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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2a91a47c-6e51-4149-ae02-0ad8f4b2fc3f | memory-maps-for-video-object-detection-and | 2303.03508 | null | https://arxiv.org/abs/2303.03508v1 | https://arxiv.org/pdf/2303.03508v1.pdf | Memory Maps for Video Object Detection and Tracking on UAVs | This paper introduces a novel approach to video object detection detection and tracking on Unmanned Aerial Vehicles (UAVs). By incorporating metadata, the proposed approach creates a memory map of object locations in actual world coordinates, providing a more robust and interpretable representation of object locations in both, image space and the real world. We use this representation to boost confidences, resulting in improved performance for several temporal computer vision tasks, such as video object detection, short and long-term single and multi-object tracking, and video anomaly detection. These findings confirm the benefits of metadata in enhancing the capabilities of UAVs in the field of temporal computer vision and pave the way for further advancements in this area. | ['Andreas Zell', 'Yitong Quan', 'Benjamin Kiefer'] | 2023-03-06 | null | null | null | null | ['video-anomaly-detection', 'video-object-detection'] | ['computer-vision', 'computer-vision'] | [ 1.14436649e-01 -7.54405439e-01 4.33139317e-02 -2.57614106e-02
1.24836743e-01 -9.60537791e-01 7.28857398e-01 2.60696024e-01
-6.08664453e-01 5.75126886e-01 -3.13594699e-01 -1.44119158e-01
-2.72857130e-01 -6.49583995e-01 -6.38978660e-01 -5.80069125e-01
-5.35309374e-01 -1.60114929e-01 8.59310687e-01 2.18334690e-01
2.13275790e-01 8.09718430e-01 -1.93029141e+00 -1.34544954e-01
2.01279253e-01 1.23831356e+00 1.73961997e-01 8.90171707e-01
4.37305212e-01 7.64884472e-01 -5.99135995e-01 -9.68091562e-02
5.15029371e-01 3.35275769e-01 -4.00493860e-01 2.66922325e-01
8.31932545e-01 -6.37867033e-01 -5.79239845e-01 1.27439225e+00
-1.97397798e-01 3.09607029e-01 3.16110045e-01 -1.50131953e+00
-3.77963185e-01 -1.48717523e-01 -6.40906155e-01 8.25611770e-01
5.09449244e-01 3.67163867e-01 4.79222983e-01 -6.31103516e-01
5.17036378e-01 9.41628516e-01 7.60070801e-01 -1.26002118e-01
-4.34883535e-01 -4.82347339e-01 4.77470130e-01 3.94045681e-01
-1.71361482e+00 -2.48826727e-01 8.46506953e-02 -7.94312298e-01
9.04391229e-01 4.26385224e-01 9.24961090e-01 5.86897075e-01
4.05613810e-01 3.35942626e-01 5.58916271e-01 -3.29863757e-01
8.17582086e-02 -1.35857120e-01 -5.88077353e-03 9.75831151e-01
7.00888038e-01 6.02516115e-01 -5.74904144e-01 -2.80636460e-01
7.96873391e-01 3.90854269e-01 -1.29097044e-01 -4.33497846e-01
-1.43927062e+00 4.68491763e-01 3.03136975e-01 2.28751257e-01
-8.06555271e-01 1.79495677e-01 3.14657837e-01 1.52392656e-01
3.96239191e-01 2.20338225e-01 -2.60259241e-01 -1.61845520e-01
-9.85868037e-01 2.64363110e-01 1.95926219e-01 1.25557673e+00
4.40403759e-01 4.08370227e-01 5.44058532e-02 -5.92737719e-02
3.97440791e-01 7.65861034e-01 4.16061014e-01 -8.66496861e-01
1.49464160e-01 6.50317252e-01 5.09913683e-01 -1.34868956e+00
-1.13641351e-01 -2.53821552e-01 -3.86289150e-01 5.09933412e-01
9.80871096e-02 4.74304296e-02 -9.51444268e-01 1.31749046e+00
4.72327054e-01 6.20332062e-01 -5.60103878e-02 7.39878118e-01
2.65845448e-01 7.15347052e-01 2.58071274e-02 -2.31922224e-01
1.28712332e+00 -6.23691916e-01 -7.36930788e-01 -2.63921142e-01
5.92311919e-01 -6.83609068e-01 2.66277134e-01 2.89503962e-01
-6.75063133e-01 -6.63301826e-01 -1.16574013e+00 5.49453914e-01
-6.86841905e-01 3.30487192e-01 6.22871101e-01 4.34390634e-01
-1.16321051e+00 4.00982499e-01 -1.09543681e+00 -6.31279349e-01
4.12669778e-01 4.65258509e-01 -5.01588166e-01 1.00508556e-02
-6.83875680e-01 9.00064766e-01 8.73757839e-01 2.62564451e-01
-1.09826875e+00 -2.99183279e-01 -9.08956230e-01 -4.57954139e-01
4.32818681e-01 -4.28855032e-01 1.02637434e+00 -7.02936053e-01
-6.83832407e-01 7.58762479e-01 -9.21470374e-02 -9.29394364e-01
3.14060330e-01 -4.00147349e-01 -7.05204189e-01 4.81333882e-02
1.23977311e-01 6.64013982e-01 1.17650092e+00 -6.90761864e-01
-1.24574590e+00 -5.85357904e-01 9.21534747e-02 8.64674374e-02
-3.35100353e-01 1.64005309e-01 -2.82997966e-01 -7.29136944e-01
1.99567139e-01 -1.09166551e+00 1.68004236e-03 3.73108447e-01
3.04903686e-01 1.37116984e-01 1.45350802e+00 -6.76386774e-01
1.19396806e+00 -2.23868728e+00 9.65859890e-02 -1.72371566e-02
9.18073878e-02 7.12286115e-01 7.38379210e-02 -5.02466317e-03
1.61668435e-02 -1.99311748e-01 1.95140779e-01 1.98283538e-01
-5.67771673e-01 1.25988200e-01 -5.07624805e-01 6.42066061e-01
3.51659544e-02 4.93368179e-01 -1.00710094e+00 -3.46200675e-01
6.60032928e-01 9.67081115e-02 -1.20791271e-01 5.93038574e-02
-2.78235435e-01 3.29639584e-01 -3.94822568e-01 1.06102669e+00
5.89783788e-01 2.76482373e-01 -1.77720845e-01 -1.41597137e-01
-7.07004130e-01 -3.33786637e-01 -1.35657501e+00 1.24399817e+00
2.32776254e-02 1.06008327e+00 -5.60067445e-02 -3.93037289e-01
7.13507354e-01 3.38465512e-01 6.01355791e-01 -2.49113917e-01
-6.74379466e-04 -2.79180586e-01 -3.34288627e-01 -3.82640421e-01
1.19164205e+00 5.30760348e-01 2.96813458e-01 6.58816770e-02
-3.68377641e-02 1.82568982e-01 3.93543571e-01 1.32428154e-01
1.06625009e+00 1.88935399e-01 5.97577631e-01 1.15876913e-01
4.24470782e-01 4.56245035e-01 2.93013930e-01 6.93124175e-01
-5.59874117e-01 -2.50011850e-02 -4.16918516e-01 -8.94291401e-01
-7.65411019e-01 -1.04209054e+00 -8.16279650e-02 9.51892257e-01
3.74919772e-01 -6.36253297e-01 -3.36577415e-01 -5.61902106e-01
-1.05795823e-03 3.95597190e-01 -4.08089012e-01 -1.59326121e-01
-2.88214296e-01 -6.51876152e-01 4.82405216e-01 6.43559992e-01
5.65043330e-01 -7.32762098e-01 -1.21985161e+00 9.27761719e-02
-1.37868226e-01 -1.38876033e+00 -2.40302220e-01 -2.74342299e-01
-1.24204361e+00 -1.24241757e+00 -2.73115188e-01 -4.50235099e-01
5.63040614e-01 8.48713636e-01 6.65105581e-01 2.40272149e-01
-6.75785422e-01 1.01336658e+00 -4.46146995e-01 -6.83226109e-01
-2.91978754e-02 -5.32020271e-01 7.18058884e-01 -1.29960194e-01
4.16698098e-01 -4.30278294e-02 -2.70658404e-01 4.79262680e-01
-9.22998190e-01 -4.76516604e-01 3.37981462e-01 3.18741560e-01
5.99396527e-01 6.35488033e-01 -3.62193406e-01 2.83598434e-02
1.09370641e-01 -4.12318349e-01 -1.51292837e+00 3.08598191e-01
-5.67606807e-01 -4.26965833e-01 2.37080187e-01 -4.86912668e-01
-5.24762213e-01 2.99424857e-01 6.43342674e-01 -7.09892631e-01
-4.47794676e-01 3.49642456e-01 3.42874527e-01 -6.03185475e-01
4.70352381e-01 3.12762320e-01 -1.29486069e-01 -2.41984148e-02
1.47681236e-01 3.75984520e-01 8.07686985e-01 2.97457092e-02
9.96172726e-01 5.01866937e-01 2.01168209e-01 -1.28811347e+00
-3.64947438e-01 -7.69008040e-01 -1.04955149e+00 -7.37243950e-01
7.35589385e-01 -1.26469469e+00 -4.22074884e-01 6.30588353e-01
-1.17075694e+00 2.22866878e-01 -1.30113408e-01 8.05602372e-01
-2.95918643e-01 5.07011414e-01 9.39319283e-03 -1.03380489e+00
1.47258088e-01 -9.05808687e-01 1.01002383e+00 2.22847193e-01
-1.50817931e-01 -8.63148749e-01 1.36600705e-02 -1.37864091e-02
1.51259229e-01 3.07474792e-01 2.67923534e-01 -2.94835091e-01
-1.27797413e+00 -6.64310396e-01 -1.01028599e-01 1.51531175e-01
1.01704434e-01 4.37827647e-01 -7.30478644e-01 -6.02572799e-01
-1.84302405e-01 2.55473256e-01 6.82398260e-01 5.10518849e-01
7.44096518e-01 -2.71207213e-01 -6.28745079e-01 5.57966530e-01
1.35250163e+00 5.77186584e-01 3.54255885e-01 6.44265652e-01
3.98270577e-01 2.33142748e-01 1.51148033e+00 7.84266055e-01
9.07192081e-02 7.92215705e-01 9.98838842e-01 1.85166448e-01
1.15986988e-01 4.93643731e-02 5.35793841e-01 3.00409675e-01
-4.48484182e-01 -3.98997903e-01 -8.08501899e-01 6.21651947e-01
-1.89108729e+00 -1.25866234e+00 -1.54701337e-01 2.37694693e+00
-1.71176195e-01 -7.53890574e-02 2.32291982e-01 2.44466923e-02
8.75137150e-01 2.68852651e-01 -4.21501160e-01 2.04370484e-01
-4.41072322e-02 -5.41073382e-01 9.14841354e-01 1.96232617e-01
-1.75857639e+00 1.03170538e+00 7.13624573e+00 2.16817304e-01
-1.04371023e+00 -2.06476733e-01 -2.13331416e-01 -9.53605250e-02
7.43771911e-01 -1.00178562e-01 -9.31549370e-01 4.19126689e-01
9.04302657e-01 -2.69097716e-01 3.73293102e-01 9.87484992e-01
8.54675621e-02 -4.71388280e-01 -7.55363107e-01 8.98109913e-01
1.64503917e-01 -1.30125165e+00 2.03672096e-01 1.11030340e-01
4.35937136e-01 6.97290525e-02 2.75442392e-01 -2.41358355e-01
2.32889131e-01 -5.02828419e-01 8.87111723e-01 6.04794264e-01
5.07452250e-01 -5.02474606e-01 1.00832093e+00 6.03037961e-02
-1.52789199e+00 -2.98260510e-01 -5.39140046e-01 -2.78527647e-01
-3.49669158e-02 -1.55397197e-02 -1.09718168e+00 5.70042312e-01
1.04225588e+00 8.77118647e-01 -8.92160833e-01 1.53335845e+00
3.42359543e-01 2.65693575e-01 -4.14829731e-01 1.28400758e-01
4.48497951e-01 -1.34444475e-01 1.05350649e+00 8.56147468e-01
7.20301032e-01 2.12682202e-01 4.41950589e-01 2.61190504e-01
3.91009659e-01 -3.37250024e-01 -1.25968921e+00 -3.37045580e-01
6.14335477e-01 9.45750415e-01 -8.77375782e-01 -1.76301971e-01
-4.14272606e-01 7.09836304e-01 -3.02497268e-01 1.07484765e-01
-1.07598579e+00 -1.05615482e-01 7.37823904e-01 6.99478909e-02
5.18540502e-01 -8.87206256e-01 2.58878112e-01 -8.82963836e-01
-8.78693361e-04 -5.14250338e-01 6.74144387e-01 -8.58550310e-01
-4.81960386e-01 7.44853199e-01 2.81857520e-01 -1.74566495e+00
-2.80352652e-01 -9.61319923e-01 -3.27095360e-01 3.33699673e-01
-1.08583546e+00 -1.30472016e+00 -5.52517474e-01 6.54548764e-01
4.79530603e-01 -6.00338340e-01 7.97510505e-01 9.60289985e-02
-4.13152784e-01 -2.48400588e-02 1.77260004e-02 -6.22248165e-02
5.29469788e-01 -7.26169229e-01 2.83999830e-01 1.63483500e+00
4.92385805e-01 4.80582505e-01 7.63752818e-01 -1.00228751e+00
-1.43479228e+00 -1.48988032e+00 1.42105743e-01 -8.14492524e-01
7.74378359e-01 2.37945527e-01 -4.74079788e-01 1.07160187e+00
-1.12997785e-01 1.23966269e-01 4.73658413e-01 -1.67418852e-01
4.26419377e-02 -4.68847044e-02 -1.11486566e+00 3.64181131e-01
8.21873128e-01 -3.50558728e-01 -5.13850808e-01 4.33520257e-01
5.88326335e-01 -4.81907994e-01 -7.42423713e-01 4.30648118e-01
6.10447586e-01 -7.21683979e-01 1.13901317e+00 -6.05661571e-01
-3.64858925e-01 -8.28861237e-01 -4.96667296e-01 -9.62471068e-01
-4.75769669e-01 -3.58374327e-01 -4.12054956e-01 9.01314020e-01
-1.08400851e-01 -4.17511493e-01 5.20834327e-01 5.12414396e-01
-1.78658754e-01 6.60912320e-02 -8.89371991e-01 -1.14209366e+00
-9.40980256e-01 -3.94311637e-01 4.06082928e-01 5.74640155e-01
-6.03651226e-01 -5.01213551e-01 -5.86191475e-01 1.05010962e+00
7.26170897e-01 -1.24145828e-01 5.17428637e-01 -1.40291142e+00
4.21010315e-01 2.08809692e-02 -1.32320595e+00 -6.40392363e-01
-4.63942736e-02 -3.88680249e-01 -1.44864947e-01 -9.21595454e-01
-2.12008342e-01 -1.23395130e-01 -1.92504913e-01 3.50737453e-01
2.22579837e-01 5.62794268e-01 1.37742609e-01 2.01777667e-01
-1.02939916e+00 1.47536308e-01 5.41773856e-01 -1.36773109e-01
6.52600676e-02 2.33739600e-01 8.67272615e-02 1.12708771e+00
6.57043457e-01 -5.54206610e-01 -2.36058801e-01 -6.00743949e-01
-3.24641407e-01 -1.37971923e-01 8.94011378e-01 -1.56354785e+00
5.98964095e-01 -4.07028675e-01 7.15617120e-01 -1.05208993e+00
5.84119976e-01 -1.36446857e+00 2.50284851e-01 5.66535056e-01
2.59159416e-01 8.49325240e-01 7.83131659e-01 9.92225587e-01
-3.27124238e-01 -2.49898463e-01 6.54958904e-01 -2.68851340e-01
-1.65070617e+00 4.72395629e-01 -7.22913206e-01 -4.20070201e-01
1.73845756e+00 -5.47989666e-01 -2.81700760e-01 -8.23441446e-02
-6.91294968e-01 -2.91610006e-02 7.31569409e-01 6.65927172e-01
9.03372169e-01 -1.16860306e+00 -1.95045039e-01 6.20282054e-01
2.52344400e-01 -3.45085323e-01 2.26121619e-01 8.00431669e-01
-8.29759061e-01 8.56295824e-01 -6.73515677e-01 -9.65550780e-01
-1.81341553e+00 6.87487960e-01 2.06465321e-03 3.01796377e-01
-4.22417462e-01 7.44848549e-01 -2.92198658e-01 2.55206048e-01
5.48017144e-01 -3.68457019e-01 -1.69403389e-01 -6.40305579e-02
8.88035357e-01 7.10170209e-01 -3.62543128e-02 -9.06778693e-01
-6.02091312e-01 6.49114013e-01 -1.03939444e-01 1.13214716e-01
9.60915387e-01 -2.92774945e-01 -2.42162615e-01 3.90422076e-01
3.26477706e-01 -1.40108526e-01 -1.26269078e+00 -6.07318394e-02
2.28069305e-01 -1.01255417e+00 4.51986343e-01 -5.15896261e-01
-7.54880548e-01 3.96873802e-01 1.13749003e+00 3.15441549e-01
1.01740742e+00 -2.84182757e-01 3.23747605e-01 6.96654797e-01
7.95252860e-01 -7.57661343e-01 -7.88312778e-02 5.54344118e-01
5.73528588e-01 -1.24617994e+00 2.54163295e-01 -1.74604535e-01
-4.94936317e-01 1.27930260e+00 7.06310689e-01 1.99119672e-01
5.01841247e-01 2.06972912e-01 -1.38285141e-02 -2.43802965e-01
-6.52851820e-01 -3.05602014e-01 3.42153430e-01 9.40996289e-01
-8.64866450e-02 1.76161656e-03 5.97581387e-01 -1.13123156e-01
2.34816015e-01 3.82226147e-02 5.25091648e-01 1.17202628e+00
-7.17979670e-01 -5.92590332e-01 -7.30206311e-01 3.17529380e-01
-2.45522663e-01 -3.22722904e-02 -3.19507718e-01 9.46976364e-01
2.79086739e-01 9.55643535e-01 2.38420501e-01 -5.60308695e-01
2.50947177e-01 -1.09376043e-01 4.77566868e-01 -3.55108529e-01
2.12604692e-03 -3.74775916e-01 -1.13690488e-01 -6.09884083e-01
-6.15865350e-01 -7.88038790e-01 -6.41616046e-01 -2.59110898e-01
-5.86802363e-01 -8.36349800e-02 8.32486808e-01 7.10872531e-01
4.38937604e-01 4.72563744e-01 2.64325619e-01 -7.68777728e-01
-3.56129616e-01 -5.56175470e-01 -4.46432233e-01 -1.60560876e-01
5.99111438e-01 -1.09925652e+00 -1.43580720e-01 3.01707447e-01] | [6.951349258422852, -1.8705164194107056] |
86bc35c5-2624-4db2-b649-926f2dff6481 | inspired-toward-sociable-recommendation | 2009.14306 | null | https://arxiv.org/abs/2009.14306v2 | https://arxiv.org/pdf/2009.14306v2.pdf | INSPIRED: Toward Sociable Recommendation Dialog Systems | In recommendation dialogs, humans commonly disclose their preference and make recommendations in a friendly manner. However, this is a challenge when developing a sociable recommendation dialog system, due to the lack of dialog dataset annotated with such sociable strategies. Therefore, we present INSPIRED, a new dataset of 1,001 human-human dialogs for movie recommendation with measures for successful recommendations. To better understand how humans make recommendations in communication, we design an annotation scheme related to recommendation strategies based on social science theories and annotate these dialogs. Our analysis shows that sociable recommendation strategies, such as sharing personal opinions or communicating with encouragement, more frequently lead to successful recommendations. Based on our dataset, we train end-to-end recommendation dialog systems with and without our strategy labels. In both automatic and human evaluation, our model with strategy incorporation outperforms the baseline model. This work is a first step for building sociable recommendation dialog systems with a basis of social science theories. | ['Dongyeop Kang', 'Weiyan Shi', 'Shirley Anugrah Hayati', 'Zhou Yu', 'Qingxiaoyang Zhu'] | 2020-09-29 | null | https://aclanthology.org/2020.emnlp-main.654 | https://aclanthology.org/2020.emnlp-main.654.pdf | emnlp-2020-11 | ['movie-recommendation'] | ['miscellaneous'] | [-4.74442482e-01 5.38675070e-01 -4.86379653e-01 -9.05447245e-01
2.02474028e-01 -7.87204444e-01 7.80642152e-01 5.12285829e-02
-4.22607183e-01 8.39853048e-01 1.06730914e+00 8.22102949e-02
-7.61265680e-02 -5.68067670e-01 6.23734072e-02 -3.70997041e-02
4.53846812e-01 6.78138673e-01 2.42070854e-01 -7.97349811e-01
1.73273459e-01 -1.12831742e-02 -1.19480217e+00 4.57092524e-01
5.79128683e-01 6.22969627e-01 2.51638480e-02 7.01672375e-01
-1.34425148e-01 9.64631140e-01 -3.96674007e-01 -1.07746768e+00
-2.45689705e-01 -5.77328682e-01 -1.18202019e+00 -2.48588949e-01
2.06345692e-01 -8.75081182e-01 -2.89251983e-01 4.30018365e-01
4.73372757e-01 7.77251720e-01 9.34060097e-01 -1.23290443e+00
-9.58577812e-01 1.41743875e+00 6.19777977e-01 -2.85354853e-01
9.01436388e-01 -2.29620740e-01 1.38674355e+00 -3.92183840e-01
9.13462222e-01 1.35457778e+00 4.63963538e-01 1.26529479e+00
-9.84555483e-01 -2.23151341e-01 3.15812856e-01 -1.81340054e-01
-6.29316449e-01 -5.83105683e-01 7.15963304e-01 -5.40138781e-01
8.54733169e-01 6.26079619e-01 5.82819104e-01 1.61544740e+00
-2.23177746e-01 6.15817130e-01 6.95327938e-01 -3.38916183e-02
1.25310123e-01 8.34549606e-01 6.84795439e-01 3.16230118e-01
-5.03511988e-02 -2.58009404e-01 -7.44735479e-01 -1.98281273e-01
4.92755234e-01 2.75204957e-01 3.61967906e-02 1.11651555e-01
-1.11414158e+00 7.97293305e-01 5.92500448e-01 5.07003307e-01
-1.44257933e-01 -1.96566164e-01 1.72107965e-01 6.93228364e-01
7.05439091e-01 1.05846953e+00 -3.81947160e-01 -4.06777143e-01
-6.53805062e-02 3.75346005e-01 1.64842558e+00 1.09633911e+00
3.71278942e-01 -4.99605447e-01 -3.73066217e-01 1.14543736e+00
6.87867463e-01 4.82030720e-01 3.05428803e-01 -1.34782958e+00
-1.01918802e-01 7.44299948e-01 5.88526785e-01 -1.09160781e+00
-7.45323479e-01 1.77228197e-01 -3.61766666e-01 -5.32071948e-01
4.71466780e-01 -6.98311567e-01 2.04965845e-01 1.79890990e+00
1.43627673e-01 -3.34765792e-01 2.48988345e-01 8.35912824e-01
1.76303065e+00 3.87061238e-01 3.27671498e-01 -2.89392442e-01
1.31826568e+00 -9.94053006e-01 -9.88147914e-01 1.47215426e-01
8.83139849e-01 -6.41060591e-01 1.25792837e+00 4.67076190e-02
-8.66126955e-01 -5.00379980e-01 -5.57491004e-01 -2.89527923e-01
-6.66031361e-01 1.38607323e-01 8.48665714e-01 6.56515300e-01
-1.13477838e+00 8.63055646e-01 -3.64046007e-01 -1.36564493e+00
-3.40604931e-01 4.84320581e-01 -1.59165800e-01 6.87164366e-01
-1.35089600e+00 1.03356993e+00 -3.89689416e-01 -3.40861678e-01
-1.97047427e-01 -2.48707846e-01 -5.99883258e-01 -1.10398665e-01
1.20995119e-01 -6.25976384e-01 1.58187580e+00 -6.47364259e-01
-2.09784889e+00 7.87608206e-01 2.57489663e-02 -2.49298766e-01
3.10177058e-01 -3.18056375e-01 -5.33042848e-01 -3.17587733e-01
-1.81798264e-01 9.60454464e-01 2.76225507e-01 -9.79862154e-01
-5.03838480e-01 6.92321733e-02 8.31713736e-01 3.36748272e-01
-8.11578631e-01 4.41086769e-01 -1.51060790e-01 -4.36602056e-01
-5.79172730e-01 -1.31325054e+00 -1.81987926e-01 -1.04779027e-01
-3.22777808e-01 -8.42233717e-01 3.38119835e-01 -2.97544748e-01
1.22408116e+00 -1.99240160e+00 5.34137189e-01 -1.25054121e-01
5.38194656e-01 -4.14842628e-02 1.43755034e-01 8.15778732e-01
8.08818519e-01 2.82782018e-01 6.41351819e-01 -9.46887672e-01
4.48119640e-01 5.63955121e-02 -1.76437140e-01 8.71173143e-02
-8.13440323e-01 7.93095767e-01 -1.11738348e+00 -3.50834727e-01
2.83208847e-01 2.58292556e-01 -8.97500396e-01 9.59559977e-01
-3.90832007e-01 7.87173629e-01 -7.54652619e-01 7.54251108e-02
-3.63139547e-02 -4.25828665e-01 5.20096362e-01 -2.00063378e-01
-3.43833901e-02 8.64055216e-01 -4.41732258e-01 1.58920717e+00
-5.48855007e-01 3.71599674e-01 -2.98830755e-02 4.36485223e-02
1.04295433e+00 5.30641437e-01 4.43475813e-01 -2.77994633e-01
5.56428552e-01 -1.85949102e-01 7.02777207e-02 -4.84425247e-01
7.83119977e-01 2.94128805e-01 -3.18531245e-01 9.38830376e-01
2.04338670e-01 -1.28085181e-01 2.83910194e-03 6.70057952e-01
9.32023406e-01 -4.21899036e-02 2.66177595e-01 -3.25039297e-01
5.93930662e-01 -3.59820217e-01 1.06811464e-01 8.54460418e-01
-2.38900393e-01 8.56608748e-02 2.16786399e-01 -2.98733950e-01
-4.40090209e-01 -5.76067448e-01 4.03707981e-01 1.85241544e+00
3.66641641e-01 -8.01036000e-01 -6.33167326e-01 -9.68363881e-01
-1.00945286e-01 9.08445060e-01 -5.67561507e-01 -1.60728693e-01
-3.40370715e-01 -4.99640107e-02 2.71419853e-01 4.70098644e-01
5.65068424e-01 -1.41526246e+00 1.00516632e-01 2.23538294e-01
-3.30333710e-01 -1.03675234e+00 -7.64029741e-01 -3.35767210e-01
-5.99460125e-01 -7.68328011e-01 -5.62021255e-01 -7.16346800e-01
4.22155648e-01 3.81406814e-01 1.26845288e+00 6.67263567e-01
7.64289796e-01 9.02660549e-01 -8.04566145e-01 3.16893831e-02
-5.75368404e-01 3.34172755e-01 4.20137107e-01 -2.53055245e-01
4.92430538e-01 -6.55325413e-01 -8.30361843e-01 1.25107908e+00
-1.27594054e-01 1.32392287e-01 -3.94763201e-01 4.36407387e-01
-4.96963620e-01 -8.25886965e-01 8.96457016e-01 -1.28700960e+00
1.15468490e+00 -6.32548511e-01 -3.11288070e-02 2.94164062e-01
-6.49910927e-01 -4.02721763e-01 6.80594206e-01 -3.81612062e-01
-1.55778325e+00 -1.73728749e-01 -2.87782401e-01 4.02080387e-01
-3.81126255e-01 1.77511528e-01 2.61423960e-02 1.09113932e-01
9.64659989e-01 -6.62834406e-01 -1.03249744e-01 -7.31863260e-01
8.69228065e-01 1.23537028e+00 2.15159748e-02 -6.85463846e-01
3.07432592e-01 8.51530060e-02 -7.50736654e-01 -7.68466771e-01
-9.74344850e-01 -6.72665536e-01 -5.30410528e-01 -7.46060193e-01
1.21891189e+00 -5.50890565e-01 -1.34291780e+00 1.60799906e-01
-1.03863168e+00 -6.69149041e-01 -2.89247204e-02 5.43007791e-01
-4.75114167e-01 1.22253537e-01 -1.17730165e+00 -8.97780478e-01
-5.66177547e-01 -9.23008084e-01 6.34192765e-01 4.14573342e-01
-1.41538787e+00 -1.43122578e+00 1.99108928e-01 7.44988143e-01
9.20434117e-01 -5.91162622e-01 4.54326034e-01 -1.27906787e+00
8.35193172e-02 4.16138755e-05 8.67467895e-02 -5.96475713e-02
4.53788668e-01 -1.51063830e-01 -8.88329566e-01 -5.84695954e-03
-2.96868414e-01 -6.99798584e-01 1.32495165e-01 -7.75142536e-02
7.54105628e-01 -6.63259208e-01 -4.60547507e-01 -8.71150196e-02
3.26800972e-01 2.16627359e-01 3.33165288e-01 -8.75216201e-02
5.11859357e-01 1.10060942e+00 7.95855343e-01 5.17494500e-01
1.02641022e+00 9.96268451e-01 -1.55104294e-01 2.46413589e-01
-1.33895829e-01 -6.27462029e-01 4.25084859e-01 8.81406546e-01
-3.64480913e-01 -9.53378677e-01 -1.17899619e-01 5.18079363e-02
-2.10828805e+00 -9.40976560e-01 1.21921696e-01 1.94999111e+00
8.53217542e-01 -1.77147716e-01 7.46537328e-01 -4.73681748e-01
5.36386549e-01 1.76637590e-01 -4.38834757e-01 -4.99298185e-01
3.44455838e-01 -5.05532861e-01 1.87559694e-01 8.62170398e-01
-9.73033965e-01 1.42191339e+00 6.20818329e+00 7.08668679e-02
-8.82983804e-01 4.69809920e-02 4.42574024e-01 2.10261613e-01
-4.30261433e-01 -3.35009098e-01 -1.04644096e+00 3.35076571e-01
9.94815469e-01 -6.11973405e-02 7.53207862e-01 8.65307212e-01
2.10432172e-01 3.29058200e-01 -1.68332756e+00 5.93582988e-01
7.10232407e-02 -1.12586606e+00 -4.07589227e-01 -1.14783876e-01
4.60302889e-01 -2.82026380e-01 -2.35164881e-01 6.36779845e-01
1.02912951e+00 -6.76138341e-01 3.38698149e-01 5.35073102e-01
2.50904113e-01 1.23975128e-01 7.09159434e-01 2.65203208e-01
-8.64196479e-01 2.51152635e-01 -2.79296219e-01 -3.68928254e-01
4.44981515e-01 -3.95448714e-01 -1.01387358e+00 -3.52827162e-02
4.86822546e-01 1.67103183e+00 -3.79856348e-01 5.45827985e-01
-2.84636587e-01 7.35951304e-01 -1.84718817e-01 -9.87596095e-01
-3.68944854e-01 -4.44049209e-01 4.43646818e-01 1.15232503e+00
6.14874326e-02 6.26736581e-01 3.39593619e-01 6.75851882e-01
-4.43931371e-01 4.26466048e-01 -6.65386736e-01 -3.11043918e-01
6.21022403e-01 1.39157832e+00 -7.45648563e-01 -1.46792263e-01
-4.67334002e-01 1.07726026e+00 3.83979291e-01 2.54143000e-01
-6.04250193e-01 1.53700352e-01 5.69552600e-01 7.00350776e-02
-4.54678893e-01 -4.77090366e-02 -1.96279347e-01 -1.26454401e+00
-6.69320881e-01 -6.79231048e-01 3.59278381e-01 -6.36019051e-01
-1.73365390e+00 6.54194832e-01 -8.63224082e-03 -1.04401207e+00
-5.41114509e-01 -2.56178945e-01 -8.38619292e-01 2.74153948e-01
-6.25047922e-01 -1.24495018e+00 -5.76497734e-01 5.46758890e-01
5.88164389e-01 -2.96032488e-01 1.16777742e+00 4.30929899e-01
-2.68713951e-01 8.37203979e-01 -3.55827987e-01 -1.16546869e-01
1.34475398e+00 -1.26132798e+00 2.46063307e-01 -1.98393166e-01
-1.66493371e-01 1.07698464e+00 9.60359633e-01 -7.83261895e-01
-1.20047116e+00 -5.35017371e-01 1.11221325e+00 -8.41590166e-01
7.98401356e-01 -4.57975984e-01 -5.28455794e-01 9.59619761e-01
5.95399439e-01 -8.44425499e-01 1.18614101e+00 8.44830871e-01
-8.17386806e-02 1.57226264e-01 -1.39610350e+00 8.74030352e-01
1.67072380e+00 -4.64610815e-01 -8.26971471e-01 9.86776352e-01
9.67159212e-01 -1.50492191e-01 -1.20946193e+00 -2.38866881e-02
7.83683062e-01 -7.56323218e-01 9.70975280e-01 -9.16643620e-01
4.46532577e-01 1.72866166e-01 -1.43778235e-01 -1.43513000e+00
-2.29772627e-01 -9.22120571e-01 5.73381968e-02 1.69859374e+00
7.05204844e-01 -6.14027202e-01 6.01820171e-01 1.51828587e+00
-7.48348832e-02 -8.99909884e-02 -2.70152658e-01 -2.74658114e-01
-7.35656321e-02 1.15067519e-01 5.05734444e-01 1.07897818e+00
9.06100810e-01 7.84545183e-01 -1.00827777e+00 -5.61991453e-01
1.29923215e-02 -9.06141624e-02 1.09871566e+00 -1.61924410e+00
-3.56156856e-01 -4.34517413e-01 1.35571763e-01 -1.55456114e+00
4.77871478e-01 -7.36710668e-01 1.23589523e-01 -1.72154570e+00
-1.30249960e-02 -6.98338091e-01 5.29264193e-03 3.70355517e-01
3.98184210e-02 8.27655494e-02 1.81968182e-01 3.42503577e-01
-1.29454446e+00 6.89813375e-01 1.14330208e+00 1.92496985e-01
-6.14081979e-01 4.36868191e-01 -1.05461514e+00 9.00295079e-01
6.68773532e-01 -1.95342943e-01 -5.43868005e-01 1.34458197e-02
4.85619098e-01 2.52626121e-01 -3.03689182e-01 -6.23393178e-01
4.46753353e-01 -3.09807539e-01 -1.68797463e-01 -1.60642236e-01
6.82631493e-01 -9.30050969e-01 9.92075950e-02 -3.43259871e-02
-1.22243714e+00 -2.51847774e-01 -4.07839209e-01 5.21365047e-01
3.65452021e-01 -4.67782170e-02 3.68784279e-01 1.02224104e-01
-3.22459310e-01 2.49313992e-02 -7.07197905e-01 -2.94476539e-01
7.29898512e-01 3.36181968e-01 -6.87103689e-01 -1.39941096e+00
-1.14849377e+00 4.86810029e-01 6.26682460e-01 8.77565145e-01
2.35031873e-01 -1.01937735e+00 -2.73624629e-01 -4.24184114e-01
2.93523431e-01 -8.98019195e-01 1.72576122e-02 4.38690662e-01
9.54457894e-02 1.79579780e-01 -1.50814608e-01 -1.89802982e-02
-1.41248345e+00 1.52205661e-01 3.88622612e-01 1.08662486e-01
-4.29446459e-01 1.03230774e+00 -3.29146869e-02 -9.79178905e-01
7.36771107e-01 -2.33195499e-01 -1.17793822e+00 1.21507682e-01
6.57211900e-01 4.30039793e-01 -4.44050729e-01 -6.22435510e-01
-4.91242446e-02 2.54638016e-01 -2.29386717e-01 -2.42487177e-01
1.17669952e+00 -5.49309075e-01 -2.04267234e-01 5.32726467e-01
8.91898632e-01 2.71599323e-01 -6.43610537e-01 -2.32290834e-01
-1.40316576e-01 -3.75043958e-01 -3.59945685e-01 -1.03695941e+00
-8.83686721e-01 1.89122662e-01 2.90197045e-01 6.22430682e-01
2.44226590e-01 2.66891956e-01 8.84261429e-01 1.08492565e+00
4.24458891e-01 -9.81018245e-01 1.54921174e-01 5.44299126e-01
1.05505073e+00 -1.54672885e+00 -6.75608292e-02 -5.99077940e-01
-1.08624327e+00 1.04296184e+00 1.02861702e+00 6.71166852e-02
1.32350194e+00 -4.46112394e-01 1.39151841e-01 -3.32812905e-01
-8.42349827e-01 3.12964544e-02 4.84379560e-01 2.21629664e-01
1.22101045e+00 2.76737481e-01 -8.40170562e-01 1.20596695e+00
-5.00638008e-01 1.14206374e-01 4.11400557e-01 5.91193080e-01
-4.46869761e-01 -1.27675605e+00 6.08078301e-01 4.90513802e-01
-4.84000146e-02 1.52157292e-01 -1.24408197e+00 1.28938690e-01
-4.80702579e-01 1.87143767e+00 -2.51193821e-01 -1.00848258e+00
5.48510075e-01 -1.64849907e-01 -1.66848257e-01 -8.05198610e-01
-1.51717436e+00 -3.88087004e-01 1.11599422e+00 -5.60546696e-01
-7.79225051e-01 -4.62431133e-01 -1.12257028e+00 -7.34239221e-01
-2.69561678e-01 4.45826501e-01 2.77523518e-01 1.08422303e+00
4.03163016e-01 8.10323954e-02 5.02025008e-01 -7.71387577e-01
-2.09321171e-01 -1.32148528e+00 -4.39152360e-01 8.25938582e-01
-2.23039135e-01 -7.29679227e-01 -5.68917811e-01 -1.01396352e-01] | [12.460420608520508, 7.647930145263672] |
ecff6006-d6ec-4e26-8927-01709fce9d5f | tsa-net-tube-self-attention-network-for | 2201.03746 | null | https://arxiv.org/abs/2201.03746v1 | https://arxiv.org/pdf/2201.03746v1.pdf | TSA-Net: Tube Self-Attention Network for Action Quality Assessment | In recent years, assessing action quality from videos has attracted growing attention in computer vision community and human computer interaction. Most existing approaches usually tackle this problem by directly migrating the model from action recognition tasks, which ignores the intrinsic differences within the feature map such as foreground and background information. To address this issue, we propose a Tube Self-Attention Network (TSA-Net) for action quality assessment (AQA). Specifically, we introduce a single object tracker into AQA and propose the Tube Self-Attention Module (TSA), which can efficiently generate rich spatio-temporal contextual information by adopting sparse feature interactions. The TSA module is embedded in existing video networks to form TSA-Net. Overall, our TSA-Net is with the following merits: 1) High computational efficiency, 2) High flexibility, and 3) The state-of-the art performance. Extensive experiments are conducted on popular action quality assessment datasets including AQA-7 and MTL-AQA. Besides, a dataset named Fall Recognition in Figure Skating (FR-FS) is proposed to explore the basic action assessment in the figure skating scene. | ['Lihua Zhang', 'Chixiao Chen', 'Peng Zhai', 'Dingkang Yang', 'Shunli Wang'] | 2022-01-11 | null | null | null | null | ['action-quality-assessment', 'action-assessment'] | ['computer-vision', 'computer-vision'] | [ 3.30171943e-01 -1.75196141e-01 -2.04680085e-01 -1.83241248e-01
-7.03460157e-01 1.56213701e-01 3.57684016e-01 -3.16074491e-01
-4.13723707e-01 4.83866751e-01 5.42307675e-01 2.64128000e-02
-1.67028412e-01 -7.59858966e-01 -4.06864196e-01 -6.72876835e-01
1.79108769e-01 -1.92342907e-01 7.06604004e-01 -1.48361757e-01
1.15202516e-01 2.98756838e-01 -1.60811388e+00 3.57784688e-01
1.19087350e+00 1.20573068e+00 3.52386348e-02 3.94915611e-01
2.88020112e-02 1.20273244e+00 -4.46800053e-01 -2.65996039e-01
1.69075280e-01 -6.61583245e-01 -8.25364172e-01 5.53389311e-01
5.18432796e-01 -8.18262041e-01 -7.83081174e-01 9.29258525e-01
4.51466084e-01 2.20768556e-01 2.28123873e-01 -1.59854639e+00
-5.41862071e-01 3.35289866e-01 -6.37896359e-01 6.42788470e-01
4.52376574e-01 7.37893105e-01 8.24330807e-01 -7.49454260e-01
2.39044935e-01 1.59529138e+00 5.71493924e-01 3.76810491e-01
-6.54289365e-01 -5.42740047e-01 6.16263986e-01 8.09445024e-01
-1.11515486e+00 -4.91455495e-01 7.76915669e-01 -3.05668443e-01
7.35824883e-01 9.78194252e-02 1.07702017e+00 1.04410768e+00
-4.91900854e-02 1.55594957e+00 9.04843748e-01 -8.08767229e-02
3.11022282e-01 -6.21577322e-01 5.34361377e-02 6.11802697e-01
7.18688443e-02 -1.55878603e-01 -6.12093389e-01 1.01078905e-01
8.85630846e-01 1.54132202e-01 -3.60213339e-01 -3.35970461e-01
-1.37532294e+00 4.07621264e-01 4.43216771e-01 1.59724772e-01
-4.97504026e-01 3.64607275e-01 5.11963010e-01 -7.23554790e-02
3.19300652e-01 -1.59475669e-01 -5.93221895e-02 -3.66030663e-01
-7.26301908e-01 3.84235173e-01 1.10763311e-01 7.15029597e-01
4.26186085e-01 2.55562425e-01 -6.53872609e-01 5.97578168e-01
4.13812369e-01 5.61758220e-01 6.63230181e-01 -1.27814031e+00
6.34333730e-01 9.66877818e-01 3.35593551e-01 -1.28582788e+00
-2.99738199e-01 -2.22202793e-01 -7.02670872e-01 3.05807710e-01
4.60568875e-01 6.07916750e-02 -8.45369697e-01 1.61204410e+00
5.73418975e-01 5.55335879e-01 -1.07446596e-01 1.11569047e+00
8.57343912e-01 5.57677507e-01 7.82159790e-02 -3.10128540e-01
1.15688396e+00 -1.35997272e+00 -9.15853083e-01 -1.93120256e-01
5.32111704e-01 -2.39772961e-01 1.23595953e+00 3.96329999e-01
-1.14898705e+00 -9.45467532e-01 -1.04757726e+00 -6.24282435e-02
5.98026998e-02 2.66611427e-01 4.78628010e-01 4.39872652e-01
-9.30832088e-01 5.54261327e-01 -1.12201524e+00 -3.13196808e-01
8.76372695e-01 1.90253139e-01 -2.75636017e-01 -3.06742936e-01
-1.03539228e+00 4.41746920e-01 3.55304778e-01 4.53227520e-01
-1.12930048e+00 -3.42002243e-01 -8.46609771e-01 4.66108583e-02
8.58493090e-01 -6.01581872e-01 1.24658632e+00 -1.21556342e+00
-1.66912174e+00 3.57477397e-01 -1.63700171e-02 -2.60511696e-01
6.28766179e-01 -6.01434648e-01 -5.68719447e-01 4.05025095e-01
2.19855726e-01 4.62313145e-01 7.10900366e-01 -8.21522713e-01
-8.78918886e-01 -4.50008988e-01 3.63152772e-01 3.66089940e-01
-3.45923960e-01 2.40491256e-01 -5.53841889e-01 -7.68934608e-01
6.86738119e-02 -6.69350624e-01 -1.90934002e-01 3.96749228e-01
-1.52424440e-01 -4.51090157e-01 9.95758116e-01 -6.39696836e-01
1.51453710e+00 -2.07093310e+00 1.22062288e-01 -2.34265417e-01
3.22157949e-01 8.28694105e-01 -3.09738338e-01 1.75550312e-01
7.93848559e-02 -2.51681298e-01 -2.85864651e-01 -8.99536461e-02
-1.08488582e-01 1.99146360e-01 9.67256129e-02 5.04681051e-01
5.55619180e-01 9.71150935e-01 -1.26818097e+00 -8.06152403e-01
3.71177435e-01 2.45124906e-01 -4.80432183e-01 2.42325395e-01
-1.65104836e-01 4.35273886e-01 -8.11627746e-01 8.76858950e-01
5.78684390e-01 -3.34853798e-01 -1.05659373e-01 -3.22567344e-01
-1.46751747e-01 2.40807403e-02 -1.53736067e+00 1.84340048e+00
1.62922412e-01 3.22077751e-01 7.37470388e-02 -8.59526634e-01
6.91055834e-01 1.94602028e-01 8.65414321e-01 -9.65408683e-01
9.16048810e-02 -3.95477302e-02 1.53096020e-01 -7.99450994e-01
3.68963778e-01 2.28413001e-01 2.45472312e-01 3.99397850e-01
5.45596071e-02 6.30477786e-01 4.23799485e-01 2.47204930e-01
1.46445680e+00 5.57094276e-01 2.69711465e-01 -4.86682616e-02
8.12056065e-01 -2.59171456e-01 1.03151405e+00 4.68902141e-01
-8.62091482e-01 5.30137777e-01 3.57461929e-01 -5.78128874e-01
-6.71379566e-01 -9.88835990e-01 2.91314572e-01 8.94069314e-01
4.96034652e-01 -5.20106435e-01 -8.41420293e-01 -8.93084228e-01
-1.82961851e-01 1.71879232e-01 -7.02948928e-01 -3.07061851e-01
-7.68301964e-01 -5.21405220e-01 4.56424206e-01 9.56893444e-01
1.27064657e+00 -1.51348567e+00 -7.91365623e-01 3.23422492e-01
-5.49134195e-01 -1.03721857e+00 -6.86077774e-01 -5.91538846e-01
-8.32023859e-01 -1.33253193e+00 -8.59577417e-01 -3.59307170e-01
3.41370672e-01 6.81999385e-01 7.95815468e-01 5.67063764e-02
-7.66435415e-02 3.77148479e-01 -5.48401058e-01 -6.27604052e-02
1.17355846e-01 -2.76711226e-01 6.08354434e-02 6.60306573e-01
4.08345282e-01 -6.01790786e-01 -9.90583241e-01 6.35230839e-01
-9.95767593e-01 1.07113957e-01 8.43305051e-01 5.63758731e-01
4.91883814e-01 8.46197158e-02 5.98334134e-01 -2.67178744e-01
3.34762752e-01 -2.44595706e-01 -1.65780813e-01 2.86275953e-01
-2.75021464e-01 -2.90340632e-01 3.63826156e-01 -3.39117050e-01
-1.23315489e+00 2.96216141e-02 -5.69484644e-02 -5.52864313e-01
-2.56329566e-01 1.91138312e-01 -7.69422472e-01 8.39246139e-02
4.12187815e-01 3.34377378e-01 -3.75880599e-02 -4.43893760e-01
1.94059029e-01 5.75876832e-01 7.56300926e-01 -2.93932140e-01
5.30404150e-01 6.98016346e-01 -1.74957111e-01 -4.75653440e-01
-1.09941363e+00 -6.12921476e-01 -7.28357434e-01 -6.92252338e-01
9.97375965e-01 -9.33037758e-01 -7.25294769e-01 1.09527826e+00
-9.69013393e-01 -3.35649192e-01 -2.46952832e-01 5.34543812e-01
-5.75419426e-01 7.27234721e-01 -4.96510029e-01 -8.87856901e-01
-3.53871495e-01 -1.20751846e+00 1.14580071e+00 4.76836264e-01
1.08846322e-01 -5.10810196e-01 -4.22763564e-02 6.31902337e-01
2.79148340e-01 4.85410303e-01 3.53345633e-01 2.38322318e-02
-7.20234931e-01 1.36037916e-01 -6.07148767e-01 5.03706753e-01
3.42989147e-01 -2.05171689e-01 -8.06058347e-01 -1.45356461e-01
-1.53387293e-01 -3.66734803e-01 9.47451949e-01 3.72180432e-01
1.20779002e+00 -2.64323741e-01 -9.04066265e-02 3.79717708e-01
9.24009860e-01 2.68250585e-01 8.83988380e-01 4.69106257e-01
1.03936589e+00 3.23437572e-01 1.07945168e+00 5.51358521e-01
4.02478337e-01 8.75820279e-01 6.83223963e-01 -1.22114636e-01
-2.34536022e-01 -1.95214093e-01 7.44202852e-01 8.02126706e-01
-4.13015425e-01 -2.05988571e-01 -6.98351204e-01 7.29558527e-01
-2.30813837e+00 -1.12972963e+00 -2.48853803e-01 1.89699256e+00
4.92745876e-01 1.52812049e-01 5.66992760e-01 4.30963784e-01
7.38236368e-01 4.24473792e-01 -8.72068346e-01 2.42541999e-01
6.36993442e-03 -2.35580325e-01 5.54694869e-02 2.50272769e-02
-1.31129992e+00 5.93235016e-01 5.27810001e+00 1.00648820e+00
-6.44891143e-01 1.18090779e-01 4.43256319e-01 -1.02740720e-01
1.51470289e-01 -1.93078458e-01 -4.88131672e-01 7.42661953e-01
5.87871015e-01 -8.94955825e-03 1.15762666e-01 8.38512003e-01
6.26225650e-01 -2.59697407e-01 -8.68079722e-01 1.05539453e+00
2.14366242e-01 -8.85207474e-01 1.63192928e-01 3.08212172e-03
4.74546939e-01 -2.47953847e-01 -1.10635243e-01 3.10658783e-01
1.09524235e-01 -7.79507041e-01 7.43007183e-01 7.37570047e-01
5.19075692e-01 -7.81625330e-01 7.47101247e-01 1.19719118e-01
-1.49965525e+00 -3.76948953e-01 -2.54589736e-01 -1.23446889e-01
4.48513597e-01 3.34505320e-01 1.40220657e-01 7.44808733e-01
9.41275716e-01 1.25090182e+00 -8.70918453e-01 1.38025761e+00
-1.05369970e-01 7.37240076e-01 -1.51351541e-01 3.36755693e-01
6.13187730e-01 -2.53808618e-01 5.33113420e-01 8.21232319e-01
2.44850233e-01 3.19435090e-01 4.41921890e-01 6.11105978e-01
1.96890041e-01 3.12948227e-02 -1.88607037e-01 3.68871391e-02
-5.50037660e-02 1.07886159e+00 -7.91155159e-01 -4.83688891e-01
-6.16833508e-01 9.92016375e-01 1.39086470e-01 1.51503935e-01
-1.03805900e+00 -2.62171477e-01 6.02723241e-01 1.42635494e-01
4.57603484e-01 -5.20977676e-02 2.79635847e-01 -1.29306769e+00
3.72796386e-01 -1.10413218e+00 4.88211244e-01 -1.02286232e+00
-9.84595776e-01 4.02296484e-01 1.18922873e-03 -1.63259327e+00
3.12463671e-01 -3.94048154e-01 -7.03833282e-01 3.07535857e-01
-1.36607301e+00 -1.26954842e+00 -8.33662331e-01 7.83800483e-01
8.65493953e-01 3.13517042e-02 2.00863034e-01 5.34367323e-01
-9.95151222e-01 4.60399956e-01 -2.06494808e-01 2.70430714e-01
5.12688458e-01 -1.04223466e+00 3.87099624e-01 1.09396088e+00
-8.79740622e-03 2.07830831e-01 1.19086884e-01 -7.12960958e-01
-1.22983420e+00 -1.33147037e+00 4.91117656e-01 -3.94169271e-01
5.59865534e-01 7.60533586e-02 -1.09825337e+00 4.06079739e-01
9.04650167e-02 3.34053338e-01 3.52960736e-01 -3.06731433e-01
-6.80469647e-02 -2.96873838e-01 -8.40921164e-01 5.49039423e-01
1.52136958e+00 -1.48339778e-01 -4.81067032e-01 1.87406123e-01
6.14038348e-01 -2.98765481e-01 -6.65103734e-01 5.37537456e-01
6.09531879e-01 -1.22017002e+00 8.55866849e-01 -5.83353698e-01
4.94142890e-01 -7.51982808e-01 -4.82353345e-02 -9.03728187e-01
-6.02758884e-01 -5.73397756e-01 -5.59635222e-01 1.20762336e+00
-3.00221920e-01 -1.90330550e-01 8.20823371e-01 4.57103312e-01
-3.69898796e-01 -8.52559090e-01 -1.03639352e+00 -8.12465906e-01
-3.98187816e-01 -5.18733501e-01 6.82553649e-01 5.91595471e-01
-2.40455076e-01 2.40484223e-01 -6.54530048e-01 1.07359048e-02
5.88184237e-01 -1.30059227e-01 7.87073493e-01 -1.07049334e+00
-3.72856677e-01 -3.38738322e-01 -7.28369653e-01 -1.18538857e+00
-2.18959093e-01 -3.50247890e-01 1.05802894e-01 -1.73797393e+00
3.45472991e-01 -5.52769713e-02 -6.15263939e-01 5.62500298e-01
-4.87056583e-01 2.52754718e-01 3.30856115e-01 1.92920804e-01
-1.41579759e+00 1.09765279e+00 1.56503737e+00 -1.50919810e-01
-8.64295959e-02 -5.24897389e-02 -4.46155787e-01 9.76181507e-01
6.03866100e-01 -2.56267190e-01 -4.71556664e-01 -4.75163072e-01
-1.47691831e-01 1.11221060e-01 6.34882331e-01 -1.52175283e+00
4.59954105e-02 -3.58757347e-01 2.31867045e-01 -7.72318840e-01
1.79700747e-01 -6.44740820e-01 -2.07895592e-01 4.24385011e-01
-1.49396554e-01 9.58579332e-02 -7.26043656e-02 9.99816358e-01
-3.38346362e-01 7.77930543e-02 7.35694706e-01 -9.60391685e-02
-1.06653023e+00 6.51806533e-01 -4.31252152e-01 1.52147099e-01
1.11695778e+00 -4.43251967e-01 -3.09700429e-01 -2.39281178e-01
-5.39677918e-01 4.33436751e-01 2.63828039e-01 5.01076698e-01
7.69809365e-01 -1.83890915e+00 -7.33902454e-01 -2.59287320e-02
3.08633059e-01 6.86024204e-02 7.24825263e-01 1.17696488e+00
-3.03902954e-01 1.81164086e-01 -4.35380369e-01 -6.54556811e-01
-1.16261089e+00 5.31410873e-01 2.38698423e-01 -2.43683994e-01
-9.03395474e-01 4.34356004e-01 2.98400670e-01 5.34521267e-02
2.65369833e-01 -4.62773383e-01 -3.98816288e-01 -1.42233044e-01
8.98020625e-01 7.83518791e-01 -2.77075201e-01 -1.00087929e+00
-4.08385456e-01 4.85567003e-01 8.55849162e-02 1.43639416e-01
1.13042378e+00 -2.64676243e-01 2.49947160e-01 2.53246188e-01
7.54929423e-01 -4.41302121e-01 -1.73658943e+00 -3.79338056e-01
-7.11107031e-02 -7.99254239e-01 1.72398895e-01 -6.44009769e-01
-1.35205054e+00 8.36434662e-01 7.70076156e-01 -3.96302193e-02
1.28763962e+00 -2.80341923e-01 1.10460103e+00 3.17783684e-01
2.70403326e-01 -1.21353507e+00 7.31475234e-01 2.34933004e-01
9.25508261e-01 -1.24615669e+00 4.05786000e-02 -1.62996069e-01
-7.26319194e-01 8.02124143e-01 1.12122750e+00 -2.51904696e-01
2.97035992e-01 -2.10374936e-01 -2.94523444e-02 -1.94817603e-01
-5.38168967e-01 -4.71685469e-01 2.53487229e-01 6.76397920e-01
-1.14897946e-02 -2.92713791e-01 -3.43996346e-01 7.50352740e-01
4.51522499e-01 3.92794847e-01 3.84706229e-01 1.06411827e+00
-5.08275926e-01 -7.80382752e-01 -1.91972896e-01 4.75501478e-01
-3.97264987e-01 2.26892292e-01 -2.78812885e-01 6.66867137e-01
2.93576181e-01 1.06054759e+00 -1.37410760e-01 -4.27320898e-01
5.26695549e-01 -2.62020677e-01 3.15341145e-01 -3.12925577e-01
-4.43014324e-01 1.14109047e-01 7.27923438e-02 -1.26189864e+00
-9.65410829e-01 -9.81709838e-01 -1.03837264e+00 -1.28871296e-02
-2.79351413e-01 -1.36064574e-01 -6.53240010e-02 1.11125052e+00
4.25691903e-01 8.17464054e-01 4.05487686e-01 -8.09643567e-01
-2.13166833e-01 -1.10133398e+00 -6.66457593e-01 4.40129429e-01
2.46215999e-01 -1.00634265e+00 -7.08159953e-02 1.32655218e-01] | [8.378626823425293, 0.7033786773681641] |
3a167890-7073-4cfa-b927-29ec8cbd98b6 | hierarchical-ranking-for-answer-selection | 2102.00677 | null | https://arxiv.org/abs/2102.00677v1 | https://arxiv.org/pdf/2102.00677v1.pdf | Hierarchical Ranking for Answer Selection | Answer selection is a task to choose the positive answers from a pool of candidate answers for a given question. In this paper, we propose a novel strategy for answer selection, called hierarchical ranking. We introduce three levels of ranking: point-level ranking, pair-level ranking, and list-level ranking. They formulate their optimization objectives by employing supervisory information from different perspectives to achieve the same goal of ranking candidate answers. Therefore, the three levels of ranking are related and they can promote each other. We take the well-performed compare-aggregate model as the backbone and explore three schemes to implement the idea of applying the hierarchical rankings jointly: the scheme under the Multi-Task Learning (MTL) strategy, the Ranking Integration (RI) scheme, and the Progressive Ranking Integration (PRI) scheme. Experimental results on two public datasets, WikiQA and TREC-QA, demonstrate that the proposed hierarchical ranking is effective. Our method achieves state-of-the-art (non-BERT) performance on both TREC-QA and WikiQA. | ['Tiegang Gao', 'Renhong Cheng', 'Mengting Hu', 'Hang Gao'] | 2021-02-01 | null | null | null | null | ['answer-selection'] | ['natural-language-processing'] | [-2.70231552e-02 -9.42168534e-02 -2.17018202e-01 -5.59601128e-01
-2.02467394e+00 -5.76190412e-01 6.48725569e-01 1.46385193e-01
-5.07795691e-01 8.64242971e-01 4.74292874e-01 -1.67691901e-01
-6.76922798e-01 -5.60675323e-01 -3.46186459e-01 -6.57880545e-01
2.06170484e-01 9.15376604e-01 8.41254115e-01 -5.57956398e-01
7.43476808e-01 -1.98689654e-01 -1.84719312e+00 1.03829360e+00
1.35660636e+00 9.26915705e-01 1.25607058e-01 4.92547154e-01
-2.98692912e-01 1.24915719e+00 -5.05480170e-01 -5.80293953e-01
3.59851904e-02 -3.89818043e-01 -1.38178027e+00 -4.01302099e-01
3.97213042e-01 -5.30395508e-02 1.96465015e-01 7.40489304e-01
8.19429040e-01 1.51041150e-01 5.45618951e-01 -8.92470717e-01
-3.22695524e-01 3.86899829e-01 -3.34089190e-01 7.75235668e-02
7.53885984e-01 -3.31260175e-01 1.58402348e+00 -1.09361076e+00
4.22713190e-01 1.39919448e+00 1.83032170e-01 3.74291867e-01
-8.74546826e-01 -2.52996325e-01 8.00416153e-03 4.87314314e-01
-1.11968839e+00 -3.30761939e-01 6.46670043e-01 -4.01702315e-01
5.10199428e-01 4.89820331e-01 -1.31534278e-01 3.21706206e-01
5.08031473e-02 9.72496629e-01 1.70230377e+00 -4.92947519e-01
1.26969278e-01 1.42122135e-01 7.03645945e-01 6.02504432e-01
-3.23839128e-01 -1.14097401e-01 -8.14637959e-01 -6.80961668e-01
-1.02443099e-01 -3.17611158e-01 -2.56192505e-01 -2.11979121e-01
-1.22727895e+00 9.05177414e-01 2.22504035e-01 -4.28644083e-02
-2.67603725e-01 -1.08366050e-01 3.79970402e-01 8.84692609e-01
5.33771038e-01 6.07664049e-01 -7.71274626e-01 1.36108503e-01
-5.44205546e-01 5.32900989e-01 7.36130476e-01 6.96020365e-01
9.14257169e-01 -5.97176671e-01 -9.67445433e-01 1.14457190e+00
3.94626021e-01 3.11757296e-01 3.89928192e-01 -1.22590828e+00
9.10224557e-01 7.48940706e-01 4.74863023e-01 -8.06485355e-01
-2.51407623e-01 -3.96331549e-01 -3.56294364e-01 -1.42856121e-01
3.41874391e-01 7.72116259e-02 -7.14927793e-01 1.43741632e+00
4.38092768e-01 -3.11518669e-01 2.92693805e-02 8.80499005e-01
1.38885868e+00 7.84704268e-01 -5.47941998e-02 -2.32604727e-01
1.50830412e+00 -1.20252192e+00 -6.98884904e-01 6.29935414e-02
7.44159222e-01 -8.33529174e-01 1.23059058e+00 3.32461268e-01
-1.19757938e+00 -7.58215904e-01 -6.35277748e-01 -1.48332179e-01
-2.40085110e-01 4.02978927e-01 3.03735942e-01 4.70253497e-01
-1.16813719e+00 2.93840528e-01 1.70440972e-01 7.26630241e-02
-2.65894175e-01 5.42442203e-01 -4.82166559e-02 -3.69184613e-02
-1.72557116e+00 9.66038942e-01 2.02246055e-01 4.59346250e-02
-7.31009483e-01 -2.39389941e-01 -3.71886551e-01 -1.10165318e-02
4.73164916e-01 -7.28135884e-01 1.36815226e+00 -4.35625046e-01
-1.54751289e+00 1.02343404e+00 -4.69816297e-01 5.84344789e-02
1.73236206e-01 -3.42088580e-01 -2.21757546e-01 1.32008851e-01
4.14381146e-01 5.86348176e-01 5.71638644e-01 -1.55339968e+00
-1.21681678e+00 -3.04702610e-01 3.70036572e-01 6.40358746e-01
-1.34406492e-01 4.04709786e-01 -6.03721559e-01 -4.60272729e-02
3.53722602e-01 -5.50595403e-01 -1.44747451e-01 -8.40861440e-01
-1.59435213e-01 -1.01474595e+00 3.53639454e-01 -5.76903820e-01
1.67311549e+00 -1.67967296e+00 3.16967398e-01 3.86942238e-01
3.28660637e-01 1.92681968e-01 -3.71949822e-01 5.41232228e-01
1.77155569e-01 3.20973806e-04 1.44182965e-01 -2.08394572e-01
-4.13448624e-02 -1.08370446e-01 -3.24383616e-01 7.22115785e-02
3.60330604e-02 8.93783152e-01 -1.19823790e+00 -9.02221978e-01
-2.97686785e-01 -2.56434560e-01 -4.75562483e-01 4.41216111e-01
-3.33378583e-01 4.58751351e-01 -8.31842422e-01 8.10253620e-01
4.40101832e-01 -2.88162351e-01 2.60946035e-01 -1.91073313e-01
-1.40046358e-01 9.04931605e-01 -1.10534108e+00 1.31249881e+00
-2.62533396e-01 -1.86686307e-01 -1.63700804e-01 -6.68040931e-01
1.09792113e+00 3.76853585e-01 3.17000657e-01 -1.03902674e+00
-9.92638245e-02 8.48512948e-01 4.19536140e-03 -5.54061294e-01
8.36898804e-01 3.89590673e-03 -4.97165710e-01 5.16306221e-01
1.56658739e-01 -6.61077797e-02 3.72478664e-01 3.00365061e-01
1.04868639e+00 7.37312585e-02 2.78721571e-01 -3.70448530e-01
1.04887557e+00 7.69795524e-03 4.89710808e-01 9.89926219e-01
-1.97489023e-01 5.53921998e-01 5.62475681e-01 -1.78022206e-01
-3.64272922e-01 -1.01427829e+00 2.98626889e-02 1.61916101e+00
2.23984763e-01 -3.60327214e-01 -2.82753646e-01 -1.14803612e+00
-1.75036207e-01 4.37990427e-01 -4.77655083e-01 8.71890858e-02
-8.31360698e-01 -5.10098994e-01 1.47714525e-01 6.62635043e-02
4.47814196e-01 -1.19643652e+00 -1.27517581e-01 9.46292132e-02
-8.95826936e-01 -4.90254909e-01 -5.51482618e-01 3.13672960e-01
-7.73847282e-01 -1.06298232e+00 -8.76690865e-01 -9.23927605e-01
4.27500546e-01 5.58419228e-01 1.64029741e+00 2.84068853e-01
8.02213371e-01 4.71807241e-01 -7.27898121e-01 5.15239574e-02
-6.36683926e-02 3.67765695e-01 -2.91132033e-01 1.31420597e-01
2.35898331e-01 -4.78928238e-02 -5.54118216e-01 6.61096036e-01
-6.03187621e-01 -2.68625170e-01 5.74555039e-01 1.01475656e+00
7.03087091e-01 -1.88551396e-01 1.21945989e+00 -1.11355019e+00
1.07683301e+00 -4.83085245e-01 -3.03701431e-01 8.63393784e-01
-7.47178972e-01 2.12776735e-01 3.63725275e-01 -2.89419834e-02
-1.39024425e+00 -1.66001081e-01 -2.35026106e-01 2.82409549e-01
3.81911576e-01 8.33037317e-01 -1.74639583e-01 -7.71951154e-02
4.22786325e-01 1.06181815e-01 -4.68703389e-01 -4.65192378e-01
5.18409967e-01 8.80581439e-01 2.70518124e-01 -8.43661010e-01
7.03761518e-01 1.38091236e-01 -3.07688385e-01 2.12989338e-02
-1.44458127e+00 -1.09695053e+00 -4.70377296e-01 -5.87345183e-01
6.79167151e-01 -8.37043405e-01 -6.60588443e-01 2.48742327e-01
-1.30860972e+00 6.05136566e-02 -4.30187136e-02 2.25045711e-01
-4.32943463e-01 3.29183072e-01 -6.06513202e-01 -8.96524608e-01
-5.34394085e-01 -1.23090255e+00 1.17709708e+00 3.16597909e-01
-8.52317736e-02 -8.41937780e-01 3.67536098e-01 9.21373129e-01
2.41104200e-01 -1.57479480e-01 1.16925430e+00 -8.88205349e-01
-8.04771364e-01 -9.92824361e-02 -1.53875008e-01 1.68190718e-01
-5.77265024e-02 -3.99130791e-01 -8.92800808e-01 -2.08216414e-01
1.54434532e-01 -9.55524743e-01 1.24145389e+00 8.13110769e-02
9.16279316e-01 -8.06060061e-02 4.53445502e-02 -2.54668266e-01
1.21854699e+00 2.20342189e-01 7.88434744e-01 4.97387171e-01
3.94872516e-01 1.22702181e+00 1.31968558e+00 8.46189857e-02
8.55096579e-01 9.39653337e-01 2.84579277e-01 3.86991762e-02
1.15616381e-01 -4.62167338e-02 4.68758702e-01 1.48887658e+00
-9.76232365e-02 -2.93888062e-01 -8.04590642e-01 5.05900502e-01
-2.17529202e+00 -8.84159327e-01 -6.46024942e-01 2.49746966e+00
9.58642006e-01 1.38821781e-01 1.52141556e-01 6.55497685e-02
6.72049701e-01 2.20167503e-01 -5.23626478e-03 -2.97211468e-01
-1.81803554e-01 2.45381147e-01 2.69351266e-02 6.19259655e-01
-1.09221315e+00 6.78289175e-01 6.09573603e+00 1.27705252e+00
-3.15084785e-01 3.54486346e-01 7.07960606e-01 3.70348454e-01
-7.13080168e-01 2.55411267e-01 -1.03158200e+00 3.92359763e-01
8.17257762e-01 -4.67670932e-02 -6.31965231e-05 4.46990043e-01
-5.33966161e-02 -2.12378561e-01 -8.11915159e-01 5.40386379e-01
8.81159306e-02 -1.15646052e+00 1.49805322e-01 -2.19420061e-01
8.81179452e-01 -1.34165555e-01 1.04830667e-01 8.87058854e-01
4.98853594e-01 -6.63379192e-01 5.67638993e-01 6.97961390e-01
5.96416056e-01 -7.00569212e-01 1.06871223e+00 3.81178588e-01
-1.35344458e+00 -1.58500329e-01 -2.44990304e-01 1.14970163e-01
2.06006840e-01 4.43243712e-01 -1.38170093e-01 1.12663054e+00
6.96747422e-01 3.07485074e-01 -8.15130651e-01 1.15777051e+00
-3.74839127e-01 7.09782243e-01 1.95702925e-01 -2.30460957e-01
2.54429728e-01 -1.76700026e-01 2.73652375e-01 6.55461013e-01
1.59240797e-01 1.68297485e-01 3.75640512e-01 1.87185258e-01
-1.69187203e-01 3.01392019e-01 5.98112540e-03 6.84770823e-01
6.11889064e-01 1.15390587e+00 -3.68753552e-01 -3.46020997e-01
-3.97245020e-01 5.78615546e-01 5.47384799e-01 2.39285365e-01
-6.79367840e-01 -4.40290779e-01 -2.02892765e-01 -2.04351068e-01
-6.84228092e-02 3.33385676e-01 -2.34575644e-01 -8.83879185e-01
2.39099383e-01 -1.30495226e+00 9.21145439e-01 -5.63608885e-01
-1.37939405e+00 5.49744964e-01 6.55445904e-02 -1.64301109e+00
-8.86378586e-02 -3.57277483e-01 -6.60812974e-01 1.04558814e+00
-2.06577086e+00 -8.99864852e-01 4.86898981e-03 4.47183847e-01
4.83980834e-01 -2.34179169e-01 6.16156757e-01 7.24181414e-01
-6.82732537e-02 4.67677176e-01 1.74506649e-01 -2.97242165e-01
1.05038536e+00 -1.47541296e+00 -3.28566372e-01 3.81230235e-01
-5.83501011e-02 5.99333704e-01 2.22738504e-01 -5.04167318e-01
-1.10446084e+00 -7.91644216e-01 1.76241720e+00 -7.56261826e-01
4.39506680e-01 4.28424589e-02 -9.23936248e-01 1.03068233e-01
1.30458638e-01 -6.54358268e-01 8.20770204e-01 5.91206133e-01
-3.29083830e-01 -4.67410088e-01 -8.19970369e-01 3.47711056e-01
4.88052934e-01 -6.51188076e-01 -9.60424840e-01 5.80318689e-01
9.30283904e-01 -2.97963351e-01 -7.93116033e-01 7.36517370e-01
4.92117405e-01 -1.12452102e+00 1.14946544e+00 -4.41082984e-01
5.28922677e-01 -6.72197342e-01 -2.60849118e-01 -1.06521118e+00
-4.39216316e-01 -4.44601327e-01 -1.41518131e-01 1.50706863e+00
7.45318234e-01 -5.84750175e-01 5.56331694e-01 2.59937286e-01
-2.57133454e-01 -1.18994093e+00 -1.04309392e+00 -5.97952247e-01
2.20492855e-01 6.61095753e-02 5.33219814e-01 6.43082678e-01
-1.75360978e-01 8.13110650e-01 -6.20264232e-01 -3.61793451e-02
3.71619672e-01 3.52736861e-01 6.92571104e-01 -1.42765272e+00
-3.16597670e-01 -4.71393347e-01 3.86091381e-01 -1.45449030e+00
-8.29110369e-02 -8.11993957e-01 2.65442163e-01 -1.81241930e+00
4.26078856e-01 -6.76597238e-01 -8.03667307e-01 1.04096718e-01
-8.19288433e-01 -5.54442145e-02 1.18619792e-01 6.56688273e-01
-1.59220099e+00 8.00481498e-01 1.30571043e+00 -3.47419195e-02
-1.57821551e-01 4.80433315e-01 -9.77309465e-01 4.57716465e-01
7.16906130e-01 -7.28632152e-01 -4.58904088e-01 -5.30660808e-01
6.32150412e-01 3.80996466e-01 -5.10620475e-02 -5.36579370e-01
4.68725622e-01 -1.60076112e-01 -2.98436224e-01 -1.18918955e+00
3.01741779e-01 -3.41923058e-01 -5.99030375e-01 3.49861145e-01
-7.58455336e-01 1.58606663e-01 -5.54619193e-01 5.50824106e-01
-7.25077987e-01 -4.20980364e-01 5.29761910e-01 -8.82277861e-02
-7.02211916e-01 1.13865249e-02 -5.24321664e-03 5.10974109e-01
4.80179965e-01 9.29356217e-02 -7.21046090e-01 -6.03777468e-01
-3.27405632e-01 1.04392862e+00 -3.12790126e-01 3.45476121e-01
8.32658708e-01 -1.56317914e+00 -1.14150953e+00 -4.69872385e-01
5.52519083e-01 -3.92068535e-01 2.57742763e-01 9.64331627e-01
3.81132998e-02 8.28578353e-01 3.77044886e-01 -5.25867403e-01
-1.37390280e+00 4.58008610e-02 2.14867935e-01 -1.36504412e+00
2.80081928e-01 9.32182074e-01 1.10203050e-01 -1.31960702e+00
2.45927885e-01 2.09528178e-01 -1.04689169e+00 3.55269551e-01
2.87720561e-01 5.71827114e-01 1.98428109e-01 -4.52272147e-01
-2.85634875e-01 6.87466860e-01 -1.21415272e-01 -1.68266624e-01
7.94336736e-01 -2.74831593e-01 -6.36395693e-01 5.38635552e-01
9.26423669e-01 1.46792740e-01 -4.94727373e-01 -6.65224314e-01
6.49108648e-01 -3.40569288e-01 -2.57904738e-01 -1.04193068e+00
-5.47278225e-01 5.97249568e-01 4.15048987e-01 3.88673902e-01
1.22859251e+00 2.15760529e-01 6.41468287e-01 5.87805688e-01
6.52945817e-01 -1.24481595e+00 5.82327306e-01 9.23189104e-01
8.52048337e-01 -1.28574324e+00 -5.03505804e-02 -5.19555271e-01
-7.75667429e-01 7.42199600e-01 8.66936207e-01 2.62463301e-01
4.29470897e-01 -6.05127394e-01 2.12752104e-01 -4.81622964e-01
-1.20022142e+00 -7.15183556e-01 8.70981038e-01 1.59066617e-01
7.09505379e-01 -2.27198973e-02 -9.75039124e-01 5.62015355e-01
1.50540054e-01 -3.62943739e-01 1.25374854e-01 8.63669276e-01
-9.27932024e-01 -1.44722068e+00 -4.38711226e-01 8.02549005e-01
-4.76979017e-01 -2.09540293e-01 -3.85688424e-01 4.24633175e-01
1.05977230e-01 1.58216298e+00 -5.19386888e-01 -6.58144534e-01
4.50419486e-01 1.52875811e-01 1.61248505e-01 -7.74897993e-01
-1.07516706e+00 4.31264006e-02 5.68528533e-01 -2.62610406e-01
-8.86737704e-01 -5.08875310e-01 -8.43364179e-01 1.23059183e-01
-6.40079558e-01 7.82386839e-01 5.36819547e-02 1.10802293e+00
2.75758326e-01 3.84839624e-01 1.13154948e+00 -1.84755817e-01
-9.54567611e-01 -8.53208899e-01 -4.04587090e-01 2.72253484e-01
1.32131577e-01 -6.01738632e-01 -4.11145359e-01 -5.04488707e-01] | [11.280601501464844, 7.985755920410156] |
359f047c-adcd-4fbe-94ef-bae62d6490f1 | 1st-place-solution-for-the-uvo-challenge-on-1 | 2110.11661 | null | https://arxiv.org/abs/2110.11661v2 | https://arxiv.org/pdf/2110.11661v2.pdf | UVO Challenge on Video-based Open-World Segmentation 2021: 1st Place Solution | In this report, we introduce our (pretty straightforard) two-step "detect-then-match" video instance segmentation method. The first step performs instance segmentation for each frame to get a large number of instance mask proposals. The second step is to do inter-frame instance mask matching with the help of optical flow. We demonstrate that with high quality mask proposals, a simple matching mechanism is good enough for tracking. Our approach achieves the first place in the UVO 2021 Video-based Open-World Segmentation Challenge. | ['Vincent Lepetit', 'Yang Xiao', 'Wen Guo', 'Yuming Du'] | 2021-10-22 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 1.53279826e-01 1.01447217e-02 -4.19421196e-01 -3.80011737e-01
-8.91808391e-01 -7.58689940e-01 4.00287569e-01 -7.44245574e-02
-5.13934791e-01 5.24695635e-01 -9.95729789e-02 -2.03252614e-01
3.54812354e-01 -4.87680405e-01 -6.26105011e-01 -1.88100651e-01
-9.47614536e-02 4.09734875e-01 1.03475606e+00 -1.27392530e-01
3.73291790e-01 4.89994228e-01 -1.28168356e+00 3.49167407e-01
5.14744222e-01 9.38787460e-01 1.41327441e-01 1.04362178e+00
-2.93124825e-01 7.71083474e-01 -4.20843422e-01 -4.24142987e-01
7.95870185e-01 -2.89282411e-01 -1.29716098e+00 4.56978321e-01
1.02987742e+00 -4.92764264e-01 -3.44097078e-01 9.58708882e-01
2.29045898e-01 4.28236365e-01 1.71743512e-01 -9.79089737e-01
2.28374317e-01 5.08088887e-01 -6.84765458e-01 6.61916375e-01
5.35547078e-01 5.79732835e-01 9.04335916e-01 -6.45547807e-01
1.25471234e+00 1.05544615e+00 6.28111482e-01 6.80839598e-01
-1.07694435e+00 -3.23408991e-01 2.40161896e-01 -2.17092857e-01
-1.14732897e+00 -5.99838614e-01 3.42338324e-01 -5.81711948e-01
7.50084281e-01 2.72747099e-01 8.45416725e-01 3.11154395e-01
-1.83318853e-01 9.28364336e-01 8.29989731e-01 -1.16994523e-01
8.68492294e-03 -3.06236833e-01 2.63825536e-01 1.04742146e+00
-5.04060723e-02 1.66236848e-01 -1.50191441e-01 -8.41373298e-03
9.16839600e-01 -3.16321492e-01 -3.82266521e-01 -1.90366030e-01
-1.38585472e+00 3.98653090e-01 5.48880219e-01 2.79687911e-01
-1.17274530e-01 5.43686450e-01 3.80143166e-01 1.81933939e-01
4.60423976e-01 3.13711524e-01 -3.60728353e-01 -2.15184420e-01
-1.61089838e+00 3.25111240e-01 8.00272465e-01 9.33976054e-01
9.21639144e-01 -5.28906211e-02 -3.22140485e-01 3.82694632e-01
2.93624908e-01 1.11468531e-01 1.74039438e-01 -1.68004978e+00
4.88845229e-01 1.73998073e-01 2.85805017e-01 -7.98200250e-01
-1.86073318e-01 -1.61711928e-02 -1.14627540e-01 4.52753097e-01
9.12283301e-01 -2.85996705e-01 -1.19238663e+00 1.27656305e+00
7.48497784e-01 7.19464898e-01 -2.83984482e-01 9.83003199e-01
8.37823451e-01 5.88745415e-01 2.28476003e-02 -1.97981000e-01
1.23926973e+00 -1.37487233e+00 -3.90895694e-01 -1.75175339e-01
4.65766132e-01 -9.41306591e-01 5.66255808e-01 2.17251807e-01
-1.41174901e+00 -6.88057721e-01 -9.65913892e-01 -2.35087797e-02
-1.13666251e-01 -1.85464248e-01 5.53673685e-01 6.50645912e-01
-1.07380176e+00 7.70915866e-01 -8.42884362e-01 -2.60905802e-01
5.95187604e-01 3.96458954e-01 -4.69013810e-01 -1.84145242e-01
-6.21563196e-01 4.77742970e-01 5.25407970e-01 -1.23573460e-01
-8.86359751e-01 -7.52996027e-01 -7.30418086e-01 -1.56159773e-01
6.49253368e-01 -7.05899835e-01 1.38716817e+00 -1.21255887e+00
-1.29130137e+00 1.02089155e+00 -4.41257089e-01 -6.39157772e-01
9.41003144e-01 -2.36216828e-01 -1.06647797e-03 5.81966937e-01
1.61785036e-01 1.22143102e+00 1.04376709e+00 -1.34409988e+00
-1.00635302e+00 -9.59757529e-03 3.81777912e-01 3.82153913e-02
3.73855829e-01 2.94358760e-01 -1.12869394e+00 -4.33456779e-01
1.30534947e-01 -8.11755538e-01 -5.21579266e-01 2.18247652e-01
-4.76725399e-01 -2.10553203e-02 8.81247580e-01 -6.26731217e-01
1.23271191e+00 -2.12849379e+00 -1.18435927e-01 2.36876950e-01
5.38972437e-01 4.53776449e-01 -1.32894874e-01 -1.90798417e-01
7.20243528e-02 2.87869629e-02 -1.93260714e-01 -6.06587350e-01
-2.07812876e-01 4.37791273e-02 2.23967452e-02 5.04687548e-01
9.06941369e-02 9.42469597e-01 -1.00873983e+00 -8.77550483e-01
5.33282280e-01 1.83466673e-01 -6.05488658e-01 9.27413329e-02
-4.84249949e-01 4.68223512e-01 -2.57417440e-01 6.94290459e-01
5.50431609e-01 -1.41736731e-01 -3.90377492e-02 -3.02865863e-01
-3.59053701e-01 -1.23679563e-02 -1.50761795e+00 1.86851180e+00
2.46109083e-01 9.83746290e-01 3.64081979e-01 -5.89517295e-01
3.78277659e-01 1.99473292e-01 1.04938447e+00 -3.11603367e-01
1.99204579e-01 1.65157244e-01 -3.54753546e-02 -3.18149716e-01
5.65727413e-01 3.31970036e-01 2.14503437e-01 2.08512962e-01
4.25271392e-02 -2.19593033e-01 7.70408750e-01 4.78308260e-01
1.16366899e+00 4.73732978e-01 1.45094946e-01 -2.99381942e-01
6.15787029e-01 4.35492247e-01 8.41395795e-01 6.81466758e-01
-7.53438950e-01 1.10909045e+00 3.72924596e-01 -7.66381443e-01
-9.89367604e-01 -7.81455934e-01 7.98043795e-03 5.86298585e-01
2.56506473e-01 -4.52800959e-01 -1.19969690e+00 -7.95342267e-01
-1.82237759e-01 -5.77829666e-02 -3.95613819e-01 6.38545096e-01
-8.37187588e-01 -3.00952464e-01 3.85701299e-01 4.55595791e-01
5.64212501e-01 -8.80113065e-01 -8.13153207e-01 3.87628406e-01
-2.06527606e-01 -1.44214189e+00 -9.45170522e-01 -1.00564711e-01
-7.59468853e-01 -1.43026674e+00 -6.73707604e-01 -7.19588935e-01
5.88549018e-01 4.59317267e-01 1.21538365e+00 5.96468151e-01
-5.03035665e-01 4.38900799e-01 -2.01821566e-01 1.46735251e-01
-3.10020804e-01 -9.62176025e-02 -3.49115252e-01 7.58226290e-02
6.76193684e-02 -1.14181936e-02 -8.12610567e-01 3.63429844e-01
-7.76173294e-01 -8.04354474e-02 2.46686060e-02 2.65338838e-01
8.51113260e-01 -3.54651004e-01 -8.02697465e-02 -7.62486398e-01
1.39983639e-01 -1.16822645e-02 -9.97485995e-01 2.78180033e-01
-1.66375473e-01 -2.18727499e-01 1.85308084e-01 -1.89304471e-01
-6.93800628e-01 6.25439048e-01 -2.85338700e-01 -6.17526114e-01
-1.88201770e-01 -3.61108184e-02 1.38539791e-01 -4.67065156e-01
3.29483658e-01 -1.49034396e-01 1.13535561e-01 -2.42573723e-01
4.86584872e-01 1.40710458e-01 9.15203631e-01 -6.43603206e-01
6.98306918e-01 7.79057086e-01 -2.05351412e-01 -6.32766068e-01
-8.00782979e-01 -8.45546782e-01 -8.47216725e-01 -4.87081051e-01
1.30322254e+00 -8.02333474e-01 -5.75324476e-01 3.33526850e-01
-1.23626804e+00 -8.69049370e-01 -5.30500412e-01 3.73263508e-01
-7.01742709e-01 5.20858645e-01 -7.88272798e-01 -4.56878930e-01
-2.55714983e-01 -1.50340283e+00 1.05362046e+00 4.06186104e-01
-3.99069116e-02 -8.69357169e-01 2.64255941e-01 4.95307297e-01
1.29317552e-01 1.45618573e-01 -3.51597294e-02 -4.64335740e-01
-1.12436044e+00 -2.90324707e-02 -3.84951621e-01 1.87876970e-01
-1.14323245e-03 5.58528066e-01 -7.48174250e-01 -1.70913875e-01
-3.39962214e-01 9.39563289e-02 9.74285007e-01 7.24465787e-01
1.04654551e+00 2.75467396e-01 -3.35460037e-01 8.80627334e-01
1.46026611e+00 7.65843317e-02 8.68759394e-01 2.82229394e-01
8.77737343e-01 2.24180162e-01 7.94413269e-01 1.37507349e-01
4.02388066e-01 7.13822365e-01 2.86496073e-01 -3.65740329e-01
-4.90225822e-01 6.67925999e-02 1.61902636e-01 2.72428602e-01
-2.37614721e-01 -5.01138996e-03 -7.85508513e-01 5.76895297e-01
-1.91030550e+00 -1.23960280e+00 -2.96396732e-01 2.03926587e+00
4.58660573e-01 3.21297646e-01 5.77081203e-01 -1.77153736e-01
8.38918090e-01 2.32904315e-01 -1.09634958e-01 -1.33145303e-01
1.45542383e-01 2.43890852e-01 6.12187684e-01 9.29426551e-01
-1.37758613e+00 1.27213967e+00 7.12151432e+00 6.46857798e-01
-8.79521370e-01 8.55839550e-02 7.32082129e-01 3.65444906e-02
6.77024722e-02 2.87082523e-01 -9.96100724e-01 6.06335342e-01
6.48055792e-01 2.45216683e-01 2.93706059e-01 5.43617606e-01
2.64558643e-01 -5.07290483e-01 -8.23065102e-01 8.23100448e-01
-1.89380407e-01 -1.88071561e+00 -2.88912237e-01 -9.17918906e-02
8.57264936e-01 2.70771921e-01 -4.37752813e-01 -1.09259836e-01
2.02569440e-01 -6.95612133e-01 7.04627097e-01 3.67975563e-01
4.32395726e-01 -4.87976551e-01 2.92004615e-01 -2.38470850e-03
-1.57601595e+00 1.97489679e-01 -2.60615379e-01 2.75736302e-01
6.55632794e-01 5.44258535e-01 -4.67746556e-01 5.18244147e-01
5.66262186e-01 6.22037947e-01 -4.10591274e-01 1.74726403e+00
1.62291765e-01 4.71804947e-01 -6.18823707e-01 5.52536488e-01
5.79347432e-01 -3.39258879e-01 5.90551257e-01 1.41692734e+00
-1.84188932e-02 1.81846485e-01 8.09607565e-01 5.36212265e-01
-1.44110650e-01 -6.09950051e-02 -3.31983060e-01 3.21650654e-01
3.30268502e-01 1.36083817e+00 -1.30400825e+00 -8.50471437e-01
-5.57393372e-01 1.03594708e+00 -1.23322140e-02 1.86005339e-01
-9.16767240e-01 -1.50352819e-02 6.62520409e-01 2.48576775e-01
4.58589613e-01 -4.88926589e-01 -1.46190137e-01 -1.23920965e+00
-2.03762755e-01 -6.89477503e-01 4.43578988e-01 -5.74874401e-01
-7.59862244e-01 4.61325377e-01 -6.17806688e-02 -1.12970781e+00
-1.28535628e-01 -4.46346641e-01 -7.97083855e-01 4.86554831e-01
-1.56736803e+00 -9.60944355e-01 -2.87087411e-01 8.17813158e-01
1.04797220e+00 2.47838989e-01 1.74176842e-01 6.48996830e-01
-6.00304127e-01 1.98518872e-01 -2.71648526e-01 5.97822011e-01
4.98458654e-01 -1.28162456e+00 6.71555936e-01 1.11085343e+00
4.86894071e-01 4.29040849e-01 6.50070429e-01 -7.59283006e-01
-9.72707748e-01 -1.00508618e+00 6.60459340e-01 -5.65877140e-01
5.42011321e-01 -6.37155324e-02 -5.86176574e-01 7.89883792e-01
2.44451746e-01 7.40295768e-01 2.04420283e-01 -4.41694558e-01
-1.02660500e-01 3.95234227e-02 -1.27980554e+00 3.43259186e-01
9.33291256e-01 -8.72949958e-02 -5.38095355e-01 4.66678739e-01
7.11434186e-01 -9.38547909e-01 -8.45814586e-01 2.03036577e-01
5.52550375e-01 -9.95864332e-01 1.17569661e+00 -5.52630186e-01
1.64741859e-01 -6.86052263e-01 -1.56531945e-01 -6.47433519e-01
2.97795958e-03 -1.30027258e+00 -1.77319467e-01 1.05805779e+00
3.73424083e-01 -1.67162225e-01 1.11672020e+00 8.19737077e-01
-1.32016391e-01 -5.56534290e-01 -8.60794067e-01 -7.55849719e-01
-2.52847254e-01 -6.04061782e-01 4.06614244e-01 7.19942033e-01
-3.19676369e-01 -2.63294309e-01 -1.39835283e-01 -1.10149868e-02
8.21885407e-01 9.25135016e-02 1.04641557e+00 -9.23488259e-01
-5.08369029e-01 -4.41655636e-01 -5.53849936e-01 -1.33473384e+00
-2.04390317e-01 -5.52968681e-01 8.00091252e-02 -1.41278219e+00
5.87972403e-02 -5.13651371e-01 -1.27094626e-01 2.12894917e-01
-3.54953200e-01 6.11921728e-01 7.52227068e-01 2.03354180e-01
-1.09280157e+00 -3.42242032e-01 1.20055246e+00 -8.59433636e-02
-2.49080718e-01 9.18848068e-02 -1.77621141e-01 6.71125710e-01
5.27165890e-01 -4.15529698e-01 -1.71052903e-01 -4.07951236e-01
-3.59621525e-01 1.84845701e-01 5.01400828e-01 -1.29660428e+00
1.43351004e-01 -2.38705620e-01 1.74427107e-01 -6.21540010e-01
3.05365831e-01 -6.47953331e-01 1.34176239e-01 5.90691566e-01
-2.39831805e-02 2.88756043e-01 1.09350763e-01 4.13087547e-01
-1.75393626e-01 -3.88304651e-01 9.23682272e-01 -5.19411206e-01
-1.22761405e+00 5.97839892e-01 -2.29192093e-01 3.41539949e-01
1.22900116e+00 -4.75885093e-01 -2.21532747e-01 -2.90855654e-02
-1.05847752e+00 4.92778748e-01 5.86478472e-01 1.67937666e-01
3.16001117e-01 -8.97814691e-01 -5.44335842e-01 -2.90931351e-02
-3.58964354e-01 1.06787071e-01 -5.01550874e-03 9.55052078e-01
-9.89843130e-01 1.10948011e-01 -7.08809942e-02 -8.87067854e-01
-1.41001666e+00 4.95838016e-01 6.07792437e-01 -2.40011990e-01
-9.28757966e-01 1.00596690e+00 4.95488942e-03 -1.84120871e-02
1.50398523e-01 -2.92998612e-01 3.21180314e-01 -5.83047457e-02
6.76343739e-01 3.85296941e-01 -1.17043957e-01 -7.67456591e-01
-3.56400907e-01 7.41911471e-01 4.67235222e-02 -3.80206883e-01
9.53628957e-01 -2.57629156e-01 1.30032718e-01 6.06361479e-02
1.01294613e+00 1.47928670e-03 -1.71929097e+00 6.36786297e-02
6.45018816e-02 -8.88310611e-01 6.29926147e-03 -4.10383999e-01
-1.44506824e+00 6.67348444e-01 4.96542215e-01 2.20361814e-01
8.74439001e-01 -5.97388633e-02 1.09709644e+00 2.33623069e-02
5.12924075e-01 -1.24335027e+00 -1.82019100e-01 3.09713900e-01
6.67927712e-02 -1.21153104e+00 5.25195636e-02 -7.41964698e-01
-3.83165717e-01 1.12715507e+00 7.23138750e-01 -4.26785320e-01
4.39861327e-01 4.21003640e-01 1.81650132e-01 -3.09027493e-01
-3.68153691e-01 -5.89369178e-01 4.37641174e-01 5.82500994e-01
3.29964042e-01 -2.52133429e-01 -3.54582757e-01 -2.78724492e-01
2.54781783e-01 9.52133089e-02 5.36518157e-01 9.38081205e-01
-8.01461577e-01 -1.15992594e+00 -4.01517034e-01 3.27988446e-01
-6.16507828e-01 1.27446070e-01 -2.47972056e-01 9.08264458e-01
3.58362198e-02 8.18374515e-01 1.86504578e-04 -2.11533103e-02
1.71820164e-01 -7.43683279e-02 7.53197432e-01 -5.80840349e-01
-8.71435106e-01 2.88444668e-01 1.64138228e-01 -1.21418774e+00
-7.87039816e-01 -7.13875473e-01 -1.52246034e+00 -3.53481770e-01
-2.14673713e-01 -2.02100612e-02 5.17386556e-01 1.08238971e+00
2.62110114e-01 3.11789542e-01 3.48760754e-01 -1.16691685e+00
5.48985340e-02 -2.94114083e-01 -1.48059562e-01 5.54249108e-01
4.26194191e-01 -3.57122153e-01 -7.55765438e-02 3.67372632e-01] | [9.134201049804688, -0.20467641949653625] |
339e8f0c-bb0e-43d5-9a65-690db2a6df39 | open-relation-and-event-type-discovery-with | 2212.00178 | null | https://arxiv.org/abs/2212.00178v1 | https://arxiv.org/pdf/2212.00178v1.pdf | Open Relation and Event Type Discovery with Type Abstraction | Conventional closed-world information extraction (IE) approaches rely on human ontologies to define the scope for extraction. As a result, such approaches fall short when applied to new domains. This calls for systems that can automatically infer new types from given corpora, a task which we refer to as type discovery. To tackle this problem, we introduce the idea of type abstraction, where the model is prompted to generalize and name the type. Then we use the similarity between inferred names to induce clusters. Observing that this abstraction-based representation is often complementary to the entity/trigger token representation, we set up these two representations as two views and design our model as a co-training framework. Our experiments on multiple relation extraction and event extraction datasets consistently show the advantage of our type abstraction approach. Code available at https://github.com/raspberryice/type-discovery-abs. | ['Jiawei Han', 'Heng Ji', 'Sha Li'] | 2022-11-30 | null | null | null | null | ['event-extraction', 'type'] | ['natural-language-processing', 'speech'] | [ 5.61596155e-02 4.95693892e-01 -4.70251948e-01 -4.74468708e-01
-4.77248311e-01 -8.23339045e-01 9.38277960e-01 4.91717517e-01
-2.74748504e-01 1.02338290e+00 2.58816928e-01 -4.19200957e-01
-2.64549395e-03 -1.13918948e+00 -7.87824571e-01 -3.46626967e-01
5.51775619e-02 4.75406379e-01 2.73074538e-01 1.77703835e-02
7.78570473e-02 3.30733627e-01 -1.78509688e+00 7.02809215e-01
7.22774506e-01 8.12048614e-01 -1.70663342e-01 2.53444463e-01
-5.23275733e-01 8.56209099e-01 -6.00783110e-01 -3.96065235e-01
8.60137939e-02 -2.60277897e-01 -1.39907610e+00 -1.10836282e-01
-1.52531639e-01 1.49014652e-01 5.08580692e-02 8.71159673e-01
2.13322282e-01 1.97478339e-01 6.28879964e-01 -1.43406320e+00
-4.78307366e-01 8.67831767e-01 -3.90076786e-01 1.17415592e-01
9.43144739e-01 -3.16783071e-01 1.33550084e+00 -8.93200696e-01
1.05984008e+00 1.04912925e+00 7.38708138e-01 7.63405144e-01
-1.53701830e+00 -5.35405338e-01 1.95072070e-01 3.18479955e-01
-1.48977542e+00 -4.71374303e-01 5.26911855e-01 -5.18934786e-01
1.05501866e+00 5.09058297e-01 3.49793166e-01 1.14448905e+00
-4.61328149e-01 8.19398820e-01 1.29789400e+00 -7.12321103e-01
4.49479252e-01 6.04213059e-01 3.47228020e-01 1.65030658e-01
8.15394640e-01 -8.37268680e-02 -4.12495553e-01 -3.72467011e-01
3.84753406e-01 -7.39177689e-02 -1.72960863e-01 -3.45404565e-01
-9.32616591e-01 6.07314050e-01 2.55868405e-01 5.15430331e-01
-4.42514062e-01 -7.83570670e-03 4.59908843e-01 2.45764568e-01
4.12840605e-01 7.33060062e-01 -9.22736168e-01 1.47930995e-01
-5.77389657e-01 5.39695799e-01 1.14329720e+00 1.26918745e+00
7.92178333e-01 -6.22240245e-01 1.50487930e-01 8.48933280e-01
2.75939971e-01 -2.51789272e-01 4.83705878e-01 -5.34475565e-01
1.57699555e-01 1.01115298e+00 2.36697167e-01 -4.26044106e-01
-4.80476618e-01 4.41489974e-03 -2.51036674e-01 2.35841125e-02
5.29838502e-01 -2.27347657e-01 -7.46492922e-01 1.70558870e+00
7.91517496e-01 3.25693846e-01 2.60394543e-01 3.74132365e-01
7.35874593e-01 2.34811768e-01 2.32545465e-01 -2.65028149e-01
1.77301967e+00 -2.65583426e-01 -7.61452496e-01 -5.20926639e-02
8.47192645e-01 -2.94899613e-01 7.31820524e-01 2.79285759e-01
-7.93537796e-01 -7.00576082e-02 -9.05970812e-01 4.75326553e-02
-1.16125560e+00 -3.89017940e-01 7.49174356e-01 5.69364250e-01
-4.57377642e-01 5.74884176e-01 -7.89676726e-01 -6.51726246e-01
3.23039263e-01 2.71295130e-01 -2.28341833e-01 1.99589044e-01
-1.38987851e+00 8.35936666e-01 7.57999539e-01 -2.29727477e-01
-2.70474017e-01 -7.42199302e-01 -6.21158183e-01 -5.08049056e-02
8.14075947e-01 -7.96789527e-01 1.45255995e+00 -6.67263746e-01
-8.58361006e-01 9.86473799e-01 -2.80013055e-01 -5.39425194e-01
-4.47588339e-02 -1.86013624e-01 -6.70276701e-01 -1.00085750e-01
3.03525329e-01 7.65924528e-02 3.33764136e-01 -1.62729096e+00
-1.14805615e+00 -3.51332277e-01 3.34137082e-01 -1.72082543e-01
-4.57426235e-02 3.53044152e-01 6.56459332e-02 -5.56261063e-01
1.44236842e-02 -7.90032566e-01 -1.16133705e-01 -6.37712657e-01
-5.99276066e-01 -7.27470100e-01 6.24411821e-01 -3.43819052e-01
1.53670037e+00 -1.89085650e+00 1.37412986e-02 3.15412730e-01
3.38711262e-01 -4.76880036e-02 4.82305735e-01 6.08429790e-01
-4.97418493e-01 5.89890480e-01 -1.81086287e-01 8.50345045e-02
2.27437481e-01 4.20615226e-01 -3.44705701e-01 -8.67574885e-02
4.16898668e-01 6.44022763e-01 -1.04436779e+00 -3.96984279e-01
-2.25169644e-01 4.73156720e-02 -6.02127016e-01 2.76521504e-01
-6.57233953e-01 1.77074075e-01 -7.49489367e-01 5.78368723e-01
4.17767942e-01 -1.69766217e-01 7.13311791e-01 -1.93173110e-01
-2.92067856e-01 7.63257086e-01 -1.46291900e+00 1.48198378e+00
-2.55056858e-01 3.36150974e-02 -3.97108942e-01 -1.18764782e+00
7.82756150e-01 7.31403470e-01 5.34706771e-01 -9.63118300e-02
-3.57040241e-02 4.42927539e-01 -1.55864164e-01 -7.99926698e-01
3.33795220e-01 -4.06985998e-01 -3.54256570e-01 3.54783028e-01
2.92144477e-01 2.36272991e-01 5.24440765e-01 1.58127427e-01
1.09427738e+00 4.88384634e-01 8.79219413e-01 -1.38521314e-01
4.30172950e-01 1.70344025e-01 9.48103070e-01 5.45721292e-01
2.84191579e-01 1.95668042e-01 7.49003470e-01 -5.88385463e-01
-7.14014053e-01 -9.46205437e-01 -4.00764674e-01 9.34627473e-01
-8.57160389e-02 -8.25577974e-01 -5.41382074e-01 -1.12383854e+00
1.41652167e-01 9.07786012e-01 -6.17488801e-01 8.24776217e-02
-5.44051647e-01 -6.91617310e-01 6.11612737e-01 5.80361784e-01
1.88401304e-02 -1.17533064e+00 -7.63671696e-01 3.13653439e-01
-4.19352740e-01 -1.06510973e+00 4.56246734e-01 4.79939282e-01
-5.78394592e-01 -1.23069632e+00 -8.42766315e-02 -5.06265104e-01
3.75732094e-01 -4.65520531e-01 1.30925691e+00 7.93130770e-02
-1.04045250e-01 2.76926100e-01 -8.33492875e-01 -6.78569078e-01
-3.52838516e-01 2.33640999e-01 1.61390483e-01 2.59576365e-02
9.93183553e-01 -7.69421637e-01 -2.30119795e-01 -8.92732963e-02
-1.00094271e+00 -1.51362181e-01 3.96128118e-01 7.44718611e-01
4.23623681e-01 1.31461337e-01 8.10962915e-01 -1.48057604e+00
4.68910187e-01 -8.31547916e-01 -4.88691390e-01 5.02939641e-01
-8.14691305e-01 4.13720995e-01 5.42186201e-01 -4.18393642e-01
-1.21671546e+00 1.70045197e-01 1.18693359e-01 -1.67673025e-02
-7.05168724e-01 9.35134232e-01 -4.74383354e-01 5.40162563e-01
7.64265895e-01 -1.26466632e-01 -5.55606067e-01 -7.71445572e-01
4.90556359e-01 7.82888234e-01 5.24202466e-01 -1.02213895e+00
7.03399897e-01 3.35889995e-01 -2.47328177e-01 -5.89010239e-01
-8.86055231e-01 -5.51836550e-01 -7.31497407e-01 1.87486455e-01
8.66931915e-01 -6.23546064e-01 -4.92701322e-01 -1.67555921e-02
-1.10291076e+00 -1.68364689e-01 -6.05204403e-01 3.48407626e-01
-5.77359200e-01 1.20824307e-01 -2.00435385e-01 -9.19767261e-01
9.63003486e-02 -5.46438813e-01 8.87397885e-01 1.78035125e-01
-6.36390805e-01 -9.76756930e-01 2.07773939e-01 -4.09992971e-02
-1.46356866e-01 6.10727966e-01 1.10977793e+00 -1.37474644e+00
-4.20689106e-01 -2.00206503e-01 -9.37848240e-02 -2.33922422e-01
3.22581232e-01 -2.48847708e-01 -1.18204021e+00 3.36643368e-01
-8.32239836e-02 9.71573144e-02 5.18337846e-01 -1.69190288e-01
1.03282440e+00 -7.08605886e-01 -7.08615661e-01 3.97074103e-01
1.45260632e+00 3.33496064e-01 6.86138272e-01 5.82495153e-01
5.49531996e-01 6.55486107e-01 5.72618127e-01 6.59451127e-01
5.88179410e-01 8.24123979e-01 -1.69574320e-01 1.01225406e-01
5.61776236e-02 -2.78799146e-01 -6.18175343e-02 2.21776888e-01
-2.30517030e-01 -1.71798900e-01 -1.10757184e+00 6.77992523e-01
-1.90077066e+00 -1.22706556e+00 -3.10326200e-02 2.16285801e+00
1.49031615e+00 -3.34776100e-03 4.71525252e-01 3.25330079e-01
5.65862179e-01 -4.04788524e-01 -2.12218747e-01 -4.78233039e-01
2.57450789e-01 3.48462701e-01 4.08484429e-01 2.54552215e-01
-9.81397092e-01 8.34979177e-01 5.42606306e+00 5.37649870e-01
-6.24046266e-01 9.84055772e-02 2.36836836e-01 2.73880750e-01
-4.64301348e-01 5.03441215e-01 -9.73734558e-01 4.05941993e-01
1.07403159e+00 -4.77228522e-01 4.39878732e-01 4.99540031e-01
-1.28439754e-01 -1.44899309e-01 -1.62914646e+00 5.30700684e-01
-3.00952792e-01 -1.23495126e+00 9.02834609e-02 1.82244614e-01
2.86066562e-01 -2.96375096e-01 -5.36461353e-01 3.05630893e-01
6.18726134e-01 -7.31619477e-01 6.78941488e-01 4.91335362e-01
7.11419046e-01 -4.09435332e-01 5.44775367e-01 2.18528911e-01
-1.43026400e+00 -1.76373869e-01 1.40337907e-02 -6.59480393e-02
3.64038013e-02 5.78168452e-01 -1.05007470e+00 1.02516699e+00
7.02500045e-01 6.02566361e-01 -5.58622479e-01 1.00153780e+00
-3.93769920e-01 4.87696856e-01 -3.89929235e-01 1.93215266e-01
-4.03411776e-01 6.14393763e-02 4.97474581e-01 1.20061016e+00
1.28647923e-01 3.15704376e-01 6.01727515e-02 1.12519324e+00
3.04509550e-02 -2.04799339e-01 -7.64364839e-01 -1.41642466e-01
8.49693835e-01 1.15076780e+00 -6.97355449e-01 -7.95047164e-01
-7.67663956e-01 5.24941564e-01 3.07445407e-01 3.67871583e-01
-6.62879467e-01 -6.12814069e-01 5.75495243e-01 4.41126116e-02
3.12026858e-01 1.70426533e-01 -3.37718517e-01 -1.46271312e+00
7.23803267e-02 -1.02183604e+00 8.70533109e-01 -5.88160336e-01
-1.25136328e+00 3.88343811e-01 5.60653865e-01 -1.16517508e+00
-5.67988038e-01 -5.19842684e-01 -3.00932199e-01 5.10053277e-01
-1.24733412e+00 -1.06580496e+00 -1.29255587e-02 3.95713419e-01
2.31478542e-01 4.16015536e-01 1.14106536e+00 3.51740718e-01
-6.26886964e-01 5.98024487e-01 -3.64408046e-01 3.50392133e-01
6.12751067e-01 -1.56095266e+00 1.91781297e-01 9.02087986e-01
2.73954898e-01 8.84593844e-01 9.01045799e-01 -7.18684256e-01
-1.13064027e+00 -9.93052781e-01 1.35443711e+00 -8.72226119e-01
9.04114187e-01 -4.96167690e-01 -1.19561517e+00 1.09095085e+00
2.19554916e-01 8.21711943e-02 9.13884223e-01 5.83135843e-01
-7.88660705e-01 9.18433294e-02 -1.21758103e+00 4.26789522e-01
1.12805188e+00 -4.19234186e-01 -1.17786670e+00 1.92582652e-01
6.86081946e-01 -5.74556738e-02 -1.33009267e+00 3.66749883e-01
5.42978466e-01 -6.68326616e-01 7.43774831e-01 -8.73519123e-01
2.32552737e-01 -4.59042639e-01 -2.03452006e-01 -1.14624929e+00
-1.94435582e-01 -5.58028162e-01 -3.59301567e-01 1.72064650e+00
8.51612508e-01 -9.70161378e-01 2.53770441e-01 9.57226157e-01
9.17266533e-02 -5.71234107e-01 -6.44232213e-01 -9.36019659e-01
2.07581408e-02 -2.68794388e-01 1.10045254e+00 1.43446648e+00
8.32423747e-01 4.31890547e-01 1.97182581e-01 3.89711708e-01
4.36846435e-01 3.16675991e-01 6.29101515e-01 -1.58184886e+00
-3.69298935e-01 -2.94869274e-01 -3.76860470e-01 -2.84218311e-01
1.51694506e-01 -9.97197449e-01 -1.25823438e-01 -1.42969394e+00
6.53246120e-02 -7.70783067e-01 -4.30978715e-01 9.17777777e-01
-1.82231441e-01 -8.19949731e-02 -1.73565403e-01 1.91767067e-01
-3.30140650e-01 -8.20970610e-02 3.68525058e-01 8.29811543e-02
-3.19147766e-01 -5.81384525e-02 -9.68536556e-01 7.53113806e-01
1.04106736e+00 -7.47833848e-01 -2.19313562e-01 -1.80469826e-02
7.25736856e-01 -1.07304476e-01 3.36183965e-01 -7.10045576e-01
5.37659861e-02 -3.80601019e-01 1.03182353e-01 -2.13374972e-01
-1.73576269e-02 -8.28369141e-01 3.59192789e-01 1.24186270e-01
-5.09199977e-01 -1.14092350e-01 1.16515607e-01 3.16178858e-01
-9.35617536e-02 -3.88752371e-01 2.83401072e-01 -2.86130518e-01
-8.81312728e-01 -1.05815604e-02 -4.12534803e-01 1.92954212e-01
9.19043362e-01 -1.80726111e-01 -2.75944114e-01 1.64535120e-01
-9.73344505e-01 -2.85885427e-02 4.85409051e-01 4.50486571e-01
2.53430933e-01 -1.23397720e+00 -4.53254223e-01 1.41691446e-01
5.12634635e-01 1.12856492e-01 -5.59629261e-01 9.29906845e-01
1.69307619e-01 1.61822677e-01 8.29927176e-02 -1.94288626e-01
-1.03061843e+00 9.56087828e-01 2.79919416e-01 -3.07928801e-01
-6.66279316e-01 6.75550461e-01 2.74463117e-01 -3.67766827e-01
-1.47190332e-01 -5.30196905e-01 -4.66430753e-01 3.42982680e-01
3.55662018e-01 8.24161321e-02 1.70826524e-01 -2.31768355e-01
-6.53897464e-01 1.28190713e-02 -1.09228209e-01 -1.74521849e-01
1.42240751e+00 -1.64183415e-02 -3.65090907e-01 8.52550507e-01
7.92492509e-01 1.22379728e-01 -5.37676454e-01 -2.98930019e-01
9.41112101e-01 -3.04082662e-01 -5.58693230e-01 -8.78719628e-01
-5.45108020e-01 2.39095628e-01 2.27299646e-01 8.30849588e-01
9.54759777e-01 3.89502257e-01 4.20809954e-01 3.23150754e-01
4.01358157e-01 -1.06785715e+00 -5.89162588e-01 3.81423354e-01
6.32953763e-01 -8.47813368e-01 4.30965647e-02 -7.09093869e-01
-4.26288664e-01 8.40189755e-01 5.44032037e-01 -2.57189255e-02
5.22478402e-01 5.89204252e-01 -1.09505288e-01 -3.75754774e-01
-1.07614625e+00 -5.82868278e-01 1.02474220e-01 6.87033534e-01
7.20865548e-01 2.62558997e-01 -5.33967614e-01 9.11647558e-01
-2.27384657e-01 3.15430015e-01 6.72530651e-01 1.23079050e+00
-7.34723955e-02 -1.49743092e+00 -2.49539912e-01 6.59968913e-01
-6.67558730e-01 -1.14667147e-01 -5.91158748e-01 8.13155711e-01
3.42570215e-01 1.06520569e+00 -1.22602381e-01 -3.72064650e-01
4.87458289e-01 7.25326598e-01 4.05532598e-01 -1.00624740e+00
-6.74260437e-01 -1.19454443e-01 6.56871021e-01 -5.09914517e-01
-6.74909115e-01 -7.93572187e-01 -1.37794960e+00 1.26344562e-01
-5.75640917e-01 4.33869332e-01 2.68972099e-01 1.06466043e+00
4.09796178e-01 4.37024623e-01 5.97759247e-01 -2.88151741e-01
-2.82894880e-01 -8.35859597e-01 -4.89905506e-01 7.36197472e-01
1.81085929e-01 -1.00104809e+00 -3.30703557e-01 3.86134863e-01] | [9.302626609802246, 8.717145919799805] |
0429a52e-c376-452e-9a7e-894e2025ab55 | variational-autoencoders-with-implicit-priors | null | null | https://openreview.net/forum?id=ryeHw1vjiQ | https://openreview.net/pdf?id=ryeHw1vjiQ | Variational Autoencoders with implicit priors for short-duration text-independent speaker verification | In this work, we exploited different strategies to provide prior knowledge to commonly used generative modeling approaches aiming to obtain speaker-dependent low dimensional representations from short-duration segments of speech data, making use of available information of speaker identities. Namely, convolutional variational autoencoders are employed, and statistics of its learned posterior distribution are used as low dimensional representations of fixed length short-duration utterances. In order to enforce speaker dependency in the latent layer, we introduced a variation of the commonly used prior within the variational autoencoders framework, i.e. the model is simultaneously trained for reconstruction of inputs along with a discriminative task performed on top of latent layers outputs. The effectiveness of both triplet loss minimization and speaker recognition are evaluated as implicit priors on the challenging cross-language NIST SRE 2016 setting and compared against fully supervised and unsupervised baselines. | ['Anonymous'] | 2018-10-22 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [ 9.63614136e-02 2.25784644e-01 6.85783327e-02 -7.24994838e-01
-1.08857155e+00 -6.05980396e-01 9.41475213e-01 -2.78128535e-01
-5.46534359e-01 5.27544320e-01 5.76058984e-01 -7.39012063e-02
1.53555155e-01 -3.74174327e-01 -9.14902210e-01 -8.91097188e-01
4.22737390e-01 4.52274263e-01 -4.86464024e-01 2.37426236e-01
-8.81413147e-02 4.33273047e-01 -1.54432619e+00 2.24280804e-01
4.44788069e-01 6.27882719e-01 1.54879034e-01 5.11007965e-01
6.52831048e-03 4.79884088e-01 -6.67787671e-01 -4.23828870e-01
-1.34722024e-01 -4.47247535e-01 -5.98154783e-01 6.68227851e-01
2.29002997e-01 -1.91584662e-01 -3.59840155e-01 7.49018490e-01
3.79572302e-01 4.06615764e-01 1.04877949e+00 -8.18508387e-01
-5.61686575e-01 8.45581770e-01 -1.70317844e-01 1.59990028e-01
1.50457263e-01 3.38512547e-02 1.05129170e+00 -1.14615345e+00
4.07452762e-01 1.20144856e+00 3.12415123e-01 7.62588441e-01
-1.92335665e+00 -3.68169129e-01 1.77528143e-01 4.17096987e-02
-1.51442623e+00 -1.09194577e+00 1.03231859e+00 -4.87917602e-01
1.04360509e+00 1.11162871e-01 1.55461550e-01 1.94770455e+00
-3.18341225e-01 7.51428485e-01 9.17622626e-01 -6.02170885e-01
3.86122376e-01 4.91519302e-01 2.16208383e-01 3.25632393e-01
-3.71758372e-01 1.71934500e-01 -6.53826118e-01 -2.49454841e-01
5.75918138e-01 -2.05891997e-01 -1.98168904e-01 -4.56559509e-01
-1.00676560e+00 1.08537102e+00 2.83012427e-02 4.67638403e-01
-6.05208993e-01 -1.46994546e-01 3.44727129e-01 5.61475120e-02
7.85803080e-01 -1.14369988e-01 -3.67699414e-01 -1.58449840e-02
-1.26873684e+00 3.23903770e-03 7.77569890e-01 6.83534384e-01
9.30207551e-01 4.66619253e-01 -4.46669996e-01 1.00476098e+00
6.84262455e-01 2.87096411e-01 5.51831186e-01 -8.51629555e-01
4.51717615e-01 1.64539441e-01 -1.20397851e-01 -2.53809571e-01
9.21876878e-02 -4.79554772e-01 -5.72836757e-01 -7.19386786e-02
2.55642265e-01 -2.63217967e-02 -1.17268622e+00 2.01224041e+00
2.91740656e-01 4.57463324e-01 4.18916345e-01 6.67620540e-01
5.54437459e-01 7.85250247e-01 1.68431744e-01 -4.30972844e-01
1.14983058e+00 -7.44222879e-01 -7.91931152e-01 -1.12683885e-01
5.43816946e-02 -7.62225866e-01 1.02968788e+00 2.49988481e-01
-1.12733459e+00 -7.43761241e-01 -8.97989631e-01 -5.39170243e-02
-2.60867745e-01 4.90455210e-01 4.67421412e-02 9.42556858e-01
-1.18920100e+00 4.19368535e-01 -1.13174891e+00 1.02141462e-01
2.09736079e-01 2.05683604e-01 -1.98768064e-01 4.05763760e-02
-9.05981183e-01 6.16733372e-01 3.57647806e-01 1.17184527e-01
-1.70593989e+00 -7.12349236e-01 -1.05989373e+00 1.54513702e-01
1.66208833e-01 -5.54996789e-01 1.28251350e+00 -8.58414888e-01
-2.08606100e+00 1.07889414e+00 -5.00773787e-01 -8.08286726e-01
4.53319013e-01 -2.64113635e-01 -2.84305334e-01 -5.90806874e-03
-2.98033744e-01 3.99986863e-01 1.58507097e+00 -1.09647155e+00
1.18493453e-01 -2.78893918e-01 -1.66794479e-01 1.75912324e-02
-3.35322678e-01 -1.87731028e-01 -2.86027521e-01 -5.37246406e-01
-3.22048329e-02 -9.83415544e-01 6.02210797e-02 -8.40584219e-01
-4.95570898e-01 -4.23978120e-01 7.24624932e-01 -1.05863070e+00
9.78452742e-01 -2.30211043e+00 6.38195455e-01 1.83465645e-01
-6.42511100e-02 1.28433704e-01 -4.81898822e-02 5.40201962e-01
-1.63345501e-01 -2.07851186e-01 -3.00153404e-01 -1.27566922e+00
3.92557025e-01 4.28140432e-01 -7.00817108e-01 5.26021838e-01
2.32007533e-01 6.16324902e-01 -4.44310158e-01 -2.47876465e-01
5.56688249e-01 1.11096442e+00 -5.12160599e-01 4.45540369e-01
-3.54428709e-01 7.98521221e-01 -1.30702332e-01 1.19588494e-01
5.79400897e-01 6.54708073e-02 3.14466774e-01 -9.56895649e-02
-1.81279764e-01 7.21912861e-01 -7.95511961e-01 2.14841390e+00
-8.66705239e-01 3.36581618e-01 2.59085685e-01 -1.15780497e+00
7.84678578e-01 8.96962762e-01 1.74012884e-01 -9.15100798e-02
1.63625300e-01 -3.53336856e-02 -3.61340135e-01 -2.36549288e-01
2.36414567e-01 -3.41625929e-01 1.18319646e-01 4.10841107e-01
7.21684873e-01 8.01561177e-02 -1.00520104e-01 1.11419976e-01
4.87796158e-01 4.06160146e-01 6.85088187e-02 -1.62089154e-01
8.28683794e-01 -6.05232298e-01 3.16591978e-01 4.21842575e-01
1.98759839e-01 5.99634409e-01 2.44129062e-01 4.19201516e-02
-1.11629486e+00 -1.38611686e+00 -2.41560653e-01 1.15173709e+00
-6.27843261e-01 -3.56157392e-01 -1.06261826e+00 -6.42327070e-01
-3.05616856e-01 1.35451746e+00 -7.92691708e-01 -1.66264221e-01
-5.84901094e-01 -4.28136915e-01 5.10157764e-01 3.79083544e-01
-1.44592643e-01 -9.98412192e-01 -2.17225909e-01 3.17675024e-01
-1.96492821e-01 -1.19161677e+00 -4.61748034e-01 2.56719112e-01
-9.51473594e-01 -5.85730553e-01 -9.44492340e-01 -4.86030042e-01
4.16167766e-01 -1.86768577e-01 1.20074081e+00 -7.03107953e-01
-1.58746108e-01 6.05494797e-01 -4.29368019e-02 -1.24996074e-01
-8.68229091e-01 -2.80084275e-02 2.70932943e-01 6.01327240e-01
4.38054621e-01 -8.27490628e-01 -2.53062874e-01 -1.06758233e-02
-8.49701464e-01 -2.53332257e-01 5.74632645e-01 9.21686530e-01
6.56170726e-01 -3.71613294e-01 4.77251530e-01 -4.99702096e-01
3.30704838e-01 -4.09002841e-01 -6.82341158e-01 1.96959347e-01
-3.47998112e-01 4.17638540e-01 5.14961004e-01 -5.55438757e-01
-1.43210423e+00 1.78779349e-01 -4.34628248e-01 -9.94219780e-01
-5.76586902e-01 2.42608562e-01 -5.12544751e-01 6.69191539e-01
3.93760920e-01 6.88096941e-01 -6.56328127e-02 -8.09402168e-01
5.53280652e-01 7.96647191e-01 4.17240173e-01 -6.92933738e-01
6.01736903e-01 3.89428914e-01 -4.48951572e-01 -1.03585017e+00
-8.87286782e-01 -4.16673392e-01 -6.54608250e-01 1.25367910e-01
1.03292799e+00 -1.08051264e+00 -4.23599184e-01 3.73697460e-01
-1.28605282e+00 -1.68662593e-01 -4.84389812e-01 7.20371723e-01
-8.05285811e-01 2.90902376e-01 -5.57689011e-01 -1.23597360e+00
-1.75156578e-01 -1.09596777e+00 1.20376265e+00 -6.90008849e-02
-4.71214131e-02 -1.13909352e+00 3.25014204e-01 4.75722730e-01
2.95072466e-01 -1.66284502e-01 7.58899093e-01 -8.80706787e-01
-4.09774542e-01 -5.07716276e-02 3.91448349e-01 9.38160598e-01
2.35862553e-01 -1.31885290e-01 -1.72383845e+00 -3.58095527e-01
3.65474939e-01 -2.38292873e-01 9.92353261e-01 5.76377332e-01
1.01572931e+00 -4.11496043e-01 -1.76823869e-01 4.63029265e-01
1.10400605e+00 -1.15199104e-01 5.15408754e-01 -4.31342900e-01
4.74852085e-01 7.24716187e-01 -2.81434767e-02 4.38651860e-01
1.36978269e-01 7.77093589e-01 1.14172392e-01 3.95314276e-01
-4.18931013e-03 -3.58124226e-01 6.06570005e-01 9.96581495e-01
-1.01016916e-01 -1.61964104e-01 -5.11580646e-01 6.61656797e-01
-1.37861991e+00 -1.02868295e+00 4.91556674e-01 2.55359650e+00
8.66830230e-01 -2.17850413e-02 2.98588932e-01 1.58564180e-01
6.49776340e-01 3.59409988e-01 -3.53540689e-01 -1.62858635e-01
-1.71302576e-02 2.59844333e-01 -5.28078601e-02 6.59436107e-01
-1.07150757e+00 7.22974360e-01 5.90422249e+00 7.28542864e-01
-1.19143951e+00 5.25844455e-01 3.38396579e-01 -2.41193160e-01
-3.43032300e-01 -1.03263624e-01 -9.47435915e-01 3.84815365e-01
1.49248004e+00 1.10502303e-01 5.48459888e-01 7.57943273e-01
2.86180109e-01 4.83903706e-01 -1.38747931e+00 9.14067090e-01
6.42688572e-02 -9.30401087e-01 -7.03357905e-02 2.78699398e-01
6.20249569e-01 1.04649730e-01 5.16131699e-01 4.57089722e-01
3.15107465e-01 -1.02612948e+00 7.27840126e-01 6.26553178e-01
6.24791503e-01 -7.57883191e-01 3.13546062e-01 6.13960385e-01
-9.29834306e-01 1.66315436e-01 -3.83896619e-01 2.49119923e-01
3.68088484e-01 4.81040210e-01 -1.14217722e+00 6.23856068e-01
3.33688736e-01 6.14804566e-01 2.63372976e-02 3.42276573e-01
-3.92093331e-01 1.16883838e+00 -3.02449405e-01 4.14180011e-01
2.24386066e-01 -3.06307852e-01 7.68139958e-01 1.30196571e+00
1.62618637e-01 -3.93044889e-01 -2.09755264e-02 1.31657529e+00
-1.78639397e-01 1.32376831e-02 -5.01667678e-01 -2.79998779e-01
3.26917142e-01 1.07695270e+00 -2.42396027e-01 -3.32351565e-01
-3.32360536e-01 1.21226692e+00 4.22881901e-01 6.73241079e-01
-9.35838997e-01 3.04471374e-01 8.28923106e-01 -1.34648234e-01
8.81039262e-01 -3.15350741e-01 1.06935956e-01 -1.30008972e+00
-1.46154940e-01 -6.27826691e-01 2.83409119e-01 -3.06135863e-01
-1.27433276e+00 7.31944203e-01 6.85622543e-02 -9.76994574e-01
-9.66477275e-01 -2.72124439e-01 -6.70986831e-01 1.48190641e+00
-1.38834131e+00 -1.49598086e+00 2.20438749e-01 7.67219841e-01
9.76773202e-01 -3.08697492e-01 1.29174459e+00 1.37700930e-01
-5.95803201e-01 6.07714832e-01 1.56567186e-01 7.51406420e-03
2.92695850e-01 -1.14078581e+00 3.45125794e-01 9.24803853e-01
8.92023206e-01 9.87253308e-01 8.45264196e-01 -2.25374922e-01
-1.09817064e+00 -9.46440637e-01 7.27737188e-01 -4.59632635e-01
3.78879875e-01 -6.99947715e-01 -1.08549750e+00 1.02035177e+00
3.65873814e-01 -1.38067991e-01 9.16015983e-01 2.58277178e-01
-5.69189131e-01 1.28242582e-01 -7.81682730e-01 1.87822357e-01
6.66725636e-01 -1.13717091e+00 -9.08994198e-01 4.36105616e-02
6.96948051e-01 -1.97790861e-01 -8.49914193e-01 -9.33916792e-02
3.42178434e-01 -6.97025478e-01 1.28059018e+00 -6.07329249e-01
2.03428362e-02 6.82094367e-03 -5.01357734e-01 -1.36766577e+00
-2.73169447e-02 -6.08460188e-01 -5.29496968e-01 1.64192259e+00
4.82794762e-01 -4.56058025e-01 6.12901866e-01 3.35538298e-01
-2.34151989e-01 -2.58704275e-01 -1.18361330e+00 -6.98773026e-01
1.05068251e-01 -4.36509281e-01 4.01533812e-01 6.98065400e-01
-4.09789085e-01 6.01285219e-01 -5.87142289e-01 4.20743942e-01
8.73623788e-01 -2.12957785e-02 6.60117626e-01 -9.65904176e-01
-7.01322019e-01 -1.60121277e-01 -1.43094793e-01 -1.00767314e+00
9.43338335e-01 -8.30188930e-01 7.22181350e-02 -1.07385588e+00
-1.54384002e-01 1.39403805e-01 -5.47256291e-01 7.57692307e-02
6.41493276e-02 3.83002646e-02 -1.37426369e-02 1.44127458e-01
-5.41459396e-02 9.79831874e-01 5.27724385e-01 -2.86565900e-01
-3.07323396e-01 4.65927035e-01 -2.44652227e-01 5.61281264e-01
4.77675021e-01 -4.77204382e-01 -5.31884611e-01 -3.24602813e-01
-3.06177527e-01 1.74399480e-01 5.44861138e-01 -7.70863056e-01
-2.16469243e-02 2.12027773e-01 3.07856649e-01 -7.49237597e-01
9.61625695e-01 -5.89977682e-01 2.59964138e-01 1.10465012e-01
-5.59005380e-01 -6.31467521e-01 3.62883545e-02 6.69384420e-01
-3.87697041e-01 -2.98979014e-01 6.64203465e-01 -4.50449483e-03
-3.40753347e-01 8.02474990e-02 -3.65314215e-01 -1.76596105e-01
6.74148202e-01 5.98756485e-02 3.95391464e-01 -3.53957325e-01
-1.27257657e+00 -4.35310453e-01 2.77254790e-01 3.66999179e-01
6.48344696e-01 -1.14107621e+00 -9.85799432e-01 3.32970649e-01
1.19262442e-01 -3.34566385e-01 5.65904379e-01 7.90862679e-01
3.61503273e-01 6.86486542e-01 1.45627901e-01 -7.78488040e-01
-1.16703486e+00 9.00188625e-01 3.14391851e-01 -1.85553893e-01
-5.51505387e-01 9.91612077e-01 4.82227802e-01 -5.17145693e-01
4.23402071e-01 -2.23899961e-01 -2.94276536e-01 8.62364173e-02
4.88163322e-01 1.34677708e-01 7.58611932e-02 -9.82409716e-01
-3.53726774e-01 1.88697785e-01 -1.28719375e-01 -7.72313476e-01
1.36457789e+00 -4.14524108e-01 2.61867166e-01 9.80118692e-01
1.40034688e+00 2.57245690e-01 -1.69406259e+00 -5.75410306e-01
-6.09874427e-02 -7.03810826e-02 2.06855953e-01 -5.58307648e-01
-8.08665216e-01 1.21854019e+00 4.85794157e-01 1.67433366e-01
8.23695481e-01 2.78659642e-01 3.99367273e-01 -3.06674540e-02
6.79968223e-02 -8.81093264e-01 -7.91231636e-03 2.74784803e-01
9.65696692e-01 -1.04754925e+00 -4.41955924e-01 -2.42021322e-01
-7.39425004e-01 1.01830804e+00 -7.86184818e-02 -1.04590364e-01
6.44663393e-01 2.65027843e-02 1.19015150e-01 -2.55715139e-02
-6.82100713e-01 -7.90679976e-02 6.15409374e-01 6.00645065e-01
6.14513397e-01 1.64463565e-01 6.38259947e-02 6.53418601e-01
-2.09755585e-01 -3.07842284e-01 1.03949063e-01 4.40393597e-01
4.14574407e-02 -1.15741026e+00 -4.11346346e-01 -4.16012518e-02
-6.20036244e-01 -1.97954595e-01 -1.64446935e-01 2.69068331e-01
-2.23373547e-01 1.00008988e+00 6.84406161e-02 -1.05156824e-01
1.14096887e-01 8.02670479e-01 4.78080243e-01 -7.00738311e-01
-3.85535240e-01 5.18373549e-01 -1.32636484e-02 -4.18072909e-01
-5.62800884e-01 -1.06993401e+00 -8.44114661e-01 2.67777681e-01
-2.91235834e-01 3.16201180e-01 1.15978944e+00 1.19240439e+00
1.21753536e-01 6.90887272e-01 7.60396361e-01 -1.06100523e+00
-1.12539458e+00 -1.43639374e+00 -5.27657509e-01 3.71870667e-01
5.59030414e-01 -5.34902334e-01 -5.83211005e-01 4.57970262e-01] | [14.901813507080078, 6.495072364807129] |
9b0924ed-87c8-4e2d-ae34-7f47778cf6d4 | learning-chess-blindfolded-evaluating | 2102.13249 | null | https://arxiv.org/abs/2102.13249v2 | https://arxiv.org/pdf/2102.13249v2.pdf | Chess as a Testbed for Language Model State Tracking | Transformer language models have made tremendous strides in natural language understanding tasks. However, the complexity of natural language makes it challenging to ascertain how accurately these models are tracking the world state underlying the text. Motivated by this issue, we consider the task of language modeling for the game of chess. Unlike natural language, chess notations describe a simple, constrained, and deterministic domain. Moreover, we observe that the appropriate choice of chess notation allows for directly probing the world state, without requiring any additional probing-related machinery. We find that: (a) With enough training data, transformer language models can learn to track pieces and predict legal moves with high accuracy when trained solely on move sequences. (b) For small training sets providing access to board state information during training can yield significant improvements. (c) The success of transformer language models is dependent on access to the entire game history i.e. "full attention". Approximating this full attention results in a significant performance drop. We propose this testbed as a benchmark for future work on the development and analysis of transformer language models. | ['Kevin Gimpel', 'Karen Livescu', 'Sam Wiseman', 'Shubham Toshniwal'] | 2021-02-26 | null | null | null | null | ['game-of-chess'] | ['playing-games'] | [ 1.07733242e-01 1.32988706e-01 -4.86743957e-01 -9.45012420e-02
-6.84014857e-01 -9.76880014e-01 8.09676588e-01 3.75462532e-01
-4.14268285e-01 6.18447185e-01 1.38653979e-01 -9.58677828e-01
1.52574569e-01 -9.46260095e-01 -7.79891074e-01 -4.26502377e-02
-9.85730886e-02 6.25861228e-01 6.96419656e-01 -5.27182102e-01
2.43722513e-01 2.67073244e-01 -1.25856483e+00 4.04460162e-01
5.33132792e-01 5.32722235e-01 2.09029838e-01 9.41584051e-01
-2.12405533e-01 1.57722259e+00 -4.47525471e-01 -3.46926421e-01
3.87737036e-01 -3.75389338e-01 -1.25352049e+00 -1.79723725e-01
8.24216977e-02 -2.04788134e-01 -6.57427609e-01 8.80605161e-01
-1.38593167e-01 1.53723404e-01 4.55095559e-01 -1.33938682e+00
-1.46973073e-01 8.97097409e-01 -3.27474743e-01 4.10126299e-01
4.23220903e-01 2.95890570e-01 1.31658137e+00 -2.34631911e-01
7.42155433e-01 9.14580166e-01 5.10179698e-01 3.79114717e-01
-1.16302335e+00 -2.82075465e-01 5.58434308e-01 1.44632727e-01
-1.12293088e+00 -4.80028421e-01 5.23706913e-01 -5.10003865e-01
1.23933041e+00 2.20830277e-01 5.16043246e-01 6.98502302e-01
3.22789192e-01 1.03050220e+00 7.96652496e-01 -4.70185161e-01
7.37411752e-02 -1.76790178e-01 5.16086400e-01 9.94314075e-01
4.17818055e-02 -2.68697441e-02 -5.78199565e-01 -1.33852094e-01
5.10639846e-01 -2.00477079e-01 3.89989391e-02 -5.65486193e-01
-9.35090065e-01 7.36326456e-01 1.26693800e-01 3.01297814e-01
-2.71933246e-02 4.13319796e-01 6.10060275e-01 6.75286055e-01
1.64146915e-01 6.56245828e-01 -5.38046896e-01 -9.60778773e-01
-1.01280808e+00 4.89683628e-01 9.39754128e-01 8.89985383e-01
5.12618363e-01 -9.37946364e-02 2.39532754e-01 3.30430090e-01
1.61530618e-02 2.07116067e-01 3.96750361e-01 -8.80865097e-01
6.80330813e-01 7.09876955e-01 3.61041069e-01 -7.86430299e-01
-6.18686318e-01 -5.21914542e-01 -2.55092204e-01 7.27312118e-02
9.80929494e-01 -3.26680429e-02 -6.96403384e-01 2.06263971e+00
2.28976682e-02 4.38063778e-02 1.58250164e-02 4.10850972e-01
5.66849038e-02 7.69518495e-01 -5.86449131e-02 -3.00699510e-02
1.35998285e+00 -9.33562338e-01 -1.84032947e-01 -8.26331854e-01
1.14991653e+00 -4.59722042e-01 1.28819168e+00 5.35052180e-01
-1.40745485e+00 -1.91515341e-01 -1.09571052e+00 -2.67556250e-01
-3.65351826e-01 -2.32562393e-01 9.37002480e-01 4.80326951e-01
-9.75684345e-01 5.13327181e-01 -1.60401857e+00 -3.21103096e-01
1.64028376e-01 3.77613664e-01 -3.29270244e-01 -1.47409439e-01
-1.09258735e+00 9.69104707e-01 3.63035858e-01 -7.06134886e-02
-8.08199346e-01 -5.37798285e-01 -1.12173724e+00 2.64000624e-01
7.68333197e-01 -4.18591201e-01 1.72873414e+00 -6.73115015e-01
-1.40024376e+00 8.25202465e-01 -2.97247350e-01 -8.20037544e-01
6.53609395e-01 -1.96870670e-01 1.40393702e-02 -1.36070073e-01
9.87841487e-02 2.42460057e-01 1.50948510e-01 -7.84001231e-01
-6.83510482e-01 -2.20837712e-01 8.99703443e-01 1.06181577e-01
-1.05749868e-01 1.52946934e-01 -6.40962243e-01 -2.55963027e-01
1.70316964e-01 -9.86826181e-01 -3.74213815e-01 -1.46604910e-01
-3.20873857e-01 -9.57567245e-02 1.46857277e-01 -5.31684399e-01
1.53236604e+00 -2.01444268e+00 1.29649043e-01 1.65980175e-01
2.46818408e-01 9.12951902e-02 -1.57732606e-01 6.58551574e-01
6.16004281e-02 1.94944263e-01 4.26195236e-03 -1.49459362e-01
3.09163570e-01 2.27140412e-01 -5.93522966e-01 4.61862385e-01
4.83587384e-02 1.13899994e+00 -9.41611946e-01 -3.47117245e-01
5.43276928e-02 -3.51184666e-01 -9.16246355e-01 -4.77089845e-02
-5.88389099e-01 8.73258337e-03 -5.22308886e-01 3.11093837e-01
2.41241883e-02 -5.07735670e-01 4.54777420e-01 2.31502339e-01
-8.30230713e-02 9.62822437e-01 -1.17920232e+00 1.72364879e+00
-6.77972734e-01 8.44115078e-01 -1.06950209e-01 -9.33418930e-01
3.78041923e-01 -2.58628819e-02 1.64506525e-01 -7.01889932e-01
3.11639514e-02 -2.76106037e-02 3.80125970e-01 -3.01747113e-01
6.64785326e-01 -2.66887188e-01 -5.30681252e-01 8.95841062e-01
-2.35030949e-01 -1.87336802e-02 6.22137666e-01 5.98469317e-01
1.37732339e+00 1.34988889e-01 3.66729110e-01 -1.78593457e-01
2.21615016e-01 4.85636383e-01 4.76952583e-01 1.06697857e+00
-8.30198526e-02 6.79543754e-03 1.02387857e+00 -6.54572785e-01
-9.21260893e-01 -8.76598299e-01 3.07960302e-01 1.33464372e+00
1.03174038e-01 -9.16881740e-01 -6.24915421e-01 -4.04660136e-01
-3.43097895e-01 8.06964755e-01 -5.79429924e-01 -3.81489962e-01
-7.88441360e-01 -5.79500794e-01 7.85684168e-01 8.37451756e-01
1.70113206e-01 -8.68626893e-01 -7.73942351e-01 3.23834836e-01
-3.89714628e-01 -1.09194112e+00 -3.63281935e-01 4.27728534e-01
-7.66047835e-01 -1.26762176e+00 -7.78490007e-02 -7.01363921e-01
3.76750350e-01 1.37792556e-02 1.33719671e+00 2.75187254e-01
-9.08983946e-02 3.49548310e-01 -7.22183734e-02 -1.63074985e-01
-6.19861841e-01 5.69894075e-01 -2.96149582e-01 -5.32219231e-01
3.33445430e-01 -6.36968136e-01 -7.37849399e-02 4.28187884e-02
-6.71355367e-01 3.71316224e-01 2.15613648e-01 7.45963275e-01
6.31933585e-02 2.72139668e-01 -9.26344469e-02 -1.05740666e+00
6.30372226e-01 -4.00269300e-01 -8.81795228e-01 2.52490312e-01
-2.62780130e-01 4.20517206e-01 8.69304299e-01 -3.10325027e-01
-7.17623651e-01 -1.78754091e-01 4.42607924e-02 9.22404900e-02
-2.21123733e-02 8.86461854e-01 1.82544664e-02 2.46445164e-01
6.89871907e-01 5.05776286e-01 -2.59329706e-01 -3.34468663e-01
2.63917387e-01 1.82175875e-01 7.22010732e-01 -1.10842144e+00
7.64993012e-01 2.69589424e-01 -9.02765766e-02 -7.28867769e-01
-8.11526954e-01 -3.02069485e-01 -5.32768130e-01 1.75998196e-01
4.05623972e-01 -8.04020584e-01 -8.44510257e-01 3.06287497e-01
-1.03151762e+00 -1.00862908e+00 -2.28077278e-01 1.47182643e-01
-7.27699578e-01 4.03691381e-01 -9.56842244e-01 -9.68903720e-01
2.27424964e-01 -1.12339067e+00 8.17746758e-01 -8.34246054e-02
-5.53135335e-01 -1.05263150e+00 1.86366409e-01 4.41072553e-01
1.10714115e-01 -1.05148628e-02 1.27833378e+00 -7.36305475e-01
-8.24707747e-01 -3.79726171e-01 1.66321784e-01 -1.70079261e-01
1.12968846e-03 -7.98118412e-02 -5.99298060e-01 -2.11580649e-01
-2.38955021e-01 -4.77230191e-01 4.60587591e-01 9.28005353e-02
9.41725612e-01 -1.06320515e-01 -3.17122549e-01 3.39896172e-01
1.16455066e+00 2.29081810e-01 4.68558311e-01 6.34761930e-01
4.21739489e-01 3.33333135e-01 4.91510451e-01 2.79750943e-01
7.45751917e-01 7.04190850e-01 1.67696968e-01 2.24000260e-01
2.05043405e-01 -7.09270298e-01 3.80074590e-01 5.24686277e-01
1.91278473e-01 -4.20517623e-01 -1.41380942e+00 7.28846669e-01
-1.92227674e+00 -1.00171244e+00 8.10058340e-02 2.09485221e+00
8.93926620e-01 9.10725772e-01 2.39592806e-01 2.47734159e-01
2.10797995e-01 2.65700907e-01 -5.78227758e-01 -4.72921520e-01
1.64080337e-01 3.08257908e-01 5.63089192e-01 9.52057779e-01
-9.50097322e-01 1.18371248e+00 6.79876328e+00 7.39368320e-01
-1.16091657e+00 -1.22209556e-01 4.00425762e-01 -2.73671031e-01
-3.42688471e-01 2.26913050e-01 -5.88839531e-01 4.38711643e-01
1.10354912e+00 -4.50344265e-01 7.14367092e-01 6.98062301e-01
8.79826099e-02 -1.53532222e-01 -1.38297975e+00 6.93790376e-01
-3.47538590e-01 -1.35143292e+00 -1.22725986e-01 1.41581312e-01
3.21063519e-01 1.44640610e-01 -1.02610350e-01 6.02528930e-01
9.38855946e-01 -1.11715353e+00 1.07772052e+00 1.43914402e-01
4.32631880e-01 -6.92353845e-01 4.21547860e-01 8.69519234e-01
-1.23138165e+00 -2.17190385e-01 -2.73447990e-01 -7.09552050e-01
2.94969529e-01 -7.68096969e-02 -6.84345484e-01 4.08681452e-01
1.03815027e-01 5.22999227e-01 -5.80654562e-01 7.34289348e-01
-2.72204727e-01 9.14791465e-01 -6.48460627e-01 -4.43291329e-02
6.38752222e-01 7.46339038e-02 2.77907491e-01 9.55505848e-01
-3.09970528e-02 1.72102094e-01 5.05274475e-01 6.48156703e-01
2.61320788e-02 -1.59399495e-01 -4.82523382e-01 -5.15409887e-01
3.79908830e-01 7.10083008e-01 -9.72050965e-01 -3.35089952e-01
-5.22996962e-01 6.34065390e-01 6.09720886e-01 1.33183569e-01
-8.44276905e-01 2.82263546e-03 6.50313795e-01 4.46362495e-01
2.08880901e-01 -6.05865896e-01 -3.69705170e-01 -1.39119470e+00
-3.34372595e-02 -9.89575028e-01 4.66271728e-01 -7.76351929e-01
-6.69434130e-01 4.77447867e-01 1.03239119e-02 -7.40910530e-01
-5.81349611e-01 -7.88897991e-01 -5.66242993e-01 5.92655003e-01
-1.21110523e+00 -8.48112404e-01 1.90289140e-01 3.43961596e-01
5.09454191e-01 2.73646321e-02 6.91387832e-01 1.49887323e-01
-4.64485973e-01 5.54126382e-01 -1.83653668e-01 3.64527583e-01
2.54178673e-01 -1.39932978e+00 1.00621295e+00 1.04516566e+00
5.88386893e-01 8.14739287e-01 9.65554774e-01 -4.18206483e-01
-1.48983932e+00 -8.01861525e-01 1.08062840e+00 -8.63315284e-01
1.07184970e+00 -5.83054245e-01 -8.18894267e-01 1.24718893e+00
-1.98069811e-01 -1.77392885e-01 5.92056274e-01 4.71966594e-01
-4.43946689e-01 2.84483314e-01 -4.83513027e-01 8.88340116e-01
1.18191361e+00 -8.62839401e-01 -7.83891439e-01 3.07838023e-01
5.29448152e-01 -7.39461660e-01 -5.50056219e-01 -1.75071910e-01
4.73600745e-01 -7.07538009e-01 7.47623384e-01 -1.32996678e+00
5.48836768e-01 -2.97366679e-01 -9.32007656e-02 -8.82959962e-01
-3.70974272e-01 -7.33812451e-01 -2.04845861e-01 8.49899590e-01
5.89694381e-01 -3.27034950e-01 1.12879527e+00 1.13442743e+00
4.10144776e-02 -7.73928165e-01 -6.45083666e-01 -6.83031142e-01
5.36082387e-01 -9.22571182e-01 4.73230451e-01 6.34910822e-01
5.51091254e-01 4.18924659e-01 -3.74351352e-01 1.90067172e-01
1.08539090e-01 4.28049713e-01 9.40855086e-01 -1.09484696e+00
-6.77927136e-01 -4.33486938e-01 -5.39142966e-01 -1.60227001e+00
3.91705722e-01 -8.81457031e-01 -8.68513621e-03 -1.39400995e+00
2.06540972e-01 -5.09095907e-01 -2.44394571e-01 7.10281909e-01
-8.60630255e-03 7.76120275e-02 3.53347629e-01 -4.23859013e-03
-8.98154378e-01 4.65908498e-02 7.94163406e-01 -8.77334028e-02
5.92350867e-03 8.16895366e-02 -8.97151649e-01 7.90488660e-01
8.23652029e-01 -3.33644062e-01 -8.03166747e-01 -6.93260908e-01
7.07768023e-01 4.39756125e-01 1.65926769e-01 -8.22551787e-01
6.20734692e-01 -5.41576028e-01 -2.38726869e-01 -2.57026434e-01
3.79724354e-01 -5.96926093e-01 -2.21137349e-02 6.39027715e-01
-4.97399986e-01 3.15584302e-01 5.13219774e-01 4.15477097e-01
-1.44036368e-01 -1.92852199e-01 5.72316051e-01 -3.66000950e-01
-7.63087571e-01 1.55634314e-01 -8.69917154e-01 5.88578403e-01
8.29982400e-01 -3.09848636e-01 -1.18842170e-01 -6.53627515e-01
-6.98040187e-01 2.54909098e-01 7.81466424e-01 2.92625278e-01
-9.11871046e-02 -8.61090958e-01 -3.04169565e-01 9.31798741e-02
-5.95881697e-03 4.20847014e-02 6.67327270e-02 5.66796720e-01
-7.00100124e-01 6.20010316e-01 -3.19136567e-02 -2.83230782e-01
-1.19007039e+00 6.76498055e-01 5.68224788e-01 -8.85902762e-01
-4.82586354e-01 6.61132514e-01 3.12619627e-01 -3.83363694e-01
1.12149850e-01 -7.36613691e-01 1.02971926e-01 -2.46670380e-01
5.54447293e-01 -4.97076213e-02 2.02590339e-02 -3.24240834e-01
-4.86787915e-01 1.68049321e-01 -4.17931855e-01 -2.38183647e-01
1.41463280e+00 7.43372142e-02 3.13168354e-02 5.80051243e-01
6.69922411e-01 2.16118276e-01 -1.20119190e+00 -2.93798357e-01
3.90811712e-01 -1.02443732e-01 -4.21463549e-02 -7.77231216e-01
-6.89554870e-01 1.02725005e+00 -1.16045035e-01 4.08319384e-01
7.20591545e-01 -1.27744213e-01 7.33819366e-01 7.50858843e-01
7.28858709e-01 -7.46162355e-01 -2.67868310e-01 8.51700246e-01
2.57755399e-01 -8.55012357e-01 -1.98761851e-01 -2.42694929e-01
-3.76168847e-01 1.08980834e+00 6.13749802e-01 -5.68072386e-02
4.06152278e-01 7.22305059e-01 6.88082597e-04 -2.37246349e-01
-1.24253654e+00 4.23880145e-02 -1.61403835e-01 2.72255361e-01
3.52004647e-01 -2.15006590e-01 1.59728482e-01 8.59067857e-01
-6.41977489e-01 1.00980945e-01 7.64449656e-01 1.28228140e+00
-4.74411219e-01 -1.30038452e+00 -1.70096964e-01 3.56430292e-01
-5.40508628e-01 -2.68144310e-01 -3.90604645e-01 9.97314036e-01
-2.86290407e-01 8.19777966e-01 1.62732154e-01 -1.34267837e-01
2.76225358e-01 2.35420048e-01 6.86393499e-01 -8.56078982e-01
-3.83453041e-01 -4.03416127e-01 3.56437802e-01 -5.93155861e-01
1.26959816e-01 -6.14732325e-01 -1.22550917e+00 -7.18142152e-01
-1.48716182e-01 4.97488379e-01 8.91531035e-02 1.19827521e+00
1.38599902e-01 4.26039159e-01 3.65486667e-02 -2.92847008e-01
-7.53767371e-01 -4.98352110e-01 -4.35886383e-01 1.05093926e-01
3.48033518e-01 -5.09014428e-01 -8.93537253e-02 2.06738710e-01] | [9.088247299194336, 7.353518009185791] |
4fd644a0-974e-4f89-b8ac-b2d8824da723 | conversational-intent-understanding-for | 1901.04899 | null | http://arxiv.org/abs/1901.04899v1 | http://arxiv.org/pdf/1901.04899v1.pdf | Conversational Intent Understanding for Passengers in Autonomous Vehicles | Understanding passenger intents and extracting relevant slots are important
building blocks towards developing a contextual dialogue system responsible for
handling certain vehicle-passenger interactions in autonomous vehicles (AV).
When the passengers give instructions to AMIE (Automated-vehicle Multimodal
In-cabin Experience), the agent should parse such commands properly and trigger
the appropriate functionality of the AV system. In our AMIE scenarios, we
describe usages and support various natural commands for interacting with the
vehicle. We collected a multimodal in-cabin data-set with multi-turn dialogues
between the passengers and AMIE using a Wizard-of-Oz scheme. We explored
various recent Recurrent Neural Networks (RNN) based techniques and built our
own hierarchical models to recognize passenger intents along with relevant
slots associated with the action to be performed in AV scenarios. Our
experimental results achieved F1-score of 0.91 on utterance-level intent
recognition and 0.96 on slot extraction models. | ['Nachman Lama', 'Esme Asli Arslan', 'Sahay Saurav', 'Kumar Shachi H', 'Okur Eda'] | 2018-12-14 | null | null | null | null | ['intent-recognition'] | ['natural-language-processing'] | [ 2.34599262e-01 6.74289227e-01 -1.97285652e-01 -9.47575808e-01
-9.88037586e-01 -5.70023179e-01 9.09263074e-01 -6.21120222e-02
-4.03213978e-01 6.77030444e-01 5.36283374e-01 -8.77165675e-01
1.98822975e-01 -5.14424205e-01 -2.97692657e-01 -3.09159338e-01
1.67564258e-01 7.91897058e-01 2.31539294e-01 -1.01082301e+00
2.70894438e-01 3.76713574e-01 -1.72718441e+00 5.89557767e-01
1.58469588e-01 6.12727642e-01 3.49509835e-01 1.44962215e+00
-5.76172769e-01 1.29713857e+00 -7.28869021e-01 -8.63621607e-02
-2.45188907e-01 -1.73281834e-01 -1.20579064e+00 1.41129181e-01
-2.36700356e-01 -2.72583634e-01 -5.41806042e-01 3.41201663e-01
3.62913981e-02 6.54385924e-01 7.54569590e-01 -1.90420532e+00
3.84318024e-01 4.40621018e-01 3.41290832e-01 8.21770728e-02
8.57457101e-01 1.80385262e-01 9.53758657e-01 -7.09514916e-01
5.61052382e-01 1.18123651e+00 2.90678144e-01 9.12212372e-01
-7.39638567e-01 -3.81999880e-01 7.53036067e-02 3.05432945e-01
-1.07774568e+00 -9.64900851e-01 6.38743758e-01 -2.77649164e-01
1.57715547e+00 6.84670568e-01 3.04847777e-01 9.95684981e-01
1.03801638e-01 1.27246332e+00 3.00061196e-01 -3.12404871e-01
2.43837796e-02 5.19999564e-01 6.27594113e-01 4.19751525e-01
-6.97086453e-01 -1.15985028e-01 -5.24473906e-01 -1.25161991e-01
1.25482515e-01 -2.79139578e-01 3.34302723e-01 2.25338399e-01
-9.40788209e-01 1.13249075e+00 -4.28156376e-01 2.40596458e-02
-5.77870309e-01 -1.03984796e-01 9.60887909e-01 4.29114789e-01
-1.06591005e-02 1.20110460e-01 -5.04650116e-01 -9.25366580e-01
-3.19705576e-01 3.50843221e-01 1.12108243e+00 1.29184318e+00
6.09843612e-01 -7.93503597e-03 -2.50618309e-01 1.02829504e+00
4.74826276e-01 2.06296682e-01 2.02505261e-01 -9.61726964e-01
6.58625364e-01 6.62043810e-01 4.39047396e-01 -7.00503469e-01
-7.63796031e-01 7.03430593e-01 -2.15416119e-01 -1.82429001e-01
4.97815832e-02 -5.01015425e-01 -8.14870298e-01 1.36534607e+00
4.59340096e-01 -1.87464863e-01 8.80570412e-01 6.49155319e-01
1.43379319e+00 1.06422687e+00 5.68426907e-01 -1.46408245e-01
1.88267350e+00 -9.17396963e-01 -1.21333587e+00 -4.93611932e-01
1.04792821e+00 -8.50062311e-01 8.49011719e-01 1.01412475e-01
-7.96990991e-01 -4.59150463e-01 -8.36978376e-01 -1.75787866e-01
-5.75455248e-01 8.31362978e-02 6.70080543e-01 5.57769060e-01
-9.50843155e-01 -4.34330165e-01 -3.13089758e-01 -6.01419926e-01
-4.50104147e-01 6.10546947e-01 -4.57978010e-01 3.42161447e-01
-1.51552486e+00 9.97505784e-01 2.26127237e-01 1.89364910e-01
-1.03755271e+00 6.79382905e-02 -1.40268660e+00 -3.11274022e-01
6.13391519e-01 -2.28879347e-01 1.83996928e+00 -3.43481988e-01
-1.68058503e+00 6.69554710e-01 -8.30184281e-01 -5.54581702e-01
-6.88776746e-02 -1.90590113e-01 -8.10280204e-01 2.44801994e-02
2.71984518e-01 9.61678684e-01 5.41569412e-01 -1.09779918e+00
-1.28112054e+00 -2.16874242e-01 2.52781361e-01 4.90974218e-01
4.33520675e-01 4.36343223e-01 -5.64180791e-01 3.82705301e-01
-1.06493928e-01 -1.16452527e+00 -8.57496187e-02 -1.28992558e+00
-5.43107867e-01 -8.85899246e-01 1.22809196e+00 -5.00176370e-01
1.20177412e+00 -2.15170813e+00 -2.22257480e-01 3.34922895e-02
2.08367966e-02 2.07482129e-01 -1.29899681e-01 8.33167195e-01
-7.59886205e-02 -2.10218474e-01 5.09307146e-01 -5.10921240e-01
3.23019087e-01 5.42619646e-01 -4.26533312e-01 -1.21256568e-01
2.45785803e-01 8.66546214e-01 -4.24443334e-01 -5.68811178e-01
6.23599231e-01 3.24292302e-01 -1.13726199e-01 8.91381025e-01
-3.30205828e-01 1.17906794e-01 -6.52943373e-01 5.41931510e-01
2.20896164e-03 5.06539106e-01 5.23896553e-02 -5.59960492e-02
-4.23679054e-01 8.90506983e-01 -7.77476311e-01 1.01099539e+00
-6.61168158e-01 1.15690660e+00 3.13352227e-01 -5.29874265e-01
7.97068119e-01 8.06996346e-01 5.41490689e-02 -4.57819700e-01
3.25903565e-01 -1.39650494e-01 -1.35217845e-01 -9.85200465e-01
1.32777548e+00 9.08322483e-02 -8.13127041e-01 2.75833160e-01
-8.86911601e-02 -1.14858784e-01 2.33394623e-01 4.69813496e-01
8.62243354e-01 -2.80998766e-01 1.92240983e-01 2.65217602e-01
8.04436088e-01 3.27969700e-01 1.48350835e-01 6.68392181e-01
-4.86217052e-01 2.45164130e-02 6.34911716e-01 -4.85177219e-01
-8.15019488e-01 -3.27658594e-01 3.04510564e-01 1.74340129e+00
1.91853270e-02 -5.35872996e-01 -7.43413687e-01 -3.94120872e-01
-5.34043670e-01 1.57980740e+00 -4.81277317e-01 4.01251167e-02
-4.77323234e-01 4.25751228e-03 7.77551293e-01 3.24163556e-01
2.45137885e-01 -1.17842519e+00 -6.34505391e-01 2.83605963e-01
-6.25549495e-01 -1.52850759e+00 -4.74618673e-01 4.05976087e-01
-1.70801640e-01 -7.96672344e-01 5.59323989e-02 -9.25395548e-01
2.47012287e-01 2.59904712e-01 8.76945317e-01 -2.46490613e-01
1.13005869e-01 6.13990486e-01 -4.02845711e-01 -5.37692368e-01
-8.55748653e-01 1.46953478e-01 -1.28023727e-02 -5.38094528e-02
1.04223907e+00 1.88054368e-01 -8.88077542e-02 7.39304781e-01
-4.17838782e-01 4.72993672e-01 1.65753797e-01 5.38175225e-01
-7.84374997e-02 -3.95785235e-02 5.24219036e-01 -5.27374029e-01
1.04964912e+00 -4.39805478e-01 -3.50248098e-01 2.39119623e-02
1.28336534e-01 -1.35623291e-01 2.19802275e-01 -2.43880171e-02
-1.19544220e+00 2.24521533e-01 -7.80506313e-01 3.02578032e-01
-9.92224097e-01 1.77480251e-01 -4.12898213e-01 4.76665020e-01
2.08938476e-02 3.41742486e-01 5.31939566e-02 1.92505985e-01
2.04525292e-01 1.32092988e+00 5.33098996e-01 -3.94605666e-01
6.36315867e-02 -5.31497374e-02 -4.47922468e-01 -1.33762062e+00
-1.01656668e-01 -1.03282094e+00 -3.13998133e-01 -6.99564040e-01
1.25743985e+00 -8.47012937e-01 -1.72859979e+00 2.85592705e-01
-1.36921954e+00 -3.52389634e-01 5.87343201e-02 3.31555575e-01
-7.87595749e-01 -6.65829256e-02 -5.88291109e-01 -1.65872526e+00
-1.39350459e-01 -1.56825197e+00 1.40963042e+00 2.88642228e-01
-7.49653876e-01 -7.05103934e-01 -1.75785691e-01 1.01816165e+00
2.46062711e-01 -5.11320174e-01 8.49565864e-01 -1.36532581e+00
-2.05202326e-01 -3.63051891e-01 -2.35815514e-02 -1.92130685e-01
-1.81771982e-02 -1.03169858e-01 -1.08256721e+00 3.90078664e-01
-1.96273550e-01 -3.55574787e-01 1.32364422e-01 1.45261720e-01
3.16459894e-01 -2.57352144e-01 -3.69339079e-01 -2.88939327e-01
7.70166457e-01 1.06150448e+00 6.96194351e-01 3.32256287e-01
3.88967961e-01 1.25022209e+00 1.06202936e+00 2.06117555e-01
9.68421280e-01 8.74592304e-01 2.69088507e-01 1.02723926e-01
4.55847025e-01 -3.43558222e-01 6.46374643e-01 5.91306627e-01
2.82779187e-01 -3.65170300e-01 -8.79288971e-01 7.04796910e-01
-1.99714100e+00 -7.32381821e-01 -4.12619799e-01 1.58083689e+00
5.51525712e-01 2.20834464e-01 4.24780160e-01 -1.54357299e-01
5.66286504e-01 1.66906059e-01 -2.02481464e-01 -1.19395196e+00
3.37218702e-01 -7.29363859e-01 3.09472322e-01 1.08344841e+00
-1.18492675e+00 1.45404494e+00 6.05538225e+00 7.19200850e-01
-5.28382301e-01 -3.87729295e-02 6.58314884e-01 1.39459074e-01
-1.71761736e-01 -1.89494923e-01 -1.49137902e+00 -1.18332036e-01
1.80068481e+00 1.81433260e-01 1.36864886e-01 9.14762735e-01
5.98961532e-01 -5.40705860e-01 -1.21440637e+00 8.34546208e-01
1.34449244e-01 -1.17120230e+00 -2.45884359e-01 -1.09372184e-01
-4.01443280e-02 3.80242355e-02 -2.79827744e-01 8.27489316e-01
3.48069578e-01 -1.05745351e+00 3.99108976e-01 3.74312401e-01
4.23978746e-01 -1.03496647e+00 9.56450462e-01 5.24059653e-01
-1.25323999e+00 8.79049487e-03 5.56494296e-02 -3.23670767e-02
6.42206371e-01 -4.15867746e-01 -1.74299252e+00 1.50708824e-01
2.62628734e-01 -1.18211336e-01 -8.66451208e-03 1.64007321e-01
1.51825383e-01 4.54391241e-01 -2.59743452e-01 -7.00611889e-01
5.99169075e-01 -2.85113633e-01 7.20154047e-01 1.47632492e+00
-2.47883797e-01 4.29861307e-01 -6.10446930e-02 3.36161911e-01
3.97157252e-01 8.94844607e-02 -1.09622586e+00 -2.28967264e-01
2.24867091e-01 1.13804138e+00 -5.73979676e-01 -5.54110646e-01
-5.31950474e-01 7.14624047e-01 -4.24000829e-01 4.87766832e-01
-8.38433146e-01 -6.52051389e-01 1.10797071e+00 -2.77621120e-01
8.32152516e-02 -2.67730564e-01 1.67260878e-02 -3.01944017e-01
-2.04440594e-01 -9.48614061e-01 1.08237460e-01 -1.00917315e+00
-4.17525917e-01 1.03429055e+00 2.54841834e-01 -8.13986897e-01
-9.29934621e-01 -5.80383599e-01 -6.90999210e-01 6.50605440e-01
-9.99171793e-01 -1.06375873e+00 1.24938883e-01 6.62146807e-01
1.40242696e+00 -5.19526899e-01 1.12644112e+00 1.56447276e-01
-5.83790421e-01 1.83019653e-01 -6.72986627e-01 2.95285851e-01
2.16323018e-01 -8.12705576e-01 5.93630314e-01 3.65836233e-01
-2.35039145e-01 6.69410110e-01 1.19920886e+00 -6.50124311e-01
-1.81718659e+00 -6.39839232e-01 1.43280137e+00 -6.89042449e-01
5.64823747e-01 -4.87262994e-01 -5.53767443e-01 1.00538623e+00
7.08732307e-01 -1.02702200e+00 9.64446723e-01 2.62886316e-01
2.98438460e-01 2.61854976e-01 -8.13846529e-01 9.15662885e-01
3.14899445e-01 -9.77461576e-01 -1.12791204e+00 3.68697286e-01
9.27371502e-01 -5.08036077e-01 -3.82374376e-01 3.15858573e-01
3.92965376e-01 -7.72240400e-01 7.54492164e-01 -1.07195711e+00
-1.16952047e-01 -2.89310932e-01 -4.93759543e-01 -7.47018754e-01
5.81681848e-01 -7.95413911e-01 3.39926660e-01 1.19844329e+00
8.63651037e-01 -5.30244648e-01 7.48703837e-01 1.52389109e+00
-3.56927365e-01 -3.67456436e-01 -8.92067492e-01 2.63207734e-01
-7.91297197e-01 -1.41833055e+00 6.10910654e-01 2.06945568e-01
5.37836730e-01 8.93125415e-01 -5.35900474e-01 3.14518154e-01
-1.43423453e-01 -3.59797478e-01 1.26186216e+00 -7.51487851e-01
4.22921270e-01 -2.40956739e-01 -3.99580300e-01 -1.51112628e+00
4.67085153e-01 -2.35611215e-01 6.58367038e-01 -1.45857382e+00
-2.05684960e-01 -2.58538388e-02 3.54148090e-01 5.35941184e-01
3.03452194e-01 -2.98255175e-01 8.61216635e-02 -1.07008398e-01
-8.96994591e-01 4.47273254e-01 4.91750598e-01 -1.78771228e-01
-5.40203929e-01 4.50131923e-01 -4.61913675e-01 7.75115192e-01
7.29839146e-01 -1.46453753e-01 -6.15693569e-01 2.32066378e-01
1.91255853e-01 9.06459093e-01 6.72413735e-04 -4.68229622e-01
7.22214758e-01 -2.27043480e-01 -2.28112727e-01 -1.20752788e+00
9.87687349e-01 -9.23042715e-01 -1.07005328e-01 -1.28420472e-01
-8.21147323e-01 1.30524978e-01 5.73815525e-01 1.28686368e-01
-4.71717477e-01 -3.63559037e-01 -1.84989236e-02 2.09885627e-01
-1.18740451e+00 -3.03648412e-01 -1.70559859e+00 -3.95058185e-01
1.13925922e+00 -2.94284344e-01 -1.13524273e-01 -1.09260309e+00
-8.82593095e-01 6.71022832e-01 -1.14283338e-01 8.08557928e-01
9.62065220e-01 -9.12786424e-01 -3.54773432e-01 6.54900551e-01
2.41932541e-01 -3.27138662e-01 3.39799255e-01 6.05639398e-01
-2.84876645e-01 1.18175387e+00 1.55502446e-02 -6.29132092e-01
-1.82544732e+00 3.07158418e-02 2.69189596e-01 1.32162303e-01
-2.61362642e-01 7.97695518e-01 1.28638119e-01 -9.08946037e-01
6.00046217e-01 -6.22540005e-02 -8.70481908e-01 2.33551219e-01
7.03817844e-01 1.02592908e-01 3.65149751e-02 -1.39855015e+00
-3.91109437e-01 2.10567005e-03 -2.31762663e-01 -4.89771187e-01
7.94171512e-01 -5.36792517e-01 8.92546996e-02 8.45637083e-01
1.23510742e+00 -1.81703851e-01 -7.53714561e-01 2.27444082e-01
4.03366424e-02 1.15047777e-02 -1.40196174e-01 -5.36527812e-01
-3.73873830e-01 5.97265363e-01 3.36299986e-01 7.54583716e-01
6.34059429e-01 2.62582630e-01 1.02543807e+00 8.03650081e-01
3.44406545e-01 -1.39773703e+00 -6.19360209e-01 1.01074982e+00
8.32483053e-01 -1.58899212e+00 -6.08461142e-01 -2.09964499e-01
-1.42614746e+00 9.95110154e-01 7.48484790e-01 6.73720241e-01
3.52828503e-01 2.55828351e-01 6.54355705e-01 -5.62874138e-01
-1.46500015e+00 -2.18257710e-01 1.22017823e-01 5.31744480e-01
4.67469871e-01 2.81139821e-01 6.53884187e-02 3.39228004e-01
-4.06991631e-01 -3.64024460e-01 4.80020612e-01 9.53796387e-01
-6.81640387e-01 -8.06428730e-01 -4.94362772e-01 3.40834796e-01
-1.22832783e-01 -1.18721277e-02 -5.82310736e-01 9.59553123e-01
-4.28127229e-01 1.79511559e+00 8.88780802e-02 -7.69562066e-01
6.41374886e-01 7.54052162e-01 -5.65226257e-01 -5.25725365e-01
-6.75524175e-01 1.14639215e-01 1.10939837e+00 -6.64671302e-01
-4.21469659e-01 -7.69849002e-01 -1.71551299e+00 -7.77658150e-02
-1.08653866e-01 6.76954091e-01 9.69995141e-01 1.33446801e+00
1.48966983e-01 5.54644227e-01 6.76588058e-01 -8.79755437e-01
3.10112774e-01 -1.07013237e+00 -2.43033499e-01 5.09302616e-02
6.39507711e-01 -3.81973505e-01 -3.49268347e-01 -1.14056319e-01] | [12.694375038146973, 7.722817420959473] |
3aa3dd23-a5e9-496a-aa45-98b4a2ed483c | tieval-an-evaluation-framework-for-temporal | 2301.04643 | null | https://arxiv.org/abs/2301.04643v2 | https://arxiv.org/pdf/2301.04643v2.pdf | tieval: An Evaluation Framework for Temporal Information Extraction Systems | Temporal information extraction (TIE) has attracted a great deal of interest over the last two decades, leading to the development of a significant number of datasets. Despite its benefits, having access to a large volume of corpora makes it difficult when it comes to benchmark TIE systems. On the one hand, different datasets have different annotation schemes, thus hindering the comparison between competitors across different corpora. On the other hand, the fact that each corpus is commonly disseminated in a different format requires a considerable engineering effort for a researcher/practitioner to develop parsers for all of them. This constraint forces researchers to select a limited amount of datasets to evaluate their systems which consequently limits the comparability of the systems. Yet another obstacle that hinders the comparability of the TIE systems is the evaluation metric employed. While most research works adopt traditional metrics such as precision, recall, and $F_1$, a few others prefer temporal awareness -- a metric tailored to be more comprehensive on the evaluation of temporal systems. Although the reason for the absence of temporal awareness in the evaluation of most systems is not clear, one of the factors that certainly weights this decision is the necessity to implement the temporal closure algorithm in order to compute temporal awareness, which is not straightforward to implement neither is currently easily available. All in all, these problems have limited the fair comparison between approaches and consequently, the development of temporal extraction systems. To mitigate these problems, we have developed tieval, a Python library that provides a concise interface for importing different corpora and facilitates system evaluation. In this paper, we present the first public release of tieval and highlight its most relevant features. | ['Ricardo Campos', 'Alípio Jorge', 'Hugo Sousa'] | 2023-01-11 | null | null | null | null | ['temporal-relation-extraction', 'event-extraction', 'temporal-relation-classification', 'temporal-information-extraction', 'temporal-tagging'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-9.40225199e-02 -9.52489004e-02 -2.34884962e-01 -3.43571454e-01
-4.21404123e-01 -8.72623563e-01 8.26073527e-01 7.97132373e-01
-7.26104915e-01 6.25751257e-01 -3.40593196e-02 -3.88417095e-01
-4.74994481e-01 -6.01679265e-01 -1.85472831e-01 -4.25249040e-01
6.03542924e-02 3.58599395e-01 7.09679842e-01 -2.23349646e-01
3.17893177e-01 4.34859484e-01 -1.78420794e+00 6.83705602e-03
5.24776101e-01 8.42750371e-01 3.57384205e-01 2.60888845e-01
-2.61218190e-01 4.44114000e-01 -5.77983499e-01 -6.61386430e-01
1.16298616e-01 -3.22646111e-01 -8.36551547e-01 -4.42699790e-01
-8.43507126e-02 1.35290891e-01 1.03752591e-01 7.09003925e-01
4.05485153e-01 1.10672325e-01 3.90140355e-01 -1.21197939e+00
-5.65910116e-02 7.71696150e-01 -2.65490413e-01 3.48667204e-01
4.63079870e-01 -2.74526291e-02 1.01430368e+00 -3.89438629e-01
1.00145388e+00 8.24945748e-01 5.31523407e-01 2.39410982e-01
-1.12111247e+00 -3.04367721e-01 7.21716732e-02 2.21336111e-01
-1.32640374e+00 -3.23499620e-01 6.01032376e-01 -4.47080702e-01
1.19061387e+00 4.02479172e-01 6.32798254e-01 1.02168167e+00
4.92403284e-02 4.00730968e-01 1.31881809e+00 -7.58545280e-01
9.61937085e-02 4.98456717e-01 2.83552498e-01 2.49142021e-01
3.99716079e-01 -5.68892136e-02 -4.73545969e-01 3.88076194e-02
1.61269739e-01 -4.82561916e-01 -1.72056943e-01 -3.85647178e-01
-9.30434644e-01 5.08415699e-01 -1.08478568e-01 1.09897089e+00
-1.77657157e-01 -2.87902415e-01 9.17533875e-01 3.51876408e-01
2.40338624e-01 5.38650215e-01 -5.44358671e-01 -7.91450500e-01
-1.01205766e+00 4.21925426e-01 9.84400570e-01 6.29388571e-01
4.00931567e-01 -5.75556993e-01 2.12396115e-01 5.51904440e-01
1.27761394e-01 -5.78120165e-02 4.74600524e-01 -5.46799660e-01
5.12680352e-01 8.39434206e-01 1.24335261e-02 -1.21980989e+00
-4.89525110e-01 -2.66748995e-01 -4.93045151e-01 1.44321591e-01
8.82070959e-01 2.47832924e-01 -1.83624819e-01 1.75753748e+00
2.59976864e-01 -5.13044894e-01 -2.42218971e-02 7.77586281e-01
3.97168785e-01 3.44071597e-01 2.17563719e-01 -5.83025992e-01
1.29245853e+00 -3.83589685e-01 -8.01504731e-01 -1.37977272e-01
7.14682579e-01 -1.09325242e+00 9.32141960e-01 4.98720765e-01
-9.34766710e-01 -2.14411870e-01 -1.16616547e+00 -5.46528362e-02
-9.11155522e-01 -2.85345882e-01 7.63473153e-01 6.43085182e-01
-6.46485090e-01 8.65339696e-01 -1.01191032e+00 -8.32107008e-01
-1.81861341e-01 3.42036664e-01 -4.96179253e-01 2.20086217e-01
-1.14177108e+00 1.33436775e+00 5.83229065e-01 8.39050785e-02
8.43078569e-02 -4.01449353e-01 -5.97463727e-01 -9.84431729e-02
5.26886821e-01 -2.67477721e-01 1.34960985e+00 -6.72855198e-01
-1.21224511e+00 7.57876754e-01 1.10503444e-02 -5.35188794e-01
8.98333371e-01 -6.31415173e-02 -5.84917188e-01 -2.79141724e-01
2.63793617e-01 1.99250057e-01 3.00182611e-01 -6.64617538e-01
-7.80176342e-01 -3.26778829e-01 4.21157330e-02 1.64853074e-02
-3.59325469e-01 3.90487820e-01 -7.91293263e-01 -5.57868302e-01
-1.34509638e-01 -7.84903646e-01 -1.92059278e-01 -5.05114853e-01
1.97222619e-03 -3.83023411e-01 8.24890792e-01 -3.71738315e-01
1.86687994e+00 -2.28471804e+00 1.59128055e-01 -3.82669792e-02
-2.00676516e-01 3.27358812e-01 2.45167106e-01 7.80264199e-01
-8.53430703e-02 3.51691216e-01 -1.39294803e-01 -2.39840090e-01
8.32045898e-02 3.20562780e-01 -2.28609875e-01 4.06315506e-01
-4.95328568e-03 4.51352894e-01 -9.02099788e-01 -6.94886029e-01
4.97391343e-01 5.35357714e-01 -1.85966700e-01 -1.33666173e-01
-2.68207550e-01 1.54888928e-01 -5.68075299e-01 1.39655858e-01
3.42790127e-01 2.19159529e-01 3.15332234e-01 -3.19015943e-02
-9.03621793e-01 8.32104564e-01 -1.41625905e+00 1.71406960e+00
-3.00514072e-01 6.85825348e-01 -4.43012148e-01 -9.39403534e-01
8.02332997e-01 7.27742672e-01 6.64062083e-01 -9.75993156e-01
1.64853036e-01 5.49368203e-01 -1.93447806e-02 -3.90662253e-01
7.87218332e-01 1.69863664e-02 -3.64562422e-01 6.24802470e-01
-1.52767003e-01 -8.74149278e-02 1.03936946e+00 2.82928981e-02
1.03147459e+00 4.35247362e-01 5.50183952e-01 -1.41797140e-01
5.02754211e-01 2.49313563e-01 8.57161343e-01 3.60719860e-01
-2.30449215e-01 5.35437822e-01 7.25834727e-01 -5.81452906e-01
-8.74625683e-01 -7.33972847e-01 -2.91158289e-01 7.95455396e-01
-3.85215618e-02 -8.36705089e-01 -6.55854225e-01 -7.29725540e-01
-2.50799835e-01 6.45916998e-01 -6.04907691e-01 2.94978678e-01
-6.90357268e-01 -7.63184369e-01 5.83095610e-01 2.65683770e-01
3.73657644e-01 -1.12003827e+00 -1.26025808e+00 4.03948903e-01
-3.67605895e-01 -1.00157535e+00 1.55164003e-01 3.90771300e-01
-9.55172479e-01 -1.23657596e+00 -2.58669972e-01 -3.02262098e-01
1.45914584e-01 -8.54445547e-02 1.09381568e+00 4.97683845e-02
-7.35900700e-02 5.59902936e-02 -6.78900301e-01 -6.61546767e-01
-6.13920510e-01 5.35404325e-01 2.65177339e-02 -4.35996532e-01
6.50318146e-01 -4.92762595e-01 -1.56620398e-01 3.15090597e-01
-1.03102946e+00 -9.23809409e-02 4.47889686e-01 2.84666896e-01
2.62389451e-01 3.31826478e-01 4.03355211e-01 -8.12118888e-01
6.42073154e-01 -2.23863065e-01 -7.92516649e-01 3.07948738e-01
-9.49567080e-01 2.12214351e-01 4.66541201e-01 -2.44882166e-01
-8.21377277e-01 -8.60703737e-02 -4.40726914e-02 3.56273293e-01
-2.51331180e-01 9.16221082e-01 -2.79630516e-02 4.35172856e-01
6.88058972e-01 -1.13443807e-01 -2.37682000e-01 -6.58742487e-01
6.89147189e-02 5.97307026e-01 3.37283432e-01 -5.11110842e-01
5.84006131e-01 1.10284500e-01 -2.01769173e-01 -7.59797454e-01
-5.36012948e-01 -5.49176753e-01 -7.07573414e-01 -1.28808945e-01
6.13724053e-01 -3.63826632e-01 -3.41972500e-01 1.98395684e-01
-1.05422425e+00 -2.03287989e-01 -2.54398584e-01 4.95419890e-01
-3.61850321e-01 4.55870688e-01 -2.60735244e-01 -7.34229386e-01
-1.47823080e-01 -1.24835801e+00 3.98996353e-01 2.82475576e-02
-8.44078243e-01 -7.70868599e-01 9.67267156e-02 2.67107308e-01
4.13391680e-01 4.18431431e-01 8.74858499e-01 -3.57841253e-01
-3.19608927e-01 -4.63701844e-01 -4.70222458e-02 1.20722815e-01
7.67917037e-02 4.89343137e-01 -7.78600633e-01 -9.43071842e-02
2.20005121e-02 -1.89555045e-02 2.68426269e-01 1.24782622e-01
4.93634850e-01 1.28696442e-01 -4.87631589e-01 -1.31447762e-01
1.42858291e+00 2.56234437e-01 5.85332751e-01 9.63958859e-01
5.45927621e-02 8.29860806e-01 1.01674020e+00 3.90614778e-01
3.95915508e-01 1.17466724e+00 1.15485169e-01 2.81048626e-01
5.48244230e-02 6.31268919e-02 3.29431593e-01 9.06289995e-01
-2.78618425e-01 -1.23709120e-01 -9.45953369e-01 7.62474716e-01
-1.83822143e+00 -8.22922468e-01 -2.07894862e-01 2.36473966e+00
8.04579556e-01 3.73585671e-01 2.55794853e-01 7.78034925e-01
3.12590241e-01 1.28219202e-01 1.69314340e-01 -6.95311964e-01
-1.81172956e-02 4.69500711e-03 1.98223948e-01 2.56374508e-01
-9.69540358e-01 6.36336386e-01 5.31756496e+00 4.30120111e-01
-1.49545288e+00 -8.24080687e-03 5.99708110e-02 3.30431387e-02
6.87737688e-02 3.74340296e-01 -6.66551650e-01 5.16691089e-01
1.18017793e+00 -5.92510104e-01 1.02196194e-01 7.08094299e-01
3.92099887e-01 -3.91240388e-01 -1.20564270e+00 6.94642901e-01
-3.11178476e-01 -9.59191978e-01 -4.10535812e-01 -3.65790166e-02
9.60224196e-02 1.20405130e-01 -1.18276469e-01 2.26288587e-01
-9.04239044e-02 -6.53884768e-01 1.18616009e+00 2.52117634e-01
4.81733829e-01 -6.60622597e-01 7.66308665e-01 4.25669223e-01
-1.20862818e+00 1.24878675e-01 -1.07432961e-01 -2.48096228e-01
2.52541989e-01 7.05265343e-01 -8.52334857e-01 9.11709368e-01
9.47768033e-01 2.89564520e-01 -5.20898759e-01 1.06580579e+00
-1.55029997e-01 4.37377602e-01 -5.11854827e-01 -1.56897068e-01
1.56304091e-01 -1.55756529e-02 5.03957033e-01 1.53270066e+00
3.23884696e-01 -2.25965247e-01 -6.50310144e-02 5.55121362e-01
4.00108278e-01 6.11827016e-01 -5.87713778e-01 -3.25205326e-01
5.40562332e-01 1.13820195e+00 -9.39249933e-01 1.47947758e-01
-7.38262057e-01 4.95031625e-01 2.67293334e-01 -1.00971103e-01
-7.99597561e-01 -3.07988465e-01 3.74021590e-01 4.19440150e-01
6.49458617e-02 -5.23882091e-01 -3.33662778e-01 -7.85151124e-01
4.70183313e-01 -9.41601515e-01 7.21145511e-01 -2.84637809e-01
-7.86255002e-01 7.56204963e-01 3.86409938e-01 -1.13436067e+00
-4.49238420e-01 -3.62876296e-01 -2.61281468e-02 8.40248644e-01
-1.14827645e+00 -8.65406513e-01 -7.39020482e-02 2.27165520e-01
2.17263386e-01 2.37910897e-01 8.60968590e-01 5.65341175e-01
-6.57711804e-01 3.76672387e-01 -4.65003043e-01 5.56940958e-02
8.72782171e-01 -1.17096603e+00 3.27342927e-01 1.09141803e+00
2.75404632e-01 8.99987459e-01 1.10585582e+00 -4.24938470e-01
-1.04564261e+00 -5.34273684e-01 1.65664613e+00 -5.49095809e-01
8.02886963e-01 -1.83045208e-01 -9.71597612e-01 4.85882550e-01
1.36386067e-01 -2.24790245e-01 3.39504600e-01 2.62006015e-01
-4.29684788e-01 -1.68995470e-01 -8.06297302e-01 5.88469267e-01
6.07983828e-01 -3.62016380e-01 -7.85425067e-01 -2.39002407e-01
2.41314441e-01 -3.20500582e-01 -1.12645555e+00 5.46472847e-01
4.75117296e-01 -1.25287831e+00 3.75119716e-01 6.45246357e-02
9.89141241e-02 -4.65088636e-01 1.07682437e-01 -8.38434935e-01
1.29348993e-01 -5.46138644e-01 2.34491274e-01 1.71306884e+00
6.56396091e-01 -7.45942891e-01 4.77290869e-01 8.62514317e-01
1.36620894e-01 -5.73814332e-01 -9.95330811e-01 -8.71515810e-01
-2.23275591e-02 -9.15160775e-01 6.03340089e-01 1.00994563e+00
4.16500419e-01 1.94100022e-01 6.46969602e-02 -1.33038193e-01
1.92852952e-02 2.16070503e-01 7.07694113e-01 -1.44060695e+00
-1.98755965e-01 -7.14899838e-01 -4.26353365e-01 -4.36715931e-01
-1.84798077e-01 -7.72383630e-01 -1.95272535e-01 -1.38739121e+00
-2.27226894e-02 -8.40614855e-01 -1.79651231e-01 5.53185344e-01
3.36420566e-01 6.39694929e-02 1.40521988e-01 2.00737342e-01
-3.38901430e-01 -1.68595817e-02 7.03454912e-01 2.61735290e-01
-4.01306897e-01 -1.23949423e-01 -5.99422336e-01 5.79514027e-01
7.54117012e-01 -7.41179347e-01 -3.49032938e-01 -4.44488704e-01
6.45043433e-01 -2.21987009e-01 7.24785700e-02 -1.11990583e+00
2.43600190e-01 -3.55687998e-02 -1.86727092e-01 -4.91926789e-01
-2.47283354e-02 -8.27839732e-01 4.11917895e-01 3.75696838e-01
1.38686690e-02 3.56614888e-01 4.00178224e-01 7.39924908e-02
-5.23012698e-01 -4.87271935e-01 5.51318645e-01 -3.41839604e-02
-8.43347967e-01 1.94205213e-02 -3.51260811e-01 -6.61617741e-02
1.03048682e+00 -3.22160393e-01 -1.24628402e-01 1.03209704e-01
-3.98561388e-01 -1.15994014e-01 7.42370069e-01 7.75897682e-01
-1.58342913e-01 -7.59576440e-01 -4.07604218e-01 -1.14238553e-01
3.19114327e-01 -9.69357938e-02 -6.37819096e-02 1.04140413e+00
-5.43603659e-01 6.19485378e-01 -1.13100834e-01 -3.98997843e-01
-1.49008238e+00 6.40330017e-01 1.25931323e-01 -6.51884973e-01
-7.16953576e-01 2.10222483e-01 -4.82847631e-01 -4.85024229e-02
9.79321152e-02 -4.71290916e-01 -2.86805093e-01 4.08119619e-01
3.99409235e-01 3.31417352e-01 5.20947874e-01 -5.70952535e-01
-4.88424063e-01 2.45508328e-01 -1.97813809e-01 -4.31922734e-01
1.45831406e+00 -1.04784526e-01 -3.09379637e-01 6.65141582e-01
7.40936518e-01 2.45143563e-01 -6.92605019e-01 -5.19050397e-02
8.45494151e-01 -3.51722449e-01 -1.05432153e-01 -7.60033250e-01
-6.00579441e-01 5.84277451e-01 4.08327609e-01 6.71256542e-01
1.14229989e+00 -1.69800460e-01 4.10424501e-01 1.66523069e-01
6.52318537e-01 -1.38065016e+00 -4.80956286e-01 4.84097451e-01
6.90534472e-01 -9.28981304e-01 1.68874785e-01 -5.27648926e-01
-3.45149487e-01 1.11720335e+00 1.24751344e-01 4.67339396e-01
3.36983383e-01 5.65753102e-01 3.78328383e-01 -1.64875194e-01
-7.62089312e-01 -2.67448604e-01 8.91811699e-02 2.27272540e-01
1.05359542e+00 -1.16017774e-01 -1.21902263e+00 5.08870482e-01
-3.76865864e-01 3.72438163e-01 4.36124384e-01 1.07698202e+00
-4.12452035e-02 -1.89521623e+00 -1.78389013e-01 1.07917283e-02
-8.44857216e-01 2.96363413e-01 -4.30920899e-01 1.34596908e+00
3.64239573e-01 1.10499001e+00 -3.01981330e-01 -1.32345080e-01
5.85916221e-01 1.92136154e-01 5.44086576e-01 -5.51896095e-01
-9.09024298e-01 -1.40590191e-01 3.57367218e-01 -5.17517567e-01
-5.77177107e-01 -1.11852264e+00 -1.12078595e+00 -3.27113360e-01
-3.03589970e-01 4.56255108e-01 8.94293785e-01 1.06851721e+00
1.48126120e-02 2.34496281e-01 7.42602199e-02 -2.86446154e-01
-2.87893683e-01 -9.81792748e-01 -3.16223562e-01 6.00259185e-01
-3.47721100e-01 -7.98732936e-01 -1.58745155e-01 -8.45835581e-02] | [9.443304061889648, 9.234559059143066] |
2af25376-53a8-4476-9b38-f0118176c073 | ischemic-stroke-lesion-segmentation-in-ct | 1811.01085 | null | http://arxiv.org/abs/1811.01085v1 | http://arxiv.org/pdf/1811.01085v1.pdf | Ischemic Stroke Lesion Segmentation in CT Perfusion Scans using Pyramid Pooling and Focal Loss | We present a fully convolutional neural network for segmenting ischemic
stroke lesions in CT perfusion images for the ISLES 2018 challenge. Treatment
of stroke is time sensitive and current standards for lesion identification
require manual segmentation, a time consuming and challenging process.
Automatic segmentation methods present the possibility of accurately
identifying lesions and improving treatment planning. Our model is based on the
PSPNet, a network architecture that makes use of pyramid pooling to provide
global and local contextual information. To learn the varying shapes of the
lesions, we train our network using focal loss, a loss function designed for
the network to focus on learning the more difficult samples. We compare our
model to networks trained using the U-Net and V-Net architectures. Our approach
demonstrates effective performance in lesion segmentation and ranked among the
top performers at the challenge conclusion. | ['S. Mazdak Abulnaga', 'Jonathan Rubin'] | 2018-11-02 | null | null | null | null | ['ischemic-stroke-lesion-segmentation'] | ['medical'] | [ 2.52946645e-01 -1.01065002e-02 -3.82884920e-01 -3.92808676e-01
-1.23386753e+00 -7.61901855e-01 2.50024498e-01 2.06554666e-01
-8.35335314e-01 5.76963544e-01 4.87806231e-01 -5.04209340e-01
3.11852656e-02 -5.29878199e-01 -5.51966906e-01 -5.28595328e-01
-3.19235444e-01 7.43910551e-01 5.88009596e-01 8.19927976e-02
6.28612936e-02 1.01171827e+00 -5.65436125e-01 8.25035453e-01
7.42838025e-01 8.84791791e-01 1.65841848e-01 8.25713098e-01
6.63532764e-02 8.75639737e-01 -4.71383512e-01 3.69964428e-02
4.85877573e-01 -5.36208868e-01 -1.19695091e+00 -3.21978092e-01
5.99083006e-01 -6.23607755e-01 -6.78340256e-01 7.16081440e-01
1.00659132e+00 5.64371422e-02 8.35688829e-01 -7.84112871e-01
8.13343562e-03 8.06439698e-01 -2.78863549e-01 1.17224836e+00
-2.24599659e-01 4.59604770e-01 5.95331967e-01 -6.47260129e-01
7.79526174e-01 8.95109177e-01 1.01745486e+00 6.37119174e-01
-1.12648129e+00 -3.00930470e-01 -1.16821624e-01 5.51555395e-01
-9.42326248e-01 -2.31765017e-01 2.37781882e-01 -6.98479295e-01
9.60406184e-01 1.73606098e-01 8.00265670e-01 1.16999948e+00
3.42867404e-01 1.11903036e+00 7.94918656e-01 -3.03550772e-02
1.44587904e-01 -4.88175333e-01 4.31545287e-01 4.13816512e-01
8.76386743e-03 7.96739683e-02 1.95967034e-01 -1.27643012e-02
9.09286320e-01 2.06044406e-01 -6.08580887e-01 -3.73664856e-01
-1.04937327e+00 7.78231025e-01 1.02393854e+00 1.69583648e-01
-6.54996574e-01 3.41612428e-01 6.39331043e-01 -2.28899270e-01
5.91834709e-02 4.89705116e-01 -4.25850332e-01 -1.43124744e-01
-1.27950597e+00 2.18160570e-01 4.03229147e-01 3.58137429e-01
-1.22029372e-01 -2.88347840e-01 -1.04903138e+00 6.32717490e-01
-1.73413500e-01 1.23436958e-01 5.09377718e-01 -9.65912223e-01
4.52831388e-01 5.99142492e-01 -1.17196672e-01 -1.49056137e-01
-1.16099572e+00 -7.39887774e-01 -5.76928556e-01 5.59216082e-01
7.91265070e-01 -6.10449553e-01 -1.59054923e+00 1.31546378e+00
-1.77481472e-01 1.25836521e-01 -2.48361334e-01 9.90153253e-01
9.99393046e-01 1.77078724e-01 5.40724277e-01 3.77643079e-01
1.15214455e+00 -1.20616078e+00 -2.38737360e-01 -3.47423851e-01
8.56405318e-01 -3.50723803e-01 9.07956123e-01 2.42058188e-02
-1.41301608e+00 -1.16941832e-01 -9.44017112e-01 -2.26784989e-01
-1.45021379e-01 2.82991379e-01 2.76124239e-01 5.92208028e-01
-1.34316766e+00 8.01179111e-01 -1.13058984e+00 -2.99589813e-01
1.48619807e+00 4.72657382e-01 -2.44805753e-01 -1.15464181e-01
-8.54398727e-01 1.51260710e+00 3.63835812e-01 4.42223586e-02
-1.08854425e+00 -1.25072384e+00 -5.62080979e-01 3.59338462e-01
1.30154127e-02 -6.85514629e-01 1.37508857e+00 -7.11125433e-01
-1.26801074e+00 8.71816516e-01 -2.21716687e-02 -1.03838146e+00
1.07440448e+00 -8.46636202e-03 1.80182010e-01 8.57465744e-01
9.08547267e-03 9.31794524e-01 3.64382178e-01 -6.85514212e-01
-5.83493412e-01 -8.16574916e-02 2.31974088e-02 1.61786124e-01
3.15560579e-01 1.98459998e-01 7.24969655e-02 -5.31966925e-01
6.87961578e-02 -6.55266404e-01 -6.49113238e-01 2.36042798e-01
-2.29376584e-01 -3.86751294e-02 9.08933580e-01 -1.10015380e+00
6.25781298e-01 -1.81650138e+00 -7.35140294e-02 3.47487599e-01
5.87227464e-01 5.76801538e-01 2.27990765e-02 -2.03928933e-01
-3.67881387e-01 2.46322989e-01 -4.92671758e-01 1.48194715e-01
-3.69151622e-01 -7.00473040e-02 -1.12098813e-01 3.99590969e-01
1.73398182e-01 1.48053277e+00 -8.08308303e-01 -3.17550391e-01
3.29558939e-01 4.94086236e-01 -5.66849768e-01 3.92532796e-02
8.77568647e-02 7.10730910e-01 -3.81327242e-01 2.01033190e-01
4.84990180e-01 -2.85012037e-01 -1.92835122e-01 -3.40028465e-01
3.59037369e-02 1.65087685e-01 -6.86218083e-01 1.71864080e+00
-1.58953488e-01 6.79018676e-01 3.64281870e-02 -1.30394888e+00
3.68290126e-01 4.82765704e-01 8.41778219e-01 -4.49241877e-01
5.72950125e-01 3.23609531e-01 4.44536865e-01 -6.29847646e-01
-2.90967137e-01 1.22215850e-02 2.70967841e-01 4.18944508e-01
-2.04854812e-02 -1.89840153e-01 2.89362460e-01 2.76940733e-01
1.41819835e+00 -4.18948010e-02 -4.46733385e-02 -4.08005416e-01
2.28982300e-01 1.65331721e-01 3.07728589e-01 1.11324108e+00
-8.95805597e-01 1.11317480e+00 8.44872713e-01 -7.36369491e-01
-9.10695910e-01 -1.26862288e+00 -3.88749242e-01 5.55524051e-01
-2.59916574e-01 1.18132241e-01 -1.01586497e+00 -9.10748422e-01
-2.59577602e-01 4.56772953e-01 -8.48699272e-01 -1.87846452e-01
-9.27939355e-01 -8.53193700e-01 7.38716245e-01 1.09644866e+00
6.49257183e-01 -1.31819558e+00 -1.09935164e+00 4.40434247e-01
-3.73592824e-01 -1.04445446e+00 -4.98398393e-01 3.11158985e-01
-9.75483477e-01 -1.39713478e+00 -1.29164624e+00 -1.04939032e+00
5.48639596e-01 -1.51235774e-01 1.06872761e+00 1.95789725e-01
-7.95532823e-01 1.35601267e-01 -2.19179466e-01 -5.27240336e-01
-1.51795447e-01 4.33365822e-01 -7.03689754e-01 -4.35612082e-01
8.47807378e-02 -4.74339962e-01 -1.20995975e+00 3.42326388e-02
-6.34035766e-01 -7.27989227e-02 5.27462482e-01 8.17727387e-01
4.86619830e-01 -5.75317264e-01 5.92726946e-01 -7.89127827e-01
5.92691183e-01 -3.57881159e-01 -1.56526938e-01 3.17198932e-01
-3.22961174e-02 -3.10974061e-01 2.82919317e-01 -1.43609256e-01
-5.23604155e-01 3.39322478e-01 -3.50294888e-01 -1.68222502e-01
-2.66147494e-01 3.91065031e-01 2.08424851e-01 -3.96217763e-01
1.15125620e+00 -1.08716503e-01 -5.92067093e-03 -2.46777937e-01
3.36361051e-01 2.50592262e-01 8.29694688e-01 -3.63549918e-01
1.92987993e-01 6.08308733e-01 2.51059502e-01 -3.58854115e-01
-9.22244787e-01 -3.95737588e-01 -9.73803580e-01 -2.81061053e-01
1.34922397e+00 -4.62003291e-01 -3.33425313e-01 4.19565946e-01
-1.10009515e+00 -8.67888927e-01 -5.94409525e-01 4.61389780e-01
-5.19096315e-01 1.45776957e-01 -7.56330669e-01 7.02310428e-02
-7.10681558e-01 -1.74017978e+00 7.37996519e-01 3.58616531e-01
-1.16921782e-01 -9.29618716e-01 -1.79915488e-01 9.82320309e-02
8.83470893e-01 5.83661377e-01 1.16637826e+00 -7.53520489e-01
-4.04512167e-01 -4.44090664e-01 -6.82833433e-01 3.14967424e-01
-5.65641709e-02 -2.48493806e-01 -7.58966684e-01 -2.36832798e-01
-4.42134112e-01 -3.43600720e-01 1.25394964e+00 1.31122077e+00
1.54629171e+00 4.16730374e-01 -4.61637855e-01 8.82754385e-01
9.94028389e-01 2.08684176e-01 7.89032459e-01 3.46045464e-01
6.15095854e-01 1.88922733e-01 -5.08740842e-01 -7.91344559e-04
2.41675004e-01 3.36474806e-01 4.09878075e-01 -5.86500168e-01
-6.44975007e-01 3.49811137e-01 -2.80816108e-01 -1.29182309e-01
6.12749830e-02 -3.31082754e-02 -1.33146250e+00 6.91710711e-01
-1.66859579e+00 -1.00194347e+00 -2.76439667e-01 1.77154660e+00
7.37216949e-01 3.02531213e-01 2.93236524e-01 -1.22261465e-01
5.89471400e-01 -8.20928663e-02 -6.21409178e-01 -3.40844542e-01
-1.55024529e-01 8.73062015e-01 8.18711102e-01 5.38547873e-01
-1.53196633e+00 8.45017314e-01 7.08503962e+00 5.83847582e-01
-1.36591208e+00 2.76956737e-01 8.70910823e-01 -2.06984907e-01
3.34196150e-01 -2.17858210e-01 -2.64347434e-01 3.02756459e-01
8.01450431e-01 -1.04916897e-02 8.53809714e-02 3.57547730e-01
3.92315447e-01 -1.20195650e-01 -9.23143864e-01 5.07209778e-01
-2.98909307e-01 -1.71844292e+00 -1.48066446e-01 -4.57460105e-01
8.68196309e-01 7.79835939e-01 1.87354535e-03 2.35467598e-01
3.47183377e-01 -1.38134825e+00 4.81988132e-01 4.92161065e-01
5.81897616e-01 -5.48669100e-01 8.80872369e-01 4.92264852e-02
-7.85096407e-01 -8.67570192e-02 -9.18974578e-02 4.30329591e-01
3.81161332e-01 1.64709061e-01 -9.07323539e-01 6.07237741e-02
8.54887009e-01 7.15144753e-01 -6.65493011e-01 2.01168680e+00
-3.50472540e-01 8.87641370e-01 -3.48829985e-01 4.65038985e-01
5.59050143e-01 1.18384637e-01 5.58771431e-01 1.42047238e+00
1.00445291e-02 2.98131466e-01 5.68239748e-01 8.12503695e-01
-1.24036022e-01 1.44334212e-01 -1.03778519e-01 3.82153213e-01
-5.68833202e-02 1.14321697e+00 -1.13735986e+00 -6.04010284e-01
-5.48665524e-02 7.28740811e-01 3.07211131e-01 4.24983621e-01
-5.78426659e-01 -3.78528863e-01 1.82880029e-01 1.66068196e-01
1.96095526e-01 -1.07108548e-01 -9.54202950e-01 -8.95352185e-01
-1.76751822e-01 -5.87155819e-01 6.90685272e-01 -6.25164926e-01
-9.58539903e-01 6.30907774e-01 -1.35834023e-01 -7.68073201e-01
7.44691491e-02 -8.32448661e-01 -1.13981783e+00 1.18791449e+00
-1.51583576e+00 -8.59253526e-01 -4.27242637e-01 5.67356467e-01
4.26476479e-01 4.95914966e-02 6.77987695e-01 4.04935181e-01
-4.76781726e-01 5.21809161e-01 -1.84422582e-02 4.73666728e-01
5.66534400e-01 -1.34306145e+00 3.88453931e-01 9.53574538e-01
-4.08343405e-01 2.02405751e-01 3.28193367e-01 -5.50812781e-01
-4.25906211e-01 -1.14588630e+00 5.21356463e-01 -4.50415194e-01
5.06409526e-01 7.42877349e-02 -8.60238612e-01 7.57433534e-01
1.07808307e-01 3.73229623e-01 2.71904975e-01 -4.46072876e-01
-1.32581100e-01 3.07152122e-01 -1.48011386e+00 5.36532938e-01
8.47572923e-01 -3.52697164e-01 -6.22891426e-01 8.68129194e-01
2.06359267e-01 -8.46707761e-01 -6.87860429e-01 6.43126011e-01
3.99850637e-01 -5.26905596e-01 1.12020874e+00 -1.00482023e+00
6.02687359e-01 1.71999812e-01 4.55015868e-01 -1.46342027e+00
-4.18293744e-01 -2.14763671e-01 4.73491490e-01 1.44222751e-01
6.90070212e-01 -4.89199996e-01 1.20298791e+00 8.31444323e-01
-6.53928995e-01 -8.62579823e-01 -1.06338799e+00 -3.65187407e-01
8.72128129e-01 -1.11236833e-01 2.46193811e-01 5.90344846e-01
-1.53214008e-01 -3.36544454e-01 3.10783774e-01 2.00196244e-02
5.62968254e-01 -3.59835148e-01 -5.07428534e-02 -1.12995791e+00
1.22775748e-01 -1.32523549e+00 -4.32818532e-01 -6.37784600e-01
1.36107691e-02 -1.48956430e+00 -2.74592005e-02 -2.22266507e+00
2.69169956e-01 -1.59924790e-01 -5.64247847e-01 6.61155581e-01
-2.50109494e-01 4.46536481e-01 1.71939418e-01 7.51058804e-03
-1.44390821e-01 -4.95633371e-02 1.57886481e+00 -3.69627118e-01
-2.25884616e-01 1.54118389e-01 -4.07900125e-01 6.92846954e-01
1.02877331e+00 -4.42567259e-01 -2.77884662e-01 -6.75023913e-01
-4.88874644e-01 -5.87795814e-03 7.28042543e-01 -1.12891304e+00
2.71941394e-01 1.74023271e-01 7.21532285e-01 -6.38482332e-01
-3.65861878e-02 -6.45316899e-01 -6.04284704e-01 8.64296496e-01
-6.80034637e-01 2.41913088e-02 4.15589362e-01 -1.41180262e-01
1.34640962e-01 -3.34926665e-01 1.17670643e+00 -3.00766736e-01
-4.45204079e-01 7.96971083e-01 -7.24246562e-01 5.27761877e-01
8.32796752e-01 9.99549311e-03 -4.41555798e-01 -2.70793438e-01
-1.34954357e+00 4.59815085e-01 -2.11449146e-01 8.80798399e-02
5.10759652e-01 -1.00434864e+00 -1.01636124e+00 -6.87463954e-02
-3.02304387e-01 2.66534276e-03 3.73133719e-01 1.26242304e+00
-1.26405454e+00 2.93537587e-01 -7.11888731e-01 -6.92582011e-01
-8.30354512e-01 1.05268219e-02 1.11219811e+00 -5.53966999e-01
-1.28401017e+00 9.43313479e-01 2.39068829e-02 -2.36394614e-01
4.08802241e-01 -6.74458027e-01 -4.82693970e-01 -1.29087806e-01
6.49260998e-01 4.39501077e-01 3.61627311e-01 -4.42324340e-01
-3.00862670e-01 2.47725710e-01 -2.78868556e-01 -1.27538711e-01
1.30442095e+00 3.50569218e-01 3.15354876e-02 -4.46639925e-01
1.30997610e+00 -8.12235773e-01 -1.47497487e+00 -2.00411826e-01
1.84520632e-01 9.19673368e-02 6.65227115e-01 -1.44398272e+00
-1.38870358e+00 1.26399922e+00 1.14825249e+00 -3.31542939e-01
7.91695535e-01 -1.90372720e-01 1.07227528e+00 1.85590521e-01
-1.08059913e-01 -8.50170255e-01 -2.63332784e-01 5.90969265e-01
9.43476498e-01 -1.10556126e+00 -2.37192616e-01 -1.71029374e-01
-4.17369723e-01 1.40895295e+00 5.35559118e-01 -4.49758887e-01
8.06068182e-01 4.52575982e-01 4.69708264e-01 -2.85291702e-01
1.08837578e-02 -1.30927473e-01 5.90942264e-01 5.78077018e-01
2.57351995e-01 2.48402998e-01 -3.89863104e-01 5.89608967e-01
5.20256571e-02 1.52005166e-01 5.58853626e-01 9.15613472e-01
-5.65576613e-01 -1.04686272e+00 1.25129163e-01 9.76304829e-01
-7.34911919e-01 -1.97903350e-01 -3.31096411e-01 5.47765613e-01
4.02447954e-02 4.99180019e-01 -2.15633720e-01 2.77766407e-01
6.78511739e-01 2.08058968e-01 6.66119337e-01 -5.14901340e-01
-9.79885042e-01 -1.55980825e-01 -1.46503404e-01 -6.11755729e-01
2.71424595e-02 -7.60695517e-01 -1.62431908e+00 2.79166967e-01
3.40719521e-01 -1.84377700e-01 6.04729593e-01 1.12465751e+00
5.22691943e-02 9.01214242e-01 1.74815878e-01 -1.14467835e+00
-6.06553078e-01 -8.57617021e-01 -2.66442001e-01 4.31879878e-01
4.89645064e-01 -5.69348335e-01 -2.58463155e-02 -1.44798100e-01] | [14.284843444824219, -2.1074516773223877] |
9cc43b46-76eb-4810-aa16-fbe0b7faa944 | reducing-bias-and-increasing-utility-by | 2101.07235 | null | https://arxiv.org/abs/2101.07235v2 | https://arxiv.org/pdf/2101.07235v2.pdf | Reducing bias and increasing utility by federated generative modeling of medical images using a centralized adversary | We introduce FELICIA (FEderated LearnIng with a CentralIzed Adversary) a generative mechanism enabling collaborative learning. In particular, we show how a data owner with limited and biased data could benefit from other data owners while keeping data from all the sources private. This is a common scenario in medical image analysis where privacy legislation prevents data from being shared outside local premises. FELICIA works for a large family of Generative Adversarial Networks (GAN) architectures including vanilla and conditional GANs as demonstrated in this work. We show that by using the FELICIA mechanism, a data owner with limited image samples can generate high-quality synthetic images with high utility while neither data owners has to provide access to its data. The sharing happens solely through a central discriminator that has access limited to synthetic data. Here, utility is defined as classification performance on a real test set. We demonstrate these benefits on several realistic healthcare scenarions using benchmark image datasets (MNIST, CIFAR-10) as well as on medical images for the task of skin lesion classification. With multiple experiments, we show that even in the worst cases, combining FELICIA with real data gracefully achieves performance on par with real data while most results significantly improves the utility. | ['Raymond T Ng', 'Juan Lavista Ferres', 'Christopher West', 'Anthony Ortiz', 'Caleb Robinson', 'Sumit Mukherjee', 'Jean-Francois Rajotte'] | 2021-01-18 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 3.56416970e-01 7.56260216e-01 -7.75754601e-02 -2.41289452e-01
-1.12460649e+00 -7.70971239e-01 5.45238495e-01 -7.42161274e-02
-5.38402557e-01 1.07673335e+00 1.75997857e-02 -2.40507036e-01
-4.02252143e-03 -1.12449384e+00 -8.88093114e-01 -1.10865462e+00
-1.84934527e-01 4.12025273e-01 -4.18435007e-01 4.12745923e-02
-6.58774137e-01 3.54493707e-01 -9.11369324e-01 5.29160202e-01
6.19417191e-01 8.64098012e-01 -3.77387762e-01 9.13082123e-01
4.22259361e-01 1.17181575e+00 -1.06210089e+00 -8.96987915e-01
8.17929029e-01 -5.31702936e-01 -7.12039292e-01 -1.21502148e-03
1.87013805e-01 -7.29824781e-01 -2.35657200e-01 8.49553466e-01
7.98322439e-01 -4.60742533e-01 4.51903343e-01 -1.74236345e+00
-6.93870306e-01 8.04619968e-01 -1.98680475e-01 -3.42523158e-01
-5.06952032e-02 6.63874745e-01 6.09927475e-01 -4.84701581e-02
8.97169352e-01 7.74450958e-01 6.02710485e-01 1.00060964e+00
-1.27730334e+00 -6.79512024e-01 -3.81068021e-01 -4.71618056e-01
-1.21973491e+00 -3.34025741e-01 5.07171154e-01 -2.59116739e-01
1.92116097e-01 6.33142948e-01 3.70978862e-01 1.59603167e+00
2.49313042e-01 8.78006756e-01 1.38328552e+00 -1.52858526e-01
5.62334657e-01 4.67188329e-01 -2.72628158e-01 4.55794930e-01
2.97902197e-01 3.33508104e-01 -3.53618115e-01 -7.68819511e-01
5.08375406e-01 2.42520168e-01 -4.74595219e-01 -4.09360349e-01
-1.10056174e+00 1.00649273e+00 6.58640683e-01 1.57941156e-03
-6.87538564e-01 3.22480351e-01 3.02151948e-01 6.16224587e-01
2.58678079e-01 1.88888952e-01 -1.75910637e-01 3.92061144e-01
-7.64018774e-01 3.06081861e-01 7.88166821e-01 8.68349493e-01
3.49761158e-01 4.77682017e-02 -4.84885365e-01 2.76137292e-01
-8.56678560e-02 7.07662165e-01 5.08259654e-01 -1.01776743e+00
2.96881825e-01 2.63925344e-01 2.04955757e-01 -5.16070485e-01
1.84433013e-01 -4.16288078e-01 -1.27334392e+00 6.86125338e-01
6.31547391e-01 -6.72382474e-01 -9.16512609e-01 2.08217692e+00
3.58766675e-01 -6.85729831e-02 4.95716929e-01 6.86692894e-01
5.35240650e-01 3.32444072e-01 6.63872063e-02 -4.24317792e-02
1.24718165e+00 -5.63365877e-01 -5.77345431e-01 3.74413490e-01
3.65391940e-01 -2.53464133e-01 5.89875400e-01 6.00646317e-01
-1.14437270e+00 -3.01550537e-01 -9.86287773e-01 2.29612008e-01
-3.71498227e-01 -2.86589473e-01 6.49229228e-01 1.17108619e+00
-1.19626105e+00 4.44459677e-01 -8.67297411e-01 -8.60012621e-02
1.17827237e+00 4.45886314e-01 -6.70733631e-01 -2.63015896e-01
-1.21732354e+00 3.07835668e-01 6.82656169e-02 -4.69932318e-01
-1.45614517e+00 -9.93841708e-01 -6.08041942e-01 -6.37629926e-02
1.30658299e-01 -1.32888675e+00 9.41574037e-01 -1.32783151e+00
-1.01496160e+00 9.76657867e-01 6.20864689e-01 -1.15450633e+00
1.35638523e+00 2.12435141e-01 -3.22375745e-01 4.16329764e-02
-8.05791244e-02 7.31497705e-01 8.94669056e-01 -1.41221929e+00
-4.77007955e-01 -4.57186669e-01 2.24259749e-01 -2.26604521e-01
-3.28680694e-01 -4.12223518e-01 2.08466277e-01 -6.43447757e-01
-8.02213073e-01 -9.00026262e-01 -3.11579257e-01 2.88938105e-01
-6.95059836e-01 3.11042696e-01 9.15378690e-01 -5.78014851e-01
4.37501043e-01 -2.13819480e+00 -2.99448818e-01 3.51870626e-01
5.36007941e-01 2.06855208e-01 -1.96251497e-02 3.18339199e-01
1.11299641e-01 3.34221035e-01 -4.53724593e-01 -4.09144312e-01
-5.46401627e-02 3.93697351e-01 -2.73187965e-01 4.63018298e-01
4.00512218e-02 1.10004318e+00 -7.73607910e-01 -3.31625164e-01
-1.03653939e-02 7.53144741e-01 -4.90092665e-01 3.89904886e-01
-5.52981831e-02 5.53189337e-01 -3.41482669e-01 4.89772171e-01
7.89860487e-01 -3.91124159e-01 3.67686123e-01 3.79963443e-02
7.45001137e-01 -5.29487848e-01 -1.01833332e+00 1.53749645e+00
-5.28684914e-01 9.86170769e-02 4.04826373e-01 -6.45323873e-01
5.07488549e-01 6.21084988e-01 5.01149833e-01 -3.81170571e-01
2.21585527e-01 1.02629833e-01 -5.02200387e-02 -2.42836446e-01
-4.02480289e-02 -2.42233530e-01 -3.41515422e-01 7.23255038e-01
1.93082869e-01 1.32384643e-01 -4.59097087e-01 5.26352108e-01
1.53363180e+00 -3.23268354e-01 3.75164330e-01 5.86266490e-03
1.57793522e-01 -1.05047703e-01 3.92444044e-01 1.05041623e+00
-2.92054117e-01 6.58944428e-01 5.25717437e-01 -2.36030146e-01
-1.15889657e+00 -1.45822287e+00 8.30406696e-02 6.11464620e-01
-2.17106521e-01 -5.79579733e-02 -1.07306743e+00 -1.12965298e+00
2.85742432e-01 6.39462948e-01 -1.13398004e+00 -1.81609616e-01
-9.34584811e-02 -7.45619655e-01 9.24325287e-01 4.11596924e-01
5.87762535e-01 -1.06789374e+00 -8.10470819e-01 -5.84600680e-02
3.99649851e-02 -8.42063665e-01 -4.78587985e-01 4.45427448e-02
-3.19346458e-01 -1.11282027e+00 -9.33627963e-01 -1.57227159e-01
7.53585637e-01 -2.33129680e-01 1.08055878e+00 -1.08859204e-01
-6.49531066e-01 5.63320994e-01 -1.92373574e-01 -8.01604748e-01
-1.04841602e+00 -1.44790351e-01 -4.36092615e-01 3.09668630e-01
-1.82166770e-01 -4.84859169e-01 -8.30123127e-01 -1.42722100e-03
-1.28481150e+00 -8.43862668e-02 3.79251540e-01 1.11053324e+00
5.54414093e-01 3.82080190e-02 7.72890389e-01 -1.62778103e+00
6.24601722e-01 -9.11456347e-01 -3.83013010e-01 2.82604486e-01
-6.08139813e-01 -1.54647484e-01 9.62998927e-01 -2.64771283e-01
-9.51703727e-01 2.88862772e-02 1.34064272e-01 -5.40796518e-01
-2.32675374e-01 -1.00546345e-01 -5.31995595e-01 4.61898558e-02
8.71776760e-01 1.91671848e-01 3.74767095e-01 -1.17600210e-01
4.97490197e-01 7.77277887e-01 7.19003499e-01 -3.81125808e-01
6.53144956e-01 9.17498708e-01 6.92559406e-02 -1.82208344e-01
-4.19144034e-01 1.70063853e-01 -1.00350037e-01 1.25643108e-02
8.93352985e-01 -9.49020505e-01 -7.98587739e-01 6.71015203e-01
-7.17196107e-01 -2.84270197e-01 -9.82938349e-01 1.80714309e-01
-6.91064835e-01 -9.27942321e-02 -4.74403948e-01 -6.97758853e-01
-7.53496587e-01 -9.92420495e-01 7.37501979e-01 1.03070565e-01
-7.33684227e-02 -1.03263855e+00 -1.46006380e-04 6.93361461e-01
6.85541689e-01 1.14466608e+00 6.60292864e-01 -9.27189231e-01
-7.40546286e-01 -3.86429131e-01 3.29739034e-01 7.42534935e-01
3.30312282e-01 -3.83355170e-01 -1.30624533e+00 -5.88079989e-01
1.36854753e-01 -6.88381612e-01 5.85124254e-01 1.38015002e-01
1.37141418e+00 -8.12827885e-01 -1.41869232e-01 8.15820932e-01
1.60877836e+00 1.19820036e-01 6.37360990e-01 -2.20190108e-01
4.62622792e-01 4.75828528e-01 1.46503270e-01 5.64631164e-01
1.38899460e-01 1.46560311e-01 6.69966221e-01 -5.21130621e-01
-1.46045640e-01 -4.01506305e-01 1.38695300e-01 -1.42638907e-01
5.33106998e-02 -6.03443384e-01 -4.61842984e-01 7.13252723e-01
-1.69949567e+00 -9.46743846e-01 3.16215217e-01 2.29332376e+00
1.02207959e+00 -2.84713358e-01 1.85573921e-01 -6.54203519e-02
5.24753034e-01 -8.31466839e-02 -9.35334802e-01 -2.36204609e-01
-2.90017515e-01 6.16973937e-01 9.46819305e-01 6.50683716e-02
-8.87324572e-01 2.16541752e-01 5.82318306e+00 7.86238909e-01
-1.04120100e+00 7.23924994e-01 1.25035000e+00 -4.69374537e-01
-4.89652067e-01 -3.77163976e-01 -5.11366613e-02 7.32465982e-01
1.09922349e+00 -4.43013102e-01 3.75210494e-01 1.00210929e+00
-1.93309382e-01 1.94524884e-01 -1.15836239e+00 8.37915838e-01
-8.92419517e-02 -1.62040854e+00 -6.06069639e-02 5.05513430e-01
9.95953679e-01 1.35472775e-01 3.61810327e-01 -7.90319592e-02
1.13246083e+00 -1.41926110e+00 4.88891453e-01 6.48313344e-01
1.08172607e+00 -1.11300123e+00 9.28563952e-01 5.55691361e-01
-2.24065319e-01 1.43209314e-02 -3.47301438e-02 7.40729928e-01
-6.70528505e-03 6.40614688e-01 -9.87279594e-01 8.22976470e-01
5.57169557e-01 -3.49493660e-02 -3.48227412e-01 5.21444380e-01
-1.13425083e-01 6.92884803e-01 -2.27222294e-01 3.84021521e-01
1.11450814e-01 2.08119541e-01 2.68393189e-01 8.60436082e-01
2.24842072e-01 2.58389376e-02 -1.05399340e-01 9.42446172e-01
-7.06230819e-01 -7.09526762e-02 -9.15951788e-01 2.15133920e-01
3.00618470e-01 1.27934492e+00 -1.04437135e-01 -3.47791225e-01
3.23711336e-02 1.18753338e+00 -1.80709176e-02 1.19319238e-01
-8.34274411e-01 -2.34668449e-01 6.33203566e-01 2.82408565e-01
1.45356715e-01 5.41404963e-01 -1.05042262e-02 -8.44989240e-01
-7.61228055e-02 -1.14042699e+00 7.25790858e-01 -6.76749051e-01
-1.70037329e+00 6.63501859e-01 -2.75965512e-01 -9.86295164e-01
-5.35243869e-01 -1.94780141e-01 -5.47365546e-01 1.11477911e+00
-1.20960379e+00 -1.67027676e+00 -3.26127946e-01 1.15631318e+00
-1.90076619e-01 -4.51775819e-01 1.12733507e+00 -1.41391736e-02
-9.49373096e-02 1.15940273e+00 2.94094324e-01 4.51309919e-01
5.88717103e-01 -1.25344753e+00 2.50587285e-01 6.75208628e-01
1.39642417e-01 4.89007086e-01 3.69796574e-01 -5.02944946e-01
-1.23232853e+00 -1.45196581e+00 2.10972786e-01 -4.50848669e-01
1.64559856e-01 -5.16006112e-01 -6.88646138e-01 6.82159841e-01
6.47028506e-01 6.03248358e-01 9.68681097e-01 -5.83543360e-01
-3.52395982e-01 -2.45842203e-01 -2.19627690e+00 1.29031494e-01
7.51642168e-01 -4.21363711e-01 5.85590415e-02 5.25023580e-01
5.64744532e-01 -1.74143195e-01 -1.14222455e+00 9.87079367e-02
2.86673814e-01 -1.03309369e+00 6.99849427e-01 -7.95783222e-01
2.30158687e-01 -1.04800902e-01 -2.00287178e-01 -1.40007269e+00
2.06173509e-01 -9.71112370e-01 7.09978268e-02 1.29769254e+00
1.71724617e-01 -9.56050754e-01 9.66622770e-01 8.27139556e-01
6.24262750e-01 -5.38494825e-01 -1.04784250e+00 -7.87725151e-01
3.81771684e-01 -1.34950697e-01 1.04251933e+00 1.08746243e+00
-5.26915669e-01 -3.40902746e-01 -6.11919165e-01 1.17512003e-01
1.13426518e+00 -1.34311721e-01 1.03057349e+00 -5.95014870e-01
-6.60441756e-01 1.78828046e-01 -4.53794897e-01 -6.74688593e-02
3.65721658e-02 -1.06039083e+00 -3.72145951e-01 -1.14913225e+00
3.03329766e-01 -6.33413851e-01 -3.88678938e-01 7.77011871e-01
1.15332976e-01 3.57716471e-01 3.58024657e-01 1.07325278e-01
-8.65837932e-02 5.83087318e-02 1.23093712e+00 -4.27509069e-01
3.07976782e-01 2.26445630e-01 -1.06145990e+00 2.46957973e-01
7.62816727e-01 -5.97680390e-01 -7.59733915e-01 -4.35853675e-02
-1.37456089e-01 3.44115525e-01 9.89905775e-01 -9.34990823e-01
1.19727239e-01 -7.11853579e-02 5.06816804e-01 -9.96470451e-02
6.89350590e-02 -9.65539277e-01 9.43948150e-01 6.66363597e-01
-5.21282017e-01 -4.24288779e-01 -3.06730211e-01 6.59240246e-01
3.08845714e-02 -3.17343773e-04 9.83891368e-01 -3.86120617e-01
-7.31270388e-02 4.22862828e-01 1.20202780e-01 5.55834398e-02
1.61357415e+00 2.02032059e-01 -5.76719463e-01 -6.43361270e-01
-1.01874757e+00 1.61149278e-01 7.67259955e-01 6.90787584e-02
4.67026800e-01 -1.20560110e+00 -1.14893794e+00 4.21426028e-01
1.02090329e-01 1.51625663e-01 4.55557376e-01 4.04495478e-01
-2.75332868e-01 -7.16936663e-02 -3.26732129e-01 -4.39104646e-01
-1.26178789e+00 7.73034513e-01 6.34091437e-01 -4.97016400e-01
-4.31115687e-01 6.68641746e-01 3.39331001e-01 -3.36800963e-01
2.42943000e-02 2.29522400e-02 5.73880494e-01 -1.70962721e-01
5.47898471e-01 1.93822324e-01 2.18512937e-01 -2.54258096e-01
-1.80011287e-01 -3.94628853e-01 -1.03896707e-01 -8.90259817e-02
1.37167311e+00 2.30717868e-01 2.31908172e-01 -9.19722170e-02
1.29282641e+00 1.53745979e-01 -1.37455404e+00 -3.43516357e-02
-8.89495850e-01 -5.28304577e-01 -1.15634874e-01 -1.44107568e+00
-1.69619918e+00 5.13842106e-01 8.75468791e-01 2.56481767e-01
1.31416774e+00 -3.24440487e-02 7.40701854e-01 -5.95245510e-02
8.00900102e-01 -6.42545044e-01 -2.27196559e-01 -6.62755072e-01
6.85500622e-01 -1.23907602e+00 -1.52249947e-01 9.19763092e-03
-1.00772953e+00 4.85493571e-01 1.96378633e-01 -1.53407276e-01
6.62609696e-01 6.75479412e-01 2.88510352e-01 -1.28880665e-02
-8.44397902e-01 3.04382533e-01 -3.24238688e-01 1.09000742e+00
-2.20156580e-01 4.93569911e-01 7.97363743e-02 7.26152718e-01
-1.79351777e-01 2.50888199e-01 7.64468014e-01 1.03568685e+00
4.07915413e-01 -1.37601674e+00 -4.25681233e-01 5.70840895e-01
-1.00507593e+00 3.83401364e-02 -2.58228540e-01 8.35676610e-01
5.92768490e-01 7.36604512e-01 2.39409562e-02 -9.95737035e-03
4.93962988e-02 -1.19700342e-01 3.15401316e-01 -2.80541122e-01
-1.18180311e+00 -2.65326589e-01 -1.78792253e-01 -5.60217619e-01
-4.63462770e-01 -5.68955004e-01 -1.04094398e+00 -4.32190359e-01
7.51087666e-02 4.26825322e-02 7.30159521e-01 3.18354279e-01
5.58996379e-01 4.65731233e-01 1.08158565e+00 1.84482947e-01
-8.73695493e-01 -4.70408618e-01 -8.09661984e-01 8.79095495e-01
3.89537394e-01 1.84241518e-01 -3.31085891e-01 2.97490895e-01] | [6.0956573486328125, 6.848149299621582] |
13a0850a-618c-472e-a166-8d2be8513547 | model-patching-closing-the-subgroup | 2008.06775 | null | https://arxiv.org/abs/2008.06775v1 | https://arxiv.org/pdf/2008.06775v1.pdf | Model Patching: Closing the Subgroup Performance Gap with Data Augmentation | Classifiers in machine learning are often brittle when deployed. Particularly concerning are models with inconsistent performance on specific subgroups of a class, e.g., exhibiting disparities in skin cancer classification in the presence or absence of a spurious bandage. To mitigate these performance differences, we introduce model patching, a two-stage framework for improving robustness that encourages the model to be invariant to subgroup differences, and focus on class information shared by subgroups. Model patching first models subgroup features within a class and learns semantic transformations between them, and then trains a classifier with data augmentations that deliberately manipulate subgroup features. We instantiate model patching with CAMEL, which (1) uses a CycleGAN to learn the intra-class, inter-subgroup augmentations, and (2) balances subgroup performance using a theoretically-motivated subgroup consistency regularizer, accompanied by a new robust objective. We demonstrate CAMEL's effectiveness on 3 benchmark datasets, with reductions in robust error of up to 33% relative to the best baseline. Lastly, CAMEL successfully patches a model that fails due to spurious features on a real-world skin cancer dataset. | ['Christopher Ré', 'Albert Gu', 'Yixuan Li', 'Karan Goel'] | 2020-08-15 | null | https://openreview.net/forum?id=9YlaeLfuhJF | https://openreview.net/pdf?id=9YlaeLfuhJF | iclr-2021-1 | ['skin-cancer-classification'] | ['medical'] | [ 1.02967036e+00 6.66731656e-01 -6.63131535e-01 -4.47934806e-01
-1.05871773e+00 -6.17445469e-01 8.60920846e-01 2.56438851e-01
-1.55452579e-01 5.49233854e-01 3.42284620e-01 -3.04267406e-01
-1.04223341e-01 -6.18891120e-01 -9.51489806e-01 -9.75564182e-01
1.37817189e-01 3.54650728e-02 2.64311880e-01 -7.59766772e-02
5.21316603e-02 2.20204726e-01 -1.34711730e+00 7.75408208e-01
8.01611543e-01 9.52687562e-01 -4.81560469e-01 5.26894212e-01
2.91120440e-01 7.06600368e-01 -3.28930229e-01 -2.48808205e-01
4.11682636e-01 -6.30740583e-01 -8.32200646e-01 4.11582053e-01
1.06579828e+00 1.28953636e-01 -1.73256814e-01 8.92137408e-01
2.99322903e-01 -3.00316095e-01 6.08716130e-01 -1.51378536e+00
-3.97767782e-01 7.35083938e-01 -5.65935194e-01 -2.12530285e-01
-4.46164459e-02 4.94988382e-01 1.03802454e+00 -2.81910032e-01
7.24988222e-01 1.14209926e+00 1.30872095e+00 1.12479806e+00
-1.74798918e+00 -7.13463664e-01 2.85521090e-01 -2.46838629e-01
-1.24362373e+00 -5.73307276e-01 4.81707513e-01 -4.73694116e-01
6.64062023e-01 7.87958205e-01 5.61337709e-01 1.46657515e+00
2.86327273e-01 5.52877665e-01 1.29826951e+00 -3.26616406e-01
2.76072562e-01 1.03121348e-01 4.69722390e-01 7.89270103e-01
1.54649109e-01 1.69613883e-01 -4.81310278e-01 -4.26416814e-01
2.29187384e-01 -1.14649355e-01 -2.61722893e-01 -4.49611098e-01
-1.15325403e+00 8.33796382e-01 5.64269185e-01 -6.43156916e-02
-7.54725486e-02 2.41747454e-01 2.23273084e-01 1.65347174e-01
4.71096128e-01 5.81791997e-01 -5.58406770e-01 4.53072220e-01
-7.69137263e-01 3.03186536e-01 5.64565122e-01 7.59286642e-01
6.31739259e-01 -5.15654445e-01 -6.15505397e-01 7.83933520e-01
1.63320675e-02 3.67988534e-02 4.03654456e-01 -7.29874671e-01
1.13529913e-01 6.86489940e-01 -4.23065335e-01 -6.51766896e-01
-6.60136700e-01 -5.55512488e-01 -1.00168300e+00 1.37210060e-02
4.72116202e-01 6.00312278e-02 -1.29768395e+00 2.23993230e+00
4.74381983e-01 4.68241692e-01 -5.28008454e-02 4.10397500e-01
8.99564862e-01 -4.00637798e-02 3.24247688e-01 1.90621331e-01
1.37501478e+00 -1.03165936e+00 -4.05001938e-01 -4.80821311e-01
1.00322783e+00 -4.27397370e-01 1.06355190e+00 2.09357873e-01
-7.62215376e-01 -1.26181930e-01 -1.02377403e+00 -4.01945859e-02
-3.92989248e-01 -1.10207148e-01 8.27802777e-01 8.14794838e-01
-1.06343961e+00 5.96993446e-01 -9.28359330e-01 -4.53687191e-01
8.91425729e-01 5.60464621e-01 -8.16574216e-01 -6.17608652e-02
-8.45046818e-01 7.36768365e-01 2.78714776e-01 -7.82165304e-02
-6.13088906e-01 -1.42944789e+00 -1.01533759e+00 -2.14831516e-01
3.29033136e-01 -9.71825123e-01 1.03270614e+00 -1.30347276e+00
-1.18838429e+00 1.01294541e+00 7.83467367e-02 -5.07794142e-01
7.68301308e-01 2.75460422e-01 -3.44403446e-01 -2.26064444e-01
1.43353805e-01 8.59636366e-01 8.97767842e-01 -1.24274170e+00
-5.98083198e-01 -3.94817144e-01 -2.17446283e-01 -8.83028805e-02
-4.62265611e-01 -2.90892959e-01 -3.97382230e-01 -6.99539721e-01
1.79011032e-01 -1.22357559e+00 -4.10505980e-01 2.34003171e-01
-9.77964938e-01 -1.96202099e-02 7.87399292e-01 -5.87696612e-01
1.05456555e+00 -2.06586385e+00 5.30860573e-02 5.18419504e-01
3.53164464e-01 3.76149267e-02 -5.07350326e-01 -6.30453508e-03
-4.52339172e-01 4.17109340e-01 -6.89063311e-01 -5.45688748e-01
-1.36524200e-01 2.79101759e-01 -1.11834139e-01 7.22385645e-01
6.07385099e-01 7.49470055e-01 -5.99710286e-01 -2.82013357e-01
-1.01015210e-01 4.32544827e-01 -1.03062940e+00 -1.28704369e-01
-2.35449284e-01 3.34352851e-01 -5.33456802e-02 7.80073702e-01
7.60156155e-01 -3.45156550e-01 1.34028390e-01 -3.94023538e-01
2.21408620e-01 2.14691579e-01 -9.05743778e-01 1.50429332e+00
-1.35550708e-01 3.55746299e-01 1.14634611e-01 -6.31008685e-01
4.48481113e-01 -5.60916364e-02 4.94341612e-01 -2.80063808e-01
-7.78648928e-02 2.40255892e-02 4.04183827e-02 -3.06316227e-01
1.70232937e-01 -9.66260135e-02 1.33963386e-02 3.09628934e-01
1.81625765e-02 -2.47506768e-01 -5.38886748e-02 7.00395405e-02
1.56141412e+00 -2.26535022e-01 2.66942084e-01 -4.12967592e-01
1.30947739e-01 4.35481481e-02 7.42812932e-01 9.66749549e-01
-2.52792895e-01 9.95430946e-01 7.77245283e-01 -3.55036736e-01
-9.46520269e-01 -8.47984254e-01 -5.03757834e-01 9.89558935e-01
7.81591758e-02 -3.09112310e-01 -8.19289148e-01 -1.28430116e+00
4.61012036e-01 5.42114139e-01 -1.45918834e+00 -6.25112414e-01
-2.12411717e-01 -1.35988438e+00 7.92036772e-01 5.03619671e-01
5.01782477e-01 -5.18672884e-01 -7.97573701e-02 -2.31370240e-01
2.05923125e-01 -9.99998987e-01 -6.31400824e-01 5.11523068e-01
-5.59671104e-01 -1.52745128e+00 -1.51893690e-01 -7.25404382e-01
1.08137643e+00 1.80329066e-02 1.06323755e+00 4.47529048e-01
-5.94798326e-01 9.79789346e-02 -1.67076007e-01 -3.05937439e-01
-6.42627060e-01 5.16331494e-01 -2.97314286e-01 2.38572910e-01
1.74477592e-01 -8.84888619e-02 -4.37637120e-01 4.10391927e-01
-1.15364516e+00 3.42211336e-01 4.07815367e-01 1.25996554e+00
5.46333373e-01 -2.45411307e-01 4.75388438e-01 -1.48471248e+00
2.00387016e-01 -6.01488948e-01 -2.05216333e-01 3.57411027e-01
-5.29436171e-01 -1.06484015e-02 2.57630199e-01 -5.51417530e-01
-7.05800891e-01 3.25987428e-01 2.90331673e-02 -8.30197856e-02
-5.23807406e-02 1.21863022e-01 -3.17480892e-01 -2.90860087e-01
1.11989021e+00 -5.21651693e-02 4.97557104e-01 -2.56462395e-01
2.44030952e-01 5.20278931e-01 4.54474777e-01 -4.47831184e-01
8.39660525e-01 6.24832034e-01 1.12879410e-01 -7.24861860e-01
-1.02145529e+00 -2.07485840e-01 -5.65164566e-01 2.60792524e-01
6.55693829e-01 -9.37798858e-01 -2.93938309e-01 8.03362489e-01
-6.73855484e-01 -8.86523962e-01 -4.75930393e-01 7.80333281e-02
-3.55736077e-01 1.01115152e-01 -4.51412797e-01 -1.29326403e-01
-2.03908741e-01 -1.17710125e+00 1.20289135e+00 1.27945289e-01
-4.33309674e-01 -1.12920833e+00 -1.41758457e-01 4.03475344e-01
4.44261432e-01 6.78443253e-01 9.82329667e-01 -8.65464389e-01
-1.17787138e-01 -2.82241374e-01 -1.92955047e-01 4.14693207e-01
5.75446665e-01 3.33152533e-01 -1.27396882e+00 -5.29591203e-01
-5.24907947e-01 -3.65814894e-01 1.21265268e+00 2.11901277e-01
1.55769372e+00 -5.08674562e-01 -7.18995750e-01 1.11509848e+00
1.00657165e+00 -3.77449363e-01 5.60188234e-01 2.33062387e-01
8.18532467e-01 6.23581946e-01 4.14624140e-02 3.38345841e-02
3.09653670e-01 6.13729179e-01 4.95767415e-01 -4.79649305e-01
-5.19940078e-01 -2.72440135e-01 -2.41094157e-02 9.06985998e-02
4.69528526e-01 1.42277002e-01 -7.61795640e-01 4.76558536e-01
-1.63477099e+00 -7.29375780e-01 3.66056822e-02 2.01523757e+00
1.18468809e+00 1.04702838e-01 1.37676552e-01 -1.76549389e-03
6.44169450e-01 6.04862534e-02 -8.33888948e-01 3.06226066e-05
-4.69445914e-01 1.98182628e-01 9.26921666e-01 5.25326729e-01
-1.42593455e+00 8.80277812e-01 6.84739494e+00 7.20450997e-01
-1.08945298e+00 -9.59805548e-02 1.03475070e+00 -5.04013561e-02
-5.83329499e-01 -3.90745476e-02 -7.47683764e-01 4.82281417e-01
4.14080054e-01 2.58503675e-01 3.02100122e-01 6.01573944e-01
-2.34262913e-01 1.66961804e-01 -1.43019724e+00 4.55703259e-01
1.72449410e-01 -1.55549133e+00 4.94398922e-02 2.00088158e-01
1.09695172e+00 -9.06808749e-02 3.42238754e-01 1.33027196e-01
4.54998136e-01 -1.39243650e+00 5.36161542e-01 3.13273489e-01
7.22838938e-01 -5.62175512e-01 5.41881800e-01 6.00666367e-02
-5.63416183e-01 -2.17413813e-01 1.83291640e-02 3.18138093e-01
-5.92212439e-01 5.36472678e-01 -1.09470928e+00 1.83778092e-01
5.46241343e-01 7.71316409e-01 -1.08393049e+00 9.80264187e-01
-1.34469315e-01 6.71505988e-01 -4.64962810e-01 2.21259370e-01
1.58891454e-02 2.62268513e-01 2.63120860e-01 1.15086269e+00
-4.16672193e-02 -2.30572239e-01 4.39100452e-02 6.77823365e-01
-1.84555128e-01 -2.13754043e-01 -4.35470164e-01 2.73960650e-01
5.94229460e-01 1.16547060e+00 -4.62055892e-01 -1.29053473e-01
-9.94988531e-02 8.15528214e-01 2.56623268e-01 1.01738088e-01
-7.80039251e-01 -7.22976923e-02 9.73779917e-01 2.92262018e-01
-2.59582214e-02 5.01507282e-01 -6.07219517e-01 -9.65001166e-01
-1.08229667e-01 -1.30471504e+00 7.78184772e-01 -2.68338650e-01
-1.50091410e+00 3.33922267e-01 -1.81338906e-01 -8.80709767e-01
4.92252931e-02 -5.60609281e-01 -5.95869958e-01 5.96764803e-01
-1.43904400e+00 -1.76021159e+00 -5.51034689e-01 5.98649621e-01
1.29728913e-01 -4.97714542e-02 9.03281510e-01 -8.17582905e-02
-9.23222899e-01 1.33210790e+00 -2.47907236e-01 4.54673953e-02
1.00878572e+00 -1.44224060e+00 5.14097750e-01 6.59204185e-01
-3.41017663e-01 4.64201927e-01 6.91037357e-01 -5.21454751e-01
-1.21770096e+00 -1.56185639e+00 6.94041491e-01 -8.14096808e-01
6.63198054e-01 -5.47355413e-01 -9.02820349e-01 9.97238457e-01
-2.67091066e-01 5.93587041e-01 1.01613712e+00 1.88250273e-01
-9.88845766e-01 -2.40676701e-01 -1.59181547e+00 8.11226189e-01
1.25087357e+00 -4.54317123e-01 -2.14854047e-01 5.95674336e-01
7.40488470e-01 -5.67576826e-01 -8.29478443e-01 6.82231963e-01
6.74602449e-01 -8.42323542e-01 7.88942218e-01 -1.08542681e+00
3.45720977e-01 -1.71583399e-01 -2.79160768e-01 -1.47490430e+00
-2.66127378e-01 -7.48147488e-01 2.66427934e-01 1.13151789e+00
5.15791357e-01 -8.36888433e-01 1.12124574e+00 7.99467266e-01
-1.93043172e-01 -7.26574063e-01 -9.05358970e-01 -5.88578820e-01
4.39399391e-01 -2.53074557e-01 8.79929960e-01 1.24201000e+00
-8.88658911e-02 -6.00567795e-02 -1.78750560e-01 2.98184901e-01
6.63097560e-01 -3.21124375e-01 9.36400354e-01 -1.04761064e+00
-3.47987652e-01 -7.01544404e-01 -4.65204775e-01 -2.43229941e-01
3.46398294e-01 -1.13637006e+00 3.95155326e-02 -9.90152359e-01
4.32894796e-01 -8.05077016e-01 -2.77658641e-01 1.22355163e+00
-4.94695842e-01 6.07366800e-01 7.97337443e-02 -1.24998689e-01
-2.99375862e-01 8.56340453e-02 1.11020255e+00 -4.62009996e-01
-1.75367042e-01 -1.07327430e-03 -1.36906612e+00 6.15828097e-01
7.57702887e-01 -4.18272913e-01 -1.35118783e-01 -2.22589001e-01
-3.67182270e-02 -5.52479327e-01 6.39766097e-01 -9.76060629e-01
2.30631694e-01 -2.61766374e-01 4.78959411e-01 7.15678185e-02
1.47586778e-01 -7.14640200e-01 2.83075064e-01 6.47672951e-01
-8.74863684e-01 -4.00745273e-01 3.27042103e-01 4.33546394e-01
1.31874889e-01 2.57950038e-01 1.20049727e+00 3.24406832e-01
-2.95217365e-01 3.64007831e-01 9.59052145e-02 5.99124841e-02
1.19655573e+00 -5.36616482e-02 -7.59838641e-01 1.84874609e-01
-6.13524735e-01 2.30483577e-01 7.35561848e-01 4.61025953e-01
1.53899128e-02 -1.33187020e+00 -8.40697587e-01 7.70124674e-01
6.09972179e-01 1.78700820e-01 3.25831681e-01 7.73416519e-01
-2.53337562e-01 -5.74223958e-02 -4.96658981e-02 -8.89710307e-01
-1.60951674e+00 2.47585312e-01 6.95474446e-01 -2.31513277e-01
-3.57519627e-01 1.25179720e+00 2.09193096e-01 -8.04327607e-01
2.61854112e-01 -5.18591166e-01 2.99532950e-01 4.19024192e-03
4.94134098e-01 4.71710749e-02 4.07496214e-01 -5.17293572e-01
-5.56485057e-01 5.76871216e-01 -3.92182529e-01 3.82352859e-01
1.15171587e+00 4.41229522e-01 -1.41396582e-01 -7.24056587e-02
1.32832289e+00 -4.66201268e-02 -1.58418679e+00 -3.24846417e-01
-1.64903048e-02 -2.98981458e-01 6.53382763e-02 -1.16113925e+00
-1.07063258e+00 2.15747908e-01 5.62449872e-01 2.65457928e-02
1.14034295e+00 -4.79997583e-02 4.46952105e-01 5.03600091e-02
-5.64030334e-02 -9.13113832e-01 -1.09498620e-01 2.13344410e-01
7.01353371e-01 -1.35014141e+00 1.50821730e-01 -7.46916354e-01
-4.31428224e-01 7.67742395e-01 6.12971663e-01 -1.76480282e-02
7.75812209e-01 3.88242811e-01 1.10961810e-01 -2.24243924e-02
-1.02464318e+00 1.59855247e-01 4.72103626e-01 7.32696056e-01
1.77514598e-01 2.01020449e-01 -6.57342598e-02 6.59029305e-01
-4.41264808e-01 -9.26644504e-02 2.03416705e-01 8.59321535e-01
-1.44537147e-02 -1.08898985e+00 -3.69109929e-01 8.26890051e-01
-4.29520100e-01 -5.99421896e-02 -6.69570625e-01 1.02242696e+00
4.30145949e-01 8.43363285e-01 2.68639833e-01 -6.82144761e-01
4.13272619e-01 -7.66258910e-02 4.62377518e-01 -6.98245943e-01
-1.03754151e+00 -2.47022942e-01 1.88966468e-01 -7.37931430e-01
-8.16825554e-02 -7.75956035e-01 -8.35167050e-01 -8.36218297e-02
-1.52467877e-01 -1.67290822e-01 6.01261139e-01 8.35556924e-01
4.44941521e-01 6.76868737e-01 6.95239842e-01 -4.39308226e-01
-8.19595814e-01 -8.40395510e-01 -4.12226707e-01 7.72550762e-01
7.02948749e-01 -4.04721290e-01 -7.18801618e-01 -4.42052893e-02] | [15.131891250610352, -2.5793685913085938] |
1efe5d87-ed96-42ba-8820-0931b6f95573 | improved-pronunciation-prediction-accuracy | null | null | https://aclanthology.org/2021.sigmorphon-1.24 | https://aclanthology.org/2021.sigmorphon-1.24.pdf | Improved pronunciation prediction accuracy using morphology | Pronunciation lexicons and prediction models are a key component in several speech synthesis and recognition systems. We know that morphologically related words typically follow a fixed pattern of pronunciation which can be described by language-specific paradigms. In this work we explore how deep recurrent neural networks can be used to automatically learn and exploit this pattern to improve the pronunciation prediction quality of words related by morphological inflection. We propose two novel approaches for supplying morphological information, using the word’s morphological class and its lemma, which are typically annotated in standard lexicons. We report improvements across a number of European languages with varying degrees of phonological and morphological complexity, and two language families, with greater improvements for languages where the pronunciation prediction task is inherently more challenging. We also observe that combining bidirectional LSTM networks with attention mechanisms is an effective neural approach for the computational problem considered, across languages. Our approach seems particularly beneficial in the low resource setting, both by itself and in conjunction with transfer learning. | ['Antoine Bruguier', 'Neha Chaudhari', 'Saumya Sahai', 'Dravyansh Sharma'] | null | null | null | null | acl-sigmorphon-2021-8 | ['morphological-inflection'] | ['natural-language-processing'] | [ 2.96527654e-01 -1.10231541e-01 -2.84879118e-01 -2.24389464e-01
-6.77257299e-01 -9.12258387e-01 4.67293203e-01 2.57656068e-01
-5.80520272e-01 6.03104651e-01 6.16015375e-01 -6.94279790e-01
7.73282303e-03 -7.52540767e-01 -5.92932522e-01 -3.90180349e-01
2.20246702e-01 5.57282686e-01 -1.94806293e-01 -4.60526049e-01
1.73839867e-01 6.00794017e-01 -1.46740079e+00 4.10446525e-01
6.72584713e-01 8.68256152e-01 5.58732808e-01 6.43139064e-01
-3.58012199e-01 7.40693212e-01 -5.56765139e-01 -6.66164577e-01
1.08108193e-01 -2.85423130e-01 -9.66132224e-01 -1.39210552e-01
4.33071673e-01 1.92863077e-01 -2.14745943e-02 9.10483837e-01
4.70275730e-01 3.84137869e-01 6.00976586e-01 -4.19864565e-01
-7.84152508e-01 1.13521540e+00 7.65094087e-02 4.88697469e-01
1.12501658e-01 2.95932978e-01 1.38249648e+00 -1.16246760e+00
5.86433649e-01 1.19112837e+00 6.64501131e-01 6.76167607e-01
-1.38051009e+00 -4.55211252e-01 2.40987867e-01 4.03756499e-01
-1.23862612e+00 -7.24469900e-01 7.27433324e-01 -3.71022761e-01
1.61691165e+00 1.74683243e-01 6.17409468e-01 1.07548475e+00
-3.74351703e-02 7.15121150e-01 1.12437832e+00 -7.22484410e-01
-1.04183979e-01 2.53622949e-01 4.42143306e-02 4.92293179e-01
-8.82350504e-02 2.22657800e-01 -6.41358495e-01 3.05544674e-01
5.71539283e-01 -4.58920926e-01 -2.37674713e-01 5.03695570e-02
-1.28136051e+00 8.44087899e-01 2.11629868e-01 6.11788809e-01
-2.99098283e-01 5.51270768e-02 6.36071146e-01 4.41845328e-01
5.46680689e-01 7.19618201e-01 -1.03365707e+00 -2.62091726e-01
-7.01677322e-01 -1.35413408e-01 7.08037615e-01 8.62476289e-01
5.49905717e-01 5.93294084e-01 -9.45417807e-02 1.42899120e+00
2.41847914e-02 2.33808294e-01 9.13532495e-01 -5.58144212e-01
4.07149822e-01 2.11444363e-01 -3.68888319e-01 -6.03162289e-01
-2.87682354e-01 -5.48376024e-01 -6.03001356e-01 -1.32398590e-01
5.62938631e-01 -4.37284037e-02 -8.10557127e-01 2.22483087e+00
-8.36879238e-02 -3.56276184e-02 4.03692663e-01 3.09259802e-01
7.69014537e-01 9.11863148e-01 3.55782151e-01 -2.87260830e-01
1.38123059e+00 -9.37844038e-01 -7.11457312e-01 -5.23025990e-01
8.13205183e-01 -9.32300389e-01 1.48512006e+00 4.45550770e-01
-1.43094313e+00 -6.33289397e-01 -1.00593877e+00 -2.91189969e-01
-7.75457323e-01 3.84991497e-01 4.93595660e-01 7.41374373e-01
-1.24886847e+00 7.77056098e-01 -4.56739277e-01 -4.00474638e-01
1.64130166e-01 4.73530233e-01 -1.95144922e-01 3.98880243e-01
-1.18613100e+00 1.47494543e+00 6.25028849e-01 -1.16410963e-01
-5.15689492e-01 -7.50487506e-01 -9.28024888e-01 1.46938130e-01
4.66206064e-03 -4.33015019e-01 1.35529816e+00 -1.12665522e+00
-1.68603837e+00 1.05740809e+00 -1.63036391e-01 -5.86915016e-01
-6.40846565e-02 -4.42754524e-03 -5.38291693e-01 -2.79139489e-01
-3.12541813e-01 7.97276437e-01 8.09183955e-01 -8.04556429e-01
-6.66423976e-01 -2.88388394e-02 -3.32412332e-01 3.05707783e-01
-4.85219657e-01 4.33166593e-01 1.06548682e-01 -1.02471280e+00
-1.98335081e-01 -7.58947432e-01 -1.46126494e-01 -5.13826549e-01
-3.04807186e-01 -6.65068150e-01 2.09849983e-01 -7.89270699e-01
1.21849263e+00 -1.91181600e+00 3.20910305e-01 5.00234812e-02
-3.44505906e-01 5.54712892e-01 -8.68084356e-02 3.32507312e-01
-1.01276018e-01 2.82593846e-01 -3.08473229e-01 -4.54457253e-01
6.03234284e-02 4.02121365e-01 -5.66511691e-01 1.24284841e-01
6.28638029e-01 1.06713903e+00 -7.36774981e-01 -1.27429157e-01
1.24057464e-01 4.77789193e-01 -5.13927579e-01 1.56223312e-01
-3.62046063e-01 3.03727120e-01 4.29713011e-01 3.79511565e-01
5.79056442e-02 1.65651277e-01 3.41210395e-01 -2.10766029e-02
-3.33557636e-01 1.44695985e+00 -8.91453266e-01 1.56622648e+00
-1.03725898e+00 6.35310888e-01 -1.65031359e-01 -1.00976241e+00
1.04799032e+00 5.42027414e-01 -2.37750232e-01 -4.74305958e-01
1.92115933e-01 5.25070786e-01 6.05640888e-01 -4.65148874e-02
6.84774876e-01 -5.36637783e-01 -2.22456437e-02 4.14924085e-01
4.79797810e-01 -2.95255005e-01 1.97463185e-01 -4.28114176e-01
7.40531921e-01 3.98432575e-02 5.76123416e-01 -5.65393209e-01
5.57186663e-01 -1.24123633e-01 3.64705443e-01 4.99712378e-01
8.26469883e-02 4.07792747e-01 1.47487268e-01 -3.37920994e-01
-1.11704290e+00 -9.70530033e-01 -2.87328809e-01 1.47610259e+00
-5.87787271e-01 -3.58731627e-01 -6.63847685e-01 -2.85521656e-01
-3.57837915e-01 1.07513809e+00 -4.33855206e-01 -3.92807692e-01
-1.12426710e+00 -3.62866759e-01 6.60954714e-01 8.02971482e-01
-1.34322211e-01 -1.87401402e+00 -1.94532201e-01 5.47835946e-01
9.20296684e-02 -9.78981197e-01 -5.38985908e-01 7.64729142e-01
-6.77377641e-01 -3.87083203e-01 -6.12780452e-01 -1.22638845e+00
9.71486270e-02 -2.05796152e-01 1.39202511e+00 2.76139155e-02
1.81565851e-01 -4.04774807e-02 -2.23345563e-01 -5.09013772e-01
-8.54325056e-01 5.43414831e-01 3.04540843e-01 -6.05313480e-02
4.49535131e-01 -5.79729676e-01 -7.23003745e-02 -1.29409596e-01
-6.72516346e-01 -8.41430128e-02 5.20737290e-01 8.80291581e-01
6.56219006e-01 -3.78148913e-01 7.01805651e-01 -7.81314671e-01
5.91789067e-01 -5.81982553e-01 -4.43917274e-01 1.43264487e-01
-3.34674329e-01 2.46578291e-01 1.03177524e+00 -6.96753323e-01
-9.45312798e-01 9.14165676e-02 -6.76553488e-01 -1.19388230e-01
-4.04150248e-01 6.14585638e-01 -4.33442384e-01 1.89543799e-01
4.79310811e-01 2.17340469e-01 -1.38065323e-01 -7.48973072e-01
5.08902490e-01 6.08434916e-01 4.73492742e-01 -5.74616849e-01
5.13175905e-01 -1.55309245e-01 -2.48691559e-01 -9.93428290e-01
-8.14287424e-01 -2.71480858e-01 -7.67817378e-01 1.91127434e-01
8.65921497e-01 -5.89163721e-01 -4.89683598e-01 3.51338118e-01
-1.49188387e+00 -4.54611927e-01 -6.12474680e-01 4.92745280e-01
-7.73973227e-01 -1.02464460e-01 -8.41279626e-01 -4.78413075e-01
-3.78172904e-01 -1.11849093e+00 7.29648590e-01 -1.77660882e-02
-5.88038325e-01 -1.19451404e+00 -1.19092157e-02 -2.12674424e-01
6.55838728e-01 -5.11588693e-01 1.53230679e+00 -1.09666753e+00
-2.60837585e-01 3.90090555e-01 3.63371000e-02 6.02466762e-01
3.91698271e-01 -9.10733640e-02 -1.04947507e+00 -3.40895317e-02
-5.42111583e-02 -2.66192466e-01 8.68808270e-01 2.54458338e-01
9.67749655e-01 -4.66176391e-01 1.01884402e-01 6.52451158e-01
1.20557117e+00 2.43532747e-01 4.62550700e-01 1.54211462e-01
6.98443472e-01 1.05270112e+00 1.88095927e-01 -7.99680874e-02
2.78380096e-01 7.48740256e-01 -7.11658821e-02 -1.28608104e-03
-5.22192597e-01 -4.32202458e-01 6.72664642e-01 1.64680600e+00
2.07400858e-01 -3.39608163e-01 -1.02982473e+00 8.62525463e-01
-1.32415152e+00 -7.92859554e-01 1.12959862e-01 2.12256646e+00
1.31510043e+00 1.56053826e-01 1.37368754e-01 3.07880998e-01
6.65370464e-01 5.68080693e-02 -1.94051862e-01 -1.18034410e+00
-4.29688573e-01 9.22416389e-01 3.38228077e-01 5.87868869e-01
-7.78868198e-01 1.64086306e+00 6.71567583e+00 1.07483852e+00
-1.34292543e+00 2.46885106e-01 5.77720821e-01 3.14979106e-02
-3.79568666e-01 -1.51209563e-01 -1.07905126e+00 1.69252649e-01
1.55095685e+00 -4.96572554e-02 4.21086997e-01 5.53901851e-01
-4.76110019e-02 4.08271104e-01 -1.21401250e+00 6.77845478e-01
-5.46716526e-03 -1.34462500e+00 2.92663038e-01 -2.08917767e-01
4.23642039e-01 1.40183598e-01 4.01857197e-01 3.53055418e-01
3.69072676e-01 -1.33181453e+00 8.33543420e-01 2.59210914e-01
9.78567362e-01 -1.06837988e+00 2.45688215e-01 1.24612503e-01
-1.19388425e+00 1.67562008e-01 -3.62827837e-01 -2.76627660e-01
1.24487035e-01 2.20136624e-02 -9.72653449e-01 -3.13551091e-02
4.57010239e-01 7.18057096e-01 -5.15255392e-01 8.97691548e-01
-5.32814622e-01 1.12327921e+00 -2.26544395e-01 -4.27165896e-01
3.42535406e-01 -1.99079782e-01 5.90355456e-01 1.65862417e+00
4.27261531e-01 -1.33735910e-01 -8.79523456e-02 7.45820761e-01
-2.01543331e-01 7.28227139e-01 -7.37462938e-01 -2.35831246e-01
5.08493066e-01 1.00441182e+00 -7.82978058e-01 -4.83577818e-01
-3.18884075e-01 7.06447601e-01 7.76735365e-01 5.08870631e-02
-3.84478390e-01 -3.70427698e-01 1.08545053e+00 -4.46926579e-02
6.38573289e-01 -4.19768214e-01 -4.50505376e-01 -9.72968280e-01
-1.95836335e-01 -9.07859325e-01 9.46341306e-02 -5.56059420e-01
-1.37160611e+00 9.98323500e-01 -2.41597459e-01 -9.01711583e-01
-5.20276606e-01 -1.07691300e+00 -5.25810480e-01 9.14938152e-01
-1.55099881e+00 -1.14559412e+00 7.06270576e-01 3.65572065e-01
9.52127218e-01 -3.83733481e-01 1.19616103e+00 2.68971354e-01
-5.17803788e-01 6.74544930e-01 -7.40979016e-02 2.68543940e-02
5.38328052e-01 -1.45129085e+00 8.55773926e-01 8.46372902e-01
8.40290964e-01 6.74847543e-01 5.65685034e-01 -3.45433474e-01
-9.12047446e-01 -1.19516826e+00 1.41221559e+00 -3.75717163e-01
9.66942728e-01 -7.08006382e-01 -1.12956417e+00 7.47595191e-01
5.63451707e-01 -1.59662783e-01 8.78189802e-01 3.55457664e-01
-3.12844306e-01 5.59830777e-02 -6.35529578e-01 9.05784547e-01
1.21185923e+00 -8.87671232e-01 -9.05465245e-01 3.80782396e-01
1.00112951e+00 -1.03703842e-01 -6.29342854e-01 3.01267982e-01
1.74837843e-01 -5.91901541e-01 7.74835825e-01 -9.16039824e-01
2.79483080e-01 2.97395084e-02 -3.88619512e-01 -1.82641947e+00
-3.97644728e-01 -7.87385225e-01 5.26126474e-02 1.48605013e+00
9.29036379e-01 -5.80439389e-01 4.33276266e-01 -5.14286198e-02
-4.54582900e-01 -6.50280476e-01 -9.51423168e-01 -8.79083037e-01
6.66435838e-01 -7.92106390e-01 7.03625560e-01 7.87164629e-01
1.74045563e-05 7.20085323e-01 -2.47099221e-01 -1.64784908e-01
-1.49783224e-01 -9.97323319e-02 2.98951585e-02 -9.93217647e-01
-4.56781149e-01 -9.66190159e-01 -3.03887486e-01 -1.00411630e+00
7.49908447e-01 -1.42173040e+00 1.58712685e-01 -1.05420101e+00
-4.22772527e-01 -4.97826427e-01 -4.61525410e-01 5.62376380e-01
-1.71924382e-02 3.50852400e-01 1.77129045e-01 -1.50066122e-01
7.62375519e-02 4.47212309e-01 7.90324569e-01 -6.26156852e-02
-2.82863438e-01 1.18764304e-01 -5.89929819e-01 8.46988738e-01
1.11274624e+00 -4.78491604e-01 -8.10928717e-02 -6.71165168e-01
3.21590871e-01 -2.65264750e-01 -1.48985341e-01 -8.23113203e-01
2.33131483e-01 -7.41598010e-03 1.28458500e-01 -4.58375812e-01
5.46267629e-01 -4.17710871e-01 -2.40628287e-01 4.57982033e-01
-5.72161019e-01 5.21599591e-01 5.54723263e-01 -6.60035312e-02
-2.88684815e-01 -4.73701596e-01 9.33794737e-01 -2.67316133e-01
-6.66202247e-01 7.78745040e-02 -9.18221653e-01 2.60088712e-01
5.31468272e-01 -1.79671213e-01 4.32074741e-02 -1.78771570e-01
-8.03429544e-01 -4.38174963e-01 2.24468827e-01 6.55499279e-01
6.33251250e-01 -1.34926379e+00 -8.39159548e-01 5.81351042e-01
4.41922806e-02 -5.68368077e-01 -3.57734829e-01 5.60378194e-01
-2.35997722e-01 5.76226652e-01 -2.78704464e-01 -1.92071781e-01
-1.08142900e+00 6.89893484e-01 4.10682321e-01 -1.62656799e-01
-2.58047253e-01 1.22398448e+00 4.00884636e-03 -7.17862844e-01
3.09698224e-01 -6.32755995e-01 -4.27775502e-01 3.55896771e-01
3.59023064e-01 1.17863275e-01 4.82513428e-01 -1.14609158e+00
-2.53430277e-01 4.43055153e-01 -3.19568545e-01 -2.67749041e-01
1.28338957e+00 -2.61408519e-02 -1.28666863e-01 8.81968081e-01
1.11749399e+00 4.25343871e-01 -8.28188121e-01 -3.67756277e-01
5.28072119e-01 1.50041699e-01 -1.73795410e-02 -9.01364565e-01
-9.50907230e-01 1.17054820e+00 1.23092048e-01 7.81094208e-02
9.26949978e-01 4.81676087e-02 1.03317499e+00 2.66901910e-01
3.91611718e-02 -1.11092222e+00 -2.61954546e-01 1.24476027e+00
8.38666975e-01 -9.76169407e-01 -7.36917436e-01 -2.54726589e-01
-5.90265751e-01 1.01751375e+00 4.17251229e-01 -2.03429714e-01
5.55058837e-01 3.58758926e-01 9.13908556e-02 2.16059074e-01
-1.00732517e+00 -5.50079107e-01 5.02443492e-01 6.19646311e-01
8.62595439e-01 1.50183544e-01 -3.78350824e-01 6.92603827e-01
-8.34025562e-01 -8.49247694e-01 4.48902369e-01 5.01310349e-01
-4.11688447e-01 -1.60988879e+00 -7.82613754e-02 5.15056908e-01
-7.31693745e-01 -8.95974040e-01 -6.26833797e-01 5.40880144e-01
2.77540475e-01 8.22508097e-01 3.15708548e-01 -2.10323423e-01
1.83790833e-01 4.63572472e-01 4.84317482e-01 -1.16133237e+00
-1.06161916e+00 -1.07629687e-01 2.68028915e-01 -2.31208727e-01
-1.61695898e-01 -7.82342374e-01 -1.17948949e+00 6.48214817e-02
-3.19051027e-01 7.15722442e-02 5.95666647e-01 1.11298227e+00
1.55745700e-01 4.22051698e-01 3.37064385e-01 -7.42347419e-01
-5.02779067e-01 -1.07415211e+00 -3.74859095e-01 7.09429607e-02
3.43156278e-01 -4.02480543e-01 -1.78448543e-01 3.72061096e-02] | [10.804110527038574, 9.781210899353027] |
ba3ca476-3bad-4268-b88b-cda63798746a | social-learning-under-platform-influence | 2202.12453 | null | https://arxiv.org/abs/2202.12453v1 | https://arxiv.org/pdf/2202.12453v1.pdf | Social Learning under Platform Influence: Consensus and Persistent Disagreement | Individuals increasingly rely on social networking platforms to form opinions. However, these platforms typically aim to maximize engagement, which may not align with social good. In this paper, we introduce an opinion dynamics model where agents are connected in a social network, and update their opinions based on their neighbors' opinions and on the content shown to them by the platform. We focus on a stochastic block model with two blocks, where the initial opinions of the individuals in different blocks are different. We prove that for large and dense enough networks the trajectory of opinion dynamics in such networks can be approximated well by a simple two-agent system. The latter admits tractable analytical analysis, which we leverage to provide interesting insights into the platform's impact on the social learning outcome in our original two-block model. Specifically, by using our approximation result, we show that agents' opinions approximately converge to some limiting opinion, which is either: consensus, where all agents agree, or persistent disagreement, where agents' opinions differ. We find that when the platform is weak and there is a high number of connections between agents with different initial opinions, a consensus equilibrium is likely. In this case, even if a persistent disagreement equilibrium arises, the polarization in this equilibrium, i.e., the degree of disagreement, is low. When the platform is strong, a persistent disagreement equilibrium is likely and the equilibrium polarization is high. A moderate platform typically leads to a persistent disagreement equilibrium with moderate polarization. Lastly, more balanced and less polarized initial opinions are more likely to lead to persistent disagreement than to consensus. | ['Jerry Anunrojwong', 'Bar Light', 'Nicole Immorlica', 'Ozan Candogan'] | 2022-02-25 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [-4.67368424e-01 5.95266283e-01 -3.32166195e-01 2.09958404e-01
7.61432424e-02 -1.04133821e+00 5.09615183e-01 1.97321489e-01
-1.13134779e-01 1.12849355e+00 1.64511740e-01 -2.85470515e-01
-6.86100051e-02 -1.05487573e+00 -4.82050776e-01 -9.74545300e-01
9.39328596e-02 5.23706675e-01 2.45759021e-02 -6.75286114e-01
-6.50502741e-02 -7.45182335e-02 -1.21073604e+00 -1.19558387e-01
1.01925921e+00 5.44833779e-01 -4.05076951e-01 6.51521742e-01
-5.04523255e-02 9.71126676e-01 -8.99932683e-01 -6.69276416e-01
5.36796570e-01 -3.38358015e-01 -4.31493253e-01 1.24622397e-01
-1.84624746e-01 3.58704403e-02 -7.86269233e-02 1.25257826e+00
3.18303883e-01 -1.87663868e-01 6.42003655e-01 -1.53003049e+00
-8.66216421e-01 9.47207153e-01 -8.89590740e-01 7.09216297e-02
4.01850194e-01 9.59535316e-02 1.12531221e+00 -2.36155763e-01
6.97770655e-01 1.15722167e+00 5.66746354e-01 3.65712196e-01
-1.27242267e+00 -6.00232780e-01 7.24602938e-01 -4.00643140e-01
-1.31724477e+00 -1.85829312e-01 7.62899160e-01 -5.05803943e-01
2.62776792e-01 8.74490216e-02 1.41419864e+00 8.73362362e-01
4.07788873e-01 3.74581397e-01 1.40337992e+00 -2.73025751e-01
4.38965619e-01 3.48050147e-01 3.84141356e-01 2.31389150e-01
9.33190286e-01 -3.09423476e-01 -7.28362918e-01 -5.40441036e-01
4.59054291e-01 5.37234731e-02 -4.36098039e-01 -1.81109831e-01
-6.02028370e-01 7.51661420e-01 4.70951647e-01 2.72039771e-01
-7.56370425e-01 -1.24007389e-01 -3.79722863e-01 1.18599963e+00
6.74745619e-01 3.60878289e-01 -2.16372073e-01 -1.63769633e-01
-3.32412928e-01 3.48628432e-01 1.43017554e+00 4.97588873e-01
1.15752661e+00 -4.10450637e-01 2.24042356e-01 4.28570986e-01
4.27886873e-01 8.35131645e-01 -1.32460147e-01 -1.17865503e+00
1.08010873e-01 1.00830090e+00 8.25408518e-01 -1.01482058e+00
-3.39775801e-01 -6.81040525e-01 -8.29864204e-01 2.53349304e-01
6.97415411e-01 -1.05234873e+00 -2.84181207e-01 2.08339238e+00
2.90002763e-01 -1.76004037e-01 3.33018869e-01 7.64438033e-01
3.08447391e-01 7.20771551e-01 -4.53336537e-01 -7.99455166e-01
8.42362463e-01 -7.03174472e-01 -6.49589360e-01 -2.26762637e-01
6.43728316e-01 -5.81124961e-01 5.14780760e-01 1.90855324e-01
-1.41639626e+00 2.66743720e-01 -8.40442777e-01 8.47748280e-01
-7.08573638e-03 -7.25260913e-01 2.79297084e-01 5.26065052e-01
-1.59759164e+00 2.80477852e-01 -3.89821231e-01 -6.67111218e-01
1.78684339e-01 3.67305517e-01 -2.58262679e-02 5.52519783e-02
-1.21466839e+00 7.17699945e-01 -6.12579525e-01 2.16177121e-01
-3.63650978e-01 -3.90819073e-01 -2.11139396e-01 -8.64909068e-02
2.72410035e-01 -1.07230401e+00 1.10682118e+00 -1.91865480e+00
-1.53989768e+00 5.77880025e-01 -2.70300090e-01 -3.84747311e-02
6.65447474e-01 1.71946526e-01 -2.78342962e-01 8.27749223e-02
3.07273120e-01 3.16885918e-01 4.16634560e-01 -1.61568010e+00
-5.67854404e-01 -3.69263828e-01 7.14006364e-01 6.44882381e-01
-5.17706692e-01 -1.74208641e-01 -1.44228831e-01 2.58601606e-01
-1.31723210e-01 -1.50146103e+00 -3.57035518e-01 -2.56518126e-01
-3.99776399e-01 -3.84532541e-01 4.49936450e-01 2.65659541e-01
1.00419247e+00 -1.93003249e+00 1.47075787e-01 8.21368992e-01
6.88308895e-01 -2.37406179e-01 -2.02369988e-01 7.16408551e-01
3.50397766e-01 5.21456659e-01 1.95633203e-01 -3.21029902e-01
-1.77038852e-02 2.57220626e-01 4.83049899e-02 7.12129295e-01
-1.45241752e-01 7.05808580e-01 -9.10499632e-01 -4.56029885e-02
-5.93539655e-01 1.60880879e-01 -6.02091610e-01 -7.26598799e-02
1.19265288e-01 4.40875858e-01 -7.63579547e-01 3.82796913e-01
6.08613968e-01 -7.78469861e-01 6.91540480e-01 4.83015358e-01
-4.42817479e-01 -9.86117795e-02 -9.93444860e-01 4.73846138e-01
-2.15399548e-01 7.36119688e-01 9.10739303e-01 -8.11481833e-01
5.06395161e-01 4.68353599e-01 3.94283175e-01 -9.15648602e-03
4.00435120e-01 3.63974690e-01 5.99490404e-01 -8.19077641e-02
2.02795893e-01 -1.72704309e-01 1.73549145e-01 9.52061057e-01
-5.91632485e-01 -1.21898934e-01 4.11023408e-01 6.32141948e-01
1.07798946e+00 -7.22012937e-01 8.81926809e-03 -6.78166807e-01
1.46343157e-01 1.17169786e-02 5.80681622e-01 9.58346128e-01
-4.13680732e-01 7.39518367e-03 1.25791645e+00 -7.42761046e-02
-9.35202003e-01 -9.08747673e-01 2.74785668e-01 6.78987622e-01
7.82254577e-01 -1.63576454e-01 -6.85308397e-01 -1.34687215e-01
2.03329310e-01 3.46017331e-02 -6.76066577e-01 3.18038873e-02
-1.52011424e-01 -8.67987990e-01 -1.79948971e-01 -1.00035444e-01
4.78155166e-01 -7.10150957e-01 1.34192601e-01 -1.14898965e-01
-2.25241467e-01 -5.59959233e-01 -4.60331738e-01 -3.87348711e-01
-6.77704334e-01 -1.25361955e+00 -1.00222790e+00 -4.51121747e-01
1.26058912e+00 5.46852827e-01 1.00066590e+00 5.39539099e-01
7.26344764e-01 7.34110117e-01 -3.01667780e-01 -4.90303785e-01
-3.71057272e-01 3.30373257e-01 3.80739421e-01 2.97520995e-01
1.46172211e-01 -7.86885440e-01 -1.19343519e+00 5.67614436e-01
-6.72698081e-01 -3.56470227e-01 3.51174802e-01 2.42058799e-01
-1.04636364e-01 7.61330547e-03 9.86053944e-01 -6.64553761e-01
1.06129777e+00 -1.03064704e+00 -5.02076387e-01 5.88171706e-02
-5.31206787e-01 -4.53201622e-01 4.68932122e-01 -4.84268427e-01
-8.61494780e-01 -5.91503084e-01 4.74357545e-01 2.18447000e-01
5.45316398e-01 8.67096364e-01 2.53441215e-01 -1.78514406e-01
6.40648663e-01 -3.66377264e-01 5.61381161e-01 8.02445225e-03
2.28500903e-01 8.96852553e-01 -4.78922009e-01 -5.43681264e-01
8.47197533e-01 7.53786504e-01 -1.85928658e-01 -7.46433318e-01
-8.63892198e-01 1.01308994e-01 -6.43939450e-02 -6.23380065e-01
4.23586041e-01 -1.04950595e+00 -1.13159633e+00 7.74253368e-01
-9.73463356e-01 -6.54453337e-01 -3.39130521e-01 2.73803413e-01
4.00594715e-03 6.25682846e-02 -7.76021838e-01 -1.25612378e+00
7.11959973e-02 -8.99518788e-01 2.74814725e-01 7.23963797e-01
-4.92606789e-01 -1.34726524e+00 3.22790444e-01 2.80644715e-01
7.61611640e-01 -1.37726903e-01 4.49329317e-01 -6.10889912e-01
-6.77450597e-01 -2.40963429e-01 7.76020288e-02 -7.26321414e-02
2.64210403e-01 5.28702199e-01 -2.79183149e-01 -4.81052816e-01
-3.79076451e-01 -2.35212192e-01 3.40282559e-01 7.19236910e-01
-1.24397196e-01 -3.87205899e-01 -5.21658003e-01 -3.08433026e-01
9.37596262e-01 5.60043566e-02 6.28704056e-02 1.14712395e-01
2.78875232e-01 9.63507116e-01 8.32948238e-02 3.34231973e-01
8.08875799e-01 1.97812065e-01 3.43469918e-01 -6.21842071e-02
3.66721243e-01 -9.14031565e-02 7.84713805e-01 1.13767600e+00
-2.54201859e-01 -7.11419523e-01 -7.95018673e-01 7.20463157e-01
-2.12284160e+00 -7.54542112e-01 -2.12724730e-01 2.10940027e+00
7.95516372e-01 4.17878568e-01 4.45312202e-01 -2.93136865e-01
1.05544782e+00 8.85625705e-02 -6.70763254e-01 -2.55501658e-01
-6.51389003e-01 -4.25801516e-01 4.21638459e-01 1.06864893e+00
-2.49457628e-01 6.32633030e-01 7.18992758e+00 1.53508008e-01
-1.12340426e+00 2.18831629e-01 8.74855101e-01 -6.05728686e-01
-8.12204838e-01 2.70500511e-01 -6.97211504e-01 4.49235767e-01
7.79334843e-01 -7.32755721e-01 4.23811495e-01 3.47055346e-01
4.74831820e-01 -3.78873646e-01 -5.68075061e-01 5.06464660e-01
-1.66420594e-01 -1.31527352e+00 -3.20315868e-01 5.14845669e-01
1.54413748e+00 6.19833469e-02 8.53231251e-02 -7.86091462e-02
1.27742457e+00 -6.02619290e-01 6.16387486e-01 4.82038021e-01
2.77689040e-01 -7.49448895e-01 7.59342909e-01 5.85366547e-01
-8.03881586e-01 -3.78178447e-01 -1.43101230e-01 -9.93779361e-01
5.55182278e-01 9.87929642e-01 1.94520965e-01 1.08265959e-01
4.08161789e-01 6.68579996e-01 7.50221610e-02 7.00189829e-01
-3.40673566e-01 5.46563685e-01 -4.39617038e-01 -3.60527247e-01
9.63602290e-02 -7.42448568e-01 7.24261105e-01 1.45928621e-01
3.83687139e-01 4.75248575e-01 2.38403514e-01 5.37397683e-01
-4.48640376e-01 1.83172792e-01 -7.34877050e-01 -1.84247550e-02
6.34228826e-01 1.25651991e+00 -6.32951558e-01 -5.18026650e-01
-4.00304526e-01 4.73904580e-01 4.46191460e-01 9.99807119e-01
-2.51837969e-01 3.10697220e-02 1.09631002e+00 3.85236889e-01
-5.01632616e-02 9.59727913e-02 -1.14615746e-02 -1.39254880e+00
3.65348603e-03 -7.18888342e-01 3.96067388e-02 -7.30652809e-01
-1.51373386e+00 4.51312363e-01 -3.07077050e-01 -7.63338864e-01
9.39195678e-02 -1.19622812e-01 -1.18633354e+00 6.84826612e-01
-1.32415545e+00 -6.05211139e-01 9.14290026e-02 3.09740096e-01
-2.08279550e-01 3.72133218e-02 1.16708726e-01 1.26450300e-01
-8.86120081e-01 2.46867701e-01 2.46443942e-01 -2.02250794e-01
6.16607964e-01 -9.43596780e-01 -3.31118286e-01 4.58112150e-01
-2.96543390e-01 8.27597320e-01 9.71261203e-01 -7.57347703e-01
-1.31462824e+00 -6.64265335e-01 8.48440349e-01 -4.32133645e-01
1.15063453e+00 -1.18655145e-01 -4.78526950e-01 7.21583128e-01
7.11413980e-01 -1.06848501e-01 9.54770803e-01 3.29543680e-01
4.62158062e-02 -1.04123123e-01 -1.03727341e+00 1.23621190e+00
1.01473367e+00 -3.01636249e-01 -7.73308519e-03 4.97403383e-01
4.05408233e-01 2.45515764e-01 -5.70696533e-01 2.58735265e-03
5.74598074e-01 -9.94579196e-01 1.08646773e-01 -5.75827301e-01
3.60144585e-01 -3.21359366e-01 1.24743164e-01 -1.94949114e+00
-2.96960920e-01 -1.23367298e+00 2.25140408e-01 8.47243428e-01
9.36749637e-01 -1.34937751e+00 9.65981960e-01 7.76826680e-01
4.58983988e-01 -8.00484240e-01 -7.71417558e-01 -4.06932414e-01
5.65176606e-01 3.67611617e-01 4.07501519e-01 9.15887296e-01
5.11336923e-01 5.22648275e-01 -5.68620116e-02 2.16113001e-01
5.93839049e-01 2.37649351e-01 7.18931496e-01 -1.55356920e+00
2.15701517e-02 -6.05668068e-01 3.48958038e-02 -1.31679058e+00
3.27995628e-01 -3.91266227e-01 -3.04092407e-01 -1.71832764e+00
3.59712780e-01 -6.70763612e-01 -5.38373142e-02 -7.88974762e-02
-1.77594423e-01 4.23265547e-01 7.76917562e-02 6.33900344e-01
-9.01776314e-01 1.98006138e-01 1.27460980e+00 -1.67999472e-02
-6.08451009e-01 1.95356950e-01 -1.40975332e+00 8.94106150e-01
8.04914594e-01 -3.25104505e-01 -3.97861898e-01 -1.79212471e-03
1.26091492e+00 1.87549949e-01 -1.59932002e-01 -2.18992382e-01
5.35352349e-01 -2.75211096e-01 -2.42668077e-01 -1.88514858e-01
2.46175945e-01 -4.65694159e-01 4.47644353e-01 6.41937196e-01
-2.55297005e-01 1.31661803e-01 -5.00476360e-01 6.98941648e-01
-4.65754382e-02 8.86310413e-02 5.94611883e-01 -1.25152497e-02
4.51758206e-01 2.39265278e-01 -1.06632197e+00 2.23411713e-02
1.20397282e+00 -1.79501623e-01 -8.46996963e-01 -1.37470925e+00
-9.61686671e-01 7.69329548e-01 1.08671522e+00 -8.85040089e-02
-6.81582913e-02 -1.27853417e+00 -9.22787786e-01 -3.47478569e-01
-1.36199072e-01 -2.71364897e-01 2.34193400e-01 1.30221415e+00
-9.14851874e-02 -2.47529536e-01 6.38979450e-02 -4.49869663e-01
-1.10516787e+00 6.76441714e-02 7.11536705e-01 -1.42462850e-01
2.19614267e-01 7.42797196e-01 3.66016865e-01 -2.87303120e-01
-1.84900030e-01 1.18685625e-01 -3.44032228e-01 5.49364984e-01
2.82432169e-01 3.57944489e-01 -5.02459586e-01 -7.22652972e-01
-1.34722620e-01 7.68109322e-01 -1.05778024e-01 -2.82495707e-01
1.09518325e+00 -5.64583182e-01 -6.54285192e-01 6.75494134e-01
5.85317373e-01 7.38055766e-01 -1.13553643e+00 -1.52565226e-01
-6.14880741e-01 -2.93841153e-01 -3.11049789e-01 -3.37012112e-01
-1.35623145e+00 1.40925199e-01 -2.04019651e-01 1.31750286e+00
5.36510110e-01 3.26042771e-01 1.57843128e-01 4.02427882e-01
3.41935337e-01 -9.41466033e-01 1.96624935e-01 5.02397954e-01
6.39878333e-01 -1.12082982e+00 -9.91344601e-02 -4.69171286e-01
-5.43870330e-01 4.35961157e-01 5.84431887e-01 -3.72252077e-01
1.21556938e+00 7.66285434e-02 3.61319095e-01 -2.41990373e-01
-1.27663064e+00 -5.99020086e-02 -3.92253846e-01 4.63809311e-01
4.01322544e-01 1.68719813e-01 -4.68172997e-01 3.95878077e-01
-3.24274331e-01 -3.53469312e-01 1.18149471e+00 6.38533592e-01
-7.05180526e-01 -1.22352254e+00 -3.37455034e-01 6.10203445e-01
-5.45411646e-01 2.03056913e-02 -1.07608140e+00 4.68306184e-01
-1.69373021e-01 1.47476399e+00 4.40009356e-01 1.28173873e-01
3.23538005e-01 -3.41477990e-01 1.61237881e-01 -5.64999998e-01
-6.34877324e-01 -4.74594906e-02 1.48913696e-01 -5.43146096e-02
-8.56597602e-01 -7.01422870e-01 -7.62546897e-01 -1.05496407e+00
-5.52869320e-01 5.82555056e-01 -5.13238013e-02 8.43947709e-01
5.83713591e-01 5.48951700e-03 8.37648630e-01 -4.05907571e-01
-2.37534150e-01 -7.53826559e-01 -9.05444264e-01 2.69804090e-01
5.55040061e-01 -4.61239129e-01 -1.03818476e+00 -4.61587101e-01] | [6.865180015563965, 5.330889701843262] |
5822555f-35f3-4206-8622-685bf5a50ed4 | a-multi-cascaded-deep-model-for-bilingual-sms | 1911.13066 | null | https://arxiv.org/abs/1911.13066v1 | https://arxiv.org/pdf/1911.13066v1.pdf | A Multi-cascaded Deep Model for Bilingual SMS Classification | Most studies on text classification are focused on the English language. However, short texts such as SMS are influenced by regional languages. This makes the automatic text classification task challenging due to the multilingual, informal, and noisy nature of language in the text. In this work, we propose a novel multi-cascaded deep learning model called McM for bilingual SMS classification. McM exploits $n$-gram level information as well as long-term dependencies of text for learning. Our approach aims to learn a model without any code-switching indication, lexical normalization, language translation, or language transliteration. The model relies entirely upon the text as no external knowledge base is utilized for learning. For this purpose, a 12 class bilingual text dataset is developed from SMS feedbacks of citizens on public services containing mixed Roman Urdu and English languages. Our model achieves high accuracy for classification on this dataset and outperforms the previous model for multilingual text classification, highlighting language independence of McM. | ['Asim Karim', 'Muhammad Haroon Shakeel', 'Imdadullah Khan'] | 2019-11-29 | null | null | null | null | ['multilingual-text-classification', 'lexical-normalization'] | ['miscellaneous', 'natural-language-processing'] | [-2.00617891e-02 -3.67312849e-01 -5.11437654e-01 -6.84545040e-01
-9.01578128e-01 -5.98145068e-01 7.22329617e-01 2.86215752e-01
-8.71424139e-01 7.38094866e-01 4.15898353e-01 -8.71000290e-01
6.38366282e-01 -5.10468483e-01 -5.86941063e-01 -2.76644230e-01
6.19941235e-01 6.44299209e-01 -2.47906059e-01 -5.59831738e-01
4.04690159e-03 -1.04318835e-01 -9.67421710e-01 1.13995147e+00
1.07417929e+00 6.83785439e-01 2.64685750e-01 7.70786405e-01
-5.54994047e-01 1.13374209e+00 -3.54123116e-01 -6.07622683e-01
-7.92877376e-02 -4.00735229e-01 -1.03173959e+00 1.41299190e-02
8.11349079e-02 -2.23063380e-01 -2.47871786e-01 7.16651618e-01
5.60351133e-01 -2.79210031e-01 7.83998311e-01 -7.86502600e-01
-3.09973508e-01 1.12145531e+00 -3.96800518e-01 5.36105819e-02
3.42833698e-01 -2.85723984e-01 1.11879063e+00 -1.11850131e+00
6.45835042e-01 1.08775735e+00 1.00715923e+00 2.83799022e-01
-1.23919976e+00 -6.97135746e-01 3.43875170e-01 4.99077663e-02
-1.26891637e+00 -6.02259099e-01 5.86296439e-01 -7.20094323e-01
1.33377433e+00 2.31607050e-01 1.89861625e-01 1.52173460e+00
4.00147974e-01 1.06615627e+00 1.37383091e+00 -8.43104661e-01
-1.89769000e-01 7.61577249e-01 8.26832280e-02 5.50769329e-01
-4.30181652e-01 -3.72130245e-01 -4.90983546e-01 7.52927363e-03
-2.58954316e-01 -1.19829461e-01 6.88442364e-02 1.92758709e-01
-8.69195223e-01 9.90406930e-01 -1.72278576e-03 5.71624637e-01
8.08382109e-02 -3.69988739e-01 7.49980152e-01 6.44642770e-01
1.06670284e+00 -1.29066631e-01 -9.99823332e-01 -1.20481648e-01
-1.09358358e+00 -2.66985387e-01 9.31425750e-01 8.07048619e-01
6.56241715e-01 -2.50485420e-01 6.48759380e-02 1.23290718e+00
2.78221250e-01 6.60881460e-01 8.09065104e-01 5.08511923e-02
1.18817282e+00 8.71502817e-01 -3.86041969e-01 -9.06942964e-01
-7.14665055e-01 -4.03604448e-01 -9.62914884e-01 -4.97504830e-01
2.18208194e-01 -4.45327163e-01 -5.50473034e-01 1.42216384e+00
5.93816582e-03 -6.98672175e-01 1.83026373e-01 2.15468928e-01
1.09983301e+00 6.28113329e-01 -7.61470348e-02 -9.81567055e-02
1.06291914e+00 -1.03728867e+00 -5.99801540e-01 -3.46073240e-01
1.22239327e+00 -1.04217887e+00 1.06895769e+00 1.28009349e-01
-5.36944568e-01 -4.17303979e-01 -5.51261365e-01 -3.46370637e-01
-8.24221253e-01 6.60046399e-01 3.77364337e-01 9.22880292e-01
-9.06309724e-01 9.34344009e-02 -5.21616995e-01 -6.19718373e-01
2.20158622e-01 4.60922778e-01 -2.90768832e-01 -6.54610917e-02
-1.32613373e+00 9.19867098e-01 5.45726418e-01 4.54967096e-02
-3.02701592e-01 -1.31115511e-01 -8.15770388e-01 -3.72630596e-01
-1.78022683e-01 -1.47516191e-01 1.25417709e+00 -1.66803670e+00
-1.51387334e+00 1.16420865e+00 -3.69769573e-01 -2.63796538e-01
9.62724745e-01 6.31240234e-02 -5.84127009e-01 -3.96082491e-01
3.47292751e-01 4.48408306e-01 7.89296329e-01 -8.87480259e-01
-9.09734964e-01 -1.93961784e-01 -6.65697604e-02 3.80960703e-01
-7.80893803e-01 4.58965272e-01 -2.20778868e-01 -6.16293907e-01
-1.68853626e-01 -9.42847908e-01 -6.43074065e-02 -7.82479405e-01
-2.98515379e-01 -9.14435536e-02 6.96923852e-01 -1.26410890e+00
1.49115431e+00 -1.91042387e+00 -1.52802721e-01 1.52183756e-01
-1.51536912e-01 -3.96085680e-02 -1.18774831e-01 6.95454776e-01
-6.34834450e-03 1.40799522e-01 -2.28642803e-02 -6.16285980e-01
-7.48513341e-02 1.79974526e-01 -2.38570362e-01 4.85298902e-01
1.63023725e-01 9.04921412e-01 -6.77529573e-01 -5.31745017e-01
3.84092666e-02 4.20032978e-01 -6.28853500e-01 -2.04887778e-01
-1.01373777e-01 8.34475935e-01 -1.49262831e-01 6.09724343e-01
4.65497762e-01 -4.68642414e-02 6.42543852e-01 -4.23837826e-02
-4.10407275e-01 7.42039442e-01 -6.98025107e-01 1.51939034e+00
-1.03362477e+00 8.56919825e-01 5.23592830e-02 -1.11674953e+00
6.01701558e-01 4.87373978e-01 4.37904358e-01 -1.00519407e+00
5.33693552e-01 5.77429652e-01 -9.60580409e-02 -3.39048326e-01
4.74052787e-01 5.02798753e-03 -4.18566942e-01 5.83114386e-01
1.13997184e-01 8.73511210e-02 2.91759282e-01 5.42538092e-02
6.24931514e-01 9.74529162e-02 3.05755526e-01 -4.30208892e-01
8.01430821e-01 2.24966882e-03 3.19784939e-01 8.22931528e-01
1.17853858e-01 3.00952077e-01 2.70801008e-01 -5.82946777e-01
-9.49978113e-01 -2.44872093e-01 -4.05686051e-01 1.74477494e+00
-5.91987073e-01 -4.44460422e-01 -7.00451195e-01 -9.74779248e-01
-1.74999982e-01 4.11191434e-01 -3.71964663e-01 1.37352124e-01
-6.95082426e-01 -1.32126749e+00 6.59144163e-01 2.99812675e-01
4.27929133e-01 -7.21049488e-01 3.38303238e-01 4.19976503e-01
-9.59061205e-01 -1.37387693e+00 -5.09153247e-01 4.80279893e-01
-7.30220854e-01 -9.20056760e-01 -4.01927650e-01 -1.09370065e+00
7.24176586e-01 -1.38346016e-01 1.23299658e+00 -1.68821380e-01
1.79907084e-01 1.16003722e-01 -5.65061152e-01 -3.01914483e-01
-9.75459039e-01 9.50285912e-01 2.27628555e-02 3.55440259e-01
6.25282884e-01 -1.85187861e-01 -3.63448523e-02 2.44376317e-01
-5.82427144e-01 4.17082638e-01 5.33678174e-01 1.10767400e+00
7.09272549e-02 8.73168260e-02 4.98098999e-01 -1.09521973e+00
4.96888191e-01 -5.98384261e-01 -3.95084023e-01 2.11104289e-01
-4.17521089e-01 -7.59952515e-02 9.08346772e-01 -3.15389484e-01
-1.15293014e+00 1.71230122e-01 -5.75835705e-01 7.25208461e-01
-9.64791924e-02 7.52366662e-01 -1.89737454e-01 2.73455097e-03
7.76972055e-01 2.43927479e-01 -4.85207587e-01 -5.70501566e-01
9.82177928e-02 1.59059262e+00 -6.25880212e-02 -4.82177347e-01
4.99956518e-01 2.98589528e-01 -5.50399840e-01 -9.88240719e-01
-8.98733616e-01 -5.78695893e-01 -1.24012494e+00 -1.04271069e-01
7.33832717e-01 -1.24573278e+00 -4.04557556e-01 7.02250481e-01
-1.22017324e+00 -4.21323091e-01 3.89004022e-01 5.11561453e-01
-2.69967973e-01 2.18443260e-01 -1.05055249e+00 -6.95133388e-01
-3.24095964e-01 -1.07966411e+00 9.44910228e-01 -4.74617332e-01
-2.72228032e-01 -1.36123013e+00 9.91771556e-03 8.46781254e-01
3.19100618e-01 -2.18769208e-01 1.01730680e+00 -7.38098681e-01
1.08676977e-01 -3.48692447e-01 -1.16190672e-01 4.65574592e-01
3.22798580e-01 -5.20300912e-03 -1.17848623e+00 -5.34578800e-01
-5.30267619e-02 -5.62506497e-01 8.49596858e-01 1.64013997e-01
8.21217179e-01 -5.04230022e-01 -1.59642205e-01 4.67126220e-01
1.17429245e+00 -4.74995039e-02 -6.75465018e-02 5.25785446e-01
9.11237955e-01 6.74012542e-01 2.08773836e-01 3.01611632e-01
9.09518361e-01 5.08646548e-01 -8.42200145e-02 -3.14458996e-01
2.46829063e-01 -1.66203648e-01 7.66908288e-01 1.63700008e+00
2.52904415e-01 -2.57828355e-01 -1.48697305e+00 4.35212195e-01
-1.80367231e+00 -6.24742270e-01 -4.95140284e-01 1.92644763e+00
1.21064639e+00 1.19475126e-01 5.83324321e-02 2.25491315e-01
6.27615809e-01 -1.44425735e-01 -8.52599144e-02 -5.33359826e-01
-5.66670716e-01 5.23884129e-03 5.95336616e-01 7.57191956e-01
-1.55719233e+00 1.21830463e+00 5.66321135e+00 8.26655149e-01
-1.32469606e+00 5.31981051e-01 8.29785466e-01 4.56120893e-02
-1.84926346e-01 -1.48947999e-01 -1.22340226e+00 5.70086837e-01
1.04428196e+00 3.07708204e-01 3.41832548e-01 4.19030756e-01
2.40393355e-01 -2.90121902e-02 -9.02473032e-01 8.81613374e-01
1.60325155e-01 -9.13218737e-01 1.34336114e-01 -1.42608061e-01
1.11146319e+00 6.49295747e-01 -1.08468765e-02 6.16705775e-01
4.34276313e-01 -9.05315161e-01 8.45976591e-01 2.55184826e-02
8.52639556e-01 -8.34252834e-01 1.02090061e+00 8.67295504e-01
-1.06397283e+00 -3.47956181e-01 8.00539088e-03 -1.15487345e-01
-1.36983693e-01 4.89169180e-01 -9.72692251e-01 5.76983869e-01
5.61549246e-01 8.68656874e-01 -8.71640146e-01 4.55336049e-02
1.69680268e-02 7.76174247e-01 -3.22844207e-01 -7.95598491e-04
4.60241467e-01 -8.18521157e-02 9.16952863e-02 1.89835703e+00
1.05320469e-01 -5.95733821e-01 5.83504617e-01 6.99892417e-02
-2.96444982e-01 9.58538532e-01 -5.86111307e-01 -1.83153391e-01
-2.12060705e-01 1.02582812e+00 -7.86861777e-01 -2.69316256e-01
-8.86937559e-01 1.07832205e+00 4.40598786e-01 3.36505830e-01
-6.23287082e-01 -1.51341632e-01 1.12781346e-01 -8.54403377e-02
7.97463506e-02 -2.66769677e-01 -4.27434742e-01 -1.62285280e+00
8.47438127e-02 -1.25071144e+00 3.79042715e-01 -1.28488809e-01
-1.24900055e+00 7.56937444e-01 -5.03082991e-01 -1.06761920e+00
-4.36515689e-01 -7.93792665e-01 -2.30159098e-03 8.96787524e-01
-1.76876318e+00 -1.66525280e+00 1.81616008e-01 7.10728407e-01
7.28567123e-01 -6.06006265e-01 8.63821745e-01 8.97735178e-01
-5.99450529e-01 8.92884374e-01 7.76695848e-01 7.95897484e-01
9.12910342e-01 -9.85748470e-01 3.84734511e-01 6.85515285e-01
1.23399526e-01 2.83381253e-01 2.38035649e-01 -7.18468308e-01
-9.86481547e-01 -1.27064025e+00 1.86797965e+00 -7.72469997e-01
9.34972048e-01 -9.92890418e-01 -4.59484458e-01 8.02247405e-01
2.71131516e-01 -3.96856844e-01 1.09773576e+00 3.54375899e-01
-3.59250844e-01 -1.47753879e-01 -7.43824661e-01 6.01321220e-01
6.51797414e-01 -1.05203533e+00 -1.18959866e-01 7.54612207e-01
2.76349783e-01 -2.97469616e-01 -6.19670808e-01 1.89796835e-01
5.57477772e-01 -5.22083998e-01 3.29321355e-01 -7.69192338e-01
3.93725127e-01 1.38681397e-01 -4.22257990e-01 -1.19708562e+00
-1.35047752e-02 -2.45619237e-01 3.44188243e-01 1.30891395e+00
9.68898177e-01 -6.43597066e-01 4.14019257e-01 1.47003546e-01
2.20388100e-02 -2.86242455e-01 -9.57316995e-01 -4.45214599e-01
5.13963878e-01 -6.34387612e-01 1.84980109e-01 1.56893122e+00
1.73449829e-01 7.57024825e-01 -5.17543733e-01 -1.64472178e-01
2.57322371e-01 4.33507822e-02 6.99715495e-01 -1.19332302e+00
-2.16223150e-01 -4.45445925e-01 -7.15960935e-02 -9.73666608e-01
6.67793930e-01 -1.57604265e+00 -7.54867569e-02 -1.25515926e+00
3.59446913e-01 -4.56422806e-01 7.44340867e-02 5.94573677e-01
3.27839591e-02 4.30565119e-01 -1.60748646e-01 1.94926649e-01
-5.34557164e-01 5.17897844e-01 5.47121763e-01 -5.07939100e-01
-2.08794743e-01 1.87672302e-01 -2.90140897e-01 7.79560506e-01
1.08661163e+00 -6.46538734e-01 -9.50506516e-03 -8.58096540e-01
6.87449276e-01 -2.40997523e-01 -1.67345464e-01 -6.53863668e-01
1.29944757e-01 3.83426398e-02 4.97007877e-01 -6.45255566e-01
-1.38429686e-01 -8.99757028e-01 -2.81430513e-01 3.75935048e-01
-3.05376440e-01 1.38009951e-01 2.18783468e-01 9.16914493e-02
-3.34955692e-01 -1.17193915e-01 6.72719896e-01 -2.40361705e-01
-1.94728792e-01 1.76232487e-01 -1.02518511e+00 9.49290171e-02
2.69475609e-01 6.42741099e-02 -1.01738609e-01 -2.89308876e-01
-6.73540652e-01 2.20401838e-01 2.24107191e-01 7.29783773e-01
-9.58065912e-02 -1.24053240e+00 -9.85446572e-01 4.31706905e-01
2.26345003e-01 -3.57538611e-01 -6.70935661e-02 1.06542671e+00
-4.95064497e-01 7.17971087e-01 4.94621657e-02 -6.87724173e-01
-1.34862137e+00 3.21659930e-02 4.62876946e-01 -5.66118836e-01
-8.38702247e-02 5.58113217e-01 -1.29639462e-01 -1.45230401e+00
1.15420729e-01 -2.56808490e-01 -3.94377619e-01 5.84888279e-01
1.75922409e-01 5.99654019e-02 6.10783517e-01 -1.11579680e+00
-2.27044165e-01 3.53143007e-01 -3.81396472e-01 -3.03636640e-01
1.15347505e+00 -5.06959021e-01 -3.84470075e-01 8.86842310e-01
1.39681911e+00 1.77006811e-01 -5.60059607e-01 -4.76169050e-01
3.71170431e-01 9.83444881e-03 2.69800853e-02 -8.09945762e-01
-7.71065712e-01 9.03446078e-01 4.02941972e-01 -1.06830813e-01
7.89292872e-01 -2.83482075e-01 6.36951685e-01 7.54653513e-01
2.55916119e-01 -1.85387683e+00 -2.97712117e-01 1.30184555e+00
4.91546482e-01 -1.84817171e+00 -2.75619835e-01 1.37535989e-01
-5.57538688e-01 1.11285341e+00 3.91651481e-01 6.20699644e-01
9.59046483e-01 2.69070208e-01 5.77604949e-01 3.71019930e-01
-6.44136369e-01 -8.37695003e-02 1.77495033e-01 3.38056207e-01
1.09831452e+00 3.66488472e-02 -4.34034467e-01 4.65588510e-01
-4.48304325e-01 -2.50963897e-01 3.27975750e-01 9.93795395e-01
-1.12363026e-01 -1.43083000e+00 -1.20733760e-01 4.26144570e-01
-9.30314183e-01 -6.92777216e-01 -5.87385893e-01 4.49405372e-01
1.67628407e-01 1.27756381e+00 -1.29047930e-02 -4.12283331e-01
-6.17756182e-03 4.85672504e-01 1.06184214e-01 -6.60523951e-01
-1.07648146e+00 6.09019548e-02 4.18210059e-01 -6.48259744e-02
-4.24829096e-01 -8.32450926e-01 -9.32565928e-01 -4.51470077e-01
-3.05624783e-01 1.60506845e-01 1.07698166e+00 1.16047704e+00
3.36161703e-02 1.33004144e-01 1.05603862e+00 -5.92944145e-01
-1.87998265e-01 -1.19608748e+00 -1.52522370e-01 2.86569715e-01
5.14018714e-01 -3.99199426e-02 -2.72578031e-01 2.57960826e-01] | [10.80510425567627, 9.891663551330566] |
12fd67e0-2cac-4309-bc5a-7c45e3cdab4d | karasinger-score-free-singing-voice-synthesis | 2110.04005 | null | https://arxiv.org/abs/2110.04005v1 | https://arxiv.org/pdf/2110.04005v1.pdf | KaraSinger: Score-Free Singing Voice Synthesis with VQ-VAE using Mel-spectrograms | In this paper, we propose a novel neural network model called KaraSinger for a less-studied singing voice synthesis (SVS) task named score-free SVS, in which the prosody and melody are spontaneously decided by machine. KaraSinger comprises a vector-quantized variational autoencoder (VQ-VAE) that compresses the Mel-spectrograms of singing audio to sequences of discrete codes, and a language model (LM) that learns to predict the discrete codes given the corresponding lyrics. For the VQ-VAE part, we employ a Connectionist Temporal Classification (CTC) loss to encourage the discrete codes to carry phoneme-related information. For the LM part, we use location-sensitive attention for learning a robust alignment between the input phoneme sequence and the output discrete code. We keep the architecture of both the VQ-VAE and LM light-weight for fast training and inference speed. We validate the effectiveness of the proposed design choices using a proprietary collection of 550 English pop songs sung by multiple amateur singers. The result of a listening test shows that KaraSinger achieves high scores in intelligibility, musicality, and the overall quality. | ['Yi-Hsuan Yang', 'Jen-Yu Liu', 'Chien-Feng Liao'] | 2021-10-08 | null | null | null | null | ['singing-voice-synthesis'] | ['speech'] | [ 1.03426866e-01 -8.20719525e-02 2.05941331e-02 -6.63536936e-02
-9.88869429e-01 -6.41689301e-01 1.59407362e-01 -4.38241482e-01
-7.88631812e-02 3.58254552e-01 4.74429250e-01 -1.53252646e-01
-3.24058942e-02 -2.75292784e-01 -8.31267834e-01 -6.49468601e-01
4.04994609e-03 1.06745139e-01 6.38671499e-03 -2.01780662e-01
-2.70501226e-02 2.55703032e-02 -1.73659849e+00 6.26653731e-01
4.38994169e-01 1.11480296e+00 5.43395281e-01 1.17900407e+00
2.87607044e-01 9.45027888e-01 -5.90688765e-01 -9.50461626e-02
2.52155423e-01 -9.63827670e-01 -5.55373728e-01 -1.57292575e-01
3.94134283e-01 -1.55605704e-01 -4.29959863e-01 7.24688113e-01
6.43948019e-01 4.92213935e-01 5.94524801e-01 -8.67245495e-01
-5.63628793e-01 9.37352300e-01 9.21957344e-02 6.66309372e-02
1.24448575e-01 1.84324697e-01 1.66203725e+00 -1.04912925e+00
3.59331459e-01 9.77040768e-01 6.76117301e-01 7.77845442e-01
-1.16312790e+00 -7.04521656e-01 -3.12853336e-01 3.87203097e-01
-1.52114272e+00 -7.01808572e-01 1.09288836e+00 -5.48019111e-01
9.87321496e-01 5.41548312e-01 7.65327394e-01 9.62549269e-01
8.37012976e-02 8.63828182e-01 4.59216386e-01 -5.74773073e-01
4.76112574e-01 8.02768953e-03 -3.89621705e-01 4.40215647e-01
-8.51205885e-01 4.48555380e-01 -9.37914610e-01 -1.70181170e-01
7.60535598e-01 -3.69779587e-01 -4.73240316e-01 -4.54341769e-02
-1.01351035e+00 8.68856788e-01 2.70086497e-01 2.95891374e-01
-4.83197510e-01 2.95043558e-01 4.30295497e-01 4.47726369e-01
1.58221290e-01 6.48310661e-01 -3.46678436e-01 -3.11665386e-01
-1.33164263e+00 2.56566197e-01 7.59228587e-01 3.13235611e-01
-3.78631726e-02 7.45437086e-01 -4.49902773e-01 1.24437368e+00
1.87641397e-01 2.27273285e-01 8.84369433e-01 -1.31252229e+00
4.20881987e-01 -2.09872723e-01 -7.07591251e-02 -8.19150805e-01
1.82686135e-01 -6.27026558e-01 -6.29583657e-01 2.02230886e-01
1.43424258e-01 -1.40163705e-01 -4.74808872e-01 1.95828485e+00
-9.93804038e-02 5.14110029e-01 -1.74180567e-02 1.17319143e+00
5.74547887e-01 1.24150681e+00 -3.18517894e-01 -3.49919021e-01
9.27003860e-01 -1.19002616e+00 -9.31071758e-01 -6.29460961e-02
4.29948010e-02 -6.87923431e-01 1.48945022e+00 5.74088395e-01
-1.49328411e+00 -1.05557489e+00 -1.24290395e+00 -4.04278561e-02
2.46280968e-01 5.61500609e-01 -5.09184040e-02 2.50357985e-01
-8.82954955e-01 1.01009202e+00 -6.87905192e-01 2.70614058e-01
-1.55992359e-01 3.90369564e-01 7.01933950e-02 6.95093691e-01
-1.30708480e+00 3.68108004e-01 3.90404254e-01 -5.10772914e-02
-1.28267848e+00 -6.02187753e-01 -7.06044137e-01 3.86170477e-01
1.13281179e-02 -3.39865148e-01 1.50561547e+00 -1.11105430e+00
-1.98671293e+00 5.60732067e-01 9.85704735e-03 -6.42605066e-01
3.61855980e-03 -1.47111237e-01 -5.50646544e-01 9.32753831e-02
-3.50351036e-01 5.58940887e-01 1.31861937e+00 -9.12935495e-01
-4.98124778e-01 2.88889438e-01 -5.42101741e-01 3.90408814e-01
-6.19589150e-01 -6.38738498e-02 -1.45687908e-01 -1.17795765e+00
2.13425271e-02 -8.34250331e-01 2.78378546e-01 -2.27053866e-01
-4.30564076e-01 -3.77903581e-01 5.59819102e-01 -1.06547749e+00
1.81856310e+00 -2.61383438e+00 6.60758138e-01 -4.33138497e-02
-1.72511056e-01 3.63676697e-01 -3.45491379e-01 4.25270110e-01
-1.55410290e-01 -3.11276227e-01 -3.99375260e-01 -5.17609537e-01
7.06321448e-02 2.15477303e-01 -9.18738484e-01 1.57555506e-01
8.35765675e-02 7.16715515e-01 -6.72168136e-01 -2.27618635e-01
3.66357677e-02 5.06469011e-01 -9.85136688e-01 6.99621260e-01
-3.83518964e-01 4.07399267e-01 1.95608944e-01 2.66395181e-01
-1.13213561e-01 2.20480129e-01 5.10960147e-02 -1.11617371e-01
-1.69816315e-01 8.86338174e-01 -1.15209901e+00 1.75345528e+00
-5.16563177e-01 6.57724261e-01 3.58592510e-01 -6.44550741e-01
9.86814618e-01 8.03152382e-01 2.29371041e-01 -3.55854005e-01
-1.92944091e-02 2.51991183e-01 1.81988478e-01 -3.28019589e-01
6.15762234e-01 -4.22102123e-01 1.73473001e-01 3.69349867e-01
3.19286883e-01 -4.38564688e-01 -2.14726463e-01 -1.91057310e-01
7.55828619e-01 1.79095864e-01 3.82140055e-02 7.59711489e-02
4.58809614e-01 -4.46248442e-01 6.42207742e-01 4.51507688e-01
4.67167906e-02 8.79580081e-01 1.45259663e-01 -5.74570894e-02
-1.26985037e+00 -1.29315722e+00 2.01147974e-01 1.37811685e+00
-4.99945134e-01 -6.05783343e-01 -9.41356361e-01 9.46966708e-02
-1.81789204e-01 9.25614059e-01 -1.31204233e-01 -2.56757438e-01
-7.66405880e-01 1.11440703e-01 6.95884347e-01 4.11540866e-01
-6.83403611e-02 -1.60883462e+00 -3.19329560e-01 5.42532802e-01
-3.28323483e-01 -7.06536829e-01 -1.18801725e+00 2.54504383e-01
-6.82393134e-01 -4.65079337e-01 -7.06451893e-01 -1.20675039e+00
-2.08191037e-01 -3.78644079e-01 8.68931770e-01 -2.07912698e-01
2.65400782e-02 2.56231390e-02 -2.77341485e-01 -2.05075532e-01
-6.25602841e-01 1.00029938e-01 5.82847178e-01 3.56035948e-01
-1.98795229e-01 -9.33698118e-01 -3.76275957e-01 6.73801452e-02
-6.84492350e-01 -5.16203139e-03 1.61259815e-01 1.05471325e+00
8.92822981e-01 -8.13405868e-03 6.72274411e-01 -1.94336295e-01
9.51100349e-01 -2.83399373e-01 -4.55991268e-01 -1.62857145e-01
-3.10539901e-01 -2.08041400e-01 1.03476381e+00 -8.49875152e-01
-5.09688318e-01 1.20092966e-01 -4.60792273e-01 -1.02004838e+00
2.40757167e-01 3.24403077e-01 -3.98032740e-02 4.45479959e-01
6.89646661e-01 5.28628647e-01 -1.11064330e-01 -6.83976650e-01
3.61878961e-01 1.09597516e+00 1.10984218e+00 -3.93757045e-01
6.69097066e-01 -9.89302993e-02 -5.91483116e-01 -1.13507831e+00
-7.39002228e-01 -2.42445394e-01 -3.07364672e-01 -2.09187463e-01
8.83637786e-01 -9.25118923e-01 -7.39643693e-01 2.82751203e-01
-1.25006497e+00 -4.91992980e-01 -8.05854976e-01 7.31255710e-01
-1.00060511e+00 1.26855463e-01 -1.00518537e+00 -1.10412037e+00
-4.75596040e-01 -8.98680449e-01 7.62892783e-01 -3.03307315e-03
-4.97758389e-01 -6.80858552e-01 3.68142724e-01 3.23446363e-01
3.50010782e-01 -1.41035154e-01 9.72732723e-01 -5.23585081e-01
-3.25286746e-01 1.09897610e-02 5.59333861e-01 1.00604916e+00
1.25437994e-02 -1.74034640e-01 -1.32471681e+00 -4.11420763e-01
2.56226599e-01 -5.91699302e-01 8.30926001e-01 6.65818036e-01
1.36334038e+00 -7.32491851e-01 4.11262006e-01 7.91478455e-01
8.64055097e-01 5.24414182e-01 4.02324229e-01 -3.37295443e-01
6.02727830e-01 4.92419928e-01 4.48770583e-01 4.39081877e-01
-7.93620758e-03 8.72846365e-01 2.73297906e-01 9.34529305e-02
-3.97393554e-01 -8.24453235e-01 8.85528564e-01 1.93738997e+00
-3.12212594e-02 -2.16882855e-01 -3.11001003e-01 7.55613267e-01
-1.61793697e+00 -1.20180058e+00 4.15195793e-01 2.19025397e+00
1.09589601e+00 2.48607285e-02 2.09534004e-01 5.68541467e-01
6.66372657e-01 4.44573641e-01 -5.87003767e-01 -7.70062983e-01
-4.08760980e-02 5.10062933e-01 -1.50733218e-01 6.74192011e-01
-8.93353820e-01 9.83217418e-01 6.05477905e+00 1.17944491e+00
-1.18537986e+00 2.26885542e-01 2.00136900e-01 -5.98931849e-01
-3.54491442e-01 -2.12653145e-01 -4.20836836e-01 4.96326417e-01
1.30021846e+00 1.82455838e-01 1.18245912e+00 8.42147887e-01
4.51640815e-01 6.55523956e-01 -1.21437407e+00 1.15636921e+00
2.60164514e-02 -1.56989336e+00 -4.90330048e-02 -2.65818566e-01
6.75438285e-01 -7.89530203e-02 4.16770071e-01 4.72846150e-01
-1.69771001e-01 -1.07823944e+00 1.33901370e+00 5.16394138e-01
1.07288361e+00 -7.01004148e-01 2.04356372e-01 3.95369291e-01
-1.32569051e+00 -2.63982564e-01 -1.91753551e-01 -2.18837857e-01
1.53507590e-01 1.39729660e-02 -8.71904969e-01 -2.18135230e-02
4.37911421e-01 4.95034039e-01 1.90345481e-01 7.35597670e-01
-2.82633394e-01 1.50225759e+00 -9.76066291e-02 7.64991716e-02
2.35721052e-01 -1.49810553e-01 8.56410742e-01 1.13880515e+00
3.23586315e-01 5.34352511e-02 2.93673631e-02 1.11230469e+00
-2.06847534e-01 1.72396466e-01 -1.98448971e-01 -4.11933929e-01
6.72691703e-01 8.09650362e-01 4.76958007e-02 -1.62975006e-02
3.61484401e-02 1.16938281e+00 1.23700432e-01 3.38844150e-01
-5.60166299e-01 -4.70023602e-01 7.16219485e-01 1.38196990e-01
6.19530678e-01 -2.93452322e-01 -8.71519521e-02 -7.90466428e-01
-1.46565005e-01 -9.48133051e-01 1.85147539e-01 -9.59922194e-01
-1.05996525e+00 7.81758249e-01 -5.44027627e-01 -1.31238914e+00
-8.45151901e-01 -3.69362831e-01 -7.90342510e-01 1.08238816e+00
-1.14556777e+00 -7.49587536e-01 4.69727784e-01 5.20543933e-01
9.67688084e-01 -6.72825813e-01 9.59752560e-01 3.21016252e-01
-3.31989199e-01 5.99653721e-01 1.65938303e-01 1.87771812e-01
4.70120460e-01 -1.26724482e+00 4.97989863e-01 5.01736820e-01
9.19861495e-01 3.72358978e-01 7.94435084e-01 -2.89145470e-01
-1.02953207e+00 -1.03668535e+00 1.02000690e+00 -1.88175917e-01
5.41586339e-01 -5.40123999e-01 -8.47399831e-01 3.65775466e-01
2.15748698e-01 -1.45778522e-01 7.69394398e-01 -2.36275092e-01
-1.82424977e-01 -2.48255417e-01 -5.36447406e-01 5.68860352e-01
6.33547604e-01 -1.31892788e+00 -9.00317311e-01 1.60792828e-01
1.09936345e+00 -3.62454414e-01 -7.57655740e-01 4.81626391e-02
6.49421930e-01 -8.82024407e-01 8.88016045e-01 -5.98607659e-01
5.48781157e-01 -4.04145151e-01 -5.21283567e-01 -1.36284268e+00
-4.42199945e-01 -8.76136601e-01 -3.85981709e-01 1.08983171e+00
4.30003792e-01 1.96089208e-01 5.02714157e-01 -1.37437388e-01
-4.60463732e-01 -6.48217022e-01 -1.22599602e+00 -7.88225353e-01
-3.86594571e-02 -6.59965694e-01 2.79881179e-01 9.23306704e-01
-1.42625254e-03 5.76060355e-01 -8.88949037e-01 1.04640290e-01
3.37172896e-01 1.91065535e-01 2.04845905e-01 -9.31596339e-01
-1.08670437e+00 -3.43529999e-01 3.88456993e-02 -1.08421743e+00
1.56508267e-01 -1.06919038e+00 2.79886752e-01 -9.45733845e-01
-3.37127209e-01 8.24579895e-02 -6.62688673e-01 4.02465373e-01
2.52873033e-01 3.30884337e-01 4.65209275e-01 3.75151038e-01
-2.04695791e-01 9.19688106e-01 1.06788862e+00 -2.02799499e-01
-6.81309700e-01 2.90914237e-01 -4.58187088e-02 7.56546438e-01
6.59672618e-01 -5.49217880e-01 -3.53743345e-01 -2.25998655e-01
-8.88522938e-02 7.73308933e-01 2.82870740e-01 -1.23166907e+00
3.87149185e-01 1.28781060e-02 -4.01706696e-02 -6.52848959e-01
9.16699290e-01 -3.95471036e-01 1.29933178e-01 5.15156329e-01
-8.89189959e-01 -2.96862811e-01 -5.50973276e-03 4.40853626e-01
-5.27447820e-01 -1.35508120e-01 8.09742272e-01 1.32773563e-01
-3.85689408e-01 1.81124762e-01 -3.93272370e-01 5.12655228e-02
4.39435780e-01 -1.06970832e-01 4.93052572e-01 -7.06696272e-01
-8.35195184e-01 -2.90585816e-01 -1.39056012e-01 5.91194034e-01
8.48034322e-01 -1.54140174e+00 -9.19758320e-01 6.25538468e-01
-9.47257206e-02 -3.48702967e-01 2.20223859e-01 2.87123233e-01
-2.14706272e-01 4.67903882e-01 3.64768356e-02 -4.20782983e-01
-1.15674281e+00 2.47462377e-01 4.68307734e-01 1.17667116e-01
-6.02088869e-01 1.14307570e+00 2.77863927e-02 -4.57119912e-01
6.62558317e-01 -4.11714971e-01 -2.32788235e-01 -1.65373161e-02
3.94084334e-01 3.35058749e-01 -1.66445971e-01 -6.14172637e-01
-2.75257602e-02 3.22763681e-01 3.54806930e-01 -6.35974467e-01
1.29084516e+00 -4.28437144e-02 5.67833781e-02 1.02386761e+00
1.05314457e+00 4.07540500e-01 -1.51703227e+00 -4.40062173e-02
-1.59649357e-01 -5.62866852e-02 2.81192273e-01 -7.29651630e-01
-9.26225483e-01 1.10244906e+00 4.53797311e-01 1.26885056e-01
1.14161313e+00 -1.26252756e-01 1.17601681e+00 7.38459006e-02
-7.39373490e-02 -1.30325747e+00 1.87315553e-01 7.75662482e-01
1.13979721e+00 -5.24757862e-01 -6.81128442e-01 3.58932614e-01
-9.81382787e-01 9.66668546e-01 1.81327641e-01 -3.59144270e-01
5.40712416e-01 1.35897130e-01 9.75825563e-02 2.62892127e-01
-9.65628982e-01 1.59869075e-01 6.31307065e-01 3.42656523e-01
3.91782463e-01 2.08717123e-01 -1.01814240e-01 1.18781030e+00
-8.06213975e-01 -2.63644844e-01 2.65133411e-01 1.95647150e-01
-6.12376273e-01 -9.05215442e-01 -4.37153459e-01 1.35450184e-01
-4.21591580e-01 -3.88650805e-01 -4.42403018e-01 5.32360189e-02
2.55176723e-01 9.59010005e-01 2.62274623e-01 -9.69961524e-01
3.41723830e-01 4.08046186e-01 1.76690891e-01 -5.51171064e-01
-9.42693293e-01 4.46650803e-01 -2.18694389e-01 -4.55931097e-01
9.01548117e-02 -3.92751426e-01 -1.15778053e+00 1.64492160e-01
-2.41748661e-01 5.40389359e-01 7.19125032e-01 6.52282536e-01
1.50790915e-01 8.94871473e-01 1.00652361e+00 -8.27567220e-01
-1.02653205e+00 -9.06182230e-01 -1.06261778e+00 2.91665703e-01
6.94315910e-01 -5.74215949e-02 -5.38856149e-01 2.54511923e-01] | [15.526739120483398, 6.142667770385742] |
a44860e2-45d1-411c-980c-4c13eb42a5a0 | alzheimers-disease-diagnostics-by-a-deeply | 1607.00556 | null | http://arxiv.org/abs/1607.00556v1 | http://arxiv.org/pdf/1607.00556v1.pdf | Alzheimer's Disease Diagnostics by a Deeply Supervised Adaptable 3D Convolutional Network | Early diagnosis, playing an important role in preventing progress and
treating the Alzheimer's disease (AD), is based on classification of features
extracted from brain images. The features have to accurately capture main
AD-related variations of anatomical brain structures, such as, e.g., ventricles
size, hippocampus shape, cortical thickness, and brain volume. This paper
proposes to predict the AD with a deep 3D convolutional neural network
(3D-CNN), which can learn generic features capturing AD biomarkers and adapt to
different domain datasets. The 3D-CNN is built upon a 3D convolutional
autoencoder, which is pre-trained to capture anatomical shape variations in
structural brain MRI scans. Fully connected upper layers of the 3D-CNN are then
fine-tuned for each task-specific AD classification. Experiments on the
\emph{ADNI} MRI dataset with no skull-stripping preprocessing have shown our
3D-CNN outperforms several conventional classifiers by accuracy and robustness.
Abilities of the 3D-CNN to generalize the features learnt and adapt to other
domains have been validated on the \emph{CADDementia} dataset. | ['Ayman El-Baz', "Georgy Gimel'farb", 'Ehsan Hosseini-Asl'] | 2016-07-02 | null | null | null | null | ['skull-stripping'] | ['medical'] | [-2.99342424e-01 1.76295012e-01 3.26826304e-01 -9.08150852e-01
-3.21524501e-01 -1.86870415e-02 3.44994396e-01 1.72694139e-02
-4.82464731e-01 5.98965168e-01 5.66218495e-02 -6.51577860e-02
-2.87570983e-01 -7.91439772e-01 -4.94160384e-01 -4.94485676e-01
-7.49393880e-01 9.17286456e-01 4.98292059e-01 -7.92813674e-02
-1.73990801e-01 8.75274897e-01 -1.31971192e+00 5.18570840e-01
4.54403043e-01 1.45743692e+00 3.32894772e-01 4.62542921e-01
-7.90065080e-02 2.93234378e-01 -4.25972611e-01 -6.97673187e-02
2.63095945e-01 -6.04363494e-02 -7.57237434e-01 1.32844910e-01
4.08256918e-01 -6.20406151e-01 -2.70304054e-01 9.00089681e-01
6.59921944e-01 -5.60238540e-01 8.57707083e-01 -8.32129002e-01
-7.50250638e-01 3.59052569e-01 -4.92874533e-02 8.96278441e-01
-1.79445907e-01 2.17817482e-02 6.82298481e-01 -7.13348806e-01
4.58376080e-01 1.30407369e+00 1.03310740e+00 7.47477233e-01
-9.69960093e-01 -5.48962355e-01 -7.90656880e-02 5.26634097e-01
-1.02129304e+00 -6.23642616e-02 4.91300911e-01 -9.12167370e-01
7.59577155e-01 2.52393316e-02 1.11216521e+00 1.13195550e+00
6.59962654e-01 4.79037285e-01 1.35735214e+00 -7.75511935e-02
3.96676600e-01 -1.86948463e-01 4.47871745e-01 7.60970414e-01
4.17916209e-01 7.74253383e-02 2.48935252e-01 -1.66519895e-01
9.03060496e-01 2.32620999e-01 -1.50664791e-01 -2.74843276e-01
-1.09181011e+00 9.30429876e-01 8.32022071e-01 4.25779283e-01
-5.61084092e-01 -1.98557064e-01 5.23774624e-01 3.03359747e-01
5.93875349e-01 1.76807165e-01 -9.43829238e-01 5.50047159e-01
-7.75976837e-01 3.83099586e-01 4.17870224e-01 4.17677969e-01
4.09499735e-01 -7.86925554e-02 -1.42850325e-01 1.10037100e+00
5.71259260e-01 3.93409401e-01 1.24317908e+00 -5.32316923e-01
1.23494230e-01 1.08242905e+00 -6.21843219e-01 -4.47360307e-01
-1.16315889e+00 -5.40116131e-01 -1.03913975e+00 4.98823076e-01
3.01388830e-01 -2.20570803e-01 -1.14278340e+00 1.25106204e+00
2.87164360e-01 -4.15341079e-01 -2.84595579e-01 1.03337181e+00
1.04408801e+00 5.77477664e-02 9.26166177e-02 1.65832058e-01
1.52077448e+00 -4.66491312e-01 -2.06636012e-01 -4.43778783e-01
5.54677427e-01 -8.71692374e-02 7.76314497e-01 1.76740691e-01
-7.26092279e-01 -4.51799154e-01 -1.03627217e+00 -6.20212629e-02
-6.38370037e-01 2.81587094e-01 3.98688793e-01 5.99611759e-01
-1.23165727e+00 5.31717420e-01 -9.87235010e-01 -4.57928419e-01
1.11552763e+00 6.02258325e-01 -6.65151298e-01 -1.24859335e-02
-1.16905248e+00 1.10862303e+00 5.46313465e-01 6.33090362e-02
-9.33982849e-01 -8.45748067e-01 -5.58668733e-01 3.18478085e-02
-4.39279437e-01 -9.27977860e-01 1.02023780e+00 -1.04782784e+00
-1.29133701e+00 1.23461330e+00 4.72253025e-01 -7.97160149e-01
3.96558404e-01 -8.81018564e-02 -6.03381336e-01 3.45647335e-01
9.06323493e-02 7.43757844e-01 7.24256456e-01 -7.95134544e-01
-2.98708767e-01 -1.21419311e+00 -3.25796396e-01 -3.55504006e-01
-4.21647012e-01 -2.46048160e-02 3.72829437e-01 -6.05494082e-01
3.23537678e-01 -8.86374831e-01 -3.91051441e-01 4.26316023e-01
-2.51156986e-01 -4.34826836e-02 9.62055087e-01 -9.53971326e-01
5.98090649e-01 -1.81936324e+00 -1.52163863e-01 1.88082725e-01
7.06868112e-01 4.24032122e-01 6.52992353e-02 -4.03003991e-01
-4.41804647e-01 -8.44926909e-02 -3.41480911e-01 7.70202950e-02
-1.13064796e-01 2.71662861e-01 4.20277566e-01 4.80962127e-01
4.33336109e-01 8.29048753e-01 -3.39513600e-01 -3.05923402e-01
5.15213087e-02 6.83188736e-01 -5.89279532e-01 1.61056817e-01
-2.27599263e-01 4.37017649e-01 -6.63918734e-01 3.94774765e-01
6.31544590e-01 -2.94007540e-01 -6.81654066e-02 -1.46718919e-01
1.36259198e-01 -2.26190668e-02 -5.17442465e-01 1.18949795e+00
-3.11011940e-01 5.12260497e-01 1.19592771e-02 -1.46601665e+00
1.09546816e+00 2.73238331e-01 5.29604137e-01 -9.10160482e-01
5.42589247e-01 3.53781521e-01 3.78976852e-01 -7.21170425e-01
-7.60833859e-01 -4.63059396e-02 3.61370921e-01 2.54248440e-01
3.84110421e-01 4.23767954e-01 4.52043787e-02 -4.32978302e-01
1.11637783e+00 -4.20664579e-01 3.41086477e-01 -5.50769567e-01
7.59927213e-01 -1.65517122e-01 3.79276037e-01 2.00312868e-01
-3.90611529e-01 5.86786270e-01 5.66872895e-01 -1.13582134e+00
-1.16417837e+00 -1.08220339e+00 -6.91628158e-01 6.80355787e-01
-5.47151804e-01 1.32056460e-01 -8.63606215e-01 -8.01711619e-01
1.52568549e-01 1.34769827e-01 -9.41125512e-01 -2.62919903e-01
-6.01546168e-01 -1.22807658e+00 4.70162094e-01 7.63922870e-01
7.42164731e-01 -9.22440588e-01 -8.71683180e-01 1.91840708e-01
4.99087155e-01 -9.75731313e-01 4.84019555e-02 4.08275783e-01
-1.53305829e+00 -1.47887290e+00 -1.08817244e+00 -1.06212926e+00
7.39933670e-01 -5.36886692e-01 1.02585757e+00 -6.50503039e-02
-7.02753901e-01 2.62666672e-01 -2.05159619e-01 -6.26777172e-01
-3.88642758e-01 1.75589725e-01 3.20114233e-02 1.17210768e-01
6.04214489e-01 -1.05389738e+00 -8.03910077e-01 4.84650694e-02
-6.11048520e-01 -4.25533563e-01 9.78618860e-01 6.19164705e-01
7.85770357e-01 -6.89536259e-02 6.79635465e-01 -6.84566259e-01
6.80326045e-01 -4.59039807e-01 -1.79375485e-01 -2.51744650e-02
-6.43970251e-01 -1.30479503e-02 5.56495309e-01 -2.99265712e-01
-6.25252128e-01 1.52144372e-01 -6.28307045e-01 -1.34924918e-01
-7.02080131e-01 2.66320407e-01 -4.56746042e-01 6.44059666e-03
7.35632539e-01 1.41307890e-01 3.52418423e-01 -8.03395748e-01
-5.57300895e-02 7.20874310e-01 3.08134794e-01 -1.85524464e-01
3.32789272e-01 3.46799701e-01 1.60567418e-01 -8.12215269e-01
-9.12193120e-01 -7.59441853e-02 -1.23450994e+00 -7.59122968e-02
1.13586783e+00 -7.65261948e-01 -1.72048151e-01 7.01283574e-01
-8.62612009e-01 -5.15881836e-01 -1.56146556e-01 6.23475552e-01
-4.02277857e-01 -2.74761412e-02 -6.40126407e-01 -8.31355825e-02
-7.89416492e-01 -1.23493862e+00 8.83248508e-01 3.41748670e-02
-9.68003795e-02 -1.27965486e+00 2.37745330e-01 1.86944231e-01
5.69222808e-01 3.00870061e-01 1.46799183e+00 -1.25219512e+00
7.39636943e-02 -3.34597945e-01 -4.42975074e-01 6.95181966e-01
4.52978052e-02 -4.73565370e-01 -8.56405795e-01 -4.43142280e-02
1.55199608e-02 -1.81234032e-01 8.95188510e-01 7.12511897e-01
1.34236121e+00 -4.61515546e-01 -2.98473418e-01 5.34162641e-01
1.12089527e+00 1.79928184e-01 5.51336765e-01 7.44736373e-01
3.80771935e-01 2.37369224e-01 -2.82492191e-01 2.99395233e-01
4.13279712e-01 4.64566380e-01 7.43412316e-01 -7.83971548e-02
-4.32501674e-01 6.07267082e-01 1.95237622e-01 5.60207486e-01
-1.77120805e-01 4.57253486e-01 -1.00537384e+00 4.08349454e-01
-1.25757492e+00 -6.05102003e-01 -2.81455487e-01 1.60220695e+00
8.45678329e-01 2.26073548e-01 3.83058578e-01 2.22564816e-01
6.56005204e-01 -1.70156687e-01 -6.05336547e-01 -3.11446905e-01
-1.18998960e-01 7.07895458e-01 2.62526304e-01 -8.71076584e-02
-1.27613938e+00 3.46047252e-01 6.05093288e+00 -3.18304300e-02
-1.42028224e+00 4.10006404e-01 6.41256034e-01 7.73142427e-02
2.75793612e-01 -7.71983624e-01 -6.53835952e-01 4.64144170e-01
9.42117095e-01 1.85171813e-01 8.07475522e-02 1.05993307e+00
7.99858570e-02 4.88131315e-01 -1.12979364e+00 7.08420753e-01
6.84300661e-02 -1.02299774e+00 2.21959621e-01 1.46180809e-01
3.54940534e-01 3.31217468e-01 6.04637377e-02 3.59592766e-01
-7.13294297e-02 -1.19377649e+00 4.26149309e-01 7.98598707e-01
6.24961495e-01 -5.13528347e-01 1.02340627e+00 2.05921177e-02
-7.63548851e-01 -4.84308898e-01 -3.82898718e-01 9.69209746e-02
-1.06449939e-01 1.03141677e+00 -1.10524905e+00 -1.30224535e-02
1.23280847e+00 7.34118104e-01 -1.02160311e+00 1.23341584e+00
-4.79187965e-02 6.12525761e-01 -1.02728628e-01 1.58114776e-01
1.33780912e-01 1.16644381e-02 3.66706967e-01 1.25125504e+00
5.56086540e-01 9.52462703e-02 4.58431877e-02 9.97985065e-01
7.79012591e-02 2.15658665e-01 -2.70939350e-01 9.21915472e-02
-2.15851530e-01 1.13885820e+00 -5.98818302e-01 -1.79765314e-01
-2.99503744e-01 7.83332348e-01 1.12319514e-01 4.60621044e-02
-3.87465268e-01 -7.23911747e-02 7.16224849e-01 5.40119588e-01
5.29313266e-01 -9.06612277e-02 -4.76476252e-01 -8.07476580e-01
-7.07765222e-02 -6.51390791e-01 4.55570400e-01 -8.02210152e-01
-1.70037174e+00 9.47808981e-01 -1.77725419e-01 -7.25492120e-01
8.21199417e-02 -1.37872267e+00 -7.18889892e-01 6.66353464e-01
-1.35701406e+00 -1.21732306e+00 -4.53972846e-01 9.30167973e-01
3.88822287e-01 -7.66047060e-01 1.09882474e+00 4.81836289e-01
-5.88951588e-01 2.45173261e-01 -5.14987856e-02 5.19880354e-01
5.12719154e-01 -1.36991441e+00 1.47028953e-01 1.50599211e-01
-3.55408251e-01 3.39040190e-01 1.65136933e-01 -5.30371308e-01
-7.32283890e-01 -1.69659710e+00 7.59501159e-01 -6.38089553e-02
5.89084685e-01 -8.85709152e-02 -8.67393672e-01 8.65845799e-01
-2.08369002e-01 5.82695067e-01 8.15093458e-01 -1.94379658e-01
-4.48195547e-01 -4.50852633e-01 -1.38546681e+00 1.42012089e-01
9.92614090e-01 -1.82533100e-01 -1.09448934e+00 5.80155492e-01
2.84496933e-01 -1.31182104e-01 -1.27118134e+00 3.75850379e-01
6.28160655e-01 -9.47189808e-01 1.34066880e+00 -8.95953238e-01
4.93729889e-01 2.71703035e-01 -8.62877592e-02 -1.43327343e+00
-5.06228685e-01 4.59747732e-01 -2.16367170e-01 6.89173400e-01
2.84050643e-01 -7.30465770e-01 5.05201817e-01 3.21919471e-01
-3.85778517e-01 -9.46906447e-01 -1.11335659e+00 -7.27681100e-01
6.04406893e-01 -1.91801623e-01 6.98723376e-01 7.30237186e-01
-5.04590333e-01 2.39077911e-01 3.81626546e-01 3.12617481e-01
5.75237036e-01 -3.45160455e-01 4.75032404e-02 -2.22215676e+00
1.70536280e-01 -6.19987130e-01 -1.05326986e+00 -2.53274292e-01
2.81115592e-01 -1.39107585e+00 -4.89977628e-01 -1.57069302e+00
1.18540324e-01 -3.61851394e-01 -4.71434414e-01 6.20222032e-01
2.21278325e-01 2.97379196e-01 -1.86613351e-01 -3.91029790e-02
-1.79026112e-01 5.20673275e-01 1.43447137e+00 -4.50715840e-01
-2.35398278e-01 1.90666467e-01 -3.81926537e-01 9.27577436e-01
1.06136119e+00 -4.28033412e-01 -2.23300770e-01 -2.69392729e-01
-2.24376410e-01 -4.87925559e-01 9.41116571e-01 -1.49964321e+00
-2.52717674e-01 4.27902550e-01 1.29011178e+00 -3.04647833e-01
1.50980935e-01 -9.00559783e-01 -1.38995349e-01 7.04726756e-01
-3.06542844e-01 5.15984334e-02 1.61224768e-01 1.87311307e-01
2.24325527e-02 -1.34525120e-01 9.71025169e-01 -5.28798461e-01
-4.76470679e-01 7.16407716e-01 -5.56310773e-01 2.84473170e-02
9.13170695e-01 -2.34881967e-01 2.38788519e-02 2.30018288e-01
-1.25024951e+00 -1.00964338e-01 -9.19229537e-02 2.49669150e-01
6.00110948e-01 -1.40829897e+00 -8.12534511e-01 3.41540724e-01
1.68461725e-02 4.19382676e-02 1.66366488e-01 9.42673922e-01
-6.67561293e-01 5.53018093e-01 -1.06566048e+00 -6.88548744e-01
-1.15500188e+00 3.82644892e-01 9.61091042e-01 -2.47692347e-01
-1.07663679e+00 7.12346375e-01 4.67995629e-02 -4.61123586e-01
2.75884002e-01 -6.51005089e-01 -6.97286665e-01 3.86484452e-02
7.28941321e-01 2.34125733e-01 5.53431749e-01 -4.67670321e-01
-5.46180189e-01 4.26461667e-01 -3.20035785e-01 3.80226612e-01
2.09891152e+00 9.07130688e-02 -3.13365847e-01 1.26069769e-01
1.37344468e+00 -8.41582417e-01 -1.04258978e+00 -5.04146144e-02
1.75485358e-01 3.24158967e-01 4.56239939e-01 -1.07743013e+00
-1.74001324e+00 1.07381928e+00 1.57034516e+00 3.70295830e-02
9.59114254e-01 2.39555061e-01 1.03019845e+00 4.63548481e-01
2.21657768e-01 -8.60853732e-01 -4.96062916e-03 4.18468297e-01
1.31438303e+00 -9.98150766e-01 -7.08134100e-02 1.92182228e-01
-2.26363599e-01 1.88622510e+00 6.18925452e-01 -4.62199658e-01
1.25115108e+00 8.43633264e-02 1.88152134e-01 -5.74765146e-01
-1.63532570e-01 -1.08359732e-01 4.47490603e-01 9.24323440e-01
2.26240829e-01 1.43800154e-01 -2.58490950e-01 1.49751055e+00
-2.87455559e-01 2.91341633e-01 1.43594995e-01 5.83202124e-01
-6.64874136e-01 -9.99785721e-01 -4.05335665e-01 9.45269048e-01
-4.24602389e-01 2.16680005e-01 -6.33103371e-01 9.35171366e-01
7.63866842e-01 8.88672546e-02 3.48550141e-01 -1.47104695e-01
4.30831522e-01 4.46467519e-01 5.86550236e-01 -5.56902170e-01
-5.38114011e-01 -4.01385903e-01 -1.09271325e-01 -4.79234576e-01
-3.66906464e-01 -7.10031807e-01 -1.28469801e+00 2.06485018e-01
2.95603514e-01 -4.95964438e-01 5.32064676e-01 1.15579641e+00
3.76828462e-01 8.10418427e-01 2.43834525e-01 -9.52959538e-01
-5.37463725e-01 -1.15993488e+00 -8.37496400e-01 1.93358257e-01
4.49202001e-01 -8.89527142e-01 -1.38467208e-01 2.04498753e-01] | [14.185647964477539, -1.7849012613296509] |
a84330d7-3306-4780-aa23-6cb25f4c9783 | robustness-of-sam-segment-anything-under | 2306.07713 | null | https://arxiv.org/abs/2306.07713v1 | https://arxiv.org/pdf/2306.07713v1.pdf | Robustness of SAM: Segment Anything Under Corruptions and Beyond | Segment anything model (SAM), as the name suggests, is claimed to be capable of cutting out any object. SAM is a vision foundation model which demonstrates impressive zero-shot transfer performance with the guidance of a prompt. However, there is currently a lack of comprehensive evaluation of its robustness performance under various types of corruptions. Prior works show that SAM is biased towards texture (style) rather than shape, motivated by which we start by investigating SAM's robustness against style transfer, which is synthetic corruption. With the effect of corruptions interpreted as a style change, we further evaluate its robustness on 15 common corruptions with 5 severity levels for each real-world corruption. Beyond the corruptions, we further evaluate the SAM robustness on local occlusion and adversarial perturbations. Overall, this work provides a comprehensive empirical study on the robustness of the SAM under corruptions and beyond. | ['Choong Seon Hong', 'Chenshuang Zhang', 'Shehbaz Tariq', 'Donghun Kim', 'Taegoo Kang', 'Chaoning Zhang', 'Yu Qiao'] | 2023-06-13 | null | null | null | null | ['style-transfer'] | ['computer-vision'] | [ 4.77720261e-01 -6.52713999e-02 7.55428150e-02 -6.26839921e-02
-8.20636749e-01 -9.50613797e-01 9.22317863e-01 -5.01215994e-01
2.19782647e-02 4.57539558e-01 3.72249395e-01 -3.59969050e-01
2.23136291e-01 -6.21686041e-01 -9.28015888e-01 -5.54014564e-01
3.73932660e-01 -1.31342590e-01 2.42716298e-01 -3.19539130e-01
4.91016030e-01 5.09851217e-01 -1.22249901e+00 2.70442605e-01
8.99346709e-01 7.50334501e-01 -3.13471228e-01 7.58657873e-01
4.38098222e-01 6.50075734e-01 -9.54767048e-01 -9.04659033e-01
6.61861360e-01 -3.62546682e-01 -7.42334783e-01 3.63456339e-01
1.01802015e+00 -7.51518965e-01 -3.70282620e-01 1.10296977e+00
5.28586626e-01 -1.51384398e-01 6.93110824e-01 -1.35293436e+00
-9.53723013e-01 7.64908940e-02 -5.39497614e-01 -7.52338693e-02
5.06977558e-01 8.42494786e-01 7.56447792e-01 -8.71770144e-01
8.50375414e-01 1.54944408e+00 9.06843126e-01 6.01774812e-01
-1.40509164e+00 -4.19767201e-01 -1.15067847e-01 -1.42327547e-01
-9.50299084e-01 -5.86741865e-01 7.92251348e-01 -5.35498023e-01
5.79691470e-01 3.09394926e-01 4.05476749e-01 1.85005748e+00
3.25535089e-01 5.93923986e-01 1.46107531e+00 -2.70372272e-01
8.41862932e-02 9.19807181e-02 -5.15304059e-02 3.80391985e-01
3.93122584e-01 6.11393034e-01 -4.70747083e-01 -1.01786852e-01
7.76820600e-01 -6.65572405e-01 -2.95789510e-01 -1.98362961e-01
-1.21767008e+00 4.95661259e-01 5.18158317e-01 -3.60071719e-01
5.24712838e-02 4.16990191e-01 4.62343216e-01 4.95952398e-01
5.70807874e-01 4.43800926e-01 -3.67553711e-01 -1.53010800e-01
-7.36266732e-01 3.95791560e-01 6.25226736e-01 1.10684061e+00
4.86120403e-01 3.89559627e-01 -5.60746133e-01 7.24430025e-01
6.35567009e-02 8.15061390e-01 -1.18291572e-01 -1.20178151e+00
2.73056656e-01 2.62749225e-01 2.47245654e-01 -1.19538212e+00
-7.44350031e-02 -2.72896647e-01 -8.15772593e-01 7.91457951e-01
5.26012003e-01 -3.81779402e-01 -1.01036716e+00 1.80992246e+00
3.27366553e-02 1.61458969e-01 7.45941279e-03 7.92546630e-01
1.00445569e+00 2.27637589e-01 5.04176039e-03 5.25823496e-02
1.19428897e+00 -8.15045416e-01 -4.55321819e-01 -1.76452562e-01
2.84603566e-01 -1.04519951e+00 1.31882679e+00 2.54850477e-01
-8.80720317e-01 -4.85603124e-01 -8.48390639e-01 -7.59442672e-02
-2.36475512e-01 -1.77161470e-01 3.93993884e-01 8.88273954e-01
-1.03578675e+00 3.93549800e-01 -4.05629516e-01 -4.79417026e-01
7.57510126e-01 -2.75530487e-01 -5.21878183e-01 -2.37293541e-01
-1.02617228e+00 1.14740682e+00 -1.09831423e-01 -1.19127631e-01
-1.15539968e+00 -8.95841956e-01 -7.35164881e-01 -4.13844705e-01
2.76534855e-01 -1.04966283e+00 9.27748144e-01 -1.25812709e+00
-1.53984272e+00 1.06572998e+00 -1.54001072e-01 -2.54874259e-01
1.13004065e+00 -5.08143783e-01 -2.43482456e-01 5.12880348e-02
6.21691197e-02 7.65590310e-01 1.27165568e+00 -1.81807446e+00
3.00445613e-02 -9.86476615e-02 3.60674381e-01 2.02001870e-01
-8.77530947e-02 3.16496491e-01 -3.60789716e-01 -1.09667087e+00
-2.34875694e-01 -1.21898472e+00 1.25202909e-01 3.32416266e-01
-7.83265710e-01 1.52135000e-01 1.06607461e+00 -6.28428638e-01
7.79783607e-01 -2.35023093e+00 -8.11389387e-02 -8.01520795e-02
4.44062576e-02 3.20986181e-01 -4.16242510e-01 3.21969867e-01
-8.90643522e-02 4.97234523e-01 -3.39375019e-01 -3.44087154e-01
-6.83077648e-02 1.44689322e-01 -7.54437447e-01 5.59559405e-01
4.20949012e-01 1.29058766e+00 -7.82629728e-01 -2.98303783e-01
1.65755734e-01 5.58958411e-01 -7.10869849e-01 5.62091619e-02
-7.42865726e-02 6.38748407e-01 -1.42356411e-01 9.52830553e-01
8.46106768e-01 8.76374692e-02 -4.25622940e-01 -4.10664767e-01
1.55487821e-01 -2.73738772e-01 -7.13312924e-01 1.51691592e+00
3.32929827e-02 9.67440844e-01 -3.13941799e-02 -3.47308666e-01
7.46839702e-01 2.48315677e-01 1.82639897e-01 -4.92036641e-01
-3.52489878e-03 7.48290345e-02 -1.66360259e-01 -4.59892541e-01
5.30902863e-01 -1.26244888e-01 -1.90726474e-01 3.21392238e-01
-2.34280199e-01 -6.19481862e-01 -1.74820885e-01 3.63418788e-01
1.26923418e+00 4.02758628e-01 -1.09935850e-01 -2.89463311e-01
8.08251649e-02 2.20215470e-02 4.90377158e-01 1.02253580e+00
-5.05894125e-01 1.28638589e+00 5.20711660e-01 -2.91532099e-01
-1.15758240e+00 -1.43747795e+00 -6.00749962e-02 8.61283720e-01
5.14621675e-01 -1.36997774e-01 -9.60931420e-01 -7.53420889e-01
1.07915379e-01 7.80802310e-01 -8.01247239e-01 -3.09612662e-01
-3.69191349e-01 -6.74309194e-01 1.05536962e+00 3.60756338e-01
9.45215762e-01 -1.23946702e+00 -5.93831480e-01 -2.96267718e-01
-3.11107606e-01 -1.13783407e+00 -4.40658063e-01 -5.24247110e-01
-5.76368988e-01 -1.15340889e+00 -7.31735229e-01 -2.77573109e-01
6.24360979e-01 4.36880291e-01 1.38162577e+00 1.32493556e-01
-2.19689548e-01 8.29213202e-01 -2.64829636e-01 -6.25337005e-01
-5.51793158e-01 -2.16731772e-01 -1.07862870e-03 -1.02966484e-02
-3.08687478e-01 -5.09646475e-01 -6.83183134e-01 4.08929527e-01
-1.12072110e+00 8.86989012e-02 3.65011394e-01 6.58789694e-01
1.66147500e-01 -3.29008400e-01 4.71655756e-01 -9.69620883e-01
8.10235441e-01 -3.19076747e-01 -2.79440284e-01 1.99888960e-01
-2.68581659e-01 -5.05475044e-01 2.49734104e-01 -4.07630920e-01
-1.30661595e+00 -3.18602324e-01 1.71884358e-01 -4.86343026e-01
-5.51868796e-01 1.23550981e-01 -3.47306997e-01 -3.44074756e-01
9.26937103e-01 1.06304497e-01 -9.84354466e-02 -3.69686306e-01
7.02594340e-01 2.74894834e-01 7.76899695e-01 -8.07550430e-01
1.24428570e+00 7.02666938e-01 -1.69491451e-02 -8.48291934e-01
-9.14723337e-01 1.03640005e-01 -5.71639061e-01 -4.30897892e-01
7.26661444e-01 -8.41880739e-01 -4.36321706e-01 1.05900013e+00
-1.29500282e+00 -4.94794160e-01 -2.53642529e-01 -1.42882541e-01
-8.91078949e-01 5.09757996e-01 -6.40717328e-01 -4.78620559e-01
-1.23772532e-01 -1.13774228e+00 1.03041446e+00 1.68892574e-02
-3.78945440e-01 -9.22395706e-01 -8.67940709e-02 5.55749834e-01
7.00111806e-01 7.83341408e-01 7.41490543e-01 4.06545810e-02
-6.15371764e-01 -1.55499950e-01 -4.28907782e-01 5.43734014e-01
1.88048795e-01 4.14624125e-01 -1.19591141e+00 -5.75770199e-01
7.08186030e-02 -3.44936579e-01 7.87152052e-01 2.78513461e-01
8.88840437e-01 -4.79518801e-01 9.06287655e-02 9.43592250e-01
1.43007135e+00 -1.90741256e-01 1.19591463e+00 5.40937006e-01
8.46836448e-01 4.62845623e-01 5.08836508e-01 1.29972324e-01
1.38142444e-02 4.97367769e-01 6.94387376e-01 -3.58013749e-01
-6.88502192e-01 -4.86024879e-02 5.93936086e-01 1.71667337e-01
-3.34462345e-01 -5.20905495e-01 -7.30226576e-01 4.16200936e-01
-1.36969554e+00 -1.05171227e+00 -4.04864289e-02 1.94037771e+00
8.05480182e-01 1.67698681e-01 -1.39897570e-01 -1.29455671e-01
5.84501624e-01 4.05102849e-01 -6.20305419e-01 -4.53498393e-01
-5.10910571e-01 -1.32555664e-01 4.86489892e-01 3.32140476e-01
-1.14669347e+00 1.29440844e+00 7.86992359e+00 6.26875103e-01
-1.07135260e+00 -2.24489465e-01 7.26506889e-01 -3.56552340e-02
-2.56829351e-01 8.26443173e-03 -3.19995105e-01 6.08631492e-01
4.05217081e-01 -1.00400802e-02 4.79993790e-01 5.90224802e-01
3.08065861e-01 -4.33330089e-02 -9.38037515e-01 6.04992628e-01
2.15056166e-01 -1.12697518e+00 4.52870488e-01 -1.50438333e-02
1.15885854e+00 -1.56820938e-02 6.13912344e-01 9.75963324e-02
5.09249628e-01 -1.17402554e+00 9.05402780e-01 5.92882216e-01
1.17547560e+00 -4.47261363e-01 4.61362362e-01 7.14639202e-02
-5.04004419e-01 8.57643336e-02 -9.55819190e-02 1.20420985e-01
8.86969939e-02 3.87563974e-01 -4.09351796e-01 5.75850427e-01
6.01176322e-01 7.09573746e-01 -8.14291000e-01 7.84092128e-01
-4.63087440e-01 8.95687103e-01 8.20517261e-03 9.42845404e-01
4.74619977e-02 8.74853879e-02 1.04495776e+00 1.33244812e+00
1.38730824e-01 1.09549519e-02 -2.24735066e-02 1.01514781e+00
-1.72954455e-01 -4.40673739e-01 -8.86700571e-01 3.55586708e-01
5.03895283e-01 7.83129871e-01 -6.10279024e-01 -2.43237436e-01
-3.51505846e-01 1.32939255e+00 -1.79173648e-02 7.43275583e-01
-9.06048954e-01 -5.45123853e-02 1.11213100e+00 5.20959124e-02
1.11769609e-01 -1.67767942e-01 -8.81664455e-01 -1.14856446e+00
-1.29056247e-02 -1.20116222e+00 -9.91676301e-02 -1.19156826e+00
-1.44993889e+00 2.34732568e-01 -1.24951592e-02 -1.34285808e+00
2.35990569e-01 -4.74902928e-01 -8.08891714e-01 7.97090173e-01
-1.44286561e+00 -1.57783461e+00 -6.28186285e-01 6.18895054e-01
5.96754491e-01 -1.41415045e-01 7.81908274e-01 -1.88823700e-01
-4.69410986e-01 8.76385629e-01 -1.96343675e-01 3.04394722e-01
1.13478887e+00 -1.02454042e+00 7.93210089e-01 1.08571386e+00
-1.55596033e-01 7.12883890e-01 1.01380587e+00 -7.96592534e-01
-1.25574565e+00 -1.29124486e+00 2.98070699e-01 -9.51632440e-01
4.66937333e-01 3.26776020e-02 -8.49869788e-01 8.02861750e-01
3.16003323e-01 1.84951723e-01 2.44787931e-01 -3.27768296e-01
-8.42534065e-01 2.41145521e-01 -1.41598582e+00 8.95057857e-01
1.26925385e+00 -6.83181047e-01 -6.68020844e-01 4.38752621e-02
7.89365947e-01 -5.31240582e-01 -8.35547030e-01 6.66774333e-01
5.57008505e-01 -1.36309040e+00 1.27156830e+00 -5.84759653e-01
8.32400739e-01 -3.53817016e-01 -2.82315403e-01 -1.44438529e+00
-5.31962633e-01 -7.47809827e-01 2.67932774e-03 1.31556547e+00
1.70556773e-02 -6.71360016e-01 4.66945529e-01 5.61388910e-01
-1.29390135e-01 -4.94629413e-01 -7.50194013e-01 -1.02909553e+00
2.23503366e-01 -4.80758309e-01 4.50914562e-01 1.19094038e+00
-6.87321246e-01 1.20807841e-01 -6.22677028e-01 1.03920490e-01
8.32571924e-01 -1.92603573e-01 1.28296816e+00 -8.83784056e-01
-6.96093515e-02 -5.88815928e-01 -4.51330662e-01 -6.30814493e-01
5.69976568e-02 -4.94114906e-01 2.94921156e-02 -1.39842176e+00
1.91254705e-01 -2.26133600e-01 -1.74055353e-01 3.52167338e-01
-3.90922189e-01 7.74203360e-01 5.67746937e-01 3.36090028e-01
-3.05656672e-01 4.04517740e-01 1.52780187e+00 -1.97806880e-01
2.16498286e-01 -1.01180904e-01 -9.80793238e-01 8.75458777e-01
7.92848527e-01 -1.65682584e-01 -3.32798123e-01 -7.20742106e-01
2.08199501e-01 -3.00129712e-01 7.65722990e-01 -8.33667219e-01
-3.53507638e-01 -3.34296763e-01 2.13270709e-01 -1.63805202e-01
2.86924094e-01 -6.28261149e-01 2.58641064e-01 2.93737024e-01
-2.63376594e-01 -1.62751466e-01 4.63091880e-01 6.44013405e-01
1.03145450e-01 2.22324327e-01 1.04297030e+00 -1.36551559e-02
-6.77538157e-01 1.98630989e-01 -1.72397137e-01 3.08268219e-01
9.66857970e-01 -5.80853283e-01 -7.44171679e-01 -3.97013187e-01
-5.66485047e-01 -2.13005915e-01 1.07500648e+00 7.12836683e-01
5.32098055e-01 -1.36638021e+00 -1.01180577e+00 1.69545740e-01
1.76854208e-01 -1.02942318e-01 5.36109842e-02 6.18680418e-01
-4.18095320e-01 -1.27447307e-01 -3.67408752e-01 -6.68651402e-01
-1.39256716e+00 5.63356519e-01 3.19923133e-01 2.09833860e-01
-7.88584292e-01 8.62805128e-01 5.58249533e-01 -1.87622726e-01
9.65354517e-02 -4.41662297e-02 1.04615040e-01 -2.80032396e-01
3.41952771e-01 5.38249016e-01 -1.90424770e-01 -7.99013853e-01
-2.30060101e-01 7.16171145e-01 2.07393348e-01 -1.37973592e-01
1.01328611e+00 -2.33376324e-01 -5.48589565e-02 3.30620110e-01
8.96864474e-01 1.03820927e-01 -1.58615351e+00 -5.08050919e-02
-3.19466382e-01 -8.16782176e-01 -2.82316953e-01 -1.08782887e+00
-1.01707637e+00 5.82209170e-01 4.33590382e-01 1.74538344e-02
1.01837611e+00 -2.21292719e-01 6.96771801e-01 7.56907882e-03
3.42716306e-01 -8.04233372e-01 3.53347570e-01 7.22181857e-01
1.34301615e+00 -1.28942263e+00 4.52926122e-02 -5.30135095e-01
-7.24584520e-01 6.81976497e-01 6.87826693e-01 -2.80875236e-01
4.61666375e-01 2.90029526e-01 3.44432116e-01 -1.38538713e-02
-6.06028795e-01 1.23189926e-01 3.14650565e-01 1.01231050e+00
9.39451903e-02 1.62971973e-01 1.08558774e-01 2.01883744e-02
-3.77245635e-01 -1.54442921e-01 6.21505916e-01 7.98512399e-01
-1.04640715e-01 -6.69715106e-01 -6.90270603e-01 2.03448817e-01
-5.61482489e-01 -1.03046067e-01 -7.94041276e-01 5.83510756e-01
1.83889449e-01 1.06734884e+00 -1.53926238e-01 -3.93311024e-01
4.64540571e-01 -2.67170072e-01 6.30353153e-01 -3.89665037e-01
-7.33967185e-01 -3.19034010e-01 5.44070266e-02 -8.68872106e-01
-2.82538325e-01 -5.65705240e-01 -4.49158072e-01 -7.40690410e-01
-2.25859627e-01 -8.31728756e-01 3.91695231e-01 7.86721945e-01
3.80232811e-01 4.22055066e-01 6.30125999e-01 -9.43815947e-01
-4.27525789e-01 -8.88969839e-01 -1.79205641e-01 1.09911144e+00
4.18177783e-01 -6.53898418e-01 -5.12388766e-01 4.64443326e-01] | [8.168388366699219, -1.228469729423523] |
bdd21aa2-f14f-4610-8de0-a97a1c580398 | contextual-information-integration-for-stance | 2211.01874 | null | https://arxiv.org/abs/2211.01874v2 | https://arxiv.org/pdf/2211.01874v2.pdf | Contextual information integration for stance detection via cross-attention | Stance detection deals with identifying an author's stance towards a target. Most existing stance detection models are limited because they do not consider relevant contextual information which allows for inferring the stance correctly. Complementary context can be found in knowledge bases but integrating the context into pretrained language models is non-trivial due to the graph structure of standard knowledge bases. To overcome this, we explore an approach to integrate contextual information as text which allows for integrating contextual information from heterogeneous sources, such as structured knowledge sources and by prompting large language models. Our approach can outperform competitive baselines on a large and diverse stance detection benchmark in a cross-target setup, i.e. for targets unseen during training. We demonstrate that it is more robust to noisy context and can regularize for unwanted correlations between labels and target-specific vocabulary. Finally, it is independent of the pretrained language model in use. | ['Iryna Gurevych', 'Andreas Waldis', 'Tilman Beck'] | 2022-11-03 | null | null | null | null | ['stance-detection'] | ['natural-language-processing'] | [ 1.64977670e-01 3.12134880e-03 -9.58339572e-01 -1.94180578e-01
-1.04320145e+00 -1.09296131e+00 1.05831826e+00 6.16986990e-01
-5.91813862e-01 8.50314319e-01 7.00045228e-01 -2.85481632e-01
2.23009199e-01 -9.03605878e-01 -6.94055676e-01 -2.42048547e-01
3.33589882e-01 6.52616680e-01 7.08913386e-01 -5.71630239e-01
2.65182376e-01 2.13879794e-02 -1.11740911e+00 4.81923282e-01
6.48189723e-01 4.67170238e-01 -6.97510913e-02 4.96270031e-01
-2.53100812e-01 8.88094783e-01 -9.52764511e-01 -6.88098013e-01
-6.86574504e-02 -3.97623330e-01 -8.51374865e-01 -1.13599360e-01
6.52342618e-01 9.27061811e-02 -6.27650544e-02 8.90312850e-01
3.96337748e-01 -2.63231814e-01 8.53324652e-01 -9.55229580e-01
-6.34347737e-01 9.20555353e-01 -3.57890040e-01 4.67810273e-01
4.32892352e-01 -1.60604313e-01 1.47798491e+00 -8.41862500e-01
1.07994616e+00 1.25283933e+00 6.52797878e-01 2.95614868e-01
-1.25170970e+00 -3.24065536e-01 5.95892072e-01 3.08553785e-01
-8.39815795e-01 -3.97301823e-01 8.97691846e-01 -5.72553217e-01
7.42302597e-01 2.43970752e-01 4.04574931e-01 1.92980611e+00
-3.25061977e-01 8.01852584e-01 1.17680931e+00 -7.87976980e-01
-9.03941318e-02 3.28130543e-01 6.64022207e-01 4.90775943e-01
5.28448462e-01 -1.08121626e-01 -7.83637047e-01 -3.90901357e-01
3.46699029e-01 -4.10656929e-01 -1.59496725e-01 -2.83456385e-01
-1.44335616e+00 9.96076643e-01 2.79608756e-01 6.42961800e-01
-1.65290549e-01 -4.29239869e-03 7.83983767e-01 3.76731694e-01
6.29762471e-01 7.89671600e-01 -5.84718823e-01 1.04544252e-01
-1.15406907e+00 4.31675404e-01 1.17069864e+00 8.99071574e-01
5.57150900e-01 -1.90663293e-01 -6.25681460e-01 6.45599186e-01
2.96025366e-01 5.73812544e-01 3.16389173e-01 -7.16292977e-01
8.52909207e-01 6.90952241e-01 2.50645101e-01 -9.11636770e-01
-1.73286378e-01 -6.01842940e-01 -1.12985566e-01 -1.68170869e-01
7.87600040e-01 -1.33595079e-01 -7.52635419e-01 2.00287056e+00
4.03162569e-01 1.23737983e-01 7.95898885e-02 5.99903762e-01
1.18201244e+00 1.99570134e-01 1.73002213e-01 -2.55043715e-01
1.62540352e+00 -8.68965209e-01 -6.57539904e-01 -8.32411766e-01
7.41062343e-01 -9.34420884e-01 1.22849965e+00 1.89205259e-01
-7.35607445e-01 -1.34765610e-01 -1.00328124e+00 -1.60340756e-01
-7.91890979e-01 -5.16651198e-02 2.84823060e-01 6.28341317e-01
-8.48396003e-01 2.25285560e-01 -4.40819770e-01 -4.22414780e-01
2.28638545e-01 -5.25652478e-03 -1.65936604e-01 -6.06458597e-02
-1.65890467e+00 1.15321898e+00 4.95244205e-01 -3.61199200e-01
-8.14909041e-01 -6.40231431e-01 -8.47555399e-01 -3.45874876e-01
8.84171665e-01 -7.57977009e-01 1.18893337e+00 -9.65852678e-01
-1.17184234e+00 1.12299144e+00 -3.69843394e-01 -4.04817432e-01
5.48420608e-01 -3.38998765e-01 -2.31616616e-01 -9.42394882e-02
4.46752191e-01 -3.57228033e-02 8.80317211e-01 -1.05404651e+00
-3.68001372e-01 -2.72115171e-01 4.41347897e-01 2.17021957e-01
-4.62551087e-01 3.98774475e-01 -5.30391991e-01 -7.41341591e-01
-1.92432046e-01 -9.75025535e-01 -6.87025934e-02 -5.28785348e-01
-5.48836172e-01 -3.13837349e-01 9.47834015e-01 -6.00429773e-01
1.41783190e+00 -1.47398508e+00 1.73997864e-01 2.94039637e-01
1.67946547e-01 3.23763192e-01 -1.30886331e-01 5.39705396e-01
2.06229612e-01 3.13744813e-01 2.05907196e-01 -2.52840579e-01
1.57644689e-01 2.94585884e-01 -3.95983398e-01 3.16566229e-01
-2.38508936e-02 1.18879712e+00 -1.01943159e+00 -6.77399576e-01
-3.20682198e-01 3.99240583e-01 -4.54593599e-01 -8.64202306e-02
-7.74305224e-01 3.81495386e-01 -5.61288238e-01 6.92499220e-01
-4.41504419e-02 -4.14697707e-01 4.43429857e-01 -1.44391552e-01
9.92063992e-03 9.27433968e-01 -9.85868633e-01 1.39189529e+00
-5.14315724e-01 6.68670118e-01 1.90281108e-01 -1.01381588e+00
6.56850874e-01 4.92993653e-01 1.44933686e-01 -4.74582255e-01
2.59493589e-01 4.34936702e-01 -1.78855106e-01 -3.56059074e-01
2.40001321e-01 -1.03160322e-01 -3.56486410e-01 5.61216652e-01
-1.21443765e-02 3.11321974e-01 5.36236465e-01 3.48764330e-01
1.11783659e+00 1.88528419e-01 4.19830203e-01 -3.49825174e-01
6.37267351e-01 9.78087336e-02 5.95327258e-01 9.91011798e-01
-4.23568115e-02 8.10106993e-02 5.77801824e-01 -5.15023135e-02
-7.51725197e-01 -6.87558711e-01 7.25186765e-02 1.61823630e+00
-6.91894889e-02 -8.57673526e-01 -3.60743344e-01 -1.11357939e+00
-3.44424024e-02 6.22680902e-01 -7.53840446e-01 4.10067700e-02
-9.21713054e-01 -5.54851353e-01 4.96118039e-01 5.52196801e-01
1.29528359e-01 -9.50569630e-01 -3.64288896e-01 3.39369357e-01
-5.77490807e-01 -1.49430895e+00 -4.06538725e-01 1.65926918e-01
-7.02478349e-01 -1.30600619e+00 -3.56130749e-01 -5.68342924e-01
1.12104565e-01 1.08272722e-02 1.79539907e+00 1.22743309e-01
2.14245036e-01 3.40987444e-01 -4.06664968e-01 -6.29648268e-01
-5.59955001e-01 3.97936344e-01 -1.64894015e-01 -4.37697589e-01
6.92851245e-01 -4.24825668e-01 -2.69828528e-01 2.36335456e-01
-6.06095672e-01 -1.93562314e-01 3.08839470e-01 8.42347324e-01
3.36054116e-01 -3.92662019e-01 5.59047461e-01 -1.52366447e+00
8.09398472e-01 -5.30689955e-01 -3.81777465e-01 4.19413179e-01
-4.85328376e-01 7.69901276e-02 2.94497102e-01 -4.92632747e-01
-9.01014984e-01 -4.37316090e-01 -2.11078878e-02 3.29860039e-02
-9.32889357e-02 8.86484683e-01 -2.40149647e-01 1.81086391e-01
1.04897773e+00 -1.66614234e-01 -4.01067823e-01 -4.85907704e-01
5.42522848e-01 4.60551590e-01 2.77692676e-01 -9.21250999e-01
9.55578685e-01 3.25263351e-01 -5.71646988e-02 -7.48611867e-01
-1.55145514e+00 -7.60626256e-01 -8.62337232e-01 9.73446574e-03
6.01460457e-01 -9.26176846e-01 -1.44663543e-01 5.54787815e-02
-1.21862245e+00 -3.27405244e-01 7.11871013e-02 4.00731593e-01
-3.42883706e-01 3.62732530e-01 -7.73387730e-01 -4.01188731e-01
-4.09221888e-01 -8.51669371e-01 9.19241905e-01 -2.53351480e-01
-7.02043176e-01 -1.44942236e+00 2.50194371e-01 6.04478300e-01
1.66732296e-01 3.87142003e-01 9.29440975e-01 -1.18374038e+00
-3.97341669e-01 -3.30472469e-01 1.75549779e-02 -8.12456571e-03
1.52619213e-01 -4.21665609e-02 -8.48262429e-01 -1.30608514e-01
-1.02435395e-01 -6.11262321e-01 9.92340267e-01 1.63583338e-01
3.33881438e-01 -6.71451151e-01 -4.72789228e-01 -4.37158858e-03
1.25435257e+00 -4.59924787e-01 2.02309400e-01 7.52323091e-01
8.80231977e-01 7.69196928e-01 6.94087684e-01 -3.72465327e-02
5.07740200e-01 8.57709765e-01 2.62028515e-01 -1.85253322e-02
-1.91449121e-01 -1.36084855e-01 5.71453750e-01 5.88791311e-01
-6.01105168e-02 -2.24527895e-01 -1.08281827e+00 6.98939919e-01
-1.85921884e+00 -1.18487763e+00 -3.50513130e-01 2.00480437e+00
1.42349660e+00 6.55700684e-01 4.79205012e-01 1.25956908e-01
4.45544362e-01 3.22471052e-01 -1.91356078e-01 -2.38237470e-01
-3.21715206e-01 1.90921053e-01 4.45847064e-01 7.99162269e-01
-1.33619714e+00 1.32010901e+00 6.52901459e+00 6.52770877e-01
-9.82176006e-01 5.03511846e-01 -5.40590920e-02 -2.40112901e-01
-4.45605904e-01 1.05352283e-01 -1.21133971e+00 3.60844731e-01
9.78927553e-01 -1.44135714e-01 -1.43710494e-01 8.03049862e-01
1.32430345e-01 7.95096308e-02 -1.20547473e+00 4.19947416e-01
2.62047082e-01 -1.39016497e+00 -9.67118219e-02 -4.23633754e-02
6.33644640e-01 2.84607202e-01 -1.74524724e-01 5.79288661e-01
7.44203210e-01 -8.48272681e-01 8.15805972e-01 2.13221461e-02
4.31420296e-01 -4.37747926e-01 6.19810283e-01 6.59385622e-01
-9.46790397e-01 8.32286254e-02 -9.03528631e-02 -4.50313464e-02
-9.96960476e-02 6.37160301e-01 -1.07533717e+00 4.42899734e-01
2.90318251e-01 8.63004029e-01 -7.51241505e-01 5.58560908e-01
-8.45452189e-01 9.59911346e-01 -4.11152601e-01 -1.47007048e-01
3.66676807e-01 1.61917582e-01 8.43792379e-01 1.60253668e+00
-1.54813647e-01 -2.30035663e-01 6.42117023e-01 5.86616576e-01
-2.56869525e-01 4.90092874e-01 -7.72378445e-01 -8.92565846e-02
5.36582112e-01 1.04462528e+00 -6.10892713e-01 -5.03208101e-01
-7.61953831e-01 6.27654254e-01 4.98706400e-01 3.70839804e-01
-6.40912771e-01 1.82092652e-01 4.09611523e-01 3.18160623e-01
3.20003688e-01 -2.88690925e-01 -2.84855187e-01 -1.39352667e+00
8.69018808e-02 -1.15805030e+00 7.60926664e-01 -2.90579677e-01
-1.34985626e+00 4.00968194e-01 6.26347810e-02 -8.40857744e-01
-3.91575605e-01 -7.17484951e-01 -5.49268365e-01 9.29373443e-01
-1.86971056e+00 -1.70327294e+00 8.48934427e-02 6.09494448e-01
5.09071648e-01 3.21866721e-02 7.96866536e-01 2.29234412e-01
-4.29484010e-01 4.90270168e-01 -3.57220054e-01 4.60221440e-01
1.26681983e+00 -1.39558959e+00 2.18118429e-01 1.10619831e+00
5.33465922e-01 9.35467243e-01 1.19038141e+00 -8.84581268e-01
-9.56973195e-01 -1.07593000e+00 1.44437075e+00 -1.21888816e+00
1.15951312e+00 -4.68289077e-01 -9.73501801e-01 8.92825663e-01
2.64624834e-01 -2.38198027e-01 1.07455456e+00 8.47824156e-01
-1.11912990e+00 3.85500848e-01 -5.24474561e-01 7.28230059e-01
1.25168431e+00 -7.44852901e-01 -1.02875280e+00 5.54015338e-01
6.24377131e-01 -4.72696006e-01 -7.39796817e-01 3.84862214e-01
1.54565185e-01 -5.41304529e-01 1.11839795e+00 -9.46421027e-01
5.52463293e-01 -1.11448459e-01 -1.43833667e-01 -1.12094903e+00
-2.31985718e-01 -3.46896052e-01 -5.59794426e-01 1.39287066e+00
7.93937743e-01 -2.16542169e-01 6.86846614e-01 3.50840896e-01
8.51943344e-02 -6.19563699e-01 -6.72541022e-01 -7.29484797e-01
3.48691106e-01 -3.52952957e-01 1.82769224e-01 1.22808051e+00
2.14209065e-01 1.06654084e+00 -4.18464959e-01 6.94917142e-02
5.88432491e-01 5.42185605e-01 7.35898137e-01 -1.64766097e+00
-4.21514660e-01 -4.51366961e-01 -1.25386873e-02 -7.65280247e-01
5.31402409e-01 -1.10160971e+00 -1.91763431e-01 -1.74745166e+00
2.35614270e-01 -3.55772108e-01 -1.64635643e-01 6.62131846e-01
-4.91369873e-01 3.15020919e-01 3.24137062e-02 3.70041758e-01
-7.41469324e-01 1.36347085e-01 1.04618442e+00 -2.26294577e-01
-2.35569060e-01 -1.83911249e-02 -9.51631606e-01 8.95162225e-01
8.35791886e-01 -6.65590942e-01 -5.10114014e-01 -3.55176419e-01
6.46901727e-01 -1.88037723e-01 4.29632187e-01 -5.57420373e-01
2.16939420e-01 -3.02829564e-01 5.05992845e-02 -4.13202554e-01
2.09860682e-01 -5.61406851e-01 -3.77173394e-01 3.70107405e-02
-4.70914394e-01 -2.16848001e-01 2.09365580e-02 8.13474715e-01
-1.50577411e-01 -3.38189721e-01 4.59956765e-01 -4.22914892e-01
-5.69190145e-01 1.24889076e-01 -3.04746479e-01 6.64566875e-01
7.39720583e-01 -4.68095206e-02 -7.08410859e-01 -2.88202733e-01
-7.58526623e-01 1.11181289e-01 5.53642094e-01 5.49284101e-01
1.46427657e-02 -1.24590003e+00 -8.47817302e-01 -4.27878171e-01
3.03305030e-01 -3.85166407e-01 -4.55062389e-01 9.32226658e-01
2.43833661e-02 5.26365817e-01 2.96504915e-01 -4.51273352e-01
-1.65227008e+00 5.72579503e-01 1.02181271e-01 -8.65016997e-01
-4.03713793e-01 7.64374793e-01 -1.04506165e-01 -3.24641377e-01
2.38151297e-01 -1.04820915e-01 -6.67499542e-01 5.91971695e-01
5.22194922e-01 -2.45865121e-01 2.27038190e-01 -8.55092168e-01
-4.87275034e-01 4.67478037e-01 -2.56888002e-01 -2.56187648e-01
1.07669365e+00 -1.30422592e-01 -1.64309323e-01 7.14457572e-01
9.11545217e-01 6.28099144e-01 -6.62497818e-01 -9.49151218e-01
6.92032218e-01 -1.95471823e-01 -1.10889906e-02 -9.30053890e-01
-5.81946790e-01 6.71654761e-01 -2.50059426e-01 1.72539219e-01
5.31971812e-01 1.96917802e-01 5.45281768e-01 5.63072443e-01
2.34611750e-01 -1.18673122e+00 4.18421447e-01 1.00646412e+00
7.86824346e-01 -1.33901763e+00 9.31041539e-02 -6.91849947e-01
-6.03662431e-01 9.70269382e-01 4.94480997e-01 1.38338327e-01
6.76865757e-01 2.02288851e-01 4.45327252e-01 -4.38029379e-01
-8.33539069e-01 -4.32760656e-01 5.56601167e-01 4.84872222e-01
8.16782594e-01 -2.13585794e-01 -5.75568795e-01 6.50279701e-01
-2.51139849e-01 -2.39813581e-01 4.43183333e-01 1.12129378e+00
-4.56626326e-01 -1.53880501e+00 -3.41931432e-01 3.01984876e-01
-9.46966588e-01 -4.36100751e-01 -8.93534482e-01 7.06706226e-01
5.68996407e-02 9.88749325e-01 -5.03270686e-01 2.99214944e-02
3.36277634e-01 4.69640881e-01 4.52668935e-01 -1.22816896e+00
-8.55463803e-01 2.12933242e-01 6.71701849e-01 -3.95208776e-01
-7.49246180e-01 -6.52348459e-01 -7.02652574e-01 -2.20006123e-01
-3.85046214e-01 2.67926514e-01 2.71415710e-01 1.29894793e+00
6.14504777e-02 4.29305762e-01 4.90313917e-02 -4.65191811e-01
-4.61271197e-01 -8.96715105e-01 -1.23731434e-01 5.16400635e-01
2.73032755e-01 -7.65408814e-01 -2.42823407e-01 1.91917807e-01] | [9.016279220581055, 9.910894393920898] |
3b556306-673a-402e-9a51-9bae66155cad | spatial-scale-aligned-network-for-fine | 2001.01211 | null | https://arxiv.org/abs/2001.01211v1 | https://arxiv.org/pdf/2001.01211v1.pdf | Spatial-Scale Aligned Network for Fine-Grained Recognition | Existing approaches for fine-grained visual recognition focus on learning marginal region-based representations while neglecting the spatial and scale misalignments, leading to inferior performance. In this paper, we propose the spatial-scale aligned network (SSANET) and implicitly address misalignments during the recognition process. Especially, SSANET consists of 1) a self-supervised proposal mining formula with Morphological Alignment Constraints; 2) a discriminative scale mining (DSM) module, which exploits the feature pyramid via a circulant matrix, and provides the Fourier solver for fast scale alignments; 3) an oriented pooling (OP) module, that performs the pooling operation in several pre-defined orientations. Each orientation defines one kind of spatial alignment, and the network automatically determines which is the optimal alignments through learning. With the proposed two modules, our algorithm can automatically determine the accurate local proposal regions and generate more robust target representations being invariant to various appearance variances. Extensive experiments verify that SSANET is competent at learning better spatial-scale invariant target representations, yielding superior performance on the fine-grained recognition task on several benchmarks. | ['Hai-Hua Xu', 'Yu-Wing Tai', 'Lizhao Gao', 'Junling Liu', 'Chong Sun'] | 2020-01-05 | null | null | null | null | ['fine-grained-visual-recognition'] | ['computer-vision'] | [ 1.62908241e-01 -3.89913946e-01 -4.68331546e-01 -5.25403261e-01
-8.11760128e-01 -5.10407567e-01 6.12693608e-01 -2.74089783e-01
-3.95144165e-01 2.58316576e-01 3.06577981e-01 5.46525344e-02
-1.68814555e-01 -5.67867041e-01 -6.44306540e-01 -8.76576841e-01
-3.93201448e-02 9.87491980e-02 5.54320872e-01 6.69737309e-02
5.94639242e-01 9.98859406e-01 -1.37748551e+00 3.43456686e-01
9.32174563e-01 1.50752628e+00 1.28264770e-01 4.12809461e-01
-2.29684472e-01 5.98938823e-01 -5.27889907e-01 3.85319591e-02
3.88852596e-01 -8.95437673e-02 -4.54263657e-01 3.03254366e-01
7.40095139e-01 -1.43581033e-01 -2.35623345e-01 1.10229564e+00
4.45440888e-01 8.76517519e-02 7.87599623e-01 -9.62532401e-01
-5.57593465e-01 2.10665196e-01 -8.62256050e-01 4.57995355e-01
-6.00366630e-02 3.28380577e-02 9.83860552e-01 -1.26717508e+00
3.41984749e-01 1.35350835e+00 4.99117732e-01 -1.87774878e-02
-1.19873607e+00 -5.77638745e-01 4.40963745e-01 1.48049355e-01
-1.81018519e+00 -3.77650291e-01 8.74501288e-01 -2.86908299e-01
1.06098723e+00 6.11686260e-02 3.93907666e-01 8.83361101e-01
2.68549979e-01 5.39618433e-01 1.31122160e+00 -3.33192408e-01
3.12516510e-01 -9.11656022e-02 1.29986405e-01 7.18773365e-01
1.80262387e-01 1.21341787e-01 -6.70889974e-01 -7.92895406e-02
1.25265992e+00 2.64410372e-03 -1.51250243e-01 -8.57016861e-01
-1.27831495e+00 7.11539268e-01 9.28369224e-01 4.18349802e-01
-3.96842122e-01 -6.32880703e-02 1.26189694e-01 1.19921774e-01
3.38631034e-01 3.62329096e-01 -4.63732451e-01 3.70808452e-01
-1.15862191e+00 -8.85745436e-02 4.37061191e-01 5.74636996e-01
9.53062713e-01 1.15918092e-01 -4.16173279e-01 1.13640642e+00
2.54831582e-01 4.13985014e-01 7.62162268e-01 -4.77258921e-01
4.89833206e-01 8.57041538e-01 2.11244654e-02 -1.23777914e+00
-4.49170262e-01 -7.32371271e-01 -1.00249028e+00 1.39444426e-01
3.75354707e-01 2.85832465e-01 -1.15030658e+00 1.62796569e+00
2.98425198e-01 2.96826243e-01 -8.85773748e-02 1.22065115e+00
6.96346641e-01 4.83856916e-01 4.62339334e-02 -6.02232665e-02
1.44273543e+00 -1.17326701e+00 -2.83348054e-01 -5.06222963e-01
2.66832530e-01 -9.40178454e-01 9.03050244e-01 1.61000505e-01
-7.69053876e-01 -8.67653131e-01 -1.24487066e+00 -7.23561570e-02
-4.55003232e-01 7.70181239e-01 5.31401277e-01 3.51893336e-01
-1.04824722e+00 5.53174436e-01 -7.43534982e-01 -2.04296023e-01
3.90457541e-01 3.88246417e-01 -4.91014749e-01 -6.29727915e-02
-7.86346853e-01 8.81147385e-01 3.51575404e-01 2.98076898e-01
-5.38660645e-01 -3.99949819e-01 -1.02602220e+00 9.24897939e-02
2.65386879e-01 -4.10956323e-01 6.71847343e-01 -1.13642919e+00
-1.27405715e+00 8.31730306e-01 -4.37410772e-01 -2.47238263e-01
2.97731698e-01 8.89665633e-02 -3.26919436e-01 2.60414094e-01
2.30277508e-01 6.80098951e-01 1.45276117e+00 -9.57547784e-01
-6.38355851e-01 -6.13909721e-01 -1.20624818e-01 2.87998468e-01
-2.60802835e-01 1.05828770e-01 -6.35280907e-01 -1.19378722e+00
6.46893084e-01 -6.19883478e-01 -2.92601973e-01 9.34065692e-03
-2.66425490e-01 -3.47974718e-01 6.34314656e-01 -8.45472991e-01
1.15963960e+00 -2.28982043e+00 3.19532603e-01 6.09069288e-01
3.21811736e-02 2.49331314e-02 -4.25203502e-01 -4.91714515e-02
-1.92073941e-01 -2.73572654e-01 3.77273634e-02 -5.17785549e-02
-7.58293942e-02 -7.66523853e-02 -3.31285298e-01 5.95058024e-01
6.08460665e-01 9.40178216e-01 -4.11115676e-01 -5.31336427e-01
2.29258835e-01 3.65968704e-01 -3.58362466e-01 1.94849864e-01
2.31686920e-01 1.10890366e-01 -5.81410289e-01 9.35986519e-01
9.14039254e-01 -3.71649891e-01 3.40165012e-02 -7.94409871e-01
-2.81279951e-01 3.63690522e-03 -1.57045257e+00 1.54773057e+00
-3.50551575e-01 4.81633469e-02 4.89943102e-02 -9.87881720e-01
1.39709246e+00 -1.58025429e-01 1.30632803e-01 -8.70467424e-01
9.07740816e-02 2.91063935e-01 -1.38655186e-01 -4.35083508e-02
3.65967900e-01 6.87440559e-02 -1.60853088e-01 1.99085489e-01
7.21076429e-02 -5.68473749e-02 -1.82938147e-02 -6.64856210e-02
8.80246401e-01 -2.04487164e-02 5.54989994e-01 -4.01911139e-01
8.57675254e-01 -3.95392776e-01 7.39632487e-01 6.04085326e-01
-4.39848155e-01 5.89056432e-01 2.35508576e-01 -8.15793514e-01
-9.75436330e-01 -1.27030981e+00 -1.26416953e-02 1.07522595e+00
5.55359304e-01 -1.70661509e-01 -5.33294439e-01 -7.34524786e-01
7.98991919e-02 2.11933330e-02 -6.65631473e-01 -1.59992769e-01
-6.22650623e-01 -5.14755487e-01 2.21342459e-01 8.62984657e-01
6.41240418e-01 -1.10412562e+00 -2.01588094e-01 -1.43006910e-02
4.61207703e-02 -9.32206511e-01 -1.01945424e+00 3.35788071e-01
-8.37816179e-01 -9.91649270e-01 -7.65828371e-01 -1.06695759e+00
1.02499533e+00 5.73344946e-01 5.31252861e-01 -2.32162267e-01
-3.99117768e-01 -2.70733945e-02 -1.30154431e-01 1.42178446e-01
2.00649828e-01 1.52554378e-01 6.32103756e-02 4.80559528e-01
2.56786466e-01 -5.17028093e-01 -7.21387744e-01 7.27991104e-01
-6.05419099e-01 2.83102579e-02 1.16818380e+00 9.98089075e-01
9.42605376e-01 -8.24284628e-02 4.64717120e-01 -3.12604666e-01
3.92714083e-01 1.13069162e-01 -6.87024355e-01 5.36783040e-01
-3.33385497e-01 2.94483304e-01 6.87564075e-01 -5.02232909e-01
-9.56400454e-01 3.28425497e-01 1.25030354e-01 -4.96473610e-01
-1.99892089e-01 1.01897992e-01 -2.65662313e-01 -6.17418826e-01
6.56041145e-01 5.79662681e-01 -2.08519667e-01 -5.37803769e-01
3.28085721e-01 4.69782352e-01 4.10399407e-01 -5.78116119e-01
1.13886046e+00 5.74152589e-01 -2.65200257e-01 -8.53474200e-01
-9.70610321e-01 -6.78903759e-01 -1.01823366e+00 4.23196740e-02
8.19027781e-01 -9.77033138e-01 -2.46862561e-01 5.52343905e-01
-8.56890261e-01 -4.20594998e-02 -1.92119986e-01 4.04104352e-01
-4.92220342e-01 3.44441533e-01 -6.53124392e-01 -4.25800234e-01
-4.18425262e-01 -1.13290274e+00 1.39913094e+00 5.90848804e-01
6.48676455e-02 -6.17676318e-01 -1.03537738e-01 2.59845614e-01
5.25270104e-01 -1.82808444e-01 7.15872228e-01 -3.50275099e-01
-7.96291173e-01 -1.28025368e-01 -8.53087664e-01 3.85478079e-01
2.00213313e-01 -1.71676576e-01 -8.55958462e-01 -6.89362466e-01
-2.87747473e-01 -3.37128818e-01 1.04554212e+00 4.73028660e-01
1.25955987e+00 -1.51300222e-01 -2.82809138e-01 9.67844605e-01
1.28131306e+00 4.91665378e-02 3.76906395e-01 3.60021889e-01
6.79292202e-01 4.73628551e-01 7.26417780e-01 2.47760504e-01
1.45101130e-01 8.63430440e-01 1.58048004e-01 -3.53733629e-01
-2.11845756e-01 -1.29225641e-01 3.85144770e-01 6.64400935e-01
1.29902950e-02 5.77396214e-01 -3.74915272e-01 3.65761995e-01
-1.69751120e+00 -8.17180276e-01 5.89255691e-01 2.07283759e+00
7.56204069e-01 5.31144775e-02 4.25374657e-02 -1.64021194e-01
7.72759795e-01 6.37867510e-01 -8.24940205e-01 -2.55699694e-01
-3.88206661e-01 4.49167937e-01 6.09047771e-01 3.04257691e-01
-1.42642426e+00 1.32624459e+00 5.42548180e+00 1.17447603e+00
-1.47529149e+00 -1.66026309e-01 6.64877772e-01 2.16329128e-01
1.24436012e-02 -1.50655836e-01 -9.64517593e-01 1.15954526e-01
1.98080570e-01 1.67747244e-01 3.73096734e-01 9.71059144e-01
1.72599122e-01 1.02463767e-01 -6.46550119e-01 9.34267819e-01
2.74986267e-01 -1.37442648e+00 1.76470116e-01 2.11284123e-02
7.50910521e-01 4.01019268e-02 -2.16197278e-02 8.46922994e-02
-4.71247397e-02 -9.80978608e-01 8.55243325e-01 6.44987762e-01
8.86157811e-01 -7.40339100e-01 6.06352270e-01 2.59550273e-01
-1.74755394e+00 -1.40902638e-01 -5.71266532e-01 3.16935450e-01
-2.76101172e-01 4.76276875e-01 -6.03608727e-01 4.98964012e-01
6.68010116e-01 7.62493193e-01 -9.32723284e-01 1.10798514e+00
-2.69487590e-01 1.53782964e-01 -2.38663658e-01 1.17615215e-01
2.42386043e-01 -1.24440379e-01 2.47162849e-01 1.28128862e+00
2.92366594e-01 -2.32137963e-01 5.33650458e-01 5.81038296e-01
2.00634971e-01 2.38365978e-01 -1.92567021e-01 1.00050457e-01
5.26116431e-01 1.54582894e+00 -1.12190616e+00 -1.66168228e-01
-2.49430254e-01 1.12507439e+00 7.19310284e-01 4.82063264e-01
-5.60237706e-01 -4.01225895e-01 6.55054629e-01 -8.93801674e-02
9.03666973e-01 -3.47968787e-01 -3.98604244e-01 -1.32346630e+00
2.60749817e-01 -9.82672334e-01 5.64301848e-01 -5.84934771e-01
-1.39209485e+00 6.83534622e-01 -3.99488032e-01 -1.20843601e+00
1.14676788e-01 -7.25165725e-01 -6.06488764e-01 9.84124124e-01
-1.83757305e+00 -1.49031734e+00 -3.88140023e-01 7.21589983e-01
6.49404168e-01 -2.06209779e-01 5.54529011e-01 1.17551848e-01
-6.65908575e-01 8.89547884e-01 -2.62145579e-01 1.64476439e-01
9.03065741e-01 -1.09687686e+00 2.79410779e-01 9.49348748e-01
2.81961918e-01 7.19139278e-01 1.99450061e-01 -5.55588603e-01
-1.30563962e+00 -1.30248654e+00 7.39381909e-01 2.28212215e-02
6.79742336e-01 -3.05167556e-01 -9.99282598e-01 3.00357461e-01
-4.46025997e-01 4.30866808e-01 4.05938625e-01 -1.32213950e-01
-8.89527380e-01 -5.03568530e-01 -9.75006938e-01 3.61809999e-01
9.95012760e-01 -6.74567640e-01 -6.77350461e-01 1.43034622e-01
4.28904533e-01 -3.46655846e-01 -7.04332709e-01 4.08521086e-01
5.89724720e-01 -8.32869291e-01 1.18516064e+00 -3.32825154e-01
7.85869062e-02 -7.58160830e-01 -2.46197015e-01 -1.11992931e+00
-7.63828754e-01 7.23560825e-02 2.71406490e-02 9.88521218e-01
2.49970764e-01 -7.52078593e-01 7.82926619e-01 -5.31526506e-02
-1.44256959e-02 -9.26932752e-01 -1.06479740e+00 -7.79114783e-01
-2.95373768e-01 -1.97443560e-01 4.39878047e-01 7.56295562e-01
-4.70475346e-01 1.50875643e-01 -1.65478513e-01 6.79019630e-01
7.07852483e-01 6.77146554e-01 6.45702302e-01 -9.58513677e-01
-3.41839969e-01 -6.72863305e-01 -5.73345006e-01 -1.31018293e+00
1.11572444e-01 -7.03722715e-01 1.78005934e-01 -1.25693536e+00
2.52532721e-01 -4.66372341e-01 -5.11364400e-01 4.58085805e-01
-8.62462819e-02 5.27684450e-01 7.51391277e-02 4.97725517e-01
-7.32640564e-01 5.62234879e-01 1.23475301e+00 -2.42538318e-01
-8.58233720e-02 -7.60481833e-03 -7.56540954e-01 8.53971958e-01
6.37326062e-01 3.60481478e-02 -1.28956258e-01 -1.02457829e-01
-3.11443418e-01 -3.16064715e-01 3.70630890e-01 -9.47242439e-01
3.36386681e-01 -2.64105409e-01 8.74114871e-01 -7.57317722e-01
3.52240503e-01 -5.44692159e-01 -3.20013911e-01 4.06203032e-01
-1.84970260e-01 -9.98124257e-02 1.66900918e-01 4.49967742e-01
-4.56818014e-01 6.77324235e-02 1.08263874e+00 1.50215089e-01
-1.05525649e+00 5.22700727e-01 1.38685837e-01 -2.61563063e-01
1.03322351e+00 -4.36126053e-01 -2.65924096e-01 5.55926301e-02
-7.12005317e-01 7.60321170e-02 5.32272637e-01 5.27489185e-01
8.42221856e-01 -1.54530191e+00 -4.30668563e-01 6.48986399e-01
2.61737674e-01 -1.88665181e-01 4.63028997e-01 7.22409904e-01
-3.96204621e-01 5.37660360e-01 -2.34917447e-01 -7.17379808e-01
-1.25180900e+00 3.99539649e-01 5.42419314e-01 -4.90333289e-01
-3.64789605e-01 1.03961134e+00 5.19489050e-01 -4.41865414e-01
3.50194633e-01 -3.88725430e-01 -3.99524122e-01 8.31350684e-02
6.22885585e-01 7.32908165e-03 1.81706548e-01 -9.59864736e-01
-5.27242362e-01 1.14584744e+00 -3.46241951e-01 2.09875077e-01
1.20710778e+00 -6.36952696e-03 -9.13725197e-02 1.74933404e-01
1.19407487e+00 -4.61910255e-02 -1.61522269e+00 -6.39953673e-01
1.54822126e-01 -6.10800743e-01 5.84622435e-02 -6.75355434e-01
-9.04855728e-01 6.58206105e-01 6.51435375e-01 -5.46104200e-02
1.31297052e+00 3.92463356e-02 4.88631338e-01 1.44787058e-01
4.61027294e-01 -1.12501657e+00 3.30033600e-01 6.14908397e-01
1.14008975e+00 -1.16912556e+00 1.01146735e-01 -4.90337521e-01
-4.05331016e-01 1.17404032e+00 8.29478860e-01 -4.07807261e-01
6.20493114e-01 1.50181279e-01 4.64851968e-02 -1.17174998e-01
-4.37854469e-01 -2.00505659e-01 9.52585518e-01 4.99188215e-01
1.99894890e-01 1.10185221e-01 -1.97376624e-01 5.77090681e-01
-1.32364169e-01 -4.55237567e-01 -2.08984867e-01 6.62279248e-01
-7.09549189e-01 -9.11152482e-01 -5.47809839e-01 3.33576947e-01
6.89725801e-02 2.85163373e-02 -5.26117861e-01 4.79141295e-01
2.41571024e-01 5.37186563e-01 2.21754521e-01 -2.68233895e-01
4.46940333e-01 -1.46604016e-01 4.09507334e-01 -5.56856751e-01
-3.86263102e-01 3.09083730e-01 -3.22745293e-01 -9.91522431e-01
-2.15456173e-01 -4.99066353e-01 -1.08529091e+00 7.56896213e-02
-3.65541309e-01 -4.36042137e-02 5.28530180e-01 1.03681064e+00
4.21488881e-01 4.03252512e-01 9.60072756e-01 -1.15606403e+00
-7.13318288e-01 -8.14924657e-01 -5.82552671e-01 2.28641391e-01
2.57583439e-01 -7.28047431e-01 -3.60707939e-01 -2.48937130e-01] | [9.659908294677734, 1.9686601161956787] |
933eae00-d4e9-47cc-aac9-7e4a04ffd90f | on-the-capability-of-neural-networks-to | 2007.08032 | null | https://arxiv.org/abs/2007.08032v3 | https://arxiv.org/pdf/2007.08032v3.pdf | When and how CNNs generalize to out-of-distribution category-viewpoint combinations | Object recognition and viewpoint estimation lie at the heart of visual understanding. Recent works suggest that convolutional neural networks (CNNs) fail to generalize to out-of-distribution (OOD) category-viewpoint combinations, ie. combinations not seen during training. In this paper, we investigate when and how such OOD generalization may be possible by evaluating CNNs trained to classify both object category and 3D viewpoint on OOD combinations, and identifying the neural mechanisms that facilitate such OOD generalization. We show that increasing the number of in-distribution combinations (ie. data diversity) substantially improves generalization to OOD combinations, even with the same amount of training data. We compare learning category and viewpoint in separate and shared network architectures, and observe starkly different trends on in-distribution and OOD combinations, ie. while shared networks are helpful in-distribution, separate networks significantly outperform shared ones at OOD combinations. Finally, we demonstrate that such OOD generalization is facilitated by the neural mechanism of specialization, ie. the emergence of two types of neurons -- neurons selective to category and invariant to viewpoint, and vice versa. | ['Frédo Durand', 'Spandan Madan', 'Xavier Boix', 'Tomotake Sasaki', 'Timothy Henry', 'Helen Ho', 'Nishchal Bhandari', 'Jamell Dozier', 'Hanspeter Pfister'] | 2020-07-15 | null | null | null | null | ['viewpoint-estimation'] | ['computer-vision'] | [ 8.20733309e-02 2.17256676e-02 -4.07031272e-03 -5.83226204e-01
1.36865139e-01 -1.00949836e+00 8.41120899e-01 -1.71019867e-01
-2.32398346e-01 1.10585958e-01 9.11381096e-02 -2.25724250e-01
-2.39654914e-01 -6.66337609e-01 -8.83396924e-01 -5.97452462e-01
-1.03997834e-01 4.07815099e-01 2.83399552e-01 -1.59556925e-01
3.15587431e-01 8.41303766e-01 -1.90519130e+00 5.55902481e-01
5.13016880e-01 9.70496714e-01 8.86465535e-02 5.80290675e-01
1.23530496e-02 3.24413657e-01 -7.27047145e-01 -7.28702620e-02
6.61070108e-01 -3.70907307e-01 -4.88969833e-01 2.25530893e-01
1.29945624e+00 -2.80328810e-01 -2.97469020e-01 9.42323446e-01
3.37983191e-01 2.34146506e-01 1.08533359e+00 -1.35878205e+00
-1.18506432e+00 1.34898812e-01 -2.16989830e-01 5.48261940e-01
5.17357066e-02 3.76724064e-01 9.65608180e-01 -9.20438290e-01
7.09926486e-01 1.35428166e+00 6.37374282e-01 7.30987430e-01
-1.41985798e+00 -2.76113063e-01 6.03766143e-01 -1.11940932e-02
-1.22853816e+00 -3.54285538e-01 6.83260918e-01 -6.30667031e-01
1.25072622e+00 1.16780885e-01 8.56580555e-01 9.68881965e-01
3.08118016e-01 8.08472991e-01 1.21960557e+00 -2.43367642e-01
2.75115222e-01 1.65742114e-01 1.95783392e-01 3.65823150e-01
6.04351282e-01 2.46350572e-01 -6.68995380e-01 3.13689768e-01
1.11277306e+00 2.34008580e-02 -3.02464604e-01 -6.25338018e-01
-1.23305774e+00 5.56844890e-01 8.14489961e-01 1.40949354e-01
-9.93656218e-02 2.97255348e-02 2.02414706e-01 6.97077394e-01
3.12989742e-01 9.67382133e-01 -7.08892941e-01 3.01326960e-01
-5.26989043e-01 3.58936578e-01 5.63156486e-01 1.09523952e+00
9.00985241e-01 2.54669487e-01 -1.19360067e-01 8.14881444e-01
-1.09547965e-01 2.22186849e-01 4.61832672e-01 -9.61614311e-01
2.93854088e-01 7.77742088e-01 -1.03459865e-01 -8.15034866e-01
-7.51461327e-01 -8.43523443e-01 -6.95607305e-01 4.36633974e-01
8.01391065e-01 7.50055462e-02 -9.99890387e-01 2.02609706e+00
-9.45736021e-02 -4.27574247e-01 -3.15228626e-02 9.02206242e-01
8.18659663e-01 1.98566779e-01 -1.72960702e-02 2.51910210e-01
1.32001221e+00 -7.75771976e-01 -2.18842000e-01 -5.31928718e-01
8.38531077e-01 -3.26562852e-01 1.21308672e+00 3.67628783e-01
-8.95216346e-01 -7.49279380e-01 -9.82363284e-01 -1.14478469e-01
-6.78673625e-01 -4.75956388e-02 5.51517844e-01 5.42396188e-01
-1.35429513e+00 7.16280103e-01 -3.48655730e-01 -5.89624047e-01
7.47047246e-01 4.50144321e-01 -3.85027587e-01 -1.02084555e-01
-7.28481174e-01 1.10821772e+00 2.99841315e-01 -1.12479039e-01
-1.05180502e+00 -8.60224187e-01 -6.79098845e-01 2.91633099e-01
1.98323857e-02 -7.21817195e-01 1.13539934e+00 -1.44378197e+00
-9.14736331e-01 1.07255280e+00 -5.15769795e-02 -2.23155379e-01
3.47665280e-01 -1.18108183e-01 -2.90584922e-01 -3.01417578e-02
-1.94838531e-02 1.38792861e+00 9.56201196e-01 -1.39143240e+00
-4.69240099e-01 -6.40447319e-01 2.87381172e-01 4.12808508e-01
-3.63066077e-01 -5.24091601e-01 1.69349506e-01 -4.69731301e-01
5.42560816e-01 -1.04461789e+00 1.77250594e-01 4.71584737e-01
-3.10858727e-01 -5.04513025e-01 8.23122442e-01 -1.63406298e-01
5.29264212e-01 -2.28914905e+00 1.86338320e-01 -8.89439061e-02
4.76945937e-01 -8.22306871e-02 -1.70157075e-01 2.27502927e-01
-2.85258442e-01 1.76552892e-01 1.63234875e-01 -3.64838764e-02
8.31331238e-02 2.39276290e-01 -1.41359508e-01 5.16810060e-01
4.57077712e-01 8.57015908e-01 -6.99735582e-01 1.30058855e-01
1.65529400e-01 1.90945029e-01 -9.18004870e-01 -5.50701134e-02
-3.58741194e-01 4.92687136e-01 7.60222301e-02 6.44706428e-01
6.28293395e-01 -3.61739665e-01 1.74371883e-01 -3.16545129e-01
-2.56574273e-01 4.23105329e-01 -8.72846067e-01 1.26780653e+00
-4.92518276e-01 1.21639132e+00 -3.11799467e-01 -1.07243621e+00
9.47726369e-01 9.64159593e-02 -1.44167602e-01 -7.11426377e-01
2.71268010e-01 3.54598671e-01 7.73254395e-01 -2.82438248e-01
4.00706291e-01 -1.32644311e-01 1.62935898e-01 3.35343301e-01
3.69682789e-01 -1.83984682e-01 -2.86809597e-02 -1.41458213e-01
6.23911619e-01 -7.78773576e-02 3.39055508e-01 -6.63315654e-01
-1.68640062e-01 -1.63176000e-01 2.16530666e-01 1.08099604e+00
-2.86368340e-01 7.36904144e-01 8.98909450e-01 -7.55630851e-01
-1.11954343e+00 -9.64528918e-01 -4.89889860e-01 1.25251603e+00
3.49184096e-01 2.16494545e-01 -4.57173586e-01 -7.39419162e-01
2.69154012e-01 8.93625915e-01 -9.73208129e-01 -4.88822341e-01
-3.69845629e-01 -4.66173619e-01 2.92623699e-01 8.43860924e-01
6.95160389e-01 -9.19595122e-01 -7.81321824e-01 -2.33512267e-01
5.12200415e-01 -9.36461985e-01 -1.97540075e-01 5.86765528e-01
-1.06010365e+00 -9.96102035e-01 -6.92645133e-01 -8.29362094e-01
8.62926483e-01 7.45045841e-01 1.17648220e+00 2.19423212e-02
-2.77541801e-02 5.08549511e-01 -7.33399987e-02 -5.00069976e-01
-1.03274584e-01 2.14354247e-01 8.32060277e-02 -1.91083431e-01
4.51437891e-01 -8.51049781e-01 -7.59830177e-01 6.83401763e-01
-7.66210675e-01 2.87308805e-02 6.30254328e-01 6.73342645e-01
5.48675805e-02 -3.22489768e-01 6.02668881e-01 -7.43176281e-01
4.01506424e-01 -5.16200900e-01 -4.29722607e-01 5.63850962e-02
-4.33630526e-01 1.43551886e-01 5.54592073e-01 -6.67545736e-01
-7.16654480e-01 -1.64295986e-01 1.02965511e-01 -5.60691535e-01
-6.27910554e-01 9.88322943e-02 -1.17676824e-01 -3.40836048e-01
1.01236439e+00 2.71826893e-01 -3.72572020e-02 -4.26584899e-01
3.27627033e-01 3.21346223e-01 1.28064558e-01 -3.79722357e-01
5.68016708e-01 6.40376747e-01 -4.06572334e-02 -8.80533814e-01
-9.04109657e-01 -1.16551347e-01 -8.35446239e-01 -1.67364910e-01
1.05462217e+00 -9.65323210e-01 -4.17060554e-01 6.07390165e-01
-9.13429499e-01 -6.19895279e-01 -3.35598022e-01 4.40422565e-01
-4.35229003e-01 -1.84999496e-01 -9.85965654e-02 -3.26948166e-01
4.54868942e-01 -1.08204615e+00 9.56893027e-01 2.46440843e-01
-3.07685554e-01 -1.00136554e+00 -3.62249643e-01 -2.19795913e-01
4.97408181e-01 -5.04163429e-02 1.23083436e+00 -1.09996963e+00
-7.57951260e-01 -2.11966559e-01 -5.48381448e-01 3.10132176e-01
1.61479652e-01 -1.57339782e-01 -1.11442459e+00 -2.85967708e-01
-1.69613451e-01 -2.36272171e-01 1.05014277e+00 5.46522558e-01
1.27488482e+00 -1.38677165e-01 -2.75796235e-01 8.50898087e-01
1.17991579e+00 2.47607201e-01 4.28845316e-01 3.24243337e-01
7.02898920e-01 5.92694879e-01 -8.70909914e-02 2.65303887e-02
2.42555752e-01 8.31324995e-01 6.27830386e-01 1.38657764e-01
-4.75156575e-01 -6.69988170e-02 -8.23042616e-02 1.96176976e-01
-1.36650175e-01 -3.48794162e-01 -7.85329998e-01 6.53500557e-01
-1.23017228e+00 -7.94934094e-01 5.84520400e-02 2.18183970e+00
4.85893339e-01 3.93940389e-01 1.33699879e-01 -1.19331904e-01
5.50838947e-01 1.89581916e-01 -8.60868752e-01 -6.23908460e-01
-3.39432985e-01 -2.20738813e-01 4.42764312e-01 1.83858216e-01
-9.50768650e-01 7.52370477e-01 6.95908403e+00 4.48087960e-01
-1.34536326e+00 4.06930186e-02 5.33240080e-01 -7.83651918e-02
-4.89773363e-01 -1.86569259e-01 -9.64434803e-01 1.62786052e-01
4.89209592e-01 4.03010845e-01 3.14466208e-01 1.03471136e+00
-3.48534942e-01 6.27892790e-04 -1.57042909e+00 7.00582504e-01
2.40519941e-01 -1.35484862e+00 3.25924486e-01 2.64550984e-01
1.00618148e+00 7.24123865e-02 3.34007919e-01 3.97623658e-01
2.20056221e-01 -9.87363696e-01 8.89056683e-01 9.29298922e-02
6.97058797e-01 -3.03583920e-01 3.77436876e-01 4.08792138e-01
-8.70045304e-01 -4.26513135e-01 -4.24016327e-01 -2.95020968e-01
-3.02488834e-01 2.37628892e-01 -6.29206479e-01 3.07014063e-02
7.96046317e-01 8.75193000e-01 -9.12739515e-01 1.06009424e+00
-3.35485227e-02 1.63474813e-01 -3.43179852e-01 -1.34729609e-01
2.61943758e-01 1.99844375e-01 6.46167517e-01 8.78031433e-01
3.01111102e-01 -6.44667819e-02 -1.40553832e-01 9.88134086e-01
-5.86483665e-02 -4.74262655e-01 -9.30104196e-01 2.10110337e-01
5.34595907e-01 7.66170502e-01 -8.52384984e-01 -3.00175577e-01
-4.21496779e-01 7.41771102e-01 5.21508574e-01 4.25580829e-01
-4.68284845e-01 -1.39944136e-01 7.51187086e-01 1.45058393e-01
5.91569424e-01 -4.28520262e-01 -6.07126892e-01 -1.18080401e+00
-1.84061483e-01 -6.51855290e-01 1.88800469e-01 -9.61344123e-01
-1.38753903e+00 6.73187613e-01 2.68875360e-01 -1.28563678e+00
1.89978912e-01 -1.21227515e+00 -6.13359451e-01 5.78242719e-01
-1.30215847e+00 -1.00997114e+00 -4.31716144e-01 2.97046244e-01
7.74468184e-01 -2.85346001e-01 5.69496751e-01 -2.32477393e-02
-3.52100313e-01 7.89449573e-01 7.93117955e-02 -2.09979271e-03
5.65743148e-01 -1.26697969e+00 3.92534852e-01 3.83785278e-01
3.22332084e-01 7.83272028e-01 5.48422873e-01 -2.43497998e-01
-1.05597007e+00 -9.12024498e-01 8.22377682e-01 -7.54946828e-01
2.29786485e-01 -6.02639675e-01 -8.40400875e-01 1.04819298e+00
1.48110539e-02 6.45716637e-02 4.71478939e-01 4.31078672e-01
-9.69952285e-01 -1.40185682e-02 -1.06335878e+00 8.65104795e-01
1.53566217e+00 -5.54915011e-01 -7.04110444e-01 4.51080263e-01
7.18264639e-01 -2.73647010e-01 -5.19325852e-01 3.54796350e-01
6.80501461e-01 -1.43584800e+00 1.00249135e+00 -7.90549159e-01
5.48881471e-01 -2.22610071e-01 -5.16452849e-01 -1.52445662e+00
-4.21218276e-01 -6.88782055e-03 1.00896470e-01 6.87821448e-01
4.83705580e-01 -1.02161646e+00 4.82894152e-01 2.20674410e-01
-6.24588132e-01 -8.33901644e-01 -1.00132513e+00 -1.16179752e+00
3.84059370e-01 -1.95477098e-01 5.18981576e-01 1.04995775e+00
-4.38727558e-01 1.03519373e-01 1.40834644e-01 1.98986843e-01
2.41885304e-01 2.04880297e-01 6.07717693e-01 -1.30714166e+00
-2.34673381e-01 -7.38735914e-01 -7.84682214e-01 -1.39797020e+00
1.48179103e-02 -9.85108912e-01 -2.32738242e-01 -1.32589734e+00
8.16645287e-03 -5.36061883e-01 -1.53481752e-01 4.59091783e-01
1.17272846e-01 4.06368434e-01 3.69556606e-01 5.59469044e-01
-4.29537773e-01 3.02317470e-01 1.45347273e+00 1.06650457e-01
-2.22712755e-01 1.49182573e-01 -1.04549348e+00 7.83702910e-01
6.84842467e-01 -2.52043784e-01 -4.37978089e-01 -9.54490542e-01
2.32946098e-01 -4.17154938e-01 9.09361124e-01 -1.17451560e+00
-1.82380095e-01 -3.04316310e-03 7.97620118e-01 -5.07256985e-01
4.36864614e-01 -7.32452989e-01 -3.04066747e-01 4.44671839e-01
-5.93662798e-01 -1.19859260e-02 3.70640248e-01 6.41310513e-01
8.72474164e-02 -1.22199468e-01 7.67090738e-01 -4.09895152e-01
-9.35874403e-01 8.58507976e-02 -4.53864127e-01 1.95702732e-01
7.60512769e-01 -7.88151026e-01 -5.82525253e-01 -3.85136366e-01
-9.55714703e-01 -1.82314157e-01 6.47762716e-01 5.13130248e-01
2.67353892e-01 -1.30382586e+00 -4.27625746e-01 4.81025189e-01
3.04773211e-01 5.47842570e-02 2.58092731e-01 8.15664411e-01
-3.49233121e-01 3.41660559e-01 -5.69528401e-01 -9.77168798e-01
-7.61739135e-01 4.99619514e-01 7.37085104e-01 4.57159907e-01
-3.73923481e-01 1.25849831e+00 8.97789061e-01 -7.24588335e-01
1.02954760e-01 -4.53279376e-01 -1.11171097e-01 1.26285911e-01
1.12510152e-01 7.71656185e-02 6.81290999e-02 -4.22518581e-01
-2.92026639e-01 8.01950216e-01 -1.38285458e-01 2.41125047e-01
1.17899644e+00 -5.47652990e-02 1.78081751e-01 6.41722798e-01
1.20978701e+00 -4.10256892e-01 -1.60137320e+00 -8.38527754e-02
-1.95440441e-01 -5.51209986e-01 -2.36995623e-01 -8.27927411e-01
-1.13488317e+00 1.21149695e+00 5.93262017e-01 4.22992349e-01
7.57711530e-01 2.95405537e-01 8.22710246e-03 6.80581927e-01
2.10851744e-01 -7.22090840e-01 3.24786603e-01 7.17240274e-01
1.03681207e+00 -1.24692917e+00 4.85627167e-02 -3.15166056e-01
-6.67249680e-01 1.21488416e+00 1.15502942e+00 -6.47462428e-01
7.25366473e-01 -3.82772565e-01 -1.43270820e-01 -4.09906954e-01
-7.04733491e-01 -1.09441543e-03 5.43716848e-01 9.41631556e-01
2.55247861e-01 -4.03379789e-03 3.26815456e-01 2.04980612e-01
-3.22903693e-01 -5.12530625e-01 4.44906354e-01 7.79577792e-01
-5.90085983e-01 -4.31034535e-01 -1.00095317e-01 6.47479534e-01
2.14407250e-01 -1.47681415e-01 -6.44304216e-01 1.23441219e+00
4.41093624e-01 3.94598365e-01 7.00314939e-01 -2.89344043e-01
2.83903092e-01 2.32450768e-01 6.90345705e-01 -9.17677701e-01
-4.95042652e-01 -2.83570349e-01 -2.75749173e-02 -2.34407648e-01
-2.40255490e-01 -6.94752216e-01 -7.24820018e-01 -1.74526069e-02
-3.87736171e-01 -5.76843202e-01 4.56719637e-01 1.01674592e+00
4.84360784e-01 5.81518710e-01 4.42226499e-01 -1.19374132e+00
-6.21360421e-01 -8.60624075e-01 -6.57099903e-01 3.71178716e-01
6.17776692e-01 -1.02546048e+00 -8.64934444e-01 -2.39747643e-01] | [9.581777572631836, 2.077199697494507] |
f2f1faff-814a-4b1a-8bba-31f6b493a2a2 | explaining-neural-network-predictions-on | 2104.04488 | null | https://arxiv.org/abs/2104.04488v2 | https://arxiv.org/pdf/2104.04488v2.pdf | Explaining Neural Network Predictions on Sentence Pairs via Learning Word-Group Masks | Explaining neural network models is important for increasing their trustworthiness in real-world applications. Most existing methods generate post-hoc explanations for neural network models by identifying individual feature attributions or detecting interactions between adjacent features. However, for models with text pairs as inputs (e.g., paraphrase identification), existing methods are not sufficient to capture feature interactions between two texts and their simple extension of computing all word-pair interactions between two texts is computationally inefficient. In this work, we propose the Group Mask (GMASK) method to implicitly detect word correlations by grouping correlated words from the input text pair together and measure their contribution to the corresponding NLP tasks as a whole. The proposed method is evaluated with two different model architectures (decomposable attention model and BERT) across four datasets, including natural language inference and paraphrase identification tasks. Experiments show the effectiveness of GMASK in providing faithful explanations to these models. | ['Yangfeng Ji', 'Sachindra Joshi', 'Chulaka Gunasekara', 'Hui Wan', 'Jatin Ganhotra', 'Song Feng', 'Hanjie Chen'] | 2021-04-09 | null | https://aclanthology.org/2021.naacl-main.306 | https://aclanthology.org/2021.naacl-main.306.pdf | naacl-2021-4 | ['paraphrase-identification'] | ['natural-language-processing'] | [ 4.41990644e-01 1.77803695e-01 -2.33453900e-01 -7.71714091e-01
-2.69657075e-01 -3.27376783e-01 9.75195348e-01 2.04583511e-01
-8.73731673e-02 5.39677501e-01 3.91116768e-01 -5.81541896e-01
-3.21852714e-01 -4.07436341e-01 -7.69985437e-01 -3.39828968e-01
3.16972911e-01 4.13710088e-01 -2.43736342e-01 1.31944865e-01
6.09588683e-01 1.76609531e-01 -1.45031643e+00 6.68929875e-01
1.12243867e+00 8.64788532e-01 1.62304312e-01 4.97111171e-01
-3.93929392e-01 1.01009190e+00 -4.96745944e-01 -7.90258288e-01
-1.14944339e-01 -7.91895092e-01 -7.93783903e-01 -1.51597545e-01
7.04071760e-01 -2.19003186e-01 -2.96197653e-01 1.11793756e+00
4.51366752e-02 1.77194312e-01 9.67801213e-01 -1.66186059e+00
-1.28262937e+00 1.34018946e+00 -4.77542460e-01 4.21273291e-01
1.97709098e-01 -1.17533661e-01 1.61856306e+00 -1.30136347e+00
-8.39230837e-04 1.47894502e+00 5.59582829e-01 6.00321829e-01
-1.34590602e+00 -1.00865638e+00 3.74071896e-01 8.87111425e-01
-9.98981714e-01 -4.49635118e-01 8.96653354e-01 -3.41521919e-01
1.31665075e+00 5.27681649e-01 3.47824544e-01 1.30435991e+00
2.83935130e-01 7.48497188e-01 7.97703624e-01 -4.11268800e-01
8.04663002e-02 2.41457880e-01 8.83179009e-01 5.59157789e-01
4.32711244e-01 -3.56928855e-02 -1.12161493e+00 -2.31842846e-01
2.93723226e-01 1.67618364e-01 -3.11034709e-01 -4.37804013e-02
-1.29484212e+00 1.01088774e+00 5.94913006e-01 3.20596814e-01
-3.91772866e-01 1.93269104e-01 2.48845622e-01 4.72297370e-01
5.13693631e-01 5.87302208e-01 -2.95379043e-01 2.69727170e-01
-8.34933043e-01 1.52239025e-01 5.60130775e-01 9.39457178e-01
6.33529007e-01 -1.37189969e-01 -3.24977070e-01 8.32577944e-01
5.09724498e-01 1.55365899e-01 9.40209270e-01 -4.16045159e-01
5.63514888e-01 6.88965976e-01 -1.06265314e-01 -1.47847378e+00
-2.81195164e-01 -4.73050386e-01 -9.67982352e-01 -3.89282227e-01
1.51216224e-01 5.37103303e-02 -4.79553699e-01 1.91810060e+00
-9.18227956e-02 4.85503495e-01 2.34431848e-01 7.77871549e-01
1.21028578e+00 4.54140931e-01 -1.31922960e-02 7.41964765e-03
1.28292978e+00 -1.27352309e+00 -8.21377397e-01 -6.45156682e-01
4.78291869e-01 -5.61239839e-01 1.04042065e+00 5.84078431e-02
-9.39569890e-01 -8.13561618e-01 -1.00425637e+00 -2.43258417e-01
-1.75835997e-01 2.45403573e-01 5.71969390e-01 2.32758731e-01
-7.43277133e-01 7.23231375e-01 -5.39842904e-01 -2.42159501e-01
3.93963844e-01 5.09281278e-01 -2.58937657e-01 1.04618572e-01
-1.47675860e+00 9.88126695e-01 3.46277028e-01 2.47929856e-01
-6.01607323e-01 -7.37226427e-01 -7.60075927e-01 6.54311061e-01
-5.35866693e-02 -7.46401191e-01 1.04867160e+00 -1.11058855e+00
-9.31976140e-01 7.03495324e-01 -6.08091295e-01 -6.06830001e-01
2.94391662e-01 -3.86605680e-01 -2.05022916e-01 -1.88867167e-01
1.36934653e-01 6.42826319e-01 8.53948712e-01 -1.11720741e+00
-2.60715395e-01 -4.11003083e-01 -6.26971796e-02 2.59476036e-01
-7.59754777e-01 7.96452314e-02 5.11287414e-02 -5.29171884e-01
3.66306216e-01 -7.26297438e-01 3.99675742e-02 -1.85015455e-01
-1.00382817e+00 -5.31381547e-01 6.64459825e-01 -6.52177036e-01
9.93998289e-01 -1.93277109e+00 8.51017889e-03 1.48136705e-01
6.05315328e-01 1.23013608e-01 -2.65816599e-01 3.54338378e-01
-4.39859062e-01 3.46554160e-01 -5.32173924e-02 -7.51266301e-01
1.36677638e-01 2.07538813e-01 -8.15407097e-01 1.51097640e-01
3.77260000e-01 1.10435343e+00 -8.30423653e-01 -4.10090268e-01
-7.11605325e-02 1.84264347e-01 -3.24818969e-01 2.89750129e-01
1.29202306e-01 -5.20828925e-02 -2.37177789e-01 2.30425626e-01
4.98327821e-01 -3.68609905e-01 8.97987336e-02 -9.19552222e-02
3.99515688e-01 6.00549161e-01 -7.02427804e-01 9.94442165e-01
-3.25561702e-01 1.02453959e+00 -6.95715070e-01 -1.06563091e+00
1.16322196e+00 2.48945907e-01 -2.68100709e-01 -3.46726269e-01
1.12656280e-01 7.92762488e-02 4.88222718e-01 -6.09476030e-01
4.20582265e-01 -1.99416429e-01 1.71046376e-01 7.96173692e-01
2.24031787e-02 3.61549854e-01 -2.38669664e-01 4.02420610e-01
9.75810707e-01 -5.00145555e-01 5.21148682e-01 -2.15268657e-01
4.85344827e-01 -3.08200508e-01 2.29682639e-01 9.11647499e-01
-1.43360943e-01 6.48803234e-01 7.87845850e-01 -3.41946393e-01
-8.30820441e-01 -7.44396865e-01 8.89137462e-02 9.21323597e-01
1.43018365e-01 -3.00034225e-01 -6.20480299e-01 -8.11492741e-01
1.97706774e-01 1.30023456e+00 -1.16130614e+00 -6.68824434e-01
-1.33430868e-01 -4.09059972e-01 5.86440623e-01 7.79815733e-01
4.98141497e-01 -1.22132778e+00 -3.17606211e-01 -1.78277597e-01
-4.21176016e-01 -1.07396019e+00 -5.61546266e-01 3.42996567e-01
-7.59470582e-01 -1.01676643e+00 -4.62287143e-02 -7.39200652e-01
8.90016496e-01 6.56132162e-01 9.99371648e-01 4.81044859e-01
1.25242084e-01 -1.98022634e-01 -2.21991315e-01 -5.08964598e-01
-4.66946125e-01 -1.51428446e-01 1.81721702e-01 1.57937661e-01
9.10188556e-01 -4.46130067e-01 -3.36280346e-01 3.23698819e-01
-6.52408421e-01 3.94109845e-01 6.29940033e-01 1.16868198e+00
-1.18008293e-02 -3.70015413e-01 7.18256831e-01 -1.02312756e+00
1.10710573e+00 -8.05207253e-01 1.16553277e-01 5.25131762e-01
-6.29193068e-01 1.99667528e-01 9.54138160e-01 -6.07475162e-01
-9.52328146e-01 -3.43531400e-01 3.11643779e-01 -5.90402365e-01
-2.43012860e-01 7.80274391e-01 -1.06311269e-01 4.27748322e-01
6.48757696e-01 3.17274392e-01 1.15857437e-01 -8.26824680e-02
2.25363880e-01 7.32612908e-01 5.07969737e-01 -1.56764582e-01
7.59159327e-01 -3.18390243e-02 -2.43230090e-01 -5.04909098e-01
-9.95540679e-01 -3.72933894e-01 -6.08209491e-01 -4.49534431e-02
5.21041930e-01 -6.28629744e-01 -6.94100142e-01 8.66023153e-02
-1.62345970e+00 1.57360882e-01 2.29982078e-01 5.18897176e-01
-2.12616861e-01 4.62237656e-01 -2.96873152e-01 -6.70975447e-01
-3.80078495e-01 -8.69893014e-01 6.64426863e-01 9.74441916e-02
-9.07459974e-01 -1.09805417e+00 -7.41085112e-02 6.12342238e-01
3.32736284e-01 -1.97476819e-01 1.32333016e+00 -1.38306201e+00
-1.23460829e-01 -2.53113598e-01 -4.91925746e-01 1.00267768e-01
2.82625500e-02 7.24792406e-02 -1.11307502e+00 1.76007494e-01
1.16024949e-01 -3.72368008e-01 1.07100451e+00 2.33842075e-01
1.31050801e+00 -6.40209019e-01 -3.89709592e-01 1.66579425e-01
7.56352842e-01 -5.86001910e-02 3.81451696e-01 4.52213697e-02
7.81341076e-01 9.47728813e-01 4.51423079e-01 2.55191684e-01
1.64741367e-01 4.70492274e-01 6.08886123e-01 -5.86048998e-02
9.85004976e-02 -4.67589200e-01 4.75943297e-01 7.32223988e-01
1.94385648e-01 -4.88436431e-01 -8.03028345e-01 5.33936977e-01
-2.25871468e+00 -1.18762517e+00 -6.49043560e-01 1.83536077e+00
5.63650846e-01 -3.17453966e-02 -4.72024024e-01 4.15466689e-02
9.72756565e-01 -3.00849992e-04 -8.75582159e-01 -5.80500484e-01
-9.93556976e-02 -7.20957369e-02 -1.36928186e-01 5.09366930e-01
-7.36011326e-01 7.96217263e-01 5.64405489e+00 5.52549422e-01
-5.67394853e-01 -4.76516318e-03 8.51939023e-01 -1.09596290e-02
-7.36016512e-01 -2.12751832e-02 -6.13620818e-01 4.18258667e-01
7.83515871e-01 -3.99860084e-01 1.73734143e-01 8.38158488e-01
3.92625213e-01 8.18303302e-02 -1.57183218e+00 8.41397822e-01
4.04235601e-01 -1.17447352e+00 5.96034288e-01 -2.27675900e-01
6.75182164e-01 -2.43049055e-01 1.20085023e-01 7.53402635e-02
2.29645953e-01 -1.31884658e+00 6.54653847e-01 3.86337608e-01
2.44220674e-01 -7.00481474e-01 9.55541432e-01 4.88866031e-01
-7.36322582e-01 -8.98224562e-02 -7.44842052e-01 -3.57925117e-01
-1.10101730e-01 4.68161047e-01 -1.08927214e+00 1.66303992e-01
4.57907856e-01 8.90279591e-01 -8.46704781e-01 6.26444101e-01
-5.97400606e-01 9.33172941e-01 2.05348730e-01 -6.28869712e-01
1.60690561e-01 4.35515307e-02 3.83341372e-01 1.09791398e+00
2.18880072e-01 -1.61818117e-01 -4.44319189e-01 1.50843596e+00
-2.57993668e-01 5.53209633e-02 -6.83297396e-01 -4.83816527e-02
6.82705820e-01 1.27750385e+00 -5.84342241e-01 -4.93058354e-01
-3.08117121e-01 1.02478755e+00 7.63392806e-01 2.50652581e-01
-1.04290295e+00 -3.16293806e-01 8.51580262e-01 -3.32077235e-01
-1.64523851e-02 7.27879852e-02 -8.19809616e-01 -1.07303083e+00
5.89974113e-02 -7.33231068e-01 1.42301112e-01 -9.89813447e-01
-1.64629185e+00 6.49409652e-01 -9.48334932e-02 -8.93685043e-01
-3.52998048e-01 -5.09944081e-01 -1.26386976e+00 1.02469957e+00
-1.24860227e+00 -1.13756382e+00 -5.42423606e-01 5.85661650e-01
7.59509146e-01 -3.15809458e-01 8.12692523e-01 -2.25801542e-01
-7.21541047e-01 8.71108532e-01 -7.78534487e-02 7.18904212e-02
5.76882124e-01 -1.17596734e+00 7.21122801e-01 1.00510597e+00
6.01108074e-01 1.33826506e+00 9.11741138e-01 -6.93906605e-01
-9.16187227e-01 -1.01801848e+00 1.64466763e+00 -4.91958618e-01
7.62329102e-01 -4.77773786e-01 -1.24229169e+00 7.23035455e-01
5.02185404e-01 -4.41969573e-01 1.10573435e+00 4.49773997e-01
-6.08110011e-01 2.42265984e-01 -7.66647279e-01 8.47292483e-01
1.12007403e+00 -7.82953143e-01 -9.89230394e-01 3.80144149e-01
6.84589386e-01 1.62459224e-01 -2.86468983e-01 1.35299489e-01
6.31019711e-01 -1.10854256e+00 7.81888664e-01 -1.19161904e+00
1.22815108e+00 -6.97417632e-02 -3.57984826e-02 -1.55558968e+00
-5.27237177e-01 -2.10700914e-01 -1.64141580e-02 1.32645726e+00
7.98280656e-01 -6.88311636e-01 6.28461659e-01 1.07620513e+00
-9.04958099e-02 -5.85963905e-01 -6.59026384e-01 -5.41961491e-01
-1.32558584e-01 -3.92006725e-01 6.50642216e-01 1.31489265e+00
5.75814068e-01 8.06245863e-01 -3.62121820e-01 2.85107076e-01
5.24630308e-01 1.89463556e-01 6.50187373e-01 -1.15751576e+00
-4.09632295e-01 -6.90579057e-01 -2.32543707e-01 -8.25885534e-01
7.53864527e-01 -1.13613474e+00 1.01031415e-01 -1.47528040e+00
8.46685290e-01 8.69238302e-02 -3.32106173e-01 6.40895844e-01
-7.49494672e-01 1.82411879e-01 9.66273397e-02 4.59396392e-01
-2.79202789e-01 6.19267464e-01 6.85433745e-01 -1.85730308e-01
1.55669570e-01 -1.34304827e-02 -9.59366798e-01 7.58349776e-01
1.08770823e+00 -5.34815788e-01 -7.12664187e-01 -4.64594930e-01
2.25292653e-01 -2.73849458e-01 8.38090301e-01 -6.22843802e-01
3.99643153e-01 -1.02096841e-01 3.88054222e-01 -4.47619826e-01
1.28436804e-01 -6.56674504e-01 2.91137639e-02 5.52368045e-01
-1.21329129e+00 3.74421865e-01 -2.26978399e-02 6.55490577e-01
-3.42013299e-01 -6.21481240e-01 4.18691605e-01 2.20115911e-02
-3.93096805e-01 -1.08381242e-01 -3.55489522e-01 -1.24279328e-01
7.78210878e-01 -2.37708896e-01 -5.75568855e-01 -6.37849867e-01
-4.87465709e-01 1.13391124e-01 -1.02956042e-01 7.61790037e-01
1.05725014e+00 -1.31604588e+00 -6.39232755e-01 2.80811459e-01
3.64940673e-01 -4.21326727e-01 1.81418970e-01 6.95866227e-01
1.05966963e-01 6.80829704e-01 -2.16160223e-01 -3.34117621e-01
-1.77874362e+00 4.34168637e-01 1.74782574e-01 -1.82356685e-01
-1.80407584e-01 1.18320644e+00 4.95532066e-01 -4.15444523e-01
2.27114737e-01 -4.70706791e-01 -5.24427056e-01 5.27463481e-02
5.91736138e-01 5.33744544e-02 -5.38232140e-02 -5.69334865e-01
-2.92027891e-01 2.17386022e-01 -4.67107534e-01 1.32263511e-01
1.15109539e+00 -9.31359082e-02 -2.32479021e-01 5.96164525e-01
1.16155589e+00 -3.95637929e-01 -8.29084337e-01 -3.88479441e-01
2.23498065e-02 -3.81821692e-01 -1.00761414e-01 -5.94325185e-01
-8.32336605e-01 1.16968548e+00 -3.79058868e-02 1.83746502e-01
5.88978112e-01 1.62906088e-02 5.18020034e-01 5.82861364e-01
-2.24266097e-01 -6.39801919e-01 2.97414929e-01 3.83171439e-01
1.19670224e+00 -1.27188742e+00 -1.39511883e-01 -3.06829989e-01
-8.73072743e-01 1.27486420e+00 9.19262886e-01 -3.54577079e-02
3.50189120e-01 -2.42465094e-01 -6.19528964e-02 -2.37361670e-01
-1.15555513e+00 2.25673467e-01 5.73141396e-01 3.39497715e-01
6.80236936e-01 7.26017952e-02 -3.22894484e-01 9.47385430e-01
-3.83826464e-01 -5.68913102e-01 4.66187954e-01 1.05018951e-01
-3.47979933e-01 -6.80582345e-01 -2.64293909e-01 8.29007745e-01
-9.66157317e-02 -6.07051671e-01 -1.15358818e+00 5.51052928e-01
-2.34030679e-01 1.20108199e+00 2.83373028e-01 -6.86879694e-01
-3.92444758e-03 3.38608146e-01 -1.27540633e-01 -5.67240059e-01
-8.68408859e-01 -6.12488508e-01 1.48437545e-01 -3.08763355e-01
-3.19393933e-01 -7.47353256e-01 -1.21282637e+00 -3.91132504e-01
-6.34801745e-01 1.50151387e-01 4.65231746e-01 1.31054735e+00
4.88406241e-01 5.08717477e-01 5.38853526e-01 -4.86949921e-01
-7.39647388e-01 -1.26365721e+00 -3.31810206e-01 7.33871877e-01
1.54570714e-01 -6.75146043e-01 -7.09988534e-01 1.38014201e-02] | [9.253952980041504, 6.286302089691162] |
e25ad364-4e47-4bd6-b15f-86bca6a1db6a | identity-preserving-realistic-talking-face | 2005.12318 | null | https://arxiv.org/abs/2005.12318v1 | https://arxiv.org/pdf/2005.12318v1.pdf | Identity-Preserving Realistic Talking Face Generation | Speech-driven facial animation is useful for a variety of applications such as telepresence, chatbots, etc. The necessary attributes of having a realistic face animation are 1) audio-visual synchronization (2) identity preservation of the target individual (3) plausible mouth movements (4) presence of natural eye blinks. The existing methods mostly address the audio-visual lip synchronization, and few recent works have addressed the synthesis of natural eye blinks for overall video realism. In this paper, we propose a method for identity-preserving realistic facial animation from speech. We first generate person-independent facial landmarks from audio using DeepSpeech features for invariance to different voices, accents, etc. To add realism, we impose eye blinks on facial landmarks using unsupervised learning and retargets the person-independent landmarks to person-specific landmarks to preserve the identity-related facial structure which helps in the generation of plausible mouth shapes of the target identity. Finally, we use LSGAN to generate the facial texture from person-specific facial landmarks, using an attention mechanism that helps to preserve identity-related texture. An extensive comparison of our proposed method with the current state-of-the-art methods demonstrates a significant improvement in terms of lip synchronization accuracy, image reconstruction quality, sharpness, and identity-preservation. A user study also reveals improved realism of our animation results over the state-of-the-art methods. To the best of our knowledge, this is the first work in speech-driven 2D facial animation that simultaneously addresses all the above-mentioned attributes of a realistic speech-driven face animation. | ['Brojeshwar Bhowmick', 'Sandika Biswas', 'Sanjana Sinha'] | 2020-05-25 | null | null | null | null | ['audio-visual-synchronization', 'audio-visual-synchronization', 'talking-face-generation'] | ['audio', 'computer-vision', 'computer-vision'] | [-4.52459753e-02 2.22217381e-01 -3.55853094e-03 -1.75351679e-01
-5.93612731e-01 -2.32007742e-01 5.85913479e-01 -5.32284677e-01
9.06496048e-02 6.15420938e-01 4.13282454e-01 3.11690480e-01
3.17329526e-01 -2.81320333e-01 -5.50055325e-01 -8.23764503e-01
-6.73645549e-03 -6.44712001e-02 2.87235808e-02 -4.18388903e-01
1.69187233e-01 7.31858253e-01 -1.94810081e+00 6.94927499e-02
4.67087686e-01 7.74589062e-01 -8.03478807e-02 5.92301488e-01
-1.78956743e-02 5.59744835e-01 -4.67864692e-01 -3.25177163e-01
1.82528660e-01 -6.87259853e-01 -5.95299423e-01 4.01196092e-01
5.42486489e-01 -4.38912123e-01 -3.16481739e-02 9.48203564e-01
7.11814463e-01 1.73500419e-01 5.38083911e-01 -1.60516334e+00
-6.27530456e-01 4.19832654e-02 -9.40830946e-01 -2.45308563e-01
6.92927957e-01 1.27102464e-01 5.46272099e-01 -6.94280386e-01
7.71483421e-01 1.62969255e+00 5.99241972e-01 9.53637064e-01
-1.08229363e+00 -1.05859125e+00 -1.23200372e-01 8.73802006e-02
-1.58888197e+00 -9.91419256e-01 1.07421112e+00 -2.43632361e-01
2.45531887e-01 2.41366357e-01 6.63276672e-01 1.10841501e+00
-1.15006017e-02 4.28941458e-01 9.97406125e-01 -5.85575521e-01
-1.54424170e-02 2.88353622e-01 -6.17507577e-01 7.36009181e-01
-3.53548169e-01 1.87940761e-01 -5.57360947e-01 -3.68197523e-02
1.03097713e+00 -3.29343528e-01 -2.93497175e-01 -1.79410875e-01
-1.05166483e+00 6.98096752e-01 1.38856936e-03 3.45697135e-01
-3.31612587e-01 1.76097900e-01 3.11111927e-01 2.72728242e-02
5.85178435e-01 5.18446006e-02 -2.87334770e-02 -4.47386205e-02
-1.03222227e+00 1.37673542e-02 5.51855624e-01 9.06810820e-01
6.71774864e-01 5.21844506e-01 3.11576687e-02 7.61750996e-01
5.52119553e-01 5.95792055e-01 4.99573708e-01 -1.28351986e+00
-2.72937775e-01 1.15011364e-01 -5.01348823e-03 -1.14802980e+00
-1.98466152e-01 1.73643768e-01 -7.90035307e-01 5.49591899e-01
1.07462443e-01 -2.13169724e-01 -6.03139937e-01 2.09870839e+00
6.77068353e-01 6.34061575e-01 7.02734292e-02 9.04364884e-01
1.12516832e+00 7.05108941e-01 9.25604429e-04 -5.87814093e-01
1.39746523e+00 -7.16772854e-01 -1.27068114e+00 2.76149750e-01
5.73225394e-02 -1.20322573e+00 1.02349424e+00 1.03606209e-01
-1.16171610e+00 -7.49298692e-01 -7.26273775e-01 -1.22574996e-02
1.23011135e-01 2.67352432e-01 3.19620878e-01 9.14212346e-01
-1.27547288e+00 2.22245559e-01 -4.52734143e-01 -5.30512512e-01
1.05934188e-01 4.64744747e-01 -7.27930963e-01 5.87494671e-01
-1.03478038e+00 7.27820396e-01 -1.57985374e-01 -1.38150603e-01
-8.20971787e-01 -5.96722186e-01 -1.13335407e+00 -2.64129847e-01
9.97737348e-02 -6.00914657e-01 1.17926598e+00 -1.76046669e+00
-2.32489157e+00 1.13035882e+00 -5.74428976e-01 -2.49004122e-02
4.30083483e-01 3.94732580e-02 -4.33455378e-01 3.29102933e-01
-3.79605480e-02 1.08844173e+00 1.26355231e+00 -1.47972727e+00
-3.74412268e-01 -6.87499419e-02 -1.59865916e-01 3.04019928e-01
-3.22275668e-01 4.67497170e-01 -5.25478959e-01 -8.20289373e-01
-1.51538655e-01 -1.03449285e+00 4.05735135e-01 4.49956656e-01
-2.37198129e-01 -1.02993235e-01 1.29820800e+00 -7.04924881e-01
8.93132329e-01 -2.45395684e+00 -8.69238342e-04 -7.98763037e-02
-1.14885502e-01 2.39519358e-01 -1.66765720e-01 2.87897885e-01
-2.62328535e-01 9.85342860e-02 1.34693682e-01 -8.81031513e-01
-2.41715252e-01 1.66516434e-02 -1.33708149e-01 8.15520644e-01
1.24571525e-01 5.55289447e-01 -5.60398817e-01 -8.44530046e-01
3.88550371e-01 1.06894040e+00 -4.70422983e-01 1.43774599e-01
9.03757736e-02 8.57936680e-01 -1.11351795e-01 6.56472623e-01
7.50968277e-01 3.87528896e-01 -7.72982538e-02 -3.93148452e-01
-2.38076746e-01 -1.40854880e-01 -1.22698510e+00 1.49458563e+00
-4.74922568e-01 7.87293613e-01 4.81621087e-01 -4.62698549e-01
1.09143865e+00 7.46022940e-01 6.88413739e-01 -4.04373676e-01
3.27090681e-01 1.44278884e-01 -2.78286636e-01 -5.71892262e-01
2.53796518e-01 -3.17912608e-01 2.36507893e-01 4.30978596e-01
-2.65895179e-03 -2.58261204e-01 -2.82415569e-01 -8.99475142e-02
-7.13478588e-03 2.58852094e-01 2.22482309e-01 -2.60533690e-01
9.67085540e-01 -6.78245425e-01 4.54559416e-01 -2.16076553e-01
-3.88727635e-01 8.24304044e-01 2.89736032e-01 -1.68990046e-01
-1.11726701e+00 -7.49894440e-01 9.89608020e-02 7.61766136e-01
2.32668281e-01 -2.51075953e-01 -1.11936080e+00 -1.16840020e-01
-3.51245880e-01 4.87503439e-01 -6.09325290e-01 -2.88754404e-02
-5.78238070e-01 2.71631815e-02 8.37480187e-01 1.09447517e-01
7.23170578e-01 -1.25909901e+00 -3.26339334e-01 -1.71517506e-01
-3.73146355e-01 -1.19070935e+00 -8.37012708e-01 -9.11582172e-01
-5.09884477e-01 -8.62558544e-01 -1.00056827e+00 -1.22295845e+00
7.47614503e-01 1.09400861e-01 4.65709686e-01 3.33673991e-02
-2.09986120e-01 4.60075736e-01 -1.33999944e-01 -3.85968238e-01
-6.79050684e-01 -4.72438574e-01 5.75073361e-01 5.96586883e-01
-1.62189960e-01 -6.21312737e-01 -4.75330174e-01 4.34762448e-01
-7.00409949e-01 1.01433247e-01 4.93046269e-02 6.00856245e-01
5.20004153e-01 7.66628981e-02 5.09784877e-01 -3.29089761e-01
5.67553639e-01 -2.98419278e-02 -3.69707644e-01 -2.70752725e-03
-1.29516162e-02 -2.63222277e-01 5.23856342e-01 -7.47782230e-01
-1.20584345e+00 1.78394184e-01 -2.85943955e-01 -6.28237247e-01
-3.32477242e-01 -3.11535329e-01 -2.95302689e-01 -3.51709604e-01
3.93312484e-01 2.76932806e-01 7.33023942e-01 -2.28261307e-01
3.19216996e-01 8.41903150e-01 6.43758833e-01 -4.18989241e-01
7.22150862e-01 7.65187383e-01 7.33430237e-02 -1.31410205e+00
-1.48497775e-01 -2.33544439e-01 -5.17635703e-01 -4.19145137e-01
8.91325414e-01 -8.54258239e-01 -1.24114358e+00 7.33440518e-01
-1.26200914e+00 -2.10994422e-01 -1.62409917e-01 4.76159304e-01
-9.55657840e-01 6.52634859e-01 -3.91922235e-01 -1.03488505e+00
-4.77008820e-01 -1.33411968e+00 1.17890704e+00 4.17310387e-01
-3.74818414e-01 -8.42410326e-01 7.79525116e-02 1.53502837e-01
3.20401341e-01 4.94661272e-01 3.84672940e-01 -3.38833854e-02
-2.03293562e-01 9.32293907e-02 1.63309857e-01 2.91394979e-01
7.16949463e-01 5.22822499e-01 -1.07479811e+00 -2.63757259e-01
-4.86967489e-02 -8.43060166e-02 1.32815793e-01 5.95363438e-01
6.38603866e-01 -5.78062117e-01 -1.13699391e-01 6.74023509e-01
9.34769273e-01 2.68223017e-01 8.68471324e-01 8.89642462e-02
4.73764896e-01 9.65763807e-01 5.89728892e-01 4.52186614e-01
2.37032682e-01 8.98028493e-01 4.40126836e-01 -5.24190247e-01
-5.66462934e-01 -3.18682700e-01 5.52833915e-01 5.52903652e-01
-1.60288259e-01 1.92112159e-02 -2.49229923e-01 6.15177691e-01
-1.43030524e+00 -1.13962317e+00 1.12615824e-01 2.03927684e+00
8.93233895e-01 -4.79156405e-01 4.01186436e-01 1.48668289e-01
1.14916515e+00 1.18490629e-01 -2.37147614e-01 -5.14442444e-01
-2.11335197e-02 1.45193160e-01 -1.17205620e-01 8.18541884e-01
-1.05104578e+00 1.12097108e+00 5.83593655e+00 7.94594228e-01
-1.63216829e+00 1.95318773e-01 5.47881246e-01 -1.06281839e-01
-1.97693929e-01 -1.32962286e-01 -6.24880075e-01 4.48350847e-01
5.03238380e-01 -1.87507063e-01 3.21033090e-01 6.88988745e-01
7.64891982e-01 1.03611402e-01 -7.14895785e-01 1.22371638e+00
4.15101230e-01 -1.22280848e+00 7.96358064e-02 3.06604672e-02
6.89890742e-01 -7.22466826e-01 3.42992544e-01 -3.16712797e-01
-2.81082004e-01 -1.03541100e+00 8.26613307e-01 4.59934950e-01
1.32856905e+00 -1.06166148e+00 4.46239322e-01 8.30235705e-02
-1.27088058e+00 2.60393798e-01 2.07997467e-02 3.40808541e-01
3.86648238e-01 -8.70748013e-02 -7.28432178e-01 1.68174580e-01
6.56026006e-01 4.55518126e-01 -2.33731605e-02 6.43553436e-01
-3.53271812e-01 2.73115128e-01 -3.26202959e-01 1.66702405e-01
1.64669342e-02 -2.58856826e-02 7.25000978e-01 1.07777858e+00
3.86060208e-01 2.22562715e-01 -1.95627853e-01 8.41191530e-01
7.66754616e-03 4.79224235e-01 -7.34644532e-01 1.71758652e-01
4.21084106e-01 1.15796840e+00 -4.48021203e-01 -6.54980242e-02
-1.39478981e-01 1.08468997e+00 -4.45312023e-01 1.32236749e-01
-9.17694390e-01 -3.54512513e-01 1.02010536e+00 4.18569416e-01
-8.35662999e-04 -1.93798423e-01 3.76866199e-02 -7.61933625e-01
-2.14519650e-01 -1.12599134e+00 -6.37429506e-02 -9.47674096e-01
-7.85484433e-01 7.97731221e-01 -1.51308313e-01 -1.25728476e+00
-3.49182367e-01 -3.49330381e-02 -8.34196746e-01 7.21590698e-01
-1.52922165e+00 -1.57083046e+00 -3.95568222e-01 1.04287541e+00
7.60406077e-01 -2.48277724e-01 9.26484942e-01 2.19667599e-01
-4.14646208e-01 9.45573092e-01 -2.39477128e-01 4.66860607e-02
1.20356703e+00 -4.95675772e-01 2.47440249e-01 7.55270302e-01
-5.24759898e-03 5.95030606e-01 7.97132373e-01 -6.17486835e-01
-1.10191572e+00 -7.48141646e-01 1.00554848e+00 1.14395484e-01
2.64725357e-01 -2.32192799e-01 -5.50686121e-01 5.74121177e-01
4.32223290e-01 -1.70856252e-01 5.58724880e-01 -5.44065952e-01
-6.23790808e-02 -3.17761362e-01 -1.46864212e+00 8.89497280e-01
6.48961604e-01 -5.02981186e-01 -2.08829612e-01 1.82445422e-01
4.65938568e-01 -2.62752414e-01 -6.73740864e-01 2.67408401e-01
7.87280798e-01 -1.05414784e+00 8.90517831e-01 -1.53811291e-01
1.24846347e-01 -4.92081970e-01 1.23077489e-01 -9.44346726e-01
-1.11731984e-01 -1.42354548e+00 3.69831502e-01 1.71662831e+00
-1.23123385e-01 -5.84392607e-01 6.68282211e-01 3.67154658e-01
1.22332312e-01 -2.75293052e-01 -9.51733828e-01 -4.54876393e-01
-2.23347262e-01 -1.48619926e-02 6.34502411e-01 1.07755089e+00
5.31161353e-02 -5.25170146e-03 -7.79851198e-01 2.55638033e-01
5.88196754e-01 -1.59732684e-01 1.10710132e+00 -1.01254392e+00
1.31002307e-01 -3.61239910e-01 -5.44877887e-01 -6.32212162e-01
5.44506431e-01 -2.74205118e-01 -5.74068874e-02 -1.09978437e+00
-1.73257470e-01 -2.63896555e-01 4.16025698e-01 4.24412161e-01
2.25242257e-01 5.93254745e-01 2.91690469e-01 1.48086339e-01
4.34738491e-03 5.74892104e-01 1.41431034e+00 1.98884562e-01
-3.86807382e-01 2.80593242e-02 -5.27847588e-01 8.91248465e-01
7.84765184e-01 -2.81891912e-01 -5.82678854e-01 -3.54277194e-02
-4.19942200e-01 2.17800334e-01 3.08405429e-01 -8.12554300e-01
1.91164985e-01 -2.30805382e-01 6.06985576e-02 -1.46054387e-01
8.40924799e-01 -6.60776734e-01 4.03479666e-01 4.46855187e-01
-1.24328516e-01 1.82561912e-02 3.97701532e-01 1.52689040e-01
-3.36252630e-01 -9.76115931e-03 1.34363735e+00 -1.29160091e-01
-5.78711987e-01 2.33411476e-01 -5.08648038e-01 -2.21724182e-01
1.28434384e+00 -5.20779848e-01 1.92530807e-02 -1.05284035e+00
-5.98949492e-01 -3.95556808e-01 7.06647456e-01 4.82843786e-01
7.51723766e-01 -1.35198915e+00 -9.63395178e-01 6.84673369e-01
-3.45470935e-01 -3.72854829e-01 2.29186863e-01 9.19754267e-01
-7.61479735e-01 9.13206786e-02 -5.98831892e-01 -6.59398496e-01
-1.91286147e+00 3.72357905e-01 3.61478865e-01 4.22161192e-01
-5.23547649e-01 8.21308911e-01 4.37650979e-01 -8.12215323e-04
3.40485573e-01 1.48593202e-01 -3.69490921e-01 8.89578238e-02
5.38247049e-01 3.90619427e-01 -3.12261164e-01 -1.36209893e+00
-4.87010032e-01 1.21085322e+00 3.64463300e-01 -2.96436429e-01
1.00347662e+00 -3.78150791e-01 -1.18462108e-01 8.19665268e-02
1.25134861e+00 5.24950504e-01 -1.30127907e+00 2.27932528e-01
-7.42604375e-01 -5.13427258e-01 -9.66894552e-02 -2.70111501e-01
-1.32840109e+00 8.97985339e-01 6.10273302e-01 -3.01778227e-01
1.33811212e+00 -1.91933796e-01 8.10942173e-01 -3.58565539e-01
2.05487892e-01 -7.20188856e-01 4.37889040e-01 1.69742554e-01
1.04670858e+00 -9.36656773e-01 -1.67034462e-01 -5.94102502e-01
-8.81978035e-01 1.22191334e+00 4.62582618e-01 4.41534678e-03
5.92781782e-01 3.62647712e-01 3.56503665e-01 1.72877580e-01
-2.93209046e-01 -2.11694300e-01 3.32942188e-01 9.55057323e-01
3.68978918e-01 -3.72175902e-01 -1.64084062e-01 1.09819464e-01
-3.89726549e-01 -9.11672860e-02 3.96396279e-01 4.62269694e-01
-2.99971461e-01 -9.95467722e-01 -7.00642765e-01 -6.95785463e-01
-6.35887980e-01 3.10544875e-02 -4.59728509e-01 9.54315543e-01
2.01770186e-01 1.10734916e+00 1.78656280e-01 -1.49412677e-01
1.44099951e-01 8.53953976e-03 5.69624662e-01 -2.70045221e-01
-4.58560526e-01 4.53278065e-01 -3.95689420e-02 -4.55190241e-01
-7.41607845e-01 -5.51454484e-01 -1.37951744e+00 -6.50367320e-01
-2.17040360e-01 -7.16361329e-02 8.04139912e-01 5.34892797e-01
4.53491598e-01 1.39820173e-01 7.72111118e-01 -1.17182744e+00
7.12985098e-02 -8.02605033e-01 -5.50436497e-01 5.72927475e-01
5.91033936e-01 -7.06081688e-01 -5.64354062e-01 4.36926216e-01] | [13.217726707458496, -0.41062334179878235] |
a6a4f16e-6fe2-469e-86cc-4f8c845ec0b6 | embedding-space-augmentation-for-weakly | 2210.17013 | null | https://arxiv.org/abs/2210.17013v1 | https://arxiv.org/pdf/2210.17013v1.pdf | Embedding Space Augmentation for Weakly Supervised Learning in Whole-Slide Images | Multiple Instance Learning (MIL) is a widely employed framework for learning on gigapixel whole-slide images (WSIs) from WSI-level annotations. In most MIL based analytical pipelines for WSI-level analysis, the WSIs are often divided into patches and deep features for patches (i.e., patch embeddings) are extracted prior to training to reduce the overall computational cost and cope with the GPUs' limited RAM. To overcome this limitation, we present EmbAugmenter, a data augmentation generative adversarial network (DA-GAN) that can synthesize data augmentations in the embedding space rather than in the pixel space, thereby significantly reducing the computational requirements. Experiments on the SICAPv2 dataset show that our approach outperforms MIL without augmentation and is on par with traditional patch-level augmentation for MIL training while being substantially faster. | ['Faisal Mahmood', 'Nasir Rajpoot', 'Guillaume Jaume', 'Imaad Zaffar'] | 2022-10-31 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 6.01949334e-01 3.58446568e-01 -3.30846310e-01 -1.15810283e-01
-1.72118258e+00 -5.27602553e-01 4.16801900e-01 1.48985133e-01
-2.49006122e-01 7.15256810e-01 2.25685641e-01 -1.58483371e-01
3.24193180e-01 -8.73778045e-01 -1.06812334e+00 -1.12716293e+00
3.84435773e-01 3.60320926e-01 1.55941054e-01 2.67082483e-01
-1.85397435e-02 4.74579901e-01 -1.23593509e+00 6.36876881e-01
4.74120051e-01 1.01859856e+00 -4.44136620e-01 1.01087296e+00
-1.67594269e-01 8.13185394e-01 -6.87551022e-01 -3.52300525e-01
3.53233397e-01 -1.64970279e-01 -7.73211360e-01 9.23049450e-02
8.92850935e-01 -2.31251836e-01 -3.30348253e-01 4.53228772e-01
5.72234035e-01 -2.27573916e-01 6.34390116e-01 -1.25751281e+00
-1.07679717e-01 1.66315883e-01 -9.86085236e-01 8.93397406e-02
-1.54867455e-01 4.60657567e-01 1.08225012e+00 -1.01208830e+00
8.30863953e-01 8.35135520e-01 9.91696775e-01 5.62301397e-01
-1.50724208e+00 -2.99313247e-01 -2.56345004e-01 -2.13177338e-01
-1.36149812e+00 -2.11438909e-01 6.31235719e-01 -2.16815352e-01
1.01639080e+00 5.52981257e-01 7.10273027e-01 1.06476092e+00
3.62497866e-01 1.14181697e+00 1.13901269e+00 -4.59488124e-01
4.61125523e-02 -2.52970904e-01 -1.01256154e-01 7.37174034e-01
-1.26047190e-02 -2.64829606e-01 -6.68469131e-01 -2.38555253e-01
9.84771013e-01 -1.64604872e-01 8.66403505e-02 -1.39073476e-01
-1.33377457e+00 6.61655664e-01 5.90851903e-01 5.59952930e-02
-2.07490817e-01 4.65879202e-01 7.33202636e-01 -3.71276215e-02
7.35036671e-01 5.94820023e-01 -3.92556727e-01 -6.10141493e-02
-1.14513302e+00 3.33697945e-01 2.68245727e-01 6.83146954e-01
9.03747380e-01 -8.79094228e-02 -5.02804160e-01 6.73794270e-01
-1.46790475e-01 1.19468875e-01 3.60148013e-01 -8.37838531e-01
2.89105505e-01 8.28017831e-01 -1.62402675e-01 -4.80937392e-01
-2.10587442e-01 -5.20280480e-01 -8.96374583e-01 3.67245317e-01
6.57291293e-01 -6.42687306e-02 -1.18409622e+00 1.42084408e+00
5.99646926e-01 6.73619866e-01 -1.61774457e-01 6.74396574e-01
1.03244352e+00 7.00930536e-01 2.81915158e-01 3.07534486e-01
1.34050500e+00 -1.35188091e+00 -3.42278272e-01 -3.58077198e-01
8.97166312e-01 -6.56328499e-01 1.46147633e+00 3.61675978e-01
-1.12416291e+00 -5.22676826e-01 -1.16633856e+00 -4.82977331e-01
-5.69361806e-01 3.16583931e-01 7.98325837e-01 4.55975354e-01
-1.08664644e+00 4.02683228e-01 -1.13067818e+00 3.84692438e-02
9.96480942e-01 4.54820484e-01 -5.67204893e-01 -7.63703287e-02
-7.12916255e-01 3.34783375e-01 7.73026049e-02 -2.09080175e-01
-8.77936244e-01 -1.59387589e+00 -8.51465523e-01 9.27830413e-02
1.32671684e-01 -6.01716995e-01 7.35768616e-01 -7.16334045e-01
-1.51558709e+00 1.22898936e+00 -5.34177870e-02 -4.02175903e-01
5.41829705e-01 8.76725838e-02 1.55024156e-01 9.74272117e-02
-7.47966096e-02 9.21045244e-01 8.93147767e-01 -8.78486812e-01
-3.93609792e-01 -2.23282367e-01 -9.91906896e-02 1.13982059e-01
-3.16617310e-01 -3.35550010e-01 -6.29122555e-01 -8.06874931e-01
-1.08139552e-01 -9.38244700e-01 -3.26811045e-01 2.68531591e-01
-7.11644292e-01 -6.21765219e-02 9.05111372e-01 -7.48444498e-01
7.04218805e-01 -2.25414419e+00 6.74644709e-02 1.91531822e-01
4.23439980e-01 2.79859066e-01 -4.03686434e-01 2.02053219e-01
-6.01067506e-02 1.95459574e-02 -1.92106336e-01 -6.50913596e-01
-2.26461425e-01 1.67015061e-01 -1.23025775e-01 4.44200963e-01
5.51236510e-01 1.37880409e+00 -7.77482748e-01 -4.78862554e-01
3.53963047e-01 6.32621109e-01 -5.80133855e-01 2.86194712e-01
-2.33864442e-01 6.77092254e-01 -1.14845321e-01 8.58972371e-01
6.13678575e-01 -4.51596826e-01 -5.86552471e-02 -4.44950670e-01
1.79467261e-01 1.48029000e-01 -9.56148326e-01 1.98142362e+00
-6.27291381e-01 7.92790353e-01 1.25972375e-01 -1.06371915e+00
6.46966696e-01 1.49857804e-01 6.31529808e-01 -4.42772090e-01
6.75717741e-02 9.68273282e-02 -3.16258192e-01 1.06200412e-01
2.81162053e-01 -2.22214554e-02 -9.73367915e-02 3.59921366e-01
1.52379230e-01 -2.54789680e-01 -1.41539900e-02 1.48610637e-01
1.44021332e+00 1.89719215e-01 1.70877457e-01 -2.54997730e-01
3.64918381e-01 3.07947576e-01 4.11982954e-01 6.29586339e-01
5.05699962e-02 9.37484562e-01 6.87732935e-01 -6.05449498e-01
-1.58541501e+00 -1.18509841e+00 -2.64209360e-01 9.70823288e-01
-2.82057434e-01 -3.78211111e-01 -9.48491871e-01 -8.55691493e-01
7.29661994e-03 1.22884236e-01 -9.11947429e-01 2.25407660e-01
-5.41106820e-01 -1.07184684e+00 7.50386357e-01 8.24178576e-01
2.06407875e-01 -8.00941288e-01 -2.15666056e-01 2.88870543e-01
1.62890464e-01 -1.11075211e+00 -3.02762479e-01 9.38436389e-02
-6.40881181e-01 -1.10697234e+00 -6.49252594e-01 -6.15618706e-01
1.04449153e+00 -1.17821604e-01 1.29268765e+00 2.41866753e-01
-1.15889084e+00 3.91923189e-01 -1.48059696e-01 -5.74902773e-01
-3.10425490e-01 3.52536857e-01 -4.64563131e-01 8.10807943e-02
2.00474471e-01 -4.72560525e-01 -9.58328307e-01 -6.47169724e-02
-1.11540163e+00 3.87995750e-01 7.42642462e-01 1.47156954e+00
1.25276828e+00 -1.27944887e-01 4.16160673e-01 -1.16205812e+00
2.37445191e-01 -2.74403393e-01 -5.68943083e-01 2.84662426e-01
-2.16174334e-01 -1.52625680e-01 5.93166232e-01 -4.02625293e-01
-9.16304648e-01 1.83829665e-02 -4.49602067e-01 -2.13383108e-01
-1.69374689e-01 2.83635765e-01 -9.74135473e-02 -5.28556824e-01
7.25214005e-01 2.67564178e-01 3.43748033e-01 -9.71777271e-03
2.85487086e-01 3.60053241e-01 6.44198120e-01 -5.31504691e-01
6.96561396e-01 5.57455838e-01 1.92569688e-01 -7.05843151e-01
-9.43849087e-01 -4.75385070e-01 -5.79240143e-01 -8.84869769e-02
7.88813949e-01 -1.03674614e+00 -5.01474857e-01 4.70902532e-01
-6.39262855e-01 -8.75276566e-01 -5.96970320e-01 -1.89958941e-02
-5.92517972e-01 -8.10134783e-03 -1.04164457e+00 -1.55208126e-01
-7.38339126e-01 -1.26750779e+00 1.61378849e+00 2.84027427e-01
-2.85550565e-01 -1.11391294e+00 1.58453345e-01 5.85021019e-01
4.21247870e-01 9.13373768e-01 8.58177423e-01 -3.39477271e-01
-7.67087281e-01 -7.30235815e-01 -3.70679319e-01 2.78562099e-01
-9.64903086e-03 1.48290098e-01 -1.17465091e+00 -3.72181982e-01
-5.03983796e-01 -7.27307260e-01 7.18517423e-01 5.98289907e-01
1.66426492e+00 -8.61106589e-02 -4.37826276e-01 8.43250155e-01
1.72336733e+00 -5.36267161e-01 1.06532359e+00 2.70424992e-01
7.56712198e-01 1.16990969e-01 6.06568277e-01 2.73524463e-01
-1.21910535e-01 5.72349608e-01 3.63127172e-01 -8.41073036e-01
-6.63014233e-01 -2.08135188e-01 -2.08867993e-02 3.17340016e-01
2.19097182e-01 -3.11357439e-01 -8.35353017e-01 7.27154195e-01
-1.53830338e+00 -4.93292183e-01 -6.67719729e-03 1.83651817e+00
9.81804848e-01 -1.57983683e-03 -1.77372992e-02 3.02163869e-01
1.88035399e-01 2.11478919e-01 -6.97875619e-01 -2.94888765e-01
-6.45539761e-02 7.50808537e-01 5.91724992e-01 4.28999543e-01
-1.14093590e+00 8.65988612e-01 6.44205523e+00 1.28071010e+00
-1.17218518e+00 3.26568961e-01 1.32136893e+00 -3.21179390e-01
-2.27547526e-01 -3.68959486e-01 -8.38408828e-01 3.75715941e-01
7.96553850e-01 5.00576615e-01 2.50809997e-01 8.45170557e-01
-3.94392699e-01 -6.48654550e-02 -9.57683861e-01 6.57123327e-01
-1.17336921e-01 -2.07686543e+00 -1.87804207e-01 1.98041365e-01
1.05751932e+00 1.64592341e-01 3.28867495e-01 2.87983209e-01
3.11338216e-01 -1.18534636e+00 -2.05452684e-02 1.80460617e-01
1.24617290e+00 -9.19061482e-01 1.03848112e+00 4.56671454e-02
-9.39811110e-01 4.29844379e-01 -2.24803835e-01 3.03126186e-01
-1.29672438e-01 7.69869506e-01 -1.24350834e+00 3.83123815e-01
5.00363529e-01 3.90711963e-01 -6.74127936e-01 6.53986454e-01
-2.26003140e-01 8.29558969e-01 -4.60534722e-01 3.42787653e-01
4.94108796e-01 1.20434158e-01 1.95795983e-01 1.31281233e+00
3.00091445e-01 -1.09798245e-01 -1.14101641e-01 7.96383321e-01
-2.58810639e-01 1.35228887e-01 -3.58892024e-01 -9.46858600e-02
1.84688047e-01 1.66042495e+00 -7.42817402e-01 -4.43860888e-01
-5.51144242e-01 1.07332325e+00 4.98717070e-01 8.48421231e-02
-7.91662633e-01 -2.01301187e-01 6.87177539e-01 3.66527379e-01
1.19889811e-01 8.89482424e-02 -6.67126894e-01 -8.78283203e-01
-1.59381419e-01 -7.20344782e-01 4.22902942e-01 -4.71783996e-01
-1.40117681e+00 3.48755985e-01 -5.74387193e-01 -1.22197521e+00
1.08113877e-01 -8.21990728e-01 -8.53662252e-01 7.76277363e-01
-1.63594234e+00 -1.88554227e+00 -6.45332694e-01 4.27000642e-01
3.50017667e-01 -2.32886747e-02 9.81306911e-01 1.77149698e-01
-6.65299177e-01 1.09193206e+00 2.67088324e-01 2.43736446e-01
7.50463307e-01 -1.31143010e+00 5.11822879e-01 5.81235826e-01
1.54622972e-01 3.73760313e-01 3.80994856e-01 -4.02777404e-01
-1.52545309e+00 -1.23516595e+00 3.27440470e-01 -5.76542318e-01
7.06781745e-01 -4.56112206e-01 -9.24280107e-01 7.81880677e-01
1.14646420e-01 6.71171427e-01 1.29742444e+00 8.12862143e-02
-1.76807404e-01 -1.60995379e-01 -1.37534499e+00 6.61876976e-01
6.31330490e-01 -4.47885364e-01 5.09050973e-02 5.86103261e-01
3.52126718e-01 -9.61138248e-01 -1.39159596e+00 3.79435569e-01
4.46259677e-01 -6.33522153e-01 1.38201153e+00 -5.23589373e-01
7.40085006e-01 -2.87700683e-01 1.67385608e-01 -1.16236866e+00
-1.04486093e-01 -3.03006679e-01 -8.75229016e-02 1.27384531e+00
3.55548948e-01 -4.55059081e-01 1.50656247e+00 5.74475884e-01
-2.48588920e-01 -1.07512796e+00 -1.19253170e+00 -3.16519529e-01
8.66904110e-02 -2.13161737e-01 6.64168239e-01 8.53632629e-01
-1.77642092e-01 -8.49756598e-03 -2.52711207e-01 1.85659632e-01
8.95385683e-01 -8.03350061e-02 1.18099535e+00 -7.34106243e-01
-4.59235847e-01 -3.29164058e-01 -4.94324684e-01 -5.08217096e-01
1.33827552e-01 -9.40792680e-01 -2.30255015e-02 -1.38561428e+00
2.19335184e-01 -6.49376869e-01 -3.63729924e-01 7.05127954e-01
-4.69419718e-01 1.14700019e+00 -2.67821819e-01 1.69042442e-02
-3.59714687e-01 1.37754366e-01 1.33234322e+00 -4.07626182e-01
3.65249366e-02 -4.42813098e-01 -5.10686159e-01 4.69428569e-01
6.35495484e-01 -2.38160804e-01 -1.66329220e-01 -3.41529340e-01
1.52881190e-01 -6.75785989e-02 4.42500889e-01 -1.11837649e+00
6.76687509e-02 8.29655007e-02 6.72887325e-01 -5.67914963e-01
4.05459404e-01 -5.35563231e-01 2.57218242e-01 2.77673513e-01
-3.91476870e-01 -5.70275187e-01 4.99372691e-01 3.05206418e-01
-3.24077547e-01 4.89475541e-02 1.00112414e+00 6.54086024e-02
-4.17431712e-01 6.37582839e-01 -1.71338499e-01 -2.12151498e-01
1.11069155e+00 -1.32809535e-01 -4.66326565e-01 8.97236988e-02
-8.05121958e-01 1.07523270e-01 5.76569676e-01 -9.47457552e-02
4.18097407e-01 -1.46833384e+00 -7.60572255e-01 5.54027140e-01
2.92219937e-01 5.42420626e-01 6.90069556e-01 1.01331782e+00
-1.03340900e+00 1.86077386e-01 -1.66381881e-01 -7.52966702e-01
-1.18710554e+00 2.71679759e-01 1.30551994e-01 -1.03102338e+00
-6.71337903e-01 1.32573783e+00 3.97770494e-01 -5.62173486e-01
1.09146042e-02 -1.38465255e-01 2.87126392e-01 -2.02240869e-01
7.10063517e-01 1.51691809e-01 2.39720866e-01 -1.91464648e-01
-6.10472783e-02 3.73429567e-01 -2.48787537e-01 2.42663592e-01
1.41537297e+00 3.20534438e-01 -9.55014750e-02 1.40207773e-02
1.28480613e+00 1.83420181e-02 -1.65512204e+00 -2.67190784e-01
-3.74577045e-01 -3.78315419e-01 3.38952243e-01 -6.80269539e-01
-1.18785608e+00 9.59301114e-01 5.23734510e-01 -1.71054602e-01
1.08915079e+00 -1.86502794e-03 1.05486977e+00 -2.59073228e-01
1.56647325e-01 -9.94690657e-01 3.15849930e-01 -1.74736828e-01
7.33765066e-01 -1.27089322e+00 2.14986712e-01 -5.62777102e-01
-5.49954712e-01 1.16148639e+00 7.08956301e-01 -4.39080238e-01
2.72732198e-01 8.75184655e-01 8.52462947e-02 -1.93874031e-01
-4.63553101e-01 2.63713598e-01 3.72511864e-01 5.43526113e-01
4.15515542e-01 4.01896704e-03 1.36587233e-03 3.64243746e-01
2.99106687e-02 -9.64968204e-02 1.24439247e-01 7.32219577e-01
1.21171139e-01 -1.23904717e+00 -2.84134328e-01 8.14355671e-01
-6.47187054e-01 -1.43335149e-01 -3.10315955e-02 8.46164465e-01
2.19593406e-01 2.43643194e-01 4.53126222e-01 -2.51213104e-01
5.64259067e-02 -1.77225359e-02 5.55532813e-01 -5.11507332e-01
-6.36291444e-01 1.27614573e-01 1.67815387e-02 -7.30403304e-01
-2.32115999e-01 -4.88539606e-01 -1.17688560e+00 -2.91374594e-01
-1.03124969e-01 -1.76729858e-01 6.33064210e-01 8.34518194e-01
3.80629003e-01 9.06610787e-01 4.82704043e-01 -1.08291173e+00
4.05506119e-02 -8.98090959e-01 -4.84058917e-01 3.27437788e-01
2.68578529e-01 -3.32214862e-01 -4.24989462e-01 -4.98328416e-04] | [15.076680183410645, -2.9110348224639893] |
a1d43c32-5a79-4100-a594-5677b2ef415e | pre-training-and-fine-tuning-neural-topic-1 | null | null | https://aclanthology.org/2022.acl-long.413 | https://aclanthology.org/2022.acl-long.413.pdf | Pre-training and Fine-tuning Neural Topic Model: A Simple yet Effective Approach to Incorporating External Knowledge | Recent years have witnessed growing interests in incorporating external knowledge such as pre-trained word embeddings (PWEs) or pre-trained language models (PLMs) into neural topic modeling. However, we found that employing PWEs and PLMs for topic modeling only achieved limited performance improvements but with huge computational overhead. In this paper, we propose a novel strategy to incorporate external knowledge into neural topic modeling where the neural topic model is pre-trained on a large corpus and then fine-tuned on the target dataset. Experiments have been conducted on three datasets and results show that the proposed approach significantly outperforms both current state-of-the-art neural topic models and some topic modeling approaches enhanced with PWEs or PLMs. Moreover, further study shows that the proposed approach greatly reduces the need for the huge size of training data. | ['Yunbo Cao', 'Qian-Wen Zhang', 'Deyu Zhou', 'Boyu Wang', 'Xuemeng Hu', 'Linhai Zhang'] | null | null | null | null | acl-2022-5 | ['topic-models'] | ['natural-language-processing'] | [-1.30689442e-01 3.81725281e-01 -3.36550266e-01 -3.89748454e-01
-5.39033651e-01 2.42688376e-02 9.42046463e-01 7.47555345e-02
-5.71271956e-01 6.10140800e-01 3.63297433e-01 -2.30648220e-01
3.99182774e-02 -1.22829986e+00 -5.73258758e-01 -4.95042741e-01
6.33499119e-04 6.12213016e-01 5.43287516e-01 1.48983166e-01
3.73411894e-01 -1.68758512e-01 -1.35706329e+00 -2.69415230e-01
9.48008776e-01 7.19495952e-01 5.80033720e-01 2.64605671e-01
-7.89788008e-01 1.17754333e-01 -6.54540360e-01 -3.64251435e-01
-1.70792416e-01 2.46982142e-01 -5.15610397e-01 -7.94037133e-02
1.88080624e-01 -2.21098259e-01 -3.16915125e-01 7.22093821e-01
3.97064030e-01 5.82396328e-01 6.19800448e-01 -1.07746208e+00
-9.07666028e-01 6.98928714e-01 -5.33206522e-01 1.00780107e-01
-4.74091589e-01 -4.15200800e-01 8.42745781e-01 -1.02577400e+00
5.28810382e-01 1.16438651e+00 7.58573174e-01 5.11235714e-01
-9.53117907e-01 -9.73921895e-01 5.27313709e-01 7.12504610e-02
-1.38062251e+00 1.49055690e-01 8.81959498e-01 -3.84224057e-01
1.28741908e+00 -4.88613605e-01 4.76399094e-01 9.61840093e-01
3.37087840e-01 7.57556081e-01 8.10301542e-01 -4.64830071e-01
4.75183517e-01 7.93395698e-01 6.17501438e-01 3.61161768e-01
4.89838928e-01 -3.60636771e-01 -4.16689605e-01 -5.60280085e-01
8.10778737e-01 -6.12669578e-03 1.59743309e-01 -3.51669699e-01
-7.43496954e-01 1.31477284e+00 1.12578005e-01 4.37627345e-01
-7.17564464e-01 1.01623222e-01 4.86025155e-01 -2.73100704e-01
1.14877498e+00 4.66066390e-01 -6.47905469e-01 -1.57726035e-01
-1.29759836e+00 1.39438495e-01 6.74758315e-01 1.01435006e+00
7.94610381e-01 1.48327246e-01 1.18065432e-01 1.18840051e+00
5.02178013e-01 2.29183793e-01 1.12378025e+00 -2.81581610e-01
4.87184197e-01 5.50205052e-01 -5.13784923e-02 -9.86080766e-01
-1.91661939e-01 -5.47621548e-01 -2.92696476e-01 -3.79430562e-01
-2.12813497e-01 -4.61180866e-01 -1.21526027e+00 1.72372580e+00
3.28189313e-01 6.51594758e-01 3.63447040e-01 1.67033121e-01
7.17247248e-01 1.30762160e+00 6.84017181e-01 2.23224778e-02
1.37717175e+00 -1.17052615e+00 -9.20720398e-01 -4.07170713e-01
5.91143131e-01 -6.46219313e-01 1.06086218e+00 4.03610706e-01
-8.06333959e-01 -5.48643887e-01 -8.80573690e-01 1.91094026e-01
-9.24713016e-01 3.72144312e-01 8.53968382e-01 1.00818598e+00
-1.14763808e+00 2.02857167e-01 -1.02678227e+00 -7.71958888e-01
5.72076917e-01 4.59018916e-01 -4.76089790e-02 -1.24559909e-01
-1.45935714e+00 8.93124104e-01 8.58759940e-01 -5.05592465e-01
-9.54327762e-01 -8.86437237e-01 -8.09617460e-01 3.00325245e-01
4.19717699e-01 -3.88947189e-01 1.08726978e+00 -3.45303267e-01
-1.59741580e+00 2.49577105e-01 -2.38201022e-01 -6.62476301e-01
-1.99786708e-01 -6.04728222e-01 -4.68340039e-01 -4.74682450e-02
-7.85352662e-02 1.02144396e+00 7.09869027e-01 -1.26472139e+00
-7.06618369e-01 -9.79044735e-02 -1.83610529e-01 2.21250683e-01
-1.24573016e+00 1.52574047e-01 -7.12685347e-01 -7.38029361e-01
-1.32965162e-01 -7.16875017e-01 -3.71743321e-01 -4.49290603e-01
-6.30248562e-02 -8.36770713e-01 1.30147839e+00 -6.14719391e-01
1.52786541e+00 -1.76956201e+00 -2.29621887e-01 1.40278906e-01
-9.40412357e-02 7.34144390e-01 -8.02387819e-02 5.33873320e-01
1.67709067e-01 2.49396816e-01 1.07307665e-01 -6.10089660e-01
1.15949459e-01 3.20440352e-01 -6.91582263e-01 -1.54114112e-01
4.59792390e-02 6.44439936e-01 -6.11105263e-01 -5.29886305e-01
3.31352562e-01 8.51923168e-01 -4.97385114e-01 2.74664938e-01
-4.34312493e-01 -5.25700927e-01 -5.49042165e-01 2.80851066e-01
5.49540758e-01 -1.36945352e-01 4.40947339e-02 9.32567492e-02
1.41851783e-01 4.87267315e-01 -1.11201990e+00 1.57339609e+00
-7.03782260e-01 6.50763750e-01 -4.22214478e-01 -8.76596808e-01
1.32621050e+00 6.43706024e-01 2.01493323e-01 -1.25112891e-01
2.57681727e-01 -1.97240293e-01 -3.20906878e-01 -1.97428748e-01
1.01913691e+00 -2.76804239e-01 5.57865538e-02 7.48283565e-01
7.08249152e-01 -1.37830852e-02 7.38823265e-02 2.13853970e-01
4.96524245e-01 1.77345574e-02 2.79913068e-01 -3.24427933e-01
2.13297844e-01 6.98396266e-02 3.80116910e-01 5.93407214e-01
1.07838605e-02 2.58179516e-01 3.59962881e-02 -1.84355557e-01
-1.00477946e+00 -8.70682299e-01 -2.41586149e-01 1.33717859e+00
-2.62285322e-01 -8.52865160e-01 -8.80359232e-01 -6.04086399e-01
-2.70659715e-01 1.02982664e+00 -7.83500969e-01 -1.35904759e-01
-2.39379421e-01 -9.89093006e-01 6.17106497e-01 8.88777316e-01
3.65913540e-01 -9.47156310e-01 -5.06953657e-01 4.98138279e-01
2.18577430e-01 -9.05196488e-01 -1.88390642e-01 1.37152523e-01
-1.44507861e+00 -3.88871998e-01 -9.19446647e-01 -8.16167772e-01
5.01983702e-01 2.98449874e-01 8.40696037e-01 -4.60304171e-01
-7.00453529e-04 2.59539187e-01 -3.11349928e-01 -8.33287954e-01
2.35825926e-01 4.61436182e-01 6.56809583e-02 -2.81582385e-01
9.88331676e-01 -5.91967702e-01 -1.84392124e-01 1.12408794e-01
-1.05298042e+00 -8.60224813e-02 3.97924244e-01 7.50815928e-01
2.65507579e-01 4.56420064e-01 9.99589980e-01 -1.14394975e+00
1.23132920e+00 -7.26761818e-01 -3.91805470e-01 2.12903708e-01
-9.35641646e-01 -6.03855364e-02 1.30513906e-01 -8.24499130e-01
-1.67590463e+00 -5.27190387e-01 -2.68792175e-02 -3.89820993e-01
-3.23636889e-01 9.73625124e-01 2.24368162e-02 3.53362799e-01
3.36939275e-01 2.82102466e-01 -3.17203373e-01 -5.44651568e-01
5.68215311e-01 8.69518399e-01 1.76089227e-01 -6.38417006e-01
5.25276780e-01 2.78142840e-01 -7.14592099e-01 -8.93713176e-01
-6.28099978e-01 -6.91672325e-01 -4.33963746e-01 2.08049342e-01
6.67355001e-01 -1.22691202e+00 2.67576128e-01 4.53685135e-01
-1.02545643e+00 -2.03492686e-01 -1.84697405e-01 7.92154491e-01
-2.66683131e-01 9.97408777e-02 -4.76899594e-01 -8.01279545e-01
-6.07834756e-01 -8.33676696e-01 7.81115532e-01 4.48146433e-01
-3.57800335e-01 -1.62296402e+00 3.80087525e-01 1.18545495e-01
1.03319430e+00 -4.86020833e-01 1.14049661e+00 -1.29463816e+00
-2.82504141e-01 -4.14148122e-01 -2.27138937e-01 2.84913242e-01
3.35362136e-01 -1.00339577e-01 -1.21201539e+00 1.36034386e-02
2.79200319e-02 -2.63655961e-01 8.53718996e-01 6.20356917e-01
9.20302153e-01 2.50204355e-02 -7.49215066e-01 1.39739588e-01
1.52076733e+00 3.45805943e-01 5.37007153e-01 6.08425140e-01
4.21431065e-01 4.55415994e-01 6.53407991e-01 3.31447214e-01
6.11227393e-01 4.27778751e-01 -1.19241007e-01 -2.38134526e-02
1.01935551e-01 -4.34001565e-01 2.69913644e-01 9.50649917e-01
-2.49082036e-02 -3.59828293e-01 -9.30754960e-01 1.06348598e+00
-1.69465446e+00 -5.02179444e-01 3.84190083e-01 1.88051975e+00
1.04726493e+00 2.33099505e-01 -1.18091851e-01 -1.60724089e-01
6.38734221e-01 2.46078864e-01 -3.40863019e-01 -5.66455781e-01
3.34694684e-01 4.46273565e-01 2.59617478e-01 3.53105962e-01
-1.24600410e+00 1.51178443e+00 6.66036415e+00 1.13158035e+00
-1.16865933e+00 6.39574468e-01 2.97918200e-01 4.59254980e-02
-1.28430471e-01 -1.67323634e-01 -1.30178332e+00 1.89091519e-01
1.36098766e+00 -5.20474374e-01 -3.73414874e-01 1.44090533e+00
-1.69062778e-01 -1.18076414e-01 -4.94473040e-01 4.01567400e-01
4.42449033e-01 -1.18694019e+00 3.24469268e-01 1.22661844e-01
9.73827958e-01 2.40605563e-01 1.60197578e-02 8.18640888e-01
6.09179080e-01 -6.99581563e-01 9.36442614e-02 2.41508156e-01
1.65694624e-01 -8.14811945e-01 1.06534278e+00 3.65441114e-01
-9.98464882e-01 2.59493917e-01 -9.35255826e-01 1.66437998e-01
3.38653713e-01 5.34454882e-01 -1.20015097e+00 3.42280805e-01
5.46656132e-01 3.71345460e-01 -4.24450696e-01 1.14438975e+00
-1.19463280e-01 1.14846730e+00 -4.00063217e-01 -9.09170434e-02
6.66463614e-01 1.04535127e-03 2.20283300e-01 1.40211535e+00
5.39926052e-01 -1.30313054e-01 6.73339665e-02 7.44183302e-01
6.46296591e-02 4.10000145e-01 -5.79811811e-01 -3.82957637e-01
3.77772063e-01 1.22575998e+00 -8.28364909e-01 -8.88270080e-01
-7.07888544e-01 2.88489252e-01 2.25549310e-01 3.25188518e-01
-7.52489090e-01 -4.37166512e-01 6.07713759e-01 -8.55652988e-02
7.04041779e-01 -3.76080185e-01 -2.05851838e-01 -9.69436169e-01
-4.69371140e-01 -2.03337982e-01 2.82417566e-01 -8.35933626e-01
-1.12576687e+00 9.38334048e-01 5.33660769e-01 -5.63103437e-01
-2.94684082e-01 -5.31429112e-01 -1.03203440e+00 9.85762715e-01
-1.61132514e+00 -1.30289161e+00 6.29527569e-02 4.10536736e-01
9.47425961e-01 -4.40878004e-01 9.11835849e-01 1.00204259e-01
-5.18164754e-01 4.84911919e-01 2.34883070e-01 -2.63216525e-01
7.53619313e-01 -1.01864767e+00 5.07218122e-01 5.20634830e-01
1.36719942e-01 1.22843778e+00 6.27041459e-01 -9.74231124e-01
-7.30787218e-01 -1.29081702e+00 1.16991282e+00 -2.20522955e-01
6.55236542e-01 -4.90393072e-01 -1.29028511e+00 9.02286232e-01
6.30823791e-01 -5.35393059e-01 1.25356734e+00 7.04942226e-01
-3.29113960e-01 2.50095993e-01 -8.83345366e-01 3.36563528e-01
1.99553698e-01 -3.48170280e-01 -1.25438309e+00 2.18095973e-01
1.02609062e+00 4.52083722e-02 -9.90803123e-01 2.04451010e-01
4.49202836e-01 -1.17707685e-01 9.10187244e-01 -7.00186431e-01
2.22301453e-01 8.34813267e-02 -1.79549560e-01 -1.55050230e+00
-5.31428009e-02 4.59966436e-02 -3.22218120e-01 1.73713338e+00
6.01808012e-01 -7.01018333e-01 9.55855966e-01 9.43319798e-01
-6.05077520e-02 -8.43338192e-01 -4.51794952e-01 -6.72916174e-01
4.25742120e-01 -6.37929678e-01 5.30267954e-01 1.04594767e+00
1.39299884e-01 1.66097403e-01 -4.15648580e-01 7.44837523e-02
1.80770546e-01 -1.66920379e-01 7.07168698e-01 -1.43911719e+00
5.99956326e-02 -1.44291148e-01 -2.34193787e-01 -8.35867226e-01
2.48813003e-01 -3.89182895e-01 -7.99804181e-02 -1.84710634e+00
3.33063751e-01 -5.79118788e-01 -6.65053785e-01 4.92289513e-01
-4.19205457e-01 -8.56735259e-02 -4.16442230e-02 -1.01549625e-01
-5.39435387e-01 9.19105291e-01 6.76947474e-01 1.15358874e-01
-2.95309573e-01 5.35627529e-02 -8.43835235e-01 8.81128311e-01
9.73215938e-01 -7.73905814e-01 -8.15520287e-01 -3.42722654e-01
-1.25161096e-01 -5.54650962e-01 -1.72485322e-01 -1.00951219e+00
5.39426744e-01 1.37326764e-02 2.97569364e-01 -1.01788414e+00
5.48317909e-01 -6.97715640e-01 -3.85686785e-01 1.52372997e-02
-4.12355721e-01 -1.59655258e-01 6.46652341e-01 7.59956300e-01
-3.72739345e-01 -5.56403816e-01 2.18777403e-01 -2.03570575e-01
-8.04424703e-01 2.49385163e-01 -6.02786362e-01 -2.44597092e-01
1.02279794e+00 -1.03581771e-01 -2.52688468e-01 -2.65314579e-01
-6.46211267e-01 2.47990251e-01 9.69137549e-02 7.03881443e-01
4.15236950e-01 -1.15568233e+00 -1.26818478e-01 7.67890289e-02
1.03766702e-01 -4.27016392e-02 2.57713318e-01 1.70827433e-01
1.53647169e-01 1.15992630e+00 -7.81080946e-02 -2.42157623e-01
-1.14477181e+00 5.50386012e-01 -3.24058354e-01 -5.71278453e-01
-4.77802157e-01 9.51910853e-01 3.91314059e-01 -7.11219132e-01
5.87329030e-01 -3.45934004e-01 -4.72307146e-01 1.52514040e-01
4.44757670e-01 1.59150511e-01 -5.43241054e-02 -4.30649072e-01
1.16922624e-01 4.91456181e-01 -6.11905277e-01 -2.62552291e-01
1.58665287e+00 -2.55009197e-02 2.78586864e-01 5.11138797e-01
9.18975472e-01 -3.96986485e-01 -1.09542334e+00 -7.42581010e-01
1.72170371e-01 -1.96904182e-01 6.56378150e-01 -6.21327281e-01
-8.56325805e-01 1.22796202e+00 5.26888847e-01 9.35022160e-02
1.00500417e+00 -2.23954692e-02 9.26158190e-01 5.78885913e-01
2.53933311e-01 -1.34366107e+00 2.57757474e-02 5.51656008e-01
4.01926577e-01 -9.79754925e-01 5.29328659e-02 -2.80601710e-01
-7.64923871e-01 8.62379014e-01 9.33098078e-01 -3.95026326e-01
1.16969085e+00 3.65779214e-02 -2.42759101e-02 -3.05412203e-01
-1.14585197e+00 8.49357322e-02 3.50686967e-01 5.87349653e-01
4.67447191e-01 5.88869676e-02 -3.51909071e-01 1.01442242e+00
-2.19247952e-01 3.16501200e-01 2.98201650e-01 1.07846332e+00
-8.54602635e-01 -1.42805660e+00 -1.21252477e-01 6.09267831e-01
-4.55485255e-01 -4.03455436e-01 2.44518556e-03 9.77072597e-01
2.79001091e-02 6.24413967e-01 1.45333886e-01 -3.18310350e-01
-1.76145524e-01 7.67597914e-01 -8.19679275e-02 -1.10561025e+00
-4.37786758e-01 3.36809248e-01 1.72762461e-02 -1.00736722e-01
-5.05679846e-01 -5.51867485e-01 -9.59409952e-01 2.79232979e-01
-1.01058745e+00 4.51424897e-01 1.05499148e+00 1.03435588e+00
5.36105275e-01 5.69895029e-01 -6.59470856e-02 -6.64739192e-01
-3.02195460e-01 -1.57496333e+00 -5.05350113e-01 -2.08270729e-01
-3.29763740e-01 -8.93280685e-01 -3.03463452e-02 1.95852309e-01] | [10.46237564086914, 6.939294815063477] |
86341710-795d-477a-942c-942ec53d7fb2 | global-local-transformer-for-brain-age | 2109.01663 | null | https://arxiv.org/abs/2109.01663v1 | https://arxiv.org/pdf/2109.01663v1.pdf | Global-Local Transformer for Brain Age Estimation | Deep learning can provide rapid brain age estimation based on brain magnetic resonance imaging (MRI). However, most studies use one neural network to extract the global information from the whole input image, ignoring the local fine-grained details. In this paper, we propose a global-local transformer, which consists of a global-pathway to extract the global-context information from the whole input image and a local-pathway to extract the local fine-grained details from local patches. The fine-grained information from the local patches are fused with the global-context information by the attention mechanism, inspired by the transformer, to estimate the brain age. We evaluate the proposed method on 8 public datasets with 8,379 healthy brain MRIs with the age range of 0-97 years. 6 datasets are used for cross-validation and 2 datasets are used for evaluating the generality. Comparing with other state-of-the-art methods, the proposed global-local transformer reduces the mean absolute error of the estimated ages to 2.70 years and increases the correlation coefficient of the estimated age and the chronological age to 0.9853. In addition, our proposed method provides regional information of which local patches are most informative for brain age estimation. Our source code is available on: \url{https://github.com/shengfly/global-local-transformer}. | ['Yangming Ou', 'P. Ellen Grant', 'Sheng He'] | 2021-09-03 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-3.16067457e-01 1.89605784e-02 8.58342573e-02 -6.35095119e-01
-6.22046947e-01 1.08091183e-01 3.60439599e-01 3.08187723e-01
-8.33282053e-01 7.74850845e-01 3.81819934e-01 1.91504896e-01
-1.48096427e-01 -8.07717621e-01 -5.30462086e-01 -9.33139980e-01
-3.26282918e-01 2.63444066e-01 2.15589851e-01 1.93900928e-01
3.26971114e-01 1.87068522e-01 -1.36818683e+00 -4.44524027e-02
1.28292418e+00 1.31606865e+00 3.04592520e-01 2.04718724e-01
7.97732696e-02 5.04219651e-01 -4.61447835e-01 -2.79905021e-01
-6.36169836e-02 -1.29482642e-01 -6.66939318e-01 -4.20744628e-01
3.45082074e-01 -5.13636470e-01 -2.85887122e-01 1.15324628e+00
7.51549006e-01 -9.83150154e-02 8.61081064e-01 -1.09741378e+00
-6.63478673e-01 5.32546282e-01 -8.07824135e-01 6.71145082e-01
-1.55516714e-01 -6.69005290e-02 4.93586630e-01 -7.67108977e-01
2.73785204e-01 9.37262177e-01 6.79943860e-01 5.60997903e-01
-5.38859189e-01 -1.15225410e+00 2.06469193e-01 7.24604905e-01
-1.38617253e+00 -3.80180895e-01 7.17938662e-01 -7.22656548e-01
5.39121985e-01 -2.33427837e-01 8.14916730e-01 8.46377790e-01
7.66795874e-01 3.24393213e-01 1.61463690e+00 1.82649307e-02
4.99084629e-02 -4.70474184e-01 4.21921492e-01 9.40496266e-01
2.80884176e-01 2.04207420e-01 -4.76303101e-01 -3.67162116e-02
8.70678365e-01 1.45553783e-01 -8.41586739e-02 2.13850170e-01
-1.23522425e+00 5.56669056e-01 8.51512194e-01 4.31747913e-01
-7.28098452e-01 -1.13255223e-02 3.13202620e-01 5.90376668e-02
7.67531991e-01 -1.22393981e-01 -5.02121270e-01 1.73645422e-01
-1.02164733e+00 1.67302251e-01 1.74531788e-01 3.84033948e-01
6.31034315e-01 -7.07861036e-02 -1.33134812e-01 9.05084908e-01
2.65943319e-01 3.88147175e-01 7.91956782e-01 -8.65308523e-01
1.77140966e-01 5.38866997e-01 -4.37647015e-01 -7.01765180e-01
-6.42855406e-01 -5.80468416e-01 -9.77931321e-01 2.13735968e-01
6.43266976e-01 -2.39039376e-01 -1.02037561e+00 2.02398658e+00
3.22566122e-01 -8.28213543e-02 -4.79584008e-01 8.66123736e-01
9.38957691e-01 3.36783141e-01 4.23857659e-01 -2.49957323e-01
1.79449487e+00 -8.63743365e-01 -4.10277396e-01 -4.02282447e-01
1.34299085e-01 -3.47681552e-01 7.18095541e-01 2.95195162e-01
-1.06685853e+00 -5.69971144e-01 -8.86102796e-01 -8.45839735e-03
-3.90000582e-01 2.95061022e-01 4.81362402e-01 3.95165205e-01
-1.12628615e+00 4.30398852e-01 -9.98535395e-01 -2.55093843e-01
6.88084424e-01 3.28851789e-01 -6.81646585e-01 -7.00675771e-02
-1.17228270e+00 7.34539390e-01 4.65738416e-01 5.69322892e-02
-8.61856520e-01 -9.67395961e-01 -5.26698351e-01 6.45253733e-02
-8.26900154e-02 -8.60040307e-01 8.08790565e-01 -8.06873024e-01
-9.22135413e-01 9.24363077e-01 -1.17136970e-01 -2.86543846e-01
3.94326478e-01 -2.51665980e-01 -1.13926448e-01 4.45937783e-01
3.78107935e-01 8.93707156e-01 7.78670609e-01 -6.17470324e-01
-3.42218071e-01 -9.62217689e-01 -3.19882661e-01 8.36921036e-02
-4.74063367e-01 3.61954719e-01 -1.95950359e-01 -9.36554193e-01
2.12262362e-01 -5.48061848e-01 3.68749574e-02 5.87504394e-02
4.00661817e-03 -3.64331514e-01 3.44873935e-01 -1.59006894e+00
1.19567764e+00 -1.89249945e+00 4.66501527e-02 -8.21766853e-02
6.37952566e-01 -3.09962809e-01 1.44594163e-01 -2.76747327e-02
-2.50808954e-01 -1.46884277e-01 -2.92227775e-01 -1.83913991e-01
-2.95417547e-01 -3.70679796e-01 4.18100923e-01 5.38514912e-01
-7.92927444e-02 6.97323382e-01 -8.13029885e-01 -8.10265481e-01
-1.77506492e-01 6.53992176e-01 -2.08886638e-01 2.72835314e-01
5.53744495e-01 6.65246665e-01 -4.54930216e-01 6.94217443e-01
6.69620991e-01 4.95648617e-03 -3.58880579e-01 -2.11778238e-01
-2.60656849e-02 -9.29787904e-02 -4.76602077e-01 1.61932480e+00
-2.38349512e-01 1.96057156e-01 4.86180931e-02 -9.91543472e-01
1.16834867e+00 2.17421010e-01 5.72510064e-01 -9.11708653e-01
4.06060725e-01 2.90717185e-01 1.51028395e-01 -3.01955044e-01
-2.31020674e-01 -1.12258993e-01 -4.74958494e-02 5.12067497e-01
2.03084737e-01 4.49393183e-01 2.18600094e-01 1.75033107e-01
1.05224407e+00 -1.09027036e-01 3.15788329e-01 -5.76935768e-01
7.08986819e-01 -7.20500171e-01 9.07614172e-01 3.29210341e-01
-6.76551521e-01 4.50473607e-01 4.70392913e-01 -5.85970879e-01
-1.04884243e+00 -1.20970690e+00 -1.61906734e-01 9.38261509e-01
-2.42915466e-01 -3.85025203e-01 -1.26087737e+00 -8.33410859e-01
-2.33376876e-01 2.33961001e-01 -9.35152531e-01 -2.60030985e-01
-3.63184035e-01 -9.68594670e-01 3.67144078e-01 9.05413985e-01
9.65869963e-01 -1.18562984e+00 -5.18300354e-01 -9.91422012e-02
-4.76567388e-01 -8.48209739e-01 -7.19388306e-01 -2.54681915e-01
-1.05892241e+00 -9.15178537e-01 -1.16375065e+00 -5.65914273e-01
9.83386934e-01 -2.87169218e-01 9.10518467e-01 2.29051203e-01
-2.97398537e-01 6.72175363e-02 -1.80158570e-01 -3.49019378e-01
1.61247939e-01 3.06674391e-01 1.38591811e-01 -5.40131181e-02
2.80515283e-01 -9.86938715e-01 -1.02228785e+00 1.04892455e-01
-4.07679915e-01 1.83303609e-01 8.88034761e-01 7.52285004e-01
4.73785669e-01 -6.15723096e-02 8.88434827e-01 -4.42786396e-01
3.70138735e-01 -5.69260538e-01 -2.94617176e-01 9.20547917e-02
-7.84603238e-01 8.07100907e-02 3.56398165e-01 -4.86731857e-01
-1.05672634e+00 -2.22740635e-01 -2.11150050e-01 -1.96634978e-02
-2.38361448e-01 4.21883374e-01 -2.87427068e-01 9.58384350e-02
2.92780280e-01 2.02918187e-01 -5.40546328e-02 -7.24795938e-01
-2.71930724e-01 5.53268671e-01 7.06604660e-01 -7.98748314e-01
2.95897633e-01 2.12095827e-01 -1.40454054e-01 -4.82215583e-01
-6.82155311e-01 -3.78832445e-02 -7.99545825e-01 -3.93477798e-01
9.91241157e-01 -9.81463313e-01 -3.53357077e-01 1.24256933e+00
-8.93240511e-01 -4.09878284e-01 2.43346021e-01 6.28207207e-01
-3.81991029e-01 3.13594460e-01 -9.01382208e-01 -5.48317075e-01
-1.06886089e+00 -1.10049772e+00 9.56363022e-01 4.20912027e-01
-3.31411995e-02 -6.90506220e-01 -1.25839472e-01 5.39900661e-01
3.39759529e-01 3.09674740e-01 1.05059254e+00 -3.31793696e-01
-8.91453624e-02 -1.73440591e-01 -5.65407872e-01 2.78736562e-01
-1.64487828e-02 -3.23163241e-01 -7.54173219e-01 -3.77346337e-01
-8.34234208e-02 1.24573186e-02 9.21041906e-01 7.61067927e-01
1.42651367e+00 -1.87213510e-01 -1.06470324e-01 5.14299035e-01
1.24501646e+00 1.84091717e-01 7.69130409e-01 3.60367537e-01
6.42896414e-01 5.81350982e-01 3.83977532e-01 4.42639172e-01
7.57914722e-01 2.51865983e-01 1.64123759e-01 1.12629578e-01
-2.72951603e-01 5.08233011e-02 3.93380314e-01 1.02138245e+00
-5.69163680e-01 5.96412718e-01 -1.19623971e+00 7.84638464e-01
-1.54370010e+00 -7.67420948e-01 1.10742114e-01 2.20700264e+00
9.43511188e-01 2.30654404e-01 2.88266748e-01 5.17531745e-02
1.01646054e+00 1.72106773e-01 -5.86067498e-01 -1.75643027e-01
1.23211235e-01 2.59616196e-01 3.88431758e-01 2.82281309e-01
-9.60541308e-01 5.00348628e-01 5.57538462e+00 8.00752044e-01
-1.05420470e+00 5.21141052e-01 1.05171502e+00 -1.01993613e-01
6.34724125e-02 -2.07929641e-01 -4.99179780e-01 8.51649463e-01
9.28847253e-01 -3.24324429e-01 3.64936203e-01 6.87047899e-01
1.24730229e-01 -3.41243833e-01 -6.64720833e-01 9.31881726e-01
1.28139287e-01 -6.71154678e-01 -2.46789783e-01 1.18756749e-01
4.77885932e-01 1.68396700e-02 2.75589451e-02 9.30002630e-02
-9.08987820e-02 -9.58916366e-01 7.88567007e-01 1.11425924e+00
9.31331098e-01 -8.77270043e-01 8.75210464e-01 3.08383703e-01
-1.29281306e+00 -2.75340170e-01 -1.75784037e-01 -1.89378157e-01
-1.28160924e-01 1.09186888e+00 -2.23172382e-01 4.46753412e-01
1.29141998e+00 6.00113809e-01 -9.80436623e-01 1.08609807e+00
-4.17213887e-01 4.52273667e-01 -9.36740339e-02 4.11640167e-01
-1.87855229e-01 -2.42625937e-01 2.26650313e-01 8.30405593e-01
5.61337948e-01 1.96872264e-01 -2.68793553e-01 7.35527098e-01
-3.50040197e-02 3.03579032e-01 -1.31426096e-01 1.59095868e-01
5.08566201e-01 1.40710211e+00 -8.99733245e-01 -3.93880367e-01
-2.54755765e-01 6.05833769e-01 5.86259007e-01 1.40451744e-01
-6.69040561e-01 -6.47002518e-01 3.32385540e-01 2.40981460e-01
3.51747759e-02 -3.83904368e-01 -4.27539110e-01 -1.02795756e+00
9.33203027e-02 -6.28313482e-01 5.10521889e-01 -9.97743666e-01
-1.40613043e+00 7.41322815e-01 1.63558662e-01 -4.78694618e-01
-2.54933946e-02 -3.72419268e-01 -8.85941684e-01 9.91142511e-01
-1.09621036e+00 -1.22754800e+00 -6.44925117e-01 5.68524539e-01
3.45312566e-01 -1.22845404e-01 5.46363592e-01 4.82059568e-01
-6.85246468e-01 5.61660945e-01 -2.66595274e-01 4.02923942e-01
8.73310268e-01 -1.21583009e+00 9.31650177e-02 9.03102040e-01
-5.84869742e-01 6.43874824e-01 2.84418195e-01 -8.27769995e-01
-5.08697927e-01 -7.31660664e-01 1.14955699e+00 -2.23170161e-01
6.71773672e-01 -2.44812712e-01 -9.41898346e-01 5.10114133e-01
7.27530196e-02 1.39347073e-02 3.30606639e-01 1.07805245e-01
-4.97172385e-01 -4.01612729e-01 -1.30000210e+00 2.48472601e-01
1.06538022e+00 -4.39769804e-01 -6.57618582e-01 -1.15467928e-01
5.90026259e-01 2.20377203e-02 -1.44968903e+00 4.78056550e-01
9.42304850e-01 -9.25864458e-01 9.60287631e-01 4.17233631e-02
5.81375718e-01 -4.11543921e-02 1.71570882e-01 -1.24484205e+00
-5.35758197e-01 1.16964214e-01 -1.91413492e-01 1.33972275e+00
4.47207224e-03 -8.66623044e-01 5.86353898e-01 4.94943410e-01
-2.53775150e-01 -1.09155059e+00 -9.83176112e-01 -5.98395228e-01
5.54709852e-01 -1.34453624e-02 8.95198882e-01 6.75847173e-01
-3.34804773e-01 5.45087345e-02 -5.48690595e-02 3.16758342e-02
9.42144990e-01 -1.69184640e-01 -5.75694628e-03 -1.49119520e+00
1.47137344e-01 -5.02667546e-01 -5.62106729e-01 -1.62118897e-01
2.58274168e-01 -7.35530257e-01 -2.57411808e-01 -1.69043446e+00
7.15876043e-01 -3.18720251e-01 -8.30861092e-01 6.40881658e-01
-3.88664752e-01 2.61740029e-01 -2.05176342e-02 -1.39930090e-02
-2.62352169e-01 5.24585545e-01 1.22050941e+00 -1.94928661e-01
1.68978214e-01 -2.98028648e-01 -7.70310998e-01 6.65261447e-01
1.26108634e+00 -5.36500335e-01 -2.62195855e-01 -3.56246173e-01
-9.33399275e-02 -1.44158244e-01 4.05776113e-01 -1.47703934e+00
9.12491977e-02 -3.53011787e-02 9.21865225e-01 -4.65931863e-01
1.03187904e-01 -2.78277457e-01 6.74271360e-02 6.67124569e-01
-4.17028874e-04 2.76843905e-01 -1.65422052e-01 1.69696495e-01
-7.79066905e-02 -8.13298970e-02 8.01878452e-01 -2.19784170e-01
-5.33257067e-01 8.82634044e-01 -2.01562479e-01 -4.88169454e-02
8.42311859e-01 -3.22150216e-02 -5.52518785e-01 -4.03523445e-02
-8.92655671e-01 7.91115686e-02 4.37453032e-01 2.72588372e-01
5.98773122e-01 -1.46532536e+00 -8.99476051e-01 1.89202905e-01
5.75414710e-02 -2.89442509e-01 6.91798508e-01 1.34819245e+00
-2.90895343e-01 1.15149960e-01 -1.03317177e+00 -3.51014107e-01
-1.31429517e+00 5.88932216e-01 3.79790932e-01 -3.25324595e-01
-5.06082118e-01 8.62671614e-01 6.15122676e-01 -2.23723948e-01
-3.23878578e-03 -1.37303755e-01 -4.13018823e-01 2.65657187e-01
8.08088720e-01 4.93491888e-01 1.88125414e-03 -7.66167283e-01
-5.55799901e-01 7.97642052e-01 -2.44090289e-01 -1.70565516e-01
1.54276323e+00 -1.39694899e-01 -6.37263715e-01 3.81559938e-01
1.12555075e+00 -2.31745526e-01 -1.19852293e+00 -2.88979918e-01
-2.16751218e-01 -1.83786988e-01 2.85427034e-01 -8.03733647e-01
-1.63678801e+00 1.08982766e+00 1.00533640e+00 -4.21757281e-01
1.54557610e+00 1.61993459e-01 9.48434830e-01 -2.85846263e-01
5.13733089e-01 -1.04618180e+00 7.22417235e-02 3.94978315e-01
8.63533676e-01 -9.78673100e-01 1.22999907e-01 -1.06543293e-02
-3.07956457e-01 9.63577449e-01 8.80468667e-01 -1.88687012e-01
8.29308152e-01 -2.31715273e-02 -1.61005989e-01 -2.25437835e-01
-4.82871264e-01 2.43436452e-02 3.47377360e-01 6.00876689e-01
3.68492693e-01 1.94646522e-01 -8.29129696e-01 1.29124331e+00
-4.12737936e-01 1.20097697e-01 1.88533604e-01 5.86151659e-01
-4.95452017e-01 -1.07660806e+00 -4.42178190e-01 8.83760154e-01
-6.81391060e-01 -1.12148866e-01 -1.35561228e-01 3.95846546e-01
5.18194258e-01 5.92905581e-01 2.53892720e-01 -2.71514684e-01
-2.28206560e-01 2.62487948e-01 7.09585309e-01 -1.00816518e-01
-3.79560947e-01 -1.47725910e-01 -7.14838877e-02 -4.61275697e-01
-2.54721373e-01 -8.97365153e-01 -1.30510676e+00 -4.43004578e-01
1.24804787e-01 -7.31888169e-04 6.28567517e-01 8.52132678e-01
2.56306291e-01 6.40787065e-01 3.85262132e-01 -7.95371532e-01
5.28720133e-02 -1.20192301e+00 -6.87525332e-01 1.82563618e-01
1.54057369e-01 -9.86459076e-01 -1.61828578e-01 4.94237281e-02] | [14.093441009521484, -1.5288227796554565] |
8c6281b0-50ba-43f6-9346-155c95326959 | understanding-how-people-rate-their | 2206.00167 | null | https://arxiv.org/abs/2206.00167v1 | https://arxiv.org/pdf/2206.00167v1.pdf | Understanding How People Rate Their Conversations | User ratings play a significant role in spoken dialogue systems. Typically, such ratings tend to be averaged across all users and then utilized as feedback to improve the system or personalize its behavior. While this method can be useful to understand broad, general issues with the system and its behavior, it does not take into account differences between users that affect their ratings. In this work, we conduct a study to better understand how people rate their interactions with conversational agents. One macro-level characteristic that has been shown to correlate with how people perceive their inter-personal communication is personality. We specifically focus on agreeableness and extraversion as variables that may explain variation in ratings and therefore provide a more meaningful signal for training or personalization. In order to elicit those personality traits during an interaction with a conversational agent, we designed and validated a fictional story, grounded in prior work in psychology. We then implemented the story into an experimental conversational agent that allowed users to opt-in to hearing the story. Our results suggest that for human-conversational agent interactions, extraversion may play a role in user ratings, but more data is needed to determine if the relationship is significant. Agreeableness, on the other hand, plays a statistically significant role in conversation ratings: users who are more agreeable are more likely to provide a higher rating for their interaction. In addition, we found that users who opted to hear the story were, in general, more likely to rate their conversational experience higher than those who did not. | ['Dilek Hakkani-Tur', 'Julia Hirschberg', 'Pankaj Rajan', 'Nicole Chartier', 'Alexandros Papangelis'] | 2022-06-01 | null | null | null | null | ['spoken-dialogue-systems'] | ['speech'] | [-2.57745147e-01 4.22969818e-01 -4.07831520e-02 -9.12452340e-01
-3.87013517e-02 -5.11983693e-01 5.99130630e-01 2.63971299e-01
-5.32706618e-01 4.89231765e-01 6.97250009e-01 -1.79667056e-01
6.89339787e-02 -5.20765364e-01 2.01816902e-01 -3.40291739e-01
1.94215178e-01 5.16747713e-01 -3.35434347e-01 -4.70891893e-01
4.68489647e-01 2.75327891e-01 -1.33455718e+00 3.18004698e-01
8.40358675e-01 1.02725632e-01 2.41399586e-01 5.88594735e-01
1.53591976e-01 6.73009694e-01 -1.18686008e+00 -4.36276913e-01
-2.71840990e-01 -6.35224998e-01 -8.85742188e-01 4.29156274e-01
2.18271479e-01 -6.65882945e-01 1.76969618e-01 6.22712314e-01
3.66547227e-01 6.61490500e-01 8.62328768e-01 -1.15993631e+00
-4.69607919e-01 8.30410182e-01 -1.56848922e-01 3.26388218e-02
8.16564381e-01 2.78757751e-01 1.04964149e+00 -3.91628444e-01
3.11792612e-01 1.47619772e+00 5.06889105e-01 8.74925733e-01
-1.35985756e+00 -6.15292430e-01 3.68502736e-02 -1.93855554e-01
-7.37663865e-01 -6.85975611e-01 4.92941946e-01 -5.80425918e-01
7.19531596e-01 3.50112051e-01 9.55425858e-01 1.03863859e+00
-7.03113675e-02 3.41781914e-01 1.42488992e+00 -1.13301069e-01
2.64668524e-01 1.10999525e+00 5.44465125e-01 1.06865376e-01
6.78700730e-02 -5.54875992e-02 -5.45018077e-01 -5.66776395e-01
6.97810650e-01 -3.14664006e-01 -4.71858233e-01 4.82576294e-03
-8.45647156e-01 1.09090948e+00 2.77932346e-01 6.27959311e-01
-6.83663905e-01 -5.08156419e-01 3.85342687e-01 6.78709745e-01
2.82454282e-01 1.31450856e+00 -3.67826194e-01 -1.00423431e+00
-1.96404189e-01 1.99785188e-01 1.43000793e+00 2.32867852e-01
4.17092979e-01 -2.06827879e-01 -2.08089039e-01 1.38093746e+00
7.62886554e-02 -8.20595920e-02 5.20772040e-01 -1.25045955e+00
-8.33339766e-02 6.02324069e-01 4.55756426e-01 -9.47546721e-01
-6.50333762e-01 -5.95763745e-03 -1.71871826e-01 4.49325323e-01
5.43326020e-01 -5.10005832e-01 -1.68050304e-01 1.83667111e+00
-2.34399457e-02 -7.48485267e-01 2.95285344e-01 1.11499918e+00
4.83282477e-01 5.72153211e-01 6.57097772e-02 -6.80101275e-01
1.47002196e+00 -5.78927457e-01 -6.38709247e-01 -3.20395738e-01
6.12920523e-01 -8.65421653e-01 1.69376779e+00 4.04991895e-01
-9.89919066e-01 -6.13374770e-01 -8.83481324e-01 3.66889238e-01
7.22093582e-02 -2.46826068e-01 7.78896093e-01 1.14133811e+00
-1.00457263e+00 5.87970257e-01 -4.53189373e-01 -1.03275871e+00
-4.85044688e-01 2.95571357e-01 -1.70332521e-01 4.65255499e-01
-1.13985717e+00 1.41900730e+00 -1.64120361e-01 -2.77731687e-01
-2.77816683e-01 -2.29759961e-01 -7.19842672e-01 1.18945017e-01
2.23135892e-02 -6.52786195e-01 1.81045818e+00 -1.12783659e+00
-1.96185184e+00 7.02554882e-01 -1.69359311e-01 5.02251610e-02
1.00885108e-01 8.11743736e-02 -2.54473418e-01 -1.96488366e-01
2.20344942e-02 5.20917356e-01 4.31286633e-01 -1.10502958e+00
-4.69914228e-01 -4.65660661e-01 6.27156496e-01 8.19097459e-01
-5.67863882e-01 1.74964488e-01 2.34978363e-01 -3.41958642e-01
-1.81543127e-01 -1.07169509e+00 1.16591386e-01 -5.96525192e-01
1.93314910e-01 -6.74027860e-01 4.33771104e-01 -4.80069339e-01
1.03399324e+00 -1.97932780e+00 5.23952395e-02 1.92603230e-01
2.54484117e-01 1.38798594e-01 1.92070931e-01 8.47406447e-01
2.15006873e-01 3.01349849e-01 4.31278199e-01 -3.60305190e-01
1.49139181e-01 8.81046280e-02 3.99406016e-01 2.12607488e-01
-2.81927288e-01 2.69732296e-01 -7.52924919e-01 -1.20593317e-01
1.06359676e-01 2.57895499e-01 -6.48467600e-01 6.30868375e-01
2.20527023e-01 5.88235676e-01 -2.50371128e-01 -1.42556623e-01
3.11059475e-01 3.73882502e-02 2.19546437e-01 2.47713611e-01
-1.77001983e-01 7.74914503e-01 -5.75883508e-01 8.19713712e-01
-7.24676967e-01 7.74642110e-01 4.21296358e-01 -4.37627435e-01
1.15145719e+00 5.06347120e-01 1.77963078e-01 -4.01949793e-01
3.19670498e-01 -8.80364776e-02 7.42884338e-01 -6.34049594e-01
1.02138710e+00 -4.47706789e-01 1.22383423e-02 9.86583889e-01
-5.45004427e-01 -4.97629583e-01 1.34474710e-01 3.88825655e-01
7.14607894e-01 -6.97424293e-01 3.95854741e-01 -1.25499815e-01
2.27754802e-01 -1.32528380e-01 4.57841307e-01 7.41883099e-01
-3.24455142e-01 2.19641235e-02 7.33799219e-01 1.54600456e-01
-8.98379445e-01 -5.48115015e-01 -2.15876084e-02 1.28083980e+00
-7.60924220e-02 -2.19286785e-01 -8.41594279e-01 -3.31116468e-01
-3.51907969e-01 1.38315654e+00 -3.69142771e-01 -2.97654837e-01
4.72885258e-02 -2.80307144e-01 2.27837428e-01 2.15284318e-01
3.88010293e-01 -1.20537293e+00 -5.34436285e-01 2.37885609e-01
-3.66801918e-01 -6.78027034e-01 -7.26150572e-01 -2.84393221e-01
-6.66985154e-01 -6.76351070e-01 -5.06746590e-01 -4.67015266e-01
5.94078362e-01 5.84062636e-01 8.50354135e-01 2.53220439e-01
3.92960489e-01 8.01153243e-01 -5.98739684e-01 -4.14166898e-01
-8.70576262e-01 8.17806832e-03 3.27439785e-01 -3.44760060e-01
5.57287633e-01 -4.25972670e-01 -3.80230099e-01 7.21901357e-01
-3.53488863e-01 -5.49651906e-02 2.65909165e-01 8.60673249e-01
-7.66428590e-01 3.05734295e-02 7.32413352e-01 -1.12277865e+00
1.65507090e+00 -2.94054151e-01 7.42986724e-02 -3.00260961e-01
-7.93879509e-01 -3.22385132e-01 5.41502476e-01 -5.26754379e-01
-1.24652243e+00 -4.80271667e-01 -1.20215565e-01 4.21351641e-01
-3.61898810e-01 6.64322674e-01 7.28507265e-02 2.14121491e-01
8.74135077e-01 -2.96595544e-01 7.15382516e-01 -9.81462598e-02
-1.65548697e-01 1.40501118e+00 4.55337912e-02 -5.99399626e-01
4.87670213e-01 -2.10199952e-01 -8.79725933e-01 -1.26310694e+00
-3.99132013e-01 -5.22199869e-01 -1.43234998e-01 -4.93685156e-01
7.74231374e-01 -7.13739693e-01 -1.25302279e+00 2.67788410e-01
-7.97074378e-01 -8.12578738e-01 3.46279860e-01 8.68969262e-01
-5.18126428e-01 1.87343642e-01 -7.47799993e-01 -1.24540806e+00
-1.44491658e-01 -1.19105816e+00 3.20525020e-01 5.39355397e-01
-1.53346896e+00 -1.02550280e+00 1.21356174e-02 8.83750975e-01
5.11973500e-01 -4.61088806e-01 9.05019581e-01 -1.00444639e+00
4.42702800e-01 -1.52807981e-01 2.34871879e-01 3.58092904e-01
5.74350715e-01 -3.32079385e-03 -8.54203105e-01 -4.07949001e-01
3.63966823e-01 -4.68583196e-01 6.81848302e-02 2.02444613e-01
3.85897875e-01 -5.38675189e-01 2.41127703e-02 -3.25090945e-01
3.70721281e-01 5.72117448e-01 4.00021791e-01 1.65051341e-01
1.07187137e-01 1.13968778e+00 9.21497285e-01 5.23644090e-01
6.06631994e-01 9.49386120e-01 -2.10570663e-01 1.06376717e-02
4.16560113e-01 -1.38091028e-01 8.96462679e-01 6.58885002e-01
-7.52723292e-02 -2.04381317e-01 -4.09120798e-01 1.06498443e-01
-1.49005818e+00 -9.75159407e-01 -1.34706944e-01 2.39270043e+00
8.47934365e-01 1.51237428e-01 7.23623693e-01 3.01293023e-02
8.01710963e-01 1.47524206e-02 -4.54017758e-01 -1.05497205e+00
4.56380337e-01 -2.77001113e-01 2.96538807e-02 9.30941224e-01
-2.58622319e-01 7.04200566e-01 6.24326420e+00 -3.46048266e-01
-9.81248081e-01 -1.61818340e-01 7.47810066e-01 3.50472145e-02
-2.30366617e-01 1.78717412e-02 -2.98496544e-01 1.11095741e-01
1.04149926e+00 -4.61385608e-01 6.82141244e-01 7.58821368e-01
8.16591978e-01 -4.10072148e-01 -1.37276888e+00 6.55847132e-01
-5.96664511e-02 -4.43028450e-01 -5.96501827e-01 3.53445530e-01
3.04699570e-01 -6.22972906e-01 -2.76821433e-03 5.33068240e-01
4.14435893e-01 -7.97554851e-01 2.32478589e-01 1.48986921e-01
2.36526817e-01 -6.13533020e-01 8.67069423e-01 5.21371841e-01
-2.56798208e-01 1.49316669e-01 -3.88410658e-01 -9.20739174e-01
2.18448594e-01 -5.56247728e-03 -1.74673355e+00 -3.91392261e-01
4.33164030e-01 1.48274183e-01 -1.54603958e-01 5.36311865e-01
-1.19702838e-01 7.10470557e-01 -9.03235078e-02 -6.47551715e-01
-1.57088280e-01 -4.89328742e-01 4.82078940e-01 7.78791606e-01
4.55734469e-02 4.99183238e-01 -2.80402000e-05 6.49295986e-01
3.48036826e-01 3.86780143e-01 -7.12835073e-01 -3.93143088e-01
7.71394491e-01 1.28903723e+00 -6.00761652e-01 -4.63416040e-01
-4.65662658e-01 1.04209125e+00 2.45037392e-01 3.16769242e-01
-4.98345286e-01 -2.45101437e-01 1.00750172e+00 1.41077042e-01
-3.45098078e-01 -1.99622586e-01 -5.03776014e-01 -8.41598749e-01
-2.96312511e-01 -1.21967125e+00 4.26404700e-02 -7.65499771e-01
-1.36462677e+00 4.48673010e-01 -1.12604931e-01 -7.55375445e-01
-6.26223564e-01 -2.50572741e-01 -6.65221989e-01 1.02760363e+00
-2.88848788e-01 -3.08607012e-01 -6.21906102e-01 1.72879636e-01
7.48838425e-01 -8.32237750e-02 9.61880744e-01 -2.15739995e-01
-5.22778511e-01 6.83982611e-01 -3.60063940e-01 -1.36841074e-01
1.22642672e+00 -1.24564564e+00 1.35190859e-01 8.48552659e-02
-3.43743801e-01 1.27397072e+00 1.11260951e+00 -6.12232506e-01
-1.06123102e+00 -1.92963466e-01 6.87550128e-01 -3.16964626e-01
5.02094626e-01 -1.14122696e-01 -8.99598062e-01 6.41344130e-01
5.81574917e-01 -1.42077768e+00 9.74378347e-01 8.71763527e-01
1.94454473e-02 1.58180609e-01 -1.24650490e+00 8.95529211e-01
8.35069895e-01 -5.75845838e-01 -8.57680738e-01 1.07691862e-01
4.34160113e-01 -8.53820667e-02 -1.14116335e+00 -4.72550727e-02
6.33174360e-01 -1.34877276e+00 4.63852435e-01 -1.39847547e-01
2.69716889e-01 2.17549011e-01 3.07873815e-01 -2.06648588e+00
-5.15537560e-01 -7.90610552e-01 7.52568781e-01 1.41174459e+00
7.31274903e-01 -1.17511022e+00 7.01088369e-01 1.48007047e+00
-2.61730075e-01 -3.79612714e-01 -2.58577973e-01 -2.61755973e-01
1.39704689e-01 -7.06297010e-02 3.87568384e-01 1.11863887e+00
1.13068604e+00 9.54229176e-01 -4.17340517e-01 1.67646166e-02
2.46117450e-02 -4.44360405e-01 1.26693559e+00 -1.35016620e+00
-3.94584417e-01 -6.32723808e-01 -8.75150934e-02 -1.19587350e+00
1.24606803e-01 -3.72606456e-01 1.65445015e-01 -1.52696097e+00
1.72948033e-01 -5.96705258e-01 4.43119794e-01 2.15672702e-01
-3.62316966e-01 -2.23232076e-01 2.21036181e-01 2.55341262e-01
6.91235065e-02 4.04158443e-01 9.97365236e-01 3.70758057e-01
-9.41417396e-01 5.61111927e-01 -1.38231134e+00 6.19896591e-01
1.13345301e+00 -3.16669531e-02 -7.64030814e-01 1.85503706e-01
-1.10695906e-01 5.16522288e-01 -1.67530179e-01 -6.36893928e-01
1.23832308e-01 -4.68316108e-01 3.44499946e-02 8.72929394e-02
7.72297800e-01 -6.35425389e-01 7.48845935e-02 2.45510191e-01
-5.54403126e-01 2.14084536e-01 -4.36423905e-03 -6.87357560e-02
-5.94459549e-02 -5.11783957e-01 6.23324096e-01 -2.70899851e-02
-1.92850217e-01 -3.65917176e-01 -1.23630559e+00 -3.71184617e-01
8.38550091e-01 -4.28069621e-01 -2.22359553e-01 -1.39797926e+00
-6.67448461e-01 2.34808460e-01 9.01489854e-01 5.68885326e-01
3.61034572e-01 -8.39727044e-01 -5.37370920e-01 -1.65138215e-01
1.96769178e-01 -7.15051472e-01 6.53395755e-03 8.42842638e-01
-1.19882390e-01 3.66579384e-01 -2.25766465e-01 -2.49422893e-01
-1.62749314e+00 8.97162780e-02 2.32562825e-01 3.45060021e-01
-2.83146054e-01 6.36282980e-01 2.43516758e-01 -3.98081571e-01
2.33789101e-01 -4.85494249e-02 -5.94221950e-01 3.52649808e-01
5.99571586e-01 3.76659483e-01 -2.46046588e-01 -5.97487211e-01
1.42015517e-01 -1.91019885e-02 -5.58473766e-01 -7.02856302e-01
1.01942432e+00 -2.81801462e-01 -1.10311426e-01 1.11693084e+00
7.88606644e-01 3.34709615e-01 -9.05397892e-01 1.11564450e-01
-3.21533352e-01 -8.01698029e-01 -3.26713711e-01 -8.00865769e-01
-4.85326320e-01 3.80376101e-01 1.37724295e-01 8.36757123e-01
5.99151909e-01 5.22380099e-02 4.01490420e-01 4.74529058e-01
4.92155194e-01 -1.15864849e+00 4.01921660e-01 5.05785227e-01
1.01808703e+00 -1.20957386e+00 -1.26599655e-01 -3.88858080e-01
-1.48858821e+00 9.46405768e-01 1.09948456e+00 4.02366668e-01
2.57281512e-01 -2.41860807e-01 5.12351096e-01 -2.26552859e-01
-9.71273899e-01 1.52706653e-01 -1.54665843e-01 8.01693082e-01
1.21602130e+00 4.46488380e-01 -9.62700248e-01 4.42760706e-01
-1.00466931e+00 -2.67885208e-01 1.15401721e+00 4.40868318e-01
-5.38469255e-01 -1.26057577e+00 -4.79752779e-01 6.78764343e-01
7.46055394e-02 1.31087780e-01 -1.08677769e+00 4.10499603e-01
-5.53567886e-01 1.71723580e+00 1.18708685e-01 -7.06730068e-01
5.97479582e-01 1.61603644e-01 1.59742743e-01 -1.06037879e+00
-1.06281757e+00 -1.89791024e-01 8.25110078e-01 -4.96698543e-02
-2.35814989e-01 -1.13112128e+00 -1.09641314e+00 -8.70048404e-01
-5.18579483e-01 7.18813360e-01 4.73333597e-01 7.28193521e-01
7.68969283e-02 1.10449947e-01 8.32872868e-01 -6.38117850e-01
-6.00159168e-01 -1.77441251e+00 -6.29195809e-01 5.87689579e-01
-6.55941665e-02 -5.68774998e-01 -6.00414634e-01 -4.91431117e-01] | [12.82366943359375, 7.893664836883545] |
f4127ed8-5581-4e95-b8ea-f79298a251b0 | discovering-picturesque-highlights-from | 1601.04406 | null | http://arxiv.org/abs/1601.04406v1 | http://arxiv.org/pdf/1601.04406v1.pdf | Discovering Picturesque Highlights from Egocentric Vacation Videos | We present an approach for identifying picturesque highlights from large
amounts of egocentric video data. Given a set of egocentric videos captured
over the course of a vacation, our method analyzes the videos and looks for
images that have good picturesque and artistic properties. We introduce novel
techniques to automatically determine aesthetic features such as composition,
symmetry and color vibrancy in egocentric videos and rank the video frames
based on their photographic qualities to generate highlights. Our approach also
uses contextual information such as GPS, when available, to assess the relative
importance of each geographic location where the vacation videos were shot.
Furthermore, we specifically leverage the properties of egocentric videos to
improve our highlight detection. We demonstrate results on a new egocentric
vacation dataset which includes 26.5 hours of videos taken over a 14 day
vacation that spans many famous tourist destinations and also provide results
from a user-study to access our results. | ['Vinay Bettadapura', 'Daniel Castro', 'Irfan Essa'] | 2016-01-18 | null | null | null | null | ['highlight-detection'] | ['computer-vision'] | [ 7.16997236e-02 -3.47766101e-01 -1.53788805e-01 -2.17722550e-01
-6.55409932e-01 -8.98286462e-01 5.51914394e-01 1.01607807e-01
-1.44471060e-02 2.85256296e-01 8.34966838e-01 2.57382631e-01
-7.10188374e-02 -6.42653048e-01 -7.12056637e-01 -3.57451230e-01
-5.15002549e-01 -2.42438570e-01 3.34460884e-02 -3.02940726e-01
7.93710113e-01 5.45390546e-01 -1.85380566e+00 4.04207289e-01
5.10953426e-01 9.33035314e-01 1.23576216e-01 8.84905994e-01
3.23059469e-01 7.29004323e-01 -5.60071170e-01 -4.15257037e-01
4.39239085e-01 -3.55656296e-01 -4.99197692e-01 5.02499521e-01
9.42952335e-01 -5.20231247e-01 -6.11226678e-01 8.10728431e-01
4.36469615e-01 7.87884116e-01 6.11303091e-01 -1.38919711e+00
-4.50825572e-01 2.43271321e-01 -6.57802701e-01 8.48729789e-01
7.55096674e-01 1.55342683e-01 1.16235626e+00 -8.52489114e-01
1.24914622e+00 9.55079377e-01 3.61672938e-01 -1.02417752e-01
-9.20170307e-01 -5.25781751e-01 1.14543170e-01 3.96124572e-01
-1.39510596e+00 -6.66269064e-01 1.06164920e+00 -6.04505956e-01
6.48323536e-01 1.86848938e-01 9.47591782e-01 9.71620381e-01
-3.36573906e-02 8.91201198e-01 5.94031930e-01 -4.16584536e-02
2.81896472e-01 -9.51423049e-02 -4.86726582e-01 6.71844661e-01
-5.41518033e-02 -3.61444175e-01 -8.36661458e-01 -2.17284337e-01
8.09892416e-01 1.72308564e-01 -3.50300133e-01 -7.61665761e-01
-1.21590078e+00 7.19022512e-01 4.26473588e-01 1.26677558e-01
-6.07179463e-01 1.76859125e-01 3.53129238e-01 1.73143387e-01
3.89071465e-01 8.94129276e-01 4.93034646e-02 -6.57179236e-01
-9.46979403e-01 2.10827708e-01 3.86487037e-01 1.11173916e+00
8.68802249e-01 -4.42750007e-02 -8.47521126e-02 7.36585617e-01
-4.82294649e-01 4.66949761e-01 1.59237742e-01 -1.30733490e+00
4.13482547e-01 5.02496183e-01 3.58717322e-01 -1.72713935e+00
-1.41192272e-01 -5.95240146e-02 8.33190978e-02 -1.70373797e-01
1.78524405e-01 -1.68493524e-01 -4.29895163e-01 1.40531576e+00
9.80541036e-02 1.66379511e-01 -2.31477097e-01 1.15953362e+00
7.82524824e-01 4.45844084e-01 6.18645437e-02 1.09645903e-01
1.38508689e+00 -7.41692722e-01 -4.61351812e-01 5.78995273e-02
7.12328494e-01 -9.05999362e-01 1.29902196e+00 3.87833506e-01
-1.32230866e+00 -4.12410706e-01 -9.97982204e-01 -3.56266797e-02
-3.81786674e-01 2.80921251e-01 4.77342784e-01 5.40703177e-01
-8.21197033e-01 6.40858054e-01 -2.51522571e-01 -6.87698662e-01
2.16683090e-01 -8.80815759e-02 -5.13175964e-01 3.16415727e-02
-8.08803380e-01 3.82234961e-01 6.02585673e-02 -5.42849362e-01
-8.96300673e-01 -6.66116476e-01 -9.31981087e-01 1.87093228e-01
3.57327849e-01 -4.71470088e-01 1.00673616e+00 -1.40827298e+00
-1.10386825e+00 9.57768083e-01 -1.49968922e-01 -1.27076298e-01
1.96377888e-01 -3.18510562e-01 -6.03308856e-01 8.90379369e-01
1.72788352e-01 7.01285601e-01 9.22323227e-01 -1.12839890e+00
-9.83225644e-01 -1.67818904e-01 8.65888536e-01 6.26104057e-01
-5.41045249e-01 9.10381302e-02 -6.89143777e-01 -7.98013926e-01
-1.27583712e-01 -9.22875702e-01 2.16618776e-01 -2.69075483e-01
-2.08133906e-01 3.11425507e-01 9.25589442e-01 -6.35433912e-01
1.19746208e+00 -2.50958967e+00 2.17820704e-02 3.43966573e-01
1.67383105e-01 -4.52674687e-01 -1.46581009e-01 6.11896157e-01
-5.31108566e-02 1.62926301e-01 3.62087220e-01 -9.06045958e-02
-1.87802061e-01 -2.98871040e-01 -2.43431196e-01 4.40031081e-01
-4.11988944e-02 6.64696455e-01 -1.33498204e+00 -5.49844563e-01
3.19724858e-01 1.39904499e-01 -7.79695511e-01 6.60529733e-02
1.74661633e-02 8.55757967e-02 -3.89604509e-01 8.27207506e-01
4.97565478e-01 -1.76946744e-02 3.53490040e-02 -3.14410806e-01
-1.22573301e-01 6.43010363e-02 -9.06084120e-01 1.79850340e+00
-5.11891901e-01 1.23654521e+00 -2.69409925e-01 -3.04832816e-01
7.40490317e-01 3.76856104e-02 6.95010006e-01 -7.25149691e-01
1.70471892e-01 -1.72604680e-01 -6.76547766e-01 -8.58838797e-01
1.33011734e+00 3.30550551e-01 -3.78321260e-01 5.87957978e-01
-2.78925627e-01 -3.58111225e-02 5.32258332e-01 6.53108239e-01
1.16675794e+00 1.95535481e-01 1.90149367e-01 -1.63332641e-01
2.46768907e-01 4.09754038e-01 2.14607909e-01 5.91434836e-01
-3.49148750e-01 1.02411890e+00 5.59090555e-01 -4.51054394e-01
-1.26229239e+00 -1.06385565e+00 3.94442856e-01 1.28037643e+00
5.85694015e-01 -8.47451806e-01 -5.49984217e-01 -3.62848520e-01
-1.50704086e-01 5.82454741e-01 -7.75377154e-01 -3.96669693e-02
-4.17873830e-01 -1.89469293e-01 5.04482388e-02 4.76055056e-01
4.22233671e-01 -9.65098619e-01 -7.51281083e-01 -2.85870254e-01
-5.08715928e-01 -1.14113986e+00 -9.53486145e-01 -6.14969671e-01
-4.93894070e-01 -1.31097412e+00 -7.21961379e-01 -5.91405272e-01
8.36372614e-01 1.13435149e+00 1.19405699e+00 -4.00288068e-02
-3.61753225e-01 8.86503458e-01 -5.93645155e-01 8.55056718e-02
1.80073440e-01 -1.08053863e-01 9.78028774e-02 2.42522612e-01
4.32050645e-01 -4.72028673e-01 -1.06459498e+00 4.60149854e-01
-6.89246476e-01 -1.69690643e-02 1.56912386e-01 3.55614811e-01
3.87042820e-01 1.20129153e-01 -5.90532757e-02 -5.87534130e-01
4.87518847e-01 -9.02981818e-01 -1.24397933e-01 -8.16556811e-02
9.25423726e-02 -5.15447497e-01 4.96185541e-01 -1.69763029e-01
-9.36039269e-01 -4.00384068e-02 7.46752203e-01 -7.25237429e-01
-1.17235323e-02 3.13347399e-01 1.21866286e-01 8.97058398e-02
6.81305110e-01 -1.00686677e-01 -4.01054293e-01 -1.12684770e-02
3.49746257e-01 4.77967709e-01 6.35752141e-01 -2.84011543e-01
6.94301426e-01 9.75600123e-01 -2.12287992e-01 -1.08574343e+00
-3.27996254e-01 -8.59097004e-01 -4.85698074e-01 -7.12022603e-01
7.49408543e-01 -1.12956035e+00 -8.02996099e-01 5.31966463e-02
-6.32523000e-01 -1.20966099e-01 -2.02123895e-01 5.26412606e-01
-7.99895763e-01 6.13551617e-01 -1.44049898e-01 -6.66620374e-01
-5.10305278e-02 -7.48127460e-01 1.01494205e+00 3.91290724e-01
-4.94433910e-01 -8.08171928e-01 -2.00361192e-01 3.22752059e-01
1.09432429e-01 4.73412514e-01 5.33880174e-01 -4.39272299e-02
-7.89698958e-01 -2.43239269e-01 -3.52143824e-01 -3.81572098e-01
3.50371748e-01 3.82526368e-01 -8.37394774e-01 -3.11061859e-01
-5.75910151e-01 -9.50949080e-03 5.56118906e-01 6.03670835e-01
1.00224876e+00 -3.07709336e-01 -2.62579858e-01 5.59924781e-01
1.30142260e+00 7.29322359e-02 7.83461392e-01 6.88887894e-01
4.78499174e-01 7.45758057e-01 1.13905251e+00 8.03399205e-01
2.39606261e-01 7.12454021e-01 5.06711543e-01 3.27139013e-02
9.66712683e-02 -5.84330261e-01 5.00491619e-01 1.51827559e-01
-2.04337209e-01 -3.66344810e-01 -4.41115171e-01 1.09636581e+00
-1.64035988e+00 -1.53001618e+00 8.97178873e-02 2.23936200e+00
-1.45713046e-01 8.29382688e-02 6.19734287e-01 -2.68801898e-01
8.79467070e-01 4.60052729e-01 -3.47942352e-01 -5.18719554e-01
-1.42962009e-01 -4.65240657e-01 6.83740377e-01 9.34899449e-02
-1.12936521e+00 8.63370240e-01 6.78618908e+00 6.05826735e-01
-8.97143841e-01 -1.96315855e-01 6.11932039e-01 -7.34073758e-01
-5.46057820e-01 -9.49586853e-02 -1.45366296e-01 5.60346842e-01
5.64008653e-01 -3.04677129e-01 4.88917440e-01 1.24487042e+00
7.65120387e-01 -5.88641822e-01 -7.49611378e-01 1.28403580e+00
5.13228297e-01 -1.63227022e+00 9.86097530e-02 4.65846173e-02
9.78580534e-01 -2.75386781e-01 4.25290763e-01 -4.96438763e-04
7.02470317e-02 -6.17144406e-01 7.19412684e-01 3.50167543e-01
6.81443572e-01 -1.08041310e+00 2.06600636e-01 -3.60420883e-01
-1.34100938e+00 -2.85644561e-01 -2.65880406e-01 -1.09322906e-01
3.32169950e-01 7.51403198e-02 -8.69469941e-01 3.22157532e-01
1.08200431e+00 1.15719593e+00 -7.65869737e-01 1.19432008e+00
-1.14162475e-01 1.20733611e-01 -1.35588124e-01 3.64940353e-02
3.32635164e-01 -4.04020369e-01 8.53626728e-01 1.20429921e+00
6.19521499e-01 2.08155394e-01 -2.84348309e-01 5.49285710e-01
-2.85293162e-02 4.26081598e-01 -1.05366349e+00 -2.53286004e-01
4.62836236e-01 1.43501103e+00 -9.65929449e-01 -3.92875910e-01
-2.13266477e-01 1.04248261e+00 -6.18155636e-02 4.27878112e-01
-7.40958035e-01 -4.86603528e-01 9.47933793e-01 3.52341652e-01
4.91353840e-01 -4.87818152e-01 9.68652964e-02 -1.07235074e+00
1.43118137e-02 -5.05945623e-01 5.30154705e-01 -1.52816558e+00
-7.41952300e-01 4.86054212e-01 5.37778586e-02 -1.67668033e+00
-3.01320881e-01 -1.47892177e-01 -9.34127450e-01 2.09289059e-01
-1.10082436e+00 -8.16986084e-01 -9.51938987e-01 4.96695071e-01
8.40973377e-01 -2.45596003e-02 2.97854543e-01 2.72295594e-01
-2.81147361e-01 3.74702126e-01 2.54596829e-01 2.90997303e-03
8.67237568e-01 -1.18685961e+00 5.66708505e-01 8.78113866e-01
7.95693994e-02 5.73807478e-01 9.29422855e-01 -6.64455354e-01
-1.46499276e+00 -7.53088653e-01 6.17260456e-01 -5.86417317e-01
7.24519789e-01 -1.10764429e-01 -1.08249478e-01 5.93202353e-01
5.95206916e-02 -6.16901338e-01 8.09686899e-01 9.60478336e-02
-6.80039078e-02 1.30085245e-01 -9.17854071e-01 1.14682436e+00
1.42716205e+00 -5.03312707e-01 -4.27093506e-01 3.04392576e-01
1.84323058e-01 -3.37601990e-01 -6.79183245e-01 -1.94163904e-01
8.05361211e-01 -1.21449423e+00 9.79860544e-01 -3.66055101e-01
7.52764940e-01 -2.99684495e-01 -2.02057600e-01 -1.44294047e+00
-3.14407438e-01 -9.72622037e-01 3.83790135e-01 9.99620557e-01
3.61047499e-02 -6.70554936e-02 9.35732007e-01 6.87057853e-01
-9.00578052e-02 -1.16232350e-01 -4.80837345e-01 -6.09042346e-01
-9.67366874e-01 -8.46103802e-02 5.73114812e-01 7.62763262e-01
2.71943539e-01 -1.45820171e-01 -5.96146882e-01 -8.61200616e-02
3.53109330e-01 4.17698771e-01 1.42344618e+00 -8.49330544e-01
1.01665437e-01 -4.52123910e-01 -7.11817861e-01 -8.43862057e-01
-2.59496212e-01 -4.88597423e-01 -3.51020604e-01 -1.14891517e+00
3.57903361e-01 -1.50759807e-02 -1.12768613e-01 -1.69860154e-01
1.44385114e-01 8.43380332e-01 2.16628939e-01 3.79965574e-01
-8.77794504e-01 2.36181095e-01 1.07156336e+00 1.28127141e-02
-4.41544414e-01 -3.17738950e-01 -6.75789237e-01 6.13947570e-01
6.75394535e-01 -1.13206990e-01 -5.65391481e-01 -1.59717903e-01
4.14638549e-01 -1.33141488e-01 2.43408471e-01 -1.10485959e+00
-1.75010577e-01 -3.10403734e-01 5.25865912e-01 -6.80553198e-01
8.89591515e-01 -5.82030773e-01 2.45227173e-01 9.67954621e-02
-3.28992039e-01 4.57066894e-01 1.07149675e-01 8.50594878e-01
-1.69582397e-01 1.09476876e-02 4.09610122e-01 -1.08993068e-01
-1.26550603e+00 9.81624573e-02 -5.49058080e-01 7.48290569e-02
1.26815951e+00 -6.69250309e-01 -3.19787294e-01 -1.12790740e+00
-4.83509272e-01 1.53834209e-01 1.20208204e+00 7.25849092e-01
6.44278109e-01 -1.33565426e+00 -4.55735147e-01 8.86082873e-02
5.10325670e-01 -8.81009996e-01 6.75922453e-01 7.42943287e-01
-7.68020689e-01 2.22302571e-01 -7.69384742e-01 -4.71770167e-01
-1.46737003e+00 8.00842345e-01 -2.25014701e-01 4.57451612e-01
-8.38597298e-01 4.81396586e-01 5.70430756e-01 5.16350150e-01
-1.40274689e-01 -3.20295878e-02 -5.12809694e-01 5.14252663e-01
5.78830779e-01 6.00807905e-01 -2.66088665e-01 -9.84283805e-01
-2.99018830e-01 6.23933196e-01 1.76326856e-01 -2.15700746e-01
1.10696495e+00 -7.00389564e-01 6.86326087e-01 2.08376586e-01
1.38244796e+00 4.61350262e-01 -1.44329572e+00 -2.77532917e-02
-3.03205341e-01 -1.22823024e+00 -5.65260723e-02 -2.43739203e-01
-1.15398371e+00 4.28193092e-01 3.01978439e-01 9.61830616e-02
1.16005743e+00 -2.38532320e-01 9.73002434e-01 7.40636736e-02
3.80978823e-01 -1.42602897e+00 5.83881438e-01 7.26904273e-02
8.21450531e-01 -1.12813687e+00 1.93112180e-01 -2.71004558e-01
-1.11842740e+00 1.13511670e+00 2.97236711e-01 -4.27062899e-01
2.96193928e-01 -5.33515692e-01 -1.05370142e-01 -5.78045845e-01
-3.68199587e-01 -3.35548729e-01 3.32136661e-01 4.38660204e-01
4.91967574e-02 -4.25923765e-02 -2.39907458e-01 4.36924659e-02
-5.31231284e-01 -3.11183572e-01 1.03477597e+00 9.16684806e-01
-4.32449996e-01 -3.53787988e-01 -3.19568545e-01 2.44916975e-01
-2.45825320e-01 6.00275584e-02 -5.41929662e-01 8.28096628e-01
-3.02758276e-01 9.80600476e-01 4.47043151e-01 -5.86351156e-01
2.23800391e-01 -2.50327080e-01 3.14797133e-01 -3.24369937e-01
-4.35751647e-01 -4.45757993e-02 4.56470460e-01 -7.16574907e-01
-5.18313646e-01 -8.64180386e-01 -8.76123726e-01 -7.60912836e-01
2.31099963e-01 1.56374350e-01 8.13569248e-01 3.73847783e-01
7.35415399e-01 1.56367481e-01 9.21345532e-01 -1.20570040e+00
5.42283356e-01 -4.79853928e-01 -7.80831277e-01 9.41858590e-01
1.42222434e-01 -9.55887258e-01 -5.07540047e-01 2.52560496e-01] | [10.299599647521973, 0.5022799968719482] |
c20de731-d330-4f16-b100-8b450af53659 | real-masks-and-fake-faces-on-the-masked-face | 2103.01546 | null | https://arxiv.org/abs/2103.01546v2 | https://arxiv.org/pdf/2103.01546v2.pdf | Real Masks and Spoof Faces: On the Masked Face Presentation Attack Detection | Face masks have become one of the main methods for reducing the transmission of COVID-19. This makes face recognition (FR) a challenging task because masks hide several discriminative features of faces. Moreover, face presentation attack detection (PAD) is crucial to ensure the security of FR systems. In contrast to the growing number of masked FR studies, the impact of face masked attacks on PAD has not been explored. Therefore, we present novel attacks with real face masks placed on presentations and attacks with subjects wearing masks to reflect the current real-world situation. Furthermore, this study investigates the effect of masked attacks on PAD performance by using seven state-of-the-art PAD algorithms under different experimental settings. We also evaluate the vulnerability of FR systems to masked attacks. The experiments show that real masked attacks pose a serious threat to the operation and security of FR systems. | ['Arjan Kuijper', 'Florian Kirchbuchner', 'Naser Damer', 'Meiling Fang'] | 2021-03-02 | null | null | null | null | ['face-presentation-attack-detection'] | ['computer-vision'] | [ 4.44686711e-01 -1.29233524e-01 3.33771557e-01 -2.08117113e-01
-1.61172554e-01 -7.87319481e-01 3.59944910e-01 -4.53030437e-01
-1.09637842e-01 5.02356589e-01 -1.21782824e-01 -1.79428741e-01
5.51321395e-02 -4.98006403e-01 -6.05815232e-01 -7.56336808e-01
-6.90865517e-01 -3.86984855e-01 2.18085006e-01 -1.12687632e-01
-3.28332111e-02 1.24392450e+00 -1.98073614e+00 5.46887815e-01
2.06507355e-01 1.09316015e+00 -2.22215712e-01 4.82132703e-01
3.10728520e-01 4.70613129e-02 -1.18173814e+00 -6.43123627e-01
4.29051667e-01 -2.56427228e-01 -4.01544534e-02 -1.14126027e-01
6.72723114e-01 -6.16663218e-01 -3.67631763e-01 9.07297492e-01
7.29142129e-01 -4.02105749e-01 3.31928283e-01 -1.36977625e+00
-4.25950915e-01 3.74021411e-01 -6.65652037e-01 4.88706648e-01
7.01342821e-01 5.74379116e-02 4.66699712e-02 -8.43408644e-01
1.65300190e-01 1.64176011e+00 4.07603234e-01 7.64211416e-01
-1.23525822e+00 -1.38810778e+00 -5.99261113e-02 1.30103350e-01
-1.82723475e+00 -9.76887286e-01 4.95485961e-01 -2.40256444e-01
4.92268026e-01 5.54763019e-01 7.42891133e-02 1.09939241e+00
2.73409545e-01 -2.92687863e-01 1.16856134e+00 -5.17585874e-01
-3.66731435e-02 6.16388798e-01 -4.98321578e-02 5.41046381e-01
6.26420856e-01 3.57179016e-01 -5.18294871e-01 -6.34400845e-01
4.41589534e-01 -3.88327092e-01 -4.50020224e-01 4.16089594e-01
-1.51088700e-01 6.48645997e-01 9.16595943e-03 4.61830974e-01
-1.25882998e-01 -2.28684306e-01 3.26355617e-03 1.39078841e-01
4.09872502e-01 9.96969566e-02 -1.56825513e-01 3.39155287e-01
-7.30579019e-01 7.12190336e-03 7.22795844e-01 4.20900553e-01
4.56755966e-01 3.34372938e-01 -3.08319569e-01 5.87503850e-01
5.18306136e-01 8.42245758e-01 -1.16615660e-01 -3.07049870e-01
1.05485775e-01 2.46300802e-01 -1.41018443e-02 -1.31870735e+00
1.32491468e-02 -1.93533953e-02 -4.34565008e-01 3.31011504e-01
3.08730990e-01 -2.79911250e-01 -7.43041456e-01 1.64865816e+00
2.07566783e-01 4.06463712e-01 -1.44070446e-01 4.55912501e-01
6.95086479e-01 3.42638701e-01 1.54269412e-01 -5.85526943e-01
1.52129400e+00 4.60348837e-02 -1.25095665e+00 -2.32572947e-02
1.04856700e-01 -1.27344525e+00 6.71682775e-01 4.41374153e-01
-8.93456161e-01 -5.61266005e-01 -1.25790370e+00 6.97620273e-01
-3.48827422e-01 1.47497833e-01 1.98100537e-01 2.09692836e+00
-1.11431563e+00 5.04767783e-02 -2.78934896e-01 -1.97015136e-01
5.47119915e-01 7.92662323e-01 -5.44853330e-01 -1.24281555e-01
-1.24166274e+00 9.50086355e-01 -3.50313276e-01 4.33491975e-01
-9.25513804e-01 -4.64566231e-01 -7.62935579e-01 -1.24457583e-01
3.22895736e-01 4.61000085e-01 8.66093814e-01 -6.39947295e-01
-1.24652457e+00 1.00036001e+00 -5.07495522e-01 -3.33201677e-01
2.75206447e-01 -4.70898926e-01 -1.20535600e+00 2.96849012e-01
-4.78158414e-01 1.50997952e-01 1.35186648e+00 -1.27229404e+00
9.50503945e-02 -4.53902483e-01 -2.16482565e-01 -5.22893012e-01
-6.09947264e-01 9.26694155e-01 -6.14401931e-03 -4.38775390e-01
-4.38384563e-01 -8.43155980e-01 3.05904597e-01 -2.82797422e-02
-2.61832356e-01 1.01353087e-01 1.66495883e+00 -6.33316815e-01
1.37761939e+00 -2.59269452e+00 -7.08994627e-01 3.79021704e-01
2.68914737e-03 8.61982465e-01 3.54085155e-02 4.25599277e-01
-2.01081604e-01 3.84004533e-01 1.76641658e-01 -2.36350656e-01
-2.11394399e-01 3.85809205e-02 -6.15566611e-01 9.42053318e-01
3.14539641e-01 3.15574288e-01 9.11975279e-02 -1.60957277e-01
7.74137378e-02 8.88617277e-01 -5.04102170e-01 4.85007614e-01
1.75719708e-01 3.97763640e-01 -2.64183044e-01 7.10793734e-01
1.33488083e+00 5.00858307e-01 2.39674151e-01 -8.30108523e-02
-4.06900607e-02 -1.35982499e-01 -1.30997705e+00 3.42752546e-01
6.70263171e-02 5.82692146e-01 3.80444437e-01 -2.68745691e-01
9.54445541e-01 6.41986012e-01 6.21632040e-02 -5.65230250e-01
5.07947028e-01 -1.29638985e-01 2.30067298e-01 -5.81370711e-01
2.56720304e-01 1.98728759e-02 2.37716675e-01 5.05328894e-01
-3.88652951e-01 4.59215075e-01 -1.62343577e-01 -2.07419693e-02
9.14238989e-01 -3.84904414e-01 -4.56846915e-02 -2.60483980e-01
8.25092316e-01 -8.66721153e-01 3.00602913e-01 4.88447696e-01
-4.60933387e-01 2.64503777e-01 4.64001924e-01 -1.34827256e-01
-7.86100850e-02 -1.12591100e+00 -3.22965026e-01 7.39957750e-01
1.03453986e-01 -3.10384810e-01 -1.22787142e+00 -5.65579832e-01
1.74019143e-01 4.84193116e-01 -5.97422540e-01 -2.78237522e-01
-6.31868541e-01 -7.31018245e-01 1.11341095e+00 1.13042206e-01
4.46010381e-01 -1.02689457e+00 -6.67616069e-01 -3.16078484e-01
9.52737182e-02 -1.66570091e+00 -5.20926774e-01 -4.13194478e-01
-1.39789626e-01 -1.35845327e+00 -2.78473020e-01 -5.93570352e-01
8.43083739e-01 5.23787498e-01 6.23421133e-01 6.37900651e-01
-8.13737929e-01 3.28611642e-01 -2.51241952e-01 -7.50503600e-01
-5.48129082e-01 -5.66369474e-01 5.25566757e-01 6.17395580e-01
5.36156237e-01 -1.77182496e-01 -4.32208389e-01 8.70258212e-01
-9.20908868e-01 -9.57930744e-01 2.76304334e-01 -6.27790242e-02
-6.42761290e-02 5.28644621e-01 6.14741445e-01 -7.95514822e-01
7.06241071e-01 -1.43107951e-01 -7.55932033e-01 1.41628206e-01
-1.22880243e-01 -3.73746097e-01 2.06242353e-01 -7.35648274e-01
-1.13137472e+00 -7.53553063e-02 -5.07794842e-02 -3.63516539e-01
-4.35709298e-01 -2.82911301e-01 -7.61783481e-01 -6.77893341e-01
6.35043919e-01 -1.43845409e-01 2.62776792e-01 -4.75202709e-01
-7.36677125e-02 8.72356951e-01 1.66393548e-01 -1.78091168e-01
1.19555426e+00 5.70555866e-01 4.61351238e-02 -1.47236180e+00
-3.62411886e-01 -3.28271352e-02 -2.95998096e-01 -5.32235444e-01
7.77990460e-01 -7.16541886e-01 -1.09119415e+00 6.84052885e-01
-1.08631909e+00 9.40632373e-02 6.08404875e-01 2.19718516e-01
1.90268815e-01 5.40623069e-01 -4.80806381e-01 -1.32759786e+00
-1.40824869e-01 -1.28917861e+00 8.97125006e-01 1.62405446e-01
-2.58946478e-01 -4.34270859e-01 -1.77185059e-01 3.69913101e-01
5.57010710e-01 4.13404256e-01 5.88920176e-01 -3.17464381e-01
-3.50400656e-01 -4.23251569e-01 -5.29157445e-02 5.18622100e-01
5.73203444e-01 2.54751027e-01 -1.68156278e+00 -7.18457103e-01
3.74674231e-01 2.19845492e-02 3.68403673e-01 2.45080978e-01
1.03957689e+00 -4.70484823e-01 -5.20103335e-01 3.28601301e-01
1.12801135e+00 4.23755586e-01 1.04087174e+00 -3.11404288e-01
1.18589193e-01 9.47182894e-01 4.98285860e-01 3.64027023e-01
-6.37469053e-01 7.55560994e-01 4.68011141e-01 -2.21996363e-02
-1.97455883e-02 7.88597465e-02 6.68074846e-01 -2.63779074e-01
2.51204967e-01 -5.09105980e-01 -5.45507848e-01 -1.04788326e-01
-7.68299103e-01 -1.10147154e+00 -1.34848177e-01 2.34279799e+00
3.16040307e-01 1.13154545e-01 1.20429538e-01 4.79627401e-01
1.24611545e+00 8.55083168e-02 7.76841715e-02 -5.29337108e-01
-8.59083608e-02 6.16635382e-01 3.98901701e-01 3.89658570e-01
-1.02143359e+00 7.55615413e-01 7.15200949e+00 3.69108021e-01
-1.33683562e+00 -1.51268514e-02 5.68557501e-01 -1.28063306e-01
3.14202756e-01 -4.53924984e-01 -1.30903029e+00 5.29709637e-01
9.87194777e-01 6.73958287e-02 3.10270935e-01 5.53318322e-01
2.90206701e-01 -1.74888909e-01 -8.46869171e-01 8.90053093e-01
5.01239777e-01 -7.23653972e-01 -1.50755987e-01 5.72802603e-01
7.76790380e-02 -7.17711329e-01 4.90817726e-01 -9.66301635e-02
-2.34714001e-01 -1.44961441e+00 3.42155188e-01 -1.56799763e-01
1.00325632e+00 -8.76889646e-01 5.96116602e-01 -7.55437762e-02
-1.22845173e+00 -2.00531051e-01 -2.76530743e-01 -2.22211987e-01
2.21072242e-01 3.38291287e-01 -9.32233334e-01 1.69570565e-01
7.47602344e-01 -2.29647547e-01 -6.33402228e-01 6.10197723e-01
-1.62803382e-01 6.87509418e-01 -3.90371501e-01 1.67004839e-01
-3.68800968e-01 3.63524437e-01 2.82854348e-01 1.04944801e+00
1.69727921e-01 1.46329165e-01 -1.61731541e-01 5.72453141e-01
2.28526369e-02 -8.77026916e-02 -9.63654816e-01 -8.72942433e-02
6.07530236e-01 1.22440732e+00 -5.47722340e-01 3.86086226e-01
-4.39490169e-01 7.04383731e-01 -2.87280172e-01 2.67642170e-01
-7.98428953e-01 -8.45634341e-02 1.14662421e+00 5.96546710e-01
4.14619148e-01 -1.63511217e-01 1.32734656e-01 -3.77687931e-01
2.24434763e-01 -1.11866212e+00 5.73520243e-01 -3.64760727e-01
-1.16086686e+00 9.93354917e-01 1.87492460e-01 -9.21250522e-01
1.04073741e-01 -5.94494402e-01 -6.22483969e-01 1.14082313e+00
-1.44721735e+00 -8.80513728e-01 -3.05619091e-01 7.43248403e-01
6.30966499e-02 -3.00928205e-01 1.01852906e+00 5.11067092e-01
-7.30216742e-01 1.19675994e+00 -4.90040690e-01 8.53253752e-02
6.91736102e-01 -9.67958793e-02 4.25145686e-01 1.08151639e+00
-1.52246416e-01 7.95213103e-01 7.02783048e-01 -7.40501881e-01
-1.44671953e+00 -7.36630619e-01 4.32495117e-01 -3.83692652e-01
2.39423633e-01 -9.35798943e-01 -1.02366102e+00 4.11086500e-01
2.79687464e-01 -1.87081527e-02 1.08190668e+00 -3.65430534e-01
-4.79300797e-01 -1.61334082e-01 -1.72213233e+00 5.91366827e-01
7.47785687e-01 -5.94932258e-01 -6.48924336e-02 5.95222116e-02
3.39501590e-01 6.47913516e-02 -4.16559219e-01 6.01180136e-01
7.28150368e-01 -9.84530389e-01 9.51384425e-01 -3.10519487e-01
-3.91148388e-01 -2.44089663e-01 -3.18569124e-01 -7.37804055e-01
5.32788448e-02 -8.41555178e-01 5.71810864e-02 1.50623620e+00
1.02226675e-01 -9.59133983e-01 7.12453127e-01 5.83040059e-01
5.66786766e-01 -8.43089893e-02 -9.67031240e-01 -8.87208283e-01
-4.94101554e-01 -2.11883098e-01 6.85564756e-01 7.46921957e-01
-3.30297440e-01 -2.34749720e-01 -4.50368017e-01 1.01854265e+00
7.86692619e-01 -6.14623725e-01 8.42350602e-01 -1.00286174e+00
-7.58584067e-02 -7.32938498e-02 -6.75888658e-01 -8.73459950e-02
2.17216626e-01 -2.27576509e-01 -1.39342576e-01 -4.61268067e-01
-1.85855716e-01 -8.86200294e-02 -8.82101208e-02 3.49502981e-01
-1.14812396e-01 7.87014186e-01 2.30314493e-01 -1.58660665e-01
3.48519862e-01 2.97496110e-01 6.38372004e-01 1.66605428e-01
1.03327140e-01 8.48887265e-02 -6.15282774e-01 5.75155437e-01
8.95844638e-01 -6.66948795e-01 -5.20244956e-01 -2.23659631e-03
-4.12340492e-01 -2.21723080e-01 2.61951774e-01 -1.05675316e+00
-1.27530709e-01 4.83062826e-02 5.55723429e-01 -4.30837363e-01
6.02480054e-01 -1.15797186e+00 2.64396757e-01 6.96661353e-01
2.46649146e-01 3.10210168e-01 8.90191317e-01 2.00901777e-01
1.33970425e-01 -6.61148205e-02 9.52208579e-01 1.98029578e-01
-4.68612500e-02 1.05325967e-01 -7.06901133e-01 -3.95390421e-01
1.43253565e+00 -4.76919502e-01 -6.44286215e-01 -2.24176213e-01
-4.12678808e-01 -1.26517802e-01 3.33768487e-01 5.19594729e-01
7.82388330e-01 -7.36170948e-01 -6.81519747e-01 1.02782333e+00
-1.40363500e-01 -8.17647576e-01 3.69097888e-01 4.07273054e-01
-5.14809906e-01 2.58809179e-01 -3.83134425e-01 -2.99570501e-01
-2.38180304e+00 8.41740668e-01 3.35439742e-01 3.29792023e-01
-2.15036497e-01 8.47565055e-01 4.39685345e-01 2.82451808e-01
5.26233494e-01 5.73521018e-01 -3.05076808e-01 -9.89799947e-02
1.21633887e+00 2.42221817e-01 3.16838890e-01 -8.35044444e-01
-6.21606350e-01 3.43607545e-01 -4.31548625e-01 3.35151106e-02
7.77877212e-01 1.51062161e-01 -2.60401726e-01 -2.18224451e-01
9.17518377e-01 5.73251009e-01 -7.77836025e-01 7.68694654e-02
-1.58999830e-01 -9.21060979e-01 -2.30132967e-01 -5.19464731e-01
-1.12893355e+00 8.12302232e-01 1.05566859e+00 3.78206819e-01
1.14107609e+00 -1.21499032e-01 4.19789523e-01 -4.83881794e-02
3.85080367e-01 -4.07135576e-01 1.84002295e-01 -1.19183294e-01
8.65469337e-01 -7.44626105e-01 7.44027793e-02 -1.24158001e+00
-2.28567183e-01 6.17552400e-01 6.89451694e-01 2.09614724e-01
1.00919187e+00 7.56373405e-01 4.55832779e-01 -2.47875810e-01
-4.85157520e-01 2.88590550e-01 1.76185712e-01 7.28566349e-01
3.44043612e-01 -1.09909214e-01 -1.77490681e-01 4.64382291e-01
-1.29220724e-01 -2.30512232e-01 5.10063231e-01 1.12304819e+00
-2.12355420e-01 -1.22499478e+00 -9.49503481e-01 3.10233712e-01
-1.06537557e+00 2.06117228e-01 -8.31590056e-01 7.12009072e-01
2.12211329e-02 1.42987299e+00 -1.75463304e-01 -5.25040448e-01
2.83777535e-01 1.61933750e-01 5.10693967e-01 -5.79755664e-01
-8.32379043e-01 -1.13628656e-01 -2.07149913e-03 -4.78687912e-01
-3.72276932e-01 -5.44290364e-01 -6.76793873e-01 -7.26507187e-01
-4.91063803e-01 7.87560269e-02 6.17135525e-01 7.87306786e-01
3.93351018e-01 2.77309179e-01 9.31457937e-01 -6.05627835e-01
-3.74095678e-01 -8.56163442e-01 -7.97802269e-01 1.26918107e-01
3.20527166e-01 -9.21886802e-01 -6.28555536e-01 3.62747982e-02] | [13.061750411987305, 1.0637691020965576] |
03f926fb-5284-434a-b1d2-efbbe1fd25fa | phoneme-aware-and-channel-wise-attentive | 2106.13514 | null | https://arxiv.org/abs/2106.13514v1 | https://arxiv.org/pdf/2106.13514v1.pdf | Phoneme-aware and Channel-wise Attentive Learning for Text DependentSpeaker Verification | This paper proposes a multi-task learning network with phoneme-aware and channel-wise attentive learning strategies for text-dependent Speaker Verification (SV). In the proposed structure, the frame-level multi-task learning along with the segment-level adversarial learning is adopted for speaker embedding extraction. The phoneme-aware attentive pooling is exploited on frame-level features in the main network for speaker classifier, with the corresponding posterior probability for the phoneme distribution in the auxiliary subnet. Further, the introduction of Squeeze and Excitation (SE-block) performs dynamic channel-wise feature recalibration, which improves the representational ability. The proposed method exploits speaker idiosyncrasies associated with pass-phrases, and is further improved by the phoneme-aware attentive pooling and SE-block from temporal and channel-wise aspects, respectively. The experiments conducted on RSR2015 Part 1 database confirm that the proposed system achieves outstanding results for textdependent SV. | ['Qingyang Hong', 'Lin Li', 'Zheng Li', 'Yan Liu'] | 2021-06-25 | null | null | null | null | ['text-dependent-speaker-verification'] | ['speech'] | [ 5.43729886e-02 -2.73258358e-01 -1.18613087e-01 -4.71387714e-01
-1.21901226e+00 -3.92899871e-01 4.58990842e-01 -1.95697978e-01
-4.19786721e-01 4.37475204e-01 6.40007317e-01 -2.18880430e-01
3.34349602e-01 -2.11080864e-01 -7.07193971e-01 -1.20111132e+00
-2.05693677e-01 -4.45790321e-01 -8.98358505e-03 -1.54895365e-01
-2.24939510e-02 3.88304740e-01 -1.17817354e+00 6.30907774e-01
4.81329083e-01 1.11820483e+00 1.26917332e-01 8.99832249e-01
-2.17668280e-01 5.84028602e-01 -7.29865670e-01 -3.04050148e-01
1.08404187e-02 -4.74779278e-01 -3.60845774e-01 -1.42836794e-01
4.20126379e-01 -1.70833066e-01 -5.47322631e-01 8.92207146e-01
1.11478376e+00 1.67425573e-01 3.96631658e-01 -1.16083097e+00
-3.86937052e-01 7.69713104e-01 -5.63958526e-01 6.48531497e-01
1.74868181e-02 3.51943791e-01 8.92366648e-01 -1.08788347e+00
-4.06437293e-02 1.51197803e+00 6.50478959e-01 7.10016131e-01
-8.69722605e-01 -1.07127583e+00 4.60612774e-01 7.21845627e-01
-1.63156140e+00 -9.16417718e-01 1.21851242e+00 -2.31840592e-02
9.16944981e-01 1.77569613e-01 3.03412825e-01 1.28622496e+00
2.68925488e-01 1.09629655e+00 9.70250130e-01 -2.22892359e-01
-7.47154653e-02 2.89248109e-01 -1.32437227e-02 6.42613232e-01
-6.71432436e-01 2.58997500e-01 -9.48390007e-01 9.64108929e-02
3.93357366e-01 -2.68188834e-01 -4.54477906e-01 1.36954576e-01
-1.02193987e+00 7.26394355e-01 3.48195434e-01 4.64660287e-01
-2.48308465e-01 2.03182083e-02 9.17214274e-01 3.01671296e-01
4.78485197e-01 -3.80819738e-01 -5.10991275e-01 -1.13066696e-02
-1.16841209e+00 -1.47370636e-01 5.11906505e-01 9.21955407e-01
4.92501736e-01 8.23673964e-01 -7.68206179e-01 9.46858764e-01
6.73157513e-01 6.27690136e-01 9.13820446e-01 -1.43191233e-01
6.65731609e-01 -1.97331160e-01 -4.00470316e-01 -8.38367164e-01
-2.32431710e-01 -8.53443325e-01 -9.31564867e-01 -8.39285329e-02
6.59710402e-03 -4.56638515e-01 -8.50958526e-01 1.90184593e+00
2.73235172e-01 7.75823772e-01 1.92158461e-01 6.78536296e-01
1.16252208e+00 1.06711709e+00 4.06587690e-01 -3.64810735e-01
1.59063673e+00 -1.16659689e+00 -1.11045289e+00 -1.26017332e-01
3.07938129e-01 -6.48668051e-01 8.16839635e-01 1.40025094e-01
-8.79844546e-01 -1.00473833e+00 -1.16661346e+00 1.14956267e-01
-3.71348202e-01 1.92127436e-01 1.26056209e-01 1.06622469e+00
-1.06803238e+00 -1.64409518e-01 -5.49074709e-01 2.82863617e-01
5.16768813e-01 3.70797038e-01 -1.50443003e-01 1.91937927e-02
-1.59948730e+00 7.57859588e-01 2.10453659e-01 5.06294191e-01
-1.39696515e+00 -6.53292120e-01 -1.19222260e+00 2.70430565e-01
-8.86772573e-02 -2.09521890e-01 1.03614974e+00 -1.10785377e+00
-1.98797333e+00 4.04962689e-01 -6.05135560e-01 -5.24406493e-01
2.83078611e-01 1.74013227e-01 -9.58881021e-01 2.41204530e-01
-4.04396445e-01 6.11833930e-01 1.81569505e+00 -9.82233763e-01
-6.44856989e-01 -1.98130205e-01 -4.34128642e-01 4.13177401e-01
-7.22288907e-01 2.73105979e-01 -3.05262059e-01 -9.08913612e-01
-1.47247404e-01 -4.15152311e-01 1.04614682e-01 -3.37221354e-01
-3.35314453e-01 -2.69835532e-01 1.42172241e+00 -1.22466862e+00
1.29906225e+00 -2.77888560e+00 7.35372603e-02 1.51547091e-02
-1.18403524e-01 3.55013311e-01 -4.03022110e-01 1.04662646e-02
-2.76580691e-01 -8.88526738e-02 -6.01477921e-02 -7.05859601e-01
-8.69605541e-02 -3.21514368e-01 -2.04387739e-01 6.01121783e-01
3.13406259e-01 9.71237659e-01 -4.81484115e-01 -7.47365236e-01
5.07949233e-01 9.03238952e-01 -2.30493650e-01 -2.83750566e-03
2.06893161e-01 5.31490088e-01 -3.48903418e-01 5.14309347e-01
1.00306845e+00 5.27410209e-01 -3.19045752e-01 -4.26915437e-01
-4.06452268e-02 3.90373886e-01 -1.06970334e+00 1.80560791e+00
-8.53587687e-01 6.36377096e-01 6.06047213e-01 -9.63440001e-01
9.51449275e-01 8.98680568e-01 1.10583894e-01 -6.59922838e-01
1.37625352e-01 -3.14916931e-02 8.46322626e-02 -5.07610977e-01
3.41503710e-01 -2.85006225e-01 -2.10465088e-01 1.89369261e-01
3.82590294e-01 5.73179498e-02 -7.14140296e-01 -2.95863599e-02
7.08328485e-01 -2.84064878e-02 -1.18574612e-01 -2.84998327e-01
1.31946337e+00 -7.34476388e-01 6.74927533e-01 4.81620908e-01
-8.20282459e-01 2.00245574e-01 1.07519263e-02 -2.06036661e-02
-6.82396531e-01 -8.83737326e-01 -1.96786866e-01 1.31491160e+00
6.12616353e-02 -8.71640965e-02 -6.85890853e-01 -7.40748107e-01
-1.66060373e-01 7.31885135e-01 -4.44097817e-01 -4.02665496e-01
-6.14317238e-01 -6.44580722e-01 1.05755222e+00 4.86757517e-01
9.18229401e-01 -1.03271019e+00 1.49558932e-01 5.17609000e-01
-2.53680080e-01 -1.03848195e+00 -8.90127301e-01 1.73566848e-01
-6.12073123e-01 -2.20098570e-01 -1.07305169e+00 -1.03975725e+00
1.84227586e-01 -2.77613439e-02 2.01642603e-01 -5.60895443e-01
4.24872637e-02 2.41662607e-01 -2.77655959e-01 -3.08435291e-01
-3.27720493e-01 9.64377522e-02 -3.66369560e-02 7.36238658e-01
4.40814734e-01 -5.01813650e-01 -4.51351315e-01 3.11676621e-01
-6.75972044e-01 -4.10788238e-01 5.92291296e-01 1.13680649e+00
2.48389617e-01 1.98244989e-01 1.11681187e+00 -1.90205768e-01
5.99128008e-01 -1.91620946e-01 -2.31379300e-01 1.13814160e-01
-4.96917055e-04 -5.22294082e-02 7.21667409e-01 -6.05274498e-01
-1.58199573e+00 1.20715998e-01 -5.89835227e-01 -5.50273359e-01
-1.97483599e-01 1.18942954e-01 -7.41402626e-01 -2.54411042e-01
3.48364472e-01 8.90085638e-01 -3.13953981e-02 -1.83484793e-01
4.05342638e-01 1.06688893e+00 4.30681288e-01 -2.05803499e-01
7.76232898e-01 2.98194855e-01 -4.52600002e-01 -9.98249888e-01
-4.08368021e-01 -3.74182165e-01 -4.10005122e-01 -3.49993497e-01
9.57502484e-01 -1.40302336e+00 -8.21547389e-01 9.45397615e-01
-1.18774390e+00 -1.22726306e-01 -8.14221799e-02 6.99855268e-01
-3.64704221e-01 3.31434101e-01 -7.60519207e-01 -9.64840531e-01
-4.87261057e-01 -1.38331795e+00 9.98492658e-01 1.97406247e-01
3.65145117e-01 -1.05697572e+00 -2.96417803e-01 4.30238575e-01
8.30657423e-01 -3.83194745e-01 6.64206862e-01 -9.54620421e-01
-5.76813936e-01 -1.34551093e-01 -2.14712881e-02 7.24695385e-01
6.27788678e-02 -3.33252579e-01 -1.73224878e+00 -4.80286062e-01
2.90996820e-01 -1.30727485e-01 1.03352714e+00 8.12175810e-01
1.46768475e+00 -3.27079505e-01 -1.95175782e-01 7.00283945e-01
1.13860333e+00 2.51720369e-01 5.57086825e-01 -9.78672653e-02
7.48568416e-01 4.17646945e-01 2.62994200e-01 2.63838232e-01
4.30960774e-01 7.57029653e-01 1.13576978e-01 -7.64576495e-02
-4.30874914e-01 -1.55774772e-01 7.11760461e-01 1.15259576e+00
3.25311005e-01 -4.63562250e-01 -4.38600063e-01 5.66374719e-01
-1.31837320e+00 -1.08556437e+00 2.30529368e-01 2.01321483e+00
7.16299057e-01 3.48645300e-02 -1.71583071e-02 2.78479904e-01
1.12374294e+00 8.08374286e-01 -6.15204990e-01 -5.53414941e-01
-5.33692718e-01 1.88754588e-01 3.33634049e-01 6.65705085e-01
-1.22422493e+00 9.54714000e-01 5.39487553e+00 1.51070321e+00
-1.50193226e+00 7.88399100e-01 5.78970730e-01 2.00789683e-02
-9.52383652e-02 -5.74116588e-01 -1.12836730e+00 5.56197762e-01
1.20308042e+00 -1.17325159e-02 1.72691345e-01 5.21255970e-01
4.05599475e-01 3.30004901e-01 -8.54159713e-01 1.14758313e+00
4.78739381e-01 -9.91566479e-01 -1.80410489e-01 -1.19253255e-01
3.82509232e-01 -3.47738601e-02 4.58199084e-01 5.20181060e-01
-2.91000783e-01 -8.68154883e-01 8.37757230e-01 2.52424926e-01
9.27223861e-01 -1.00561118e+00 8.29076707e-01 1.64254643e-02
-1.68248379e+00 -3.16695213e-01 3.43480371e-02 5.01888454e-01
4.03974742e-01 2.97858626e-01 -7.80951262e-01 5.87949038e-01
6.17414057e-01 7.16759205e-01 -2.61746764e-01 8.35793793e-01
-3.59282196e-01 9.42319155e-01 -2.27784216e-01 7.81171629e-03
1.55669913e-01 4.21964109e-01 8.85491014e-01 1.72676599e+00
-1.12673581e-01 -1.57669336e-01 -2.27844656e-01 3.58965963e-01
-1.67500407e-01 1.75249532e-01 -2.75583267e-01 3.00183445e-01
5.24193823e-01 1.26307762e+00 -1.79004192e-01 -2.91526496e-01
-3.87944579e-01 1.16906726e+00 -2.58151591e-01 6.91211343e-01
-1.02519989e+00 -7.20526278e-01 5.79500616e-01 -5.00948787e-01
8.34691107e-01 -1.16627544e-01 -2.46678725e-01 -9.47072685e-01
-3.20740603e-02 -8.34387779e-01 2.65989959e-01 -4.23450500e-01
-1.14323020e+00 7.27034092e-01 -3.96325260e-01 -1.11246979e+00
-5.53067662e-02 -2.89749771e-01 -9.64487076e-01 1.30222404e+00
-2.03661346e+00 -1.41215909e+00 1.52726576e-01 1.13049877e+00
9.71530616e-01 -4.44390535e-01 7.62391508e-01 6.45302296e-01
-9.50790763e-01 1.46177506e+00 6.93406910e-02 2.64750063e-01
7.20745385e-01 -7.41423309e-01 2.51372337e-01 1.02715027e+00
-1.18381158e-01 2.15763554e-01 3.79321903e-01 -2.66226292e-01
-1.14683771e+00 -1.25305855e+00 9.24852848e-01 9.79008377e-02
4.25381213e-01 -7.36689806e-01 -8.44058871e-01 5.03764808e-01
3.32695901e-01 8.36509243e-02 6.58645093e-01 -1.52355775e-01
-5.51737428e-01 -5.12753367e-01 -1.27897358e+00 1.67735875e-01
4.96303529e-01 -1.18140435e+00 -5.24171531e-01 2.38614023e-01
9.46400225e-01 -2.75265276e-01 -6.55365765e-01 1.43523917e-01
5.08019447e-01 -5.30221403e-01 1.11007750e+00 -3.20139915e-01
-7.65703171e-02 -3.63558590e-01 -1.26737043e-01 -1.32521403e+00
-2.59642214e-01 -7.07056224e-01 -1.34075418e-01 1.68546867e+00
6.25309527e-01 -6.61481321e-01 5.12559295e-01 -2.56708656e-02
-5.63219070e-01 -3.65299255e-01 -1.50840211e+00 -7.51945198e-01
-1.01267926e-01 -5.56547284e-01 5.86006880e-01 8.36641252e-01
-1.31176800e-01 2.20091283e-01 -6.11412644e-01 6.00360572e-01
5.29000759e-01 -4.20363426e-01 2.41284236e-01 -5.03259778e-01
-2.91060418e-01 -3.63106072e-01 -3.87906909e-01 -1.20183051e+00
5.67720175e-01 -7.87772179e-01 3.21422935e-01 -8.24790895e-01
-2.44308084e-01 -1.39970422e-01 -6.37859106e-01 2.86654323e-01
-1.40375540e-01 1.19713685e-02 1.29565477e-01 -2.34506115e-01
-3.63600433e-01 9.16564107e-01 1.21691859e+00 -4.80805248e-01
-3.30563903e-01 2.66029626e-01 -2.97619849e-01 3.49396437e-01
7.00461149e-01 -3.26557726e-01 -1.58109829e-01 -2.98316568e-01
-6.81601346e-01 1.94964290e-01 3.50093991e-01 -8.72335434e-01
4.39906359e-01 5.08324385e-01 3.74380410e-01 -6.63131952e-01
7.88396955e-01 -7.87116289e-01 -4.39720988e-01 5.96476734e-01
-5.63250422e-01 -3.51867527e-01 7.00421512e-01 6.92650855e-01
-5.37110746e-01 2.03491282e-02 1.02339888e+00 2.80466080e-01
-7.54435182e-01 4.64649141e-01 -5.90317786e-01 -2.40234450e-01
8.22259665e-01 -9.68926623e-02 -6.84509352e-02 -5.24361074e-01
-7.27002323e-01 2.00711682e-01 -5.49124122e-01 5.29277861e-01
6.64247453e-01 -1.31310892e+00 -9.51464772e-01 6.48590624e-01
3.74123231e-02 -3.55754226e-01 9.61558342e-01 7.91667223e-01
2.02825382e-01 4.23123181e-01 -4.83458973e-02 -6.51838958e-01
-1.23785388e+00 2.66979158e-01 5.74055493e-01 -2.04700008e-02
-2.13545099e-01 1.54661465e+00 2.22230330e-01 -2.04657018e-01
5.97719073e-01 -2.09623829e-01 -1.94943026e-01 2.11914182e-01
6.71133399e-01 2.12415323e-01 1.20140076e-01 -1.00005293e+00
-5.32288194e-01 3.62344861e-01 -3.02416056e-01 -4.86747712e-01
1.12448990e+00 -4.30297315e-01 2.76897073e-01 3.73200178e-01
1.32618105e+00 1.74450591e-01 -1.38305175e+00 -4.01203871e-01
-6.31840229e-01 -1.18178315e-01 6.39778078e-01 -8.49298596e-01
-1.19891691e+00 1.42385268e+00 9.34452713e-01 -1.73661038e-01
1.34609759e+00 -1.83489799e-01 9.07206476e-01 -1.09422639e-01
-9.15664136e-02 -9.61616099e-01 -8.15391317e-02 3.41320008e-01
7.99844205e-01 -1.05770671e+00 -5.65803051e-01 -1.45947367e-01
-5.51896930e-01 8.60134363e-01 5.27266920e-01 2.77761221e-01
9.36696708e-01 2.60595530e-01 1.98682174e-02 3.05226028e-01
-4.47002828e-01 1.22765817e-01 2.12809429e-01 7.54485309e-01
1.20351523e-01 -6.53783511e-03 7.83386603e-02 6.14436269e-01
1.95470620e-02 -5.45771241e-01 5.30160554e-02 4.56559986e-01
-3.98600519e-01 -9.32235479e-01 -5.35986841e-01 3.91991548e-02
-4.53575790e-01 -5.28773189e-01 2.01355740e-01 2.94103116e-01
1.55391574e-01 1.12008178e+00 5.66293374e-02 -5.08418202e-01
7.38084763e-02 4.16187972e-01 1.51093557e-01 -3.25681627e-01
-9.18046832e-01 3.26910198e-01 -2.84713581e-02 -3.71554643e-01
-3.82032126e-01 -7.61417031e-01 -1.01469350e+00 -1.32892374e-02
-5.04403234e-01 9.80315730e-02 1.11034238e+00 1.00243163e+00
4.89164114e-01 8.31537783e-01 1.18275487e+00 -8.58640134e-01
-4.39660102e-01 -1.29170728e+00 -6.51912451e-01 -2.68608704e-02
1.00979125e+00 -2.80606806e-01 -6.64747357e-01 1.67065471e-01] | [14.369524955749512, 6.080367088317871] |
997f8042-cf67-4859-9c20-b807101c8a63 | knowledge-grounded-conversational-symptom | 2101.09773 | null | https://arxiv.org/abs/2101.09773v1 | https://arxiv.org/pdf/2101.09773v1.pdf | Knowledge Grounded Conversational Symptom Detection with Graph Memory Networks | In this work, we propose a novel goal-oriented dialog task, automatic symptom detection. We build a system that can interact with patients through dialog to detect and collect clinical symptoms automatically, which can save a doctor's time interviewing the patient. Given a set of explicit symptoms provided by the patient to initiate a dialog for diagnosing, the system is trained to collect implicit symptoms by asking questions, in order to collect more information for making an accurate diagnosis. After getting the reply from the patient for each question, the system also decides whether current information is enough for a human doctor to make a diagnosis. To achieve this goal, we propose two neural models and a training pipeline for the multi-step reasoning task. We also build a knowledge graph as additional inputs to further improve model performance. Experiments show that our model significantly outperforms the baseline by 4%, discovering 67% of implicit symptoms on average with a limited number of questions. | ['James Glass', 'Shang-Wen Li', 'Hongyin Luo'] | 2021-01-24 | null | https://aclanthology.org/2020.clinicalnlp-1.16 | https://aclanthology.org/2020.clinicalnlp-1.16.pdf | emnlp-clinicalnlp-2020-11 | ['goal-oriented-dialog'] | ['natural-language-processing'] | [ 1.48810908e-01 9.84363139e-01 -4.17967737e-01 -8.16784918e-01
-8.96156013e-01 -4.79941517e-01 -8.87654051e-02 6.52224064e-01
-3.40681970e-01 6.97232842e-01 4.51389551e-01 -4.87040102e-01
-4.85013165e-02 -8.29630077e-01 1.54524118e-01 -3.43089461e-01
1.12196498e-01 1.12430060e+00 3.12580615e-01 -1.49440721e-01
-1.54991686e-01 8.45427215e-02 -6.08565032e-01 6.62981510e-01
1.01081264e+00 7.12589920e-01 2.93177694e-01 9.26865935e-01
-2.17537329e-01 1.58093703e+00 -7.97302723e-01 -2.86603242e-01
-2.57091701e-01 -7.03148842e-01 -1.39507270e+00 4.12515581e-01
-3.54101241e-01 -1.02481568e+00 -3.95290591e-02 6.43887103e-01
2.73403674e-01 1.20528955e-02 3.50719690e-01 -8.88628960e-01
-4.19788450e-01 8.42850387e-01 3.00618857e-01 -1.66989908e-01
7.38308847e-01 4.32128340e-01 9.28615332e-01 -4.11805898e-01
6.55709565e-01 1.30858588e+00 3.90494287e-01 1.14605057e+00
-8.27998757e-01 -2.70617515e-01 2.54068673e-01 8.81452113e-02
-8.40570271e-01 -2.60008752e-01 4.63910550e-01 -2.37131163e-01
7.24249125e-01 5.87819934e-01 4.92962271e-01 8.68249238e-01
-2.02173635e-01 7.25766182e-01 9.05386508e-01 -8.79140049e-02
4.32448208e-01 5.63681304e-01 7.46963561e-01 9.88402724e-01
-1.84498921e-01 -5.44402659e-01 -1.58617780e-01 -6.07256591e-01
4.80978072e-01 2.57334679e-01 -3.54580522e-01 2.57964790e-01
-8.93194377e-01 9.24365938e-01 6.98422611e-01 4.26155001e-01
-6.01425409e-01 -4.99802053e-01 1.20481849e-01 4.41833377e-01
2.10045919e-01 5.76399863e-01 -7.04638779e-01 1.09163642e-01
-4.37824667e-01 3.11251312e-01 1.52230442e+00 4.63656098e-01
3.14141899e-01 -6.53791547e-01 -7.11213946e-01 7.30644286e-01
2.78261632e-01 3.52116704e-01 3.66801590e-01 -1.21420956e+00
2.82960504e-01 1.37109244e+00 4.53076094e-01 -5.68684220e-01
-8.28703523e-01 -1.68897323e-02 -6.13523483e-01 -2.41786495e-01
7.06764102e-01 -5.20421624e-01 -9.28413510e-01 1.60457659e+00
6.37858033e-01 -8.28246400e-02 3.78621072e-01 1.15387583e+00
1.22340381e+00 2.80936986e-01 2.91617990e-01 -3.91444892e-01
1.89715219e+00 -1.07073164e+00 -9.86867309e-01 -2.47611403e-01
1.14827681e+00 -2.89122909e-01 7.50669122e-01 5.18117309e-01
-1.04326785e+00 -2.20494106e-01 -4.82663989e-01 -2.31911197e-01
-1.34008989e-01 1.65681466e-01 5.42743981e-01 1.18369125e-01
-1.08144641e+00 4.72015232e-01 -8.44308138e-01 -4.58047986e-01
-7.60080991e-04 5.30027926e-01 4.37976830e-02 -1.76765040e-01
-1.25095725e+00 7.97786355e-01 4.19577628e-01 -3.76245799e-03
-5.48479855e-01 -5.93176961e-01 -8.58408809e-01 2.19886079e-01
7.60322928e-01 -1.08808625e+00 1.77816105e+00 -5.08164823e-01
-1.45418286e+00 1.03258097e+00 -4.49787974e-01 -4.97760057e-01
5.04640520e-01 -2.13446587e-01 -2.57429928e-01 5.80260754e-01
4.00396325e-02 7.52776146e-01 3.48483741e-01 -7.97340274e-01
-6.18905544e-01 -4.35284376e-01 4.41561788e-01 3.44297469e-01
-1.07400596e-01 -3.90475169e-02 -8.27364028e-01 -1.98028907e-02
-1.52161601e-03 -9.27239716e-01 -6.44786954e-01 -3.83525752e-02
-8.06631148e-01 -6.87078297e-01 6.70535386e-01 -9.64906752e-01
1.20852196e+00 -1.74654543e+00 -2.18947306e-02 7.49245882e-02
9.40253913e-01 4.19775218e-01 2.40731630e-02 3.28784138e-01
1.60863474e-01 -5.19583225e-02 -3.05912971e-01 -3.55374664e-01
-2.99783438e-01 4.59276468e-01 -1.89212561e-01 -3.44485849e-01
5.59004903e-01 1.12437701e+00 -9.84875381e-01 -6.39232516e-01
3.15124728e-02 1.56854227e-01 -5.67603588e-01 9.13888872e-01
-7.39225149e-01 8.64004374e-01 -9.52175379e-01 5.11903107e-01
4.79309410e-01 -1.15598679e+00 7.68597960e-01 2.02869579e-01
4.78876114e-01 8.56609464e-01 -6.02597654e-01 1.22758973e+00
-1.65747061e-01 -1.60726145e-01 5.32551229e-01 -9.17975903e-01
6.92312062e-01 6.36017442e-01 5.37367225e-01 -3.12691331e-01
2.99082696e-01 -6.57410920e-02 1.56842783e-01 -9.90823328e-01
-2.35549062e-02 6.18955940e-02 -9.41404030e-02 6.51132166e-01
-3.93834233e-01 1.01245061e-01 1.35432363e-01 6.03981912e-01
1.72849870e+00 -6.30219460e-01 3.32222283e-01 2.37974137e-01
8.26645732e-01 3.67002040e-01 6.10437870e-01 7.75774598e-01
-1.83727413e-01 1.29324034e-01 6.10277414e-01 -5.19749820e-01
-2.33305499e-01 -8.46184254e-01 2.22368672e-01 8.46976697e-01
3.44961137e-02 -5.02463937e-01 -8.05339634e-01 -1.21440673e+00
1.16296023e-01 7.20294476e-01 -5.08856654e-01 -1.01294838e-01
-5.55024445e-01 -4.54987556e-01 1.10751443e-01 5.73038340e-01
5.35252810e-01 -1.38680184e+00 -6.59070611e-01 4.21817869e-01
-6.93029284e-01 -1.17231667e+00 -3.02682757e-01 -5.16012050e-02
-9.42245543e-01 -1.48817158e+00 -5.43509960e-01 -8.12741995e-01
7.54918814e-01 -4.53348458e-03 1.14433420e+00 7.59521723e-01
-2.56990939e-01 2.62293637e-01 -2.79353708e-01 -1.81983158e-01
-7.83617139e-01 5.20819686e-02 -4.49655533e-01 -3.84135723e-01
8.36685896e-01 -1.58894971e-01 -8.98948193e-01 2.69201279e-01
-6.77678943e-01 1.63921148e-01 5.19884646e-01 6.93546534e-01
1.86538368e-01 -3.98048550e-01 6.48465812e-01 -1.37242973e+00
1.04231262e+00 -6.73733294e-01 -2.55672812e-01 2.42031559e-01
-5.04793227e-01 2.68244445e-01 5.12012720e-01 -3.52680445e-01
-1.10534620e+00 7.51842409e-02 -5.80690503e-01 -1.19512118e-02
-4.38687414e-01 7.19407856e-01 2.06201732e-01 5.28395653e-01
5.60504675e-01 -1.30970791e-01 -4.55998397e-03 -7.35083163e-01
2.75648087e-01 1.04122603e+00 5.09020627e-01 -1.09770205e-02
3.53857100e-01 1.72110200e-01 -4.78023380e-01 -3.87951523e-01
-1.47166789e+00 -8.63795161e-01 -3.42843294e-01 1.86577275e-01
1.25779963e+00 -7.27012753e-01 -1.39433849e+00 -7.39675772e-04
-1.23639786e+00 -5.89844167e-01 -7.94547349e-02 3.17031384e-01
-1.18000723e-01 1.32262439e-01 -1.10541666e+00 -9.75862205e-01
-6.29547536e-01 -9.94796813e-01 1.01669371e+00 2.02570528e-01
-7.91553020e-01 -1.16797400e+00 8.24342966e-02 7.95170069e-01
2.96545714e-01 -8.14431533e-02 1.00021422e+00 -1.24303555e+00
-5.77943742e-01 1.41513220e-03 -3.23213309e-01 -1.36663094e-01
4.66691613e-01 -7.01226890e-01 -8.76106977e-01 -7.33513269e-04
5.19074321e-01 -6.97947383e-01 6.77632272e-01 1.57311440e-01
1.05571938e+00 -7.49608099e-01 -6.66101575e-01 1.87822082e-03
6.65334344e-01 3.94974679e-01 1.60757095e-01 -2.86469251e-01
4.09358948e-01 8.38405848e-01 6.72869742e-01 4.07832921e-01
9.40089107e-01 4.58053738e-01 1.73797682e-01 -2.17463240e-01
7.03113750e-02 1.80318356e-01 -1.90748591e-02 8.13004375e-01
4.63025987e-01 -2.28884891e-01 -8.55558693e-01 5.05307674e-01
-1.99237204e+00 -4.07701224e-01 -1.98520318e-01 1.77446544e+00
1.35448098e+00 1.65720414e-02 1.25129759e-01 -1.20528124e-01
3.07664990e-01 -3.95165741e-01 -8.86879325e-01 -3.01929384e-01
5.80359340e-01 9.06532854e-02 3.36670242e-02 9.75757301e-01
-8.20036113e-01 1.03214359e+00 6.31170559e+00 1.08901234e-02
-1.13056755e+00 -7.80837908e-02 7.75458038e-01 2.25178227e-01
-1.55349135e-01 -1.82133511e-01 -7.34409094e-01 1.88741758e-01
1.12886894e+00 9.29454267e-02 1.40257388e-01 8.49669755e-01
3.73523623e-01 -2.93909907e-01 -1.30034733e+00 6.49474800e-01
-8.77600685e-02 -1.01937020e+00 -1.76733375e-01 -1.54543817e-01
1.58027023e-01 -1.85743973e-01 -3.76822174e-01 4.44072306e-01
8.62175107e-01 -9.31781769e-01 -3.67258072e-01 5.24022639e-01
3.37650478e-01 -1.87470526e-01 7.65059769e-01 8.04285884e-01
-8.53863418e-01 -1.30493611e-01 -5.92462756e-02 -1.33278653e-01
2.24310443e-01 5.89414954e-01 -1.85035360e+00 3.82867366e-01
4.36012119e-01 3.77934426e-01 -4.61617827e-01 7.16592908e-01
-6.61870480e-01 7.75429249e-01 -2.59939253e-01 -1.02482229e-01
1.76170111e-01 6.30205870e-02 3.03985655e-01 1.03583384e+00
-1.70696765e-01 8.60465646e-01 7.40872622e-01 1.05289721e+00
-2.87015975e-01 1.03681363e-01 -2.57703036e-01 -1.66779712e-01
2.33314738e-01 1.58937919e+00 -6.15556836e-01 -9.31289077e-01
-1.07661776e-01 1.14449573e+00 5.57857335e-01 2.56787032e-01
-4.89307761e-01 -1.35478154e-01 3.51680726e-01 -1.15669310e-01
-6.77329823e-02 3.00747395e-01 -5.72887771e-02 -1.05467856e+00
4.50713411e-02 -8.74813735e-01 8.38010550e-01 -7.98905015e-01
-1.16230536e+00 5.93450427e-01 -5.10658443e-01 -5.96749604e-01
-9.21854258e-01 -6.32117689e-01 -6.39586806e-01 9.08682287e-01
-1.44291091e+00 -6.72170579e-01 -6.39206648e-01 6.71426713e-01
5.16763866e-01 4.70396549e-01 1.27632678e+00 9.28546488e-03
-3.12925607e-01 2.59349942e-01 -8.42035115e-01 5.69326520e-01
9.61713731e-01 -1.47987151e+00 2.00620621e-01 8.60216469e-02
-4.71372515e-01 8.14696312e-01 5.07753313e-01 -1.00955021e+00
-1.18902171e+00 -1.10705912e+00 1.31037605e+00 -7.29904771e-01
3.37051153e-01 -1.10063627e-01 -1.47944343e+00 7.34477580e-01
1.18659250e-01 -5.07136405e-01 9.21119094e-01 1.29945174e-01
-1.39464885e-02 2.98293620e-01 -1.44111788e+00 5.66084802e-01
7.33485758e-01 -4.37797129e-01 -9.88009691e-01 7.59146631e-01
1.19138324e+00 -5.06336570e-01 -9.50760782e-01 3.44476461e-01
1.51822239e-01 -6.29602253e-01 6.87032878e-01 -8.23544085e-01
3.32625389e-01 9.78851691e-02 6.22635305e-01 -1.11623192e+00
-1.74231544e-01 -5.73674679e-01 -3.47428113e-01 6.67775273e-01
6.15279913e-01 -6.77399278e-01 1.01258385e+00 1.44355941e+00
1.19228780e-01 -8.83310139e-01 -2.69281387e-01 -7.97747746e-02
-3.68369430e-01 -1.36763155e-01 4.06563997e-01 1.02654731e+00
6.68877840e-01 7.08578527e-01 -1.44499779e-01 2.65444428e-01
1.05521962e-01 2.87045807e-01 6.26920581e-01 -1.41981316e+00
-4.47257847e-01 -1.44129498e-02 3.64189595e-01 -1.41309083e+00
-2.75621861e-01 -8.84838164e-01 1.24189302e-01 -1.94762039e+00
3.62011105e-01 -3.46979201e-01 -1.96258754e-01 9.92311537e-01
-7.66960919e-01 -2.54722685e-01 -7.85437226e-02 2.21971884e-01
-9.71949697e-01 3.75297926e-02 1.29185104e+00 -4.89620492e-02
-7.21311033e-01 2.35311493e-01 -1.04405689e+00 8.44287872e-01
8.86413753e-01 -4.85135764e-01 -4.19081867e-01 -1.74453735e-01
-4.51685339e-02 8.32255125e-01 2.06661776e-01 -5.40928781e-01
4.25042540e-01 -1.26616374e-01 2.53620118e-01 -6.94335282e-01
6.11104846e-01 -7.99435556e-01 -4.24907416e-01 1.05980659e+00
-7.78556049e-01 -1.78827763e-01 -2.63742562e-02 3.43483180e-01
-7.13088289e-02 -1.14902392e-01 3.07881147e-01 -3.86486650e-01
-2.52660334e-01 1.09747536e-01 -5.20563602e-01 6.75969794e-02
8.03813159e-01 4.67959940e-01 -4.45098221e-01 -7.57280469e-01
-1.36261559e+00 1.04644525e+00 1.11348502e-01 1.93779930e-01
7.56707013e-01 -7.36214519e-01 -5.96055150e-01 -4.12640460e-02
9.47264880e-02 1.36084288e-01 2.40376800e-01 1.06812716e+00
-3.42497766e-01 6.66618764e-01 4.11231905e-01 -5.68474591e-01
-1.61273587e+00 7.81262577e-01 2.32266143e-01 -9.40033138e-01
-6.75506771e-01 6.21664107e-01 2.95126408e-01 -6.83508635e-01
4.58693683e-01 -8.51903439e-01 -4.28661585e-01 -1.70062363e-01
9.74626660e-01 -1.00972608e-01 -1.01984628e-02 2.05500443e-02
-3.32887441e-01 -6.26794100e-02 -4.91186798e-01 -1.62309915e-01
1.08899093e+00 -1.03122167e-01 -1.96667999e-01 1.66539937e-01
8.52738321e-01 -8.46294165e-02 -8.11893582e-01 -5.62035322e-01
2.19653770e-01 5.86314797e-02 -3.77346754e-01 -1.53916407e+00
-7.55983055e-01 6.78095579e-01 2.78157860e-01 5.42710304e-01
1.07214284e+00 3.12628388e-01 1.10332513e+00 1.01163125e+00
-9.36902910e-02 -7.87667394e-01 2.71017194e-01 4.67950821e-01
7.04447031e-01 -1.47782552e+00 -4.21725929e-01 -8.51122260e-01
-9.29187059e-01 9.56710100e-01 5.86208105e-01 2.29698777e-01
4.26002145e-01 1.17642000e-01 6.76664770e-01 -6.39465034e-01
-1.24161029e+00 -2.49932110e-01 3.48534524e-01 3.33347410e-01
3.95979375e-01 1.16821393e-01 -2.06205726e-01 7.54863739e-01
-1.28484190e-01 3.29736710e-01 2.47060433e-01 9.68653202e-01
-7.09906816e-01 -1.39053524e+00 -2.21313581e-01 6.07423782e-01
-5.34034431e-01 -4.94306423e-02 -1.14238262e+00 3.86386245e-01
-2.52520204e-01 1.48215222e+00 -1.69604078e-01 -2.28096902e-01
2.87085861e-01 5.05460799e-01 -1.09440647e-02 -1.05875266e+00
-9.78143036e-01 -5.06823547e-02 5.63650668e-01 -7.57384658e-01
-2.47284666e-01 -5.75863011e-02 -1.83958328e+00 1.41333714e-01
-1.68381352e-02 5.82643211e-01 1.70776367e-01 1.15105855e+00
4.99425352e-01 7.16518879e-01 4.06905234e-01 9.06627029e-02
-6.04025006e-01 -9.91489291e-01 -1.90351792e-02 7.01468825e-01
4.45317149e-01 -4.91149276e-02 -1.04121663e-01 9.04683471e-02] | [12.392280578613281, 8.417745590209961] |
636ecc72-1fc3-46c7-af20-d01128f3a152 | solving-quantitative-reasoning-problems-with | 2206.14858 | null | https://arxiv.org/abs/2206.14858v2 | https://arxiv.org/pdf/2206.14858v2.pdf | Solving Quantitative Reasoning Problems with Language Models | Language models have achieved remarkable performance on a wide range of tasks that require natural language understanding. Nevertheless, state-of-the-art models have generally struggled with tasks that require quantitative reasoning, such as solving mathematics, science, and engineering problems at the college level. To help close this gap, we introduce Minerva, a large language model pretrained on general natural language data and further trained on technical content. The model achieves state-of-the-art performance on technical benchmarks without the use of external tools. We also evaluate our model on over two hundred undergraduate-level problems in physics, biology, chemistry, economics, and other sciences that require quantitative reasoning, and find that the model can correctly answer nearly a third of them. | ['Vedant Misra', 'Guy Gur-Ari', 'Behnam Neyshabur', 'Yuhuai Wu', 'Theo Gutman-Solo', 'Imanol Schlag', 'Cem Anil', 'Ambrose Slone', 'Vinay Ramasesh', 'Henryk Michalewski', 'Ethan Dyer', 'David Dohan', 'Anders Andreassen', 'Aitor Lewkowycz'] | 2022-06-29 | null | null | null | null | ['math-word-problem-solving', 'multi-task-language-understanding', 'arithmetic-reasoning', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'methodology', 'reasoning', 'reasoning', 'time-series'] | [-4.51947033e-01 2.80891240e-01 -2.26435393e-01 -2.72948951e-01
-9.08521533e-01 -8.34543765e-01 5.04064918e-01 6.16830826e-01
-4.19253558e-01 6.80169582e-01 -1.87248796e-01 -1.13374662e+00
-7.40741491e-02 -1.13055968e+00 -7.87057161e-01 2.37430975e-01
1.18746869e-01 4.85263377e-01 7.46148825e-02 -4.37798053e-01
5.23318708e-01 5.95264733e-01 -1.07648993e+00 3.05822402e-01
1.05380511e+00 5.28753936e-01 -3.07531625e-01 6.80920899e-01
-8.05742383e-01 1.31659400e+00 -5.32700777e-01 -6.37198985e-01
-8.86373743e-02 6.41124323e-02 -1.27251041e+00 -5.67188144e-01
6.60728574e-01 -2.43518390e-02 -2.94597626e-01 9.50681329e-01
2.26846159e-01 1.33614987e-01 4.72109944e-01 -9.90137100e-01
-9.79305625e-01 6.23424768e-01 -4.43799198e-01 2.81096250e-01
6.48160040e-01 1.17984451e-01 1.19103050e+00 -1.03695858e+00
3.70610446e-01 1.53835440e+00 5.94964325e-01 5.30281067e-01
-1.10835648e+00 -6.87267721e-01 4.27485347e-01 7.14933500e-02
-1.25553882e+00 1.23714749e-02 5.96077323e-01 -6.10534966e-01
1.01367593e+00 -2.05633894e-01 3.59007269e-01 6.16440654e-01
4.64821815e-01 6.71418786e-01 1.11372268e+00 -2.74659485e-01
4.07619849e-02 1.32394075e-01 4.43936169e-01 1.35126388e+00
1.83279008e-01 -3.75172436e-01 -4.91967380e-01 2.24454049e-02
5.97136676e-01 -9.66400281e-02 1.53071344e-01 -1.36124775e-01
-1.27031493e+00 8.65822673e-01 2.01571196e-01 4.65809017e-01
2.34572187e-01 2.66675770e-01 1.03196360e-01 6.74046814e-01
1.81651518e-01 9.59299922e-01 -8.77226889e-01 -1.61579415e-01
-8.55896413e-01 2.17590228e-01 1.25721729e+00 9.15002823e-01
7.48034358e-01 -5.58578372e-02 7.02478271e-03 4.48968112e-01
2.73471981e-01 2.80131072e-01 2.77658224e-01 -1.14764106e+00
6.76737189e-01 1.07319403e+00 -1.74350888e-01 -8.94790471e-01
-2.30808049e-01 -5.22734404e-01 -7.88270652e-01 1.02393351e-01
5.11119425e-01 2.92588901e-02 -7.38144457e-01 1.53161883e+00
3.43881198e-03 2.89510131e-01 5.68411052e-02 5.29273808e-01
1.40546346e+00 7.22406507e-01 3.31333309e-01 1.80344626e-01
1.22460961e+00 -1.09109485e+00 -3.27321261e-01 -6.12851083e-01
9.71202374e-01 -6.67223990e-01 1.28230107e+00 7.37446785e-01
-1.46679509e+00 -4.89662468e-01 -7.83753455e-01 -7.30657935e-01
-7.89851844e-01 -1.03186548e-01 8.60338569e-01 4.26952630e-01
-1.15886867e+00 5.91767192e-01 -4.29465175e-01 -1.08936116e-01
1.34095222e-01 3.00585181e-01 -4.50004160e-01 -2.79826969e-01
-1.17240965e+00 1.02659762e+00 1.33766428e-01 -1.66995794e-01
-7.69790173e-01 -1.37838328e+00 -1.13928747e+00 4.07931179e-01
4.53700215e-01 -5.67081928e-01 1.43045509e+00 1.21773608e-01
-1.46220767e+00 1.33224547e+00 -1.13999672e-01 -5.74799478e-02
4.76432711e-01 -2.02604428e-01 -1.30976915e-01 -5.99533319e-02
3.17121409e-02 4.11030442e-01 9.94223282e-02 -7.74179041e-01
-6.35476932e-02 -6.23570383e-02 6.27785623e-01 -2.90602863e-01
-4.25481349e-01 -8.39332342e-02 -4.40125108e-01 -3.52799147e-01
1.15517549e-01 -6.97031975e-01 -3.66042614e-01 2.75006354e-01
-1.70700416e-01 -7.79386640e-01 3.94859761e-01 -3.31324250e-01
1.20914352e+00 -1.72681415e+00 2.60623574e-01 1.44074991e-01
5.62462389e-01 2.02125490e-01 -3.68921876e-01 1.39064327e-01
-1.54170215e-01 7.35942304e-01 2.83967657e-03 -1.55576110e-01
3.30565780e-01 -7.90199563e-02 -5.51415324e-01 7.98412859e-02
3.26947540e-01 1.30749083e+00 -1.15623438e+00 -6.98117971e-01
1.38944745e-01 -2.29583830e-02 -6.74120188e-01 1.97684616e-01
-5.54854691e-01 2.24078670e-01 -6.40933812e-01 5.72032630e-01
9.97790322e-02 -6.95104599e-01 2.44334459e-01 4.99533176e-01
3.62931401e-03 7.47964382e-01 -9.76867080e-01 1.93423629e+00
-9.09391761e-01 5.55343509e-01 1.87387601e-01 -1.16976237e+00
8.98411870e-01 1.44383952e-01 3.64963681e-01 -5.70374548e-01
-3.13557871e-02 1.60619721e-01 9.50294510e-02 -3.63627583e-01
1.35121435e-01 -7.20903650e-02 -3.45322967e-01 7.23513663e-01
7.10894838e-02 -9.61562514e-01 4.88205880e-01 4.84120190e-01
1.26130962e+00 5.42783998e-02 2.12005973e-01 -4.84603226e-01
9.18161035e-01 -1.25353262e-01 2.71110386e-01 8.92896295e-01
-8.30828696e-02 -8.39198455e-02 5.21654427e-01 -7.21155286e-01
-8.13801825e-01 -1.04947007e+00 2.19487146e-01 1.47211063e+00
-9.57725719e-02 -7.52387762e-01 -4.11329865e-01 -5.13265908e-01
2.39697639e-02 5.78042746e-01 -3.43109637e-01 -1.97519958e-01
-5.95402539e-01 -2.63374239e-01 5.60997546e-01 5.61334133e-01
4.84674066e-01 -9.58510518e-01 -2.75870960e-04 1.84448645e-01
-3.41395624e-02 -1.54467165e+00 1.33459672e-01 -3.25246118e-02
-8.76798987e-01 -1.02843237e+00 -2.81682581e-01 -9.85700667e-01
5.68535388e-01 4.17044573e-02 1.79774761e+00 7.07388520e-01
-6.06267810e-01 6.44733012e-01 2.34313145e-01 -4.63639587e-01
-4.00674790e-01 5.32609761e-01 -1.50819331e-01 -1.00489485e+00
3.89104396e-01 -6.47040784e-01 -4.38220566e-03 -1.86387315e-01
-7.45777667e-01 -6.61107525e-02 4.62960809e-01 4.31081861e-01
1.67172030e-01 -5.30712269e-02 3.55216593e-01 -8.23038101e-01
7.52842247e-01 -3.51266682e-01 -5.68759501e-01 6.06894672e-01
-4.71562088e-01 4.25030559e-01 9.38964307e-01 -3.35046798e-01
-5.33657014e-01 -4.21125591e-01 -1.32010087e-01 -7.00163767e-02
1.61328353e-02 8.59216094e-01 8.10119212e-02 -4.26168919e-01
6.67071640e-01 1.25129387e-01 -6.45266414e-01 -2.59074092e-01
1.85015142e-01 7.62322471e-02 4.99108076e-01 -1.56366014e+00
1.23616052e+00 -7.02401325e-02 3.42891961e-01 -6.17014468e-01
-1.73236942e+00 -3.11936378e-01 -4.45252389e-01 7.11165592e-02
6.37059152e-01 -1.01734507e+00 -1.24087143e+00 1.98167935e-02
-1.35793293e+00 -7.45296955e-01 1.58835873e-01 3.05431426e-01
-3.39423686e-01 3.45625043e-01 -1.00072181e+00 -6.34483635e-01
-3.46327305e-01 -9.91204321e-01 1.09044743e+00 2.46959031e-01
-3.26795131e-01 -1.47623622e+00 -2.65645415e-01 5.85093141e-01
3.96230936e-01 2.71456689e-02 1.40521252e+00 -4.29912657e-01
-8.88867378e-01 -1.39964759e-01 -2.26232454e-01 4.97776791e-02
-2.99213201e-01 9.54412743e-02 -6.84267282e-01 -5.27379625e-02
-3.21661264e-01 -8.86341035e-01 8.41339111e-01 -2.35202745e-01
1.88293576e+00 2.31218748e-02 -6.56762570e-02 3.50947201e-01
1.03716743e+00 -4.94153798e-01 3.92725497e-01 2.75609225e-01
5.53463638e-01 4.85301018e-01 3.15578550e-01 -2.52038315e-02
6.31914675e-01 7.50124902e-02 1.10009484e-01 5.25087379e-02
2.76264638e-01 -4.04135108e-01 2.58099914e-01 8.82085562e-01
2.81668827e-02 4.92489748e-02 -1.37222505e+00 6.36079311e-01
-1.63253236e+00 -7.72627652e-01 -9.96744260e-02 1.60943913e+00
1.30574083e+00 6.96481824e-01 -3.76662344e-01 8.36593434e-02
8.27788860e-02 -2.25451514e-02 -3.47783417e-01 -7.49985576e-01
2.85378516e-01 8.82375479e-01 2.86151916e-02 7.07603693e-01
-7.93745100e-01 1.44037664e+00 7.24723959e+00 7.72593021e-01
-1.11033678e+00 -2.76251733e-01 5.89280069e-01 3.49467486e-01
-4.88432318e-01 -1.48141503e-01 -7.89988756e-01 -1.35822341e-01
1.03987670e+00 -4.89476830e-01 5.33235788e-01 9.09325957e-01
5.54582011e-03 1.59360871e-01 -1.36981022e+00 7.48668075e-01
-2.81980842e-01 -1.42037380e+00 1.47858709e-01 -9.62516665e-02
8.71364534e-01 -3.62288296e-01 1.46284088e-01 9.30463672e-01
9.39098299e-01 -1.64181435e+00 3.15197349e-01 4.85668987e-01
6.95111215e-01 -4.12463635e-01 4.68070626e-01 7.89301395e-01
-1.18258631e+00 5.02366200e-02 -1.73970789e-01 -7.40716636e-01
-4.52874362e-01 7.01508462e-01 -6.99001849e-01 2.39876613e-01
5.43370366e-01 5.50462425e-01 -6.98671997e-01 6.33164883e-01
-7.68897712e-01 7.59484649e-01 -2.54524082e-01 -3.12148601e-01
3.66673112e-01 -1.73707381e-02 -1.11029699e-01 1.25457704e+00
1.71477392e-01 5.03453553e-01 6.29119217e-01 1.28329575e+00
-5.44107139e-01 -1.39129069e-02 -3.91845584e-01 -6.87137783e-01
3.33624989e-01 1.36429262e+00 -4.63474005e-01 -4.80503589e-01
-4.51259822e-01 4.04025376e-01 7.45159626e-01 2.80027449e-01
-6.97761297e-01 -4.03405935e-01 5.65799534e-01 1.19513169e-01
-2.95994818e-01 -8.79539192e-01 -7.38307774e-01 -1.30117047e+00
-2.32077446e-02 -9.52972591e-01 3.25177699e-01 -8.84187639e-01
-1.32489717e+00 -4.32916135e-02 -2.23806888e-01 -4.20651317e-01
-1.67614281e-01 -1.24347472e+00 -7.58564353e-01 1.06787777e+00
-1.81799662e+00 -9.71862495e-01 -4.79305118e-01 4.68487382e-01
3.57386410e-01 -2.17187300e-01 9.57539558e-01 2.38493755e-01
-3.52263391e-01 4.94478047e-01 -6.07321799e-01 5.97955406e-01
6.13917351e-01 -1.42075312e+00 6.55766845e-01 4.38445479e-01
-7.89007619e-02 1.30844080e+00 5.19483268e-01 -3.35537016e-01
-1.72892416e+00 -9.39457595e-01 1.45830584e+00 -7.58921385e-01
1.27434850e+00 -3.69904578e-01 -8.51647198e-01 7.56228924e-01
-1.99539408e-01 1.78058237e-01 7.40541399e-01 3.56943995e-01
-8.33354056e-01 1.92519933e-01 -8.58213484e-01 6.53208911e-01
1.08177137e+00 -1.03403544e+00 -9.84836400e-01 6.03476882e-01
9.07843947e-01 -6.83351755e-01 -1.11677015e+00 2.97493070e-01
3.72810870e-01 -3.08932364e-01 1.21584916e+00 -1.32607651e+00
1.17105627e+00 9.36137512e-03 2.29564205e-01 -9.09872055e-01
-2.80512542e-01 -6.73561633e-01 -3.69527996e-01 8.43364894e-01
4.71020103e-01 -2.27553248e-01 8.04203033e-01 7.92658508e-01
-9.52875391e-02 -8.69392276e-01 -1.82249039e-01 -6.27622604e-01
8.49408984e-01 -5.19996107e-01 3.35176915e-01 1.16918910e+00
2.32698664e-01 6.98733926e-01 2.41072133e-01 -6.55549467e-02
3.91770869e-01 5.14860213e-01 8.93242300e-01 -1.81227636e+00
-2.06888765e-01 -7.31090188e-01 -1.83097199e-01 -1.13899195e+00
1.03745377e+00 -1.01320827e+00 -1.13975935e-01 -1.89947224e+00
6.37057573e-02 -2.19189480e-01 -4.30642627e-02 7.64305472e-01
-2.12129474e-01 2.29775310e-01 6.22094348e-02 -2.91108549e-01
-8.64201069e-01 2.28118792e-01 1.51174712e+00 -4.98243213e-01
2.76052862e-01 -2.18024850e-01 -1.11212981e+00 1.00105000e+00
6.75585330e-01 -1.16471671e-01 -1.51407391e-01 -6.13069534e-01
8.05292189e-01 -3.58638056e-02 3.67637187e-01 -9.85569477e-01
4.27236080e-01 -8.29379201e-01 2.97072709e-01 -6.40891045e-02
1.81023329e-01 -5.84616423e-01 -7.13393152e-01 5.19683242e-01
-6.96702063e-01 1.79702625e-01 4.65958923e-01 2.45423820e-02
-2.10163966e-01 -2.57131785e-01 6.18449926e-01 -5.73409557e-01
-7.04634726e-01 3.27283353e-01 -3.21521074e-01 5.30761600e-01
7.78942227e-01 4.36210126e-01 -5.59052706e-01 -4.15674150e-01
-5.39859533e-01 6.87094092e-01 6.46858364e-02 4.72432464e-01
3.99768382e-01 -8.95934343e-01 -6.43841267e-01 -3.04968655e-01
-6.21339306e-02 3.28019768e-01 -1.82625726e-01 4.53157663e-01
-8.19245517e-01 8.74042571e-01 1.52342796e-01 -1.33265778e-01
-9.45303261e-01 5.32562792e-01 3.43030155e-01 -9.08687413e-01
-1.68281421e-01 7.61800647e-01 1.05256494e-02 -1.10699832e+00
3.46036434e-01 -7.48272598e-01 6.01276197e-02 -5.15420616e-01
5.85842907e-01 1.42036423e-01 1.82466328e-01 2.15378061e-01
-3.75577748e-01 6.30111992e-01 1.44403562e-01 1.60133407e-01
1.34339523e+00 4.28691775e-01 -3.29336077e-01 5.46138644e-01
9.28624392e-01 1.68283463e-01 -3.36541146e-01 -4.14224684e-01
3.31670851e-01 1.61865234e-01 3.11205704e-02 -9.38787818e-01
-6.69534266e-01 1.40239632e+00 -2.36008257e-01 2.17371628e-01
6.22807443e-01 -1.48181185e-01 7.90480673e-01 1.33344698e+00
5.78485847e-01 -8.63179326e-01 4.25147951e-01 1.25005102e+00
6.23200357e-01 -1.40986848e+00 2.80983269e-01 -7.26484537e-01
7.07331002e-02 1.40958846e+00 8.90631318e-01 -6.33082017e-02
6.49504960e-01 5.95961630e-01 -2.42175490e-01 -2.02654153e-01
-1.04328287e+00 -9.03231278e-02 3.68106395e-01 9.59053822e-03
1.15992045e+00 -7.96786323e-02 -8.75237510e-02 6.46981776e-01
-6.04217708e-01 3.24694365e-01 5.11431456e-01 1.15580845e+00
-5.79099119e-01 -1.08731747e+00 -2.53364295e-01 7.37162381e-02
-6.03033423e-01 -4.47267205e-01 -8.99592102e-01 5.46024978e-01
-1.54210791e-01 8.63328457e-01 -2.10745901e-01 -1.09403357e-02
1.24224596e-01 3.37810278e-01 7.39763677e-01 -9.74180579e-01
-7.73948729e-01 -8.22060764e-01 -1.53822675e-01 -3.81201953e-01
-1.22919843e-01 1.43146560e-01 -1.68332791e+00 -9.14931238e-01
3.60723048e-01 3.89290661e-01 5.15901923e-01 1.31918526e+00
1.62960812e-01 6.74090326e-01 -8.77545998e-02 -1.78918377e-01
-8.63325715e-01 -7.57659793e-01 -1.06300481e-01 1.25645921e-01
2.01144591e-01 -3.34359258e-01 -1.86308548e-01 4.42016684e-02] | [9.622613906860352, 7.311420917510986] |
b8b461dc-19d1-4705-9e6d-e6feeb49f8db | asr-based-features-for-emotion-recognition-a | 1805.09197 | null | http://arxiv.org/abs/1805.09197v3 | http://arxiv.org/pdf/1805.09197v3.pdf | ASR-based Features for Emotion Recognition: A Transfer Learning Approach | During the last decade, the applications of signal processing have
drastically improved with deep learning. However areas of affecting computing
such as emotional speech synthesis or emotion recognition from spoken language
remains challenging. In this paper, we investigate the use of a neural
Automatic Speech Recognition (ASR) as a feature extractor for emotion
recognition. We show that these features outperform the eGeMAPS feature set to
predict the valence and arousal emotional dimensions, which means that the
audio-to-text mapping learning by the ASR system contain information related to
the emotional dimensions in spontaneous speech. We also examine the
relationship between first layers (closer to speech) and last layers (closer to
text) of the ASR and valence/arousal. | ['Noé Tits', 'Thierry Dutoit', 'Kevin El Haddad'] | 2018-05-23 | asr-based-features-for-emotion-recognition-a-1 | https://aclanthology.org/W18-3307 | https://aclanthology.org/W18-3307.pdf | ws-2018-7 | ['emotional-speech-synthesis'] | ['speech'] | [ 3.24629396e-02 1.19789734e-01 2.10388109e-01 -9.16234493e-01
-5.79596877e-01 -2.83569664e-01 4.96758521e-01 -9.19168144e-02
-2.41936579e-01 6.03275299e-01 4.42861527e-01 2.27258489e-01
5.18612675e-02 -3.17293882e-01 -2.51744300e-01 -5.84963202e-01
-2.98376232e-01 -2.05344304e-01 -5.51600635e-01 -5.84478557e-01
-5.98583966e-02 7.08730757e-01 -1.97728419e+00 7.99955904e-01
2.67164320e-01 1.78300059e+00 -1.01824090e-01 1.10624814e+00
-3.17550540e-01 8.37619245e-01 -1.09047997e+00 3.46457399e-02
-2.19095320e-01 -3.87602746e-01 -7.56735921e-01 3.19590271e-02
-1.94267988e-01 1.38516426e-01 -3.35121304e-01 9.81854856e-01
7.68014967e-01 4.43203688e-01 5.34853041e-01 -1.37001705e+00
-5.17528594e-01 5.52475870e-01 -1.29101038e-01 1.99768394e-01
5.41382909e-01 -5.10819256e-01 1.13228512e+00 -1.29959381e+00
3.18570316e-01 1.33960295e+00 3.99795383e-01 7.01846480e-01
-7.33342767e-01 -6.06314182e-01 2.21539512e-02 3.50397408e-01
-1.34944546e+00 -9.57920969e-01 1.18192315e+00 -1.37167871e-01
1.62974405e+00 6.16288483e-01 7.58001447e-01 1.41810596e+00
1.91877380e-01 1.02347589e+00 1.07514679e+00 -3.37762505e-01
3.00829113e-01 5.75654447e-01 4.45587933e-02 1.61245614e-01
-8.49030793e-01 1.98162690e-01 -8.61328602e-01 -2.74227653e-02
2.02549294e-01 -4.45250064e-01 -2.14535862e-01 4.90489930e-01
-6.81830585e-01 8.56579721e-01 2.32402608e-01 5.92071652e-01
-8.11103463e-01 2.24989969e-02 7.91992426e-01 8.79766822e-01
7.09116697e-01 6.96067631e-01 -6.92078590e-01 -7.01534986e-01
-9.99312401e-01 -2.89115697e-01 8.99631739e-01 5.06072938e-01
5.20787835e-01 7.44011402e-01 6.71546906e-02 1.38523114e+00
3.08620840e-01 4.47219312e-01 8.44440043e-01 -8.22880447e-01
-1.25290498e-01 2.76970983e-01 -3.40867996e-01 -1.03155839e+00
-6.38273001e-01 -2.22748071e-01 -9.52638865e-01 3.30529176e-02
-4.40007657e-01 -4.31944251e-01 -6.97817624e-01 1.61176753e+00
-5.41118272e-02 -1.59148172e-01 7.22206533e-01 8.96064520e-01
1.12622356e+00 1.31780291e+00 -1.08003661e-01 -4.63516384e-01
1.18389237e+00 -4.62122500e-01 -1.27392459e+00 -4.82462794e-01
3.45139742e-01 -6.72726810e-01 9.53732371e-01 7.08892524e-01
-1.16288435e+00 -5.77374756e-01 -9.79754031e-01 7.63725191e-02
-7.99686134e-01 2.17433929e-01 6.69896185e-01 5.83675921e-01
-1.19864893e+00 3.67688417e-01 -5.03631651e-01 -2.78927803e-01
-1.51225133e-02 4.46197569e-01 -3.74162376e-01 7.82590330e-01
-1.78771889e+00 9.36936259e-01 2.37380922e-01 1.99729472e-01
-4.73442137e-01 -2.57511497e-01 -9.83554661e-01 3.32544237e-01
-2.36336663e-01 1.44162402e-01 1.22207963e+00 -1.74411416e+00
-2.07784033e+00 7.82241046e-01 -1.84852913e-01 -2.84720749e-01
-4.49826956e-01 -8.98737758e-02 -1.00333238e+00 3.03191602e-01
-7.44251251e-01 7.80380487e-01 8.86135757e-01 -9.44233418e-01
-3.39486808e-01 -4.79113698e-01 -5.38074493e-01 1.92557052e-01
-6.66650116e-01 5.10303855e-01 2.12088734e-01 -6.37837470e-01
-4.19937707e-02 -5.80567181e-01 1.22992121e-01 -4.38170999e-01
-1.54012501e-01 -3.67837399e-01 8.50177228e-01 -6.92368329e-01
1.26961291e+00 -2.46804762e+00 1.07661039e-01 4.53489304e-01
-2.57148117e-01 1.18346982e-01 -2.77791709e-01 5.44670165e-01
-7.32248127e-01 -4.40628752e-02 1.84554011e-01 -1.69200793e-01
2.34621704e-01 3.03951502e-01 -6.47892118e-01 1.85701758e-01
5.72080255e-01 8.33859921e-01 -4.48737830e-01 -2.20123753e-01
1.25252992e-01 9.47551787e-01 -2.64002502e-01 5.17436504e-01
9.97598022e-02 8.68808664e-03 -9.79479924e-02 4.65854764e-01
2.44233519e-01 3.84554803e-01 -2.39145875e-01 -9.11984965e-02
-1.75005391e-01 5.99656761e-01 -8.38060141e-01 1.43514931e+00
-7.30737567e-01 1.22085142e+00 4.26250041e-01 -1.06034136e+00
1.63044131e+00 9.20066595e-01 4.35513645e-01 -8.77423406e-01
4.97742087e-01 7.87803382e-02 1.12784192e-01 -5.96115708e-01
4.92169350e-01 -4.40631479e-01 -2.37937346e-01 1.88982189e-01
3.79484177e-01 -2.96021134e-01 -3.54148865e-01 -1.27955928e-01
6.58308685e-01 -6.47579610e-01 3.03453952e-01 7.57762277e-03
6.19552016e-01 -5.97909629e-01 3.35729212e-01 1.37599215e-01
-2.91837096e-01 2.99716055e-01 5.40254176e-01 -1.60789326e-01
-6.55675828e-01 -7.91416705e-01 -8.75480622e-02 1.34398520e+00
-4.55050051e-01 -5.20882428e-01 -6.21012092e-01 -1.59477949e-01
-4.10241932e-01 8.58015060e-01 -6.39572144e-01 -6.48430765e-01
4.25420440e-04 -4.34167176e-01 8.69401872e-01 7.27232337e-01
-1.09069966e-01 -1.48062611e+00 -4.89354819e-01 1.99507773e-01
-2.00718209e-01 -1.12301719e+00 -5.19024245e-02 7.82144964e-01
-4.42973673e-01 -5.63187823e-02 -4.25430924e-01 -7.22039759e-01
6.71289116e-02 -4.59024012e-01 9.29067671e-01 -3.80352199e-01
-2.78784543e-01 6.32434011e-01 -5.00952780e-01 -6.40563667e-01
-5.73538899e-01 -2.31516108e-01 2.91177273e-01 3.29883397e-01
5.78085899e-01 -8.51632953e-01 -2.29347497e-01 6.96775457e-03
-8.39561701e-01 -3.60610753e-01 3.98603320e-01 6.75760269e-01
3.61862898e-01 6.58139735e-02 1.12519276e+00 -5.80351762e-02
1.16971231e+00 -2.55684853e-01 1.64607093e-02 -2.30424777e-02
-2.43504390e-01 1.78313497e-02 7.51522660e-01 -5.87717950e-01
-1.09209859e+00 1.42767474e-01 -7.33255386e-01 -5.79540908e-01
-4.90945727e-01 6.67420149e-01 -2.96365976e-01 2.09324345e-01
4.95038658e-01 2.14655727e-01 -1.07136630e-01 -6.57589287e-02
2.20168531e-01 1.69962358e+00 4.32170779e-01 -2.86138475e-01
4.11973298e-02 -2.86903847e-02 -3.46575916e-01 -1.59475923e+00
-5.78303754e-01 -3.25882018e-01 -2.66831517e-01 -5.63299417e-01
7.73011208e-01 -9.59379196e-01 -8.11242938e-01 1.80137530e-01
-1.22671640e+00 4.13718168e-03 -4.54214573e-01 7.27058887e-01
-7.26172924e-01 5.20171830e-03 -8.90151978e-01 -1.40136063e+00
-7.43935525e-01 -8.11876833e-01 1.29285741e+00 1.38885945e-01
-8.29996169e-01 -7.65198410e-01 8.46318975e-02 -1.49856806e-01
5.95402777e-01 6.45274147e-02 6.49875641e-01 -9.34159279e-01
5.50370276e-01 -3.24950695e-01 2.03660473e-01 7.98144519e-01
-2.34427527e-02 1.48976892e-01 -1.55913544e+00 8.93301815e-02
3.98090720e-01 -6.65267169e-01 6.03853047e-01 4.12959039e-01
1.24400127e+00 -2.95436561e-01 3.36129785e-01 4.99606460e-01
6.54838085e-01 6.90669477e-01 6.90143287e-01 -4.75735441e-02
5.29652350e-02 9.01431620e-01 5.54485917e-01 7.12658286e-01
6.25811517e-02 3.57251287e-01 1.23697110e-01 -1.83626845e-01
3.02631795e-01 1.27112865e-01 9.48028982e-01 1.47113204e+00
3.70184571e-01 -1.62680224e-01 -7.58088589e-01 4.26221073e-01
-1.40933657e+00 -9.81433332e-01 2.50033468e-01 1.63841319e+00
9.53877568e-01 -9.82512757e-02 -3.20988940e-03 5.12902975e-01
4.31465477e-01 2.51592189e-01 -3.65318924e-01 -1.58453429e+00
-3.48853648e-01 3.76408458e-01 -3.23133469e-01 4.46954101e-01
-9.77685690e-01 8.44865263e-01 6.95047712e+00 6.71745718e-01
-1.54151630e+00 -3.38034749e-01 6.88374102e-01 -3.70068729e-01
-1.17816590e-01 -7.68226743e-01 -3.14831495e-01 1.14486225e-01
1.72123063e+00 -1.93456963e-01 8.75310540e-01 1.01670098e+00
4.01986301e-01 3.01546544e-01 -1.32759178e+00 1.39363515e+00
3.06645602e-01 -4.83560205e-01 -1.05900794e-01 -2.67212093e-01
3.14506173e-01 2.23067254e-02 3.50224853e-01 6.31078839e-01
-3.06392044e-01 -1.29238939e+00 5.70236444e-01 5.06048560e-01
6.82765186e-01 -1.25892293e+00 6.52622342e-01 2.03645691e-01
-1.05766439e+00 -2.48358935e-01 -3.03774148e-01 -1.90368280e-01
2.73661013e-03 5.07626891e-01 -8.93442512e-01 2.47313431e-03
7.56710172e-01 5.49882710e-01 -1.33026153e-01 2.48581409e-01
-4.29539531e-02 6.36072516e-01 -2.42489204e-01 -3.63664418e-01
3.10187161e-01 -2.90082376e-02 5.48443079e-01 1.71176839e+00
4.12760586e-01 2.57425159e-01 -2.57850379e-01 7.13921666e-01
-2.37042785e-01 3.74142587e-01 -6.90511644e-01 -8.14858437e-01
2.41474360e-01 1.46477520e+00 -4.49973941e-01 -3.56620938e-01
-1.60968274e-01 9.98778462e-01 -4.32685427e-02 4.74107474e-01
-6.59931719e-01 -8.03096354e-01 1.13301432e+00 -6.08197391e-01
8.23276117e-02 -1.64962281e-02 -1.69891819e-01 -8.88811886e-01
-1.08339787e-01 -9.82249439e-01 2.45895013e-02 -1.20921838e+00
-1.24979937e+00 1.13117611e+00 -4.76530701e-01 -9.53283370e-01
-6.67505860e-01 -7.68585145e-01 -5.95923066e-01 8.16004694e-01
-1.31892729e+00 -4.01088208e-01 1.29011711e-02 7.47157156e-01
6.42769396e-01 -3.36075246e-01 1.22737992e+00 9.68534276e-02
-3.62201393e-01 5.05222976e-01 -1.08631544e-01 2.37548292e-01
5.64946771e-01 -1.15099347e+00 9.38515961e-02 3.02574724e-01
1.95690587e-01 3.93275708e-01 7.30823636e-01 -1.47284893e-02
-1.41339087e+00 -8.22265208e-01 9.82121587e-01 -4.73566204e-02
7.38056242e-01 -8.63272727e-01 -9.26034689e-01 3.96010965e-01
5.66456497e-01 -1.71494707e-01 1.06739151e+00 2.05424026e-01
-2.75642723e-01 -2.99771905e-01 -9.77358103e-01 3.53903562e-01
3.48301232e-01 -1.21515667e+00 -8.37450624e-01 -1.32241607e-01
7.43695319e-01 -1.83727611e-02 -1.12918520e+00 2.96898216e-01
5.87693155e-01 -6.54090285e-01 9.01773214e-01 -8.34532022e-01
4.72094655e-01 2.27783307e-01 -6.83502734e-01 -1.77046728e+00
2.90731620e-02 -6.01419866e-01 -2.19767466e-01 1.26712883e+00
4.44103688e-01 -4.22233164e-01 3.82177114e-01 4.59956497e-01
-1.65884897e-01 -7.56722748e-01 -1.06426275e+00 -5.40283918e-01
-1.88009962e-01 -7.25202084e-01 3.88656408e-01 1.03931201e+00
6.73885584e-01 6.94574773e-01 -3.69273096e-01 5.61955720e-02
-2.32161120e-01 3.57536040e-02 5.04489727e-02 -1.37851930e+00
-4.52816933e-02 -5.35263300e-01 -5.42046547e-01 -6.15084112e-01
6.03426814e-01 -8.94324243e-01 3.31775546e-01 -1.20040894e+00
-2.97465026e-01 4.33273256e-01 -6.18161738e-01 4.44910288e-01
3.12323302e-01 -6.22902177e-02 4.21921313e-02 -6.90730810e-01
-4.81944144e-01 1.06875288e+00 6.12562060e-01 -6.62009120e-02
-3.57548863e-01 -1.31601244e-01 -5.55849433e-01 7.40782082e-01
8.94962966e-01 -2.77605981e-01 -2.82201171e-01 1.88122004e-01
5.17895818e-01 4.26771790e-01 -5.98676465e-02 -8.42201591e-01
1.30662337e-01 2.51614422e-01 6.65762901e-01 -4.98129100e-01
1.08292496e+00 -9.63703513e-01 -2.46363372e-01 -1.85902059e-01
-8.14837039e-01 -1.13413483e-01 4.46832716e-01 3.22397128e-02
-9.18471217e-01 3.09206545e-03 7.11298704e-01 1.42201617e-01
-5.10383964e-01 -1.39478683e-01 -1.05827487e+00 -1.82843328e-01
6.03171468e-01 -1.48545608e-01 5.12070358e-02 -9.61080670e-01
-1.07492447e+00 6.19135518e-03 -2.86471575e-01 8.01656783e-01
1.09289193e+00 -1.40139341e+00 -5.07293284e-01 4.95952249e-01
-1.17891654e-02 -6.30954742e-01 2.10148171e-01 9.40267801e-01
4.07488108e-01 5.16131818e-01 -2.98472792e-01 -3.97380739e-01
-1.51547194e+00 3.70346963e-01 5.35525858e-01 2.41037652e-01
-8.60857517e-02 8.77807081e-01 8.45444649e-02 -4.21098083e-01
6.30132079e-01 -2.78352976e-01 -6.02217376e-01 6.64297044e-01
7.32094526e-01 1.95964739e-01 2.65883982e-01 -8.69599462e-01
-4.58046347e-01 1.05531611e-01 1.44129544e-01 -4.51543152e-01
1.80012012e+00 -1.96860865e-01 -2.20754683e-01 1.18323410e+00
1.65615690e+00 -2.71345705e-01 -5.60038984e-01 8.37224647e-02
1.57203808e-01 7.79987276e-02 5.09090841e-01 -9.88491058e-01
-1.08680069e+00 1.37031615e+00 7.54838467e-01 5.45476437e-01
1.60646391e+00 3.31321135e-02 6.82709336e-01 8.38232934e-01
-2.29423121e-01 -1.85078609e+00 1.11342572e-01 8.90671074e-01
1.27112782e+00 -1.05552840e+00 -7.25316167e-01 1.83791354e-01
-1.16391861e+00 1.51557362e+00 3.12574893e-01 1.40681416e-01
8.07442725e-01 6.49688840e-01 3.59797657e-01 -2.73392200e-01
-1.28905082e+00 -1.05495214e-01 5.13619363e-01 4.25548345e-01
8.26365590e-01 5.71863391e-02 1.28335297e-01 1.24458051e+00
-6.76170290e-01 -3.54323715e-01 3.51236224e-01 6.02567315e-01
-5.45888782e-01 -6.25576854e-01 -4.34917569e-01 3.10528368e-01
-6.76480055e-01 -5.12272157e-02 -1.19136381e+00 2.40123242e-01
-4.53998387e-01 1.21965694e+00 2.55996197e-01 -7.85206199e-01
4.28302795e-01 7.34166324e-01 4.92136963e-02 -4.45762098e-01
-8.57427120e-01 4.03728366e-01 3.73416454e-01 -6.88968480e-01
-3.78688306e-01 -4.70968455e-01 -1.62051320e+00 3.27061236e-01
-2.47905225e-01 3.52786243e-01 1.13414228e+00 7.73588538e-01
3.88149172e-01 7.38271117e-01 1.14069247e+00 -7.89289296e-01
-3.16842705e-01 -1.16748250e+00 -9.80621576e-01 1.45591199e-01
5.51846087e-01 -2.11229682e-01 -6.50705516e-01 -1.52815878e-01] | [13.628520965576172, 5.720970153808594] |
cd3a7a5c-8b45-44a3-83d5-c80b1d8ae5cc | non-local-intrinsic-decomposition-with-near | null | null | http://openaccess.thecvf.com/content_ICCV_2019/html/Cheng_Non-Local_Intrinsic_Decomposition_With_Near-Infrared_Priors_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Cheng_Non-Local_Intrinsic_Decomposition_With_Near-Infrared_Priors_ICCV_2019_paper.pdf | Non-Local Intrinsic Decomposition With Near-Infrared Priors | Intrinsic image decomposition is a highly under-constrained problem that has been extensively studied by computer vision researchers. Previous methods impose additional constraints by exploiting either empirical or data-driven priors. In this paper, we revisit intrinsic image decomposition with the aid of near-infrared (NIR) imagery. We show that NIR band is considerably less sensitive to textures and can be exploited to reduce ambiguity caused by reflectance variation, promoting a simple yet powerful prior for shading smoothness. With this observation, we formulate intrinsic decomposition as an energy minimisation problem. Unlike existing methods, our energy formulation decouples reflectance and shading estimation, into a convex local shading component based on NIR-RGB image pair, and a reflectance component that encourages reflectance homogeneity both locally and globally. We further show the minimisation process can be approached by a series of multi-dimensional kernel convolutions, each within linear time complexity. To validate the proposed algorithm, a NIR-RGB dataset is captured over real-world objects, where our NIR-assisted approach demonstrates clear superiority over RGB methods.
| [' Imari Sato', ' Shaodi You', ' Yinqiang Zheng', 'Ziang Cheng'] | 2019-10-01 | null | null | null | iccv-2019-10 | ['intrinsic-image-decomposition'] | ['computer-vision'] | [ 9.87867653e-01 -4.35774811e-02 3.10539067e-01 -4.31013435e-01
-6.26925051e-01 -5.65561771e-01 5.28429389e-01 -4.22822446e-01
-3.40486228e-01 5.55602431e-01 1.90290883e-01 -3.85048464e-02
-1.30463526e-01 -6.39065087e-01 -5.84469259e-01 -1.18693149e+00
4.93169367e-01 -3.22636813e-01 -1.59696206e-01 -8.64243060e-02
2.06607953e-01 5.87630391e-01 -1.60361946e+00 7.58835450e-02
9.63009238e-01 9.61843729e-01 3.59894395e-01 6.04591846e-01
-8.39384738e-03 4.53160524e-01 -8.50381181e-02 2.54499987e-02
7.75907815e-01 -2.53837705e-01 -5.55994272e-01 5.57192862e-01
4.30240273e-01 -3.44785005e-01 -1.99876782e-02 1.06002080e+00
4.50214326e-01 3.52769017e-01 4.60659534e-01 -7.74852395e-01
-6.33432269e-01 -1.79858714e-01 -9.63744998e-01 -2.80430675e-01
2.86103487e-01 1.08466521e-01 9.06256020e-01 -9.69612062e-01
1.85887381e-01 1.00031769e+00 6.03508234e-01 2.46945187e-01
-1.54074550e+00 -8.91442746e-02 8.08869600e-02 -1.90884352e-01
-1.22345173e+00 -5.15603483e-01 1.31485713e+00 -3.02275419e-01
7.06220925e-01 6.01442516e-01 5.01596749e-01 7.23015666e-01
-1.83188319e-01 5.05970240e-01 1.70915484e+00 -6.51631057e-01
3.18679303e-01 1.18935242e-01 1.31109571e-02 5.31289518e-01
1.45451352e-01 1.58633173e-01 -5.77310622e-01 -2.46923804e-01
7.61040211e-01 5.10759167e-02 -7.38664865e-01 -4.21339393e-01
-1.03530741e+00 6.00353241e-01 4.49330211e-01 -1.84852898e-01
-5.41277111e-01 6.67677894e-02 -2.64099926e-01 6.17692918e-02
7.60035574e-01 3.31500871e-03 -4.98427838e-01 2.62173533e-01
-7.58164167e-01 -1.80227190e-01 5.71668267e-01 6.88461363e-01
1.19510829e+00 -9.00946837e-03 2.89891452e-01 1.12232876e+00
7.73957789e-01 7.32913017e-01 -2.20968444e-02 -1.12247264e+00
6.89351410e-02 4.38897848e-01 3.17732722e-01 -9.50022995e-01
-2.66704828e-01 -3.41475040e-01 -9.11973178e-01 5.83783448e-01
3.25155735e-01 4.55731153e-02 -7.06311405e-01 1.53600991e+00
5.67554176e-01 2.55054757e-02 1.18809670e-01 1.16350341e+00
4.85368639e-01 5.09675860e-01 -3.07995886e-01 -4.25190300e-01
1.32440245e+00 -8.20324540e-01 -6.09767795e-01 -1.81037173e-01
1.27043635e-01 -1.00825596e+00 1.20669878e+00 5.65853119e-01
-1.16414237e+00 -3.17954898e-01 -8.68877470e-01 -3.99188578e-01
-1.59110144e-01 3.61664742e-01 7.60960221e-01 9.33529615e-01
-1.27286863e+00 4.07593429e-01 -7.18489587e-01 -4.76583838e-01
4.98809069e-02 2.19465911e-01 -1.69066846e-01 -1.66104525e-01
-5.81663430e-01 6.10203564e-01 -4.61579068e-03 5.99601507e-01
-4.43494886e-01 -6.81539834e-01 -7.49859273e-01 -2.11100146e-01
4.50101018e-01 -6.02278292e-01 6.80894613e-01 -1.10231841e+00
-1.94127119e+00 1.10944533e+00 -4.18039173e-01 3.42098549e-02
3.82992148e-01 -3.12547028e-01 -1.99656226e-02 2.42832989e-01
-3.57416719e-01 1.77579716e-01 1.14362109e+00 -1.79769003e+00
-7.96610042e-02 -4.92468536e-01 2.40907952e-01 3.52161467e-01
-2.82511055e-01 8.24591815e-02 -4.98536497e-01 -5.30235827e-01
4.94213909e-01 -7.84707189e-01 -8.76712874e-02 3.18839401e-01
-4.48058128e-01 3.50977272e-01 3.89591098e-01 -7.10573196e-01
7.17085004e-01 -2.20684695e+00 -3.82917863e-03 3.16258758e-01
1.12751655e-01 1.20047329e-03 -1.31910801e-01 2.04735994e-01
-3.00534725e-01 -1.02133118e-01 -7.64488637e-01 -4.38411325e-01
-2.98635811e-02 3.40536833e-01 -1.83047533e-01 9.85495448e-01
1.82071820e-01 5.79371870e-01 -6.70153975e-01 -5.13452552e-02
4.65434283e-01 1.07289302e+00 -3.76409620e-01 1.81632072e-01
6.18510395e-02 4.66660321e-01 -2.19047695e-01 7.53297389e-01
1.19258070e+00 -2.00124346e-02 1.44500941e-01 -6.04769886e-01
-4.05264884e-01 -5.80361374e-02 -1.17368746e+00 1.88136864e+00
-5.37020504e-01 4.25306827e-01 6.75081551e-01 -1.00526416e+00
1.13538432e+00 -5.11769503e-02 7.21723914e-01 -6.93822801e-01
7.17635229e-02 6.71696290e-02 -5.95557392e-01 -3.50610852e-01
3.39304149e-01 -2.65270382e-01 7.54117370e-01 5.23162782e-01
-3.66081864e-01 -1.54775649e-01 -3.43846589e-01 -1.69987813e-01
8.01598072e-01 6.94159746e-01 7.99219981e-02 -4.43501115e-01
7.71040618e-01 -1.71360955e-01 3.52145821e-01 5.25678575e-01
1.20097427e-02 7.92050838e-01 8.07843432e-02 -3.68784726e-01
-7.50495076e-01 -1.05319440e+00 -2.92559773e-01 9.51526284e-01
2.46609733e-01 9.70770866e-02 -7.29424417e-01 -9.00805593e-02
-2.10512313e-03 5.66392839e-01 -6.43758416e-01 1.83960691e-01
-4.33582634e-01 -1.19014049e+00 -6.59387931e-02 1.65593222e-01
6.63095891e-01 -6.28995717e-01 -8.71414840e-01 -8.54687616e-02
-1.11390561e-01 -1.16712999e+00 -2.19576329e-01 2.65749127e-01
-9.88439798e-01 -1.05386281e+00 -8.35655749e-01 -2.97564298e-01
9.77016449e-01 9.60464716e-01 1.17771959e+00 7.65012503e-02
-5.83223403e-01 1.03105795e+00 -3.32535297e-01 -1.87086761e-01
1.58644155e-01 -4.98914123e-01 -2.52436519e-01 5.38005710e-01
2.63227761e-01 -7.03768551e-01 -7.92912483e-01 2.18308836e-01
-1.02922404e+00 3.62676173e-01 5.39144695e-01 6.76374257e-01
7.63502061e-01 -3.89932580e-02 -2.36662805e-01 -6.41982794e-01
2.62187302e-01 -1.37241915e-01 -8.48311841e-01 2.47952744e-01
-5.57292521e-01 -4.65013273e-02 2.20998168e-01 -1.01027288e-01
-1.67583585e+00 5.66400111e-01 2.14233920e-01 -2.15086311e-01
-3.40966851e-01 2.72629201e-01 -3.21875542e-01 -6.22420847e-01
4.94588077e-01 4.19748217e-01 1.22133799e-01 -7.30310023e-01
6.01875484e-01 5.15064061e-01 5.89036226e-01 -7.90686488e-01
1.02859282e+00 1.15489137e+00 1.52420357e-01 -1.32505441e+00
-7.70617068e-01 -7.46208727e-01 -8.10563385e-01 -3.32517982e-01
9.21879590e-01 -9.29216444e-01 -8.95255029e-01 4.55525488e-01
-1.06136298e+00 -5.77526391e-01 -2.60386229e-01 3.20407152e-01
-6.10854983e-01 5.83892941e-01 -2.84088761e-01 -1.24329293e+00
-1.44670695e-01 -8.24311793e-01 1.31476009e+00 7.24045187e-02
4.51404959e-01 -1.12468421e+00 1.11047573e-01 5.34609318e-01
4.64455634e-01 5.79689920e-01 3.77933681e-01 4.87386465e-01
-7.70429254e-01 1.83131829e-01 -7.14252472e-01 4.89964366e-01
3.90080392e-01 2.27812082e-02 -1.45597541e+00 -2.80821025e-01
5.92775106e-01 -9.74739939e-02 9.88262355e-01 4.61563379e-01
1.13145530e+00 -1.35262325e-01 2.14208156e-01 1.21176636e+00
2.09155822e+00 -2.83986419e-01 7.90866137e-01 4.95802402e-01
8.26937616e-01 1.00341356e+00 3.73162776e-01 6.21091187e-01
3.56375605e-01 5.32019854e-01 7.61075914e-01 -7.97457516e-01
-1.48918211e-01 4.54516709e-01 2.25329950e-01 5.39945722e-01
-5.48398018e-01 -8.27042013e-02 -6.08119071e-01 3.05758804e-01
-1.75652826e+00 -6.18599772e-01 -5.96748292e-01 2.34897208e+00
7.44570255e-01 -5.84305108e-01 -1.32249385e-01 1.64224416e-01
3.96582931e-01 2.45915622e-01 -4.29340333e-01 -1.22998260e-01
-5.12572348e-01 2.01681390e-01 7.11136699e-01 7.08091497e-01
-9.12740648e-01 5.63451886e-01 6.53942823e+00 3.42251092e-01
-1.05719352e+00 1.42512202e-01 5.49240530e-01 -3.84205095e-02
-6.06643081e-01 7.29802549e-02 -1.68395504e-01 1.04159288e-01
2.98798949e-01 3.32705587e-01 9.48968470e-01 4.42532152e-01
5.42631388e-01 -5.24899483e-01 -7.45561838e-01 1.13466299e+00
1.83637634e-01 -7.56945610e-01 -2.90529937e-01 2.26137683e-01
6.41098917e-01 5.28128855e-02 3.28282565e-02 -5.11893034e-01
1.84239134e-01 -9.13144827e-01 4.92603362e-01 6.68744087e-01
7.56269693e-01 -4.70958054e-01 3.66061300e-01 3.43390256e-02
-1.20330405e+00 1.13084048e-01 -6.10357046e-01 -1.46962881e-01
5.15904836e-02 8.92483473e-01 -4.12197188e-02 8.01070511e-01
7.42893875e-01 6.24481738e-01 -3.17820817e-01 5.83098412e-01
-3.83309633e-01 3.76020372e-01 -5.72194219e-01 4.55778301e-01
3.19759436e-02 -1.06005061e+00 3.56705844e-01 1.09454155e+00
5.63682877e-02 5.56921661e-01 5.11864759e-02 1.23766959e+00
3.07698876e-01 1.32517949e-01 -3.42409402e-01 3.84878784e-01
-3.61500949e-01 1.65293562e+00 -6.74643815e-01 8.84441510e-02
-6.76533043e-01 1.36344993e+00 1.10524274e-01 9.97020125e-01
-6.50012255e-01 -5.23689464e-02 6.94128036e-01 -1.79026440e-01
7.89648965e-02 -4.86015588e-01 -6.15673721e-01 -1.36108398e+00
2.41840079e-01 -4.95260328e-01 -3.14250290e-02 -1.10090601e+00
-1.08279467e+00 2.70973176e-01 -2.10550517e-01 -9.99062240e-01
4.13888007e-01 -9.46798503e-01 -5.22176325e-01 1.27410507e+00
-2.35308623e+00 -1.30413806e+00 -8.38418424e-01 9.05548811e-01
2.91774958e-01 4.63654906e-01 7.20058441e-01 1.19710349e-01
-5.88254869e-01 5.10758013e-02 3.23494941e-01 -3.09274346e-01
5.36461830e-01 -1.33717453e+00 -2.25009233e-01 9.75128233e-01
-1.69211894e-01 7.01861978e-01 7.13325858e-01 -3.43157589e-01
-1.95751870e+00 -8.10435891e-01 3.52191836e-01 -9.82826501e-02
5.33711612e-01 -2.68959433e-01 -8.49835873e-01 3.70797098e-01
2.91013539e-01 3.03381652e-01 8.38760018e-01 -1.78494334e-01
-4.23401952e-01 -2.45241955e-01 -1.24657595e+00 3.47483754e-01
1.06483543e+00 -6.87202990e-01 -6.88401386e-02 5.19858539e-01
3.37535679e-01 -1.90277651e-01 -8.28653216e-01 3.57099473e-01
6.40446782e-01 -1.42085361e+00 1.24690485e+00 5.78261204e-02
1.97661325e-01 -5.81516325e-01 -4.77162898e-01 -9.31045175e-01
-3.51193488e-01 -8.73499095e-01 4.86109257e-02 1.21759558e+00
-4.90210252e-03 -8.45213830e-01 7.03668833e-01 7.95022845e-01
-7.42266998e-02 -5.13951540e-01 -5.81391275e-01 -6.83608592e-01
-3.75112861e-01 -4.44878101e-01 2.57215053e-01 9.01871145e-01
-4.68517870e-01 -2.22946972e-01 -4.20358568e-01 4.71591383e-01
1.21471155e+00 1.72977030e-01 8.76257241e-01 -1.12887037e+00
-4.43302482e-01 -5.05007684e-01 1.36744693e-01 -1.15140331e+00
-2.66492628e-02 -5.94683170e-01 4.35562938e-01 -1.45329738e+00
2.32144773e-01 -3.94825071e-01 -2.32696459e-01 4.59386379e-01
-1.79861203e-01 5.29396832e-01 -6.25991970e-02 2.74225205e-01
-5.63322790e-02 6.62070215e-01 1.02940357e+00 1.30846677e-02
-2.21077368e-01 -2.89918512e-01 -6.63875103e-01 8.72289240e-01
7.58622169e-01 -3.98161300e-02 -5.16559184e-01 -6.95726931e-01
4.32512224e-01 -2.57619590e-01 5.21708190e-01 -4.92366344e-01
-5.40929250e-02 -3.50138277e-01 3.29465121e-01 -3.41542453e-01
6.09536529e-01 -1.14341497e+00 2.18396619e-01 4.01522685e-03
4.01204415e-02 -4.10457641e-01 5.10886461e-02 6.17864072e-01
5.01782037e-02 -2.02848446e-02 8.04920733e-01 -1.87781528e-01
-3.86520177e-01 8.18936974e-02 -7.88289234e-02 -4.96218026e-01
7.69220889e-01 -5.92642546e-01 -1.89168707e-01 -1.61914989e-01
-4.19091046e-01 -2.22057283e-01 8.24869156e-01 -2.19353020e-01
6.10173345e-01 -1.00135064e+00 -6.11062348e-01 2.52119422e-01
-5.53937815e-02 4.24667671e-02 2.47659430e-01 1.22189224e+00
-6.15599036e-01 8.27590562e-03 3.26896906e-02 -7.81691670e-01
-1.24778569e+00 4.06183571e-01 3.13296914e-01 1.49737775e-01
-7.28141129e-01 8.35763156e-01 4.17218000e-01 -4.91191477e-01
1.17797852e-01 -3.31710696e-01 9.99314040e-02 -2.13809446e-01
4.31560785e-01 5.67423403e-01 9.65425186e-03 -7.86246896e-01
-3.28372419e-01 1.12715137e+00 5.58538437e-01 -1.15721658e-01
1.55969012e+00 -6.51911020e-01 -5.30648291e-01 2.96434820e-01
1.10974252e+00 9.56051201e-02 -1.61697268e+00 -2.94023186e-01
-2.35224649e-01 -7.84009993e-01 5.46549261e-01 -7.44143069e-01
-1.19760835e+00 8.66838336e-01 5.53222299e-01 3.30862731e-01
1.76380897e+00 -3.79757524e-01 3.97363991e-01 1.87231973e-01
2.25498050e-01 -9.73593056e-01 -2.14891937e-02 1.16491422e-01
1.11272216e+00 -1.41165149e+00 3.54305685e-01 -9.29127932e-01
-2.41892815e-01 1.38154018e+00 1.36068612e-01 -1.36495665e-01
5.53982139e-01 1.76642418e-01 2.93102533e-01 -1.27864465e-01
-2.20109507e-01 -4.75032121e-01 4.51711535e-01 7.63001680e-01
4.69137132e-01 -1.02026582e-01 -1.48288295e-01 -3.08442824e-02
2.56298006e-01 -2.47207418e-01 3.97876620e-01 8.06731224e-01
-4.21103001e-01 -9.95653629e-01 -7.16310143e-01 -5.47775030e-02
-1.89383924e-01 -2.58919895e-01 -3.23377967e-01 4.58815396e-01
-1.08597510e-01 1.07058024e+00 -1.60667524e-01 9.45442021e-02
4.82134894e-02 -1.46621138e-01 5.27660370e-01 -3.38427842e-01
-2.15609699e-01 6.41348004e-01 -3.20876569e-01 -8.37329388e-01
-1.12927425e+00 -7.12085903e-01 -8.89612973e-01 -7.19622597e-02
-5.09531200e-01 -4.39482838e-01 1.05200195e+00 7.57468820e-01
9.39473137e-02 4.55286980e-01 8.35053861e-01 -1.13448954e+00
-2.24095151e-01 -5.94036579e-01 -8.01255941e-01 3.42749864e-01
5.56538403e-01 -5.37592411e-01 -7.80189395e-01 3.35787088e-01] | [10.088234901428223, -2.8284285068511963] |
a9e4948b-05ab-46ac-9d9a-0808ee4e160a | sendd-sparse-efficient-neural-depth-and | 2305.06477 | null | https://arxiv.org/abs/2305.06477v1 | https://arxiv.org/pdf/2305.06477v1.pdf | SENDD: Sparse Efficient Neural Depth and Deformation for Tissue Tracking | Deformable tracking and real-time estimation of 3D tissue motion is essential to enable automation and image guidance applications in robotically assisted surgery. Our model, Sparse Efficient Neural Depth and Deformation (SENDD), extends prior 2D tracking work to estimate flow in 3D space. SENDD introduces novel contributions of learned detection, and sparse per-point depth and 3D flow estimation, all with less than half a million parameters. SENDD does this by using graph neural networks of sparse keypoint matches to estimate both depth and 3D flow. We quantify and benchmark SENDD on a comprehensively labelled tissue dataset, and compare it to an equivalent 2D flow model. SENDD performs comparably while enabling applications that 2D flow cannot. SENDD can track points and estimate depth at 10fps on an NVIDIA RTX 4000 for 1280 tracked (query) points and its cost scales linearly with an increasing/decreasing number of points. SENDD enables multiple downstream applications that require 3D motion estimation. | ['Septimiu E. Salcudean', 'Simon DiMaio', 'Omid Mohareri', 'Adam Schmidt'] | 2023-05-10 | null | null | null | null | ['motion-estimation'] | ['computer-vision'] | [-2.14829430e-01 2.01414928e-01 -4.82584596e-01 2.28386119e-01
-7.14238405e-01 -8.78841400e-01 2.76084542e-01 8.56743231e-02
-5.58012664e-01 3.56811613e-01 4.05672491e-01 -3.41013789e-01
1.33914530e-01 -4.56789792e-01 -4.35641766e-01 -4.38013643e-01
-5.45427263e-01 4.40893054e-01 4.73363131e-01 1.95233807e-01
9.99098048e-02 1.10199785e+00 -9.04860556e-01 -6.76850453e-02
7.17080161e-02 9.65720594e-01 2.18704715e-01 1.26565373e+00
-1.59040794e-01 7.40808368e-01 -2.74423301e-01 1.01694427e-01
8.35381031e-01 -7.67439678e-02 -8.30413640e-01 -3.33533645e-01
9.44060504e-01 -8.85326326e-01 -7.69214571e-01 5.09676993e-01
9.45185244e-01 5.58478124e-02 5.48842132e-01 -1.04367280e+00
-4.07620333e-02 8.23078379e-02 -5.84142029e-01 5.78242481e-01
2.98260212e-01 5.54154694e-01 1.79686025e-01 -7.13061690e-01
1.26125646e+00 9.96698320e-01 1.23323560e+00 9.29937243e-01
-9.74975049e-01 -2.79866874e-01 -2.52307981e-01 -7.30832696e-01
-8.48797381e-01 -4.42356944e-01 3.62703294e-01 -7.47291744e-01
1.10834122e+00 5.00483215e-02 1.24122679e+00 5.38887501e-01
8.92630756e-01 5.89031994e-01 3.36736113e-01 1.10436045e-01
6.62271306e-02 -5.48134267e-01 -1.01859756e-01 1.33058107e+00
2.74528593e-01 3.13590229e-01 -5.65000057e-01 -3.89159173e-01
1.95650530e+00 1.14180399e-02 -7.20471561e-01 -6.45103395e-01
-1.80877852e+00 6.30555987e-01 5.30326605e-01 8.03815424e-02
-4.64649796e-01 1.02980852e+00 6.86727345e-01 1.71526521e-01
4.06536996e-01 4.08015668e-01 -2.24102467e-01 -6.00218832e-01
-7.25340545e-01 2.29292251e-02 9.83478844e-01 1.12817395e+00
4.78647262e-01 2.07782954e-01 -2.23646581e-01 9.24291015e-02
2.48101607e-01 5.42526007e-01 5.45767903e-01 -1.81675184e+00
-4.98288795e-02 4.87945229e-01 3.55218984e-02 -1.09928155e+00
-8.88405085e-01 -2.78787434e-01 -8.54326308e-01 6.99544549e-01
6.88666105e-01 -3.65595996e-01 -1.02138293e+00 1.18417609e+00
8.23073983e-01 4.37330991e-01 -2.18703508e-01 1.13340652e+00
1.27772236e+00 2.77558982e-01 -2.56469697e-01 -7.18377829e-02
1.02491915e+00 -8.63874197e-01 -4.46014404e-01 -1.93427384e-01
1.15772057e+00 -5.63929975e-01 7.21421182e-01 1.68509871e-01
-1.46174860e+00 -1.12942941e-01 -5.62696397e-01 -4.86944377e-01
1.66686043e-01 -1.79234788e-01 1.15622377e+00 6.13553405e-01
-1.58533955e+00 6.91298962e-01 -1.56997836e+00 -9.11297500e-02
7.39389300e-01 9.26397085e-01 -6.39561772e-01 -5.50946966e-02
-3.53814542e-01 7.56846428e-01 -1.54157892e-01 2.46205896e-01
-8.58650029e-01 -1.58014452e+00 -1.04210055e+00 -4.45043981e-01
-1.59063786e-01 -1.39973962e+00 1.14786196e+00 -1.85306668e-01
-1.45055056e+00 1.22860885e+00 -2.06330389e-01 -4.90833700e-01
8.79035175e-01 -1.61442772e-01 4.56351668e-01 5.91775119e-01
-7.71089196e-02 1.12222826e+00 5.47516525e-01 -5.96892059e-01
-2.16585085e-01 -5.21737337e-01 -1.06508963e-01 3.41248751e-01
-1.67563945e-01 -3.75592172e-01 -5.48785090e-01 -4.31887001e-01
2.79248118e-01 -1.05825424e+00 -6.13699377e-01 1.39292550e+00
-1.47354737e-01 1.67668104e-01 9.01473641e-01 -3.49167407e-01
5.93452930e-01 -1.73505461e+00 2.36502856e-01 -9.52560976e-02
9.01333213e-01 4.36842674e-03 -5.14244176e-02 -1.97546691e-01
4.90972936e-01 -2.45431960e-01 5.75396791e-02 -3.83586466e-01
-5.15850425e-01 2.83178717e-01 1.98009789e-01 1.03149748e+00
-2.07371056e-01 1.24523771e+00 -1.22854829e+00 -8.21433544e-01
4.33809608e-01 8.16821516e-01 -8.40665162e-01 -5.53260781e-02
2.72080898e-01 7.64812827e-01 -6.34981632e-01 8.33164454e-01
6.49414599e-01 -5.62251866e-01 -1.07544966e-01 -6.16084158e-01
-1.88824266e-01 -1.92171320e-01 -7.53256083e-01 2.55572891e+00
-3.83267462e-01 8.30206394e-01 3.64190727e-01 -4.84739155e-01
8.01510513e-01 2.90986747e-01 1.37526953e+00 -1.94851890e-01
3.67593050e-01 3.99081767e-01 -4.07866091e-01 -4.63909149e-01
4.29159641e-01 9.85511169e-02 1.88226119e-01 3.58423412e-01
1.42551631e-01 -6.97152555e-01 -2.72968531e-01 4.26506191e-01
1.63689113e+00 3.32771689e-01 9.09247175e-02 -3.78911346e-01
4.83228266e-02 5.31598747e-01 3.59944582e-01 6.80371881e-01
-7.10330784e-01 5.43999493e-01 5.35506845e-01 -1.04383457e+00
-1.01345325e+00 -9.43537831e-01 -2.88602468e-02 4.30720687e-01
4.54987049e-01 6.46059169e-03 -2.11961180e-01 -7.36718893e-01
5.09771883e-01 -4.10405099e-01 -7.05762744e-01 2.68899202e-01
-1.00104427e+00 -2.28259519e-01 4.93269414e-01 5.17003536e-01
-1.87607296e-02 -7.30368793e-01 -1.29101145e+00 2.93495744e-01
8.26500356e-02 -1.24988961e+00 -7.62308538e-01 1.44722760e-01
-1.39771497e+00 -1.10492206e+00 -1.19587743e+00 -8.95157158e-01
7.40077674e-01 4.60750282e-01 1.11915398e+00 1.99834526e-01
-8.45693767e-01 7.44763315e-01 -3.26623693e-02 -2.15601608e-01
-4.49838966e-01 9.37303305e-02 -7.54667148e-02 -9.15009618e-01
-1.24704547e-01 -5.63239396e-01 -1.11128724e+00 2.73760229e-01
-3.92419368e-01 5.15549518e-02 1.37903735e-01 7.28325427e-01
8.19132328e-01 -1.03882706e+00 5.44865504e-02 -6.17876589e-01
1.92744419e-01 -1.51637197e-01 -8.40033948e-01 -2.69508958e-01
-1.65414169e-01 -7.38497749e-02 1.52908966e-01 -5.94847381e-01
-4.21697140e-01 7.82487512e-01 -6.98261261e-02 -1.15179849e+00
3.08430314e-01 -1.35564972e-02 9.90607858e-01 -9.89457369e-01
1.01366770e+00 -1.84652694e-02 8.24075341e-01 1.59865886e-01
2.55520880e-01 -1.60181448e-02 8.16008151e-01 -2.50974685e-01
4.84633058e-01 1.10177577e+00 5.99390507e-01 -5.66531956e-01
-3.79030347e-01 -7.23420560e-01 -8.08223903e-01 -5.06895483e-01
7.25402057e-01 -8.18224430e-01 -1.25885165e+00 4.32273716e-01
-1.18476737e+00 -9.25139427e-01 -5.91058671e-01 8.46743762e-01
-9.96828437e-01 2.33964100e-01 -1.08309269e+00 -4.57921356e-01
-8.39111924e-01 -1.33625042e+00 1.56006253e+00 -8.45044926e-02
-3.11721653e-01 -1.39764094e+00 1.37602925e-01 -1.89620450e-01
6.14971042e-01 8.48563969e-01 2.04135776e-01 1.12635605e-01
-7.70720005e-01 -2.53243089e-01 -1.56974524e-01 -3.82057756e-01
4.06047583e-01 -2.53486156e-01 -7.02099383e-01 -4.65842962e-01
-1.54176831e-01 -1.14322476e-01 5.73513150e-01 1.30800879e+00
9.06236529e-01 -1.46497726e-01 -7.75480747e-01 1.28577769e+00
1.26978672e+00 -1.80165425e-01 3.27260822e-01 1.45576760e-01
9.38625693e-01 3.38932157e-01 4.59185869e-01 3.38835835e-01
8.56599882e-02 5.70264280e-01 7.02936649e-01 -3.93389374e-01
-5.98290622e-01 1.73315465e-01 1.58652961e-02 5.96768498e-01
-2.36247972e-01 1.19128183e-01 -9.10567284e-01 5.03948748e-01
-1.58822942e+00 -6.85106397e-01 -1.90497935e-01 2.05866289e+00
7.21339345e-01 -1.97324693e-01 1.98432982e-01 -4.34426337e-01
3.70082527e-01 -3.35324407e-02 -9.13067222e-01 1.18164197e-02
2.40117013e-01 2.19351649e-01 1.03356993e+00 6.61830544e-01
-9.17053401e-01 8.15557599e-01 6.81369305e+00 4.14505631e-01
-1.31827235e+00 -1.40680159e-02 5.45229733e-01 -5.95237851e-01
-3.21544617e-01 -3.40684474e-01 -7.65505791e-01 9.25843883e-03
3.99378598e-01 -2.45398954e-01 3.90074514e-02 6.31955266e-01
3.65054786e-01 -1.60612717e-01 -1.04085386e+00 1.25608134e+00
-9.23523232e-02 -2.13993526e+00 -1.41574740e-01 4.79434371e-01
5.10016203e-01 4.80188102e-01 -2.79253006e-01 -3.36892694e-01
3.62166911e-01 -9.33342755e-01 1.52923748e-01 5.31943083e-01
1.19623017e+00 -3.80618602e-01 5.15249431e-01 2.97135115e-01
-1.08705211e+00 3.14862430e-01 -3.39052051e-01 3.67797017e-01
4.16943938e-01 5.19733787e-01 -1.17292452e+00 2.47537777e-01
7.18089223e-01 1.15540850e+00 5.42459972e-02 1.33627439e+00
3.82449090e-01 8.70533753e-03 -6.50986969e-01 -2.46928073e-02
2.48932540e-01 2.54024446e-01 8.52044642e-01 1.14673042e+00
4.82930869e-01 1.07666321e-01 6.64613023e-02 4.42491382e-01
1.30869923e-02 -1.59197360e-01 -8.46606672e-01 3.34465325e-01
4.68521267e-01 1.46261871e+00 -9.53719199e-01 -1.99825287e-01
6.87740464e-03 9.26871598e-01 1.18911818e-01 -1.48647130e-01
-4.62054223e-01 -2.68424451e-01 7.69114792e-01 3.09275538e-01
-1.45501997e-02 -6.87997937e-01 -1.09707534e-01 -9.55060780e-01
-2.90979207e-01 -7.11944252e-02 2.58524269e-01 -8.72985661e-01
-1.00574446e+00 4.49978948e-01 -4.44948018e-01 -1.63208199e+00
-4.65527773e-01 -7.58650959e-01 -1.50760084e-01 7.54418194e-01
-1.63542867e+00 -9.45421755e-01 -8.88885260e-01 8.41013908e-01
2.88854986e-01 3.67407978e-01 9.76223230e-01 1.20259769e-01
1.39279142e-01 3.16138357e-01 -4.73406106e-01 4.51027781e-01
7.02119887e-01 -1.08873713e+00 6.83683932e-01 3.40937525e-01
-2.87337273e-01 3.61198723e-01 1.94836706e-01 -7.85749376e-01
-2.17862248e+00 -9.81673598e-01 2.65809864e-01 -8.21709871e-01
4.92540628e-01 -2.93450616e-02 -5.30416489e-01 8.00459325e-01
-3.55199426e-01 1.18116283e+00 3.91193092e-01 -6.12553954e-01
1.68675736e-01 2.08971038e-01 -1.59579623e+00 3.84294122e-01
1.50212991e+00 -1.13035463e-01 -5.09327017e-02 6.90039396e-01
9.86332178e-01 -1.56543064e+00 -1.48845541e+00 1.38875380e-01
9.97308493e-01 -7.53329933e-01 1.22529447e+00 -1.71568647e-01
2.65959680e-01 -1.46409154e-01 2.90152878e-01 -9.58471835e-01
-2.64117301e-01 -1.15899611e+00 -5.49855530e-01 -3.87247689e-02
-1.23139717e-01 -7.40631759e-01 1.51884556e+00 6.31386161e-01
-6.50963724e-01 -8.53309453e-01 -1.29704916e+00 -6.45305037e-01
9.37744752e-02 -2.02530503e-01 1.92982733e-01 9.46370125e-01
6.11647628e-02 -5.52370906e-01 2.54096054e-02 -3.88996303e-02
7.39326835e-01 -3.87712754e-02 8.95297647e-01 -1.04569519e+00
-6.57506566e-03 -3.25075328e-01 -1.01480639e+00 -1.46034300e+00
-1.86109528e-01 -1.03653026e+00 -1.30747601e-01 -1.68203640e+00
-3.47188622e-01 -6.84395194e-01 3.19014460e-01 5.77857912e-01
3.43120903e-01 5.28397739e-01 -1.04957916e-01 5.22052109e-01
-7.25410432e-02 1.35793956e-02 2.06564212e+00 1.02235608e-01
-4.71946716e-01 -1.30902529e-01 -1.87436894e-01 6.43574238e-01
3.31170321e-01 -4.06482846e-01 -2.68501103e-01 -8.80528450e-01
-2.76210122e-02 7.49982357e-01 7.58693337e-01 -1.04567659e+00
6.91006839e-01 7.00469613e-02 6.41096830e-01 -6.10616982e-01
6.25161111e-01 -7.10990131e-01 2.80006677e-02 1.03064561e+00
-1.77810162e-01 1.97480962e-01 4.48608220e-01 4.66493875e-01
2.14720070e-01 1.60560459e-01 6.59838736e-01 -5.11607170e-01
-6.22899354e-01 9.78538513e-01 -1.57089308e-01 2.07455128e-01
1.03377104e+00 -8.26388419e-01 -3.03596973e-01 -3.04355443e-01
-9.05536950e-01 1.25503108e-01 5.30729771e-01 1.50840998e-01
9.14537013e-01 -1.19165289e+00 -7.44477093e-01 5.91175370e-02
-2.22410128e-01 5.77993631e-01 4.02536958e-01 1.15032709e+00
-1.39231861e+00 2.24233478e-01 -2.21856475e-01 -1.19347036e+00
-1.10695016e+00 1.25244528e-01 7.80126810e-01 -7.15975836e-02
-1.17236042e+00 9.98760521e-01 6.93692884e-04 -3.90621513e-01
1.24838158e-01 -5.29239178e-01 1.07829504e-01 -5.13775766e-01
5.12297213e-01 3.42988968e-01 1.53849229e-01 -3.26350778e-01
-4.21111971e-01 1.11513448e+00 2.76183993e-01 1.60472825e-01
1.21952629e+00 6.68947846e-02 1.37885243e-01 1.35246003e-02
1.36465275e+00 -1.36106104e-01 -1.64452553e+00 1.35963246e-01
-6.79535806e-01 -5.81147909e-01 5.04441142e-01 -2.87992746e-01
-1.49629498e+00 7.18598485e-01 6.22830987e-01 -2.77328417e-02
7.33398914e-01 5.49836904e-02 1.08546937e+00 1.60388008e-01
6.47907794e-01 -2.36226246e-01 -7.23852068e-02 3.11087430e-01
5.77127457e-01 -1.12428391e+00 2.96654999e-01 -6.91839457e-01
-2.24424094e-01 1.42405188e+00 4.77979094e-01 -4.02870089e-01
6.73365057e-01 9.93729413e-01 1.84294760e-01 -4.20386523e-01
-4.42792565e-01 3.64141285e-01 2.77519494e-01 9.91570413e-01
4.29766119e-01 -3.90050858e-01 1.61635906e-01 -6.77058935e-01
-1.01989858e-01 5.60598135e-01 5.05000353e-01 1.19796598e+00
-1.95348650e-01 -4.98892993e-01 3.63134742e-02 5.03581464e-01
-4.21132714e-01 4.59251739e-02 3.95774543e-01 8.90983582e-01
-5.05224466e-01 2.45446220e-01 4.36236292e-01 7.13509247e-02
2.08305284e-01 -8.01567554e-01 8.57486963e-01 -4.32811588e-01
-8.80940914e-01 4.48417068e-02 -1.40263170e-01 -1.34072781e+00
-5.64806461e-01 -4.16750222e-01 -1.52019966e+00 -3.35556358e-01
-1.86309703e-02 -2.67446280e-01 9.01168227e-01 2.38646135e-01
7.54989564e-01 3.84403974e-01 3.22241247e-01 -1.48883295e+00
-1.49458081e-01 -2.75728911e-01 -2.91119248e-01 -1.20526634e-01
7.85895109e-01 -5.29479325e-01 -5.01272917e-01 -6.07688166e-03] | [14.036113739013672, -3.0984606742858887] |
500ae70c-3f49-4f1f-b48b-5517b241f189 | retraining-a-graph-based-recommender-with | 2305.03624 | null | https://arxiv.org/abs/2305.03624v1 | https://arxiv.org/pdf/2305.03624v1.pdf | Retraining A Graph-based Recommender with Interests Disentanglement | In a practical recommender system, new interactions are continuously observed. Some interactions are expected, because they largely follow users' long-term preferences. Some other interactions are indications of recent trends in user preference changes or marketing positions of new items. Accordingly, the recommender needs to be periodically retrained or updated to capture the new trends, and yet not to forget the long-term preferences. In this paper, we propose a novel and generic retraining framework called Disentangled Incremental Learning (DIL) for graph-based recommenders. We assume that long-term preferences are well captured in the existing model, in the form of model parameters learned from past interactions. New preferences can be learned from the user-item bipartite graph constructed using the newly observed interactions. In DIL, we design an Information Extraction Module to extract historical preferences from the existing model. Then we blend the historical and new preferences in the form of node embeddings in the new graph, through a Disentanglement Module. The essence of the disentanglement module is to decorrelate the historical and new preferences so that both can be well captured, via carefully designed losses. Through experiments on three benchmark datasets, we show the effectiveness of DIL in capturing dynamics of useritem interactions. We also demonstrate the robustness of DIL by attaching it to two base models - LightGCN and NGCF. | ['Jie Zhang', 'Aixin Sun', 'Yitong Ji'] | 2023-05-05 | null | null | null | null | ['incremental-learning', 'marketing'] | ['methodology', 'miscellaneous'] | [-9.05294344e-02 -7.90934637e-02 -4.39820647e-01 -4.89503860e-01
-9.97975916e-02 -6.69173121e-01 4.83562201e-01 6.58859983e-02
-3.33559543e-01 6.09259307e-01 5.77245653e-01 -8.34966525e-02
-6.05248988e-01 -8.12176347e-01 -6.43103957e-01 -6.08431280e-01
-3.92044127e-01 5.23369610e-01 4.05479334e-02 -4.25769061e-01
-1.12558007e-01 2.90762305e-01 -1.14647055e+00 5.08695766e-02
8.89955342e-01 6.40599191e-01 1.04474410e-01 5.83198607e-01
2.24835843e-01 4.57965612e-01 2.00798940e-02 -5.67359924e-01
2.55253345e-01 -1.67740975e-02 -4.60674554e-01 7.77169839e-02
-2.05588131e-03 -3.39490622e-01 -7.74781644e-01 7.33367383e-01
3.68702471e-01 4.68790054e-01 3.95395130e-01 -1.17861068e+00
-8.21856022e-01 1.07905960e+00 -4.18068141e-01 2.10824236e-01
3.26422989e-01 -4.13082615e-02 1.57190704e+00 -9.78725314e-01
5.03300369e-01 1.15067828e+00 4.55352277e-01 4.30534720e-01
-1.45002961e+00 -5.58916807e-01 1.01697183e+00 1.92335099e-01
-9.27518427e-01 -1.94870859e-01 1.16234207e+00 -3.50767910e-01
3.29750121e-01 4.72275496e-01 8.32184732e-01 1.42261744e+00
1.02159329e-01 9.22087669e-01 5.39660990e-01 -4.36657295e-02
5.09803072e-02 3.61272782e-01 6.33991122e-01 4.66532737e-01
3.23473573e-01 2.39204466e-01 -4.31200832e-01 -5.21041274e-01
3.78522635e-01 5.78631282e-01 -4.41691875e-01 -8.03719580e-01
-9.18076992e-01 7.73487449e-01 5.49765766e-01 2.15538114e-01
-4.08587277e-01 -3.08838576e-01 4.61090766e-02 5.81611454e-01
6.01071477e-01 3.64796013e-01 -7.81662822e-01 1.13496363e-01
-3.64972681e-01 3.21757235e-03 8.28157425e-01 9.23812866e-01
8.49989474e-01 -2.47192383e-01 -7.50490651e-02 8.88751328e-01
4.74086672e-01 3.12928438e-01 5.69591641e-01 -2.84008682e-01
2.16296554e-01 7.43253291e-01 2.33547091e-01 -1.34481084e+00
-3.62034291e-01 -8.78265619e-01 -8.57627749e-01 -5.34230351e-01
9.19392034e-02 -3.13767046e-01 -6.24563098e-01 2.00664687e+00
3.44043761e-01 5.17246723e-01 -1.78202704e-01 6.57923102e-01
4.49675441e-01 6.80877268e-01 -3.11790138e-01 -3.31940979e-01
8.61026108e-01 -9.64011610e-01 -4.99888152e-01 -3.71415854e-01
3.37084472e-01 -3.48891735e-01 1.08459067e+00 3.72902930e-01
-7.93075562e-01 -5.15008986e-01 -9.74836051e-01 2.90181905e-01
-1.90085709e-01 1.84381735e-02 6.37382746e-01 2.70912319e-01
-6.47492349e-01 7.82267332e-01 -6.77201271e-01 -2.65193433e-01
4.25266922e-02 5.49921930e-01 -7.64517859e-02 -2.15706617e-01
-1.53354585e+00 4.97626781e-01 1.16692208e-01 3.95871460e-01
-6.30805492e-01 -7.58957684e-01 -6.85094833e-01 2.81532168e-01
7.38440931e-01 -7.18580008e-01 1.03808904e+00 -8.49466085e-01
-1.29458797e+00 3.84136066e-02 6.07216954e-02 -1.46565000e-02
3.07086974e-01 -2.30862081e-01 -9.31305766e-01 -4.94142354e-01
-4.41024274e-01 -1.48734599e-01 7.19145954e-01 -1.32262015e+00
-6.98280394e-01 -4.43829775e-01 3.25375676e-01 2.99683005e-01
-7.93625176e-01 -6.12497628e-01 -6.46696627e-01 -6.93383098e-01
-1.14197657e-01 -1.28664684e+00 -3.75272900e-01 -3.46840680e-01
-2.43968859e-01 -1.15623102e-01 7.03262806e-01 -5.25739253e-01
1.49954677e+00 -2.18021202e+00 4.50166315e-01 5.29361844e-01
3.99891347e-01 2.06175953e-01 -5.29750407e-01 6.93728685e-01
-3.26211415e-02 -1.35989740e-01 1.23704456e-01 -3.36103112e-01
9.28825736e-02 5.94970405e-01 -3.87924165e-01 2.86506504e-01
-2.40158781e-01 7.85108984e-01 -1.25117481e+00 2.69565284e-01
-7.15033710e-02 3.98137927e-01 -7.50058830e-01 4.28989708e-01
-3.14178854e-01 3.77368957e-01 -6.67570412e-01 1.78654552e-01
6.29096031e-01 -2.63188452e-01 7.14220047e-01 -3.34586054e-01
1.53249949e-01 2.56724149e-01 -1.21557283e+00 1.60250044e+00
-5.87782741e-01 6.03223918e-03 -4.20425758e-02 -8.32205355e-01
6.81928515e-01 1.43984199e-01 5.59520483e-01 -4.57000703e-01
3.84124815e-02 -1.06375754e-01 2.64499281e-02 -4.25834179e-01
4.93850201e-01 3.30285609e-01 -9.71389040e-02 5.16872704e-01
1.33491129e-01 7.16685653e-01 2.44859129e-01 5.66820085e-01
9.61663187e-01 -1.79688290e-01 1.64306998e-01 -7.33158439e-02
5.53898394e-01 -4.25827295e-01 8.26373100e-01 6.23093843e-01
2.17618361e-01 2.64593780e-01 4.11968738e-01 -4.06309068e-01
-5.81944287e-01 -1.01789522e+00 2.71875650e-01 1.35277450e+00
1.79209873e-01 -6.56397820e-01 3.59542556e-02 -1.15898323e+00
2.05204114e-01 6.00657701e-01 -7.52917171e-01 -6.54501200e-01
-5.17453492e-01 -9.56818163e-01 -2.76872486e-01 3.80830944e-01
-5.27620763e-02 -6.93664789e-01 1.89710826e-01 4.50561076e-01
1.91460084e-02 -7.04961181e-01 -1.16908062e+00 -3.97623926e-02
-8.68515074e-01 -1.13214898e+00 -4.23403323e-01 -5.61917424e-01
8.02711666e-01 4.90198225e-01 9.65717018e-01 9.86441299e-02
2.21222043e-01 5.53659081e-01 -5.54095626e-01 2.03841746e-01
3.02891457e-03 1.65420100e-01 3.98756981e-01 6.27028823e-01
1.50050208e-01 -1.10614121e+00 -8.04450274e-01 4.53158170e-01
-8.37449789e-01 -1.58255801e-01 7.18381226e-01 1.11145854e+00
3.13437819e-01 5.83657548e-02 5.58083534e-01 -1.46040690e+00
8.26289952e-01 -9.29671824e-01 -4.42774564e-01 4.33373272e-01
-1.12970233e+00 2.52849311e-01 8.42225552e-01 -9.63743091e-01
-1.08739591e+00 -1.43913805e-01 -9.15948674e-02 -3.65819633e-01
2.71229178e-01 1.04920626e+00 -4.35694426e-01 2.64902413e-01
2.70268261e-01 -3.04642152e-02 -3.63391578e-01 -8.50680828e-01
7.87100434e-01 5.13381839e-01 2.99260259e-01 -5.27490616e-01
9.25761998e-01 1.95674554e-01 -5.26086390e-01 -4.00919646e-01
-9.53957796e-01 -6.25097930e-01 -5.81341922e-01 -1.79257931e-03
1.22297183e-01 -7.32495666e-01 -6.22006536e-01 2.02945754e-01
-7.03273356e-01 -7.66230598e-02 -3.96736681e-01 7.19150424e-01
-3.43760587e-02 3.12004179e-01 -5.56746721e-01 -6.59540296e-01
-6.15813285e-02 -7.94161618e-01 4.12406623e-01 3.11866790e-01
-7.13945180e-02 -1.47061038e+00 6.22901976e-01 -9.67072323e-02
2.63669670e-01 -3.40569243e-02 1.09222519e+00 -9.69530582e-01
-2.52289802e-01 -3.99662793e-01 8.62931833e-02 2.90760934e-01
5.24515688e-01 -1.21489726e-02 -6.65906429e-01 -8.05189371e-01
-2.08974987e-01 1.55359358e-01 8.85849714e-01 8.31671357e-02
8.15632343e-01 -7.33926713e-01 -4.88488734e-01 4.20735240e-01
1.13622093e+00 2.71064073e-01 1.07493319e-01 -1.74691036e-01
9.03217852e-01 4.90779668e-01 3.16512853e-01 5.00641346e-01
5.88300586e-01 6.42884552e-01 3.37943554e-01 1.19958490e-01
2.00688511e-01 -7.40170181e-01 6.26835108e-01 1.41275597e+00
1.20357223e-01 -5.09188056e-01 -3.20888698e-01 3.31216931e-01
-2.31925964e+00 -7.36231983e-01 6.34626597e-02 2.35910535e+00
5.75817466e-01 2.37371311e-01 3.00386012e-01 -2.62453794e-01
6.12694502e-01 3.19818884e-01 -9.71172392e-01 -1.99737892e-01
1.45767510e-01 -1.31252304e-01 3.07744473e-01 6.76646054e-01
-7.62759268e-01 5.46529233e-01 5.29308224e+00 3.55832100e-01
-9.40005779e-01 -1.56994667e-02 3.13119560e-01 -2.78751999e-01
-7.74274230e-01 1.61448732e-01 -5.99400640e-01 4.58338648e-01
8.30578029e-01 -4.18632865e-01 7.19414592e-01 7.60527611e-01
8.09342042e-02 4.92649615e-01 -1.25826764e+00 7.41492808e-01
6.19062781e-02 -9.46641982e-01 9.08463821e-02 7.29055181e-02
6.23917401e-01 -2.53286958e-02 2.68272817e-01 6.52339756e-01
8.31823885e-01 -4.17010486e-01 2.45737717e-01 8.49179983e-01
3.94643843e-01 -8.32593203e-01 5.02928376e-01 4.66255814e-01
-1.25437331e+00 -4.08812433e-01 -4.31012332e-01 -5.41234203e-02
1.60046369e-01 8.77602935e-01 -5.92764318e-01 1.01672316e+00
4.32715893e-01 1.36739480e+00 -5.02432406e-01 9.33101118e-01
-2.11189836e-01 7.00316370e-01 -1.25812799e-01 9.67347696e-02
4.24717553e-02 -6.56261861e-01 5.86547077e-01 7.39203572e-01
3.36353391e-01 1.92242682e-01 3.43006402e-01 5.11387765e-01
-2.66282260e-01 1.40030041e-01 -4.79818702e-01 -1.81305066e-01
4.49640334e-01 1.51718056e+00 -4.81873266e-02 -1.04968868e-01
-7.06690073e-01 1.00345016e+00 4.79797810e-01 6.34410799e-01
-6.82457209e-01 -8.98995176e-02 8.79373431e-01 -1.92669630e-02
3.28523695e-01 -7.85635933e-02 5.07033587e-01 -1.72829938e+00
2.74214782e-02 -7.78209984e-01 7.59328246e-01 -4.28201973e-01
-1.84663236e+00 7.32619047e-01 -1.18978590e-01 -1.12154293e+00
-1.52417421e-02 -3.39169890e-01 -7.31583476e-01 6.36392474e-01
-1.31322598e+00 -1.02233434e+00 -4.75290008e-02 6.75011814e-01
3.95633399e-01 -4.13337573e-02 6.35971963e-01 4.93764579e-01
-9.25814152e-01 7.10754514e-01 4.31426793e-01 -1.99565843e-01
7.55880952e-01 -1.16824245e+00 2.88995832e-01 7.20563531e-01
5.15967667e-01 8.48366678e-01 6.15346730e-01 -6.41918600e-01
-1.56397462e+00 -1.19154823e+00 6.75198138e-01 -4.32777226e-01
9.04534519e-01 -6.87934756e-01 -9.15371537e-01 9.82830226e-01
8.90806541e-02 -1.93168968e-02 8.12195301e-01 6.12047911e-01
-4.31053936e-01 -3.76394778e-01 -7.30947018e-01 7.39788473e-01
1.16030085e+00 -3.41712415e-01 -6.05426788e-01 2.16403335e-01
9.07519519e-01 -4.01093289e-02 -7.16971219e-01 4.67797428e-01
8.98139060e-01 -5.14874160e-01 8.98117900e-01 -1.03273523e+00
-1.18146203e-01 7.84180779e-03 -2.34346818e-02 -1.99381113e+00
-8.76994848e-01 -8.73415470e-01 -5.54836512e-01 1.31672323e+00
6.45389616e-01 -7.50846982e-01 7.66924620e-01 6.42534614e-01
3.78328376e-02 -7.77218163e-01 -3.97696614e-01 -6.59828484e-01
-3.27527434e-01 -1.22341745e-01 9.07778203e-01 1.18160129e+00
2.21298054e-01 7.63664603e-01 -1.11070299e+00 2.50884145e-01
3.51774305e-01 4.34658498e-01 6.20783448e-01 -1.52516639e+00
-6.78992391e-01 -1.50293559e-01 4.85868379e-02 -1.35436952e+00
4.15291525e-02 -9.57536995e-01 -3.58416140e-01 -1.30011451e+00
3.63945693e-01 -4.17911798e-01 -8.98141205e-01 2.72265375e-01
-3.39082539e-01 -2.39774212e-01 -6.83617890e-02 2.15739638e-01
-5.37746549e-01 8.68687451e-01 1.20786321e+00 -1.97395444e-01
-5.63220978e-01 5.09033799e-01 -1.08489501e+00 6.03720546e-01
6.14639878e-01 -5.75911224e-01 -9.36571598e-01 -3.06093216e-01
6.06808841e-01 1.14533454e-02 -1.72242090e-01 -3.10009748e-01
3.41223091e-01 -2.64129311e-01 1.17047757e-01 -4.33491468e-01
1.57986701e-01 -1.08032644e+00 4.18357283e-01 1.12860210e-01
-4.67025340e-01 2.79028118e-01 -2.24840969e-01 1.25233948e+00
3.15945633e-02 2.23283842e-01 2.09078506e-01 2.65238911e-01
-5.44725120e-01 9.41398740e-01 -3.68754007e-02 -1.85941949e-01
7.55503535e-01 2.25904271e-01 -1.50985628e-01 -4.13249969e-01
-1.31962252e+00 7.41367579e-01 2.59176314e-01 8.02561224e-01
6.54673159e-01 -1.54375327e+00 -3.94246370e-01 3.68138582e-01
2.46658102e-01 -3.86519462e-01 3.97794783e-01 7.76081741e-01
3.76087636e-01 3.64561714e-02 7.46934488e-02 -4.02612723e-02
-1.14148951e+00 9.66762006e-01 6.87473044e-02 -5.63564003e-01
-4.82481569e-01 6.61098301e-01 2.26174086e-01 -8.20954740e-01
3.79650801e-01 -2.56976336e-01 -5.68122804e-01 3.20427626e-01
3.35087031e-01 3.13235641e-01 -6.87060058e-02 -3.65238100e-01
-2.23417535e-01 2.95459002e-01 -5.20529926e-01 1.75362140e-01
1.67721546e+00 -4.30477560e-01 7.99920782e-03 7.36566901e-01
1.15827715e+00 5.28213918e-01 -1.34995520e+00 -7.00253963e-01
7.88872987e-02 -6.04527473e-01 -5.59282415e-02 -8.26214552e-01
-1.44480002e+00 4.90208775e-01 3.48690242e-01 3.76062095e-01
1.04525304e+00 -1.31718263e-01 8.91496599e-01 3.63901824e-01
2.91700691e-01 -7.92331338e-01 4.97749783e-02 3.69094789e-01
7.00652838e-01 -7.78365135e-01 -1.11252740e-01 -2.78639913e-01
-6.58191800e-01 9.12937403e-01 5.42978525e-01 -1.29834548e-01
1.23966444e+00 -1.87530607e-01 -2.65830904e-01 -1.27236307e-01
-1.14830124e+00 1.97945326e-03 6.97099268e-01 2.33052239e-01
4.92483005e-02 1.75574437e-01 -2.27898464e-01 1.09634459e+00
1.36106074e-01 -2.72983879e-01 5.47465086e-01 6.43407106e-01
-4.39543799e-02 -1.63601875e+00 4.13186044e-01 5.86961687e-01
4.50037159e-02 -3.47816609e-02 -4.69424754e-01 5.13030350e-01
-6.72724172e-02 9.20764446e-01 -1.53056934e-01 -1.00627077e+00
7.52507925e-01 -3.36982049e-02 1.50223926e-01 -7.41252542e-01
-3.70094776e-01 2.04324275e-02 1.02769479e-01 -5.93046784e-01
-1.44386783e-01 -5.73728204e-01 -6.29402459e-01 -3.77298623e-01
-6.20015442e-01 6.98276460e-01 3.59280780e-02 6.75660610e-01
5.71023881e-01 5.85995674e-01 1.29190266e+00 -5.68119764e-01
-7.92827249e-01 -7.52460837e-01 -8.61890972e-01 4.98804659e-01
3.59485507e-01 -7.17861295e-01 -5.72790682e-01 -2.54502028e-01] | [10.232356071472168, 5.55294942855835] |
f2397fbb-93c7-4ef2-8cb7-16efab16f58c | chinese-zero-pronoun-resolution-with-deep | null | null | https://aclanthology.org/D17-1135 | https://aclanthology.org/D17-1135.pdf | Chinese Zero Pronoun Resolution with Deep Memory Network | Existing approaches for Chinese zero pronoun resolution typically utilize only syntactical and lexical features while ignoring semantic information. The fundamental reason is that zero pronouns have no descriptive information, which brings difficulty in explicitly capturing their semantic similarities with antecedents. Meanwhile, representing zero pronouns is challenging since they are merely gaps that convey no actual content. In this paper, we address this issue by building a deep memory network that is capable of encoding zero pronouns into vector representations with information obtained from their contexts and potential antecedents. Consequently, our resolver takes advantage of semantic information by using these continuous distributed representations. Experiments on the OntoNotes 5.0 dataset show that the proposed memory network could substantially outperform the state-of-the-art systems in various experimental settings. | ['Wei-Nan Zhang', 'Ting Liu', 'Qingyu Yin', 'Yu Zhang'] | 2017-09-01 | null | null | null | emnlp-2017-9 | ['chinese-zero-pronoun-resolution'] | ['natural-language-processing'] | [-5.10900728e-02 -8.34658518e-02 -5.72859645e-01 -4.07422394e-01
-8.34429145e-01 -5.74746311e-01 5.58547318e-01 3.41006219e-01
-6.44949615e-01 8.30884337e-01 6.81495428e-01 -1.25355005e-01
-8.05724226e-03 -1.09579813e+00 -5.09897053e-01 -4.24585432e-01
2.88623571e-01 4.63084012e-01 7.16287345e-02 -5.09701073e-01
3.44939381e-01 -1.21395732e-03 -1.33521748e+00 4.17018563e-01
9.65014517e-01 7.33734548e-01 4.97977257e-01 -1.73313797e-01
-9.69517350e-01 5.50436497e-01 -9.45509970e-01 -6.64992571e-01
-1.43855706e-01 -7.07399175e-02 -8.55230987e-01 -4.85114247e-01
3.53996634e-01 -4.72850919e-01 -5.06992996e-01 1.30854869e+00
2.53507823e-01 3.70145172e-01 7.35806048e-01 -7.57260323e-01
-1.39609718e+00 9.30269420e-01 -3.13193977e-01 3.29021484e-01
4.88601238e-01 -5.45586288e-01 1.46728516e+00 -1.00932515e+00
6.03148222e-01 1.54392898e+00 5.14041603e-01 7.42736697e-01
-9.39848661e-01 -7.09850609e-01 3.83362651e-01 3.66663426e-01
-1.57196498e+00 -4.70685273e-01 7.72531629e-01 -1.19930044e-01
1.30202699e+00 2.91918695e-01 1.45752072e-01 1.27923322e+00
4.19201553e-02 8.77608538e-01 6.77433312e-01 -5.09264648e-01
1.28085136e-01 -4.00039777e-02 7.65062153e-01 3.54232192e-01
4.62262064e-01 -4.72809136e-01 -5.44144869e-01 -1.08184963e-01
7.37130642e-01 2.74286300e-01 -3.70789558e-01 2.36320227e-01
-9.89872158e-01 8.50418568e-01 3.02891463e-01 4.88962919e-01
-3.74482900e-01 5.60795777e-02 2.93889552e-01 6.67349994e-02
2.29069024e-01 3.42506826e-01 -4.59573686e-01 -1.15057215e-01
-6.70238972e-01 3.49980384e-01 7.07608998e-01 1.15518916e+00
7.94757485e-01 5.88489473e-02 -1.46456510e-02 1.15156686e+00
1.16090588e-01 2.56088465e-01 8.96480262e-01 -9.25793529e-01
7.46594667e-01 6.44592464e-01 1.24820635e-01 -1.05609071e+00
-2.33896494e-01 -2.73387313e-01 -6.84443057e-01 -4.81832325e-01
1.26467675e-01 -3.32431197e-02 -6.39769733e-01 1.92632139e+00
-1.90639153e-01 2.83262521e-01 5.66908836e-01 1.03743601e+00
1.33436394e+00 7.32827127e-01 3.81332785e-01 -1.72639102e-01
1.57036889e+00 -6.52235985e-01 -1.38895237e+00 -4.00257111e-01
2.43794933e-01 -7.39802122e-01 1.26254237e+00 -2.64107645e-01
-1.06905079e+00 -3.27391684e-01 -1.12624216e+00 -4.86351401e-01
-4.62828189e-01 -1.66582819e-02 9.22459900e-01 2.94864118e-01
-4.70479935e-01 6.31811917e-01 -6.73654675e-01 -2.12858409e-01
1.42086148e-01 2.36099795e-01 -3.65288228e-01 1.88656777e-01
-1.74204075e+00 7.19838262e-01 4.07421798e-01 -5.62171191e-02
-1.63555488e-01 -6.36475086e-01 -1.02129471e+00 4.11466122e-01
4.07009542e-01 -5.32760322e-01 1.31810904e+00 -7.50404954e-01
-1.24887133e+00 7.22126365e-01 -6.57980680e-01 -3.47265035e-01
9.21367388e-03 -3.79745752e-01 -4.38641906e-01 -1.45801216e-01
5.25213838e-01 4.98287231e-01 3.46722066e-01 -1.11927915e+00
-7.84716308e-01 -3.44471931e-01 5.83140254e-01 2.91534424e-01
-6.52541101e-01 -8.94656554e-02 -5.00195384e-01 -7.86990404e-01
4.26881969e-01 -5.97767472e-01 7.61064375e-03 -5.62229037e-01
-3.98314089e-01 -8.00906241e-01 3.32839966e-01 -5.20821571e-01
1.45575845e+00 -2.14123726e+00 6.76174387e-02 -2.68341422e-01
1.69732317e-01 1.75688237e-01 -9.83803943e-02 4.14560050e-01
2.57974446e-01 3.71520191e-01 5.15297353e-02 -2.47285634e-01
2.91137844e-01 4.99025822e-01 -8.11827242e-01 5.46087176e-02
1.79072097e-01 9.00058031e-01 -9.34506297e-01 -4.20433879e-01
-3.88291329e-02 6.10415518e-01 -4.96405005e-01 1.12941653e-01
-4.45936471e-02 -3.11542392e-01 -6.11591578e-01 7.38818944e-01
6.75127745e-01 -4.65655178e-02 7.07675993e-01 -2.44147062e-01
1.76711082e-02 9.88913953e-01 -1.12204897e+00 1.64354169e+00
-3.28698456e-01 4.21670049e-01 -2.84434408e-01 -9.31723177e-01
9.37492430e-01 3.72659236e-01 1.76924512e-01 -8.36529672e-01
-7.53284544e-02 2.25793749e-01 -2.38187373e-01 -1.16275363e-01
8.19603443e-01 -3.59577417e-01 -4.19800848e-01 8.33519697e-02
1.03390053e-01 4.94107902e-01 8.10939074e-03 6.20248020e-02
9.39858735e-01 -8.07825625e-02 4.99314040e-01 -2.07014680e-01
3.94247800e-01 -1.34988740e-01 1.27621377e+00 5.83059251e-01
-1.45081446e-01 5.80076158e-01 5.89609206e-01 -4.79462802e-01
-5.27358890e-01 -1.41543531e+00 -3.28325510e-01 1.01874042e+00
5.90922534e-01 -7.56617486e-01 -7.35892177e-01 -5.54759681e-01
1.14301279e-01 9.25967336e-01 -4.75168854e-01 7.33860582e-02
-9.29860234e-01 -5.24296999e-01 5.06444216e-01 8.80222619e-01
1.98674738e-01 -1.10063624e+00 -1.54576272e-01 5.27973354e-01
-6.46888733e-01 -1.22274339e+00 -3.92503232e-01 2.18064755e-01
-6.83450460e-01 -1.00341797e+00 -3.16640854e-01 -1.09729207e+00
4.74695235e-01 5.59386909e-01 1.22637606e+00 3.41275297e-02
1.33073717e-01 -1.41928261e-02 -1.50835350e-01 -4.17186111e-01
1.06879525e-01 2.69763827e-01 2.36952111e-01 -4.19785947e-01
1.22406685e+00 -5.79839110e-01 -3.58065873e-01 -2.51386613e-01
-7.85114408e-01 -1.36910081e-01 4.04398322e-01 1.05525398e+00
7.20006883e-01 2.35576415e-03 9.52481031e-01 -1.17473328e+00
9.70502377e-01 -5.89860916e-01 -3.37741673e-01 2.19383806e-01
-3.82228464e-01 1.42995626e-01 5.53032458e-01 -3.88614327e-01
-1.21955228e+00 -2.61380762e-01 -5.70844896e-02 -2.52509475e-01
4.87260446e-02 7.03579843e-01 -5.55267215e-01 5.23063123e-01
2.02260390e-01 1.23699948e-01 -4.08147186e-01 -7.82911420e-01
3.24246764e-01 6.97170675e-01 7.25784481e-01 -9.70191896e-01
4.50436354e-01 2.13837296e-01 -3.24957430e-01 -7.60013819e-01
-1.13170135e+00 -4.56292838e-01 -6.38173044e-01 4.60440606e-01
6.84822500e-01 -1.00139582e+00 -6.89924359e-01 5.68482876e-02
-1.57916188e+00 5.27308702e-01 -1.31056532e-01 2.54695147e-01
-2.39344284e-01 4.41993475e-01 -9.73495603e-01 -5.03493071e-01
-4.10065353e-01 -9.60175574e-01 9.50912178e-01 4.45220292e-01
-4.73719537e-01 -1.05101109e+00 -2.80865550e-01 3.39200556e-01
3.53654772e-01 -1.27518594e-01 1.39470959e+00 -6.81464672e-01
-6.02221787e-01 -7.50411302e-02 -2.79330879e-01 1.37982830e-01
4.19118047e-01 -5.15995502e-01 -7.89613187e-01 -3.83425690e-03
-1.60887033e-01 -3.74961570e-02 1.06522238e+00 1.13786809e-01
1.18468618e+00 -5.87714374e-01 -3.37269545e-01 6.20199978e-01
1.44033754e+00 3.61208878e-02 5.93182504e-01 6.17285132e-01
7.52071381e-01 4.55029011e-01 5.36327899e-01 3.29466015e-01
9.32591379e-01 6.80782557e-01 2.55296648e-01 1.78523988e-01
-1.43661842e-01 -4.11398023e-01 2.09234700e-01 1.05078959e+00
-1.56256720e-01 -3.25261392e-02 -7.66642988e-01 7.78213322e-01
-1.89153051e+00 -1.08851290e+00 7.15350062e-02 1.86397362e+00
1.32987070e+00 9.41894129e-02 -7.02368617e-01 -7.68216550e-02
7.99185634e-01 3.16795588e-01 -2.26642936e-01 -3.82060498e-01
-4.19129074e-01 3.97130132e-01 3.02849472e-01 5.48992276e-01
-1.35903776e+00 1.47951281e+00 6.11736870e+00 7.19511271e-01
-9.40028667e-01 1.25877028e-02 1.37144595e-01 3.00707538e-02
-7.22984612e-01 1.97335109e-02 -1.27014637e+00 4.46751028e-01
6.90801561e-01 -6.37613654e-01 3.53960633e-01 8.79995823e-01
-2.29510441e-01 3.66115004e-01 -1.18845677e+00 7.94625163e-01
7.62964562e-02 -1.62492955e+00 6.40637994e-01 -1.22722603e-01
5.69571495e-01 -3.26544881e-01 -5.88748977e-02 5.95094979e-01
1.09646715e-01 -1.21180832e+00 5.81889033e-01 6.14656866e-01
7.18633413e-01 -9.08596337e-01 9.66206074e-01 2.37086982e-01
-1.06870365e+00 1.28659174e-01 -9.81003106e-01 -3.29969406e-01
-2.03907974e-02 1.05251215e-01 -5.35340726e-01 3.63915324e-01
4.45462406e-01 7.36054063e-01 -1.21416055e-01 3.98464680e-01
-3.36073726e-01 5.44502974e-01 1.70006566e-02 -1.36839792e-01
4.04303402e-01 -9.22133997e-02 4.51144934e-01 1.54118073e+00
3.34760219e-01 4.97382075e-01 2.88890630e-01 1.02095520e+00
-5.32761812e-01 2.64816344e-01 -4.76074010e-01 -1.10029867e-02
1.16025233e+00 1.01989436e+00 -2.42289677e-01 -3.61546338e-01
-8.60581815e-01 8.61951053e-01 8.00092340e-01 4.76791888e-01
-5.70136070e-01 -6.27273560e-01 1.43816817e+00 1.06303087e-02
1.57390863e-01 -2.54388452e-01 -3.82804602e-01 -1.46816194e+00
1.89694121e-01 -6.58789396e-01 4.16878074e-01 -4.32235718e-01
-1.65649319e+00 4.98270780e-01 -2.64170617e-01 -9.59720969e-01
-2.86938339e-01 -6.37772322e-01 -4.35783982e-01 1.10358810e+00
-1.81509340e+00 -1.20801783e+00 -8.69218633e-02 5.21411300e-01
8.16707432e-01 -3.84277433e-01 1.43249512e+00 3.16323310e-01
-4.33014303e-01 9.01728630e-01 3.78218703e-02 6.00698829e-01
6.88712597e-01 -1.39209235e+00 3.46001893e-01 5.80805957e-01
1.92621291e-01 1.49529576e+00 3.96632403e-01 -5.78081071e-01
-1.62064528e+00 -9.43075955e-01 1.49048233e+00 -3.42793226e-01
7.47750700e-01 -1.69256255e-01 -1.33695078e+00 8.54996443e-01
2.64739603e-01 -1.29988700e-01 9.02614832e-01 6.00428045e-01
-5.62992930e-01 -3.40577029e-02 -6.85797572e-01 7.90243685e-01
1.09774864e+00 -7.19958127e-01 -1.46414578e+00 1.39796421e-01
9.28246498e-01 -4.92193997e-01 -6.51763141e-01 2.70523190e-01
3.82644445e-01 -3.73339742e-01 1.09731066e+00 -9.44470167e-01
4.97818530e-01 -1.71846077e-01 -7.24176347e-01 -1.03467858e+00
-3.75068307e-01 -1.58994123e-01 -1.61450610e-01 1.67767572e+00
2.65288800e-01 -3.67566913e-01 5.85244894e-01 7.46032357e-01
-2.11585000e-01 -3.74124527e-01 -1.23725069e+00 -5.77850103e-01
4.31204289e-01 -3.16597134e-01 8.57794464e-01 1.34774566e+00
3.01059484e-01 7.26431072e-01 -1.63412243e-01 1.39592931e-01
5.16878009e-01 3.22325319e-01 2.18217894e-01 -1.40817142e+00
1.82610884e-01 -4.39089715e-01 -5.20289660e-01 -1.17567253e+00
9.96489465e-01 -1.13268447e+00 -2.11406738e-01 -1.55044568e+00
4.14940774e-01 -5.91438532e-01 -8.10047746e-01 5.99647164e-01
-5.77071905e-01 -9.32696462e-02 1.48990452e-01 3.94106537e-01
-6.29811585e-01 5.07884920e-01 8.77252519e-01 -3.18999410e-01
-5.00187017e-02 -4.12573278e-01 -1.15106416e+00 9.96669114e-01
6.90587044e-01 -3.57251287e-01 -2.19373196e-01 -9.64080691e-01
3.28141525e-02 9.49550122e-02 1.92483574e-01 -5.02803266e-01
4.04221892e-01 -4.49105561e-01 3.04783583e-01 -5.49088120e-01
5.75771749e-01 -6.63120031e-01 -3.43641490e-01 -1.22773536e-01
-4.55156118e-01 1.60923883e-01 1.56409860e-01 5.65091074e-01
-5.53521156e-01 -3.89913529e-01 1.88287616e-01 -2.83068985e-01
-1.26465106e+00 -3.22336406e-02 -3.37595940e-01 2.29126826e-01
4.52525020e-01 1.75870731e-01 -4.65035468e-01 -1.41140580e-01
-4.47746724e-01 1.14451582e-02 3.33781123e-01 8.64694417e-01
7.90066004e-01 -1.68455970e+00 -5.95469177e-01 2.57895440e-01
3.51320729e-02 -1.60253033e-01 6.27765507e-02 -3.59467417e-02
-1.37402460e-01 6.39446557e-01 -1.81284368e-01 -2.70261526e-01
-1.18171144e+00 4.14354801e-01 1.67991742e-01 3.71240713e-02
-8.98024797e-01 6.45971477e-01 4.90965605e-01 -5.11033118e-01
4.15550947e-01 -2.23550677e-01 -6.77073359e-01 1.62857428e-01
7.56328762e-01 1.20907880e-01 4.34686951e-02 -7.99285173e-01
-6.72845304e-01 4.51463878e-01 -2.47117043e-01 1.06729180e-01
1.47076285e+00 -1.05027117e-01 -3.29418242e-01 4.46988374e-01
1.00046015e+00 3.50883067e-01 -6.43989146e-01 -6.69844806e-01
4.77265388e-01 -5.76685786e-01 3.62278000e-02 -6.17105544e-01
-8.45339119e-01 8.96736383e-01 3.80521454e-02 -2.66111434e-01
8.98938656e-01 -3.09315845e-02 1.09227800e+00 7.51484573e-01
5.13297260e-01 -1.22919846e+00 -2.78491050e-01 9.32920694e-01
7.92171001e-01 -1.07073760e+00 -2.66140699e-01 -5.37513733e-01
-6.75866663e-01 1.27812231e+00 7.51429617e-01 -1.27308220e-01
4.03377414e-01 4.62652266e-01 3.87574553e-01 1.97257288e-02
-9.35347795e-01 -2.38803640e-01 3.04356694e-01 5.87858319e-01
1.04297829e+00 4.48368341e-01 -8.35134447e-01 1.76954126e+00
-3.77152979e-01 -3.55017394e-01 5.98407745e-01 6.72984958e-01
-5.64833999e-01 -1.34840882e+00 -7.25753382e-02 2.05509812e-01
-8.97041857e-01 -4.70541000e-01 -3.39683384e-01 6.97007418e-01
4.39865924e-02 9.26626861e-01 2.47357443e-01 -1.26203269e-01
4.47575688e-01 1.31590247e-01 1.01134509e-01 -7.76052237e-01
-2.93350250e-01 -2.31191441e-02 2.17649505e-01 -4.56414461e-01
-8.87480453e-02 -7.60667145e-01 -1.86274445e+00 -5.69204271e-01
-1.98013484e-01 2.04350099e-01 5.04643857e-01 9.84669805e-01
2.70184964e-01 7.06780612e-01 2.26860940e-01 -2.46194556e-01
-8.26390982e-01 -7.90741324e-01 -5.85598826e-01 6.63911104e-01
1.57601357e-01 -8.33846033e-01 -5.50363436e-02 -1.36887863e-01] | [10.223319053649902, 9.231841087341309] |
9349e9e3-c8bc-40d9-86bb-e583758ee017 | online-gesture-recognition-using-transformer | 2305.03407 | null | https://arxiv.org/abs/2305.03407v1 | https://arxiv.org/pdf/2305.03407v1.pdf | Online Gesture Recognition using Transformer and Natural Language Processing | The Transformer architecture is shown to provide a powerful machine transduction framework for online handwritten gestures corresponding to glyph strokes of natural language sentences. The attention mechanism is successfully used to create latent representations of an end-to-end encoder-decoder model, solving multi-level segmentation while also learning some language features and syntax rules. The additional use of a large decoding space with some learned Byte-Pair-Encoding (BPE) is shown to provide robustness to ablated inputs and syntax rules. The encoder stack was directly fed with spatio-temporal data tokens potentially forming an infinitely large input vocabulary, an approach that finds applications beyond that of this work. Encoder transfer learning capabilities is also demonstrated on several languages resulting in faster optimisation and shared parameters. A new supervised dataset of online handwriting gestures suitable for generic handwriting recognition tasks was used to successfully train a small transformer model to an average normalised Levenshtein accuracy of 96% on English or German sentences and 94% in French. | ['M. Ramo', 'O. Akinremi', 'F. Balado', 'G. C. M. Silvestre'] | 2023-05-05 | null | null | null | null | ['gesture-recognition', 'handwriting-recognition'] | ['computer-vision', 'computer-vision'] | [ 7.84467876e-01 4.51162606e-01 -7.62272999e-02 -5.16749263e-01
-9.03161943e-01 -7.93266773e-01 7.62629926e-01 -5.41262984e-01
-5.52544475e-01 4.10704285e-01 1.80235311e-01 -4.44088072e-01
-5.72494641e-02 -4.57214594e-01 -6.97457671e-01 -7.83680260e-01
-1.81953743e-01 7.20797956e-01 1.18730359e-01 -1.66533068e-01
3.16545457e-01 6.76045954e-01 -1.35820627e+00 7.24450707e-01
3.53259414e-01 7.53738701e-01 6.90596938e-01 1.36846960e+00
-1.35255218e-01 9.02073622e-01 -7.19160676e-01 -6.09238386e-01
3.93070310e-01 -2.47813597e-01 -8.48399878e-01 3.41508150e-01
6.99975550e-01 -4.77658987e-01 -4.10989970e-01 3.94815326e-01
5.50309479e-01 -1.91997573e-01 6.35802984e-01 -5.43187678e-01
-7.75160968e-01 6.39269590e-01 2.93551409e-03 1.64404605e-02
3.72987241e-01 3.97365421e-01 1.14145660e+00 -6.15552962e-01
9.53144252e-01 1.14567900e+00 6.35399938e-01 7.01998949e-01
-1.04612291e+00 -2.51101911e-01 -5.36916256e-02 -1.20207660e-01
-7.19796002e-01 -1.48262396e-01 3.54790539e-01 -3.79551440e-01
1.75667524e+00 3.95682365e-01 6.51237249e-01 1.27951491e+00
3.74847144e-01 1.07177901e+00 1.12035918e+00 -5.07395744e-01
-8.68906677e-02 -3.76613401e-02 -2.49682404e-02 8.69420350e-01
-7.04788625e-01 1.49684474e-01 -6.83349311e-01 3.42522055e-01
1.14384651e+00 -3.19449663e-01 9.64666381e-02 -2.88252644e-02
-1.29498279e+00 6.11249208e-01 3.79160196e-01 4.35278744e-01
-5.04759178e-02 4.09916729e-01 5.55006921e-01 7.60263324e-01
1.03731826e-01 5.04908085e-01 -4.78559911e-01 -5.62306762e-01
-7.98711479e-01 4.44087461e-02 9.74469781e-01 1.09763288e+00
4.92951065e-01 6.57464191e-02 -2.71722496e-01 7.96108007e-01
1.39762178e-01 5.65937340e-01 6.33256316e-01 -8.33585620e-01
8.82802486e-01 5.71805298e-01 -2.53179312e-01 -2.86338151e-01
-2.59292245e-01 -1.47346497e-01 -5.19913971e-01 2.91507930e-01
6.47828281e-01 -9.89235193e-03 -1.25768280e+00 1.37172782e+00
-3.22652459e-01 -2.70242393e-01 2.08671272e-01 9.00602281e-01
2.12296113e-01 8.98829997e-01 -1.21195085e-01 1.49423238e-02
1.25610292e+00 -7.42536128e-01 -4.70918566e-01 -2.49618277e-01
5.45001388e-01 -8.15596282e-01 1.33623970e+00 6.15782678e-01
-1.25259686e+00 -5.18544495e-01 -9.17961717e-01 -3.69650394e-01
-2.57519901e-01 3.93151581e-01 8.71485472e-01 5.84859610e-01
-1.06452084e+00 7.98063934e-01 -1.03236449e+00 -3.09200585e-01
5.16683221e-01 8.05140138e-01 -3.59046072e-01 1.63239151e-01
-8.35054815e-01 9.71840084e-01 2.61473179e-01 2.77951986e-01
-1.01249230e+00 -4.54403788e-01 -7.67004967e-01 -1.82663858e-01
-1.44293576e-01 -1.95348665e-01 1.33189070e+00 -1.22564304e+00
-2.14530778e+00 9.68364179e-01 -9.66868829e-04 -6.64158940e-01
7.32791960e-01 -6.18547425e-02 2.20935028e-02 2.75304824e-01
-1.29048556e-01 9.21212196e-01 1.06680834e+00 -5.24521291e-01
-4.59425569e-01 -3.64376694e-01 -1.20488331e-01 3.00224692e-01
-3.51216733e-01 3.11608016e-01 -3.41822326e-01 -8.07503343e-01
-2.12456211e-02 -7.02421248e-01 -7.51087293e-02 -2.73339331e-01
-1.95056766e-01 -2.56222665e-01 8.56087565e-01 -1.09293687e+00
7.66340137e-01 -1.80038357e+00 7.03408837e-01 1.67876288e-01
-4.25990403e-01 2.41106629e-01 -4.07651633e-01 6.72659874e-01
1.72625422e-01 -2.32278362e-01 -4.76876855e-01 -2.57677734e-01
2.18353257e-01 6.69281185e-01 -8.55128467e-01 2.35089734e-01
6.52640581e-01 1.29035854e+00 -7.44278967e-01 -1.56130522e-01
2.97294587e-01 6.08375609e-01 -3.41783255e-01 1.90205589e-01
-4.51867014e-01 4.03942823e-01 -1.77848235e-01 7.76155889e-01
-1.95805728e-01 2.26989202e-03 1.29052654e-01 2.79050678e-01
4.59877029e-02 5.11824608e-01 -7.17311025e-01 2.13446760e+00
-7.79012263e-01 1.08080208e+00 -1.08685652e-02 -6.66801929e-01
1.02947199e+00 3.39773387e-01 1.56992063e-01 -7.57766902e-01
-1.56788563e-03 4.94124711e-01 4.98104841e-03 -7.19323516e-01
2.29912683e-01 -2.35843703e-01 -2.06125304e-01 5.37745714e-01
3.80627602e-01 -4.48219925e-01 1.80688620e-01 -2.53892597e-02
1.09042120e+00 7.07128406e-01 -3.85975510e-01 -1.05783813e-01
3.63180131e-01 6.81541637e-02 -2.90823672e-02 6.13199174e-01
2.36872151e-01 6.24542534e-01 5.47095299e-01 -5.59085011e-01
-1.40364373e+00 -9.11546826e-01 -7.50829726e-02 1.30043650e+00
-3.01065832e-01 -2.53387600e-01 -7.86012828e-01 -4.76805240e-01
-1.80633992e-01 4.65117395e-01 -3.69438320e-01 2.79340982e-01
-1.01055193e+00 -3.95679981e-01 1.08426380e+00 8.34711552e-01
3.71305734e-01 -1.40098107e+00 -1.03816414e+00 4.64687079e-01
2.39384264e-01 -9.64139879e-01 -3.05546403e-01 6.77488506e-01
-1.09679258e+00 -8.92456651e-01 -1.13463068e+00 -1.06784797e+00
5.55959821e-01 -8.28253627e-01 7.10764289e-01 -4.46334720e-01
-7.56487191e-01 4.59223151e-01 -2.01186180e-01 -1.69020250e-01
-7.68687487e-01 2.66948938e-01 -2.83995450e-01 -2.42919743e-01
3.21425885e-01 -4.89814490e-01 -1.10291056e-01 2.22043231e-01
-7.88405061e-01 1.95307910e-01 8.23189259e-01 1.22214949e+00
3.43296945e-01 -6.15273178e-01 1.98979944e-01 -4.91183132e-01
6.73786998e-01 3.65638822e-01 -7.08601654e-01 2.71591842e-01
-1.54483289e-01 2.97125190e-01 6.04468107e-01 -6.84758782e-01
-9.77373838e-01 3.54464173e-01 -2.91800588e-01 -3.54220867e-01
-2.70374715e-01 3.86556014e-02 -3.39491479e-02 6.81966394e-02
4.91375059e-01 4.46889848e-01 2.76023895e-01 -4.57049310e-01
3.93799633e-01 1.07741892e+00 7.36678004e-01 -5.26980579e-01
2.67486691e-01 9.05165151e-02 -4.32171710e-02 -9.55229461e-01
-7.45039731e-02 -2.08634615e-01 -1.05152905e+00 1.08501203e-01
8.65461528e-01 -5.65299511e-01 -6.40654027e-01 7.85218537e-01
-1.21996784e+00 -1.01763916e+00 -5.80634356e-01 2.10826129e-01
-1.13263059e+00 3.04893702e-01 -1.18957615e+00 -8.16290200e-01
-2.36173823e-01 -1.29324889e+00 1.61783278e+00 -2.31722191e-01
-3.40432763e-01 -9.95177746e-01 -1.37434408e-01 2.85005987e-01
2.42365956e-01 -4.86688269e-03 8.93785954e-01 -4.92524058e-01
-7.08427727e-01 -3.03165078e-01 5.76763637e-02 4.46696222e-01
8.08207244e-02 -1.13359168e-01 -1.18034697e+00 -4.07122672e-01
-1.38562799e-01 -5.97846150e-01 9.94779885e-01 -8.55451152e-02
8.80544782e-01 -3.55772018e-01 -5.18209562e-02 5.83091438e-01
1.11498475e+00 2.36817300e-01 8.35631490e-01 1.60840347e-01
5.14291465e-01 6.77084386e-01 2.73281187e-01 4.39639464e-02
-1.56747237e-01 6.89301491e-01 1.82979852e-01 1.53629586e-01
-3.06747586e-01 -3.13272625e-01 8.63336682e-01 7.29485512e-01
-2.62314826e-01 -1.17387511e-01 -1.10676670e+00 5.10888875e-01
-1.51519966e+00 -8.23216915e-01 1.91554651e-01 1.87140572e+00
8.97827625e-01 5.55076897e-02 2.85857432e-02 2.14567110e-01
3.15697670e-01 4.67016846e-02 -4.39357519e-01 -8.37191522e-01
-2.22631946e-01 6.98517561e-01 8.49720836e-01 7.88168609e-01
-1.13709664e+00 1.34898722e+00 6.87644720e+00 7.22620845e-01
-1.29576886e+00 -1.13576226e-01 4.13998634e-01 -1.24079935e-01
-1.49248481e-01 -2.22656712e-01 -8.20249438e-01 3.57599556e-01
1.04664683e+00 5.64261973e-01 7.88494289e-01 4.11774695e-01
-1.91966265e-01 1.93157136e-01 -1.31232440e+00 9.35776293e-01
9.97879654e-02 -1.30904150e+00 1.36081010e-01 2.47596353e-01
4.19970602e-01 1.72001705e-01 1.45829275e-01 3.71070430e-02
1.64859027e-01 -1.41764033e+00 6.95990920e-01 4.92402524e-01
1.39262128e+00 -3.93804908e-01 4.27126467e-01 4.02193367e-01
-7.87511945e-01 -4.28281993e-01 -2.55689323e-01 -2.49931648e-01
2.23723233e-01 -3.01059067e-01 -1.13678300e+00 1.46402672e-01
2.55928904e-01 8.36164057e-01 -4.14916277e-01 3.68664503e-01
-3.49452376e-01 6.88872337e-01 -5.21752954e-01 -3.25598747e-01
7.07571626e-01 -6.18220307e-02 5.48282385e-01 1.73777461e+00
1.53384775e-01 -1.09280042e-01 -3.86462241e-01 7.14428365e-01
2.71209121e-01 -1.44084513e-01 -8.29325855e-01 -4.06749338e-01
-1.64192319e-01 8.67820919e-01 -6.62222028e-01 -1.76117212e-01
-2.61394605e-02 1.59625900e+00 9.49742272e-02 4.45041567e-01
-4.95831937e-01 -5.16692638e-01 3.91285866e-01 -3.09355911e-02
5.35049260e-01 -4.31631446e-01 -2.86031067e-01 -1.11259544e+00
1.46927550e-01 -1.01153576e+00 -4.22899462e-02 -7.22710907e-01
-1.00363803e+00 5.59990644e-01 -1.74244076e-01 -8.27245772e-01
-7.86920726e-01 -1.45956135e+00 -3.73805881e-01 1.14975202e+00
-1.00635445e+00 -1.50637567e+00 8.92315432e-02 4.81233597e-01
9.10090983e-01 -7.77613521e-01 1.08457220e+00 -5.02797142e-02
-1.10480428e-01 5.78427255e-01 -9.89423692e-03 3.87818545e-01
2.94848770e-01 -1.54584467e+00 7.74106383e-01 6.81685805e-01
5.91173112e-01 5.86000204e-01 4.00135964e-01 -4.84204352e-01
-1.73143303e+00 -7.07759142e-01 1.04880011e+00 -7.04428375e-01
7.02161074e-01 -7.44930685e-01 -6.36687517e-01 9.41155732e-01
4.02673364e-01 -4.56257492e-01 4.14431363e-01 -3.11612099e-01
-2.83873737e-01 1.09632485e-01 -7.97038972e-01 5.39694011e-01
1.22233462e+00 -9.21679258e-01 -8.98633659e-01 4.32402402e-01
2.53650367e-01 -8.57043684e-01 -8.25303555e-01 1.81702942e-01
8.20373714e-01 -7.04656243e-01 7.25792766e-01 -8.58068407e-01
6.42778993e-01 1.37478977e-01 -2.26687104e-01 -9.05697882e-01
9.24138054e-02 -1.05997431e+00 -1.29269481e-01 9.27056909e-01
4.57567573e-01 -5.37735939e-01 8.62313092e-01 4.92349505e-01
-1.45364136e-01 -8.35942805e-01 -1.17850709e+00 -6.76042974e-01
3.02203715e-01 -6.13153398e-01 2.54418552e-01 3.46539140e-01
1.65256903e-01 3.36385697e-01 -2.26512089e-01 -2.39002228e-01
2.33177841e-01 1.68461472e-01 6.59102142e-01 -7.22106516e-01
-6.47625864e-01 -6.21599913e-01 -7.74328828e-01 -1.63398862e+00
2.38337502e-01 -1.19686127e+00 -7.12736025e-02 -1.30513644e+00
-1.87769040e-01 -2.76780576e-01 2.29493454e-02 7.82660246e-01
4.49808985e-01 3.32931489e-01 3.09280187e-01 2.69299209e-01
3.98612488e-03 1.47798672e-01 1.17892337e+00 -4.41297561e-01
-1.12751082e-01 -7.35477507e-02 -2.87948269e-02 3.57253253e-01
6.07927620e-01 9.28555988e-03 -3.71383667e-01 -7.90433228e-01
1.67103216e-01 -3.82707082e-02 5.43626130e-01 -6.96025014e-01
1.68846324e-01 2.48946890e-01 3.49715978e-01 -4.11181271e-01
4.94109094e-01 -7.64715016e-01 -1.45949632e-01 6.11787319e-01
-7.40602911e-01 1.65600143e-02 3.36999625e-01 3.42230886e-01
-2.44325250e-01 -2.43125871e-01 3.50484580e-01 -3.12940657e-01
-9.08343256e-01 -2.46102646e-01 -5.67459643e-01 -3.61098140e-01
9.18226957e-01 -6.30992174e-01 6.48639947e-02 -2.09048718e-01
-9.72769737e-01 -1.95831675e-02 2.81180799e-01 5.37014484e-01
6.87476933e-01 -9.96778071e-01 -6.83231354e-01 7.47626781e-01
-1.48102552e-01 -1.77670076e-01 -3.71326923e-01 6.18565321e-01
-1.02369165e+00 8.55252028e-01 -4.71051425e-01 -8.91720235e-01
-1.43543184e+00 7.72134662e-02 6.21691108e-01 -3.05099219e-01
-9.52976644e-01 1.13125288e+00 -2.73700237e-01 -5.70778728e-01
5.31287968e-01 -8.12311471e-01 3.97346884e-01 -5.28324395e-02
4.06880528e-01 8.82300958e-02 6.35619760e-02 -4.74202812e-01
-3.72590274e-02 7.75145411e-01 6.36727661e-02 -5.36134839e-01
1.40209460e+00 3.50834638e-01 -3.28760743e-02 5.00890017e-01
1.12328553e+00 -2.58693099e-01 -1.74665105e+00 1.79104358e-01
2.23424017e-01 -1.65498167e-01 -3.06367576e-01 -1.10550785e+00
-6.41178906e-01 1.42510116e+00 6.24184728e-01 -1.54362932e-01
9.91827667e-01 -3.62014249e-02 8.18547308e-01 8.43295515e-01
5.87673366e-01 -1.22775912e+00 4.44234014e-02 8.54219496e-01
1.11027074e+00 -9.59914386e-01 -5.37197292e-01 -9.85216871e-02
-8.20594966e-01 1.67298079e+00 3.82875465e-02 -2.52641350e-01
-9.59710702e-02 8.27102184e-01 2.38670662e-01 -7.98752233e-02
-6.24123931e-01 -5.33923730e-02 1.98519051e-01 6.29527032e-01
5.39913237e-01 -7.62191713e-02 3.41954492e-02 1.25355721e-01
-2.34257489e-01 6.97713271e-02 3.76453362e-02 9.08443153e-01
-2.05124356e-02 -1.42899060e+00 -1.66817456e-02 3.21690500e-01
-3.95101517e-01 -4.74857502e-02 -8.47693622e-01 8.31551552e-01
-6.39743060e-02 2.94694662e-01 4.36501950e-01 -2.80207098e-01
1.12256616e-01 5.26580036e-01 1.04139650e+00 -6.17641151e-01
-1.15378475e+00 9.62223336e-02 2.59599313e-02 -5.94852030e-01
-1.97023749e-01 -6.64006233e-01 -1.28547382e+00 4.48680013e-01
-1.05957985e-01 -2.76687622e-01 7.49033153e-01 9.34686482e-01
9.72763896e-02 4.95890021e-01 2.08571792e-01 -9.82897222e-01
-9.34295952e-01 -1.12749684e+00 -2.57649392e-01 3.41027796e-01
3.73566717e-01 -4.16224338e-02 -1.11851934e-02 4.15015399e-01] | [9.231735229492188, -6.5418291091918945] |
2ddb1cf5-32cb-4e61-882b-abb4fd4ffd81 | toward-an-intelligent-tutoring-system-for | 2210.13635 | null | https://arxiv.org/abs/2210.13635v1 | https://arxiv.org/pdf/2210.13635v1.pdf | Toward an Intelligent Tutoring System for Argument Mining in Legal Texts | We propose an adaptive environment (CABINET) to support caselaw analysis (identifying key argument elements) based on a novel cognitive computing framework that carefully matches various machine learning (ML) capabilities to the proficiency of a user. CABINET supports law students in their learning as well as professionals in their work. The results of our experiments focused on the feasibility of the proposed framework are promising. We show that the system is capable of identifying a potential error in the analysis with very low false positives rate (2.0-3.5%), as well as of predicting the key argument element type (e.g., an issue or a holding) with a reasonably high F1-score (0.74). | ['Karim Benyekhlef', 'Kevin D. Ashley', 'Vern R. Walker', 'Jaromir Savelka', 'Hannes Westermann'] | 2022-10-24 | null | null | null | null | ['argument-mining'] | ['natural-language-processing'] | [ 1.73268721e-01 4.06880528e-01 -2.84970757e-02 -4.65516932e-02
-8.27728689e-01 -8.07660818e-01 7.19086885e-01 9.18928564e-01
-3.82788301e-01 8.34293425e-01 -2.72473127e-01 -1.18396461e+00
-6.40228510e-01 -1.02996528e+00 -4.58160818e-01 -3.40024024e-01
6.07508540e-01 5.56666970e-01 4.66805905e-01 -4.02274281e-01
1.09994459e+00 7.13033319e-01 -1.78474283e+00 6.27828181e-01
1.13311684e+00 6.94954634e-01 4.30706143e-01 7.87078440e-01
-1.08032532e-01 1.21663547e+00 -9.53154027e-01 -7.81602144e-01
-1.04850933e-01 1.27780974e-01 -1.19730914e+00 -5.05265474e-01
2.37932906e-01 -1.34935647e-01 3.94435614e-01 1.29319763e+00
3.89393002e-01 -4.47285362e-02 5.72289646e-01 -1.19423616e+00
-4.31865215e-01 5.90839148e-01 -3.35271537e-01 6.88777506e-01
5.07021785e-01 -3.87150139e-01 8.65176916e-01 -9.09894347e-01
5.80508947e-01 9.74792898e-01 5.42815745e-01 -3.78963277e-02
-6.32994592e-01 -6.14293635e-01 -1.80944592e-01 6.91239417e-01
-1.06861830e+00 -3.77496600e-01 8.98522973e-01 -7.78315246e-01
9.15363908e-01 3.49674910e-01 4.58186328e-01 7.48836040e-01
5.63718677e-01 5.22208869e-01 1.27253723e+00 -1.03321373e+00
4.11567777e-01 7.86050856e-01 7.94555902e-01 6.69713736e-01
3.32386941e-01 -3.15720767e-01 -3.60558629e-01 -5.04205883e-01
3.24323148e-01 -3.18786323e-01 4.94686067e-01 3.06290627e-01
-6.71367526e-01 9.49445963e-01 -2.29511261e-01 8.16893876e-01
-6.32988393e-01 -5.08610070e-01 3.65252852e-01 2.18663022e-01
1.26049602e-02 4.16862398e-01 -3.70730758e-01 -4.31867123e-01
-4.96291161e-01 5.75320661e-01 7.76026726e-01 7.92209864e-01
2.03249738e-01 -3.29029888e-01 -4.32719022e-01 5.99714279e-01
1.71343938e-01 3.14160585e-01 3.40521812e-01 -7.14241326e-01
5.55189252e-01 1.06001449e+00 2.74040878e-01 -1.16570520e+00
-2.89960325e-01 -4.48914737e-01 -2.04015911e-01 3.87911707e-01
4.39563125e-01 -9.88979787e-02 -1.21956788e-01 1.31449533e+00
4.38792408e-01 -6.43466925e-03 2.94220716e-01 2.06345022e-01
7.10905254e-01 3.88151467e-01 2.35705227e-01 -4.54740763e-01
1.73500693e+00 -4.53555286e-01 -9.71632838e-01 2.17365343e-02
5.37085474e-01 -1.05746841e+00 1.17202628e+00 6.75860763e-01
-1.16355562e+00 -5.58093727e-01 -7.79260576e-01 9.25434604e-02
-3.62346590e-01 3.49634290e-01 8.39999795e-01 8.83797348e-01
-5.66409051e-01 2.66427577e-01 -2.97931939e-01 -1.01025395e-01
1.64697826e-01 1.33590221e-01 -4.18981276e-02 2.68290639e-01
-1.10612297e+00 9.55459833e-01 4.51206326e-01 -1.57191500e-01
-5.05996823e-01 -7.32640803e-01 -2.17259839e-01 2.35069484e-01
6.33544445e-01 -1.29607469e-01 1.17697227e+00 -4.16542262e-01
-9.85753953e-01 9.39500153e-01 -1.13101043e-02 -2.90188998e-01
7.35739410e-01 -2.25472927e-01 -5.36652863e-01 2.22260743e-01
2.69460559e-01 -4.39132482e-01 4.40818459e-01 -9.16369319e-01
-1.09551609e+00 -3.76833111e-01 7.20583439e-01 -8.11733603e-02
-4.39586073e-01 4.88305449e-01 4.21746969e-01 -4.74798232e-01
-1.04105836e-02 -7.10147262e-01 9.52753797e-02 -6.03949189e-01
-5.34957089e-02 -7.85742819e-01 7.69012690e-01 -7.29268014e-01
1.55453873e+00 -1.62488043e+00 -5.89685082e-01 5.75511575e-01
9.23882872e-02 5.84057450e-01 6.55541778e-01 4.97060806e-01
-1.34146109e-01 2.33187690e-01 2.58609653e-01 4.78721291e-01
-1.49835601e-01 -1.50293440e-01 -2.40656137e-01 -1.14462629e-01
-1.76133454e-01 6.05765224e-01 -8.52396429e-01 -7.82090008e-01
9.64897349e-02 -1.19565744e-02 -4.52331603e-01 3.33840519e-01
1.77068651e-01 -1.76383585e-01 -8.70403230e-01 7.11741507e-01
4.33249384e-01 -1.50740266e-01 4.31073308e-01 -5.49778491e-02
-2.21665159e-01 3.35349619e-01 -1.32120872e+00 8.60374212e-01
-5.15337706e-01 5.76809943e-01 2.55705826e-02 -1.21171844e+00
1.02705169e+00 5.75690448e-01 1.26443207e-01 -4.25666630e-01
1.39460683e-01 1.03466801e-01 1.54968664e-01 -1.04985070e+00
3.71622950e-01 -3.96059640e-02 1.81417763e-01 7.82498002e-01
-1.68890715e-01 5.21918535e-01 3.42091262e-01 2.20260233e-01
1.11184573e+00 -3.38073045e-01 7.41840005e-01 -5.61022639e-01
1.03391469e+00 6.45807087e-02 4.10378456e-01 9.73627567e-01
-1.96544468e-01 -4.26674366e-01 6.31374180e-01 -4.30667460e-01
-8.05192471e-01 -6.55811965e-01 -1.54929549e-01 1.11560464e+00
-1.87643632e-01 -2.69358635e-01 -9.40917432e-01 -7.01727509e-01
-1.63029552e-01 1.00793946e+00 -3.05238217e-01 6.90315962e-02
-4.82792675e-01 -4.07971412e-01 3.92698586e-01 3.77346724e-01
5.30477107e-01 -1.27793193e+00 -9.56127226e-01 2.98741490e-01
-3.08426619e-01 -1.09250402e+00 3.89934748e-01 1.30833253e-01
-6.92432046e-01 -1.44651020e+00 3.82180363e-01 -9.27663386e-01
6.21112585e-01 6.04199916e-02 8.80492210e-01 5.48017085e-01
-2.26958945e-01 2.50087500e-01 -5.12948453e-01 -8.78073514e-01
-8.05488884e-01 -1.78387910e-01 -1.36264175e-01 -2.41498142e-01
6.85159564e-01 -4.60428834e-01 8.78740028e-02 1.48268333e-02
-6.26397252e-01 -3.93450931e-02 3.44857335e-01 6.65135682e-01
1.63221136e-01 5.99859357e-01 9.30577517e-01 -1.29652274e+00
1.60417652e+00 -3.68248075e-01 -7.95534432e-01 9.57472682e-01
-1.16050684e+00 -3.54317456e-01 5.97374260e-01 -4.54340726e-01
-1.33097780e+00 -4.40577269e-01 -2.21179485e-01 3.36141855e-01
-4.54475820e-01 4.47901279e-01 -3.73179913e-02 -7.24566504e-02
8.51540625e-01 -1.93861425e-01 -2.71798432e-01 -3.66354734e-01
-3.56236339e-01 1.09820354e+00 5.17636657e-01 -1.04019606e+00
6.69338048e-01 -2.77486205e-01 6.80070221e-02 -4.11429107e-01
-8.60522032e-01 -3.03283215e-01 -5.60975373e-01 -6.80420041e-01
4.06772494e-01 -5.10217905e-01 -1.29991674e+00 1.71248068e-03
-1.10499918e+00 3.64210695e-01 1.78761154e-01 6.82633817e-01
-3.09550911e-01 2.75439650e-01 -5.47498345e-01 -1.36826849e+00
-5.74167252e-01 -1.07394862e+00 4.20011640e-01 2.83831567e-01
-3.56716573e-01 -9.24922585e-01 -2.28400245e-01 1.16306460e+00
9.00059119e-02 2.13624224e-01 1.51535869e+00 -1.12459219e+00
-2.52896659e-02 -3.93254697e-01 2.96033118e-02 1.07360415e-01
-1.87439725e-01 8.90454426e-02 -1.07401097e+00 9.59138647e-02
3.59994859e-01 -2.06226185e-01 1.97266545e-02 1.53027609e-01
1.02709627e+00 -4.76998746e-01 -2.48735711e-01 -3.73298973e-01
1.19495690e+00 8.62274766e-01 5.36866069e-01 7.59702623e-01
4.69726771e-02 5.73561192e-01 1.15334797e+00 3.85648310e-01
1.16477050e-01 5.97263932e-01 2.01518387e-01 2.56700397e-01
3.23157400e-01 -2.23412916e-01 -4.10145111e-02 2.74555296e-01
-4.00710911e-01 -8.18320513e-02 -1.27693903e+00 2.84013331e-01
-2.00596857e+00 -1.16928184e+00 -3.80070686e-01 1.79128313e+00
7.05501199e-01 6.89988017e-01 1.11908168e-01 9.21590626e-01
9.14122701e-01 -6.36929095e-01 1.36838779e-02 -8.64726663e-01
3.15334320e-01 3.41049433e-01 -7.85277262e-02 7.31671989e-01
-8.56170535e-01 7.41254091e-01 5.92623138e+00 9.81827617e-01
-5.23110390e-01 1.86734945e-01 5.24192870e-01 4.66562569e-01
-2.94611722e-01 -5.65031683e-03 -7.74685085e-01 4.06569034e-01
8.14052939e-01 -5.84861040e-01 2.48714447e-01 1.10564125e+00
1.39480829e-01 -1.71403199e-01 -6.73067749e-01 6.59010708e-01
2.29827594e-02 -1.61369061e+00 -1.52796373e-01 1.05058357e-01
2.54903048e-01 -9.25883770e-01 -2.71231472e-01 4.07778472e-01
7.45326281e-02 -5.30644059e-01 9.36492264e-01 4.07760054e-01
1.31482199e-01 -9.61972594e-01 1.18751931e+00 1.08635223e+00
-7.01777101e-01 -4.35543835e-01 -2.00592175e-01 -4.08443689e-01
-2.89621353e-01 6.34229779e-01 -1.40833080e+00 3.66734177e-01
5.72618604e-01 -5.61941490e-02 -6.17444396e-01 6.16925359e-01
-3.86600971e-01 7.29359448e-01 -7.91648850e-02 -4.77526009e-01
-8.74460861e-02 3.39148231e-02 4.19344753e-01 1.09586465e+00
3.08743209e-01 6.27976656e-01 -4.99279089e-02 5.78575373e-01
1.75414652e-01 5.17410576e-01 -5.51148832e-01 1.33183703e-01
9.41096783e-01 1.25738823e+00 -9.18519020e-01 -5.19972742e-01
2.74890456e-02 4.41425629e-02 2.98213542e-01 -1.59915179e-01
-5.69486976e-01 -4.34714168e-01 3.39024402e-02 3.20327163e-01
1.28643238e-03 3.58023226e-01 -4.00709629e-01 -5.86119235e-01
2.60415375e-01 -1.02998185e+00 5.61915398e-01 -7.79821932e-01
-1.01345110e+00 6.30187154e-01 1.92769632e-01 -9.16074574e-01
-3.48472357e-01 -7.56393671e-01 -7.37869680e-01 7.31836200e-01
-1.10119414e+00 -1.13738430e+00 -1.60013467e-01 5.93782544e-01
2.74507374e-01 -8.10257792e-01 8.97732854e-01 2.12848276e-01
-4.05767709e-01 5.45550823e-01 -3.89878929e-01 3.12551558e-02
2.21586928e-01 -1.04420304e+00 -4.17922646e-01 8.21121871e-01
3.54047678e-03 9.96810496e-01 6.68894768e-01 -6.62665606e-01
-1.05885863e+00 -3.71567398e-01 1.62141466e+00 -4.66678470e-01
6.50799096e-01 -1.27097070e-01 -7.44242966e-01 2.31098205e-01
-7.49989077e-02 -4.49072957e-01 9.35476005e-01 4.11816955e-01
-1.01128466e-01 -2.94310693e-02 -1.69094324e+00 3.38845611e-01
6.68501019e-01 -6.37442887e-01 -1.16994822e+00 6.61543190e-01
2.16618925e-01 -2.52313763e-01 -9.12268639e-01 2.32426077e-01
6.87823176e-01 -1.22750652e+00 9.30547118e-01 -9.66995239e-01
2.84621924e-01 -9.24685672e-02 -1.76932551e-02 -3.97533000e-01
-4.85946983e-01 -3.62706721e-01 -1.68322012e-01 1.39334416e+00
5.01821458e-01 -5.05322397e-01 6.36435628e-01 9.14749861e-01
2.87142135e-02 -7.72334695e-01 -7.40621328e-01 -4.94330168e-01
-2.75021136e-01 -6.52833641e-01 6.21243298e-01 1.13357675e+00
4.27414507e-01 4.57988858e-01 9.81520787e-02 2.90813416e-01
3.07251036e-01 3.97791982e-01 4.29579526e-01 -1.79693770e+00
-9.35472101e-02 -2.72233456e-01 -3.41541022e-01 -1.24113448e-01
4.10292596e-01 -8.70742381e-01 -5.82390487e-01 -1.34381950e+00
3.34924400e-01 -5.16607225e-01 -4.29295212e-01 6.19493544e-01
-3.47166389e-01 -3.37069720e-01 6.99768309e-04 -2.36312486e-02
-6.57291114e-02 -4.21230644e-01 7.49762833e-01 1.19468428e-01
-2.31642559e-01 4.16919082e-01 -1.14701974e+00 1.12543023e+00
1.01865089e+00 -5.48349440e-01 -4.23000872e-01 1.81805920e-02
5.31242073e-01 2.88821459e-01 1.79936379e-01 -7.70282269e-01
4.87195790e-01 -5.69981277e-01 1.97405204e-01 -6.07350945e-01
-2.77265221e-01 -1.13149643e+00 4.14098985e-02 6.02312684e-01
-6.46254838e-01 4.63593066e-01 7.79682994e-02 2.22616330e-01
7.11096972e-02 -1.09754658e+00 3.43072802e-01 -9.69631504e-03
-4.99980569e-01 -6.52008712e-01 -3.07613909e-01 -2.64895558e-01
1.38656926e+00 -1.54176399e-01 -6.19797707e-01 -1.22791797e-01
-4.41273361e-01 -1.39739886e-01 -1.41631112e-01 2.62706339e-01
4.96854544e-01 -9.34698343e-01 -5.78484416e-01 1.68935299e-01
1.55765470e-02 -6.34300649e-01 -1.56922281e-01 5.44189095e-01
-5.75906754e-01 5.39579511e-01 -3.95990342e-01 -5.60073256e-02
-1.87568533e+00 4.17455196e-01 -1.80280060e-01 -6.06836915e-01
-4.60169643e-01 5.75233340e-01 -6.33652687e-01 -5.32657169e-02
3.82991791e-01 1.21273339e-01 -9.09279108e-01 1.58438340e-01
7.26324677e-01 7.13349879e-01 4.99030650e-01 -3.36263210e-01
-4.65914577e-01 2.11599827e-01 -1.68135062e-01 3.73322144e-02
1.36246121e+00 4.00050998e-01 -1.67066842e-01 1.56760201e-01
3.99008125e-01 6.37888551e-01 -1.77938417e-01 -9.61593091e-02
2.96477228e-01 -5.66847503e-01 1.12160251e-01 -1.15970290e+00
-3.96527588e-01 6.22192979e-01 5.90629339e-01 5.93969882e-01
8.98132443e-01 -2.52882332e-01 1.68580830e-01 9.00849640e-01
4.71160740e-01 -1.46867037e+00 -1.09784506e-01 3.65509152e-01
7.75835395e-01 -1.09307110e+00 3.54512691e-01 -5.15315890e-01
-5.32342017e-01 1.40259850e+00 5.28429329e-01 2.54786491e-01
7.00898230e-01 5.93035042e-01 4.91751656e-02 -4.15539473e-01
-9.89375055e-01 1.60659924e-01 1.37440428e-01 4.53150451e-01
6.70937955e-01 5.44358082e-02 -9.49119568e-01 8.44529033e-01
-4.11143899e-01 1.29485324e-01 5.96356094e-01 1.27932966e+00
-7.76733637e-01 -1.06556475e+00 -8.01684916e-01 4.21011090e-01
-8.96236598e-01 1.30325690e-01 -6.42225683e-01 7.75469303e-01
4.38826352e-01 1.41179609e+00 -1.57364801e-01 -3.62104446e-01
5.51518738e-01 1.88201040e-01 2.43230641e-01 -4.19856906e-01
-1.03735888e+00 -3.09230201e-03 3.55568916e-01 -1.31815791e-01
-4.12644029e-01 -5.32473743e-01 -1.14001644e+00 -4.49075043e-01
-4.61033344e-01 6.44030511e-01 7.29534924e-01 1.05593538e+00
-4.55392599e-02 4.49275404e-01 5.32550633e-01 1.90569177e-01
-4.69625145e-01 -1.03917241e+00 -3.95783007e-01 3.27943146e-01
-1.11109301e-01 -8.15505683e-01 -1.73995972e-01 1.44517735e-01] | [9.733309745788574, 9.048041343688965] |
aae4e2d4-5e95-481e-a7c3-af1d593310a0 | realtabformer-generating-realistic-relational | 2302.02041 | null | https://arxiv.org/abs/2302.02041v1 | https://arxiv.org/pdf/2302.02041v1.pdf | REaLTabFormer: Generating Realistic Relational and Tabular Data using Transformers | Tabular data is a common form of organizing data. Multiple models are available to generate synthetic tabular datasets where observations are independent, but few have the ability to produce relational datasets. Modeling relational data is challenging as it requires modeling both a "parent" table and its relationships across tables. We introduce REaLTabFormer (Realistic Relational and Tabular Transformer), a tabular and relational synthetic data generation model. It first creates a parent table using an autoregressive GPT-2 model, then generates the relational dataset conditioned on the parent table using a sequence-to-sequence (Seq2Seq) model. We implement target masking to prevent data copying and propose the $Q_{\delta}$ statistic and statistical bootstrapping to detect overfitting. Experiments using real-world datasets show that REaLTabFormer captures the relational structure better than a baseline model. REaLTabFormer also achieves state-of-the-art results on prediction tasks, "out-of-the-box", for large non-relational datasets without needing fine-tuning. | ['Olivier Dupriez', 'Aivin V. Solatorio'] | 2023-02-04 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 2.11763501e-01 6.75691009e-01 -2.70248771e-01 -7.62930274e-01
-1.33982289e+00 -6.76887333e-01 5.59241593e-01 1.97811142e-01
3.75648022e-01 9.96581554e-01 2.67987520e-01 -6.00181162e-01
-3.45314387e-03 -1.23270988e+00 -1.48740125e+00 -3.82959068e-01
-3.58085483e-02 1.32663131e+00 1.68035477e-01 -2.26249725e-01
4.25727367e-02 -9.41319112e-03 -1.62196052e+00 1.21631825e+00
9.62534785e-01 7.57796526e-01 1.81246884e-02 7.38259017e-01
-2.28052914e-01 1.40691268e+00 -7.87911952e-01 -8.65647554e-01
4.81498331e-01 -2.87471175e-01 -5.25254846e-01 -3.90977532e-01
3.75680566e-01 9.88573357e-02 -2.61871994e-01 4.74770069e-01
4.28239137e-01 -5.45188971e-02 5.64520955e-01 -1.45317519e+00
-7.96026528e-01 1.31270063e+00 -4.22043562e-01 -2.53598779e-01
4.20818627e-01 1.65697262e-01 1.08986843e+00 -8.18700016e-01
8.84004712e-01 1.61568844e+00 9.78465855e-01 2.20020592e-01
-2.21235251e+00 -7.14273095e-01 -6.10874444e-02 -4.33950365e-01
-1.25009382e+00 -3.43484253e-01 1.26323491e-01 -6.00919724e-01
1.22556031e+00 5.26234925e-01 2.78347671e-01 1.48354173e+00
5.10910690e-01 7.52728522e-01 7.63514102e-01 4.06418368e-02
1.68499619e-01 1.79052994e-01 -5.02095884e-03 2.54663050e-01
3.06452513e-01 1.54991046e-01 -6.97118819e-01 -4.16054100e-01
5.21633983e-01 -1.33647397e-01 3.64506960e-01 -6.71499312e-01
-1.44548869e+00 5.44753969e-01 1.86013594e-01 -5.94581723e-01
-3.16036373e-01 3.77277434e-01 5.71071506e-01 3.55702370e-01
3.72795254e-01 4.23860371e-01 -7.87905693e-01 -3.90570134e-01
-6.65659010e-01 8.91989529e-01 9.78118598e-01 1.76917136e+00
5.67182362e-01 5.60124554e-02 -5.92506707e-01 9.47026908e-01
-1.34486230e-02 6.58581734e-01 1.97423056e-01 -8.58170331e-01
1.09536648e+00 7.10005164e-01 1.15783036e-01 -6.07166469e-01
-2.65136480e-01 4.74550799e-02 -8.51347685e-01 -2.34675556e-01
3.45994055e-01 -7.63347894e-02 -9.20038223e-01 1.46190155e+00
2.69726098e-01 -4.61782694e-01 2.32892483e-01 1.35718703e-01
8.28813314e-01 9.26037431e-01 9.22548622e-02 7.79702440e-02
1.07494938e+00 -6.85819268e-01 -4.38669473e-01 -4.09592777e-01
7.87962675e-01 -5.16772926e-01 1.39992559e+00 6.01970375e-01
-1.08017683e+00 -7.28035092e-01 -6.35497689e-01 -3.77307594e-01
-5.00286996e-01 1.45747568e-02 7.22759485e-01 6.74448371e-01
-8.80828857e-01 4.73232418e-01 -8.50475311e-01 -8.60235617e-02
2.56690919e-01 8.04520696e-02 -5.69007754e-01 2.85208877e-02
-1.06165695e+00 4.80646133e-01 6.96503520e-01 -8.79936144e-02
-6.54534340e-01 -1.26179338e+00 -8.91449928e-01 5.48666865e-02
5.72115123e-01 -7.31811404e-01 1.32757342e+00 -1.19617477e-01
-8.81562769e-01 5.67157924e-01 -3.31360877e-01 -7.77583599e-01
5.59317768e-01 -1.85727939e-01 -2.87973911e-01 -7.04857349e-01
1.69338316e-01 7.32511878e-01 3.35167259e-01 -1.38586056e+00
-3.70265275e-01 -2.51867741e-01 -5.65697849e-01 -4.94162403e-02
5.55986702e-01 -1.27540588e-01 -5.83073139e-01 -9.88281965e-01
8.79670680e-02 -9.83103335e-01 -2.96784431e-01 -5.62724710e-01
-1.07306254e+00 1.33010298e-01 2.28278160e-01 -5.72829008e-01
1.32811987e+00 -1.88098490e+00 -2.31583193e-02 1.12664036e-01
-4.91334759e-02 -3.15426499e-01 -3.68112810e-02 8.83701801e-01
-4.78594393e-01 5.09689808e-01 -1.62681431e-01 -2.75552392e-01
1.05868809e-01 3.60136092e-01 -7.26350725e-01 -3.30861896e-01
2.71608442e-01 1.06213522e+00 -6.54834211e-01 -4.37081873e-01
-2.22114980e-01 1.64494857e-01 -9.45750892e-01 3.05592775e-01
-9.45596159e-01 -1.85380891e-01 1.31569244e-02 6.27669096e-01
5.66162109e-01 -2.83590376e-01 5.13704240e-01 -9.52324495e-02
1.73362821e-01 9.08730209e-01 -1.32945168e+00 1.20363605e+00
-8.66778046e-02 1.86208040e-01 -6.33087277e-01 -5.10736942e-01
1.34932745e+00 -8.32320750e-02 1.72867104e-01 -7.39146888e-01
-4.05372590e-01 -1.42027484e-02 -1.68655679e-01 -1.85521975e-01
1.00482655e+00 6.04395531e-02 -6.01823568e-01 3.60548198e-01
-2.14165285e-01 -4.22860801e-01 5.71119070e-01 3.82092863e-01
1.30391157e+00 5.81712365e-01 7.14571774e-02 -1.36776403e-01
-2.62313306e-01 3.82874906e-01 6.19003892e-01 1.11600661e+00
7.57169187e-01 9.91498947e-01 1.39074194e+00 -5.40281355e-01
-1.34467757e+00 -1.35185087e+00 -8.78038928e-02 9.59989548e-01
-4.89311159e-01 -7.12927580e-01 -4.79421258e-01 -5.84569991e-01
6.33951545e-01 1.40910363e+00 -1.00210094e+00 -2.04441726e-01
-3.26685011e-01 -1.07065392e+00 6.46983206e-01 6.26631737e-01
-1.75079092e-01 -1.29458237e+00 -7.31600076e-02 3.73956829e-01
-2.59662509e-01 -8.08501780e-01 -2.56159425e-01 7.64833391e-01
-8.48130763e-01 -8.65546405e-01 3.21933091e-01 -2.14513272e-01
4.32979465e-01 -1.91835299e-01 1.72702110e+00 -6.52029574e-01
-1.09817339e-02 -4.84400839e-01 -2.13449463e-01 -5.53228974e-01
-9.10001397e-01 3.40737462e-01 -1.78174138e-01 -3.27470779e-01
3.00029904e-01 -3.49527001e-01 9.62827951e-02 3.68372917e-01
-8.33531857e-01 5.19844949e-01 5.08053839e-01 9.99156833e-01
9.99950588e-01 -8.95586535e-02 4.51135635e-01 -1.84422457e+00
5.43904424e-01 -6.78932011e-01 -7.76313484e-01 3.97190720e-01
-5.69584787e-01 5.11635423e-01 8.31917763e-01 -3.29958588e-01
-9.81382132e-01 -9.91917849e-02 1.55145153e-01 -5.01728415e-01
1.19633831e-01 6.69441044e-01 -3.22299838e-01 1.12146652e+00
9.03489947e-01 3.15788865e-01 1.08058721e-01 -6.47581339e-01
4.56569254e-01 4.29741561e-01 7.51884878e-01 -7.87539899e-01
8.21466804e-01 8.41474608e-02 -6.97915182e-02 -4.71263267e-02
-6.58434749e-01 2.18002751e-01 -6.47046864e-01 3.74782950e-01
4.84809488e-01 -1.08022976e+00 -9.74027336e-01 2.09023654e-01
-9.09528911e-01 -1.00237429e+00 -5.70558667e-01 -1.15467481e-01
-8.97769928e-01 -3.61187190e-01 -6.14424169e-01 -8.30674052e-01
-3.56776536e-01 -1.16398382e+00 1.28494310e+00 -3.95603955e-01
-5.47480047e-01 -5.92713833e-01 7.62525648e-02 2.82532305e-01
1.34560630e-01 3.11675012e-01 1.31244111e+00 -9.46826041e-01
-1.02815723e+00 -1.63783416e-01 -2.11902142e-01 -4.80576009e-02
-1.30777564e-02 4.76568580e-01 -6.68853521e-01 1.26142457e-01
-6.43383026e-01 -5.13353944e-01 6.09256625e-01 1.00928031e-01
1.53767705e+00 -5.74336886e-01 -4.24812198e-01 6.52155757e-01
1.03070080e+00 5.78165591e-01 1.05972731e+00 1.13168545e-01
7.94729292e-01 7.28081048e-01 9.49030757e-01 4.78711337e-01
8.16248953e-01 5.51276565e-01 1.69491231e-01 9.60970595e-02
2.97426164e-01 -1.14557922e+00 4.22526985e-01 3.23580682e-01
8.09036553e-01 -5.90402961e-01 -1.27435243e+00 3.69775146e-01
-1.86745882e+00 -1.26195657e+00 -5.57935834e-01 2.56450748e+00
1.25724900e+00 5.83778620e-01 1.18681967e-01 -4.64026295e-02
2.63877332e-01 -2.74314642e-01 -5.76890528e-01 -4.96028334e-01
-2.97909617e-01 5.98380864e-02 6.46281660e-01 4.38640743e-01
-8.80204201e-01 1.08860505e+00 6.88245726e+00 7.77595699e-01
-5.89965999e-01 -5.23898065e-01 1.32947612e+00 -2.28713438e-01
-7.67717600e-01 1.11543253e-01 -1.20741796e+00 6.96923792e-01
1.45277739e+00 -1.46544382e-01 5.19519329e-01 1.07640588e+00
6.92795068e-02 -2.04652622e-01 -1.46908784e+00 6.87282264e-01
-2.10657045e-01 -1.49986422e+00 3.91390830e-01 4.27983962e-02
4.98324722e-01 -2.13823952e-02 1.36589855e-01 7.60801315e-01
1.26873064e+00 -1.29451597e+00 1.02308202e+00 8.77738178e-01
9.43304241e-01 -7.94897020e-01 4.02811795e-01 3.15823615e-01
-7.69523561e-01 1.17913142e-01 -4.85615194e-01 3.38629425e-01
-4.32016551e-02 5.67212760e-01 -1.39779305e+00 6.31251514e-01
9.93664622e-01 7.50413418e-01 -1.02824783e+00 4.60688680e-01
1.66488498e-01 5.70279479e-01 -3.02291155e-01 9.17381793e-02
-4.75763381e-01 -2.80363202e-01 1.46269545e-01 1.03600788e+00
4.32998598e-01 -1.99807689e-01 -1.24867931e-01 1.20869064e+00
-1.39115587e-01 -1.69593111e-01 -8.46128225e-01 -1.38020992e-01
8.85236621e-01 7.89374411e-01 -2.41244853e-01 -5.61149955e-01
-1.13192834e-01 3.16989392e-01 3.34214926e-01 1.92449287e-01
-7.94250190e-01 -2.86636055e-01 6.26347661e-01 5.03707349e-01
5.05123913e-01 2.63232350e-01 -7.97482908e-01 -1.00936246e+00
6.36365414e-02 -1.59886515e+00 5.61639428e-01 -1.47959781e+00
-1.30826592e+00 7.34093487e-01 4.11729842e-01 -1.12528527e+00
-8.01714242e-01 -3.69999170e-01 1.00191809e-01 1.19064105e+00
-4.40050304e-01 -1.06758535e+00 -1.55597985e-01 3.63192558e-02
3.46357405e-01 -1.78521425e-01 8.02437007e-01 -5.51154427e-02
-6.05850577e-01 8.61820281e-01 2.40666091e-01 1.40079126e-01
8.45335364e-01 -1.54887509e+00 1.52845883e+00 5.59652030e-01
-1.83966458e-01 1.04297459e+00 9.38073933e-01 -1.18129635e+00
-1.78721678e+00 -1.39518368e+00 1.06527662e+00 -1.26031375e+00
6.61802530e-01 -1.20077884e+00 -1.15549934e+00 1.31411183e+00
-1.25141934e-01 -1.40148163e-01 5.18493474e-01 3.19498211e-01
-8.82204533e-01 -6.59109652e-01 -1.08762145e+00 7.84044385e-01
1.01229525e+00 -3.31776649e-01 -3.14099818e-01 1.94723025e-01
1.28093898e+00 -6.45974696e-01 -1.26741683e+00 4.14366394e-01
6.85357094e-01 -1.06235051e+00 1.03650963e+00 -7.39833057e-01
8.91988695e-01 -3.25616539e-01 -3.55705738e-01 -1.33542395e+00
-1.50067210e-01 -7.29266107e-01 -9.46621597e-02 1.45197618e+00
1.20761323e+00 -6.16470039e-01 1.01626861e+00 7.89306521e-01
-4.02408279e-02 -5.53920209e-01 -4.65151727e-01 -6.50117636e-01
5.08370474e-02 -6.01295173e-01 1.07472956e+00 7.26821542e-01
-6.63443133e-02 4.22127128e-01 -4.15607870e-01 -1.17291622e-01
4.26756233e-01 8.35113227e-02 1.55484533e+00 -1.06888556e+00
-4.85932171e-01 -2.14101635e-02 8.61362666e-02 -8.83361161e-01
-3.66049588e-01 -8.66930544e-01 1.50870606e-02 -1.29978120e+00
1.56717628e-01 -9.38982010e-01 1.86199963e-01 6.87211752e-01
-3.92112911e-01 -2.64839023e-01 -1.94139965e-02 1.85153484e-02
-3.25159222e-01 5.12357056e-01 7.66194522e-01 -5.47061525e-02
-1.94383577e-01 -1.35461301e-01 -9.36226428e-01 -2.03733996e-01
6.74380243e-01 -7.58185327e-01 -7.23896086e-01 -1.99856490e-01
8.19273829e-01 8.17466319e-01 -9.19584278e-03 -8.68810713e-01
-8.18769410e-02 -3.01101625e-01 4.47593361e-01 -1.42388892e+00
3.53357404e-01 -4.83365148e-01 1.10532653e+00 1.03272714e-01
-7.39490509e-01 8.16097260e-01 5.26584268e-01 5.75820625e-01
-8.53384137e-02 2.99809426e-01 4.21988547e-01 1.20312173e-03
1.17581397e-01 6.19338900e-02 -2.64457822e-01 2.92062581e-01
8.12409759e-01 -1.53892608e-02 -8.06827009e-01 -2.20644310e-01
-4.09770489e-01 4.33539450e-01 6.13328159e-01 6.53608501e-01
2.84422398e-01 -1.16070640e+00 -6.86663508e-01 5.75808823e-01
4.50262070e-01 5.10811329e-01 -2.39362754e-02 2.91022390e-01
-7.29718447e-01 5.12197137e-01 5.27755134e-02 -5.67394018e-01
-1.01817989e+00 1.12505829e+00 2.62998760e-01 -5.95504999e-01
-1.95046857e-01 7.90533066e-01 1.92770481e-01 -1.33649540e+00
9.61162150e-02 -7.27250099e-01 4.05896813e-01 2.05230564e-01
3.34345549e-01 1.12920791e-01 2.87035912e-01 -6.69785589e-02
-5.84779866e-03 -3.95761698e-01 -3.35405737e-01 -1.43688083e-01
1.23525560e+00 3.32845539e-01 -2.94379801e-01 1.03518128e+00
6.75347507e-01 2.61156820e-02 -1.18228090e+00 -3.64048369e-02
2.40567058e-01 -4.32866842e-01 -8.75624776e-01 -1.09160972e+00
-5.26466310e-01 7.04640806e-01 -1.89824060e-01 4.29659843e-01
6.34030938e-01 -2.53508151e-01 1.80285782e-01 4.36012626e-01
5.43657243e-01 -5.61689198e-01 -1.02027901e-01 5.69999635e-01
1.09262514e+00 -9.99863863e-01 1.24886725e-03 -6.75337374e-01
-1.06999612e+00 6.14254653e-01 8.40133190e-01 2.11318612e-01
4.26335156e-01 7.43651748e-01 7.33843446e-03 9.03928727e-02
-1.91897810e+00 3.03814977e-01 -1.21480525e-02 7.23441958e-01
7.39836514e-01 1.67563409e-01 4.22872782e-01 6.07564628e-01
-9.31517959e-01 -4.10052650e-02 5.61670899e-01 7.17682004e-01
2.20435724e-01 -1.40963817e+00 -5.02059758e-01 1.08968329e+00
-3.21372807e-01 -5.63623369e-01 -3.67081404e-01 1.04584086e+00
-3.39192718e-01 6.12547696e-01 5.47408283e-01 -7.84927368e-01
7.14625418e-01 3.38003397e-01 7.37213567e-02 -7.09988892e-01
-8.44097674e-01 2.04496812e-02 5.01546085e-01 -5.60547173e-01
5.77475309e-01 -9.90992486e-01 -8.90905261e-01 -7.54781246e-01
3.48884553e-01 1.23682164e-01 1.93606198e-01 8.85316133e-02
8.67211580e-01 6.40653670e-01 1.65221006e-01 -2.82352000e-01
-5.68983376e-01 -1.22411191e+00 -2.70768881e-01 4.66400802e-01
-4.40231198e-03 -3.84028822e-01 2.06689507e-01 2.74457961e-01] | [9.761838912963867, 7.879315376281738] |
502b42c8-320f-453a-b5d1-28fdedfa6678 | bayesian-experimental-design-for-symbolic | 2211.15860 | null | https://arxiv.org/abs/2211.15860v1 | https://arxiv.org/pdf/2211.15860v1.pdf | Bayesian Experimental Design for Symbolic Discovery | This study concerns the formulation and application of Bayesian optimal experimental design to symbolic discovery, which is the inference from observational data of predictive models taking general functional forms. We apply constrained first-order methods to optimize an appropriate selection criterion, using Hamiltonian Monte Carlo to sample from the prior. A step for computing the predictive distribution, involving convolution, is computed via either numerical integration, or via fast transform methods. | ['Nimrod Megiddo', 'Lior Horesh', 'Joao Goncalves', 'Sanjeeb Dash', 'Cristina Cornelio', 'Kenneth L. Clarkson'] | 2022-11-29 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [ 5.26035666e-01 -2.69904137e-01 -5.69110692e-01 -3.67873669e-01
-4.46777254e-01 -3.81933779e-01 7.16943860e-01 -1.64947599e-01
-4.86910164e-01 1.34131086e+00 -3.94438863e-01 -9.21308935e-01
-7.06560791e-01 -6.17883325e-01 -6.08876824e-01 -7.79033840e-01
-2.14768350e-01 7.40380764e-01 -7.31175626e-03 3.57538849e-01
7.55378366e-01 9.77332950e-01 -1.35654163e+00 -4.64370638e-01
8.65055859e-01 9.00715232e-01 1.57464758e-01 6.19075954e-01
5.37477791e-01 3.49465460e-01 -5.13584651e-02 -7.81436358e-03
-1.18440084e-01 -7.90219605e-01 -5.91402292e-01 -1.44788325e-01
-3.70608270e-01 -2.83132732e-01 7.02843070e-02 1.33179462e+00
5.26288927e-01 6.56119168e-01 1.34982777e+00 -9.89997983e-01
-3.74405593e-01 4.00808156e-01 -8.56692120e-02 1.20185383e-01
3.82709414e-01 3.77065301e-01 9.03287411e-01 -9.36282575e-01
4.14644867e-01 1.41381025e+00 3.96060616e-01 1.63219962e-02
-2.09188271e+00 -4.86026615e-01 -5.74410260e-01 1.72313198e-01
-2.07239580e+00 -7.21466720e-01 5.49171567e-01 -7.93780923e-01
8.97020578e-01 9.57788527e-02 6.78382397e-01 8.78106952e-01
5.46251297e-01 5.46602271e-02 1.40743589e+00 -7.11043656e-01
9.10829544e-01 2.03962363e-02 2.87354905e-02 6.78035617e-01
4.49630827e-01 1.10546947e+00 -3.57155770e-01 -6.20688081e-01
7.54253685e-01 -3.13830137e-01 -4.21440229e-02 -5.36156632e-02
-7.21739948e-01 1.28220546e+00 -2.79642165e-01 -2.86904097e-01
-7.99804330e-01 3.89976114e-01 -2.13815290e-02 -1.35738567e-01
4.26604599e-01 7.66086459e-01 -5.55349648e-01 1.82323530e-01
-8.62613916e-01 6.67054474e-01 9.57475722e-01 5.31803489e-01
4.84462231e-01 -9.90909263e-02 -1.39536023e-01 2.36083090e-01
8.67499292e-01 9.84746039e-01 -1.05651140e-01 -1.38353360e+00
-4.21929538e-01 -2.24610522e-01 5.75445175e-01 -5.20853519e-01
-2.83518463e-01 3.54657881e-02 -3.95870149e-01 2.76524007e-01
4.07588214e-01 -2.17534423e-01 -9.37821388e-01 1.56699765e+00
5.29132545e-01 5.24553210e-02 -7.03427345e-02 6.03580058e-01
1.88580826e-01 8.32932651e-01 2.75866866e-01 -1.07407999e+00
1.09036994e+00 3.90758645e-03 -9.30554211e-01 2.23721072e-01
3.70017797e-01 -7.56348014e-01 5.80700457e-01 6.22301280e-01
-9.91679907e-01 -1.62657127e-01 -1.04885924e+00 2.42380023e-01
-2.27211133e-01 9.18646231e-02 6.59874737e-01 5.73917270e-01
-5.17057419e-01 1.25229192e+00 -9.01114285e-01 1.45402431e-01
1.25664666e-01 5.52508950e-01 2.02869371e-01 3.33664328e-01
-1.10522258e+00 1.18044508e+00 5.64845264e-01 9.24855918e-02
-1.07137609e+00 -7.16778219e-01 -7.09802508e-01 -1.27639875e-01
3.82495880e-01 -7.78244019e-01 1.29416788e+00 -3.00086945e-01
-2.09903145e+00 3.75024736e-01 -3.72641742e-01 -3.29413533e-01
-1.03422906e-02 -8.92296259e-04 -1.56786561e-01 1.58081368e-01
-4.98982519e-02 2.22436532e-01 9.82905447e-01 -6.69007003e-01
8.80945027e-02 -2.84966141e-01 -5.67483902e-01 -2.49211296e-01
6.47939622e-01 2.75796026e-01 3.50682944e-01 -2.71565467e-01
2.05126062e-01 -9.33738708e-01 -7.10742414e-01 -2.15097830e-01
-5.33794343e-01 -2.99621642e-01 1.86736763e-01 -9.11940396e-01
1.18375981e+00 -1.98289132e+00 3.81451547e-01 7.28594244e-01
-2.12286726e-01 -4.23420131e-01 3.56105477e-01 3.51032495e-01
-1.29457831e-01 6.83594793e-02 -3.88946682e-01 2.94459641e-01
1.56975508e-01 7.01962560e-02 -4.19880152e-01 8.36802006e-01
3.53628099e-01 5.33880234e-01 -6.73580408e-01 -6.68410242e-01
1.90204978e-01 1.89921841e-01 -9.14294183e-01 2.77126193e-01
-6.45213723e-01 5.36690831e-01 -9.24450934e-01 3.31008613e-01
4.06160861e-01 -1.87352240e-01 3.52516353e-01 -2.31529567e-02
-7.61724412e-01 4.95298564e-01 -1.23914266e+00 1.21159256e+00
3.66162620e-02 2.04982236e-01 -1.07525699e-01 -1.32614839e+00
6.05196476e-01 3.55744541e-01 3.19296539e-01 -2.63621867e-01
4.89844322e-01 3.84677261e-01 1.00014180e-01 -6.89181030e-01
2.49801725e-01 -5.05492270e-01 -1.33935913e-01 2.20971823e-01
1.26693472e-01 -6.12966061e-01 1.54778242e-01 -2.64712751e-01
5.44832826e-01 3.99787813e-01 9.00495350e-01 -9.30613458e-01
1.38067618e-01 1.40039846e-01 2.49096677e-01 9.28858042e-01
1.79472223e-01 -1.65563524e-01 7.24415898e-01 -5.71305901e-02
-9.96759415e-01 -9.10251200e-01 -8.45655262e-01 5.01191676e-01
-2.65555501e-01 -1.04340946e-03 -6.39709830e-01 1.79122388e-01
1.41470999e-01 1.29528522e+00 -3.36518794e-01 -3.91761392e-01
-2.29170904e-01 -1.32338643e+00 1.13683399e-02 1.17514290e-01
-9.70242918e-02 -6.78152919e-01 -6.42531812e-01 4.50534344e-01
2.52403140e-01 -5.52177787e-01 -1.05022363e-01 7.93420494e-01
-1.06197834e+00 -8.99467170e-01 -1.56818852e-01 1.81115478e-01
3.14331889e-01 -4.73735660e-01 7.29785860e-01 -3.76430929e-01
-4.47346359e-01 -5.34755504e-03 1.99300036e-01 -2.93917418e-01
-4.58268344e-01 -4.33415771e-01 2.90077925e-01 -3.11337441e-01
2.34302521e-01 -6.95079923e-01 -2.74882525e-01 1.05632797e-01
-2.63866246e-01 -2.63636321e-01 4.33951259e-01 1.04990375e+00
8.95236194e-01 2.81030424e-02 3.43078464e-01 -3.10974836e-01
5.81843615e-01 -4.08134639e-01 -1.56012964e+00 2.39770249e-01
-8.41749489e-01 5.59004903e-01 2.64153123e-01 -5.31579733e-01
-9.84089613e-01 1.40078068e-01 8.88718367e-02 -2.36138359e-01
-1.26023352e-01 9.52144384e-01 -2.60402858e-01 -1.54042661e-01
6.27663136e-01 -2.16434956e-01 9.87234563e-02 -5.39734721e-01
3.74104261e-01 4.31933939e-01 2.63932675e-01 -1.14663994e+00
2.08901823e-01 -7.47087225e-02 8.41562808e-01 -1.00259340e+00
-6.94497168e-01 1.79666549e-01 -4.22061920e-01 -1.34177968e-01
1.09234536e+00 -3.14301789e-01 -1.21099985e+00 -5.95195219e-03
-1.18650424e+00 -3.57080549e-01 -2.91395485e-01 1.17463136e+00
-7.73297250e-01 -1.76587388e-01 -2.84842938e-01 -1.38083220e+00
7.91493133e-02 -1.32412159e+00 9.01678741e-01 2.74976909e-01
-3.24108213e-01 -8.50209713e-01 3.21751148e-01 -2.90234983e-01
-7.28256404e-02 1.83847845e-01 1.13677537e+00 -5.53395033e-01
-6.98253274e-01 -3.31123769e-01 -8.05971306e-03 -9.65040401e-02
-5.15591264e-01 6.05873764e-01 -8.96347404e-01 -4.37596329e-02
1.52125493e-01 -9.54924822e-02 4.25143927e-01 1.25896168e+00
1.22676516e+00 -3.25031340e-01 -7.35925078e-01 5.86273015e-01
1.46060145e+00 7.14820325e-01 2.54392892e-01 -3.34602416e-01
1.51824549e-01 4.83351827e-01 5.64167798e-01 7.20567167e-01
-3.88974309e-01 7.80059695e-01 -8.49469751e-02 5.91020465e-01
4.98871893e-01 -3.26840788e-01 1.45608425e-01 2.62110591e-01
-3.43543351e-01 -8.76939483e-03 -7.36975193e-01 -3.77809256e-02
-1.61462009e+00 -1.08351707e+00 -2.74001956e-01 2.54983830e+00
1.13894176e+00 8.24186280e-02 -3.91724929e-02 -1.55454844e-01
6.77541792e-01 -6.30771935e-01 -6.26724362e-01 -3.27845454e-01
3.67443353e-01 7.47394383e-01 9.11013544e-01 8.70735347e-01
-8.89719903e-01 6.60150528e-01 8.97643471e+00 1.18860269e+00
-6.68484569e-01 4.85099144e-02 3.91010880e-01 1.96034938e-01
-7.47445002e-02 6.89692259e-01 -1.14575040e+00 4.64109242e-01
1.52283549e+00 -3.82183880e-01 7.87405491e-01 4.72639471e-01
8.72628927e-01 -6.64257526e-01 -1.15178168e+00 3.23198915e-01
-6.35879338e-01 -1.18290782e+00 -2.50175387e-01 3.68344009e-01
6.25612378e-01 -4.71852630e-01 -2.05631375e-01 -7.72474781e-02
6.81341112e-01 -1.02767336e+00 7.85575330e-01 1.01025593e+00
7.63718545e-01 -7.17675328e-01 2.78033793e-01 2.67200738e-01
-6.84039712e-01 1.06164813e-01 -3.21173102e-01 -4.48355258e-01
2.31809586e-01 9.40442502e-01 -7.51533151e-01 1.24956109e-01
2.30125591e-01 3.38686973e-01 1.18751481e-01 1.12065816e+00
-3.30406189e-01 8.09615314e-01 -7.29664445e-01 -4.99277294e-01
-1.34547740e-01 -7.86995888e-01 6.45752728e-01 7.53393650e-01
5.64687252e-01 4.13296551e-01 2.80992061e-01 1.60268223e+00
6.36251926e-01 -2.01191559e-01 -6.96902096e-01 -4.03223157e-01
8.18741143e-01 4.71512914e-01 -7.57966757e-01 -2.30498701e-01
1.07865948e-02 6.81168064e-02 -1.91167787e-01 6.81029022e-01
-7.17733204e-01 -4.21029180e-01 4.12005484e-01 -2.85945535e-01
3.69195312e-01 -3.78554940e-01 -3.40403110e-01 -7.06611991e-01
-5.28607965e-01 -7.04261720e-01 3.14558297e-01 -7.50733256e-01
-9.71507311e-01 -3.12720150e-01 1.06349838e+00 -4.55438256e-01
-6.54998481e-01 -9.42420185e-01 -4.32322055e-01 1.28250468e+00
-7.48781681e-01 -3.91857654e-01 7.90695190e-01 2.16846466e-01
-3.15256938e-02 -4.30875942e-02 8.67765069e-01 -7.42764845e-02
-8.71900380e-01 3.67711186e-02 4.39771771e-01 -7.03833222e-01
3.74206714e-02 -9.87969279e-01 -3.07896852e-01 7.23673999e-01
-6.59231663e-01 7.39671648e-01 1.53490222e+00 -8.91252220e-01
-1.42159247e+00 -4.82569933e-01 8.02106678e-01 -3.75218466e-02
9.65253830e-01 5.59782162e-02 -5.40867507e-01 4.94211555e-01
-1.61653429e-01 -2.53158480e-01 4.47207779e-01 1.19292863e-01
3.61361861e-01 3.33124697e-01 -1.11552501e+00 5.51797450e-01
6.82872355e-01 -4.85462248e-01 -5.02567410e-01 4.36287761e-01
6.86221004e-01 -7.55139291e-02 -1.13296890e+00 5.34342527e-01
5.69836259e-01 -1.74817338e-03 9.93246555e-01 -8.75687003e-01
3.38557214e-01 -3.85650247e-01 -4.51838464e-01 -1.16087627e+00
-2.37567216e-01 -1.13664222e+00 -3.20150740e-02 6.82576895e-01
4.45501149e-01 -7.01394260e-01 1.68472528e-01 5.77379227e-01
3.31601761e-02 -4.10973430e-01 -1.06549168e+00 -9.09432113e-01
2.11759120e-01 -5.33295691e-01 3.53079557e-01 5.09229660e-01
1.31580606e-01 4.20493037e-01 -1.95255756e-01 4.17281389e-01
1.16491926e+00 1.59989059e-01 2.65610576e-01 -1.30860829e+00
-6.52099609e-01 -5.33367455e-01 -1.86833486e-01 -9.24993992e-01
2.98714221e-01 -6.91267729e-01 4.14914250e-01 -6.30995572e-01
9.61687118e-02 -1.68562010e-02 1.41307920e-01 3.15608680e-02
1.33318856e-01 -4.28326368e-01 -5.36111236e-01 -1.96335077e-01
1.09001279e-01 7.19662845e-01 1.11431384e+00 2.49770507e-01
-2.01611936e-01 1.43922597e-01 -2.11041257e-01 8.18222880e-01
5.42755604e-01 -9.72427070e-01 -2.90780723e-01 5.44377327e-01
5.07770121e-01 6.95414782e-01 7.97985017e-01 -4.14834708e-01
3.00269504e-03 -8.32769036e-01 2.94803411e-01 -7.04110682e-01
2.25681946e-01 -4.73652869e-01 6.12614989e-01 6.70712352e-01
-5.31531513e-01 -1.60230160e-01 8.20228606e-02 4.45771903e-01
2.31384844e-01 -1.08087492e+00 8.66187751e-01 4.00168113e-02
-7.14152679e-02 -6.44618273e-02 -8.41686130e-01 -1.17035769e-01
6.50219977e-01 3.33063096e-01 1.64096713e-01 6.48834780e-02
-1.17978823e+00 -4.07471173e-02 2.29384899e-01 -7.56030679e-01
4.80791688e-01 -1.34068394e+00 -4.13225353e-01 1.82356182e-02
-4.75249201e-01 -5.37810445e-01 -1.71898574e-01 1.10207570e+00
-4.76942182e-01 5.30086577e-01 2.14427024e-01 -4.04206932e-01
-5.34740388e-01 6.62463546e-01 7.29721665e-01 -8.69317204e-02
-1.63347155e-01 4.95520979e-01 -1.98475376e-01 -7.31133623e-03
-1.29748121e-01 -1.79988131e-01 7.50048608e-02 -1.22137897e-01
4.76066977e-01 6.91431582e-01 -4.72844362e-01 -7.05425069e-02
-4.04341698e-01 4.09285843e-01 6.05225742e-01 -7.70792067e-01
1.00125337e+00 4.99873571e-02 -4.11652654e-01 7.01089263e-01
1.08212507e+00 -5.94924271e-01 -1.18845844e+00 2.06386950e-02
2.58554190e-01 -3.91675830e-01 6.63566113e-01 -6.05522871e-01
-3.18321317e-01 5.95086753e-01 4.34167087e-01 1.97241306e-01
6.11992180e-01 -6.27398714e-02 -2.33534455e-01 6.78459942e-01
1.14918910e-01 -1.09787214e+00 -5.00757277e-01 2.39714712e-01
9.15355325e-01 -9.56997931e-01 7.14438081e-01 -3.37306261e-01
1.66299164e-01 1.06372762e+00 -9.40096080e-02 -1.23865351e-01
1.26214218e+00 2.14173675e-01 -1.01244748e+00 -3.98430347e-01
-7.76304483e-01 -1.22601494e-01 5.20272970e-01 2.32510671e-01
3.78564864e-01 4.55829382e-01 -9.39602017e-01 5.01109719e-01
-4.19000387e-02 4.61530060e-01 1.89452186e-01 7.48131573e-01
-5.64585149e-01 -7.51103044e-01 -5.80076873e-01 5.11206567e-01
-4.23592895e-01 -2.95046031e-01 4.91243191e-02 5.98052979e-01
-8.33971575e-02 1.10583663e+00 -6.70944452e-02 6.41446561e-02
3.87548655e-02 5.03803194e-01 8.62065077e-01 -2.20332131e-01
3.45449179e-01 5.26478052e-01 3.62249255e-01 -3.95420015e-01
-4.81858939e-01 -1.28895676e+00 -9.05295551e-01 -3.38705808e-01
-8.18604648e-01 4.65961695e-01 8.26116025e-01 1.44028783e+00
-1.35432854e-01 4.28890288e-01 6.49512410e-01 -1.02832389e+00
-1.09030044e+00 -8.69104564e-01 -8.94094884e-01 -6.43840849e-01
4.98716608e-02 -1.11080956e+00 -6.75306022e-01 -4.82019335e-02] | [6.663104057312012, 3.939760684967041] |
e80ba69f-9625-4ea1-98ae-f6faf3d7121a | what-makes-visual-place-recognition-easy-or | 2106.12671 | null | https://arxiv.org/abs/2106.12671v1 | https://arxiv.org/pdf/2106.12671v1.pdf | What makes visual place recognition easy or hard? | Visual place recognition is a fundamental capability for the localization of mobile robots. It places image retrieval in the practical context of physical agents operating in a physical world. It is an active field of research and many different approaches have been proposed and evaluated in many different experiments. In the following, we argue that due to variations of this practical context and individual design decisions, place recognition experiments are barely comparable across different papers and that there is a variety of properties that can change from one experiment to another. We provide an extensive list of such properties and give examples how they can be used to setup a place recognition experiment easier or harder. This might be interesting for different involved parties: (1) people who just want to select a place recognition approach that is suitable for the properties of their particular task at hand, (2) researchers that look for open research questions and are interested in particularly difficult instances, (3) authors that want to create reproducible papers on this topic, and (4) also reviewers that have the task to identify potential problems in papers under review. | ['Peer Neubert', 'Stefan Schubert'] | 2021-06-23 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [-8.79114680e-03 -2.37336367e-01 -2.80969441e-01 -3.29381585e-01
-2.06579253e-01 -7.78372109e-01 8.28056216e-01 1.75601259e-01
-5.16556501e-01 6.35856926e-01 -2.10077360e-01 -4.62804168e-01
-2.42323399e-01 -5.13762712e-01 -5.13572812e-01 -7.53540158e-01
-4.38775867e-01 5.49329579e-01 5.20608127e-01 -2.48492584e-01
9.88451600e-01 9.44424868e-01 -1.77530694e+00 -2.15669513e-01
4.15162981e-01 4.55574840e-01 6.67169154e-01 5.43293715e-01
2.52601337e-02 4.37331945e-01 -4.35335845e-01 9.06709954e-02
3.87072444e-01 -1.76608771e-01 -9.33774114e-01 1.68624111e-02
9.81923342e-02 3.96790713e-01 -7.86903128e-02 1.27936113e+00
4.08136517e-01 2.91228354e-01 7.32846200e-01 -1.45303261e+00
-7.78089643e-01 1.27388775e-01 -4.32264984e-01 3.03265452e-01
6.41514122e-01 3.67576466e-03 4.89094019e-01 -7.82113552e-01
8.36247146e-01 1.15420926e+00 4.56818581e-01 1.69417143e-01
-8.22693169e-01 -4.75862995e-02 1.84746966e-01 5.52128971e-01
-1.55818033e+00 -4.21689004e-01 8.03173244e-01 -2.24320143e-01
8.10011327e-01 4.37257826e-01 2.52839923e-01 1.05871439e+00
1.51918486e-01 6.51662588e-01 1.19028473e+00 -7.63187230e-01
5.10725975e-01 4.13338333e-01 -2.04457790e-01 5.48132360e-01
4.11080807e-01 -8.94976780e-02 -3.32973093e-01 -1.12376697e-01
7.64259815e-01 1.02546766e-01 -3.14434618e-01 -1.08459449e+00
-1.79663873e+00 6.88898742e-01 5.50774813e-01 7.92430580e-01
-1.66174471e-01 -5.55356182e-02 7.60275349e-02 4.30898309e-01
-2.89643407e-01 8.47875059e-01 -1.73232496e-01 -4.22775060e-01
-5.49959421e-01 3.81672025e-01 1.07195711e+00 1.18949878e+00
8.14481735e-01 -4.48059797e-01 3.56034994e-01 6.62170529e-01
4.67394441e-01 4.41669166e-01 6.52268052e-01 -9.48998630e-01
3.36673379e-01 4.66205180e-01 5.71007907e-01 -1.32106483e+00
-4.54419136e-01 2.26268992e-01 -3.96097779e-01 4.55026984e-01
5.51638126e-01 9.70860105e-03 -5.79071403e-01 1.28263354e+00
1.44062683e-01 -2.87104905e-01 7.60721490e-02 1.06332624e+00
6.96421564e-01 6.37305796e-01 -1.98014572e-01 4.26977351e-02
1.53431237e+00 -1.19298589e+00 -4.94844437e-01 -5.85256755e-01
4.91845220e-01 -1.02458549e+00 8.83706033e-01 1.65232979e-02
-8.19135964e-01 -3.42082888e-01 -1.37397933e+00 -3.86532247e-02
-9.10383105e-01 6.93003684e-02 7.04518557e-01 5.24433076e-01
-1.19876122e+00 5.62810898e-01 -8.64108205e-01 -1.24907637e+00
-1.63101524e-01 4.84115452e-01 -4.27488089e-01 1.08971982e-03
-8.18166733e-01 1.35553515e+00 8.78876597e-02 1.34482518e-01
-4.75528747e-01 1.15545459e-01 -9.19214785e-01 -3.64975691e-01
2.91362822e-01 -4.19522882e-01 1.10317135e+00 -5.23543239e-01
-1.26249051e+00 1.01716912e+00 -4.61943120e-01 -2.21811116e-01
5.31044245e-01 2.54533648e-01 -3.74275714e-01 -1.26695976e-01
4.27059323e-01 5.16835153e-01 3.57305914e-01 -1.26602530e+00
-8.19235504e-01 -5.19980073e-01 2.28902012e-01 5.96724451e-01
3.14192086e-01 1.43347755e-01 -4.10186708e-01 -4.26035911e-01
5.51130235e-01 -1.27799821e+00 -4.95997310e-01 -1.46858278e-03
-1.47109106e-01 -2.93530762e-01 8.95652175e-01 1.19598277e-01
4.31692779e-01 -2.21228456e+00 5.79513470e-03 1.53370768e-01
-2.89686412e-01 1.60239607e-01 -5.28448112e-02 8.14186394e-01
1.49544641e-01 1.36512563e-01 -1.49091594e-02 -6.80858716e-02
2.02798411e-01 1.07213311e-01 -1.79828122e-01 9.86431420e-01
-2.55435169e-01 7.28881717e-01 -1.10582173e+00 -1.77072674e-01
4.17883992e-01 3.30224663e-01 2.54602909e-01 -1.20519541e-01
2.88620353e-01 4.72821206e-01 -7.46831357e-01 7.14369774e-01
5.41386724e-01 -1.02081507e-01 2.76303757e-02 2.38648966e-01
-5.84858835e-01 2.86624819e-01 -1.48484015e+00 1.36995721e+00
-1.62808031e-01 1.01636934e+00 2.29237881e-03 -1.03668511e+00
1.02315867e+00 2.22921178e-01 2.80193925e-01 -7.61344314e-01
8.35986957e-02 4.66318250e-01 -2.36959253e-02 -5.85213602e-01
8.67813706e-01 1.78477988e-02 -1.27824038e-01 4.51071203e-01
-2.97397405e-01 -3.08114868e-02 1.57700837e-01 -5.46359837e-01
1.11256325e+00 6.33521453e-02 5.33796132e-01 -4.62666869e-01
5.16229212e-01 1.08347349e-01 1.51515111e-01 1.05328798e+00
-8.54987025e-01 6.03859842e-01 5.80866709e-02 -8.03548217e-01
-1.02550089e+00 -7.24645197e-01 -2.16535524e-01 1.01949453e+00
9.85183060e-01 1.42359138e-02 -3.03654850e-01 -1.97201118e-01
-1.93734080e-01 2.57318050e-01 -5.90318441e-01 1.82687297e-01
-6.43466473e-01 -3.94659698e-01 4.93340731e-01 4.39669907e-01
5.10713995e-01 -1.47576833e+00 -1.11091936e+00 -1.88965760e-02
-1.09042920e-01 -1.01046741e+00 -3.63296014e-03 4.69013095e-01
-7.16478527e-01 -8.51008177e-01 -1.06300926e+00 -1.32702541e+00
8.10724080e-01 9.45492268e-01 7.41660178e-01 1.76182643e-01
2.90702302e-02 5.88060856e-01 -5.90844870e-01 -7.00677514e-01
-6.62082806e-02 4.27817255e-01 1.96653903e-01 -3.05854261e-01
6.47284329e-01 -4.18628991e-01 -4.12206054e-01 7.63471842e-01
-7.55922079e-01 -5.73102951e-01 4.39198941e-01 4.53533441e-01
3.10690373e-01 -6.31473586e-02 6.39906302e-02 -3.26479226e-01
6.57713592e-01 -4.70416903e-01 -6.47374451e-01 4.40397650e-01
-4.49179560e-01 1.43426716e-01 2.53467541e-02 -3.30399454e-01
-5.01689315e-01 2.10492194e-01 2.17532545e-01 1.98473960e-01
-7.64490783e-01 2.15844750e-01 -1.88083977e-01 -6.90312564e-01
8.46454978e-01 3.33420008e-01 -3.55671383e-02 -2.70742178e-01
1.52282804e-01 7.59624898e-01 2.70989120e-01 -3.47501606e-01
7.57476091e-01 7.23051310e-01 1.35451421e-01 -1.02798474e+00
-8.14763382e-02 -8.30684423e-01 -7.19973326e-01 2.78416499e-02
6.01698041e-01 -4.59664375e-01 -5.67642272e-01 2.82004654e-01
-1.18957841e+00 -3.73799354e-01 4.71923649e-02 5.39699018e-01
-7.80218840e-01 4.76512313e-01 -1.50856167e-01 -7.64943063e-01
3.55420142e-01 -1.50245583e+00 1.00130081e+00 6.83246613e-01
-2.83676267e-01 -1.24722505e+00 2.09074035e-01 6.04539551e-03
4.99998152e-01 -6.16471358e-02 2.86161423e-01 -5.98823071e-01
-7.86700368e-01 -3.58836710e-01 -1.04389988e-01 -5.29400229e-01
3.78879398e-01 1.19040450e-02 -1.03864193e+00 -3.42812926e-01
1.47661820e-01 1.85696557e-01 4.41612571e-01 3.55735958e-01
5.17764509e-01 -6.30400181e-02 -9.19143438e-01 1.74811900e-01
1.30138779e+00 7.41939187e-01 1.00750208e+00 1.24276614e+00
3.25549006e-01 7.56809771e-01 8.87861252e-01 -1.96257029e-02
5.72812438e-01 1.08194804e+00 4.75381285e-01 -1.52401790e-01
3.39315742e-01 6.35434613e-02 4.15006548e-01 3.12056869e-01
-2.46755615e-01 -2.79555649e-01 -9.49890018e-01 9.86732543e-01
-2.04016066e+00 -1.05544794e+00 -1.58412829e-01 2.30129099e+00
1.23717889e-01 -2.84651577e-01 2.21255079e-01 -1.91884190e-02
9.19905365e-01 3.72476161e-01 -7.72702694e-02 -3.21242571e-01
-1.33700952e-01 -4.49207217e-01 7.98834205e-01 4.30176109e-01
-1.29034436e+00 8.60136807e-01 6.91342592e+00 3.67146701e-01
-1.39894629e+00 -2.56743908e-01 1.57373965e-01 5.69061220e-01
1.59970909e-01 3.64648789e-01 -5.82451999e-01 3.15129071e-01
5.11133134e-01 -2.23331973e-01 4.09619272e-01 1.09778798e+00
-8.59578848e-02 -4.89545882e-01 -1.20805836e+00 1.10568082e+00
1.59476563e-01 -9.90849793e-01 -4.45486277e-01 2.41042659e-01
5.77665448e-01 4.03767496e-01 2.13991135e-01 -9.15336013e-02
-1.18388496e-01 -1.01841640e+00 7.28766859e-01 3.41679573e-01
-1.22616068e-01 -2.56233424e-01 8.95800114e-01 4.77449894e-01
-1.13080156e+00 -3.98107953e-02 -6.49926722e-01 -5.28034568e-01
1.79476559e-01 4.33706045e-02 -1.00832868e+00 5.06536961e-01
7.98543155e-01 3.89977306e-01 -7.11669326e-01 1.86969626e+00
-7.60135800e-02 -1.33505929e-02 -3.89754415e-01 -6.95785880e-01
3.84012073e-01 -9.45896059e-02 7.65595913e-01 1.18852854e+00
4.95975554e-01 -4.42702085e-01 -1.98373601e-01 6.45372450e-01
6.23696685e-01 6.56049699e-02 -1.35762727e+00 1.14466958e-01
5.76310277e-01 1.14814854e+00 -1.37781119e+00 1.49888441e-01
-3.12863141e-01 1.08495247e+00 3.76916416e-02 3.44504982e-01
-4.19388980e-01 -7.37672627e-01 5.50825655e-01 7.60506392e-02
4.60585535e-01 -5.86463630e-01 -2.23350406e-01 -8.60105276e-01
2.75270760e-01 -5.74835420e-01 5.67528233e-02 -8.95291448e-01
-9.91810858e-01 5.63603938e-01 -9.85595305e-03 -1.51346314e+00
-3.84915441e-01 -7.63949573e-01 -5.38572252e-01 5.61363518e-01
-1.58023024e+00 -7.13254690e-01 -2.04417050e-01 1.75546974e-01
4.38648522e-01 -3.02592278e-01 8.17097902e-01 1.99182630e-01
-5.06568924e-02 1.86034933e-01 8.89199078e-02 1.01845220e-01
7.31893241e-01 -9.12658393e-01 4.26051199e-01 8.47864866e-01
4.81721967e-01 1.07724476e+00 9.62003291e-01 -2.78294772e-01
-1.44865203e+00 -6.18035436e-01 1.43039536e+00 -8.92605364e-01
4.04357940e-01 -1.14025950e-01 -5.46614230e-01 7.43780851e-01
2.72850186e-01 -5.55241853e-02 3.84768814e-01 -5.47056347e-02
7.04706982e-02 1.13100022e-01 -1.20104849e+00 9.11750972e-01
7.17965484e-01 -4.47747529e-01 -8.61369967e-01 2.20770866e-01
5.83037250e-02 -5.94784498e-01 -2.10862830e-01 2.35499993e-01
4.50808227e-01 -1.07219899e+00 1.02028835e+00 -1.97786152e-01
-2.24063694e-01 -9.87407506e-01 -2.08515361e-01 -1.14518118e+00
-4.73874569e-01 -4.62119073e-01 6.32139802e-01 9.21634376e-01
5.54708779e-01 -7.36386716e-01 5.10981679e-01 5.69335461e-01
1.30566522e-01 -4.49885607e-01 -8.78379345e-01 -9.93814349e-01
-1.08053543e-01 -2.95851052e-01 6.21376157e-01 9.79396343e-01
4.33163822e-01 1.35695964e-01 -1.32026926e-01 4.13846135e-01
3.00847143e-01 1.60346881e-01 8.91665339e-01 -1.15780842e+00
3.61138076e-01 -5.13137102e-01 -9.99223232e-01 -1.28136742e+00
-1.44897550e-01 -4.45567429e-01 5.37752509e-01 -1.74116254e+00
-1.30646676e-01 -7.55682468e-01 -1.19201079e-01 1.67575926e-01
1.84602082e-01 2.20905975e-01 2.24706575e-01 5.42421579e-01
-9.22215641e-01 1.35320470e-01 1.18955493e+00 -1.79562628e-01
-2.33321771e-01 2.90540695e-01 -7.23842084e-01 8.58133554e-01
7.37057626e-01 -4.38434839e-01 -3.72062385e-01 -3.18277299e-01
2.83261478e-01 -3.16738486e-01 5.01028478e-01 -1.24343419e+00
7.05551386e-01 -5.69060683e-01 6.67513907e-02 -3.59394014e-01
3.23616594e-01 -1.03600395e+00 1.18050426e-01 4.20349628e-01
2.42802538e-02 4.48679060e-01 -6.81684241e-02 3.54786992e-01
-1.77177146e-01 -7.45734930e-01 4.58339661e-01 -4.91650611e-01
-1.33490717e+00 -1.36634588e-01 -5.78215957e-01 -3.73205602e-01
1.30076146e+00 -7.19608545e-01 -2.52260983e-01 -5.82005203e-01
-4.69312936e-01 -6.82877302e-02 1.08760798e+00 6.75967991e-01
2.93591499e-01 -1.39503336e+00 -1.38656378e-01 -5.62743247e-02
4.89021689e-01 -2.55452693e-01 -3.20230901e-01 9.15215492e-01
-9.71905172e-01 7.17657566e-01 -4.04934645e-01 -6.42810941e-01
-9.55718875e-01 7.20053375e-01 2.42502436e-01 1.27305239e-01
-5.70694685e-01 6.52633607e-01 -7.96514675e-02 -6.05531871e-01
3.23651671e-01 -3.67371142e-01 -2.39774629e-01 -4.29587066e-02
5.63168585e-01 4.46876347e-01 8.16308856e-02 -9.68067348e-01
-8.37938845e-01 8.22109759e-01 2.67496020e-01 -2.59279311e-01
1.17052770e+00 -4.60671663e-01 -1.17806695e-01 7.00374484e-01
8.91748130e-01 1.07628845e-01 -9.99486983e-01 1.10964514e-01
2.84723610e-01 -3.78608167e-01 -4.58820879e-01 -2.80299187e-01
-6.25493050e-01 6.65464401e-01 8.74058723e-01 6.24915659e-01
5.15413046e-01 2.97800243e-01 2.06441469e-02 8.10962498e-01
1.13809454e+00 -1.15531600e+00 -1.15004264e-01 8.03256810e-01
9.60370541e-01 -1.28990817e+00 6.25639036e-02 -2.94993341e-01
-5.53183973e-01 1.09167337e+00 1.30132854e-01 -2.27257758e-01
4.40468043e-01 1.44629151e-01 5.06847084e-01 -2.56402761e-01
-2.42216997e-02 -3.00444275e-01 8.72368664e-02 1.25241160e+00
3.93095404e-01 -2.83323806e-02 -4.01017994e-01 -8.79099667e-02
-2.47721270e-01 -8.09205770e-02 4.94053096e-01 1.22507715e+00
-6.70636952e-01 -1.28992939e+00 -6.41127050e-01 -2.53294796e-01
1.16702810e-01 4.14469570e-01 -6.90149546e-01 8.93300951e-01
1.01683646e-01 1.01441419e+00 -7.86861107e-02 -9.59195197e-02
3.67179841e-01 -1.46610320e-01 4.95350450e-01 -4.03369486e-01
-5.14174625e-02 -3.96204054e-01 -1.71169087e-01 -5.04304230e-01
-7.86578476e-01 -9.26734388e-01 -1.14342558e+00 -1.99300110e-01
-1.45744145e-01 3.49749207e-01 9.05077338e-01 1.03445733e+00
4.07112241e-01 -7.14154467e-02 3.23262334e-01 -1.31432235e+00
-9.60278809e-02 -8.02241325e-01 -8.20740461e-01 9.82903391e-02
3.77970576e-01 -1.05771053e+00 -2.51145154e-01 -3.34418043e-02] | [7.46663761138916, -1.936576008796692] |
a9d8dd2f-efa0-4ec2-8f25-81d29db3a212 | machine-learning-models-and-facial-regions | 2202.08913 | null | https://arxiv.org/abs/2202.08913v1 | https://arxiv.org/pdf/2202.08913v1.pdf | Machine learning models and facial regions videos for estimating heart rate: a review on Patents, Datasets and Literature | Estimating heart rate is important for monitoring users in various situations. Estimates based on facial videos are increasingly being researched because it makes it possible to monitor cardiac information in a non-invasive way and because the devices are simpler, requiring only cameras that capture the user's face. From these videos of the user's face, machine learning is able to estimate heart rate. This study investigates the benefits and challenges of using machine learning models to estimate heart rate from facial videos, through patents, datasets, and articles review. We searched Derwent Innovation, IEEE Xplore, Scopus, and Web of Science knowledge bases and identified 7 patent filings, 11 datasets, and 20 articles on heart rate, photoplethysmography, or electrocardiogram data. In terms of patents, we note the advantages of inventions related to heart rate estimation, as described by the authors. In terms of datasets, we discovered that most of them are for academic purposes and with different signs and annotations that allow coverage for subjects other than heartbeat estimation. In terms of articles, we discovered techniques, such as extracting regions of interest for heart rate reading and using Video Magnification for small motion extraction, and models such as EVM-CNN and VGG-16, that extract the observed individual's heart rate, the best regions of interest for signal extraction and ways to process them. | ['Erick Giovani Sperandio Nascimento', 'Ingrid Winkler', 'Lian Filipe Santana Nascimento', 'Paulo Henrique Miranda Sá', 'José Vinícius Dantas Paranhos', 'Yasmin da Silva Bonfim', 'Victor Rocha Santos', 'Lucas Lemos Ortega', 'Tiago Palma Pagano'] | 2022-02-17 | null | null | null | null | ['heart-rate-estimation'] | ['medical'] | [ 1.23243131e-01 1.50879934e-01 -8.43487442e-01 3.76759768e-02
-1.85222372e-01 -4.36386466e-01 -2.43028596e-01 -1.09217726e-01
-2.73017257e-01 6.13564372e-01 -1.73934363e-02 -3.89872879e-01
-2.22804341e-02 -4.41344887e-01 -1.97296247e-01 -6.32383049e-01
-1.11094750e-01 -4.54329282e-01 -4.59199101e-01 3.18458408e-01
5.48487425e-01 6.22071683e-01 -1.12919033e+00 -5.62939793e-02
5.16243756e-01 1.44570577e+00 -4.96503681e-01 6.04396760e-01
-3.22488695e-02 5.09347141e-01 -9.01673436e-01 -2.77576149e-01
7.15566650e-02 -6.63611591e-01 -3.19444150e-01 -4.31484103e-01
4.52228159e-01 -6.08036458e-01 -4.01297271e-01 6.54019058e-01
8.49902987e-01 -6.87201381e-01 3.18153232e-01 -1.23767841e+00
-5.56505322e-01 2.22902179e-01 -7.42657602e-01 5.28142512e-01
5.19704580e-01 2.70547420e-01 2.55018741e-01 -3.70062858e-01
5.35617590e-01 7.13659883e-01 9.45112288e-01 6.92604005e-01
-8.01889420e-01 -1.02958548e+00 -4.28503573e-01 1.11973561e-01
-1.38484550e+00 -5.75464904e-01 9.72060442e-01 -6.18708968e-01
7.25838184e-01 3.37635607e-01 1.16045499e+00 1.21711981e+00
6.65737629e-01 2.13966936e-01 9.87293601e-01 -3.76218200e-01
5.26437983e-02 2.55482733e-01 -2.67276555e-01 7.19544172e-01
6.05262399e-01 1.62695751e-01 -6.02580428e-01 -6.93297088e-02
1.23466659e+00 -1.09482035e-02 -5.41684806e-01 9.30009633e-02
-1.19260573e+00 5.49878359e-01 -2.06966549e-01 3.25684518e-01
-3.14785868e-01 3.25710803e-01 3.92748624e-01 2.37659141e-01
4.01357204e-01 6.67489767e-01 -4.74589616e-01 -4.94921982e-01
-1.10401475e+00 -3.51533800e-01 9.21486497e-01 7.37012506e-01
1.91897169e-01 3.11511844e-01 -9.59599018e-03 3.15753132e-01
3.65179896e-01 6.87323630e-01 7.33618975e-01 -1.17561638e+00
4.74592708e-02 5.70883453e-01 -1.47438198e-01 -1.09270191e+00
-5.49986362e-01 -1.56077981e-01 -9.88215208e-01 2.49204170e-02
3.49485815e-01 -4.27639037e-01 -4.21054363e-01 1.34456408e+00
2.37802878e-01 4.84535635e-01 -1.75203815e-01 7.46978104e-01
1.35729110e+00 2.05723539e-01 1.66905046e-01 -8.45720410e-01
1.64666998e+00 -3.23268265e-01 -9.85269308e-01 2.29182974e-01
3.43846053e-01 -4.41101283e-01 6.53951287e-01 6.42283022e-01
-1.07302153e+00 -6.67327642e-01 -9.57865238e-01 1.66328430e-01
-3.08390051e-01 3.90174061e-01 6.53117776e-01 1.13924491e+00
-7.29037285e-01 8.92567813e-01 -6.09964192e-01 -2.52385259e-01
7.79680014e-01 2.15445027e-01 -9.17695239e-02 5.71542859e-01
-1.19775200e+00 8.76713514e-01 -1.69532344e-01 7.77803659e-02
-4.09120440e-01 -1.03731763e+00 -7.14293003e-01 -7.35676438e-02
1.90102249e-01 -5.53453445e-01 8.49153638e-01 -8.54839623e-01
-1.86678576e+00 9.00725305e-01 -2.32932746e-01 -3.04910421e-01
4.05186534e-01 -9.59517732e-02 -7.68383324e-01 6.23874962e-01
-3.52537513e-01 3.75368029e-01 1.21906888e+00 -1.50905341e-01
-2.94836164e-01 -4.57816809e-01 -2.34578773e-01 -2.75319099e-01
-4.59181488e-01 -3.97057645e-02 -3.38590831e-01 -4.42462772e-01
-1.27843663e-01 -8.19902241e-01 1.87007725e-01 2.64235049e-01
-1.33546395e-02 9.13726445e-03 9.34367537e-01 -9.72033262e-01
1.52261543e+00 -2.08467364e+00 -3.25034618e-01 1.57490730e-01
7.08055556e-01 4.79203880e-01 2.33994916e-01 -2.04900473e-01
-9.54466239e-02 7.39334047e-01 3.22108686e-01 4.18040872e-01
-5.37863731e-01 -3.84432346e-01 1.34946212e-01 6.54035747e-01
-1.28898397e-01 1.19055343e+00 -6.00912690e-01 -5.66456079e-01
3.75749707e-01 7.43635237e-01 -3.35462135e-03 -1.84047475e-01
5.35932243e-01 6.37942255e-01 -4.59354132e-01 9.59463775e-01
4.17273790e-01 -2.92035699e-01 1.81253433e-01 -6.09643996e-01
-2.01191664e-01 4.57416847e-02 -8.59224916e-01 1.42606330e+00
-5.25189638e-01 1.16224694e+00 -2.60737568e-01 -7.56408751e-01
1.07629561e+00 6.94304407e-01 9.48821664e-01 -6.87571108e-01
3.76960009e-01 2.41788235e-02 -1.74621716e-01 -9.91573155e-01
-1.78399265e-01 3.93789895e-02 3.09971571e-01 2.01356679e-01
9.31206811e-03 -1.19565308e-01 8.48056632e-04 -2.54756749e-01
8.40761960e-01 1.85906842e-01 6.42992616e-01 9.09564048e-02
1.89274579e-01 -5.63369930e-01 3.90527070e-01 4.06828165e-01
-5.28396308e-01 3.83133143e-01 6.48055911e-01 -6.79078817e-01
-7.14062452e-01 -5.74181080e-01 -2.83040375e-01 2.07608357e-01
-1.48405656e-01 -2.11244285e-01 -7.49689519e-01 -3.12824398e-01
7.44690895e-02 7.02062948e-03 -4.90913868e-01 -1.29025102e-01
-3.89875710e-01 -4.50345248e-01 7.25712538e-01 4.54345703e-01
4.89879370e-01 -1.03405452e+00 -1.36334717e+00 3.57535444e-02
-1.90417320e-01 -1.24676132e+00 -3.92094612e-01 -3.50495756e-01
-1.27269530e+00 -1.24386203e+00 -9.07209277e-01 -2.92949826e-01
1.95149153e-01 -8.85748565e-02 9.21203792e-01 -1.20116271e-01
-1.05579770e+00 6.35774970e-01 2.44914591e-02 -9.03478563e-01
5.88432364e-02 -4.51426245e-02 -5.21108396e-02 -1.38471827e-01
5.01662374e-01 -5.37286937e-01 -8.78320634e-01 1.85440183e-02
-1.95288762e-01 -3.45420033e-01 6.49935067e-01 1.93041824e-02
4.29758012e-01 -2.42917791e-01 7.44729221e-01 -5.74538767e-01
6.87395990e-01 -3.99865836e-01 -5.89023709e-01 6.82316767e-03
-1.09621000e+00 -4.87031907e-01 4.37100194e-02 -6.47395551e-01
-5.53337932e-01 -8.53649899e-02 3.38157475e-01 -7.01279759e-01
-1.93633139e-01 2.60849327e-01 2.51855046e-01 -2.62030751e-01
7.35421240e-01 -3.17085609e-02 5.59890509e-01 -7.81508386e-02
-6.71448857e-02 7.46510327e-01 4.90351558e-01 5.64393550e-02
1.38036609e-01 5.57717264e-01 3.56190115e-01 -1.18740439e+00
-3.91852260e-01 -3.86946797e-01 -4.38040316e-01 -7.21878648e-01
9.75867450e-01 -9.83460128e-01 -1.50836611e+00 4.33675110e-01
-1.01336145e+00 1.69336140e-01 -2.71408528e-01 9.37722504e-01
-3.15009087e-01 3.66808742e-01 -5.27737498e-01 -9.09388483e-01
-8.91399264e-01 -6.62123263e-01 6.72676861e-01 7.56829619e-01
-6.35467291e-01 -8.73562217e-01 -1.38970064e-02 4.30435449e-01
7.58669794e-01 7.64226437e-01 4.74894911e-01 -4.67020366e-03
-2.41576642e-01 -4.48758423e-01 -1.87101518e-03 2.44756550e-01
3.93678486e-01 3.82525086e-01 -1.20672679e+00 -8.06422457e-02
1.17971927e-01 2.06738755e-01 4.00497288e-01 1.12962794e+00
1.61247671e+00 -3.86794835e-01 -4.86115158e-01 7.45276749e-01
1.19263101e+00 6.59809649e-01 1.20713353e+00 -9.50193405e-02
5.40392816e-01 3.99250597e-01 -9.39601380e-03 5.44046044e-01
-9.54029858e-02 3.87685150e-01 2.40944237e-01 -2.57357031e-01
6.02549762e-02 4.20741625e-02 2.95092762e-01 4.65039015e-01
-8.72047007e-01 2.45010853e-01 -4.10147101e-01 1.79795429e-01
-1.09369588e+00 -1.14901268e+00 -2.35318944e-01 2.28822470e+00
5.33476114e-01 -9.30640474e-02 4.37878579e-01 3.32473218e-02
7.97928810e-01 1.01890473e-03 -8.68952930e-01 -4.89115149e-01
2.22690597e-01 6.00650609e-01 7.14611053e-01 -9.13472548e-02
-7.41981804e-01 3.07043016e-01 6.97975826e+00 1.86931148e-01
-1.69430256e+00 -8.31358973e-03 8.81982863e-01 -5.25254488e-01
9.69902724e-02 -1.67449787e-01 -6.41391993e-01 6.26782477e-01
1.27614605e+00 -1.50198862e-01 2.08603054e-01 7.57152855e-01
7.14694619e-01 -1.19760185e-01 -7.61976957e-01 1.55343378e+00
7.75078908e-02 -1.64925563e+00 -4.60995167e-01 2.56181687e-01
1.22134566e-01 -5.80796540e-01 7.80211315e-02 1.51931308e-02
-9.16729212e-01 -1.04201877e+00 1.05297364e-01 9.25870895e-01
1.32756937e+00 -4.53235149e-01 6.43294096e-01 -1.55237645e-01
-9.30276573e-01 -9.05829817e-02 -1.02218457e-01 -3.45344320e-02
-1.33001357e-01 7.99646556e-01 -5.43943048e-01 3.05899262e-01
8.10792983e-01 9.80982780e-01 -2.67343909e-01 9.52655673e-01
-1.27396017e-01 7.15878129e-01 -1.48230135e-01 -1.49044544e-01
-6.01646304e-01 -9.38141868e-02 1.58858582e-01 9.39871490e-01
6.67812169e-01 2.60917157e-01 -5.29718220e-01 1.02418983e+00
-5.84668927e-02 6.37419373e-02 -7.22112179e-01 -3.43586713e-01
4.49960142e-01 1.52833402e+00 -6.41319156e-01 -1.48652494e-01
-6.27920389e-01 6.14500821e-01 -6.76966667e-01 2.96136141e-01
-1.14758956e+00 -8.06029439e-01 4.60909873e-01 6.04782522e-01
-1.76574618e-01 -1.38333011e-02 -4.10565168e-01 -9.69840646e-01
7.77054802e-02 -7.19293118e-01 2.92484671e-01 -8.67001951e-01
-5.08374512e-01 1.93456247e-01 -7.62641877e-02 -1.41648793e+00
-2.24397257e-01 -5.37854254e-01 -6.61806405e-01 8.09282303e-01
-1.42583835e+00 -5.42210817e-01 -6.85296595e-01 3.79062504e-01
3.00946325e-01 -1.71163037e-01 6.56904936e-01 4.42905396e-01
-6.12899125e-01 4.86542463e-01 -6.91792011e-01 2.92698026e-01
6.54982269e-01 -6.71046317e-01 7.74973780e-02 8.35835934e-02
-1.44151285e-01 5.46790659e-01 1.59963489e-01 -4.92829531e-01
-1.55446696e+00 -5.76972783e-01 7.89297283e-01 -5.18895626e-01
1.47568777e-01 1.03985645e-01 -4.82458889e-01 1.97898254e-01
1.02847174e-01 1.57495424e-01 9.12120998e-01 -1.37027055e-01
5.91607951e-02 -4.56761062e-01 -1.23083985e+00 3.21674675e-01
5.75458407e-01 -4.26286191e-01 -1.34378552e-01 2.32533645e-02
1.66883871e-01 -5.39873719e-01 -1.38391423e+00 4.26561266e-01
1.37435317e+00 -6.68021917e-01 9.81508374e-01 -4.70660001e-01
4.18743163e-01 4.52764295e-02 6.05522096e-01 -8.67521822e-01
-1.00715220e-01 -9.82818961e-01 -5.31416774e-01 1.02424610e+00
4.12850827e-01 -7.47661889e-01 9.39159930e-01 7.73735225e-01
3.07022125e-01 -8.43229413e-01 -8.02358389e-01 -6.33214891e-01
-4.03552294e-01 -1.22220419e-01 3.25126946e-01 9.71267760e-01
4.04064178e-01 2.25708440e-01 -4.23044622e-01 -3.38787317e-01
4.16585743e-01 -1.11825749e-01 6.31754875e-01 -1.44876790e+00
1.64563339e-02 -4.66533989e-01 -5.50286829e-01 -4.86306041e-01
-1.72791526e-01 -3.81133080e-01 -7.54966795e-01 -1.40551591e+00
9.52863693e-02 1.46247029e-01 -2.88646638e-01 4.97405559e-01
1.85307309e-01 3.90095383e-01 -6.54136911e-02 2.81622648e-01
4.65235896e-02 -3.06721538e-01 1.38639295e+00 3.20319599e-03
-6.13287270e-01 1.24103673e-01 -8.18630338e-01 7.40387440e-01
8.75232279e-01 -2.45607913e-01 -3.61120671e-01 2.85562307e-01
3.47628355e-01 4.50524569e-01 5.09966373e-01 -1.01551008e+00
-8.18703771e-02 -1.13263980e-01 9.49687600e-01 -1.10183343e-01
1.93822056e-01 -5.65436065e-01 3.74332130e-01 7.03464568e-01
-2.37518862e-01 -4.90640178e-02 2.99560219e-01 2.08280370e-01
-9.07175094e-02 -1.32699221e-01 8.50214362e-01 -4.25282747e-01
-3.70125115e-01 3.77051055e-01 -4.98885274e-01 6.79614767e-02
1.12648904e+00 -6.53966129e-01 -4.57257539e-01 -7.30092883e-01
-7.92119205e-01 -3.02060425e-01 -1.35821745e-01 3.87617916e-01
6.79030180e-01 -1.06285346e+00 -5.39498687e-01 2.00334474e-01
-6.18519187e-02 -5.91555297e-01 4.04449165e-01 1.32427847e+00
-4.31992590e-01 4.34543014e-01 -4.65654135e-01 -5.42704284e-01
-1.36693668e+00 4.39032882e-01 5.09441912e-01 2.41729438e-01
-5.11818051e-01 2.28367135e-01 -2.97091126e-01 4.77613628e-01
2.77774870e-01 -6.82455063e-01 -5.92060506e-01 4.54264015e-01
5.85070372e-01 7.14258611e-01 9.67688039e-02 -3.27528380e-02
-3.45613867e-01 9.35272753e-01 4.12823528e-01 3.12981009e-01
9.42198157e-01 -1.49023384e-01 7.90102035e-02 3.38227719e-01
1.17996836e+00 -1.23982914e-02 -8.17938983e-01 5.71055651e-01
-5.38926780e-01 -3.86756599e-01 2.02844087e-02 -6.48282826e-01
-1.64123297e+00 9.73279119e-01 1.06962228e+00 1.61348984e-01
1.40351522e+00 -3.44493896e-01 6.20100319e-01 -9.90708619e-02
1.45909622e-01 -1.32453036e+00 1.72837377e-01 -1.71200529e-01
5.49767435e-01 -9.53203142e-01 3.39924276e-01 -2.90554702e-01
-4.62931454e-01 1.60680568e+00 3.94295126e-01 3.53738576e-01
9.18183804e-01 2.19496280e-01 1.92247123e-01 -2.24965081e-01
-4.20893371e-01 2.52749920e-01 2.18597889e-01 5.83592892e-01
7.16185153e-01 -1.78622119e-02 -5.35762906e-01 4.08750713e-01
1.30445203e-02 7.90898561e-01 6.11415923e-01 7.55751729e-01
-2.19719738e-01 -5.29881597e-01 -1.91529214e-01 8.06116879e-01
-9.05297935e-01 -1.76006109e-02 -3.32063764e-01 8.22275162e-01
1.96977615e-01 9.36535299e-01 -6.85277656e-02 -2.11637482e-01
3.54509771e-01 2.80293465e-01 6.38261080e-01 -1.54511586e-01
-3.07722598e-01 6.18575551e-02 -1.39239222e-01 -7.34687030e-01
-5.94781876e-01 -4.52602923e-01 -1.01228476e+00 -2.55202264e-01
-3.00489902e-01 -2.12522477e-01 1.13033330e+00 7.72450089e-01
6.52900636e-01 5.22360802e-01 6.09370112e-01 -4.85934168e-01
4.23876867e-02 -9.31880772e-01 -7.34909117e-01 6.01529144e-02
3.31007898e-01 -4.27362382e-01 -4.71130699e-01 4.34881479e-01] | [13.856539726257324, 2.833507537841797] |
83ef974c-2c72-4afc-81cf-a82509c361f8 | best-of-both-worlds-robust-accented-speech | 2103.05834 | null | https://arxiv.org/abs/2103.05834v1 | https://arxiv.org/pdf/2103.05834v1.pdf | Best of Both Worlds: Robust Accented Speech Recognition with Adversarial Transfer Learning | Training deep neural networks for automatic speech recognition (ASR) requires large amounts of transcribed speech. This becomes a bottleneck for training robust models for accented speech which typically contains high variability in pronunciation and other semantics, since obtaining large amounts of annotated accented data is both tedious and costly. Often, we only have access to large amounts of unannotated speech from different accents. In this work, we leverage this unannotated data to provide semantic regularization to an ASR model that has been trained only on one accent, to improve its performance for multiple accents. We propose Accent Pre-Training (Acc-PT), a semi-supervised training strategy that combines transfer learning and adversarial training. Our approach improves the performance of a state-of-the-art ASR model by 33% on average over the baseline across multiple accents, training only on annotated samples from one standard accent, and as little as 105 minutes of unannotated speech from a target accent. | ['Duen Horng Chau', 'Sundararajan Srinivasan', 'Monica Sunkara', 'Sravan Bodapati', 'Nilaksh Das'] | 2021-03-10 | null | null | null | null | ['accented-speech-recognition'] | ['speech'] | [ 2.08941057e-01 3.34621519e-01 1.19586341e-01 -6.68251574e-01
-1.23254681e+00 -9.30311620e-01 3.75264376e-01 -1.77428901e-01
-7.35583484e-01 6.44177854e-01 3.36975694e-01 -5.78247488e-01
7.66058624e-01 -3.07712406e-01 -7.80803978e-01 -4.17713702e-01
3.76214892e-01 8.61528099e-01 -1.68876126e-01 -4.11760658e-01
-5.40392637e-01 4.54222679e-01 -9.32813823e-01 8.02445486e-02
9.05558228e-01 8.33574235e-01 3.44769210e-01 7.44144440e-01
-1.51086792e-01 8.36906910e-01 -1.22217643e+00 -4.31574851e-01
2.18449175e-01 -3.59909356e-01 -8.55599582e-01 2.24793404e-01
5.38527608e-01 -1.46195427e-01 -4.12820518e-01 9.73981917e-01
5.72349429e-01 4.83745903e-01 2.28825420e-01 -7.86819220e-01
-7.32313216e-01 1.06645453e+00 -1.22478064e-02 1.53695971e-01
-1.81016997e-02 2.60619491e-01 9.08494771e-01 -9.26865757e-01
2.77137190e-01 1.20777702e+00 3.56743574e-01 9.44174290e-01
-1.28920662e+00 -7.23503888e-01 3.30276191e-01 -1.77329034e-01
-1.33808982e+00 -1.08105958e+00 8.44722688e-01 -8.33210424e-02
9.57160175e-01 2.82875270e-01 2.99353212e-01 1.48680544e+00
-6.02899373e-01 9.21847284e-01 1.11356890e+00 -3.23488414e-01
4.61752027e-01 -3.60583737e-02 6.72302097e-02 1.19599476e-01
-4.79435921e-01 -4.10131365e-03 -5.14055729e-01 2.64785409e-01
4.86814708e-01 -3.77723634e-01 -2.88076520e-01 2.62289524e-01
-1.22215772e+00 7.68183768e-01 3.49003255e-01 5.33581786e-02
-3.74752432e-01 -3.27094883e-01 4.44174349e-01 5.11417806e-01
7.04616666e-01 7.35918880e-01 -9.08724725e-01 -4.43944931e-01
-1.02677524e+00 -7.59750083e-02 7.89053321e-01 9.26994085e-01
7.00299799e-01 7.18460083e-01 2.96592619e-02 1.47376084e+00
-5.14229275e-02 9.15156007e-01 6.25384808e-01 -7.08846033e-01
6.30831301e-01 2.38706619e-01 -3.39566432e-02 -6.88085854e-02
-4.08865027e-02 -4.65142041e-01 -8.64615560e-01 7.90803656e-02
5.12384057e-01 -5.44816375e-01 -1.64856553e+00 1.94765401e+00
4.05899473e-02 1.06935650e-01 3.89255345e-01 1.01441717e+00
5.47454119e-01 9.75614369e-01 2.98187137e-01 -1.37341976e-01
1.11499786e+00 -1.39411759e+00 -9.12821531e-01 -7.42425978e-01
4.87468153e-01 -9.66882467e-01 1.55956984e+00 1.73829496e-01
-1.06752503e+00 -6.92719996e-01 -8.87460828e-01 -1.47341555e-02
-5.48072278e-01 1.19669810e-01 1.10144548e-01 6.23290479e-01
-1.16563070e+00 7.00909942e-02 -8.75323951e-01 6.90070763e-02
3.20894271e-01 4.43773806e-01 -4.48775887e-01 -8.98555815e-02
-1.49561346e+00 8.93127859e-01 3.74798954e-01 1.28442138e-01
-9.69176650e-01 -9.51501191e-01 -1.12248874e+00 4.62120585e-02
4.24986452e-01 -1.27934352e-01 1.64718616e+00 -1.36566770e+00
-1.90956700e+00 6.30823374e-01 -1.98581532e-01 -7.33990133e-01
3.75191331e-01 -5.62144279e-01 -6.84681416e-01 -2.51087219e-01
-2.11977169e-01 6.28274322e-01 8.42755437e-01 -1.09248245e+00
-2.12383896e-01 -3.44968855e-01 -1.85791120e-01 3.46198380e-01
-1.98574200e-01 3.44469070e-01 -1.94552720e-01 -1.05225682e+00
-2.71579504e-01 -1.02909493e+00 -2.74362624e-01 -8.93677592e-01
-5.43327153e-01 -1.79751515e-01 9.03268814e-01 -1.00202394e+00
1.01201510e+00 -2.21120644e+00 1.26743004e-01 -1.20896734e-02
-1.69977203e-01 8.28919649e-01 -3.11273485e-01 -7.68275410e-02
-1.53117880e-01 1.48867950e-01 -3.32791150e-01 -6.74986780e-01
5.54130524e-02 3.89320701e-01 -5.69770575e-01 -6.30645826e-03
3.81121874e-01 7.89896011e-01 -9.04698670e-01 -9.41973552e-02
4.29192632e-01 5.98408639e-01 -2.63369739e-01 8.51637661e-01
-3.62050951e-01 6.64860427e-01 3.46081071e-02 4.89379406e-01
6.24705851e-01 2.20439017e-01 1.16546057e-01 1.03443265e-01
1.71616375e-01 9.49887753e-01 -9.18925762e-01 1.66254556e+00
-9.53144491e-01 4.00110304e-01 3.50863874e-01 -8.63925517e-01
1.13038170e+00 6.81525826e-01 1.99564546e-02 -5.20084858e-01
-8.08503851e-02 4.15336430e-01 1.05757289e-01 1.78297833e-01
4.55847472e-01 -3.53926867e-01 -3.50440055e-01 3.41086626e-01
3.51390302e-01 -4.00057167e-01 -1.37164906e-01 -7.29921274e-03
1.20865428e+00 -4.10999209e-01 4.40220237e-02 5.55010792e-03
4.87072468e-01 -3.20461214e-01 6.71433389e-01 6.24659240e-01
-3.56241137e-01 8.63643646e-01 2.62835193e-02 -3.70115280e-01
-1.26513970e+00 -1.23352671e+00 1.86232820e-01 1.41611135e+00
-4.86683130e-01 -1.78064704e-01 -1.10812497e+00 -9.18216288e-01
-2.88882017e-01 1.06363523e+00 -2.97148317e-01 -9.77730379e-02
-9.36210513e-01 -4.00688797e-01 7.80051470e-01 7.26995349e-01
3.54590505e-01 -1.47635174e+00 2.53536582e-01 2.57476479e-01
-1.95011750e-01 -1.42459929e+00 -8.83861840e-01 4.42605674e-01
-2.40737230e-01 -3.32108945e-01 -7.85466790e-01 -7.30886340e-01
5.34356117e-01 -2.55148798e-01 1.39541161e+00 -2.88995683e-01
3.42766136e-01 -1.26756608e-01 -4.43280727e-01 -5.02451420e-01
-1.07769036e+00 6.70184612e-01 3.53002638e-01 2.29144901e-01
4.91075456e-01 -4.31215912e-01 -8.89028758e-02 3.92727435e-01
-7.44281292e-01 -1.16730787e-01 6.49029911e-01 9.90679026e-01
4.69333082e-01 -4.65396076e-01 1.03406048e+00 -1.10580921e+00
3.47657859e-01 -2.49203980e-01 -6.68515146e-01 9.41985175e-02
-2.12333396e-01 1.01094186e-01 1.14837420e+00 -6.34237170e-01
-1.10317922e+00 1.91965118e-01 -6.30663395e-01 -8.67492259e-01
-6.13926232e-01 3.25989544e-01 -6.43720806e-01 5.25289118e-01
5.25844693e-01 2.25781262e-01 1.07474349e-01 -5.42619169e-01
6.55502677e-01 1.18806779e+00 7.66097486e-01 -4.73019034e-01
7.15659022e-01 -3.08604419e-01 -7.30696142e-01 -1.04231787e+00
-1.25055790e+00 -4.11933035e-01 -6.62710428e-01 1.14020951e-01
7.27452278e-01 -1.03876889e+00 -1.15368903e-01 7.72630870e-01
-1.20093119e+00 -8.11398268e-01 -3.51735741e-01 4.20481473e-01
-4.09266919e-01 -1.65472582e-01 -6.51285887e-01 -7.93260515e-01
-4.45730805e-01 -1.25854194e+00 9.46404576e-01 -2.82285828e-02
-4.33996350e-01 -9.24306512e-01 5.52739128e-02 7.72304356e-01
7.16728866e-01 -3.72864246e-01 4.43282753e-01 -1.37930822e+00
-3.66339713e-01 -1.41988963e-01 2.62273997e-01 1.05897629e+00
4.88367677e-01 -2.71373242e-01 -1.29879653e+00 -3.88548851e-01
-1.30650491e-01 -8.00474346e-01 7.39353299e-01 1.53551787e-01
1.16222548e+00 -5.64532697e-01 2.18058065e-01 4.58948135e-01
6.83626771e-01 2.74967819e-01 5.48318326e-01 2.17663515e-02
9.42853689e-01 4.11740273e-01 4.94303256e-01 -1.45198435e-01
4.48935404e-02 6.80958271e-01 -1.15520309e-03 -2.81401783e-01
-2.71852255e-01 -3.15850317e-01 5.39652467e-01 1.49453151e+00
2.37944886e-01 -4.55908120e-01 -9.48668540e-01 8.87246549e-01
-1.41449177e+00 -7.16752827e-01 6.54999375e-01 2.33478785e+00
1.44729543e+00 1.24353707e-01 1.48090854e-01 -7.17659518e-02
1.02110612e+00 4.78385568e-01 -5.87526441e-01 -5.26497543e-01
-2.36420572e-01 6.42968833e-01 3.05899948e-01 8.94307077e-01
-1.34063923e+00 1.51101732e+00 6.20685482e+00 8.80218923e-01
-1.50670171e+00 1.59426093e-01 8.94162357e-01 -8.93575177e-02
-1.40438512e-01 -2.63215482e-01 -7.00715721e-01 6.78830922e-01
1.71376300e+00 1.21125191e-01 8.20368230e-01 1.12663639e+00
1.81243718e-01 5.58787405e-01 -9.34356093e-01 7.08443999e-01
-7.74505828e-03 -9.13545609e-01 -2.36910004e-02 -1.38377085e-01
6.30921960e-01 4.54142362e-01 1.41695008e-01 6.09565794e-01
7.90210783e-01 -1.30222702e+00 6.09301090e-01 -2.04422474e-01
9.82863367e-01 -9.55503881e-01 8.24912190e-01 2.00109616e-01
-7.93773532e-01 3.49344343e-01 -1.86673209e-01 1.16364419e-01
3.72881144e-01 4.78069872e-01 -1.24326801e+00 2.08993599e-01
2.72177815e-01 1.88474327e-01 -2.90736884e-01 2.69534260e-01
-4.52566564e-01 1.37379062e+00 -4.54373628e-01 1.88074023e-01
1.87692180e-01 -9.94087234e-02 4.98433620e-01 1.34721303e+00
-1.36714401e-02 -1.24435499e-01 3.69379967e-01 5.86154521e-01
-6.18367910e-01 2.33038832e-02 -5.67979336e-01 -4.04773533e-01
9.73149598e-01 1.12776220e+00 -2.60599643e-01 -4.36023772e-01
-4.35062915e-01 1.18354452e+00 7.35336602e-01 5.29207766e-01
-6.68993473e-01 -4.77713674e-01 9.75285828e-01 -5.37976027e-02
3.16426039e-01 -3.45166802e-01 -2.82494962e-01 -1.11575568e+00
-1.03342541e-01 -1.31815839e+00 2.17920959e-01 -5.97462416e-01
-1.36636245e+00 1.11047804e+00 -5.68708420e-01 -6.67342782e-01
-6.43746793e-01 -5.55920541e-01 -4.89519119e-01 1.32424402e+00
-1.33680391e+00 -1.22498322e+00 5.90521172e-02 4.47312653e-01
1.07950914e+00 -5.22355378e-01 1.04041147e+00 3.07863176e-01
-6.28242254e-01 9.99070525e-01 -2.12069631e-01 7.01181769e-01
8.98544312e-01 -1.67667449e+00 1.01430535e+00 1.18683088e+00
4.25633788e-01 4.69636619e-01 5.51363587e-01 -3.79366428e-01
-9.28869486e-01 -1.53922224e+00 8.54985654e-01 -5.76062024e-01
7.69700944e-01 -8.06727290e-01 -1.26353478e+00 1.16285908e+00
4.36930299e-01 2.70545393e-01 6.93961263e-01 4.34594244e-01
-3.66630584e-01 -1.83497995e-01 -7.51417816e-01 7.11312354e-01
7.16276765e-01 -9.28933263e-01 -8.31417203e-01 9.82795283e-02
1.31281590e+00 -5.85842550e-01 -8.02816272e-01 2.79059887e-01
-2.14905851e-02 -3.45668286e-01 8.16270471e-01 -7.85292029e-01
-8.44134092e-02 -3.00583482e-01 -2.80066192e-01 -2.08208704e+00
1.67963058e-01 -9.66357291e-01 3.48232649e-02 1.50239050e+00
9.13629472e-01 -5.31364262e-01 5.50123334e-01 7.00080156e-01
-6.65961266e-01 -3.27642143e-01 -9.14344311e-01 -9.91506457e-01
2.27587670e-01 -4.73776698e-01 7.39982963e-01 1.21248710e+00
-2.25578457e-01 7.85941541e-01 -2.86833912e-01 3.06024343e-01
2.06499159e-01 -3.49796414e-01 7.97600567e-01 -7.79172003e-01
-1.17902160e-01 -5.09687103e-02 -3.14143926e-01 -1.03051519e+00
8.43322456e-01 -8.20360899e-01 4.79956061e-01 -9.02863741e-01
-5.22339404e-01 -5.96299767e-01 -4.50802624e-01 8.51692438e-01
-4.84003276e-01 1.32416382e-01 1.46768749e-01 -1.39834180e-01
-4.32516694e-01 7.61311233e-01 9.35565412e-01 -2.07930714e-01
-2.79064864e-01 9.97825265e-02 -4.68258440e-01 5.95101237e-01
9.74542558e-01 -3.86523813e-01 -4.92291272e-01 -5.94234586e-01
-5.11827171e-01 -6.46074340e-02 -9.69023630e-02 -8.99553895e-01
6.72131106e-02 -2.43121400e-01 1.52255818e-01 -2.92229027e-01
6.72465503e-01 -4.94092464e-01 -1.98372260e-01 -2.28105277e-01
-6.28408015e-01 2.05705455e-03 3.28441381e-01 1.83889627e-01
-6.03840053e-01 -2.88492385e-02 9.75043297e-01 -1.43657759e-01
-5.35731673e-01 2.20767319e-01 -3.24347526e-01 5.75056851e-01
4.56879020e-01 4.42157954e-01 -2.87111461e-01 -5.01618266e-01
-9.77408350e-01 1.53446896e-03 4.19422925e-01 6.87588274e-01
3.44454408e-01 -1.24765325e+00 -8.12559426e-01 5.10528982e-01
-1.36082247e-01 4.03014094e-01 1.70438737e-01 2.42884889e-01
-2.05199450e-01 3.58143240e-01 5.48281968e-02 -3.55337679e-01
-1.12761366e+00 6.65856659e-01 3.56392175e-01 -1.40995160e-01
-3.17800909e-01 9.59894836e-01 2.28554696e-01 -1.08437121e+00
2.47763470e-01 -1.59354806e-01 2.23746285e-01 -2.69465029e-01
4.50132787e-01 6.98476285e-02 3.86037976e-01 -8.36087048e-01
-3.46975565e-01 -6.16579205e-02 -5.00539303e-01 -5.19506514e-01
1.00939322e+00 -3.17755230e-02 1.61604658e-01 5.93069017e-01
1.10337484e+00 3.79120618e-01 -1.45652437e+00 -1.64090484e-01
-2.13176176e-01 -1.02267608e-01 2.10652918e-01 -1.09001863e+00
-1.01899314e+00 9.29118693e-01 1.87995926e-01 1.91322699e-01
9.02462840e-01 1.83309913e-01 9.98722792e-01 5.08536339e-01
9.26577896e-02 -1.19949996e+00 -9.62888747e-02 9.05363977e-01
9.55552638e-01 -1.41234219e+00 -8.69915307e-01 -2.89298654e-01
-1.04615235e+00 7.11882353e-01 7.93541968e-01 9.11140516e-02
5.78977108e-01 4.85448629e-01 9.02773917e-01 3.51284236e-01
-6.91123128e-01 -1.61092699e-01 1.56472072e-01 6.47723377e-01
5.46864450e-01 5.21164477e-01 4.64173615e-01 6.85323119e-01
-6.99593842e-01 -5.32637894e-01 4.25443769e-01 7.09493697e-01
-2.11602762e-01 -1.20300150e+00 -2.24644139e-01 1.61237746e-01
-6.06692851e-01 -4.83176619e-01 -5.12517512e-01 4.52881515e-01
-4.02227789e-01 1.03254545e+00 3.39912206e-01 -2.94634044e-01
3.53136301e-01 5.65193415e-01 -1.29586399e-01 -8.91276360e-01
-4.25384492e-01 4.01323028e-02 3.55119556e-01 -3.04583549e-01
-5.88225685e-02 -5.49213409e-01 -1.13447738e+00 3.55475694e-02
-1.52817577e-01 3.15297484e-01 7.57468343e-01 1.02899194e+00
4.11867499e-01 5.50395370e-01 8.07118058e-01 -5.66348255e-01
-7.60294974e-01 -1.42653871e+00 -4.92201269e-01 5.68843365e-01
6.37293220e-01 -3.64965290e-01 -5.51744461e-01 3.02703381e-01] | [14.42362117767334, 6.700058937072754] |
76a47ead-b2e9-46c7-8dfe-8a1b228020f2 | model-based-versus-model-free-feeding-control | 2306.09915 | null | https://arxiv.org/abs/2306.09915v1 | https://arxiv.org/pdf/2306.09915v1.pdf | Model-based versus model-free feeding control and water quality monitoring for fish growth tracking in aquaculture systems | The high concentration level of the environmental factors, such as a high ammonia concentration and pH level, affect the water quality, affecting fish's survival and mass death. Therefore, there is a critical need to develop control strategies to determine optimal, efficient, and reliable feeding and water quality monitoring processes. In this paper, we revisit the representative fish growth model describing the total biomass change by incorporating the fish population density and mortality. Since the measurement data of the total biomass and population from the aquaculture systems are limited and difficult to obtain, we validate the new dynamic population model with the individual fish growth data for tracking control purposes. We specifically focus on relative feeding as a manipulated variable to design traditional and optimal control to track the desired weight reference within the sub-optimal temperature and dissolved oxygen profiles under different levels of unionized ammonia exposure. Then, we propose a Q-learning approach that learns an optimal feeding control policy from the simulated data of the fish growth weight trajectories while managing the ammonia effects. The proposed Q-learning feeding control prevents fish mortality and achieves good tracking errors of the fish weight under the different levels of unionized ammonia. However, it maintains a relative food consumption that potentially underfeeds the fish. Finally, we propose an optimal algorithm that optimizes the feeding and water quality of the dynamic fish population growth process. We also show that the model predictive control decreases fish mortality and reduces food consumption in all different cases of unionized ammonia exposure. | ['Taous-Meriem Laleg-Kirati', "Ibrahima N'Doye", 'Fahad Aljehani'] | 2023-06-14 | null | null | null | null | ['q-learning'] | ['methodology'] | [-3.79974097e-01 -2.30432674e-01 4.11977544e-02 2.39599898e-01
2.03016013e-01 -4.38591093e-01 -1.59845725e-01 5.22359371e-01
-4.56816256e-01 7.58910716e-01 -6.16914928e-02 1.19078293e-01
-5.23808241e-01 -9.21286047e-01 -7.83137143e-01 -1.38601184e+00
-4.19718862e-01 -2.51010228e-02 -9.33112055e-02 -1.69837087e-01
3.00749511e-01 4.24118847e-01 -1.68248653e+00 -6.05793238e-01
9.48091388e-01 7.58030295e-01 5.99816859e-01 7.91095495e-01
-3.35471444e-02 2.16615468e-01 -7.59119630e-01 2.07460880e-01
5.83885193e-01 -6.64395690e-01 7.33819157e-02 1.79222316e-01
-1.51157543e-01 -3.90775323e-01 2.48157829e-02 1.29255545e+00
6.84977710e-01 2.65093356e-01 5.71768105e-01 -1.22515261e+00
-4.87737060e-01 8.09273183e-01 -5.71320832e-01 9.86393392e-02
-4.52604562e-01 9.48135853e-01 2.07428947e-01 -9.85847190e-02
-3.42343479e-01 1.47400677e+00 6.22508705e-01 7.81289876e-01
-1.24185419e+00 -9.01677787e-01 4.37234819e-01 -2.07898870e-01
-1.30725884e+00 -1.77451093e-02 -6.47655204e-02 -5.41321278e-01
5.15656471e-01 -1.83038473e-01 1.46747696e+00 2.19397306e-01
9.32842016e-01 1.22156911e-01 8.52323472e-01 -2.30725154e-01
4.31101233e-01 -3.61163616e-01 -3.29186499e-01 3.09740812e-01
8.94544661e-01 4.92045820e-01 -1.12904958e-01 -4.01666798e-02
9.26381350e-01 4.07417417e-01 -5.08094192e-01 -1.11007862e-01
-6.92252874e-01 7.56903946e-01 2.78101891e-01 -2.28318945e-01
-5.59608459e-01 4.52421367e-01 6.41201884e-02 6.95051193e-01
2.98141420e-01 5.52574039e-01 -1.14082086e+00 3.29800844e-01
-4.17906523e-01 4.03905720e-01 1.23254907e+00 9.28442657e-01
7.08197236e-01 3.73041093e-01 -2.21333161e-01 5.07912219e-01
9.36486304e-01 1.65695155e+00 3.72285485e-01 -9.24701273e-01
-1.77653566e-01 5.97783267e-01 6.66527450e-01 -7.02149808e-01
-4.73668307e-01 -4.02131885e-01 -7.91015744e-01 3.80181968e-01
4.01010334e-01 -8.99704754e-01 -7.41262197e-01 1.74334741e+00
3.93051893e-01 8.66885930e-02 5.77251732e-01 8.66029918e-01
2.40822598e-01 1.02121484e+00 4.49872941e-01 -1.29843104e+00
1.30121362e+00 -3.39111388e-01 -1.32449079e+00 1.95864782e-01
4.89202291e-01 -2.97544718e-01 7.20113337e-01 1.21698752e-01
-1.11925495e+00 -2.17422545e-02 -9.70204592e-01 1.12545204e+00
-2.85518199e-01 -1.90734595e-01 -1.80666968e-02 2.77134806e-01
-7.74026871e-01 9.36781406e-01 -1.29314780e+00 -4.54006970e-01
1.20746559e-02 5.16524255e-01 2.80476898e-01 2.21841037e-01
-1.06153822e+00 9.72319186e-01 2.87906826e-01 5.64762712e-01
-1.43995750e+00 -1.11601603e+00 -7.41966963e-01 -1.07493429e-02
2.23140106e-01 -4.67050552e-01 1.46408129e+00 -6.04637563e-01
-1.87038016e+00 -5.78653663e-02 4.29643989e-01 -2.53368020e-01
3.12873542e-01 -4.26285118e-01 7.28145018e-02 -1.30734935e-01
-2.03871012e-01 3.95958811e-01 4.21002150e-01 -9.98238146e-01
-9.69217718e-01 -2.98138231e-01 -2.09982529e-01 6.46143734e-01
-4.65580165e-01 -2.61961490e-01 3.92801762e-01 -2.53636777e-01
-1.60912216e-01 -7.26821542e-01 -6.81592047e-01 2.43971467e-01
3.71839374e-01 -1.37001455e-01 5.63291967e-01 -1.36467651e-01
9.83250380e-01 -1.96435237e+00 1.69882059e-01 -1.43372938e-01
-4.13062543e-01 4.42318060e-02 -1.53046072e-01 5.53453147e-01
6.06442690e-01 5.76283894e-02 -1.16771668e-01 3.60428184e-01
-2.96238273e-01 5.78816712e-01 3.45651120e-01 7.81217396e-01
4.15914387e-01 2.29232997e-01 -1.31592798e+00 -2.38187239e-01
1.47021199e-02 3.78745705e-01 -6.07290268e-01 7.23429084e-01
-3.30770671e-01 3.89171481e-01 -5.07475257e-01 4.64514881e-01
7.48932958e-01 3.39598894e-01 7.40160197e-02 -1.40410796e-01
-9.23712432e-01 -6.90139592e-01 -1.30264592e+00 8.83860171e-01
-3.01462024e-01 -1.37385771e-01 6.18211925e-01 -7.03204572e-01
1.15760100e+00 4.32669431e-01 8.10562372e-01 -2.86177546e-01
4.04907346e-01 1.99031785e-01 2.74100065e-01 -9.72291946e-01
-3.11229285e-02 -7.07522273e-01 3.37485045e-01 7.38691688e-02
-1.26514137e-01 -2.49637336e-01 1.74849212e-01 -4.23823148e-01
6.04015708e-01 -7.94708654e-02 3.56682390e-01 -1.22287941e+00
5.22237122e-01 6.85543567e-02 1.05798173e+00 2.75227547e-01
-4.27452326e-01 -3.82121861e-01 4.31807280e-01 -2.16788232e-01
-8.97563636e-01 -5.45159101e-01 -2.78072774e-01 1.06186593e+00
7.14853168e-01 5.32617092e-01 -5.01181722e-01 9.92052034e-02
3.74490798e-01 3.94653022e-01 -5.71650028e-01 -4.81847167e-01
-6.45235360e-01 -1.32850468e+00 1.44964620e-01 2.88004041e-01
3.41743797e-01 -1.10491502e+00 -1.00949502e+00 5.92413247e-01
3.80262703e-01 -1.65296331e-01 -3.97620320e-01 6.33270919e-01
-1.12107539e+00 -1.27718627e+00 -6.06487036e-01 -8.25077355e-01
7.05822647e-01 -3.27732116e-02 7.11602986e-01 3.33469570e-01
2.07968913e-02 2.68340349e-01 -4.01609272e-01 -9.88564610e-01
-7.35812247e-01 -1.89579234e-01 4.72514987e-01 -3.20741892e-01
4.06444371e-01 -2.68995553e-01 -1.08632863e+00 3.58713001e-01
-8.55294168e-01 -4.47109252e-01 5.72704256e-01 8.14921379e-01
5.26659966e-01 2.51808256e-01 8.63933682e-01 -1.37830839e-01
3.13303351e-01 -5.38888991e-01 -1.16212404e+00 2.06303209e-01
-7.79505253e-01 1.32624820e-01 5.68245649e-01 -8.89118075e-01
-6.78900599e-01 2.60951370e-01 2.74064004e-01 -2.02831686e-01
-3.12880017e-02 3.89961094e-01 -5.64728118e-02 1.63526639e-01
3.72340143e-01 1.13747783e-01 7.21962571e-01 -4.59205866e-01
-3.30418736e-01 4.07213032e-01 -5.33362068e-02 -2.81293541e-01
7.90669560e-01 -3.72076742e-02 3.82467777e-01 -7.87373841e-01
-5.51104665e-01 -2.44148880e-01 -4.03363466e-01 -4.78040040e-01
7.00670958e-01 -1.08085310e+00 -1.40436435e+00 8.38322699e-01
-6.17235541e-01 -8.07024896e-01 -3.75981748e-01 8.81421387e-01
-5.89837134e-01 7.01615401e-03 -7.69812047e-01 -1.31069982e+00
-6.25758648e-01 -1.08425617e+00 4.55446184e-01 5.50120473e-01
6.37132168e-01 -9.99985337e-01 4.18767869e-01 -9.20450509e-01
6.40757561e-01 2.88877517e-01 6.10628009e-01 -1.11364491e-01
-2.69953370e-01 2.89278299e-01 3.53531986e-01 1.86735526e-01
4.75545079e-01 4.29059178e-01 -1.64261296e-01 -8.66783679e-01
6.75253337e-03 -1.33694336e-01 3.39388549e-01 9.31285203e-01
2.14473888e-01 -6.55622721e-01 -6.29643947e-02 6.61706567e-01
1.92742872e+00 7.19851732e-01 7.80967101e-02 4.78401780e-01
1.10147625e-01 9.18676436e-01 8.10582161e-01 1.08937049e+00
-5.56090586e-02 7.71581903e-02 1.15328026e+00 7.91486204e-02
4.77684200e-01 -5.32506555e-02 6.84539080e-01 8.47420692e-01
-2.14853529e-02 -3.70519429e-01 -6.26365125e-01 6.82611108e-01
-1.64650428e+00 -8.21257591e-01 -2.78328657e-01 2.25102115e+00
1.00415480e+00 -8.43720615e-01 2.06741944e-01 -1.98747158e-01
8.84796977e-01 -6.78614974e-01 -8.94552231e-01 -4.51495424e-02
5.74111426e-03 -3.47745538e-01 1.22849774e+00 4.22165006e-01
-7.25387394e-01 5.15125573e-01 6.66550064e+00 -1.01997487e-01
-1.11058056e+00 -3.67198795e-01 1.07631594e-01 -2.37732530e-01
1.97975755e-01 -4.00245756e-01 -1.14263177e+00 3.43355000e-01
1.13585329e+00 -5.43239295e-01 3.70388925e-01 4.38774228e-01
1.09594059e+00 7.14844605e-03 -8.04592133e-01 2.71725923e-01
-4.66876537e-01 -4.05724496e-01 -9.52503085e-02 1.31191820e-01
7.01545298e-01 -5.13115386e-03 -4.19138849e-01 2.90090889e-01
7.83338487e-01 -4.75927293e-01 8.24695945e-01 6.59150660e-01
4.50739801e-01 -6.96213901e-01 1.20547175e+00 6.81670785e-01
-1.26667225e+00 -8.02418649e-01 -8.44310582e-01 -4.87015665e-01
1.87101483e-01 1.15488008e-01 -2.47407705e-01 -3.16786617e-02
7.46959746e-01 7.13080943e-01 -8.96329880e-02 1.59696043e+00
1.48805445e-02 4.71891940e-01 -5.99129617e-01 -5.17098248e-01
5.74734621e-02 -3.63862962e-01 3.91560316e-01 8.97511065e-01
7.58696020e-01 7.06070483e-01 4.37740624e-01 6.13798797e-01
3.43700349e-01 4.83820230e-01 -2.41470814e-01 1.07875466e-01
7.01060236e-01 6.64640009e-01 -4.78199214e-01 -3.54373753e-02
-2.96814870e-02 -3.51381488e-02 -2.68792689e-01 2.64970630e-01
-5.33026934e-01 -1.46795377e-01 1.29691887e+00 -8.01163614e-02
1.49862066e-01 -5.83094172e-03 2.27137536e-01 -8.69350374e-01
-6.86820388e-01 -3.83060843e-01 4.31458026e-01 -4.31743860e-02
-1.46046686e+00 2.50915252e-02 4.22274619e-01 -1.44507933e+00
1.20107800e-01 -6.45149767e-01 -5.14190316e-01 8.07539582e-01
-1.80187571e+00 -2.19389796e-01 -3.67981732e-01 2.78267682e-01
6.71629310e-01 1.08495623e-01 5.11559963e-01 -5.35148159e-02
-9.61328268e-01 3.25315803e-01 6.21502101e-01 -3.66639972e-01
4.21501487e-01 -1.04730749e+00 -6.53758228e-01 6.62450254e-01
-1.29253125e+00 2.92457849e-01 1.25129282e+00 -9.24346209e-01
-1.85669744e+00 -1.40945399e+00 9.94504094e-02 3.08542222e-01
8.76791060e-01 1.91965535e-01 -9.90263939e-01 1.78164631e-01
3.01455647e-01 -4.07629460e-02 3.48447442e-01 -7.78340816e-01
7.65289783e-01 -5.01925230e-01 -1.39899135e+00 3.72735083e-01
4.58003521e-01 6.59085810e-01 -1.82012826e-01 1.51268631e-01
9.80098367e-01 -2.55332142e-01 -1.33286607e+00 6.54491007e-01
6.22554183e-01 1.54308841e-01 3.95004898e-01 -6.65843427e-01
8.65271688e-02 -5.73817194e-01 2.42082644e-02 -1.73823810e+00
-5.06528080e-01 -3.74736458e-01 1.40411720e-01 1.14246356e+00
2.35059887e-01 -6.18692279e-01 4.55884993e-01 3.91792625e-01
-1.72861859e-01 -5.44476569e-01 -5.72819471e-01 -1.08614635e+00
6.36776626e-01 5.57333589e-01 6.30032003e-01 5.42047977e-01
1.29595041e-01 -1.39228478e-01 -4.04186875e-01 1.04148698e+00
9.31511760e-01 -2.63234168e-01 6.29031539e-01 -1.29176223e+00
-1.62456126e-03 -2.26534337e-01 -2.30968937e-01 -6.73389435e-01
-1.94934607e-01 -1.10519931e-01 8.58368218e-01 -1.41412497e+00
6.50096610e-02 -1.50615359e-02 -3.91366750e-01 4.49616373e-01
-2.94193983e-01 -2.37396166e-01 7.00174049e-02 2.27220833e-01
-2.99483892e-02 9.88262355e-01 1.51592362e+00 -1.82476819e-01
-8.51164639e-01 1.04107797e-01 -3.32959056e-01 3.07572961e-01
8.04420173e-01 -6.87586010e-01 -3.10475290e-01 -6.03774786e-01
7.83743858e-02 5.46244383e-01 -2.48303816e-01 -9.46960449e-01
2.33883932e-01 -1.13677728e+00 -6.01782799e-02 -4.36460257e-01
-3.96116674e-01 -1.30582869e+00 4.81574625e-01 1.57679570e+00
-3.28820676e-01 2.05781475e-01 3.98084581e-01 8.79617512e-01
8.56252760e-02 -4.20424730e-01 1.30185509e+00 -2.55727559e-01
-9.21936855e-02 6.33849680e-01 -9.04184163e-01 -2.29641661e-01
1.32670867e+00 -6.00570403e-02 -4.02132690e-01 7.35729709e-02
-4.56217080e-01 1.11266935e+00 4.39545572e-01 1.34057984e-01
3.07188988e-01 -8.47251058e-01 -1.13858390e+00 2.00591177e-01
-8.45262855e-02 6.60554916e-02 2.82615066e-01 7.63782918e-01
-8.99525881e-01 -2.74679810e-01 -3.41102481e-01 -7.00154006e-01
-1.01856446e+00 6.80682778e-01 8.59534621e-01 2.32525125e-01
-4.41939890e-01 4.52443242e-01 1.96361601e-01 5.91964051e-02
1.97465196e-01 -6.67129755e-01 -4.87869948e-01 1.41998306e-01
7.74456620e-01 6.50871873e-01 -3.53321582e-01 -2.96383709e-01
6.29494414e-02 7.62738168e-01 6.19509816e-01 4.45465237e-01
1.54966378e+00 -5.83293140e-01 -1.04225211e-01 5.09742677e-01
6.19085252e-01 -7.32194960e-01 -2.18925357e+00 1.71133056e-01
-1.93704247e-01 -3.72728586e-01 7.81049207e-02 -5.09240866e-01
-1.31727338e+00 4.43877757e-01 1.03055930e+00 4.70592380e-01
1.21775091e+00 -3.60803574e-01 1.51833162e-01 5.62175453e-01
3.11221093e-01 -7.27833092e-01 -9.74611640e-02 5.38372159e-01
6.76967084e-01 -9.02134538e-01 3.22592538e-03 1.31641909e-01
-3.03904623e-01 1.09901786e+00 1.06515992e+00 -2.14891613e-01
8.62067163e-01 8.46385956e-01 5.72479129e-01 -7.83679336e-02
-1.24918365e+00 -1.16774254e-01 -8.21711481e-01 6.61042869e-01
2.22708032e-01 1.27815321e-01 -7.09291101e-01 1.06273718e-01
1.80948347e-01 -1.92869917e-01 6.88614130e-01 8.33749831e-01
-1.21807325e+00 -6.85650885e-01 -6.10319078e-01 3.50522012e-01
-6.75822079e-01 2.44050592e-01 3.45094353e-01 7.43941247e-01
1.62835196e-01 1.00534427e+00 2.03029677e-01 2.38818690e-01
8.63959849e-01 -6.67790174e-02 7.68049285e-02 -5.75495422e-01
-8.01966727e-01 4.22923893e-01 -8.00881267e-01 1.09110080e-01
-9.35969293e-01 -7.81089485e-01 -1.29820979e+00 -4.42992210e-01
-9.90487099e-01 6.55960679e-01 6.41631544e-01 5.95380068e-01
-3.12992603e-01 3.93589973e-01 1.18531466e+00 -9.01141822e-01
-1.11550784e+00 -1.44291580e+00 -1.18545842e+00 7.25042820e-02
4.00837690e-01 -8.74082029e-01 -1.09377325e+00 1.28570110e-01] | [4.6811604499816895, 2.1591053009033203] |
1bd21129-4747-4865-a95c-540543025791 | from-scratch-to-sketch-deep-decoupled | 2208.04833 | null | https://arxiv.org/abs/2208.04833v1 | https://arxiv.org/pdf/2208.04833v1.pdf | From Scratch to Sketch: Deep Decoupled Hierarchical Reinforcement Learning for Robotic Sketching Agent | We present an automated learning framework for a robotic sketching agent that is capable of learning stroke-based rendering and motor control simultaneously. We formulate the robotic sketching problem as a deep decoupled hierarchical reinforcement learning; two policies for stroke-based rendering and motor control are learned independently to achieve sub-tasks for drawing, and form a hierarchy when cooperating for real-world drawing. Without hand-crafted features, drawing sequences or trajectories, and inverse kinematics, the proposed method trains the robotic sketching agent from scratch. We performed experiments with a 6-DoF robot arm with 2F gripper to sketch doodles. Our experimental results show that the two policies successfully learned the sub-tasks and collaborated to sketch the target images. Also, the robustness and flexibility were examined by varying drawing tools and surfaces. | ['Byoung-Tak Zhang', 'Minsu Lee', 'Minji Kim', 'Ganghun Lee'] | 2022-08-09 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [-1.89393312e-01 1.09384269e-01 -2.12739289e-01 1.16257787e-01
-1.51958734e-01 -7.97302961e-01 8.25404704e-01 -9.14293110e-01
-1.90076753e-01 6.17930353e-01 -3.70505929e-01 -1.75046816e-01
-4.73805845e-01 -6.87645197e-01 -8.06212246e-01 -6.55004025e-01
-1.36651173e-01 7.40707159e-01 1.22326128e-01 -2.85113543e-01
6.32636309e-01 1.13551247e+00 -1.43922567e+00 2.11038604e-01
5.03623247e-01 4.65936750e-01 7.85512686e-01 1.13280249e+00
-1.37532830e-01 1.12939417e+00 -4.29815292e-01 4.27594557e-02
6.59184873e-01 -1.69833049e-01 -4.06119317e-01 4.35339630e-01
1.75948247e-01 -8.55643809e-01 -9.35573876e-01 5.21201372e-01
2.66915232e-01 1.17520303e-01 9.20421898e-01 -1.74866354e+00
-6.00323498e-01 5.25824487e-01 -3.51495743e-01 -9.85014737e-01
-4.79973257e-02 4.79759067e-01 6.96379364e-01 -6.41471744e-01
9.33360517e-01 1.48583639e+00 3.69684130e-01 1.03006661e+00
-7.52155542e-01 -8.62022281e-01 1.39996842e-01 -2.52002984e-01
-8.26244354e-01 1.70293167e-01 1.02166605e+00 -5.52529097e-01
6.11108422e-01 -1.56750754e-01 8.31236124e-01 1.23564792e+00
2.61739373e-01 1.01840270e+00 1.06136489e+00 -2.09607795e-01
3.71985286e-01 -2.81870097e-01 -4.51274723e-01 1.05551255e+00
3.70207708e-04 6.00006938e-01 -3.29185486e-01 -1.27434790e-01
2.20122623e+00 2.75925189e-01 1.94499850e-01 -1.00496101e+00
-1.30682266e+00 6.59575045e-01 2.66840398e-01 -1.58473067e-02
-4.89088565e-01 8.93660665e-01 2.17747450e-01 6.29182935e-01
-5.13545156e-01 6.85155213e-01 -3.62711072e-01 4.58045723e-03
-3.22714686e-01 7.77131140e-01 8.98663223e-01 1.58979154e+00
3.53600264e-01 1.93892792e-01 -1.13784730e-01 7.91781723e-01
1.32092983e-01 5.69408000e-01 -8.78785700e-02 -1.72640216e+00
5.57157278e-01 3.64221483e-01 7.18599617e-01 -6.83624446e-01
-2.60014772e-01 4.39323485e-01 -8.08986127e-01 1.13062608e+00
2.94451445e-01 -2.60539651e-01 -8.25838208e-01 1.44787574e+00
-1.07038714e-01 -1.48583487e-01 -1.70001630e-02 1.09031117e+00
3.00793350e-01 5.01497447e-01 -2.89355479e-02 2.77835965e-01
8.54096830e-01 -1.29281104e+00 -6.64703012e-01 1.65812954e-01
-1.99343413e-02 -5.43595672e-01 1.29506564e+00 7.61460662e-01
-1.19114578e+00 -8.22771728e-01 -1.26181936e+00 6.27213046e-02
2.52597518e-02 6.61100864e-01 7.93261826e-01 -1.39405802e-01
-7.19824374e-01 1.19439471e+00 -7.22834885e-01 -6.15065619e-02
3.49889129e-01 2.79768884e-01 -2.61990309e-01 9.85517800e-02
-6.62194371e-01 1.01288915e+00 3.93591225e-02 -1.22097455e-01
-1.34699690e+00 -2.95068353e-01 -7.27354661e-02 -1.51550218e-01
2.22143322e-01 -7.99935341e-01 1.45939326e+00 -5.77869833e-01
-2.60034037e+00 6.03240192e-01 5.99342883e-01 2.72191077e-01
1.07397926e+00 -4.70161706e-01 2.16339901e-01 2.28100538e-01
-1.40693948e-01 8.56754005e-01 1.26800287e+00 -1.87727499e+00
-5.21665454e-01 -9.68633369e-02 3.06924403e-01 3.68028641e-01
1.06342189e-01 -5.37492573e-01 -3.10478330e-01 -8.20039511e-01
-5.93539961e-02 -1.15684164e+00 -2.84714311e-01 6.56914473e-01
-2.30684131e-01 -4.05733407e-01 1.07659602e+00 -3.18876833e-01
4.24954116e-01 -1.79711235e+00 7.74838328e-01 1.83091804e-01
8.95710438e-02 1.54677010e-03 -5.92696488e-01 1.00864160e+00
5.43446124e-01 -1.58959687e-01 7.14653777e-03 1.00336805e-01
3.07729334e-01 6.94201827e-01 -6.92680240e-01 2.40257736e-02
8.32782611e-02 1.05811846e+00 -1.13455093e+00 -2.78152615e-01
1.89680502e-01 1.19824149e-01 -2.79938489e-01 8.13028991e-01
-5.02886474e-01 4.52566803e-01 -9.10681069e-01 5.17479062e-01
4.18371409e-01 3.02922159e-01 5.73407233e-01 2.29537755e-01
-4.50520851e-02 -4.52996910e-01 -1.13905811e+00 1.85020769e+00
-6.84745371e-01 3.60159814e-01 5.98981440e-01 -7.80914783e-01
1.47063935e+00 1.70583621e-01 2.07808703e-01 -5.67296922e-01
-1.05427563e-01 2.25012392e-01 -3.52719963e-01 -8.42125237e-01
2.12282866e-01 -3.86255048e-02 -1.88834071e-01 5.37143052e-01
1.91828199e-02 -1.32095599e+00 -2.95079082e-01 -6.21978976e-02
1.16344476e+00 9.03088033e-01 -1.82293177e-01 -1.86152607e-01
1.47203639e-01 3.11972070e-02 1.51036367e-01 1.01183522e+00
3.63360137e-01 4.00346279e-01 6.09209418e-01 -6.07079029e-01
-1.68264663e+00 -1.48044097e+00 6.57178581e-01 1.09827924e+00
4.17224348e-01 5.36807962e-02 -3.97218734e-01 -6.10141158e-01
5.12087941e-01 5.41023791e-01 -4.15217072e-01 1.64671674e-01
-1.01326215e+00 5.29391825e-01 4.20380116e-01 7.15627015e-01
3.28794777e-01 -1.68564010e+00 -1.14601815e+00 3.70485157e-01
6.13206267e-01 -7.76140809e-01 -3.90586793e-01 9.91501263e-04
-8.22253764e-01 -1.43456304e+00 -1.05304658e+00 -1.15879679e+00
6.44383788e-01 3.50061923e-01 6.87978387e-01 3.98311079e-01
-3.88287812e-01 8.27644289e-01 -8.39015692e-02 -3.60640824e-01
-6.22342587e-01 -1.34845674e-01 -1.71459123e-01 -6.21836901e-01
-7.86850452e-01 -8.29963803e-01 -6.16335809e-01 5.38820922e-01
-6.50353611e-01 5.80783427e-01 1.05141616e+00 1.12171876e+00
2.37979770e-01 -2.66347110e-01 5.29192090e-01 -2.62378305e-01
1.23085880e+00 2.05777764e-01 -7.85836697e-01 3.77371758e-01
-7.83289149e-02 2.32578665e-01 9.68119502e-01 -8.60120833e-01
-9.98766422e-01 4.74633932e-01 4.84648257e-01 -9.30316508e-01
-2.27847755e-01 -3.34352106e-01 -7.74357393e-02 -1.39901370e-01
2.79362112e-01 1.09257333e-01 4.17247325e-01 -3.05437535e-01
8.15344691e-01 5.60122132e-01 7.02610135e-01 -1.39455783e+00
8.23924422e-01 3.07318032e-01 2.33647645e-01 -6.75589025e-01
5.96049987e-02 2.96144515e-01 -9.32718754e-01 -4.26926672e-01
6.22590363e-01 -5.20052016e-01 -1.47256303e+00 7.40519762e-01
-1.48596251e+00 -1.13153815e+00 -2.21562237e-01 8.90360847e-02
-1.48443913e+00 2.96734989e-01 -1.00177479e+00 -9.86458778e-01
-2.52077639e-01 -1.21945953e+00 1.11348903e+00 1.02757104e-01
-3.50080654e-02 -2.18967184e-01 9.59863290e-02 -3.50479662e-01
4.15770471e-01 3.86716723e-01 1.19080102e+00 1.82788387e-01
-8.56568873e-01 -5.03781177e-02 -3.05776507e-01 -2.81230006e-02
9.88819450e-02 2.25264877e-01 -5.09530544e-01 -3.05223256e-01
-4.07773882e-01 -7.62902558e-01 4.28966552e-01 -1.10394992e-01
1.33267152e+00 -4.75278735e-01 -2.18311176e-01 1.97321534e-01
1.10227776e+00 5.97409725e-01 5.14052093e-01 1.80920929e-01
6.79868877e-01 8.23257148e-01 8.31315041e-01 4.93894458e-01
2.01728456e-02 6.96775377e-01 4.95059848e-01 3.09077770e-01
-1.82107329e-01 -8.54442179e-01 8.39876160e-02 5.33810377e-01
-5.53517044e-01 1.45468920e-01 -5.10659754e-01 1.31552354e-01
-2.25395036e+00 -7.33553529e-01 2.03371748e-01 1.79340363e+00
5.36004663e-01 -7.42661953e-02 2.39896834e-01 -1.54489249e-01
7.19193995e-01 -2.77234733e-01 -1.00877106e+00 -3.62892509e-01
4.38155562e-01 2.31695712e-01 2.63958335e-01 4.58137929e-01
-6.58504367e-01 1.27492547e+00 6.42312098e+00 6.33765221e-01
-1.08696139e+00 -4.96978164e-01 -1.11344263e-01 -2.32826499e-03
-2.39588693e-01 -1.23106919e-01 -1.83555316e-02 2.48699769e-01
-1.51761070e-01 3.58311504e-01 1.25342131e+00 9.66528714e-01
-3.85838933e-02 1.76576719e-01 -1.24110234e+00 1.14527667e+00
-5.91545641e-01 -1.41136515e+00 2.48300448e-01 -1.50919795e-01
6.29486799e-01 -4.29345757e-01 -5.58674671e-02 5.41109204e-01
9.45648849e-01 -1.05646026e+00 1.00316310e+00 8.53844464e-01
9.73554492e-01 -7.78536737e-01 -6.88731223e-02 7.07148612e-01
-1.04496968e+00 -4.51279968e-01 -4.51379448e-01 -3.57474357e-01
5.08542843e-02 -2.47373968e-01 -5.52298188e-01 2.59027481e-01
1.85576156e-01 4.99458462e-01 3.03141177e-01 6.68821454e-01
-7.40231693e-01 -2.20459908e-01 1.47466920e-02 -5.90438664e-01
4.47253197e-01 -5.45846581e-01 4.61748153e-01 8.50898504e-01
3.06836575e-01 3.43536884e-01 2.49493346e-01 1.10502505e+00
7.06396475e-02 -4.57756460e-01 -7.99104452e-01 -1.70387089e-01
6.88486099e-01 1.32355845e+00 -5.86910427e-01 -4.17427748e-01
2.23477617e-01 1.20508206e+00 6.17735684e-01 5.69593489e-01
-7.81716466e-01 -7.58606791e-01 6.54767096e-01 -2.76078641e-01
4.05957222e-01 -9.33093548e-01 -2.58696020e-01 -8.80175591e-01
-1.29626215e-01 -5.75614512e-01 -4.13709790e-01 -1.20251191e+00
-1.38029325e+00 1.94633156e-01 -1.61770910e-01 -1.08836424e+00
-2.50286520e-01 -9.47052538e-01 -9.80302274e-01 6.19622350e-01
-1.21190834e+00 -1.26250887e+00 -5.90712011e-01 5.38392425e-01
8.10609221e-01 -5.62978625e-01 8.29815328e-01 -3.61013472e-01
-1.04003236e-01 2.42482275e-01 1.38979955e-02 9.76446196e-02
5.62680066e-01 -1.29341233e+00 5.01573682e-01 -1.45877466e-01
-1.84095860e-01 3.57780814e-01 2.91083187e-01 -7.07909584e-01
-2.15056419e+00 -7.23688126e-01 -1.55151069e-01 -1.70824662e-01
6.85553789e-01 -5.67640662e-01 -5.02072811e-01 4.02469248e-01
2.03460276e-01 -1.52329594e-01 -5.09458423e-01 -4.12423193e-01
-2.65141338e-01 1.91442341e-01 -1.13771367e+00 8.73333454e-01
1.60672569e+00 -4.25283998e-01 -6.63089573e-01 2.12418020e-01
5.01436830e-01 -5.96478760e-01 -7.88873672e-01 6.28719628e-02
1.25861812e+00 -4.76638436e-01 1.02482009e+00 -9.80845034e-01
6.43748701e-01 -2.60180056e-01 8.53318423e-02 -1.38158166e+00
-5.70731580e-01 -9.15514588e-01 -1.89559907e-01 7.13530481e-01
-2.52397627e-01 -3.53398234e-01 8.79255176e-01 2.85746038e-01
-9.29001123e-02 -8.12047601e-01 -4.85588670e-01 -9.85359788e-01
3.28427225e-01 1.49906248e-01 5.12370050e-01 7.87464857e-01
3.23106647e-02 1.02086075e-01 -6.39923334e-01 6.39498383e-02
5.46139836e-01 4.03231770e-01 1.31866062e+00 -1.18470144e+00
-3.97503018e-01 -7.87220001e-01 2.46111393e-01 -1.15803552e+00
3.88830990e-01 -7.57024825e-01 1.93025902e-01 -1.61526656e+00
-1.60630286e-01 -8.22294295e-01 2.00745180e-01 6.10545933e-01
3.43546361e-01 -5.27915835e-01 7.84414411e-01 4.71616149e-01
-2.34974623e-01 5.51349044e-01 1.95568419e+00 -1.94784343e-01
-3.03415596e-01 1.95464995e-02 -2.18336787e-02 5.73446035e-01
8.88380587e-01 -1.81297913e-01 -6.31939113e-01 -5.34567356e-01
-1.22618355e-01 9.02050316e-01 5.95891356e-01 -5.82468390e-01
2.94413447e-01 -5.75085938e-01 4.42528635e-01 -5.46415925e-01
3.76394987e-01 -1.01460838e+00 5.81745021e-02 9.52385485e-01
-6.17513418e-01 7.66936541e-02 -3.42586152e-02 4.92393017e-01
3.66299331e-01 -1.13849849e-01 5.74532270e-01 -4.47174639e-01
-5.89094222e-01 2.50973225e-01 -5.16323388e-01 -3.30046415e-01
1.13350427e+00 7.53432214e-02 -5.20712957e-02 -4.00972068e-01
-7.04636335e-01 2.87995815e-01 5.35620213e-01 7.24490464e-01
9.21481431e-01 -1.55540204e+00 -5.37131488e-01 1.00077726e-01
-3.58762085e-01 -5.46948984e-02 1.97027232e-02 -2.07825929e-01
-7.38561571e-01 1.35098308e-01 -8.97009075e-01 -3.70486915e-01
-7.81763256e-01 7.24929094e-01 2.08170637e-01 -1.66580677e-01
-9.71940458e-01 3.45327467e-01 8.30848962e-02 -1.03699851e+00
5.10495245e-01 -3.56848478e-01 2.85795271e-01 -5.41791260e-01
1.28502697e-01 6.99511349e-01 -6.28623903e-01 6.24078572e-01
1.33474022e-01 8.63680184e-01 2.40507022e-01 -3.95418257e-01
1.54857326e+00 3.24737906e-01 1.39405549e-01 8.20138976e-02
6.99162066e-01 -5.02677619e-01 -2.03849864e+00 3.31763595e-01
-6.30936697e-02 -4.50609148e-01 -5.28984785e-01 -8.66516113e-01
-8.59302878e-01 8.84502232e-01 5.34682609e-02 -1.25650495e-01
4.03135389e-01 -2.67010391e-01 4.83272195e-01 1.16556633e+00
8.92074287e-01 -1.28290391e+00 7.93016076e-01 5.87076664e-01
1.74461889e+00 -6.41235113e-01 -1.48543507e-01 -1.47441909e-01
-9.44719136e-01 1.79697025e+00 9.03196096e-01 -1.01588011e+00
2.06965804e-01 7.06309557e-01 -6.94101378e-02 -1.27302498e-01
-6.76780820e-01 3.99861991e-01 3.34251294e-04 7.90526211e-01
-1.91819176e-01 3.01938802e-01 6.42136410e-02 5.59501529e-01
-1.02515459e-01 3.86580676e-01 3.74194562e-01 1.18304777e+00
-5.32850623e-01 -1.10554051e+00 -1.18821979e-01 1.02482803e-01
6.60228550e-01 7.93529630e-01 -5.42125344e-01 1.22865558e+00
-2.30723321e-01 2.12444350e-01 -2.49059498e-02 -3.80768478e-01
7.26975262e-01 -2.10960940e-01 1.18566334e+00 -1.39118657e-01
-2.64645338e-01 -1.38676956e-01 -2.38072835e-02 -7.47180998e-01
2.26018369e-01 -3.12235206e-01 -1.44509542e+00 -6.64389729e-02
2.00427681e-01 -1.48296401e-01 6.91659272e-01 4.87493664e-01
3.99386466e-01 4.99234527e-01 8.38167608e-01 -1.74481058e+00
-1.00777900e+00 -6.71577573e-01 -6.98131859e-01 2.74306983e-01
2.18530789e-01 -9.06784415e-01 6.61180317e-02 -2.24839360e-01] | [4.727212429046631, 0.588460385799408] |
1864c07e-71df-4eb6-86df-7b64f6885991 | self-supervised-learning-via-inter-modal | 2307.03008 | null | https://arxiv.org/abs/2307.03008v1 | https://arxiv.org/pdf/2307.03008v1.pdf | Self-supervised learning via inter-modal reconstruction and feature projection networks for label-efficient 3D-to-2D segmentation | Deep learning has become a valuable tool for the automation of certain medical image segmentation tasks, significantly relieving the workload of medical specialists. Some of these tasks require segmentation to be performed on a subset of the input dimensions, the most common case being 3D-to-2D. However, the performance of existing methods is strongly conditioned by the amount of labeled data available, as there is currently no data efficient method, e.g. transfer learning, that has been validated on these tasks. In this work, we propose a novel convolutional neural network (CNN) and self-supervised learning (SSL) method for label-efficient 3D-to-2D segmentation. The CNN is composed of a 3D encoder and a 2D decoder connected by novel 3D-to-2D blocks. The SSL method consists of reconstructing image pairs of modalities with different dimensionality. The approach has been validated in two tasks with clinical relevance: the en-face segmentation of geographic atrophy and reticular pseudodrusen in optical coherence tomography. Results on different datasets demonstrate that the proposed CNN significantly improves the state of the art in scenarios with limited labeled data by up to 8% in Dice score. Moreover, the proposed SSL method allows further improvement of this performance by up to 23%, and we show that the SSL is beneficial regardless of the network architecture. | ['Hrvoje Bogunović', 'Ursula Schmidt-Erfurth', 'Julia Mai', 'Dmitrii Lachinov', 'Guilherme Aresta', 'José Morano'] | 2023-07-06 | null | null | null | null | ['self-supervised-learning', 'medical-image-segmentation', 'transfer-learning'] | ['computer-vision', 'medical', 'miscellaneous'] | [ 2.45058745e-01 3.12959701e-01 -2.23361000e-01 -4.85507786e-01
-6.63201392e-01 -3.58754396e-01 2.36180544e-01 -1.63782369e-02
-6.85703337e-01 8.71482074e-01 -5.00509106e-02 -2.77338624e-01
-2.24371538e-01 -4.65137750e-01 -4.51790333e-01 -6.97755873e-01
5.35421148e-02 8.31740975e-01 2.70474762e-01 1.68456957e-01
-9.04549584e-02 6.97951674e-01 -1.38300145e+00 2.37186477e-01
9.72561061e-01 1.08417940e+00 3.86302173e-01 5.23855388e-01
-1.21797003e-01 5.65169156e-01 -2.60917544e-01 -4.74342667e-02
5.23139060e-01 -6.17633224e-01 -1.04860246e+00 2.80574530e-01
3.98400605e-01 -5.36962211e-01 6.18841462e-02 7.60038793e-01
7.03406572e-01 -2.49467596e-01 6.64462805e-01 -7.11026073e-01
1.68331002e-03 3.38465512e-01 -5.07965744e-01 5.69237582e-02
-8.32336172e-02 9.72399488e-02 4.90081578e-01 -5.32544792e-01
8.28016043e-01 8.28573108e-01 5.96512198e-01 4.64817137e-01
-1.23105109e+00 -3.36290151e-01 -4.84488964e-01 -2.18805037e-02
-1.21653605e+00 4.41685272e-03 5.46216369e-01 -8.70812356e-01
7.83591032e-01 1.11309096e-01 8.22250426e-01 5.54813623e-01
9.10609290e-02 6.55137420e-01 1.49228752e+00 -4.04709578e-01
3.06215525e-01 6.74358308e-02 -7.16537312e-02 6.73617721e-01
2.91758895e-01 -8.47847983e-02 5.06059341e-02 1.93153247e-01
8.79823744e-01 -1.72378033e-01 -2.64687538e-01 -5.07676780e-01
-1.08476257e+00 8.10904503e-01 5.59980333e-01 5.84385514e-01
-3.97863388e-01 -7.21152574e-02 5.07421076e-01 5.61730042e-02
7.40178764e-01 5.00932038e-01 -4.45509404e-01 2.13642269e-01
-1.33954024e+00 1.51595948e-02 5.42296946e-01 4.45143193e-01
4.81132418e-01 -2.93781042e-01 -1.23250447e-01 8.05975080e-01
3.03236932e-01 4.43076730e-01 5.42740941e-01 -9.56624508e-01
1.53383404e-01 8.31744194e-01 4.20253202e-02 -6.11602962e-01
-1.05944300e+00 -7.55524635e-01 -1.03854430e+00 4.55409139e-01
6.10000253e-01 -3.08260590e-01 -1.00439811e+00 1.36113429e+00
4.85926867e-01 -1.38415426e-01 -1.26566261e-01 1.20387685e+00
6.74305618e-01 2.80851960e-01 -1.74479052e-01 -2.08686009e-01
1.21843493e+00 -9.05727386e-01 -6.36942685e-01 -8.14045668e-02
9.44567621e-01 -7.61202812e-01 7.66325355e-01 4.29686964e-01
-1.10833967e+00 -4.16796446e-01 -9.25371528e-01 -1.65662408e-01
-1.72905356e-01 5.32712758e-01 4.81312543e-01 7.24229634e-01
-1.14781916e+00 6.97715461e-01 -9.34043884e-01 -3.98971617e-01
8.79922032e-01 7.76279211e-01 -5.55131078e-01 -2.81912863e-01
-8.79979849e-01 9.02540326e-01 5.88908732e-01 1.34617984e-01
-5.72559536e-01 -6.86048627e-01 -5.10234058e-01 -5.93670607e-02
2.07837731e-01 -8.87030303e-01 9.47506487e-01 -9.60995436e-01
-1.56377137e+00 1.33514214e+00 2.40965247e-01 -7.57729888e-01
1.02861726e+00 -1.05264738e-01 -2.98630893e-02 4.08517331e-01
1.28132373e-01 9.90345180e-01 7.58012235e-01 -1.11474931e+00
-4.50449735e-01 -6.70981765e-01 -1.11439057e-01 1.40805706e-01
-9.70518738e-02 -2.75679797e-01 -3.94981474e-01 -3.95224452e-01
1.21901959e-01 -1.04865336e+00 -2.32167438e-01 3.53841126e-01
-5.40961802e-01 1.78976536e-01 5.95092416e-01 -9.07268345e-01
7.46102273e-01 -1.89152884e+00 2.29790062e-01 1.04639724e-01
4.28797215e-01 5.71397007e-01 4.43957932e-03 3.19907553e-02
-1.48061901e-01 -6.32041991e-02 -6.86343729e-01 -5.74601948e-01
-3.67442489e-01 1.73461094e-01 3.43851894e-01 6.78928554e-01
2.89045535e-02 8.37160408e-01 -7.53345668e-01 -6.71328068e-01
6.01382136e-01 5.96030056e-01 -4.74316359e-01 3.64238262e-01
-1.43016830e-01 1.16296601e+00 -1.66659519e-01 3.19344163e-01
8.85448992e-01 -6.07906640e-01 2.85077095e-01 -2.66164273e-01
-1.52927935e-01 -2.47092601e-02 -8.57725322e-01 2.00124669e+00
-5.75061023e-01 5.16764641e-01 -8.21952522e-02 -1.14949071e+00
6.75320089e-01 3.09303224e-01 9.40929651e-01 -8.08172345e-01
3.95410538e-01 6.73906565e-01 3.04618746e-01 -7.47303724e-01
-1.34944543e-01 -2.77321756e-01 4.75820929e-01 3.81995052e-01
1.27166286e-01 -1.27173990e-01 3.42474997e-01 -2.17672572e-01
9.32454348e-01 6.86753243e-02 2.17585370e-01 -2.01593861e-01
7.02162325e-01 2.00279206e-02 2.98526436e-01 5.27596712e-01
-1.09115966e-01 8.95986378e-01 6.49392366e-01 -4.30919439e-01
-1.29081023e+00 -7.17410445e-01 -3.97317290e-01 2.06869498e-01
1.33341579e-02 1.10434137e-01 -1.05792439e+00 -7.43524194e-01
-3.20222259e-01 3.99728030e-01 -5.17158687e-01 2.32632086e-01
-4.89908814e-01 -8.73420954e-01 3.56931239e-01 2.82629997e-01
7.09711611e-01 -9.14724827e-01 -8.89493108e-01 2.21477464e-01
-1.05393074e-01 -1.47913420e+00 -2.17302181e-02 7.16370791e-02
-1.29582345e+00 -1.30239809e+00 -1.21404147e+00 -6.67491615e-01
7.10836470e-01 -9.29486305e-02 9.64500248e-01 -1.01614118e-01
-4.54193085e-01 1.11938938e-02 -3.23015273e-01 -1.73587091e-02
-4.17020947e-01 4.06798869e-01 -3.79939675e-01 1.53083742e-01
-4.54875641e-02 -5.26404321e-01 -1.02328527e+00 3.54129374e-01
-9.33803082e-01 3.92787635e-01 8.44473422e-01 8.03890824e-01
7.64043272e-01 -6.43894002e-02 3.65089864e-01 -1.10256541e+00
2.39918351e-01 -2.01376870e-01 -6.65543258e-01 2.69081611e-02
-6.60097778e-01 7.10824355e-02 6.05759025e-01 3.15080546e-02
-7.87628233e-01 5.06620586e-01 -3.43598247e-01 -3.54616106e-01
-3.74064803e-01 4.64938164e-01 2.41707042e-01 -2.10751221e-01
5.77600956e-01 2.17592437e-02 3.85521770e-01 -5.75745702e-01
2.39883646e-01 6.90590322e-01 2.45959952e-01 -7.35874381e-03
2.58790255e-01 8.60989213e-01 4.20982718e-01 -6.79458141e-01
-8.69471014e-01 -5.43905556e-01 -9.64347303e-01 -2.72321850e-01
1.32186925e+00 -6.94329023e-01 -3.21885139e-01 6.74741566e-01
-1.04699910e+00 -4.77724642e-01 -3.70941699e-01 7.25856960e-01
-7.11493373e-01 4.49405998e-01 -4.02342528e-01 -4.23664421e-01
-4.95893359e-01 -1.66360569e+00 1.03707302e+00 2.99402396e-03
-1.31093040e-02 -9.43631411e-01 9.22958851e-02 6.17522955e-01
4.48774397e-01 3.77289593e-01 1.03928339e+00 -6.77237749e-01
-5.52237749e-01 -2.00994357e-01 -3.92103672e-01 7.27805555e-01
1.05200246e-01 -4.77056831e-01 -8.72158945e-01 -3.28995734e-01
3.06188148e-02 -3.33699077e-01 7.69501209e-01 9.22181249e-01
1.17966795e+00 1.53677166e-01 -2.16299161e-01 5.84258378e-01
1.53728330e+00 6.10850900e-02 7.45831728e-01 1.72559723e-01
6.23097241e-01 6.56230032e-01 2.95822352e-01 2.88827777e-01
1.14571542e-01 7.97905445e-01 6.07625306e-01 -6.43119812e-01
-6.52509987e-01 2.66807377e-01 -2.89773315e-01 6.37217522e-01
-9.66309607e-02 -7.80409500e-02 -1.00306582e+00 5.16637206e-01
-1.67503953e+00 -3.51799577e-01 -2.84889907e-01 2.22473836e+00
6.45333171e-01 -3.01675429e-03 8.62386301e-02 1.78825736e-01
6.12264335e-01 -1.25176832e-01 -6.11096501e-01 -8.23679194e-02
8.24091658e-02 5.42428792e-01 5.66967070e-01 4.73379850e-01
-1.13576531e+00 6.16033018e-01 5.39212847e+00 7.04891980e-01
-1.36210287e+00 3.78844768e-01 7.93571770e-01 2.50402577e-02
1.93884879e-01 -3.21586072e-01 -4.97525424e-01 5.41220307e-01
7.37605095e-01 6.70672238e-01 2.16412559e-01 5.90279520e-01
3.95581394e-01 -4.82391864e-01 -9.85788405e-01 1.13435423e+00
6.72633797e-02 -1.36439025e+00 -9.46231708e-02 1.84455484e-01
7.95957029e-01 2.14291051e-01 -2.92284936e-02 -1.54799983e-01
-5.11606216e-01 -1.12089527e+00 8.58722702e-02 5.10847926e-01
1.02143419e+00 -5.66292107e-01 1.05406845e+00 2.20211476e-01
-5.42396605e-01 1.20930858e-01 -1.01807177e-01 4.29658949e-01
3.05447966e-01 9.37210441e-01 -1.04635596e+00 5.29221654e-01
4.83213574e-01 7.06595957e-01 -4.42742348e-01 1.37792552e+00
-3.68860364e-02 4.16165680e-01 -3.02227795e-01 3.83609772e-01
3.61291021e-01 -3.30432862e-01 3.80181730e-01 9.90205526e-01
3.67287576e-01 -6.26874119e-02 -1.83878317e-01 8.58345628e-01
-1.17033854e-01 3.21020693e-01 -3.90834153e-01 4.37150560e-02
-1.79798499e-01 1.25655878e+00 -1.12603545e+00 -2.21758798e-01
-3.17179650e-01 1.03164577e+00 5.16300052e-02 2.71323882e-02
-6.40878081e-01 -1.57007575e-01 1.17408574e-01 5.66388428e-01
4.12741661e-01 -6.29891008e-02 -3.91758204e-01 -8.71150076e-01
-1.80770755e-02 -6.51545823e-01 2.11526632e-01 -8.27340484e-01
-1.07573366e+00 8.45727682e-01 -2.68217057e-01 -1.35630977e+00
-2.92431235e-01 -7.45141208e-01 1.18461803e-01 8.68406415e-01
-1.69544673e+00 -1.06404936e+00 -3.90346766e-01 4.86265212e-01
2.89803594e-01 -2.13611409e-01 7.28751957e-01 6.10208571e-01
-2.71430939e-01 1.76662460e-01 3.10290605e-01 7.39073455e-02
7.01627791e-01 -1.24489629e+00 -1.10687405e-01 5.34092009e-01
-2.69118756e-01 6.62247837e-02 2.87783474e-01 -4.90983099e-01
-8.35787654e-01 -1.07097483e+00 7.10578501e-01 4.05746605e-03
1.92764655e-01 1.41569972e-02 -7.06294954e-01 4.08125073e-01
6.56998977e-02 3.63234580e-01 6.17802680e-01 -3.70392948e-01
1.03695787e-01 -1.15000904e-01 -1.43864584e+00 2.34346196e-01
6.12304330e-01 -2.62439936e-01 -2.33680993e-01 6.08532488e-01
3.99475217e-01 -6.84267700e-01 -1.00395989e+00 5.44997454e-01
5.05852997e-01 -1.24054360e+00 9.02409494e-01 -3.24346393e-01
7.14494884e-01 -2.77066767e-01 -3.12897284e-03 -1.07867146e+00
1.54605746e-01 -1.10456832e-01 3.39901149e-02 5.20884097e-01
2.79591978e-01 -6.31479979e-01 9.50398684e-01 2.75338441e-01
-2.87185520e-01 -9.37735677e-01 -1.16628468e+00 -5.63593686e-01
1.02128498e-01 -3.01713109e-01 1.82822078e-01 8.42310071e-01
-6.19569421e-01 2.96359301e-01 -2.82336831e-01 -2.07599998e-02
5.92256606e-01 1.66148797e-01 4.76564586e-01 -1.33278501e+00
-3.01461130e-01 -3.84657234e-01 -5.34622729e-01 -1.02223217e+00
1.60439238e-02 -1.13642800e+00 -3.18275869e-01 -1.78353775e+00
1.43574551e-01 -4.93813634e-01 -9.37698185e-02 3.28622192e-01
2.83218265e-01 5.36105573e-01 -5.50437486e-03 3.66174012e-01
-3.50321412e-01 2.54840702e-01 1.69257224e+00 -1.90002739e-01
-2.44333535e-01 6.80688843e-02 -1.79443091e-01 6.84713840e-01
8.74163032e-01 -5.93115330e-01 -4.22728121e-01 -3.86064172e-01
1.62519217e-01 2.15884194e-01 4.37500477e-01 -1.19300473e+00
-2.65401546e-02 5.02731264e-01 4.61023510e-01 -7.30761170e-01
2.98695147e-01 -1.02542877e+00 -1.04469936e-02 8.26123536e-01
-2.72684693e-01 -4.10538405e-01 1.09060653e-01 2.31724456e-01
-2.95646608e-01 -3.51937890e-01 1.09607601e+00 -2.29311645e-01
-3.70588928e-01 3.49690676e-01 -1.26183495e-01 8.80121738e-02
9.77426887e-01 -1.78138271e-01 9.93469730e-02 -1.31857648e-01
-1.08157277e+00 -7.31484918e-03 1.85376212e-01 -6.45850524e-02
3.94632071e-01 -9.20999467e-01 -7.51807272e-01 1.39870718e-01
-1.10433571e-01 2.75574535e-01 5.04951358e-01 1.52085817e+00
-1.03022504e+00 7.32003689e-01 -4.93357867e-01 -1.06351364e+00
-1.09697247e+00 3.26884419e-01 8.32690299e-01 -5.36738217e-01
-7.30464995e-01 5.74563801e-01 2.61459798e-02 -4.87991005e-01
5.34221791e-02 -4.36696798e-01 -3.12313139e-01 8.43482465e-02
2.66253471e-01 3.31709176e-01 5.10477126e-01 -5.98540843e-01
-2.39364356e-01 7.76261210e-01 -1.87338918e-01 6.20775670e-02
1.43441033e+00 -1.16200924e-01 -3.13023210e-01 2.83511698e-01
1.33489907e+00 -5.18601358e-01 -1.17551947e+00 -1.02141105e-01
-2.48428509e-01 -3.07725102e-01 3.41018736e-01 -1.05843329e+00
-1.40192008e+00 1.06594956e+00 1.29587340e+00 1.78733133e-02
1.28380775e+00 -7.65950829e-02 9.62948799e-01 1.05041028e-04
3.19472909e-01 -8.31454933e-01 -1.46764755e-01 1.27327189e-01
6.64754033e-01 -1.53446114e+00 2.44608847e-04 -5.22341430e-01
-4.63069826e-01 1.15521657e+00 3.69193196e-01 -8.32314640e-02
6.79724932e-01 2.20154468e-02 6.53539151e-02 -3.65823269e-01
9.89468070e-04 -3.94524485e-01 4.07052308e-01 4.60665166e-01
4.12300915e-01 7.16208890e-02 -7.37712979e-01 1.80131912e-01
1.47848055e-01 4.24222291e-01 4.01372641e-01 6.13985479e-01
-3.02863959e-02 -1.18008232e+00 -1.40458211e-01 5.92417121e-01
-4.61857170e-01 -1.32903233e-02 -9.94937196e-02 9.42294121e-01
5.01224101e-01 6.48898065e-01 8.88382122e-02 -2.50357781e-02
1.23594210e-01 -9.98416096e-02 6.79539025e-01 -6.24429762e-01
-5.82997918e-01 2.12041855e-01 -1.45770654e-01 -5.93360364e-01
-8.51370811e-01 -4.24325675e-01 -1.21799934e+00 8.15775916e-02
-6.12883307e-02 -2.44732693e-01 8.37507665e-01 1.09385264e+00
4.18124795e-01 5.29490471e-01 4.93376255e-01 -7.61936367e-01
-2.15245917e-01 -9.00882125e-01 -6.61038935e-01 3.72513533e-01
3.56913358e-01 -6.06759548e-01 -9.92309526e-02 -2.67992122e-03] | [14.410202980041504, -2.4842326641082764] |
43a7769a-1fd2-4afd-b2ed-84fe3b389ff3 | new-approach-for-solar-tracking-systems-based | 1809.07048 | null | http://arxiv.org/abs/1809.07048v1 | http://arxiv.org/pdf/1809.07048v1.pdf | New approach for solar tracking systems based on computer vision, low cost hardware and deep learning | In this work, a new approach for Sun tracking systems is presented. Due to
the current system limitations regarding costs and operational problems, a new
approach based on low cost, computer vision open hardware and deep learning has
been developed. The preliminary tests carried out successfully in Plataforma
solar de Almeria (PSA), reveal the great potential and show the new approach as
a good alternative to traditional systems. The proposed approach can provide
key variables for the Sun tracking system control like cloud movements
prediction, block and shadow detection, atmospheric attenuation or measures of
concentrated solar radiation, which can improve the control strategies of the
system and therefore the system performance. | ['Ginés García', 'Jesús Fernández-Reche', 'Jose A. Carballo', 'Manuel Berenguel', 'Javier Bonilla'] | 2018-09-19 | null | null | null | null | ['shadow-detection'] | ['computer-vision'] | [-2.14257032e-01 -4.29265201e-01 9.86921489e-02 2.12222159e-01
5.14109969e-01 -6.30191505e-01 7.18252659e-01 -1.02566108e-01
7.91822821e-02 7.67696381e-01 -3.83104950e-01 -4.36259210e-01
-1.10173084e-01 -8.55559051e-01 -1.89431578e-01 -1.01654363e+00
1.22087978e-01 1.25976220e-01 3.03321719e-01 -3.16002220e-01
3.14526379e-01 9.60486591e-01 -1.88689649e+00 -5.91281474e-01
9.61005390e-01 9.01838005e-01 3.90558064e-01 1.05278707e+00
9.82205644e-02 6.55283153e-01 -7.14593887e-01 8.15267265e-01
3.48952830e-01 -1.69237833e-02 -6.23188466e-02 -1.32070035e-01
6.31330132e-01 -1.14017062e-01 1.37721524e-02 8.27926993e-01
6.52930558e-01 2.98878960e-02 4.87796009e-01 -9.10246730e-01
-2.53578722e-01 -1.63731381e-01 -2.97371268e-01 5.62887013e-01
1.86435550e-01 3.44376653e-01 3.91029656e-01 -5.67031145e-01
2.45764166e-01 6.65794551e-01 8.56799781e-01 -6.83007613e-02
-6.85190022e-01 -4.66106862e-01 -2.99963087e-01 4.08588558e-01
-1.23724127e+00 3.47239971e-02 2.30370641e-01 -6.42815292e-01
9.48274493e-01 6.34511232e-01 1.33749259e+00 4.02203441e-01
4.66799498e-01 -1.78899914e-02 1.65379822e+00 -8.94018173e-01
4.57732201e-01 4.30401236e-01 2.04843462e-01 4.69141603e-01
6.47728384e-01 7.15948761e-01 -2.91371979e-02 1.67649090e-01
3.58642280e-01 -2.33936712e-01 -3.03704113e-01 -1.90463066e-01
-5.55140376e-01 8.39944243e-01 8.31328869e-01 7.12809205e-01
-2.35433444e-01 -2.65459018e-03 -7.10782185e-02 6.29729629e-02
9.53198373e-02 1.84030578e-01 -2.89322108e-01 1.70616880e-01
-1.05451572e+00 2.07091078e-01 8.36856961e-01 4.85726744e-01
4.47565377e-01 6.88932896e-01 -1.87116742e-01 3.44255678e-02
5.84237933e-01 1.25570095e+00 3.60094756e-01 -4.72850084e-01
-3.80726427e-01 5.57750583e-01 3.79935056e-01 -6.56995416e-01
-7.07743585e-01 -4.16761220e-01 -5.01767099e-01 8.11347783e-01
-1.11422867e-01 -4.31824595e-01 -1.39845860e+00 7.37509191e-01
5.63356280e-01 5.88946454e-02 2.15689629e-01 9.61561739e-01
6.35086179e-01 1.03129005e+00 -3.23873609e-01 -3.17830056e-01
1.06634617e+00 -9.85880017e-01 -1.02108729e+00 -6.70602471e-02
7.35666901e-02 -9.79708970e-01 5.17998993e-01 3.03022414e-01
-4.20475364e-01 -6.40151620e-01 -1.31053424e+00 3.64676565e-01
-1.07970405e+00 2.69700319e-01 8.18791211e-01 1.21908844e+00
-1.24037218e+00 3.12654257e-01 -6.41946793e-01 -9.45741236e-01
5.03950082e-02 2.95112282e-01 6.89816833e-01 5.16227067e-01
-9.02708709e-01 1.44935429e+00 3.83206487e-01 3.64106387e-01
-1.01351106e+00 -4.28381622e-01 -3.95733744e-01 3.47803026e-01
6.66679963e-02 -4.85296518e-01 1.36020315e+00 -1.11332560e+00
-1.80037773e+00 2.58755505e-01 1.35057405e-01 -8.34349632e-01
3.36416572e-01 -1.08584382e-01 -6.61594391e-01 -5.58602735e-02
-4.78703439e-01 4.74587232e-02 7.74812281e-01 -9.99911606e-01
-9.66883361e-01 -2.68698961e-01 8.97125378e-02 2.45283991e-01
-2.93066114e-01 6.38577864e-02 2.92210877e-01 -9.30774510e-02
-4.17474508e-01 -1.27574313e+00 -2.26302147e-01 9.14073214e-02
5.78293726e-02 5.31481253e-03 1.39172101e+00 -6.71573341e-01
1.01072550e+00 -1.65987229e+00 -1.69948831e-01 5.57772875e-01
-3.93574476e-01 9.61789548e-01 7.80666530e-01 3.35862249e-01
5.87271973e-02 -4.17871594e-01 -8.00230280e-02 3.92745227e-01
-4.17618632e-01 4.29546863e-01 2.10578546e-01 3.89531881e-01
-4.68036264e-01 2.46656284e-01 -5.26502132e-01 3.66403051e-02
9.18813646e-01 6.19628847e-01 5.14820337e-01 -3.64917554e-02
-4.99778003e-01 3.37025702e-01 -2.52252549e-01 7.25052953e-01
9.15737569e-01 6.44290745e-02 -7.74637461e-02 1.14002779e-01
-1.01529491e+00 -2.05760747e-01 -1.25452936e+00 9.56737399e-01
-5.46927571e-01 1.12948549e+00 2.14254305e-01 -4.44073677e-01
1.03771842e+00 9.88337547e-02 2.13962138e-01 -8.74836028e-01
5.04926801e-01 6.36908934e-02 -2.35330328e-01 -7.58570671e-01
4.55701977e-01 -4.99699488e-02 8.23526561e-01 -1.98038712e-01
-4.11835641e-01 -1.86865658e-01 -1.06341474e-01 -1.56663537e-01
5.56285024e-01 2.43194818e-01 4.40663040e-01 -6.38925433e-01
9.28817272e-01 3.41352046e-01 3.71683955e-01 2.20923424e-01
-3.19934875e-01 7.70067945e-02 -2.31625974e-01 -3.39945674e-01
-1.01545262e+00 -5.58408678e-01 -4.54084516e-01 4.41744387e-01
1.78565174e-01 1.03758752e-01 -8.37810099e-01 -1.54186293e-01
1.42896876e-01 7.65355825e-01 -2.00341493e-01 6.84971213e-02
-1.25349894e-01 -8.83477569e-01 1.67997897e-01 3.30496937e-01
4.86680329e-01 -7.25397944e-01 -1.18810749e+00 -9.66815203e-02
3.93361509e-01 -9.06510413e-01 3.44745129e-01 2.85739839e-01
-9.42463994e-01 -1.07794440e+00 -6.00878179e-01 -4.90599662e-01
3.84231031e-01 7.27423012e-01 9.07220304e-01 2.38445774e-01
-7.34514475e-01 3.77402723e-01 -4.20539141e-01 -1.13875437e+00
-3.44624192e-01 -3.86172742e-01 1.14899594e-02 -2.50790000e-01
1.66488349e-01 -1.56132013e-01 -6.04071558e-01 -6.12307377e-02
-4.55471247e-01 -2.46497765e-01 5.11295736e-01 4.55471605e-01
2.34604374e-01 3.10686409e-01 2.25431755e-01 -7.41532326e-01
1.67645201e-01 -1.33912131e-01 -1.85334396e+00 2.93165475e-01
-1.44026959e+00 -3.61043185e-01 3.60943556e-01 2.05685094e-01
-1.43392158e+00 2.17848301e-01 2.90072739e-01 -4.26354647e-01
-2.41336703e-01 2.07261741e-01 6.94279671e-02 -9.84598160e-01
8.20076823e-01 1.66977882e-01 -4.72089350e-02 -3.02231848e-01
3.30586433e-02 7.83759832e-01 3.12502623e-01 2.47797400e-01
1.20872974e+00 7.40507483e-01 4.58316952e-01 -1.58053708e+00
-5.40096343e-01 -6.00083172e-01 -6.10571146e-01 -5.86199164e-01
1.01000106e+00 -1.16897249e+00 -5.36599338e-01 6.81294799e-01
-8.55966985e-01 -1.02054439e-01 6.03191219e-02 4.04723912e-01
1.38549255e-02 2.37932891e-01 1.22123979e-01 -1.45784736e+00
-8.14913929e-01 -5.89710832e-01 5.10016620e-01 1.35786843e+00
5.47848046e-01 -1.23739505e+00 4.45769459e-01 3.20664823e-01
9.08042967e-01 4.17371988e-01 2.83907443e-01 7.52825513e-02
-1.13795996e+00 -2.15430260e-01 -1.05924606e-01 6.48087561e-01
2.53466945e-02 6.48347139e-01 -1.24670708e+00 -5.75607121e-01
-3.72966379e-02 2.42829651e-01 5.59127808e-01 4.67732489e-01
6.73659444e-01 1.72091573e-01 -4.03126091e-01 7.97097564e-01
2.18792248e+00 6.01063490e-01 5.53793550e-01 9.09952164e-01
4.52081889e-01 -1.88838765e-01 1.05558491e+00 6.28228009e-01
3.69226746e-02 4.89554197e-01 8.38583589e-01 -4.29495275e-01
-2.97344178e-01 4.29544061e-01 3.15742105e-01 2.02199280e-01
-4.07173693e-01 -1.84120521e-01 -7.12269247e-01 3.68453264e-01
-1.56372631e+00 -8.50222051e-01 -7.58721292e-01 2.34991002e+00
-6.83197603e-02 5.30295186e-02 -2.00125471e-01 1.18731767e-01
5.97146094e-01 1.47555143e-01 -2.37571165e-01 -7.46843874e-01
-2.32861146e-01 3.87039542e-01 1.13678491e+00 6.03298068e-01
-8.79931271e-01 7.43412018e-01 7.14647055e+00 2.40830287e-01
-1.61248589e+00 1.78798363e-01 -3.02281260e-01 2.04172940e-03
9.83389392e-02 4.63252753e-01 -1.03993118e+00 7.06305206e-01
1.26578486e+00 -1.53852746e-01 4.63476002e-01 6.40116453e-01
6.88977301e-01 -9.03092921e-01 -3.11265528e-01 5.75792134e-01
2.55590022e-01 -1.35827875e+00 -4.24716890e-01 -5.52141201e-03
9.57220018e-01 3.59592229e-01 -3.76219541e-01 2.67619900e-02
-9.18093547e-02 -4.57403809e-01 4.28103745e-01 1.00691271e+00
1.10393800e-01 -7.83702254e-01 9.06021059e-01 1.75192535e-01
-1.06777215e+00 -3.89832288e-01 -5.92336774e-01 -6.04535043e-01
-2.44800359e-01 5.12697875e-01 -1.10551953e+00 9.37626243e-01
1.04718244e+00 2.65258133e-01 -9.07019079e-01 1.58074999e+00
-2.86515087e-01 5.67245007e-01 -8.69431913e-01 -6.33626699e-01
2.25748435e-01 -5.49302399e-01 5.65891922e-01 1.00071383e+00
3.95659238e-01 -1.98405430e-01 -7.25111216e-02 4.88194585e-01
8.54151845e-01 -1.05689064e-01 -8.65319192e-01 1.52545467e-01
1.90087125e-01 1.47579408e+00 -5.55327296e-01 -3.94244850e-01
-5.23102820e-01 4.71883953e-01 -4.80486035e-01 2.11674348e-01
-1.00680351e+00 -1.98921502e-01 3.89338464e-01 1.08009361e-01
5.30896604e-01 -6.18713088e-02 -5.39948761e-01 -7.31686771e-01
-5.45273870e-02 -4.46861386e-01 3.41424197e-01 -1.20719552e+00
-3.83228928e-01 1.46032780e-01 -1.65947378e-01 -1.18111718e+00
2.22198546e-01 -9.46117699e-01 -1.25774419e+00 9.51890409e-01
-1.71172345e+00 -1.37960839e+00 -7.78757215e-01 4.53653336e-01
4.56618756e-01 -5.13073742e-01 7.31528699e-01 1.88457370e-01
-8.29050541e-01 -8.33099335e-02 1.01511991e+00 -7.54184783e-01
2.66676396e-01 -1.46060026e+00 -2.15190828e-01 1.34008062e+00
-1.13733128e-01 -2.07679808e-01 1.06137848e+00 -5.00273585e-01
-1.33423066e+00 -6.88418090e-01 4.68692213e-01 -1.19262330e-01
6.60639167e-01 -9.80860516e-02 -5.36800027e-01 5.10813355e-01
8.53562653e-01 -3.79099488e-01 3.10537666e-01 1.03037290e-01
4.65149999e-01 -6.51171684e-01 -1.11423481e+00 3.73791248e-01
-3.22250426e-02 -5.44164889e-02 -3.37743521e-01 9.37380552e-01
3.15112650e-01 -5.88961780e-01 -6.71566010e-01 3.65956217e-01
4.25724179e-01 -1.22025466e+00 6.49812281e-01 -2.14801934e-02
-5.55026054e-01 -6.79458380e-01 2.09502161e-01 -1.21873295e+00
-2.61859298e-01 -6.51894689e-01 -5.09772360e-01 1.29660499e+00
3.79760176e-01 -4.99721140e-01 7.39526629e-01 5.83839640e-02
-1.52149454e-01 -2.07821772e-01 -8.25447738e-01 -7.44162083e-01
-3.44155520e-01 1.07789084e-01 2.29223207e-01 7.21243739e-01
-6.84666455e-01 3.58887494e-01 -2.84440994e-01 1.12113822e+00
5.82452059e-01 3.54125530e-01 7.43946195e-01 -1.45294952e+00
-1.68119674e-03 -2.47291550e-02 -3.95064622e-01 1.00984715e-01
-3.06324989e-01 -2.53831267e-01 -1.04225188e-01 -1.69987893e+00
-1.33743137e-01 -3.46379101e-01 -2.16321275e-01 2.34364077e-01
8.25254545e-02 -2.73780137e-01 1.86783940e-01 -4.27752696e-02
-4.97132502e-02 4.84132886e-01 9.19004321e-01 -1.61095098e-01
-5.20582795e-02 5.35964668e-01 1.37586370e-02 6.70920312e-01
9.47116196e-01 -1.68982387e-01 -1.99148312e-01 -3.07517141e-01
5.35547808e-02 -1.51146844e-01 4.61749524e-01 -1.92151511e+00
3.62218976e-01 -3.51630867e-01 6.84997678e-01 -9.03129160e-01
2.29988918e-01 -1.46158600e+00 2.86255509e-01 1.06261539e+00
6.97260559e-01 -3.48204046e-01 4.18917686e-01 4.41219717e-01
6.33374462e-03 -5.64626157e-01 1.26825976e+00 -1.98942963e-02
-1.02539921e+00 -6.36355579e-02 -7.31221080e-01 -1.00215507e+00
1.69454467e+00 -4.66427118e-01 -5.31579614e-01 -3.82649750e-02
-6.46222234e-01 5.83860099e-01 5.86452365e-01 4.42143142e-01
1.17509611e-01 -8.89357567e-01 -2.67773062e-01 2.23452330e-01
-7.31419548e-02 -1.23361252e-01 2.43608043e-01 6.08970046e-01
-1.30124009e+00 7.22044408e-01 -6.40203595e-01 -7.45388091e-01
-1.73502219e+00 7.13540912e-01 8.81175995e-01 1.62465692e-01
-5.58179319e-01 2.85180062e-01 -4.38268453e-01 -1.24699101e-02
1.37500033e-01 -2.06869692e-01 -5.02362192e-01 -1.58468947e-01
4.08062488e-01 6.54328406e-01 3.44149053e-01 -3.94882023e-01
-1.57607302e-01 7.08474636e-01 5.52815497e-01 3.57825339e-01
1.14754319e+00 -3.57321411e-01 -2.72182494e-01 5.77832937e-01
1.86297581e-01 1.04312282e-02 -1.08588839e+00 3.66009891e-01
-1.02860048e-01 -6.06731236e-01 8.33811164e-01 -1.24227536e+00
-9.35355604e-01 6.51354134e-01 1.77663493e+00 3.58808994e-01
1.39145958e+00 -7.17319250e-01 1.55025855e-01 6.33932590e-01
1.68450505e-01 -1.43033588e+00 -5.01491845e-01 6.29751503e-01
5.28845847e-01 -1.32678413e+00 3.60409409e-01 -2.72186875e-01
-2.24840224e-01 1.34298956e+00 9.52579856e-01 -7.27842422e-03
7.79458165e-01 4.33884829e-01 4.59208012e-01 -2.71513820e-01
-3.87069166e-01 -5.97557604e-01 -2.58994430e-01 8.68964791e-01
2.19469056e-01 3.08727831e-01 -4.83371705e-01 -3.95279735e-01
1.18366957e-01 1.05151050e-01 9.43927467e-01 1.13436902e+00
-1.18042552e+00 -6.57548368e-01 -1.29348266e+00 2.37012550e-01
-1.62931621e-01 2.37809807e-01 -4.55249578e-01 1.06995046e+00
4.91978139e-01 7.53979146e-01 -6.41487092e-02 7.40476176e-02
3.43181133e-01 -1.66322902e-01 2.67926186e-01 -1.87359065e-01
-6.16236091e-01 -4.22444381e-02 -1.89787563e-04 -5.04312888e-02
-4.20966417e-01 -6.33016825e-01 -1.06081057e+00 -2.20347837e-01
-7.13127077e-01 8.53437930e-02 1.36585450e+00 9.31457043e-01
8.88185427e-02 8.16925883e-01 1.07892132e+00 -7.74082839e-01
-3.38686228e-01 -1.19992089e+00 -6.61605954e-01 -4.81737614e-01
5.24617910e-01 -7.78793693e-01 -2.75434643e-01 -2.63369232e-01] | [9.637843132019043, -1.6985565423965454] |
ec7db6ed-a162-4ddf-93a6-3370ad85a24b | direct-superpoints-matching-for-fast-and | 2307.01362 | null | https://arxiv.org/abs/2307.01362v1 | https://arxiv.org/pdf/2307.01362v1.pdf | Direct Superpoints Matching for Fast and Robust Point Cloud Registration | Although deep neural networks endow the downsampled superpoints with discriminative feature representations, directly matching them is usually not used alone in state-of-the-art methods, mainly for two reasons. First, the correspondences are inevitably noisy, so RANSAC-like refinement is usually adopted. Such ad hoc postprocessing, however, is slow and not differentiable, which can not be jointly optimized with feature learning. Second, superpoints are sparse and thus more RANSAC iterations are needed. Existing approaches use the coarse-to-fine strategy to propagate the superpoints correspondences to the point level, which are not discriminative enough and further necessitates the postprocessing refinement. In this paper, we present a simple yet effective approach to extract correspondences by directly matching superpoints using a global softmax layer in an end-to-end manner, which are used to determine the rigid transformation between the source and target point cloud. Compared with methods that directly predict corresponding points, by leveraging the rich information from the superpoints matchings, we can obtain more accurate estimation of the transformation and effectively filter out outliers without any postprocessing refinement. As a result, our approach is not only fast, but also achieves state-of-the-art results on the challenging ModelNet and 3DMatch benchmarks. Our code and model weights will be publicly released. | ['Huaizu Jiang', 'Hanumant Singh', 'Yiming Xie', 'Aniket Gupta'] | 2023-07-03 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [-2.08677158e-01 -3.88791859e-01 -8.60046595e-02 -4.94673222e-01
-8.61833096e-01 -3.90731871e-01 5.60506165e-01 9.80124101e-02
-4.22264814e-01 2.61929989e-01 -1.25797868e-01 1.01451799e-01
-7.98662975e-02 -5.95575154e-01 -9.30829585e-01 -6.51895642e-01
3.05285990e-01 6.16664767e-01 2.95969516e-01 -2.79820323e-01
3.14046115e-01 7.01197505e-01 -1.54496336e+00 -6.45000041e-02
1.07010651e+00 1.20259416e+00 2.62667358e-01 9.53143910e-02
-1.88961495e-02 1.75204277e-01 -1.25427559e-01 -2.78099984e-01
5.43764114e-01 -5.26426733e-03 -4.99321908e-01 2.01524675e-01
9.03688908e-01 -3.80579501e-01 -4.19389099e-01 1.23148656e+00
2.31333122e-01 2.71634907e-01 4.73411262e-01 -1.07076991e+00
-6.45698309e-01 1.80663720e-01 -7.94665039e-01 -2.82295465e-01
1.72294185e-01 3.68534289e-02 1.05653989e+00 -1.29401696e+00
4.57172990e-01 1.17326820e+00 8.75695229e-01 1.42769396e-01
-1.18714762e+00 -9.33821797e-01 3.20738256e-01 2.99009144e-01
-1.66640902e+00 -4.85563606e-01 9.53639627e-01 -4.11933869e-01
7.44429350e-01 1.42419830e-01 7.31629014e-01 9.06484604e-01
-4.57543954e-02 6.47892296e-01 6.32273972e-01 6.28847778e-02
6.30102912e-03 -2.73991108e-01 -1.83330938e-01 7.04073906e-01
5.03257141e-02 3.33334565e-01 -4.86914128e-01 -2.50687785e-02
1.05008471e+00 5.22681832e-01 -2.67466158e-01 -6.13554418e-01
-1.49825799e+00 7.13435471e-01 1.01380026e+00 2.37834662e-01
-6.46143675e-01 1.77116960e-01 8.50475132e-02 1.90700576e-01
4.87723768e-01 3.15933168e-01 -4.64695662e-01 -4.40031514e-02
-1.31541324e+00 3.58811587e-01 2.56888092e-01 1.16750312e+00
1.02882564e+00 -1.69144630e-01 1.47044525e-01 7.10390329e-01
4.13586676e-01 4.85814482e-01 4.37505096e-01 -7.22715557e-01
5.91059327e-01 6.47620678e-01 2.25091845e-01 -1.28822565e+00
-5.54190934e-01 -4.80756342e-01 -1.12581336e+00 2.64506727e-01
3.30500722e-01 1.06034301e-01 -1.19864583e+00 1.40525460e+00
4.35992002e-01 6.64439559e-01 -3.36537987e-01 1.21332872e+00
5.56149065e-01 4.31153417e-01 -3.10984284e-01 4.01906371e-02
1.02212036e+00 -1.10867429e+00 -4.90765244e-01 -2.94884950e-01
4.65596914e-01 -1.10208488e+00 9.00618434e-01 2.27600351e-01
-9.24259365e-01 -9.14102077e-01 -1.13008094e+00 -3.10267031e-01
-1.50908098e-01 6.11686409e-02 5.70772529e-01 -1.85309663e-01
-7.19475627e-01 8.72657955e-01 -1.27925527e+00 -7.73841664e-02
6.28377497e-01 4.70515788e-01 -5.77688694e-01 -2.81296894e-02
-6.84134722e-01 9.19485331e-01 3.17738593e-01 3.66525799e-01
-3.75101566e-01 -8.55318308e-01 -1.05763483e+00 1.02114551e-01
3.91762495e-01 -6.99365377e-01 1.20021558e+00 -7.29897499e-01
-1.44137800e+00 5.37034512e-01 -4.57121491e-01 -2.00305745e-01
7.17868388e-01 -4.94412005e-01 -1.89742446e-01 -2.34429717e-01
1.73508942e-01 5.81036508e-01 1.04800022e+00 -1.26760340e+00
-9.05782342e-01 -3.74801934e-01 -3.08638453e-01 9.52051729e-02
-6.90344572e-02 -1.94592968e-01 -8.03684950e-01 -8.00848067e-01
8.38157535e-01 -1.00187838e+00 -5.22669554e-01 2.21055403e-01
-4.95998442e-01 -2.19410211e-01 1.02494133e+00 -4.53796834e-01
8.35084975e-01 -2.26825166e+00 2.34203070e-01 4.44427937e-01
2.05705911e-01 1.65899888e-01 -1.69393614e-01 7.63889179e-02
-8.33603516e-02 -2.05928624e-01 -3.19872618e-01 -7.52042770e-01
9.89730507e-02 -4.49573621e-02 -4.84084129e-01 8.18370581e-01
3.08723718e-01 8.05925608e-01 -8.36062908e-01 -2.29946315e-01
6.83953047e-01 7.58976340e-01 -6.37492478e-01 8.59180689e-02
-9.37917456e-02 7.45263517e-01 -5.59938788e-01 6.26450539e-01
1.03545988e+00 -2.78352022e-01 -4.96168166e-01 -6.21287525e-01
-2.03342482e-01 1.10100962e-01 -1.45872450e+00 2.15606976e+00
-5.01023769e-01 3.75460356e-01 -2.19684631e-01 -8.10206771e-01
1.05528498e+00 7.83741474e-02 7.95739889e-01 -5.68889022e-01
9.61627066e-02 4.81173933e-01 -1.41475633e-01 -1.63351968e-01
4.73592132e-01 -1.14405081e-01 6.47859275e-02 -2.00614184e-02
3.61052081e-02 -2.92186022e-01 -9.28253457e-02 -3.31559986e-01
6.45855844e-01 5.35060585e-01 2.57493854e-01 9.85517278e-02
5.44332683e-01 -8.14355072e-03 8.63210678e-01 3.28852445e-01
1.29051656e-01 1.08146107e+00 -1.68636873e-01 -5.45233905e-01
-9.82521057e-01 -9.85089719e-01 -2.38223508e-01 6.16442800e-01
6.15598619e-01 -3.81508529e-01 -5.08139789e-01 -5.75118661e-01
3.04443508e-01 3.27254325e-01 -2.73482263e-01 -1.82088122e-01
-8.67001116e-01 -3.54532838e-01 4.13808189e-02 6.93128347e-01
4.90057349e-01 -7.21578956e-01 -4.02592391e-01 3.91144753e-01
-4.58572581e-02 -1.32779241e+00 -7.39473462e-01 -4.96781543e-02
-1.09497774e+00 -1.04159999e+00 -7.22145259e-01 -7.31060922e-01
8.81970525e-01 4.05347168e-01 8.18156898e-01 2.91913807e-01
1.25695050e-01 -3.66142213e-01 -2.98706502e-01 -1.99970007e-01
2.40400285e-02 1.55895084e-01 2.58291215e-01 1.23262301e-01
4.55539167e-01 -7.00497210e-01 -6.47353649e-01 4.57012773e-01
-6.39366388e-01 1.73472345e-01 6.94586217e-01 8.72718751e-01
1.12572491e+00 -5.50642759e-02 2.28554115e-01 -5.44480085e-01
-3.91183933e-03 -1.03068724e-01 -1.02371955e+00 3.22126113e-02
-3.91769022e-01 1.03507027e-01 9.34903264e-01 -3.47840875e-01
-8.27695966e-01 6.15226686e-01 -4.05193925e-01 -1.03074884e+00
-3.07543129e-01 5.18476963e-01 -6.65621683e-02 -4.28753585e-01
3.78357887e-01 6.99312314e-02 -1.38355181e-01 -7.99441814e-01
2.42586017e-01 4.61764127e-01 8.26956272e-01 -3.25587392e-01
1.48782861e+00 7.43265331e-01 8.11309963e-02 -6.49170160e-01
-1.00381458e+00 -7.80290306e-01 -1.00201344e+00 2.38566324e-01
6.55961275e-01 -9.30977285e-01 -5.48259079e-01 5.89193225e-01
-1.28575909e+00 -2.86555733e-03 -9.41445827e-02 8.40487361e-01
-7.15544879e-01 4.23255801e-01 -4.50387895e-01 -2.32247442e-01
-2.17563108e-01 -1.44932926e+00 1.36266005e+00 2.67013341e-01
-3.61571200e-02 -6.48066640e-01 -1.10637695e-01 1.26267806e-01
2.56969124e-01 2.83125699e-01 3.74867201e-01 -5.05744338e-01
-8.33453059e-01 -5.88688850e-01 -2.59927124e-01 1.34130940e-01
3.15369397e-01 1.37834758e-01 -8.06342840e-01 -5.51298320e-01
5.88586666e-02 -4.69950810e-02 6.87883735e-01 2.67820805e-01
1.44688666e+00 -7.26084262e-02 -3.89139682e-01 1.31386018e+00
1.36803079e+00 -2.92401701e-01 4.81506348e-01 4.96616513e-01
1.28720725e+00 3.05584818e-01 9.92438436e-01 3.65491897e-01
5.06374776e-01 9.14233446e-01 6.54017866e-01 -2.75488824e-01
7.87220746e-02 -4.18741107e-01 8.00310746e-02 8.76942337e-01
-3.05119514e-01 3.59275460e-01 -7.45399475e-01 4.15208906e-01
-2.12210560e+00 -7.17333853e-01 -1.57032683e-01 2.42002249e+00
7.00732768e-01 2.60486491e-02 -6.59820139e-02 -2.05763560e-02
7.19221771e-01 1.52914479e-01 -6.95506811e-01 3.38741094e-01
8.93695727e-02 2.69112557e-01 6.38735056e-01 4.48101282e-01
-1.19128072e+00 1.15074301e+00 4.59834290e+00 7.33584404e-01
-1.42084098e+00 -1.31178170e-01 1.22530475e-01 -1.52481303e-01
-5.92566235e-03 1.51233062e-01 -8.14385116e-01 5.33546507e-01
3.98285717e-01 1.08137414e-01 4.21666533e-01 9.28738177e-01
1.84540823e-01 3.96184981e-01 -1.39653277e+00 1.27107739e+00
-2.12220803e-01 -1.45934510e+00 -3.32805291e-02 -5.43329818e-03
7.46717930e-01 4.77183282e-01 -6.19022995e-02 1.82010844e-01
1.05187416e-01 -8.86920214e-01 9.35379863e-01 5.53918183e-01
5.66399217e-01 -9.47435319e-01 7.53763855e-01 4.09135312e-01
-1.31936514e+00 1.20600477e-01 -6.85146749e-01 2.85096139e-01
4.04060334e-01 6.32810771e-01 -4.08935547e-01 7.79488623e-01
9.67816770e-01 1.00198603e+00 -3.89065474e-01 1.44435227e+00
-3.13667834e-01 4.38704938e-02 -7.09895432e-01 4.42405671e-01
3.67008299e-01 -2.86367506e-01 4.92186815e-01 8.87988806e-01
7.65840113e-01 -1.92129090e-01 4.60650712e-01 9.08384979e-01
-9.78276432e-02 -1.20198373e-02 -1.80036098e-01 4.12362814e-01
5.75337350e-01 1.35316527e+00 -6.13156140e-01 -2.43541196e-01
-5.82267284e-01 1.08063877e+00 4.01099056e-01 3.05461287e-01
-7.60729134e-01 -5.00264585e-01 8.75391424e-01 6.46895468e-02
5.90647995e-01 -4.83271390e-01 -4.43835616e-01 -1.45251870e+00
3.33664656e-01 -8.01163316e-01 1.98038772e-01 -4.04391885e-01
-1.48724139e+00 5.85082114e-01 -2.61311382e-01 -1.77792740e+00
-2.96098918e-01 -4.43919808e-01 -6.20391726e-01 9.33374524e-01
-1.67715120e+00 -1.35999882e+00 -6.29890740e-01 6.63053811e-01
4.89890277e-01 1.85584083e-01 5.22355199e-01 4.57714498e-01
-4.71194565e-01 7.60560870e-01 1.05181150e-01 2.26759076e-01
8.44662368e-01 -1.00841403e+00 6.15110397e-01 9.96217608e-01
3.39476973e-01 7.38214850e-01 5.99349737e-01 -5.29660106e-01
-1.28729963e+00 -1.17225182e+00 6.25459492e-01 -2.58346200e-01
6.61683023e-01 -2.93916970e-01 -1.24935853e+00 6.72865033e-01
-2.01531291e-01 4.84622896e-01 1.98093191e-01 1.11085869e-01
-3.63717556e-01 -1.71798512e-01 -1.04661894e+00 6.94560587e-01
1.10879838e+00 -4.37937289e-01 -6.12935364e-01 2.37822071e-01
7.00600684e-01 -9.06730354e-01 -9.57857609e-01 6.59484267e-01
2.96339750e-01 -9.32434678e-01 1.16864073e+00 -3.72940063e-01
2.94132084e-01 -7.25825906e-01 -6.04182445e-02 -1.38227856e+00
-4.70623523e-01 -5.83480537e-01 -2.22632632e-01 1.02388716e+00
2.91153252e-01 -6.37525499e-01 8.53919506e-01 5.46465993e-01
-4.13010329e-01 -8.55953574e-01 -1.05996037e+00 -8.68604898e-01
-4.66909669e-02 -4.84114170e-01 1.05983150e+00 9.95489657e-01
-3.88570458e-01 1.17946282e-01 -3.41473311e-01 5.82592547e-01
7.05507874e-01 4.90991741e-01 9.07926023e-01 -1.32978189e+00
-1.35151520e-01 -5.73207498e-01 -6.81708574e-01 -1.33680534e+00
2.04581454e-01 -7.66570032e-01 4.13003117e-01 -1.34028602e+00
-2.58345276e-01 -7.48741925e-01 -3.80214453e-01 4.31068987e-01
-3.40409458e-01 3.52489024e-01 1.97466537e-01 5.50439239e-01
-3.09517562e-01 7.39121795e-01 1.35112894e+00 1.11766625e-02
-2.86019266e-01 1.33109525e-01 -5.03126204e-01 1.13425863e+00
6.21812999e-01 -4.49188322e-01 -2.65773945e-02 -6.80549026e-01
-8.90315324e-02 -2.55333871e-01 6.34634495e-01 -1.19369173e+00
4.99675363e-01 -3.55116248e-01 6.01099968e-01 -9.98709440e-01
5.33330500e-01 -1.20835161e+00 8.45149681e-02 1.47200719e-01
5.39963320e-03 1.11165129e-01 -5.22125289e-02 5.34005225e-01
-4.58778024e-01 -1.09850772e-01 8.42862487e-01 1.30892560e-01
-6.67250097e-01 1.12421298e+00 4.09830779e-01 -4.22465622e-01
9.15988326e-01 -3.21703106e-01 -1.55725107e-02 -2.23023772e-01
-3.61198515e-01 2.99498826e-01 8.82851958e-01 6.41980827e-01
7.27645040e-01 -1.56570423e+00 -6.76041245e-01 4.30895835e-01
2.14263678e-01 9.23081279e-01 2.92604774e-01 1.06597793e+00
-3.61716062e-01 8.51894543e-02 -6.69828504e-02 -9.31768596e-01
-9.50774789e-01 6.44755065e-01 2.11965173e-01 1.20427892e-01
-1.02693868e+00 7.05239415e-01 2.01788276e-01 -5.47080338e-01
2.76318759e-01 -5.18338501e-01 6.69843704e-02 -1.68621212e-01
4.21536565e-01 1.17605641e-01 2.29352400e-01 -9.93834853e-01
-4.11947787e-01 1.10965228e+00 -2.61329651e-01 2.51001120e-01
1.47689104e+00 -3.56485066e-03 -3.30496021e-02 8.02662596e-02
1.52850926e+00 -9.88472067e-03 -1.60402179e+00 -5.73938251e-01
-7.97527563e-03 -9.18468654e-01 1.64128810e-01 -3.92091013e-02
-1.43487847e+00 8.67960334e-01 4.19897079e-01 -4.05200049e-02
9.53406453e-01 -5.90159819e-02 1.23447263e+00 5.00648141e-01
3.73448133e-01 -1.13552928e+00 -1.48147434e-01 5.80194652e-01
1.00553882e+00 -1.41114390e+00 6.80982694e-02 -5.47709644e-01
-2.80354351e-01 1.20163822e+00 7.46453166e-01 -4.88967150e-01
6.60728872e-01 7.08023235e-02 1.22556351e-01 -1.44671187e-01
-2.02597797e-01 -1.70717075e-01 5.91569960e-01 4.28852469e-01
2.10343897e-01 -1.85244426e-01 2.25853100e-01 3.99742365e-01
-4.91096348e-01 3.89466882e-02 9.59583744e-03 7.36726284e-01
-2.67103851e-01 -1.07212722e+00 -5.08814096e-01 4.74376976e-01
-2.49357775e-01 2.48240270e-02 -7.59322383e-03 6.84208691e-01
1.10152431e-01 6.54402137e-01 2.33712941e-01 -5.59739590e-01
7.08773255e-01 -4.53953862e-01 3.46106380e-01 -4.42167372e-01
-3.27164441e-01 2.74174869e-01 -4.28061336e-01 -9.24598336e-01
-3.51364970e-01 -6.53471589e-01 -1.48941863e+00 -4.54269558e-01
-5.27131498e-01 -1.78908139e-01 5.25661826e-01 1.03145862e+00
4.45563376e-01 2.49299422e-01 8.48069668e-01 -1.48481143e+00
-7.56270587e-01 -7.71203876e-01 -1.17718995e-01 5.89367390e-01
5.66149056e-01 -8.46901476e-01 -3.04486364e-01 -1.52401179e-01] | [7.6676716804504395, -3.0945184230804443] |
5301b032-0d99-48bc-bb5d-cf31490c29a3 | rmpe-regional-multi-person-pose-estimation | 1612.00137 | null | http://arxiv.org/abs/1612.00137v5 | http://arxiv.org/pdf/1612.00137v5.pdf | RMPE: Regional Multi-person Pose Estimation | Multi-person pose estimation in the wild is challenging. Although
state-of-the-art human detectors have demonstrated good performance, small
errors in localization and recognition are inevitable. These errors can cause
failures for a single-person pose estimator (SPPE), especially for methods that
solely depend on human detection results. In this paper, we propose a novel
regional multi-person pose estimation (RMPE) framework to facilitate pose
estimation in the presence of inaccurate human bounding boxes. Our framework
consists of three components: Symmetric Spatial Transformer Network (SSTN),
Parametric Pose Non-Maximum-Suppression (NMS), and Pose-Guided Proposals
Generator (PGPG). Our method is able to handle inaccurate bounding boxes and
redundant detections, allowing it to achieve a 17% increase in mAP over the
state-of-the-art methods on the MPII (multi person) dataset.Our model and
source codes are publicly available. | ['Yu-Wing Tai', 'Hao-Shu Fang', 'Shuqin Xie', 'Cewu Lu'] | 2016-12-01 | rmpe-regional-multi-person-pose-estimation-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Fang_RMPE_Regional_Multi-Person_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Fang_RMPE_Regional_Multi-Person_ICCV_2017_paper.pdf | iccv-2017-10 | ['2d-human-pose-estimation'] | ['computer-vision'] | [ 1.74595322e-02 5.93229011e-02 4.83305126e-01 -2.58092135e-01
-9.55082834e-01 -3.98312956e-01 5.31493962e-01 -4.05177921e-02
-8.64347577e-01 7.86613286e-01 -1.44294295e-02 4.47433889e-01
3.14496845e-01 -4.95275944e-01 -7.95148075e-01 -2.22602561e-01
-1.32849747e-02 1.00759304e+00 6.66290402e-01 -1.50618151e-01
-3.01085442e-01 6.37876153e-01 -1.53265572e+00 4.16340493e-02
4.61458236e-01 7.00140715e-01 1.25090629e-02 6.12626076e-01
4.99196351e-01 1.71499050e-04 -9.52149928e-01 -4.69474345e-01
3.53683889e-01 -2.45966092e-02 -2.59191483e-01 -1.94465876e-01
7.94702947e-01 -5.73731542e-01 -1.81656167e-01 6.65122092e-01
1.01125050e+00 5.14923371e-02 5.98197579e-01 -1.50719762e+00
3.98236483e-01 1.03639051e-01 -8.32804978e-01 -6.07641414e-02
6.49720132e-01 9.79822204e-02 3.20689708e-01 -1.16945410e+00
6.89910889e-01 1.58401453e+00 1.16811705e+00 5.63591897e-01
-1.28060067e+00 -8.20704520e-01 -3.34189460e-02 4.85051796e-02
-1.88807797e+00 -3.57859671e-01 2.47896433e-01 -4.40315813e-01
8.53977859e-01 8.04539770e-02 7.68313050e-01 1.36808825e+00
3.08344550e-02 6.53450787e-01 9.50698793e-01 -3.71524245e-01
2.12733969e-02 2.87916571e-01 -2.81043589e-01 5.90132773e-01
7.55231321e-01 2.30635539e-01 -8.22999835e-01 -1.04796007e-01
7.76426375e-01 -1.93282902e-01 9.84078497e-02 -6.34395361e-01
-1.22706962e+00 6.26492679e-01 6.60840571e-01 -6.32747561e-02
-5.70120990e-01 9.35968012e-02 3.62285942e-01 -4.36842889e-01
4.21790838e-01 1.96410030e-01 -1.18908212e-01 -9.19878110e-02
-1.28084123e+00 9.11417425e-01 4.98020589e-01 9.87900615e-01
2.81850129e-01 -2.48658225e-01 -4.41662878e-01 5.97733915e-01
3.72354805e-01 9.25466895e-01 -8.59321803e-02 -4.73686665e-01
5.71217000e-01 4.65112150e-01 4.94620532e-01 -1.19182944e+00
-9.72494721e-01 -5.73546112e-01 -7.11897016e-01 1.00369006e-01
6.85566008e-01 -3.62331063e-01 -8.32390785e-01 1.75071359e+00
7.00666547e-01 -2.56567419e-01 -6.16699159e-01 1.23179805e+00
7.30337441e-01 3.83444399e-01 2.71315247e-01 9.27087292e-02
1.48318565e+00 -8.05722475e-01 -4.45224285e-01 -9.31312978e-01
1.70772314e-01 -5.95148563e-01 4.09062982e-01 2.64227778e-01
-8.46705496e-01 -7.60483265e-01 -9.97483194e-01 2.00624481e-01
-2.94067085e-01 6.68333471e-01 3.69355679e-01 8.68756711e-01
-9.18294847e-01 2.69824505e-01 -7.74047673e-01 -8.47146571e-01
3.15268546e-01 4.88565028e-01 -6.36557221e-01 1.59217179e-01
-9.77977335e-01 1.15296948e+00 4.62428629e-01 4.71149445e-01
-7.46977448e-01 -5.29514730e-01 -8.57496858e-01 -2.75498420e-01
6.27752483e-01 -8.44558418e-01 1.03286994e+00 -2.96676606e-01
-1.00506902e+00 8.24245751e-01 -1.33790821e-01 -5.57903409e-01
1.28236902e+00 -6.78418040e-01 -1.26203105e-01 8.55931416e-02
3.38596255e-01 1.09280336e+00 8.66235435e-01 -1.18491423e+00
-6.86047256e-01 -6.71690285e-01 -5.59732556e-01 2.53892154e-01
-7.54858106e-02 3.97181809e-01 -6.78707242e-01 -5.62924266e-01
1.33367345e-01 -1.20632517e+00 -2.77664542e-01 2.91716546e-01
-7.01168776e-01 1.00043109e-02 6.80635929e-01 -1.12070334e+00
1.05691338e+00 -1.79061532e+00 2.40663871e-01 2.43572176e-01
9.36571807e-02 3.29998821e-01 4.47506495e-02 2.71441072e-01
2.62610286e-01 -3.56801659e-01 7.38145784e-02 -9.18095767e-01
1.86168626e-02 -2.55560458e-01 1.80687845e-01 8.35477293e-01
2.66768243e-02 8.05858850e-01 -5.39482355e-01 -6.77424252e-01
4.45204377e-01 7.05798388e-01 -3.49498391e-01 1.19463108e-01
1.68317184e-01 5.04504681e-01 -1.81982771e-01 7.83731759e-01
8.01826894e-01 1.24931231e-01 1.34297609e-01 -3.36739004e-01
-2.49136746e-01 -1.46337971e-01 -1.78685319e+00 1.43854535e+00
-2.70696655e-02 3.32845896e-01 2.68801212e-01 -1.57891572e-01
7.92203784e-01 1.51900023e-01 3.49576145e-01 -1.71036512e-01
3.03456336e-01 2.00314641e-01 -1.87874243e-01 2.10374713e-01
8.14759433e-01 1.70389175e-01 -2.28350490e-01 1.17645465e-01
9.73455533e-02 3.66250634e-01 3.94824982e-01 1.48102716e-01
9.35275435e-01 4.54017997e-01 4.14855063e-01 -1.23614781e-01
4.37083393e-01 -1.35951281e-01 5.38405657e-01 9.34192896e-01
-6.46554410e-01 7.84014940e-01 1.20267682e-01 -3.98521185e-01
-1.22378826e+00 -1.14419651e+00 -7.01017231e-02 1.01411474e+00
1.90009382e-02 -4.93295103e-01 -8.05216551e-01 -5.56006908e-01
1.59772784e-01 2.12764338e-01 -5.01015544e-01 1.82532683e-01
-7.51937568e-01 -6.99231923e-01 7.64803767e-01 7.15374470e-01
6.07955754e-01 -9.73387599e-01 -9.23543870e-01 2.47991711e-01
-3.73962849e-01 -1.42646623e+00 -4.17483866e-01 -1.20067857e-01
-3.57049346e-01 -9.88006115e-01 -1.13987339e+00 -2.76790828e-01
5.85705996e-01 8.37257653e-02 8.30616713e-01 -3.91736776e-01
-6.46914124e-01 2.95039922e-01 -9.96039659e-02 -4.97010589e-01
-7.60235712e-02 1.87947348e-01 5.47235668e-01 -1.44409627e-01
2.78732389e-01 -1.53670952e-01 -7.03715801e-01 6.11262858e-01
-1.20660648e-01 -2.09733043e-02 5.87409735e-01 5.95960200e-01
6.21795595e-01 -3.47259104e-01 2.06021994e-01 -5.55803835e-01
2.33236149e-01 8.72569308e-02 -7.01443911e-01 9.39423367e-02
-3.51822048e-01 -3.31604898e-01 1.22995257e-01 -3.37251812e-01
-1.00648141e+00 5.89766324e-01 -1.09311000e-01 -2.75224596e-01
-3.43268454e-01 -2.25663364e-01 -1.07902855e-01 -3.27634752e-01
9.35974360e-01 8.11702162e-02 -7.45588765e-02 -5.64813018e-01
2.16573700e-01 5.07140517e-01 8.10202479e-01 -3.89729083e-01
9.67100561e-01 4.64148879e-01 1.26559302e-01 -8.46767664e-01
-6.78689599e-01 -8.08363557e-01 -1.05420303e+00 -6.84241295e-01
9.03383195e-01 -1.24245572e+00 -7.58557439e-01 6.80693090e-01
-1.33155954e+00 1.10175885e-01 1.52510464e-01 2.49081105e-01
-2.93751717e-01 3.09161335e-01 -4.29461300e-01 -1.27308893e+00
-4.49350119e-01 -8.71303082e-01 1.53625536e+00 3.34450245e-01
-4.26504672e-01 -3.22121292e-01 -1.32353893e-02 4.70622271e-01
3.28044802e-01 6.44071162e-01 -3.28550488e-01 -4.19841856e-01
-4.20451224e-01 -7.00376213e-01 -3.02787185e-01 -7.42872953e-02
-5.03917694e-01 -3.19278389e-01 -9.55852926e-01 -7.00617015e-01
-6.65577650e-01 -3.09820712e-01 8.02445769e-01 6.16649449e-01
4.67841357e-01 4.42238376e-02 -8.73894691e-01 3.86918038e-01
1.10426974e+00 -4.82223809e-01 3.84642899e-01 4.61501330e-01
7.43611932e-01 6.78869963e-01 7.47499764e-01 7.27681935e-01
4.65892047e-01 1.35571373e+00 2.78016508e-01 -1.74364313e-01
-2.87378430e-01 -4.69345182e-01 3.16200733e-01 -3.05499911e-01
-4.76235092e-01 -6.29918799e-02 -9.79687691e-01 3.58523428e-01
-1.90954721e+00 -8.71296167e-01 -3.18402350e-01 2.42023945e+00
3.70286524e-01 2.58356839e-01 8.41857433e-01 1.20640127e-02
1.04168785e+00 -3.96549068e-02 -2.07382634e-01 3.12992692e-01
7.49559402e-02 -7.39161149e-02 8.46728146e-01 2.12826177e-01
-1.71958709e+00 9.70982194e-01 6.20768452e+00 5.82714856e-01
-4.43916649e-01 3.38043958e-01 1.47606000e-01 -1.76269576e-01
7.75726259e-01 -3.98666352e-01 -1.59728122e+00 3.38374674e-01
6.14278615e-01 3.68484020e-01 3.29105668e-02 1.14477503e+00
1.40034854e-01 -4.39102739e-01 -8.62589836e-01 1.17409396e+00
2.32563540e-01 -7.42163002e-01 -4.51870292e-01 7.62292445e-02
3.48382354e-01 -2.29983881e-01 -1.92673624e-01 3.91102195e-01
-2.00437561e-01 -7.16515720e-01 1.09786832e+00 4.42839265e-01
7.43954301e-01 -9.11346257e-01 9.61439729e-01 5.00271559e-01
-1.57922006e+00 1.83925091e-03 -2.89532125e-01 1.15947947e-01
5.32582343e-01 5.00484645e-01 -9.98863697e-01 4.53091890e-01
8.99998307e-01 2.01344505e-01 -1.01610279e+00 1.36034870e+00
-3.11328441e-01 1.24592781e-01 -8.34463894e-01 -8.33155364e-02
-2.25014240e-01 3.39633018e-01 8.10142934e-01 1.43700755e+00
2.68019587e-01 -2.92642236e-01 6.83447599e-01 8.55951369e-01
2.08267704e-01 -4.20279466e-02 -1.21408045e-01 4.51215446e-01
6.15604758e-01 1.42815578e+00 -9.35992897e-01 -2.03331903e-01
3.61165591e-02 1.22632790e+00 2.94831753e-01 -1.13679860e-02
-1.06484234e+00 -8.19617063e-02 5.46354830e-01 5.71790516e-01
2.77218163e-01 -3.20636809e-01 -6.07884862e-02 -8.58452916e-01
3.50820988e-01 -6.88662529e-01 3.91081154e-01 -5.29066324e-01
-1.08981109e+00 5.00510812e-01 3.15111697e-01 -1.14915454e+00
-4.03134704e-01 -5.82682490e-01 -1.13901049e-01 8.52254152e-01
-8.53335857e-01 -1.69206858e+00 -5.36635876e-01 3.56391668e-01
2.81191885e-01 -2.79211998e-02 7.08730996e-01 5.75065672e-01
-5.89484453e-01 8.45092297e-01 -5.41262448e-01 2.88247645e-01
9.08444822e-01 -1.02049720e+00 5.55586874e-01 1.19044483e+00
-1.22631051e-01 5.18186271e-01 1.06763732e+00 -1.13845587e+00
-1.09325278e+00 -1.19628441e+00 9.21803832e-01 -6.78641438e-01
1.08894385e-01 -8.39727104e-01 -3.34371090e-01 6.63869977e-01
-5.27504206e-01 2.59287301e-02 1.64340466e-01 2.97795802e-01
-2.90302843e-01 -1.03518078e-02 -1.22650409e+00 4.89098579e-01
1.11282206e+00 -2.24955436e-02 -2.42457911e-01 4.97493744e-01
9.29110572e-02 -7.52435505e-01 -3.78809988e-01 4.43671435e-01
7.73111045e-01 -1.13870108e+00 1.41858256e+00 6.73502833e-02
-2.10786447e-01 -6.45875394e-01 8.40089470e-02 -9.70003963e-01
-4.62994516e-01 -2.76683956e-01 -1.88071907e-01 1.01398015e+00
2.14939684e-01 -4.31113094e-01 9.38413560e-01 5.50710499e-01
2.22304970e-01 -2.01282546e-01 -1.27365994e+00 -1.03836834e+00
-5.35956383e-01 -3.58545274e-01 1.21476501e-01 1.18387669e-01
-2.29358941e-01 2.80773908e-01 -7.83961415e-01 3.99894923e-01
1.11132646e+00 -3.92895073e-01 1.14651167e+00 -1.20545518e+00
-3.16523910e-01 -1.17804967e-01 -7.45830953e-01 -7.53178954e-01
-2.04568654e-01 -3.25739831e-01 4.28891271e-01 -1.48081934e+00
4.92239922e-01 -9.41345915e-02 2.78073400e-01 5.60551763e-01
-1.82917088e-01 6.46054626e-01 4.13319528e-01 4.59171049e-02
-8.32067549e-01 4.12713438e-01 6.31520391e-01 1.56649828e-01
-1.06800705e-01 2.02071577e-01 -1.28529713e-01 7.37238944e-01
4.47531283e-01 -5.72297037e-01 2.49090806e-01 2.13239565e-01
-6.25551194e-02 -1.26368538e-01 9.81665611e-01 -1.77396631e+00
4.89141822e-01 4.43140626e-01 1.00475097e+00 -1.07754302e+00
8.38859499e-01 -5.25277019e-01 3.87039274e-01 7.75150716e-01
1.66544706e-01 -2.83079757e-03 3.58989626e-01 4.58584160e-01
1.26939371e-01 2.01951697e-01 9.01835680e-01 -1.18353076e-01
-5.16925991e-01 1.62930235e-01 -1.17914394e-01 -1.00501701e-01
1.01601470e+00 -2.06719607e-01 -1.97210833e-01 -2.79142678e-01
-7.10782230e-01 1.48713768e-01 2.89943725e-01 4.49904829e-01
5.94347954e-01 -1.28085864e+00 -9.82670069e-01 1.12266466e-01
1.67983323e-01 -2.14158788e-01 3.88578445e-01 8.98967445e-01
-5.28773069e-01 6.19854152e-01 -2.19834998e-01 -6.24431193e-01
-1.57941806e+00 1.50745347e-01 4.48671997e-01 -4.25085664e-01
-5.39432049e-01 9.96980250e-01 -1.75843537e-01 -5.48008144e-01
3.87140572e-01 2.37696588e-01 7.93952402e-03 1.01750933e-01
8.14724028e-01 7.04856217e-01 1.73786059e-02 -9.94076490e-01
-8.78820956e-01 4.92010146e-01 -7.19196200e-02 -2.63326436e-01
1.00891280e+00 6.26944602e-02 2.33868986e-01 -8.20301753e-03
5.17201304e-01 -1.66893348e-01 -1.34992766e+00 4.56008082e-03
-2.34615251e-01 -5.19626915e-01 -1.76234856e-01 -8.99858177e-01
-4.76810783e-01 4.98175353e-01 1.00021052e+00 -3.29975158e-01
5.33777654e-01 1.34451151e-01 6.51688933e-01 2.61346102e-01
8.72294903e-01 -1.33113170e+00 -6.91080615e-02 3.54538947e-01
1.02147985e+00 -1.26436675e+00 3.97093594e-01 -7.67212331e-01
-4.72244114e-01 8.28212798e-01 8.98721099e-01 -7.77436793e-02
3.35631222e-01 3.18954170e-01 -1.73715994e-01 -6.71045557e-02
-1.51104378e-02 -3.92936826e-01 5.08503437e-01 8.92649889e-01
2.66438186e-01 2.03830838e-01 -2.36036509e-01 7.29453564e-01
-3.02329361e-01 -1.73256934e-01 1.24148168e-01 8.10722053e-01
-5.03225386e-01 -8.24816167e-01 -1.05187082e+00 1.71764642e-01
-7.46680498e-02 1.87575445e-01 -4.27443117e-01 1.10420501e+00
3.46985757e-01 8.37442279e-01 -2.29325399e-01 -3.76791209e-01
7.46020079e-01 4.09535617e-02 7.15992689e-01 -5.87595403e-01
-6.69721723e-01 2.46984258e-01 2.24883467e-01 -9.55200672e-01
-1.92465112e-01 -1.03776395e+00 -8.69312584e-01 -3.22260380e-01
-2.82668233e-01 -4.14142698e-01 6.90124333e-01 8.62314582e-01
2.05984339e-01 4.81348723e-01 -1.77206039e-01 -1.54791379e+00
-5.04454255e-01 -1.20008993e+00 -5.83233654e-01 1.96070015e-01
-1.82743803e-01 -1.04106939e+00 -2.78295260e-02 -2.27516219e-01] | [7.137802600860596, -0.8205591440200806] |
ffd39386-dbff-40c6-898e-eb00c3defdda | good-great-excellent-global-inference-of | null | null | https://aclanthology.org/Q13-1023 | https://aclanthology.org/Q13-1023.pdf | Good, Great, Excellent: Global Inference of Semantic Intensities | Adjectives like good, great, and excellent are similar in meaning, but differ in intensity. Intensity order information is very useful for language learners as well as in several NLP tasks, but is missing in most lexical resources (dictionaries, WordNet, and thesauri). In this paper, we present a primarily unsupervised approach that uses semantics from Web-scale data (e.g., phrases like good but not excellent) to rank words by assigning them positions on a continuous scale. We rely on Mixed Integer Linear Programming to jointly determine the ranks, such that individual decisions benefit from global information. When ranking English adjectives, our global algorithm achieves substantial improvements over previous work on both pairwise and rank correlation metrics (specifically, 70{\%} pairwise accuracy as compared to only 56{\%} by previous work). Moreover, our approach can incorporate external synonymy information (increasing its pairwise accuracy to 78{\%}) and extends easily to new languages. We also make our code and data freely available. | ['Gerard de Melo', 'Mohit Bansal'] | 2013-01-01 | null | null | null | tacl-2013-1 | ['subjectivity-analysis'] | ['natural-language-processing'] | [-1.57063767e-01 1.31566331e-01 -5.56135535e-01 -6.30236447e-01
-9.25598562e-01 -1.11252809e+00 4.00405824e-01 9.75257933e-01
-9.78220463e-01 7.70872116e-01 4.62513298e-01 -1.43065378e-01
-4.75263685e-01 -8.93088043e-01 -2.58863300e-01 -3.01372617e-01
5.55200428e-02 7.56495953e-01 1.45319253e-01 -5.39884746e-01
4.11091000e-01 1.42968193e-01 -1.60925508e+00 2.18478173e-01
1.18743289e+00 7.71569312e-01 7.62422085e-02 9.28047299e-02
-5.63091099e-01 5.00657201e-01 -4.11016226e-01 -8.15987289e-01
-4.59563583e-02 7.83144012e-02 -8.57911825e-01 -4.52979088e-01
4.19100761e-01 -1.47174731e-01 4.64083076e-01 1.11295128e+00
2.93355405e-01 4.03310686e-01 8.59777629e-01 -7.15975344e-01
-9.00239408e-01 1.15947449e+00 -4.77535576e-01 -3.37142944e-01
6.65836513e-01 -5.93863010e-01 1.95735562e+00 -1.00059485e+00
4.25265491e-01 1.06823778e+00 4.96835917e-01 2.27827922e-01
-1.23326993e+00 -7.62281597e-01 4.59595859e-01 4.74680364e-02
-1.34155810e+00 -2.51268130e-02 6.52807474e-01 -4.28521574e-01
9.85492647e-01 4.40856427e-01 6.52775705e-01 6.23075962e-01
-4.58241642e-01 6.30811572e-01 1.42668760e+00 -5.93662024e-01
1.11703068e-01 4.07964975e-01 5.24888694e-01 4.58801419e-01
3.02214533e-01 -3.25972855e-01 -5.74376523e-01 -1.08143173e-01
3.25667173e-01 -9.48358253e-02 5.39034652e-03 -3.92233759e-01
-1.24178886e+00 9.62579489e-01 3.78982604e-01 3.10255915e-01
-1.34028867e-01 -2.52073675e-01 1.20415688e-01 4.69872624e-01
4.06192958e-01 1.01699507e+00 -7.73745656e-01 -3.05078924e-01
-7.52782285e-01 2.85622448e-01 1.06895137e+00 8.25170755e-01
8.93032491e-01 -5.35028756e-01 2.40416378e-01 1.61814618e+00
3.87586206e-01 2.78516471e-01 7.31941760e-01 -8.84370685e-01
3.26903075e-01 6.87801242e-01 2.84987926e-01 -9.84392285e-01
-6.35547698e-01 -3.00996274e-01 -3.29142898e-01 -1.57116458e-01
3.90723556e-01 -1.06311003e-02 -6.05890810e-01 1.82034361e+00
1.72253072e-01 -4.39178675e-01 1.45316990e-02 8.04766059e-01
1.22158146e+00 4.78232533e-01 3.69861811e-01 -2.38914713e-01
1.56565964e+00 -5.30052125e-01 -3.82354289e-01 -3.06295276e-01
7.88939714e-01 -8.54821444e-01 1.40098846e+00 6.53851628e-01
-9.53681588e-01 -1.89244494e-01 -8.44834924e-01 -2.17520505e-01
-6.22847795e-01 -4.45662178e-02 7.44558156e-01 8.71464014e-01
-1.11956346e+00 5.42404950e-01 -4.22939420e-01 -5.00231802e-01
-1.36541605e-01 6.26255274e-01 -3.83785039e-01 6.57719448e-02
-1.48378336e+00 1.08412695e+00 5.15809536e-01 -4.45756018e-01
7.04912236e-03 -6.03951871e-01 -9.38458562e-01 -7.95883760e-02
3.00666124e-01 -3.00974429e-01 1.15935433e+00 -7.48519301e-01
-1.58238029e+00 1.13642180e+00 -5.47986887e-02 -1.52127549e-01
1.24465888e-02 -3.07082474e-01 -3.76560926e-01 -2.05153227e-01
2.93073565e-01 9.33653116e-01 1.04254626e-01 -1.09195411e+00
-8.48799467e-01 -1.96116641e-01 4.88814026e-01 6.94526076e-01
-9.88533258e-01 2.61286885e-01 -3.30774844e-01 -7.98511028e-01
2.69563794e-01 -1.03318608e+00 -2.25485653e-01 -2.45909736e-01
-4.84818406e-02 -7.88027644e-01 -1.22020572e-01 -5.46483040e-01
1.41120970e+00 -1.85908353e+00 1.95380226e-01 5.42080820e-01
2.27535024e-01 -1.62013113e-01 -8.91650766e-02 5.19726515e-01
1.69248302e-02 2.93134272e-01 -2.66642094e-01 -6.70449808e-02
3.71467143e-01 3.35130155e-01 -4.85156663e-02 -3.33304778e-02
-1.29045159e-01 6.38990879e-01 -1.14457655e+00 -4.58105654e-01
9.73285176e-03 8.54391754e-02 -7.08202481e-01 -1.82780445e-01
5.22870943e-02 -1.39953032e-01 -4.37943161e-01 6.20826364e-01
1.68107495e-01 -7.91703910e-02 4.30923402e-01 -3.20374593e-02
-2.76067436e-01 7.09110975e-01 -1.24893737e+00 1.36486673e+00
-9.02811110e-01 1.85358822e-01 -1.35183766e-01 -8.24072719e-01
1.08367872e+00 1.90770864e-01 5.39253414e-01 -5.89020491e-01
-1.04878850e-01 6.27818525e-01 6.41291812e-02 -6.03327788e-02
8.96278024e-01 -2.98074335e-01 -5.82260072e-01 3.67537111e-01
-1.57129820e-02 -5.24922371e-01 6.76659882e-01 1.93133593e-01
8.42742264e-01 -3.99764962e-02 7.25334108e-01 -6.45386040e-01
4.65879083e-01 4.48830016e-02 3.29195052e-01 6.34056151e-01
9.75280777e-02 4.53843862e-01 5.52592158e-01 -1.74066946e-01
-7.91331351e-01 -1.16459167e+00 -5.16890824e-01 1.55446172e+00
1.98085815e-01 -8.19816053e-01 -2.65754402e-01 -6.35362983e-01
1.88456744e-01 8.84145081e-01 -4.44181174e-01 7.75182173e-02
-3.17704022e-01 -8.11398029e-01 1.18110709e-01 6.30454957e-01
-3.17683592e-02 -1.07091105e+00 2.01930087e-02 3.31777453e-01
-2.34702751e-01 -9.72352386e-01 -2.91169494e-01 5.70764840e-01
-6.42122269e-01 -5.99142551e-01 -5.72889507e-01 -1.02845562e+00
4.27677542e-01 2.58288868e-02 1.59744966e+00 6.77233264e-02
2.34356925e-01 2.06166416e-01 -6.72914207e-01 -4.25290585e-01
1.04035132e-01 2.15946496e-01 2.40861326e-01 -4.25654024e-01
6.96419835e-01 -5.86325943e-01 -3.69721413e-01 3.00273120e-01
-6.85962200e-01 -3.20856303e-01 6.50651336e-01 6.61994278e-01
5.95620632e-01 -7.48434588e-02 5.38919866e-01 -1.22424805e+00
9.56545591e-01 -5.07177234e-01 -3.73029232e-01 -1.48044135e-02
-1.01970375e+00 2.10050523e-01 5.89848161e-01 -2.80494869e-01
-6.32142663e-01 -1.64636627e-01 -3.83668303e-01 4.22787696e-01
-2.35011056e-02 8.94832909e-01 -3.08144949e-02 -4.51763682e-02
5.97283006e-01 -1.68212146e-01 -4.49098021e-01 -4.75193411e-01
5.24874687e-01 7.09582508e-01 1.37779057e-01 -8.44009399e-01
6.81036890e-01 1.02877498e-01 -3.55730623e-01 -6.92493260e-01
-1.21153152e+00 -9.19693410e-01 -6.84147418e-01 2.39520878e-01
6.77762628e-01 -1.08752787e+00 -7.10532904e-01 -2.76429988e-02
-7.82950580e-01 1.40040442e-02 -2.10525542e-01 6.91264570e-01
-2.37183169e-01 2.77149022e-01 -5.52631497e-01 -4.42008197e-01
-7.20543489e-02 -8.76013041e-01 9.61199343e-01 3.25769261e-02
-8.26662242e-01 -1.18569767e+00 1.21103339e-01 4.67930555e-01
2.06523716e-01 -6.13291375e-02 1.18089449e+00 -9.75277066e-01
7.96405450e-02 -1.33100957e-01 -1.98439091e-01 2.52418250e-01
2.83719271e-01 -1.13328785e-01 -5.24064183e-01 1.33791836e-02
-4.27893251e-01 -4.68396783e-01 9.53690648e-01 8.32575262e-02
8.89562726e-01 -5.07256746e-01 -5.18734641e-02 1.46198258e-01
1.31343067e+00 -2.57253889e-02 1.54402494e-01 6.17002010e-01
7.32991695e-01 9.41625416e-01 6.59732044e-01 4.34605628e-01
8.21080208e-01 7.13321209e-01 1.75345600e-01 7.67668039e-02
2.82288492e-01 -2.84488760e-02 4.79969561e-01 1.14705789e+00
-8.72237459e-02 -2.19651192e-01 -8.91879141e-01 7.60338247e-01
-1.52633464e+00 -8.11083257e-01 -1.37153104e-01 2.33184314e+00
1.46306968e+00 2.00900197e-01 2.33193249e-01 1.13377772e-01
5.24407864e-01 1.82295427e-01 -5.48593476e-02 -8.18600297e-01
-1.88404322e-01 5.91202080e-01 5.57174921e-01 7.64443815e-01
-1.09005797e+00 1.21083891e+00 6.19205046e+00 9.01189446e-01
-7.29157805e-01 -2.35241771e-01 3.36830288e-01 -9.85328332e-02
-8.15724969e-01 5.54626472e-02 -7.73471653e-01 2.89786249e-01
6.54621542e-01 -3.49807948e-01 3.73159140e-01 7.55323887e-01
-1.27663717e-01 1.35273822e-02 -9.04105723e-01 9.33117390e-01
-1.19471867e-02 -6.36480629e-01 -3.40758637e-02 -4.07977477e-02
8.80783379e-01 -6.82088286e-02 1.19637266e-01 5.00098586e-01
1.03706110e+00 -9.86937702e-01 5.56787848e-01 8.79150406e-02
7.03186393e-01 -9.92014587e-01 7.86133051e-01 2.06708908e-01
-1.23169148e+00 6.46582097e-02 -3.59502226e-01 -3.74105036e-01
-8.15240964e-02 9.21945632e-01 -6.04268372e-01 3.16382557e-01
7.71414995e-01 8.28483522e-01 -7.47635424e-01 7.17416644e-01
-6.72321022e-01 6.51465118e-01 -5.70496321e-01 -7.45871544e-01
2.63528615e-01 -4.43394899e-01 2.76406795e-01 1.22684777e+00
4.11925167e-01 1.15469232e-01 5.81970513e-01 2.22532466e-01
-3.05035919e-01 1.02695417e+00 -2.46182799e-01 -4.94473763e-02
5.32565534e-01 1.54537761e+00 -6.55482113e-01 -1.81156158e-01
-6.13634109e-01 7.36323774e-01 6.26271427e-01 1.18489459e-01
-4.93626982e-01 -6.51000023e-01 6.82338476e-01 5.53025156e-02
-5.58751337e-02 -2.92493373e-01 -5.55866063e-01 -1.02797341e+00
3.47430892e-02 -6.76379859e-01 6.51518524e-01 -5.23551822e-01
-1.60203207e+00 4.90338862e-01 2.59830393e-02 -1.14132500e+00
-3.04091692e-01 -9.00316954e-01 -4.82542627e-02 7.31203198e-01
-1.36234093e+00 -8.97501349e-01 -1.08567126e-01 2.77996838e-01
6.80349469e-02 1.60191339e-02 1.03222513e+00 1.88257813e-01
-5.44040725e-02 6.55573845e-01 1.28703222e-01 1.02704041e-01
1.14829850e+00 -1.86018491e+00 7.53085464e-02 3.58670592e-01
5.85679650e-01 1.01503253e+00 7.06695974e-01 -4.32552278e-01
-7.32870460e-01 -6.92932367e-01 1.43601334e+00 -5.63412905e-01
1.11217833e+00 -2.31339484e-01 -7.53971636e-01 3.90756160e-01
2.53293395e-01 -3.19810569e-01 1.23354244e+00 9.62979436e-01
-6.32310688e-01 -1.11609541e-01 -1.00353360e+00 8.02132964e-01
1.05728853e+00 -3.94272357e-01 -8.11693013e-01 3.67768973e-01
6.95741773e-01 -2.25529566e-01 -1.16149974e+00 3.09261650e-01
7.48291314e-01 -6.49498343e-01 9.20593202e-01 -4.67444927e-01
4.80098873e-01 -1.74699843e-01 -3.96998525e-01 -1.53001535e+00
-4.96983618e-01 -3.51298809e-01 1.65526196e-01 1.27143252e+00
8.73909056e-01 -7.10449636e-01 6.62206173e-01 6.29973173e-01
-4.38692309e-02 -6.89533472e-01 -6.97197020e-01 -6.43041313e-01
4.78926331e-01 -4.81291115e-01 4.93025392e-01 1.12698913e+00
4.30223525e-01 5.17689586e-01 -5.06865233e-02 -1.01044081e-01
3.91599178e-01 3.00400376e-01 9.10078585e-02 -1.64343393e+00
-1.23679899e-01 -7.70643115e-01 -4.18612897e-01 -1.01193619e+00
3.92304182e-01 -1.26393902e+00 -9.40220430e-02 -1.72516859e+00
2.00163022e-01 -8.47316325e-01 -7.70377338e-01 7.28347063e-01
-3.62694293e-01 5.76093316e-01 7.81517550e-02 1.81941107e-01
-7.22791553e-01 3.22252482e-01 7.73747027e-01 1.19602717e-02
-3.00150990e-01 -1.32196903e-01 -1.25326860e+00 8.24914575e-01
6.72047615e-01 -5.08354962e-01 -1.75497621e-01 -5.50162494e-01
8.95345628e-01 -4.87697482e-01 -1.83605812e-02 -6.81021392e-01
1.22420684e-01 -3.83541703e-01 2.49160290e-01 -1.65780798e-01
1.64077237e-01 -7.54978240e-01 -3.24635714e-01 5.63658290e-02
-7.25502610e-01 3.31847101e-01 -4.74627316e-02 1.85100302e-01
-4.92573470e-01 -4.43890303e-01 5.26596129e-01 -1.90238282e-01
-8.81438792e-01 5.93448356e-02 -1.45154476e-01 3.80701274e-01
7.38571763e-01 5.21792322e-02 -5.07336408e-02 -4.61168736e-01
-6.89191818e-01 2.75972933e-01 5.11575699e-01 6.85296774e-01
3.54560792e-01 -1.35911369e+00 -7.58946717e-01 -2.42484316e-01
6.45914853e-01 -2.78952360e-01 -2.67331660e-01 5.62770724e-01
-3.88384879e-01 4.20352876e-01 5.12604490e-02 -1.86459690e-01
-1.21020818e+00 2.51640618e-01 -2.10226074e-01 -3.90693605e-01
-8.30620080e-02 1.04493690e+00 3.15913670e-02 -8.82083595e-01
1.08722113e-01 -3.50669205e-01 -8.85531127e-01 7.54938960e-01
2.60872930e-01 1.24626324e-01 2.00164989e-01 -6.92804873e-01
-4.37183857e-01 8.41556966e-01 -2.39030749e-01 -5.03674328e-01
1.34804690e+00 3.09688933e-02 -4.98935372e-01 7.66618550e-01
1.18109334e+00 7.50423729e-01 -4.05637950e-01 -3.19413275e-01
3.75871569e-01 -1.28859743e-01 -1.34510309e-01 -9.82969582e-01
-5.74083447e-01 4.77904677e-01 1.76301181e-01 3.71060759e-01
9.35986400e-01 1.21091090e-01 6.60666168e-01 5.40169477e-01
4.55931485e-01 -1.28164554e+00 -1.34841830e-01 8.83515835e-01
5.32601058e-01 -1.17654037e+00 2.37141669e-01 -5.42122543e-01
-8.04362357e-01 9.47372019e-01 4.36208308e-01 -1.00351825e-01
5.70294261e-01 8.09388831e-02 1.19006872e-01 3.42368223e-02
-6.71346903e-01 -5.24725795e-01 5.83730400e-01 3.01032692e-01
1.16283441e+00 3.36641401e-01 -1.04846931e+00 9.03695822e-01
-7.54432023e-01 -5.69155931e-01 4.86833304e-01 7.29151726e-01
-7.14167953e-01 -1.64729965e+00 -5.44157028e-02 7.94782817e-01
-5.63858271e-01 -7.93606877e-01 -6.58804178e-01 6.24651372e-01
2.68228427e-02 1.09370160e+00 -2.51060706e-02 -3.27406794e-01
3.78136873e-01 -8.34175274e-02 2.53581554e-01 -9.83171582e-01
-8.47001195e-01 -1.51394859e-01 3.74777704e-01 -2.46786684e-01
-3.94648463e-01 -7.72880495e-01 -1.40297949e+00 -6.54126033e-02
-1.98677614e-01 4.68533248e-01 5.56918800e-01 9.49849665e-01
-1.48922890e-01 1.89452693e-01 4.99326229e-01 -2.02642351e-01
-5.13057649e-01 -8.23344290e-01 -6.58433795e-01 5.27839124e-01
-1.04878724e-01 -6.55600190e-01 -4.65341538e-01 -3.40203494e-01] | [10.324832916259766, 9.193887710571289] |
f8a3f840-633f-4afd-8451-f27d0f0da64e | towards-stable-co-saliency-detection-and | 2209.12138 | null | https://arxiv.org/abs/2209.12138v2 | https://arxiv.org/pdf/2209.12138v2.pdf | Towards Stable Co-saliency Detection and Object Co-segmentation | In this paper, we present a novel model for simultaneous stable co-saliency detection (CoSOD) and object co-segmentation (CoSEG). To detect co-saliency (segmentation) accurately, the core problem is to well model inter-image relations between an image group. Some methods design sophisticated modules, such as recurrent neural network (RNN), to address this problem. However, order-sensitive problem is the major drawback of RNN, which heavily affects the stability of proposed CoSOD (CoSEG) model. In this paper, inspired by RNN-based model, we first propose a multi-path stable recurrent unit (MSRU), containing dummy orders mechanisms (DOM) and recurrent unit (RU). Our proposed MSRU not only helps CoSOD (CoSEG) model captures robust inter-image relations, but also reduces order-sensitivity, resulting in a more stable inference and training process. { Moreover, we design a cross-order contrastive loss (COCL) that can further address order-sensitive problem by pulling close the feature embedding generated from different input orders.} We validate our model on five widely used CoSOD datasets (CoCA, CoSOD3k, Cosal2015, iCoseg and MSRC), and three widely used datasets (Internet, iCoseg and PASCAL-VOC) for object co-segmentation, the performance demonstrates the superiority of the proposed approach as compared to the state-of-the-art (SOTA) methods. | ['Shouhong Ding', 'Mofei Song', 'Senyun Kuang', 'Lv Tang', 'Bo Li'] | 2022-09-25 | null | null | null | null | ['saliency-detection'] | ['computer-vision'] | [ 1.74397245e-01 -1.64396197e-01 -6.59395978e-02 -7.17618987e-02
-4.11719441e-01 -3.07888210e-01 4.47047323e-01 -1.82688087e-01
-3.00213426e-01 3.73251557e-01 1.20775327e-01 -1.02360792e-01
-1.16509549e-01 -4.72408444e-01 -8.98872018e-01 -4.11388516e-01
2.88676023e-01 5.46112061e-02 9.45745707e-01 -2.63352871e-01
3.93991649e-01 4.70421404e-01 -1.83833790e+00 1.87868178e-01
1.23860800e+00 9.90067780e-01 6.74389541e-01 3.94644022e-01
-8.11373256e-03 6.54971778e-01 -4.57220912e-01 -2.74813563e-01
3.23724262e-02 -4.13710684e-01 -8.13635588e-01 3.00176535e-02
5.10777414e-01 8.56332779e-02 -2.30417728e-01 1.23310101e+00
5.61329067e-01 -1.56856608e-02 7.08509743e-01 -1.36004162e+00
-5.96505344e-01 6.17809474e-01 -8.04754257e-01 3.62754434e-01
1.46404179e-02 -1.06738238e-02 1.06830466e+00 -9.93587136e-01
6.21671975e-01 1.45072985e+00 5.29175818e-01 4.17654485e-01
-7.62732804e-01 -5.81279516e-01 2.97880292e-01 6.13799870e-01
-1.25863361e+00 -4.17788029e-02 8.22462320e-01 -1.17962465e-01
7.42122769e-01 5.29586613e-01 6.91587746e-01 7.44588554e-01
1.64776016e-02 1.39065278e+00 1.10595095e+00 -3.63735884e-01
-1.89538032e-01 1.93096429e-01 3.90393019e-01 5.93608558e-01
1.76893651e-01 -9.60342884e-02 -2.99694091e-01 4.87809867e-01
8.77805412e-01 3.07824984e-02 -2.12701157e-01 -2.22942337e-01
-1.17668164e+00 4.48151767e-01 8.61861527e-01 4.13267106e-01
-7.38899931e-02 -6.69974927e-03 3.50262225e-01 -5.90918679e-03
2.11990312e-01 2.45896906e-01 -3.03004146e-01 6.61195442e-02
-9.06503379e-01 2.44105458e-01 3.54415953e-01 1.09765112e+00
6.54520154e-01 -8.43172744e-02 -3.80972952e-01 8.75591457e-01
3.88775349e-01 4.03252840e-01 6.72019362e-01 -6.85150027e-01
4.25090581e-01 8.92863691e-01 -7.86528438e-02 -1.15907156e+00
-3.98133934e-01 -7.27307498e-01 -8.25368404e-01 -1.53102800e-01
5.89222871e-02 3.44716221e-01 -1.03298914e+00 1.67613399e+00
3.37397188e-01 3.50323200e-01 -1.22534178e-01 1.08255756e+00
1.20216668e+00 7.12030292e-01 -3.37282307e-02 -2.65912771e-01
1.33969104e+00 -1.39810228e+00 -6.83443785e-01 -3.27336550e-01
3.28870744e-01 -1.03999901e+00 1.21868002e+00 1.19093858e-01
-1.02373123e+00 -8.90170097e-01 -1.10910249e+00 -2.22090885e-01
-5.40541112e-01 3.46461326e-01 4.92250472e-01 5.98438159e-02
-9.88898754e-01 2.90175050e-01 -5.40755630e-01 -3.81226331e-01
3.41536045e-01 3.58432174e-01 5.99400327e-02 1.51591972e-01
-1.33662665e+00 9.16868627e-01 5.06018102e-01 4.08374608e-01
-7.43313193e-01 -4.46530670e-01 -7.59745717e-01 -1.20471939e-01
6.08131588e-01 -4.20301199e-01 8.60858142e-01 -1.09259701e+00
-1.20430005e+00 8.58538389e-01 -2.02742562e-01 -4.91274506e-01
5.11590004e-01 -4.36036646e-01 -4.05334949e-01 1.72203466e-01
2.60905713e-01 1.01024544e+00 7.52793908e-01 -1.39733660e+00
-6.34002209e-01 -2.89402366e-01 -2.13959944e-02 5.20054340e-01
-1.55244082e-01 3.91490489e-01 -7.64305174e-01 -7.17901170e-01
3.62260580e-01 -9.07233119e-01 1.07581213e-01 -1.91320047e-01
-6.79057479e-01 -5.94429612e-01 1.13549185e+00 -6.19998574e-01
1.47970974e+00 -2.30656338e+00 2.64915913e-01 -1.33494794e-01
4.04509455e-02 7.15049207e-01 -9.37548280e-02 -8.16453174e-02
-1.70601979e-01 7.54527748e-02 -3.05240780e-01 -4.30868030e-01
-1.55263647e-01 1.70015112e-01 -1.72428951e-01 2.20910162e-01
2.62659580e-01 1.19221973e+00 -7.87423015e-01 -8.73060405e-01
2.83014148e-01 1.52419150e-01 -2.41019607e-01 2.05747515e-01
-2.73636311e-01 -4.06866670e-02 -2.97839582e-01 8.02154362e-01
8.26733291e-01 -3.20009381e-01 -2.59008020e-01 -6.55781329e-01
-3.03219974e-01 -8.97361338e-03 -1.41883278e+00 1.30903947e+00
-2.32933566e-01 3.83231789e-01 -2.77992308e-01 -9.33377624e-01
9.37909424e-01 -5.44270948e-02 1.34024709e-01 -8.26194644e-01
2.25960523e-01 3.69792968e-01 -2.00509608e-01 -5.81283331e-01
6.10920191e-01 4.27061409e-01 2.24115819e-01 2.27821413e-02
4.89744246e-02 1.37932509e-01 1.88448682e-01 3.05911452e-01
5.91244817e-01 2.60558099e-01 1.37788221e-01 -3.54099333e-01
7.73575366e-01 -1.48246855e-01 7.93721259e-01 4.42430913e-01
-4.55391794e-01 9.53617156e-01 5.54418981e-01 -4.60861437e-03
-8.57242346e-01 -1.02568626e+00 -1.75429001e-01 9.20159876e-01
9.87404287e-01 -1.86242834e-01 -7.21332967e-01 -8.94824922e-01
-1.49163306e-01 5.17634153e-01 -6.54279351e-01 -1.28257409e-01
-7.42319405e-01 -7.68731356e-01 5.32437861e-01 6.07998013e-01
1.03559303e+00 -1.37907863e+00 -4.45710868e-01 8.23240206e-02
-4.21746075e-01 -1.18671763e+00 -8.67972195e-01 -1.48277342e-01
-7.38341093e-01 -1.06972003e+00 -9.07672703e-01 -1.17368639e+00
5.28870404e-01 5.59495211e-01 9.84202623e-01 5.08322529e-02
-2.60542989e-01 -1.33762538e-01 -2.38684937e-01 -2.70926148e-01
5.38683645e-02 7.25064054e-02 -1.34804919e-01 2.29581684e-01
2.50621170e-01 -3.57494056e-01 -6.28634751e-01 7.90425539e-01
-1.00203240e+00 3.71556252e-01 7.49892533e-01 6.85639441e-01
7.50920296e-01 -2.23923430e-01 7.04153836e-01 -7.82237768e-01
3.24789882e-01 -2.36798719e-01 -6.10582948e-01 5.33580124e-01
-8.06998491e-01 -1.86154321e-01 5.21592617e-01 -3.19402516e-01
-1.02709579e+00 -1.63528711e-01 1.22543052e-01 -7.74338126e-01
1.26261353e-01 3.04595262e-01 -4.42379534e-01 6.78380281e-02
3.91136199e-01 4.80056465e-01 -1.20301612e-01 -4.45827663e-01
2.46103242e-01 7.26112604e-01 7.96402037e-01 -1.10438913e-01
5.20345390e-01 2.19095230e-01 -4.73338328e-02 -6.19274139e-01
-9.45831776e-01 -6.00519478e-01 -5.78473687e-01 -2.40347952e-01
9.14605677e-01 -9.33616400e-01 -5.16081989e-01 7.77367294e-01
-1.33818543e+00 1.75756756e-02 -1.12928458e-01 3.40914726e-01
-2.80803174e-01 3.86028826e-01 -5.35111248e-01 -6.89372897e-01
-4.72436279e-01 -1.27150285e+00 1.11205542e+00 6.59899294e-01
3.00317019e-01 -7.91575789e-01 -3.19213450e-01 4.87761259e-01
2.71962166e-01 2.68648386e-01 5.47858298e-01 -4.25538510e-01
-9.37469780e-01 2.97330767e-01 -8.09427083e-01 5.61465740e-01
8.76014829e-02 1.29423767e-01 -8.90305519e-01 -8.38726014e-02
-1.83869645e-01 -2.42062151e-01 1.02182400e+00 5.38323283e-01
1.10571444e+00 -1.63805366e-01 -3.48620713e-01 5.49816728e-01
1.42141175e+00 5.21611013e-02 9.23074841e-01 4.05761808e-01
9.56747770e-01 4.34554011e-01 1.00886357e+00 -1.46113887e-01
6.23441577e-01 7.35626042e-01 4.42261130e-01 -3.43113363e-01
-4.67381388e-01 -3.85307044e-01 4.16323543e-01 1.11853325e+00
-6.58392534e-02 -2.14811265e-01 -4.36890543e-01 7.75920630e-01
-2.19935918e+00 -6.62894666e-01 -4.50042933e-01 1.76784253e+00
8.27340364e-01 3.34496617e-01 2.58329898e-01 2.08183557e-01
1.00350642e+00 3.64452660e-01 -5.14036715e-01 -2.29423001e-01
-6.33464575e-01 -1.34685457e-01 3.70413870e-01 2.99855977e-01
-1.17659771e+00 1.21326578e+00 5.12321854e+00 1.35030818e+00
-1.17764115e+00 1.78903461e-01 5.86352646e-01 3.30233544e-01
-3.25134873e-01 7.50355306e-04 -1.03383589e+00 6.97178423e-01
2.63555110e-01 1.84514467e-02 2.34152585e-01 8.46533835e-01
1.07852079e-01 -3.33569735e-01 -7.87278533e-01 1.02115750e+00
3.31399858e-01 -1.19835401e+00 1.66250482e-01 -3.59710693e-01
7.27652788e-01 -3.84938717e-02 2.14092374e-01 2.34757811e-01
-1.71693906e-01 -8.28191578e-01 7.71888793e-01 3.58141512e-01
7.88702130e-01 -5.89177608e-01 8.91738951e-01 1.88908055e-01
-1.42772746e+00 1.12667471e-01 -5.05515754e-01 3.80324513e-01
-5.00905998e-02 6.36306047e-01 -5.49139738e-01 8.32491815e-01
9.02478635e-01 1.10076785e+00 -1.03504288e+00 1.09749734e+00
-3.14676672e-01 5.33989072e-01 -3.41320634e-01 -3.07442874e-01
3.72188061e-01 -3.20124850e-02 7.91832805e-01 1.22148788e+00
-1.02481216e-01 -1.39442787e-01 -8.12343508e-02 1.15410984e+00
-1.07877150e-01 7.35226199e-02 -1.54658556e-01 2.46142477e-01
4.04516906e-01 1.40933239e+00 -9.94277775e-01 -3.18076104e-01
-1.00031666e-01 1.04598725e+00 1.49561733e-01 2.64611781e-01
-1.02926242e+00 -4.58298087e-01 8.10159221e-02 3.73379290e-02
4.63273972e-01 9.49996337e-02 -2.91678637e-01 -1.08712673e+00
3.51826310e-01 -8.09955537e-01 3.88275355e-01 -9.34716821e-01
-1.06089306e+00 6.67689979e-01 9.62223336e-02 -1.38802266e+00
2.26832733e-01 -2.35218912e-01 -7.00561106e-01 7.83247232e-01
-1.76162469e+00 -1.47820282e+00 -3.97702724e-01 5.76488853e-01
7.81337619e-01 5.51848076e-02 5.07448427e-02 5.33747196e-01
-7.86839008e-01 7.10159838e-01 -2.75470734e-01 -7.28387088e-02
5.25258541e-01 -1.14138162e+00 2.03840077e-01 1.06025577e+00
1.44475788e-01 4.68868881e-01 5.20413101e-01 -6.75438225e-01
-9.92889106e-01 -1.31295323e+00 8.84527862e-01 -2.33532533e-01
3.17325145e-01 -2.58468926e-01 -9.89719868e-01 5.54208457e-01
7.61655346e-02 9.67263952e-02 -1.88497216e-01 -2.15771556e-01
-1.40909702e-01 -3.26513618e-01 -9.35289025e-01 6.79693341e-01
1.25187564e+00 -4.11339730e-01 -7.81551600e-01 1.55799344e-01
1.18208289e+00 -3.92299682e-01 -2.79715806e-01 8.57863903e-01
3.86802852e-01 -1.10759127e+00 9.60224748e-01 -6.17623590e-02
5.29479980e-01 -8.60351562e-01 4.85951174e-03 -9.41986382e-01
-3.60715747e-01 -5.72036922e-01 -1.23239622e-01 1.46962333e+00
3.19470644e-01 -3.77441972e-01 5.48260391e-01 -4.00781631e-03
-3.06593508e-01 -1.04512274e+00 -7.82853723e-01 -7.52640426e-01
-3.67553622e-01 -1.55178934e-01 5.44975340e-01 6.95650280e-01
-4.56140041e-01 4.05865222e-01 -2.67366886e-01 2.07168773e-01
5.27061284e-01 1.86807036e-01 5.28150499e-01 -8.61794829e-01
-4.33200449e-02 -4.48690504e-01 -5.27938366e-01 -1.32934964e+00
-2.32648715e-01 -6.37125254e-01 -3.62586491e-02 -1.50090158e+00
3.12951714e-01 -5.07893741e-01 -5.29226422e-01 2.92713732e-01
-5.31815648e-01 3.62720996e-01 4.08611983e-01 3.17254871e-01
-1.05389917e+00 6.90798342e-01 1.59980905e+00 -5.74580021e-02
-2.16929585e-01 8.74673761e-03 -5.96326768e-01 7.00041711e-01
4.63559031e-01 -1.92313373e-01 -3.45559418e-01 -3.16822052e-01
2.38931421e-02 -3.31662804e-01 5.17810047e-01 -1.06401157e+00
4.12144005e-01 4.77715582e-02 2.20892698e-01 -1.21534264e+00
7.00998381e-02 -4.39434648e-01 -1.02059066e-01 4.07281101e-01
-1.94782421e-01 1.40451506e-01 1.97968081e-01 3.80774081e-01
-4.59210068e-01 -1.95311710e-01 9.89272118e-01 -2.51880269e-02
-9.28303599e-01 1.23461768e-01 2.26565257e-01 2.34276593e-01
9.42753792e-01 -3.74035418e-01 -3.84110540e-01 -1.52362799e-02
-4.39067334e-01 5.50436199e-01 2.20050082e-01 6.61114693e-01
8.53759170e-01 -1.29138243e+00 -4.80908513e-01 1.79090738e-01
9.51413736e-02 3.74556124e-01 2.61945248e-01 9.96666431e-01
-6.03163898e-01 2.68058360e-01 -1.06092796e-01 -8.86210561e-01
-1.24692786e+00 4.52081352e-01 3.39934260e-01 -2.45115206e-01
-3.44836980e-01 1.05891407e+00 3.16324890e-01 -4.25244272e-01
3.88701051e-01 -4.65736419e-01 -6.05976164e-01 1.01380304e-01
2.04183057e-01 4.46647912e-01 -4.80928533e-02 -7.30141580e-01
-4.26300436e-01 8.45014215e-01 -3.07164401e-01 1.70853540e-01
1.01125038e+00 -2.19840080e-01 -3.32356691e-01 4.81894523e-01
1.21827579e+00 -3.09389591e-01 -1.06861997e+00 -2.80650884e-01
-8.31896216e-02 -2.35834897e-01 -8.83901566e-02 -7.45621085e-01
-1.09283686e+00 8.34696412e-01 8.90200138e-01 1.43813595e-01
1.13152134e+00 1.24154404e-01 1.05137610e+00 -5.21047786e-02
7.50364736e-02 -1.36881876e+00 3.72833490e-01 4.53494400e-01
9.86631513e-01 -1.17385137e+00 -8.98420140e-02 -7.19050348e-01
-7.57133543e-01 8.17696035e-01 1.00812161e+00 -2.29320467e-01
6.55778587e-01 -1.90027013e-01 -1.18059017e-01 -2.38320023e-01
-3.55871499e-01 -4.34477329e-01 5.21304369e-01 1.34301543e-01
7.05665350e-02 -5.25393412e-02 -6.04386866e-01 5.76665103e-01
-1.71141148e-01 1.29771084e-01 3.96637142e-01 7.54455447e-01
-4.68050033e-01 -7.09794998e-01 -2.50798017e-01 3.81112307e-01
-3.32661122e-01 -1.29116192e-01 -2.95676768e-01 6.79895878e-01
5.32043159e-01 8.37636769e-01 1.65660068e-01 -5.20970345e-01
3.72395307e-01 -5.05906940e-01 1.91562444e-01 -4.06182259e-01
-5.47302842e-01 2.88858593e-01 -2.00006768e-01 -5.12699783e-01
-7.18699336e-01 -5.53920507e-01 -1.28010058e+00 1.27348512e-01
-7.85475314e-01 2.08666790e-02 4.56410557e-01 9.41079319e-01
3.92893881e-01 8.26333463e-01 8.32952559e-01 -5.73135972e-01
-2.37471983e-01 -1.02631962e+00 -5.54870844e-01 5.62301278e-01
1.22905783e-01 -8.67318511e-01 -4.12708431e-01 -1.32644460e-01] | [9.741229057312012, -0.2849617302417755] |
c7ce82fe-fda3-45b6-bb79-1e7faa83272f | hyperspectral-image-super-resolution-via-dual | 2304.04589 | null | https://arxiv.org/abs/2304.04589v8 | https://arxiv.org/pdf/2304.04589v8.pdf | Hyperspectral Image Super-Resolution via Dual-domain Network Based on Hybrid Convolution | Since the number of incident energies is limited, it is difficult to directly acquire hyperspectral images (HSI) with high spatial resolution. Considering the high dimensionality and correlation of HSI, super-resolution (SR) of HSI remains a challenge in the absence of auxiliary high-resolution images. Furthermore, it is very important to extract the spatial features effectively and make full use of the spectral information. This paper proposes a novel HSI super-resolution algorithm, termed dual-domain network based on hybrid convolution (SRDNet). Specifically, a dual-domain network is designed to fully exploit the spatial-spectral and frequency information among the hyper-spectral data. To capture inter-spectral self-similarity, a self-attention learning mechanism (HSL) is devised in the spatial domain. Meanwhile the pyramid structure is applied to increase the acceptance field of attention, which further reinforces the feature representation ability of the network. Moreover, to further improve the perceptual quality of HSI, a frequency loss(HFL) is introduced to optimize the model in the frequency domain. The dynamic weighting mechanism drives the network to gradually refine the generated frequency and excessive smoothing caused by spatial loss. Finally, In order to better fully obtain the mapping relationship between high-resolution space and low-resolution space, a hybrid module of 2D and 3D units with progressive upsampling strategy is utilized in our method. Experiments on a widely used benchmark dataset illustrate that the proposed SRDNet method enhances the texture information of HSI and is superior to state-of-the-art methods. | ['Yuan Liyin', 'Qian Chen', 'Xiubao Sui', 'Chuncheng Zhang', 'YuAn Liu', 'Tingting Liu'] | 2023-04-10 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 6.52153730e-01 -4.71588969e-01 1.30894005e-01 -1.75150767e-01
-6.34268999e-01 -9.60262939e-02 2.01862678e-01 -2.77116001e-01
-1.51446477e-01 8.19272757e-01 2.41826385e-01 2.32136071e-01
-6.43232763e-01 -1.08476019e+00 -3.85899454e-01 -1.12954712e+00
1.01245426e-01 -5.65211773e-01 1.79691806e-01 -4.12149161e-01
3.63960378e-02 5.40402174e-01 -1.72571015e+00 1.52900115e-01
1.59426856e+00 1.19622171e+00 8.32493663e-01 2.32727692e-01
3.50396261e-02 3.75521481e-01 -1.62679017e-01 3.57866287e-01
3.11982036e-01 -4.53198642e-01 -4.53812838e-01 2.28051856e-01
1.99624583e-01 -4.03201073e-01 -4.76661444e-01 1.50470543e+00
6.25769377e-01 5.68976998e-01 3.60186815e-01 -5.58543563e-01
-1.02812791e+00 2.84247845e-01 -1.04371643e+00 5.14493048e-01
2.64745224e-02 2.06874102e-01 7.24035561e-01 -9.10787106e-01
1.26291811e-01 9.80422080e-01 3.87494206e-01 -6.05770096e-04
-1.22596478e+00 -7.42453933e-01 8.76625031e-02 3.73139620e-01
-1.61902785e+00 -1.56053409e-01 1.15625215e+00 -8.02919120e-02
4.36645806e-01 1.85982093e-01 5.32422245e-01 6.13499284e-01
-1.25122637e-01 2.09422052e-01 1.23630214e+00 -2.92944372e-01
-1.11560769e-01 1.04615251e-02 -1.35808781e-01 3.70702028e-01
1.34261042e-01 4.27432954e-01 -4.41298664e-01 3.72016579e-01
1.23117495e+00 1.18064828e-01 -7.81169176e-01 5.08465990e-02
-9.54092979e-01 5.81548214e-01 1.04608893e+00 6.97256267e-01
-7.56804824e-01 -5.27257621e-01 7.90060908e-02 -1.99149817e-01
5.42905033e-01 4.90648031e-01 -4.41114530e-02 4.27224696e-01
-9.28456545e-01 -3.03841919e-01 -1.13752209e-01 5.74617565e-01
9.05098617e-01 4.00269896e-01 -3.04525703e-01 1.14337099e+00
-2.32780240e-02 3.19726855e-01 4.80930835e-01 -8.93450916e-01
2.41608202e-01 4.91952896e-01 7.61429593e-02 -1.03887212e+00
-3.15792739e-01 -8.53424191e-01 -1.52191007e+00 2.25716814e-01
-1.03662848e-01 5.85414283e-02 -6.74063981e-01 1.62092507e+00
3.13882947e-01 4.05413747e-01 2.59028226e-01 1.44188583e+00
8.31621706e-01 9.49937403e-01 8.41419995e-02 -5.52085519e-01
1.30464613e+00 -7.42929399e-01 -7.06585944e-01 -1.23967648e-01
-2.25730702e-01 -5.50589979e-01 1.12089360e+00 7.01398924e-02
-1.05833471e+00 -1.09009552e+00 -1.22227621e+00 -1.46952555e-01
-2.33122304e-01 4.48054641e-01 5.07665455e-01 7.00546801e-02
-6.41413152e-01 6.82765841e-01 -6.07541203e-01 -9.27837715e-02
6.47299767e-01 1.15457058e-01 -1.41299531e-01 -4.00647283e-01
-1.48396325e+00 5.23741364e-01 7.23992944e-01 3.31415951e-01
-2.93862760e-01 -9.46941733e-01 -7.69797504e-01 2.56781906e-01
2.84229577e-01 -3.21018845e-01 3.73269051e-01 -8.93508792e-01
-1.35561478e+00 3.27145070e-01 -1.18761271e-01 4.58415262e-02
1.29184186e-01 1.25298008e-01 -8.86005998e-01 6.14471316e-01
6.00491799e-02 3.10056418e-01 1.08525133e+00 -1.17386782e+00
-7.80284286e-01 -5.26149929e-01 -2.19580811e-02 5.84405005e-01
-7.56359279e-01 -3.81628454e-01 -1.34607047e-01 -6.50960684e-01
2.70506889e-01 -2.82743245e-01 -1.31970465e-01 -1.36058539e-01
-2.72975594e-01 2.55557001e-01 8.70056152e-01 -9.04422104e-01
1.25729966e+00 -2.51353264e+00 2.12675571e-01 -9.41970944e-03
2.44422659e-01 5.00337481e-01 -2.73531556e-01 -1.09902978e-01
-3.75877172e-01 -9.31955054e-02 -4.84254152e-01 3.93782675e-01
-3.82347167e-01 -1.51328921e-01 -1.13771252e-01 3.56896251e-01
4.34760898e-01 5.93900502e-01 -8.65430772e-01 -3.49584430e-01
4.57654983e-01 9.68791544e-01 -2.41160184e-01 1.79258898e-01
1.11541674e-01 7.04723895e-01 -6.44384086e-01 5.05437315e-01
1.24144566e+00 -4.33756590e-01 -1.93264201e-01 -9.67285275e-01
-7.55108595e-01 -1.54935405e-01 -1.16711700e+00 1.61099780e+00
-5.11932373e-01 1.60183296e-01 3.43450516e-01 -9.07329321e-01
1.06277835e+00 1.12819709e-01 5.35883963e-01 -1.02111351e+00
-9.34322998e-02 9.79589000e-02 -2.45148540e-01 -5.38910210e-01
4.59859848e-01 -3.05545479e-01 5.51303148e-01 -4.48880792e-02
-1.76226258e-01 3.84158641e-02 6.31108135e-03 -3.50539237e-01
4.47561264e-01 1.64994262e-02 3.01415980e-01 -2.45590240e-01
9.68730927e-01 -9.14240777e-02 6.02901638e-01 3.43741298e-01
-1.25486227e-02 4.45532709e-01 -1.02323517e-01 -1.39923319e-01
-1.16832471e+00 -8.69472802e-01 -5.03288150e-01 9.32187855e-01
4.74459797e-01 2.69659966e-01 -4.11762923e-01 -5.82449399e-02
-1.71466738e-01 6.08507812e-01 -4.84958082e-01 -3.95131081e-01
-3.45607489e-01 -1.10927081e+00 2.54260805e-02 4.61521268e-01
1.38427341e+00 -8.39676201e-01 -5.11263609e-01 2.05207124e-01
-2.91752785e-01 -1.03499699e+00 -4.28840727e-01 2.11232249e-02
-7.46198297e-01 -8.81757855e-01 -7.37653613e-01 -6.66741252e-01
4.62066859e-01 8.53794038e-01 5.60338080e-01 -2.02394277e-01
-4.98528510e-01 -1.63191721e-01 -3.29516679e-01 -3.59096713e-02
3.13620418e-02 1.02270678e-01 -3.12859178e-01 2.83305198e-01
2.61473626e-01 -8.13184261e-01 -8.67281735e-01 6.46281838e-02
-1.21383679e+00 3.18318129e-01 9.51758385e-01 1.12212026e+00
7.19286561e-01 1.03400898e+00 6.79152012e-01 -4.05074328e-01
4.89701897e-01 -4.15442377e-01 -6.86685503e-01 1.01948753e-01
-4.53139067e-01 -1.79492816e-01 8.89931560e-01 -5.01700222e-01
-1.71456480e+00 -1.12337299e-01 1.70718282e-02 -4.69128847e-01
-1.55567959e-01 6.43562257e-01 -3.78319234e-01 -3.40062439e-01
5.33923268e-01 6.83043063e-01 -2.45009921e-02 -5.01836181e-01
1.27107650e-01 6.08880460e-01 6.84086442e-01 -2.51497179e-01
1.03925502e+00 5.67361772e-01 2.17389405e-01 -1.08011794e+00
-1.06691754e+00 -3.53415191e-01 -5.44584215e-01 -3.35028730e-02
9.33088660e-01 -1.13015544e+00 -7.01056957e-01 4.67382431e-01
-6.84071243e-01 -1.23750739e-01 -2.25240171e-01 6.69487417e-01
-8.93352032e-02 3.45399350e-01 -6.58161402e-01 -7.68242896e-01
-3.50677341e-01 -9.06780779e-01 7.89677739e-01 8.21305931e-01
6.72571540e-01 -6.77094221e-01 -4.59709406e-01 2.77777523e-01
8.77069294e-01 1.77851096e-01 1.01775420e+00 2.05812350e-01
-7.55161285e-01 2.76734233e-01 -9.94315267e-01 5.01864731e-01
5.43087900e-01 -2.79766500e-01 -9.65871096e-01 -2.98249036e-01
2.14603648e-01 -1.62841409e-01 9.60670114e-01 6.38177514e-01
1.59327710e+00 -1.19048208e-01 5.37755415e-02 9.31529164e-01
1.95361125e+00 1.42281175e-01 6.30601764e-01 3.11465591e-01
7.86659122e-01 6.35835826e-01 6.76730931e-01 6.23673499e-01
8.74084234e-02 4.31294113e-01 4.54989374e-01 -6.30817115e-01
-1.37798831e-01 -9.92408395e-02 -1.75267130e-01 3.64302963e-01
-3.40163261e-01 1.41561791e-01 -3.30181539e-01 2.54192024e-01
-1.33184445e+00 -1.08903933e+00 -3.47458385e-03 2.14636064e+00
8.64352822e-01 -2.69708961e-01 -1.67647257e-01 2.41650000e-01
1.03510427e+00 4.82938856e-01 -8.97713125e-01 3.92547369e-01
-5.24960399e-01 2.57042915e-01 5.03129661e-01 3.62407297e-01
-1.15096045e+00 6.32934213e-01 5.00317860e+00 1.05241930e+00
-1.47030735e+00 5.09689003e-03 5.76641083e-01 8.90599936e-02
-2.46739641e-01 -4.24345165e-01 -4.93900418e-01 5.17481923e-01
4.45564032e-01 -2.76605070e-01 9.02750134e-01 4.46433574e-01
4.16324943e-01 -1.16374455e-02 -3.27708423e-01 9.57472086e-01
-2.01841742e-01 -1.15258753e+00 -9.99555066e-02 -1.04208291e-01
7.07883179e-01 -2.53538281e-01 4.18292642e-01 1.03967555e-01
-1.16425805e-01 -9.77479219e-01 8.67268220e-02 6.13025188e-01
1.41642225e+00 -1.05850387e+00 5.61831594e-01 2.67257929e-01
-1.68685436e+00 -5.03143728e-01 -7.80850768e-01 2.53812283e-01
-6.37576031e-03 8.35226655e-01 -1.95250705e-01 9.69922960e-01
7.27172136e-01 9.27234888e-01 -4.53400433e-01 7.65454829e-01
-1.44522190e-01 3.02013844e-01 -8.72232243e-02 4.55419242e-01
3.13040406e-01 -6.94369733e-01 5.45829833e-01 8.76028955e-01
5.81948578e-01 8.18743646e-01 1.55353606e-01 1.20102775e+00
2.80390978e-02 -1.00540407e-01 -2.44951412e-01 -1.54435122e-02
4.98374492e-01 1.45540583e+00 -3.30370456e-01 -8.86057094e-02
-4.55777466e-01 8.91228855e-01 -1.55848349e-02 6.99960649e-01
-6.64524078e-01 -5.31357646e-01 4.61365044e-01 2.15988979e-02
4.15919632e-01 -4.47841100e-02 -1.13304533e-01 -1.04805005e+00
-4.04558070e-02 -7.91948855e-01 2.89192885e-01 -1.13159835e+00
-1.35163403e+00 7.89288104e-01 -3.04479420e-01 -1.43344653e+00
3.24811399e-01 -2.11022258e-01 -4.80698556e-01 1.49945819e+00
-2.24977183e+00 -1.07052457e+00 -1.07356644e+00 7.85454750e-01
3.95234048e-01 9.89471078e-02 5.78339159e-01 4.67586994e-01
-7.25297689e-01 1.30059958e-01 2.69370317e-01 -1.91523269e-01
5.08692026e-01 -9.38571632e-01 -3.75621468e-01 9.48890328e-01
-7.41891205e-01 3.77787471e-01 4.72494513e-01 -5.40028572e-01
-1.24280441e+00 -1.47870243e+00 9.96247754e-02 5.12157440e-01
4.92678344e-01 1.50309771e-01 -1.55150950e+00 6.79417551e-02
-9.52724218e-02 2.52734572e-01 5.30024469e-01 -3.32590461e-01
-2.72406697e-01 -6.14677191e-01 -1.26000249e+00 3.94776553e-01
8.20002794e-01 -7.04684377e-01 -2.31078610e-01 2.40862399e-01
9.77192938e-01 -2.55548447e-01 -1.10282683e+00 7.59730697e-01
2.94364184e-01 -1.08503008e+00 1.21418059e+00 -2.07827426e-02
7.14504182e-01 -6.17381215e-01 -2.62300313e-01 -1.42295432e+00
-9.22665477e-01 -1.14196740e-01 8.15804303e-02 1.29916918e+00
-2.69503109e-02 -6.81265235e-01 4.66066718e-01 1.92269400e-01
-1.65384695e-01 -5.22367299e-01 -6.72835231e-01 -5.31715572e-01
-2.97525197e-01 1.50201961e-01 8.46251845e-01 1.08375776e+00
-3.03998709e-01 2.50275940e-01 -2.26193950e-01 7.79758751e-01
9.93454099e-01 4.63257611e-01 -3.87527868e-02 -1.18961167e+00
-3.37102935e-02 -5.35229087e-01 8.83451477e-03 -8.49152207e-01
5.20975366e-02 -6.40503347e-01 -1.32508039e-01 -1.41871095e+00
1.68709904e-01 -3.09524298e-01 -6.61445737e-01 1.16025336e-01
-4.71129596e-01 3.58638853e-01 -6.83564767e-02 2.39752397e-01
-1.69969589e-01 9.51719403e-01 1.75245643e+00 -1.35038167e-01
-3.82400364e-01 -1.09837092e-01 -9.64795887e-01 4.79289144e-01
7.85124540e-01 1.12469263e-01 -5.26056588e-01 -3.63870382e-01
-2.03187510e-01 3.78655314e-01 5.28444529e-01 -1.10027027e+00
2.79432148e-01 -3.50965410e-01 8.48315060e-01 -6.45664096e-01
3.89764160e-01 -7.82901287e-01 2.31608860e-02 1.25675395e-01
-3.20774883e-01 -6.96434736e-01 1.83334798e-01 4.86025572e-01
-4.50845003e-01 1.60442114e-01 1.29152811e+00 -1.03129022e-01
-8.26673329e-01 6.22186303e-01 1.35453761e-01 -3.21611136e-01
7.30184913e-01 -2.39040449e-01 -3.18306714e-01 6.41267449e-02
-3.77901644e-01 2.79097587e-01 4.52660382e-01 9.68450606e-02
8.58583033e-01 -1.39816523e+00 -7.46433258e-01 4.15063083e-01
8.97750109e-02 1.45930082e-01 1.25707424e+00 7.40014672e-01
-1.56149149e-01 1.70071170e-01 -5.01171649e-01 -3.47514600e-01
-9.10721898e-01 6.86826110e-01 4.10742700e-01 -7.04879239e-02
-7.76183903e-01 5.82657337e-01 3.48288715e-01 -7.21929967e-02
-1.78451613e-01 5.33741862e-02 -6.32299006e-01 1.07998718e-02
9.87614989e-01 5.52690268e-01 -1.70774773e-01 -6.66533172e-01
-5.48087470e-02 7.13242888e-01 1.79467618e-01 2.09930956e-01
1.63260138e+00 -4.10039753e-01 -3.66416126e-01 1.31844297e-01
1.11260629e+00 -1.10941119e-01 -1.67999816e+00 -6.18713558e-01
-5.73113501e-01 -7.92228162e-01 7.21285403e-01 -6.21297061e-01
-1.19960451e+00 9.33244646e-01 7.48960316e-01 2.52975404e-01
1.96665478e+00 -4.61717725e-01 6.87756181e-01 -7.48948008e-02
3.13071422e-02 -1.10190153e+00 8.60228539e-02 2.77056932e-01
8.08999896e-01 -1.24971032e+00 1.24814473e-01 -6.14242136e-01
-4.66163784e-01 1.10459137e+00 7.31570899e-01 4.27031070e-02
3.77396256e-01 1.84511207e-02 -3.84570658e-01 -3.39381732e-02
-1.89779460e-01 -3.63995612e-01 3.72731417e-01 7.28659570e-01
2.86490023e-01 -9.33468491e-02 -1.06331287e-02 6.89052224e-01
1.67621717e-01 1.60244524e-01 4.10569191e-01 3.51336658e-01
-5.89917183e-01 -3.94583583e-01 -5.53012192e-01 4.56427693e-01
-2.00293019e-01 -3.04962218e-01 3.95147473e-01 4.11632091e-01
3.42509806e-01 1.00461864e+00 -8.60470161e-02 -3.35389614e-01
3.22831035e-01 -3.56193483e-01 2.42738590e-01 -3.53697091e-01
6.54386654e-02 3.78493667e-01 -4.56868708e-01 -4.76231277e-01
-6.09530449e-01 -2.55780190e-01 -1.14080119e+00 -2.15498209e-01
-3.56647164e-01 1.63093269e-01 4.38316882e-01 6.93377256e-01
2.92953759e-01 1.01471496e+00 1.01907837e+00 -9.39431369e-01
-4.17695612e-01 -1.01259971e+00 -1.24140561e+00 1.11024275e-01
5.70929766e-01 -5.93591630e-01 -5.41791320e-01 -1.00860953e-01] | [10.21865177154541, -1.8928794860839844] |
9ec9bb01-e9d3-47fe-b05a-94fccbd0e881 | code-structure-guided-transformer-for-source | 2104.09340 | null | https://arxiv.org/abs/2104.09340v2 | https://arxiv.org/pdf/2104.09340v2.pdf | Code Structure Guided Transformer for Source Code Summarization | Code summaries help developers comprehend programs and reduce their time to infer the program functionalities during software maintenance. Recent efforts resort to deep learning techniques such as sequence-to-sequence models for generating accurate code summaries, among which Transformer-based approaches have achieved promising performance. However, effectively integrating the code structure information into the Transformer is under-explored in this task domain. In this paper, we propose a novel approach named SG-Trans to incorporate code structural properties into Transformer. Specifically, we inject the local symbolic information (e.g., code tokens and statements) and global syntactic structure (e.g., data flow graph) into the self-attention module of Transformer as inductive bias. To further capture the hierarchical characteristics of code, the local information and global structure are designed to distribute in the attention heads of lower layers and high layers of Transformer. Extensive evaluation shows the superior performance of SG-Trans over the state-of-the-art approaches. Compared with the best-performing baseline, SG-Trans still improves 1.4% and 2.0% in terms of METEOR score, a metric widely used for measuring generation quality, respectively on two benchmark datasets. | ['Michael R. Lyu', 'Xin Xia', 'Lun Yiu Nie', 'Jichuan Zeng', 'Yulan He', 'Cuiyun Gao', 'Shuzheng Gao'] | 2021-04-19 | null | null | null | null | ['code-summarization'] | ['computer-code'] | [ 8.62157270e-02 2.66820770e-02 -2.64626861e-01 -3.74766022e-01
-9.57247853e-01 -3.74306351e-01 2.70629287e-01 4.39000219e-01
1.03370570e-01 3.23127806e-01 4.39800262e-01 -4.42451268e-01
3.68708342e-01 -7.05905497e-01 -1.03105009e+00 -2.84462810e-01
1.23849593e-01 -1.93218365e-01 2.36430019e-01 -6.11741096e-02
5.85350990e-01 -2.63015062e-01 -1.47033048e+00 5.07489622e-01
1.49641669e+00 6.70229495e-01 4.29229736e-01 4.33794379e-01
-4.56117362e-01 1.30415082e+00 -6.28458977e-01 -5.98929644e-01
-4.55738366e-01 -4.51434255e-01 -7.50683129e-01 -2.33338863e-01
2.86379576e-01 -3.07909876e-01 -2.45293289e-01 1.30391526e+00
2.47167751e-01 -1.87817931e-01 4.32773858e-01 -9.95257676e-01
-9.87600505e-01 1.08302820e+00 -8.34633768e-01 2.38503695e-01
3.41145605e-01 3.03057432e-01 1.22021222e+00 -7.78245389e-01
4.27918345e-01 1.05169773e+00 6.84647322e-01 2.27498814e-01
-1.01512134e+00 -5.09450555e-01 3.62284452e-01 4.09599513e-01
-1.14644611e+00 -4.31928426e-01 8.64324927e-01 -7.39413321e-01
1.22289622e+00 1.05228737e-01 3.82737279e-01 8.17136168e-01
2.98093438e-01 1.04955721e+00 4.99818385e-01 -2.02780098e-01
2.32921205e-02 6.80157309e-03 3.82710606e-01 1.07522535e+00
1.42133579e-01 -3.48866105e-01 -3.67697388e-01 -4.83684950e-02
2.37758964e-01 -1.14278654e-02 -3.27809036e-01 2.08141357e-02
-1.00559330e+00 6.50433421e-01 7.54428685e-01 3.78484309e-01
-4.29307431e-01 3.37429941e-01 8.05550158e-01 -9.02484506e-02
4.66097027e-01 4.18313354e-01 -3.64513457e-01 -4.67927784e-01
-9.51282740e-01 3.02551061e-01 4.78222400e-01 1.36766362e+00
7.73584306e-01 2.38435045e-01 -7.50430226e-01 6.61189914e-01
4.40757781e-01 2.03852519e-01 6.86162233e-01 -5.65706909e-01
1.07906234e+00 1.17406988e+00 -9.11449417e-02 -1.20252430e+00
1.27044576e-03 -8.56074810e-01 -6.78209722e-01 -2.00179204e-01
-1.80298388e-01 7.95212165e-02 -6.23776138e-01 1.72359633e+00
-8.91351849e-02 1.44612238e-01 -1.29123092e-01 5.69458723e-01
8.54536712e-01 7.89018214e-01 -3.84675860e-02 4.87788096e-02
1.25036132e+00 -1.42123926e+00 -5.35995245e-01 -4.35940057e-01
8.12049568e-01 -5.79565763e-01 1.38791323e+00 2.41636932e-01
-1.16421950e+00 -7.55382717e-01 -1.09961689e+00 -2.08620846e-01
-9.17074829e-02 5.37421703e-01 4.43646073e-01 3.13542813e-01
-9.60422456e-01 5.39070249e-01 -1.05806243e+00 3.18323523e-02
6.99099839e-01 -2.59447675e-02 4.45358176e-03 -1.25041887e-01
-9.07522798e-01 5.23916423e-01 3.37330610e-01 5.46194427e-02
-1.20813119e+00 -9.01724517e-01 -1.05691159e+00 5.15018225e-01
2.47384533e-01 -5.53735018e-01 1.48818827e+00 -6.93798721e-01
-1.16948569e+00 5.61764836e-01 -4.43661183e-01 -3.96942675e-01
1.43660501e-01 -3.70399714e-01 2.40722694e-03 -3.66858214e-01
3.10242504e-01 2.09565327e-01 4.65612769e-01 -9.96435881e-01
-5.33755004e-01 -2.26379097e-01 4.00607973e-01 -1.10907473e-01
-4.77738529e-01 2.10074663e-01 -6.15505219e-01 -6.51713133e-01
-3.84057969e-01 -6.27854407e-01 -3.39994766e-02 -4.36577767e-01
-6.38301551e-01 -4.98137385e-01 5.51956415e-01 -1.16668308e+00
1.92359197e+00 -2.14851522e+00 3.27432215e-01 -4.94777650e-01
3.89172316e-01 3.77310187e-01 -3.74237821e-02 4.19144690e-01
5.31166047e-02 3.26223612e-01 -3.58861148e-01 -3.79572004e-01
3.21735740e-01 -1.86270177e-01 -3.23049515e-01 2.54106104e-01
3.48929584e-01 1.04616678e+00 -1.00266349e+00 -4.90132809e-01
-6.75598383e-02 2.33913407e-01 -9.04374957e-01 5.16177595e-01
-5.78830421e-01 8.59571546e-02 -6.35272682e-01 4.40546572e-01
5.19994259e-01 -4.80037779e-01 -1.25120297e-01 -1.93449765e-01
-2.10902944e-01 9.86509204e-01 -3.32109809e-01 1.92551827e+00
-9.34500635e-01 6.46782935e-01 -2.66238123e-01 -9.99568045e-01
9.87526178e-01 1.65867463e-01 -1.81665063e-01 -8.04066658e-01
-6.23630844e-02 -1.14048431e-02 5.57933487e-02 -7.38360465e-01
4.14654374e-01 2.25406975e-01 -3.81295741e-01 5.01855493e-01
6.33470295e-03 3.78280193e-01 2.81979680e-01 3.03620607e-01
1.33489215e+00 4.65398461e-01 3.33014071e-01 -2.52562761e-01
8.63087654e-01 -1.25230432e-01 6.62731946e-01 4.73398447e-01
1.64554268e-01 3.92955154e-01 9.11687970e-01 -1.13138691e-01
-9.19769526e-01 -5.93512714e-01 4.74246919e-01 1.22111869e+00
-2.91485228e-02 -8.90498519e-01 -1.18602812e+00 -8.57476830e-01
-2.87440538e-01 9.45896864e-01 -7.19789267e-01 -4.87351179e-01
-7.83759296e-01 -6.44971013e-01 6.60277009e-01 9.32452857e-01
5.98564029e-01 -1.04584181e+00 -5.71345806e-01 4.12441164e-01
-3.56007218e-01 -8.87303472e-01 -7.38345325e-01 -1.18619926e-01
-6.86338305e-01 -9.07955766e-01 -4.77060914e-01 -6.44304097e-01
7.40193903e-01 -2.83832308e-02 1.42612064e+00 4.12937820e-01
-2.14892458e-02 -4.04139429e-01 -5.30624509e-01 -8.43308270e-02
-5.46365976e-01 4.87071007e-01 -7.80392945e-01 -1.58308551e-01
3.19276005e-01 -5.18600047e-01 -4.80636060e-01 4.23328346e-03
-7.20848560e-01 3.81046534e-01 8.99676323e-01 7.58446574e-01
2.24429399e-01 4.15433496e-02 4.03703153e-01 -1.08088636e+00
5.27934492e-01 -6.98726475e-01 -6.64302349e-01 3.79594505e-01
-4.50622946e-01 4.87052947e-01 1.11938775e+00 2.41625719e-02
-1.31100726e+00 -2.35362291e-01 -2.78417468e-01 -2.07182840e-01
9.48074684e-02 9.64016855e-01 -5.19097447e-01 2.90008575e-01
4.90754873e-01 6.33636653e-01 -4.38665628e-01 -6.69786870e-01
5.83207272e-02 7.61314094e-01 5.58772206e-01 -7.42102265e-01
6.96477592e-01 -6.13919869e-02 -5.22855937e-01 -5.75361662e-02
-1.01388907e+00 -1.24730386e-01 -3.17984253e-01 1.48980081e-01
7.61084855e-01 -9.30715263e-01 -3.66860569e-01 5.49390674e-01
-1.59342825e+00 -1.89893529e-01 2.70448804e-01 -1.00745939e-01
-3.79412323e-01 3.91742706e-01 -6.19614720e-01 -5.51508069e-01
-5.62670588e-01 -1.60675299e+00 1.21760476e+00 1.93252861e-01
-5.49952611e-02 -8.44286501e-01 -2.87843067e-02 3.10759515e-01
6.30882978e-01 3.67997706e-01 1.34437144e+00 -4.64015752e-01
-8.21142912e-01 -1.16376132e-02 -4.63839203e-01 3.59216809e-01
2.33864650e-01 1.60351261e-01 -9.08954382e-01 -1.62854791e-01
3.03882845e-02 -8.84550139e-02 6.93788588e-01 -6.68086037e-02
1.55161965e+00 -6.67654693e-01 -3.67013067e-01 5.85467577e-01
1.45360935e+00 2.11580440e-01 5.53888500e-01 1.47235110e-01
1.03820968e+00 4.16758776e-01 6.25632823e-01 5.10630310e-01
7.45700002e-01 6.90333426e-01 6.67679369e-01 2.13577121e-01
-2.09612399e-02 -6.06943309e-01 6.79339588e-01 1.26513541e+00
3.43862027e-01 -2.09151819e-01 -1.12168014e+00 7.07830906e-01
-1.80445600e+00 -7.31985748e-01 -2.18597665e-01 2.02960706e+00
1.05669856e+00 3.88443440e-01 -1.35846257e-01 -2.69098189e-02
7.61302829e-01 2.74592727e-01 -5.94973564e-01 -3.37675005e-01
3.83695632e-01 -1.38022691e-01 9.86076742e-02 2.88763732e-01
-9.03128564e-01 7.59494364e-01 4.70958662e+00 8.85742903e-01
-9.68944430e-01 1.93917170e-01 4.68365103e-01 2.99000651e-01
-6.68289959e-01 7.92494938e-02 -8.63061547e-01 8.49798262e-01
1.14450920e+00 -6.75355673e-01 3.66708130e-01 1.05462849e+00
7.99500346e-02 2.42823079e-01 -1.33940387e+00 5.71055412e-01
1.91338882e-01 -1.26726532e+00 1.41757891e-01 -1.81222618e-01
8.09481025e-01 1.01652183e-03 -3.94731686e-02 7.08430171e-01
2.29796261e-01 -9.71702516e-01 1.03275836e+00 4.48228151e-01
6.81303918e-01 -8.28239441e-01 9.26644802e-01 4.88453954e-01
-1.50280702e+00 -1.29016057e-01 -3.83725733e-01 4.08247821e-02
-1.06433183e-01 7.42784023e-01 -6.18565917e-01 7.27685571e-01
5.54468036e-01 1.19838989e+00 -9.51188564e-01 1.02510893e+00
-4.41342324e-01 8.15590143e-01 2.23792776e-01 -9.00265947e-02
4.43743795e-01 1.99047029e-01 2.00837955e-01 1.39118779e+00
3.98804486e-01 -4.17549521e-01 -2.61036661e-02 1.59926236e+00
-3.14075619e-01 -4.63790819e-02 -4.59199220e-01 -3.67027014e-01
4.40103054e-01 1.03857565e+00 -2.10377097e-01 -5.18931270e-01
-6.51635408e-01 6.60594463e-01 5.03719032e-01 9.24318582e-02
-1.12856758e+00 -8.95982265e-01 5.38134634e-01 9.11201090e-02
4.94351685e-01 3.88573948e-03 -2.89560795e-01 -1.12899613e+00
4.95369285e-01 -1.09758198e+00 3.61024961e-02 -8.60211730e-01
-7.43994236e-01 8.71151686e-01 -1.16442889e-01 -1.21226978e+00
-3.54220569e-01 -2.49929488e-01 -9.72371101e-01 1.09886158e+00
-1.39168847e+00 -1.00154996e+00 -6.51648045e-01 -4.39826772e-02
7.76493728e-01 -1.72104910e-01 4.79682267e-01 4.40355539e-01
-7.89051592e-01 9.04621661e-01 -9.01599154e-02 2.52625465e-01
2.73723662e-01 -1.38462830e+00 1.11775875e+00 1.27923679e+00
-1.87931389e-01 1.13399982e+00 5.13254821e-01 -5.55667400e-01
-1.49256980e+00 -1.39482033e+00 8.32805812e-01 -3.32453519e-01
6.64681435e-01 -5.47156096e-01 -1.23248076e+00 6.27069592e-01
3.55374038e-01 -2.54483879e-01 3.76250595e-01 -1.37870431e-01
-5.38492799e-01 -1.16678976e-01 -6.94054902e-01 4.12631124e-01
8.85918677e-01 -7.05510437e-01 -5.25609612e-01 7.53339753e-02
1.14759004e+00 -5.64368784e-01 -8.72252226e-01 2.92650878e-01
2.13776931e-01 -1.10047138e+00 4.72213477e-01 -3.75889242e-01
1.15617728e+00 -5.37161350e-01 4.17355411e-02 -1.28612530e+00
-3.14212531e-01 -5.27603269e-01 -2.33611718e-01 1.73742557e+00
3.75119746e-01 -2.02120319e-01 6.14310324e-01 1.49947569e-01
-6.95980966e-01 -9.22011852e-01 -3.73235554e-01 -5.08305550e-01
2.27719848e-03 -1.87380001e-01 9.27127361e-01 6.12375557e-01
4.74398434e-02 4.94591147e-01 -7.68893287e-02 -1.01777703e-01
3.80247146e-01 4.45039809e-01 6.34514809e-01 -9.37134683e-01
-5.51711023e-01 -6.07004404e-01 -2.21843466e-01 -1.23752737e+00
5.38657546e-01 -1.15803325e+00 2.95591325e-01 -1.65657842e+00
4.98279542e-01 -2.19309971e-01 -2.36780241e-01 4.54171091e-01
-6.53202355e-01 -4.34676915e-01 -9.24075302e-03 -6.02015406e-02
-6.85390711e-01 7.24018753e-01 9.67258751e-01 -5.03418744e-01
1.62095904e-01 -8.91483426e-02 -1.07228994e+00 5.19770741e-01
7.57124364e-01 -5.90414524e-01 -5.60634553e-01 -9.94981647e-01
3.81071150e-01 2.39715755e-01 1.85796738e-01 -1.05073941e+00
3.47601235e-01 9.48485062e-02 -1.55419096e-01 -6.01970077e-01
-3.71084511e-01 -2.72660792e-01 -5.25987968e-02 5.11307538e-01
-5.19602954e-01 3.46719772e-01 3.13898236e-01 4.58694577e-01
-5.45146704e-01 -5.07210732e-01 6.62353754e-01 -2.53575265e-01
-5.68564832e-01 2.70540267e-01 -8.69727209e-02 3.81601304e-01
5.84885895e-01 2.04273090e-01 -6.88729227e-01 7.95348138e-02
-3.03847101e-02 2.73371488e-01 6.57812476e-01 6.24637663e-01
5.00865757e-01 -1.28489602e+00 -8.46879482e-01 1.79212317e-01
4.63238358e-01 1.53829902e-01 3.35594505e-01 9.52095032e-01
-5.01640499e-01 6.55620635e-01 -7.72844627e-02 -4.20216233e-01
-1.06527519e+00 5.61642170e-01 3.18561673e-01 -5.46372235e-01
-5.37173629e-01 8.60036552e-01 5.66476107e-01 -2.72845745e-01
1.35679752e-01 -6.24453545e-01 -1.50784642e-01 -2.39212081e-01
6.22450054e-01 2.22148284e-01 2.00666502e-01 -4.37226772e-01
-5.17839193e-01 4.76289004e-01 -4.01118755e-01 4.39630479e-01
1.36381221e+00 1.43945500e-01 -5.33338428e-01 3.15292507e-01
1.50477517e+00 -2.27118470e-02 -1.09589648e+00 -2.30112791e-01
3.41707110e-01 -3.30295175e-01 2.12050192e-02 -7.50282049e-01
-1.24753046e+00 1.34703195e+00 3.90168168e-02 4.58132215e-02
1.06558108e+00 -8.19988772e-02 7.73112595e-01 1.12939142e-01
3.43881696e-01 -4.37294692e-01 3.11182942e-02 5.57160079e-01
9.55035210e-01 -8.61413836e-01 -3.00950140e-01 -2.67546505e-01
-4.24803406e-01 8.18416119e-01 1.04447103e+00 -1.22228682e-01
8.73788223e-02 4.63508636e-01 -4.41909462e-01 -6.41676784e-02
-1.17370248e+00 1.86295703e-01 3.82489681e-01 3.43108028e-01
9.56644714e-01 -1.71491370e-01 -6.26438409e-02 8.96727800e-01
-6.77666739e-02 6.38726428e-02 5.76916635e-01 7.67016530e-01
-3.81345600e-01 -9.17457402e-01 1.07421890e-01 6.10224605e-01
-5.77347815e-01 -4.19770271e-01 -1.67547688e-01 3.75978321e-01
7.72630274e-02 6.70840561e-01 -1.70187160e-01 -5.15538037e-01
4.23177987e-01 -8.27806443e-02 1.53710306e-01 -9.45903301e-01
-9.97386336e-01 -2.72359252e-01 -1.24817332e-02 -6.63201928e-01
-1.38196960e-01 -5.62851191e-01 -1.25109971e+00 -8.19482282e-02
-2.68915296e-01 3.52048844e-01 4.03744966e-01 7.30113268e-01
5.58410525e-01 1.22807109e+00 5.73376775e-01 -5.87506831e-01
-5.93027174e-01 -1.11665320e+00 1.98414195e-02 1.96774527e-01
5.58389187e-01 -4.82693702e-01 -2.68339843e-01 2.81869233e-01] | [7.569018840789795, 7.963604927062988] |
ea0b8da3-4e9f-42d6-a396-b8809d67c253 | minvis-a-minimal-video-instance-segmentation | 2208.02245 | null | https://arxiv.org/abs/2208.02245v1 | https://arxiv.org/pdf/2208.02245v1.pdf | MinVIS: A Minimal Video Instance Segmentation Framework without Video-based Training | We propose MinVIS, a minimal video instance segmentation (VIS) framework that achieves state-of-the-art VIS performance with neither video-based architectures nor training procedures. By only training a query-based image instance segmentation model, MinVIS outperforms the previous best result on the challenging Occluded VIS dataset by over 10% AP. Since MinVIS treats frames in training videos as independent images, we can drastically sub-sample the annotated frames in training videos without any modifications. With only 1% of labeled frames, MinVIS outperforms or is comparable to fully-supervised state-of-the-art approaches on YouTube-VIS 2019/2021. Our key observation is that queries trained to be discriminative between intra-frame object instances are temporally consistent and can be used to track instances without any manually designed heuristics. MinVIS thus has the following inference pipeline: we first apply the trained query-based image instance segmentation to video frames independently. The segmented instances are then tracked by bipartite matching of the corresponding queries. This inference is done in an online fashion and does not need to process the whole video at once. MinVIS thus has the practical advantages of reducing both the labeling costs and the memory requirements, while not sacrificing the VIS performance. Code is available at: https://github.com/NVlabs/MinVIS | ['Anima Anandkumar', 'Zhiding Yu', 'De-An Huang'] | 2022-08-03 | null | null | null | null | ['video-instance-segmentation'] | ['computer-vision'] | [ 9.16414186e-02 8.89045298e-02 -6.47312939e-01 -1.61555856e-01
-1.18518257e+00 -7.69097924e-01 3.22685778e-01 -4.04984616e-02
-6.73229575e-01 4.38385874e-01 -2.59355426e-01 -1.07403882e-01
2.73670107e-01 -5.72518826e-01 -1.24366260e+00 -5.11908114e-01
1.60515774e-02 6.76296949e-01 9.39268291e-01 4.41695094e-01
-5.87649643e-02 3.12875628e-01 -1.68177211e+00 4.01633680e-01
3.44922125e-01 1.03489304e+00 2.75282115e-01 9.22259927e-01
-1.45315498e-01 9.98222530e-01 -3.49213630e-01 -3.95046562e-01
5.61509252e-01 -5.76237500e-01 -1.09645915e+00 6.11916423e-01
1.14442790e+00 -5.48606813e-01 -8.38154137e-01 9.64100182e-01
2.44626030e-01 4.01015967e-01 2.67783284e-01 -1.31456697e+00
-1.63619742e-01 5.54553211e-01 -6.56820714e-01 4.66716915e-01
4.28702682e-01 3.21930826e-01 1.01446867e+00 -8.71251941e-01
1.21086049e+00 9.92226005e-01 3.27966690e-01 4.91021693e-01
-1.34549654e+00 -3.36205065e-01 3.79899085e-01 5.89503109e-01
-1.48249400e+00 -5.32515168e-01 4.01342124e-01 -5.28965354e-01
7.83489108e-01 3.46983850e-01 1.03003561e+00 8.39891374e-01
-3.63747418e-01 1.29197395e+00 8.47763598e-01 -1.76278532e-01
2.84463912e-01 -1.48309827e-01 2.43557125e-01 1.03112483e+00
5.09655997e-02 -4.84413430e-02 -5.78230679e-01 1.23922437e-01
6.78849161e-01 2.39382274e-02 -3.38134080e-01 -5.73724747e-01
-1.37569320e+00 4.04237330e-01 2.00085625e-01 -4.89463694e-02
-4.32846069e-01 6.94146633e-01 5.77899754e-01 1.95495054e-01
3.20467561e-01 -3.82310823e-02 -5.61700225e-01 -2.73865372e-01
-1.52625251e+00 1.85359240e-01 7.35822022e-01 1.17385364e+00
1.02006161e+00 -5.43072112e-02 -2.80981183e-01 4.92924094e-01
1.30509973e-01 5.99111736e-01 -1.52421623e-01 -1.67448068e+00
2.51184553e-01 3.28735799e-01 -5.06640933e-02 -7.24937916e-01
-1.65420864e-02 -1.14612550e-01 -1.24139085e-01 3.71887088e-02
6.76402867e-01 -8.21506456e-02 -1.21946383e+00 1.38484704e+00
7.08920181e-01 6.22646451e-01 -1.61840662e-01 9.87298131e-01
8.38674486e-01 8.22858751e-01 1.29415765e-01 -3.63523602e-01
1.45070362e+00 -1.35059047e+00 -5.06081045e-01 -2.85979658e-01
4.61042851e-01 -7.90596962e-01 9.15516734e-01 3.78376424e-01
-1.49886680e+00 -5.32777250e-01 -7.40310133e-01 -1.98770165e-01
-1.42017290e-01 1.24008179e-01 4.07429963e-01 3.05844218e-01
-1.12673450e+00 4.07516748e-01 -1.16276371e+00 -3.32730740e-01
7.28598714e-01 3.80076319e-01 -3.67015183e-01 -3.51703048e-01
-6.87618017e-01 2.61421859e-01 5.71644485e-01 -1.29532432e-02
-1.38006878e+00 -8.05251777e-01 -9.38930571e-01 -8.29378143e-02
1.09696150e+00 -6.22685850e-01 1.31951332e+00 -1.41257989e+00
-1.17590940e+00 1.00727236e+00 -6.64893568e-01 -5.59159040e-01
7.70766020e-01 -2.47283146e-01 -2.30397861e-02 7.51150191e-01
1.13535628e-01 1.13232112e+00 8.63793433e-01 -1.22612226e+00
-8.58306348e-01 7.86347762e-02 4.61912304e-01 1.51875123e-01
8.21321905e-02 -2.91460697e-02 -1.73344457e+00 -3.72575819e-01
-1.49908438e-02 -1.30138171e+00 -1.14921764e-01 2.31955305e-01
-3.57629329e-01 -1.30960271e-01 9.69775438e-01 -6.73750699e-01
1.07222927e+00 -2.08618116e+00 1.76473245e-01 6.98448420e-02
1.42517209e-01 3.32687378e-01 -3.02093506e-01 2.58994579e-01
2.71729410e-01 -1.72768757e-01 -2.57915646e-01 -4.62876827e-01
-1.43692404e-01 2.92424142e-01 4.90176007e-02 6.38787568e-01
-1.21654972e-01 9.89421785e-01 -1.03060186e+00 -9.24685955e-01
5.60203850e-01 4.07313436e-01 -7.40288675e-01 1.63461685e-01
-7.36969352e-01 3.35516214e-01 -7.70216510e-02 7.73641884e-01
5.51630020e-01 -4.84496057e-01 3.57662708e-01 -4.98325825e-01
1.46550432e-01 -4.90473844e-02 -1.25067234e+00 1.91496134e+00
-1.06220767e-01 1.02296114e+00 -2.15273704e-02 -1.12463856e+00
9.80911776e-02 2.56733000e-01 8.93730938e-01 -6.26454592e-01
7.65370876e-02 3.17819864e-02 -3.95741522e-01 -7.26740301e-01
2.82513469e-01 6.56866908e-01 2.47403100e-01 3.71105820e-01
3.67199153e-01 1.20995328e-01 9.05897081e-01 5.11035740e-01
9.92455840e-01 5.45359135e-01 5.83380684e-02 -4.57897708e-02
5.09885490e-01 3.96586299e-01 7.08427548e-01 6.34749830e-01
-2.75329769e-01 5.89382708e-01 5.60320437e-01 -3.95898551e-01
-8.85665476e-01 -1.02418590e+00 6.94571109e-03 1.20872426e+00
5.53826988e-01 -5.97405016e-01 -1.07245910e+00 -8.63381445e-01
-1.02391474e-01 3.93256038e-01 -5.50709963e-01 4.21969175e-01
-6.35010958e-01 -1.58604905e-01 3.31066638e-01 5.04267037e-01
4.90901709e-01 -1.01265156e+00 -6.23673916e-01 7.04148635e-02
-3.96708548e-01 -1.67913342e+00 -7.05658019e-01 -1.04190633e-01
-7.45373011e-01 -1.53716111e+00 -6.90616131e-01 -7.67085552e-01
8.30703557e-01 5.10428429e-01 1.32029927e+00 2.57720441e-01
-3.51564765e-01 8.00943375e-01 -3.40758622e-01 9.64242145e-02
-8.31973404e-02 -3.03475857e-02 -3.80951941e-01 -5.83419651e-02
4.15040821e-01 -3.28462273e-02 -8.91842008e-01 3.89026105e-01
-9.47945118e-01 1.26438051e-01 2.49860764e-01 6.27273500e-01
1.25845218e+00 -2.46858150e-01 1.35681462e-02 -1.14384699e+00
-5.01810074e-01 -3.41405332e-01 -9.23960447e-01 2.53971010e-01
-2.31238842e-01 -3.05868149e-01 2.46085808e-01 -3.51707101e-01
-7.92080283e-01 5.77561617e-01 1.13259733e-01 -8.94329488e-01
-2.29659423e-01 2.29410648e-01 1.37689814e-01 -8.61235783e-02
2.97900468e-01 2.51946539e-01 -1.04218356e-01 -1.67281568e-01
5.36772668e-01 2.63000220e-01 7.35388994e-01 -5.04830778e-01
5.24884641e-01 7.18976915e-01 -1.67888761e-01 -8.63758743e-01
-9.63388562e-01 -8.96782577e-01 -6.54012442e-01 -5.92569053e-01
1.22217941e+00 -1.11135364e+00 -7.38218546e-01 1.72543332e-01
-9.47975516e-01 -8.33704352e-01 -4.38976377e-01 4.03034121e-01
-8.69792879e-01 5.14935195e-01 -8.13145936e-01 -4.20348763e-01
-9.36084092e-02 -1.33259892e+00 1.14843905e+00 9.14925709e-02
-8.98570195e-02 -8.80844414e-01 -2.47607246e-01 6.99318409e-01
-1.91651210e-01 1.81165978e-01 2.63999403e-01 -4.45435822e-01
-1.28406799e+00 -3.64722125e-02 -2.78439075e-01 3.44919503e-01
-8.77646580e-02 2.03639299e-01 -8.96809697e-01 -3.70448500e-01
-5.36793113e-01 -3.41560721e-01 9.34151292e-01 7.03620732e-01
1.28977883e+00 -1.41916931e-01 -3.18711430e-01 7.11634517e-01
1.53111601e+00 1.63764641e-01 5.77867031e-01 2.84270674e-01
9.38883305e-01 3.07662040e-01 9.34888005e-01 2.04121783e-01
3.78396690e-01 7.26674974e-01 4.77646381e-01 -2.21635699e-01
-4.40668076e-01 -7.40867555e-02 5.67736387e-01 5.24695456e-01
-1.04393706e-01 -3.25157940e-01 -7.54110992e-01 8.26279461e-01
-2.17400789e+00 -1.22597742e+00 -1.39590502e-01 2.08500838e+00
5.82590163e-01 -4.38823365e-02 3.83187622e-01 -1.22961730e-01
6.87303960e-01 2.67219096e-01 -6.33283973e-01 1.36751801e-01
2.11573523e-02 1.18674219e-01 7.91977882e-01 5.46317697e-01
-1.40967262e+00 1.31130707e+00 6.10130739e+00 8.13011050e-01
-9.44270730e-01 2.39975646e-01 7.54453421e-01 -5.85852563e-01
3.67820784e-02 2.71373212e-01 -6.43217504e-01 4.37422067e-01
7.09415972e-01 7.40063488e-02 6.63375020e-01 8.71003747e-01
3.88182700e-01 -4.13051009e-01 -1.22533095e+00 1.11682045e+00
1.47000700e-01 -1.67458844e+00 1.83994733e-02 -9.41280425e-02
9.63205397e-01 3.88756275e-01 -2.78495461e-01 3.43432613e-02
7.71449283e-02 -5.15347302e-01 1.05842733e+00 4.45642442e-01
7.86666751e-01 -5.04189551e-01 5.59057117e-01 1.03037404e-02
-1.41191208e+00 1.64838284e-01 -3.13056678e-01 4.36131746e-01
4.86819208e-01 2.70783752e-01 -4.50864911e-01 4.73659277e-01
9.08857942e-01 9.37514603e-01 -5.23662210e-01 1.21897233e+00
-1.16477542e-01 8.26858222e-01 -3.43706787e-01 4.96069193e-01
4.33868289e-01 -2.60255694e-01 4.07074571e-01 1.34211624e+00
5.58893532e-02 2.05576181e-01 8.02843928e-01 4.66239989e-01
-2.40686506e-01 -1.54888462e-02 -2.30245098e-01 -7.35939816e-02
4.65699643e-01 1.15445340e+00 -1.08673191e+00 -8.65444064e-01
-7.09688187e-01 1.22582889e+00 -3.64572182e-03 6.56009257e-01
-1.11789000e+00 8.41422006e-02 5.21227002e-01 2.90527225e-01
7.50740945e-01 -1.69078693e-01 2.58270085e-01 -1.18057430e+00
6.11702912e-02 -8.96196783e-01 6.31173313e-01 -7.20223725e-01
-6.43691480e-01 4.90808398e-01 1.36558086e-01 -1.18936670e+00
-2.54454613e-01 -3.39775652e-01 -1.82967439e-01 1.22057013e-01
-1.51425183e+00 -1.03387558e+00 -5.45990288e-01 8.88640702e-01
1.02257788e+00 2.42601797e-01 2.81781703e-01 5.48014283e-01
-7.05097914e-01 3.23138922e-01 -1.12240173e-01 4.24271971e-01
6.54886425e-01 -1.11227441e+00 2.74354964e-01 1.18866324e+00
5.85408092e-01 1.87833712e-01 5.71980238e-01 -5.18386006e-01
-1.59609473e+00 -1.27131426e+00 4.91286397e-01 -3.38937193e-01
5.40997922e-01 -1.93089500e-01 -6.45037472e-01 9.58197653e-01
4.57784921e-01 6.48663282e-01 3.91855329e-01 -3.64194542e-01
-2.44934306e-01 -1.66773528e-01 -7.67430365e-01 4.80446965e-01
1.32517028e+00 -5.28079450e-01 -1.37200266e-01 6.93470180e-01
6.53508246e-01 -6.42268777e-01 -8.46468627e-01 7.59279355e-02
3.14333469e-01 -8.88576448e-01 1.09547508e+00 -4.08050835e-01
2.07039580e-01 -6.90912068e-01 -1.95447385e-01 -5.72362185e-01
4.60850596e-02 -6.80846989e-01 -5.19957721e-01 1.07551360e+00
3.02927047e-01 -2.91095346e-01 1.08401501e+00 5.76803088e-01
-7.65647665e-02 -7.43657649e-01 -7.94450700e-01 -8.02026153e-01
-6.09325111e-01 -6.62393630e-01 1.41832858e-01 6.98998570e-01
-2.58868128e-01 -6.55975938e-02 -3.24295104e-01 1.88122064e-01
8.87260914e-01 2.38414556e-01 9.90792811e-01 -6.79723561e-01
-5.38433373e-01 -1.24171689e-01 -5.56850493e-01 -1.29865098e+00
2.68531650e-01 -7.54666030e-01 1.93225965e-02 -1.62428379e+00
3.30246985e-01 -3.68143499e-01 -2.65648812e-01 7.01484621e-01
-7.50680193e-02 7.20281601e-01 4.73607361e-01 3.46196055e-01
-1.25875378e+00 4.64050099e-02 1.07485497e+00 -3.25176030e-01
-8.94827470e-02 -2.55158603e-01 -1.48252234e-01 7.03934789e-01
4.58466202e-01 -4.89832908e-01 -6.24786973e-01 -4.85727340e-01
-3.98927033e-02 3.68178248e-01 4.46578979e-01 -1.05977547e+00
3.18678647e-01 -1.70699447e-01 1.78951979e-01 -6.21169448e-01
4.68724042e-01 -8.94858181e-01 4.16763365e-01 4.87625390e-01
-1.72335342e-01 -3.10622845e-02 1.68626711e-01 7.60098159e-01
-3.77392709e-01 -1.77114308e-01 8.28545809e-01 -1.93514854e-01
-1.24615574e+00 8.02482903e-01 -3.21882457e-01 2.67614335e-01
1.48590744e+00 -3.41162115e-01 -2.23064095e-01 -2.63689935e-01
-8.81319821e-01 3.53848666e-01 7.98425674e-01 2.62925833e-01
3.43249053e-01 -9.24254835e-01 -5.01204371e-01 -2.27091283e-01
3.89372334e-02 1.53636634e-01 5.44023335e-01 1.06050360e+00
-9.38893914e-01 2.67432868e-01 1.45609304e-01 -1.13693929e+00
-1.52602577e+00 7.75784671e-01 2.43255392e-01 1.02690227e-01
-8.43703985e-01 7.66616583e-01 2.36171961e-01 6.64622113e-02
3.27341974e-01 -1.26550794e-01 1.37987196e-01 1.88879013e-01
4.62012023e-01 4.94959295e-01 -6.79324642e-02 -8.16340446e-01
-3.90470326e-01 6.63543284e-01 -7.13407621e-02 -8.60643834e-02
1.04733300e+00 -1.43652707e-01 1.24126658e-01 2.20930979e-01
1.30042207e+00 -9.27924514e-02 -1.68771458e+00 -2.65599072e-01
-1.53067589e-01 -7.08830178e-01 1.18154347e-01 -4.52074587e-01
-1.70587671e+00 4.66803968e-01 4.67550904e-01 -3.81203033e-02
1.14805710e+00 1.95108935e-01 1.00110447e+00 2.84256399e-01
3.69496673e-01 -1.14467537e+00 6.76634684e-02 1.49336800e-01
2.47646943e-01 -1.19440782e+00 1.73773661e-01 -6.50050879e-01
-7.37945199e-01 9.52460051e-01 4.90239322e-01 -3.57508659e-01
4.47208196e-01 2.43225515e-01 2.03183591e-01 -2.21221432e-01
-7.59533763e-01 -4.88827348e-01 3.92993540e-01 3.89985353e-01
2.14336008e-01 -2.15277612e-01 -2.21838310e-01 -2.41209194e-01
3.73652160e-01 1.84859157e-01 4.38117921e-01 8.18864703e-01
-2.85056919e-01 -1.05521226e+00 -1.53484240e-01 4.73166049e-01
-5.68323016e-01 -8.17568451e-02 -8.44798386e-02 9.38243985e-01
-1.04735540e-02 7.71549344e-01 2.66974479e-01 -5.71061112e-02
1.46987826e-01 -6.30345270e-02 5.49477994e-01 -4.08172071e-01
-3.40394139e-01 3.87056410e-01 1.28498495e-01 -1.29397166e+00
-9.52953160e-01 -9.41973448e-01 -1.45464909e+00 -3.84120077e-01
-2.84570336e-01 5.02040647e-02 3.52147132e-01 8.31130564e-01
5.58383524e-01 4.71007735e-01 2.72429854e-01 -1.13422394e+00
1.77314997e-01 -3.55116457e-01 -1.70685187e-01 5.05463183e-01
3.11972976e-01 -5.75699031e-01 -1.57139465e-01 6.81548178e-01] | [9.115948677062988, -0.053552862256765366] |
a9267d88-8a3d-47b4-bc94-23dc3cb8ee62 | gm-tcnet-gated-multi-scale-temporal | 2210.15834 | null | https://arxiv.org/abs/2210.15834v1 | https://arxiv.org/pdf/2210.15834v1.pdf | GM-TCNet: Gated Multi-scale Temporal Convolutional Network using Emotion Causality for Speech Emotion Recognition | In human-computer interaction, Speech Emotion Recognition (SER) plays an essential role in understanding the user's intent and improving the interactive experience. While similar sentimental speeches own diverse speaker characteristics but share common antecedents and consequences, an essential challenge for SER is how to produce robust and discriminative representations through causality between speech emotions. In this paper, we propose a Gated Multi-scale Temporal Convolutional Network (GM-TCNet) to construct a novel emotional causality representation learning component with a multi-scale receptive field. GM-TCNet deploys a novel emotional causality representation learning component to capture the dynamics of emotion across the time domain, constructed with dilated causal convolution layer and gating mechanism. Besides, it utilizes skip connection fusing high-level features from different gated convolution blocks to capture abundant and subtle emotion changes in human speech. GM-TCNet first uses a single type of feature, mel-frequency cepstral coefficients, as inputs and then passes them through the gated temporal convolutional module to generate the high-level features. Finally, the features are fed to the emotion classifier to accomplish the SER task. The experimental results show that our model maintains the highest performance in most cases compared to state-of-the-art techniques. | ['Kun-Hong Liu', 'Li-Yan Chen', 'Chang-Li Wu', 'Yan Luo', 'Yong Xu', 'Xuan-Ze Wang', 'Xin-Cheng Wen', 'Jia-Xin Ye'] | 2022-10-28 | null | null | null | null | ['speech-emotion-recognition'] | ['speech'] | [-9.68828723e-02 -3.33451897e-01 6.87416494e-02 -5.77117920e-01
-3.33157063e-01 -3.14311415e-01 2.52847105e-01 -1.52335867e-01
-8.52563232e-02 4.50995833e-01 5.46498299e-01 -8.72392431e-02
1.77147403e-01 -4.51339781e-01 -2.63484001e-01 -5.61809778e-01
-4.94517267e-01 -5.64642966e-01 2.08695270e-02 -4.79456723e-01
-9.84145105e-02 2.41333500e-01 -1.51175141e+00 6.72652483e-01
7.29029834e-01 1.53626180e+00 -4.42212336e-02 9.67014968e-01
-1.95895761e-01 1.35096049e+00 -6.27953410e-01 -2.35285699e-01
-5.17769217e-01 -6.13416970e-01 -7.43573010e-01 -1.46394789e-01
-4.81820852e-01 2.12621484e-02 -3.16092372e-01 8.31538796e-01
5.63920319e-01 5.30008793e-01 2.74976104e-01 -1.22415769e+00
-8.34939480e-01 2.81914055e-01 -2.20685586e-01 5.78300893e-01
6.46347761e-01 -9.09937918e-02 1.17065382e+00 -9.57536936e-01
2.18493566e-01 1.40449405e+00 6.20524824e-01 5.44212699e-01
-5.54689527e-01 -8.57825577e-01 6.77484691e-01 4.66532588e-01
-1.02341497e+00 -3.40832770e-01 1.06860554e+00 -1.41585886e-01
1.23440957e+00 3.48941088e-01 7.89649367e-01 1.43290901e+00
4.21507746e-01 9.35750127e-01 8.69950533e-01 -1.69302467e-02
1.65707737e-01 -1.47738960e-02 1.19600378e-01 5.91040909e-01
-1.04816473e+00 1.12527341e-01 -8.70910525e-01 -2.82308429e-01
4.07095909e-01 -7.50219822e-02 -4.12971020e-01 4.78228360e-01
-9.45495486e-01 9.05888915e-01 5.44425547e-01 6.63134634e-01
-8.01898062e-01 2.15338930e-01 7.64251530e-01 4.45943862e-01
6.98319435e-01 -1.20187057e-02 -7.68511295e-01 -5.86457312e-01
-3.69677395e-01 -2.20306799e-01 6.47756040e-01 4.02392894e-01
2.17928991e-01 3.46076012e-01 -1.15820073e-01 8.67137849e-01
2.67343968e-01 2.09354132e-01 8.48500192e-01 -2.87745684e-01
5.24596982e-02 5.59085608e-01 -2.21273780e-01 -1.22689211e+00
-4.39526528e-01 -5.19463122e-01 -9.98141170e-01 -3.13267767e-01
-5.25885105e-01 -4.64109361e-01 -5.41141212e-01 1.91186595e+00
2.55636364e-01 7.13630974e-01 4.97771949e-01 1.02212429e+00
1.11859941e+00 1.17881870e+00 4.05305356e-01 -5.90001523e-01
1.47852647e+00 -7.64813185e-01 -1.19691491e+00 -8.68299231e-02
3.06358337e-01 -7.00143695e-01 1.10442305e+00 3.77557069e-01
-7.58748889e-01 -5.76890349e-01 -9.90613043e-01 9.78709534e-02
-5.21657586e-01 2.88858443e-01 9.18798864e-01 2.82989442e-01
-1.01145053e+00 3.75645876e-01 -4.48215812e-01 -2.11807087e-01
9.25262719e-02 3.86314631e-01 -1.85862407e-01 6.68568373e-01
-2.12975144e+00 6.58312619e-01 1.60081074e-01 5.45411170e-01
-7.40475237e-01 -5.05418420e-01 -1.03089476e+00 1.67469621e-01
-1.96750507e-01 -2.77587771e-01 1.48573482e+00 -1.56253195e+00
-1.98454571e+00 2.33757138e-01 -4.03543174e-01 -2.40332425e-01
-7.90217817e-02 -1.84080303e-01 -1.19680786e+00 1.37208417e-01
-2.57917732e-01 1.70119286e-01 1.14999878e+00 -7.87347972e-01
-6.75134778e-01 -4.97240834e-02 -1.62503004e-01 3.24222833e-01
-7.45406568e-01 5.86373687e-01 -2.81402737e-01 -9.06333804e-01
-7.20786080e-02 -5.66546381e-01 -1.08962469e-01 -5.46036422e-01
7.99948797e-02 -6.26033068e-01 1.30573106e+00 -6.77466929e-01
1.60791373e+00 -2.44196916e+00 1.19954564e-01 -6.12463430e-02
4.96819690e-02 1.24520071e-01 -1.09226957e-01 3.21850747e-01
-6.11413002e-01 4.31626663e-02 1.71642676e-02 -2.56478548e-01
-1.61382720e-01 1.23409964e-01 -7.61131644e-01 2.71496177e-01
7.23864436e-01 9.47535932e-01 -1.08621407e+00 -2.00054333e-01
2.45434403e-01 8.68031859e-01 -2.26183668e-01 4.34248656e-01
-2.32909191e-02 5.92499077e-01 -6.77831590e-01 2.90771186e-01
3.53708923e-01 -2.52154261e-01 1.06429234e-01 -1.52698025e-01
-1.91128463e-01 6.56425178e-01 -9.58240449e-01 1.52281022e+00
-9.20021474e-01 8.28983963e-01 4.04576072e-03 -9.37001705e-01
1.16320395e+00 9.27260935e-01 4.81644899e-01 -8.30755830e-01
4.58987743e-01 7.19707506e-03 -1.15356073e-01 -7.57034898e-01
4.03683424e-01 -5.28313279e-01 -4.55010623e-01 3.14539135e-01
1.89641386e-01 1.97571144e-01 -5.78618467e-01 9.07338932e-02
1.01205516e+00 -3.06735247e-01 2.48941168e-01 -3.81509587e-02
1.02305508e+00 -5.10388017e-01 8.76628697e-01 -1.73616424e-01
-4.68991041e-01 8.34914669e-02 5.26817203e-01 -5.57773948e-01
-3.86943877e-01 -7.66681969e-01 1.50312915e-01 1.42538309e+00
2.39495903e-01 -2.92602152e-01 -4.17881668e-01 -7.80789256e-01
-3.61270547e-01 5.65343261e-01 -9.78350580e-01 -5.44657171e-01
-3.69379640e-01 -7.52336562e-01 5.10679901e-01 6.88776374e-01
5.43791950e-01 -1.62493038e+00 -5.76380730e-01 3.52096915e-01
-4.92538989e-01 -1.05975044e+00 -5.15120566e-01 3.98165315e-01
-4.65750962e-01 -6.06208742e-01 -3.58118147e-01 -8.65141034e-01
1.62929863e-01 8.75614863e-03 8.71447086e-01 -8.39935616e-02
-7.78075308e-02 2.03084514e-01 -7.18087852e-01 -4.11690056e-01
-9.32254717e-02 -2.25826308e-01 1.20766751e-01 6.00709677e-01
3.85536015e-01 -6.87697232e-01 -5.14781952e-01 1.11349583e-01
-6.42641664e-01 -2.79565424e-01 2.89716929e-01 9.42754269e-01
2.68918931e-01 4.66939330e-01 1.06807518e+00 -3.05151224e-01
1.05331957e+00 -7.36174762e-01 1.30495980e-01 5.60032986e-02
1.57708172e-02 -1.86801776e-01 1.01362872e+00 -7.42481768e-01
-1.47323108e+00 -1.85998663e-01 -2.91236579e-01 -6.17185652e-01
3.63648906e-02 6.58822000e-01 -3.88671644e-02 3.70473027e-01
2.80765861e-01 2.84054071e-01 -3.45622391e-01 -5.77618629e-02
4.07257825e-01 8.75154674e-01 4.91547793e-01 -2.51262754e-01
2.43424356e-01 3.12514782e-01 -7.12314963e-01 -9.19283986e-01
-8.34290147e-01 -5.25872409e-01 -1.93007708e-01 -4.60906118e-01
1.21857834e+00 -1.07041919e+00 -1.02161801e+00 4.40326542e-01
-1.43677568e+00 -2.00114116e-01 5.41490117e-05 6.32945120e-01
-3.58709306e-01 -4.64443713e-02 -8.74266505e-01 -1.19509900e+00
-5.87544084e-01 -8.40155244e-01 1.05514503e+00 4.75649029e-01
-3.70422095e-01 -1.04625010e+00 -1.06280468e-01 -1.57482803e-01
3.41228127e-01 8.11808184e-02 6.44629002e-01 -4.58622038e-01
2.21645534e-01 -4.48867455e-02 -5.10394163e-02 4.28178787e-01
2.41887540e-01 9.05128494e-02 -1.23406434e+00 1.24083742e-01
3.45590025e-01 -3.53507042e-01 7.28966117e-01 3.36543113e-01
1.40248382e+00 -2.10575446e-01 -1.63749047e-02 3.87047708e-01
8.63280535e-01 6.53756201e-01 5.25985122e-01 -2.21482009e-01
5.39128423e-01 7.30548322e-01 6.87614858e-01 5.86067915e-01
5.25234699e-01 3.16296101e-01 3.64101201e-01 -3.01599830e-01
1.79445058e-01 -1.33526415e-01 7.32283771e-01 1.52658367e+00
9.96126980e-02 -1.44605830e-01 -4.92229491e-01 4.90901679e-01
-2.10549831e+00 -1.21130466e+00 -6.37887120e-02 1.42327285e+00
7.68933058e-01 4.06951718e-02 -9.34901237e-02 1.69452563e-01
7.38125920e-01 4.17339861e-01 -5.71326196e-01 -9.10164535e-01
-9.99889299e-02 3.24786365e-01 -2.46516690e-01 3.55836332e-01
-1.25036693e+00 1.11911309e+00 5.81570435e+00 6.94426358e-01
-1.75749493e+00 2.82314479e-01 7.64932752e-01 1.06237710e-01
-2.67307639e-01 -4.18666691e-01 -2.00432777e-01 3.93816233e-01
1.20934558e+00 -1.81351453e-01 4.21333909e-01 8.63951147e-01
5.36235750e-01 3.63493383e-01 -5.76068342e-01 1.14307296e+00
1.78603195e-02 -9.58446383e-01 -4.10368830e-01 -4.96607304e-01
4.70523179e-01 -1.83901638e-01 2.13172823e-01 5.34050226e-01
4.21886235e-01 -1.14491248e+00 7.62022316e-01 4.60861266e-01
7.60584414e-01 -1.15521181e+00 6.90731108e-01 6.39457256e-02
-1.85810459e+00 -3.35775822e-01 2.21481323e-02 -3.51387501e-01
1.21578492e-01 4.43832487e-01 -5.74744821e-01 5.31953752e-01
1.06965995e+00 1.10772824e+00 6.44545928e-02 7.18757883e-02
-4.27033156e-01 7.33831227e-01 8.15838948e-02 -5.09694636e-01
3.67340147e-01 1.67744607e-01 3.19072694e-01 1.54950368e+00
2.05697343e-01 5.09696543e-01 -2.82313284e-02 5.74216664e-01
-9.95644033e-02 2.51206040e-01 -3.40865731e-01 -3.12623739e-01
3.62821758e-01 1.53189981e+00 -4.11593378e-01 -3.84423763e-01
-4.36586469e-01 1.37178743e+00 3.05811167e-01 5.08667946e-01
-1.28981340e+00 -4.79774088e-01 1.13442075e+00 -7.00243354e-01
3.66473705e-01 -7.75276348e-02 -8.15894753e-02 -1.24626803e+00
-5.32607846e-02 -8.28971744e-01 4.38900739e-01 -8.89473736e-01
-1.44958150e+00 1.10106766e+00 -6.67541385e-01 -1.12483108e+00
-1.77039593e-01 -4.11616385e-01 -1.18803501e+00 1.15198123e+00
-1.67086554e+00 -1.21520972e+00 -2.61656106e-01 9.59851921e-01
8.50580752e-01 -8.27947538e-03 1.05848837e+00 1.89996913e-01
-6.93697274e-01 4.52535778e-01 -5.51413238e-01 2.25151375e-01
5.76915979e-01 -1.02249599e+00 2.14705914e-01 8.66138220e-01
-1.62228569e-01 6.82829201e-01 4.14632559e-01 -4.40683126e-01
-1.31130922e+00 -1.14800406e+00 1.01832712e+00 1.85052864e-02
9.68966544e-01 -4.10941482e-01 -8.97784770e-01 4.38104033e-01
4.95217294e-01 2.98625052e-01 9.12667394e-01 3.01662415e-01
-4.21064615e-01 -1.77206825e-02 -8.06987464e-01 4.30553049e-01
6.95859075e-01 -1.07648087e+00 -6.88462794e-01 -1.64204195e-01
1.17136526e+00 -2.21960723e-01 -9.33260381e-01 4.45933729e-01
5.79973578e-01 -7.99523294e-01 7.07414508e-01 -6.87538385e-01
5.36072254e-01 -2.66514987e-01 -2.52874225e-01 -1.51989722e+00
-3.34578425e-01 -8.34109485e-01 -1.55290008e-01 1.15717697e+00
3.03411841e-01 -6.44734085e-01 1.88098595e-01 1.26126111e-01
-3.68668497e-01 -1.02925801e+00 -8.65818560e-01 -2.63829082e-01
-3.28986645e-01 -9.15078700e-01 7.31233776e-01 1.22716069e+00
4.08610493e-01 6.76238835e-01 -4.12620276e-01 4.76595372e-01
-8.59053582e-02 1.23182349e-01 1.37761131e-01 -1.10431206e+00
-3.62685025e-02 -5.28192043e-01 -2.41345137e-01 -8.56680989e-01
6.10677481e-01 -6.06788218e-01 1.91558108e-01 -1.14558303e+00
-2.84021795e-01 -1.66707203e-01 -7.85824656e-01 4.02694732e-01
-3.98723900e-01 -8.44078436e-02 -1.80859029e-01 -2.34752640e-01
-6.37535572e-01 1.13938057e+00 1.14303088e+00 4.03585844e-02
-4.74979281e-01 -1.88889503e-01 -5.90811074e-01 8.17619622e-01
6.86362982e-01 -5.99301495e-02 -5.27052760e-01 2.45671324e-03
2.84615546e-01 2.58093923e-01 4.20419365e-01 -6.79285526e-01
2.78359950e-01 -2.42646486e-01 4.69156563e-01 -5.24513721e-01
4.67305541e-01 -7.95913398e-01 -2.22591862e-01 2.95478851e-01
-3.96546930e-01 1.22417249e-01 4.08720970e-01 5.91605723e-01
-8.26218903e-01 4.36525762e-01 6.07319832e-01 7.89814666e-02
-1.13130915e+00 4.40448463e-01 -6.23051047e-01 -1.71280801e-01
9.83005285e-01 2.08319828e-01 3.58849689e-02 -5.33275604e-01
-8.57598782e-01 2.25071490e-01 -4.99777049e-01 1.00442505e+00
8.87145340e-01 -1.60159111e+00 -3.79755378e-01 3.64233166e-01
8.36320668e-02 -4.34650630e-01 6.14659250e-01 8.29927027e-01
2.60101646e-01 2.52480328e-01 1.67084783e-01 -4.34932530e-01
-1.31218565e+00 2.88927585e-01 5.64107418e-01 -1.53888255e-01
-3.59055340e-01 1.26591265e+00 4.38571870e-01 -4.07841176e-01
2.22688481e-01 -4.97222722e-01 -5.38385808e-01 4.05468911e-01
6.60941243e-01 -1.32738456e-01 9.19086412e-02 -8.13138545e-01
-6.37546778e-01 2.31202364e-01 1.57444417e-01 -1.31352141e-01
1.48036766e+00 -2.37744108e-01 -2.51175761e-01 8.45033586e-01
1.47379804e+00 -1.81090206e-01 -1.11175299e+00 -1.20229647e-01
-1.80412054e-01 4.48765419e-02 3.60478133e-01 -8.02390158e-01
-1.12347853e+00 1.05990589e+00 6.12548411e-01 3.28759581e-01
1.71841800e+00 5.54013345e-03 1.07539809e+00 -1.54913411e-01
-4.62103859e-02 -1.17633855e+00 5.74578285e-01 8.67149293e-01
1.16844046e+00 -1.02889347e+00 -6.66127682e-01 -3.30071747e-01
-1.00544441e+00 1.22444081e+00 5.96728861e-01 -6.30516708e-02
9.34952080e-01 3.70164692e-01 3.38952631e-01 -3.08785200e-01
-1.23048997e+00 -2.12732941e-01 3.89444321e-01 1.94160372e-01
7.90634453e-01 2.39502296e-01 -8.55786875e-02 1.22863472e+00
-2.24938154e-01 -4.72915210e-02 -6.89488798e-02 5.31105161e-01
-1.38577476e-01 -8.78562868e-01 -1.25357568e-01 1.06826052e-01
-5.35748303e-01 -2.68559754e-01 -3.87578309e-01 2.76309878e-01
1.61179796e-01 1.29644549e+00 2.12622672e-01 -1.01325512e+00
3.73008758e-01 1.97893515e-01 -1.60290837e-01 -4.38091993e-01
-9.97832716e-01 1.58687532e-01 9.32574794e-02 -7.40487695e-01
-5.38700283e-01 -5.69139838e-01 -1.83287787e+00 2.24901699e-02
-1.89815372e-01 3.11403722e-01 5.07009566e-01 9.94055629e-01
4.39024359e-01 1.25831056e+00 1.21676874e+00 -5.44700563e-01
-1.08080521e-01 -1.08184016e+00 -4.18425888e-01 4.63178754e-01
7.01312900e-01 -6.24839187e-01 -4.34698880e-01 1.71884581e-01] | [13.417219161987305, 5.7325053215026855] |
464b163e-67ad-4b20-802e-8cf953584d7e | micro-expression-detection-in-long-videos | 1903.10765 | null | http://arxiv.org/abs/1903.10765v1 | http://arxiv.org/pdf/1903.10765v1.pdf | Micro-expression detection in long videos using optical flow and recurrent neural networks | Facial micro-expressions are subtle and involuntary expressions that can
reveal concealed emotions. Micro-expressions are an invaluable source of
information in application domains such as lie detection, mental health,
sentiment analysis and more. One of the biggest challenges in this field of
research is the small amount of available spontaneous micro-expression data.
However, spontaneous data collection is burdened by time-consuming and
expensive annotation. Hence, methods are needed which can reduce the amount of
data that annotators have to review. This paper presents a novel
micro-expression spotting method using a recurrent neural network (RNN) on
optical flow features. We extract Histogram of Oriented Optical Flow (HOOF)
features to encode the temporal changes in selected face regions. Finally, the
RNN spots short intervals which are likely to contain occurrences of relevant
facial micro-movements. The proposed method is evaluated on the SAMM database.
Any chance of subject bias is eliminated by training the RNN using
Leave-One-Subject-Out cross-validation. Comparing the spotted intervals with
the labeled data shows that the method produced 1569 false positives while
obtaining a recall of 0.4654. The initial results show that the proposed method
would reduce the video length by a factor of 3.5, while still retaining almost
half of the relevant micro-movements. Lastly, as the model gets more data, it
becomes better at detecting intervals, which makes the proposed method suitable
for supporting the annotation process. | ['Michiel Verburg', 'Vlado Menkovski'] | 2019-03-26 | null | null | null | null | ['micro-expression-spotting'] | ['computer-vision'] | [ 2.92725861e-01 1.49775803e-01 -2.50474393e-01 -6.06866121e-01
-3.85618418e-01 -2.18596324e-01 3.31646413e-01 -1.52311064e-02
-5.00129759e-01 7.86173940e-01 1.64680585e-01 2.64970839e-01
1.59535766e-01 -3.71935666e-01 -1.81147307e-01 -8.32467020e-01
-2.25137800e-01 -4.21284765e-01 -1.32892296e-01 -2.32637450e-01
2.79962629e-01 6.63495541e-01 -2.00503659e+00 4.78775948e-01
4.06374484e-01 1.08532774e+00 -2.05428809e-01 3.74071151e-01
1.58324204e-02 1.23285830e+00 -8.52898300e-01 -5.04277766e-01
-3.12456656e-02 -7.63492942e-01 -6.21831656e-01 1.19317554e-01
3.08565676e-01 -2.75809467e-01 3.05845626e-02 1.03943932e+00
6.48531139e-01 3.27794194e-01 2.93076813e-01 -1.19148195e+00
-2.13395283e-01 -1.53846936e-02 -8.56037140e-01 4.84564483e-01
5.42586327e-01 -1.93095282e-01 6.44038081e-01 -8.98264647e-01
9.14230883e-01 1.08600068e+00 5.45118570e-01 6.51593447e-01
-7.98554897e-01 -8.49363804e-01 -2.82366157e-01 4.31682348e-01
-1.45897198e+00 -7.48917043e-01 9.43652332e-01 -3.49533290e-01
8.31289709e-01 3.43824387e-01 6.21297240e-01 1.06536245e+00
1.19364083e-01 4.87327635e-01 1.17952776e+00 -4.36960846e-01
1.05294082e-02 4.38793093e-01 -9.03493688e-02 7.86212325e-01
-6.39619231e-02 -4.73570600e-02 -7.10015059e-01 -3.85321327e-04
5.74862599e-01 -3.72673348e-02 -2.57469654e-01 1.57062113e-01
-7.26539433e-01 8.04651558e-01 -6.17864123e-03 6.41498208e-01
-5.36670625e-01 -2.09524676e-01 8.50093424e-01 3.18028152e-01
8.83811891e-01 2.41520137e-01 -9.32464749e-02 -5.87007403e-01
-1.20358837e+00 -6.21251122e-04 4.68978375e-01 3.54166985e-01
6.61273539e-01 2.62082964e-01 1.55206202e-02 9.75309193e-01
-1.47598922e-01 3.77645373e-01 6.67256534e-01 -9.06470060e-01
3.27013694e-02 6.59550190e-01 2.96104401e-02 -1.59790719e+00
-5.30029774e-01 -1.53246328e-01 -8.45278919e-01 2.53282547e-01
2.24953115e-01 -3.56981754e-01 -4.25429404e-01 1.73723876e+00
3.76001984e-01 7.23454654e-02 -1.97119892e-01 1.01570654e+00
6.18413210e-01 7.16599107e-01 -5.46580479e-02 -8.90929759e-01
1.47472954e+00 -5.94933987e-01 -1.27096725e+00 1.33510098e-01
6.87299073e-01 -8.19082916e-01 7.29764819e-01 3.97276729e-01
-8.00042212e-01 -5.20305932e-01 -8.59302640e-01 2.51828581e-01
-1.94139794e-01 4.29689318e-01 5.92120290e-01 6.90432489e-01
-9.29539502e-01 5.80058753e-01 -5.53187668e-01 -4.47885841e-01
3.88756871e-01 4.19946015e-01 -6.30853772e-01 4.48180318e-01
-1.09330451e+00 7.78165340e-01 6.10294007e-02 4.56556112e-01
-3.32522631e-01 -1.98914677e-01 -9.12013173e-01 -3.39193135e-01
2.53961861e-01 2.67917395e-01 1.11316919e+00 -1.55192053e+00
-1.49303401e+00 1.11446333e+00 -7.14468122e-01 -4.37039077e-01
3.20126593e-01 -2.54622877e-01 -7.84023762e-01 5.98054767e-01
4.82235029e-02 6.48833215e-01 1.04757833e+00 -6.27575934e-01
-4.14353251e-01 -4.43044066e-01 -2.46654436e-01 7.42814615e-02
-3.66487145e-01 6.50210261e-01 -1.80324405e-01 -5.94006300e-01
-1.94277078e-01 -1.03234923e+00 2.16003776e-01 -9.22736898e-02
-2.05540329e-01 -3.05155843e-01 1.13680375e+00 -7.51499712e-01
1.30136001e+00 -2.38290048e+00 -2.93434203e-01 2.17094287e-01
1.53438002e-01 4.60702300e-01 4.52631339e-02 1.63490951e-01
-4.77273732e-01 1.73618458e-02 1.51584163e-01 -2.19385788e-01
-5.75567365e-01 -1.17108142e-02 -8.97828341e-02 7.33389497e-01
1.82034597e-01 5.81042945e-01 -6.23643041e-01 -5.13847172e-01
2.08734080e-01 7.23840058e-01 -2.41459340e-01 2.24863634e-01
2.94097185e-01 5.14773130e-01 -1.21058993e-01 7.04118669e-01
4.34679627e-01 1.48749352e-01 -1.37820214e-01 -2.36465499e-01
-1.00942612e-01 -1.34556098e-02 -1.01615047e+00 1.13532090e+00
-3.94010484e-01 1.31822300e+00 -4.80117202e-02 -8.67238700e-01
1.23062217e+00 6.33135140e-01 7.42311358e-01 -7.55376279e-01
4.73030984e-01 1.53971696e-03 -1.49181515e-01 -9.37042952e-01
4.73949939e-01 -1.41111329e-01 1.62651300e-01 5.43753386e-01
-1.19610548e-01 5.19798875e-01 2.78112411e-01 -1.11292303e-01
8.00912619e-01 4.44203541e-02 5.20992696e-01 1.69190288e-01
7.48249710e-01 -4.54793394e-01 7.77149796e-01 7.79073387e-02
-5.71094751e-01 2.10720614e-01 7.36804545e-01 -6.77405953e-01
-8.05996537e-01 -1.67179719e-01 -2.59964671e-02 9.58729029e-01
-2.41780698e-01 -4.33043182e-01 -9.06307697e-01 -4.96463120e-01
-4.86424655e-01 4.66926068e-01 -9.14756954e-01 -1.17998838e-01
-3.91977727e-01 -6.56460583e-01 7.11543679e-01 2.29798555e-01
5.57325184e-01 -1.62341917e+00 -1.21249020e+00 -3.43703898e-03
-4.16466862e-01 -1.20792723e+00 -1.95749700e-01 -2.21400619e-01
-7.43548036e-01 -1.02813196e+00 -6.66884005e-01 -5.28838694e-01
7.85629392e-01 3.03272545e-01 6.58657670e-01 -6.34676032e-03
-6.18826926e-01 9.63811874e-02 -4.88122433e-01 -4.77891207e-01
-2.42789909e-01 -3.15506607e-01 2.91422188e-01 5.09411156e-01
8.97692621e-01 -4.12566990e-01 -4.19730425e-01 2.75086850e-01
-5.50413549e-01 -2.54318684e-01 3.84715527e-01 7.46402979e-01
5.38908124e-01 2.78752800e-02 6.50369763e-01 -8.39501739e-01
7.96429873e-01 -1.95500404e-01 -4.19566661e-01 -1.21484429e-01
-3.53502572e-01 -3.04733843e-01 4.99417454e-01 -5.83223164e-01
-1.18374717e+00 1.75791979e-02 1.01284802e-01 -4.71559852e-01
-2.70385325e-01 2.17754498e-01 2.26895750e-01 -1.42839000e-01
7.62014985e-01 -5.79896793e-02 4.16794211e-01 -9.15600732e-02
-1.15029573e-01 7.75907993e-01 3.38446856e-01 6.86781555e-02
1.89391136e-01 5.37595689e-01 5.09603173e-02 -1.35134041e+00
-8.65226924e-01 -6.56148076e-01 -5.38312972e-01 -6.15347743e-01
8.34467769e-01 -6.85776412e-01 -1.00214243e+00 3.73497576e-01
-1.16922319e+00 1.32708520e-01 -8.98919255e-02 5.34412980e-01
-4.44358975e-01 3.66028816e-01 -5.12250960e-01 -1.26675582e+00
-5.64791024e-01 -9.98155892e-01 8.89946342e-01 2.96134263e-01
-6.99695051e-01 -5.73630929e-01 2.17718165e-03 3.40318948e-01
2.74309814e-01 4.53247219e-01 3.30990851e-01 -5.54392099e-01
6.23364598e-02 -5.83682299e-01 -1.11796707e-01 4.67418134e-01
3.70298982e-01 1.16057873e-01 -1.22363424e+00 8.47796425e-02
2.90623218e-01 -4.12252426e-01 4.15300757e-01 3.37135583e-01
1.25886941e+00 -5.51780105e-01 1.25821978e-01 3.14146817e-01
1.08666885e+00 5.66345751e-01 7.96263516e-01 2.24841163e-01
4.81230587e-01 1.01045620e+00 8.15299213e-01 6.95447624e-01
-2.08793551e-01 7.52996683e-01 2.83992976e-01 -2.07417771e-01
2.08761200e-01 7.70499855e-02 5.92727244e-01 5.07915854e-01
-4.72376198e-01 1.60299182e-01 -4.77842897e-01 4.88533527e-01
-1.67945528e+00 -1.60013771e+00 -7.53270835e-02 2.04459476e+00
7.56266773e-01 -1.45581126e-01 2.66514778e-01 4.86290962e-01
7.74998486e-01 3.28233004e-01 -3.19451094e-01 -9.32845116e-01
3.57622392e-02 1.52119279e-01 1.27778679e-01 2.54078239e-01
-1.12256992e+00 7.42643178e-01 5.43242168e+00 6.56719506e-01
-1.58468211e+00 -1.26842946e-01 7.56024122e-01 -2.80725479e-01
3.21529984e-01 -4.88903016e-01 -6.68796778e-01 5.38974524e-01
1.20872009e+00 -6.06032386e-02 2.55232155e-01 8.85636628e-01
8.86895239e-01 -4.84484464e-01 -6.72785938e-01 1.43346059e+00
5.14088929e-01 -1.05018818e+00 -2.51028240e-01 2.67019980e-02
4.05282348e-01 -4.23238546e-01 -1.35854498e-01 -1.14614107e-01
-6.11839116e-01 -1.07050037e+00 2.94321358e-01 4.86018062e-01
9.47598219e-01 -9.96998906e-01 1.14428508e+00 8.97376910e-02
-9.58196640e-01 4.44139689e-02 -2.83618003e-01 -2.31770262e-01
4.52560931e-02 6.61439955e-01 -9.70712900e-01 1.03777938e-01
7.13340402e-01 7.24181771e-01 -4.08881634e-01 7.01840222e-01
-2.83848047e-01 6.82474256e-01 -1.61732346e-01 -3.47258210e-01
1.85636431e-02 -2.18658403e-01 4.70886856e-01 1.36453950e+00
2.60556221e-01 9.44650769e-02 -3.78160715e-01 4.44261193e-01
-5.59185036e-02 5.26182532e-01 -1.00080967e+00 -8.01744834e-02
2.24855170e-01 1.51255596e+00 -8.12031448e-01 -1.86182663e-01
-2.93092787e-01 1.03750479e+00 3.59427035e-02 2.21162915e-01
-7.37560868e-01 -5.67135513e-01 4.36579496e-01 1.27634868e-01
-6.69783950e-02 1.24648765e-01 8.13833773e-02 -8.69536579e-01
2.20704213e-01 -9.79751706e-01 2.77773798e-01 -8.15222681e-01
-7.62714028e-01 8.70615542e-01 -1.40823081e-01 -1.32144356e+00
-8.10455561e-01 -3.25766534e-01 -4.80670750e-01 5.99484205e-01
-1.06767631e+00 -7.91653633e-01 -5.18204570e-01 6.08135223e-01
6.80870652e-01 -1.76613063e-01 1.05222535e+00 4.38281298e-01
-6.96830690e-01 7.47133136e-01 -4.73389924e-01 3.66099238e-01
9.41211820e-01 -6.46572113e-01 -4.26061243e-01 8.56907606e-01
1.99264795e-01 6.36195242e-01 8.27437580e-01 -4.17813778e-01
-7.50253022e-01 -9.11751390e-01 1.14812589e+00 6.55198842e-02
3.91938031e-01 -1.28230169e-01 -7.91103184e-01 4.57909942e-01
1.31469622e-01 1.65088937e-01 9.74337637e-01 -1.21351995e-01
-1.45855963e-01 -3.21834445e-01 -1.29193485e+00 3.52709830e-01
5.39515793e-01 -7.06623733e-01 -3.48988086e-01 1.35417640e-01
1.51251301e-01 -1.66871488e-01 -6.17035329e-01 2.35773951e-01
7.91180849e-01 -1.24246812e+00 4.59000766e-01 -3.49483997e-01
4.90399718e-01 -4.21770513e-02 1.39946654e-01 -1.05148828e+00
1.72139019e-01 -8.47045004e-01 3.87745760e-02 1.36697364e+00
3.09614539e-01 -3.71771514e-01 8.47993731e-01 6.29088759e-01
3.57720137e-01 -8.48220110e-01 -9.89898801e-01 -4.39152092e-01
-8.54673624e-01 -3.81388396e-01 1.21275857e-01 1.06628728e+00
2.67540097e-01 2.91753203e-01 -7.91794538e-01 -4.08189118e-01
3.17532033e-01 -1.77351907e-01 6.97523475e-01 -1.10608828e+00
2.61941761e-01 -1.84972882e-01 -8.16067636e-01 -2.99181312e-01
2.93946087e-01 -3.91228080e-01 -9.77745801e-02 -8.44458759e-01
6.28652871e-02 1.79726079e-01 -8.66916403e-02 6.12095654e-01
9.40942690e-02 5.59070706e-01 -2.32005510e-02 1.52285546e-01
-3.95647645e-01 4.07270670e-01 9.57580388e-01 1.47529483e-01
-4.74335551e-01 4.56879660e-02 -3.88850927e-01 1.03202784e+00
9.25065339e-01 -3.74824107e-01 -5.69194674e-01 2.85652012e-01
1.55730262e-01 1.07166961e-01 1.89110950e-01 -7.66750097e-01
5.02498401e-03 -2.13684309e-02 4.98693973e-01 -5.24367690e-01
5.16052067e-01 -7.24324882e-01 -8.16959143e-02 2.56336004e-01
-3.30044210e-01 3.00779074e-01 3.23302597e-01 2.72926122e-01
-5.06188273e-01 -3.99117678e-01 8.74206722e-01 -1.23326309e-01
-9.05424356e-01 6.16226383e-02 -5.69647729e-01 -2.40373239e-01
1.23889267e+00 -4.51402307e-01 5.18632606e-02 -7.80279398e-01
-6.27504885e-01 -3.98280382e-01 1.57333419e-01 4.21530873e-01
7.05038488e-01 -1.11954641e+00 -4.17180806e-01 3.17617506e-01
3.24997157e-02 -4.83447522e-01 2.76594192e-01 9.99795616e-01
-4.16920453e-01 3.52129370e-01 -4.18949097e-01 -5.07081330e-01
-1.94474697e+00 2.58925498e-01 2.42133528e-01 3.02446820e-02
-6.50643766e-01 7.51807213e-01 -1.83590397e-01 2.26638198e-01
2.49339640e-01 8.91531706e-02 -8.35297227e-01 7.16418564e-01
9.39553738e-01 5.35686314e-01 7.13109411e-03 -1.14138424e+00
-3.52220684e-01 5.50805032e-01 -1.40728652e-01 -6.22428320e-02
1.33626556e+00 -1.60467446e-01 -4.45397317e-01 5.58476031e-01
1.46965206e+00 2.30164245e-01 -8.58541965e-01 1.61305755e-01
1.61903221e-02 -4.59624738e-01 -2.39617229e-02 -5.20563900e-01
-1.09721279e+00 9.92722988e-01 8.06675255e-01 3.20501849e-02
1.45613217e+00 -3.95043045e-01 5.65668166e-01 3.36163670e-01
2.06062198e-01 -1.27558374e+00 7.41347745e-02 1.29606694e-01
8.23309541e-01 -1.10303199e+00 -1.22974999e-02 -3.12868953e-01
-8.77883911e-01 1.32706523e+00 5.14689982e-01 5.66823855e-02
4.16967154e-01 1.08388051e-01 3.34986418e-01 -4.15301472e-01
-6.77272499e-01 2.68298723e-02 1.83545664e-01 3.63039911e-01
6.90112054e-01 -1.07713662e-01 -5.37225842e-01 2.57040471e-01
-2.29906037e-01 3.22814584e-01 7.22176611e-01 7.51283586e-01
-3.12214702e-01 -6.45217478e-01 -4.07123059e-01 4.81963485e-01
-1.08577812e+00 2.77973413e-01 -3.69408429e-01 7.23179698e-01
7.87863955e-02 9.84179676e-01 1.28286868e-01 -2.98369855e-01
1.52274281e-01 2.71318704e-01 1.19839884e-01 -3.13699126e-01
-3.17424238e-01 1.39502019e-01 3.34162533e-01 -8.84542048e-01
-1.06621480e+00 -5.69512784e-01 -1.21708775e+00 -2.50718057e-01
-2.27789015e-01 2.11304381e-01 5.44624269e-01 8.45472336e-01
3.88063461e-01 3.04307699e-01 8.23669612e-01 -6.72184408e-01
-1.03688389e-02 -1.04513705e+00 -5.24995327e-01 6.40553534e-01
4.27976161e-01 -7.71938086e-01 -4.16930884e-01 3.45851719e-01] | [13.61422348022461, 1.8958269357681274] |
c1ea68f0-0f39-4af4-b452-90e48e4077b3 | a-large-scale-homography-benchmark | 2302.09997 | null | https://arxiv.org/abs/2302.09997v1 | https://arxiv.org/pdf/2302.09997v1.pdf | A Large Scale Homography Benchmark | We present a large-scale dataset of Planes in 3D, Pi3D, of roughly 1000 planes observed in 10 000 images from the 1DSfM dataset, and HEB, a large-scale homography estimation benchmark leveraging Pi3D. The applications of the Pi3D dataset are diverse, e.g. training or evaluating monocular depth, surface normal estimation and image matching algorithms. The HEB dataset consists of 226 260 homographies and includes roughly 4M correspondences. The homographies link images that often undergo significant viewpoint and illumination changes. As applications of HEB, we perform a rigorous evaluation of a wide range of robust estimators and deep learning-based correspondence filtering methods, establishing the current state-of-the-art in robust homography estimation. We also evaluate the uncertainty of the SIFT orientations and scales w.r.t. the ground truth coming from the underlying homographies and provide codes for comparing uncertainty of custom detectors. The dataset is available at \url{https://github.com/danini/homography-benchmark}. | ['Jiri Matas', 'Wolfgang Förstner', 'Michal Polic', 'Dmytro Mishkin', 'Daniel Barath'] | 2023-02-20 | null | null | null | null | ['homography-estimation'] | ['computer-vision'] | [-1.67623475e-01 -2.28669927e-01 4.98973532e-03 -2.21150130e-01
-1.04302061e+00 -7.11299837e-01 7.76372254e-01 -3.40531170e-01
2.52623092e-02 1.29616246e-01 2.76143074e-01 2.67292142e-01
-1.51802883e-01 -7.15796649e-01 -1.27570152e+00 -4.00824159e-01
-1.47199556e-01 9.20143187e-01 3.62336129e-01 -1.49608016e-01
6.51760101e-01 7.16403961e-01 -1.49946511e+00 -1.62462797e-02
2.81263322e-01 1.16640866e+00 -1.36563346e-01 5.97053945e-01
5.48970401e-01 -7.07006529e-02 -3.04978371e-01 -6.04163945e-01
9.76985931e-01 5.93990050e-02 -3.79425496e-01 2.69837558e-01
1.59549534e+00 -7.75880814e-01 -9.16782260e-01 1.17993772e+00
5.07565677e-01 -1.76181257e-01 7.07382381e-01 -1.45630157e+00
-4.91816640e-01 -8.93158987e-02 -6.56226575e-01 -3.13104577e-02
6.73960745e-01 3.40893209e-01 9.59327638e-01 -1.08888423e+00
1.22994971e+00 1.51396394e+00 8.56955051e-01 1.31246924e-01
-1.09452283e+00 -6.33932292e-01 -4.25234735e-01 1.48434862e-01
-1.49939454e+00 -4.73151684e-01 5.49050629e-01 -6.74150407e-01
1.14993203e+00 -1.13965228e-01 6.37125373e-01 1.19217288e+00
4.84244585e-01 5.55649936e-01 7.48634219e-01 -2.79198855e-01
-7.59511590e-02 -3.03637475e-01 -8.44269991e-02 7.65462220e-01
6.25972986e-01 6.08452260e-01 -1.05576050e+00 -1.54126570e-01
1.41387451e+00 -1.60925776e-01 -3.50249410e-01 -1.08871245e+00
-1.51974237e+00 4.71033543e-01 4.32299435e-01 -5.42555392e-01
-1.04138497e-02 2.25037307e-01 1.06967859e-01 5.17900825e-01
1.91696465e-01 4.92228657e-01 -2.00900406e-01 -5.31166419e-02
-6.19616747e-01 4.58152145e-01 7.11034119e-01 1.88569522e+00
1.03212690e+00 -3.89952719e-01 4.19480443e-01 4.96812046e-01
2.57902950e-01 9.13408101e-01 -6.51969388e-02 -1.47298026e+00
6.70436561e-01 3.45934331e-01 8.42916071e-02 -1.32382905e+00
-2.76319355e-01 -2.16786247e-02 -6.20333970e-01 3.25890988e-01
5.01888514e-01 4.16386396e-01 -7.46531367e-01 1.13277698e+00
3.59805495e-01 -2.30513215e-02 -3.72927725e-01 9.49274838e-01
7.20540285e-01 1.52592987e-01 -1.02078104e+00 5.18747568e-01
1.03714705e+00 -7.78060973e-01 -1.86911985e-01 -2.69520819e-01
2.83014923e-01 -1.28617871e+00 6.91078246e-01 6.54236436e-01
-1.18352151e+00 -3.91212016e-01 -1.28315234e+00 -5.92231810e-01
-2.85858184e-01 6.23711338e-03 4.64261591e-01 1.51457667e-01
-1.00832820e+00 8.48319769e-01 -7.21560836e-01 -5.00161052e-01
2.69328415e-01 2.96488941e-01 -6.95746839e-01 -3.51242453e-01
-7.91745186e-01 7.58328199e-01 2.36163866e-02 -2.27913216e-01
-7.66902030e-01 -8.95371675e-01 -9.96690989e-01 -2.57581472e-01
1.83834195e-01 -8.92821848e-01 1.16579008e+00 -2.29554176e-01
-1.10492063e+00 1.49701631e+00 6.92257006e-03 -3.10368478e-01
9.37935472e-01 -5.57602942e-01 -1.56163394e-01 3.20655316e-01
1.61325872e-01 7.16900051e-01 8.15743744e-01 -1.02984560e+00
-5.68918526e-01 -5.89251101e-01 -2.78067291e-01 1.87233761e-01
3.68356675e-01 -3.37162971e-01 -8.03538799e-01 -3.94515514e-01
6.56739533e-01 -9.20839071e-01 1.60230532e-01 3.59393805e-01
-6.42753720e-01 2.32757568e-01 4.17619646e-01 -5.72825253e-01
5.43266773e-01 -2.18781447e+00 1.27758637e-01 4.46777999e-01
3.58338594e-01 -4.74699944e-01 -5.73013760e-02 6.29810989e-01
1.02414638e-01 -2.66571432e-01 2.17008427e-01 -4.68695134e-01
3.90786141e-01 -1.86798185e-01 -3.31771523e-01 1.12256896e+00
-1.49837196e-01 7.02446580e-01 -5.79261482e-01 -1.12361513e-01
6.33143544e-01 5.13787031e-01 -5.63100815e-01 1.46750450e-01
-5.15755266e-02 2.53451675e-01 1.27414716e-02 9.77473497e-01
1.11664283e+00 -2.72220910e-01 -2.82960594e-01 -7.97171950e-01
-1.79260492e-01 3.70068252e-01 -1.39605737e+00 1.93555462e+00
-5.52558377e-02 1.03768575e+00 -9.52373166e-03 -2.36105680e-01
9.34296131e-01 -2.59407163e-01 6.41901255e-01 -6.77173615e-01
2.88360924e-01 4.72440541e-01 -2.89143145e-01 1.23982839e-01
6.12634897e-01 5.57994962e-01 9.76616964e-02 -1.82911567e-02
3.16016436e-01 -8.43700051e-01 2.26844884e-02 1.91862375e-01
1.11208582e+00 2.10477293e-01 4.68798399e-01 -4.47471797e-01
8.40037838e-02 7.42777213e-02 3.67144942e-01 6.18531764e-01
1.11294650e-02 1.10979068e+00 3.20017040e-01 -5.37757277e-01
-1.57126749e+00 -1.43585503e+00 -5.18427372e-01 -4.83049788e-02
6.65749013e-01 -5.21409810e-01 -3.80141467e-01 6.42093420e-02
7.21068680e-01 -9.15959030e-02 -4.58447337e-01 -1.20452002e-01
-2.72413492e-01 -1.35106474e-01 1.65250391e-01 3.58352482e-01
7.73949265e-01 -3.40078861e-01 -3.95717263e-01 -2.17232123e-01
2.09859863e-01 -1.46252167e+00 -7.12979972e-01 -2.76716501e-01
-8.12004983e-01 -1.51877022e+00 -4.57008064e-01 -5.16159654e-01
6.66956902e-01 4.94367599e-01 1.60165226e+00 -1.48522601e-01
-5.51454127e-01 7.89114833e-01 1.71779282e-02 4.85906377e-02
-1.47878945e-01 -1.91101551e-01 1.84675068e-01 -4.48969215e-01
5.78615487e-01 -6.51745141e-01 -8.27253401e-01 9.72128570e-01
-4.10553604e-01 -1.03256190e-02 4.66734380e-01 6.44748569e-01
9.18121636e-01 -2.35219628e-01 -5.56059241e-01 -4.75675315e-01
-1.27262607e-01 -1.60810158e-01 -1.58322036e+00 5.37889600e-02
-4.18037385e-01 -1.01685345e-01 -2.61272602e-02 -1.31808385e-01
-5.48558354e-01 1.30833328e-01 1.44212157e-01 -8.68090987e-01
-9.48022753e-02 -1.76759988e-01 -3.40993762e-01 -7.10271716e-01
7.14904904e-01 -1.23277381e-01 -9.67353806e-02 -5.38993299e-01
3.49314064e-01 2.33649746e-01 9.94893551e-01 -5.25700510e-01
1.25437236e+00 9.25486565e-01 4.91092741e-01 -9.29151475e-01
-5.25395691e-01 -8.35052907e-01 -1.01690149e+00 -5.45536093e-02
5.82665563e-01 -1.23679864e+00 -7.78693616e-01 8.10958564e-01
-1.23171473e+00 -3.20665896e-01 3.07273655e-03 6.49340212e-01
-9.68102396e-01 5.21590292e-01 -5.98921716e-01 -2.09964700e-02
-1.18743658e-01 -1.34705484e+00 1.64288509e+00 -4.40222882e-02
-1.80436179e-01 -7.76193798e-01 1.06739081e-01 2.49251589e-01
-1.38933644e-01 4.02457744e-01 5.21560311e-01 -3.05834673e-02
-1.46578991e+00 -2.88982838e-01 -3.26490521e-01 1.13389172e-01
1.21116443e-02 3.05795312e-01 -8.36608589e-01 -4.82900828e-01
-1.82246745e-01 -2.57502973e-01 7.18826413e-01 5.34187615e-01
5.80797493e-01 5.21715879e-02 -2.35582873e-01 1.47511303e+00
1.59453273e+00 -1.74186721e-01 8.40299189e-01 8.40332687e-01
8.67585719e-01 5.06680489e-01 7.30433643e-01 5.20519853e-01
2.59996891e-01 1.00011349e+00 5.42153239e-01 1.24500342e-01
-3.65596145e-01 -4.88752425e-01 1.01302557e-01 5.79130709e-01
5.70590533e-02 6.31128550e-02 -9.47716892e-01 3.75141263e-01
-1.57277536e+00 -6.11545026e-01 -2.67093807e-01 2.63580728e+00
5.18749714e-01 1.48443550e-01 -1.10901386e-01 -2.74145961e-01
6.14720762e-01 8.31386223e-02 -7.69601405e-01 2.13182926e-01
-3.79276842e-01 -4.83110808e-02 1.12492681e+00 7.05493808e-01
-1.29314911e+00 1.01368010e+00 6.16322899e+00 5.46370864e-01
-8.21312070e-01 -4.07350153e-01 4.09881771e-02 -7.38751590e-02
-1.34432450e-01 9.72566307e-02 -1.33817911e+00 1.67664856e-01
3.22587013e-01 -2.65827805e-01 5.51127911e-01 8.69848251e-01
-3.42230141e-01 -1.89036787e-01 -1.47248518e+00 1.45594859e+00
3.10421795e-01 -1.56450093e+00 2.05954053e-02 4.30878550e-01
1.13066053e+00 5.87207854e-01 1.94837496e-01 -3.51568252e-01
3.41939807e-01 -6.06815755e-01 7.45200336e-01 4.67350215e-01
9.94637132e-01 -5.01277626e-01 4.42012638e-01 -1.64124459e-01
-1.11264932e+00 4.74150091e-01 -8.80571604e-01 3.06655139e-01
1.93389624e-01 6.50541186e-01 -5.93044400e-01 5.41942000e-01
1.03809273e+00 1.24368656e+00 -7.24321425e-01 1.35216630e+00
-2.70623058e-01 -1.33167624e-01 -6.05654597e-01 5.47118485e-01
-7.54603371e-02 -4.07288969e-01 7.79829085e-01 7.34657943e-01
5.24749577e-01 -1.06746212e-01 -9.87109169e-02 8.37886870e-01
-3.43749046e-01 -2.99276203e-01 -9.56641495e-01 2.92995393e-01
6.60447836e-01 9.58744884e-01 -4.77937579e-01 -1.05355181e-01
-4.51963961e-01 9.22857463e-01 -3.58783267e-02 2.84531623e-01
-5.55002272e-01 -2.92230964e-01 1.14685166e+00 3.10192972e-01
2.60127157e-01 -5.75489283e-01 -1.65463328e-01 -1.42213333e+00
1.80472881e-01 -9.25800323e-01 1.08064897e-01 -1.03842092e+00
-1.49156427e+00 2.85994142e-01 1.85478091e-01 -1.58362257e+00
-1.33099079e-01 -1.12787497e+00 -3.32327336e-01 7.13933289e-01
-1.25272346e+00 -9.01521504e-01 -8.73867691e-01 5.87610483e-01
4.02714580e-01 -2.88101524e-01 2.66194016e-01 3.02785695e-01
9.04348865e-03 5.11751056e-01 3.41956377e-01 5.98580316e-02
1.13379991e+00 -1.14692724e+00 1.23839474e+00 8.70228887e-01
1.38567820e-01 6.67464256e-01 7.87616014e-01 -7.73383260e-01
-1.92861772e+00 -6.47963524e-01 4.17548031e-01 -1.04604852e+00
8.17237139e-01 -6.61686182e-01 -6.25254273e-01 1.04121518e+00
-1.10817470e-01 3.16358581e-02 -3.20803002e-02 -2.71878809e-01
-5.77711463e-01 -1.23087399e-01 -1.08213365e+00 5.27487457e-01
1.60751379e+00 -7.28093326e-01 -3.52159500e-01 5.53010225e-01
5.77497423e-01 -1.14127672e+00 -1.15962183e+00 5.26020467e-01
7.99332559e-01 -1.43883526e+00 1.48036993e+00 -9.73384529e-02
4.80012894e-01 -1.94971174e-01 -5.75969696e-01 -1.05450201e+00
-2.27972940e-01 -7.72172391e-01 6.20527752e-02 8.16472054e-01
-5.20265587e-02 -7.46971071e-01 9.62158382e-01 4.76371258e-01
-1.99870899e-01 -3.16900492e-01 -1.00893068e+00 -1.15665162e+00
-1.92273021e-01 -3.39234364e-03 6.07697368e-01 7.96408355e-01
-5.15624166e-01 -1.27458528e-01 -3.50970626e-01 5.30686021e-01
1.44357777e+00 2.49389753e-01 1.47598839e+00 -1.05546093e+00
-3.81590128e-01 -4.98207062e-01 -1.05952752e+00 -1.52058733e+00
-7.41693974e-02 -6.09901786e-01 -1.52818173e-01 -1.04885936e+00
-1.26405060e-01 -1.28122037e-02 5.14358878e-01 -4.02237847e-02
5.06981671e-01 2.16320917e-01 -1.01415627e-01 5.13005078e-01
-2.33153984e-01 4.50461417e-01 1.17621291e+00 -1.03907533e-01
1.51678041e-01 -9.17201117e-02 1.07231669e-01 7.77249873e-01
5.57480991e-01 -1.16924167e-01 -1.65204689e-01 -7.71136045e-01
2.80445069e-01 -7.61010870e-02 6.15890861e-01 -1.14077997e+00
4.55454201e-01 1.38040975e-01 5.65902531e-01 -1.10058975e+00
5.74007332e-01 -7.96955764e-01 5.22490442e-01 2.27612242e-01
3.42070535e-02 3.75600338e-01 2.09543034e-01 3.46308410e-01
-1.86911240e-01 1.92989439e-01 7.78145134e-01 -2.05147818e-01
-9.56946194e-01 6.79702222e-01 4.78017837e-01 3.52752239e-01
5.53824127e-01 -4.95886385e-01 -8.45506489e-01 -3.83405238e-01
-1.73606321e-01 8.15614462e-02 1.38626218e+00 4.72575545e-01
8.52217734e-01 -1.44361746e+00 -4.20033216e-01 5.76306522e-01
4.14419413e-01 1.68493092e-01 6.97643608e-02 6.40232623e-01
-1.05192459e+00 3.49631190e-01 -5.64073861e-01 -9.66106594e-01
-1.11666799e+00 2.89986581e-01 4.67627287e-01 3.87094527e-01
-9.19377744e-01 8.05319905e-01 4.46950883e-01 -3.73670101e-01
4.46097344e-01 -4.05127764e-01 6.32159472e-01 -4.33402777e-01
3.01770329e-01 6.50972962e-01 1.86567664e-01 -6.73503578e-01
-4.17973429e-01 1.29419303e+00 1.73717245e-01 -7.65985921e-02
1.19606626e+00 -1.71362907e-01 -1.05971977e-01 3.43063995e-02
1.50123417e+00 2.78683826e-02 -1.59667635e+00 -1.88456357e-01
-1.14505887e-01 -1.16204262e+00 -5.20489104e-02 -3.53948653e-01
-9.53585505e-01 8.28123510e-01 5.50343037e-01 -5.70315361e-01
7.22827733e-01 1.28115177e-01 6.38918340e-01 5.95146954e-01
9.20815527e-01 -7.61243701e-01 8.24595839e-02 5.06506681e-01
1.14260972e+00 -1.44156623e+00 4.57241774e-01 -7.16356337e-01
-1.58411056e-01 1.27039099e+00 6.40179217e-01 -5.81398427e-01
5.78180850e-01 2.89746165e-01 4.75389101e-02 -5.41212738e-01
-4.07989115e-01 1.39121979e-01 6.01202607e-01 6.75496340e-01
-9.07775238e-02 -2.98242331e-01 3.48944724e-01 -1.90827876e-01
-6.56940520e-01 -4.07390445e-01 4.57362592e-01 6.49885118e-01
-2.00833961e-01 -9.67944801e-01 -5.71459293e-01 4.96032983e-02
2.86337465e-01 -2.67701894e-02 -5.63022435e-01 1.12103379e+00
-3.76776725e-01 4.18760210e-01 1.38587296e-01 -5.77455759e-01
6.50454819e-01 -6.27028823e-01 9.70240474e-01 -2.91485578e-01
4.41578701e-02 6.43964335e-02 5.74810728e-02 -1.24842215e+00
-1.67710647e-01 -8.15714061e-01 -6.86428010e-01 -7.00102389e-01
7.07584620e-02 -4.20209408e-01 6.48002982e-01 2.85657406e-01
3.18163693e-01 -2.61969775e-01 4.93370593e-01 -1.33678639e+00
-6.04632318e-01 -5.31726062e-01 -7.59631515e-01 6.34369850e-01
4.08427209e-01 -9.53895152e-01 -8.11898112e-01 -1.04961939e-01] | [8.057168006896973, -2.322669506072998] |
da578cd3-60d9-4617-b37d-565e013626b7 | online-functional-connectivity-analysis-of | 2303.03279 | null | https://arxiv.org/abs/2303.03279v1 | https://arxiv.org/pdf/2303.03279v1.pdf | Online functional connectivity analysis of large all-to-all networks | The analysis of EEG/MEG functional connectivity has become an important tool in neural research. Especially the high time resolution of EEG/MEG enables important insight into the functioning of the human brain. To date, functional connectivity is commonly estimated offline, i.e., after the conclusion of the experiment. However, online computation of functional connectivity has the potential to enable unique experimental paradigms. For example, changes of functional connectivity due to learning processes could be tracked in real time and the experiment be adjusted based on these observations. Furthermore, the connectivity estimates can be used for neurofeedback applications or the instantaneous inspection of measurement results. In this study, we present the implementation and evaluation of online sensor and source space functional connectivity estimation in the open-source software MNE Scan. Online capable implementations of several functional connectivity metrics were established in the Connectivity library within MNE-CPP and made available as a plugin in MNE Scan. Online capability was achieved by enforcing multithreading and high efficiency for all computations, so that repeated computations were avoided wherever possible, which allows for a major speed-up in the case of overlapping intervalls. We present comprehensive performance evaluations of these implementations proving the online capability for the computation of large all-to-all functional connectivity networks. As a proof of principle, we demonstrate the feasibility of online functional connectivity estimation in the evaluation of somatosensory evoked brain activity. | ['Johannes Vorwerk', 'Jens Haueisen', 'Daniel Baumgarten', 'Matti Hämäläinen', 'Jinlong Dong', 'Lorenz Esch'] | 2023-03-06 | null | null | null | null | ['connectivity-estimation', 'eeg', 'eeg'] | ['graphs', 'methodology', 'time-series'] | [-1.45023525e-01 -1.86645702e-01 3.90079618e-01 -2.27880284e-01
-1.91265315e-01 -5.51386952e-01 1.20844625e-01 4.63666946e-01
-6.29488051e-01 7.97448933e-01 -2.09134206e-01 -2.49649659e-01
-5.85591793e-01 -7.07741916e-01 -6.32174730e-01 -5.28994024e-01
-9.23176706e-01 2.88030505e-01 1.85133398e-01 1.53046648e-03
2.69595146e-01 9.84998465e-01 -1.48891413e+00 -5.45023999e-04
8.11559618e-01 1.15000284e+00 5.29433250e-01 3.12897384e-01
2.94360429e-01 -1.40825734e-02 -5.09153128e-01 6.52456284e-02
5.81644513e-02 -2.34088361e-01 -5.32141387e-01 -4.77037549e-01
-2.88949281e-01 -2.06199273e-01 -1.25296161e-01 8.75484288e-01
7.97920287e-01 1.90731943e-01 1.21049486e-01 -1.21374822e+00
3.10597062e-01 7.12989807e-01 -2.88959235e-01 7.14082778e-01
3.77129436e-01 1.62765756e-01 4.71769691e-01 -7.03502536e-01
5.45799851e-01 5.45305729e-01 4.38076526e-01 9.75991972e-03
-1.47238863e+00 -8.56532216e-01 2.72702537e-02 4.27339911e-01
-1.56260943e+00 -4.74041820e-01 4.92584556e-01 -4.86854643e-01
1.15879893e+00 2.36132935e-01 1.28228986e+00 8.48049164e-01
5.71949899e-01 -2.53957827e-02 1.27253938e+00 -1.94338083e-01
4.30911779e-01 -2.38172755e-01 2.36131132e-01 2.26977617e-01
3.45323592e-01 9.12966430e-02 -6.16973162e-01 -1.66271925e-01
8.53451252e-01 -3.11956376e-01 -6.03685379e-01 -1.51359260e-01
-1.42342961e+00 2.44297758e-01 5.25100470e-01 7.26199865e-01
-7.60380507e-01 1.47173285e-01 5.85048735e-01 3.76345128e-01
4.04780030e-01 4.66677517e-01 -4.27314132e-01 -5.00333607e-01
-1.16485202e+00 1.57036036e-02 5.10645211e-01 2.74573386e-01
6.52754486e-01 -1.97005957e-01 -2.51953658e-02 5.20495832e-01
-7.24054575e-02 2.32418180e-01 7.27083802e-01 -8.35951269e-01
1.01093546e-01 2.28515103e-01 -6.48775250e-02 -9.23305273e-01
-9.55825508e-01 -5.50861835e-01 -7.88117826e-01 3.15495193e-01
2.24949107e-01 -3.18268120e-01 -1.86450318e-01 1.67790627e+00
3.17414969e-01 1.43182024e-01 -6.00388765e-01 8.24383795e-01
1.59441262e-01 -3.25145270e-03 -3.35882008e-02 -4.68147248e-01
1.12325644e+00 3.21387313e-02 -7.63896644e-01 7.36132041e-02
5.39986670e-01 -3.48952889e-01 5.87520778e-01 4.92769867e-01
-9.40329850e-01 -3.65136683e-01 -9.04270411e-01 5.27901769e-01
-4.52496231e-01 -1.96231917e-01 8.73799860e-01 5.36716819e-01
-1.36396646e+00 9.00874734e-01 -1.31268954e+00 -3.28765720e-01
2.67717779e-01 7.81183243e-01 -7.31100261e-01 3.38000625e-01
-1.10606742e+00 1.10543787e+00 4.81525183e-01 5.63289821e-01
-5.26422799e-01 -7.87201047e-01 -4.28224474e-01 2.93404281e-01
-1.51521489e-01 -3.98198605e-01 6.63112044e-01 -1.03465998e+00
-1.47348118e+00 6.38515234e-01 -1.89233513e-03 -4.52728629e-01
4.13849086e-01 1.67496130e-01 -3.62323374e-01 2.84240812e-01
2.43715886e-02 5.90857625e-01 3.78014117e-01 -3.67366970e-01
1.13101080e-01 -6.16418302e-01 -2.26638317e-02 1.04370505e-01
-2.35807672e-01 3.38302217e-02 6.93739112e-03 -3.35689455e-01
2.59234875e-01 -6.06508136e-01 -6.90841675e-02 -9.27177370e-02
-6.73142225e-02 2.81376064e-01 7.42460862e-02 -7.42435157e-01
9.27935123e-01 -2.15221691e+00 -4.17262316e-03 6.80993974e-01
1.40671104e-01 -2.03170225e-01 -1.70127470e-02 4.33523297e-01
-6.15299940e-01 -1.32421836e-01 -1.34438664e-01 3.10595900e-01
-3.07868987e-01 -3.30116421e-01 2.30387822e-01 8.87985408e-01
-1.39041945e-01 7.72363782e-01 -6.77998126e-01 -1.20173514e-01
2.84712583e-01 5.03455043e-01 -4.26853150e-01 1.36751279e-01
3.50708991e-01 7.89237440e-01 -1.72474012e-01 1.00817546e-01
5.61480284e-01 2.89706774e-02 4.38221067e-01 -3.22637409e-01
-6.14837646e-01 4.32983488e-01 -1.13675487e+00 1.95657694e+00
-6.06492221e-01 6.69652820e-01 3.89033616e-01 -1.23012364e+00
5.53572834e-01 5.55087090e-01 9.01760936e-01 -9.08073664e-01
3.76380891e-01 2.32415214e-01 6.45626128e-01 -3.20546746e-01
3.74198221e-02 2.84133375e-01 4.27440256e-01 6.09354377e-01
7.68310353e-02 -5.89854121e-02 3.48180711e-01 6.76965714e-02
1.23871505e+00 9.04624015e-02 2.81541288e-01 -7.61692643e-01
1.97424248e-01 -2.17045620e-01 5.71921431e-02 5.27307451e-01
-6.96587749e-03 2.00550377e-01 5.19272804e-01 9.03112739e-02
-4.95592237e-01 -9.69472051e-01 -6.08080804e-01 5.33800662e-01
-1.84494361e-01 -3.04609597e-01 -8.72223973e-01 -8.91112443e-03
-1.24113999e-01 3.89093339e-01 -5.02452850e-01 -1.18261017e-01
-2.71547228e-01 -7.23210633e-01 1.91286191e-01 3.99007201e-01
2.39885077e-01 -1.17865813e+00 -9.98868227e-01 5.47618449e-01
8.47422704e-02 -9.60230172e-01 -1.17950141e-03 3.94570410e-01
-1.26742327e+00 -1.21708107e+00 -4.88096267e-01 -2.43796140e-01
6.95829451e-01 -4.01978195e-02 6.28723681e-01 4.61976081e-02
-5.02080023e-01 5.85486770e-01 -3.20208371e-02 5.60790189e-02
2.03960061e-01 2.74082031e-02 1.35575026e-01 -1.04955472e-01
-1.22140475e-01 -1.25577474e+00 -8.70184958e-01 1.26755074e-01
-7.69241989e-01 1.14868348e-03 2.35134140e-01 4.89757985e-01
6.26099467e-01 -2.68304460e-02 8.51222634e-01 -4.62237597e-01
9.20497537e-01 -6.26033187e-01 -9.87780631e-01 -1.88283566e-02
-4.85269696e-01 -4.52748984e-02 5.56507409e-01 -2.36442462e-01
-6.53475463e-01 -4.55394313e-02 -3.39409351e-01 -2.80115679e-02
-1.04277901e-01 9.66289341e-01 -1.57507479e-01 -3.19442809e-01
4.82448161e-01 -6.68283030e-02 1.99011099e-02 -2.21605554e-01
1.38457879e-01 5.24441481e-01 4.09688801e-01 -5.07304847e-01
-9.01604351e-03 2.96408534e-01 1.40385628e-02 -7.37301290e-01
1.00622416e-01 -3.28955621e-01 -7.68948674e-01 -5.27136862e-01
3.97214800e-01 -7.24003792e-01 -1.11595023e+00 5.08721709e-01
-1.01098990e+00 -4.84011203e-01 -4.46534902e-02 8.12560380e-01
-4.64737833e-01 1.10832803e-01 -3.19062620e-01 -5.67323446e-01
-5.50229728e-01 -1.11631143e+00 5.43421626e-01 9.71197933e-02
-4.73282248e-01 -1.02369595e+00 -2.30516437e-02 -4.36510742e-01
6.19189560e-01 3.09365422e-01 6.16610765e-01 -3.90016586e-01
-2.26139784e-01 -3.40005189e-01 -3.05435210e-02 2.65627857e-02
6.33376464e-02 -1.21645957e-01 -9.07235503e-01 -3.12970817e-01
1.23723507e-01 1.47495672e-01 1.99662670e-01 6.34932876e-01
1.33802831e+00 2.60937840e-01 -4.32879627e-01 3.50210458e-01
1.23682952e+00 -6.91555068e-02 6.84678078e-01 -4.19096351e-02
1.19143598e-01 6.76725626e-01 2.79586256e-01 4.50603217e-01
-2.03237608e-02 7.03697145e-01 3.86604846e-01 1.23570479e-01
1.13837533e-01 3.28236341e-01 1.60728395e-01 8.42057347e-01
-1.64467439e-01 3.00449789e-01 -8.78410757e-01 5.08556902e-01
-1.56895912e+00 -8.16223502e-01 -3.80960017e-01 2.82347751e+00
6.78577602e-01 1.74192980e-01 -7.71907493e-02 2.60237902e-01
8.85317385e-01 -3.10027421e-01 -4.41983670e-01 -2.85726011e-01
2.99836010e-01 6.72810376e-01 4.07168448e-01 3.57373506e-01
-2.90575832e-01 3.11641932e-01 6.22557974e+00 3.86393130e-01
-1.48943794e+00 5.32513559e-01 2.83376276e-01 -4.14242774e-01
-5.63079305e-02 8.94737244e-03 -2.22755402e-01 5.16047060e-01
1.36632919e+00 -2.96303421e-01 8.64545226e-01 4.18730915e-01
7.51692891e-01 -8.20788980e-01 -9.49357152e-01 9.55243170e-01
-4.99401480e-01 -1.17693627e+00 -7.63485014e-01 1.28517583e-01
1.10932678e-01 5.60816884e-01 -5.94444394e-01 -2.34277144e-01
-5.19147933e-01 -6.59553349e-01 6.35727286e-01 6.87235177e-01
9.83564794e-01 -7.00624764e-01 4.59483027e-01 3.64717931e-01
-1.10355246e+00 1.25782967e-01 -8.77821445e-02 -2.96212882e-01
2.39004195e-01 1.32328820e+00 -2.11464167e-01 4.24695700e-01
6.27084553e-01 5.23761570e-01 -4.30740803e-01 1.45893657e+00
-2.35152930e-01 4.21600878e-01 -7.42086291e-01 9.98080373e-02
-4.25766557e-01 -2.16988668e-01 4.06510770e-01 9.48764205e-01
5.70720673e-01 6.14568181e-02 -2.58288175e-01 9.93417442e-01
9.58959684e-02 9.86735821e-02 -4.63410318e-01 -4.55638394e-02
5.03926277e-01 1.53557694e+00 -1.07262480e+00 1.06344052e-01
-8.17586631e-02 8.08100820e-01 5.27391613e-01 3.11563194e-01
-7.56958365e-01 -5.67095280e-01 5.24424195e-01 3.74513984e-01
-1.10108607e-01 -6.29360676e-01 -2.52087057e-01 -1.09377551e+00
2.31939897e-01 -5.02809942e-01 -3.82207334e-02 -8.31967294e-01
-7.28810966e-01 6.46342278e-01 2.29145810e-01 -6.01598620e-01
-2.27565035e-01 -5.54173470e-01 -6.81971192e-01 1.09586334e+00
-1.05250347e+00 -1.84755623e-01 -3.15207362e-01 7.07053542e-01
-1.92170367e-01 5.65654755e-01 1.08888555e+00 4.80350703e-01
-6.03461504e-01 1.00757875e-01 -9.13453996e-02 -2.73158103e-01
4.01117563e-01 -8.07195365e-01 1.37163073e-01 9.14279580e-01
7.70296231e-02 7.51816690e-01 4.44400579e-01 -6.36910200e-01
-1.30491912e+00 -4.13955331e-01 6.15665495e-01 2.57457137e-01
9.04025614e-01 -6.58190727e-01 -8.66109788e-01 3.49760622e-01
1.76575288e-01 2.25208908e-01 3.98651540e-01 1.24069586e-01
3.04554403e-01 -3.89488786e-01 -1.03652894e+00 3.05348456e-01
9.94777858e-01 -5.83388329e-01 -1.57183409e-01 3.91698092e-01
-2.85405219e-02 -4.11356807e-01 -1.23880267e+00 1.30665421e-01
6.57879353e-01 -1.11496007e+00 5.58522165e-01 8.31191167e-02
-2.55319685e-01 -1.17042407e-01 4.10982430e-01 -1.49612689e+00
-7.62006566e-02 -6.27898097e-01 7.73343518e-02 1.00480926e+00
2.37545744e-01 -1.16741896e+00 9.38340127e-02 7.00360417e-01
-1.44054428e-01 -5.17061949e-01 -1.29421830e+00 -5.09306967e-01
-3.38133752e-01 -7.35519052e-01 3.80693614e-01 5.60214281e-01
7.48481512e-01 -1.98456392e-01 3.68685186e-01 2.69228131e-01
1.84081212e-01 -3.94406468e-02 5.50320707e-02 -1.15144646e+00
-2.74929851e-01 -3.78363490e-01 -6.01258397e-01 -3.02626789e-01
3.57158184e-01 -1.06041491e+00 -1.90255381e-02 -1.50232637e+00
-9.52248573e-02 -3.23379576e-01 -5.45993149e-01 6.03550792e-01
-4.82004881e-03 2.89293170e-01 -1.99318901e-01 -5.19057177e-02
1.09519981e-01 1.62090182e-01 8.59435678e-01 3.77472222e-01
-2.77165174e-01 -1.05954215e-01 -1.20928533e-01 2.02650324e-01
1.01333809e+00 -6.45646930e-01 -3.60061526e-01 -1.66804969e-01
3.11033338e-01 2.09268838e-01 5.13530612e-01 -1.35437632e+00
2.39927322e-01 4.24292654e-01 3.77349019e-01 -1.51247635e-01
2.72737682e-01 -9.88741219e-01 4.58642870e-01 5.76327562e-01
-3.31751071e-03 2.05883726e-01 4.88308012e-01 1.10497206e-01
-1.61326677e-01 -1.62800640e-01 6.06632292e-01 1.88400552e-01
-3.80312145e-01 2.15211675e-01 -7.40529835e-01 -1.07915126e-01
9.15218115e-01 -1.72045574e-01 -1.65326312e-01 -1.61216676e-01
-9.54590738e-01 -2.44130734e-02 9.82787460e-02 4.22169752e-02
3.90401393e-01 -9.65572119e-01 -2.98178881e-01 4.24178809e-01
-3.02367181e-01 -5.35661280e-01 4.74449307e-01 1.67685997e+00
-4.26685810e-01 3.02709341e-01 -7.03755498e-01 -4.14905101e-01
-8.37279141e-01 2.25220129e-01 6.62104189e-01 -2.58506909e-02
-6.56614721e-01 4.10616487e-01 -2.26194888e-01 1.39496848e-01
-3.95779498e-03 -3.99179935e-01 -1.51196882e-01 2.84236223e-01
5.11478066e-01 2.92717874e-01 7.92828619e-01 -1.49277925e-01
-6.17789984e-01 2.41608582e-02 3.38004172e-01 -3.75309438e-01
1.41083801e+00 -1.89515203e-01 -7.08826959e-01 5.50207675e-01
9.43794012e-01 -4.82484661e-02 -1.02163732e+00 5.58270812e-01
-4.00274023e-02 -8.77377614e-02 4.86376613e-01 -8.43866289e-01
-1.24138248e+00 9.79108036e-01 9.48694766e-01 1.97800338e-01
1.39556682e+00 -3.84586036e-01 2.67929375e-01 1.97810635e-01
8.20713937e-01 -9.97867644e-01 -6.30412459e-01 9.00043696e-02
9.05722260e-01 -5.34546733e-01 1.28452986e-01 -1.41891316e-01
-9.52498391e-02 1.18433273e+00 2.91046649e-01 -1.53699219e-01
9.48619187e-01 5.88976920e-01 -3.54323238e-01 -3.58021945e-01
-3.95540178e-01 -3.71017084e-02 1.35005951e-01 5.61639130e-01
7.22698689e-01 1.65060610e-01 -7.08242655e-01 5.47790289e-01
-9.53140929e-02 2.40257293e-01 3.49617541e-01 7.83621669e-01
-1.38728589e-01 -9.54667270e-01 -3.30243289e-01 7.29861677e-01
-4.44534183e-01 -3.69918436e-01 1.63517147e-02 7.58899152e-01
-8.71533528e-02 8.26321363e-01 1.78051397e-01 4.73719314e-02
2.79997617e-01 1.08743399e-01 8.53806376e-01 -5.53835452e-01
-9.82373595e-01 -1.14760295e-01 5.55360578e-02 -1.06862772e+00
-3.89410138e-01 -6.40895486e-01 -1.51957464e+00 8.05056468e-03
-4.84501421e-01 3.39926779e-01 1.21932948e+00 1.02700448e+00
7.09498882e-01 7.17534721e-01 2.70935118e-01 -1.08807099e+00
4.35485691e-02 -9.96190488e-01 -8.85045230e-01 -6.71966523e-02
-2.97446579e-01 -7.41883337e-01 -2.26689145e-01 -2.92116880e-01] | [13.030318260192871, 3.395862102508545] |
71c2c9ef-4a2d-4c98-b1ce-f0d42dfed29c | afdp-an-automated-function-description | 1910.06965 | null | http://arxiv.org/abs/1910.06965v1 | http://arxiv.org/pdf/1910.06965v1.pdf | AFDP: An Automated Function Description Prediction Approach to Improve Accuracy of Protein Function Predictions | With the rapid growth in high-throughput biological sequencing technologies
and subsequently the amount of produced omics data, it is essential to develop
automated methods to annotate the functionality of unknown genes and proteins.
There are developed tools such as AHRD applying known proteins characterization
to annotate unknown ones. Some other algorithms such as eggNOG apply
orthologous groups of proteins to detect the most probable function. However,
while the available tools focus on the detection of the most similar
characterization, they are not able to generalize and integrate information
from multiple homologs while maintaining accuracy. Here, we devise AFDP, an
integrated approach for protein function prediction which benefits from the
combination of two available tools, AHRD and eggNOG, to predict the
functionality of novel proteins and produce more precise human readable
descriptions by applying our stCFExt algorithm. StCFExt creates function
descriptions applying available manually curated descriptions in swiss-prot.
Using a benchmark dataset we show that the annotations predicted by our
approach are more accurate than eggNOG and AHRD annotations. | [] | 2019-10-15 | null | null | null | null | ['protein-function-prediction'] | ['medical'] | [ 2.42378071e-01 -3.13389078e-02 2.57166117e-01 -3.90381157e-01
-2.02684030e-01 -9.52667654e-01 1.35289565e-01 7.51422584e-01
-2.27681935e-01 1.45921004e+00 -1.59226045e-01 -2.22510442e-01
-2.83800751e-01 -4.93937105e-01 -6.47988558e-01 -5.60005605e-01
8.06108266e-02 8.33233535e-01 7.18710542e-01 -3.34065139e-01
2.42427051e-01 7.09312677e-01 -1.72995937e+00 5.29511094e-01
1.04292679e+00 3.98695171e-01 8.01986217e-01 6.14489377e-01
-3.60290408e-01 1.11851841e-01 -4.24868941e-01 -3.59702826e-01
8.45918357e-02 -5.72549403e-01 -8.87803376e-01 -3.66817951e-01
-3.18175256e-01 2.48886392e-01 5.61349809e-01 7.39989758e-01
4.69456077e-01 -1.31855235e-01 6.01276934e-01 -8.52689326e-01
-4.01964277e-01 1.34321496e-01 1.36991531e-01 -1.91590831e-01
8.82743955e-01 1.48295924e-01 7.93032765e-01 -7.68202245e-01
1.26898801e+00 1.14266694e+00 6.88292325e-01 3.85820866e-01
-1.30177391e+00 -2.32198849e-01 -4.77039546e-01 2.88143843e-01
-1.21609592e+00 6.08386509e-02 -9.79009345e-02 -6.68900609e-01
1.54849148e+00 2.80838549e-01 4.81873274e-01 4.40816700e-01
2.80108869e-01 9.29785892e-02 1.00176251e+00 -3.48533154e-01
4.36194181e-01 2.57478118e-01 -5.51039446e-03 6.92933202e-01
5.05751848e-01 -4.27493125e-01 -4.04919893e-01 -5.54500878e-01
-5.28320037e-02 1.04068935e-01 -3.68831605e-01 -3.71641099e-01
-1.09451783e+00 3.90261948e-01 3.03468741e-02 3.14733773e-01
-5.82715869e-01 -3.25816244e-01 6.45783484e-01 1.84271410e-01
6.26180708e-01 8.49688351e-01 -1.18664527e+00 -1.21923238e-01
-5.06729424e-01 2.08334535e-01 1.16375589e+00 8.81145656e-01
1.12669802e+00 -5.87292194e-01 4.69812334e-01 6.48034632e-01
7.02305809e-02 -1.23628929e-01 8.80451381e-01 -7.35218525e-01
-3.29355657e-01 9.38797951e-01 2.98122734e-01 -6.00539386e-01
-3.95874143e-01 1.79033175e-01 -1.22301705e-01 1.62150815e-01
4.99453902e-01 9.40385088e-02 -7.66782224e-01 1.39561951e+00
6.07325017e-01 -1.17113836e-01 3.86350840e-01 6.58265948e-01
4.73613292e-01 5.68760157e-01 2.81019151e-01 -4.05265063e-01
1.11639857e+00 -4.65838760e-01 -5.36900580e-01 6.20922208e-01
7.37854004e-01 -9.77985740e-01 5.44273257e-01 6.50453985e-01
-5.85828543e-01 -5.41527927e-01 -1.13314545e+00 3.39105129e-02
-1.04509020e+00 1.32848425e-02 5.80069840e-01 3.49639446e-01
-9.57299173e-01 1.19300449e+00 -8.82995009e-01 -9.59051311e-01
-2.21549511e-01 4.80202764e-01 -8.90921474e-01 1.44832537e-01
-8.16646874e-01 1.36483204e+00 1.18732190e+00 -3.05990815e-01
-6.57313287e-01 -4.35740769e-01 -3.17962587e-01 3.88960005e-03
2.95583785e-01 -6.82452857e-01 8.79854620e-01 -6.65336430e-01
-1.23092663e+00 8.79216433e-01 -3.15889508e-01 -5.69739580e-01
3.17884326e-01 -1.61433052e-02 -2.12086424e-01 3.22960943e-01
1.45439729e-01 6.35664165e-01 -7.18267187e-02 -7.94483662e-01
-6.50911510e-01 -3.70545596e-01 -2.17415705e-01 -2.74777174e-01
-3.50017212e-02 1.42634347e-01 3.25715423e-01 -1.13585487e-01
2.61288345e-01 -7.37559557e-01 -2.93040752e-01 2.01389402e-01
-1.80195943e-01 -5.78609586e-01 7.55075336e-01 -8.20800364e-01
6.76046968e-01 -1.92202771e+00 3.24287057e-01 -1.16622880e-01
2.46522471e-01 3.77951831e-01 7.18734181e-03 1.00722742e+00
-4.32395041e-01 2.12481409e-01 -2.80848980e-01 4.01113778e-01
-1.01165988e-01 5.30146003e-01 -7.54159465e-02 4.87052411e-01
4.60152894e-01 3.28385562e-01 -8.17132592e-01 -2.11933121e-01
2.23938718e-01 4.73177582e-01 -4.71592844e-01 2.88085788e-01
-5.55942774e-01 3.95795256e-01 -4.67118829e-01 6.13682032e-01
7.12416649e-01 -2.07128003e-01 5.08424878e-01 -1.04441382e-01
-5.56972027e-01 2.60047704e-01 -7.65852451e-01 1.55825901e+00
1.39131531e-01 1.11000896e-01 -2.34415546e-01 -1.17335355e+00
1.22297418e+00 4.93581504e-01 5.53559065e-01 3.06603372e-01
-1.53632835e-01 6.92263663e-01 1.42295808e-01 -5.98338485e-01
1.33338198e-01 -9.11899954e-02 4.32300925e-01 -6.79302812e-02
4.43846822e-01 -6.27947971e-03 6.50567234e-01 -2.11545005e-02
1.24001133e+00 1.03692997e+00 1.06233382e+00 -4.73519087e-01
8.14948618e-01 6.08839333e-01 8.21523786e-01 1.37610450e-01
-1.01227917e-01 3.36271018e-01 4.97244537e-01 -7.12932289e-01
-1.54156220e+00 -5.96674621e-01 -1.82413250e-01 1.12138343e+00
-8.21552873e-02 -5.34977973e-01 -8.24285865e-01 -4.67548817e-01
5.02802692e-02 4.41049069e-01 -1.66427806e-01 2.09970381e-02
-2.12369338e-01 -9.57950532e-01 5.41718125e-01 -1.15555443e-01
1.42184481e-01 -9.17477548e-01 -6.77345395e-01 7.07934499e-01
2.24857479e-02 -8.08654785e-01 3.20118755e-01 5.84657490e-01
-5.19165099e-01 -1.39311588e+00 -7.00395942e-01 -6.05397522e-01
3.50062251e-01 2.03272495e-02 6.96372807e-01 -9.94684547e-02
-7.38043487e-01 -4.09245104e-01 -7.04993248e-01 -6.45468175e-01
-9.26875889e-01 -1.73992604e-01 2.06418261e-01 -4.09016788e-01
5.17791748e-01 -8.53963614e-01 -1.89476252e-01 4.99949127e-01
-7.85504103e-01 1.94486827e-01 5.57909250e-01 6.42912209e-01
7.77344465e-01 -1.92244008e-01 9.77298975e-01 -8.39296818e-01
4.11648899e-01 -6.65674269e-01 -5.33644974e-01 5.99881351e-01
-7.74733901e-01 2.84209758e-01 8.95976841e-01 -3.53219628e-01
-6.90915227e-01 6.82935953e-01 -5.35759032e-01 1.80210561e-01
-4.84261692e-01 4.28242087e-01 -1.61997631e-01 -1.04960598e-01
7.03904569e-01 4.07849997e-01 1.28618062e-01 -8.83690298e-01
5.92222810e-02 7.60100663e-01 3.55833411e-01 -5.00915945e-01
3.46691340e-01 7.52318352e-02 1.04327254e-01 -7.31713891e-01
-3.82031053e-01 -8.49658549e-01 -7.77093410e-01 2.73584872e-01
9.32448864e-01 -6.30427778e-01 -1.03524184e+00 -1.13781467e-01
-1.35611641e+00 4.32594985e-01 -1.33845797e-02 5.45765400e-01
-7.71935761e-01 7.96409726e-01 -2.80824810e-01 -6.03748798e-01
-5.45045495e-01 -1.14840138e+00 1.01435637e+00 1.21547477e-02
-4.30382401e-01 -3.85809362e-01 3.03066701e-01 4.65812460e-02
9.82439667e-02 4.04708356e-01 1.22964013e+00 -1.48237526e+00
-3.44667107e-01 1.08548805e-01 -8.05421993e-02 3.56158048e-01
3.46724510e-01 2.37578556e-01 -4.47412550e-01 1.59617439e-01
-3.23824018e-01 -3.44456621e-02 5.32592356e-01 -2.30256677e-01
7.34420359e-01 -1.08335704e-01 -5.13305545e-01 3.38479728e-01
1.60912502e+00 4.94744152e-01 5.09363949e-01 4.92328525e-01
9.14177820e-02 7.95092463e-01 1.11519849e+00 2.36269042e-01
-3.47580016e-01 6.48531437e-01 4.74701375e-01 2.55890340e-01
9.22582373e-02 2.58163027e-02 4.96466048e-02 1.64711863e-01
-5.50382853e-01 -4.85958010e-01 -9.99996245e-01 3.25010508e-01
-2.08194613e+00 -8.87220562e-01 -4.05742615e-01 2.04804921e+00
8.83959711e-01 -2.81090021e-01 7.97857791e-02 6.99856356e-02
8.72653842e-01 -1.06244123e+00 -7.13371217e-01 -6.25180066e-01
-3.98175210e-01 4.23464686e-01 4.07652289e-01 5.85148484e-02
-7.59891868e-01 7.68041432e-01 6.48787403e+00 5.05650222e-01
-6.86351776e-01 -5.75446598e-02 1.53668106e-01 4.03775185e-01
7.12973997e-02 3.33521694e-01 -8.75143170e-01 4.21734929e-01
1.18614423e+00 -6.53111756e-01 3.49788070e-01 1.14116597e+00
4.35621321e-01 -1.12380460e-01 -9.39601660e-01 4.35564637e-01
-1.79666191e-01 -1.45155859e+00 -8.36912841e-02 7.33101368e-02
2.98134655e-01 1.44110397e-01 -1.04379892e+00 -1.23000816e-01
4.36812133e-01 -6.53028786e-01 3.70049387e-01 5.94745100e-01
3.35986882e-01 -5.24523675e-01 1.00253415e+00 4.83873487e-01
-9.34110820e-01 1.76535398e-01 -9.52459335e-01 -8.50890800e-02
-1.18472297e-02 4.90021557e-01 -1.74335730e+00 9.37093437e-01
5.81373274e-01 6.22695446e-01 -5.29626369e-01 1.21890950e+00
-5.63183278e-02 2.42126718e-01 -2.69925445e-01 -4.65212226e-01
-1.75700054e-01 -2.58117318e-01 4.86159712e-01 1.22961020e+00
6.01869702e-01 5.70592619e-02 4.66927648e-01 7.65057623e-01
4.77979481e-01 6.87262535e-01 -3.13571006e-01 -5.14510393e-01
1.15992777e-01 1.38829005e+00 -6.15987122e-01 -6.67613685e-01
-1.70955092e-01 1.29031515e+00 1.06771357e-01 -2.80011836e-02
-4.68718916e-01 -4.44895387e-01 9.58037257e-01 2.75433630e-01
2.16050610e-01 -2.64294475e-01 5.66322386e-01 -7.64519989e-01
-3.99957061e-01 -8.36831212e-01 4.53568518e-01 -1.15152335e+00
-1.45031965e+00 5.20807564e-01 -7.29624368e-03 -1.08600545e+00
-4.67857748e-01 -9.96843696e-01 -1.85050908e-02 1.04625845e+00
-9.69227850e-01 -1.13102579e+00 -1.27520934e-02 -5.04228398e-02
5.70271611e-01 -2.81200279e-02 1.28346813e+00 1.86435357e-02
-2.02829659e-01 -3.45119387e-01 4.23064977e-01 -6.53495669e-01
1.04184616e+00 -1.22388041e+00 1.39908060e-01 3.55620980e-01
-7.04023778e-01 7.67210782e-01 1.16139960e+00 -1.00947487e+00
-1.10800231e+00 -9.49480653e-01 1.09004092e+00 -1.26105830e-01
8.30699563e-01 -3.27791810e-01 -1.10558450e+00 4.62148666e-01
8.72123688e-02 -3.84247005e-01 8.46506834e-01 -4.77594674e-01
-1.29856214e-01 3.76572520e-01 -1.36446929e+00 -1.24840261e-02
9.17488396e-01 -6.27284823e-03 -7.29329467e-01 7.25155234e-01
9.78097320e-01 1.53480306e-01 -1.14740992e+00 5.26412904e-01
6.44939125e-01 -1.00939214e+00 7.74241507e-01 -8.93732369e-01
1.61500648e-01 -8.99939656e-01 -1.90579474e-01 -9.04734194e-01
-1.82015404e-01 -3.85578603e-01 5.10467947e-01 1.07602882e+00
5.79300880e-01 -8.47916007e-01 3.50400627e-01 2.23310575e-01
-4.39978689e-01 -4.28914070e-01 -8.30202103e-01 -7.55831182e-01
-4.12549466e-01 4.79496747e-01 5.01822710e-01 1.02962875e+00
3.47099543e-01 3.16167504e-01 -1.54649615e-01 -4.36986499e-02
1.38959000e-02 9.71757248e-02 6.90323651e-01 -1.70198905e+00
-4.73327667e-01 2.30559945e-01 -9.49895442e-01 -2.53026217e-01
4.38303314e-03 -8.30518961e-01 2.69487388e-02 -1.64734030e+00
2.83565104e-01 1.30082414e-01 -1.52820826e-01 7.89836586e-01
1.32982314e-01 8.15249234e-02 -4.87715095e-01 3.22882042e-05
-4.01380599e-01 -4.78942655e-02 6.61097646e-01 2.85915941e-01
-2.51584668e-02 -3.54990810e-01 -3.15397292e-01 6.90503597e-01
6.75898910e-01 -5.76920629e-01 -6.35214597e-02 4.29826856e-01
5.33556342e-01 -4.18812782e-01 1.47059754e-01 -9.59287405e-01
-1.03388123e-01 -2.97370523e-01 5.04741788e-01 -7.35816002e-01
1.89564407e-01 -6.82582140e-01 1.03217936e+00 1.02266645e+00
7.44140660e-03 1.39828727e-01 1.40287027e-01 5.88307381e-01
-1.37735978e-01 -5.62571585e-01 6.31256700e-01 -6.29142463e-01
-9.40821946e-01 -9.83961299e-02 -6.58435285e-01 -9.41273034e-01
1.18084455e+00 -3.98413360e-01 -4.20726538e-01 7.51745999e-02
-1.06007946e+00 -2.79570311e-01 8.61443102e-01 2.77948797e-01
3.77664804e-01 -3.81806314e-01 -6.53573811e-01 -5.37511834e-04
4.09879923e-01 -5.53859770e-01 -1.66318566e-01 6.46686196e-01
-1.21178842e+00 5.92544615e-01 -5.76662064e-01 -4.93751884e-01
-1.61487567e+00 1.00711966e+00 1.99646413e-01 2.29384862e-02
-1.49123967e-01 4.00765866e-01 1.54597417e-01 -6.02772236e-01
-5.35222590e-01 -7.03356341e-02 -3.14474016e-01 -2.58905917e-01
3.95139575e-01 1.81066871e-01 3.09678435e-01 -5.40172577e-01
-4.92700934e-01 3.16708177e-01 1.28625840e-01 4.32924598e-01
1.68827605e+00 -6.90924888e-03 -6.50140762e-01 2.60959983e-01
1.01556110e+00 -1.29677966e-01 -4.66706693e-01 2.86040932e-01
4.58464116e-01 -3.27305436e-01 -7.41593063e-01 -9.68271434e-01
2.29825824e-01 4.39397573e-01 7.06934929e-01 3.32737453e-02
1.12548172e+00 1.01175018e-01 2.31909260e-01 8.60149086e-01
5.49032331e-01 -7.95327663e-01 -4.97233212e-01 3.22293460e-01
5.32770336e-01 -7.96680629e-01 -1.34094618e-02 -8.69806349e-01
-1.23077378e-01 1.32599986e+00 3.36583406e-01 2.43685856e-01
1.60018243e-02 1.32310167e-01 -4.99187946e-01 -5.02412347e-03
-1.09396064e+00 -3.79349202e-01 -1.80897713e-02 6.13942444e-01
8.79772604e-01 -1.07409403e-01 -1.35060966e+00 3.34537894e-01
-5.13060279e-02 2.42211863e-01 5.23010790e-01 1.00555909e+00
-1.01575804e+00 -1.68047547e+00 -2.05804437e-01 1.78103015e-01
-7.02254593e-01 -5.57329245e-02 -1.12138247e+00 6.50546670e-01
4.65394795e-01 8.24916303e-01 -4.61858124e-01 -2.98636705e-01
2.40525961e-01 6.54515266e-01 2.67538756e-01 -9.20663416e-01
-3.91824722e-01 2.13548198e-01 4.80219036e-01 -1.88776419e-01
-5.57876170e-01 -4.58631009e-01 -1.57229543e+00 3.52213532e-02
-6.72550023e-01 7.63678193e-01 1.23041928e+00 8.51633847e-01
6.89327002e-01 2.56267399e-01 2.54399925e-01 -5.37157893e-01
-2.68827170e-01 -1.06340802e+00 -5.00377059e-01 3.75491112e-01
-4.13744211e-01 -4.78632569e-01 -5.56752719e-02 5.34115016e-01] | [4.754537105560303, 5.439800262451172] |
97afa044-a444-4df0-a4fe-86709dc94402 | your-diffusion-model-is-secretly-a-zero-shot | 2303.16203 | null | https://arxiv.org/abs/2303.16203v2 | https://arxiv.org/pdf/2303.16203v2.pdf | Your Diffusion Model is Secretly a Zero-Shot Classifier | The recent wave of large-scale text-to-image diffusion models has dramatically increased our text-based image generation abilities. These models can generate realistic images for a staggering variety of prompts and exhibit impressive compositional generalization abilities. Almost all use cases thus far have solely focused on sampling; however, diffusion models can also provide conditional density estimates, which are useful for tasks beyond image generation. In this paper, we show that the density estimates from large-scale text-to-image diffusion models like Stable Diffusion can be leveraged to perform zero-shot classification without any additional training. Our generative approach to classification, which we call Diffusion Classifier, attains strong results on a variety of benchmarks and outperforms alternative methods of extracting knowledge from diffusion models. Although a gap remains between generative and discriminative approaches on zero-shot recognition tasks, we find that our diffusion-based approach has stronger multimodal relational reasoning abilities than competing discriminative approaches. Finally, we use Diffusion Classifier to extract standard classifiers from class-conditional diffusion models trained on ImageNet. Even though these models are trained with weak augmentations and no regularization, they approach the performance of SOTA discriminative classifiers. Overall, our results are a step toward using generative over discriminative models for downstream tasks. Results and visualizations at https://diffusion-classifier.github.io/ | ['Deepak Pathak', 'Ellis Brown', 'Shivam Duggal', 'Mihir Prabhudesai', 'Alexander C. Li'] | 2023-03-28 | null | null | null | null | ['relational-reasoning'] | ['natural-language-processing'] | [ 2.79778540e-01 2.86154449e-01 -5.09676516e-01 -4.41959500e-01
-9.30469453e-01 -4.38870162e-01 1.32689035e+00 -1.72772989e-01
-1.00442700e-01 4.86339629e-01 7.75069952e-01 -2.38389388e-01
1.68654576e-01 -8.91688883e-01 -5.90809703e-01 -5.34026980e-01
4.02117938e-01 6.51267529e-01 4.91146557e-02 -2.45445341e-01
2.88786590e-02 1.26751751e-01 -1.45199203e+00 6.57339215e-01
5.99014044e-01 9.56213892e-01 7.83172622e-02 8.96901906e-01
-3.19761097e-01 1.44405568e+00 -5.83070219e-01 -5.45138121e-01
3.07117682e-03 -7.50812352e-01 -7.12806404e-01 2.49391243e-01
6.92014635e-01 -6.81603968e-01 -6.08506262e-01 7.92388856e-01
5.30138135e-01 1.80243388e-01 1.22657681e+00 -1.30132604e+00
-1.46266520e+00 8.66058648e-01 -5.24097860e-01 3.37739855e-01
3.35414201e-01 5.44879735e-01 1.23585665e+00 -1.02346230e+00
1.17253673e+00 1.44053721e+00 5.05985975e-01 7.50133753e-01
-1.51366603e+00 -5.84783316e-01 -8.98780823e-02 3.26964781e-02
-1.17677295e+00 -5.96970856e-01 6.37376904e-01 -7.37613440e-01
1.09275961e+00 5.44320559e-03 6.64465964e-01 1.86042416e+00
-1.16506875e-01 1.27427018e+00 1.28137243e+00 -2.17337012e-01
1.33423880e-01 6.79512098e-02 2.09107976e-02 8.78623784e-01
-1.23760536e-01 2.45207641e-03 -6.85063303e-01 -1.27376720e-01
7.34321535e-01 1.65207654e-01 -1.67542934e-01 -2.70506948e-01
-1.22388959e+00 1.34743214e+00 4.40162092e-01 2.70415336e-01
-2.19880655e-01 4.85389262e-01 1.74662486e-01 1.93570226e-01
7.92097449e-01 3.36967081e-01 1.86415613e-01 -3.72913182e-01
-1.14682376e+00 2.05840930e-01 9.13663030e-01 1.08051336e+00
7.42314517e-01 2.86604941e-01 -6.90132916e-01 9.64841008e-01
3.00875772e-02 6.30324423e-01 7.34006941e-01 -1.00593340e+00
1.95984229e-01 3.41499776e-01 -3.59294266e-01 -8.87311816e-01
8.00217316e-02 -2.03718126e-01 -7.57686555e-01 2.40313020e-02
3.46308708e-01 -2.57790923e-01 -1.27868724e+00 1.53542244e+00
-8.31082016e-02 -3.89961638e-02 2.70128846e-02 7.29836106e-01
9.39395666e-01 8.11581314e-01 2.73041457e-01 2.55071461e-01
1.02267420e+00 -1.12987554e+00 -4.85784948e-01 -2.80500144e-01
6.34448588e-01 -7.26716995e-01 1.29936540e+00 1.89279333e-01
-1.16114700e+00 -3.87364626e-01 -8.38746428e-01 -2.57702619e-01
-4.46210712e-01 1.36866756e-02 9.09644127e-01 5.14065683e-01
-1.23588431e+00 4.21978474e-01 -6.28498435e-01 -7.10599422e-01
8.11523080e-01 -2.76648432e-01 -1.92915753e-01 -4.40464646e-01
-8.81942153e-01 8.32696617e-01 -2.60048900e-02 -5.32689095e-01
-1.29407370e+00 -6.53400600e-01 -9.54527259e-01 -1.80056840e-01
1.61605090e-01 -8.25071335e-01 1.41976678e+00 -1.03893244e+00
-1.19549525e+00 7.93694079e-01 -1.27462417e-01 -5.48503339e-01
6.25811696e-01 2.04150798e-04 2.83011552e-02 2.82507479e-01
3.02390486e-01 1.33879340e+00 1.12921274e+00 -1.15159202e+00
-2.49864265e-01 5.77394553e-02 6.05061725e-02 1.23944342e-01
-6.62903130e-01 -2.08548367e-01 -4.62772071e-01 -7.59668350e-01
-3.26454163e-01 -6.90710366e-01 -1.19058587e-01 1.31636068e-01
-4.31709737e-01 -3.25117260e-01 8.89509618e-01 -3.92776251e-01
9.13683116e-01 -1.93629980e+00 2.44213566e-01 -1.08746268e-01
4.45133775e-01 -1.54458597e-01 -4.31340694e-01 6.24076426e-01
1.62107214e-01 2.43610188e-01 -2.61738956e-01 -6.10001266e-01
2.30376557e-01 2.24432781e-01 -6.65063679e-01 1.53009608e-01
4.21897143e-01 1.52825129e+00 -8.72619987e-01 -6.25631392e-01
1.90602794e-01 5.58285952e-01 -6.39790654e-01 8.69445279e-02
-6.12938285e-01 6.66207448e-02 -2.25964993e-01 7.67257571e-01
2.33663544e-01 -6.01083755e-01 -6.64896145e-02 -1.76515840e-02
4.44197565e-01 -7.30728656e-02 -5.14357090e-01 1.88264215e+00
-4.40980941e-01 9.96687949e-01 -3.21751654e-01 -7.72559762e-01
7.89276779e-01 -2.22668219e-02 2.94619560e-01 -7.47712791e-01
4.10871953e-02 -4.78675328e-02 -1.37956381e-01 -4.47351158e-01
6.80798411e-01 -3.74470085e-01 -1.30556762e-01 8.61510754e-01
5.69599688e-01 -4.67899501e-01 3.76519889e-01 9.59276021e-01
1.26371384e+00 2.28427961e-01 -1.04871936e-01 2.32444126e-02
-3.03650349e-01 3.08301568e-01 -1.15922108e-01 1.05404341e+00
4.90883216e-02 8.17811430e-01 5.74332774e-01 -9.08350572e-02
-1.16769564e+00 -1.18073571e+00 -6.08464368e-02 1.20206070e+00
-1.70218050e-01 -5.86585999e-01 -6.49417937e-01 -7.42472053e-01
2.14912340e-01 7.90380538e-01 -8.41647804e-01 -1.82458535e-01
6.20039627e-02 -8.98741841e-01 8.81688297e-01 7.58985519e-01
2.82750547e-01 -9.38407898e-01 -1.18440412e-01 -4.00349200e-02
-1.66949719e-01 -1.14981937e+00 -5.06852806e-01 2.84373797e-02
-6.99403703e-01 -7.39229798e-01 -1.17151618e+00 -4.98384118e-01
6.26773119e-01 3.88549477e-01 1.31246054e+00 -1.52907118e-01
-5.63980341e-01 1.06690407e+00 -4.68688637e-01 -3.67671400e-01
-6.13321900e-01 1.17676832e-01 -3.42059761e-01 -3.87182049e-02
4.52334583e-01 -4.64511544e-01 -6.25418484e-01 9.95024815e-02
-1.05946064e+00 2.13672310e-01 7.60665536e-01 9.30878699e-01
2.73501486e-01 -4.49266762e-01 5.44916630e-01 -1.04847705e+00
1.17140472e+00 -9.06870484e-01 3.50532122e-02 -2.53152251e-02
-6.54752314e-01 1.19180538e-01 2.81628519e-01 -7.64026642e-01
-1.17307508e+00 -2.06320122e-01 -3.91017199e-02 -8.38496327e-01
-2.04093114e-01 4.90453571e-01 4.60616767e-01 3.09210718e-01
1.15955746e+00 3.92056406e-01 1.66948959e-01 -2.37963691e-01
1.00064087e+00 6.30299926e-01 2.79311389e-01 -7.06212878e-01
5.99419773e-01 6.31325364e-01 -3.68169934e-01 -9.65687513e-01
-9.90022779e-01 -2.88117081e-01 -3.38313609e-01 -3.24634016e-01
9.93874371e-01 -1.05448222e+00 -1.19923830e-01 4.00489748e-01
-9.93456244e-01 -8.51623833e-01 -6.95737362e-01 1.19265348e-01
-5.33063650e-01 2.95340985e-01 -1.05085254e+00 -6.89323425e-01
-2.30505526e-01 -9.41218674e-01 1.18701673e+00 2.54030451e-02
-4.68653172e-01 -1.27430868e+00 4.37171161e-02 3.93459082e-01
7.30604112e-01 9.77514982e-02 8.12355697e-01 -5.72733939e-01
-7.46016145e-01 -1.95057005e-01 -4.76436883e-01 2.52862662e-01
-1.81825727e-01 5.45374602e-02 -1.07045138e+00 4.78987489e-03
-3.46818119e-01 -9.88836706e-01 1.38039994e+00 3.32693994e-01
8.93473446e-01 -1.94704339e-01 -3.37333977e-01 5.20303547e-01
1.17241728e+00 -1.79425910e-01 6.30792618e-01 -1.33188888e-01
7.79770911e-01 4.89438683e-01 2.93950170e-01 3.96809697e-01
5.54537058e-01 3.04864794e-01 6.76328242e-02 -2.16830805e-01
-7.80585885e-01 -5.90197563e-01 5.67615807e-01 7.35516310e-01
4.08002101e-02 -3.86453569e-01 -9.50170100e-01 6.08886957e-01
-1.82050693e+00 -1.30753803e+00 -6.74211755e-02 1.52889979e+00
9.49621558e-01 1.85774695e-02 1.24847554e-01 -4.53191608e-01
4.23846930e-01 4.79785621e-01 -6.22134626e-01 -2.55653620e-01
-2.87023246e-01 3.07474047e-01 2.23701671e-01 4.49273527e-01
-9.16699231e-01 1.21611369e+00 7.26473856e+00 1.22528875e+00
-1.08741713e+00 4.30741310e-01 9.47431266e-01 -3.86232048e-01
-7.00590670e-01 2.49164142e-02 -7.88873017e-01 1.94291905e-01
8.68266761e-01 -1.33987054e-01 3.47102284e-01 1.06320095e+00
-1.79453373e-01 -1.72185272e-01 -1.17320967e+00 1.03294194e+00
5.28576255e-01 -1.74320269e+00 4.60588425e-01 2.32780859e-01
1.06858277e+00 4.02799189e-01 3.61120373e-01 5.69375098e-01
1.00907338e+00 -1.23548496e+00 8.37233961e-01 6.38557076e-01
9.73669052e-01 -2.97099739e-01 2.92032957e-01 3.00007641e-01
-7.45687783e-01 8.57026353e-02 -4.99702126e-01 6.23830594e-02
1.45514354e-01 6.35587573e-01 -7.83589602e-01 -2.44162623e-02
3.45640123e-01 1.01373661e+00 -7.37524152e-01 5.62227607e-01
-2.71031439e-01 7.59967268e-01 -8.50339457e-02 -2.69696921e-01
4.00222957e-01 1.31285386e-02 3.12132597e-01 1.60127032e+00
3.46991450e-01 -9.86100063e-02 5.12991883e-02 1.33680069e+00
-3.17058802e-01 7.82709420e-02 -1.16212511e+00 -5.67249477e-01
1.01873621e-01 1.36627126e+00 -6.82239950e-01 -7.50311434e-01
-4.01892811e-01 1.20407522e+00 6.09789729e-01 5.80199480e-01
-8.61315787e-01 -1.35704011e-01 3.43142509e-01 7.81867355e-02
3.32994431e-01 -3.29300106e-01 -1.40470356e-01 -1.44232309e+00
-4.16846544e-01 -8.77433062e-01 3.38135779e-01 -1.10592401e+00
-1.76203787e+00 4.57790583e-01 1.35270223e-01 -7.94093072e-01
-5.43236256e-01 -4.82729912e-01 -6.31918013e-01 6.84754491e-01
-1.20258689e+00 -1.51065314e+00 -4.29283470e-01 6.58350468e-01
9.44069445e-01 -2.28942856e-01 7.66919076e-01 5.07760840e-03
-3.24833661e-01 3.04607093e-01 -5.76269254e-02 2.55496114e-01
8.08790147e-01 -1.29285181e+00 4.35222179e-01 6.92606449e-01
5.75532377e-01 5.75199187e-01 4.75155622e-01 -8.43044162e-01
-1.50389874e+00 -1.05399728e+00 4.43564087e-01 -8.57022643e-01
9.64988351e-01 -3.86987925e-01 -5.60903966e-01 9.20729280e-01
6.18892908e-01 -1.17843866e-01 7.23564625e-01 1.90352961e-01
-6.55082524e-01 3.22865963e-01 -6.99429214e-01 8.16181898e-01
1.20880902e+00 -8.22830319e-01 -3.98241818e-01 4.79873031e-01
5.91296971e-01 -6.46707490e-02 -9.29104567e-01 -1.72809288e-01
4.63987380e-01 -9.40998614e-01 8.82621109e-01 -6.26235902e-01
1.13915217e+00 3.21813315e-01 -2.12060645e-01 -1.30616009e+00
-3.29576075e-01 -5.99544406e-01 -3.24660271e-01 1.35180068e+00
5.91048539e-01 -4.86136585e-01 7.62494802e-01 7.54661322e-01
8.20721164e-02 -7.05763102e-01 -3.46573174e-01 -6.93020523e-01
3.26253444e-01 -6.71899676e-01 9.48889479e-02 1.06502748e+00
7.67687336e-02 8.38342607e-01 -4.11967218e-01 -7.43097305e-01
7.07523406e-01 9.74173751e-03 9.53768134e-01 -7.52314627e-01
-5.69699645e-01 -7.70743787e-01 -2.23005086e-01 -1.34016395e+00
2.38764644e-01 -1.30977666e+00 -1.32841259e-01 -1.85644424e+00
5.65534353e-01 -3.39977652e-01 1.09006040e-01 6.30475640e-01
-1.54463751e-02 6.46572351e-01 3.15481067e-01 4.83302951e-01
-6.01054192e-01 6.70142889e-01 1.49526823e+00 -6.34522796e-01
6.76618665e-02 -4.72432107e-01 -8.90144169e-01 5.63779712e-01
6.58165395e-01 -3.23407918e-01 -7.68523157e-01 -4.90420103e-01
1.42695352e-01 -9.55092981e-02 4.16768283e-01 -8.63118589e-01
2.00087100e-01 -1.65304407e-01 5.48361897e-01 -2.28877708e-01
6.50439382e-01 -1.02387160e-01 -1.44211978e-01 1.42496362e-01
-6.27002895e-01 -1.40061110e-01 -5.79624029e-04 7.88088739e-01
-1.94688320e-01 -2.14265034e-01 5.34295619e-01 -3.84140640e-01
-7.89663315e-01 3.85066748e-01 -6.74898982e-01 3.36509198e-01
9.98904765e-01 -1.01553842e-01 -7.87930429e-01 -1.02541804e+00
-7.86345780e-01 -1.90155372e-01 5.59941053e-01 5.49172640e-01
8.06247056e-01 -1.36669600e+00 -6.98434293e-01 -5.51005676e-02
3.18390638e-01 -3.50666493e-01 1.22448310e-01 1.02344024e+00
-1.59022257e-01 2.51807600e-01 -4.74326685e-02 -6.92526519e-01
-8.94480407e-01 5.54078221e-01 5.26833609e-02 -6.69201091e-02
-8.26183975e-01 9.53835964e-01 4.04245019e-01 -7.72751644e-02
-2.31303181e-02 -1.84377313e-01 1.47647321e-01 3.35669190e-01
3.85004312e-01 7.67356381e-02 -3.93924177e-01 -6.05667055e-01
2.08344609e-01 1.91460907e-01 -1.39874995e-01 -6.46010935e-01
1.25776327e+00 1.09065339e-01 1.67588115e-01 5.68466127e-01
1.14726627e+00 -7.05043301e-02 -1.34563482e+00 -1.87062874e-01
-3.59680086e-01 -3.46724272e-01 8.21814165e-02 -8.83188307e-01
-1.03871274e+00 1.31407392e+00 1.56682715e-01 3.51030499e-01
6.32928431e-01 4.52633739e-01 7.77052224e-01 2.74970055e-01
1.48004100e-01 -9.58188653e-01 7.03438282e-01 4.78313625e-01
9.22371805e-01 -1.30453360e+00 -1.57249525e-01 -6.18788227e-02
-1.21047771e+00 9.83709514e-01 5.96398711e-01 -1.53164834e-01
5.16768277e-01 4.14732069e-01 -5.79750985e-02 -3.80085438e-01
-1.17581582e+00 -5.55283606e-01 3.40402275e-01 9.24497366e-01
5.17141223e-01 -5.01152426e-02 1.12783529e-01 2.67860115e-01
-1.67517871e-01 8.92782491e-03 5.65559506e-01 8.22064698e-01
-4.37676400e-01 -9.12157178e-01 -8.72042924e-02 7.44105220e-01
-1.22474611e-01 -4.62369561e-01 -7.16821313e-01 6.70082092e-01
-2.22735062e-01 8.10881436e-01 2.23268911e-01 -4.19013441e-01
-1.58722863e-01 2.87812412e-01 6.62726760e-01 -6.82760060e-01
-2.72596717e-01 8.02977309e-02 2.83311218e-01 -5.21048605e-01
-3.30424815e-01 -7.14086473e-01 -8.87856245e-01 -4.36605662e-01
-4.51633669e-02 -2.37946033e-01 5.70530176e-01 6.87588334e-01
4.86915261e-01 4.36627448e-01 5.70433363e-02 -1.00877118e+00
-4.80275631e-01 -1.23823953e+00 -4.13502246e-01 6.58385456e-01
-7.35986233e-02 -5.66387594e-01 -2.11832359e-01 3.26789290e-01] | [11.250561714172363, -0.015052932314574718] |
ae29fab4-e576-4c9d-9b02-edde459185fe | solving-the-false-positives-problem-in-fraud | 1710.07709 | null | http://arxiv.org/abs/1710.07709v1 | http://arxiv.org/pdf/1710.07709v1.pdf | Solving the "false positives" problem in fraud prediction | In this paper, we present an automated feature engineering based approach to
dramatically reduce false positives in fraud prediction. False positives plague
the fraud prediction industry. It is estimated that only 1 in 5 declared as
fraud are actually fraud and roughly 1 in every 6 customers have had a valid
transaction declined in the past year. To address this problem, we use the Deep
Feature Synthesis algorithm to automatically derive behavioral features based
on the historical data of the card associated with a transaction. We generate
237 features (>100 behavioral patterns) for each transaction, and use a random
forest to learn a classifier. We tested our machine learning model on data from
a large multinational bank and compared it to their existing solution. On an
unseen data of 1.852 million transactions, we were able to reduce the false
positives by 54% and provide a savings of 190K euros. We also assess how to
deploy this solution, and whether it necessitates streaming computation for
real time scoring. We found that our solution can maintain similar benefits
even when historical features are computed once every 7 days. | ['Santiago Moral Rubio', 'Kalyan Veeramachaneni', 'Roy Wedge', 'Sergio Iglesias Perez', 'James Max Kanter'] | 2017-10-20 | null | null | null | null | ['automated-feature-engineering'] | ['methodology'] | [-1.38863340e-01 3.36558856e-02 3.15709342e-03 -7.42903173e-01
-7.82982409e-01 -4.38325614e-01 2.64859229e-01 5.02250910e-01
-6.22481525e-01 8.45820963e-01 -3.06450650e-02 -3.90660554e-01
9.64374095e-02 -1.18879950e+00 -6.84907556e-01 -7.11151678e-03
-3.25529546e-01 6.37595236e-01 6.35894537e-02 -2.43767992e-01
6.65420055e-01 4.11397815e-01 -1.32237971e+00 9.82425988e-01
6.16113722e-01 1.24036694e+00 -6.24733388e-01 4.79589999e-01
1.59176812e-01 9.62992191e-01 -8.78709495e-01 -1.00938034e+00
5.22782445e-01 -1.39188424e-01 -6.35390401e-01 -1.03909984e-01
1.17320091e-01 -9.50214565e-01 -1.67282254e-01 5.11538208e-01
-7.34821707e-02 -5.11385620e-01 5.64419627e-01 -1.24257481e+00
-7.60230944e-02 6.22637510e-01 -7.93585122e-01 4.84637827e-01
4.26268816e-01 1.89560372e-02 1.23697793e+00 -7.85673499e-01
6.19727969e-01 5.94505012e-01 1.24565768e+00 -2.98304707e-02
-1.40608025e+00 -7.12288201e-01 -3.43159765e-01 1.31189376e-01
-9.27217901e-01 -3.02017212e-01 2.67740756e-01 -3.58971179e-01
1.59519124e+00 2.06971556e-01 1.18675566e+00 7.75990546e-01
4.52837288e-01 6.88270092e-01 6.00256324e-01 -3.95085335e-01
1.52247816e-01 2.73616642e-01 4.69216347e-01 5.19363880e-01
1.03369617e+00 9.63463038e-02 -6.26684844e-01 -1.06087661e+00
6.31515741e-01 3.53384674e-01 3.93755972e-01 -7.59582669e-02
-9.70798790e-01 1.32542157e+00 2.51003485e-02 1.30614206e-01
-5.82017541e-01 1.99257672e-01 5.92141867e-01 7.17952371e-01
2.51672238e-01 5.96679211e-01 -6.61792934e-01 -4.35785413e-01
-1.16750705e+00 8.26356590e-01 1.29028106e+00 5.63616037e-01
7.49279618e-01 -4.20470648e-02 3.32506120e-01 5.93930721e-01
-1.54914513e-01 5.93974441e-02 6.90388501e-01 -1.13326001e+00
6.91385090e-01 8.15983713e-01 4.68285978e-01 -1.16244555e+00
-3.73747796e-01 -1.04229733e-01 -4.79074657e-01 -7.21627846e-02
6.84362531e-01 -3.19135576e-01 -3.63492578e-01 9.27614570e-01
-1.94519296e-01 -2.18639001e-01 -2.33940884e-01 5.64116180e-01
-3.60330343e-01 2.59235770e-01 -3.24515969e-01 -2.95012712e-01
1.00418544e+00 -2.60712653e-01 -3.66978079e-01 7.64698312e-02
1.10176146e+00 -5.90714216e-01 6.54772937e-01 1.15880919e+00
-1.01161432e+00 3.02927420e-02 -1.14795959e+00 4.41739589e-01
-6.94137737e-02 -2.49519736e-01 1.16911280e+00 9.14015889e-01
-5.60298979e-01 9.19235945e-01 -1.15129066e+00 -2.07518771e-01
6.29648745e-01 4.23780084e-01 -1.98884323e-01 3.45940217e-02
-9.89812493e-01 8.32060754e-01 1.84976697e-01 -2.52489597e-01
-2.42236540e-01 -4.58002567e-01 -3.67728978e-01 1.98862970e-01
2.75860447e-02 -5.50657332e-01 1.42471957e+00 -1.21530426e+00
-6.15450621e-01 5.84511578e-01 3.44110839e-02 -1.20902264e+00
8.13504815e-01 -3.74013931e-01 -4.55474555e-01 -7.58886635e-02
1.00016400e-01 -1.26531497e-01 5.35469592e-01 -5.44335783e-01
-1.09195185e+00 -4.03330177e-01 -3.85962278e-01 -2.16344565e-01
-4.98838991e-01 2.40771964e-01 1.53804526e-01 -8.69260013e-01
1.65939167e-01 -6.19639099e-01 -1.35818586e-01 -4.69676435e-01
-7.92683288e-02 1.45163491e-01 2.54008591e-01 -7.95696378e-01
1.35863209e+00 -1.81571186e+00 -7.74737954e-01 6.51362479e-01
-8.66986290e-02 -9.48669761e-02 4.34495628e-01 6.29119873e-01
1.35764599e-01 2.17735305e-01 -1.96588248e-01 2.53264189e-01
-2.00722635e-01 4.55842838e-02 -6.96817458e-01 4.28904206e-01
4.67606634e-01 5.97824752e-01 -5.45331657e-01 -1.40954122e-01
-1.09183148e-01 3.46968845e-02 -1.04286802e+00 8.07097331e-02
1.41198233e-01 -6.86226726e-01 -1.92550689e-01 1.01937556e+00
5.52332163e-01 -2.54730642e-01 6.46511853e-01 -2.09975126e-03
2.36495882e-01 3.50838661e-01 -1.18321204e+00 9.89063919e-01
-6.40921816e-02 3.59084785e-01 -4.35765952e-01 -1.27755356e+00
1.02996886e+00 1.85158588e-02 7.55567193e-01 -7.11461604e-01
-2.18896523e-01 5.53515613e-01 -1.43740222e-01 -1.36330321e-01
5.88267744e-01 -3.69395554e-01 -4.89035189e-01 8.69295180e-01
-1.56051815e-01 3.43206644e-01 3.02543771e-02 3.91301811e-02
1.73249733e+00 -2.71596402e-01 4.76786286e-01 5.76862767e-02
-2.08921880e-01 5.27565598e-01 8.82232249e-01 8.54745686e-01
-1.87387407e-01 5.62244117e-01 9.17111695e-01 -1.28394401e+00
-1.00810087e+00 -9.81790066e-01 -1.11090682e-01 7.76822090e-01
-3.74345541e-01 -3.60245079e-01 -5.11032104e-01 -7.20074296e-01
9.11735356e-01 6.01555824e-01 -4.38752562e-01 -3.70011143e-02
-6.40852749e-01 -1.50296855e+00 7.25144506e-01 5.40968478e-01
2.61932164e-01 -8.14583421e-01 -1.10462272e+00 7.81446755e-01
-3.53881530e-02 -5.89505792e-01 5.47805876e-02 3.46160740e-01
-1.17195320e+00 -1.32263863e+00 -9.96561274e-02 -2.24072546e-01
3.18710923e-01 -1.18198402e-01 1.30423522e+00 2.32818350e-01
-6.61834836e-01 -3.44155222e-01 -2.25247115e-01 -4.74919260e-01
-1.48331687e-01 -5.20320982e-02 1.96914464e-01 -1.51217508e-03
1.14369452e+00 -4.38761324e-01 -4.68397498e-01 9.08402205e-02
-6.13874018e-01 -4.00390983e-01 5.96780062e-01 1.18097639e+00
2.36963809e-01 6.20196722e-02 9.04305816e-01 -1.35948288e+00
6.41541243e-01 -8.61107111e-01 -5.44313669e-01 3.80662456e-02
-1.20719576e+00 -2.54483800e-02 3.63762408e-01 -6.22656085e-02
-7.94415712e-01 3.12753826e-01 2.20081672e-01 -5.56490421e-02
1.36894181e-01 5.76561630e-01 5.59890449e-01 2.68425405e-01
6.87145174e-01 -8.92469734e-02 7.58933499e-02 -5.29874623e-01
-1.31756693e-01 7.70018280e-01 3.46546978e-01 -3.52403969e-01
3.40381205e-01 5.00968874e-01 -4.95059192e-01 -4.80247051e-01
-3.93149585e-01 -3.08780551e-01 -3.07029605e-01 2.38138795e-01
-4.83783111e-02 -1.05218339e+00 -8.87567043e-01 6.43295228e-01
-9.28680897e-01 5.96141890e-02 -2.38374591e-01 3.79036188e-01
-6.44616783e-01 2.37535551e-01 -1.10020208e+00 -8.64992797e-01
-4.66457278e-01 -2.49754801e-01 4.87665296e-01 -1.49611130e-01
-7.91947782e-01 -3.86484891e-01 2.28775337e-01 4.10662413e-01
3.02971691e-01 5.05081117e-01 8.46928954e-01 -1.15674365e+00
-3.27804357e-01 -6.91237807e-01 -3.24106932e-01 3.10752869e-01
1.06131017e-01 9.40672681e-02 -7.46291339e-01 -4.58116621e-01
1.31381080e-01 -3.83193433e-01 8.96280885e-01 1.11189395e-01
9.68585372e-01 -5.69778800e-01 -1.45436764e-01 5.05089343e-01
1.65083325e+00 2.31510982e-01 5.47441661e-01 7.67510712e-01
1.51708558e-01 5.59174359e-01 9.74832594e-01 1.14789760e+00
2.07164466e-01 5.07238328e-01 6.00033030e-02 3.32424462e-01
7.70517051e-01 -2.73077190e-01 3.67514938e-01 4.14057821e-01
-2.32575014e-01 2.30651543e-01 -1.13258195e+00 7.40656853e-01
-1.95777714e+00 -1.28154612e+00 -1.19417623e-01 2.27755404e+00
7.19219267e-01 7.28782475e-01 5.28087318e-01 3.57950300e-01
5.32372534e-01 -5.84037304e-01 -7.34467983e-01 -7.77038276e-01
9.60098282e-02 4.41784173e-01 9.33499336e-01 1.07503831e-01
-8.77884686e-01 4.65311915e-01 7.10626030e+00 6.91091195e-02
-7.64667928e-01 -2.13306576e-01 1.03932059e+00 -7.52413094e-01
-2.46801257e-01 4.89633419e-02 -7.17485607e-01 7.46388137e-01
1.36693716e+00 -4.53537345e-01 2.55758077e-01 1.09246397e+00
3.99708226e-02 -3.77542675e-01 -1.18163908e+00 9.63072836e-01
-7.73088709e-02 -1.61230540e+00 1.32751286e-01 5.81584036e-01
4.77767318e-01 1.01517133e-01 -4.09795433e-01 4.29137409e-01
4.54661787e-01 -1.00949860e+00 5.58509648e-01 7.58207262e-01
4.08643514e-01 -1.33219337e+00 1.15166700e+00 3.48952115e-01
-5.68587959e-01 -4.59949732e-01 -2.71502435e-01 -6.00736916e-01
4.09253910e-02 1.03104973e+00 -1.28507841e+00 1.90767363e-01
1.03788841e+00 6.33446693e-01 -5.35500228e-01 9.07839298e-01
6.75948620e-01 8.08559895e-01 -6.85178280e-01 9.93159115e-02
-1.03006423e-01 -5.33868074e-02 -1.34717599e-01 1.16698039e+00
5.57452321e-01 -1.84044731e-03 -2.89789855e-01 6.18854582e-01
-1.50019944e-01 -4.24452052e-02 -6.50149584e-01 1.41881164e-02
4.31414872e-01 7.63102829e-01 -4.19843376e-01 -4.43662941e-01
-5.90906560e-01 9.31819618e-01 3.36823255e-01 -1.05719686e-01
-6.66136861e-01 -8.17617059e-01 6.43638670e-01 6.09816313e-01
4.55412269e-01 1.74821094e-01 -6.65118217e-01 -1.43833208e+00
3.90496016e-01 -9.86176670e-01 6.71692312e-01 -1.88633248e-01
-1.60467017e+00 1.25920191e-01 -4.41935331e-01 -1.09597242e+00
-8.87868881e-01 -4.71347958e-01 -5.01040876e-01 8.36781919e-01
-1.02162397e+00 -4.57351476e-01 1.86979711e-01 3.83573830e-01
9.08611417e-02 -3.53466511e-01 9.13231075e-01 2.30740219e-01
-3.54084522e-01 7.78536618e-01 2.95855969e-01 2.72749096e-01
6.73652470e-01 -1.12244856e+00 8.74656737e-01 5.28093815e-01
1.28832966e-01 7.26639569e-01 5.14907598e-01 -8.04055989e-01
-1.32585454e+00 -8.26729715e-01 1.31360114e+00 -6.40202880e-01
7.03181922e-01 -2.50453323e-01 -1.01953578e+00 9.24684882e-01
-3.98477703e-01 4.48801629e-02 8.66279185e-01 2.17371464e-01
-3.54874521e-01 -3.25422645e-01 -1.59886420e+00 -1.81500673e-01
6.32436991e-01 -2.78869092e-01 -8.87205899e-01 5.04394546e-02
9.60207731e-02 4.32084650e-02 -1.11231196e+00 1.89705282e-01
1.10276198e+00 -1.56834257e+00 6.03343904e-01 -8.30520272e-01
3.19870323e-01 2.86409318e-01 -2.06981212e-01 -8.05029929e-01
-1.85473174e-01 -5.58175862e-01 -4.26318437e-01 9.17299628e-01
4.89502519e-01 -8.60806763e-01 1.28376567e+00 9.44561541e-01
4.76133615e-01 -8.81624043e-01 -8.71939182e-01 -6.88474417e-01
9.56417695e-02 -4.91648763e-01 7.98440635e-01 1.14247084e+00
5.73637307e-01 -2.44850710e-01 -2.09836438e-01 -2.95550048e-01
6.22786582e-01 3.90132964e-01 7.49955416e-01 -1.32510924e+00
-6.58522666e-01 -1.96279958e-01 -5.95918477e-01 -1.99511573e-01
-4.48364615e-01 -5.44429600e-01 -6.45846963e-01 -9.44087267e-01
4.76608098e-01 -5.11642575e-01 -5.69059849e-01 6.70964003e-01
2.69886345e-01 1.68464020e-01 -7.36500472e-02 4.51935172e-01
-6.53906018e-02 -2.66088754e-01 4.29817475e-02 9.76732820e-02
-1.01952195e-01 1.17809489e-01 -9.61694062e-01 7.22146690e-01
8.65439892e-01 -6.31629527e-01 1.34765744e-01 -1.88859954e-01
6.55336082e-01 2.36227930e-01 2.64832258e-01 -6.94357634e-01
-1.71139911e-01 -2.41424769e-01 8.89328122e-01 -5.80480218e-01
2.54208650e-02 -5.76790273e-01 2.48863488e-01 9.36327040e-01
-2.75714938e-02 2.12796584e-01 2.02068617e-03 6.34795308e-01
-3.16124350e-01 -1.27912179e-01 4.09842283e-01 -1.84823200e-01
-2.66995072e-01 -2.74429619e-01 -4.69522297e-01 -1.40701026e-01
1.09876132e+00 -1.64300397e-01 -2.41777182e-01 -2.49552265e-01
-4.64088053e-01 3.32421139e-02 8.07674408e-01 7.21440986e-02
5.72944164e-01 -1.27939939e+00 -7.47355044e-01 5.79378605e-01
1.33218929e-01 -4.48142260e-01 -2.76878595e-01 5.04028440e-01
-8.80737185e-01 4.63828236e-01 -4.37815070e-01 -2.45787680e-01
-8.39394510e-01 1.69400290e-01 2.01254427e-01 -4.62720424e-01
-5.48290491e-01 5.26941717e-01 -6.54655039e-01 -8.15407038e-02
-4.05778214e-02 -5.64042687e-01 2.10243061e-01 3.43237102e-01
5.56921124e-01 5.02180457e-01 5.90391159e-01 -1.75543632e-02
-4.62743849e-01 -2.10979238e-01 -6.98837936e-01 -2.52999902e-01
1.75401664e+00 5.24288595e-01 1.38827235e-01 4.00597602e-01
1.11537087e+00 -9.80039611e-02 -1.10727525e+00 1.68464363e-01
6.16873682e-01 -7.35160232e-01 -2.80474186e-01 -7.48098195e-01
-9.65497494e-01 3.93117011e-01 3.27671289e-01 6.10108197e-01
9.00566220e-01 -5.72147012e-01 1.18950772e+00 6.99423969e-01
8.90312970e-01 -1.42243648e+00 -3.48033130e-01 1.97153285e-01
3.90820801e-01 -1.13547122e+00 2.80867189e-01 1.45865873e-01
-7.60874450e-01 1.14481664e+00 3.23848039e-01 -8.13390553e-01
5.94967246e-01 2.99649447e-01 -2.00791046e-01 -1.68054625e-01
-1.17561948e+00 7.21387625e-01 -6.37503743e-01 1.88844711e-01
5.70452571e-01 3.16823423e-01 -4.78280485e-01 1.10495472e+00
-4.43023533e-01 4.20502633e-01 9.53578949e-01 1.39163160e+00
-5.59256852e-01 -1.09293258e+00 -2.82657325e-01 1.48486102e+00
-8.63427281e-01 5.16602732e-02 -3.87396097e-01 7.27116406e-01
-1.28470346e-01 7.82608509e-01 5.60495794e-01 -6.20293498e-01
5.13425589e-01 3.49990845e-01 2.90170670e-01 -6.26529276e-01
-1.02774811e+00 -2.50869215e-01 4.82975960e-01 -8.82942379e-01
4.95162122e-02 -1.39930618e+00 -1.20670927e+00 -7.95130968e-01
-2.48109683e-01 3.50035429e-02 5.17207921e-01 4.54219550e-01
4.32265937e-01 -1.49066344e-01 1.00584912e+00 -3.75555336e-01
-1.10639691e+00 -7.67780364e-01 -9.28007305e-01 5.36191344e-01
2.00006992e-01 -3.73608828e-01 -6.18886292e-01 2.24078819e-01] | [7.679744243621826, 5.4446611404418945] |
adb7d2ad-f02d-4120-95ca-8eca90af446c | deep-denoising-entity-pre-training-for-neural | 2111.07393 | null | https://arxiv.org/abs/2111.07393v1 | https://arxiv.org/pdf/2111.07393v1.pdf | DEEP: DEnoising Entity Pre-training for Neural Machine Translation | It has been shown that machine translation models usually generate poor translations for named entities that are infrequent in the training corpus. Earlier named entity translation methods mainly focus on phonetic transliteration, which ignores the sentence context for translation and is limited in domain and language coverage. To address this limitation, we propose DEEP, a DEnoising Entity Pre-training method that leverages large amounts of monolingual data and a knowledge base to improve named entity translation accuracy within sentences. Besides, we investigate a multi-task learning strategy that finetunes a pre-trained neural machine translation model on both entity-augmented monolingual data and parallel data to further improve entity translation. Experimental results on three language pairs demonstrate that \method results in significant improvements over strong denoising auto-encoding baselines, with a gain of up to 1.3 BLEU and up to 9.2 entity accuracy points for English-Russian translation. | ['Graham Neubig', 'Kyunghyun Cho', 'Hiroaki Hayashi', 'Junjie Hu'] | 2021-11-14 | null | https://aclanthology.org/2022.acl-long.123 | https://aclanthology.org/2022.acl-long.123.pdf | acl-2022-5 | ['transliteration'] | ['natural-language-processing'] | [ 1.48790345e-01 1.30490866e-02 -4.36410040e-01 -4.50799763e-01
-1.68979931e+00 -8.31794679e-01 5.43765068e-01 -2.43422911e-01
-4.50434029e-01 1.11741376e+00 5.94037116e-01 -6.93416834e-01
7.56065607e-01 -7.11711884e-01 -1.08042407e+00 -2.44765371e-01
6.82594836e-01 6.71375573e-01 -4.88735825e-01 -3.96663517e-01
-1.26626477e-01 -7.82929510e-02 -5.25003135e-01 4.20765460e-01
1.28468311e+00 3.72442514e-01 1.88357402e-02 1.71809077e-01
-3.28548104e-01 3.23223382e-01 -6.81746364e-01 -1.02059376e+00
3.53155702e-01 -7.46729374e-01 -7.93031812e-01 -2.55297929e-01
4.64963555e-01 -2.77492851e-01 -1.72847673e-01 9.68699574e-01
8.53398085e-01 -2.40852311e-01 5.36339760e-01 -5.46700478e-01
-1.13274348e+00 1.03190148e+00 -4.14711297e-01 2.58012086e-01
1.27853587e-01 1.18077574e-02 1.10986423e+00 -1.38055432e+00
8.63202333e-01 8.79787624e-01 8.95708680e-01 7.69605279e-01
-1.52475941e+00 -7.00295448e-01 -2.58007884e-01 -6.18262477e-02
-1.27678573e+00 -7.86188126e-01 4.40503210e-01 -1.39423043e-01
1.46210730e+00 -1.47984087e-01 2.00221598e-01 1.63310182e+00
4.52461541e-01 7.82684803e-01 1.18046725e+00 -4.93517488e-01
-6.15512617e-02 6.24338388e-02 -2.63388664e-01 2.07001537e-01
6.24707416e-02 1.04854122e-01 -5.72792709e-01 -3.95720154e-02
5.68689227e-01 -5.69288790e-01 -1.06648922e-01 1.62508950e-01
-1.51186442e+00 7.93853939e-01 1.56500280e-01 3.36231202e-01
-6.20687246e-01 1.41763082e-02 5.96452773e-01 6.15640402e-01
8.37517321e-01 6.70188308e-01 -7.71160305e-01 -4.21709627e-01
-1.05351651e+00 -1.43975124e-01 7.55248189e-01 1.12898159e+00
7.84018934e-01 3.37423682e-01 -2.82928437e-01 9.54875827e-01
-1.86555728e-01 1.11920583e+00 6.79515600e-01 -4.01269734e-01
1.03342497e+00 3.22114229e-01 6.78750046e-04 -5.02388537e-01
-8.05922821e-02 -8.63880396e-01 -6.92581117e-01 -5.52763522e-01
7.87074491e-02 -5.46993792e-01 -9.91920948e-01 1.98676991e+00
2.22820327e-01 3.66889946e-02 4.69256192e-01 7.11097598e-01
6.93690240e-01 7.55281329e-01 1.29772231e-01 -1.12020627e-01
1.23195875e+00 -1.19405150e+00 -8.35155666e-01 -5.00724554e-01
9.69814658e-01 -1.12994778e+00 1.27380908e+00 -2.32158542e-01
-1.21140873e+00 -5.24383664e-01 -8.44265997e-01 -2.96261281e-01
-1.65007129e-01 5.09478629e-01 4.23978448e-01 5.05509079e-01
-9.44490016e-01 4.33427602e-01 -1.07661796e+00 -2.84126312e-01
2.10744426e-01 3.43876630e-01 -4.22927529e-01 -5.02384594e-03
-1.42540228e+00 1.23018682e+00 2.20436588e-01 7.44471550e-02
-7.22249746e-01 -9.34416354e-01 -8.24248910e-01 3.76977585e-02
1.98510531e-02 -1.10796857e+00 1.33834350e+00 -1.07764721e+00
-1.66070366e+00 8.12968791e-01 -5.31907380e-01 -5.02003491e-01
4.76974577e-01 -4.09251422e-01 -4.59786981e-01 -1.71601817e-01
3.82726520e-01 6.06095791e-01 4.17917758e-01 -8.72176051e-01
-5.46603441e-01 -2.32533157e-01 -3.82550031e-01 4.50566918e-01
-3.08255404e-01 3.43896508e-01 -2.86423385e-01 -1.00232816e+00
-1.58290163e-01 -1.00156808e+00 -3.08184594e-01 -9.06840026e-01
-4.72617596e-01 5.85383959e-02 2.79432237e-01 -1.15475833e+00
1.13499784e+00 -1.72958589e+00 2.64160901e-01 -2.22112209e-01
-2.35238224e-01 2.80719072e-01 -5.57676256e-01 6.15363538e-01
9.80844945e-02 2.73370355e-01 -2.90133268e-01 -6.54331326e-01
-9.51853618e-02 1.31218150e-01 -5.85607469e-01 1.08578213e-01
6.25360548e-01 1.43524170e+00 -7.35671580e-01 -2.25767523e-01
-2.28286698e-01 7.15862274e-01 -3.21782440e-01 4.02982756e-02
-4.43968959e-02 8.26889515e-01 -3.09811175e-01 6.40764832e-01
6.01956785e-01 -6.60528988e-02 2.30269417e-01 1.19209457e-02
-1.90029815e-02 1.30370474e+00 -4.68924552e-01 2.02918053e+00
-9.10001457e-01 5.14361978e-01 -3.06575447e-01 -5.83895802e-01
9.17383075e-01 6.85512125e-01 2.57154524e-01 -7.95535922e-01
9.33658332e-02 7.76391864e-01 -7.11152256e-02 -2.06135258e-01
6.51262164e-01 -4.39489543e-01 -2.67539799e-01 6.07225358e-01
2.98783839e-01 -8.73344368e-04 4.65779454e-02 -1.78232118e-01
1.13651061e+00 4.11711723e-01 1.11846469e-01 -1.77914854e-02
1.82853580e-01 3.48349780e-01 1.00141060e+00 4.27166671e-01
-8.05832259e-03 7.29086459e-01 2.90556736e-02 -2.44254261e-01
-1.51842380e+00 -1.03395641e+00 -3.23090144e-02 1.09232759e+00
-2.26995036e-01 -2.71327198e-01 -8.78827274e-01 -8.25060785e-01
-3.15214962e-01 1.03718913e+00 -2.79339731e-01 -2.95509905e-01
-1.22119927e+00 -1.00576937e+00 9.58002090e-01 7.19924867e-01
3.53813350e-01 -1.01321316e+00 2.08256796e-01 5.09217560e-01
-8.82695138e-01 -1.40256107e+00 -6.80727899e-01 1.51905224e-01
-1.10231662e+00 -3.02955955e-01 -8.05302083e-01 -9.60597932e-01
5.20188034e-01 -5.96226491e-02 1.55610216e+00 -4.97550577e-01
4.70133066e-01 -2.57178992e-01 -3.28818113e-01 -2.48485625e-01
-9.10121918e-01 8.73314977e-01 2.37558201e-01 -2.43564337e-01
8.79470766e-01 -6.73488021e-01 -2.22168341e-01 1.92496136e-01
-4.44203287e-01 9.06163454e-02 1.06749058e+00 1.21612692e+00
7.44100988e-01 -7.30682850e-01 9.77940321e-01 -8.04820716e-01
5.72902918e-01 -4.83634740e-01 -2.95800507e-01 4.39850539e-01
-7.54802227e-01 1.62314162e-01 6.83661163e-01 -5.20070553e-01
-1.17366493e+00 -8.09445754e-02 -2.76713163e-01 -1.08061291e-01
3.08246184e-02 8.17658901e-01 -3.21491987e-01 2.36272886e-01
8.32749367e-01 4.84418154e-01 -3.06104153e-01 -4.55723882e-01
4.88211781e-01 8.49667549e-01 5.97503185e-01 -5.46307802e-01
8.82316291e-01 3.07752416e-02 -4.86941665e-01 -2.43292838e-01
-7.45923936e-01 -2.57051826e-01 -7.69474030e-01 3.59040231e-01
8.07883441e-01 -1.38966513e+00 1.13113225e-01 1.60092667e-01
-1.43655682e+00 -3.62264395e-01 -1.75615177e-01 8.82984757e-01
-5.33527792e-01 6.70740977e-02 -1.01097381e+00 -2.43412763e-01
-8.78215492e-01 -9.96241629e-01 1.26428866e+00 -1.72932521e-01
-2.03577027e-01 -1.00099099e+00 3.32235426e-01 5.99671483e-01
6.53562188e-01 -1.25709400e-01 8.78267586e-01 -8.82513702e-01
-5.94154656e-01 -2.42225565e-02 -1.64321419e-02 4.80651945e-01
1.83072507e-01 -3.82388800e-01 -6.38240695e-01 -3.40688109e-01
9.37950145e-03 -3.16247344e-01 7.62573898e-01 2.26348378e-02
1.32419243e-01 -3.15221995e-01 -2.55155265e-01 6.59549713e-01
1.19326258e+00 -8.91274512e-02 6.45788133e-01 3.41524363e-01
7.76133358e-01 3.27444643e-01 4.44575638e-01 -6.59134239e-02
5.52744985e-01 6.57455325e-01 -9.00171995e-02 -3.98306161e-01
-3.71409297e-01 -5.27471483e-01 6.93271041e-01 1.62655485e+00
1.66288838e-02 -1.11444227e-01 -1.04484510e+00 9.91905510e-01
-1.54607379e+00 -7.62065649e-01 -1.18782721e-01 2.11110067e+00
1.48711836e+00 -1.67859301e-01 -1.64079130e-01 -6.12006903e-01
8.49215686e-01 -1.34904683e-01 -4.98690456e-01 -4.34419572e-01
-5.34319282e-01 5.98792076e-01 6.71331525e-01 4.14321274e-01
-9.21252489e-01 1.61776972e+00 6.24658060e+00 8.29773128e-01
-1.32236683e+00 5.65723360e-01 4.76112127e-01 8.26297179e-02
-5.47777057e-01 1.43851742e-01 -9.72376883e-01 3.28116298e-01
1.44522882e+00 -2.44382769e-01 3.75344127e-01 5.71538866e-01
3.59582305e-01 6.28503799e-01 -1.12628353e+00 6.14919484e-01
4.13069241e-02 -1.26688862e+00 7.63274506e-02 1.66612953e-01
1.11963010e+00 5.72112083e-01 8.64209533e-02 5.58474660e-01
4.55948710e-01 -9.63814259e-01 6.40529633e-01 1.09572753e-01
9.61389184e-01 -8.18299353e-01 9.98620629e-01 2.95873940e-01
-8.10692370e-01 3.92935514e-01 -4.55292046e-01 1.43325686e-01
6.04419708e-01 5.65796852e-01 -1.17389357e+00 7.61470139e-01
2.49205574e-01 6.50102377e-01 -2.09796309e-01 6.61731780e-01
-5.55886686e-01 1.09297156e+00 -2.79645830e-01 2.23539904e-01
2.10590303e-01 -2.82115400e-01 5.08635461e-01 1.38476813e+00
5.87710500e-01 -2.34572768e-01 -6.82750121e-02 8.98617208e-01
-5.97700000e-01 5.09610057e-01 -5.58641374e-01 -3.75511199e-01
6.11554384e-01 9.87262726e-01 -1.64251655e-01 -4.28099960e-01
-6.12099409e-01 1.50198603e+00 5.39194286e-01 5.27355075e-01
-9.24358070e-01 -2.94623971e-01 8.16959441e-01 -2.09140778e-01
5.17438948e-01 -4.24955577e-01 -6.55865431e-01 -1.64389455e+00
2.41530418e-01 -1.23003411e+00 -2.43363276e-01 -3.30195844e-01
-1.35009909e+00 9.59240913e-01 -6.63700402e-01 -1.23860610e+00
-5.42100966e-01 -2.93274105e-01 -3.49971861e-01 1.27897835e+00
-1.51562512e+00 -1.65703988e+00 4.22947198e-01 1.24903157e-01
6.12072468e-01 -3.05209756e-01 1.03420055e+00 6.59938574e-01
-6.27008319e-01 1.06462026e+00 5.16768336e-01 4.87444222e-01
1.12326419e+00 -1.16774571e+00 1.29545379e+00 1.07989526e+00
4.65906769e-01 9.41380680e-01 3.65257382e-01 -8.81954193e-01
-1.26491296e+00 -1.38627326e+00 1.85373425e+00 -8.38305235e-01
6.42071545e-01 -5.29300988e-01 -8.63784552e-01 1.02016068e+00
5.52621722e-01 -3.43060374e-01 9.56248641e-01 3.75878513e-01
-5.65416157e-01 1.38525113e-01 -8.67539406e-01 6.80055618e-01
1.14752281e+00 -8.80057216e-01 -7.92801738e-01 4.50900584e-01
8.77005458e-01 -6.19060636e-01 -1.02019608e+00 5.45724034e-01
3.05495381e-01 -2.97299385e-01 7.32656419e-01 -1.05460942e+00
6.68167472e-01 -1.24667078e-01 -3.09599221e-01 -1.68262827e+00
-1.30440995e-01 -6.87095523e-01 1.09222814e-01 1.42179561e+00
1.14860725e+00 -5.66965342e-01 6.27893031e-01 2.64846772e-01
-3.76854658e-01 -6.56082511e-01 -1.07342911e+00 -9.11669850e-01
6.50964141e-01 -2.19278708e-01 6.62124276e-01 1.25398898e+00
-1.48495957e-01 1.04093087e+00 -4.83578116e-01 3.90047841e-02
1.92977861e-01 1.74484581e-01 7.52325594e-01 -5.72028577e-01
-3.04788232e-01 -1.94731042e-01 -1.66070685e-02 -1.28520632e+00
6.06139839e-01 -1.24246097e+00 1.11338019e-01 -1.62642837e+00
3.07794720e-01 -3.46567780e-01 -1.06190562e-01 6.18223846e-01
-5.64271390e-01 5.08142352e-01 -1.22983173e-01 4.11698490e-01
-1.09126687e-01 7.82989144e-01 1.02236581e+00 -5.53459674e-02
-2.08514184e-01 -9.72310826e-02 -6.69674754e-01 1.46576524e-01
7.90893137e-01 -7.10441113e-01 -4.90325456e-03 -1.19157469e+00
2.70080745e-01 -8.57967660e-02 -1.62574351e-01 -6.49392486e-01
-1.55490205e-01 1.21560681e-03 1.67226717e-01 -3.92926127e-01
1.87206030e-01 -4.08446431e-01 1.18887730e-01 1.21084265e-01
-3.91527474e-01 4.89359558e-01 2.26950988e-01 2.89436996e-01
-4.31909561e-01 6.74794018e-02 4.32564318e-01 -7.55988881e-02
-2.74374604e-01 9.58314538e-02 -3.05779487e-01 2.75689930e-01
5.06064415e-01 1.36675850e-01 -3.20905775e-01 -2.25105241e-01
-4.80515331e-01 -5.99059649e-02 5.15623331e-01 5.30344605e-01
1.37015790e-01 -1.70245326e+00 -1.32052195e+00 5.58160152e-03
-6.38735155e-03 -3.05756599e-01 -1.21096849e-01 9.96748805e-01
-3.31798881e-01 5.08000433e-01 -7.73755163e-02 -4.06289876e-01
-9.17226493e-01 3.63234729e-01 3.21897209e-01 -4.31718260e-01
-3.63631159e-01 8.44905138e-01 -1.43991113e-01 -1.07117176e+00
-3.28549683e-01 -3.89163196e-01 3.58843923e-01 -2.69092202e-01
2.38192946e-01 1.48902118e-01 4.70999688e-01 -1.01023483e+00
-2.89871752e-01 4.77509141e-01 -1.84697777e-01 -5.13415277e-01
1.23539221e+00 -3.32619011e-01 -8.86456594e-02 3.49183679e-01
1.25500679e+00 5.21075189e-01 -7.98518419e-01 -5.60044050e-01
1.70815825e-01 -2.45647594e-01 -1.29966035e-01 -1.15916049e+00
-7.56627858e-01 9.23683226e-01 1.73294380e-01 -5.98284543e-01
1.02693784e+00 -1.56059459e-01 1.55939424e+00 5.34541547e-01
4.25007045e-01 -1.08767521e+00 -3.27873647e-01 1.02755034e+00
6.60679102e-01 -1.48242664e+00 -5.18142581e-01 -3.48733038e-01
-7.29181767e-01 8.38331878e-01 3.83639514e-01 1.72524214e-01
1.44153342e-01 3.67204517e-01 7.30151296e-01 3.15837115e-01
-6.26052797e-01 -1.01040095e-01 3.00421208e-01 4.95216966e-01
8.33414316e-01 1.12994835e-01 -6.21666312e-01 7.27893114e-01
-5.25252640e-01 -5.33767976e-02 2.15857327e-01 6.34954870e-01
-3.64257023e-02 -1.55014157e+00 -7.96943456e-02 6.88921139e-02
-9.12669659e-01 -9.10334945e-01 -5.89477181e-01 4.27333504e-01
-1.14048362e-01 9.77477074e-01 -1.97665393e-02 -4.58082348e-01
2.89060771e-01 4.47840393e-01 3.39255959e-01 -7.22861290e-01
-9.83813405e-01 1.18824720e-01 4.63022053e-01 -3.92457306e-01
-2.31398329e-01 -6.36233568e-01 -1.22630012e+00 -3.47888559e-01
-5.27676344e-01 2.36105233e-01 9.77369249e-01 1.03100598e+00
9.81717288e-01 3.35762978e-01 6.02583230e-01 -3.33798796e-01
-9.34654474e-01 -1.24581158e+00 1.59650192e-01 2.44268298e-01
8.95325914e-02 -9.22268108e-02 -8.18312094e-02 2.01517791e-01] | [11.638959884643555, 10.267261505126953] |
5a7c03e6-5323-4709-8c88-c38f963199d0 | neural-architecture-search-for-visual-anomaly | 2304.08975 | null | https://arxiv.org/abs/2304.08975v2 | https://arxiv.org/pdf/2304.08975v2.pdf | Neural Architecture Search for Visual Anomaly Segmentation | This paper presents the first application of neural architecture search to the complex task of segmenting visual anomalies. Measurement of anomaly segmentation performance is challenging due to imbalanced anomaly pixels, varying region areas, and various types of anomalies. First, the region-weighted Average Precision (rwAP) metric is proposed as an alternative to existing metrics, which does not need to be limited to a specific maximum false positive rate. Second, the AutoPatch neural architecture search method is proposed, which enables efficient segmentation of visual anomalies without any training. By leveraging a pre-trained supernet, a black-box optimization algorithm can directly minimize computational complexity and maximize performance on a small validation set of anomalous examples. Finally, compelling results are presented on the widely studied MVTec dataset, demonstrating that AutoPatch outperforms the current state-of-the-art with lower computational complexity, using only one example per type of anomaly. The results highlight the potential of automated machine learning to optimize throughput in industrial quality control. The code for AutoPatch is available at: https://github.com/tommiekerssies/AutoPatch | ['Joaquin Vanschoren', 'Tommie Kerssies'] | 2023-04-18 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 4.37112510e-01 -1.73571065e-01 -1.67312294e-01 -1.62375867e-01
-7.18395531e-01 -6.45906746e-01 1.68692261e-01 6.77053928e-01
-2.02720352e-02 2.13482782e-01 -7.74295747e-01 -5.03761768e-01
-3.78390312e-01 -5.49260437e-01 -6.91357493e-01 -6.98698699e-01
-1.04557373e-01 4.10754353e-01 1.51575774e-01 1.65913373e-01
6.21311665e-01 7.23366737e-01 -1.54552209e+00 4.67672825e-01
8.25057209e-01 1.55914164e+00 9.32930782e-03 7.02946842e-01
-9.38721895e-02 3.39728683e-01 -1.01557934e+00 -1.72537282e-01
5.53412080e-01 -5.49040020e-01 -4.56288040e-01 2.96012789e-01
5.30493975e-01 -8.22203383e-02 2.23546460e-01 1.03233790e+00
3.47028196e-01 5.84118180e-02 5.75821817e-01 -1.48245060e+00
-3.82588476e-01 2.76007503e-01 -8.12855124e-01 7.52463698e-01
-1.09245591e-01 6.42754972e-01 1.10573065e+00 -7.10594833e-01
3.56021255e-01 8.17100346e-01 5.19624174e-01 1.54408380e-01
-1.30056381e+00 -5.61108410e-01 3.97241771e-01 5.02239287e-01
-1.23736680e+00 1.10680901e-01 7.46129513e-01 -6.82851255e-01
1.25459421e+00 4.04590786e-01 6.49710715e-01 8.33175898e-01
4.04700428e-01 6.12875521e-01 8.19466531e-01 -3.55549872e-01
4.16889548e-01 -9.57894772e-02 1.48504779e-01 6.40376568e-01
5.96132398e-01 3.90735678e-02 -4.31571603e-01 -1.16206415e-01
5.15472293e-01 9.39350128e-02 6.65037483e-02 -5.64006686e-01
-1.02422631e+00 8.19545805e-01 2.92640984e-01 1.90700009e-01
-5.70179522e-01 -2.11591087e-02 7.75613844e-01 3.40959787e-01
5.43550849e-01 1.16712892e+00 -6.84740663e-01 -1.42895162e-01
-1.03151429e+00 8.35405216e-02 4.47104841e-01 6.05473936e-01
4.44361389e-01 4.93770301e-01 -3.67947847e-01 7.74937749e-01
2.65126199e-01 2.66127855e-01 2.58783877e-01 -8.71522725e-01
3.00849676e-01 1.02040601e+00 -2.96514541e-01 -9.42690253e-01
-4.69523191e-01 -5.62687099e-01 -5.88721812e-01 8.05236936e-01
7.46456683e-01 -5.21767652e-03 -1.29600370e+00 1.01925337e+00
2.84440309e-01 1.19380578e-02 -3.18504989e-01 7.27541685e-01
1.89791188e-01 5.27349532e-01 -2.01013446e-01 -2.41243333e-01
1.19633412e+00 -8.55417967e-01 -5.04697859e-01 -2.28834361e-01
6.55949354e-01 -7.70122766e-01 1.05113029e+00 9.65340972e-01
-9.25655723e-01 -2.75780231e-01 -1.39011669e+00 5.17878532e-01
-7.42779791e-01 1.62026256e-01 3.21059585e-01 7.09838867e-01
-6.05870247e-01 6.85789347e-01 -8.41163754e-01 -2.66828835e-01
6.88114166e-01 2.82446027e-01 1.54780179e-01 1.85108095e-01
-4.85773712e-01 6.50940180e-01 5.38791001e-01 2.53761768e-01
-9.08987761e-01 -8.18844676e-01 -6.63562238e-01 -5.84669858e-02
8.34566951e-01 -1.45312786e-01 1.17082143e+00 -1.01465130e+00
-1.22413158e+00 6.46159768e-01 1.80641964e-01 -7.78768837e-01
4.70727593e-01 -3.72518778e-01 -5.20447791e-01 1.70949504e-01
-2.16014072e-01 3.00739914e-01 7.07733750e-01 -1.16164577e+00
-6.86910510e-01 -3.42055470e-01 -3.07209402e-01 -2.21616894e-01
-7.94906691e-02 -2.01126523e-02 -2.65638500e-01 -6.38743818e-01
2.11934341e-04 -6.34665012e-01 -4.13085938e-01 -2.20198914e-01
-7.30900943e-01 -1.18819930e-01 9.15887713e-01 -6.71817183e-01
1.41317225e+00 -1.99287915e+00 -1.95142582e-01 7.45797515e-01
1.40555203e-01 3.94889295e-01 -6.06585741e-02 2.29858682e-01
-4.00249839e-01 3.70156318e-02 -5.81337094e-01 2.85396695e-01
7.14487806e-02 -8.56654570e-02 -2.75382310e-01 5.66062629e-01
6.09705389e-01 6.38499439e-01 -6.37272179e-01 -1.94952026e-01
6.08349800e-01 -1.27723575e-01 -3.84615511e-01 9.90396366e-03
-5.17310798e-01 2.71509200e-01 -1.76580593e-01 1.19471979e+00
4.98952121e-01 -4.80753332e-01 8.27385336e-02 4.73497398e-02
-1.48672938e-01 -8.69059488e-02 -1.11461008e+00 1.46732330e+00
-1.43984303e-01 6.28120244e-01 -2.37290457e-01 -1.22189343e+00
1.08514833e+00 1.46800112e-02 8.55424225e-01 -1.01559234e+00
1.82568699e-01 2.62737602e-01 8.33528265e-02 -3.84295970e-01
2.81824023e-01 6.23098195e-01 3.51439305e-02 3.75836760e-01
-2.22855099e-02 1.04116753e-01 5.38906932e-01 -5.22794686e-02
1.41246462e+00 7.04449117e-02 4.29032981e-01 -1.81094334e-01
2.71844506e-01 4.55851048e-01 5.68037629e-01 7.49016821e-01
-3.11491787e-01 7.45698869e-01 9.19664502e-01 -5.98694205e-01
-1.25742865e+00 -1.09847224e+00 -1.30774334e-01 8.92724156e-01
1.23949677e-01 -2.18802169e-01 -7.54242003e-01 -9.54944193e-01
1.71500251e-01 8.83417010e-01 -5.06609917e-01 -1.06019996e-01
-5.87043464e-01 -9.62878883e-01 4.17218924e-01 4.43769574e-01
1.11694470e-01 -1.24907434e+00 -1.10713577e+00 1.08627364e-01
2.82751948e-01 -9.32454824e-01 -9.10234600e-02 3.65592271e-01
-9.52431500e-01 -1.40507197e+00 -3.47905308e-01 -2.98878878e-01
7.20502019e-01 -2.27624863e-01 1.45019174e+00 1.10683016e-01
-9.22168791e-01 4.17989939e-01 -2.53874660e-01 -7.55574882e-01
-2.56190211e-01 -6.34029321e-03 -2.59263664e-01 -1.53280422e-01
6.20344758e-01 -2.41015598e-01 -6.96493626e-01 3.61022323e-01
-8.08863640e-01 -4.29286629e-01 8.22753489e-01 8.54691446e-01
1.03473234e+00 -4.30767946e-02 4.86085624e-01 -7.26014316e-01
4.79895979e-01 -5.53264678e-01 -1.22383642e+00 1.38264209e-01
-1.21335566e+00 -1.89802989e-01 3.93084407e-01 -4.55797821e-01
-6.40687823e-01 1.86178088e-01 1.72948241e-01 -7.47292757e-01
-4.15785134e-01 2.26963595e-01 1.41757369e-01 2.65674502e-01
7.70821869e-01 -8.07891563e-02 1.97303712e-01 -4.19421792e-01
8.87895748e-02 2.30406672e-01 4.50264871e-01 -2.53698260e-01
5.24298906e-01 2.00960487e-01 1.84150293e-01 -5.51880896e-01
-4.47350919e-01 -6.86397254e-01 -6.52321100e-01 -4.54420209e-01
1.03465867e+00 -4.41311419e-01 -7.30319500e-01 3.79545361e-01
-8.73081565e-01 -3.95441055e-01 -6.68858290e-01 2.25485221e-01
-5.29418290e-01 3.36905658e-01 -4.17820215e-01 -9.19980586e-01
-4.77575630e-01 -1.14975286e+00 8.97881031e-01 5.72072268e-02
-3.62218559e-01 -5.98943710e-01 -1.11169897e-01 1.38034433e-01
2.02123016e-01 7.62831926e-01 9.97473359e-01 -1.25248885e+00
-8.66082549e-01 -3.50342125e-01 -1.76664799e-01 5.33137023e-01
4.04666178e-02 3.04159760e-01 -7.50099361e-01 -1.48773164e-01
-1.94597006e-01 -3.46033089e-02 7.13704646e-01 6.22893333e-01
1.57778275e+00 -2.42775843e-01 -2.33026937e-01 4.27473605e-01
1.41936493e+00 5.96413970e-01 5.15106380e-01 7.90635705e-01
6.16002321e-01 4.58366960e-01 9.98547256e-01 6.66419268e-01
-2.79190302e-01 5.29824555e-01 1.06311715e+00 -3.25803041e-01
1.53464168e-01 3.94807160e-01 1.60819903e-01 2.15197563e-01
-5.84110105e-03 -3.14939082e-01 -1.04747617e+00 6.19075596e-01
-1.78065193e+00 -8.72269750e-01 -1.98058575e-01 2.40059900e+00
2.53244430e-01 2.58912772e-01 4.49378133e-01 2.87704229e-01
7.55415618e-01 5.74205108e-02 -9.22875285e-01 -8.36181104e-01
8.59561712e-02 1.25525430e-01 7.06130087e-01 -2.94377524e-02
-1.20342064e+00 4.33041364e-01 6.38150454e+00 6.67920589e-01
-9.29534018e-01 -6.92209527e-02 8.67198825e-01 -4.68683630e-01
2.31805757e-01 -3.86872828e-01 -5.92750609e-01 6.80854201e-01
1.06989706e+00 5.37336580e-02 3.55289549e-01 1.04139137e+00
7.63312280e-02 -9.36653763e-02 -1.01943016e+00 7.43344784e-01
1.29938096e-01 -1.21480680e+00 -2.35732839e-01 2.52222747e-01
6.72022343e-01 9.83543098e-02 2.99492329e-01 -2.43817288e-02
2.18048692e-01 -9.82324302e-01 5.19061685e-01 4.57182318e-01
3.75988483e-01 -9.51549232e-01 7.70759165e-01 -2.70193219e-01
-9.00606036e-01 -6.06347740e-01 -1.32338375e-01 4.94350642e-01
4.46579941e-02 8.94227028e-01 -1.17594624e+00 5.99963665e-01
1.10900700e+00 3.79990637e-01 -7.48257875e-01 1.46449482e+00
2.12299794e-01 7.19901383e-01 -3.71315777e-01 4.02389802e-02
2.72549838e-01 -1.99217424e-01 8.38064373e-01 1.27142894e+00
6.06238067e-01 -5.92122793e-01 1.47199571e-01 1.01322281e+00
2.50749916e-01 1.76993534e-01 -6.15456879e-01 -2.13041186e-01
2.60857493e-01 1.23567808e+00 -1.14169800e+00 -1.65747717e-01
-2.89396733e-01 7.88638055e-01 -2.02327386e-01 2.92243600e-01
-8.79018545e-01 -5.10879278e-01 6.94392443e-01 6.72552437e-02
5.77229023e-01 7.82232061e-02 -8.84419918e-01 -3.09956968e-01
1.72508478e-01 -1.15167809e+00 9.13832903e-01 -5.82395136e-01
-1.28050554e+00 5.44790089e-01 7.13792667e-02 -1.62655711e+00
-4.87059683e-01 -1.03657103e+00 -6.49908066e-01 5.37217975e-01
-1.11456347e+00 -6.96384609e-01 -3.27686518e-01 2.25872934e-01
9.53512132e-01 -6.19033039e-01 4.76498157e-01 6.04074635e-02
-1.00804996e+00 5.53911150e-01 6.65462837e-02 -1.77956894e-01
5.64405262e-01 -1.55773997e+00 4.12073791e-01 1.38422573e+00
1.28209874e-01 1.01764984e-02 9.11577761e-01 -7.07816839e-01
-9.69168007e-01 -1.17706215e+00 4.73481491e-02 -5.43261051e-01
6.75684214e-01 -1.17681213e-01 -1.06469941e+00 5.76693356e-01
2.70868868e-01 -4.25034985e-02 5.53009987e-01 9.91398543e-02
-2.00034633e-01 -2.03548580e-01 -1.15018749e+00 3.89362574e-01
6.31650984e-01 -3.11492290e-02 -1.36492148e-01 5.81671953e-01
6.29575610e-01 -6.14971757e-01 -8.59180689e-01 5.30100942e-01
2.84565687e-01 -1.01712823e+00 9.72401738e-01 -3.29484493e-01
4.48067009e-01 -5.50076365e-01 5.36297336e-02 -1.16416490e+00
-8.71093422e-02 -4.32752073e-01 -4.86549735e-01 1.08685923e+00
9.08261657e-01 -7.11387932e-01 7.62957990e-01 2.47805893e-01
-5.05927145e-01 -1.04918861e+00 -8.55537891e-01 -1.07332647e+00
-2.00457945e-01 -5.64452171e-01 6.72280669e-01 8.86930048e-01
-2.12260813e-01 -3.32704246e-01 -1.26545206e-01 4.66801852e-01
4.59539354e-01 3.61878648e-02 5.00205398e-01 -1.27356958e+00
-3.06559205e-01 -7.02220559e-01 -5.76510489e-01 -2.87815422e-01
-3.40242803e-01 -6.74829900e-01 8.98294002e-02 -1.27179885e+00
-3.14617246e-01 -5.36455289e-02 -7.89064646e-01 5.63507259e-01
4.02133055e-02 3.50656420e-01 -3.16635077e-03 3.56780435e-03
-6.12471282e-01 -4.04778719e-02 9.50820208e-01 -3.00862134e-01
-3.93040240e-01 1.57392696e-01 -4.37757999e-01 6.32180810e-01
1.19134367e+00 -5.08269131e-01 -3.56888741e-01 -3.30045037e-02
1.63734302e-01 -5.06286681e-01 4.07118350e-01 -1.09994280e+00
3.50793935e-02 -1.66869253e-01 6.59740210e-01 -9.05598044e-01
1.06426284e-01 -8.63693893e-01 -9.73963663e-02 5.18761158e-01
-1.32359162e-01 8.21694851e-01 5.29700696e-01 6.23228133e-01
-6.43978342e-02 -3.00908923e-01 6.81428194e-01 -3.47160548e-02
-9.67208147e-01 8.11714604e-02 -2.74906039e-01 -2.41562620e-01
1.49429584e+00 -4.80846316e-01 -3.66854370e-01 2.11165160e-01
-4.55586135e-01 1.79597735e-01 3.69593114e-01 4.20930892e-01
5.13290823e-01 -1.07997847e+00 -4.20255274e-01 2.33494520e-01
3.85907501e-01 2.92748455e-02 3.67744625e-01 8.59997451e-01
-6.92391276e-01 2.91852802e-01 -3.79471660e-01 -1.09604657e+00
-1.39433599e+00 8.40416253e-01 4.19404477e-01 -4.80942845e-01
-7.39292979e-01 5.66076994e-01 -1.19411029e-01 -1.25956029e-01
2.12725803e-01 -2.98628628e-01 -1.26534536e-01 5.04499003e-02
3.91948164e-01 7.48840451e-01 4.38902140e-01 -1.46728501e-01
-2.93079168e-01 3.34051341e-01 -1.16886675e-01 1.95319518e-01
1.08717978e+00 1.19309187e-01 9.91219655e-02 5.12023270e-01
6.32893384e-01 -2.81365007e-01 -1.29367840e+00 1.15070358e-01
3.14561546e-01 -4.70494896e-01 -3.48965712e-02 -1.12342513e+00
-1.39685559e+00 7.09038198e-01 1.20212173e+00 5.30880630e-01
1.56487370e+00 -1.66169614e-01 4.83046681e-01 2.66042352e-01
-2.37325817e-01 -1.44651330e+00 3.64125967e-01 1.64102584e-01
7.60234177e-01 -1.18865812e+00 1.51790306e-01 -1.72122926e-01
-9.18994248e-01 1.05674839e+00 9.60611761e-01 -5.02176508e-02
3.04654777e-01 4.00875688e-01 2.46382132e-01 -5.27778864e-01
-6.86251581e-01 -1.12062521e-01 4.21564817e-01 5.21592379e-01
7.45635703e-02 -1.24305531e-01 7.97885060e-02 1.36922061e-01
6.93035647e-02 -4.57341164e-01 3.51450115e-01 8.90270114e-01
-4.18510437e-01 -8.18564355e-01 -5.16195476e-01 9.69452858e-01
-7.91879058e-01 1.31934613e-01 -3.72131884e-01 9.12746906e-01
1.02552839e-01 9.15808141e-01 5.87225974e-01 -3.98696005e-01
4.91658300e-01 2.42786631e-01 8.09037015e-02 -2.96278894e-01
-6.45728111e-01 1.01336367e-01 -5.01459315e-02 -1.04079473e+00
1.68188028e-02 -8.30404818e-01 -1.03011966e+00 6.92645013e-02
-2.63507813e-01 -1.13794208e-01 8.02945197e-01 7.50629425e-01
5.34784496e-01 1.05489528e+00 4.50469285e-01 -6.74104810e-01
-3.90005887e-01 -8.00145626e-01 -4.13070887e-01 2.40118966e-01
3.38460118e-01 -5.29888988e-01 -3.71318758e-01 -5.84900565e-02] | [7.629028797149658, 2.015986919403076] |
2c4086dd-109e-49e8-a29d-c62c46bd7b8e | prescriptive-business-process-monitoring-for | 2008.08693 | null | https://arxiv.org/abs/2008.08693v1 | https://arxiv.org/pdf/2008.08693v1.pdf | Prescriptive Business Process Monitoring for Recommending Next Best Actions | Predictive business process monitoring (PBPM) techniques predict future process behaviour based on historical event log data to improve operational business processes. Concerning the next activity prediction, recent PBPM techniques use state-of-the-art deep neural networks (DNNs) to learn predictive models for producing more accurate predictions in running process instances. Even though organisations measure process performance by key performance indicators (KPIs), the DNN`s learning procedure is not directly affected by them. Therefore, the resulting next most likely activity predictions can be less beneficial in practice. Prescriptive business process monitoring (PrBPM) approaches assess predictions regarding their impact on the process performance (typically measured by KPIs) to prevent undesired process activities by raising alarms or recommending actions. However, none of these approaches recommends actual process activities as actions that are optimised according to a given KPI. We present a PrBPM technique that transforms the next most likely activities into the next best actions regarding a given KPI. Thereby, our technique uses business process simulation to ensure the control-flow conformance of the recommended actions. Based on our evaluation with two real-life event logs, we show that our technique`s next best actions can outperform next activity predictions regarding the optimisation of a KPI and the distance from the actual process instances. | ['Martin Matzner', 'Sven Weinzierl', 'Sebastian Dunzer', 'Sandra Zilker'] | 2020-08-19 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 4.34917927e-01 2.98213750e-01 9.37177986e-03 -2.63124853e-01
-7.89895207e-02 -1.45028144e-01 9.89355981e-01 7.33019650e-01
-2.80357510e-01 5.95327139e-01 3.21881443e-01 -6.31511927e-01
-8.16265941e-01 -1.32461178e+00 -5.00684321e-01 -4.18779522e-01
-2.11482763e-01 8.06813598e-01 1.10919893e-01 2.23372340e-01
5.31475008e-01 9.02678370e-01 -1.33310962e+00 6.25334680e-01
4.72819030e-01 1.23339522e+00 -2.30946038e-02 8.68438959e-01
-6.36290491e-01 1.43099535e+00 -8.71104956e-01 -1.39847249e-01
4.28683758e-01 -3.54378730e-01 -6.63322687e-01 -2.81743377e-01
-3.37710917e-01 -3.91174108e-01 -4.92443740e-02 6.69276178e-01
-1.27998129e-01 2.35124961e-01 6.39570832e-01 -1.34972942e+00
-3.09327453e-01 7.99456418e-01 -1.91245601e-01 5.53743303e-01
1.98762119e-01 6.25427127e-01 8.50043535e-01 -3.07240218e-01
1.44044161e-01 1.30630350e+00 6.79413617e-01 1.61074683e-01
-1.07279527e+00 -4.10603732e-01 4.08513665e-01 4.06810641e-01
-8.31463695e-01 7.49457181e-02 5.93954086e-01 -2.68026292e-01
1.37150097e+00 1.52119607e-01 9.52337325e-01 1.04816186e+00
9.50558662e-01 6.39977753e-01 7.14286804e-01 -3.42305034e-01
6.68536425e-01 -3.47958535e-01 -1.84184045e-01 -1.53388858e-01
2.13622898e-01 2.56507039e-01 -3.90646726e-01 -1.05471998e-01
9.30859208e-01 7.52043545e-01 9.21766385e-02 1.65894598e-01
-1.24335229e+00 4.13033634e-01 1.59785420e-01 4.85359192e-01
-1.39968681e+00 4.26095098e-01 3.86325479e-01 4.37740684e-01
-9.70291197e-02 7.93912351e-01 -6.69336140e-01 -6.41379416e-01
-7.36159325e-01 2.09347710e-01 1.28919697e+00 6.20127320e-01
4.94551629e-01 2.42482424e-01 -7.55248964e-01 1.71358347e-01
4.72473383e-01 7.71303922e-02 3.44647646e-01 -1.11293256e+00
4.41804230e-01 8.87933791e-01 6.30492270e-01 -7.99061179e-01
-3.42054814e-01 -2.62139946e-01 -7.73240089e-01 1.11745216e-01
5.26588082e-01 -1.33947521e-01 -7.36437500e-01 9.46966529e-01
-2.72307426e-01 4.61023062e-01 1.15879670e-01 2.86778361e-01
-2.03445584e-01 1.18842220e+00 5.84896445e-01 -5.82787216e-01
1.01062751e+00 -6.77631259e-01 -8.77242804e-01 -2.95590878e-01
2.82018691e-01 -1.74751744e-01 5.88216901e-01 6.95846558e-01
-9.85675693e-01 -7.90572882e-01 -5.97110271e-01 1.03184557e+00
-1.22933172e-01 -4.39244747e-01 5.36113501e-01 3.11647236e-01
-6.86221302e-01 1.25113320e+00 -1.25936198e+00 -1.79375306e-01
3.71657908e-01 3.21795434e-01 8.80616829e-02 1.45844609e-01
-1.16954076e+00 1.00918245e+00 8.51716399e-01 3.85395795e-01
-1.40982783e+00 -8.83084297e-01 -3.04061413e-01 7.60512590e-01
6.55359209e-01 -3.46233040e-01 1.75022161e+00 -9.20561612e-01
-1.51070797e+00 -2.22634315e-01 1.96982622e-01 -1.23146915e+00
8.11354756e-01 -4.70539838e-01 -6.94654644e-01 -1.17586963e-02
-4.77571219e-01 2.03104749e-01 6.66367590e-01 -9.55041409e-01
-1.38563824e+00 -5.61298691e-02 -1.20132819e-01 1.87844679e-01
-2.59752154e-01 -1.64607957e-01 2.06680089e-01 8.83234739e-02
-8.59485865e-02 -4.38495576e-01 -6.95447505e-01 -6.43542290e-01
-1.43245578e-01 -3.53030711e-01 5.34665763e-01 -6.28303528e-01
1.40257263e+00 -1.76398981e+00 -4.83113945e-01 5.25168538e-01
1.02486260e-01 2.54183054e-01 2.62677103e-01 8.54043245e-01
-1.31937891e-01 1.87241986e-01 7.37110749e-02 1.92736402e-01
1.46364555e-01 4.14197803e-01 -3.46872151e-01 2.33586431e-02
3.36310714e-01 6.85549319e-01 -8.29288483e-01 8.44506919e-02
6.95107698e-01 7.89410546e-02 1.40145114e-02 6.64223731e-01
-4.38022196e-01 4.85878676e-01 -2.55859137e-01 5.66850483e-01
1.16441913e-01 6.17131814e-02 2.39790171e-01 2.01686770e-01
-1.49196535e-01 2.74083614e-01 -1.25084591e+00 5.04980445e-01
-6.68633044e-01 2.62260914e-01 -3.03193569e-01 -1.08033264e+00
1.39534926e+00 6.32474184e-01 9.79015708e-01 -7.69381702e-01
-2.81255275e-01 -1.00717410e-01 2.57271320e-01 -2.30578244e-01
1.37040332e-01 -2.28290647e-01 2.59263992e-01 5.14028490e-01
-2.33464718e-01 3.88013124e-01 4.15124148e-01 -5.24219155e-01
1.76919043e+00 2.07538575e-01 7.48458445e-01 6.08726144e-02
7.17944384e-01 -2.38113254e-01 9.10426974e-01 1.05080366e+00
-4.37597334e-01 -1.07595041e-01 9.61583674e-01 -1.04120207e+00
-1.09961748e+00 -1.11143744e+00 4.60873574e-01 1.00920129e+00
-2.90997297e-01 -1.18875057e-01 -2.39317179e-01 -6.86781108e-01
3.71675603e-02 1.43132246e+00 -5.51812053e-01 -2.99776137e-01
-5.67103088e-01 -7.15729892e-01 2.50719219e-01 8.90863955e-01
3.82774234e-01 -1.70127881e+00 -9.26280499e-01 1.10106897e+00
6.05730176e-01 -7.03872859e-01 3.10520053e-01 4.53499287e-01
-1.01215708e+00 -1.17156410e+00 -2.14016169e-01 -3.48609984e-02
3.04218680e-01 -5.22534430e-01 1.17745352e+00 -3.87673825e-01
4.36471790e-01 2.60515511e-01 -5.72157539e-02 -9.68849540e-01
-1.11786067e+00 -2.48411492e-01 -1.38807878e-01 1.20560139e-01
9.56145167e-01 -6.29112482e-01 -5.17639697e-01 1.70701489e-01
-6.50417209e-01 -1.95803493e-01 8.11761081e-01 4.26337689e-01
6.51457787e-01 6.22107327e-01 9.05406654e-01 -6.77026868e-01
1.14319098e+00 -3.35700393e-01 -5.31603038e-01 3.80901456e-01
-1.16747010e+00 -1.83522552e-02 9.33012545e-01 -4.35999542e-01
-1.50621223e+00 -2.50358224e-01 5.30419983e-02 -4.88835722e-01
-7.72254229e-01 5.55373430e-01 -1.64401889e-01 9.46714342e-01
5.57244122e-01 3.65524709e-01 -3.20781231e-01 -2.04593256e-01
-1.52750209e-01 3.13910723e-01 3.46951216e-01 -4.56969053e-01
2.74783611e-01 3.50455523e-01 1.59472749e-01 -1.43289104e-01
-5.66636860e-01 -4.97026801e-01 -4.45557296e-01 -6.54086053e-01
6.40894830e-01 -4.42774534e-01 -1.35624897e+00 1.95214733e-01
-1.02228904e+00 -2.49897525e-01 -6.55864358e-01 4.65418518e-01
-7.92554677e-01 -5.74286804e-02 -9.24109519e-01 -1.26076782e+00
-5.95585585e-01 -7.64398873e-01 4.83270168e-01 3.40890944e-01
-7.80282974e-01 -9.97275591e-01 9.62397084e-02 -2.11311057e-02
5.37098527e-01 2.60931700e-01 8.55971038e-01 -1.48586333e+00
-6.02179110e-01 -6.18970573e-01 9.17186067e-02 5.98637521e-01
3.71610343e-01 2.71587849e-01 -5.07538557e-01 -3.07348240e-02
2.46227205e-01 8.14395308e-01 1.75558090e-01 4.43847954e-01
1.27706766e+00 -6.18840575e-01 -3.04812163e-01 1.92019984e-01
1.39223361e+00 1.12555599e+00 9.21988487e-01 9.80915070e-01
5.58802426e-01 7.02369571e-01 8.42672229e-01 8.48456383e-01
-9.08756480e-02 -6.90364391e-02 7.39531100e-01 5.18597782e-01
6.49936676e-01 -3.15055639e-01 6.49233639e-01 3.34292144e-01
-5.02406120e-01 -1.88145876e-01 -1.48132443e+00 5.20973802e-01
-2.03775072e+00 -1.35965371e+00 -2.85857707e-01 2.24706864e+00
3.52905542e-01 8.09595942e-01 -6.90976828e-02 2.75588810e-01
6.55573726e-01 -1.64239138e-01 -6.23906195e-01 -9.73375499e-01
3.57529879e-01 7.47046173e-02 6.59344673e-01 -1.87249109e-02
-8.40453207e-01 3.91020179e-01 5.40956879e+00 2.52510548e-01
-5.99151313e-01 -1.14650130e-01 8.20755124e-01 -2.60627031e-01
1.14061637e-02 -4.21399437e-02 -1.03893864e+00 5.17302752e-01
1.88124382e+00 -4.39852059e-01 2.71625936e-01 1.15675867e+00
9.06774104e-01 6.98851347e-02 -1.62523007e+00 6.11891925e-01
-7.84373105e-01 -1.51905990e+00 1.01769410e-01 2.00752139e-01
7.34788895e-01 -2.75649607e-01 -4.96637911e-01 6.96188211e-01
6.73057258e-01 -1.12862563e+00 4.87176478e-01 1.17003143e+00
-3.14960331e-01 -1.37805378e+00 1.15934181e+00 3.68590385e-01
-1.10764539e+00 -9.71266985e-01 -1.70537367e-01 -3.50804269e-01
3.23379576e-01 7.26315737e-01 -1.51348329e+00 3.86034876e-01
7.30961502e-01 5.40446818e-01 -2.29172856e-01 8.57475400e-01
-4.15979981e-01 1.00430000e+00 -1.72885641e-01 -6.61172345e-02
5.22505760e-01 -2.51932144e-01 2.77865380e-01 1.18892133e+00
4.24276382e-01 -5.55140495e-01 1.58127084e-01 9.42837775e-01
3.42267573e-01 -1.47489342e-03 -3.56266260e-01 -2.82157987e-01
5.07403553e-01 8.43410611e-01 -6.85434163e-01 -5.39665461e-01
-2.97634900e-01 5.30531228e-01 -1.48359910e-01 3.82776320e-01
-4.32569146e-01 4.90703136e-02 7.97588348e-01 3.71805757e-01
2.58963853e-01 8.88818130e-02 -5.37892103e-01 -1.50395155e-01
-1.26572907e-01 -8.26253355e-01 4.97974515e-01 -5.43198287e-01
-1.54298866e+00 1.28850043e-01 -1.50061548e-01 -1.15366316e+00
-6.91565275e-01 -5.78579664e-01 -9.95841861e-01 1.02320611e+00
-1.20276904e+00 -6.89501822e-01 -2.28079855e-01 1.32863387e-01
4.34065729e-01 -1.02515608e-01 6.17541671e-01 -3.28031182e-01
-4.36436176e-01 -4.10547912e-01 3.16549540e-01 -8.57732967e-02
4.10734177e-01 -1.60153234e+00 4.85223532e-01 8.45184803e-01
-2.84844846e-01 4.06242788e-01 8.11650038e-01 -1.02244258e+00
-1.01789975e+00 -1.21202755e+00 8.56004298e-01 -3.71827334e-01
1.07503247e+00 2.93898314e-01 -1.39389408e+00 9.59139943e-01
1.31460264e-01 -4.91811693e-01 6.81287467e-01 1.61893070e-02
5.07002056e-01 -6.04148388e-01 -1.09511375e+00 5.70253611e-01
6.88055575e-01 -5.07591292e-02 -8.91908944e-01 -4.85450514e-02
3.72834206e-01 1.40836865e-01 -1.42923641e+00 5.17528296e-01
1.91158012e-01 -1.02254212e+00 5.95461309e-01 -7.86223412e-01
4.83389527e-01 -2.19101042e-01 1.57318935e-01 -1.50049019e+00
-5.06272376e-01 -5.47664940e-01 -9.51356351e-01 1.09251595e+00
9.87701118e-02 -4.84322369e-01 8.90726447e-01 8.87233675e-01
-7.36193135e-02 -7.06881344e-01 -9.03576314e-01 -6.75822616e-01
-4.18821782e-01 -8.18511844e-01 1.20844829e+00 8.56967092e-01
-2.65734911e-01 7.31160678e-03 -2.25373790e-01 2.32593745e-01
5.34199417e-01 -8.60927925e-02 7.17080951e-01 -1.44841170e+00
-3.39399636e-01 -6.62994623e-01 -4.98676419e-01 -5.53003132e-01
-2.37294972e-01 -3.08437824e-01 -3.16603743e-02 -2.09818721e+00
-2.20086694e-01 8.37677345e-02 -9.87659097e-01 5.31535864e-01
-4.32605669e-02 -9.88913774e-01 3.69893491e-01 1.38549969e-01
-3.78084123e-01 4.44478869e-01 1.16345179e+00 -4.49464144e-03
-7.04833031e-01 5.58335423e-01 -4.54611957e-01 7.33823478e-01
9.94089901e-01 -3.77338141e-01 -3.56994629e-01 3.45007062e-01
4.66436028e-01 4.14470494e-01 2.45850533e-01 -1.34852040e+00
3.96577954e-01 -7.50817060e-01 6.74125254e-01 -7.19470024e-01
-1.18930548e-01 -1.18582225e+00 7.40253747e-01 9.19836223e-01
-6.14208758e-01 1.55951872e-01 4.54741381e-02 9.13981676e-01
-3.23369503e-01 -2.18286976e-01 3.41495216e-01 -2.89861917e-01
-1.01135719e+00 2.82802373e-01 -1.04385626e+00 -7.22031593e-01
1.28119838e+00 -5.55570185e-01 5.54097779e-02 -3.96466643e-01
-9.83329296e-01 2.70800531e-01 -2.13353306e-01 3.23957056e-01
3.91841590e-01 -9.39205408e-01 -7.29015350e-01 -1.00041956e-01
-3.19432206e-02 2.50620663e-01 1.61906645e-01 6.20425999e-01
-5.77659547e-01 5.21080494e-01 -3.77951384e-01 -4.54256296e-01
-6.52375340e-01 7.01630414e-01 7.54847527e-01 -9.99923348e-01
-7.01852798e-01 2.02791512e-01 -1.57626316e-01 -2.60908514e-01
1.19791828e-01 -5.56698799e-01 -3.67440104e-01 6.27385601e-02
8.84561241e-01 7.55064487e-01 2.14401603e-01 3.00172836e-01
1.73302218e-02 -6.94919705e-01 -1.38969757e-02 1.10467769e-01
1.69003522e+00 1.94308236e-01 -2.32472643e-02 6.45034432e-01
2.26593524e-01 -5.96224606e-01 -1.99669254e+00 -2.08099306e-01
9.08119738e-01 -4.11365271e-01 -8.27151686e-02 -1.09591186e+00
-8.09047878e-01 6.56362534e-01 4.59019810e-01 6.01263165e-01
1.14258695e+00 -2.55917579e-01 4.18337852e-01 6.53373599e-01
1.28522009e-01 -1.49614477e+00 -3.50816473e-02 5.76326549e-01
7.52716720e-01 -6.90347850e-01 -3.32112044e-01 1.14933386e-01
-7.31462240e-01 1.38925540e+00 9.17419314e-01 4.36033532e-02
6.24477863e-01 3.20628256e-01 -1.59373999e-01 -2.09978968e-01
-1.06191325e+00 4.80736911e-01 -2.50951707e-01 7.70879030e-01
1.68237507e-01 4.19267774e-01 1.18526936e-01 6.68326497e-01
-5.99777326e-03 1.25461638e-01 6.28588080e-01 1.00463700e+00
-4.52589214e-01 -6.10796034e-01 -6.84293449e-01 1.07058942e+00
-5.46184421e-01 1.69933587e-01 -8.62952471e-02 6.75774097e-01
-1.14331439e-01 7.82709777e-01 6.40331030e-01 -5.81714064e-02
9.31504250e-01 5.38831532e-01 -4.56677079e-02 -6.17404938e-01
-8.94428611e-01 -1.41462117e-01 2.12912589e-01 -5.81073165e-01
-9.78289843e-02 -8.25300872e-01 -1.33622932e+00 -6.42668009e-01
3.64456236e-01 -2.21808806e-01 4.64675665e-01 1.30424333e+00
1.24662772e-01 1.31778181e+00 4.70555097e-01 -5.04053533e-01
-8.72426331e-01 -1.29633820e+00 -5.09168208e-01 5.48007071e-01
-3.18005115e-01 -2.28896812e-01 -4.85005677e-02 -2.63704341e-02] | [8.606925964355469, 5.955133438110352] |
9ffe7572-9ee7-4ac9-a736-8be13a69947e | on-the-relation-between-prediction-and | 2112.05248 | null | https://arxiv.org/abs/2112.05248v1 | https://arxiv.org/pdf/2112.05248v1.pdf | On the Relation between Prediction and Imputation Accuracy under Missing Covariates | Missing covariates in regression or classification problems can prohibit the direct use of advanced tools for further analysis. Recent research has realized an increasing trend towards the usage of modern Machine Learning algorithms for imputation. It originates from their capability of showing favourable prediction accuracy in different learning problems. In this work, we analyze through simulation the interaction between imputation accuracy and prediction accuracy in regression learning problems with missing covariates when Machine Learning based methods for both, imputation and prediction are used. In addition, we explore imputation performance when using statistical inference procedures in prediction settings, such as coverage rates of (valid) prediction intervals. Our analysis is based on empirical datasets provided by the UCI Machine Learning repository and an extensive simulation study. | ['Markus Pauly', 'Justus Tulowietzki', 'Burim Ramosaj'] | 2021-12-09 | null | null | null | null | ['prediction-intervals'] | ['miscellaneous'] | [ 3.89391154e-01 -2.08617952e-02 -7.42949784e-01 -7.35160887e-01
-9.45439339e-01 -1.12478189e-01 4.22866106e-01 4.06926513e-01
-5.41627586e-01 1.34061885e+00 3.58367443e-01 -5.88671744e-01
-5.34436822e-01 -7.02632368e-01 -6.35839164e-01 -5.03828108e-01
-1.69761494e-01 4.72523332e-01 -5.63237727e-01 3.04853886e-01
3.04186344e-01 2.26117164e-01 -1.40242660e+00 2.05490500e-01
1.22462356e+00 5.44767439e-01 -3.19141328e-01 2.78825641e-01
1.14434913e-01 7.51618266e-01 -3.48880917e-01 -6.28980756e-01
1.80207655e-01 -5.00965893e-01 -2.97329187e-01 -4.17865276e-01
-8.82432610e-03 -8.11183229e-02 1.50640056e-01 3.75120878e-01
5.09561062e-01 2.46836040e-02 1.00860310e+00 -1.30399621e+00
-2.30277196e-01 7.20142066e-01 -6.68491066e-01 -7.58917853e-02
4.92758721e-01 -1.08264774e-01 5.97354591e-01 -4.01914924e-01
4.61832196e-01 1.03816736e+00 1.08816159e+00 1.60059705e-01
-1.61488700e+00 -8.29659462e-01 -4.13846672e-01 1.05556376e-01
-1.22311664e+00 -5.34120023e-01 4.00803387e-01 -6.76254570e-01
5.66046357e-01 3.74047577e-01 1.32063478e-01 1.04761446e+00
4.35674936e-01 3.66489381e-01 1.43133307e+00 -5.92943549e-01
3.85611176e-01 2.53907353e-01 3.03557783e-01 -5.65354805e-03
6.81277156e-01 5.83617926e-01 -2.77641088e-01 -5.99232674e-01
6.83503509e-01 1.63716923e-05 -6.60196617e-02 -1.85323417e-01
-9.70572770e-01 1.07415378e+00 -5.67037351e-02 -1.66835278e-01
-5.51173627e-01 -3.81301297e-03 3.47832769e-01 5.19330084e-01
6.56387329e-01 8.17264542e-02 -6.68776929e-01 -2.05904365e-01
-1.07044327e+00 2.99007177e-01 8.58235240e-01 4.50404406e-01
3.91944766e-01 -1.87057823e-01 -1.56374916e-01 1.06469429e+00
2.17326626e-01 2.87114739e-01 3.97025526e-01 -9.03392076e-01
5.53015709e-01 7.33763874e-01 4.98183012e-01 -7.76421666e-01
-5.90120494e-01 -3.68955463e-01 -1.23208439e+00 2.53967047e-01
7.79356837e-01 -4.51239377e-01 -4.54528242e-01 1.51306665e+00
1.69967771e-01 2.53658384e-01 8.90729576e-02 4.63561356e-01
2.40502328e-01 1.99890539e-01 3.47033054e-01 -6.27466559e-01
1.00274980e+00 -1.95763603e-01 -5.89299440e-01 1.68494508e-01
9.80194509e-01 -5.50836742e-01 6.20685279e-01 6.79009616e-01
-9.49459195e-01 -4.16708440e-01 -4.84038115e-01 5.75877726e-02
-1.57977998e-01 1.64659470e-01 7.51265287e-01 8.16255212e-01
-3.74138951e-01 5.64876139e-01 -7.26005077e-01 -1.21952273e-01
5.78114986e-01 5.28949916e-01 -3.17923456e-01 -2.62660027e-01
-8.99144709e-01 9.28320944e-01 9.02514979e-02 1.75824121e-01
4.67806961e-03 -1.09434450e+00 -7.84712195e-01 1.11296792e-02
2.01542210e-03 -7.40715563e-01 8.22756350e-01 -1.11781013e+00
-8.66602123e-01 6.44640267e-01 -3.92698765e-01 -6.11564994e-01
8.72105181e-01 -7.81580508e-02 -1.84848160e-01 -7.37741530e-01
-7.17377365e-02 9.38452780e-02 4.77913350e-01 -9.12184119e-01
-6.24547243e-01 -6.88931167e-01 -4.46028471e-01 -2.16456980e-01
2.58088946e-01 8.67097378e-02 2.64320731e-01 -8.11585426e-01
-4.27289195e-02 -6.64790332e-01 -5.63625276e-01 -7.12550640e-01
-2.81004488e-01 -2.20428303e-01 9.32203606e-02 -8.35133255e-01
1.34044027e+00 -1.93504441e+00 -2.47917641e-02 4.78224397e-01
-3.43122452e-01 -2.42963389e-01 8.16775784e-02 5.20090997e-01
-1.93758994e-01 -1.19011314e-03 -3.74570191e-01 -2.92915434e-01
-1.59605771e-01 1.53883740e-01 -2.74066150e-01 6.56289101e-01
-1.36203378e-01 6.79890633e-01 -2.16561079e-01 -4.86135334e-01
4.21143770e-01 5.72445810e-01 -4.44847226e-01 4.51406948e-02
2.68205404e-01 7.54837215e-01 -1.83453768e-01 5.75531960e-01
6.99452519e-01 -6.94199800e-02 3.96517187e-01 2.31088698e-01
-1.82638094e-01 1.59843519e-01 -1.11653745e+00 1.13915920e+00
-4.57865804e-01 3.11719567e-01 -2.29556710e-01 -1.23086727e+00
1.06672347e+00 1.41391426e-01 3.00779462e-01 -6.54919446e-01
-2.20414391e-03 6.00866787e-02 9.18796360e-02 -4.21713293e-01
-9.56706703e-02 -2.39293814e-01 -1.00888930e-01 2.79147804e-01
-4.03618425e-01 3.87857467e-01 7.99446031e-02 -3.85596037e-01
9.55608070e-01 1.92148209e-01 5.24324059e-01 -2.12844089e-01
4.69995946e-01 6.87665567e-02 6.80634022e-01 1.01906610e+00
3.98845375e-01 5.09517491e-01 6.68188572e-01 -4.39268798e-01
-9.51301575e-01 -8.59213173e-01 -9.37096715e-01 8.03510845e-01
-5.68036079e-01 2.29592668e-03 -3.08198452e-01 -3.91302854e-01
5.08095086e-01 8.69913280e-01 -7.91305602e-01 1.55762017e-01
-3.48205835e-01 -1.38054895e+00 3.36508274e-01 6.18928730e-01
-2.86419056e-02 -8.90145302e-01 -5.72943449e-01 4.23169106e-01
-2.95230448e-02 -6.71260238e-01 3.87664527e-01 2.56200194e-01
-1.40150142e+00 -1.36367226e+00 -5.61356664e-01 -1.90662578e-01
5.58152437e-01 -3.19661885e-01 1.13415742e+00 2.82915890e-01
-1.81438364e-02 9.40061212e-02 -3.41907620e-01 -6.22830510e-01
-3.60607624e-01 2.29106739e-01 -1.70622915e-01 -1.22397862e-01
4.77652848e-01 -6.51168525e-01 -3.69706064e-01 3.00969988e-01
-6.78916395e-01 7.62373134e-02 4.32438850e-01 9.39408243e-01
3.54477078e-01 1.49721010e-02 1.12169731e+00 -1.14999592e+00
5.43881595e-01 -9.10441577e-01 -9.19415593e-01 2.33347207e-01
-8.56436074e-01 3.12449653e-02 3.18170339e-01 -2.22712249e-01
-1.09760034e+00 9.08759236e-02 -1.95316896e-01 2.05487981e-01
-4.92076874e-01 9.51369822e-01 -9.16692521e-03 2.84338593e-01
4.17487204e-01 -1.43048659e-01 1.41935632e-01 -6.49431348e-01
-3.51777256e-01 7.89148271e-01 2.38310590e-01 -4.13934410e-01
4.96314138e-01 2.09007695e-01 3.41920704e-01 -6.63037360e-01
-6.49334371e-01 -1.32256478e-01 -6.65921152e-01 2.31203027e-02
4.12505180e-01 -9.25442636e-01 -8.88981640e-01 3.89444619e-01
-5.99075973e-01 -4.96747315e-01 -2.57224943e-02 9.71293330e-01
-5.60445428e-01 1.97402269e-01 -1.00415878e-01 -1.28296959e+00
-5.36305904e-02 -8.24906707e-01 6.29094243e-01 1.60496637e-01
-6.39142990e-01 -1.23970449e+00 3.46204430e-01 5.45598805e-01
2.29900524e-01 7.37763584e-01 1.07443321e+00 -7.13122785e-01
-1.45031050e-01 -4.56334472e-01 -6.81525916e-02 -3.72453704e-02
-3.22953090e-02 -1.87335107e-02 -8.06520164e-01 -1.79597616e-01
-2.68895984e-01 -4.00689729e-02 8.16229165e-01 9.90420103e-01
1.42980909e+00 -3.24816585e-01 -3.32538098e-01 4.35833246e-01
1.61214757e+00 1.29721299e-01 8.73277247e-01 4.53593403e-01
1.31234780e-01 9.33273971e-01 6.70538664e-01 7.01724887e-01
3.62305254e-01 7.94992924e-01 4.95360717e-02 -2.40079574e-02
5.25960386e-01 -2.11407587e-01 -1.39826879e-01 -1.50490403e-01
-2.57513911e-01 -5.26236407e-02 -1.13832331e+00 2.75275797e-01
-2.02947378e+00 -1.02979147e+00 -5.90940058e-01 2.94064069e+00
8.44609380e-01 7.34813884e-02 3.03283185e-01 3.67269129e-01
5.52179396e-01 -5.26820719e-01 -3.62027198e-01 -6.71536565e-01
-4.87860069e-02 2.48723820e-01 5.78291893e-01 3.35986286e-01
-7.80495942e-01 2.12883092e-02 7.00192595e+00 5.14162719e-01
-5.34750938e-01 -5.49142957e-02 1.06875324e+00 -1.44923031e-01
-2.62650922e-02 -1.53203323e-01 -5.75344861e-01 7.08609879e-01
1.20147359e+00 4.73262817e-02 2.31905743e-01 4.49692219e-01
7.68618703e-01 -6.80267215e-01 -1.15609419e+00 7.10439384e-01
-3.27245474e-01 -1.24455059e+00 -4.58674937e-01 2.07813442e-01
7.24008739e-01 -2.00189218e-01 2.99520046e-02 4.64855403e-01
3.35795105e-01 -1.46041393e+00 2.27253092e-03 7.69243956e-01
7.29453862e-01 -9.56477582e-01 1.25485349e+00 6.40673280e-01
-2.38688767e-01 -3.32270473e-01 -2.51058966e-01 -6.48217201e-01
1.76089387e-02 9.19703901e-01 -6.19301379e-01 5.90774119e-01
6.09670460e-01 5.04596472e-01 -4.12420869e-01 1.40997708e+00
1.36570856e-01 1.09619665e+00 -2.93841898e-01 3.29180360e-01
-1.91727698e-01 -5.58519185e-01 -1.16174057e-01 1.00369942e+00
3.83981943e-01 1.57393426e-01 -2.68317819e-01 5.42902172e-01
1.76107183e-01 1.92655370e-01 -6.54520690e-01 4.95393813e-01
4.12293345e-01 6.57814264e-01 -3.21504921e-01 -4.22050571e-03
-6.26807392e-01 2.25984976e-01 2.01211169e-01 3.34108084e-01
-7.19771504e-01 1.48714921e-02 5.23346066e-01 2.32980013e-01
3.87368351e-02 -1.61999352e-02 -1.11984265e+00 -8.79844964e-01
-9.58578363e-02 -6.46619499e-01 6.45626664e-01 -5.94027042e-01
-1.44080031e+00 -5.43912314e-02 2.67620027e-01 -8.21497321e-01
-5.02118051e-01 -2.45071843e-01 -7.43819296e-01 1.17358947e+00
-1.31875384e+00 -8.57039511e-01 3.50322895e-04 3.69888306e-01
5.27816042e-02 -5.88884205e-02 7.88676679e-01 1.65549830e-01
-6.14539146e-01 7.28917360e-01 7.46801555e-01 8.80981684e-02
7.25923717e-01 -1.04447114e+00 -1.23194396e-01 2.98780799e-01
-2.30415285e-01 3.33061337e-01 9.18345213e-01 -7.85364032e-01
-1.16352737e+00 -8.99934530e-01 1.24953544e+00 -7.01594532e-01
2.41071701e-01 -1.29052609e-01 -9.01902080e-01 7.83403873e-01
-2.07993284e-01 -4.69966829e-01 1.06087124e+00 7.68035233e-01
-1.01093560e-01 -1.06672369e-01 -1.46307123e+00 1.50520802e-01
5.24951220e-01 8.84991065e-02 -4.46898580e-01 2.57754512e-02
4.64444607e-03 -1.96332186e-01 -1.23176622e+00 8.11169326e-01
9.67660308e-01 -9.87860978e-01 9.74348247e-01 -7.14094877e-01
7.75143325e-01 1.85152024e-01 -7.37420321e-02 -1.03392446e+00
-1.26491860e-01 -1.05895601e-01 6.47796914e-02 1.37907338e+00
5.94307184e-01 -7.18937874e-01 9.19282317e-01 1.04012537e+00
5.33057153e-01 -8.44274223e-01 -1.33821130e+00 -5.93311906e-01
4.37237382e-01 -6.63915277e-01 7.04350591e-01 9.54918146e-01
-9.33271870e-02 8.23660344e-02 -5.81904531e-01 1.40984822e-03
8.81879985e-01 2.70311147e-01 9.83702421e-01 -1.50857258e+00
-1.70069888e-01 -8.72357786e-02 -3.61464590e-01 -1.35610804e-01
4.17536378e-01 -6.21276259e-01 -5.47008932e-01 -1.37490368e+00
4.41891164e-01 -8.86651635e-01 -2.11553976e-01 6.40749991e-01
-1.05775997e-01 1.93731770e-01 -7.25790858e-02 3.82410176e-02
-1.71228573e-02 2.04249606e-01 6.98602498e-01 2.92042404e-01
-5.43323815e-01 6.61803484e-01 -6.27585113e-01 5.98301291e-01
1.18855727e+00 -6.97070658e-01 -2.20671669e-01 -5.99371493e-02
3.57003242e-01 7.17615962e-01 5.75168908e-01 -8.03336203e-01
-2.09440365e-01 -3.98653746e-01 9.47729290e-01 -6.39940202e-01
4.68035154e-02 -1.12663221e+00 7.83282042e-01 5.15381992e-01
-5.57154357e-01 2.40870025e-02 1.27657101e-01 5.12700438e-01
7.67665878e-02 -2.39551470e-01 4.68994915e-01 2.60213166e-01
-1.17712267e-01 -2.11371392e-01 -4.08089876e-01 -2.28159353e-01
8.86560321e-01 -4.47693050e-01 -1.28023118e-01 -3.14098984e-01
-9.90628242e-01 3.62424374e-01 3.96155119e-01 1.44827008e-01
3.64738911e-01 -1.09700167e+00 -1.20558226e+00 2.57326901e-01
4.97442111e-02 -4.82767493e-01 7.90822729e-02 1.25435090e+00
-4.60595737e-04 4.41783011e-01 -1.82134569e-01 -2.03759804e-01
-1.28587925e+00 4.71943021e-01 1.40548557e-01 -3.00684124e-01
-5.05701423e-01 5.95517233e-02 -2.92745799e-01 -3.87750030e-01
2.26587772e-01 -4.52959724e-02 -1.43599182e-01 4.92646173e-02
4.10413623e-01 9.31184590e-01 8.34211931e-02 -2.37116516e-01
-2.20884100e-01 8.68894383e-02 2.52019733e-01 4.63504493e-02
1.45298576e+00 -2.07813442e-01 -1.16178229e-01 5.71905494e-01
8.18389297e-01 -1.05708860e-01 -8.77558112e-01 1.73987851e-01
2.94174254e-01 -6.45837486e-01 -7.44733512e-02 -1.21573174e+00
-6.18573546e-01 5.04491746e-01 6.15405858e-01 1.33107036e-01
1.12637222e+00 -3.79336894e-01 8.80033299e-02 7.15993866e-02
3.90154630e-01 -7.86867976e-01 -1.00846767e+00 1.50343895e-01
6.41305864e-01 -1.52185106e+00 1.95771053e-01 -1.44224733e-01
-4.43163753e-01 8.54523480e-01 2.31859103e-01 -1.68730214e-01
7.50556648e-01 2.95835614e-01 -1.05232865e-01 3.19640189e-01
-9.64583695e-01 2.04734176e-01 1.84897646e-01 7.94674218e-01
7.99818993e-01 2.65307277e-01 -9.05464828e-01 9.51279104e-01
-2.00063229e-01 5.84590435e-01 6.68750346e-01 4.90231752e-01
1.89665314e-02 -1.38358366e+00 -6.56456947e-01 1.20588875e+00
-7.60399520e-01 7.57839903e-02 -1.07507147e-01 9.23310935e-01
5.40741086e-02 1.21115339e+00 2.20541716e-01 1.92497402e-01
1.60565481e-01 1.95273086e-01 3.08046997e-01 -1.61658332e-01
-6.85492754e-01 -2.87639380e-01 2.32269183e-01 -2.75780648e-01
-5.10132372e-01 -1.20434356e+00 -8.42106283e-01 -5.61995149e-01
-3.46807212e-01 3.56670171e-01 5.90059340e-01 9.55464959e-01
2.15151384e-01 1.96202204e-01 5.77668786e-01 -3.77685815e-01
-6.73730075e-01 -9.21402395e-01 -6.49041831e-01 1.58275664e-01
2.12705299e-01 -6.52815282e-01 -3.48465681e-01 -2.13331908e-01] | [7.826644420623779, 4.975827217102051] |
4a111909-c8d6-4d2d-984a-846ae4f47f6d | text-detection-recognition-in-the-wild-for | 2205.08565 | null | https://arxiv.org/abs/2205.08565v2 | https://arxiv.org/pdf/2205.08565v2.pdf | Text Detection & Recognition in the Wild for Robot Localization | Signage is everywhere and a robot should be able to take advantage of signs to help it localize (including Visual Place Recognition (VPR)) and map. Robust text detection & recognition in the wild is challenging due to such factors as pose, irregular text, illumination, and occlusion. We propose an end-to-end scene text spotting model that simultaneously outputs the text string and bounding boxes. This model is more suitable for VPR. Our central contribution is introducing utilizing an end-to-end scene text spotting framework to adequately capture the irregular and occluded text regions in different challenging places. To evaluate our proposed architecture's performance for VPR, we conducted several experiments on the challenging Self-Collected Text Place (SCTP) benchmark dataset. The initial experimental results show that the proposed method outperforms the SOTA methods in terms of precision and recall when tested on this benchmark. | ['John Zelek', 'Zobeir Raisi'] | 2022-05-17 | null | null | null | null | ['text-spotting', 'visual-place-recognition'] | ['computer-vision', 'computer-vision'] | [ 4.05335546e-01 -6.11307442e-01 1.54459462e-01 -5.46324909e-01
-5.93556821e-01 -5.93720555e-01 8.47627342e-01 -7.36465156e-02
-5.26317835e-01 2.13225469e-01 1.92680016e-01 -1.24975659e-01
3.08335423e-01 -1.83571890e-01 -7.08950520e-01 -2.83827752e-01
2.25999266e-01 6.97469771e-01 6.72913015e-01 -9.38515067e-02
7.24455595e-01 4.72966373e-01 -1.66491592e+00 1.62292495e-01
9.24950600e-01 7.60336518e-01 5.97294450e-01 7.51520514e-01
-7.16056526e-02 5.87445080e-01 -5.98567069e-01 -1.66431386e-02
3.56523126e-01 4.93736602e-02 -2.06966460e-01 3.56416613e-01
1.04991174e+00 -5.06953776e-01 -3.10780615e-01 8.23163569e-01
6.60183430e-01 2.71405131e-01 6.53301775e-01 -1.14977252e+00
-3.24965328e-01 -1.72466226e-02 -7.74014413e-01 -2.88821608e-01
6.80928051e-01 2.59777635e-01 7.19719648e-01 -1.29260480e+00
8.42385411e-01 1.19457376e+00 7.25113690e-01 2.53994465e-01
-8.94923270e-01 -3.53808641e-01 2.40590870e-01 -8.70026201e-02
-1.41365898e+00 -5.02317011e-01 7.67229140e-01 -3.69276762e-01
8.53637695e-01 2.72264928e-01 2.09981740e-01 1.11833835e+00
7.76745602e-02 1.36823881e+00 1.08104289e+00 -7.11712122e-01
3.17557514e-01 1.53067470e-01 -1.82450972e-02 7.39499331e-01
2.94512570e-01 -2.51222998e-01 -9.18656945e-01 8.67504254e-02
5.73459029e-01 -3.84266339e-02 -1.47938237e-01 -5.90836406e-01
-1.58898997e+00 2.84118384e-01 4.30519730e-01 1.75001353e-01
-1.62359074e-01 3.34951252e-01 3.10426563e-01 -1.09531991e-01
1.03763662e-01 3.14344019e-02 -1.89986885e-01 -2.69419402e-01
-1.22460973e+00 5.31975105e-02 5.53984463e-01 1.10054600e+00
3.81777555e-01 -5.80314770e-02 -2.99884647e-01 1.10467494e+00
6.50560021e-01 1.25092351e+00 4.33841944e-01 -1.40498161e-01
9.56463873e-01 6.75988615e-01 3.33542585e-01 -9.94427979e-01
-3.04924071e-01 1.64390847e-01 -4.50293720e-01 4.06376839e-01
4.53832626e-01 -4.48712036e-02 -1.63913763e+00 8.82990003e-01
2.75859356e-01 -1.26450270e-01 -2.18960553e-01 1.16882527e+00
7.62298226e-01 6.01622581e-01 -1.08093493e-01 6.26395702e-01
1.21097946e+00 -1.02245426e+00 -7.28029668e-01 -9.03107345e-01
6.62785769e-01 -1.02326584e+00 1.25607145e+00 3.02149594e-01
-4.27769333e-01 -4.48370576e-02 -1.14673972e+00 -6.04341984e-01
-6.53327584e-01 9.67923284e-01 7.87716880e-02 4.79159236e-01
-1.06059873e+00 -5.66534661e-02 -7.11228848e-01 -1.06598318e+00
3.38691235e-01 1.95476472e-01 -4.13087130e-01 -2.34478801e-01
-4.81926352e-01 8.75551522e-01 2.99938768e-01 4.00607526e-01
-6.18331075e-01 5.93909360e-02 -7.69354522e-01 -1.96767420e-01
2.08141714e-01 -3.59850794e-01 8.98636997e-01 -5.65934122e-01
-1.56610036e+00 9.73422766e-01 -3.82924587e-01 -2.82260299e-01
1.15806651e+00 -4.28614855e-01 -1.30074441e-01 1.21668108e-01
1.41374797e-01 8.26184690e-01 9.42201734e-01 -1.23292494e+00
-7.17237771e-01 -5.11723876e-01 -6.43639445e-01 6.33745074e-01
4.31722365e-02 2.00459749e-01 -6.58401668e-01 -5.37725806e-01
5.08399367e-01 -9.43266869e-01 7.88117498e-02 5.44018030e-01
-6.94641411e-01 9.66618508e-02 1.49542952e+00 -7.95697451e-01
5.95752001e-01 -1.97370303e+00 -2.91006386e-01 1.68918997e-01
-1.61424175e-01 1.84131429e-01 1.69742275e-02 5.40793300e-01
4.50079113e-01 -1.19377762e-01 -1.64050624e-01 -7.73813069e-01
3.64873976e-01 -1.30433599e-02 -4.20028090e-01 8.03226233e-01
-1.76086724e-01 8.42576444e-01 -3.81383806e-01 -6.94160521e-01
1.00127947e+00 6.14153862e-01 2.04954352e-02 -2.18074709e-01
-5.52574992e-01 1.36034325e-01 -4.69139069e-01 1.20697069e+00
7.38435090e-01 1.59600042e-02 -1.62823468e-01 1.88983038e-01
-3.43661964e-01 -2.91267801e-02 -1.20019567e+00 1.65597308e+00
-1.22510538e-01 1.48032248e+00 6.79533407e-02 -1.23174645e-01
1.10138941e+00 -7.90248662e-02 1.06292523e-01 -9.91101682e-01
3.15144897e-01 3.42829734e-01 -6.71070039e-01 -4.78317976e-01
9.45817888e-01 5.23893833e-01 2.24608798e-02 2.06977084e-01
-4.80489284e-01 5.49960583e-02 -1.15568917e-02 1.09159313e-01
9.44851160e-01 2.72887677e-01 5.03975265e-02 -9.43281576e-02
4.00551140e-01 3.62156689e-01 -7.40398392e-02 7.69997954e-01
-3.71257752e-01 8.45530927e-01 3.25094312e-02 -4.31592494e-01
-1.12001395e+00 -7.47739077e-01 -1.61433309e-01 1.26092958e+00
5.10203719e-01 -1.52187988e-01 -3.35983515e-01 -7.01727569e-01
1.77701563e-02 7.34512091e-01 -5.31394541e-01 6.57762587e-01
-6.43185139e-01 -2.67046034e-01 6.05217874e-01 4.91741598e-01
8.97770703e-01 -1.03272641e+00 -1.01546812e+00 -2.74149626e-01
-3.32580000e-01 -1.37925971e+00 -5.57509601e-01 1.88346311e-01
-6.89278960e-01 -8.17504644e-01 -1.12444997e+00 -1.09050870e+00
8.95034969e-01 5.59108675e-01 3.15993071e-01 -2.56056875e-01
-3.08560252e-01 4.38557446e-01 -4.29618716e-01 -4.34939921e-01
-1.19541220e-01 -4.27629389e-02 -3.31215680e-01 1.75395131e-01
1.63021892e-01 4.73024957e-02 -6.03166401e-01 6.94847822e-01
-7.70071983e-01 2.24011123e-01 7.29259491e-01 5.96776962e-01
5.86154878e-01 -2.92261481e-01 -2.55960256e-01 -3.47195983e-01
4.03691471e-01 1.86950132e-01 -8.50148976e-01 5.05141795e-01
-2.16615260e-01 -2.62164157e-02 2.09846407e-01 -3.38294059e-01
-8.19040656e-01 6.33644462e-01 2.77985036e-01 -2.47264758e-01
-3.56384307e-01 1.68980613e-01 -1.01957776e-01 -4.30595964e-01
7.32396662e-01 5.81867993e-01 -2.13021263e-01 -5.24782896e-01
3.06706637e-01 1.14639950e+00 5.28110504e-01 -2.01049641e-01
6.71541929e-01 7.92303503e-01 -1.49340183e-01 -1.22752047e+00
-1.50459379e-01 -8.51037025e-01 -8.71309102e-01 -4.72519934e-01
6.29118323e-01 -7.11858332e-01 -6.74050987e-01 7.17432439e-01
-1.08284640e+00 -4.96446073e-01 2.78994381e-01 2.57699817e-01
-5.21214306e-01 6.45800292e-01 -8.52943584e-02 -1.05973828e+00
-5.17626345e-01 -9.55553174e-01 1.88275969e+00 2.69063413e-01
2.26827353e-01 -8.06127369e-01 2.10791398e-02 3.03226590e-01
3.66708636e-01 2.74027020e-01 2.18782395e-01 -5.04942060e-01
-8.17589939e-01 -5.79225421e-01 -4.74903584e-01 -2.08908290e-01
-2.60067046e-01 -1.17114000e-01 -1.09688795e+00 -1.76279679e-01
-5.50112247e-01 -1.34519964e-01 9.29991901e-01 2.43235961e-01
4.96475041e-01 -1.13278948e-01 -5.77944577e-01 4.36442286e-01
1.55723631e+00 3.83525304e-02 6.28475845e-01 8.25440228e-01
9.58127141e-01 3.66889089e-01 8.88503015e-01 3.14393669e-01
3.48061532e-01 9.01256979e-01 2.85835862e-01 -2.11464539e-01
-3.68341893e-01 -5.21146655e-01 6.35412276e-01 3.62066627e-02
5.12149215e-01 -5.50271392e-01 -1.33776033e+00 6.93132818e-01
-2.11669707e+00 -5.85319400e-01 -3.58665138e-01 2.14775062e+00
2.23160863e-01 -1.14289798e-01 1.25406003e-02 1.50156200e-01
9.67846155e-01 1.14216723e-01 -5.25616169e-01 -1.86614856e-01
-3.38691533e-01 -4.86949265e-01 1.06484067e+00 4.46943581e-01
-1.30102825e+00 1.41924798e+00 5.94859552e+00 5.76720357e-01
-1.40690148e+00 -2.24990904e-01 5.70577383e-02 7.67103657e-02
5.04956782e-01 -1.81160077e-01 -8.92566204e-01 3.91309947e-01
3.20789158e-01 2.98126251e-01 3.66870552e-01 1.06885564e+00
2.61809111e-01 -5.43867588e-01 -9.58638549e-01 1.26394486e+00
5.26733398e-01 -9.50951457e-01 -1.83923006e-01 -1.71312720e-01
7.98169553e-01 6.39627934e-01 1.41478106e-01 -8.09713900e-02
1.52268827e-01 -8.81490171e-01 1.15365434e+00 2.91831791e-01
1.06902015e+00 -1.01074368e-01 4.89532471e-01 4.34603870e-01
-1.13120568e+00 6.83597028e-02 -2.78116822e-01 2.81531304e-01
1.27097040e-01 1.93902463e-01 -1.84363163e+00 9.28200260e-02
3.53369802e-01 7.96477258e-01 -1.18571067e+00 1.62565124e+00
-3.51357311e-01 1.81330502e-01 -8.38905692e-01 -6.83269382e-01
3.04905295e-01 1.67186797e-01 6.78039789e-01 1.43768311e+00
2.24320397e-01 -6.06410265e-01 2.27776945e-01 6.64395332e-01
1.32453606e-01 2.54665941e-01 -8.01907837e-01 1.06161639e-01
3.89819622e-01 9.90836859e-01 -1.27394927e+00 -2.51238942e-01
-3.67423543e-03 1.42545438e+00 -1.06869608e-01 4.60454583e-01
-6.91581368e-01 -7.53788114e-01 -5.74668758e-02 9.38774496e-02
4.95348632e-01 -6.15819454e-01 -7.94301271e-01 -1.21674383e+00
5.69763184e-01 -5.83299577e-01 8.21141303e-02 -1.35969615e+00
-7.24012315e-01 2.95461386e-01 -3.72872263e-01 -1.35608125e+00
2.38810882e-01 -8.22761476e-01 -2.93941468e-01 7.04642892e-01
-1.61174750e+00 -1.70607519e+00 -7.60363877e-01 6.41619384e-01
8.85123491e-01 1.25854030e-01 4.29658681e-01 2.09300488e-01
-3.60391408e-01 6.58250034e-01 5.52948236e-01 3.83619457e-01
9.72164571e-01 -1.02162838e+00 5.70777476e-01 1.21614814e+00
3.08602244e-01 3.14982802e-01 7.37308264e-01 -1.05476248e+00
-1.69735205e+00 -1.12098789e+00 9.45330083e-01 -8.78368437e-01
3.51209521e-01 -8.03018093e-01 -4.08135533e-01 8.45081925e-01
-1.23569518e-01 -2.61863321e-01 -1.86691925e-01 -1.83047250e-01
-2.32209340e-01 3.16339582e-02 -1.21873951e+00 9.37715769e-01
8.66530955e-01 -3.77658963e-01 -6.09515548e-01 5.28730273e-01
7.54378363e-02 -7.17566609e-01 1.71597563e-02 3.69585976e-02
7.15078175e-01 -6.19864762e-01 6.69051111e-01 2.90750593e-01
8.36328268e-02 -6.93504989e-01 -2.98412442e-01 -7.97071576e-01
3.89508545e-01 -4.75402802e-01 3.12830359e-01 1.10678780e+00
3.77506793e-01 -6.68634892e-01 8.73851836e-01 6.94041491e-01
-6.24745935e-02 -3.91598716e-02 -1.03604996e+00 -6.52714252e-01
-5.87490559e-01 -5.83458900e-01 2.10439801e-01 7.07931936e-01
2.08228920e-02 2.88935244e-01 -6.38673365e-01 3.42108369e-01
5.94264209e-01 8.27547088e-02 1.17295063e+00 -1.07902956e+00
3.36005360e-01 -3.66201341e-01 -8.77572179e-01 -1.31317186e+00
-4.69185948e-01 -7.14548349e-01 7.24753618e-01 -1.97077096e+00
-1.01249248e-01 -4.01386142e-01 3.50970596e-01 5.91431856e-01
1.19387984e-01 2.39001378e-01 4.88056213e-01 3.97588640e-01
-1.12076616e+00 5.93518496e-01 1.02141488e+00 -1.89587235e-01
-3.29714000e-01 -2.57319331e-01 -8.18311144e-03 6.59572423e-01
6.35469913e-01 -3.24668676e-01 9.07722488e-02 -8.23915303e-01
9.27988216e-02 -1.87463686e-01 5.33013046e-01 -1.19295204e+00
8.03913057e-01 6.76146448e-02 6.08399451e-01 -1.47550941e+00
4.66513216e-01 -9.81959581e-01 -3.54127407e-01 1.57823607e-01
-2.72490710e-01 -2.41261601e-01 2.28999943e-01 7.37115979e-01
1.93694159e-01 -3.89354713e-02 7.05850542e-01 3.95669490e-01
-9.88801599e-01 -2.07886130e-01 -3.18007380e-01 -2.97706008e-01
1.00355351e+00 -7.45731115e-01 -5.51223755e-01 -2.33516499e-01
-1.33721545e-01 5.13252378e-01 9.67295408e-01 6.44152045e-01
9.28084552e-01 -9.76619184e-01 -5.81173420e-01 3.92475009e-01
4.78869915e-01 6.60464466e-02 -8.35919753e-02 7.92375624e-01
-1.18441927e+00 8.45542848e-01 -1.09504543e-01 -1.02832520e+00
-1.57137060e+00 2.08174318e-01 1.80965587e-01 1.00701019e-01
-1.08475220e+00 4.48170722e-01 -8.61057788e-02 -5.25036752e-01
8.18493187e-01 -6.62063837e-01 2.16697827e-01 -5.05740643e-01
5.70272505e-01 4.93253827e-01 1.60392430e-02 -7.99202859e-01
-5.29526412e-01 1.07336938e+00 -1.81320366e-02 -4.13500994e-01
9.74841774e-01 -4.01471853e-01 2.98873782e-01 2.98957705e-01
8.45841646e-01 6.70986623e-02 -1.31194484e+00 -1.93709135e-01
1.94131732e-01 -7.56993711e-01 7.60234669e-02 -1.31657410e+00
-3.73793185e-01 8.79158080e-01 8.05282652e-01 -3.36294383e-01
7.46079147e-01 -3.04085851e-01 6.07066572e-01 8.37600946e-01
6.04639888e-01 -1.38189065e+00 1.84884444e-02 8.13267708e-01
9.88558233e-01 -1.42479801e+00 2.40443530e-03 -1.94352731e-01
-9.35693741e-01 1.14209473e+00 3.70635509e-01 -1.87397022e-02
5.82803003e-02 2.13054344e-01 2.32902244e-01 -2.75020182e-01
-2.20329717e-01 -2.99088210e-01 3.75028908e-01 6.50537312e-01
-9.80429202e-02 -3.07716839e-02 3.88051942e-02 -2.53372461e-01
-5.65945394e-02 -1.08970646e-02 3.78372222e-01 1.18872654e+00
-8.08169246e-01 -6.40585780e-01 -8.29254150e-01 1.13892667e-01
-6.66882694e-02 8.44924971e-02 -1.09906137e+00 7.76573241e-01
-3.55406016e-01 9.35020149e-01 -2.48786762e-01 -2.55072802e-01
3.89138937e-01 1.78021133e-01 2.80378729e-01 -2.88697600e-01
-2.67007798e-01 2.96747327e-01 9.56729352e-02 -4.59485710e-01
-5.16767092e-02 -6.49676383e-01 -1.28110027e+00 -1.22687891e-01
-4.42704260e-01 -6.52380824e-01 1.42749500e+00 7.49005854e-01
4.04863060e-01 1.22376084e-01 4.40737426e-01 -1.15119982e+00
-2.21987844e-01 -1.08791053e+00 -4.47915614e-01 2.32744128e-01
4.04663414e-01 -5.99777162e-01 -1.59236908e-01 2.60781646e-01] | [11.945646286010742, 2.2488279342651367] |
224e954b-a461-43f0-97a8-065fe6a43b3b | robust-decision-focused-learning-for-reward | 2304.03365 | null | https://arxiv.org/abs/2304.03365v1 | https://arxiv.org/pdf/2304.03365v1.pdf | Robust Decision-Focused Learning for Reward Transfer | Decision-focused (DF) model-based reinforcement learning has recently been introduced as a powerful algorithm which can focus on learning the MDP dynamics which are most relevant for obtaining high rewards. While this approach increases the performance of agents by focusing the learning towards optimizing for the reward directly, it does so by learning less accurate dynamics (from a MLE standpoint), and may thus be brittle to changes in the reward function. In this work, we develop the robust decision-focused (RDF) algorithm which leverages the non-identifiability of DF solutions to learn models which maximize expected returns while simultaneously learning models which are robust to changes in the reward function. We demonstrate on a variety of toy example and healthcare simulators that RDF significantly increases the robustness of DF to changes in the reward function, without decreasing the overall return the agent obtains. | ['Finale Doshi-Velez', 'Omer Gottesman', 'Sonali Parbhoo', 'Abhishek Sharma'] | 2023-04-06 | null | null | null | null | ['model-based-reinforcement-learning'] | ['reasoning'] | [-2.54640758e-01 2.95264453e-01 -3.21972400e-01 2.25000843e-01
-9.02742624e-01 -5.53065538e-01 6.28940582e-01 3.84924501e-01
-5.99900007e-01 1.08512425e+00 -1.35889873e-02 -3.80933851e-01
-5.83237052e-01 -7.21125305e-01 -7.73587406e-01 -7.94700325e-01
-5.66171706e-01 5.48896134e-01 -1.56318128e-01 -2.32327297e-01
1.80462033e-01 3.80285054e-01 -1.18579125e+00 -3.83576125e-01
8.96331608e-01 8.66707861e-01 1.43643171e-01 5.63854516e-01
4.92161304e-01 1.06716776e+00 -5.33253789e-01 2.80323923e-02
4.23109829e-01 -2.79858202e-01 -3.75810534e-01 -1.94053650e-01
-3.60794246e-01 -7.53669083e-01 -1.89998761e-01 1.03969157e+00
4.58758712e-01 1.28659740e-01 5.40234089e-01 -1.24272096e+00
4.87342030e-02 7.03736901e-01 -4.86443818e-01 1.50110126e-01
1.01591051e-01 6.32752657e-01 1.05822182e+00 -3.96863073e-02
4.47124779e-01 1.41064095e+00 2.88057327e-01 5.54188728e-01
-1.46113348e+00 -6.74347579e-01 2.00926900e-01 -2.02815309e-01
-8.69264841e-01 -1.01618148e-01 4.79758531e-01 -3.79104674e-01
9.34660435e-01 -4.42512222e-02 7.71024644e-01 9.10693109e-01
6.44974232e-01 6.31059587e-01 1.31913495e+00 -5.98079413e-02
6.56486273e-01 -7.03001320e-02 -1.35115430e-01 4.15126950e-01
5.24384022e-01 1.14855838e+00 -1.93053722e-01 -2.38565296e-01
8.73204350e-01 -4.47662957e-02 3.16908620e-02 -7.02123821e-01
-1.00104380e+00 1.04166555e+00 3.82680178e-01 -1.55708015e-01
-7.54575968e-01 6.46076977e-01 1.50501430e-01 6.88698411e-01
-3.94269563e-02 1.03715754e+00 -4.27466929e-01 -4.14401114e-01
-4.62160051e-01 8.80432665e-01 6.80530548e-01 4.62198198e-01
6.56402707e-01 3.96635234e-01 -3.06965172e-01 1.64775066e-02
3.39214891e-01 6.56014621e-01 2.24720210e-01 -1.37581635e+00
5.75874746e-01 4.33492452e-01 8.57872069e-01 -6.19939983e-01
-6.60294712e-01 -7.03593850e-01 -9.05068442e-02 6.93530202e-01
5.41069567e-01 -8.85560632e-01 -7.77398646e-01 2.01794910e+00
3.34564328e-01 -1.86286941e-01 2.44396240e-01 6.98226094e-01
-1.53249413e-01 6.05267107e-01 1.25971634e-03 -4.98528749e-01
4.79527503e-01 -5.55442929e-01 -5.70755959e-01 -2.09934711e-01
7.56965458e-01 -1.63381666e-01 9.49617028e-01 3.70939434e-01
-1.25262582e+00 -1.84392221e-02 -1.04989517e+00 8.79944205e-01
1.44463256e-01 -4.98183846e-01 3.16724956e-01 4.57667887e-01
-8.32969725e-01 8.41776848e-01 -1.04786670e+00 -2.61213090e-02
4.42377090e-01 6.77338779e-01 2.94982731e-01 1.49322957e-01
-1.07892370e+00 1.30142939e+00 4.00899529e-01 -3.70657235e-01
-1.49185395e+00 -8.59486401e-01 -4.86967593e-01 1.85225204e-01
9.22550201e-01 -5.60148358e-01 1.61259329e+00 -8.40378642e-01
-1.84436655e+00 -1.28911898e-01 4.86924648e-01 -9.66956496e-01
7.68897712e-01 -4.00463156e-02 1.02892645e-01 2.17591166e-01
-6.76499903e-02 6.30193949e-01 1.07258964e+00 -1.16994572e+00
-7.63937116e-01 -1.83721304e-01 3.70442241e-01 3.20498288e-01
7.47515410e-02 -3.36377621e-01 6.66119576e-01 -3.86653155e-01
-6.39412761e-01 -1.13208699e+00 -6.42309070e-01 -4.35125947e-01
-1.19208746e-01 -1.10810734e-01 5.66137314e-01 -3.80834937e-01
1.01969433e+00 -1.79441917e+00 2.25050598e-01 5.20889580e-01
4.58448492e-02 -3.38986963e-02 -9.84668508e-02 7.36594498e-01
2.87806541e-01 -1.55633893e-02 -6.15177490e-03 1.83280587e-01
1.90822989e-01 4.05617386e-01 -3.32646549e-01 4.64938998e-01
4.13196802e-01 1.02056348e+00 -9.71150756e-01 -1.80122838e-03
1.07871741e-01 -4.85715419e-02 -7.43548095e-01 4.07349616e-01
-4.76388514e-01 5.51958263e-01 -8.01638842e-01 2.67304242e-01
2.79378623e-01 -1.92789026e-02 3.58236879e-01 5.73068380e-01
-1.10546358e-01 9.04135481e-02 -1.16239357e+00 8.47525060e-01
-4.22653198e-01 -1.01401359e-02 1.65982395e-01 -1.14112878e+00
6.38356745e-01 -5.34914108e-03 8.17467868e-01 -7.33392656e-01
1.84864670e-01 1.50375322e-01 4.09972250e-01 -3.80447626e-01
3.10430765e-01 -4.48422283e-01 -2.62843132e-01 6.92890108e-01
-1.25599131e-01 -2.32771337e-01 1.82208180e-01 1.10596381e-01
1.16550076e+00 1.77474782e-01 2.91293293e-01 -5.62593341e-01
-1.59326661e-02 1.54986799e-01 4.99995112e-01 8.98163319e-01
-1.45004854e-01 -3.16255391e-01 9.11979020e-01 2.69015152e-02
-9.84249115e-01 -1.21930254e+00 1.18404977e-01 5.82824409e-01
2.04881296e-01 -1.96546108e-01 -5.00838518e-01 -6.74485803e-01
5.39711893e-01 9.40249681e-01 -6.82790279e-01 -3.97486806e-01
-4.65590090e-01 -7.80855954e-01 9.85501409e-02 3.66155058e-01
1.14783160e-01 -8.39932799e-01 -1.25115347e+00 5.01677155e-01
2.09766820e-01 -5.55770993e-01 -3.51787210e-01 3.97280276e-01
-8.92895758e-01 -1.16006303e+00 -5.97199798e-01 2.03233911e-03
6.61175251e-01 -2.15374425e-01 7.67840803e-01 -3.51107031e-01
1.10745626e-02 8.34392726e-01 -1.10942952e-01 -6.19354784e-01
-5.37387431e-01 -6.33314922e-02 2.11350709e-01 -3.27941954e-01
-3.38504761e-01 -2.93880254e-01 -5.03684759e-01 2.98813246e-02
-7.24276364e-01 -3.12532634e-01 4.86961603e-01 8.98834348e-01
4.75703984e-01 3.09707224e-01 1.25677252e+00 -6.36540949e-01
1.12880313e+00 -6.87388122e-01 -1.16818070e+00 -4.68508303e-02
-1.05161524e+00 6.70640051e-01 8.03919137e-01 -5.89929163e-01
-7.79785156e-01 -3.88903283e-02 3.17779809e-01 -3.25816900e-01
2.98681587e-01 5.12580395e-01 2.41492867e-01 -7.59685971e-03
5.18914282e-01 -3.44723724e-02 4.53593969e-01 -2.19965428e-01
9.21481848e-02 8.87762010e-02 -1.99081860e-02 -8.52310836e-01
8.54545116e-01 9.96224359e-02 4.75252777e-01 -2.56917953e-01
-5.09443462e-01 6.80417642e-02 4.29664291e-02 -2.90727973e-01
3.88094753e-01 -8.66668701e-01 -1.06604314e+00 1.81970641e-01
-3.90016913e-01 -6.91768467e-01 -6.74681425e-01 5.39296806e-01
-1.09469664e+00 -1.55641750e-01 -2.83038318e-01 -1.13160610e+00
1.29077211e-02 -1.14975488e+00 3.65115970e-01 3.94442588e-01
-6.89213201e-02 -9.63295221e-01 2.66351610e-01 -2.88868576e-01
5.31229794e-01 5.47095776e-01 9.84675229e-01 -3.94611418e-01
-5.95801711e-01 -6.77605020e-03 3.23312014e-01 1.08367197e-01
-1.79165546e-02 -2.02520385e-01 -5.16528189e-01 -7.84415185e-01
6.29556328e-02 -5.52798271e-01 4.87849087e-01 6.51870608e-01
4.08314645e-01 -7.63266325e-01 -2.16030329e-01 1.46026492e-01
1.51484311e+00 5.55116653e-01 2.46471658e-01 6.30352378e-01
1.49102524e-01 6.49294972e-01 1.05103588e+00 7.94150293e-01
5.11596382e-01 5.82123160e-01 7.62610435e-01 2.29367927e-01
5.66880584e-01 -4.07355100e-01 7.97358155e-01 -2.76548625e-03
1.14189714e-01 1.93852618e-01 -6.65589988e-01 3.90404999e-01
-2.12013769e+00 -1.04500759e+00 5.69589317e-01 2.45535231e+00
9.76306975e-01 3.25480253e-01 9.72616732e-01 -2.08931416e-01
3.58050019e-01 -2.80769616e-01 -1.13909519e+00 -5.90667725e-01
2.14884043e-01 1.58122256e-01 6.81926847e-01 5.05325079e-01
-5.75871885e-01 5.43139935e-01 6.75385714e+00 5.14666975e-01
-1.19202328e+00 -1.99297667e-01 6.65021122e-01 -4.30612415e-01
-1.66514650e-01 -4.99694282e-03 -6.79613888e-01 5.27008533e-01
1.29147279e+00 -7.59970307e-01 6.90892518e-01 9.66236353e-01
8.22424710e-01 -4.63286132e-01 -1.08804667e+00 5.34837425e-01
-7.45808363e-01 -1.29731202e+00 -4.02774930e-01 3.78257155e-01
8.13406408e-01 1.63802307e-03 1.80993125e-01 5.85774422e-01
8.22179496e-01 -1.14508975e+00 7.18291283e-01 6.16176426e-01
2.45743021e-01 -1.40229785e+00 6.95114613e-01 7.08508074e-01
-6.09300256e-01 -7.37277269e-01 3.93632650e-02 -3.78912985e-01
-3.28496732e-02 1.91260427e-01 -1.00060976e+00 1.26176909e-01
3.72130096e-01 4.03266370e-01 -1.99656427e-01 9.76660728e-01
-5.18992841e-02 6.10395968e-01 -4.17687058e-01 -2.84290314e-01
5.55079520e-01 -1.82554990e-01 6.17029786e-01 6.35635853e-01
2.16252431e-01 -2.30700701e-01 5.60196280e-01 9.94345665e-01
2.72493124e-01 -1.68123618e-01 -6.30634546e-01 -2.22381234e-01
2.90892601e-01 9.16181147e-01 -5.25266945e-01 -2.87517421e-02
1.76800132e-01 1.44342065e-01 1.78648964e-01 3.71485174e-01
-7.44009435e-01 -1.22429021e-01 7.85138547e-01 6.44904673e-02
4.54394042e-01 -3.68417650e-01 4.38855290e-02 -8.28397691e-01
-2.43364781e-01 -1.01461649e+00 4.16791618e-01 -1.68116510e-01
-1.06768978e+00 -6.48292527e-02 2.46437892e-01 -1.21495450e+00
-9.99890029e-01 -3.71381491e-01 -4.88109797e-01 6.53381526e-01
-1.75416684e+00 -5.13395607e-01 6.10231102e-01 5.62243879e-01
2.00559080e-01 4.92671877e-02 3.21653366e-01 -2.84134597e-01
-6.10808790e-01 4.23302680e-01 4.59524751e-01 -4.63712245e-01
3.72428030e-01 -1.24532628e+00 -3.01730484e-01 5.86358726e-01
-3.68555993e-01 3.81914884e-01 1.09123731e+00 -6.10138059e-01
-1.83094418e+00 -9.14830863e-01 -5.27119823e-02 -1.71350643e-01
9.25907671e-01 7.60771930e-02 -3.19111347e-01 3.56008530e-01
-1.12198561e-01 -3.46593231e-01 -5.52749559e-02 -2.52157182e-01
2.41505861e-01 -2.56514400e-01 -1.47139275e+00 6.19518101e-01
4.35997516e-01 -9.10784975e-02 -2.33430043e-01 5.86200915e-02
6.65312171e-01 -3.13721597e-01 -9.56915438e-01 2.21696928e-01
2.61405647e-01 -7.17421412e-01 8.14606726e-01 -7.66257465e-01
3.24116230e-01 -1.73507616e-01 1.01943696e-02 -1.77067137e+00
-8.28847140e-02 -1.14981890e+00 -6.90397739e-01 8.29428613e-01
3.07312459e-01 -7.91111887e-01 6.04550898e-01 5.22596896e-01
2.46033758e-01 -9.36135173e-01 -1.10660994e+00 -1.23457575e+00
4.65519458e-01 -9.41452160e-02 4.69130754e-01 5.36312342e-01
4.02501941e-01 7.74200261e-02 -4.35704380e-01 1.57927990e-01
8.85673761e-01 1.68428674e-01 5.86263061e-01 -7.20978975e-01
-7.05410302e-01 -4.70179141e-01 -7.94817209e-02 -6.97431564e-01
2.76243165e-02 -5.22522449e-01 4.62012365e-02 -1.16287816e+00
-1.68811455e-02 -6.10799134e-01 -3.05830747e-01 3.99219453e-01
-1.10927992e-01 -6.60365224e-01 5.61645389e-01 2.05992430e-01
-5.19507051e-01 6.09346449e-01 1.36914003e+00 -1.06707998e-02
-4.88987863e-01 4.82300311e-01 -8.37162971e-01 3.72930914e-01
9.44273472e-01 -6.59784436e-01 -7.04364836e-01 4.64321524e-02
2.64826924e-01 5.32878816e-01 3.44546974e-01 -7.36733079e-01
-1.22739591e-01 -7.68802643e-01 3.90722416e-02 -3.24608684e-01
4.37865965e-02 -7.25992858e-01 -9.11538675e-03 1.20370698e+00
-5.94154477e-01 2.84057766e-01 2.44006500e-01 8.35632205e-01
2.03497574e-01 -2.72223800e-01 9.42860246e-01 -1.91582039e-01
-3.10462683e-01 2.11179286e-01 -5.84749043e-01 3.69042307e-01
1.17884350e+00 1.80324480e-01 -1.25891849e-01 -9.02285218e-01
-4.60294366e-01 6.50572956e-01 4.01438683e-01 2.80325282e-02
3.48434418e-01 -9.34283376e-01 -5.30231535e-01 -2.31700554e-01
-3.62913281e-01 -2.40311548e-01 -6.72661066e-02 8.19618702e-01
-9.36751720e-03 5.53580284e-01 -3.28200132e-01 -3.66813719e-01
-5.95687628e-01 6.76666379e-01 6.92249298e-01 -7.14851975e-01
-5.07154226e-01 2.39229649e-01 -7.15339109e-02 -2.89400399e-01
2.45667338e-01 -4.94364262e-01 5.95664531e-02 6.33678287e-02
3.27674001e-01 4.50078785e-01 -2.13460982e-01 -9.85582024e-02
-1.33795604e-01 1.33990526e-01 -8.42808634e-02 -3.32344323e-01
1.42125297e+00 -6.72950894e-02 6.27451897e-01 2.11417019e-01
6.26522779e-01 -3.13278913e-01 -2.11894345e+00 -3.82103175e-02
4.12714571e-01 -2.89005369e-01 3.18004668e-01 -1.00210512e+00
-8.87880027e-01 5.18408895e-01 6.09048545e-01 3.11396837e-01
8.97862554e-01 -4.46313351e-01 5.59993744e-01 6.41776383e-01
6.06788218e-01 -1.29988492e+00 3.88948888e-01 4.39302474e-01
4.61670697e-01 -9.88886178e-01 7.99288228e-02 4.45882142e-01
-9.71966445e-01 1.03375912e+00 5.33474386e-01 -5.32428622e-01
5.27467906e-01 5.86948454e-01 -3.05116922e-01 -3.75296251e-04
-1.07123041e+00 -2.62963831e-01 -3.29064041e-01 8.01408291e-01
-3.45684558e-01 1.66488677e-01 -2.23152936e-01 3.63702446e-01
8.27088207e-02 -1.46960914e-02 8.87432814e-01 1.18541574e+00
-5.84064126e-01 -1.23722160e+00 -3.37376565e-01 6.08670771e-01
-2.95714766e-01 3.42880875e-01 -7.04566687e-02 9.87251699e-01
-4.10054147e-01 8.90645385e-01 -3.83652225e-02 -9.36000198e-02
2.66071051e-01 -2.96341062e-01 6.65410578e-01 -3.99091333e-01
-6.24422789e-01 2.05388859e-01 2.17923999e-01 -6.52064025e-01
-2.20376953e-01 -9.36896622e-01 -1.52273071e+00 -3.03072721e-01
-2.52189010e-01 8.09839442e-02 3.52097243e-01 1.14675379e+00
4.15896356e-01 3.74159515e-01 1.11876643e+00 -7.03415334e-01
-1.61688554e+00 -3.39959860e-01 -5.92442095e-01 1.39892837e-02
6.43479764e-01 -9.29682374e-01 -3.09684157e-01 -6.77672863e-01] | [4.157845497131348, 2.37015438079834] |
617501be-4788-4a52-821e-5c7ee02613d1 | a-compact-sequence-encoding-scheme-for-online | 2012.00873 | null | https://arxiv.org/abs/2012.00873v1 | https://arxiv.org/pdf/2012.00873v1.pdf | A compact sequence encoding scheme for online human activity recognition in HRI applications | Human activity recognition and analysis has always been one of the most active areas of pattern recognition and machine intelligence, with applications in various fields, including but not limited to exertion games, surveillance, sports analytics and healthcare. Especially in Human-Robot Interaction, human activity understanding plays a crucial role as household robotic assistants are a trend of the near future. However, state-of-the-art infrastructures that can support complex machine intelligence tasks are not always available, and may not be for the average consumer, as robotic hardware is expensive. In this paper we propose a novel action sequence encoding scheme which efficiently transforms spatio-temporal action sequences into compact representations, using Mahalanobis distance-based shape features and the Radon transform. This representation can be used as input for a lightweight convolutional neural network. Experiments show that the proposed pipeline, when based on state-of-the-art human pose estimation techniques, can provide a robust end-to-end online action recognition scheme, deployable on hardware lacking extreme computing capabilities. | ['Stefanos Kollias', 'Kostas Karpouzis', 'Georgios Tsatiris'] | 2020-12-01 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [ 4.45297241e-01 2.22511161e-02 -1.92981780e-01 -2.13512406e-01
-1.94315121e-01 -1.51948184e-01 4.34724659e-01 1.85632586e-01
-6.83980882e-01 5.29343009e-01 1.31931126e-01 -1.02672994e-01
-2.64134973e-01 -6.84792995e-01 -4.29811060e-01 -4.43929166e-01
-2.37206802e-01 4.13331032e-01 4.59304959e-01 -2.52410084e-01
2.00935915e-01 8.00053418e-01 -1.93272710e+00 2.21385822e-01
4.51255530e-01 1.39365351e+00 5.10352775e-02 7.67134666e-01
2.08997548e-01 9.49855983e-01 -2.08834171e-01 -1.27225608e-01
3.11342299e-01 -2.72905111e-01 -5.43403149e-01 1.94399834e-01
-9.25959125e-02 -4.61322695e-01 -6.76106155e-01 4.97671813e-01
4.35963511e-01 2.94613063e-01 2.01548338e-01 -1.12075698e+00
-5.93289435e-02 5.50670251e-02 -4.05481547e-01 1.90324277e-01
6.91887259e-01 1.85035124e-01 4.94287670e-01 -5.10686040e-01
4.01352644e-01 7.85867691e-01 6.41405642e-01 2.55518019e-01
-5.08317888e-01 -2.81992942e-01 -2.34703913e-01 6.62474751e-01
-1.10598016e+00 -2.10267991e-01 9.46651995e-01 -2.38336220e-01
1.03623343e+00 2.36730486e-01 1.05746424e+00 8.83368313e-01
3.12476456e-01 1.03230608e+00 6.45773232e-01 -3.67818385e-01
2.92317271e-01 -2.78479218e-01 -2.99086601e-01 7.70395875e-01
1.88430563e-01 -2.76865363e-01 -6.57570124e-01 1.72904044e-01
8.70599389e-01 4.34565514e-01 -7.33100027e-02 -7.36052752e-01
-1.45048416e+00 5.51811457e-01 3.74976784e-01 4.29819435e-01
-8.84805083e-01 3.50977838e-01 5.73192835e-01 8.11450854e-02
3.74143839e-01 1.39837831e-01 -2.60316402e-01 -1.00211358e+00
-7.46433318e-01 4.37011421e-01 6.98980868e-01 5.71752727e-01
4.30389047e-01 -2.32453793e-01 6.58321157e-02 4.09277111e-01
1.17455050e-01 3.33833426e-01 6.26720130e-01 -9.19744551e-01
4.67457801e-01 1.04391170e+00 2.19034925e-01 -1.09787393e+00
-7.63961554e-01 -1.66213661e-01 -9.02643859e-01 2.56670386e-01
4.96901691e-01 3.10774073e-02 -2.16494739e-01 1.09736419e+00
5.97968638e-01 1.16186440e-01 -1.65086329e-01 1.27014136e+00
2.97391534e-01 2.77826607e-01 2.34779436e-02 6.64604530e-02
1.54452777e+00 -8.30515444e-01 -5.54621100e-01 -4.05949116e-01
8.77512813e-01 -4.67659712e-01 7.77575254e-01 6.86851084e-01
-8.80649328e-01 -5.43407321e-01 -1.22087121e+00 -1.14424005e-01
-3.46187025e-01 3.87183815e-01 9.78264630e-01 7.44350612e-01
-5.61970830e-01 8.27785194e-01 -1.24807930e+00 -6.75130248e-01
4.31956679e-01 3.94317120e-01 -6.38344586e-01 -8.85417163e-02
-8.29619169e-01 1.04292810e+00 4.33551013e-01 2.44263023e-01
-4.65085626e-01 -2.32931852e-01 -9.42182899e-01 8.94079059e-02
4.99283016e-01 -3.93041551e-01 1.13158989e+00 -8.67385328e-01
-1.68682241e+00 6.14421248e-01 1.70094535e-01 -7.50572383e-01
6.17349684e-01 -6.17706478e-01 -4.23381001e-01 3.40578258e-01
-2.99696326e-01 3.08945775e-01 6.15060806e-01 -3.04563642e-01
-6.58130169e-01 -6.72069430e-01 2.11415943e-02 3.13091785e-01
-4.48625058e-01 1.00590244e-01 -1.15455016e-01 -3.74865621e-01
1.52023032e-01 -9.74804103e-01 -2.29388297e-01 4.60741669e-01
1.03191532e-01 -2.80420482e-01 1.07081032e+00 -7.90509641e-01
1.01770329e+00 -2.09513378e+00 1.79882333e-01 8.91200006e-02
-2.52282947e-01 5.45264244e-01 3.46487552e-01 3.73064131e-01
1.73874751e-01 -6.87753499e-01 -1.98753327e-01 -1.11120060e-01
1.10096380e-01 1.58319056e-01 6.78722933e-02 8.48185301e-01
1.17928289e-01 7.76371419e-01 -8.77121270e-01 -2.95456409e-01
7.29919434e-01 5.29266953e-01 -3.71492147e-01 2.47435823e-01
1.94654390e-02 5.09374499e-01 -4.24778491e-01 4.74391013e-01
9.57045034e-02 -5.50776310e-02 1.66805252e-01 5.01199029e-02
-1.70807108e-01 6.96638450e-02 -1.37574637e+00 2.21715498e+00
-3.67603809e-01 6.98713303e-01 -6.20004628e-03 -1.30568695e+00
9.59026515e-01 3.52844447e-01 1.00397468e+00 -7.84448266e-01
3.30332577e-01 3.08666050e-01 -9.85690206e-02 -7.15092778e-01
5.92864156e-01 1.61298245e-01 -1.19114868e-01 4.59437609e-01
-1.69253930e-01 2.32587337e-01 2.12669149e-01 -2.74969995e-01
1.41353643e+00 6.51465654e-01 5.13257563e-01 1.59258500e-01
6.57552481e-01 4.98736389e-02 3.24217796e-01 1.75380990e-01
-3.69840413e-01 4.87758905e-01 1.69990778e-01 -8.92711759e-01
-1.13405359e+00 -7.55746186e-01 3.76089633e-01 9.74375963e-01
5.89122437e-03 -1.29570961e-01 -8.46741319e-01 -3.36579174e-01
-2.15157524e-01 1.59192413e-01 -2.95651436e-01 -2.17399612e-01
-7.98548341e-01 -2.69172043e-02 6.69092774e-01 8.25335681e-01
8.00230682e-01 -1.34199750e+00 -1.82738662e+00 5.27648091e-01
-8.95716250e-03 -1.17302370e+00 -1.38457611e-01 -3.07991449e-03
-7.94043541e-01 -1.17832899e+00 -8.69987726e-01 -5.29048800e-01
3.01952422e-01 1.99034303e-01 6.90099716e-01 -2.42053792e-02
-7.26829588e-01 4.82549578e-01 -5.11518180e-01 -3.91781777e-01
-5.25390729e-02 8.49874392e-02 3.07676345e-01 3.28738630e-01
5.48549771e-01 -6.91184163e-01 -9.91323113e-01 3.76398087e-01
-8.15929294e-01 -1.60695892e-02 8.09859872e-01 4.79807943e-01
4.10124719e-01 -1.36908561e-01 2.08155587e-01 -3.23988825e-01
2.80479372e-01 -1.13169871e-01 -3.83988321e-01 1.78036913e-01
-1.03195503e-01 1.53326811e-02 6.02677047e-01 -4.07048136e-01
-8.98848057e-01 5.47823250e-01 -2.25218311e-01 -4.41060245e-01
-3.71784568e-01 4.47250187e-01 -2.13744074e-01 6.35324493e-02
6.17266834e-01 3.11874092e-01 2.91445732e-01 -4.47948545e-01
3.37212712e-01 9.33067203e-01 7.41342962e-01 -2.31095284e-01
2.85498351e-01 6.54130816e-01 3.00067991e-01 -1.13680351e+00
-4.44327861e-01 -9.85909224e-01 -9.43698466e-01 -5.56329072e-01
9.79497254e-01 -6.45226896e-01 -1.01264572e+00 6.36767447e-01
-1.23889279e+00 -1.78959131e-01 -3.30928177e-01 7.21297979e-01
-9.88486528e-01 5.61114132e-01 -3.26395541e-01 -8.50292623e-01
-3.16406250e-01 -9.48607862e-01 1.17204154e+00 2.26380616e-01
-4.64933306e-01 -5.49330652e-01 1.43327639e-01 7.13839054e-01
3.48882765e-01 4.62820888e-01 2.26639837e-01 -4.81331825e-01
-5.11680484e-01 -7.28043556e-01 1.17298029e-02 1.66100144e-01
3.32529424e-03 -4.83221143e-01 -6.61448419e-01 -1.00061014e-01
-6.90786615e-02 -5.15964925e-01 3.69247496e-01 2.19065219e-01
1.23210084e+00 -1.62205219e-01 -2.36644536e-01 2.96184629e-01
8.30590308e-01 3.70880142e-02 1.14499950e+00 3.38039488e-01
6.66318417e-01 6.27671421e-01 8.27537775e-01 7.86637783e-01
2.37957299e-01 9.09507811e-01 4.61482733e-01 2.41102856e-02
1.17788829e-01 -1.12851575e-01 3.36951345e-01 4.71519291e-01
-6.35520399e-01 3.00590485e-01 -8.65069151e-01 3.87708336e-01
-2.06691980e+00 -1.33087683e+00 2.13110857e-02 2.37444878e+00
2.45270669e-01 1.02256546e-02 3.32915217e-01 7.02456594e-01
3.36122900e-01 8.11085403e-02 -4.52678859e-01 -3.27764750e-01
4.41932559e-01 2.42872953e-01 5.36135018e-01 -2.08951443e-01
-1.22424793e+00 4.96164262e-01 4.50483274e+00 4.91491616e-01
-1.08668268e+00 1.17454939e-01 2.02814341e-01 2.96117496e-02
5.53136230e-01 -3.72984976e-01 -2.11627722e-01 2.93975294e-01
1.02202725e+00 3.04899633e-01 3.34476084e-01 1.38847697e+00
3.43527079e-01 -2.90387034e-01 -1.05491567e+00 1.36547112e+00
2.50503868e-01 -1.07841325e+00 -6.17044806e-01 8.73607099e-02
1.96420550e-01 -1.90154105e-01 -2.50313610e-01 2.39734203e-01
-3.74382079e-01 -1.01907492e+00 6.96125925e-01 5.28350413e-01
5.23797512e-01 -1.03407633e+00 6.35831535e-01 6.51619494e-01
-1.47185326e+00 -2.15159774e-01 -2.89154232e-01 -4.08139706e-01
3.93899769e-01 3.69344205e-01 -6.71376228e-01 4.90753025e-01
8.97949338e-01 5.48467934e-01 -2.94734031e-01 1.01041472e+00
-5.82174771e-02 1.30568832e-01 -4.12258327e-01 -3.84697884e-01
3.47729400e-02 -1.33658007e-01 2.70204544e-01 9.91013229e-01
5.02678812e-01 2.25386247e-01 3.89083117e-01 2.83725739e-01
1.63461342e-01 1.98544636e-02 -7.69080639e-01 -3.41056645e-01
-1.09462522e-01 1.20809627e+00 -1.08370125e+00 -4.70397435e-02
-4.63028967e-01 1.31511831e+00 2.50556350e-01 -3.27707797e-01
-8.69575083e-01 -6.84916735e-01 7.45350182e-01 4.36507583e-01
1.72251970e-01 -7.31611609e-01 -1.48480698e-01 -8.92090261e-01
3.87309045e-01 -8.43223035e-01 2.45357752e-01 -4.79062408e-01
-6.10092700e-01 8.06295797e-02 -1.14176981e-01 -1.51249135e+00
-5.76961756e-01 -9.50700462e-01 -3.26005876e-01 2.93505549e-01
-9.15503085e-01 -1.18381524e+00 -4.95405763e-01 6.83839083e-01
5.95517933e-01 -8.12008828e-02 8.38745236e-01 3.50682288e-01
-2.45168969e-01 1.81810051e-01 -2.47846767e-01 3.03925216e-01
2.63720512e-01 -6.98443949e-01 2.27079600e-01 8.43845963e-01
2.86811948e-01 2.86216229e-01 5.00826776e-01 -4.54574376e-01
-1.93533075e+00 -8.19031298e-01 8.14735651e-01 -2.16025636e-01
4.26852345e-01 -2.37234741e-01 -6.18032038e-01 4.87294823e-01
-1.87126011e-01 3.29537451e-01 4.60285455e-01 -7.26273432e-02
1.34816170e-01 -2.18020886e-01 -1.02195990e+00 5.58049738e-01
1.14418924e+00 -4.49853957e-01 -3.63551676e-01 4.04414028e-01
1.94605231e-01 -5.53346276e-01 -7.55526304e-01 3.34841400e-01
8.44625950e-01 -1.08491540e+00 9.45696294e-01 -4.97332752e-01
1.83560550e-01 -3.54978800e-01 3.76347862e-02 -8.20145488e-01
-2.70560142e-02 -5.08162498e-01 -4.83763158e-01 5.71749151e-01
-5.91869093e-02 -3.48385185e-01 1.26926303e+00 5.75481117e-01
-1.53237313e-01 -8.71563077e-01 -1.41956294e+00 -7.46940672e-01
-7.12430716e-01 -7.99075186e-01 4.78568137e-01 4.20232147e-01
3.70939851e-01 -9.01764201e-04 -4.91346627e-01 -1.25107527e-01
4.10437018e-01 -1.09789744e-01 9.22449052e-01 -1.23263752e+00
-4.22672480e-01 -2.24717036e-01 -1.25769782e+00 -1.01026058e+00
-6.78908080e-02 -3.73415679e-01 1.46775767e-01 -1.74277437e+00
-2.22265571e-01 -4.46130596e-02 -1.16561741e-01 4.75201875e-01
4.00968313e-01 4.57371324e-01 9.58895236e-02 -1.57526648e-03
-7.44756401e-01 6.40726447e-01 8.81432950e-01 -9.80274230e-02
-1.36913478e-01 1.10143073e-01 2.53693704e-02 7.58696198e-01
6.73836291e-01 -1.71544090e-01 -3.96023452e-01 -9.41040218e-02
2.79641360e-01 1.67475104e-01 4.27859724e-01 -1.87677085e+00
3.84578228e-01 -6.87027276e-02 3.86567205e-01 -3.35111141e-01
7.30444789e-01 -1.13574982e+00 2.44947642e-01 8.72794449e-01
-9.21184793e-02 9.12130624e-02 -1.44761130e-01 5.62753320e-01
-1.26150385e-01 5.70530854e-02 5.25173426e-01 -1.53704867e-01
-9.11982834e-01 2.92762160e-01 -4.79540199e-01 -3.21588248e-01
1.50401163e+00 -5.67092955e-01 3.60729098e-02 -3.27965081e-01
-3.52212608e-01 -1.52806580e-01 2.68977106e-01 5.58438838e-01
7.24705338e-01 -1.22873056e+00 -5.39449453e-01 1.04995191e-01
1.90621018e-01 -7.76076540e-02 4.46778446e-01 9.34938550e-01
-1.01424170e+00 6.16486132e-01 -5.22016287e-01 -5.58028340e-01
-1.33599591e+00 4.96553570e-01 2.82661133e-02 -5.12078643e-01
-8.87298465e-01 3.01773012e-01 -5.62239766e-01 -2.34090477e-01
3.80870283e-01 -2.86794126e-01 -2.32531473e-01 -2.35853389e-01
9.14330840e-01 6.88901782e-01 3.33625764e-01 -6.82363451e-01
-4.23508197e-01 2.69280106e-01 4.16048288e-01 -2.08208784e-02
1.27306807e+00 1.86946541e-01 1.74509510e-01 3.77183199e-01
8.56195867e-01 -5.90060711e-01 -1.26622283e+00 6.41327798e-02
2.20336527e-01 -4.42352265e-01 -1.19734325e-01 -3.39361846e-01
-8.74570608e-01 1.12624335e+00 9.20324981e-01 8.48639663e-03
1.07159674e+00 -2.11288676e-01 1.03919780e+00 8.14522564e-01
9.67000306e-01 -1.45276237e+00 2.27191716e-01 3.62835467e-01
7.53798604e-01 -1.01973879e+00 2.47135818e-01 -2.70554245e-01
-6.43353879e-01 1.12860763e+00 3.53543848e-01 -1.26485273e-01
5.32509208e-01 1.18058033e-01 -1.81404561e-01 -1.03093810e-01
-1.46873906e-01 -3.20452213e-01 1.29710138e-01 6.46195590e-01
4.24880475e-01 8.53271633e-02 -4.97080833e-01 4.66845661e-01
-7.51786912e-03 4.32608902e-01 1.68512732e-01 1.34005094e+00
-5.26914895e-01 -9.10721838e-01 -4.10043299e-01 4.13399488e-01
-3.92615467e-01 5.75548768e-01 4.76208143e-02 5.71982086e-01
9.92511809e-02 7.64753699e-01 1.39559522e-01 -4.32168961e-01
5.69988728e-01 2.33848751e-01 6.70486987e-01 -3.84900421e-01
-4.81604189e-01 -5.78380108e-01 1.87724411e-01 -1.00906873e+00
-5.49735963e-01 -6.87954545e-01 -1.31781507e+00 -3.15037996e-01
2.30156913e-01 -3.09785813e-01 1.06645453e+00 1.20241964e+00
4.35633898e-01 2.95023620e-01 3.44269663e-01 -1.41859746e+00
-5.44715524e-01 -9.95850861e-01 -5.78430116e-01 5.24407029e-01
-2.10885286e-01 -9.11848545e-01 3.13037604e-01 5.21849804e-02] | [7.7573466300964355, 0.39160603284835815] |
55247958-ad8d-4047-a467-4b14bc114d7d | improving-sign-language-translation-with | 2105.12397 | null | https://arxiv.org/abs/2105.12397v1 | https://arxiv.org/pdf/2105.12397v1.pdf | Improving Sign Language Translation with Monolingual Data by Sign Back-Translation | Despite existing pioneering works on sign language translation (SLT), there is a non-trivial obstacle, i.e., the limited quantity of parallel sign-text data. To tackle this parallel data bottleneck, we propose a sign back-translation (SignBT) approach, which incorporates massive spoken language texts into SLT training. With a text-to-gloss translation model, we first back-translate the monolingual text to its gloss sequence. Then, the paired sign sequence is generated by splicing pieces from an estimated gloss-to-sign bank at the feature level. Finally, the synthetic parallel data serves as a strong supplement for the end-to-end training of the encoder-decoder SLT framework. To promote the SLT research, we further contribute CSL-Daily, a large-scale continuous SLT dataset. It provides both spoken language translations and gloss-level annotations. The topic revolves around people's daily lives (e.g., travel, shopping, medical care), the most likely SLT application scenario. Extensive experimental results and analysis of SLT methods are reported on CSL-Daily. With the proposed sign back-translation method, we obtain a substantial improvement over previous state-of-the-art SLT methods. | ['Houqiang Li', 'Junfu Pu', 'Weizhen Qi', 'Wengang Zhou', 'Hao Zhou'] | 2021-05-26 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zhou_Improving_Sign_Language_Translation_With_Monolingual_Data_by_Sign_Back-Translation_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zhou_Improving_Sign_Language_Translation_With_Monolingual_Data_by_Sign_Back-Translation_CVPR_2021_paper.pdf | cvpr-2021-1 | ['sign-language-recognition', 'sign-language-translation'] | ['computer-vision', 'computer-vision'] | [ 4.72781450e-01 -1.86709389e-01 -2.36775026e-01 -6.35562897e-01
-1.27585351e+00 -5.13338566e-01 5.53573310e-01 -7.71460474e-01
-3.63396972e-01 5.75407505e-01 6.73817992e-01 -1.98941007e-01
5.23874760e-01 -2.13368878e-01 -7.19004214e-01 -5.77014804e-01
4.89085793e-01 7.56725669e-01 7.92759806e-02 -2.52839416e-01
-6.29878491e-02 -1.57343835e-01 -1.20406151e+00 5.22526979e-01
1.00433409e+00 6.90765619e-01 1.46170110e-01 5.89249551e-01
-3.08566004e-01 6.42524958e-01 -4.54448521e-01 -6.93838060e-01
2.67924666e-01 -1.18292344e+00 -4.40357149e-01 9.80669856e-02
7.14662790e-01 -6.61291838e-01 -2.95924932e-01 7.89432824e-01
8.04361701e-01 -4.14073050e-01 5.19467652e-01 -1.24534440e+00
-6.89139962e-01 6.64117277e-01 -2.11216867e-01 -6.05650723e-01
4.13640469e-01 3.94087493e-01 1.14608169e+00 -1.36480594e+00
1.10022187e+00 1.32241762e+00 4.84469891e-01 7.91414917e-01
-7.29955256e-01 -5.14073133e-01 7.97971487e-02 7.80160651e-02
-1.11276841e+00 -6.87324584e-01 4.26723361e-01 -4.55845110e-02
8.97927642e-01 1.73644543e-01 9.44873869e-01 1.24245358e+00
-1.60021201e-01 1.66625440e+00 1.27472329e+00 -6.61333144e-01
-1.61763988e-02 -3.23458582e-01 -1.99976787e-01 8.17839146e-01
-2.07555771e-01 -8.35582614e-02 -1.12812805e+00 2.54850626e-01
5.29425621e-01 -4.04008210e-01 -2.26128817e-01 -1.01008870e-01
-1.57154346e+00 3.53578717e-01 5.57182264e-03 1.76430941e-01
-2.41269469e-01 1.41342610e-01 5.55338025e-01 7.12075591e-01
2.42005065e-01 -3.85725915e-01 -5.16975343e-01 -4.66503948e-01
-9.90141392e-01 2.46624634e-01 9.35872018e-01 1.39908504e+00
1.35407388e-01 1.36601785e-02 -4.53005880e-01 9.50591087e-01
6.10187590e-01 1.21367002e+00 4.81307030e-01 -5.02782047e-01
1.01378500e+00 3.11711192e-01 -1.35514781e-01 -3.89301330e-02
8.05747509e-02 -4.33450192e-02 -6.13833487e-01 -1.73327625e-02
6.65435433e-01 -5.70863485e-02 -1.21022403e+00 1.60027874e+00
1.98228374e-01 -4.24247980e-02 3.02314967e-01 1.25155044e+00
7.56607056e-01 4.19222713e-01 -1.35328397e-01 -3.55732292e-02
1.45147073e+00 -1.42000830e+00 -9.59699512e-01 -2.57309139e-01
8.35170567e-01 -1.14549088e+00 1.33201897e+00 2.05697015e-01
-1.10493982e+00 -1.35561854e-01 -7.48934150e-01 -4.53855962e-01
-2.40408227e-01 6.63280249e-01 1.79232389e-01 3.25257033e-01
-8.46160710e-01 1.72005132e-01 -8.95606041e-01 -6.93420172e-01
3.03231776e-01 -1.37364596e-01 -1.94092855e-01 -5.88847399e-01
-1.02062392e+00 9.27037120e-01 -7.08061755e-02 4.17814642e-01
-4.49221224e-01 -2.86317557e-01 -7.76743174e-01 -5.36531687e-01
3.14979136e-01 -8.94345701e-01 1.74767458e+00 -9.53825057e-01
-2.06542730e+00 9.92423058e-01 -6.26457453e-01 -1.54669762e-01
1.30916357e+00 -4.66155767e-01 -3.38030308e-01 -1.53564870e-01
7.05746561e-02 5.51497519e-01 9.39561427e-01 -9.84392464e-01
-7.44841337e-01 -1.74391225e-01 -6.37425303e-01 4.92402464e-01
1.24122515e-01 4.09116328e-01 -8.06818783e-01 -7.68718362e-01
2.32330322e-01 -1.10686624e+00 2.43018374e-01 3.72015387e-01
-4.35711443e-01 -2.40979329e-01 7.61167705e-01 -1.17588139e+00
1.12089324e+00 -2.06238174e+00 3.49373192e-01 8.31539482e-02
-2.57302940e-01 3.88726532e-01 -4.64865416e-01 8.27454686e-01
4.19203192e-01 -1.63807809e-01 -5.76357484e-01 -9.90813375e-01
4.69520718e-01 5.93105674e-01 -4.24248576e-01 1.65271446e-01
2.31103271e-01 1.51892066e+00 -9.39722657e-01 -6.63144410e-01
5.06914444e-02 3.57122868e-01 -3.17419916e-01 2.09851563e-01
-3.97475779e-01 6.61222577e-01 -2.98494637e-01 9.86703277e-01
3.90251517e-01 6.06219247e-02 3.02682638e-01 8.95367656e-03
-3.10690403e-01 6.60150230e-01 -6.65530503e-01 2.09085083e+00
-4.85638887e-01 6.41825497e-01 -1.05116293e-01 -5.36375523e-01
7.37671733e-01 4.03396666e-01 1.73782319e-01 -9.84240770e-01
9.54778343e-02 9.61074412e-01 -1.24621950e-01 -8.12701821e-01
2.86108196e-01 -2.16363758e-01 -1.94232583e-01 7.07015693e-01
-5.86369634e-02 -3.08185309e-01 3.81460935e-01 -4.74292636e-02
9.70040798e-01 7.47892857e-01 4.63969186e-02 3.49311680e-01
2.41471589e-01 1.55327380e-01 4.18906122e-01 2.17167422e-01
-1.65000528e-01 7.66356707e-01 2.33096763e-01 -5.77737018e-02
-1.14949930e+00 -1.07275009e+00 3.05606633e-01 1.03405428e+00
-2.59610325e-01 -3.78059477e-01 -8.06405425e-01 -7.70370245e-01
-1.77782103e-01 7.05545783e-01 -1.74428850e-01 1.28037468e-01
-1.19399393e+00 -4.03377175e-01 9.48708892e-01 5.81494629e-01
8.19960296e-01 -1.30147529e+00 -4.38986868e-01 1.45528406e-01
-6.45704508e-01 -1.33788812e+00 -1.26294887e+00 -3.40381980e-01
-7.29862690e-01 -6.75816834e-01 -1.24805379e+00 -1.25956440e+00
6.75149143e-01 9.16948467e-02 8.66630852e-01 -3.99066024e-02
-8.11953619e-02 7.45338276e-02 -6.57057226e-01 -2.34897032e-01
-7.96811044e-01 -1.53033778e-01 -4.01305035e-02 1.54526249e-01
5.02906144e-01 -3.39647740e-01 -4.16555405e-01 4.20443505e-01
-6.73162460e-01 4.76875097e-01 9.78208184e-01 1.04553151e+00
5.59452236e-01 -9.56215322e-01 3.83664012e-01 -3.14989299e-01
5.47922134e-01 1.43445194e-01 -3.07458133e-01 6.34227276e-01
-4.04505551e-01 2.43356213e-01 4.00401980e-01 -5.47216058e-01
-9.82649028e-01 5.55829331e-02 -3.71625781e-01 -1.30354658e-01
1.28166795e-01 5.52527845e-01 -3.18798810e-01 2.76787549e-01
1.05474271e-01 8.87508273e-01 1.88299462e-01 -6.87956452e-01
7.28394866e-01 1.02404547e+00 6.23967052e-01 -3.58371675e-01
8.64322960e-01 1.90379784e-01 -3.19570780e-01 -8.04090202e-01
-6.03524148e-01 -2.07355037e-01 -8.37455750e-01 -2.72373438e-01
6.21339738e-01 -7.08081007e-01 -4.90700662e-01 9.89379525e-01
-1.22057378e+00 -7.55346477e-01 -3.13393801e-01 5.36297143e-01
-7.90144682e-01 4.53636140e-01 -7.13422716e-01 -7.47153580e-01
-5.28159142e-01 -9.80690241e-01 1.62187338e+00 -1.38146147e-01
-2.94396490e-01 -6.00867629e-01 2.01855347e-01 6.04975820e-01
3.75250548e-01 -1.28111735e-01 6.32025182e-01 -2.31833175e-01
-7.29155600e-01 -2.50876099e-01 -3.40868354e-01 6.03171527e-01
1.46199673e-01 -2.37029657e-01 -6.62520468e-01 -1.90846652e-01
-5.55301368e-01 -7.77753174e-01 7.85897374e-01 -1.08374655e-02
2.37328604e-01 -4.06424433e-01 -2.27394328e-02 5.73714972e-01
9.76526856e-01 -1.36776343e-01 6.48468435e-01 -9.40156803e-02
6.87920272e-01 5.53432941e-01 7.40774989e-01 1.91255569e-01
8.94656718e-01 9.08947587e-01 -3.42459530e-01 -7.68546909e-02
-9.65702593e-01 -9.60423648e-01 9.04365063e-01 1.66265607e+00
-2.30984479e-01 -3.53923947e-01 -7.44570255e-01 5.33810437e-01
-1.98610902e+00 -6.46231771e-01 -1.89849973e-01 2.02446914e+00
1.02978015e+00 -2.59882569e-01 2.64793597e-02 8.51192921e-02
1.69965178e-01 4.71610948e-02 -5.63727558e-01 -6.36110753e-02
-3.54384094e-01 1.49143770e-01 4.61982936e-01 5.61329544e-01
-6.55981541e-01 1.47213900e+00 5.78144217e+00 7.27068722e-01
-1.18069589e+00 1.95898935e-01 -3.12500060e-01 -1.02663338e-01
-1.49183095e-01 -6.33200929e-02 -7.93736339e-01 5.00247359e-01
7.52947330e-01 7.69748837e-02 5.87921441e-01 4.20412868e-01
5.94931066e-01 4.69688736e-02 -1.11174595e+00 9.65522647e-01
3.87611061e-01 -7.93636382e-01 1.88979134e-01 9.82492715e-02
6.99511111e-01 3.71573508e-01 -1.47289023e-01 4.14451152e-01
4.05787945e-01 -6.33421421e-01 1.23387003e+00 5.56341529e-01
1.46117580e+00 -8.08365941e-02 6.43406093e-01 4.15187776e-01
-1.34590638e+00 2.50087500e-01 2.66755968e-01 1.29523933e-01
9.73649085e-01 2.19672740e-01 -7.80081630e-01 4.75437403e-01
3.18589896e-01 9.12694991e-01 -1.00891903e-01 8.40847671e-01
-8.32547009e-01 8.90518486e-01 -6.75363421e-01 -2.63339609e-01
2.26251602e-01 -4.08502817e-01 5.51497936e-01 1.38036478e+00
6.38722956e-01 -1.06417887e-01 1.22978821e-01 5.71023583e-01
-1.87712803e-01 3.29694122e-01 -5.52824736e-01 -3.03482592e-01
8.86469781e-02 5.80736578e-01 -3.02493751e-01 -6.20408177e-01
-5.79039693e-01 1.94181800e+00 6.73301984e-03 6.43436134e-01
-5.79214454e-01 -2.34155521e-01 7.65126646e-01 -1.55498758e-01
2.99176306e-01 -5.33948004e-01 -2.04907566e-01 -1.52518225e+00
7.95476139e-01 -1.10008669e+00 -4.69670817e-02 -8.03687155e-01
-1.26126540e+00 3.64023536e-01 -3.02837521e-01 -1.72000015e+00
-5.87133110e-01 -5.56864738e-01 -2.79950321e-01 8.84143353e-01
-1.67176843e+00 -1.90514731e+00 -2.14091793e-01 5.37904680e-01
8.79627109e-01 -3.28850225e-02 7.47890592e-01 4.83524442e-01
-1.80217206e-01 9.74383473e-01 2.53672153e-01 5.06016970e-01
9.72706079e-01 -8.99219811e-01 9.10041809e-01 8.98813426e-01
3.35326076e-01 4.05931503e-01 3.84770393e-01 -8.71300340e-01
-1.54973090e+00 -1.04017949e+00 1.73606479e+00 -5.71891665e-01
7.45352030e-01 -5.71898460e-01 -4.39082503e-01 7.75377810e-01
1.56977206e-01 -2.41728842e-01 2.42660224e-01 -5.53625762e-01
-4.01779711e-01 1.94588676e-01 -7.50701368e-01 1.08878446e+00
1.73080504e+00 -6.59959137e-01 -6.86954439e-01 3.33516091e-01
5.08630455e-01 -6.39456987e-01 -5.28879523e-01 6.66560903e-02
1.19327283e+00 -4.64104146e-01 5.30812323e-01 -4.86572117e-01
4.65859771e-01 -3.77660394e-01 -2.49160290e-01 -1.45094216e+00
5.09916663e-01 -9.51880157e-01 -1.46411201e-02 9.83066857e-01
5.18164694e-01 -6.52869523e-01 6.45226896e-01 3.19856346e-01
-4.62172389e-01 -5.73336780e-01 -1.16020095e+00 -1.01709151e+00
-2.47718696e-03 -5.30935049e-01 4.24639195e-01 4.56371456e-01
2.63346620e-02 4.46253359e-01 -8.51060331e-01 -3.56034100e-01
8.23208451e-01 2.14654386e-01 1.12036848e+00 -6.25889719e-01
-3.05357069e-01 -4.30766433e-01 5.18329442e-02 -1.93354440e+00
-1.12521671e-01 -1.10760283e+00 6.52407050e-01 -1.92828512e+00
-7.71737322e-02 -7.71149695e-02 3.73767585e-01 7.61333466e-01
1.27890572e-01 3.04100782e-01 5.12871206e-01 4.36782628e-01
-3.95247549e-01 8.58727276e-01 1.81602001e+00 -9.31139886e-02
-1.65375635e-01 1.67600904e-02 -7.01179951e-02 5.44329941e-01
5.80470562e-01 -3.77373189e-01 -2.55654398e-02 -8.48777413e-01
-2.14443170e-02 7.82613531e-02 2.27399379e-01 -5.09278059e-01
2.24540100e-01 -5.03991060e-02 -3.84582520e-01 -7.37241089e-01
4.01277512e-01 -5.74525833e-01 -3.83230865e-01 4.50888574e-01
-3.37187648e-01 -9.52053443e-02 -3.15448374e-01 2.10299015e-01
-3.93471032e-01 1.26372054e-01 5.37335813e-01 3.02059520e-02
-4.44841236e-01 1.69855893e-01 -4.57035542e-01 1.22491829e-01
4.49033380e-01 -2.51732796e-01 -3.56660992e-01 -5.32861948e-01
-4.64212298e-01 4.15477484e-01 3.23322594e-01 5.92570901e-01
7.05015182e-01 -1.53730989e+00 -1.17394459e+00 4.90294695e-01
4.85825300e-01 -6.82903603e-02 -1.75522625e-01 1.23531508e+00
-4.61578548e-01 4.80809957e-01 -3.15875164e-03 -4.02035475e-01
-1.50745809e+00 -2.39824772e-01 1.04545265e-01 -2.62483984e-01
-1.00559640e+00 6.92015231e-01 -3.75041157e-01 -7.52262533e-01
3.01634043e-01 -8.30781519e-01 5.77599108e-01 -1.73176348e-01
4.97847855e-01 1.02143861e-01 -1.38047457e-01 -8.75337780e-01
-2.02043951e-01 9.26995456e-01 2.15211466e-01 -8.25980186e-01
1.15118361e+00 -2.40164250e-01 3.81445885e-02 5.16637564e-01
1.12649119e+00 -2.37398621e-04 -1.12654281e+00 -5.53081393e-01
1.44633681e-01 -3.56038988e-01 -5.47060132e-01 -1.21827066e+00
-7.09715784e-01 9.93493974e-01 3.20351332e-01 -7.30752468e-01
9.44325447e-01 2.38756202e-02 1.47094274e+00 6.29738450e-01
6.10887229e-01 -1.29172301e+00 7.69066066e-02 9.57946420e-01
1.35430157e+00 -1.22986042e+00 -4.75766331e-01 -2.97790796e-01
-8.45227540e-01 9.65686321e-01 1.81786582e-01 1.75268829e-01
3.35502654e-01 3.22038203e-01 7.56046951e-01 1.59600064e-01
-5.67009985e-01 -3.61139655e-01 2.45937988e-01 4.68160033e-01
4.21030760e-01 1.77139565e-01 -6.76151395e-01 2.81170845e-01
-4.74096835e-01 6.30449474e-01 1.10842332e-01 1.12724423e+00
-1.64591968e-01 -1.39006376e+00 -2.36252069e-01 1.82869166e-01
1.38767228e-01 -4.35054302e-01 -7.46718943e-01 6.15985155e-01
-4.34201621e-02 7.09669173e-01 -4.02287304e-01 -1.94800749e-01
6.83785439e-01 6.30186796e-01 7.11497545e-01 -3.24407279e-01
-3.07679236e-01 2.63630331e-01 4.03891504e-01 -5.99906445e-01
-4.03512776e-01 -1.03214979e+00 -1.44970763e+00 4.61348891e-02
9.36636031e-02 -4.37999874e-01 8.38696837e-01 1.31362486e+00
8.73870626e-02 2.13640109e-01 1.10668331e-01 -7.31826842e-01
-7.50532985e-01 -1.31312394e+00 -3.74752373e-01 5.38263619e-01
2.97495067e-01 -2.39141747e-01 -2.43203148e-01 4.45753247e-01] | [9.207501411437988, -6.529805660247803] |
c2511266-2eab-4f58-a06f-fc34379d7d0f | robust-real-time-tracking-of-multiple-objects | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Possegger_Robust_Real-Time_Tracking_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Possegger_Robust_Real-Time_Tracking_2013_CVPR_paper.pdf | Robust Real-Time Tracking of Multiple Objects by Volumetric Mass Densities | Combining foreground images from multiple views by projecting them onto a common ground-plane has been recently applied within many multi-object tracking approaches. These planar projections introduce severe artifacts and constrain most approaches to objects moving on a common 2D ground-plane. To overcome these limitations, we introduce the concept of an occupancy volume exploiting the full geometry and the objects' center of mass and develop an efficient algorithm for 3D object tracking. Individual objects are tracked using the local mass density scores within a particle filter based approach, constrained by a Voronoi partitioning between nearby trackers. Our method benefits from the geometric knowledge given by the occupancy volume to robustly extract features and train classifiers on-demand, when volumetric information becomes unreliable. We evaluate our approach on several challenging real-world scenarios including the public APIDIS dataset. Experimental evaluations demonstrate significant improvements compared to state-of-theart methods, while achieving real-time performance. | ['Thomas Mauthner', 'Horst Bischof', 'Peter M. Roth', 'Horst Possegger', 'Sabine Sternig'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['3d-object-tracking'] | ['computer-vision'] | [ 8.86443704e-02 -4.06282604e-01 1.68916471e-02 9.17801186e-02
-8.98684204e-01 -7.86069632e-01 7.33064830e-01 2.29795098e-01
-3.56174827e-01 4.66315389e-01 -3.06142181e-01 2.78298277e-02
1.94966704e-01 -5.68135798e-01 -8.11757088e-01 -7.45091498e-01
-1.80107027e-01 8.92259419e-01 1.16452265e+00 4.93714869e-01
5.23892269e-02 8.29408824e-01 -1.78374374e+00 1.02097921e-01
4.10203934e-01 8.06183338e-01 2.02654526e-01 1.05038464e+00
6.65251240e-02 2.68206775e-01 -6.29631341e-01 -1.57243416e-01
5.42712450e-01 7.06757382e-02 -3.50940190e-02 4.52616543e-01
9.93344724e-01 -3.08094531e-01 2.03959807e-03 7.81161964e-01
3.37822348e-01 5.21392785e-02 5.41308701e-01 -1.30672455e+00
-2.07798332e-02 -1.71412781e-01 -8.37337136e-01 6.46738827e-01
3.73713434e-01 1.67796329e-01 6.38582408e-01 -9.71782267e-01
7.86005914e-01 1.24828482e+00 1.08362234e+00 3.37896824e-01
-1.46156991e+00 -4.41618532e-01 4.85951275e-01 -2.33442530e-01
-1.50886595e+00 -2.86659420e-01 4.02320653e-01 -7.25221455e-01
8.21466744e-01 4.98369902e-01 9.01490510e-01 5.89710832e-01
4.11770672e-01 6.64506137e-01 9.49130177e-01 -1.55902207e-01
9.08541009e-02 2.26007313e-01 -9.08849835e-02 7.49406636e-01
7.29986668e-01 1.13933951e-01 -5.46794474e-01 -7.08536029e-01
8.05433095e-01 1.87591702e-01 -3.25717121e-01 -1.34302509e+00
-1.25167382e+00 5.91617405e-01 7.66901597e-02 -5.48839979e-02
-1.70149878e-01 1.54609770e-01 4.44982983e-02 -4.90792453e-01
7.28310347e-01 -1.82628840e-01 -2.73011237e-01 7.36204907e-02
-1.07866371e+00 6.25199974e-01 6.46385133e-01 1.10003328e+00
4.97377485e-01 -2.70431340e-01 -2.28546560e-01 2.09252149e-01
8.04030955e-01 1.08687449e+00 -2.88477600e-01 -8.59356642e-01
3.43843281e-01 4.16880071e-01 5.17293274e-01 -8.17081153e-01
-4.08032596e-01 -3.84993792e-01 -1.97900549e-01 5.27227163e-01
7.93925643e-01 -4.37845150e-03 -8.26915622e-01 1.31786251e+00
1.19415855e+00 7.00488389e-01 -4.12084073e-01 9.08639371e-01
3.73280019e-01 3.73834997e-01 -1.91843182e-01 -2.56228596e-01
1.41079998e+00 -6.31590068e-01 -3.24980944e-01 -3.06169712e-03
2.37117618e-01 -6.87050641e-01 2.65158355e-01 5.05240381e-01
-1.10114312e+00 -4.67871159e-01 -9.94729936e-01 3.94950688e-01
-2.08381921e-01 -8.24692696e-02 3.65648687e-01 1.03236341e+00
-9.03729856e-01 5.23232996e-01 -1.36383486e+00 -4.49844152e-01
5.96014202e-01 5.49422741e-01 1.19839989e-01 1.50141478e-01
-3.21066976e-01 9.03506517e-01 1.67461097e-01 -2.63845712e-01
-8.69938910e-01 -1.14301395e+00 -5.98382473e-01 -2.62065977e-01
5.02881646e-01 -7.75631607e-01 1.13374615e+00 -2.27495357e-02
-1.20900428e+00 7.97593057e-01 -1.71677515e-01 -3.59867275e-01
8.87634635e-01 -5.38394153e-01 -1.58696517e-01 2.14004144e-01
1.14068516e-01 2.64143676e-01 7.87048697e-01 -1.53706038e+00
-1.10247838e+00 -5.76363325e-01 -2.32871756e-01 2.60638118e-01
8.87983590e-02 3.81792709e-02 -7.17214704e-01 -2.82413185e-01
2.49953076e-01 -1.13164413e+00 -1.92898124e-01 4.03586954e-01
-2.71234840e-01 6.54184744e-02 1.29052615e+00 -2.83035308e-01
7.11650968e-01 -1.84850240e+00 4.26566154e-02 1.75946966e-01
4.87052619e-01 1.21334285e-01 3.84880394e-01 -5.60132265e-02
5.88668704e-01 -2.84244686e-01 1.16120107e-01 -5.62201500e-01
-3.43209505e-02 4.32299674e-02 -1.51805341e-01 1.14459360e+00
1.58508971e-01 5.94284832e-01 -1.12797141e+00 -6.65501654e-01
8.06756079e-01 8.10216963e-01 -4.00370449e-01 -4.31048498e-02
-3.16222042e-01 4.70229089e-01 -6.05400980e-01 6.19513452e-01
1.12503111e+00 -4.36198413e-01 1.50903985e-01 -3.04017514e-02
-3.63567293e-01 -5.01724146e-02 -1.56175971e+00 1.46473098e+00
5.44393212e-02 3.80504102e-01 2.41596729e-01 -2.86564112e-01
4.63416964e-01 1.66418225e-01 1.00735891e+00 -5.62822120e-03
1.49918776e-02 1.66398846e-02 -3.22790086e-01 2.14640945e-01
4.51255471e-01 -4.35985252e-02 2.12502241e-01 2.44566619e-01
-1.20150499e-01 -4.28500362e-02 1.44533545e-01 1.51763380e-01
1.14466107e+00 6.03448212e-01 2.72864670e-01 -3.44844460e-01
3.72935385e-01 1.25046417e-01 5.67362070e-01 9.86852586e-01
-2.84542054e-01 8.11334729e-01 -2.26677686e-01 -3.96969497e-01
-9.04234111e-01 -1.63588929e+00 -3.72738689e-01 7.87659466e-01
4.79763418e-01 -3.95070136e-01 -5.84772468e-01 -7.30586410e-01
4.56185848e-01 2.21731514e-01 -5.73646665e-01 3.47154707e-01
-7.58512080e-01 -8.16653073e-01 1.61162451e-01 4.88352448e-01
5.14907762e-03 -2.78021067e-01 -1.41664231e+00 3.93919230e-01
1.51766583e-01 -1.34415495e+00 -3.71460795e-01 7.19555840e-02
-9.54555809e-01 -1.26145840e+00 -7.48375654e-01 5.55120991e-04
4.97104257e-01 6.88391089e-01 1.33352256e+00 5.67823350e-02
-5.89717090e-01 8.84119391e-01 -1.67629927e-01 -6.78920984e-01
-1.19451709e-01 -3.57719719e-01 2.95071095e-01 -4.09952402e-02
3.56855035e-01 -1.99086607e-01 -5.33940494e-01 5.45575559e-01
-2.88348585e-01 -4.00233865e-02 1.05705060e-01 2.11724699e-01
9.73515391e-01 -1.76846281e-01 -2.34979853e-01 -5.49107969e-01
-2.76811838e-01 -4.10646200e-01 -1.46061563e+00 1.39401004e-01
-1.03836067e-01 -3.10441554e-01 -3.55611332e-02 -6.65392935e-01
-9.41777110e-01 4.05255258e-01 5.92779934e-01 -8.90775323e-01
-1.78658012e-02 -3.56241018e-01 -1.70652255e-01 -4.19947892e-01
3.37305009e-01 2.72748321e-02 -2.59384125e-01 -4.89232242e-01
4.44961220e-01 7.38502741e-02 5.28818786e-01 -5.59120059e-01
1.12086093e+00 1.19397640e+00 3.18376064e-01 -7.54316092e-01
-7.71032453e-01 -9.68690276e-01 -9.58837688e-01 -6.81334257e-01
8.94937158e-01 -9.71055686e-01 -8.60017359e-01 1.50436759e-01
-1.18380308e+00 8.01120922e-02 -3.55708390e-01 7.18433380e-01
-5.52427053e-01 3.34978461e-01 -1.94146052e-01 -1.25221813e+00
-6.21504942e-03 -9.92610574e-01 1.51953888e+00 3.10893923e-01
1.91757567e-02 -9.12920654e-01 4.13614839e-01 9.82505903e-02
-1.74161233e-02 5.92767656e-01 6.45079613e-02 -4.15920168e-01
-1.30625498e+00 -1.76381528e-01 -1.14343069e-01 -3.24983805e-01
1.19770481e-03 1.61193416e-01 -1.04503810e+00 -4.84943956e-01
2.64177248e-02 3.55211765e-01 6.77088261e-01 8.08328271e-01
6.16975367e-01 3.41403753e-01 -1.19776261e+00 4.62653935e-01
1.57797992e+00 -7.69877946e-03 2.35643778e-02 2.23032236e-01
7.86914349e-01 1.55112758e-01 7.82886267e-01 7.32363999e-01
3.50965083e-01 1.00020111e+00 4.84200060e-01 1.31449312e-01
-3.67871553e-01 8.64910707e-02 3.21523488e-01 3.32196325e-01
-1.84933960e-01 -3.62691313e-01 -1.07629132e+00 4.56234872e-01
-1.90624249e+00 -1.04880881e+00 -6.07839048e-01 2.54115105e+00
2.55906641e-01 3.17520469e-01 4.13516611e-01 -3.31411421e-01
8.40566516e-01 -8.95106643e-02 -5.34781039e-01 5.01160860e-01
1.17065407e-01 -1.01537652e-01 9.43894267e-01 5.22471011e-01
-1.49949801e+00 5.40210187e-01 6.58515310e+00 3.49299312e-01
-6.61290050e-01 4.59313989e-01 -1.03513792e-01 -6.03701174e-01
2.15462565e-01 -1.17133230e-01 -1.62984407e+00 5.39383829e-01
7.95187593e-01 -7.43280128e-02 -1.17591463e-01 6.63579583e-01
3.85009386e-02 -4.72461969e-01 -1.15997899e+00 8.93254340e-01
1.57690108e-01 -1.31724370e+00 -3.68052423e-01 5.35227895e-01
6.29876077e-01 4.19305533e-01 7.15115219e-02 -1.57468894e-03
5.15790820e-01 -4.10514623e-01 8.76804233e-01 4.94092166e-01
3.32370311e-01 -5.55158138e-01 3.64021063e-01 4.37390596e-01
-1.94493222e+00 3.30893666e-01 -4.00565505e-01 1.80578187e-01
5.54711401e-01 7.01676846e-01 -9.48336244e-01 6.52765632e-01
8.75195205e-01 5.88894546e-01 -6.04594171e-01 1.54400134e+00
4.09314513e-01 1.73397839e-01 -1.05016267e+00 2.26946309e-01
-1.32144779e-01 -8.89539495e-02 1.04904544e+00 1.22856259e+00
3.55943829e-01 -1.75271109e-02 6.38951182e-01 8.58252048e-01
3.06585163e-01 -1.64127573e-01 -6.21394575e-01 4.16979194e-01
3.66344601e-01 1.26644564e+00 -1.31216836e+00 -4.55697238e-01
-7.03189313e-01 4.77384090e-01 -1.46948770e-02 -5.41841425e-02
-1.21269989e+00 3.08687806e-01 7.03320384e-01 4.91963953e-01
9.21821952e-01 -2.92654514e-01 7.97735527e-02 -1.18320668e+00
2.61963189e-01 -3.32062989e-01 3.10672432e-01 -4.24844414e-01
-1.17703223e+00 2.26571053e-01 4.67559397e-01 -1.45419145e+00
6.43129796e-02 -6.16878808e-01 -3.01671654e-01 6.66254580e-01
-1.31345510e+00 -1.32534385e+00 -2.92660177e-01 4.34772938e-01
3.58266354e-01 1.51032254e-01 5.28370500e-01 3.72826248e-01
-1.08232804e-01 -4.73817289e-02 3.26747864e-01 -2.66796529e-01
4.24769461e-01 -1.38395512e+00 4.28852439e-01 8.69127929e-01
4.03849542e-01 2.95022190e-01 8.57886791e-01 -1.10780430e+00
-1.49921572e+00 -1.18091285e+00 3.17713857e-01 -1.37399197e+00
4.71236974e-01 -7.97765911e-01 -9.34761405e-01 6.98078513e-01
-1.15424044e-01 6.39476418e-01 7.01817393e-01 6.03235178e-02
-8.19639266e-02 3.09767842e-01 -1.08525074e+00 1.80888504e-01
1.18480456e+00 -5.28761186e-02 -5.14089525e-01 5.28294861e-01
3.35292846e-01 -9.43485200e-01 -7.92771578e-01 4.48635072e-01
6.55871511e-01 -1.05491912e+00 1.31431723e+00 -1.12424724e-01
-6.28962755e-01 -8.69776130e-01 -1.85376137e-01 -8.92048299e-01
-2.26483494e-01 -4.84443635e-01 -7.08059072e-01 1.02807462e+00
-1.11267149e-01 -3.26259226e-01 1.08352590e+00 4.15621400e-01
3.95100862e-02 -5.11310875e-01 -1.14643288e+00 -1.02615118e+00
-2.72379935e-01 -4.80355442e-01 2.61128604e-01 4.65894103e-01
-5.17994463e-01 -1.24532834e-01 -1.36944592e-01 8.44700158e-01
1.48740840e+00 2.84805506e-01 1.05731654e+00 -1.78184032e+00
-3.18879932e-01 -1.26089349e-01 -6.02727354e-01 -1.03513396e+00
-1.38569206e-01 -6.54383481e-01 2.15916529e-01 -1.25086153e+00
5.88999510e-01 -4.04646814e-01 -5.66370189e-02 1.99204367e-02
-2.42700666e-01 3.17242056e-01 5.91138661e-01 1.79315850e-01
-1.12811732e+00 2.95818031e-01 8.84191155e-01 4.17269655e-02
3.35863512e-03 8.83293599e-02 2.79627740e-02 1.02489984e+00
3.01323354e-01 -7.82401025e-01 -1.20904222e-01 -3.05286139e-01
-2.07276270e-01 -2.42940292e-01 6.45440996e-01 -1.45020628e+00
2.22329944e-01 -7.02264309e-02 8.40622425e-01 -1.39864600e+00
8.40382576e-01 -1.09684074e+00 4.59465623e-01 3.86877745e-01
4.55965817e-01 6.37854710e-02 5.43095887e-01 9.61195290e-01
4.26551014e-01 8.85392725e-02 9.30262327e-01 -9.04598311e-02
-3.76998991e-01 4.10111278e-01 -1.73125386e-01 4.84803878e-02
1.58322763e+00 -5.10316432e-01 -9.56381187e-02 1.48495048e-01
-7.18768597e-01 3.07309508e-01 9.26172614e-01 3.97527337e-01
5.20912707e-01 -1.30305207e+00 -5.79473317e-01 1.56090915e-01
-7.98702538e-02 -2.35075094e-02 3.22862412e-04 9.49195564e-01
-4.20986146e-01 5.30073464e-01 -1.80995278e-02 -1.44094193e+00
-1.78346872e+00 5.87979853e-01 4.86914635e-01 -3.42125326e-01
-1.03582251e+00 7.81973839e-01 4.13673013e-01 -1.63338229e-01
1.12016827e-01 -4.79603916e-01 2.33154386e-01 -1.22649528e-01
7.14194894e-01 5.24346948e-01 -4.41666320e-02 -8.09467018e-01
-7.60364413e-01 8.59822690e-01 -3.87899578e-02 -2.72002488e-01
1.09462965e+00 -2.69725114e-01 5.20581126e-01 4.93602157e-01
6.27248645e-01 3.12268078e-01 -1.84634864e+00 -2.43827775e-01
3.16763557e-02 -1.06223357e+00 3.78048094e-03 -3.08257818e-01
-9.09408629e-01 6.90300763e-01 1.00189114e+00 1.93292007e-01
2.93419480e-01 3.12099159e-01 4.04535800e-01 -3.88027318e-02
7.62682319e-01 -6.96970522e-01 -1.09906428e-01 2.97527790e-01
2.34324694e-01 -1.12322426e+00 5.59394360e-01 -6.91358089e-01
-1.68769628e-01 8.11265111e-01 6.73463523e-01 -1.18332267e-01
6.91796064e-01 6.77015603e-01 -8.48869234e-02 -3.53524745e-01
-5.73282182e-01 -3.57151061e-01 3.77296418e-01 8.81497681e-01
3.90079990e-02 -1.06088541e-01 3.85004014e-01 1.32731358e-02
3.01613152e-01 -1.57661229e-01 2.46488005e-01 1.34630561e+00
-6.16739094e-01 -9.32044029e-01 -1.07878840e+00 4.59166020e-01
-4.27559018e-01 3.97302508e-01 -3.84267196e-02 9.63125527e-01
3.66112798e-01 7.21797228e-01 4.95367199e-01 2.54092932e-01
3.70614588e-01 -7.28643984e-02 1.01050663e+00 -8.52988243e-01
-3.91917497e-01 4.45896804e-01 -2.27798268e-01 -7.73124337e-01
-7.46089935e-01 -1.28373909e+00 -1.20223284e+00 -6.11910261e-02
-7.14267612e-01 -1.35882437e-01 6.21795714e-01 7.50050008e-01
3.68992418e-01 4.91808027e-01 -4.59045405e-03 -1.59451759e+00
-3.21609408e-01 -5.13346255e-01 -5.16915858e-01 3.06633085e-01
4.67580587e-01 -1.39431798e+00 -1.92469925e-01 1.99468106e-01] | [6.65265417098999, -2.1780431270599365] |
7a974a8b-9ca6-445c-b2a4-297365893203 | qa-is-the-new-kr-question-answer-pairs-as | 2207.00630 | null | https://arxiv.org/abs/2207.00630v1 | https://arxiv.org/pdf/2207.00630v1.pdf | QA Is the New KR: Question-Answer Pairs as Knowledge Bases | In this position paper, we propose a new approach to generating a type of knowledge base (KB) from text, based on question generation and entity linking. We argue that the proposed type of KB has many of the key advantages of a traditional symbolic KB: in particular, it consists of small modular components, which can be combined compositionally to answer complex queries, including relational queries and queries involving "multi-hop" inferences. However, unlike a traditional KB, this information store is well-aligned with common user information needs. | ['John Wieting', 'Pat Verga', 'Alessandro Presta', 'Nitish Gupta', 'Michiel de Jong', 'William W. Cohen', 'Wenhu Chen'] | 2022-07-01 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [-2.68754750e-01 1.07036805e+00 -2.85328209e-01 -3.24223101e-01
-1.11846828e+00 -6.05237007e-01 5.84685683e-01 7.66592562e-01
-2.46782139e-01 1.21795428e+00 2.06524432e-01 -5.53180814e-01
-4.97788221e-01 -1.28431571e+00 -6.66206121e-01 3.24309647e-01
2.47104853e-01 8.69245291e-01 1.10109746e+00 -9.40542161e-01
2.87194788e-01 4.95478332e-01 -1.51676428e+00 4.13898140e-01
1.03303075e+00 1.06669462e+00 -7.69316852e-02 3.57896924e-01
-7.59597123e-01 1.52649283e+00 -6.59263611e-01 -1.00914001e+00
-5.12511015e-01 -1.51664719e-01 -1.52961278e+00 -6.96183681e-01
2.52683252e-01 -1.52349263e-01 -4.38359022e-01 8.46003354e-01
4.40140486e-01 4.36013527e-02 4.79980677e-01 -1.22712100e+00
-6.42538190e-01 1.03076184e+00 3.60805988e-01 4.01940383e-02
1.21110964e+00 -5.90917945e-01 1.26339447e+00 -8.57366383e-01
9.10261571e-01 1.29310656e+00 7.27132797e-01 2.87625343e-01
-1.06288743e+00 -1.23202316e-01 -9.39909294e-02 5.54156303e-01
-1.58656955e+00 -5.69200695e-01 3.43637049e-01 -1.84633613e-01
1.29249597e+00 7.68278241e-01 5.96136153e-01 5.13627231e-01
-1.68017179e-01 5.07264972e-01 7.23390043e-01 -6.85026526e-01
1.54151574e-01 8.53074253e-01 2.91820586e-01 8.17780375e-01
5.71806490e-01 -6.25308156e-01 -6.18623793e-01 -6.98550522e-01
6.09420061e-01 -6.35130107e-01 -2.47973502e-01 -2.78293252e-01
-8.83543968e-01 7.57613122e-01 5.85216545e-02 3.87781382e-01
-2.04012260e-01 1.05940044e-01 2.85057873e-01 2.97726452e-01
2.12210879e-01 8.01437318e-01 -6.84298277e-01 -6.38854578e-02
-4.33588982e-01 6.59122288e-01 1.67162144e+00 1.46861458e+00
9.87418771e-01 -4.16152805e-01 -2.56355464e-01 5.05113244e-01
5.25681674e-01 5.38630962e-01 3.73238117e-01 -1.02552128e+00
6.04154229e-01 1.06316066e+00 4.88682538e-01 -1.23324192e+00
-1.82363033e-01 7.79756904e-02 -1.65623173e-01 -4.77132916e-01
1.26064703e-01 1.78591888e-02 -7.93717802e-02 1.34482634e+00
4.58848119e-01 -3.57801199e-01 5.62200189e-01 3.55885714e-01
1.53056145e+00 4.22031611e-01 -1.76414680e-02 -1.71983749e-01
1.71880960e+00 -4.99698222e-01 -1.11923587e+00 -5.89628480e-02
7.12642014e-01 -5.05622208e-01 8.68628085e-01 -6.73916638e-02
-1.29978168e+00 -3.91773224e-01 -8.64587545e-01 -5.36739588e-01
-1.05909157e+00 -1.92488655e-01 7.09771037e-01 6.15773559e-01
-1.22492206e+00 -1.00058675e-01 -1.30961388e-01 -4.23423380e-01
-2.83804685e-01 1.51764274e-01 -2.15419248e-01 2.43151203e-01
-1.82824349e+00 1.21693206e+00 8.93826127e-01 -2.33890861e-02
-2.07548708e-01 -5.48533559e-01 -9.37785089e-01 1.07275434e-01
9.60232973e-01 -9.57954526e-01 1.43475032e+00 -7.82462582e-02
-1.40316796e+00 6.88920736e-01 -3.41922879e-01 -3.48924667e-01
1.59614891e-01 -4.00977880e-01 -6.34662390e-01 5.03568172e-01
2.58853078e-01 3.18194956e-01 2.64908344e-01 -1.14002120e+00
-4.84808117e-01 -1.97645038e-01 7.00513542e-01 9.47140828e-02
-1.18227087e-01 4.42347020e-01 -6.27531946e-01 -4.50475633e-01
7.25124404e-02 -4.09704804e-01 -7.69694671e-02 -2.76781350e-01
-6.22393787e-01 -4.91354674e-01 5.36932707e-01 -7.23059416e-01
1.72062492e+00 -1.51168740e+00 3.10114529e-02 4.32772875e-01
2.80410230e-01 3.74284834e-02 4.83301103e-01 1.09143579e+00
2.36381635e-01 4.38751340e-01 7.68455788e-02 3.37831229e-01
4.11330253e-01 2.99126923e-01 -8.47315073e-01 -4.41214263e-01
3.68949771e-01 1.38933408e+00 -9.03869629e-01 -1.18312275e+00
-4.52499360e-01 3.25109214e-02 -3.66012216e-01 1.79021105e-01
-8.20516706e-01 -1.45014614e-01 -6.94201887e-01 6.84528828e-01
1.95880443e-01 -3.15112531e-01 4.87801105e-01 -2.82472998e-01
1.07946470e-01 6.85288846e-01 -1.41706467e+00 1.31963336e+00
-6.00653999e-02 2.57540882e-01 -2.88002007e-02 -5.84075153e-01
9.99832273e-01 6.82849646e-01 1.51778817e-01 -5.83708704e-01
-3.09616238e-01 2.77006656e-01 -7.92922974e-01 -9.55713153e-01
1.03583837e+00 -5.02520837e-02 -4.69087929e-01 5.22074640e-01
7.56607875e-02 -5.87564051e-01 3.98522764e-01 6.54258370e-01
9.83270526e-01 2.23335341e-01 5.23070633e-01 -9.13671628e-02
8.33155990e-01 3.80257308e-01 3.44797701e-01 8.10768843e-01
4.45962995e-01 -2.46121064e-02 7.53630161e-01 -1.03112943e-01
-7.30559468e-01 -9.77046192e-01 -3.89678441e-02 9.48854089e-01
3.12092155e-01 -9.02188599e-01 -5.88397086e-01 -5.85663140e-01
1.14697263e-01 7.54603744e-01 -2.25830153e-01 6.58149794e-02
-6.30288303e-01 -1.42603457e-01 1.09532857e+00 4.87864017e-01
2.96461612e-01 -1.14720678e+00 -5.01391888e-01 2.58869052e-01
-7.13464916e-01 -1.26570034e+00 1.76909808e-02 -1.78470507e-01
-7.86665916e-01 -1.30780685e+00 -8.77468064e-02 -6.87037230e-01
3.92201513e-01 9.78501141e-02 1.61540020e+00 3.46674263e-01
6.15446679e-02 6.88656867e-01 -6.02442086e-01 -4.67298985e-01
-4.94446546e-01 3.31476808e-01 -2.61922508e-01 -2.99298108e-01
2.45099321e-01 -3.06137651e-01 -1.85552076e-01 1.39294848e-01
-1.25979245e+00 -1.33442864e-01 5.07146657e-01 4.47751254e-01
5.28345346e-01 1.05774917e-01 9.73716974e-01 -1.30246198e+00
1.10028934e+00 -6.59949064e-01 -5.41006327e-01 9.76210475e-01
-8.74290109e-01 2.09251300e-01 2.69093931e-01 -6.07114956e-02
-1.41424775e+00 -4.99022722e-01 -1.49951994e-01 4.08179879e-01
9.50276926e-02 1.08335614e+00 -2.80505091e-01 -5.79362884e-02
7.54313767e-01 1.66776583e-01 -1.55438438e-01 -5.52616417e-01
9.22529697e-01 5.28250813e-01 2.93095976e-01 -1.00570381e+00
7.73522973e-01 1.84208632e-01 -1.21764675e-01 -5.60721874e-01
-5.00825703e-01 -4.22317982e-01 -7.24515855e-01 -1.46732450e-01
5.81930935e-01 -7.53490269e-01 -9.20276165e-01 -1.23637907e-01
-1.26501441e+00 -2.69041490e-03 -7.27181196e-01 1.80095375e-01
-6.90686226e-01 5.76246500e-01 -6.49467111e-01 -9.34793055e-01
-3.29575002e-01 -4.56623644e-01 8.65185618e-01 1.52200237e-01
-5.17785668e-01 -9.51751232e-01 6.85434267e-02 5.73053658e-01
4.72890556e-01 3.11645772e-02 1.43789804e+00 -8.72550011e-01
-7.24483669e-01 -3.67155343e-01 -4.00861800e-01 -3.96174155e-02
-1.35927647e-01 -2.78227311e-02 -5.97276747e-01 2.28177756e-01
-1.73615113e-01 -6.64492428e-01 8.43411386e-02 -5.52195132e-01
6.95826828e-01 -7.31750190e-01 -4.23139960e-01 -9.84443948e-02
1.42192173e+00 2.66119480e-01 1.00937068e+00 3.90748739e-01
2.95289129e-01 9.01707470e-01 6.71980679e-01 1.65447339e-01
1.16969752e+00 7.85898805e-01 8.18812624e-02 2.00926915e-01
2.14332603e-02 -5.16052127e-01 2.97715310e-02 8.98255527e-01
2.88619325e-02 1.76645830e-01 -9.62690771e-01 5.08905172e-01
-2.10616899e+00 -9.97686088e-01 -2.18566105e-01 1.76036668e+00
1.44325924e+00 -7.13865235e-02 1.20720096e-01 5.77982739e-02
3.86324614e-01 -2.13574708e-01 -2.29816258e-01 -2.29251742e-01
-4.14033324e-01 4.11825955e-01 2.12351561e-01 6.33808196e-01
-3.47147107e-01 1.00107920e+00 8.17153549e+00 5.76334059e-01
-5.72456479e-01 1.02324925e-01 -9.52277705e-02 3.29355687e-01
-8.61567974e-01 3.05152506e-01 -1.14272618e+00 1.08606704e-02
1.03710341e+00 -6.35784984e-01 1.41274974e-01 7.14616537e-01
-3.76876682e-01 -1.41571850e-01 -1.02700579e+00 4.38800454e-01
8.97456706e-02 -1.63558531e+00 3.09536815e-01 -2.66559511e-01
7.59612992e-02 -5.28356791e-01 -5.13100982e-01 5.19803643e-01
5.62153995e-01 -7.89049685e-01 8.74445856e-01 1.12167823e+00
5.54841876e-01 -5.93880355e-01 8.61144602e-01 4.73980874e-01
-1.17054582e+00 -3.82316075e-02 -3.03710073e-01 1.70609131e-01
1.89301893e-01 5.30422986e-01 -9.42359209e-01 1.23410547e+00
7.29891181e-01 -2.34356999e-01 -8.86642814e-01 7.98633039e-01
-3.61187458e-01 1.97457060e-01 -1.03985228e-01 -3.84201080e-01
-3.80718231e-01 -1.24841839e-01 2.27185994e-01 1.27260256e+00
3.13986748e-01 3.61987323e-01 -1.32175744e-01 9.33240950e-01
-4.91183810e-03 4.13884044e-01 -7.85482228e-01 -4.33401525e-04
9.17372227e-01 1.01995718e+00 -3.38690370e-01 -9.38478827e-01
-4.65999007e-01 4.59164143e-01 5.38881600e-01 4.02256161e-01
-5.71683586e-01 -8.80660236e-01 -9.59839001e-02 2.23372743e-01
3.03684026e-01 -8.96244347e-02 8.90959278e-02 -1.15978515e+00
5.29407918e-01 -9.55162823e-01 6.69965088e-01 -1.10140800e+00
-1.22978842e+00 5.46215653e-01 5.27624607e-01 -7.01389492e-01
-5.89398384e-01 -3.01211715e-01 -9.32512283e-02 8.03504884e-01
-1.51169682e+00 -1.22692704e+00 -1.87601373e-01 6.94527149e-01
-2.88678199e-01 6.73048049e-02 1.25360107e+00 4.21419263e-01
-2.70878226e-01 4.49807674e-01 -4.87151444e-01 1.79173589e-01
5.62111139e-01 -1.33493507e+00 8.18243772e-02 4.63601112e-01
6.75938502e-02 1.05237567e+00 4.78665888e-01 -8.74559402e-01
-1.74118698e+00 -9.37901139e-01 1.52366149e+00 -7.04314768e-01
9.40202475e-01 -2.42096096e-01 -1.19263601e+00 9.10411417e-01
1.56874850e-01 -3.79738718e-01 9.22725260e-01 1.26809150e-01
-5.82301855e-01 -4.10282820e-01 -1.19706631e+00 4.78495300e-01
8.42293799e-01 -1.01311266e+00 -1.25438488e+00 4.71395284e-01
1.15105498e+00 -4.30509508e-01 -1.41667557e+00 4.70167071e-01
4.53514516e-01 -8.01699102e-01 1.04900920e+00 -7.35273123e-01
-1.19795524e-01 -5.30363679e-01 -3.35132897e-01 -5.45757890e-01
-1.12420157e-01 -7.84344375e-01 -6.54052138e-01 1.41749668e+00
8.04433644e-01 -1.07377172e+00 6.37791455e-01 1.04929006e+00
-1.84927564e-02 -6.84387803e-01 -1.01520443e+00 -6.14801824e-01
-2.47801483e-01 -4.42922860e-01 9.53441739e-01 8.81613910e-01
6.73266232e-01 6.50978923e-01 3.54001857e-02 1.82327151e-01
1.50436342e-01 1.14477426e-01 7.81905055e-01 -1.40762377e+00
-3.95837396e-01 -1.67366952e-01 -1.73434839e-01 -1.07603741e+00
4.67144251e-02 -7.78040946e-01 -1.24390341e-01 -1.94186795e+00
-1.16555318e-01 -8.00121486e-01 1.68532789e-01 4.80109185e-01
-6.81479052e-02 -1.14139564e-01 -1.81404874e-01 2.86381632e-01
-7.73767471e-01 3.10649842e-01 9.66752529e-01 2.13312660e-03
-4.78960872e-02 -1.27520546e-01 -1.17186284e+00 4.22507524e-01
4.83628839e-01 -4.43578213e-01 -5.90733230e-01 -1.46942466e-01
1.31327462e+00 4.86925244e-01 2.72451073e-01 -5.38274884e-01
7.73174405e-01 -2.71083176e-01 -2.90977776e-01 -8.03601980e-01
5.08825898e-01 -6.44122183e-01 4.13019985e-01 2.54135281e-01
-3.01876962e-01 3.58477026e-01 6.98301271e-02 3.86620045e-01
-6.74371243e-01 -7.01734483e-01 -8.14592093e-02 -3.41246992e-01
-7.48496532e-01 -1.96107209e-01 -4.37347114e-01 3.23088735e-01
9.39850926e-01 -6.55011460e-02 -6.89251184e-01 -5.11339307e-01
-4.53684896e-01 2.61536986e-01 2.05904812e-01 1.64931893e-01
7.45995760e-01 -1.34576356e+00 -4.50809330e-01 -1.84029624e-01
3.67827475e-01 2.11479917e-01 -1.80969685e-01 6.06593132e-01
-7.27527380e-01 1.01375544e+00 1.46582946e-01 1.02744460e-01
-7.28590727e-01 6.13360405e-01 5.47385700e-02 -3.73016924e-01
-4.71849650e-01 4.27275270e-01 -6.29318118e-01 -7.01684415e-01
2.98297226e-01 -3.35007399e-01 -5.10032356e-01 2.78973758e-01
4.93147820e-01 3.92300427e-01 1.98199257e-01 -4.24202442e-01
-2.85391480e-01 3.20062608e-01 2.84951091e-01 -3.61042559e-01
9.93370414e-01 -2.33926147e-01 -9.95671570e-01 6.20342135e-01
4.62347656e-01 5.10163009e-01 1.24829426e-01 -5.66594958e-01
6.18750393e-01 -1.30820349e-01 -5.57620764e-01 -7.74648368e-01
-3.12385499e-01 2.40938708e-01 -2.93755025e-01 9.09878731e-01
8.12142909e-01 5.57577133e-01 7.69432843e-01 1.15967607e+00
6.29278004e-01 -1.16920626e+00 4.11550291e-02 5.72088480e-01
1.17451322e+00 -7.35206842e-01 1.06946617e-01 -8.17441404e-01
-3.61637086e-01 1.16161799e+00 5.89526117e-01 5.18501341e-01
5.42226195e-01 2.65198708e-01 6.26561865e-02 -7.01610088e-01
-9.17182326e-01 -3.45844060e-01 1.97639987e-01 7.40964592e-01
2.94481874e-01 -1.93068787e-01 -3.43323886e-01 6.69819891e-01
-4.27145690e-01 2.63983518e-01 4.40042377e-01 1.24003983e+00
-7.10066199e-01 -1.36665785e+00 -3.42328727e-01 3.00265789e-01
-5.02610326e-01 -2.40539402e-01 -7.28344083e-01 1.06541741e+00
-6.04009144e-02 9.88539159e-01 -3.56570363e-01 -3.20835412e-01
7.77543128e-01 4.82532501e-01 4.12000746e-01 -8.81377935e-01
-5.92648864e-01 -6.82791471e-01 9.21613336e-01 -3.57387930e-01
-4.52690899e-01 -2.45341793e-01 -1.31720650e+00 -2.41907880e-01
-4.74388480e-01 7.55944967e-01 5.01074612e-01 8.80132377e-01
6.08888209e-01 1.49529994e-01 2.05839992e-01 1.95980862e-01
-5.80862999e-01 -9.14776981e-01 -3.67186666e-01 2.51146853e-01
-1.85693026e-01 -5.06678224e-01 7.76794553e-02 8.96317512e-02] | [10.322897911071777, 7.937964916229248] |
1db8b398-7ab3-46e5-84c4-a3a66a928d92 | learning-music-audio-representations-via-weak | 2112.04214 | null | https://arxiv.org/abs/2112.04214v2 | https://arxiv.org/pdf/2112.04214v2.pdf | Learning music audio representations via weak language supervision | Audio representations for music information retrieval are typically learned via supervised learning in a task-specific fashion. Although effective at producing state-of-the-art results, this scheme lacks flexibility with respect to the range of applications a model can have and requires extensively annotated datasets. In this work, we pose the question of whether it may be possible to exploit weakly aligned text as the only supervisory signal to learn general-purpose music audio representations. To address this question, we design a multimodal architecture for music and language pre-training (MuLaP) optimised via a set of proxy tasks. Weak supervision is provided in the form of noisy natural language descriptions conveying the overall musical content of the track. After pre-training, we transfer the audio backbone of the model to a set of music audio classification and regression tasks. We demonstrate the usefulness of our approach by comparing the performance of audio representations produced by the same audio backbone with different training strategies and show that our pre-training method consistently achieves comparable or higher scores on all tasks and datasets considered. Our experiments also confirm that MuLaP effectively leverages audio-caption pairs to learn representations that are competitive with audio-only and cross-modal self-supervised methods in the literature. | ['Gyorgy Fazekas', 'Elio Quinton', 'Emmanouil Benetos', 'Ilaria Manco'] | 2021-12-08 | null | null | null | null | ['music-information-retrieval'] | ['music'] | [ 5.06144047e-01 1.14457399e-01 -2.05203503e-01 -2.55068362e-01
-1.60221660e+00 -8.36041510e-01 7.28669524e-01 9.56022069e-02
-3.91337276e-01 3.86034727e-01 5.02088547e-01 1.86155841e-01
-2.89814800e-01 -3.90386939e-01 -9.14676428e-01 -5.08997202e-01
-1.38157338e-01 4.51242715e-01 -1.43952593e-01 -3.27016681e-01
9.45376977e-02 2.85944771e-02 -1.84861016e+00 8.70544553e-01
8.51225629e-02 1.01760650e+00 1.24138212e-02 6.83318198e-01
-3.47404629e-02 7.10359752e-01 -5.95968783e-01 -1.49469167e-01
1.99822128e-01 -4.89648432e-01 -9.32614803e-01 4.92441133e-02
7.37017989e-01 3.45744863e-02 -1.64932996e-01 5.78016162e-01
6.92324877e-01 3.31877694e-02 5.85973442e-01 -1.11379623e+00
-2.60372519e-01 1.05689609e+00 -3.52942467e-01 -3.52848694e-02
5.90885520e-01 -3.19739571e-04 1.65465915e+00 -7.38041043e-01
5.19981802e-01 1.15824187e+00 7.61343837e-01 5.79954267e-01
-1.72046006e+00 -6.92108810e-01 8.46643671e-02 -4.99543771e-02
-1.31998098e+00 -8.69848073e-01 9.17097390e-01 -3.94172400e-01
7.07565904e-01 3.89668912e-01 2.92651713e-01 1.42065740e+00
-4.78924572e-01 9.27267969e-01 7.71449983e-01 -6.99756265e-01
-6.16931356e-03 1.55601054e-01 -2.10109755e-01 3.16343129e-01
-4.16284680e-01 7.62465745e-02 -1.34070504e+00 -3.57819647e-01
4.53957736e-01 -5.57067573e-01 -2.29846835e-01 -4.17802751e-01
-1.56270492e+00 6.76302671e-01 3.16304147e-01 5.67902207e-01
-2.54109383e-01 4.33028102e-01 9.58684921e-01 6.21168256e-01
2.44560659e-01 7.21775055e-01 -4.32078183e-01 -2.31850505e-01
-1.31222415e+00 2.96497613e-01 6.43303514e-01 7.33503401e-01
4.31937635e-01 2.14700758e-01 -1.27132222e-01 1.11006415e+00
2.24073648e-01 2.65435606e-01 6.09417856e-01 -1.05357766e+00
5.26568234e-01 1.05076909e-01 -1.72095269e-01 -4.54233319e-01
-1.41851604e-01 -7.03526378e-01 -4.56011593e-01 -6.97625056e-02
3.83819938e-01 6.33029193e-02 -4.07125741e-01 1.92600203e+00
1.17051499e-02 2.58499891e-01 1.84994549e-01 9.99553204e-01
7.15714514e-01 5.68652928e-01 1.07151262e-01 1.17237363e-02
1.22095191e+00 -9.26350892e-01 -4.46410686e-01 -1.16607577e-01
4.92102861e-01 -1.02975261e+00 1.33324909e+00 6.16959691e-01
-1.05637074e+00 -7.70274818e-01 -1.03087771e+00 3.60629112e-02
3.13190483e-02 4.12702173e-01 3.72878194e-01 3.79176766e-01
-9.04059410e-01 6.19708836e-01 -5.22005856e-01 -2.51613349e-01
1.14826843e-01 3.61252546e-01 -4.16771829e-01 2.18442172e-01
-9.96069014e-01 3.52767825e-01 4.87677634e-01 -1.48740456e-01
-1.09179652e+00 -7.68463254e-01 -7.35906661e-01 -7.46739749e-03
2.75561571e-01 -6.45737112e-01 1.57181787e+00 -1.43782878e+00
-1.47988260e+00 9.32969093e-01 1.44462243e-01 -5.31925499e-01
2.79969752e-01 -4.44595397e-01 -2.33772933e-01 2.93699026e-01
2.38460690e-01 1.03481293e+00 1.04042411e+00 -1.33546770e+00
-4.31934386e-01 -1.17667042e-01 -8.16563144e-02 1.58171445e-01
-4.65922982e-01 1.79871410e-01 -3.87294441e-01 -1.00210011e+00
-1.01878352e-01 -1.14120448e+00 1.19755752e-01 -2.06977472e-01
-3.91759276e-01 -6.62593618e-02 5.77203453e-01 -4.14952308e-01
1.12949240e+00 -2.65623617e+00 2.61325419e-01 1.91443652e-01
-5.37518859e-01 -1.39328003e-01 -5.60559452e-01 7.18629777e-01
-2.89878577e-01 -1.42821074e-01 -3.42928469e-01 -6.11525059e-01
2.53996730e-01 1.21718705e-01 -8.66136312e-01 1.75430685e-01
3.98245007e-01 6.79510057e-01 -8.85366619e-01 -3.92796636e-01
-1.58705160e-01 5.97438931e-01 -6.82574868e-01 2.95140922e-01
-5.29621005e-01 6.57837391e-01 -1.85673222e-01 5.31931758e-01
-1.87659949e-01 -3.08769047e-02 2.64527947e-01 -1.50079086e-01
5.09251244e-02 8.43287468e-01 -1.23795879e+00 2.35956573e+00
-4.93726969e-01 7.51249015e-01 1.43056825e-01 -1.01925790e+00
8.78814697e-01 8.12325060e-01 6.42192364e-01 -4.34933931e-01
-1.46501362e-01 2.73089767e-01 -8.25577751e-02 -3.60618323e-01
3.96808982e-01 -4.45387572e-01 -2.75965899e-01 7.72561610e-01
5.22152185e-01 -7.72386938e-02 -7.35009387e-02 4.22215611e-02
1.10209012e+00 4.62972224e-01 -8.55488256e-02 8.47991109e-02
5.16327143e-01 -1.91951171e-02 1.45960733e-01 7.80123472e-01
1.63776368e-01 9.11114395e-01 2.32149318e-01 -7.66254738e-02
-1.02020216e+00 -9.92866158e-01 -1.82106644e-01 1.90913045e+00
-5.02108812e-01 -9.19816792e-01 -4.93575692e-01 -4.74462330e-01
-1.07858479e-02 4.92712319e-01 -5.07422447e-01 -8.13251082e-03
-4.97489780e-01 -3.36398154e-01 1.07216871e+00 4.06607032e-01
-3.54458159e-03 -1.28097928e+00 -3.92574370e-01 3.57294858e-01
-2.87129700e-01 -1.05004108e+00 -3.71176302e-01 4.08133060e-01
-8.82910848e-01 -8.88127983e-01 -6.84260309e-01 -6.57636762e-01
2.17783265e-02 1.45652741e-01 1.21259642e+00 -2.17783421e-01
-1.30483598e-01 8.21742892e-01 -4.72802728e-01 -4.40846533e-01
-5.48416555e-01 4.59538698e-01 2.16632560e-02 2.47449443e-01
5.85854724e-02 -8.63310754e-01 -3.01452547e-01 1.37153029e-01
-1.02331662e+00 -2.08950505e-01 6.80129886e-01 7.44086862e-01
6.08466566e-01 -2.29708254e-01 8.38743269e-01 -7.01229930e-01
7.00242877e-01 -4.26035613e-01 -8.27105194e-02 -8.14079121e-02
-2.40613461e-01 1.88534498e-01 3.72928590e-01 -7.15538621e-01
-5.90446353e-01 4.19802368e-01 -4.39683087e-02 -6.62378430e-01
-3.56466651e-01 6.59971416e-01 2.16277353e-02 1.88012794e-01
9.39394891e-01 1.22094482e-01 1.02473013e-02 -8.46061647e-01
5.24652302e-01 8.41506779e-01 8.47575009e-01 -9.24050629e-01
7.69957900e-01 3.87933373e-01 -9.62793604e-02 -8.21709454e-01
-1.14660048e+00 -7.20352232e-01 -5.71252108e-01 -2.97840804e-01
5.88540018e-01 -1.24328208e+00 -3.81284386e-01 -1.93448797e-01
-8.86178017e-01 -2.66678423e-01 -5.64193487e-01 4.71063226e-01
-9.86243129e-01 1.67328909e-01 -5.28038263e-01 -9.67030644e-01
-3.41454208e-01 -7.69321442e-01 1.66992867e+00 -5.96155643e-01
-6.85732722e-01 -6.31513059e-01 4.50530738e-01 5.46222389e-01
1.66311190e-01 1.32984787e-01 8.32701087e-01 -7.21766055e-01
-4.60344285e-01 -4.45897244e-02 1.78048819e-01 3.49438787e-01
-8.66964087e-02 -2.39559144e-01 -1.66796505e+00 -3.37059617e-01
-3.88920337e-01 -9.05020297e-01 1.05168021e+00 -1.48815766e-03
1.09146273e+00 -2.28773549e-01 1.90583080e-01 3.48902136e-01
1.16835260e+00 -5.06311119e-01 3.45354766e-01 5.78080952e-01
4.22217488e-01 8.40766370e-01 5.48648715e-01 3.09613377e-01
1.56150945e-02 1.08777499e+00 3.19630682e-01 1.01428308e-01
-2.37680539e-01 -5.61655521e-01 7.72418976e-01 8.53645205e-01
1.28515720e-01 2.87752271e-01 -8.32219660e-01 6.60572529e-01
-1.96516454e+00 -1.04575312e+00 2.22739130e-01 2.16125250e+00
1.20433474e+00 6.04511648e-02 6.50128007e-01 7.10713506e-01
3.59890699e-01 1.62933171e-01 -2.05229595e-01 -1.36740938e-01
-4.52919677e-02 3.16401184e-01 -8.02149475e-02 1.89870954e-01
-1.30136895e+00 7.15623379e-01 6.44773388e+00 7.14197636e-01
-1.16866934e+00 1.36063129e-01 -1.10902321e-02 -4.59305495e-01
-2.49445438e-01 -1.29934177e-01 -3.36140424e-01 6.28404021e-02
1.29086518e+00 2.20358416e-01 5.42403996e-01 5.87971985e-01
9.74198207e-02 4.76865888e-01 -1.66819656e+00 1.09408212e+00
1.50428161e-01 -1.06337202e+00 2.53778845e-01 -6.28902465e-02
4.67903614e-01 1.95934266e-01 2.20679805e-01 4.82858896e-01
-2.17029303e-01 -9.84545529e-01 1.21105051e+00 4.61103708e-01
7.93341398e-01 -5.45684397e-01 4.18170244e-01 2.52052426e-01
-1.13190615e+00 -1.60681725e-01 7.15456679e-02 -8.75770003e-02
-1.58690661e-02 8.85971487e-02 -9.53129828e-01 6.40270889e-01
6.18911445e-01 7.62436569e-01 -5.70599318e-01 1.04279065e+00
-1.17909387e-01 9.80240166e-01 -3.36795628e-01 3.91575247e-01
3.51957917e-01 2.59543329e-01 6.88354611e-01 1.51362371e+00
1.38366014e-01 -4.63867188e-01 2.20189661e-01 6.61993623e-01
-9.16786119e-02 4.47964936e-01 -8.12595725e-01 -3.16204488e-01
2.46020183e-01 9.88862693e-01 -2.79740423e-01 -7.34382942e-02
-2.62174159e-01 6.60264134e-01 2.99777329e-01 3.11409384e-01
-4.10986036e-01 -1.44419298e-01 4.66816753e-01 3.22713137e-01
3.34954143e-01 -1.33453697e-01 -1.24703147e-01 -1.05067801e+00
8.80134478e-02 -1.20210516e+00 6.95542336e-01 -9.13984895e-01
-1.45745933e+00 5.32242179e-01 -8.87807012e-02 -1.52934849e+00
-6.75143957e-01 -3.85730952e-01 -4.29538608e-01 5.71669042e-01
-1.48420012e+00 -1.27091289e+00 7.14676902e-02 5.58504224e-01
6.22823238e-01 -5.42593002e-01 1.23625004e+00 3.82275939e-01
-5.02942540e-02 6.08187914e-01 -1.41883358e-01 1.81132466e-01
1.19426143e+00 -1.24521244e+00 -1.36700705e-01 1.94101736e-01
1.14483011e+00 4.49218363e-01 7.44053125e-01 -8.64466131e-02
-1.33820307e+00 -9.50461209e-01 6.18664086e-01 -5.14111340e-01
8.97086978e-01 -4.67189223e-01 -9.08065200e-01 6.17948711e-01
2.57788092e-01 -3.30597997e-01 1.11717856e+00 5.32795191e-01
-7.05498993e-01 -2.11486235e-01 -3.45723331e-01 2.16437578e-01
7.96611905e-01 -1.19274998e+00 -7.93732464e-01 2.47856289e-01
6.29305065e-01 -9.69739929e-02 -8.47107530e-01 4.24089402e-01
6.78765953e-01 -6.44553542e-01 1.07721472e+00 -9.07873094e-01
5.03983736e-01 -3.38369638e-01 -4.85356271e-01 -1.00359654e+00
7.01582283e-02 -7.70603716e-01 1.12551406e-01 1.52836525e+00
5.69047213e-01 5.99297062e-02 5.86298943e-01 -1.81697562e-01
-7.79629201e-02 -4.03945416e-01 -9.99438465e-01 -7.04865873e-01
-7.01699182e-02 -8.44293416e-01 2.52911091e-01 1.00586402e+00
1.71891302e-01 7.80150652e-01 -3.58201951e-01 1.13433540e-01
5.38195908e-01 3.16116035e-01 1.14031231e+00 -1.26611960e+00
-7.97430992e-01 -4.31864738e-01 -4.62838590e-01 -7.47163653e-01
5.70425928e-01 -1.26739228e+00 1.63646713e-01 -9.90731299e-01
6.40688092e-02 -3.96444410e-01 -6.44431174e-01 7.30203927e-01
2.80769050e-01 6.00019813e-01 3.50699842e-01 5.62009037e-01
-7.40723550e-01 4.98434216e-01 6.98396087e-01 -4.05892670e-01
-3.52407366e-01 2.73861885e-02 -7.22193897e-01 5.48098505e-01
6.58060253e-01 -5.69243312e-01 -4.04980004e-01 -4.52411532e-01
4.18065995e-01 -3.40583138e-02 5.16247094e-01 -1.05561328e+00
1.33780420e-01 3.74720901e-01 8.16083625e-02 -2.39180639e-01
6.71134710e-01 -6.77691340e-01 1.18285008e-01 -1.14806697e-01
-1.01593542e+00 -2.35881567e-01 4.13813591e-01 7.27120817e-01
-6.72508001e-01 -3.35117906e-01 3.80326152e-01 2.71623619e-02
-1.94234967e-01 -2.39917085e-01 -2.31801942e-01 9.98388603e-02
3.23917329e-01 1.19122691e-01 2.05998644e-01 -6.28330588e-01
-1.00714111e+00 -9.75124836e-02 1.55560479e-01 5.89593351e-01
3.32733661e-01 -1.45030880e+00 -9.07053113e-01 7.36995339e-02
5.75749218e-01 -2.70314187e-01 -2.38693491e-01 5.94310045e-01
1.14873275e-01 5.12674868e-01 -3.27931605e-02 -9.12165582e-01
-1.36526549e+00 3.18591118e-01 2.60206342e-01 -6.18257485e-02
-6.05112016e-01 4.99404460e-01 6.44035339e-02 -5.46719432e-01
7.90136218e-01 4.70133200e-02 -5.34059480e-02 2.14535490e-01
3.92356247e-01 -1.29235625e-01 1.31909445e-01 -7.68455207e-01
-6.57062009e-02 4.51525331e-01 1.55805379e-01 -7.58340180e-01
1.58373058e+00 1.53354406e-01 8.78828838e-02 1.13438463e+00
1.03051329e+00 1.30403191e-01 -1.01267815e+00 -3.55576873e-01
3.13853562e-01 -2.10831866e-01 1.81324437e-01 -8.11766863e-01
-6.31789982e-01 9.92990196e-01 4.96689826e-01 1.75112277e-01
1.09839189e+00 2.70218372e-01 5.05797327e-01 6.55873537e-01
1.96276501e-01 -9.88957763e-01 3.56150955e-01 2.96889037e-01
1.17634296e+00 -9.84724402e-01 -2.06961319e-01 -7.31448457e-02
-6.38359785e-01 1.12983811e+00 1.67299062e-01 -9.69940871e-02
3.15619022e-01 1.31092340e-01 3.53467852e-01 -1.91115335e-01
-1.03466439e+00 -3.90477836e-01 7.31929243e-01 3.77879024e-01
8.12539160e-01 -1.37224570e-01 2.66488224e-01 7.05527961e-01
-5.97700953e-01 -1.41848609e-01 5.50723113e-02 8.75588238e-01
-3.10501456e-01 -1.40498209e+00 -4.82149035e-01 7.30464309e-02
-5.87262094e-01 2.23410800e-02 -7.85305798e-01 6.09066784e-01
-3.51566151e-02 9.03752267e-01 -1.62973508e-01 -2.15570822e-01
3.86676282e-01 5.80138087e-01 5.45948327e-01 -8.43661785e-01
-9.73849177e-01 6.20477080e-01 2.89069355e-01 -4.92438793e-01
-7.69050658e-01 -6.93387568e-01 -9.47602570e-01 4.45962340e-01
-2.10030615e-01 3.60735625e-01 6.83999717e-01 8.28918099e-01
2.46274844e-01 6.65306389e-01 5.23255348e-01 -1.02667308e+00
-6.31053269e-01 -9.83701885e-01 -4.55175430e-01 5.87960780e-01
5.17926753e-01 -5.33013880e-01 -2.88528115e-01 3.74955028e-01] | [15.539666175842285, 5.104557991027832] |
115e64b0-0cee-4a98-836f-5c9915df8b0a | multi-label-learning-based-deep-transfer | 1805.01282 | null | http://arxiv.org/abs/1805.01282v1 | http://arxiv.org/pdf/1805.01282v1.pdf | Multi-label Learning Based Deep Transfer Neural Network for Facial Attribute Classification | Deep Neural Network (DNN) has recently achieved outstanding performance in a
variety of computer vision tasks, including facial attribute classification.
The great success of classifying facial attributes with DNN often relies on a
massive amount of labelled data. However, in real-world applications, labelled
data are only provided for some commonly used attributes (such as age, gender);
whereas, unlabelled data are available for other attributes (such as
attraction, hairline). To address the above problem, we propose a novel deep
transfer neural network method based on multi-label learning for facial
attribute classification, termed FMTNet, which consists of three sub-networks:
the Face detection Network (FNet), the Multi-label learning Network (MNet) and
the Transfer learning Network (TNet). Firstly, based on the Faster Region-based
Convolutional Neural Network (Faster R-CNN), FNet is fine-tuned for face
detection. Then, MNet is fine-tuned by FNet to predict multiple attributes with
labelled data, where an effective loss weight scheme is developed to explicitly
exploit the correlation between facial attributes based on attribute grouping.
Finally, based on MNet, TNet is trained by taking advantage of unsupervised
domain adaptation for unlabelled facial attribute classification. The three
sub-networks are tightly coupled to perform effective facial attribute
classification. A distinguishing characteristic of the proposed FMTNet method
is that the three sub-networks (FNet, MNet and TNet) are constructed in a
similar network structure. Extensive experimental results on challenging face
datasets demonstrate the effectiveness of our proposed method compared with
several state-of-the-art methods. | ['Si Chen', 'Ni Zhuang', 'Chunhua Shen', 'Yan Yan', 'Hanzi Wang'] | 2018-05-03 | null | null | null | null | ['facial-attribute-classification'] | ['computer-vision'] | [ 3.30271631e-01 5.06116338e-02 -1.69390246e-01 -7.26187646e-01
-3.21611643e-01 -6.75572827e-02 4.69320208e-01 -4.78421431e-03
-3.04898858e-01 6.02264524e-01 -1.54359251e-01 1.06404416e-01
-1.64979398e-01 -9.18368161e-01 -4.51500416e-01 -1.00023687e+00
4.98114601e-02 5.04205346e-01 -7.62492046e-02 -1.40194356e-01
-5.67343608e-02 5.71949184e-01 -1.68798530e+00 1.02927007e-01
7.47724295e-01 1.75667536e+00 -3.16840291e-01 2.52796412e-02
-1.99664995e-01 7.25247562e-01 -2.04022169e-01 -5.22336364e-01
2.33876109e-01 -3.29829603e-01 -7.39422739e-01 1.22195691e-01
3.62541258e-01 -4.05365229e-01 -9.27661583e-02 1.03728521e+00
7.21495867e-01 1.22652434e-01 8.55222583e-01 -1.57481039e+00
-4.77738291e-01 1.71681508e-01 -7.87028193e-01 -1.38119623e-01
2.93850750e-02 -1.61485657e-01 8.15049946e-01 -9.90534842e-01
1.63065821e-01 1.60761905e+00 7.81092763e-01 5.69499969e-01
-9.62245047e-01 -1.21858025e+00 1.47104576e-01 1.72484562e-01
-1.71737897e+00 -5.12788415e-01 7.85980761e-01 -3.03930372e-01
3.02349299e-01 -1.77302331e-01 5.12156188e-01 1.00114930e+00
-3.09697958e-03 6.45619869e-01 1.00092077e+00 -2.26425260e-01
-1.21548474e-01 1.22502506e-01 -2.33160764e-01 8.67955089e-01
-1.70620710e-01 5.86674437e-02 -1.67878613e-01 -1.96470618e-01
8.40958297e-01 2.53372699e-01 8.73037353e-02 -3.52672905e-01
-7.40388036e-01 9.67630565e-01 5.92684686e-01 -4.08599433e-03
-5.55809975e-01 -1.20419428e-01 7.98248112e-01 2.05283523e-01
8.16652358e-01 -2.89291710e-01 -5.08844495e-01 2.52687365e-01
-5.54118931e-01 -5.39622642e-02 4.31434035e-01 8.98709476e-01
1.14458358e+00 7.25020170e-02 -2.16402367e-01 1.46592760e+00
3.76675963e-01 3.52358311e-01 4.59414810e-01 -5.88567197e-01
1.92819923e-01 8.26397717e-01 -3.88948858e-01 -1.03271544e+00
-4.93879735e-01 -3.13441187e-01 -1.26380503e+00 2.45434061e-01
2.11558759e-01 -2.68386245e-01 -8.20961535e-01 1.80725706e+00
6.14080191e-01 3.48510504e-01 1.89441256e-02 7.61175990e-01
9.81192112e-01 5.21102786e-01 6.31074548e-01 -1.09321214e-01
1.37962961e+00 -7.97849834e-01 -5.94002843e-01 -1.00488186e-01
5.88118494e-01 -6.16861880e-01 5.63368440e-01 1.65179998e-01
-6.73883736e-01 -8.29911947e-01 -7.60982513e-01 2.08330557e-01
-3.79712731e-01 3.14746797e-01 7.06224680e-01 4.57912743e-01
-9.00525808e-01 4.14914131e-01 -1.85449272e-01 -5.31445980e-01
1.01915407e+00 9.51521158e-01 -7.11882889e-01 -1.24085456e-01
-1.26252115e+00 5.63380063e-01 6.09082699e-01 3.63501072e-01
-9.71730947e-01 -3.73824984e-01 -1.01053786e+00 6.02162108e-02
4.34757680e-01 -5.04842639e-01 1.05395710e+00 -1.70455301e+00
-1.61602283e+00 1.09392512e+00 6.97933212e-02 1.05234943e-01
2.96627849e-01 -7.15235481e-04 -7.52372622e-01 1.54530182e-01
1.16984896e-01 8.31597984e-01 9.79004443e-01 -1.02259564e+00
-8.35407674e-01 -6.69197977e-01 -1.72422111e-01 1.76418945e-01
-8.03805470e-01 2.11784437e-01 -1.80428386e-01 -5.15906572e-01
-7.14104995e-02 -8.23161662e-01 1.19623505e-01 3.35205734e-01
-3.28453302e-01 -7.39462078e-01 1.19678199e+00 -3.91820997e-01
9.33840096e-01 -2.28758025e+00 -4.70681023e-03 3.67994100e-01
4.12265569e-01 5.66141963e-01 -3.41244608e-01 1.48370760e-02
-3.25551808e-01 -1.18564747e-01 -1.83425158e-01 -2.75010228e-01
-1.86986253e-01 3.54429603e-01 2.17941463e-01 4.83448416e-01
5.61810017e-01 5.99883914e-01 -7.50390232e-01 -9.55303192e-01
9.77902263e-02 5.44759393e-01 -2.43296355e-01 3.63738388e-01
-1.26788884e-01 4.92849082e-01 -8.47685456e-01 1.03680027e+00
9.09247756e-01 -3.05283405e-02 -8.22670236e-02 -3.01570922e-01
4.68272194e-02 -3.81589413e-01 -1.05453348e+00 1.23795688e+00
-4.64478284e-01 9.65553075e-02 3.26060891e-01 -1.13662708e+00
1.31166267e+00 5.09472847e-01 5.68763614e-01 -6.52373135e-01
4.47904885e-01 2.87066281e-01 -2.96547543e-02 -4.44232225e-01
-9.73978117e-02 -3.27596486e-01 1.27398863e-01 3.54740769e-01
3.77285659e-01 5.27195394e-01 -1.67892739e-01 -1.49067700e-01
5.50353706e-01 1.42653227e-01 4.07149673e-01 -2.37866759e-01
1.13126707e+00 -5.75337708e-01 9.11555231e-01 1.21593721e-01
-4.14946586e-01 2.55318373e-01 4.37964886e-01 -8.93629968e-01
-1.01349878e+00 -5.97951055e-01 -4.07668501e-01 1.58624411e+00
2.29901802e-02 -5.15380800e-02 -8.42331588e-01 -9.96478617e-01
1.62372202e-01 -7.61808604e-02 -8.69828820e-01 -2.45510817e-01
-3.71286005e-01 -8.22915375e-01 6.98720574e-01 7.49786556e-01
1.01946318e+00 -1.47631633e+00 4.38563153e-03 2.18298271e-01
1.97135471e-02 -1.04656231e+00 -4.22750771e-01 5.63376909e-03
-6.00139558e-01 -9.75745618e-01 -6.90322757e-01 -8.86473298e-01
7.64245331e-01 -3.38022187e-02 8.26335311e-01 3.64230841e-01
-1.27421275e-01 1.04507655e-01 -2.94599831e-01 -6.31157994e-01
-2.21869543e-01 7.99508840e-02 2.86114484e-01 1.06849480e+00
6.50544345e-01 -7.18832135e-01 -6.31451905e-01 5.17233551e-01
-7.66454458e-01 -3.11424494e-01 1.09553695e+00 1.07843196e+00
6.78600729e-01 -7.29259029e-02 9.91824090e-01 -1.19378436e+00
2.14306414e-01 -5.65587759e-01 -3.56288105e-01 1.23351164e-01
-4.84452158e-01 -1.84084535e-01 7.17633545e-01 -5.96505225e-01
-1.17160404e+00 2.35328168e-01 -4.77898329e-01 -5.21624148e-01
-3.49784911e-01 2.46490061e-01 -6.23915136e-01 -4.70038086e-01
2.16044843e-01 1.87841386e-01 4.23449188e-01 -2.50742525e-01
5.16056977e-02 9.45772827e-01 4.66216117e-01 -5.76482296e-01
6.11859381e-01 3.21953475e-01 2.92518675e-01 -5.97036242e-01
-1.06068575e+00 -4.11192358e-01 -8.52545440e-01 -3.60287040e-01
9.66065764e-01 -9.50111151e-01 -1.08244753e+00 8.75093043e-01
-8.68973255e-01 -6.27116719e-03 1.37999326e-01 3.85692269e-01
-5.32987654e-01 2.82230787e-02 -5.21812797e-01 -8.44393015e-01
-5.55438936e-01 -9.44655180e-01 1.18319297e+00 4.75264639e-01
7.62723163e-02 -1.14875400e+00 -3.97192657e-01 2.40013152e-01
4.65959728e-01 4.96444702e-01 1.02777445e+00 -1.06893659e+00
-1.20794401e-01 -2.57644385e-01 -6.45831585e-01 5.55998027e-01
3.09294850e-01 -1.99219603e-02 -1.27207088e+00 -3.58666241e-01
-3.43262136e-01 -7.95773685e-01 7.29714751e-01 3.28551680e-01
1.53677237e+00 -2.96021849e-01 -5.15720963e-01 8.35737169e-01
1.29950464e+00 2.34368473e-01 6.27848148e-01 3.07545155e-01
9.70125973e-01 6.91673160e-01 6.75719023e-01 5.29171288e-01
2.77136266e-01 7.25175560e-01 7.34465063e-01 -5.10684371e-01
1.45278335e-01 -6.60302443e-03 3.74269485e-03 3.95232797e-01
-3.76468629e-01 8.73746574e-02 -5.76106429e-01 2.96217591e-01
-1.71282399e+00 -7.14159966e-01 2.24722847e-01 2.18893456e+00
6.93909287e-01 -4.01664525e-03 4.23393101e-01 3.51422504e-02
9.69540894e-01 4.73651551e-02 -6.74037993e-01 -4.61930245e-01
-1.45321488e-01 2.35734195e-01 1.51438028e-01 -1.42485052e-01
-1.62332761e+00 8.98599267e-01 4.86637449e+00 1.14872277e+00
-1.06142545e+00 6.16929308e-02 9.88174140e-01 3.91849399e-01
3.03552121e-01 -4.31672245e-01 -1.03558290e+00 4.52059567e-01
7.05864668e-01 1.81324095e-01 1.25842586e-01 9.16786432e-01
-4.77414466e-02 3.27153236e-01 -1.01387632e+00 9.81828094e-01
1.69389978e-01 -8.25111330e-01 3.14022422e-01 7.24137127e-02
3.63093823e-01 -4.69072789e-01 1.92587852e-01 5.35126269e-01
1.73430294e-01 -1.17431998e+00 3.21586609e-01 3.82246852e-01
1.16004395e+00 -1.27534223e+00 1.02485478e+00 8.93608015e-03
-1.44297028e+00 -4.45566267e-01 -6.23446584e-01 1.86653629e-01
-2.92800546e-01 4.52943474e-01 -4.62277949e-01 6.24316156e-01
7.39604890e-01 1.03506243e+00 -4.50076669e-01 8.00643384e-01
-6.30598068e-02 3.74213398e-01 -1.13143727e-01 2.33174786e-01
4.37974304e-01 -3.84394556e-01 4.21036072e-02 1.00252151e+00
2.31150821e-01 3.53906810e-01 3.83856505e-01 5.22604644e-01
-3.96131396e-01 3.94526541e-01 -5.72682261e-01 2.19533890e-01
5.35197079e-01 1.85159862e+00 -4.31711555e-01 -1.04930602e-01
-5.95106006e-01 8.63352299e-01 4.58795816e-01 1.22947440e-01
-6.40236139e-01 -4.33794975e-01 7.84979880e-01 -7.72200152e-02
2.56243378e-01 4.57476705e-01 2.06001565e-01 -4.32255745e-01
-2.06914842e-01 -7.50716448e-01 6.26023412e-01 -5.11505485e-01
-1.54345798e+00 8.91982675e-01 -2.97637135e-01 -1.28830087e+00
-6.61803037e-02 -7.78723776e-01 -6.02535248e-01 1.02817905e+00
-1.59591770e+00 -1.72584558e+00 -5.76411366e-01 9.91540790e-01
1.89748049e-01 -5.94943464e-01 9.47010875e-01 7.02783406e-01
-8.22669923e-01 1.02576029e+00 -5.17285317e-02 5.99609911e-01
9.38595474e-01 -9.70461190e-01 -1.37725845e-01 1.45219222e-01
-3.66127610e-01 3.10183644e-01 -8.64604861e-02 -4.20050055e-01
-8.30457926e-01 -1.70708358e+00 6.61232948e-01 -7.39491805e-02
4.02514875e-01 -3.37143540e-01 -9.37250674e-01 7.41582394e-01
-7.47795850e-02 6.66312397e-01 8.27957273e-01 1.38634488e-01
-4.75547940e-01 -7.15551138e-01 -1.29808903e+00 5.30150980e-02
7.83021867e-01 -3.97455007e-01 4.75825258e-02 4.75284010e-01
3.48477691e-01 -2.22005188e-01 -1.11182487e+00 6.89976096e-01
7.22826660e-01 -7.77166009e-01 9.99203742e-01 -5.28808653e-01
5.84947050e-01 -8.35216790e-02 5.80997765e-02 -1.35458362e+00
-4.86619323e-01 -1.91454813e-01 2.69536138e-01 1.57465041e+00
6.65678382e-02 -7.24531233e-01 7.65987992e-01 2.41139174e-01
-2.05478430e-01 -1.13483727e+00 -9.37409282e-01 -6.51217997e-01
-9.57552269e-02 3.21712762e-01 8.35583866e-01 1.01578760e+00
-6.36024773e-01 5.45155942e-01 -5.14988363e-01 1.61913652e-02
7.12751031e-01 -9.84345898e-02 4.94184464e-01 -1.83892560e+00
2.57817030e-01 -4.13181275e-01 -7.68696070e-01 -6.29603446e-01
5.56402028e-01 -7.76018560e-01 -2.05831543e-01 -1.06043863e+00
3.79057407e-01 -8.05119395e-01 -5.64584970e-01 7.89873540e-01
-2.04042703e-01 5.70605814e-01 -1.48357153e-01 1.42928779e-01
-6.59092784e-01 7.84662485e-01 1.31218541e+00 -1.74480647e-01
1.58063695e-01 2.57642299e-01 -6.11248732e-01 8.38510752e-01
5.86363077e-01 -3.07954103e-01 -2.41949022e-01 -9.21610743e-02
-3.31948549e-01 -2.43914425e-02 4.54146653e-01 -9.19830799e-01
1.99691594e-01 1.15625151e-01 8.72770011e-01 -3.25430065e-01
3.00048977e-01 -9.75601315e-01 -2.10544422e-01 1.12297229e-01
-1.98300824e-01 9.46314726e-03 2.80639142e-01 4.43107247e-01
-4.26579714e-01 5.26462793e-02 1.10714984e+00 4.69383113e-02
-1.01240063e+00 1.00548029e+00 2.19584741e-02 -1.41736463e-01
1.28231752e+00 -4.24484819e-01 -9.87521335e-02 -2.42754668e-01
-7.15248764e-01 3.96552533e-01 -9.22441334e-02 2.89693326e-01
7.44051218e-01 -1.81436181e+00 -9.47013259e-01 4.66942519e-01
3.78675669e-01 4.18774076e-02 3.00321966e-01 7.79909432e-01
-1.51676849e-01 8.90230015e-02 -7.30927467e-01 -6.48208618e-01
-1.37337089e+00 5.62242270e-01 4.24610943e-01 -3.72797213e-02
-3.23307306e-01 1.02294803e+00 5.73459387e-01 -7.13248312e-01
1.40096128e-01 5.12083948e-01 -7.61615336e-01 2.60625571e-01
5.26605666e-01 1.74498975e-01 1.06900498e-01 -1.02260268e+00
-3.63995969e-01 8.82370353e-01 -3.41261148e-01 5.32099903e-01
1.32871616e+00 -7.43005201e-02 -4.87480581e-01 8.78653396e-03
1.29107809e+00 -5.54381490e-01 -1.30623448e+00 -4.65279818e-01
-2.41676822e-01 -1.69579610e-01 7.80587271e-03 -6.44208074e-01
-1.52009928e+00 9.25395370e-01 8.18488479e-01 -1.91412613e-01
1.68597281e+00 -1.92103952e-01 7.25181341e-01 1.78239956e-01
2.36516178e-01 -1.01482880e+00 1.42918840e-01 3.96545410e-01
5.27870953e-01 -1.43921328e+00 -2.06272513e-01 -5.22787809e-01
-5.60348988e-01 1.22220898e+00 1.04153085e+00 1.65166080e-01
8.18322361e-01 -4.38679717e-02 1.64520100e-01 -2.09162578e-01
-3.24139625e-01 -5.27726710e-02 1.16537139e-01 6.86198592e-01
3.18209589e-01 -8.30217674e-02 1.33650571e-01 6.44206524e-01
1.62185684e-01 2.26383097e-02 -1.77614968e-02 5.64933121e-01
-3.79885614e-01 -1.11495674e+00 -3.04904968e-01 7.13573992e-01
-5.23410678e-01 -2.78426986e-02 -2.61790752e-01 9.00080800e-01
5.38452923e-01 7.41480649e-01 2.39728943e-01 -5.25447428e-01
1.98616415e-01 4.05650586e-02 2.62934595e-01 -4.91766632e-01
-6.06905282e-01 -3.88471819e-02 -1.06996350e-01 -4.59289730e-01
-6.34493887e-01 -4.97903407e-01 -9.23724174e-01 -3.73518139e-01
-2.10081384e-01 1.39764190e-01 2.49268070e-01 1.07638752e+00
1.43940821e-01 4.64455813e-01 9.50561166e-01 -7.77235210e-01
-3.94782513e-01 -1.11111462e+00 -8.51684749e-01 5.22837222e-01
3.64791393e-01 -1.09539127e+00 -1.37957767e-01 -2.61494610e-02] | [13.508076667785645, 0.8130905032157898] |
f108787d-f613-4949-91d6-8efb14ea6226 | efficient-and-scalable-recommendation-via | 2207.05959 | null | https://arxiv.org/abs/2207.05959v2 | https://arxiv.org/pdf/2207.05959v2.pdf | Fine-tuning Partition-aware Item Similarities for Efficient and Scalable Recommendation | Collaborative filtering (CF) is widely searched in recommendation with various types of solutions. Recent success of Graph Convolution Networks (GCN) in CF demonstrates the effectiveness of modeling high-order relationships through graphs, while repetitive graph convolution and iterative batch optimization limit their efficiency. Instead, item similarity models attempt to construct direct relationships through efficient interaction encoding. Despite their great performance, the growing item numbers result in quadratic growth in similarity modeling process, posing critical scalability problems. In this paper, we investigate the graph sampling strategy adopted in latest GCN model for efficiency improving, and identify the potential item group structure in the sampled graph. Based on this, we propose a novel item similarity model which introduces graph partitioning to restrict the item similarity modeling within each partition. Specifically, we show that the spectral information of the original graph is well in preserving global-level information. Then, it is added to fine-tune local item similarities with a new data augmentation strategy acted as partition-aware prior knowledge, jointly to cope with the information loss brought by partitioning. Experiments carried out on 4 datasets show that the proposed model outperforms state-of-the-art GCN models with 10x speed-up and item similarity models with 95\% parameter storage savings. | ['Tommy W. S. Chow', 'Jianghong Ma', 'Tianjun Wei'] | 2022-07-13 | null | null | null | null | ['graph-sampling', 'graph-partitioning'] | ['graphs', 'graphs'] | [-3.53724957e-02 -1.84071973e-01 -3.33263665e-01 -1.91638514e-01
2.57320285e-01 -3.68024588e-01 2.51519144e-01 4.87559468e-01
-3.31140272e-02 3.66406083e-01 3.41176301e-01 -2.46065140e-01
-7.50707746e-01 -1.10152042e+00 -6.22246861e-01 -4.11136776e-01
-3.87498796e-01 2.59504408e-01 1.25105843e-01 -3.38449568e-01
2.50914037e-01 3.23435634e-01 -1.26767349e+00 2.71066308e-01
1.28923655e+00 1.07782328e+00 3.62621635e-01 1.73855454e-01
-4.88449097e-01 4.78871584e-01 -2.16665372e-01 -5.68666756e-01
5.95999956e-01 -5.15625656e-01 -5.86635470e-01 2.37754777e-01
4.58737344e-01 -3.09819076e-02 -7.25787103e-01 1.22030544e+00
4.62199718e-01 6.62958443e-01 3.82735640e-01 -1.04098594e+00
-1.13345110e+00 9.74317491e-01 -4.93702263e-01 3.01390737e-01
3.41350257e-01 -8.50671902e-02 1.45646441e+00 -7.60300338e-01
3.45706999e-01 1.05466259e+00 9.44476902e-01 1.99744198e-02
-1.18254995e+00 -5.25993824e-01 6.66614413e-01 1.24924749e-01
-1.55932546e+00 1.48705274e-01 9.22734320e-01 -1.07114855e-02
1.05520654e+00 3.31454813e-01 9.94854033e-01 5.97392917e-01
-4.33420949e-03 5.25896907e-01 5.67700505e-01 -2.62313455e-01
3.30749527e-02 1.37590379e-01 3.89431536e-01 9.12285864e-01
5.54886103e-01 -1.40060768e-01 -4.83127236e-01 -2.71311164e-01
1.09318304e+00 4.23631728e-01 -5.12575805e-01 -5.23156822e-01
-7.62175500e-01 7.00180292e-01 9.45558846e-01 3.79625559e-01
-3.42992604e-01 -8.94675322e-04 2.40262866e-01 5.96878469e-01
5.31934381e-01 4.16637301e-01 -2.50646949e-01 5.12617528e-01
-6.39966547e-01 1.20030366e-01 8.84984553e-01 1.31542635e+00
5.59646130e-01 7.60309538e-03 -2.81097174e-01 9.78532612e-01
3.78514796e-01 -2.63032895e-02 4.57239389e-01 -3.50491524e-01
5.12295485e-01 1.05410683e+00 -2.43403435e-01 -1.60764194e+00
-4.82337415e-01 -1.29465079e+00 -1.25567365e+00 -4.76872772e-01
2.36756042e-01 2.47843251e-01 -6.98268175e-01 1.49967492e+00
3.20658892e-01 4.64325339e-01 -4.42979664e-01 1.00137269e+00
7.65432656e-01 4.32274550e-01 3.28608393e-03 -2.20075399e-01
1.15298510e+00 -1.05951130e+00 -7.01190412e-01 1.22142345e-01
7.32988596e-01 -6.02849901e-01 1.01350152e+00 4.31057841e-01
-1.01478028e+00 -8.82261932e-01 -9.72712398e-01 2.68824786e-01
-3.99149746e-01 -7.69777745e-02 1.18712473e+00 8.87466550e-01
-1.07321262e+00 8.33934486e-01 -4.17586416e-01 -3.66824716e-01
5.14887214e-01 6.41803861e-01 -2.16001701e-02 -1.78884968e-01
-1.32464159e+00 3.57978880e-01 4.74340588e-01 1.07375331e-01
-1.00965098e-01 -9.65900898e-01 -6.03632390e-01 5.34839988e-01
6.07613802e-01 -9.88313198e-01 4.52678919e-01 -8.95070732e-01
-1.15675259e+00 9.96717811e-02 2.83250749e-01 -5.89648902e-01
1.96682021e-01 -3.37235816e-02 -7.69141257e-01 -7.62253553e-02
-3.09521466e-01 1.87753737e-01 6.14236057e-01 -1.05935264e+00
-5.83976924e-01 -3.64487380e-01 2.37784848e-01 5.18106222e-01
-7.74084151e-01 -2.01163694e-01 -7.85091519e-01 -9.34107900e-01
3.82988960e-01 -6.22684300e-01 -5.25077939e-01 -9.31356400e-02
-7.34359249e-02 -2.87575901e-01 3.44603270e-01 -4.92115378e-01
1.92693770e+00 -1.96637130e+00 -6.01029135e-02 6.82975054e-01
4.79716897e-01 3.54630351e-01 -2.88309216e-01 5.84218919e-01
4.97748703e-02 3.06527112e-02 1.35416090e-01 -2.17797533e-01
-1.56614911e-02 2.22520664e-01 2.46350560e-02 4.03709859e-01
-1.74214527e-01 9.41046476e-01 -7.71281600e-01 -2.72443920e-01
9.91960615e-02 3.63356978e-01 -9.50252116e-01 2.35142857e-02
-1.71228439e-01 4.48708870e-02 -4.36085045e-01 5.69171965e-01
9.59634483e-01 -6.08485818e-01 4.12104756e-01 -5.53289831e-01
2.51327425e-01 1.98570192e-01 -1.64852667e+00 1.75971222e+00
-2.95082003e-01 -3.02184016e-01 1.84940085e-01 -1.31482649e+00
1.02153909e+00 -7.54114464e-02 5.16923606e-01 -6.47382915e-01
1.35340601e-01 1.70827117e-02 1.83208257e-01 -8.53721723e-02
5.12005687e-01 3.20313573e-01 2.71961331e-01 3.03084910e-01
1.84188381e-01 4.11105156e-01 2.48491287e-01 3.82734179e-01
8.25477004e-01 -2.23111480e-01 1.87143326e-01 -4.90879893e-01
5.92092752e-01 -3.14844370e-01 2.15127766e-01 1.02128768e+00
9.14898217e-02 3.31185430e-01 -3.23869810e-02 -4.43141520e-01
-5.75244486e-01 -7.97237635e-01 1.56001434e-01 1.15081143e+00
5.06334245e-01 -6.47104323e-01 -3.22057098e-01 -8.10360968e-01
1.92015648e-01 2.82999068e-01 -5.43917537e-01 -3.09882015e-01
-5.45450211e-01 -9.06949282e-01 5.08764796e-02 4.48032022e-01
6.41987920e-01 -7.34660745e-01 3.15543383e-01 4.34149712e-01
1.66337088e-01 -6.91715777e-01 -9.44140375e-01 1.06610268e-01
-1.03849685e+00 -9.95690644e-01 -3.43575716e-01 -9.34269071e-01
6.98996007e-01 7.64730275e-01 1.19120014e+00 4.54437017e-01
-1.13346003e-01 3.65257114e-01 -6.39242411e-01 1.18882515e-01
6.97568133e-02 3.23193520e-02 3.22153494e-02 1.69911802e-01
4.24648523e-01 -8.45254719e-01 -9.01569366e-01 4.56400901e-01
-8.14870000e-01 -2.26025641e-01 5.28985083e-01 8.59499335e-01
6.25984430e-01 4.82050896e-01 6.95012867e-01 -1.03629339e+00
1.06913829e+00 -7.53630757e-01 -4.41988528e-01 4.34921265e-01
-1.17371953e+00 -1.08627565e-01 1.06772709e+00 -5.68044126e-01
-9.23749268e-01 -2.71358877e-01 1.19016849e-01 -4.99209732e-01
1.38846338e-01 8.03866327e-01 -2.22544387e-01 -3.84382010e-01
4.85848755e-01 2.70598143e-01 -4.33144458e-02 -7.06155360e-01
5.64829946e-01 2.85917580e-01 2.02280641e-01 -4.19221193e-01
5.56824088e-01 2.04776347e-01 6.51038215e-02 -5.02589822e-01
-7.63207614e-01 -8.08592856e-01 -4.52970326e-01 1.43317208e-02
3.33160698e-01 -8.05253327e-01 -7.24733889e-01 3.82011607e-02
-5.56653976e-01 1.79489359e-01 -2.71627635e-01 5.44603288e-01
-9.07649379e-03 8.19225967e-01 -6.24772131e-01 -6.75613940e-01
-6.00905895e-01 -5.88783026e-01 3.53577226e-01 1.67308316e-01
1.17735676e-01 -1.18706977e+00 -1.56529248e-01 3.10796857e-01
5.52685142e-01 -3.99246842e-01 8.83424103e-01 -9.96694148e-01
-6.55464828e-01 -1.76445976e-01 -5.04229903e-01 2.13691890e-01
3.29636365e-01 -6.23064876e-01 -2.00353265e-01 -5.02023757e-01
-1.17153861e-01 2.76688129e-01 9.29240346e-01 3.85731459e-01
1.44113755e+00 -4.21847671e-01 -3.94715279e-01 8.64784479e-01
1.58471847e+00 1.76244184e-01 3.33981484e-01 -6.30175695e-02
1.05123997e+00 3.10812503e-01 5.13399065e-01 5.60081780e-01
2.02673256e-01 5.41461468e-01 4.47101802e-01 -5.61507829e-02
-3.66469115e-01 -3.32356125e-01 -1.60717830e-01 1.11592412e+00
-2.10079655e-01 -7.73741126e-01 -3.01118970e-01 2.83490270e-01
-2.10856819e+00 -7.37489223e-01 -3.44867408e-01 2.31382561e+00
4.54028517e-01 1.36835545e-01 2.48708025e-01 -7.22112954e-02
7.48857021e-01 -8.59602820e-04 -3.44789594e-01 -2.05859885e-01
-3.06232542e-01 2.41245702e-01 5.96399128e-01 2.55850554e-01
-1.06114864e+00 7.00970113e-01 5.64430046e+00 1.13369036e+00
-6.55540347e-01 -8.80468786e-02 3.61787587e-01 -5.85324131e-02
-3.94211560e-01 -1.34210691e-01 -8.33675265e-01 5.88264585e-01
6.86171770e-01 -5.60966618e-02 6.78503335e-01 5.24041295e-01
1.25763398e-02 3.09982061e-01 -8.19252908e-01 9.93084192e-01
1.29476279e-01 -1.37496006e+00 5.47278345e-01 9.30382833e-02
8.31606209e-01 -1.94570392e-01 2.34885979e-02 4.51514095e-01
1.63615838e-01 -8.55439067e-01 2.23130733e-01 5.36542416e-01
3.45042527e-01 -1.04371393e+00 7.58005798e-01 2.52402574e-01
-1.66684353e+00 -2.33953044e-01 -7.18231201e-01 -1.78401738e-01
-1.64312366e-02 5.59317708e-01 -6.53779626e-01 1.09638834e+00
6.14147782e-01 8.57058942e-01 -6.52344227e-01 1.32574689e+00
1.99747190e-01 6.96336329e-01 -3.94942224e-01 -2.04720244e-01
2.64046073e-01 -7.48400390e-01 3.38001609e-01 1.15824151e+00
3.66261512e-01 3.05323809e-01 5.00029385e-01 8.14888299e-01
-2.15927213e-01 5.30152738e-01 -4.25105542e-01 -6.37393519e-02
5.37089586e-01 1.08772945e+00 -7.99767852e-01 -1.77485183e-01
-7.82069027e-01 9.83326435e-01 4.73878860e-01 2.42341548e-01
-7.73515761e-01 -2.73344100e-01 4.29173201e-01 3.02370429e-01
6.32333338e-01 -1.80606112e-01 -3.07088554e-01 -1.02441013e+00
-1.60149306e-01 -7.29441226e-01 7.01608956e-01 -1.61313489e-01
-1.72373843e+00 3.23166966e-01 -1.52373672e-01 -1.33814657e+00
3.27982128e-01 -3.88412535e-01 -5.12358844e-01 8.64181161e-01
-1.08266914e+00 -1.30079007e+00 -2.80074894e-01 7.42944896e-01
4.10972476e-01 -1.35380119e-01 5.47987878e-01 6.36886716e-01
-3.86554569e-01 1.13207340e+00 2.25692272e-01 -1.97964147e-01
3.98655921e-01 -1.06630778e+00 3.41001451e-01 7.25969017e-01
3.71715367e-01 1.08455777e+00 3.93159121e-01 -9.92388666e-01
-1.67292261e+00 -1.17990375e+00 6.03190958e-01 6.80512637e-02
8.04910243e-01 -3.65383238e-01 -1.19661272e+00 3.10964167e-01
5.24005666e-02 2.42092282e-01 7.98139632e-01 4.71353918e-01
-3.39829981e-01 -1.39294416e-01 -1.13531840e+00 5.92055559e-01
1.81093180e+00 -2.23524004e-01 -2.11630285e-01 3.37029368e-01
8.78315806e-01 -1.96928270e-02 -1.14693654e+00 4.12853628e-01
4.52930719e-01 -9.04775679e-01 1.20026958e+00 -7.06018150e-01
-4.41449061e-02 -2.36151040e-01 -1.63973063e-01 -1.19369721e+00
-9.92600858e-01 -6.14563704e-01 -5.07945895e-01 1.13193393e+00
2.59333700e-01 -6.94989502e-01 1.00082743e+00 3.93683702e-01
-3.21657687e-01 -1.03153181e+00 -4.43378806e-01 -8.66747022e-01
-3.75602424e-01 -1.07463695e-01 7.31482089e-01 1.14508700e+00
1.00951567e-01 4.23893154e-01 -4.23962623e-01 2.68159509e-01
6.47832096e-01 3.97327125e-01 4.46615517e-01 -1.27349687e+00
-7.34614849e-01 -6.35336757e-01 -3.32853854e-01 -1.48156321e+00
-3.34438026e-01 -1.22202826e+00 -6.53575182e-01 -1.55544138e+00
1.55232102e-01 -6.40588939e-01 -7.84464419e-01 1.52884662e-01
-3.20632398e-01 1.60918087e-01 1.26490623e-01 1.91259116e-01
-7.79797971e-01 4.94136244e-01 1.49964535e+00 -9.97518227e-02
-5.15636921e-01 2.95225501e-01 -9.00262773e-01 3.40843916e-01
6.89085305e-01 -3.51044275e-02 -8.67759287e-01 -3.02611142e-01
4.25988346e-01 -4.41158749e-02 3.38033959e-02 -8.30474436e-01
4.92345780e-01 8.85103270e-02 4.68255907e-01 -7.56162941e-01
4.63535078e-02 -1.05472958e+00 3.89940143e-01 4.28263783e-01
-2.33638451e-01 -1.86629996e-01 3.82057112e-03 1.16641748e+00
-4.66528162e-02 -2.00167671e-01 4.97938663e-01 -1.41675875e-01
-5.78685582e-01 6.78729773e-01 3.04552708e-02 -3.01313609e-01
5.99550545e-01 -5.54925382e-01 9.17563122e-03 -2.65209615e-01
-9.07454252e-01 2.83179492e-01 1.47320375e-01 3.53362918e-01
5.39930761e-01 -1.34186304e+00 -4.59083170e-01 3.91133517e-01
-7.27196634e-02 -2.01556131e-01 5.25145650e-01 7.64159143e-01
-1.87518030e-01 3.51157606e-01 6.23887666e-02 -1.92645997e-01
-1.12124002e+00 9.68650341e-01 2.80891526e-02 -5.75204432e-01
-6.38321936e-01 1.07539368e+00 2.45516837e-01 -2.83297211e-01
3.87790143e-01 -3.20107430e-01 -4.99806851e-01 8.43663886e-03
2.45291039e-01 4.40505296e-01 2.28996843e-01 -2.10821494e-01
-9.98981073e-02 3.89731735e-01 -3.91073674e-01 7.24268079e-01
1.18826509e+00 -4.34027016e-01 -1.14919804e-01 -1.15020685e-01
1.06667161e+00 6.30843081e-03 -7.28441417e-01 -5.64179242e-01
-2.39214435e-01 -6.51556313e-01 1.38297528e-01 -7.82838821e-01
-1.16983795e+00 2.62441516e-01 6.11452401e-01 6.29916728e-01
1.30115581e+00 -1.98922992e-01 8.95477653e-01 3.55801612e-01
5.46938956e-01 -9.19364274e-01 -7.39154071e-02 4.08890307e-01
6.82105541e-01 -1.03756642e+00 2.64157623e-01 -9.20816064e-01
-2.90766954e-01 8.83927822e-01 7.15794504e-01 -4.72328603e-01
1.07181525e+00 -1.73323572e-01 -7.40361512e-01 -3.76880258e-01
-5.17163575e-01 -2.41317049e-01 6.91267848e-01 5.40910304e-01
3.90378207e-01 2.80013513e-02 -6.25193954e-01 1.02397013e+00
-1.14052603e-02 -2.64158070e-01 -5.30101173e-02 6.35736883e-01
-5.39497495e-01 -1.07083428e+00 3.33679095e-02 1.00718451e+00
-3.20766598e-01 -4.92760420e-01 -2.25203320e-01 6.28088176e-01
3.09836030e-01 8.90634239e-01 7.22133890e-02 -6.08967841e-01
3.54012907e-01 -3.62035632e-01 5.66688538e-01 -5.80578983e-01
-1.10032487e+00 2.96349704e-01 1.63980290e-01 -5.68680525e-01
-2.49405220e-01 -3.24821442e-01 -1.02843106e+00 -2.35138282e-01
-6.88729823e-01 3.78715664e-01 3.14306438e-01 5.40489376e-01
5.36729753e-01 5.37641287e-01 7.41289198e-01 -4.33179557e-01
-6.33477867e-01 -1.06082547e+00 -9.49747026e-01 5.13809919e-01
-2.16669530e-01 -5.45513570e-01 -4.42552835e-01 -3.57818246e-01] | [10.190051078796387, 5.626008033752441] |
e475cbd0-2009-436b-b9ec-caaa3a0b2dc1 | rethinking-table-parsing-using-graph-neural | 1905.13391 | null | https://arxiv.org/abs/1905.13391v2 | https://arxiv.org/pdf/1905.13391v2.pdf | Rethinking Table Recognition using Graph Neural Networks | Document structure analysis, such as zone segmentation and table recognition, is a complex problem in document processing and is an active area of research. The recent success of deep learning in solving various computer vision and machine learning problems has not been reflected in document structure analysis since conventional neural networks are not well suited to the input structure of the problem. In this paper, we propose an architecture based on graph networks as a better alternative to standard neural networks for table recognition. We argue that graph networks are a more natural choice for these problems, and explore two gradient-based graph neural networks. Our proposed architecture combines the benefits of convolutional neural networks for visual feature extraction and graph networks for dealing with the problem structure. We empirically demonstrate that our method outperforms the baseline by a significant margin. In addition, we identify the lack of large scale datasets as a major hindrance for deep learning research for structure analysis and present a new large scale synthetic dataset for the problem of table recognition. Finally, we open-source our implementation of dataset generation and the training framework of our graph networks to promote reproducible research in this direction. | ['Faisal Shafait', 'Shah Rukh Qasim', 'Hassan Mahmood'] | 2019-05-31 | null | null | null | null | ['table-recognition'] | ['computer-vision'] | [ 2.97137231e-01 8.76121148e-02 -1.03987828e-01 -3.01717728e-01
-4.37918156e-01 -7.88377047e-01 5.52010059e-01 4.92167681e-01
-1.50573909e-01 3.35020304e-01 2.93018907e-01 -6.79872155e-01
-1.17768407e-01 -1.19140434e+00 -8.09418261e-01 -2.87006766e-01
-6.73180968e-02 6.29358649e-01 1.88199326e-01 -4.33614582e-01
4.21508998e-01 8.21470916e-01 -1.05876780e+00 6.33904576e-01
5.50587177e-01 9.97095644e-01 -1.75623074e-01 7.59754896e-01
-5.15282571e-01 1.13110530e+00 -7.88766146e-01 -7.36166060e-01
1.87287748e-01 -4.70503181e-01 -1.14266467e+00 5.50393939e-01
6.10080063e-01 -2.09753126e-01 -4.24971074e-01 9.40466762e-01
4.27422464e-01 1.07503533e-01 5.79273939e-01 -1.31813836e+00
-7.26007402e-01 9.98899281e-01 -5.51529169e-01 1.82217821e-01
4.15691435e-01 -1.53901190e-01 1.26890504e+00 -4.92078543e-01
7.99124241e-01 1.31261158e+00 7.42640436e-01 2.52614886e-01
-9.30699468e-01 -2.23524105e-02 2.35384628e-01 4.01887476e-01
-1.09724236e+00 -2.05071643e-01 9.92572367e-01 -2.46036857e-01
9.68792439e-01 2.35785037e-01 5.30699790e-01 1.03672481e+00
1.06103502e-01 1.16460288e+00 6.06978178e-01 -6.26115084e-01
1.02769747e-01 -1.64151579e-01 3.48654956e-01 1.16441083e+00
5.31000197e-01 -5.99597692e-01 -2.05358014e-01 1.21983007e-01
7.51898348e-01 -9.27188024e-02 -1.00375831e-01 -6.66292787e-01
-1.16724932e+00 8.66205513e-01 7.12156951e-01 3.87207597e-01
-9.92327407e-02 1.60154015e-01 5.37636161e-01 4.37775493e-01
1.79123446e-01 6.62817955e-01 -1.47799626e-01 -1.84392929e-02
-9.42597508e-01 3.17771912e-01 9.47027087e-01 9.27217543e-01
4.01663452e-01 2.09647030e-01 -3.89738172e-01 6.39252484e-01
1.99979708e-01 1.57970050e-03 1.55501381e-01 -5.33296227e-01
7.34413207e-01 1.28026223e+00 -2.77510494e-01 -1.34988749e+00
-6.04856610e-01 -2.86786795e-01 -1.03073299e+00 1.58057660e-01
8.22941005e-01 8.52035731e-02 -9.51545775e-01 1.02275383e+00
4.97096637e-03 -6.10559762e-01 -5.70286959e-02 5.64687490e-01
8.28172326e-01 5.94252050e-01 -5.94131052e-01 1.93339780e-01
1.34208322e+00 -1.10596299e+00 -7.43816435e-01 -2.51963407e-01
7.05216110e-01 -6.65620029e-01 1.17122924e+00 4.48756665e-01
-1.04669714e+00 -2.76129514e-01 -1.16241372e+00 -1.66097730e-01
-7.82504022e-01 1.33473486e-01 1.02368557e+00 7.81649053e-01
-1.20854771e+00 6.62912071e-01 -6.58456445e-01 -7.20939100e-01
7.84639001e-01 3.81594539e-01 -4.25185263e-01 -5.32035083e-02
-7.70303011e-01 5.70925474e-01 3.75465572e-01 3.61586839e-01
-3.74190599e-01 6.45795166e-02 -9.44498301e-01 3.70682746e-01
6.77452445e-01 -3.99704754e-01 9.97756720e-01 -6.19256377e-01
-1.00124407e+00 9.92129266e-01 2.20661566e-01 -6.80236816e-01
5.21800637e-01 -1.59835592e-02 -8.35982114e-02 1.26431435e-01
-1.74407959e-01 3.06767493e-01 6.78293586e-01 -1.06864178e+00
-1.92428872e-01 -4.17488277e-01 1.21484756e-01 4.08587791e-02
-4.09661382e-01 6.77565113e-02 -6.65129364e-01 -6.34553611e-01
1.09201178e-01 -6.60080910e-01 -2.37191707e-01 -1.43198535e-01
-9.45409060e-01 -1.59204870e-01 6.58155620e-01 -4.71707910e-01
1.25055087e+00 -1.82430315e+00 -9.18752924e-02 4.02284652e-01
4.09283996e-01 2.67366529e-01 -2.15982199e-01 7.07228780e-01
4.76165153e-02 4.50452447e-01 -3.77774239e-01 -3.29790786e-02
1.18646160e-01 -9.95301381e-02 -2.45510712e-01 3.70022774e-01
3.54478598e-01 1.20374203e+00 -6.01822138e-01 -4.69538331e-01
-5.46649285e-02 2.42450342e-01 -4.78077739e-01 7.72156417e-02
-5.23316503e-01 -2.71646529e-01 -3.54533911e-01 7.32392073e-01
6.77071095e-01 -5.93846381e-01 5.33955514e-01 -1.07347481e-01
3.42175752e-01 1.93196371e-01 -1.42813742e+00 1.48828197e+00
8.89872834e-02 8.15724909e-01 2.75790747e-02 -1.33776295e+00
1.03800380e+00 -3.30241211e-02 2.65257806e-01 -7.60007143e-01
3.31269830e-01 -9.53740850e-02 2.68237770e-01 -3.28999788e-01
7.98237324e-01 4.02748376e-01 -9.41959694e-02 6.75290585e-01
-1.93234291e-02 -1.96951494e-01 6.29589677e-01 4.67741519e-01
1.29141366e+00 -8.91970098e-02 1.24857821e-01 -1.75718457e-01
4.50125635e-01 2.01220259e-01 3.52784432e-02 8.59002471e-01
-1.76953420e-01 7.31451154e-01 1.11551595e+00 -8.45889211e-01
-1.03862309e+00 -6.03704333e-01 2.65773743e-01 9.52265620e-01
-1.07297160e-01 -6.28119707e-01 -1.10460579e+00 -8.86580765e-01
-4.53952514e-02 2.64416933e-01 -7.16419816e-01 7.25778788e-02
-7.96447337e-01 -8.40025067e-01 7.30709672e-01 7.32336223e-01
5.82089663e-01 -1.47737992e+00 -3.84431809e-01 2.17364728e-01
9.94147733e-02 -1.26793635e+00 -2.61455506e-01 3.74142021e-01
-8.81236196e-01 -1.39364922e+00 -4.43116695e-01 -1.13334227e+00
6.73944473e-01 2.25995615e-01 1.40350759e+00 3.98464143e-01
-2.64529914e-01 2.57909834e-01 -3.24460804e-01 -4.46662456e-01
-5.74036658e-01 5.29676318e-01 -5.35878420e-01 4.54812236e-02
3.49560261e-01 -2.02972725e-01 -3.65971595e-01 1.65235344e-02
-1.14115238e+00 -5.37604243e-02 4.45799470e-01 7.28847742e-01
3.60105932e-01 2.73158252e-01 1.42508388e-01 -1.40881681e+00
1.29208887e+00 -1.19811065e-01 -7.81521440e-01 4.08368319e-01
-5.74823618e-01 4.99986470e-01 7.11818218e-01 5.89141101e-02
-5.92617631e-01 2.85828430e-02 -2.13207766e-01 9.90338251e-02
-1.35358423e-01 7.45779037e-01 -2.86934137e-01 8.96353945e-02
6.83584571e-01 1.69684604e-01 -1.16088071e-04 -4.80276912e-01
4.61981207e-01 4.52409297e-01 3.22644502e-01 -4.87900764e-01
7.60830402e-01 5.74096203e-01 1.66674882e-01 -5.74129760e-01
-8.22412789e-01 -2.01459870e-01 -8.14084411e-01 4.24678996e-02
9.81171548e-01 -4.18581188e-01 -7.15005100e-01 5.79200387e-01
-1.12768400e+00 -4.59076732e-01 -1.15962669e-01 -2.23003298e-01
-4.74334925e-01 4.89418477e-01 -7.67626345e-01 -6.18042111e-01
-4.08957571e-01 -1.19815969e+00 1.02350271e+00 1.97264418e-01
-4.20315266e-02 -1.26181209e+00 -8.06736201e-02 4.04212922e-01
4.38204497e-01 6.25974834e-01 1.29398143e+00 -7.17267454e-01
-8.97555292e-01 -3.43608767e-01 -5.05648673e-01 1.42099261e-02
1.70582294e-01 3.40890676e-01 -8.07119012e-01 -2.14127108e-01
-4.04480994e-01 -4.96803969e-01 1.10035360e+00 2.95504630e-01
1.38857102e+00 -2.25260928e-01 -1.89253286e-01 6.77238703e-01
1.33251309e+00 1.77088797e-01 8.21427464e-01 8.32053363e-01
1.19940352e+00 6.54327631e-01 4.65755258e-03 3.21801096e-01
4.17112231e-01 5.20712376e-01 6.39546573e-01 -4.24797297e-01
-4.53964025e-02 -2.15351388e-01 -2.88862325e-02 6.07955694e-01
8.48229080e-02 -7.75896132e-01 -1.31141102e+00 4.31291074e-01
-1.81732571e+00 -8.96677852e-01 -2.54150242e-01 1.71021008e+00
4.58704352e-01 5.86290598e-01 3.11083674e-01 7.54155397e-01
6.51005983e-01 3.80206674e-01 -1.83478341e-01 -7.86718071e-01
-2.87527472e-01 1.98180392e-01 3.81390989e-01 2.88455427e-01
-1.30784631e+00 1.13260555e+00 6.51286316e+00 5.42992592e-01
-1.00438046e+00 -6.90440536e-01 8.43059182e-01 3.91395241e-01
-2.83252329e-01 -3.16969872e-01 -6.45386159e-01 -2.86199041e-02
8.16384435e-01 6.63128961e-03 5.16474605e-01 8.48217666e-01
-2.71978587e-01 1.99871436e-01 -1.24070537e+00 9.85240519e-01
1.18491292e-01 -1.84340823e+00 4.41693813e-01 1.63698524e-01
4.21809971e-01 -2.57228702e-01 5.48036136e-02 1.14835493e-01
3.93632174e-01 -1.36191356e+00 7.24662900e-01 1.03922568e-01
4.74126816e-01 -8.86472523e-01 6.98703825e-01 -3.41881625e-03
-1.22389650e+00 7.64941871e-02 -4.58204746e-01 -1.17867552e-01
-2.87518352e-01 3.01066339e-01 -1.05425107e+00 4.81543690e-01
6.64179862e-01 7.02534735e-01 -1.14858842e+00 1.02233219e+00
-1.71845436e-01 6.94003284e-01 -1.54831447e-02 -3.24132591e-01
5.25887609e-01 -2.17710525e-01 2.88887948e-01 1.34386027e+00
-1.67427063e-01 -5.47506511e-01 4.35269251e-02 9.20975089e-01
-5.40416658e-01 2.06728131e-01 -9.76255596e-01 -5.53829849e-01
-4.08545509e-02 1.34382939e+00 -1.58700621e+00 -1.93774492e-01
-4.46653545e-01 9.09724355e-01 6.35828137e-01 2.20353648e-01
-3.38592738e-01 -8.26386154e-01 8.99722725e-02 1.75891802e-01
5.08770406e-01 -3.59511584e-01 -5.36655605e-01 -1.19713104e+00
1.58237383e-01 -1.25110495e+00 6.44859612e-01 -5.57850242e-01
-1.11767185e+00 9.35688317e-01 -3.01973492e-01 -8.46774280e-01
-2.93555528e-01 -1.00183403e+00 -6.96562171e-01 5.29745281e-01
-1.29957151e+00 -1.05030215e+00 -4.15219873e-01 6.92574620e-01
3.89927149e-01 -2.79073536e-01 7.20498621e-01 -4.43579582e-03
-6.53784931e-01 7.45530784e-01 1.76058099e-01 9.42670763e-01
3.16637844e-01 -1.70584357e+00 1.12100446e+00 1.07925999e+00
7.00009465e-01 4.89873141e-01 4.33385640e-01 -4.44462597e-01
-1.66964173e+00 -9.87621844e-01 7.50818133e-01 -6.44662797e-01
7.81536937e-01 -1.02788639e+00 -1.03861427e+00 7.39519000e-01
3.50995511e-01 -1.02769949e-01 4.69717294e-01 1.15431152e-01
-3.69827271e-01 -6.91001341e-02 -1.02696121e+00 7.83259988e-01
1.02993548e+00 -4.55133379e-01 -3.48831058e-01 4.91188228e-01
5.84336281e-01 -4.25885916e-01 -5.51452279e-01 2.89897434e-02
2.68505633e-01 -1.16283453e+00 7.78427601e-01 -6.30290389e-01
6.79016531e-01 -1.59584552e-01 -3.58056091e-02 -9.92336214e-01
-3.03752780e-01 -6.30506337e-01 -8.39020088e-02 1.21870565e+00
6.26523137e-01 -2.09478095e-01 1.49023771e+00 2.04061881e-01
4.06408533e-02 -7.91800082e-01 -3.52611601e-01 -7.07510591e-01
1.50400236e-01 -4.15170461e-01 7.65330434e-01 9.08278644e-01
-1.57134205e-01 8.00106943e-01 -1.27206564e-01 -2.11840793e-01
4.47761983e-01 4.80634153e-01 8.95502448e-01 -1.22789395e+00
-2.00621516e-01 -7.16709197e-01 -4.58732069e-01 -8.34482610e-01
2.40561161e-02 -1.11673069e+00 -3.25441360e-01 -2.39789414e+00
9.28360373e-02 1.21978978e-02 -1.31377548e-01 3.78103256e-01
-1.40396561e-02 1.97332099e-01 3.68729115e-01 -7.86969662e-02
-6.58969402e-01 2.29371160e-01 1.45767784e+00 -7.19234288e-01
-1.57201320e-01 -1.35315701e-01 -1.06557798e+00 4.40158784e-01
8.41521084e-01 -2.73145288e-01 -4.50574696e-01 -5.76762259e-01
5.91135561e-01 -8.20795894e-02 -7.63750225e-02 -9.26168919e-01
8.47012773e-02 3.58988419e-02 6.93882167e-01 -7.36716926e-01
-1.51337460e-01 -6.88091457e-01 -5.70602655e-01 3.41617107e-01
-3.66088510e-01 4.40008610e-01 3.46474528e-01 3.39048773e-01
-3.17317665e-01 -1.22811645e-01 4.32503819e-01 -4.32230234e-01
-6.32726550e-01 3.30380648e-01 -5.06694019e-01 1.57132372e-01
5.26063204e-01 -5.46666086e-01 -5.01795232e-01 -5.09149075e-01
-4.80419397e-01 1.84493512e-01 5.52218735e-01 4.12429601e-01
7.37212002e-01 -1.14989340e+00 -5.88220298e-01 2.24536180e-01
2.26717979e-01 1.09399647e-01 -4.30689842e-01 2.56098688e-01
-1.17516112e+00 4.97814983e-01 -4.84569639e-01 -4.33666706e-01
-1.17190874e+00 8.15879583e-01 2.76841015e-01 -8.47575128e-01
-6.21266127e-01 8.38931620e-01 -9.46733281e-02 -4.85479325e-01
4.98376459e-01 -6.55925870e-01 -5.11350751e-01 4.57251258e-02
4.49527055e-01 1.81254491e-01 5.19077241e-01 -2.68223047e-01
-4.82751936e-01 2.23783448e-01 -4.29441035e-01 2.87253112e-01
1.44898748e+00 3.17801565e-01 -1.31724283e-01 2.34843984e-01
9.52317357e-01 1.44546414e-02 -8.32469165e-01 3.01383738e-03
5.79901874e-01 -8.53385255e-02 -1.89017028e-01 -7.52994061e-01
-1.07591188e+00 9.82361257e-01 2.26524383e-01 8.38517547e-01
1.04643130e+00 -2.56427199e-01 5.78002036e-01 9.35949326e-01
-4.45919717e-03 -1.10077250e+00 4.46107745e-01 7.90115833e-01
9.16540444e-01 -1.37424564e+00 1.49726748e-01 -4.19821024e-01
-4.06188071e-01 1.42387235e+00 5.95814168e-01 -2.74487555e-01
3.24678838e-01 3.99632603e-01 2.09902108e-01 -5.97288311e-01
-6.06329441e-01 -4.13226008e-01 3.97106141e-01 7.09852636e-01
7.48174012e-01 -1.85571074e-01 6.78262732e-04 3.25564533e-01
-3.49663824e-01 -3.18410546e-01 8.87091398e-01 1.25648475e+00
-2.48454437e-01 -1.28854835e+00 -4.30707186e-01 6.21972084e-01
-6.66034281e-01 -2.16238618e-01 -1.12100005e+00 9.66371298e-01
-3.85935247e-01 7.85648286e-01 5.23679331e-02 -3.60081702e-01
4.24674332e-01 -2.30395384e-02 6.14329875e-01 -6.18142307e-01
-7.91169167e-01 -1.29068494e-01 5.11368476e-02 -4.91944671e-01
-1.74174264e-01 -2.16862366e-01 -1.07973635e+00 -4.72602010e-01
4.36426364e-02 2.18543913e-02 5.17696559e-01 6.21356547e-01
2.59826183e-01 7.06164002e-01 2.81716764e-01 -5.01617908e-01
-3.28960150e-01 -6.96425438e-01 -7.47335076e-01 5.95800281e-01
2.08381206e-01 -3.17023903e-01 4.25911956e-02 1.44240990e-01] | [11.668123245239258, 2.924440622329712] |
14e726f8-f297-4984-b95c-88d7fd1d216b | adversarial-self-supervised-learning-for-semi | 2007.05934 | null | https://arxiv.org/abs/2007.05934v1 | https://arxiv.org/pdf/2007.05934v1.pdf | Adversarial Self-Supervised Learning for Semi-Supervised 3D Action Recognition | We consider the problem of semi-supervised 3D action recognition which has been rarely explored before. Its major challenge lies in how to effectively learn motion representations from unlabeled data. Self-supervised learning (SSL) has been proved very effective at learning representations from unlabeled data in the image domain. However, few effective self-supervised approaches exist for 3D action recognition, and directly applying SSL for semi-supervised learning suffers from misalignment of representations learned from SSL and supervised learning tasks. To address these issues, we present Adversarial Self-Supervised Learning (ASSL), a novel framework that tightly couples SSL and the semi-supervised scheme via neighbor relation exploration and adversarial learning. Specifically, we design an effective SSL scheme to improve the discrimination capability of learned representations for 3D action recognition, through exploring the data relations within a neighborhood. We further propose an adversarial regularization to align the feature distributions of labeled and unlabeled samples. To demonstrate effectiveness of the proposed ASSL in semi-supervised 3D action recognition, we conduct extensive experiments on NTU and N-UCLA datasets. The results confirm its advantageous performance over state-of-the-art semi-supervised methods in the few label regime for 3D action recognition. | ['Jiashi Feng', 'Xuecheng Nie', 'Liang Wang', 'Chenyang Si', 'Tieniu Tan', 'Wei Wang'] | 2020-07-12 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/123_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123520035.pdf | eccv-2020-8 | ['3d-human-action-recognition'] | ['computer-vision'] | [ 4.62794751e-01 -6.02646172e-03 -7.50280261e-01 -4.24662352e-01
-6.87486887e-01 -3.50169003e-01 6.27390087e-01 -7.28968084e-01
-2.00434074e-01 8.01105499e-01 5.17182112e-01 -1.53603449e-01
6.50840998e-02 -2.12898403e-01 -6.44218147e-01 -9.62224722e-01
4.79225479e-02 3.19891751e-01 1.32856384e-01 1.39387503e-01
7.90719092e-02 6.74501240e-01 -1.29115200e+00 1.18293114e-01
7.81995118e-01 9.51141536e-01 -2.17771739e-01 3.63089353e-01
-1.46789819e-01 1.36042452e+00 -4.33065057e-01 3.58925223e-01
4.84651059e-01 -7.12527096e-01 -9.48154986e-01 7.12529302e-01
4.14963663e-01 -3.85449946e-01 -7.63233185e-01 9.55534577e-01
4.32042390e-01 1.61568582e-01 1.18826520e+00 -1.49033892e+00
-6.99511051e-01 1.54263869e-01 -7.36321032e-01 3.56781960e-01
3.66929978e-01 1.81395575e-01 5.53895712e-01 -8.46821427e-01
6.38135195e-01 1.33389580e+00 3.34544599e-01 8.28455687e-01
-9.49801385e-01 -5.94960749e-01 1.12098172e-01 1.73861846e-01
-1.21164620e+00 -4.29784328e-01 1.04866135e+00 -4.69371498e-01
8.41308057e-01 -6.67473078e-02 4.37355161e-01 1.37457645e+00
-3.38329598e-02 1.36387038e+00 1.37547255e+00 -4.90482509e-01
3.36065590e-01 -2.30978400e-01 -4.39514145e-02 7.59047806e-01
-1.21861428e-01 4.61549014e-01 -6.72669649e-01 6.92033470e-02
1.09552824e+00 1.26438603e-01 -8.40537250e-02 -9.04740751e-01
-1.28844500e+00 5.72112381e-01 3.84107679e-01 1.63685292e-01
-3.27928290e-02 -9.04893968e-03 4.82475370e-01 2.43289888e-01
7.07981467e-01 1.08958542e-01 -1.86556786e-01 -1.82800800e-01
-5.46090901e-01 -2.32977465e-01 2.08293378e-01 1.01779628e+00
5.56215286e-01 4.70488995e-01 -2.05689877e-01 7.28296340e-01
4.22956377e-01 7.64820457e-01 7.96536744e-01 -8.24122369e-01
4.95812058e-01 8.19382489e-01 -9.40511823e-02 -6.81948841e-01
-2.09680364e-01 -1.54052719e-01 -9.36920047e-01 4.40501481e-01
4.38665926e-01 -1.14991173e-01 -1.12313437e+00 1.43555486e+00
4.21360046e-01 4.70048219e-01 4.99179780e-01 1.05939162e+00
8.11566889e-01 4.92267728e-01 1.48544610e-01 -3.53761166e-01
5.28350830e-01 -1.25800729e+00 -6.77292287e-01 -3.18231106e-01
9.98330891e-01 -3.87022823e-01 7.95397401e-01 -8.54983330e-02
-6.13395810e-01 -8.00618768e-01 -1.11566567e+00 8.65174383e-02
-1.90997168e-01 2.35882133e-01 5.54517031e-01 4.58541900e-01
-4.60320354e-01 4.72786039e-01 -1.10188055e+00 -3.24262977e-01
9.33997571e-01 2.49830335e-01 -8.05512667e-01 -2.31940255e-01
-1.03708124e+00 6.97573245e-01 3.62829089e-01 5.42972758e-02
-1.18009555e+00 -1.34605214e-01 -1.16953218e+00 -5.30739486e-01
6.14602387e-01 -1.81990847e-01 9.08169568e-01 -1.08710384e+00
-1.52097940e+00 1.21466231e+00 5.03930636e-02 -3.82991284e-01
3.77582997e-01 -1.93740129e-01 -4.58783567e-01 2.55989909e-01
2.80051172e-01 5.88311076e-01 1.19370711e+00 -1.03287601e+00
-2.62585074e-01 -7.89832473e-01 -1.23100415e-01 6.03790045e-01
-2.11632997e-01 -2.56730169e-01 -2.08662748e-01 -9.30508912e-01
4.98273909e-01 -1.11046636e+00 -2.51959950e-01 2.20519483e-01
-4.14889872e-01 -2.74204761e-01 1.12898207e+00 -1.49646923e-01
6.08214676e-01 -2.39856052e+00 2.22418457e-01 -7.56203234e-02
-1.28575027e-01 6.81077063e-01 -1.60574377e-01 1.10393919e-01
-2.81508505e-01 -2.58665830e-01 -3.67075890e-01 -2.93161243e-01
-1.47490174e-01 6.21876001e-01 -3.20872664e-01 7.07141638e-01
4.10445929e-01 1.22642028e+00 -1.03730977e+00 -7.74674952e-01
4.11816955e-01 1.72435746e-01 -8.14754888e-02 5.90033293e-01
-7.68109411e-02 9.19768035e-01 -8.96207035e-01 1.09379625e+00
4.37522411e-01 -2.52509385e-01 -1.07217893e-01 -2.70080455e-02
4.16596740e-01 -1.17030732e-01 -1.10867691e+00 1.95305371e+00
-4.44953144e-02 3.39054525e-01 -4.03514236e-01 -1.58738291e+00
1.20643461e+00 1.42864719e-01 6.84653938e-01 -5.40493309e-01
3.51316296e-02 1.33376196e-01 -4.40151930e-01 -7.86030710e-01
-1.23064846e-01 -2.51446962e-01 -7.21749812e-02 6.99114084e-01
2.90696502e-01 -7.84475952e-02 -2.87524521e-01 2.36586276e-02
9.34967935e-01 8.68424773e-01 4.63678658e-01 2.28507947e-02
7.03525186e-01 1.40999198e-01 6.35688543e-01 5.27502418e-01
-8.24765265e-01 6.19590163e-01 3.13531578e-01 -4.25928414e-01
-5.99083543e-01 -1.04409432e+00 -1.74518097e-02 6.96558654e-01
3.93340200e-01 1.58883542e-01 -4.13894951e-01 -1.62173259e+00
1.22963257e-01 2.60659993e-01 -7.57394969e-01 -4.40307498e-01
-4.12763804e-01 -4.36501235e-01 6.27555311e-01 9.02364790e-01
9.23800707e-01 -1.18903089e+00 -3.23498666e-01 -1.15595989e-01
2.04292476e-01 -1.32082951e+00 -4.86249238e-01 3.59991580e-01
-1.07479393e+00 -1.28461182e+00 -8.98778796e-01 -9.84741211e-01
1.07082510e+00 4.91116405e-01 6.16188288e-01 -2.97112525e-01
-1.50335148e-01 5.43982267e-01 -6.79704905e-01 -7.59314150e-02
-4.45210934e-01 -1.59692541e-01 5.06401002e-01 2.86258101e-01
6.04243040e-01 -5.30825138e-01 -2.85329640e-01 7.67520249e-01
-9.00976658e-01 -1.82263479e-01 6.83468163e-01 1.07058489e+00
8.54774892e-01 -9.22678411e-03 6.86668873e-01 -9.81466055e-01
-5.70091382e-02 -3.63951057e-01 -1.28187507e-01 2.08054498e-01
-4.55048323e-01 2.37017483e-01 5.46881497e-01 -5.85606933e-01
-1.07584918e+00 6.33050442e-01 5.67414351e-02 -9.94112372e-01
-7.27866888e-01 1.60992652e-01 -4.21351045e-01 -3.02537888e-01
7.31915355e-01 4.29612696e-01 4.31524515e-01 -3.18315983e-01
4.00392413e-01 9.26157176e-01 5.82354903e-01 -3.53738427e-01
8.96028578e-01 7.03418612e-01 1.20792843e-01 -7.73278296e-01
-1.29853320e+00 -5.94247103e-01 -1.18681645e+00 -2.36175239e-01
1.06765151e+00 -9.57976043e-01 -1.12318896e-01 7.20634639e-01
-4.62912083e-01 -4.56869185e-01 -6.76165581e-01 7.93219090e-01
-1.06145978e+00 7.35539436e-01 -1.59377426e-01 -7.62648404e-01
5.39490208e-02 -1.06597853e+00 1.14053702e+00 1.19192570e-01
-3.38646700e-03 -1.02768064e+00 1.31359830e-01 7.98147321e-01
-8.81680325e-02 4.74432528e-01 4.86959130e-01 -7.99481273e-01
-2.93244481e-01 -2.99732029e-01 -5.39932176e-02 7.90097117e-01
6.29359663e-01 -5.37994802e-01 -8.37467909e-01 -2.25445390e-01
4.46727909e-02 -1.13398278e+00 8.97753477e-01 2.24778235e-01
1.04998147e+00 1.27481213e-02 -3.25080186e-01 6.95907354e-01
9.29670572e-01 2.25650996e-01 6.34034395e-01 1.74173668e-01
1.04353368e+00 4.73553896e-01 1.17798555e+00 3.18861902e-01
-1.13285780e-01 5.27653039e-01 3.44412088e-01 -1.81257382e-01
-1.43419847e-01 -6.20716333e-01 4.75273788e-01 6.01944268e-01
-8.34689513e-02 1.42238483e-01 -6.51564181e-01 1.67724729e-01
-2.03083897e+00 -8.64596903e-01 2.87563026e-01 1.93035650e+00
6.41239107e-01 3.12040299e-01 1.67908505e-01 3.21333677e-01
5.37432611e-01 6.19503856e-01 -9.42153990e-01 1.49067968e-01
-2.62677044e-01 -2.62463707e-02 4.57131386e-01 1.92275450e-01
-1.76555967e+00 1.09083664e+00 6.08632088e+00 9.47580397e-01
-8.36453319e-01 -1.71376601e-01 6.24184668e-01 1.69765055e-01
2.92939454e-01 -1.66092038e-01 -7.63028562e-01 1.68510959e-01
2.97953010e-01 1.99519128e-01 -6.31064847e-02 1.07916474e+00
2.71826148e-01 -8.15521255e-02 -1.23522484e+00 1.19996870e+00
4.67811018e-01 -1.06505764e+00 2.68383145e-01 -9.80759263e-02
1.06959856e+00 -2.63704538e-01 -6.13481812e-02 4.18881953e-01
1.02006190e-01 -1.11547887e+00 1.21537194e-01 4.60408121e-01
9.43766594e-01 -5.84093213e-01 5.96911371e-01 6.22652471e-01
-1.02480888e+00 -7.99527094e-02 -3.66910487e-01 -6.88055828e-02
1.79400414e-01 4.65320423e-02 -4.57170278e-01 5.14909208e-01
4.03919786e-01 1.53309929e+00 -4.96321976e-01 4.99123245e-01
-4.63389307e-01 5.85608542e-01 1.58126503e-01 1.21240079e-01
4.99216825e-01 -1.91026196e-01 4.83278573e-01 6.60703957e-01
-1.37731448e-01 3.92943740e-01 4.63291109e-01 5.56207001e-01
1.67502537e-01 7.31939971e-02 -1.11421001e+00 -3.84681284e-01
-1.33943977e-02 5.59967756e-01 -5.45607388e-01 -2.15835661e-01
-4.02370781e-01 1.18041658e+00 3.78942907e-01 3.98043931e-01
-5.90294302e-01 2.97145341e-02 4.56949562e-01 -3.06419395e-02
2.50721127e-01 -4.40267861e-01 -1.38658918e-02 -1.29697585e+00
-7.72216395e-02 -9.49746430e-01 4.40297902e-01 -7.87870228e-01
-1.37666011e+00 3.43008041e-01 6.85003325e-02 -1.95027852e+00
-4.13498521e-01 -6.96050406e-01 -3.51272702e-01 3.13957423e-01
-1.36820543e+00 -1.19176650e+00 -2.07266971e-01 9.64053214e-01
7.90108323e-01 -7.15154111e-01 7.82831073e-01 -1.82864014e-02
-5.58883429e-01 5.55217147e-01 1.87831342e-01 3.38629603e-01
7.10918188e-01 -1.09676158e+00 1.83196813e-01 7.07671463e-01
6.12602592e-01 -7.78336897e-02 9.25209224e-02 -7.29487300e-01
-1.51360750e+00 -1.24406409e+00 3.57062578e-01 -4.19659555e-01
4.41739261e-01 -3.91717069e-02 -7.41124928e-01 6.81393027e-01
-3.72660667e-01 6.77821696e-01 7.61927128e-01 -4.44571525e-01
-2.10388660e-01 1.61785949e-02 -9.67446387e-01 3.94915253e-01
1.57105660e+00 -6.77702427e-01 -7.77348518e-01 6.33600950e-01
5.06108165e-01 -4.17813838e-01 -6.76105440e-01 7.44908035e-01
2.77274877e-01 -8.28712881e-01 1.06038761e+00 -1.07864153e+00
3.50753367e-01 -3.95196229e-01 -2.81448036e-01 -1.22395968e+00
-4.48880810e-03 -4.40205306e-01 -3.25773209e-01 8.90505731e-01
-3.32138548e-03 -3.32413703e-01 1.29178238e+00 1.80579782e-01
-7.15032667e-02 -8.24440539e-01 -9.51621830e-01 -1.10052896e+00
5.39813831e-04 -2.64029294e-01 6.57606591e-03 1.05924845e+00
5.59517108e-02 2.76268691e-01 -7.08475292e-01 5.38623929e-02
9.01162803e-01 1.94850087e-01 9.98193562e-01 -7.13403106e-01
-3.75551432e-01 8.17081332e-02 -9.48512673e-01 -1.52161348e+00
6.99864864e-01 -9.01547015e-01 1.54799625e-01 -1.30582607e+00
6.65141568e-02 -4.11819965e-01 -2.25810379e-01 6.39341533e-01
-1.16851725e-01 4.50528890e-01 -1.22639284e-01 5.43778539e-01
-7.91249931e-01 1.08526707e+00 1.67007244e+00 -4.56338137e-01
-1.75265789e-01 2.35175803e-01 -1.63838327e-01 8.13863099e-01
5.91534197e-01 -3.83913249e-01 -7.71277189e-01 -1.82117909e-01
-7.51511455e-01 1.73243061e-01 2.55087346e-01 -1.02702725e+00
2.60215178e-02 -3.59929353e-01 5.04159749e-01 -6.77251697e-01
1.76523432e-01 -8.12502742e-01 -3.79258156e-01 3.14337999e-01
-5.52932441e-01 -6.78196311e-01 -1.15297154e-01 9.06280279e-01
-4.86445695e-01 -1.42543510e-01 7.93109357e-01 -1.68524772e-01
-1.26015329e+00 5.34688234e-01 -1.87964842e-01 3.10350448e-01
1.38177168e+00 -4.24201012e-01 1.49752563e-02 -3.19056600e-01
-1.14165163e+00 3.03377271e-01 2.85376906e-01 5.08582056e-01
8.82400215e-01 -1.73342478e+00 -4.34201151e-01 6.53930128e-01
4.44767147e-01 9.60618705e-02 2.05129936e-01 5.77694893e-01
-3.27177554e-01 3.25197428e-01 -5.02333105e-01 -8.18491697e-01
-1.12378919e+00 6.80240870e-01 3.43856663e-01 -1.22005448e-01
-5.95973492e-01 6.83005989e-01 1.14817590e-01 -6.01225734e-01
5.29810786e-01 4.15929332e-02 -2.70539612e-01 -3.06600124e-01
3.24639320e-01 3.53562862e-01 -4.50738400e-01 -9.93294775e-01
-4.71084535e-01 8.02326679e-01 9.31441560e-02 1.44094735e-01
1.01774800e+00 -6.06917180e-02 4.23462570e-01 6.01627290e-01
1.49822354e+00 -5.92368245e-01 -1.74205899e+00 -4.90443975e-01
-1.71837494e-01 -6.03794038e-01 -1.57130241e-01 -2.69990951e-01
-9.98841345e-01 8.26342642e-01 7.01445818e-01 -2.51172125e-01
9.93758500e-01 1.66321620e-01 5.71119368e-01 5.01317322e-01
4.44258690e-01 -1.07563543e+00 7.37674057e-01 4.56093132e-01
6.76729023e-01 -1.76352012e+00 1.19210705e-01 -4.57189381e-01
-9.98057067e-01 9.26529467e-01 9.46824074e-01 -3.48746181e-01
7.77262986e-01 -2.00795941e-03 3.98239702e-01 -1.69037566e-01
-2.07037523e-01 -4.38046694e-01 3.91696572e-01 8.28896403e-01
1.75385941e-02 -1.64746732e-01 2.73234341e-02 2.16353104e-01
4.89642650e-01 1.12548202e-01 1.45577967e-01 1.35626292e+00
-2.62111813e-01 -1.06290984e+00 -2.14087650e-01 2.85814852e-01
-4.46117185e-02 4.79184389e-01 -5.25219321e-01 7.57345974e-01
-4.30009253e-02 7.05329776e-01 1.72658742e-03 -5.31466067e-01
2.27680370e-01 9.18417871e-02 5.26860416e-01 -7.48516142e-01
2.84660786e-01 6.61004335e-02 -2.60357074e-02 -7.13688076e-01
-1.20316696e+00 -7.21498430e-01 -1.38019848e+00 4.90283489e-01
-1.77793100e-01 -2.93587539e-02 8.68881270e-02 1.20014417e+00
1.82907507e-01 1.54402316e-01 1.16243231e+00 -8.79628360e-01
-9.98319328e-01 -9.13388073e-01 -9.82560992e-01 6.78772807e-01
2.08055779e-01 -1.15941191e+00 -5.67488670e-01 1.47836789e-01] | [8.138470649719238, 0.657090961933136] |
e7a9a292-9fc8-4656-90a7-631c069a42f6 | kgplm-knowledge-guided-language-model-pre | 2012.03551 | null | https://arxiv.org/abs/2012.03551v1 | https://arxiv.org/pdf/2012.03551v1.pdf | KgPLM: Knowledge-guided Language Model Pre-training via Generative and Discriminative Learning | Recent studies on pre-trained language models have demonstrated their ability to capture factual knowledge and applications in knowledge-aware downstream tasks. In this work, we present a language model pre-training framework guided by factual knowledge completion and verification, and use the generative and discriminative approaches cooperatively to learn the model. Particularly, we investigate two learning schemes, named two-tower scheme and pipeline scheme, in training the generator and discriminator with shared parameter. Experimental results on LAMA, a set of zero-shot cloze-style question answering tasks, show that our model contains richer factual knowledge than the conventional pre-trained language models. Furthermore, when fine-tuned and evaluated on the MRQA shared tasks which consists of several machine reading comprehension datasets, our model achieves the state-of-the-art performance, and gains large improvements on NewsQA (+1.26 F1) and TriviaQA (+1.56 F1) over RoBERTa. | ['Qun Liu', 'Jinghui Xiao', 'Xin Jiang', 'Bin He'] | 2020-12-07 | null | null | null | null | ['triviaqa'] | ['miscellaneous'] | [ 7.48119205e-02 3.26486647e-01 2.23250855e-02 -4.96033013e-01
-1.53686714e+00 -7.67309904e-01 7.40393937e-01 7.81468004e-02
-3.90656978e-01 8.69158745e-01 4.24075365e-01 -7.11881518e-01
-6.27832934e-02 -8.66673350e-01 -9.71677184e-01 -3.43127996e-01
3.26335043e-01 6.72300994e-01 3.39291751e-01 -6.16560638e-01
1.72664046e-01 -2.75799602e-01 -1.16962051e+00 7.67067075e-01
1.36996508e+00 8.16404402e-01 2.46452764e-01 9.70812261e-01
-3.41109723e-01 1.64819241e+00 -6.27378106e-01 -1.02363575e+00
-2.86280125e-01 -6.39574409e-01 -1.44093084e+00 -4.20221150e-01
6.30030930e-01 -5.26610672e-01 -1.80358499e-01 6.45525634e-01
5.71898937e-01 3.09914887e-01 5.70351720e-01 -7.61160612e-01
-1.38819075e+00 9.39282894e-01 -2.30494007e-01 2.75155306e-01
5.91262460e-01 2.87626207e-01 1.35859537e+00 -9.24397886e-01
3.25867265e-01 1.33712935e+00 3.43327016e-01 8.96513581e-01
-8.39152396e-01 -4.86120611e-01 2.49016091e-01 5.44179261e-01
-1.09920943e+00 -5.46136081e-01 5.57749927e-01 -3.40532094e-01
1.25752974e+00 -5.60213551e-02 -9.07045789e-03 1.33705461e+00
2.61935852e-02 1.26254523e+00 1.25092828e+00 -5.92975259e-01
9.53793600e-02 3.16653252e-02 5.43403029e-01 1.05143833e+00
-7.25541264e-02 -2.21288264e-01 -8.11762631e-01 -2.17262328e-01
3.50723803e-01 -4.21470076e-01 -5.28974354e-01 -7.42114335e-02
-1.06266594e+00 8.80210161e-01 1.60454273e-01 7.22149909e-02
-3.08942616e-01 4.03461382e-02 2.42531016e-01 5.54830968e-01
4.57129538e-01 7.07458317e-01 -7.92051494e-01 -3.19549888e-01
-7.99350560e-01 2.98697710e-01 9.82664287e-01 1.09081817e+00
4.91172612e-01 -1.14074245e-01 -8.64545405e-01 1.01819384e+00
1.95574462e-01 7.12230384e-01 6.60111785e-01 -9.71178472e-01
7.70449102e-01 7.02194273e-01 7.29818717e-02 -2.53648639e-01
-8.51266906e-02 -4.84787166e-01 -7.19293416e-01 -5.68172872e-01
4.26633000e-01 -4.10615429e-02 -1.07580638e+00 1.83769441e+00
2.94758696e-02 1.99558735e-01 7.11339474e-01 5.88150978e-01
1.21561956e+00 8.23146462e-01 5.10656595e-01 -4.76756655e-02
1.50745714e+00 -1.43696964e+00 -8.53508890e-01 -2.47647554e-01
8.40039194e-01 -6.38134360e-01 1.68360567e+00 5.91196299e-01
-1.40652561e+00 -6.38743937e-01 -7.66194463e-01 -6.03651702e-01
-4.12241101e-01 1.66232288e-01 4.98947710e-01 5.40927231e-01
-1.01216662e+00 6.28567412e-02 -5.45796692e-01 6.28565345e-03
5.85576773e-01 -2.81727016e-01 -1.70160066e-02 -5.54666638e-01
-1.40696037e+00 9.16235387e-01 4.87719625e-01 -1.65737301e-01
-1.64186621e+00 -9.77374375e-01 -8.51990223e-01 2.44843960e-01
6.36622250e-01 -1.00072360e+00 1.78146470e+00 -4.97404277e-01
-1.74905550e+00 1.03271532e+00 -4.30238158e-01 -5.20946085e-01
3.72503251e-01 -7.87608027e-01 -3.64890635e-01 1.64760947e-02
1.57606885e-01 4.77877885e-01 4.91242379e-01 -1.06899726e+00
-4.75665331e-01 -1.21652342e-01 5.78709424e-01 2.45294407e-01
-7.23013431e-02 -7.52105713e-02 -2.51030266e-01 -3.64326864e-01
-1.96101829e-01 -2.70550013e-01 1.39634416e-01 -6.04121745e-01
-3.53087634e-01 -8.70262742e-01 2.71325797e-01 -1.04593945e+00
1.33884239e+00 -1.63088155e+00 9.83345136e-02 -3.04500014e-01
1.77910373e-01 5.45030177e-01 -4.81449723e-01 4.74192083e-01
3.56088787e-01 2.99868695e-02 -3.91919136e-01 -2.71933854e-01
1.34571984e-01 3.44930202e-01 -9.22449708e-01 -3.56030315e-01
4.70900208e-01 1.44058371e+00 -1.28523660e+00 -3.28084022e-01
-3.28955024e-01 5.35885468e-02 -4.72268432e-01 9.10731435e-01
-8.83288562e-01 2.21786261e-01 -5.93438923e-01 7.19172776e-01
4.06362593e-01 -5.84262192e-01 1.61881968e-02 2.67013073e-01
3.90571028e-01 9.60909307e-01 -5.44982374e-01 2.31660390e+00
-7.11841106e-01 1.76866800e-01 -2.95997709e-01 -5.75518250e-01
8.68750691e-01 5.09838581e-01 -4.39362973e-01 -8.61337125e-01
1.81230381e-02 1.34816304e-01 4.53215167e-02 -8.75387907e-01
5.35508037e-01 -7.97345415e-02 -2.65034311e-03 6.60478175e-01
5.64166963e-01 -1.31255284e-01 1.22834101e-01 6.67544603e-01
1.21452367e+00 3.75428259e-01 2.57680357e-01 -1.84714451e-01
8.78052235e-01 1.24772340e-01 1.97409764e-01 1.04046404e+00
-1.63693577e-01 4.20675457e-01 5.10258555e-01 5.67533411e-02
-5.55731297e-01 -1.26940656e+00 2.08538041e-01 1.80210435e+00
-4.73027267e-02 -3.50065023e-01 -7.80552149e-01 -1.05459273e+00
-3.42402697e-01 1.37295473e+00 -5.51371455e-01 -3.02329898e-01
-5.38798153e-01 -6.24486208e-01 9.24953222e-01 7.07742274e-01
9.86829042e-01 -1.16077864e+00 -3.20232570e-01 2.20939904e-01
-6.33423924e-01 -1.06002188e+00 -2.00361684e-01 -1.83701143e-01
-6.50437534e-01 -1.09972358e+00 -6.37126565e-01 -8.63167524e-01
3.44754845e-01 2.03767121e-01 1.76294565e+00 1.19661339e-01
2.15943471e-01 6.09757125e-01 -7.13576257e-01 -3.67312789e-01
-4.11501765e-01 2.60371357e-01 -5.77377200e-01 -1.76679909e-01
6.37370825e-01 -2.97044426e-01 -4.35763478e-01 4.29268330e-02
-7.13374436e-01 2.08939493e-01 6.53959930e-01 1.07819593e+00
4.92651433e-01 -5.71628511e-01 1.07097328e+00 -1.16390669e+00
1.21260345e+00 -7.43945956e-01 -2.76309431e-01 1.01227653e+00
-5.22081316e-01 1.62540257e-01 5.90522826e-01 -2.53335685e-01
-1.75860906e+00 -4.79016453e-01 -2.37855479e-01 -8.21200609e-02
-2.45984301e-01 6.30332828e-01 -3.88922513e-01 4.33498919e-01
8.53012919e-01 5.94308794e-01 -5.34037650e-01 -6.69697583e-01
1.04762900e+00 7.42699802e-01 8.52403641e-01 -1.01346326e+00
4.86410618e-01 2.93492377e-02 -8.41585755e-01 -2.30012104e-01
-1.70231259e+00 -5.92797577e-01 -5.41956544e-01 2.62892097e-01
9.10058796e-01 -1.26087976e+00 -5.54206133e-01 4.42924082e-01
-1.46531582e+00 -5.37748992e-01 -1.52824163e-01 4.98990193e-02
-5.76982498e-01 7.71413147e-02 -7.42944717e-01 -8.31686258e-01
-8.24353516e-01 -7.67800748e-01 9.68227804e-01 3.89420986e-01
3.17354733e-03 -1.12889755e+00 3.98527116e-01 1.16285038e+00
5.63425779e-01 -3.28584582e-01 1.34679925e+00 -1.04127264e+00
-8.16351891e-01 3.14565718e-01 -1.48358226e-01 5.57557523e-01
-2.41463438e-01 -5.50199389e-01 -1.25697207e+00 -1.43430471e-01
8.05199891e-02 -1.29949188e+00 1.00588405e+00 -1.37250423e-01
1.25003839e+00 -3.46804619e-01 1.12474106e-01 2.69368231e-01
1.11686897e+00 6.42653704e-02 7.41054654e-01 2.65700929e-02
6.63627684e-01 5.71041524e-01 4.83259469e-01 -8.47283602e-02
1.07292235e+00 1.50132224e-01 -1.33840377e-02 3.44578803e-01
-4.54570949e-01 -7.97028661e-01 4.92638528e-01 1.22474301e+00
7.54098594e-02 -4.52966839e-01 -1.09754372e+00 1.01471806e+00
-1.89359212e+00 -9.21556115e-01 -7.42905214e-02 1.72667432e+00
1.56187296e+00 -5.92663027e-02 -3.77349854e-01 -4.54282880e-01
2.17135370e-01 1.80480525e-01 -6.02805376e-01 -5.96027493e-01
-2.90132701e-01 6.78349137e-01 -1.72995433e-01 7.03272939e-01
-7.37185240e-01 1.46554506e+00 6.27376795e+00 1.00798821e+00
-5.10186613e-01 6.81410134e-01 6.26638651e-01 2.20495388e-01
-5.20310700e-01 2.60685086e-02 -7.98184097e-01 1.03138164e-01
1.14080048e+00 -1.99860647e-01 1.93436757e-01 7.09239066e-01
-2.44524926e-01 -1.54956371e-01 -1.13608742e+00 5.75381279e-01
5.24644792e-01 -1.22734833e+00 5.51139653e-01 -6.23149812e-01
1.02311528e+00 -8.97160172e-03 -1.30830392e-01 1.13261080e+00
7.85447478e-01 -1.35861468e+00 4.13020670e-01 6.58897519e-01
5.23359835e-01 -4.79219705e-01 8.04187894e-01 7.41271913e-01
-5.58037102e-01 8.10406059e-02 -4.47464854e-01 -2.39212230e-01
3.53452533e-01 2.71034330e-01 -9.52402472e-01 6.82455778e-01
5.65561295e-01 2.91444927e-01 -6.56711459e-01 6.54167235e-01
-1.07208109e+00 1.38545132e+00 2.31121540e-01 -3.32771331e-01
2.82975167e-01 2.89633125e-01 -3.51357996e-03 1.18550849e+00
-2.44634971e-01 4.92167622e-01 -3.90656181e-02 1.20753264e+00
-6.48450673e-01 1.01875722e-01 -1.08272888e-01 -4.52526510e-02
4.98998344e-01 1.04423642e+00 3.04007947e-01 -7.43352592e-01
-5.31788886e-01 9.92515087e-01 8.61063063e-01 4.73939300e-01
-6.27730966e-01 -5.52307248e-01 2.78262615e-01 -1.28168821e-01
2.19411090e-01 -1.61869720e-01 -6.27093017e-02 -1.45406842e+00
1.21534489e-01 -1.25683749e+00 6.83274329e-01 -1.11815429e+00
-1.66347325e+00 5.38849711e-01 -3.53746377e-02 -3.52640867e-01
-4.17550564e-01 -6.34430528e-01 -7.78913915e-01 1.16164374e+00
-2.03742313e+00 -1.56978893e+00 -3.21721882e-01 8.00194144e-01
7.80950248e-01 -3.84223968e-01 9.98657942e-01 5.51632158e-02
-2.58892149e-01 7.76028872e-01 -2.78653521e-02 4.35512960e-01
6.70442998e-01 -1.41810060e+00 4.32441860e-01 9.13054287e-01
3.84384423e-01 8.03219259e-01 1.63644314e-01 -6.17914379e-01
-1.31763113e+00 -1.01591551e+00 1.34882438e+00 -1.15543807e+00
7.80860066e-01 -3.87144089e-01 -1.41505051e+00 9.79056418e-01
6.19723260e-01 -5.09026229e-01 8.28500688e-01 3.04011762e-01
-8.35777640e-01 1.22028783e-01 -1.01897764e+00 2.96739519e-01
1.11772716e+00 -9.41173255e-01 -1.55744588e+00 4.90852058e-01
1.10778904e+00 -5.39725959e-01 -8.49527121e-01 3.82772624e-01
9.06050652e-02 -6.90595448e-01 7.81030416e-01 -1.18093705e+00
8.10421884e-01 -1.26191467e-01 -7.32173845e-02 -1.30012941e+00
-2.90625483e-01 -6.45380616e-01 -3.65010381e-01 1.45875418e+00
6.35557055e-01 -4.35244650e-01 3.83053929e-01 3.42570931e-01
-2.36587018e-01 -9.18556035e-01 -9.76454198e-01 -6.10915065e-01
6.32210732e-01 -3.04238051e-01 5.16220391e-01 1.00906217e+00
-3.39255296e-02 9.36876893e-01 -1.33464426e-01 1.67729914e-01
2.38733977e-01 2.23975554e-01 6.61120415e-01 -5.49604595e-01
-5.32364905e-01 -2.62307525e-01 2.81193346e-01 -1.63879955e+00
3.13803732e-01 -1.00044024e+00 1.16352797e-01 -1.61993945e+00
4.48571533e-01 -2.82454103e-01 -4.04145688e-01 8.10232878e-01
-7.54932523e-01 -3.59393716e-01 -2.14346275e-01 -1.28838373e-02
-1.07675123e+00 8.36459696e-01 1.33629167e+00 -1.72861174e-01
3.52012366e-02 -2.90027738e-01 -1.00833452e+00 4.45691496e-01
7.17302680e-01 -1.34465694e-01 -8.80804062e-01 -1.06647420e+00
2.66826451e-01 -6.57164603e-02 4.28127080e-01 -5.53732097e-01
3.31158102e-01 -7.15246871e-02 -4.24093194e-02 -4.66719031e-01
2.67820239e-01 -6.27418309e-02 -8.41054559e-01 1.81975424e-01
-8.88097167e-01 -1.15987351e-02 2.21061081e-01 5.10133862e-01
-3.90073985e-01 -3.72500569e-01 3.28435063e-01 -3.43535304e-01
-9.03270066e-01 -1.59326568e-02 2.44403835e-02 9.78065133e-01
5.42489111e-01 5.66215038e-01 -1.02861738e+00 -5.62404990e-01
-3.93103600e-01 7.63447821e-01 -3.29640746e-01 7.57131040e-01
6.25869155e-01 -1.05774963e+00 -1.16759908e+00 -9.01009589e-02
3.87356371e-01 3.21196318e-01 3.98464680e-01 5.20642877e-01
-3.20490569e-01 7.41277754e-01 1.24388926e-01 -3.84673059e-01
-9.02108371e-01 3.07751775e-01 4.41307634e-01 -6.83690012e-01
-7.55494237e-02 1.36865771e+00 2.21122816e-01 -7.20919907e-01
1.80395469e-01 -3.00237149e-01 -2.74937361e-01 -2.75655478e-01
8.85965466e-01 3.01952541e-01 1.10823981e-01 -1.48557752e-01
-1.15160950e-01 2.87761241e-01 -4.66794193e-01 5.15044015e-03
9.98417139e-01 -3.12842950e-02 -1.88410416e-01 3.91597927e-01
7.97780275e-01 -5.12581244e-02 -1.07117391e+00 -6.75269365e-01
8.50454047e-02 -5.44842845e-03 -7.88419396e-02 -1.84788895e+00
-4.63307261e-01 1.35189056e+00 1.71535864e-01 -1.46480739e-01
9.57161248e-01 3.62612277e-01 9.61502254e-01 7.99777448e-01
3.70011181e-01 -7.52667248e-01 2.99452990e-01 9.19263065e-01
1.08999753e+00 -1.36246347e+00 -4.19694126e-01 -2.78530806e-01
-7.99300194e-01 6.50404751e-01 1.10607839e+00 1.16743229e-01
1.92690790e-01 -3.47240508e-01 3.58119130e-01 -2.44777992e-01
-1.32380760e+00 -3.16391557e-01 5.63274622e-01 3.94590199e-01
7.31375933e-01 2.00627580e-01 -2.63070077e-01 1.25757670e+00
-4.73261863e-01 6.34054020e-02 2.85294324e-01 9.38268363e-01
-6.32435799e-01 -8.81223321e-01 6.31413981e-02 1.96872577e-01
-4.14128721e-01 -6.36354148e-01 -4.42048043e-01 5.64890683e-01
3.79271917e-02 1.24944592e+00 -2.12091953e-01 -1.34429455e-01
5.99523604e-01 6.67074621e-01 6.46171749e-01 -8.19077611e-01
-8.25283408e-01 -6.15252137e-01 3.13417315e-01 -5.34129560e-01
-2.26728842e-01 3.85772921e-02 -1.23426807e+00 -8.15275535e-02
-3.44480008e-01 3.51042241e-01 1.68583617e-01 1.23252749e+00
5.03003478e-01 5.10418952e-01 6.44969642e-02 3.83425176e-01
-9.47202504e-01 -1.29420173e+00 -6.30988851e-02 5.59458494e-01
2.12464914e-01 -3.02384615e-01 -1.55704737e-01 2.83457130e-01] | [11.162311553955078, 8.054126739501953] |
8a0abeab-9260-471b-8d69-a17f2dd51293 | metric-learning-improves-the-ability-of | 2302.14616 | null | https://arxiv.org/abs/2302.14616v1 | https://arxiv.org/pdf/2302.14616v1.pdf | Metric Learning Improves the Ability of Combinatorial Coverage Metrics to Anticipate Classification Error | Machine learning models are increasingly used in practice. However, many machine learning methods are sensitive to test or operational data that is dissimilar to training data. Out-of-distribution (OOD) data is known to increase the probability of error and research into metrics that identify what dissimilarities in data affect model performance is on-going. Recently, combinatorial coverage metrics have been explored in the literature as an alternative to distribution-based metrics. Results show that coverage metrics can correlate with classification error. However, other results show that the utility of coverage metrics is highly dataset-dependent. In this paper, we show that this dataset-dependence can be alleviated with metric learning, a machine learning technique for learning latent spaces where data from different classes is further apart. In a study of 6 open-source datasets, we find that metric learning increased the difference between set-difference coverage metrics (SDCCMs) calculated on correctly and incorrectly classified data, thereby demonstrating that metric learning improves the ability of SDCCMs to anticipate classification error. Paired t-tests validate the statistical significance of our findings. Overall, we conclude that metric learning improves the ability of coverage metrics to anticipate classifier error and identify when OOD data is likely to degrade model performance. | ['Laura Freeman', 'Tyler Cody'] | 2023-02-28 | null | null | null | null | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 2.38968551e-01 -1.20996982e-01 -4.95897114e-01 -7.24026740e-01
-1.06874204e+00 -6.75098181e-01 2.92872578e-01 6.21546090e-01
-2.54333377e-01 6.65898561e-01 9.00151283e-02 -4.58043993e-01
-6.27584934e-01 -8.22137773e-01 -5.65100014e-01 -3.41194898e-01
9.25499573e-02 5.19231558e-01 -7.74883479e-02 4.02074128e-01
6.56234622e-01 4.06258315e-01 -1.83289099e+00 5.41852415e-01
9.57039535e-01 9.20273185e-01 -4.53378111e-02 4.32170570e-01
-1.61184400e-01 7.16526508e-01 -1.07860494e+00 -1.08413160e-01
2.51620173e-01 -3.29553366e-01 -5.29874921e-01 -1.95125371e-01
5.80960333e-01 2.09271368e-02 3.19209814e-01 7.02398837e-01
4.16062385e-01 -3.61409605e-01 1.16427517e+00 -1.59078467e+00
-3.99419576e-01 5.28716147e-01 -2.18254358e-01 3.35649520e-01
3.28374773e-01 2.00284928e-01 1.13867939e+00 -5.28429806e-01
4.70294833e-01 9.82961833e-01 1.13502491e+00 4.76348922e-02
-1.45371652e+00 -6.99694753e-01 -3.69462788e-01 2.71860547e-02
-1.38561964e+00 -3.33630815e-02 4.06291723e-01 -1.02306855e+00
1.05982435e+00 4.82345045e-01 5.00720322e-01 8.19546580e-01
5.88908970e-01 3.92430663e-01 1.27801585e+00 -6.56745672e-01
3.53683859e-01 3.39819461e-01 3.24069381e-01 2.67489940e-01
8.45783830e-01 3.32263857e-01 -3.19721282e-01 -3.99086952e-01
1.46351054e-01 7.48493597e-02 -1.87885910e-02 -3.59564275e-01
-7.80606449e-01 1.07961619e+00 2.90293824e-02 3.29259098e-01
7.02780783e-02 -1.43278822e-01 3.21989179e-01 5.47084272e-01
6.04655981e-01 1.18500745e+00 -7.53057659e-01 -4.74618912e-01
-1.02414167e+00 2.71015495e-01 9.61595476e-01 6.29051864e-01
6.63468421e-01 -2.31469706e-01 -2.41921708e-01 1.03535688e+00
1.70826107e-01 3.15656036e-01 6.69206798e-01 -6.76007032e-01
4.18204099e-01 1.06434250e+00 -1.00243084e-01 -9.95825529e-01
-5.94452500e-01 -4.77569997e-01 -3.99864912e-02 1.42354414e-01
5.99555969e-01 -7.50289634e-02 -6.84965611e-01 1.56046033e+00
2.89132819e-02 -2.71104664e-01 -5.43847643e-02 3.20766360e-01
3.52294415e-01 2.04707235e-01 1.98806506e-02 -1.86046436e-01
7.38461733e-01 -2.49557078e-01 -4.97370213e-01 -1.84713021e-01
1.47625995e+00 -8.44926536e-01 1.38854575e+00 5.23894072e-01
-5.57337642e-01 -3.61783564e-01 -1.26183975e+00 3.16515595e-01
-4.89757657e-01 -3.28014433e-01 5.06121814e-01 1.08951247e+00
-3.76180917e-01 8.63089681e-01 -7.05697358e-01 -2.70157427e-01
6.83295071e-01 2.66310602e-01 -2.52000809e-01 -3.82225662e-01
-7.99043834e-01 1.11696649e+00 4.57155019e-01 -6.19090319e-01
-6.31534457e-01 -1.04669845e+00 -8.21756124e-01 -6.13061227e-02
3.63539122e-02 6.05836604e-03 1.05011272e+00 -8.17654371e-01
-4.24255997e-01 6.09608233e-01 3.00334305e-01 -2.46931747e-01
2.12241143e-01 3.49201448e-02 -5.68094909e-01 -4.61321950e-01
1.75718978e-01 2.95277625e-01 3.43256682e-01 -1.00538373e+00
-7.45401084e-01 -3.79225820e-01 -2.52550006e-01 -1.27062738e-01
-4.43807244e-01 -1.09759994e-01 4.14904624e-01 -5.63401937e-01
1.10714726e-01 -7.38429606e-01 1.75315827e-01 -3.90309930e-01
-9.55012068e-02 -3.98558527e-01 6.44500971e-01 -3.44686210e-01
1.77188945e+00 -2.18049121e+00 -4.86070126e-01 3.94009441e-01
1.28022790e-01 8.64016339e-02 -5.52645400e-02 5.30680895e-01
-1.84334755e-01 5.14505446e-01 -3.22991818e-01 3.56925428e-01
-7.97446817e-03 1.62701428e-01 8.65231082e-02 4.74044055e-01
3.59470546e-01 4.57300365e-01 -6.03814363e-01 -3.24278653e-01
5.09654135e-02 1.45074964e-01 -7.39253521e-01 6.69978485e-02
-9.10524726e-02 -5.84459491e-02 9.58779231e-02 7.29068458e-01
5.38095772e-01 -2.90667683e-01 9.82847512e-02 7.90919214e-02
5.60473911e-02 5.51209271e-01 -1.01055682e+00 1.05615163e+00
-4.49680299e-01 1.00493693e+00 -8.97581756e-01 -8.26207519e-01
1.14508462e+00 -2.00627781e-02 4.81953382e-01 -7.47931063e-01
4.47352715e-02 4.64568287e-01 5.75563133e-01 -4.55796480e-01
1.02104098e-01 -1.29134685e-01 -3.36827934e-02 6.73429728e-01
-8.30237493e-02 -3.34291220e-01 1.88548982e-01 -2.87729830e-01
1.40991926e+00 -1.07880548e-01 3.30349833e-01 -3.97773743e-01
-6.72358423e-02 3.03282022e-01 5.86264729e-01 7.55134642e-01
-4.14796025e-02 6.12947404e-01 7.30268061e-01 -1.59355938e-01
-1.12020481e+00 -1.02279365e+00 -8.69227290e-01 6.60623491e-01
-3.79648536e-01 -6.94191039e-01 -8.12899947e-01 -7.95591772e-01
5.69525182e-01 1.10529673e+00 -5.90994596e-01 -5.43682814e-01
-5.57856774e-03 -7.77880609e-01 6.75063372e-01 6.36581302e-01
-1.69395357e-01 -6.95093691e-01 -7.27296352e-01 9.23596025e-02
8.41011479e-02 -6.09655678e-01 -1.71313718e-01 4.95653182e-01
-1.19102037e+00 -1.55328023e+00 -1.61621481e-01 -4.23578590e-01
6.25550091e-01 1.91044081e-02 1.30264461e+00 1.15631439e-01
-4.47713614e-01 4.15062338e-01 -4.64124620e-01 -8.28820467e-01
-5.36088884e-01 1.85123473e-01 -1.09006837e-01 -4.92073625e-01
1.07969081e+00 -2.23708406e-01 -1.28842175e-01 5.84801733e-01
-1.06177747e+00 -4.54194099e-01 4.37889427e-01 8.59209359e-01
5.47884345e-01 2.65023232e-01 8.67220044e-01 -1.04227662e+00
8.33537757e-01 -7.91675150e-01 -4.68741119e-01 2.59395838e-01
-1.33308876e+00 2.66428411e-01 1.34157583e-01 -6.32390320e-01
-4.13074374e-01 -4.67066675e-01 1.48242593e-01 -1.57066837e-01
-8.78383964e-02 9.66875732e-01 1.20579405e-02 1.76577881e-01
1.21633017e+00 -4.44592565e-01 2.32704863e-01 -3.34572703e-01
-4.81035262e-01 1.12731194e+00 -2.08849341e-01 -5.30929863e-01
2.64744610e-01 -8.78683701e-02 -2.84641944e-02 -4.97467339e-01
-1.02179885e+00 -4.82025653e-01 -4.98087734e-01 -9.72784907e-02
5.72749376e-01 -7.08134592e-01 -2.07474887e-01 2.33516648e-01
-4.68785733e-01 -6.81918800e-01 -4.35754210e-01 7.24415839e-01
-4.30593044e-01 -1.70004398e-01 -1.82929896e-02 -6.70293510e-01
1.63366243e-01 -1.20034301e+00 6.77729011e-01 1.19728083e-02
-8.42187226e-01 -1.19613206e+00 2.44546205e-01 2.94348180e-01
2.57214248e-01 5.90185821e-01 1.26241946e+00 -1.27427351e+00
-5.65623045e-02 -5.82951844e-01 -2.41484139e-02 7.01934040e-01
4.53630865e-01 1.58381946e-02 -9.52629089e-01 -3.00379366e-01
-1.46951852e-02 -2.66744196e-01 5.20758450e-01 4.88573194e-01
1.29032028e+00 -1.71974704e-01 -3.19400728e-01 2.78228819e-01
1.47902524e+00 3.64486277e-01 5.57596266e-01 5.02109468e-01
4.23128814e-01 8.79554689e-01 1.05834842e+00 4.57838774e-01
9.30991918e-02 5.21642327e-01 1.78264499e-01 1.86144337e-01
6.08613528e-03 -2.02264965e-01 3.40434313e-01 2.34367713e-01
5.42302430e-01 -1.82503924e-01 -1.40416813e+00 3.80368680e-01
-1.37462819e+00 -9.07664597e-01 -4.60748643e-01 2.53032613e+00
8.77810359e-01 4.77361679e-01 1.58575490e-01 6.65642262e-01
5.03935814e-01 -3.12881172e-01 -5.05402982e-01 -6.44113839e-01
4.88961823e-02 1.56491429e-01 4.49850470e-01 2.99204320e-01
-6.67849243e-01 2.88829625e-01 6.99771118e+00 6.72951579e-01
-9.48388875e-01 -3.10145952e-02 6.22736275e-01 -3.27379912e-01
-4.83328909e-01 -4.27953899e-02 -7.39291131e-01 6.70480847e-01
1.32425272e+00 -1.83921799e-01 8.61434564e-02 9.10456240e-01
3.74945179e-02 -4.33755010e-01 -1.62445998e+00 8.02597284e-01
3.04328829e-01 -1.07428300e+00 -2.96788067e-01 3.46050024e-01
9.40247178e-01 -7.25652054e-02 1.83144450e-01 3.61090600e-01
1.41191080e-01 -1.25093055e+00 5.20866632e-01 5.77767134e-01
7.96928287e-01 -1.00798333e+00 1.05357790e+00 4.50157613e-01
-5.90167999e-01 -3.04078519e-01 -3.96109223e-01 -2.99750090e-01
-4.34148252e-01 1.03546286e+00 -1.31796288e+00 -2.88940426e-02
7.42649078e-01 6.57955527e-01 -1.03352427e+00 1.28137255e+00
2.79709458e-01 9.21565890e-01 -1.53578237e-01 1.04100034e-02
-1.08254984e-01 1.74310520e-01 4.24636900e-02 1.06574953e+00
4.56477821e-01 -3.66776794e-01 -1.46939307e-01 7.67061830e-01
3.07705164e-01 -4.66580084e-03 -9.15251315e-01 -2.86607921e-01
7.07548320e-01 5.49084485e-01 -5.47331750e-01 2.64657792e-02
-4.86944854e-01 1.66818708e-01 3.50453667e-02 -6.81642964e-02
-6.67715371e-01 -5.28774619e-01 6.95790231e-01 4.56130981e-01
5.81680005e-03 5.62110655e-02 -8.65944803e-01 -6.70387387e-01
2.35789828e-03 -1.11837101e+00 6.64983034e-01 -5.62439978e-01
-1.47062635e+00 -7.76201859e-03 1.68587252e-01 -1.38976955e+00
-2.06232011e-01 -7.31449366e-01 -2.59891212e-01 8.61230433e-01
-9.59459007e-01 -5.74029565e-01 -2.57404953e-01 5.66189140e-02
3.65603268e-01 -2.23934352e-01 7.84880340e-01 2.05358565e-01
-2.70594925e-01 8.45117748e-01 3.98330659e-01 -5.93113378e-02
1.05817497e+00 -1.30091822e+00 -5.73408082e-02 4.05258924e-01
1.68807924e-01 5.28409898e-01 5.13594985e-01 -8.94893289e-01
-8.33952785e-01 -9.55508590e-01 1.00896692e+00 -9.97644544e-01
2.94544429e-01 -1.33438632e-01 -9.01364982e-01 5.60002089e-01
-4.45475847e-01 -2.97644407e-01 1.51964211e+00 5.72926939e-01
-6.28582895e-01 -3.46003205e-01 -1.46765065e+00 1.03875250e-01
7.89327323e-01 -5.15988708e-01 -6.00629628e-01 2.46386349e-01
4.26113874e-01 -1.92664936e-01 -1.32871485e+00 6.24808133e-01
7.11957514e-01 -1.13498652e+00 4.63133901e-01 -6.00807905e-01
5.12120485e-01 -7.96720758e-03 -4.97621953e-01 -1.38831282e+00
-2.24144850e-02 1.29092887e-01 -1.26931863e-02 1.23394358e+00
8.87151897e-01 -6.63572609e-01 6.31347775e-01 9.10720229e-01
-3.31013530e-01 -9.74512756e-01 -6.57567322e-01 -1.05842030e+00
3.21465164e-01 -8.75519872e-01 7.78618515e-01 1.07606733e+00
1.72020659e-01 -4.01326008e-02 2.61928737e-01 -1.52437137e-02
3.06530982e-01 -3.44024479e-01 6.81946814e-01 -1.70011187e+00
-1.87707290e-01 -5.42607844e-01 -7.46975005e-01 -1.38581902e-01
-3.11284605e-02 -1.13996851e+00 -1.96510062e-01 -1.41047597e+00
1.49164215e-01 -8.72632504e-01 -4.10156995e-01 4.37393367e-01
-9.35333893e-02 5.78665249e-02 1.28349006e-01 1.65655077e-01
-8.98020193e-02 -8.09845626e-02 8.88392389e-01 -8.03253204e-02
-2.90453076e-01 1.26688629e-01 -8.18551362e-01 5.24441838e-01
8.46860290e-01 -9.83292997e-01 -7.00000882e-01 -2.20576927e-01
3.79129738e-01 -3.65993738e-01 9.51441303e-02 -1.37699223e+00
-1.59770757e-01 -2.99984694e-01 6.21851742e-01 -5.12464702e-01
-2.54244059e-01 -9.22861516e-01 2.78118372e-01 6.13706470e-01
-5.81879079e-01 2.91907400e-01 4.52802420e-01 3.37351829e-01
-7.62613937e-02 -6.80607259e-01 5.26599884e-01 2.73563415e-01
-2.28325263e-01 -1.72440350e-01 -2.28282407e-01 3.45069945e-01
1.06901872e+00 -4.76835489e-01 -2.65845925e-01 1.32586975e-02
-3.33113641e-01 -1.12427562e-01 5.20224869e-01 5.06560147e-01
3.71744573e-01 -1.25201833e+00 -6.34814739e-01 3.09780419e-01
5.63424706e-01 -2.64344931e-01 -9.91346911e-02 7.92928278e-01
-4.99208421e-01 6.87429726e-01 -7.24303350e-02 -8.76440823e-01
-1.20146036e+00 3.59239817e-01 3.23757529e-01 -2.47780770e-01
1.29845450e-02 6.52462125e-01 -3.52197796e-01 -5.64144850e-01
2.43325412e-01 -5.84735513e-01 2.51329765e-02 3.62836719e-01
3.38437915e-01 7.99566507e-01 4.24888074e-01 5.47365919e-02
-1.96732432e-01 2.09085554e-01 -2.46780425e-01 -3.22467797e-02
1.17048013e+00 3.34742904e-01 1.86974406e-01 9.14102077e-01
1.25647378e+00 -1.42962396e-01 -1.07266772e+00 1.69138387e-01
5.22408009e-01 -7.80600786e-01 1.57163027e-04 -1.19700062e+00
-4.76017594e-01 6.88842356e-01 9.25902724e-01 4.72201169e-01
9.78966117e-01 -8.66394043e-02 1.08776599e-01 2.14512333e-01
3.31059992e-01 -1.33024621e+00 2.10516125e-01 3.69319916e-01
7.23070145e-01 -1.44099104e+00 1.57272413e-01 -1.37619572e-02
-5.46015203e-01 8.98117244e-01 7.41534472e-01 2.21793696e-01
7.55690038e-01 4.07699704e-01 -2.89338399e-02 -6.98522478e-02
-6.55074835e-01 1.46696985e-01 3.00087035e-01 9.29799914e-01
6.93484545e-01 2.07889006e-01 -4.94296968e-01 4.99529660e-01
-5.08941889e-01 -1.21569134e-01 4.78413314e-01 1.05941141e+00
-4.48995352e-01 -1.35082889e+00 -4.88767356e-01 1.28634655e+00
-3.08045447e-01 -6.61434084e-02 -4.94864136e-01 1.02963197e+00
3.47686648e-01 1.19620502e+00 4.81463283e-01 -5.94971478e-01
2.50510961e-01 4.50126946e-01 3.36462080e-01 -1.02489889e+00
-4.80915934e-01 -4.74835008e-01 1.42179161e-01 -3.90090227e-01
-8.25383440e-02 -9.90034223e-01 -1.06726027e+00 -3.22642177e-01
-7.74287820e-01 4.96794693e-02 7.99672186e-01 9.85736728e-01
4.42152053e-01 5.00240326e-01 5.92387140e-01 7.64589980e-02
-8.49335551e-01 -1.14145446e+00 -5.32905877e-01 3.39233726e-01
2.43489429e-01 -8.93686950e-01 -6.73423648e-01 -2.42133602e-01] | [8.873300552368164, 4.605966091156006] |
520c511b-1f17-4b58-a2af-3c9ad9605275 | multi-view-vision-to-geometry-knowledge | 2207.03128 | null | https://arxiv.org/abs/2207.03128v4 | https://arxiv.org/pdf/2207.03128v4.pdf | PointMCD: Boosting Deep Point Cloud Encoders via Multi-view Cross-modal Distillation for 3D Shape Recognition | As two fundamental representation modalities of 3D objects, 3D point clouds and multi-view 2D images record shape information from different domains of geometric structures and visual appearances. In the current deep learning era, remarkable progress in processing such two data modalities has been achieved through respectively customizing compatible 3D and 2D network architectures. However, unlike multi-view image-based 2D visual modeling paradigms, which have shown leading performance in several common 3D shape recognition benchmarks, point cloud-based 3D geometric modeling paradigms are still highly limited by insufficient learning capacity, due to the difficulty of extracting discriminative features from irregular geometric signals. In this paper, we explore the possibility of boosting deep 3D point cloud encoders by transferring visual knowledge extracted from deep 2D image encoders under a standard teacher-student distillation workflow. Generally, we propose PointMCD, a unified multi-view cross-modal distillation architecture, including a pretrained deep image encoder as the teacher and a deep point encoder as the student. To perform heterogeneous feature alignment between 2D visual and 3D geometric domains, we further investigate visibility-aware feature projection (VAFP), by which point-wise embeddings are reasonably aggregated into view-specific geometric descriptors. By pair-wisely aligning multi-view visual and geometric descriptors, we can obtain more powerful deep point encoders without exhausting and complicated network modification. Experiments on 3D shape classification, part segmentation, and unsupervised learning strongly validate the effectiveness of our method. The code and data will be publicly available at https://github.com/keeganhk/PointMCD. | ['Yue Qian', 'Junhui Hou', 'Qijian Zhang'] | 2022-07-07 | null | null | null | null | ['3d-shape-retrieval', '3d-shape-recognition'] | ['computer-vision', 'computer-vision'] | [-1.83273152e-01 -1.21470638e-01 -3.05473562e-02 -5.03056109e-01
-6.89524174e-01 -8.15647542e-01 8.16598415e-01 -3.14395353e-02
1.19565642e-02 -9.14864391e-02 -9.32758749e-02 -2.61890322e-01
1.79643761e-02 -9.87374783e-01 -1.05404139e+00 -4.88975734e-01
2.60609269e-01 7.70744979e-01 7.72174820e-02 -1.45315558e-01
1.46623686e-01 1.01933014e+00 -1.56037676e+00 1.49074584e-01
4.81429011e-01 1.16395915e+00 1.81037933e-01 5.14627516e-01
-3.78256202e-01 9.50613245e-02 -3.32116522e-02 -4.58388418e-01
6.48726106e-01 2.06772998e-01 -5.63193262e-01 4.12334651e-01
8.88515532e-01 -4.50475961e-01 -5.07047474e-01 9.38265979e-01
6.89423501e-01 -9.80394185e-02 5.41376412e-01 -1.38028657e+00
-1.09654832e+00 -5.06370049e-03 -7.00205088e-01 -1.80322438e-01
2.24957734e-01 1.44096494e-01 9.40458834e-01 -1.21728694e+00
6.89671218e-01 1.31535614e+00 5.94898939e-01 5.17751634e-01
-1.18639481e+00 -4.14772183e-01 2.64991283e-01 -2.14860160e-02
-1.30287051e+00 -1.20852552e-01 1.42858386e+00 -5.44371486e-01
1.11734927e+00 4.29236665e-02 8.52690995e-01 1.10814202e+00
-7.37254769e-02 9.25314248e-01 9.19242620e-01 -1.25955373e-01
-1.90119445e-01 5.49397282e-02 -2.14088693e-01 9.30749118e-01
6.13918714e-02 1.87427282e-01 -2.95217812e-01 -5.75716533e-02
1.07332742e+00 5.38944483e-01 -2.21901193e-01 -1.10445273e+00
-1.07490075e+00 7.25502193e-01 6.83016658e-01 1.13608487e-01
-2.13932946e-01 2.15948135e-01 2.45262399e-01 2.22298741e-01
7.71571815e-01 4.96349633e-02 -6.54140055e-01 -3.49846929e-02
-6.68515265e-01 1.15877017e-01 3.41599852e-01 1.18845856e+00
1.03844023e+00 5.93553148e-02 2.99597681e-01 7.30246425e-01
5.36378980e-01 8.68461728e-01 3.56041789e-02 -9.31281388e-01
5.06605268e-01 1.18318939e+00 -3.14400375e-01 -8.35120559e-01
-3.11337680e-01 -4.30674046e-01 -7.71313548e-01 4.15413052e-01
1.61589235e-02 3.20744157e-01 -1.13310242e+00 1.47558296e+00
5.16216815e-01 4.10035066e-03 -1.60884082e-01 9.03532386e-01
9.44561183e-01 4.63330597e-01 -3.70297313e-01 5.13906538e-01
1.32942486e+00 -8.04768145e-01 1.96283963e-02 -3.70499901e-02
5.38106263e-01 -7.00127065e-01 9.71792579e-01 1.02957673e-01
-1.34982264e+00 -6.88897491e-01 -1.03722382e+00 -6.31125391e-01
-5.60611606e-01 -2.50936393e-02 6.85873806e-01 3.77377480e-01
-9.58154082e-01 3.93121839e-01 -1.08934081e+00 -1.46564707e-01
8.98833156e-01 3.12265337e-01 -7.43728697e-01 -2.85230368e-01
-5.49713016e-01 6.87742591e-01 -3.96106429e-02 -2.95314156e-02
-9.19550359e-01 -1.14253473e+00 -1.10147619e+00 1.70012772e-01
6.78232089e-02 -1.17750108e+00 9.65492308e-01 -4.83362883e-01
-1.28885901e+00 1.35295784e+00 -1.02321748e-02 8.89408588e-02
4.17949826e-01 -4.87708114e-02 -5.57722598e-02 1.60518155e-01
2.43463600e-03 7.66497552e-01 7.70541430e-01 -1.68186796e+00
-3.10239851e-01 -8.51112366e-01 2.43237495e-01 3.69705260e-01
-7.10188523e-02 -4.77060497e-01 -5.95517635e-01 -2.66352773e-01
5.03115356e-01 -8.30354691e-01 -8.77405554e-02 3.88821721e-01
-3.16868663e-01 -3.79438311e-01 1.08248997e+00 -2.31714934e-01
2.85291642e-01 -2.29126549e+00 2.85475612e-01 8.92040730e-02
4.15416867e-01 9.52821523e-02 -2.68377870e-01 3.45899552e-01
-1.54767379e-01 1.00704007e-01 -1.03621028e-01 -6.11622274e-01
1.43549532e-01 2.59202093e-01 -2.75595963e-01 5.38330436e-01
3.46696675e-01 1.23414552e+00 -6.90546513e-01 -2.76789933e-01
6.31816804e-01 8.28770518e-01 -6.87182665e-01 3.69928032e-01
-2.33764693e-01 2.72428781e-01 -5.42539597e-01 7.49936223e-01
1.00050628e+00 -6.09534860e-01 -3.02741706e-01 -4.14578080e-01
-2.23899633e-02 1.80564001e-01 -7.36719668e-01 2.26753569e+00
-7.43395209e-01 3.90408814e-01 2.09271405e-02 -1.19574130e+00
1.13380504e+00 2.48032480e-01 7.11964071e-01 -5.50553322e-01
7.16214478e-02 1.27291620e-01 -4.14136916e-01 -3.10749024e-01
3.01544815e-01 -1.19143292e-01 8.58343542e-02 3.31325501e-01
4.49315071e-01 -7.55870283e-01 -4.60830539e-01 6.13671392e-02
7.50137925e-01 4.08839136e-01 -1.45970853e-02 2.22467825e-01
4.68902946e-01 -1.15457907e-01 1.50712460e-01 2.07708403e-01
-4.93409066e-03 9.94272232e-01 3.03363979e-01 -7.23249853e-01
-1.27122140e+00 -1.51886737e+00 -2.15747550e-01 6.44492924e-01
2.75410503e-01 -2.45596468e-01 -1.96128637e-01 -8.25535953e-01
3.09429228e-01 3.20057690e-01 -4.19485927e-01 -1.33482972e-03
-6.33683145e-01 -1.76474109e-01 3.07074577e-01 6.25057280e-01
4.26199704e-01 -4.23853368e-01 -4.44178849e-01 -1.37048975e-01
3.59499365e-01 -1.32860136e+00 -4.56619948e-01 2.24945202e-01
-1.06116414e+00 -1.12247121e+00 -7.66522408e-01 -7.55931199e-01
7.67505169e-01 7.77215838e-01 1.15103936e+00 9.48831141e-02
-1.59816578e-01 9.13533151e-01 -2.25438118e-01 -3.81201863e-01
-1.60330012e-01 1.06476448e-01 -4.45221923e-02 -1.52314648e-01
4.49483693e-01 -1.17841768e+00 -6.96193039e-01 9.95781794e-02
-8.11594546e-01 3.49486768e-01 5.77537179e-01 7.01951265e-01
8.97520304e-01 -5.70890069e-01 3.69476750e-02 -6.48108244e-01
-2.33521163e-02 -2.67059267e-01 -8.00734758e-01 1.63955331e-01
-4.01303500e-01 -1.00988802e-02 6.46251738e-01 -1.30411899e-02
-7.17956245e-01 2.02152863e-01 -4.06873375e-01 -1.23940539e+00
-5.73144197e-01 1.23561896e-01 -3.79281819e-01 -2.48269930e-01
2.87048638e-01 5.54851532e-01 1.84551701e-01 -6.14933670e-01
7.04389453e-01 5.36710262e-01 1.56623214e-01 -5.00947952e-01
1.19457257e+00 8.08859885e-01 1.02887325e-01 -7.21645713e-01
-7.64404118e-01 -4.01024520e-01 -9.52094734e-01 -8.15518126e-02
1.08066130e+00 -1.22176063e+00 -7.22305357e-01 3.49425316e-01
-1.38522959e+00 -5.75390086e-02 -2.98954636e-01 3.24172527e-01
-8.57461393e-01 4.67598081e-01 -2.61736691e-01 -3.15501511e-01
-2.36857593e-01 -1.33897614e+00 1.64505875e+00 3.93777750e-02
3.36649477e-01 -1.06272054e+00 8.28536674e-02 5.95805943e-01
1.26367003e-01 3.13238889e-01 1.12134051e+00 -5.84452391e-01
-9.51881945e-01 -1.53541178e-01 -3.52315634e-01 4.62608367e-01
1.05488725e-01 -1.53752759e-01 -1.11893129e+00 -3.44153136e-01
1.66592002e-02 -4.07043904e-01 6.48114145e-01 2.31568754e-01
1.39891887e+00 2.39781469e-01 -3.50718141e-01 1.21504843e+00
1.56757164e+00 1.64036080e-02 2.35768586e-01 9.67658684e-02
1.21854568e+00 3.99532020e-01 -6.42137155e-02 2.57678479e-01
7.04289913e-01 7.97403395e-01 8.13654840e-01 -2.59364635e-01
-1.86598644e-01 -6.03722095e-01 7.07962587e-02 1.13804185e+00
-2.18786657e-01 3.01704574e-02 -9.91608322e-01 4.47249025e-01
-1.33661902e+00 -8.25560331e-01 1.07481882e-01 1.83073246e+00
3.06798726e-01 -9.76203829e-02 -1.59339607e-01 -7.08112195e-02
3.69306833e-01 4.51235294e-01 -7.26901591e-01 -3.58790457e-01
-3.70911881e-02 1.51086107e-01 3.48287553e-01 3.40741277e-01
-1.14232469e+00 7.23585129e-01 4.50457907e+00 7.47113466e-01
-1.35344398e+00 1.43950105e-01 4.87123609e-01 -7.29476288e-02
-8.39534819e-01 -7.86016956e-02 -6.71581089e-01 1.26800761e-01
3.54723722e-01 1.60619482e-01 2.16757834e-01 1.02413106e+00
-7.27468356e-02 5.77016234e-01 -1.41178179e+00 1.43375683e+00
2.02201426e-01 -1.62872291e+00 4.01134878e-01 3.74216676e-01
7.27519274e-01 5.20746768e-01 2.44732961e-01 1.41743615e-01
5.01968004e-02 -9.46398854e-01 7.17136443e-01 3.07397455e-01
8.92467976e-01 -5.52565634e-01 2.37369955e-01 3.08251262e-01
-1.25767505e+00 1.89874306e-01 -4.79376704e-01 1.74546570e-01
2.37805337e-01 4.91421044e-01 -4.58335996e-01 1.01484394e+00
6.63425922e-01 1.13299727e+00 -4.60567683e-01 7.72546351e-01
-1.25431595e-02 2.32704561e-02 -3.66157532e-01 3.61473441e-01
3.86608034e-01 -3.41268927e-01 6.08841717e-01 6.23456836e-01
4.97087628e-01 8.43349099e-03 1.01302400e-01 1.26458502e+00
-1.57033652e-01 -2.16744289e-01 -1.11414874e+00 1.39577225e-01
3.75978500e-01 1.31211567e+00 -5.78686953e-01 -1.68351069e-01
-9.14397717e-01 8.87215376e-01 5.13954997e-01 2.72523969e-01
-6.91571474e-01 -3.50039415e-02 9.27268267e-01 1.10434480e-01
5.87080240e-01 -6.18673146e-01 -3.95816416e-01 -1.47356856e+00
5.86245544e-02 -4.37265843e-01 -8.69854465e-02 -8.68629873e-01
-1.50002730e+00 5.10787904e-01 4.40136977e-02 -1.60702121e+00
9.94340330e-02 -9.49731886e-01 -5.08612216e-01 8.67488265e-01
-1.59838247e+00 -1.51791918e+00 -3.17183614e-01 6.63501799e-01
4.95064020e-01 -1.33039132e-01 7.41591930e-01 3.08970839e-01
-6.09246157e-02 4.87640113e-01 -3.98342162e-02 1.51702985e-01
3.19827467e-01 -1.32219911e+00 6.77120447e-01 4.29673672e-01
6.01713061e-01 2.56702930e-01 -1.01996269e-02 -1.69305936e-01
-1.91087401e+00 -1.11097836e+00 5.70180416e-01 -8.02693665e-01
2.92347848e-01 -7.31856823e-01 -8.80414009e-01 6.87091470e-01
9.49986577e-02 4.31604981e-01 6.39265478e-01 8.09926018e-02
-5.74463427e-01 2.97204684e-03 -8.22707415e-01 2.99975514e-01
1.43294120e+00 -8.53896320e-01 -5.26781857e-01 3.02964121e-01
7.75039256e-01 -6.99729979e-01 -1.21062875e+00 5.93717575e-01
3.77981454e-01 -1.06896126e+00 1.40257597e+00 -5.64325988e-01
5.91790855e-01 -3.01382482e-01 -4.65090096e-01 -1.35307240e+00
-2.00306267e-01 -3.60875502e-02 -7.78815150e-02 9.37173903e-01
1.43487483e-01 -3.89589131e-01 9.93321598e-01 2.42333859e-01
-6.04812324e-01 -1.19058251e+00 -1.15177798e+00 -6.53497577e-01
3.62168044e-01 -5.63836277e-01 8.81102622e-01 9.44317400e-01
-6.38509154e-01 3.60667408e-01 9.10860002e-02 4.19929475e-01
5.29064000e-01 7.74455667e-01 1.02789283e+00 -1.27862000e+00
-2.06091210e-01 -6.32260084e-01 -8.48722279e-01 -1.41324413e+00
2.62640238e-01 -1.37179387e+00 -4.69303817e-01 -1.59204650e+00
1.72416806e-01 -4.50461507e-01 -2.78876591e-02 3.38064641e-01
1.74022064e-01 3.20832551e-01 4.38133001e-01 7.66802281e-02
-3.37562084e-01 9.87474024e-01 1.69984448e+00 -3.11355084e-01
1.66284051e-02 -3.92322429e-02 -4.29694653e-01 7.04316556e-01
5.13831079e-01 -3.83287929e-02 -4.96089667e-01 -1.11685920e+00
2.48884365e-01 7.43305311e-02 7.29214847e-01 -6.28873229e-01
1.63882583e-01 5.62544242e-02 5.85900366e-01 -9.45027471e-01
6.86042607e-01 -1.19761550e+00 3.95902172e-02 5.43997921e-02
6.05339774e-05 1.98753521e-01 1.89175859e-01 5.29172540e-01
-3.30854535e-01 1.73941836e-01 5.83887935e-01 -3.16123664e-01
-5.39432406e-01 9.61513937e-01 2.82114148e-01 -5.44184074e-02
9.12144959e-01 -5.31745672e-01 -2.97761351e-01 -1.32429704e-01
-7.16320395e-01 1.57089308e-01 8.28786790e-01 5.84353387e-01
1.01938128e+00 -1.60148370e+00 -4.53925431e-01 6.53973103e-01
2.03091368e-01 5.48250496e-01 4.94759947e-01 6.26173794e-01
-5.54212034e-01 3.95288229e-01 -2.71100879e-01 -1.11629128e+00
-9.90762055e-01 5.84375978e-01 5.27406096e-01 1.04494058e-01
-9.12564397e-01 8.36501122e-01 5.53884506e-01 -1.01523125e+00
8.31157789e-02 -4.31988418e-01 1.29405081e-01 -2.71845788e-01
-3.65403034e-02 1.15181610e-01 2.45001540e-01 -7.17600465e-01
-3.38301480e-01 1.15098917e+00 4.86124642e-02 2.53414571e-01
1.64482284e+00 -2.19116062e-02 2.66804397e-02 3.74845982e-01
1.71010971e+00 -1.44948915e-01 -1.43359435e+00 -2.06988841e-01
-3.53468716e-01 -5.21005452e-01 1.10068075e-01 -2.84993380e-01
-1.33741307e+00 1.43882930e+00 5.90139031e-01 9.14032981e-02
9.79545891e-01 3.67030919e-01 8.01049471e-01 2.35914826e-01
4.26655948e-01 -4.87029076e-01 -3.74003290e-03 6.42513871e-01
8.95329297e-01 -1.35576534e+00 -3.86039242e-02 -3.91232133e-01
-1.77661881e-01 1.16348922e+00 7.81008244e-01 -3.18074465e-01
1.05104744e+00 -4.08494063e-02 5.56040369e-02 -6.01770103e-01
-8.15208912e-01 -1.08284526e-01 4.38322634e-01 7.03362346e-01
2.72242874e-01 4.47286572e-03 4.35591012e-01 3.48957866e-01
-2.49266952e-01 -3.61675113e-01 2.33769476e-01 6.07115805e-01
-3.23712565e-02 -1.17956614e+00 -2.88517755e-02 2.89970368e-01
-1.86406337e-02 1.50288254e-01 -1.88626036e-01 9.20927465e-01
2.76891500e-01 1.62390947e-01 3.98956567e-01 -4.54643726e-01
5.55961192e-01 7.61856809e-02 7.87721992e-01 -7.05576837e-01
-3.02342445e-01 -1.13069095e-01 -4.27375257e-01 -7.04364598e-01
-5.70585668e-01 -5.03203452e-01 -1.00988626e+00 -3.23918164e-01
-1.03704661e-01 -2.57948577e-01 6.86896205e-01 7.94605911e-01
6.49357259e-01 1.85079962e-01 8.73849213e-01 -1.33217645e+00
-5.53346574e-01 -4.40825969e-01 -4.83873695e-01 4.98161077e-01
2.99461901e-01 -8.84723663e-01 -2.61065930e-01 -1.85819045e-01] | [8.186348915100098, -3.416020631790161] |
28f3ad2f-1614-4a3b-b49b-e02a1863e98a | on-the-cross-modal-transfer-from-natural | 2204.08653 | null | https://arxiv.org/abs/2204.08653v1 | https://arxiv.org/pdf/2204.08653v1.pdf | On The Cross-Modal Transfer from Natural Language to Code through Adapter Modules | Pre-trained neural Language Models (PTLM), such as CodeBERT, are recently used in software engineering as models pre-trained on large source code corpora. Their knowledge is transferred to downstream tasks (e.g. code clone detection) via fine-tuning. In natural language processing (NLP), other alternatives for transferring the knowledge of PTLMs are explored through using adapters, compact, parameter efficient modules inserted in the layers of the PTLM. Although adapters are known to facilitate adapting to many downstream tasks compared to fine-tuning the model that require retraining all of the models' parameters -- which owes to the adapters' plug and play nature and being parameter efficient -- their usage in software engineering is not explored. Here, we explore the knowledge transfer using adapters and based on the Naturalness Hypothesis proposed by Hindle et. al \cite{hindle2016naturalness}. Thus, studying the bimodality of adapters for two tasks of cloze test and code clone detection, compared to their benchmarks from the CodeXGLUE platform. These adapters are trained using programming languages and are inserted in a PTLM that is pre-trained on English corpora (N-PTLM). Three programming languages, C/C++, Python, and Java, are studied along with extensive experiments on the best setup used for adapters. Improving the results of the N-PTLM confirms the success of the adapters in knowledge transfer to software engineering, which sometimes are in par with or exceed the results of a PTLM trained on source code; while being more efficient in terms of the number of parameters, memory usage, and inference time. Our results can open new directions to build smaller models for more software engineering tasks. We open source all the scripts and the trained adapters. | ['Fatemeh H. Fard', 'Ramansh Grover', 'Divyam Goel'] | 2022-04-19 | null | null | null | null | ['cloze-test'] | ['natural-language-processing'] | [ 1.40820220e-02 2.98635632e-01 -5.58475368e-02 -2.03901425e-01
-4.89301503e-01 -7.33885169e-01 3.91977429e-01 6.56492040e-02
-4.88671780e-01 2.41048500e-01 6.94936067e-02 -7.89382637e-01
2.48129033e-02 -6.43859804e-01 -1.15020955e+00 -2.32998848e-01
-2.25842193e-01 1.55631810e-01 3.49608302e-01 -1.44370437e-01
4.69622850e-01 1.01924703e-01 -1.55518234e+00 7.50098228e-01
8.45125616e-01 4.01556760e-01 4.65704411e-01 9.25279796e-01
-3.91321808e-01 9.76249456e-01 -4.35629964e-01 -6.45892262e-01
1.92564294e-01 -2.43970990e-01 -1.05810726e+00 -5.20651519e-01
1.87923580e-01 9.32165906e-02 2.79389709e-01 1.06217968e+00
3.59063625e-01 -3.40531081e-01 5.96631825e-01 -1.08578527e+00
-7.45853364e-01 1.08110785e+00 -4.46697384e-01 9.09004509e-02
2.67583370e-01 2.57789969e-01 7.56154537e-01 -7.86185980e-01
6.97293818e-01 1.02929068e+00 1.14533758e+00 7.34314442e-01
-1.24977696e+00 -5.60878515e-01 -2.34517142e-01 1.30148724e-01
-1.41444778e+00 -4.71251220e-01 3.24597627e-01 -9.35889781e-01
1.63899744e+00 7.63249993e-02 4.19471189e-02 1.19411647e+00
1.12193502e-01 6.55528545e-01 7.83838153e-01 -7.84035802e-01
1.21542849e-01 9.29124177e-01 -1.29389297e-02 8.79997134e-01
-5.83087876e-02 1.07697152e-01 -3.30179781e-01 -1.89319625e-01
4.55506265e-01 -3.54791194e-01 -3.55583668e-01 -3.08253109e-01
-1.22320831e+00 8.46427321e-01 2.69959718e-01 7.62507439e-01
-7.54866004e-02 1.69601187e-01 9.11117256e-01 7.44728625e-01
2.77527213e-01 7.55309463e-01 -1.05707943e+00 -5.37778199e-01
-9.40010607e-01 -1.12515084e-01 1.24899864e+00 1.48333013e+00
1.01761377e+00 -1.29725918e-01 -2.84600575e-02 9.72719252e-01
3.73486459e-01 1.13472790e-01 9.94078338e-01 -6.59236670e-01
6.50831580e-01 9.36751068e-01 -4.66713637e-01 -5.26525676e-01
-3.39488059e-01 -3.85751158e-01 -4.66999710e-01 2.96716243e-01
4.56300110e-01 -9.65380222e-02 -4.82215971e-01 1.62938797e+00
-7.57926181e-02 9.21953411e-04 1.30302221e-01 3.14588696e-01
5.93545616e-01 6.07214570e-01 1.43564060e-01 2.54330486e-01
1.33595693e+00 -1.17848492e+00 2.42361836e-02 -2.87651330e-01
1.41581118e+00 -8.81281435e-01 1.65284336e+00 4.50416476e-01
-8.58850360e-01 -7.69333422e-01 -9.60987687e-01 -1.27270356e-01
-7.33368158e-01 3.72236460e-01 4.90489155e-01 9.51657295e-01
-1.43732178e+00 7.70444036e-01 -7.38073647e-01 -6.85155332e-01
3.49378258e-01 3.19651067e-01 -2.71877646e-01 6.22624829e-02
-8.09142709e-01 9.29645717e-01 6.02402151e-01 -2.57892251e-01
-9.10241365e-01 -8.98790359e-01 -8.06454003e-01 2.63407797e-01
1.97980940e-01 -5.16792834e-01 1.38413227e+00 -1.31316829e+00
-1.74783731e+00 1.17398155e+00 2.45193064e-01 -2.62936950e-01
4.32459652e-01 -1.62977263e-01 -1.77028194e-01 -4.10812944e-01
-6.59932196e-03 5.88014066e-01 7.20036983e-01 -9.31854486e-01
-3.74859631e-01 1.08452745e-01 2.96322763e-01 -5.24232268e-01
-4.47249591e-01 4.26524937e-01 -2.57124931e-01 -4.81125444e-01
-8.24560940e-01 -8.69556189e-01 -3.01350523e-02 -2.41655618e-01
-1.96825430e-01 -2.46816739e-01 4.03509766e-01 -8.12170625e-01
1.38368273e+00 -2.35635328e+00 3.42033565e-01 6.94539696e-02
-3.52965519e-02 4.35793638e-01 -5.37229002e-01 4.81202394e-01
-3.93125027e-01 3.38174194e-01 -4.92154509e-01 -1.81957811e-01
2.87081629e-01 1.07105039e-01 -3.68243046e-02 2.34107003e-01
4.36859429e-01 9.62595820e-01 -6.23176575e-01 -3.41933995e-01
-5.86600453e-02 3.12603801e-01 -9.72138524e-01 3.44635785e-01
-3.57056886e-01 1.11159310e-01 -8.88701901e-02 3.12948763e-01
3.32106262e-01 -1.75818026e-01 -2.26351485e-01 3.19309443e-01
-2.97829688e-01 4.62513387e-01 -9.08855319e-01 1.92288649e+00
-1.20997679e+00 6.99595153e-01 4.03393731e-02 -9.21175420e-01
8.28389108e-01 4.01828319e-01 -3.46790254e-01 -4.41432148e-01
-2.02448536e-02 4.50651884e-01 3.13137352e-01 -1.02762306e+00
1.66425511e-01 3.91772157e-03 -7.61007518e-02 5.27359605e-01
4.65415239e-01 -1.92119941e-01 4.45533156e-01 7.56728202e-02
1.40290999e+00 6.45292401e-01 3.58605772e-01 -6.15945995e-01
9.19225991e-01 -8.31815079e-02 3.80424224e-02 7.21168876e-01
2.57438004e-01 3.52787882e-01 7.81596959e-01 -1.59157172e-01
-1.16401541e+00 -6.99592352e-01 -6.19511977e-02 1.68983924e+00
-6.76056623e-01 -7.76779294e-01 -1.10844421e+00 -8.17389667e-01
-2.33625546e-01 9.56461370e-01 -6.03840530e-01 -3.98148626e-01
-6.20140672e-01 -5.78966856e-01 1.05821884e+00 5.16193092e-01
1.76797539e-01 -1.42783737e+00 -7.88290918e-01 1.96090207e-01
1.28015250e-01 -8.42365623e-01 -3.26241642e-01 5.15840113e-01
-8.86607528e-01 -1.01106763e+00 -4.66230482e-01 -7.75331736e-01
5.18540084e-01 -4.87930179e-01 1.52288008e+00 3.70710194e-01
-3.28696549e-01 3.69552106e-01 -4.96677250e-01 -4.17587042e-01
-1.26670229e+00 6.67410195e-01 -3.74490499e-01 -4.14375782e-01
6.91443861e-01 -7.70057201e-01 -6.14502095e-02 1.60875320e-01
-1.03973985e+00 -1.03863254e-02 1.02107799e+00 6.40361369e-01
-2.29387850e-01 -3.20981294e-01 3.47990125e-01 -1.26341116e+00
5.47316492e-01 -6.35253489e-01 -7.72216737e-01 2.49486879e-01
-5.56836605e-01 4.11510050e-01 8.79824817e-01 -6.03598058e-01
-1.11029124e+00 -1.02414958e-01 -5.49176633e-01 -8.96741971e-02
-4.70426291e-01 6.52817726e-01 -9.24667791e-02 -2.95889437e-01
1.15948701e+00 1.67178974e-01 -1.96177363e-01 -7.33178377e-01
4.13419247e-01 9.09427106e-01 2.91750997e-01 -8.85235965e-01
8.04796398e-01 -2.22931430e-01 -5.64405978e-01 -7.45711446e-01
-2.97868580e-01 -3.60584587e-01 -7.27072716e-01 2.82229066e-01
6.79914176e-01 -7.03973591e-01 -5.94166160e-01 3.80894423e-01
-1.41661131e+00 -6.70744956e-01 -3.12881589e-01 3.08904439e-01
-6.61491632e-01 3.04708093e-01 -8.87134552e-01 -3.81569713e-01
-1.44698545e-01 -1.36925650e+00 1.05898046e+00 -8.17068573e-03
-4.88484353e-01 -1.19239628e+00 3.69499713e-01 3.24528888e-02
9.02029932e-01 -3.12661856e-01 1.50207484e+00 -7.70693481e-01
-3.97968709e-01 2.28909515e-02 -2.50965446e-01 6.81036055e-01
-2.23266691e-01 1.10038102e-01 -1.26145995e+00 -2.65248507e-01
8.09172392e-02 -3.28284651e-01 7.14002609e-01 -1.44766062e-01
1.09713328e+00 -3.03954303e-01 -2.65257388e-01 6.70540035e-01
1.51207078e+00 -1.54171199e-01 7.29611635e-01 5.78494310e-01
5.97486794e-01 7.31727481e-01 -1.41307442e-02 1.26509324e-01
2.69947767e-01 4.40456003e-01 3.40626121e-01 1.20659746e-01
-6.20784312e-02 -8.21982548e-02 9.44371283e-01 1.19336605e+00
-8.87390897e-02 1.86976075e-01 -1.17152178e+00 6.36841893e-01
-1.52554691e+00 -5.36999464e-01 -2.96394497e-01 2.05872631e+00
1.25696218e+00 1.12898216e-01 -4.12768684e-02 -9.15260240e-02
3.68384153e-01 -4.02227849e-01 -2.02126130e-01 -8.68861794e-01
2.59321630e-01 4.01362777e-01 2.68050194e-01 4.90935057e-01
-8.02048981e-01 8.49325478e-01 5.19251060e+00 8.22254956e-01
-1.09571445e+00 5.63352704e-01 8.79965425e-02 2.62956887e-01
-6.65428191e-02 2.73835033e-01 -8.61414015e-01 4.36709255e-01
1.47079396e+00 -1.33933509e-02 7.53730476e-01 1.25204766e+00
-1.98330656e-01 -8.27260837e-02 -1.76258278e+00 6.28759146e-01
5.98053932e-02 -1.07995605e+00 -2.41277486e-01 -2.18095064e-01
6.26079142e-01 4.82830703e-01 -3.02925825e-01 1.04163694e+00
2.14689150e-01 -9.58258390e-01 8.15218925e-01 3.29036176e-01
6.16201937e-01 -3.37432086e-01 9.97516096e-01 5.47708511e-01
-8.38542521e-01 -1.33159056e-01 -5.00998020e-01 -1.70992300e-01
-3.90210271e-01 4.09178287e-01 -9.90037143e-01 3.32881600e-01
9.27573979e-01 5.60409129e-01 -1.20754218e+00 9.13977385e-01
-2.84496367e-01 8.27949762e-01 -1.94408804e-01 -1.53886005e-01
1.09004036e-01 1.31038025e-01 2.62342632e-01 1.98643279e+00
2.97113478e-01 -6.46667957e-01 -2.16306090e-01 1.49805021e+00
-7.73231462e-02 2.95315325e-01 -6.34633243e-01 -3.37503493e-01
5.10509014e-02 1.21448731e+00 -3.69189978e-01 -2.94607490e-01
-7.61771500e-01 8.85199308e-01 5.51600277e-01 2.04442397e-01
-8.37873340e-01 -8.46054375e-01 3.99439633e-01 1.44679546e-01
4.04654771e-01 1.34203583e-01 -1.37536556e-01 -1.12641692e+00
2.49981984e-01 -1.16558802e+00 2.14034006e-01 -6.02823436e-01
-1.32221508e+00 8.24152470e-01 4.72168922e-02 -9.26959813e-01
-3.54761839e-01 -1.09807193e+00 -6.06436670e-01 9.93309259e-01
-1.51723301e+00 -1.09635949e+00 -1.31631836e-01 4.27095145e-01
5.51838815e-01 -3.29371035e-01 9.39798951e-01 6.38980687e-01
-2.84187436e-01 7.60180116e-01 -4.99970317e-02 2.28690878e-01
7.57111251e-01 -1.39486384e+00 5.48860133e-01 7.65387118e-01
-4.44808370e-03 1.03090727e+00 5.75178385e-01 -2.37426713e-01
-1.13987494e+00 -1.10827279e+00 9.45954919e-01 -7.22033381e-01
9.07478333e-01 -6.90132082e-01 -1.30293119e+00 8.38859797e-01
4.59728062e-01 -2.90091634e-01 6.04433000e-01 1.62156627e-01
-6.58129811e-01 2.18267903e-01 -9.13612068e-01 3.14949572e-01
9.55361903e-01 -7.48177528e-01 -6.63630962e-01 2.14570731e-01
7.52804160e-01 -1.55494094e-01 -1.05347073e+00 1.46992743e-01
3.23089689e-01 -1.15131426e+00 6.33460999e-01 -6.88014090e-01
8.01335752e-01 -2.27976158e-01 -4.99489456e-02 -1.39083111e+00
-1.10412374e-01 -4.14856195e-01 2.46540219e-01 1.63745809e+00
8.29670608e-01 -6.76563382e-01 3.19843829e-01 2.91038275e-01
-2.12572083e-01 -4.37586665e-01 -5.89700401e-01 -7.49827206e-01
4.77311909e-01 -6.00784361e-01 4.16062564e-01 1.07048833e+00
2.52624691e-01 5.06071687e-01 2.26822421e-01 -7.64397755e-02
7.31709972e-02 -3.01139146e-01 9.46085393e-01 -1.18145239e+00
-9.68559921e-01 -7.69128025e-01 -2.84744173e-01 -7.72262514e-01
4.94312555e-01 -1.37292624e+00 2.23304167e-01 -8.60386789e-01
3.25293601e-01 -4.47178423e-01 1.54646575e-01 8.24064434e-01
-1.81484427e-02 -1.15476318e-01 1.56829879e-01 -3.75747159e-02
-3.27877164e-01 1.39180407e-01 4.84732449e-01 -7.90255591e-02
-2.59774089e-01 7.42464289e-02 -5.24375737e-01 9.12349403e-01
6.26580417e-01 -7.93148398e-01 -1.20504789e-01 -6.32014632e-01
8.12133968e-01 -2.80203611e-01 2.55801857e-01 -1.02534461e+00
1.66892380e-01 4.55099434e-01 -1.08248606e-01 1.85970724e-01
-3.87076199e-01 -8.42979431e-01 4.11076546e-02 5.35072565e-01
-3.96774739e-01 1.61000043e-01 7.73869693e-01 3.16492319e-02
-5.34864366e-02 -1.03271890e+00 8.46411765e-01 -5.53600550e-01
-6.51470244e-01 -3.59824747e-01 -5.82764387e-01 2.40831241e-01
7.04755902e-01 -1.08720273e-01 -3.98311645e-01 1.01206541e-01
-5.86419284e-01 -8.67380202e-02 5.65780878e-01 8.44227493e-01
-2.54421029e-02 -8.20557475e-01 -7.40661204e-01 3.90587240e-01
4.78870571e-01 -2.72801816e-01 -5.23095913e-02 1.12266886e+00
-6.96922958e-01 3.78979057e-01 -3.05425733e-01 -6.80876553e-01
-1.03555787e+00 9.01137054e-01 4.50678498e-01 -3.71994138e-01
-3.22566658e-01 8.91673803e-01 2.65651017e-01 -1.28255475e+00
2.04875559e-01 -6.74920857e-01 5.59614748e-02 -3.37523431e-01
5.64060152e-01 3.00686628e-01 3.52411717e-01 -1.15628205e-01
-3.41497511e-01 6.74184978e-01 -1.59527101e-02 2.97539026e-01
1.57746923e+00 2.65274256e-01 -7.58844197e-01 5.76392949e-01
1.38367355e+00 2.67459124e-01 -7.85760343e-01 -1.07269309e-01
6.35695696e-01 -1.07580900e-01 -3.77610296e-01 -9.67115402e-01
-8.12520087e-01 1.18584538e+00 5.44679821e-01 1.45186111e-01
9.17578220e-01 1.80689812e-01 7.90330768e-02 5.04025221e-01
4.97362256e-01 -8.11425686e-01 -1.85610250e-01 7.04254091e-01
9.37897444e-01 -1.02021790e+00 -5.38945258e-01 -2.45023698e-01
-2.90993303e-01 1.43597889e+00 7.69649029e-01 1.82359163e-02
4.77335602e-01 7.69709766e-01 6.90518841e-02 -4.53460775e-02
-7.61217713e-01 7.46785253e-02 2.47718915e-02 5.59401751e-01
9.43100572e-01 -2.40765736e-01 -1.93838447e-01 7.19525218e-01
-2.45077297e-01 1.85932174e-01 6.35197222e-01 8.21033001e-01
-1.08671501e-01 -1.25294447e+00 -3.51044685e-01 3.83841515e-01
-4.89643842e-01 -4.82110351e-01 -3.99715960e-01 8.36255670e-01
4.66700166e-01 6.28123820e-01 -2.93443620e-01 -2.80627400e-01
3.98003727e-01 4.87273276e-01 5.33899903e-01 -1.00552213e+00
-1.35768867e+00 -4.54891086e-01 4.04800568e-03 -6.49606824e-01
-1.97709769e-01 -4.35495347e-01 -9.05571520e-01 -1.45657793e-01
-4.77936983e-01 1.99354440e-01 6.89526498e-01 9.79463995e-01
4.61034894e-01 7.08145499e-01 -9.54705998e-02 -8.94778430e-01
-7.09137022e-01 -1.36705172e+00 -1.76543653e-01 2.02739030e-01
8.04841518e-02 -1.92688331e-01 -4.56126034e-01 3.97705197e-01] | [7.660797119140625, 7.8557353019714355] |
a81f459f-cd7a-44cc-abfa-a166877e6244 | getting-more-data-schoolkids-as-annotators | null | null | https://aclanthology.org/L12-1495 | https://aclanthology.org/L12-1495.pdf | Getting more data -- Schoolkids as annotators | We present a new way to get more morphologically and syntactically annotated data. We have developed an annotation editor tailored to school children to involve them in text annotation. Using this editor, they practice morphology and dependency-based syntax in the same way as they normally do at (Czech) schools, without any special training. Their annotation is then automatically transformed into the target annotation schema. The editor is designed to be language independent, however the subsequent transformation is driven by the annotation framework we are heading for. In our case, the object language is Czech and the target annotation scheme corresponds to the Prague Dependency Treebank annotation framework. | ["Barbora Hladk{\\'a}", 'Jirka Hana'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['text-annotation'] | ['natural-language-processing'] | [-2.34861076e-01 8.48187268e-01 -5.88974059e-02 -6.39393747e-01
-2.56889522e-01 -7.52815127e-01 2.52480537e-01 7.28675008e-01
-7.80369937e-01 7.62571037e-01 3.14020216e-01 -3.24620783e-01
7.22981766e-02 -7.79688358e-01 -5.21767419e-03 -1.30034715e-01
5.22823393e-01 8.31946254e-01 8.00192177e-01 -4.61748213e-01
-2.06364453e-01 3.17604780e-01 -1.22507977e+00 2.02690363e-01
9.46668684e-01 -3.17730904e-02 2.84509182e-01 5.11316061e-01
-5.20952284e-01 6.33956432e-01 -7.96435058e-01 -9.02124584e-01
-1.62983701e-01 -4.29676503e-01 -1.40870476e+00 1.36853037e-02
7.72054717e-02 8.14879462e-02 2.68642157e-01 1.15976942e+00
9.85894427e-02 3.61997820e-02 4.90315706e-01 -7.14848936e-01
-2.51876265e-01 1.29712212e+00 2.33404204e-01 -6.89330474e-02
4.55259264e-01 -4.41940039e-01 9.53022718e-01 -5.13323128e-01
1.18461454e+00 1.16718900e+00 2.50477284e-01 8.04348528e-01
-1.15143633e+00 -1.81884065e-01 2.51890123e-01 -1.21194318e-01
-1.14916968e+00 -1.54286146e-01 6.20749712e-01 -5.79516709e-01
1.11766005e+00 9.51748639e-02 8.85814071e-01 6.61903441e-01
-2.85798997e-01 5.27987003e-01 9.08390760e-01 -1.17398167e+00
1.13533065e-02 5.99021018e-01 4.36956018e-01 3.68133008e-01
4.27827477e-01 -4.05875564e-01 -1.59173697e-01 3.86742026e-01
6.94610357e-01 -8.39014947e-01 -3.48570123e-02 -2.60393351e-01
-9.00782168e-01 9.49319720e-01 -3.45308810e-01 1.12784171e+00
3.39134693e-01 -6.84409142e-02 8.86551499e-01 3.66255283e-01
2.04920903e-01 2.36258104e-01 -9.51682925e-01 -1.35694414e-01
-6.16756737e-01 2.26723120e-01 1.19532585e+00 1.20316541e+00
2.50723392e-01 6.33942708e-02 4.26556051e-01 1.11346281e+00
8.37180972e-01 1.99319631e-01 5.26192546e-01 -8.05052996e-01
3.60435456e-01 8.62284184e-01 -5.71446382e-02 -9.14826468e-02
-4.07227874e-01 -1.31864145e-01 -1.13798201e-01 2.03128427e-01
9.08270955e-01 -8.30865800e-02 -6.15200460e-01 1.63361490e+00
7.81001627e-01 -8.50666106e-01 3.80910605e-01 1.93825155e-01
1.14972067e+00 1.87444732e-01 6.09055996e-01 -4.90384907e-01
1.69364059e+00 -6.98658168e-01 -1.08572960e+00 9.28185061e-02
1.47963178e+00 -1.00465417e+00 9.90585506e-01 5.00484824e-01
-1.36159861e+00 -2.58610249e-01 -7.37317502e-01 -2.85418302e-01
-6.06525481e-01 1.32544562e-01 5.08451521e-01 1.07631087e+00
-9.98390436e-01 4.29372579e-01 -7.27944732e-01 -5.60666323e-01
-1.73843369e-01 4.08316016e-01 -6.08737886e-01 5.22665739e-01
-1.17853796e+00 1.10006666e+00 1.23387027e+00 -2.31520012e-01
1.62405282e-01 -1.76641881e-01 -1.25364101e+00 -1.37919605e-01
1.41653240e-01 -2.41397411e-01 1.78836346e+00 -1.05934179e+00
-1.78555179e+00 1.67851067e+00 2.52231687e-01 -7.98353106e-02
5.36512733e-01 1.61091574e-02 -3.94043803e-01 -8.41610655e-02
2.65813023e-01 5.11585295e-01 1.13914385e-01 -1.12327731e+00
-9.78380501e-01 -3.49446923e-01 1.26717925e-01 1.51409641e-01
-1.53398126e-01 9.98756588e-01 -5.38236856e-01 -9.64331269e-01
1.72320321e-01 -5.75401843e-01 -1.72197998e-01 -6.30715966e-01
3.26045342e-02 -7.71698356e-01 4.04923707e-01 -8.12355876e-01
1.64329886e+00 -1.80618453e+00 4.56974432e-02 1.43229708e-01
1.08907772e-02 6.42308772e-01 3.15748721e-01 6.34922624e-01
-2.27153569e-01 2.04905316e-01 -2.32210189e-01 -3.41193855e-01
2.50925004e-01 8.79053831e-01 2.02227503e-01 2.15435922e-01
-2.49282606e-02 5.68306327e-01 -9.27422762e-01 -9.05657649e-01
2.47109339e-01 2.51540035e-01 -4.42791522e-01 3.40361297e-02
-4.55398679e-01 7.01534212e-01 -5.11769950e-01 4.01168317e-01
4.53404337e-01 5.61970413e-01 9.02977943e-01 3.45783412e-01
-6.73826039e-01 6.19820714e-01 -1.41101587e+00 1.77442861e+00
-6.62914991e-01 1.81213230e-01 1.51389092e-01 -9.02251601e-01
9.39013243e-01 7.70286798e-01 2.94277025e-03 -9.91175324e-02
3.23135942e-01 6.30499244e-01 3.18885863e-01 -5.94818473e-01
5.09963155e-01 7.90371373e-02 -4.07109559e-01 3.61200929e-01
3.92317832e-01 -3.56504768e-01 6.85114682e-01 -1.50534287e-02
4.42282349e-01 7.12600589e-01 9.91473556e-01 -4.29784209e-01
1.02536368e+00 2.71575153e-01 6.82876110e-01 1.53421849e-01
1.50351133e-02 9.43256356e-03 6.06102705e-01 -3.84932429e-01
-1.07888103e+00 -6.82341754e-01 -5.13885379e-01 1.18332374e+00
-6.58511460e-01 -8.97586405e-01 -1.10775340e+00 -9.82843876e-01
-6.85890913e-01 8.45499635e-01 -3.17962080e-01 6.29633009e-01
-1.15256619e+00 1.22400351e-01 5.93246102e-01 2.86979526e-01
1.46714270e-01 -1.41499043e+00 -6.90456808e-01 7.31169343e-01
2.44927227e-01 -1.13867056e+00 -1.66518018e-01 2.72190720e-01
-5.52632391e-01 -1.03461587e+00 -1.75281942e-01 -1.52105772e+00
5.57260633e-01 -8.90058577e-01 1.18195283e+00 4.06977355e-01
2.56435633e-01 2.81055629e-01 -7.84449100e-01 -8.59240413e-01
-1.13804960e+00 2.45221972e-01 -4.23399419e-01 -6.92825913e-01
3.18158776e-01 -5.13027012e-01 1.90101340e-01 -2.89557248e-01
-1.10353673e+00 -1.38756320e-01 -1.12645000e-01 4.06445771e-01
2.01284915e-01 -1.27735108e-01 3.11717093e-01 -1.50578237e+00
2.62451261e-01 2.06982896e-01 -8.83808970e-01 1.74461916e-01
-3.09463859e-01 9.95064341e-03 6.41683161e-01 -3.28167677e-01
-1.21394372e+00 5.85246801e-01 -9.45259511e-01 4.05840278e-01
-9.39998567e-01 4.48292494e-01 -9.10561979e-01 -1.14885783e-02
2.27927253e-01 -3.92588794e-01 -6.57011092e-01 -9.12727416e-01
3.65485609e-01 5.35573363e-01 4.58250195e-01 -7.84307301e-01
6.52228117e-01 -2.29198307e-01 -4.63030413e-02 -5.43822587e-01
-5.45063794e-01 -3.90557975e-01 -1.42997718e+00 -1.32044598e-01
1.19927013e+00 -5.72823167e-01 -7.05613852e-01 3.18558723e-01
-1.78519487e+00 -4.46357965e-01 -9.01034772e-01 6.76994681e-01
-4.70008433e-01 4.74022597e-01 -6.96445882e-01 -7.38933086e-01
-1.58868060e-01 -1.02366567e+00 6.96366251e-01 -2.61262339e-02
-5.61060846e-01 -1.63689876e+00 4.79998648e-01 9.53933895e-02
-2.38779768e-01 2.15896964e-01 1.36763239e+00 -1.32106209e+00
2.31451228e-01 4.86586727e-02 2.36074895e-01 3.89676481e-01
-1.69098511e-01 2.63761729e-01 -7.86126375e-01 2.70451427e-01
9.52311829e-02 7.16924965e-02 -2.06040703e-02 -1.06818587e-01
3.04167628e-01 -2.95538634e-01 -3.17561217e-02 1.57963008e-01
1.62715638e+00 1.48944885e-01 3.00082922e-01 5.57778716e-01
5.74603260e-01 1.24324799e+00 6.41361296e-01 -1.07804328e-01
7.87817240e-01 9.47975039e-01 1.59580484e-02 9.06031653e-02
-3.93843129e-02 -2.04636708e-01 4.25556302e-01 1.23098791e+00
5.94894625e-02 4.31354120e-02 -1.33269584e+00 7.01699615e-01
-2.09234762e+00 -4.94780630e-01 -1.02490544e+00 2.15944314e+00
1.38629842e+00 2.10163996e-01 2.96849102e-01 3.65359724e-01
6.11606777e-01 -3.06871742e-01 7.74710715e-01 -1.09114838e+00
4.72593606e-02 7.70640314e-01 1.87431127e-01 9.10376310e-01
-9.66248155e-01 1.61613274e+00 6.24974632e+00 7.61410177e-01
-7.76603639e-01 3.85623157e-01 -2.51706272e-01 5.65681219e-01
-3.91149551e-01 3.64317954e-01 -1.09912539e+00 1.68027878e-01
1.09957552e+00 -9.31744799e-02 -2.00512484e-01 7.17980385e-01
2.82712519e-01 -2.38992721e-01 -1.19441903e+00 4.64167386e-01
-1.65565163e-01 -1.08002710e+00 -6.45671114e-02 5.79455160e-02
2.96737075e-01 -3.36020201e-01 -8.35428059e-01 1.89026937e-01
5.35941422e-01 -8.15571010e-01 1.13592744e+00 9.40494537e-02
7.11611450e-01 -6.28102481e-01 9.13588464e-01 3.78545314e-01
-1.26555276e+00 3.43114823e-01 -1.28950492e-01 -1.68449640e-01
3.33589017e-01 2.95933664e-01 -7.53894746e-01 6.57277405e-01
3.68508577e-01 1.86206907e-01 -6.26379967e-01 1.00729442e+00
-1.04008710e+00 7.43496597e-01 -1.83975518e-01 -4.86526825e-02
1.25944346e-01 -4.67316151e-01 8.16019475e-01 1.61831999e+00
-9.30241644e-02 2.13918447e-01 4.14460391e-01 1.95925578e-01
2.59670794e-01 9.77050304e-01 -4.66176957e-01 4.15350407e-01
2.84076929e-01 1.29148889e+00 -1.18244612e+00 -4.68084484e-01
-6.26346171e-01 6.53034627e-01 3.02797616e-01 -1.71214089e-01
-2.94307441e-01 -4.69481200e-01 1.99844375e-01 3.15359622e-01
4.01670247e-01 -2.40127325e-01 -1.19364001e-01 -1.00260723e+00
-7.45810717e-02 -7.43241251e-01 6.10450268e-01 -5.70596874e-01
-8.94449353e-01 6.75385475e-01 4.50763881e-01 -6.31197870e-01
-3.90539289e-01 -8.69423628e-01 -6.82408750e-01 8.67330790e-01
-1.11121523e+00 -1.48689508e+00 4.75762069e-01 5.40773928e-01
4.67516333e-01 2.09619001e-01 1.44235253e+00 2.22874656e-01
-3.65159065e-01 3.89298409e-01 -3.52330506e-01 5.73376060e-01
5.90954125e-01 -1.76298177e+00 1.55386299e-01 8.70162785e-01
2.09869653e-01 4.40954238e-01 9.17142332e-01 -7.02804804e-01
-5.09220123e-01 -9.16517198e-01 1.96548772e+00 -4.73822653e-01
1.10766816e+00 -3.99457574e-01 -1.08313751e+00 7.96018302e-01
7.97873974e-01 6.53566327e-03 9.99887824e-01 -8.13877955e-02
-1.32330269e-01 2.61391729e-01 -1.21053112e+00 1.97564989e-01
9.18563366e-01 -1.89863637e-01 -1.19585133e+00 3.91719669e-01
8.27171564e-01 -8.12777519e-01 -1.50126755e+00 9.17223692e-02
1.73368037e-01 -4.79554743e-01 1.78286210e-01 -7.29570329e-01
-1.88745677e-01 -3.81747961e-01 3.06480199e-01 -1.05536544e+00
8.92116874e-02 -7.54181981e-01 4.77737904e-01 1.97598028e+00
5.96745312e-01 -6.71625257e-01 6.61604881e-01 5.26707232e-01
-4.70341504e-01 -1.13669448e-01 -9.30131078e-01 -6.73553348e-01
5.20167649e-01 -6.65015638e-01 6.19208097e-01 1.15895927e+00
6.27253175e-01 5.25487363e-01 4.67161208e-01 4.78316471e-02
3.02653611e-01 -3.10283482e-01 5.22250712e-01 -1.76531303e+00
-3.02156836e-01 -5.40218830e-01 -4.16822910e-01 -1.56656697e-01
4.59507734e-01 -1.02770007e+00 -5.52271567e-02 -1.70489097e+00
-6.56791449e-01 -5.96257210e-01 4.22285408e-01 6.22107327e-01
1.80979922e-01 6.08927943e-02 3.09234917e-01 -1.37682140e-01
-4.32678044e-01 -6.82869703e-02 8.76270175e-01 5.12731493e-01
-5.44300258e-01 1.00916743e-01 -4.33758676e-01 1.20765030e+00
8.72915626e-01 -7.71737337e-01 -5.90165071e-02 -3.61596614e-01
5.05544543e-01 -1.89799905e-01 -1.80117890e-01 -6.23758852e-01
1.84127226e-01 -2.64167190e-01 4.83108461e-02 -3.89754415e-01
-4.33265090e-01 -1.01656294e+00 2.36766130e-01 3.96762848e-01
-1.86199278e-01 1.86628252e-01 2.82475561e-01 -4.28820372e-01
-2.71528691e-01 -1.32800972e+00 7.83623040e-01 -3.95022929e-01
-4.77610648e-01 -9.43068191e-02 -8.33575368e-01 2.38582119e-01
1.18199980e+00 -4.31738287e-01 1.03178896e-01 1.79133579e-01
-1.44402921e+00 5.05707152e-02 6.40502810e-01 -1.10253982e-01
-1.17081534e-02 -1.08702946e+00 -6.35597110e-01 2.96932030e-02
2.44388908e-01 3.54060203e-01 -2.98094720e-01 5.91624200e-01
-8.22214663e-01 4.71541315e-01 -8.52374658e-02 -1.92061216e-01
-1.57604551e+00 3.45204085e-01 1.09093711e-01 -7.37397075e-01
-6.36439919e-01 4.10933793e-01 -2.81384081e-01 -8.77936482e-01
2.53521144e-01 -4.24497664e-01 -1.07894123e+00 3.10084194e-01
1.18500374e-01 1.06336504e-01 3.85714024e-01 -1.09784329e+00
-2.29647636e-01 4.93585378e-01 2.38937929e-01 -3.57251734e-01
1.36780703e+00 -1.63974464e-01 -5.24627805e-01 3.86800319e-01
5.67322791e-01 9.61467326e-01 -4.54657912e-01 -1.34441197e-01
7.08733797e-01 -6.22474737e-02 -4.75650609e-01 -5.39647400e-01
-6.34109259e-01 4.95558113e-01 -2.50522885e-02 5.20316958e-01
8.71082246e-01 3.10403436e-01 3.11726123e-01 1.06338337e-01
1.48709744e-01 -1.36810338e+00 -7.13506818e-01 1.06342638e+00
6.44662559e-01 -5.65917253e-01 -2.32053578e-01 -1.11680412e+00
-4.56756264e-01 1.54029047e+00 5.73150635e-01 5.04531385e-03
6.43042445e-01 6.27554476e-01 2.00512350e-01 -2.92679310e-01
-5.56817889e-01 -7.02423394e-01 1.30588934e-01 1.15549147e+00
1.04746616e+00 5.51579446e-02 -1.29540932e+00 8.81173134e-01
-6.57000005e-01 -1.39143124e-01 9.37825561e-01 9.34595406e-01
-4.64367270e-01 -2.18791890e+00 -4.75959182e-01 -2.26948500e-01
-9.34419572e-01 -2.07443327e-01 -4.96554345e-01 1.48471403e+00
6.27539456e-01 7.32096016e-01 2.00914196e-03 3.64104182e-01
5.88944077e-01 3.76824498e-01 8.48414600e-01 -1.62081265e+00
-1.42530632e+00 3.95118684e-01 6.75936580e-01 1.22697521e-02
-7.23990262e-01 -7.33628929e-01 -1.74095643e+00 9.34208483e-02
-2.03013226e-01 6.95062816e-01 6.36259675e-01 1.16433024e+00
-8.12055051e-01 2.28456825e-01 6.93046004e-02 -4.51128095e-01
3.80463488e-02 -7.90979445e-01 -3.98616731e-01 -5.61616058e-03
-2.69394815e-01 -1.85309187e-01 1.27514303e-01 4.17149991e-01] | [10.394126892089844, 10.120427131652832] |
0c54777b-a765-4d8a-b843-1dafc384685e | non-uniform-speaker-disentanglement-for | 2306.01861 | null | https://arxiv.org/abs/2306.01861v2 | https://arxiv.org/pdf/2306.01861v2.pdf | Non-uniform Speaker Disentanglement For Depression Detection From Raw Speech Signals | While speech-based depression detection methods that use speaker-identity features, such as speaker embeddings, are popular, they often compromise patient privacy. To address this issue, we propose a speaker disentanglement method that utilizes a non-uniform mechanism of adversarial SID loss maximization. This is achieved by varying the adversarial weight between different layers of a model during training. We find that a greater adversarial weight for the initial layers leads to performance improvement. Our approach using the ECAPA-TDNN model achieves an F1-score of 0.7349 (a 3.7% improvement over audio-only SOTA) on the DAIC-WoZ dataset, while simultaneously reducing the speaker-identification accuracy by 50%. Our findings suggest that identifying depression through speech signals can be accomplished without placing undue reliance on a speaker's identity, paving the way for privacy-preserving approaches of depression detection. | ['Abeer Alwan', 'Vijay Ravi', 'Jinhan Wang'] | 2023-06-02 | null | null | null | null | ['disentanglement', 'speaker-identification'] | ['methodology', 'speech'] | [ 3.33241045e-01 3.71606916e-01 -1.61001489e-01 -6.18115962e-01
-1.30504537e+00 -6.78214669e-01 3.41033906e-01 3.25425953e-01
-5.76587915e-01 6.22685373e-01 4.08733577e-01 -4.99352008e-01
1.75865427e-01 -4.77062404e-01 -2.32540473e-01 -7.53236592e-01
-1.63521498e-01 1.03807254e-02 -5.19722462e-01 2.17754859e-03
-1.53903574e-01 4.80730295e-01 -8.47347975e-01 3.39631110e-01
6.52127087e-01 7.31222570e-01 -9.45577085e-01 6.72098696e-01
1.69349521e-01 4.43232477e-01 -9.51083302e-01 -8.07087123e-01
1.99807897e-01 -3.70824724e-01 -5.73537529e-01 -3.66776377e-01
4.91633832e-01 -7.17707098e-01 -5.45036197e-01 9.47334349e-01
9.48965311e-01 -2.51810580e-01 4.37297612e-01 -1.20872009e+00
-3.47063363e-01 5.57788670e-01 -4.98509794e-01 1.88686475e-01
3.35650444e-01 -4.41371687e-02 8.77055168e-01 -4.69878465e-01
3.64173092e-02 1.15161395e+00 8.51303458e-01 1.00715292e+00
-1.84313893e+00 -1.29360712e+00 -1.32684946e-01 -2.36716777e-01
-1.83909357e+00 -1.01841283e+00 7.77672529e-01 -3.52736443e-01
5.10714948e-01 5.44556260e-01 3.69193703e-01 1.28129566e+00
4.86430749e-02 4.75459039e-01 7.69480228e-01 -1.90807670e-01
1.35402694e-01 3.69598746e-01 5.24506681e-02 6.28821969e-01
1.97241798e-01 2.94225849e-02 -7.24601924e-01 -9.52500701e-01
3.80606860e-01 -6.19027130e-02 -2.20247716e-01 -2.32521176e-01
-7.92405486e-01 1.01258767e+00 -6.15932904e-02 1.47886574e-01
-1.80815563e-01 -1.20165730e-02 7.04711258e-01 4.17779088e-01
5.83026052e-01 3.85642111e-01 -1.68717667e-01 6.56308085e-02
-9.90674436e-01 2.08804801e-01 8.12174499e-01 3.35502207e-01
6.86795190e-02 1.65913492e-01 -1.97878554e-01 7.10169613e-01
1.68821752e-01 4.72322375e-01 5.01006067e-01 -8.37646008e-01
3.90672505e-01 1.72631606e-01 -9.63070095e-02 -1.04530239e+00
-1.93603024e-01 -2.69287527e-01 -7.45382309e-01 2.26729408e-01
4.10468847e-01 -6.79068148e-01 -5.78477323e-01 2.20582509e+00
3.06579620e-01 1.48405030e-01 1.92477465e-01 3.36019754e-01
4.58767146e-01 1.50158405e-01 2.42237836e-01 -2.97361966e-02
1.34750926e+00 -2.59196639e-01 -7.85523593e-01 -9.96592566e-02
5.49000204e-01 -5.59413373e-01 6.29387617e-01 3.00733060e-01
-9.55519617e-01 9.73676071e-02 -1.20868433e+00 1.02653034e-01
-1.58668727e-01 -2.42080569e-01 5.31458557e-01 1.49787891e+00
-1.06878972e+00 3.38066518e-01 -9.53557312e-01 -1.26579076e-01
7.81914771e-01 7.76194990e-01 -7.67769456e-01 7.13718086e-02
-1.37408090e+00 6.65517271e-01 -1.15163401e-01 -2.38687783e-01
-7.73545980e-01 -8.11133444e-01 -9.03241456e-01 2.10770071e-01
-1.66568503e-01 -5.62565029e-01 1.03775072e+00 -8.92501831e-01
-1.33289790e+00 1.08384323e+00 -2.36010671e-01 -8.14840078e-01
6.46420360e-01 1.50127858e-01 -7.68806219e-01 2.49979138e-01
-7.95722380e-02 5.04716039e-01 1.06132019e+00 -8.46089542e-01
-3.58329266e-01 -7.19757259e-01 -2.38917530e-01 1.03367055e-02
-6.89380586e-01 2.51776457e-01 3.40257555e-01 -7.04126477e-01
-1.62791103e-01 -8.37100625e-01 -8.06135461e-02 3.21895957e-01
-6.93070590e-01 2.40697369e-01 6.78009033e-01 -1.00374138e+00
1.20840371e+00 -2.53782582e+00 -4.63202268e-01 3.56094122e-01
6.33670211e-01 5.07272780e-01 -7.07989335e-02 2.09915534e-01
-3.26222718e-01 3.96502912e-01 -2.73671210e-01 -5.89564264e-01
-7.07174763e-02 -3.83626878e-01 -4.16721366e-02 8.41018617e-01
1.42236158e-01 4.28482026e-01 -3.98048222e-01 -2.53045738e-01
-2.20699329e-03 7.60077178e-01 -8.46795201e-01 -1.63695123e-02
6.46473408e-01 1.69311181e-01 -2.69430518e-01 5.44304788e-01
9.33563232e-01 3.61308604e-01 4.04155761e-01 1.85156286e-01
2.86474407e-01 5.16447663e-01 -8.40517938e-01 1.51981020e+00
-1.73092276e-01 7.68319547e-01 4.98092771e-01 -9.44295168e-01
9.63008761e-01 6.89958453e-01 5.70074975e-01 -3.02859962e-01
2.20869049e-01 -6.00848049e-02 2.69671857e-01 -2.75418878e-01
2.82402765e-02 -5.50187826e-01 -1.31722465e-01 2.54656672e-01
-1.50502712e-01 4.14362580e-01 -5.78975618e-01 7.81429037e-02
1.36793649e+00 -9.51048076e-01 1.57947853e-01 -3.46166998e-01
3.15940350e-01 -4.73037809e-01 6.59434974e-01 5.96269786e-01
-8.43153656e-01 6.32794678e-01 6.24176264e-01 -7.20406175e-02
-7.05464542e-01 -1.24741781e+00 -4.80035961e-01 7.51676977e-01
-5.96210837e-01 -2.88098127e-01 -9.14708555e-01 -8.40920866e-01
5.30963182e-01 8.16318452e-01 -8.54578137e-01 -6.54222727e-01
-2.83552080e-01 -7.07520902e-01 1.45539296e+00 2.73127407e-01
2.23609969e-01 -2.35948190e-01 -9.53775421e-02 1.91453382e-01
-1.04264662e-01 -7.07886338e-01 -8.75346780e-01 1.44215599e-01
-5.17249048e-01 -8.39219809e-01 -8.34455907e-01 -5.21776974e-01
4.62425888e-01 -1.40430838e-01 6.37951910e-01 -4.86062825e-01
-4.49535817e-01 4.39406395e-01 2.34929323e-01 -6.45762861e-01
-6.43627882e-01 1.43579647e-01 3.23635221e-01 2.75219470e-01
8.26953053e-01 -7.36623645e-01 -6.72662199e-01 4.59962562e-02
-6.03912175e-01 -5.64742923e-01 2.45858394e-02 8.01439524e-01
1.16876416e-01 -2.19964877e-01 1.18015337e+00 -9.31199253e-01
8.58083844e-01 -4.19634104e-01 -1.69736892e-01 8.43653362e-03
-7.20623374e-01 -1.06626011e-01 3.18429440e-01 -6.54701948e-01
-8.35808039e-01 1.22280389e-01 -4.30658937e-01 -3.93015534e-01
-1.46732971e-01 1.29568279e-01 -3.25015366e-01 -2.48149544e-01
7.68481851e-01 1.16272688e-01 6.40438437e-01 -3.21511835e-01
1.48462281e-01 1.06229627e+00 5.31981289e-01 -1.17307857e-01
5.22828877e-01 4.63141233e-01 -3.69968623e-01 -9.24982488e-01
-4.11938369e-01 -3.40864360e-01 -8.17747861e-02 3.19566667e-01
9.90547836e-01 -1.09346175e+00 -1.09357834e+00 3.53258640e-01
-9.28836644e-01 1.61722928e-01 1.08075272e-02 5.68238914e-01
-1.23743266e-01 2.25570232e-01 -6.45603836e-01 -8.64856422e-01
-6.13188684e-01 -8.41828644e-01 6.33670747e-01 -3.22896540e-01
-8.04672897e-01 -9.55021203e-01 2.15161145e-01 5.03820360e-01
4.10704017e-01 5.47229946e-01 9.29473042e-01 -1.23787534e+00
1.74516574e-01 -7.17783511e-01 9.40752774e-02 6.23623610e-01
5.22699237e-01 -3.35256487e-01 -1.44776738e+00 -4.09429967e-01
2.42000490e-01 -2.76671916e-01 5.87618947e-01 3.29652935e-01
1.07905507e+00 -6.12215281e-01 -2.64270276e-01 6.26791358e-01
9.35678065e-01 3.95244718e-01 5.69666684e-01 -5.23620211e-02
6.05131805e-01 5.43274641e-01 -1.81938425e-01 5.69731951e-01
2.86753774e-01 5.14769852e-01 7.52194524e-02 -3.13373119e-01
1.65450349e-01 -3.81756693e-01 4.60293710e-01 2.38170475e-01
6.02371931e-01 -3.84024605e-02 -8.51990223e-01 5.27795672e-01
-1.23519421e+00 -9.73868489e-01 3.34710449e-01 2.44288921e+00
1.09343970e+00 -4.73429859e-02 5.43153882e-01 2.53097028e-01
8.38674486e-01 2.08784208e-01 -6.84254229e-01 -7.08354473e-01
-6.72671851e-03 1.83313549e-01 6.52883470e-01 5.40402532e-01
-1.03467441e+00 4.58453387e-01 6.96476030e+00 4.88334149e-01
-1.26141870e+00 7.33587891e-02 8.02650809e-01 -6.28202975e-01
-2.47518167e-01 -6.22160554e-01 -5.94595015e-01 5.54613531e-01
1.23630190e+00 -6.16619587e-01 3.13114285e-01 5.31205535e-01
3.28039080e-01 2.98779935e-01 -1.32663727e+00 1.08769870e+00
2.73389846e-01 -1.14808059e+00 -1.59245148e-01 4.10836250e-01
2.04011410e-01 -2.06493586e-01 4.62561339e-01 -3.84636149e-02
2.14373454e-01 -1.27365077e+00 2.89541632e-01 1.27135202e-01
1.10087323e+00 -1.27257359e+00 5.47643423e-01 -1.60383154e-02
-4.89926487e-01 7.61806220e-02 3.97504587e-03 3.70279700e-01
-1.28250927e-01 5.43245554e-01 -1.30428386e+00 8.11256841e-02
4.37732011e-01 1.47865653e-01 -1.72487602e-01 6.26381278e-01
1.56091094e-01 9.84321177e-01 -3.66276890e-01 2.00065985e-01
-1.58372074e-01 2.14407012e-01 5.48174977e-01 1.21932280e+00
1.61389038e-01 -8.47639814e-02 -2.49060154e-01 8.02300096e-01
-4.07418042e-01 5.52345999e-02 -9.43823278e-01 -2.86781877e-01
6.70085073e-01 8.13833237e-01 1.30170286e-01 4.20759581e-02
-2.70536870e-01 1.10784340e+00 2.64757294e-02 1.16895176e-01
-5.94160080e-01 -6.78632915e-01 1.38519716e+00 1.81833208e-01
-8.99337232e-02 1.76930875e-01 -4.94613886e-01 -9.31182206e-01
-1.16669662e-01 -1.30087817e+00 5.01369476e-01 5.29168397e-02
-1.18992662e+00 3.39848638e-01 -4.91257787e-01 -8.54691446e-01
-3.93651068e-01 -1.21604040e-01 -5.28524756e-01 1.20399737e+00
-1.05603063e+00 -9.43859637e-01 4.71317977e-01 7.11809218e-01
-9.17810351e-02 -2.65403122e-01 1.41041088e+00 6.13359451e-01
-8.56496811e-01 1.62984574e+00 1.34908065e-01 4.90052819e-01
9.65396523e-01 -9.18348551e-01 3.80269110e-01 5.21363676e-01
-4.25903499e-02 9.40975249e-01 5.72681665e-01 -2.31765524e-01
-9.09759700e-01 -9.68195081e-01 1.11801815e+00 -4.59002256e-01
5.01403928e-01 -8.32592785e-01 -1.00614595e+00 7.41991103e-01
7.85349384e-02 -2.86854386e-01 1.46145248e+00 3.97017449e-01
-6.85677469e-01 -3.14160854e-01 -1.89345694e+00 5.78874648e-01
5.40191472e-01 -1.09979939e+00 -3.41807365e-01 9.57111418e-02
7.75284827e-01 -2.71641277e-02 -1.06101239e+00 -3.46437320e-02
7.91822970e-01 -6.49975419e-01 1.19401276e+00 -8.11255276e-01
5.90312593e-02 2.04597950e-01 -2.66089439e-01 -1.18541837e+00
-1.60583869e-01 -8.13407302e-01 1.15941860e-01 1.56738079e+00
7.14745581e-01 -1.02833438e+00 1.06024766e+00 1.17259312e+00
4.51135933e-01 -4.16763961e-01 -1.15628278e+00 -6.04872167e-01
4.00514007e-01 -3.21595192e-01 6.42493367e-01 1.42287946e+00
3.98851305e-01 2.02798888e-01 -3.99305105e-01 3.88864160e-01
6.83952153e-01 -7.35308349e-01 4.64483321e-01 -1.06544006e+00
-1.27115831e-01 -2.24338651e-01 -7.00405478e-01 -1.51704937e-01
4.09245074e-01 -9.75863636e-01 -3.80693674e-01 -7.53822088e-01
7.68140256e-02 -3.03830504e-01 -5.51381826e-01 6.63615048e-01
-9.05658975e-02 9.57812369e-02 -8.75698105e-02 -3.61942649e-01
7.95858875e-02 3.61187965e-01 5.20851910e-01 -3.08581024e-01
-2.74464875e-01 2.97022104e-01 -1.29137349e+00 5.41022420e-01
1.08718276e+00 -8.29414427e-01 -4.87074167e-01 -1.43886775e-01
-5.09169638e-01 7.29022399e-02 2.89616525e-01 -6.74885094e-01
1.09801076e-01 1.53200105e-01 3.64784241e-01 -6.54213727e-02
5.46690524e-01 -5.73788762e-01 3.10811438e-02 6.80767894e-01
-7.42020965e-01 -7.24828616e-02 6.10599697e-01 5.65801859e-01
-5.09266071e-02 4.22309153e-02 9.78994370e-01 2.64632851e-01
3.80488396e-01 1.47146702e-01 -6.51707709e-01 1.43557921e-01
8.28261852e-01 1.00792065e-01 -1.56148359e-01 -5.10343134e-01
-8.57547939e-01 3.32494043e-02 2.17041329e-01 1.12043425e-01
4.27956849e-01 -1.12933815e+00 -8.20880294e-01 5.35756290e-01
1.53519481e-01 -5.73139668e-01 2.77572036e-01 6.21937394e-01
-1.50897607e-01 1.73453212e-01 1.07891690e-02 -2.60860562e-01
-1.79498374e+00 4.09723997e-01 6.38069630e-01 1.07072771e-01
-4.42925930e-01 9.82990384e-01 2.25964129e-01 -4.09766734e-01
5.29564142e-01 1.21019930e-01 2.60885328e-01 1.11616552e-01
8.63852859e-01 4.48046207e-01 1.87384412e-01 -3.18718702e-01
-7.28478193e-01 -2.56285012e-01 -4.84345496e-01 -4.57399189e-01
1.05881464e+00 -6.01068102e-02 2.03387067e-01 1.93194583e-01
1.66884506e+00 4.87366408e-01 -9.10561204e-01 -1.72513183e-02
-2.76149750e-01 -4.14862275e-01 2.51768738e-01 -8.77332270e-01
-9.88260090e-01 7.77003109e-01 1.16554248e+00 3.34934801e-01
9.17659938e-01 -2.16567606e-01 7.11732209e-01 2.93617487e-01
-7.48430043e-02 -7.30351329e-01 -4.40401256e-01 -2.18873814e-01
4.96258974e-01 -1.14802754e+00 -3.03861380e-01 -1.59373090e-01
-5.98368585e-01 5.25556564e-01 1.21896684e-01 2.10628256e-01
7.02380955e-01 5.08908212e-01 4.09348160e-01 5.11230975e-02
-4.33674961e-01 4.21475470e-01 7.59492442e-03 9.21783805e-01
4.25511509e-01 4.67191547e-01 -1.16060309e-01 8.32754493e-01
-3.00439298e-01 -2.26521730e-01 3.25118363e-01 7.12158680e-01
1.56320408e-02 -1.25198400e+00 -3.02491575e-01 5.60067117e-01
-1.05972528e+00 -2.79576719e-01 -6.34293199e-01 4.46708888e-01
-1.97851330e-01 1.11044407e+00 1.31628618e-01 -4.00029302e-01
4.21054929e-01 4.58158225e-01 1.29778221e-01 -4.07269537e-01
-7.83034921e-01 -9.44177955e-02 4.27479744e-01 -4.22537327e-01
6.07586466e-02 -1.06357455e+00 -9.61200774e-01 -8.69017541e-01
-8.55234936e-02 7.12845698e-02 7.33869910e-01 5.30550659e-01
7.30349779e-01 3.92838210e-01 9.25863206e-01 -8.34887400e-02
-8.15687954e-01 -6.30602121e-01 -9.31168616e-01 8.38942975e-02
8.95481884e-01 -1.82664543e-01 -6.09630466e-01 -2.81351030e-01] | [14.004011154174805, 5.885194301605225] |
df22bf3b-1959-4b6b-a6ab-2de8547a127b | interaction-part-mining-a-mid-level-approach | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Zhou_Interaction_Part_Mining_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Zhou_Interaction_Part_Mining_2015_CVPR_paper.pdf | Interaction Part Mining: A Mid-Level Approach for Fine-Grained Action Recognition | Modeling human-object interactions and manipulating motions lies in the heart of fine-grained action recognition. Previous methods heavily rely on explicit detection of the object being interacted, which requires intensive human labour on object annotation. To bypass this constraint and achieve better classification performance, in this work, we propose a novel fine-grained action recognition pipeline by interaction part proposal and discriminative mid-level part mining. Firstly, we generate a large number of candidate object regions using off-the-shelf object proposal tool, e.g., BING. Secondly, these object regions are matched and tracked across frames to form a large spatio-temporal graph based on the appearance matching and the dense motion trajectories through them. We then propose an efficient approximate graph segmentation algorithm to partition and filter the graph into consistent local dense sub-graphs. These sub-graphs, which are spatio-temporal sub-volumes, represent our candidate interaction parts. Finally, we mine discriminative mid-level part detectors from the features computed over the candidate interaction parts. Bag-of-detection scores based on a novel Max-N pooling scheme are computed as the action representation for a video sample. We conduct extensive experiments on human-object interaction datasets including MPII Cooking and MSR Daily Activity 3D. The experimental results demonstrate that the proposed framework achieves consistent improvements over the state-of-the-art action recognition accuracies on the benchmarks, without using any object annotation. | ['Yang Zhou', 'Richang Hong', 'Bingbing Ni', 'Qi Tian', 'Meng Wang'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['fine-grained-action-recognition'] | ['computer-vision'] | [ 3.91440511e-01 -1.33069336e-01 -3.61567646e-01 -1.94173366e-01
-4.67639208e-01 -2.21677899e-01 4.40216005e-01 1.20287485e-01
-2.57500052e-01 3.77043873e-01 3.55323732e-01 3.34233642e-01
-1.69155329e-01 -7.81697333e-01 -6.38012052e-01 -6.87807322e-01
-2.33597890e-01 4.52003419e-01 9.82532203e-01 1.18994400e-01
3.22134584e-01 5.52983761e-01 -1.59792840e+00 6.93568408e-01
6.22859538e-01 1.20818007e+00 8.34707245e-02 6.76525295e-01
9.30090249e-03 8.69777143e-01 -4.88804162e-01 -3.80489528e-02
2.16033801e-01 -5.79863966e-01 -8.88186336e-01 7.68060625e-01
3.71228784e-01 -4.74152297e-01 -3.71092409e-01 9.77707446e-01
1.27695426e-01 4.96007383e-01 6.47385895e-01 -1.23804545e+00
-1.16239391e-01 3.13328832e-01 -8.71582270e-01 3.68095368e-01
6.27248764e-01 3.24086964e-01 8.26498449e-01 -9.70994830e-01
6.89683735e-01 1.38098788e+00 4.57169443e-01 4.18749660e-01
-9.55445409e-01 -4.22585458e-01 5.45538962e-01 4.23220456e-01
-1.47841978e+00 -2.23345757e-01 7.29535341e-01 -4.98355776e-01
1.11676097e+00 2.22973406e-01 1.03525209e+00 7.21033216e-01
2.70865619e-01 1.07058096e+00 5.44290900e-01 -3.15136045e-01
1.92766681e-01 -5.03149211e-01 1.81585118e-01 1.12345958e+00
1.37872040e-01 -3.51330251e-01 -5.65756977e-01 -1.92020953e-01
9.65579987e-01 3.79523158e-01 -2.42693841e-01 -6.22014046e-01
-1.26450682e+00 4.62576598e-01 3.17811728e-01 2.97548145e-01
-5.89032590e-01 1.79729268e-01 4.57427233e-01 -3.00027519e-01
3.78794491e-01 -4.30104613e-01 -3.68197620e-01 -1.91259190e-01
-8.14110518e-01 1.63640454e-01 5.50189376e-01 9.10862446e-01
7.39784122e-01 -3.67505789e-01 -5.82124531e-01 6.06001675e-01
3.82436484e-01 -3.75927351e-02 4.21296030e-01 -8.22210848e-01
7.33613610e-01 1.33256483e+00 1.21666351e-02 -1.06277299e+00
-2.76507318e-01 7.36850947e-02 -7.45703876e-01 2.03612849e-01
6.02038801e-01 2.75180727e-01 -9.77004468e-01 1.05305576e+00
9.34841335e-01 4.42008883e-01 -3.46452296e-01 1.02962148e+00
6.90492332e-01 4.11439270e-01 3.81888658e-01 -8.00848827e-02
1.57217753e+00 -1.25743842e+00 -4.61457193e-01 -1.45427138e-01
7.38477170e-01 -3.37925464e-01 8.35798442e-01 1.72904596e-01
-9.37112331e-01 -8.38119745e-01 -7.76859105e-01 7.86831677e-02
-2.01248839e-01 3.31479192e-01 7.58195043e-01 3.70255113e-01
-4.05139685e-01 8.51646304e-01 -1.14510787e+00 -3.55702788e-01
8.57846797e-01 2.94461519e-01 -6.00971341e-01 7.57222176e-02
-5.36370099e-01 3.43142211e-01 7.04525590e-01 2.82245249e-01
-9.95687008e-01 -2.17097998e-01 -1.03150010e+00 7.87415802e-02
6.71226561e-01 -4.95955735e-01 9.15111601e-01 -9.44278538e-01
-1.22435522e+00 8.96290362e-01 -1.45325705e-01 -3.80979121e-01
4.39876378e-01 -2.85447150e-01 -1.43611372e-01 3.32035631e-01
9.07995030e-02 4.41245139e-01 8.37397337e-01 -8.15750241e-01
-1.04527974e+00 -6.18237257e-01 1.93042718e-02 2.34332561e-01
-8.92628636e-03 2.26281121e-01 -8.95449996e-01 -7.43548214e-01
2.99786806e-01 -8.63619745e-01 -3.52995694e-01 -6.44923151e-02
-2.90039957e-01 -7.70807564e-01 1.08678782e+00 -7.67633379e-01
1.31142914e+00 -2.21207571e+00 1.88519225e-01 2.89640665e-01
1.50618687e-01 2.11433440e-01 -5.72061799e-02 1.29127070e-01
6.84418231e-02 -2.49359235e-01 -8.95855278e-02 -1.56777337e-01
-3.64080593e-02 1.22056097e-01 1.81509107e-01 7.34386265e-01
1.39356047e-01 1.00743568e+00 -8.91439497e-01 -9.77374077e-01
5.69406033e-01 1.54278502e-01 -3.85165989e-01 2.51543641e-01
-3.31427038e-01 4.62652475e-01 -9.36379910e-01 9.91063058e-01
3.11724454e-01 -3.41253519e-01 1.87838033e-01 -4.54772532e-01
1.22588806e-01 -5.05373962e-02 -1.55954444e+00 1.95319200e+00
3.03557605e-01 9.88333672e-02 -3.06882150e-02 -1.24211478e+00
7.98023582e-01 8.26371536e-02 8.93995285e-01 -4.28111047e-01
2.50168383e-01 -5.38493618e-02 -1.41583949e-01 -5.27109087e-01
2.55159110e-01 3.74675184e-01 -2.37738982e-01 2.39417493e-01
9.09326002e-02 3.33717287e-01 4.91826832e-01 1.76910028e-01
1.37897182e+00 6.39948189e-01 5.99995852e-01 6.40621260e-02
7.79269636e-01 1.97482929e-01 7.71589935e-01 5.37626982e-01
-5.22283375e-01 3.92719895e-01 1.61702543e-01 -6.24575019e-01
-4.58346277e-01 -7.67310917e-01 4.06934351e-01 1.14993358e+00
3.39915961e-01 -5.12922287e-01 -9.77861941e-01 -1.26662266e+00
-4.37394716e-02 1.13022782e-01 -6.93336010e-01 4.93740886e-02
-9.99878705e-01 -4.43113923e-01 1.60048425e-01 8.56175840e-01
7.00502396e-01 -1.53390145e+00 -8.53437424e-01 2.88477182e-01
-1.76474303e-01 -1.08601403e+00 -8.24712753e-01 -3.07392310e-02
-9.49489057e-01 -1.43028736e+00 -5.41383266e-01 -9.03035223e-01
7.74795413e-01 3.64728659e-01 9.42449808e-01 1.93212926e-01
-7.09679782e-01 5.00138104e-01 -5.18112957e-01 7.67645836e-02
2.28375699e-02 -4.41131145e-01 -1.14959851e-01 4.91270274e-01
5.83525777e-01 -4.13457006e-01 -8.59543085e-01 5.25943756e-01
-5.54044366e-01 2.96875220e-02 7.11104155e-01 5.34329593e-01
1.20283067e+00 2.34858006e-01 3.15626934e-02 -4.64105040e-01
1.07788809e-01 -1.09187856e-01 -5.35728276e-01 3.31707150e-01
7.11234733e-02 -9.18801278e-02 2.24996120e-01 -6.50901020e-01
-1.01203966e+00 7.36252069e-01 3.33740771e-01 -6.62784517e-01
-5.28634310e-01 2.05386966e-01 -4.31432396e-01 1.69788040e-02
3.13995749e-01 2.87788391e-01 -3.80258292e-01 -5.09494841e-01
4.09865826e-01 4.57079321e-01 5.12312472e-01 -3.60004067e-01
5.19351840e-01 6.00180686e-01 2.20727697e-02 -7.27980077e-01
-7.02424288e-01 -9.00641799e-01 -1.13913369e+00 -5.09576201e-01
1.51234341e+00 -7.81578362e-01 -7.57715166e-01 4.60713476e-01
-1.03421056e+00 -4.00905132e-01 -4.04317081e-01 6.09089911e-01
-6.76202714e-01 6.19979858e-01 -4.98959035e-01 -8.71505558e-01
-3.25765878e-01 -9.55048025e-01 1.54772198e+00 2.38470629e-01
-4.22257155e-01 -5.26962817e-01 7.70148262e-02 6.88013911e-01
-4.86352354e-01 5.42176366e-01 3.78763437e-01 -4.81814414e-01
-8.56634080e-01 -3.51098508e-01 -1.46171406e-01 1.62631512e-01
2.06407100e-01 -1.51186809e-01 -5.39955616e-01 -9.78447422e-02
-3.92640918e-01 -1.58243462e-01 8.52353454e-01 5.24638712e-01
1.37860298e+00 -1.48111030e-01 -8.58620465e-01 3.61339599e-01
1.06698501e+00 2.96855122e-01 6.14465237e-01 -3.27437557e-02
1.15775239e+00 6.01390839e-01 1.12932813e+00 7.77190864e-01
1.54857889e-01 8.57411742e-01 2.93588281e-01 8.03371668e-02
-2.34736189e-01 -2.67586291e-01 4.47414726e-01 3.12268347e-01
-5.82615256e-01 -7.96704553e-03 -5.38145781e-01 5.13330340e-01
-2.26431179e+00 -1.27130270e+00 -3.02697778e-01 1.93268549e+00
5.37653089e-01 2.24151984e-01 5.17971039e-01 2.45667920e-01
7.28614688e-01 1.49932429e-01 -4.23149347e-01 1.59349412e-01
3.51123929e-01 2.67584205e-01 3.13263774e-01 -8.64796899e-03
-1.53888357e+00 1.08158302e+00 4.58858633e+00 9.12961543e-01
-4.14706051e-01 9.08347815e-02 3.54114622e-01 -6.24991208e-02
5.59798539e-01 -1.13908395e-01 -9.16270852e-01 3.29206407e-01
4.41955447e-01 4.84528214e-01 1.29116222e-01 1.05848336e+00
4.31319326e-01 -4.26246226e-01 -1.23953700e+00 1.11015773e+00
6.18734695e-02 -1.24269521e+00 -8.86882022e-02 7.54570290e-02
5.26033878e-01 -2.61517733e-01 -7.41607368e-01 2.61738449e-01
1.84917748e-01 -7.11042941e-01 8.41181219e-01 6.28756702e-01
4.29888904e-01 -6.23723030e-01 3.98158252e-01 4.89419460e-01
-2.09145141e+00 -1.59992650e-01 -1.76745832e-01 -4.15942781e-02
1.99596211e-01 2.94557780e-01 -4.30795431e-01 5.20572484e-01
9.35416579e-01 8.17184210e-01 -5.51612079e-01 1.02110684e+00
-9.12595913e-03 4.42652941e-01 -1.86691523e-01 -1.11797571e-01
1.75105438e-01 -2.88096339e-01 2.52653718e-01 1.29266953e+00
1.18072219e-01 6.24356687e-01 7.60494113e-01 5.55446982e-01
1.45581111e-01 3.01070306e-02 -4.22699362e-01 -1.13655761e-01
-3.65297981e-02 1.37660336e+00 -1.18663955e+00 -5.35720050e-01
-6.78824663e-01 1.34610546e+00 4.76532102e-01 1.16493076e-01
-9.92040992e-01 -3.55518609e-01 5.71074963e-01 2.58108139e-01
7.53651321e-01 -3.34760547e-01 1.19794004e-01 -1.09971774e+00
3.34332079e-01 -8.24602008e-01 8.44799995e-01 -4.00162876e-01
-1.07269216e+00 1.53211743e-01 8.74026418e-02 -1.36694825e+00
-8.51385593e-02 -4.33307201e-01 -6.34910762e-01 5.14506876e-01
-6.79321349e-01 -1.32018685e+00 -7.43389666e-01 1.01810265e+00
9.31029439e-01 1.42847121e-01 6.15761459e-01 2.54075170e-01
-6.79274023e-01 2.33373448e-01 -5.72710872e-01 5.42140722e-01
8.91089588e-02 -9.17365253e-01 3.41640949e-01 7.86994696e-01
4.62876529e-01 1.79547668e-01 8.18259344e-02 -1.18096089e+00
-1.31568539e+00 -1.31395376e+00 5.94708979e-01 -5.50880969e-01
4.51728821e-01 -4.36356038e-01 -8.41123879e-01 5.73750675e-01
-2.78923899e-01 5.53867936e-01 3.11632454e-01 -2.87418723e-01
3.21188271e-02 1.01724409e-01 -1.00906098e+00 4.50849414e-01
1.69902134e+00 -2.07221672e-01 -6.92677736e-01 5.11297822e-01
2.29471028e-01 -3.17137092e-01 -7.89604127e-01 4.98142749e-01
6.24406755e-01 -8.70791256e-01 9.97560084e-01 -7.76145101e-01
-1.53826950e-02 -6.33353591e-01 -3.30019593e-02 -6.15652263e-01
-5.88943839e-01 -4.54049706e-01 -5.84020138e-01 9.81951535e-01
-2.20692098e-01 7.22228258e-04 1.16219187e+00 3.88098717e-01
-2.37153918e-01 -8.73747408e-01 -8.58957589e-01 -7.59264886e-01
-8.08011174e-01 -4.89306778e-01 3.58224243e-01 4.82105821e-01
1.42695904e-01 1.14218824e-01 -1.11756437e-01 2.05804333e-01
7.56722271e-01 3.10154498e-01 1.17512202e+00 -1.11354697e+00
-4.70052004e-01 -3.23655158e-01 -8.77167225e-01 -1.24752045e+00
1.80216506e-01 -6.26948953e-01 3.22093308e-01 -1.81428516e+00
4.98228282e-01 3.02330847e-03 -1.37037933e-01 6.62894785e-01
-2.01857448e-01 4.24354166e-01 -2.67106462e-02 2.72737533e-01
-1.21069849e+00 4.45061088e-01 1.32712817e+00 -2.75737047e-01
-4.12708759e-01 5.89658171e-02 -6.55021071e-02 1.02449751e+00
5.62362015e-01 -3.61914426e-01 -3.06840062e-01 1.96791574e-01
-6.28209829e-01 3.91792767e-02 6.25504375e-01 -1.32297993e+00
4.98165786e-02 -4.31503236e-01 6.28745317e-01 -8.89361918e-01
4.41482693e-01 -8.54880929e-01 1.96532384e-01 5.59380472e-01
6.59210160e-02 -2.74406254e-01 1.96448453e-02 8.68198574e-01
-1.46623030e-01 2.97105149e-03 7.21515477e-01 -4.04670328e-01
-1.16956353e+00 6.60099030e-01 -2.51332939e-01 6.04940113e-03
1.60561764e+00 -6.46520257e-01 1.76234931e-01 1.81748137e-01
-9.87314939e-01 2.32093483e-02 2.08059520e-01 5.47808707e-01
7.30067372e-01 -1.48860478e+00 -4.86235082e-01 1.88510895e-01
2.07111627e-01 -2.45684069e-02 3.81221175e-01 9.79932010e-01
-2.66997278e-01 3.27679813e-01 -1.50882483e-01 -7.37576604e-01
-1.84712911e+00 5.09092748e-01 1.79399267e-01 -4.46072876e-01
-1.09087074e+00 8.35018575e-01 4.69378054e-01 1.82470471e-01
3.73570442e-01 -7.91536689e-01 -3.14562142e-01 -4.74307090e-02
6.16813064e-01 6.14773095e-01 -2.38270655e-01 -9.38351691e-01
-8.04404914e-01 7.29778707e-01 2.56350368e-01 1.35296404e-01
1.18322766e+00 1.81151554e-02 4.40683439e-02 6.52946755e-02
9.06419218e-01 -2.27157697e-01 -1.57659984e+00 -1.27160579e-01
3.14034447e-02 -7.30197489e-01 -2.16200396e-01 -3.45162839e-01
-1.06539512e+00 5.44945776e-01 6.58805907e-01 1.43075213e-01
1.16001558e+00 4.27819967e-01 9.06829417e-01 1.03366829e-01
5.85314274e-01 -1.21105909e+00 4.01262194e-01 3.06642979e-01
8.22123826e-01 -1.06989801e+00 2.07170218e-01 -6.54265583e-01
-5.62556684e-01 1.03597999e+00 8.32386017e-01 -2.61082739e-01
6.18730366e-01 -4.94699031e-02 -3.79850477e-01 -4.86573964e-01
-2.41624638e-01 -3.94957066e-01 6.68568671e-01 5.00430226e-01
2.02563241e-01 1.53562754e-01 -2.93696225e-01 6.10297382e-01
4.25522327e-01 1.19410589e-01 -1.23301394e-01 1.14707315e+00
-6.13672376e-01 -8.74249101e-01 -4.05140162e-01 5.47992289e-01
-3.48872274e-01 3.78912717e-01 -5.11295974e-01 7.01838791e-01
5.80538630e-01 9.28939939e-01 5.57503849e-02 -4.33385104e-01
5.85094392e-01 1.81989633e-02 7.98652351e-01 -7.98484802e-01
-5.40958285e-01 1.47161677e-01 1.32951692e-01 -1.22861183e+00
-6.38578951e-01 -9.66546774e-01 -1.68764031e+00 1.94543287e-01
-4.23524976e-01 -1.50472179e-01 1.00930743e-01 1.09977555e+00
3.16697329e-01 5.22548199e-01 7.49698877e-02 -1.33852220e+00
-2.38343626e-01 -1.03095281e+00 -6.49186075e-01 7.09409237e-01
-9.80147198e-02 -9.59696531e-01 -2.59714369e-02 3.87004077e-01] | [8.175671577453613, 0.46758249402046204] |
4f98f28a-baeb-406b-9756-41547d651496 | mbore-multi-objective-bayesian-optimisation | 2203.16912 | null | https://arxiv.org/abs/2203.16912v1 | https://arxiv.org/pdf/2203.16912v1.pdf | MBORE: Multi-objective Bayesian Optimisation by Density-Ratio Estimation | Optimisation problems often have multiple conflicting objectives that can be computationally and/or financially expensive. Mono-surrogate Bayesian optimisation (BO) is a popular model-based approach for optimising such black-box functions. It combines objective values via scalarisation and builds a Gaussian process (GP) surrogate of the scalarised values. The location which maximises a cheap-to-query acquisition function is chosen as the next location to expensively evaluate. While BO is an effective strategy, the use of GPs is limiting. Their performance decreases as the problem input dimensionality increases, and their computational complexity scales cubically with the amount of data. To address these limitations, we extend previous work on BO by density-ratio estimation (BORE) to the multi-objective setting. BORE links the computation of the probability of improvement acquisition function to that of probabilistic classification. This enables the use of state-of-the-art classifiers in a BO-like framework. In this work we present MBORE: multi-objective Bayesian optimisation by density-ratio estimation, and compare it to BO across a range of synthetic and real-world benchmarks. We find that MBORE performs as well as or better than BO on a wide variety of problems, and that it outperforms BO on high-dimensional and real-world problems. | ['Alma A. M. Rahat', 'Tinkle Chugh', 'George De Ath'] | 2022-03-31 | null | null | null | null | ['density-ratio-estimation', 'bayesian-optimisation'] | ['methodology', 'methodology'] | [ 1.23255059e-01 -2.15731487e-01 -1.20512143e-01 -3.27562571e-01
-1.42590868e+00 -4.43406403e-01 8.03838670e-01 3.68384570e-01
-7.24869549e-01 1.00182700e+00 -4.95697968e-02 -2.15779126e-01
-8.73601019e-01 -7.30354071e-01 -6.27570093e-01 -1.16933286e+00
-2.87960142e-01 1.11628318e+00 3.08261842e-01 -1.68116391e-02
5.33178568e-01 3.61375749e-01 -1.51483619e+00 -2.01207280e-01
6.52697146e-01 1.35064006e+00 4.87998687e-02 1.02061248e+00
2.68035531e-01 2.04530731e-01 -8.74784827e-01 -5.66919386e-01
2.36690328e-01 -1.71602704e-02 -7.45353699e-01 -2.82750577e-01
-1.45127729e-01 2.06572354e-01 2.82796562e-01 6.49440706e-01
9.57525074e-01 5.16221702e-01 9.96443570e-01 -1.38700819e+00
-1.40267119e-01 1.41654551e-01 -4.93311286e-01 4.12659556e-01
3.47009122e-01 2.65140265e-01 9.96544540e-01 -5.69221497e-01
7.06668720e-02 1.57755601e+00 8.81864607e-01 3.34949307e-02
-1.74939930e+00 -1.26996577e-01 -6.16165362e-02 -1.31657608e-02
-1.51538324e+00 -3.91612351e-01 1.74339771e-01 -3.92359495e-01
1.08094549e+00 5.87776899e-01 5.40130377e-01 8.11078966e-01
4.97937649e-01 6.79129362e-01 1.15158629e+00 -2.88165122e-01
6.09716833e-01 -3.72050144e-02 -3.86336982e-01 2.10297763e-01
5.54580614e-02 4.10175234e-01 -5.18902302e-01 -6.00867212e-01
3.16806406e-01 -3.18054974e-01 -2.17064455e-01 -4.07250404e-01
-1.15956783e+00 1.11919355e+00 1.52920559e-01 -1.92333266e-01
-5.09852052e-01 5.75471222e-01 2.02773735e-01 2.04497371e-02
6.49618685e-01 8.66115749e-01 -6.53033078e-01 -8.38336170e-01
-1.19216108e+00 5.95717371e-01 1.13805580e+00 5.57704389e-01
4.10137266e-01 -3.43206167e-01 -3.97999048e-01 9.77675974e-01
6.93893015e-01 4.09727186e-01 2.75557399e-01 -1.05848563e+00
3.90906870e-01 6.69684559e-02 2.73211420e-01 -9.22999084e-01
-3.18987042e-01 -5.82006514e-01 -5.70264816e-01 2.67239332e-01
5.84834993e-01 -9.29690003e-02 -8.88290465e-01 1.30168557e+00
4.75811750e-01 -1.43982740e-02 -7.68754482e-02 7.44184911e-01
2.93488264e-01 7.79016614e-01 1.22694299e-01 -6.41307235e-02
1.27320886e+00 -7.76660442e-01 -3.51408631e-01 -4.23004389e-01
4.16352451e-01 -7.94589818e-01 6.02207184e-01 7.86862671e-01
-1.07087040e+00 -1.41745552e-01 -8.54859173e-01 5.25029242e-01
-2.87024200e-01 -3.79451334e-01 5.40279508e-01 1.10931039e+00
-1.10584867e+00 8.96957636e-01 -8.32156122e-01 -3.87908667e-02
5.22404850e-01 6.53293133e-01 1.13537377e-02 -5.18045798e-02
-1.04458082e+00 1.26616037e+00 7.46467590e-01 7.39774778e-02
-8.73529613e-01 -7.73657858e-01 -7.82072544e-01 -1.05892550e-02
5.39116621e-01 -7.62491822e-01 1.30579209e+00 -3.68577600e-01
-1.62860799e+00 3.57557327e-01 5.05601466e-02 -4.54119980e-01
6.05454147e-01 -3.01530510e-01 -6.51640892e-02 -3.20337266e-01
-5.89000657e-02 7.74303257e-01 9.26389515e-01 -1.06795263e+00
-9.24174070e-01 -2.55242616e-01 -1.27362341e-01 4.32557106e-01
1.61377519e-01 1.94175512e-01 -4.17359531e-01 -4.89929676e-01
-9.13527533e-02 -8.27901661e-01 -5.21000206e-01 -3.16042751e-01
-2.22243950e-01 -2.85830885e-01 3.84549290e-01 -5.96192539e-01
1.41827691e+00 -1.82559311e+00 4.54177916e-01 3.99862885e-01
-7.86490664e-02 -4.00468968e-02 2.31894389e-01 2.95426548e-01
1.96548611e-01 2.88477838e-01 -6.49726093e-01 -4.71085876e-01
3.14428031e-01 5.19336700e-01 1.18661083e-01 7.10684896e-01
4.25720274e-01 7.63786316e-01 -1.07248616e+00 -6.56599939e-01
2.22042903e-01 4.56798762e-01 -4.62370574e-01 4.25594524e-02
-1.84277028e-01 3.49923044e-01 -2.51141459e-01 8.04853857e-01
4.49045986e-01 -4.39959876e-02 -2.46580556e-01 3.29570174e-01
2.19163090e-01 3.03774387e-01 -1.65375447e+00 1.33127761e+00
-4.83121604e-01 3.87212217e-01 6.46134019e-02 -1.06938326e+00
9.97727811e-01 5.79365380e-02 4.57900822e-01 -4.14655536e-01
3.95811647e-02 2.97054380e-01 -4.68923636e-02 -3.78294289e-03
4.46038038e-01 -1.81640327e-01 -2.43093312e-01 2.35806614e-01
-2.13012118e-02 -7.31267631e-01 2.55609512e-01 -4.20568109e-01
1.26511896e+00 4.25179899e-01 4.56827521e-01 -4.18594867e-01
3.25809479e-01 -6.92901164e-02 4.21098858e-01 9.34430659e-01
-2.17680801e-02 5.50727427e-01 5.79765439e-01 -2.11932003e-01
-7.73640871e-01 -1.04165947e+00 -5.81305563e-01 1.32010400e+00
-1.12801291e-01 -2.68982410e-01 -5.39975762e-01 -4.42407608e-01
4.18871731e-01 8.96252096e-01 -4.09513026e-01 -2.17004791e-02
-3.30587506e-01 -1.56237745e+00 4.23256308e-01 3.47887009e-01
9.67374966e-02 -7.67256796e-01 -7.65831232e-01 7.02162087e-01
5.81748039e-02 -5.56173563e-01 2.06217263e-02 5.46169460e-01
-8.14707398e-01 -8.04666162e-01 -8.50623429e-01 -8.12828913e-02
1.25233307e-01 -6.19974971e-01 1.47648501e+00 -4.10786271e-01
-2.62566835e-01 2.95332581e-01 -5.27613647e-02 -8.07900310e-01
-3.24332952e-01 1.16172820e-01 -2.00163916e-01 -1.50581092e-01
2.25630835e-01 -2.98520863e-01 -5.52908540e-01 7.27423191e-01
-6.89120352e-01 -4.87839371e-01 5.04599094e-01 1.07674336e+00
6.76460862e-01 4.02459800e-01 2.89093256e-01 -2.35813275e-01
1.10596263e+00 -7.81265378e-01 -9.14300978e-01 2.95559391e-02
-7.17633307e-01 3.88445288e-01 4.87249680e-02 -4.38638330e-01
-6.02077782e-01 -1.20357744e-01 -1.80395409e-01 -1.99967518e-01
-7.05401078e-02 6.89888537e-01 1.12807974e-01 1.20283207e-02
6.23630583e-01 -5.77674881e-02 -3.49706933e-02 -5.02563119e-01
2.54637659e-01 7.19169497e-01 4.34269905e-01 -9.20778155e-01
3.53763402e-01 1.86460316e-01 4.66000140e-01 -5.59423566e-01
-7.47930169e-01 -6.63537085e-01 -2.90560812e-01 -1.94932938e-01
5.96347868e-01 -5.16990244e-01 -8.91195178e-01 2.74459600e-01
-9.10845518e-01 -3.46739203e-01 -3.70657831e-01 4.69090998e-01
-8.30457032e-01 1.68908715e-01 1.94061641e-02 -1.32157147e+00
-1.00497268e-01 -1.52635348e+00 1.50106287e+00 -3.57253477e-02
-3.51953685e-01 -1.20925093e+00 3.08967322e-01 2.43563265e-01
5.06602645e-01 6.01643682e-01 4.69776273e-01 -5.65427423e-01
-3.31399918e-01 -3.78791928e-01 3.43525074e-02 3.82296115e-01
-2.89279044e-01 2.87552504e-03 -8.67541194e-01 -2.96749741e-01
2.29188904e-01 -9.72546041e-02 7.66901851e-01 5.97947896e-01
1.19100416e+00 -1.89331502e-01 -4.32923168e-01 5.29263079e-01
1.40681374e+00 1.58222839e-01 6.96542144e-01 8.60656321e-01
2.64740914e-01 6.67727411e-01 1.00063121e+00 5.75817108e-01
4.01430786e-01 1.05220783e+00 6.46132886e-01 1.05315782e-01
5.27633369e-01 3.55895340e-01 3.44092757e-01 2.92415142e-01
-1.58069760e-01 -4.37438726e-01 -1.41848552e+00 6.49903357e-01
-2.10519290e+00 -7.30927110e-01 -1.43214837e-01 2.45547724e+00
9.65148091e-01 3.51057917e-01 4.13499326e-01 4.13010716e-01
6.78316355e-01 -3.27721750e-03 -2.87541121e-01 -8.05980921e-01
1.60334393e-01 3.55516523e-01 8.90857935e-01 4.01136249e-01
-1.37253273e+00 2.23580137e-01 6.73637915e+00 1.38023329e+00
-5.95401525e-01 2.98883736e-01 9.90739167e-01 -4.15220946e-01
2.94596225e-01 -7.36714676e-02 -9.94323969e-01 9.08836782e-01
1.56602240e+00 8.04810748e-02 5.87794602e-01 9.37845409e-01
2.55444441e-02 -7.31510580e-01 -1.21491385e+00 9.43050921e-01
-1.70766607e-01 -9.59513307e-01 -7.27976918e-01 3.25831771e-01
6.76051795e-01 1.91119120e-01 -4.30887379e-03 2.65327781e-01
6.76525712e-01 -1.49533057e+00 7.73604691e-01 7.66066372e-01
3.60993892e-01 -1.08662713e+00 1.23462141e+00 5.05544901e-01
-9.02017355e-01 -3.23060870e-01 -2.28476256e-01 8.30227509e-02
3.81537110e-01 8.54645491e-01 -8.73647511e-01 6.49452448e-01
1.15353823e+00 2.24697173e-01 -5.70106745e-01 1.68299520e+00
-1.60271227e-01 6.30682230e-01 -1.03451717e+00 -2.97642350e-01
6.03105962e-01 -1.65805042e-01 7.77235329e-01 1.36714387e+00
3.85037065e-01 -4.31849688e-01 5.36299236e-02 7.04147995e-01
5.39377451e-01 -1.00278348e-01 -2.37374082e-01 1.23673268e-01
3.67834091e-01 9.48723316e-01 -5.99865079e-01 -8.00499544e-02
4.41952311e-02 4.02341127e-01 3.80171910e-02 -1.89328827e-02
-7.51387537e-01 -3.63983840e-01 5.93398750e-01 -1.90826714e-01
5.84322870e-01 -8.19046125e-02 -3.80711436e-01 -3.10663760e-01
-1.26092201e-02 -8.27764213e-01 5.77852905e-01 -5.79265594e-01
-1.40123045e+00 2.32786611e-01 5.66011786e-01 -8.44730854e-01
-5.24233341e-01 -6.17891490e-01 -3.01107705e-01 1.21196449e+00
-1.55994165e+00 -5.13561904e-01 -4.92877886e-02 4.97593954e-02
2.26642326e-01 5.84045276e-02 8.29012930e-01 1.04649529e-01
-4.07308161e-01 2.28404149e-01 6.03457749e-01 -5.15807390e-01
5.33099234e-01 -1.61149669e+00 5.36122203e-01 3.70599717e-01
-2.06564561e-01 1.70578510e-01 8.96401763e-01 -4.42687482e-01
-1.17889440e+00 -8.10479581e-01 5.40323436e-01 -1.03021824e+00
6.62651837e-01 -1.73603922e-01 -6.41889930e-01 4.77827266e-02
-1.52604893e-01 -1.44658908e-01 5.67598045e-01 1.41991049e-01
4.03681517e-01 -4.76913620e-03 -1.38927865e+00 3.87148768e-01
5.72226286e-01 -6.26265779e-02 -6.79709673e-01 4.12176460e-01
2.37353727e-01 -5.37572503e-01 -1.52771330e+00 5.31203330e-01
4.75682467e-01 -8.09499860e-01 1.26310885e+00 -2.28687480e-01
1.90697208e-01 -2.70910949e-01 -2.79330969e-01 -1.60082448e+00
-1.36476919e-01 -9.66263115e-01 -4.28935379e-01 1.10715973e+00
4.57415968e-01 -9.10369217e-01 5.42967498e-01 6.99104428e-01
1.79240927e-01 -1.38746321e+00 -1.47126043e+00 -1.06945395e+00
-5.13569638e-03 -8.00019383e-01 9.08090115e-01 3.68468404e-01
-4.84693021e-01 3.29964459e-02 -2.54872870e-02 -3.63576226e-02
6.95994139e-01 -4.91973281e-01 8.58237684e-01 -1.39836478e+00
-6.37440324e-01 -7.86245108e-01 -7.83481956e-01 -8.37131262e-01
-3.56251448e-01 -5.14622390e-01 3.97454679e-01 -1.49006271e+00
-2.76779413e-01 -7.23805249e-01 -2.91117579e-01 2.03406811e-01
-3.77935052e-01 7.58310854e-02 -2.37199757e-02 1.79847479e-01
-6.29323483e-01 5.64324975e-01 8.39620113e-01 -3.19752516e-03
-1.98245227e-01 3.49221468e-01 -5.26589453e-01 6.04252756e-01
7.64248192e-01 -9.36868906e-01 -6.05135001e-02 -5.80683276e-02
4.28890854e-01 -6.35284632e-02 3.15348744e-01 -9.52188492e-01
1.04154602e-01 -2.41786957e-01 3.65805894e-01 -5.83179355e-01
6.87001944e-01 -6.82124317e-01 2.55353838e-01 2.25953951e-01
-1.07514299e-01 3.28120291e-01 3.23376268e-01 7.77158141e-01
-7.46835619e-02 -6.49661720e-01 7.85594642e-01 -1.29225627e-01
-3.03054869e-01 -1.02299012e-01 -1.26516595e-01 1.56965017e-01
1.16770470e+00 -3.86674106e-01 -3.74192484e-02 -2.55724430e-01
-6.04848742e-01 6.25975251e-01 2.76537478e-01 2.08817586e-01
3.13001513e-01 -9.98490393e-01 -8.68557811e-01 -2.14761421e-01
-7.13474862e-03 3.99712235e-01 -4.28260148e-01 1.16750395e+00
-6.07486904e-01 5.71821272e-01 3.06870192e-01 -1.12789762e+00
-1.00912857e+00 2.46815637e-01 5.01937687e-01 -6.83696747e-01
-1.33626714e-01 1.13338280e+00 -2.98807919e-01 -5.96112847e-01
3.80640149e-01 -1.78676188e-01 -4.90601659e-02 1.41069055e-01
2.66798705e-01 8.19341660e-01 2.25151211e-01 -5.14785230e-01
-4.13842738e-01 3.30944866e-01 4.93367046e-01 -5.73290408e-01
1.54693592e+00 4.47032601e-02 -1.20598160e-01 4.63689536e-01
8.90992463e-01 -6.03971004e-01 -1.29739571e+00 1.56388488e-02
4.21708673e-01 -6.71618283e-01 5.95116556e-01 -8.68219912e-01
-5.32883406e-01 6.14420414e-01 5.26451528e-01 3.50748867e-01
9.47937727e-01 -6.47418052e-02 2.14522898e-01 2.91812778e-01
3.72699320e-01 -1.23818505e+00 -2.60577828e-01 3.25464368e-01
9.94287550e-01 -1.22069824e+00 3.62683266e-01 -1.17217064e-01
-6.50775850e-01 8.26427817e-01 6.23919815e-02 8.83669257e-02
7.22469866e-01 2.11227849e-01 -4.89735186e-01 -2.56876558e-01
-7.88281322e-01 -1.99104980e-01 5.29833674e-01 7.28764415e-01
-1.01556107e-01 1.28684998e-01 -1.00308798e-01 2.43852854e-01
-4.71470535e-01 -1.88895941e-01 -8.95113572e-02 1.18136346e+00
-1.73727229e-01 -1.10041225e+00 -1.05022490e+00 8.12505007e-01
-6.86935723e-01 -1.05687432e-01 1.16541814e-02 7.03559220e-01
2.53705904e-02 1.12929416e+00 2.06913546e-01 3.43320481e-02
2.34681234e-01 1.29421949e-01 3.86315733e-01 -5.55190980e-01
-7.75259256e-01 5.33665381e-02 5.05520046e-01 -5.92406273e-01
-4.03657585e-01 -9.32471156e-01 -4.88965482e-01 -2.75703102e-01
-7.06085920e-01 1.39378384e-01 1.28764963e+00 9.44918334e-01
1.89827114e-01 5.49764514e-01 3.47937375e-01 -1.31517637e+00
-9.44307983e-01 -9.29512262e-01 -2.84490168e-01 -9.32929292e-02
1.58501089e-01 -1.15639377e+00 -6.11280739e-01 -4.67360646e-01] | [6.305703163146973, 3.8288803100585938] |
871560f3-31a7-4a8a-927e-f0a77d111666 | evaluating-temporal-observation-based-causal | 2302.00064 | null | https://arxiv.org/abs/2302.00064v2 | https://arxiv.org/pdf/2302.00064v2.pdf | Evaluating Temporal Observation-Based Causal Discovery Techniques Applied to Road Driver Behaviour | Autonomous robots are required to reason about the behaviour of dynamic agents in their environment. The creation of models to describe these relationships is typically accomplished through the application of causal discovery techniques. However, as it stands observational causal discovery techniques struggle to adequately cope with conditions such as causal sparsity and non-stationarity typically seen during online usage in autonomous agent domains. Meanwhile, interventional techniques are not always feasible due to domain restrictions. In order to better explore the issues facing observational techniques and promote further discussion of these topics we carry out a benchmark across 10 contemporary observational temporal causal discovery methods in the domain of autonomous driving. By evaluating these methods upon causal scenes drawn from real world datasets in addition to those generated synthetically we highlight where improvements need to be made in order to facilitate the application of causal discovery techniques to the aforementioned use-cases. Finally, we discuss potential directions for future work that could help better tackle the difficulties currently experienced by state of the art techniques. | ['Lars Kunze', 'Rhys Howard'] | 2023-01-31 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 4.63519782e-01 3.24635029e-01 -3.43826145e-01 -4.69559669e-01
-8.36294293e-02 -2.45719954e-01 1.38300657e+00 3.28639716e-01
-1.84092373e-01 1.09285867e+00 5.33076942e-01 -5.73394537e-01
-7.64129639e-01 -7.78746426e-01 -7.63479233e-01 -5.50304651e-01
-7.79320061e-01 6.00782275e-01 2.71205187e-01 -1.36694610e-01
3.59731883e-01 4.18125987e-01 -1.96273053e+00 -1.89287484e-01
7.41848290e-01 3.20040047e-01 2.41544783e-01 4.42284644e-01
3.78141105e-01 1.08851886e+00 -3.18393856e-01 1.39520898e-01
2.42838368e-01 -4.18900102e-01 -6.80215299e-01 6.49380535e-02
3.48482341e-01 -3.44650388e-01 -4.24344569e-01 9.05747652e-01
4.86844815e-02 3.13358396e-01 7.14010119e-01 -1.58232391e+00
-1.94535568e-01 5.73047698e-01 -3.45290184e-01 5.47847927e-01
5.24774790e-01 2.94130802e-01 9.27924693e-01 -3.91655654e-01
7.85434425e-01 1.54934168e+00 3.23877394e-01 4.95407768e-02
-1.31985462e+00 -8.26252997e-01 5.32167554e-01 7.00041413e-01
-8.33179593e-01 -6.37429655e-01 6.19894743e-01 -4.68111306e-01
8.76769364e-01 -1.78885236e-02 4.25789446e-01 1.50950897e+00
2.09221691e-01 4.59635079e-01 1.27536345e+00 -1.83597907e-01
3.96202654e-01 -3.07886094e-01 6.62666745e-03 5.49208939e-01
4.51770574e-01 8.98123801e-01 -9.31094766e-01 -4.43638772e-01
6.75190628e-01 -5.43924510e-01 1.54105164e-02 -7.17521727e-01
-1.31769407e+00 1.13543463e+00 2.54746646e-01 5.31848893e-02
-6.31557107e-01 4.42114264e-01 4.61069942e-01 2.87757903e-01
2.77231723e-01 5.09399235e-01 -5.45179129e-01 -2.54759938e-01
-5.39995074e-01 8.35374951e-01 7.91124701e-01 9.60648298e-01
4.19480532e-01 1.08944606e-02 3.65151823e-01 4.25741225e-01
5.06299734e-01 3.36753070e-01 -6.38234690e-02 -1.22903800e+00
2.56112069e-01 4.47091758e-01 3.69533181e-01 -9.02739704e-01
-4.52051461e-01 -6.22599944e-02 -3.84103000e-01 4.23108399e-01
5.71159840e-01 -1.53070152e-01 -6.88368618e-01 1.66020894e+00
5.52296221e-01 5.34267724e-01 2.44588807e-01 9.06055152e-01
4.73531395e-01 2.49372870e-01 2.93246955e-01 -5.63166261e-01
1.23411262e+00 -3.59293252e-01 -7.99288273e-01 -4.25082624e-01
3.74994129e-01 -6.36823297e-01 6.72065556e-01 2.73296207e-01
-6.26714468e-01 -2.15562314e-01 -1.06300259e+00 2.86546558e-01
-1.60769612e-01 -3.88414711e-01 9.28356528e-01 3.36899638e-01
-5.24963558e-01 3.18339109e-01 -1.05615747e+00 -7.58765280e-01
2.36338690e-01 2.76582390e-01 -2.70296693e-01 3.32140774e-02
-1.28179157e+00 1.06392491e+00 3.65077138e-01 7.08042309e-02
-1.45433927e+00 -8.99982572e-01 -8.19255471e-01 -3.57620955e-01
8.29070747e-01 -7.19968796e-01 1.51874840e+00 -3.66711020e-01
-1.14907014e+00 4.24523234e-01 -2.76811153e-01 -8.54377031e-01
6.21789694e-01 -2.50186473e-01 -4.20139521e-01 -2.84727722e-01
4.77989167e-01 3.84877086e-01 3.78229111e-01 -1.41647029e+00
-1.05524504e+00 -2.26614341e-01 3.35609108e-01 7.29233697e-02
1.98833823e-01 1.51913598e-01 1.51279375e-01 -3.88421386e-01
-9.99262333e-02 -1.25465596e+00 -5.82884669e-01 -4.38581914e-01
-2.56927937e-01 -3.37976903e-01 1.10412455e+00 -2.20461451e-02
7.26640880e-01 -2.00041890e+00 1.68617472e-01 9.92886797e-02
1.04611859e-01 -2.46797040e-01 5.66036031e-02 7.77795911e-01
-1.53709278e-01 -1.50664479e-01 -2.58803993e-01 -4.57157530e-02
-2.72550415e-02 6.84032798e-01 -5.22170424e-01 7.29688525e-01
2.60517716e-01 3.49461317e-01 -1.35109544e+00 -4.76821661e-01
5.21695256e-01 -4.94810119e-02 -4.46131140e-01 1.11764237e-01
-7.56850719e-01 8.27051640e-01 -6.93014443e-01 2.54224598e-01
1.22194514e-01 1.20298184e-01 5.29672503e-01 4.89482343e-01
-5.61274290e-01 7.63832033e-01 -1.29622769e+00 1.43436372e+00
-3.34707648e-01 7.51741409e-01 -5.05746678e-02 -1.14307690e+00
7.14387298e-01 4.69048023e-01 5.94448686e-01 -8.10013771e-01
-1.23418635e-02 9.13622379e-02 3.84452909e-01 -8.73527944e-01
5.06741345e-01 -1.58789411e-01 -1.05092518e-01 3.55618536e-01
-3.99502963e-01 -1.07712345e-02 5.06120443e-01 1.13102682e-01
1.46255708e+00 1.75135449e-01 4.50270891e-01 -4.86225337e-01
2.27066427e-01 7.35325873e-01 5.66828847e-01 8.43077302e-01
-2.16488108e-01 6.17013313e-02 6.87886536e-01 -6.51128531e-01
-1.04204869e+00 -8.22535694e-01 -4.46161538e-01 7.13136077e-01
2.53522664e-01 -2.92439461e-01 -1.48173003e-02 -5.44809699e-01
5.15301898e-02 1.27411735e+00 -9.42472577e-01 -5.86665198e-02
-6.77298784e-01 -1.13998568e+00 4.96140212e-01 3.37348938e-01
3.23824733e-01 -9.39830363e-01 -1.25176406e+00 3.71605366e-01
-1.91408679e-01 -1.28601062e+00 6.23067975e-01 4.14846241e-01
-7.50653982e-01 -1.41537666e+00 1.36480138e-01 -1.76179841e-01
3.11390728e-01 3.19823414e-01 1.01274014e+00 -2.04089060e-01
-2.13377461e-01 2.73400396e-01 -4.91050452e-01 -7.05083728e-01
-7.13636696e-01 -3.00886929e-02 3.02388370e-01 -3.99671048e-01
4.37528402e-01 -8.94554079e-01 -3.81839871e-01 4.69610542e-01
-6.33934200e-01 -1.58277914e-01 3.44290406e-01 7.82315850e-01
1.37156129e-01 5.23364246e-01 5.19373655e-01 -7.55587041e-01
6.91278934e-01 -1.08798409e+00 -8.69806707e-01 -2.01373458e-01
-7.23607898e-01 2.13240936e-01 3.07569414e-01 -4.13231015e-01
-1.37717056e+00 1.07565068e-01 4.98506248e-01 -1.09215409e-01
-4.97850209e-01 8.47994208e-01 2.82534033e-01 5.53401649e-01
9.43905532e-01 -4.25913841e-01 1.31166738e-03 -1.22493073e-01
5.08860707e-01 1.90598339e-01 5.21727741e-01 -6.77191377e-01
7.64955759e-01 7.27437973e-01 3.15742940e-01 -8.04386437e-01
-5.59840918e-01 -3.42125952e-01 -4.43888694e-01 -3.07629734e-01
6.24683082e-01 -9.04863119e-01 -6.13501847e-01 -7.34979436e-02
-1.00630641e+00 -7.42422163e-01 -1.69031903e-01 6.13937378e-01
-8.89930665e-01 1.43948756e-02 -1.00799099e-01 -1.06839848e+00
6.89692199e-01 -1.10973644e+00 9.04829502e-01 -3.42392661e-02
-7.88666189e-01 -1.00179541e+00 3.18210721e-01 3.29345852e-01
3.64878252e-02 3.15106720e-01 1.03063107e+00 -2.89399832e-01
-5.98942339e-01 2.21053392e-01 -1.10985383e-01 -5.80628574e-01
2.70112395e-01 2.40147516e-01 -7.41413176e-01 8.95220414e-02
-1.10943042e-01 -1.73186794e-01 6.08473659e-01 5.62918484e-01
5.32794058e-01 -8.40853825e-02 -5.29078364e-01 -2.27168858e-01
1.03422499e+00 2.94913352e-01 4.73838896e-01 7.16128111e-01
3.66656482e-01 1.20295620e+00 1.25028610e+00 6.32734418e-01
7.18000710e-01 8.00944269e-01 9.72577870e-01 1.59986779e-01
-9.65633616e-02 -3.14570636e-01 1.80126920e-01 -1.09353207e-01
-2.27200925e-01 -2.47762039e-01 -1.12386036e+00 9.48665917e-01
-2.32758212e+00 -1.18092573e+00 -1.01559627e+00 2.00416136e+00
4.96290118e-01 1.63721532e-01 3.06249708e-01 1.75464034e-01
3.48961353e-01 8.94770175e-02 -6.33155942e-01 -2.59473771e-01
4.87441197e-02 -3.27100307e-01 8.43688011e-01 4.58584040e-01
-1.11866212e+00 8.40805769e-01 6.36611700e+00 1.60953805e-01
-6.88907027e-01 6.38148561e-02 6.66485429e-02 -1.94714844e-01
-4.07421254e-02 5.00159442e-01 -4.65487599e-01 1.69085935e-01
1.19570541e+00 -1.77579254e-01 3.48037928e-01 5.73135197e-01
9.80739772e-01 -6.64732575e-01 -1.26703751e+00 3.56780678e-01
-5.72102487e-01 -1.19566560e+00 -5.53691208e-01 3.40148896e-01
8.68176937e-01 3.96738470e-01 -5.97379766e-02 -4.07483540e-02
1.19977033e+00 -1.09023046e+00 8.27949286e-01 2.33910978e-01
2.40017906e-01 -3.42601091e-01 4.66157287e-01 3.10548723e-01
-7.78862059e-01 -3.72405469e-01 -1.59119979e-01 -8.37815464e-01
3.98401856e-01 6.34978652e-01 -1.31760633e+00 7.42039263e-01
9.69955623e-01 8.63009155e-01 -1.59685031e-01 1.01313627e+00
-5.08838236e-01 7.98981369e-01 -5.01742005e-01 -8.69662762e-02
3.26455921e-01 2.00059656e-02 7.86657333e-01 8.20827782e-01
1.14264019e-01 5.97137250e-02 2.29565352e-02 7.53915012e-01
6.18329823e-01 -3.59140575e-01 -1.08385062e+00 -4.44793701e-02
5.36749601e-01 6.48499072e-01 -6.61365926e-01 -2.32763767e-01
-4.61115301e-01 2.45251432e-01 1.60521761e-01 2.54202545e-01
-9.17132974e-01 6.23846233e-01 9.78903532e-01 1.44932017e-01
4.58825892e-03 -7.18068957e-01 -4.87916231e-01 -7.10743845e-01
-1.67695433e-01 -1.09722400e+00 3.48328590e-01 -6.15372002e-01
-1.11765182e+00 4.82401811e-02 8.27580273e-01 -1.04422605e+00
-7.18017757e-01 -2.35409141e-01 -3.49215806e-01 4.13206995e-01
-1.41466904e+00 -8.01162958e-01 -1.26085684e-01 4.05917525e-01
8.17886353e-01 9.41426903e-02 8.44471633e-01 1.82490483e-01
-5.13334632e-01 -4.72139984e-01 -2.62453020e-01 -5.78069091e-01
7.58091033e-01 -1.15534127e+00 5.56805670e-01 1.08266914e+00
7.54945651e-02 5.64272344e-01 1.60647166e+00 -1.00460076e+00
-1.27834964e+00 -1.06568134e+00 8.22611928e-01 -4.86371040e-01
1.34302557e+00 -1.30821347e-01 -5.80014110e-01 7.97273338e-01
2.15073630e-01 -2.56560951e-01 3.65953177e-01 7.59251535e-01
-1.64900348e-01 4.05508816e-01 -5.85623801e-01 8.05524230e-01
1.35847139e+00 -1.55338585e-01 -7.92026222e-01 4.27597284e-01
3.83301824e-01 -1.80444777e-01 -7.01209605e-01 6.68882608e-01
3.82793099e-01 -9.27012146e-01 6.69223130e-01 -6.26744270e-01
6.44758821e-01 -6.27335250e-01 -1.55059874e-01 -1.34338641e+00
-1.28119275e-01 -5.94537139e-01 1.54541314e-01 9.77448702e-01
3.91484559e-01 -7.14034557e-01 6.77983403e-01 4.95999664e-01
-1.62234213e-02 -1.69899389e-01 -1.01786661e+00 -7.71909595e-01
-8.23410749e-02 -9.13963020e-01 4.92127359e-01 1.13605988e+00
3.01429406e-02 4.28483337e-01 -3.84192228e-01 6.29257441e-01
8.34107757e-01 3.69992927e-02 1.04785264e+00 -1.47444940e+00
5.33916168e-02 -2.44591057e-01 -4.43803579e-01 -6.76413417e-01
4.03395146e-01 -2.31278747e-01 2.37318918e-01 -1.18722856e+00
3.15234773e-02 -1.06555521e+00 3.04964989e-01 2.94679910e-01
-1.76100321e-02 1.54306423e-02 -1.59375697e-01 2.62910396e-01
-1.90481633e-01 6.59962773e-01 9.01420295e-01 1.75576042e-02
-2.44598508e-01 -9.36279371e-02 -4.39916998e-01 8.09451282e-01
1.04037702e+00 -7.52631426e-01 -8.70136917e-01 -2.36334383e-01
4.45538491e-01 3.79158825e-01 8.31936955e-01 -6.57367051e-01
3.83526862e-01 -8.16454113e-01 -3.53526950e-01 -4.12219435e-01
2.65251756e-01 -9.49193358e-01 4.99512523e-01 4.21242654e-01
-4.20310080e-01 1.79345623e-01 1.40581802e-01 9.46926951e-01
-1.90056935e-01 -1.88350439e-01 1.82158440e-01 -9.36483815e-02
-1.08277786e+00 -1.08643681e-01 -8.10450077e-01 -2.92866290e-01
1.28751194e+00 1.05462819e-01 -3.33814979e-01 -3.76422048e-01
-5.28055847e-01 7.09129393e-01 4.25660551e-01 7.33799338e-01
1.63195997e-01 -8.86182129e-01 -9.05732334e-01 -1.91026703e-01
3.43193740e-01 2.59451587e-02 1.06027380e-01 1.06628430e+00
-1.50251478e-01 6.18126810e-01 -1.08201660e-01 -4.68129486e-01
-1.12941980e+00 7.21314251e-01 1.63024724e-01 -1.27061605e-01
-8.29885066e-01 3.83501321e-01 3.39240313e-01 -2.78560549e-01
1.13391437e-01 -1.39372006e-01 -3.17110896e-01 -2.02668179e-02
1.49011627e-01 5.47723055e-01 2.70050205e-02 -4.76967663e-01
-4.38901246e-01 -2.42473900e-01 2.53351748e-01 -4.45038915e-01
1.34263694e+00 -3.16337913e-01 3.82288694e-02 6.84171677e-01
3.77524436e-01 -1.28033817e-01 -1.55569422e+00 1.41443580e-01
3.70535403e-01 -4.09284174e-01 5.06778806e-02 -8.49207819e-01
-6.22355282e-01 3.96710753e-01 4.27117556e-01 5.97374618e-01
7.29525328e-01 1.28282174e-01 1.03994928e-01 1.05607025e-01
6.66936576e-01 -8.41597795e-01 -3.89942378e-01 3.72958690e-01
9.72299457e-01 -1.34607625e+00 2.36763328e-01 -6.42779231e-01
-3.16662759e-01 8.67350280e-01 2.51251906e-01 -3.37170571e-01
4.25382584e-01 3.39580268e-01 -6.49415106e-02 -7.09367454e-01
-1.29315007e+00 -4.58014280e-01 -2.47110546e-01 7.04806209e-01
3.02940369e-01 3.23744714e-01 -5.09500444e-01 -4.49590772e-01
-4.75145429e-01 -5.02184853e-02 8.18316877e-01 8.28217208e-01
9.45210233e-02 -1.29422224e+00 -5.66875517e-01 4.09146428e-01
-2.03983068e-01 1.99040189e-01 -3.89089406e-01 1.16852617e+00
1.90138504e-01 1.64784765e+00 3.99668142e-02 -1.53115362e-01
4.12122399e-01 -3.08127970e-01 4.05362338e-01 -5.78578949e-01
-5.09504676e-02 -2.55847901e-01 8.83903265e-01 -6.49001241e-01
-9.20835018e-01 -1.44472301e+00 -1.19585299e+00 -3.62865835e-01
-8.01516697e-02 -9.40633491e-02 6.84108436e-01 1.19343746e+00
-1.29822055e-02 8.92664194e-01 2.62426734e-02 -7.51324952e-01
-2.42045969e-01 -8.22854996e-01 -1.73280671e-01 1.61026210e-01
2.92500228e-01 -1.51734972e+00 -4.27284002e-01 2.84071565e-01] | [7.8574347496032715, 5.328951358795166] |
9dd6bad4-b6e9-45bf-888b-101874672619 | crossing-the-gap-domain-generalization-for | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ren_Crossing_the_Gap_Domain_Generalization_for_Image_Captioning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ren_Crossing_the_Gap_Domain_Generalization_for_Image_Captioning_CVPR_2023_paper.pdf | Crossing the Gap: Domain Generalization for Image Captioning | Existing image captioning methods are under the assumption that the training and testing data are from the same domain or that the data from the target domain (i.e., the domain that testing data lie in) are accessible. However, this assumption is invalid in real-world applications where the data from the target domain is inaccessible. In this paper, we introduce a new setting called Domain Generalization for Image Captioning (DGIC), where the data from the target domain is unseen in the learning process. We first construct a benchmark dataset for DGIC, which helps us to investigate models' domain generalization (DG) ability on unseen domains. With the support of the new benchmark, we further propose a new framework called language-guided semantic metric learning (LSML) for the DGIC setting. Experiments on multiple datasets demonstrate the challenge of the task and the effectiveness of our newly proposed benchmark and LSML framework. | ['Wanli Ouyang', 'Yongdong Zhang', 'Hao Du', 'Tong He', 'Yan Lu', 'Shancheng Fang', 'Zhendong Mao', 'Yuchen Ren'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['metric-learning', 'metric-learning'] | ['computer-vision', 'methodology'] | [ 5.04838407e-01 1.41834691e-01 -4.38749552e-01 -4.69340146e-01
-7.49300599e-01 -7.02356458e-01 6.50990725e-01 -2.59263456e-01
-2.64717966e-01 9.73448098e-01 1.10435456e-01 -2.50732422e-01
2.09336400e-01 -6.28279626e-01 -1.12006664e+00 -6.50613844e-01
4.58640844e-01 6.09896302e-01 1.21853583e-01 -1.67261772e-02
-4.21679467e-02 6.37171566e-02 -1.33451962e+00 4.16464001e-01
9.34653938e-01 1.12563205e+00 4.10774797e-01 1.80188641e-01
-2.85613537e-03 5.44585466e-01 -5.46690464e-01 -2.99491554e-01
3.75838161e-01 -6.59430623e-01 -9.97158825e-01 3.63123953e-01
5.41926444e-01 -4.27371800e-01 -3.64277750e-01 1.22979498e+00
2.08598197e-01 4.96493950e-02 6.24203622e-01 -1.88816142e+00
-1.16732419e+00 1.40056968e-01 -2.13292301e-01 -7.72243142e-02
4.69093561e-01 1.61981340e-02 7.79929280e-01 -7.42841363e-01
9.47439492e-01 1.21597052e+00 1.58309802e-01 1.10829306e+00
-1.02915609e+00 -6.84326172e-01 4.03711319e-01 2.73620635e-01
-1.22478330e+00 -3.14837128e-01 8.10718954e-01 -4.48321670e-01
1.60596043e-01 -1.16217487e-01 -1.35939181e-01 1.65825260e+00
-4.01303321e-01 8.96834135e-01 1.16826773e+00 -3.85613412e-01
5.90655088e-01 4.92023319e-01 3.53392921e-02 1.52550980e-01
4.40056235e-01 1.22107804e-01 -3.72981101e-01 -3.99247855e-02
6.40296102e-01 -2.62647092e-01 -5.28292179e-01 -9.29140747e-01
-1.39962530e+00 8.84756625e-01 3.05145353e-01 7.93378204e-02
-1.35718836e-02 -1.34642303e-01 3.81301880e-01 4.01210874e-01
4.82739508e-01 3.43633056e-01 -4.90206569e-01 1.31779641e-01
-5.62915504e-01 1.27058700e-01 8.67249668e-01 1.39965820e+00
7.01834977e-01 -3.94608021e-01 2.67164465e-02 7.50331521e-01
2.26420581e-01 6.91716194e-01 6.77900970e-01 -7.49523699e-01
8.61843109e-01 5.29864252e-01 5.37476957e-01 -6.14492774e-01
1.59901202e-01 -1.22941978e-01 -6.02235794e-01 -1.36065498e-01
6.42580152e-01 -6.42944574e-02 -9.66571212e-01 2.29028344e+00
3.47991437e-01 4.96411562e-01 6.39432728e-01 1.04476452e+00
8.93281698e-01 7.71877766e-01 1.93189815e-01 -4.10582349e-02
9.54071283e-01 -9.10693109e-01 -5.02375782e-01 -5.77627540e-01
8.14523578e-01 -2.58764803e-01 1.34552860e+00 1.81710362e-01
-5.71340561e-01 -6.19361222e-01 -1.19577694e+00 -1.58980321e-02
-3.22977573e-01 -3.91496979e-02 2.41060987e-01 3.45852762e-01
-9.03362095e-01 2.28207354e-02 -4.31062698e-01 -5.97890258e-01
5.20560086e-01 5.99101484e-02 -6.02514684e-01 -6.45356655e-01
-1.31066549e+00 7.85464704e-01 7.44711459e-01 -1.73627079e-01
-1.26492608e+00 -5.91588438e-01 -1.12770343e+00 -2.31982768e-01
4.34833229e-01 -4.98978585e-01 1.20409977e+00 -1.61453283e+00
-8.46817315e-01 1.25135720e+00 -1.51146561e-01 -4.10785228e-01
7.08889008e-01 -1.66833207e-01 -5.80514193e-01 2.59859085e-01
3.51717889e-01 8.97684693e-01 7.60577798e-01 -1.64502764e+00
-4.97380167e-01 -4.16575313e-01 3.01547110e-01 2.12584332e-01
-3.69343370e-01 -3.59175831e-01 -3.05342555e-01 -4.35312390e-01
-4.44666371e-02 -9.59707201e-01 1.39620140e-01 2.09171638e-01
-3.44820619e-01 -2.54395664e-01 1.18866897e+00 -5.99251449e-01
8.42376769e-01 -2.51061368e+00 3.64584848e-02 -5.57583757e-02
-1.18938848e-01 4.73600775e-01 -5.28872252e-01 3.16780478e-01
-3.80241722e-01 1.20405905e-01 -4.56452936e-01 -9.90247577e-02
7.44808614e-02 5.95769644e-01 -6.52748227e-01 2.78980792e-01
2.53813118e-01 9.41468537e-01 -1.09014130e+00 -5.29371321e-01
6.63431641e-03 1.37676254e-01 -3.76571655e-01 4.98212397e-01
-5.61403930e-01 8.05565417e-01 -7.30386317e-01 4.44525659e-01
8.74564767e-01 -5.39465010e-01 3.84950936e-02 -2.68037003e-02
4.02912647e-01 6.51937276e-02 -1.07848561e+00 1.88671684e+00
-4.48580772e-01 5.10548294e-01 -2.95257121e-01 -1.19304514e+00
8.65851760e-01 3.26786965e-01 1.95150122e-01 -8.72586727e-01
-3.29964101e-01 2.56907314e-01 -2.77386129e-01 -6.82548821e-01
8.05724710e-02 -2.41092697e-01 -1.13984972e-01 4.69205081e-01
8.12399108e-03 1.23097107e-01 7.13592768e-02 1.46926910e-01
7.68527985e-01 2.69717395e-01 1.92090005e-01 -1.50199145e-01
5.53635061e-01 1.77011222e-01 6.06979847e-01 6.15712345e-01
-5.14353156e-01 7.58059144e-01 5.79401791e-01 -2.88823277e-01
-1.19924772e+00 -1.28520060e+00 -2.57042527e-01 9.01449502e-01
6.09043241e-01 7.69545659e-02 -9.40678060e-01 -1.17969835e+00
-1.26695201e-01 8.53817523e-01 -6.80418730e-01 -3.49724859e-01
-3.03823739e-01 -3.01400512e-01 1.82894200e-01 5.26633680e-01
8.29804718e-01 -9.98694062e-01 -2.85378307e-01 -1.46496475e-01
-5.21078944e-01 -1.67797840e+00 -4.65738356e-01 -3.42531741e-01
-7.29994535e-01 -1.13100171e+00 -7.54377425e-01 -1.16205013e+00
8.24632883e-01 3.06815237e-01 1.27522480e+00 -1.70434728e-01
2.86987484e-01 6.81665540e-01 -5.48598349e-01 -3.54178190e-01
-6.61509812e-01 -1.49041519e-01 -3.39186266e-02 1.68934450e-01
6.81619883e-01 -2.58988857e-01 -5.71314573e-01 5.75095773e-01
-1.20936227e+00 2.24949285e-01 3.86267126e-01 7.55228400e-01
6.14962816e-01 -9.70392749e-02 9.98843789e-01 -8.60661983e-01
3.44555229e-01 -8.41901779e-01 -5.52110910e-01 5.44290900e-01
-3.62693518e-01 1.67155206e-01 6.59351349e-01 -6.85550570e-01
-1.03301430e+00 1.53301470e-02 2.84142703e-01 -4.39012945e-01
-4.01265651e-01 4.34833556e-01 -8.41034591e-01 1.48239195e-01
5.28376877e-01 3.93479437e-01 -2.24539898e-02 -3.90083671e-01
2.19534755e-01 8.88591528e-01 6.73342407e-01 -8.05319190e-01
7.80562162e-01 6.39624596e-01 -2.48772770e-01 -5.75149953e-01
-1.12747014e+00 -3.02000582e-01 -5.40841043e-01 2.43979157e-03
8.12452793e-01 -1.06144512e+00 -1.15076736e-01 3.36671919e-01
-1.30753314e+00 -1.80897772e-01 -3.23807031e-01 3.89972717e-01
-8.11728835e-01 4.69478458e-01 8.73090699e-02 -5.21629155e-01
1.42366394e-01 -1.27096796e+00 1.14930558e+00 1.09583087e-01
4.67010625e-02 -1.30349493e+00 -9.79342405e-03 4.99362171e-01
3.92693989e-02 4.60661650e-01 1.09345913e+00 -1.20075464e+00
-5.26283562e-01 -3.93863581e-02 -4.73188877e-01 1.00226116e+00
1.35061979e-01 -6.93472207e-01 -9.95398879e-01 -4.19911563e-01
2.18427569e-01 -7.36191988e-01 5.71433127e-01 3.07735428e-02
1.23660922e+00 -2.18753591e-01 -2.83583730e-01 4.52734649e-01
1.63841140e+00 1.79072142e-01 6.33404970e-01 4.20152128e-01
4.72311288e-01 7.69726992e-01 8.62093329e-01 2.96720982e-01
4.25207317e-01 6.02702618e-01 6.30422831e-01 -2.33414203e-01
-1.52498588e-01 -6.52162492e-01 4.37160581e-01 2.30930284e-01
5.48791707e-01 -5.84236443e-01 -9.54226375e-01 8.84305954e-01
-1.92207825e+00 -6.26575589e-01 6.47692457e-02 2.35982275e+00
8.87872159e-01 -1.06751643e-01 -9.34402868e-02 -2.40188956e-01
9.98275220e-01 -9.04613286e-02 -9.66288388e-01 -2.25491866e-01
-1.72751427e-01 -8.89128447e-02 2.59176791e-01 2.23229691e-01
-1.30174994e+00 7.38380551e-01 5.64191198e+00 5.95050871e-01
-9.09054518e-01 1.44821033e-01 7.47991860e-01 2.80690223e-01
-3.48962903e-01 -5.27912304e-02 -7.76281655e-01 7.32856095e-01
8.94904792e-01 -3.69817853e-01 2.33046204e-01 1.09665942e+00
-7.78179476e-03 5.45661673e-02 -1.74797630e+00 1.12028348e+00
2.96712011e-01 -7.91165829e-01 3.99881810e-01 5.30369654e-02
7.49703586e-01 -5.78266161e-04 3.12191039e-01 2.82311827e-01
1.97202355e-01 -9.14181828e-01 5.57213068e-01 1.10798590e-01
1.13940203e+00 -4.18822438e-01 7.84200311e-01 4.06415373e-01
-5.74803829e-01 -1.90238059e-02 -4.71438676e-01 1.67440459e-01
-6.90134540e-02 3.34891409e-01 -9.09175038e-01 5.64269602e-01
4.42952424e-01 9.08358097e-01 -5.77517211e-01 9.38362539e-01
-2.90705770e-01 5.91501951e-01 1.20498873e-02 3.92380327e-01
4.10206020e-01 -1.69068575e-01 4.20557439e-01 8.69007707e-01
3.62904817e-01 -1.61069304e-01 1.21315323e-01 9.29783225e-01
-2.46217892e-01 -1.06591731e-01 -9.69790936e-01 -1.26332760e-01
4.31488484e-01 7.04508662e-01 -2.06238106e-01 -2.92791724e-01
-7.41534352e-01 1.09437954e+00 1.53311163e-01 8.29594135e-01
-9.49504495e-01 -1.04797803e-01 8.04838359e-01 -3.53505500e-02
1.86216787e-01 -1.45255588e-04 5.49087673e-02 -1.37827957e+00
4.33019280e-01 -9.49657321e-01 5.36991715e-01 -9.36444640e-01
-1.59116805e+00 3.77462536e-01 1.72825113e-01 -1.60450912e+00
-3.37056339e-01 -8.94020140e-01 -2.21960694e-01 8.81464601e-01
-1.90685344e+00 -1.20688951e+00 -4.59814548e-01 8.86519492e-01
4.42692220e-01 -2.41129212e-02 6.82937682e-01 1.52946815e-01
-2.25355238e-01 4.52563792e-01 3.24776828e-01 2.70011932e-01
8.99599671e-01 -1.02555645e+00 1.70183554e-01 8.53197277e-01
-5.69722727e-02 3.17427218e-01 5.25051475e-01 -4.55438375e-01
-1.12191010e+00 -1.55016148e+00 7.85006225e-01 -6.33319497e-01
5.85734725e-01 -6.87175632e-01 -1.08693600e+00 9.65398848e-01
5.92836216e-02 1.25254571e-01 7.96339571e-01 -3.25325817e-01
-6.77185774e-01 -1.04175312e-02 -1.32625091e+00 3.61329645e-01
1.06394732e+00 -5.88711739e-01 -8.16406548e-01 6.54010534e-01
9.28007603e-01 -2.27526277e-01 -6.08454585e-01 3.75543952e-01
2.31489375e-01 -4.99861985e-01 9.80161190e-01 -9.47640777e-01
4.89825070e-01 -2.88896888e-01 -6.07654870e-01 -1.32121789e+00
1.32249370e-01 -1.07072324e-01 -4.30157688e-03 1.25745547e+00
3.49634409e-01 -5.75925171e-01 6.40198052e-01 7.14434326e-01
-6.52165040e-02 -2.03979537e-01 -1.15668583e+00 -1.21298969e+00
3.56415361e-01 -4.77682531e-01 8.29191387e-01 1.13961494e+00
-2.87514299e-01 2.36875847e-01 -2.26759940e-01 4.06849414e-01
6.07342243e-01 2.59035170e-01 6.32761955e-01 -1.17293179e+00
-8.95951167e-02 1.87435672e-01 -5.58284879e-01 -1.14584994e+00
5.02763331e-01 -9.10174608e-01 2.09210012e-02 -1.57128024e+00
4.45405334e-01 -3.21369350e-01 -4.18312877e-01 3.37765664e-01
-1.20489307e-01 5.48268557e-02 4.48250435e-02 2.81029552e-01
-8.57158899e-01 6.83551192e-01 1.51043808e+00 -2.67040014e-01
1.38372853e-01 -1.14801295e-01 -8.57577682e-01 4.72898722e-01
6.96265578e-01 -3.85104120e-01 -1.01333201e+00 -7.20393479e-01
-1.56786609e-02 -2.62525678e-01 6.76749349e-01 -7.35732615e-01
-1.23346619e-01 -3.69269133e-01 1.04666010e-01 -1.52311236e-01
1.58588946e-01 -1.00902069e+00 -1.87819064e-01 1.19458094e-01
-5.33672690e-01 -3.77225220e-01 1.21030599e-01 6.57802701e-01
-4.34804142e-01 -2.94193178e-01 8.75376940e-01 1.19562246e-01
-1.16212893e+00 4.28142130e-01 3.85955811e-01 6.85388207e-01
1.33420491e+00 -3.14615697e-01 -5.86310029e-01 -5.00868142e-01
-7.62757957e-01 4.32888597e-01 7.67477632e-01 7.55793691e-01
7.45536149e-01 -1.56216729e+00 -8.13273311e-01 2.06146121e-01
9.43534791e-01 5.18465303e-02 4.74829637e-02 3.44180524e-01
-1.57855257e-01 4.98474836e-01 -1.39728963e-01 -6.92390025e-01
-7.45462000e-01 1.06752467e+00 2.45661229e-01 1.01863921e-01
-3.72466892e-01 5.97243369e-01 1.08004618e+00 -4.74125445e-01
1.58681214e-01 -2.53391951e-01 6.08777367e-02 -3.28995764e-01
5.71464241e-01 -3.28307375e-02 -1.75747618e-01 -7.20655322e-01
-2.47329190e-01 2.27722645e-01 -1.72261909e-01 -1.13926105e-01
1.09225094e+00 -3.53363365e-01 7.95606151e-02 3.86563480e-01
1.47222006e+00 -6.55332804e-01 -1.45320797e+00 -5.43813050e-01
9.29924995e-02 -4.64800328e-01 -2.42529944e-01 -9.33251500e-01
-6.99564338e-01 9.37806606e-01 6.76362336e-01 -3.92610654e-02
1.28386927e+00 3.86499614e-01 8.04636061e-01 2.77146399e-01
4.48300987e-01 -1.13418424e+00 9.71069559e-02 2.43631795e-01
1.03201437e+00 -1.52542019e+00 -5.98672271e-01 -2.46032521e-01
-8.63970041e-01 7.23861277e-01 8.68226707e-01 1.58682436e-01
4.31230247e-01 -3.26751083e-01 2.73919087e-02 1.18818924e-01
-6.02577150e-01 -6.24044016e-02 1.33300647e-01 1.09701371e+00
7.81178996e-02 -7.45028779e-02 -4.99265920e-03 6.97076917e-01
5.50407209e-02 2.05402687e-01 5.30791759e-01 9.47193027e-01
-2.69628853e-01 -1.15954912e+00 -2.62037903e-01 5.61531924e-04
8.66942399e-04 -3.92894670e-02 -5.88327348e-01 1.00575340e+00
1.79876402e-01 1.12512529e+00 5.88810630e-03 -1.37488857e-01
3.26475143e-01 1.46963343e-01 3.11823457e-01 -7.59253681e-01
3.62081558e-01 -5.09582520e-01 -1.45236284e-01 -5.75358748e-01
-5.61884642e-01 -4.36880291e-01 -9.69644368e-01 1.57132030e-01
2.58724056e-02 3.53716612e-01 6.39424562e-01 9.50812399e-01
4.23610061e-01 7.02261971e-03 7.77571738e-01 -1.11969337e-01
-5.75203359e-01 -8.20895672e-01 -5.60470402e-01 9.99062479e-01
5.78265131e-01 -6.94140375e-01 -4.94648516e-01 4.62847650e-01] | [10.337489128112793, 2.882642984390259] |
efd9486b-9cef-4cfe-9b8e-a06eda52e936 | view-inter-prediction-gan-unsupervised | 1811.02744 | null | http://arxiv.org/abs/1811.02744v1 | http://arxiv.org/pdf/1811.02744v1.pdf | View Inter-Prediction GAN: Unsupervised Representation Learning for 3D Shapes by Learning Global Shape Memories to Support Local View Predictions | In this paper we present a novel unsupervised representation learning
approach for 3D shapes, which is an important research challenge as it avoids
the manual effort required for collecting supervised data. Our method trains an
RNN-based neural network architecture to solve multiple view inter-prediction
tasks for each shape. Given several nearby views of a shape, we define view
inter-prediction as the task of predicting the center view between the input
views, and reconstructing the input views in a low-level feature space. The key
idea of our approach is to implement the shape representation as a
shape-specific global memory that is shared between all local view
inter-predictions for each shape. Intuitively, this memory enables the system
to aggregate information that is useful to better solve the view
inter-prediction tasks for each shape, and to leverage the memory as a
view-independent shape representation. Our approach obtains the best results
using a combination of L_2 and adversarial losses for the view inter-prediction
task. We show that VIP-GAN outperforms state-of-the-art methods in unsupervised
3D feature learning on three large scale 3D shape benchmarks. | ['Yu-Shen Liu', 'Matthias Zwicker', 'Zhizhong Han', 'Mingyang Shang'] | 2018-11-07 | null | null | null | null | ['3d-point-cloud-linear-classification'] | ['computer-vision'] | [ 7.74932653e-03 2.61353225e-01 -9.73798987e-03 -5.33547640e-01
-9.13098931e-01 -7.07968473e-01 5.31437874e-01 -3.33690733e-01
3.76954943e-01 6.06892966e-02 4.42049801e-01 -5.82059892e-03
1.28026903e-01 -1.03490281e+00 -9.53129709e-01 -7.42693365e-01
3.84981871e-01 8.94631207e-01 -4.73086052e-02 9.35532227e-02
4.86347452e-02 9.62808490e-01 -1.26133215e+00 5.70101440e-01
2.62125134e-01 1.26871419e+00 1.87789544e-01 5.85324883e-01
-3.11912969e-02 6.40177548e-01 -3.16070884e-01 -4.86909837e-01
6.44638538e-01 -1.40561700e-01 -7.56897986e-01 4.41964239e-01
7.48960137e-01 -5.13234437e-01 -3.30960244e-01 7.32723534e-01
5.74777007e-01 5.99704608e-02 9.13263798e-01 -1.00541437e+00
-1.00280106e+00 1.36523873e-01 -6.28298402e-01 -3.70714635e-01
9.54525992e-02 8.25034752e-02 1.12897003e+00 -1.16303861e+00
8.35008144e-01 1.26590276e+00 5.47424257e-01 6.05472386e-01
-1.44768727e+00 -3.21984589e-01 1.96855426e-01 -1.51623324e-01
-1.07115161e+00 -4.07853723e-01 1.25595021e+00 -5.47446489e-01
1.15800345e+00 2.65519251e-03 6.70782745e-01 1.08064711e+00
3.14660221e-01 9.06357944e-01 8.41528952e-01 4.89039309e-02
1.54048637e-01 -2.15438813e-01 -3.77728760e-01 7.03750193e-01
-2.63027549e-01 2.56464630e-01 -4.89733905e-01 -1.22484148e-01
1.18761504e+00 5.29491484e-01 -8.41681585e-02 -1.12415862e+00
-9.76960599e-01 8.56133401e-01 6.04407489e-01 -1.30890701e-02
-4.21652138e-01 6.22435808e-02 1.93986326e-01 4.73777622e-01
7.89069474e-01 1.90288231e-01 -6.96085274e-01 1.25445381e-01
-7.71071792e-01 2.43065637e-02 7.94194162e-01 9.11421239e-01
6.29360199e-01 2.99021304e-01 8.80727321e-02 6.82578385e-01
5.22771776e-01 7.86229312e-01 2.39754352e-03 -1.04697716e+00
5.66543162e-01 8.60602558e-01 -2.68178940e-01 -1.00852072e+00
-2.18122959e-01 -2.70046055e-01 -1.09477580e+00 6.23703897e-01
2.18955293e-01 2.48697028e-01 -9.61857796e-01 1.65174305e+00
2.46733069e-01 1.68342724e-01 8.98451284e-02 7.22728908e-01
9.97122943e-01 5.67971289e-01 -3.64792585e-01 1.92082956e-01
8.04016948e-01 -1.00445116e+00 -8.13212991e-03 -4.41048369e-02
2.27567330e-01 -7.51592755e-01 7.54850388e-01 4.13130149e-02
-1.34291947e+00 -6.61110282e-01 -8.93213511e-01 -2.73714483e-01
-3.34508419e-01 1.07612878e-01 3.24283898e-01 2.89334595e-01
-1.14120722e+00 6.16282940e-01 -8.70505571e-01 -1.52468532e-01
7.37560272e-01 3.59682590e-01 -7.63006270e-01 -9.46772099e-02
-2.34421566e-01 7.02601254e-01 -2.74878621e-01 -5.54790124e-02
-1.03541338e+00 -8.06547582e-01 -1.00739992e+00 -6.65294304e-02
2.36097559e-01 -1.00609291e+00 8.31909776e-01 -7.88010538e-01
-1.46425891e+00 1.26264131e+00 -2.95443088e-01 -2.31953219e-01
3.35798085e-01 -1.01994783e-01 4.84162057e-03 4.00002534e-03
9.49817598e-02 7.56549060e-01 1.25597978e+00 -1.76640797e+00
-1.16680093e-01 -8.12631190e-01 -8.58614892e-02 1.95684850e-01
3.67447346e-01 -5.59074998e-01 -4.50413167e-01 -8.45298588e-01
5.15805423e-01 -9.36869085e-01 -1.74739420e-01 2.98709750e-01
-3.97282392e-01 -2.33737342e-02 9.74208593e-01 -6.52152717e-01
2.43414178e-01 -2.13070703e+00 6.08006597e-01 2.45101139e-01
3.78419310e-01 -1.03458986e-01 -5.00520051e-01 2.95247614e-01
-1.04720227e-01 -3.85857970e-02 -1.81139976e-01 -8.83923769e-01
-1.10555440e-01 5.45825303e-01 -8.38147402e-01 1.99219614e-01
8.96940380e-02 1.39690518e+00 -5.55527151e-01 1.75521284e-01
3.57862771e-01 5.67357898e-01 -6.46575272e-01 6.88381374e-01
-2.39919677e-01 7.71358788e-01 -2.20573932e-01 6.12994909e-01
8.72390032e-01 -3.93564910e-01 4.78694215e-02 -4.50407326e-01
3.87088865e-01 2.19015673e-01 -8.01105499e-01 2.10820246e+00
-6.75212145e-01 1.80708274e-01 -1.09332375e-01 -1.02880633e+00
1.28853762e+00 2.33909041e-01 6.96674764e-01 -5.08821785e-01
-6.56666458e-02 1.83568969e-02 -5.84285676e-01 -8.56862143e-02
9.79437605e-02 -1.51626512e-01 -4.81988005e-02 8.75235677e-01
3.92949224e-01 -5.22503436e-01 -7.96140313e-01 3.17187607e-02
9.88297880e-01 3.44235986e-01 2.53377885e-01 9.40002054e-02
3.33064824e-01 -7.36182392e-01 4.24127460e-01 4.97115284e-01
1.58738971e-01 1.04479599e+00 2.68417805e-01 -9.50011492e-01
-1.17079782e+00 -1.70369685e+00 4.16976660e-01 8.64288211e-01
-8.14142376e-02 -1.62693530e-01 -1.89298958e-01 -1.24744844e+00
2.62129784e-01 5.69205642e-01 -8.60195756e-01 -1.69887662e-01
-4.85289156e-01 2.66464800e-01 1.39814168e-01 8.87198627e-01
2.41357923e-01 -1.23481488e+00 -3.98154974e-01 -7.17077032e-02
1.10290565e-01 -1.00896645e+00 -6.67668462e-01 2.47276574e-01
-1.18357885e+00 -9.44813311e-01 -6.20029509e-01 -8.18464518e-01
9.85400677e-01 4.39858824e-01 1.54958797e+00 -6.35803416e-02
2.65655722e-02 9.74224031e-01 -4.65357564e-02 -3.07997227e-01
-4.54024464e-01 -5.28483465e-02 -1.58079326e-01 1.67442352e-01
5.43301627e-02 -1.26662397e+00 -4.83399779e-01 3.08488220e-01
-6.57303154e-01 1.14693254e-01 5.33562064e-01 9.59662199e-01
1.15515947e+00 -5.66990733e-01 5.59884906e-01 -9.17871714e-01
5.06784543e-02 -3.12407374e-01 -5.82854271e-01 3.43309432e-01
-2.17573330e-01 3.68258744e-01 6.57903612e-01 -1.58435509e-01
-8.12537372e-01 3.83184910e-01 -1.94126174e-01 -1.06901169e+00
-2.49108508e-01 3.88363935e-02 -5.95283389e-01 1.67787392e-02
1.92443937e-01 5.74342489e-01 2.43043751e-01 -6.76262021e-01
6.22582495e-01 4.50204425e-02 4.22851831e-01 -5.95868468e-01
9.69973683e-01 7.11457551e-01 2.07544401e-01 -4.20452356e-01
-1.17015612e+00 -1.91164225e-01 -1.22067010e+00 -5.37790805e-02
1.01392734e+00 -9.97408032e-01 -6.87309504e-01 4.22301233e-01
-1.28425312e+00 -2.96282083e-01 -7.20770478e-01 -1.08002901e-01
-1.10300648e+00 2.60531574e-01 -2.80621290e-01 -3.95917863e-01
-5.68339050e-01 -1.09206700e+00 1.51881969e+00 -1.26197040e-01
2.54373706e-04 -1.13647103e+00 2.33483598e-01 5.00180006e-01
3.40978652e-01 2.79992491e-01 1.05125391e+00 -6.55229926e-01
-9.87414122e-01 -3.33301648e-02 -1.42711014e-01 5.11451423e-01
4.11812723e-01 -4.57886159e-01 -1.04648769e+00 -4.82110947e-01
2.95906633e-01 -6.28937185e-01 9.00126517e-01 3.23384732e-01
1.45472383e+00 -3.30944002e-01 3.00937686e-02 8.72465312e-01
1.28818703e+00 -4.31974744e-03 5.78788579e-01 -3.41000944e-01
9.42703426e-01 3.42413694e-01 1.50407568e-01 3.36186618e-01
5.14442027e-01 6.46821797e-01 7.84141839e-01 3.65705206e-03
-1.30854949e-01 -6.54012322e-01 1.70491368e-01 1.08369780e+00
-2.34705657e-01 -5.02779372e-02 -5.86211979e-01 4.26132917e-01
-1.89750957e+00 -9.56244171e-01 5.90361714e-01 2.24395132e+00
3.30963820e-01 -6.60456195e-02 -4.60830852e-02 -3.07238877e-01
2.00259209e-01 5.91793299e-01 -7.12751746e-01 -4.49678570e-01
-6.20294847e-02 3.00535560e-01 -1.66577864e-02 3.80148560e-01
-1.13959610e+00 7.44951129e-01 6.05780315e+00 5.28451324e-01
-1.11657703e+00 4.11561541e-02 7.78640032e-01 -8.94214436e-02
-5.53757191e-01 -2.17402995e-01 -4.76876199e-01 4.89473082e-02
3.40125591e-01 2.20389485e-01 6.20368659e-01 9.00772035e-01
-2.49561578e-01 3.31311882e-01 -1.49002349e+00 1.11982584e+00
3.88243347e-01 -1.55443180e+00 5.35933614e-01 2.70595044e-01
1.02783346e+00 1.68514341e-01 2.17700005e-01 9.08988267e-02
2.92612702e-01 -1.22188187e+00 4.40512627e-01 8.40133011e-01
7.12729156e-01 -6.75616622e-01 4.97554421e-01 4.35249746e-01
-1.39220750e+00 2.54465908e-01 -6.37882411e-01 2.07723677e-01
1.42438784e-01 4.82562959e-01 -5.74708223e-01 7.82631576e-01
4.70901221e-01 1.15277243e+00 -3.34081203e-01 5.32865882e-01
-2.33646676e-01 2.09957018e-01 -2.44801134e-01 5.99098086e-01
-6.44425601e-02 -2.54539281e-01 7.46597111e-01 5.13387680e-01
4.57983702e-01 -8.10327232e-02 3.18947136e-01 1.25274432e+00
-4.42543328e-01 -2.77308732e-01 -1.33205414e+00 2.00256079e-01
3.22495043e-01 1.06021535e+00 -4.43396151e-01 -6.79259673e-02
-5.64162850e-01 1.25330257e+00 7.98103631e-01 1.63365513e-01
-4.82293457e-01 4.68754143e-01 6.97859824e-01 1.99342892e-02
9.03747439e-01 -2.58537561e-01 -5.38972199e-01 -1.13105106e+00
1.79558992e-01 -7.66450047e-01 3.22477102e-01 -8.36282313e-01
-1.71264374e+00 4.69324559e-01 -5.02488256e-01 -1.64389050e+00
-2.64033854e-01 -6.14679933e-01 -8.21773469e-01 7.95732677e-01
-1.10414064e+00 -1.80238986e+00 -1.85352877e-01 6.51455462e-01
6.99870229e-01 -6.34648681e-01 1.24777770e+00 -1.17351249e-01
9.40678790e-02 6.28323972e-01 -1.00310050e-01 2.62290180e-01
5.24913967e-01 -1.32620978e+00 6.89877152e-01 5.04651427e-01
8.19637239e-01 2.06410244e-01 -8.01995322e-02 -6.33785009e-01
-1.60429502e+00 -1.25978017e+00 7.29243338e-01 -8.14022422e-01
2.62870938e-01 -2.93684632e-01 -8.17066908e-01 9.76434767e-01
9.10026431e-02 5.12208819e-01 7.77773142e-01 -1.09686613e-01
-8.25500846e-01 -1.31632462e-01 -1.08858228e+00 3.70697051e-01
1.32373226e+00 -9.47699845e-01 -6.98493004e-01 -1.34739816e-01
6.82745457e-01 -5.31829357e-01 -1.17569220e+00 3.37561816e-01
6.14646852e-01 -1.06857932e+00 1.37288737e+00 -7.95059800e-01
8.74090791e-01 -8.23250487e-02 -6.36195242e-01 -1.51038289e+00
-3.15046370e-01 -2.73792982e-01 -4.75072384e-01 1.08850634e+00
3.28796178e-01 -4.61475253e-01 9.39099491e-01 5.51315784e-01
-7.94028770e-03 -1.24573910e+00 -1.04379272e+00 -6.32843316e-01
2.15148628e-01 -3.32753927e-01 7.49762058e-01 7.12319374e-01
-7.58847117e-01 4.50021684e-01 -5.96243918e-01 2.47142032e-01
8.19455028e-01 1.02115178e+00 1.10511613e+00 -1.19365299e+00
-5.22268772e-01 -2.31054679e-01 -6.01946712e-01 -1.42489767e+00
3.63920510e-01 -1.28501856e+00 -3.05343211e-01 -1.44677460e+00
2.17677698e-01 -3.11705381e-01 -2.44480997e-01 6.20675385e-01
2.68811435e-01 2.98155457e-01 5.83865643e-01 5.31465262e-02
-6.07584536e-01 8.99449944e-01 1.59131420e+00 -5.11055946e-01
-1.05648257e-01 3.45183104e-01 -6.01198316e-01 7.86976576e-01
4.47206616e-01 -4.25966799e-01 -3.91156882e-01 -7.39731014e-01
1.42798662e-01 2.43325561e-01 6.77150011e-01 -7.07424104e-01
2.07245111e-01 5.42670190e-02 8.45592082e-01 -1.14017045e+00
6.47448182e-01 -1.25541461e+00 1.94135621e-01 7.60806426e-02
-2.36920461e-01 1.03940621e-01 8.35098475e-02 7.94560373e-01
-2.02002287e-01 1.77790031e-01 6.49425805e-01 -2.96814233e-01
-3.85043114e-01 7.60571361e-01 2.72714078e-01 2.09016517e-01
8.75016510e-01 -3.05085182e-01 1.38814589e-02 -7.12298810e-01
-1.08559656e+00 -5.41316085e-02 6.41658425e-01 6.22530520e-01
1.08644331e+00 -1.78299749e+00 -6.19718790e-01 7.32915938e-01
9.19638127e-02 3.27339411e-01 3.86869818e-01 4.09741879e-01
4.16769236e-02 1.71108041e-02 -4.04606014e-01 -8.73667657e-01
-1.20750380e+00 5.27273715e-01 3.46270114e-01 -6.32545114e-01
-8.87024462e-01 8.32436502e-01 5.66941798e-01 -1.10073924e+00
1.33967698e-01 -1.29105493e-01 5.06556407e-02 -2.47585058e-01
1.59552991e-01 1.00866795e-01 -1.19689759e-02 -7.71515787e-01
-9.77438390e-02 1.09161484e+00 1.04887396e-01 2.02974528e-01
1.81652641e+00 3.02171372e-02 -1.87108159e-01 6.72163785e-01
1.53742623e+00 1.05543479e-01 -1.52486229e+00 -3.93657595e-01
-4.94741678e-01 -4.58139360e-01 -2.07823992e-01 -7.69532323e-01
-1.66676688e+00 1.03582966e+00 4.13368046e-01 -7.33106807e-02
1.04865062e+00 4.16247040e-01 8.30717444e-01 4.91330922e-01
5.55009842e-01 -5.98916829e-01 3.42893898e-01 8.18490446e-01
1.47870362e+00 -1.25043166e+00 -3.90637182e-02 -2.93354094e-01
-6.57155037e-01 1.16492939e+00 6.67328835e-01 -5.27760804e-01
1.14561367e+00 2.29070827e-01 6.36370331e-02 -4.62615371e-01
-8.47470880e-01 6.13901839e-02 7.72020519e-01 7.44814813e-01
-1.93809625e-02 1.75314784e-01 8.02022278e-01 7.86852181e-01
-1.70579895e-01 -3.91330540e-01 -5.81489457e-03 6.07434273e-01
5.27977124e-02 -1.20254958e+00 8.97955224e-02 6.67315662e-01
-2.57094540e-02 2.14308485e-01 -6.81941509e-01 3.58527273e-01
-9.16949138e-02 3.08981568e-01 2.80694008e-01 -6.12406313e-01
3.98487896e-01 1.51978642e-01 6.58367932e-01 -6.05221510e-01
-4.31926310e-01 -7.92265236e-02 -3.99274796e-01 -8.64741266e-01
-4.12412882e-01 -4.58973885e-01 -9.25813675e-01 -2.23792687e-01
1.89494193e-01 -5.25701702e-01 1.36553854e-01 9.20373321e-01
8.53737891e-01 2.64079660e-01 9.94279623e-01 -1.33062494e+00
-5.58639407e-01 -4.42340255e-01 -5.07215083e-01 7.71404624e-01
2.89512217e-01 -7.31632411e-01 -1.52127072e-01 1.07229203e-02] | [8.338624954223633, -3.460123062133789] |
31eebbd3-6f0c-4300-8d44-0217cd5f710a | a-muze-net-music-generation-by-composing-the | 2111.12986 | null | https://arxiv.org/abs/2111.12986v1 | https://arxiv.org/pdf/2111.12986v1.pdf | A-Muze-Net: Music Generation by Composing the Harmony based on the Generated Melody | We present a method for the generation of Midi files of piano music. The method models the right and left hands using two networks, where the left hand is conditioned on the right hand. This way, the melody is generated before the harmony. The Midi is represented in a way that is invariant to the musical scale, and the melody is represented, for the purpose of conditioning the harmony, by the content of each bar, viewed as a chord. Finally, notes are added randomly, based on this chord representation, in order to enrich the generated audio. Our experiments show a significant improvement over the state of the art for training on such datasets, and demonstrate the contribution of each of the novel components. | ['Lior Wolf', 'Eliya Nachmani', 'Or Goren'] | 2021-11-25 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 3.82116050e-01 3.43996704e-01 1.45753203e-02 1.84472710e-01
-4.33575749e-01 -8.79530311e-01 3.43446016e-01 -1.27813384e-01
-2.57134378e-01 6.07086003e-01 5.04547298e-01 7.40788206e-02
-5.36933020e-02 -8.21201444e-01 -7.53539920e-01 -8.00723255e-01
-9.36002359e-02 5.88645756e-01 1.07503742e-01 -4.28900599e-01
8.28494504e-02 2.34219596e-01 -1.58620751e+00 5.47999322e-01
2.30707690e-01 9.86598134e-01 -7.27633759e-02 8.21639597e-01
1.81562662e-01 7.87155688e-01 -7.85838485e-01 -3.07115465e-01
4.58080560e-01 -8.38603735e-01 -6.55451238e-01 -2.46715978e-01
8.51586685e-02 -8.20956901e-02 -2.50383317e-01 7.48600423e-01
4.78951067e-01 4.70437348e-01 6.22249603e-01 -9.05924201e-01
-1.02297142e-01 1.42123961e+00 -3.27715099e-01 -4.27024871e-01
2.65185446e-01 -1.37502313e-01 1.51355624e+00 -5.70054531e-01
9.09084439e-01 9.67641056e-01 6.46800816e-01 5.67914188e-01
-1.32047188e+00 -8.02593946e-01 -1.79092273e-01 -1.73320949e-01
-1.28735471e+00 -3.95092726e-01 1.02048385e+00 -4.04878199e-01
5.71347535e-01 2.98484534e-01 9.16614652e-01 7.13424325e-01
-1.28042057e-01 6.14599109e-01 4.97793466e-01 -8.01848769e-01
2.06155464e-01 -1.93417400e-01 -1.44755185e-01 3.06822091e-01
-3.89951289e-01 3.05836022e-01 -6.98327839e-01 -1.47469357e-01
8.40877712e-01 -3.52652103e-01 -5.56739420e-02 -3.81574541e-01
-1.11639202e+00 6.03178859e-01 3.38926613e-01 4.36892033e-01
-5.19589186e-01 2.54256994e-01 2.87709922e-01 1.73471838e-01
-1.57937780e-01 8.04985344e-01 -2.25346148e-01 -3.32532823e-01
-1.28818762e+00 6.39255345e-01 1.03890145e+00 4.74220693e-01
4.05455291e-01 1.02525197e-01 -5.00255764e-01 6.84055150e-01
2.98767388e-02 2.03321248e-01 7.04400837e-01 -1.13639188e+00
3.59392881e-01 1.72166511e-01 6.83515519e-02 -6.70941830e-01
-2.74362355e-01 -5.68707824e-01 -8.16261411e-01 5.30962050e-01
6.10578418e-01 -2.23894998e-01 -7.26960897e-01 1.98143375e+00
-3.39707802e-03 2.73310304e-01 -6.94265217e-02 6.30208731e-01
3.69142741e-01 6.89931929e-01 -2.58131057e-01 -1.92837253e-01
1.24284649e+00 -8.61318171e-01 -5.80223501e-01 2.92488188e-01
1.05907083e-01 -9.20911551e-01 1.07645047e+00 7.32004881e-01
-1.43938899e+00 -8.68912339e-01 -1.19961226e+00 1.77822441e-01
-9.94467512e-02 3.31929117e-01 1.83704972e-01 2.55405903e-01
-6.54342949e-01 1.07660306e+00 -5.59614539e-01 4.03791696e-01
-8.23694989e-02 3.28995287e-01 -9.62167978e-02 6.97287917e-01
-1.29334283e+00 3.83116841e-01 6.17085516e-01 -8.86335745e-02
-7.98972487e-01 -5.98917961e-01 -5.98004818e-01 5.33955157e-01
-4.80137169e-02 -5.58454037e-01 1.50179029e+00 -9.36146319e-01
-1.96744990e+00 7.12969899e-01 1.37275904e-01 -8.03211391e-01
5.44225872e-01 7.64366016e-02 -1.28210843e-01 1.31653398e-01
-1.93495795e-01 8.06637704e-01 9.75894392e-01 -1.20445096e+00
-7.56117821e-01 -1.34640619e-01 4.21164632e-02 1.25323862e-01
-2.21786246e-01 -2.90754735e-01 -5.80849946e-01 -9.31487858e-01
4.49798293e-02 -1.36008084e+00 -1.15385419e-02 -6.46844089e-01
-5.54244399e-01 -2.95577925e-02 3.79543692e-01 -8.12226057e-01
1.78594887e+00 -2.72252154e+00 3.73913467e-01 6.66689873e-01
3.45156640e-02 -9.77730379e-02 -1.55432492e-01 4.86874044e-01
-2.79044598e-01 -7.00834468e-02 -1.57166809e-01 -3.95956963e-01
2.09704012e-01 -3.64522710e-02 -6.86011851e-01 -1.46671906e-01
-3.07162583e-01 6.69065356e-01 -6.76339507e-01 -1.46704659e-01
-9.91608649e-02 4.02131885e-01 -8.51550102e-01 2.93106377e-01
-4.31875348e-01 4.56201553e-01 2.22890377e-02 -1.87665060e-01
1.55549720e-01 7.73320571e-02 4.94647145e-01 -2.31174618e-01
-2.92270124e-01 7.94293106e-01 -1.50124443e+00 1.59460807e+00
-4.22052681e-01 3.98216069e-01 -1.51193947e-01 -5.73317945e-01
9.50662374e-01 7.64005244e-01 5.20321906e-01 -3.15885186e-01
5.94898239e-02 9.06259790e-02 3.80170524e-01 3.68272588e-02
6.36282086e-01 -3.40953559e-01 -2.45581657e-01 7.42078602e-01
-3.48339677e-02 -4.87908542e-01 7.57944465e-01 8.36929455e-02
5.66292763e-01 4.01351362e-01 2.93139577e-01 2.76693434e-01
2.82581896e-01 -2.86735147e-01 3.54271531e-01 7.65320301e-01
3.57296258e-01 7.69402385e-01 6.22417152e-01 -3.87009889e-01
-1.00591457e+00 -1.09973824e+00 3.51049215e-01 1.36504197e+00
-3.16181481e-01 -7.41648376e-01 -9.43804681e-01 -2.60303766e-01
-1.04166411e-01 8.10328841e-01 -7.21465886e-01 -1.61226094e-01
-9.16727722e-01 1.84172671e-02 5.83580315e-01 5.85079312e-01
1.39387339e-01 -1.66839981e+00 -6.53090835e-01 4.42269802e-01
-2.68252343e-01 -5.50662041e-01 -6.76335096e-01 3.30343962e-01
-7.24140286e-01 -7.84761071e-01 -5.68582594e-01 -8.42058957e-01
3.30118239e-02 -2.83775955e-01 1.12279320e+00 -1.06193267e-01
-7.62008056e-02 -1.74566790e-01 -1.47088677e-01 -5.17182827e-01
-7.23664343e-01 2.29108185e-01 -3.56763117e-02 1.17861092e-01
-2.14457408e-01 -1.00147748e+00 -4.90911692e-01 -7.60099217e-02
-8.87314260e-01 2.14281440e-01 3.32617790e-01 8.38104367e-01
8.08528841e-01 2.06779525e-01 4.41890717e-01 -7.69858479e-01
6.23133898e-01 2.49098353e-02 -4.83872116e-01 -1.12613432e-01
-1.32323923e-02 1.12382837e-01 7.21509278e-01 -6.44459486e-01
-7.38384366e-01 3.64661396e-01 -2.46610224e-01 -3.20056438e-01
-1.27394006e-01 5.94679952e-01 -6.90712929e-02 4.49587435e-01
8.02324474e-01 -1.01828858e-01 -2.06395462e-01 -6.45500720e-01
7.77721226e-01 5.47060966e-01 1.13203120e+00 -6.57342434e-01
8.16692293e-01 2.05594704e-01 -3.33748236e-02 -6.04449153e-01
-7.55865037e-01 -2.62470901e-01 -8.45743418e-01 -1.72237635e-01
7.88338900e-01 -5.80121100e-01 -7.49183118e-01 2.92030990e-01
-1.00711846e+00 -4.33561623e-01 -1.08079195e+00 5.59795558e-01
-8.52431357e-01 -2.18649637e-02 -7.52193034e-01 -7.57055521e-01
-4.08571243e-01 -7.51106381e-01 5.16495585e-01 3.07101250e-01
-9.65397179e-01 -5.59327483e-01 3.30755025e-01 1.43992454e-02
2.29029953e-02 2.46169806e-01 1.24106169e+00 -6.78610146e-01
-2.73295581e-01 -3.50321859e-01 3.54540884e-01 4.09476936e-01
1.09598570e-01 1.74992323e-01 -1.02981544e+00 -4.35976610e-02
-2.09700659e-01 -2.59817988e-01 7.47897685e-01 3.00517172e-01
9.11953092e-01 -2.90565073e-01 2.20625959e-02 3.82670492e-01
7.76238024e-01 3.74252260e-01 6.68473125e-01 2.94912428e-01
4.18992788e-01 3.05226713e-01 1.71688959e-01 6.84848189e-01
6.30360693e-02 7.81814158e-01 2.12337449e-01 -6.33938536e-02
-4.03376788e-01 -6.96059644e-01 2.50306636e-01 8.23348701e-01
-3.43244612e-01 1.58264358e-02 -4.82639819e-01 5.52796364e-01
-1.69216311e+00 -1.24624431e+00 3.25171560e-01 2.37913966e+00
1.20215988e+00 3.34746540e-01 5.89485765e-01 7.87581086e-01
5.76218486e-01 1.14644490e-01 -1.13079406e-01 -2.37085596e-01
-7.77519569e-02 7.73310900e-01 -1.83992848e-01 4.77087796e-01
-9.21316922e-01 8.65361273e-01 6.65572453e+00 6.32005870e-01
-1.19044840e+00 -4.04800892e-01 1.37249202e-01 -2.56234646e-01
-2.36229122e-01 2.87386347e-02 -6.02097273e-01 3.19544703e-01
5.87125719e-01 -8.35990012e-02 9.79704797e-01 6.75003707e-01
1.37262240e-01 4.32624161e-01 -1.18730164e+00 7.41536081e-01
-1.55000582e-01 -1.27117074e+00 2.16839865e-01 -1.16903663e-01
8.40766072e-01 -4.65096146e-01 2.63222665e-01 4.27060246e-01
2.57447541e-01 -8.30086172e-01 1.29643512e+00 4.72824037e-01
8.01712632e-01 -1.02129006e+00 3.29154104e-01 3.16907138e-01
-1.19259071e+00 -1.51270300e-01 2.89890449e-02 -4.22504634e-01
1.91692889e-01 2.46582672e-01 -8.90824616e-01 3.38469893e-01
4.36594486e-01 1.97603956e-01 -1.52085319e-01 1.14826024e+00
-5.22356689e-01 1.01178622e+00 -2.62601852e-01 3.50461870e-01
5.92771731e-02 -2.30443493e-01 6.35952532e-01 1.10721254e+00
4.75733161e-01 -1.38714552e-01 1.84710890e-01 7.58689344e-01
-3.05255413e-01 2.04326957e-01 -3.58910710e-01 -1.15218706e-01
6.19042516e-01 1.23076451e+00 -5.95530510e-01 -4.98828799e-01
1.73658863e-01 8.86567473e-01 5.17383404e-02 3.27452213e-01
-5.58391452e-01 -6.33953989e-01 2.19345018e-01 1.27914429e-01
5.78744292e-01 1.15973420e-01 -3.45733374e-01 -6.68942928e-01
-1.50816172e-01 -1.09067798e+00 6.11138105e-01 -8.30853879e-01
-1.01788604e+00 6.60458744e-01 -3.13529074e-01 -1.30196118e+00
-9.91386414e-01 -2.50428438e-01 -7.54867375e-01 8.75724912e-01
-7.26950824e-01 -9.93166745e-01 7.19612613e-02 3.95222276e-01
-3.32069732e-02 -1.09120160e-01 1.21389711e+00 1.69312656e-01
-3.76067869e-02 5.74524760e-01 -2.41990597e-03 2.82467395e-01
8.28725636e-01 -1.27601027e+00 3.93995196e-01 3.95950109e-01
8.32247436e-01 6.27268791e-01 7.19260633e-01 -4.37972099e-01
-5.86101115e-01 -7.35037923e-01 1.22474039e+00 -1.61811441e-01
6.27294123e-01 -4.45688248e-01 -7.32843459e-01 6.02266848e-01
1.57521367e-01 -5.73386490e-01 7.98002899e-01 2.60493100e-01
-4.40569222e-01 -2.55707592e-01 -5.94426751e-01 7.40222394e-01
4.42667782e-01 -6.13307953e-01 -9.37603831e-01 -2.12175623e-01
6.08884573e-01 -5.97989261e-01 -6.09304368e-01 2.47051194e-01
9.38609898e-01 -9.28239763e-01 9.32879090e-01 -5.84027290e-01
6.02153361e-01 -5.90704322e-01 -1.96892954e-02 -1.43601751e+00
-6.20148480e-01 -8.99287462e-01 -2.47529835e-01 1.24829304e+00
5.11843503e-01 1.52945369e-01 7.19654083e-01 7.54269287e-02
-1.45457387e-01 -2.58951366e-01 -7.76308239e-01 -5.48950791e-01
-7.39757940e-02 -3.97634894e-01 9.21559691e-01 6.99589968e-01
2.81347364e-01 7.56068110e-01 -6.23551369e-01 -3.25919718e-01
1.16629988e-01 4.61301327e-01 8.56208980e-01 -1.33776057e+00
-8.28887224e-01 -7.73181796e-01 -2.36197095e-02 -7.47147858e-01
-5.63419759e-02 -8.85174513e-01 2.29258612e-01 -1.09829819e+00
2.11894102e-02 -1.60870165e-01 -7.33569503e-01 3.85286629e-01
-1.31334588e-01 7.08741367e-01 7.45768547e-01 5.37102856e-02
-2.18191221e-01 1.80695862e-01 1.14475167e+00 -2.22839490e-01
-7.56609559e-01 3.54665279e-01 -4.87880081e-01 1.00108504e+00
8.24198544e-01 -3.50281745e-01 -3.12582761e-01 9.35397819e-02
3.05334300e-01 2.58793443e-01 1.54926315e-01 -1.31284189e+00
1.06775865e-01 1.49930656e-01 4.68056858e-01 -7.21315145e-01
4.52059150e-01 -5.93670905e-01 4.21847224e-01 4.36286300e-01
-8.36500823e-01 -1.95776045e-01 1.13306820e-01 7.13362694e-02
-3.59414577e-01 -4.37331289e-01 8.02246392e-01 8.48203059e-03
2.20808666e-02 4.43050712e-02 -2.72790611e-01 6.00454863e-03
3.57379556e-01 4.15191129e-02 5.62163115e-01 -7.31606781e-01
-1.12817228e+00 -4.21669573e-01 1.61720768e-01 2.06623375e-01
8.31089765e-02 -1.53096962e+00 -5.27498424e-01 3.71791691e-01
-1.19444706e-01 -2.50570595e-01 2.39250258e-01 2.69338012e-01
-6.26977533e-02 1.68604493e-01 -3.50002408e-01 -7.25477934e-02
-1.21876204e+00 4.69803363e-01 1.86423436e-01 -7.58592248e-01
-5.47697783e-01 5.81806242e-01 4.54935074e-01 -2.42332637e-01
5.59620559e-01 -2.93117315e-01 -4.00653571e-01 2.82673717e-01
6.25835180e-01 1.98246822e-01 -2.01283634e-01 -4.33287024e-01
2.08857641e-01 4.63016808e-01 3.17551136e-01 -8.62534106e-01
1.19155180e+00 1.81975037e-01 -2.23952368e-01 7.74411023e-01
5.30242741e-01 7.80087829e-01 -1.04977894e+00 -3.14510390e-02
-1.41928107e-01 2.65571941e-02 -5.22049665e-01 -1.00729692e+00
-6.03220940e-01 8.42023671e-01 2.15121105e-01 2.98112452e-01
1.07926798e+00 -3.42201144e-01 7.31735051e-01 3.13010305e-01
1.30488247e-01 -1.08299446e+00 1.83394060e-01 8.77541721e-01
9.06672895e-01 -2.21395701e-01 -4.37844038e-01 4.54488620e-02
-6.19543254e-01 1.21757925e+00 1.97172165e-01 -1.98966712e-01
5.52769542e-01 3.78251612e-01 2.23349541e-01 2.06182346e-01
-4.54679787e-01 7.91550353e-02 4.72023100e-01 3.70935798e-01
6.79981828e-01 2.73178846e-01 -6.23896718e-02 1.26030159e+00
-1.06549811e+00 1.78053722e-01 3.52713078e-01 5.32508552e-01
-2.30091512e-01 -1.29562008e+00 -5.18253505e-01 3.50115411e-02
-6.54729724e-01 -2.75687695e-01 -5.16506851e-01 6.24229908e-01
5.37362456e-01 6.79089606e-01 2.70666480e-01 -4.49003994e-01
5.92498839e-01 5.09880304e-01 4.95668054e-01 -5.29544413e-01
-9.51136410e-01 5.20322502e-01 -3.70240100e-02 -2.42969364e-01
1.38905421e-01 -2.86692291e-01 -1.52762675e+00 4.17916588e-02
-1.51290081e-03 6.80641174e-01 4.45224673e-01 6.12554252e-01
8.12419131e-02 7.58529484e-01 7.49022484e-01 -1.25851333e+00
-6.15136385e-01 -1.19901502e+00 -8.96624804e-01 7.69952953e-01
2.11611718e-01 -2.27642581e-01 -1.74237877e-01 4.97730762e-01] | [16.002500534057617, 5.493551731109619] |
add560c5-985f-4eb4-9e62-642c30cffabb | semantic-graph-convolutional-network-for | 1910.09183 | null | https://arxiv.org/abs/1910.09183v1 | https://arxiv.org/pdf/1910.09183v1.pdf | Semantic Graph Convolutional Network for Implicit Discourse Relation Classification | Implicit discourse relation classification is of great importance for discourse parsing, but remains a challenging problem due to the absence of explicit discourse connectives communicating these relations. Modeling the semantic interactions between the two arguments of a relation has proven useful for detecting implicit discourse relations. However, most previous approaches model such semantic interactions from a shallow interactive level, which is inadequate on capturing enough semantic information. In this paper, we propose a novel and effective Semantic Graph Convolutional Network (SGCN) to enhance the modeling of inter-argument semantics on a deeper interaction level for implicit discourse relation classification. We first build an interaction graph over representations of the two arguments, and then automatically extract in-depth semantic interactive information through graph convolution. Experimental results on the English corpus PDTB and the Chinese corpus CDTB both demonstrate the superiority of our model to previous state-of-the-art systems. | ['Jie zhou', 'Wei Cheng', 'Ruiying Geng', 'Yingxue Zhang', 'Ping Jian', 'Fandong Meng'] | 2019-10-21 | null | null | null | null | ['implicit-discourse-relation-classification'] | ['natural-language-processing'] | [ 4.76228356e-01 1.14081407e+00 -3.42934400e-01 -3.67542088e-01
-2.39650577e-01 -6.63076162e-01 8.75332057e-01 6.30625069e-01
-7.41815194e-02 5.61860740e-01 7.32235134e-01 -8.41801465e-01
-1.44479182e-02 -1.10192692e+00 -5.57966709e-01 -1.91937968e-01
-1.59005776e-01 5.34540594e-01 5.54511189e-01 -6.89364374e-01
2.68492430e-01 -2.01391533e-01 -1.28718925e+00 1.10308468e+00
1.14402926e+00 8.18034172e-01 1.53062895e-01 6.16163909e-01
-7.38269687e-01 1.80433750e+00 -8.88427734e-01 -4.75245565e-01
-6.00287139e-01 -7.67811656e-01 -1.73997855e+00 -2.52790928e-01
-4.80829366e-02 -2.17030078e-01 -3.06071162e-01 9.96433437e-01
-8.64515547e-03 1.30479425e-01 5.38340986e-01 -9.63139176e-01
-7.17993736e-01 1.23129940e+00 -1.01372689e-01 2.53390461e-01
5.16215324e-01 -3.43295604e-01 1.59966350e+00 -1.78619668e-01
9.19501781e-01 1.62183416e+00 5.17432630e-01 5.30628204e-01
-1.05647731e+00 -2.34472066e-01 7.23908842e-01 7.31800556e-01
-5.02615154e-01 -1.51055127e-01 1.02670038e+00 -3.89495760e-01
1.43861949e+00 5.13266265e-01 6.52640998e-01 9.97409880e-01
-5.13487816e-01 1.02606690e+00 6.54383838e-01 -6.84596300e-01
-6.41480684e-02 -3.69661331e-01 9.73860562e-01 5.76449275e-01
-1.19937874e-01 -4.07799065e-01 -3.58589143e-01 3.80644426e-02
5.02465546e-01 -5.92234612e-01 -2.83437133e-01 2.73498520e-02
-7.31675446e-01 1.07891154e+00 8.42094541e-01 6.34011149e-01
-6.09289184e-02 1.34453863e-01 7.98755348e-01 4.61292237e-01
8.28293979e-01 5.09839654e-01 -2.88719058e-01 -3.47651750e-01
-6.42921254e-02 2.90132523e-01 1.16345966e+00 8.41998875e-01
3.69295806e-01 -4.63789910e-01 -1.08989939e-01 8.21849644e-01
2.56508738e-01 -3.10654610e-01 6.70616180e-02 -9.75982428e-01
8.76979589e-01 1.47510135e+00 -2.01754183e-01 -1.24573302e+00
-5.80014467e-01 8.50775316e-02 -4.57368433e-01 -3.47680561e-02
5.91596723e-01 1.06548063e-01 -3.82754743e-01 1.65146875e+00
4.20038342e-01 1.44765712e-02 4.18301791e-01 9.65350747e-01
1.64299607e+00 5.55027246e-01 3.11535537e-01 -5.17126210e-02
1.74523842e+00 -1.26013803e+00 -1.18251348e+00 -4.30550814e-01
1.24243021e+00 -3.07981163e-01 9.09214675e-01 -1.54142871e-01
-9.65007603e-01 -2.28501469e-01 -1.07613850e+00 -6.47183537e-01
-3.27533692e-01 -1.12135604e-01 1.17540371e+00 2.56643534e-01
-6.49143279e-01 3.94382626e-01 -7.56115675e-01 3.23745562e-03
7.09971011e-01 2.20104679e-01 -2.16435179e-01 2.09234044e-01
-1.59233081e+00 1.03093934e+00 8.18477392e-01 1.35987580e-01
-1.95610747e-01 -3.91616732e-01 -1.35215902e+00 2.41604939e-01
7.78453469e-01 -4.45011258e-01 1.40367925e+00 -1.21401584e+00
-1.59074378e+00 1.01075327e+00 -2.48830363e-01 -7.16031015e-01
2.90006429e-01 -4.26168591e-01 -2.76281331e-02 2.94694722e-01
-6.99156299e-02 1.41027138e-01 2.24668205e-01 -1.13790417e+00
-4.08454597e-01 -3.34586203e-01 1.04155493e+00 5.92280686e-01
-1.35393068e-01 3.50626200e-01 -1.53958410e-01 -3.82484078e-01
2.78014541e-01 -5.25649250e-01 -1.31952852e-01 -4.96700823e-01
-6.18852735e-01 -9.18674886e-01 1.06401455e+00 -9.41832006e-01
1.42502964e+00 -2.11715126e+00 3.53251278e-01 -3.35064650e-01
5.18330812e-01 5.23487449e-01 8.51559490e-02 4.66161817e-01
-1.08898804e-01 1.85570508e-01 -2.09296688e-01 -1.42534554e-01
5.33275455e-02 4.63612497e-01 -3.70261133e-01 2.67565511e-02
4.71633911e-01 1.20880079e+00 -1.16152978e+00 -4.40238595e-01
5.39459437e-02 2.33908013e-01 -5.42247415e-01 5.59595346e-01
-8.67405534e-01 4.21445817e-01 -8.04705024e-01 1.49566054e-01
3.34125906e-01 -6.71686351e-01 1.00283599e+00 -1.89543515e-01
1.41006976e-01 1.51173818e+00 -6.40861809e-01 1.51815987e+00
-2.76126742e-01 9.34888005e-01 1.59963131e-01 -1.62928617e+00
7.61599481e-01 5.08976758e-01 1.46765513e-02 -7.79246926e-01
4.20489013e-01 1.83627419e-02 4.87032235e-01 -5.75585008e-01
4.42565382e-01 7.07776994e-02 -1.72959045e-01 4.76959974e-01
-9.41724628e-02 2.95030773e-02 3.42112720e-01 4.89136875e-01
1.14816070e+00 2.22814411e-01 4.62338001e-01 -1.30837902e-01
7.95063615e-01 2.37244219e-01 3.14377964e-01 3.82438302e-01
1.78310230e-01 4.03699398e-01 1.34904957e+00 -4.52530354e-01
-4.63997096e-01 -6.76900804e-01 2.24557474e-01 1.05993605e+00
4.89992887e-01 -7.74323821e-01 -9.13688183e-01 -1.05713487e+00
-2.46884510e-01 8.16402078e-01 -7.27638543e-01 7.52464607e-02
-1.32077622e+00 -5.18602729e-01 6.83929145e-01 6.27039015e-01
7.71179616e-01 -1.26009762e+00 -5.69487453e-01 1.71905816e-01
-7.26603687e-01 -1.35282612e+00 2.47515023e-01 9.86745954e-02
-6.65714800e-01 -1.66070843e+00 1.92100078e-01 -9.82131183e-01
4.71189559e-01 -2.23376025e-02 1.46205008e+00 7.57836044e-01
1.56225950e-01 -1.27393678e-01 -6.38451576e-01 -2.59712160e-01
-7.42331386e-01 3.10628772e-01 -8.66363525e-01 -4.20564473e-01
4.19480681e-01 -2.68599540e-01 -3.00055325e-01 -7.09277689e-02
-4.93995100e-01 5.84474146e-01 -2.34776840e-01 9.42593634e-01
7.98563734e-02 -2.15067528e-02 5.67879438e-01 -1.65360785e+00
8.57658148e-01 -3.93373579e-01 -2.49738663e-01 1.36557758e-01
-7.93657377e-02 -8.38392675e-02 4.64868009e-01 -2.72748023e-01
-1.50050771e+00 -8.21862519e-01 -2.70981371e-01 4.20694828e-01
-6.61251172e-02 9.10450995e-01 -4.09018189e-01 5.97373664e-01
5.01730919e-01 -5.67648947e-01 4.24601622e-02 -2.88628519e-01
7.31890678e-01 5.10405660e-01 3.70570093e-01 -8.95331860e-01
1.70543820e-01 2.41789758e-01 -1.14164650e-01 -8.36183250e-01
-1.37824440e+00 -3.83144319e-01 -5.93084633e-01 1.56214520e-01
1.27331245e+00 -6.41256869e-01 -1.04223442e+00 3.24053675e-01
-1.74951088e+00 -6.65866137e-01 -3.22817653e-01 4.52989107e-03
-5.39331377e-01 5.12990236e-01 -1.11005330e+00 -7.97137141e-01
-2.97091812e-01 -1.00791240e+00 7.05159307e-01 6.29448295e-02
-6.89303637e-01 -1.39207852e+00 -2.16195613e-01 7.34152973e-01
1.53397827e-03 4.42305624e-01 1.40039062e+00 -9.56596792e-01
-5.01541495e-01 2.31929362e-01 -6.65144026e-01 1.60640061e-01
2.54922271e-01 -3.63513082e-01 -9.50386167e-01 3.43770444e-01
-1.65852308e-02 -4.87676769e-01 7.89971352e-01 2.15550840e-01
9.84745264e-01 -5.43493867e-01 -5.69479287e-01 3.10337216e-01
7.42202818e-01 8.40054303e-02 9.29375947e-01 5.85758805e-01
8.92116845e-01 1.01094532e+00 6.17036283e-01 -1.81255728e-01
6.46778822e-01 6.10367715e-01 5.16732097e-01 -2.37196326e-01
-3.50900650e-01 3.69650461e-02 5.65730855e-02 7.34636009e-01
-3.89840543e-01 -4.13880318e-01 -9.39383626e-01 4.61391836e-01
-2.25337839e+00 -9.62256432e-01 -7.81780541e-01 1.31665623e+00
1.15810454e+00 3.07264596e-01 -1.76102772e-01 2.86199152e-01
6.39312029e-01 4.86972511e-01 -8.39449242e-02 -5.78968883e-01
-3.06775093e-01 2.98933029e-01 -2.10169286e-01 7.48319745e-01
-1.32146263e+00 1.18736768e+00 5.21584845e+00 6.06464624e-01
-6.72023237e-01 1.25476524e-01 6.10924006e-01 6.41943932e-01
-3.62927079e-01 3.22290391e-01 -4.08652693e-01 2.41002813e-02
7.11809099e-01 7.07641691e-02 1.08176790e-01 6.66587710e-01
-3.76962304e-01 -5.12367189e-02 -1.27403176e+00 4.28464413e-01
-2.18928471e-01 -1.83638871e+00 2.62276791e-02 -2.13259399e-01
3.80938679e-01 -4.24737692e-01 -4.42101330e-01 3.85323137e-01
5.18421233e-01 -1.21578288e+00 5.95209777e-01 -8.59442651e-02
5.19966900e-01 -4.04354155e-01 8.67951512e-01 3.80798817e-01
-1.22504127e+00 4.60140854e-02 6.22162502e-03 -7.54367471e-01
2.07157195e-01 6.54949322e-02 -8.30628514e-01 7.84278810e-01
3.95286947e-01 8.20076108e-01 -2.05907658e-01 8.88259262e-02
-1.05221069e+00 8.15831542e-01 -1.26453653e-01 -1.74970150e-01
3.10470611e-01 -1.93878084e-01 5.35568476e-01 1.16620529e+00
-3.73933017e-01 6.22603238e-01 1.26252338e-01 9.88398194e-01
-2.33865306e-01 -1.49249937e-02 -4.08842683e-01 -4.04895157e-01
3.11284006e-01 9.77163672e-01 -7.66568422e-01 -4.84574765e-01
-5.60894012e-01 7.38617897e-01 9.26327407e-01 2.69774646e-01
-6.70939803e-01 -1.72091201e-01 6.59529924e-01 -5.03974818e-02
1.26352578e-01 -2.18135849e-01 -3.39916021e-01 -9.47467625e-01
5.39690256e-02 -8.67833674e-01 7.52312779e-01 -4.72325146e-01
-1.17235529e+00 6.13517821e-01 4.63394262e-02 -5.60357392e-01
-3.35861742e-01 -8.49691033e-01 -8.28281641e-01 9.61170852e-01
-1.31883931e+00 -1.53352344e+00 -3.75281900e-01 1.82342172e-01
6.01032853e-01 1.89652458e-01 1.16994584e+00 -7.28325248e-02
-1.61684901e-01 2.43059546e-01 -7.34090626e-01 7.31653094e-01
2.09754519e-02 -1.40089571e+00 6.86957657e-01 4.58150446e-01
-2.48537567e-02 7.34737754e-01 5.65007925e-01 -6.32715404e-01
-1.00912142e+00 -6.25028551e-01 9.50662792e-01 -6.00025833e-01
8.58030379e-01 -4.07217920e-01 -1.35609531e+00 9.39878702e-01
6.28670454e-01 -3.21727276e-01 6.67817116e-01 7.93906152e-01
-5.03843069e-01 6.94106221e-01 -5.55753887e-01 6.37619078e-01
1.37025869e+00 -6.57433033e-01 -1.32590747e+00 2.00539276e-01
1.28575635e+00 -8.33537400e-01 -7.26167738e-01 5.40680885e-01
9.17913988e-02 -8.08026612e-01 1.04811883e+00 -1.02305400e+00
9.59914267e-01 3.33354925e-03 -1.41183706e-02 -1.02552187e+00
9.42882895e-02 -4.03417498e-01 -4.96105909e-01 1.30060291e+00
5.23682952e-01 -5.70414901e-01 7.69019485e-01 4.66659188e-01
-4.41661239e-01 -6.82703912e-01 -6.95995510e-01 -4.18649852e-01
1.08709216e-01 -4.97407317e-01 4.82628614e-01 1.24354863e+00
7.83873737e-01 1.17771482e+00 3.47508460e-01 6.37905672e-03
1.62518874e-01 4.74533379e-01 6.83036029e-01 -1.29232562e+00
-3.96871448e-01 -7.60086000e-01 -2.86578894e-01 -1.27211785e+00
8.04920733e-01 -1.10797584e+00 -2.08707005e-01 -2.19814897e+00
-2.99122767e-03 -5.94191253e-01 8.27665851e-02 2.92570174e-01
-5.30506492e-01 -3.15317929e-01 -7.69561604e-02 -3.71861737e-03
-6.23426378e-01 6.70064986e-01 1.30634105e+00 -4.78057206e-01
-1.74336985e-01 -2.57097274e-01 -5.90809524e-01 1.06947637e+00
6.80936813e-01 -1.57892238e-02 -7.42741585e-01 -6.38384759e-01
3.25691938e-01 1.40580222e-01 2.64885426e-01 -2.31037244e-01
5.29021546e-02 -1.94975153e-01 -3.01933676e-01 -4.73500192e-01
3.57577384e-01 -4.02846634e-01 -3.97668660e-01 2.79614657e-01
-6.11689985e-01 -3.80415022e-01 2.25400403e-01 3.82242143e-01
-6.38073802e-01 -3.33898038e-01 1.77029178e-01 -1.81168169e-01
-8.31223845e-01 -3.22075456e-01 -1.95658341e-01 5.18674493e-01
7.18749762e-01 5.13334125e-02 -1.04756296e+00 -3.23954761e-01
-8.15070033e-01 2.84473628e-01 1.94548175e-01 4.70262855e-01
5.08284748e-01 -9.37822104e-01 -6.19928420e-01 -1.72443569e-01
-2.86573153e-02 5.87536871e-01 2.72812657e-02 6.37983620e-01
-7.43093848e-01 1.73538059e-01 1.69345945e-01 -4.62277353e-01
-1.70547593e+00 4.13801610e-01 2.98008293e-01 -7.62508214e-01
-1.00810456e+00 1.05818236e+00 6.91156089e-01 -4.48108643e-01
5.69014922e-02 -4.68323022e-01 -8.87089968e-01 8.39181021e-02
4.87875700e-01 -1.24367289e-02 -9.86787528e-02 -6.95023477e-01
-2.96779782e-01 8.74887928e-02 -6.45402223e-02 2.16407925e-01
1.33487427e+00 -2.96474434e-02 -7.13259995e-01 3.23087484e-01
8.25338721e-01 -1.68532774e-01 -8.42400372e-01 -1.96584493e-01
5.44308543e-01 -1.53431505e-01 -3.01389277e-01 -4.56067920e-01
-6.03209376e-01 1.01709640e+00 -4.36657727e-01 9.51938152e-01
7.76107788e-01 4.55984116e-01 6.70608401e-01 4.72984016e-01
7.84040987e-02 -9.56024528e-01 2.84637511e-01 1.13374984e+00
9.64327216e-01 -1.05660319e+00 5.20447195e-02 -1.39575386e+00
-5.58727205e-01 1.25660026e+00 8.98825228e-01 -1.28917843e-01
3.80441219e-01 3.00954998e-01 6.49699420e-02 -7.01953411e-01
-8.11347365e-01 -2.75359273e-01 3.37471634e-01 5.34041643e-01
1.08861566e+00 1.80282742e-01 -5.44558823e-01 8.24595392e-01
-1.81568801e-01 -4.82250750e-01 1.80880412e-01 9.27894354e-01
-1.63019314e-01 -1.31363714e+00 7.48253316e-02 1.04777254e-01
-4.32693034e-01 -2.85069972e-01 -8.09756160e-01 1.00860608e+00
-1.25543654e-01 1.12730181e+00 2.52701432e-01 -5.05151488e-02
3.81063193e-01 -3.13053280e-02 6.16488993e-01 -8.82068634e-01
-8.85858595e-01 -3.13212126e-01 1.22696900e+00 -5.36696792e-01
-8.10913086e-01 -4.10469681e-01 -1.99234235e+00 -2.11445808e-01
-4.19384420e-01 3.29771727e-01 1.66341722e-01 1.27037632e+00
6.36956692e-02 1.12562287e+00 -1.01618096e-01 -3.62604976e-01
-1.08077288e-01 -1.09346735e+00 3.96557190e-02 9.66516137e-01
1.33971110e-01 -7.60257244e-01 -8.30512866e-02 3.39050293e-02] | [10.71588134765625, 9.254990577697754] |
fb9d9d25-6c90-40bd-9408-737a1eddaf08 | reconstructing-humpty-dumpty-multi-feature | 2212.06023 | null | https://arxiv.org/abs/2212.06023v1 | https://arxiv.org/pdf/2212.06023v1.pdf | Reconstructing Humpty Dumpty: Multi-feature Graph Autoencoder for Open Set Action Recognition | Most action recognition datasets and algorithms assume a closed world, where all test samples are instances of the known classes. In open set problems, test samples may be drawn from either known or unknown classes. Existing open set action recognition methods are typically based on extending closed set methods by adding post hoc analysis of classification scores or feature distances and do not capture the relations among all the video clip elements. Our approach uses the reconstruction error to determine the novelty of the video since unknown classes are harder to put back together and thus have a higher reconstruction error than videos from known classes. We refer to our solution to the open set action recognition problem as "Humpty Dumpty", due to its reconstruction abilities. Humpty Dumpty is a novel graph-based autoencoder that accounts for contextual and semantic relations among the clip pieces for improved reconstruction. A larger reconstruction error leads to an increased likelihood that the action can not be reconstructed, i.e., can not put Humpty Dumpty back together again, indicating that the action has never been seen before and is novel/unknown. Extensive experiments are performed on two publicly available action recognition datasets including HMDB-51 and UCF-101, showing the state-of-the-art performance for open set action recognition. | ['Christopher Funk', 'Anthony Hoogs', 'Ameya Shringi', 'Dawei Du'] | 2022-12-12 | null | null | null | null | ['open-set-action-recognition'] | ['computer-vision'] | [ 3.99839103e-01 9.99231860e-02 -2.26610646e-01 -2.04447180e-01
-5.69919348e-01 -4.67742860e-01 2.54072875e-01 -1.46923244e-01
-1.32730246e-01 8.14078093e-01 4.04363215e-01 2.86522955e-01
-3.96506906e-01 -6.68174386e-01 -1.01627994e+00 -8.17317069e-01
-2.73785204e-01 6.15465939e-01 3.84601772e-01 -5.69975078e-02
8.80117342e-02 2.60392606e-01 -1.85987246e+00 7.88080812e-01
2.23375723e-01 1.10104752e+00 -1.04132369e-01 6.06128335e-01
3.36434543e-01 1.12291610e+00 -9.42721725e-01 -4.55915660e-01
5.48673928e-01 -6.81457996e-01 -7.97894001e-01 5.49156487e-01
7.27187097e-01 -4.95303959e-01 -7.68329799e-01 9.68380213e-01
2.19754755e-01 5.45841157e-01 6.86263800e-01 -1.54545867e+00
-5.96497893e-01 6.77052438e-01 -3.35782081e-01 2.81260043e-01
7.93037534e-01 1.69328108e-01 1.11419916e+00 -7.72213221e-01
1.01219654e+00 1.03395641e+00 7.06353962e-01 4.88723814e-01
-1.02410531e+00 -4.88125861e-01 2.57107079e-01 1.14149415e+00
-1.33897150e+00 -3.46242279e-01 7.10818231e-01 -3.09949845e-01
1.17861772e+00 4.50198621e-01 9.59615171e-01 1.48949766e+00
-2.34843045e-03 1.05982196e+00 8.03968787e-01 -2.64168799e-01
3.75920087e-01 -1.78253621e-01 3.19910273e-02 4.94303674e-01
1.84122339e-01 1.67694852e-01 -4.03225094e-01 1.42072365e-01
6.61950052e-01 3.89100462e-01 -6.54585779e-01 -5.88052630e-01
-1.37667227e+00 6.11006975e-01 4.61246222e-01 3.99023235e-01
-4.62272972e-01 -1.83718070e-01 4.57020909e-01 5.30063927e-01
3.98438387e-02 3.60247254e-01 -5.38928032e-01 -4.15959775e-01
-6.08874798e-01 3.46147753e-02 9.10514534e-01 7.80578613e-01
5.05893409e-01 -1.31620765e-01 -2.03542873e-01 7.52480805e-01
-1.58313826e-01 1.79277062e-01 7.82824993e-01 -9.70014811e-01
5.34375429e-01 9.16081250e-01 -1.68227732e-01 -1.04175472e+00
-1.34452611e-01 -1.32666647e-01 -4.66100216e-01 3.20772290e-01
5.34313738e-01 1.28616065e-01 -7.57694125e-01 1.45360065e+00
2.17142507e-01 7.78825283e-01 3.18994105e-01 9.20386851e-01
9.65016723e-01 5.96943796e-01 -4.30925131e-01 -1.81129277e-01
9.24889982e-01 -9.67962086e-01 -5.83476245e-01 -1.04495071e-01
7.05183864e-01 -3.00336868e-01 7.58315325e-01 6.44099653e-01
-5.95868945e-01 -6.58652782e-01 -1.20513940e+00 3.04615051e-01
-4.85200644e-01 1.02798730e-01 3.01310211e-01 2.72677094e-01
-5.33348680e-01 1.02601981e+00 -6.87318623e-01 -3.51288736e-01
7.98624456e-01 1.22987777e-01 -9.90652621e-01 -4.51174766e-01
-8.65101337e-01 7.53964543e-01 7.56694376e-01 -1.53330714e-01
-8.70406806e-01 -3.78714472e-01 -9.46191490e-01 3.73544730e-02
8.95553708e-01 -2.76311308e-01 8.27486217e-01 -1.41632199e+00
-1.14924216e+00 5.25531292e-01 3.24459195e-01 -7.03431189e-01
5.16119063e-01 4.03442187e-03 -8.47312272e-01 2.35427618e-01
5.01819961e-02 3.81568253e-01 1.00176203e+00 -1.01225030e+00
-6.90692425e-01 -5.17914355e-01 1.71934798e-01 2.63453215e-01
-3.21576983e-01 -3.78343046e-01 -2.82104284e-01 -6.76439047e-01
3.72128993e-01 -1.09781480e+00 1.68530077e-01 -4.41590622e-02
-3.66805404e-01 -7.81675987e-03 1.03636634e+00 -5.30591011e-01
1.24271929e+00 -2.14958763e+00 2.80290157e-01 2.70073637e-02
1.68761522e-01 3.90796959e-01 -1.74093381e-01 3.75082970e-01
-5.61340034e-01 -1.70070544e-01 -1.59933716e-01 2.82543242e-01
-1.46773621e-01 6.49134576e-01 -1.23016506e-01 4.50084388e-01
3.00594233e-02 7.29273260e-01 -9.47758019e-01 -2.24078968e-01
3.79942864e-01 1.53022572e-01 -4.57096428e-01 -2.61109807e-02
-2.08501160e-01 2.17624843e-01 -1.05046399e-01 8.32252324e-01
3.55427533e-01 -9.45097879e-02 4.54094037e-02 -2.85659730e-01
2.97948390e-01 -1.45559207e-01 -1.57887435e+00 1.40507853e+00
1.72924861e-01 8.27305973e-01 -6.04402483e-01 -1.29835737e+00
6.33706689e-01 2.81170279e-01 7.76295543e-01 -5.38187265e-01
1.62768885e-01 -4.89107221e-02 2.23049492e-01 -8.10188353e-01
3.26975375e-01 1.03904977e-01 1.44292399e-01 1.69948608e-01
3.00418854e-01 3.65295291e-01 4.86058146e-01 6.78981394e-02
1.56763101e+00 3.14646870e-01 5.21915972e-01 3.45994174e-01
4.47439224e-01 -2.14750860e-02 8.35378349e-01 8.19696486e-01
-4.65097964e-01 6.79782808e-01 5.92868567e-01 -5.39962590e-01
-7.10118949e-01 -1.00416470e+00 -4.93047684e-02 7.43457198e-01
7.91741163e-02 -4.81585741e-01 -4.69928145e-01 -1.14192975e+00
-1.77317336e-02 5.17147183e-01 -9.23632979e-01 -4.05055195e-01
-4.45660204e-01 -1.16622029e-02 5.42977393e-01 8.27346325e-01
5.07740319e-01 -1.26226485e+00 -6.62541509e-01 1.96232110e-01
-4.04266804e-01 -9.45323765e-01 -2.40855455e-01 6.28255382e-02
-8.97100627e-01 -1.83704627e+00 -4.76188242e-01 -5.24675310e-01
6.71205759e-01 1.88564077e-01 1.03562081e+00 -1.23583572e-02
-3.43789667e-01 4.93717015e-01 -8.94515336e-01 -3.74686688e-01
-3.05463254e-01 -7.91280806e-01 1.70539945e-01 5.75684547e-01
4.16014463e-01 -3.50436956e-01 -4.66836810e-01 7.29791462e-01
-7.96504557e-01 -3.06121647e-01 4.69738513e-01 1.16086245e+00
8.56208265e-01 5.22454083e-01 2.57706821e-01 -5.34396887e-01
-5.14448341e-03 -5.66557646e-01 5.15145808e-02 3.46199095e-01
-3.62992398e-02 -1.79409325e-01 5.91982305e-01 -1.03858364e+00
-7.08909273e-01 1.57300428e-01 1.19204387e-01 -1.04034472e+00
-3.89156133e-01 5.17899632e-01 -3.44891578e-01 1.01765223e-01
7.66335487e-01 3.43424380e-01 2.17010989e-03 -2.72029847e-01
1.90633293e-02 3.91106844e-01 5.30201614e-01 -1.07060954e-01
2.54270583e-01 7.49983251e-01 -9.22204852e-02 -8.40865612e-01
-7.65920341e-01 -4.02663916e-01 -8.37304413e-01 -7.51179576e-01
7.57880270e-01 -6.86804771e-01 -4.59062696e-01 4.36684638e-01
-5.37814856e-01 -2.94798553e-01 -8.36203039e-01 7.56491482e-01
-7.61015117e-01 5.96290290e-01 -1.28566027e-01 -4.51236278e-01
3.16212684e-01 -1.05092001e+00 8.86686087e-01 3.79486084e-02
-3.38126183e-01 -6.90218031e-01 -3.40928771e-02 5.89798272e-01
-3.84731978e-01 5.32198310e-01 6.03007495e-01 -1.11290669e+00
-3.93806964e-01 -7.35627890e-01 2.90723234e-01 6.58994913e-01
8.58210325e-02 4.28599492e-02 -6.01498604e-01 -3.58959496e-01
-6.23405017e-02 -3.62565637e-01 9.67983186e-01 2.19764531e-01
1.06305993e+00 -5.12111545e-01 -3.28027815e-01 2.51467049e-01
1.23092747e+00 4.79438424e-01 1.08325577e+00 2.01909199e-01
6.86588526e-01 3.10493708e-01 7.62307644e-01 7.66796231e-01
-8.22796896e-02 6.98814750e-01 4.80029345e-01 3.19532752e-01
-1.34280354e-01 -3.02675486e-01 7.56034195e-01 4.09593999e-01
-9.62288007e-02 -4.27192271e-01 -6.90529346e-01 4.84815031e-01
-2.12817693e+00 -1.53492153e+00 -2.36908615e-01 2.36627936e+00
4.07027662e-01 2.28670254e-01 1.18235551e-01 6.26242816e-01
5.42562783e-01 -2.54367013e-02 -6.93840504e-01 1.09536322e-02
-2.29212850e-01 3.40352915e-02 1.07600406e-01 -3.71129178e-02
-1.21167064e+00 6.20833814e-01 5.54706287e+00 8.97527695e-01
-6.58313930e-01 -1.47589609e-01 1.84592605e-01 -4.66290832e-01
4.25708860e-01 -8.97177774e-03 -5.57539284e-01 4.49195743e-01
6.45910740e-01 1.81349933e-01 2.66260594e-01 9.48407769e-01
-2.18920738e-01 -3.81077886e-01 -1.75120556e+00 1.12573290e+00
6.53986812e-01 -1.23988068e+00 1.34705424e-01 -1.57861188e-02
6.41383588e-01 -1.43171385e-01 -3.52458596e-01 8.24657798e-01
5.38395718e-02 -6.84058905e-01 6.87412500e-01 7.39651322e-01
4.49137241e-01 -4.67970222e-01 7.06149340e-01 4.88683343e-01
-8.83644581e-01 -5.33011138e-01 -4.52509135e-01 -3.43511432e-01
-1.53846145e-01 -1.55334768e-03 -6.95925355e-01 4.35864002e-01
8.58462691e-01 1.43282890e+00 -7.08573103e-01 1.16169250e+00
-7.32799917e-02 4.65281099e-01 -1.24461375e-01 3.23166922e-02
2.46612336e-02 -1.27621397e-01 9.37491000e-01 5.59584975e-01
1.88052818e-01 2.84623921e-01 4.88016874e-01 3.51506263e-01
1.44162074e-01 -1.36333883e-01 -9.48129892e-01 -1.14097863e-01
-4.12824593e-04 6.23183429e-01 -7.58383572e-01 -5.52142620e-01
-6.39850914e-01 1.11232436e+00 2.08303645e-01 1.88179046e-01
-9.80757415e-01 -1.87314838e-01 7.24316597e-01 3.11470591e-02
6.83677137e-01 2.53219873e-01 1.86643466e-01 -1.43404794e+00
1.84966847e-01 -1.09294498e+00 9.55453396e-01 -9.54445660e-01
-1.16442919e+00 3.44267666e-01 7.91384354e-02 -1.94051206e+00
-7.21203014e-02 -6.87741101e-01 -2.49916539e-01 -2.59089451e-02
-6.81388021e-01 -7.29316413e-01 -2.99564421e-01 9.73520815e-01
9.13901150e-01 -4.30376083e-01 9.14857924e-01 2.72866160e-01
-5.32031655e-01 4.67320412e-01 2.58781910e-01 4.60639417e-01
4.48120832e-01 -8.94591630e-01 -1.86515838e-01 8.25471222e-01
8.35345387e-01 -5.00670746e-02 4.88897353e-01 -8.23005736e-01
-1.33320320e+00 -1.01150870e+00 4.71397161e-01 -8.22542310e-01
7.44157791e-01 -1.27444193e-01 -1.06152964e+00 8.85957956e-01
-3.38066459e-01 3.14748555e-01 8.45876038e-01 -6.86051995e-02
-4.97915655e-01 4.68013994e-03 -1.02428818e+00 6.43319547e-01
1.33683610e+00 -3.19942683e-01 -9.68031049e-01 4.32149857e-01
2.42871627e-01 -2.48809293e-01 -8.77015233e-01 6.38270140e-01
7.09020674e-01 -1.18186724e+00 9.41256046e-01 -1.10423374e+00
6.28892183e-01 -2.85469472e-01 -4.57795441e-01 -1.30555725e+00
-3.95884305e-01 2.07584491e-03 -5.87832332e-01 8.23129237e-01
1.73882708e-01 -2.92270154e-01 7.94529259e-01 3.92236948e-01
-1.25684440e-01 -7.76235163e-01 -1.24541712e+00 -1.23364604e+00
-4.24211472e-01 -7.08660066e-01 4.22193944e-01 1.03006899e+00
2.26481676e-01 5.13937846e-02 -4.98896390e-01 -1.02434456e-01
3.35377723e-01 1.78446919e-01 7.21802473e-01 -1.17969167e+00
-6.26223028e-01 -3.67809147e-01 -1.25136554e+00 -8.29045415e-01
2.62358695e-01 -8.01783562e-01 -7.38558080e-03 -1.28786850e+00
2.81067640e-01 6.88888878e-02 -4.57855999e-01 7.89109468e-01
1.63832352e-01 4.02097344e-01 1.32807791e-01 2.83987939e-01
-9.79022741e-01 5.80998480e-01 1.27295446e+00 -4.97343153e-01
-2.29731068e-01 6.81948289e-02 -3.43335599e-01 9.13987756e-01
3.91482979e-01 -4.91418749e-01 -5.12327731e-01 -1.53070077e-01
-9.23428386e-02 4.33664203e-01 5.96574783e-01 -1.52729261e+00
5.93278334e-02 -1.09944373e-01 6.72556102e-01 -4.82505918e-01
5.56829870e-01 -1.32053077e+00 4.56275165e-01 3.93705040e-01
-5.16146123e-01 -3.96972060e-01 -2.80975737e-03 1.20413005e+00
-1.59926251e-01 -2.91160703e-01 4.74031061e-01 -1.94702715e-01
-1.18360555e+00 4.01280105e-01 -3.66178691e-01 5.93334399e-02
1.46677268e+00 -1.04717779e+00 -8.53895918e-02 -2.56359249e-01
-1.24681509e+00 -5.97160496e-02 2.07956389e-01 8.38169932e-01
9.31585908e-01 -1.52683127e+00 -6.18117154e-01 3.74202520e-01
5.45118093e-01 -5.25912762e-01 4.34041500e-01 8.27114224e-01
-1.67780831e-01 -4.99370927e-03 -4.13950354e-01 -4.77460235e-01
-1.68802214e+00 7.15672553e-01 2.42363214e-01 -8.99060443e-02
-1.08731496e+00 8.99763882e-01 -5.14686257e-02 -1.46852031e-01
5.50995708e-01 -2.98974931e-01 -3.85495871e-01 3.46878082e-01
5.69975495e-01 6.26253068e-01 7.11895674e-02 -1.07559836e+00
-4.04707313e-01 1.07781611e-01 3.32179628e-02 3.03479403e-01
1.34645963e+00 3.40674698e-01 2.64574289e-01 7.86597490e-01
1.23963296e+00 -5.42036831e-01 -1.30609035e+00 -3.94323438e-01
-2.68644452e-01 -1.10356331e+00 -1.06291525e-01 -9.36039329e-01
-1.00894177e+00 3.52196217e-01 7.08745658e-01 1.83355343e-02
1.00514185e+00 4.46607471e-02 6.15128994e-01 7.30618656e-01
6.04103029e-01 -1.40943515e+00 4.70453650e-01 4.87104326e-01
1.15401173e+00 -1.42396402e+00 5.23360260e-02 -2.25735605e-01
-9.77272630e-01 1.11006415e+00 7.09173858e-01 -1.05475374e-01
6.38514340e-01 -8.98985788e-02 -3.27038348e-01 -1.69386938e-01
-7.62767434e-01 -2.34584093e-01 4.13070291e-01 7.80123711e-01
-1.64268136e-01 2.54914939e-01 -6.03715749e-03 7.18986452e-01
6.08730800e-02 6.60285130e-02 5.96427381e-01 1.01646674e+00
-3.09115022e-01 -7.01363266e-01 -4.56900924e-01 9.49917376e-01
-2.20898003e-03 3.49861503e-01 -6.58425450e-01 7.20495760e-01
3.40245008e-01 8.77018631e-01 2.09254578e-01 -8.82582307e-01
5.26847124e-01 2.00031102e-01 6.83759451e-01 -5.29007077e-01
-2.53128976e-01 -2.24444702e-01 1.71143115e-01 -9.08598900e-01
-4.83839244e-01 -1.08658242e+00 -1.31304348e+00 4.12258804e-02
-4.89682019e-01 -2.23242968e-01 1.64127022e-01 1.02120996e+00
3.66957694e-01 5.66595733e-01 3.87162745e-01 -6.04967475e-01
-5.81615329e-01 -7.42098689e-01 -7.03929007e-01 1.02168930e+00
9.48737115e-02 -1.12556374e+00 -2.79361874e-01 2.11851895e-01] | [8.464584350585938, 0.89312344789505] |
7a6c417f-66fa-4c17-a9b4-9eecb48f25e2 | a-convolution-recurrent-autoencoder-for | 1904.12413 | null | http://arxiv.org/abs/1904.12413v1 | http://arxiv.org/pdf/1904.12413v1.pdf | A convolution recurrent autoencoder for spatio-temporal missing data imputation | When sensors collect spatio-temporal data in a large geographical area, the
existence of missing data cannot be escaped. Missing data negatively impacts
the performance of data analysis and machine learning algorithms. In this
paper, we study deep autoencoders for missing data imputation in
spatio-temporal problems. We propose a convolution bidirectional-LSTM for
capturing spatial and temporal patterns. Moreover, we analyze an autoencoder's
latent feature representation in spatio-temporal data and illustrate its
performance for missing data imputation. Traffic flow data are used for
evaluation of our models. The result shows that the proposed convolution
recurrent neural network outperforms state-of-the-art methods. | ['Amelia Regan', 'Reza Asadi'] | 2019-04-29 | null | null | null | null | ['multivariate-time-series-imputation'] | ['time-series'] | [-1.71867311e-01 -4.48736668e-01 -5.00237465e-01 -7.12669730e-01
-3.58238965e-01 9.24933404e-02 2.99036264e-01 -2.09863752e-01
-3.51313502e-01 1.13498640e+00 9.13018286e-01 -5.35386920e-01
-4.29322779e-01 -1.02843750e+00 -9.23320770e-01 -6.01094306e-01
-3.47343981e-01 1.57798558e-01 -3.27109665e-01 -1.05227210e-01
-1.57672301e-01 4.28035140e-01 -1.55089760e+00 7.34320760e-01
7.79451311e-01 8.72535110e-01 6.23580627e-02 3.88954878e-01
-3.16608608e-01 1.43716133e+00 -6.56339467e-01 -7.98633918e-02
8.13821405e-02 -1.94171861e-01 -4.86971319e-01 -1.23127475e-01
8.14375281e-02 -7.67598331e-01 -9.96528625e-01 3.99522603e-01
2.42452338e-01 2.24614188e-01 5.03654599e-01 -1.65699446e+00
-9.77460325e-01 6.58660889e-01 -4.46050286e-01 5.59091270e-01
-3.71104211e-01 3.10574919e-02 2.43890271e-01 -6.73393190e-01
7.29617029e-02 8.53374362e-01 1.10510516e+00 3.45929652e-01
-8.98734391e-01 -8.68982136e-01 1.50732815e-01 2.57381141e-01
-1.36250627e+00 -5.35787582e-01 8.78645897e-01 -4.85889852e-01
1.34857309e+00 -1.30111929e-02 3.17889690e-01 1.60491741e+00
4.60144341e-01 7.67361343e-01 4.73061234e-01 2.44357824e-01
1.31778777e-01 -3.41407448e-01 -1.15716517e-01 5.78109594e-03
1.74616933e-01 4.82398063e-01 -4.61888909e-01 -1.08433016e-01
8.70270133e-01 9.90225136e-01 3.50367337e-01 1.46892369e-01
-1.26255381e+00 7.75112987e-01 5.03623128e-01 3.64421904e-01
-8.49079072e-01 5.12883663e-01 4.34845477e-01 4.33276057e-01
7.81472862e-01 -5.30101895e-01 -4.21113223e-01 -2.12188572e-01
-1.28208530e+00 -7.81846493e-02 3.06118488e-01 9.34847593e-01
4.29047346e-01 9.03801024e-01 -2.31182098e-01 6.98952436e-01
-2.01849222e-01 7.65622437e-01 4.31804299e-01 -9.76133168e-01
8.76355767e-01 7.49161720e-01 3.96706551e-01 -1.05595624e+00
-6.44365489e-01 -3.70096028e-01 -1.87100399e+00 1.35650979e-02
2.27839217e-01 -6.12006187e-01 -9.03828800e-01 1.60936368e+00
-1.66735366e-01 7.44832337e-01 1.84286520e-01 9.84124839e-01
6.33565485e-01 7.70961940e-01 3.39973927e-01 -1.36955306e-01
5.76689303e-01 -5.69225252e-01 -1.20700192e+00 -4.70730327e-02
5.39581895e-01 -9.34340507e-02 3.34623426e-01 -4.05705608e-02
-8.59124362e-01 -6.98247969e-01 -4.30370837e-01 -9.78034213e-02
-5.49694896e-01 4.77531075e-01 7.06521571e-01 1.65080443e-01
-6.95640206e-01 4.01856810e-01 -1.15604675e+00 -9.83085185e-02
4.99611139e-01 3.24415952e-01 -4.17980671e-01 2.11843148e-01
-1.29348028e+00 4.90710646e-01 3.29237550e-01 6.57694757e-01
-9.02766585e-01 -7.87392855e-01 -8.81798863e-01 1.69728428e-01
6.83582574e-03 -7.57515132e-01 8.68910730e-01 -6.81964517e-01
-6.24538064e-01 1.37035444e-01 -7.62897849e-01 -8.08828771e-01
5.11641681e-01 -2.13687941e-02 -1.14513755e+00 -5.85909545e-01
3.91143292e-01 4.14759547e-01 6.84773445e-01 -8.68185222e-01
-7.88142145e-01 -3.69453847e-01 -5.01680791e-01 -2.65892833e-01
-2.39690602e-01 -2.06532717e-01 3.07941318e-01 -9.31410730e-01
-1.46395853e-02 -3.59911114e-01 -3.97085488e-01 -3.73588264e-01
-3.16951573e-01 -4.27228510e-02 1.01121080e+00 -9.01469350e-01
1.41903317e+00 -2.19216371e+00 -4.17133838e-01 -1.10088944e-01
1.78455070e-01 -2.67265327e-02 -1.16764545e-01 6.48643136e-01
-1.83876410e-01 1.03155747e-01 -4.15983468e-01 -6.78613245e-01
-1.78344369e-01 7.06933022e-01 -8.03265095e-01 5.06611645e-01
1.69648588e-01 1.09153569e+00 -4.27166343e-01 -5.99883907e-02
6.83251023e-01 8.23065162e-01 -8.75553340e-02 1.00824647e-01
7.98625499e-02 5.41320980e-01 -3.39938849e-01 5.59115112e-01
8.68024766e-01 -3.80827457e-01 -7.62493536e-02 1.52352080e-01
-4.80037540e-01 2.75379509e-01 -9.86160278e-01 1.64633191e+00
-6.23202801e-01 7.99583554e-01 -3.94116521e-01 -9.00336206e-01
9.94988441e-01 4.37657714e-01 8.52870584e-01 -1.24847591e+00
-8.51140171e-02 -1.73805207e-01 -4.46953893e-01 -7.89946973e-01
6.37060225e-01 2.35989079e-01 -1.76705912e-01 6.91758811e-01
-2.54432321e-01 1.17362785e+00 -2.05718905e-01 -1.21806547e-01
1.15642059e+00 -1.84849858e-01 -3.52090150e-01 2.37697557e-01
1.52744815e-01 2.48228073e-01 9.47931051e-01 9.84902918e-01
-2.10375994e-01 4.66223240e-01 1.18526489e-01 -1.37048805e+00
-1.37502015e+00 -1.00682271e+00 -6.45297095e-02 1.02741277e+00
-1.95418999e-01 2.50778615e-01 -2.38587871e-01 -3.73132586e-01
3.95353198e-01 6.36307478e-01 -9.89777565e-01 -5.89475408e-02
-6.96832955e-01 -9.69531476e-01 9.25993204e-01 1.07680547e+00
6.50769234e-01 -1.25131297e+00 -6.70561671e-01 5.10270119e-01
-6.76802576e-01 -9.68904793e-01 8.26377124e-02 3.76855247e-02
-1.04683912e+00 -7.61318326e-01 -3.86223763e-01 -4.79056001e-01
4.07463431e-01 5.26604831e-01 1.07913792e+00 -6.67285472e-02
4.41435091e-02 -2.85066903e-01 -1.39254108e-01 -2.81716973e-01
6.07634224e-02 3.40511590e-01 1.01036564e-01 1.74826652e-01
1.04968464e+00 -9.72807109e-01 -6.61847115e-01 2.87690103e-01
-9.74595249e-01 5.53758023e-03 4.27171230e-01 7.65966594e-01
3.85838181e-01 2.02598393e-01 1.09338343e+00 -2.88798720e-01
6.78389549e-01 -1.15625823e+00 -3.00599128e-01 1.95230562e-02
-5.68220913e-02 1.73781440e-02 6.18133605e-01 -3.07864219e-01
-1.15950739e+00 5.94458319e-02 -1.04968622e-01 -6.69365883e-01
-7.07963347e-01 5.82081199e-01 6.98366165e-02 6.42779469e-01
2.36622840e-01 4.21596587e-01 -2.29058623e-01 -7.07182169e-01
-5.52376546e-02 9.76871133e-01 7.50561476e-01 1.08989932e-01
3.28267694e-01 9.79486406e-01 -3.67506266e-01 -8.31172049e-01
-6.49057448e-01 -5.39877057e-01 -7.94238150e-01 -1.34632841e-01
7.91840136e-01 -1.37393534e+00 -8.67736638e-01 4.87222373e-01
-1.32525885e+00 -4.08761740e-01 -3.72138798e-01 8.41568410e-01
-4.59401160e-01 -3.10531437e-01 -7.47903883e-01 -1.19709480e+00
-2.71388590e-02 -6.39518201e-01 9.87757146e-01 -2.66067207e-01
-1.36036679e-01 -1.15682387e+00 1.50924921e-01 1.25649765e-01
7.83169627e-01 4.79887187e-01 4.58434135e-01 -3.43499482e-01
-6.32956743e-01 -2.75851991e-02 -3.77160579e-01 -1.19607694e-01
2.68231869e-01 -2.94351578e-01 -9.00223911e-01 -1.33219101e-02
-7.56747797e-02 1.35409907e-01 1.30382764e+00 9.58025634e-01
1.63334537e+00 -7.80505896e-01 -3.35706443e-01 8.04926932e-01
1.33037889e+00 2.12236438e-02 9.11363482e-01 2.39548922e-01
6.78856909e-01 6.07449591e-01 5.00404276e-02 8.34021151e-01
7.04509854e-01 9.40846279e-02 6.33994341e-01 -3.37146699e-01
1.70413733e-01 -4.93798971e-01 4.95918170e-02 7.96957433e-01
-2.09046211e-02 -4.58507210e-01 -9.84932542e-01 1.37534976e+00
-2.40009189e+00 -1.65289378e+00 -7.78595686e-01 1.85452056e+00
1.93836346e-01 -3.22153687e-01 3.93492222e-01 3.04462075e-01
6.70295298e-01 2.93814659e-01 -7.17919767e-01 -3.46688956e-01
-5.16977370e-01 -2.11673111e-01 8.86161208e-01 3.02029818e-01
-1.29543424e+00 6.99139357e-01 6.75611353e+00 3.41836512e-01
-9.19479370e-01 3.09898019e-01 5.63376427e-01 -3.42622250e-01
-2.74544090e-01 -4.16306108e-01 -3.70374888e-01 9.05750275e-01
1.40955520e+00 4.31407988e-01 3.15798342e-01 4.74714518e-01
9.57105756e-01 2.77592838e-02 -6.59031928e-01 9.40584600e-01
-2.28559420e-01 -1.62243211e+00 -2.06554253e-02 2.63692766e-01
8.20665538e-01 6.37255907e-01 1.70826629e-01 1.70529947e-01
6.77441120e-01 -1.48530138e+00 9.85021144e-02 1.04816914e+00
5.73514700e-01 -1.03670406e+00 9.98655736e-01 6.37481093e-01
-8.79464805e-01 -5.01017630e-01 -5.12618423e-01 -5.85200787e-01
6.40430510e-01 8.69809985e-01 -2.32376382e-01 3.77283394e-01
9.89837170e-01 1.02987504e+00 -1.93394661e-01 1.07676363e+00
1.75178364e-01 9.22160149e-01 -2.96230078e-01 4.15549934e-01
4.03572083e-01 -2.53446668e-01 2.33490944e-01 1.19519567e+00
5.60504675e-01 1.86785068e-02 -5.82863577e-02 1.18708479e+00
-1.97556347e-01 -7.15519130e-01 -1.09895158e+00 3.10728878e-01
5.32070577e-01 4.36442107e-01 3.50688584e-02 -2.39781737e-01
-4.99578118e-01 8.35978150e-01 1.11065112e-01 8.53195429e-01
-9.90617216e-01 -1.44957289e-01 9.62170243e-01 7.56256953e-02
2.59290993e-01 -4.54187393e-01 -7.84523487e-01 -1.26980519e+00
1.59039408e-01 -1.75851732e-01 4.45781350e-01 -1.03397572e+00
-1.60075498e+00 4.20345366e-01 -1.62242338e-01 -1.33965027e+00
-5.84801137e-01 4.44553345e-02 -5.96398413e-01 8.42106223e-01
-1.56302762e+00 -1.32750583e+00 -3.99617285e-01 1.09364629e+00
3.52708608e-01 -3.35608333e-01 6.75407290e-01 8.14524055e-01
-7.25580633e-01 1.81507140e-01 5.99113941e-01 8.05246651e-01
2.22171992e-01 -7.27185547e-01 9.83987391e-01 1.14711487e+00
7.65793249e-02 4.12715256e-01 4.51148957e-01 -9.26809967e-01
-1.15935659e+00 -1.84331226e+00 1.55131912e+00 -5.47071040e-01
5.44772923e-01 -2.60393679e-01 -1.16432536e+00 1.17587924e+00
1.07198082e-01 1.58722267e-01 7.52910912e-01 1.08478330e-01
-1.64283812e-01 -3.13742429e-01 -1.07168865e+00 1.20191038e-01
1.06326103e+00 -6.25108838e-01 -5.22229910e-01 1.86957465e-03
7.49953389e-01 5.08762226e-02 -7.43298709e-01 4.10798043e-01
4.97080266e-01 -9.18626487e-01 9.12167430e-01 -1.13844502e+00
5.01734078e-01 -2.99413204e-01 -1.72127783e-01 -1.32473588e+00
-5.46881378e-01 -1.14414141e-01 -4.52850163e-01 8.89075458e-01
3.86108339e-01 -2.97318131e-01 1.06014061e+00 6.87549531e-01
-1.65087357e-01 3.58407348e-02 -1.28365183e+00 -6.83109879e-01
1.87500551e-01 -9.79715884e-01 1.25047994e+00 1.29848647e+00
-3.55692297e-01 -6.19713403e-02 -1.30728948e+00 4.01087224e-01
6.98849738e-01 2.00445697e-01 6.68686509e-01 -1.21843696e+00
3.73007357e-01 5.65400347e-02 -3.02162528e-01 -7.89462507e-01
3.31961870e-01 -4.68590826e-01 -6.30000591e-01 -1.61868310e+00
5.42010330e-02 -2.41671160e-01 -6.17516160e-01 9.09491301e-01
5.37953079e-01 1.65045813e-01 -3.21935803e-01 1.10625021e-01
-4.93136674e-01 8.73753011e-01 5.59442937e-01 -4.36426044e-01
-5.03267407e-01 1.42644793e-01 -2.62361467e-01 2.60915697e-01
1.39476752e+00 -7.79997170e-01 -1.90997154e-01 -1.21533132e+00
3.17305773e-01 4.36341137e-01 9.43901300e-01 -1.07118034e+00
5.01514494e-01 -4.78179365e-01 9.24132347e-01 -1.26507866e+00
2.17348382e-01 -1.39309895e+00 4.45417821e-01 2.51779377e-01
-3.93419743e-01 5.62804937e-01 2.79103637e-01 7.07444847e-01
-4.00357097e-01 4.93237883e-01 -2.91799843e-01 6.00117929e-02
-6.90747261e-01 5.10841846e-01 -8.19067478e-01 -1.94088817e-01
6.54542744e-01 -4.93349701e-01 -4.58571762e-01 -5.11812747e-01
-8.85899186e-01 6.13982201e-01 -8.65512043e-02 7.58778811e-01
9.23603594e-01 -1.76937401e+00 -9.22918737e-01 5.91415823e-01
1.56861302e-02 -2.95947731e-01 9.61821795e-01 8.23491395e-01
6.24013096e-02 5.08045793e-01 -4.31900561e-01 -5.22398412e-01
-4.76057947e-01 5.91745555e-01 4.11671788e-01 1.74927190e-01
-7.35609949e-01 -8.51016641e-02 -4.19638872e-01 -6.55404329e-01
3.33955646e-01 -5.74948072e-01 -4.36341166e-02 -9.57258493e-02
5.94530404e-01 7.85969079e-01 2.00335935e-01 -7.06514060e-01
-2.28570759e-01 -1.13671012e-01 5.62281311e-01 4.89464961e-02
1.88705862e+00 -5.37453592e-01 8.82659778e-02 8.19224179e-01
1.05456448e+00 -5.27412534e-01 -1.32881308e+00 -1.75876424e-01
-9.05820355e-02 -4.93100822e-01 2.46356815e-01 -8.97616029e-01
-1.28925252e+00 1.16985977e+00 6.58554435e-01 1.70058832e-01
9.85055625e-01 -4.13049549e-01 1.15723991e+00 2.80851603e-01
1.67541027e-01 -1.10980427e+00 -6.59972847e-01 6.58285320e-01
5.84148109e-01 -1.54857445e+00 -6.17164135e-01 3.27222347e-01
-6.48771703e-01 7.43096888e-01 5.19481182e-01 -2.24806473e-01
8.41632664e-01 5.20344794e-01 5.55684650e-03 -8.16854164e-02
-1.05446398e+00 -4.24547285e-01 2.70187166e-02 6.99044466e-01
2.41893008e-01 2.48012945e-01 4.85338956e-01 6.32311761e-01
-1.30625367e-01 4.99606729e-01 5.70859671e-01 6.68611109e-01
-1.90215886e-01 -6.49413586e-01 -2.27831095e-01 5.97982585e-01
-4.48503017e-01 2.01294869e-02 -1.53907865e-01 8.20526600e-01
2.79126674e-01 1.29048550e+00 8.37754309e-01 -3.52463305e-01
2.06946924e-01 1.20273940e-02 -4.45628464e-01 1.95436791e-01
-5.20172119e-01 -2.09518567e-01 -2.17489675e-01 -5.65886855e-01
-7.23640084e-01 -4.04876411e-01 -9.04865265e-01 -1.09867060e+00
3.32492709e-01 1.22868478e-01 5.62115431e-01 9.25741017e-01
8.98080230e-01 7.46252120e-01 5.16199231e-01 -2.52042741e-01
2.18126327e-02 -9.84707355e-01 -3.93052787e-01 3.69048685e-01
1.19831157e+00 -4.02510911e-01 -1.46492019e-01 1.65515617e-02] | [6.7580060958862305, 2.662545919418335] |
bc4aa5ed-1aa7-416b-b95c-2db01e4eb88d | contextual-mask-auto-encoder-for-dense | 2208.07670 | null | https://arxiv.org/abs/2208.07670v3 | https://arxiv.org/pdf/2208.07670v3.pdf | ConTextual Masked Auto-Encoder for Dense Passage Retrieval | Dense passage retrieval aims to retrieve the relevant passages of a query from a large corpus based on dense representations (i.e., vectors) of the query and the passages. Recent studies have explored improving pre-trained language models to boost dense retrieval performance. This paper proposes CoT-MAE (ConTextual Masked Auto-Encoder), a simple yet effective generative pre-training method for dense passage retrieval. CoT-MAE employs an asymmetric encoder-decoder architecture that learns to compress the sentence semantics into a dense vector through self-supervised and context-supervised masked auto-encoding. Precisely, self-supervised masked auto-encoding learns to model the semantics of the tokens inside a text span, and context-supervised masked auto-encoding learns to model the semantical correlation between the text spans. We conduct experiments on large-scale passage retrieval benchmarks and show considerable improvements over strong baselines, demonstrating the high efficiency of CoT-MAE. Our code is available at https://github.com/caskcsg/ir/tree/main/cotmae. | ['Songlin Hu', 'Zhongyuan Wang', 'Zijia Lin', 'Meng Lin', 'Guangyuan Ma', 'Xing Wu'] | 2022-08-16 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [-1.60383880e-01 -2.52294868e-01 -3.76454264e-01 -3.01829249e-01
-1.70617819e+00 -5.31847715e-01 8.20817769e-01 3.43612939e-01
-3.45565796e-01 6.40619397e-01 1.09338820e+00 3.51643488e-02
1.46157995e-01 -9.10589516e-01 -9.20625746e-01 -4.52364147e-01
-6.36456162e-02 5.72650611e-01 -1.17847547e-01 -4.43781912e-01
1.85366273e-01 -1.34485275e-01 -1.19720995e+00 7.64941573e-01
6.45339310e-01 6.08019292e-01 3.09483200e-01 9.58763421e-01
-2.08612218e-01 9.48460460e-01 -6.87602997e-01 -2.44780540e-01
7.79973119e-02 -8.46862257e-01 -1.09865415e+00 -3.68296325e-01
3.09203833e-01 -5.71821570e-01 -1.09566772e+00 5.10367453e-01
4.66662169e-01 3.45840126e-01 7.69039154e-01 -7.04429507e-01
-1.22248018e+00 8.16479504e-01 -1.53530717e-01 5.72608173e-01
6.54048800e-01 -1.49147600e-01 1.60732424e+00 -1.02465987e+00
6.10170543e-01 1.01787674e+00 2.94923306e-01 4.84132558e-01
-1.04356730e+00 -2.45840207e-01 -1.62830085e-01 1.91344261e-01
-1.65650129e+00 -4.00400192e-01 4.50245798e-01 -7.18513802e-02
1.49122036e+00 3.17790359e-01 4.32625830e-01 1.28984809e+00
1.25679344e-01 1.20411611e+00 3.63950163e-01 -5.10200441e-01
1.25488788e-01 -3.89688201e-02 3.33482563e-01 6.03127956e-01
-4.52964986e-03 -4.72806673e-03 -4.35633302e-01 -4.29280818e-01
5.06361663e-01 3.58102769e-01 -3.40740055e-01 1.26533806e-01
-9.34060574e-01 1.04305279e+00 6.86065257e-01 4.30644691e-01
-4.76631045e-01 1.70764595e-01 5.82476556e-01 5.44661224e-01
4.92456943e-01 5.56051552e-01 -2.32122198e-01 -1.74588069e-01
-1.24247527e+00 4.52673316e-01 8.21204126e-01 1.35768378e+00
7.96772778e-01 -3.31128180e-01 -8.67612660e-01 9.95394945e-01
2.19149083e-01 5.95776021e-01 6.49963915e-01 -4.42444563e-01
6.10553205e-01 3.69550496e-01 -2.94101909e-02 -6.07225001e-01
1.95684791e-01 -5.16289711e-01 -6.87738299e-01 -7.77461469e-01
-2.73585051e-01 1.54566299e-02 -8.76303017e-01 1.65931141e+00
-2.36623496e-01 6.11069761e-02 2.76064366e-01 9.11430240e-01
9.31525707e-01 1.29615128e+00 2.08999723e-01 1.44013986e-01
1.18789852e+00 -1.42583776e+00 -5.03425539e-01 -3.30071241e-01
8.39650571e-01 -8.92443538e-01 1.24668372e+00 -4.13402826e-01
-1.38181317e+00 -3.58968288e-01 -9.03611839e-01 -8.61501753e-01
-3.23461711e-01 2.40915611e-01 3.67789358e-01 -1.55346245e-01
-1.01858079e+00 2.33231544e-01 -7.43527949e-01 -1.63412005e-01
3.09463769e-01 -9.15869772e-02 -1.53165683e-01 -3.91227186e-01
-1.47727132e+00 5.14806747e-01 5.00290036e-01 -1.84318855e-01
-1.08987296e+00 -7.62563407e-01 -9.32462573e-01 6.33948684e-01
-6.93981275e-02 -1.09789026e+00 1.39826691e+00 -5.70716202e-01
-1.06095016e+00 7.71239102e-01 -5.08824587e-01 -7.21529067e-01
-1.53896526e-01 -6.03047013e-01 -3.19637418e-01 5.92992425e-01
2.91866392e-01 7.55064368e-01 6.49763823e-01 -9.07142580e-01
-2.18450382e-01 1.01820052e-01 3.89770903e-02 3.04404497e-01
-2.49260008e-01 -2.89831068e-02 -9.24588621e-01 -8.89080465e-01
-4.67752337e-01 -7.73743510e-01 -1.27516016e-01 -4.69380558e-01
-4.73085701e-01 -4.27461028e-01 4.37557101e-01 -7.81721711e-01
1.63410366e+00 -2.20774603e+00 4.37812991e-02 -1.43649394e-03
1.14734754e-01 1.99945375e-01 -6.31027341e-01 1.20356095e+00
2.98403084e-01 -1.44055148e-03 -2.06041127e-01 -5.84054649e-01
2.49593914e-01 1.20580576e-01 -9.25130844e-01 1.57064706e-01
1.85625494e-01 1.41638553e+00 -1.07450163e+00 -5.17210484e-01
-6.25904426e-02 6.44226134e-01 -6.91944063e-01 6.31652355e-01
-5.81267774e-01 -1.34564430e-01 -7.22675681e-01 3.43946069e-01
1.09300941e-01 -6.05821013e-01 -3.59674275e-01 2.10137576e-01
4.54398125e-01 1.05899465e+00 -1.52534574e-01 2.27548313e+00
-6.45778060e-01 7.81075120e-01 -4.10890073e-01 -5.90766191e-01
7.11756766e-01 4.66253668e-01 2.46526688e-01 -1.09926641e+00
-6.50522672e-03 1.34428486e-01 -7.67897487e-01 -4.67063874e-01
1.14959311e+00 -1.56246321e-02 -3.95642966e-01 8.16747546e-01
2.97449797e-01 -4.09377739e-02 3.47007573e-01 9.87058461e-01
1.35319245e+00 -5.80873201e-03 6.67493269e-02 2.97991578e-02
2.84115911e-01 1.57436639e-01 8.12345743e-02 1.00923407e+00
4.51974601e-01 9.28533077e-01 1.09817997e-01 2.12197821e-03
-1.33965111e+00 -1.16107237e+00 1.93365797e-01 1.10938954e+00
8.28716457e-02 -9.34798479e-01 -4.72679883e-01 -4.65490162e-01
-2.57164035e-02 7.92285264e-01 -5.11922717e-01 -6.31893337e-01
-6.00202739e-01 -2.87819088e-01 6.50615990e-01 7.16848850e-01
1.95589334e-01 -1.16723764e+00 -9.75800902e-02 2.71535486e-01
-4.96890068e-01 -9.96963859e-01 -1.01231670e+00 -9.86933559e-02
-7.30867445e-01 -6.61869287e-01 -8.83761406e-01 -1.09823418e+00
5.76633215e-01 5.50423920e-01 1.63359964e+00 4.93447304e-01
-4.91092116e-01 5.71232438e-01 -8.44887674e-01 8.86125863e-02
-2.23524168e-01 5.06787539e-01 -5.52195489e-01 -5.03076196e-01
7.69969046e-01 -4.67257917e-01 -9.84559178e-01 -7.06428215e-02
-1.21587384e+00 -2.00092793e-01 6.08406305e-01 9.77170050e-01
8.23589861e-01 -4.92516875e-01 7.37730324e-01 -7.19221175e-01
9.67468023e-01 -9.22914982e-01 -2.03972071e-01 4.42478269e-01
-4.01556373e-01 2.46093765e-01 4.67753619e-01 -2.32562557e-01
-7.39513159e-01 -6.27286375e-01 -4.39932227e-01 -5.81706941e-01
4.66858447e-02 6.33495331e-01 2.88481742e-01 6.58585787e-01
6.13364577e-01 6.52966201e-01 -3.32196385e-01 -6.07286215e-01
6.70338035e-01 8.19464982e-01 4.57282454e-01 -6.33845866e-01
6.69052780e-01 2.18326733e-01 -5.50913155e-01 -5.83728194e-01
-1.14575863e+00 -9.72473860e-01 -1.82240218e-01 2.82517850e-01
7.37797618e-01 -1.40298438e+00 9.50984880e-02 -1.57139525e-01
-1.00361645e+00 -5.18089175e-01 -6.61801398e-01 4.42696810e-01
-5.20549238e-01 2.50131071e-01 -1.10451543e+00 -2.96015561e-01
-1.08547592e+00 -5.96702337e-01 1.45633256e+00 2.54509300e-01
-3.11624497e-01 -1.06506324e+00 5.23809731e-01 3.84510905e-01
6.20288789e-01 -3.42090935e-01 9.03453231e-01 -9.00961697e-01
-7.75561929e-01 -6.13750517e-01 -1.54890180e-01 2.95923680e-01
-1.22583836e-01 -3.05157393e-01 -7.66773820e-01 -6.43842578e-01
-3.73653144e-01 -7.74227440e-01 1.24522305e+00 1.69314474e-01
1.09610915e+00 -6.28189504e-01 -2.57879078e-01 3.88180345e-01
1.62887824e+00 -1.33419424e-01 8.99415076e-01 2.40138724e-01
3.86000574e-01 -5.57152927e-02 3.74158025e-01 4.34796125e-01
4.97414500e-01 4.88103360e-01 -1.26940206e-01 4.53756787e-02
-3.19470257e-01 -9.04494524e-01 4.19769675e-01 1.29976809e+00
5.35096347e-01 -6.01024866e-01 -5.03106236e-01 9.94914114e-01
-1.62922966e+00 -1.16361511e+00 3.98318529e-01 1.94561315e+00
1.10075855e+00 -1.31263748e-01 -4.25427258e-02 -3.79004687e-01
3.35488737e-01 3.99183989e-01 -2.07618117e-01 -3.59980077e-01
-1.58865094e-01 5.48355997e-01 7.73741752e-02 7.53138483e-01
-9.30890620e-01 1.01085234e+00 5.84782934e+00 1.00531077e+00
-8.14595163e-01 1.74700528e-01 3.38616610e-01 -6.01780474e-01
-7.82572508e-01 -5.68033755e-02 -9.19501543e-01 3.79978746e-01
1.14644742e+00 -5.55057585e-01 2.37935185e-01 7.17585385e-01
-7.89924487e-02 1.95709929e-01 -1.19671524e+00 6.19523823e-01
4.44236368e-01 -1.62267220e+00 4.02733386e-01 -1.32171944e-01
9.21965837e-01 4.57247555e-01 7.72173032e-02 8.96875620e-01
1.29591733e-01 -8.23810220e-01 4.34647381e-01 6.03352070e-01
7.66772687e-01 -4.50113297e-01 6.67780995e-01 2.39221364e-01
-1.03684938e+00 1.90349191e-01 -6.99633896e-01 1.58442900e-01
2.49222815e-01 3.48481059e-01 -7.71833479e-01 5.55982590e-01
4.83542860e-01 9.25629377e-01 -5.44561744e-01 1.12291718e+00
-3.55261892e-01 7.92440116e-01 -2.61920363e-01 -1.92082241e-01
6.05580807e-01 -3.40478867e-02 5.82889676e-01 1.72688484e+00
2.54715353e-01 2.30381563e-01 -3.82126751e-03 1.04823935e+00
-5.55466115e-01 1.61395997e-01 -5.65345347e-01 -4.97460753e-01
5.48616171e-01 8.28610837e-01 -8.09829310e-02 -6.62533224e-01
-5.66396415e-01 1.20400226e+00 4.55967009e-01 7.46875346e-01
-5.37507236e-01 -6.08179390e-01 6.01394296e-01 6.26030937e-02
6.97784722e-01 -8.07206184e-02 7.07009584e-02 -1.55728710e+00
2.04442531e-01 -6.73457146e-01 6.20204687e-01 -9.30043459e-01
-1.45070076e+00 7.27862597e-01 -5.58802858e-02 -1.26177990e+00
-6.21688902e-01 1.05225801e-01 -7.09263384e-01 1.02639782e+00
-1.59841311e+00 -9.15971696e-01 3.30500528e-02 7.22460508e-01
9.97452676e-01 -1.02469586e-01 1.25918031e+00 3.43164265e-01
-2.48034552e-01 8.04902852e-01 4.89038706e-01 5.71535051e-01
5.78629792e-01 -1.08573687e+00 5.84974289e-01 7.59581089e-01
6.73844457e-01 9.79259729e-01 3.04640740e-01 -4.94357973e-01
-1.50483632e+00 -1.11795068e+00 1.44976830e+00 -4.75317180e-01
5.77580869e-01 -5.22153139e-01 -1.21413028e+00 8.89991164e-01
7.09668398e-01 -1.58032060e-01 9.81891811e-01 1.19162381e-01
-6.04994118e-01 9.79898050e-02 -5.75314283e-01 5.30685067e-01
6.11662149e-01 -1.18759811e+00 -1.19455254e+00 6.02597773e-01
1.16977084e+00 -3.59527051e-01 -8.66428077e-01 -3.38815562e-02
2.95295477e-01 -4.90163088e-01 1.23596275e+00 -7.37176895e-01
9.10006881e-01 6.72560036e-02 -1.20844759e-01 -1.19948065e+00
-3.09196830e-01 -4.04721022e-01 -5.44791460e-01 1.01912308e+00
6.18592441e-01 -1.42636061e-01 5.54795027e-01 3.50667715e-01
-2.89144754e-01 -1.02847850e+00 -5.87705672e-01 -6.14283979e-01
4.06023741e-01 -1.28617734e-02 6.55277789e-01 5.17278194e-01
2.43908495e-01 6.38987720e-01 -2.55702645e-01 6.39810413e-02
2.97505885e-01 5.57863533e-01 5.90657473e-01 -3.83036822e-01
-4.17624891e-01 -3.93455446e-01 -1.69778109e-01 -1.78990376e+00
2.14382768e-01 -1.17135799e+00 1.49487838e-01 -1.96626878e+00
5.20164609e-01 -1.87558532e-01 -3.61516148e-01 4.01223421e-01
-3.61857742e-01 9.48967263e-02 -4.24171686e-02 3.51750940e-01
-1.16907370e+00 1.05145395e+00 1.09565377e+00 -3.07228625e-01
-4.13334332e-02 -1.48199201e-01 -6.88323855e-01 -1.42080784e-01
7.59575605e-01 -6.31521344e-01 -5.41003406e-01 -9.45388794e-01
1.98727161e-01 2.47209042e-01 3.15660864e-01 -7.00092494e-01
4.85651523e-01 3.87223840e-01 2.88357228e-01 -8.16000640e-01
4.39511836e-01 -4.38573390e-01 -3.48231465e-01 7.98548833e-02
-1.04206836e+00 1.71488255e-01 1.93634063e-01 5.90340376e-01
-8.12364101e-01 -2.83247262e-01 9.95392278e-02 -2.61450380e-01
-4.93006408e-01 5.38254023e-01 -4.04947460e-01 5.86153269e-01
3.10070068e-01 4.06685472e-01 -4.90116239e-01 -7.05511391e-01
-5.06669283e-01 5.25903702e-01 2.25089505e-01 6.30722821e-01
9.11024272e-01 -1.37796986e+00 -9.44251835e-01 2.24673465e-01
3.33977699e-01 2.85446979e-02 4.00898099e-01 2.43505746e-01
-4.08036649e-01 1.01654184e+00 2.68124819e-01 -1.86823308e-01
-1.02217746e+00 6.45080447e-01 1.22609101e-01 -5.29438198e-01
-9.77713287e-01 1.04371440e+00 1.05569921e-02 2.23430414e-02
3.22672755e-01 -2.19117299e-01 1.54006094e-01 -3.76333833e-01
8.60115409e-01 6.32652789e-02 -9.06124040e-02 -4.01861906e-01
1.98788121e-02 2.45954975e-01 -5.57932138e-01 -1.76729843e-01
1.27815282e+00 -3.72216225e-01 -1.80189162e-01 1.90944850e-01
2.00345588e+00 -1.12428844e-01 -8.36811543e-01 -8.08351994e-01
-4.14667875e-02 -3.25951248e-01 2.34404370e-01 -7.31123149e-01
-8.02322865e-01 9.30748582e-01 -7.23295137e-02 -1.28546610e-01
1.21516323e+00 3.62964392e-01 1.40291560e+00 7.30712175e-01
2.80463338e-01 -7.60014296e-01 1.58560082e-01 6.57642841e-01
1.00970948e+00 -1.02358305e+00 -1.16042696e-01 5.40589727e-02
-6.15855396e-01 8.05115521e-01 3.23452890e-01 -5.18715978e-01
6.56897604e-01 -9.62581951e-03 -2.14931462e-02 -3.48710299e-01
-1.27100861e+00 -2.87250638e-01 5.52334964e-01 -6.77271634e-02
5.19948840e-01 -5.84176481e-02 -1.58596098e-01 5.15164256e-01
-3.01080793e-01 3.45159099e-02 8.71857181e-02 1.10754895e+00
-3.93687159e-01 -1.05852962e+00 1.01724297e-01 5.95778704e-01
-5.15212119e-01 -6.48941755e-01 -3.16342205e-01 3.60227585e-01
-6.37139380e-01 8.88118744e-01 2.04238147e-01 -2.17825547e-01
1.40425682e-01 2.42678314e-01 3.63969117e-01 -8.79030406e-01
-7.31976509e-01 2.42830753e-01 -1.98089629e-02 -4.68555003e-01
-2.26977617e-02 -4.18945312e-01 -1.23846924e+00 -1.00726746e-01
-4.65892963e-02 9.44755018e-01 2.52731383e-01 7.56329536e-01
6.30372941e-01 2.96135932e-01 7.88696289e-01 -2.62106419e-01
-5.39346337e-01 -1.18839514e+00 -3.93452317e-01 5.48286974e-01
5.86840510e-01 3.54252420e-02 -4.19828415e-01 -1.47091076e-02] | [11.42757797241211, 7.7344651222229] |
6977f633-ff73-4abf-884a-2f212cf0a080 | multi-domain-pose-network-for-multi-person | 1810.08338 | null | http://arxiv.org/abs/1810.08338v1 | http://arxiv.org/pdf/1810.08338v1.pdf | Multi-Domain Pose Network for Multi-Person Pose Estimation and Tracking | Multi-person human pose estimation and tracking in the wild is important and
challenging. For training a powerful model, large-scale training data are
crucial. While there are several datasets for human pose estimation, the best
practice for training on multi-dataset has not been investigated. In this
paper, we present a simple network called Multi-Domain Pose Network (MDPN) to
address this problem. By treating the task as multi-domain learning, our
methods can learn a better representation for pose prediction. Together with
prediction heads fine-tuning and multi-branch combination, it shows significant
improvement over baselines and achieves the best performance on PoseTrack ECCV
2018 Challenge without additional datasets other than MPII and COCO. | ['Yongchen Lu', 'Linfu Wen', 'Tang Tang', 'Hengkai Guo', 'Guozhong Luo', 'Riwei Chen'] | 2018-10-19 | null | null | null | null | ['multi-person-pose-estimation-and-tracking'] | ['computer-vision'] | [-2.67468482e-01 -3.79954726e-02 -2.25919560e-01 -3.83139729e-01
-1.05300152e+00 -4.77698326e-01 2.32181370e-01 -4.05281037e-01
-6.55984342e-01 8.64213765e-01 4.33412075e-01 3.63295138e-01
3.38531137e-01 -2.53491372e-01 -8.81591558e-01 -3.29462796e-01
4.21105623e-02 9.27459955e-01 5.31795204e-01 -3.09086978e-01
-6.35202587e-01 1.87498167e-01 -1.11554241e+00 2.00454578e-01
4.40138340e-01 9.32137489e-01 -1.57254070e-01 6.61297083e-01
6.56556368e-01 3.11009198e-01 -7.65161395e-01 -6.57230437e-01
5.09291053e-01 -1.25446588e-01 -7.74897814e-01 -2.17635319e-01
1.06499696e+00 -5.33528030e-01 -4.47333813e-01 6.17760122e-01
1.15982652e+00 9.59597304e-02 4.95201856e-01 -1.36866570e+00
6.70549572e-02 3.42396408e-01 -7.49952972e-01 1.03460565e-01
4.45841193e-01 3.38011593e-01 7.57645845e-01 -6.25062108e-01
6.37014925e-01 1.65571320e+00 1.18108559e+00 9.28190231e-01
-9.99616861e-01 -7.43898809e-01 1.49154723e-01 1.47840127e-01
-1.30908716e+00 -2.21301451e-01 5.76627612e-01 -2.99497724e-01
7.44270563e-01 1.10907048e-01 7.36142576e-01 1.85538423e+00
2.19998248e-02 1.20743942e+00 9.18277621e-01 7.12575158e-04
-3.18598092e-01 -9.05798823e-02 -4.19632047e-02 4.72596705e-01
5.05320251e-01 2.42266014e-01 -6.05262041e-01 1.98033303e-02
6.63802445e-01 -1.64118052e-01 -1.07131213e-01 -4.45950300e-01
-1.08454931e+00 5.95363915e-01 7.10494459e-01 -1.43616870e-01
-1.05966210e-01 4.97451723e-01 6.62127495e-01 -7.98184574e-02
4.41874623e-01 4.20460492e-01 -9.38092947e-01 -2.81126231e-01
-9.07686651e-01 9.35160100e-01 7.32071877e-01 8.71872365e-01
1.55094668e-01 -3.06936860e-01 -5.46439886e-01 6.75115049e-01
4.22621995e-01 7.64021873e-01 9.15285498e-02 -6.66673958e-01
8.38588417e-01 4.11097705e-01 2.83378601e-01 -9.23121750e-01
-9.43638921e-01 -6.64109945e-01 -6.01147771e-01 -6.34245574e-02
9.29547191e-01 -6.20309114e-01 -8.76404047e-01 1.92782080e+00
6.06451809e-01 1.99319303e-01 -4.05474186e-01 1.20584619e+00
1.07356155e+00 4.66141582e-01 3.20110232e-01 2.84219563e-01
1.50301218e+00 -1.38179946e+00 -5.16345561e-01 -7.90261567e-01
5.54999590e-01 -4.44991112e-01 6.31023288e-01 3.36316288e-01
-8.38873327e-01 -7.76889265e-01 -8.65620911e-01 -3.28351170e-01
-1.50769278e-01 5.15261948e-01 7.60785878e-01 6.94231749e-01
-7.44390786e-01 4.02845651e-01 -8.67666245e-01 -4.59638268e-01
5.67829072e-01 4.08262670e-01 -5.35657465e-01 -8.20177123e-02
-1.20102644e+00 1.15975463e+00 2.98622042e-01 3.69939476e-01
-7.55866051e-01 -6.97092116e-01 -8.76998901e-01 -4.02920783e-01
5.95709741e-01 -9.18659449e-01 1.32602072e+00 -3.02059174e-01
-1.23488784e+00 7.99490392e-01 -2.16171052e-02 -7.40668714e-01
1.12334609e+00 -9.14173365e-01 -2.55320966e-01 -6.14432693e-02
2.03300808e-02 1.05591214e+00 7.50164270e-01 -1.02103698e+00
-4.77875501e-01 -6.60442472e-01 -1.73134923e-01 2.04226196e-01
-1.93420440e-01 1.27465785e-01 -1.02700317e+00 -5.72419167e-01
-2.85793215e-01 -1.30937207e+00 -3.58462751e-01 1.74738035e-01
-5.03748238e-01 -3.39674592e-01 8.36822510e-01 -1.13942039e+00
1.19462585e+00 -1.80624914e+00 4.09505129e-01 -1.49133995e-01
2.90077597e-01 4.59570736e-01 -1.58510193e-01 1.28705397e-01
2.01652333e-01 -3.64051342e-01 2.10211813e-01 -7.58489370e-01
1.92435786e-01 -1.28829658e-01 1.53938904e-01 6.55186653e-01
1.32877342e-02 1.37388110e+00 -6.46364629e-01 -7.69340694e-01
1.23979956e-01 7.32967377e-01 -8.10113072e-01 2.25647479e-01
-3.86860102e-01 8.95389318e-01 -4.39987183e-01 6.21792018e-01
7.41756976e-01 -3.90191138e-01 1.64937660e-01 -5.05928755e-01
3.14879924e-01 -4.73631062e-02 -1.34692991e+00 1.90000510e+00
-1.36491522e-01 5.03472030e-01 -8.06245953e-02 -4.37663883e-01
5.79890370e-01 1.71440363e-01 6.02447629e-01 -5.12960553e-01
3.81022602e-01 -5.84091283e-02 -3.79679948e-02 -3.39617133e-01
5.00920951e-01 1.40698567e-01 -4.57799315e-01 3.10487412e-02
2.45247811e-01 3.16904545e-01 7.22984374e-02 -4.23924327e-02
9.46403921e-01 6.31375611e-01 1.21600360e-01 1.07468747e-01
2.83578038e-01 -1.90847978e-01 7.68152714e-01 4.39000756e-01
-8.28086138e-01 7.35493124e-01 2.13850766e-01 -5.97258151e-01
-8.82311463e-01 -8.12083244e-01 -1.21333487e-02 1.56591582e+00
1.63003236e-01 -7.01810837e-01 -7.41818964e-01 -7.84863234e-01
3.19886357e-01 -1.07493937e-01 -7.15622306e-01 1.18973218e-01
-9.50964630e-01 -8.22687745e-01 8.45365524e-01 9.17416036e-01
5.90755045e-01 -9.20448601e-01 -5.28247058e-01 4.85721864e-02
-5.59748650e-01 -1.44322097e+00 -8.16925049e-01 -1.17359757e-01
-5.56936800e-01 -1.14272296e+00 -1.07212889e+00 -5.51658332e-01
9.63072106e-02 -1.50391236e-01 1.22809768e+00 -3.75081182e-01
-4.30519015e-01 2.40440190e-01 -8.01124871e-02 -5.68930864e-01
2.23987266e-01 4.44190174e-01 2.45999649e-01 -3.12467664e-01
3.72832358e-01 -2.83423692e-01 -7.45393634e-01 4.07709539e-01
-3.42158135e-03 -1.50339723e-01 7.20308125e-01 5.36696732e-01
5.17687559e-01 -5.10000408e-01 4.10728693e-01 -6.99841321e-01
3.40159148e-01 -8.83413851e-03 -4.71348345e-01 2.88143784e-01
-4.49652076e-02 -2.24772617e-02 2.25932062e-01 -4.10159707e-01
-1.00612652e+00 4.71508861e-01 -4.14463013e-01 -4.40906376e-01
-1.97758809e-01 -7.08862394e-02 -4.57740158e-01 -1.62392586e-01
8.36717606e-01 -2.63823628e-01 -1.41836941e-01 -8.04854870e-01
3.69198203e-01 3.19413930e-01 8.05301785e-01 -6.49672329e-01
9.04029846e-01 2.19016343e-01 4.75695990e-02 -5.75741768e-01
-1.25015330e+00 -6.40010834e-01 -9.64296997e-01 -2.89262384e-01
1.22147870e+00 -1.44020641e+00 -8.51630092e-01 5.88584363e-01
-1.22262907e+00 -2.69896030e-01 7.99013153e-02 3.92107427e-01
-3.44854236e-01 2.16354027e-01 -5.95902324e-01 -7.30761647e-01
-4.94129062e-01 -9.82976794e-01 1.57150507e+00 1.44900501e-01
-3.90690893e-01 -7.85458863e-01 3.29490721e-01 7.48916030e-01
2.95323700e-01 6.87589347e-01 -2.17902735e-01 -6.32447958e-01
-5.01727760e-01 -4.96317118e-01 -1.15060180e-01 1.27290795e-02
-3.56515765e-01 -4.32706654e-01 -1.03180170e+00 -6.78439617e-01
-4.94430125e-01 -9.15651083e-01 1.04583514e+00 5.32989800e-01
1.28812230e+00 -5.81449680e-02 -7.68053412e-01 9.04068232e-01
9.00200725e-01 -6.17737293e-01 4.13391083e-01 2.26146638e-01
1.06289458e+00 3.59640241e-01 6.36687517e-01 4.94730830e-01
8.14279675e-01 1.21736050e+00 3.76603544e-01 -1.22727580e-01
-3.10617089e-01 -3.74128938e-01 2.38874972e-01 1.91447526e-01
-4.78062361e-01 -1.23959780e-02 -8.72231722e-01 1.32881597e-01
-1.83677208e+00 -1.05377448e+00 -5.94018288e-02 1.95923102e+00
7.96704590e-01 1.62262037e-01 9.15413737e-01 -2.69308776e-01
4.93821353e-01 3.54650110e-01 -5.53481221e-01 4.07389045e-01
-1.03775494e-01 1.21393949e-01 5.96807718e-01 2.22953767e-01
-1.90598249e+00 1.06874454e+00 6.35248137e+00 6.88126445e-01
-8.96266520e-01 3.07508469e-01 3.92295033e-01 -5.07833838e-01
4.88273472e-01 -5.20870090e-01 -1.54950082e+00 3.26747775e-01
7.16727197e-01 4.48923230e-01 2.14889180e-02 1.00356424e+00
-9.21784043e-02 1.37532771e-01 -1.22217810e+00 1.42594409e+00
8.02295208e-02 -7.85236657e-01 -4.75310743e-01 1.38217524e-01
6.55995369e-01 1.53263554e-01 -4.95849103e-02 7.16285825e-01
1.48831472e-01 -1.06773627e+00 7.51749575e-01 2.66510159e-01
6.32888019e-01 -6.92361593e-01 7.76211202e-01 4.02141929e-01
-1.46575737e+00 -3.54856178e-02 -3.21167141e-01 8.74611810e-02
4.03073579e-01 2.63778895e-01 -4.20563072e-01 4.72209096e-01
8.78533423e-01 8.27369153e-01 -1.14574873e+00 1.36457253e+00
-2.50463009e-01 3.88621598e-01 -4.76187915e-01 -2.24939156e-02
-1.37285501e-01 3.35611820e-01 4.30169940e-01 1.23855412e+00
-3.39208581e-02 -2.13942975e-01 6.89281344e-01 4.17736143e-01
-1.97425142e-01 -2.12220907e-01 -2.67346352e-01 2.25184739e-01
2.93434024e-01 1.36601973e+00 -3.05377543e-01 -9.53196641e-03
-2.84904361e-01 1.10701871e+00 6.01944089e-01 3.87767330e-02
-1.24341929e+00 2.63526112e-01 8.35376918e-01 1.38090566e-01
2.82174468e-01 -3.50981325e-01 -1.46039858e-01 -1.11494887e+00
1.55468062e-01 -1.03650701e+00 6.76691890e-01 -2.06335336e-01
-1.37129951e+00 5.60717106e-01 1.70996502e-01 -1.23678684e+00
-3.01843911e-01 -8.50883842e-01 -2.97223032e-01 7.50628769e-01
-1.15762174e+00 -1.77093637e+00 -4.35475349e-01 6.14538372e-01
2.44853765e-01 -1.34061677e-02 5.66264987e-01 5.93914270e-01
-7.21657217e-01 1.16623330e+00 -4.27426875e-01 5.46602547e-01
1.21080554e+00 -1.23538077e+00 5.13688445e-01 6.81694746e-01
-3.39649469e-02 4.81189251e-01 8.15158725e-01 -8.55223715e-01
-1.44881546e+00 -1.28052247e+00 7.91042149e-01 -1.07729006e+00
2.69073814e-01 -7.54601836e-01 -4.92148042e-01 8.76716673e-01
3.94504704e-02 2.65270114e-01 6.14031255e-01 6.52257085e-01
-3.56036365e-01 -3.91480587e-02 -9.71033812e-01 3.25497985e-01
1.43707108e+00 -1.50434077e-01 -4.29900467e-01 6.71003699e-01
7.19859123e-01 -1.10672426e+00 -8.62143397e-01 4.60624069e-01
7.96200991e-01 -6.31654859e-01 1.42463064e+00 -7.25897431e-01
2.86414981e-01 -1.97080329e-01 5.54407015e-02 -1.25532496e+00
-3.51605833e-01 -3.84709299e-01 -5.31504512e-01 9.17894602e-01
3.67531300e-01 -1.95729092e-01 1.15885973e+00 5.86232722e-01
2.15197548e-01 -7.98207581e-01 -7.94015586e-01 -9.91597831e-01
1.79871053e-01 -2.85829008e-01 4.84925896e-01 3.49203467e-01
-4.65911984e-01 5.99972665e-01 -1.19012690e+00 2.21195996e-01
1.01661444e+00 -1.65012494e-01 1.43630540e+00 -1.34641743e+00
-6.14765763e-01 -7.82595500e-02 -3.74743193e-01 -1.30017471e+00
1.47769958e-01 -6.19981229e-01 1.64407045e-01 -1.48946106e+00
4.53626901e-01 -1.63635701e-01 1.33681417e-01 5.18739998e-01
-4.69894528e-01 3.93205792e-01 5.31545341e-01 -6.36931101e-04
-1.21497607e+00 6.60369456e-01 1.43761826e+00 -1.37450144e-01
-6.20210581e-02 3.21343839e-01 -6.00456536e-01 6.51214421e-01
4.42669034e-01 -2.81663179e-01 -1.01163380e-01 -4.44843888e-01
1.38290167e-01 -4.54075560e-02 6.56295538e-01 -1.55683219e+00
3.92652452e-01 1.92652047e-01 9.29395258e-01 -1.13506079e+00
8.44332397e-01 -6.23504400e-01 4.11871374e-02 4.65728015e-01
-1.06157407e-01 4.64067981e-02 2.50183374e-01 7.30389833e-01
3.52723320e-04 4.16492313e-01 8.40476990e-01 -1.42579928e-01
-9.31283116e-01 8.22050929e-01 4.97684538e-01 5.42301416e-01
1.03748381e+00 2.51284912e-02 -2.65774667e-01 -2.08817199e-01
-9.86554265e-01 6.59771800e-01 1.19394593e-01 7.77454674e-01
3.21399957e-01 -1.56772852e+00 -8.49525690e-01 -2.31103629e-01
2.64609188e-01 1.65083453e-01 3.40368867e-01 7.76973724e-01
-3.38194370e-01 5.50817907e-01 -2.91573197e-01 -6.77825809e-01
-1.36858654e+00 3.53349894e-01 5.94676137e-01 -7.14135587e-01
-5.18872082e-01 1.31206000e+00 -2.99261250e-02 -8.10300589e-01
5.53479910e-01 -9.42565873e-02 -2.65825331e-01 1.33606210e-01
6.90367937e-01 4.28623945e-01 -6.01194426e-02 -9.82791901e-01
-7.03588367e-01 6.37499094e-01 6.42424747e-02 1.09815709e-01
1.34416068e+00 6.77509829e-02 2.77045339e-01 1.10318579e-01
1.20735109e+00 -2.34230116e-01 -1.63334107e+00 -2.56987453e-01
-1.79791495e-01 -4.47347015e-01 -3.42868209e-01 -8.78407598e-01
-1.12000990e+00 8.20120931e-01 7.90256798e-01 -5.52875280e-01
7.16465950e-01 1.76791176e-01 9.91430700e-01 4.74888325e-01
6.45737529e-01 -1.37242067e+00 5.04270673e-01 6.90191388e-01
1.01417017e+00 -1.67072093e+00 1.17550164e-01 -4.72606063e-01
-7.57091343e-01 6.95802689e-01 1.22986960e+00 -9.39230621e-02
6.24369085e-01 2.33296126e-01 1.26477592e-02 -1.65291220e-01
-4.29281741e-01 -3.57958764e-01 8.49414945e-01 7.23784566e-01
6.02288485e-01 2.14182928e-01 -1.67274818e-01 1.03135395e+00
-4.46988225e-01 9.06675160e-02 -2.89689004e-01 6.72291994e-01
-3.55759084e-01 -1.15385759e+00 -6.61351264e-01 3.71934682e-01
-4.98197973e-01 3.16998243e-01 -5.07213712e-01 8.57378006e-01
3.46674562e-01 6.62767529e-01 -4.00478393e-01 -5.78050971e-01
5.66859961e-01 -2.73895711e-02 8.72330964e-01 -4.54641968e-01
-7.04457402e-01 -4.27748226e-02 3.22252154e-01 -8.82379234e-01
-4.14673775e-01 -8.68278503e-01 -8.65411460e-01 -4.34689969e-01
-6.36613667e-02 -3.34217191e-01 3.03658038e-01 1.08518791e+00
2.40317285e-01 6.24132752e-01 -5.28432466e-02 -1.34977424e+00
-7.46893644e-01 -1.17001104e+00 -3.13740164e-01 5.66256583e-01
1.74628779e-01 -1.06045914e+00 2.40702361e-01 -3.20217729e-01] | [7.04201602935791, -0.8608893752098083] |
58b8ad78-0b63-47c7-bfca-80e309c898b1 | learnable-spatio-temporal-map-embeddings-for | 2211.07635 | null | https://arxiv.org/abs/2211.07635v1 | https://arxiv.org/pdf/2211.07635v1.pdf | Learnable Spatio-Temporal Map Embeddings for Deep Inertial Localization | Indoor localization systems often fuse inertial odometry with map information via hand-defined methods to reduce odometry drift, but such methods are sensitive to noise and struggle to generalize across odometry sources. To address the robustness problem in map utilization, we propose a data-driven prior on possible user locations in a map by combining learned spatial map embeddings and temporal odometry embeddings. Our prior learns to encode which map regions are feasible locations for a user more accurately than previous hand-defined methods. This prior leads to a 49% improvement in inertial-only localization accuracy when used in a particle filter. This result is significant, as it shows that our relative positioning method can match the performance of absolute positioning using bluetooth beacons. To show the generalizability of our method, we also show similar improvements using wheel encoder odometry. | ['Kris Kitani', 'Vivek Roy', 'Karnik Ram', 'Dennis Melamed'] | 2022-11-14 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-1.66433156e-01 6.44877702e-02 -1.90875009e-01 -5.02252877e-01
-8.37899745e-01 -5.73995352e-01 5.80252051e-01 2.88704842e-01
-6.64523423e-01 1.08966887e+00 5.71249068e-01 -2.33605042e-01
-3.69981825e-01 -1.08389354e+00 -1.06601930e+00 -3.22501242e-01
-3.60565752e-01 4.51807290e-01 1.64854512e-01 -3.11880589e-01
2.61355609e-01 3.44430000e-01 -1.35867417e+00 -6.64198339e-01
7.89299488e-01 6.03962898e-01 2.42961552e-02 7.81425357e-01
5.13862669e-01 3.01175147e-01 -5.97107053e-01 4.97919768e-02
2.08967030e-01 6.89863116e-02 -4.56407040e-01 -3.60796243e-01
5.69760025e-01 -4.54267412e-01 -6.97224617e-01 6.75249755e-01
8.21643949e-01 3.99980873e-01 3.27953041e-01 -1.36965013e+00
-5.76016486e-01 1.24177657e-01 -5.22753075e-02 4.03535336e-01
8.46581697e-01 -1.57616377e-01 6.63016319e-01 -6.03622079e-01
5.40298522e-01 9.97541487e-01 1.25814867e+00 5.33490404e-02
-1.45663214e+00 -6.09880745e-01 1.15459003e-01 -2.97227800e-01
-1.76444674e+00 -5.37240207e-01 3.33796859e-01 -3.55838180e-01
1.05425096e+00 1.69571593e-01 4.72657502e-01 1.10068345e+00
6.26389146e-01 2.15144694e-01 7.44430423e-01 -8.98575336e-02
3.58073413e-01 -5.07711340e-03 2.82042921e-02 6.28343523e-01
8.36985171e-01 1.40358135e-01 -6.28758788e-01 -1.14501745e-01
7.85676360e-01 2.94391990e-01 -2.26009041e-01 -7.69196212e-01
-1.11304152e+00 6.14006162e-01 9.48626339e-01 7.00030103e-02
-3.19629252e-01 6.12557590e-01 -1.62829474e-01 -3.23841125e-02
4.76541758e-01 7.24451840e-01 -5.38946927e-01 -7.03315735e-01
-8.65415990e-01 4.40592140e-01 9.02317226e-01 1.22215462e+00
1.17954338e+00 -2.39579067e-01 3.26740026e-01 3.95254612e-01
3.82030129e-01 1.02458024e+00 3.12498748e-01 -9.26738441e-01
5.82527518e-01 3.65758657e-01 6.56360269e-01 -1.28903389e+00
-8.83230448e-01 -3.96464139e-01 -3.17494273e-01 -1.74078777e-01
3.03837419e-01 -5.18623471e-01 -1.02984190e+00 1.84958589e+00
2.41138086e-01 5.57182968e-01 -1.66773498e-01 7.01185942e-01
2.21359387e-01 3.15766037e-01 -1.32434681e-01 4.64815050e-01
1.14391541e+00 -3.71937603e-01 -5.84159195e-01 -7.00198948e-01
7.31820703e-01 -4.56230342e-01 6.93452239e-01 -2.43251011e-01
-5.74474692e-01 -4.75217879e-01 -1.59883916e+00 -2.27540389e-01
-4.58736062e-01 -2.23372534e-01 6.77335203e-01 1.18319333e+00
-1.38775003e+00 3.17743450e-01 -1.35708034e+00 -6.66073501e-01
-4.07922044e-02 8.00296068e-01 -3.80127221e-01 -2.08776012e-01
-1.27372050e+00 1.27415061e+00 7.65570179e-02 -1.90169916e-01
-3.53590995e-01 -6.43613517e-01 -1.39852870e+00 -1.47465408e-01
-2.51369327e-01 -8.85756731e-01 1.03697443e+00 1.80356234e-01
-1.42260838e+00 2.58183211e-01 -6.64608717e-01 -6.39208436e-01
2.78440237e-01 -4.78344977e-01 -5.44216275e-01 -4.75840211e-01
5.47365665e-01 4.83200341e-01 2.19084531e-01 -1.13813269e+00
-7.96761513e-01 -2.73674607e-01 2.24188983e-01 4.67602998e-01
2.15770211e-02 -8.99321973e-01 -5.41209340e-01 -5.75686507e-02
8.13712239e-01 -1.34709239e+00 -4.03345615e-01 -3.49473119e-01
-1.20434590e-01 4.05037761e-01 3.72485340e-01 -4.24420089e-01
1.17625701e+00 -2.05693150e+00 -4.44313049e-01 4.33204144e-01
4.05424759e-02 -4.46232945e-01 3.39226663e-01 3.75395864e-01
4.22303647e-01 2.69166604e-02 1.72621116e-01 -7.37785697e-01
3.45351875e-01 6.41866565e-01 -2.88717926e-01 8.62548530e-01
-1.38829440e-01 8.53318393e-01 -1.23500967e+00 5.76545671e-02
5.87511361e-01 6.38158083e-01 -6.18895113e-01 -2.99702793e-01
5.56258202e-01 4.71421212e-01 -2.46478736e-01 2.27510110e-01
8.75908554e-01 -8.13555121e-02 2.46177375e-01 2.39508552e-03
-2.55020112e-01 9.94965136e-01 -1.30558443e+00 1.93165445e+00
-8.73054981e-01 8.10619175e-01 -2.39820436e-01 -2.37691551e-01
8.62741172e-01 1.08528286e-02 5.78940690e-01 -7.83669233e-01
-1.51444301e-01 1.94725439e-01 -4.31434721e-01 -2.71369517e-02
1.10077643e+00 5.32134324e-02 -4.29453403e-01 3.08095127e-01
-2.55714840e-04 -2.90772170e-01 -6.93230852e-02 -7.22078681e-02
1.39607596e+00 2.42925778e-01 2.23963693e-01 -4.83010173e-01
1.03913032e-01 -3.00393049e-02 6.07642829e-01 1.08641374e+00
-2.96297997e-01 6.73923373e-01 -1.79488674e-01 -5.61029494e-01
-8.37662876e-01 -1.46417856e+00 -5.11171162e-01 9.07821834e-01
7.94914424e-01 -6.52148485e-01 -2.82054842e-01 -4.38668609e-01
6.61651134e-01 5.23571432e-01 -6.46004558e-01 -1.75554618e-01
-6.64685130e-01 -9.95317578e-01 7.84439802e-01 6.88333809e-01
3.78037095e-01 1.35950923e-01 -4.02785748e-01 3.84598941e-01
-1.81719780e-01 -9.91573334e-01 -3.37215692e-01 4.62697476e-01
-7.48695910e-01 -8.25186968e-01 -4.62925881e-01 -5.46498299e-01
7.82760799e-01 4.60960031e-01 8.53057981e-01 -3.32771391e-01
3.80722046e-01 4.72500026e-01 7.39825442e-02 -1.93524018e-01
2.26354465e-01 5.11189938e-01 8.03613245e-01 -4.85763550e-01
4.37900454e-01 -7.31137335e-01 -7.70940661e-01 5.82177162e-01
-6.64659590e-02 -2.90122628e-01 2.89452255e-01 6.59709275e-01
2.98681974e-01 -7.08139464e-02 1.04317211e-01 -3.85137945e-01
6.62860036e-01 -7.37208664e-01 -6.34578884e-01 -3.86247039e-01
-9.49212968e-01 3.88421357e-01 8.87808669e-03 -7.70385098e-03
-5.21965981e-01 2.55147666e-01 -3.32406670e-01 4.93293494e-01
1.52243031e-02 2.97174454e-01 -3.68149281e-02 -4.90568668e-01
9.62573886e-01 5.98991662e-02 -2.50770241e-01 -2.72055805e-01
4.62807506e-01 5.68086386e-01 5.07341683e-01 -4.88390207e-01
9.07295942e-01 5.99436104e-01 -1.65400840e-02 -6.36309087e-01
-3.77108008e-01 -4.61448252e-01 -5.56024134e-01 3.48601282e-01
5.64995825e-01 -1.31334507e+00 -1.04256201e+00 -7.22995922e-02
-8.80115509e-01 -3.24137092e-01 2.19151080e-02 8.13077986e-01
-4.89350200e-01 -4.22358233e-03 -4.85120207e-01 -8.10663760e-01
4.78406429e-01 -1.10269558e+00 1.25010490e+00 4.85850722e-01
-4.49002326e-01 -1.16504574e+00 5.84428549e-01 -2.70723075e-01
4.86809611e-01 2.36514702e-01 -1.33273736e-01 -2.61632949e-01
-7.71973729e-01 -5.66706836e-01 -1.86240777e-01 -3.99500310e-01
5.25955617e-01 -6.54126287e-01 -1.06338990e+00 -5.18009901e-01
-2.11717322e-01 3.59058470e-01 5.85107684e-01 5.20507574e-01
1.75503150e-01 -2.16617242e-01 -9.08558726e-01 8.73669505e-01
1.41280389e+00 1.46964062e-02 6.90832376e-01 7.44510472e-01
7.05066800e-01 -2.18874067e-01 4.87477005e-01 3.43400717e-01
1.19369018e+00 8.48556340e-01 3.00138116e-01 -7.38186389e-03
1.95564270e-01 -6.81922615e-01 2.71099210e-01 4.03328210e-01
3.67985889e-02 -3.17497909e-01 -1.10727894e+00 6.93506479e-01
-1.91365755e+00 -6.92983329e-01 1.14064282e-02 2.36429334e+00
4.51352507e-01 4.56652850e-01 -2.29011163e-01 3.76707874e-02
4.60786581e-01 2.95554042e-01 -2.74743468e-01 -3.30653824e-02
2.05134884e-01 1.59350768e-01 1.66246772e+00 1.23546660e+00
-1.33248043e+00 7.29292989e-01 7.36133242e+00 5.08843288e-02
-9.99149859e-01 2.34725028e-01 -8.23076144e-02 3.11733335e-02
-2.73098320e-01 -4.22907136e-02 -1.07009721e+00 6.90422773e-01
1.27571082e+00 -5.40458038e-02 3.29023868e-01 9.84517336e-01
1.84417456e-01 -5.70465982e-01 -9.17823613e-01 1.03223479e+00
-9.30547416e-02 -1.39773154e+00 -7.80431330e-01 4.12992626e-01
9.18761253e-01 6.55943394e-01 1.20022610e-01 4.21582341e-01
7.47953415e-01 -7.40052402e-01 5.59508085e-01 3.24337393e-01
4.91927505e-01 -7.22071052e-01 8.58413815e-01 2.82102853e-01
-1.42977583e+00 1.11414820e-01 -3.55791897e-01 -8.47940981e-01
5.01223147e-01 5.57031572e-01 -1.52889287e+00 5.74546278e-01
5.70798934e-01 6.15688503e-01 -6.79730296e-01 1.30749369e+00
-4.64884311e-01 1.73349336e-01 -9.57867324e-01 -6.23021834e-02
1.77345321e-01 2.00136364e-01 2.91329801e-01 1.09362268e+00
6.24733090e-01 -5.76784551e-01 2.30276227e-01 4.73875612e-01
1.90143436e-01 -6.74104989e-01 -1.10509384e+00 4.66870099e-01
9.93502617e-01 6.46940589e-01 -5.45578182e-01 -2.06494123e-01
-8.69085826e-03 1.11438191e+00 4.09883559e-02 5.37405074e-01
-8.56788158e-01 -6.86347842e-01 1.30481994e+00 1.77630246e-01
2.73978803e-02 -1.14613903e+00 -6.68865085e-01 -1.09480131e+00
-6.81124181e-02 9.17219520e-02 -1.48313865e-01 -5.98272443e-01
-6.91396832e-01 1.91807643e-01 -1.66784748e-01 -1.19520128e+00
-5.32992005e-01 -4.96044159e-01 -2.93243155e-02 1.10070610e+00
-1.20959771e+00 -7.95470357e-01 -5.50508142e-01 2.88942814e-01
-2.01656520e-01 5.22096574e-01 9.56933558e-01 7.15825081e-01
-2.57791486e-02 6.25319362e-01 5.01991570e-01 1.77459314e-01
9.78101313e-01 -1.62505996e+00 1.29327095e+00 1.04402423e+00
2.90769517e-01 1.38455045e+00 1.03066742e+00 -7.56059527e-01
-1.50069416e+00 -9.62828875e-01 1.18999803e+00 -1.32983375e+00
4.21124071e-01 -7.10785866e-01 -3.05982947e-01 1.18349266e+00
-3.86173069e-01 1.12055711e-01 3.98298591e-01 6.96456313e-01
-4.26993519e-02 -2.51530647e-01 -1.18379748e+00 5.11486053e-01
1.19157374e+00 -8.64279389e-01 -4.99901295e-01 1.54561177e-01
6.31375134e-01 -1.05959296e+00 -8.56984198e-01 2.09934324e-01
7.68440783e-01 -8.03210855e-01 1.25054276e+00 1.98540732e-01
-5.78505695e-01 -7.88362622e-01 -3.31286311e-01 -1.44929636e+00
-5.27458310e-01 -5.79781711e-01 1.33498879e-02 8.49318922e-01
3.11170936e-01 -1.09428144e+00 1.03660214e+00 9.92062628e-01
-1.75215542e-01 -1.16587661e-01 -1.19339609e+00 -9.35075521e-01
-3.32384080e-01 -6.44400239e-01 1.00764871e+00 7.95920491e-01
1.07662179e-01 6.54803813e-02 -4.45733696e-01 8.89288187e-01
4.32100058e-01 -5.60849845e-01 1.15509820e+00 -1.08806205e+00
1.39444396e-01 -6.47725463e-02 -1.07991982e+00 -1.77244079e+00
-2.00719133e-01 -5.36335349e-01 4.49629515e-01 -1.78951323e+00
-5.89824855e-01 -8.36803138e-01 -4.01240289e-01 2.44123697e-01
-1.13994904e-01 6.39361441e-01 -1.77754641e-01 -4.75159697e-02
-6.06485069e-01 2.76319653e-01 5.23505569e-01 -1.21702831e-02
-5.94409585e-01 1.45442545e-01 -6.02395892e-01 4.27183658e-01
5.61590314e-01 -5.85568964e-01 -4.74236369e-01 -7.36406922e-01
4.31580275e-01 -2.81612456e-01 3.23441356e-01 -1.56559050e+00
5.16650379e-01 1.60222918e-01 7.77044475e-01 -4.33237165e-01
6.10265136e-01 -6.73264027e-01 2.10459843e-01 5.40508807e-01
3.54335040e-01 4.28850740e-01 4.15827632e-01 7.92853117e-01
8.13727751e-02 4.00547445e-01 2.34026358e-01 3.82328510e-01
-9.44070220e-01 1.03407800e-01 -5.92707634e-01 -2.26508304e-01
4.77404565e-01 -4.20665503e-01 -3.95215392e-01 -6.72046542e-01
-5.27413666e-01 2.60176182e-01 1.01836467e+00 4.45411861e-01
2.82167792e-01 -1.59594536e+00 5.76820970e-02 4.85476136e-01
2.22679749e-01 -8.64475146e-02 -6.99475110e-02 9.63284731e-01
-8.21321785e-01 6.81162775e-01 -1.75638229e-01 -7.55277872e-01
-5.12372553e-01 9.09393132e-02 4.36002672e-01 9.86003429e-02
-4.84699100e-01 9.61645544e-01 -3.91132534e-01 -8.29852581e-01
2.01981425e-01 -8.12083185e-01 4.66318280e-01 -3.48443627e-01
4.18525875e-01 4.44562018e-01 1.80545628e-01 -6.30549192e-01
-1.00894248e+00 5.61707139e-01 4.26819444e-01 -6.24731421e-01
8.74096930e-01 -7.71903634e-01 4.81931090e-01 4.63470131e-01
1.19280279e+00 5.10503709e-01 -1.56864309e+00 4.56641950e-02
3.71153593e-01 -6.12843931e-01 -5.56262396e-02 -5.17145395e-01
-1.00414813e-01 4.16582197e-01 9.49052036e-01 1.06567591e-01
3.27126443e-01 -4.43166047e-01 6.36812091e-01 6.26382768e-01
1.01514757e+00 -9.32197034e-01 -6.25509202e-01 9.78332996e-01
1.36182815e-01 -1.35688853e+00 2.78962493e-01 3.40006575e-02
-2.30753496e-01 7.32544959e-01 3.39965403e-01 -5.07522583e-01
7.85302699e-01 3.84834349e-01 8.56729895e-02 1.89686909e-01
-5.44242151e-02 -3.90816987e-01 1.86127633e-01 1.09127510e+00
3.89507890e-01 7.56975412e-02 4.38999496e-02 2.97422826e-01
-6.48005128e-01 7.02948822e-03 3.81484270e-01 1.22950506e+00
-6.30258381e-01 -1.05853236e+00 -5.26261568e-01 3.65962982e-02
1.28457353e-01 2.13979743e-02 1.87972248e-01 9.72392678e-01
2.30396777e-01 1.02007735e+00 2.41172165e-01 -8.64315867e-01
3.97667289e-01 -2.82418896e-02 5.76841593e-01 -4.96330261e-01
5.86394034e-02 -3.96394163e-01 1.69189215e-01 -9.15908933e-01
3.07254624e-02 -5.22578895e-01 -1.34506392e+00 -7.02037692e-01
-8.21695700e-02 1.82042062e-01 9.97980416e-01 8.52388084e-01
7.94844270e-01 5.15910566e-01 2.73816556e-01 -1.38275707e+00
-6.14580512e-02 -7.11708546e-01 -3.92957211e-01 1.35359138e-01
8.20510507e-01 -1.07985091e+00 -1.87321946e-01 -3.91421348e-01] | [6.463015079498291, 0.7126482725143433] |
bdfdb8c9-0f49-4f4e-955e-53647c4f82ca | entropy-difference-based-stereo-error | 1711.10412 | null | http://arxiv.org/abs/1711.10412v1 | http://arxiv.org/pdf/1711.10412v1.pdf | Entropy-difference based stereo error detection | Stereo depth estimation is error-prone; hence, effective error detection
methods are desirable. Most such existing methods depend on characteristics of
the stereo matching cost curve, making them unduly dependent on functional
details of the matching algorithm. As a remedy, we propose a novel error
detection approach based solely on the input image and its depth map. Our
assumption is that, entropy of any point on an image will be significantly
higher than the entropy of its corresponding point on the image's depth map. In
this paper, we propose a confidence measure, Entropy-Difference (ED) for stereo
depth estimates and a binary classification method to identify incorrect
depths. Experiments on the Middlebury dataset show the effectiveness of our
method. Our proposed stereo confidence measure outperforms 17 existing measures
in all aspects except occlusion detection. Established metrics such as
precision, accuracy, recall, and area-under-curve are used to demonstrate the
effectiveness of our method. | ['Ram Mohana Reddy Guddeti', 'Subhayan Mukherjee', 'Irene Cheng', 'Anup Basu'] | 2017-11-28 | null | null | null | null | ['stereo-depth-estimation'] | ['computer-vision'] | [ 3.38966072e-01 -1.32025659e-01 -1.24029517e-02 -4.92500603e-01
-4.69880670e-01 -1.92431986e-01 2.66897082e-01 3.56690168e-01
-4.80624944e-01 7.39833415e-01 1.58142820e-01 -2.54382901e-02
1.03035986e-01 -8.89447510e-01 -4.41125125e-01 -5.10103822e-01
1.70609400e-01 -9.63552073e-02 6.10974967e-01 1.86399743e-01
8.88033986e-01 3.19271654e-01 -1.68180478e+00 -1.11842953e-01
1.25493836e+00 1.40648687e+00 1.87457353e-01 4.64190274e-01
7.52915768e-03 5.06864429e-01 -5.17610610e-01 -3.67451370e-01
4.64434594e-01 -4.05557632e-01 -5.20061553e-01 3.08623631e-02
4.54723030e-01 -7.05562830e-01 -3.69802922e-01 1.28375030e+00
4.31266993e-01 -3.47115435e-02 8.93188715e-01 -1.00577426e+00
1.31783903e-01 -2.53075212e-01 -6.73950493e-01 2.24409088e-01
6.36840820e-01 -5.61511293e-02 6.34950221e-01 -7.01822460e-01
4.94477898e-01 7.92552352e-01 8.41807485e-01 2.90157527e-01
-1.01596534e+00 -6.49546325e-01 -1.30425289e-01 1.90018177e-01
-1.40553057e+00 -3.57687563e-01 8.01952302e-01 -6.82751119e-01
6.83005691e-01 4.00849618e-02 5.93781948e-01 3.95782053e-01
5.00421047e-01 3.35902482e-01 1.21450436e+00 -4.81492251e-01
4.32128906e-01 8.83461162e-02 -1.18137650e-01 6.86229646e-01
4.48270619e-01 4.07220274e-01 -4.57631320e-01 -1.52097374e-01
9.28287506e-01 -6.26504123e-02 -5.02422392e-01 -6.79246485e-01
-7.79451787e-01 6.49760008e-01 4.25893962e-01 8.96285325e-02
-1.33051828e-01 -1.26069337e-01 2.73313850e-01 1.50827691e-01
5.18657684e-01 1.80818722e-01 -9.03796181e-02 -2.69798726e-01
-9.46489751e-01 2.18743518e-01 5.26594698e-01 8.22294891e-01
8.74410629e-01 -4.16207641e-01 1.94215193e-01 6.74035788e-01
2.14846328e-01 2.38806412e-01 6.07159197e-01 -1.03783429e+00
5.29557824e-01 7.68452525e-01 1.34793684e-01 -1.47466946e+00
-2.56725043e-01 -1.22323595e-01 -6.83011830e-01 4.78474826e-01
4.32058156e-01 1.27868906e-01 -6.81305707e-01 1.37901199e+00
4.00208920e-01 -1.05105769e-02 -5.21220639e-02 9.60925996e-01
5.65521777e-01 4.18023974e-01 -2.95117021e-01 -3.57020766e-01
6.16051793e-01 -4.50369388e-01 -5.66330135e-01 -2.88398653e-01
4.39271361e-01 -7.91199327e-01 7.42246270e-01 3.42315495e-01
-1.02436054e+00 -6.04585290e-01 -1.27288604e+00 -1.77872423e-02
2.96454113e-02 -1.71864703e-01 4.85856503e-01 5.58787942e-01
-7.80295551e-01 7.82889009e-01 -6.50022209e-01 -2.57250905e-01
6.16426729e-02 3.14955264e-01 -3.16335171e-01 -1.95067786e-02
-7.25289524e-01 7.49586880e-01 4.54732567e-01 -1.18458167e-01
-1.85220242e-01 -3.21211219e-01 -9.12661374e-01 -1.27097622e-01
-7.91388527e-02 -5.39339364e-01 1.11067760e+00 -8.72185230e-01
-1.20910871e+00 9.15696025e-01 -2.81999350e-01 -3.23692977e-01
9.01677191e-01 -1.13422737e-01 -1.38411209e-01 2.69584030e-01
2.08573878e-01 6.36879385e-01 6.85101449e-01 -1.14073420e+00
-1.05461419e+00 -6.36989832e-01 1.00129135e-01 2.23190680e-01
8.72715004e-03 -5.04462302e-01 -2.64370173e-01 -4.26972598e-01
1.03510797e+00 -5.69019914e-01 -1.39320657e-01 3.37446600e-01
-1.93059251e-01 1.65208250e-01 6.18132412e-01 -6.56741321e-01
1.49502718e+00 -2.30549693e+00 -2.18513876e-01 3.32186699e-01
1.44788787e-01 -4.04451825e-02 5.15472591e-01 1.42691121e-01
6.35035485e-02 -1.36306614e-01 -4.78233367e-01 -6.60515949e-02
-3.46732169e-01 -1.11010201e-01 1.27749266e-02 6.84973955e-01
-5.41307665e-02 1.78461179e-01 -8.18999469e-01 -7.59496331e-01
5.61089694e-01 3.05406839e-01 -6.28856063e-01 2.59369969e-01
2.28634268e-01 4.13892567e-01 -3.32920432e-01 5.84687710e-01
9.63200271e-01 7.98379555e-02 -1.01288773e-01 -3.07533354e-01
-2.49299213e-01 2.51800954e-01 -1.26152408e+00 1.49010623e+00
-4.79687244e-01 7.39307404e-01 -4.63782072e-01 -5.66650450e-01
1.06026840e+00 1.41422525e-01 3.44898105e-01 -8.75465572e-01
7.34272748e-02 5.88630557e-01 -1.81540176e-01 -3.97750288e-01
5.43879747e-01 -1.06976144e-01 1.60538211e-01 2.10026521e-02
-4.62046415e-01 -2.48064235e-01 -3.53147574e-02 -2.45572384e-02
1.03343403e+00 1.09314919e-01 7.63318360e-01 -2.02816710e-01
6.39262021e-01 -3.73739660e-01 6.99921787e-01 5.26752532e-01
-6.32205725e-01 9.63173568e-01 3.54753584e-01 -3.42441797e-01
-1.14646196e+00 -8.65844131e-01 -6.51981592e-01 -3.63589562e-02
8.75041783e-01 -2.44339824e-01 -7.56018519e-01 -6.55658960e-01
1.53391644e-01 2.44064614e-01 -3.81715983e-01 -3.11383996e-02
-3.80787224e-01 -3.42346162e-01 2.77916528e-02 4.77139920e-01
1.04143405e+00 -5.85648477e-01 -9.87202048e-01 1.87184528e-01
-4.11611587e-01 -1.20755970e+00 -3.68451089e-01 -2.73769528e-01
-1.40148377e+00 -1.19592702e+00 -6.44156039e-01 -5.98002136e-01
8.34006011e-01 3.39229017e-01 8.83594990e-01 8.65304023e-02
-1.75510496e-01 4.25463207e-02 -9.69260186e-02 4.73517030e-02
-1.43427268e-01 -2.47790381e-01 -2.71967769e-01 -1.25617012e-01
3.73980433e-01 -6.16712749e-01 -1.07775593e+00 5.68570018e-01
-6.87243760e-01 -1.41966278e-02 3.68143320e-01 7.38931656e-01
5.68994045e-01 1.92038581e-01 1.36095956e-01 -5.02746224e-01
3.50041091e-01 -7.99050331e-02 -8.96583140e-01 -7.54589960e-02
-8.09868693e-01 1.52417466e-01 1.34855375e-01 -3.39102233e-03
-9.45959270e-01 2.24230602e-01 1.66542847e-02 -2.90099997e-02
1.28900055e-02 1.11478589e-01 -1.33848563e-01 -3.04797620e-01
5.31340480e-01 2.29316115e-01 -2.58708447e-01 -2.33294263e-01
-4.02705103e-01 8.20235133e-01 6.93890274e-01 -2.32159942e-01
4.11437750e-01 6.37735307e-01 3.01860511e-01 -7.00256288e-01
-5.18184662e-01 -8.15689445e-01 -4.70153242e-01 -3.89014661e-01
6.55865610e-01 -9.05797184e-01 -7.18448400e-01 6.75694287e-01
-1.28505778e+00 3.29153717e-01 2.02463895e-01 7.06327677e-01
-6.47082031e-01 6.87153876e-01 -2.89522678e-01 -1.03992689e+00
-1.06642477e-01 -1.13814557e+00 1.08991778e+00 3.12323958e-01
-2.17824951e-01 -1.01915479e+00 6.57482892e-02 2.66699970e-01
2.22599819e-01 4.43136960e-01 5.26604176e-01 4.51322757e-02
-7.11304367e-01 -3.84218663e-01 -5.94813526e-01 5.26288509e-01
1.45768791e-01 5.69273345e-03 -1.09407198e+00 -9.44198221e-02
3.58044595e-01 -1.46842783e-03 8.97450149e-01 5.47167122e-01
1.26237059e+00 -2.73903497e-02 -3.22478980e-01 7.88127482e-01
1.89449918e+00 4.75955576e-01 9.56120193e-01 5.93753219e-01
3.53391945e-01 6.41092956e-01 9.15088296e-01 6.08190894e-01
2.89707631e-01 7.28463531e-01 4.74743396e-01 1.81494504e-01
7.01672211e-02 -3.85526329e-01 9.16620940e-02 5.89810014e-01
-2.02585638e-01 -1.01196095e-01 -7.44295537e-01 4.05150801e-01
-1.71618319e+00 -8.02399993e-01 -1.56920284e-01 2.71248436e+00
6.02356970e-01 4.40147430e-01 -2.68885732e-01 6.90040946e-01
8.83908331e-01 6.18879013e-02 -4.43129569e-01 -2.28459105e-01
-1.45141978e-03 -6.98859096e-02 6.80637121e-01 6.97310567e-01
-1.10441959e+00 3.90670449e-01 6.06894588e+00 6.47069871e-01
-1.00141132e+00 -2.67319918e-01 5.66664398e-01 2.53635228e-01
-1.67114258e-01 6.71708807e-02 -7.98875391e-01 8.64413083e-01
3.32078367e-01 -1.92077890e-01 -5.53144775e-02 8.83363247e-01
1.44998297e-01 -9.00509238e-01 -1.10921955e+00 1.33975399e+00
1.20174468e-01 -8.63101184e-01 -2.04500556e-01 1.72549516e-01
7.77937174e-01 -4.39759910e-01 -8.28633234e-02 -2.04900935e-01
-3.24799448e-01 -6.75180614e-01 4.87602890e-01 4.07661647e-01
7.56703496e-01 -7.19765604e-01 9.55331564e-01 1.47854522e-01
-1.07311404e+00 1.48181543e-01 -3.81172717e-01 -2.98307002e-01
-2.60238647e-02 8.20111454e-01 -6.13498926e-01 3.94228280e-01
6.34588838e-01 6.08243048e-01 -3.50651443e-01 1.45576870e+00
-1.75404072e-01 -3.47728953e-02 -2.64974982e-01 1.04730211e-01
-7.24015683e-02 -1.34406716e-01 4.42981809e-01 8.98966908e-01
5.67227304e-01 1.89491525e-01 -1.98811308e-01 5.30205905e-01
2.32140824e-01 2.78825194e-01 -8.13631952e-01 5.41124940e-01
5.70891321e-01 6.52069390e-01 -5.89700103e-01 -2.93497533e-01
-4.95714933e-01 1.07466197e+00 3.07408813e-02 1.16843827e-01
-6.89287424e-01 -5.73306441e-01 5.56502819e-01 2.31135160e-01
1.12733781e-01 -9.11527500e-02 -7.23593354e-01 -1.16656959e+00
5.34514785e-01 -4.69864964e-01 2.58257836e-01 -7.45289207e-01
-9.16363597e-01 4.13009137e-01 -1.81244165e-01 -1.80325699e+00
-2.67072141e-01 -4.14797664e-01 -3.35560769e-01 7.55635083e-01
-1.53789818e+00 -2.52430409e-01 -8.56085896e-01 3.32560629e-01
4.59788889e-01 1.46122172e-01 3.84144992e-01 4.22537416e-01
-2.36859232e-01 5.94877064e-01 2.14445665e-02 -1.21217603e-02
5.28976202e-01 -9.19927955e-01 1.65193170e-01 9.14070308e-01
-4.01030421e-01 1.56782523e-01 7.79251158e-01 -6.64001524e-01
-6.61481857e-01 -4.88777667e-01 1.04459560e+00 -1.31330434e-02
2.42888451e-01 1.01899207e-01 -8.12245011e-01 3.02564710e-01
-3.76886994e-01 -1.29912421e-01 1.37174353e-01 -2.43807331e-01
-1.94236174e-01 -2.82031059e-01 -1.62586200e+00 4.02821749e-01
1.20854449e+00 -5.96948862e-01 -5.06725669e-01 -2.56506726e-02
2.75630176e-01 -6.27323747e-01 -6.26782477e-01 9.51560318e-01
8.31718683e-01 -1.88325322e+00 8.39704096e-01 2.63608694e-01
6.39100194e-01 -3.06687534e-01 -4.81914431e-01 -8.95548463e-01
-2.70855445e-02 -1.02836654e-01 8.71368498e-02 8.53466868e-01
1.91740021e-01 -8.01380813e-01 8.93440664e-01 6.86881065e-01
3.42340185e-03 -6.04056060e-01 -1.17389774e+00 -8.35031450e-01
-3.80119354e-01 -4.00629371e-01 3.90258968e-01 7.29376614e-01
2.19971031e-01 -5.08714430e-02 -3.11295362e-03 2.53916383e-01
9.01894271e-01 -2.64610834e-02 7.29585707e-01 -1.24505496e+00
-2.29888126e-01 -3.82121384e-01 -1.13427520e+00 -1.24575555e+00
-2.66510725e-01 -2.28898585e-01 7.91435838e-02 -1.36987400e+00
3.34840477e-01 -4.86635178e-01 2.84940358e-02 -1.75570026e-01
-2.05615669e-01 1.82992473e-01 -1.23103283e-01 4.04754430e-01
-1.20522730e-01 4.93785769e-01 1.10647380e+00 1.21350020e-01
-8.73721689e-02 1.39246553e-01 -8.11755881e-02 1.02933192e+00
7.15875387e-01 -4.03562784e-01 -3.72136146e-01 -2.95386285e-01
2.96833664e-02 2.98930347e-01 3.19704771e-01 -1.55845082e+00
6.31871447e-02 3.38100083e-02 3.80673200e-01 -5.94514430e-01
4.66618776e-01 -9.32733178e-01 4.40349057e-02 6.59294367e-01
-6.95015714e-02 6.68222308e-02 -7.27064684e-02 5.41813433e-01
-5.65114081e-01 -5.17681718e-01 1.08425975e+00 -1.92738958e-02
-8.51337075e-01 5.86948656e-02 1.40992582e-01 -3.83308716e-03
1.06887865e+00 -7.99724102e-01 -2.77927577e-01 -4.71828431e-01
-2.38705099e-01 -1.05426289e-01 1.03472805e+00 -6.42494336e-02
8.84855509e-01 -1.22065198e+00 -3.59452814e-01 3.91299248e-01
3.87299210e-01 -2.83103660e-02 1.20611623e-01 6.15415096e-01
-1.01395905e+00 4.02753264e-01 -1.78561628e-01 -8.14772248e-01
-1.28310919e+00 2.42689386e-01 4.85053778e-01 -3.28046829e-02
-4.09211040e-01 5.63404083e-01 4.09765571e-01 1.50595568e-02
3.90063405e-01 -6.30258501e-01 1.35162706e-02 -2.77114898e-01
4.49278653e-01 5.25943041e-01 4.72435243e-02 -4.68044937e-01
-3.09729457e-01 1.01940143e+00 1.83266059e-01 -2.35364318e-01
6.85234845e-01 -3.90720874e-01 1.40584841e-01 3.74392629e-01
1.30087054e+00 -7.99373761e-02 -1.30932796e+00 -1.07995376e-01
-6.12655729e-02 -1.22333694e+00 2.38362148e-01 -4.17554200e-01
-8.41171026e-01 9.43283856e-01 1.03267145e+00 9.88967940e-02
1.38050449e+00 -2.89700210e-01 8.27290714e-01 -1.03123829e-01
6.62494540e-01 -1.17133558e+00 -5.63682057e-02 9.91982669e-02
6.65288329e-01 -1.71837878e+00 1.51709929e-01 -7.11269140e-01
-2.35570237e-01 1.09165311e+00 7.79105306e-01 2.62840744e-02
6.09867454e-01 -2.49731401e-03 3.47649753e-02 1.27580866e-01
-2.42173523e-01 -8.69550556e-02 1.33602455e-01 3.97640854e-01
4.85392451e-01 -2.07806438e-01 -7.24500597e-01 -1.89500898e-01
-1.87029451e-01 2.10412025e-01 4.83362377e-01 1.09580553e+00
-7.70242691e-01 -7.10673034e-01 -3.52685422e-01 2.34200060e-01
-2.81195521e-01 5.79350404e-02 -9.32821184e-02 6.89962149e-01
6.67427033e-02 8.72233212e-01 3.68275404e-01 -4.64302808e-01
4.26187634e-01 -3.56499642e-01 6.20104313e-01 -1.87825114e-01
3.69638689e-02 -2.58058041e-01 -1.06995098e-01 -7.26115346e-01
-5.89087427e-01 -6.99847162e-01 -1.17876744e+00 -4.39543515e-01
-4.04223651e-01 -1.50341317e-01 9.09667790e-01 7.23475575e-01
1.60339355e-01 -1.28350541e-01 9.16536987e-01 -6.47643864e-01
-3.40899587e-01 -7.78152466e-01 -4.96935725e-01 4.65791464e-01
4.31644768e-01 -7.43925512e-01 -8.52008820e-01 -1.88375354e-01] | [9.075057029724121, -2.4789514541625977] |
e2709cb5-cf42-4f03-ae38-05be4e15df1f | 2d-shapley-a-framework-for-fragmented-data | 2306.10473 | null | https://arxiv.org/abs/2306.10473v1 | https://arxiv.org/pdf/2306.10473v1.pdf | 2D-Shapley: A Framework for Fragmented Data Valuation | Data valuation -- quantifying the contribution of individual data sources to certain predictive behaviors of a model -- is of great importance to enhancing the transparency of machine learning and designing incentive systems for data sharing. Existing work has focused on evaluating data sources with the shared feature or sample space. How to valuate fragmented data sources of which each only contains partial features and samples remains an open question. We start by presenting a method to calculate the counterfactual of removing a fragment from the aggregated data matrix. Based on the counterfactual calculation, we further propose 2D-Shapley, a theoretical framework for fragmented data valuation that uniquely satisfies some appealing axioms in the fragmented data context. 2D-Shapley empowers a range of new use cases, such as selecting useful data fragments, providing interpretation for sample-wise data values, and fine-grained data issue diagnosis. | ['Ruoxi Jia', 'Xi Chen', 'Xiangyu Chang', 'Hoang Anh Just', 'Zhihong Liu'] | 2023-06-18 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 2.50777423e-01 5.04883766e-01 -1.00679982e+00 -5.65695286e-01
-7.05699265e-01 -6.52102172e-01 4.65984762e-01 2.40792632e-01
-3.76495957e-01 1.08798969e+00 5.63100219e-01 -2.57696301e-01
-5.07364571e-01 -8.86606276e-01 -5.50114751e-01 -5.49649239e-01
-2.47221589e-01 1.33764192e-01 -2.95880437e-01 1.21974833e-01
2.74190843e-01 3.94155055e-01 -1.55502582e+00 4.88724858e-01
7.35390127e-01 1.25347447e+00 -1.66266635e-01 8.37408677e-02
-3.27166289e-01 7.81407475e-01 -5.96564353e-01 -1.07783651e+00
9.13658381e-01 -1.01249434e-01 -6.08767509e-01 -3.69345516e-01
2.17200756e-01 -4.51944619e-01 1.36688411e-01 1.25305367e+00
1.77348584e-01 -2.51228929e-01 5.01511574e-01 -2.28375340e+00
-8.62456858e-01 1.23267531e+00 -6.23528421e-01 -2.27126870e-02
-1.03596216e-02 1.41439527e-01 1.40646100e+00 -4.57589358e-01
8.06802213e-01 1.20220661e+00 5.08637667e-01 4.05238330e-01
-1.35897565e+00 -7.35340774e-01 -7.92828351e-02 -7.72119639e-03
-1.13292968e+00 -4.73848790e-01 9.22078133e-01 -6.67404890e-01
4.89401758e-01 9.69198287e-01 7.53323436e-01 6.55843139e-01
3.54594886e-01 8.42709422e-01 1.09121656e+00 -8.25437345e-03
6.11841142e-01 4.69277442e-01 2.19332889e-01 5.14415577e-02
1.21995533e+00 2.69498765e-01 -8.66729200e-01 -8.75185013e-01
3.51350874e-01 3.37047279e-01 6.14086464e-02 -1.04923773e+00
-1.11420858e+00 1.06100583e+00 2.88192898e-01 -1.29894540e-01
-6.35015845e-01 6.51650056e-02 4.41924870e-01 4.80799645e-01
6.01650655e-01 6.75831318e-01 -7.84883320e-01 2.25014165e-01
-9.00590122e-01 6.06003702e-01 5.79619765e-01 1.09647667e+00
7.78164685e-01 -2.05129743e-01 -4.28696901e-01 1.41697079e-01
1.50711685e-01 5.66602290e-01 2.19390437e-01 -1.51822889e+00
6.68838859e-01 9.03350890e-01 5.62646210e-01 -7.46641695e-01
-2.20406935e-01 -1.22054704e-01 -9.19840693e-01 3.95696074e-01
3.14244330e-01 -3.78984451e-01 3.62832360e-02 1.94286954e+00
4.40231979e-01 -5.84501803e-01 4.64867115e-01 8.23165953e-01
4.30958003e-01 1.02041006e-01 1.64948016e-01 -5.60529709e-01
1.04063070e+00 -2.23143965e-01 -8.81139874e-01 4.30948049e-01
6.35149419e-01 -6.14811257e-02 8.72161329e-01 1.36687160e-01
-1.24096906e+00 9.96197313e-02 -7.21259058e-01 -1.66664138e-01
-5.05858660e-01 -3.66267473e-01 8.50551426e-01 1.09089744e+00
-7.75816917e-01 5.58669388e-01 -1.32259786e-01 2.29893491e-01
1.24756098e+00 1.85106665e-01 -4.58588183e-01 1.94871515e-01
-1.20698690e+00 6.35655463e-01 2.40489617e-01 -5.27691662e-01
-6.72945738e-01 -1.48702431e+00 -6.09165013e-01 5.22318840e-01
5.10613561e-01 -8.87792110e-01 1.06293130e+00 -1.04403639e+00
-7.24977493e-01 7.47080326e-01 2.59649694e-01 -7.54119158e-01
8.32581758e-01 1.69116989e-01 -2.61362135e-01 -4.23899055e-01
3.34639579e-01 4.23735619e-01 5.35858393e-01 -1.27110696e+00
-9.89310801e-01 -7.54247367e-01 2.97068954e-01 1.72934905e-01
-4.32359964e-01 -5.66348806e-02 4.69660729e-01 -6.74484193e-01
-5.17483056e-01 -4.97035533e-01 -5.65698504e-01 3.98386449e-01
-2.98594505e-01 -1.95738822e-02 5.76748729e-01 -3.47668022e-01
1.21022010e+00 -2.09829545e+00 -3.04933131e-01 2.72370368e-01
5.28205812e-01 -1.23572081e-01 6.57810941e-02 3.33018750e-01
8.76029506e-02 7.28255391e-01 -2.82319367e-01 3.44985351e-02
3.21798056e-01 -2.64130503e-01 -5.69997191e-01 4.98314112e-01
5.17895035e-02 8.63848031e-01 -7.51399636e-01 -3.75680804e-01
-1.98036805e-01 -1.78731978e-01 -6.58952653e-01 -8.61245915e-02
-1.61749423e-01 -3.40698212e-01 -5.10965407e-01 6.08207345e-01
1.20408165e+00 -9.93685275e-02 3.69465411e-01 -3.52612771e-02
-2.53692806e-01 -2.35597119e-02 -1.42316914e+00 1.35472751e+00
-3.34723876e-03 2.87817478e-01 3.05691808e-01 -5.30892134e-01
7.14171410e-01 -5.29945530e-02 8.51564407e-01 -6.68886900e-01
-9.16680917e-02 1.00324728e-01 -1.16634540e-01 -3.53519261e-01
8.58826399e-01 -7.60775730e-02 -4.10151452e-01 9.19261932e-01
-3.09456348e-01 2.93592304e-01 -1.32513061e-01 3.86498630e-01
9.03432548e-01 -2.78429329e-01 1.02338922e+00 -6.44651592e-01
7.40770251e-02 2.61544526e-01 8.37580740e-01 6.83373094e-01
-4.45810497e-01 1.30416974e-01 9.29359376e-01 -4.13443357e-01
-1.06810701e+00 -1.02748179e+00 -1.48848966e-01 9.59501624e-01
2.23372236e-01 -7.89065585e-02 -5.75341463e-01 -9.02830660e-01
9.92464185e-01 9.75834072e-01 -8.68057668e-01 -1.77355424e-01
3.66918296e-01 -6.26738429e-01 2.98692822e-01 3.94741684e-01
3.12053233e-01 -6.87914670e-01 -1.21294391e+00 -1.63181350e-01
-2.22995188e-02 -3.17027748e-01 -6.38049901e-01 5.73528931e-02
-7.37963498e-01 -1.22753417e+00 -3.97475928e-01 2.60384142e-01
5.09535432e-01 5.58103263e-01 1.00907767e+00 -2.68357217e-01
-9.54148732e-03 2.12283522e-01 -7.82474950e-02 -1.07908809e+00
-2.73725480e-01 -2.68122554e-01 -5.21105193e-02 7.29315951e-02
4.45674539e-01 -1.21924244e-01 -5.00463367e-01 1.11818783e-01
-9.22872066e-01 1.21656293e-02 4.14163381e-01 6.10557616e-01
4.53516722e-01 -6.43619448e-02 8.87330770e-01 -1.35127115e+00
8.71552408e-01 -9.64163840e-01 -7.21356273e-01 6.49954796e-01
-1.00725150e+00 1.82442665e-01 5.61041713e-01 9.78762656e-03
-1.22829139e+00 -1.65182054e-01 8.18634033e-01 -3.12218100e-01
2.54092991e-01 4.94934350e-01 -7.49916196e-01 2.73040980e-01
7.44791687e-01 -4.36655700e-01 3.22675616e-01 -3.66323233e-01
8.33614528e-01 6.83191538e-01 1.28869697e-01 -5.13488412e-01
4.16601509e-01 5.93001246e-01 2.20714942e-01 -1.41814306e-01
-4.94382441e-01 -2.30595097e-01 -3.99090588e-01 5.68558015e-02
2.70618796e-01 -9.15605843e-01 -1.00014961e+00 5.63940965e-02
-7.49656796e-01 9.92962196e-02 -1.18050325e+00 2.28210494e-01
-6.48960590e-01 -6.17726259e-02 3.06111991e-01 -1.09344566e+00
-2.59196699e-01 -8.82611811e-01 5.25574028e-01 -1.78165540e-01
1.07313693e-01 -6.36738658e-01 4.56401780e-02 4.83956039e-01
5.38995445e-01 4.49064076e-01 8.09385538e-01 -8.72652948e-01
-7.18175948e-01 -1.35563463e-01 -3.38908166e-01 -3.45376320e-02
1.12474933e-01 -1.31218672e-01 -1.14126253e+00 -1.14645056e-01
8.23260173e-02 -1.47437051e-01 5.66938102e-01 5.28506756e-01
1.43221223e+00 -9.16719615e-01 -2.47093379e-01 2.70994604e-01
1.43348229e+00 8.04748014e-02 3.27580214e-01 1.49617493e-02
1.56701550e-01 9.98934329e-01 9.52810407e-01 1.19195604e+00
6.30696774e-01 4.93871421e-01 6.04550600e-01 6.22138306e-02
3.28122407e-01 -2.57390261e-01 1.08644404e-02 -4.18533981e-01
4.82230149e-02 1.94250438e-02 -5.74764848e-01 6.61758542e-01
-2.08346558e+00 -1.38726759e+00 -9.59688239e-03 2.56238890e+00
7.97110260e-01 -3.33424985e-01 5.48671067e-01 3.28254588e-02
7.92237818e-01 5.46561144e-02 -1.05703938e+00 -3.80470306e-01
-1.86072692e-01 -3.94485950e-01 1.11204720e+00 1.59580007e-01
-7.69940913e-01 2.64578670e-01 6.61929083e+00 6.72112226e-01
-7.35764921e-01 2.53239900e-01 8.64583910e-01 -6.63989544e-01
-1.24332273e+00 1.14862628e-01 -5.47912598e-01 6.07071638e-01
9.17526841e-01 -1.43671703e+00 2.79909164e-01 1.13788021e+00
3.26097041e-01 -1.24972381e-01 -1.25268781e+00 8.14879179e-01
-2.96098322e-01 -1.96118796e+00 2.49645293e-01 5.44344127e-01
9.29866254e-01 -1.91001803e-01 1.66405842e-01 -2.15422302e-01
9.94623244e-01 -7.24216342e-01 1.07874560e+00 7.74821401e-01
9.56168890e-01 -8.31517160e-01 6.52549446e-01 3.57025206e-01
-7.77381182e-01 -4.89932507e-01 -4.70674366e-01 -1.42001033e-01
-2.05771282e-01 9.50470865e-01 -5.78692675e-01 8.34109068e-01
6.52848482e-01 4.53847140e-01 -4.75047380e-01 9.15435970e-01
6.95208132e-01 1.13605261e-01 2.68428046e-02 2.09312618e-01
-4.24154878e-01 -5.03721833e-02 2.81374961e-01 6.27742171e-01
3.29204530e-01 -7.89474100e-02 -2.21191928e-01 1.10843623e+00
-5.96545041e-01 1.34450749e-01 -1.01446915e+00 4.62124869e-02
9.27261949e-01 1.11705875e+00 -1.11227550e-01 -2.74668157e-01
-5.22111952e-01 6.48097321e-02 2.13640243e-01 1.50009349e-01
-3.84121537e-01 -4.58500907e-02 1.11931205e+00 2.19323963e-01
-7.44819567e-02 4.64006871e-01 -9.94800150e-01 -1.22722065e+00
5.58741502e-02 -7.93635964e-01 7.19988942e-01 -5.50121307e-01
-1.54058468e+00 -1.91975579e-01 2.70720184e-01 -1.41760111e+00
3.59710790e-02 -1.42491758e-01 -2.82693386e-01 9.33196783e-01
-1.40927434e+00 -1.02180946e+00 7.23024085e-02 7.33065367e-01
-2.03166500e-01 -4.50696498e-01 5.60554087e-01 -1.87764674e-01
-2.85310179e-01 7.44931817e-01 3.55622292e-01 -4.60172743e-01
5.78566790e-01 -1.10077238e+00 2.49013543e-01 7.99631357e-01
-2.25305617e-01 8.03991377e-01 5.28558552e-01 -7.98250556e-01
-1.42465746e+00 -1.28771889e+00 9.82773423e-01 -5.50880134e-01
4.99196470e-01 -1.43093377e-01 -5.87295353e-01 5.21078825e-01
1.15139857e-01 5.10553084e-02 1.27865517e+00 9.91448835e-02
-4.75487679e-01 -7.38091528e-01 -2.18700361e+00 3.78585547e-01
1.16378438e+00 -3.13700557e-01 -5.07048309e-01 -1.40501380e-01
8.19952369e-01 1.69632539e-01 -1.04749572e+00 2.96656881e-02
6.79823697e-01 -1.03108740e+00 7.03913033e-01 -1.00574529e+00
4.36833590e-01 8.14559758e-02 -6.56886697e-01 -1.04445326e+00
-3.82313550e-01 -5.92733681e-01 6.89843521e-02 1.32536864e+00
4.29603398e-01 -8.22935641e-01 7.35761702e-01 1.56589532e+00
2.48828143e-01 -3.59611452e-01 -1.01864958e+00 -6.72734141e-01
-7.60155963e-03 -2.37365723e-01 1.67779386e+00 1.28806412e+00
2.61779785e-01 -5.52446604e-01 -4.28953856e-01 -1.75005183e-01
1.14493251e+00 5.21311879e-01 8.82710338e-01 -1.64464664e+00
2.91715384e-01 -4.30168808e-01 -1.65959805e-01 -1.44210774e-02
4.69500721e-02 -1.12113369e+00 -6.98569953e-01 -1.16917431e+00
6.79951370e-01 -5.72090864e-01 -5.31070888e-01 6.73891604e-01
5.69048412e-02 -2.16454536e-01 6.57040656e-01 4.03723657e-01
-4.89193380e-01 3.74199927e-01 1.09699535e+00 -5.18229567e-02
-1.44888103e-01 -1.58174261e-02 -1.69139504e+00 3.10988843e-01
6.37708426e-01 -6.66420579e-01 -6.19839013e-01 2.56146565e-02
5.10140359e-01 3.32006127e-01 6.10665262e-01 -1.83061659e-01
4.57704179e-02 -1.16522193e+00 7.44313300e-02 -5.46874404e-01
-2.73734093e-01 -1.37365413e+00 7.90220320e-01 4.95388180e-01
-1.00740314e+00 -1.71003282e-01 -2.32577726e-01 6.15080297e-01
9.83700529e-02 -2.29239330e-01 4.99699235e-01 -2.45018750e-01
-5.03998220e-01 4.58605349e-01 1.00341804e-01 2.23898336e-01
1.51976335e+00 -9.08159018e-02 -7.89676666e-01 8.11548531e-02
-2.31125146e-01 3.84737909e-01 7.64345109e-01 3.76651436e-01
2.65370280e-01 -1.62110054e+00 -7.15335667e-01 3.83587569e-01
3.00057858e-01 -2.23000050e-01 3.40139687e-01 3.16433877e-01
1.41099215e-01 4.26599383e-01 -7.63851225e-01 -4.05027345e-02
-1.12892628e+00 8.79201293e-01 -7.65806884e-02 -6.01072758e-02
-6.50543198e-02 2.81719565e-01 2.43295372e-01 -2.12761581e-01
-1.60614684e-01 -3.10414135e-01 -1.30816866e-02 5.78569293e-01
5.37097752e-01 7.81858444e-01 -8.09842572e-02 -2.91732997e-01
-4.39573675e-01 -3.42690080e-01 3.08414623e-02 -4.85497825e-02
1.45808101e+00 -4.02225643e-01 -9.23487172e-02 4.19410735e-01
8.66423190e-01 1.60128083e-02 -1.34193051e+00 -2.28500634e-01
-1.14905834e-01 -1.05239475e+00 7.46745616e-02 -9.94888246e-01
-1.12119234e+00 4.12315935e-01 4.17600274e-01 5.33812284e-01
1.05313182e+00 -5.21824099e-02 -7.36027062e-02 1.08536728e-01
7.61137545e-01 -1.11062253e+00 -5.20540118e-01 -4.42270428e-01
1.18404782e+00 -1.24176705e+00 3.14095199e-01 1.40889287e-02
-1.19912386e+00 5.50571799e-01 3.55098695e-01 2.12566704e-01
7.47048140e-01 3.30461591e-01 -1.57604650e-01 -1.02097258e-01
-1.02523661e+00 1.23378232e-01 9.55619588e-02 8.36628616e-01
9.65298340e-02 6.80347025e-01 -6.62140310e-01 1.29999781e+00
-2.39576116e-01 4.02839392e-01 8.95171762e-01 8.98993790e-01
-1.67653888e-01 -9.08263266e-01 -4.98388976e-01 1.20606565e+00
-3.62577945e-01 1.45018427e-02 -5.89201689e-01 6.11875415e-01
2.97361284e-01 7.11088061e-01 1.44706830e-01 -1.12353601e-01
2.92102009e-01 -9.19312239e-03 -1.01494290e-01 -2.69454002e-01
-7.42699146e-01 -2.16943100e-01 -5.55863641e-02 -6.88165128e-01
-6.83061540e-01 -1.21660471e+00 -8.48015666e-01 -7.63960600e-01
-2.58059889e-01 2.18874559e-01 6.66788697e-01 4.50732946e-01
8.76063406e-01 1.13987766e-01 9.62163270e-01 -2.57804096e-01
-1.10201907e+00 -4.03303921e-01 -1.15317786e+00 6.09151185e-01
4.58682567e-01 -6.49459839e-01 -2.61939704e-01 1.89176071e-02] | [8.686422348022461, 5.404581546783447] |
49c1e2b2-f08b-4ffe-801a-5ba2a3c6034c | medical-scientific-table-to-text-generation | 2205.12368 | null | https://arxiv.org/abs/2205.12368v2 | https://arxiv.org/pdf/2205.12368v2.pdf | Medical Scientific Table-to-Text Generation with Human-in-the-Loop under the Data Sparsity Constraint | Structured (tabular) data in the preclinical and clinical domains contains valuable information about individuals and an efficient table-to-text summarization system can drastically reduce manual efforts to condense this data into reports. However, in practice, the problem is heavily impeded by the data paucity, data sparsity and inability of the state-of-the-art natural language generation models (including T5, PEGASUS and GPT-Neo) to produce accurate and reliable outputs. In this paper, we propose a novel table-to-text approach and tackle these problems with a novel two-step architecture which is enhanced by auto-correction, copy mechanism and synthetic data augmentation. The study shows that the proposed approach selects salient biomedical entities and values from structured data with improved precision (up to 0.13 absolute increase) of copying the tabular values to generate coherent and accurate text for assay validation reports and toxicology reports. Moreover, we also demonstrate a light-weight adaptation of the proposed system to new datasets by fine-tuning with as little as 40\% training examples. The outputs of our model are validated by human experts in the Human-in-the-Loop scenario. | ['Yike Guo', 'Vibhor Gupta', 'Bingyuan Chen', 'Tong Li', 'Julia Ive', 'Jingqing Zhang', 'Heng-Yi Wu'] | 2022-05-24 | null | null | null | null | ['table-to-text-generation'] | ['natural-language-processing'] | [ 6.48806632e-01 5.09819269e-01 -1.89455450e-01 -3.21946979e-01
-1.15336680e+00 -6.03009880e-01 5.42415679e-01 9.91953611e-01
-2.89141625e-01 1.54885304e+00 5.06286502e-01 -2.04584792e-01
-1.47167176e-01 -6.19461954e-01 -6.54932976e-01 -1.83870614e-01
1.61340371e-01 6.54140711e-01 -1.45087183e-01 -2.87108719e-01
3.37846518e-01 1.00823954e-01 -1.36414623e+00 6.48873746e-01
1.59249806e+00 5.17077804e-01 -9.04302970e-02 6.78888440e-01
-3.83199096e-01 7.53019691e-01 -1.07060778e+00 -6.06243610e-01
-1.31251626e-02 -6.25535667e-01 -4.65479225e-01 -1.27168491e-01
3.58128309e-01 3.70612219e-02 2.00212777e-01 7.80247033e-01
8.56923580e-01 -1.92706019e-01 7.63248742e-01 -8.63980472e-01
-6.08478665e-01 7.60792434e-01 -6.92154586e-01 -1.14741083e-02
6.99239016e-01 2.03898579e-01 7.72702634e-01 -9.07964587e-01
9.95181501e-01 1.06103146e+00 6.16125584e-01 4.90952879e-01
-1.30192459e+00 -6.18657231e-01 -2.94427108e-02 -2.49476269e-01
-1.28146005e+00 -5.83110154e-01 4.74141836e-01 -4.66315299e-01
1.09475052e+00 6.46080494e-01 4.74231720e-01 9.09578919e-01
4.95202839e-01 3.50394011e-01 8.76283944e-01 -2.56264180e-01
4.50972676e-01 3.90997618e-01 -6.40221536e-02 4.93609071e-01
1.00681722e+00 -5.08533418e-01 -7.64883876e-01 -4.42298770e-01
2.03335136e-01 -1.45407841e-01 -1.77202016e-01 3.17060649e-01
-1.28517044e+00 7.00413942e-01 9.80314612e-02 1.75077811e-01
-7.35586107e-01 -2.83482701e-01 5.63808560e-01 2.17427853e-02
5.83331347e-01 9.00110185e-01 -4.42790717e-01 6.12269416e-02
-1.35099983e+00 5.81408203e-01 8.16985011e-01 1.27045727e+00
1.12982653e-01 8.21569562e-02 -6.70504630e-01 6.97180092e-01
-1.18266597e-01 4.67409641e-01 6.28650188e-01 -4.33976769e-01
1.00707793e+00 9.08986866e-01 2.88156182e-01 -9.04810667e-01
-5.94865918e-01 -6.21397853e-01 -1.06645966e+00 -3.71420413e-01
1.17533259e-01 -5.70760131e-01 -9.50303257e-01 1.51583064e+00
5.29780269e-01 -3.05918843e-01 3.32541734e-01 5.00622153e-01
1.23331654e+00 7.86836684e-01 4.30918187e-01 -7.19838142e-01
1.56379330e+00 -5.77460051e-01 -1.17702079e+00 -1.16067871e-01
7.87964642e-01 -8.42076778e-01 6.77704990e-01 4.80706722e-01
-1.17995906e+00 -3.56284261e-01 -1.19650459e+00 -2.15842053e-01
-5.58037877e-01 4.47119653e-01 4.55540270e-01 5.63102424e-01
-6.59020662e-01 5.74336708e-01 -3.93148720e-01 -4.77785707e-01
7.02507138e-01 4.69751179e-01 -5.41774631e-01 1.72675446e-01
-1.20453942e+00 8.53729963e-01 7.65928149e-01 -9.82440412e-02
-2.88544446e-01 -1.17519224e+00 -8.19436669e-01 -3.47425463e-04
4.44557309e-01 -1.11037922e+00 8.92627418e-01 -2.30895147e-01
-1.17815471e+00 4.99266416e-01 -1.91120133e-01 -6.01456642e-01
6.44668818e-01 -9.81662646e-02 -2.86361128e-01 2.04378307e-01
2.77212888e-01 7.09870160e-01 2.54890472e-01 -8.34545851e-01
-5.49387038e-01 -3.74531299e-01 -3.26231748e-01 9.98355970e-02
-1.18813641e-01 -1.11576289e-01 -9.45738778e-02 -9.30413127e-01
-2.01649740e-01 -5.01390100e-01 -4.52363849e-01 -2.10893512e-01
-8.00592005e-01 -1.90969899e-01 2.66231269e-01 -8.28063846e-01
1.61503327e+00 -1.46134126e+00 -1.83663487e-01 -2.28133556e-02
3.26027066e-01 3.97889584e-01 6.27015308e-02 8.52942705e-01
-2.21735820e-01 3.39660704e-01 -4.22227949e-01 -1.96831897e-01
-1.84663400e-01 -1.77330226e-01 -2.53293812e-01 1.36925071e-01
6.51921332e-01 9.22735393e-01 -9.42356169e-01 -7.82292843e-01
-1.61897257e-01 2.47345105e-01 -6.06325150e-01 1.78989843e-01
-4.30672258e-01 2.96927661e-01 -5.96641243e-01 7.65047193e-01
5.89979887e-01 -1.71199843e-01 3.81955415e-01 -2.27863058e-01
-2.29820058e-01 2.15866223e-01 -1.13737178e+00 1.55467570e+00
-1.49652109e-01 9.95159447e-02 -4.39368308e-01 -7.85425186e-01
9.68206823e-01 3.78051192e-01 5.07252872e-01 -5.89019418e-01
1.09729029e-01 2.23341450e-01 -1.63284048e-01 -6.28290355e-01
7.07494140e-01 -1.53462455e-01 -2.40740404e-01 3.72907817e-01
-2.66544316e-02 -1.46198571e-01 6.59574747e-01 3.41553062e-01
9.88144159e-01 -1.91453286e-02 7.92014956e-01 -1.15219250e-01
6.52138829e-01 4.33333755e-01 5.27708530e-01 6.15014136e-01
3.39998037e-01 5.71186185e-01 7.85776675e-01 -2.06898332e-01
-1.13034272e+00 -5.07079005e-01 -6.43562153e-02 5.04288077e-01
-5.72772920e-01 -7.25094438e-01 -9.73761797e-01 -6.00799680e-01
9.06541594e-04 1.15491259e+00 -7.96878219e-01 -1.30608335e-01
-2.69082457e-01 -1.16256022e+00 6.40456021e-01 2.62605965e-01
2.74119854e-01 -8.80119979e-01 -5.17999530e-01 6.38089836e-01
-2.09073111e-01 -1.15913045e+00 -4.59687293e-01 2.41215490e-02
-9.70764339e-01 -1.11110318e+00 -8.16303611e-01 -5.12067676e-01
9.82841730e-01 -3.80607307e-01 9.40634608e-01 -4.19403464e-01
-3.66747469e-01 -2.81105042e-01 -3.06761861e-01 -9.51674044e-01
-7.87186980e-01 1.53884336e-01 -1.21471919e-01 -9.64704230e-02
1.36063740e-01 -2.35288024e-01 -5.78292131e-01 -2.54035264e-01
-1.10808122e+00 3.63653779e-01 8.22155118e-01 8.60698283e-01
7.40732014e-01 -2.49809504e-01 1.21748412e+00 -1.54044700e+00
1.17522275e+00 -5.76554179e-01 -4.03679490e-01 3.85161072e-01
-8.68254960e-01 3.63073409e-01 1.03779781e+00 -3.34835827e-01
-8.67734134e-01 2.77461618e-01 -1.25516623e-01 1.28718391e-01
1.02694981e-01 8.33886206e-01 4.08418626e-02 3.50082755e-01
9.45158362e-01 4.38632995e-01 2.55559962e-02 -4.48416948e-01
5.38563251e-01 7.72940278e-01 4.49622780e-01 -3.20713282e-01
5.30300200e-01 5.54742245e-03 1.34304568e-01 -5.61951339e-01
-6.24278724e-01 -1.68796703e-01 -4.40606266e-01 9.95198637e-02
6.45771086e-01 -1.05452240e+00 -7.06490695e-01 -5.90858497e-02
-1.22761345e+00 3.00447792e-01 -4.73305345e-01 1.64724559e-01
-3.07457685e-01 3.18891764e-01 -1.83808714e-01 -8.27848911e-01
-1.26074958e+00 -7.80939698e-01 1.11683953e+00 2.72320032e-01
-6.06678367e-01 -4.84128356e-01 3.98744568e-02 3.56241286e-01
2.82338887e-01 8.80327940e-01 1.04338729e+00 -1.14916945e+00
-1.58496022e-01 -3.89154911e-01 -4.14769799e-02 -1.75118044e-01
2.35830590e-01 -1.56562179e-01 -6.55418396e-01 6.03138171e-02
-2.42891341e-01 -2.78170109e-01 6.28578365e-01 3.62219065e-01
8.71090949e-01 -8.16559494e-01 -3.73010784e-01 8.60400647e-02
1.33560002e+00 2.43284792e-01 4.86349225e-01 -1.02461852e-01
5.01401842e-01 7.12378144e-01 6.89347327e-01 7.55930781e-01
4.84120339e-01 4.86908913e-01 -1.47059754e-01 -1.81223422e-01
-1.30121246e-01 -4.26596045e-01 -9.10246745e-02 7.37931073e-01
1.50660604e-01 -3.12085629e-01 -6.28743172e-01 5.86537957e-01
-1.76189077e+00 -8.22338223e-01 -2.57134020e-01 2.15769076e+00
1.54202402e+00 2.05334708e-01 1.31451249e-01 1.11225307e-01
4.16228086e-01 -2.45099142e-01 -6.12873495e-01 -5.43895185e-01
-7.28582144e-02 3.86139631e-01 4.28220153e-01 1.95111409e-01
-7.36614704e-01 7.40481973e-01 6.32166958e+00 8.73817980e-01
-9.48355556e-01 -1.88934922e-01 9.73764598e-01 -1.39022827e-01
-1.99486896e-01 -4.92420107e-01 -8.60410571e-01 5.46060264e-01
1.29939246e+00 -7.34498560e-01 -1.79909214e-01 6.11091673e-01
6.89361751e-01 -1.88628823e-01 -1.15568852e+00 8.94848883e-01
1.99209839e-01 -1.85129833e+00 5.74775100e-01 -1.15074717e-01
8.36440682e-01 -6.24719203e-01 -2.73956984e-01 1.19513065e-01
-1.33575769e-02 -9.47304845e-01 5.59010208e-01 6.76050246e-01
1.06888449e+00 -7.52657235e-01 9.48422551e-01 3.27098250e-01
-8.75120223e-01 2.20222063e-02 -3.00235003e-01 2.82495230e-01
2.94768587e-02 8.19449067e-01 -1.49081683e+00 1.10266221e+00
6.09310679e-02 4.91894871e-01 -7.50343025e-01 9.32307541e-01
1.34221360e-01 4.25862402e-01 -1.54425859e-01 -4.85983819e-01
2.74802395e-03 5.46604022e-03 3.11523825e-01 1.43432462e+00
5.03647983e-01 4.10946012e-01 5.93514787e-03 9.64538813e-01
-3.12304467e-01 6.00260913e-01 -6.71829283e-01 -3.94918948e-01
3.74234170e-01 1.06174314e+00 -4.86623585e-01 -7.51866579e-01
-5.52618615e-02 4.38194722e-01 -9.45699122e-03 7.01364502e-02
-6.25443935e-01 -7.25206971e-01 5.06511554e-02 3.47095788e-01
9.16284993e-02 2.68368363e-01 -8.02549064e-01 -8.52597892e-01
2.69409791e-02 -1.13576031e+00 5.46818435e-01 -7.23218322e-01
-1.01540196e+00 7.70128429e-01 1.11337788e-01 -1.44657600e+00
-4.77149367e-01 -1.42452881e-01 -3.22966188e-01 8.53721440e-01
-1.08283663e+00 -9.10999000e-01 -5.12774065e-02 1.86671853e-01
7.37177193e-01 -3.68098974e-01 7.75418580e-01 2.87885278e-01
-5.63037634e-01 8.22357297e-01 1.34679362e-01 -1.27807871e-01
7.92729616e-01 -1.23663998e+00 4.77972716e-01 6.01556838e-01
-3.41360807e-01 9.70962644e-01 1.08039439e+00 -1.10756803e+00
-1.36746776e+00 -1.43636417e+00 1.35036516e+00 -3.43373924e-01
4.26608354e-01 -4.48760390e-01 -7.69419849e-01 7.10848495e-02
3.50449175e-01 -5.14844835e-01 9.50897872e-01 -3.85619283e-01
1.47559926e-01 -2.20506623e-01 -1.51623595e+00 6.84014440e-01
6.18965626e-01 5.08598424e-02 -6.57248855e-01 7.93373346e-01
8.00750375e-01 -6.42893255e-01 -1.03641760e+00 1.78371459e-01
3.81025136e-01 -2.42354840e-01 6.37822270e-01 -8.63014936e-01
6.85781538e-01 -4.44927573e-01 2.93143183e-01 -1.25086129e+00
5.53719141e-02 -9.12583709e-01 -6.60347715e-02 1.27014387e+00
9.16451216e-01 -4.88714695e-01 7.36403763e-01 6.06859624e-01
-1.71046808e-01 -1.03785074e+00 -8.20142865e-01 -2.62635142e-01
-1.07700229e-01 7.95410573e-02 5.89675844e-01 7.25653827e-01
4.12960649e-01 6.49268746e-01 -3.73228282e-01 -1.44061163e-01
5.71214497e-01 -7.60362148e-02 6.98212445e-01 -1.06112158e+00
5.30604497e-02 -7.32464194e-02 -1.81185558e-01 -4.51028138e-01
-5.28308809e-01 -8.33707154e-01 -8.71961713e-02 -1.87424612e+00
4.56280917e-01 2.13956479e-02 7.01190252e-03 5.30859411e-01
-4.54292297e-01 2.27141213e-02 1.93149060e-01 -1.06954999e-01
-5.84893107e-01 5.16627789e-01 1.18664479e+00 -3.30023050e-01
-3.96377116e-01 -1.54473498e-01 -1.25929832e+00 1.75163269e-01
7.27681637e-01 -7.67233193e-01 -5.35685003e-01 2.19748512e-01
3.68669778e-01 3.90308917e-01 -1.95682749e-01 -9.26439345e-01
3.12490523e-01 -1.21730871e-01 7.20925152e-01 -9.57852721e-01
-1.29006326e-01 -3.75400633e-01 4.59712744e-01 5.78604102e-01
-7.13174939e-01 2.63411641e-01 6.21997774e-01 6.13343179e-01
1.17410123e-02 -1.56513944e-01 3.92866045e-01 -2.09056303e-01
1.52009264e-01 -7.55839795e-03 -2.84725636e-01 1.94232553e-01
8.92894864e-01 -1.61411479e-01 -6.30077958e-01 -1.65201038e-01
-6.15072608e-01 4.69820797e-01 -4.49438915e-02 2.44043931e-01
6.14599168e-01 -1.11034334e+00 -1.14190459e+00 2.13098284e-02
2.21980110e-01 8.77145380e-02 2.83822894e-01 7.96175897e-01
-6.40452683e-01 9.21989381e-01 -2.06332102e-01 -2.78302044e-01
-1.30501294e+00 4.13170695e-01 -1.02565467e-01 -7.92050362e-01
-2.87431926e-01 3.58795166e-01 -2.96449363e-02 -1.99473649e-01
6.16261512e-02 -6.80794954e-01 -5.06062865e-01 3.70644391e-01
6.98154986e-01 2.10451290e-01 4.20523256e-01 -2.17239410e-01
-4.07033056e-01 1.81530863e-01 -4.78120893e-01 -4.74812649e-02
1.57688224e+00 3.23002994e-01 4.93805036e-02 1.44997969e-01
7.62897015e-01 2.02744767e-01 -6.32915378e-01 6.36761561e-02
5.96357435e-02 9.39451456e-02 -2.79711008e-01 -1.41018856e+00
-4.32989776e-01 4.87277508e-01 2.81837791e-01 -4.18398790e-02
1.02631533e+00 -3.09928566e-01 7.25980759e-01 3.98172915e-01
-6.42344207e-02 -1.20526445e+00 -2.14642808e-01 -4.11638469e-02
1.23972142e+00 -1.02941751e+00 6.79947793e-01 -4.47584003e-01
-1.02954006e+00 1.01472461e+00 3.47574621e-01 3.69387835e-01
1.54821828e-01 7.76021332e-02 -2.28041738e-01 -1.76571921e-01
-1.15582609e+00 1.53001353e-01 4.66910988e-01 5.51893115e-01
6.01038814e-01 -4.64783721e-02 -9.20781612e-01 8.72353792e-01
-3.02555531e-01 3.34741235e-01 7.98764646e-01 9.58560050e-01
-3.30222338e-01 -8.92902195e-01 -4.73117769e-01 8.38025272e-01
-7.51272619e-01 -2.34135926e-01 -5.67652047e-01 7.62466550e-01
5.94515987e-02 1.08128941e+00 -4.38771099e-01 -2.01799631e-01
7.62188435e-01 1.54845864e-01 1.46009713e-01 -7.45570362e-01
-9.64673877e-01 2.17981860e-01 4.71221596e-01 -2.16564447e-01
-2.32972130e-01 -6.82136834e-01 -1.31686902e+00 -6.90786690e-02
-8.77849534e-02 4.15337741e-01 6.60887361e-01 6.62724197e-01
8.62166703e-01 8.87779236e-01 3.58761966e-01 -2.62012005e-01
-5.51422119e-01 -1.30528128e+00 -2.17433661e-01 3.05823445e-01
1.64302751e-01 -2.49282017e-01 1.90496877e-01 3.81853789e-01] | [12.233797073364258, 9.520767211914062] |
6070d4fe-8a67-4d8a-8377-4b8ca8e3c33a | locally-weighted-mean-phase-angle-lwmpa-based | 2109.08774 | null | https://arxiv.org/abs/2109.08774v1 | https://arxiv.org/pdf/2109.08774v1.pdf | Locally Weighted Mean Phase Angle (LWMPA) Based Tone Mapping Quality Index (TMQI-3) | High Dynamic Range (HDR) images are the ones that contain a greater range of luminosity as compared to the standard images. HDR images have a higher detail and clarity of structure, objects, and color, which the standard images lack. HDR images are useful in capturing scenes that pose high brightness, darker areas, and shadows, etc. An HDR image comprises multiple narrow-range-exposure images combined into one high-quality image. As these HDR images cannot be displayed on standard display devices, the real challenge comes while converting these HDR images to Low dynamic range (LDR) images. The conversion of HDR image to LDR image is performed using Tone-mapped operators (TMOs). This conversion results in the loss of much valuable information in structure, color, naturalness, and exposures. The loss of information in the LDR image may not directly be visible to the human eye. To calculate how good an LDR image is after conversion, various metrics have been proposed previously. Some are not noise resilient, some work on separate color channels (Red, Green, and Blue one by one), and some lack capacity to identify the structure. To deal with this problem, we propose a metric in this paper called the Tone Mapping Quality Index (TMQI-3), which evaluates the quality of the LDR image based on its objective score. TMQI-3 is noise resilient, takes account of structure and naturalness, and works on all three color channels combined into one luminosity component. This eliminates the need to use multiple metrics at the same time. We compute results for several HDR and LDR images from the literature and show that our quality index metric performs better than the baseline models. | ['Sarwan Ali', 'Abdul Haseeb', 'Inaam Ul Hassan'] | 2021-09-17 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 3.18553686e-01 -4.79921073e-01 2.80190289e-01 -8.92874300e-02
-5.15894055e-01 -6.16793871e-01 4.57560241e-01 -1.75101623e-01
-2.11650342e-01 6.62083328e-01 2.12377924e-02 -2.16235578e-01
6.14207312e-02 -9.36916471e-01 -2.98840970e-01 -6.91772044e-01
1.55109763e-01 -1.75503418e-01 4.99808520e-01 -5.10077953e-01
5.13475947e-03 7.06267476e-01 -1.69422841e+00 2.03968897e-01
9.16290700e-01 9.79370654e-01 4.84995723e-01 7.63552248e-01
-1.68409720e-01 8.16118896e-01 -7.35912800e-01 -2.20604986e-01
5.36853015e-01 -6.38218820e-01 -4.63137060e-01 1.44956902e-01
3.08135271e-01 -4.58251774e-01 -3.84860039e-01 1.29048431e+00
4.55606252e-01 4.36400017e-03 4.64404136e-01 -1.08791828e+00
-8.88965130e-01 -7.92985857e-02 -8.30670774e-01 1.94419324e-01
6.38503134e-01 2.43999451e-01 6.00696385e-01 -7.79242814e-01
6.31080270e-01 1.27725458e+00 2.50723153e-01 2.62386024e-01
-1.29682481e+00 -6.93633616e-01 -3.57854515e-01 2.55738199e-01
-1.43252170e+00 -1.10984951e-01 7.64868259e-01 -1.79107055e-01
5.07341981e-01 6.75635934e-01 6.49057627e-01 6.43693268e-01
2.14029506e-01 1.83687553e-01 1.84634912e+00 -4.06045884e-01
4.25239243e-02 1.89607531e-01 -1.83594093e-01 4.68980640e-01
2.49680951e-01 5.10279477e-01 -2.89111316e-01 3.61674428e-01
7.76893139e-01 1.92702949e-01 -5.91709375e-01 -3.49082612e-02
-9.38198388e-01 3.81770492e-01 4.11028028e-01 3.79315078e-01
5.45701422e-02 -3.17048341e-01 -7.91233853e-02 6.62399709e-01
2.12331504e-01 4.50475514e-01 5.22948205e-02 -8.32424760e-02
-7.17009127e-01 -2.82422483e-01 3.49841177e-01 6.93357885e-01
1.05534244e+00 -1.03710435e-01 -9.14591700e-02 1.16478956e+00
3.02227549e-02 9.94635284e-01 1.49396196e-01 -9.56051052e-01
1.65865377e-01 7.03626096e-01 -3.86422426e-02 -9.73396599e-01
-3.05363417e-01 -1.06013203e-02 -1.11084127e+00 9.02685821e-01
1.83162823e-01 2.04589665e-01 -1.03519225e+00 1.25904930e+00
-1.07912952e-02 -4.94629294e-01 9.37359482e-02 1.12502229e+00
9.31596279e-01 1.11824691e+00 -2.45620117e-01 -2.06350058e-01
1.23455167e+00 -6.12020016e-01 -9.52263594e-01 -1.70641720e-01
-6.80587515e-02 -1.29803085e+00 1.39684951e+00 6.60389364e-01
-1.18373394e+00 -8.42566967e-01 -1.33750272e+00 -2.09334671e-01
-5.68705797e-01 6.62216172e-02 2.11078197e-01 8.66115212e-01
-1.23141921e+00 4.76346195e-01 -7.82564282e-02 -2.50535578e-01
-2.48339176e-01 -1.12189427e-01 -4.72045720e-01 -4.87959296e-01
-1.19792557e+00 1.11997378e+00 1.35512382e-01 -9.07986076e-04
-4.44304347e-01 -3.81659478e-01 -6.13172829e-01 -6.36820570e-02
3.13221693e-01 -1.40699893e-01 6.41923845e-01 -1.03941596e+00
-1.57781839e+00 9.43461716e-01 1.83491245e-01 1.52123228e-01
5.35163224e-01 3.17084119e-02 -9.49371934e-01 3.55139613e-01
-1.94882601e-01 3.68591994e-01 8.93288910e-01 -1.58598423e+00
-5.73931992e-01 -1.07345432e-01 6.73174933e-02 1.27876177e-01
3.99891026e-02 2.20914975e-01 -6.88032508e-01 -5.84309280e-01
1.76879913e-01 -8.29744160e-01 2.95878679e-01 1.86551020e-01
-2.76056617e-01 3.86842161e-01 1.11766398e+00 -7.18528509e-01
1.28944969e+00 -2.41325092e+00 -3.51882488e-01 2.50209302e-01
8.81128535e-02 4.53341514e-01 -1.95631161e-01 2.68411487e-01
-2.67095298e-01 1.73271492e-01 -2.17858136e-01 3.69952589e-01
-2.03354731e-01 6.11482672e-02 -1.40687823e-01 3.49984825e-01
1.10119857e-01 5.32035589e-01 -5.72728515e-01 -5.79753339e-01
5.02974033e-01 7.37770736e-01 1.44116387e-01 3.65754634e-01
2.34753445e-01 2.50099123e-01 1.31039947e-01 6.94650650e-01
1.07662213e+00 9.80080143e-02 -1.50203541e-01 -5.77281296e-01
-3.06816757e-01 -2.32024044e-01 -1.34579456e+00 9.65620935e-01
-5.32573581e-01 9.58635092e-01 -1.69228330e-01 -2.44255051e-01
1.38588154e+00 4.44823354e-02 2.98102081e-01 -1.44435859e+00
-1.45576611e-01 1.93927631e-01 -5.90971075e-02 -4.66966420e-01
6.78217709e-01 -1.38454333e-01 1.08108334e-02 4.89758790e-01
-4.88788337e-01 -3.45219195e-01 3.31585646e-01 1.53637812e-01
8.73938382e-01 -1.66686386e-01 2.03699976e-01 -1.24820154e-02
4.95591849e-01 -4.62878644e-01 3.38420093e-01 6.16918564e-01
-2.02577151e-02 1.11217594e+00 4.51985747e-01 -3.21723580e-01
-1.35439217e+00 -1.40289736e+00 -1.39222383e-01 8.26972306e-01
6.24431074e-01 -1.63264237e-02 -3.02196175e-01 -6.63684309e-02
-2.78899163e-01 4.15170848e-01 -3.86524916e-01 -2.00434938e-01
-4.95714784e-01 -6.96195006e-01 3.02160949e-01 1.02910630e-01
9.35021937e-01 -9.67133164e-01 -5.85897803e-01 -9.56676826e-02
-3.64154428e-01 -9.37894225e-01 -4.67247069e-01 9.64929536e-02
-6.08266950e-01 -9.66744125e-01 -8.71868193e-01 -4.97743160e-01
5.44599295e-01 6.53148115e-01 1.15048432e+00 6.68943673e-02
-6.97451413e-01 9.88571346e-02 -7.80172646e-01 3.64196463e-03
-5.70728838e-01 -5.21774292e-01 -3.37273002e-01 9.26436186e-02
-9.66867723e-04 -1.20695964e-01 -7.18973100e-01 6.60268307e-01
-1.32042122e+00 1.53397426e-01 8.67359281e-01 5.66662133e-01
7.31806159e-01 4.68793839e-01 1.39115229e-01 -6.88630641e-01
4.48781997e-01 1.47281602e-01 -6.54416978e-01 3.87295902e-01
-7.05643177e-01 -7.29918778e-02 6.80873156e-01 -2.56218672e-01
-1.33124709e+00 -1.97262809e-01 7.11944327e-02 -1.50860533e-01
-6.84442967e-02 -1.47721052e-01 -3.74403059e-01 -3.22359502e-01
6.04864657e-01 2.83169150e-01 2.63657793e-02 -5.45338511e-01
2.50213325e-01 8.54392409e-01 6.40055478e-01 -1.29369915e-01
1.13232350e+00 4.80064601e-01 5.48711047e-03 -9.42375183e-01
-4.75113899e-01 -4.13965374e-01 -2.24356204e-01 -3.99294585e-01
7.64030814e-01 -6.64901018e-01 -4.72971290e-01 5.37372649e-01
-7.62962103e-01 -2.70250797e-01 -2.21004933e-01 3.17364484e-01
-1.88173205e-01 4.92217332e-01 -6.44526899e-01 -7.03143299e-01
-1.81544021e-01 -1.04866064e+00 8.09028327e-01 5.22509694e-01
2.46408910e-01 -5.00869691e-01 -1.27093032e-01 6.64450377e-02
6.77207947e-01 1.57623067e-01 1.04570735e+00 3.93089831e-01
-7.03339159e-01 -6.73541520e-03 -6.95646048e-01 5.08176744e-01
3.35236847e-01 3.01548332e-01 -9.01868284e-01 -2.02369928e-01
-1.52200293e-02 2.28659417e-02 8.96404505e-01 2.38714725e-01
1.10892093e+00 7.16350507e-03 1.65714830e-01 6.67946339e-01
1.94004416e+00 6.84504449e-01 1.42287838e+00 5.40173650e-01
4.28543925e-01 4.82224166e-01 9.06624675e-01 -9.09090787e-03
-3.58601063e-02 8.58303964e-01 3.40627879e-01 -1.05849159e+00
-5.93019724e-01 3.12252957e-02 5.78849494e-01 5.03103912e-01
-2.07729831e-01 -2.79448718e-01 -5.17223775e-01 8.21289867e-02
-1.10093188e+00 -9.95149016e-01 -3.09735954e-01 2.29491377e+00
8.05586755e-01 -3.01447287e-02 1.21427901e-01 2.85615683e-01
9.02245998e-01 3.16239074e-02 -3.33823293e-01 -6.62057221e-01
-7.53366947e-01 1.87239945e-01 6.69910908e-01 3.54844123e-01
-8.59077632e-01 5.08284926e-01 6.46071815e+00 7.65215933e-01
-1.38865387e+00 -1.98494062e-01 7.76116192e-01 6.23993203e-02
-3.24316472e-01 -8.10901970e-02 -3.55113208e-01 5.78517079e-01
4.13030684e-01 -8.76885559e-03 6.76588297e-01 4.41819400e-01
3.47278208e-01 -5.96551239e-01 -6.67986929e-01 1.43601501e+00
1.36630312e-01 -6.46788418e-01 -1.05906263e-01 1.10901207e-01
7.40212560e-01 -3.61561090e-01 5.17683625e-01 -1.25702888e-01
1.03679545e-01 -1.04577529e+00 6.75985813e-01 6.10337734e-01
1.24187267e+00 -7.70342708e-01 7.77232766e-01 -3.30572933e-01
-1.27477753e+00 -1.21284751e-02 -6.10940754e-01 2.51461416e-01
1.47947669e-01 6.69053614e-01 -1.97192356e-01 4.36650395e-01
1.05083585e+00 3.13776821e-01 -1.00142074e+00 1.09904921e+00
-1.79476902e-01 7.89134502e-02 -1.51801199e-01 2.53237098e-01
-1.74966708e-01 -5.15756726e-01 3.84042442e-01 1.13318801e+00
4.92220104e-01 1.22191496e-01 -4.40737307e-02 1.00524819e+00
1.30980819e-01 7.84629062e-02 -6.09668612e-01 1.35815933e-01
2.96805114e-01 1.40736973e+00 -8.22568715e-01 -1.62340954e-01
-5.98193884e-01 1.18337953e+00 -4.54676300e-01 4.64476079e-01
-8.13969433e-01 -7.78488040e-01 3.50637764e-01 2.99503714e-01
4.35548723e-02 -9.10700262e-02 -2.75391966e-01 -8.19374144e-01
-3.34215583e-04 -1.04989529e+00 2.28775367e-01 -1.25226891e+00
-1.22411752e+00 8.18463266e-01 -2.01641440e-01 -1.67682540e+00
2.30696440e-01 -5.30157328e-01 -3.58384788e-01 9.28257048e-01
-1.70611727e+00 -7.59002447e-01 -7.36079633e-01 6.00985348e-01
2.17407569e-01 1.55919358e-01 5.54520190e-01 5.66261292e-01
-4.20006335e-01 3.77086312e-01 1.55908078e-01 -6.47805724e-03
9.99623239e-01 -1.35949123e+00 5.19232377e-02 1.06003499e+00
-1.82629198e-01 3.89620692e-01 6.54870152e-01 -4.73943293e-01
-1.21092224e+00 -9.50430095e-01 3.30100030e-01 2.01000385e-02
3.31835710e-02 -1.29186973e-01 -1.04302609e+00 -5.25135212e-02
5.20079471e-02 -2.12038875e-01 4.27639157e-01 -3.21860164e-01
-4.53934789e-01 -6.48030221e-01 -1.20768702e+00 4.84659374e-01
5.84221959e-01 -6.77985549e-01 -2.17516154e-01 -2.46078506e-01
5.32760680e-01 -9.20671672e-02 -9.61262584e-01 2.49233887e-01
6.65179968e-01 -1.54065025e+00 1.03427207e+00 4.77000773e-01
3.61636311e-01 -9.04793620e-01 -1.36719763e-01 -1.15036750e+00
-3.52284461e-01 -4.52110231e-01 5.63338518e-01 1.24449849e+00
4.14547741e-01 -5.42026579e-01 1.44994870e-01 4.12466168e-01
9.99156013e-02 -2.28080168e-01 -4.95769978e-01 -9.82124984e-01
-3.35073054e-01 -5.21770492e-02 6.70268178e-01 8.02588642e-01
-3.11686039e-01 8.87402296e-02 -5.89143813e-01 1.77993737e-02
6.08854473e-01 4.48312908e-01 5.67573965e-01 -1.17311883e+00
-1.62827492e-01 -3.21484476e-01 -3.94703627e-01 -4.25579458e-01
-5.81403196e-01 -4.83172625e-01 4.51076478e-02 -1.61516786e+00
4.75954652e-01 -4.40026730e-01 -1.05466977e-01 4.39286649e-01
-1.50833905e-01 8.67416799e-01 5.14555097e-01 2.40929753e-01
-5.17116189e-01 2.28193879e-01 1.46132684e+00 -2.02462390e-01
-2.86614120e-01 -4.26413506e-01 -5.07969856e-01 5.44007123e-01
7.27134645e-01 -1.82819262e-01 -1.29169419e-01 -1.82479367e-01
1.79342866e-01 -7.61441002e-03 3.76519471e-01 -1.20700681e+00
-5.71124479e-02 -1.89321265e-01 7.90978432e-01 -5.09301603e-01
3.10220718e-01 -8.97294700e-01 7.22145617e-01 5.27534127e-01
-1.37528740e-02 5.22692315e-02 -2.09176511e-01 1.98898792e-01
-3.96918535e-01 -1.65002588e-02 1.48840249e+00 -1.85053140e-01
-1.12818229e+00 5.91940507e-02 -2.80472666e-01 -2.23953903e-01
8.96909654e-01 -7.03667045e-01 -7.06487298e-01 -6.15276158e-01
-3.82657528e-01 -3.77890795e-01 1.08463049e+00 4.41571593e-01
8.10986400e-01 -1.28925681e+00 -4.75928366e-01 3.46157640e-01
2.34176025e-01 -4.09811825e-01 5.39858580e-01 5.11567116e-01
-1.01059544e+00 5.21725677e-02 -6.91068828e-01 -4.24364001e-01
-1.40878749e+00 6.28995121e-01 3.35598707e-01 -6.17197417e-02
-7.89824665e-01 1.89669952e-01 4.56917226e-01 1.55042827e-01
8.21109954e-03 -1.28106937e-01 -3.26620430e-01 -8.16900358e-02
7.60694265e-01 6.60981715e-01 5.51987402e-02 -7.81343877e-01
-1.05664238e-01 1.03132915e+00 1.96973115e-01 -1.24147311e-01
1.09330559e+00 -6.36461556e-01 -2.90401459e-01 3.35482538e-01
1.33576679e+00 1.96686402e-01 -1.03801537e+00 1.31377084e-02
-4.32491094e-01 -9.31722224e-01 8.88853893e-02 -1.01667130e+00
-1.24836195e+00 8.83519650e-01 1.27709663e+00 5.35263121e-01
1.79975414e+00 -1.48392677e-01 6.82376862e-01 -1.43960506e-01
2.63120234e-01 -1.30486906e+00 4.13940161e-01 2.42887482e-01
8.20678353e-01 -1.06613290e+00 -2.27924390e-03 -5.63757420e-01
-7.25380182e-01 1.26759934e+00 5.59956908e-01 2.24941865e-01
2.24710271e-01 3.59433442e-01 3.81327420e-01 3.89523655e-02
-3.10259819e-01 -6.94639623e-01 3.06661785e-01 8.91937256e-01
2.36945719e-01 -5.44251949e-02 -2.47499466e-01 -3.98185179e-02
-6.23304173e-02 -2.54048884e-01 8.62209678e-01 5.06463826e-01
-7.91921377e-01 -1.00345528e+00 -7.20819116e-01 3.23970795e-01
-5.31780660e-01 3.22679840e-02 -4.27007049e-01 6.00960255e-01
2.27565736e-01 1.14278972e+00 4.33038808e-02 -4.55473214e-01
4.33794349e-01 -4.70932037e-01 4.59087044e-01 -1.49099872e-01
-1.02159902e-01 2.32621297e-01 -1.76926881e-01 -7.12748647e-01
-3.63420278e-01 1.15894200e-02 -9.77600992e-01 -5.55241883e-01
-2.16059700e-01 -2.30629534e-01 7.83274710e-01 3.13412726e-01
-5.29088229e-02 5.32403588e-01 9.28194284e-01 -4.45794255e-01
7.67566711e-02 -6.55429900e-01 -1.18659127e+00 7.97189713e-01
4.12375808e-01 -5.01262784e-01 -4.57924694e-01 1.13694862e-01] | [10.913724899291992, -2.39638352394104] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.