 -7.59056389e-01 -1.42500699e+00 1.37713794e-02
-4.93983686e-01 -2.11882114e-01 8.19314659e-01 4.88821775e-01
5.42383552e-01 5.41574121e-01 6.81922495e-01 -9.63960469e-01
-5.72555363e-01 -3.72282654e-01 -4.66105968e-01 -4.50399965e-01
3.85169297e-01 -8.25678766e-01 -2.74864256e-01 2.58651197e-01] | [7.2083821296691895, -1.75824773311615] |
13fee72e-548a-4daa-9638-c971184147bd | explaining-and-adapting-graph-conditional | 2306.03256 | null | https://arxiv.org/abs/2306.03256v1 | https://arxiv.org/pdf/2306.03256v1.pdf | Explaining and Adapting Graph Conditional Shift | Graph Neural Networks (GNNs) have shown remarkable performance on graph-structured data. However, recent empirical studies suggest that GNNs are very susceptible to distribution shift. There is still significant ambiguity about why graph-based models seem more vulnerable to these shifts. In this work we provide a thorough theoretical analysis on it by quantifying the magnitude of conditional shift between the input features and the output label. Our findings show that both graph heterophily and model architecture exacerbate conditional shifts, leading to performance degradation. To address this, we propose an approach that involves estimating and minimizing the conditional shift for unsupervised domain adaptation on graphs. In our controlled synthetic experiments, our algorithm demonstrates robustness towards distribution shift, resulting in up to 10% absolute ROC AUC improvement versus the second-best algorithm. Furthermore, comprehensive experiments on both node classification and graph classification show its robust performance under various distribution shifts. | ['Bryan Perozzi', 'Jiawei Han', 'Natalia Ponomareva', 'Yizhu Jiao', 'Qi Zhu'] | 2023-06-05 | null | null | null | null | ['graph-classification', 'unsupervised-domain-adaptation'] | ['graphs', 'methodology'] | [ 3.97162169e-01 3.63312215e-01 -3.61642241e-01 -3.56739044e-01
-1.71569034e-01 -5.75145006e-01 5.24222553e-01 5.24544120e-01
-1.49315894e-01 7.89209187e-01 -9.48475674e-02 -5.05038679e-01
-3.34201694e-01 -9.05209303e-01 -7.94123232e-01 -5.20701349e-01
-4.30520177e-01 4.01054740e-01 1.60441384e-01 3.96488905e-02
1.17424078e-01 4.07838464e-01 -1.00753164e+00 -1.13932349e-01
8.92655551e-01 6.25805736e-01 -3.01199377e-01 5.83561897e-01
-4.55054902e-02 6.13283694e-01 -7.76990294e-01 -5.32548666e-01
2.75860280e-01 -3.56463432e-01 -5.53987384e-01 -3.67789604e-02
5.80502093e-01 9.93397180e-03 -3.43159437e-01 1.29270351e+00
4.38063890e-01 -1.97282005e-02 9.33396578e-01 -1.53770971e+00
-7.78406680e-01 1.02421951e+00 -4.82251853e-01 3.81442934e-01
-2.44792216e-02 1.72869340e-01 1.16061223e+00 -4.84218508e-01
7.53058016e-01 1.05990231e+00 1.13283050e+00 2.57766455e-01
-1.68412781e+00 -5.54284990e-01 1.55937567e-01 -2.09914431e-01
-1.51780248e+00 -2.37877890e-01 9.07868445e-01 -4.06831235e-01
9.08808768e-01 -4.94977795e-02 4.41063106e-01 1.15255010e+00
4.05328304e-01 3.79855901e-01 9.61389601e-01 -5.09203970e-01
2.64805555e-01 3.54001462e-01 2.92608231e-01 5.34349680e-01
9.11815107e-01 2.09228345e-03 -3.32617074e-01 -1.73717111e-01
6.92070544e-01 -1.63555026e-01 -6.37115687e-02 -6.99662507e-01
-7.06099153e-01 9.02628958e-01 5.88793337e-01 3.43747258e-01
-2.37746462e-01 2.48993158e-01 3.79209965e-01 5.79744160e-01
6.50933504e-01 6.04434907e-01 -3.75896364e-01 3.22284669e-01
-6.14197493e-01 -9.78753045e-02 9.77730215e-01 9.39088285e-01
7.76650667e-01 1.86729610e-01 1.19460905e-02 7.81863451e-01
2.96104759e-01 1.85662851e-01 1.48757324e-01 -4.55307186e-01
5.04526079e-01 8.64387333e-01 -3.85207951e-01 -1.49192429e+00
-7.53220320e-01 -8.81424367e-01 -1.10635960e+00 -1.61253676e-01
6.27679408e-01 -2.08413467e-01 -9.31725085e-01 2.11751366e+00
8.69053826e-02 -1.08992904e-01 -8.76898840e-02 3.57975930e-01
6.48773015e-01 2.24368811e-01 4.77753818e-01 -3.02158762e-02
7.55374014e-01 -5.17365515e-01 -5.45622051e-01 -3.70015323e-01
8.83774996e-01 -3.16040963e-01 1.08304083e+00 3.24487723e-02
-6.77087963e-01 -1.90830663e-01 -1.27706861e+00 4.91627961e-01
-4.27263945e-01 -2.77072400e-01 6.14946127e-01 1.12342751e+00
-1.22443998e+00 9.22649741e-01 -5.21641493e-01 -9.08296406e-01
3.87745678e-01 5.44545233e-01 -2.97689140e-01 -7.16088042e-02
-1.19060791e+00 6.38393581e-01 5.28917074e-01 -2.54449368e-01
-3.94979864e-01 -6.10644102e-01 -7.22531557e-01 2.64333755e-01
4.18214202e-01 -6.27032220e-01 7.94852793e-01 -1.00383639e+00
-1.11312938e+00 6.52966261e-01 3.09793234e-01 -7.04779804e-01
3.57050389e-01 4.35840249e-01 -4.66571420e-01 -1.84150890e-01
-2.62042563e-02 3.56287301e-01 7.16118097e-01 -1.29700673e+00
-2.86433529e-02 -3.84393245e-01 -1.14277601e-01 8.10767934e-02
-8.17003608e-01 -2.75411904e-01 -1.30488500e-01 -6.99839532e-01
1.12561077e-01 -9.62122142e-01 -2.91254401e-01 -4.91191268e-01
-6.33991778e-01 -3.11918467e-01 3.91555429e-01 -2.88615912e-01
1.55620372e+00 -2.09188461e+00 -3.37080568e-01 6.66490972e-01
6.98514581e-01 -2.85133664e-02 -2.21092284e-01 6.08322501e-01
-1.73773676e-01 4.77998286e-01 -2.41575092e-01 5.95024377e-02
6.64836243e-02 8.92452747e-02 3.32333669e-02 5.37459314e-01
9.56666917e-02 8.48465979e-01 -6.73144281e-01 -3.29094499e-01
-2.56487161e-01 2.31798992e-01 -4.79812264e-01 -2.37279192e-01
-1.00226313e-01 3.86709832e-02 -1.97917417e-01 7.31039047e-01
6.78311348e-01 -8.49331856e-01 8.49097013e-01 2.00483799e-02
5.93054712e-01 7.89107159e-02 -9.49982703e-01 8.08045626e-01
-2.06930991e-02 7.31717050e-01 -7.91156068e-02 -1.13865006e+00
1.06931913e+00 -1.94892928e-01 3.54017794e-01 -6.17681801e-01
1.49352312e-01 -1.02376090e-02 4.74119455e-01 -7.08037317e-02
4.06368166e-01 -1.54669881e-02 -1.60092652e-01 6.05118573e-01
1.69533834e-01 1.32219553e-01 2.19056457e-01 3.98657620e-01
1.49254155e+00 -5.14520228e-01 4.33853596e-01 -6.54435098e-01
2.20187884e-02 -4.97690812e-02 3.69930625e-01 1.07040393e+00
-4.41575140e-01 2.00387657e-01 9.81343687e-01 -4.72345799e-01
-9.36707735e-01 -1.09173882e+00 1.05863377e-01 1.11640275e+00
9.79713276e-02 -5.01573384e-01 -9.07365978e-01 -9.18773592e-01
2.66631991e-01 7.45211661e-01 -7.72893012e-01 -5.02011657e-01
-3.15584779e-01 -1.04789746e+00 6.94328725e-01 5.81631541e-01
1.99574128e-01 -8.60091269e-01 -9.40303579e-02 6.99469969e-02
1.68821588e-01 -1.09312427e+00 -3.20397288e-01 3.94279659e-01
-1.09678066e+00 -9.75888789e-01 -4.14804548e-01 -7.04584837e-01
1.00521839e+00 9.72865820e-02 1.41038692e+00 3.58311862e-01
-2.26100367e-02 4.10712838e-01 -1.40607610e-01 -2.91701972e-01
-7.32972205e-01 6.90154493e-01 8.25831220e-02 -2.57704496e-01
3.38974297e-01 -8.01379442e-01 -4.06262130e-01 3.20741475e-01
-8.85709405e-01 -3.82820249e-01 6.38404310e-01 6.82233095e-01
2.23693371e-01 2.56476611e-01 7.66337991e-01 -1.37289345e+00
1.43594563e+00 -7.30711222e-01 -4.61011887e-01 2.98736960e-01
-1.43283737e+00 1.03078865e-01 7.26044476e-01 -3.56675506e-01
-6.86343551e-01 -3.67818139e-02 3.81938159e-01 -1.36323243e-01
-7.23642334e-02 8.68820906e-01 -3.48007120e-02 -2.16833681e-01
1.27351606e+00 -1.40405670e-01 1.54464841e-01 -6.12197556e-02
2.30198726e-01 3.70327562e-01 1.18704967e-01 -4.08019423e-01
8.97952616e-01 1.38206333e-01 1.50583327e-01 -7.86751032e-01
-4.41639870e-01 -2.65479255e-02 -5.94360948e-01 -3.87595862e-01
3.59631598e-01 -5.24214327e-01 -2.66637683e-01 3.50647002e-01
-7.17963755e-01 -7.12946057e-01 6.69156015e-03 9.93414149e-02
-2.06466809e-01 4.47951049e-01 -4.69367385e-01 -5.99651098e-01
-2.73539484e-01 -8.02626014e-01 2.26184756e-01 2.26966426e-01
-3.77138257e-01 -1.53464079e+00 8.31859186e-03 -2.01390311e-02
6.39445186e-01 4.26767379e-01 1.27912152e+00 -1.12214100e+00
-4.24257010e-01 -2.24852338e-01 -4.41042423e-01 1.69753596e-01
2.39090264e-01 6.71675727e-02 -6.61110997e-01 -6.00060523e-01
-5.56437135e-01 -2.64600199e-03 7.80763626e-01 6.18651032e-01
8.59091997e-01 -2.21540406e-01 -6.13883913e-01 4.20760334e-01
1.59334767e+00 4.93888892e-02 3.78298730e-01 2.09289938e-01
7.40202129e-01 6.29534960e-01 1.28821343e-01 8.09737518e-02
2.40839422e-01 2.59521127e-01 3.40092629e-01 -1.37377411e-01
-2.05479667e-01 -4.12004262e-01 1.43079266e-01 6.68547034e-01
1.96059570e-01 -8.07941198e-01 -1.33459473e+00 5.65818250e-01
-1.58132017e+00 -4.64071155e-01 -2.47802153e-01 2.21838355e+00
5.06087959e-01 5.32703161e-01 2.81377316e-01 -1.01226047e-01
1.00592124e+00 1.68131769e-01 -7.42301047e-01 -3.24256510e-01
-1.25710070e-01 -4.12568799e-04 8.00237477e-01 2.03351021e-01
-9.58722353e-01 9.89718378e-01 7.64799213e+00 6.12919867e-01
-9.68444765e-01 -2.57563859e-01 8.58360052e-01 3.19658339e-01
-3.43970180e-01 -5.99669665e-02 -3.10257703e-01 3.27709854e-01
1.17494011e+00 -4.41595942e-01 4.29164708e-01 9.34746325e-01
-2.39688382e-01 4.15242873e-02 -1.15246153e+00 5.67494452e-01
4.61241938e-02 -1.23537135e+00 2.82832272e-02 3.42972338e-01
8.66169870e-01 2.71527320e-01 5.25961667e-02 2.54789799e-01
7.57611990e-01 -1.12892377e+00 1.02150112e-01 1.93364918e-01
8.52496743e-01 -7.28172123e-01 7.19725668e-01 1.69075266e-01
-9.57295954e-01 1.97147112e-02 -2.23745048e-01 -1.30161524e-01
-2.98924923e-01 7.34383941e-01 -1.38777006e+00 2.82664746e-01
5.97257853e-01 4.95274931e-01 -1.00359774e+00 7.83984601e-01
-7.24156722e-02 8.95402730e-01 -3.20183009e-01 -3.69863510e-01
-9.24470648e-02 -5.11424206e-02 5.16961992e-01 1.12548590e+00
1.73766404e-01 -4.80702043e-01 5.66983595e-02 7.77158201e-01
-3.95430297e-01 1.86809406e-01 -1.09789038e+00 -4.13750380e-01
6.14041328e-01 1.06700933e+00 -1.36171806e+00 -1.65859044e-01
-2.07151845e-01 6.85035646e-01 6.59286141e-01 5.66172063e-01
-6.86305404e-01 -5.08491337e-01 2.61101186e-01 2.96989083e-01
6.50543272e-02 -1.75292149e-01 -5.67988634e-01 -8.45572531e-01
-9.85206291e-02 -7.73169458e-01 6.91542625e-01 -2.34995708e-01
-1.67198563e+00 4.91458535e-01 -5.48587069e-02 -8.02533746e-01
-5.12645021e-02 -6.46498263e-01 -6.27604306e-01 4.22749519e-01
-1.08871377e+00 -8.60274494e-01 -3.26059312e-01 2.62651563e-01
8.85533169e-03 -2.79295534e-01 6.10349000e-01 2.06156820e-01
-6.79570735e-01 1.03422868e+00 2.35651314e-01 2.18709886e-01
8.37893426e-01 -1.35747623e+00 6.64383411e-01 8.40334117e-01
1.33648604e-01 7.67668962e-01 8.24314475e-01 -8.91557693e-01
-1.03889322e+00 -1.15365589e+00 6.57282174e-01 -4.53883648e-01
7.86693037e-01 -3.84784818e-01 -9.31265593e-01 8.06437492e-01
9.99400690e-02 -1.40038967e-01 8.44820023e-01 4.32279289e-01
-5.65939367e-01 -9.95051414e-02 -1.31006730e+00 7.22100437e-01
1.39670217e+00 -3.45207244e-01 -1.10709280e-01 2.72027493e-01
5.29884040e-01 -3.62569641e-04 -1.08650970e+00 5.71255922e-01
4.04868901e-01 -1.01796615e+00 7.05322802e-01 -5.41647315e-01
3.28545332e-01 1.58404205e-02 -8.35158527e-02 -1.53450191e+00
-5.62413454e-01 -5.70942104e-01 6.68100566e-02 1.13758421e+00
7.77755558e-01 -9.50489759e-01 1.02045262e+00 6.46944463e-01
3.28282088e-01 -5.23735464e-01 -7.15468466e-01 -9.06754553e-01
2.37293378e-01 -1.83101192e-01 4.36341971e-01 1.26236570e+00
-1.23607705e-03 5.94215751e-01 -7.70206898e-02 7.70206004e-02
8.02251995e-01 -1.95163161e-01 8.93419445e-01 -1.64231634e+00
-6.41656294e-02 -6.27207398e-01 -4.78219301e-01 -6.14968240e-01
2.42057368e-01 -1.21556866e+00 -1.38958916e-01 -1.50507009e+00
2.24368706e-01 -4.65623409e-01 -5.65059423e-01 3.12584847e-01
-1.02928989e-01 1.06963567e-01 -4.41321842e-02 1.28585801e-01
-5.93072832e-01 1.81150615e-01 8.48034620e-01 -1.64700598e-01
-3.60544264e-01 -7.51529187e-02 -9.14932787e-01 5.70534170e-01
1.27069402e+00 -7.83665895e-01 -6.88474238e-01 -2.66369134e-01
5.22305727e-01 -2.83590257e-01 1.32299379e-01 -1.04375589e+00
1.48962513e-01 -1.01707742e-01 4.06121165e-01 -1.84682027e-01
-2.53167421e-01 -6.35788321e-01 1.43508419e-01 5.79539061e-01
-4.63207155e-01 1.86388403e-01 3.20010394e-01 9.24552321e-01
2.13026866e-01 -6.31406344e-03 7.07639575e-01 4.11107875e-02
-3.90633583e-01 2.27656499e-01 -4.26194876e-01 1.93931341e-01
8.16755474e-01 -3.48651797e-01 -4.65538532e-01 -6.89251721e-01
-6.95057094e-01 9.34833363e-02 6.50944769e-01 3.31488818e-01
2.25634620e-01 -1.09136188e+00 -4.36009139e-01 2.16023311e-01
2.12910846e-01 -4.46682423e-01 5.57041653e-02 7.56011367e-01
-4.36219186e-01 1.09633058e-01 -1.51491225e-01 -5.73008597e-01
-1.22644305e+00 4.74037856e-01 3.84392917e-01 -4.91343707e-01
-3.30953717e-01 8.62488031e-01 4.30938043e-03 -7.97546983e-01
1.86564863e-01 -3.01497996e-01 6.12123981e-02 -6.31609261e-02
-2.82731175e-01 3.69095892e-01 6.36205897e-02 -2.94775754e-01
-3.97823036e-01 1.96519673e-01 -1.70551613e-01 2.27296218e-01
1.18912590e+00 -1.04838319e-01 -1.23448446e-01 2.64896482e-01
1.01272154e+00 6.18762057e-03 -1.06897271e+00 -2.48343691e-01
4.25603688e-01 -1.01935051e-01 -1.64525077e-01 -8.04870784e-01
-9.52976227e-01 5.15178740e-01 4.08558577e-01 9.49288785e-01
9.67216492e-01 2.86904685e-02 3.08111370e-01 4.90392774e-01
2.13730678e-01 -1.18006563e+00 1.04482174e-01 3.95639032e-01
5.01253545e-01 -1.12251198e+00 2.36454442e-01 -5.47885120e-01
-3.62687856e-01 9.62258279e-01 8.84174407e-01 -1.30447879e-01
7.47048438e-01 1.43139601e-01 6.32462353e-02 -4.41770583e-01
-6.89990461e-01 1.76275268e-01 9.66619700e-02 9.19996440e-01
2.60717541e-01 2.70279318e-01 -1.29271448e-01 3.99801940e-01
-2.53111482e-01 -3.68368506e-01 5.79487860e-01 6.27837241e-01
-1.55869588e-01 -9.51258421e-01 -7.16249421e-02 7.09933817e-01
-3.64154845e-01 -8.55061710e-02 -9.60640907e-01 1.13048697e+00
-3.83855075e-01 8.69738519e-01 1.29480632e-02 -6.56861365e-01
1.77727789e-01 2.23300457e-01 3.63746136e-01 -4.48290467e-01
-5.19576311e-01 -2.33165652e-01 2.98587590e-01 -2.79678911e-01
-1.17640667e-01 -4.31597352e-01 -1.03364635e+00 -3.93601447e-01
-5.04783809e-01 -9.35120285e-02 4.12068367e-01 6.12148523e-01
5.99620819e-01 5.95381260e-01 3.85993093e-01 -2.71404058e-01
-6.66829228e-01 -1.03870308e+00 -6.23243272e-01 5.69830298e-01
3.90064828e-02 -6.02242291e-01 -7.14792669e-01 -2.98751146e-01] | [6.91973876953125, 6.119566917419434] |
d291aa4d-900b-4429-8cfc-6dc21ac2d2d4 | siminet-a-novel-method-for-quantifying-brain | 1709.07211 | null | http://arxiv.org/abs/1709.07211v1 | http://arxiv.org/pdf/1709.07211v1.pdf | SimiNet: a Novel Method for Quantifying Brain Network Similarity | Quantifying the similarity between two networks is critical in many
applications. A number of algorithms have been proposed to compute graph
similarity, mainly based on the properties of nodes and edges. Interestingly,
most of these algorithms ignore the physical location of the nodes, which is a
key factor in the context of brain networks involving spatially defined
functional areas. In this paper, we present a novel algorithm called "SimiNet"
for measuring similarity between two graphs whose nodes are defined a priori
within a 3D coordinate system. SimiNet provides a quantified index (ranging
from 0 to 1) that accounts for node, edge and spatiality features. Complex
graphs were simulated to evaluate the performance of SimiNet that is compared
with eight state-of-art methods. Results show that SimiNet is able to detect
weak spatial variations in compared graphs in addition to computing similarity
using both nodes and edges. SimiNet was also applied to real brain networks
obtained during a visual recognition task. The algorithm shows high performance
to detect spatial variation of brain networks obtained during a naming task of
two categories of visual stimuli: animals and tools. A perspective to this work
is a better understanding of object categorization in the human brain. | [] | 2017-09-21 | null | null | null | null | ['object-categorization', 'graph-similarity'] | ['computer-vision', 'graphs'] | [-2.59786267e-02 5.38176820e-02 3.70958507e-01 -2.51532048e-01
5.76081514e-01 -6.64986968e-01 7.69016743e-01 7.04199374e-01
-5.37168443e-01 2.14600921e-01 -2.40750298e-01 1.07511980e-02
-6.73506141e-01 -9.84638095e-01 -1.47598639e-01 -3.35280091e-01
-5.04846871e-01 4.22235072e-01 4.99484211e-01 -3.50949526e-01
3.42930913e-01 1.02256966e+00 -1.48764408e+00 -2.18491793e-01
5.84122241e-01 7.27399886e-01 4.24913347e-01 4.65195805e-01
-1.91717386e-01 1.31863028e-01 -6.63285494e-01 -7.54438043e-02
1.91982433e-01 -5.44129491e-01 -6.80068433e-01 -8.07285085e-02
4.90237057e-01 5.90849519e-01 -9.80195850e-02 1.39291143e+00
2.56453991e-01 2.01148123e-01 9.94631529e-01 -1.41870689e+00
-4.64480191e-01 5.88136375e-01 -2.48520136e-01 5.53093553e-01
5.90666294e-01 -1.13074921e-01 9.56222236e-01 -5.04094303e-01
6.85487926e-01 1.32482159e+00 6.25361383e-01 -4.54106703e-02
-1.46226096e+00 -3.99933577e-01 -8.16754103e-02 4.24399942e-01
-1.67559159e+00 -7.87580106e-03 8.76478553e-01 -8.37131143e-01
7.69097626e-01 3.30843270e-01 8.60489130e-01 4.98670012e-01
4.30729449e-01 -1.09130576e-01 1.29441822e+00 -4.83901858e-01
4.39165026e-01 -2.43443362e-02 3.62650126e-01 6.15999937e-01
4.51765656e-01 -1.65752143e-01 -3.21020663e-01 2.54665595e-02
9.11017537e-01 -2.31724843e-01 -2.66636938e-01 -7.23328173e-01
-1.52555180e+00 6.18829072e-01 9.52306092e-01 1.27896476e+00
-3.88767868e-01 -1.45824850e-01 3.06429744e-01 4.47493702e-01
1.99753657e-01 7.68987954e-01 2.14220420e-01 2.59620249e-01
-7.92542338e-01 -1.21170558e-01 8.53322506e-01 6.24864340e-01
6.33500993e-01 -7.98402876e-02 7.93455169e-02 6.19169950e-01
2.27541015e-01 3.35393608e-01 4.49797511e-01 -4.12162721e-01
1.11226095e-02 8.90629470e-01 -5.19637287e-01 -1.79486644e+00
-1.03129411e+00 -5.14629841e-01 -1.10432124e+00 4.96472478e-01
6.54630482e-01 4.38751549e-01 -5.57018042e-01 1.55368066e+00
1.81316510e-01 -5.79888783e-02 -3.09059858e-01 9.95105624e-01
9.70056176e-01 1.29362255e-01 -6.64620474e-02 -6.15807325e-02
1.30189502e+00 -3.64695579e-01 -5.59715867e-01 1.45232812e-01
3.41854185e-01 -5.99553943e-01 8.24749947e-01 1.78366184e-01
-9.28116739e-01 -6.15653336e-01 -9.41848695e-01 3.97752017e-01
-9.23656940e-01 -1.38002768e-01 1.21685252e-01 6.61837995e-01
-1.48536766e+00 9.17528868e-01 -3.77158046e-01 -8.99902225e-01
7.20460489e-02 2.85265476e-01 -7.73318708e-01 2.37758905e-01
-1.00333786e+00 1.21545494e+00 4.59722757e-01 5.23561016e-02
-1.57505825e-01 -2.90752143e-01 -7.43144512e-01 2.15024382e-01
4.26385030e-02 -4.57679749e-01 3.86508912e-01 -9.32069540e-01
-1.15332437e+00 1.05421627e+00 2.84375131e-01 -2.67257214e-01
6.69599354e-01 6.26172125e-01 -5.85284531e-01 3.53515476e-01
6.25363365e-02 4.52119112e-01 7.98606813e-01 -1.07910383e+00
1.62772939e-01 -5.05121171e-01 -1.24064684e-02 7.63776246e-03
-1.75161466e-01 -4.29212339e-02 1.63220510e-01 -5.48879921e-01
5.63037455e-01 -8.23673725e-01 6.55966774e-02 3.52907896e-01
-4.24503744e-01 -2.94328153e-01 6.61348760e-01 -4.39526528e-01
9.34596777e-01 -1.99753618e+00 4.44848359e-01 9.09520984e-01
6.76046014e-01 1.81315631e-01 -2.23586544e-01 5.66168487e-01
-4.49062049e-01 6.89554587e-02 -3.34812462e-01 2.52177298e-01
3.36155072e-02 -1.84089299e-02 3.89630824e-01 7.93163657e-01
-6.75627515e-02 6.88094676e-01 -9.90846395e-01 -4.46600556e-01
4.92600918e-01 5.85205197e-01 -2.55405181e-03 -4.07424755e-02
2.67978102e-01 4.24772352e-01 -1.28399253e-01 1.89614594e-01
6.46321654e-01 -1.72638535e-01 3.03791344e-01 -3.92735332e-01
-1.83455423e-01 -2.08392024e-01 -1.21488595e+00 1.27608037e+00
-1.64347202e-01 8.86570513e-01 2.82316133e-02 -1.16538882e+00
1.31623423e+00 1.18605003e-01 3.18075985e-01 -7.55186915e-01
5.01233280e-01 7.22805262e-02 6.41647518e-01 -2.33584002e-01
1.57016560e-01 1.50479406e-01 2.79357135e-01 3.93153340e-01
1.46335661e-01 -3.16670239e-01 5.03974438e-01 6.83529302e-02
1.23466229e+00 -5.35671175e-01 7.57167220e-01 -9.79476511e-01
9.13851738e-01 -4.17855799e-01 -1.80355296e-01 4.31910992e-01
-3.50014985e-01 4.33780819e-01 8.77004504e-01 -2.59511918e-01
-9.18889284e-01 -1.36061895e+00 -2.96783805e-01 5.58173239e-01
2.73932636e-01 -2.87458837e-01 -8.25189292e-01 -2.37378508e-01
6.37304708e-02 4.06830132e-01 -8.61602128e-01 -1.99333236e-01
-3.07655275e-01 -1.47150218e-01 2.32446626e-01 3.46105434e-02
4.05416191e-01 -1.09021449e+00 -9.15507019e-01 -6.34169877e-02
2.65322268e-01 -1.01065993e+00 -2.38726661e-01 -1.46631882e-01
-8.42764497e-01 -1.32307804e+00 -6.88177228e-01 -6.96901381e-01
8.28500330e-01 2.79483378e-01 1.14949286e+00 3.53814572e-01
-5.41973352e-01 5.80066919e-01 -3.48809868e-01 1.04604736e-01
-5.43367445e-01 8.97044912e-02 8.79495293e-02 1.47029638e-01
-1.40586002e-02 -9.54623103e-01 -4.00602847e-01 6.16666198e-01
-9.20057356e-01 -2.01601699e-01 5.81874430e-01 4.74801987e-01
2.02668682e-01 2.57495027e-02 3.28903466e-01 -4.78070825e-01
1.09092486e+00 -5.35700202e-01 -5.87355018e-01 2.32807115e-01
-5.15734196e-01 1.79726645e-01 7.64754593e-01 -4.10293967e-01
-2.15382174e-01 -1.92305997e-01 2.28643566e-01 -1.23063944e-01
-3.03701162e-01 5.13435602e-01 -8.73010606e-02 -5.69137394e-01
7.58090198e-01 6.04643859e-02 3.08236599e-01 -2.85382450e-01
2.68031299e-01 2.29004204e-01 4.00644273e-01 -1.83017358e-01
6.98473930e-01 2.24645764e-01 5.64268887e-01 -1.20898604e+00
1.55979946e-01 -5.01475751e-01 -1.18778956e+00 -6.45203412e-01
7.09298015e-01 -8.60978663e-02 -9.09407794e-01 3.90462816e-01
-1.06272507e+00 -2.85982303e-02 -1.90249234e-01 5.34761429e-01
-3.93303633e-01 5.10999322e-01 -1.43816546e-01 -5.22632599e-01
-4.92773168e-02 -8.49209785e-01 6.12355411e-01 8.41583163e-02
-4.03494477e-01 -1.44922042e+00 5.24213575e-02 -3.77197266e-01
5.78992009e-01 5.67828000e-01 1.10685456e+00 -6.57071114e-01
-2.38838479e-01 -1.56845883e-01 -4.05227005e-01 4.16250266e-02
3.30239713e-01 5.97493686e-02 -5.94561219e-01 -2.25656807e-01
-1.99380010e-01 4.26062852e-01 5.22134662e-01 2.70908207e-01
6.68983996e-01 -6.88509364e-03 -4.34221894e-01 3.50763127e-02
1.48622537e+00 2.01695070e-01 4.15810555e-01 -2.03680177e-03
5.18879592e-01 9.22343194e-01 1.52614135e-02 2.23931044e-01
-1.34928122e-01 7.86749959e-01 5.20102322e-01 1.28620258e-02
-2.39489049e-01 2.40521416e-01 -1.82481110e-01 9.03191328e-01
-1.70516655e-01 -1.69554070e-01 -1.16313386e+00 5.14319479e-01
-1.49743342e+00 -9.64709282e-01 -4.89135742e-01 2.29462576e+00
1.67428151e-01 2.19693542e-01 3.29400390e-01 4.63687420e-01
1.20695949e+00 2.22739890e-01 -2.57293046e-01 -3.41463119e-01
-1.09077252e-01 2.37494022e-01 2.35204577e-01 4.52810049e-01
-6.14177883e-01 4.20699924e-01 6.66011953e+00 4.57071811e-01
-1.23275054e+00 6.03294224e-02 1.46606922e-01 3.66729796e-01
-5.63546680e-02 -2.44905457e-01 7.91521147e-02 3.90970379e-01
6.80952132e-01 -5.46500325e-01 4.51870024e-01 5.80763459e-01
4.32986096e-02 -2.67181456e-01 -1.10880625e+00 1.22416699e+00
2.35343516e-01 -9.14367795e-01 -9.11531225e-02 -3.53102572e-02
2.21934289e-01 -1.25441819e-01 -1.13986857e-01 -4.92517889e-01
-1.95572123e-01 -1.20573735e+00 6.96258962e-01 7.72320449e-01
5.82476079e-01 -4.75755423e-01 5.90635240e-01 6.43041581e-02
-1.59662271e+00 1.90693781e-01 -3.15709800e-01 -2.08658144e-01
-3.70409600e-02 7.12154031e-01 -9.52098906e-01 4.73609030e-01
4.44872528e-01 7.66975343e-01 -8.98161769e-01 1.45572495e+00
-3.93345319e-02 -4.06895876e-02 -3.69646937e-01 -4.52119380e-01
1.09680310e-01 -5.75641572e-01 8.39145243e-01 1.19886768e+00
3.65206003e-01 -3.10720176e-01 -1.72753528e-01 1.23863578e+00
1.58637121e-01 4.03413683e-01 -1.00356555e+00 -1.97729059e-02
4.48477268e-01 1.44541252e+00 -1.53344393e+00 -3.55108343e-02
-6.39190674e-02 8.37079525e-01 2.87744254e-01 1.40598848e-01
-5.21838605e-01 -4.77457821e-01 5.00985503e-01 3.60212862e-01
-1.08470589e-01 -6.56208932e-01 -1.22614399e-01 -5.89967489e-01
1.00164395e-02 -2.35027418e-01 1.40863374e-01 -9.02560115e-01
-1.36702371e+00 1.06215894e+00 2.97836185e-01 -1.11492658e+00
-2.11612880e-02 -8.82070780e-01 -5.63387096e-01 9.57808375e-01
-8.47055018e-01 -6.02157414e-01 -8.22000921e-01 7.16976702e-01
-6.45491257e-02 -1.00922488e-01 6.80922568e-01 6.56855032e-02
-1.65637583e-01 1.71828285e-01 -5.94491698e-02 1.94505766e-01
3.12249362e-01 -1.29255533e+00 4.16083634e-01 6.08884752e-01
3.24334174e-01 5.20239294e-01 7.47581184e-01 -4.75422800e-01
-9.31589901e-01 -6.07812345e-01 8.33938539e-01 -1.33976057e-01
8.04972053e-01 -3.89299273e-01 -8.88638377e-01 2.76920557e-01
2.97215551e-01 2.09970787e-01 2.82385767e-01 -2.09203660e-01
-3.82365435e-01 -1.28499866e-01 -1.29208088e+00 4.53001410e-01
1.26467121e+00 -5.90650380e-01 -6.92196667e-01 3.00376952e-01
1.42846152e-01 4.48526815e-02 -1.12518966e+00 1.01743452e-01
5.43173134e-01 -1.39343500e+00 9.28833008e-01 -9.63658318e-02
-1.43022090e-01 -4.21452820e-01 1.94060877e-01 -1.76842809e+00
-6.05852306e-01 -2.62286425e-01 4.77694511e-01 7.73072124e-01
1.18198857e-01 -1.17080975e+00 3.50450188e-01 9.46011245e-02
1.82638451e-01 -1.75813332e-01 -1.08574438e+00 -1.15127397e+00
-1.33763984e-01 -4.44978252e-02 4.42595035e-01 1.15378022e+00
2.39441842e-01 4.23972279e-01 4.66611773e-01 -1.80504411e-01
6.10857248e-01 -1.32660687e-01 3.11422795e-01 -1.95115256e+00
2.38449842e-01 -1.10045588e+00 -1.50717998e+00 -1.26440212e-01
1.94522649e-01 -1.10940850e+00 -3.58289480e-01 -1.55107498e+00
-1.64941117e-01 -3.49660158e-01 -1.90342113e-01 1.24837697e-01
4.99221206e-01 2.27558658e-01 3.78193468e-01 5.39412498e-02
-1.90111667e-01 2.41702378e-01 1.22194052e+00 -1.18726142e-01
2.56808586e-02 -2.98602819e-01 -5.44360057e-02 6.38773322e-01
8.65781307e-01 -3.61534446e-01 -4.03108567e-01 3.54916789e-02
2.06999928e-01 -2.58635730e-01 6.26150787e-01 -1.53665495e+00
3.47826123e-01 1.49589464e-01 1.96118221e-01 -2.59514064e-01
1.95149258e-01 -1.22014081e+00 5.89409947e-01 7.74661958e-01
-1.79586887e-01 6.36402667e-01 -7.73562305e-03 3.92786533e-01
-2.83045888e-01 -3.93294871e-01 9.84440804e-01 -1.10273421e-01
-8.07111025e-01 5.16145665e-04 -3.40165466e-01 -8.88514295e-02
1.25563145e+00 -4.18493807e-01 -4.05109704e-01 -2.84587383e-01
-1.02058566e+00 -2.77438104e-01 5.96344233e-01 3.42168778e-01
6.64303124e-01 -1.35618198e+00 -5.47866762e-01 1.42305776e-01
1.38981491e-01 -6.34287715e-01 -1.35901337e-02 1.04192519e+00
-9.81888056e-01 3.75723779e-01 -9.19916809e-01 -8.20609987e-01
-1.59666979e+00 7.38879323e-01 5.91950953e-01 -2.86154100e-03
-4.04025674e-01 3.75653297e-01 2.30826670e-03 -2.70082593e-01
-4.01646197e-02 -6.23635590e-01 -5.21480083e-01 2.71326691e-01
2.22524881e-01 5.68491399e-01 6.98594004e-02 -1.00624406e+00
-6.84507847e-01 8.88027370e-01 5.42308569e-01 -6.84226900e-02
1.15123677e+00 7.73300510e-03 -5.54670215e-01 7.72589028e-01
1.24788630e+00 -1.64570734e-01 -5.15674770e-01 -1.58609390e-01
2.41261795e-01 -4.80258912e-01 -1.90372691e-01 -5.09397864e-01
-1.10927951e+00 8.69685650e-01 8.63905430e-01 1.03318322e+00
1.02998877e+00 6.91240132e-02 -4.50579561e-02 2.82747120e-01
5.09585381e-01 -7.93699622e-01 1.50419354e-01 3.22964996e-01
1.33193862e+00 -8.70640576e-01 4.33517732e-02 -6.68756247e-01
-1.25300780e-01 1.41230893e+00 3.31363052e-01 -4.14289176e-01
9.06641304e-01 2.74827634e-03 -2.51677483e-01 -4.70381767e-01
3.61073725e-02 -3.90907198e-01 6.32030904e-01 8.18713188e-01
4.70944524e-01 2.04784974e-01 -7.12623537e-01 -2.43292436e-01
-3.79586697e-01 -4.86004949e-01 4.39226747e-01 5.97073197e-01
-5.09913921e-01 -8.07474375e-01 -5.32258213e-01 3.84982318e-01
1.64569527e-01 9.44463164e-03 -7.86100090e-01 1.09697628e+00
4.93886732e-02 8.59853387e-01 2.83710390e-01 -4.42064255e-01
5.59785903e-01 -2.00675383e-01 7.63801873e-01 -3.00263435e-01
-7.21964836e-01 -6.27518356e-01 -2.16757849e-01 -5.23904324e-01
-6.48555636e-01 -6.19942844e-01 -1.03479254e+00 -3.85212958e-01
-2.20429510e-01 1.54708639e-01 8.48981917e-01 7.21117795e-01
2.35916957e-01 6.10279024e-01 3.74633998e-01 -9.02463198e-01
3.35438550e-02 -9.81252551e-01 -9.45461571e-01 6.07306600e-01
2.46315412e-02 -9.86023843e-01 -4.40355480e-01 -2.71173745e-01] | [12.25837516784668, 3.414670944213867] |
0fb27f6a-b52a-47a7-8268-e5e8456c4ddb | covid-19-detection-and-analysis-from-lung-ct | 2209.10963 | null | https://arxiv.org/abs/2209.10963v2 | https://arxiv.org/pdf/2209.10963v2.pdf | COVID-19 Detection and Analysis From Lung CT Images using Novel Channel Boosted CNNs | In December 2019, the global pandemic COVID-19 in Wuhan, China, affected human life and the worldwide economy. Therefore, an efficient diagnostic system is required to control its spread. However, the automatic diagnostic system poses challenges with a limited amount of labeled data, minor contrast variation, and high structural similarity between infection and background. In this regard, a new two-phase deep convolutional neural network (CNN) based diagnostic system is proposed to detect minute irregularities and analyze COVID-19 infection. In the first phase, a novel SB-STM-BRNet CNN is developed, incorporating a new channel Squeezed and Boosted (SB) and dilated convolutional-based Split-Transform-Merge (STM) block to detect COVID-19 infected lung CT images. The new STM blocks performed multi-path region-smoothing and boundary operations, which helped to learn minor contrast variation and global COVID-19 specific patterns. Furthermore, the diverse boosted channels are achieved using the SB and Transfer Learning concepts in STM blocks to learn texture variation between COVID-19-specific and healthy images. In the second phase, COVID-19 infected images are provided to the novel COVID-CB-RESeg segmentation CNN to identify and analyze COVID-19 infectious regions. The proposed COVID-CB-RESeg methodically employed region-homogeneity and heterogeneity operations in each encoder-decoder block and boosted-decoder using auxiliary channels to simultaneously learn the low illumination and boundaries of the COVID-19 infected region. The proposed diagnostic system yields good performance in terms of accuracy: 98.21 %, F-score: 98.24%, Dice Similarity: 96.40 %, and IOU: 98.85 % for the COVID-19 infected region. The proposed diagnostic system would reduce the burden and strengthen the radiologist's decision for a fast and accurate COVID-19 diagnosis. | ['Saddam Hussain Khan'] | 2022-09-22 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 2.40197256e-01 -4.44976270e-01 -6.04365356e-02 -3.81404683e-02
-7.02194452e-01 -3.69507462e-01 1.55002281e-01 -6.20813668e-02
-4.14459944e-01 7.10407913e-01 -2.34711543e-01 -4.42680985e-01
7.49096125e-02 -8.10905457e-01 -4.71784681e-01 -1.11426520e+00
1.64184370e-04 4.49102223e-01 4.72772330e-01 1.17436327e-01
-2.09283695e-01 7.04746902e-01 -8.88354301e-01 3.41922015e-01
9.61174905e-01 1.00903702e+00 6.60848737e-01 8.82100821e-01
1.76069424e-01 4.75302458e-01 -4.54738170e-01 1.92821577e-01
1.95393592e-01 -5.66752434e-01 -5.48801959e-01 -2.28408366e-01
1.28860295e-01 -5.55433035e-01 -7.53104240e-02 8.40929627e-01
4.77144539e-01 -2.12356627e-01 1.05128634e+00 -8.21400762e-01
-5.04906535e-01 -6.02326505e-02 -9.71022546e-01 7.81917930e-01
-2.63938636e-01 4.18264955e-01 2.43318722e-01 -8.82524788e-01
7.98067868e-01 9.46277440e-01 9.74950075e-01 4.74130452e-01
-9.08649385e-01 -9.77533877e-01 -2.00025141e-01 2.73030788e-01
-1.44051039e+00 3.88093531e-01 3.79743487e-01 -8.27582121e-01
8.00270855e-01 3.86281550e-01 8.49052846e-01 8.72424006e-01
8.15189958e-01 3.00722718e-01 1.00116909e+00 6.15344681e-02
-2.26662293e-01 -1.38306037e-01 -7.54633695e-02 1.06651652e+00
5.50396144e-01 7.06176981e-02 4.02149290e-01 8.74376372e-02
1.09076643e+00 5.45398772e-01 -3.87359589e-01 5.07597513e-02
-1.27901888e+00 7.78904736e-01 7.80969858e-01 5.16483724e-01
-3.86660069e-01 -1.10801011e-01 6.17496848e-01 -1.80184275e-01
3.69599491e-01 3.49558592e-02 -3.75237018e-01 3.44100088e-01
-7.84329593e-01 -1.52861029e-01 7.01805502e-02 4.90530670e-01
4.94834125e-01 1.63760349e-01 -5.47282696e-01 6.76290512e-01
2.82114416e-01 1.28219211e+00 3.35516542e-01 -4.42430317e-01
2.50792742e-01 4.03754592e-01 -2.64949650e-01 -1.02735054e+00
-7.03027368e-01 -6.89946771e-01 -1.57333505e+00 -1.81554660e-01
9.54021960e-02 -2.56924719e-01 -1.42899871e+00 1.53228951e+00
4.93930846e-01 4.22471106e-01 -3.08138341e-01 9.98243511e-01
8.92511129e-01 8.46325099e-01 6.10489352e-03 -4.21148390e-01
1.72741544e+00 -8.19109261e-01 -5.94475746e-01 1.83428466e-01
5.77777982e-01 -8.45142782e-01 7.78310180e-01 -4.00699452e-02
-9.67276871e-01 -5.18928468e-01 -1.01943088e+00 2.17709810e-01
-1.60689116e-01 2.41113156e-01 1.79213002e-01 5.58513224e-01
-9.13200915e-01 -4.80071232e-02 -8.87167692e-01 -3.25270355e-01
6.05850697e-01 2.98922390e-01 -5.36769181e-02 -4.11418021e-01
-1.26036167e+00 7.91312397e-01 2.47895405e-01 2.91619629e-01
-1.24707437e+00 -8.80445719e-01 -7.33906209e-01 -5.91245629e-02
1.37025356e-01 -1.17888665e+00 7.96113133e-01 -7.16606379e-01
-9.26679492e-01 1.00378823e+00 -1.35914013e-01 -6.54005632e-02
3.91770601e-01 4.84818101e-01 -4.48702872e-01 4.04903561e-01
1.73785418e-01 5.07449150e-01 9.58139002e-01 -1.18912590e+00
-7.71202147e-01 -5.20019233e-01 -6.07757986e-01 3.28000896e-02
2.73294866e-01 2.19690636e-01 -4.09283876e-01 -9.28296924e-01
-5.01868986e-02 -9.47721899e-01 -8.70822668e-02 -5.25366776e-02
-2.61768252e-01 1.14299558e-01 1.22553039e+00 -1.12919092e+00
9.94525492e-01 -1.94802952e+00 -3.23667943e-01 3.35843116e-01
5.44866562e-01 6.46568894e-01 -6.94177076e-02 -4.86787945e-01
-3.03090084e-02 1.13853648e-01 -4.72946048e-01 1.46377236e-01
-6.42725825e-01 7.66485631e-02 1.95016637e-01 7.48281121e-01
3.34385008e-01 1.02064621e+00 -7.77252138e-01 -8.32000434e-01
2.97302037e-01 7.88133681e-01 -4.89362568e-01 3.49947155e-01
1.03803594e-02 9.46065843e-01 -4.12590116e-01 8.76981258e-01
1.17025852e+00 -4.54315722e-01 -7.59376362e-02 -2.39405558e-01
-1.24005927e-02 -4.13631409e-01 -7.63461530e-01 1.10144842e+00
-6.21033609e-01 5.84963679e-01 4.67373103e-01 -9.52555239e-01
7.61458278e-01 3.98464411e-01 8.56992364e-01 -8.28643978e-01
3.01121712e-01 2.91192621e-01 1.15701400e-01 -8.08902085e-01
1.53382167e-01 -4.23525512e-01 1.17562301e-01 3.62847954e-01
-4.11773354e-01 -1.17542267e-01 -1.55191228e-01 -9.94437709e-02
8.18567812e-01 -5.38349688e-01 1.35953382e-01 -3.28320622e-01
8.46208870e-01 1.08039245e-01 7.62174487e-01 5.81575632e-01
-3.34195554e-01 7.81726897e-01 6.43989742e-02 -5.01942992e-01
-9.61582780e-01 -1.37087727e+00 -4.08473521e-01 4.95870143e-01
2.50628233e-01 5.14979243e-01 -7.04034090e-01 -7.03448892e-01
-4.04110074e-01 1.32968709e-01 -5.51309884e-01 5.31910639e-03
-9.64920342e-01 -8.64536226e-01 5.60439527e-01 5.93964100e-01
7.79263854e-01 -1.10469890e+00 -5.74696124e-01 7.62042627e-02
-5.08408427e-01 -8.32989275e-01 -7.33336747e-01 5.05442023e-02
-6.83445871e-01 -1.15667367e+00 -1.26005900e+00 -1.17995822e+00
7.87068486e-01 3.50312978e-01 9.13622022e-01 2.69206852e-01
-8.43536377e-01 -1.65639445e-01 5.87264374e-02 -3.89069796e-01
-3.10830116e-01 -3.76653969e-02 -2.90110886e-01 -2.82355487e-01
-2.56231390e-02 -7.23095313e-02 -1.01203465e+00 4.94461089e-01
-7.63452888e-01 1.59638301e-01 6.76449418e-01 1.08915842e+00
8.09791148e-01 -7.70129859e-02 5.08515179e-01 -5.92400670e-01
3.76562297e-01 -5.93068242e-01 -4.78571892e-01 2.73838073e-01
-5.72269261e-01 -5.80071151e-01 6.71173751e-01 -2.86588192e-01
-1.22120774e+00 -1.83663383e-01 -1.85579479e-01 -5.48103511e-01
-4.52510007e-02 1.75653785e-01 1.62870675e-01 9.52303559e-02
4.12613809e-01 4.04522032e-01 3.72962840e-02 4.63598631e-02
2.06303615e-02 6.83087289e-01 6.52560830e-01 -1.99548900e-01
6.25150561e-01 7.37879157e-01 -3.86809781e-02 -7.28776753e-01
-5.35942554e-01 -6.29087865e-01 -4.31111455e-01 -2.92998940e-01
1.71428907e+00 -9.69994009e-01 -7.00949609e-01 7.98588097e-01
-1.26905954e+00 -2.37791866e-01 -1.02648884e-01 7.41499603e-01
-1.63036510e-01 3.44956934e-01 -9.65272069e-01 -1.96264759e-01
-8.98044884e-01 -1.64865589e+00 9.33806956e-01 4.31005508e-01
1.32444084e-01 -9.03999090e-01 9.02959779e-02 4.29216921e-01
7.02302694e-01 3.22629720e-01 1.18668926e+00 -4.80891198e-01
-6.73927248e-01 7.46426806e-02 -9.45278525e-01 4.10098404e-01
4.06081051e-01 -4.58425730e-02 -7.55213022e-01 -4.55475718e-01
2.03251634e-02 -8.30723252e-03 9.38726127e-01 1.02000809e+00
1.33463275e+00 -3.49996649e-02 -4.78668690e-01 1.11974180e+00
1.30666721e+00 8.62745762e-01 3.79547983e-01 -6.55110367e-03
1.01553881e+00 1.48619831e-01 4.42679644e-01 3.50737751e-01
1.31033674e-01 3.52971971e-01 3.46323043e-01 -8.18372250e-01
-3.89680475e-01 2.07334816e-01 -8.20625275e-02 9.71166372e-01
-2.32353583e-01 -1.93669230e-01 -1.04812574e+00 6.45470679e-01
-1.21420228e+00 -8.73092353e-01 -3.41049582e-01 1.70395899e+00
5.44102132e-01 -2.30049133e-01 -8.95423368e-02 -4.26834434e-01
1.00738764e+00 -5.65400161e-02 -6.73021317e-01 -3.86406064e-01
-2.69621816e-02 3.21224332e-01 4.37395424e-01 3.83489460e-01
-1.10392237e+00 3.37092161e-01 5.61032963e+00 9.37740624e-01
-1.63714910e+00 4.27064568e-01 1.02345824e+00 2.05632359e-01
-2.50413060e-01 -6.86178088e-01 -6.15912676e-01 5.96303403e-01
3.63349646e-01 2.42684945e-01 4.95484062e-02 4.76304293e-01
4.04997140e-01 4.41561826e-02 -4.37323272e-01 9.68263030e-01
1.04462773e-01 -1.54975057e+00 -2.23337352e-01 1.34769320e-01
1.04478693e+00 1.78513929e-01 2.15599984e-01 1.87853619e-01
-5.87020256e-02 -1.16146946e+00 1.11998722e-01 4.47861582e-01
1.47274303e+00 -8.18980336e-01 1.21115589e+00 6.96520731e-02
-1.58404040e+00 1.71017423e-01 -1.98799685e-01 6.80622399e-01
8.87589976e-02 5.74580610e-01 -1.24642491e+00 5.40713489e-01
8.48965645e-01 5.63768744e-01 -2.31420845e-01 7.44278967e-01
2.95046475e-02 5.62002182e-01 -1.54474735e-01 4.24538776e-02
3.34716320e-01 -2.47826099e-01 4.76829618e-01 1.78365588e+00
4.28719521e-01 1.10890515e-01 1.99468791e-01 9.36269164e-01
5.79516701e-02 1.02970615e-01 -4.66206342e-01 4.74440992e-01
1.29054666e-01 1.16464865e+00 -1.03855550e+00 -6.39587283e-01
-2.49040648e-01 7.26675153e-01 -3.58932734e-01 3.25395912e-01
-1.28513575e+00 -1.91301361e-01 4.90134716e-01 2.15639651e-01
4.91025537e-01 1.57057747e-01 -4.48992223e-01 -9.62628245e-01
-3.70473653e-01 -8.18984032e-01 6.53889894e-01 -6.37178957e-01
-1.23877788e+00 6.49215758e-01 3.63047305e-03 -1.15435064e+00
-1.16987094e-01 -4.09740776e-01 -8.05337667e-01 1.04079580e+00
-1.52010918e+00 -1.16483438e+00 -6.98643327e-01 8.31753075e-01
5.81154406e-01 -8.92374218e-02 6.27196968e-01 3.99473608e-01
-7.45489061e-01 5.35777271e-01 3.82522732e-01 1.23076983e-01
6.89969540e-01 -9.32625353e-01 -9.17672589e-02 8.70555758e-01
-8.34788203e-01 5.03901303e-01 -4.78469431e-02 -9.22303975e-01
-8.41087043e-01 -1.45414376e+00 5.52592456e-01 1.66151613e-01
2.91645139e-01 -2.44239848e-02 -9.17816341e-01 4.02776927e-01
1.16344079e-01 3.10928732e-01 3.25717449e-01 -7.35192657e-01
-4.37084064e-02 -1.28134117e-01 -1.57900655e+00 3.69703501e-01
5.60490251e-01 -4.47955757e-01 -2.74670541e-01 3.22543800e-01
8.31017554e-01 -4.72006291e-01 -8.02074969e-01 6.17551267e-01
5.23431420e-01 -7.51269996e-01 1.26223886e+00 -1.74633265e-01
3.70510191e-01 -3.81666124e-01 9.45639089e-02 -9.95029032e-01
-4.59022045e-01 -2.84296926e-02 3.82837474e-01 5.83592117e-01
2.58641362e-01 -7.38900423e-01 7.26274848e-01 -1.82084933e-01
-3.81632864e-01 -9.96941566e-01 -9.09815073e-01 -4.82992828e-01
9.34914201e-02 -1.57989994e-01 3.49242061e-01 1.00260687e+00
-8.23175311e-01 -5.13590872e-02 -6.53371438e-02 1.94410056e-01
4.90022510e-01 1.28289729e-01 2.78682828e-01 -9.18400884e-01
-5.92178255e-02 -4.21902061e-01 -6.31088689e-02 -7.86368132e-01
-3.41417015e-01 -9.77916300e-01 7.00102327e-03 -1.69830751e+00
6.16057098e-01 -5.49194157e-01 -5.88007867e-01 2.70169318e-01
-2.96182722e-01 5.34612656e-01 1.78929400e-02 2.19556093e-01
-1.12363756e-01 1.32846236e-01 1.96509862e+00 -5.38556576e-01
-8.04449469e-02 5.68651743e-02 -1.01395063e-01 6.85011864e-01
8.75355065e-01 -5.04670620e-01 -4.28895772e-01 -1.59702897e-01
-1.70999601e-01 1.97892413e-01 5.04351735e-01 -7.86513805e-01
-4.44936082e-02 -2.73211837e-01 7.50396311e-01 -1.05236316e+00
1.25901863e-01 -7.61757135e-01 1.73749268e-01 1.24868250e+00
2.52280235e-01 2.27117091e-01 2.56304711e-01 3.44494820e-01
-1.09483033e-01 1.07104957e-01 1.24933898e+00 -8.78706947e-02
-5.44383705e-01 5.38734376e-01 -7.47625768e-01 2.42073178e-01
1.30261958e+00 -3.11304241e-01 -3.72410446e-01 6.93743080e-02
-4.31657404e-01 3.08222137e-03 2.26510987e-01 -4.98995110e-02
1.00151074e+00 -9.70321119e-01 -9.03866947e-01 5.91962636e-01
-1.85928755e-02 4.35511559e-01 9.83276844e-01 1.40321958e+00
-1.24028313e+00 4.71423268e-01 -3.78683120e-01 -1.13121951e+00
-1.47718525e+00 5.38272142e-01 6.57294750e-01 -7.32087851e-01
-6.75995946e-01 9.21696782e-01 8.34355772e-01 -2.30223060e-01
-9.74531844e-02 -5.87748528e-01 6.00902177e-03 -8.91827345e-02
4.07831103e-01 2.39619106e-01 1.05844736e-01 -7.21182883e-01
-5.70206404e-01 8.70277166e-01 -1.46304488e-01 4.20190394e-01
9.93654788e-01 -1.65393874e-01 -1.09046623e-01 -1.13894992e-01
1.38792121e+00 -1.92666903e-01 -1.02757895e+00 -8.00299793e-02
-6.81683362e-01 -2.87462324e-01 -1.54916039e-02 -8.88747096e-01
-1.31085205e+00 9.29328620e-01 1.24517965e+00 -3.46989572e-01
1.38198376e+00 1.93729885e-02 1.21925652e+00 -2.78264284e-01
-1.12570181e-01 -5.68453372e-01 1.35888338e-01 5.34211397e-01
5.59858203e-01 -1.28386497e+00 -9.28736255e-02 -3.08500320e-01
-5.20576298e-01 8.71563792e-01 6.57362401e-01 -1.02053781e-03
7.49278426e-01 5.18223226e-01 2.39646554e-01 -4.54866111e-01
-3.93155247e-01 1.41843989e-01 4.17781472e-01 8.78988683e-01
2.81209230e-01 4.23773706e-01 -1.97103426e-01 4.09139425e-01
3.20771724e-01 -1.65568590e-01 1.30094573e-01 6.34001017e-01
-3.34686875e-01 -3.59277934e-01 -7.15246797e-01 6.01815999e-01
-5.54030895e-01 -8.49088877e-02 2.07560524e-01 1.00217748e+00
7.49567866e-01 6.77175462e-01 3.27563375e-01 -3.21734160e-01
5.52268848e-02 -4.21991795e-01 3.05324048e-01 -4.44518685e-01
-7.38280296e-01 4.16079253e-01 -3.41202736e-01 -2.68721163e-01
-2.72424310e-01 -4.04525131e-01 -1.52934003e+00 -2.69891649e-01
-1.41328812e-01 -2.08366007e-01 6.76121891e-01 8.23386967e-01
1.35152683e-01 9.13807034e-01 7.48998404e-01 -4.78972137e-01
-1.46519423e-01 -7.19105065e-01 -5.32843292e-01 1.75913855e-01
5.46680510e-01 -4.13552076e-01 -1.09002978e-01 1.15663335e-01] | [15.513184547424316, -1.7724653482437134] |
ddef22ec-369a-4a46-98d9-6bd3d5457e73 | cocolot-combining-complementary-trackers-in | 2205.04261 | null | https://arxiv.org/abs/2205.04261v1 | https://arxiv.org/pdf/2205.04261v1.pdf | CoCoLoT: Combining Complementary Trackers in Long-Term Visual Tracking | How to combine the complementary capabilities of an ensemble of different algorithms has been of central interest in visual object tracking. A significant progress on such a problem has been achieved, but considering short-term tracking scenarios. Instead, long-term tracking settings have been substantially ignored by the solutions. In this paper, we explicitly consider long-term tracking scenarios and provide a framework, named CoCoLoT, that combines the characteristics of complementary visual trackers to achieve enhanced long-term tracking performance. CoCoLoT perceives whether the trackers are following the target object through an online learned deep verification model, and accordingly activates a decision policy which selects the best performing tracker as well as it corrects the performance of the failing one. The proposed methodology is evaluated extensively and the comparison with several other solutions reveals that it competes favourably with the state-of-the-art on the most popular long-term visual tracking benchmarks. | ['Christian Micheloni', 'Matteo Dunnhofer'] | 2022-05-09 | null | null | null | null | ['visual-tracking', 'visual-object-tracking'] | ['computer-vision', 'computer-vision'] | [-2.80733198e-01 -3.03135812e-01 -1.77524030e-01 4.21667472e-02
-3.61709774e-01 -9.06135917e-01 8.55142772e-01 -1.30533176e-02
-5.18791258e-01 5.36700487e-01 -3.29204291e-01 -2.49574706e-01
-8.29543769e-02 -1.29280254e-01 -8.11184466e-01 -8.40389669e-01
-1.24786317e-01 4.63486761e-01 9.37921405e-01 -2.27485085e-03
1.21046744e-01 5.75528324e-01 -1.79869890e+00 -2.62727402e-02
5.27721822e-01 1.05601168e+00 1.58668742e-01 8.09049308e-01
2.02183966e-02 6.39351368e-01 -7.39950657e-01 -4.53501523e-01
5.02401412e-01 -1.80766851e-01 -3.21290523e-01 9.50995646e-03
1.18289375e+00 -1.57183111e-01 -1.47636950e-01 1.07625711e+00
5.55237412e-01 -3.48621428e-01 2.79427648e-01 -1.55049813e+00
-3.62564832e-01 1.77117050e-01 -5.74315846e-01 5.70140183e-01
2.15075806e-01 7.33919859e-01 7.06979811e-01 -7.09215164e-01
6.41185880e-01 1.07845545e+00 8.10218871e-01 5.73240161e-01
-1.15524197e+00 -7.78398335e-01 3.85875791e-01 8.70109908e-03
-1.20181096e+00 -5.24613798e-01 4.08948243e-01 -7.75469482e-01
7.57285953e-01 6.28369227e-02 7.52954662e-01 1.16428828e+00
3.93680334e-01 7.67083228e-01 1.07681894e+00 -2.69113451e-01
-5.96409962e-02 2.01086178e-01 2.44054884e-01 6.92513764e-01
5.40488482e-01 8.16970766e-01 -3.75220299e-01 -8.26383606e-02
4.25006598e-01 -2.76069343e-01 -1.36458769e-01 -8.61219645e-01
-1.36003971e+00 3.44376236e-01 5.97851455e-01 4.80743855e-01
-3.09725553e-01 4.79440480e-01 4.02526528e-01 3.01764905e-01
1.79431975e-01 2.42971599e-01 -3.04038972e-01 -6.27650321e-02
-1.21930921e+00 4.58154798e-01 4.21073467e-01 7.66589165e-01
2.78480083e-01 2.70710349e-01 -7.83756554e-01 7.98465535e-02
6.29098296e-01 7.14339137e-01 4.91298065e-02 -6.73034906e-01
1.48817331e-01 5.34262061e-01 5.49589694e-01 -7.56371498e-01
-2.13698551e-01 -9.81543303e-01 -2.22425848e-01 9.96171355e-01
7.83467054e-01 -8.27927664e-02 -9.13119256e-01 1.74986541e+00
5.72625339e-01 4.85176414e-01 -1.21274114e-01 1.07624936e+00
6.23664021e-01 2.31282413e-01 2.93365151e-01 -1.28135741e-01
1.29604065e+00 -1.04375625e+00 -8.51755261e-01 -2.10149646e-01
8.85261223e-02 -8.72177839e-01 3.53986174e-01 2.81761408e-01
-9.16404784e-01 -1.09708929e+00 -1.23555017e+00 4.78754908e-01
-5.13035595e-01 3.60340059e-01 3.96039903e-01 8.38129461e-01
-1.38118649e+00 5.06638765e-01 -9.23157096e-01 -5.66979647e-01
2.99341112e-01 4.59579289e-01 -5.62293530e-02 4.20248330e-01
-8.63074899e-01 1.10880351e+00 3.19496542e-01 2.37095624e-01
-1.25631487e+00 -5.65812945e-01 -4.08654362e-01 -1.46426082e-01
7.75767565e-01 -7.49667764e-01 1.35466933e+00 -8.08241785e-01
-1.30056357e+00 8.64250064e-01 -1.65765882e-02 -7.36276686e-01
1.15415144e+00 -3.84323955e-01 -3.46367598e-01 -3.54795069e-01
3.13411430e-02 8.01318884e-01 1.18157709e+00 -1.29548538e+00
-9.16747391e-01 -2.99169928e-01 8.16337392e-02 -1.02510452e-01
-1.76306069e-01 2.60994136e-01 -7.66379178e-01 -6.06605053e-01
-3.92719507e-01 -1.11063194e+00 6.15912378e-02 5.63125730e-01
-1.44490615e-01 -4.67249244e-01 1.25477481e+00 -1.56155571e-01
1.21106374e+00 -2.08078885e+00 1.40015721e-01 -7.24478513e-02
3.44812870e-01 8.00243139e-01 -2.83688470e-03 2.04690218e-01
4.01605554e-02 -2.65536278e-01 3.56179655e-01 -5.11342049e-01
9.40556303e-02 4.74681938e-03 -3.68787825e-01 7.18798399e-01
-2.52293004e-03 1.08262491e+00 -1.01571691e+00 -4.60471243e-01
5.26318491e-01 4.57460940e-01 1.21178724e-01 2.78879195e-01
-4.29108322e-01 5.53473175e-01 -3.11314464e-01 6.29720390e-01
7.22574353e-01 -5.21246672e-01 6.52034357e-02 -2.53345668e-01
-6.10671997e-01 -3.29007804e-01 -1.16363621e+00 1.40605569e+00
1.37316376e-01 7.31002212e-01 4.76642810e-02 -3.10597271e-01
7.35381424e-01 3.71123105e-01 6.50320411e-01 -5.62789738e-01
2.23735958e-01 1.90553412e-01 1.74511880e-01 -3.08195442e-01
5.56493819e-01 3.67951483e-01 3.44321370e-01 1.16817161e-01
2.73374766e-02 6.96590424e-01 4.06649202e-01 2.01228440e-01
1.03933370e+00 7.01824307e-01 1.21655047e-01 -1.36118591e-01
7.41351128e-01 2.25011677e-01 4.91771430e-01 1.11498630e+00
-8.09432387e-01 2.53874302e-01 3.74917686e-02 -3.65413576e-01
-7.28230178e-01 -1.02826643e+00 -1.47841973e-02 1.06340039e+00
4.90416050e-01 -4.16104287e-01 -4.45808977e-01 -1.03453088e+00
4.93360996e-01 2.11323500e-01 -7.77052283e-01 3.97777781e-02
-6.40512764e-01 -2.99950093e-01 5.84534943e-01 5.30854285e-01
3.22055876e-01 -1.06033289e+00 -1.06708956e+00 2.61501372e-01
2.53981531e-01 -1.23215497e+00 -4.21072006e-01 7.05075562e-02
-6.28573596e-01 -1.28136241e+00 -8.36403012e-01 -3.11422467e-01
2.33001173e-01 4.74686742e-01 1.07149398e+00 2.31239527e-01
-9.09413695e-02 5.27832806e-01 -3.38921577e-01 -4.31707263e-01
-5.20660043e-01 -1.02354832e-01 1.88378006e-01 1.79605737e-01
1.14240184e-01 1.18076734e-01 -6.15455449e-01 6.56537831e-01
-4.41642195e-01 -3.13034564e-01 5.40213406e-01 5.11492431e-01
3.18112582e-01 -2.79819161e-01 3.30017954e-01 -5.28308392e-01
9.40937996e-02 -9.65107083e-02 -1.40104985e+00 4.98882025e-01
-8.08195710e-01 1.99722573e-02 3.66582334e-01 -7.31436968e-01
-7.48176396e-01 2.50171304e-01 5.30102141e-02 -9.38741982e-01
6.14042170e-02 -8.80826786e-02 2.12853067e-02 -7.20939696e-01
5.29007673e-01 5.99268377e-02 3.80233265e-02 -5.40860355e-01
1.14972338e-01 2.32212782e-01 7.61804521e-01 -2.93433577e-01
1.18729281e+00 5.64470887e-01 -7.25611448e-02 -4.19997513e-01
-7.74388313e-01 -6.56965256e-01 -5.35436809e-01 -9.62619245e-01
8.89214575e-01 -7.29169369e-01 -1.07564032e+00 6.47772551e-01
-1.16176105e+00 -1.32232666e-01 -2.44927034e-01 3.93291950e-01
-2.70271868e-01 2.45392174e-01 -1.33594826e-01 -9.65456605e-01
-1.87289178e-01 -1.33856189e+00 1.19604361e+00 4.49290246e-01
1.68459892e-01 -8.02944541e-01 4.54021931e-01 -1.21066710e-02
6.30368412e-01 3.86229575e-01 -1.93456449e-02 -4.55680072e-01
-1.11189497e+00 -2.09746614e-01 -1.85392961e-01 3.17506753e-02
-1.59278452e-01 3.02674949e-01 -1.10039282e+00 -8.35314691e-01
-5.17722130e-01 -7.63041154e-02 8.78542244e-01 4.79900420e-01
6.36273324e-01 2.84941345e-01 -8.87181401e-01 4.99560148e-01
1.46595132e+00 1.27051756e-01 3.20172250e-01 5.44133425e-01
5.54094970e-01 2.14551732e-01 9.81839299e-01 1.35048628e-01
1.35040909e-01 1.27395666e+00 8.59095871e-01 5.58122806e-03
-5.53665280e-01 -1.30243316e-01 6.08415484e-01 1.55541161e-02
9.49518103e-03 -3.44632953e-01 -7.53905892e-01 5.05977094e-01
-2.16315842e+00 -1.01269996e+00 -3.31672043e-01 2.23918796e+00
1.07228935e-01 4.73103195e-01 4.00876850e-01 -7.27404132e-02
8.09115589e-01 3.16659123e-01 -6.64127171e-01 2.30010778e-01
-2.08792575e-02 -1.41082212e-01 6.87436223e-01 1.16492830e-01
-1.44328105e+00 9.84207571e-01 6.76873922e+00 6.51670814e-01
-1.25072289e+00 3.05480212e-01 -2.33015880e-01 1.54289231e-01
5.24332643e-01 1.20323397e-01 -1.22223759e+00 5.63423634e-01
8.71256709e-01 1.76441185e-02 -7.82960504e-02 8.07191372e-01
1.00699358e-01 -2.01799329e-02 -1.16316330e+00 8.20797145e-01
2.22421251e-02 -1.20533442e+00 -3.96505088e-01 2.28670090e-01
5.13435125e-01 1.88409179e-01 3.24545920e-01 4.55707461e-01
3.28435540e-01 -5.63766301e-01 1.18326163e+00 6.51738644e-01
4.12943423e-01 -1.42211959e-01 6.74168229e-01 2.76861042e-01
-1.79364824e+00 -1.36099339e-01 -6.81071281e-02 3.48470777e-01
2.60056496e-01 1.95306554e-01 -7.49960899e-01 8.70725632e-01
8.50979805e-01 7.57171392e-01 -1.29989958e+00 1.79446924e+00
-1.72699302e-01 3.42860013e-01 -2.51929283e-01 1.44311681e-01
3.30523342e-01 2.61084884e-01 1.05898964e+00 1.04571497e+00
1.01128019e-01 -5.88466763e-01 5.46319604e-01 6.55460656e-01
2.73960173e-01 -2.97268093e-01 -5.44004261e-01 6.57074004e-02
1.84182510e-01 1.30554676e+00 -7.55320966e-01 -5.54621935e-01
-4.55155373e-01 6.26228511e-01 3.59251618e-01 1.17360458e-01
-1.27899158e+00 3.88329417e-01 6.61683619e-01 1.00976788e-01
8.12116861e-01 -8.24532285e-02 2.04219937e-01 -9.76751864e-01
1.39845252e-01 -7.70792842e-01 5.71740448e-01 -8.22427690e-01
-1.08825183e+00 7.65649915e-01 -8.99711326e-02 -1.60651934e+00
-1.63452566e-01 -6.34953916e-01 -4.10908908e-01 7.32983887e-01
-1.52746010e+00 -1.39538276e+00 -4.54078883e-01 5.23120105e-01
3.85840863e-01 -1.84063271e-01 2.45767981e-01 5.18363178e-01
-4.28401381e-01 6.69761360e-01 -1.00844368e-01 1.43952034e-02
9.44053769e-01 -1.23922873e+00 4.12756681e-01 1.18934190e+00
1.69460848e-01 4.25756097e-01 1.11010730e+00 -8.58238101e-01
-1.42439640e+00 -1.19845009e+00 5.87459564e-01 -7.28381991e-01
6.68536067e-01 -1.72981352e-01 -9.42369223e-01 5.33941150e-01
4.94495243e-01 4.72822458e-01 1.59502234e-02 -1.59916669e-01
-3.87625486e-01 -3.37566078e-01 -8.25466931e-01 2.55174309e-01
1.03439653e+00 -8.22388753e-02 -5.36406636e-01 1.97739989e-01
3.66061807e-01 -5.40871680e-01 -4.78654027e-01 6.44120693e-01
8.74080598e-01 -1.11348128e+00 1.03672671e+00 -4.84251559e-01
-5.27045786e-01 -1.02380359e+00 3.29303116e-01 -8.33110094e-01
-3.36134404e-01 -5.21453023e-01 -6.53962016e-01 1.11981535e+00
1.47795668e-02 -5.50173640e-01 8.40274811e-01 1.48568358e-02
8.02078694e-02 -4.18598652e-01 -9.05721068e-01 -1.23582351e+00
-3.27931046e-01 -4.73118350e-02 2.52619982e-01 5.93004107e-01
-8.05379152e-01 4.45779376e-02 -6.36632621e-01 3.78425926e-01
1.00346100e+00 3.57033670e-01 1.18954217e+00 -1.44038498e+00
-3.76624703e-01 -5.84916413e-01 -6.33301258e-01 -1.16458488e+00
2.75325570e-02 -7.78684020e-01 2.28608161e-01 -1.14154947e+00
1.34393081e-01 -3.48105103e-01 -5.28355300e-01 4.47067350e-01
-3.33752483e-01 2.65500367e-01 7.13086605e-01 2.34345719e-01
-1.05633843e+00 3.33013982e-01 1.07521904e+00 -1.97722316e-01
-1.26097342e-02 2.97188908e-01 -3.34350288e-01 3.65312636e-01
2.97906697e-01 -6.51391327e-01 3.82571109e-02 -3.04798841e-01
-3.87362503e-02 1.21443152e-01 8.24456871e-01 -1.53027427e+00
7.44206846e-01 2.22750902e-01 4.42811012e-01 -1.03723931e+00
2.14504257e-01 -1.21908319e+00 5.53900063e-01 7.73762405e-01
-1.73449919e-01 4.77413207e-01 3.45553339e-01 8.44897985e-01
-8.47834498e-02 -4.18758398e-04 1.00331378e+00 1.63021505e-01
-9.31486070e-01 3.97409558e-01 -9.70076919e-02 -2.19410524e-01
1.43776166e+00 -4.54486251e-01 -4.27623779e-01 1.44570470e-01
-7.71086931e-01 5.05976915e-01 5.48716009e-01 8.79604340e-01
3.22967470e-01 -1.27465034e+00 -5.82739532e-01 -1.05417222e-02
4.07697976e-01 -7.74739683e-01 -5.16041070e-02 1.03125381e+00
-7.08142892e-02 5.94280481e-01 -2.54721493e-01 -1.20645022e+00
-1.70650864e+00 9.71696377e-01 5.94634295e-01 -4.70787555e-01
-7.50185728e-01 4.67648327e-01 5.57185616e-03 4.49380167e-02
6.05293334e-01 -1.03404917e-01 -3.85527462e-01 1.10634072e-02
5.56973279e-01 2.34574437e-01 5.52300103e-02 -9.99190032e-01
-6.72370851e-01 6.42651439e-01 -8.78983811e-02 1.73788108e-02
9.18594599e-01 -9.25032720e-02 3.22052091e-01 3.33537161e-01
6.10515177e-01 -3.93451601e-02 -1.62880862e+00 -2.06717893e-01
1.68840289e-01 -7.47994900e-01 -6.14235029e-02 -9.35764313e-01
-1.29197013e+00 5.81017792e-01 1.20038223e+00 2.97724664e-01
6.90329492e-01 -1.89216793e-01 4.04204071e-01 7.26396218e-02
6.10704958e-01 -6.11424148e-01 9.89153087e-02 4.50770706e-01
5.12136936e-01 -1.40145051e+00 -4.05178964e-03 -6.93786964e-02
-3.49333942e-01 9.47335303e-01 9.32309866e-01 -3.30208123e-01
4.96669710e-01 3.51848215e-01 2.19296142e-01 -2.81551808e-01
-8.26330066e-01 -6.91883206e-01 4.43243742e-01 4.96037006e-01
1.15739338e-01 -3.13502043e-01 -1.75218508e-01 -2.10537747e-01
4.73277867e-01 9.13111120e-02 -7.42782876e-02 9.51528370e-01
-3.24561834e-01 -1.09658849e+00 -8.27909887e-01 2.14760359e-02
-3.59531313e-01 4.65776592e-01 -4.89280492e-01 1.05593586e+00
3.54624480e-01 8.50020707e-01 -2.67264128e-01 -2.45046541e-01
3.55985522e-01 -1.06603883e-01 5.84842145e-01 -1.89668804e-01
-1.28572059e+00 2.81002074e-01 -2.54545510e-01 -7.55600631e-01
-9.12573218e-01 -7.56600797e-01 -6.75734401e-01 -6.23185374e-02
-5.10513365e-01 1.98689494e-02 7.02329457e-01 8.57735276e-01
2.83675849e-01 8.52171302e-01 2.76088417e-01 -9.86124992e-01
-6.96984529e-01 -5.99347413e-01 -2.13111397e-02 2.63585240e-01
7.42163599e-01 -1.35143042e+00 -2.23336056e-01 -2.70682216e-01] | [6.335813045501709, -2.073403835296631] |
a7183276-f986-403d-af9c-8f3ec30f60d5 | general-purpose-tagging-of-freesound-audio | 1807.09902 | null | http://arxiv.org/abs/1807.09902v3 | http://arxiv.org/pdf/1807.09902v3.pdf | General-purpose Tagging of Freesound Audio with AudioSet Labels: Task Description, Dataset, and Baseline | This paper describes Task 2 of the DCASE 2018 Challenge, titled
"General-purpose audio tagging of Freesound content with AudioSet labels". This
task was hosted on the Kaggle platform as "Freesound General-Purpose Audio
Tagging Challenge". The goal of the task is to build an audio tagging system
that can recognize the category of an audio clip from a subset of 41 diverse
categories drawn from the AudioSet Ontology. We present the task, the dataset
prepared for the competition, and a baseline system. | ['Xavier Serra', 'Xavier Favory', 'Manoj Plakal', 'Eduardo Fonseca', 'Daniel P. W. Ellis', 'Jordi Pons', 'Frederic Font'] | 2018-07-26 | null | null | null | null | ['audio-tagging'] | ['audio'] | [ 5.42292669e-02 1.43415317e-01 -5.29647507e-02 -3.41168225e-01
-1.75180447e+00 -9.12078559e-01 1.23239145e-01 1.38657046e-02
-2.19386742e-01 4.48287129e-01 8.99112821e-01 4.54962760e-01
2.22899780e-01 -6.55989796e-02 -6.35817230e-01 -3.32735121e-01
-4.42868054e-01 2.56242663e-01 3.22402537e-01 -5.65896370e-02
-2.41262406e-01 -2.43666187e-01 -1.88578594e+00 9.39792216e-01
-2.00278416e-01 1.58925879e+00 -7.79711157e-02 1.09673810e+00
1.11522980e-01 7.04535961e-01 -8.06738675e-01 -1.10084742e-01
-1.83497682e-01 -2.01933786e-01 -1.35856187e+00 -4.36639905e-01
8.58119249e-01 8.40436593e-02 -4.70306724e-01 7.71949291e-01
1.05386651e+00 2.99754947e-01 5.50650001e-01 -1.44058061e+00
-1.31899178e-01 1.64785790e+00 2.81560451e-01 6.38453424e-01
7.07806468e-01 -4.82154489e-01 1.71021938e+00 -7.46121526e-01
7.42011964e-01 1.08688533e+00 9.46068704e-01 4.74284589e-01
-9.40015793e-01 -1.06151056e+00 -7.72938430e-02 3.69815260e-01
-1.84054673e+00 -9.73739624e-01 6.33372128e-01 -8.23441267e-01
7.77317047e-01 4.88190770e-01 6.13130212e-01 1.27016759e+00
-4.58137929e-01 8.98034692e-01 4.74936366e-01 -4.52219933e-01
1.60765469e-01 -2.04746634e-01 3.84833783e-01 3.21274787e-01
-4.47328925e-01 -4.09973770e-01 -1.36139965e+00 -4.10370946e-01
9.82955992e-02 -9.94229078e-01 -3.86841893e-01 2.76013929e-02
-1.18300092e+00 5.51339567e-01 4.22833227e-02 4.88101691e-01
-6.58943355e-02 4.69243735e-01 1.23814344e+00 2.92750835e-01
7.23361075e-01 5.61056435e-01 -7.03301847e-01 -7.47179389e-01
-8.65831733e-01 3.28063905e-01 7.98654318e-01 1.00134408e+00
1.38358474e-01 5.08231670e-02 -3.75769436e-01 1.31939530e+00
2.84480363e-01 3.19809467e-01 4.42877263e-01 -1.12318182e+00
4.91271168e-01 -2.31265530e-01 -3.68976151e-03 -3.47225785e-01
-5.38261533e-01 -3.21727246e-01 -1.38060778e-01 -4.57367092e-01
-9.10276361e-03 -3.37357968e-01 -7.68143117e-01 1.93697262e+00
1.43331200e-01 7.08169281e-01 -5.59057482e-02 7.96326756e-01
1.70547378e+00 7.66594648e-01 6.76205933e-01 2.09172353e-01
1.63954985e+00 -6.29630744e-01 -9.19540703e-01 2.26906136e-01
5.93283474e-01 -9.03010547e-01 1.19429922e+00 4.81676638e-01
-8.62570703e-01 -5.90867996e-01 -8.72033238e-01 -1.08024456e-01
-3.42926294e-01 2.11844698e-01 5.28929710e-01 4.36471730e-01
-1.00800788e+00 2.55848289e-01 -3.34833652e-01 -4.33913678e-01
1.86441019e-01 -2.07666025e-01 -4.82346714e-01 5.58734953e-01
-1.69269168e+00 1.21859178e-01 8.53000581e-01 -5.14914036e-01
-1.32115006e+00 -1.25666440e+00 -8.45557630e-01 2.79273000e-02
1.80655092e-01 -1.22613817e-01 2.03121924e+00 -8.19628239e-01
-1.25782549e+00 1.30036187e+00 1.98359281e-01 -7.30204165e-01
-1.96758628e-01 -5.76219141e-01 -1.07342577e+00 3.77974331e-01
2.98056811e-01 9.53382850e-01 6.91354215e-01 -6.08573198e-01
-1.02376652e+00 2.52774388e-01 1.69314981e-01 2.10652128e-01
-3.78576279e-01 5.34695148e-01 -1.56310350e-01 -1.03456616e+00
-2.53848940e-01 -1.04158163e+00 5.80781400e-01 -6.25869036e-01
-7.71411121e-01 -6.71431303e-01 6.24660790e-01 -8.12478125e-01
1.53677261e+00 -2.95152926e+00 -1.91408068e-01 -2.46219441e-01
-3.98152135e-02 -3.14201415e-01 -3.51736218e-01 3.97109985e-01
-5.14572382e-01 1.43973053e-01 4.52404529e-01 -3.01374465e-01
5.16874969e-01 -3.45950663e-01 -9.16719496e-01 -1.15529731e-01
-2.16852471e-01 4.38709319e-01 -1.12038815e+00 -4.65176195e-01
-4.62867171e-01 2.41131470e-01 -6.22268796e-01 4.13537532e-01
-2.00262770e-01 2.66974896e-01 6.96856380e-02 5.52701890e-01
6.47645146e-02 3.09659809e-01 -1.99069843e-01 -5.39762974e-01
-5.51862903e-02 1.14316189e+00 -1.13228226e+00 2.34040213e+00
-3.03406119e-01 9.45505798e-01 1.79722756e-02 -4.40593928e-01
5.19824147e-01 1.13076603e+00 8.68273675e-01 -2.12885961e-02
5.37504256e-02 2.88175106e-01 -2.39050403e-01 -5.16796231e-01
5.57586730e-01 -1.52227819e-01 -8.82495582e-01 3.87051523e-01
9.27463114e-01 -4.78229165e-01 3.07032734e-01 3.15254211e-01
1.43741095e+00 7.93276634e-03 1.03753440e-01 -2.63774693e-01
2.35105485e-01 -2.02108979e-01 4.73217905e-01 5.34384072e-01
-2.37413898e-01 8.45920265e-01 2.43327007e-01 -2.43996978e-01
-6.22791290e-01 -1.25469077e+00 -3.45972985e-01 1.84813154e+00
-5.30434012e-01 -1.19956744e+00 -5.48393846e-01 -6.22469664e-01
-4.82054800e-02 4.55547154e-01 -5.32338500e-01 -1.01791285e-01
-1.10642560e-01 6.00520819e-02 1.43944943e+00 3.47979307e-01
2.46808618e-01 -1.33678973e+00 -1.26657993e-01 3.93955290e-01
-1.00539339e+00 -1.20744300e+00 -8.13517094e-01 5.20849049e-01
-2.54963711e-02 -1.01431227e+00 -4.02899712e-01 -9.75395322e-01
-5.91621339e-01 -1.37997702e-01 1.59921503e+00 -6.56452119e-01
-9.71613303e-02 7.21514106e-01 -7.99786687e-01 -8.78190339e-01
-2.25147501e-01 6.94449663e-01 1.23937458e-01 -5.04812635e-02
5.22440434e-01 -5.87111712e-01 -2.50588655e-01 4.18360680e-02
-4.72724587e-01 -3.98410112e-01 -2.14930713e-01 2.34207854e-01
8.19168329e-01 4.88852896e-03 1.03443933e+00 -7.09536672e-01
3.97482187e-01 -6.80531085e-01 1.06752969e-01 -2.37760946e-01
2.43508786e-01 -4.83562142e-01 2.41105556e-01 -5.60181081e-01
-5.49983680e-01 2.24922523e-01 -6.78330839e-01 -2.90591210e-01
-2.60746539e-01 5.99662066e-01 -4.43451911e-01 3.72979105e-01
6.98906779e-01 -1.09108940e-01 -9.46264923e-01 -1.16957092e+00
3.68328661e-01 1.35122120e+00 1.15578866e+00 -7.66661108e-01
3.71483773e-01 5.25277071e-02 -3.96033734e-01 -7.11509645e-01
-1.49006712e+00 -9.15126503e-01 -5.73710680e-01 -4.37963277e-01
8.84650052e-01 -1.48143148e+00 -2.77865618e-01 3.27230781e-01
-9.68991876e-01 -3.65274638e-01 -8.15597773e-01 4.13741708e-01
-7.78644800e-01 -2.11342946e-01 -6.23242438e-01 -6.76498711e-01
-3.33398491e-01 -3.11358720e-01 1.46017241e+00 -2.59720951e-01
-8.93249094e-01 -5.98615766e-01 6.01646960e-01 2.22732469e-01
-5.40842302e-02 5.34887835e-02 5.96508026e-01 -1.13862860e+00
1.29786193e-01 -1.15944996e-01 1.41262665e-01 4.40811723e-01
-9.39680934e-02 -3.22251990e-02 -1.64988089e+00 -6.14431575e-02
-6.43017292e-01 -8.90609384e-01 8.65105808e-01 2.18551934e-01
1.37549829e+00 -1.99182257e-01 -1.23970173e-01 4.73196089e-01
8.18561673e-01 1.67597711e-01 3.32252651e-01 3.09070855e-01
4.73246396e-01 3.38364571e-01 6.73801422e-01 6.50745094e-01
3.97500634e-01 1.17198920e+00 1.07096568e-01 3.42649579e-01
-7.50926137e-01 -6.43817067e-01 4.15897489e-01 1.37533498e+00
4.28779870e-01 -2.41821423e-01 -1.03582871e+00 9.55984056e-01
-1.45357585e+00 -1.01684117e+00 -9.21442211e-02 1.75002575e+00
1.40945745e+00 -6.57909289e-02 5.29559374e-01 5.15802622e-01
7.07062185e-01 1.16465904e-01 1.53881773e-01 -2.84630597e-01
-1.34143159e-02 4.23299730e-01 -2.24692523e-01 9.38412792e-04
-1.96954441e+00 6.19623959e-01 7.55127907e+00 1.11306512e+00
-8.63234401e-01 7.18182504e-01 -8.32599774e-02 -3.13570440e-01
5.93319647e-02 -1.24354079e-01 -9.59080637e-01 7.52552032e-01
1.82258666e+00 -2.29633197e-01 1.54734790e-01 6.92288578e-01
-1.00378813e-02 3.06525558e-01 -1.20403683e+00 1.11655784e+00
6.62145987e-02 -1.01270008e+00 -1.21568345e-01 -4.87529963e-01
4.02792484e-01 3.14232498e-01 -8.77128169e-02 7.21211374e-01
2.63173252e-01 -5.92015386e-01 1.47626388e+00 3.89650911e-01
1.46396196e+00 -5.63828766e-01 6.37989819e-01 -4.00470167e-01
-1.65138078e+00 -3.78014743e-02 7.50815421e-02 1.18111067e-01
2.94379205e-01 6.16949558e-01 -7.82348871e-01 2.15607777e-01
1.32631147e+00 8.96498024e-01 -3.87205958e-01 1.54980385e+00
-2.36737251e-01 1.36567402e+00 -4.45395917e-01 4.84875441e-01
-1.27877638e-01 8.11928451e-01 9.85286355e-01 1.71911597e+00
4.34286833e-01 -1.00708015e-01 2.95357168e-01 1.40884873e-02
-5.88076055e-01 2.93644816e-01 -3.76912653e-01 -3.73686671e-01
1.00880826e+00 1.16475844e+00 -3.55854869e-01 -3.64941835e-01
3.23039368e-02 3.54478121e-01 -6.42638355e-02 -2.10665297e-02
-8.60065579e-01 -9.22886252e-01 9.12986875e-01 1.38881326e-01
1.75272882e-01 2.93644935e-01 1.43584520e-01 -1.09759617e+00
-4.30616170e-01 -5.17981231e-01 1.05206466e+00 -1.37431848e+00
-1.28781974e+00 7.24566996e-01 1.67910263e-01 -1.62259364e+00
-4.23807919e-01 7.48428479e-02 -2.25922287e-01 3.95643443e-01
-1.26449811e+00 -1.12411702e+00 -2.93957025e-01 6.08520925e-01
5.83651781e-01 -5.11041582e-01 1.33574355e+00 9.55364108e-01
6.76651597e-02 8.22179437e-01 -1.80484474e-01 3.23779315e-01
1.30075979e+00 -1.37672913e+00 2.35952497e-01 3.33384335e-01
6.67697489e-01 2.92905886e-02 8.59894037e-01 -3.48540187e-01
-4.88518894e-01 -1.42780983e+00 1.30609035e+00 -6.70006514e-01
1.27716696e+00 -7.95992136e-01 -5.98563910e-01 7.00078189e-01
2.25446135e-01 -5.20475879e-02 1.10638022e+00 5.96405566e-01
-8.71031284e-01 -1.48119077e-01 -6.77225351e-01 8.12887773e-03
1.32885087e+00 -1.21514893e+00 -7.82757521e-01 5.02879262e-01
1.21954858e+00 -4.94944781e-01 -1.34483695e+00 2.27581695e-01
5.42281866e-01 8.44610333e-02 9.11610663e-01 -6.02394104e-01
1.79346651e-01 -3.32745016e-01 -4.89853740e-01 -1.52060866e+00
-3.02185655e-01 -9.37246740e-01 -1.97071895e-01 2.09560537e+00
3.00930977e-01 1.34502232e-01 3.01214188e-01 -3.44413638e-01
-7.91095495e-01 3.11039388e-01 -1.62266159e+00 -1.05161226e+00
-2.50544012e-01 -9.21481967e-01 5.20112932e-01 9.35096204e-01
4.46539283e-01 8.16819251e-01 -2.51881272e-01 -4.08203267e-02
2.80055702e-01 -2.75748014e-01 5.80436826e-01 -1.74773860e+00
-2.07920477e-01 -6.91379681e-02 -7.33108819e-01 -4.72179204e-01
4.91669595e-01 -1.18014121e+00 4.67557520e-01 -9.89866316e-01
1.96390495e-01 -6.52851343e-01 -7.33249545e-01 1.18866062e+00
3.05397391e-01 9.69352067e-01 3.09809148e-01 3.27992886e-01
-1.18598723e+00 2.32988432e-01 2.20100373e-01 -5.80673516e-01
-8.57833251e-02 5.24900369e-02 -7.80302405e-01 4.89466250e-01
5.82712293e-01 -8.27054501e-01 -2.91813076e-01 -3.49777758e-01
4.25663561e-01 -1.44075990e-01 9.91861969e-02 -1.47916031e+00
-1.83216795e-01 3.30664426e-01 -3.10240805e-01 -6.26055658e-01
5.99710941e-01 -6.13533735e-01 5.03255785e-01 -3.35768968e-01
-8.44827771e-01 -3.36823016e-01 4.16031599e-01 2.12790400e-01
-6.69710517e-01 -4.03203696e-01 5.47222078e-01 3.33044112e-01
-8.07476461e-01 -4.23396975e-02 -6.19954705e-01 6.32724226e-01
4.60251480e-01 5.16245306e-01 -2.99096167e-01 -5.06437302e-01
-1.38968480e+00 -2.58026540e-01 -1.49924383e-01 8.40903103e-01
1.48910619e-02 -1.94145656e+00 -8.97528410e-01 -2.23116606e-01
6.79426074e-01 -5.87332010e-01 2.52427608e-01 4.53665853e-01
2.87760030e-02 7.75049329e-01 1.26352822e-02 -3.86325896e-01
-1.36175990e+00 5.60508110e-02 1.24599636e-01 -1.01064488e-01
-5.79773605e-01 1.22049797e+00 -1.09847076e-02 -1.40911445e-01
7.00571954e-01 -2.74611890e-01 -5.37061632e-01 4.61605579e-01
8.07223856e-01 1.25986487e-01 4.13275689e-01 -7.83399105e-01
-5.48979521e-01 8.33113790e-02 4.77204919e-01 -5.63567519e-01
1.35227537e+00 -1.86890259e-01 1.36875838e-01 1.35049105e+00
1.21601415e+00 3.23697209e-01 -7.88574696e-01 -1.90631449e-01
8.98657367e-02 2.38472987e-02 1.89460739e-01 -1.14860415e+00
-7.58016646e-01 7.36171603e-01 9.07657206e-01 5.62621832e-01
1.04892302e+00 3.82077187e-01 1.07212961e+00 2.19494715e-01
3.96507472e-01 -1.29047000e+00 -1.22159995e-01 9.47649121e-01
9.76360798e-01 -4.96896923e-01 -4.62151706e-01 -3.85977566e-01
-5.55432558e-01 7.99883246e-01 2.89791822e-01 5.28344661e-02
9.69460428e-01 3.23874861e-01 2.60117322e-01 -1.31917417e-01
-9.41565931e-01 -4.86543775e-01 7.29234338e-01 9.27664697e-01
6.34056509e-01 1.58047661e-01 2.31124610e-02 1.38622820e+00
-7.25946665e-01 9.33661312e-02 3.59147012e-01 6.53277993e-01
-4.40494925e-01 -9.23024654e-01 -1.01050548e-01 2.50750959e-01
-1.05486822e+00 -1.94948927e-01 -6.15262091e-01 2.96443135e-01
3.65800500e-01 1.02718222e+00 2.03275040e-01 -6.84127510e-01
5.38993299e-01 6.56442702e-01 1.38296261e-02 -1.12263274e+00
-7.54055560e-01 1.86330929e-01 6.96930349e-01 -4.29253697e-01
-4.73102033e-01 -7.78447866e-01 -1.16851521e+00 3.46690863e-01
4.38267067e-02 8.73052537e-01 6.63118720e-01 5.46625614e-01
5.03511488e-01 7.60702074e-01 4.68972802e-01 -8.04631531e-01
1.76184580e-01 -1.33544183e+00 -1.00373495e+00 5.24929106e-01
2.83730537e-01 -8.72708261e-01 -4.18510646e-01 5.45261443e-01] | [15.232882499694824, 5.064076900482178] |
7741255d-dbf9-4d1e-b2d9-aca451f27133 | why-existing-multimodal-crowd-counting | 2304.06401 | null | https://arxiv.org/abs/2304.06401v1 | https://arxiv.org/pdf/2304.06401v1.pdf | Why Existing Multimodal Crowd Counting Datasets Can Lead to Unfulfilled Expectations in Real-World Applications | More information leads to better decisions and predictions, right? Confirming this hypothesis, several studies concluded that the simultaneous use of optical and thermal images leads to better predictions in crowd counting. However, the way multimodal models extract enriched features from both modalities is not yet fully understood. Since the use of multimodal data usually increases the complexity, inference time, and memory requirements of the models, it is relevant to examine the differences and advantages of multimodal compared to monomodal models. In this work, all available multimodal datasets for crowd counting are used to investigate the differences between monomodal and multimodal models. To do so, we designed a monomodal architecture that considers the current state of research on monomodal crowd counting. In addition, several multimodal architectures have been developed using different multimodal learning strategies. The key components of the monomodal architecture are also used in the multimodal architectures to be able to answer whether multimodal models perform better in crowd counting in general. Surprisingly, no general answer to this question can be derived from the existing datasets. We found that the existing datasets hold a bias toward thermal images. This was determined by analyzing the relationship between the brightness of optical images and crowd count as well as examining the annotations made for each dataset. Since answering this question is important for future real-world applications of crowd counting, this paper establishes criteria for a potential dataset suitable for answering whether multimodal models perform better in crowd counting in general. | ['Elke Hergenröther', 'Martin Thißen'] | 2023-04-13 | null | null | null | null | ['crowd-counting'] | ['computer-vision'] | [-4.65387031e-02 -2.17621744e-01 5.85154332e-02 -3.12941432e-01
-2.61021584e-01 -5.66042125e-01 8.32643926e-01 3.84713411e-01
-8.66435468e-01 8.17614734e-01 6.71720132e-02 -1.69306949e-01
-9.94443670e-02 -7.39392340e-01 -3.12721789e-01 -7.55416870e-01
3.25842440e-01 5.80894291e-01 4.09991115e-01 -1.89349711e-01
5.75437844e-01 5.36872685e-01 -2.19928718e+00 5.54538786e-01
6.15335941e-01 7.26069152e-01 5.03766060e-01 8.31753731e-01
-4.40445036e-01 8.07557046e-01 -7.27715731e-01 -6.96693838e-01
-1.39339715e-01 -2.21015796e-01 -8.01918924e-01 -1.10758767e-01
6.29128695e-01 -2.98900187e-01 8.70527774e-02 7.20229208e-01
6.58011973e-01 1.10631242e-01 7.78631508e-01 -1.05034697e+00
-4.04772729e-01 2.02533662e-01 -2.59447217e-01 5.77915430e-01
7.35372782e-01 3.37466002e-01 6.42043769e-01 -8.29445779e-01
3.31689060e-01 1.25455713e+00 4.90638465e-01 4.92501676e-01
-1.09688330e+00 -2.82408476e-01 -7.70660862e-02 1.58110753e-01
-1.16070366e+00 -4.66128767e-01 5.24906933e-01 -7.40047634e-01
1.05743682e+00 3.79712969e-01 8.17734063e-01 9.79265034e-01
-1.54736951e-01 6.37351513e-01 1.51549602e+00 -7.82668114e-01
1.04132220e-01 5.15393257e-01 1.56460568e-01 6.80970311e-01
5.82580984e-01 4.50444482e-02 -6.83123589e-01 -4.87074107e-02
4.01566684e-01 -3.13277006e-01 3.63310613e-02 -6.20128773e-02
-1.09313619e+00 9.18164253e-01 -3.69945727e-02 7.79401064e-01
-7.20808133e-02 -6.90966994e-02 4.36951607e-01 -4.46612015e-02
2.61848181e-01 5.82077682e-01 -1.54600069e-01 -2.30017528e-01
-1.05719292e+00 1.51265085e-01 6.19535923e-01 3.67861658e-01
9.76474226e-01 -2.12041304e-01 -1.12480722e-01 7.06266105e-01
3.90667588e-01 8.05285096e-01 2.56114513e-01 -1.00474286e+00
5.89168489e-01 8.72180343e-01 2.12940529e-01 -1.28838789e+00
-8.56912792e-01 3.26664150e-01 -4.53347683e-01 2.53277719e-02
1.01135540e+00 -5.60633698e-03 -2.25424498e-01 1.45868027e+00
-4.01418656e-03 -1.89677522e-01 -2.69784275e-02 9.96351600e-01
1.04674256e+00 3.86883229e-01 3.62455100e-01 -9.53815505e-03
1.53860748e+00 -3.26322496e-01 -6.30827904e-01 -3.31798196e-01
6.58341646e-01 -9.31270301e-01 1.13967872e+00 9.73011106e-02
-1.07249284e+00 -6.92376792e-01 -7.49199986e-01 -1.07347213e-01
-9.17974114e-01 4.67980742e-01 7.93655634e-01 1.14405322e+00
-8.39826286e-01 9.26012397e-02 -6.05391562e-01 -8.77476275e-01
7.18421787e-02 4.36239630e-01 -4.25530910e-01 9.88401547e-02
-1.12396419e+00 1.45032179e+00 4.12634879e-01 2.05299154e-01
-4.21537071e-01 -6.25431165e-02 -7.63750792e-01 -2.50859231e-01
3.88826095e-02 -7.92854309e-01 9.20250356e-01 -9.86407995e-01
-1.29116189e+00 1.24107385e+00 -7.17490613e-01 -1.00382224e-01
4.50560093e-01 3.11527371e-01 -1.78090096e-01 3.18396538e-01
6.95632175e-02 7.24749625e-01 4.47732568e-01 -1.54734910e+00
-8.59018028e-01 -5.48897445e-01 1.63953990e-01 6.55626208e-02
-2.56055355e-01 6.69473559e-02 -1.66065142e-01 -2.66286228e-02
-3.30278575e-01 -7.64108479e-01 2.25107908e-01 -6.10915720e-01
3.88739184e-02 -3.46510023e-01 5.16187191e-01 -5.06147742e-01
1.18368816e+00 -1.84969985e+00 -4.55630794e-02 1.67908564e-01
1.05494715e-01 4.29754317e-01 -2.36821309e-01 5.43321073e-01
2.69140065e-01 1.48911640e-01 -1.34455236e-02 -4.65395570e-01
1.05623871e-01 3.87647480e-01 -8.24830681e-03 3.72691691e-01
3.10869157e-01 7.85953999e-01 -8.35828483e-01 -9.55186367e-01
7.29570746e-01 5.56189537e-01 -1.94128230e-01 6.13818429e-02
8.29327479e-02 7.34028101e-01 -7.73922354e-02 6.04261220e-01
7.16463804e-01 8.67729783e-02 1.93807989e-01 -1.99504927e-01
-4.09787089e-01 -2.58212239e-01 -8.50831032e-01 1.13442695e+00
-5.29929042e-01 1.00685620e+00 -3.40284944e-01 -8.88900578e-01
9.20250177e-01 2.42518395e-01 1.71120331e-01 -9.71600771e-01
4.14402515e-01 5.70140064e-01 -1.39800817e-01 -9.62497175e-01
9.56999838e-01 -3.27188611e-01 1.24705777e-01 2.15445384e-01
6.86822906e-02 -1.36306196e-01 8.11342061e-01 -2.27368653e-01
5.37687361e-01 -8.49465728e-02 2.61651963e-01 -3.96607589e-04
8.99937749e-01 6.36281520e-02 -2.69489625e-04 8.17190647e-01
-3.40555072e-01 4.47125375e-01 5.45226216e-01 -5.67865252e-01
-8.50991905e-01 -8.69774759e-01 -2.08961710e-01 1.20411837e+00
2.84753218e-02 1.12507492e-02 -7.69240797e-01 -3.92876655e-01
-2.38249153e-01 6.46196127e-01 -6.59828067e-01 3.68342578e-01
-3.49861890e-01 -9.07676280e-01 7.61490643e-01 4.27342892e-01
3.55055183e-01 -9.76159871e-01 -1.14005578e+00 -2.79546022e-01
-6.63381934e-01 -1.21115661e+00 2.45579392e-01 -2.99227983e-02
-9.93841648e-01 -1.33308601e+00 -6.73474967e-01 -4.41834927e-01
4.87905711e-01 3.13329905e-01 1.27886033e+00 3.86039585e-01
9.05965120e-02 8.51649046e-01 -3.62752110e-01 -5.43406785e-01
-3.35390180e-01 2.50026584e-01 -2.08173096e-01 -1.51068330e-01
9.30661798e-01 -1.48594752e-02 -3.08477849e-01 4.05618876e-01
-1.21738243e+00 -4.73608822e-01 4.98011887e-01 5.95259070e-01
7.26128146e-02 -2.81068981e-01 2.73334056e-01 -3.08000416e-01
7.38000572e-01 -2.88448662e-01 -6.49413824e-01 3.92888367e-01
-8.21394920e-02 4.56483029e-02 2.49846861e-01 -2.18658939e-01
-1.24681914e+00 9.76967290e-02 1.69495344e-02 1.45445913e-01
-5.72122514e-01 4.64607298e-01 7.97219127e-02 -2.05341443e-01
5.86781502e-01 1.82494655e-01 7.03474283e-02 -9.58749056e-02
-3.30052041e-02 6.29655838e-01 9.03605595e-02 -6.47847950e-01
3.46632481e-01 7.02973068e-01 1.86849490e-01 -1.30723906e+00
-6.03412867e-01 -6.13525391e-01 -9.48858440e-01 -6.05695426e-01
1.17809737e+00 -5.32796383e-01 -1.25471640e+00 3.71109694e-01
-1.42923450e+00 -2.34843213e-02 4.58665378e-02 6.11567378e-01
-4.88650292e-01 7.25946128e-01 -7.96295404e-02 -1.49846780e+00
2.42429018e-01 -1.28817225e+00 1.21528625e+00 2.66727597e-01
-3.45790654e-01 -1.34929979e+00 3.97003949e-01 9.10496712e-01
3.58194023e-01 -3.30832340e-02 9.19754267e-01 -5.65060377e-01
-2.96064913e-01 -3.24776679e-01 -3.15943867e-01 1.61329672e-01
-3.55342358e-01 1.89693376e-01 -1.31806123e+00 9.73305851e-02
-3.39969039e-01 -3.64393502e-01 9.84727383e-01 3.30646038e-01
6.58679664e-01 2.84225136e-01 -1.01502620e-01 -1.14243820e-01
1.40294695e+00 5.63594932e-03 8.19954753e-01 4.86338377e-01
3.85074198e-01 1.26487517e+00 5.23172617e-01 4.02348727e-01
7.51752257e-01 6.32754147e-01 5.00196397e-01 1.28524706e-01
1.72928676e-01 -9.83244926e-02 3.22594345e-01 5.24125814e-01
-6.98334038e-01 -3.35089177e-01 -1.24350417e+00 4.88361210e-01
-1.74244630e+00 -1.30609286e+00 -4.32717115e-01 2.36390519e+00
1.50778055e-01 -3.68483692e-01 4.42946076e-01 2.49864966e-01
6.62828982e-01 1.16136581e-01 2.52534211e-01 -5.72082222e-01
-5.86293340e-01 -6.97683096e-02 3.71003121e-01 4.69352126e-01
-1.10987437e+00 5.66135406e-01 6.79786301e+00 5.50351918e-01
-1.09839261e+00 -1.83928236e-02 5.41787505e-01 6.04673894e-03
-1.60138890e-01 -2.83076484e-02 -8.12896192e-01 5.37256539e-01
9.29130554e-01 6.11798823e-01 3.97773623e-01 2.90700912e-01
3.54410321e-01 -1.01003528e+00 -9.90449071e-01 1.04472268e+00
4.05381769e-01 -9.81031775e-01 3.62800546e-02 2.53756940e-02
4.24799025e-01 -2.48813555e-01 -4.85375486e-02 2.25425661e-01
-2.81956911e-01 -1.11286819e+00 7.89237976e-01 1.16251624e+00
1.91246122e-01 -4.79578078e-01 1.43322670e+00 4.03648525e-01
-1.05869448e+00 -1.40401319e-01 -4.59883183e-01 -3.91280204e-01
1.22492597e-01 2.60857224e-01 -7.45327234e-01 5.32206774e-01
5.66778362e-01 1.08936001e-02 -1.09032393e+00 1.00853014e+00
7.56639317e-02 2.53560811e-01 -3.66777897e-01 -3.51215720e-01
1.51933998e-01 -1.68436646e-01 2.57600814e-01 1.54547322e+00
3.80420357e-01 -4.59671795e-01 -1.96630150e-01 6.88701153e-01
6.20770037e-01 3.77710819e-01 -8.78002286e-01 -1.21390015e-01
2.63962686e-01 1.08801961e+00 -8.81513715e-01 -1.93842158e-01
-7.82927573e-01 5.66164970e-01 3.53504628e-01 3.18446517e-01
-7.46439397e-01 7.03959689e-02 -4.67764437e-02 1.81568891e-01
-1.82315502e-02 -2.73074627e-01 -4.48958814e-01 -9.26413417e-01
-1.32101953e-01 -6.22949004e-01 4.20570314e-01 -8.87774527e-01
-1.47870779e+00 2.39683151e-01 3.80715013e-01 -7.93272734e-01
-1.03092432e-01 -9.84313726e-01 -4.06048596e-01 9.40700948e-01
-1.35779345e+00 -1.24257922e+00 -5.04065394e-01 4.76488829e-01
3.80027778e-02 -2.21948996e-01 7.65604138e-01 2.28932455e-01
-4.15440321e-01 1.90576971e-01 -3.01237702e-01 7.71707147e-02
6.16431296e-01 -1.14453804e+00 -5.71414709e-01 5.43202102e-01
2.58409195e-02 5.77495992e-01 6.38088584e-01 -4.44075972e-01
-1.03887188e+00 -3.78214002e-01 1.30650234e+00 -9.84764099e-01
4.66856271e-01 1.62056983e-01 -6.25863373e-01 2.75613934e-01
3.87606829e-01 -3.28999996e-01 9.38867211e-01 1.33416429e-01
-1.54801443e-01 4.98984829e-02 -1.10069227e+00 3.10289502e-01
4.34063286e-01 -6.39934838e-01 -6.76046073e-01 -8.64841789e-02
-2.45754734e-01 7.32981265e-02 -6.79244578e-01 1.89318910e-01
5.64033747e-01 -1.60316849e+00 8.32121849e-01 -3.19918007e-01
4.63737756e-01 -3.38504404e-01 -4.12840962e-01 -9.57323551e-01
1.44383907e-01 2.25648671e-01 2.30544969e-01 1.06985068e+00
6.64498329e-01 -8.06113780e-01 6.18426383e-01 6.10719085e-01
1.38297945e-01 -3.87242615e-01 -1.06441975e+00 -5.42222023e-01
3.24225813e-01 -5.31751692e-01 5.14563322e-01 7.69978523e-01
1.26654074e-01 2.48425648e-01 -2.62461960e-01 7.75050223e-02
1.92564771e-01 -1.95988894e-01 1.02529955e+00 -1.27801466e+00
1.35240793e-01 -7.23134398e-01 -6.48768723e-01 -6.48725808e-01
3.96945477e-01 -5.74757159e-01 -2.48496279e-01 -1.58868766e+00
4.90359008e-01 2.26708837e-02 2.49321416e-01 8.71586129e-02
-1.86828047e-01 4.27895874e-01 5.73728502e-01 8.30721036e-02
-7.02018917e-01 1.64862081e-01 1.13289583e+00 -8.21833238e-02
5.57294749e-02 -1.34103224e-01 -2.01076105e-01 9.37366366e-01
9.69979882e-01 -7.68289901e-03 -1.59864157e-01 -5.94003320e-01
6.97456241e-01 -6.14304394e-02 6.18819535e-01 -1.00114357e+00
2.39054054e-01 -5.95284328e-02 4.27919030e-01 -7.53365457e-01
6.78981781e-01 -8.32614720e-01 -1.73059404e-01 2.58765072e-01
1.34043014e-02 1.17469922e-01 3.28320295e-01 3.75818968e-01
-4.69115436e-01 -7.46752560e-01 6.68199241e-01 -3.24401051e-01
-7.25776732e-01 -5.32501161e-01 -8.17040563e-01 -1.21402018e-01
9.46927845e-01 -7.47435570e-01 -5.30110419e-01 -2.68033206e-01
-6.71861291e-01 -1.77603140e-02 4.83994633e-01 1.55185178e-01
4.16958302e-01 -9.68818665e-01 -4.35895920e-01 -1.57787189e-01
3.02312046e-01 -5.41227341e-01 3.63417536e-01 1.42781591e+00
-6.23409152e-01 8.81740093e-01 -2.92902529e-01 -6.61392510e-01
-1.36743736e+00 4.96945947e-01 4.54025149e-01 -2.86382765e-01
4.86271024e-01 4.44779038e-01 -2.20000863e-01 -6.15630329e-01
4.71783467e-02 -1.62621036e-01 -7.07015276e-01 7.06365943e-01
3.37106556e-01 1.01089978e+00 3.28306146e-02 -1.17507076e+00
-3.98380727e-01 7.74929821e-01 6.26570106e-01 -2.11735487e-01
8.84919167e-01 -4.19211090e-01 -4.02662933e-01 8.96787345e-01
8.34890306e-01 1.17637932e-01 -6.76221311e-01 2.57955939e-01
1.45708010e-01 -6.13046944e-01 -2.73384899e-01 -6.40944898e-01
-8.76109958e-01 1.21915960e+00 8.55823994e-01 5.35652459e-01
9.79702950e-01 -4.18030992e-02 4.93853912e-02 4.57830906e-01
3.89555424e-01 -1.25482631e+00 1.24078818e-01 6.79343760e-01
6.40102267e-01 -1.74135375e+00 -9.19947177e-02 -1.03172153e-01
-9.20432746e-01 1.43060648e+00 4.74802405e-01 3.95890564e-01
1.90819278e-01 2.20082533e-02 2.27738079e-02 -4.57487702e-01
-2.51336575e-01 -6.83590472e-01 3.38099808e-01 9.45080757e-01
8.69398773e-01 1.73344478e-01 -5.86314917e-01 3.86484295e-01
-9.09169614e-02 4.49975133e-02 4.19273645e-01 7.55903542e-01
-6.98228717e-01 -1.03128755e+00 -1.12880027e+00 3.80633920e-01
-2.66222328e-01 9.38606933e-02 -7.84121692e-01 9.92545128e-01
6.11269057e-01 1.47348714e+00 1.12291135e-01 -2.69672275e-01
3.97407681e-01 2.59926438e-01 7.68697858e-01 -2.52611876e-01
-6.59005761e-01 -3.74318808e-01 2.35566437e-01 -1.59232929e-01
-1.24408245e+00 -8.19820702e-01 -6.96407020e-01 -6.70112610e-01
-3.53946567e-01 7.05413288e-03 6.14778936e-01 1.15229285e+00
-1.46862671e-01 1.01620242e-01 9.87764150e-02 -1.17162001e+00
3.42154801e-02 -1.00286531e+00 -2.08587423e-01 2.56975323e-01
-1.64568618e-01 -8.24721754e-01 -4.70345080e-01 -5.65315187e-02] | [12.975988388061523, 5.216159820556641] |
70f3ccf7-a054-49c8-a4d0-cc574431f299 | performance-evaluation-and-hybrid-application | 2302.11740 | null | https://arxiv.org/abs/2302.11740v1 | https://arxiv.org/pdf/2302.11740v1.pdf | Performance Evaluation and Hybrid Application of the Greedy and Predictive UAV Trajectory Optimization Methods for Localizing a Target Mobile Device | This study investigates unmanned aerial vehicle (UAV) trajectory planning strategies for localizing a target mobile device in emergency situations. The global navigation satellite system (GNSS)-based accurate position information of a target mobile device in an emergency may not be always available to first responders. For example, 1) GNSS positioning accuracy may be degraded in harsh signal environments and 2) in countries where emergency positioning service is not mandatory, some mobile devices may not report their locations. Under the cases mentioned above, one way to find the target mobile device is to use UAVs. Dispatched UAVs may search the target directly on the emergency site by measuring the strength of the signal (e.g., LTE wireless communication signal) from the target mobile device. To accurately localize the target mobile device in the shortest time possible, UAVs should fly in the most efficient way possible. The two popular trajectory optimization strategies of UAVs are greedy and predictive approaches. However, the research on localization performances of the two approaches has been evaluated only under favorable settings (i.e., under good UAV geometries and small received signal strength (RSS) errors); more realistic scenarios still remain unexplored. In this study, we compare the localization performance of the greedy and predictive approaches under realistic RSS errors (i.e., up to 6 dB according to the ITU-R channel model). | ['Jiwon Seo', 'Halim Lee'] | 2023-02-23 | null | null | null | null | ['trajectory-planning'] | ['robots'] | [-2.00746413e-02 -2.59755135e-01 -6.87619224e-02 3.16548347e-01
-3.75414044e-01 -8.18919420e-01 1.45149067e-01 3.31115365e-01
-4.67756808e-01 1.00619805e+00 -5.46041608e-01 -7.71863759e-01
-6.38697326e-01 -8.77315700e-01 -5.16764283e-01 -9.43802238e-01
-3.74080569e-01 8.06744397e-02 1.87227920e-01 -2.97294885e-01
1.97834969e-02 1.00111234e+00 -1.21195316e+00 -7.27392495e-01
8.79522979e-01 1.24252319e+00 5.27887225e-01 6.77302063e-01
6.18288755e-01 6.40894175e-02 -9.56710458e-01 1.90695003e-01
5.27553439e-01 -2.20290348e-01 -5.07533133e-01 2.91457418e-02
-5.08066297e-01 -4.32278752e-01 -9.91537049e-02 9.06071424e-01
5.30826688e-01 2.17554986e-01 4.15437788e-01 -1.30578053e+00
4.44183022e-01 2.54152805e-01 -3.86330426e-01 3.22502762e-01
5.68179548e-01 -1.44078612e-01 2.77040988e-01 -3.42528641e-01
4.17561799e-01 3.13673526e-01 7.12109923e-01 -1.81432694e-01
-6.10376596e-01 -2.14616463e-01 9.32150036e-02 -6.77225813e-02
-1.87690878e+00 -2.39926860e-01 9.18947458e-02 -1.28996372e-01
3.83522004e-01 3.90196204e-01 6.45173967e-01 7.09923446e-01
5.07543862e-01 3.91672663e-02 4.35287148e-01 -2.48533890e-01
3.25031012e-01 -2.48512272e-02 -4.21021670e-01 4.03297186e-01
7.51527667e-01 2.68622441e-03 7.84081146e-02 -4.29204881e-01
2.95358032e-01 -2.73905718e-03 -8.66897047e-01 -1.56373858e-01
-1.52844763e+00 2.93212622e-01 6.56368971e-01 6.40582502e-01
-1.13024211e+00 -1.77007779e-01 2.49255411e-02 4.01291370e-01
1.79525018e-01 6.28930271e-01 -5.20246327e-01 -1.92683652e-01
-1.13275778e+00 5.10238521e-02 5.74738979e-01 1.24664545e+00
3.36661875e-01 2.07204118e-01 2.41158083e-01 1.10896267e-01
2.45044619e-01 9.52495456e-01 1.04444794e-01 -3.90480757e-01
6.28586471e-01 9.79356989e-02 7.66930342e-01 -1.49403036e+00
-5.90304911e-01 -9.55168903e-01 -9.23392534e-01 -3.47678602e-01
4.74499375e-01 -1.03055954e+00 -3.12071174e-01 1.23763955e+00
4.56390172e-01 4.31155354e-01 2.20325649e-01 9.39059913e-01
2.35645130e-01 7.70103514e-01 -2.53290921e-01 -7.24404991e-01
8.67617726e-01 -1.71704054e-01 -4.61662233e-01 -2.23258480e-01
8.83155942e-01 -8.03968906e-01 2.28532881e-01 2.69996375e-01
-6.07657433e-01 -1.21017620e-01 -1.08980954e+00 1.17631733e+00
-2.01861486e-01 6.30333722e-01 -4.50906204e-03 7.12472379e-01
-1.00643349e+00 4.12402928e-01 -7.65757143e-01 -7.35777915e-01
-2.93325782e-01 5.43167293e-01 -9.48924869e-02 -3.05771351e-01
-1.29513061e+00 5.62840879e-01 1.07717775e-01 4.07888502e-01
-6.16096795e-01 -1.59861192e-01 -5.52080333e-01 -9.87264663e-02
4.74597573e-01 -6.14272833e-01 9.03272569e-01 -6.72507405e-01
-1.15789795e+00 1.56668406e-02 -1.73724853e-02 -6.04848325e-01
2.80971438e-01 1.59932002e-01 -4.90478307e-01 9.54820439e-02
2.87682086e-01 2.57084519e-03 5.40434778e-01 -1.09133887e+00
-9.84428048e-01 -2.55625725e-01 4.13191199e-01 4.48319972e-01
-1.37218356e-01 -1.13576382e-01 4.98699546e-02 -3.81843358e-01
4.51652527e-01 -1.39079797e+00 -3.47657561e-01 -6.44811630e-01
-5.80141306e-01 3.78529161e-01 8.09395015e-01 -3.22662890e-01
1.33170176e+00 -1.98456252e+00 -8.01013857e-02 5.18721700e-01
-3.72642964e-01 3.94079953e-01 2.10518226e-01 6.80911720e-01
1.39461398e-01 3.93588729e-02 -7.51412613e-03 1.17035516e-01
-3.78975630e-01 -1.81890652e-01 -4.09416288e-01 6.94497108e-01
-6.83240891e-01 3.18788797e-01 -1.01067734e+00 3.59773301e-02
-3.74509883e-03 1.80169493e-01 3.59803177e-02 -9.28635523e-03
4.13782537e-01 7.02571630e-01 -8.86287451e-01 9.32863772e-01
7.03463793e-01 1.77382559e-01 7.18209818e-02 5.01685590e-03
-7.19732463e-01 -7.73428520e-03 -1.21921730e+00 9.97254729e-01
-5.83797336e-01 5.41466892e-01 4.03514832e-01 -7.84825265e-01
8.27201486e-01 4.61637646e-01 7.20820129e-01 -7.58258402e-02
2.41466820e-01 4.69806314e-01 -1.34707108e-01 -1.61931038e-01
6.82048976e-01 2.24154651e-01 -2.30005950e-01 2.24492207e-01
-5.54540396e-01 5.46400622e-02 -7.93163553e-02 -8.37715566e-02
1.30583477e+00 -2.97384262e-01 5.78182220e-01 -9.88736749e-02
3.85315567e-01 2.13065520e-01 4.12399650e-01 7.53687918e-01
-3.36120538e-02 1.35624886e-01 1.28389999e-01 -6.04290776e-02
-2.15192184e-01 -4.98385519e-01 -5.02418019e-02 4.35387433e-01
6.60151064e-01 -3.43885452e-01 -6.49944305e-01 -4.81890082e-01
-2.59129971e-01 7.38807082e-01 4.15126346e-02 -3.77379768e-02
-1.61565468e-01 -8.36761236e-01 6.17020071e-01 -1.39982924e-01
7.77175844e-01 -2.45226935e-01 -9.80666459e-01 4.14296269e-01
-4.26300585e-01 -1.22967505e+00 -2.45956834e-02 -4.62702475e-03
-6.42803252e-01 -1.22862101e+00 -8.94125640e-01 -3.37413460e-01
8.20943773e-01 1.16160536e+00 3.74299914e-01 2.59291500e-01
4.59975749e-01 8.84240806e-01 -6.97463632e-01 -3.53602022e-01
1.41461134e-01 1.52567551e-01 5.02912998e-01 2.39729047e-01
-2.83391982e-01 -1.40544981e-01 -6.24924242e-01 9.68259096e-01
-4.50564623e-01 -6.30291760e-01 2.76099831e-01 4.30305719e-01
5.78495800e-01 9.60034072e-01 2.43670985e-01 -2.44874470e-02
6.56141400e-01 -9.46028650e-01 -8.47970068e-01 2.55789995e-01
-3.51393610e-01 -7.82962501e-01 6.61700249e-01 -1.76162109e-01
-4.21170890e-01 8.25891271e-02 8.39394480e-02 1.79928672e-02
-3.50751519e-01 9.33147073e-01 -2.45528847e-01 -5.97522438e-01
3.97456318e-01 2.72534102e-01 -6.07413113e-01 6.92299232e-02
-3.70998025e-01 8.65195870e-01 4.45159115e-02 -7.61018619e-02
9.84042943e-01 3.05321723e-01 4.25487012e-01 -1.44739711e+00
-9.84548479e-02 -4.94077504e-01 -4.92170811e-01 -4.51054841e-01
3.96221757e-01 -8.33780289e-01 -6.42463744e-01 3.62656206e-01
-1.22957981e+00 2.63833869e-02 2.68943012e-01 9.51975226e-01
-2.94905841e-01 4.53052640e-01 1.79356009e-01 -1.01266491e+00
8.21806788e-02 -1.32941914e+00 8.51037323e-01 1.95510775e-01
-1.43375218e-01 -8.00741315e-01 -4.47313637e-01 -3.66797537e-01
6.08119607e-01 7.03446329e-01 2.20975757e-01 -3.76479506e-01
-7.17354834e-01 -7.44683802e-01 1.92145005e-01 -3.64794493e-01
2.89830118e-01 -1.46545649e-01 -2.17532992e-01 -5.99150777e-01
8.05884674e-02 5.14609039e-01 5.84677979e-02 8.27897549e-01
4.96395618e-01 -2.85749823e-01 -8.12429607e-01 6.43741727e-01
1.18980706e+00 6.68263733e-01 3.79543543e-01 6.08563960e-01
1.43308386e-01 3.82676631e-01 1.58780515e+00 7.72219360e-01
2.49634534e-01 9.61457372e-01 9.39674735e-01 1.56233087e-01
6.98942125e-01 2.74669994e-02 3.09517086e-01 2.29730695e-01
-2.75429994e-01 -1.18844676e+00 -9.49741483e-01 3.81357670e-01
-1.58272159e+00 -6.32758081e-01 -3.16000134e-01 2.66817784e+00
-3.20624262e-01 -1.75804764e-01 -2.40933634e-02 3.99223149e-01
8.46481860e-01 3.35888378e-02 -2.68284708e-01 9.22736004e-02
-3.02658156e-02 -6.43289387e-01 1.17968416e+00 4.01067048e-01
-1.07651377e+00 5.18781304e-01 4.31439829e+00 4.71621841e-01
-1.31387532e+00 -1.06579959e-01 5.59584573e-02 3.50367188e-01
9.03307125e-02 1.52507722e-01 -7.16471910e-01 5.79422712e-01
8.79747748e-01 -1.67772412e-01 1.65415570e-01 6.97842598e-01
6.82083130e-01 -6.06830955e-01 -2.43984744e-01 1.00737774e+00
-3.06897819e-01 -7.51578271e-01 -5.53412676e-01 4.69523251e-01
4.15539145e-01 -1.06276020e-01 6.23627342e-02 -2.38408118e-01
-2.94809192e-01 -2.94967055e-01 4.81680244e-01 6.07169747e-01
6.77692354e-01 -1.07532787e+00 1.25452828e+00 7.89130330e-01
-1.45120502e+00 -3.69607002e-01 -3.81850123e-01 3.23790349e-02
6.72922432e-01 6.93034530e-01 -1.11244273e+00 1.03144670e+00
6.05290771e-01 4.14289474e-01 -1.34686783e-01 1.67790639e+00
-4.11716580e-01 6.12554848e-01 -6.71214879e-01 -1.68396175e-01
6.60316288e-01 -1.91452876e-01 1.28121066e+00 4.78219330e-01
1.49945486e+00 4.57261503e-01 7.21898228e-02 1.33582689e-02
4.39925879e-01 1.92475647e-01 -1.12590909e+00 2.03506276e-01
7.73214757e-01 8.63826752e-01 -9.71353114e-01 2.41156355e-01
-4.54483507e-03 8.97429645e-01 -6.41799152e-01 6.12707555e-01
-8.01366687e-01 -7.49390304e-01 5.66870034e-01 5.54202735e-01
1.24943934e-01 -8.58186483e-01 2.06728816e-01 -6.85245872e-01
-1.48450941e-01 -5.09275973e-01 -4.60432395e-02 -6.81954145e-01
-2.31317773e-01 8.52983415e-01 4.93886061e-02 -1.90553868e+00
-4.93945509e-01 -1.49908438e-01 -5.45212567e-01 6.96451306e-01
-1.05558288e+00 -7.04622805e-01 -4.00759786e-01 4.59955573e-01
1.38289928e-01 -3.25369179e-01 8.60375702e-01 2.25591004e-01
-4.34407771e-01 2.27236748e-01 3.28092545e-01 -3.71488005e-01
3.66029412e-01 -6.36982977e-01 1.69063762e-01 1.08534050e+00
-5.00480272e-02 7.77385950e-01 8.37091267e-01 -9.53862667e-01
-1.57603908e+00 -1.02556562e+00 8.80299985e-01 3.44943330e-02
3.70535970e-01 9.65839848e-02 -2.10700229e-01 4.95919108e-01
-3.29796039e-02 -2.33189180e-01 4.58214700e-01 -3.37635159e-01
8.94386232e-01 -7.94451907e-02 -1.28155208e+00 6.73924863e-01
7.45092511e-01 1.12636909e-01 2.71363676e-01 5.49484372e-01
2.53739208e-01 -5.39455950e-01 -6.32904410e-01 6.66397989e-01
3.46174598e-01 -8.21854651e-01 8.53820145e-01 1.64046213e-02
-8.27310681e-01 -7.45071828e-01 -4.96216029e-01 -1.81630552e+00
-1.80995055e-02 -9.41885948e-01 2.98485488e-01 9.39834356e-01
4.39120114e-01 -9.05318618e-01 3.36778313e-01 2.28188828e-01
3.45632285e-02 -6.45749807e-01 -1.12968206e+00 -8.97254229e-01
-8.47937286e-01 -5.15927970e-01 9.40426946e-01 5.13986826e-01
-2.00411022e-01 -1.37488261e-01 -5.37541866e-01 1.15308583e+00
2.83066928e-01 -2.32562721e-01 8.17729831e-01 -1.06839144e+00
1.03098065e-01 6.73995316e-02 -5.94476044e-01 -1.37214947e+00
-6.67305067e-02 -2.89155602e-01 1.56048685e-02 -1.58097744e+00
-1.05615604e+00 -9.15156066e-01 2.18188122e-01 3.87986563e-02
3.92966211e-01 -1.06882952e-01 -6.32224381e-02 1.49720892e-01
-3.48962456e-01 3.66120875e-01 6.75991237e-01 3.58899906e-02
-4.52206552e-01 1.29236102e+00 -3.71990830e-01 7.96610057e-01
9.96521115e-01 -5.31369090e-01 -6.75488055e-01 -3.99935871e-01
4.43992615e-01 7.15091586e-01 2.64628291e-01 -1.42233014e+00
6.63669944e-01 -1.62062034e-01 6.06757663e-02 -5.79534769e-01
4.25361246e-01 -1.40569806e+00 3.57782066e-01 5.63010216e-01
4.97919917e-01 4.33730215e-01 1.51481062e-01 8.77977669e-01
-2.69900650e-01 -4.22699034e-01 5.01933455e-01 2.26239190e-01
-4.67157811e-01 3.57684225e-01 -9.32097316e-01 -6.62872553e-01
1.44618928e+00 -3.60252261e-01 -7.67664760e-02 -1.02065885e+00
-5.00411630e-01 1.98036134e-01 4.75548297e-01 1.19160013e-02
7.23869622e-01 -1.00829971e+00 -1.42440185e-01 3.28919925e-02
-1.17769599e-01 -3.14241022e-01 2.11865708e-01 1.38653588e+00
-4.61715162e-01 7.84874141e-01 1.72541350e-01 -4.12532896e-01
-1.39819419e+00 2.08807275e-01 3.57346594e-01 2.83922285e-01
1.10973865e-01 6.75004959e-01 -4.20076489e-01 -9.54893231e-02
4.46080528e-02 -4.15815771e-01 -3.81931394e-01 1.93420932e-01
5.37067235e-01 7.53734112e-01 3.06837291e-01 -1.10572100e+00
-6.97871089e-01 8.67869556e-01 5.80205023e-01 -3.14099155e-02
8.92857611e-01 -7.37481356e-01 7.14291353e-03 -5.39011173e-02
6.01302326e-01 4.79758501e-01 -6.02865994e-01 2.05040276e-01
-1.97470374e-02 -6.14147007e-01 2.88460761e-01 -2.87679046e-01
-7.59041309e-01 4.31243896e-01 4.16873693e-01 7.02376962e-01
1.26290715e+00 -4.65057880e-01 7.19012260e-01 6.30840182e-01
1.33810592e+00 -6.72610879e-01 -7.12845206e-01 5.35798848e-01
5.54461181e-01 -7.40818381e-01 -5.44268787e-02 -4.49430078e-01
-6.27043664e-01 1.16868389e+00 9.49859917e-02 2.90432185e-01
8.37505400e-01 -2.77745184e-02 -1.54215433e-02 1.36266407e-02
-5.25756963e-02 -4.61889952e-01 -1.24635026e-02 7.67753422e-01
-1.91340894e-02 1.46068543e-01 -3.27022642e-01 4.49921876e-01
-2.18140587e-01 -4.01248038e-01 7.36750245e-01 1.05767858e+00
-5.67445397e-01 -8.32067311e-01 -7.40641177e-01 2.82364935e-01
-3.95186782e-01 9.28484276e-02 -2.66771972e-01 8.68674099e-01
2.59088099e-01 1.58398390e+00 -3.59782308e-01 -6.57855988e-01
7.96874464e-01 -5.78432143e-01 1.21777551e-02 -2.79136777e-01
-3.29772495e-02 -7.45765641e-02 1.63946912e-01 -4.17552531e-01
-3.49379063e-01 -7.21410215e-01 -1.05539882e+00 -2.70027876e-01
-5.22849500e-01 5.71958721e-01 9.29369152e-01 9.17964935e-01
3.99506807e-01 2.83153862e-01 9.78766799e-01 -9.79537606e-01
-2.10686713e-01 -4.73416001e-01 -9.24881339e-01 -9.54923570e-01
3.51096988e-01 -8.98755610e-01 -3.75862300e-01 -6.49950087e-01] | [6.139701843261719, 1.269937515258789] |
f9fd7ed9-a227-4973-bc30-1c3fd31ee9f3 | pay-attention-to-the-activations-a-modular | 1907.13075 | null | https://arxiv.org/abs/1907.13075v1 | https://arxiv.org/pdf/1907.13075v1.pdf | Pay attention to the activations: a modular attention mechanism for fine-grained image recognition | Fine-grained image recognition is central to many multimedia tasks such as search, retrieval and captioning. Unfortunately, these tasks are still challenging since the appearance of samples of the same class can be more different than those from different classes. Attention has been typically implemented in neural networks by selecting the most informative regions of the image that improve classification. In contrast, in this paper, attention is not applied at the image level but to the convolutional feature activations. In essence, with our approach, the neural model learns to attend to lower-level feature activations without requiring part annotations and uses those activations to update and rectify the output likelihood distribution. The proposed mechanism is modular, architecture-independent and efficient in terms of both parameters and computation required. Experiments demonstrate that well-known networks such as Wide Residual Networks and ResNeXt, when augmented with our approach, systematically improve their classification accuracy and become more robust to changes in deformation and pose and to the presence of clutter. As a result, our proposal reaches state-of-the-art classification accuracies in CIFAR-10, the Adience gender recognition task, Stanford Dogs, and UEC-Food100 while obtaining competitive performance in ImageNet, CIFAR-100, CUB200 Birds, and Stanford Cars. In addition, we analyze the different components of our model, showing that the proposed attention modules succeed in finding the most discriminative regions of the image. Finally, as a proof of concept, we demonstrate that with only local predictions, an augmented neural network can successfully classify an image before reaching any fully connected layer, thus reducing the computational amount up to 10%. | ['Jordi Gonzàlez Sabaté', 'Josep M. Gonfaus', 'Guillem Cucurull Preixens', 'Pau Rodríguez López', 'F. Xavier Roca Marva', 'Diego Velazquez Dorta'] | 2019-07-30 | null | null | null | null | ['fine-grained-image-recognition'] | ['computer-vision'] | [ 2.58996099e-01 -8.60152021e-02 4.64713143e-04 -3.08976829e-01
-5.95385313e-01 -5.46107531e-01 4.35457438e-01 4.37617376e-02
-7.66053081e-01 6.21829271e-01 -2.15877399e-01 1.08427823e-01
1.95985846e-02 -6.82732463e-01 -1.21847272e+00 -6.16029263e-01
1.06432550e-01 4.08356518e-01 3.62125278e-01 -1.97688162e-01
9.25626606e-02 8.39379191e-01 -1.78829658e+00 4.29163247e-01
5.99569619e-01 1.36567712e+00 2.92886555e-01 4.74796921e-01
1.60904795e-01 6.47169650e-01 -5.81448972e-01 -4.50271517e-01
1.52446628e-01 1.02435779e-02 -7.66555667e-01 2.43134931e-01
8.67165446e-01 -2.27715805e-01 -2.87410378e-01 1.14273226e+00
2.91745842e-01 1.58656061e-01 7.40410149e-01 -9.33573961e-01
-7.69320548e-01 3.98917496e-01 -5.07101893e-01 1.51849926e-01
3.64268906e-02 1.62878871e-01 9.38091695e-01 -1.16809928e+00
5.84307730e-01 1.11979747e+00 4.25003797e-01 5.48470080e-01
-1.29998004e+00 -5.75433910e-01 6.28079653e-01 4.97368515e-01
-1.35836005e+00 -3.49477470e-01 7.50626564e-01 -4.09540862e-01
7.58834958e-01 2.47077629e-01 4.29480672e-01 1.09653544e+00
3.44218500e-02 7.10930645e-01 8.36010218e-01 -4.20954287e-01
1.15836956e-01 2.85083681e-01 1.79116040e-01 6.50229454e-01
3.07875216e-01 -6.46692216e-02 -1.41709581e-01 2.61509001e-01
5.95713139e-01 1.98681846e-01 -3.29824269e-01 -6.20595276e-01
-1.08518231e+00 6.58519149e-01 9.77151752e-01 5.30639708e-01
-4.77143079e-01 3.42431426e-01 1.21362381e-01 -5.02014905e-03
3.89563471e-01 4.19008434e-01 -5.24150312e-01 3.71638834e-01
-7.80782521e-01 8.72408673e-02 5.32202840e-01 7.48064101e-01
9.41293597e-01 1.12560391e-01 -1.52145013e-01 8.68107915e-01
1.53769404e-01 3.84942800e-01 6.01516187e-01 -6.03303730e-01
2.92556018e-01 5.97186565e-01 1.97471902e-02 -1.14662230e+00
-5.24323344e-01 -1.01704788e+00 -1.02838087e+00 2.75407344e-01
3.96988958e-01 1.90116346e-01 -1.13136351e+00 1.88285697e+00
1.24841198e-01 1.04971863e-01 1.76416393e-02 1.03033745e+00
7.82815278e-01 4.23633814e-01 2.40439266e-01 1.62590265e-01
1.47356343e+00 -1.11745918e+00 -2.38421783e-01 -6.71933651e-01
2.40749761e-01 -5.80670416e-01 1.02440512e+00 2.81571925e-01
-9.26396668e-01 -9.19305861e-01 -1.07653606e+00 -1.93005856e-02
-6.60985768e-01 4.06194568e-01 5.11529207e-01 3.70157570e-01
-1.04636300e+00 7.02614486e-01 -5.04862368e-01 -3.68546486e-01
5.52563250e-01 4.64682788e-01 -6.21417701e-01 -9.00821462e-02
-9.02186811e-01 7.77571023e-01 4.21594411e-01 3.77405554e-01
-8.84468377e-01 -4.89722311e-01 -7.57474542e-01 5.18157184e-01
2.87041217e-01 -5.93954980e-01 8.98875117e-01 -1.35464633e+00
-1.22360969e+00 8.85704517e-01 -2.33743321e-02 -7.45729029e-01
3.97926569e-01 -2.92695701e-01 -1.45359576e-01 1.60107493e-01
-1.89834088e-01 1.03986788e+00 1.04659545e+00 -1.15287411e+00
-4.64092255e-01 -4.22526419e-01 1.54984578e-01 -7.83882588e-02
-2.82141328e-01 -1.42026454e-01 -6.40174031e-01 -7.71300256e-01
-3.69273592e-03 -1.05479908e+00 -3.75648975e-01 3.70918252e-02
-1.05413251e-01 -1.55106574e-01 5.27033091e-01 -4.01421726e-01
7.41015375e-01 -2.21104193e+00 2.79254943e-01 2.10818708e-01
3.70881148e-02 2.56450832e-01 -4.85121459e-01 -7.84242526e-02
-2.07670853e-01 1.13010854e-01 -1.07752249e-01 -2.00717404e-01
-2.59665418e-02 1.28341228e-01 -1.25490606e-01 5.43915629e-01
4.18175876e-01 1.01718521e+00 -4.45255548e-01 -2.22208381e-01
4.62282896e-01 5.12426853e-01 -7.08721817e-01 3.62405553e-02
-2.58644283e-01 2.73044735e-01 -3.20136040e-01 4.33056951e-01
6.26410961e-01 -5.40086687e-01 3.19230594e-02 -5.56062341e-01
1.06130071e-01 -2.49249637e-01 -1.12372971e+00 1.58503044e+00
-5.63522160e-01 5.07535219e-01 2.91853473e-02 -1.34934473e+00
7.82011926e-01 -8.29465166e-02 1.42313018e-01 -8.68317187e-01
1.99143231e-01 2.01628372e-01 4.53668609e-02 -2.98433661e-01
5.00917017e-01 2.61947304e-01 -2.16689870e-01 -2.75280084e-02
3.08909237e-01 2.46466249e-01 1.65986761e-01 -7.39682317e-02
6.37553036e-01 1.02819279e-01 3.64632964e-01 -1.65190935e-01
7.63610303e-01 -1.74573645e-01 3.40860069e-01 9.95380044e-01
-5.43468259e-02 6.12275779e-01 1.44585818e-01 -4.70173120e-01
-7.32159019e-01 -7.08896995e-01 -1.24387234e-01 1.27278376e+00
3.01331252e-01 1.97286587e-02 -8.43039453e-01 -6.72904253e-01
-6.42442331e-02 5.40966153e-01 -9.18409765e-01 -2.54554272e-01
-6.84963584e-01 -6.11563563e-01 2.68952608e-01 5.72939396e-01
4.67903525e-01 -1.30468416e+00 -7.28343427e-01 1.46814540e-01
-4.96631265e-02 -1.17920792e+00 -3.16870332e-01 3.56622964e-01
-6.57180727e-01 -1.18194151e+00 -8.84345353e-01 -8.41695428e-01
7.95727551e-01 1.41474053e-01 1.00278842e+00 1.23611204e-01
-5.31868219e-01 3.56795698e-01 -3.18556666e-01 -2.38201637e-02
-3.69178951e-02 2.59950966e-01 -3.44255686e-01 2.83390611e-01
7.07104281e-02 -3.96704644e-01 -7.03958869e-01 4.20245975e-01
-8.83107901e-01 -6.21538386e-02 8.69266272e-01 1.10677779e+00
7.10745990e-01 -2.33499080e-01 5.60920417e-01 -8.28652382e-01
-2.49320716e-02 -3.09607744e-01 -6.61502481e-01 3.25802475e-01
-2.63003021e-01 2.38327727e-01 7.85384476e-01 -5.04293740e-01
-8.77970397e-01 4.11843866e-01 -2.64946342e-01 -6.06199801e-01
-5.06526291e-01 2.73990303e-01 -6.30360991e-02 -3.47803265e-01
6.49631917e-01 1.79277480e-01 -2.94128418e-01 -5.51098406e-01
4.18677628e-01 2.43152991e-01 5.43982685e-01 -3.58860046e-01
6.97248638e-01 3.51879805e-01 -3.19338329e-02 -8.07113290e-01
-8.75291586e-01 -3.89899164e-01 -6.62773728e-01 -1.59689754e-01
9.60926414e-01 -9.38106477e-01 -8.54056299e-01 4.56890672e-01
-1.14929938e+00 -9.13136378e-02 -2.05910772e-01 3.64956379e-01
-3.71155769e-01 6.05840981e-02 -4.50593919e-01 -5.01253784e-01
-2.09062368e-01 -1.37164533e+00 1.07661402e+00 4.04584676e-01
1.29844755e-01 -6.83931589e-01 -4.42942500e-01 2.92978257e-01
5.04583836e-01 1.33721158e-01 9.23030138e-01 -7.68056333e-01
-8.36276770e-01 -2.89454579e-01 -4.31975245e-01 3.17477345e-01
-1.39675230e-01 -1.54051542e-01 -1.15615869e+00 -3.81145537e-01
-3.15877944e-01 -3.29889923e-01 1.19181693e+00 3.79904985e-01
1.57396746e+00 -2.83132553e-01 -3.10357124e-01 6.60873532e-01
1.40231216e+00 6.19995147e-02 5.42321920e-01 3.07005823e-01
5.55136085e-01 6.82708621e-01 4.22738910e-01 2.04461321e-01
1.15359738e-01 9.70829546e-01 8.42925847e-01 -2.80270845e-01
-2.68130213e-01 2.00271164e-03 1.10198371e-01 1.80137947e-01
-9.51673985e-02 -2.78952152e-01 -4.77613479e-01 5.51598608e-01
-1.74868155e+00 -9.77878332e-01 1.83563337e-01 2.05013466e+00
4.77755606e-01 1.06612798e-02 -1.26488656e-01 -8.57030973e-02
6.15625083e-01 8.29261765e-02 -6.65164411e-01 -1.24861352e-01
-2.75841087e-01 3.88804495e-01 5.54403961e-01 3.18721861e-01
-1.32530749e+00 1.05262053e+00 5.51573563e+00 8.63296151e-01
-1.37885618e+00 1.57529321e-02 9.10396993e-01 -7.99321160e-02
1.25662629e-02 -4.69873399e-01 -6.27440035e-01 1.79816231e-01
5.78601003e-01 3.37084353e-01 4.14756030e-01 1.13845181e+00
-2.04427525e-01 -4.89293784e-03 -1.23272681e+00 9.97956455e-01
2.72068918e-01 -1.43372607e+00 9.46997032e-02 -9.28518772e-02
5.61921000e-01 1.57749340e-01 2.69351125e-01 4.84484166e-01
4.36982978e-03 -1.22739089e+00 1.05167067e+00 4.99101251e-01
6.93750262e-01 -6.40718400e-01 8.47024024e-01 2.39825219e-01
-1.03178632e+00 -1.92561626e-01 -4.85059202e-01 2.48368233e-01
-1.58570275e-01 2.02233404e-01 -6.50866210e-01 4.16370153e-01
9.39481735e-01 5.01735687e-01 -8.94078016e-01 1.00966668e+00
-1.29508495e-01 2.78605223e-01 -1.76355451e-01 -5.78447431e-02
2.96111465e-01 1.52331442e-01 2.69568115e-01 1.20208716e+00
2.39761218e-01 -1.13522686e-01 2.09026128e-01 8.36263120e-01
-3.45377475e-01 1.64476801e-02 -3.04688245e-01 2.34328702e-01
7.09429830e-02 1.34495676e+00 -7.10983872e-01 -2.27763593e-01
-2.84635961e-01 1.08315086e+00 5.71462154e-01 5.87746143e-01
-8.98331046e-01 -2.83601135e-01 6.06835425e-01 4.82188575e-02
8.87209654e-01 1.25605613e-01 3.41472658e-03 -1.20015824e+00
4.97589447e-02 -8.20999324e-01 2.73475915e-01 -6.99250817e-01
-1.09171188e+00 9.88642216e-01 -2.65769690e-01 -9.86383975e-01
-3.27222705e-01 -9.54136670e-01 -2.32440084e-01 7.25938499e-01
-1.63569212e+00 -1.19489050e+00 -5.04938543e-01 6.96650505e-01
5.81654906e-01 3.74650843e-02 7.40844727e-01 5.35053790e-01
-5.26329577e-01 7.40895510e-01 4.80138287e-02 1.95922747e-01
5.91055036e-01 -9.59836066e-01 1.79392800e-01 8.18516135e-01
4.33493763e-01 5.02602279e-01 5.63357055e-01 -1.76869243e-01
-1.07415032e+00 -1.17968500e+00 5.21863461e-01 4.81651090e-02
5.49293160e-01 -5.37509680e-01 -9.02386308e-01 4.91271883e-01
4.79491800e-02 4.94925946e-01 1.31274983e-01 3.84734347e-02
-5.18068552e-01 -4.15947527e-01 -1.03840363e+00 4.15254772e-01
9.73231494e-01 -4.63315338e-01 -3.69950682e-01 2.50245869e-01
5.52928269e-01 -3.15538675e-01 -5.44401586e-01 4.02848750e-01
6.65284872e-01 -9.68588650e-01 1.23760271e+00 -6.73654616e-01
2.04536468e-01 -3.30283850e-01 -3.19444060e-01 -1.21904933e+00
-5.14289677e-01 -4.22660261e-02 2.18954086e-01 1.08286405e+00
4.72609997e-01 -5.46964467e-01 8.27377796e-01 2.86336541e-01
-1.41833842e-01 -6.56453013e-01 -9.37343538e-01 -5.22198737e-01
5.50010949e-02 -3.22667092e-01 4.06964421e-01 4.61045861e-01
-6.03981376e-01 2.25193709e-01 -4.24345613e-01 2.98813730e-01
5.01185119e-01 3.51516664e-01 6.92865431e-01 -1.32655942e+00
-5.20425141e-01 -4.94986445e-01 -5.30462742e-01 -1.09093606e+00
3.83960903e-01 -8.41342688e-01 2.82982826e-01 -1.25651193e+00
2.22990364e-01 -5.68773150e-01 -5.43319881e-01 5.84348679e-01
-8.83480906e-02 6.29584670e-01 5.54943204e-01 1.37789831e-01
-6.22302055e-01 4.34853584e-01 1.03678071e+00 -4.45374101e-01
1.82257399e-01 5.47754914e-02 -6.37721896e-01 7.26923466e-01
5.93509614e-01 -3.50144744e-01 -6.52355179e-02 -4.07060266e-01
-3.60431969e-02 -2.17146039e-01 8.70598376e-01 -1.05841398e+00
1.19495369e-01 7.58366585e-02 8.31072092e-01 -4.99952376e-01
5.20719528e-01 -1.08961523e+00 2.32868806e-01 5.32635510e-01
-3.98745537e-01 -2.62605131e-01 4.00273561e-01 5.47020912e-01
-3.79284471e-01 -3.84716272e-01 9.81701314e-01 -1.12130143e-01
-9.90032434e-01 3.36837292e-01 -1.54742062e-01 -1.32225081e-01
1.04725909e+00 -1.31150499e-01 -4.58348125e-01 -1.65734336e-01
-8.65422964e-01 -2.99192294e-02 3.68124455e-01 4.14425611e-01
4.35266197e-01 -1.14127791e+00 -5.50212860e-01 3.06833953e-01
1.96372315e-01 -4.89264391e-02 6.65097535e-01 7.29989529e-01
-3.15777600e-01 6.01924658e-01 -3.12180758e-01 -7.68743992e-01
-1.19911289e+00 7.56045938e-01 4.70949203e-01 -3.91973704e-01
-3.08392853e-01 9.42538142e-01 8.24592054e-01 -1.60304248e-01
4.15610284e-01 -4.39707577e-01 -3.98575217e-01 4.26826440e-02
4.49287474e-01 -6.88274354e-02 2.56861180e-01 -7.72465169e-01
-5.92869341e-01 8.35228980e-01 -2.66001463e-01 3.80630583e-01
1.33662188e+00 3.21615897e-02 1.89412713e-01 2.82217599e-02
1.25787795e+00 -8.16500038e-02 -1.29898298e+00 -1.89129919e-01
-1.02240175e-01 -1.82692870e-01 -4.33684103e-02 -8.88437927e-01
-1.33675039e+00 9.67314899e-01 8.29073310e-01 -3.29239145e-02
1.24417424e+00 2.57533401e-01 2.76320219e-01 4.20095176e-01
3.62945229e-01 -6.95422053e-01 1.00855581e-01 4.44920778e-01
1.04733765e+00 -1.26518929e+00 -2.25735888e-01 -2.73252606e-01
-5.19480228e-01 1.09327888e+00 7.70241082e-01 -2.54919827e-01
4.46948022e-01 1.68842394e-02 -1.63624138e-01 4.43000793e-02
-7.04087317e-01 -2.82576799e-01 6.01704419e-01 4.97152269e-01
3.52498084e-01 -1.27639070e-01 1.41235799e-01 6.06079996e-01
1.09315269e-01 -2.40734622e-01 2.17241734e-01 4.28617209e-01
-4.51177716e-01 -8.27844381e-01 -3.38921487e-01 2.70935088e-01
-5.36243737e-01 -1.49109304e-01 -1.80620834e-01 8.81500065e-01
2.03844994e-01 6.91183627e-01 1.91900104e-01 -1.98493958e-01
4.38007265e-01 -9.64404941e-02 4.98519063e-01 -4.45937365e-01
-6.98198497e-01 1.07539952e-01 -9.23844427e-02 -7.50752211e-01
-5.27203023e-01 -4.94760960e-01 -9.28973436e-01 1.13612071e-01
-4.34346616e-01 9.53277275e-02 6.43854499e-01 8.53542566e-01
4.26244348e-01 7.57411361e-01 3.81712914e-01 -1.09811318e+00
-5.41441381e-01 -8.95230234e-01 -2.81079948e-01 4.20677572e-01
3.74451399e-01 -8.17635834e-01 -1.95130900e-01 3.00386958e-02] | [9.467645645141602, 2.1157848834991455] |
66ec84f7-dbc8-4bec-bb2a-d49368da256f | continuous-time-analog-filters-for-audio-edge | 2206.02639 | null | https://arxiv.org/abs/2206.02639v2 | https://arxiv.org/pdf/2206.02639v2.pdf | Continuous-Time Analog Filters for Audio Edge Intelligence: Review on Circuit Designs | Edge audio devices can reduce data bandwidth requirements by pre-processing input speech on the device before transmission to the cloud. As edge devices are required to ensure always-on operation, their stringent power constraints pose several design challenges and force IC designers to look for solutions that use low standby power. One promising bio-inspired approach is to combine the continuous-time analog filter channels with a small memory footprint deep neural network that is trained on edge tasks such as keyword spotting, thereby allowing all blocks to be embedded in an IC. This paper reviews the historical background of the continuous-time analog filter circuits that have been used as feature extractors for current edge audio devices. Starting from the interpretation of a basic biquad filter as a two-integrator-loop topology, we introduce the progression in the design of second-order low-pass and band-pass filters ranging from OTA-based to source-follower-based architectures. We also derive and analyze the small-signal transfer function and discuss their usage in edge audio applications. | ['Shih-Chii Liu', 'Kwantae Kim'] | 2022-06-06 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 2.96126872e-01 -3.23615134e-01 -1.23698197e-01 -2.25002438e-01
-4.32005286e-01 -7.70922661e-01 -1.64774284e-01 -5.96514121e-02
-2.85012066e-01 3.76726985e-01 1.41549826e-01 -6.56848490e-01
-1.81147605e-01 -4.99648869e-01 -3.27636927e-01 -1.77954897e-01
1.43459663e-01 -2.41013616e-01 1.49997413e-01 -1.02709539e-01
-1.91441447e-01 5.52480042e-01 -1.42251194e+00 7.64775932e-01
4.73205119e-01 1.66278052e+00 2.80535351e-02 1.02004278e+00
3.56725335e-01 5.30283868e-01 -1.13137472e+00 -2.64911860e-01
9.60474759e-02 -5.20665109e-01 -1.81278482e-01 -4.64137971e-01
1.36333592e-02 -5.48805118e-01 -7.10611403e-01 8.16349149e-01
9.60296273e-01 -1.98691040e-01 7.36392081e-01 -1.21893895e+00
-2.53725290e-01 8.02686930e-01 5.73594198e-02 4.78116632e-01
6.00725532e-01 -3.30923535e-02 7.64360368e-01 -8.44600379e-01
1.45120218e-01 7.63355613e-01 1.11899197e+00 2.00025856e-01
-1.15230238e+00 -7.19538033e-01 -1.99840471e-01 5.73271632e-01
-1.48729563e+00 -8.04784894e-01 1.00833523e+00 -5.15683144e-02
1.68778217e+00 4.97135669e-01 1.09957242e+00 7.00295866e-01
3.28946114e-01 7.62951016e-01 6.55525863e-01 -4.66931701e-01
4.90250915e-01 2.38325968e-02 9.44013447e-02 2.59426653e-01
3.62219065e-02 -4.49940003e-02 -1.03033400e+00 -2.68984377e-01
7.69856513e-01 -9.02521238e-02 -6.05190039e-01 4.38329726e-01
-3.91783983e-01 2.69316941e-01 6.48654640e-01 3.34588110e-01
-5.15474558e-01 5.72735250e-01 5.58761299e-01 7.32299745e-01
1.82622835e-01 2.72265553e-01 -5.15125573e-01 -3.64501804e-01
-1.25557721e+00 -3.19489539e-01 7.96071112e-01 7.96912909e-01
-1.85237713e-02 5.71494877e-01 7.56653696e-02 2.77953148e-01
4.09793019e-01 1.67018995e-01 4.18137252e-01 -3.51072133e-01
3.18693109e-02 2.62529910e-01 -1.98165458e-02 -3.99506450e-01
-6.11145973e-01 -6.58228576e-01 -5.69633842e-01 2.60046363e-01
2.29728147e-01 -2.16756716e-01 -7.99990118e-01 1.28119457e+00
3.62722501e-02 -1.71289220e-02 -1.09962970e-01 9.38778639e-01
9.54937279e-01 7.40162790e-01 -3.55622739e-01 -1.58178911e-01
1.74946070e+00 -6.02328360e-01 -1.03779542e+00 -2.39831284e-01
-9.13272649e-02 -8.15142572e-01 9.97170687e-01 8.38773370e-01
-1.22796714e+00 -7.63649583e-01 -1.64208615e+00 -2.73947537e-01
-2.44401082e-01 3.95128995e-01 3.99310321e-01 1.29574788e+00
-1.17569423e+00 6.50427580e-01 -7.66310632e-01 3.79646897e-01
2.59128928e-01 8.22012544e-01 1.91308737e-01 8.54014933e-01
-1.18700910e+00 3.19865018e-01 3.10447700e-02 1.13011874e-01
-5.58493614e-01 -1.03088129e+00 -1.71951950e-01 6.76470399e-01
-6.76101297e-02 -8.78033757e-01 1.53648996e+00 -1.20122838e+00
-2.08319426e+00 3.83143783e-01 -1.89938754e-01 -7.04912424e-01
2.82456219e-01 -3.64810318e-01 -8.93842995e-01 3.82230550e-01
-7.08311677e-01 1.87813997e-01 1.34495187e+00 -1.86404735e-01
-8.62803400e-01 -8.51409808e-02 -8.71247351e-02 -3.78042251e-01
-6.83434129e-01 -3.60241160e-02 -1.74418390e-02 -6.49581015e-01
1.15553051e-01 -4.59358156e-01 1.36138886e-01 1.16803810e-01
-9.25763696e-02 1.02497779e-01 9.29527879e-01 -7.07970560e-01
1.75500047e+00 -2.42812490e+00 -2.15716213e-01 1.27926618e-01
1.29844680e-01 4.27188218e-01 4.76426721e-01 4.92735714e-01
-8.03155750e-02 -3.13691348e-01 4.01947647e-01 7.71498978e-02
-1.93449423e-01 -7.08384573e-01 -4.48834360e-01 5.45541883e-01
1.54734582e-01 7.02427864e-01 -3.28256160e-01 2.92521805e-01
3.48155320e-01 7.94087768e-01 -6.52806640e-01 -1.63181037e-01
8.67967978e-02 2.85403311e-01 -1.06043510e-01 6.57248437e-01
4.76071715e-01 -1.14033064e-02 1.30876610e-02 -8.06318700e-01
-1.38436481e-01 1.08124876e+00 -1.19967151e+00 1.64225447e+00
-9.17110443e-01 1.31380403e+00 7.03620970e-01 -5.22172034e-01
7.95236886e-01 9.72578049e-01 2.01869711e-01 -7.62505770e-01
6.18806958e-01 6.21611059e-01 6.43165931e-02 -2.70215571e-01
3.22150767e-01 -1.51093543e-01 -5.16390242e-02 1.38823748e-01
2.54264057e-01 -1.64436027e-01 -3.35911453e-01 -1.37750491e-01
1.07979476e+00 -2.48560518e-01 1.99352413e-01 -4.01484489e-01
5.35289824e-01 -4.62645084e-01 3.45020950e-01 3.13611627e-01
-1.51976809e-01 3.45426410e-01 8.63990411e-02 -4.48663354e-01
-8.25051069e-01 -1.13077533e+00 -7.88438395e-02 1.09108806e+00
-2.50070333e-01 -7.65754938e-01 -8.21218193e-01 1.20612174e-01
-2.31330976e-01 4.74883586e-01 1.68264806e-01 -4.65169370e-01
-4.57179368e-01 -1.71504721e-01 8.46314371e-01 7.21415818e-01
4.48711902e-01 -3.93665910e-01 -1.34211099e+00 6.39910281e-01
1.73428148e-01 -8.10578585e-01 -4.36701804e-01 9.56268191e-01
-8.16129684e-01 -3.10089260e-01 -5.22087753e-01 -8.20488274e-01
1.39025196e-01 -3.16272080e-01 8.88111711e-01 -3.40117455e-01
-6.55073762e-01 4.69161183e-01 -7.19890818e-02 -8.28896940e-01
-5.11568710e-02 2.49020644e-02 4.16941613e-01 1.07098809e-02
5.81553519e-01 -9.38795984e-01 -1.04389846e+00 1.64551198e-01
-4.78927910e-01 -2.48284757e-01 3.98011178e-01 6.30328834e-01
3.28125000e-01 9.04520750e-02 9.54230249e-01 -4.90376949e-02
8.60816956e-01 1.54126242e-01 -6.85337186e-01 4.74354289e-02
-1.87027380e-01 -1.84767276e-01 1.05855608e+00 -6.71669722e-01
-8.06760132e-01 5.82974017e-01 -7.09502876e-01 -3.73764962e-01
4.74220604e-01 1.96562827e-01 -4.31934655e-01 -2.26713240e-01
8.79864633e-01 -1.63749494e-02 -4.22739774e-01 -3.64182681e-01
1.47200078e-01 1.51843059e+00 7.01932251e-01 -1.10701416e-02
1.24912903e-01 5.54057002e-01 -1.16821803e-01 -1.18765020e+00
-6.94737956e-02 -3.26294541e-01 -3.96996103e-02 -4.11051214e-01
5.22414923e-01 -1.19459999e+00 -1.11371052e+00 3.41319919e-01
-1.21066582e+00 -1.23553187e-01 -2.95038998e-01 6.64083898e-01
-6.63227379e-01 -3.23701382e-01 -1.00726759e+00 -1.15200448e+00
-1.09985983e+00 -1.02601695e+00 9.05347824e-01 6.76737428e-01
-7.10740745e-01 -6.70295060e-01 -5.45780122e-01 -2.56286144e-01
6.05535924e-01 -2.87727386e-01 7.52264738e-01 -1.44503489e-01
-4.09059376e-01 -3.98835778e-01 3.71287316e-01 1.57751486e-01
9.76080522e-02 -2.09028542e-01 -1.79965532e+00 -3.16526711e-01
7.94580460e-01 1.83177814e-01 6.27520025e-01 8.05886984e-01
8.92386079e-01 -1.32923285e-02 -5.41573644e-01 9.07164335e-01
1.28213942e+00 6.20217025e-01 5.84344208e-01 -4.09535348e-01
1.13940842e-01 -1.23441428e-01 -7.08686262e-02 3.86976510e-01
-4.78489727e-01 3.60295802e-01 3.09760384e-02 -3.88933420e-02
-2.56926179e-01 -3.45346183e-01 4.89362448e-01 8.68316650e-01
2.53294975e-01 -4.10122216e-01 -3.13078672e-01 4.08739597e-01
-1.23096824e+00 -5.59033453e-01 -3.79920900e-01 2.24380970e+00
5.26275516e-01 3.93947661e-01 2.95393735e-01 8.75973940e-01
6.60933435e-01 -4.38411862e-01 -7.35850394e-01 -8.16401124e-01
8.97878110e-02 9.41087186e-01 6.76435113e-01 3.94948781e-01
-7.05388665e-01 4.21980590e-01 6.77513933e+00 8.67187381e-01
-1.84962058e+00 3.12397361e-01 4.86400962e-01 -4.42482650e-01
-4.38417010e-02 -3.08503091e-01 -7.25652993e-01 3.00336897e-01
1.50258493e+00 -3.46183270e-01 3.35914254e-01 9.41009939e-01
9.40540582e-02 9.22543630e-02 -1.59046292e+00 1.48833764e+00
-5.88220835e-01 -1.28655863e+00 -5.66788077e-01 -4.14837480e-01
2.63193935e-01 -3.45231652e-01 3.21645528e-01 -6.54160697e-03
-7.57686019e-01 -8.73179376e-01 1.24757314e+00 2.55632669e-01
1.39010382e+00 -8.70692909e-01 1.45963803e-01 1.27283067e-01
-1.77583718e+00 -4.61736023e-01 -2.22162515e-01 -8.06976378e-01
4.21489835e-01 8.50559533e-01 -8.13714564e-01 9.43735335e-03
7.82971025e-01 9.66614187e-02 1.79586783e-01 1.27933776e+00
-1.62772551e-01 8.05543005e-01 -8.60201180e-01 -5.34504354e-01
-1.33003220e-01 2.44477585e-01 4.32583034e-01 1.15144420e+00
4.28027838e-01 2.42078528e-01 -4.53991741e-01 9.02257621e-01
-1.16046160e-01 -3.68350953e-01 -1.95442945e-01 -1.22921705e-01
6.88203752e-01 1.27129257e+00 -7.87319899e-01 -2.78455019e-01
-5.82518637e-01 1.11535633e+00 -5.27370274e-01 1.19255573e-01
-8.18057716e-01 -1.23985147e+00 7.90520012e-01 5.12994111e-01
2.78419018e-01 -2.50762045e-01 -5.40851772e-01 -4.64368820e-01
2.89943159e-01 -7.69497633e-01 1.24511756e-01 -8.08981657e-01
-8.50409269e-01 9.00417030e-01 -6.31273270e-01 -1.46532798e+00
-3.01350802e-01 -7.97300041e-01 -7.34568298e-01 9.66125786e-01
-9.28336084e-01 -5.42753220e-01 1.15403742e-01 6.90322638e-01
4.97286499e-01 7.84360841e-02 7.34065771e-01 8.06228042e-01
2.50486970e-01 8.63675117e-01 2.40034788e-04 -2.69946381e-02
3.47995937e-01 -6.57255888e-01 7.45185852e-01 6.43800080e-01
1.69375822e-01 6.16589844e-01 6.91806734e-01 -5.57197690e-01
-1.89440560e+00 -5.27463973e-01 6.37486994e-01 3.26754451e-01
5.07100165e-01 -8.22004437e-01 -5.78815818e-01 1.78633198e-01
3.56674969e-01 -1.60363361e-01 6.28341615e-01 -4.15615529e-01
6.14561886e-02 -5.63593984e-01 -1.15214598e+00 7.57314384e-01
8.99786294e-01 -1.06435323e+00 -1.85618669e-01 -2.78152227e-02
6.12288833e-01 -4.71939981e-01 -4.90137607e-01 -1.96099970e-02
7.98292458e-01 -7.99022675e-01 8.32645655e-01 -1.86181486e-01
-1.38112709e-01 -4.66772556e-01 -5.82300685e-02 -1.27253258e+00
-3.20984095e-01 -1.41999793e+00 -2.17244312e-01 7.56901979e-01
5.81907392e-01 -7.26239562e-01 9.53837633e-01 3.00182223e-01
-3.93742681e-01 -6.20783508e-01 -1.44996727e+00 -9.08507645e-01
-3.88137758e-01 -5.05027413e-01 3.78654152e-01 3.85710374e-02
6.84082091e-01 6.89633012e-01 -2.13812828e-01 1.72722667e-01
-4.21637669e-02 -3.27396840e-01 -3.18244025e-02 -1.09542263e+00
-6.31830573e-01 -4.72669661e-01 -6.96876526e-01 -1.50746429e+00
-5.38519800e-01 -5.65852821e-01 -1.40164539e-01 -1.10264456e+00
-7.09173322e-01 6.64300546e-02 -6.27329201e-02 -1.77675977e-01
2.18963489e-01 3.44110489e-01 -9.87550095e-02 -4.65105176e-01
1.33896649e-01 1.18354306e-01 4.31814313e-01 -2.04556495e-01
-5.67198336e-01 4.29655761e-01 -2.05847278e-01 8.31068277e-01
5.56458533e-01 -3.07864845e-01 -5.60522556e-01 -4.38703328e-01
6.34029806e-01 3.77084076e-01 5.54184914e-02 -1.74491644e+00
8.66305709e-01 6.56250119e-01 5.65277874e-01 -6.10073209e-01
7.36357272e-01 -1.37375700e+00 5.02862513e-01 8.73316824e-01
8.00938159e-02 1.66065142e-01 4.07693416e-01 1.41947016e-01
-2.76774466e-01 -1.88657612e-01 6.82907641e-01 2.69496650e-01
-4.01940286e-01 -2.14802191e-01 -9.33821142e-01 -3.03592503e-01
6.58086479e-01 -4.63631958e-01 6.69168234e-02 -6.93806589e-01
-8.92422438e-01 -3.36954147e-01 -1.95514575e-01 1.59705766e-02
7.14667678e-01 -1.07935345e+00 -1.75141558e-01 5.88372886e-01
-5.21090567e-01 -4.99989510e-01 4.20991957e-01 5.51891267e-01
-6.64059043e-01 6.89187109e-01 -4.33805972e-01 -7.81162381e-01
-1.33504319e+00 5.06140709e-01 8.78980100e-01 3.71720821e-01
-5.73123753e-01 1.24871624e+00 -3.26440006e-01 7.77949393e-01
5.09299874e-01 -6.84750676e-01 4.49201733e-01 1.28125176e-01
9.45156693e-01 5.42347968e-01 5.94525933e-01 2.20353603e-01
-6.16859376e-01 5.84893346e-01 3.28216106e-01 -4.60995525e-01
1.20734084e+00 9.47650373e-02 3.26055229e-01 5.14734030e-01
1.23459756e+00 8.36104900e-02 -9.46341753e-01 1.46777540e-01
-2.85290301e-01 -1.35111243e-01 4.87044960e-01 -7.72095323e-01
-1.12768292e+00 1.25778949e+00 1.03785849e+00 3.30655903e-01
1.84989059e+00 -3.02544713e-01 9.53404963e-01 3.09499949e-01
4.06167179e-01 -1.48031354e+00 2.67132185e-02 1.28461689e-01
6.73009336e-01 -2.42139906e-01 -7.95830861e-02 -5.56372523e-01
1.55347824e-01 1.49393833e+00 1.82569027e-02 -2.90734440e-01
1.00651217e+00 1.09646487e+00 -4.83867116e-02 2.38865927e-01
-7.46918261e-01 9.08249766e-02 3.34426850e-01 6.09564424e-01
6.82901323e-01 -6.59203008e-02 -4.42399561e-01 1.24869239e+00
-3.99742573e-01 5.28534114e-01 3.74976903e-01 8.70499849e-01
-4.20237988e-01 -8.80951643e-01 -4.46227878e-01 5.59821427e-01
-7.96289206e-01 -6.17856048e-02 -6.63628459e-01 -6.46515936e-03
-6.09317757e-02 1.24271536e+00 5.12095451e-01 -6.68738723e-01
3.70682299e-01 2.32026592e-01 6.15660429e-01 2.48081684e-02
-1.29772246e+00 4.71569359e-01 3.17863002e-02 -3.21433067e-01
4.46646541e-01 -1.03512011e-01 -1.27227044e+00 -2.51789302e-01
-6.52878106e-01 -1.59252435e-01 1.10391796e+00 5.10477126e-01
6.26652479e-01 1.24192107e+00 2.96410233e-01 -7.12611556e-01
-3.13423604e-01 -6.81965590e-01 -5.37293673e-01 -4.55251098e-01
7.93496370e-01 -1.47970065e-01 -2.47923270e-01 -3.78981680e-02] | [14.573966979980469, 5.573640823364258] |
57988744-fb62-42a1-a5ce-3b548c01e59d | explanations-from-large-language-models-make | 2210.06726 | null | https://arxiv.org/abs/2210.06726v1 | https://arxiv.org/pdf/2210.06726v1.pdf | Explanations from Large Language Models Make Small Reasoners Better | Integrating free-text explanations to in-context learning of large language models (LLM) is shown to elicit strong reasoning capabilities along with reasonable explanations. In this paper, we consider the problem of leveraging the explanations generated by LLM to improve the training of small reasoners, which are more favorable in real-production deployment due to their low cost. We systematically explore three explanation generation approaches from LLM and utilize a multi-task learning framework to facilitate small models to acquire strong reasoning power together with explanation generation capabilities. Experiments on multiple reasoning tasks show that our method can consistently and significantly outperform finetuning baselines across different settings, and even perform better than finetuning/prompting a 60x larger GPT-3 (175B) model by up to 9.5% in accuracy. As a side benefit, human evaluation further shows that our method can generate high-quality explanations to justify its predictions, moving towards the goal of explainable AI. | ['Xifeng Yan', 'Wenhu Chen', 'Yi Mao', 'Baolin Peng', 'Jing Qian', 'Hong Wang', 'Zekun Li', 'Xinlu Zhang', 'Zhiyu Chen', 'Yelong Shen', 'Jianshu Chen', 'Shiyang Li'] | 2022-10-13 | null | null | null | null | ['explanation-generation'] | ['natural-language-processing'] | [ 3.95597458e-01 1.03880680e+00 -4.29004192e-01 -5.22895992e-01
-1.20093846e+00 -3.78997266e-01 8.60685349e-01 -4.42489348e-02
8.34476575e-02 9.47822392e-01 6.72571719e-01 -9.90231216e-01
-1.09350756e-01 -4.61359382e-01 -1.04532301e+00 9.19408947e-02
1.81604475e-01 9.82311487e-01 -1.55825809e-01 -2.96562225e-01
3.70103896e-01 -5.36160581e-02 -1.34525692e+00 9.49802756e-01
1.33067441e+00 5.37070572e-01 1.56162575e-01 8.52457106e-01
-3.63550514e-01 1.17416596e+00 -5.65918505e-01 -6.69488609e-01
-2.52503101e-02 -2.91031629e-01 -1.24848211e+00 -1.61097899e-01
7.91417837e-01 -4.71118152e-01 2.46971130e-01 3.68048102e-01
2.26295758e-02 1.66496828e-01 5.99400461e-01 -1.38644862e+00
-1.12260258e+00 1.42389297e+00 -3.18370521e-01 -1.25819352e-02
5.48605144e-01 4.00475323e-01 1.48411429e+00 -8.50291848e-01
4.03074712e-01 1.55664325e+00 6.73056245e-01 9.92599726e-01
-1.22172773e+00 -5.93812466e-01 4.06318784e-01 7.18604550e-02
-5.53599179e-01 -3.94105256e-01 2.53683448e-01 8.24388117e-03
1.67987704e+00 4.55087990e-01 3.72129381e-01 1.45263612e+00
1.81016073e-01 9.44092810e-01 9.17181253e-01 -5.50363839e-01
1.88412830e-01 7.11728707e-02 3.46356601e-01 8.39608312e-01
5.64889312e-01 5.26408106e-02 -7.29912996e-01 -2.66546696e-01
5.31672537e-01 3.81889101e-03 -2.05147229e-02 -6.26343414e-02
-1.33764708e+00 1.02933073e+00 6.81742013e-01 -8.95633828e-03
-4.04753774e-01 5.07845223e-01 9.41973478e-02 1.39894322e-01
2.66381532e-01 1.31688237e+00 -9.37123299e-01 -2.33813167e-01
-6.36317432e-01 6.50512278e-01 8.52163911e-01 1.20682955e+00
5.14113426e-01 2.56988347e-01 -3.92889559e-01 5.84299505e-01
3.29263240e-01 6.01832092e-01 6.39743030e-01 -1.29388261e+00
8.89658272e-01 8.16610456e-01 1.14510342e-01 -4.66259688e-01
-4.46143121e-01 -3.81532401e-01 -4.77309555e-01 3.46768387e-02
3.92038703e-01 -3.26709002e-01 -7.61283338e-01 1.79298019e+00
1.14038484e-02 5.90700693e-02 5.60311437e-01 6.62955105e-01
7.48436809e-01 5.97263396e-01 3.85640532e-01 3.11475396e-01
1.22006285e+00 -1.61370790e+00 -1.53131053e-01 -1.03889000e+00
9.59244311e-01 -3.88800979e-01 1.61379421e+00 3.06935042e-01
-1.20384824e+00 -5.81834733e-01 -8.76000762e-01 -2.54444897e-01
-4.74909656e-02 6.72254786e-02 1.32991993e+00 2.05378339e-01
-9.57180917e-01 4.89186913e-01 -6.31407678e-01 -3.67105722e-01
5.18933654e-01 3.22368950e-01 -1.39019206e-01 -2.36849353e-01
-1.03984761e+00 1.09931409e+00 3.39475572e-01 -4.04834598e-01
-7.71359622e-01 -1.04954517e+00 -9.96836007e-01 3.52117360e-01
5.39422154e-01 -1.19465601e+00 1.73926938e+00 -7.49692619e-01
-1.14944530e+00 3.83930266e-01 -3.43849212e-01 -8.91300797e-01
4.36529309e-01 -5.65547705e-01 -4.39580530e-02 -3.16188782e-01
2.72889078e-01 1.20554852e+00 3.47896367e-01 -1.08917034e+00
-5.92606902e-01 1.28947064e-01 6.42825603e-01 1.40240759e-01
-6.81180507e-02 -3.64632875e-01 -3.38522010e-02 -3.18031549e-01
-3.54526602e-02 -1.09204650e+00 -4.93829668e-01 -5.79812765e-01
-6.43507600e-01 -4.56392169e-01 2.22074986e-01 -4.71496046e-01
7.69798160e-01 -1.52114654e+00 -3.94648500e-02 -1.15035444e-01
3.32451880e-01 1.49644643e-01 -4.61142629e-01 2.69495100e-01
4.14563948e-03 4.65153843e-01 -5.68456315e-02 -5.03796875e-01
3.71569067e-01 3.59064728e-01 -8.35169792e-01 -5.44692755e-01
4.29270029e-01 1.48355365e+00 -9.27561462e-01 -3.83720666e-01
4.83229905e-02 1.48519248e-01 -1.11152470e+00 3.15666348e-01
-9.32597756e-01 3.14184368e-01 -5.94982982e-01 5.25541782e-01
1.10469639e-01 -8.21465254e-01 2.64831245e-01 3.97400945e-01
3.65996838e-01 8.12115788e-01 -5.87762356e-01 1.69556403e+00
-9.20556188e-01 5.65308630e-01 -5.97732604e-01 -4.42319274e-01
7.08893001e-01 1.82337359e-01 -3.64819050e-01 -5.82002938e-01
-1.91361427e-01 3.50593925e-01 2.66817719e-01 -3.82762641e-01
5.78570187e-01 -1.43895775e-01 -7.30056018e-02 7.44578660e-01
-1.39971644e-01 -2.86944985e-01 1.65396065e-01 5.28866947e-01
1.17753112e+00 1.72284722e-01 3.85562211e-01 -2.17128426e-01
2.10069314e-01 3.12228113e-01 2.28227884e-01 1.23506999e+00
3.12931299e-01 3.52741271e-01 4.71950889e-01 -6.54717326e-01
-9.32468176e-01 -6.09187007e-01 2.31880084e-01 1.17285502e+00
-1.28660023e-01 -6.42288148e-01 -5.96003890e-01 -1.08141208e+00
2.82925963e-01 1.56145155e+00 -5.89448154e-01 -1.13318421e-01
-4.76439476e-01 -3.57687771e-01 4.63922471e-01 9.72268581e-01
3.43017817e-01 -1.47770274e+00 -8.43833387e-01 8.37390050e-02
-3.52709025e-01 -1.13640118e+00 -2.51360387e-01 3.19173127e-01
-9.84177351e-01 -8.30254555e-01 -1.98230430e-01 -3.66319239e-01
6.94132328e-01 2.89619863e-01 1.63640213e+00 7.52351284e-01
2.17077047e-01 1.35136798e-01 -2.78149337e-01 -5.67740679e-01
-6.97284222e-01 4.33668762e-01 -9.07237679e-02 -9.68002081e-01
3.24888170e-01 -5.35421491e-01 -2.21224159e-01 1.81196183e-01
-5.70481598e-01 7.67769992e-01 9.26475346e-01 9.05750692e-01
2.17984051e-01 -3.33141506e-01 6.11038148e-01 -1.45645523e+00
7.82674313e-01 -4.17686731e-01 -2.72870332e-01 5.45476139e-01
-1.20074296e+00 6.91603959e-01 7.71857381e-01 -3.95610094e-01
-1.30457544e+00 -3.69620860e-01 7.78819025e-02 1.75486635e-02
-1.23967849e-01 4.35341746e-01 1.92676038e-02 3.37883979e-01
9.77631330e-01 -2.21957982e-01 -3.71270120e-01 -1.43128201e-01
8.85455787e-01 1.65404886e-01 4.92056221e-01 -1.04456902e+00
8.66880119e-01 3.01879682e-02 -1.69140175e-01 6.44087568e-02
-1.42380059e+00 1.41938135e-01 -1.28091022e-01 3.58856112e-01
5.43410540e-01 -9.12126064e-01 -6.79398656e-01 -5.12961507e-01
-1.28672183e+00 -9.07555759e-01 -2.22368374e-01 3.17928404e-01
-6.10099256e-01 -1.32880554e-01 -6.08379722e-01 -6.70770407e-01
-5.09036541e-01 -1.17142344e+00 1.14465880e+00 2.99033076e-01
-9.71737266e-01 -1.18481112e+00 -1.42225608e-01 8.76997709e-01
5.44963598e-01 -1.58301637e-01 1.27218318e+00 -1.07780516e+00
-8.62573624e-01 1.16905130e-01 -2.50949383e-01 -1.68214515e-01
-1.97776422e-01 -1.72295600e-01 -9.64672208e-01 1.60576046e-01
-4.82940197e-01 -8.30297053e-01 9.33672965e-01 2.42183357e-01
1.09993470e+00 -6.25001907e-01 -4.50815141e-01 4.71217632e-01
9.82424200e-01 -1.90307453e-01 4.04417843e-01 4.37577546e-01
6.80407941e-01 5.25381684e-01 7.66055286e-01 1.82703152e-01
5.39554000e-01 4.27172929e-01 3.22680324e-01 -7.96839073e-02
-3.26765329e-01 -7.98260272e-01 3.92698467e-01 2.65939862e-01
9.65320133e-03 -3.41329366e-01 -1.16014636e+00 3.66477996e-01
-2.11869311e+00 -9.40891445e-01 -2.49633178e-01 1.59844124e+00
8.22365940e-01 4.71569628e-01 -3.34162444e-01 -3.74264091e-01
5.04751727e-02 -4.48925011e-02 -7.73678422e-01 -8.38338614e-01
2.97880545e-02 5.56521975e-02 2.07090199e-01 9.02033269e-01
-4.77430522e-01 1.19525337e+00 6.86434555e+00 4.54696983e-01
-6.22840822e-01 -9.53290462e-02 7.40106106e-01 -8.53701830e-02
-1.03741455e+00 2.30112538e-01 -9.12981451e-01 -5.17953932e-02
1.14975011e+00 -1.52794778e-01 5.98727107e-01 1.27862120e+00
-6.08588494e-02 1.15665749e-01 -1.43657279e+00 5.41865647e-01
3.69128361e-02 -1.63626206e+00 5.09285629e-01 1.06521733e-01
1.18798780e+00 -1.19076185e-01 1.46949172e-01 9.68260229e-01
8.75170290e-01 -1.33675599e+00 7.25187480e-01 1.59277141e-01
5.64152896e-01 -5.15480340e-01 8.26988518e-01 6.67650998e-01
-7.28734791e-01 -3.85429621e-01 -5.76510787e-01 -6.92613125e-01
1.35752991e-01 3.48328769e-01 -1.52104878e+00 2.61744261e-01
1.41581938e-01 5.11457443e-01 -8.13591719e-01 3.97295445e-01
-9.48750019e-01 7.06212044e-01 9.41578373e-02 -2.19097659e-01
3.40222210e-01 4.19649333e-01 7.30271190e-02 1.03610122e+00
2.70810485e-01 2.85786480e-01 9.24121365e-02 1.35652471e+00
-3.69597286e-01 -1.46187112e-01 -4.48504359e-01 -1.03940524e-01
5.52166224e-01 1.29003263e+00 -3.57425034e-01 -7.55609930e-01
-2.92393267e-01 7.38689244e-01 7.36508191e-01 2.56530792e-01
-1.13541937e+00 3.64271104e-01 5.47302127e-01 -1.52228221e-01
5.95731214e-02 2.07213432e-01 -7.61453688e-01 -1.17903626e+00
-1.52959287e-01 -1.30441034e+00 3.08262855e-01 -1.29564917e+00
-1.23790920e+00 9.67271328e-01 1.06939346e-01 -4.68021423e-01
-8.96236658e-01 -6.63883626e-01 -8.64922523e-01 9.71653104e-01
-1.49656200e+00 -1.51125360e+00 -3.48162264e-01 3.26810032e-02
1.06661928e+00 -2.66081870e-01 9.23324883e-01 -3.80787790e-01
-3.38239700e-01 5.95381856e-01 -4.78583694e-01 -4.75846499e-01
5.83659470e-01 -1.66376138e+00 1.21484840e+00 7.63639569e-01
5.35089552e-01 1.10957682e+00 9.06300843e-01 -7.58185744e-01
-1.19769919e+00 -9.64446962e-01 1.31732392e+00 -1.18054700e+00
5.10522485e-01 -1.77235782e-01 -7.89679825e-01 1.25118971e+00
2.75903702e-01 -4.32113677e-01 4.93795693e-01 9.61645961e-01
-5.84231019e-01 3.81460816e-01 -8.32702100e-01 8.87323439e-01
1.20277798e+00 -3.39180857e-01 -1.12971413e+00 5.07313132e-01
1.10750926e+00 -5.85874856e-01 -3.18921983e-01 3.18842232e-01
5.41840672e-01 -9.95721102e-01 1.08775556e+00 -1.07430315e+00
1.26036143e+00 9.85766277e-02 -1.09214902e-01 -1.32982552e+00
-3.62902552e-01 -6.38424873e-01 -1.63807720e-01 1.19366276e+00
1.08785129e+00 -5.99992037e-01 7.78853178e-01 1.30314684e+00
-1.96439713e-01 -9.80169892e-01 -3.23443413e-01 -5.27590454e-01
2.02074602e-01 -5.95150173e-01 1.07232630e+00 5.80797911e-01
4.49913889e-01 7.59024918e-01 -3.07051629e-01 1.44105047e-01
4.11570907e-01 5.30413210e-01 1.11994720e+00 -1.10913050e+00
-7.03898668e-01 -3.34588677e-01 4.47724402e-01 -1.34085000e+00
6.81588650e-01 -1.05583501e+00 1.83185592e-01 -1.89961684e+00
5.75856447e-01 -5.25350213e-01 -8.95282254e-02 9.06209648e-01
-6.71369195e-01 -1.26121774e-01 2.65967786e-01 1.55486122e-01
-8.04703534e-01 2.47178927e-01 1.14455736e+00 -3.02520357e-02
7.67795667e-02 -1.73412748e-02 -1.53551364e+00 8.99187684e-01
6.81053281e-01 -3.90401095e-01 -8.48304927e-01 -8.08640540e-01
3.12481821e-01 1.21700265e-01 3.23056519e-01 -8.83498609e-01
-2.75433566e-02 -4.29805219e-01 3.13756704e-01 -1.70027167e-01
1.94815263e-01 -5.13870120e-01 -1.11330032e-01 3.67422551e-01
-1.11563897e+00 2.52038956e-01 4.35793132e-01 4.12634760e-01
2.40602776e-01 -1.77351415e-01 1.94369972e-01 -2.56976992e-01
-6.33530319e-01 -9.78243425e-02 1.28126098e-02 1.60767213e-01
4.46737736e-01 3.91583405e-02 -8.48680019e-01 -5.67482412e-01
-1.85566396e-01 4.72494125e-01 3.79421771e-01 4.91808176e-01
5.11835098e-01 -1.09827936e+00 -7.36008227e-01 2.77707912e-02
3.02975416e-01 -3.12856287e-02 1.32091224e-01 4.92041945e-01
-1.11833349e-01 1.03227615e+00 -2.80321613e-02 -3.58388186e-01
-1.15659630e+00 6.34597838e-01 5.61707802e-02 -7.14705348e-01
-7.62883902e-01 1.19266236e+00 3.96008313e-01 -6.78967774e-01
1.22826494e-01 -8.63581002e-01 1.47986099e-01 -7.55162060e-01
6.28709853e-01 3.02728899e-02 -2.65732348e-01 1.75663516e-01
-2.86662459e-01 1.72096100e-02 -1.34134501e-01 -1.47650436e-01
1.33922327e+00 1.59005150e-01 3.43312234e-01 2.10525274e-01
4.49498445e-01 2.86391973e-01 -1.40222299e+00 -9.52963382e-02
2.65597135e-01 -2.73244798e-01 -4.37283546e-01 -1.54816675e+00
-4.66394186e-01 8.08369696e-01 -3.52826238e-01 5.45579121e-02
6.25651538e-01 3.26185733e-01 7.78669596e-01 8.52336287e-01
3.44416142e-01 -3.94184589e-01 3.43571693e-01 4.63195384e-01
1.09233046e+00 -1.51691532e+00 4.50467505e-03 -2.03838959e-01
-1.12401903e+00 1.06508958e+00 1.13116169e+00 1.60150543e-01
-2.91653514e-01 1.40690967e-01 1.65949047e-01 -2.28920087e-01
-1.62497449e+00 1.68715701e-01 4.60127145e-01 3.98132712e-01
7.94327796e-01 1.78273797e-01 1.96188778e-01 1.00229800e+00
-6.79872811e-01 -1.74745262e-01 4.09031183e-01 3.35385144e-01
-5.18526673e-01 -1.20175564e+00 -2.29506001e-01 4.02937710e-01
-9.85168573e-03 -6.11068726e-01 -6.46666169e-01 9.54920590e-01
3.32550108e-02 1.16304255e+00 -3.36085796e-01 -1.56628996e-01
1.52393803e-01 3.08362097e-01 3.04772705e-01 -8.51969004e-01
-6.76174402e-01 -5.51828444e-01 4.63177621e-01 -7.11544096e-01
-5.11783548e-02 -5.21014512e-01 -1.71189189e+00 -5.10284841e-01
-2.79674172e-01 3.61143261e-01 2.50441283e-01 1.30511093e+00
4.72542852e-01 7.95833707e-01 -1.01278000e-01 -4.91560251e-01
-9.71680343e-01 -1.04248250e+00 1.32724131e-02 4.55204815e-01
2.00580299e-01 -4.18194234e-01 -3.54602873e-01 -1.10721298e-01] | [9.61400032043457, 7.002117156982422] |
e860652f-c29c-408e-b6a5-1b0c6211a101 | somtimes-self-organizing-maps-for-time-series | 2108.11523 | null | https://arxiv.org/abs/2108.11523v1 | https://arxiv.org/pdf/2108.11523v1.pdf | SOMTimeS: Self Organizing Maps for Time Series Clustering and its Application to Serious Illness Conversations | There is an increasing demand for scalable algorithms capable of clustering and analyzing large time series datasets. The Kohonen self-organizing map (SOM) is a type of unsupervised artificial neural network for visualizing and clustering complex data, reducing the dimensionality of data, and selecting influential features. Like all clustering methods, the SOM requires a measure of similarity between input data (in this work time series). Dynamic time warping (DTW) is one such measure, and a top performer given that it accommodates the distortions when aligning time series. Despite its use in clustering, DTW is limited in practice because it is quadratic in runtime complexity with the length of the time series data. To address this, we present a new DTW-based clustering method, called SOMTimeS (a Self-Organizing Map for TIME Series), that scales better and runs faster than other DTW-based clustering algorithms, and has similar performance accuracy. The computational performance of SOMTimeS stems from its ability to prune unnecessary DTW computations during the SOM's training phase. We also implemented a similar pruning strategy for K-means for comparison with one of the top performing clustering algorithms. We evaluated the pruning effectiveness, accuracy, execution time and scalability on 112 benchmark time series datasets from the University of California, Riverside classification archive. We showed that for similar accuracy, the speed-up achieved for SOMTimeS and K-means was 1.8x on average; however, rates varied between 1x and 18x depending on the dataset. SOMTimeS and K-means pruned 43% and 50% of the total DTW computations, respectively. We applied SOMtimeS to natural language conversation data collected as part of a large healthcare cohort study of patient-clinician serious illness conversations to demonstrate the algorithm's utility with complex, temporally sequenced phenomena. | ['Robert Gramling', 'Byung Suk Lee', 'Donna M. Rizzo', 'Ali Javed'] | 2021-08-26 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [ 8.04033205e-02 -1.23474620e-01 1.08930461e-01 -3.92476439e-01
-3.20226699e-01 -5.78374207e-01 4.12096083e-01 6.76667035e-01
-6.33413732e-01 1.71262875e-01 4.49787766e-01 -3.67428660e-01
-9.08972740e-01 -7.18278289e-01 2.85884328e-02 -9.39533949e-01
-6.65697515e-01 8.67051780e-01 2.53556609e-01 -4.44099456e-02
3.14656347e-01 5.61874747e-01 -1.82450926e+00 4.21889633e-01
7.68106878e-01 7.71235883e-01 1.00650517e-02 6.25171721e-01
-2.09425136e-01 6.06701314e-01 -6.85676157e-01 4.66987997e-01
2.60769248e-01 -3.52638334e-01 -4.74451214e-01 -9.61264130e-03
-4.07500386e-01 4.06006932e-01 -9.66416895e-02 4.66143280e-01
6.06974363e-01 6.39463186e-01 6.15750968e-01 -1.45436585e+00
4.25877748e-03 6.49320662e-01 -3.11644435e-01 5.98673284e-01
6.51918501e-02 -2.83100545e-01 4.92165715e-01 -6.09425783e-01
7.90527344e-01 9.07219350e-01 1.05021393e+00 3.46679658e-01
-1.38357735e+00 -3.98746073e-01 -3.94242942e-01 2.67722189e-01
-1.35979497e+00 -2.03530788e-01 7.76566923e-01 -5.15209734e-01
1.53024018e+00 7.07773626e-01 8.98666978e-01 6.76878810e-01
3.27158958e-01 3.74139637e-01 9.65374947e-01 -4.29187506e-01
8.33567083e-01 -1.65432289e-01 4.47795928e-01 -3.83938663e-02
-3.05810601e-01 -4.06488478e-02 -4.19921577e-01 -6.87162876e-01
2.81333119e-01 5.55516541e-01 4.47633006e-02 -2.69657522e-01
-1.39106464e+00 7.35317051e-01 -6.23055063e-02 6.83843195e-01
-5.21493256e-01 -2.46283680e-01 7.31732666e-01 5.43120027e-01
6.37761056e-01 5.60582221e-01 -3.66643965e-01 -7.04353034e-01
-1.19135118e+00 -5.01984693e-02 7.08115399e-01 5.44305563e-01
1.63030326e-01 1.55218625e-02 1.12118550e-01 1.01536155e+00
-3.50789160e-01 -1.38540484e-03 9.95025814e-01 -1.09572887e+00
8.75812098e-02 8.88387680e-01 -2.63880730e-01 -1.14514351e+00
-1.01082516e+00 -4.24402654e-02 -1.06705630e+00 9.80568677e-02
2.28834182e-01 9.23407227e-02 -1.06887794e+00 1.44229877e+00
2.96736985e-01 3.25611345e-02 9.81331170e-02 4.99641508e-01
2.73536533e-01 9.07179832e-01 -1.33588076e-01 -8.96972299e-01
1.09956419e+00 -4.63107318e-01 -7.77269483e-01 4.22371000e-01
6.46979094e-01 -6.62727773e-01 7.56951869e-01 5.11614323e-01
-9.39314246e-01 -2.70287335e-01 -9.26431835e-01 3.66662621e-01
-6.82395756e-01 -6.15513146e-01 4.81784403e-01 4.48252410e-01
-1.34141338e+00 1.00789964e+00 -1.46898389e+00 -8.77993703e-01
3.79853211e-02 5.46217859e-01 -2.02699959e-01 3.27763021e-01
-8.95357966e-01 8.16942275e-01 5.72592020e-01 -3.94801259e-01
-2.14568496e-01 -8.39522779e-01 -3.68456960e-01 1.59048468e-01
-1.25368461e-01 -1.25613809e-01 8.89936209e-01 -4.94255573e-01
-1.07695007e+00 5.37437916e-01 -3.39375138e-01 -6.62777424e-01
2.94808298e-01 4.73868668e-01 -8.63270521e-01 2.37684697e-01
1.57555584e-02 3.68049026e-01 6.28851771e-01 -8.00691187e-01
-4.63335514e-01 -5.18141329e-01 -7.50074148e-01 2.86376495e-02
-7.97786176e-01 1.33567601e-01 -1.30012199e-01 -8.30251396e-01
5.40137947e-01 -1.03871095e+00 -4.03206050e-01 -5.32745421e-01
-9.87050533e-02 -4.37429458e-01 1.22346163e+00 -4.65363711e-01
1.59730828e+00 -2.40375447e+00 4.48184423e-02 6.50564730e-01
4.18085128e-01 -7.53832757e-02 1.99965343e-01 9.75524724e-01
-3.94502312e-01 -2.29593534e-02 -5.26260614e-01 -1.71816230e-01
-2.92307377e-01 5.40869176e-01 -4.07434285e-01 3.92646074e-01
-2.24066526e-01 4.47076529e-01 -8.68031025e-01 -4.59538013e-01
3.32782596e-01 6.02047205e-01 -2.78319120e-01 -2.67080486e-01
3.63737196e-01 1.71566308e-01 1.47875145e-01 5.05781472e-01
1.48780495e-01 -1.42293960e-01 2.78649509e-01 2.01828793e-01
-4.83216047e-01 8.58885497e-02 -1.04614139e+00 1.29380977e+00
8.84966105e-02 8.98335397e-01 -4.61247087e-01 -1.00037241e+00
1.12599075e+00 6.24325573e-01 1.25362897e+00 -6.32017434e-01
1.92952063e-02 1.46057487e-01 1.03342436e-01 -4.53267068e-01
2.78033018e-01 -6.50230199e-02 1.07422397e-01 8.92485976e-01
-9.70223248e-02 4.00442444e-02 5.01749158e-01 5.49891731e-03
1.53991282e+00 -6.42346144e-01 2.06811637e-01 -7.27571845e-01
-5.16618090e-03 5.03314376e-01 6.50540709e-01 3.73021901e-01
-2.00316966e-01 6.39737427e-01 2.96508700e-01 -9.98922646e-01
-1.30795062e+00 -1.09536290e+00 -2.69576490e-01 1.20254326e+00
-2.66943395e-01 -8.00971508e-01 -6.73200130e-01 -2.31421720e-02
2.60394383e-02 7.95047462e-01 -7.73563385e-01 -4.92087305e-02
-6.59765780e-01 -1.12087977e+00 5.05116105e-01 5.86147130e-01
9.31222690e-04 -1.23073721e+00 -1.14259756e+00 6.15862727e-01
-2.24431470e-01 -4.95889574e-01 -2.27831334e-01 7.12804317e-01
-1.41038346e+00 -9.41382289e-01 -2.64554769e-01 -6.97194636e-01
8.05068612e-01 2.30397493e-01 8.57544303e-01 -1.67806089e-01
-5.63467920e-01 2.46702477e-01 -3.52230966e-01 -7.51013756e-01
-4.14684653e-01 -5.30429445e-02 5.15697241e-01 -3.20148528e-01
8.98207307e-01 -1.11213899e+00 -5.22747695e-01 4.49794829e-01
-9.99210775e-01 -1.53075472e-01 -7.55661055e-02 7.45088398e-01
6.15091562e-01 9.31183755e-01 2.60559529e-01 -5.73758304e-01
1.01524019e+00 -5.03452659e-01 -2.66507626e-01 -1.41739830e-01
-9.09125447e-01 -1.61245972e-01 7.50175595e-01 -7.82350957e-01
-4.31153327e-01 2.11395860e-01 3.80746871e-01 -6.11026525e-01
2.91499658e-03 6.02625489e-01 6.09440625e-01 3.68308961e-01
9.63028312e-01 3.95395875e-01 2.85202980e-01 -4.28898066e-01
-6.39520884e-02 8.00127804e-01 6.30354404e-01 -2.00388193e-01
4.80345398e-01 8.52956474e-01 -1.72747180e-01 -9.64994669e-01
2.46306825e-02 -9.27881896e-01 -8.00880730e-01 -2.65060931e-01
6.99984193e-01 -3.26407284e-01 -7.25431740e-01 1.29991904e-01
-7.30725110e-01 -1.82371169e-01 -5.85038126e-01 5.04975319e-01
-6.32829785e-01 1.08673707e-01 -5.23999393e-01 -8.98558259e-01
-4.53605592e-01 -6.06226027e-01 5.18921852e-01 -2.89166182e-01
-9.49453592e-01 -1.04241049e+00 3.70991498e-01 -9.68028903e-02
5.04159927e-01 8.06138098e-01 1.13002729e+00 -8.96079600e-01
3.79255295e-01 -2.30325043e-01 4.21712190e-01 -1.55129492e-01
2.34338418e-01 1.91917568e-01 -7.15900123e-01 -3.78898650e-01
4.92225617e-01 3.43479127e-01 7.19280183e-01 5.42319179e-01
1.10374916e+00 -4.08895791e-01 -4.96893346e-01 3.25977832e-01
1.12148643e+00 9.80600357e-01 5.44767439e-01 5.84706664e-01
3.88921082e-01 9.52729702e-01 4.63429958e-01 7.07354367e-01
1.46150574e-01 3.83521140e-01 -1.67984944e-02 -8.10450390e-02
5.43334723e-01 2.47554794e-01 5.94060011e-02 1.51568186e+00
-5.42912185e-01 8.76801535e-02 -1.52488685e+00 7.69969046e-01
-2.13693523e+00 -1.36164784e+00 -8.96233916e-02 2.18562245e+00
7.33445525e-01 2.39018828e-01 6.04855239e-01 9.87926006e-01
7.54836440e-01 -1.17439292e-01 -5.77718318e-01 -6.31599009e-01
-3.18471566e-02 4.27785814e-02 3.44010293e-01 1.31239250e-01
-1.06216800e+00 2.29456604e-01 6.68570995e+00 5.57862818e-01
-1.08412111e+00 1.24114245e-01 4.18668360e-01 -6.19426608e-01
1.77837864e-01 -4.35315877e-01 2.26652417e-02 6.18525207e-01
1.40010262e+00 -6.92511618e-01 6.48363829e-01 8.35358441e-01
6.56591296e-01 1.90173775e-01 -9.19213057e-01 1.16639030e+00
-1.74266934e-01 -1.32758665e+00 -3.43591273e-01 7.24953413e-02
6.26626551e-01 2.08768874e-01 -1.14720710e-01 4.18484546e-02
3.33787173e-01 -1.02800286e+00 3.60039830e-01 1.81701258e-01
3.61786395e-01 -1.10849786e+00 6.37851000e-01 3.69158536e-01
-1.25628817e+00 -3.99465173e-01 -2.98549503e-01 -2.53894538e-01
7.56838396e-02 7.31197000e-01 -9.91545379e-01 1.33032680e-01
1.29359949e+00 6.93294585e-01 -2.70359218e-01 9.61171389e-01
5.30630946e-01 6.58910751e-01 -7.51133084e-01 3.62137333e-02
3.66303891e-01 -1.52308688e-01 6.11189663e-01 1.35776436e+00
4.84045804e-01 4.15545374e-01 7.22395256e-02 3.82558674e-01
5.83518565e-01 -1.74704995e-02 -5.83399296e-01 -1.90558493e-01
9.85256791e-01 1.09302354e+00 -1.40169859e+00 -5.43610871e-01
1.18694462e-01 5.24674416e-01 -1.96601808e-01 2.38966510e-01
-5.75452030e-01 -6.60672307e-01 6.40718520e-01 3.63798887e-01
2.53464073e-01 -4.42319036e-01 -5.25579691e-01 -6.19067848e-01
-2.57937191e-03 -6.93469286e-01 8.72413754e-01 -5.92381299e-01
-1.29123712e+00 8.35772216e-01 1.14572033e-01 -1.70662045e+00
-6.47766948e-01 -2.36852199e-01 -5.67704856e-01 4.77059841e-01
-5.19662619e-01 -3.63881588e-01 -3.05666149e-01 8.95327091e-01
5.80828607e-01 -1.25119209e-01 1.14799702e+00 9.36159790e-02
-3.11178148e-01 1.58836991e-01 7.78660059e-01 -1.30551038e-02
4.59625423e-01 -1.32720196e+00 4.25357670e-01 6.01224005e-01
-1.29735202e-01 9.65578020e-01 8.88046145e-01 -6.14508569e-01
-1.26682687e+00 -1.16878879e+00 1.04122686e+00 -5.22803366e-01
7.27209210e-01 -3.70567471e-01 -1.11242533e+00 4.19448882e-01
4.40661088e-02 -4.46865976e-01 1.08898497e+00 1.09775931e-01
-2.38551214e-01 1.89879294e-02 -1.26976514e+00 5.80724001e-01
9.36269224e-01 -4.81439322e-01 -9.26135659e-01 3.76222670e-01
6.23589039e-01 -1.76923662e-01 -1.36943126e+00 2.59291351e-01
4.69362140e-01 -1.17421019e+00 9.66250718e-01 -3.98963362e-01
2.01842055e-01 -3.86571646e-01 8.52890406e-03 -1.33530819e+00
-6.74193084e-01 -8.26973796e-01 -1.87433898e-01 8.69432509e-01
1.61318421e-01 -6.82669163e-01 7.49713123e-01 6.05511367e-01
-1.42406449e-01 -7.13705897e-01 -1.09785855e+00 -1.08722520e+00
-3.02869320e-01 -6.14384174e-01 6.34557903e-01 1.60653126e+00
7.48638451e-01 -5.13496948e-03 6.47449046e-02 -1.63131490e-01
6.61451340e-01 9.54243541e-02 2.30870932e-01 -1.70469546e+00
5.39444461e-02 -5.92046678e-01 -7.71907449e-01 1.56873152e-01
-3.59228641e-01 -9.09594059e-01 -1.88263431e-01 -1.32686234e+00
-5.83491512e-02 -2.45149910e-01 -6.06708884e-01 8.58934760e-01
3.27505082e-01 8.70600268e-02 3.09325326e-02 8.10930371e-01
-3.36536348e-01 -3.26359761e-03 4.28510875e-01 -1.18496329e-01
-9.47966278e-01 -2.19663352e-01 -1.49659246e-01 5.85784435e-01
9.82347846e-01 -7.06368983e-01 -7.44649291e-01 -5.02906181e-02
-2.95575291e-01 8.15480500e-02 -1.38426855e-01 -1.11647093e+00
7.78204322e-01 -9.96233448e-02 4.65316147e-01 -1.05828619e+00
2.17895612e-01 -9.43362892e-01 7.81075895e-01 7.01596141e-01
-2.81759799e-01 6.04094028e-01 4.25091326e-01 4.13595349e-01
-2.79253751e-01 4.02611256e-01 6.17140055e-01 -2.66572572e-02
-5.97706079e-01 -4.50066701e-02 -8.90944958e-01 -2.91905463e-01
1.09772956e+00 -6.93764508e-01 -2.34705225e-01 -2.28595063e-01
-1.03669000e+00 1.74629256e-01 4.46985781e-01 4.59352255e-01
5.51994205e-01 -1.39179611e+00 -3.66702139e-01 3.14116716e-01
2.06229061e-01 -1.39412776e-01 8.14066753e-02 1.07516718e+00
-5.53137898e-01 4.23211962e-01 -5.05848825e-01 -8.96972597e-01
-1.52180183e+00 9.17182446e-01 -6.19278848e-02 -7.02686757e-02
-1.05317795e+00 2.79438645e-01 -3.92558157e-01 -4.46315885e-01
4.54735786e-01 -3.12398851e-01 -2.12541670e-01 3.61782402e-01
9.05611038e-01 9.22218919e-01 3.76994282e-01 -2.10454524e-01
-5.65222204e-01 4.96664017e-01 1.11972280e-01 -1.92803532e-01
1.82126975e+00 3.89741696e-02 -4.95674163e-01 9.62698400e-01
1.13161659e+00 -5.49802959e-01 -7.75890708e-01 8.73958766e-02
5.80282450e-01 -7.18295649e-02 -1.04417361e-01 -4.25173193e-01
-6.04213297e-01 5.45030534e-01 6.31731987e-01 9.62745965e-01
1.48795998e+00 -1.94924414e-01 6.77057445e-01 5.59234202e-01
2.90260792e-01 -1.41980696e+00 -2.66812056e-01 3.64005655e-01
5.67800820e-01 -7.11284280e-01 -1.00387134e-01 -9.05580148e-02
-4.79576558e-01 1.27505398e+00 3.50721516e-02 -5.48138954e-02
6.68926120e-01 6.51754081e-01 2.85591304e-01 -3.62074524e-01
-1.00314832e+00 3.88172358e-01 -6.76032156e-02 6.83647454e-01
2.33962998e-01 1.75031513e-01 -2.12670729e-01 2.45442525e-01
-7.33335972e-01 -8.19257461e-03 3.11958432e-01 1.30342388e+00
-4.55953479e-01 -6.70657635e-01 -5.17237306e-01 7.49980688e-01
-1.26933455e-01 1.25526175e-01 -5.33512712e-01 6.93393528e-01
-9.51464400e-02 1.11443090e+00 4.77481157e-01 -7.66710997e-01
3.49617451e-01 3.65224659e-01 -1.53896227e-01 -2.79746920e-01
-8.42217445e-01 2.44158626e-01 -1.51127428e-01 -7.05102801e-01
-5.70583761e-01 -7.99939632e-01 -1.60203362e+00 -7.70730257e-01
1.49070367e-01 4.38472003e-01 6.36103392e-01 6.11315966e-01
6.44098043e-01 4.16010290e-01 7.41957009e-01 -7.66233861e-01
3.38526326e-03 -8.14770937e-01 -7.27105975e-01 5.95372856e-01
1.11350164e-01 -4.61326569e-01 -5.88980794e-01 3.39815825e-01] | [7.252604007720947, 3.3931961059570312] |
9543a6f3-6cce-4b2a-9a58-9179ea7267d2 | global-bilateral-symmetry-detection-using | null | null | https://hal.archives-ouvertes.fr/ujm-01387193 | https://hal-ujm.archives-ouvertes.fr/ujm-01387193v2/document | Global Bilateral Symmetry Detection Using Multiscale Mirror Histograms | In recent years, there has been renewed interest in bilateral symmetry detection in images. It consists in detecting the main bilateral symmetry axis inside artificial or natural images. State-of-the-art methods combine feature point detection, pairwise comparison and voting in Hough-like space. In spite of their good performance, they fail to give reliable results over challenging real-world and artistic images. In this paper, we propose a novel symmetry detection method using multi-scale edge features combined with local orientation histograms. An experimental evaluation is conducted on public datasets plus a new aesthetic-oriented dataset. The results show that our approach outperforms all other concurrent methods. | ['Philippe Colantoni', 'Christophe Ducottet', 'Cécile Barat', 'Mohamed Elawady'] | 2016-10-01 | null | null | null | null | ['symmetry-detection'] | ['computer-vision'] | [ 1.39331609e-01 -5.25050163e-01 5.27717099e-02 -1.39373466e-01
-5.47862589e-01 -5.91975808e-01 7.52179325e-01 3.89102474e-02
-4.54093874e-01 4.56262946e-01 2.54790962e-01 1.17773287e-01
-1.99475124e-01 -6.36775732e-01 -2.99452931e-01 -4.78729963e-01
-9.84972492e-02 2.20468104e-01 8.27487350e-01 -2.10694477e-01
1.03992951e+00 9.74480152e-01 -1.79448378e+00 3.31169933e-01
4.23008800e-01 1.11402524e+00 -2.24361017e-01 5.22751331e-01
4.07079190e-01 3.00173849e-01 -3.16278040e-01 -5.12712061e-01
4.98387367e-01 -6.32170379e-01 -5.17020822e-01 3.39074165e-01
6.62466168e-01 -2.89528240e-02 2.10320652e-02 9.53661978e-01
8.39548886e-01 -2.01738968e-01 8.10130715e-01 -1.36533391e+00
-3.59763205e-01 -2.15779483e-01 -1.04284525e+00 2.98500746e-01
5.29647708e-01 -1.28351524e-01 1.16874325e+00 -1.06269836e+00
1.06931520e+00 8.36335719e-01 7.16222286e-01 -9.43809599e-02
-1.06858969e+00 -5.93955040e-01 -4.57697600e-01 6.09210432e-01
-1.63739550e+00 -2.34620586e-01 1.09704089e+00 -3.00105006e-01
9.34159935e-01 4.16098148e-01 7.53150105e-01 4.16101158e-01
3.96638125e-01 7.11790502e-01 1.44927227e+00 -7.89315641e-01
2.98735201e-02 -1.56893089e-01 -5.34525275e-01 7.91630328e-01
2.44883358e-01 1.69287533e-01 -6.70722187e-01 -8.63190144e-02
8.66288245e-01 -1.33474886e-01 -2.29913771e-01 -1.01045108e+00
-1.30938673e+00 7.20301747e-01 4.52810794e-01 5.41116893e-01
-2.99171150e-01 -3.69566649e-01 3.75967264e-01 4.49648499e-02
4.80183125e-01 3.92755032e-01 -2.03552619e-01 -1.81684747e-01
-1.12444973e+00 3.95012081e-01 4.92669255e-01 4.84435767e-01
4.47429210e-01 -4.96590346e-01 3.74502987e-02 5.86048663e-01
9.83792841e-02 5.37797660e-02 5.72203875e-01 -5.55985630e-01
2.43186459e-01 9.36836064e-01 -3.40465754e-02 -1.55943513e+00
-5.58460712e-01 -3.61390144e-01 -8.85462761e-01 5.62537253e-01
4.25201535e-01 4.06364441e-01 -6.05546892e-01 9.34185922e-01
2.84395784e-01 -5.83632514e-02 -2.06379905e-01 8.88954878e-01
5.87125123e-01 3.21712971e-01 -6.21493340e-01 1.25062257e-01
1.31078112e+00 -8.47040474e-01 -3.44722599e-01 -3.69782522e-02
1.78935856e-01 -1.37439156e+00 5.86846769e-01 6.49392366e-01
-1.02487993e+00 -3.43202859e-01 -1.06394005e+00 2.62899115e-03
-5.16512573e-01 2.61932135e-01 4.53038543e-01 6.45842254e-01
-8.97441566e-01 4.94400948e-01 -3.94639611e-01 -6.24633610e-01
1.66934863e-01 1.53687716e-01 -7.42955208e-01 8.75245109e-02
-5.92984974e-01 7.94473648e-01 3.02190721e-01 5.80806984e-03
1.65121421e-01 -2.45314941e-01 -9.09339666e-01 -7.70550445e-02
1.82256833e-01 -5.24826288e-01 8.36767495e-01 -5.30863583e-01
-1.28047824e+00 1.39588058e+00 -1.47517711e-01 -2.67547995e-01
9.77122307e-01 1.38099089e-01 -4.96244490e-01 2.87086904e-01
-4.24527489e-02 6.92092180e-01 7.66827464e-01 -1.28820133e+00
-8.78837824e-01 -4.98420715e-01 -4.61647540e-01 1.75881848e-01
-1.10346057e-01 3.16417336e-01 -3.66967857e-01 -7.75758564e-01
7.47007430e-01 -7.91754782e-01 1.31758854e-01 2.55053192e-01
-4.99780774e-01 -2.53075302e-01 1.11279142e+00 -4.90411341e-01
1.18096912e+00 -2.02621126e+00 -8.06003883e-02 6.04384542e-01
-2.61590071e-03 2.41024643e-02 1.98206231e-01 5.92912138e-01
-1.60443589e-01 -4.38882224e-02 -6.56731203e-02 -5.28815575e-02
2.38514948e-03 -2.11187571e-01 3.51475775e-02 8.66497517e-01
1.96276922e-02 8.33369076e-01 -5.41138470e-01 -8.60770047e-01
5.20524561e-01 4.12189066e-01 -4.36733723e-01 -1.04849875e-01
6.99110031e-01 1.89977929e-01 3.87247242e-02 7.68751562e-01
8.77928734e-01 -5.58719561e-02 -1.29979685e-01 -3.57564181e-01
-5.11571884e-01 -2.25542128e-01 -1.65227091e+00 1.44471121e+00
-2.68193036e-02 8.25175881e-01 -3.43949795e-01 -8.26165199e-01
9.46131706e-01 7.67755136e-02 3.86368096e-01 -9.24101830e-01
1.42099038e-01 4.58119839e-01 4.64906171e-02 -2.63001174e-01
5.92561901e-01 -1.71211526e-01 5.62860668e-02 2.56270438e-01
-2.42433578e-01 -5.38881481e-01 6.09273732e-01 -1.84905827e-01
6.90841794e-01 -9.72085260e-03 7.34375954e-01 -2.67487437e-01
9.80680406e-01 -3.09575588e-01 4.16567892e-01 2.93492585e-01
-2.82188207e-01 1.16005564e+00 4.90534335e-01 -6.17987514e-01
-1.25025713e+00 -1.18077278e+00 -3.05283695e-01 4.78202015e-01
3.78905565e-01 -6.30356789e-01 -5.53023696e-01 -6.06450081e-01
5.59207126e-02 1.15608111e-01 -6.98640645e-01 1.06992126e-01
-5.51317215e-01 -6.08068526e-01 1.27632856e-01 3.72838616e-01
9.05100405e-01 -1.03292775e+00 -1.09110558e+00 -1.71750516e-01
-9.73796397e-02 -1.06492865e+00 -6.06519520e-01 -3.39018375e-01
-7.80003309e-01 -1.27699554e+00 -1.05995011e+00 -9.80211198e-01
8.10596824e-01 4.78797436e-01 1.04688978e+00 -2.08426237e-01
-1.03013229e+00 2.45576799e-01 -3.17943603e-01 -1.07282870e-01
-1.71273556e-02 -1.51772335e-01 -2.58376390e-01 2.61830360e-01
2.37709567e-01 -4.04297113e-01 -8.95373702e-01 8.72985423e-01
-7.90277064e-01 -6.39431551e-02 9.67831135e-01 7.67102540e-01
5.06087065e-01 2.54900694e-01 -1.53272420e-01 -2.21920222e-01
5.70401132e-01 2.68835932e-01 -7.15098619e-01 2.80748904e-01
-4.96402055e-01 1.44668877e-01 1.39232129e-01 7.44560137e-02
-7.28690386e-01 2.05190182e-01 2.24684954e-01 5.98838143e-02
-1.59377620e-01 1.80456921e-01 1.48934633e-01 -6.44806564e-01
5.46045125e-01 4.32936221e-01 -3.69545296e-02 -1.53006524e-01
4.40704897e-02 6.41747892e-01 6.03566825e-01 -1.62759945e-02
9.18715358e-01 7.73749650e-01 4.12825018e-01 -9.69281256e-01
-2.67100215e-01 -9.71500278e-01 -7.83901334e-01 -3.18561316e-01
5.68963826e-01 -4.45256799e-01 -7.36880243e-01 6.58578515e-01
-1.05031252e+00 5.16745985e-01 -1.81708828e-01 4.46950793e-01
-7.75154412e-01 9.52269137e-01 -2.15501755e-01 -7.31824934e-01
-2.94168919e-01 -9.39879417e-01 1.32977414e+00 3.30766529e-01
-2.20012560e-01 -7.03568876e-01 2.16348931e-01 3.31188589e-01
4.71868485e-01 3.38167697e-01 3.08745652e-01 -1.97080880e-01
-4.85816479e-01 -7.86668718e-01 -6.40685081e-01 1.32652387e-01
6.70295879e-02 3.47061634e-01 -6.39489472e-01 -7.90375993e-02
-3.28878522e-01 -1.88176081e-01 7.59968162e-01 2.96006292e-01
1.03357577e+00 1.05149664e-01 -1.72326893e-01 4.24195886e-01
1.41120398e+00 -1.26068192e-02 8.68653178e-01 7.11792707e-01
9.27757770e-02 6.05564952e-01 6.69498324e-01 5.95819294e-01
9.68360603e-02 8.05928528e-01 3.24354023e-01 -4.07695860e-01
-6.74859360e-02 -1.13691792e-01 -8.04483984e-03 2.12384805e-01
-4.95142847e-01 1.93028729e-02 -8.91725123e-01 4.32162076e-01
-1.66762400e+00 -1.32167828e+00 -1.39343098e-01 2.41557479e+00
4.12859201e-01 3.49630773e-01 3.42893332e-01 7.32484281e-01
9.53420162e-01 2.15849876e-02 -1.94217470e-02 -2.77208447e-01
-4.88188773e-01 -3.43890488e-02 6.04106784e-01 1.34728819e-01
-1.46114433e+00 4.18427676e-01 6.70177460e+00 8.87644231e-01
-1.28775072e+00 -2.46139616e-01 5.41533589e-01 5.09343684e-01
1.10357091e-01 -1.70743823e-01 -5.82900882e-01 2.88225532e-01
4.89985347e-02 -2.97984749e-01 -2.06779882e-01 6.88515127e-01
-9.08389017e-02 -6.05314016e-01 -5.24287164e-01 1.35875583e+00
6.62002921e-01 -1.10949564e+00 -2.71829128e-01 2.52508670e-01
8.78875256e-01 -1.51804641e-01 2.11812273e-01 -4.83237743e-01
-3.08490694e-01 -7.32256472e-01 6.10799491e-01 2.38515466e-01
4.94385332e-01 -8.84722412e-01 8.93555045e-01 1.30362049e-01
-1.46430671e+00 5.25016226e-02 -9.83742997e-02 -1.61651701e-01
2.27320120e-01 4.47820783e-01 -8.42975140e-01 7.05100417e-01
7.50904441e-01 6.46739006e-01 -8.85256529e-01 2.00197768e+00
-3.33150148e-01 -1.14086151e-01 -5.12657166e-01 -5.22453189e-02
1.96082622e-01 -2.89336979e-01 3.53467315e-01 1.23737383e+00
4.20515001e-01 -2.72289991e-01 -1.31271154e-01 4.61296886e-01
3.34038168e-01 5.60459375e-01 -6.91310048e-01 2.28438079e-01
-2.12953076e-01 1.53305995e+00 -1.48956740e+00 -1.14559665e-01
-4.65042144e-01 1.34701419e+00 -1.42518505e-01 -1.32551372e-01
-6.64319932e-01 -6.50994718e-01 8.05369020e-02 2.99375355e-01
3.81467581e-01 -3.97980303e-01 -6.51094541e-02 -1.06128311e+00
3.75674963e-01 -4.57625508e-01 5.49782932e-01 -7.82831371e-01
-1.28910637e+00 4.17839170e-01 5.03231725e-03 -1.76606762e+00
-2.98571616e-01 -8.66379917e-01 -8.00141394e-01 4.91492242e-01
-1.28162265e+00 -1.14211023e+00 -4.93854284e-01 4.17962432e-01
3.85529935e-01 -1.59576803e-01 6.23790503e-01 2.05172241e-01
-1.07610829e-01 3.52652997e-01 2.43432477e-01 1.58080012e-01
9.48692441e-01 -1.31269884e+00 4.39599305e-01 9.69861209e-01
2.65728295e-01 2.43949831e-01 9.78363752e-01 -3.53263587e-01
-9.65786457e-01 -2.82589883e-01 1.27923930e+00 -1.57830030e-01
5.15883565e-01 -2.97515184e-01 -3.75407636e-01 1.54800653e-01
3.89302343e-01 3.49542350e-01 5.23437023e-01 -2.83027470e-01
-3.95968825e-01 -8.25223774e-02 -1.28395855e+00 4.34572101e-01
1.01693356e+00 -1.91330761e-01 -5.70536137e-01 2.18073726e-01
-4.84309971e-01 -3.39266449e-01 -5.13311505e-01 7.12331414e-01
9.05673504e-01 -1.64248383e+00 1.04992139e+00 -1.24188550e-02
2.59720117e-01 -4.90092069e-01 1.08834960e-01 -1.04089892e+00
-2.22513914e-01 -5.29408097e-01 5.17471731e-01 8.30796003e-01
1.10803515e-01 -6.45204127e-01 1.16457367e+00 9.06171799e-02
2.48962522e-01 -5.14463603e-01 -1.16017103e+00 -1.04237688e+00
-4.45872605e-01 5.46898246e-02 1.16243646e-01 6.89634264e-01
2.21182793e-01 1.00404538e-01 -2.65664220e-01 -2.25939348e-01
7.52000332e-01 6.98862970e-01 9.61547494e-01 -1.42890346e+00
-3.85727622e-02 -9.62489128e-01 -1.49871945e+00 -5.46277702e-01
-2.91170180e-01 -4.55387115e-01 -2.38393694e-01 -1.41231096e+00
3.36145043e-01 1.44680515e-01 -1.09109990e-01 1.39495462e-01
1.24328017e-01 1.12412632e+00 3.23564634e-02 -2.19651386e-02
-6.77363038e-01 4.84562933e-01 1.11623943e+00 3.26363593e-02
8.49460363e-02 1.97170395e-02 -8.62545818e-02 7.86742806e-01
9.67714727e-01 -2.10111469e-01 7.38652349e-02 4.20635492e-01
2.89211005e-01 -5.59305668e-01 5.40431142e-01 -1.33519864e+00
4.34636801e-01 1.10973231e-01 6.94811046e-01 -1.00182784e+00
2.33163804e-01 -7.83164322e-01 -1.12427048e-01 7.57116973e-01
1.92966387e-01 5.26553214e-01 1.40949190e-02 2.92552769e-01
-5.82562625e-01 -1.67860374e-01 8.26388180e-01 4.78244992e-03
-9.60139096e-01 -9.39596668e-02 -1.27568729e-02 -1.14213057e-01
1.48760569e+00 -6.62462056e-01 -1.13093548e-01 -3.00300747e-01
-3.80562127e-01 -2.14547038e-01 8.36255968e-01 5.32071471e-01
8.99628818e-01 -1.22453189e+00 -8.37838471e-01 5.44465601e-01
6.22526705e-01 -4.38597351e-01 2.18331933e-01 1.05652928e+00
-1.13147056e+00 4.78122711e-01 -7.29269981e-01 -7.04652011e-01
-1.92525816e+00 4.74987507e-01 2.66070783e-01 -9.07505378e-02
-5.03113806e-01 6.44309878e-01 1.53081380e-02 -2.19320074e-01
4.59896997e-02 -1.87637076e-01 -2.24983186e-01 1.30853668e-01
3.25978845e-01 6.41873360e-01 3.47572088e-01 -9.79734838e-01
-4.79124427e-01 1.25661755e+00 7.12267309e-02 -3.24131578e-01
1.05995643e+00 5.01601100e-02 -2.13673651e-01 2.16827884e-01
1.31576192e+00 3.88078749e-01 -7.11209714e-01 1.01727555e-02
8.50561038e-02 -1.04853737e+00 -9.19986367e-02 -5.03595948e-01
-9.81225789e-01 7.62982368e-01 8.65737081e-01 4.38243002e-01
1.27403140e+00 -4.00828794e-02 4.28435653e-01 2.84534600e-02
4.24400091e-01 -1.26981401e+00 1.18973851e-01 1.00965381e-01
1.12667382e+00 -1.44364309e+00 4.04772311e-01 -7.15115607e-01
-4.91553962e-01 1.50229788e+00 3.92275572e-01 -2.46864080e-01
6.32590890e-01 8.36213157e-02 5.35692759e-02 -9.78533775e-02
-1.94392368e-01 -3.54494125e-01 7.41527259e-01 3.55933785e-01
5.57858944e-01 -2.28390023e-01 -8.61948550e-01 -2.35034674e-01
-2.41951779e-01 -4.80011739e-02 4.42267030e-01 1.25384104e+00
-5.22043049e-01 -1.20538867e+00 -6.66421056e-01 1.46998718e-01
-4.20166016e-01 2.66820580e-01 -7.13751197e-01 9.74278629e-01
1.04880035e-01 7.26696730e-01 1.50527775e-01 -1.06766470e-01
5.67621529e-01 -2.31594115e-01 7.24758744e-01 3.57428528e-02
-4.24055368e-01 2.00198561e-01 -2.20164478e-01 -6.95327401e-01
-5.58641195e-01 -9.51327920e-01 -8.81847680e-01 -1.78929076e-01
-2.53616452e-01 1.61101863e-01 8.06769013e-01 4.22510654e-01
2.74700850e-01 -5.62091246e-02 7.28707135e-01 -1.15610600e+00
-3.41373533e-01 -6.87521100e-01 -8.68494093e-01 6.19467616e-01
3.49177085e-02 -7.60039926e-01 -6.22927666e-01 -1.25134721e-01] | [8.991484642028809, -2.0091652870178223] |
6f641c33-9374-4c47-acd8-de4342c0ff70 | review-of-face-presentation-attack-detection | 2112.11290 | null | https://arxiv.org/abs/2112.11290v1 | https://arxiv.org/pdf/2112.11290v1.pdf | Review of Face Presentation Attack Detection Competitions | Face presentation attack detection (PAD) has received increasing attention ever since the vulnerabilities to spoofing have been widely recognized. The state of the art in unimodal and multi-modal face anti-spoofing has been assessed in eight international competitions organized in conjunction with major biometrics and computer vision conferences in 2011, 2013, 2017, 2019, 2020 and 2021, each introducing new challenges to the research community. In this chapter, we present the design and results of the five latest competitions from 2019 until 2021. The first two challenges aimed to evaluate the effectiveness of face PAD in multi-modal setup introducing near-infrared (NIR) and depth modalities in addition to colour camera data, while the latest three competitions focused on evaluating domain and attack type generalization abilities of face PAD algorithms operating on conventional colour images and videos. We also discuss the lessons learnt from the competitions and future challenges in the field in general. | ['Guoying Zhao', 'Xiaobai Li', 'Jukka Komulainen', 'Zitong Yu'] | 2021-12-21 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 5.03537118e-01 -3.19205433e-01 6.43800274e-02 -4.05972302e-02
-4.49708939e-01 -7.74941683e-01 7.91867733e-01 -4.05657440e-01
-3.69904786e-01 4.62113649e-01 6.87713698e-02 -1.54887483e-01
-1.56851470e-01 -2.83456445e-01 -2.40157545e-01 -8.03591073e-01
-3.48116159e-01 -7.40625262e-02 2.65848786e-02 -4.10475612e-01
3.36743921e-01 8.19036186e-01 -1.93797827e+00 6.77517176e-01
2.98660994e-01 1.01421988e+00 -6.81260645e-01 8.01126003e-01
3.69494289e-01 2.15792596e-01 -7.28805602e-01 -1.15161860e+00
3.47418398e-01 -1.53816327e-01 -8.51520121e-01 -3.46706867e-01
9.70198572e-01 -4.94240642e-01 -3.07670712e-01 1.02853560e+00
1.04700625e+00 -2.81829119e-01 3.55083108e-01 -1.63553941e+00
-5.71875334e-01 -9.68144313e-02 -8.44242871e-01 5.21115839e-01
9.61798608e-01 9.91071090e-02 1.31406903e-01 -8.63264859e-01
4.81901079e-01 1.68545282e+00 7.35063970e-01 1.15172231e+00
-8.33998024e-01 -1.13011181e+00 -4.49150771e-01 6.05978906e-01
-1.26254690e+00 -9.61870849e-01 6.78585529e-01 -3.06223720e-01
8.79630029e-01 2.72695631e-01 2.64347464e-01 1.68660891e+00
7.28302076e-02 4.26855177e-01 1.39956856e+00 -5.89764833e-01
-3.27651232e-01 4.44005638e-01 -1.96182832e-01 6.49763167e-01
5.34210324e-01 6.05028152e-01 -8.77572775e-01 -5.98750770e-01
4.14370209e-01 -4.74431604e-01 -1.34342313e-01 -1.27883792e-01
-6.88948870e-01 8.77685070e-01 -1.27514973e-01 4.23497915e-01
-1.40119344e-01 -4.20894533e-01 6.83675766e-01 3.48276496e-01
3.34426880e-01 6.87611103e-03 -2.09118903e-01 -2.80346293e-02
-9.40063179e-01 2.63132248e-02 7.16871321e-01 3.72760206e-01
2.92166680e-01 -4.39574430e-03 2.43112981e-01 8.54443371e-01
6.32461488e-01 8.42116535e-01 9.24794972e-02 -7.50332832e-01
3.29050839e-01 -2.58139282e-01 -8.60714447e-03 -1.35696900e+00
-2.49376014e-01 3.05253118e-01 -5.23441315e-01 1.63729623e-01
4.15487677e-01 -2.55681664e-01 -8.70233953e-01 1.38006175e+00
4.12856340e-01 5.94039679e-01 -9.97064337e-02 3.20691258e-01
1.02948391e+00 1.44432038e-01 1.04072094e-01 -4.02436018e-01
1.64581490e+00 -3.97975355e-01 -7.99419641e-01 -7.47778192e-02
2.17155829e-01 -1.29699337e+00 2.34642103e-01 6.28323078e-01
-7.93069899e-01 -5.43989480e-01 -8.48474681e-01 4.85351473e-01
-5.96024990e-01 -2.57500350e-01 3.70514780e-01 2.20388174e+00
-1.38165486e+00 1.35245070e-01 -4.12023872e-01 -7.62895525e-01
8.13001812e-01 6.66516364e-01 -9.45651531e-01 -3.55710745e-01
-1.53562832e+00 9.76722777e-01 -3.58909220e-02 -5.85212894e-02
-9.33926761e-01 -7.12006748e-01 -6.75875187e-01 -5.00330925e-01
-1.27818421e-01 -7.09373504e-04 6.93537414e-01 -7.05812454e-01
-1.55671656e+00 1.49222887e+00 -2.81043142e-01 -1.92287371e-01
1.55622333e-01 -2.09637508e-01 -1.27044261e+00 5.97274363e-01
-1.83534384e-01 5.25856912e-01 1.20215225e+00 -1.14357758e+00
-4.93339866e-01 -6.82068408e-01 -1.15985453e-01 -4.63952869e-01
-6.28323495e-01 8.78347337e-01 -6.64244667e-02 -3.55137944e-01
-3.50752354e-01 -8.58209789e-01 6.36070728e-01 -2.55457908e-01
-8.67279693e-02 -2.42570236e-01 1.56022489e+00 -8.74422133e-01
8.86782825e-01 -2.33014274e+00 -4.14490700e-01 2.69145280e-01
-1.19186260e-01 1.10137904e+00 -3.03059816e-01 5.10535896e-01
-4.77254957e-01 3.45379114e-01 -1.71140712e-02 -2.59791583e-01
-2.03973249e-01 -2.46212766e-01 -1.85018733e-01 9.78966713e-01
1.13010146e-01 3.98153782e-01 -8.23795021e-01 -5.18285811e-01
5.90309083e-01 1.17447984e+00 -4.08749819e-01 -1.26678467e-01
6.99233234e-01 5.98643124e-01 -8.79314449e-03 1.03940547e+00
1.54991853e+00 3.43456119e-01 -2.50616577e-02 -2.07069352e-01
1.85777470e-01 -7.01129362e-02 -1.08450961e+00 1.14579999e+00
-2.24344268e-01 8.35413456e-01 5.57447314e-01 -1.01671886e+00
6.72440708e-01 8.85546386e-01 7.09250748e-01 -7.07876146e-01
9.34679285e-02 3.23904604e-01 -1.72564104e-01 -7.59537995e-01
1.43605089e-02 -2.15464339e-01 6.31240964e-01 3.16025704e-01
4.14557487e-01 2.13461116e-01 -3.73246111e-02 -5.92149794e-02
8.14226627e-01 -1.29644766e-01 -6.38710260e-02 -2.69607186e-01
1.39853513e+00 -6.79552734e-01 8.22245255e-02 5.65645635e-01
-1.04643714e+00 2.97808558e-01 1.97343126e-01 -3.92268032e-01
-3.18668336e-01 -1.01439166e+00 -5.70525587e-01 9.69305694e-01
-6.46579117e-02 -3.83441299e-01 -9.71457779e-01 -8.80746543e-01
1.36905955e-02 -2.47172993e-02 -8.40761006e-01 -1.45017847e-01
-6.04737937e-01 -8.14729631e-01 1.30882752e+00 -1.48000076e-01
9.06337023e-01 -1.08929431e+00 -3.46813560e-01 -4.39781219e-01
-4.42907602e-01 -1.36162245e+00 -3.38822566e-02 -6.46498442e-01
-3.71983290e-01 -1.56569898e+00 -9.58715856e-01 -8.32024097e-01
3.73961121e-01 3.58888984e-01 7.26860821e-01 2.10185722e-01
-7.81158209e-01 1.04426730e+00 -3.01960737e-01 -6.27187371e-01
-2.89069623e-01 -3.71222377e-01 5.07790029e-01 5.15242577e-01
9.16984260e-01 -8.61295834e-02 -7.13245094e-01 6.20858908e-01
-8.23862255e-01 -9.73682940e-01 2.38889337e-01 5.27580619e-01
-4.91835266e-01 -5.74104078e-02 4.49925333e-01 -6.04810774e-01
4.88262266e-01 -3.85095388e-01 -3.52095276e-01 2.66148001e-01
-3.39511424e-01 -5.85915506e-01 -9.38236788e-02 -2.43423581e-01
-1.21498120e+00 -1.64842531e-01 -3.59444410e-01 -1.48323789e-01
-5.28321028e-01 -2.43791774e-01 -1.91898569e-01 -1.03201890e+00
9.59035575e-01 6.27522403e-03 3.55317354e-01 -2.33983994e-01
-6.80082589e-02 9.51507986e-01 4.56701010e-01 -4.80295300e-01
1.05852056e+00 9.62589502e-01 7.76795149e-02 -1.48338950e+00
4.58414666e-02 -6.86781585e-01 -5.56430995e-01 -6.00408494e-01
7.48696744e-01 -6.53506219e-01 -1.24224710e+00 1.34211910e+00
-1.10274291e+00 3.93070132e-01 4.86449987e-01 3.50755900e-01
-3.93393636e-02 9.88182366e-01 -6.18841052e-01 -1.12169075e+00
-3.49540979e-01 -1.23575175e+00 1.12008882e+00 2.99267173e-01
1.46866783e-01 -1.15104496e+00 1.67144239e-01 7.51379430e-01
7.46429145e-01 3.97252172e-01 2.65220940e-01 -4.82562929e-01
2.16398194e-01 -5.26331723e-01 -3.40537816e-01 5.39736032e-01
4.00687099e-01 2.18000650e-01 -1.78182852e+00 -7.89686978e-01
1.19101644e-01 -3.27002198e-01 7.68549383e-01 4.60929900e-01
7.88551152e-01 -5.19726910e-02 -5.92525780e-01 5.51264524e-01
1.18644130e+00 3.15601498e-01 1.16237402e+00 4.16386455e-01
2.73635894e-01 1.24495554e+00 3.35742384e-01 3.84831697e-01
-1.91542193e-01 5.86320639e-01 5.19426167e-01 1.88174650e-01
3.03044114e-02 2.09805563e-01 6.39977217e-01 -7.21302778e-02
-2.89311975e-01 -1.31362900e-01 -9.09933269e-01 2.11159468e-01
-8.66015732e-01 -1.43398535e+00 -1.04228780e-02 2.52234006e+00
2.37282559e-01 -4.05051440e-01 5.29428422e-01 5.31887949e-01
1.22291851e+00 2.56215364e-01 9.50368643e-02 -5.11623800e-01
-3.21700454e-01 6.41781509e-01 3.61591816e-01 4.85344976e-01
-1.53746998e+00 8.56211364e-01 7.22628069e+00 6.48516774e-01
-1.35512221e+00 2.10190341e-01 5.71849942e-01 6.21453710e-02
3.18363160e-01 -2.92562574e-01 -9.32235062e-01 4.16145414e-01
1.35740221e+00 2.60073572e-01 5.19792080e-01 3.99894744e-01
-2.94959158e-01 2.37795170e-02 -6.01875007e-01 1.32097864e+00
6.91960752e-01 -1.08454669e+00 -1.73572525e-01 5.12109041e-01
5.30558288e-01 7.81366304e-02 6.52083933e-01 -1.21783227e-01
-2.41162464e-01 -1.17501342e+00 5.49127124e-02 -6.01138771e-02
1.01145804e+00 -7.64635265e-01 7.08113849e-01 -4.00422335e-01
-1.13557053e+00 -3.17443132e-01 -2.09195837e-01 4.76430178e-01
-6.04215860e-02 1.49700418e-01 -4.91496176e-01 6.18787348e-01
1.08763731e+00 5.17602265e-01 -5.83833635e-01 9.22744989e-01
1.74624678e-02 5.14139891e-01 -3.88971835e-01 3.22187454e-01
-2.13765189e-01 4.41706568e-01 4.83066440e-01 1.40569401e+00
1.32356778e-01 -8.86284336e-02 -3.92492324e-01 -3.60042863e-02
1.70367956e-01 -1.27607092e-01 -6.11620486e-01 -1.06509276e-01
4.18899059e-01 1.30828393e+00 -4.26481724e-01 1.34369925e-01
-6.08241141e-01 9.14922714e-01 -6.31248415e-01 3.06804717e-01
-6.58117890e-01 -3.31846058e-01 9.93219674e-01 1.16175734e-01
1.83239087e-01 1.23217046e-01 1.47276580e-01 -8.85000706e-01
-2.62615710e-01 -1.28900719e+00 8.38634312e-01 -3.08563292e-01
-1.22119617e+00 5.98154843e-01 1.94502667e-01 -9.29003060e-01
-1.27037987e-01 -1.29055536e+00 -3.59540790e-01 9.81866062e-01
-1.67238283e+00 -1.13539505e+00 -2.00496927e-01 9.90616500e-01
-1.13959931e-01 -6.90315902e-01 1.11856163e+00 6.72812879e-01
-5.97463965e-01 1.21564829e+00 -1.10656299e-01 3.17043334e-01
1.14410436e+00 -4.87335235e-01 4.60015684e-01 7.54288435e-01
-9.39883664e-02 8.69139254e-01 5.10822237e-01 -4.68726158e-01
-1.52159727e+00 -3.92683297e-01 7.61350453e-01 -5.29018044e-01
3.31424475e-01 -1.55217350e-01 -4.08986896e-01 2.57400274e-01
3.58726174e-01 3.32396954e-01 1.03552628e+00 -1.75155565e-01
-9.26528156e-01 -1.41915068e-01 -2.10427928e+00 -1.04629673e-01
6.64326727e-01 -1.08515298e+00 1.12281851e-01 3.89360189e-01
-2.53181487e-01 -9.53077897e-02 -6.80291414e-01 6.35946929e-01
1.06989038e+00 -1.33797204e+00 1.45700896e+00 -5.33494651e-01
-9.51981172e-02 1.61805019e-01 -1.45658970e-01 -6.89729393e-01
-1.06182992e-01 -1.09173429e+00 -1.12419657e-01 1.21005011e+00
-1.31718457e-01 -1.04219234e+00 9.50249851e-01 8.90004858e-02
6.10551894e-01 -1.49281994e-01 -1.29662073e+00 -7.58739531e-01
1.14090957e-01 -2.03766823e-01 5.27603567e-01 8.72687340e-01
-9.82664749e-02 -4.33731437e-01 -6.03330076e-01 5.53817749e-01
1.00260651e+00 -8.39014351e-01 6.80569410e-01 -1.41891146e+00
1.10787392e-01 -3.91535014e-01 -9.15223420e-01 -3.24330181e-01
6.52252883e-02 -3.80791992e-01 -5.10500729e-01 -7.94064045e-01
3.46242897e-02 2.65087873e-01 -4.34020191e-01 2.29184031e-01
-1.19624436e-01 9.16304290e-01 1.23460054e-01 -2.23159995e-02
7.36657307e-02 -2.49453291e-01 8.83862257e-01 -1.21413685e-01
4.10897344e-01 -4.12550457e-02 -5.04751265e-01 3.07053804e-01
6.90088570e-01 -1.07712507e-01 -2.42584035e-01 -1.08804367e-01
-3.00513238e-01 -2.80084997e-01 5.97494900e-01 -1.12696469e+00
-4.38461080e-04 9.63370427e-02 6.19561315e-01 -1.93346068e-01
6.99854076e-01 -8.00615370e-01 -1.90225676e-01 7.42218316e-01
2.89403886e-01 3.90517712e-02 6.65657163e-01 2.33661845e-01
-4.27084900e-02 -4.17047441e-02 1.29864931e+00 5.38268350e-02
-7.60557413e-01 4.39664096e-01 -3.80633593e-01 -1.51333204e-02
1.24079263e+00 -6.43630207e-01 -7.59144068e-01 -2.34754220e-01
-6.33213997e-01 -3.62880319e-01 1.43849537e-01 7.63622046e-01
7.23063469e-01 -9.37521458e-01 -8.86556625e-01 8.97385061e-01
7.99077377e-02 -1.25494432e+00 5.17560303e-01 7.89823055e-01
-6.30471289e-01 7.56962240e-01 -7.68237412e-01 -5.64820528e-01
-1.95573997e+00 6.11928403e-01 4.25926298e-01 9.63722467e-02
2.37571523e-01 1.34852159e+00 -1.86837509e-01 -2.56909370e-01
2.74337769e-01 1.09243429e+00 -4.70172793e-01 1.02166720e-01
1.20550454e+00 7.63767183e-01 2.63440251e-01 -1.27056599e+00
-9.45425689e-01 6.61274850e-01 -1.97492763e-01 -2.80442506e-01
1.00703621e+00 2.73150858e-02 -4.11850303e-01 -5.08996725e-01
1.40674984e+00 -4.85353405e-03 -6.14967823e-01 1.27566501e-01
1.83353797e-01 -9.59061861e-01 -1.59259826e-01 -7.69595325e-01
-1.24606884e+00 1.17015243e+00 1.39769888e+00 3.39631408e-01
1.22981763e+00 -2.39277139e-01 6.52583361e-01 -1.90176293e-01
4.53103155e-01 -8.60133886e-01 2.21242592e-01 2.06607416e-01
6.85492575e-01 -1.26403534e+00 -1.39931187e-01 -7.93959200e-01
-1.91184804e-01 1.22512150e+00 4.58422095e-01 2.21389890e-01
1.13106358e+00 8.16016421e-02 1.98905513e-01 -2.53878176e-01
8.82760286e-02 1.15767337e-01 2.75163621e-01 1.56377447e+00
6.13490522e-01 -3.38328511e-01 -7.44815543e-03 -2.75886744e-01
1.21122360e-01 -9.01154280e-02 8.21715072e-02 1.05194485e+00
-2.71185040e-02 -1.58931732e+00 -1.02304125e+00 2.51187146e-01
-9.64321613e-01 1.57763615e-01 -8.15256596e-01 5.20868719e-01
2.87237138e-01 1.60534775e+00 -3.24575156e-01 -7.66325474e-01
3.83510925e-02 2.12841436e-01 6.65078580e-01 -2.27172002e-01
-6.34766400e-01 -3.06058675e-01 5.72204292e-02 -6.38270080e-01
-7.83398807e-01 -9.06009018e-01 -3.97798270e-01 -9.42794323e-01
-5.27201653e-01 5.28216362e-02 9.35975313e-01 7.85544693e-01
3.99034262e-01 -1.64490879e-01 8.88891399e-01 -8.61450970e-01
-1.72971323e-01 -8.44256699e-01 -4.35944080e-01 2.70076603e-01
7.53421962e-01 -7.98168421e-01 -6.41772687e-01 1.26285730e-02] | [13.055747985839844, 1.1148267984390259] |
c1c4ee95-cd92-4996-9458-8f71ae786cc6 | multi-level-and-multi-scale-feature-1 | 1706.06810 | null | http://arxiv.org/abs/1706.06810v1 | http://arxiv.org/pdf/1706.06810v1.pdf | Multi-Level and Multi-Scale Feature Aggregation Using Sample-level Deep Convolutional Neural Networks for Music Classification | Music tag words that describe music audio by text have different levels of
abstraction. Taking this issue into account, we propose a music classification
approach that aggregates multi-level and multi-scale features using pre-trained
feature extractors. In particular, the feature extractors are trained in
sample-level deep convolutional neural networks using raw waveforms. We show
that this approach achieves state-of-the-art results on several music
classification datasets. | ['Juhan Nam', 'Jongpil Lee'] | 2017-06-21 | null | null | null | null | ['music-classification'] | ['music'] | [ 1.91392899e-01 -4.44014192e-01 -7.26058632e-02 -2.49721348e-01
-1.07996070e+00 -8.29688787e-01 3.54754746e-01 3.54063399e-02
-2.11981729e-01 1.36493832e-01 3.11599523e-01 3.04819793e-01
-4.87644643e-01 -8.34536493e-01 -5.64006090e-01 -4.07853514e-01
-3.71305466e-01 2.01138034e-02 -2.04332937e-02 -9.74577963e-02
3.38106662e-01 5.43959856e-01 -1.85963786e+00 7.46919811e-01
2.28817701e-01 1.53980935e+00 -3.59000117e-01 9.18026030e-01
-6.10883571e-02 7.17118025e-01 -8.67741585e-01 -2.66924858e-01
2.72818208e-01 -3.89985740e-01 -6.29483700e-01 -1.70421228e-01
7.42198169e-01 -1.62083864e-01 -4.96044636e-01 8.92097056e-01
6.07008696e-01 1.47691384e-01 4.69908834e-01 -9.67973650e-01
-4.46444511e-01 1.30114985e+00 -1.85419947e-01 -7.40347654e-02
2.57780939e-01 -1.14232346e-01 1.74567688e+00 -7.71501362e-01
1.20962016e-01 9.71785188e-01 1.10012674e+00 1.20432563e-01
-1.26535118e+00 -8.88186216e-01 -2.68025190e-01 3.73794287e-01
-1.57101977e+00 -3.84088606e-01 1.05497670e+00 -4.77800280e-01
9.39379334e-01 1.91593304e-01 8.79739761e-01 8.74860346e-01
9.19714645e-02 8.15082371e-01 6.16498649e-01 -4.40589726e-01
1.16050847e-01 -6.13843322e-01 2.93456823e-01 3.90224904e-01
1.53373927e-01 -1.57888189e-01 -1.19576228e+00 -4.45495874e-01
6.72394276e-01 -2.45450839e-01 -9.06766858e-03 5.55250198e-02
-1.27260256e+00 8.05426598e-01 9.72673818e-02 6.77960575e-01
-5.95696449e-01 8.17955434e-01 7.90700138e-01 3.53952795e-01
2.55559891e-01 7.31515884e-01 -5.45969784e-01 -4.77669150e-01
-1.53452098e+00 3.33184123e-01 7.50230074e-01 6.75225317e-01
4.56206888e-01 4.59569424e-01 -1.40426144e-01 8.39443624e-01
-9.49892104e-02 3.48756790e-01 6.93022311e-01 -8.26561689e-01
1.35825112e-01 2.97090322e-01 -3.31585139e-01 -8.10070992e-01
-5.62747359e-01 -1.11972046e+00 -7.16713369e-01 2.32681617e-01
1.23909831e-01 2.06865489e-01 -6.13702536e-01 1.61779714e+00
-2.38930270e-01 5.98620832e-01 2.47694086e-02 8.39039445e-01
8.73622835e-01 4.36023206e-01 -2.83043623e-01 2.12881088e-01
1.65851486e+00 -7.77006507e-01 -7.24030018e-01 2.67619193e-01
1.38069600e-01 -8.74482751e-01 1.00027275e+00 9.55440581e-01
-1.14042580e+00 -9.47134912e-01 -1.51683223e+00 -1.07592709e-01
-4.62900043e-01 7.21379995e-01 5.27981341e-01 6.22433841e-01
-6.32306039e-01 1.15923929e+00 -5.00338912e-01 6.49091676e-02
3.19897085e-01 3.88899833e-01 -7.09444210e-02 7.61613667e-01
-1.17747641e+00 5.46272732e-02 3.98736507e-01 -1.70905262e-01
-8.70951653e-01 -7.71150947e-01 -5.19510329e-01 5.65134704e-01
4.05611284e-02 -4.31005508e-01 1.53855169e+00 -8.45905542e-01
-1.72797537e+00 5.19568264e-01 1.29064873e-01 -4.90749359e-01
1.57944486e-02 -5.09407103e-01 -7.01545298e-01 2.65185952e-01
-1.97094738e-01 1.37510806e-01 1.19850814e+00 -8.67644250e-01
-7.48196542e-01 -8.79771635e-02 5.80697693e-02 -2.22484395e-01
-6.00041628e-01 1.95662737e-01 -1.85886428e-01 -1.20864308e+00
5.89958988e-02 -9.09922183e-01 2.93893963e-01 -2.10850716e-01
-6.12326443e-01 -2.74286300e-01 5.66485226e-01 -3.65710258e-01
1.51300633e+00 -2.43009210e+00 2.21279621e-01 3.74275953e-01
1.40898600e-01 1.78194456e-02 -4.06805873e-01 3.82514954e-01
-1.06538944e-01 -2.63426080e-02 4.35810462e-02 -2.11590543e-01
5.55653393e-01 -2.77068615e-01 -5.68409741e-01 6.22001179e-02
1.47500798e-01 7.16787577e-01 -5.99623561e-01 -2.05909684e-01
1.65157214e-01 6.25950277e-01 -6.76265001e-01 -9.69642475e-02
-6.47853836e-02 9.39940959e-02 -3.25078487e-01 5.73166311e-01
2.61070579e-01 -9.03705060e-02 3.94376367e-02 -6.12736762e-01
-1.46423087e-01 9.08651531e-01 -1.50636601e+00 2.09607387e+00
-3.59171838e-01 9.34215486e-01 -2.53105521e-01 -8.24529648e-01
8.68114471e-01 6.01552904e-01 8.81331503e-01 -1.00187130e-01
2.59259194e-01 5.12580156e-01 -1.99265638e-03 -1.42395362e-01
7.08367705e-01 -1.97878122e-01 -6.46342814e-01 6.51443839e-01
6.23493969e-01 -1.40118703e-01 -1.01723401e-02 -3.11200649e-01
1.14798725e+00 6.39258549e-02 3.98004800e-01 -3.43669765e-02
4.73560840e-01 -3.35528672e-01 4.44125384e-01 8.99922311e-01
2.20799223e-01 7.51355410e-01 3.39368224e-01 -5.79751551e-01
-8.85270536e-01 -9.86109197e-01 -1.37211382e-01 1.51017070e+00
-5.64504206e-01 -1.07905817e+00 -6.95187032e-01 -2.99079508e-01
1.24454454e-01 1.87057957e-01 -5.61593056e-01 -6.05764054e-02
-5.76376736e-01 -2.86321163e-01 1.37020421e+00 6.33468866e-01
1.44532546e-01 -1.21388972e+00 -6.03319764e-01 5.78950465e-01
-2.70696152e-02 -1.04169989e+00 -3.41431022e-01 6.76339507e-01
-8.07699144e-01 -1.02361786e+00 -5.56978285e-01 -6.89543307e-01
-3.00219595e-01 -1.33535802e-01 1.18863690e+00 -1.72840625e-01
-3.52847844e-01 1.09870836e-01 -6.71508372e-01 -6.87491477e-01
-1.50234878e-01 6.63663924e-01 3.38255346e-01 1.67712629e-01
5.58992326e-01 -1.05755413e+00 -3.19134504e-01 -9.68282446e-02
-1.03195941e+00 -3.30656648e-01 5.64220190e-01 5.17269790e-01
6.37533426e-01 1.95309252e-01 8.49529624e-01 -2.59642154e-01
7.71473467e-01 1.21569902e-01 -4.33743566e-01 -6.13352656e-02
-2.21445307e-01 1.29034624e-01 6.84748113e-01 -8.01850080e-01
-4.30391133e-02 2.84168929e-01 -3.61709595e-01 -4.82368201e-01
-2.66694993e-01 5.74916661e-01 1.07598118e-02 4.47590246e-05
5.96973181e-01 1.47861615e-01 -6.99620128e-01 -8.52522910e-01
4.74113554e-01 1.08718860e+00 8.12328517e-01 -5.67128599e-01
9.16949093e-01 3.46093208e-01 2.06535682e-02 -6.66131735e-01
-1.02061689e+00 -4.08876300e-01 -8.57416213e-01 -3.64172965e-01
8.25389326e-01 -9.14441884e-01 -9.33719635e-01 5.91327727e-01
-8.91544938e-01 2.09033061e-02 -7.26957798e-01 8.13377559e-01
-9.76434171e-01 2.71952013e-03 -8.25342894e-01 -6.58948541e-01
-6.52594149e-01 -7.75951385e-01 1.36102235e+00 -7.36644566e-02
-5.51626980e-01 -3.03942412e-01 2.40478948e-01 -2.19177499e-01
3.30530941e-01 -1.66079234e-02 1.04832745e+00 -9.59035337e-01
-5.54917395e-01 -1.93332687e-01 2.87038796e-02 4.61874783e-01
1.16586544e-01 -1.27058446e-01 -1.53284645e+00 -1.85469631e-02
-2.50420123e-01 -3.96096677e-01 9.50065434e-01 4.03081864e-01
1.38955498e+00 -5.41717783e-02 2.32824191e-01 8.60051394e-01
1.31521749e+00 3.60055231e-02 4.33174223e-01 4.16006088e-01
6.75184071e-01 9.44247693e-02 1.14094548e-01 8.44289422e-01
-1.87770024e-01 1.02970195e+00 1.43341199e-01 2.97138631e-01
-2.31304079e-01 -1.63990200e-01 3.83757681e-01 1.04862344e+00
-1.25349268e-01 2.31006429e-01 -5.91755509e-01 5.65842509e-01
-1.83766758e+00 -1.18925560e+00 2.31007356e-02 1.87188637e+00
7.59029925e-01 1.51736766e-01 4.73609358e-01 1.03523004e+00
5.13902426e-01 3.36748928e-01 -4.78831202e-01 -3.48660082e-01
-1.81157768e-01 9.20215249e-01 1.89063683e-01 8.41238303e-04
-1.31760371e+00 7.65226364e-01 7.00658607e+00 1.04649425e+00
-1.29412794e+00 1.74882710e-01 -3.40225548e-01 -3.58641565e-01
3.86609547e-02 -3.31575006e-01 -2.96007633e-01 1.21483080e-01
1.06460369e+00 1.86891165e-02 6.51336312e-01 6.77404463e-01
-2.60834862e-02 7.38603771e-01 -1.11771107e+00 1.35381770e+00
7.99909160e-02 -1.30350864e+00 1.64676398e-01 -3.40230094e-04
4.01888579e-01 1.07910305e-01 1.14536732e-01 3.17232400e-01
-1.44687459e-01 -8.40262830e-01 1.25684094e+00 7.14866400e-01
9.27777886e-01 -1.02178335e+00 5.83318949e-01 -3.09972882e-01
-1.73625672e+00 -1.47826523e-01 -1.32839680e-01 -2.38758087e-01
8.67806524e-02 7.55999982e-01 -3.73069644e-01 7.60181963e-01
7.57079899e-01 8.91131938e-01 -3.15503508e-01 1.11083758e+00
-4.21264730e-02 8.87503564e-01 -3.19271117e-01 1.61199048e-01
9.98888090e-02 2.17768669e-01 5.73644400e-01 1.47243261e+00
7.58040607e-01 -1.83270022e-01 3.93012352e-02 7.88191974e-01
-2.17720211e-01 1.15189932e-01 -2.79597908e-01 -4.99471664e-01
2.38446370e-01 1.35415196e+00 -6.06103003e-01 -4.95067328e-01
-1.76136836e-01 8.43298972e-01 -1.22196980e-01 3.89664210e-02
-6.60634458e-01 -1.03241765e+00 1.08928561e+00 -7.26682842e-02
6.64839745e-01 -2.36851200e-01 -3.52431267e-01 -1.13350236e+00
-5.83570041e-02 -9.00658846e-01 2.80911624e-01 -8.26702654e-01
-1.42808270e+00 6.04568124e-01 -4.25698370e-01 -1.78535545e+00
-4.91175532e-01 -7.60206759e-01 -5.32230556e-01 4.45901275e-01
-1.21121073e+00 -1.14535069e+00 -1.04212448e-01 5.65340638e-01
3.66954058e-01 -6.40852511e-01 1.19414473e+00 5.90896845e-01
1.17255516e-01 7.86072016e-01 9.76315215e-02 6.31286263e-01
6.53396606e-01 -1.19051290e+00 6.10485137e-01 1.98057085e-01
1.09865892e+00 4.86595601e-01 4.29941356e-01 -1.62735298e-01
-1.27455664e+00 -9.86999452e-01 9.55619276e-01 -2.45625779e-01
9.29733098e-01 -4.13332671e-01 -6.23671472e-01 5.12353659e-01
1.16096146e-01 -2.00411022e-01 1.20158052e+00 5.48321962e-01
-8.09372306e-01 -2.64305890e-01 -5.35317302e-01 1.34042606e-01
1.02447581e+00 -1.24819911e+00 -7.82518387e-01 -1.25402316e-01
5.31307936e-01 -1.73231944e-01 -1.20185459e+00 3.99227500e-01
1.27702284e+00 -6.29538357e-01 1.08534288e+00 -6.71321034e-01
3.12129050e-01 -4.03995961e-01 -3.82152915e-01 -1.18725443e+00
-4.41225052e-01 -8.22880566e-01 -2.39906877e-01 1.01243126e+00
3.46112579e-01 -5.95061220e-02 4.71978933e-01 -5.51662862e-01
-2.51884043e-01 -3.66024613e-01 -1.06418669e+00 -1.02491832e+00
-1.00776643e-01 -9.97625589e-01 9.40456390e-01 9.47194874e-01
3.48018914e-01 4.19761389e-01 -5.26759148e-01 1.09709157e-02
3.75851154e-01 4.33316231e-01 7.39462614e-01 -1.85036361e+00
-5.08413792e-01 -7.70986915e-01 -9.79682565e-01 -6.55091703e-01
3.24760407e-01 -1.06758070e+00 -7.47280195e-02 -1.25459146e+00
-7.13293478e-02 -7.64487535e-02 -1.02617645e+00 5.97153723e-01
2.93036103e-01 9.23599482e-01 4.97083008e-01 2.16206461e-01
-6.52351737e-01 3.34920228e-01 6.01993918e-01 -3.71655911e-01
-2.39228025e-01 3.86680365e-02 -4.62330431e-01 8.49231958e-01
8.66903305e-01 -7.18685567e-01 3.18562537e-02 -2.66734362e-01
5.69344282e-01 -3.22675347e-01 4.29662436e-01 -1.80095148e+00
1.59548029e-01 2.55075574e-01 3.11785519e-01 -7.63808072e-01
6.97251916e-01 -6.02957189e-01 2.78075397e-01 2.45498613e-01
-8.64497244e-01 -1.44123852e-01 2.41266713e-01 1.54390484e-01
-5.54500163e-01 -2.97533303e-01 4.27498549e-01 3.05310845e-01
-1.88658193e-01 1.18075227e-02 -5.60509980e-01 -3.22861493e-01
8.62157717e-02 1.26798123e-01 -1.26971118e-02 -5.09651124e-01
-9.91637468e-01 -6.70757949e-01 -7.28062317e-02 6.07671797e-01
3.06237578e-01 -1.90629697e+00 -7.08090246e-01 2.29658082e-01
4.43480134e-01 -7.08011627e-01 9.74840000e-02 4.07066345e-01
-1.09744646e-01 4.74592894e-01 -3.23380947e-01 -5.02267897e-01
-1.16871214e+00 1.91435248e-01 3.18943650e-01 -8.73176605e-02
-7.24390209e-01 5.46269178e-01 -3.87413293e-01 -2.33914092e-01
4.22764957e-01 -8.70810688e-01 -2.24788859e-01 4.88622695e-01
7.92630136e-01 1.29015863e-01 1.57601178e-01 -6.86709762e-01
-2.75921643e-01 1.00519490e+00 4.70507652e-01 -5.00304699e-01
1.46223867e+00 3.68837953e-01 5.26530482e-02 1.07080793e+00
1.10274351e+00 3.62751544e-01 -6.88365519e-01 -2.32658088e-01
3.09474356e-02 -1.46148011e-01 2.20948324e-01 -7.56959677e-01
-1.11569548e+00 9.45799828e-01 7.44088233e-01 5.58254778e-01
1.31769216e+00 -5.31575233e-02 1.02105916e+00 7.44875371e-01
3.09184819e-01 -1.23914194e+00 8.26981813e-02 7.16538787e-01
7.72877455e-01 -4.92854893e-01 -1.58573911e-01 4.98454012e-02
7.38844424e-02 1.49937201e+00 -1.45150334e-01 -6.63675964e-01
8.21929574e-01 4.85070884e-01 1.75248325e-01 -3.64587873e-01
-4.60372061e-01 -5.93440294e-01 8.74983370e-01 1.56163275e-01
6.06876373e-01 2.47581944e-01 -2.58773059e-01 1.45047009e+00
-7.58247316e-01 2.71036536e-01 3.50984007e-01 7.56711602e-01
-5.80000460e-01 -1.24018049e+00 -4.16775674e-01 4.08783883e-01
-9.39848125e-01 -2.83217877e-01 -5.72139382e-01 4.16851163e-01
2.82060683e-01 9.12237883e-01 5.61536290e-02 -9.27127302e-01
4.48681056e-01 5.35536349e-01 5.08537889e-01 -4.79455173e-01
-1.18780780e+00 4.36404020e-01 8.70967135e-02 -5.36215246e-01
-5.46843886e-01 -3.76013935e-01 -1.24888277e+00 2.04144996e-02
-2.93452293e-01 1.97820202e-01 1.02598834e+00 8.33233595e-01
4.71366107e-01 1.10340512e+00 5.33139825e-01 -1.19614768e+00
-4.33532983e-01 -1.27949691e+00 -1.00482714e+00 3.83099377e-01
4.98185813e-01 -4.52591121e-01 -1.73210964e-01 1.54929355e-01] | [15.774018287658691, 5.228280067443848] |
5f26f95a-3772-4259-a133-8d67e808952c | pareto-self-supervised-training-for-few-shot | 2104.07841 | null | https://arxiv.org/abs/2104.07841v2 | https://arxiv.org/pdf/2104.07841v2.pdf | Pareto Self-Supervised Training for Few-Shot Learning | While few-shot learning (FSL) aims for rapid generalization to new concepts with little supervision, self-supervised learning (SSL) constructs supervisory signals directly computed from unlabeled data. Exploiting the complementarity of these two manners, few-shot auxiliary learning has recently drawn much attention to deal with few labeled data. Previous works benefit from sharing inductive bias between the main task (FSL) and auxiliary tasks (SSL), where the shared parameters of tasks are optimized by minimizing a linear combination of task losses. However, it is challenging to select a proper weight to balance tasks and reduce task conflict. To handle the problem as a whole, we propose a novel approach named as Pareto self-supervised training (PSST) for FSL. PSST explicitly decomposes the few-shot auxiliary problem into multiple constrained multi-objective subproblems with different trade-off preferences, and here a preference region in which the main task achieves the best performance is identified. Then, an effective preferred Pareto exploration is proposed to find a set of optimal solutions in such a preference region. Extensive experiments on several public benchmark datasets validate the effectiveness of our approach by achieving state-of-the-art performance. | ['Donglin Wang', 'Siteng Huang', 'Heshen Zhan', 'Jixie Ge', 'Zhengyu Chen'] | 2021-04-16 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Chen_Pareto_Self-Supervised_Training_for_Few-Shot_Learning_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Chen_Pareto_Self-Supervised_Training_for_Few-Shot_Learning_CVPR_2021_paper.pdf | cvpr-2021-1 | ['auxiliary-learning'] | ['methodology'] | [ 3.55996996e-01 1.11559018e-01 -5.92354178e-01 -4.21837777e-01
-9.99864995e-01 -1.92683965e-01 3.52658927e-01 1.36857972e-01
-4.62049305e-01 8.76100600e-01 1.03214927e-01 1.65871114e-01
-5.44468403e-01 -4.44763571e-01 -5.35748482e-01 -1.05321658e+00
2.36988395e-01 4.96563792e-01 1.97287723e-01 -1.00847617e-01
2.48143166e-01 -1.25716031e-01 -1.48063183e+00 3.09274010e-02
1.51175117e+00 1.10018706e+00 6.32235467e-01 -1.23550199e-01
-2.30039954e-01 5.25417626e-01 -3.74039769e-01 -2.15753645e-01
2.79487371e-01 -4.96877015e-01 -8.27084363e-01 3.19157511e-01
-1.81294397e-01 2.78867006e-01 3.15963745e-01 1.12660944e+00
6.27476871e-01 5.85765898e-01 5.45547187e-01 -1.49128628e+00
-4.80860233e-01 6.22996867e-01 -5.17486453e-01 7.42649660e-02
-1.19756132e-01 2.10150048e-01 1.15761876e+00 -1.04472637e+00
5.14827907e-01 1.06584680e+00 2.34483078e-01 6.29527867e-01
-1.68865228e+00 -6.51492715e-01 5.24851978e-01 1.92328334e-01
-1.30229235e+00 -4.12311941e-01 1.08046114e+00 -2.41590872e-01
5.98125279e-01 -1.13287725e-01 4.29756165e-01 1.11423123e+00
-1.74786195e-01 1.12549376e+00 1.22537673e+00 -3.58164996e-01
7.85455704e-01 5.14473498e-01 3.43570918e-01 5.97743094e-01
5.13546169e-02 -7.34392703e-02 -7.06189692e-01 -1.67634889e-01
1.86164185e-01 2.73711354e-01 -3.43683153e-01 -9.41328228e-01
-1.14605916e+00 8.33662093e-01 3.73160750e-01 3.19720268e-01
-3.59509140e-01 -2.87109494e-01 4.68223810e-01 4.01992232e-01
6.42714798e-01 6.40588701e-01 -5.37389278e-01 1.69252437e-02
-7.27504432e-01 -3.32443900e-02 4.49742287e-01 9.33792293e-01
1.10363901e+00 7.60040581e-02 -5.84066212e-01 1.04388869e+00
3.13523673e-02 1.11664876e-01 6.35811210e-01 -7.47042954e-01
5.48316777e-01 7.03808725e-01 7.80678391e-02 -4.84193206e-01
-1.87587634e-01 -8.17742467e-01 -5.69568574e-01 3.81557718e-02
9.79506448e-02 -4.12904084e-01 -8.20192695e-01 1.96034014e+00
3.17626566e-01 3.09492886e-01 1.33951291e-01 1.00492680e+00
4.80910122e-01 5.81106305e-01 1.92856461e-01 -6.28960073e-01
1.00939894e+00 -1.06246674e+00 -6.25920951e-01 -5.51691234e-01
6.15468442e-01 -2.28900477e-01 1.25176752e+00 1.71934575e-01
-7.61936426e-01 -5.38723886e-01 -1.00610566e+00 3.37374985e-01
-2.67647475e-01 -2.65575051e-02 4.23341006e-01 4.84336287e-01
-7.15768874e-01 5.57736933e-01 -4.87574846e-01 -1.10774457e-01
7.44380236e-01 3.20156246e-01 9.51859448e-03 -4.28035520e-02
-1.29372394e+00 6.74596250e-01 7.03442335e-01 -3.01376909e-01
-1.08583367e+00 -7.99647212e-01 -8.33998144e-01 2.77974218e-01
1.27839851e+00 -3.91170025e-01 1.12108719e+00 -1.05781090e+00
-1.50650895e+00 6.34487510e-01 -1.00406677e-01 -4.21782762e-01
2.79951870e-01 -1.20375700e-01 -2.27820739e-01 -1.30339772e-01
3.39420170e-01 5.17506540e-01 1.12364018e+00 -1.46746469e+00
-8.60234201e-01 -4.27544445e-01 -7.07571283e-02 4.62128103e-01
-9.13865328e-01 -1.96869180e-01 -5.71940422e-01 -6.04905486e-01
-2.15748057e-01 -7.96844482e-01 -3.88186067e-01 -2.83312201e-01
-3.00773591e-01 -5.21689355e-01 7.79740453e-01 -8.14319551e-02
1.37852943e+00 -2.04623580e+00 5.72483122e-01 9.49923694e-02
1.73392281e-01 4.09122646e-01 -3.48182887e-01 6.35977760e-02
3.17570381e-02 -2.22316623e-01 -3.95794123e-01 -5.94196141e-01
1.34663954e-02 2.98826277e-01 -2.64638543e-01 2.28956640e-01
1.47282526e-01 9.63667691e-01 -1.22872782e+00 -7.35878825e-01
4.00736406e-02 -9.59465876e-02 -3.56540024e-01 3.47500920e-01
-3.22349519e-01 4.17269439e-01 -8.52576256e-01 9.13703978e-01
2.63847023e-01 -4.54872489e-01 2.60236021e-02 -1.46853272e-02
-1.05703957e-02 -3.77603471e-02 -1.16377747e+00 1.97415507e+00
-5.31417131e-01 1.03701673e-01 1.11094117e-01 -1.42124319e+00
1.16949522e+00 1.92284226e-01 5.97163022e-01 -6.08914018e-01
2.39444509e-01 1.36894658e-01 -1.31334573e-01 -2.65563518e-01
1.22847833e-01 -4.00707364e-01 -1.87863931e-01 6.41840816e-01
4.64364052e-01 1.11948997e-01 2.26660714e-01 1.49971038e-01
7.30632722e-01 2.41333783e-01 5.16415834e-01 -4.32013720e-01
5.86072147e-01 -2.23141350e-02 9.83770490e-01 7.00568616e-01
-4.64187622e-01 2.77975410e-01 3.93515021e-01 -1.58153698e-01
-4.63967413e-01 -9.65982139e-01 1.54566243e-01 1.48666120e+00
3.83748084e-01 -2.19707415e-01 -5.49002945e-01 -1.06611049e+00
-2.87158899e-02 9.24156964e-01 -7.28866637e-01 -5.81509352e-01
-2.33033851e-01 -7.65019119e-01 -1.01009414e-01 4.62300777e-01
2.57937253e-01 -1.18840492e+00 -8.95301044e-01 2.31754661e-01
-1.81888372e-01 -8.01296175e-01 -7.38579452e-01 8.42571080e-01
-9.52741921e-01 -9.15121913e-01 -1.05203032e+00 -9.13049102e-01
7.61488676e-01 5.16413093e-01 9.41099286e-01 -3.05182576e-01
-1.63525641e-01 2.21327245e-01 -4.47685331e-01 -3.86892110e-01
8.64277408e-02 1.68522030e-01 2.07451850e-01 4.49734837e-01
4.44366813e-01 -5.97862005e-01 -1.80124938e-01 4.70718443e-01
-7.69420207e-01 -6.82147294e-02 7.39032209e-01 1.20387936e+00
8.25478137e-01 5.99725284e-02 8.96758020e-01 -9.58898544e-01
8.91498983e-01 -6.13039911e-01 -4.73641813e-01 6.89255178e-01
-1.00706434e+00 5.09557486e-01 7.75850892e-01 -6.53365672e-01
-1.30689490e+00 1.26716137e-01 6.41622305e-01 -9.80550289e-01
7.01461658e-02 5.50326765e-01 -3.42922151e-01 1.09066360e-01
5.37455142e-01 3.87821794e-01 1.28610760e-01 -4.09871906e-01
2.71194905e-01 5.16885459e-01 3.00753146e-01 -7.58367658e-01
8.53292227e-01 4.19225544e-01 -2.20079154e-01 -4.87847924e-01
-1.41640842e+00 -6.56490088e-01 -6.62478983e-01 -3.58094305e-01
5.26756346e-01 -6.24706447e-01 -5.55281997e-01 1.47598177e-01
-6.29052997e-01 -2.63483554e-01 -7.54934251e-01 3.73937964e-01
-6.33356571e-01 9.77065340e-02 1.21945351e-01 -1.08344519e+00
-2.49200493e-01 -1.20183086e+00 9.56442177e-01 3.21312815e-01
-1.16953790e-01 -9.64150190e-01 1.09705582e-01 3.96133184e-01
2.21540153e-01 1.31971668e-02 9.70467091e-01 -7.57542849e-01
-2.69675612e-01 1.02990218e-01 -1.38834596e-01 3.07500422e-01
3.36926162e-01 -7.67402232e-01 -8.34906280e-01 -4.68880415e-01
2.62603730e-01 -9.16376829e-01 1.08742869e+00 3.30234915e-01
1.16079664e+00 -1.07865900e-01 -3.45541358e-01 5.99705756e-01
1.35493565e+00 2.57952690e-01 1.24475092e-01 3.59161884e-01
4.87092793e-01 7.55758464e-01 1.14334297e+00 5.50496638e-01
1.73339155e-02 6.35418892e-01 4.23745126e-01 1.94818035e-01
2.37166137e-01 -2.66737133e-01 4.23221201e-01 4.80455399e-01
4.07044180e-02 1.06144942e-01 -7.63166904e-01 5.07484436e-01
-2.18420577e+00 -7.97520876e-01 6.06775939e-01 2.47390532e+00
8.92220199e-01 3.02481681e-01 2.24487901e-01 4.14239876e-02
7.96739876e-01 3.04224223e-01 -1.07054353e+00 -3.72685716e-02
1.49495965e-02 -9.44793224e-02 2.04086721e-01 2.01917112e-01
-1.22021413e+00 9.70352590e-01 4.88701868e+00 1.23097825e+00
-9.41418886e-01 3.23032290e-01 6.34489119e-01 -4.98737305e-01
-2.72736073e-01 2.35952977e-02 -1.01677549e+00 5.63888848e-01
5.13890028e-01 -4.38292801e-01 4.84102994e-01 9.37666535e-01
1.01997919e-01 -5.26170917e-02 -1.08583140e+00 9.47256923e-01
2.20402822e-01 -1.05769241e+00 -5.78308664e-02 -1.21446714e-01
9.85262156e-01 -1.26006663e-01 1.48731083e-01 7.90750861e-01
3.44949245e-01 -6.44792140e-01 6.50237679e-01 3.54898363e-01
7.86257684e-01 -8.23338449e-01 5.20388544e-01 7.12289214e-01
-1.21232867e+00 -5.60781658e-01 -3.84876370e-01 1.18294887e-01
9.16213989e-02 4.32625294e-01 -5.25608063e-01 5.00693142e-01
5.81155717e-01 8.68953526e-01 -4.80154842e-01 8.11960936e-01
-2.80622691e-01 3.96438301e-01 1.18313938e-01 -2.21851304e-01
5.03399312e-01 -4.48045790e-01 7.94505775e-01 8.11578393e-01
8.00504759e-02 4.65250835e-02 6.05979085e-01 9.28018928e-01
-3.16166729e-02 2.68477499e-01 -4.18367118e-01 -1.20091014e-01
4.63272154e-01 1.20904541e+00 -7.95544505e-01 -2.35713050e-01
-3.02834481e-01 7.94643939e-01 7.67751634e-01 4.63863343e-01
-6.04591727e-01 -3.32911134e-01 4.89385784e-01 -2.66148567e-01
2.63307154e-01 2.26013198e-01 -2.66765893e-01 -1.09395397e+00
4.63118311e-03 -6.75146282e-01 6.59840763e-01 -2.76163518e-01
-1.36739182e+00 3.73466372e-01 7.04830512e-02 -1.60095394e+00
-1.02636971e-01 -1.50089532e-01 -8.61781001e-01 7.46658385e-01
-1.58949649e+00 -8.30043793e-01 -2.71664113e-01 6.89173341e-01
9.71369624e-01 -5.00259459e-01 5.91731012e-01 -7.36730471e-02
-8.02741706e-01 4.34300601e-01 3.12822193e-01 -3.69711250e-01
7.01195955e-01 -1.20787418e+00 -2.35363543e-01 6.62895620e-01
4.43359837e-03 3.68214399e-01 5.94527006e-01 -5.93449771e-01
-1.35378611e+00 -1.19650662e+00 6.50797963e-01 -1.99000016e-02
5.51173806e-01 -3.49293321e-01 -9.28441226e-01 2.74433196e-01
-9.16431323e-02 2.89744027e-02 7.37925410e-01 2.94294119e-01
-2.40923271e-01 -2.84694046e-01 -8.92196834e-01 5.97699404e-01
1.05161488e+00 -2.09283188e-01 -7.47050583e-01 4.38654870e-01
9.94507313e-01 9.81911123e-02 -4.34997320e-01 2.27788806e-01
5.72659373e-02 -7.45657980e-01 8.61641228e-01 -8.41375649e-01
5.14747381e-01 -5.76933622e-02 -4.59053740e-02 -1.78483129e+00
-3.92594397e-01 -6.37477160e-01 -3.33065271e-01 1.15504229e+00
3.14037859e-01 -5.64478159e-01 7.70010233e-01 4.88365293e-01
-2.93225378e-01 -1.14886320e+00 -8.21476102e-01 -1.28223395e+00
-2.75806397e-01 -1.71862975e-01 2.60263294e-01 1.00237572e+00
1.73597604e-01 6.81950212e-01 -5.65742075e-01 -2.97602475e-01
1.04089653e+00 5.61758578e-01 5.29582500e-01 -1.54906166e+00
-3.24045539e-01 -4.50343013e-01 1.86391011e-01 -7.65085280e-01
5.61544776e-01 -1.09652090e+00 4.54180598e-01 -1.32407320e+00
3.79223824e-01 -4.91663843e-01 -8.25848460e-01 6.77974463e-01
-4.63972986e-01 -2.66850770e-01 3.72254759e-01 2.40680650e-01
-1.15686774e+00 1.12736762e+00 1.19831479e+00 -3.06706935e-01
-7.40153551e-01 2.58681297e-01 -1.06839359e+00 7.10363984e-01
7.30705023e-01 -6.54821754e-01 -7.78943062e-01 -2.01388478e-01
-2.09316477e-01 7.21883550e-02 -7.63299316e-02 -8.95350039e-01
3.96934867e-01 -4.21189606e-01 7.30416402e-02 -4.28723991e-01
3.76251101e-01 -8.92275393e-01 -3.53444517e-01 5.70557892e-01
-6.00537956e-01 -6.33065164e-01 -1.90613568e-01 9.94252384e-01
-1.85173720e-01 -5.69170773e-01 9.45299327e-01 -6.17040843e-02
-9.24012244e-01 4.04006064e-01 1.22374542e-01 3.63326699e-01
1.36106622e+00 -1.73461989e-01 -3.34002860e-02 -6.99500591e-02
-6.27113819e-01 8.79662991e-01 1.66963503e-01 6.11278415e-01
6.87250674e-01 -1.24158978e+00 -5.84606230e-01 -1.09745441e-02
5.27840078e-01 6.77000508e-02 3.40840071e-01 8.66155267e-01
6.20892048e-01 5.11715591e-01 -2.28695795e-01 -4.55272019e-01
-1.07329512e+00 8.43041599e-01 -1.54472189e-02 -5.10396421e-01
-5.24527311e-01 9.13906515e-01 2.75637001e-01 -4.01698440e-01
5.59256375e-01 1.65088922e-01 -4.78986800e-01 5.46145678e-01
4.62373346e-01 6.08549535e-01 -4.20256965e-02 -3.79292369e-01
-2.75746882e-01 3.32472742e-01 -1.00316949e-01 1.41400723e-02
1.57892632e+00 -1.81477174e-01 2.14626715e-01 6.41051412e-01
1.18033504e+00 -4.74859297e-01 -1.55023181e+00 -9.49534953e-01
2.22630844e-01 -4.64793801e-01 2.40472719e-01 -5.96626759e-01
-1.06235135e+00 7.30088651e-01 4.09288555e-01 -9.26729441e-02
1.25817049e+00 -1.32173663e-02 6.89675152e-01 5.67903996e-01
5.31959355e-01 -1.56608796e+00 5.67871571e-01 5.15865803e-01
6.18029714e-01 -1.49389040e+00 -2.05472738e-01 -1.48947075e-01
-1.06184506e+00 6.67752504e-01 8.93118739e-01 9.58879665e-02
5.82419097e-01 -3.49366032e-02 -3.38909566e-01 -1.62723780e-01
-9.35100257e-01 -5.08332312e-01 3.55909437e-01 5.10468841e-01
-1.14608444e-02 7.98552576e-03 -3.53656948e-01 1.03435850e+00
3.88682187e-01 9.16692168e-02 2.55266149e-02 1.25889301e+00
-7.44565487e-01 -1.00066888e+00 -3.07880849e-01 7.34569252e-01
8.73422101e-02 -4.54717455e-03 -1.70682117e-01 2.22068518e-01
1.09351292e-01 8.78832757e-01 -1.71686918e-01 -2.85387307e-01
2.35427201e-01 4.19739157e-01 2.63476577e-02 -1.02904153e+00
-5.17522216e-01 2.24773630e-01 -3.20279092e-01 -5.71703732e-01
-4.06103641e-01 -5.26079535e-01 -1.07561195e+00 4.90402728e-01
-3.73358250e-01 3.97128016e-01 2.44481564e-01 1.15093863e+00
2.86074549e-01 6.99592412e-01 1.02865052e+00 -7.64895380e-01
-1.02577710e+00 -7.28908241e-01 -8.70403349e-01 3.62790763e-01
1.26740158e-01 -1.06027484e+00 -3.48607719e-01 -1.10815033e-01] | [10.05685806274414, 3.2963805198669434] |
aa6de50b-3d68-45eb-b764-3dd25ea7a3f0 | density-open-domain-dialogue-evaluation | 2305.04720 | null | https://arxiv.org/abs/2305.04720v2 | https://arxiv.org/pdf/2305.04720v2.pdf | DEnsity: Open-domain Dialogue Evaluation Metric using Density Estimation | Despite the recent advances in open-domain dialogue systems, building a reliable evaluation metric is still a challenging problem. Recent studies proposed learnable metrics based on classification models trained to distinguish the correct response. However, neural classifiers are known to make overly confident predictions for examples from unseen distributions. We propose DEnsity, which evaluates a response by utilizing density estimation on the feature space derived from a neural classifier. Our metric measures how likely a response would appear in the distribution of human conversations. Moreover, to improve the performance of DEnsity, we utilize contrastive learning to further compress the feature space. Experiments on multiple response evaluation datasets show that DEnsity correlates better with human evaluations than the existing metrics. Our code is available at https://github.com/ddehun/DEnsity. | ['Seungil Chad Lee', 'Jaegul Choo', 'Daniel Rim', 'ChaeHun Park'] | 2023-05-08 | null | null | null | null | ['dialogue-evaluation'] | ['natural-language-processing'] | [-2.19941005e-01 9.23198611e-02 -2.39631727e-01 -1.05743623e+00
-1.07953894e+00 -5.28706074e-01 6.33449256e-01 2.28415236e-01
-4.65974241e-01 1.18052971e+00 3.94315481e-01 -1.22579344e-01
1.29505217e-01 -7.75715590e-01 -2.21649751e-01 -4.32534218e-01
2.38568857e-01 7.56093800e-01 9.78882611e-02 -1.25121608e-01
3.15348268e-01 -1.86144263e-01 -1.15207696e+00 4.82653528e-01
9.90116358e-01 1.13979244e+00 -2.82654185e-02 1.00514042e+00
-2.63199687e-01 9.54487145e-01 -1.11167574e+00 -6.47320449e-01
-1.85493231e-01 -6.71945691e-01 -1.10703599e+00 -3.91606897e-01
2.10651755e-01 -6.72922313e-01 -4.93959635e-01 9.78473902e-01
3.80141199e-01 2.99493402e-01 1.06470895e+00 -1.35617685e+00
-1.04724181e+00 7.21140563e-01 1.41308354e-02 2.38955379e-01
8.87817740e-01 9.16205999e-03 1.06838703e+00 -9.68822360e-01
1.78282991e-01 1.31116414e+00 4.21549290e-01 1.00043046e+00
-1.09698021e+00 -5.77532887e-01 -1.99993372e-01 3.40212554e-01
-1.16136301e+00 -4.13190305e-01 4.44666058e-01 -2.99585402e-01
6.16675794e-01 2.67649800e-01 1.97810717e-02 1.41095686e+00
-1.13188162e-01 1.05452240e+00 1.24606025e+00 -2.53187478e-01
2.57592052e-01 5.14489353e-01 4.09222692e-01 4.76464719e-01
-2.19926029e-01 -1.66181073e-01 -4.67868209e-01 -5.03925145e-01
4.09891278e-01 -1.99344791e-02 -5.65651059e-01 1.49399072e-01
-6.73365772e-01 1.18090844e+00 3.26363266e-01 1.03731982e-01
-2.20131874e-01 -3.15910965e-01 3.60004127e-01 3.72315139e-01
8.20270956e-01 4.69215065e-01 -6.68137729e-01 -8.22773874e-01
-5.24572909e-01 3.02203357e-01 1.52489388e+00 7.09583759e-01
9.10491943e-01 -3.86145115e-01 -5.03341377e-01 1.29165053e+00
1.84463430e-02 2.87494093e-01 5.95497191e-01 -1.01264095e+00
4.01951730e-01 7.75074720e-01 1.01412080e-01 -8.82766902e-01
-1.51856720e-01 3.42566781e-02 -6.88457191e-01 -3.52710545e-01
8.44293654e-01 -5.18981814e-01 -3.45974028e-01 1.69739497e+00
2.92542011e-01 -1.14539333e-01 2.05289703e-02 9.59231615e-01
7.97351360e-01 7.59567082e-01 -2.76622381e-02 3.22115957e-03
8.23730767e-01 -7.45082617e-01 -6.54995143e-01 -2.25701630e-02
7.82109320e-01 -4.92723584e-01 1.35572493e+00 2.67696053e-01
-7.83384144e-01 -2.60096133e-01 -7.28216946e-01 3.22759934e-02
-2.42117807e-01 -2.19757944e-01 5.36286950e-01 6.39102459e-01
-8.00674796e-01 5.10949671e-01 -4.02723879e-01 -1.44757569e-01
1.69070080e-01 1.27009973e-01 -2.72777647e-01 -2.08481848e-01
-1.43547225e+00 1.02782142e+00 1.82205796e-01 -1.76708072e-01
-6.64779365e-01 -3.36890936e-01 -9.29156423e-01 1.76383972e-01
1.79876998e-01 -3.47376406e-01 1.88996053e+00 -7.15086043e-01
-1.66303873e+00 6.34042561e-01 -1.88629121e-01 -5.16860485e-01
5.25800467e-01 -2.49447376e-01 -2.16244340e-01 7.50747025e-02
-2.32850373e-01 4.93865401e-01 3.70222569e-01 -8.94273281e-01
-6.47006214e-01 -2.79929191e-01 2.40588114e-01 1.56715453e-01
-6.01346970e-01 -8.05973411e-02 -3.50832455e-02 -1.36487558e-01
-3.10749680e-01 -8.90461683e-01 4.18587103e-02 -1.94460407e-01
-3.00208569e-01 -9.99158680e-01 5.55244565e-01 -6.26213014e-01
1.44028568e+00 -1.94272006e+00 -1.54906988e-01 -8.86927769e-02
4.22035515e-01 2.17410520e-01 -7.77038634e-02 4.15538192e-01
5.13954878e-01 1.22578144e-01 -1.23965353e-01 -1.67731658e-01
2.94876426e-01 1.36973277e-01 -8.83877277e-02 3.97584319e-01
2.32074037e-01 4.90537107e-01 -9.35359359e-01 -3.70601714e-01
-7.44777918e-02 3.58932048e-01 -5.03991544e-01 7.37241089e-01
-3.12787265e-01 3.30511332e-01 -4.46236849e-01 3.65031481e-01
5.30517757e-01 -4.78079319e-01 6.76388517e-02 2.34096289e-01
4.00722951e-01 6.22342885e-01 -5.62966764e-01 1.40455556e+00
-5.06938934e-01 6.83944821e-01 -1.27248079e-01 -1.02508819e+00
1.23841751e+00 2.33262718e-01 6.59329295e-02 -5.56866229e-01
3.94454658e-01 1.23853311e-01 1.09236665e-01 -3.76190871e-01
7.53151238e-01 1.12529650e-01 -4.15682614e-01 7.51030743e-01
2.60633707e-01 -6.36152923e-02 1.37568116e-01 4.08620089e-01
1.30391073e+00 -3.33175451e-01 2.32513040e-01 3.05625349e-02
5.70236206e-01 -1.34009421e-01 5.54065704e-01 8.04518044e-01
-5.01526058e-01 6.44658387e-01 7.15599656e-01 -3.15452278e-01
-8.68452549e-01 -1.00070369e+00 -3.33693236e-01 1.28104985e+00
7.41564343e-03 -4.60126370e-01 -9.27611828e-01 -9.98370111e-01
6.98109716e-02 8.24953556e-01 -5.80648065e-01 -2.54793316e-01
-2.71246731e-01 -4.92367923e-01 5.52903354e-01 4.37177747e-01
4.43929046e-01 -8.31073761e-01 -1.81645662e-01 -4.05308977e-02
-5.86804152e-01 -8.66670430e-01 -6.07373297e-01 7.89248124e-02
-2.97709584e-01 -1.00987732e+00 -9.06358242e-01 -5.92986703e-01
3.85032505e-01 4.47195023e-02 1.28726411e+00 1.49768427e-01
-2.48157438e-02 4.75187778e-01 -6.40202522e-01 -2.93306470e-01
-8.23153317e-01 2.09118411e-01 1.83637902e-01 -2.62653768e-01
1.06378353e+00 -1.46732047e-01 -6.66589499e-01 5.48051536e-01
-4.69780356e-01 -3.07325214e-01 1.71846598e-01 1.15583563e+00
9.28541347e-02 -4.40269053e-01 9.44122493e-01 -9.27414894e-01
1.45419157e+00 -9.10043001e-01 -1.67834908e-01 3.32864434e-01
-6.33326113e-01 2.17424408e-01 6.03300393e-01 -5.50236046e-01
-9.85363185e-01 -4.15254116e-01 -2.08777562e-01 -5.07926568e-02
-4.41808641e-01 4.70846593e-01 2.51829624e-01 3.62781733e-01
9.17178631e-01 1.40595838e-01 1.66153163e-01 -3.44441503e-01
1.73191786e-01 1.36692417e+00 3.39919508e-01 -7.71369815e-01
1.50959760e-01 -2.40237132e-01 -7.46684194e-01 -6.47556424e-01
-1.07842207e+00 -5.52845776e-01 -2.48877347e-01 -3.52874160e-01
5.60752690e-01 -5.52865207e-01 -8.76348138e-01 3.17729652e-01
-1.25076568e+00 -3.83916140e-01 1.41599849e-01 5.44080675e-01
-5.15295565e-01 3.19242716e-01 -9.46881652e-01 -1.13259482e+00
-4.18131799e-01 -7.64403701e-01 5.99097073e-01 5.23646832e-01
-8.51888597e-01 -1.04407704e+00 3.72549772e-01 3.61931831e-01
5.63084900e-01 -2.33280957e-01 7.20158041e-01 -1.34918904e+00
-8.24524313e-02 -4.62279856e-01 -1.32345811e-01 7.27875054e-01
2.13203225e-02 -7.44743496e-02 -1.11807632e+00 -5.36013134e-02
1.49494797e-01 -9.65975225e-01 6.43571317e-01 1.36286169e-01
1.28744709e+00 -6.57755852e-01 1.32585187e-02 9.71067399e-02
6.95013463e-01 1.13145076e-01 5.07766962e-01 8.18974450e-02
1.04502447e-01 7.70484030e-01 7.15566576e-01 8.76321375e-01
6.04065955e-01 4.04719710e-01 6.48146942e-02 3.62092346e-01
2.26298392e-01 -4.90607500e-01 3.26338172e-01 9.51976359e-01
3.96740139e-01 -5.21761537e-01 -9.27687049e-01 4.04276878e-01
-1.74258101e+00 -8.56149077e-01 1.85611084e-01 2.24881911e+00
1.22421956e+00 1.25577047e-01 2.85049409e-01 3.89403105e-02
8.14204156e-01 -7.34303519e-02 -6.43942416e-01 -6.79775059e-01
-1.00205122e-02 9.14318487e-02 -6.13773707e-03 6.71290755e-01
-8.83021116e-01 5.58447778e-01 6.33558607e+00 6.53562129e-01
-8.36777389e-01 6.54647499e-02 9.69269276e-01 -1.06204785e-02
-2.80052006e-01 -2.79833823e-01 -8.62688184e-01 6.70067966e-01
1.25581038e+00 -5.76860666e-01 1.60485700e-01 1.05758703e+00
-2.50252694e-01 -2.13570938e-01 -1.25725818e+00 9.42377329e-01
6.81325942e-02 -9.04037774e-01 -1.83965564e-01 -3.69692370e-02
4.55556244e-01 -4.92909588e-02 8.55534524e-02 8.94004881e-01
6.68263555e-01 -1.05492985e+00 1.08175948e-01 5.57494521e-01
6.16338253e-01 -8.09567213e-01 8.84630859e-01 8.13970864e-01
-4.78481174e-01 1.12740584e-01 -6.43795013e-01 -2.60040224e-01
-1.47803679e-01 4.98079538e-01 -1.52428842e+00 -1.14852019e-01
4.08651680e-01 4.19471920e-01 -4.14809376e-01 9.69601631e-01
-2.87708908e-01 7.50927031e-01 -2.20336914e-01 -7.30768144e-01
1.19448587e-01 1.10247033e-02 2.03705773e-01 1.33321702e+00
2.65086770e-01 2.62234688e-01 2.79645294e-01 8.73465776e-01
-2.58260429e-01 2.00977087e-01 -5.95152617e-01 -1.18642963e-01
7.53229320e-01 1.24127543e+00 -2.31532697e-02 -2.27805465e-01
-4.80523348e-01 9.41109717e-01 7.35374033e-01 1.92551330e-01
-9.36225295e-01 -5.09587228e-01 6.05217814e-01 -1.29375741e-01
-2.22186103e-01 -2.00029444e-02 -1.29358500e-01 -1.18706858e+00
9.36875045e-02 -8.42941761e-01 4.37107146e-01 -4.45573390e-01
-1.80983090e+00 8.32002103e-01 -6.87393099e-02 -1.12862933e+00
-8.39917839e-01 -6.21703148e-01 -4.99501973e-01 1.02296412e+00
-9.95604396e-01 -2.86043733e-01 -3.46141756e-01 5.00490129e-01
7.00321138e-01 -1.99623883e-01 1.28938711e+00 1.79980963e-01
-4.80315834e-01 1.03105354e+00 3.51658277e-02 3.87670845e-01
1.00706816e+00 -1.38538098e+00 1.23470046e-01 7.83931240e-02
-7.25066513e-02 5.80057919e-01 5.92066646e-01 -2.88736969e-01
-9.84773636e-01 -5.80568671e-01 8.84768128e-01 -5.40799499e-01
6.18885040e-01 -2.27477506e-01 -1.28110456e+00 2.70626664e-01
3.97363752e-01 -2.72620618e-01 1.19497514e+00 4.96282518e-01
-5.01662374e-01 1.46928489e-01 -1.22776079e+00 3.32056999e-01
5.61670482e-01 -8.09818685e-01 -5.73280334e-01 3.47218245e-01
4.36385721e-01 -3.83826047e-01 -9.52436507e-01 1.93530232e-01
5.56895733e-01 -1.02559519e+00 5.12682378e-01 -7.04801261e-01
6.35727763e-01 3.67792040e-01 -3.10058177e-01 -1.65009451e+00
-1.33289084e-01 -5.80140233e-01 -4.67872620e-01 1.17345119e+00
7.06658840e-01 -5.69251180e-01 8.30442965e-01 1.17254102e+00
6.94313496e-02 -9.07052457e-01 -8.32357168e-01 -6.26545548e-01
4.22959626e-01 -7.58322701e-02 4.93698806e-01 9.55028117e-01
6.95188999e-01 6.24242902e-01 -4.63732928e-01 -2.49844968e-01
3.49270046e-01 -2.25141615e-01 9.29111719e-01 -1.23319137e+00
-3.87338847e-01 -3.41977298e-01 -3.09604079e-01 -1.49614000e+00
5.54001689e-01 -7.11702108e-01 3.19445789e-01 -1.36326134e+00
4.74494517e-01 -3.56883407e-01 -2.66885221e-01 2.46646047e-01
-3.83773983e-01 -1.55978054e-01 -9.12211761e-02 1.12974577e-01
-8.41855049e-01 8.13988447e-01 1.13272655e+00 -1.47007063e-01
5.83832338e-02 2.18969166e-01 -6.62282944e-01 7.04963803e-01
1.23358977e+00 -4.83808607e-01 -3.66226256e-01 -3.86965424e-01
-8.60489383e-02 4.22252148e-01 -3.47599667e-03 -8.55884075e-01
3.95005912e-01 -1.66471899e-01 3.19714278e-01 -5.18748224e-01
5.03544211e-01 -2.01196119e-01 -5.75909674e-01 4.46535870e-02
-8.92163575e-01 8.76787826e-02 -1.44218579e-01 5.07531226e-01
-4.80600595e-01 -4.84738797e-01 7.29456663e-01 -5.41928373e-02
-5.77198267e-01 2.78066248e-01 -2.62834072e-01 4.54778910e-01
6.16898239e-01 1.40025824e-01 -6.84735835e-01 -9.48756099e-01
-4.84359980e-01 3.71234685e-01 2.21593887e-01 4.68976974e-01
8.27154756e-01 -1.17716706e+00 -9.71506178e-01 -1.71190470e-01
2.16516316e-01 -1.92899913e-01 1.90330341e-01 4.70108986e-01
-3.10135126e-01 3.89821708e-01 6.06858544e-02 -6.01445079e-01
-1.21070206e+00 8.48829672e-02 3.33254635e-01 -1.89106360e-01
-1.05165362e-01 1.01107419e+00 5.62375970e-02 -7.34431624e-01
4.67213750e-01 -4.15604264e-02 -1.65341496e-01 -2.20084429e-01
9.57117617e-01 4.35640633e-01 -1.51599497e-01 -3.23047817e-01
-1.37743413e-01 -2.83931345e-01 -4.71998543e-01 -2.62590408e-01
9.35809612e-01 -8.82467106e-02 1.18026994e-01 6.60610914e-01
1.51008177e+00 -3.08098048e-01 -1.17319894e+00 -5.70521057e-01
1.06402583e-01 -6.99248075e-01 -2.29969054e-01 -9.09758747e-01
-4.80993599e-01 9.40478563e-01 3.07562917e-01 5.38629234e-01
7.12778986e-01 -1.89835522e-02 8.04497600e-01 6.69893384e-01
4.04070675e-01 -1.13730085e+00 2.92967588e-01 1.02495110e+00
8.82804632e-01 -1.66624975e+00 -3.60839516e-01 -8.39175954e-02
-9.23914135e-01 1.06263185e+00 1.00079978e+00 3.83882714e-03
6.55939043e-01 2.79141933e-01 2.40350038e-01 1.63915932e-01
-1.07111704e+00 2.42374390e-01 2.02495486e-01 5.99125743e-01
9.32496727e-01 3.17382276e-01 -3.90982658e-01 8.42991650e-01
-3.43295157e-01 -1.20159350e-01 6.28884196e-01 6.29557610e-01
-6.11525416e-01 -1.09622467e+00 -2.55722012e-02 6.38711035e-01
-4.43084627e-01 -2.61077136e-02 -6.87738955e-01 2.88169444e-01
-6.25755429e-01 1.32793725e+00 1.58916926e-03 -8.12391102e-01
1.62651405e-01 2.92426437e-01 3.27515125e-01 -7.89902508e-01
-3.41976345e-01 -3.98296326e-01 4.43889976e-01 -2.49329120e-01
-1.28041953e-01 -4.78584528e-01 -1.05923617e+00 -6.75579131e-01
-3.77799124e-01 6.00563169e-01 4.09154773e-01 6.99416697e-01
1.29066437e-01 1.10911757e-01 9.39152658e-01 -3.56367916e-01
-1.34979582e+00 -1.51808679e+00 -6.00964487e-01 5.38513660e-01
1.76699042e-01 -5.99132061e-01 -6.79474235e-01 -4.35811073e-01] | [12.75041675567627, 7.991774082183838] |
a84f3e3c-3131-4a30-96be-ca4c362f3706 | idpl-intra-subdomain-adaptation-adversarial | 2210.03435 | null | https://arxiv.org/abs/2210.03435v2 | https://arxiv.org/pdf/2210.03435v2.pdf | IDPL: Intra-subdomain adaptation adversarial learning segmentation method based on Dynamic Pseudo Labels | Unsupervised domain adaptation(UDA) has been applied to image semantic segmentation to solve the problem of domain offset. However, in some difficult categories with poor recognition accuracy, the segmentation effects are still not ideal. To this end, in this paper, Intra-subdomain adaptation adversarial learning segmentation method based on Dynamic Pseudo Labels(IDPL) is proposed. The whole process consists of 3 steps: Firstly, the instance-level pseudo label dynamic generation module is proposed, which fuses the class matching information in global classes and local instances, thus adaptively generating the optimal threshold for each class, obtaining high-quality pseudo labels. Secondly, the subdomain classifier module based on instance confidence is constructed, which can dynamically divide the target domain into easy and difficult subdomains according to the relative proportion of easy and difficult instances. Finally, the subdomain adversarial learning module based on self-attention is proposed. It uses multi-head self-attention to confront the easy and difficult subdomains at the class level with the help of generated high-quality pseudo labels, so as to focus on mining the features of difficult categories in the high-entropy region of target domain images, which promotes class-level conditional distribution alignment between the subdomains, improving the segmentation performance of difficult categories. For the difficult categories, the experimental results show that the performance of IDPL is significantly improved compared with other latest mainstream methods. | ['Xuzhou Fu', 'Jie Gao', 'Jian Yu', 'Weilun Zhang', 'XueWei Li'] | 2022-10-07 | null | null | null | null | ['subdomain-adaptation'] | ['methodology'] | [ 4.58174765e-01 2.49712855e-01 -3.85233372e-01 -3.07933360e-01
-5.95014393e-01 -2.69033492e-01 1.07944034e-01 -7.17400983e-02
-3.37328434e-01 5.26270330e-01 -9.03448313e-02 5.48400879e-02
-1.11986808e-01 -9.59829092e-01 -2.42433444e-01 -9.52799857e-01
4.41190511e-01 7.05500901e-01 4.91389781e-01 -7.27303848e-02
1.48498774e-01 -5.71030863e-02 -1.32959533e+00 2.15748295e-01
1.56585133e+00 1.36118567e+00 1.54917002e-01 4.90229689e-02
-4.17127281e-01 5.22017241e-01 -5.64935267e-01 -1.29576623e-01
1.44055620e-01 -6.95546865e-01 -8.39647353e-01 2.60375410e-01
2.30516233e-02 -1.19196318e-01 -3.51325274e-02 1.45325708e+00
4.42700952e-01 2.50104398e-01 1.02192140e+00 -1.22524774e+00
-5.36849678e-01 4.40189809e-01 -6.31990135e-01 3.16425115e-01
-5.93301319e-02 2.44834185e-01 5.66303372e-01 -4.27123904e-01
3.45464826e-01 1.27411485e+00 2.72147477e-01 7.04094291e-01
-9.95288610e-01 -1.16739416e+00 3.29008132e-01 4.58815783e-01
-1.35728049e+00 2.96351593e-02 1.11065412e+00 -4.41023648e-01
2.65576899e-01 -1.96985248e-02 5.11585593e-01 1.06398463e+00
-6.82604462e-02 9.68486905e-01 1.42734480e+00 -2.40770370e-01
3.14247459e-01 3.92307788e-01 2.40632012e-01 2.60533810e-01
-9.93958488e-02 1.06322587e-01 1.41491413e-01 1.67695493e-01
5.87316871e-01 -1.32100999e-01 -1.27042443e-01 -3.25571209e-01
-8.23416173e-01 8.48717749e-01 6.70934498e-01 4.46951509e-01
-2.33670518e-01 -7.48680770e-01 5.09187579e-01 7.68285841e-02
6.33870065e-01 1.33840561e-01 -5.81163764e-01 9.96850058e-02
-8.12674820e-01 1.43107980e-01 4.87346858e-01 8.51540208e-01
6.97289169e-01 -5.65782003e-02 -3.12291920e-01 1.07895231e+00
2.80862570e-01 3.13251495e-01 1.00818896e+00 -5.69071174e-01
5.81234157e-01 9.94448125e-01 -3.64476860e-01 -1.04651570e+00
-1.93716198e-01 -4.67575669e-01 -1.01676488e+00 1.73978537e-01
3.17672431e-01 -1.79735988e-01 -1.24736571e+00 1.77604353e+00
6.81061924e-01 4.61215466e-01 2.17242062e-01 9.03955519e-01
9.36675608e-01 7.05120683e-01 5.07903636e-01 -4.14857537e-01
1.40082622e+00 -8.65216076e-01 -6.82095706e-01 -4.68434185e-01
2.08114564e-01 -4.93719012e-01 1.05325150e+00 2.41984144e-01
-6.69771850e-01 -1.07005000e+00 -1.05505359e+00 2.35134378e-01
-4.93713319e-01 -8.56776461e-02 2.26242617e-01 5.17704308e-01
-4.41319793e-01 2.60668010e-01 -4.28416699e-01 3.96491848e-02
1.01745975e+00 2.28803843e-01 1.13982543e-01 -7.57340342e-03
-1.49604106e+00 5.25631547e-01 1.12194276e+00 -2.64898181e-01
-7.95121729e-01 -5.53437769e-01 -1.01144409e+00 -1.58781018e-02
4.48418170e-01 -3.33182186e-01 1.01516247e+00 -1.56447434e+00
-1.50973141e+00 1.03267050e+00 1.57558233e-01 -2.93469459e-01
6.05372250e-01 2.73395091e-01 -3.98248166e-01 1.12353973e-01
3.78891826e-01 9.85294342e-01 1.12900746e+00 -1.26472712e+00
-9.04900014e-01 -6.29009426e-01 -2.90569514e-01 7.47550488e-01
-3.42136085e-01 -4.31260526e-01 -3.42355877e-01 -7.75433481e-01
2.80376792e-01 -6.08420253e-01 -1.67568311e-01 -4.70368028e-01
-3.72719169e-01 -5.68108141e-01 1.02903569e+00 -7.47586012e-01
1.15761292e+00 -2.36149096e+00 2.23838463e-01 3.69288951e-01
1.42015457e-01 4.89686430e-01 1.44171134e-01 -6.98233724e-01
-3.22470009e-01 -2.99697299e-03 -8.51658463e-01 3.18905741e-01
-2.33105302e-01 1.00080661e-01 -1.21816121e-01 7.88511485e-02
3.77178073e-01 6.23792827e-01 -9.28142667e-01 -8.54953766e-01
3.56400400e-01 -2.22038068e-02 -3.03058654e-01 3.64682555e-01
-3.26703638e-01 8.13352168e-01 -7.43074119e-01 7.02075481e-01
1.01101887e+00 5.19840345e-02 -8.31615329e-02 -1.14307895e-01
3.89645517e-01 -4.62853676e-03 -1.20130908e+00 1.34178054e+00
-3.22569728e-01 -2.66195331e-02 1.07980430e-01 -1.30578935e+00
1.11929202e+00 5.22354543e-02 4.77272213e-01 -9.82239425e-01
4.18790221e-01 3.07289422e-01 8.03964883e-02 -5.64632058e-01
-2.23463271e-02 -3.89825135e-01 -3.76095176e-01 6.65255636e-02
4.98013571e-02 -3.41378540e-01 -1.21580645e-01 -3.09916977e-02
4.82595950e-01 2.54023284e-01 1.51417106e-01 -1.49513870e-01
9.30064023e-01 4.87660430e-02 8.40172529e-01 1.17007822e-01
-6.61122501e-01 5.19224405e-01 2.40264460e-01 4.72524464e-02
-8.51998389e-01 -1.03747916e+00 -3.77822727e-01 7.65946567e-01
9.05997336e-01 4.30088103e-01 -1.29098332e+00 -1.15149367e+00
-1.17349230e-01 6.78097248e-01 -6.60656095e-01 -7.58493006e-01
-3.92590135e-01 -9.41209137e-01 3.79852176e-01 5.92241466e-01
1.33180344e+00 -1.51484728e+00 -1.12140380e-01 2.59777427e-01
-3.79445851e-01 -9.26914096e-01 -5.08829653e-01 5.87340221e-02
-8.40648890e-01 -1.03157711e+00 -9.48161066e-01 -1.24331093e+00
7.45415032e-01 -8.28046799e-02 6.72729492e-01 -2.39828765e-01
-7.31494948e-02 -1.05642594e-01 -3.41048986e-01 -2.53880292e-01
-4.82201725e-01 1.14449561e-01 -1.53155833e-01 3.88292044e-01
6.50699735e-01 -4.75765973e-01 -4.43572462e-01 5.30670166e-01
-9.14771020e-01 -1.32118845e-02 6.72733724e-01 9.21339095e-01
9.17879939e-01 5.82487822e-01 8.88271809e-01 -1.00842810e+00
4.97226506e-01 -7.97692835e-01 -3.27297240e-01 1.03612002e-02
-8.22482109e-01 -3.06284338e-01 9.03654218e-01 -7.86668181e-01
-1.41194355e+00 -1.85282864e-02 -2.91476548e-01 -3.48333746e-01
-5.73754728e-01 1.58792600e-01 -8.90772939e-01 2.64403731e-01
5.19986391e-01 5.16319454e-01 2.00210452e-01 -1.83612451e-01
2.54097849e-01 1.00392580e+00 6.04501903e-01 -4.95818764e-01
6.24211848e-01 1.89204499e-01 -5.17091334e-01 -3.37971330e-01
-8.73773634e-01 -3.78361195e-01 -6.10958993e-01 -1.78583637e-01
1.24392259e+00 -8.06308985e-01 -2.23769456e-01 9.54158366e-01
-6.10505462e-01 -4.84728783e-01 -5.42303622e-01 3.29925120e-01
-5.81118584e-01 3.79617780e-01 -4.75296170e-01 -4.37329441e-01
-2.99106807e-01 -1.36211181e+00 7.95823455e-01 1.04537463e+00
2.38200333e-02 -8.98673534e-01 -4.08869833e-01 6.91974282e-01
4.50705849e-02 2.83856273e-01 8.71597588e-01 -9.77892101e-01
-2.36689761e-01 -3.89025137e-02 -4.37920272e-01 8.32073808e-01
1.63554519e-01 -4.86515790e-01 -9.23707962e-01 -7.79059008e-02
2.51227021e-01 -4.93859679e-01 6.90214336e-01 5.51564455e-01
1.52874243e+00 -2.10355967e-01 -4.80541050e-01 5.89074433e-01
1.14724398e+00 6.12336159e-01 6.50079966e-01 3.98066968e-01
8.08790684e-01 6.30735874e-01 1.19765401e+00 -4.08514589e-02
3.12217146e-01 4.20625120e-01 3.75997365e-01 -1.55769825e-01
-2.96836078e-01 -2.19851673e-01 2.28960156e-01 6.95271969e-01
3.85358393e-01 -3.21613736e-02 -7.03768253e-01 6.80312634e-01
-1.49298203e+00 -7.59383500e-01 1.49825230e-01 1.97622120e+00
1.11158729e+00 6.06409132e-01 1.60433620e-01 2.42455453e-01
1.15242028e+00 1.28214180e-01 -1.02098954e+00 -1.61331490e-01
2.50327121e-02 2.64370352e-01 3.43745798e-01 2.87513316e-01
-1.42239726e+00 1.00037003e+00 4.06630754e+00 1.75206387e+00
-1.05228162e+00 2.39754066e-01 1.07109189e+00 4.92242903e-01
-7.71624967e-02 -1.54699147e-01 -7.72851646e-01 1.17548454e+00
4.85981494e-01 -6.40093759e-02 2.66529083e-01 1.09687340e+00
-3.83040868e-02 -2.73034424e-02 -4.54583555e-01 8.13763440e-01
2.33900119e-02 -5.75947225e-01 3.32146809e-02 -9.86888036e-02
8.26792955e-01 -4.05612230e-01 2.85095841e-01 6.42149627e-01
2.11556286e-01 -8.41511250e-01 4.55805749e-01 1.19851328e-01
1.10299146e+00 -1.07418191e+00 8.34821880e-01 6.87377989e-01
-1.11408484e+00 -3.02179158e-01 -3.27031732e-01 3.56707215e-01
-1.64945945e-01 5.12384713e-01 -6.84594214e-01 6.85444355e-01
7.24260688e-01 7.50701964e-01 -4.22104567e-01 7.25255728e-01
-2.56083161e-01 7.50565469e-01 -3.12274247e-01 2.41838545e-01
3.34283561e-01 -4.60784763e-01 4.69402730e-01 9.35926080e-01
-8.19302350e-02 2.54454494e-01 3.96651208e-01 8.35219920e-01
-2.51015034e-02 2.15211138e-01 -2.90074676e-01 3.52968067e-01
5.31228662e-01 1.13353097e+00 -9.14526641e-01 -6.50350332e-01
1.21329315e-01 1.07692850e+00 2.83217616e-02 1.63871914e-01
-1.00986338e+00 -6.27196252e-01 2.38543794e-01 -8.57494846e-02
1.41161010e-01 4.34703141e-01 -6.88160956e-01 -9.78371024e-01
1.85303744e-02 -9.77173924e-01 8.79031539e-01 -5.34177780e-01
-1.37768149e+00 4.77128893e-01 7.43767619e-02 -1.43537259e+00
6.34513274e-02 -3.20719063e-01 -7.80093431e-01 9.79138553e-01
-1.52079809e+00 -1.05743539e+00 -5.48598766e-01 6.74714923e-01
9.65777338e-01 -2.22538680e-01 5.68369389e-01 4.26059216e-01
-7.76189029e-01 9.17474508e-01 1.15431540e-01 3.11306506e-01
6.42740130e-01 -1.17349386e+00 -5.82342967e-02 6.27581775e-01
-4.84156817e-01 7.35096186e-02 1.73516884e-01 -8.22847664e-01
-1.50470421e-01 -1.35892582e+00 3.11756641e-01 -8.07139054e-02
1.81455269e-01 -1.13748640e-01 -1.21405900e+00 2.82583207e-01
-2.35797420e-01 -6.21453449e-02 4.10547733e-01 -3.50309104e-01
-3.29386964e-02 -2.83779144e-01 -1.68526089e+00 5.15248716e-01
8.20619106e-01 -3.23215097e-01 -8.55907083e-01 2.43917376e-01
7.83076346e-01 -6.60491407e-01 -8.24641228e-01 7.78053045e-01
-1.19244019e-02 -5.57872713e-01 9.53783214e-01 -2.15571061e-01
5.65235019e-01 -4.25895602e-01 3.68247181e-01 -1.30581343e+00
-3.81940544e-01 1.61652103e-01 1.54367670e-01 1.58197272e+00
1.00931846e-01 -7.52938092e-01 8.98533106e-01 2.13943630e-01
-1.57151759e-01 -8.51021647e-01 -9.91886139e-01 -5.55434704e-01
2.91120142e-01 -2.65274107e-01 6.23125076e-01 1.17039764e+00
-4.21161622e-01 3.95360321e-01 3.38777542e-01 1.85288414e-01
5.01922786e-01 1.96801558e-01 3.34070891e-01 -1.04586744e+00
-8.44849050e-02 -5.86210668e-01 -2.85724819e-01 -9.99464095e-01
3.02894354e-01 -1.08679843e+00 1.35697722e-01 -1.05770540e+00
1.19040988e-01 -9.11143661e-01 -4.69797164e-01 3.73229861e-01
-7.16263831e-01 2.22919777e-01 -1.18813440e-01 2.36107543e-01
-4.93094772e-01 6.91786706e-01 1.63920712e+00 -5.60483634e-01
-2.41272286e-01 2.16030225e-01 -8.46784294e-01 7.31931031e-01
8.72536182e-01 -3.80480677e-01 -6.24204457e-01 1.94443360e-01
-7.47688830e-01 -8.74849632e-02 1.74420819e-01 -1.12451935e+00
-1.70230672e-01 -3.06326360e-01 5.96310258e-01 -4.62546527e-01
-1.16293982e-01 -8.04517746e-01 -3.38614553e-01 5.06619096e-01
-3.39100122e-01 -7.74744213e-01 2.25789890e-01 5.84880650e-01
-3.31685781e-01 -3.33027750e-01 1.41042864e+00 -1.96832910e-01
-1.06846905e+00 4.88076866e-01 -6.23068810e-02 5.94910860e-01
1.47071719e+00 -5.03973782e-01 1.61504284e-01 8.48882347e-02
-8.29692245e-01 5.83918929e-01 2.90779173e-01 4.39931303e-01
4.80690628e-01 -1.34009361e+00 -6.30469799e-01 3.86168569e-01
1.50447518e-01 7.64004886e-01 7.47092426e-01 4.98244166e-01
-3.16921800e-01 -2.49369815e-01 -3.53357077e-01 -7.22721636e-01
-9.95860577e-01 7.22460270e-01 5.03185332e-01 -5.29374301e-01
-2.92559534e-01 1.01100910e+00 4.98398036e-01 -6.24260008e-01
2.80925240e-02 1.66871533e-01 -7.83868551e-01 3.20031106e-01
3.43959689e-01 1.97817162e-01 -2.18416750e-01 -7.16642678e-01
-2.43955284e-01 7.28736579e-01 -2.71048576e-01 2.80626088e-01
8.28465819e-01 -1.59741893e-01 3.89715061e-02 1.08183399e-01
1.22506225e+00 -2.12781638e-01 -1.46064627e+00 -1.50934279e-01
-2.98955202e-01 -2.11989194e-01 -8.58910009e-02 -1.05185604e+00
-1.23062992e+00 9.62339818e-01 1.14687026e+00 7.94443861e-02
1.58047509e+00 -3.42990458e-03 1.26193655e+00 -5.49286664e-01
1.74455643e-01 -1.40448499e+00 9.99207199e-02 3.87241066e-01
3.13392550e-01 -1.28631735e+00 -4.17638361e-01 -5.19553125e-01
-9.03199971e-01 6.58745408e-01 1.26675510e+00 -2.97871917e-01
5.32164335e-01 -4.33751829e-02 1.38011605e-01 1.43824324e-01
6.05583452e-02 -1.78937986e-01 2.11741641e-01 9.23551559e-01
-2.83781856e-01 2.51175076e-01 -6.06396675e-01 1.08337045e+00
9.51724406e-03 -1.99985594e-01 -1.89352930e-02 6.03273034e-01
-4.32611674e-01 -1.02127218e+00 -5.20275712e-01 5.55620670e-01
-3.36452961e-01 2.45738495e-02 -2.32059017e-01 4.49679703e-01
9.63204086e-01 7.52952576e-01 2.54899681e-01 -6.54756665e-01
1.98712498e-01 1.18863866e-01 9.92648154e-02 -6.05146527e-01
-4.00085479e-01 -4.51430418e-02 -3.42433423e-01 -1.43753216e-01
-2.48512611e-01 -5.94970405e-01 -1.60820127e+00 2.88264722e-01
-4.72436726e-01 2.15839058e-01 1.37690514e-01 1.09347200e+00
7.13303983e-02 7.08967149e-01 9.50101972e-01 -6.08763158e-01
-5.54660559e-01 -1.04030192e+00 -5.01817226e-01 7.16612279e-01
4.37893234e-02 -8.53127539e-01 -3.52413684e-01 1.08952641e-01] | [9.691176414489746, 1.378183364868164] |
15b8d38c-dd79-4407-8eb8-9871c5f41719 | adapting-pretrained-text-to-text-models-for | 2209.10052 | null | https://arxiv.org/abs/2209.10052v2 | https://arxiv.org/pdf/2209.10052v2.pdf | Adapting Pretrained Text-to-Text Models for Long Text Sequences | We present an empirical study of adapting an existing pretrained text-to-text model for long-sequence inputs. Through a comprehensive study along three axes of the pretraining pipeline -- model architecture, optimization objective, and pretraining corpus, we propose an effective recipe to build long-context models from existing short-context models. Specifically, we replace the full attention in transformers with pooling-augmented blockwise attention, and pretrain the model with a masked-span prediction task with spans of varying length. In terms of the pretraining corpus, we find that using randomly concatenated short-documents from a large open-domain corpus results in better performance than using existing long document corpora which are typically limited in their domain coverage. With these findings, we build a long-context model that achieves competitive performance on long-text QA tasks and establishes the new state of the art on five long-text summarization datasets, often outperforming previous methods with larger model sizes. Our code has been released at https://github.com/facebookresearch/bart_ls. | ['Wen-tau Yih', 'Yashar Mehdad', 'Shubham Toshniwal', 'Anchit Gupta', 'Wenhan Xiong'] | 2022-09-21 | null | null | null | null | ['long-range-modeling'] | ['natural-language-processing'] | [ 5.07340789e-01 3.07694018e-01 -2.87526786e-01 -4.77767259e-01
-1.44828415e+00 -6.70158744e-01 6.01892829e-01 3.80436182e-02
-5.59986472e-01 9.06982124e-01 8.96620870e-01 -5.23681104e-01
2.28149101e-01 -2.72438705e-01 -9.44363296e-01 -2.59498447e-01
2.51454592e-01 6.78724527e-01 -7.96120837e-02 -3.83939922e-01
4.40346122e-01 -2.79663414e-01 -1.02777696e+00 7.03991234e-01
1.24472821e+00 5.77626288e-01 4.35058624e-01 1.03845119e+00
-1.80266231e-01 8.25503826e-01 -8.52382123e-01 -6.69029593e-01
3.67539078e-02 -5.96166134e-01 -1.39692199e+00 -8.25959966e-02
8.71337652e-01 -6.14438117e-01 -3.49532098e-01 5.67081332e-01
7.38673508e-01 1.77899197e-01 5.23499548e-01 -5.58054328e-01
-9.88766968e-01 1.18923151e+00 -3.06340218e-01 3.05199832e-01
2.29286030e-01 8.62739533e-02 1.62778163e+00 -7.86226928e-01
6.38718367e-01 1.24945211e+00 7.01828122e-01 6.42841637e-01
-1.37918258e+00 -2.43352279e-01 3.03713679e-01 1.75086051e-01
-7.74644613e-01 -7.28758395e-01 4.81081098e-01 -1.30063817e-01
1.73460066e+00 1.59067094e-01 2.50730932e-01 1.42228675e+00
3.46328169e-01 1.11092806e+00 5.28569341e-01 -5.63012838e-01
-2.13884175e-01 -4.57528293e-01 4.62513626e-01 3.82278323e-01
1.63175642e-01 -3.78578454e-01 -4.69239771e-01 -7.69437850e-02
9.18937102e-02 -3.40922087e-01 -3.11558634e-01 1.70844689e-01
-1.26944339e+00 8.50388229e-01 2.61091977e-01 2.35773116e-01
-1.89608619e-01 3.21047753e-01 7.94845104e-01 3.92864764e-01
7.43613303e-01 8.44266415e-01 -8.80146563e-01 -4.34228599e-01
-1.15963471e+00 3.79428446e-01 9.19110358e-01 9.97691274e-01
4.37370747e-01 -9.34201702e-02 -5.12149692e-01 1.12832057e+00
-2.41773143e-01 4.16524380e-01 7.74726033e-01 -9.45939660e-01
1.00970352e+00 3.27265203e-01 -8.66009202e-03 -3.50821793e-01
-3.03627640e-01 -5.35425842e-01 -5.94040036e-01 -6.31255746e-01
3.04613829e-01 -4.61968392e-01 -9.14364040e-01 1.69364274e+00
-1.89407483e-01 -1.51609123e-01 2.75549591e-01 4.63822365e-01
8.90157163e-01 9.98663962e-01 -1.40711740e-02 -1.77434281e-01
1.22299087e+00 -1.72866046e+00 -7.95303404e-01 -6.35890782e-01
1.00530541e+00 -7.73590624e-01 1.52088583e+00 2.25521013e-01
-1.34080660e+00 -5.32073617e-01 -9.75490332e-01 -7.75174916e-01
-2.13230312e-01 4.71529901e-01 3.05136979e-01 1.21696055e-01
-1.20014727e+00 7.17778206e-01 -6.56285465e-01 -4.99663264e-01
3.89750212e-01 -3.27434987e-02 -3.45394947e-02 -1.60982370e-01
-1.11708939e+00 1.06739593e+00 6.97509229e-01 -1.20527528e-01
-6.77914560e-01 -7.23599195e-01 -8.19979310e-01 2.68506229e-01
4.74892855e-01 -9.77182508e-01 1.92090261e+00 -9.94331062e-01
-1.46178794e+00 6.77749515e-01 -4.53246593e-01 -7.69003570e-01
2.59147078e-01 -7.30017304e-01 -9.64666232e-02 1.38490334e-01
1.69352889e-01 7.60310054e-01 7.66694665e-01 -8.64201605e-01
-3.68761748e-01 1.34781739e-02 9.08313915e-02 5.08091331e-01
-2.17778489e-01 3.03126257e-02 -4.08997864e-01 -8.47782075e-01
-4.56079185e-01 -6.59545600e-01 -1.41583920e-01 -7.78798401e-01
-6.26334667e-01 -4.25299048e-01 4.63292658e-01 -1.02156937e+00
1.53976583e+00 -1.78843927e+00 2.86784977e-01 -6.31380677e-01
-7.05392361e-02 3.42380553e-01 -5.49433172e-01 9.56422448e-01
-1.91466194e-02 1.95862457e-01 -5.76027393e-01 -7.36565292e-01
1.29930675e-01 1.32214725e-01 -6.19534135e-01 -1.67731300e-01
4.97449636e-01 1.39899468e+00 -8.11226606e-01 -4.79551584e-01
-2.49809235e-01 -4.41055298e-02 -6.94915831e-01 3.15276831e-01
-6.90126777e-01 2.56623477e-01 -2.85365254e-01 2.72987634e-01
2.13728070e-01 -4.78531361e-01 1.71183422e-01 1.19387738e-01
1.58725515e-01 1.00711727e+00 -2.77461380e-01 2.29778171e+00
-4.70405310e-01 8.25763345e-01 3.75141278e-02 -1.03851461e+00
4.70315486e-01 2.21510097e-01 1.28152385e-01 -5.55488169e-01
1.44359708e-01 1.94181770e-01 1.09360427e-01 -6.13056302e-01
1.22647631e+00 1.17810220e-01 -1.34094253e-01 7.27262557e-01
4.64251131e-01 -2.75066227e-01 6.60403609e-01 5.75628459e-01
1.25038230e+00 2.33628482e-01 1.90531284e-01 -1.91251531e-01
2.85165131e-01 1.56977102e-01 1.24184176e-01 8.46406341e-01
1.03188969e-01 7.23755836e-01 7.09980190e-01 -5.59193864e-02
-1.21965623e+00 -6.71886146e-01 -9.90239382e-02 1.67732561e+00
-3.74017775e-01 -8.65431070e-01 -9.32226062e-01 -9.01154757e-01
-7.49366060e-02 9.48327184e-01 -5.53320229e-01 8.74345526e-02
-8.72031510e-01 -5.83309054e-01 7.69917905e-01 6.85087800e-01
2.73808151e-01 -1.17933345e+00 -3.15039843e-01 3.79947990e-01
-6.87785864e-01 -1.07808948e+00 -9.47709501e-01 2.56892204e-01
-1.08412266e+00 -8.29724491e-01 -7.89504290e-01 -9.25137460e-01
3.47488612e-01 3.24632078e-02 1.71702409e+00 1.90008774e-01
3.58708575e-02 2.37839758e-01 -5.70477366e-01 -4.04569685e-01
-5.02594948e-01 7.58188725e-01 -4.76006895e-01 -5.91211736e-01
1.63531974e-01 -3.62455606e-01 -5.70531905e-01 -1.03448376e-01
-8.42326045e-01 2.76022822e-01 9.65763748e-01 1.25111771e+00
4.10188228e-01 -5.86386383e-01 9.98027742e-01 -1.01223660e+00
1.04391706e+00 -5.06248951e-01 -1.46661595e-01 4.46326554e-01
-5.37766814e-01 1.92717284e-01 6.85487032e-01 -2.68493295e-01
-1.12205362e+00 -5.35509109e-01 -5.32654583e-01 6.68518841e-02
-8.81821513e-02 8.04080367e-01 -3.94064263e-02 6.83430552e-01
7.25372851e-01 2.24403575e-01 -1.81874201e-01 -6.67886734e-01
8.01299989e-01 7.72063673e-01 6.07552946e-01 -7.06844509e-01
1.40170351e-01 5.29094450e-02 -5.95206976e-01 -6.09571457e-01
-1.24467790e+00 -4.38992500e-01 -6.54361844e-01 3.67991775e-01
8.75228882e-01 -9.10287499e-01 -4.88301292e-02 2.94991463e-01
-1.32469022e+00 -9.79297042e-01 -3.22347313e-01 1.46319419e-01
-5.32646120e-01 5.34843743e-01 -9.78532910e-01 -3.32850188e-01
-9.76723433e-01 -9.08594966e-01 1.24465322e+00 -9.06724483e-02
-3.84783357e-01 -1.16909301e+00 1.81012362e-01 5.17792881e-01
5.15618205e-01 -2.43170992e-01 9.43935275e-01 -9.96362805e-01
-2.49248415e-01 3.11506558e-02 -1.49892971e-01 5.59096396e-01
-1.33622766e-01 -1.53689235e-01 -8.64647388e-01 -4.61416274e-01
-1.88306391e-01 -8.49438190e-01 1.39015889e+00 6.31595016e-01
1.28340280e+00 -5.77130079e-01 -2.17162460e-01 5.73509336e-01
1.14475322e+00 -2.37381190e-01 4.55336601e-01 2.78639704e-01
5.75850248e-01 5.54939568e-01 5.94015419e-01 1.34341285e-01
5.62221169e-01 2.83177286e-01 1.02770671e-01 1.62700057e-01
-2.05903292e-01 -4.39782798e-01 5.53541243e-01 1.14705718e+00
2.54808575e-01 -7.05911338e-01 -8.54152560e-01 7.69030213e-01
-1.83514094e+00 -9.88100111e-01 1.12735502e-01 1.76063478e+00
1.20769954e+00 2.28330642e-01 7.29342103e-02 -4.42979515e-01
4.04315561e-01 6.12599313e-01 -5.70311427e-01 -7.90045500e-01
-8.85428488e-02 2.96004832e-01 3.62902105e-01 6.65350616e-01
-1.05610538e+00 1.15974689e+00 6.93211651e+00 8.98297846e-01
-9.33425784e-01 2.11778924e-01 6.73968256e-01 -5.52734911e-01
-3.91875684e-01 -5.47966361e-02 -9.34861481e-01 4.54492033e-01
1.43938112e+00 -3.73873383e-01 3.05021018e-01 6.30840421e-01
8.60812664e-02 -9.10553485e-02 -1.33178520e+00 4.60291266e-01
4.59027082e-01 -1.40171266e+00 1.69426143e-01 -2.06689328e-01
1.02880168e+00 5.28620899e-01 -1.83215111e-01 7.07062006e-01
4.42207396e-01 -1.17789817e+00 5.84696829e-01 2.14015797e-01
9.66633797e-01 -4.08086240e-01 7.60077953e-01 4.43606853e-01
-7.14944243e-01 -2.14955598e-01 -4.54130679e-01 -1.84318155e-01
4.14770633e-01 3.64197791e-01 -7.77264774e-01 7.49367833e-01
4.75600988e-01 9.15506899e-01 -6.29865706e-01 7.32855678e-01
-3.44959021e-01 9.97402072e-01 -1.26813665e-01 -1.72168780e-02
4.35815513e-01 -4.94719967e-02 5.30191898e-01 1.67887855e+00
1.51207149e-01 1.82205979e-02 3.50360423e-02 6.73274279e-01
-5.97568333e-01 2.10007057e-01 -4.29351002e-01 -2.16287389e-01
3.66510153e-01 1.04735470e+00 -7.13968128e-02 -7.21388936e-01
-5.17640352e-01 9.45562959e-01 7.51639426e-01 5.19201815e-01
-6.49177194e-01 -5.59345603e-01 3.13441783e-01 -2.28012756e-01
5.55604219e-01 -2.59571761e-01 -4.17885184e-01 -1.35606372e+00
2.38299519e-01 -1.11533737e+00 5.37712812e-01 -9.74856198e-01
-1.26891816e+00 7.15986907e-01 9.55636352e-02 -7.54148841e-01
-4.99879271e-01 -4.56758469e-01 -9.49340045e-01 9.51198339e-01
-1.61920536e+00 -1.28166151e+00 9.58055183e-02 2.75871992e-01
1.25312817e+00 4.78514545e-02 7.33027101e-01 -5.59008010e-02
-6.05925679e-01 7.04995751e-01 5.15491366e-01 1.71764836e-01
1.11281013e+00 -1.48391247e+00 9.13536727e-01 9.95684683e-01
1.50769114e-01 7.44919538e-01 5.53768754e-01 -7.19627798e-01
-1.14981437e+00 -1.01429355e+00 1.14199936e+00 -7.45510697e-01
8.35371673e-01 -4.30274040e-01 -1.00001907e+00 1.20202267e+00
1.02131486e+00 -5.52270472e-01 6.41171396e-01 3.84238720e-01
-1.51240438e-01 1.40301242e-01 -5.13061762e-01 5.53879023e-01
1.07477498e+00 -4.67950553e-01 -1.29578674e+00 3.78322989e-01
1.24081612e+00 -6.12091780e-01 -8.12492013e-01 1.60378069e-01
4.55590218e-01 -4.59151387e-01 7.08071649e-01 -9.02363181e-01
1.11363316e+00 3.65642160e-01 -1.30724460e-02 -1.62409341e+00
-2.23531917e-01 -7.58002996e-01 -3.27683181e-01 1.21109176e+00
7.33774185e-01 -6.00543082e-01 5.06986618e-01 1.83721870e-01
-8.24879766e-01 -1.12127101e+00 -6.69319630e-01 -6.58468008e-01
9.38112736e-01 -2.53212929e-01 4.28566754e-01 6.92959964e-01
2.58726269e-01 9.43205595e-01 -3.31070632e-01 -3.36949140e-01
2.08146483e-01 2.10225657e-01 7.94073284e-01 -7.30399430e-01
-4.21524048e-01 -6.61909163e-01 4.99225378e-01 -1.79280365e+00
3.87292892e-01 -1.09868050e+00 2.67381161e-01 -1.97474027e+00
4.56146032e-01 -2.28016116e-02 -6.30135685e-02 5.75927794e-01
-6.24017179e-01 -6.82520568e-02 9.11142901e-02 1.68401003e-01
-9.01874781e-01 8.41995120e-01 1.23280084e+00 -1.83246195e-01
-7.50970840e-02 -3.23429286e-01 -9.40356493e-01 2.96730876e-01
1.03085649e+00 -2.09359333e-01 -4.10963297e-01 -1.12373018e+00
1.59387678e-01 2.26742812e-02 -4.13320363e-02 -6.10871732e-01
6.16893917e-02 3.12752686e-02 9.29375663e-02 -9.55332041e-01
2.36592725e-01 -1.85875744e-01 -6.27643347e-01 1.53365701e-01
-9.45643544e-01 3.31175148e-01 4.11487967e-01 4.65820134e-01
-2.14570671e-01 -4.70066220e-01 4.01517123e-01 -1.92640945e-01
-5.17549753e-01 -6.16068346e-03 -3.36927921e-01 6.24601007e-01
3.48812312e-01 1.87851623e-01 -9.29724813e-01 -4.71777201e-01
-4.96053249e-01 6.56715035e-01 4.91807997e-01 4.15068358e-01
2.18905792e-01 -9.20074701e-01 -1.09331620e+00 -1.42703690e-02
4.01978977e-02 2.36454144e-01 2.84827530e-01 8.77379179e-01
-3.17763746e-01 8.93400967e-01 7.70251080e-02 -4.71985281e-01
-1.14719963e+00 4.31792200e-01 1.43899277e-01 -8.25234652e-01
-5.88950813e-01 8.91200662e-01 3.87538001e-02 -5.46975434e-01
2.41990030e-01 -6.41827464e-01 -5.99223599e-02 -3.17255664e-03
4.78107721e-01 1.47970840e-01 2.06247702e-01 -3.76268804e-01
1.49432570e-01 3.41417938e-01 -4.47695941e-01 -1.74284980e-01
1.31942105e+00 -4.12949592e-01 -8.56776461e-02 4.10419285e-01
1.17338645e+00 9.59864706e-02 -1.21019638e+00 -3.81276518e-01
8.60899761e-02 -1.05133668e-01 -2.23431841e-01 -1.16105890e+00
-5.15531480e-01 1.04579055e+00 -1.66582599e-01 1.34005010e-01
1.02787757e+00 1.51987866e-01 1.09362292e+00 8.59852076e-01
-2.32791439e-01 -1.22374022e+00 3.71686578e-01 1.14475632e+00
1.25111437e+00 -1.29931545e+00 1.14754841e-01 1.55082777e-01
-9.50681627e-01 9.56299245e-01 7.06948638e-01 -8.63610730e-02
1.42645225e-01 2.43396629e-02 -7.65675381e-02 -7.50377104e-02
-1.34180582e+00 -1.21891133e-01 3.80174071e-01 2.38517955e-01
8.13256502e-01 -1.33883581e-01 -4.93478119e-01 8.15694988e-01
-6.22717500e-01 -2.21639127e-01 6.20919466e-01 8.27100098e-01
-5.38366079e-01 -1.17705512e+00 -4.60701063e-02 7.45771646e-01
-7.96099603e-01 -5.90066910e-01 -5.17612755e-01 5.79883397e-01
-4.37614202e-01 1.03779268e+00 1.99127689e-01 -5.65752946e-02
2.15368569e-01 5.47704041e-01 3.93636018e-01 -9.56713080e-01
-6.64636731e-01 -1.13336910e-02 7.06929088e-01 -4.41602677e-01
-2.14123040e-01 -6.35871172e-01 -1.01331639e+00 -2.99515396e-01
-3.77665520e-01 1.81536853e-01 4.75121796e-01 9.33943987e-01
6.88009560e-01 6.84026599e-01 2.70433396e-01 -1.07347763e+00
-9.11595881e-01 -1.64899814e+00 -4.75903116e-02 4.25783217e-01
4.61357236e-01 -9.31207463e-02 -1.07744426e-01 2.49051362e-01] | [11.709339141845703, 8.948652267456055] |
5af459c4-1602-40f5-ac80-75d00e03aafa | measuring-your-aste-models-in-the-wild-a | 2305.17448 | null | https://arxiv.org/abs/2305.17448v1 | https://arxiv.org/pdf/2305.17448v1.pdf | Measuring Your ASTE Models in The Wild: A Diversified Multi-domain Dataset For Aspect Sentiment Triplet Extraction | Aspect Sentiment Triplet Extraction (ASTE) is widely used in various applications. However, existing ASTE datasets are limited in their ability to represent real-world scenarios, hindering the advancement of research in this area. In this paper, we introduce a new dataset, named DMASTE, which is manually annotated to better fit real-world scenarios by providing more diverse and realistic reviews for the task. The dataset includes various lengths, diverse expressions, more aspect types, and more domains than existing datasets. We conduct extensive experiments on DMASTE in multiple settings to evaluate previous ASTE approaches. Empirical results demonstrate that DMASTE is a more challenging ASTE dataset. Further analyses of in-domain and cross-domain settings provide promising directions for future research. Our code and dataset are available at https://github.com/NJUNLP/DMASTE. | ['Xinyu Dai', 'Fei Zhao', 'Jiaze Chen', 'Zhen Wu', 'Huiyun Yang', 'Ting Xu'] | 2023-05-27 | null | null | null | null | ['aspect-sentiment-triplet-extraction'] | ['natural-language-processing'] | [-1.93434089e-01 -5.38437009e-01 -5.27914762e-01 -6.90509915e-01
-7.49794066e-01 -7.99025893e-01 5.13929069e-01 -2.10008353e-01
-2.41510257e-01 6.26750052e-01 2.34418273e-01 -1.93176314e-01
2.30216548e-01 -5.55538595e-01 -1.76681146e-01 -2.79790342e-01
5.16810358e-01 5.22841156e-01 9.05002803e-02 -4.77764010e-01
1.26844615e-01 8.76643509e-02 -1.19891012e+00 3.59581918e-01
8.15645039e-01 8.25831473e-01 -1.03209361e-01 1.43044367e-01
-4.13583294e-02 2.83963293e-01 -8.46012712e-01 -1.02382874e+00
4.42977130e-01 -2.94513404e-01 -3.98622096e-01 4.08681661e-01
3.09635904e-02 -6.68618381e-02 -7.59270415e-02 9.26281095e-01
5.60580790e-01 3.45041789e-02 5.97636342e-01 -1.57825482e+00
-9.42055702e-01 7.48708546e-01 -8.04094613e-01 2.67957877e-02
3.77175003e-01 1.91123918e-01 1.36667514e+00 -9.70044196e-01
5.49380362e-01 1.13416243e+00 6.07962966e-01 4.03588206e-01
-7.71120608e-01 -1.02742088e+00 4.06719178e-01 3.69683281e-02
-1.15514457e+00 -3.00831884e-01 8.40331495e-01 -9.56157818e-02
1.05053425e+00 1.90488175e-01 6.79661572e-01 1.61146927e+00
-1.35565981e-01 1.09277582e+00 1.49908781e+00 -2.43604377e-01
1.47271723e-01 2.96875298e-01 4.23945278e-01 1.27240941e-01
6.68744743e-01 -2.28847831e-01 -4.38232601e-01 -2.08214357e-01
4.02149171e-01 -6.75303340e-02 -1.64894015e-01 -1.55450135e-01
-9.65154171e-01 8.90314281e-01 -5.92550896e-02 2.24732891e-01
-2.62737274e-01 -4.06791806e-01 5.68681479e-01 3.45261455e-01
6.46408796e-01 6.25101507e-01 -8.79607737e-01 -5.78563094e-01
-4.75377232e-01 3.28867137e-01 1.11173594e+00 1.23269212e+00
4.92302209e-01 -7.53415897e-02 2.66251955e-02 1.16866815e+00
6.32741302e-02 4.92037535e-01 5.49299002e-01 -8.06110501e-01
7.82438934e-01 9.16920602e-01 1.47017643e-01 -8.62407207e-01
-2.24819332e-01 -4.09833103e-01 -6.11263156e-01 -3.96817923e-01
3.02697182e-01 -4.09781694e-01 -6.42290771e-01 1.54013872e+00
3.39399099e-01 -1.21928245e-01 1.47012800e-01 8.98609281e-01
8.97892237e-01 5.46854556e-01 -1.71953663e-01 1.84068695e-01
1.74462295e+00 -1.25916839e+00 -7.76243389e-01 -6.90001905e-01
6.18946493e-01 -1.07485449e+00 1.62803471e+00 5.09504020e-01
-6.44893050e-01 4.82123680e-02 -7.95175493e-01 -2.26290636e-02
-5.03250062e-01 5.37628233e-01 9.35211122e-01 5.36958814e-01
-6.14362478e-01 -2.35572457e-01 -5.98004460e-01 -4.16374862e-01
3.75672668e-01 -1.36030257e-01 -2.72808462e-01 -3.32875580e-01
-1.39022827e+00 5.68436444e-01 2.12687194e-01 -3.03754974e-02
-3.13544929e-01 -5.79007268e-01 -9.99275625e-01 -1.43650621e-01
7.38957465e-01 -6.09748781e-01 1.68107784e+00 -7.76950359e-01
-1.18416357e+00 9.39338446e-01 -2.59443402e-01 -1.18798763e-01
2.63844669e-01 -2.89951354e-01 -6.50757551e-01 -3.22091162e-01
1.61926165e-01 3.29972029e-01 5.15957117e-01 -1.10612774e+00
-3.72983724e-01 -2.76388526e-01 4.60099488e-01 2.09123284e-01
-5.46462655e-01 2.68375576e-01 -8.59475315e-01 -9.92132068e-01
-4.07228410e-01 -1.18868649e+00 -2.24840015e-01 -2.29597613e-01
-6.50690258e-01 -2.62746394e-01 6.44474864e-01 -1.93938375e-01
1.52443421e+00 -2.24056816e+00 -3.56271684e-01 -1.47733584e-01
1.20132379e-01 3.94349456e-01 -3.98817033e-01 5.93932688e-01
-6.24123104e-02 2.60434091e-01 -1.87778950e-01 -4.90450293e-01
2.69353896e-01 2.50233650e-01 -1.79056063e-01 -1.21463843e-01
1.70292646e-01 9.50950563e-01 -6.64799988e-01 -3.55731189e-01
-3.09222400e-01 2.73031861e-01 -4.41590399e-01 1.82697531e-02
-3.36316764e-01 1.44569486e-01 -8.68531168e-01 9.44304645e-01
7.45193064e-01 -4.26306009e-01 5.41062579e-02 -2.39573553e-01
1.74909115e-01 3.93470496e-01 -9.37641025e-01 1.42095864e+00
-5.66565573e-01 8.59159708e-01 -2.17962638e-01 -7.02037752e-01
1.10227144e+00 7.00054616e-02 3.43972981e-01 -7.70595312e-01
4.19119865e-01 2.36947045e-01 -6.45371061e-03 -4.90435541e-01
8.42234373e-01 -1.36352822e-01 -4.81527925e-01 5.68323612e-01
-3.46453786e-01 -1.41174927e-01 7.30122268e-01 2.53022701e-01
1.03530562e+00 -1.95359617e-01 3.87639225e-01 5.91760837e-02
6.29426166e-02 1.25100210e-01 9.14790213e-01 2.01457232e-01
-5.50991476e-01 6.18574977e-01 7.67694771e-01 -2.03293353e-01
-1.02368331e+00 -7.07221627e-01 -2.15407893e-01 8.38486433e-01
9.19662565e-02 -8.34893048e-01 -5.06886423e-01 -7.91550577e-01
-1.15337625e-01 8.51314664e-01 -4.81402040e-01 6.31831288e-02
-2.20746309e-01 -1.20334542e+00 4.44939017e-01 5.58535933e-01
6.96859837e-01 -9.99750078e-01 -2.53106862e-01 -7.33885840e-02
-5.42606771e-01 -1.57958305e+00 -7.31672585e-01 -2.13789463e-01
-5.26939750e-01 -1.20080280e+00 -4.99556005e-01 -5.39864480e-01
5.11067212e-01 4.93466049e-01 1.60148847e+00 -3.15499365e-01
-3.25764455e-02 2.35490903e-01 -8.22716892e-01 -6.59436643e-01
-1.93278834e-01 3.49053800e-01 -1.69433817e-01 -2.88557678e-01
1.20081043e+00 -3.24003726e-01 -6.68229043e-01 7.26925969e-01
-1.08891582e+00 6.94937259e-02 7.59339690e-01 6.49647772e-01
6.28668845e-01 1.40543550e-01 6.00326896e-01 -1.45200634e+00
1.15739262e+00 -8.03579688e-01 -5.57643294e-01 1.91113696e-01
-7.82152712e-01 -3.29927772e-01 5.84704340e-01 -5.28436005e-01
-1.08569026e+00 -4.93554741e-01 -2.03590259e-01 -1.59954458e-01
-1.17147587e-01 8.20819914e-01 -2.64883190e-01 4.92193848e-01
3.19368571e-01 -4.59234044e-02 -3.65515143e-01 -5.33603430e-01
2.63319146e-02 9.35879767e-01 -1.27122896e-02 -7.07203984e-01
6.53732717e-01 1.13844797e-01 -4.88203466e-01 -5.16824663e-01
-1.13094199e+00 -5.61228871e-01 -1.57380313e-01 1.58794969e-01
1.66531593e-01 -1.22751307e+00 -8.13496336e-02 6.73949420e-01
-8.26467216e-01 -2.66134471e-01 -1.01528347e-01 3.92681569e-01
-4.81551476e-02 2.04742208e-01 -6.11580610e-01 -5.07061899e-01
-4.55476463e-01 -1.25984836e+00 1.22645640e+00 2.85244167e-01
-4.32238698e-01 -1.02658665e+00 2.01945603e-01 7.48685896e-01
3.23564440e-01 2.67795652e-01 4.61985409e-01 -8.79326999e-01
-3.87690872e-01 -3.62243146e-01 -1.61042854e-01 4.32502091e-01
3.33181709e-01 3.80701512e-01 -7.63846934e-01 -9.24574286e-02
-9.56737772e-02 -5.22290647e-01 6.50477707e-01 2.50230990e-02
1.07809865e+00 -2.59910196e-01 -2.31963277e-01 3.70352834e-01
1.31808639e+00 1.57802820e-01 6.27561748e-01 7.48841047e-01
4.86278892e-01 4.62925732e-01 1.28072441e+00 7.88010240e-01
8.83283675e-01 6.87565863e-01 3.66658792e-02 -5.42690903e-02
9.26670339e-03 -2.24860404e-02 6.33405387e-01 1.21369445e+00
2.87457287e-01 -7.04749703e-01 -1.07706976e+00 8.83514225e-01
-1.61685920e+00 -6.11487985e-01 -1.02352977e-01 1.58881915e+00
1.07015646e+00 2.55570740e-01 3.03980082e-01 3.47582623e-02
5.09166658e-01 2.24340245e-01 -7.37330556e-01 -6.34198070e-01
-3.27293485e-01 1.13877758e-01 2.30934266e-02 -5.94628267e-02
-1.08194101e+00 1.00622809e+00 5.45589876e+00 9.87356186e-01
-9.48268652e-01 3.52803692e-02 7.96930432e-01 -8.83895904e-02
-6.00692630e-01 6.91414997e-02 -9.70736265e-01 5.97957373e-01
7.31885791e-01 -4.51944977e-01 1.14446066e-01 9.01860356e-01
1.26401424e-01 -8.35179165e-03 -8.62940013e-01 9.19705629e-01
1.47090495e-01 -6.69894576e-01 -9.76785496e-02 -9.67183411e-02
8.83214056e-01 2.12441146e-01 3.01001012e-01 6.34828746e-01
5.20338893e-01 -6.78642511e-01 2.46916190e-01 1.21829184e-02
7.68471181e-01 -6.23136401e-01 9.49324667e-01 2.17516273e-01
-1.14412975e+00 1.13221988e-01 -3.14497709e-01 -3.80454189e-03
8.46436396e-02 9.79901612e-01 -6.97452307e-01 6.17225468e-01
8.13613892e-01 1.11078441e+00 -7.81955242e-01 9.31286693e-01
-2.91077048e-01 7.83131719e-01 -4.06712294e-01 -2.69173890e-01
2.80608773e-01 -3.92474562e-01 4.63495463e-01 1.19084728e+00
4.23638254e-01 -1.32133052e-01 1.91651151e-01 4.86880511e-01
-3.18683237e-01 1.56303048e-01 -6.98116243e-01 -3.90336871e-01
7.23209321e-01 1.48130572e+00 -6.43943310e-01 -2.21301764e-01
-8.47170949e-01 6.71943307e-01 2.88145483e-01 5.08972228e-01
-1.08007252e+00 -4.40978885e-01 9.97441888e-01 -3.36132981e-02
3.12629253e-01 -1.65872425e-01 -1.66090995e-01 -1.75481498e+00
3.12188864e-01 -1.45678329e+00 4.73268002e-01 -7.78700531e-01
-1.84327686e+00 8.72190297e-01 1.13119192e-01 -1.54598367e+00
-1.71961427e-01 -7.29437113e-01 -5.41724920e-01 4.48370218e-01
-1.67472017e+00 -1.08721316e+00 -4.77107882e-01 3.40392500e-01
8.07021022e-01 -1.85079291e-01 7.00582862e-01 5.51126063e-01
-1.03009462e+00 9.44418371e-01 2.93789078e-02 1.93206400e-01
1.11303818e+00 -1.01425171e+00 6.43387198e-01 6.75020158e-01
1.24538027e-01 7.70307004e-01 6.61869526e-01 -4.44443017e-01
-1.20151722e+00 -1.18186414e+00 8.46685588e-01 -6.59099579e-01
1.04836547e+00 -4.79097128e-01 -7.98288822e-01 9.42724228e-01
4.70449418e-01 -2.25104809e-01 1.09134519e+00 3.21804166e-01
-5.74726462e-01 -1.62903950e-01 -9.79600251e-01 8.09955835e-01
8.29667211e-01 -3.34381938e-01 -3.84693623e-01 1.74530834e-01
5.75324833e-01 -5.58810532e-01 -1.01091945e+00 4.99059170e-01
5.84786355e-01 -8.24072838e-01 5.97842515e-01 -3.76240075e-01
7.02335417e-01 -8.47956836e-02 -1.81700319e-01 -1.60182452e+00
9.83836204e-02 -3.70377541e-01 -2.83641070e-02 1.79183662e+00
7.76874483e-01 -8.27361345e-01 6.59863532e-01 7.62659967e-01
9.12622437e-02 -1.32140303e+00 -4.77305174e-01 -8.21547031e-01
8.60543400e-02 -5.32078624e-01 9.65218127e-01 1.11214316e+00
-1.83225125e-01 6.68972075e-01 -2.06257746e-01 -2.27868706e-01
1.80622786e-01 4.80431139e-01 9.09216762e-01 -7.64810562e-01
-2.37061620e-01 -3.19701552e-01 -1.19379625e-01 -1.10223639e+00
2.80228585e-01 -7.01416969e-01 -3.29094887e-01 -1.36673462e+00
4.76538986e-01 -6.34352326e-01 -1.04339018e-01 7.18024671e-01
-4.94940221e-01 4.10034269e-01 1.03000380e-01 4.09060456e-02
-7.41417170e-01 9.11651790e-01 1.39636338e+00 -1.51672423e-01
1.33299395e-01 2.29899347e-01 -1.15553129e+00 7.08423734e-01
1.39423501e+00 -4.19232368e-01 -6.98031723e-01 -6.26815021e-01
4.93439078e-01 -5.12447417e-01 -1.20925091e-01 -4.06792432e-01
-1.51694760e-01 -1.56733349e-01 3.69896814e-02 -5.26072681e-01
3.66506428e-01 -8.32266331e-01 -1.08223699e-01 -2.78646588e-01
-1.25234425e-01 6.41450703e-01 4.20408458e-01 3.16793591e-01
-6.10879540e-01 -3.16141903e-01 3.42745125e-01 -1.89637784e-02
-8.64735305e-01 4.27393913e-01 -1.57072440e-01 6.86140537e-01
9.77536142e-01 -1.45110125e-02 -5.96663535e-01 -6.39212430e-01
-2.19351843e-01 3.85196686e-01 7.42372155e-01 7.34919608e-01
5.53138018e-01 -1.31190443e+00 -6.43011451e-01 -6.37795106e-02
6.65722311e-01 -1.19787743e-02 -2.85777338e-02 6.48877740e-01
-2.71072596e-01 2.16584399e-01 3.96526195e-02 -2.52596438e-01
-1.30542839e+00 3.94167066e-01 3.14972773e-02 -6.64376140e-01
-3.01705986e-01 5.28495908e-01 1.82730988e-01 -7.71164834e-01
-1.00755155e-01 -2.98386395e-01 -2.93120265e-01 8.42177495e-03
2.39498407e-01 2.19420284e-01 -1.22110173e-01 -5.18592238e-01
-3.09478879e-01 4.59772497e-01 -3.94405663e-01 9.31179002e-02
1.26374400e+00 -2.72320151e-01 -3.22934575e-02 4.83617723e-01
1.11740029e+00 2.55256474e-01 -7.90178239e-01 -3.28643352e-01
1.23268060e-01 -6.04695916e-01 -3.50274920e-01 -9.59181011e-01
-1.33190858e+00 5.73688626e-01 7.80239627e-02 7.64514208e-02
1.56643391e+00 8.82097345e-04 1.19354117e+00 4.22925562e-01
3.19144428e-01 -1.04098511e+00 1.84210703e-01 6.66062355e-01
8.85199904e-01 -1.42640698e+00 1.37735888e-01 -6.32004023e-01
-1.26515031e+00 5.63144565e-01 1.01863348e+00 2.65220255e-01
6.49981558e-01 5.39013326e-01 5.02982855e-01 -2.39282802e-01
-1.13789630e+00 -1.54856563e-01 6.19615465e-02 3.46852213e-01
7.15945601e-01 1.70765445e-01 -5.06010830e-01 1.18300140e+00
-5.99810541e-01 -1.29243974e-02 6.37400389e-01 8.00159752e-01
2.02419013e-01 -1.55586541e+00 -2.94877607e-02 6.54651046e-01
-6.00391150e-01 -2.51191318e-01 -7.68366337e-01 1.10592806e+00
-1.04190126e-01 1.10761940e+00 -2.83155739e-01 -3.36494982e-01
5.69025874e-01 -2.86644697e-01 1.44267246e-01 -6.88378274e-01
-4.27435517e-01 3.86039913e-02 5.81933498e-01 -3.62265706e-01
-3.86410892e-01 -7.76239693e-01 -9.56423163e-01 -4.66223150e-01
-3.61213118e-01 2.47807384e-01 4.61813927e-01 5.60084164e-01
6.03311062e-01 4.39497977e-01 7.79873490e-01 -1.08609192e-01
-3.29381526e-01 -1.02564883e+00 -6.15169048e-01 7.21801102e-01
-6.79013729e-02 -7.83362091e-01 -3.05477053e-01 -2.15225354e-01] | [11.41590690612793, 6.712143898010254] |
8f0e16ad-2b32-4a21-95ca-3a66841583f2 | video-dialog-as-conversation-about-objects | 2207.03656 | null | https://arxiv.org/abs/2207.03656v1 | https://arxiv.org/pdf/2207.03656v1.pdf | Video Dialog as Conversation about Objects Living in Space-Time | It would be a technological feat to be able to create a system that can hold a meaningful conversation with humans about what they watch. A setup toward that goal is presented as a video dialog task, where the system is asked to generate natural utterances in response to a question in an ongoing dialog. The task poses great visual, linguistic, and reasoning challenges that cannot be easily overcome without an appropriate representation scheme over video and dialog that supports high-level reasoning. To tackle these challenges we present a new object-centric framework for video dialog that supports neural reasoning dubbed COST - which stands for Conversation about Objects in Space-Time. Here dynamic space-time visual content in videos is first parsed into object trajectories. Given this video abstraction, COST maintains and tracks object-associated dialog states, which are updated upon receiving new questions. Object interactions are dynamically and conditionally inferred for each question, and these serve as the basis for relational reasoning among them. COST also maintains a history of previous answers, and this allows retrieval of relevant object-centric information to enrich the answer forming process. Language production then proceeds in a step-wise manner, taking into the context of the current utterance, the existing dialog, the current question. We evaluate COST on the DSTC7 and DSTC8 benchmarks, demonstrating its competitiveness against state-of-the-arts. | ['Truyen Tran', 'Tu Minh Phuong', 'Vuong Le', 'Thao Minh Le', 'Hoang-Anh Pham'] | 2022-07-08 | null | null | null | null | ['video-question-answering', 'visual-dialogue', 'relational-reasoning', 'visual-dialogue'] | ['computer-vision', 'computer-vision', 'natural-language-processing', 'natural-language-processing'] | [ 6.58841804e-02 2.77739912e-01 -1.41811922e-01 -4.79672611e-01
-4.65656608e-01 -6.85513854e-01 1.09099078e+00 8.67303908e-02
-2.42975533e-01 5.11685908e-01 7.43781209e-01 -9.21990350e-02
1.62741944e-01 -6.71750426e-01 -5.12028933e-01 -3.03389817e-01
1.14245519e-01 7.15678453e-01 5.97268760e-01 -5.13731658e-01
1.83199719e-01 4.99863297e-01 -1.73670781e+00 1.02837348e+00
2.45092258e-01 9.37037110e-01 5.20806849e-01 1.08227050e+00
-6.00338817e-01 1.60160255e+00 -7.27312624e-01 -4.48717028e-01
-4.02050652e-02 -6.35698676e-01 -1.23467267e+00 5.32658517e-01
3.97998244e-01 -6.59643292e-01 -6.16752088e-01 7.03137994e-01
3.31672095e-02 4.60862815e-01 4.47487712e-01 -1.44763637e+00
-3.56117785e-01 5.93989074e-01 2.01269999e-01 2.75148034e-01
9.93157148e-01 6.46868169e-01 9.89079535e-01 -9.22612369e-01
1.16068876e+00 1.75091290e+00 -3.13994214e-02 7.99740732e-01
-9.28309977e-01 -1.87230781e-01 4.51821923e-01 4.23115671e-01
-8.05732846e-01 -5.55961668e-01 6.96092129e-01 -4.04552907e-01
1.20603263e+00 4.42598492e-01 9.64011192e-01 1.11457455e+00
-6.79822043e-02 1.09011304e+00 5.51782727e-01 -3.41095448e-01
3.80376548e-01 2.05798328e-01 1.35914728e-01 8.33591104e-01
-6.42052829e-01 -2.30511591e-01 -1.09785104e+00 5.47783189e-02
5.77108085e-01 -1.33678377e-01 -1.65477604e-01 -1.70938149e-01
-1.39878881e+00 6.23526812e-01 4.18856055e-01 4.28927809e-01
-4.10483181e-01 2.09487349e-01 3.06684881e-01 4.84571129e-01
1.12402692e-01 2.93418288e-01 2.56944392e-02 -4.26146746e-01
-6.31627917e-01 4.43750024e-01 9.91362512e-01 1.12000561e+00
5.30590892e-01 -2.80417968e-03 -5.99277139e-01 4.81434643e-01
3.12518090e-01 1.15229718e-01 2.03396812e-01 -1.68077052e+00
3.51251185e-01 8.94184053e-01 1.66753024e-01 -7.68034935e-01
-1.41149163e-01 3.36248398e-01 -2.80348569e-01 3.83781075e-01
6.43864334e-01 -1.17928416e-01 -5.58362782e-01 1.76202607e+00
7.04093814e-01 1.07920624e-01 3.32503498e-01 8.65352392e-01
1.14633858e+00 1.01039410e+00 9.20659583e-03 -4.35990602e-01
1.70350266e+00 -1.08065999e+00 -1.03244209e+00 -1.00484200e-01
3.67629051e-01 -5.16912341e-01 1.14617240e+00 3.04820925e-01
-1.47571397e+00 -6.82475090e-01 -6.31526291e-01 -3.63969117e-01
-3.37903678e-01 -2.38252640e-01 3.10633391e-01 1.29949018e-01
-1.32030463e+00 8.30538422e-02 -4.97206569e-01 -5.89679360e-01
1.47164036e-02 2.04328880e-01 -2.83903480e-01 8.69595557e-02
-1.00161719e+00 9.96131539e-01 2.48755544e-01 5.83948791e-02
-1.14874983e+00 -3.35429579e-01 -8.11718464e-01 1.07594788e-01
8.05548489e-01 -6.65564120e-01 1.68624866e+00 -8.71801138e-01
-1.72989666e+00 8.17947865e-01 -3.69091541e-01 -6.51201129e-01
4.27031040e-01 -2.26312801e-02 -2.47369289e-01 6.57287300e-01
-1.97535679e-01 1.28611183e+00 7.92114735e-01 -1.23788416e+00
-8.79361331e-01 -3.34847093e-01 8.97131085e-01 6.08372152e-01
-1.40557587e-01 3.17441881e-01 -9.76430357e-01 -2.57896692e-01
-1.85626104e-01 -1.03808928e+00 5.13758175e-02 3.00759941e-01
-8.99250358e-02 -5.32751381e-01 1.28395319e+00 -4.54579771e-01
9.59476888e-01 -2.07527828e+00 5.78693449e-01 -3.61844003e-01
1.89415038e-01 4.75405082e-02 -4.51519080e-02 6.96090639e-01
3.18551242e-01 -2.47516841e-01 2.20443428e-01 -4.77146924e-01
5.04181124e-02 1.62135854e-01 -4.56906080e-01 3.28980833e-02
2.73669779e-01 9.33213592e-01 -8.70408714e-01 -7.49514341e-01
3.69299948e-01 3.78362209e-01 -7.20128119e-01 6.10811770e-01
-1.14124048e+00 4.95978504e-01 -3.35464180e-01 3.79391223e-01
3.12095508e-02 -4.05058503e-01 3.65711033e-01 -2.95557410e-01
-2.43522450e-01 2.16390386e-01 -8.64098132e-01 1.86643827e+00
-3.07589263e-01 9.03493941e-01 3.56682777e-01 -6.51557922e-01
7.48295248e-01 5.43608606e-01 3.54229808e-01 -4.66134787e-01
8.33335742e-02 -2.70290881e-01 -6.05812930e-02 -9.11850810e-01
6.04203582e-01 1.31059542e-01 -1.86142921e-01 4.94644612e-01
2.25573748e-01 -5.15708864e-01 4.47858155e-01 7.94847488e-01
1.09243846e+00 1.99148968e-01 -2.41513085e-02 1.24596342e-01
8.69097412e-01 2.52636850e-01 -9.92393941e-02 5.38237512e-01
-1.51171982e-01 2.75494307e-01 7.35816717e-01 -4.05386567e-01
-7.02837884e-01 -1.08008611e+00 4.55312759e-01 1.43499124e+00
2.72403538e-01 -4.37345177e-01 -8.31147611e-01 -6.56087577e-01
-3.62909734e-01 8.15723479e-01 -4.69251037e-01 -4.63931635e-02
-7.61469662e-01 6.01783432e-02 3.37058336e-01 3.03311050e-01
5.54040432e-01 -1.63237071e+00 -9.88889813e-01 1.24849230e-01
-5.86634755e-01 -1.41709924e+00 -3.55115801e-01 -1.49584204e-01
-5.07608235e-01 -8.32712114e-01 -5.36420345e-01 -7.69993603e-01
4.65790182e-01 2.36894071e-01 1.15488172e+00 3.22019279e-01
-1.18810214e-01 1.05254519e+00 -3.42918754e-01 -8.82992893e-02
-8.35606813e-01 -5.07320106e-01 -2.48850256e-01 2.47429341e-01
7.29922056e-02 -1.33623391e-01 -5.30581415e-01 3.33445132e-01
-9.06827569e-01 3.57152939e-01 1.08326457e-01 5.16504943e-01
2.02848196e-01 -2.61281610e-01 4.67945129e-01 -4.79563057e-01
5.92994153e-01 -4.28859383e-01 -5.54588497e-01 4.16584879e-01
1.75735474e-01 6.39388412e-02 3.99530351e-01 -3.88540506e-01
-1.59630978e+00 2.24277675e-01 5.11347549e-03 -3.74841303e-01
-2.30677858e-01 2.80395329e-01 -1.11610569e-01 4.37544584e-01
5.45012295e-01 1.76853210e-01 1.75522923e-01 -9.45994779e-02
7.75133908e-01 5.72443783e-01 9.09357548e-01 -6.54198170e-01
4.47194070e-01 6.27754450e-01 -2.12053403e-01 -1.07999647e+00
-6.34139061e-01 -4.65215892e-01 -6.12968206e-01 -9.62018013e-01
1.29433584e+00 -7.07823455e-01 -1.28851390e+00 2.79195428e-01
-1.41413927e+00 -7.09292233e-01 -4.44750547e-01 7.45701343e-02
-8.01989853e-01 2.10308656e-01 -6.60138786e-01 -1.01948118e+00
1.62759230e-01 -1.27802122e+00 1.01917887e+00 1.89587742e-01
-4.37320948e-01 -8.41821015e-01 -1.43043071e-01 7.69083083e-01
1.81315526e-01 -7.10345060e-02 9.11199868e-01 -5.75252414e-01
-1.16148734e+00 1.52135104e-01 -1.94223076e-01 -7.10969642e-02
-4.93518785e-02 2.51474716e-02 -8.87598097e-01 -4.27502627e-03
2.18982145e-01 -6.08151495e-01 3.04175913e-01 7.20640048e-02
8.98526371e-01 -4.64565575e-01 -1.38210475e-01 -9.63704586e-02
8.07298660e-01 5.92170715e-01 4.91725266e-01 -1.36962950e-01
3.58852804e-01 1.14748847e+00 7.11071312e-01 2.74506450e-01
6.23833954e-01 1.13676631e+00 5.83762705e-01 2.14005634e-01
-4.46766376e-01 -1.40799925e-01 5.53970158e-01 4.43006873e-01
1.33304950e-02 -6.20247662e-01 -7.39879668e-01 3.75623554e-01
-2.06026387e+00 -1.43319833e+00 -1.31754562e-01 1.73881233e+00
8.92444313e-01 9.34638828e-02 3.10009062e-01 -2.15362191e-01
5.67803323e-01 2.64304906e-01 -5.23929238e-01 -2.82497466e-01
4.78880666e-02 -3.92350644e-01 -4.22694147e-01 7.10240185e-01
-6.91027522e-01 1.16910648e+00 6.06047297e+00 3.16505730e-01
-9.73987401e-01 -7.52438307e-02 4.42661583e-01 -2.89402664e-01
-2.00169981e-01 -1.68035217e-02 -7.23548055e-01 1.42136365e-01
1.10161722e+00 -3.84452567e-02 6.34246945e-01 5.45788229e-01
3.14076513e-01 -4.57397223e-01 -1.40121055e+00 7.43319631e-01
2.42486075e-01 -1.84963441e+00 1.17850222e-01 -1.57352433e-01
2.78665841e-01 -2.75552422e-01 3.61711644e-02 4.33228254e-01
5.02153099e-01 -6.83151662e-01 8.86399508e-01 4.68622983e-01
5.13850749e-01 -2.63230145e-01 1.00096583e-01 5.82257450e-01
-1.31138325e+00 -2.27408975e-01 1.05005212e-01 -3.75557244e-02
5.14043033e-01 -2.75055200e-01 -1.10385478e+00 2.22603917e-01
6.73558414e-01 3.97172749e-01 -2.95279503e-01 4.67970073e-01
-7.74005428e-02 9.82341170e-02 -1.32024944e-01 -3.00946593e-01
3.62945378e-01 4.86479551e-02 7.46910095e-01 1.29797721e+00
8.86836052e-02 5.28281510e-01 2.22481042e-01 7.21371651e-01
3.49343792e-02 -2.02071562e-01 -7.04620421e-01 3.82488929e-02
3.66959333e-01 1.13227808e+00 -8.60365629e-01 -7.68748641e-01
-5.49542844e-01 1.07031155e+00 1.91547409e-01 3.79193723e-01
-8.22393715e-01 2.68552393e-01 4.03224230e-01 1.26807407e-01
2.56093353e-01 -3.06167096e-01 4.20922965e-01 -1.15321529e+00
-1.65454760e-01 -1.15229166e+00 5.57356715e-01 -1.32689011e+00
-7.34313965e-01 7.57031918e-01 4.18661088e-01 -9.71709430e-01
-7.93958843e-01 -5.14122009e-01 -4.56471026e-01 2.58297414e-01
-9.05913413e-01 -1.05764782e+00 -7.31265426e-01 9.34474170e-01
1.39683771e+00 -2.62175679e-01 6.67295575e-01 -1.52421612e-02
-2.43596286e-01 -1.21115267e-01 -7.60449588e-01 -3.16304341e-03
4.53340083e-01 -1.14509904e+00 1.35517374e-01 6.12490177e-01
4.88378733e-01 2.71489859e-01 8.32707167e-01 -7.56970882e-01
-1.76293302e+00 -6.77273035e-01 8.60387862e-01 -6.47736609e-01
8.20272386e-01 -6.28510714e-01 -8.66670787e-01 6.97017908e-01
6.71205759e-01 -1.97515413e-01 2.52276152e-01 -3.57715756e-01
-2.12234586e-01 5.86818531e-02 -9.93787706e-01 9.89288688e-01
1.14407766e+00 -7.51463652e-01 -8.97162259e-01 5.30344307e-01
1.05744326e+00 -5.31882823e-01 -3.40377361e-01 -5.28987981e-02
4.18978930e-01 -1.15812016e+00 1.07516396e+00 -7.17806220e-01
4.79778081e-01 -3.03812563e-01 -3.45888019e-01 -7.83841729e-01
1.05723731e-01 -9.33866441e-01 -3.53048384e-01 1.18277514e+00
3.03681046e-01 9.25978348e-02 8.92248333e-01 8.78125846e-01
-2.14992866e-01 -3.92920613e-01 -8.19214940e-01 -4.10198092e-01
-5.15748739e-01 -5.37117422e-01 4.54767704e-01 4.49680775e-01
4.22345906e-01 7.38100052e-01 -3.27049464e-01 -7.54581764e-02
1.05943769e-01 9.73116010e-02 1.13035345e+00 -1.03060937e+00
-2.80892253e-01 -3.28109652e-01 -9.90485102e-02 -1.38600934e+00
3.60222280e-01 -5.74948490e-01 2.54906714e-01 -1.69432855e+00
-2.86373729e-03 4.50945832e-02 3.32461894e-01 1.28579170e-01
2.67032504e-01 -4.62719649e-02 6.81637406e-01 3.30102116e-01
-9.57107246e-01 4.87792015e-01 1.25836813e+00 -2.08739221e-01
-3.93137336e-01 -1.22536682e-01 -1.42715290e-01 8.72990072e-01
4.11363870e-01 -4.30541746e-02 -7.63535559e-01 -2.17941150e-01
1.24934606e-01 9.98534024e-01 4.96500373e-01 -1.00592673e+00
6.71590745e-01 -3.55963856e-01 -1.35129476e-02 -8.72531354e-01
1.23465979e+00 -1.07368064e+00 2.64006585e-01 4.55198228e-01
-7.99695492e-01 2.05368653e-01 5.96967712e-02 7.02895820e-01
-3.10301930e-01 -2.23565206e-01 4.91430938e-01 -3.10433477e-01
-9.75901067e-01 1.03450678e-01 -8.50876808e-01 -7.44354576e-02
1.20940733e+00 -1.96498364e-01 -4.34586078e-01 -9.99355435e-01
-1.01964307e+00 3.98034096e-01 2.15806350e-01 6.08670175e-01
7.47021079e-01 -1.13192296e+00 -5.30526400e-01 -2.18923360e-01
5.77850640e-02 -1.28139243e-01 3.27183187e-01 5.27207136e-01
-4.03464913e-01 3.95461142e-01 -2.89305955e-01 -7.81137645e-01
-1.55631924e+00 6.95929468e-01 2.77357131e-01 -1.48092493e-01
-7.66049087e-01 8.40303779e-01 5.17292619e-01 -5.04205376e-02
5.79610050e-01 -4.37630504e-01 -6.04054987e-01 2.89210379e-01
6.76427186e-01 6.78060725e-02 -3.20185691e-01 -7.58057058e-01
-1.37584671e-01 1.61829486e-01 1.78052709e-01 -7.99503207e-01
1.03084600e+00 -4.48264003e-01 -5.60756102e-02 7.16305196e-01
8.06029201e-01 -5.51386178e-02 -1.57613409e+00 -2.66010433e-01
-8.01987350e-02 -3.20015103e-01 -3.83894980e-01 -6.96308076e-01
-7.55129635e-01 8.23063791e-01 1.92512497e-01 7.02638388e-01
9.56621528e-01 5.05286634e-01 5.41228592e-01 6.77586854e-01
2.33344853e-01 -9.33061302e-01 9.66800988e-01 6.31480932e-01
1.35925937e+00 -1.08014965e+00 -2.96835274e-01 -4.14244115e-01
-1.03024864e+00 1.17285097e+00 8.20664525e-01 2.99786985e-01
2.75948197e-01 2.51681447e-01 -2.29167026e-02 -4.94499236e-01
-1.38838279e+00 -2.40440816e-01 1.94421515e-01 5.17705679e-01
1.53491527e-01 -3.65381867e-01 2.92344719e-01 2.45682076e-02
-7.57731274e-02 -6.38550594e-02 5.23095489e-01 9.07688379e-01
-7.31831670e-01 -9.65967655e-01 -3.72959048e-01 2.87767146e-02
-9.10268649e-02 2.54659504e-01 -7.06762195e-01 7.51139462e-01
-1.46248892e-01 1.22864711e+00 3.74525309e-01 -1.79019138e-01
2.07358018e-01 3.04114878e-01 3.74383003e-01 -7.81725645e-01
-7.81481981e-01 1.16522051e-01 4.60805863e-01 -6.78688705e-01
-5.85485697e-01 -7.20556378e-01 -1.53955567e+00 -1.19267054e-01
3.49199921e-02 1.46249771e-01 4.99782801e-01 9.78303134e-01
2.65012294e-01 4.61264014e-01 3.31332535e-01 -8.59716952e-01
-1.27135694e-01 -5.97971559e-01 2.65594602e-01 4.96472925e-01
4.03354079e-01 -4.51568604e-01 -1.94239482e-01 4.96357292e-01] | [10.817216873168945, 1.1336114406585693] |
a96405a9-5402-4c94-89f6-be4f69fb5fb1 | openfwi-benchmark-seismic-datasets-for | 2111.02926 | null | https://arxiv.org/abs/2111.02926v6 | https://arxiv.org/pdf/2111.02926v6.pdf | OpenFWI: Large-Scale Multi-Structural Benchmark Datasets for Seismic Full Waveform Inversion | Full waveform inversion (FWI) is widely used in geophysics to reconstruct high-resolution velocity maps from seismic data. The recent success of data-driven FWI methods results in a rapidly increasing demand for open datasets to serve the geophysics community. We present OpenFWI, a collection of large-scale multi-structural benchmark datasets, to facilitate diversified, rigorous, and reproducible research on FWI. In particular, OpenFWI consists of 12 datasets (2.1TB in total) synthesized from multiple sources. It encompasses diverse domains in geophysics (interface, fault, CO2 reservoir, etc.), covers different geological subsurface structures (flat, curve, etc.), and contains various amounts of data samples (2K - 67K). It also includes a dataset for 3D FWI. Moreover, we use OpenFWI to perform benchmarking over four deep learning methods, covering both supervised and unsupervised learning regimes. Along with the benchmarks, we implement additional experiments, including physics-driven methods, complexity analysis, generalization study, uncertainty quantification, and so on, to sharpen our understanding of datasets and methods. The studies either provide valuable insights into the datasets and the performance, or uncover their current limitations. We hope OpenFWI supports prospective research on FWI and inspires future open-source efforts on AI for science. All datasets and related information can be accessed through our website at https://openfwi-lanl.github.io/ | ['Yinpeng Chen', 'Hanchen Wang', 'Youzuo Lin', 'Qili Zeng', 'Xitong Zhang', 'Peng Jin', 'Shihang Feng', 'Yinan Feng', 'Chengyuan Deng'] | 2021-11-04 | null | null | null | null | ['geophysics'] | ['miscellaneous'] | [-1.42422274e-01 -3.26113909e-01 8.55750591e-02 -3.04047048e-01
-1.24818325e+00 -6.10433638e-01 5.92514873e-01 -1.57320559e-01
-1.12495713e-01 1.02321625e+00 4.86591876e-01 -6.02163970e-01
-2.76819766e-01 -1.20576441e+00 -7.91712582e-01 -1.02050579e+00
-5.79426944e-01 6.14003241e-01 3.36381137e-01 -3.22647005e-01
5.47703087e-01 5.03991663e-01 -1.25941670e+00 3.88050079e-02
9.30069268e-01 1.08285511e+00 1.34547532e-01 3.88904423e-01
6.28345064e-04 3.60335737e-01 -2.04449058e-01 1.17670409e-01
1.58415377e-01 -1.61901087e-01 -8.39869618e-01 -7.16603339e-01
1.95832059e-01 -1.69431165e-01 -5.93550801e-01 7.81691670e-01
8.61001134e-01 2.85893768e-01 7.76399136e-01 -9.33477461e-01
-5.87938547e-01 6.07661247e-01 -6.37703896e-01 6.07140243e-01
1.76162459e-02 2.88659215e-01 1.12542355e+00 -1.22231674e+00
5.74139953e-01 9.22926128e-01 8.13419998e-01 2.29164079e-01
-8.74726534e-01 -8.70472312e-01 -2.92728662e-01 2.80285805e-01
-1.21419477e+00 -6.73919439e-01 7.58978426e-01 -5.85906625e-01
8.37035835e-01 1.72522739e-01 5.42886615e-01 1.14909613e+00
9.62226093e-02 7.15130568e-01 1.07448626e+00 -8.93585235e-02
4.62154865e-01 -6.56350255e-01 1.69184908e-01 5.37118614e-01
6.00128949e-01 3.95014733e-01 -7.50761092e-01 -3.58125895e-01
8.49564016e-01 -4.06110406e-01 -6.12839401e-01 3.18256259e-01
-1.20813918e+00 9.26606417e-01 1.40378147e-01 2.03929216e-01
-1.12090103e-01 3.50703567e-01 3.76734406e-01 4.33198124e-01
7.67104387e-01 6.09739125e-01 -6.43930793e-01 -4.19002116e-01
-9.65061188e-01 8.78872097e-01 8.82517636e-01 7.61303961e-01
9.34710801e-01 4.99469340e-01 3.03078413e-01 8.58431399e-01
4.84013945e-01 1.13277745e+00 4.33003515e-01 -1.17225230e+00
5.71295559e-01 9.40870419e-02 1.27532527e-01 -1.04280722e+00
-6.70674562e-01 -1.58690363e-01 -9.16705430e-01 -3.74349132e-02
3.02305967e-01 -3.35962772e-01 -7.67140150e-01 1.39131737e+00
-5.30249253e-03 7.13266850e-01 1.87878534e-01 9.78100300e-01
1.31527722e+00 9.85204041e-01 -4.85504448e-01 2.16957018e-01
1.04340994e+00 -4.90631610e-01 -4.39370662e-01 -2.52232432e-01
5.66370666e-01 -4.80957627e-01 1.06762731e+00 3.32873017e-01
-1.08912730e+00 -1.39938682e-01 -9.19759154e-01 5.37151434e-02
-3.72567952e-01 -3.09608310e-01 7.40443170e-01 1.54232070e-01
-8.10360551e-01 7.90617347e-01 -1.10532224e+00 9.22846571e-02
4.96607602e-01 -2.64698178e-01 -1.41769499e-01 -2.98926570e-02
-1.76301539e+00 6.18090451e-01 1.99024081e-01 3.17535013e-01
-1.22620904e+00 -1.16990638e+00 -9.86467183e-01 -1.40555084e-01
1.12923250e-01 -3.42546135e-01 1.15779448e+00 2.55697817e-01
-1.25149989e+00 5.22992313e-01 4.08175215e-02 -3.96638632e-01
4.43628073e-01 -1.08733788e-01 -3.69088739e-01 7.69076496e-02
1.60319984e-01 1.57353580e-01 4.75021958e-01 -1.10291088e+00
-3.06219429e-01 -2.30782107e-01 -3.74233544e-01 -3.83356363e-01
-9.95455608e-02 -2.50071704e-01 -3.15460801e-01 -7.33794987e-01
2.41079271e-01 -6.43164992e-01 -2.07006842e-01 -1.92674324e-01
-4.04460132e-01 -6.37900382e-02 7.22220540e-01 -7.92379916e-01
1.13985515e+00 -1.93938518e+00 1.26806632e-01 3.06236863e-01
2.94853598e-01 1.34688824e-01 -1.27756223e-01 6.87616587e-01
5.35765141e-02 3.72211993e-01 -1.02690756e+00 -1.99819788e-01
-1.39367402e-01 4.83290046e-01 -5.51047444e-01 5.44752598e-01
3.34801286e-01 9.38911498e-01 -7.88960755e-01 -2.07348809e-01
4.56616729e-02 5.81188649e-02 -5.41498959e-01 6.94623515e-02
-1.60038546e-01 7.60188282e-01 -7.27201700e-01 9.77428496e-01
1.07994235e+00 -3.46565545e-01 -3.78196448e-01 -1.38254330e-01
-4.68292654e-01 2.77539641e-01 -1.14038205e+00 1.76372683e+00
-4.39816505e-01 6.09915137e-01 2.70045642e-02 -1.37086976e+00
1.12744141e+00 1.70212239e-01 4.86019611e-01 -8.48778903e-01
-1.64664596e-01 9.03549492e-01 1.49799958e-02 -8.50973964e-01
3.45817626e-01 9.06021800e-04 -1.66332468e-01 5.35723329e-01
4.25252728e-02 -3.51247936e-01 3.34539145e-01 7.23564625e-02
9.97381449e-01 1.02018699e-01 -2.36286923e-01 -8.44137549e-01
1.69556364e-01 1.07399061e-01 7.13185310e-01 8.32122207e-01
-3.49523611e-02 9.11647975e-01 3.38192821e-01 -5.33421516e-01
-1.11055064e+00 -1.11001670e+00 -7.71703362e-01 5.24374783e-01
1.00263201e-01 -9.70657319e-02 -2.47483719e-02 6.95919171e-02
3.72001082e-01 3.51067334e-01 -6.81341112e-01 -9.12126601e-02
-7.88611054e-01 -1.35374653e+00 1.09863341e+00 8.19722176e-01
8.00520718e-01 -9.54046369e-01 -5.64455569e-01 1.79285288e-01
-5.58959901e-01 -8.38760614e-01 1.09008625e-01 7.52151981e-02
-8.78144324e-01 -9.59191918e-01 -6.93868041e-01 -3.76391143e-01
-6.68404996e-03 -1.47298928e-02 1.18142545e+00 -1.20824398e-02
-2.68037945e-01 -3.37507874e-02 -4.53668654e-01 -4.62619960e-01
-1.76338077e-01 1.19397424e-01 -4.29286622e-03 -4.25778419e-01
-1.08511500e-01 -7.67969728e-01 -6.21311069e-01 3.05187911e-01
-7.55454540e-01 -1.27774671e-01 1.19700097e-01 9.95837152e-01
4.93135452e-01 -1.49770468e-01 8.75855088e-01 -6.55054450e-01
5.43462157e-01 -1.06707990e+00 -8.57431889e-01 1.64604131e-02
-3.19002301e-01 1.59963414e-01 1.97880328e-01 5.46337664e-02
-1.23174262e+00 -9.28087771e-01 -6.32389724e-01 -1.06358103e-01
1.02792263e-01 1.18118465e+00 1.44303992e-01 -2.54666597e-01
8.27315807e-01 2.64803201e-01 -1.13373637e-01 -6.69928014e-01
6.18340559e-02 6.88181400e-01 6.28331900e-01 -1.07002628e+00
7.25020528e-01 5.91690302e-01 -5.56518473e-02 -1.23817289e+00
-6.78761482e-01 -2.65742511e-01 -1.04756750e-01 -1.43593594e-01
5.02596319e-01 -9.36876118e-01 -3.13615531e-01 8.42454016e-01
-8.75557780e-01 -6.80459917e-01 -2.41995245e-01 5.97022831e-01
-1.70660138e-01 2.82464862e-01 -8.09048712e-01 -8.20954442e-01
-5.20446956e-01 -1.14592624e+00 1.17130387e+00 2.63533592e-01
1.25578731e-01 -1.36068749e+00 3.61027002e-01 1.63086474e-01
8.78302991e-01 5.88918865e-01 7.50222027e-01 -4.43796158e-01
-5.47429562e-01 1.20526701e-01 -3.65487546e-01 1.45259738e-01
-1.41937032e-01 2.58234948e-01 -1.08549356e+00 -1.47124365e-01
-3.09617305e-03 -7.47626185e-01 1.50820458e+00 6.51372790e-01
1.29436123e+00 1.25950217e-01 -2.30565906e-01 9.79828894e-01
1.31732249e+00 2.49074817e-01 7.01698661e-01 4.83358473e-01
4.92060930e-01 3.05117667e-01 3.92893106e-01 8.98142815e-01
2.92561501e-01 2.28065439e-02 4.79929745e-01 -6.15180321e-02
1.48758873e-01 1.15123600e-01 -1.45684019e-01 1.10992336e+00
-6.17966890e-01 -2.60347366e-01 -1.63798594e+00 6.47811472e-01
-1.69597864e+00 -9.05072570e-01 -4.01360065e-01 2.03371000e+00
8.42387795e-01 -4.24407572e-02 -4.96621579e-01 1.73776999e-01
2.39630163e-01 4.88516510e-01 -7.69172251e-01 7.27352872e-02
-4.53571498e-01 5.30905724e-01 5.58647156e-01 4.60916370e-01
-1.04615474e+00 7.47056544e-01 6.34110308e+00 7.72503555e-01
-1.33608627e+00 7.44910985e-02 6.77286386e-01 2.84554541e-01
-8.80983055e-01 -1.11000970e-01 -7.16611862e-01 4.38720554e-01
1.07080233e+00 -3.34295690e-01 2.44332045e-01 3.94282162e-01
2.83197075e-01 -2.78817099e-02 -5.62960267e-01 6.24040306e-01
-3.50778848e-01 -2.11171913e+00 -2.38644734e-01 8.66361186e-02
7.59214044e-01 8.70356858e-01 -2.06289366e-01 3.02010834e-01
1.82573587e-01 -8.85689020e-01 5.93103647e-01 7.79223382e-01
1.04351461e+00 -4.99060869e-01 8.50331664e-01 1.33500144e-01
-1.28497386e+00 6.62408844e-02 -6.15254581e-01 -3.15879256e-01
3.32721949e-01 8.76518428e-01 -1.03725396e-01 9.92648304e-01
1.04595709e+00 1.15427065e+00 -1.06852144e-01 1.14983130e+00
2.12458093e-02 1.22725487e+00 -5.38698673e-01 1.98497340e-01
3.42653930e-01 -3.43748927e-01 5.73782206e-01 1.12466264e+00
6.12419963e-01 2.80428290e-01 3.45145613e-02 9.72479820e-01
-8.92510340e-02 -2.98885524e-01 -6.53700173e-01 -6.46718517e-02
8.86265814e-01 9.79816556e-01 -4.11854208e-01 -2.26477116e-01
-4.32990074e-01 9.73062292e-02 -1.28892399e-02 4.75184143e-01
-8.70212555e-01 -4.57797110e-01 8.00508976e-01 1.37147993e-01
2.52395179e-02 -4.95427251e-01 -3.83571029e-01 -1.29781616e+00
-1.55837253e-01 -4.75692987e-01 4.58831191e-01 -7.74191976e-01
-1.46970499e+00 3.58197689e-01 4.12343621e-01 -1.02119279e+00
6.98789731e-02 -8.77890885e-01 -9.59788561e-01 8.44532967e-01
-1.96674573e+00 -6.44638896e-01 -7.05716550e-01 3.03962559e-01
3.85533035e-01 -2.69433141e-01 8.08524907e-01 6.58819795e-01
-7.14946806e-01 1.49598867e-01 6.34209216e-01 2.76898921e-01
5.88097453e-01 -9.82225955e-01 9.94310677e-01 6.85392082e-01
-1.57267720e-01 3.03454101e-01 5.49588203e-01 -7.98218489e-01
-1.67704809e+00 -1.15156519e+00 3.50866914e-01 -2.05187201e-01
1.20670950e+00 -1.02637477e-01 -1.50737464e+00 5.04923224e-01
-8.77222270e-02 4.47408199e-01 4.95962888e-01 -2.23999679e-01
2.30978504e-02 1.67848729e-02 -8.87961805e-01 1.26689702e-01
1.18312538e+00 -1.96545079e-01 -5.47203541e-01 3.13948542e-01
6.34961426e-01 -6.98462546e-01 -1.20735466e+00 7.32622504e-01
4.11606729e-01 -8.07176709e-01 9.50416207e-01 -5.02540112e-01
9.89976108e-01 2.77545806e-02 -3.13431084e-01 -1.48077118e+00
-2.28795230e-01 -2.68148184e-01 -2.72729427e-01 8.54768276e-01
6.50088191e-01 -1.10289538e+00 5.10750949e-01 2.56770372e-01
-6.88070297e-01 -1.05942988e+00 -1.17924798e+00 -8.72345984e-01
6.90372467e-01 -6.64752841e-01 7.48080075e-01 9.12976921e-01
-2.46965796e-01 -2.54219115e-01 -3.40846360e-01 3.34427446e-01
7.51345515e-01 5.21438241e-01 3.59616399e-01 -1.45251274e+00
-3.07877176e-02 -4.81632978e-01 -8.37721378e-02 -8.87151659e-01
-1.99637283e-02 -9.50469136e-01 1.77966524e-02 -1.45475364e+00
-2.55341828e-01 -9.35497284e-01 -1.49798840e-01 5.72810709e-01
2.73392182e-02 3.04300636e-01 -3.81884396e-01 5.79774439e-01
1.13614693e-01 8.83751869e-01 1.33828306e+00 -1.81299120e-01
1.60225913e-01 -2.10964650e-01 -4.16164666e-01 7.07557857e-01
1.18884516e+00 -3.68459672e-01 -1.77150473e-01 -8.95083606e-01
2.74605870e-01 1.99260920e-01 3.86358857e-01 -1.02040493e+00
3.46076265e-02 -3.97643626e-01 1.44364297e-01 -6.06401265e-01
1.47515073e-01 -7.59337693e-02 -9.51559469e-02 3.51308495e-01
3.31167765e-02 -9.74316746e-02 3.92309010e-01 4.20316309e-01
-5.17030656e-01 -3.50708246e-01 6.83663189e-01 -3.64937693e-01
-9.05164182e-01 6.25253856e-01 -1.67517275e-01 8.75741839e-01
4.38754559e-01 1.88805535e-01 -5.78314781e-01 -1.10395268e-01
-4.05444860e-01 5.17863154e-01 1.15131289e-01 7.41323978e-02
8.01970899e-01 -1.06622887e+00 -1.23955536e+00 4.60205883e-01
1.68359458e-01 5.65186024e-01 3.55370611e-01 8.49126160e-01
-8.65260363e-01 1.62426606e-01 -5.54699488e-02 -7.25166559e-01
-2.23921165e-01 -2.85226166e-01 4.74026114e-01 -8.84011686e-02
-9.90913332e-01 1.14711761e+00 -2.03268483e-01 -6.55631959e-01
-1.44414663e-01 -6.69174969e-01 -5.29007651e-02 -2.09786054e-02
6.18691146e-01 7.72150755e-01 2.41863877e-01 -2.51485109e-01
-3.38094652e-01 6.76653206e-01 3.50259751e-01 -1.14844896e-01
1.75850248e+00 1.60294786e-01 -1.03625052e-01 7.42846966e-01
1.15130842e+00 -2.10693106e-01 -1.19716501e+00 -3.32228601e-01
1.17355620e-03 -3.87933403e-01 2.44162455e-01 -5.52390516e-01
-1.45789683e+00 1.11474919e+00 1.78162649e-01 7.36445859e-02
8.69964123e-01 2.35898849e-02 9.72842574e-01 5.39411247e-01
5.16591966e-01 -7.77776599e-01 -1.41628245e-02 8.11983883e-01
1.27635872e+00 -1.37292683e+00 1.69491395e-02 7.45574087e-02
-4.03281748e-01 1.24410975e+00 5.40475428e-01 -2.17562363e-01
1.00177097e+00 7.90205002e-01 -3.58099006e-02 -4.78895009e-01
-6.65120125e-01 2.44403154e-01 -1.22116834e-01 1.58157766e-01
4.47791994e-01 -9.88071412e-02 -1.62926987e-01 6.91748321e-01
-2.75688946e-01 1.75970681e-02 6.15305603e-01 1.12483263e+00
-5.99979222e-01 -8.42871666e-01 -3.93440247e-01 7.58859813e-01
-2.40556728e-02 -2.10067794e-01 4.52751458e-01 6.64111197e-01
-1.04354411e-01 5.26735604e-01 3.96179967e-02 3.90282720e-02
1.51984423e-01 -1.40329987e-01 1.17446333e-01 -1.95039213e-01
9.60490778e-02 -1.41566381e-01 2.44774073e-01 -4.23547745e-01
-3.17423344e-01 -8.31106722e-01 -1.58700299e+00 -4.23612088e-01
-1.82067811e-01 4.19058114e-01 5.58881521e-01 8.58920991e-01
2.76963294e-01 5.80338359e-01 3.32555562e-01 -1.24333501e+00
-4.45882022e-01 -1.16407359e+00 -6.95270181e-01 3.06776632e-02
2.92080939e-01 -1.12227607e+00 -7.47549534e-01 -3.35696280e-01] | [6.857238292694092, 2.4919168949127197] |
9f80c773-d000-48c8-bd03-8918807605fc | on-the-relevance-of-bandwidth-extension-for | 2202.13865 | null | https://arxiv.org/abs/2202.13865v1 | https://arxiv.org/pdf/2202.13865v1.pdf | On the relevance of bandwidth extension for speaker identification | In this paper we discuss the relevance of bandwidth extension for speaker identification tasks. Mainly we want to study if it is possible to recognize voices that have been bandwith extended. For this purpose, we created two different databases (microphonic and ISDN) of speech signals that were bandwidth extended from telephone bandwidth ([300, 3400] Hz) to full bandwidth ([100, 8000] Hz). We have evaluated different parameterizations, and we have found that the MELCEPST parameterization can take advantage of the bandwidth extension algorithms in several situations. | ['W. Bastiaan Kleijn', 'Mattias Nilsson', 'Marcos Faundez-Zanuy'] | 2022-02-24 | null | null | null | null | ['bandwidth-extension', 'bandwidth-extension', 'speaker-identification'] | ['audio', 'speech', 'speech'] | [ 6.13078400e-02 -9.39199626e-02 -8.90273526e-02 -3.31623495e-01
-5.70693910e-01 -5.56924164e-01 4.49518025e-01 -4.20054287e-01
-5.80253959e-01 1.01759136e+00 2.89147347e-01 -5.58095098e-01
-3.22978348e-01 -2.28197202e-01 2.33877217e-03 -4.27788109e-01
-3.28707486e-01 2.18308792e-01 3.44943017e-01 -2.03285918e-01
1.64945751e-01 7.99338579e-01 -1.86101305e+00 2.18899235e-01
3.00906032e-01 6.66832089e-01 -3.56430635e-02 1.15017796e+00
3.31003107e-02 -1.66210800e-01 -1.21243060e+00 -1.61496121e-02
7.32031390e-02 -3.44732404e-01 -1.08312070e+00 9.79682580e-02
-1.19681515e-01 -1.34048998e-01 -4.74004820e-02 1.01353264e+00
7.21291184e-01 3.31575423e-01 5.79266131e-01 -1.09683704e+00
1.21077456e-01 8.71287048e-01 -1.79196566e-01 9.73890126e-01
6.01150393e-01 -2.27394208e-01 5.34851551e-01 -5.52615643e-01
3.20980966e-01 1.24831545e+00 5.57534933e-01 4.90160674e-01
-1.46004975e+00 -7.50660598e-01 -3.29806894e-01 8.05638358e-02
-1.73305845e+00 -9.80257094e-01 1.02093983e+00 -3.09910446e-01
1.16496468e+00 5.84105015e-01 6.75537884e-01 1.25804698e+00
-2.63117313e-01 1.22534022e-01 7.88230300e-01 -1.03837776e+00
1.86353344e-02 4.20261085e-01 5.08719265e-01 -1.14886425e-01
-5.95652498e-02 1.55513361e-01 -4.53352123e-01 -5.86963892e-01
5.84933579e-01 -9.38936174e-01 -7.49593675e-01 3.97009075e-01
-1.08989656e+00 9.09741104e-01 -3.03986311e-01 8.59746099e-01
-3.25574666e-01 -1.19284622e-01 6.31949663e-01 6.68801129e-01
5.04377842e-01 6.43815756e-01 -2.55517483e-01 -5.82870364e-01
-9.77791190e-01 1.60869397e-02 1.24788499e+00 9.87348080e-01
5.03884435e-01 5.06207108e-01 2.91541994e-01 1.27982080e+00
1.46621168e-01 1.66111887e-01 7.71512628e-01 -8.24767828e-01
3.17385197e-01 -5.14576316e-01 2.60835528e-01 -3.63910377e-01
-5.09143353e-01 -8.65202248e-02 -5.62668264e-01 -4.71773684e-01
5.00265658e-01 -5.72127163e-01 -4.59054500e-01 1.46099591e+00
1.69783816e-01 1.19470112e-01 1.53785110e-01 6.63710415e-01
4.86011088e-01 7.12734461e-01 -1.14644125e-01 -7.27679372e-01
1.35042536e+00 -4.33226377e-01 -1.18132126e+00 1.54713109e-01
2.47263238e-01 -1.25282359e+00 1.14271510e+00 6.80792570e-01
-8.96216094e-01 -8.93297851e-01 -1.03627586e+00 5.15797496e-01
-2.00037986e-01 -7.97954276e-02 3.72530073e-01 1.49344552e+00
-1.17334020e+00 6.37764692e-01 -4.24414158e-01 -4.19823021e-01
-4.84457731e-01 3.65780562e-01 -2.96732694e-01 6.56346440e-01
-1.62737238e+00 5.34637034e-01 4.98681456e-01 -2.00848415e-01
-2.35911399e-01 -3.77270818e-01 -4.99340892e-01 -5.51526062e-02
-1.50838256e-01 -9.03746579e-03 1.57465684e+00 -8.93642306e-01
-1.64209414e+00 6.35388851e-01 -1.37489960e-01 -5.57952225e-01
1.24527179e-01 -2.58549601e-02 -1.19879115e+00 4.32987571e-01
-4.90781248e-01 1.67074084e-01 1.11295915e+00 -6.41881406e-01
-4.33645964e-01 -4.90478687e-02 -3.22443038e-01 -7.77294636e-02
-5.60193837e-01 4.39570099e-01 -8.90173689e-02 -7.65715480e-01
-8.55205208e-02 -8.36401880e-01 4.99764442e-01 -8.86538446e-01
-4.59882200e-01 2.27018129e-02 7.51415074e-01 -7.76540756e-01
1.35207665e+00 -2.50836825e+00 -1.47736728e-01 5.49570620e-01
-5.46521306e-01 4.43529576e-01 1.11283161e-01 6.04192317e-01
-3.79594892e-01 2.09764197e-01 2.33990893e-01 -9.14176926e-02
-1.11578122e-01 2.75155783e-01 -4.40843165e-01 4.56987798e-01
-3.20774704e-01 -2.91911401e-02 -6.24883249e-02 -3.53155255e-01
2.22946912e-01 7.31935143e-01 -4.83331233e-01 -1.84691977e-02
4.44680482e-01 2.56191313e-01 -1.19091183e-01 2.88580477e-01
4.91259813e-01 4.29141790e-01 -4.51103039e-02 -8.43872502e-02
-3.45052451e-01 5.50457954e-01 -1.25109780e+00 1.24056923e+00
-3.47346038e-01 1.02422833e+00 3.49187732e-01 -8.00179362e-01
1.03700304e+00 1.02386785e+00 3.44552666e-01 -9.74880084e-02
1.09865069e-01 2.68352896e-01 4.31004226e-01 -3.80849183e-01
4.80311453e-01 -4.50343668e-01 2.75295317e-01 1.72716692e-01
2.34247506e-01 -5.27817488e-01 2.32126117e-01 -4.73901093e-01
4.49108958e-01 -6.36198103e-01 4.00671870e-01 -6.62874460e-01
9.14104939e-01 -4.43624139e-01 7.43343160e-02 5.47101617e-01
-4.38439935e-01 2.37091586e-01 4.20231730e-01 1.92008987e-01
-1.10479319e+00 -9.32681024e-01 -6.93746746e-01 1.13382936e+00
-3.67247462e-01 -3.27266067e-01 -1.02645242e+00 1.93401247e-01
-1.85164332e-01 9.60945070e-01 -8.13064426e-02 -4.86240089e-02
-6.92291379e-01 -6.58094168e-01 1.16106284e+00 3.42120618e-01
3.27604979e-01 -8.12103033e-01 -3.24147910e-01 3.53426456e-01
-2.73514122e-01 -9.80791092e-01 -7.93761432e-01 4.08086061e-01
-8.34740698e-01 -5.49721420e-01 -8.32710683e-01 -5.91702759e-01
-2.24652007e-01 -8.58793221e-03 9.36874032e-01 -3.24331373e-01
-2.05026969e-01 6.21027827e-01 -6.18248582e-01 -4.90324795e-01
-8.52201283e-01 3.25181037e-01 3.32026601e-01 -5.34671023e-02
3.47023785e-01 -7.31159389e-01 -1.71314135e-01 6.09027505e-01
-6.83592558e-01 -5.41670382e-01 7.26223812e-02 7.36515582e-01
1.65127013e-02 1.58408031e-01 1.11357474e+00 -6.06367171e-01
1.24685037e+00 -1.80011690e-01 -4.93840039e-01 -1.59253716e-01
-2.67322570e-01 -2.18213946e-01 3.76785338e-01 -8.43774021e-01
-1.10834980e+00 -1.24979630e-01 -8.20771039e-01 -4.15213674e-01
-5.91018260e-01 3.01822752e-01 -2.23763838e-01 -1.94931865e-01
8.21156502e-01 2.80147076e-01 4.54148948e-02 -7.70660520e-01
1.94089472e-01 1.30049264e+00 4.66280460e-01 -5.54687262e-01
1.85411885e-01 -1.93483680e-01 -5.79235733e-01 -1.68253458e+00
1.57888412e-01 -7.67201185e-01 -3.23412478e-01 -2.86064088e-01
5.62288284e-01 -6.37081265e-01 -5.11217356e-01 3.88934523e-01
-1.10889888e+00 -1.28075276e-02 -3.98674488e-01 1.10690236e+00
-6.93404675e-01 5.98150373e-01 -7.78424084e-01 -1.20429683e+00
-2.33552918e-01 -6.77239060e-01 5.76726615e-01 1.95175871e-01
-5.75586796e-01 -8.58039796e-01 8.37942809e-02 -3.32762040e-02
5.23279488e-01 -3.12020212e-01 6.68926060e-01 -9.46515560e-01
4.86242712e-01 -2.17269048e-01 3.51890773e-01 5.59649706e-01
3.76597553e-01 7.84426779e-02 -1.58297288e+00 -4.18815136e-01
4.41139996e-01 1.13340840e-01 5.34555733e-01 4.73515481e-01
1.05563414e+00 -3.48664165e-01 -2.56247129e-02 4.08421427e-01
7.06067979e-01 7.95096695e-01 7.06495583e-01 4.88221049e-02
-2.26400331e-01 6.82530940e-01 3.07866901e-01 4.05584127e-01
-5.07737935e-01 8.17001343e-01 -2.62557924e-01 2.14435890e-01
-1.79044232e-02 9.23051611e-02 2.45667070e-01 9.20145869e-01
-2.64134914e-01 -2.73289055e-01 -6.64798081e-01 6.17875874e-01
-9.37263668e-01 -1.01609361e+00 7.72184730e-02 2.40697765e+00
9.66197610e-01 3.23871732e-01 6.00577176e-01 8.27687025e-01
1.12715805e+00 9.49627906e-02 -1.10926457e-01 -9.07821298e-01
1.77264571e-01 1.56366304e-01 3.00069690e-01 6.86105907e-01
-1.08096325e+00 3.79401177e-01 7.90039110e+00 9.76821482e-01
-1.02291787e+00 -5.73124252e-02 3.14455897e-01 7.01033399e-02
4.65403683e-02 -4.08409089e-01 -8.52997184e-01 4.44627494e-01
1.64744985e+00 -4.00249779e-01 5.38858175e-01 7.91146159e-01
1.26031309e-01 4.17453274e-02 -1.15022135e+00 1.00749230e+00
-1.41651183e-01 -6.89940631e-01 -2.28740394e-01 2.07523972e-01
1.52747050e-01 -2.93391705e-01 -7.13988543e-02 2.12713435e-01
-4.73749638e-01 -7.88048744e-01 4.91113722e-01 1.60091370e-01
1.00001955e+00 -9.92636681e-01 7.41300881e-01 4.88345742e-01
-1.29182053e+00 8.39591697e-02 -3.94196689e-01 4.91795987e-02
3.17292809e-01 5.81517994e-01 -1.37046504e+00 4.25850749e-01
4.75028008e-01 -1.18846834e-01 -9.08013247e-03 1.13209069e+00
4.03867602e-01 8.46345901e-01 -5.38647711e-01 -3.87611240e-02
-4.57097664e-02 5.31722456e-02 6.92225337e-01 1.44903409e+00
6.64276361e-01 3.62036452e-02 -1.87491313e-01 5.60100019e-01
3.50969493e-01 9.19924211e-03 -6.47418916e-01 -2.42268428e-01
7.15579271e-01 6.89978838e-01 -4.97871786e-01 -1.61644831e-01
-3.66146356e-01 6.69026792e-01 -3.84234607e-01 5.52258909e-01
-5.84857821e-01 -1.01281190e+00 6.26040697e-01 -7.55434185e-02
1.96040958e-01 -1.20114014e-01 2.81626850e-01 -8.59022737e-01
-5.41311860e-01 -8.27580035e-01 3.46248209e-01 -5.15687466e-01
-1.12659073e+00 8.19273174e-01 4.42216814e-01 -9.56310928e-01
-9.06395972e-01 -4.69984263e-01 -4.50356185e-01 1.30161262e+00
-8.16146195e-01 -3.39213401e-01 4.41962540e-01 7.57443249e-01
6.44863605e-01 -2.74913818e-01 1.03292537e+00 3.86156648e-01
-1.52149320e-01 5.76890767e-01 -2.19859350e-02 -4.19225357e-02
4.84958023e-01 -1.09394670e+00 3.16909522e-01 4.25286829e-01
2.32663751e-01 8.03949416e-01 1.17791319e+00 -2.15015993e-01
-7.91971207e-01 -6.14052951e-01 9.91137207e-01 1.98490635e-01
4.78392601e-01 -3.24399740e-01 -1.13501477e+00 5.03499150e-01
2.11584374e-01 -3.50759268e-01 8.96626592e-01 2.37547368e-01
-7.24922866e-02 -1.26626968e-01 -1.26507890e+00 2.66207218e-01
4.85097796e-01 -8.62358332e-01 -8.42983723e-01 6.26921356e-02
7.03032911e-01 -2.05855191e-01 -9.65618789e-01 8.80082473e-02
5.44773638e-01 -1.15178990e+00 1.10840416e+00 -4.22181308e-01
-7.56826937e-01 1.60986573e-01 -3.02588880e-01 -1.45090628e+00
-1.01898480e-02 -9.40265179e-01 2.12561861e-01 1.44136798e+00
4.36296672e-01 -1.19085693e+00 3.54721487e-01 2.99349070e-01
-1.80223770e-02 -2.14132816e-02 -1.46428990e+00 -1.05895019e+00
-1.88402086e-02 -6.49080217e-01 7.60569751e-01 5.54123580e-01
5.64816177e-01 3.18427175e-01 -5.14133990e-01 -3.85773443e-02
1.77602053e-01 -3.18461955e-01 2.98379540e-01 -1.16668987e+00
-4.85868692e-01 -5.22177219e-01 -2.88466692e-01 -1.05043697e+00
1.46030724e-01 -1.91341445e-01 -5.93368970e-02 -5.35181105e-01
-3.61959636e-01 -5.43095469e-02 -2.40668654e-01 -4.69752178e-02
2.82258511e-01 -5.05639948e-02 4.02644016e-02 -6.27686754e-02
5.06952286e-01 2.54138559e-01 7.95344353e-01 -1.75893831e-03
-3.51094991e-01 6.02383971e-01 -2.16177717e-01 8.10887456e-01
9.85015988e-01 -9.83253345e-02 -7.01668024e-01 1.67720526e-01
-4.57511842e-01 5.83730042e-01 -1.21714901e-02 -9.98634934e-01
-7.77985454e-02 1.10444121e-01 1.16832718e-01 -5.50622582e-01
8.41922939e-01 -6.57896519e-01 4.64842170e-01 3.57904255e-01
-5.03748834e-01 -1.75947830e-01 5.78080833e-01 4.16632056e-01
-4.84613568e-01 -7.13479638e-01 8.86805117e-01 1.26401410e-01
-5.21516502e-01 -3.53218675e-01 -8.12453985e-01 -1.96149960e-01
6.42523587e-01 -3.10960978e-01 -2.05699876e-02 -6.34259582e-01
-9.81122792e-01 -4.61204320e-01 -3.78054130e-04 1.78920120e-01
4.55442995e-01 -1.10691679e+00 -4.10144031e-01 6.31794155e-01
-1.55609831e-01 -8.27141941e-01 1.84895679e-01 7.69052923e-01
-4.82811928e-01 8.47994983e-01 -2.18623713e-01 -5.24225295e-01
-1.65669668e+00 4.51954931e-01 4.68818724e-01 3.29328448e-01
-4.71067667e-01 8.33141446e-01 -2.21154183e-01 -1.24848187e-01
5.27363479e-01 -3.74080092e-01 -4.88152593e-01 1.19842306e-01
5.77226102e-01 6.09858990e-01 3.03258747e-01 -6.22017860e-01
-4.22001600e-01 3.83908987e-01 2.59670228e-01 -8.44490707e-01
7.28496611e-01 -3.26359749e-01 1.80072457e-01 8.74660194e-01
1.16692758e+00 2.48972818e-01 -5.90958238e-01 1.88781932e-01
2.30481386e-01 -5.34146667e-01 -1.31130321e-02 -3.72893453e-01
-4.83242035e-01 8.86477113e-01 6.67300403e-01 1.08988059e+00
1.30270886e+00 -1.37471095e-01 3.87442887e-01 2.89210379e-01
5.04485011e-01 -1.16277969e+00 -3.41656059e-01 4.51776117e-01
9.91219759e-01 -5.03753901e-01 -3.00722927e-01 -6.14835680e-01
-4.84973013e-01 1.36369312e+00 -2.32743588e-03 2.36891076e-01
9.35812950e-01 2.49310419e-01 3.29265483e-02 2.75940508e-01
-6.07439518e-01 -1.22441381e-01 3.08697850e-01 8.98734212e-01
7.48024106e-01 1.55487165e-01 -3.69717568e-01 6.90352619e-01
-6.53837800e-01 -3.34005594e-01 6.63859248e-01 4.30537879e-01
-7.44514406e-01 -1.04688632e+00 -1.06031513e+00 3.53464335e-01
-6.68275476e-01 1.66086510e-01 -3.21884751e-01 1.05950296e+00
-2.69008249e-01 1.60464239e+00 1.07978038e-01 -2.58809894e-01
4.16458011e-01 7.03940928e-01 2.77385324e-01 -4.22700107e-01
-4.44712490e-01 9.76866722e-01 7.91696429e-01 -7.21792951e-02
-4.04106945e-01 -7.58857846e-01 -8.40907991e-01 -2.50779808e-01
-7.53934920e-01 7.71425247e-01 8.05149734e-01 6.97435617e-01
-2.38759086e-01 7.20706880e-01 6.04968131e-01 -8.43017220e-01
-8.28848839e-01 -1.51131749e+00 -1.38163137e+00 4.81016748e-02
5.41171610e-01 -4.36348379e-01 -7.54553497e-01 1.80233613e-01] | [14.73827075958252, 5.970371723175049] |
85c45bd7-0287-4ac4-9b84-743d743a59f1 | optimization-of-passive-chip-components | 2001.09612 | null | https://arxiv.org/abs/2001.09612v1 | https://arxiv.org/pdf/2001.09612v1.pdf | Optimization of Passive Chip Components Placement with Self-Alignment Effect for Advanced Surface Mounting Technology | Surface mount technology (SMT) is an enhanced method in electronic packaging in which electronic components are placed directly on soldered printing circuit board (PCB) and are permanently attached on PCB with the aim of reflow soldering process. During reflow process, once deposited solder pastes start melting, electronic components move in a direction that achieve their highest symmetry. This motion is known as self-alignment since can correct potential mounting misalignment. In this study, two noticeable machine learning algorithms, including support vector regression (SVR) and random forest regression (RFR) are proposed as a prediction technique to (1) diagnose the relation among component self-alignment, deposited solder paste status and placement machining parameters, (2) predict the final component position on PCB in x, y, and rotational directions before entering in the reflow process. Based on the prediction result, a non-linear optimization model (NLP) is developed to optimize placement parameters at initial stage. Resultantly, RFR outperforms in terms of prediction model fitness and error. The optimization model is run for 6 samples in which the minimum Euclidean distance from component position after reflow process from ideal position (i.e., the center of pads) is outlined as 25.57 ({\mu}m) regarding defined boundaries in model. | ['Hae-Yong Yang', 'Irandokht Parviziomran', 'Shun Cao', 'Seungbae Park', 'Daehan Won'] | 2020-01-27 | null | null | null | null | ['2d-human-pose-estimation'] | ['computer-vision'] | [ 1.86243296e-01 -1.29342198e-01 -4.25502174e-02 -3.11032623e-01
-8.80796015e-02 -3.34655017e-01 -1.19590508e-02 -2.73025334e-01
1.95466444e-01 6.05629563e-01 -1.97083965e-01 2.04702988e-02
-6.00388944e-01 -5.88350236e-01 -4.73949969e-01 -8.33124638e-01
4.31944877e-01 6.45024598e-01 1.76003456e-01 1.20304301e-01
6.06747150e-01 7.88561523e-01 -1.05148292e+00 1.04483500e-01
7.55552232e-01 8.12304199e-01 6.22813821e-01 1.44640446e-01
3.22647452e-01 1.87848404e-01 -6.56929612e-01 -2.91302890e-01
-5.95657118e-02 -9.55831334e-02 -9.09996107e-02 3.31053361e-02
-4.82435077e-01 -1.66545838e-01 -7.58162141e-02 7.29838312e-01
-1.18766678e-02 -2.60680139e-01 9.58113134e-01 -1.17441285e+00
-3.36497575e-01 4.11472321e-01 -4.23413754e-01 -4.59666222e-01
4.04625446e-01 -1.77323028e-01 4.81837869e-01 -8.67822468e-01
7.10127592e-01 8.41673315e-01 7.53881395e-01 3.89130324e-01
-9.32079196e-01 -4.20159757e-01 -3.45579505e-01 2.82042801e-01
-1.20157433e+00 2.38464829e-02 1.17215919e+00 -5.60006022e-01
7.66710997e-01 6.88863218e-01 5.29615998e-01 6.94426298e-01
1.42658985e+00 -1.87801987e-01 1.03680420e+00 -3.98704022e-01
3.67729515e-01 2.55853236e-01 3.33808780e-01 1.16041966e-01
4.90939796e-01 1.65481605e-02 -1.51174411e-01 7.88282901e-02
7.85308123e-01 -1.57660201e-01 1.64443150e-01 -1.37172595e-01
-7.47992277e-01 4.42835778e-01 5.33035584e-02 2.33354777e-01
-3.82631987e-01 -3.35097432e-01 -6.23318308e-04 -7.43609220e-02
2.34379858e-01 7.10523784e-01 -7.54744112e-01 -2.07067236e-01
-5.64532280e-01 -1.21712886e-01 1.04762137e+00 7.48119533e-01
2.96433181e-01 -6.86391592e-02 4.24560681e-02 9.23154533e-01
8.12640786e-01 3.26075107e-01 4.44510251e-01 -7.47866809e-01
2.28091985e-01 7.12617695e-01 2.10904449e-01 -1.16801310e+00
-2.93504089e-01 -6.39828444e-02 -6.40568912e-01 2.54560530e-01
-1.95518732e-01 -4.22441483e-01 -4.46706533e-01 8.24758410e-01
5.96411288e-01 8.47570300e-02 -9.88678187e-02 9.70798969e-01
6.70097113e-01 8.05730999e-01 -1.82670921e-01 -5.46519756e-01
1.30165672e+00 -5.56173503e-01 -9.92843032e-01 -2.05526710e-01
1.95905939e-01 -1.32044423e+00 5.15748322e-01 5.38624823e-01
-5.32033682e-01 -4.38026160e-01 -1.74176741e+00 5.73319554e-01
-2.68424815e-03 7.02724755e-01 4.88165766e-01 6.20352149e-01
-4.67782617e-01 1.08427453e+00 -9.28936720e-01 -4.02313434e-02
-1.90001458e-01 7.77809322e-01 -4.09710646e-01 4.56659310e-02
-5.81576824e-01 1.23058057e+00 -5.21153986e-01 8.09429109e-01
-1.70149565e-01 -6.07216239e-01 -4.91987705e-01 -5.90544224e-01
2.59501133e-02 -2.84561962e-01 7.87458479e-01 -2.19353572e-01
-1.97361481e+00 5.31065047e-01 7.70219490e-02 6.98097944e-02
3.98478240e-01 -7.22096980e-01 -4.88386631e-01 -5.33685863e-01
-3.21942195e-03 -4.98276949e-02 7.93601036e-01 -1.30502832e+00
2.15044856e-01 -4.94812936e-01 -7.81068683e-01 -1.49271697e-01
1.92347795e-01 2.02445075e-01 -1.39469266e-01 -4.19564158e-01
9.88795400e-01 -8.32288921e-01 -2.36063376e-01 -3.38097274e-01
-9.04607892e-01 -3.17707151e-01 1.09800744e+00 -1.09553158e+00
9.62429881e-01 -1.86263049e+00 1.11127621e-03 3.07259202e-01
-3.60973418e-01 -1.76469564e-01 4.83135641e-01 6.51387990e-01
-1.87143013e-01 -3.22276086e-01 9.09252316e-02 2.13315159e-01
-4.58318666e-02 1.18055686e-01 3.39926571e-01 7.71518111e-01
3.60365421e-01 4.97614890e-01 -2.17551112e-01 -2.05583230e-01
4.88573581e-01 2.04634815e-01 -1.42903790e-01 3.62459242e-01
-4.31534797e-02 1.90894619e-01 -5.87802827e-01 8.31200004e-01
1.02931285e+00 3.41191739e-01 3.58781338e-01 -8.89242113e-01
-4.12866950e-01 -1.10169038e-01 -1.36639130e+00 8.32072258e-01
-2.20169827e-01 3.93862188e-01 1.09906644e-01 -6.27309620e-01
1.76659739e+00 2.08657905e-01 8.28299642e-01 -4.05211121e-01
3.19631606e-01 5.23614064e-02 -7.46439695e-02 -7.93965995e-01
3.02964598e-01 8.98878202e-02 5.71811199e-03 9.60197076e-02
-4.46469307e-01 -1.72132865e-01 -3.04026634e-01 -5.47544777e-01
8.77264798e-01 4.35572147e-01 -1.88571557e-01 -3.52464169e-01
3.78143936e-01 -3.48327123e-02 9.70030904e-01 -1.15040690e-01
5.16354084e-01 6.76345587e-01 4.96289343e-01 -2.02651352e-01
-1.24006176e+00 -1.00645709e+00 -2.95267850e-01 1.41243994e-01
5.44433296e-01 3.05317342e-02 -8.68420124e-01 -2.28373688e-02
8.48077163e-02 6.48213685e-01 -1.24927841e-01 -2.70831078e-01
-7.95065522e-01 -9.05147672e-01 -4.11612660e-01 3.16259384e-01
1.02602974e-01 -7.13905573e-01 -3.80283564e-01 4.06364352e-01
6.31535590e-01 -1.05773902e+00 1.18681304e-02 2.84522831e-01
-1.17150342e+00 -1.27200353e+00 1.38086930e-01 -9.13317144e-01
9.35162544e-01 -1.36397123e-01 6.47777319e-01 -9.38085839e-02
-7.26839244e-01 -5.59551790e-02 -4.40961838e-01 -4.30571698e-02
-6.16120875e-01 -2.61417955e-01 2.97302961e-01 -2.31163248e-01
4.53045443e-02 -3.84741992e-01 -4.92950439e-01 9.62209404e-01
-1.58529326e-01 1.15454584e-01 8.26262414e-01 6.19381964e-01
4.66886044e-01 4.40192431e-01 4.92410243e-01 -5.30050457e-01
3.10029358e-01 -5.01688421e-01 -6.44265592e-01 2.52545387e-01
-6.17651939e-01 -3.04508954e-01 4.02518362e-01 -3.79967332e-01
-8.21539104e-01 -8.07769671e-02 -1.89586759e-01 -1.28746152e-01
-9.31732580e-02 2.15834573e-01 -9.19802189e-01 7.60056749e-02
-8.39971602e-02 -2.10003614e-01 2.84586132e-01 -7.99227834e-01
-1.85178384e-01 7.72374570e-01 3.59110236e-01 -4.41426367e-01
1.06109416e+00 -4.98937607e-01 1.97321311e-01 -6.93670750e-01
9.57165211e-02 1.35152474e-01 -7.75204360e-01 -5.97119331e-01
9.22263145e-01 -2.76291221e-01 -9.82696056e-01 7.18154669e-01
-1.04932141e+00 3.42781916e-02 4.14338022e-01 6.31858706e-01
-1.05581380e-01 1.44497484e-01 -7.69687057e-01 -1.00126839e+00
-5.02334595e-01 -1.16422558e+00 7.40313411e-01 4.05780077e-01
-8.04902613e-01 -5.20713449e-01 -3.52387816e-01 4.75343645e-01
1.30292445e-01 4.27719802e-01 1.30232930e+00 -5.11126459e-01
-5.33254564e-01 -7.79632926e-01 2.43429989e-01 5.03936052e-01
4.26142395e-01 5.66192746e-01 -3.50342274e-01 -2.29194999e-01
5.70612669e-01 7.12182283e-01 -2.71734923e-01 5.52200794e-01
4.49495465e-01 -2.25272551e-01 -6.48684502e-01 2.45348841e-01
1.63293970e+00 9.38859761e-01 7.17151999e-01 5.52176058e-01
4.02721614e-01 7.09321618e-01 1.27837193e+00 3.17496568e-01
-9.25676897e-02 9.13988173e-01 5.81012011e-01 1.45178154e-01
-6.95038214e-02 -9.83203948e-02 5.54581702e-01 1.13236701e+00
1.81083620e-01 1.56570002e-01 -5.50544739e-01 -2.21986532e-01
-1.31549489e+00 -8.34721848e-02 -5.71199477e-01 2.17785501e+00
4.59033757e-01 4.07184899e-01 -3.57301414e-01 3.11136365e-01
1.03310716e+00 -5.18272400e-01 -4.96451169e-01 -6.31911099e-01
1.60036162e-01 2.39940733e-02 6.32637084e-01 4.50994164e-01
-6.83635414e-01 6.48319840e-01 5.41995192e+00 4.56397861e-01
-1.19089317e+00 -3.32355589e-01 6.30507469e-01 1.38532147e-01
-8.29838142e-02 -2.27200501e-02 -7.94785798e-01 8.87842834e-01
3.65898550e-01 3.39202911e-01 2.16071442e-01 7.34348297e-01
2.72401839e-01 -3.13241661e-01 -1.28921163e+00 7.60496974e-01
-1.66051909e-01 -9.77795005e-01 -5.79675019e-01 -1.34242177e-01
3.86150956e-01 -7.80497015e-01 -8.26495811e-02 -2.85483271e-01
-1.86728090e-01 -8.76845956e-01 5.75332403e-01 5.47970533e-01
4.43191409e-01 -5.06689429e-01 1.14365363e+00 -6.29431233e-02
-8.58602703e-01 -4.01562862e-02 -2.84131199e-01 -1.11844800e-01
7.57558346e-02 8.71250153e-01 -1.24110115e+00 9.66222286e-01
6.09238148e-01 4.35564160e-01 -2.71545321e-01 9.00733471e-01
-1.34642616e-01 4.54693735e-01 -3.23182434e-01 -5.67101657e-01
-4.76562411e-01 -9.32740748e-01 6.39560461e-01 3.76585573e-01
6.84202909e-01 -1.04130253e-01 -3.80581141e-01 9.40480769e-01
6.03378415e-01 -2.11841688e-02 -1.48399144e-01 4.53290820e-01
1.13608122e+00 1.10229015e+00 -8.66892755e-01 7.67639518e-01
-6.02799817e-04 6.78348541e-01 -5.63117981e-01 1.35948867e-01
-1.00145650e+00 -2.58962870e-01 8.08646917e-01 4.06426102e-01
1.87119648e-01 -4.85613972e-01 -9.52651143e-01 -1.60353750e-01
4.96477306e-01 -2.08967179e-01 -3.80254745e-01 -1.11318898e+00
-1.23397565e+00 2.85890013e-01 -1.45223498e-01 -1.27711093e+00
6.97565451e-02 -9.27797854e-01 -7.72315383e-01 7.71197915e-01
-4.00095820e-01 -1.00329506e+00 -5.93956597e-02 -4.18045044e-01
7.67232955e-01 -1.03649519e-01 6.32737160e-01 1.35465279e-01
-1.20016325e+00 5.45714498e-01 4.17646676e-01 -1.70332551e-01
4.48392123e-01 -7.14641213e-01 6.93768561e-02 5.47664046e-01
-4.55413401e-01 5.80462694e-01 1.09661293e+00 -1.15165782e+00
-2.31531715e+00 -8.29702795e-01 6.63047254e-01 -3.25053662e-01
5.17696738e-01 -4.10453111e-01 -6.26534641e-01 2.43387967e-01
-1.19815275e-01 -7.97524333e-01 5.89991629e-01 -1.30869582e-01
6.39472604e-01 -7.32917666e-01 -1.42993987e+00 5.69799542e-01
1.85571641e-01 1.86792791e-01 -3.99309099e-01 1.58378676e-01
3.82624567e-01 -3.75081480e-01 -1.54347277e+00 7.24944055e-01
8.28459024e-01 -4.15685624e-01 1.00104928e+00 1.05999120e-01
3.85936111e-01 -1.77859887e-01 -1.60095274e-01 -7.07131505e-01
-2.34825850e-01 -6.99792624e-01 1.66232795e-01 1.80356181e+00
4.92741853e-01 -8.05114925e-01 1.09262156e+00 9.29763734e-01
-3.84549946e-01 -1.43143904e+00 -1.19259858e+00 -5.24871588e-01
-3.74780536e-01 -6.81191012e-02 6.43661320e-01 6.03731334e-01
-2.75648654e-01 1.30592674e-01 -2.50488341e-01 6.33163452e-01
5.03163338e-01 -9.70325023e-02 3.31881613e-01 -1.39250875e+00
1.35079458e-01 -2.13225721e-03 -5.75650752e-01 -5.75302243e-01
-4.79775704e-02 -3.80921096e-01 1.57271788e-01 -1.52421129e+00
-1.42860219e-01 -8.86577189e-01 -1.83471441e-01 6.04360253e-02
8.67928341e-02 -3.27971250e-01 -1.88079372e-01 1.52476251e-01
2.19352305e-01 2.60932267e-01 1.08457720e+00 3.12026814e-02
-2.66770631e-01 4.09398973e-01 -1.22100972e-01 4.11307037e-01
7.42351294e-01 -4.70418900e-01 -1.85553189e-02 -2.03353539e-01
9.24190953e-02 4.90091145e-01 1.35041028e-01 -9.70630646e-01
2.82949209e-01 -1.84574965e-02 8.55979502e-01 -8.22793126e-01
3.33797514e-01 -1.34660327e+00 8.78050387e-01 6.55946016e-01
2.55875796e-01 2.68781602e-01 -1.61285177e-02 9.22182798e-02
2.56276608e-01 -6.08428180e-01 2.91055977e-01 4.88570243e-01
-1.78141639e-01 -1.56755790e-01 -6.47778988e-01 -1.07582438e+00
1.41345978e+00 -8.34267378e-01 -1.57477289e-01 4.25547898e-01
-5.56128502e-01 -1.08113833e-01 5.84551632e-01 6.39120579e-01
7.76081800e-01 -1.42510951e+00 -4.01152223e-01 4.41851825e-01
-7.11082697e-01 -1.81484431e-01 3.87239814e-01 6.47301078e-01
-9.60786045e-01 1.17090821e-01 -1.55013546e-01 -7.43696749e-01
-1.46206760e+00 1.71530917e-01 1.52773917e-01 8.84486511e-02
-6.11784399e-01 7.32570708e-01 -8.43095005e-01 -2.45175570e-01
-2.41656065e-01 -2.98940182e-01 -2.91198879e-01 -1.58313602e-01
-3.80854130e-01 7.52499580e-01 5.06890595e-01 -1.96780547e-01
-6.44175112e-01 1.04214013e+00 1.01431765e-01 1.97241500e-01
1.51357532e+00 2.58127898e-01 -5.76463163e-01 2.62200952e-01
1.10470879e+00 -8.79321396e-02 -1.24702334e+00 7.15390801e-01
3.92252713e-01 -4.24049377e-01 -3.74970287e-01 -8.42252493e-01
-8.57557416e-01 1.40338987e-01 8.60417247e-01 -1.25363871e-01
7.86070526e-01 8.98654982e-02 7.08824039e-01 3.30002643e-02
4.68115956e-01 -1.18481100e+00 -8.35742503e-02 -7.98465535e-02
9.04505193e-01 -6.44436181e-01 3.88719589e-01 -1.10869253e+00
-2.93879092e-01 1.34536171e+00 9.32062685e-01 -3.73905659e-01
4.96263802e-01 8.08347762e-01 -1.43793076e-01 -4.98082638e-02
-5.01635253e-01 1.07708442e+00 1.70486525e-01 7.32003629e-01
3.94450366e-01 2.42461756e-01 -2.65275598e-01 1.30641866e+00
-5.30375004e-01 -4.95282829e-01 1.25912055e-01 1.19199669e+00
-4.24129397e-01 -1.21065962e+00 -7.58063912e-01 6.49507284e-01
-7.72137940e-02 4.06916887e-01 -2.19400138e-01 8.01855743e-01
3.38244513e-02 1.17587912e+00 2.70212829e-01 -7.55093932e-01
5.47394037e-01 -3.09311241e-01 5.40877998e-01 -2.82877326e-01
-4.31031704e-01 1.10989094e-01 1.81612387e-01 -3.62771809e-01
3.32389206e-01 -8.03086579e-01 -1.02705991e+00 5.68125099e-02
-5.73764920e-01 2.45890826e-01 1.19262075e+00 8.47690701e-01
7.47581571e-02 7.93104589e-01 1.32443488e+00 -6.72077179e-01
-3.97820950e-01 -1.04962039e+00 -6.36696935e-01 -2.08516985e-01
-4.66772795e-01 -8.25734317e-01 -2.80115932e-01 -2.88350195e-01] | [6.492500305175781, 2.75191593170166] |
98708aa0-62a9-46d5-9ac8-b313eeee236d | why-so-pessimistic-estimating-uncertainties | null | null | https://openreview.net/forum?id=wQ7RCayXUSl | https://openreview.net/pdf?id=wQ7RCayXUSl | Why so pessimistic? Estimating uncertainties for offline RL through ensembles, and why their independence matters. | In order to achieve strong performance in offline reinforcement learning (RL), it is necessary to act conservatively with respect to confident lower-bounds on anticipated values of actions. Thus, a valuable approach would be to obtain high quality uncertainty estimates on action values. In current supervised learning literature, state-of-the-art approaches to uncertainty estimation and calibration rely on ensembling methods. In this work, we aim to transfer the success of ensembles from supervised learning to the setting of batch RL. We propose, MSG, a model-free dynamic programming based offline RL method that trains an ensemble of independent Q-functions, and updates a policy to act conservatively with respect to the uncertainties derived from the ensemble. Theoretically, by referring to the literature on infinite-width neural networks, we demonstrate the crucial dependence of the quality of uncertainty on the manner in which ensembling is performed, a phenomenon that arises due to the dynamic programming nature of RL and overlooked by existing offline RL methods. Our theoretical predictions are corroborated by pedagogical examples on toy MDPs, as well as empirical comparisons in benchmark continuous control domains. In the more challenging domains of the D4RL offline RL benchmark, MSG significantly surpasses highly well-tuned state-of-the-art methods in batch RL. Motivated by the success of MSG, we investigate whether efficient approximations to ensembles can be as effective. We demonstrate that while efficient variants outperform current state-of-the-art, they do not match MSG with deep ensembles. We hope our work engenders increased focus into deep network uncertainty estimation techniques directed for reinforcement learning. | ['Ofir Nachum', 'Shixiang Shane Gu', 'Seyed Kamyar Seyed Ghasemipour'] | 2021-09-29 | null | null | null | null | ['d4rl'] | ['robots'] | [-1.53202936e-02 4.12474722e-01 -1.56658083e-01 -1.68232396e-01
-1.07839537e+00 -6.54788673e-01 5.33159256e-01 1.85127124e-01
-5.94942808e-01 1.24872255e+00 -7.84664899e-02 -6.05189025e-01
-4.98628825e-01 -7.16296196e-01 -1.12303925e+00 -7.57302940e-01
-2.51515239e-01 6.40892982e-01 6.03684001e-02 -3.84755552e-01
1.45576522e-01 3.43305558e-01 -1.46466446e+00 -2.48010099e-01
9.67270613e-01 1.16233575e+00 -2.41294742e-01 5.22130966e-01
8.43650755e-03 8.85770321e-01 -5.06159544e-01 -2.01107055e-01
2.50189751e-01 -5.63342333e-01 -5.87907374e-01 -2.25438833e-01
-1.40258903e-02 -6.06598079e-01 -2.56833524e-01 1.02009034e+00
5.49508274e-01 5.37856996e-01 7.27790475e-01 -1.01428926e+00
-2.26753727e-01 1.06040776e+00 -1.85945988e-01 1.14728287e-02
-5.72289340e-03 3.34669411e-01 9.09403026e-01 -3.54750842e-01
2.98000336e-01 1.15028083e+00 6.68932199e-01 7.13178039e-01
-1.33390677e+00 -4.67843831e-01 3.70050609e-01 -5.48814461e-02
-9.00745809e-01 -3.95960182e-01 5.30068576e-01 -3.70687515e-01
7.58976400e-01 -2.27152720e-01 5.42607009e-01 1.14185226e+00
1.94719613e-01 8.46167207e-01 1.32773626e+00 -5.36281168e-01
8.70603740e-01 4.47651893e-02 -2.71624684e-01 5.78481793e-01
2.13926807e-02 8.52853775e-01 -4.22170162e-01 -2.01653107e-03
7.85911977e-01 -5.46171725e-01 -9.32218507e-02 -6.87174916e-01
-7.77443886e-01 9.60508168e-01 1.30998537e-01 6.56642988e-02
-2.52978623e-01 6.61989748e-01 5.26040316e-01 4.31132138e-01
4.71418202e-01 6.35148108e-01 -5.40566385e-01 -7.65633583e-01
-8.13579261e-01 4.41424876e-01 9.27293122e-01 7.35671222e-01
3.04350108e-01 3.17848176e-01 -3.28147233e-01 4.85752910e-01
2.81117618e-01 3.17719162e-01 2.98213542e-01 -1.27431726e+00
3.36759716e-01 1.18207037e-01 4.82181698e-01 -4.39978242e-01
-2.57440656e-01 -6.50618434e-01 -3.45481128e-01 6.79059267e-01
8.04022610e-01 -4.79006499e-01 -7.03163564e-01 2.01642942e+00
3.61307859e-01 1.99770570e-01 2.02467531e-01 5.96552134e-01
-2.09426805e-02 3.24184209e-01 3.95516492e-02 -3.66678774e-01
6.26995564e-01 -5.68137228e-01 -4.21087265e-01 -4.89534773e-02
6.59803450e-01 -1.06881179e-01 1.00366962e+00 7.89950430e-01
-1.22102046e+00 -4.91089761e-01 -1.08005190e+00 5.64333081e-01
-9.24075693e-02 -9.28345919e-02 5.42069018e-01 8.61193597e-01
-8.08656037e-01 1.41240966e+00 -1.06810415e+00 7.85904974e-02
4.27286953e-01 3.97330880e-01 -6.53610379e-02 3.04381609e-01
-1.31587720e+00 1.25366151e+00 6.46708012e-01 1.98619619e-01
-1.37971938e+00 -8.34533513e-01 -6.76469088e-01 4.72101606e-02
9.53792989e-01 -3.34094614e-01 1.89125884e+00 -1.04628730e+00
-2.28135657e+00 1.24026984e-01 6.07748210e-01 -9.84821737e-01
1.00897300e+00 -2.31878117e-01 4.63613756e-02 1.00566395e-01
-4.06415045e-01 4.95042533e-01 9.14713025e-01 -1.21283710e+00
-5.20993948e-01 -1.87996596e-01 2.71975428e-01 7.25578219e-02
-6.19275421e-02 -5.03214478e-01 2.11637422e-01 -2.46151716e-01
-4.16101873e-01 -8.98479462e-01 -4.83751982e-01 -2.22346425e-01
8.31049383e-02 -2.83946753e-01 3.69074494e-01 -2.87772924e-01
1.22699940e+00 -1.79245973e+00 2.97857881e-01 2.71086872e-01
-1.60255000e-01 2.54367590e-01 -7.14190304e-03 5.11812866e-01
1.50350006e-02 2.86543556e-02 -3.40439975e-01 -3.27689946e-01
7.18225598e-01 5.12429655e-01 -6.21423244e-01 5.04401982e-01
1.76362827e-01 8.28170836e-01 -1.10514474e+00 -3.28708053e-01
3.60830396e-01 2.15423062e-01 -5.82650363e-01 2.59619504e-01
-8.02933812e-01 5.64381301e-01 -5.07411480e-01 8.88364241e-02
2.02299520e-01 -1.30353905e-02 2.97608644e-01 1.10285662e-01
-8.71646851e-02 2.36271009e-01 -1.24085903e+00 1.70539939e+00
-7.23733544e-01 1.74896896e-01 1.82183422e-02 -1.31663883e+00
6.79300845e-01 1.85226977e-01 4.60616887e-01 -4.95815635e-01
3.39490116e-01 2.94019282e-01 6.05783984e-03 -1.63718477e-01
3.05834472e-01 -2.89133817e-01 -1.90622658e-01 4.95951831e-01
3.47885489e-01 -6.35154605e-01 2.78232157e-01 -9.05361678e-03
9.58498538e-01 8.47454071e-01 2.44475842e-01 -2.77334005e-01
4.37322199e-01 -2.66001523e-01 4.02813554e-01 1.05117142e+00
-2.90909201e-01 1.57188952e-01 8.90447021e-01 -6.85422793e-02
-1.01349425e+00 -1.13095605e+00 -1.46980897e-01 1.10766923e+00
-3.59527290e-01 -1.83404252e-01 -8.86955440e-01 -8.58757555e-01
2.15890437e-01 1.01695657e+00 -7.24348903e-01 -3.40456843e-01
-3.86803925e-01 -4.45418894e-01 4.52543080e-01 6.51621044e-01
3.28958869e-01 -9.02417183e-01 -7.37527728e-01 6.75969303e-01
4.05125588e-01 -8.68631959e-01 -2.48783398e-02 5.18523991e-01
-8.47655892e-01 -8.92651260e-01 -6.47216499e-01 -4.66341851e-03
2.61408508e-01 -6.40042603e-01 1.17508662e+00 -3.25176835e-01
1.05680816e-01 8.39641571e-01 -3.96166630e-02 -5.18989086e-01
-8.49891484e-01 -6.98164552e-02 3.38632822e-01 -4.18766558e-01
1.37443598e-02 -8.20435524e-01 -4.57716137e-01 1.44486234e-01
-9.04523909e-01 -3.61740470e-01 4.45318818e-01 1.03135908e+00
5.05739570e-01 1.13012582e-01 8.30780685e-01 -6.92071617e-01
7.50747025e-01 -3.50577503e-01 -1.36157644e+00 2.92163998e-01
-9.29779708e-01 6.11306131e-01 8.27355325e-01 -5.16982973e-01
-1.03100717e+00 -9.48996022e-02 -1.89424068e-01 -4.91400480e-01
4.10508178e-02 5.15324712e-01 2.13697359e-01 -2.00308591e-01
7.06383407e-01 1.59442738e-01 2.65523016e-01 -1.45868093e-01
5.64511597e-01 3.53355020e-01 3.73925209e-01 -1.42048192e+00
5.49069524e-01 5.94723932e-02 2.49839991e-01 -3.84905517e-01
-1.04554117e+00 1.46008134e-01 -3.43130320e-01 -4.03404862e-01
5.04575908e-01 -7.18622267e-01 -1.16037738e+00 2.15548411e-01
-6.58440948e-01 -8.89260888e-01 -7.07701385e-01 4.83852178e-01
-1.34069026e+00 3.03367227e-01 -5.37446558e-01 -1.39383924e+00
-6.16431795e-02 -1.23450053e+00 6.98654234e-01 2.59486198e-01
1.61387280e-01 -1.10714352e+00 3.31286550e-01 3.30280550e-02
4.91061807e-01 1.86205521e-01 7.77756155e-01 -7.55702555e-01
-3.36795360e-01 2.11953387e-01 2.92930007e-01 7.20043659e-01
-2.44958341e-01 1.23068251e-01 -8.58198583e-01 -1.90590054e-01
-2.67840885e-02 -8.96788955e-01 8.52030158e-01 3.99414152e-01
1.29056036e+00 -2.03635201e-01 2.08144099e-01 2.43397862e-01
1.40769541e+00 2.12737799e-01 4.22936976e-01 3.49780202e-01
9.18137655e-02 5.47415316e-01 7.26548135e-01 7.85015583e-01
1.47882119e-01 4.15912360e-01 6.52494967e-01 5.83497405e-01
4.52887297e-01 -5.27250886e-01 5.52004993e-01 5.58912456e-01
-1.56682581e-01 3.20236869e-02 -7.49185205e-01 3.43011945e-01
-2.11645913e+00 -9.34064269e-01 5.43054283e-01 2.47216010e+00
1.25676811e+00 5.41735113e-01 3.40377420e-01 -1.21087655e-02
2.60560453e-01 -1.53790023e-02 -8.52943063e-01 -5.32207787e-01
3.34806114e-01 5.26004970e-01 5.06975174e-01 5.12135744e-01
-1.00150609e+00 7.03472674e-01 5.91721010e+00 1.13297439e+00
-8.14346075e-01 -1.99907720e-02 7.89955616e-01 -1.38116434e-01
-2.91518450e-01 -4.50573601e-02 -8.09884965e-01 4.42818135e-01
1.38482416e+00 -1.39552504e-01 8.07797432e-01 1.05172002e+00
1.85569972e-01 -3.32555622e-01 -1.26788795e+00 5.56978941e-01
-4.18215305e-01 -1.22487307e+00 -6.68911397e-01 -7.07943365e-02
1.02845037e+00 -1.38279691e-01 1.46824956e-01 1.03133106e+00
8.93269897e-01 -1.21155167e+00 6.90250576e-01 6.12242877e-01
5.19587100e-01 -1.11309588e+00 6.00739121e-01 5.63533604e-01
-6.81738436e-01 -3.88142258e-01 -4.16006505e-01 2.49504372e-02
1.23350965e-02 5.97335637e-01 -6.39024079e-01 4.31939662e-01
2.60171533e-01 3.98118913e-01 1.93717685e-02 7.04088926e-01
-4.66169029e-01 8.00370276e-01 -5.84777236e-01 -4.01635170e-01
6.32934868e-01 -3.42029214e-01 2.88182735e-01 8.47079158e-01
1.61952600e-01 3.50115113e-02 1.01739928e-01 1.13835049e+00
8.67996067e-02 -1.99242085e-01 -5.30336559e-01 -4.01376933e-01
2.86851913e-01 9.43642914e-01 -3.02403927e-01 -1.53031513e-01
1.51572786e-02 3.59298944e-01 5.90820670e-01 2.22345978e-01
-1.08417594e+00 -1.09637946e-01 4.70512658e-01 -4.12835121e-01
5.00861645e-01 -4.34613436e-01 7.13462196e-03 -9.12379563e-01
-1.52019799e-01 -1.07782912e+00 2.33573034e-01 -3.60505223e-01
-1.29591107e+00 2.28755534e-01 2.46502757e-01 -1.12836599e+00
-8.95708025e-01 -6.54671013e-01 -4.10866171e-01 5.70367754e-01
-1.39157379e+00 -5.90757310e-01 3.10806662e-01 3.44850510e-01
3.06625903e-01 -4.13477756e-02 7.94123530e-01 -2.23072395e-01
-4.53974873e-01 5.05234182e-01 5.43552458e-01 -2.25652307e-01
4.91582811e-01 -1.60890412e+00 1.35947857e-02 5.51709414e-01
1.15839034e-01 2.92701900e-01 9.98114049e-01 -3.49233806e-01
-1.39558268e+00 -6.77391589e-01 -5.10774739e-02 -3.78567100e-01
1.14538229e+00 -2.29220793e-01 -6.89028025e-01 5.27624190e-01
5.54387504e-03 9.52345580e-02 2.51983017e-01 1.88584730e-01
-8.11243653e-02 -1.95128173e-01 -1.16206872e+00 6.08424604e-01
8.27486694e-01 -3.75532538e-01 -6.63757682e-01 2.21579447e-01
6.70964122e-01 -6.41571224e-01 -1.05458629e+00 4.16991711e-01
5.64255297e-01 -1.06806719e+00 7.49837935e-01 -9.08436656e-01
5.96960127e-01 1.30350217e-01 -9.19316262e-02 -1.77516687e+00
3.14848900e-01 -8.99460375e-01 -5.86155117e-01 9.83522654e-01
2.98321337e-01 -7.19441831e-01 7.53308356e-01 8.45578790e-01
-5.05431332e-02 -1.11695862e+00 -1.17569995e+00 -9.65289593e-01
7.42863536e-01 -6.94856882e-01 4.54617471e-01 3.88688445e-01
2.32255444e-01 -1.25184208e-01 -2.06541404e-01 -3.75071764e-02
8.58796597e-01 -3.53651531e-02 6.14632547e-01 -9.01502609e-01
-9.26246464e-01 -6.04964256e-01 -2.54458636e-02 -9.47484434e-01
5.46593428e-01 -2.82203048e-01 3.80851686e-01 -1.01648271e+00
-3.60517770e-01 -7.05872416e-01 -4.78439182e-01 3.46411496e-01
6.60975873e-02 -3.72845531e-01 2.34500796e-01 -4.64052081e-01
-6.63177729e-01 9.46346819e-01 1.11946619e+00 5.90016805e-02
-2.78328300e-01 2.89772719e-01 -3.62479806e-01 6.75818026e-01
9.46264327e-01 -3.13925445e-01 -7.26334155e-01 3.98784950e-02
4.85340416e-01 5.16682923e-01 2.49479592e-01 -1.15662384e+00
1.96587861e-01 -4.29711461e-01 1.72082528e-01 -2.75580257e-01
2.83480197e-01 -7.51351774e-01 -1.76523253e-01 4.10249352e-01
-6.15970612e-01 -2.72287220e-01 4.39515114e-01 7.06171155e-01
2.24829689e-02 -5.66721618e-01 7.23098874e-01 -3.71775240e-01
-4.99646395e-01 2.06620067e-01 -3.24834615e-01 3.88571173e-01
9.03434634e-01 2.32706502e-01 -1.65682629e-01 -5.51419318e-01
-7.85485983e-01 3.82827461e-01 1.99642479e-02 -2.57975869e-02
3.23883057e-01 -1.03072977e+00 -4.14037645e-01 -4.20782827e-02
-2.14405969e-01 -4.39037103e-04 1.93388611e-01 8.01688731e-01
-8.17440525e-02 4.52126741e-01 -6.42606989e-02 -4.04588163e-01
-4.52033937e-01 5.48821807e-01 7.23390162e-01 -5.87814152e-01
-2.41407290e-01 6.85985208e-01 -2.79444754e-01 -5.22425115e-01
5.18261254e-01 -5.43128908e-01 1.42854184e-01 -1.12082042e-01
3.68433625e-01 3.99862081e-01 1.25375614e-01 2.54534513e-01
1.25427600e-02 2.63324171e-01 1.48849458e-01 -5.05621254e-01
1.33630395e+00 1.04348242e-01 4.99817044e-01 5.59314072e-01
7.71592021e-01 -1.25726864e-01 -1.95755017e+00 6.04242496e-02
2.19479632e-02 -1.21680044e-01 9.79315490e-02 -1.06063986e+00
-7.51654744e-01 8.67279649e-01 4.75932866e-01 9.62342694e-02
9.48618233e-01 -4.00218189e-01 2.79776603e-01 5.95526457e-01
7.61383235e-01 -1.54627752e+00 1.05707437e-01 6.45680249e-01
7.13235795e-01 -1.38024819e+00 7.35999793e-02 3.19350004e-01
-6.54118121e-01 1.26619589e+00 6.26533031e-01 -3.48358899e-01
4.68992919e-01 4.21112150e-01 -2.80980021e-01 2.98118889e-01
-1.02250421e+00 -2.47118071e-01 1.12557568e-01 5.48313916e-01
7.58620128e-02 1.67664210e-03 -2.24092647e-01 6.18151307e-01
-3.95125858e-02 3.18482906e-01 5.35062432e-01 8.35311651e-01
-5.84573567e-01 -1.21592808e+00 -1.29597530e-01 1.54126301e-01
-4.29684669e-01 1.37174547e-01 1.57319486e-01 1.01181173e+00
-2.65247434e-01 7.55949318e-01 -1.48322970e-01 -8.22402537e-02
1.21328883e-01 1.37217075e-01 9.79246736e-01 -3.03980738e-01
-6.88554525e-01 -2.19296485e-01 3.65226954e-01 -7.60716856e-01
-2.75850832e-01 -4.60111827e-01 -1.31942213e+00 -1.52843550e-01
-2.48957634e-01 3.32500339e-01 8.23214591e-01 1.16427732e+00
1.09920606e-01 4.60200459e-01 4.84192640e-01 -8.90879750e-01
-1.87241578e+00 -8.55010629e-01 -6.05534971e-01 -2.63653193e-02
3.00384015e-01 -9.24494505e-01 -6.46378934e-01 -6.01312518e-01] | [4.156290054321289, 2.4037327766418457] |
4f6dd3be-50cf-41f9-9ebb-9a36eb370227 | scalekd-distilling-scale-aware-knowledge-in | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhu_ScaleKD_Distilling_Scale-Aware_Knowledge_in_Small_Object_Detector_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhu_ScaleKD_Distilling_Scale-Aware_Knowledge_in_Small_Object_Detector_CVPR_2023_paper.pdf | ScaleKD: Distilling Scale-Aware Knowledge in Small Object Detector | Despite the prominent success of general object detection, the performance and efficiency of Small Object Detection (SOD) are still unsatisfactory. Unlike existing works that struggle to balance the trade-off between inference speed and SOD performance, in this paper, we propose a novel Scale-aware Knowledge Distillation (ScaleKD), which transfers knowledge of a complex teacher model to a compact student model. We design two novel modules to boost the quality of knowledge transfer in distillation for SOD: 1) a scale-decoupled feature distillation module that disentangled teacher's feature representation into multi-scale embedding that enables explicit feature mimicking of the student model on small objects. 2) a cross-scale assistant to refine the noisy and uninformative bounding boxes prediction student models, which can mislead the student model and impair the efficacy of knowledge distillation. A multi-scale cross-attention layer is established to capture the multi-scale semantic information to improve the student model. We conduct experiments on COCO and VisDrone datasets with diverse types of models, i.e., two-stage and one-stage detectors, to evaluate our proposed method. Our ScaleKD achieves superior performance on general detection performance and obtains spectacular improvement regarding the SOD performance. | ['Jian Tang', 'Xiaofeng Mou', 'Zhicai Ou', 'Zhiyuan Xu', 'Ning Liu', 'Qiqi Zhou', 'Yichen Zhu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['small-object-detection'] | ['computer-vision'] | [-2.66669095e-01 9.50973630e-02 4.39722501e-02 -3.99194956e-01
-7.01930583e-01 -5.07866025e-01 4.25746799e-01 1.31280236e-02
-7.04928637e-01 4.24197972e-01 6.56531379e-02 -6.77602366e-02
-1.16883777e-03 -8.65032196e-01 -8.49007249e-01 -6.98543787e-01
4.87288415e-01 2.20847175e-01 9.64359701e-01 -2.17020586e-02
-8.40662569e-02 4.47704196e-01 -1.41839027e+00 2.46660620e-01
1.23256910e+00 1.03081179e+00 3.96202922e-01 6.31176829e-01
-1.85725950e-02 6.20343328e-01 -7.12883055e-01 -5.37260413e-01
1.54895425e-01 -1.40051842e-01 -3.80734146e-01 -2.63703078e-01
8.00884366e-01 -5.60462952e-01 -6.48525417e-01 1.33978355e+00
7.99825668e-01 -6.73070401e-02 6.32011235e-01 -1.22739482e+00
-1.10896826e+00 6.93729222e-01 -7.77172267e-01 6.01252437e-01
-2.70697534e-01 4.63380575e-01 9.54271317e-01 -1.13165784e+00
1.76195234e-01 1.63383198e+00 6.16940856e-01 3.59157741e-01
-1.28521729e+00 -1.13405371e+00 4.07649040e-01 4.84375626e-01
-1.58696699e+00 -1.82085499e-01 5.98347545e-01 -4.33075964e-01
6.49349749e-01 -3.69870141e-02 6.23302460e-01 9.15131807e-01
-3.00744593e-01 1.18296754e+00 9.96427178e-01 -1.45053968e-01
-8.23966265e-02 5.73114455e-01 3.33845496e-01 1.01272511e+00
7.11709738e-01 1.05592966e-01 -4.81023341e-01 1.89287394e-01
8.91959786e-01 8.44232142e-02 -1.16168283e-01 -5.80640256e-01
-8.36170673e-01 8.79742682e-01 9.62454677e-01 4.62666675e-02
-7.11880773e-02 1.07452720e-02 3.42826515e-01 1.29844308e-01
3.33444774e-01 3.60615969e-01 -4.55036759e-01 2.88611084e-01
-8.38043392e-01 1.79969102e-01 4.59748387e-01 1.09521365e+00
7.30390787e-01 1.32010020e-02 -6.28419399e-01 6.38342738e-01
4.85018164e-01 6.48547649e-01 5.14817119e-01 -3.37682903e-01
6.15042686e-01 9.51610506e-01 -1.29613876e-01 -8.27138126e-01
-2.72301704e-01 -8.17277968e-01 -4.26915556e-01 1.02757379e-01
2.45888770e-01 -1.20256796e-01 -1.03115916e+00 1.72852206e+00
8.02674472e-01 4.99779791e-01 2.82050818e-02 1.08781779e+00
1.22237372e+00 5.14744937e-01 2.41497457e-01 2.18960196e-01
1.61463928e+00 -1.46732020e+00 -4.46049958e-01 -2.39663884e-01
6.62027001e-01 -5.72115541e-01 1.06072509e+00 1.06204614e-01
-1.08349299e+00 -8.08623910e-01 -1.16326427e+00 -5.27839899e-01
-3.76281798e-01 5.75518489e-01 4.31376666e-01 3.85612667e-01
-5.62085867e-01 2.80581355e-01 -9.25291479e-01 -1.40932217e-01
1.00486636e+00 2.19854355e-01 -1.69838905e-01 -7.55615905e-02
-1.02199078e+00 1.03353477e+00 6.17170751e-01 1.22702308e-01
-1.10863471e+00 -1.04517198e+00 -8.04123342e-01 4.08431470e-01
6.89216018e-01 -6.32815599e-01 1.29918742e+00 -4.79617178e-01
-1.17879367e+00 6.14700437e-01 2.27377415e-01 -2.38436237e-01
5.85999310e-01 -4.84822333e-01 -1.23920195e-01 1.00620821e-01
5.94301224e-02 8.18304837e-01 8.26269865e-01 -9.66632187e-01
-9.58811164e-01 -5.50331891e-01 1.60373837e-01 3.70555282e-01
-6.21488273e-01 -1.74301311e-01 -6.81938887e-01 -7.11842060e-01
-2.76897512e-02 -6.31985426e-01 6.53316155e-02 4.04603124e-01
-3.46715838e-01 -6.57783866e-01 9.84214962e-01 -5.58450699e-01
1.36191213e+00 -2.19272423e+00 8.30608830e-02 -2.33166635e-01
5.20068765e-01 8.01079869e-01 -3.89720917e-01 -2.06317872e-01
1.36392549e-01 -2.47241318e-01 1.40014291e-01 -2.13911638e-01
9.80835781e-02 2.20308125e-01 -2.87246227e-01 3.84070873e-01
6.76585853e-01 1.20925260e+00 -1.01324511e+00 -6.51402235e-01
8.88378024e-02 4.53062952e-01 -7.66069651e-01 5.05103469e-01
-1.15461528e-01 -8.43765214e-02 -6.73503339e-01 5.34088910e-01
9.10477519e-01 -2.98376918e-01 -2.51863122e-01 -5.38347960e-01
1.55191813e-02 3.51939380e-01 -1.40185285e+00 1.47839081e+00
-2.32006550e-01 3.93754572e-01 2.30047464e-01 -9.63543832e-01
6.03431225e-01 -5.60367443e-02 -2.60038853e-01 -6.20372117e-01
1.33797258e-01 6.54719770e-03 6.05203360e-02 -6.18902564e-01
3.00132453e-01 -5.06948158e-02 1.72612265e-01 7.33040869e-02
4.24716026e-01 -2.92778164e-02 -2.56705750e-02 3.66863132e-01
9.06769395e-01 1.28079444e-01 2.22394183e-01 -4.43243861e-01
4.37011868e-01 -2.03517035e-01 6.91202760e-01 6.73986077e-01
-4.65795666e-01 2.67599583e-01 4.43929940e-01 -7.48860538e-02
-6.57145619e-01 -1.36336780e+00 -1.67913333e-01 1.48990524e+00
4.54981029e-01 -1.75814033e-01 -7.22325802e-01 -1.10513067e+00
4.17979896e-01 3.84829909e-01 -6.14045143e-01 -5.49652219e-01
-5.12622654e-01 -7.04689503e-01 6.99995160e-01 9.87115741e-01
6.39419675e-01 -7.78240442e-01 -3.97141069e-01 -3.17611396e-02
2.40692526e-01 -1.21710467e+00 -7.51237869e-01 2.36778006e-01
-5.63488722e-01 -1.20656860e+00 -5.16865015e-01 -9.47568536e-01
8.69112849e-01 4.83150423e-01 7.75363326e-01 2.87187696e-02
-4.17368770e-01 7.80932158e-02 -5.38231507e-02 -7.54190922e-01
-1.64577775e-02 2.05005541e-01 8.01075101e-02 -2.79873908e-01
5.76705158e-01 -2.90080935e-01 -7.72609174e-01 3.14143062e-01
-9.45461750e-01 8.01251978e-02 1.07458901e+00 8.87251854e-01
3.74796391e-01 -4.26028818e-02 6.30600154e-01 -6.38183355e-01
4.42713410e-01 -3.29027325e-01 -8.14087749e-01 3.03492546e-01
-7.55011439e-01 3.35321784e-01 6.00717187e-01 -7.95113087e-01
-1.14759231e+00 -6.43821657e-02 -4.45704311e-02 -7.14531720e-01
4.87254858e-02 1.44158810e-01 -3.31118345e-01 -1.57323197e-01
6.45964384e-01 2.00271398e-01 -2.76298821e-01 -7.88676441e-01
5.83596468e-01 5.29367447e-01 6.87424481e-01 -5.58742642e-01
1.21333873e+00 3.61957073e-01 -4.56432700e-01 -5.17126560e-01
-1.45348001e+00 -4.83989567e-01 -6.34481549e-01 3.09189260e-01
8.17706168e-01 -1.38262749e+00 -5.90967536e-01 4.96890187e-01
-9.97880816e-01 -8.57391655e-02 -4.14951950e-01 4.66033250e-01
2.97023002e-02 1.07540056e-01 -5.93148828e-01 -4.74505901e-01
-3.76803249e-01 -1.08330536e+00 1.16539669e+00 6.98376894e-01
4.50774491e-01 -6.49561346e-01 1.10218301e-03 4.58693683e-01
2.54035532e-01 -3.84048343e-01 7.37250984e-01 -8.34990144e-01
-7.79323518e-01 4.32103546e-03 -9.91432071e-01 4.23065424e-01
-2.42948189e-01 -2.23619059e-01 -1.19323206e+00 -4.25295323e-01
-3.29456031e-01 -6.43310785e-01 1.23119712e+00 -5.47296964e-02
1.14518428e+00 -3.50944161e-01 -5.36195815e-01 6.05346441e-01
1.17593670e+00 -2.81650841e-01 1.61205947e-01 3.33154500e-02
1.05037260e+00 2.16956347e-01 5.59811175e-01 1.25730634e-01
6.60573959e-01 6.72138512e-01 2.84589529e-01 -9.38425809e-02
-5.13203740e-01 -6.59671187e-01 4.26607728e-01 8.32767844e-01
3.57066065e-01 2.21914291e-01 -5.69205344e-01 6.58890009e-01
-1.83634293e+00 -6.58151448e-01 1.71603143e-01 1.70273530e+00
1.16110682e+00 4.58692193e-01 1.28288537e-01 -2.69985497e-01
6.02910042e-01 9.04349983e-03 -8.21365654e-01 4.89256904e-02
1.42945975e-01 2.85670720e-02 3.49800617e-01 3.57682824e-01
-1.19158804e+00 1.25408828e+00 4.94683123e+00 1.31479549e+00
-1.02832818e+00 4.00829494e-01 1.40446484e-01 -3.16153824e-01
-1.10624894e-01 -3.62298161e-01 -1.44870353e+00 3.89875233e-01
5.24189353e-01 -6.35907101e-03 -3.24442685e-02 1.27426958e+00
-5.90455048e-02 9.30312872e-02 -1.10865891e+00 8.55228662e-01
-9.84209217e-03 -1.21194458e+00 1.41693205e-01 -2.16813087e-01
8.16521406e-01 1.18409187e-01 5.34008965e-02 1.00464940e+00
4.13339376e-01 -6.95175290e-01 8.70327830e-01 2.38643199e-01
5.03677607e-01 -6.03486538e-01 5.54126859e-01 5.29565811e-01
-1.34137893e+00 -3.08832645e-01 -5.09411812e-01 9.72740352e-02
-3.00172865e-01 3.78256738e-01 -8.45143676e-01 2.62390196e-01
7.67573357e-01 4.62430239e-01 -9.29663777e-01 1.20961452e+00
-6.27823353e-01 6.80460155e-01 -3.69400978e-01 -1.47654623e-01
3.88494253e-01 3.28153819e-02 4.62179422e-01 1.38993955e+00
-4.08428237e-02 1.90573543e-01 1.55255541e-01 1.27909994e+00
-3.55621755e-01 -2.60305405e-01 -1.51626945e-01 -2.32078843e-02
8.28193486e-01 1.52719808e+00 -5.12598753e-01 -5.32809138e-01
-6.30598128e-01 8.04152250e-01 8.71356785e-01 1.89775109e-01
-1.13799143e+00 -5.12060463e-01 8.72628331e-01 1.71301290e-02
8.74941230e-01 -1.53612345e-02 -1.12729967e-01 -1.33629251e+00
1.05492480e-01 -5.21036983e-01 5.74565649e-01 -6.50567293e-01
-1.35253096e+00 1.24889553e-01 -9.69210342e-02 -6.28033042e-01
5.46485901e-01 -6.56657696e-01 -7.34986901e-01 8.33090425e-01
-1.91330719e+00 -1.38520765e+00 -4.05921847e-01 4.49932247e-01
5.15699446e-01 1.38790876e-01 2.73963362e-01 5.23041844e-01
-1.03023624e+00 1.12312841e+00 1.23838931e-01 4.04006958e-01
7.60497868e-01 -1.27736998e+00 1.13384590e-01 9.40224290e-01
1.83152080e-01 5.78085721e-01 3.39957237e-01 -5.17711520e-01
-1.31012070e+00 -1.53206038e+00 3.47887605e-01 -6.79875135e-01
7.28339970e-01 -5.28710961e-01 -1.07753539e+00 6.25843227e-01
-2.95511961e-01 4.12525803e-01 2.08549961e-01 2.85394024e-03
-7.57972240e-01 -2.92335778e-01 -1.01725340e+00 5.10475278e-01
1.04357529e+00 -6.49630785e-01 -9.51980114e-01 -2.83513237e-02
1.12030578e+00 -4.36822891e-01 -6.09984696e-01 4.43426430e-01
4.79182959e-01 -5.75821817e-01 1.15180290e+00 -8.48183513e-01
2.60589540e-01 -4.24992293e-01 1.15512215e-01 -1.13053811e+00
-4.79761660e-01 -1.79008767e-01 -6.54569030e-01 1.25074458e+00
1.52949214e-01 -2.89838016e-01 8.58949244e-01 3.06477010e-01
-1.99085265e-01 -1.00532973e+00 -8.08403134e-01 -8.56554031e-01
1.25131130e-01 -1.30443960e-01 5.72654605e-01 8.79941463e-01
-2.68949270e-01 8.25496197e-01 5.41097745e-02 6.60243273e-01
5.40945470e-01 2.70819724e-01 7.61700988e-01 -1.16911113e+00
-3.03586900e-01 -5.37640214e-01 -5.14571786e-01 -1.56068683e+00
-1.66594267e-01 -9.16057587e-01 -1.95540283e-02 -1.30955207e+00
5.93749046e-01 -3.62122387e-01 -5.61124444e-01 5.88683724e-01
-7.81050265e-01 2.46746659e-01 1.36140823e-01 -1.36473209e-01
-1.02927375e+00 8.97451222e-01 1.50603426e+00 -1.11428201e-01
-1.33091211e-01 -2.69448519e-01 -9.57056165e-01 8.26225698e-01
2.41472408e-01 -6.07590139e-01 -4.87458944e-01 -5.82015812e-01
-3.51793095e-02 -4.44480658e-01 6.35766566e-01 -7.81027198e-01
4.06514466e-01 -2.72048600e-02 5.63206553e-01 -6.84512734e-01
1.69796228e-01 -7.24938273e-01 -7.89407372e-01 5.50920606e-01
-2.51471877e-01 -4.50341225e-01 3.61267656e-01 8.06957781e-01
5.78214601e-02 -1.47915587e-01 1.11335623e+00 1.22094408e-01
-8.41701627e-01 4.90589678e-01 3.19947392e-01 2.91823417e-01
1.08241868e+00 -7.27766231e-02 -6.17995203e-01 2.85739422e-01
-6.14837050e-01 6.80532813e-01 2.65455507e-02 6.65941834e-01
5.98718822e-01 -1.43523347e+00 -5.91050029e-01 4.04359192e-01
2.13179454e-01 5.14516950e-01 2.72962540e-01 8.83543730e-01
-1.92864686e-01 3.49341333e-01 6.14971668e-02 -5.53430438e-01
-1.23327327e+00 7.42982805e-01 3.23367536e-01 -1.63521588e-01
-5.77995718e-01 1.32728589e+00 7.86946177e-01 -5.05403221e-01
5.52791834e-01 -6.68784261e-01 -9.77156758e-02 1.54968515e-01
7.12240756e-01 4.54185992e-01 -9.55892205e-02 -2.87101388e-01
-3.56394202e-01 2.96460748e-01 -5.52081466e-01 5.14064908e-01
1.25146937e+00 -1.84333920e-01 2.99528420e-01 3.69769372e-02
1.13256764e+00 -1.57467932e-01 -1.63083422e+00 -7.45483994e-01
-1.43033475e-01 -3.79435420e-01 3.33048761e-01 -8.43311846e-01
-1.00405419e+00 1.11344314e+00 7.21375883e-01 -1.67796686e-01
8.67378652e-01 2.66632974e-01 8.91801953e-01 5.34329355e-01
-1.01420350e-01 -1.09338260e+00 5.09024978e-01 4.68882471e-01
6.83914721e-01 -1.57796955e+00 3.18094864e-02 -4.42576230e-01
-4.46923077e-01 7.31786728e-01 1.34889841e+00 -9.28080902e-02
5.63216567e-01 2.99075603e-01 -1.81352764e-01 -3.24798673e-01
-6.61864638e-01 -3.62881392e-01 6.57955587e-01 3.94613057e-01
-3.83813046e-02 -1.95373446e-02 1.59950912e-01 1.24065876e+00
2.31546685e-02 -1.24312833e-01 1.98946968e-02 7.31514096e-01
-9.80289698e-01 -3.51266205e-01 -3.18623006e-01 4.44138348e-01
-1.14382669e-01 -2.09309131e-01 -2.67448664e-01 8.19623053e-01
5.26112139e-01 5.00392437e-01 -1.23294510e-01 -2.57060498e-01
5.72202027e-01 -7.25260936e-03 5.55310011e-01 -8.28835845e-01
-4.93726730e-01 -3.80082801e-02 -2.39834160e-01 -5.42439818e-01
1.26396939e-01 -2.52555668e-01 -1.31372380e+00 -7.36609474e-02
-9.96808052e-01 1.50341988e-01 3.28720331e-01 9.96546924e-01
4.27001417e-01 6.64678693e-01 2.67018974e-01 -4.68095511e-01
-1.19193506e+00 -1.08383322e+00 -5.56807637e-01 2.62094617e-01
4.87962246e-01 -9.06695068e-01 -6.28785864e-02 -1.92791447e-01] | [9.344827651977539, 1.357892394065857] |
c67313e7-6139-4034-957f-2e4adab9140d | textual-explanations-for-automated-commentary | 2304.08178 | null | https://arxiv.org/abs/2304.08178v1 | https://arxiv.org/pdf/2304.08178v1.pdf | Textual Explanations for Automated Commentary Driving | The provision of natural language explanations for the predictions of deep-learning-based vehicle controllers is critical as it enhances transparency and easy audit. In this work, a state-of-the-art (SOTA) prediction and explanation model is thoroughly evaluated and validated (as a benchmark) on the new Sense--Assess--eXplain (SAX). Additionally, we developed a new explainer model that improved over the baseline architecture in two ways: (i) an integration of part of speech prediction and (ii) an introduction of special token penalties. On the BLEU metric, our explanation generation technique outperformed SOTA by a factor of 7.7 when applied on the BDD-X dataset. The description generation technique is also improved by a factor of 1.3. Hence, our work contributes to the realisation of future explainable autonomous vehicles. | ['Lars Kunze', 'Daniel Omeiza', 'Marc Alexander Kühn'] | 2023-04-12 | null | null | null | null | ['explanation-generation'] | ['natural-language-processing'] | [ 7.23464936e-02 1.03251231e+00 -3.74936521e-01 -5.84018528e-01
-6.68838382e-01 -1.95132449e-01 1.37917137e+00 6.47641122e-02
-1.27371654e-01 8.56956601e-01 3.34371090e-01 -6.92446172e-01
1.01224169e-01 -5.21283150e-01 -9.49836433e-01 -1.49141476e-01
2.65419662e-01 8.14474940e-01 3.48880917e-01 -4.36471522e-01
7.86544383e-02 3.88643026e-01 -2.06922960e+00 4.15469974e-01
8.58400524e-01 8.59062135e-01 9.55261216e-02 5.44662476e-01
-2.46716484e-01 7.91013956e-01 -5.03396809e-01 -5.75246930e-01
3.32076065e-02 -3.62834573e-01 -7.27334440e-01 -1.07273430e-01
5.70420384e-01 -2.12126508e-01 -5.26830219e-02 5.94519734e-01
-7.46395811e-02 9.70177539e-03 5.00638068e-01 -1.76603603e+00
-5.40173531e-01 7.43202806e-01 1.39329329e-01 -4.14040625e-01
-1.27767231e-02 4.45841104e-01 1.11565268e+00 -7.53220499e-01
7.28002727e-01 1.28044116e+00 6.36928797e-01 1.02419770e+00
-1.10315275e+00 -8.04414749e-01 1.99645862e-01 5.61698854e-01
-8.18656504e-01 -5.93555927e-01 4.02578741e-01 -4.26075578e-01
1.86404335e+00 1.43253848e-01 7.70853639e-01 1.32234633e+00
5.37349343e-01 7.75782704e-01 7.66300380e-01 -1.38143390e-01
3.07074785e-01 3.75434995e-01 1.09026797e-01 6.87233329e-01
5.56796610e-01 7.43674517e-01 -5.63969135e-01 2.19768137e-01
-1.22060545e-01 -3.57097924e-01 3.14597607e-01 -4.12873179e-01
-1.16126955e+00 1.03916359e+00 3.71809989e-01 -1.05135866e-01
-5.76625228e-01 6.68075800e-01 6.16070926e-01 2.90783465e-01
4.12973881e-01 5.84644496e-01 -8.71159732e-01 -3.09064895e-01
-8.72301280e-01 6.25185132e-01 8.97367477e-01 1.02435350e+00
8.92876327e-01 5.74580729e-01 -1.24904893e-01 1.64326906e-01
7.03682840e-01 7.41012156e-01 3.08639497e-01 -1.03797138e+00
3.38205189e-01 5.88063359e-01 1.02599293e-01 -5.27499616e-01
-4.75407571e-01 -4.76207703e-01 -5.70605993e-01 6.06691837e-01
-1.32189542e-01 -2.02864796e-01 -1.11696565e+00 1.69780588e+00
-7.84961060e-02 1.51005432e-01 5.18571019e-01 7.75588632e-01
9.67251658e-01 5.36508977e-01 3.05389762e-01 1.84488390e-02
1.22724783e+00 -1.22437584e+00 -9.86300468e-01 -6.87733769e-01
6.65357471e-01 -4.55898970e-01 8.24437261e-01 3.50548476e-01
-7.71533191e-01 -7.77020454e-01 -1.29394841e+00 -1.86912511e-02
-5.16961575e-01 1.30010381e-01 8.44655871e-01 6.75933123e-01
-1.23509943e+00 3.11474711e-01 -7.28912771e-01 -4.05324101e-01
2.78043002e-01 4.70250070e-01 -5.28506398e-01 3.06111276e-01
-1.18755174e+00 1.29679418e+00 4.71134424e-01 -4.87193465e-01
-1.03201830e+00 -8.02566290e-01 -1.15050733e+00 5.41014113e-02
3.78952622e-01 -9.59491432e-01 1.47713101e+00 -9.83350635e-01
-1.77020371e+00 6.45832896e-01 -3.50680619e-01 -1.37929428e+00
4.74328697e-01 -4.35573459e-01 -4.52113986e-01 -2.76332110e-01
7.96189755e-02 1.27560115e+00 5.92320144e-01 -1.40911901e+00
-8.54877174e-01 1.80390812e-02 1.30367696e-01 -2.85847068e-01
3.75523478e-01 -5.64132154e-01 -1.41073931e-02 -2.36153215e-01
-4.88238364e-01 -1.21598697e+00 -1.81364030e-01 -1.88594446e-01
-6.06528997e-01 -2.31338829e-01 9.05689061e-01 -6.42500162e-01
8.82703483e-01 -1.96810722e+00 -6.45539016e-02 3.07879373e-02
4.27563310e-01 6.10466599e-01 -2.50651956e-01 5.16777754e-01
-1.34071335e-01 2.80696601e-01 -5.63389913e-05 -7.65021563e-01
5.73240817e-01 4.60806370e-01 -5.57285428e-01 1.10519864e-01
5.45974314e-01 1.02813864e+00 -6.89435303e-01 -7.34937936e-02
7.00440168e-01 5.29709756e-01 -7.33012497e-01 7.51034021e-02
-7.93559313e-01 8.00564587e-02 -1.58756196e-01 8.57308879e-02
3.92783582e-01 1.60633042e-01 -7.46610984e-02 1.02628708e-01
-3.80012125e-01 6.76259100e-01 -7.21188486e-01 1.48842812e+00
-4.14155304e-01 1.18346310e+00 -5.59385836e-01 -4.04601097e-01
1.19965196e+00 3.95912558e-01 1.87515125e-01 -6.97461307e-01
7.42114484e-02 2.65918195e-01 1.96172222e-01 -3.54032487e-01
8.58388305e-01 -6.63117245e-02 -7.00090230e-02 2.00713769e-01
8.36223140e-02 -1.38287827e-01 7.56695494e-02 2.30489016e-01
9.48681116e-01 1.14286020e-01 4.88511711e-01 -1.25378370e-01
6.68172777e-01 3.23200673e-01 5.01508415e-01 7.06017554e-01
-3.14571828e-01 1.99771121e-01 5.24048984e-01 -7.28570044e-01
-1.10076141e+00 -6.42536521e-01 1.22309461e-01 6.27889335e-01
1.22385733e-02 -7.97233164e-01 -9.16760683e-01 -9.26174462e-01
2.62954921e-01 1.71861029e+00 -5.58813751e-01 -2.95007616e-01
-1.08270004e-01 2.23794535e-01 6.06953382e-01 6.86190188e-01
3.63955587e-01 -1.14613307e+00 -8.74990523e-01 9.14838314e-02
1.00327924e-01 -1.25060034e+00 -1.34254452e-02 3.88345331e-01
-4.71403897e-01 -9.15322423e-01 2.45158970e-01 -2.44752482e-01
9.22218338e-03 2.20252082e-01 1.18526495e+00 2.98927844e-01
4.49754715e-01 2.48196810e-01 -3.87506723e-01 -9.51061070e-01
-9.86230671e-01 2.70882785e-01 1.84610590e-01 -4.12260741e-01
6.57252967e-01 -2.63870135e-02 -1.60922900e-01 9.62268636e-02
-7.10441828e-01 5.06413817e-01 7.70822048e-01 7.17296243e-01
5.34065723e-01 -4.04372603e-01 5.14705896e-01 -8.26338410e-01
4.65947002e-01 -3.00342709e-01 -8.20072711e-01 -8.58610943e-02
-1.19585192e+00 5.52303135e-01 4.94032115e-01 -2.85050962e-02
-1.12320089e+00 2.87568897e-01 -3.42080712e-01 -2.30032399e-01
-5.72535038e-01 3.63023281e-01 -1.74134821e-01 5.59217520e-02
5.31810105e-01 -1.43202499e-01 3.00201058e-01 -3.31286825e-02
8.39296877e-01 5.28237104e-01 3.73234957e-01 9.99019966e-02
8.84045959e-01 1.42003864e-01 1.51023209e-01 -5.83025992e-01
-2.16086030e-01 -1.06875328e-02 -2.89305210e-01 -1.95131555e-01
1.03663981e+00 -1.03003979e+00 -6.53892100e-01 -8.33932683e-02
-1.53783810e+00 -5.77461720e-01 -3.26556653e-01 5.81812203e-01
-7.53462613e-01 7.62898773e-02 -2.04438210e-01 -8.18822384e-01
-2.94927835e-01 -1.34265959e+00 1.17497408e+00 1.43456906e-02
-6.60479128e-01 -6.66516006e-01 -5.45886979e-02 4.28851664e-01
6.36412799e-01 1.00829884e-01 9.31467474e-01 -1.16946483e+00
-6.33311272e-01 -1.13981843e-01 -2.27073953e-01 1.51760563e-01
-3.43507648e-01 9.03242752e-02 -1.16829741e+00 1.63058743e-01
-3.67098123e-01 -4.99072038e-02 1.01907349e+00 3.25113624e-01
5.71323276e-01 -3.74633312e-01 -4.01218772e-01 2.58983523e-01
1.06120765e+00 -6.18214533e-02 6.18059456e-01 6.34810328e-01
4.61207807e-01 7.08261311e-01 8.95083010e-01 3.20957035e-01
7.03968585e-01 9.04673219e-01 1.16033852e+00 1.53702460e-02
-5.65015256e-01 -4.23837721e-01 5.40490925e-01 4.97279793e-01
1.42356977e-01 -4.69166487e-01 -9.43301976e-01 6.45630836e-01
-2.09600592e+00 -9.45223212e-01 -7.08950460e-01 1.86193120e+00
2.56598890e-01 4.29534227e-01 -6.21081665e-02 2.81257983e-02
2.20907092e-01 3.71228196e-02 -4.84465122e-01 -1.05071652e+00
3.89150418e-02 1.16161034e-02 6.39465213e-01 8.76443028e-01
-8.28966737e-01 1.37265384e+00 5.96266747e+00 5.93376040e-01
-1.06388783e+00 1.03023062e-02 1.38702154e-01 1.52043864e-01
-4.83552366e-01 3.66925374e-02 -9.84055936e-01 1.94250405e-01
1.50185752e+00 -9.35270041e-02 3.28612715e-01 1.12718081e+00
4.14064497e-01 1.53516307e-01 -1.08811057e+00 5.41393757e-01
1.48537427e-01 -1.53265417e+00 2.68776536e-01 3.03779900e-01
7.17053294e-01 4.42160904e-01 1.26087278e-01 5.77848136e-01
3.25019032e-01 -9.74964976e-01 1.31849945e+00 4.58037317e-01
4.47797894e-01 -8.20732236e-01 1.16599214e+00 2.21740291e-01
-9.84685421e-01 -1.78599551e-01 -2.87553817e-01 -1.98874637e-01
3.20815980e-01 3.80334288e-01 -1.29113305e+00 5.32831252e-01
1.74170136e-01 7.36097217e-01 -4.37810183e-01 8.30006123e-01
-8.08450639e-01 7.97898710e-01 -9.41784680e-03 -2.88255606e-02
6.14897192e-01 3.82299453e-01 7.43370295e-01 1.36665463e+00
3.09713513e-01 -3.42915744e-01 -1.41756490e-01 1.02971685e+00
1.40347376e-01 -4.19564128e-01 -6.60571635e-01 -2.00728271e-02
5.13731837e-01 9.11002457e-01 -3.04885320e-02 -4.34435159e-01
-2.53896296e-01 6.97728872e-01 -3.44716385e-02 1.53629884e-01
-1.15178668e+00 2.80152559e-02 1.15206754e+00 1.19467840e-01
5.99201024e-01 -1.88136827e-02 -4.55229759e-01 -4.94702369e-01
-1.94991782e-01 -1.01303399e+00 -1.25823542e-01 -8.75292897e-01
-7.27435768e-01 1.06385851e+00 2.65492555e-02 -1.20685875e+00
-1.04009187e+00 -6.98300362e-01 -3.54063928e-01 5.79055846e-01
-1.86000419e+00 -1.46506333e+00 -3.44807774e-01 -3.81885096e-02
8.55690718e-01 -7.12372780e-01 1.01388824e+00 -1.91170815e-02
-2.23750263e-01 4.78290141e-01 -3.67194682e-01 -6.34739637e-01
4.67600584e-01 -1.27601254e+00 1.07600594e+00 7.44141936e-01
2.27456704e-01 4.16265965e-01 1.35309696e+00 -6.66057110e-01
-1.36079168e+00 -1.49678254e+00 1.57427108e+00 -5.92453480e-01
5.17310977e-01 -9.68964472e-02 -7.21122503e-01 9.35123801e-01
6.13674641e-01 -4.25520420e-01 5.70192218e-01 1.27964348e-01
-4.31382358e-01 -1.06885187e-01 -1.10584939e+00 4.34227914e-01
8.43506455e-01 -4.89879966e-01 -7.34590292e-01 1.08446658e-01
1.13228154e+00 -3.27310175e-01 -4.45002526e-01 2.23403782e-01
7.09613621e-01 -1.27239358e+00 5.65620959e-01 -7.02202499e-01
6.09395146e-01 -5.49479067e-01 -1.68979332e-01 -1.40434802e+00
-2.33262017e-01 -7.14990497e-01 -2.10419893e-01 1.06294692e+00
8.16720247e-01 -7.45823622e-01 6.42157316e-01 6.21882081e-01
-7.35874653e-01 -6.10632122e-01 -9.69067812e-01 -6.96407080e-01
-2.20083699e-01 -1.05070615e+00 9.68462765e-01 3.26289266e-01
-8.37673917e-02 4.77896214e-01 -4.56737936e-01 1.77514568e-01
3.60463381e-01 -3.24882984e-01 1.26188195e+00 -1.26917255e+00
-9.81615335e-02 -5.64297915e-01 -7.70991564e-01 -9.87733662e-01
6.44014180e-01 -6.51946485e-01 3.36557537e-01 -1.64303350e+00
-2.71843672e-02 7.46671110e-02 6.08941764e-02 7.19286561e-01
4.29386385e-02 -4.38160636e-03 4.77772564e-01 -1.13486275e-01
-6.24349535e-01 7.12867677e-01 6.02825820e-01 -1.21082239e-01
5.15949726e-02 -8.64393041e-02 -6.64480507e-01 5.64437151e-01
1.01618147e+00 -5.84121764e-01 -4.53401744e-01 -4.74844009e-01
2.57012576e-01 -1.60809487e-01 5.56596756e-01 -1.19142890e+00
2.62626484e-02 9.76495147e-02 -2.64568776e-01 -8.22344959e-01
4.66780514e-01 -8.74581635e-01 4.66800183e-01 7.68367052e-01
-5.19234598e-01 8.81631300e-02 7.64171839e-01 6.68062329e-01
-4.91240807e-02 7.01604709e-02 3.20121497e-01 3.99180114e-01
-1.03385508e+00 -1.18632101e-01 -6.61788762e-01 -5.65725505e-01
1.00976098e+00 -3.20465416e-01 -5.96656442e-01 -6.19716644e-01
-7.67573863e-02 2.02348217e-01 4.88988996e-01 7.84182549e-01
5.33682585e-01 -1.20929384e+00 -7.30406046e-01 3.65178943e-01
4.13459390e-01 -5.69702327e-01 -2.03978181e-01 6.44624531e-01
-3.26823562e-01 1.25036740e+00 -2.97475964e-01 -4.49081153e-01
-1.35467660e+00 3.41027617e-01 2.18355715e-01 -1.50312498e-01
-5.28462529e-01 4.96143788e-01 9.08605382e-02 -5.64794242e-01
2.42333457e-01 -6.98685765e-01 -2.02683151e-01 -4.96315747e-01
3.46767545e-01 2.83626199e-01 1.06930234e-01 -9.57069099e-01
-4.54963237e-01 1.89372886e-03 5.51876277e-02 -1.91460088e-01
1.23521054e+00 -6.73867390e-02 2.90044695e-01 2.51935303e-01
6.98991835e-01 -1.77748710e-01 -1.46414411e+00 2.69208431e-01
1.92936972e-01 4.69798781e-03 1.98905647e-01 -1.20296752e+00
-9.33943033e-01 7.69709826e-01 5.42421818e-01 2.65596271e-01
5.92562795e-01 7.75080547e-03 8.14024389e-01 5.47115803e-01
2.26252437e-01 -6.01162612e-01 -4.00748521e-01 8.52085471e-01
9.24802482e-01 -1.41801512e+00 -2.25086674e-01 -1.54853657e-01
-1.01326609e+00 1.11243653e+00 5.05145490e-01 2.11800411e-01
1.02023065e-01 2.44842097e-01 1.18277468e-01 -3.56255502e-01
-1.31957066e+00 -4.31162864e-01 5.70520580e-01 9.34276819e-01
4.43385839e-01 3.07199091e-01 -3.18327188e-01 9.17029142e-01
-5.44663191e-01 1.61721874e-02 3.03376466e-01 1.43362284e-01
-6.58549905e-01 -1.10396588e+00 9.90455970e-02 1.94522426e-01
4.55173440e-02 -8.65045488e-02 -8.52575839e-01 1.16158044e+00
3.82143110e-01 1.40686131e+00 -7.77779147e-02 -7.51525342e-01
4.09791827e-01 1.07176863e-01 -1.92510292e-01 -4.95771915e-01
-7.83134043e-01 -5.59607625e-01 6.59815967e-01 -9.59119380e-01
-3.09480190e-01 -6.71207130e-01 -1.63934278e+00 -5.21104395e-01
-3.14057052e-01 8.97176936e-02 1.12265730e+00 1.30749619e+00
5.87991416e-01 8.88455272e-01 -6.11297600e-02 -8.66485417e-01
-4.70703334e-01 -8.81036580e-01 -5.23802415e-02 4.92119649e-03
5.38439691e-01 -6.28693402e-01 -2.98947871e-01 5.38320914e-02] | [5.920041561126709, 0.9222071766853333] |
a4d957d8-b1fd-44d3-b12f-c8a530f19876 | graghvqa-language-guided-graph-neural | 2104.10283 | null | https://arxiv.org/abs/2104.10283v2 | https://arxiv.org/pdf/2104.10283v2.pdf | GraghVQA: Language-Guided Graph Neural Networks for Graph-based Visual Question Answering | Images are more than a collection of objects or attributes -- they represent a web of relationships among interconnected objects. Scene Graph has emerged as a new modality for a structured graphical representation of images. Scene Graph encodes objects as nodes connected via pairwise relations as edges. To support question answering on scene graphs, we propose GraphVQA, a language-guided graph neural network framework that translates and executes a natural language question as multiple iterations of message passing among graph nodes. We explore the design space of GraphVQA framework, and discuss the trade-off of different design choices. Our experiments on GQA dataset show that GraphVQA outperforms the state-of-the-art model by a large margin (88.43% vs. 94.78%). | ['Zixuan Liu', 'Yanhao Jiang', 'Weixin Liang'] | 2021-04-20 | null | https://aclanthology.org/2021.maiworkshop-1.12 | https://aclanthology.org/2021.maiworkshop-1.12.pdf | naacl-maiworkshop-2021-6 | ['graph-question-answering'] | ['graphs'] | [ 9.88363773e-02 3.94145101e-01 3.48428963e-03 -6.79774284e-01
-4.04119223e-01 -7.36871779e-01 6.84441566e-01 5.35520792e-01
-1.33159354e-01 1.34695709e-01 4.16697651e-01 -7.91419864e-01
-9.92870554e-02 -1.26872134e+00 -8.49198997e-01 -2.00655475e-01
-2.83500761e-01 5.05822301e-01 2.99512804e-01 -1.97421983e-01
8.82846564e-02 3.48644286e-01 -1.10316277e+00 5.20335734e-01
5.49699724e-01 9.10626590e-01 2.43227743e-02 9.45432603e-01
-7.06488967e-01 1.59770906e+00 -5.50761223e-01 -7.95178950e-01
5.45150787e-02 -4.75196779e-01 -1.23844624e+00 3.17057818e-01
9.70231414e-01 -1.98214188e-01 -7.67099738e-01 1.05028224e+00
8.70362595e-02 6.00484200e-02 5.12412727e-01 -1.64366615e+00
-1.49583805e+00 6.37931764e-01 -4.47577238e-01 2.92839438e-01
7.20016062e-01 2.61889189e-01 1.54748440e+00 -6.48939073e-01
7.14299440e-01 1.59549546e+00 2.20501602e-01 2.09706470e-01
-1.05664563e+00 -2.16896757e-01 3.46996367e-01 4.30784345e-01
-1.41975737e+00 -5.10217063e-02 7.40182877e-01 -4.05171782e-01
1.33739233e+00 4.60536480e-01 8.35672796e-01 5.66996276e-01
3.91614735e-01 7.12900341e-01 8.53002787e-01 -1.70130044e-01
1.50581852e-01 -3.73556167e-01 5.38845301e-01 1.41280854e+00
2.45504737e-01 -4.83717918e-01 -5.00976980e-01 -2.85540700e-01
7.85643935e-01 -1.71297640e-01 -1.57010049e-01 -5.85137665e-01
-1.14841712e+00 1.08348751e+00 1.13220453e+00 -1.79714710e-01
-2.32008561e-01 8.15141678e-01 2.11901098e-01 4.66163695e-01
1.79543808e-01 6.80481553e-01 2.41655577e-02 4.22243506e-01
-9.14557949e-02 1.26373142e-01 8.89608622e-01 1.20273781e+00
9.29597139e-01 6.88956007e-02 -3.80704999e-01 5.06657243e-01
6.39621258e-01 5.43140471e-01 -1.48359656e-01 -1.00707543e+00
5.22686303e-01 1.25732505e+00 -4.67745274e-01 -1.56467056e+00
-2.77832150e-01 -1.32862836e-01 -9.50334907e-01 -2.15398997e-01
3.11730564e-01 2.90501416e-01 -1.06876659e+00 1.61565924e+00
3.88021111e-01 -5.28524891e-02 -2.21375991e-02 9.09826159e-01
1.49201941e+00 9.89423156e-01 4.27866727e-01 3.13366622e-01
1.63910067e+00 -1.04910421e+00 -7.13001847e-01 -3.76676559e-01
6.11302912e-01 -2.68454045e-01 1.42679310e+00 1.44447818e-01
-1.02137458e+00 -4.20452297e-01 -6.81896627e-01 -4.64156598e-01
-5.92675865e-01 -4.87914234e-01 9.18203533e-01 4.45947081e-01
-1.59431112e+00 2.48306505e-02 -3.94572765e-01 -5.77309370e-01
6.30789280e-01 2.65646130e-01 -4.98796046e-01 -1.57026887e-01
-7.72941411e-01 4.15082484e-01 3.19774747e-01 -1.75396763e-02
-7.98720300e-01 -4.54668343e-01 -1.30094862e+00 1.64767891e-01
5.95667541e-01 -1.23861170e+00 1.14798510e+00 -5.91250539e-01
-1.08269775e+00 1.23550320e+00 -1.06139541e-01 -4.75061655e-01
-4.00166698e-02 -9.51471329e-02 -2.94003516e-01 4.47443098e-01
4.70817573e-02 7.28339255e-01 5.19699275e-01 -1.24836373e+00
-2.47104496e-01 -4.77869332e-01 7.39816129e-01 2.18570679e-01
-1.96749829e-02 -7.24314302e-02 -8.80494952e-01 -2.92681664e-01
1.75529450e-01 -8.53143215e-01 -4.00471151e-01 2.15916097e-01
-6.64575040e-01 -6.75361156e-01 7.54501224e-01 -5.00829041e-01
1.05453146e+00 -1.85138059e+00 -8.95408466e-02 2.28332713e-01
8.52278292e-01 -7.77819753e-02 -5.47859967e-01 6.55038297e-01
-8.33113194e-02 3.99904639e-01 -8.34615752e-02 1.12661749e-01
3.51653509e-02 4.26498622e-01 -3.28721195e-01 3.36329728e-01
4.49019700e-01 1.58735943e+00 -1.18168581e+00 -6.99942291e-01
9.83044282e-02 2.47049272e-01 -6.68138742e-01 3.99446487e-01
-7.09846258e-01 -1.99281722e-02 -6.86352968e-01 5.92290342e-01
5.43325305e-01 -1.11206734e+00 3.07816595e-01 -3.02361578e-01
5.28074026e-01 2.09852800e-01 -8.73330593e-01 1.88666844e+00
-8.28747172e-03 8.02422225e-01 -2.28871658e-01 -7.99728692e-01
9.78801548e-01 -7.72923455e-02 1.07620165e-01 -1.02224600e+00
8.46158192e-02 -5.10612607e-01 1.91391390e-02 -8.64624918e-01
6.26769483e-01 3.33839029e-01 -2.46465549e-01 5.41716337e-01
1.34523168e-01 -3.49944621e-01 2.25092873e-01 1.02927673e+00
1.37631547e+00 -1.59357727e-01 4.13656950e-01 -2.98740059e-01
4.07513291e-01 1.03423633e-01 -1.36705860e-01 1.01444900e+00
-7.82986283e-02 4.68342990e-01 7.86424875e-01 -6.94962204e-01
-8.02037001e-01 -1.39533091e+00 5.60008228e-01 1.19844019e+00
2.76847750e-01 -7.54364371e-01 -5.27799547e-01 -8.02949727e-01
1.83504820e-02 6.92423284e-01 -7.07952440e-01 -1.48285866e-01
-5.50706267e-01 -2.66994506e-01 3.29695255e-01 4.31499511e-01
5.08570254e-01 -1.25377023e+00 -4.34850961e-01 -1.55380845e-01
-1.21167988e-01 -1.49434662e+00 -4.40754563e-01 -4.18445855e-01
-5.96234560e-01 -1.42188561e+00 -1.28722578e-01 -1.01725578e+00
8.86466801e-01 7.48554885e-01 2.01471710e+00 6.73421085e-01
-1.10250451e-01 1.01961541e+00 -5.47090828e-01 -1.62815094e-01
-4.87779111e-01 -5.94539568e-02 -5.84968328e-01 3.67450491e-02
1.64958000e-01 -4.74190742e-01 -7.15355515e-01 -2.24343478e-03
-1.05965900e+00 3.39120835e-01 3.56467336e-01 2.56787181e-01
6.82792306e-01 -3.04325253e-01 2.68886685e-01 -1.29156828e+00
9.02954102e-01 -5.34729004e-01 -6.76784158e-01 6.65844381e-01
-2.74261653e-01 2.67554432e-01 4.92641628e-01 -7.26883560e-02
-5.40779114e-01 -1.88709766e-01 1.12244375e-01 -3.10642868e-01
-1.76250681e-01 7.64790714e-01 -2.90916950e-01 -1.01803377e-01
6.53563380e-01 3.49629186e-02 -1.44992054e-01 9.94289946e-03
1.16833591e+00 1.62921712e-01 6.29651546e-01 -4.88897592e-01
7.67355084e-01 5.44486761e-01 2.03896746e-01 -8.60197365e-01
-7.47941077e-01 -5.25427520e-01 -3.90595108e-01 -3.79073173e-01
1.24198020e+00 -7.45680273e-01 -1.11210239e+00 1.35903522e-01
-1.38452768e+00 -4.65762824e-01 -3.77982914e-01 -7.86531717e-02
-4.50245380e-01 4.41932231e-01 -5.80522835e-01 -4.50680614e-01
-4.65015799e-01 -8.90676439e-01 1.05982089e+00 1.54445797e-01
-1.30224064e-01 -9.33915973e-01 -1.61304399e-01 6.24448001e-01
1.91409841e-01 4.19282317e-01 1.26539612e+00 -5.50655723e-01
-1.22475863e+00 3.54206078e-02 -7.96667933e-01 -1.90245211e-01
1.13874070e-01 -6.50851950e-02 -5.49290955e-01 -3.08916390e-01
-5.42609215e-01 -5.21375895e-01 6.42059028e-01 1.07421502e-01
1.59297013e+00 -5.75121880e-01 -1.03489012e-01 5.89170277e-01
1.62674391e+00 2.92475466e-02 7.15940356e-01 6.46038949e-02
1.16205251e+00 3.39506447e-01 2.89185494e-02 1.09402813e-01
9.58941519e-01 1.59438699e-01 9.51516032e-01 -2.14245260e-01
-4.06067550e-01 -5.31240106e-01 -4.27523404e-02 9.94768322e-01
2.87269354e-01 -8.23312342e-01 -1.23517060e+00 4.99604613e-01
-1.84278464e+00 -5.33502102e-01 -6.61939979e-01 1.66610610e+00
2.76051104e-01 -1.33646250e-01 3.73878852e-02 -5.01395762e-01
4.70619500e-01 5.53251982e-01 -4.55083907e-01 -4.91843343e-01
-1.96042821e-01 3.45747098e-02 2.76450723e-01 6.08425736e-01
-9.25594866e-01 1.24514210e+00 6.80503559e+00 2.74256438e-01
-6.49631798e-01 -1.98995382e-01 6.90991640e-01 3.83213431e-01
-6.27655804e-01 2.35758275e-02 -4.27089632e-02 -1.22589014e-01
7.13224351e-01 -2.55735844e-01 4.09883052e-01 6.39326394e-01
-2.10535184e-01 -1.85379107e-02 -1.19549167e+00 1.14922881e+00
2.80382276e-01 -1.76175761e+00 8.58727336e-01 -1.44517213e-01
6.10381484e-01 1.62958369e-01 1.19604403e-02 1.50574878e-01
9.53672349e-01 -1.28719020e+00 6.35039628e-01 4.44093198e-01
5.78155696e-01 -2.87722558e-01 2.82279372e-01 -5.69228381e-02
-1.40131748e+00 1.56257108e-01 -4.20789272e-01 -1.45395607e-01
1.91897571e-01 2.37774566e-01 -1.00339067e+00 7.09084094e-01
8.43493283e-01 5.67321062e-01 -1.06624234e+00 9.54074621e-01
-4.18717742e-01 7.49492168e-01 3.80440764e-02 -3.69231373e-01
5.33658087e-01 -3.98616642e-01 3.29854339e-01 1.17747092e+00
7.65421838e-02 5.52978992e-01 1.97102636e-01 1.14622533e+00
-5.04973352e-01 5.57989590e-02 -1.04034829e+00 -4.12358105e-01
2.20269397e-01 1.23478830e+00 -8.90259862e-01 -4.71184134e-01
-5.89876533e-01 9.32026088e-01 8.06931138e-01 6.42861784e-01
-7.23744631e-01 -7.37662762e-02 3.92810136e-01 1.02281444e-01
8.27369243e-02 -5.57365060e-01 9.30138975e-02 -9.98952508e-01
-9.98286307e-02 -1.05385959e+00 8.69029105e-01 -1.47123528e+00
-1.47600663e+00 7.18348324e-01 -1.16289973e-01 -7.67916441e-01
1.84907634e-02 -6.38680637e-01 -4.40098345e-01 5.01223922e-01
-1.26720250e+00 -1.44645333e+00 -7.26584554e-01 8.64076495e-01
3.53146642e-01 -1.18667506e-01 9.11620796e-01 -2.02837616e-01
-1.25035509e-01 2.24705309e-01 -5.00870228e-01 5.36103964e-01
-3.39363366e-02 -1.42465591e+00 1.05447388e+00 7.28705287e-01
9.09107089e-01 5.11121631e-01 4.33285475e-01 -5.14378190e-01
-2.02366829e+00 -1.22808468e+00 7.81438828e-01 -6.83481634e-01
8.71104121e-01 -6.14678204e-01 -1.13086128e+00 8.81227255e-01
8.01790416e-01 1.54442742e-01 6.67634547e-01 8.72438103e-02
-8.49501967e-01 5.99293821e-02 -7.76212633e-01 8.15549195e-01
1.30879283e+00 -9.73593712e-01 -4.30566639e-01 4.40441370e-01
1.38685763e+00 -2.58288145e-01 -7.04015195e-01 2.10128486e-01
-9.55489092e-03 -9.05423760e-01 1.06186986e+00 -1.01756012e+00
3.49874407e-01 -4.47329551e-01 -3.95273745e-01 -1.09537196e+00
-6.51291847e-01 -6.30167067e-01 -1.76698148e-01 9.31985497e-01
2.84761518e-01 -5.03623188e-01 1.02840304e+00 5.76329470e-01
7.24855661e-02 -7.12992966e-01 -6.06013775e-01 -4.10071731e-01
-3.00427556e-01 -5.17224550e-01 7.91819632e-01 1.09785700e+00
-3.07677150e-01 8.26053977e-01 8.64428878e-02 4.10202682e-01
5.43330848e-01 4.08648372e-01 1.06905353e+00 -9.56968904e-01
-2.23526046e-01 -4.03606325e-01 -7.62305737e-01 -1.22490084e+00
7.81360790e-02 -1.10908937e+00 -3.28422040e-01 -2.34173155e+00
3.19163352e-01 -1.73214003e-01 -1.89268321e-01 6.10252142e-01
-2.29925290e-01 1.91338837e-01 5.55518866e-01 -1.43941820e-01
-1.19909406e+00 4.04509246e-01 1.60838497e+00 -6.38120234e-01
1.30558416e-01 -5.20189345e-01 -8.61214638e-01 5.94099641e-01
6.91017747e-01 -2.42158324e-01 -8.98012042e-01 -1.00429523e+00
8.82991016e-01 2.41584435e-01 6.43563092e-01 -5.99451900e-01
3.58881235e-01 -2.64701724e-01 -9.21635255e-02 -6.33901477e-01
2.68587142e-01 -6.71937108e-01 1.35985203e-02 2.48996064e-01
-5.36512077e-01 6.17487073e-01 1.15165971e-01 7.86828399e-01
-2.50851154e-01 3.91615748e-01 2.73389518e-01 -3.35743815e-01
-1.06048453e+00 5.99083424e-01 -2.62343407e-01 3.56120378e-01
7.89362192e-01 -2.18726009e-01 -7.10999608e-01 -9.15778875e-01
-5.23905575e-01 4.93240446e-01 1.63020745e-01 6.57061756e-01
1.00552142e+00 -1.29713213e+00 -7.38170266e-01 -1.20793261e-01
3.88844371e-01 1.95980772e-01 7.07606301e-02 1.84207648e-01
-8.46580148e-01 3.99022073e-01 -3.62637341e-02 -7.07646608e-01
-1.34129488e+00 5.94237685e-01 3.52177918e-01 -1.02649704e-01
-5.52008927e-01 1.12662816e+00 6.39494061e-01 -5.69090962e-01
1.36613518e-01 -5.59325755e-01 -2.87319630e-01 -2.95496225e-01
2.79878974e-01 3.43265235e-02 -2.00870529e-01 -7.44713008e-01
-3.58411431e-01 4.23664570e-01 8.67762715e-02 2.81922609e-01
1.11327887e+00 -1.25106737e-01 -6.34076357e-01 3.11973482e-01
1.32886100e+00 -3.80523175e-01 -6.71733558e-01 -4.21789467e-01
-1.07741185e-01 -4.92525548e-01 -2.22050980e-01 -6.59398019e-01
-1.08379149e+00 8.80392849e-01 3.43616486e-01 7.65432715e-01
9.12854075e-01 6.94910705e-01 5.58333874e-01 7.39307165e-01
2.98302144e-01 -3.61808866e-01 7.99981654e-01 5.05844831e-01
1.19516718e+00 -1.29707730e+00 4.24441993e-02 -8.24583054e-01
-6.61489248e-01 9.22165930e-01 7.43112504e-01 -2.81626940e-01
6.44725621e-01 -3.13340038e-01 1.69654787e-01 -1.05254364e+00
-7.08636999e-01 -2.85013974e-01 6.63212419e-01 6.36731267e-01
2.90123671e-01 2.74038523e-01 2.19002843e-01 9.90791172e-02
-3.40797782e-01 -2.55418003e-01 3.84960562e-01 6.19320154e-01
-2.92420864e-01 -7.83272624e-01 -2.40404643e-02 5.96089303e-01
-5.77864423e-02 -2.07163513e-01 -8.99893463e-01 8.57419908e-01
-2.51416087e-01 1.24131942e+00 3.84856790e-01 -3.78482312e-01
4.03021574e-01 -3.46557975e-01 4.54827845e-01 -6.37233734e-01
-4.49060768e-01 -5.00592589e-01 2.43822232e-01 -9.31269526e-01
-4.32754010e-01 5.17173558e-02 -1.46021938e+00 -3.80548775e-01
-1.19177565e-01 1.29292384e-01 3.06947649e-01 6.63643420e-01
5.68713009e-01 6.21705472e-01 3.35632026e-01 5.42793376e-03
5.33214286e-02 -4.28146213e-01 -4.09651726e-01 7.94558346e-01
3.08343917e-01 1.18230339e-02 9.86143276e-02 8.41293558e-02] | [10.52593994140625, 1.6839485168457031] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.