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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
3c49634d-e24d-4421-910e-965cf8aa7ded | self-supervised-monocular-scene-flow | 2004.04143 | null | https://arxiv.org/abs/2004.04143v2 | https://arxiv.org/pdf/2004.04143v2.pdf | Self-Supervised Monocular Scene Flow Estimation | Scene flow estimation has been receiving increasing attention for 3D environment perception. Monocular scene flow estimation -- obtaining 3D structure and 3D motion from two temporally consecutive images -- is a highly ill-posed problem, and practical solutions are lacking to date. We propose a novel monocular scene flow method that yields competitive accuracy and real-time performance. By taking an inverse problem view, we design a single convolutional neural network (CNN) that successfully estimates depth and 3D motion simultaneously from a classical optical flow cost volume. We adopt self-supervised learning with 3D loss functions and occlusion reasoning to leverage unlabeled data. We validate our design choices, including the proxy loss and augmentation setup. Our model achieves state-of-the-art accuracy among unsupervised/self-supervised learning approaches to monocular scene flow, and yields competitive results for the optical flow and monocular depth estimation sub-tasks. Semi-supervised fine-tuning further improves the accuracy and yields promising results in real-time. | ['Junhwa Hur', 'Stefan Roth'] | 2020-04-08 | self-supervised-monocular-scene-flow-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Hur_Self-Supervised_Monocular_Scene_Flow_Estimation_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Hur_Self-Supervised_Monocular_Scene_Flow_Estimation_CVPR_2020_paper.pdf | cvpr-2020-6 | ['scene-flow-estimation'] | ['computer-vision'] | [-1.31659895e-01 -5.26811957e-01 -4.55733448e-01 -3.65185648e-01
-5.02189696e-01 -5.79967141e-01 4.65236366e-01 -6.69772923e-01
-3.69098872e-01 7.69696057e-01 4.34495836e-01 -1.70786589e-01
2.53133357e-01 -4.72605526e-01 -6.60471618e-01 -4.25601363e-01
-4.36049774e-02 1.75453559e-01 1.49252802e-01 4.38788146e-01
4.39519256e-01 6.04217649e-01 -1.53726268e+00 3.80910710e-02
7.42010176e-01 1.06243801e+00 1.34342924e-01 1.18007708e+00
-9.66047943e-02 1.42819476e+00 -9.77524742e-02 -6.00630715e-02
6.69219494e-01 -4.06263411e-01 -1.10045016e+00 2.12489948e-01
1.30998850e+00 -1.04843140e+00 -8.79179478e-01 6.95502877e-01
5.92076957e-01 7.26485848e-02 5.41365087e-01 -1.32500827e+00
-4.81433988e-01 -2.12349355e-01 -5.38552761e-01 3.40038240e-01
6.92800164e-01 5.09572983e-01 9.66317475e-01 -8.97735894e-01
8.51774693e-01 1.36550558e+00 4.63955671e-01 5.73703110e-01
-1.09737039e+00 -5.05817652e-01 2.45515302e-01 2.11909533e-01
-1.01588869e+00 -6.02566361e-01 8.28379333e-01 -8.11793208e-01
1.16952062e+00 -2.57076085e-01 6.90102637e-01 9.17066872e-01
1.62274793e-01 1.11171162e+00 9.30029273e-01 -1.16765954e-01
9.00895447e-02 -1.26248181e-01 -2.94033647e-01 9.24525917e-01
2.05863610e-01 4.40506250e-01 -7.05334842e-01 3.60667408e-01
1.26322591e+00 -7.42714256e-02 -4.95016843e-01 -7.94664979e-01
-1.17248642e+00 6.44214094e-01 6.29510999e-01 -6.25690222e-02
7.70622715e-02 6.08266294e-01 3.35793197e-01 3.13825488e-01
7.84967959e-01 5.16420603e-01 -6.64509773e-01 -4.06012505e-01
-9.03691411e-01 3.03210080e-01 6.82926297e-01 1.03870142e+00
1.05735648e+00 3.05981636e-01 9.38699618e-02 5.23043513e-01
1.14313222e-01 5.88753879e-01 1.15215741e-01 -1.68939114e+00
4.99318510e-01 4.88480359e-01 3.32144350e-01 -9.32141602e-01
-5.10990679e-01 -4.42763954e-01 -5.99997163e-01 5.40558815e-01
7.18492448e-01 -2.31365606e-01 -6.44703031e-01 1.66225529e+00
4.39233541e-01 5.43837428e-01 -9.56039876e-02 1.39373267e+00
9.73767340e-01 5.20279408e-01 -3.08380216e-01 -3.41815576e-02
6.90366805e-01 -1.32150710e+00 -4.69104528e-01 -4.61720645e-01
7.65878379e-01 -7.34602809e-01 1.03173721e+00 2.39590153e-01
-1.25482357e+00 -6.86781108e-01 -8.66783023e-01 -6.41175449e-01
1.02407135e-01 2.33310200e-02 8.29212010e-01 5.54365098e-01
-1.07529593e+00 4.79883313e-01 -8.32167149e-01 -1.83083609e-01
5.54621994e-01 1.83888450e-01 -4.95959520e-01 -4.75033581e-01
-7.08738863e-01 7.09289372e-01 -3.78747322e-02 -6.57190606e-02
-8.49739313e-01 -1.12806487e+00 -1.32738590e+00 -1.83485955e-01
2.33707368e-01 -1.28789890e+00 1.12303567e+00 -5.45739532e-01
-1.78341556e+00 1.07359135e+00 -4.72313493e-01 -3.10588807e-01
8.15530598e-01 -6.18799150e-01 1.51696414e-01 5.15585005e-01
3.35217744e-01 1.21418285e+00 8.54896665e-01 -1.09613991e+00
-7.80652881e-01 -2.87640691e-01 2.66180009e-01 3.90111715e-01
3.96379083e-02 -3.88334125e-01 -3.61067802e-01 -2.28759095e-01
2.23604351e-01 -8.80260468e-01 -1.48539677e-01 6.96617782e-01
-2.12390319e-01 1.62970096e-01 7.89364994e-01 -2.89394110e-01
8.22827756e-01 -1.64462268e+00 4.01499793e-02 -6.20687306e-01
4.03566897e-01 1.98121503e-01 1.24045037e-01 -1.14388689e-01
1.23184510e-01 -2.54935801e-01 -2.51160771e-01 -6.13187671e-01
-2.35465020e-01 6.48736358e-02 -2.01755986e-01 7.14058161e-01
4.35366660e-01 1.22314572e+00 -1.24944258e+00 -5.20394146e-01
9.44200337e-01 4.50120658e-01 -1.07722259e+00 5.79031527e-01
-1.45892516e-01 8.47807825e-01 -1.35160610e-01 6.51370347e-01
8.25307846e-01 -6.28393888e-01 -3.23096663e-01 -2.29749709e-01
-1.73411429e-01 3.85785609e-01 -1.20135784e+00 2.23517942e+00
-7.33345151e-01 1.18445194e+00 -3.68363596e-02 -6.38720334e-01
7.16124892e-01 -5.33990934e-03 7.26685643e-01 -6.14265501e-01
6.71020597e-02 2.03806117e-01 -4.05690044e-01 -8.07034254e-01
5.97260356e-01 -2.42286175e-02 2.60911345e-01 4.45709616e-01
2.21726760e-01 -6.51052594e-01 1.31334573e-01 6.71201125e-02
1.04123127e+00 8.20932865e-01 2.87366629e-01 -1.06447779e-01
6.64632320e-01 8.20779521e-03 5.07688761e-01 5.91081142e-01
-6.49175167e-01 9.53661799e-01 3.85078669e-01 -8.02332878e-01
-9.69298124e-01 -1.00884461e+00 1.93125696e-03 5.50343990e-01
5.68510592e-01 5.42278076e-03 -1.99444026e-01 -6.51522458e-01
2.46795788e-01 1.68961823e-01 -5.17548978e-01 1.47779167e-01
-7.24142373e-01 -3.14447582e-01 1.52704358e-01 5.40094793e-01
7.42522418e-01 -9.38750923e-01 -9.87364948e-01 2.86666781e-01
-3.99343848e-01 -1.90111744e+00 -6.27557099e-01 -9.09836069e-02
-8.96452188e-01 -1.16324914e+00 -8.56786609e-01 -6.66514218e-01
3.95459294e-01 7.14933634e-01 1.27168667e+00 -2.84868509e-01
-3.61570001e-01 3.84926111e-01 2.52822619e-02 2.15525106e-01
-2.53484529e-02 1.42693907e-01 -1.34646401e-01 8.10613334e-02
1.04845174e-01 -7.28782713e-01 -1.17188835e+00 3.30583096e-01
-5.91845453e-01 1.23067610e-01 3.48961987e-02 6.53875709e-01
9.35229585e-02 -4.95884389e-01 2.85786718e-01 -6.47127807e-01
-9.80590507e-02 -5.07744923e-02 -8.88443470e-01 -4.05412555e-01
-4.24809307e-01 2.96967924e-02 5.75826287e-01 -1.44805104e-01
-1.31181407e+00 2.99490333e-01 1.10060070e-02 -9.05309737e-01
-3.91250312e-01 -1.18022926e-01 8.66632760e-02 -4.61211026e-01
7.54759371e-01 7.54761994e-02 -4.85066697e-02 -1.25489518e-01
6.37969732e-01 3.07925463e-01 7.59965599e-01 -4.27184463e-01
7.80748487e-01 1.11627924e+00 2.58852750e-01 -7.70429373e-01
-1.27666676e+00 -8.35398078e-01 -1.05144596e+00 -4.98287082e-01
1.11726427e+00 -1.41365647e+00 -9.27580655e-01 6.56454980e-01
-1.39566779e+00 -5.49935579e-01 -1.28235847e-01 7.28806973e-01
-9.04486954e-01 4.37172383e-01 -7.64399290e-01 -7.89800406e-01
-2.28525877e-01 -1.15775144e+00 1.39755285e+00 6.86416700e-02
-6.51270151e-02 -1.31719518e+00 1.80459201e-01 5.09555638e-01
2.88878709e-01 4.02448535e-01 3.00152153e-01 5.54605424e-01
-1.25258720e+00 2.71124393e-01 -7.65195608e-01 3.36768568e-01
2.01218709e-01 -1.39800325e-01 -1.50780404e+00 -2.16631413e-01
-1.90977871e-01 -6.27435327e-01 1.03974283e+00 8.57151091e-01
8.50837231e-01 1.43694788e-01 -4.73450422e-02 1.30674100e+00
1.41069496e+00 2.40825582e-02 4.40416396e-01 1.66368365e-01
1.25511646e+00 7.45690167e-01 4.38510954e-01 5.18095613e-01
5.83095789e-01 5.33340752e-01 4.92197752e-01 -1.64786726e-01
-4.26694721e-01 -2.94209868e-01 1.22523695e-01 5.67725360e-01
2.56269515e-01 -1.55098245e-01 -6.63354039e-01 7.51228929e-01
-1.78663325e+00 -1.05778027e+00 -5.08221015e-02 1.92140770e+00
5.71497619e-01 1.69119000e-01 -7.19081759e-02 5.01422025e-02
1.53533474e-01 4.76730198e-01 -8.33083749e-01 -4.28045727e-02
-2.79626429e-01 -3.72982137e-02 5.32444715e-01 1.03248549e+00
-1.32697356e+00 1.26170766e+00 6.39387846e+00 1.63937926e-01
-1.39658797e+00 -7.73094073e-02 7.91025579e-01 -2.78763503e-01
-1.96621910e-01 -1.24094971e-02 -8.11221123e-01 1.11660212e-01
2.63660342e-01 1.63739637e-01 3.05994660e-01 6.76914215e-01
4.01912153e-01 -3.62219959e-01 -1.31963003e+00 1.61241806e+00
1.74251929e-01 -1.69695497e+00 -1.49358034e-01 1.74300727e-02
1.27342820e+00 2.48864248e-01 -1.27885774e-01 -1.74912408e-01
2.97946125e-01 -8.16598058e-01 6.30608261e-01 3.13621849e-01
1.01930881e+00 -3.75332206e-01 3.92232865e-01 2.03049585e-01
-1.36594069e+00 -7.26015270e-02 -2.27661595e-01 -5.18104672e-01
6.32192135e-01 6.97168529e-01 -3.93097997e-01 4.66344923e-01
7.84449339e-01 1.64743268e+00 -2.79593259e-01 1.23641419e+00
-1.58002764e-01 1.60204127e-01 -1.29738733e-01 2.78212368e-01
3.69640231e-01 -2.00461239e-01 5.42233586e-01 1.16129076e+00
-2.04860009e-02 -1.07893571e-01 1.76808983e-01 1.14389586e+00
-1.28873527e-01 -2.59994447e-01 -9.08430994e-01 4.75615174e-01
1.70630485e-01 1.05355740e+00 -4.14690465e-01 -2.36081138e-01
-5.70175231e-01 1.03766239e+00 4.76849675e-01 4.70533609e-01
-5.07560968e-01 -5.25982939e-02 1.15393519e+00 5.92299402e-02
1.95744604e-01 -5.49471974e-01 -6.09931767e-01 -1.63944888e+00
8.81374627e-02 -1.14792645e-01 9.97876823e-02 -1.16863477e+00
-1.25507581e+00 4.35639590e-01 -3.95489410e-02 -1.49214292e+00
-4.28423285e-01 -8.34298670e-01 -3.66956174e-01 6.29257202e-01
-2.25598073e+00 -7.67501473e-01 -8.42029691e-01 5.06781101e-01
6.26528621e-01 5.22346124e-02 2.36551866e-01 5.08961558e-01
-3.36693645e-01 3.11155885e-01 -3.53451073e-01 9.38560739e-02
8.44138086e-01 -1.19706213e+00 6.77512228e-01 9.61457610e-01
-8.44827816e-02 1.04948729e-01 3.15479904e-01 -3.12554538e-01
-1.51254892e+00 -1.23035502e+00 9.07241881e-01 -7.65678644e-01
5.46309769e-01 -2.67600358e-01 -5.72490394e-01 5.37016332e-01
-4.75875512e-02 5.61026454e-01 1.00007147e-01 -4.60638195e-01
-3.90226752e-01 -1.01912938e-01 -9.98802543e-01 6.17478251e-01
1.71838868e+00 -8.71015429e-01 -8.49589333e-02 1.95821166e-01
9.05329585e-01 -6.66737020e-01 -5.99147975e-01 4.12047803e-01
6.21324062e-01 -1.36267722e+00 1.26964188e+00 -3.42050731e-01
8.92814398e-01 -2.85842866e-01 -3.90655473e-02 -1.06464291e+00
-1.35359198e-01 -9.95426774e-01 -3.67414027e-01 5.06452680e-01
2.55720541e-02 -5.25946558e-01 1.29976594e+00 4.31697667e-01
-1.50819898e-01 -5.85703552e-01 -8.37264240e-01 -7.46126473e-01
-2.40024235e-02 -6.93726242e-01 1.41805723e-01 9.78435934e-01
-2.80013859e-01 4.42546040e-01 -6.50749266e-01 -5.53888902e-02
9.04491365e-01 4.11475629e-01 1.31694388e+00 -1.04949200e+00
-1.11760251e-01 -4.23285633e-01 -5.30931771e-01 -2.08511448e+00
4.31275785e-01 -7.07654178e-01 -6.60607666e-02 -1.47360969e+00
-1.86718270e-01 -2.79404640e-01 2.84040928e-01 1.49896117e-02
-2.58053273e-01 4.26548749e-01 1.33853659e-01 1.81208402e-01
-6.29071593e-01 7.96116173e-01 1.86709547e+00 1.14536569e-01
-3.90885144e-01 -2.61322349e-01 -2.57781953e-01 6.14822805e-01
5.08357406e-01 -1.47193581e-01 -6.76379979e-01 -8.18666697e-01
-2.07609180e-02 3.79157931e-01 5.65203011e-01 -1.03031456e+00
2.49022380e-01 -2.32650325e-01 4.84752983e-01 -7.64826536e-01
5.85304976e-01 -6.21498644e-01 -4.52272892e-01 3.41032863e-01
-2.46088415e-01 -2.40262225e-02 2.70803943e-02 4.42234904e-01
-1.98429614e-01 2.79464751e-01 8.83122206e-01 -2.94356763e-01
-9.18362498e-01 8.47956121e-01 -1.20301079e-02 6.92030132e-01
6.01600647e-01 -4.34338540e-01 -3.46675366e-01 -7.61309624e-01
-3.26169640e-01 2.26226047e-01 4.25252289e-01 4.85043585e-01
7.74548054e-01 -1.23946762e+00 -7.06161022e-01 3.87764633e-01
1.52036563e-01 2.66553521e-01 9.45068970e-02 5.40537119e-01
-1.17467916e+00 6.93596363e-01 -3.13290864e-01 -1.10862195e+00
-7.49776185e-01 3.48999351e-01 6.58234477e-01 -7.15746311e-03
-6.86522424e-01 8.28144491e-01 6.24941707e-01 -6.68411791e-01
2.33868212e-01 -4.99336660e-01 1.32004812e-01 -3.49478573e-01
3.64749968e-01 5.90913892e-01 -2.45878994e-01 -5.81336498e-01
-2.75076389e-01 1.10045850e+00 3.37198794e-01 -1.61980897e-01
1.03755999e+00 -5.63375950e-01 2.36777142e-01 3.67810845e-01
1.83762276e+00 -4.15459633e-01 -2.13359523e+00 -3.93925995e-01
-4.05097693e-01 -1.06697679e+00 2.86610991e-01 -3.23609769e-01
-1.29764140e+00 1.22644031e+00 3.05639982e-01 -3.26821297e-01
9.05765116e-01 -2.00955316e-01 9.50563014e-01 2.82329857e-01
3.48872632e-01 -5.57180941e-01 3.62492561e-01 8.10451627e-01
4.53237802e-01 -1.77504230e+00 2.17334982e-02 -6.73969805e-01
-2.90101707e-01 1.06545055e+00 9.89448190e-01 -3.56173128e-01
7.91893482e-01 9.41144079e-02 3.63821536e-01 3.66589241e-02
-7.68741965e-01 -2.60039985e-01 4.90808755e-01 5.80421329e-01
3.29683900e-01 -2.46981353e-01 4.03254837e-01 -5.31000912e-01
-1.36372671e-01 1.71743453e-01 5.61767280e-01 7.88113177e-01
-1.84212759e-01 -6.80521667e-01 5.10253850e-03 2.62543652e-02
-7.98943564e-02 -1.63114518e-02 -1.07189745e-01 7.10753918e-01
-1.36626840e-01 1.00403512e+00 1.72034338e-01 -1.86220348e-01
1.50218815e-01 -3.34305972e-01 8.41940045e-01 -3.73715758e-01
-1.88832760e-01 6.48995265e-02 1.70430280e-02 -9.85250533e-01
-8.00735235e-01 -4.63592231e-01 -9.89189565e-01 -5.46857595e-01
-2.03885268e-02 -4.46452349e-01 4.04632896e-01 1.00537205e+00
4.02653694e-01 2.34961361e-01 8.82493794e-01 -1.37308526e+00
-1.54944672e-03 -5.22888899e-01 -1.50281265e-01 5.21143079e-01
9.19179142e-01 -7.15528190e-01 -6.87590361e-01 2.76672274e-01] | [8.645561218261719, -1.9650688171386719] |
1b12023d-78c5-46f0-b69e-d80c0d2436b2 | steerable-wavelet-scattering-for-3d-atomic | 1812.02320 | null | http://arxiv.org/abs/1812.02320v2 | http://arxiv.org/pdf/1812.02320v2.pdf | Steerable Wavelet Scattering for 3D Atomic Systems with Application to Li-Si Energy Prediction | A general machine learning architecture is introduced that uses wavelet
scattering coefficients of an inputted three dimensional signal as features.
Solid harmonic wavelet scattering transforms of three dimensional signals were
previously introduced in a machine learning framework for the regression of
properties of small organic molecules. Here this approach is extended for
general steerable wavelets which are equivariant to translations and rotations,
resulting in a sparse model of the target function. The scattering coefficients
inherit from the wavelets invariance to translations and rotations. As an
illustration of this approach a linear regression model is learned for the
formation energy of amorphous lithium-silicon material states trained over a
database generated using plane-wave Density Functional Theory methods.
State-of-the-art results are produced as compared to other machine learning
approaches over similarly generated databases. | ['Xavier Brumwell', 'Paul Sinz', 'Kwang Jin Kim', 'Yue Qi', 'Matthew Hirn'] | 2018-11-21 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 4.36467648e-01 -2.57604085e-02 -1.74329296e-01 -5.44222653e-01
-8.80398333e-01 8.59831853e-05 8.57219934e-01 1.44158214e-01
-4.87269700e-01 7.86748528e-01 1.40793100e-01 1.29098788e-01
-1.30001396e-01 -1.06470239e+00 -5.46448648e-01 -1.30471921e+00
-2.68336892e-01 8.68650377e-01 1.36110127e-01 -2.80234247e-01
2.69601494e-01 7.93042660e-01 -1.42156410e+00 6.87915683e-01
1.08016804e-01 1.10501671e+00 -2.16959119e-01 5.99805832e-01
1.94427326e-01 4.05865520e-01 -1.90511808e-01 -1.13491580e-01
2.05560878e-01 -3.16609859e-01 -4.59256798e-01 -1.73215479e-01
1.17687412e-01 3.24068904e-01 -4.94849771e-01 6.05612099e-01
3.86756927e-01 -3.08810733e-03 1.62706769e+00 -7.36180663e-01
-7.18700767e-01 3.53897572e-01 -1.44332916e-01 7.36335441e-02
5.69376469e-01 -1.74531534e-01 1.14607680e+00 -1.12697744e+00
6.68851495e-01 1.18106890e+00 7.13627815e-01 5.31648993e-01
-1.76848519e+00 -3.85596067e-01 -5.90773761e-01 3.32245350e-01
-1.15700245e+00 -4.02580947e-01 1.07464087e+00 -6.14650309e-01
1.39496815e+00 2.47987896e-01 7.93922603e-01 1.26537943e+00
7.13295996e-01 2.18131617e-01 1.10383165e+00 -8.38924825e-01
3.29887211e-01 7.84186050e-02 1.97982162e-01 7.91344762e-01
6.00947976e-01 3.13460618e-01 -7.70883262e-01 -4.22695965e-01
4.15672690e-01 -5.17894588e-02 -3.23236687e-03 -8.31881046e-01
-1.29219770e+00 1.17443609e+00 3.02769750e-01 5.24743438e-01
-6.16944611e-01 3.10331713e-02 4.08530116e-01 4.49198395e-01
6.55355573e-01 3.31873000e-01 -5.14980137e-01 1.78577781e-01
-8.51549029e-01 2.61884630e-01 9.47694838e-01 4.85603988e-01
8.74829650e-01 3.05755913e-01 2.89505005e-01 3.06218386e-01
6.11236870e-01 7.05442548e-01 6.17487371e-01 -5.89221001e-01
-4.48070355e-02 3.84145111e-01 9.56588797e-03 -5.76157033e-01
-5.80322385e-01 9.58841816e-02 -9.68413115e-01 2.90490478e-01
2.49081962e-02 2.14610010e-01 -9.51552927e-01 1.12586784e+00
1.80059046e-01 -7.27567524e-02 4.55622286e-01 5.04645944e-01
9.11098719e-01 7.86585748e-01 -2.64681339e-01 -6.96620584e-01
9.40107465e-01 -2.37860739e-01 -5.49857974e-01 3.46918613e-01
2.72388667e-01 -6.37944043e-01 4.71467704e-01 5.66164494e-01
-1.19621539e+00 -3.94911081e-01 -1.30761206e+00 4.63835616e-03
-5.40725589e-01 -1.75706565e-01 6.24573112e-01 6.64370954e-01
-7.03588247e-01 8.51236343e-01 -1.21230471e+00 -1.50348647e-02
2.12159336e-01 6.49236381e-01 -4.84628618e-01 1.53016523e-01
-9.47345495e-01 1.00779331e+00 1.11810312e-01 -1.86062917e-01
-6.97956026e-01 -8.53308618e-01 -7.15361297e-01 -3.71358246e-01
-5.54705024e-01 -5.97330868e-01 1.01400876e+00 -6.03012502e-01
-1.81152737e+00 1.02748311e+00 -2.96860605e-01 -5.00515997e-01
1.83277428e-01 3.76138330e-01 -6.90055072e-01 2.28305787e-01
-7.03481883e-02 6.14582859e-02 1.25630844e+00 -1.24735248e+00
-1.20454049e-02 -3.96276057e-01 -6.12740040e-01 -4.74911779e-01
-1.34478882e-01 -2.74810106e-01 5.72374344e-01 -4.21102464e-01
5.28044522e-01 -7.23590136e-01 -1.88562274e-01 -4.26351726e-01
-1.23748496e-01 -2.69582093e-01 9.21672761e-01 -2.78413147e-01
6.93518221e-01 -1.85727680e+00 6.87539816e-01 5.73043346e-01
-3.61132175e-02 -1.86961830e-01 1.02077596e-01 8.99198055e-01
-4.08055335e-01 -1.61323801e-01 -5.01416385e-01 -5.75291216e-02
-6.96463734e-02 9.42272618e-02 -2.26343483e-01 9.48989093e-01
3.07862312e-01 5.06906807e-01 -4.36156332e-01 -4.72779945e-02
2.22638905e-01 8.56820643e-01 -6.09059989e-01 -5.52032776e-02
-2.21873745e-01 3.90063047e-01 -5.64739764e-01 6.42273962e-01
4.90046114e-01 -1.47679016e-01 -6.77133501e-02 -4.85453278e-01
-6.53290004e-02 3.13204437e-01 -7.79590905e-01 1.52443290e+00
-3.95337433e-01 1.95023000e-01 -1.78321391e-01 -1.69481158e+00
1.47504508e+00 5.32863021e-01 1.07835162e+00 -6.35313392e-01
-6.23054206e-02 4.57325816e-01 1.34238735e-01 -4.71542865e-01
-1.63781628e-01 -9.63125825e-01 -5.47584817e-02 3.02630275e-01
4.23946530e-01 -6.40380263e-01 -3.40280421e-02 -1.75487682e-01
1.13846970e+00 1.99093401e-01 3.63474607e-01 -5.92950523e-01
8.83250475e-01 -9.05027241e-02 -3.83765772e-02 2.78898597e-01
4.00705904e-01 5.22128820e-01 9.39545929e-02 -1.19052267e+00
-1.46333766e+00 -1.16595721e+00 -6.78218186e-01 5.01296282e-01
-2.98095226e-01 -4.95597780e-01 -3.93445462e-01 -1.90210015e-01
1.62660480e-01 4.59095955e-01 -7.47161150e-01 -3.90733242e-01
-6.47153795e-01 -1.13935590e+00 1.03295527e-01 1.61500394e-01
5.97957708e-02 -1.26653242e+00 -4.84130025e-01 3.09158266e-01
4.49763000e-01 -1.01357591e+00 3.07780117e-01 5.80953956e-01
-9.80844080e-01 -1.28396404e+00 -4.60787624e-01 -5.85345268e-01
4.75159049e-01 -1.76869109e-01 1.03085256e+00 -3.41374189e-01
-7.65024662e-01 6.07883394e-01 -7.04744160e-02 -6.03445947e-01
-9.03823733e-01 -1.26379386e-01 6.12304866e-01 7.69461244e-02
7.31143296e-01 -1.02034938e+00 -3.96351367e-01 3.88815664e-02
-7.42265761e-01 -4.29334909e-01 5.91899455e-01 8.46636713e-01
8.54365766e-01 -1.16593473e-01 4.47220147e-01 -6.18073285e-01
4.49566424e-01 -2.59057254e-01 -5.54047823e-01 -1.65463090e-01
-5.95342338e-01 5.09495974e-01 7.05040872e-01 -3.76278728e-01
-7.42933452e-01 2.90990919e-01 -2.02587858e-01 -1.14845052e-01
-1.55947864e-01 3.55065525e-01 2.94553846e-01 -3.66868496e-01
9.43779707e-01 3.89893174e-01 1.76900804e-01 -3.70382786e-01
7.42608756e-02 5.23200750e-01 2.14220598e-01 -9.00297105e-01
9.55191255e-01 7.64490545e-01 7.94297814e-01 -1.46235740e+00
-4.57988679e-01 -3.34941804e-01 -9.92066622e-01 3.32959741e-02
8.66224825e-01 -7.93347120e-01 -8.25964212e-01 8.98830518e-02
-1.04451656e+00 1.22131951e-01 -6.47775948e-01 8.42313707e-01
-1.11392999e+00 2.57814229e-01 -3.48458409e-01 -7.37050116e-01
-4.01923448e-01 -9.74997163e-01 1.13397491e+00 -3.79830360e-01
-1.63714573e-01 -1.08833218e+00 6.06355786e-01 1.36437528e-02
3.16969097e-01 4.73565131e-01 1.40535688e+00 -6.75808609e-01
-3.71358216e-01 -5.24635136e-01 4.59797263e-01 3.88428599e-01
1.64288685e-01 -2.59142704e-02 -9.01920378e-01 -4.30939227e-01
4.05773401e-01 -3.51433992e-01 1.20958006e+00 6.57825887e-01
7.01424956e-01 -1.85564682e-01 -3.98261607e-01 5.03816128e-01
1.45376086e+00 1.66711852e-01 5.95750511e-01 3.65280807e-01
2.61245370e-01 3.13173652e-01 1.49849966e-01 4.54872429e-01
-4.15783167e-01 6.37428522e-01 2.88401157e-01 3.97477672e-02
1.34495661e-01 1.39366180e-01 6.44092679e-01 9.06071663e-01
-7.94599473e-01 2.82551229e-01 -7.97039151e-01 1.51308551e-01
-1.38459122e+00 -1.28525460e+00 -2.08970383e-01 2.15532327e+00
6.75625086e-01 1.99321046e-01 2.16734573e-01 5.04692972e-01
2.69993037e-01 3.01090807e-01 -4.97037619e-01 -4.08290148e-01
-1.66705906e-01 1.03220797e+00 3.78906816e-01 6.49278224e-01
-9.40644920e-01 4.60483909e-01 7.78686857e+00 4.80241150e-01
-1.17516577e+00 -1.32115791e-03 9.27064046e-02 -1.24899168e-02
-3.35594267e-01 -2.15981841e-01 -7.41597950e-01 1.99882090e-01
1.24303806e+00 -5.28848246e-02 2.35659242e-01 4.57543910e-01
1.33089885e-01 3.04896742e-01 -1.25664592e+00 8.97352755e-01
-3.11071966e-02 -1.55786908e+00 2.78496295e-01 1.27002746e-01
5.26372135e-01 1.37420580e-01 5.96242324e-02 1.84498914e-02
-1.47751406e-01 -1.13588512e+00 4.26391155e-01 8.43526840e-01
7.74601758e-01 -7.90107906e-01 4.16410953e-01 1.66491672e-01
-9.36090648e-01 1.09094717e-01 -5.74858069e-01 -4.95057888e-02
2.15074103e-02 7.26279914e-01 -7.27240145e-01 3.52213323e-01
5.35057783e-01 8.49387228e-01 -4.06645089e-02 4.43659484e-01
2.41555914e-01 5.40476024e-01 -3.99877101e-01 -1.86811134e-01
1.41741350e-01 -6.76191032e-01 5.82319081e-01 1.20800388e+00
5.88577151e-01 1.45307571e-01 -6.11366779e-02 6.74532235e-01
8.61031655e-03 7.10190460e-02 -9.72164214e-01 -1.17811896e-01
-1.82803929e-01 1.16222739e+00 -5.54301083e-01 2.18730532e-02
-5.96117318e-01 5.26552022e-01 -1.26495391e-01 4.00158167e-01
-2.84228832e-01 8.45701527e-03 6.11457527e-01 6.44776464e-01
6.65899396e-01 -5.52617729e-01 1.72623023e-01 -1.08969903e+00
-1.27951100e-01 -5.13620079e-01 5.49149215e-02 -4.23035771e-01
-1.52029383e+00 4.77766693e-01 2.61064976e-01 -8.46694529e-01
-2.95880020e-01 -1.26204562e+00 -5.62087953e-01 7.22924650e-01
-1.47720909e+00 -1.06848168e+00 2.27548420e-01 7.32271791e-01
3.01317543e-01 -8.69985402e-01 1.51029944e+00 -2.89714366e-01
1.38893157e-01 -1.55942872e-01 4.26352501e-01 -3.30088586e-01
4.70649958e-01 -1.21102726e+00 2.55611956e-01 4.07349244e-02
4.81329441e-01 3.49783659e-01 9.62388098e-01 -4.29003537e-01
-1.74910128e+00 -6.47015929e-01 5.54602265e-01 -2.34983370e-01
7.59548903e-01 -4.20010060e-01 -6.32176340e-01 5.83561540e-01
1.53890550e-01 4.15469438e-01 9.88668382e-01 -1.23952791e-01
-4.11685914e-01 -4.72540502e-03 -1.27981055e+00 -8.17587078e-02
6.48705840e-01 -6.54414773e-01 -6.32729948e-01 7.83388317e-01
2.44211346e-01 -5.01940101e-02 -1.30227852e+00 5.41679382e-01
7.20445693e-01 -1.03085947e+00 1.48654389e+00 -9.62943435e-01
3.69300187e-01 8.37779194e-02 -5.66988349e-01 -1.07010162e+00
-4.18251783e-01 -7.54740238e-01 -3.56152833e-01 1.66063994e-01
5.57364225e-01 -5.25994956e-01 9.67856646e-01 -1.56592522e-02
3.84757556e-02 -8.62087607e-01 -1.54502356e+00 -7.03150451e-01
3.56399566e-01 -7.13150129e-02 4.40589115e-02 7.60460258e-01
-9.51869786e-03 7.26862609e-01 -7.65985176e-02 7.18907416e-02
7.32946038e-01 1.26620620e-01 1.98684588e-01 -1.92667937e+00
-1.63142964e-01 -8.71174037e-02 -9.50437844e-01 -3.96365196e-01
5.57984591e-01 -1.20329428e+00 -4.32428420e-01 -1.14040422e+00
2.27415636e-01 -2.06255764e-01 -3.16148728e-01 1.23651743e-01
7.19378471e-01 4.14965063e-01 -1.96210310e-01 4.06889081e-01
-5.45892632e-04 6.81264222e-01 8.49055886e-01 -3.23074877e-01
-1.04097791e-01 3.97325337e-01 -8.97097886e-02 8.13554764e-01
7.38732874e-01 -7.04196513e-01 1.42938104e-02 4.27669168e-01
3.89734834e-01 -4.67215246e-03 2.75774986e-01 -1.17839050e+00
-1.16522826e-01 1.05643943e-01 5.73959887e-01 -4.62052077e-01
8.42396319e-01 -1.05413520e+00 1.89848199e-01 7.72819996e-01
-1.69557154e-01 -4.67093550e-02 -6.48581758e-02 6.53844059e-01
-3.15607786e-01 -3.98571879e-01 1.01089036e+00 -3.54488581e-01
-4.79519973e-03 3.17843646e-01 -4.66045618e-01 -3.75790358e-01
9.83879805e-01 -2.71912575e-01 1.45353869e-01 -2.76627690e-01
-1.02534127e+00 -7.34939277e-01 3.41409475e-01 -1.54456094e-01
8.93495023e-01 -1.41872120e+00 -8.56761992e-01 5.85034430e-01
6.93087815e-04 -2.98076272e-01 -3.25106770e-01 8.10503423e-01
-4.82541114e-01 5.79337776e-01 -3.95251840e-01 -6.93157554e-01
-1.33072519e+00 5.57262063e-01 5.09768724e-01 -2.26586968e-01
-7.40893960e-01 5.03787041e-01 -2.98867553e-01 -2.75340885e-01
-3.31495404e-01 -3.40242565e-01 6.51245750e-03 5.00116535e-02
1.93912819e-01 9.94700417e-02 5.04791379e-01 -9.41479743e-01
-3.20603222e-01 1.02051115e+00 2.28567179e-02 -8.12961832e-02
2.04835582e+00 6.26105428e-01 -2.21561179e-01 6.48991764e-01
1.58234572e+00 -8.40364210e-03 -8.88495505e-01 -2.40228504e-01
3.41792312e-03 6.86829910e-02 7.11479336e-02 -3.86600852e-01
-6.18290067e-01 9.44537282e-01 6.84953511e-01 5.36000609e-01
7.63173461e-01 1.97867259e-01 2.98294783e-01 9.97388780e-01
3.46457332e-01 -7.09349453e-01 2.71860421e-01 2.75486112e-01
1.02845240e+00 -1.22755194e+00 4.49136257e-01 -3.58460754e-01
-1.54982015e-01 1.84736335e+00 -3.77898961e-01 -5.55558026e-01
9.05044377e-01 4.44020957e-01 -4.74267006e-01 -5.35499156e-01
-4.34830070e-01 1.74050719e-01 3.40834469e-01 9.05801535e-01
7.69337416e-01 -6.38043135e-02 -2.40748808e-01 1.94539919e-01
-2.23653525e-01 -2.16973886e-01 5.00357807e-01 8.99921477e-01
-6.11631036e-01 -1.54606891e+00 -3.51992816e-01 3.69474202e-01
-3.78378540e-01 -6.36414886e-02 -4.48827863e-01 8.74405682e-01
-2.64440536e-01 4.39714789e-01 -7.09953681e-02 -8.18730667e-02
5.25491416e-01 6.54187202e-01 1.08951795e+00 -5.57651579e-01
-1.15079634e-01 7.63086826e-02 -1.21507555e-01 -3.27083409e-01
-1.11551476e+00 -7.91453898e-01 -1.32503235e+00 4.11469601e-02
-1.56882405e-01 3.30234140e-01 9.61508811e-01 8.21643472e-01
-7.40680099e-02 4.14349705e-01 6.65103555e-01 -1.39288247e+00
-6.11276865e-01 -9.03533936e-01 -8.02887082e-01 3.07710439e-01
6.63715422e-01 -7.30568051e-01 -5.04698694e-01 1.70960218e-01] | [5.339661121368408, 5.266302108764648] |
ea42c26c-6a31-4f3b-be6d-5045e462aa5e | n-singer-a-non-autoregressive-korean-singing | 2106.15205 | null | https://arxiv.org/abs/2106.15205v2 | https://arxiv.org/pdf/2106.15205v2.pdf | N-Singer: A Non-Autoregressive Korean Singing Voice Synthesis System for Pronunciation Enhancement | Recently, end-to-end Korean singing voice systems have been designed to generate realistic singing voices. However, these systems still suffer from a lack of robustness in terms of pronunciation accuracy. In this paper, we propose N-Singer, a non-autoregressive Korean singing voice system, to synthesize accurate and pronounced Korean singing voices in parallel. N-Singer consists of a Transformer-based mel-generator, a convolutional network-based postnet, and voicing-aware discriminators. It can contribute in the following ways. First, for accurate pronunciation, N-Singer separately models linguistic and pitch information without other acoustic features. Second, to achieve improved mel-spectrograms, N-Singer uses a combination of Transformer-based modules and convolutional network-based modules. Third, in adversarial training, voicing-aware conditional discriminators are used to capture the harmonic features of voiced segments and noise components of unvoiced segments. The experimental results prove that N-Singer can synthesize a natural singing voice in parallel with a more accurate pronunciation than the baseline model. | ['Hoon-Young Cho', 'Young-Ik Kim', 'Min-Ji Lee', 'Hanbin Bae', 'Tae-Woo Kim', 'Gyeong-Hoon Lee'] | 2021-06-29 | null | null | null | null | ['singing-voice-synthesis'] | ['speech'] | [-3.71811867e-01 -3.90044481e-01 2.52444923e-01 1.57210782e-01
-1.08214700e+00 -7.24443674e-01 3.61509137e-02 -6.67984903e-01
-3.07287369e-02 5.00243008e-01 4.22915250e-01 -1.00007527e-01
5.48189998e-01 -5.59087336e-01 -5.05343080e-01 -7.98010945e-01
7.21550658e-02 -3.85114141e-02 -9.51597989e-02 -3.42311382e-01
-3.46419483e-01 3.37792546e-01 -1.45146561e+00 3.57280165e-01
8.85579586e-01 8.61018062e-01 2.45201454e-01 1.23697841e+00
3.94145429e-01 4.42747295e-01 -8.60608399e-01 -1.87636390e-01
2.96588302e-01 -9.90015686e-01 -1.89834878e-01 -2.98973173e-01
3.57307166e-01 -4.65093255e-01 -4.96438861e-01 9.92486000e-01
1.13797140e+00 1.65779695e-01 6.18085861e-01 -7.60283232e-01
-4.97515112e-01 8.80180836e-01 3.45856994e-01 -1.44587055e-01
1.82584658e-01 6.95042491e-01 1.02825570e+00 -1.15035307e+00
1.29006878e-01 1.16474164e+00 8.13089430e-01 1.01746953e+00
-9.47548270e-01 -1.06734478e+00 -5.57778180e-01 9.96692851e-02
-1.16342700e+00 -6.77283525e-01 1.26449311e+00 -2.03923523e-01
7.53541768e-01 6.18380547e-01 6.63381279e-01 1.18872476e+00
-6.11792915e-02 6.72015190e-01 7.50668705e-01 -3.70316774e-01
2.64896285e-02 -1.42357633e-01 -5.59879601e-01 2.49061182e-01
-6.32545352e-01 5.52002013e-01 -5.03849685e-01 1.10041630e-02
8.02018940e-01 -5.10392666e-01 -4.90988374e-01 3.24825525e-01
-1.14963400e+00 4.85759407e-01 4.35842335e-01 5.68234205e-01
-4.06810522e-01 1.88395143e-01 5.39568782e-01 2.69010335e-01
1.28913328e-01 7.61532426e-01 -1.32958695e-01 -2.17856780e-01
-1.29447663e+00 5.47508657e-01 7.67173409e-01 5.27453542e-01
1.25643626e-01 1.12026227e+00 -3.49624097e-01 1.04159629e+00
1.09818570e-01 6.65925443e-01 9.15498197e-01 -1.03055441e+00
2.83187270e-01 -2.93959737e-01 -4.17726971e-02 -7.08823800e-01
-7.10222200e-02 -6.82749391e-01 -9.40214992e-01 1.69898480e-01
1.48340911e-01 -4.16718602e-01 -6.81859791e-01 1.95185995e+00
4.96105738e-02 6.07331872e-01 2.73697823e-01 1.33863688e+00
8.04019749e-01 9.51979399e-01 -1.63894221e-01 -4.44865137e-01
9.90486741e-01 -1.33823824e+00 -1.09578204e+00 2.25629583e-01
-1.87491566e-01 -1.13481653e+00 1.47666717e+00 5.00403464e-01
-1.41356289e+00 -1.23561430e+00 -1.11514115e+00 -3.66931334e-02
2.60622636e-03 4.44021881e-01 -2.12448999e-01 7.25681484e-01
-1.00109112e+00 8.52882981e-01 -3.80203396e-01 3.72985005e-01
-1.66408122e-01 1.16952710e-01 -1.16400525e-01 6.00843608e-01
-1.58919430e+00 3.23703051e-01 3.36768419e-01 1.92577094e-01
-1.59245932e+00 -8.86092067e-01 -7.26774037e-01 2.23644644e-01
-9.95855927e-02 -7.21531212e-01 1.67586493e+00 -1.02548075e+00
-2.19956326e+00 2.34559789e-01 -3.55278440e-02 -3.90044421e-01
4.22077715e-01 -2.46879414e-01 -8.87820780e-01 1.15361981e-01
-2.42474347e-01 2.80260921e-01 1.27547157e+00 -1.21902430e+00
-3.44747156e-01 -1.19768055e-02 -3.87166798e-01 3.98208022e-01
-4.59750205e-01 2.37740832e-03 1.53189630e-03 -1.20494163e+00
-2.18581900e-01 -7.27725685e-01 -1.84068397e-01 -5.66379905e-01
-6.35220230e-01 7.19836331e-04 8.62322807e-01 -9.75221276e-01
1.53249037e+00 -2.33978009e+00 8.60329196e-02 -2.47823566e-01
-3.42122503e-02 9.57150817e-01 -3.47341210e-01 4.41492617e-01
-1.49977088e-01 9.99390706e-02 -2.03209430e-01 -5.20633340e-01
-6.98984936e-02 -5.71098737e-02 -8.17491055e-01 -1.59590375e-02
3.61681074e-01 8.08256447e-01 -9.59653437e-01 -1.84573352e-01
3.10861170e-01 6.14416361e-01 -6.87536001e-01 7.35999048e-01
-2.80048013e-01 6.68386936e-01 4.03121673e-02 4.27522302e-01
4.57153797e-01 8.52731645e-01 -2.64439937e-02 -1.85500294e-01
-1.37601092e-01 7.79414952e-01 -1.05461180e+00 1.47112703e+00
-8.67468357e-01 2.59886265e-01 6.23151064e-01 -5.06301939e-01
1.08324695e+00 9.53524828e-01 -4.36593369e-02 -2.31434748e-01
7.57681951e-02 7.59920418e-01 2.22078219e-01 -4.32027727e-01
6.76860154e-01 -6.56837821e-01 6.90956712e-02 9.00795534e-02
2.46811762e-01 -6.43112361e-01 -4.44650710e-01 -4.50660616e-01
6.76121354e-01 -2.12806705e-02 -2.02716924e-02 5.73004074e-02
7.91614413e-01 -4.21641290e-01 7.26224244e-01 1.95872277e-01
-7.43590891e-02 1.04646230e+00 -5.21637723e-02 1.02419563e-01
-1.16044116e+00 -1.28762543e+00 1.41670972e-01 8.53233695e-01
-3.19935024e-01 -3.18182707e-01 -1.16699970e+00 -2.29754373e-01
-2.81601787e-01 9.33692813e-01 2.09177826e-02 -3.46134275e-01
-1.08254564e+00 3.06812059e-02 1.33002687e+00 5.56628227e-01
2.72087753e-01 -1.55035532e+00 9.49380621e-02 5.59061527e-01
-2.73000568e-01 -7.24394262e-01 -1.31730974e+00 1.27815548e-02
-6.73230648e-01 -6.07523620e-01 -1.08627295e+00 -1.07712734e+00
-4.38527241e-02 7.97121525e-02 8.44974756e-01 -2.86817431e-01
6.36699945e-02 -1.83279872e-01 -2.04969481e-01 -3.48149866e-01
-1.06496096e+00 9.81383608e-04 5.87452710e-01 1.13499619e-01
-9.75449160e-02 -9.55748856e-01 -6.41502976e-01 3.56529057e-01
-8.06739688e-01 -2.60844231e-01 2.34391704e-01 9.81862485e-01
7.21056402e-01 6.67088926e-02 1.12559450e+00 -3.61067384e-01
1.01081467e+00 -2.34313115e-01 -3.10799509e-01 -2.60095835e-01
-2.36457929e-01 -4.09836411e-01 1.67529809e+00 -8.45510006e-01
-8.06846321e-01 -1.00316614e-01 -8.02751720e-01 -1.14995420e+00
-1.59172982e-01 2.26956859e-01 -7.49202907e-01 5.16170800e-01
8.05770278e-01 5.27201355e-01 3.33358236e-02 -7.70373642e-01
4.50688094e-01 1.23362744e+00 1.12919044e+00 -3.15648228e-01
9.55468893e-01 -1.94690183e-01 -4.52934176e-01 -1.00413227e+00
-4.02219117e-01 -3.86162966e-01 -3.19210470e-01 -1.49187014e-01
7.00593352e-01 -1.06298196e+00 -7.06674516e-01 1.00451779e+00
-1.21684873e+00 -3.44003975e-01 -6.54640734e-01 7.85656691e-01
-9.25187230e-01 3.26750696e-01 -9.49192703e-01 -9.85281706e-01
-8.93142343e-01 -1.14303422e+00 7.02626765e-01 3.81397963e-01
-4.67641791e-03 -5.83243966e-01 1.91191375e-01 4.03459728e-01
6.60559535e-01 4.93243486e-02 6.56230628e-01 -6.19119823e-01
-1.97850749e-01 -9.79297087e-02 5.92366099e-01 1.27209961e+00
4.00346905e-01 -3.14392964e-03 -1.41522038e+00 -2.56747395e-01
2.87971944e-01 -2.94585109e-01 6.94333375e-01 3.03045720e-01
9.52811301e-01 -7.48125076e-01 3.84244263e-01 6.80071950e-01
8.65999222e-01 4.39711362e-01 6.09537244e-01 -6.63344383e-01
8.25770497e-01 2.59166896e-01 5.66467822e-01 1.06761754e-01
9.61888023e-03 6.77597702e-01 3.39503199e-01 -4.39674258e-02
-7.77054012e-01 -9.34359074e-01 8.10463607e-01 1.83393633e+00
-1.45760760e-01 -2.37457037e-01 -2.32211933e-01 8.81258845e-01
-1.21181560e+00 -1.13302743e+00 -4.55975309e-02 2.36324883e+00
1.20011675e+00 -1.48262680e-01 4.61640507e-01 5.92342257e-01
8.79905999e-01 3.94548476e-01 -6.74325585e-01 -6.36716425e-01
-9.74863768e-02 6.58387303e-01 -9.24599767e-02 5.58467031e-01
-8.34789455e-01 1.19824684e+00 5.74411201e+00 1.28879702e+00
-1.54257953e+00 1.73050210e-01 2.80940562e-01 -7.79116452e-02
-5.09429634e-01 -2.82003284e-01 -6.01969361e-01 4.74705219e-01
1.13080561e+00 -7.98676834e-02 9.64801610e-01 8.95532250e-01
5.73422849e-01 8.49100411e-01 -8.59047353e-01 7.46162057e-01
3.21225412e-02 -1.14379287e+00 1.84643090e-01 -3.18317443e-01
6.05803549e-01 -4.47957933e-01 4.07910943e-01 4.86046046e-01
-1.56340569e-01 -1.35793066e+00 1.13915241e+00 5.36169350e-01
1.28587449e+00 -1.06527674e+00 4.70991760e-01 5.55297375e-01
-1.28227973e+00 4.35746945e-02 -2.52924353e-01 -5.08469380e-02
2.80226737e-01 3.06586474e-01 -1.16524565e+00 6.19458914e-01
3.53687316e-01 2.94648498e-01 2.75727272e-01 9.14642453e-01
-4.34321254e-01 1.36570835e+00 -1.06453372e-03 3.43675129e-02
1.06436431e-01 -2.09503435e-02 9.72822607e-01 1.21224272e+00
5.56013525e-01 -7.49195460e-03 -8.51842836e-02 1.06858492e+00
-3.13444048e-01 2.41741896e-01 -3.95325243e-01 -3.76676440e-01
6.22900605e-01 1.29125893e+00 3.02264988e-01 -7.49249160e-02
2.24632502e-01 9.86584008e-01 -1.97772890e-01 3.28367293e-01
-8.33872795e-01 -6.63545072e-01 9.07570064e-01 2.89967149e-01
2.80462861e-01 -4.55223620e-01 8.95494595e-02 -9.11152542e-01
-1.64479718e-01 -1.34226155e+00 -2.49276653e-01 -8.73838425e-01
-1.20329475e+00 9.18735206e-01 -7.84709096e-01 -1.54098237e+00
-6.73154294e-01 -4.36968297e-01 -1.18283355e+00 1.40116060e+00
-1.45875025e+00 -1.30045819e+00 3.70419845e-02 6.46039963e-01
9.35609221e-01 -3.98994595e-01 1.12806940e+00 4.80892152e-01
-3.54326099e-01 8.70334148e-01 -5.74285425e-02 4.23618257e-01
8.97547424e-01 -1.29696131e+00 6.61818147e-01 7.70895779e-01
2.75883615e-01 4.21442062e-01 6.77135646e-01 -2.83966482e-01
-1.19642603e+00 -1.28882146e+00 9.21288073e-01 2.28881501e-02
4.16907251e-01 -2.91348815e-01 -9.79949117e-01 -2.02440880e-02
2.97784120e-01 -6.01714961e-02 7.67924368e-01 -5.65226316e-01
-2.31378034e-01 -4.07655120e-01 -9.34085965e-01 6.92758858e-01
6.65161014e-01 -9.52015698e-01 -7.31264651e-01 1.51502974e-02
1.10498285e+00 -3.68176848e-01 -7.81781495e-01 3.70645434e-01
4.85173047e-01 -1.02628541e+00 8.25534642e-01 -5.81428170e-01
4.11981910e-01 -5.58229506e-01 -1.32686585e-01 -1.80494809e+00
-1.41214609e-01 -1.23281038e+00 -3.09534613e-02 1.60539651e+00
3.17484766e-01 -2.41182789e-01 2.81954587e-01 -2.56936550e-01
-5.98850489e-01 -3.77841443e-01 -9.61689651e-01 -1.07669318e+00
5.20026207e-01 -5.24854481e-01 8.16739500e-01 7.85831690e-01
-1.33677468e-01 5.13999104e-01 -7.80620754e-01 2.07301676e-01
2.08162889e-01 -2.48849243e-02 7.31654882e-01 -6.37854338e-01
-5.03379643e-01 -3.87105048e-01 2.11213022e-01 -9.86847997e-01
1.92065716e-01 -8.36951256e-01 4.45787400e-01 -9.77788091e-01
-6.03136837e-01 -3.34999770e-01 -4.60055321e-01 8.94914716e-02
-3.80478501e-01 2.50369817e-01 4.99494225e-01 1.30223602e-01
1.87099308e-01 8.97538722e-01 1.71026981e+00 -2.70957202e-02
-5.00125825e-01 5.37206352e-01 -1.56396478e-01 7.98265398e-01
9.35082674e-01 -2.73430854e-01 -1.80034846e-01 1.26942601e-02
-5.47765493e-01 7.18471587e-01 3.15879017e-01 -1.26860988e+00
1.94992781e-01 1.34411141e-01 6.15059249e-02 -5.51821768e-01
7.43089259e-01 -4.59423572e-01 2.13857010e-01 5.75232863e-01
-3.96963179e-01 -2.89040327e-01 1.20820560e-01 1.96609259e-01
-7.04620957e-01 -1.99469909e-01 1.03113782e+00 -1.41327634e-01
-1.74714103e-02 2.40344375e-01 -3.80087137e-01 2.47581825e-02
5.06888568e-01 7.10557476e-02 1.11333154e-01 -5.93560278e-01
-6.42378449e-01 -2.49288797e-01 1.06221795e-01 4.95663494e-01
6.66626096e-01 -1.58630157e+00 -9.69829917e-01 5.05041659e-01
-3.26496661e-01 -9.56776217e-02 5.79910159e-01 3.29640210e-01
-4.13551539e-01 2.81268924e-01 -6.92460760e-02 -8.81020054e-02
-1.20510662e+00 5.96774638e-01 6.68074012e-01 -1.55015111e-01
-2.08497524e-01 7.01795638e-01 9.24100652e-02 -7.73815453e-01
2.82639772e-01 -3.14343721e-01 -2.63559759e-01 -6.02218807e-02
4.55680072e-01 5.08691907e-01 -1.21539362e-01 -7.71567583e-01
-2.77061313e-02 4.41983312e-01 4.97331828e-01 -3.18007827e-01
8.83076131e-01 2.08904728e-01 7.58604258e-02 5.30967355e-01
9.65336323e-01 8.22726786e-01 -1.09619236e+00 5.65996692e-02
-5.69067776e-01 -1.22847371e-01 6.56225234e-02 -8.10098052e-01
-1.06799829e+00 1.24949336e+00 3.20293665e-01 2.28659049e-01
1.38719249e+00 -5.05439341e-01 1.50783098e+00 -1.12989165e-01
5.72010539e-02 -8.96538734e-01 1.65496930e-01 5.90129197e-01
1.24588490e+00 -5.83094358e-01 -8.70720863e-01 -1.76234931e-01
-8.35178018e-01 1.20557344e+00 6.56397700e-01 -4.30372953e-01
5.92340946e-01 3.21424305e-01 4.97786045e-01 6.91460252e-01
-5.67494750e-01 -7.83487335e-02 4.14682180e-01 6.52992189e-01
3.93751621e-01 3.58454883e-01 -2.72611916e-01 1.34711242e+00
-8.16037476e-01 -1.41228691e-01 2.85893202e-01 -2.84536809e-01
-3.43813568e-01 -1.21725845e+00 -5.71149051e-01 -2.08499074e-01
-6.77844703e-01 -4.64570969e-01 -4.05122280e-01 2.94557601e-01
2.40243688e-01 1.14236856e+00 -4.79920618e-02 -1.01501799e+00
5.09435415e-01 2.46982202e-01 -2.15866300e-03 -5.47178328e-01
-1.20660889e+00 5.65348744e-01 5.79089411e-02 -2.06699148e-01
1.47723466e-01 -2.15023160e-01 -1.25803721e+00 -7.92883337e-02
-4.03853089e-01 4.01097924e-01 5.97705960e-01 4.07792419e-01
2.27592617e-01 8.62423122e-01 1.31199229e+00 -8.63975346e-01
-1.22036064e+00 -1.24959576e+00 -9.84570146e-01 2.04485998e-01
5.97708166e-01 1.14207275e-01 -5.72127342e-01 3.89301889e-02] | [15.506766319274902, 6.179930210113525] |
ff41ac60-7dad-438d-a8ba-d22aefda7d9b | analyzing-koopman-approaches-to-physics | 2010.00399 | null | https://arxiv.org/abs/2010.00399v2 | https://arxiv.org/pdf/2010.00399v2.pdf | Analyzing Koopman approaches to physics-informed machine learning for long-term sea-surface temperature forecasting | Accurately predicting sea-surface temperature weeks to months into the future is an important step toward long term weather forecasting. Standard atmosphere-ocean coupled numerical models provide accurate sea-surface forecasts on the scale of a few days to a few weeks, but many important weather systems require greater foresight. In this paper we propose machine-learning approaches sea-surface temperature forecasting that are accurate on the scale of dozens of weeks. Our approach is based in Koopman operator theory, a useful tool for dynamical systems modelling. With this approach, we predict sea surface temperature in the Gulf of Mexico up to 180 days into the future based on a present image of thermal conditions and three years of historical training data. We evaluate the combination of a basic Koopman method with a convolutional autoencoder, and a newly proposed "consistent Koopman" method, in various permutations. We show that the Koopman approach consistently outperforms baselines, and we discuss the utility of our additional assumptions and methods in this sea-surface temperature domain. | ['Andrew August', 'Wenwei Xu', 'Julian Rice'] | 2020-09-15 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [-4.48523045e-01 -4.27890062e-01 4.07944292e-01 -4.35815632e-01
-4.47656900e-01 -6.34142160e-01 8.90329838e-01 -1.84574515e-01
-4.03085858e-01 9.47968841e-01 8.39558691e-02 -6.75851941e-01
6.31318539e-02 -9.23211813e-01 -3.34694177e-01 -9.70995605e-01
-5.85682809e-01 1.85147211e-01 1.62688084e-02 -7.98433006e-01
8.60501155e-02 5.81124187e-01 -1.47825420e+00 -2.12173596e-01
7.64791489e-01 1.01753414e+00 -1.36071950e-01 1.09478080e+00
1.15867890e-01 4.19884145e-01 -2.25439280e-01 1.55598596e-01
4.51143205e-01 -3.08236063e-01 -6.12832248e-01 -2.39304930e-01
4.04598922e-01 -3.95059854e-01 1.23315878e-01 7.36618280e-01
2.28127524e-01 5.26009858e-01 5.47076941e-01 -7.99714625e-01
4.29526120e-02 1.08589098e-01 -1.09528482e-01 2.90245175e-01
-2.98398882e-01 1.38377666e-01 9.03065920e-01 -6.61477923e-01
1.60774648e-01 9.48188305e-01 1.23332381e+00 3.70705128e-01
-1.16576874e+00 -3.76079947e-01 1.05292737e-01 -3.22178930e-01
-1.20390213e+00 -4.49828804e-01 5.19464254e-01 -7.71257997e-01
1.40304399e+00 4.25309956e-01 9.03736055e-01 3.82761180e-01
9.01138842e-01 1.67560559e-02 1.25199628e+00 -5.00768125e-01
4.57495570e-01 1.51351944e-01 -2.84854136e-02 5.21623790e-01
8.97976160e-02 7.74191618e-01 -3.56107712e-01 -4.80292588e-01
4.75871146e-01 -2.00615734e-01 -3.22184771e-01 2.64543712e-01
-9.34192479e-01 9.61470902e-01 3.51295799e-01 3.34552348e-01
-3.94180864e-01 3.07356209e-01 9.63893682e-02 4.13451076e-01
1.23301280e+00 8.43803763e-01 -1.19639170e+00 -8.98345709e-02
-1.49776745e+00 6.62669480e-01 1.01919353e+00 1.51656121e-01
9.81317222e-01 5.34555495e-01 8.19167793e-01 2.69726157e-01
5.29142976e-01 9.75158751e-01 4.46548253e-01 -9.40278709e-01
-1.15498640e-01 -5.85129410e-02 6.51222289e-01 -6.55999005e-01
-5.68737924e-01 -3.18699270e-01 -9.91476059e-01 6.55934870e-01
1.31571263e-01 -1.02280152e+00 -7.99755216e-01 1.36027288e+00
2.94759572e-01 4.63031054e-01 6.09251380e-01 8.55754972e-01
1.23711247e-02 1.53006029e+00 -2.42130890e-01 -3.62647206e-01
1.02782834e+00 -6.58795178e-01 -8.97650659e-01 -5.08377671e-01
8.46054375e-01 -5.38688540e-01 3.01723421e-01 2.35911012e-01
-6.19357944e-01 -3.26094180e-01 -9.54314888e-01 2.94862866e-01
-8.68438005e-01 -4.30928245e-02 7.93068588e-01 3.23027223e-01
-1.63850498e+00 1.26691997e+00 -1.38041055e+00 -2.71020025e-01
-6.98522508e-01 5.97487651e-02 -2.34270364e-01 9.58836675e-01
-1.58090198e+00 1.23024070e+00 2.15915471e-01 6.98352337e-01
-6.31086409e-01 -8.50169599e-01 -1.07667148e+00 1.13897637e-01
-5.10269463e-01 -5.59248984e-01 1.53161800e+00 -1.10580528e+00
-1.76452613e+00 2.58265227e-01 -4.29209977e-01 -5.05965412e-01
2.29326040e-01 -3.94866109e-01 -6.09886408e-01 -2.95733273e-01
-2.92334825e-01 4.18532580e-01 5.42786181e-01 -9.96643186e-01
-6.70531392e-01 -1.30308449e-01 -4.61874664e-01 2.46930748e-01
4.66336049e-02 -4.68808897e-02 2.53436178e-01 -5.05214572e-01
1.56387940e-01 -1.24890435e+00 -7.93146551e-01 1.34211881e-02
2.81482816e-01 -4.15886790e-02 6.86491251e-01 -9.90579963e-01
1.02628863e+00 -1.77481067e+00 -6.52874708e-02 2.23672837e-01
-3.12543124e-01 -6.19073212e-03 1.61868736e-01 5.83501697e-01
-2.43087754e-01 1.99621767e-01 -7.82230377e-01 -4.89129186e-01
-1.60692394e-01 4.52074945e-01 -7.34398723e-01 5.48160255e-01
3.02430779e-01 4.60703462e-01 -6.73720002e-01 1.70680016e-01
3.55816096e-01 5.07539749e-01 -2.41769388e-01 3.70642096e-01
-2.30239168e-01 4.11693484e-01 -5.45397177e-02 4.25204858e-02
7.91046858e-01 1.11516919e-02 1.94406118e-02 3.22025120e-01
-1.00381267e+00 4.66719478e-01 -1.04175627e+00 1.28986943e+00
-5.91636598e-01 1.15758491e+00 1.79972112e-01 -6.59925699e-01
9.98315036e-01 5.85600555e-01 2.44144425e-01 -4.02572185e-01
-1.99079767e-01 4.51737791e-01 -1.43277526e-01 -3.01306754e-01
6.72752321e-01 -8.85304749e-01 9.52622518e-02 4.28515345e-01
-1.15729123e-01 -7.04627573e-01 -1.85868427e-01 -1.72256544e-01
1.86431885e-01 2.52945751e-01 2.40611419e-01 -9.56265926e-01
3.41912150e-01 1.80869028e-01 6.33948207e-01 4.99168068e-01
-1.91387638e-01 3.84191066e-01 3.90792012e-01 -1.31804383e+00
-1.15632915e+00 -3.90733391e-01 -5.07245243e-01 9.41399336e-01
-3.37720275e-01 -6.68388754e-02 -5.66463232e-01 3.49254608e-02
-4.15269360e-02 7.26861537e-01 -9.84113455e-01 1.29606307e-01
-4.16086167e-01 -1.26270854e+00 6.07950270e-01 4.65071678e-01
4.92883176e-01 -8.16539943e-01 -5.76612711e-01 3.48707259e-01
6.12867437e-02 -6.77533090e-01 1.54210269e-01 3.84845346e-01
-1.30201685e+00 -3.88234764e-01 -8.08832288e-01 -4.98660207e-01
1.78799286e-01 -2.84952164e-01 1.17086411e+00 -1.82112306e-01
2.28799492e-01 -1.35560259e-01 -9.86570865e-02 -5.55250227e-01
-4.08647388e-01 -1.98887631e-01 5.70665777e-01 -8.61615613e-02
1.24610037e-01 -6.82021558e-01 -4.02397573e-01 -1.09785624e-01
-5.91250241e-01 -5.71176820e-02 8.34773779e-02 6.97755098e-01
1.18826188e-01 -3.41526158e-02 1.89384833e-01 -4.29684192e-01
2.76299328e-01 -6.24052167e-01 -1.44944918e+00 7.78089538e-02
-8.54848027e-01 4.63123053e-01 6.06515169e-01 4.45149653e-02
-1.24985909e+00 2.29448959e-01 -4.29182917e-01 -7.80364424e-02
-2.09293857e-01 1.00043046e+00 5.63255370e-01 -1.91092595e-01
7.03149498e-01 5.83822131e-01 4.90436405e-02 -7.00195193e-01
2.13145167e-01 5.38521945e-01 4.32484269e-01 -3.96553457e-01
7.26770878e-01 5.70414424e-01 6.00170195e-02 -1.11732399e+00
-4.98229980e-01 -3.14504147e-01 -8.37418258e-01 -2.96471477e-01
8.18685353e-01 -1.17685151e+00 -2.95480669e-01 9.38147187e-01
-1.10955906e+00 -7.52172351e-01 4.75514457e-02 7.43504345e-01
-7.35298544e-02 1.10111728e-01 -6.00408673e-01 -1.04816306e+00
-5.20659208e-01 -6.74115598e-01 9.95612264e-01 2.46318579e-01
9.48581174e-02 -1.79939044e+00 1.08817661e+00 -5.41858673e-01
7.91005731e-01 3.62433642e-01 3.50008905e-01 -9.16886702e-02
3.17923236e-03 -2.62085021e-01 2.26452217e-01 5.10158539e-01
6.83445483e-02 4.26791966e-01 -1.23478305e+00 -3.59433353e-01
2.18798861e-01 1.00097794e-03 1.15246844e+00 7.00516582e-01
4.97806042e-01 -5.49586117e-01 -2.89793219e-02 9.54232752e-01
1.64460027e+00 6.56332076e-02 4.26392287e-01 3.46112937e-01
2.28281587e-01 6.41927838e-01 7.85675943e-02 6.27077401e-01
3.24281514e-01 -8.90413765e-03 2.34624520e-01 -2.11972281e-01
7.77564526e-01 3.12305480e-01 4.32251900e-01 9.48171437e-01
-4.72680956e-01 -4.14363258e-02 -1.35512471e+00 8.36554110e-01
-1.47862887e+00 -8.82982314e-01 -1.78681031e-01 1.98324108e+00
6.35426939e-01 -1.57139316e-01 -5.33700824e-01 -9.40472856e-02
2.26967245e-01 6.02952659e-01 -2.44522318e-01 -9.79856133e-01
7.32239187e-02 3.01850021e-01 8.91329288e-01 1.17600179e+00
-1.22513258e+00 9.29508448e-01 7.37107372e+00 -1.92099899e-01
-1.80207980e+00 -1.36230484e-01 4.39787865e-01 4.56344604e-01
-2.08024070e-01 2.73936987e-01 -8.40311706e-01 7.79808685e-02
1.74725115e+00 -8.31899047e-02 3.40570778e-01 8.30276847e-01
7.74586380e-01 -3.13095063e-01 -6.37096584e-01 3.25202614e-01
-2.95797884e-01 -1.52820706e+00 -3.11393321e-01 -6.89469874e-02
1.06539118e+00 4.90677834e-01 -1.38095528e-01 1.27913624e-01
4.26225811e-01 -8.44867945e-01 5.49732208e-01 9.05785143e-01
6.43535078e-01 -6.25001609e-01 9.62374985e-01 5.99850297e-01
-1.37961042e+00 2.40908384e-01 -3.88603956e-01 -8.83426249e-01
7.82169923e-02 7.13785946e-01 -3.95759434e-01 2.23621055e-01
9.74353135e-01 9.37335491e-01 -1.60524637e-01 6.95537746e-01
-3.28947082e-02 9.82336640e-01 -9.03511107e-01 -6.50350302e-02
6.83992326e-01 -3.12124968e-01 2.78995603e-01 1.18968451e+00
6.13774300e-01 6.92246020e-01 -1.84612095e-01 6.07542634e-01
4.02599931e-01 -1.85366094e-01 -7.38460064e-01 2.41708189e-01
1.71881363e-01 1.01042259e+00 -9.35016945e-02 -6.52176261e-01
-3.78969729e-01 7.41062284e-01 -6.35805577e-02 5.52919149e-01
-2.95428574e-01 -4.94773984e-01 1.26103616e+00 -2.55193323e-01
2.37365082e-01 -6.54989719e-01 -1.98790178e-01 -1.39319468e+00
-2.37734437e-01 -3.32434833e-01 9.95297059e-02 -9.19905007e-01
-9.95361984e-01 6.00300014e-01 -1.06403865e-01 -1.27228081e+00
-5.34765005e-01 -8.11990559e-01 -1.10034037e+00 1.38663411e+00
-1.96960652e+00 -7.28300214e-01 -3.85916233e-02 7.36356825e-02
2.27485254e-01 2.52019465e-01 1.32060969e+00 -2.90277213e-01
-5.10771573e-01 -1.64729938e-01 1.00505972e+00 1.12050086e-01
3.95381957e-01 -1.48124278e+00 1.03809810e+00 9.94152427e-01
-3.62552881e-01 4.93727088e-01 1.04527867e+00 -5.72404444e-01
-1.20832098e+00 -1.21163952e+00 1.36291981e+00 -1.15571640e-01
9.29470420e-01 -1.18185967e-01 -1.37201011e+00 1.01439440e+00
5.98382056e-01 4.38531309e-01 4.57419693e-01 2.14433834e-01
-3.07373907e-02 -2.89095253e-01 -6.66259170e-01 1.95240200e-01
-2.66051412e-01 -5.41298032e-01 -7.74323225e-01 3.32345396e-01
4.40933049e-01 -4.14431006e-01 -1.04023659e+00 4.70376581e-01
6.52699888e-01 -9.06819522e-01 3.21314871e-01 -6.20598853e-01
2.34611616e-01 -3.87137771e-01 -3.05886157e-02 -1.90263915e+00
-3.88815552e-01 -9.00088906e-01 2.60329217e-01 3.96677256e-01
6.93459690e-01 -9.71776426e-01 4.90076691e-01 7.62332141e-01
-2.66075015e-01 -3.07305485e-01 -1.09642386e+00 -6.41858399e-01
9.14977312e-01 -6.01575255e-01 4.39636022e-01 1.19979250e+00
1.14561550e-01 -2.17047736e-01 -7.14098155e-01 1.00911868e+00
4.73495185e-01 2.81555444e-01 3.51250291e-01 -1.34624970e+00
-3.32888067e-02 -2.77534842e-01 -6.87274784e-02 -7.54494071e-01
1.38017550e-01 -1.72437236e-01 6.14095986e-01 -1.17495930e+00
-4.37322706e-01 -8.87565613e-02 -2.98895210e-01 4.48606700e-01
-1.00551501e-01 5.32423705e-02 -2.75271386e-01 -6.79131448e-02
4.04711545e-01 6.37622476e-01 7.09094942e-01 2.79255569e-01
-3.99729639e-01 -1.68247446e-01 1.74261853e-01 7.79398084e-01
8.93003285e-01 -3.19119662e-01 2.51278102e-01 -6.76685691e-01
5.59266508e-01 3.34801227e-01 1.80184200e-01 -1.25985920e+00
2.13830531e-01 -5.38196206e-01 3.88231874e-01 -3.83128583e-01
1.98044777e-01 -5.44948518e-01 1.59343779e-01 6.21094346e-01
-2.04974785e-01 3.08261335e-01 7.99894452e-01 3.54173213e-01
-4.59118605e-01 9.46455300e-02 9.77518737e-01 -2.10903004e-01
-9.69775677e-01 1.66381270e-01 -1.05145502e+00 -3.92462164e-01
6.57499015e-01 3.48339349e-01 1.59376580e-02 -4.94820118e-01
-7.30189681e-01 4.31300730e-01 3.94940794e-01 2.19851255e-01
2.01813281e-01 -7.81931102e-01 -1.01549983e+00 5.32182693e-01
-1.87743738e-01 -1.52640119e-01 1.54151633e-01 6.55767202e-01
-1.20698094e+00 5.27782142e-01 4.95951585e-02 -4.81487393e-01
-7.82319725e-01 1.87882390e-02 1.27418101e+00 -1.07876725e-01
-5.47380984e-01 1.01118696e+00 -6.46420941e-02 -7.40997434e-01
-3.33813608e-01 -8.49258006e-01 -1.37693107e-01 1.08126655e-01
7.37861514e-01 5.54285571e-02 1.06641062e-01 -6.73392236e-01
-2.97674984e-01 8.71562183e-01 5.43997347e-01 -7.11385846e-01
1.38599551e+00 -1.72453150e-01 -2.21351147e-01 9.25249636e-01
1.19104695e+00 -3.02972168e-01 -1.57487750e+00 2.88394336e-02
-1.64699644e-01 6.28981367e-02 7.59977639e-01 -7.90601194e-01
-9.66515005e-01 1.39069307e+00 7.05769420e-01 1.53014392e-01
1.01996613e+00 -5.20037293e-01 7.74144590e-01 8.41226399e-01
-1.55229360e-01 -8.64009321e-01 -8.94874334e-01 1.04720986e+00
7.30117500e-01 -1.35986304e+00 1.40896559e-01 4.71904308e-01
-4.52822417e-01 1.55212522e+00 8.95430148e-02 -1.74671516e-01
1.28813946e+00 4.55803841e-01 7.20710695e-01 1.05989039e-01
-1.25477564e+00 -3.45416702e-02 2.62463123e-01 -7.67554343e-02
5.88695943e-01 3.40770185e-01 7.46108815e-02 8.54395330e-02
-3.99880052e-01 -2.03943744e-01 7.54360557e-01 9.53607619e-01
-7.97268212e-01 -6.52638733e-01 -7.22194493e-01 1.03624813e-01
-3.62500280e-01 -3.87030929e-01 3.30070183e-02 5.67654014e-01
-1.71909168e-01 7.66027212e-01 5.84877431e-01 -1.65907934e-01
-3.62935483e-01 5.54225028e-01 -2.95908362e-01 -1.87323838e-01
-4.78662163e-01 1.80700719e-01 9.15522426e-02 -3.52671355e-01
-5.24247587e-01 -9.50504005e-01 -1.08441210e+00 -6.08462036e-01
-4.93656844e-01 7.12430596e-01 1.00679290e+00 1.24881208e+00
5.50695322e-02 6.83176965e-02 9.73556638e-01 -1.44517970e+00
-2.94907808e-01 -1.26142406e+00 -9.17350531e-01 -2.80444026e-01
1.15115643e+00 -8.70405212e-02 -9.45101261e-01 3.23742419e-01] | [6.47027063369751, 3.0392255783081055] |
9a91b01b-e31d-48b5-8a79-7778cea97149 | are-we-making-real-progress-in-simulated | 1912.06321 | null | https://arxiv.org/abs/1912.06321v2 | https://arxiv.org/pdf/1912.06321v2.pdf | Sim2Real Predictivity: Does Evaluation in Simulation Predict Real-World Performance? | Does progress in simulation translate to progress on robots? If one method outperforms another in simulation, how likely is that trend to hold in reality on a robot? We examine this question for embodied PointGoal navigation, developing engineering tools and a research paradigm for evaluating a simulator by its sim2real predictivity. First, we develop Habitat-PyRobot Bridge (HaPy), a library for seamless execution of identical code on simulated agents and robots, transferring simulation-trained agents to a LoCoBot platform with a one-line code change. Second, we investigate the sim2real predictivity of Habitat-Sim for PointGoal navigation. We 3D-scan a physical lab space to create a virtualized replica, and run parallel tests of 9 different models in reality and simulation. We present a new metric called Sim-vs-Real Correlation Coefficient (SRCC) to quantify predictivity. We find that SRCC for Habitat as used for the CVPR19 challenge is low (0.18 for the success metric), suggesting that performance differences in this simulator-based challenge do not persist after physical deployment. This gap is largely due to AI agents learning to exploit simulator imperfections, abusing collision dynamics to 'slide' along walls, leading to shortcuts through otherwise non-navigable space. Naturally, such exploits do not work in the real world. Our experiments show that it is possible to tune simulation parameters to improve sim2real predictivity (e.g. improving $SRCC_{Succ}$ from 0.18 to 0.844), increasing confidence that in-simulation comparisons will translate to deployed systems in reality. | ['Abhishek Kadian', 'Sonia Chernova', 'Erik Wijmans', 'Stefan Lee', 'Manolis Savva', 'Dhruv Batra', 'Alexander Clegg', 'Joanne Truong', 'Aaron Gokaslan'] | 2019-12-13 | null | null | null | null | ['pointgoal-navigation'] | ['robots'] | [-3.23500425e-01 6.32249191e-02 2.81561464e-01 -6.02391772e-02
-4.60522115e-01 -8.18855584e-01 6.47471666e-01 -1.01168185e-01
-7.15562940e-01 8.67278516e-01 -1.48085818e-01 -8.36090267e-01
-1.80172876e-01 -7.64632404e-01 -9.33532178e-01 -3.04182708e-01
-8.79722297e-01 6.29800618e-01 5.99642396e-01 -8.00394356e-01
1.38888761e-01 5.69089472e-01 -1.69086707e+00 -3.78868759e-01
7.97106981e-01 1.85054675e-01 3.90404671e-01 9.89543319e-01
4.16454911e-01 7.30831742e-01 -7.58242607e-01 1.53321549e-01
4.56648111e-01 -4.33416337e-01 -6.80271745e-01 -6.30292892e-01
-1.87103018e-01 -2.20564216e-01 -2.41833627e-01 6.51448548e-01
5.02087951e-01 1.89724024e-02 4.41059619e-01 -1.74763870e+00
3.45325857e-01 4.73329395e-01 -2.37228766e-01 -7.80320587e-03
7.43658066e-01 7.53861129e-01 3.92644137e-01 -3.51109117e-01
7.91052222e-01 1.24356174e+00 1.14050591e+00 1.26489058e-01
-1.10011971e+00 -6.61234081e-01 -1.46414712e-01 -2.87171513e-01
-1.60211885e+00 -1.85666516e-01 4.49211411e-02 -5.79336107e-01
1.36862612e+00 2.33525291e-01 1.01824176e+00 1.13645196e+00
9.12975192e-01 -2.76185665e-02 1.02965498e+00 -1.29861698e-01
6.60410345e-01 -5.81284948e-02 -4.26600456e-01 7.15375483e-01
5.87883890e-01 7.13054359e-01 -1.52273685e-01 -2.47205988e-01
9.48784709e-01 -6.08204782e-01 -1.90356225e-01 -4.76444930e-01
-1.35087776e+00 4.55531240e-01 5.11791170e-01 6.23429418e-02
-4.00686413e-01 6.56999886e-01 2.11494625e-01 5.69409370e-01
-4.75636214e-01 1.18730509e+00 -4.47665989e-01 -7.02498078e-01
-2.90973902e-01 8.85672808e-01 1.18614912e+00 1.06862664e+00
5.58916628e-01 2.12874040e-01 5.99068165e-01 2.53439844e-01
2.05452889e-01 7.21603274e-01 3.07036310e-01 -1.44272268e+00
6.64165244e-02 2.11313218e-01 5.20819485e-01 -9.41050172e-01
-9.74738419e-01 -2.72375137e-01 1.46713285e-02 1.11819148e+00
4.61553097e-01 -5.18253088e-01 -6.15766406e-01 1.66283226e+00
1.67102247e-01 -3.18886042e-02 2.90502012e-01 1.01233459e+00
2.49746755e-01 5.70392728e-01 7.67634064e-02 2.86664009e-01
9.08558309e-01 -7.44348943e-01 5.57851680e-02 -5.36785185e-01
1.17150700e+00 -5.60671628e-01 1.25468135e+00 3.56535226e-01
-7.27833569e-01 -3.12087953e-01 -1.51050699e+00 7.47711301e-01
-2.54740506e-01 -6.09948516e-01 7.42432475e-01 6.67531908e-01
-1.23009646e+00 8.28915536e-01 -1.27067423e+00 -7.84782231e-01
-2.17579484e-01 5.82475126e-01 -4.24406797e-01 1.23068072e-01
-7.87372172e-01 1.43891203e+00 1.87275186e-01 -3.53998214e-01
-1.25768495e+00 -4.78651285e-01 -7.05297828e-01 -4.42404240e-01
7.02590495e-02 -7.20936537e-01 1.41253793e+00 -3.78005713e-01
-1.74931467e+00 4.16199744e-01 4.84792262e-01 -4.40498620e-01
6.26040220e-01 -4.63905409e-02 -3.73501837e-01 -3.00177068e-01
3.31946164e-01 6.63743258e-01 -1.38747804e-02 -1.63112378e+00
-6.59378946e-01 -4.65255172e-04 4.45227534e-01 4.49804664e-01
4.88219649e-01 -2.14965597e-01 -1.25387311e-01 8.81331563e-02
1.48245201e-01 -1.35552907e+00 -7.05912232e-01 -1.83466196e-01
1.82753593e-01 3.27185333e-01 3.28741938e-01 -3.08144361e-01
3.27101827e-01 -1.84865212e+00 -2.29684450e-02 2.95764446e-01
-3.84018980e-02 -3.22024196e-01 -3.28334600e-01 8.76406014e-01
2.13598549e-01 1.43445775e-01 3.37417200e-02 1.64029628e-01
8.10874104e-02 2.47341231e-01 -5.72939357e-03 6.78495109e-01
-1.41282573e-01 7.06411958e-01 -1.20581579e+00 -1.67675108e-01
1.28788576e-01 2.82311410e-01 -7.03766108e-01 -8.97734910e-02
-5.76430336e-02 4.57731545e-01 -3.34421307e-01 4.60954607e-01
4.68991995e-01 1.58814058e-01 2.60297149e-01 5.33959031e-01
-7.17144728e-01 3.74746770e-01 -1.06655979e+00 1.68107665e+00
-6.62969768e-01 5.27124345e-01 2.84374654e-01 -3.33823532e-01
8.58209610e-01 -1.92953780e-01 3.64673764e-01 -1.02601624e+00
2.23942265e-01 5.34890771e-01 3.93779337e-01 -4.76338208e-01
8.44019890e-01 -1.24707691e-01 -4.80315328e-01 2.29763508e-01
-2.31918380e-01 -9.02244925e-01 -2.03897133e-02 -6.76206425e-02
1.76099169e+00 4.31105733e-01 -9.28646252e-02 -6.69928014e-01
-2.14609399e-01 9.33199584e-01 2.68047810e-01 1.16414857e+00
-1.96103841e-01 4.04814243e-01 3.91981661e-01 -6.90761656e-02
-1.40259814e+00 -1.24721467e+00 1.44615546e-01 8.62924874e-01
9.79887068e-01 -4.53064620e-01 -6.58845186e-01 -2.37740889e-01
1.56236365e-01 1.07629490e+00 -3.30557317e-01 -3.49650621e-01
-7.05107749e-01 -4.61110860e-01 1.03817511e+00 3.67167681e-01
1.79682761e-01 -7.32815862e-01 -1.43489647e+00 3.75266671e-01
3.45915973e-01 -6.89663708e-01 2.48853847e-01 4.06642139e-01
-5.79717398e-01 -8.56206954e-01 -2.45738775e-01 -5.55474639e-01
3.24576378e-01 5.32463729e-01 1.11687708e+00 4.19631034e-01
-8.27239174e-03 4.66654569e-01 -4.67881531e-01 -2.28654519e-01
-6.62111700e-01 6.66016713e-02 4.78999317e-01 -1.40511239e+00
-2.47649759e-01 -7.10144758e-01 -6.22214377e-01 8.06752682e-01
-4.31138635e-01 -2.47852970e-02 6.30102932e-01 6.59624100e-01
-1.53206009e-02 1.44883558e-01 2.66591311e-01 -1.63196787e-01
7.96511889e-01 -5.93730032e-01 -1.00033712e+00 -1.93612069e-01
-5.61178505e-01 -2.17373855e-02 4.94070202e-01 -3.59676540e-01
-4.23092842e-01 -1.15079775e-01 -2.11734638e-01 8.07923600e-02
-5.40560745e-02 7.59249032e-01 2.86090016e-01 -3.69836420e-01
1.12930357e+00 -2.14392304e-01 2.84539223e-01 9.78740454e-02
1.57097116e-01 2.65028745e-01 5.22312403e-01 -7.78480053e-01
8.50154102e-01 3.68039757e-01 -1.20610110e-01 -6.06946766e-01
3.38413656e-01 7.55607933e-02 1.24333762e-01 -2.56124139e-01
5.07470548e-01 -8.95306349e-01 -1.27088428e+00 2.71004170e-01
-7.41627872e-01 -1.31577182e+00 -1.87562719e-01 7.53724039e-01
-7.99753785e-01 -1.70845166e-01 -4.09192681e-01 -8.33385944e-01
3.56718659e-01 -1.45514452e+00 8.01659942e-01 4.56431746e-01
-6.23816609e-01 -6.31771505e-01 5.44269621e-01 -2.52712071e-01
5.86179435e-01 2.80177414e-01 3.54258478e-01 -4.10808414e-01
-5.09174049e-01 -1.53915837e-01 1.19910598e-01 -6.10225976e-01
-3.69876355e-01 -1.38029426e-01 -4.66048539e-01 -6.33510113e-01
-1.23470306e-01 -2.53601849e-01 1.00259922e-01 -1.69504639e-02
4.71600182e-02 8.82441103e-02 -7.19793737e-01 3.54160905e-01
1.38160944e+00 5.71337581e-01 8.04426074e-01 1.13223040e+00
3.52026016e-01 3.53729755e-01 9.68282759e-01 2.78561503e-01
5.58926225e-01 8.19065869e-01 5.82666397e-01 1.38450533e-01
1.86908901e-01 -4.07519609e-01 8.15213203e-01 5.08446276e-01
-9.02386159e-02 -2.03041300e-01 -1.42324519e+00 4.31292653e-01
-1.66108739e+00 -6.55581117e-01 -2.59714693e-01 2.29555941e+00
2.22718522e-01 6.45406246e-01 1.71087071e-01 -3.26752394e-01
3.69556993e-01 -3.68136078e-01 -4.11797255e-01 -3.09519053e-01
1.15657747e-01 4.81255352e-02 1.04334879e+00 7.73692131e-01
-4.66061592e-01 1.00844920e+00 6.81075573e+00 2.82448918e-01
-1.27149487e+00 -3.31471227e-02 -1.57307852e-02 1.61210671e-01
-1.47795618e-01 4.94131982e-01 -4.45748478e-01 2.11081788e-01
1.28327036e+00 -3.43870819e-01 6.23441756e-01 1.05372596e+00
3.34272861e-01 -6.37940168e-01 -9.81743991e-01 4.83227283e-01
-2.79666841e-01 -9.99687314e-01 -5.67081749e-01 3.08512688e-01
5.19601882e-01 5.52457273e-01 -3.61141600e-02 7.76300609e-01
1.03881466e+00 -1.33304822e+00 1.12034774e+00 3.04880768e-01
5.09050548e-01 -7.18099475e-01 8.51424038e-01 4.70272392e-01
-9.59332407e-01 9.07142386e-02 -1.72050983e-01 -6.10476673e-01
4.22385722e-01 -1.62599891e-01 -1.45961857e+00 4.49054807e-01
7.30283439e-01 -6.22278005e-02 -4.20621932e-01 1.27141631e+00
5.54675013e-02 2.79018015e-01 -7.97937989e-01 -4.53645140e-01
3.80603373e-01 -1.24932803e-01 6.25538528e-01 8.78210068e-01
5.57370603e-01 -8.20212737e-02 1.99814126e-01 7.82721817e-01
5.88185370e-01 -3.44486803e-01 -9.19386804e-01 2.02766493e-01
6.72811151e-01 9.39015329e-01 -9.22264934e-01 5.02482392e-02
3.12419664e-02 7.37264872e-01 -3.92838158e-02 1.88439637e-01
-1.25875485e+00 -5.00302672e-01 8.43637407e-01 1.59045547e-01
-8.61131400e-02 -8.35528374e-01 -4.29857463e-01 -6.20034516e-01
-2.54272968e-01 -1.00222933e+00 -3.43146116e-01 -1.00772512e+00
-6.90005958e-01 7.07713068e-01 4.19263281e-02 -1.37274504e+00
-4.93184298e-01 -4.73255306e-01 -3.42112005e-01 7.18206763e-01
-9.18290675e-01 -6.95411563e-01 -6.21112049e-01 -1.18615171e-02
1.04321040e-01 -8.68554339e-02 7.59407282e-01 -2.44318955e-02
-3.36142927e-02 5.22614062e-01 3.76230255e-02 -5.25507808e-01
5.25506735e-01 -8.99714708e-01 9.36213851e-01 5.50132453e-01
-4.19565737e-01 7.66150713e-01 1.51521516e+00 -1.04812920e+00
-1.76589930e+00 -5.99928081e-01 -6.81096315e-02 -8.80545378e-01
8.86750102e-01 -3.04573447e-01 -5.85514784e-01 6.93594992e-01
-2.25695875e-02 -4.74517316e-01 1.64101385e-02 5.82785085e-02
-5.23378551e-02 3.21834803e-01 -1.15336382e+00 1.28282225e+00
1.38951457e+00 -5.12594655e-02 -3.32407415e-01 -4.86038113e-03
8.99945199e-01 -7.13603377e-01 -8.75946403e-01 5.39055407e-01
8.99155498e-01 -1.05670881e+00 8.06548893e-01 -1.63774833e-01
1.52637780e-01 -6.35456383e-01 -2.94665664e-01 -1.63188195e+00
-1.15854800e-01 -7.13035405e-01 6.53908432e-01 5.74511588e-01
6.52840793e-01 -8.47058058e-01 6.89656019e-01 4.25522804e-01
-4.57024783e-01 -4.13766474e-01 -8.90895247e-01 -1.23201847e+00
3.34942371e-01 -5.68567514e-01 6.18931532e-01 8.19898665e-01
4.26318973e-01 1.80503756e-01 1.22633740e-01 4.80855197e-01
3.42716128e-01 -5.92279911e-01 1.35745692e+00 -8.87354851e-01
-5.45611322e-01 -5.83330750e-01 -6.56892121e-01 -8.82369161e-01
-8.87191370e-02 -6.49329245e-01 5.47649860e-01 -1.42085004e+00
-2.03322485e-01 -1.07241476e+00 3.22460204e-01 2.74949580e-01
2.65548646e-01 -6.75603598e-02 2.64970541e-01 3.02398413e-01
-5.82103431e-01 4.76382643e-01 9.44607854e-01 2.42061704e-01
-6.13806784e-01 -3.78261507e-01 -3.69333327e-01 7.78934896e-01
8.33331645e-01 -5.03881633e-01 -4.30218309e-01 -4.50475037e-01
3.92567366e-01 3.25155497e-01 6.44459963e-01 -1.86405337e+00
2.98647732e-01 -3.84834439e-01 4.71067615e-02 -1.29417822e-01
4.72662419e-01 -7.36953020e-01 7.48490512e-01 8.77583623e-01
9.74443704e-02 5.28117359e-01 6.53863549e-01 3.38846236e-01
3.85412067e-01 -2.29933798e-01 3.70828152e-01 -9.77728516e-02
-7.76298046e-01 -4.90376383e-01 -1.00487745e+00 -8.10494646e-02
1.13871181e+00 -4.23436254e-01 -6.16641402e-01 -4.68615890e-01
-3.15958232e-01 4.59313899e-01 1.37632763e+00 3.84406954e-01
3.18163306e-01 -8.86853874e-01 -4.09699589e-01 -4.74546961e-02
1.65236369e-01 -3.17400545e-01 -1.09577984e-01 7.75915027e-01
-1.41790307e+00 8.85990728e-03 -5.17643273e-01 -6.82173491e-01
-7.79074252e-01 1.63269326e-01 6.10571802e-01 6.73268512e-02
-6.76075459e-01 6.74076200e-01 -9.91061330e-02 -9.64837432e-01
-1.49048135e-01 -2.77891278e-01 3.36108536e-01 -7.63393641e-01
1.37959421e-01 1.73960164e-01 -1.05010048e-01 -3.90218854e-01
-6.28008962e-01 3.28022838e-01 2.65264243e-01 -9.50481296e-01
1.25798452e+00 -1.87364724e-02 2.50320494e-01 4.60068494e-01
6.46662652e-01 1.02126658e-01 -1.46139562e+00 8.81766260e-01
2.77242251e-02 -2.42518231e-01 -2.66238987e-01 -7.38417864e-01
-2.70368934e-01 9.31575745e-02 8.01007569e-01 1.81273267e-01
2.71809399e-01 -1.25492737e-01 4.64664310e-01 7.76185095e-01
1.32604277e+00 -9.51176465e-01 4.07892950e-02 8.81056607e-01
8.78412426e-01 -8.10066879e-01 3.17567885e-02 -2.55572617e-01
-6.60875916e-01 6.63677335e-01 9.39675093e-01 -4.70148832e-01
1.61723733e-01 8.00936580e-01 1.17407717e-01 -2.01769814e-01
-6.62801266e-01 -4.60555367e-02 -9.06859338e-01 9.52661693e-01
2.34751962e-02 2.10621655e-01 -1.27649099e-01 3.59379917e-01
-9.67321754e-01 -7.26333335e-02 1.02615464e+00 1.56748414e+00
-6.82263434e-01 -8.53954613e-01 -6.44792557e-01 -4.72004190e-02
3.06427419e-01 3.40224028e-01 -2.08366171e-01 1.36003685e+00
-7.55281150e-02 9.46946859e-01 1.74110964e-01 -9.30589318e-01
6.40738130e-01 -4.05193388e-01 6.24955714e-01 -3.71468246e-01
-5.72425306e-01 -3.68556619e-01 5.91961622e-01 -7.09055245e-01
1.92292169e-01 -6.51700377e-01 -1.76079369e+00 -6.62304223e-01
-2.30801895e-01 3.41347098e-01 9.92049634e-01 6.14973068e-01
4.94575173e-01 6.95047915e-01 2.72918761e-01 -1.30329716e+00
-3.70318323e-01 -6.07240617e-01 -2.44180977e-01 -4.29692976e-02
-1.94304232e-02 -1.02844954e+00 -3.82743686e-01 -4.97555643e-01] | [4.639543056488037, 0.9841850399971008] |
3f61135c-b9e1-49e7-a49c-c196d8962439 | risk-aware-reward-shaping-of-reinforcement | 2306.03220 | null | https://arxiv.org/abs/2306.03220v1 | https://arxiv.org/pdf/2306.03220v1.pdf | Risk-Aware Reward Shaping of Reinforcement Learning Agents for Autonomous Driving | Reinforcement learning (RL) is an effective approach to motion planning in autonomous driving, where an optimal driving policy can be automatically learned using the interaction data with the environment. Nevertheless, the reward function for an RL agent, which is significant to its performance, is challenging to be determined. The conventional work mainly focuses on rewarding safe driving states but does not incorporate the awareness of risky driving behaviors of the vehicles. In this paper, we investigate how to use risk-aware reward shaping to leverage the training and test performance of RL agents in autonomous driving. Based on the essential requirements that prescribe the safety specifications for general autonomous driving in practice, we propose additional reshaped reward terms that encourage exploration and penalize risky driving behaviors. A simulation study in OpenAI Gym indicates the advantage of risk-aware reward shaping for various RL agents. Also, we point out that proximal policy optimization (PPO) is likely to be the best RL method that works with risk-aware reward shaping. | ['Zhiyong Sun', 'Zhiqiang Ma', 'Sofie Haesaert', 'Zengjie Zhang', 'Lin-Chi Wu'] | 2023-06-05 | null | null | null | null | ['openai-gym', 'motion-planning'] | ['playing-games', 'robots'] | [-2.59270757e-01 3.75457078e-01 -6.27009988e-01 -2.38284484e-01
-5.75760186e-01 -4.10124421e-01 5.13928890e-01 -3.00881058e-01
-6.21286154e-01 9.25734758e-01 1.29845649e-01 -6.08932793e-01
-3.35555404e-01 -6.59034491e-01 -7.02202201e-01 -6.96789145e-01
-5.04998088e-01 1.38793096e-01 1.74135268e-01 -6.17538631e-01
3.27036411e-01 6.38601840e-01 -1.85599637e+00 -6.19821131e-01
1.12311649e+00 6.43236637e-01 5.12407422e-01 7.27189422e-01
2.63985306e-01 9.84714866e-01 -2.81769514e-01 8.53504315e-02
5.70394695e-01 -3.23175341e-01 -4.88857448e-01 -2.09423468e-01
-4.15596664e-01 -8.12290668e-01 -4.49899495e-01 6.69296265e-01
3.22672188e-01 4.96528864e-01 5.38073003e-01 -1.70334160e+00
1.71168894e-02 5.74506998e-01 -3.31486642e-01 2.92125523e-01
-2.53270417e-02 8.08871627e-01 8.51221085e-01 -2.72384405e-01
3.43988657e-01 1.12536812e+00 4.82377392e-04 7.21903503e-01
-6.87198758e-01 -6.72498465e-01 4.91481870e-01 2.39330575e-01
-1.14304686e+00 -1.88689947e-01 6.77878559e-01 -2.87618339e-01
8.78160715e-01 -5.59047051e-02 7.87894726e-01 9.19687927e-01
6.02180958e-01 1.02738130e+00 9.62231100e-01 -6.83058202e-02
5.80260873e-01 -1.03698425e-01 -2.27029622e-01 4.33183640e-01
2.11236164e-01 9.87739503e-01 -1.66546106e-01 -1.83146849e-01
4.96268839e-01 -2.63825685e-01 1.96833596e-01 -6.38532996e-01
-8.50971282e-01 1.05428243e+00 2.23909244e-01 -3.29090685e-01
-6.89515829e-01 5.38890600e-01 3.77139091e-01 2.66366929e-01
-1.14956699e-01 7.83272088e-01 -3.61339688e-01 -7.46165812e-01
-4.49154794e-01 8.27453792e-01 5.74391425e-01 1.01397860e+00
8.80281031e-01 5.99349976e-01 -3.58360559e-01 5.43776870e-01
4.53208238e-01 6.67358637e-01 -1.85302179e-02 -1.66444397e+00
3.64581019e-01 2.21499011e-01 6.65508211e-01 -4.47319806e-01
-4.32239115e-01 -2.41574168e-01 1.74688324e-01 8.85221064e-01
2.34948277e-01 -6.95534706e-01 -7.13177919e-01 1.69760728e+00
3.05274785e-01 2.74855405e-01 4.83190984e-01 9.19816434e-01
1.36299446e-01 5.74677229e-01 3.38300526e-01 -1.28362969e-01
8.50398898e-01 -8.38864505e-01 -8.21441472e-01 -5.49133241e-01
7.44605720e-01 -2.67655909e-01 1.09588683e+00 3.42767298e-01
-9.32003677e-01 -2.80182689e-01 -1.26123405e+00 5.06349146e-01
-3.25759090e-02 -2.51471072e-01 7.30518043e-01 6.08230650e-01
-7.65318334e-01 5.26223838e-01 -9.30721879e-01 -1.55083790e-01
2.98885256e-01 4.36456859e-01 2.18176797e-01 1.41373863e-02
-1.29700637e+00 1.36026311e+00 3.55997890e-01 -8.63283873e-02
-1.62123775e+00 -3.44852358e-01 -9.15196955e-01 -1.53287545e-01
8.79745305e-01 -1.84629232e-01 1.55908513e+00 -4.38278884e-01
-1.87851381e+00 -1.05141781e-01 2.18921855e-01 -6.49927139e-01
4.64683145e-01 -3.18282276e-01 -1.67261600e-01 -9.98327211e-02
1.32361099e-01 1.07361031e+00 8.03561807e-01 -1.42582226e+00
-9.27015960e-01 1.31535456e-01 4.59709972e-01 7.76238739e-01
5.54909259e-02 -2.63583750e-01 -4.18697409e-02 -1.81792110e-01
-5.70165992e-01 -1.30354512e+00 -8.66848946e-01 -4.81683284e-01
4.87958528e-02 -3.85690629e-01 9.55194116e-01 -1.32943794e-01
1.10760975e+00 -2.10946655e+00 -2.87457556e-01 3.47492039e-01
-1.53984159e-01 3.18362080e-02 -2.82816291e-01 4.65317428e-01
5.37471294e-01 -1.64666757e-01 1.10684745e-02 1.33706778e-01
2.49875575e-01 8.23876262e-01 -5.59329450e-01 3.29159200e-01
2.78179169e-01 9.11603808e-01 -1.24196112e+00 -2.99754947e-01
4.23731029e-01 3.67776081e-02 -7.31616557e-01 5.45443892e-01
-3.53923291e-01 5.83449841e-01 -1.00246572e+00 5.52933097e-01
3.96072090e-01 5.19257963e-01 -6.78434968e-02 5.97915351e-01
-5.53758323e-01 2.45600581e-01 -1.03350580e+00 1.09341908e+00
-3.92996818e-01 3.43634814e-01 1.99569300e-01 -5.54998517e-01
9.74145114e-01 3.12389452e-02 7.49075353e-01 -7.97888994e-01
1.81925952e-01 1.20783634e-01 3.28719258e-01 -6.10288680e-01
1.01276469e+00 1.25484630e-01 -3.64727139e-01 4.89027768e-01
-4.90952373e-01 -4.65787292e-01 9.68719125e-02 -6.48360103e-02
1.19173765e+00 4.77801710e-01 1.36893108e-01 -3.48265588e-01
1.48897499e-01 3.85442108e-01 8.45767558e-01 9.25266206e-01
-8.67853463e-01 -1.63958505e-01 5.86042702e-01 -1.20939277e-01
-9.28503931e-01 -7.53091991e-01 1.14350647e-01 1.20326424e+00
5.94032347e-01 -7.50492811e-02 -6.95758700e-01 -6.52012229e-01
1.49642050e-01 1.26649940e+00 -4.75695461e-01 -3.47565979e-01
-7.54502594e-01 -3.86565626e-01 4.27396685e-01 5.06473601e-01
2.75550663e-01 -1.26767719e+00 -1.25230896e+00 3.29832375e-01
9.95449573e-02 -9.32965040e-01 -5.11741877e-01 4.09854710e-01
-5.56066334e-01 -8.66078675e-01 -3.93169492e-01 -2.30313003e-01
4.42946225e-01 4.25207049e-01 5.52746534e-01 3.14225517e-02
1.54394895e-01 6.39502168e-01 -3.25908750e-01 -7.77089596e-01
-5.43567479e-01 4.32243245e-03 1.89807937e-01 -2.85215527e-01
2.69378960e-01 -1.48117065e-01 -7.63976276e-01 5.45279920e-01
-5.51911950e-01 -2.33761311e-01 6.75217211e-01 7.34538972e-01
4.40471888e-01 2.73407131e-01 8.13976228e-01 -3.22457165e-01
9.50216889e-01 -5.70108175e-01 -9.12390411e-01 -1.44987479e-01
-9.93768454e-01 3.46397489e-01 4.88993585e-01 -4.90492553e-01
-1.01494575e+00 1.56884447e-01 -5.29952385e-02 -2.96147019e-01
-1.67761967e-01 2.22901970e-01 -6.17541410e-02 -1.59786671e-01
4.86580819e-01 -9.01021659e-02 2.61081666e-01 2.84314096e-01
4.79172498e-01 3.35253209e-01 1.27863809e-01 -8.89956057e-01
7.31447756e-01 2.55804121e-01 1.48934675e-02 -5.91089129e-01
-3.42776060e-01 -2.35413104e-01 -2.61814129e-02 -6.94901526e-01
8.30055594e-01 -9.14958000e-01 -1.12381256e+00 -9.00913179e-02
-6.64579153e-01 -9.50379670e-01 -4.55185950e-01 6.93244398e-01
-1.23630857e+00 1.22249998e-01 -6.56737685e-02 -1.34391463e+00
1.69415712e-01 -1.64959979e+00 6.65562689e-01 4.21170384e-01
-1.12362392e-01 -6.13396525e-01 2.09680181e-02 1.74245700e-01
5.41657507e-01 8.02379027e-02 5.75817466e-01 -1.09947644e-01
-8.69642675e-01 2.98581511e-01 2.10238025e-01 7.42923245e-02
-1.77861035e-01 -4.29213904e-02 -7.49857724e-01 -3.33604068e-01
-1.28298417e-01 -4.50817347e-01 3.78085732e-01 6.65174842e-01
8.69595408e-01 -2.82388926e-01 -3.03323656e-01 1.61931083e-01
1.15014291e+00 7.51108587e-01 6.92723155e-01 7.69144952e-01
4.05394822e-01 9.08791304e-01 1.51958907e+00 6.86712027e-01
7.95951605e-01 5.96024215e-01 1.00446367e+00 1.37770429e-01
4.66427624e-01 -4.90049928e-01 8.40039730e-01 5.60402088e-02
-3.74533758e-02 5.75389750e-02 -7.81119227e-01 5.19360065e-01
-2.17539525e+00 -9.86302912e-01 8.43382329e-02 2.20133567e+00
5.56125879e-01 4.07447577e-01 3.50909382e-01 -1.99214011e-01
3.00532609e-01 6.50173286e-03 -1.11228800e+00 -8.64100516e-01
3.31611007e-01 -3.46189082e-01 1.17995334e+00 7.01626480e-01
-8.10581028e-01 1.08288038e+00 7.56016731e+00 7.08331168e-01
-8.86956036e-01 -1.17901050e-01 4.64744091e-01 -1.59202203e-01
-4.32810307e-01 2.68397361e-01 -9.92815316e-01 2.77800083e-01
1.17880285e+00 -4.22733188e-01 7.27442443e-01 1.20619929e+00
9.98125315e-01 -5.07489800e-01 -9.27049637e-01 4.78825390e-01
-5.63345790e-01 -8.81496251e-01 -6.70334220e-01 3.47457200e-01
6.67848051e-01 2.45624304e-01 1.83356404e-01 9.20407712e-01
1.06131387e+00 -1.10084176e+00 8.96203399e-01 4.04426932e-01
1.97207794e-01 -1.35081959e+00 5.70364594e-01 6.53383493e-01
-1.02163315e+00 -6.92156494e-01 -2.65544862e-01 -2.54639268e-01
3.75285536e-01 7.38376565e-03 -1.16430438e+00 1.85454085e-01
4.74496931e-01 4.28289503e-01 -1.73890680e-01 8.20778728e-01
-4.00181293e-01 6.33211792e-01 -1.04182772e-01 -4.44965065e-01
7.82781303e-01 -2.94869125e-01 6.66448653e-01 7.04598308e-01
3.35468769e-01 5.59752211e-02 6.51723683e-01 7.81458199e-01
6.70466125e-01 -1.73144221e-01 -1.17517996e+00 6.63720444e-02
5.21564007e-01 1.12384856e+00 -2.99189836e-01 5.58313616e-02
-2.75501579e-01 1.22826025e-01 1.90842941e-01 4.26291466e-01
-1.01683760e+00 -1.44075304e-01 1.22493064e+00 -9.00891796e-03
2.77382433e-01 -5.07805347e-01 -3.37560415e-01 -3.16305786e-01
-3.30726713e-01 -7.17943907e-01 9.02827606e-02 -6.91294134e-01
-8.23353529e-01 3.74708414e-01 3.78398538e-01 -1.48877478e+00
-7.31914639e-01 -3.72721046e-01 -5.73626578e-01 6.81976259e-01
-2.04429412e+00 -5.47297597e-01 5.15502430e-02 3.24892074e-01
5.98831415e-01 -2.27426782e-01 2.58523792e-01 -7.15073794e-02
-5.61233699e-01 3.19957852e-01 -1.59997836e-01 -4.92284745e-01
3.87971729e-01 -1.12313366e+00 1.97990432e-01 7.41246104e-01
-7.43737102e-01 2.71220386e-01 1.09246612e+00 -7.98372507e-01
-1.72311592e+00 -1.07597935e+00 1.11332409e-01 -3.40371937e-01
6.22912765e-01 -3.36573571e-02 -7.81495154e-01 4.48263377e-01
3.25240999e-01 -1.81977123e-01 2.27447361e-01 -1.84088603e-01
4.23734039e-01 -1.11473305e-02 -1.04739475e+00 1.19982183e+00
8.86433840e-01 -9.70202759e-02 -2.72928178e-01 -9.37851444e-02
8.18504035e-01 -5.76313019e-01 -6.40655994e-01 5.44174016e-01
3.59704137e-01 -5.63846886e-01 7.07257330e-01 -4.55405176e-01
-3.48379463e-02 -5.02483904e-01 -4.02532406e-02 -1.46067250e+00
-2.38036603e-01 -8.57476830e-01 -9.45975929e-02 6.99667275e-01
4.68821824e-01 -6.42453194e-01 9.70257819e-01 1.01751924e+00
-4.68828917e-01 -8.63642633e-01 -9.80230451e-01 -1.10961258e+00
2.20528841e-01 -7.07591951e-01 7.57404029e-01 4.07932192e-01
4.56661582e-02 -5.39945923e-02 -6.41039371e-01 4.24969077e-01
5.67214251e-01 -3.68476719e-01 1.03152800e+00 -6.36654377e-01
-1.27754480e-01 -4.55084920e-01 9.95962322e-02 -1.15190458e+00
5.46002924e-01 -4.77732509e-01 8.13776016e-01 -1.49369609e+00
-2.27579296e-01 -8.27556789e-01 -2.00230673e-01 4.26974446e-01
-2.00202629e-01 -6.12310648e-01 1.65954411e-01 5.79335820e-03
-5.57920754e-01 8.85287762e-01 1.56224537e+00 1.54575193e-02
-5.57852983e-01 1.73656806e-01 -7.40449548e-01 4.91653681e-01
1.12025046e+00 -4.50983137e-01 -1.00909531e+00 -8.63048881e-02
1.22660334e-04 3.90279591e-01 9.07908380e-02 -7.55150616e-01
1.35284305e-01 -1.17414749e+00 -4.85158563e-01 -8.66451681e-01
3.49629939e-01 -8.75747204e-01 1.48396501e-02 7.61284590e-01
-4.72873271e-01 3.00351650e-01 2.84897029e-01 7.53531277e-01
8.81620869e-02 -4.32803214e-01 6.59804225e-01 1.02203079e-01
-1.14429045e+00 4.03246939e-01 -1.03724957e+00 3.97290215e-02
1.41020846e+00 -3.28476965e-01 -1.37369365e-01 -6.07945085e-01
-1.50372386e-01 1.08331823e+00 3.11633378e-01 5.45908391e-01
7.21449852e-01 -1.34203327e+00 -6.31732225e-01 -8.33528116e-03
1.65042028e-01 -2.01259851e-01 4.97077517e-02 6.08847618e-01
-1.53743297e-01 3.80054653e-01 -3.09169859e-01 -4.45374727e-01
-7.39616930e-01 5.87310433e-01 3.77691567e-01 -2.87508368e-01
-6.22690439e-01 2.11118147e-01 1.11262098e-01 -4.11459327e-01
3.08118165e-01 -1.70011073e-01 -3.78566593e-01 -4.22994047e-01
3.15446317e-01 5.59265375e-01 -2.35961571e-01 -4.39161479e-01
-1.59260228e-01 6.40445799e-02 8.61826539e-03 -4.86780107e-01
8.93044829e-01 -2.86402434e-01 6.25751555e-01 3.19776177e-01
4.55079049e-01 -2.73486376e-01 -2.20443916e+00 2.40835935e-01
1.71296701e-01 -4.93663907e-01 3.13797683e-01 -5.07756531e-01
-7.47261107e-01 4.08063024e-01 5.26046693e-01 7.97419176e-02
7.49992311e-01 -3.26711327e-01 7.74256051e-01 4.36108530e-01
8.30915332e-01 -1.64203501e+00 1.35205045e-01 7.61956632e-01
7.59116828e-01 -1.39658368e+00 -1.93090498e-01 -6.47296235e-02
-1.32446194e+00 7.22320080e-01 1.05843735e+00 -1.22021355e-01
4.68439698e-01 4.32288229e-01 2.16695786e-01 -4.53264825e-02
-1.03976095e+00 -5.07059515e-01 -1.63108364e-01 8.10851216e-01
-1.35978952e-01 4.18340355e-01 -3.81029576e-01 1.37763932e-01
-9.89568308e-02 1.24534359e-02 8.39857996e-01 1.21269500e+00
-8.41264188e-01 -1.20844269e+00 -4.65529352e-01 4.59247440e-01
-8.78033228e-03 3.93330455e-01 1.49569735e-01 7.93419898e-01
-2.13545263e-02 1.31834710e+00 -1.11788891e-01 -3.41126919e-01
4.56713885e-01 -3.79497141e-01 1.20086610e-01 -4.99504656e-01
-3.37066472e-01 -1.77201927e-02 4.12756592e-01 -7.94690132e-01
-1.19653262e-01 -8.57934535e-01 -1.85622692e+00 -2.86073148e-01
-1.28907755e-01 1.83590770e-01 6.33527637e-01 1.02094591e+00
3.49998385e-01 5.87579608e-01 1.06317365e+00 -8.06183636e-01
-8.41201484e-01 -3.74817252e-01 -5.41904271e-01 -1.91183224e-01
4.54440594e-01 -1.20032704e+00 -2.94506460e-01 -7.08941936e-01] | [4.916833400726318, 1.622194528579712] |
66c689f1-b1cd-4ec9-b869-ec86207a83a8 | mesonet-a-compact-facial-video-forgery | 1809.00888 | null | http://arxiv.org/abs/1809.00888v1 | http://arxiv.org/pdf/1809.00888v1.pdf | MesoNet: a Compact Facial Video Forgery Detection Network | This paper presents a method to automatically and efficiently detect face
tampering in videos, and particularly focuses on two recent techniques used to
generate hyper-realistic forged videos: Deepfake and Face2Face. Traditional
image forensics techniques are usually not well suited to videos due to the
compression that strongly degrades the data. Thus, this paper follows a deep
learning approach and presents two networks, both with a low number of layers
to focus on the mesoscopic properties of images. We evaluate those fast
networks on both an existing dataset and a dataset we have constituted from
online videos. The tests demonstrate a very successful detection rate with more
than 98% for Deepfake and 95% for Face2Face. | ['Junichi Yamagishi', 'Isao Echizen', 'Vincent Nozick', 'Darius Afchar'] | 2018-09-04 | null | null | null | null | ['image-forensics', 'fake-image-detection'] | ['computer-vision', 'computer-vision'] | [ 9.51888040e-02 -1.21725872e-01 2.11349964e-01 1.97924048e-01
-2.72254854e-01 -3.42132896e-01 6.90472066e-01 -4.79976535e-01
-3.13179702e-01 7.37498701e-01 -1.64005339e-01 -1.25684813e-01
8.47085863e-02 -8.07039917e-01 -8.51300120e-01 -9.43427265e-01
-5.66989362e-01 -1.91330370e-02 2.65887856e-01 -1.05587514e-02
4.75071728e-01 7.75617659e-01 -1.86379981e+00 6.04443192e-01
1.70538709e-01 9.04829860e-01 -3.77373129e-01 8.60443354e-01
2.39101976e-01 8.45303595e-01 -1.00041187e+00 -9.67198491e-01
4.89051551e-01 -3.84579062e-01 -6.65130913e-01 -8.86338800e-02
7.21422613e-01 -9.57632422e-01 -8.26774776e-01 1.02760279e+00
7.51588166e-01 -2.74672061e-01 5.93566298e-01 -1.40429211e+00
-4.72927719e-01 2.46630162e-01 -6.70979559e-01 6.44377589e-01
4.44124490e-01 5.73962152e-01 8.03354234e-02 -9.37940180e-01
8.04572165e-01 1.50955820e+00 7.68069148e-01 8.12492251e-01
-8.14892650e-01 -9.84624684e-01 -8.39398801e-01 5.22486627e-01
-1.45246959e+00 -8.08606982e-01 8.90744090e-01 -3.30345303e-01
8.60424459e-01 -5.38411513e-02 3.70126456e-01 1.62856841e+00
2.92868555e-01 2.52347976e-01 1.25348961e+00 -4.33192641e-01
1.16335764e-03 7.69226998e-02 -3.95868868e-01 8.06186497e-01
5.54186344e-01 6.42711401e-01 -7.75080085e-01 -3.48587155e-01
8.20556819e-01 -3.84613991e-01 -1.21393643e-01 7.51269981e-02
-6.23524189e-01 8.59816551e-01 4.89541925e-02 4.34147239e-01
-1.87466890e-01 3.34228843e-01 6.72559500e-01 5.10482013e-01
2.29483739e-01 8.29599053e-02 3.87395799e-01 -3.52340303e-02
-1.27759266e+00 3.49587888e-01 8.21424246e-01 5.04649878e-01
5.28196454e-01 3.59589636e-01 8.35387707e-02 3.45099032e-01
1.35608122e-01 5.36120832e-01 3.12576354e-01 -1.07691205e+00
2.39869386e-01 -3.93363163e-02 2.11062089e-01 -1.58198702e+00
1.98806524e-01 1.92346662e-01 -5.52748561e-01 6.28798306e-01
3.88843268e-01 -2.36335993e-01 -6.87573493e-01 1.29131913e+00
4.05127048e-01 5.47468483e-01 6.23115599e-02 7.37079144e-01
1.00534070e+00 6.65407062e-01 -2.09995806e-01 -2.94910073e-01
1.21882701e+00 -8.05241823e-01 -7.40781963e-01 1.68135732e-01
5.03476322e-01 -7.65493870e-01 6.42600775e-01 7.78911591e-01
-1.16924644e+00 -3.79747450e-01 -1.16457427e+00 2.56792992e-01
-4.75279540e-01 -2.49733612e-01 3.47870827e-01 1.35588169e+00
-1.38208532e+00 9.07606661e-01 -2.71478176e-01 -1.13967389e-01
8.71533036e-01 5.16804636e-01 -6.88340366e-01 -1.61557555e-01
-1.08882666e+00 6.63204730e-01 3.61780494e-01 -1.27965985e-02
-1.40164804e+00 -3.42289329e-01 -4.44669425e-01 -1.80087350e-02
9.22349095e-02 -1.58705488e-01 7.70560443e-01 -1.10613739e+00
-1.42063940e+00 1.15985870e+00 1.40906513e-01 -7.32483804e-01
8.85032713e-01 -2.06622239e-02 -6.66703343e-01 9.86419976e-01
-4.28817213e-01 4.96157676e-01 1.51458585e+00 -1.43885243e+00
-8.86328816e-02 -2.83928871e-01 -8.92200470e-02 -4.66455966e-01
-7.26180792e-01 5.95189035e-01 -2.04827972e-02 -5.30747712e-01
-7.77113795e-01 -5.52564204e-01 5.07496953e-01 1.37580201e-01
-2.50223905e-01 1.02461711e-01 1.46208251e+00 -9.04159307e-01
9.02779758e-01 -1.93650973e+00 -3.94483328e-01 4.04409580e-02
4.10243899e-01 1.01289141e+00 -3.51166517e-01 6.53726995e-01
-1.71351939e-01 3.00415903e-01 -4.62588780e-02 -4.13146287e-01
-4.55602050e-01 -4.16610278e-02 -4.02215123e-01 8.96755397e-01
2.69693226e-01 6.00540400e-01 -9.25765574e-01 -6.73461556e-01
2.97281742e-01 8.51564527e-01 -3.97108346e-01 1.94625780e-01
7.30308965e-02 1.48234040e-01 1.91041864e-02 8.01712871e-01
1.34637213e+00 2.79181033e-01 6.95671961e-02 -2.70876259e-01
2.00524796e-02 -1.58138648e-01 -6.32536232e-01 1.01113236e+00
-3.84534821e-02 8.48171949e-01 3.41296017e-01 -1.01111698e+00
7.60571122e-01 3.69214684e-01 4.77336913e-01 -7.18578637e-01
3.87937963e-01 3.15085500e-01 -6.95722103e-02 -1.11718988e+00
2.89900839e-01 -6.21182360e-02 6.44706488e-01 6.06686950e-01
3.47297788e-01 2.76653528e-01 1.87948167e-01 9.76213589e-02
1.25423765e+00 5.51701225e-02 -3.71121556e-01 -2.93759197e-01
7.26138055e-01 -5.24100482e-01 2.60168295e-02 7.68858075e-01
-4.44901258e-01 5.37060320e-01 7.11093962e-01 -7.87570953e-01
-1.21755219e+00 -6.48109257e-01 7.57324100e-02 4.55672592e-01
1.70704409e-01 -3.71794909e-01 -1.31266224e+00 -8.84264171e-01
8.97758175e-03 1.91860005e-01 -6.62797749e-01 -3.16521078e-01
-6.38223767e-01 -7.30757475e-01 1.28541124e+00 -1.43322289e-01
9.18640554e-01 -1.35865331e+00 -6.56488478e-01 -1.78262770e-01
-1.82075556e-02 -1.35418141e+00 5.65000027e-02 -6.22603178e-01
-5.26308119e-01 -1.43722785e+00 -5.63149512e-01 -7.38485277e-01
3.56049895e-01 5.20105004e-01 8.90627742e-01 7.87318408e-01
-5.29604077e-01 2.64768869e-01 -2.97921509e-01 -4.85795811e-02
-9.95375395e-01 -4.77816075e-01 2.48161688e-01 3.86521637e-01
5.22091389e-01 -5.95721900e-01 -7.15228975e-01 2.53458291e-01
-1.14268720e+00 -4.82449412e-01 3.92731339e-01 5.78412473e-01
-1.67345300e-01 2.94560850e-01 3.02694917e-01 -5.30204594e-01
2.64097840e-01 -5.10420322e-01 -5.51176190e-01 4.32415605e-02
-3.16469252e-01 -3.20366323e-01 7.25400686e-01 -4.75023776e-01
-8.90325069e-01 -1.99723735e-01 -2.47119829e-01 -8.75348032e-01
-1.82314143e-01 -1.73815638e-01 1.60446502e-02 -9.58327711e-01
6.90512598e-01 3.22884977e-01 3.33841681e-01 -1.69941396e-01
-2.53275782e-01 6.86800361e-01 6.27875507e-01 -2.98679709e-01
1.16168892e+00 9.90310550e-01 1.34857610e-01 -1.28181636e+00
-1.98805928e-02 -6.08584397e-02 -3.75168234e-01 -7.09217072e-01
5.68170905e-01 -7.44083107e-01 -1.02124631e+00 1.17104590e+00
-1.37484348e+00 -2.66422927e-01 2.41128162e-01 2.56601125e-01
-5.41023791e-01 1.09946692e+00 -7.89056540e-01 -9.95708346e-01
-4.43632156e-01 -1.06502402e+00 1.02596641e+00 -3.93473282e-02
3.70508701e-01 -1.06115794e+00 -5.84965013e-02 4.48413283e-01
6.05230153e-01 8.08192074e-01 5.86552024e-01 -3.45647812e-01
-6.33185804e-01 -2.05477878e-01 -4.74437803e-01 4.64448422e-01
3.83733859e-04 4.03717220e-01 -1.49454570e+00 -6.46626711e-01
3.95223558e-01 -6.45693481e-01 9.33185399e-01 -1.86917726e-02
1.33186519e+00 -6.33556366e-01 -2.89725870e-01 7.50989139e-01
1.51643515e+00 1.66141130e-02 1.31910264e+00 2.66731739e-01
5.45080841e-01 7.18304873e-01 1.68553755e-01 4.86562461e-01
-3.56724679e-01 5.32641530e-01 8.21505308e-01 1.24525324e-01
-3.25575501e-01 -2.30489373e-01 8.08078706e-01 3.16108853e-01
-2.46715039e-01 -6.65400565e-01 -5.89190722e-01 3.47926527e-01
-1.22063458e+00 -1.42780828e+00 -2.55989581e-01 2.23477077e+00
4.82970804e-01 -7.86322132e-02 2.27741435e-01 2.65052855e-01
1.06741524e+00 3.34116489e-01 -2.09601045e-01 -4.31040615e-01
-3.37725461e-01 3.41472805e-01 4.30242568e-01 2.94576585e-01
-1.10630918e+00 9.55032349e-01 7.27180290e+00 1.10425997e+00
-1.32617259e+00 2.09689483e-01 7.50797212e-01 -1.12716570e-01
9.98863280e-02 -2.51211971e-01 -6.02117658e-01 8.11139941e-01
1.30459034e+00 1.97886810e-01 4.46237683e-01 4.82236475e-01
2.46904820e-01 -8.80416185e-02 -7.32318819e-01 1.40321934e+00
4.82580453e-01 -1.75856435e+00 3.90784562e-01 6.15827382e-01
5.02675116e-01 -2.04435870e-01 2.39975557e-01 -1.66803151e-01
-7.70908669e-02 -1.28989530e+00 6.17341101e-01 2.32700914e-01
1.04865289e+00 -8.89863849e-01 7.61600912e-01 2.52290210e-05
-6.77787960e-01 -2.49486536e-01 -5.74308932e-01 1.87052995e-01
-1.21789776e-01 4.57114786e-01 -7.12863564e-01 2.37723589e-01
9.73908961e-01 5.30437410e-01 -5.74278176e-01 9.89902437e-01
2.23706335e-01 5.98775983e-01 -3.26044738e-01 2.29046613e-01
2.03884885e-01 1.32273212e-01 6.15890563e-01 1.26482022e+00
6.49938524e-01 -4.38239455e-01 -4.16542441e-01 8.67350996e-01
-4.20886874e-01 -8.76071602e-02 -1.12204230e+00 -2.11489454e-01
4.95636523e-01 1.37911499e+00 -7.48588622e-01 -1.83607042e-01
-3.12443405e-01 9.60563064e-01 -8.96041542e-02 4.83771879e-03
-8.89391124e-01 -3.33156437e-01 5.93710005e-01 4.22881752e-01
4.60656703e-01 -1.66224718e-01 4.77537453e-01 -1.07575524e+00
-6.77597374e-02 -1.01144302e+00 2.09422529e-01 -6.17301285e-01
-1.16157544e+00 5.88008761e-01 -7.08425269e-02 -9.46121514e-01
-1.65155470e-01 -9.23557937e-01 -7.64294267e-01 3.65674376e-01
-1.68924713e+00 -1.12913156e+00 -5.12791097e-01 7.65359223e-01
2.06419691e-01 -5.19179106e-01 5.37058473e-01 5.13994515e-01
-5.08420289e-01 8.52010667e-01 2.74474412e-01 3.07319522e-01
7.29490578e-01 -6.01255417e-01 5.21219492e-01 8.61371338e-01
-1.08168609e-01 3.81922275e-01 7.43284643e-01 -5.89762926e-01
-1.63488460e+00 -7.37191021e-01 5.83927810e-01 -3.80535424e-01
5.44620097e-01 -5.66069424e-01 -8.98899734e-01 7.37425983e-02
5.37047505e-01 1.43658042e-01 3.97830933e-01 -9.54054594e-01
-5.77439308e-01 1.87059660e-02 -1.88610756e+00 2.93745548e-01
1.01518142e+00 -6.51264191e-01 -1.05147675e-01 5.90292573e-01
2.26030633e-01 4.16112915e-02 -6.45998895e-01 1.32444456e-01
5.97699761e-01 -1.88695502e+00 1.10082507e+00 -3.29751343e-01
5.54937720e-01 -2.16328576e-02 1.39369115e-01 -7.17821479e-01
1.67399958e-01 -1.23597467e+00 -6.63679779e-01 1.16556072e+00
-4.04039800e-01 -4.99746591e-01 1.04113925e+00 -1.01672061e-01
3.99768263e-01 -2.33674437e-01 -1.13460767e+00 -1.09106159e+00
-1.96412094e-02 1.59858558e-02 6.40515924e-01 9.16691840e-01
-5.01443386e-01 -3.65789682e-01 -7.88410425e-01 7.26981163e-02
9.08536673e-01 -3.34888458e-01 8.12551558e-01 -9.81980145e-01
-4.76007834e-02 -2.21865177e-01 -8.06408226e-01 -2.69443899e-01
2.45179474e-01 -2.41180986e-01 -3.37836027e-01 -4.07903939e-01
2.44996980e-01 1.39945462e-01 1.08035564e-01 9.20668766e-02
3.76036793e-01 9.19779420e-01 -6.18028454e-02 2.31510729e-01
-1.60641462e-01 3.09321493e-01 8.66069019e-01 5.81381209e-02
4.78121340e-01 -4.46415216e-01 -3.00406873e-01 5.69722235e-01
6.34728134e-01 -5.85350931e-01 -1.37388602e-01 -3.75157505e-01
4.57049049e-02 -1.97828799e-01 8.62417758e-01 -1.42200291e+00
-1.61098406e-01 1.14946753e-01 5.24325907e-01 -1.53978631e-01
4.82330710e-01 -5.72936475e-01 1.02403678e-01 8.67230654e-01
5.71808033e-02 -1.03915250e-02 9.33947191e-02 4.66089636e-01
-1.89762935e-01 -4.85236526e-01 1.13641524e+00 -4.26468790e-01
-4.99355406e-01 3.13265979e-01 -4.49287266e-01 -2.01432526e-01
1.10937560e+00 -5.34137428e-01 -6.34879351e-01 -5.41520536e-01
-2.92448848e-01 -4.66669798e-01 7.92873025e-01 3.63531202e-01
1.00795114e+00 -1.15396285e+00 -6.45382643e-01 2.99070448e-01
-1.95812255e-01 -8.18963885e-01 3.85673642e-01 4.23717916e-01
-1.16790175e+00 2.15734020e-01 -7.04169333e-01 -2.97072351e-01
-1.47969508e+00 9.61358964e-01 5.97148895e-01 1.36713192e-01
-5.14227152e-01 6.99695528e-01 -8.83349329e-02 1.10345520e-01
-4.62510772e-02 4.86467808e-01 -2.24777088e-01 -9.82903317e-02
1.13385630e+00 6.85321271e-01 1.85479969e-01 -1.06327724e+00
-2.88789541e-01 4.49583620e-01 -4.26293537e-02 1.37962639e-01
1.32767749e+00 4.45895828e-02 -2.87843585e-01 -4.08481658e-01
1.41498971e+00 1.48358896e-01 -1.45612025e+00 1.87300786e-01
-9.37801152e-02 -1.04433715e+00 1.79460421e-02 -2.87008226e-01
-1.32846856e+00 1.06721902e+00 8.44801009e-01 4.24896479e-01
8.72147381e-01 -3.03103775e-01 1.06957710e+00 3.14744771e-01
4.13754880e-01 -1.01954436e+00 6.88089132e-01 6.64443672e-02
6.95413828e-01 -1.02991712e+00 2.30252557e-02 -4.80427682e-01
-5.66170849e-02 1.51738012e+00 5.24549127e-01 -4.20775890e-01
5.04897475e-01 2.78152138e-01 -2.26076484e-01 -4.51925755e-01
-4.80578184e-01 2.80612886e-01 -4.36479837e-01 9.81311262e-01
-5.74367084e-02 -5.67268610e-01 -5.42951263e-02 -2.28852779e-01
-2.50707269e-02 6.85212761e-02 1.00650024e+00 8.01026464e-01
-2.92026013e-01 -9.68676031e-01 -5.83448529e-01 6.09095059e-02
-8.91996026e-01 2.30372265e-01 -7.70227253e-01 8.53167236e-01
3.10789824e-01 9.60095823e-01 1.44120296e-02 -6.28436148e-01
-3.82977128e-01 -3.32132690e-02 8.44870567e-01 5.32977562e-03
-6.35728896e-01 -3.20793241e-01 -1.27458721e-02 -7.28589118e-01
-7.22833812e-01 -4.95733410e-01 -7.45859742e-01 -1.07577884e+00
-2.96349734e-01 5.33372089e-02 5.74558198e-01 8.14702809e-01
4.39674050e-01 -3.88677686e-01 1.04298687e+00 -1.19777846e+00
-3.24446827e-01 -7.01529026e-01 -5.47529936e-01 5.06838799e-01
5.59025586e-01 -7.63846278e-01 -7.28572369e-01 7.88019374e-02] | [12.573691368103027, 1.1084764003753662] |
47e2110a-2c22-4e5f-a8cf-b23e06fb2af1 | semi-supervised-learning-for-neural-keyphrase | 1808.06773 | null | https://arxiv.org/abs/1808.06773v2 | https://arxiv.org/pdf/1808.06773v2.pdf | Semi-Supervised Learning for Neural Keyphrase Generation | We study the problem of generating keyphrases that summarize the key points for a given document. While sequence-to-sequence (seq2seq) models have achieved remarkable performance on this task (Meng et al., 2017), model training often relies on large amounts of labeled data, which is only applicable to resource-rich domains. In this paper, we propose semi-supervised keyphrase generation methods by leveraging both labeled data and large-scale unlabeled samples for learning. Two strategies are proposed. First, unlabeled documents are first tagged with synthetic keyphrases obtained from unsupervised keyphrase extraction methods or a selflearning algorithm, and then combined with labeled samples for training. Furthermore, we investigate a multi-task learning framework to jointly learn to generate keyphrases as well as the titles of the articles. Experimental results show that our semi-supervised learning-based methods outperform a state-of-the-art model trained with labeled data only. | ['Hai Ye', 'Lu Wang'] | 2018-08-21 | semi-supervised-learning-for-neural-keyphrase-1 | https://aclanthology.org/D18-1447 | https://aclanthology.org/D18-1447.pdf | emnlp-2018-10 | ['keyphrase-generation'] | ['natural-language-processing'] | [ 4.10129637e-01 -9.10418630e-02 -6.40399575e-01 -9.65741202e-02
-1.54462171e+00 -9.27488387e-01 9.10693467e-01 3.54298800e-01
-4.91821975e-01 1.22729850e+00 5.43028712e-01 -2.07585812e-01
2.54927099e-01 -5.87800980e-01 -9.30549979e-01 -4.72783834e-01
1.92860812e-01 4.79452074e-01 2.09414661e-02 -1.32895306e-01
5.09923637e-01 2.93011274e-02 -1.36438417e+00 5.23756027e-01
1.11897743e+00 5.62881649e-01 4.07844186e-01 8.09660316e-01
-6.62209988e-01 1.16040516e+00 -6.32839262e-01 -2.40045592e-01
1.46542147e-01 -7.11975276e-01 -7.58616388e-01 7.16534555e-02
5.92268527e-01 -5.49451828e-01 -2.97847182e-01 9.56319690e-01
4.49236453e-01 -7.72091448e-02 7.90606797e-01 -1.15471983e+00
-6.01879716e-01 1.08751106e+00 -5.02981663e-01 3.98501493e-02
3.89786601e-01 -1.63541049e-01 1.48001420e+00 -1.07453585e+00
7.86013067e-01 9.20892000e-01 2.28576124e-01 4.88014638e-01
-1.02227855e+00 -6.66177809e-01 1.41061172e-01 6.45653382e-02
-1.10968864e+00 -4.04685944e-01 9.22390938e-01 -2.79408395e-01
3.88896465e-01 -2.11400958e-03 5.70286036e-01 1.17483592e+00
1.30637243e-01 1.56562471e+00 1.20612180e+00 -5.65963686e-01
1.98998764e-01 4.44967337e-02 5.61448187e-02 6.87428057e-01
1.75871789e-01 -1.06690593e-01 -7.69671500e-01 -4.10377324e-01
5.43014228e-01 2.55614575e-02 -4.44720946e-02 -1.74277604e-01
-1.81635165e+00 8.50117922e-01 -8.63272101e-02 2.16342375e-01
-5.97186148e-01 -1.09442420e-01 5.49820781e-01 4.18140739e-01
7.35901117e-01 8.36864054e-01 -6.86333001e-01 -1.79545999e-01
-1.42828119e+00 6.72721148e-01 1.01444268e+00 1.12619150e+00
9.62288558e-01 -8.23389813e-02 -6.23044074e-01 7.68231094e-01
-1.57840520e-01 6.18321419e-01 6.37499988e-01 -5.66620588e-01
7.29625940e-01 5.66966891e-01 3.49406362e-01 -6.73883259e-01
-6.46787658e-02 -2.04278797e-01 -7.23092258e-01 -4.62302357e-01
2.75644451e-01 -5.39378822e-01 -9.98384535e-01 1.33588815e+00
4.16575596e-02 2.06548408e-01 2.57750034e-01 4.28994268e-01
9.40372050e-01 1.13108969e+00 -1.73106954e-01 -5.51351905e-01
1.16175783e+00 -1.41150153e+00 -8.65918100e-01 -3.61747555e-02
7.16096818e-01 -8.47999990e-01 9.88340676e-01 5.09827137e-01
-9.74171400e-01 -6.25652313e-01 -7.89170027e-01 3.19294073e-02
-3.59083384e-01 5.55082202e-01 3.55955511e-01 1.25862077e-01
-6.35557413e-01 3.77374321e-01 -4.75398451e-01 -9.22259614e-02
3.44722539e-01 -8.65054876e-02 2.94170957e-02 -1.48179322e-01
-1.17290223e+00 4.06613916e-01 7.86082506e-01 -3.81806940e-01
-1.19195819e+00 -9.49012399e-01 -6.65289462e-01 -4.02825885e-02
9.74554479e-01 -5.32405674e-01 1.60113478e+00 -7.78753281e-01
-1.48189127e+00 6.18460417e-01 -1.69583768e-01 -4.72984493e-01
4.55086261e-01 -5.73831141e-01 -3.20704997e-01 4.05348271e-01
4.69957203e-01 7.17910707e-01 1.09009290e+00 -1.33214009e+00
-1.05456972e+00 3.90124694e-02 -6.40706941e-02 8.30088556e-02
-5.97748041e-01 4.69251573e-02 -5.08109868e-01 -1.14754832e+00
-4.42505777e-01 -9.08198357e-01 -3.82509172e-01 -5.61060786e-01
-8.94467771e-01 -3.64812583e-01 6.51946306e-01 -7.24354088e-01
1.34937596e+00 -1.65607703e+00 1.79657526e-02 4.42160591e-02
1.12319753e-01 3.89273703e-01 -4.51713383e-01 8.44868362e-01
1.84647694e-01 1.74865231e-01 -1.20148845e-01 -1.63090378e-01
-9.84689519e-02 -2.08927602e-01 -8.83408487e-01 -1.87793598e-01
2.12823257e-01 1.09586096e+00 -1.35456979e+00 -8.69028509e-01
-2.93834686e-01 -1.46709979e-01 -1.93626523e-01 5.53625166e-01
-8.42042804e-01 4.22146589e-01 -8.57091725e-01 5.86811483e-01
2.89053053e-01 -4.37682331e-01 1.28169462e-01 -2.15527251e-01
-1.51210666e-01 3.91644418e-01 -1.04771769e+00 1.76469910e+00
-5.02237201e-01 2.66281575e-01 -3.10136646e-01 -9.42316592e-01
9.40623701e-01 4.24336165e-01 6.15431786e-01 -3.79877895e-01
-2.14740768e-01 3.40433776e-01 -4.57547128e-01 -4.14137721e-01
9.10361886e-01 1.42224476e-01 -3.27880025e-01 9.69257116e-01
2.52058446e-01 -3.25716347e-01 7.32313931e-01 6.50951922e-01
9.50126171e-01 2.35232577e-01 4.47013944e-01 6.25973288e-03
5.64036787e-01 2.68921524e-01 3.29711765e-01 1.05740428e+00
4.53924567e-01 5.01129746e-01 6.54087901e-01 -2.43065104e-01
-1.37806046e+00 -6.29385173e-01 3.76347303e-01 1.11082494e+00
-2.34548822e-01 -9.19741273e-01 -7.07420468e-01 -1.19253695e+00
-1.01706959e-01 7.20752478e-01 -4.49507684e-01 1.63330939e-02
-5.42691469e-01 -4.56352055e-01 5.75245976e-01 3.54934067e-01
2.41411090e-01 -1.15146589e+00 -2.11688608e-01 4.06078011e-01
-2.69129962e-01 -1.20241761e+00 -8.23082387e-01 3.56748961e-02
-6.75093770e-01 -8.94804001e-01 -1.30246675e+00 -8.15862596e-01
6.98911250e-01 3.27416182e-01 1.18564653e+00 -2.88012564e-01
9.11546778e-03 4.37343508e-01 -7.42183030e-01 -5.14924526e-01
-7.15235174e-01 5.58030188e-01 -4.84741330e-02 8.83907154e-02
1.57003745e-01 -1.86374947e-01 -1.36297226e-01 -1.90532371e-01
-1.05844581e+00 3.28319609e-01 1.14164221e+00 1.06648123e+00
7.77208090e-01 6.76397160e-02 8.73966753e-01 -1.25248456e+00
1.00365281e+00 -3.25943291e-01 -5.51487386e-01 5.96822143e-01
-7.50064254e-01 4.18166459e-01 9.89519656e-01 -5.59880018e-01
-1.08258343e+00 -2.86322217e-02 3.34393501e-01 -3.83623272e-01
-8.36729631e-02 7.95066118e-01 5.66158537e-03 3.84116679e-01
4.09839123e-01 8.27674270e-01 -1.18964247e-01 -4.38418299e-01
6.97185755e-01 7.75352299e-01 4.50720638e-01 -7.37982690e-01
1.14765573e+00 2.02605620e-01 -2.42602900e-01 -8.54842782e-01
-1.36882961e+00 -5.30525267e-01 -7.36796498e-01 -6.11785948e-02
5.63404739e-01 -1.15845120e+00 -3.15467715e-02 4.67713624e-01
-1.16208994e+00 -2.75350243e-01 -3.33753616e-01 4.59361613e-01
-5.15093684e-01 5.36506951e-01 -4.90865350e-01 -5.13951540e-01
-7.48994231e-01 -7.77123868e-01 1.29924858e+00 2.66785413e-01
-2.31129825e-01 -8.89612079e-01 2.97489524e-01 2.62837350e-01
-7.25673735e-02 -2.96740979e-02 8.89388263e-01 -1.22854781e+00
-6.26536906e-01 -4.19148207e-01 -1.01109594e-01 3.60655218e-01
4.51385647e-01 -7.33014336e-03 -4.17046487e-01 -1.62992194e-01
-3.99434835e-01 -1.02413583e+00 1.07909179e+00 6.54763952e-02
1.33097982e+00 -6.49164081e-01 -4.74737763e-01 1.25542313e-01
1.04874587e+00 -3.57248671e-02 1.42771468e-01 2.83925444e-01
9.88567114e-01 6.21531606e-01 9.53880131e-01 5.94705939e-01
5.47334969e-01 2.92941272e-01 -2.23198965e-01 3.22165377e-02
1.48321226e-01 -8.37077916e-01 4.04996663e-01 1.07199192e+00
2.87855715e-01 -2.79499561e-01 -7.75328755e-01 6.46607518e-01
-1.88367224e+00 -9.48055148e-01 9.45154279e-02 1.97701168e+00
1.49080718e+00 2.99458653e-01 3.51702124e-01 3.42784896e-02
6.01823628e-01 5.35720766e-01 -4.02778596e-01 2.07736954e-01
-1.02135576e-01 1.92018926e-01 4.59383547e-01 1.58220738e-01
-1.21604216e+00 1.29248214e+00 5.85631752e+00 1.08836508e+00
-1.03609812e+00 -2.43182734e-01 4.11841154e-01 7.16212168e-02
-6.30246937e-01 3.98875400e-02 -1.12865233e+00 5.39140284e-01
8.85657012e-01 -4.34901059e-01 1.86193764e-01 8.09593141e-01
1.43915623e-01 1.09543335e-02 -1.02820039e+00 8.51784110e-01
2.97843248e-01 -1.59947109e+00 4.45966303e-01 -1.08572178e-01
1.25103748e+00 -2.22252414e-01 -1.49471804e-01 3.22445005e-01
6.39238358e-01 -6.41001701e-01 5.88523507e-01 4.97136950e-01
6.94085836e-01 -7.68923938e-01 4.97713238e-01 6.91545188e-01
-1.01477659e+00 9.31301564e-02 -1.78146899e-01 3.69362533e-01
1.86756760e-01 8.50114703e-01 -1.00862312e+00 7.69126952e-01
2.92811453e-01 1.02023470e+00 -5.90006411e-01 8.18669558e-01
-6.42072320e-01 9.39206541e-01 2.06407420e-02 -5.78305483e-01
4.39319849e-01 -1.20890610e-01 3.77448499e-01 1.39734459e+00
2.47539118e-01 -3.44971311e-04 7.09776759e-01 6.94192231e-01
-4.27857101e-01 3.67230684e-01 -6.08032703e-01 -8.52205575e-01
5.08568227e-01 1.57877612e+00 -8.39567125e-01 -8.55806828e-01
-4.90456849e-01 9.43145931e-01 1.27652064e-01 4.23072636e-01
-2.79418856e-01 -5.72116792e-01 -1.89496018e-02 -1.95298165e-01
5.16636848e-01 -3.09045851e-01 -1.00726701e-01 -1.44832253e+00
-8.53599757e-02 -1.29753149e+00 2.79807419e-01 -6.60332382e-01
-1.33014989e+00 3.86417866e-01 6.18357770e-02 -1.48053956e+00
-6.00214183e-01 -2.03594655e-01 -6.36106491e-01 6.87449038e-01
-1.68145919e+00 -1.38660467e+00 2.86917333e-02 2.39833817e-01
8.25041831e-01 -5.69726944e-01 7.22215176e-01 -2.11299613e-01
-2.81567663e-01 4.71418768e-01 3.91753227e-01 4.44919527e-01
1.01424921e+00 -1.45259285e+00 6.02886736e-01 7.58828163e-01
5.12026548e-01 6.64503634e-01 3.73616397e-01 -9.44801211e-01
-1.57734048e+00 -1.23012173e+00 1.00677824e+00 -2.51820505e-01
8.19635808e-01 -4.80658233e-01 -7.95458794e-01 4.70251292e-01
4.79040623e-01 -1.07373491e-01 8.55626702e-01 -3.20857875e-02
-3.28321844e-01 -9.59025472e-02 -4.25762773e-01 8.36753249e-01
5.88525712e-01 -6.85461044e-01 -7.66194880e-01 5.75605690e-01
8.47054183e-01 -3.27328026e-01 -6.22032762e-01 2.77113229e-01
4.13755864e-01 -3.94349366e-01 7.50627398e-01 -8.17807913e-01
9.55084026e-01 -1.61681980e-01 2.12149501e-01 -1.65469587e+00
-4.65358384e-02 -9.19204235e-01 -4.95495200e-01 1.39977872e+00
5.59560597e-01 -3.96601819e-02 7.79965401e-01 -2.19937563e-01
8.41689855e-02 -7.11922348e-01 -2.66243666e-01 -7.36225665e-01
4.42841649e-02 -9.03880447e-02 4.77724820e-01 9.00508046e-01
-2.45082416e-02 9.43179488e-01 -7.34754860e-01 -2.11878598e-01
5.95844388e-01 5.46263933e-01 1.12150609e+00 -1.07559621e+00
-6.44472912e-02 -1.94759816e-01 4.45533037e-01 -1.21895766e+00
5.34230351e-01 -8.94412398e-01 1.77130193e-01 -1.75865030e+00
3.59647274e-01 -1.77565649e-01 -1.72794819e-01 4.45753098e-01
-8.15408051e-01 -2.18006685e-01 8.38608444e-02 4.57079202e-01
-9.26363230e-01 6.88132524e-01 1.43316376e+00 -3.59884024e-01
-3.08908582e-01 1.41838685e-01 -7.43012846e-01 4.28618580e-01
6.11388922e-01 -5.55164039e-01 -5.79143345e-01 1.73088238e-02
3.50807190e-01 1.94391176e-01 1.05821095e-01 -6.61671400e-01
3.50257456e-01 -3.75724494e-01 2.96059221e-01 -1.02915084e+00
-1.35222152e-01 -3.77407789e-01 -6.25079811e-01 3.41011554e-01
-9.10854757e-01 -7.73350000e-02 -6.19927831e-02 6.99056566e-01
-3.46071929e-01 -4.05914456e-01 3.06964070e-01 -5.36604524e-01
-6.98135436e-01 4.22849327e-01 -4.31075275e-01 4.65266407e-01
7.12054312e-01 3.40766102e-01 -1.77899927e-01 -6.08386993e-01
-9.18388143e-02 3.72390777e-01 2.15931758e-01 4.98396397e-01
5.96444309e-01 -1.27795517e+00 -1.16073060e+00 -7.55368844e-02
4.33728099e-01 1.18299611e-01 -2.12412059e-01 4.10514206e-01
-8.21398273e-02 6.98039174e-01 -1.38918879e-02 -2.16558799e-01
-8.92391801e-01 7.23660469e-01 -4.50684518e-01 -7.20632732e-01
-3.41661662e-01 6.13482654e-01 -1.41315281e-01 -6.33996427e-01
1.46441162e-01 -2.54343003e-01 -3.72644842e-01 5.97661078e-01
7.11847484e-01 1.47063494e-01 -4.06193212e-02 -1.28835186e-01
1.69798732e-01 3.04239124e-01 -8.14500690e-01 -3.38548064e-01
1.28495216e+00 2.36372113e-01 -9.03646201e-02 6.28537178e-01
1.07526314e+00 2.47476488e-01 -1.02593160e+00 -8.03064704e-01
2.88350731e-01 -8.25632587e-02 -1.01611800e-01 -7.29191005e-01
-6.08130932e-01 6.12439811e-01 -1.34461209e-01 1.69575557e-01
8.35454762e-01 1.91108473e-02 1.11810207e+00 8.42789054e-01
2.09732816e-01 -1.43696785e+00 7.72909701e-01 6.72913849e-01
6.80159807e-01 -1.32537472e+00 1.92189127e-01 -1.10934146e-01
-8.41098368e-01 1.14655435e+00 5.22664070e-01 -4.34302818e-03
4.39123064e-01 2.54478484e-01 1.24084003e-01 1.76981136e-01
-9.28761244e-01 -2.25612924e-01 4.25650388e-01 3.69042665e-01
3.93848628e-01 -2.26861045e-01 -3.70861202e-01 6.59542322e-01
-1.68453470e-01 2.46027097e-01 6.12447441e-01 1.25536728e+00
-5.16843617e-01 -1.37776005e+00 -2.72955000e-01 8.43105137e-01
-6.09374166e-01 -4.83558685e-01 -6.96204126e-01 3.52789462e-01
-2.65164822e-01 6.95705414e-01 -2.39907458e-01 -3.74468923e-01
4.04188298e-02 3.59686077e-01 2.94720143e-01 -1.01155949e+00
-4.40939367e-01 3.16692829e-01 1.01697721e-01 4.58256043e-02
-5.19704342e-01 -6.36302710e-01 -1.02081025e+00 9.59572345e-02
-1.78386375e-01 6.58901572e-01 3.42726141e-01 1.15866613e+00
2.80078202e-01 6.18816197e-01 1.04092646e+00 -7.74394274e-01
-8.93648922e-01 -1.20672905e+00 -5.12931108e-01 5.79519309e-02
5.21809936e-01 -9.41289812e-02 -1.31436449e-03 5.83144426e-01] | [12.31902027130127, 8.874542236328125] |
b8f69584-e91d-4f13-aec5-78512543b914 | unsupervised-feature-learning-of-human | 1812.02592 | null | http://arxiv.org/abs/1812.02592v1 | http://arxiv.org/pdf/1812.02592v1.pdf | Unsupervised Feature Learning of Human Actions as Trajectories in Pose Embedding Manifold | An unsupervised human action modeling framework can provide useful
pose-sequence representation, which can be utilized in a variety of pose
analysis applications. In this work we propose a novel temporal pose-sequence
modeling framework, which can embed the dynamics of 3D human-skeleton joints to
a continuous latent space in an efficient manner. In contrast to end-to-end
framework explored by previous works, we disentangle the task of individual
pose representation learning from the task of learning actions as a trajectory
in pose embedding space. In order to realize a continuous pose embedding
manifold with improved reconstructions, we propose an unsupervised, manifold
learning procedure named Encoder GAN, (or EnGAN). Further, we use the pose
embeddings generated by EnGAN to model human actions using a bidirectional RNN
auto-encoder architecture, PoseRNN. We introduce first-order gradient loss to
explicitly enforce temporal regularity in the predicted motion sequence. A
hierarchical feature fusion technique is also investigated for simultaneous
modeling of local skeleton joints along with global pose variations. We
demonstrate state-of-the-art transfer-ability of the learned representation
against other supervisedly and unsupervisedly learned motion embeddings for the
task of fine-grained action recognition on SBU interaction dataset. Further, we
show the qualitative strengths of the proposed framework by visualizing
skeleton pose reconstructions and interpolations in pose-embedding space, and
low dimensional principal component projections of the reconstructed pose
trajectories. | ['R. Venkatesh Babu', 'Maharshi Gor', 'Jogendra Nath Kundu', 'Phani Krishna Uppala'] | 2018-12-06 | null | null | null | null | ['fine-grained-action-recognition'] | ['computer-vision'] | [ 0.24826929 0.19790398 -0.34004217 -0.2559172 -0.82741874 -0.23672159
0.8975464 -0.6069137 -0.26939297 0.543449 0.8146414 0.33966014
-0.19418672 -0.4509422 -0.9044548 -0.6988937 0.00723821 0.60052
-0.06601509 -0.21758309 -0.09703316 0.5547472 -1.3666073 0.2818844
0.44215962 0.60061234 -0.02457077 0.6385552 0.25631672 0.8048941
-0.1985292 -0.15193282 0.28414476 -0.5859193 -0.7813207 0.5037571
0.31629243 -0.45651788 -0.62191164 0.59148103 0.27635425 0.4091432
0.9689457 -1.2524866 -0.42302445 0.13820572 -0.5165117 -0.2702935
0.63049346 0.25958198 1.1737503 -0.7882072 0.92027354 1.467316
0.46158433 0.70693994 -1.4695194 -0.2665065 -0.05184259 0.29106686
-1.1161032 -0.1482241 1.1059649 -0.696478 0.907642 0.08685511
0.7858867 1.6652596 0.44393966 0.94541395 0.6943118 -0.3198849
-0.05015584 -0.6139155 -0.30027553 0.81025124 -0.34776166 0.10219242
-0.8762553 0.03773097 1.1567571 0.5039139 -0.14862603 -1.0390397
-1.6267425 0.93089604 0.49598715 0.05324828 -0.5425888 0.71388245
0.39809558 0.012539 0.46843863 0.1210272 -0.36853403 -0.4967786
-0.57308286 0.26051438 0.36249188 0.7321446 0.6123496 0.19659847
-0.31148335 0.5308781 0.61818427 0.28451756 0.7225559 -1.2627409
0.61599123 0.601383 0.02532376 -0.8741169 -0.22619422 -0.10075172
-0.76568663 0.26964375 0.27562186 0.3624839 -0.73732436 1.820766
0.48017845 0.3084316 -0.16594042 0.83430326 -0.02655882 0.5120078
0.09707115 0.04721827 1.2349298 -1.1564953 -0.7752172 0.04038567
0.59133905 -0.30445307 1.1536987 0.26280972 -1.0327069 -0.83001435
-1.0056334 -0.42831218 0.06324583 0.32500422 0.35447112 0.04692366
-0.6827191 0.88625556 -1.609968 -0.44180897 0.29715043 0.21323524
-0.74512887 0.2211304 -0.8690761 0.87201065 0.23334542 0.13635843
-1.136472 -0.45534256 -1.2749388 -0.20367306 0.28480417 -0.9857521
0.98490435 -0.41579494 -1.8810197 0.64142245 -0.14395122 -0.52948606
0.7178784 -0.81000483 -0.05421869 0.3597659 0.09422886 0.6480697
1.1451236 -0.9238214 -0.17936498 -0.5789596 -0.11231722 0.4126515
-0.41524085 -0.38342792 -0.2936037 -0.9853505 0.14041966 -1.1002862
-0.29360375 0.420283 -0.05922161 -0.12067139 0.9773225 -0.98003757
1.0568922 -2.0279148 1.2436743 -0.07936377 0.03033403 -0.02749499
-0.13841623 0.71456796 -0.1310433 -0.26168835 -0.39746502 -0.74834794
0.29738408 0.6769396 -0.24472757 0.70397127 0.34144 1.213592
-0.9113509 -0.23262995 0.5106501 0.93443346 -0.702082 0.42082947
-0.295098 1.0694026 -0.5307748 0.2922568 -0.095805 -0.17863432
0.03923507 -0.5178022 0.26452962 0.17619486 -0.99142706 2.4940765
-0.5306856 0.2268162 -0.12565514 -1.0671376 0.74668705 0.39096463
0.8498825 -0.20795557 0.03422776 0.03885613 -0.29958192 -0.48946172
0.26006663 -0.1392346 -0.15783516 0.46492916 0.45370713 0.11483968
-0.17774758 -0.01307013 1.0126773 1.0277034 0.30608946 0.18989801
0.59394807 -0.39997697 0.23175085 -0.05022234 -0.09262791 0.6388357
0.30481946 -0.33228043 -1.2186702 -1.326048 0.28573012 0.94635445
-0.23970778 -0.51785874 -0.71657 -0.7829393 -0.00509911 0.49757162
-0.6828233 -0.31584674 -0.96437263 -0.10125712 0.44939458 0.9657326
0.3478447 -1.2180245 -0.8036854 0.25214535 -0.223228 -1.1770488
-0.6762228 -0.13569099 -1.0612652 -0.9270276 -0.7608749 -0.5810922
0.51038986 -0.24565543 0.6727609 -0.42851663 -0.44662705 1.0641619
-0.4789588 0.47812173 -0.39459395 -0.23027521 0.27433234 0.35395393
0.13001499 -0.9409844 -0.71277845 0.33365363 -0.948433 0.21020104
0.5786923 1.1631083 0.70273125 -0.4960536 0.24369514 -0.36740825
0.28660274 -0.23604283 -0.0435058 0.10206523 -0.17015646 0.46379262
0.3095801 -0.3389483 -1.0005705 0.521463 -0.20436405 -0.9749432
-0.2502139 0.30786705 -0.28201413 0.21051118 0.56770724 0.46967563
0.5304022 -0.70241153 0.9493028 0.20187698 0.6523307 -0.7477072
1.0223423 0.67855066 0.25072172 -0.7299929 -0.46009555 -0.55551815
-1.3902818 -0.200757 1.389823 -0.8850132 -0.57452667 0.34530374
-1.211596 -0.3186241 -0.53038067 0.6336393 -1.4127998 0.7533268
-0.73007774 -0.64523363 -0.10695951 -1.1191809 1.6950362 -0.5290596
-0.5481547 -1.0030246 0.37543875 0.5403716 -0.23791523 0.87915516
0.6349857 -0.25033456 -0.5379821 -0.2866236 0.39657304 0.50277734
0.22583565 -0.32991028 -0.58237 -0.3979797 0.10217519 -0.43389887
0.7342941 0.19491097 1.0097408 -0.3840375 -0.19526327 0.58419156
1.0969297 -0.21845493 0.73391926 0.14747114 1.2388765 0.7752374
0.66700715 0.5191559 0.22226734 1.0791937 0.30724615 0.28455743
-0.2150593 -0.7931258 0.76328903 0.9933286 -0.46924984 0.32291538
-0.63799363 0.328037 -2.14748 -0.9488658 0.36941996 1.9784831
0.617341 -0.07018021 0.30916846 0.21896532 0.35719347 0.4748298
-0.46098882 -0.03341249 0.44344246 0.2899059 0.07509518 0.5351207
-0.97693664 0.9042141 5.0283504 0.77685404 -0.74370956 0.2286409
0.01034074 -0.04939581 -0.22113901 -0.10746686 -0.3952372 0.20997012
0.74938095 0.2977031 0.3100094 0.76903963 0.29962274 0.48406547
-1.5769503 1.0378449 0.11428314 -1.1451398 0.35263556 0.24955016
0.51889527 -0.44132066 0.08310171 0.24350335 0.03923685 -0.96417326
0.7093911 0.7440158 0.8761054 -0.63343275 0.15335745 0.5013864
-1.4466096 -0.13740285 0.00955264 -0.16470909 0.6471537 -0.13572161
-0.6039803 0.8391566 0.381697 1.2854419 -0.33484086 0.37228882
-0.43326792 0.27689224 -0.07182682 0.28791454 0.21406129 -0.40257898
0.6843347 0.8211374 0.25358495 -0.06299549 0.17299278 0.8553887
0.30322364 -0.12209062 -0.74538 -0.24561009 -0.08196229 1.0201434
-0.39355806 -0.03339517 -0.25340793 1.5737215 0.39631683 0.4232711
-0.8873601 0.2011141 0.78076303 0.16645049 0.3885048 -0.8314759
0.15723816 -1.2976402 0.25880444 -0.806742 0.19981915 -0.61534417
-1.0993602 0.18630002 0.31165504 -1.6948688 -0.8767637 -0.7016218
-0.48464575 0.62008345 -0.85592794 -1.6304442 -0.03723232 0.70158076
0.94495094 -0.13584365 0.9411492 0.16029106 -0.19659032 0.42845717
-0.12872127 0.13566993 0.48748296 -1.1902304 0.46333763 0.5780969
0.61704487 0.6771329 0.5319968 -0.68795377 -1.5342132 -1.1942905
0.43137905 -0.82374084 0.7300335 -0.45968372 -0.61899185 0.99654824
-0.09428845 0.08750203 0.54740536 -0.21916603 -0.25121894 0.20961069
-0.87657315 0.80952317 1.4421858 -0.78955764 -0.77395725 0.26576233
0.8655808 -0.3321495 -1.0886956 0.40577534 0.84637934 -0.63194823
1.2322559 -0.8995558 0.687102 -0.29501197 -0.3301597 -1.1529198
-0.35304207 -0.6925734 -0.70894974 0.8532464 -0.18271571 -0.1794887
0.9876611 0.18387191 -0.20105912 -1.0270876 -1.1216242 -0.77904254
-0.06418432 -0.41974276 0.19094555 0.5453487 -0.03972917 0.31128982
-0.8851486 -0.0963129 0.7564502 -0.22951038 1.0825984 -0.8215315
-0.71814203 -0.15869129 -0.959888 -1.4949859 0.44877303 -0.7891936
0.03315976 -1.4492481 -0.04749325 0.48526293 -0.1346483 0.30543125
0.13611226 0.15450479 0.27675208 0.35031947 -0.44476852 1.3703209
1.4785032 -0.02632968 0.02467113 -0.1365133 0.17560084 0.76709527
0.25933388 -0.15460297 -0.6541732 -0.27527788 -0.16128853 0.29253438
0.75579613 -1.0834913 -0.02562368 -0.02823392 0.35246497 -0.6632199
0.85289407 -0.9881021 0.38616017 0.6246264 -0.51426595 -0.08052181
-0.20528877 1.0594026 -0.28545496 0.26888543 0.41975245 -0.06800272
-0.6831743 0.4780204 -0.10488159 -0.26554692 1.0634873 -0.5019006
0.3367406 -0.29915634 -1.245785 0.01120981 0.4270429 0.5866033
0.79840153 -1.9277056 -0.5113105 0.31623703 0.17968288 -0.07342396
0.38398013 0.97833097 -0.46523526 0.29092342 -0.5892412 -0.85430187
-1.0003432 0.48122892 0.12938373 -0.41652805 -0.9877149 0.539012
0.2903641 -0.6523142 0.24479637 -0.29045117 -0.16213343 -0.21035536
0.19295132 0.6050462 -0.40519556 -1.120359 -0.13902012 0.8811585
0.3259285 -0.47586256 1.396058 -0.05716763 0.07664157 0.66178805
1.6314276 -0.3240053 -1.7763433 -0.13504446 -0.09466708 -0.4455857
-0.5472631 -0.15683024 -0.7288939 1.0638847 0.5720065 -0.44999352
0.796611 0.08543327 0.9506223 0.27519643 0.50896937 -0.98007756
0.7833198 0.3303015 1.4123212 -1.0025529 -0.07181937 -0.2693832
-0.7694557 1.1004974 0.42289722 -0.49730125 0.6094705 -0.31773207
-0.23296353 -0.26448998 -0.5026653 0.01989667 0.755646 0.6408642
0.36940724 0.12156796 -0.42003775 0.45323327 -0.1269848 -0.08412088
0.02565555 0.9454932 -0.17269672 -1.3275083 -0.20634231 0.06019421
-0.08287177 0.47662148 -0.19694734 0.7496973 -0.09637178 0.33970657
-0.07669815 -0.63057345 0.36505285 0.37614518 0.8511597 -0.80123216
-0.07481729 0.3499182 -0.04745498 -1.1123927 -0.5924179 -0.8137552
-1.2561663 0.34672686 0.21076566 -0.21383613 0.36206403 1.0101695
0.20820785 0.6134288 0.36666244 -1.4992291 -1.0218825 -1.0201849
-0.63879144 0.92472714 0.37732428 -1.135613 -0.24866046 0.42269096] | [7.330533504486084, -0.3137703537940979] |
d0a79a4d-0e9a-4437-b3ff-10318e50b71c | icd-coding-from-clinical-text-using-multi | 1912.00862 | null | https://arxiv.org/abs/1912.00862v1 | https://arxiv.org/pdf/1912.00862v1.pdf | ICD Coding from Clinical Text Using Multi-Filter Residual Convolutional Neural Network | Automated ICD coding, which assigns the International Classification of Disease codes to patient visits, has attracted much research attention since it can save time and labor for billing. The previous state-of-the-art model utilized one convolutional layer to build document representations for predicting ICD codes. However, the lengths and grammar of text fragments, which are closely related to ICD coding, vary a lot in different documents. Therefore, a flat and fixed-length convolutional architecture may not be capable of learning good document representations. In this paper, we proposed a Multi-Filter Residual Convolutional Neural Network (MultiResCNN) for ICD coding. The innovations of our model are two-folds: it utilizes a multi-filter convolutional layer to capture various text patterns with different lengths and a residual convolutional layer to enlarge the receptive field. We evaluated the effectiveness of our model on the widely-used MIMIC dataset. On the full code set of MIMIC-III, our model outperformed the state-of-the-art model in 4 out of 6 evaluation metrics. On the top-50 code set of MIMIC-III and the full code set of MIMIC-II, our model outperformed all the existing and state-of-the-art models in all evaluation metrics. The code is available at https://github.com/foxlf823/Multi-Filter-Residual-Convolutional-Neural-Network. | ['Hong Yu', 'Fei Li'] | 2019-11-25 | null | null | null | null | ['medical-code-prediction'] | ['medical'] | [-1.36445239e-01 -1.64307907e-01 -2.38943666e-01 -5.65947652e-01
-5.46364605e-01 -3.66321295e-01 5.63377738e-02 3.32842588e-01
-2.51044780e-01 2.81038314e-01 4.40896034e-01 -6.44761622e-01
-8.37329328e-02 -8.87531221e-01 -5.76330900e-01 -4.70061693e-03
2.19027661e-02 3.04173321e-01 6.58749640e-02 -1.20671406e-01
-5.27541190e-02 2.85100669e-01 -1.21026528e+00 1.08165407e+00
8.78616869e-01 1.26484239e+00 8.25825483e-02 5.30873001e-01
-4.17831540e-01 1.16666889e+00 -2.31320173e-01 -5.10069132e-01
1.15064569e-01 -3.41219664e-01 -9.17219162e-01 -1.54321060e-01
1.32873461e-01 -5.92486441e-01 -8.35020006e-01 8.11731339e-01
3.94081801e-01 -3.83352876e-01 6.28604352e-01 -7.17072904e-01
-1.09056687e+00 6.98559105e-01 -3.51924747e-01 1.04093760e-01
2.78149992e-01 -3.91223371e-01 8.81915569e-01 -6.15716636e-01
2.74300039e-01 1.18520081e+00 1.04309714e+00 4.38620389e-01
-9.10017014e-01 -9.30277526e-01 -6.86201006e-02 -1.13221206e-01
-1.51536858e+00 -6.24341555e-02 4.31351900e-01 -6.90643311e-01
1.10450232e+00 9.59489569e-02 3.68569553e-01 8.38001549e-01
5.26715875e-01 7.54776835e-01 5.08608341e-01 -4.37830269e-01
-5.96479550e-02 -6.40210882e-02 2.34915614e-01 1.06818855e+00
4.24749464e-01 -1.64183546e-02 2.77599603e-01 -4.67352003e-01
7.47339725e-01 8.91195953e-01 -1.34834021e-01 -4.10762914e-02
-8.72851253e-01 1.01343310e+00 6.03302062e-01 5.98863244e-01
-3.57014358e-01 2.25576073e-01 6.70210361e-01 3.00103277e-02
3.70197535e-01 1.25642836e-01 -5.77055097e-01 -1.58934802e-01
-9.44673657e-01 1.83757082e-01 7.99903035e-01 1.05821526e+00
2.90403426e-01 -2.63384014e-01 -5.69928885e-01 1.13184953e+00
2.90342957e-01 2.79391319e-01 7.32588530e-01 -5.55208862e-01
8.45655084e-01 1.10284710e+00 -1.86012298e-01 -1.00371480e+00
-7.82011151e-01 -5.29232383e-01 -1.31496823e+00 -3.16574216e-01
3.77024449e-02 -2.37920627e-01 -1.22468030e+00 1.22888291e+00
-3.46371382e-01 -1.21756300e-01 -1.27383294e-02 5.46407342e-01
1.16999984e+00 6.59103692e-01 -7.66872764e-02 7.64337033e-02
1.42457855e+00 -1.00234723e+00 -4.53516990e-01 5.61897121e-02
9.61789846e-01 -6.61887586e-01 7.00156033e-01 2.75120586e-01
-9.40423131e-01 -8.11900795e-01 -7.91847467e-01 -1.09640025e-01
-2.95004457e-01 6.29799843e-01 6.52114093e-01 5.85538566e-01
-8.36172223e-01 4.71897513e-01 -7.31262207e-01 -3.94982636e-01
6.79113686e-01 3.06644648e-01 -2.33599976e-01 -5.59178293e-01
-1.23578525e+00 2.08918095e-01 4.78806406e-01 -3.15072298e-01
-5.22503793e-01 -5.09080350e-01 -7.07783937e-01 5.80733001e-01
-1.08074978e-01 -6.42461479e-01 1.43201220e+00 -8.36297929e-01
-9.43965554e-01 7.72557259e-01 9.57948267e-02 -4.71549243e-01
3.97416592e-01 -1.00843392e-01 -7.36064196e-01 2.64987648e-01
6.72761276e-02 5.96701980e-01 4.13539767e-01 -8.77638042e-01
-5.42028427e-01 -1.58242732e-01 2.61643142e-01 -1.91913635e-01
-6.05535984e-01 1.85561720e-02 -7.74292111e-01 -1.08876836e+00
3.13889310e-02 -1.03664005e+00 -1.91085070e-01 -2.00285786e-03
-1.75022408e-01 -2.70117968e-01 3.28344524e-01 -4.31801826e-01
1.89308321e+00 -2.45590448e+00 -5.06580591e-01 1.91228483e-02
1.71196654e-01 3.89690012e-01 -9.20377150e-02 7.20386565e-01
-3.95959318e-01 4.08434182e-01 -1.71914384e-01 -2.24869147e-01
-2.26344258e-01 2.12958813e-01 -3.39385480e-01 2.75442570e-01
-2.06065327e-02 7.57180810e-01 -6.87654078e-01 -4.89355862e-01
6.52604401e-02 5.07008493e-01 -9.35876310e-01 2.57213354e-01
5.67213185e-02 -4.48712632e-02 -6.03421330e-01 6.02586269e-01
9.19215560e-01 -7.59946167e-01 1.30250469e-01 8.72080103e-02
1.33776128e-01 2.45529652e-01 -9.20570910e-01 1.82039154e+00
-4.52320427e-01 2.74345934e-01 -4.96113956e-01 -9.45142925e-01
8.86345565e-01 4.55825925e-01 8.06993842e-01 -6.19038284e-01
2.42316961e-01 3.81778389e-01 9.89281461e-02 -5.00156879e-01
1.85353622e-01 2.09878296e-01 -2.13444382e-01 3.00592870e-01
-1.17602013e-01 2.75588006e-01 1.82505921e-01 1.36049062e-01
1.50946343e+00 -5.19314647e-01 2.94342965e-01 -1.78744137e-01
4.70314890e-01 -3.35076332e-01 5.68161845e-01 8.20130885e-01
1.32783532e-01 9.92663443e-01 4.51352715e-01 -9.95294929e-01
-7.09800422e-01 -7.59080589e-01 -5.67396998e-01 7.83309340e-01
-2.72707164e-01 -8.06778431e-01 -6.13479614e-01 -8.87351334e-01
3.65316749e-01 3.32488149e-01 -7.06223488e-01 -6.95124418e-02
-3.49122971e-01 -5.39388955e-01 9.65924561e-01 8.82423997e-01
6.46101534e-01 -1.23480952e+00 -6.18574917e-01 3.68302405e-01
-1.80506319e-01 -8.58972013e-01 -6.51366115e-01 2.45120928e-01
-8.63299727e-01 -1.34366333e+00 -7.50700355e-01 -9.91392553e-01
8.02392125e-01 1.14338689e-01 1.14251769e+00 5.15616655e-01
-3.36117208e-01 3.86259034e-02 -7.90273249e-01 -4.86386746e-01
-3.40873182e-01 3.88714701e-01 -3.23019832e-01 -7.20557198e-02
4.78363574e-01 -1.76560953e-01 -7.09319532e-01 7.47072473e-02
-1.25662017e+00 1.36120394e-01 6.22220516e-01 8.07802141e-01
3.54198843e-01 1.31523609e-01 3.66935968e-01 -1.15588808e+00
8.34856033e-01 -6.54758930e-01 -3.79538804e-01 1.34615645e-01
-6.86930895e-01 5.87069467e-02 1.06896377e+00 -7.76177570e-02
-6.70755208e-01 3.15992311e-02 -5.51295340e-01 -4.44372773e-01
-1.62182584e-01 7.43448794e-01 3.08311820e-01 3.47851753e-01
6.31334007e-01 4.49019559e-02 -5.00823975e-01 -6.29799843e-01
-1.96155831e-01 1.25328517e+00 1.39175802e-01 -3.27419341e-01
1.71776146e-01 4.05555308e-01 -2.53779620e-01 -2.19935507e-01
-7.81855822e-01 -6.21073127e-01 -3.36840838e-01 4.14104939e-01
8.27571273e-01 -1.04619586e+00 -6.05432034e-01 5.09480000e-01
-1.36719263e+00 -7.67920166e-02 -9.47441813e-03 4.91573215e-01
-2.27678239e-01 2.91830957e-01 -9.10084665e-01 -3.48754615e-01
-4.97986048e-01 -1.28755784e+00 8.84263039e-01 -9.15912315e-02
-2.37073556e-01 -7.54124403e-01 2.14296598e-02 1.31629825e-01
4.76708740e-01 1.33605078e-01 1.18617570e+00 -6.56336308e-01
3.31106670e-02 -4.11446542e-01 -5.82483888e-01 4.56323534e-01
3.93992186e-01 -2.09624544e-01 -8.29540730e-01 -3.42652977e-01
-3.44994336e-01 -3.06854188e-01 1.27474213e+00 5.23550570e-01
2.26773548e+00 -3.12055618e-01 -5.00866830e-01 9.22793508e-01
1.51903224e+00 5.82465887e-01 7.81258464e-01 1.90442115e-01
5.55831492e-01 -5.21533601e-02 6.65827543e-02 7.83538222e-01
4.93513525e-01 5.54067373e-01 4.72403407e-01 -3.28642160e-01
1.22673221e-01 -1.55652121e-01 -8.78006741e-02 7.64699936e-01
5.93535714e-02 -4.53369349e-01 -1.29293847e+00 5.49297035e-01
-1.95782125e+00 -8.08426738e-01 -1.57012060e-01 1.75256717e+00
7.30622292e-01 2.27559075e-01 -1.13808349e-01 1.46886215e-01
5.30520737e-01 2.80509405e-02 -4.82195884e-01 -6.67527795e-01
3.66439819e-01 3.06928337e-01 5.28196335e-01 -3.74938659e-02
-1.38790631e+00 4.04269964e-01 6.18893194e+00 6.20794237e-01
-8.71958435e-01 5.79063594e-02 7.96081781e-01 4.48674299e-02
-1.36975333e-01 -4.70768780e-01 -8.13855946e-01 9.37599957e-01
1.12252796e+00 2.40759358e-01 2.87962168e-01 1.06686485e+00
-9.35530141e-02 4.05335158e-01 -9.04285789e-01 1.24427903e+00
8.17307383e-02 -1.59800494e+00 2.34542385e-01 1.55638143e-01
8.76082063e-01 3.68481755e-01 -1.10140786e-01 5.46461105e-01
2.86481708e-01 -1.08107924e+00 5.11740685e-01 4.47009414e-01
1.31691003e+00 -8.04974020e-01 1.17286205e+00 2.51650572e-01
-1.50516582e+00 -6.46255732e-01 -5.53223193e-01 6.72289683e-03
-2.52320409e-01 6.49296582e-01 -3.09662342e-01 4.77568984e-01
1.01868725e+00 9.14371312e-01 -6.20351791e-01 1.11028481e+00
2.20532775e-01 7.13348150e-01 5.66876261e-03 2.06331491e-01
4.00661528e-01 3.16353112e-01 -3.57408375e-01 1.72415948e+00
5.14253318e-01 2.74304986e-01 4.04822409e-01 5.64244211e-01
-6.48052275e-01 2.45819598e-01 -4.35542583e-01 -1.23576470e-01
2.72624224e-01 9.13496733e-01 -6.20211005e-01 -5.43503225e-01
-8.69925320e-01 9.06476498e-01 3.59043777e-01 1.47151515e-01
-9.89767790e-01 -7.26284087e-01 4.13176805e-01 2.97046185e-01
6.27762258e-01 2.01704845e-01 -1.87879205e-01 -1.26446021e+00
3.57425064e-02 -9.08868194e-01 6.94061518e-01 -5.86793900e-01
-1.31222975e+00 9.74461496e-01 -1.26758352e-01 -1.50450873e+00
-1.41657978e-01 -7.14556217e-01 -2.48012215e-01 5.87647736e-01
-1.37253559e+00 -1.00747085e+00 -5.37837148e-01 7.27809846e-01
4.88909602e-01 -4.16888565e-01 1.21986651e+00 7.68254995e-01
-2.44442046e-01 9.05641496e-01 6.01169765e-01 7.93484569e-01
5.40836632e-01 -8.23335290e-01 4.95137930e-01 4.10850137e-01
-2.77755801e-02 7.56560862e-01 -1.95803434e-01 -4.86673504e-01
-9.08264041e-01 -1.47642922e+00 9.41376805e-01 -2.83532012e-02
2.70100892e-01 -1.97329521e-01 -8.60589981e-01 6.76868081e-01
8.45129043e-02 3.91001195e-01 8.57281744e-01 3.46368626e-02
-4.39750612e-01 -3.24576467e-01 -1.05367112e+00 2.59781480e-02
1.06150925e+00 -4.00074869e-01 -4.33045894e-01 2.81677783e-01
7.81533360e-01 -5.78231573e-01 -8.66029978e-01 3.36218357e-01
8.18372905e-01 -9.08637583e-01 7.58070469e-01 -5.50835907e-01
9.19634521e-01 1.83963478e-02 -2.91451514e-01 -1.02615142e+00
-8.58891904e-01 1.91944931e-02 -9.55619514e-02 9.41419482e-01
2.81162828e-01 -6.25798702e-01 4.97627050e-01 3.34957242e-01
-3.51018339e-01 -1.24263000e+00 -6.04861319e-01 -6.70555174e-01
2.16913477e-01 -5.29698789e-01 9.26262677e-01 8.56692970e-01
2.64957398e-02 1.76371932e-02 -2.38820195e-01 -2.82491118e-01
-7.70861432e-02 1.39605001e-01 2.59002477e-01 -1.48013937e+00
-3.82129252e-01 -4.01797563e-01 -3.63153696e-01 -9.85140383e-01
-6.69013783e-02 -1.16777706e+00 -3.52669626e-01 -1.90683115e+00
4.29523349e-01 -4.59932417e-01 -6.84043825e-01 9.29785609e-01
1.02209695e-01 3.29519138e-02 3.00193042e-01 3.61876309e-01
-4.94704813e-01 9.22157168e-02 9.79139745e-01 -3.74315083e-01
-1.39631942e-01 8.35832953e-02 -9.33104396e-01 7.00107515e-01
9.56023097e-01 -7.74283171e-01 -4.20879692e-01 -7.80554354e-01
3.67974609e-01 4.96652424e-02 -4.85415192e-04 -1.25822186e+00
9.10779536e-02 5.29575273e-02 6.98283851e-01 -7.47801602e-01
-1.91375047e-01 -1.15711904e+00 1.35903895e-01 8.51918101e-01
-5.87079525e-01 2.92190969e-01 4.20450360e-01 3.55469555e-01
-2.68570632e-01 -4.37779307e-01 6.35759652e-01 -3.87263834e-01
-1.97574064e-01 5.81552982e-01 -4.09904152e-01 2.12322176e-02
7.07776785e-01 6.29719123e-02 -1.86474174e-01 -1.66226998e-01
-2.16932520e-01 5.71627133e-02 2.59216160e-01 5.76912045e-01
7.94916213e-01 -1.45269942e+00 -7.34882474e-01 5.52241147e-01
4.74416971e-01 1.55289639e-02 2.28847608e-01 3.92397851e-01
-9.87363160e-01 9.37656105e-01 -6.08885065e-02 -4.36926395e-01
-1.06220686e+00 6.40838504e-01 4.00119007e-01 -7.93675363e-01
-6.77174449e-01 6.51068807e-01 3.44615340e-01 -4.73009676e-01
3.43233019e-01 -1.04972434e+00 -2.06887484e-01 -2.71475792e-01
7.05789924e-01 9.18940678e-02 3.06741506e-01 -3.26215625e-01
-5.09892523e-01 4.62886870e-01 -4.39788669e-01 5.95242023e-01
1.43141997e+00 2.05096379e-01 -5.12628295e-02 -2.76242271e-02
1.53997886e+00 -2.83563912e-01 -6.71233594e-01 -7.60745630e-02
-2.92866349e-01 -3.46442103e-01 8.99345949e-02 -8.71623099e-01
-1.28748596e+00 8.67798150e-01 7.71083236e-01 3.28877628e-01
1.38859725e+00 -1.51216477e-01 1.04062581e+00 3.96185160e-01
2.41548434e-01 -8.26665699e-01 -3.16942707e-02 7.39333570e-01
8.54580045e-01 -1.33836830e+00 -1.72775939e-01 -3.46735381e-02
-5.35049677e-01 1.33679593e+00 4.81972277e-01 -1.97213203e-01
1.10524023e+00 2.29318917e-01 3.74163054e-02 -3.01074296e-01
-6.40507579e-01 9.47299376e-02 5.20839334e-01 1.14430711e-01
9.84856308e-01 2.00946167e-01 -3.86669815e-01 9.71382260e-01
-3.05615235e-02 5.85623860e-01 3.57041955e-01 1.00887787e+00
-1.51391894e-01 -1.08942330e+00 -2.49347046e-01 9.88980174e-01
-8.68733287e-01 -3.48992348e-01 -1.42735392e-01 5.81294537e-01
4.96554524e-01 1.01931274e+00 1.94198444e-01 -5.43680727e-01
4.48605597e-01 -4.76989113e-02 -4.01077457e-02 -8.48607838e-01
-9.18205142e-01 -1.01554409e-01 -2.13495433e-01 -4.99153465e-01
-1.67129859e-01 -4.11352217e-01 -1.49988866e+00 -4.59995866e-01
-4.45021987e-02 -1.47203561e-02 2.27029458e-01 5.89457035e-01
7.61898875e-01 8.90941799e-01 4.54567611e-01 -2.79506832e-01
-4.31217164e-01 -1.10873675e+00 -7.38210917e-01 5.33758700e-01
4.98720795e-01 -4.50295299e-01 1.17504187e-01 6.77917749e-02] | [8.000649452209473, 6.825688362121582] |
485b044a-3360-42d0-b3c4-25fd033cbead | edassistant-supporting-exploratory-data | 2112.07858 | null | https://arxiv.org/abs/2112.07858v1 | https://arxiv.org/pdf/2112.07858v1.pdf | EDAssistant: Supporting Exploratory Data Analysis in Computational Notebooks with In-Situ Code Search and Recommendation | Using computational notebooks (e.g., Jupyter Notebook), data scientists rationalize their exploratory data analysis (EDA) based on their prior experience and external knowledge such as online examples. For novices or data scientists who lack specific knowledge about the dataset or problem to investigate, effectively obtaining and understanding the external information is critical to carry out EDA. This paper presents EDAssistant, a JupyterLab extension that supports EDA with in-situ search of example notebooks and recommendation of useful APIs, powered by novel interactive visualization of search results. The code search and recommendation are enabled by state-of-the-art machine learning models, trained on a large corpus of EDA notebooks collected online. A user study is conducted to investigate both EDAssistant and data scientists' current practice (i.e., using external search engines). The results demonstrate the effectiveness and usefulness of EDAssistant, and participants appreciated its smooth and in-context support of EDA. We also report several design implications regarding code recommendation tools. | ['Jian Zhao', 'Chengnian Sun', 'Justin Leung', 'Yizhi Zhang', 'Xingjun Li'] | 2021-12-15 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-9.40636098e-01 -1.25946984e-01 -2.75314480e-01 -2.78449237e-01
-3.92609358e-01 -9.50157940e-01 4.23624665e-01 7.15053439e-01
-2.28610784e-01 -1.19731344e-01 2.69999295e-01 -9.27374363e-01
-4.79477912e-01 -5.45990646e-01 -3.64191890e-01 1.32798985e-01
1.22754768e-01 2.19409838e-01 -2.59793788e-01 7.91540295e-02
8.57973099e-01 5.27194142e-01 -1.70029342e+00 6.33203089e-01
1.09981167e+00 5.73899329e-01 5.37074208e-01 7.11560130e-01
-6.01073861e-01 9.62866724e-01 -6.98153794e-01 -9.80637968e-02
1.50318429e-01 3.04033667e-01 -7.34273970e-01 -5.52712619e-01
5.34039974e-01 -4.17200595e-01 2.59310454e-01 6.89236045e-01
2.55807757e-01 2.67439991e-01 -8.48230068e-03 -1.45946574e+00
-9.23467636e-01 6.21122360e-01 -1.37945384e-01 2.57765502e-01
7.10157633e-01 3.63329649e-01 1.02701259e+00 -1.13074839e+00
7.18027890e-01 6.48429453e-01 7.42669582e-01 1.45678774e-01
-1.36823130e+00 -9.33098257e-01 -6.32464066e-02 4.32086214e-02
-1.11076677e+00 -2.77338535e-01 6.15105212e-01 -1.11394739e+00
1.44246280e+00 6.68076992e-01 1.00572181e+00 6.57742798e-01
-2.19592795e-01 5.70333362e-01 7.65625119e-01 -6.89820111e-01
4.73210454e-01 9.92794454e-01 8.40672255e-01 6.43903255e-01
3.55041951e-01 -1.92547202e-01 -8.16949368e-01 -6.69178128e-01
4.88285422e-01 2.94770867e-01 -8.82800072e-02 -3.60347033e-01
-1.14242876e+00 5.23893774e-01 1.88642502e-01 3.50596964e-01
-4.16720152e-01 -2.59283930e-01 3.29251677e-01 4.70582843e-01
2.10011259e-01 1.22726119e+00 -7.41114378e-01 -8.87103796e-01
-9.05015707e-01 2.82359660e-01 1.19765627e+00 1.38778079e+00
7.00387716e-01 -4.90867615e-01 4.63134497e-02 8.66907775e-01
5.60296297e-01 4.24158387e-02 7.41833448e-01 -7.58638382e-01
4.49580550e-01 1.20654786e+00 2.57226348e-01 -8.69596243e-01
-2.59962291e-01 -1.75824717e-01 1.90266714e-01 6.75753295e-01
6.36414826e-01 -1.21214546e-01 -2.11591616e-01 8.21116686e-01
1.88307390e-01 -6.09354436e-01 -2.69003302e-01 5.61416090e-01
9.53175128e-01 2.37457693e-01 4.24254566e-01 2.08620951e-01
1.13555133e+00 -9.11319494e-01 -5.53300023e-01 -1.25605971e-01
1.04138970e+00 -9.79000092e-01 1.99068463e+00 7.15406537e-01
-1.02150345e+00 -6.63648963e-01 -8.70793402e-01 -4.55999434e-01
-7.56101847e-01 4.02171373e-01 7.87772775e-01 6.55235708e-01
-7.27837265e-01 7.63233900e-01 -1.00692427e+00 -7.48594940e-01
4.20427948e-01 -1.01863988e-01 -6.97885780e-03 2.60622799e-01
-2.91099817e-01 4.05776739e-01 9.94601101e-02 -4.83550191e-01
-4.35481191e-01 -1.73259401e+00 -3.99151474e-01 3.72005075e-01
1.72005579e-01 -3.68340671e-01 1.65454423e+00 -3.70894015e-01
-8.10599267e-01 6.85291529e-01 1.53663665e-01 2.38820165e-01
4.82039779e-01 -8.46753180e-01 -3.96616787e-01 -2.83615798e-01
2.16003075e-01 -9.37789381e-02 1.83361471e-01 -9.12530541e-01
-5.35313010e-01 -4.20839459e-01 2.33229231e-02 -1.98596179e-01
-9.73125458e-01 3.63556534e-01 -5.02861261e-01 -5.40041685e-01
-2.92739093e-01 -5.40022552e-01 7.00907633e-02 2.67928660e-01
-3.79109412e-01 -3.07988882e-01 9.13904905e-01 -7.88211584e-01
2.11313772e+00 -2.15748334e+00 -6.00973129e-01 5.77060759e-01
5.20694554e-01 9.79830101e-02 3.17855418e-01 9.58861649e-01
-1.45381540e-01 5.52237213e-01 6.21120572e-01 6.78225979e-02
2.11660668e-01 -5.58515012e-01 -2.02661067e-01 -5.70381321e-02
-5.14644742e-01 2.94673115e-01 -1.01927471e+00 -5.40808320e-01
1.88477159e-01 3.81556302e-01 -8.09217572e-01 5.49678743e-01
-3.69592071e-01 1.42407849e-01 -7.18474805e-01 8.76203895e-01
3.39018375e-01 -1.11447155e+00 1.57692224e-01 -1.20250970e-01
-9.22765076e-01 5.84296942e-01 -9.83432472e-01 1.78186131e+00
-9.16740835e-01 1.02282226e+00 5.11934273e-02 -7.90725052e-02
8.89384091e-01 -1.64412651e-02 3.30653936e-01 -4.24012870e-01
-4.19910222e-01 1.73842743e-01 -1.68695658e-01 -1.15156090e+00
4.59947944e-01 8.81110609e-01 4.68206763e-01 1.15801251e+00
-2.75780618e-01 3.19624484e-01 2.24894971e-01 6.46935523e-01
1.25965178e+00 4.10911500e-01 4.37191218e-01 -5.29989779e-01
-1.79692090e-01 4.95865434e-01 -5.45716286e-02 9.62409735e-01
3.39281768e-01 6.17660675e-03 2.55148619e-01 -5.04634857e-01
-9.65093493e-01 -4.74387676e-01 -2.79238939e-01 1.54676223e+00
-4.13236201e-01 -1.22470462e+00 -5.94003737e-01 -6.27975345e-01
3.37756395e-01 1.25580907e+00 -3.55455041e-01 3.44678819e-01
-1.71658173e-01 7.90460184e-02 1.05742030e-01 5.97939193e-01
4.83808547e-01 -1.01984298e+00 -1.29679632e+00 1.15564719e-01
9.59326327e-02 -4.07195270e-01 -5.68523824e-01 -6.14987575e-02
-9.78935361e-01 -1.35656810e+00 -9.05429646e-02 -2.45293066e-01
6.58504009e-01 4.94943172e-01 1.48320627e+00 6.29320025e-01
-7.34492004e-01 8.67291272e-01 -5.15729725e-01 -5.63923657e-01
-2.68237859e-01 1.60882752e-02 -4.16379690e-01 -9.76187229e-01
9.92936969e-01 -7.45499432e-01 -5.33369362e-01 2.83217847e-01
-5.63129365e-01 4.09139216e-01 3.70399684e-01 2.22334296e-01
1.01802647e-01 -1.83442026e-01 6.04281202e-02 -1.27573705e+00
1.21744335e+00 -8.33942175e-01 -9.22090054e-01 5.71087897e-01
-1.45663595e+00 -6.05808906e-02 4.89090413e-01 -4.61109668e-01
-1.16076159e+00 -1.91991225e-01 3.00595403e-01 -4.75617290e-01
-2.57587224e-01 9.93550003e-01 1.08791269e-01 -2.27185618e-02
1.15957093e+00 -2.62548983e-01 -2.07785785e-01 -1.26562977e+00
1.94983765e-01 9.21067953e-01 9.51197278e-03 -8.36259604e-01
4.21217769e-01 -2.45473776e-02 -7.43634284e-01 -8.53948295e-01
-1.60136804e-01 -6.28541648e-01 -5.16350746e-01 -3.09544683e-01
3.54303151e-01 -6.14032328e-01 -1.26145875e+00 -4.48424816e-02
-8.59216750e-01 -7.07573354e-01 -1.10564165e-01 6.41396582e-01
-1.08491629e-01 -8.14180523e-02 -1.36022344e-01 -8.39062870e-01
-4.30392027e-01 -8.24380040e-01 4.83367145e-01 3.37344736e-01
-9.88250315e-01 -1.07712042e+00 1.70750692e-01 4.98017639e-01
6.93993032e-01 1.19819511e-02 1.18841255e+00 -1.09329295e+00
-8.22814584e-01 -3.43737215e-01 -2.94851989e-01 -3.08773547e-01
2.62877405e-01 6.70287549e-01 -1.20282769e+00 -1.30565822e-01
-4.01350588e-01 -1.33838326e-01 -1.52395278e-01 -8.94112512e-02
1.63487482e+00 -6.62641585e-01 -6.47129297e-01 5.03037691e-01
1.30104661e+00 3.63269031e-01 -1.32166408e-02 8.95799875e-01
5.58967054e-01 5.32636583e-01 7.82077074e-01 1.03834569e+00
1.66276678e-01 3.12408656e-01 1.95747077e-01 1.78839788e-01
1.11534752e-01 -4.79287595e-01 -1.65885329e-01 6.08444571e-01
-2.78935153e-02 1.63346946e-01 -1.49668515e+00 6.45425022e-01
-1.71119547e+00 -7.14577258e-01 -1.42970607e-01 2.36753774e+00
7.63985574e-01 -5.14988489e-02 2.84827203e-01 -3.11077535e-01
2.61531740e-01 -3.87023211e-01 -8.33284259e-01 -2.50265151e-01
7.50168383e-01 -6.91001303e-03 1.38055056e-01 1.41268149e-01
-3.69252741e-01 1.89692199e-01 5.90272713e+00 3.79828662e-01
-8.76004934e-01 -1.20535269e-01 1.65198386e-01 -5.61227202e-01
-4.93017614e-01 1.56490639e-01 -7.15890765e-01 6.18749142e-01
8.51444721e-01 -5.76693594e-01 8.28577936e-01 1.85605478e+00
4.08490002e-01 -5.33485450e-02 -1.66870713e+00 1.03409517e+00
-4.88829225e-01 -1.91880870e+00 -5.13768017e-01 2.31691673e-01
2.61002868e-01 9.09666494e-02 -2.62721278e-05 4.93723899e-01
7.54849851e-01 -7.14141846e-01 5.94068646e-01 7.05206811e-01
5.48290133e-01 -3.90302718e-01 -5.75773269e-02 3.78199250e-01
-8.39257360e-01 -2.92451173e-01 3.17744650e-02 -2.76517779e-01
-4.98698175e-01 4.49673206e-01 -1.13006198e+00 -1.10031813e-01
1.40209174e+00 7.49550521e-01 -9.23246264e-01 1.15393376e+00
2.40039632e-01 6.45601511e-01 -2.91746199e-01 -4.03878391e-01
-4.29947287e-01 -1.99797630e-01 3.34504008e-01 1.42628932e+00
5.64293027e-01 1.48704320e-01 -4.52911928e-02 1.42820537e+00
-8.62882938e-04 2.81666011e-01 -4.42549556e-01 -6.23029172e-01
1.02529252e+00 1.50764155e+00 -4.15870398e-01 -3.43170017e-01
-6.82280719e-01 2.15546101e-01 4.30266023e-01 4.67361242e-01
-1.94381252e-01 -8.11915100e-01 7.32537031e-01 7.42340922e-01
-6.36355281e-02 -2.56180286e-01 -6.56820238e-01 -8.61344814e-01
2.34509245e-01 -1.15537715e+00 4.64472651e-01 -1.24898517e+00
-1.22400880e+00 2.79536694e-01 1.64619312e-01 -1.23058879e+00
-2.01593280e-01 -4.54595000e-01 -9.59663808e-01 1.02508724e+00
-7.32382596e-01 -9.14193332e-01 -9.10077274e-01 4.18688983e-01
4.19878781e-01 -2.76667893e-01 7.24763989e-01 4.64731216e-01
-5.99688113e-01 3.93424630e-01 3.50494117e-01 3.58274318e-02
9.18346941e-01 -1.38972342e+00 4.07251924e-01 2.33988166e-01
6.75872788e-02 1.79545593e+00 7.86419630e-01 -1.02849901e+00
-1.76533568e+00 -6.08609915e-01 5.62034309e-01 -7.62062609e-01
9.18307900e-01 -3.51596683e-01 -1.08297348e+00 8.82792890e-01
2.95722157e-01 -1.10547870e-01 1.38227022e+00 9.41809893e-01
-4.90523875e-01 9.38194245e-02 -1.09864295e+00 5.15743256e-01
9.25357103e-01 -7.35011935e-01 -3.88174206e-01 4.65965867e-01
4.41656053e-01 -2.30244279e-01 -1.13837254e+00 -3.75606328e-01
9.52515423e-01 -9.34058368e-01 8.28297436e-01 -9.89818037e-01
6.29999101e-01 1.92929376e-02 1.34801075e-01 -1.10213709e+00
-8.91527757e-02 -8.01562130e-01 -2.90259361e-01 1.36607051e+00
4.96019721e-01 -2.69067228e-01 7.13948309e-01 1.81957102e+00
-2.04513803e-01 -3.63690764e-01 5.82206734e-02 -6.27854943e-01
-3.47055584e-01 -6.52128100e-01 7.89289117e-01 1.39418256e+00
8.89761329e-01 -1.93102717e-01 1.86363220e-01 5.91073884e-04
4.59566385e-01 3.18582505e-01 1.24899530e+00 -1.54647303e+00
-6.55375004e-01 -4.23324585e-01 3.32945853e-01 -9.17841196e-01
-6.18553519e-01 -8.25778306e-01 -5.84619880e-01 -1.39097917e+00
9.76026133e-02 -8.19137096e-01 -6.32457361e-02 7.41300225e-01
1.25256956e-01 -6.67570055e-01 2.14701191e-01 8.27649176e-01
-5.26646674e-01 -2.62244433e-01 8.18381608e-01 2.25029558e-01
-8.66221368e-01 4.77818996e-02 -9.88352716e-01 7.23652601e-01
6.75976992e-01 -3.90208095e-01 -5.11986494e-01 -2.62073696e-01
7.79814184e-01 -1.52816504e-01 3.77727896e-01 -9.49507117e-01
6.31563246e-01 -3.69879335e-01 5.22006512e-01 -6.78982496e-01
-5.79379946e-02 -9.46795166e-01 2.69849271e-01 1.20518841e-01
-8.91392887e-01 2.58631736e-01 5.43344736e-01 2.06485584e-01
2.63682693e-01 -4.74915355e-01 5.73888794e-02 -1.94645315e-01
-5.13450623e-01 -2.88358539e-01 -3.70821208e-01 -5.70849031e-02
7.74869025e-01 -1.29542410e-01 -8.53747904e-01 -2.10109249e-01
-7.40112424e-01 1.73355177e-01 6.11634791e-01 3.09063017e-01
4.02056843e-01 -9.52214956e-01 4.61807176e-02 4.69498426e-01
4.30535197e-01 -3.43785614e-01 1.70654719e-04 6.08546674e-01
-5.56841135e-01 4.40510213e-01 -3.10637981e-01 -1.84714347e-01
-1.57347727e+00 6.69233382e-01 -2.07258329e-01 1.83921844e-01
-8.55065465e-01 6.52227461e-01 -3.06548625e-01 -4.74374026e-01
5.20438433e-01 -5.27860880e-01 -1.11172572e-01 -4.13402505e-02
9.48354483e-01 9.75928724e-01 2.80515730e-01 5.33570111e-01
-1.59897164e-01 -5.16685396e-02 -3.28310162e-01 5.82449697e-02
1.46762598e+00 1.57323480e-01 3.08698192e-02 6.12075806e-01
8.72035682e-01 5.16264677e-01 -1.01451480e+00 -1.33477628e-01
2.61012644e-01 -1.16079092e+00 -2.29306077e-03 -1.39533091e+00
-6.60563886e-01 8.07453394e-01 5.66326857e-01 6.39808238e-01
6.12358093e-01 7.26041496e-02 9.70362201e-02 1.05827785e+00
9.16593894e-02 -1.28398025e+00 2.02895060e-01 -2.15476453e-01
1.11684239e+00 -1.41188943e+00 3.78924966e-01 -1.32508010e-01
-3.88567328e-01 1.49817491e+00 9.46261823e-01 4.12075639e-01
9.92170095e-01 2.10039362e-01 1.45232871e-01 -7.10346103e-01
-9.70675707e-01 7.33109295e-01 6.19710125e-02 3.30987871e-01
1.12302399e+00 -2.73108900e-01 -8.53645988e-03 8.68490934e-01
-3.18241566e-01 6.06707633e-01 3.46313536e-01 1.33689857e+00
-3.29796016e-01 -9.60299909e-01 -5.17787993e-01 9.01664972e-01
-2.51281977e-01 -2.22324118e-01 -2.83553332e-01 1.09310758e+00
-3.48941863e-01 8.99507880e-01 1.62568912e-01 -2.99329638e-01
3.62546504e-01 1.52857110e-01 -1.78562194e-01 -7.51739085e-01
-1.15996993e+00 -1.64172184e-02 2.47007847e-01 -6.45462573e-01
1.76451772e-01 -7.57331848e-01 -1.22278702e+00 -4.70776111e-01
-1.88281938e-01 5.53820431e-01 1.25629520e+00 3.94944876e-01
9.83124912e-01 3.70143920e-01 -5.07129021e-02 -2.92519331e-01
-3.49276721e-01 -1.03257668e+00 -2.15491191e-01 4.15175676e-01
9.48776305e-02 -3.77688169e-01 -2.83455789e-01 2.79424608e-01] | [8.466347694396973, 7.260227203369141] |
c4f5e49e-0469-4078-b574-076a484e0886 | p-2-sdf-for-neural-indoor-scene | 2303.00236 | null | https://arxiv.org/abs/2303.00236v1 | https://arxiv.org/pdf/2303.00236v1.pdf | P$^2$SDF for Neural Indoor Scene Reconstruction | Given only a set of images, neural implicit surface representation has shown its capability in 3D surface reconstruction. However, as the nature of per-scene optimization is based on the volumetric rendering of color, previous neural implicit surface reconstruction methods usually fail in low-textured regions, including the floors, walls, etc., which commonly exist for indoor scenes. Being aware of the fact that these low-textured regions usually correspond to planes, without introducing additional ground-truth supervisory signals or making additional assumptions about the room layout, we propose to leverage a novel Pseudo Plane-regularized Signed Distance Field (P$^2$SDF) for indoor scene reconstruction. Specifically, we consider adjacent pixels with similar colors to be on the same pseudo planes. The plane parameters are then estimated on the fly during training by an efficient and effective two-step scheme. Then the signed distances of the points on the planes are regularized by the estimated plane parameters in the training phase. As the unsupervised plane segments are usually noisy and inaccurate, we propose to assign different weights to the sampled points on the plane in plane estimation as well as the regularization loss. The weights come by fusing the plane segments from different views. As the sampled rays in the planar regions are redundant, leading to inefficient training, we further propose a keypoint-guided rays sampling strategy that attends to the informative textured regions with large color variations, and the implicit network gets a better reconstruction, compared with the original uniform ray sampling strategy. Experiments show that our P$^2$SDF achieves competitive reconstruction performance in Manhattan scenes. Further, as we do not introduce any additional room layout assumption, our P$^2$SDF generalizes well to the reconstruction of non-Manhattan scenes. | ['Shenghua Gao', 'Lina Cao', 'Zhengyu Zhang', 'Zhengxin Li', 'Ruoyu Wang', 'Jinpeng Yu', 'Jing Li'] | 2023-03-01 | null | null | null | null | ['indoor-scene-reconstruction'] | ['computer-vision'] | [ 4.56059247e-01 2.21081510e-01 8.33233446e-02 -3.83616120e-01
-6.70874596e-01 -1.84732541e-01 2.46364474e-01 -5.46640716e-02
-1.19106956e-01 6.59488380e-01 -2.93170270e-02 4.59104888e-02
-1.46103248e-01 -1.12472463e+00 -1.12331176e+00 -8.51973414e-01
3.16430703e-02 5.11519849e-01 2.13386044e-01 -1.76662073e-01
2.52227068e-01 5.74341297e-01 -1.46977472e+00 2.06366614e-01
9.97314274e-01 1.03239787e+00 2.58726627e-01 3.78629655e-01
-4.88273054e-01 3.21443707e-01 -3.78550649e-01 -2.13881671e-01
4.81909156e-01 -2.17292547e-01 -3.89292628e-01 4.17098731e-01
7.68937111e-01 -3.78308028e-01 -1.47439584e-01 1.10532653e+00
1.35763109e-01 3.19482982e-01 5.93859673e-01 -6.68640971e-01
-2.59346724e-01 -2.07844839e-01 -7.76011229e-01 -5.70956349e-01
3.22640806e-01 -2.13672034e-02 8.13260794e-01 -1.08322275e+00
5.33561051e-01 1.05138779e+00 7.32446671e-01 4.25090879e-01
-1.34434414e+00 -5.30830562e-01 8.55302215e-01 -1.96203575e-01
-1.55529988e+00 -4.15175676e-01 1.33250296e+00 -2.97813058e-01
3.44027936e-01 4.09059495e-01 7.08318412e-01 8.11095655e-01
-2.16682971e-01 7.26485968e-01 1.06440830e+00 -3.31390500e-01
4.32015985e-01 2.30666175e-01 -1.10039443e-01 1.01761878e+00
3.32072020e-01 -1.19998753e-01 -6.32262826e-01 1.73353718e-03
1.26829922e+00 1.96889356e-01 -7.78427184e-01 -8.84831607e-01
-1.06702518e+00 6.87995791e-01 7.06541538e-01 1.99768990e-02
-2.85849482e-01 1.08114205e-01 -9.15001333e-02 -1.48905218e-01
6.48949564e-01 4.47107911e-01 -4.24701840e-01 3.68570656e-01
-9.92333412e-01 2.18794927e-01 6.28257573e-01 1.00385690e+00
1.34573698e+00 3.87000069e-02 1.70758143e-01 9.25978184e-01
4.12309527e-01 4.40658391e-01 -2.40454555e-01 -1.30463624e+00
7.44492114e-01 7.23311245e-01 3.19448233e-01 -1.28402436e+00
-3.36156636e-02 -3.97875279e-01 -1.03027630e+00 4.21979368e-01
5.46851575e-01 1.37732044e-01 -1.09375668e+00 1.47992063e+00
5.68870366e-01 1.85683772e-01 -2.85019130e-01 9.48852181e-01
4.48707640e-01 6.33909404e-01 -5.57560563e-01 -3.63391303e-02
1.11893535e+00 -7.79093921e-01 -4.32350129e-01 -2.89322317e-01
3.07274580e-01 -5.15595794e-01 1.20205319e+00 5.58449566e-01
-1.03330362e+00 -3.98965955e-01 -1.04620469e+00 -8.09038579e-02
-9.99639742e-03 -5.53665683e-02 6.98993981e-01 4.43637788e-01
-7.79477060e-01 5.75129092e-01 -8.85703444e-01 1.38393492e-01
6.45336270e-01 2.50166059e-01 -2.00221598e-01 -4.80072439e-01
-5.69843531e-01 2.17751399e-01 -1.12631276e-01 4.20592397e-01
-6.77816272e-01 -7.91098058e-01 -9.20299113e-01 -1.84505939e-01
6.20807111e-01 -7.50141025e-01 6.77859664e-01 -1.19202733e+00
-1.51636577e+00 6.52459025e-01 -4.61045712e-01 2.07588643e-01
5.73103666e-01 -2.33793825e-01 1.30364493e-01 2.72903860e-01
1.26086563e-01 4.74344492e-01 8.53475809e-01 -2.04138279e+00
-3.40157419e-01 -5.35158575e-01 2.24974647e-01 4.29882318e-01
5.22140786e-02 -8.89991283e-01 -7.51506925e-01 -5.15721142e-01
9.05614495e-01 -7.81837881e-01 -4.65445638e-01 6.33185446e-01
-7.00886667e-01 2.72160530e-01 4.96965170e-01 -5.80506504e-01
7.81267464e-01 -2.21176505e+00 1.28777847e-01 7.93713808e-01
2.35091537e-01 -4.12568122e-01 9.39376503e-02 -8.21433663e-02
2.27276713e-01 -1.11262575e-02 -8.34086180e-01 -6.61034763e-01
-1.02549702e-01 4.48058665e-01 -1.27714932e-01 5.18973708e-01
3.82026024e-02 3.64671737e-01 -8.76505971e-01 -3.57161015e-01
3.21449995e-01 6.87320650e-01 -7.75660515e-01 1.63061827e-01
-3.44479501e-01 6.81557477e-01 -8.45933855e-01 6.32429421e-01
1.02577388e+00 -2.03264147e-01 1.62107542e-01 -3.01041156e-01
-1.53368369e-01 3.25408101e-01 -1.48267496e+00 2.01396251e+00
-6.16677046e-01 2.88636267e-01 3.88997823e-01 -8.80310714e-01
9.55375612e-01 1.02420270e-01 5.15084803e-01 -6.20097339e-01
-2.34750226e-01 2.41348460e-01 -5.12144268e-01 -2.34442577e-01
3.08106631e-01 -1.14896253e-01 3.70919377e-01 -3.31875868e-02
-4.20472145e-01 -6.93882763e-01 -3.05190861e-01 -1.14962302e-01
7.27516949e-01 4.17482913e-01 -1.94792435e-01 -4.67116050e-02
3.78089637e-01 -1.53216019e-01 7.62707174e-01 4.85267967e-01
2.43661940e-01 1.06806493e+00 2.96914756e-01 -4.88411516e-01
-9.61858690e-01 -1.16479027e+00 -2.65017182e-01 3.97540003e-01
4.23327714e-01 -3.03511620e-01 -8.24167311e-01 -6.35257483e-01
-3.08933705e-01 6.05734646e-01 -5.42122543e-01 1.89573228e-01
-8.74302268e-01 -4.90526557e-01 -1.86011910e-01 2.67378002e-01
6.74517870e-01 -5.94758511e-01 -5.02736807e-01 2.00107500e-01
-3.21247637e-01 -1.00306046e+00 -3.39504778e-01 4.56767917e-01
-1.11772394e+00 -1.12112343e+00 -8.82129252e-01 -5.71228981e-01
1.29446733e+00 4.26922113e-01 1.20341492e+00 1.43452376e-01
-8.30488950e-02 3.85741413e-01 -1.13724522e-01 -7.07981065e-02
3.27910632e-01 -3.31469476e-01 -3.53262573e-01 2.89315879e-01
-1.99036926e-01 -7.92836249e-01 -9.15725172e-01 3.91400069e-01
-5.03481507e-01 5.32165349e-01 2.97028840e-01 8.00592840e-01
1.15303409e+00 1.45236909e-01 -1.96717843e-01 -1.09203517e+00
-2.42293611e-01 -3.19800973e-01 -7.69993544e-01 -1.61934309e-02
-3.86089414e-01 1.44035667e-01 5.69667041e-01 -1.80235729e-01
-1.07729399e+00 3.06922048e-01 -1.16491362e-01 -6.12981439e-01
-1.95829123e-01 1.43409893e-01 -4.66384023e-01 -2.05910727e-01
4.12001252e-01 2.90846348e-01 -3.58547777e-01 -5.54407537e-01
7.52162263e-02 -6.47452548e-02 3.54660675e-02 -8.14868391e-01
1.12213683e+00 9.32273805e-01 1.68455064e-01 -1.02458203e+00
-8.73845220e-01 -2.88115352e-01 -4.99496788e-01 -2.07718030e-01
8.18753600e-01 -8.04592609e-01 -5.34198463e-01 3.05934191e-01
-1.12480044e+00 -6.45760059e-01 -3.90244842e-01 3.55195761e-01
-4.63373184e-01 4.51650083e-01 -2.99486637e-01 -1.01164186e+00
8.84533376e-02 -1.18908489e+00 1.30764866e+00 5.95981851e-02
7.02544907e-03 -8.92006755e-01 -3.90468426e-02 4.63685572e-01
2.27238573e-02 3.02508622e-01 1.13787687e+00 2.95786560e-01
-1.29718900e+00 -2.53904536e-02 -1.07481711e-01 2.96869189e-01
2.24353060e-01 -2.06797197e-01 -1.18764043e+00 -3.71488519e-02
1.06314883e-01 -1.37622476e-01 7.01274097e-01 6.04181647e-01
1.61783206e+00 -3.20949793e-01 -2.15977922e-01 1.16191471e+00
1.49633658e+00 2.34793010e-03 6.45613074e-01 3.64093930e-01
1.08080745e+00 9.34047580e-01 3.32384974e-01 3.58012557e-01
5.18002450e-01 6.57436252e-01 8.04483712e-01 -2.61262149e-01
-2.88172774e-02 -3.85012060e-01 1.27378166e-01 6.99412107e-01
-4.43703622e-01 -1.34169534e-01 -4.74959522e-01 2.00094908e-01
-1.49805999e+00 -5.95481157e-01 -1.40379578e-01 2.64299798e+00
8.34919035e-01 1.27894714e-01 -3.68491292e-01 2.17525557e-01
3.59241813e-01 4.34365839e-01 -5.51768064e-01 -8.40781629e-02
-1.86513290e-01 3.70525658e-01 5.57744622e-01 7.24499941e-01
-8.64663064e-01 6.03630602e-01 4.96587181e+00 7.29176641e-01
-1.04021621e+00 -1.28475398e-01 8.96634042e-01 -1.62052855e-01
-8.69797468e-01 1.77779309e-02 -6.09600067e-01 3.47560793e-01
1.99964166e-01 6.57641470e-01 6.71031237e-01 6.56823277e-01
1.65768340e-01 -2.81309277e-01 -1.05904460e+00 1.08671927e+00
-1.12837590e-01 -1.29187858e+00 1.02591261e-01 1.00772753e-01
9.95647550e-01 -1.54600725e-01 3.31605822e-02 -2.82905828e-02
1.13749593e-01 -9.72338259e-01 9.18370008e-01 4.97892290e-01
8.78699720e-01 -5.89351058e-01 2.47557789e-01 4.76720572e-01
-1.27394474e+00 2.63795286e-01 -4.32460666e-01 2.99840391e-01
5.57591878e-02 9.16754186e-01 -4.73587334e-01 6.63363695e-01
7.78220296e-01 6.25557721e-01 -8.40812400e-02 8.20634246e-01
-3.39164466e-01 3.60858172e-01 -4.96789426e-01 3.44983369e-01
1.54516101e-01 -7.09141374e-01 4.08677906e-01 6.83689773e-01
4.18248802e-01 9.87515599e-02 3.26165944e-01 1.15870810e+00
-3.23630907e-02 -5.38615473e-02 -5.63487589e-01 4.38095868e-01
2.16235265e-01 1.00800872e+00 -7.92228520e-01 -2.07713068e-01
-4.72793609e-01 1.15769970e+00 3.79631400e-01 8.39765787e-01
-5.50825179e-01 -9.41829309e-02 6.21710241e-01 6.28337502e-01
2.36331329e-01 -4.51868057e-01 -6.74726784e-01 -1.16694498e+00
3.27413052e-01 -6.73332512e-01 -1.33626267e-01 -5.43788135e-01
-1.08930635e+00 4.08341169e-01 -1.17631584e-01 -1.25693882e+00
3.06181997e-01 -5.76552689e-01 -4.58259106e-01 9.77602124e-01
-1.90677285e+00 -8.41922700e-01 -5.56414962e-01 6.95150852e-01
6.41926587e-01 4.86027479e-01 8.05625081e-01 3.17497671e-01
-4.16266292e-01 3.50798160e-01 4.90629971e-02 6.30827099e-02
3.45339596e-01 -1.25226820e+00 4.68934067e-02 6.61022782e-01
-4.91821766e-02 7.79303908e-01 3.81335944e-01 -5.88560581e-01
-1.45076084e+00 -1.04436088e+00 1.97552204e-01 -5.11445552e-02
-1.07377447e-01 -6.70105159e-01 -1.04846454e+00 3.69390249e-01
-3.44722837e-01 2.67216682e-01 4.13453251e-01 1.28625706e-01
-3.50783974e-01 -2.54880965e-01 -1.10269856e+00 8.06835413e-01
1.20158017e+00 -4.85423833e-01 -8.51659998e-02 3.45023483e-01
5.05323350e-01 -5.26761293e-01 -4.39471900e-01 3.98407161e-01
2.93760985e-01 -1.14055467e+00 1.33484840e+00 1.23827145e-01
5.49971998e-01 -4.62059826e-01 -4.29794341e-01 -1.13919640e+00
-1.77096799e-01 -2.69127339e-01 -9.00016576e-02 9.31368113e-01
3.75600696e-01 -5.99056661e-01 1.24774301e+00 6.99956059e-01
-5.24247527e-01 -9.94333804e-01 -1.04771256e+00 -4.24203306e-01
-2.94910640e-01 -5.73142290e-01 5.96701324e-01 1.01679134e+00
-7.86276400e-01 3.73388566e-02 -1.68549061e-01 5.49007654e-01
1.00560606e+00 2.52689451e-01 7.83625782e-01 -1.25471616e+00
-4.42173451e-01 -3.01117420e-01 4.26135547e-02 -1.63593113e+00
-4.08767015e-02 -6.76008344e-01 3.97339046e-01 -1.61477959e+00
-1.17673092e-01 -1.12828553e+00 -1.45327717e-01 1.65833667e-01
-1.91993177e-01 1.78784102e-01 -1.46419257e-01 2.09598705e-01
-1.75114587e-01 8.90903831e-01 1.74337673e+00 -2.52664983e-01
-3.08370888e-01 1.38867691e-01 -4.27113086e-01 1.11146367e+00
4.05516028e-01 -2.42356345e-01 -5.68971455e-01 -7.55436242e-01
3.24504048e-01 5.55986771e-04 4.46060359e-01 -8.94173682e-01
-7.35154450e-02 -3.79744112e-01 7.46305406e-01 -5.83450377e-01
1.00049198e+00 -1.21449351e+00 1.53121889e-01 1.25764772e-01
-4.00226116e-02 -4.63176221e-01 -1.22963868e-01 7.31578767e-01
-1.67454965e-02 -2.18465731e-01 7.65403748e-01 -5.20841777e-01
-3.28641951e-01 5.97316265e-01 2.35709772e-02 -9.28087085e-02
6.92415953e-01 -9.14245367e-01 1.93662539e-01 -2.55715221e-01
-3.35540473e-01 3.05856969e-02 8.49138081e-01 -1.64377287e-01
9.52324629e-01 -1.08606970e+00 -2.22634599e-01 5.74136853e-01
4.28820355e-03 1.14000094e+00 3.18038464e-01 6.73898160e-01
-6.82805181e-01 -6.77997917e-02 2.56568104e-01 -7.61323988e-01
-7.96391010e-01 1.05685055e-01 4.98562843e-01 -1.38537940e-02
-1.04190564e+00 1.04366255e+00 8.91404629e-01 -5.78088760e-01
2.94335723e-01 -5.95851481e-01 8.89974833e-02 -1.96195498e-01
7.99539387e-02 1.89886093e-01 2.82586645e-03 -3.52112263e-01
-1.28847167e-01 9.97860968e-01 7.01299906e-02 -9.94508043e-02
1.45851660e+00 -7.97233656e-02 -7.31692836e-02 7.66251564e-01
1.05868185e+00 4.92135048e-01 -1.73863912e+00 -1.22830182e-01
-3.76919240e-01 -8.28906834e-01 1.90565437e-01 -5.09977818e-01
-1.22455215e+00 9.86498415e-01 4.71296817e-01 -2.32814118e-01
9.70039964e-01 -3.35766226e-01 7.91524768e-01 2.69909352e-01
7.10521102e-01 -1.00748646e+00 -1.76809747e-02 4.22344744e-01
7.38530695e-01 -1.17058241e+00 2.25350514e-01 -8.77747416e-01
-2.54665732e-01 1.13326037e+00 6.11976445e-01 -2.19843894e-01
4.90102977e-01 6.23831637e-02 -9.31459963e-02 -3.64062130e-01
-1.82570294e-01 2.54977494e-01 3.02194297e-01 4.92885828e-01
3.24203610e-01 9.66447219e-02 3.89523894e-01 1.65599257e-01
-9.28311422e-02 -3.25971633e-01 1.84174910e-01 8.45879495e-01
-4.10832673e-01 -9.44107175e-01 -6.40544057e-01 3.79347295e-01
1.02744380e-04 -9.61534083e-02 -4.63276394e-02 7.04170823e-01
2.43028894e-01 4.25356567e-01 1.42543584e-01 -3.12846377e-02
4.35704082e-01 -3.00409079e-01 6.39674664e-01 -7.04677820e-01
-1.72024280e-01 3.25413585e-01 -1.42458985e-02 -8.06094885e-01
-2.26140365e-01 -5.39726019e-01 -1.48778391e+00 -7.45164752e-02
-1.65906936e-01 5.00687994e-02 7.27327049e-01 7.23517478e-01
1.34275720e-01 5.68114758e-01 8.73178184e-01 -1.31240034e+00
-3.53578217e-02 -4.10095632e-01 -6.11533225e-01 2.76757896e-01
5.18797100e-01 -8.30388963e-01 -5.21875560e-01 -2.11411506e-01] | [8.875767707824707, -3.0162830352783203] |
2838d20a-7b29-40e5-84a1-4db77ed5a879 | voice-privacy-with-smart-digital-assistants | 2104.11038 | null | https://arxiv.org/abs/2104.11038v1 | https://arxiv.org/pdf/2104.11038v1.pdf | Voice Privacy with Smart Digital Assistants in Educational Settings | The emergence of voice-assistant devices ushers in delightful user experiences not just on the smart home front, but also in diverse educational environments from classrooms to personalized-learning/tutoring. However, the use of voice as an interaction modality also could result in exposure of user's identity, and hinders the broader adoption of voice interfaces; this is especially important in environments where children are present and their voice privacy needs to be protected. To this end, building on state-of-the-art techniques proposed in the literature, we design and evaluate a practical and efficient framework for voice privacy at the source. The approach combines speaker identification (SID) and speech conversion methods to randomly disguise the identity of users right on the device that records the speech, while ensuring that the transformed utterances of users can still be successfully transcribed by Automatic Speech Recognition (ASR) solutions. We evaluate the ASR performance of the conversion in terms of word error rate and show the promise of this framework in preserving the content of the input speech. | ['Ravi Kokku', 'Aditya Vempaty', 'Mohammad Niknazar'] | 2021-03-24 | null | null | null | null | ['speaker-identification'] | ['speech'] | [ 3.62535149e-01 5.39882779e-01 2.34476477e-01 -2.82136083e-01
-6.18223071e-01 -1.08498645e+00 4.48871553e-01 9.35820788e-02
-2.68589735e-01 5.54775536e-01 3.81008059e-01 -5.90083838e-01
-1.96963027e-02 -4.36177403e-01 -3.37195128e-01 -5.03900170e-01
5.84339976e-01 9.30445418e-02 1.47566885e-01 6.81515709e-02
6.87739104e-02 7.71052778e-01 -1.88965940e+00 1.82756968e-02
8.09864223e-01 6.21620893e-01 -2.35120296e-01 7.46660709e-01
-2.36035347e-01 3.59952599e-01 -1.11577928e+00 -6.90695107e-01
8.78268778e-02 -1.36084318e-01 -7.13997185e-01 -3.12662311e-02
5.66477716e-01 -5.57857335e-01 -4.44579244e-01 1.18843663e+00
9.04257417e-01 1.40424550e-01 5.06989211e-02 -1.18998909e+00
-2.81351954e-01 5.46127439e-01 8.46917927e-02 -1.95381954e-01
7.79619098e-01 -1.58554196e-01 4.55711901e-01 -3.68841559e-01
5.00334918e-01 9.86606121e-01 2.96877712e-01 9.76742208e-01
-1.27999735e+00 -9.03574705e-01 -2.52405554e-01 -1.06896423e-01
-1.47754300e+00 -1.14126575e+00 4.74390626e-01 -3.62248451e-01
6.14117086e-01 9.28140521e-01 4.66670454e-01 1.36456084e+00
-2.99736410e-01 6.07124209e-01 8.40594471e-01 -7.05125391e-01
2.96785772e-01 9.18088794e-01 8.29892084e-02 3.13842267e-01
1.59688041e-01 -1.62457973e-01 -6.25997365e-01 -2.08704576e-01
3.25403899e-01 -1.78006738e-01 -6.40562713e-01 -4.58464861e-01
-6.46119475e-01 4.19659942e-01 -6.67639554e-01 5.76779485e-01
-1.48984045e-01 -5.01139939e-01 4.49006975e-01 3.71119201e-01
3.06719452e-01 3.91074151e-01 -1.91451654e-01 -6.27110183e-01
-1.09892452e+00 6.20909110e-02 1.12275648e+00 1.25775731e+00
7.41381273e-02 -1.48893133e-01 -6.52470291e-02 8.32249939e-01
3.18575829e-01 3.48021597e-01 5.22859633e-01 -6.99096441e-01
3.36766571e-01 3.28494877e-01 4.25054617e-02 -7.05399215e-01
2.40343109e-01 -1.65894508e-01 -4.44115072e-01 2.44634762e-01
5.22499561e-01 -1.97804421e-01 -4.67447013e-01 1.61838448e+00
6.03278399e-01 7.79959047e-03 1.84428647e-01 4.03890520e-01
8.07003796e-01 4.40644890e-01 -2.70955205e-01 -3.58541340e-01
1.32029402e+00 -6.16366863e-01 -1.25749111e+00 2.00707331e-01
2.47587144e-01 -1.06222510e+00 9.42825794e-01 4.29895252e-01
-9.45042312e-01 -9.71287042e-02 -9.51395690e-01 -1.75528210e-02
-4.18553293e-01 1.12758409e-02 1.03703804e-01 1.76779878e+00
-1.11968601e+00 2.61605978e-01 -7.68464506e-01 -5.94800532e-01
2.83217639e-01 8.00691307e-01 -7.02128351e-01 6.96554855e-02
-8.50566030e-01 5.31067669e-01 -3.02665055e-01 -1.14817105e-01
-3.30455899e-01 -6.55211091e-01 -6.55783534e-01 1.71812326e-01
1.70827881e-01 -3.01026274e-03 1.41719520e+00 -7.01472104e-01
-2.10067654e+00 8.80479991e-01 -6.87087849e-02 -1.77781984e-01
1.00949717e+00 -4.07703429e-01 -7.27202117e-01 1.29755527e-01
-5.29491566e-02 4.33444604e-02 9.17412758e-01 -1.00856650e+00
-7.14206278e-01 -4.19543356e-01 -1.76745296e-01 -8.67865980e-02
-7.33254790e-01 4.10394818e-01 -2.60239065e-01 -3.23873729e-01
-7.08480030e-02 -8.86393130e-01 2.73839295e-01 -1.51443586e-01
-5.02314210e-01 1.91285446e-01 1.21402860e+00 -1.03611982e+00
1.52540052e+00 -2.60717249e+00 -3.27051014e-01 3.08017462e-01
-8.25712979e-02 8.97074759e-01 3.82838547e-01 5.26227951e-01
-7.22564161e-02 1.11772835e-01 2.28621826e-01 -5.65962732e-01
7.46186674e-02 1.39554828e-01 -4.85209346e-01 5.99376202e-01
-4.83976275e-01 2.40298554e-01 -4.97298568e-01 -1.41977534e-01
4.65142429e-01 7.53779709e-01 -4.55404997e-01 6.11531913e-01
4.28293854e-01 5.31506956e-01 -2.33198553e-01 4.57551718e-01
7.23400891e-01 6.66868508e-01 1.74114138e-01 3.82595718e-01
-5.01048565e-01 7.13729918e-01 -1.20296025e+00 1.22403300e+00
-4.59581852e-01 7.06193745e-01 8.20561707e-01 -3.17949861e-01
7.73828328e-01 9.55014169e-01 2.17211172e-01 -2.71761984e-01
1.64578110e-01 1.98897377e-01 -1.99891329e-01 -7.14384854e-01
5.67092657e-01 3.03282142e-01 1.56516209e-01 4.32551235e-01
6.49871528e-02 -4.08686101e-02 -5.29711187e-01 -1.58807650e-01
8.84055316e-01 -1.76528841e-01 1.59626946e-01 -3.08452219e-01
6.90821588e-01 -7.45185614e-01 1.39957204e-01 5.58082938e-01
-3.41296732e-01 5.07949173e-01 1.91684842e-01 2.42309675e-01
-6.35892153e-01 -8.36386323e-01 1.22500472e-02 1.08075726e+00
-4.08586293e-01 -3.08177739e-01 -1.33696413e+00 -7.58757174e-01
-3.01725328e-01 1.29865789e+00 -6.25832975e-02 -9.88493264e-02
-4.48370069e-01 2.66729236e-01 1.01652515e+00 -1.46138534e-01
7.09886029e-02 -7.86887646e-01 -7.03532517e-01 1.87782109e-01
-5.38795926e-02 -1.08748353e+00 -8.29149425e-01 -2.31917128e-02
-3.83761048e-01 -5.40704072e-01 -5.46544313e-01 -6.43119216e-01
5.98194718e-01 2.04371467e-01 3.30330133e-01 -1.98557973e-01
-7.47221559e-02 9.17541504e-01 -3.67807835e-01 -5.38217545e-01
-1.01390660e+00 2.97316670e-01 2.11504534e-01 2.30374515e-01
3.69239569e-01 -9.07434523e-01 -2.40265116e-01 2.47214794e-01
-8.19276392e-01 -4.81700450e-01 -8.00594762e-02 4.59684610e-01
-1.71071425e-01 1.75879955e-01 5.38057446e-01 -8.00289690e-01
6.18913710e-01 -1.30455002e-01 -5.61968565e-01 3.90834332e-01
-6.38107777e-01 -2.94364244e-01 5.79036713e-01 -6.34516120e-01
-9.26104128e-01 8.47980008e-02 -5.76509476e-01 -3.42897356e-01
-7.71445632e-01 -3.17832768e-01 -8.97506714e-01 -2.19641656e-01
2.18119293e-01 4.22180176e-01 1.89445272e-01 -5.90201616e-01
2.86715180e-01 1.59909880e+00 7.05377817e-01 -2.61857510e-01
4.94410008e-01 6.00427389e-02 -4.05165166e-01 -1.43034828e+00
-1.08837172e-01 -5.06874502e-01 -5.53028405e-01 -2.51613349e-01
6.06723309e-01 -6.06671870e-01 -9.33053255e-01 6.86778665e-01
-1.20776510e+00 2.16399536e-01 -1.89846709e-01 3.95908266e-01
-2.73700237e-01 3.44702661e-01 -1.56733453e-01 -1.35751998e+00
-4.03932333e-01 -1.31900120e+00 7.92164683e-01 2.93934017e-01
-6.67032123e-01 -7.95107901e-01 -2.97381639e-01 8.12044263e-01
4.59760606e-01 -9.86045152e-02 7.49848425e-01 -1.15471280e+00
-2.51514733e-01 -5.94387352e-01 5.04424572e-01 3.50530833e-01
4.11931187e-01 8.04186519e-03 -1.55022466e+00 -5.14236808e-01
2.71633118e-01 1.86365783e-01 -1.90801844e-01 -8.78164247e-02
7.98842013e-01 -8.89845371e-01 -2.73567587e-02 3.49292159e-01
8.04182351e-01 4.49678391e-01 6.75976455e-01 7.25948587e-02
5.72507322e-01 1.13628483e+00 2.77553767e-01 2.59832352e-01
9.50695053e-02 7.78612494e-01 1.48693785e-01 2.99478680e-01
3.46511900e-02 -6.40519261e-01 4.14054811e-01 8.67518842e-01
4.37882930e-01 -2.64135391e-01 -6.44976854e-01 6.49218380e-01
-1.42146480e+00 -1.00212145e+00 2.82761035e-03 2.67352891e+00
8.50072205e-01 -2.68016696e-01 1.87996611e-01 3.92747819e-01
1.07200730e+00 -1.33048177e-01 -1.25715733e-01 -7.81844020e-01
4.12428349e-01 3.88056308e-01 5.62346995e-01 5.79542756e-01
-9.01407182e-01 8.10945332e-01 5.89951944e+00 6.64479911e-01
-1.46105003e+00 1.57329902e-01 1.55297652e-01 -1.97270989e-01
-3.79303873e-01 -3.24468672e-01 -5.95611811e-01 4.67185795e-01
9.96136844e-01 -2.07659572e-01 6.36818409e-01 8.54407847e-01
1.67556927e-01 2.28897035e-01 -1.12595701e+00 1.04735398e+00
1.27088986e-02 -7.33945966e-01 -3.26212227e-01 3.51069689e-01
1.17097735e-01 -5.46905458e-01 1.45299286e-01 -7.53780231e-02
-8.73784050e-02 -8.79978299e-01 9.11102295e-01 -2.53822412e-02
8.85171056e-01 -9.45715725e-01 4.75941807e-01 4.24228013e-01
-7.68558860e-01 6.40224442e-02 9.22222584e-02 1.63017765e-01
-7.60448352e-02 1.94622815e-01 -1.29502571e+00 2.66084313e-01
5.17549396e-01 -1.85023487e-01 -2.31635392e-01 8.61925602e-01
-3.33119988e-01 8.10596645e-01 -3.53503942e-01 -1.57450199e-01
-3.63821805e-01 -1.65603966e-01 8.52903247e-01 1.13885212e+00
6.79661512e-01 1.79845512e-01 -4.31327969e-01 4.87333655e-01
9.25424416e-03 3.69264066e-01 -8.09790015e-01 -2.94876128e-01
1.04879057e+00 1.04746914e+00 -3.79327893e-01 2.34929577e-01
-3.01303536e-01 1.25709498e+00 -2.19921261e-01 6.95041791e-02
-2.62084365e-01 -5.84009171e-01 1.06900549e+00 3.82597148e-01
1.87761322e-01 -5.47959246e-02 -1.07323378e-01 -9.01158929e-01
9.70372558e-02 -1.26812398e+00 3.96427587e-02 -2.89071858e-01
-7.09605515e-01 4.21869010e-01 -3.34600717e-01 -9.66858506e-01
-4.99815673e-01 -1.33581817e-01 -5.60402811e-01 7.45372117e-01
-7.39090919e-01 -1.04768276e+00 2.23659024e-01 4.68780011e-01
3.13253909e-01 -2.73029178e-01 1.20643449e+00 6.25667036e-01
-5.05996406e-01 1.39756823e+00 1.77991241e-01 9.01827365e-02
6.92335665e-01 -9.76392865e-01 3.45427662e-01 9.56940830e-01
1.35901183e-01 9.78151560e-01 9.80797410e-01 -2.28550985e-01
-1.60452569e+00 -7.05224514e-01 1.20269358e+00 -4.02711779e-01
3.58540237e-01 -8.23440909e-01 -8.57796669e-01 5.83583236e-01
3.24176669e-01 -5.10595441e-01 1.15762305e+00 -8.74249190e-02
-1.67448685e-01 -4.87547331e-02 -1.63687086e+00 6.43139005e-01
6.86330497e-01 -1.05854547e+00 -4.73245412e-01 -3.33450019e-01
6.98121786e-01 -4.09383118e-01 -7.69996524e-01 -1.23254679e-01
8.37571144e-01 -9.40688312e-01 5.17482102e-01 -2.77764678e-01
-1.21120848e-01 -1.25301734e-01 -2.72697359e-01 -9.23568487e-01
2.92855442e-01 -1.39183962e+00 1.27045348e-01 2.33733702e+00
3.64905298e-01 -8.63179862e-01 8.55255485e-01 1.36148262e+00
3.29131112e-02 -1.94577560e-01 -1.47020817e+00 -4.32350606e-01
-1.19450122e-01 -3.37405533e-01 9.26023424e-01 1.08016372e+00
2.68505663e-01 1.08885750e-01 -4.25820082e-01 7.04882562e-01
2.95201749e-01 -5.13785720e-01 9.54257905e-01 -9.98509407e-01
-1.21085517e-01 -2.04456493e-01 -6.06216311e-01 -7.29617536e-01
2.28238478e-01 -5.49651623e-01 5.29341325e-02 -9.63475823e-01
-5.20968139e-01 -1.22432239e-01 2.40905106e-01 3.08059573e-01
2.69441038e-01 -4.22997028e-01 2.30956405e-01 -2.64505565e-01
-8.04283172e-02 3.72151077e-01 5.95500708e-01 2.17501968e-01
-4.75602508e-01 7.73968160e-01 -6.49780571e-01 4.66920674e-01
6.41316175e-01 -4.61654782e-01 -5.40886223e-01 -6.38674274e-02
-4.59155560e-01 1.11504242e-01 -1.20628521e-01 -9.98050392e-01
4.09777522e-01 1.01356886e-01 -2.34166652e-01 -2.86830366e-01
1.42662361e-01 -1.50493252e+00 3.07978719e-01 2.60353297e-01
-4.10107017e-01 -3.80513400e-01 1.97332740e-01 8.72967988e-02
-1.18956203e-02 -4.24339354e-01 7.75766253e-01 3.65474075e-01
1.09554477e-01 -1.04873024e-01 -8.64210069e-01 -2.88345158e-01
9.34590578e-01 -4.57080871e-01 -2.45399162e-01 -5.98633945e-01
-4.73216951e-01 -7.20104128e-02 7.14150548e-01 6.48206830e-01
4.55352843e-01 -7.79619992e-01 -7.94929639e-02 8.43340755e-01
-1.07948683e-01 -2.16638893e-01 1.59600645e-01 4.03969795e-01
-2.98575670e-01 3.01874757e-01 5.33380285e-02 -2.01011002e-01
-2.14727950e+00 3.67189467e-01 1.26992434e-01 4.08704519e-01
-5.39233267e-01 6.97483599e-01 -1.72756642e-01 -6.82404280e-01
8.69284630e-01 -9.50312838e-02 -2.20702499e-01 1.60374895e-01
8.05278063e-01 6.80778801e-01 3.54634076e-01 -7.05830336e-01
-2.90441960e-01 -6.06584959e-02 -8.33080187e-02 -4.51790333e-01
1.16130912e+00 -5.18459618e-01 -1.74562976e-01 3.28344792e-01
1.22925186e+00 9.77680385e-01 -7.05422163e-01 1.36949763e-01
-9.54849795e-02 -6.88679934e-01 -7.82655627e-02 -7.61235118e-01
-7.78372526e-01 8.94916236e-01 8.20379555e-01 6.50778234e-01
9.05732632e-01 -2.93706566e-01 8.13939750e-01 4.21022251e-02
4.82893407e-01 -1.05783737e+00 -6.77092075e-01 2.25319520e-01
5.19000590e-01 -6.61526442e-01 -4.15286422e-01 -7.50655472e-01
-3.60823035e-01 1.07238257e+00 2.30427518e-01 6.82292759e-01
6.19655967e-01 3.93455327e-01 4.03039098e-01 2.36720592e-01
-3.22491825e-01 2.55991161e-01 6.86530173e-02 9.37118888e-01
5.29759526e-01 2.20588073e-01 -1.87037855e-01 9.05254424e-01
-7.49904394e-01 -1.35517940e-01 7.98375726e-01 1.05899072e+00
-2.46120110e-01 -1.36828029e+00 -6.46326721e-01 3.02009583e-02
-1.05335033e+00 6.32985756e-02 -8.38552177e-01 2.55246282e-01
5.83158992e-02 1.47380733e+00 -2.02115208e-01 -3.85557681e-01
6.48572862e-01 4.60529059e-01 1.63941279e-01 -7.22097754e-01
-1.00034750e+00 3.44304740e-02 3.35029393e-01 -2.56740302e-01
3.32043953e-02 -1.00553679e+00 -8.40758145e-01 -4.06426311e-01
-5.20283937e-01 4.17817861e-01 1.06727815e+00 7.70754874e-01
4.65383738e-01 2.73985207e-01 8.62399697e-01 -2.47829750e-01
-7.43486226e-01 -5.82174897e-01 -7.22390771e-01 -1.30535886e-02
5.62375247e-01 -4.19308282e-02 -3.55261266e-01 -1.59061417e-01] | [14.007795333862305, 5.840723514556885] |
feffbfa9-983d-4308-a820-a6a703abc7ef | point-supervised-single-cell-segmentation-via | 2304.10671 | null | https://arxiv.org/abs/2304.10671v2 | https://arxiv.org/pdf/2304.10671v2.pdf | Point-supervised Single-cell Segmentation via Collaborative Knowledge Sharing | Despite their superior performance, deep-learning methods often suffer from the disadvantage of needing large-scale well-annotated training data. In response, recent literature has seen a proliferation of efforts aimed at reducing the annotation burden. This paper focuses on a weakly-supervised training setting for single-cell segmentation models, where the only available training label is the rough locations of individual cells. The specific problem is of practical interest due to the widely available nuclei counter-stain data in biomedical literature, from which the cell locations can be derived programmatically. Of more general interest is a proposed self-learning method called collaborative knowledge sharing, which is related to but distinct from the more well-known consistency learning methods. This strategy achieves self-learning by sharing knowledge between a principal model and a very light-weight collaborator model. Importantly, the two models are entirely different in their architectures, capacities, and model outputs: In our case, the principal model approaches the segmentation problem from an object-detection perspective, whereas the collaborator model a sematic segmentation perspective. We assessed the effectiveness of this strategy by conducting experiments on LIVECell, a large single-cell segmentation dataset of bright-field images, and on A431 dataset, a fluorescence image dataset in which the location labels are generated automatically from nuclei counter-stain data. Implementing code is available at https://github.com/jiyuuchc/lacss | ['Ji Yu'] | 2023-04-20 | null | null | null | null | ['cell-segmentation', 'self-learning'] | ['medical', 'natural-language-processing'] | [ 1.15435258e-01 1.64213985e-01 -9.22740698e-02 -3.65191221e-01
-1.07553959e+00 -7.67318070e-01 3.53691936e-01 4.23049659e-01
-8.29161823e-01 1.00405717e+00 -2.63253689e-01 3.55250351e-02
-5.90115339e-02 -6.23462796e-01 -7.72777855e-01 -1.24088573e+00
3.27028602e-01 7.16071606e-01 3.70112598e-01 1.57603368e-01
2.21592098e-01 3.85192156e-01 -1.19565713e+00 5.31322956e-02
8.25774372e-01 8.05805683e-01 3.30489218e-01 6.33601129e-01
-8.21528584e-02 5.66904426e-01 -3.57380629e-01 -2.15474084e-01
7.35066012e-02 -5.35020888e-01 -9.89257693e-01 2.86298431e-02
3.78767014e-01 1.06579408e-01 1.44768760e-01 1.05127907e+00
6.37873948e-01 -2.06950352e-01 5.12566209e-01 -1.24143314e+00
-3.17178190e-01 4.22307372e-01 -4.30316806e-01 3.85839194e-01
-2.59757400e-01 7.31208250e-02 1.06389809e+00 -5.33707201e-01
9.50461864e-01 4.53345567e-01 7.44853795e-01 7.53205001e-01
-1.53186989e+00 -3.75927866e-01 -1.96772814e-02 -6.50265068e-02
-1.53427577e+00 -3.03608447e-01 5.69294095e-01 -6.47906840e-01
5.44030368e-01 1.15188479e-01 7.12951720e-01 8.55349779e-01
5.23920879e-02 8.28414559e-01 1.43718088e+00 -3.45568717e-01
5.39749384e-01 2.11370811e-01 1.87198877e-01 6.68255568e-01
1.94862023e-01 -3.74716073e-01 -3.92171800e-01 -2.53264070e-01
8.38984430e-01 3.00918192e-01 -3.56692970e-01 -7.45543718e-01
-1.25508094e+00 6.05328023e-01 2.52001613e-01 7.06553280e-01
7.05956444e-02 1.12395227e-01 3.43672484e-01 5.24178408e-02
5.15554011e-01 2.73183912e-01 -6.94556177e-01 -4.55421177e-05
-9.31680083e-01 1.31950408e-01 6.65971875e-01 6.24295771e-01
9.73818243e-01 -5.07323802e-01 7.80114159e-02 6.03670537e-01
2.98709184e-01 1.17834821e-01 5.70400953e-01 -1.07676089e+00
-1.21167682e-01 7.84586549e-01 6.23256192e-02 -7.13163793e-01
-5.54995358e-01 -4.62945044e-01 -7.82755315e-01 2.64884561e-01
1.11012280e+00 -2.74697274e-01 -6.45817816e-01 1.90481222e+00
6.51596725e-01 2.73824900e-01 2.26558205e-02 9.22470629e-01
6.38234377e-01 2.73465663e-01 -5.53665496e-02 -1.63672388e-01
1.15587151e+00 -8.42388749e-01 -5.94134688e-01 7.69653395e-02
1.20722556e+00 -4.30463254e-01 7.67137229e-01 2.67800391e-01
-9.75335598e-01 -2.30784386e-01 -7.26424515e-01 -1.89763710e-01
-6.55556321e-01 1.14950299e-01 7.53399372e-01 2.84756750e-01
-1.26579356e+00 5.82646370e-01 -9.40383494e-01 -5.32727838e-01
1.02875245e+00 3.63408566e-01 -3.93536478e-01 2.24321052e-01
-6.80329561e-01 8.05445313e-01 3.03216815e-01 -3.89868021e-02
-6.79995477e-01 -8.85035574e-01 -4.57123220e-01 6.13040775e-02
3.06616783e-01 -8.75385046e-01 1.05271363e+00 -1.15590370e+00
-1.35567713e+00 1.39558291e+00 -2.11185943e-02 -3.19902122e-01
7.20081568e-01 2.23105460e-01 1.84023321e-01 1.94438353e-01
1.72701590e-02 8.35136354e-01 3.61432284e-01 -1.42223489e+00
-7.03015506e-01 -6.11216009e-01 -1.09541826e-01 -2.96271201e-02
-3.45262229e-01 -2.84519076e-01 -2.96392113e-01 -4.13071245e-01
-9.51487646e-02 -1.05205250e+00 -3.28995854e-01 3.56995642e-01
-3.12726438e-01 -2.22629905e-01 8.67594600e-01 -2.79480875e-01
8.29809129e-01 -2.07361531e+00 2.23577544e-01 1.52798951e-01
4.59128052e-01 2.17782840e-01 6.15148880e-02 2.43259668e-01
8.17011744e-02 1.03494510e-01 -4.67343956e-01 -4.59795326e-01
-2.37564087e-01 2.76604116e-01 -3.43350843e-02 7.21610188e-01
-2.04631891e-02 9.26731765e-01 -1.05569685e+00 -8.42595458e-01
-7.00165778e-02 5.61975360e-01 -3.37963432e-01 -4.24234681e-02
-1.97378099e-01 7.49143720e-01 -3.69212002e-01 6.51149094e-01
3.32710057e-01 -8.11221063e-01 4.37744468e-01 -2.65048802e-01
-1.79569408e-01 -2.48590007e-01 -9.87464547e-01 1.77642930e+00
-1.03909887e-01 5.28172553e-01 3.05731028e-01 -1.21966505e+00
5.91682911e-01 3.15328926e-01 8.52807879e-01 -3.40092272e-01
1.65608153e-01 3.88726085e-01 1.97495520e-02 -1.51115015e-01
-4.07283939e-02 -1.45464182e-01 1.29690975e-01 6.77153230e-01
2.86212564e-01 -3.28005143e-02 3.62655371e-01 3.59177977e-01
1.21809876e+00 2.21028283e-01 3.86950940e-01 -3.90064031e-01
4.20668334e-01 9.87866223e-02 8.80304515e-01 5.16390920e-01
-5.18142819e-01 7.68476605e-01 6.15938425e-01 -3.60625088e-01
-9.39536870e-01 -9.26486969e-01 -3.36612821e-01 8.99787962e-01
2.12320000e-01 -2.20159844e-01 -9.23409224e-01 -8.28388870e-01
6.53845072e-02 -2.36671278e-03 -8.29446018e-01 2.30068952e-01
-3.91808301e-01 -7.69543767e-01 6.20557964e-01 4.54393506e-01
1.92957312e-01 -8.39124322e-01 -7.94707000e-01 3.81369740e-02
-7.16962516e-02 -9.08507466e-01 -3.43531728e-01 4.01137173e-01
-8.89358640e-01 -1.30478120e+00 -9.04989481e-01 -9.71902072e-01
1.13157320e+00 6.11175783e-02 8.94413650e-01 4.51629966e-01
-4.92222399e-01 5.63979268e-01 -7.66095668e-02 -4.75513577e-01
-1.15139753e-01 1.35427520e-01 -2.24735066e-01 1.11284301e-01
2.56412864e-01 -5.10627687e-01 -6.61633670e-01 2.82088608e-01
-1.01179111e+00 2.75283098e-01 3.99666905e-01 9.98783052e-01
1.23029995e+00 -1.23802878e-01 8.18336904e-01 -1.33811128e+00
-5.76406419e-02 -4.34116930e-01 -7.65443504e-01 2.55712807e-01
-5.83711207e-01 -3.61444265e-01 6.83907032e-01 -2.34406397e-01
-1.01957750e+00 3.37564766e-01 8.46928433e-02 -6.88409135e-02
-3.65866065e-01 4.10627782e-01 -3.51726823e-02 -2.55495995e-01
4.96132642e-01 2.30203912e-01 3.77540201e-01 -2.08492965e-01
9.52663273e-02 4.50096279e-01 4.33598310e-01 -5.59409618e-01
5.26679218e-01 9.88712132e-01 1.52520314e-01 -6.19582713e-01
-9.72109258e-01 -6.69381440e-01 -7.98228383e-01 -2.27782771e-01
7.97990680e-01 -7.15567589e-01 -7.56234646e-01 7.92207837e-01
-9.41331029e-01 -8.61685455e-01 -4.10719335e-01 2.49945238e-01
-9.42618430e-01 2.88904965e-01 -8.16247344e-01 -4.67503041e-01
-2.08553150e-01 -8.93279910e-01 1.02282393e+00 4.53727335e-01
-7.07560703e-02 -1.48381710e+00 4.45347071e-01 5.64467728e-01
2.98261374e-01 2.88360059e-01 9.44635332e-01 -9.47207153e-01
-7.09861100e-01 -3.87298875e-02 -1.12938032e-01 1.40547559e-01
2.13531822e-01 6.72840178e-02 -1.04177523e+00 -2.26806611e-01
-2.65439332e-01 -4.44104284e-01 7.81591356e-01 4.79862332e-01
1.21602058e+00 8.33512619e-02 -7.17720270e-01 6.31091595e-01
1.56773901e+00 -9.29743871e-02 3.79821658e-01 2.07063526e-01
8.00305963e-01 6.24557376e-01 4.45153922e-01 3.04392606e-01
4.56952065e-01 6.09531641e-01 1.96568489e-01 -5.00196576e-01
-1.33224905e-01 1.13833793e-01 -2.77001798e-01 7.24883258e-01
-1.22029632e-01 -1.74296081e-01 -1.08197784e+00 6.85279608e-01
-2.34739876e+00 -9.41142797e-01 -5.15951542e-04 1.93733871e+00
1.20262992e+00 -1.96390137e-01 1.54171899e-01 4.24007736e-02
5.83545566e-01 -1.97907194e-01 -7.55066454e-01 2.29992047e-01
-2.85878927e-01 -1.45999804e-01 3.83883297e-01 2.44924605e-01
-1.10300303e+00 7.32157648e-01 5.48084927e+00 9.05447125e-01
-1.06436849e+00 3.31876218e-01 1.05823123e+00 -1.99903429e-01
-1.58375606e-01 9.66752693e-02 -8.12554955e-01 6.63560212e-01
6.43321216e-01 -1.13993146e-01 1.81855232e-01 5.93183637e-01
5.17521054e-02 -3.18590999e-01 -1.32482398e+00 8.74462783e-01
-5.85936718e-02 -1.85810697e+00 -3.11948031e-01 2.41389781e-01
8.44483197e-01 2.69212097e-01 -8.55133682e-02 -3.08851991e-02
4.12403494e-01 -7.52025247e-01 3.58048886e-01 8.58106852e-01
4.19403553e-01 -4.84354347e-01 8.72617066e-01 5.29469311e-01
-7.83612847e-01 1.36261463e-01 -1.26309067e-01 1.38837829e-01
7.72458315e-02 7.54966319e-01 -5.76395512e-01 2.37635866e-01
8.04988086e-01 8.04350257e-01 -5.29849827e-01 1.16219246e+00
1.70378044e-01 6.83723688e-01 -4.24960434e-01 1.28542304e-01
2.85809115e-02 -2.77772456e-01 1.89175993e-01 1.29210508e+00
1.17862143e-01 6.34351969e-02 1.39607519e-01 9.93514478e-01
-1.60353795e-01 8.70392174e-02 -4.36634928e-01 2.42698528e-02
3.19860876e-01 1.80148637e+00 -1.41431022e+00 -2.85295635e-01
-3.72821629e-01 6.26529038e-01 6.76781058e-01 4.16530967e-01
-7.29767680e-01 -1.29116133e-01 4.23858851e-01 8.49807784e-02
3.34759593e-01 7.15295076e-02 -4.70609486e-01 -9.89444435e-01
-3.29322278e-01 -4.74754930e-01 4.00342017e-01 -5.11716783e-01
-1.54054415e+00 2.78057512e-02 -2.71995425e-01 -9.33189213e-01
2.52240330e-01 -6.03342950e-01 -5.39134860e-01 6.14174902e-01
-1.58148849e+00 -1.25893629e+00 -3.43753844e-01 4.21670139e-01
1.36429176e-01 1.00460522e-01 8.27385485e-01 2.39118889e-01
-7.52566516e-01 3.93777072e-01 3.03294599e-01 1.87640190e-01
7.84832776e-01 -1.48329544e+00 -2.64780045e-01 4.40711200e-01
1.80163160e-01 6.13987505e-01 3.32179815e-01 -4.69409019e-01
-1.25088251e+00 -1.01456392e+00 7.98014581e-01 -6.98527575e-01
7.13802218e-01 -3.42970908e-01 -1.03981209e+00 6.38172328e-01
-5.80812283e-02 6.89486504e-01 1.22898412e+00 -4.52574417e-02
-2.82074697e-02 -6.03679717e-02 -1.20675242e+00 4.85093653e-01
8.64318609e-01 -4.26979661e-01 -6.56015426e-02 5.11613131e-01
1.53879955e-01 -3.98992747e-01 -1.09037912e+00 -1.48874195e-02
3.22811872e-01 -9.23230469e-01 6.72147453e-01 -3.16486269e-01
2.59810746e-01 -6.04173005e-01 1.04703173e-01 -1.16636753e+00
-1.91005141e-01 -4.10507590e-01 -7.10253971e-06 1.39791930e+00
3.70403916e-01 -8.01624417e-01 1.04043651e+00 5.27908504e-01
-2.57612526e-01 -1.08617783e+00 -1.00462294e+00 -5.97584665e-01
3.47045183e-01 1.42192796e-01 2.66220570e-01 1.17235434e+00
2.19506249e-01 2.99399525e-01 1.54214159e-01 -6.82495311e-02
5.95084012e-01 2.49579772e-01 8.18342328e-01 -1.33122504e+00
-3.48119557e-01 -5.35778403e-01 -3.86560529e-01 -7.71964312e-01
2.64140189e-01 -1.10712194e+00 1.22273602e-01 -1.44737065e+00
5.43650925e-01 -6.95727766e-01 -5.37187457e-01 7.32464373e-01
1.20954746e-02 4.62957323e-01 -4.41822223e-02 4.33172852e-01
-1.09182084e+00 1.50462627e-01 1.27859104e+00 -8.15237463e-02
1.60322152e-03 -2.30692774e-01 -6.94094658e-01 9.47463870e-01
8.08360219e-01 -4.27159935e-01 -3.24688911e-01 -3.32244068e-01
4.39541668e-01 -1.45036340e-01 6.31160200e-01 -8.01496625e-01
7.21942604e-01 -9.33320224e-02 2.91110963e-01 -3.32178921e-01
8.12846199e-02 -8.64190757e-01 2.27298126e-01 2.42464796e-01
-4.65557337e-01 -4.35195804e-01 -1.87090766e-02 6.60282016e-01
-1.32778510e-01 -1.95313141e-01 8.55968714e-01 -3.25992823e-01
-4.13996667e-01 3.14903259e-01 -2.74247378e-01 2.02787071e-01
1.19756722e+00 -4.23418105e-01 -6.39886200e-01 1.15113303e-01
-6.30067348e-01 1.74807236e-01 8.17206025e-01 -2.50536561e-01
2.12002844e-01 -1.02439463e+00 -3.72848898e-01 -9.04953927e-02
2.20962018e-01 4.34093088e-01 3.86760831e-01 1.32134628e+00
-3.37931216e-01 2.19311193e-01 -1.72499076e-01 -8.01167488e-01
-1.04502392e+00 3.01695108e-01 5.16004562e-01 -3.74323249e-01
-4.69281167e-01 8.33819926e-01 3.86471927e-01 -5.56644917e-01
-2.34644711e-02 -1.10133313e-01 -1.53788134e-01 -8.31228495e-03
2.91873068e-01 4.44339514e-01 3.46343704e-02 -6.60289764e-01
-3.73318434e-01 5.92268825e-01 -2.91483939e-01 3.17732334e-01
1.35632288e+00 -1.23289905e-01 -2.91358560e-01 7.32029796e-01
1.06689370e+00 -1.52686715e-01 -1.59938717e+00 -2.67125368e-01
1.70905516e-01 -9.37332362e-02 1.13190681e-01 -8.61406326e-01
-1.23200297e+00 5.22731662e-01 5.18870294e-01 1.91295251e-01
8.13305497e-01 2.23636121e-01 7.38955319e-01 4.01563823e-01
4.51637894e-01 -1.30494642e+00 1.14367139e-02 3.44417244e-01
2.82500058e-01 -1.48665547e+00 -1.16420560e-01 -3.20523500e-01
-3.25570226e-01 7.97039211e-01 6.99990571e-01 -5.35515994e-02
7.22287595e-01 5.99344492e-01 3.62206787e-01 -3.66806239e-01
-7.71059036e-01 -1.24967828e-01 -2.48762384e-01 7.07121909e-01
5.24118066e-01 -1.01899445e-01 -1.85901016e-01 7.93262959e-01
2.04189464e-01 3.47743571e-01 4.53231931e-01 1.00692475e+00
-2.86434203e-01 -1.03483021e+00 3.57204638e-02 4.82344300e-01
-4.98205006e-01 1.92388982e-01 -3.92735153e-01 5.94342053e-01
2.12576002e-01 5.15689552e-01 2.34672621e-01 2.38030046e-01
-1.59247965e-01 2.30430365e-02 4.87451196e-01 -7.02788532e-01
-5.03845036e-01 1.03597358e-01 -4.50809717e-01 -4.22354490e-01
-7.41770208e-01 -8.51477981e-01 -1.77901196e+00 4.62378785e-02
-3.43086869e-01 1.73476830e-01 4.99880373e-01 1.03311849e+00
4.84176487e-01 3.47886205e-01 3.39713931e-01 -7.47446895e-01
-1.68650880e-01 -5.27308881e-01 -8.03374469e-01 5.07179976e-01
1.76298752e-01 -4.83284384e-01 -4.98123020e-01 4.30452704e-01] | [14.549227714538574, -3.0220470428466797] |
bc051876-493e-47a9-8340-224ca5c5b3e9 | falsification-before-extrapolation-in-causal | 2209.13708 | null | https://arxiv.org/abs/2209.13708v3 | https://arxiv.org/pdf/2209.13708v3.pdf | Falsification before Extrapolation in Causal Effect Estimation | Randomized Controlled Trials (RCTs) represent a gold standard when developing policy guidelines. However, RCTs are often narrow, and lack data on broader populations of interest. Causal effects in these populations are often estimated using observational datasets, which may suffer from unobserved confounding and selection bias. Given a set of observational estimates (e.g. from multiple studies), we propose a meta-algorithm that attempts to reject observational estimates that are biased. We do so using validation effects, causal effects that can be inferred from both RCT and observational data. After rejecting estimators that do not pass this test, we generate conservative confidence intervals on the extrapolated causal effects for subgroups not observed in the RCT. Under the assumption that at least one observational estimator is asymptotically normal and consistent for both the validation and extrapolated effects, we provide guarantees on the coverage probability of the intervals output by our algorithm. To facilitate hypothesis testing in settings where causal effect transportation across datasets is necessary, we give conditions under which a doubly-robust estimator of group average treatment effects is asymptotically normal, even when flexible machine learning methods are used for estimation of nuisance parameters. We illustrate the properties of our approach on semi-synthetic and real world datasets, and show that it compares favorably to standard meta-analysis techniques. | ['David Sontag', 'Ming-Chieh Shih', 'Michael Oberst', 'Zeshan Hussain'] | 2022-09-27 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [ 4.48489398e-01 1.48730725e-01 -1.28747046e+00 -3.66099894e-01
-9.62345660e-01 -5.40494919e-01 4.87573296e-01 6.32210314e-01
-4.78393495e-01 1.20558953e+00 5.05190909e-01 -8.82838666e-01
-4.57436293e-01 -7.09675491e-01 -1.01458430e+00 -5.59627235e-01
-5.28815687e-01 3.32837462e-01 -2.64663041e-01 6.51222706e-01
2.78115213e-01 2.66769886e-01 -1.31808555e+00 1.92605779e-01
1.28635609e+00 3.71652871e-01 -3.91498357e-01 3.72771055e-01
3.45171332e-01 4.48292047e-01 -5.01019120e-01 -9.21655148e-02
-9.95340422e-02 -6.24928534e-01 -3.60580534e-01 -1.89168915e-01
4.05718416e-01 -3.39545548e-01 1.91117987e-01 1.01139116e+00
5.40927649e-01 -2.38417163e-01 1.29884160e+00 -1.50003231e+00
-5.40354788e-01 1.01103580e+00 -8.59804273e-01 -1.41100228e-01
2.38762140e-01 1.04512788e-01 7.25186944e-01 -6.62757695e-01
4.77623761e-01 1.34618139e+00 6.46644533e-01 3.11463326e-01
-1.67542076e+00 -1.00565112e+00 1.11115374e-01 -3.39174777e-01
-1.07519782e+00 -4.63424057e-01 7.69082904e-02 -5.69992959e-01
3.48088533e-01 3.05955172e-01 4.72901762e-01 1.29800105e+00
6.33508980e-01 5.41023985e-02 1.28918314e+00 -5.33763766e-01
7.36166477e-01 -7.14314058e-02 7.68977776e-02 2.05347240e-01
1.00264668e+00 6.67424738e-01 -2.24820495e-01 -9.93486881e-01
6.57712817e-01 1.63630441e-01 -2.67492384e-01 -4.24446791e-01
-1.42275226e+00 1.32310140e+00 1.35068297e-01 -2.04248562e-01
-7.11545467e-01 2.61746552e-02 5.46761572e-01 1.41026333e-01
5.14768004e-01 1.27326876e-01 -5.78479826e-01 3.39235753e-01
-8.48317683e-01 4.39326704e-01 7.30612695e-01 7.15270042e-01
4.22613248e-02 -3.41362894e-01 -2.37220436e-01 7.28108108e-01
6.79720044e-02 9.25278783e-01 3.60197335e-01 -1.02805293e+00
3.70171010e-01 3.49449098e-01 5.92917204e-01 -4.28526849e-01
-4.69097883e-01 1.20877527e-01 -1.11715722e+00 1.77169487e-01
5.10860324e-01 -5.05091965e-01 -8.68836999e-01 2.15247869e+00
5.51862717e-01 -3.32300030e-02 -4.92815673e-02 6.56799972e-01
5.47129571e-01 2.32797608e-01 6.13805234e-01 -9.80693340e-01
1.18858206e+00 -1.04249813e-01 -8.70064795e-01 1.46265745e-01
1.02650285e+00 -4.24539804e-01 8.86063159e-01 2.48192459e-01
-1.35910094e+00 -4.49383073e-02 -7.59461761e-01 4.03441995e-01
-1.82506219e-02 -1.36808529e-01 3.91146511e-01 7.21224546e-01
-4.57984179e-01 4.47033316e-01 -5.22999823e-01 -1.28246412e-01
4.81806099e-01 4.86141354e-01 -3.70018542e-01 -2.97220442e-02
-1.08429658e+00 6.85005963e-01 2.28108928e-01 -9.37222913e-02
-8.45349729e-01 -1.29605174e+00 -7.27065325e-01 1.73660755e-01
6.28179967e-01 -9.93399560e-01 1.21719158e+00 -8.94932449e-01
-7.51040876e-01 4.97754127e-01 -3.34549367e-01 -4.65849608e-01
6.14893079e-01 1.68765694e-01 -4.16754186e-01 -1.14660207e-02
4.47426856e-01 2.36720756e-01 6.14081860e-01 -9.55276668e-01
-9.11661446e-01 -6.31837726e-01 -3.89802247e-01 -8.44492950e-03
4.97423112e-02 3.58216316e-01 2.32427403e-01 -6.65237784e-01
-7.03870803e-02 -9.26399827e-01 -6.41873002e-01 -2.04392090e-01
-5.24236739e-01 -2.77017802e-01 1.40125588e-01 -3.15573722e-01
1.43478465e+00 -1.83869565e+00 -2.99447298e-01 3.25607210e-01
1.97398305e-01 -2.70634145e-01 2.40654405e-02 3.67704690e-01
-5.32431543e-01 5.76658845e-01 -5.00944197e-01 3.07670355e-01
-2.31181204e-01 -1.57160059e-01 -1.26885384e-01 1.11605012e+00
6.11103401e-02 5.35966396e-01 -6.76616609e-01 -5.76248050e-01
1.80686742e-01 -1.01383477e-01 -6.81311309e-01 1.49257913e-01
-8.92930478e-02 5.03042579e-01 -4.71363097e-01 8.32903087e-02
7.94766665e-01 -2.64676601e-01 5.43331206e-01 1.68053389e-01
-3.36421758e-01 4.75388616e-01 -1.35210288e+00 9.13794696e-01
-4.89794463e-01 6.93684667e-02 -1.27214178e-01 -1.38424778e+00
4.00055468e-01 5.64770103e-01 3.77994865e-01 -3.61040473e-01
1.81420758e-01 3.78845692e-01 2.93827057e-01 -5.30942321e-01
-3.58800799e-01 -6.62934780e-01 -2.45098099e-01 5.89693189e-01
-3.33358556e-01 1.69719294e-01 -8.25338885e-02 -1.09259255e-01
1.04531872e+00 -5.17902911e-01 8.13469231e-01 -4.82437670e-01
1.86582878e-01 3.40340263e-03 6.42523885e-01 1.28005862e+00
2.11223125e-01 5.16275346e-01 8.20938170e-01 -2.89964050e-01
-1.16911399e+00 -1.20433426e+00 -7.30459511e-01 5.07850528e-01
-3.73869926e-01 7.32601359e-02 -4.02090967e-01 -7.42746055e-01
3.17326397e-01 7.70964205e-01 -1.04333961e+00 -3.02506417e-01
-1.77745700e-01 -1.25771713e+00 3.89723003e-01 5.50543964e-01
-1.33504108e-01 -7.11872399e-01 -5.45503914e-01 2.28229702e-01
1.22337177e-01 -6.03593767e-01 -3.07800174e-01 7.03828037e-02
-1.20529985e+00 -1.51540732e+00 -9.07561600e-01 -4.53777462e-01
6.35148048e-01 4.74575981e-02 9.76951778e-01 -1.01655900e-01
1.97586536e-01 -1.95466485e-02 7.41181150e-02 -6.74785078e-01
-7.06604660e-01 -3.30526620e-01 3.06082159e-01 -3.38730633e-01
4.03880388e-01 -3.22444856e-01 -8.21618140e-01 2.23811299e-01
-7.78286278e-01 -2.66965002e-01 4.53463912e-01 9.91324127e-01
4.97192979e-01 -1.45776868e-01 1.25022578e+00 -1.12618101e+00
4.61050212e-01 -9.71590042e-01 -9.80151951e-01 2.63938844e-01
-8.52756143e-01 -4.53869440e-03 4.90236789e-01 -8.38230908e-01
-1.03448904e+00 -5.38817227e-01 4.44509625e-01 -1.61131918e-01
-2.71475196e-01 9.14249361e-01 -1.73303232e-01 5.21734655e-01
8.74172807e-01 -6.08907521e-01 1.69290677e-01 -3.12422276e-01
1.06152244e-01 1.02152336e+00 9.98378694e-02 -5.81071436e-01
4.49447567e-03 5.50909877e-01 3.20894033e-01 -5.88623285e-01
-7.35979915e-01 -2.04296976e-01 -4.82319929e-02 3.34590971e-01
6.34138465e-01 -9.94764864e-01 -1.19907820e+00 -2.70497531e-01
-7.71156371e-01 -4.50183600e-01 -2.04487443e-01 1.32714641e+00
-6.27789557e-01 -1.00592576e-01 -7.56942108e-02 -1.15466356e+00
-9.79597718e-02 -1.10879278e+00 8.08059335e-01 -5.82172237e-02
-6.66187704e-01 -1.22570264e+00 2.20035672e-01 -2.23230168e-01
-2.28761688e-01 4.98036951e-01 1.26469302e+00 -7.55083621e-01
1.86985478e-01 -3.23689669e-01 -1.47383422e-01 -1.60776541e-01
4.23725545e-01 3.37467551e-01 -5.29494643e-01 -2.56702751e-01
-3.83342169e-02 -1.38013616e-01 5.57760835e-01 1.64108729e+00
1.51365244e+00 -7.94647455e-01 -7.26375461e-01 1.30851284e-01
1.29854167e+00 3.71527910e-01 2.64754295e-01 -1.95192937e-02
2.78230548e-01 1.01833332e+00 6.67942226e-01 5.24291635e-01
1.78043902e-01 5.20738721e-01 -2.30443273e-02 -1.12550810e-01
4.84068871e-01 -1.44840181e-01 5.32308705e-02 -1.19840428e-01
3.45197730e-02 -1.95286676e-01 -6.77310526e-01 8.41912985e-01
-1.88241982e+00 -9.07242894e-01 -7.28101492e-01 3.22915316e+00
1.06518006e+00 2.38856748e-01 6.34350657e-01 -1.58545107e-01
1.10699677e+00 -5.56525290e-01 -7.06305742e-01 -8.50127935e-01
-1.01674885e-01 2.68811971e-01 9.91450429e-01 3.93506974e-01
-8.18382084e-01 1.13312085e-03 7.49800110e+00 4.99513060e-01
-7.68400609e-01 4.52342592e-02 7.50091672e-01 -9.02186111e-02
-3.96781653e-01 2.07782075e-01 -3.92551214e-01 5.13799727e-01
1.57450593e+00 -5.34505308e-01 -2.81615436e-01 3.09585601e-01
1.10937011e+00 -4.99198228e-01 -1.32519853e+00 2.09097981e-01
-5.98630548e-01 -1.28273928e+00 -2.20773354e-01 3.83478582e-01
1.04846025e+00 -2.42002562e-01 8.32283422e-02 -9.89072770e-02
7.71930993e-01 -1.18753672e+00 5.46713948e-01 2.54172292e-02
1.14953232e+00 -7.55716503e-01 7.78670371e-01 4.41101730e-01
-3.64040762e-01 -2.08478883e-01 -4.46749061e-01 -1.40750796e-01
-9.73920301e-02 1.09871781e+00 -6.79798305e-01 4.45036173e-01
5.31026721e-01 6.32184982e-01 1.05106115e-01 1.23461378e+00
-2.30584562e-01 1.03361571e+00 -2.78638750e-01 1.59049883e-01
-7.53472149e-02 -1.97658643e-01 2.43283629e-01 8.57927263e-01
4.84198570e-01 1.58300951e-01 -2.41175830e-01 8.89752269e-01
-1.42444223e-01 2.13812679e-01 -9.17392552e-01 1.05130889e-01
8.05258095e-01 5.14946163e-01 -4.87092197e-01 -5.49543738e-01
-7.76117980e-01 2.22980678e-02 -6.00325465e-02 4.99455422e-01
-7.27581859e-01 2.26847634e-01 3.30240428e-01 8.47131759e-02
6.74154609e-02 5.56582272e-01 -6.65935218e-01 -9.30478573e-01
-2.14299366e-01 -1.13873816e+00 9.70524490e-01 -5.08012533e-01
-1.26969492e+00 -4.30852383e-01 6.43534243e-01 -1.11980355e+00
-3.59999239e-01 -1.68406054e-01 -4.78256494e-01 1.21715629e+00
-8.14275682e-01 -5.26230335e-01 3.80374640e-01 1.45522326e-01
3.46564978e-01 4.44299310e-01 7.57040799e-01 -1.71259418e-01
-6.32519305e-01 4.06787485e-01 2.90462941e-01 -4.26447600e-01
9.43490565e-01 -1.19163930e+00 -1.29221082e-01 3.22532117e-01
-5.60851693e-01 7.15175331e-01 8.54747176e-01 -1.20092285e+00
-1.01660299e+00 -9.22602236e-01 7.39540398e-01 -1.46571234e-01
8.24078381e-01 5.64617012e-03 -8.26622069e-01 8.78206730e-01
-1.23978436e-01 -1.70033157e-01 7.90410221e-01 6.78609610e-01
-3.13004643e-01 2.02058047e-01 -1.13465583e+00 8.75693560e-01
7.68939018e-01 8.31367448e-02 -6.86318576e-01 2.67718822e-01
4.86254930e-01 1.87244900e-02 -1.16019297e+00 7.77016580e-01
9.80084300e-01 -6.42899215e-01 8.10011804e-01 -1.25953507e+00
6.26219809e-01 1.23988621e-01 1.94644868e-01 -1.39284313e+00
1.89974830e-02 -3.20042729e-01 2.53847569e-01 8.63477826e-01
7.01152086e-01 -7.31249750e-01 5.24320006e-01 6.79794312e-01
3.64458352e-01 -3.49360973e-01 -8.77717137e-01 -8.19786131e-01
6.61646068e-01 -3.11028957e-01 6.49353445e-01 1.13251507e+00
1.72546148e-01 3.37800264e-01 -2.20456764e-01 4.04510558e-01
1.01991355e+00 4.10315454e-01 5.31822145e-01 -1.32612932e+00
1.29169717e-01 -3.29296201e-01 1.83155343e-01 -7.62903914e-02
2.22002625e-01 -3.15984398e-01 -1.43876925e-01 -1.34578121e+00
7.30325937e-01 -5.80990076e-01 -8.93278345e-02 2.71888345e-01
-5.65498173e-01 -3.30382288e-01 -5.40095270e-01 1.67259052e-02
1.36556953e-01 3.95101219e-01 1.07088947e+00 9.48936865e-02
-4.72447842e-01 3.98710102e-01 -7.97633529e-01 9.10501301e-01
8.44844043e-01 -1.13158250e+00 -4.92194176e-01 2.21135870e-01
2.07354501e-01 6.11761153e-01 4.54629809e-01 -1.77525371e-01
-1.93868786e-01 -7.49793887e-01 3.50836158e-01 -5.02122819e-01
-5.38048089e-01 -7.75365770e-01 5.20542324e-01 8.58733416e-01
-8.77811611e-01 9.00992453e-02 1.69769570e-01 6.26817465e-01
1.26213998e-01 -2.66341865e-01 5.63967586e-01 1.59381002e-01
3.21533263e-01 1.40809581e-01 -5.75168133e-01 1.93555579e-01
6.93816900e-01 1.40447259e-01 -4.24166650e-01 -4.14292783e-01
-6.51982725e-01 3.78953516e-01 4.12223041e-01 7.24920854e-02
3.67480963e-01 -1.27664399e+00 -1.05868077e+00 -1.49663851e-01
3.47566307e-01 -3.10140103e-01 3.45447302e-01 1.38284445e+00
1.58608958e-01 3.69331062e-01 2.34985650e-01 -4.86233324e-01
-1.14339542e+00 1.12635171e+00 1.20377935e-01 -2.06381362e-02
-5.35719097e-01 -8.76752883e-02 8.78490865e-01 -2.86914557e-01
1.36521444e-01 -5.65321326e-01 -3.98685485e-02 4.51860726e-02
7.35725701e-01 5.95357418e-01 -5.41137867e-02 -9.52721611e-02
-3.79113972e-01 1.50522217e-01 2.15994880e-01 -4.16033059e-01
1.25937080e+00 -2.26885915e-01 -1.23740129e-01 7.48970449e-01
1.20785153e+00 2.74881572e-01 -9.31814790e-01 9.33016166e-02
6.56103529e-03 -2.98182547e-01 -3.74629465e-03 -7.52147257e-01
-4.29987460e-01 5.25918901e-01 4.59901094e-01 4.57577437e-01
9.93774414e-01 -7.09827244e-02 -3.19640517e-01 -2.94438213e-01
8.19580927e-02 -7.67967284e-01 -7.71893382e-01 -5.67698956e-01
8.29607129e-01 -1.40723920e+00 3.79283935e-01 -4.71954763e-01
-3.50346118e-01 6.32937908e-01 9.68324579e-03 -1.29696146e-01
6.36713564e-01 1.70050323e-01 -4.04667467e-01 -1.22709058e-01
-9.93566692e-01 1.38813630e-01 3.52929711e-01 6.15208924e-01
5.40010810e-01 4.48129177e-01 -1.22620106e+00 5.71522176e-01
1.08641930e-01 4.39490676e-01 7.78734326e-01 7.58679032e-01
-5.64453378e-02 -8.08461845e-01 -7.43953407e-01 9.80751693e-01
-9.64819968e-01 -1.11462101e-01 -4.33522612e-02 1.20772910e+00
-2.68625677e-01 1.18460703e+00 2.32253373e-01 4.61323708e-01
4.63618666e-01 -2.63093829e-01 3.93777788e-01 -5.36774516e-01
-1.52365372e-01 3.82563561e-01 4.18128997e-01 -3.31606269e-01
-8.51286530e-01 -8.48475099e-01 -9.30200934e-01 -5.46123743e-01
-6.07093513e-01 4.84350055e-01 4.53389853e-01 9.77451503e-01
8.67705271e-02 3.04261923e-01 9.07586277e-01 -6.86295554e-02
-7.76539385e-01 -8.54823589e-01 -6.42744303e-01 1.37739465e-01
6.18009567e-01 -8.18893671e-01 -7.42928267e-01 -1.04648449e-01] | [7.9599385261535645, 5.274258136749268] |
35fe9e2e-2421-4f54-a5c0-f71cb20feb70 | learning-augmented-algorithms-for-online | 2112.05353 | null | https://arxiv.org/abs/2112.05353v2 | https://arxiv.org/pdf/2112.05353v2.pdf | Learning-Augmented Algorithms for Online Steiner Tree | This paper considers the recently popular beyond-worst-case algorithm analysis model which integrates machine-learned predictions with online algorithm design. We consider the online Steiner tree problem in this model for both directed and undirected graphs. Steiner tree is known to have strong lower bounds in the online setting and any algorithm's worst-case guarantee is far from desirable. This paper considers algorithms that predict which terminal arrives online. The predictions may be incorrect and the algorithms' performance is parameterized by the number of incorrectly predicted terminals. These guarantees ensure that algorithms break through the online lower bounds with good predictions and the competitive ratio gracefully degrades as the prediction error grows. We then observe that the theory is predictive of what will occur empirically. We show on graphs where terminals are drawn from a distribution, the new online algorithms have strong performance even with modestly correct predictions. | ['Benjamin Moseley', 'Chenyang Xu'] | 2021-12-10 | null | null | null | null | ['steiner-tree-problem'] | ['graphs'] | [ 1.16113424e-01 6.32321954e-01 -8.85805905e-01 -2.16173992e-01
-7.38025665e-01 -1.11557078e+00 -3.36028516e-01 3.84735465e-01
-4.01676707e-02 7.48490810e-01 -3.27576280e-01 -9.21038032e-01
-6.51490510e-01 -7.10008323e-01 -1.06417847e+00 -5.52414477e-01
-5.59992671e-01 1.23319948e+00 4.57331419e-01 -5.08661792e-02
4.58533019e-02 6.93725586e-01 -1.07254875e+00 -2.83372253e-02
7.64270127e-01 9.24171388e-01 -1.11162968e-01 1.38212860e+00
-2.81031400e-01 3.33912551e-01 -1.93291366e-01 -9.00244534e-01
8.81471694e-01 -3.91558141e-01 -8.75266194e-01 1.04513876e-01
2.75408864e-01 -2.07925111e-01 -4.60521549e-01 1.00456822e+00
2.34875865e-02 -4.75802690e-01 2.77102470e-01 -1.89952135e+00
-3.14373255e-01 1.14187682e+00 -5.95455468e-01 4.56799150e-01
3.10535401e-01 -5.07085323e-02 1.52801943e+00 -1.29322588e-01
7.53396630e-01 8.02208185e-01 8.51463795e-01 3.74518752e-01
-1.29888439e+00 -2.85053104e-01 4.83640313e-01 1.47906199e-01
-1.19863164e+00 -9.64758247e-02 3.18005204e-01 -9.71853361e-02
8.09531569e-01 6.03748441e-01 4.21661198e-01 4.87888128e-01
3.27689767e-01 8.62084389e-01 3.58018696e-01 -4.20472264e-01
2.35029519e-01 2.52201036e-02 4.69111092e-02 7.32663810e-01
7.36418068e-01 4.82491314e-01 -3.67833346e-01 -3.42148811e-01
3.97627890e-01 1.45178929e-01 -2.56014347e-01 -6.83938563e-01
-7.09330440e-01 5.69517434e-01 2.34229684e-01 -5.44361956e-03
-7.51181766e-02 2.92529821e-01 1.32100448e-01 8.67536187e-01
4.08025712e-01 1.03989750e-01 -1.15099454e+00 -1.78508237e-01
-9.24751759e-01 1.37620434e-01 1.49187207e+00 1.83582211e+00
3.42282861e-01 -3.18242043e-01 9.45774466e-02 2.39071906e-01
-1.53136432e-01 4.99138743e-01 -2.64639646e-01 -9.76678610e-01
8.95815253e-01 3.52322340e-01 7.71703064e-01 -8.95034969e-01
-5.41031182e-01 -8.99494588e-01 -1.34505183e-01 -6.05965294e-02
9.00097191e-01 -2.57402539e-01 -5.32119751e-01 1.56854308e+00
4.59951043e-01 -2.73706347e-01 -9.43108872e-02 7.53819466e-01
-5.25204837e-01 6.63632393e-01 -3.50402415e-01 -8.83666635e-01
3.17035079e-01 -1.15400791e+00 -4.97267991e-01 -1.28808424e-01
1.21439862e+00 -4.96039301e-01 4.54892009e-01 4.67854738e-01
-1.33365917e+00 1.51815966e-01 -7.95979679e-01 4.53110367e-01
-1.72048528e-02 -5.57949543e-01 7.28926778e-01 9.01991963e-01
-1.26800644e+00 9.07117963e-01 -9.45087314e-01 -4.99210864e-01
2.21124813e-01 4.74553764e-01 -8.78791418e-03 -1.42748296e-01
-2.99098134e-01 5.11941016e-01 2.59267628e-01 -1.77595064e-01
-5.33172607e-01 -4.66651410e-01 -2.33082876e-01 2.63072252e-01
7.16672301e-01 -7.43096888e-01 1.64577746e+00 -1.35792077e+00
-8.62880170e-01 4.47120249e-01 -4.02369648e-01 -8.14067245e-01
9.09059346e-01 3.32926333e-01 -2.42413312e-01 1.66810498e-01
-1.65024757e-01 -1.70630172e-01 2.63849616e-01 -1.07359600e+00
-1.20751238e+00 -7.10618854e-01 3.12722027e-01 1.16939440e-01
-2.63963133e-01 -4.21435148e-01 -1.32601202e-01 -2.92325728e-02
4.59088475e-01 -1.01450944e+00 -5.58364332e-01 4.23088014e-01
-3.39477867e-01 -1.79046273e-01 2.85605073e-01 -1.53286770e-01
1.28163421e+00 -1.80716705e+00 -4.09373552e-01 4.90011871e-01
2.78473228e-01 -4.06221300e-01 -1.33506998e-01 9.81728196e-01
1.54994205e-01 5.88636220e-01 3.77294719e-01 -2.00133882e-02
2.57092774e-01 3.51825327e-01 -2.97901839e-01 7.48026669e-01
-6.25505924e-01 5.27483165e-01 -7.95898795e-01 -1.61867350e-01
-2.64630675e-01 -7.19359577e-01 -8.05681407e-01 -1.18125856e-01
-5.40291250e-01 -2.24353835e-01 -4.57954466e-01 5.55408299e-01
8.10840428e-01 -6.83185399e-01 5.49693525e-01 7.16563582e-01
3.01773194e-02 6.03262037e-02 -9.12460804e-01 1.13880157e+00
-4.29703921e-01 3.28973621e-01 3.35423887e-01 -7.24012911e-01
3.63384157e-01 5.50280921e-02 4.36288595e-01 -2.25515425e-01
9.49754417e-02 4.49680746e-01 9.58142132e-02 -3.63013625e-01
7.84640089e-02 -4.18687984e-03 1.65059194e-01 4.52257246e-01
-4.96909589e-01 5.31902611e-01 1.34650499e-01 4.67962861e-01
1.62876689e+00 -3.24919403e-01 -8.23508799e-02 -1.35065034e-01
-2.51206726e-01 4.97668654e-01 7.45211065e-01 1.25950503e+00
-2.57681042e-01 2.56344110e-01 8.98719013e-01 -4.50238734e-01
-1.26898348e+00 -1.14589834e+00 2.48536348e-01 1.33278060e+00
5.02574205e-01 -3.00146073e-01 -6.05674863e-01 -7.00750589e-01
5.14128327e-01 6.93438113e-01 -6.66645110e-01 1.72526971e-01
3.10897678e-01 -1.38243839e-01 -9.00429636e-02 5.65371752e-01
-2.94472903e-01 1.96522232e-02 -8.59088302e-02 3.54798794e-01
2.88210094e-01 -1.26884341e+00 -6.81373239e-01 1.88335449e-01
-1.16780448e+00 -1.38434649e+00 -8.81510153e-02 -7.61499465e-01
8.33802700e-01 6.29725397e-01 8.94390404e-01 3.85998100e-01
7.98438713e-02 4.55041111e-01 -5.88570833e-01 -2.73495436e-01
-4.25216913e-01 2.17519641e-01 3.07537545e-03 -4.89979684e-01
8.60503539e-02 -6.98759377e-01 -4.95424300e-01 3.47119063e-01
-3.81851107e-01 5.84319904e-02 3.97783101e-01 6.82691753e-01
5.49858928e-01 3.66200924e-01 3.47596020e-01 -1.43438745e+00
1.15595803e-01 -6.53846502e-01 -1.09083390e+00 5.38949251e-01
-1.00577152e+00 2.07577068e-02 1.29250252e+00 -6.20411187e-02
-3.96168053e-01 1.00419939e-01 3.19555014e-01 -3.91115457e-01
1.96364313e-01 2.10022748e-01 -1.16643079e-01 -2.27441758e-01
5.02188265e-01 2.47408617e-02 -1.82150766e-01 -2.24161640e-01
8.35343301e-02 3.70255947e-01 1.66677892e-01 -4.63169664e-01
9.44992363e-01 3.52073282e-01 3.85743231e-01 -9.39772874e-02
-1.08182716e+00 -4.77192819e-01 -2.65140474e-01 -1.66539215e-02
-9.33615640e-02 -5.45607328e-01 -1.22478628e+00 -1.38132542e-01
-9.87853825e-01 -5.12433946e-01 -1.91288114e-01 -7.93118589e-03
-9.57224548e-01 2.72694886e-01 -6.09170973e-01 -1.50712144e+00
-1.91878304e-02 -4.87829894e-01 4.79512066e-01 1.59331650e-01
1.78945318e-01 -1.19346905e+00 -3.92383367e-01 1.66843429e-01
1.60998583e-01 1.30540729e-01 1.00184143e+00 -1.20699441e+00
-9.62085962e-01 -5.10349154e-01 -1.15542687e-01 -3.96742821e-01
-3.71358812e-01 1.47764266e-01 -3.54183257e-01 -5.85326970e-01
-5.44395506e-01 2.80759335e-01 2.44364589e-01 3.57644200e-01
1.51179576e+00 -9.00966108e-01 -7.92462885e-01 5.67636192e-01
1.85866046e+00 7.51941502e-02 -5.96347339e-02 9.81427506e-02
1.66248098e-01 4.22425181e-01 8.81302238e-01 8.83614540e-01
2.18040541e-01 1.15399413e-01 5.33701897e-01 6.30771220e-01
4.74524587e-01 -7.63244629e-01 8.11714306e-03 5.53593278e-01
3.81974906e-01 -9.12303150e-01 -1.06689835e+00 7.62108803e-01
-2.25569797e+00 -8.54881227e-01 -4.54771310e-01 2.63160324e+00
5.14253855e-01 8.38939071e-01 5.25355160e-01 2.18936399e-01
8.37218761e-01 -4.59248066e-01 -9.61512804e-01 -1.04738712e+00
1.27081260e-01 -5.21511398e-02 1.65329754e+00 6.56901360e-01
-3.71235967e-01 7.69147456e-01 7.50483513e+00 5.84736645e-01
-4.50730532e-01 1.41743302e-01 7.91112304e-01 -2.79589087e-01
-6.46311283e-01 3.30692947e-01 -8.32770526e-01 4.89497334e-01
1.52723598e+00 -8.01414430e-01 7.20823824e-01 1.29842412e+00
2.86886096e-01 6.41898345e-03 -1.41614640e+00 7.46070027e-01
-4.04827178e-01 -1.52944493e+00 -3.54687423e-01 3.56698155e-01
9.99183178e-01 -1.00654870e-01 -5.28029837e-02 3.98599282e-02
9.29395497e-01 -6.74548805e-01 4.71424162e-01 3.95704955e-01
5.88236213e-01 -1.30892491e+00 6.53230369e-01 9.06895816e-01
-8.29669952e-01 -7.99864471e-01 -3.04588795e-01 -2.23912746e-01
1.61730405e-02 6.55200779e-01 -1.10917318e+00 3.39881957e-01
4.22467291e-01 2.34527007e-01 -1.01625182e-01 1.72020125e+00
1.68231308e-01 6.79222584e-01 -8.63409817e-01 -3.20087552e-01
1.56272069e-01 -6.66748807e-02 4.80957627e-01 7.71545291e-01
2.25685298e-01 3.03947717e-01 5.64454556e-01 3.01450878e-01
-1.58344805e-01 1.31051391e-01 -6.06067181e-01 -2.87675381e-01
8.75837505e-01 8.27170312e-01 -7.61824191e-01 5.28353117e-02
-2.49907404e-01 1.01283944e+00 3.20560664e-01 2.96869695e-01
-8.93694282e-01 -3.33769888e-01 5.64905167e-01 3.52547973e-01
5.55469871e-01 -3.17839533e-01 -6.66724861e-01 -5.86656034e-01
2.43583843e-01 -1.85543120e-01 5.85709333e-01 -6.39452875e-01
-1.42599392e+00 1.17310755e-01 -4.73352522e-01 -8.82209361e-01
-9.91998836e-02 -7.03914940e-01 -4.45752472e-01 3.77966672e-01
-1.20348692e+00 -5.99961817e-01 2.54019380e-01 3.76565121e-02
4.44216609e-01 3.81755680e-01 3.36828768e-01 -2.89805513e-02
-3.20403963e-01 1.17432547e+00 7.65397131e-01 -4.06829745e-01
9.20376331e-02 -1.45579624e+00 3.06845069e-01 8.99315476e-01
7.30483904e-02 -3.28453183e-02 1.23519218e+00 -5.19849896e-01
-1.82989049e+00 -9.58115458e-01 9.95551944e-01 -4.43567544e-01
1.11098719e+00 -1.34443328e-01 -2.91892916e-01 9.57695246e-01
-3.96007031e-01 2.86972910e-01 7.67003775e-01 5.27607858e-01
-1.93314731e-01 -3.69727492e-01 -1.28318775e+00 6.28628314e-01
1.86681402e+00 -9.15411264e-02 1.86062574e-01 1.00706112e+00
9.94092345e-01 -3.80591810e-01 -6.27224684e-01 -3.68718640e-03
4.76776361e-01 -8.86191845e-01 4.92925085e-02 -1.21458805e+00
-2.20801146e-03 2.26894036e-01 -6.71764687e-02 -1.01922297e+00
-2.89346784e-01 -1.19512546e+00 -4.60601091e-01 7.46048629e-01
1.09555852e+00 -7.70605683e-01 1.49043953e+00 8.51123869e-01
8.43873024e-02 -1.11284876e+00 -8.78554463e-01 -1.28819597e+00
8.55239183e-02 -4.22653615e-01 5.62855124e-01 5.34164906e-01
3.38945925e-01 -1.12532653e-01 4.31594551e-02 6.41135871e-01
7.41112173e-01 6.56701446e-01 7.75702477e-01 -1.10244656e+00
-7.22359657e-01 -4.41241920e-01 -4.21486259e-01 -1.52942038e+00
4.32106070e-02 -9.22841251e-01 -1.47102058e-01 -1.40574074e+00
2.60304511e-01 -8.20323706e-01 -4.93301935e-02 2.61462271e-01
4.02619541e-01 -6.90409303e-01 -1.85921378e-02 -9.19312388e-02
-1.05996180e+00 1.17467918e-01 1.22324955e+00 3.18481892e-01
-6.33396208e-03 7.44354904e-01 -7.29028404e-01 4.34404552e-01
8.26672792e-01 -5.51086783e-01 -4.22533572e-01 -4.77800906e-01
7.06135273e-01 7.46739924e-01 -2.11528525e-01 -6.58665657e-01
6.27000451e-01 -6.02578163e-01 8.09770077e-02 -6.31898046e-01
-1.76743090e-01 -1.38705182e+00 2.02837884e-01 6.94474697e-01
-7.15780079e-01 4.18410361e-01 -2.08758608e-01 1.43725979e+00
6.38009608e-01 -7.93348998e-02 5.47470450e-01 3.78158480e-01
-1.02047123e-01 8.64629984e-01 -1.77760631e-01 2.44055882e-01
1.47708642e+00 -6.10825658e-01 -3.58879775e-01 -7.35150814e-01
-1.08947456e+00 8.64581585e-01 9.23362315e-01 -1.53414622e-01
3.19455802e-01 -7.84155190e-01 -2.21339658e-01 -2.20334113e-01
1.39995098e-01 -3.01773906e-01 1.11499473e-01 1.00076234e+00
-7.36509919e-01 2.09175676e-01 3.76812369e-01 -3.11840385e-01
-1.20552123e+00 1.14977264e+00 2.65276253e-01 -2.24271700e-01
-2.67889589e-01 9.10819352e-01 -5.30720115e-01 -2.48460006e-02
4.49280202e-01 2.55551981e-03 7.95238078e-01 -6.78146541e-01
2.09298089e-01 3.76120150e-01 -1.45427406e-01 7.41842166e-02
-2.07493037e-01 -2.57365376e-01 -3.63715708e-01 1.48729682e-01
1.41412890e+00 -5.19992113e-01 2.82436788e-01 4.44656193e-01
9.57035780e-01 5.20347431e-02 -1.22127855e+00 -3.25465649e-01
3.37912291e-01 -9.76882458e-01 -3.09363693e-01 -8.94299209e-01
-1.05619645e+00 5.40885091e-01 2.69561201e-01 7.84983993e-01
1.02251494e+00 5.93985841e-02 1.11183596e+00 4.02368993e-01
1.20172715e+00 -1.35680246e+00 -6.02477670e-01 4.35999453e-01
8.57517049e-02 -8.71308982e-01 -1.31046906e-01 -8.87596607e-01
-1.82436571e-01 1.17247307e+00 5.51093876e-01 -1.74486339e-01
5.39872766e-01 4.45088416e-01 -6.49298131e-01 3.48870665e-01
-1.60842860e+00 -2.48196483e-01 -4.99502689e-01 3.79755020e-01
-1.34953380e-01 5.28262734e-01 -4.58954722e-01 8.70747030e-01
-3.18713456e-01 -3.61877203e-01 7.49077857e-01 7.63435304e-01
-9.49425757e-01 -1.11683559e+00 -5.76978251e-02 7.48309612e-01
-5.41823983e-01 1.05002634e-01 -5.95002234e-01 5.87334037e-01
-4.25572783e-01 1.12964857e+00 1.06457569e-01 -4.70047325e-01
-1.34767666e-02 -1.55314103e-01 7.97653735e-01 -4.56528068e-01
-2.96880424e-01 -3.55902284e-01 3.60433757e-01 -6.71350718e-01
3.37451607e-01 -5.31401396e-01 -1.28299117e+00 -1.15981710e+00
-6.52750075e-01 4.54184949e-01 5.57429075e-01 9.40416396e-01
6.23235226e-01 -5.77300936e-02 1.31961298e+00 1.75945416e-01
-1.14172459e+00 -6.32163212e-02 -8.49814415e-01 -8.49448666e-02
1.90383688e-01 -2.42706418e-01 -6.28666282e-01 -3.74851763e-01] | [4.700952529907227, 3.4894657135009766] |
efb05fb6-c01c-4f1f-b4ff-5cf5d3cd8116 | off-the-shelf-unsupervised-nmt | 1811.02278 | null | http://arxiv.org/abs/1811.02278v1 | http://arxiv.org/pdf/1811.02278v1.pdf | Off-the-Shelf Unsupervised NMT | We frame unsupervised machine translation (MT) in the context of multi-task
learning (MTL), combining insights from both directions. We leverage
off-the-shelf neural MT architectures to train unsupervised MT models with no
parallel data and show that such models can achieve reasonably good
performance, competitive with models purpose-built for unsupervised MT.
Finally, we propose improvements that allow us to apply our models to
English-Turkish, a truly low-resource language pair. | ['John Glover', 'Sebastian Ruder', 'Chris Hokamp'] | 2018-11-06 | null | null | null | null | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 2.86626965e-01 4.04700667e-01 -6.12586021e-01 -3.52985412e-01
-1.73416460e+00 -7.11133659e-01 9.47205007e-01 -3.82293552e-01
-6.28045201e-01 9.34501946e-01 4.72196728e-01 -7.48944402e-01
5.28693855e-01 -2.30313361e-01 -1.02792919e+00 -1.68167442e-01
4.76759255e-01 1.09726739e+00 -1.19366303e-01 -4.25323278e-01
-1.36713743e-01 -1.41033798e-01 -6.96085989e-01 6.82274818e-01
1.01076829e+00 1.90925136e-01 2.08935902e-01 5.45476139e-01
-7.86708295e-02 7.93144047e-01 -2.23903075e-01 -7.68183947e-01
2.04699174e-01 -2.13453516e-01 -1.12633550e+00 -2.67518252e-01
5.81110120e-01 -1.70297742e-01 -2.27188677e-01 5.09567499e-01
5.56716323e-01 -1.81560591e-01 8.26044679e-01 -8.73321831e-01
-1.03010595e+00 8.46546471e-01 -3.02862138e-01 1.33800298e-01
1.82740793e-01 6.23849370e-02 1.09958065e+00 -1.35869861e+00
9.74204957e-01 1.19622052e+00 6.44523144e-01 6.45104349e-01
-1.30721164e+00 -4.15453315e-01 -1.21182531e-01 1.60056636e-01
-8.18747640e-01 -1.00498438e+00 4.22495097e-01 -3.40145230e-01
1.48621190e+00 -1.15650326e-01 -2.07972783e-03 1.61271024e+00
4.86954778e-01 1.10514259e+00 1.25099754e+00 -9.95023191e-01
-3.74110848e-01 -1.13024019e-01 -1.91810563e-01 6.47119224e-01
-1.81923464e-01 -1.95005611e-02 -7.68123209e-01 4.25859587e-03
5.10227859e-01 -4.80981231e-01 2.03313947e-01 -3.51256989e-02
-1.93158984e+00 7.27917314e-01 -3.76706831e-02 5.88722527e-01
-1.58875674e-01 4.10877228e-01 5.56636214e-01 8.10929000e-01
9.79888439e-01 4.07691956e-01 -9.69614208e-01 -4.16599631e-01
-9.22636628e-01 -2.70692796e-01 7.95218766e-01 1.28229439e+00
8.58452499e-01 1.47291079e-01 -1.67972907e-01 1.08520746e+00
-7.45839477e-02 8.32209826e-01 7.23129034e-01 -7.92969584e-01
1.10954773e+00 5.92131205e-02 -1.13440856e-01 -9.35427248e-02
-1.30480617e-01 -2.33221665e-01 -5.20671546e-01 -5.79654932e-01
2.51134127e-01 -4.07369077e-01 -7.55648196e-01 1.68330443e+00
-1.14916906e-01 -7.01565370e-02 3.69252175e-01 4.31551963e-01
4.45739746e-01 5.55921495e-01 -1.13674693e-01 -3.25265527e-01
1.07713258e+00 -1.58107758e+00 -6.07426524e-01 -6.48974180e-01
1.17889285e+00 -1.07104266e+00 1.20077324e+00 1.13884218e-01
-1.28599250e+00 -4.04289484e-01 -7.33828306e-01 -5.16178787e-01
-4.49114144e-01 5.03954709e-01 5.89492083e-01 3.23347509e-01
-1.44435620e+00 4.29094017e-01 -9.51610982e-01 -7.37911046e-01
5.78126460e-02 4.98494565e-01 -4.21342075e-01 -4.84588593e-02
-1.03641856e+00 1.54861271e+00 2.44498029e-01 4.84204181e-02
-1.12735224e+00 -2.39960819e-01 -8.33387196e-01 -4.78865474e-01
7.19449744e-02 -7.51011491e-01 1.71509087e+00 -1.16921997e+00
-1.71252990e+00 1.02574086e+00 -7.75936723e-01 -4.35047477e-01
3.59135568e-01 -4.87390220e-01 -3.46612006e-01 1.90423394e-03
3.34498972e-01 6.81484461e-01 6.60098970e-01 -8.77137721e-01
-3.40299308e-01 -1.17676064e-01 -3.02529871e-01 3.03878367e-01
-5.72104454e-01 6.10852361e-01 -1.18261702e-01 -8.28663468e-01
-1.60437986e-01 -9.89360988e-01 -2.07216531e-01 -7.09038436e-01
-2.79727250e-01 -2.08052799e-01 6.26540959e-01 -9.09404099e-01
8.51265252e-01 -1.61491561e+00 5.20734191e-01 -4.50822830e-01
-5.81312412e-03 1.84821282e-02 -6.28253281e-01 7.00627983e-01
2.21748114e-01 2.18826219e-01 -1.53818801e-01 -7.94222236e-01
1.61914807e-02 3.37817639e-01 -4.02238071e-01 1.82713002e-01
4.94409263e-01 1.75077784e+00 -9.27484214e-01 -5.90348184e-01
-1.76201984e-01 -1.57966286e-01 -2.41181955e-01 1.79306641e-01
-5.75774670e-01 5.89707971e-01 -3.50680411e-01 7.83066928e-01
7.82989636e-02 -2.37683475e-01 6.31387591e-01 2.58883864e-01
-2.24201575e-01 8.28245103e-01 -3.81393731e-02 2.11641049e+00
-7.93613672e-01 8.74904037e-01 -2.17022732e-01 -9.43258226e-01
6.51842833e-01 7.13011384e-01 2.78354496e-01 -7.13862002e-01
-1.61366120e-01 8.02268326e-01 4.68974523e-02 -3.92955929e-01
5.93159020e-01 -1.33798331e-01 -2.34429479e-01 1.04072332e+00
5.77159464e-01 -2.83261210e-01 -2.84816604e-03 -2.27869246e-02
9.11065698e-01 5.81382930e-01 1.98248968e-01 -4.37412411e-01
1.01927832e-01 2.91145980e-01 2.78637022e-01 6.26819611e-01
2.20610388e-02 3.61946553e-01 -1.33936023e-02 -6.00309968e-01
-1.34024417e+00 -1.19620144e+00 1.39378622e-01 1.67802036e+00
-5.53380251e-01 -2.82385826e-01 -6.69891894e-01 -9.73423839e-01
-3.79323363e-01 4.99799192e-01 -4.83320653e-01 2.06345543e-01
-1.24903536e+00 -7.21418321e-01 9.89092588e-01 4.46424037e-01
1.07321620e-01 -1.03282225e+00 5.96795529e-02 4.13064837e-01
-7.91119814e-01 -1.42753732e+00 -6.97943330e-01 4.70781177e-01
-1.12678647e+00 -5.18454909e-01 -8.05785060e-01 -1.31569898e+00
3.34765404e-01 1.85682058e-01 1.50711679e+00 -2.03165337e-01
5.35610676e-01 1.38316736e-01 -1.97236121e-01 -4.70746011e-01
-8.58618379e-01 7.88879871e-01 4.80852425e-01 -3.62304330e-01
4.71894771e-01 -6.90866828e-01 1.21716477e-01 2.60585576e-01
-4.82843816e-01 3.68099123e-01 8.88936937e-01 1.00296593e+00
3.56969148e-01 -1.12239242e+00 8.34707141e-01 -7.45952010e-01
5.18313885e-01 -4.69059587e-01 -1.81851774e-01 7.10416257e-01
-4.40149099e-01 2.52386510e-01 7.14841425e-01 -5.96222222e-01
-8.50183666e-01 -1.89711303e-01 1.51977494e-01 -3.19443554e-01
1.09619893e-01 6.38943613e-01 1.76510960e-01 -1.15843430e-01
5.37037551e-01 4.56130564e-01 -1.12761661e-01 -6.05186224e-01
7.65180469e-01 8.23439002e-01 3.94095153e-01 -9.99017775e-01
7.50239491e-01 1.27229646e-01 -2.60150433e-01 -5.27062774e-01
-7.80040085e-01 -1.02375433e-01 -1.16926229e+00 1.37272179e-01
7.39775777e-01 -1.25325453e+00 5.80081269e-02 1.55239671e-01
-1.50628626e+00 -9.34485555e-01 1.08921014e-01 5.67863166e-01
-9.98131514e-01 2.46580452e-01 -1.16122127e+00 -3.07574093e-01
-6.33032441e-01 -1.09083700e+00 1.33381426e+00 -4.71085727e-01
-2.40088388e-01 -1.30005372e+00 3.08335900e-01 4.19575095e-01
5.33359826e-01 -3.49310368e-01 1.11820221e+00 -9.47985232e-01
-5.11650026e-01 5.14398515e-01 -6.96930736e-02 4.52147610e-02
2.02931300e-01 -1.79512143e-01 -9.07400727e-01 -3.04229170e-01
-1.79978877e-01 -8.49531293e-01 6.82644188e-01 1.99468210e-01
3.87866527e-01 -5.09900451e-01 -4.28851962e-01 5.57680309e-01
1.01770437e+00 -4.14935648e-01 3.48757535e-01 5.06482780e-01
8.55063260e-01 5.12589335e-01 4.84153181e-01 -2.79647112e-01
9.95969832e-01 7.91777194e-01 -2.29120910e-01 -2.46170610e-01
-2.42359787e-01 -3.71583343e-01 9.33666945e-01 1.99563539e+00
-1.75575875e-02 -9.44479369e-03 -1.29317892e+00 8.80491436e-01
-2.01795864e+00 -6.39167190e-01 -1.67734429e-01 1.75456989e+00
1.20669734e+00 6.94519142e-03 8.59759524e-02 -8.01465452e-01
5.02118349e-01 5.22995964e-02 -2.39293590e-01 -9.20237899e-01
-4.43407029e-01 5.74792504e-01 4.23605412e-01 5.46277225e-01
-1.09226131e+00 1.77019644e+00 8.21600533e+00 8.40645492e-01
-1.05079436e+00 9.21660900e-01 5.71544707e-01 -7.01398477e-02
-4.49581057e-01 2.28469923e-01 -5.63677788e-01 1.05062954e-01
1.64145839e+00 -2.41841808e-01 6.58222616e-01 4.07040775e-01
2.69887805e-01 4.49186891e-01 -1.44142616e+00 5.02800345e-01
1.31066650e-01 -1.29890132e+00 2.40810469e-01 2.45598555e-01
1.07822013e+00 1.00749516e+00 7.41488263e-02 5.99670708e-01
9.15206313e-01 -1.02197003e+00 7.08105683e-01 1.45626858e-01
1.11668038e+00 -2.82336056e-01 4.73376542e-01 5.72684288e-01
-9.05765116e-01 1.50365889e-01 -4.20242518e-01 -1.80941537e-01
1.57411426e-01 2.47025996e-01 -1.00397408e+00 8.39813232e-01
1.03798307e-01 9.59086776e-01 -6.42062962e-01 2.61530042e-01
-4.22019988e-01 7.99526691e-01 -2.25957289e-01 1.24365643e-01
4.38492298e-01 -2.98952848e-01 3.66411865e-01 1.48898411e+00
4.55613315e-01 -3.95662278e-01 2.61726081e-01 5.26861727e-01
-4.88193482e-01 4.31204945e-01 -1.12214673e+00 -2.05239192e-01
2.86398232e-01 9.60498750e-01 -1.10293306e-01 -6.55255258e-01
-6.06679201e-01 1.31642187e+00 9.37842846e-01 4.31389421e-01
-6.32414401e-01 2.72195071e-01 3.86450291e-01 -4.63514984e-01
7.54684508e-02 -6.99673414e-01 -4.56302404e-01 -1.83023846e+00
1.33013114e-01 -1.25552452e+00 2.33415335e-01 -6.84637487e-01
-1.39948142e+00 6.81380928e-01 -2.77307302e-01 -1.30613673e+00
-9.98883426e-01 -5.50163090e-01 -4.54128683e-01 9.09783542e-01
-1.57879853e+00 -1.99386537e+00 7.76410520e-01 3.76878381e-01
1.02473652e+00 -4.07301873e-01 1.07138836e+00 1.64839447e-01
-3.80546689e-01 8.82695436e-01 5.46827614e-01 2.91918546e-01
1.14501846e+00 -1.17470181e+00 1.23208821e+00 1.06187451e+00
8.09300423e-01 6.05879784e-01 1.10166363e-01 -5.89536488e-01
-1.76913059e+00 -1.18044889e+00 1.65653014e+00 -1.19258821e+00
9.13538098e-01 -7.55962610e-01 -3.91725838e-01 1.33524060e+00
6.91008210e-01 -2.46702597e-01 6.59203768e-01 4.84595448e-01
-4.72442210e-01 4.54435796e-01 -6.77195907e-01 8.76452208e-01
1.12490845e+00 -1.04961956e+00 -9.16187227e-01 5.32286763e-01
1.07643247e+00 -2.02045649e-01 -1.02996075e+00 3.22303563e-01
7.61855602e-01 -2.29399458e-01 6.89241350e-01 -9.19244707e-01
7.83134937e-01 1.68705150e-01 -4.08651561e-01 -1.64978147e+00
-1.76068321e-01 -9.77042973e-01 -2.37419456e-01 9.35589433e-01
9.68480527e-01 -7.44333863e-01 3.80432367e-01 6.04826026e-02
-4.42113549e-01 -4.98189449e-01 -1.19882572e+00 -9.01313841e-01
9.00563180e-01 -3.94068122e-01 4.29764360e-01 1.28539288e+00
4.54290152e-01 7.48918235e-01 -5.32865942e-01 -1.93834350e-01
4.53609318e-01 1.10101346e-02 9.22852635e-01 -9.44252014e-01
-4.70814645e-01 -3.82758349e-01 2.63875604e-01 -1.15348971e+00
7.01335013e-01 -1.37162876e+00 7.41049573e-02 -1.48255384e+00
4.24831092e-01 -2.22900152e-01 -2.09000379e-01 8.68089855e-01
-1.74074844e-01 5.21797419e-01 2.77635723e-01 7.15019226e-01
-5.98139167e-01 6.03489101e-01 1.23165035e+00 -2.88570255e-01
7.29843304e-02 -5.03489494e-01 -5.54099083e-01 4.48535770e-01
7.33079731e-01 -6.11894906e-01 1.58345476e-02 -1.55212045e+00
5.49966469e-02 6.40722066e-02 -2.06312507e-01 -5.11143625e-01
1.58814281e-01 -2.82030731e-01 1.31963938e-01 -3.26062381e-01
2.11323634e-01 -3.14677536e-01 -2.64485806e-01 1.43072709e-01
-5.16615272e-01 6.58879042e-01 2.61949182e-01 4.80263785e-04
-9.76650566e-02 8.53287727e-02 2.47924924e-01 -2.42380634e-01
-3.07448089e-01 1.99517235e-01 -6.36412680e-01 2.03853458e-01
3.82326871e-01 1.53922200e-01 -6.02757215e-01 -4.01195675e-01
-4.59921181e-01 1.76201358e-01 6.15941048e-01 6.03744090e-01
2.88601220e-01 -1.43012226e+00 -1.19198215e+00 7.64645170e-03
4.38809812e-01 -6.06004536e-01 -6.44231260e-01 1.05035186e+00
-8.70836452e-02 7.86467016e-01 3.48728709e-02 -5.46867669e-01
-6.95103645e-01 4.01243985e-01 1.07151657e-01 -5.99009216e-01
-2.86189497e-01 3.82432371e-01 -1.51839018e-01 -1.20829248e+00
-3.76996785e-01 -2.65926719e-01 3.70479554e-01 -3.32408100e-01
1.20477259e-01 3.32137905e-02 2.00304240e-01 -7.22064376e-01
-3.13520074e-01 4.49102193e-01 -2.36723378e-01 -8.64247501e-01
1.23072553e+00 -2.69831359e-01 -4.53191459e-01 8.50541115e-01
1.09229457e+00 1.35612324e-01 -5.99283397e-01 -5.67404747e-01
6.38169825e-01 1.53744817e-01 -3.22188169e-01 -1.02066040e+00
-2.79009610e-01 1.03549814e+00 8.76036063e-02 -6.01199150e-01
7.59731531e-01 4.85831127e-02 1.04959917e+00 1.00476432e+00
8.37526023e-01 -1.32710004e+00 9.93436351e-02 1.28186429e+00
4.86101836e-01 -1.30663574e+00 -2.72126645e-01 -3.40806060e-02
-5.43525696e-01 1.09558332e+00 3.81562442e-01 1.44319952e-01
1.68067887e-01 3.16005409e-01 3.79660815e-01 1.16075799e-01
-1.30821013e+00 -3.07997197e-01 3.82373720e-01 7.05355942e-01
8.98426116e-01 3.37945521e-01 -3.38952035e-01 2.52104104e-01
-3.01087976e-01 8.16937089e-02 4.52200472e-01 8.38887990e-01
-1.17245510e-01 -1.62122166e+00 -3.79866920e-02 3.34582835e-01
-7.02259302e-01 -8.74731004e-01 -5.66347718e-01 6.58640862e-01
-3.10570747e-01 1.13681519e+00 -1.37267768e-01 -4.07200754e-01
-7.73200095e-02 6.47011697e-01 7.84596622e-01 -8.48303854e-01
-8.42118800e-01 1.92151368e-01 6.02030158e-01 -2.66702712e-01
-2.91052729e-01 -6.44470274e-01 -6.17590785e-01 -1.18465751e-01
-2.76994973e-01 -3.49761620e-02 7.02904701e-01 1.34762716e+00
5.61220169e-01 -7.31421113e-02 6.55198872e-01 -8.44697535e-01
-5.32345593e-01 -1.48479342e+00 3.03328902e-01 -6.94793910e-02
2.50043720e-01 -7.84511045e-02 7.31444657e-02 3.03363025e-01] | [11.59795093536377, 10.380215644836426] |
7108b22d-05c0-464c-9f2b-20ed09e48421 | 3d-hand-pose-tracking-and-estimation-using | 1610.07214 | null | http://arxiv.org/abs/1610.07214v1 | http://arxiv.org/pdf/1610.07214v1.pdf | 3D Hand Pose Tracking and Estimation Using Stereo Matching | 3D hand pose tracking/estimation will be very important in the next
generation of human-computer interaction. Most of the currently available
algorithms rely on low-cost active depth sensors. However, these sensors can be
easily interfered by other active sources and require relatively high power
consumption. As a result, they are currently not suitable for outdoor
environments and mobile devices. This paper aims at tracking/estimating hand
poses using passive stereo which avoids these limitations. A benchmark with
18,000 stereo image pairs and 18,000 depth images captured from different
scenarios and the ground-truth 3D positions of palm and finger joints (obtained
from the manual label) is thus proposed. This paper demonstrates that the
performance of the state-of-the art tracking/estimation algorithms can be
maintained with most stereo matching algorithms on the proposed benchmark, as
long as the hand segmentation is correct. As a result, a novel stereo-based
hand segmentation algorithm specially designed for hand tracking/estimation is
proposed. The quantitative evaluation demonstrates that the proposed algorithm
is suitable for the state-of-the-art hand pose tracking/estimation algorithms
and the tracking quality is comparable to the use of active depth sensors under
different challenging scenarios. | ['Xiaobin Xu', 'Jianbo Jiao', 'Mingliang Chen', 'Qingxiong Yang', 'Liangqiong Qu', 'Jiawei Zhang'] | 2016-10-23 | null | null | null | null | ['hand-segmentation'] | ['computer-vision'] | [ 4.43865247e-02 -5.91495395e-01 -1.39466956e-01 4.43590954e-02
-4.37183470e-01 -6.30700648e-01 2.36414984e-01 -1.26784503e-01
-6.71588480e-01 4.90725964e-01 -2.99987167e-01 2.70460993e-01
-7.62744918e-02 -5.00906944e-01 -4.15848404e-01 -8.36652815e-01
2.56416947e-01 1.02669859e+00 7.90310085e-01 -1.09182680e-02
3.55344445e-01 9.35294688e-01 -1.84659827e+00 -3.63199413e-01
4.93137151e-01 8.68322015e-01 6.36554182e-01 6.60133600e-01
-3.82441543e-02 -6.98300602e-04 -7.51001954e-01 -3.43527794e-01
4.36070144e-01 4.53675464e-02 -4.59697813e-01 1.48006767e-01
7.76982367e-01 -7.55848408e-01 -3.31928909e-01 1.08355606e+00
1.17447162e+00 -1.02363631e-01 3.87945861e-01 -1.02581930e+00
4.46311057e-01 5.86028174e-02 -7.94506133e-01 -4.92818244e-02
7.82099605e-01 1.44133106e-01 4.38491315e-01 -7.66888678e-01
9.04831231e-01 1.15706563e+00 5.31582654e-01 7.24857986e-01
-1.02941942e+00 -7.10624516e-01 9.46170986e-02 1.67979091e-01
-1.61230338e+00 -1.26792967e-01 8.95272970e-01 -5.23884535e-01
6.06488049e-01 3.59807312e-01 9.70764279e-01 1.20413220e+00
1.45566523e-01 8.53717446e-01 1.09164572e+00 -5.93790591e-01
1.17674060e-01 7.78058097e-02 1.20668560e-01 7.50415623e-01
4.31272298e-01 1.25002846e-01 -7.71452248e-01 -7.54146203e-02
1.14427507e+00 1.66662317e-02 -3.47087473e-01 -8.45941842e-01
-1.34539068e+00 5.17615937e-02 2.71003246e-01 5.35252512e-01
-4.64866549e-01 -1.11431532e-01 2.88097948e-01 -1.60048828e-01
1.28736675e-01 -1.36027753e-01 -1.45426542e-01 -2.70207226e-01
-1.11914897e+00 4.37038928e-01 6.34897768e-01 1.25955749e+00
2.12637648e-01 -2.48008221e-01 -1.01339042e-01 5.07029831e-01
6.37982011e-01 8.84309649e-01 1.42845616e-01 -7.45844245e-01
6.65760696e-01 5.62588930e-01 3.99836838e-01 -8.81564975e-01
-5.32973707e-01 -3.93544227e-01 -6.90907300e-01 5.60007632e-01
1.03981709e+00 1.00250401e-01 -8.33287239e-01 1.20055723e+00
5.28533578e-01 -3.12837958e-01 -6.16031826e-01 1.23130012e+00
6.12831652e-01 5.48872799e-02 -1.12964667e-01 -2.53576189e-01
1.35927904e+00 -5.14826775e-01 -9.68248546e-01 -2.35386327e-01
-1.60376072e-01 -1.21994209e+00 9.58656192e-01 7.32049763e-01
-9.64320421e-01 -7.03515828e-01 -9.31795895e-01 3.28290105e-01
-1.04098596e-01 1.84283718e-01 4.50647265e-01 1.04286253e+00
-7.34482765e-01 5.04472375e-01 -1.07353294e+00 -5.51065028e-01
-1.75451469e-02 6.35821223e-01 -3.20503086e-01 1.42161831e-01
-8.96975577e-01 7.71503508e-01 2.77125269e-01 4.25692737e-01
-3.53040934e-01 -3.41912985e-01 -2.74133265e-01 -4.88165498e-01
1.48440123e-01 -5.30909419e-01 1.20770872e+00 -2.62297422e-01
-1.59706581e+00 1.17556059e+00 -2.72331029e-01 2.11738199e-01
1.06895888e+00 -4.37475026e-01 -8.10820460e-02 1.98087856e-01
-5.47985137e-02 2.55128652e-01 9.12308097e-01 -1.24268520e+00
-6.54752314e-01 -1.18505347e+00 -1.34573787e-01 2.19847456e-01
-3.39679122e-01 -3.72374011e-03 -8.98194492e-01 -5.83726704e-01
4.72771287e-01 -1.14012671e+00 1.57568783e-01 3.66309077e-01
-4.32975292e-01 -1.73044801e-01 9.12229896e-01 -8.18323731e-01
1.02497029e+00 -1.59808958e+00 3.11618298e-01 4.63027000e-01
-2.40750071e-02 3.88145596e-01 2.79867828e-01 1.10929668e-01
4.07504857e-01 -7.12226033e-01 1.22551747e-01 -3.82993102e-01
-1.78956926e-01 -1.67800516e-01 2.18199849e-01 8.16572785e-01
-7.94172585e-01 3.96662563e-01 -8.30657661e-01 -8.65987480e-01
7.25477278e-01 8.14996421e-01 -1.30391959e-02 4.03080910e-01
5.46547733e-02 8.98316860e-01 -4.83348191e-01 9.03855443e-01
8.03835571e-01 1.71926707e-01 2.68986076e-01 -4.18996215e-01
-2.38878444e-01 -2.32831582e-01 -1.81077826e+00 1.97988117e+00
-3.32753718e-01 4.68620718e-01 3.30986947e-01 -2.19586700e-01
8.90282631e-01 6.45235240e-01 6.58237875e-01 -4.84828353e-01
4.45877880e-01 5.56691766e-01 -2.42733896e-01 -4.13449645e-01
2.13364735e-01 2.06559509e-01 4.26560074e-01 4.47051555e-01
-2.19580263e-01 -1.50135666e-01 1.35544653e-03 -3.34042639e-01
5.69630265e-01 5.09862900e-01 2.12934241e-01 -1.80081740e-01
7.35463142e-01 -1.31799698e-01 3.05100828e-01 4.33565468e-01
-3.26974064e-01 6.50644064e-01 -2.83480912e-01 -1.35374546e-01
-9.78010595e-01 -1.20507383e+00 -2.17977360e-01 6.92984939e-01
3.13379347e-01 -1.17239609e-01 -1.12479317e+00 -2.89551049e-01
1.76849589e-01 -1.83606267e-01 -1.25503168e-01 5.30285776e-01
-8.40282857e-01 -3.08040142e-01 4.05745566e-01 7.55428016e-01
7.58559942e-01 -8.52048457e-01 -1.02364385e+00 3.84079069e-01
-1.58357725e-01 -1.16558528e+00 -4.88932043e-01 -3.47427189e-01
-1.07787478e+00 -1.31096971e+00 -1.49681437e+00 -8.44235480e-01
6.08257711e-01 3.70514452e-01 7.51697600e-01 -1.62788108e-01
-3.87167335e-01 6.50417030e-01 -1.76204950e-01 -3.65695775e-01
-4.43572514e-02 2.32093170e-01 3.90477389e-01 -1.52573526e-01
4.30621326e-01 -4.16909486e-01 -9.99113977e-01 7.29659796e-01
-1.38261318e-01 -3.39670390e-01 2.69724280e-01 5.31464636e-01
5.63327312e-01 -2.07406238e-01 -1.24880694e-01 -4.75723386e-01
4.01102751e-01 5.22606909e-01 -1.01645291e+00 2.33603448e-01
-5.22102416e-01 -1.79165564e-02 2.02038988e-01 -5.26928604e-01
-1.28181314e+00 6.02980018e-01 -3.62250119e-01 -3.01843464e-01
-3.06523204e-01 -3.79213870e-01 -5.53318858e-01 -3.41294497e-01
6.78663433e-01 1.56593576e-01 1.03889838e-01 -8.97503078e-01
-9.97435600e-02 1.18924069e+00 6.48945212e-01 -4.73406166e-01
6.87018514e-01 6.37854695e-01 1.83671206e-01 -9.63220596e-01
-3.52084696e-01 -8.14283490e-01 -1.17141759e+00 -7.28244483e-01
6.53118074e-01 -6.14679813e-01 -1.22784555e+00 1.28583455e+00
-1.45055282e+00 5.56582212e-02 1.11408032e-01 7.28352189e-01
-6.13634825e-01 7.36364365e-01 -4.63631988e-01 -1.35203159e+00
-5.18878996e-01 -1.31610858e+00 1.36510158e+00 2.22161174e-01
-3.75231624e-01 -7.22464561e-01 -2.26837080e-02 4.18971926e-01
6.53068870e-02 2.46077538e-01 1.77804351e-01 1.94295168e-01
-6.15322888e-01 -6.20095730e-01 1.08895287e-01 1.44897057e-02
2.96005636e-01 -8.59222040e-02 -1.16640210e+00 -5.93115509e-01
-8.68084133e-02 1.85229927e-01 1.94902435e-01 4.73490715e-01
9.17753875e-01 2.58377194e-01 -4.78733212e-01 1.49799377e-01
1.28663588e+00 3.83494526e-01 4.83456701e-01 4.97209042e-01
8.27896059e-01 7.85401523e-01 9.94053781e-01 5.70479512e-01
-2.14770739e-03 1.22171617e+00 4.77272421e-01 2.58828908e-01
-1.68969542e-01 -8.14724341e-02 5.51909506e-02 6.33782506e-01
-8.28038871e-01 7.93918880e-05 -1.05311847e+00 3.19481820e-01
-1.61172903e+00 -7.95253873e-01 -4.05917019e-01 2.61682749e+00
6.78499997e-01 1.86392456e-01 5.54962814e-01 9.70271409e-01
1.02549517e+00 7.19853910e-03 -5.97582281e-01 2.46697083e-01
2.18696803e-01 4.48893905e-01 6.47031844e-01 3.85271400e-01
-1.06087995e+00 6.83892071e-01 5.76394892e+00 5.10442913e-01
-8.72609794e-01 1.51616067e-01 -4.10886794e-01 2.77724359e-02
3.01401168e-01 -5.40902376e-01 -1.11447942e+00 3.52762461e-01
1.47679135e-01 3.58308613e-01 2.16836587e-01 8.14809680e-01
1.34441331e-01 -2.27818862e-01 -9.65547562e-01 1.46795964e+00
1.30141852e-02 -5.70706785e-01 -2.35543057e-01 6.40206560e-02
3.48253250e-01 -4.40460712e-01 -1.04989149e-01 -6.50171518e-01
-5.76510787e-01 -4.86404568e-01 8.21741998e-01 4.93027151e-01
7.83246279e-01 -6.12755299e-01 7.36017525e-01 6.39514506e-01
-1.51220906e+00 2.88291842e-01 -1.95715830e-01 1.75805315e-01
2.56530911e-01 4.18133616e-01 -6.75507843e-01 2.64033556e-01
8.25477481e-01 2.82390326e-01 -3.08003843e-01 1.24012434e+00
-2.08974481e-01 -6.52999207e-02 -4.21632707e-01 -2.78801829e-01
-3.35578084e-01 1.04313254e-01 8.34210515e-01 1.07543814e+00
1.92521885e-01 -1.75998673e-01 -1.46232575e-01 4.79054630e-01
4.01107252e-01 3.45836133e-02 -3.47916037e-01 2.99520612e-01
5.14510930e-01 9.94316697e-01 -8.67172420e-01 -1.42172813e-01
-2.25157380e-01 1.26973403e+00 -3.04542452e-01 3.82445641e-02
-5.09603739e-01 -4.16804761e-01 5.64125836e-01 6.15508676e-01
-7.98876062e-02 -6.72133982e-01 -2.71504879e-01 -1.05079174e+00
4.49265420e-01 -6.97080672e-01 1.80005968e-01 -5.48480630e-01
-9.87359107e-01 4.70043600e-01 -4.32161689e-02 -1.54717374e+00
-4.39165801e-01 -9.63943541e-01 -3.11822053e-02 9.40845013e-01
-1.03182614e+00 -9.92057085e-01 -8.90218318e-01 8.67002606e-01
6.32621169e-01 -2.13558972e-01 8.97245646e-01 4.56224710e-01
-2.45322704e-01 7.97671735e-01 6.32888302e-02 1.49081528e-01
7.93649137e-01 -1.17643416e+00 2.90932626e-01 6.73886240e-01
-2.37387419e-02 6.75117016e-01 8.39954972e-01 -7.05559909e-01
-1.71250498e+00 -4.33870912e-01 6.41653895e-01 -5.29503167e-01
6.08440861e-02 -3.99930894e-01 -5.12192428e-01 3.50347340e-01
-1.84767842e-01 -1.41559690e-01 1.93171903e-01 -7.03689381e-02
-2.28351466e-02 -2.49395236e-01 -1.43288994e+00 2.83691674e-01
1.43267429e+00 -4.63157803e-01 -5.26178300e-01 2.57166147e-01
-1.62301898e-01 -9.65308666e-01 -9.78092611e-01 3.65819961e-01
1.38717484e+00 -9.88943815e-01 1.26611507e+00 -2.92898715e-03
-4.94997025e-01 -2.85270482e-01 1.73805505e-02 -7.36846983e-01
-6.98282942e-02 -3.99345547e-01 -4.22482669e-01 1.24758172e+00
-3.05257380e-01 -4.93759543e-01 1.40717244e+00 5.90321958e-01
6.10915184e-01 -2.47690052e-01 -1.19166195e+00 -1.09734344e+00
-5.07253408e-01 -2.31919140e-01 4.05509531e-01 3.52153927e-01
-2.20284685e-01 -4.56371717e-02 -2.43140563e-01 1.26643479e-01
1.46908176e+00 1.78988308e-01 8.58678758e-01 -1.65336645e+00
-1.29963845e-01 -2.02829704e-01 -7.80204535e-01 -1.25439620e+00
-2.40128720e-03 -2.93082207e-01 -6.24714606e-02 -1.44329476e+00
1.77981287e-01 -3.50410044e-01 1.15132958e-01 -1.04777165e-01
1.44321308e-01 4.03993487e-01 1.79549351e-01 4.30143416e-01
8.46407712e-02 2.69843899e-02 1.59133852e+00 -7.53287598e-02
-1.25103429e-01 5.59270918e-01 2.81528234e-01 8.47401619e-01
7.01770544e-01 -2.08028212e-01 -7.38827214e-02 -2.11934090e-01
-1.21199347e-01 1.92281306e-01 6.06803298e-01 -1.18587518e+00
4.92451310e-01 1.14832826e-01 4.39297080e-01 -1.06931388e+00
5.31698883e-01 -1.31980228e+00 3.10381889e-01 8.54287505e-01
2.67143905e-01 -1.15959771e-01 2.38735229e-02 2.97580659e-01
-9.52781588e-02 -2.16024205e-01 9.26939845e-01 -2.53824919e-01
-7.89783537e-01 2.40342021e-01 -1.88605651e-01 -2.75369227e-01
9.53628540e-01 -8.76059890e-01 9.83129218e-02 -1.62658080e-01
-7.78323948e-01 -1.51552439e-01 5.02763271e-01 5.51346838e-01
6.26577318e-01 -1.25897706e+00 -4.41164523e-01 2.76622057e-01
-3.66578028e-02 3.00975423e-02 9.83829051e-02 6.73786700e-01
-7.65099883e-01 6.94934547e-01 -5.67198038e-01 -1.00936818e+00
-1.81949210e+00 2.31556341e-01 3.27716410e-01 1.15863718e-01
-5.51878929e-01 5.01274705e-01 -4.76169407e-01 -2.75560141e-01
8.84564757e-01 -7.83746988e-02 -6.95738345e-02 1.04708873e-01
4.23309267e-01 9.62918758e-01 1.85096011e-01 -6.89305842e-01
-6.94634855e-01 1.39435422e+00 4.31834131e-01 -3.12073916e-01
8.84510100e-01 -9.41228792e-02 6.82256669e-02 2.82865137e-01
7.95108080e-01 1.45700425e-01 -1.11419237e+00 -1.01176001e-01
-1.26263961e-01 -9.26029086e-01 -3.61126252e-02 -5.68633080e-01
-1.19330943e+00 1.24568820e+00 1.39416611e+00 -2.05144376e-01
1.00093782e+00 -1.21348567e-01 8.97254825e-01 2.69810349e-01
1.22333622e+00 -1.27721941e+00 -1.81831673e-01 1.44229785e-01
8.21313620e-01 -1.06429493e+00 -1.05571831e-02 -8.85274053e-01
-5.67912273e-02 1.22982490e+00 6.91846609e-01 9.80918780e-02
5.62975109e-01 5.59663296e-01 1.05590411e-01 1.57549474e-02
4.84459937e-01 -3.70405316e-01 3.56320798e-01 7.38133252e-01
6.15440547e-01 5.22690900e-02 -5.76757729e-01 2.35991944e-02
2.24414971e-02 1.57952651e-01 -1.75238013e-01 1.23220861e+00
-3.12643975e-01 -1.52746737e+00 -9.98221219e-01 1.63210019e-01
-3.43655467e-01 6.08225942e-01 -4.30300504e-01 9.30339873e-01
1.07842959e-01 7.51742303e-01 -1.63760871e-01 -4.22322899e-01
9.13166881e-01 -5.38462326e-02 1.36340404e+00 -3.04467291e-01
-5.72245181e-01 2.87455082e-01 -2.98470259e-01 -6.81098104e-01
-6.93679512e-01 -7.82170892e-01 -1.06335914e+00 -4.95456338e-01
-4.95981514e-01 -2.67571121e-01 8.21328878e-01 5.32109618e-01
2.10158993e-02 1.09654412e-01 3.41017038e-01 -1.31380546e+00
-6.97019696e-01 -8.81985962e-01 -1.00015879e+00 1.53719530e-01
4.36580107e-02 -1.21871710e+00 -6.06353246e-02 -1.01195313e-01] | [6.559937477111816, -0.6529669761657715] |
282a7aa9-8ff5-4c0c-80dc-2041f6ac730d | a-multi-level-deep-ensemble-model-for-skin | 1807.08488 | null | http://arxiv.org/abs/1807.08488v1 | http://arxiv.org/pdf/1807.08488v1.pdf | A Multi-Level Deep Ensemble Model for Skin Lesion Classification in Dermoscopy Images | A multi-level deep ensemble (MLDE) model that can be trained in an 'end to
end' manner is proposed for skin lesion classification in dermoscopy images. In
this model, four pre-trained ResNet-50 networks are used to characterize the
multiscale information of skin lesions and are combined by using an adaptive
weighting scheme that can be learned during the error back propagation. The
proposed MLDE model achieved an average AUC value of 86.5% on the ISIC-skin
2018 official validation dataset, which is substantially higher than the
average AUC values achieved by each of four ResNet-50 networks. | ['Yutong Xie', 'Jianpeng Zhang', 'Yong Xia'] | 2018-07-23 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 5.36630988e-01 1.40108481e-01 -1.71804488e-01 -2.29904950e-01
-7.53082633e-01 -2.31251463e-01 4.79421049e-01 4.44976926e-01
-5.65941632e-01 3.54986012e-01 1.42123222e-01 -1.99980870e-01
-1.62379220e-01 -5.90364397e-01 -3.19755375e-01 -6.59274518e-01
-6.15713634e-02 -1.44544721e-01 2.90153533e-01 1.84463318e-02
-6.35094345e-02 3.60874683e-01 -1.20887804e+00 7.48376131e-01
1.29010189e+00 1.37146318e+00 -2.66797096e-01 1.14946747e+00
3.12514827e-02 9.71890569e-01 -4.39858645e-01 -6.41283035e-01
5.65801114e-02 -2.12933958e-01 -9.76672649e-01 -1.45461306e-01
8.28337908e-01 -2.24131703e-01 -2.35573322e-01 9.64700639e-01
7.29192376e-01 -1.71555459e-01 8.89819503e-01 -2.77375430e-01
-4.61535573e-01 2.34874517e-01 -5.24045408e-01 2.79491484e-01
8.38289186e-02 8.82760510e-02 7.14832664e-01 -5.61833918e-01
7.71785617e-01 8.57694566e-01 1.06537187e+00 7.32276559e-01
-1.14613640e+00 -3.45856816e-01 -1.63945571e-01 2.22937301e-01
-1.29715943e+00 2.30777398e-01 2.42560953e-01 -4.45431441e-01
7.38312304e-01 4.38215494e-01 7.44859755e-01 1.26659942e+00
7.70937026e-01 6.53616428e-01 1.58367765e+00 -2.13980675e-01
1.77034125e-01 1.28742963e-01 1.34535640e-01 7.52016962e-01
-1.67914312e-02 -3.58869918e-02 -1.66241080e-01 -1.26975685e-01
8.25458884e-01 -2.40727335e-01 2.33461604e-01 4.27262872e-01
-5.42953789e-01 6.28985465e-01 8.79636705e-01 1.77206069e-01
-7.43077815e-01 2.46517695e-02 6.49064839e-01 1.13287978e-01
9.40210998e-01 4.18399960e-01 1.66525859e-02 3.38989556e-01
-8.57568026e-01 -8.94767717e-02 5.85881293e-01 -1.31889954e-01
1.47405237e-01 -2.56562740e-01 -5.24183214e-01 1.17454255e+00
9.25216004e-02 1.42630666e-01 2.45808542e-01 -6.82776034e-01
-9.09941941e-02 7.95417011e-01 -1.65228069e-01 -5.15476644e-01
-6.83768451e-01 -8.34019184e-01 -1.27685845e+00 4.20917273e-01
3.20069462e-01 -3.17542046e-01 -1.57668972e+00 1.27860177e+00
5.06866932e-01 4.32444423e-01 -1.70951039e-02 7.53632069e-01
1.23958445e+00 1.76177487e-01 6.67620897e-01 2.43238404e-01
1.28842866e+00 -1.02455032e+00 -6.25535548e-01 2.07437947e-01
5.96114039e-01 -7.35953271e-01 5.04922569e-01 5.77226520e-01
-1.04933465e+00 -5.36840260e-01 -1.13255382e+00 6.51383996e-02
-3.26327920e-01 3.07276368e-01 6.52170360e-01 5.92509568e-01
-1.25976229e+00 6.88573658e-01 -8.11076403e-01 -4.27956194e-01
7.00187385e-01 2.53760636e-01 -3.64934683e-01 -1.16127566e-01
-1.15521193e+00 1.36329651e+00 9.20276642e-02 3.05017650e-01
-8.59089494e-01 -9.05651152e-01 -5.69444835e-01 -4.17113394e-01
-2.02942327e-01 -9.04953063e-01 8.73200774e-01 -9.63891566e-01
-1.54693747e+00 1.07555223e+00 -4.79511432e-02 -7.05183327e-01
8.38398576e-01 -1.56826358e-02 -6.78892970e-01 5.74252546e-01
-4.11303252e-01 3.94977152e-01 5.94362676e-01 -1.03923476e+00
-6.17261827e-01 -2.68248588e-01 2.40025729e-01 1.88836545e-01
-2.34300300e-01 7.38855302e-02 -2.50723451e-01 -4.27377343e-01
-3.39977592e-01 -9.11639035e-01 -7.33476043e-01 2.10474536e-01
-6.21393502e-01 -1.93301380e-01 1.52413979e-01 -1.08689678e+00
1.45517457e+00 -1.80969608e+00 4.21484513e-03 6.07808888e-01
5.54622352e-01 8.98605585e-01 -4.58194435e-01 2.20262438e-01
-2.46193573e-01 2.91663945e-01 -3.97445522e-02 -2.01767087e-01
-3.77043784e-01 -4.33527559e-01 3.88764560e-01 4.04437691e-01
4.47165579e-01 8.36342990e-01 -7.86810279e-01 -4.10190970e-01
7.71629572e-01 8.57204795e-01 -9.18826908e-02 8.75506364e-03
2.21448429e-02 3.19061875e-01 -1.23097651e-01 7.23841190e-01
1.03895628e+00 -5.98125100e-01 3.49343389e-01 -5.57189822e-01
3.30717415e-01 -1.18761472e-01 -7.44831562e-01 1.49391389e+00
-7.83774734e-01 4.09863383e-01 1.28781021e-01 -2.63905674e-01
4.93231356e-01 3.47444713e-01 5.92883050e-01 -4.34211731e-01
2.26852715e-01 3.68495062e-02 1.48111373e-01 -5.12660742e-01
4.87142988e-02 -7.08352402e-02 3.51408809e-01 -5.11128828e-02
3.30477715e-01 1.70846686e-01 1.86655328e-01 -1.92780346e-02
1.36696422e+00 -1.66209340e-01 1.46765128e-01 -2.25098833e-01
7.41648495e-01 -2.80308008e-01 1.96831793e-01 7.74195254e-01
-3.58951777e-01 5.25132716e-01 3.63570660e-01 -4.91628557e-01
-9.21939671e-01 -1.27400649e+00 -6.75617158e-01 9.20435190e-01
7.17535615e-02 -2.41065234e-01 -1.07209408e+00 -8.91927183e-01
-8.83238539e-02 3.14363778e-01 -1.25848126e+00 2.33660787e-02
3.67393792e-02 -1.13030887e+00 6.94471419e-01 5.65675795e-01
4.94492918e-01 -7.16704488e-01 2.15244964e-02 2.64180958e-01
1.02643594e-01 -1.03152227e+00 -1.37424469e-01 -1.08752807e-03
-3.97041172e-01 -1.36554861e+00 -7.95178473e-01 -6.25672877e-01
9.00914609e-01 -2.92840153e-01 7.90585458e-01 2.40550354e-01
-8.30255449e-01 5.02752140e-02 -3.03625226e-01 -5.33418655e-01
-5.60601830e-01 1.31703049e-01 -3.15819770e-01 3.03595454e-01
3.25933963e-01 -3.39950681e-01 -1.25234449e+00 -4.22474854e-02
-1.03378296e+00 5.75223416e-02 9.49146748e-01 1.02108109e+00
6.44718468e-01 -1.07381389e-01 6.57190084e-01 -1.09069836e+00
7.25066423e-01 -5.13849139e-01 -1.46333678e-02 5.93488991e-01
-4.33949262e-01 -4.78209883e-01 5.32818973e-01 -3.79443109e-01
-1.15121603e+00 8.38673040e-02 -8.83202076e-01 -1.20232306e-01
-2.21168119e-02 4.87209886e-01 8.48330200e-01 -7.45070219e-01
8.89608383e-01 -4.21822444e-03 2.36697569e-01 -2.26612538e-01
6.80946261e-02 7.71153986e-01 3.39906007e-01 1.07669562e-01
3.03106040e-01 2.65231103e-01 3.33504826e-01 -6.91674411e-01
-1.25160372e+00 -4.64027941e-01 -4.59446371e-01 -4.41633046e-01
1.21116865e+00 -8.91718566e-01 -5.46584249e-01 1.14200389e+00
-8.59795988e-01 -4.15328473e-01 -8.85333642e-02 1.81441233e-01
-1.02521665e-01 3.41391236e-01 -1.04678512e+00 -7.63215601e-01
-9.94044423e-01 -8.94362211e-01 8.54403436e-01 5.77483237e-01
-3.36075395e-01 -1.59548461e+00 1.12175211e-01 3.60415161e-01
6.67585909e-01 8.40081990e-01 6.43498957e-01 -5.95025241e-01
1.81366235e-01 -6.89822555e-01 -4.84129280e-01 8.30465138e-01
1.36941746e-01 4.33779329e-01 -1.11769986e+00 -3.00472617e-01
-4.40767646e-01 -4.98536229e-01 1.33687091e+00 7.68086314e-01
1.70838094e+00 5.38852997e-02 -1.94663957e-01 7.78324366e-01
1.77524054e+00 -1.52206913e-01 8.86110067e-01 -1.16863489e-01
3.66730779e-01 3.78838331e-01 1.46366954e-01 1.43128693e-01
3.87136281e-01 8.86399671e-02 6.81573093e-01 -8.62019420e-01
-5.60356498e-01 -2.20063269e-01 -2.57071167e-01 5.28948128e-01
-4.94538546e-01 -1.22405685e-01 -7.38758206e-01 3.62575680e-01
-1.28928697e+00 -6.64090753e-01 -1.89448237e-01 1.85474026e+00
7.82420993e-01 2.23175704e-01 1.03339255e-01 -2.91056603e-01
8.37337255e-01 3.06203544e-01 -7.24260986e-01 -7.28524864e-01
-7.28075206e-02 8.36426318e-01 6.19592369e-01 5.40565133e-01
-1.49675786e+00 6.36624992e-01 7.90877008e+00 1.18921065e+00
-1.31296742e+00 1.12643048e-01 8.41378152e-01 1.14813425e-01
-2.01407686e-01 -5.28972089e-01 -3.76320243e-01 4.53630865e-01
1.14792705e+00 4.76184897e-02 1.69150054e-01 3.85834783e-01
-1.05227962e-01 -2.02827752e-01 -4.11416799e-01 4.45038259e-01
-6.93243295e-02 -1.58274519e+00 -9.00003463e-02 1.29493713e-01
1.07063770e+00 2.06766680e-01 5.43403387e-01 -2.92669050e-03
4.81127143e-01 -1.37954271e+00 -4.45226908e-01 9.13661659e-01
1.38858616e+00 -7.18618929e-01 1.09125888e+00 1.26728211e-02
-9.79673207e-01 9.53681581e-03 -1.43992171e-01 4.86638099e-01
-1.81428879e-01 6.91029191e-01 -1.15058827e+00 6.43981218e-01
4.46726084e-01 5.25138855e-01 -9.35368836e-01 1.24528754e+00
-1.30824700e-01 8.75154614e-01 -9.26218927e-02 -2.58641988e-01
4.04517263e-01 7.64047429e-02 2.93605566e-01 1.52415943e+00
-1.33920729e-01 -2.49855533e-01 -2.65747577e-01 5.27592063e-01
-9.21170637e-02 2.71500289e-01 -1.69263691e-01 1.31923646e-01
1.29787773e-01 1.89425719e+00 -4.33401048e-01 -3.06983531e-01
-6.00447729e-02 9.24008250e-01 1.80700302e-01 2.87642211e-01
-7.70190597e-01 -6.30123913e-01 4.13147569e-01 8.94029345e-03
9.92784626e-04 4.70304728e-01 -2.34114856e-01 -7.40115285e-01
-2.68108189e-01 -6.09516621e-01 4.50455368e-01 -4.28139061e-01
-1.86475718e+00 8.93384397e-01 -6.46324158e-01 -1.05159140e+00
6.33972092e-03 -8.68120670e-01 -1.00193131e+00 1.00822473e+00
-1.56804335e+00 -1.49322188e+00 -5.56816697e-01 2.25864515e-01
1.55269876e-01 7.39052240e-03 1.22061718e+00 -3.50975897e-03
-6.78298473e-01 9.59549963e-01 2.61892468e-01 1.27020016e-01
5.98684430e-01 -1.78020084e+00 2.38813803e-01 5.83619177e-01
-6.48053885e-01 5.25091469e-01 3.01051289e-01 -5.66451967e-01
-8.28533590e-01 -1.51213241e+00 5.60740948e-01 -2.86728323e-01
8.22450161e-01 4.13224585e-02 -8.68950665e-01 2.01203957e-01
4.61335570e-01 2.36599535e-01 1.32794237e+00 2.89459020e-01
-3.95028651e-01 -2.31177956e-01 -1.59522319e+00 3.87164325e-01
6.12395227e-01 -5.45231164e-01 -1.24715501e-02 6.01799190e-01
2.24703208e-01 -7.02428520e-01 -1.76547658e+00 8.85811090e-01
8.23711574e-01 -9.05106008e-01 9.83875751e-01 -7.50381351e-01
8.42701793e-01 2.22371385e-01 2.05706209e-01 -1.46759152e+00
-4.26859856e-01 -2.71061838e-01 -1.58710420e-01 6.44555509e-01
4.69487816e-01 -6.26417994e-01 8.41911435e-01 2.23164439e-01
2.00544417e-01 -1.52740431e+00 -8.90678525e-01 -2.33427510e-01
3.35210234e-01 -1.39723957e-01 1.42779410e-01 5.90502918e-01
-2.78075248e-01 5.18315360e-02 -3.46558452e-01 -1.94195863e-02
8.78451943e-01 -5.72914064e-01 1.52854979e-01 -1.08795047e+00
6.91057816e-02 -5.84675431e-01 -5.39311469e-01 -4.63714808e-01
2.21829955e-02 -1.07951128e+00 -4.06373411e-01 -1.85458910e+00
5.07049024e-01 -2.68083990e-01 -1.05848074e+00 2.28267685e-01
-5.63628674e-01 7.21787870e-01 -1.18773133e-01 -2.06955016e-01
-7.53504336e-01 -7.25520775e-02 1.57499468e+00 -1.95396960e-01
1.54447258e-01 1.43848628e-01 -7.36100078e-01 7.91192651e-01
6.95331573e-01 1.67500451e-01 1.76416524e-02 2.98930518e-02
-2.22729668e-02 5.51399738e-02 4.40431774e-01 -1.06072736e+00
2.06495330e-01 -7.68202990e-02 9.53944147e-01 -4.69606817e-01
2.68961668e-01 -2.88125753e-01 1.50204226e-01 6.63891256e-01
-7.00056851e-01 -8.21727753e-01 3.05536717e-01 4.89090711e-01
-1.96135029e-01 -9.48085040e-02 1.22636664e+00 -1.56803951e-02
-3.48653972e-01 5.27109385e-01 -4.34220761e-01 -4.24741298e-01
1.06289625e+00 -7.16391876e-02 -4.82377648e-01 -6.93591088e-02
-1.13515079e+00 2.68238634e-01 1.38519555e-01 4.20431904e-02
5.60223222e-01 -1.30262029e+00 -1.07880688e+00 -1.74541742e-01
1.35903984e-01 -2.71720141e-01 1.22961044e+00 1.19695747e+00
-8.17168057e-01 3.64567451e-02 -3.33963066e-01 -5.90254188e-01
-1.25914371e+00 -5.40805571e-02 9.01601672e-01 -9.30975676e-01
-3.62151682e-01 1.02196407e+00 -2.48142764e-01 -4.53242391e-01
2.81725433e-02 1.06325068e-01 -4.59518492e-01 -1.99254021e-01
7.14799345e-01 5.22645891e-01 2.62607723e-01 -3.70551378e-01
-2.17513248e-01 5.60179591e-01 -5.95631480e-01 4.35772449e-01
1.24847579e+00 2.54774690e-01 -1.62842348e-01 -6.73104823e-03
1.22759974e+00 -2.81488836e-01 -1.05179012e+00 -1.98892772e-01
-4.10895258e-01 -2.85842661e-02 3.66054833e-01 -1.57851493e+00
-8.75730991e-01 8.25288892e-01 1.26570773e+00 1.74951211e-01
1.38746369e+00 -4.29527670e-01 8.19265664e-01 -1.68296248e-01
-1.69426396e-01 -1.31205153e+00 1.62681311e-01 1.63702458e-01
7.51354218e-01 -1.21484303e+00 -8.27378631e-02 -5.52383780e-01
-8.06535423e-01 1.11949217e+00 7.24953949e-01 -6.64187908e-01
7.20700324e-01 2.83427089e-01 2.04438746e-01 -8.20259750e-02
-9.40951824e-01 -2.48958841e-01 8.70011747e-01 5.46715379e-01
5.38769603e-01 1.75005108e-01 -6.24782205e-01 3.17396253e-01
3.68649572e-01 1.76340938e-01 2.05589861e-01 3.69325489e-01
-2.41765171e-01 -9.05563593e-01 2.65413016e-01 1.12109149e+00
-1.14855790e+00 4.17981483e-02 -5.66935956e-01 5.35119295e-01
3.40930581e-01 8.08351934e-01 -2.29364019e-02 -8.61631572e-01
1.52070001e-01 -2.56795883e-01 6.74277425e-01 -4.77348357e-01
-9.72266614e-01 3.73788364e-02 3.59705031e-01 -6.37884557e-01
-3.33214849e-01 -4.10863049e-02 -9.41268861e-01 -1.79740474e-01
-3.27495903e-01 -2.89520055e-01 7.79396355e-01 6.82906747e-01
2.08806977e-01 8.07832718e-01 9.12404478e-01 -3.50388139e-01
-6.91574335e-01 -1.27298236e+00 -6.67960227e-01 4.56375629e-01
2.80486763e-01 -3.21455598e-01 -3.74781460e-01 -4.60443467e-01] | [15.718025207519531, -2.9992690086364746] |
5e97b2d3-59a7-4752-9f52-56949e8734c2 | few-shot-object-counting-and-detection | 2207.10988 | null | https://arxiv.org/abs/2207.10988v2 | https://arxiv.org/pdf/2207.10988v2.pdf | Few-shot Object Counting and Detection | We tackle a new task of few-shot object counting and detection. Given a few exemplar bounding boxes of a target object class, we seek to count and detect all objects of the target class. This task shares the same supervision as the few-shot object counting but additionally outputs the object bounding boxes along with the total object count. To address this challenging problem, we introduce a novel two-stage training strategy and a novel uncertainty-aware few-shot object detector: Counting-DETR. The former is aimed at generating pseudo ground-truth bounding boxes to train the latter. The latter leverages the pseudo ground-truth provided by the former but takes the necessary steps to account for the imperfection of pseudo ground-truth. To validate the performance of our method on the new task, we introduce two new datasets named FSCD-147 and FSCD-LVIS. Both datasets contain images with complex scenes, multiple object classes per image, and a huge variation in object shapes, sizes, and appearance. Our proposed approach outperforms very strong baselines adapted from few-shot object counting and few-shot object detection with a large margin in both counting and detection metrics. The code and models are available at https://github.com/VinAIResearch/Counting-DETR. | ['Minh Hoai', 'Khoi Nguyen', 'Chau Pham', 'Thanh Nguyen'] | 2022-07-22 | null | null | null | null | ['object-counting'] | ['computer-vision'] | [ 2.12793618e-01 -3.27239782e-01 -1.51564023e-02 -3.35707724e-01
-9.92019653e-01 -3.25235754e-01 7.32248604e-01 1.40870467e-01
-7.68491447e-01 6.06631041e-01 -2.94559598e-01 3.63856316e-01
5.30716419e-01 -8.05801690e-01 -8.47696960e-01 -5.43561161e-01
2.27549434e-01 6.38341725e-01 8.21647942e-01 2.64973223e-01
1.68282226e-01 1.83750242e-01 -1.68966508e+00 1.59471944e-01
4.68773723e-01 1.12335467e+00 3.53853911e-01 9.72237706e-01
-1.48756295e-01 1.09161735e+00 -6.12699270e-01 -5.29588461e-01
4.14856851e-01 -2.21867770e-01 -3.33192676e-01 2.03654662e-01
8.14384639e-01 -7.54207194e-01 -3.10231090e-01 1.21617830e+00
4.78353858e-01 2.89806485e-01 8.40413511e-01 -1.28203750e+00
-5.17991006e-01 2.47266561e-01 -9.50033426e-01 6.57143533e-01
2.92332377e-02 6.18659616e-01 8.44305277e-01 -1.15378571e+00
4.26813930e-01 1.22817993e+00 4.65983540e-01 7.43077040e-01
-1.08093333e+00 -8.12323332e-01 -5.96207604e-02 -8.07084590e-02
-1.62058747e+00 -5.05836248e-01 1.28568932e-01 -6.57066643e-01
6.63766146e-01 -4.53278273e-02 5.68242311e-01 1.04241192e+00
-2.63126075e-01 7.38645792e-01 7.65213370e-01 -4.41947490e-01
5.47845602e-01 2.77885914e-01 6.08370960e-01 7.45089293e-01
7.46801376e-01 1.53623641e-01 -3.23178053e-01 -9.84990895e-02
4.20644075e-01 3.71692717e-01 1.80072740e-01 -2.56487727e-01
-1.07699144e+00 8.83999407e-01 3.21567178e-01 1.57902494e-01
-1.34955019e-01 4.67977256e-01 4.38362569e-01 -3.87136698e-01
5.89437306e-01 5.55062294e-02 -7.37385675e-02 3.82561833e-02
-1.04320633e+00 3.43529463e-01 7.94308245e-01 1.34521413e+00
7.38964796e-01 -1.83932129e-02 -9.55530226e-01 6.70010567e-01
-5.14784344e-02 7.25011706e-01 -1.01068830e-02 -8.95203114e-01
5.46868443e-01 5.56507528e-01 5.61268747e-01 -7.32633591e-01
-1.23072751e-01 -1.69353262e-01 -5.43522358e-01 9.35851634e-02
6.14331067e-01 -5.70608396e-03 -1.20194864e+00 1.57334828e+00
7.05052376e-01 4.47082877e-01 -5.14568329e-01 9.51738775e-01
8.81153941e-01 6.49990499e-01 2.62489289e-01 -1.17616698e-01
1.63157701e+00 -9.13936198e-01 -4.44633216e-01 -5.27465522e-01
4.43580717e-01 -2.98856884e-01 9.80029285e-01 -1.22693568e-01
-1.00149560e+00 -5.10083139e-01 -1.12861252e+00 -8.44437703e-02
-3.86317581e-01 4.33978260e-01 4.55132067e-01 5.89211524e-01
-3.53493690e-01 5.11283994e-01 -7.75915504e-01 -2.54405469e-01
1.01623929e+00 -5.80556951e-02 1.01777896e-01 -2.97790647e-01
-7.05801249e-01 8.70116055e-01 7.19507992e-01 -3.54085654e-01
-1.21943080e+00 -6.88014209e-01 -9.08976376e-01 1.47649318e-01
8.33184600e-01 -5.25412083e-01 1.42402589e+00 -2.53452986e-01
-8.20057333e-01 1.21601677e+00 -1.15392372e-01 -4.62518930e-01
7.99539208e-01 -2.17128858e-01 -4.37835678e-02 -1.07747335e-02
6.93388879e-01 6.39627635e-01 8.11571896e-01 -1.13492668e+00
-9.46661294e-01 -3.71142358e-01 -1.52065635e-01 -2.51276642e-01
-5.90167977e-02 1.44282147e-01 -5.38643837e-01 -3.34315538e-01
-3.73137861e-01 -5.97319186e-01 -6.97919680e-03 3.79438490e-01
-4.92594332e-01 -2.40946829e-01 6.13158584e-01 -1.90327853e-01
9.60229158e-01 -2.08668828e+00 -2.51066774e-01 -5.38527906e-01
3.79231334e-01 4.36157733e-01 -1.44535080e-01 4.76441532e-02
3.86517465e-01 -1.35031298e-01 -3.28348547e-01 -8.40248644e-01
3.45751308e-02 2.30602652e-01 -2.61243373e-01 7.38598943e-01
6.14415586e-01 9.70042169e-01 -1.22400987e+00 -9.40255761e-01
4.56860900e-01 3.29132527e-01 -2.35114142e-01 1.98271573e-01
-4.03415710e-01 2.63735116e-01 -2.08437234e-01 8.83224845e-01
8.87609184e-01 -4.58705008e-01 -3.73018980e-01 5.95248602e-02
2.45713238e-02 -2.89205223e-01 -1.44032872e+00 1.21865618e+00
-1.33167908e-01 3.31466705e-01 -2.84662068e-01 -5.83247244e-01
6.50968552e-01 -1.43222868e-01 2.70903379e-01 -5.46399295e-01
6.13336027e-01 1.87952399e-01 -2.69340098e-01 -3.30344588e-01
5.80081642e-01 -3.42933178e-01 -1.69576138e-01 4.31326538e-01
5.08556247e-01 -2.63733745e-01 7.67776966e-01 4.83396709e-01
1.28268898e+00 -1.47720724e-01 8.04817259e-01 1.64180428e-01
2.65841424e-01 8.58174860e-02 6.48576915e-01 1.15901840e+00
-5.75024784e-01 8.79261732e-01 2.68447548e-01 -3.96627754e-01
-1.22638869e+00 -9.74526405e-01 -2.18107224e-01 1.24102581e+00
2.97610879e-01 -1.83977857e-01 -7.21292555e-01 -7.17652857e-01
1.96702465e-01 8.97506237e-01 -9.36372280e-01 -1.89029258e-02
-3.73930484e-01 -6.42818689e-01 4.24995542e-01 6.68974638e-01
4.10608500e-01 -9.43927050e-01 -9.59740043e-01 9.93642583e-02
-1.37819901e-01 -1.58059359e+00 -6.62999630e-01 1.27430260e-01
-4.59251404e-01 -1.37301362e+00 -7.79425144e-01 -3.28863472e-01
5.75388610e-01 6.48058951e-01 1.22624791e+00 2.50963718e-01
-9.79047954e-01 2.58496404e-01 -3.04706424e-01 -9.46473539e-01
-3.34546626e-01 -2.92287380e-01 -5.64525239e-02 -5.44876568e-02
7.42510915e-01 -1.52776027e-02 -5.62038481e-01 2.53597647e-01
-6.92968905e-01 -1.56562164e-01 3.08177590e-01 6.65859520e-01
7.15620637e-01 -3.14006895e-01 2.78891474e-01 -8.35452437e-01
2.71895416e-02 -5.89650571e-01 -9.15362597e-01 1.55059427e-01
-4.82424386e-02 -2.30474010e-01 4.36316401e-01 -6.80635512e-01
-8.08884442e-01 2.33323634e-01 4.07748818e-01 -7.37821221e-01
-1.58076391e-01 -4.26454484e-01 6.47640377e-02 2.00910091e-01
8.17507744e-01 1.14102870e-01 -4.31717694e-01 -4.29176599e-01
3.66569817e-01 4.26106244e-01 7.63078749e-01 -3.37251842e-01
9.71412539e-01 8.88216019e-01 -1.09685808e-01 -6.66072547e-01
-1.54662085e+00 -1.16380990e+00 -7.41240919e-01 -2.44547457e-01
9.97859180e-01 -1.09796250e+00 -5.25574565e-01 4.92352933e-01
-1.46497595e+00 -1.45826295e-01 -7.72453785e-01 3.62525254e-01
-5.50146580e-01 1.26558572e-01 -5.36124170e-01 -1.23662090e+00
-4.79849130e-01 -5.87573826e-01 1.56722975e+00 2.72935122e-01
1.19808473e-01 -3.30455840e-01 1.38799310e-01 8.51338208e-02
1.95914991e-02 3.68763059e-01 3.78219724e-01 -7.80932009e-01
-6.93233430e-01 -6.88592911e-01 -7.08568573e-01 4.02027845e-01
-1.76510528e-01 -3.43921930e-02 -1.07815218e+00 -1.78601190e-01
1.71616133e-02 -6.55728877e-01 1.20410430e+00 5.35779595e-01
1.02996922e+00 4.12708335e-02 -3.47649574e-01 4.72129405e-01
1.70019794e+00 -1.23762280e-01 2.69057333e-01 -5.61289676e-02
6.43068492e-01 2.55626291e-01 9.18069363e-01 7.73580074e-01
2.34290630e-01 5.81258237e-01 4.71882880e-01 2.77032226e-01
-3.10007900e-01 -2.56714821e-01 -8.80862121e-03 3.86945397e-01
-1.55707836e-01 -2.96569139e-01 -8.66140485e-01 6.16379082e-01
-1.91296601e+00 -1.39937139e+00 -1.16175257e-01 2.19723868e+00
6.06425762e-01 2.56217599e-01 4.04455930e-01 -6.83679208e-02
1.06773448e+00 1.68079987e-01 -6.89755499e-01 2.15095937e-01
7.89349750e-02 -2.09581461e-02 5.91589332e-01 9.53086540e-02
-1.22245741e+00 8.11209440e-01 5.39951277e+00 8.97294700e-01
-4.62364584e-01 5.67596674e-01 5.59270382e-01 -4.79086787e-01
6.18664563e-01 -1.86110958e-01 -1.49307263e+00 7.56352842e-01
6.05638981e-01 -1.88398838e-01 2.37975538e-01 1.10210979e+00
-1.65010452e-01 -5.12843370e-01 -1.24142849e+00 9.77748394e-01
2.80156463e-01 -1.24852729e+00 -1.41641200e-01 -2.99986809e-01
7.54263699e-01 1.45696342e-01 -3.56100619e-01 7.47596204e-01
4.17924196e-01 -6.93030715e-01 1.07904243e+00 4.49287176e-01
1.02862251e+00 -3.83701146e-01 7.22172618e-01 7.76054800e-01
-1.43011224e+00 -2.72543639e-01 -7.26421654e-01 -2.00577259e-01
3.29294235e-01 5.41546643e-01 -5.81479430e-01 -1.76351760e-02
5.90942264e-01 4.72184509e-01 -6.66954637e-01 1.38240075e+00
-1.87170208e-01 3.18392426e-01 -3.81897926e-01 -2.20191136e-01
3.65759544e-02 3.49820219e-02 6.08823299e-01 1.41022360e+00
1.42295241e-01 3.20625246e-01 1.92651093e-01 1.42945647e+00
-2.81663090e-01 -2.63125658e-01 -4.91742432e-01 2.78955009e-02
7.48256505e-01 1.39334810e+00 -1.11717951e+00 -8.13264847e-01
-5.15648127e-01 7.18042195e-01 5.70003808e-01 4.38291393e-03
-1.29036391e+00 -1.54960141e-01 3.06386322e-01 2.32807174e-01
4.60748792e-01 3.48342098e-02 -1.74592271e-01 -1.28959966e+00
7.67523237e-03 -4.32685077e-01 4.43274051e-01 -5.98240554e-01
-1.50401688e+00 1.62502855e-01 2.85868615e-01 -1.26169574e+00
6.95756078e-02 -6.47764683e-01 -8.94527853e-01 5.22088945e-01
-1.26975393e+00 -1.18089485e+00 -8.44481587e-01 2.75791228e-01
8.83109093e-01 7.59872645e-02 3.98444593e-01 3.75905484e-01
-7.27541268e-01 3.54800522e-01 -9.85447690e-02 3.77828568e-01
5.83501935e-01 -1.13371623e+00 6.27796471e-01 8.78402233e-01
1.16451606e-01 2.06224501e-01 6.30224228e-01 -6.59421504e-01
-1.05907631e+00 -1.52424896e+00 3.83616418e-01 -9.00028944e-01
4.51872259e-01 -7.64959335e-01 -7.83214808e-01 5.92758894e-01
-5.79260588e-01 8.00328195e-01 1.12716325e-01 -3.65583867e-01
-3.55198950e-01 2.78809190e-01 -1.19364643e+00 1.92664459e-01
1.14953685e+00 -2.09186673e-01 -5.06549358e-01 5.81488371e-01
7.73249388e-01 -3.63390356e-01 -1.90513641e-01 2.49762133e-01
4.02244031e-01 -9.11050022e-01 1.12923813e+00 -4.48881567e-01
5.11547983e-01 -2.95368791e-01 -2.97168016e-01 -7.86299467e-01
-2.55486131e-01 1.55340314e-01 -5.68985045e-01 1.16046894e+00
-9.59258899e-02 -2.41157085e-01 7.41440773e-01 3.74968231e-01
2.12943405e-01 -5.63574493e-01 -8.82621050e-01 -1.20114052e+00
-9.68769118e-02 -4.15659457e-01 4.75668460e-01 5.62769115e-01
-5.21887779e-01 3.92443657e-01 -4.32570577e-01 6.40154034e-02
1.11646450e+00 4.10426930e-02 1.05119538e+00 -1.17094421e+00
-2.45070696e-01 -1.09679878e-01 -5.63231349e-01 -7.26216137e-01
-1.92808211e-01 -6.71970427e-01 4.15802628e-01 -1.35396957e+00
1.09349978e+00 -8.11169967e-02 7.30894157e-04 4.09440100e-01
-5.15199900e-01 5.39949954e-01 4.49027658e-01 2.58836329e-01
-1.20753574e+00 4.72347736e-01 1.02805054e+00 -2.72073328e-01
4.50732633e-02 1.68871418e-01 -2.50792474e-01 8.28882396e-01
6.11483037e-01 -9.46960926e-01 4.42069136e-02 -1.69022605e-01
-1.58355162e-01 -9.38843340e-02 8.87922764e-01 -1.28635788e+00
2.65526921e-01 -1.25605285e-01 4.90838647e-01 -8.19674730e-01
4.99287188e-01 -5.31401694e-01 -2.99461424e-01 5.79405963e-01
2.85097752e-02 -6.51520729e-01 1.09656200e-01 9.76417840e-01
2.30763137e-01 -5.14633000e-01 1.31544113e+00 -4.37434942e-01
-7.39215434e-01 5.47540665e-01 1.25040323e-01 4.71554190e-01
1.45377171e+00 -4.40537035e-01 -5.11749983e-01 1.19246878e-01
-4.70817298e-01 3.30967665e-01 3.96143943e-01 3.13333660e-01
4.11319435e-01 -1.26575387e+00 -8.94002438e-01 -5.11690304e-02
5.80183268e-01 2.41804957e-01 1.17125668e-01 8.10770452e-01
-2.67977118e-01 2.17037559e-01 -1.09438896e-01 -7.99734592e-01
-1.18269491e+00 9.05891359e-01 3.21190596e-01 -1.12864397e-01
-6.59973443e-01 1.08668911e+00 3.95032704e-01 -2.41071552e-01
3.06182742e-01 -2.48310149e-01 1.68143108e-01 9.20670554e-02
9.80872512e-01 8.39151502e-01 -2.49384254e-01 -5.34433544e-01
-3.84352446e-01 3.45716745e-01 -5.85867465e-02 2.27124527e-01
1.26921105e+00 5.20344591e-03 4.10734236e-01 8.32305431e-01
9.33852911e-01 -3.88864279e-01 -1.47711968e+00 -5.27562737e-01
-1.06481202e-02 -9.19909477e-01 -1.09588265e-01 -5.38380563e-01
-8.23102355e-01 9.39191163e-01 5.73508978e-01 -1.41652897e-01
4.57887948e-01 3.40146303e-01 5.15150368e-01 1.81674868e-01
8.04674149e-01 -1.08901644e+00 5.30885875e-01 4.97505784e-01
5.58328807e-01 -1.86744714e+00 1.76705673e-01 -3.86985540e-01
-5.54816306e-01 6.87717140e-01 8.90279830e-01 -2.50060052e-01
2.70837963e-01 5.11869550e-01 -5.05659103e-01 -4.34844971e-01
-5.85927069e-01 -6.48146868e-01 2.10160956e-01 4.23506290e-01
-2.94301398e-02 6.36307299e-02 1.78258009e-02 6.40775502e-01
3.77412409e-01 2.09650025e-01 3.85574639e-01 1.02392530e+00
-8.57296646e-01 -1.92497432e-01 -4.27731335e-01 8.46655846e-01
-3.04363012e-01 2.11229958e-02 -2.44813919e-01 8.57507765e-01
3.43677759e-01 8.02480519e-01 2.45260969e-01 -4.53437380e-02
3.35032761e-01 -8.20724070e-02 5.32287478e-01 -1.26804411e+00
-7.30412379e-02 -4.11130637e-01 -1.87344715e-01 -5.56541562e-01
-1.67523161e-01 -5.53897560e-01 -1.13460231e+00 -4.40875173e-01
-9.92827892e-01 -4.24236655e-01 4.13445592e-01 8.23229134e-01
-1.45343766e-01 4.83244687e-01 2.51081228e-01 -1.30138505e+00
-1.06410384e+00 -1.14542890e+00 -8.76764119e-01 5.18411040e-01
2.97903568e-01 -9.84355330e-01 -6.23248994e-01 -8.37942734e-02] | [9.005167961120605, 0.5657517313957214] |
5dfd99a0-bcdb-4299-b38c-0dff05814cd4 | real-time-detection-tracking-and | 1803.06077 | null | http://arxiv.org/abs/1803.06077v2 | http://arxiv.org/pdf/1803.06077v2.pdf | Real-time Detection, Tracking, and Classification of Moving and Stationary Objects using Multiple Fisheye Images | The ability to detect pedestrians and other moving objects is crucial for an
autonomous vehicle. This must be done in real-time with minimum system
overhead. This paper discusses the implementation of a surround view system to
identify moving as well as static objects that are close to the ego vehicle.
The algorithm works on 4 views captured by fisheye cameras which are merged
into a single frame. The moving object detection and tracking solution uses
minimal system overhead to isolate regions of interest (ROIs) containing moving
objects. These ROIs are then analyzed using a deep neural network (DNN) to
categorize the moving object. With deployment and testing on a real car in
urban environments, we have demonstrated the practical feasibility of the
solution. The video demos of our algorithm have been uploaded to Youtube:
https://youtu.be/vpoCfC724iA, https://youtu.be/2X4aqH2bMBs | ['Ragunathan', 'Iljoo Baek', 'Geng Yan', 'Rajkumar', 'Albert Davies'] | 2018-03-16 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [-3.52193415e-01 -1.02804996e-01 2.22738177e-01 -3.78295690e-01
-2.75656223e-01 -6.23718739e-01 4.02790815e-01 -2.92814493e-01
-4.91369635e-01 5.98580778e-01 -1.09881498e-01 -4.24134105e-01
3.42896819e-01 -6.35834157e-01 -7.04563260e-01 -6.45253837e-01
-4.29910608e-02 1.58568144e-01 1.02666271e+00 7.57321641e-02
1.91746220e-01 6.47477686e-01 -1.72138262e+00 1.81375816e-01
1.98632300e-01 7.12148786e-01 6.41218603e-01 1.04337847e+00
3.71214390e-01 7.38746226e-01 -4.84347671e-01 -1.65264420e-02
7.30943441e-01 7.51246605e-03 -4.66977507e-01 9.06988829e-02
7.09363341e-01 -9.28094327e-01 -5.24400592e-01 1.05781543e+00
4.03942972e-01 2.72393376e-01 3.20812166e-01 -1.51035583e+00
8.87648761e-02 -1.25275310e-02 -4.86247122e-01 8.83186460e-01
1.12728827e-01 3.38536918e-01 3.23373377e-01 -1.00896037e+00
6.69204652e-01 1.10717714e+00 3.98292780e-01 4.96763289e-01
-7.82476783e-01 -9.00631905e-01 -3.71350087e-02 6.33979619e-01
-1.35494339e+00 -9.97624278e-01 4.99453962e-01 -4.20600623e-01
8.71224523e-01 1.40891790e-01 7.63134539e-01 7.87115633e-01
3.90529335e-01 8.36979866e-01 4.70746636e-01 -1.85666993e-01
2.88202763e-01 3.09172899e-01 2.24362791e-01 4.84802485e-01
5.41574121e-01 1.55551627e-01 4.10235189e-02 2.80248106e-01
6.85288846e-01 2.55002767e-01 -3.08880359e-01 -5.55943787e-01
-9.84745622e-01 6.91045940e-01 2.55369842e-01 3.51671055e-02
-2.98902035e-01 4.50968683e-01 2.53406972e-01 -2.76654623e-02
3.40896130e-01 -3.75676513e-01 -2.01871365e-01 -1.18654504e-01
-8.56582165e-01 1.41005248e-01 3.20327729e-01 1.15087688e+00
6.34657383e-01 2.16094017e-01 4.36809570e-01 4.15787905e-01
5.42753041e-01 6.58958495e-01 7.22159222e-02 -1.33021379e+00
2.63069063e-01 5.14489532e-01 3.70411903e-01 -9.36925411e-01
-2.82382190e-01 5.71057573e-02 -7.11806267e-02 9.57338691e-01
6.24959588e-01 -4.25133675e-01 -8.35616827e-01 1.11298764e+00
6.29132628e-01 2.57956833e-01 5.51177636e-02 1.06480229e+00
9.11971986e-01 8.29976797e-01 3.26355696e-02 3.17652225e-01
1.33207548e+00 -8.18773389e-01 -5.14407873e-01 -2.10577220e-01
5.93772531e-01 -7.48858869e-01 4.61647958e-01 9.03881937e-02
-1.10909200e+00 -6.57270432e-01 -9.66318369e-01 2.44690385e-02
-4.49914038e-01 1.80418283e-01 -4.77227867e-02 7.12706566e-01
-1.11996126e+00 1.26735851e-01 -1.00580060e+00 -7.19766200e-01
2.95188755e-01 4.50134993e-01 -5.46454847e-01 2.65808105e-02
-1.02563596e+00 9.87406492e-01 4.48675901e-01 2.43311316e-01
-1.34637296e+00 -3.66013438e-01 -8.74560297e-01 -1.08708054e-01
3.22234660e-01 -4.16767210e-01 1.31182086e+00 -1.06235480e+00
-1.09690917e+00 6.53593123e-01 -4.86639440e-01 -6.71803117e-01
6.83870137e-01 -3.03359717e-01 -5.28939843e-01 5.60775638e-01
2.58780241e-01 9.26223993e-01 6.57693028e-01 -1.28741777e+00
-1.43156779e+00 -3.05800468e-01 1.38361529e-01 2.12604105e-01
9.79015678e-02 3.12586814e-01 -5.01569331e-01 -9.27569196e-02
-2.01579705e-01 -9.23110545e-01 4.53471430e-02 5.06098866e-02
-8.66715983e-02 -5.05603775e-02 1.62885213e+00 -6.21008098e-01
8.09047997e-01 -2.10854840e+00 -4.78205621e-01 -8.69628713e-02
1.87523410e-01 5.70376158e-01 7.78267905e-02 3.06108534e-01
-1.57102659e-01 -2.08969399e-01 4.11668152e-01 -7.86781535e-02
-4.36172038e-01 -1.73390970e-01 9.21006128e-02 8.64386916e-01
-1.05455182e-02 6.71291947e-01 -8.36061239e-01 -3.68776739e-01
8.29231799e-01 4.80892569e-01 -7.26985037e-02 -1.88713938e-01
2.16962948e-01 1.30929723e-01 -4.13580447e-01 6.01698518e-01
8.73023212e-01 2.83926129e-01 -1.76645175e-01 -1.69533476e-01
-5.39601684e-01 -2.29007661e-01 -1.48818636e+00 8.40059400e-01
3.10395518e-03 1.41710722e+00 4.98528689e-01 -7.43662894e-01
6.65562034e-01 3.26376140e-01 5.02654612e-01 -5.43509841e-01
3.72356117e-01 -1.75025277e-02 -4.08135094e-02 -6.29869878e-01
6.76789939e-01 2.01916605e-01 2.15394899e-01 2.22135141e-01
-1.60613790e-01 3.96681994e-01 3.96208584e-01 3.06054950e-01
1.14038455e+00 7.77301788e-02 1.60533682e-01 -2.58173704e-01
5.08257627e-01 3.49435776e-01 4.14780945e-01 3.32468212e-01
-7.97372878e-01 4.45750862e-01 2.25930940e-02 -5.08235097e-01
-1.23988509e+00 -9.53424096e-01 -9.63056758e-02 5.86216927e-01
6.43710971e-01 -1.40627936e-01 -7.60822892e-01 -5.07435083e-01
-1.30118519e-01 8.79076719e-01 -3.92882854e-01 1.19025968e-01
-6.79763794e-01 -1.03844598e-01 2.67478198e-01 6.00804567e-01
4.24727589e-01 -8.80961716e-01 -1.53749824e+00 1.62627861e-01
1.76280439e-01 -1.12819552e+00 -2.45873019e-01 -3.84035632e-02
-4.97142524e-01 -1.27957726e+00 -6.07505441e-01 -6.65744185e-01
8.13974261e-01 1.19228542e+00 6.47261500e-01 -9.80611295e-02
-3.14859152e-01 7.10869670e-01 -2.74243087e-01 -5.34924626e-01
-2.75738031e-01 -4.00089413e-01 2.60118961e-01 -1.22466184e-01
8.23575377e-01 -9.09705907e-02 -9.71894085e-01 5.65996051e-01
-4.62118983e-01 7.54490420e-02 1.86753616e-01 1.47911265e-01
1.31785780e-01 2.59119093e-01 1.78781614e-01 -1.88379303e-01
1.19527370e-01 -5.88111281e-01 -1.11532891e+00 -2.45222524e-01
-2.59366184e-02 -7.29548275e-01 4.66591567e-01 -1.52965128e-01
-9.45090592e-01 5.30978084e-01 1.03075378e-01 -6.10796332e-01
-7.14413464e-01 -1.80926338e-01 -3.15214157e-01 3.12914252e-02
3.77012283e-01 -1.54775772e-02 1.39492154e-01 -3.79213169e-02
3.64027053e-01 1.00610721e+00 4.19793338e-01 3.38971972e-01
4.98661071e-01 9.63576496e-01 -3.23486954e-01 -1.12261248e+00
-5.58497943e-02 -7.93144286e-01 -8.52460265e-01 -8.79754126e-01
9.88443255e-01 -1.15701020e+00 -6.98849797e-01 1.84278518e-01
-9.39581811e-01 -2.91674823e-01 1.07018333e-02 6.74677014e-01
-4.13226396e-01 3.08529317e-01 -2.42719322e-01 -8.18081498e-01
-2.29948759e-01 -1.07915735e+00 8.90449107e-01 5.73441505e-01
-9.32789743e-02 -8.32218289e-01 -2.16391772e-01 3.73182088e-01
2.35305727e-01 1.94260046e-01 -2.95764860e-03 -5.92667639e-01
-8.74120414e-01 -4.07576561e-01 -1.93104655e-01 2.07109839e-01
2.84528341e-02 3.58678281e-01 -9.02872860e-01 -1.47054106e-01
-1.62901938e-01 4.53987002e-01 8.24988008e-01 9.54043865e-01
1.58344343e-01 1.09060138e-01 -8.05459619e-01 3.26961726e-01
1.44899213e+00 8.98678660e-01 6.36988699e-01 5.97383440e-01
5.81001163e-01 5.86195648e-01 7.87177384e-01 1.42060205e-01
2.96772480e-01 6.83137238e-01 6.71969295e-01 -2.44846400e-02
-2.99935937e-01 2.12878644e-01 6.84233248e-01 8.74644965e-02
8.38605464e-02 -2.92026699e-01 -1.13147545e+00 8.96919310e-01
-1.75269389e+00 -1.36894333e+00 -6.21274829e-01 1.94308090e+00
-2.33300373e-01 2.06301838e-01 4.32802796e-01 -1.32808462e-01
1.09182394e+00 -1.14298105e-01 -3.50436300e-01 -3.51709485e-01
4.27177936e-01 -4.96823430e-01 8.79015684e-01 6.50986969e-01
-1.38942027e+00 8.22620094e-01 5.10418558e+00 4.04746503e-01
-1.28209865e+00 1.66610286e-01 4.41408485e-01 -5.07954061e-01
4.47874218e-01 -5.99758476e-02 -1.22264028e+00 6.03928268e-01
1.35405338e+00 -1.11731343e-01 -4.54895571e-02 1.09332037e+00
7.89139807e-01 -6.48796380e-01 -7.41444826e-01 7.12093711e-01
3.09582520e-03 -1.26447475e+00 -5.62380075e-01 2.28803664e-01
2.41925105e-01 4.35160607e-01 -2.25663424e-01 -3.40772159e-02
2.58320123e-01 -5.67030549e-01 1.02665555e+00 3.11422020e-01
4.13032204e-01 -8.53821456e-01 7.40634918e-01 3.60537052e-01
-1.37415457e+00 -3.50711793e-01 -5.27309775e-01 6.18956462e-02
5.43357909e-01 1.67988911e-02 -1.21437466e+00 1.86736047e-01
9.57339644e-01 7.32406437e-01 -4.98957843e-01 1.23745978e+00
9.45039690e-02 2.32538313e-01 -3.56365234e-01 -1.00941055e-01
2.61963844e-01 -2.12467566e-01 8.30500603e-01 1.09456623e+00
5.46415389e-01 1.60739988e-01 -1.38788953e-01 6.93532169e-01
5.39768159e-01 -3.03087533e-01 -8.13111305e-01 4.76333469e-01
4.49755490e-01 1.42295730e+00 -1.16660893e+00 -4.62596714e-01
-6.22983038e-01 6.11265182e-01 -2.20840394e-01 2.62006670e-01
-9.93804216e-01 -4.88016129e-01 7.40206659e-01 6.94735229e-01
9.83685851e-01 -3.69386435e-01 4.34018850e-01 -8.17193389e-01
-1.66104157e-02 -3.56494278e-01 2.45076343e-02 -9.91966069e-01
-4.79729056e-01 6.06087327e-01 3.33034098e-01 -1.63024199e+00
-2.01214850e-01 -7.08590269e-01 -7.13490725e-01 6.51796103e-01
-9.51638579e-01 -9.15345371e-01 -5.01837432e-01 6.16109550e-01
8.74703109e-01 -2.56028026e-01 3.12758088e-01 5.25622964e-01
-5.97006977e-01 -8.42333883e-02 3.31753224e-01 1.75307661e-01
4.80853140e-01 -9.03651774e-01 4.02436852e-01 1.30071747e+00
-3.52649629e-01 3.55526507e-01 7.56321609e-01 -8.05098772e-01
-1.13563001e+00 -1.31363261e+00 4.22873378e-01 -6.56716943e-01
4.38639462e-01 -3.55579197e-01 -5.79520285e-01 7.96063066e-01
5.05862176e-01 1.56293601e-01 2.72757292e-01 -7.13967204e-01
5.10836005e-01 -1.58547327e-01 -1.14734197e+00 6.50501490e-01
6.40862644e-01 -1.71231590e-02 -4.18465167e-01 -2.32461803e-02
7.90948346e-02 -4.52526778e-01 -2.82622308e-01 1.19924024e-01
7.43277967e-01 -1.14772940e+00 9.48991239e-01 -2.76647508e-01
2.28805691e-02 -8.91068697e-01 -1.10023282e-01 -9.14827824e-01
-1.36505321e-01 -4.54171225e-02 -3.93067151e-02 1.03796959e+00
4.29583937e-02 -6.49092197e-01 8.26524019e-01 6.71088398e-01
-8.35753232e-02 -2.90038049e-01 -1.06012237e+00 -6.64094329e-01
-3.66348743e-01 -4.58699167e-01 -1.92788255e-04 4.44088608e-01
-2.46987864e-01 1.80326357e-01 -2.73126721e-01 4.75482613e-01
5.39549291e-01 -1.87423512e-01 1.02977693e+00 -1.01065814e+00
4.43333775e-01 -2.83177525e-01 -1.03080952e+00 -7.19867826e-01
-1.23161897e-01 -4.04439151e-01 -8.02413747e-02 -1.65016663e+00
1.47777736e-01 3.13637704e-02 1.35650516e-01 7.77818486e-02
1.16281383e-01 1.87248304e-01 2.70891845e-01 1.63805053e-01
-6.26952350e-01 -6.61525801e-02 6.36105001e-01 9.00517106e-02
-1.62261003e-03 2.22502097e-01 -1.90864265e-01 8.54099452e-01
1.22732067e+00 -5.57685077e-01 -4.29863483e-01 -3.27095032e-01
-4.22708094e-01 -8.49741697e-02 7.16609478e-01 -1.23494387e+00
5.09261429e-01 1.44971251e-01 7.08605707e-01 -1.18425095e+00
6.48081183e-01 -1.24892402e+00 4.33357447e-01 7.32728779e-01
2.24714667e-01 3.86494100e-01 5.25460362e-01 6.17936373e-01
-4.51308908e-04 -3.88935775e-01 8.33043694e-01 -2.94165015e-01
-1.43561780e+00 -5.03073484e-02 -9.47856963e-01 -5.01132727e-01
1.97400415e+00 -8.66694212e-01 -4.06928331e-01 -5.94459713e-01
-7.25244641e-01 4.12752151e-01 6.06548607e-01 4.63354528e-01
7.99393356e-01 -1.06509030e+00 -4.98410523e-01 3.39536667e-01
-2.21582904e-01 -1.98650435e-01 4.07595575e-01 6.74895704e-01
-1.07102728e+00 7.51591682e-01 -5.86357534e-01 -7.61852026e-01
-1.96294165e+00 7.61079788e-01 5.93140781e-01 6.01496339e-01
-9.57654357e-01 6.37928367e-01 1.59263059e-01 1.58742994e-01
1.27184480e-01 -3.21933329e-01 -4.46156114e-01 7.62785301e-02
8.04680228e-01 6.99844420e-01 -1.96420118e-01 -1.21207643e+00
-6.78603232e-01 4.22299355e-01 -7.04184398e-02 -2.50867277e-01
1.25705683e+00 -5.47717273e-01 3.01235050e-01 1.43549040e-01
1.18116117e+00 8.59078243e-02 -1.47444320e+00 2.39012033e-01
-7.17249438e-02 -7.21218348e-01 1.18726000e-01 -9.42233205e-02
-1.11491871e+00 7.62500286e-01 1.22478926e+00 3.81713390e-01
8.87653112e-01 9.13617946e-03 6.10113919e-01 1.27412885e-01
3.20449233e-01 -1.01818955e+00 -4.33183372e-01 2.41174325e-01
4.57477033e-01 -1.23606133e+00 9.04861763e-02 -1.74074844e-01
-6.63175106e-01 1.40080345e+00 7.27374732e-01 -2.38060802e-01
7.24464357e-01 3.75447899e-01 2.55835414e-01 -1.88140675e-01
-7.46940255e-01 -3.89218599e-01 -1.59617558e-01 6.60782158e-01
-5.11012273e-03 -7.45246932e-02 5.07464595e-02 -1.61894232e-01
1.33873016e-01 -1.31094724e-01 7.42895901e-01 1.12219572e+00
-7.29340255e-01 -5.34012794e-01 -5.98267138e-01 1.28847376e-01
-5.33975840e-01 2.44960964e-01 -2.12541386e-01 1.13204026e+00
2.46064708e-01 1.07226229e+00 4.44093943e-01 -2.42545128e-01
3.42502177e-01 -2.05722943e-01 7.07563311e-02 -4.48830336e-01
-2.54011869e-01 2.39044026e-01 2.10398123e-01 -5.73006988e-01
-5.36010981e-01 -9.19727564e-01 -1.31102514e+00 -3.10346782e-01
-2.32344881e-01 -4.51306738e-02 7.30342507e-01 4.47289586e-01
4.72847462e-01 4.78015363e-01 3.80656421e-01 -1.20984626e+00
1.84521288e-01 -6.31559193e-01 -4.52234805e-01 -4.25871946e-02
5.45248330e-01 -6.22131705e-01 -2.76085645e-01 3.39388460e-01] | [8.015373229980469, -1.2895967960357666] |
ef8a60ad-7d08-4883-a5b7-ddbc664fd532 | dataci-a-platform-for-data-centric-ai-on | 2306.15538 | null | https://arxiv.org/abs/2306.15538v2 | https://arxiv.org/pdf/2306.15538v2.pdf | DataCI: A Platform for Data-Centric AI on Streaming Data | We introduce DataCI, a comprehensive open-source platform designed specifically for data-centric AI in dynamic streaming data settings. DataCI provides 1) an infrastructure with rich APIs for seamless streaming dataset management, data-centric pipeline development and evaluation on streaming scenarios, 2) an carefully designed versioning control function to track the pipeline lineage, and 3) an intuitive graphical interface for a better interactive user experience. Preliminary studies and demonstrations attest to the easy-to-use and effectiveness of DataCI, highlighting its potential to revolutionize the practice of data-centric AI in streaming data contexts. | ['Yuanming Li', 'Yizheng Huang', 'Huaizheng Zhang'] | 2023-06-27 | null | null | null | null | ['management'] | ['miscellaneous'] | [-1.93939470e-02 -3.99437189e-01 -7.17085674e-02 -4.77294683e-01
-4.60444570e-01 -7.16857791e-01 4.14822936e-01 5.97850621e-01
-2.85834968e-01 -7.17200860e-02 5.47251523e-01 -3.63889933e-02
-1.61638647e-01 -6.59135401e-01 -4.10593003e-01 -5.24633288e-01
-8.57670605e-01 6.71582401e-01 7.01268911e-01 -4.80523109e-01
1.97246701e-01 5.41365087e-01 -2.13526845e+00 6.60152018e-01
3.71096820e-01 1.08646178e+00 1.32816494e-01 1.43436420e+00
-1.91414684e-01 1.02825212e+00 -1.06580567e+00 -1.27615690e-01
2.99288154e-01 -2.57177860e-01 -7.02204525e-01 -2.77754426e-01
2.56717950e-01 -5.47366023e-01 -2.62947828e-01 2.77273208e-01
7.39220500e-01 -3.14140499e-01 -2.08549425e-01 -1.66729760e+00
1.16377771e-01 1.07629788e+00 -2.00747415e-01 8.41487110e-01
6.05047464e-01 1.04076254e+00 9.78207946e-01 -4.10650849e-01
9.42494333e-01 8.93832028e-01 4.22992140e-01 2.33753353e-01
-8.43229234e-01 -6.22371614e-01 -9.43947658e-02 2.61967868e-01
-1.12063813e+00 -7.87582219e-01 2.03631952e-01 -5.35077870e-01
9.92424130e-01 6.37379944e-01 1.31726074e+00 6.93179965e-01
1.02801956e-01 1.14170372e+00 7.63645843e-02 8.56706128e-02
7.48132408e-01 -4.41398561e-01 1.33890167e-01 -1.90922655e-02
-2.75826454e-02 5.65781817e-02 -1.47454953e+00 -2.81628728e-01
4.23661202e-01 -3.53695065e-01 -1.42195418e-01 -1.34458855e-01
-1.68681312e+00 3.14309210e-01 1.68911234e-01 1.83600053e-01
-7.17468441e-01 3.79527241e-01 1.25408924e+00 3.96222770e-01
2.81275123e-01 3.90198261e-01 -4.72439498e-01 -1.00935268e+00
-1.01331532e+00 8.86351168e-01 9.43795145e-01 1.55677509e+00
7.52866194e-02 2.87596762e-01 -3.19906414e-01 2.05880702e-01
-1.31116390e-01 2.63165921e-01 3.64664197e-01 -1.15351868e+00
2.73585290e-01 4.93714809e-01 -3.70505787e-02 -4.46614295e-01
-4.97628242e-01 -3.31532121e-01 -4.04739618e-01 1.76070407e-01
1.62800834e-01 -1.02875359e-01 -3.23236287e-01 1.27567863e+00
5.06864250e-01 2.08187178e-01 2.52513587e-01 1.02953255e+00
9.25504208e-01 6.58781350e-01 1.46974206e-01 -1.10153086e-01
1.09615767e+00 -5.54239333e-01 -6.82351649e-01 9.48921293e-02
6.25569642e-01 -4.20826226e-01 1.31831002e+00 8.13340306e-01
-1.78544593e+00 -3.27336758e-01 -1.02374482e+00 4.57653031e-02
-2.14859754e-01 -5.72938979e-01 8.35447967e-01 3.31760198e-01
-1.17167306e+00 4.83420640e-01 -1.27484107e+00 -4.84505862e-01
6.38047636e-01 1.49825305e-01 -1.58271283e-01 2.38973036e-01
-7.63727188e-01 1.83069542e-01 5.92021406e-01 -3.77004445e-01
-1.13677704e+00 -1.45900130e+00 -6.85136497e-01 2.32219726e-01
2.13730380e-01 -6.71554089e-01 1.70825195e+00 -7.43501604e-01
-1.33296800e+00 7.21554995e-01 1.84854805e-01 -5.74270666e-01
4.63937014e-01 -3.70801210e-01 -2.21882671e-01 6.07317805e-01
-6.20487472e-03 5.54129839e-01 3.08934540e-01 -5.88471413e-01
-1.13951039e+00 -2.25767374e-01 -2.73553312e-01 2.87040532e-01
-4.27432835e-01 4.40111816e-01 -6.77054286e-01 -6.14249647e-01
-4.27051067e-01 -3.53085428e-01 -1.23324461e-01 3.87219340e-01
7.54954144e-02 -1.81142733e-01 1.00055838e+00 -3.62781018e-01
1.30387473e+00 -2.21357846e+00 4.61391918e-03 -1.98890433e-01
1.87806755e-01 4.71824706e-01 -3.22069526e-02 1.05733991e+00
1.99748993e-01 3.12192962e-02 -1.09659351e-01 -5.15441783e-02
-9.89774764e-02 1.26714915e-01 -1.17500894e-01 1.67772561e-01
4.54483509e-01 5.28057516e-01 -1.42943668e+00 -4.13034648e-01
8.19911435e-02 1.30835578e-01 -7.64748216e-01 6.99050009e-01
-5.14223814e-01 3.48256558e-01 -2.72632539e-01 8.43252063e-01
6.56096816e-01 -1.82485193e-01 -7.44920038e-03 1.33723803e-02
-6.48141742e-01 2.18943670e-01 -1.23654652e+00 2.17830610e+00
-1.17326304e-02 8.13837290e-01 4.64750081e-01 -4.07759964e-01
7.87806988e-01 4.48591024e-01 6.73101366e-01 -5.35255134e-01
-2.08518341e-01 -3.62296514e-02 -8.17031413e-02 -8.09062898e-01
7.23859191e-01 4.60740447e-01 1.43909957e-02 6.29371703e-01
7.01535717e-02 -2.01494545e-01 7.62308657e-01 7.91099072e-01
1.74328685e+00 3.16131324e-01 -5.15173525e-02 -4.15077955e-01
-2.53048092e-02 5.04480004e-01 6.65189683e-01 7.00919032e-01
-7.85589889e-02 6.68583095e-01 4.53186691e-01 -5.50402343e-01
-1.12195289e+00 -1.08240259e+00 -9.84733850e-02 1.51043689e+00
-2.64986046e-02 -9.30755734e-01 -4.68529761e-01 -7.09742755e-02
5.28232232e-02 8.75642121e-01 -5.16376674e-01 1.14292778e-01
-4.33967382e-01 -1.35132611e-01 8.52226973e-01 6.21692061e-01
2.41304666e-01 -1.45450008e+00 -1.47629023e+00 4.26107705e-01
4.37617116e-02 -8.77547741e-01 -5.47570646e-01 8.25253651e-02
-7.49158740e-01 -1.02227461e+00 -2.81528562e-01 -4.10266846e-01
-6.06806539e-02 2.90537447e-01 1.54627240e+00 1.07307799e-01
-7.10194051e-01 6.96244240e-01 -6.88750684e-01 -6.82447255e-01
-5.04947603e-01 -6.02647513e-02 -3.53034377e-01 -4.13416594e-01
6.61958978e-02 -5.71990073e-01 -7.81789124e-01 2.17442647e-01
-1.47292030e+00 2.15949789e-01 -2.52265185e-01 2.55519122e-01
4.30682331e-01 -3.53314012e-01 8.93018246e-01 -7.93186009e-01
6.42868936e-01 -8.51998329e-01 -8.55706334e-01 -6.65309802e-02
-5.46538353e-01 -3.79566669e-01 6.99256957e-01 -1.68556094e-01
-8.83512318e-01 6.09463714e-02 -3.76042008e-01 -4.56692219e-01
-3.58518213e-01 6.38545692e-01 -1.38793990e-01 5.60963094e-01
8.07458878e-01 1.95029557e-01 2.60368139e-01 -3.45121056e-01
4.18286830e-01 6.96129739e-01 9.77298498e-01 -4.40598220e-01
2.27245495e-01 6.46180928e-01 -2.92339295e-01 -8.64995122e-01
-2.52224863e-01 -6.74673975e-01 -5.43425262e-01 -2.78177530e-01
6.54669881e-01 -1.01530945e+00 -1.04367340e+00 5.20731390e-01
-9.47135746e-01 -6.40155077e-01 -7.91566312e-01 1.09709449e-01
-8.21868539e-01 7.55493641e-02 -5.87748766e-01 -5.63391626e-01
-1.01233256e+00 -9.70822752e-01 9.71649587e-01 2.90437371e-01
-2.79437274e-01 -7.65791833e-01 3.52652192e-01 -1.32271722e-01
7.17644274e-01 4.91088361e-01 3.34760845e-01 -7.76643336e-01
-5.69332600e-01 -1.38398260e-01 5.55097871e-02 -1.09113380e-01
-2.77662814e-01 8.24776888e-01 -9.49781239e-01 -1.91774890e-01
-6.10879421e-01 -3.84418547e-01 1.27804518e-01 1.41064808e-01
1.15841794e+00 -1.61504358e-01 -1.71220168e-01 9.74242687e-01
1.13952065e+00 1.63274601e-01 5.05943954e-01 3.84667695e-01
2.70715743e-01 4.90242839e-01 1.10977590e+00 1.49263942e+00
3.91098469e-01 3.72029305e-01 6.81415319e-01 -1.12012185e-01
4.28125262e-02 -5.84829152e-02 1.85166344e-01 7.71926522e-01
5.84395468e-01 -6.09810829e-01 -9.85550523e-01 7.88932383e-01
-1.97654390e+00 -9.58987236e-01 -4.48157042e-01 2.26488113e+00
7.66059935e-01 1.08311251e-01 7.44051576e-01 1.04840815e-01
2.67962486e-01 -1.42323568e-01 -8.92365038e-01 -6.66714489e-01
-1.92309722e-01 1.17894754e-01 1.45223349e-01 -1.91685200e-01
-6.01643920e-01 8.68447423e-01 7.63556433e+00 4.15433973e-01
-1.07467151e+00 -4.28201444e-02 6.81557357e-02 -6.66477621e-01
-3.30462843e-01 -1.99120138e-02 -5.29345691e-01 6.36447072e-01
1.45052302e+00 -9.33938444e-01 1.97193608e-01 7.70305991e-01
5.57950497e-01 -2.02736631e-01 -1.32637489e+00 8.49178433e-01
-3.75516087e-01 -2.14036703e+00 -2.11174153e-02 -1.53164461e-01
1.04665644e-01 5.37063360e-01 -5.08122146e-01 -2.01501057e-01
4.20141876e-01 -5.38075387e-01 1.19174194e+00 3.22597027e-01
7.40287840e-01 -9.06051040e-01 4.17899489e-01 2.10420728e-01
-1.02726865e+00 1.05702402e-02 -2.06960052e-01 -1.02470644e-01
4.66438293e-01 4.08575207e-01 -8.53834093e-01 5.20696223e-01
1.25807703e+00 8.34141791e-01 -4.66627091e-01 1.44407582e+00
4.36960965e-01 9.75825191e-01 -4.89604115e-01 1.62243322e-01
-1.52268827e-01 3.82215232e-01 7.53282607e-01 1.96649575e+00
7.99404234e-02 1.27673551e-01 -2.24118531e-02 4.48785663e-01
4.32276785e-01 2.09297910e-02 -6.32872164e-01 -2.25771621e-01
8.83476436e-01 1.27776575e+00 -5.57475626e-01 -6.01208687e-01
-2.83031583e-01 6.32393420e-01 -1.38274983e-01 -8.24722648e-02
-5.83038032e-01 -7.11838663e-01 1.02024233e+00 2.09322423e-01
3.61159533e-01 -3.12999874e-01 -3.64988446e-01 -8.38219523e-01
-1.70933023e-01 -1.00578129e+00 9.69009876e-01 -7.55073845e-01
-1.04566562e+00 6.85992181e-01 2.25764513e-01 -1.17097676e+00
-4.38961804e-01 3.50519754e-02 -9.98654842e-01 4.39880192e-01
-9.49348629e-01 -7.56351709e-01 -8.47639561e-01 6.30141079e-01
6.37797773e-01 -1.49193972e-01 7.39592016e-01 2.96573222e-01
-6.07894301e-01 5.22185802e-01 -9.63459536e-02 -3.91427547e-01
5.64047575e-01 -1.25290918e+00 8.80896211e-01 9.59926844e-01
-4.57548350e-01 1.73368558e-01 1.04044604e+00 -5.50338626e-01
-2.45111847e+00 -8.31907988e-01 1.67050391e-01 -2.83445328e-01
7.88192272e-01 -4.79179502e-01 -1.01080704e+00 6.51992261e-01
4.37032014e-01 5.25428765e-02 9.80739534e-01 -1.33640781e-01
-1.52722701e-01 -5.95570743e-01 -1.08705497e+00 3.01884443e-01
1.01413369e+00 2.32845217e-01 -2.19793711e-02 3.95208091e-01
9.68024790e-01 -3.68711591e-01 -1.47833812e+00 1.75559476e-01
6.32915199e-01 -1.14944732e+00 6.64601505e-01 -3.78925055e-01
3.78492415e-01 -3.88375342e-01 1.46811411e-01 -1.03635502e+00
-1.47139654e-01 -1.34654701e+00 -2.36864880e-01 1.53233874e+00
1.06804609e-01 -1.22601144e-01 8.32522154e-01 5.09692848e-01
-6.46501303e-01 -2.69171447e-01 -6.77715123e-01 -5.23704588e-01
-4.26141143e-01 -6.67867899e-01 8.93588424e-01 5.67656875e-01
5.46544194e-01 -1.42911142e-02 1.03579760e-01 3.37016881e-01
4.49681342e-01 -1.29434736e-02 1.24126053e+00 -1.03886390e+00
-3.93305331e-01 -4.70678210e-01 -5.80953419e-01 -1.00800359e+00
-8.45114231e-01 -6.56961560e-01 4.65314500e-02 -1.39361525e+00
-1.61214590e-01 -6.20520353e-01 1.21348895e-01 3.31568211e-01
1.74231157e-01 -1.85833015e-02 3.36896241e-01 3.60428423e-01
-1.01528740e+00 4.78695005e-01 9.61598396e-01 3.04845899e-01
-3.33477110e-01 -1.10181846e-01 -6.37585223e-01 1.55236438e-01
6.13122642e-01 -5.68003356e-01 -5.87952316e-01 -6.99437380e-01
1.77287653e-01 3.28734338e-01 -7.18105733e-02 -1.22889352e+00
6.05162978e-01 -4.22070265e-01 -3.12937237e-02 -8.35225224e-01
1.01273321e-01 -5.48990011e-01 3.94516647e-01 5.71400762e-01
-5.68739295e-01 4.89108711e-01 5.11630297e-01 1.50527701e-01
-1.57271832e-01 -1.17548451e-01 4.88872439e-01 -2.16033868e-02
-7.49618948e-01 4.68336463e-01 -5.82471550e-01 3.98338497e-01
1.24532282e+00 -1.47923768e-01 -5.38804948e-01 -4.33407843e-01
-5.15640259e-01 7.06445992e-01 7.53191233e-01 5.27724028e-01
5.42963922e-01 -7.97409952e-01 -1.04836619e+00 4.10138965e-01
5.62068582e-01 4.10109907e-01 7.13737980e-02 4.79791582e-01
-1.08037853e+00 -2.78281365e-02 -4.37929451e-01 -9.87898231e-01
-1.17202091e+00 6.49258673e-01 -9.46707353e-02 1.98755026e-01
-1.12196064e+00 8.20462704e-01 -1.57148048e-01 -4.40855548e-02
6.92543566e-01 -2.75597066e-01 3.05879802e-01 1.24099128e-01
1.23696983e+00 6.79649532e-01 2.76483476e-01 -9.16616544e-02
-5.64081669e-01 -3.39952379e-01 1.23623097e-02 -3.22650671e-01
1.55389106e+00 -1.26138896e-01 1.14979103e-01 7.58368194e-01
5.46074569e-01 -3.06011170e-01 -1.76044071e+00 1.55127048e-01
5.44115789e-02 -5.27232587e-01 2.29692347e-02 -8.42886150e-01
-9.95169282e-01 6.84083164e-01 3.82638335e-01 5.16401291e-01
1.43947566e+00 -1.58851132e-01 7.26084530e-01 -8.69349614e-02
2.91791469e-01 -7.78633177e-01 -1.38653085e-01 1.28361419e-01
9.55293119e-01 -7.43376791e-01 -2.13280302e-02 -5.02424948e-02
-1.08305192e+00 1.34694409e+00 6.39081299e-01 1.54221058e-01
5.40943444e-01 1.04879582e+00 4.06606913e-01 -3.46575886e-01
-1.81097317e+00 -6.50530681e-02 -5.00239074e-01 9.07940626e-01
4.84244794e-01 -2.32625112e-01 -4.89777327e-02 3.57891709e-01
-2.30481967e-01 5.13772726e-01 7.20832705e-01 1.60167682e+00
-3.40779960e-01 -8.46911967e-01 -8.24012905e-02 1.83642447e-01
-3.00255686e-01 1.94114342e-01 -8.64845216e-02 4.78064775e-01
-1.62397832e-01 1.09144580e+00 6.97827160e-01 -1.89810142e-01
4.97674018e-01 -4.72364157e-01 5.77988550e-02 -4.74586785e-01
-1.50056887e+00 2.08414868e-02 1.34797618e-01 -9.52330828e-01
2.12272927e-01 -7.98476577e-01 -1.41603494e+00 -4.66749787e-01
2.32712910e-01 2.53363043e-01 1.14159358e+00 2.28745237e-01
1.14416540e+00 4.25767034e-01 5.47114909e-01 -1.07512307e+00
-1.67429417e-01 -7.60472119e-01 -4.32205915e-01 4.23756927e-01
2.52797931e-01 3.59689705e-02 7.74282729e-03 4.24462944e-01] | [8.495954513549805, -0.28072142601013184] |
843fd2ae-8f33-46ad-8561-73a5988b18ab | speech-driven-video-editing-via-an-audio | 2301.04474 | null | https://arxiv.org/abs/2301.04474v3 | https://arxiv.org/pdf/2301.04474v3.pdf | Speech Driven Video Editing via an Audio-Conditioned Diffusion Model | Taking inspiration from recent developments in visual generative tasks using diffusion models, we propose a method for end-to-end speech-driven video editing using a denoising diffusion model. Given a video of a talking person, and a separate auditory speech recording, the lip and jaw motions are re-synchronized without relying on intermediate structural representations such as facial landmarks or a 3D face model. We show this is possible by conditioning a denoising diffusion model on audio mel spectral features to generate synchronised facial motion. Proof of concept results are demonstrated on both single-speaker and multi-speaker video editing, providing a baseline model on the CREMA-D audiovisual data set. To the best of our knowledge, this is the first work to demonstrate and validate the feasibility of applying end-to-end denoising diffusion models to the task of audio-driven video editing. | ['Hugh Jordan', 'Maciej Zięba', 'Michał Stypułkowski', 'Peter Corcoran', 'Rachel McDonnell', 'Shubhajit Basak', 'Dan Bigioi'] | 2023-01-10 | null | null | null | null | ['face-model'] | ['computer-vision'] | [ 3.2647008e-01 3.5292909e-01 2.4786524e-01 -2.8974035e-01
-1.0061537e+00 -4.9002612e-01 9.5776671e-01 -5.7430899e-01
-1.2749852e-01 3.4380171e-01 5.9641683e-01 1.6875149e-01
2.0470966e-01 -1.5321331e-01 -6.7446202e-01 -5.4697239e-01
-7.2175890e-02 1.0539145e-01 -5.6440972e-02 -1.6992371e-01
4.5665773e-03 3.5321602e-01 -1.5724801e+00 5.6893545e-01
2.1664079e-01 6.2061220e-01 2.0388722e-01 1.2315065e+00
1.1953320e-01 7.1685755e-01 -5.3966951e-01 -2.1669778e-01
8.5012980e-02 -8.2555223e-01 -4.6207166e-01 5.5039555e-01
5.7108945e-01 -5.6368619e-01 -4.5174983e-01 7.3314232e-01
8.6032104e-01 2.5273946e-01 7.4678332e-01 -1.2249254e+00
-6.7596132e-01 3.7190443e-01 -4.4577816e-01 -1.5396612e-03
8.5534739e-01 -5.1755559e-02 6.7225063e-01 -1.2087892e+00
1.1678143e+00 1.4622730e+00 6.2715060e-01 8.6790991e-01
-1.3840886e+00 -4.2152745e-01 1.6591197e-02 -2.2442335e-02
-1.3281614e+00 -1.2798759e+00 1.0964469e+00 -5.0382620e-01
7.5570190e-01 4.4387687e-02 7.5501382e-01 1.4097401e+00
-8.0034733e-02 6.9997042e-01 7.3862910e-01 -5.9496206e-01
1.6048239e-01 -1.0174715e-01 -5.9271520e-01 5.6172562e-01
-6.5116626e-01 2.1891245e-01 -9.8644805e-01 -1.6000003e-01
7.7025473e-01 -5.9044027e-01 -2.7581143e-01 -4.1635838e-01
-1.0981535e+00 7.1636814e-01 -1.5928344e-01 3.9191911e-01
-3.4530768e-01 4.5585707e-01 2.3894987e-01 4.0218425e-01
7.7012867e-01 -3.7228033e-01 2.7068374e-01 -3.8043377e-01
-1.5188937e+00 6.2216964e-02 7.8377849e-01 9.2139751e-01
1.3380660e-01 5.2605689e-01 -3.0810468e-02 7.1804273e-01
5.5601406e-01 5.0839257e-01 3.4178895e-01 -1.4716344e+00
7.6157331e-02 -4.5107868e-01 -2.6270447e-02 -7.1879870e-01
-4.5721501e-02 1.5732001e-01 -5.7208216e-01 5.1301366e-01
3.2586744e-01 -2.2239971e-01 -9.9850357e-01 1.9229209e+00
3.3282003e-01 6.2708443e-01 2.3901882e-03 9.3273729e-01
7.1773130e-01 7.5267333e-01 -1.6865368e-01 -4.9130070e-01
1.0837944e+00 -6.1489028e-01 -1.0029205e+00 1.9720212e-01
1.2487337e-01 -1.0231783e+00 6.5285641e-01 6.0797322e-01
-1.5515592e+00 -7.6609844e-01 -8.5926276e-01 -3.2223278e-01
2.2422576e-01 1.3273908e-01 -4.7981799e-02 7.1502048e-01
-1.4413135e+00 2.3443259e-01 -9.3534672e-01 -5.4645187e-01
1.7970847e-01 -1.8590404e-02 -5.4164934e-01 1.6069759e-01
-9.2934906e-01 5.1879156e-01 -3.0456841e-01 -4.0062781e-02
-1.6029782e+00 -5.9852076e-01 -8.0017161e-01 -2.9136208e-01
6.9329530e-02 -8.4878057e-01 1.5284309e+00 -1.2807244e+00
-1.9899734e+00 9.1451168e-01 -6.7280632e-01 -4.9647388e-01
7.8168410e-01 -9.8605350e-02 -3.5594112e-01 5.2270550e-01
-1.8510279e-01 8.8350546e-01 1.5859959e+00 -1.4746007e+00
-2.3932733e-01 -2.0337865e-01 -2.9430014e-01 3.2507008e-01
-1.5401806e-01 2.2698033e-01 -5.5335188e-01 -1.0128225e+00
-8.2292527e-02 -8.8463616e-01 1.9366513e-01 4.9200645e-01
-2.4527611e-01 3.1711739e-01 1.0455481e+00 -1.0953671e+00
8.2064158e-01 -2.4531960e+00 4.8777860e-01 1.4300242e-01
1.7769496e-01 1.7289991e-02 -3.5264230e-01 3.8132575e-01
-2.7406517e-01 -3.1187687e-02 -1.7563440e-01 -9.5428318e-01
-1.3429363e-01 -6.4761110e-02 -3.9860532e-01 5.8186609e-01
2.4895590e-02 6.0837829e-01 -6.7585814e-01 -4.4068256e-01
2.5585821e-01 1.2217211e+00 -6.4570051e-01 1.6947776e-01
-1.2216938e-01 6.6616094e-01 2.5991997e-02 3.4040275e-01
3.6727306e-01 4.4103295e-01 4.3127246e-02 -1.9530046e-01
-4.3562237e-02 4.4820946e-02 -1.2484303e+00 2.2915151e+00
-5.5639261e-01 1.0534880e+00 7.8367484e-01 -7.3136955e-01
7.5715733e-01 8.7182450e-01 4.7424179e-01 -3.9658508e-01
9.7812125e-03 5.2454535e-02 -3.2889456e-01 -4.5333934e-01
3.6521319e-01 -4.3462625e-01 2.6920572e-01 6.2797326e-01
3.9985740e-01 -5.1636595e-01 3.9218934e-03 4.6565947e-01
6.4622170e-01 4.0730822e-01 -1.0940157e-01 5.7906177e-02
3.7930021e-01 -4.3446815e-01 -4.0066369e-02 3.3004525e-01
-1.2032590e-01 1.2028644e+00 2.9519805e-01 1.6761951e-01
-1.1588775e+00 -1.3015479e+00 2.4325684e-01 1.0596094e+00
-4.5319319e-01 -5.6924093e-01 -1.1565207e+00 -1.0088886e-01
-2.2806410e-01 8.2008958e-01 -6.1254442e-01 -3.0531175e-02
-3.6545780e-01 1.8457185e-02 7.6366651e-01 3.0815342e-01
5.0275836e-02 -8.8793582e-01 -2.2049350e-01 2.9437381e-01
-3.2003120e-01 -1.1264434e+00 -8.9535564e-01 -3.5867137e-01
-6.3888228e-01 -6.6834843e-01 -1.3081365e+00 -8.6327124e-01
5.0236547e-01 1.6341060e-01 7.7589941e-01 -3.3271161e-01
-4.8577350e-01 1.0462801e+00 3.0982606e-02 -2.6263189e-01
-9.0658402e-01 -4.9893117e-01 3.3165893e-01 2.9153922e-01
-2.7271980e-01 -8.0014825e-01 -5.2570653e-01 2.3911981e-01
-7.8425163e-01 3.5844937e-02 1.5669666e-02 6.4869249e-01
5.1280409e-01 -2.7276981e-01 5.4763889e-01 -2.8265280e-01
8.9014620e-01 -2.4945055e-01 -3.2448947e-01 -3.0732844e-02
-8.5692525e-02 -2.1695319e-01 2.1808614e-01 -6.3123327e-01
-1.4292415e+00 2.6443234e-01 -2.3514476e-01 -8.3673137e-01
-1.5099555e-01 2.0668173e-01 -6.9884658e-02 9.2167147e-02
7.4675310e-01 2.7795151e-01 4.9171010e-01 -4.1970521e-01
9.2684609e-01 7.7531511e-01 8.8707525e-01 -4.0048763e-01
5.6539267e-01 8.7014389e-01 -1.7599860e-01 -1.4125741e+00
-5.7867911e-02 -2.1212701e-01 -6.4288920e-01 -4.4674379e-01
1.0490755e+00 -1.0905479e+00 -6.1387181e-01 5.9920359e-01
-1.3464516e+00 -5.4378551e-01 -4.3467614e-01 4.9427003e-01
-1.1394087e+00 5.5186862e-01 -5.8558977e-01 -9.3893200e-01
-9.0566337e-02 -1.0436605e+00 1.2565677e+00 -1.9665977e-01
-4.2450923e-01 -1.0442674e+00 2.4429446e-01 2.4230875e-01
3.2621554e-01 1.6587460e-02 3.5155943e-01 -4.6370089e-02
-4.0903088e-01 7.2617322e-02 4.3028387e-01 5.7143104e-01
2.6797819e-01 4.3051013e-01 -1.2116355e+00 -2.6869512e-01
2.2100233e-01 -1.6580793e-01 7.3262525e-01 7.2946548e-01
5.2904952e-01 -7.6862335e-02 -1.5176432e-01 5.7453310e-01
7.1588802e-01 2.1555558e-01 5.7353336e-01 -2.7058491e-01
5.3429705e-01 6.1754286e-01 2.2675714e-01 5.1116055e-01
2.5174773e-01 7.8867722e-01 1.2266868e-01 -7.4741840e-02
-8.8942122e-01 -4.9045134e-01 6.5997589e-01 7.8874660e-01
-8.5309960e-02 -3.4591600e-01 -6.7010891e-01 8.1052226e-01
-1.4060303e+00 -1.2388470e+00 1.2942438e-01 2.0267270e+00
7.0440227e-01 -2.1194030e-01 3.0391577e-01 8.4279962e-02
7.9494202e-01 2.5629538e-01 -3.0276799e-01 -4.2534348e-01
-9.4185464e-02 2.2764485e-01 -1.7124143e-01 9.6325809e-01
-8.0471516e-01 8.5517240e-01 6.6878924e+00 6.2897772e-01
-1.1483370e+00 2.3225853e-01 3.8023105e-01 -5.3954649e-01
-3.7385041e-01 -7.8821527e-03 -4.4923267e-01 1.4625219e-01
1.0041263e+00 -3.6467248e-01 6.8686581e-01 4.1468534e-01
8.1639075e-01 7.0585556e-02 -1.2251211e+00 1.0809381e+00
5.2347273e-01 -1.2346239e+00 3.1696610e-02 4.0210757e-02
5.6792831e-01 -1.4521508e-01 3.3509347e-01 -3.2026348e-01
1.4421268e-01 -1.0252663e+00 1.1293170e+00 7.2331852e-01
1.0795884e+00 -5.8251089e-01 -1.2804070e-01 2.6750112e-01
-1.1600797e+00 2.9216078e-01 1.6181429e-01 4.2058629e-01
7.1856815e-01 1.5917762e-01 -8.6812013e-01 2.8309900e-01
5.7057482e-01 7.5505340e-01 7.4085146e-02 7.4576509e-01
-1.8296388e-01 7.0157850e-01 -3.7814555e-01 6.0903180e-01
-1.8029605e-01 3.5650771e-02 9.7802448e-01 1.2500473e+00
7.3294121e-01 8.4995180e-02 -3.5842094e-01 7.9302937e-01
1.0965335e-02 -3.9808832e-02 -8.3768332e-01 -1.6856487e-01
3.9472464e-02 1.0391386e+00 -3.9762294e-01 -1.9028515e-01
-3.0567077e-01 1.3267403e+00 -2.2079985e-01 6.6069084e-01
-8.0804485e-01 -9.9230841e-02 6.7907375e-01 2.6003438e-01
5.6970739e-01 -6.2511361e-01 1.3180618e-01 -9.9735165e-01
-9.0801209e-02 -1.0124052e+00 -3.8339432e-02 -1.2088985e+00
-9.7167021e-01 5.4951227e-01 -4.7196027e-02 -9.9352419e-01
-7.2088116e-01 -1.6280879e-01 -5.6485593e-01 8.5716963e-01
-1.0536370e+00 -1.1331421e+00 -1.5308768e-01 1.0583867e+00
6.6816568e-01 -1.3793410e-01 7.3241860e-01 1.9702718e-01
2.6119065e-02 4.7968581e-01 1.2942782e-01 -7.3262207e-02
9.4573885e-01 -6.7809200e-01 5.9173852e-01 7.3261285e-01
4.8162296e-01 4.6908006e-01 9.2457128e-01 -6.3922906e-01
-1.3632048e+00 -6.8951106e-01 7.3561579e-01 -2.5389588e-01
4.2334023e-01 -6.5747052e-01 -8.0238110e-01 7.0975631e-01
5.1247340e-01 -8.6917907e-02 6.3812023e-01 -3.9271817e-01
-8.9281380e-02 5.5850826e-02 -1.1077347e+00 6.0585988e-01
1.2493175e+00 -1.0181056e+00 -5.5066258e-01 2.2892173e-01
3.9206174e-01 -4.7930229e-01 -7.7577263e-01 -1.3828385e-01
6.4859295e-01 -9.6493763e-01 1.0677668e+00 -2.7304953e-01
1.3916717e-01 -2.6919654e-01 -1.6619945e-01 -1.4059739e+00
1.8125238e-01 -1.4339023e+00 -1.0992696e-01 1.5147895e+00
1.4215761e-01 -1.4291540e-01 5.5511183e-01 3.2443830e-01
-5.7287107e-04 1.0125907e-01 -1.2278323e+00 -5.5014718e-01
-8.9976408e-02 -7.3802996e-01 3.4104098e-02 6.9433373e-01
-1.3917954e-01 7.2121359e-02 -7.1079177e-01 4.5486666e-02
6.8842912e-01 -3.5149094e-01 1.0309503e+00 -9.9089128e-01
-3.2893777e-01 -2.2626393e-01 -3.4368467e-01 -1.0907798e+00
2.8582951e-01 -6.5616357e-01 1.6215403e-01 -1.4704772e+00
-2.4476753e-01 1.8884143e-01 3.1553197e-01 1.1915536e-01
4.4727731e-01 2.9702768e-01 3.7014082e-01 1.3444272e-01
-6.3034549e-02 7.7177352e-01 1.1017394e+00 -1.9381404e-02
-2.6693302e-01 -1.6049203e-01 -4.2932585e-01 7.4252433e-01
4.5356044e-01 -4.2631200e-01 -7.7325261e-01 -3.6760947e-01
-1.2077511e-01 6.2852591e-01 5.2836376e-01 -7.4099368e-01
3.8138095e-01 2.0017211e-01 2.5395587e-01 -1.5601858e-01
9.4337279e-01 -6.2088102e-01 4.5457932e-01 1.6183938e-01
-5.1420772e-01 -9.1651797e-02 3.2979649e-01 7.8575116e-01
-3.5371625e-01 1.7687051e-01 7.0419782e-01 1.2130600e-01
-4.9381849e-01 1.0692197e-01 -8.7298715e-01 -2.1174395e-02
1.0645002e+00 -3.5445198e-01 3.6630932e-02 -1.3185686e+00
-1.4224131e+00 -3.9704150e-01 4.6420756e-01 5.2764833e-01
8.4401262e-01 -1.3342205e+00 -9.8147333e-01 5.1968426e-01
-3.5690948e-01 -3.9888108e-01 3.9455640e-01 8.6353940e-01
-2.6098552e-01 7.2725132e-02 -1.0734817e-01 -8.1425661e-01
-1.7281052e+00 3.8136613e-01 5.2163440e-01 5.6232500e-01
-6.9255656e-01 8.7987047e-01 2.6992661e-01 1.2938833e-01
3.3405125e-01 -1.5814495e-01 2.7984136e-01 2.2573270e-01
4.6476963e-01 2.8978378e-01 -3.5694133e-02 -9.9753296e-01
-2.7841893e-01 6.7992663e-01 2.2573633e-01 -1.1656302e+00
1.1721795e+00 -4.7805965e-01 3.5553044e-01 4.4543171e-01
1.2022177e+00 4.4946668e-01 -1.5841362e+00 -6.9334134e-02
-5.9164709e-01 -3.9282504e-01 1.9181083e-01 -4.9536610e-01
-1.1065922e+00 1.2215440e+00 5.9862918e-01 -5.3244770e-02
1.0674136e+00 -6.6772930e-02 5.8579165e-01 -2.7679799e-03
8.6144678e-02 -1.0520309e+00 4.4098413e-01 1.8126580e-01
1.3066545e+00 -7.0577347e-01 -2.9668680e-01 -2.3524274e-01
-9.0718192e-01 1.1199430e+00 -5.9316926e-02 -3.5200290e-02
8.5721666e-01 5.6422079e-01 3.3806202e-01 6.9008224e-02
-7.8309852e-01 3.6760934e-02 3.1684515e-01 9.1131914e-01
5.8239669e-01 -3.7347931e-01 9.2837982e-02 1.0089080e-01
-1.4468825e-01 3.5960916e-01 5.1845115e-01 9.1697121e-01
-2.0633984e-01 -9.4352293e-01 -4.2849454e-01 -4.3189520e-01
-3.4925324e-01 -1.3461857e-01 -5.8316219e-01 6.0151839e-01
-3.3463681e-01 1.1919526e+00 1.3023312e-01 -5.1815093e-02
2.1436158e-01 4.8504344e-01 7.5357485e-01 -2.5347352e-01
-2.8038052e-01 6.3105559e-01 8.5261315e-02 -5.2347714e-01
-6.5996790e-01 -8.8465977e-01 -1.2447132e+00 -2.7006403e-01
1.9387705e-02 -1.9993556e-01 9.6561444e-01 5.6134373e-01
6.2465143e-01 3.9743349e-01 4.5115715e-01 -1.2640158e+00
-3.0598730e-01 -8.5275519e-01 -6.0999042e-01 3.8785645e-01
6.3059020e-01 -4.6585935e-01 -5.6278467e-01 7.6064575e-01] | [13.229667663574219, -0.4376731812953949] |
2f2b0383-d0a3-4d00-929c-88a5b008123e | a-fast-and-lightweight-network-for-low-light | 2304.02978 | null | https://arxiv.org/abs/2304.02978v1 | https://arxiv.org/pdf/2304.02978v1.pdf | A Fast and Lightweight Network for Low-Light Image Enhancement | Low-light images often suffer from severe noise, low brightness, low contrast, and color deviation. While several low-light image enhancement methods have been proposed, there remains a lack of efficient methods that can simultaneously solve all of these problems. In this paper, we introduce FLW-Net, a Fast and LightWeight Network for low-light image enhancement that significantly improves processing speed and overall effect. To achieve efficient low-light image enhancement, we recognize the challenges of the lack of an absolute reference and the need for a large receptive field to obtain global contrast. Therefore, we propose an efficient global feature information extraction component and design loss functions based on relative information to overcome these challenges. Finally, we conduct comparative experiments to demonstrate the effectiveness of the proposed method, and the results confirm that FLW-Net can significantly reduce the complexity of supervised low-light image enhancement networks while improving processing effect. Code is available at https://github.com/hitzhangyu/FLW-Net | ['Chunhui Wang', 'Guohui YANG', 'Yanwu Xu', 'Yue Wang', 'Yong Li', 'Rao Fu', 'Junde Wu', 'Xiaoguang Di', 'Yu Zhang'] | 2023-04-06 | null | null | null | null | ['image-enhancement', 'low-light-image-enhancement'] | ['computer-vision', 'computer-vision'] | [ 3.42415124e-01 -7.14026570e-01 6.45305812e-02 -4.11522478e-01
-5.79712093e-01 -8.43570977e-02 1.26071349e-01 -2.26089701e-01
-6.13977730e-01 6.73310459e-01 -4.56913114e-02 -1.09057061e-01
-4.04749438e-02 -8.61109436e-01 -4.36267167e-01 -8.75352442e-01
4.67607439e-01 -7.36951470e-01 3.94804180e-01 -2.02781558e-01
2.41226643e-01 3.79329801e-01 -1.60841537e+00 1.21707447e-01
1.04814756e+00 9.76618767e-01 5.80067515e-01 3.96864325e-01
5.29147033e-03 6.92364812e-01 -2.56065547e-01 -8.58534351e-02
4.75300163e-01 -4.00840342e-01 -3.41044337e-01 -1.16743660e-02
4.91140455e-01 -6.70435250e-01 -3.11206430e-01 1.38267851e+00
1.02206779e+00 2.58573562e-01 2.26183638e-01 -9.99207020e-01
-7.21358120e-01 1.17824025e-01 -9.12053764e-01 3.68461996e-01
-4.83444557e-02 3.23807150e-01 5.93548298e-01 -1.09064043e+00
3.19185853e-01 8.62321913e-01 7.12767422e-01 4.44864959e-01
-9.57575262e-01 -8.63394499e-01 -3.35901752e-02 2.69696742e-01
-1.18239868e+00 -5.31142294e-01 8.83770466e-01 8.67964327e-02
7.39908338e-01 2.84940377e-02 5.42660296e-01 4.96594906e-01
1.83955282e-01 4.78168517e-01 1.48102355e+00 -5.21325529e-01
-2.65082233e-02 -4.59116697e-02 5.37871495e-02 8.22308958e-01
4.52053547e-01 2.32298344e-01 -5.21464229e-01 3.86115879e-01
6.90426230e-01 2.66179919e-01 -4.90189463e-01 7.79163092e-02
-7.39832103e-01 2.92550802e-01 6.59436345e-01 2.75559843e-01
-3.14826071e-01 9.24221799e-02 2.03751966e-01 1.12143224e-02
5.33460021e-01 1.81879893e-01 -1.67087331e-01 3.04989386e-02
-6.68189466e-01 -1.27813056e-01 8.95579904e-02 5.12763619e-01
1.02453208e+00 1.16768092e-01 -1.77518472e-01 1.06614292e+00
1.61993727e-01 6.60556853e-01 1.70381486e-01 -9.97848809e-01
3.89074445e-01 4.97056097e-01 2.77828313e-02 -1.26890051e+00
-4.16217119e-01 -2.88404673e-01 -1.07002389e+00 5.21436930e-01
2.06647635e-01 -1.55830890e-01 -7.08459973e-01 1.65255702e+00
3.26159894e-01 4.65925187e-02 4.44981307e-02 1.15434706e+00
1.03203797e+00 8.29143584e-01 5.06653450e-03 -3.94873291e-01
1.27314293e+00 -1.07939112e+00 -9.95175183e-01 -1.64033309e-01
2.37278253e-01 -1.07204342e+00 1.28895974e+00 1.40456587e-01
-1.47986913e+00 -6.37120247e-01 -1.17422807e+00 -3.88380378e-01
-2.94928938e-01 4.45915908e-01 7.50592232e-01 5.78372955e-01
-9.97434080e-01 4.79270488e-01 -6.58359706e-01 2.82169916e-02
4.76135939e-01 3.87097657e-01 -8.38550329e-02 -3.78137797e-01
-1.08604860e+00 6.04367733e-01 2.55768925e-01 4.92451042e-01
-4.82250780e-01 -6.32388830e-01 -7.48215199e-01 1.55628910e-02
4.34260607e-01 -3.96779269e-01 9.62335944e-01 -7.54651189e-01
-1.48990083e+00 4.32170153e-01 -3.28970075e-01 1.20057151e-01
1.94039404e-01 -5.77944070e-02 -3.87777507e-01 2.55577475e-01
-1.20303869e-01 4.56537783e-01 7.00292528e-01 -1.30721760e+00
-6.49452984e-01 -2.47215182e-01 9.85963196e-02 2.81995267e-01
-6.67143881e-01 2.58941501e-01 -6.49953663e-01 -5.03572524e-01
1.05699353e-01 -5.42699397e-01 -1.00110628e-01 2.66064346e-01
-3.33839655e-02 -2.71080360e-02 9.14334357e-01 -5.71300685e-01
1.23330081e+00 -2.26408029e+00 -6.60665214e-01 -8.50152150e-02
3.51034343e-01 6.99912667e-01 -3.43767852e-01 -1.88496839e-02
7.03520849e-02 -1.90210100e-02 -2.78016001e-01 -1.98735088e-01
-3.59046698e-01 -3.96080874e-03 -3.26441065e-03 5.09463370e-01
1.80906534e-01 8.60368013e-01 -7.89100111e-01 -5.38591564e-01
4.90484565e-01 1.03512609e+00 -2.77277470e-01 1.28200695e-01
1.64930820e-01 6.67825863e-02 -3.14345598e-01 7.36896276e-01
1.16944051e+00 -1.52056068e-01 -7.20907077e-02 -7.26165771e-01
-5.05548477e-01 -9.19829533e-02 -1.32095945e+00 1.27174723e+00
-6.32110894e-01 6.05754495e-01 3.21088493e-01 -6.00162148e-01
8.94913018e-01 -3.13502625e-02 3.33896369e-01 -1.23511589e+00
5.29860079e-01 2.60106802e-01 -2.37526298e-01 -5.78120649e-01
4.83425975e-01 -2.49029115e-01 7.50465572e-01 5.41469336e-01
-2.91827768e-01 1.32276610e-01 3.53802145e-01 -2.11612567e-01
7.27541268e-01 9.46454033e-02 1.24503553e-01 2.81972880e-03
6.84139013e-01 -5.92504203e-01 7.58275270e-01 5.72276294e-01
-4.15102184e-01 5.06478250e-01 -1.54073060e-01 -4.73357558e-01
-7.44111180e-01 -8.15331817e-01 -1.51149660e-01 1.04369760e+00
6.53878093e-01 -3.05497110e-01 -5.68683863e-01 -2.03136370e-01
-4.55250204e-01 2.95222104e-01 -1.44496411e-01 -1.15857489e-01
-6.27166510e-01 -9.94466066e-01 1.48392513e-01 3.85198474e-01
1.09212458e+00 -1.05495024e+00 -6.34870708e-01 -6.65746629e-02
-4.81419653e-01 -1.18662333e+00 -5.66766858e-01 3.62980440e-02
-7.47347534e-01 -1.03216290e+00 -6.51499808e-01 -9.76042926e-01
8.91762733e-01 9.16474164e-01 7.87238717e-01 2.94171959e-01
-5.38909674e-01 4.40563746e-02 -1.52475089e-01 -3.49659353e-01
-1.05219662e-01 -1.99095994e-01 -1.04427397e-01 2.13942584e-02
3.32397938e-01 -4.57556307e-01 -1.12907588e+00 4.28461552e-01
-1.09061635e+00 1.91961646e-01 7.94840276e-01 8.10875475e-01
7.96100259e-01 5.62930405e-01 2.82023221e-01 -5.30502319e-01
7.44057000e-01 2.65925109e-01 -7.88415432e-01 2.63753563e-01
-9.52689886e-01 -2.49842793e-01 7.78905571e-01 -2.89931715e-01
-1.50636971e+00 -1.17989235e-01 -2.43654311e-01 -3.76862660e-02
-1.08707845e-01 9.87596214e-02 -1.69190019e-01 -6.19573116e-01
4.67911452e-01 2.56517023e-01 4.49936539e-02 -3.70352805e-01
1.24802224e-01 6.97260916e-01 3.98168772e-01 -2.93880701e-01
8.15479696e-01 6.52177572e-01 2.31952325e-01 -8.57012331e-01
-9.88439441e-01 -3.73833746e-01 -2.47424349e-01 -4.16741163e-01
7.76528835e-01 -9.92703259e-01 -9.01350558e-01 7.72928059e-01
-9.28223431e-01 -3.76502842e-01 -9.04952586e-02 4.98576850e-01
-2.15111986e-01 5.84308982e-01 -8.36565673e-01 -6.82085395e-01
-7.01571465e-01 -1.15493774e+00 6.16850734e-01 6.29623353e-01
6.08277321e-01 -8.39577675e-01 -2.60351986e-01 3.96494418e-01
8.03541124e-01 4.11424832e-03 5.65960169e-01 6.50564075e-01
-7.99208403e-01 1.73501968e-02 -8.89887869e-01 7.02257872e-01
4.55802441e-01 -3.89704816e-02 -1.02177298e+00 -3.49140584e-01
7.88152069e-02 -3.79725277e-01 9.66442287e-01 6.39915407e-01
1.33058882e+00 -3.07987612e-02 -1.61077641e-02 1.05676687e+00
1.84508467e+00 2.70534843e-01 7.21855223e-01 3.91606241e-01
5.09612858e-01 3.67891699e-01 6.32883489e-01 3.52894515e-01
2.51808882e-01 4.83000875e-01 4.17212218e-01 -8.01777065e-01
-5.45704007e-01 4.40485068e-02 2.17323303e-01 9.33373153e-01
-3.83586466e-01 5.33222081e-03 -4.69031751e-01 4.04865742e-01
-1.42738748e+00 -9.79319513e-01 -1.31916583e-01 1.99731231e+00
9.49740112e-01 -3.10892351e-02 -1.83457345e-01 1.46409437e-01
7.78824151e-01 2.74355322e-01 -5.17042875e-01 -3.26199643e-02
-2.99464345e-01 8.79027173e-02 5.65786302e-01 5.88206887e-01
-9.35117602e-01 6.47526681e-01 6.05139542e+00 9.01399374e-01
-1.35194266e+00 2.04132631e-01 6.14562929e-01 -2.05490693e-01
-1.78743720e-01 -2.00853497e-01 -8.93099666e-01 5.44270217e-01
2.28916511e-01 5.72933964e-02 6.05040133e-01 4.58978891e-01
5.36770821e-01 -2.95770079e-01 -3.24913919e-01 1.20549798e+00
2.35933796e-01 -1.02887762e+00 -1.44179195e-01 -1.55183524e-01
7.26318657e-01 1.17134508e-02 2.14020416e-01 -1.34440899e-01
-2.10825294e-01 -5.85042953e-01 2.59297460e-01 2.96321839e-01
9.38575983e-01 -7.58194506e-01 6.85073256e-01 5.17019928e-02
-1.32379961e+00 -1.44351229e-01 -6.89353585e-01 -3.96792172e-03
1.38141781e-01 8.53799522e-01 -1.82667688e-01 4.24474597e-01
8.25359404e-01 7.46514797e-01 -6.05759382e-01 1.05227613e+00
-3.88728112e-01 4.26549107e-01 -2.22258240e-01 -8.14169794e-02
1.23212174e-01 -3.42664421e-01 1.98179409e-01 1.17329931e+00
2.72397518e-01 2.98217803e-01 1.84748217e-01 8.63858640e-01
-1.50133461e-01 1.79897502e-01 -3.87111008e-01 2.17179641e-01
7.69937783e-02 1.60090017e+00 -7.90105462e-01 4.97645475e-02
-6.95239604e-01 7.09690690e-01 -2.64913589e-02 4.73396182e-01
-8.09680760e-01 -8.39619339e-01 4.37120885e-01 3.03288788e-01
-7.78868347e-02 -2.27962941e-01 -2.62905836e-01 -1.08607221e+00
3.33335221e-01 -6.53347969e-01 1.38523042e-01 -1.10641468e+00
-9.68257487e-01 6.13048732e-01 -3.51027757e-01 -1.25850773e+00
6.57789826e-01 -7.18150020e-01 -5.34545124e-01 8.51900458e-01
-2.45545411e+00 -1.07402086e+00 -8.82789671e-01 7.36118734e-01
4.68395501e-01 1.49573451e-02 2.37749115e-01 6.62922680e-01
-6.42180741e-01 4.40338224e-01 1.15332440e-01 1.42340614e-02
8.88421535e-01 -8.09735298e-01 -1.21603146e-01 1.24296737e+00
-3.92194003e-01 6.11834884e-01 2.38891751e-01 -3.64394814e-01
-1.24399817e+00 -1.04939771e+00 6.49824858e-01 2.45738640e-01
2.45924056e-01 -1.05602205e-01 -8.32829535e-01 9.12842602e-02
2.71899283e-01 2.97299236e-01 5.21736920e-01 -2.29560018e-01
-1.84343547e-01 -6.43452585e-01 -1.09932208e+00 7.61943460e-01
9.60567534e-01 -3.90763849e-01 -3.51919271e-02 1.62872121e-01
4.78169858e-01 -2.79649198e-01 -6.52835965e-01 6.00526333e-01
5.55950046e-01 -1.28285122e+00 1.00398827e+00 2.92233437e-01
6.37610555e-01 -6.21360958e-01 -5.87171689e-02 -1.06887376e+00
-4.08893198e-01 -5.02030373e-01 2.19286323e-01 1.29180050e+00
1.21751577e-01 -9.09679890e-01 6.40811861e-01 3.44667256e-01
-5.44737950e-02 -9.11218166e-01 -5.02478957e-01 -6.41395807e-01
-4.08867210e-01 -2.74971515e-01 4.49615151e-01 7.53764212e-01
-3.85591835e-01 2.53504038e-01 -4.28796887e-01 1.90979078e-01
9.30211127e-01 3.23603749e-01 4.01593417e-01 -8.10219765e-01
-7.10976869e-02 -3.21388006e-01 -2.08017379e-02 -9.76383626e-01
-2.08649747e-02 -6.91433311e-01 3.05360883e-01 -1.83297896e+00
6.38410270e-01 -3.52723211e-01 -5.42274177e-01 7.04729140e-01
-3.74693155e-01 6.86634660e-01 1.68006092e-01 1.07491709e-01
-6.72683239e-01 6.18715703e-01 1.58651125e+00 -1.24499768e-01
-2.46107310e-01 -2.79832751e-01 -8.10579479e-01 7.84116387e-01
9.51856315e-01 -2.67460436e-01 -4.63981509e-01 -7.22232878e-01
3.31637651e-01 -2.99189508e-01 3.58355343e-01 -1.10198903e+00
3.46181065e-01 -2.09210128e-01 5.47339737e-01 -5.81686199e-01
3.37954015e-01 -8.77851129e-01 -3.22816670e-01 3.93671721e-01
-2.74935421e-02 -7.66867474e-02 1.71527222e-01 1.95369363e-01
-3.56222183e-01 -2.20835671e-01 1.26671922e+00 -1.29916564e-01
-8.50423992e-01 4.14247185e-01 -9.57797244e-02 -1.07062615e-01
9.22864258e-01 -3.24586928e-01 -7.41612375e-01 -2.75356591e-01
-1.02247238e-01 1.22557819e-01 4.52553511e-01 2.00177446e-01
9.71868336e-01 -1.19267738e+00 -4.39936429e-01 4.91017908e-01
3.41278762e-02 -2.38829523e-01 5.75396180e-01 9.37081814e-01
-6.87046468e-01 1.18333682e-01 -5.11871099e-01 -2.58695334e-01
-1.63494682e+00 4.12533611e-01 4.69668686e-01 4.19873372e-02
-4.77473050e-01 6.43712401e-01 3.08750182e-01 -1.98420539e-01
5.46074212e-02 -2.08391473e-01 -9.84003395e-02 -4.30611640e-01
8.72414410e-01 6.35004342e-01 4.61690687e-02 -4.02081758e-01
-1.18956350e-01 9.79276359e-01 -3.17972824e-02 4.26203847e-01
1.45572054e+00 -4.79139864e-01 -3.80022168e-01 -2.33896840e-02
1.06382167e+00 4.53565978e-02 -1.33605826e+00 -2.14324385e-01
-7.33079970e-01 -8.28925550e-01 5.33294737e-01 -7.41216898e-01
-1.37752390e+00 1.07039404e+00 1.00139046e+00 8.40032380e-03
1.85996032e+00 -4.55685556e-01 1.04843044e+00 3.75621825e-01
1.11381851e-01 -1.27215922e+00 2.84215838e-01 3.46845597e-01
5.75329185e-01 -1.37778282e+00 2.70726234e-01 -6.43185854e-01
-2.24982113e-01 9.94877100e-01 8.37496400e-01 2.64478803e-01
6.64057195e-01 5.52172363e-01 3.31061661e-01 -1.05369642e-01
-4.17595297e-01 -4.82865542e-01 2.54568160e-02 5.46734869e-01
4.95631695e-01 -4.09945041e-01 -4.70065325e-01 2.34211624e-01
4.33535725e-01 2.84441352e-01 4.83345330e-01 8.37442398e-01
-5.87351799e-01 -1.01273787e+00 -2.98513293e-01 3.24080884e-01
-8.08936894e-01 -2.53853559e-01 1.38376042e-01 4.72290635e-01
3.87392074e-01 1.07248223e+00 -2.60380089e-01 -2.73030072e-01
3.28007281e-01 -4.64731157e-01 4.38651294e-01 -1.28702611e-01
-2.41718337e-01 2.46250153e-01 -3.05703461e-01 -5.95923066e-01
-8.46520782e-01 -6.08108975e-02 -1.01125693e+00 -3.84745002e-01
-5.22975683e-01 -2.83901036e-01 7.22281873e-01 4.56309795e-01
3.59231085e-01 5.19780397e-01 7.87544429e-01 -8.15878451e-01
-2.51923144e-01 -7.10885525e-01 -7.80661225e-01 4.91213083e-01
3.69082451e-01 -6.08365953e-01 -4.30164367e-01 8.85735154e-02] | [10.790146827697754, -2.472576856613159] |
217f057f-87ea-4233-b6c6-24ce778cf298 | lps-lt-edi-acl2022-an-ensemble-approach-about | null | null | https://aclanthology.org/2022.ltedi-1.24 | https://aclanthology.org/2022.ltedi-1.24.pdf | LPS@LT-EDI-ACL2022:An Ensemble Approach about Hope Speech Detection | The task shared by sponsor about Hope Speech Detection for Equality, Diversity, and Inclusion at LT-EDI-ACL-2022.The goal of this task is to identify whether a given comment contains hope speech or not,and hope is considered significant for the well-being, recuperation and restoration of human life.Our work aims to change the prevalent way of thinking by moving away from a preoccupation with discrimination, loneliness or the worst things in life to building the confidence, support and good qualities based on comments by individuals. In response to the need to detect equality, diversity and inclusion of hope speech in a multilingual environment, we built an integration model and achieved well performance on multiple datasets presented by the sponsor and the specific results can be referred to the experimental results section. | ['Yue Zhu'] | null | null | null | null | ltedi-acl-2022-5 | ['hope-speech-detection'] | ['natural-language-processing'] | [-5.00387728e-01 3.80916297e-01 -6.24540687e-01 -3.92817706e-01
-8.35227132e-01 6.87869266e-02 7.47154832e-01 5.78817725e-01
-3.96203816e-01 9.73898590e-01 1.51563215e+00 2.55033281e-02
-1.06496751e-01 -4.21073109e-01 1.39569700e-01 -1.77724868e-01
6.35409653e-02 3.97317529e-01 -4.53803539e-01 -6.05522215e-01
3.39078337e-01 1.39055967e-01 -1.17909014e+00 4.02933240e-01
7.45473742e-01 3.09855044e-01 -6.03154659e-01 5.82157135e-01
3.83555889e-02 1.04637671e+00 -4.97489691e-01 -6.95720196e-01
-3.68105620e-01 -4.60166931e-01 -1.03277528e+00 -2.66830683e-01
3.86857152e-01 -2.59075999e-01 1.05477065e-01 9.45440948e-01
9.03567851e-01 2.35851079e-01 5.12752712e-01 -1.31501234e+00
-6.89188838e-01 1.11903632e+00 -8.12264085e-02 1.27403289e-01
8.57187033e-01 -1.23438098e-01 1.14182377e+00 -1.17932069e+00
8.98260176e-01 1.55603230e+00 8.89124513e-01 7.25131273e-01
-9.31926847e-01 -7.05380917e-01 -3.49400669e-01 1.72971115e-01
-1.14855886e+00 -1.30458820e+00 7.32552290e-01 -4.11344826e-01
1.04444158e+00 7.03170717e-01 9.48295474e-01 1.33991659e+00
1.14638664e-01 5.66132247e-01 9.02242124e-01 -3.79038364e-01
2.30128765e-01 3.94452155e-01 3.00025612e-01 3.22462440e-01
5.93490675e-02 -3.41446787e-01 -1.09389746e+00 -3.24345410e-01
-3.60299908e-02 -3.25119197e-01 -2.28437006e-01 3.70959461e-01
-1.36689270e+00 1.06268668e+00 1.90325215e-01 6.03173137e-01
-2.20375434e-01 -3.09199542e-01 8.36653233e-01 3.57737422e-01
8.55063856e-01 1.55241892e-01 1.90463766e-01 -8.36588621e-01
-1.02485263e+00 2.02212214e-01 9.49797273e-01 2.21271396e-01
9.74328816e-02 -1.46852374e-01 -4.18016791e-01 1.12453032e+00
4.41309661e-01 4.99489129e-01 4.10774380e-01 -1.11034071e+00
3.24965209e-01 6.01506591e-01 -1.96252406e-01 -1.12076151e+00
-7.28150547e-01 -8.90715778e-01 -8.76116931e-01 1.91958144e-01
2.60051101e-01 -2.54093289e-01 3.09119076e-02 1.91532183e+00
1.46892220e-01 -5.04332483e-01 3.38079512e-01 7.90819049e-01
1.20242798e+00 2.76704550e-01 1.13760412e-01 -5.02014339e-01
1.19429433e+00 -8.17389250e-01 -1.21172333e+00 -3.28133374e-01
1.02055466e+00 -1.07648325e+00 1.07457280e+00 3.31334442e-01
-1.18760896e+00 -2.06905439e-01 -1.14610064e+00 -3.79067838e-01
-6.95706978e-02 -6.61708042e-02 4.19311821e-01 8.77714753e-01
-1.09935164e+00 4.42917168e-01 -1.37306895e-04 -8.50617290e-01
4.14491653e-01 -1.68254122e-01 -7.71231771e-01 1.42133087e-01
-1.41749966e+00 9.11980867e-01 -9.95941609e-02 -1.49074599e-01
-4.74738121e-01 -3.12475443e-01 -7.24905610e-01 -2.32518360e-01
-2.13850766e-01 -6.30448163e-01 6.86443985e-01 -8.60450327e-01
-1.00618207e+00 1.34045982e+00 -1.51310369e-01 -4.99473989e-01
7.03879416e-01 -3.41733396e-01 -6.54586971e-01 -4.18222278e-01
3.86390626e-01 5.92402995e-01 3.03780794e-01 -8.18694711e-01
-4.34510261e-01 -5.46880305e-01 -9.13681984e-02 3.50803494e-01
-7.79354870e-01 6.49714351e-01 5.96136451e-01 -3.34069639e-01
2.93559488e-02 -5.41236222e-01 2.21626535e-01 -1.70980543e-01
-5.84882855e-01 -4.12477285e-01 7.69282222e-01 -9.24232244e-01
1.48726833e+00 -2.36671782e+00 -2.04846784e-01 -2.75625795e-01
5.11001706e-01 -7.87027478e-02 3.21332932e-01 8.50860298e-01
-5.97011261e-02 5.21709263e-01 3.16541910e-01 -7.03649938e-01
1.88561901e-01 7.83370435e-02 -1.02089807e-01 7.27032423e-01
-2.79912055e-01 5.39034545e-01 -9.51528311e-01 -7.40691483e-01
-2.74863571e-01 5.13026237e-01 -7.08625197e-01 -1.39483079e-01
5.21398246e-01 3.03128153e-01 3.05988807e-02 5.57993710e-01
5.06924331e-01 -3.17096747e-02 1.07811250e-01 9.09324512e-02
-5.19950509e-01 8.20872247e-01 -8.49674940e-01 1.41003501e+00
-9.98740941e-02 7.56636798e-01 2.03409165e-01 -7.78961420e-01
1.16875756e+00 4.18658227e-01 3.09407264e-01 -7.61971354e-01
9.42957029e-02 3.74393553e-01 2.53345907e-01 -4.33070689e-01
7.27647960e-01 -4.14115429e-01 -2.37732261e-01 3.66254270e-01
-4.63250101e-01 -1.57099310e-02 -1.38857663e-01 5.06436229e-01
8.29596460e-01 -5.50907314e-01 4.93475318e-01 -4.90615755e-01
5.91006577e-01 -2.43844643e-01 8.45380485e-01 3.39340955e-01
-7.50746667e-01 3.58615488e-01 5.45454025e-01 -1.56252131e-01
-9.97614861e-01 -9.93708253e-01 -3.79158169e-01 1.11131811e+00
-5.05379736e-01 -6.77382946e-01 -3.27900141e-01 -2.76060075e-01
-4.59404737e-01 1.26201594e+00 -4.44902271e-01 -2.16980413e-01
-3.47259015e-01 -2.42191270e-01 7.32790232e-01 4.12866101e-02
5.79872668e-01 -7.61232913e-01 -4.64149296e-01 3.14160511e-02
-9.42197025e-01 -8.35849762e-01 -3.68709862e-01 -1.83573961e-01
-5.54029405e-01 -5.89788079e-01 -7.53951907e-01 -6.09668732e-01
1.47435069e-02 -2.14073658e-01 1.03422487e+00 1.51337996e-01
-3.61073986e-02 3.17226678e-01 -3.67913514e-01 -2.92764783e-01
-7.56626785e-01 -1.00347102e-01 2.50590593e-01 -3.84896666e-01
1.12049967e-01 -6.20187819e-01 -4.71466929e-01 3.50860469e-02
1.81980301e-02 2.27735817e-01 -1.73744842e-01 8.44202101e-01
-2.14190990e-01 -2.84300715e-01 1.05649507e+00 -5.59157968e-01
9.70229983e-01 -6.15250647e-01 8.32067072e-01 -1.84610233e-01
-7.75928319e-01 -5.09058177e-01 -1.53760016e-01 -4.26550470e-02
-9.35823739e-01 -4.89220768e-01 -6.31320477e-01 4.00911212e-01
1.17448561e-01 5.75913310e-01 -1.36425346e-01 5.55534959e-01
8.18567216e-01 -5.96317500e-02 1.49393976e-01 -2.07357839e-01
3.32083195e-01 1.12862992e+00 5.62815905e-01 -4.28341210e-01
7.62924030e-02 3.79469454e-01 -2.76189804e-01 -1.14028156e+00
-8.01383078e-01 -5.71568191e-01 -3.15796316e-01 -7.04461336e-01
8.64598036e-01 -1.03859663e+00 -6.72582805e-01 3.18909317e-01
-1.20803344e+00 8.29369873e-02 -4.89744961e-01 5.27449548e-01
-4.86131847e-01 4.93618518e-01 -5.19594967e-01 -1.31456757e+00
-7.91365743e-01 -6.55533195e-01 5.51557720e-01 1.49544895e-01
-9.58597302e-01 -1.06064308e+00 2.48086497e-01 9.99052227e-01
4.59387124e-01 3.79890263e-01 9.53178823e-01 -1.00684023e+00
4.68143821e-01 -5.14770076e-02 -2.55961157e-02 2.94457376e-01
7.25624636e-02 -1.13139562e-01 -8.06984544e-01 -1.62329331e-01
2.04328448e-01 -8.93175304e-01 5.14425635e-01 -4.85108234e-02
3.06611508e-02 -7.62999833e-01 4.12769839e-02 -6.08551316e-02
8.21763337e-01 -2.42071971e-01 8.45107555e-01 3.47957075e-01
9.17416364e-02 9.21944201e-01 5.80091476e-01 9.99513209e-01
7.34621763e-01 7.50337660e-01 4.62440491e-01 2.17936710e-02
-4.96419579e-01 -3.60399812e-01 6.69313908e-01 1.11709559e+00
-1.59762660e-03 -1.71138212e-01 -1.11822212e+00 7.87854135e-01
-1.88462687e+00 -1.47494316e+00 -8.22881639e-01 2.28653288e+00
8.60450685e-01 3.77624691e-01 4.51503307e-01 4.57894713e-01
4.41862464e-01 3.92314523e-01 1.18660375e-01 -8.91028523e-01
-4.29305077e-01 -4.44320858e-01 -1.24752000e-01 6.07331276e-01
-5.60242832e-01 6.06250584e-01 6.67370462e+00 7.06484377e-01
-8.90007138e-01 5.23022294e-01 8.64285886e-01 3.89844808e-03
-4.87840086e-01 3.41724977e-03 -7.41964042e-01 1.14554621e-01
9.83930886e-01 -6.61973000e-01 8.69184956e-02 4.89409447e-01
7.22873271e-01 -1.88384935e-01 -9.49315429e-01 8.47192585e-01
2.90895730e-01 -1.07238841e+00 -6.10788763e-01 -8.84681046e-02
3.27759266e-01 4.51724678e-02 1.79080561e-01 4.11356300e-01
1.75113708e-01 -1.05511272e+00 9.69618022e-01 6.06164277e-01
4.88125861e-01 -5.25040150e-01 5.83585143e-01 6.62490249e-01
-6.49428189e-01 -6.78147003e-02 -4.86611761e-02 -6.02742076e-01
2.63900936e-01 8.31469238e-01 -8.63066494e-01 2.18287826e-01
4.62381393e-01 9.02433336e-01 -3.62201869e-01 8.90485942e-01
-1.48234472e-01 6.97288394e-01 -9.91691947e-02 -2.16088265e-01
-7.80642629e-02 3.77501875e-01 1.06004930e+00 1.36627579e+00
2.35623136e-01 -4.13481385e-01 5.94433062e-02 6.62257314e-01
-1.08860455e-01 7.51138687e-01 -4.77147877e-01 -3.06730419e-01
2.58567303e-01 1.07856071e+00 -3.06924134e-01 -2.05923110e-01
-3.51586267e-02 6.82754874e-01 2.87670374e-01 -1.90625846e-01
-5.24631977e-01 5.63918762e-02 5.04364669e-01 4.74076033e-01
-6.61663771e-01 5.04534394e-02 -6.10215843e-01 -9.48694527e-01
-2.35885471e-01 -9.65037942e-01 4.50769365e-01 -7.48144090e-01
-1.16246343e+00 3.49598557e-01 -2.96141863e-01 -9.07891154e-01
-3.45841110e-01 3.29861566e-02 -9.05024290e-01 8.33045542e-01
-8.72259498e-01 -1.24974358e+00 -2.13411450e-02 3.79148781e-01
6.37929678e-01 -1.10225029e-01 1.07742906e+00 5.16090453e-01
-2.31461257e-01 6.44684792e-01 -1.58578411e-01 7.75921568e-02
1.16292727e+00 -8.69239390e-01 -1.77535281e-01 5.79110682e-01
-1.89978629e-01 3.61826122e-01 1.16291320e+00 -7.70660758e-01
-8.89920175e-01 -3.13478559e-01 2.00945139e+00 -2.42342368e-01
7.63810158e-01 -4.20341313e-01 -5.46073794e-01 1.87209129e-01
5.52493274e-01 -5.49470723e-01 9.77457762e-01 8.55302691e-01
-3.33954096e-01 1.67859778e-01 -1.32811296e+00 5.43267190e-01
1.21033216e+00 -6.06360853e-01 -6.94453835e-01 6.33501589e-01
6.92445099e-01 -8.09797794e-02 -9.52099741e-01 6.06606528e-02
6.47774935e-01 -1.30099034e+00 9.09996927e-01 -6.62088692e-01
7.81740487e-01 4.48492587e-01 -4.15237993e-01 -1.12103868e+00
-2.60795236e-01 -5.00632584e-01 5.14172554e-01 1.94082904e+00
4.62762117e-01 -5.59880972e-01 6.56716943e-01 6.91849768e-01
-3.69556636e-01 -4.71608043e-01 -1.45325887e+00 -3.72242153e-01
4.48349655e-01 -4.99904841e-01 1.02211677e-01 1.27247870e+00
8.31855357e-01 9.23530936e-01 -6.86281741e-01 -4.46130127e-01
5.09807527e-01 -4.50302929e-01 7.53417313e-01 -1.36892700e+00
-4.91565391e-02 -6.55410886e-01 -4.51755792e-01 -2.90292084e-01
4.48103137e-02 -1.11288655e+00 -3.31242472e-01 -1.93198621e+00
5.92453778e-01 -3.45290631e-01 -8.20899196e-03 4.51995045e-01
2.44466692e-01 5.69686368e-02 3.82829607e-01 3.84155601e-01
-7.90213823e-01 5.80558836e-01 1.03982890e+00 -2.41443753e-01
2.44441368e-02 8.26207846e-02 -1.18949115e+00 6.74283206e-01
7.65885293e-01 -4.68202293e-01 -2.79657364e-01 1.76869750e-01
7.72661030e-01 5.39057970e-01 3.72620076e-01 -1.10075271e+00
2.74820507e-01 -3.44007909e-02 -1.70663279e-02 -6.46286607e-01
7.07971990e-01 -3.47246677e-01 2.49510527e-01 7.27643788e-01
-5.91499984e-01 -5.30904494e-02 -1.25947461e-01 1.52239250e-02
2.24001147e-02 -3.33668858e-01 6.98755980e-01 1.75677255e-01
-1.05774507e-01 -4.32966679e-01 -3.62118900e-01 4.37258005e-01
5.60766280e-01 -1.06709287e-01 -7.68501639e-01 -9.91294742e-01
-1.02441442e+00 1.15266375e-01 4.54395741e-01 6.10696673e-01
7.12017357e-01 -1.51335967e+00 -1.42390907e+00 -1.54523358e-01
1.85007900e-01 -1.07736576e+00 2.99934387e-01 1.25599134e+00
-6.84140250e-02 9.26054269e-02 -2.49206141e-01 -1.54250860e-01
-1.77047026e+00 1.36778817e-01 2.09496915e-01 -2.21590221e-01
-6.74862385e-01 7.67325580e-01 -4.68322515e-01 -3.01771879e-01
4.21527892e-01 3.35467070e-01 -5.66379488e-01 2.69280404e-01
7.75151491e-01 9.32784438e-01 -1.42598674e-01 -1.33727360e+00
-7.11508095e-01 -3.74181825e-03 8.58785361e-02 -4.61934477e-01
1.22012210e+00 -4.08446044e-01 -4.48833436e-01 8.84701073e-01
1.25269318e+00 6.30817652e-01 -2.14703098e-01 2.47573420e-01
-5.39564043e-02 -5.21523297e-01 1.42267525e-01 -1.01676321e+00
-4.58667219e-01 8.86217892e-01 8.26010168e-01 3.83890569e-01
5.79550207e-01 9.04819146e-02 6.83215797e-01 -6.08171634e-02
4.20016237e-02 -1.25716150e+00 2.14590564e-01 5.86322606e-01
1.32131422e+00 -1.13972390e+00 -5.16674742e-02 -1.98388442e-01
-7.43505061e-01 1.01942456e+00 3.58092368e-01 4.19807106e-01
5.86495817e-01 1.39223009e-01 1.01271279e-01 -1.65427685e-01
-1.02977526e+00 1.73307404e-01 -1.07387891e-02 7.20095873e-01
1.13461113e+00 3.76156360e-01 -1.20195496e+00 7.74325728e-01
-5.43648601e-01 -1.04475692e-02 5.51293254e-01 2.46864483e-01
-8.02255392e-01 -1.09050751e+00 -2.91325957e-01 2.41425037e-01
-6.95556343e-01 -1.51428401e-01 -7.12952256e-01 4.01364923e-01
3.54230404e-01 1.55894423e+00 -2.50271350e-01 -5.23212671e-01
1.67501062e-01 2.74818242e-01 -2.50409842e-01 -2.69205689e-01
-9.20218527e-01 1.36731667e-02 1.24767911e+00 -1.09356724e-01
-4.81064230e-01 -1.01643050e+00 -1.27386272e+00 -8.58241558e-01
4.81869123e-04 2.95686424e-01 7.33238578e-01 8.24160457e-01
1.85783669e-01 3.61654103e-01 3.78142029e-01 -1.30980358e-01
-5.67589283e-01 -9.59834874e-01 -4.85920072e-01 3.01708668e-01
2.28107080e-01 -2.19893634e-01 -5.34174323e-01 -5.02160132e-01] | [9.010466575622559, 10.712047576904297] |
be2cfd1f-3c6e-439c-9a7b-5345f1738be2 | improving-candidate-retrieval-with-entity-1 | 2202.13404 | null | https://arxiv.org/abs/2202.13404v3 | https://arxiv.org/pdf/2202.13404v3.pdf | Improving Candidate Retrieval with Entity Profile Generation for Wikidata Entity Linking | Entity linking (EL) is the task of linking entity mentions in a document to referent entities in a knowledge base (KB). Many previous studies focus on Wikipedia-derived KBs. There is little work on EL over Wikidata, even though it is the most extensive crowdsourced KB. The scale of Wikidata can open up many new real-world applications, but its massive number of entities also makes EL challenging. To effectively narrow down the search space, we propose a novel candidate retrieval paradigm based on entity profiling. Wikidata entities and their textual fields are first indexed into a text search engine (e.g., Elasticsearch). During inference, given a mention and its context, we use a sequence-to-sequence (seq2seq) model to generate the profile of the target entity, which consists of its title and description. We use the profile to query the indexed search engine to retrieve candidate entities. Our approach complements the traditional approach of using a Wikipedia anchor-text dictionary, enabling us to further design a highly effective hybrid method for candidate retrieval. Combined with a simple cross-attention reranker, our complete EL framework achieves state-of-the-art results on three Wikidata-based datasets and strong performance on TACKBP-2010. | ['ChengXiang Zhai', 'Heng Ji', 'Tuan Manh Lai'] | 2022-02-27 | null | https://aclanthology.org/2022.findings-acl.292 | https://aclanthology.org/2022.findings-acl.292.pdf | findings-acl-2022-5 | ['pgtask'] | ['natural-language-processing'] | [-4.06601727e-01 1.09464526e-01 -6.52277589e-01 4.68301401e-02
-1.24783409e+00 -7.87027657e-01 5.32450378e-01 5.69287419e-01
-9.95615244e-01 1.12307537e+00 5.66800654e-01 1.51643381e-01
-7.92397335e-02 -9.64648485e-01 -8.07862699e-01 -2.44833827e-01
1.82716995e-01 1.06323552e+00 7.92878568e-01 -4.30601269e-01
1.56988516e-01 -4.25727665e-02 -1.35145128e+00 3.15126091e-01
1.24553561e+00 7.66329706e-01 3.04149956e-01 3.17680180e-01
-6.64129019e-01 8.29778612e-01 -6.87473178e-01 -9.66941476e-01
-1.79964006e-01 1.27086133e-01 -1.10231364e+00 -7.87563026e-01
5.78528345e-01 2.02018134e-02 -4.29675788e-01 9.93348300e-01
9.66632307e-01 1.19345061e-01 5.01718342e-01 -9.70795631e-01
-1.00832796e+00 1.05248511e+00 -3.28591079e-01 1.64979130e-01
6.36798978e-01 -2.66004235e-01 1.62009811e+00 -1.21778560e+00
1.12402070e+00 1.02835929e+00 7.41284370e-01 3.75889897e-01
-9.05251741e-01 -3.52305591e-01 -3.29846255e-02 4.42094773e-01
-1.68202317e+00 -3.31033677e-01 2.24008337e-01 -3.62548620e-01
1.29941809e+00 2.98304766e-01 5.25222600e-01 7.13177264e-01
-4.23058182e-01 1.05287528e+00 4.90699559e-01 -5.08037090e-01
6.64027929e-02 1.10291205e-01 2.39494234e-01 7.38490701e-01
4.22210991e-01 -5.44997513e-01 -6.28037930e-01 -6.39658213e-01
4.32615168e-02 -4.79504138e-01 -2.83075064e-01 -2.60240495e-01
-1.28684664e+00 5.97835064e-01 3.17279488e-01 6.92484900e-02
-4.24928755e-01 1.46260664e-01 6.51973844e-01 -1.58228070e-01
4.49658126e-01 8.14693809e-01 -7.30963707e-01 1.37332201e-01
-8.99260283e-01 6.11059248e-01 1.13071311e+00 1.09798813e+00
9.00043607e-01 -7.22844660e-01 -7.67372668e-01 1.14861035e+00
4.57448423e-01 6.63754344e-01 5.88483572e-01 -5.05975485e-01
8.64586413e-01 8.09572518e-01 3.79964650e-01 -1.10465443e+00
-2.63329268e-01 -1.90252244e-01 -1.50068954e-01 -7.05832839e-01
2.55523890e-01 -2.29308158e-01 -6.16365314e-01 1.40305090e+00
6.22295558e-01 2.40479857e-01 2.51768380e-01 9.13593709e-01
1.43589914e+00 7.05883086e-01 3.79170358e-01 2.03957725e-02
1.61542320e+00 -1.10813463e+00 -6.13127351e-01 -2.99788892e-01
9.41966534e-01 -5.80992162e-01 8.65665674e-01 -1.95254654e-01
-8.27156842e-01 1.86253358e-02 -6.22159779e-01 -4.96633083e-01
-6.69747829e-01 3.44682068e-01 3.26910168e-01 2.45628640e-01
-6.57075047e-01 3.12842458e-01 -2.53168225e-01 -2.87829280e-01
7.62737542e-02 5.51475361e-02 -4.71474707e-01 -7.74839893e-02
-2.00959492e+00 1.06890106e+00 9.11919177e-01 2.15318445e-02
-4.02559161e-01 -8.67150843e-01 -8.98683727e-01 6.55790642e-02
9.63202357e-01 -8.40067565e-01 1.09269702e+00 -2.65762657e-01
-9.81045127e-01 8.61056626e-01 -2.15387747e-01 -3.92646462e-01
1.09350257e-01 -3.80891949e-01 -6.81973219e-01 8.29655081e-02
6.13092124e-01 4.96723592e-01 2.93023288e-01 -9.97363329e-01
-1.05154753e+00 -6.51285723e-02 1.10769071e-01 3.01545113e-01
-4.67049956e-01 2.49436766e-01 -1.32679129e+00 -6.21221185e-01
-3.60133201e-01 -9.85303521e-01 -2.91819014e-02 -4.68352169e-01
-8.29225957e-01 -6.60760283e-01 1.15064368e-01 -8.20753932e-01
1.91162872e+00 -1.64707482e+00 1.28525510e-01 1.44187316e-01
3.28182280e-01 5.49106658e-01 -3.56313959e-02 7.75838673e-01
3.92625809e-01 2.41409287e-01 5.57192303e-02 2.03261040e-02
2.79484957e-01 -1.89163145e-02 -4.17889237e-01 7.02260574e-03
4.16257093e-03 1.34805667e+00 -1.47441900e+00 -8.96465242e-01
-6.38165414e-01 2.37271845e-01 -3.97131413e-01 -5.79710677e-02
-4.95595247e-01 -2.68727928e-01 -7.49537647e-01 7.01923192e-01
1.89389154e-01 -2.97449261e-01 3.75185370e-01 -4.13855076e-01
-2.68435832e-02 6.55456662e-01 -1.19091916e+00 1.37675464e+00
-3.06469887e-01 3.97325605e-01 -4.19558197e-01 -4.42444861e-01
6.07048690e-01 2.96428770e-01 1.69001237e-01 -6.43256843e-01
-4.47791159e-01 6.68982744e-01 -5.35677254e-01 -7.94057131e-01
9.94759202e-01 3.01805675e-01 -5.22154570e-01 1.70331612e-01
1.02901720e-01 4.15712446e-01 7.11841106e-01 5.70212007e-01
1.24668837e+00 6.62531555e-02 3.97522539e-01 -1.13944262e-01
5.87183297e-01 5.33464313e-01 6.67853653e-01 7.62425303e-01
3.80619794e-01 2.89173622e-04 3.28635931e-01 -3.00692260e-01
-8.95988822e-01 -7.32647419e-01 8.93286765e-02 1.35707140e+00
2.35909507e-01 -7.71593690e-01 -4.83875275e-01 -1.10389483e+00
6.17916584e-01 3.98508519e-01 -5.87259591e-01 6.20615319e-04
-6.85303390e-01 -6.50124609e-01 1.09339809e+00 3.26911926e-01
3.01516473e-01 -1.12412620e+00 -5.33136502e-02 5.10966122e-01
-6.78538740e-01 -1.06902719e+00 -8.79914403e-01 -8.97997096e-02
-2.27152437e-01 -9.85269427e-01 -8.08294535e-01 -9.05300140e-01
4.18563515e-01 -7.71919712e-02 1.57129598e+00 -6.23004418e-03
-8.13357532e-02 2.69731700e-01 -5.23033559e-01 -2.45453805e-01
-1.72525361e-01 6.17259800e-01 1.18296258e-01 -3.49775255e-01
6.90005362e-01 1.07082427e-01 -5.78024447e-01 1.80204377e-01
-7.34968007e-01 -2.69726455e-01 6.00615799e-01 8.35449159e-01
6.46717489e-01 -2.08524287e-01 9.13179457e-01 -1.19157708e+00
8.65937769e-01 -8.37702096e-01 -5.72515905e-01 7.69096375e-01
-7.74125338e-01 2.84929156e-01 4.52738494e-01 -4.78869468e-01
-1.06984460e+00 -3.91817763e-02 -1.79613270e-02 -3.58081958e-03
4.22233790e-01 1.37787604e+00 -3.63950729e-02 2.08990470e-01
7.68454492e-01 2.23987848e-01 -7.08465457e-01 -6.51193738e-01
6.32641673e-01 7.01838553e-01 6.52663827e-01 -7.99937129e-01
7.03583181e-01 -6.26638308e-02 -4.82593358e-01 -3.90182197e-01
-1.26629138e+00 -9.19962227e-01 -4.25892323e-01 4.97020893e-02
6.76779151e-01 -1.24734902e+00 -7.37098932e-01 6.34979606e-02
-1.14640403e+00 -4.17330414e-02 -1.32558987e-01 2.98597306e-01
1.23939224e-01 3.78247857e-01 -6.38590574e-01 -4.15515810e-01
-5.38013577e-01 -7.37313032e-01 1.15829062e+00 1.98480934e-01
-1.10798568e-01 -9.01593804e-01 6.26420856e-01 4.94551688e-01
1.59044877e-01 -2.90332556e-01 8.64708006e-01 -1.11596608e+00
-6.34701431e-01 -5.02931714e-01 -2.94672459e-01 -1.95304781e-01
-1.80906296e-01 -9.88595337e-02 -5.81396341e-01 -5.71099706e-02
-1.10139322e+00 -4.21728402e-01 1.03215110e+00 -2.77160645e-01
7.18138933e-01 -5.97470880e-01 -7.74979413e-01 1.69781357e-01
1.38180232e+00 -2.12882861e-01 5.00251830e-01 6.82811379e-01
8.28536808e-01 5.40604532e-01 9.00053859e-01 3.12717348e-01
1.00479829e+00 1.13214326e+00 -7.26994425e-02 1.59797639e-01
-4.94268164e-02 -6.18192434e-01 2.40552098e-01 9.70769405e-01
8.76628458e-02 -6.08044386e-01 -1.21483219e+00 9.62671816e-01
-1.95122945e+00 -1.17630625e+00 -4.58011627e-02 2.07382703e+00
1.55454147e+00 -8.18998590e-02 2.41397563e-02 -4.61425900e-01
8.87125313e-01 -1.35102704e-01 -3.66896778e-01 3.39433342e-01
-2.38248035e-01 -9.99727845e-02 7.91079223e-01 4.98960197e-01
-1.13257813e+00 1.32457507e+00 5.48488140e+00 1.35094476e+00
-8.36445034e-01 4.35834080e-02 5.89083247e-02 -1.93068366e-02
-3.98030221e-01 3.44738476e-02 -1.72744405e+00 8.08742762e-01
7.28602350e-01 -7.25964308e-01 1.20460153e-01 8.15809965e-01
-3.16203654e-01 -1.09398074e-01 -7.66725004e-01 5.61815560e-01
1.85100779e-01 -1.70477688e+00 1.90975457e-01 -1.68294638e-01
6.84347034e-01 3.40775579e-01 -3.95658165e-01 8.12325656e-01
5.18634379e-01 -6.60454512e-01 7.67026484e-01 6.06372297e-01
8.33464324e-01 -5.83021641e-01 8.56528401e-01 3.57465267e-01
-1.36478066e+00 5.97657822e-02 -3.70231241e-01 5.70006847e-01
3.05398703e-01 8.67526114e-01 -1.10739982e+00 6.52820110e-01
6.86667204e-01 3.72190326e-01 -4.90470856e-01 1.30843246e+00
-4.47874367e-01 5.16209304e-01 -3.44066501e-01 -5.41279852e-01
1.98671177e-01 3.39935243e-01 6.99632287e-01 1.68291879e+00
1.55162945e-01 -7.31971413e-02 3.24392706e-01 4.14087296e-01
-8.44308972e-01 6.79087698e-01 -3.46787810e-01 -4.81767431e-02
1.15839410e+00 1.39576054e+00 -3.86025220e-01 -6.72133327e-01
-4.04318392e-01 9.54768002e-01 9.51483190e-01 1.58146873e-01
-7.29936898e-01 -8.37994218e-01 3.51641238e-01 -9.67100337e-02
6.10376000e-01 1.79386690e-01 5.92368841e-01 -1.21638501e+00
3.63203913e-01 -7.41037965e-01 7.21891940e-01 -6.61364496e-01
-1.46547294e+00 4.66498643e-01 7.92883113e-02 -9.26747859e-01
-1.49206147e-01 -6.25598282e-02 -1.09427370e-01 8.44636559e-01
-1.80740142e+00 -1.09265804e+00 -1.14941485e-02 1.94269910e-01
9.56080854e-02 -7.61229321e-02 7.11178482e-01 9.07597780e-01
-6.00722492e-01 7.47869849e-01 3.84163022e-01 6.46977663e-01
1.18460512e+00 -1.35368848e+00 5.61152160e-01 5.95915020e-01
2.87498891e-01 1.00284660e+00 4.47025210e-01 -1.28398573e+00
-1.36259460e+00 -1.27154160e+00 1.52675235e+00 -8.86161804e-01
8.89743745e-01 -1.96414515e-01 -1.01644778e+00 5.54264486e-01
-3.42917182e-02 1.83134973e-01 8.19637418e-01 3.03179085e-01
-5.51493466e-01 6.82128370e-02 -7.85504937e-01 6.52137995e-01
9.34018791e-01 -6.30837917e-01 -8.48270059e-01 2.66288429e-01
9.56077397e-01 -6.49345517e-01 -1.04912496e+00 2.80591518e-01
6.10665321e-01 -2.95323040e-02 1.20978558e+00 -7.84859538e-01
2.79878557e-01 -7.24672198e-01 9.50079635e-02 -1.35257971e+00
-2.64370590e-01 -3.78194362e-01 -6.59016550e-01 1.68116248e+00
1.11887050e+00 -4.85406220e-01 6.76867127e-01 6.09569132e-01
-1.61098000e-02 -8.76407266e-01 -8.96710396e-01 -8.06080997e-01
-1.73443779e-01 3.93576100e-02 6.68413699e-01 1.04726326e+00
4.11479652e-01 7.06048787e-01 -2.96457022e-01 3.69500458e-01
3.44230503e-01 1.75297186e-01 5.62102854e-01 -1.40454543e+00
-1.58452764e-01 -1.45385534e-01 -7.96948150e-02 -1.09806931e+00
3.51837307e-01 -1.24352694e+00 3.24583858e-01 -1.72354257e+00
4.20776635e-01 -9.36194599e-01 -1.40779436e-01 7.23677516e-01
-8.35243583e-01 2.73027718e-01 -1.58644125e-01 4.76678461e-01
-1.34670997e+00 4.80907828e-01 7.36762404e-01 -1.63217068e-01
-1.29411176e-01 -4.57955718e-01 -6.77781463e-01 3.98169160e-01
2.53061980e-01 -8.35842907e-01 -1.29396930e-01 -6.58805251e-01
1.12246597e+00 4.20227349e-02 4.31638323e-02 -4.52178180e-01
8.81340921e-01 -5.14339320e-02 -5.33759780e-02 -5.00210345e-01
-1.21132610e-02 -4.81230408e-01 8.11773241e-02 3.08533013e-02
-6.02395892e-01 1.01570571e-02 3.65915410e-02 8.18974197e-01
-3.26819837e-01 -6.59888864e-01 1.21926256e-01 -1.74139991e-01
-1.06489849e+00 4.12902772e-01 -2.61288345e-01 7.57513285e-01
6.90859258e-01 3.32410574e-01 -7.67035484e-01 6.25672266e-02
-4.45521802e-01 7.00022161e-01 4.25578743e-01 3.90063882e-01
2.02659547e-01 -1.56584096e+00 -8.24277997e-01 -5.95815897e-01
6.80392087e-01 4.62129377e-02 4.84189056e-02 6.87601686e-01
-3.33167881e-01 6.10258579e-01 3.10060471e-01 -2.78253276e-02
-1.14121187e+00 4.84051585e-01 -6.88439002e-03 -7.09719300e-01
-4.20441359e-01 9.71119761e-01 -3.11083138e-01 -6.17977262e-01
1.90974221e-01 1.40309289e-01 -5.99210799e-01 5.90985060e-01
7.21411288e-01 3.93348634e-01 3.14342380e-01 -5.84730446e-01
-5.18795013e-01 2.87417620e-01 -2.33549461e-01 -2.78906804e-02
1.20487535e+00 -1.33788228e-01 -4.48263586e-01 1.50741413e-01
9.19142485e-01 7.39543259e-01 -4.13421959e-01 -7.07842946e-01
7.42640018e-01 -2.22561330e-01 -5.44552840e-02 -9.28114772e-01
-7.80836582e-01 1.54776528e-01 -1.45845190e-01 9.04180706e-02
5.47936857e-01 2.95633405e-01 1.05688119e+00 9.29876268e-01
4.23448920e-01 -1.35628676e+00 -2.43160367e-01 8.49704027e-01
5.39364398e-01 -1.15440869e+00 2.44697817e-02 -4.51349109e-01
-6.82917118e-01 6.87052071e-01 6.70581937e-01 3.24084550e-01
2.64272392e-01 3.24858986e-02 -1.33510083e-01 -4.35699731e-01
-8.77979696e-01 -5.86600482e-01 7.74265647e-01 2.95430452e-01
4.53762203e-01 -2.05453768e-01 -4.41872090e-01 8.15204322e-01
1.32963777e-01 2.00721268e-02 1.92841277e-01 7.68830061e-01
-6.64372563e-01 -1.14740849e+00 -1.95674673e-02 6.12900734e-01
-7.57888496e-01 -7.22326219e-01 -2.92712837e-01 5.29369950e-01
1.01322189e-01 5.09644389e-01 -2.60540515e-01 -2.06736058e-01
4.74481165e-01 1.74921542e-01 1.26210996e-03 -8.92318308e-01
-6.94252491e-01 -4.00758862e-01 7.64655650e-01 -2.08231717e-01
-3.16599429e-01 -4.80834842e-01 -1.20981014e+00 -1.91244557e-01
-7.76730657e-01 6.67037904e-01 3.51686895e-01 8.41419399e-01
6.24731719e-01 1.12272426e-01 1.48186550e-01 -2.40962207e-01
-2.14193285e-01 -7.50973165e-01 -2.03906760e-01 2.66864896e-01
7.42076710e-02 -7.82036126e-01 1.29988343e-01 -9.83371213e-02] | [9.481307029724121, 8.80034351348877] |
11ad6358-dc14-48b8-b034-92a1e0dc8a66 | machine-unlearning-learning-polluting-and | 2111.14609 | null | https://arxiv.org/abs/2111.14609v2 | https://arxiv.org/pdf/2111.14609v2.pdf | An Investigation on Learning, Polluting, and Unlearning the Spam Emails for Lifelong Learning | Machine unlearning for security is studied in this context. Several spam email detection methods exist, each of which employs a different algorithm to detect undesired spam emails. But these models are vulnerable to attacks. Many attackers exploit the model by polluting the data, which are trained to the model in various ways. So to act deftly in such situations model needs to readily unlearn the polluted data without the need for retraining. Retraining is impractical in most cases as there is already a massive amount of data trained to the model in the past, which needs to be trained again just for removing a small amount of polluted data, which is often significantly less than 1%. This problem can be solved by developing unlearning frameworks for all spam detection models. In this research, unlearning module is integrated into spam detection models that are based on Naive Bayes, Decision trees, and Random Forests algorithms. To assess the benefits of unlearning over retraining, three spam detection models are polluted and exploited by taking attackers' positions and proving models' vulnerability. Reduction in accuracy and true positive rates are shown in each case showing the effect of pollution on models. Then unlearning modules are integrated into the models, and polluted data is unlearned; on testing the models after unlearning, restoration of performance is seen. Also, unlearning and retraining times are compared with different pollution data sizes on all models. On analyzing the findings, it can be concluded that unlearning is considerably superior to retraining. Results show that unlearning is fast, easy to implement, easy to use, and effective. | ['Ripon Patgiri', 'Nithish Bhupathi', 'Kyathi Puppaala', 'Nishchal Parne'] | 2021-11-26 | null | null | null | null | ['spam-detection'] | ['natural-language-processing'] | [ 1.12772211e-01 6.66772574e-03 9.80232731e-02 -2.99828518e-02
-2.99622342e-02 -6.64330304e-01 6.97453737e-01 -1.00244798e-01
-6.42391145e-01 8.53956282e-01 -1.46890491e-01 -9.36513007e-01
-1.04218505e-01 -9.70593691e-01 -4.63047534e-01 -5.75281560e-01
2.94460386e-01 4.51957345e-01 6.35762155e-01 -2.84849197e-01
5.02155840e-01 6.62323117e-01 -1.60612857e+00 4.55729961e-01
8.96275938e-01 2.05640286e-01 -2.35888995e-02 8.10844958e-01
-4.93618935e-01 5.86876333e-01 -1.06418085e+00 -5.22681177e-01
3.15774947e-01 -1.51131421e-01 -8.50962639e-01 -1.24234967e-01
3.76770854e-01 -5.04413962e-01 -1.73562497e-01 1.04514766e+00
3.66700172e-01 9.37435124e-03 3.84942889e-01 -1.20285487e+00
-3.66633266e-01 4.84891951e-01 -3.18712771e-01 3.04869860e-01
4.23272282e-01 3.00068557e-01 4.74812627e-01 -4.51893687e-01
3.55405331e-01 1.48405838e+00 6.39694393e-01 7.60072172e-01
-1.24066532e+00 -9.42814171e-01 1.32060498e-01 6.76753670e-02
-5.05054832e-01 1.56923123e-02 4.16750848e-01 -1.87396735e-01
7.77261674e-01 6.45025134e-01 5.37878156e-01 1.38695407e+00
2.53942102e-01 3.84246856e-01 1.53992355e+00 -6.26810789e-01
1.94728687e-01 8.15915465e-01 7.78211057e-01 3.85773391e-01
8.17647815e-01 5.25705636e-01 -3.33439440e-01 -5.60284317e-01
1.67919293e-01 3.48350376e-01 5.39185256e-02 2.44304705e-02
-3.25216562e-01 8.94542038e-01 4.77908522e-01 4.46827203e-01
-1.49481043e-01 -3.82149220e-01 3.76942903e-01 5.96811712e-01
3.98905307e-01 5.96995115e-01 -6.05814874e-01 -4.84998040e-02
-6.79581046e-01 7.01496303e-02 1.06275773e+00 4.42141473e-01
8.04109514e-01 1.44679427e-01 6.95554838e-02 5.31241059e-01
2.84073204e-01 6.08751893e-01 7.79549539e-01 -2.14161366e-01
1.51794836e-01 1.05813551e+00 1.30471691e-01 -8.37385058e-01
-2.16560662e-01 -5.06921828e-01 -5.08605123e-01 5.55563271e-01
7.01955855e-01 -2.45237023e-01 -1.16490269e+00 1.28878748e+00
2.58886486e-01 3.63010764e-01 -1.41858727e-01 5.87635577e-01
3.99932504e-01 4.31172818e-01 4.77451563e-01 -2.49391764e-01
1.10227871e+00 -7.39039660e-01 -6.94990277e-01 -2.99690515e-01
5.67965925e-01 -7.42491722e-01 1.33204293e+00 6.99221611e-01
-3.70210975e-01 -4.23948199e-01 -9.26074862e-01 6.08349800e-01
-9.44324672e-01 -6.64933920e-01 5.47206581e-01 1.43287635e+00
-7.01944053e-01 9.06659126e-01 -4.30445790e-01 -4.01600868e-01
3.57447177e-01 4.67352897e-01 -1.00090794e-01 3.62729095e-02
-1.50771737e+00 1.30924249e+00 2.41224363e-01 -7.74976388e-02
-8.34568679e-01 -3.85550708e-01 -1.41226098e-01 3.90312672e-02
1.17708907e-01 -3.61878246e-01 1.02131891e+00 -1.07246590e+00
-1.10617936e+00 4.09832895e-01 6.81301132e-02 -5.53430974e-01
8.72050762e-01 -5.19147158e-01 -6.76333427e-01 -1.67291135e-01
-1.32231638e-02 -2.28053764e-01 1.21069908e+00 -1.51318634e+00
-6.02121949e-01 -5.50686002e-01 -1.28797041e-02 -3.05265099e-01
-7.47486591e-01 2.41699666e-01 1.92538589e-01 -4.36873913e-01
-4.49810252e-02 -7.19313741e-01 -2.65974104e-01 -6.12054467e-01
-1.27844945e-01 1.02917947e-01 1.43070889e+00 -6.81597829e-01
1.58907855e+00 -1.73006415e+00 -6.80884361e-01 5.90349615e-01
1.38960481e-02 1.10569382e+00 -1.14323035e-01 4.18616593e-01
-4.47732471e-02 4.23485160e-01 -1.19980320e-01 3.28483760e-01
-1.80872008e-01 2.91919172e-01 -6.33983195e-01 1.48493111e-01
-9.46979672e-02 5.75182438e-01 -7.65936792e-01 -2.48639688e-01
3.32128912e-01 2.55457669e-01 -3.27243537e-01 2.15753645e-01
-1.24763828e-02 5.15743233e-02 -3.84727836e-01 4.59204644e-01
9.48812902e-01 1.41244993e-01 2.60931611e-01 3.41138422e-01
1.60768002e-01 2.21976653e-01 -1.05837953e+00 4.24282074e-01
-5.21797180e-01 3.91363621e-01 -1.46824066e-02 -8.58368218e-01
9.07329679e-01 1.86192378e-01 -1.74341217e-01 -3.73385280e-01
2.68727303e-01 2.73618519e-01 3.66929211e-02 -7.46686637e-01
1.66638151e-01 -2.40749478e-01 2.25160941e-01 7.67753720e-01
-1.91592529e-01 2.42470711e-01 2.35456321e-02 4.10183430e-01
1.37753034e+00 -1.53092638e-01 1.95402190e-01 3.78130411e-04
7.99497843e-01 4.09833714e-02 2.34874472e-01 1.25393653e+00
-3.74342173e-01 -2.27412194e-01 3.68983626e-01 -4.56943750e-01
-6.15319133e-01 -1.22445691e+00 -2.26570159e-01 1.45132208e+00
6.15809262e-02 -1.51420355e-01 -9.80550349e-01 -1.56870735e+00
3.80963326e-01 9.45089281e-01 -5.28845966e-01 -2.29833886e-01
-5.55524528e-01 -1.07636464e+00 6.79874003e-01 -1.43681502e-03
6.25742853e-01 -1.32263803e+00 -4.24979329e-01 2.44038701e-01
2.72351205e-01 -4.33953166e-01 2.66222715e-01 3.33225965e-01
-1.22840619e+00 -1.36605692e+00 3.33897211e-02 -5.28317571e-01
8.40948522e-01 5.32637358e-01 8.45218956e-01 7.26874828e-01
-3.86474043e-01 1.23183146e-01 -5.01373768e-01 -7.56765544e-01
-9.45715070e-01 6.07134961e-02 2.09247455e-01 -2.74173409e-01
9.05032456e-01 -5.81343472e-01 -9.55491811e-02 4.14763570e-01
-1.03422570e+00 -7.03484952e-01 8.57326388e-01 9.56231296e-01
-3.67129117e-01 4.17758286e-01 1.10135448e+00 -1.57499623e+00
9.89338756e-01 -3.80857229e-01 -4.28934246e-01 2.94706285e-01
-1.00942850e+00 1.53407240e-02 7.93616891e-01 -7.54243731e-01
-1.26828349e+00 -1.51788324e-01 -5.65621331e-02 1.44861728e-01
-4.21937704e-01 -2.71366276e-02 -2.72839844e-01 -1.71410039e-01
1.18020058e+00 -5.57410438e-03 3.03971440e-01 -6.48046196e-01
2.35455543e-01 1.23213565e+00 -8.19926932e-02 -2.04169527e-01
1.20704412e+00 2.27653429e-01 -4.59860831e-01 -7.70807385e-01
-8.44801009e-01 -3.68280560e-01 -4.76861119e-01 -4.93895829e-01
2.63463885e-01 -2.44053286e-02 -5.38918495e-01 5.29016674e-01
-8.22888255e-01 7.84243047e-02 2.77784746e-02 2.81499028e-01
2.56840765e-01 6.38908923e-01 -5.00432611e-01 -1.12700570e+00
-5.23137093e-01 -8.39751840e-01 1.18173450e-01 2.71636367e-01
-3.91052336e-01 -9.77547169e-01 -6.55501783e-02 6.29303992e-01
6.78923786e-01 -2.47266427e-01 1.14764380e+00 -1.34667635e+00
-2.24803597e-01 -7.92919636e-01 -1.41549900e-01 7.21750557e-01
2.51888186e-01 -1.31250754e-01 -1.12228608e+00 -5.29589117e-01
4.22891855e-01 -5.16713709e-02 8.11512947e-01 -2.51864672e-01
8.38071048e-01 -6.22228682e-01 -5.50863624e-01 -1.34464756e-01
1.22227526e+00 2.67029226e-01 5.91610193e-01 8.17442894e-01
1.91272080e-01 7.28710175e-01 4.87623274e-01 -7.07535595e-02
-5.11514783e-01 3.64606455e-02 6.23487234e-01 1.28232792e-01
9.06881019e-02 -2.47733369e-01 6.40420735e-01 5.01796424e-01
3.00805360e-01 1.18899204e-01 -7.46982872e-01 8.30191448e-02
-1.61256075e+00 -1.13504744e+00 -4.72585171e-01 2.22918248e+00
7.99099803e-01 6.55190229e-01 1.34828836e-01 4.18971092e-01
9.08110380e-01 -1.54856607e-01 -4.56677884e-01 -7.06939936e-01
2.30484799e-01 4.78821546e-01 4.20939922e-01 8.57576489e-01
-1.11267889e+00 1.04055560e+00 6.65353012e+00 7.73416638e-01
-9.87616539e-01 6.43302947e-02 2.51912981e-01 1.23881949e-02
-1.30476773e-01 3.83088976e-01 -8.87162447e-01 6.68741465e-01
1.07937694e+00 -4.62684780e-02 4.74175394e-01 1.08257151e+00
2.23631412e-01 -3.79623145e-01 -5.42662740e-01 2.76848525e-01
-9.49851424e-02 -9.05830503e-01 4.15533662e-01 1.09848445e-02
3.74643117e-01 -3.69296938e-01 -1.47481367e-01 7.33339190e-01
5.56097746e-01 -8.79419267e-01 -4.24444070e-03 4.94321167e-01
-2.00795960e-02 -7.65112400e-01 1.10442114e+00 8.23661089e-01
-2.14482635e-01 -5.32951593e-01 -4.49828744e-01 -3.45823258e-01
-2.80037582e-01 5.59103608e-01 -1.12691259e+00 3.47606719e-01
8.38862062e-01 -5.60519472e-03 -1.13783693e+00 8.35483372e-01
-4.93861973e-01 1.10344577e+00 -1.32440358e-01 -4.43086475e-01
9.86217633e-02 -2.13276237e-01 6.12923384e-01 1.25733817e+00
-9.48712900e-02 -3.15151513e-01 -1.32774532e-01 5.77291369e-01
2.97630489e-01 4.96896319e-02 -7.88191438e-01 3.27881455e-01
5.57255685e-01 1.24419332e+00 -5.40397108e-01 -2.68948793e-01
-2.55636573e-01 6.95365608e-01 1.32042035e-01 3.00479978e-01
-6.69666827e-01 -5.74039757e-01 3.12277883e-01 5.13732791e-01
-2.65216351e-01 2.66036391e-01 -5.10673165e-01 -7.08385766e-01
-2.21647337e-01 -1.21713960e+00 4.24020439e-01 -4.04324263e-01
-1.69814503e+00 5.68824828e-01 -2.36084357e-01 -9.40380275e-01
-7.97528848e-02 -7.87560225e-01 -7.06601977e-01 1.01029372e+00
-1.11134863e+00 -8.76673639e-01 -1.71822593e-01 3.87991101e-01
3.11688215e-01 -6.30110860e-01 8.70946586e-01 4.00101058e-02
-6.24986231e-01 5.28444588e-01 5.98543435e-02 1.07264102e-01
8.26305866e-01 -1.14841259e+00 1.62062660e-01 1.02910554e+00
-2.28795893e-02 1.09641159e+00 7.14489102e-01 -1.08705199e+00
-8.33630621e-01 -1.08561707e+00 9.15931344e-01 -8.15215707e-01
7.30452716e-01 -3.99456918e-01 -1.33225691e+00 4.65139896e-01
4.56818677e-02 -5.01856863e-01 7.81943738e-01 2.01374799e-01
-4.62283611e-01 -5.04430197e-02 -1.51119304e+00 6.64938629e-01
5.04785001e-01 -3.16950619e-01 -9.21426058e-01 1.46740571e-01
3.83234382e-01 2.67636448e-01 -2.63284028e-01 1.94855511e-01
5.52602351e-01 -1.24914777e+00 8.08870435e-01 -1.17987573e+00
-2.19362393e-01 -1.16953991e-01 2.75078237e-01 -1.00745881e+00
-2.24773034e-01 -5.23274601e-01 -1.45422190e-01 1.24913001e+00
5.87920785e-01 -1.12028003e+00 1.05234492e+00 6.05758190e-01
4.55444157e-01 -4.96909827e-01 -7.46799409e-01 -8.28377247e-01
3.01235557e-01 -4.19840008e-01 5.26096463e-01 8.85456920e-01
5.34095168e-02 5.83047748e-01 -3.21354389e-01 1.78585976e-01
6.58796370e-01 -2.09787816e-01 1.04030287e+00 -1.48702943e+00
-4.64949161e-02 -2.09801331e-01 -1.00647464e-01 -5.18235087e-01
1.89370111e-01 -7.94029772e-01 -3.71387005e-01 -1.28133798e+00
1.61268294e-01 -3.08706433e-01 -1.76321954e-01 5.71955681e-01
-6.59550905e-01 -8.62902105e-02 1.86113015e-01 3.03725630e-01
-9.55759659e-02 -1.18465900e-01 1.05822527e+00 -1.00879438e-01
-3.94954950e-01 5.72962523e-01 -7.58882284e-01 1.00334108e+00
1.14502490e+00 -7.60466993e-01 -2.88320065e-01 1.70330539e-01
-1.60872132e-01 -4.72305954e-01 4.72036183e-01 -8.98870766e-01
4.25994769e-02 -5.65276220e-02 6.18013918e-01 -3.59516233e-01
-1.26189023e-01 -1.12008834e+00 -2.20830411e-01 8.98408294e-01
-1.61077544e-01 -7.11129084e-02 1.63030729e-01 7.52430260e-01
2.71852016e-01 -8.36123765e-01 1.02500618e+00 -2.07441404e-01
-3.98534715e-01 -2.65002519e-01 -8.70226800e-01 -1.62109837e-01
1.05381346e+00 -4.18041974e-01 -6.10790431e-01 -3.66769403e-01
-9.57871318e-01 -9.85620357e-03 1.61983654e-01 5.55629432e-01
4.25358295e-01 -7.72191942e-01 -3.43254924e-01 6.26954436e-01
-3.82665277e-01 -6.98899627e-01 8.20232034e-02 4.06504899e-01
-5.02494693e-01 1.66092798e-01 -1.94894910e-01 -1.95729628e-01
-1.57746911e+00 8.30697477e-01 2.35795140e-01 -5.63262045e-01
-3.23680252e-01 3.76308352e-01 -3.94716769e-01 -8.23902667e-01
4.94386822e-01 3.11291665e-01 -3.67866248e-01 -1.50457341e-02
5.04150867e-01 7.69533455e-01 7.63948336e-02 -1.68620154e-01
-3.83907676e-01 -1.83155909e-02 -5.80591381e-01 9.10699666e-02
8.71525168e-01 2.40550786e-01 -4.32069719e-01 6.11007959e-02
7.58013546e-01 3.55325699e-01 -5.98152578e-01 -8.11625570e-02
4.67406571e-01 -6.58094883e-01 -3.53059292e-01 -1.31009972e+00
-5.48417509e-01 8.14731240e-01 7.27129519e-01 6.76236153e-01
9.04571235e-01 -5.48439264e-01 5.24046063e-01 5.49120009e-01
2.48384029e-01 -1.23583913e+00 2.23003685e-01 5.41953444e-01
5.89384794e-01 -1.15315533e+00 1.43283293e-01 -4.06106591e-01
-4.53898281e-01 1.09573734e+00 9.67456698e-01 -1.30215034e-01
9.30059016e-01 2.60990620e-01 3.91339570e-01 7.29234293e-02
-5.50832510e-01 -4.69686016e-02 -2.10046142e-01 7.66964138e-01
8.11245590e-02 -2.87285428e-02 -5.24034381e-01 6.35092676e-01
-2.49053985e-01 -3.68764520e-01 5.05465090e-01 1.05780530e+00
-1.00778484e+00 -1.50898492e+00 -8.78767610e-01 8.79107058e-01
-4.60897624e-01 6.44874945e-02 -9.05051053e-01 9.72823143e-01
1.83405697e-01 1.40886283e+00 -4.69909847e-01 -7.18383908e-01
5.26960671e-01 4.74302053e-01 -6.01406321e-02 -6.85682237e-01
-1.26134980e+00 -2.65739381e-01 1.01735473e-01 -1.03805237e-01
-1.82077941e-02 -4.30173635e-01 -9.15477574e-01 -5.70015788e-01
-9.19742465e-01 6.03389621e-01 3.63699377e-01 7.65947104e-01
7.91603625e-02 3.83123666e-01 7.85891652e-01 -3.25430304e-01
-1.21045554e+00 -1.31605947e+00 -4.84005660e-01 5.64256608e-01
4.29604128e-02 -5.35817087e-01 -1.06421590e+00 -3.67735595e-01] | [7.731541633605957, 9.923995018005371] |
82c36afb-af81-45fa-bb49-de2e826c825c | zero-shot-aspect-based-scientific-document | null | null | https://openreview.net/forum?id=iEqUccu6yII | https://openreview.net/pdf?id=iEqUccu6yII | Zero-Shot Aspect-Based Scientific Document Summarization using Self-Supervised Pre-training | We study the zero-shot setting for the aspect-based scientific document summarization task. Summarizing scientific documents with respect to an aspect can remarkably improve document assistance systems and readers experience. However, existing large-scale datasets contain a limited variety of aspects, causing summarization models to over-fit to a small set of aspects. We establish baseline results in zero-shot performance (over unseen aspects and the presence of domain shift), paraphrasing, leave-one-out, and limited supervised samples experimental setups. We propose a self-supervised pre-training approach to enhance the zero-shot performance. Experimental results on the FacetSum and PubMed aspect-based datasets show promising performance when the model is pre-trained using unlabeled in-domain data. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['scientific-article-summarization'] | ['natural-language-processing'] | [ 3.30571294e-01 3.92703980e-01 -7.41393387e-01 -3.19998354e-01
-1.42563808e+00 -5.21625698e-01 7.05950558e-01 6.46303177e-01
-3.59181404e-01 9.38890159e-01 9.34659064e-01 -5.05655818e-02
-6.99127242e-02 -4.50155675e-01 -7.20773757e-01 -4.16725844e-01
3.50678444e-01 7.93334842e-01 3.46259147e-01 -1.43469915e-01
5.75028062e-01 -4.44240086e-02 -1.27861261e+00 5.89600921e-01
1.28020036e+00 1.16871446e-01 -6.23500310e-02 9.18552637e-01
-5.19077063e-01 4.99023944e-01 -1.15001035e+00 -4.49391991e-01
-1.75414070e-01 -6.95446372e-01 -8.97319853e-01 1.37994945e-01
5.93338192e-01 -6.22420795e-02 -9.36003327e-02 9.24950838e-01
9.64901626e-01 2.60996073e-01 9.00875926e-01 -6.88657880e-01
-7.64023960e-01 5.86182356e-01 -7.00899482e-01 3.48903418e-01
5.32615662e-01 3.08171492e-02 1.06408930e+00 -6.21327162e-01
1.34672284e+00 1.13638878e+00 5.48556268e-01 6.90210581e-01
-1.26251185e+00 -2.21044064e-01 3.11448164e-02 1.73643366e-01
-8.63283038e-01 -6.12798989e-01 4.38135117e-01 -1.10056520e-01
1.24206507e+00 3.75929594e-01 5.57937920e-01 1.33019412e+00
3.35841119e-01 8.88193429e-01 6.07795000e-01 -2.22198144e-01
5.84272087e-01 1.14344053e-01 9.42715049e-01 3.02786678e-01
7.82668173e-01 -7.72682965e-01 -6.78580463e-01 -6.72618270e-01
1.74023416e-02 -7.19972923e-02 -3.65925491e-01 -1.38923272e-01
-1.18885565e+00 6.63922966e-01 3.38935223e-03 2.14761764e-01
-2.26202711e-01 -4.69317883e-01 8.04372787e-01 2.57902622e-01
8.30437481e-01 1.09312558e+00 -4.28436190e-01 -1.46895081e-01
-1.45654809e+00 4.24641192e-01 1.06116891e+00 1.37674820e+00
4.34673101e-01 -4.19839531e-01 -1.12672269e+00 1.09649026e+00
-6.27523184e-01 2.71340907e-01 8.25934649e-01 -9.02656376e-01
4.21253115e-01 5.57124496e-01 3.86722125e-02 -3.38452816e-01
-4.01201725e-01 -5.84122717e-01 -8.75740945e-01 -5.78147650e-01
-8.33589211e-03 -1.13169149e-01 -1.25549817e+00 1.23916364e+00
3.24652553e-01 1.12429112e-01 4.30846483e-01 5.60261726e-01
1.67951906e+00 6.60818577e-01 8.17383826e-02 -5.31501949e-01
1.70142984e+00 -1.21455896e+00 -1.20130026e+00 -8.45566094e-02
7.57508039e-01 -6.32704318e-01 1.17441392e+00 2.19662264e-01
-1.17977989e+00 -1.93398058e-01 -1.00791764e+00 -6.22353733e-01
-4.50400174e-01 5.32812998e-02 4.23490644e-01 2.34945267e-01
-8.92899573e-01 7.16174543e-01 -5.86638391e-01 -7.51096666e-01
7.37208009e-01 -9.83291045e-02 -3.68921101e-01 -3.90267074e-01
-8.02676260e-01 6.13730907e-01 3.95716757e-01 -8.97083044e-01
-6.16689920e-01 -1.43337262e+00 -6.77670181e-01 4.03189272e-01
6.15093648e-01 -1.27099347e+00 1.19520688e+00 -4.79109213e-02
-9.81436074e-01 9.46021557e-01 -6.87497199e-01 -3.98770362e-01
2.95368791e-01 -1.94652945e-01 -1.68452233e-01 4.86756653e-01
5.80517769e-01 4.42987323e-01 5.09349823e-01 -1.01858020e+00
-2.66764253e-01 -5.16489863e-01 -2.04541370e-01 2.88114816e-01
-3.15271199e-01 -6.76991493e-02 -7.05508471e-01 -5.92459977e-01
-2.37859845e-01 -4.94790852e-01 -3.54401916e-01 -2.82515854e-01
-8.23156059e-01 -3.39474291e-01 5.39558113e-01 -5.83912730e-01
1.27628589e+00 -1.68468761e+00 6.43088073e-02 -5.61138630e-01
3.57855409e-01 3.64985913e-01 -3.43133956e-01 5.04056573e-01
-1.50054507e-02 3.05187434e-01 -4.84898299e-01 -4.88021493e-01
-5.02947927e-01 1.19132340e-01 -1.81446850e-01 7.89225772e-02
1.38211146e-01 1.04707313e+00 -1.26329923e+00 -8.29966962e-01
-4.66638088e-01 -8.12767148e-02 -5.77559650e-01 1.35668099e-01
-3.86866033e-01 1.14525557e-01 -6.05858564e-01 7.69124806e-01
6.67411804e-01 -4.65267032e-01 -2.14513704e-01 -6.47578314e-02
1.80793151e-01 2.74814248e-01 -5.49700856e-01 2.15588903e+00
-4.85218205e-02 4.39695299e-01 -3.88698131e-01 -8.73304904e-01
5.50450087e-01 3.93706650e-01 4.55225855e-01 -2.86110520e-01
-1.40008762e-01 6.53859451e-02 -1.44257039e-01 -7.01302588e-01
9.67169046e-01 -1.45142734e-01 -1.33798406e-01 5.07852077e-01
5.79177141e-01 -2.06640735e-01 7.14601994e-01 9.32520807e-01
1.34388459e+00 -2.73677766e-01 9.27110076e-01 -3.86876464e-01
1.46287307e-01 5.57372630e-01 5.44538796e-01 1.19764805e+00
-7.70157799e-02 1.07901657e+00 9.80887473e-01 5.38068935e-02
-1.08500457e+00 -6.54217899e-01 -2.90244281e-01 1.05795074e+00
-5.50774708e-02 -7.95671225e-01 -7.27570593e-01 -8.85003805e-01
-6.41517416e-02 1.29939282e+00 -7.38379896e-01 -4.56092149e-01
8.78967792e-02 -8.37559879e-01 5.69366634e-01 3.76419604e-01
6.30784929e-02 -6.69021487e-01 -2.91587442e-01 1.44293964e-01
-2.09289446e-01 -1.03581786e+00 -6.92932010e-01 1.30972311e-01
-9.32830632e-01 -1.08070576e+00 -1.27050710e+00 -5.39906383e-01
6.70477211e-01 7.11342454e-01 1.22724152e+00 -3.50567669e-01
-3.13079149e-01 3.89424324e-01 -3.42368275e-01 -8.09368730e-01
-2.83716857e-01 3.74491543e-01 -2.87443727e-01 -6.94278002e-01
4.73054767e-01 -3.79395902e-01 -6.23127520e-01 -3.03046823e-01
-9.33606446e-01 3.69750820e-02 7.05120385e-01 1.08095562e+00
6.94425285e-01 -6.37789726e-01 7.79965162e-01 -1.59032822e+00
1.04673803e+00 -6.79066658e-01 3.18971604e-01 6.90867722e-01
-6.66604459e-01 2.15195417e-01 5.12647033e-01 -5.00190556e-01
-1.24145615e+00 -6.33588135e-01 -4.37047100e-03 -3.29399914e-01
-4.39217128e-02 6.46933496e-01 -1.17047012e-01 5.13150871e-01
1.02945948e+00 4.51820284e-01 -4.08761054e-02 -6.12384379e-01
6.04597569e-01 7.90897131e-01 6.25936449e-01 -4.12129790e-01
1.93658918e-01 4.83851701e-01 -1.35119185e-01 -1.08998859e+00
-1.36940181e+00 -1.12592804e+00 -4.55646068e-01 2.92012155e-01
6.44511640e-01 -1.03192568e+00 -2.84141880e-02 -8.75603259e-02
-1.37618113e+00 3.50741297e-01 -7.85160542e-01 1.30832195e-01
-3.93264920e-01 6.76346302e-01 -4.60250914e-01 -3.70086312e-01
-1.09501982e+00 -9.39427733e-01 1.49692917e+00 4.58732009e-01
-5.90103269e-01 -6.75126851e-01 4.17153895e-01 4.13775504e-01
3.55218276e-02 2.24075958e-01 1.17618060e+00 -1.41323078e+00
-5.44777699e-03 -3.88894230e-01 -7.70123154e-02 -1.52555227e-01
1.53676882e-01 -2.73679197e-01 -9.03062284e-01 -6.33226708e-02
-1.06506944e-01 -4.37787086e-01 1.42999506e+00 5.17285049e-01
1.08895087e+00 -4.64026213e-01 -5.75150251e-01 4.71177429e-01
9.90032911e-01 -1.36076272e-01 5.05324244e-01 1.54884711e-01
6.21179104e-01 5.23651242e-01 6.11453533e-01 5.15182316e-01
1.64501369e-01 1.80962339e-01 -5.02161741e-01 -5.50951026e-02
-4.33337569e-01 -2.38953486e-01 -1.43760607e-01 8.54793608e-01
1.42605320e-01 -5.26198030e-01 -5.72866380e-01 7.87727475e-01
-1.85551548e+00 -1.07539642e+00 -8.64424556e-02 2.11198378e+00
1.29923439e+00 -6.61048517e-02 1.77597538e-01 -3.54722559e-01
4.87174898e-01 3.37140769e-01 -8.20712388e-01 -4.43113148e-01
-3.51432890e-01 1.68967620e-01 3.45320672e-01 2.11822033e-01
-7.49737382e-01 1.02115524e+00 6.65425730e+00 1.19211113e+00
-6.76552355e-01 3.71960849e-01 6.67907119e-01 -4.90290165e-01
-6.39952481e-01 -3.97209935e-02 -9.34515893e-01 3.35521340e-01
9.49712753e-01 -9.71210122e-01 -2.03185380e-01 9.39037561e-01
1.99738339e-01 -2.11660609e-01 -1.14477980e+00 6.75408959e-01
5.47050595e-01 -1.66706395e+00 6.16388083e-01 -1.36625022e-01
1.02962673e+00 4.54355590e-02 -4.09972787e-01 5.66646695e-01
2.17756391e-01 -6.45187140e-01 2.69159842e-02 6.22908950e-01
9.07998502e-01 -4.98770982e-01 7.64530241e-01 5.36092877e-01
-2.95741946e-01 5.52062035e-01 -8.55625272e-01 5.12843132e-01
4.23976146e-02 9.19025540e-01 -8.94084871e-01 8.79653931e-01
4.47579831e-01 9.10345376e-01 -7.10674524e-01 1.20761943e+00
-6.35160506e-02 6.30895078e-01 -5.07278554e-03 -2.45536849e-01
3.54817166e-04 -1.35941476e-01 1.11942971e+00 1.42170513e+00
1.77346170e-01 3.28990251e-01 -8.87045264e-02 6.57652617e-01
-4.30431813e-01 3.06385726e-01 -6.50983214e-01 -2.65317708e-01
1.44961029e-01 1.28002763e+00 -6.92804635e-01 -1.05980146e+00
-3.78276169e-01 1.01171732e+00 3.61598313e-01 4.34236944e-01
-2.41488814e-01 -6.97481573e-01 4.03937936e-01 3.44153121e-02
2.85187870e-01 3.22835147e-01 -7.14242816e-01 -1.62091863e+00
-1.64734796e-01 -8.06494594e-01 7.19414175e-01 -6.62668526e-01
-1.16864145e+00 3.32022369e-01 2.53733397e-01 -1.15067160e+00
-2.20203653e-01 -5.67762069e-02 -1.06351936e+00 5.49870253e-01
-1.41196680e+00 -8.51161540e-01 -1.21957906e-01 -3.42863612e-02
1.08649349e+00 -2.37448603e-01 8.35052311e-01 -3.29339921e-01
-6.78839922e-01 7.33983099e-01 6.16894722e-01 -4.62695420e-01
1.30344820e+00 -1.37254250e+00 3.88349652e-01 7.40273774e-01
-2.18607664e-01 1.02637780e+00 1.07836175e+00 -1.04870522e+00
-1.32533920e+00 -1.17098594e+00 1.15577245e+00 -5.69566429e-01
4.19789195e-01 -1.07641421e-01 -1.35884094e+00 3.87604654e-01
5.69949031e-01 -3.26791793e-01 1.19364786e+00 5.99899709e-01
-3.42660695e-01 3.51659536e-01 -1.20400393e+00 6.98834002e-01
1.12040210e+00 -2.54807293e-01 -1.28077364e+00 6.92657769e-01
1.01306069e+00 -3.16787601e-01 -8.75060856e-01 2.14178935e-01
3.83861333e-01 -4.65459794e-01 8.97283852e-01 -1.28479338e+00
1.02732575e+00 3.24879512e-02 3.62618238e-01 -1.58268857e+00
-3.63413393e-01 -6.39640749e-01 -3.75231743e-01 1.43503773e+00
4.19818342e-01 -1.44951493e-01 7.05480576e-01 6.11881196e-01
-3.80880177e-01 -7.53728926e-01 -7.29875863e-01 -6.74868584e-01
2.27526903e-01 1.87731415e-01 3.93334061e-01 8.12632382e-01
6.18295252e-01 1.07297289e+00 -1.96279034e-01 -4.93996352e-01
6.34409726e-01 3.25365663e-01 7.87850201e-01 -1.10683000e+00
-3.25532049e-01 -4.11651373e-01 -1.60335258e-01 -8.49821627e-01
1.61185861e-01 -1.01044297e+00 1.82331670e-02 -2.15459204e+00
1.19532502e+00 5.02745271e-01 -7.69623742e-02 3.89853448e-01
-1.11842525e+00 -2.69127786e-01 -2.42872313e-01 4.28111225e-01
-1.24164414e+00 8.78535092e-01 1.34315574e+00 -5.07119596e-01
-2.67557114e-01 -1.31609598e-02 -1.28117335e+00 3.23221594e-01
4.95687693e-01 -5.57873487e-01 -5.10632753e-01 -9.18500423e-02
-3.14298898e-01 5.09815402e-02 -4.66937959e-01 -7.31739163e-01
5.34778714e-01 -4.86849807e-02 3.50836754e-01 -1.01668775e+00
1.78556629e-02 9.37720314e-02 -4.01463121e-01 2.86232203e-01
-8.99477303e-01 -4.67830151e-01 3.31396878e-01 8.88807595e-01
4.07737307e-02 -5.06564379e-01 5.88323832e-01 -3.56616676e-01
-3.84122789e-01 1.21514887e-01 -3.48895341e-01 7.93142915e-01
6.94824815e-01 -4.56165709e-02 -8.27790022e-01 -4.57807571e-01
-3.43047440e-01 5.49744368e-01 4.94439214e-01 2.66280055e-01
3.96349847e-01 -8.96261036e-01 -8.25511277e-01 -2.95432150e-01
7.34331071e-01 1.33878529e-01 6.20683134e-01 6.87588751e-01
-1.19997382e-01 6.18039787e-01 -8.85292441e-02 -4.23988253e-01
-1.40818572e+00 8.10892105e-01 -4.49803084e-01 -6.50439918e-01
-7.17127442e-01 7.02017426e-01 2.99038172e-01 -4.69856769e-01
5.24031967e-02 -2.54281074e-01 -3.74031723e-01 3.11745942e-01
9.01108861e-01 5.58353543e-01 2.72355616e-01 -8.30496252e-02
-9.36824754e-02 2.26954252e-01 -7.83514082e-01 -7.35597089e-02
1.34907711e+00 9.78918746e-02 -6.11449825e-03 5.44973075e-01
1.33499861e+00 2.88819964e-03 -5.08880794e-01 -4.33717787e-01
2.76745334e-02 -1.22355871e-01 1.61287129e-01 -8.32893074e-01
-3.25223655e-01 6.56481981e-01 -1.22505732e-01 -2.48034090e-01
7.12012708e-01 2.71027446e-01 7.79991329e-01 7.04756796e-01
-1.34950161e-01 -1.16067481e+00 -1.55406632e-02 4.49066371e-01
8.46399307e-01 -1.37970245e+00 6.57883286e-01 -4.27172661e-01
-9.95474279e-01 7.95038402e-01 3.99029940e-01 1.70505792e-01
2.86359996e-01 -3.09956297e-02 -2.39808813e-01 -3.78623843e-01
-1.13401031e+00 -2.46649563e-01 5.72314382e-01 6.62965775e-01
5.27380943e-01 -9.84790698e-02 -9.30286944e-01 1.06112814e+00
-1.82388410e-01 4.88795936e-02 8.35280180e-01 1.02449250e+00
-6.30097866e-01 -7.01464891e-01 -1.43378094e-01 1.19016516e+00
-6.58058286e-01 -3.08319777e-01 -7.73763001e-01 4.02190387e-01
-5.11683524e-01 7.85729051e-01 -9.47321281e-02 1.34052649e-01
5.68913460e-01 2.18431324e-01 2.37178087e-01 -1.09104502e+00
-6.47958457e-01 3.22932750e-01 3.04497302e-01 -3.44694793e-01
-1.12042002e-01 -6.64213121e-01 -1.01337397e+00 -1.13319039e-01
-6.42291233e-02 5.17339110e-01 2.24857360e-01 8.07341814e-01
1.02430630e+00 7.97468126e-01 1.53890803e-01 -5.15216172e-01
-6.36932790e-01 -1.35414481e+00 -5.63993573e-01 5.46211541e-01
2.50305712e-01 -3.27551097e-01 -1.88137978e-01 1.03979990e-01] | [12.39635944366455, 9.467986106872559] |
980e71fd-64cd-498a-819c-e1d8b700e60a | semantic-segmentation-assisted-instance | 2208.04766 | null | https://arxiv.org/abs/2208.04766v1 | https://arxiv.org/pdf/2208.04766v1.pdf | Semantic Segmentation-Assisted Instance Feature Fusion for Multi-Level 3D Part Instance Segmentation | Recognizing 3D part instances from a 3D point cloud is crucial for 3D structure and scene understanding. Several learning-based approaches use semantic segmentation and instance center prediction as training tasks and fail to further exploit the inherent relationship between shape semantics and part instances. In this paper, we present a new method for 3D part instance segmentation. Our method exploits semantic segmentation to fuse nonlocal instance features, such as center prediction, and further enhances the fusion scheme in a multi- and cross-level way. We also propose a semantic region center prediction task to train and leverage the prediction results to improve the clustering of instance points. Our method outperforms existing methods with a large-margin improvement in the PartNet benchmark. We also demonstrate that our feature fusion scheme can be applied to other existing methods to improve their performance in indoor scene instance segmentation tasks. | ['Yang Liu', 'Xin Tong', 'ChunYu Sun'] | 2022-08-09 | null | null | null | null | ['3d-instance-segmentation-1', '3d-part-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.33135968e-01 1.96280852e-01 -3.08177799e-01 -6.15459800e-01
-7.86264300e-01 -6.43771172e-01 4.65528846e-01 3.03333431e-01
1.45402461e-01 6.50989562e-02 2.04708111e-02 3.88591439e-02
-9.58350599e-02 -8.11611712e-01 -9.17552173e-01 -2.27654696e-01
8.53669569e-02 7.51040637e-01 8.85206580e-01 -1.31123990e-01
3.27148020e-01 8.74496639e-01 -1.72186661e+00 3.69757146e-01
8.72876942e-01 1.20387113e+00 2.59777069e-01 2.81657219e-01
-4.99936342e-01 3.47064942e-01 -3.50298494e-01 5.64878865e-04
7.04524577e-01 -6.02533389e-03 -1.00446463e+00 4.98631775e-01
6.14691019e-01 8.89615715e-03 8.78808424e-02 8.16130579e-01
3.20074230e-01 2.02766150e-01 5.61091840e-01 -1.19872832e+00
5.23144193e-02 3.16829860e-01 -6.21682584e-01 -3.45291734e-01
2.75065303e-01 -1.03382796e-01 9.36609685e-01 -7.26689160e-01
5.71654260e-01 1.23771930e+00 9.02881980e-01 3.21787596e-01
-9.95834172e-01 -5.38008332e-01 5.09944558e-01 1.08018003e-01
-1.30810440e+00 -2.24444240e-01 1.18307698e+00 -3.23719710e-01
1.15134645e+00 9.92986783e-02 8.46777380e-01 4.14990038e-01
-3.72636825e-01 1.30165887e+00 1.13365078e+00 -2.86761522e-01
2.34921426e-01 -7.51783401e-02 2.50175178e-01 8.93826723e-01
9.69188586e-02 -1.84752464e-01 -2.39878237e-01 -3.79808694e-02
8.85316372e-01 3.53066444e-01 7.30761215e-02 -8.31925452e-01
-1.12325549e+00 6.15579009e-01 8.75155449e-01 2.14382246e-01
-3.94314051e-01 2.92677969e-01 1.12718515e-01 -7.12350905e-02
7.53921866e-01 4.14672852e-01 -1.01515281e+00 2.72892546e-02
-1.16664886e+00 2.44907498e-01 8.21725428e-01 1.22625828e+00
1.21324599e+00 -4.75712806e-01 -1.57839417e-01 9.03046012e-01
4.05838370e-01 4.22862589e-01 3.38747390e-02 -1.23117256e+00
2.81982213e-01 1.28926706e+00 -2.19016820e-01 -4.93323237e-01
-5.59402168e-01 -2.88314879e-01 -1.55581892e-01 8.45931098e-02
2.87617087e-01 4.04337108e-01 -1.46217334e+00 1.17980647e+00
8.27238739e-01 5.83138525e-01 -2.82384902e-01 8.16337109e-01
8.73468339e-01 1.62699893e-01 4.60735373e-02 5.22461116e-01
1.16784072e+00 -1.09435356e+00 1.12457145e-02 -2.36961946e-01
7.68454850e-01 -7.53606856e-01 8.51122856e-01 7.16010407e-02
-9.25196111e-01 -6.74398720e-01 -9.24896061e-01 -2.77777821e-01
-4.05133039e-01 1.12868719e-01 7.53746867e-01 5.72619736e-01
-7.48641431e-01 6.14274442e-01 -1.15532160e+00 -4.60243911e-01
1.01229560e+00 6.19188726e-01 -2.70696551e-01 -2.62733012e-01
-3.90954912e-01 5.11533380e-01 6.20404124e-01 -4.78525251e-01
-4.70380515e-01 -1.04880393e+00 -9.97565746e-01 -8.92809406e-02
5.24573505e-01 -8.88237774e-01 1.15709925e+00 -6.09896719e-01
-1.34797728e+00 1.05570650e+00 -1.69853419e-01 -3.99571270e-01
2.74394006e-01 -2.97872424e-01 2.53292769e-01 3.03825617e-01
1.74776047e-01 1.08722329e+00 6.72097087e-01 -1.65582705e+00
-8.51279020e-01 -7.33891249e-01 2.87071347e-01 3.07811171e-01
2.75428414e-01 -6.43024266e-01 -9.11033332e-01 -4.33036387e-01
8.02163124e-01 -1.07244146e+00 -4.97529984e-01 -3.95370647e-02
-5.55910349e-01 -3.44403505e-01 1.08084118e+00 -2.01437458e-01
4.59379375e-01 -1.95017314e+00 -8.21347013e-02 4.92608279e-01
1.31287172e-01 1.03790238e-01 -5.35017997e-02 1.43614590e-01
2.21842140e-01 1.22502623e-02 -6.00310981e-01 -5.93657911e-01
2.13722676e-01 4.72639412e-01 6.11118637e-02 2.61532336e-01
2.38157615e-01 1.21438038e+00 -6.73311353e-01 -5.92531502e-01
6.51330650e-01 5.49218833e-01 -9.21133757e-01 8.02841387e-04
-5.88314295e-01 4.98614252e-01 -6.84983432e-01 9.21844482e-01
7.93630362e-01 -4.01023686e-01 -1.00766473e-01 -3.15059215e-01
1.32762283e-01 3.15266073e-01 -1.16440630e+00 2.35348845e+00
-4.62573200e-01 1.64053828e-01 -9.25302580e-02 -1.11946392e+00
8.34278882e-01 -1.04681766e-02 9.71101046e-01 -4.15129900e-01
-1.02643199e-01 2.13618174e-01 -3.67054045e-01 -1.75848588e-01
4.34572577e-01 -1.03047207e-01 -6.63828775e-02 2.20696315e-01
1.40832886e-01 -7.24060535e-01 -8.94434676e-02 1.16354316e-01
1.00307751e+00 6.72374189e-01 1.67778343e-01 -2.36507624e-01
4.98755723e-01 3.63993645e-01 4.83499378e-01 7.79253006e-01
-1.46617725e-01 8.03046584e-01 1.76039010e-01 -1.69476330e-01
-7.62415051e-01 -9.53518271e-01 -2.61851072e-01 8.23760629e-01
6.24357998e-01 -4.42137361e-01 -8.00002217e-01 -1.35070777e+00
4.47242886e-01 6.12924576e-01 -3.47331852e-01 1.74285041e-03
-5.62350094e-01 -3.29757452e-01 1.46545038e-01 9.25813913e-01
6.27192497e-01 -7.40022004e-01 -4.67067659e-01 1.11575112e-01
-6.20271340e-02 -1.54837704e+00 -2.16148466e-01 4.41537946e-01
-1.26999593e+00 -1.15642369e+00 -4.97738451e-01 -7.11043239e-01
7.89571106e-01 5.60520947e-01 1.25937402e+00 3.05377066e-01
-2.91236699e-01 9.66718793e-01 -5.55247366e-01 -5.93890190e-01
-1.65572628e-01 3.35354060e-01 -2.32425034e-01 -3.71630251e-01
4.55575019e-01 -6.52214408e-01 -7.34092295e-01 4.35176194e-01
-8.55796874e-01 1.23817533e-01 7.88687170e-01 4.40101117e-01
1.11496735e+00 -4.83096093e-02 2.33318448e-01 -1.09429801e+00
-2.77582496e-01 -2.49517202e-01 -5.26493847e-01 1.01217046e-01
-3.99187624e-01 2.28386283e-01 6.64381729e-03 2.13510182e-04
-9.65195596e-01 5.99031270e-01 -3.39188069e-01 -7.38031626e-01
-7.38588750e-01 6.72827810e-02 -6.15660906e-01 -3.54809821e-01
2.54031658e-01 3.79858091e-02 -1.98374882e-01 -8.27871859e-01
6.23296320e-01 2.43643209e-01 1.12700805e-01 -7.11601198e-01
9.90385711e-01 7.81055629e-01 2.51252025e-01 -6.76187754e-01
-1.07308185e+00 -1.12709558e+00 -1.29428303e+00 6.83384091e-02
9.58567679e-01 -1.04058075e+00 -5.06385863e-01 2.06413224e-01
-9.95440423e-01 -5.46468079e-01 -5.96468031e-01 3.77396673e-01
-8.52148473e-01 3.35702926e-01 -3.66803646e-01 -5.10109961e-01
-5.89306839e-03 -1.12545538e+00 1.90218830e+00 7.25738555e-02
2.23638341e-01 -9.59391952e-01 -1.52137220e-01 7.01108873e-01
-6.04032725e-02 2.35255584e-01 7.61798382e-01 -8.88311327e-01
-1.10958290e+00 -2.31882632e-01 -3.14742684e-01 1.98776290e-01
1.89101025e-01 -3.63788098e-01 -1.14306188e+00 8.40227306e-02
-1.38988271e-01 -6.73214272e-02 1.05764997e+00 5.17014921e-01
1.43321419e+00 3.01276557e-02 -7.20387995e-01 6.75755382e-01
1.41231227e+00 -1.63432330e-01 2.77714610e-01 9.57007408e-02
1.12731874e+00 6.32712662e-01 8.17521870e-01 2.94058979e-01
6.00508451e-01 7.59534121e-01 6.51628435e-01 -3.07080686e-01
-3.99804294e-01 -4.74038929e-01 -1.74163699e-01 4.79737133e-01
-1.36018544e-01 1.75521821e-01 -9.52325404e-01 4.97706205e-01
-1.89259577e+00 -8.00010383e-01 -2.12178111e-01 1.94832015e+00
4.35657114e-01 2.38665879e-01 1.92874327e-01 1.65223673e-01
4.93891656e-01 -3.71401384e-02 -5.79434514e-01 1.89386964e-01
1.91163033e-01 4.93385404e-01 7.78864861e-01 3.31264883e-01
-1.44409311e+00 1.34478056e+00 5.79831219e+00 8.10355306e-01
-6.51162386e-01 6.67039976e-02 3.83151591e-01 1.65162235e-01
-2.23843113e-01 1.15071222e-01 -8.63783181e-01 -4.03431877e-02
3.19629043e-01 5.21119595e-01 3.47778909e-02 1.19709229e+00
-1.57292902e-01 -9.97619182e-02 -1.29897964e+00 9.90824044e-01
2.98105441e-02 -1.29543591e+00 -6.23838082e-02 2.80993313e-01
9.14498150e-01 3.11273485e-01 -2.83931226e-01 2.97462374e-01
3.28515112e-01 -7.48308182e-01 6.04280710e-01 3.71307522e-01
1.05598204e-01 -7.63191104e-01 5.10850370e-01 4.47097152e-01
-1.58816504e+00 9.48256776e-02 -2.89256692e-01 1.96962684e-01
1.41330972e-01 7.34688938e-01 -1.06843424e+00 9.11536396e-01
6.96291387e-01 1.06489646e+00 -7.94689834e-01 1.32367575e+00
-1.90742493e-01 5.03504038e-01 -7.11889863e-01 2.69376934e-01
2.60645419e-01 -1.46606207e-01 5.97852647e-01 9.97316182e-01
1.39239311e-01 1.49645299e-01 6.98197067e-01 1.00694895e+00
-6.30783141e-02 -1.24464976e-02 -5.35140157e-01 2.28979051e-01
2.44796500e-01 1.31558168e+00 -1.18443501e+00 -4.85453665e-01
-4.99871433e-01 1.04937625e+00 1.15447387e-01 1.34031773e-01
-7.36271501e-01 -1.06189646e-01 9.29377794e-01 2.28647649e-01
7.82276690e-01 -4.72100437e-01 -5.49356818e-01 -1.11448801e+00
-3.90884578e-02 -3.32599819e-01 2.05109566e-01 -5.05388737e-01
-1.34017992e+00 1.02503940e-01 3.14875543e-01 -1.29884970e+00
2.62476429e-02 -6.77024603e-01 -3.20219010e-01 3.32917154e-01
-1.90828025e+00 -1.74840176e+00 -3.76022369e-01 7.17661619e-01
7.86656857e-01 3.54956895e-01 5.01306891e-01 1.32712692e-01
-1.38348415e-01 1.39686003e-01 -3.76793414e-01 1.60912469e-01
2.43162438e-01 -1.55293882e+00 4.14554566e-01 5.45653999e-01
5.79685628e-01 3.39565665e-01 2.26263687e-01 -7.11699724e-01
-1.19422591e+00 -1.36504734e+00 4.26660061e-01 -8.77888858e-01
1.08689576e-01 -4.78252470e-01 -6.88102663e-01 7.21145749e-01
-3.10279757e-01 3.03412884e-01 7.35339344e-01 2.49548152e-01
-4.30785060e-01 1.02277666e-01 -1.26083028e+00 2.43714839e-01
1.57947397e+00 -4.68122721e-01 -6.58863604e-01 3.98584664e-01
9.89542961e-01 -6.32605970e-01 -1.11601388e+00 8.89897645e-01
3.09791207e-01 -9.85488236e-01 1.36904216e+00 -3.57314706e-01
1.61812097e-01 -4.02740806e-01 -4.86117661e-01 -9.00706112e-01
-1.45362854e-01 -9.90708396e-02 -2.26897806e-01 1.13828516e+00
2.34339610e-01 -3.10331345e-01 1.24141717e+00 6.40300095e-01
-7.20082641e-01 -6.40184104e-01 -8.30568910e-01 -7.77367651e-01
-4.04712521e-02 -8.70173335e-01 9.65859830e-01 6.07681513e-01
-4.55965668e-01 6.14795573e-02 3.49229574e-01 3.92788589e-01
7.60344803e-01 7.18573153e-01 1.06686890e+00 -1.51331186e+00
-5.28264008e-02 -5.25334716e-01 -8.24190378e-01 -1.53171051e+00
3.34885776e-01 -1.32044590e+00 1.37820184e-01 -1.85611916e+00
1.80767558e-03 -8.80206704e-01 -4.08087373e-01 7.71233559e-01
-1.64736167e-01 5.08226573e-01 4.45739120e-01 3.20706844e-01
-9.49609518e-01 5.59908509e-01 1.28676236e+00 -2.08588526e-01
-4.96066034e-01 3.28380078e-01 -3.60631108e-01 8.76045167e-01
5.88422537e-01 -4.33971882e-01 -4.13276136e-01 -2.37330675e-01
-4.37953621e-01 -2.81212926e-01 5.96051216e-01 -1.41145492e+00
1.17220245e-02 1.11198276e-01 7.23894000e-01 -1.18858707e+00
6.71924949e-01 -1.20247650e+00 -2.23827019e-01 2.51494944e-01
3.11186463e-02 -4.96127337e-01 2.54400074e-01 7.24190593e-01
-3.68903019e-02 8.95790383e-02 5.89678526e-01 -4.75281388e-01
-9.39618349e-01 5.99243879e-01 2.97181159e-01 -6.35646731e-02
1.12059712e+00 -6.18917704e-01 1.22324280e-01 1.65046826e-01
-7.90551364e-01 3.14230531e-01 8.08558166e-01 5.15708566e-01
6.56790733e-01 -1.24816895e+00 -2.35218719e-01 2.83183873e-01
3.43021661e-01 5.07164121e-01 2.39452839e-01 1.02517164e+00
-4.12344486e-01 3.92464757e-01 9.81625319e-02 -1.26057374e+00
-1.14617133e+00 4.10910219e-01 3.81692559e-01 -1.50792658e-01
-8.56464982e-01 8.21191549e-01 3.75321597e-01 -9.21447635e-01
1.11107051e-01 -6.56923890e-01 2.82760374e-02 -2.41457626e-01
-2.26861715e-01 1.68020919e-01 2.74503559e-01 -7.52878070e-01
-5.30545950e-01 1.01089191e+00 1.54568598e-01 1.63397148e-01
1.44479334e+00 -8.87503996e-02 9.68681425e-02 3.56304079e-01
1.29451084e+00 2.89535383e-04 -1.45846593e+00 -3.12133491e-01
3.22467327e-01 -5.31038284e-01 2.60142945e-02 -8.26476693e-01
-1.23241985e+00 7.70619571e-01 5.83498955e-01 5.24295308e-02
9.86606300e-01 6.97752297e-01 9.90027964e-01 3.73438537e-01
5.92852533e-01 -1.05771708e+00 8.18800032e-02 3.86097282e-01
3.46590608e-01 -1.53854859e+00 1.04043521e-01 -8.97219121e-01
-4.47233975e-01 9.54621971e-01 6.14299774e-01 -2.89756924e-01
9.81224716e-01 1.01714231e-01 -9.02128592e-02 -5.50206900e-01
-2.12759972e-01 -8.97212625e-01 8.56563389e-01 7.48139501e-01
2.22025171e-01 9.67458040e-02 8.34197626e-02 4.84408170e-01
-1.18278526e-01 -2.23728925e-01 -1.81225672e-01 1.09639907e+00
-6.08014524e-01 -1.50634861e+00 -3.08436245e-01 4.81443495e-01
-2.65526533e-01 1.99713469e-01 -5.39400876e-01 9.30514574e-01
4.29616988e-01 7.32987404e-01 1.53664142e-01 -3.72641295e-01
4.54553455e-01 1.88541934e-01 7.00344205e-01 -9.00822759e-01
-5.70563912e-01 2.24131301e-01 -2.42228806e-01 -1.01705992e+00
-8.87060940e-01 -6.39067769e-01 -1.70830572e+00 2.06293732e-01
-3.47265869e-01 -6.83788657e-02 9.06368613e-01 1.05020702e+00
6.70308828e-01 3.98532569e-01 5.34970105e-01 -1.23672271e+00
6.09673671e-02 -5.99940658e-01 -4.43942636e-01 6.02905869e-01
-1.65585801e-02 -7.54418552e-01 -2.08496545e-02 -1.10833943e-01] | [8.010801315307617, -3.1636037826538086] |
f95c72a2-ccc5-4c95-8b66-ae99b1a94943 | towards-streaming-image-understanding | 2005.10420 | null | https://arxiv.org/abs/2005.10420v2 | https://arxiv.org/pdf/2005.10420v2.pdf | Towards Streaming Perception | Embodied perception refers to the ability of an autonomous agent to perceive its environment so that it can (re)act. The responsiveness of the agent is largely governed by latency of its processing pipeline. While past work has studied the algorithmic trade-off between latency and accuracy, there has not been a clear metric to compare different methods along the Pareto optimal latency-accuracy curve. We point out a discrepancy between standard offline evaluation and real-time applications: by the time an algorithm finishes processing a particular frame, the surrounding world has changed. To these ends, we present an approach that coherently integrates latency and accuracy into a single metric for real-time online perception, which we refer to as "streaming accuracy". The key insight behind this metric is to jointly evaluate the output of the entire perception stack at every time instant, forcing the stack to consider the amount of streaming data that should be ignored while computation is occurring. More broadly, building upon this metric, we introduce a meta-benchmark that systematically converts any single-frame task into a streaming perception task. We focus on the illustrative tasks of object detection and instance segmentation in urban video streams, and contribute a novel dataset with high-quality and temporally-dense annotations. Our proposed solutions and their empirical analysis demonstrate a number of surprising conclusions: (1) there exists an optimal "sweet spot" that maximizes streaming accuracy along the Pareto optimal latency-accuracy curve, (2) asynchronous tracking and future forecasting naturally emerge as internal representations that enable streaming perception, and (3) dynamic scheduling can be used to overcome temporal aliasing, yielding the paradoxical result that latency is sometimes minimized by sitting idle and "doing nothing". | ['Yu-Xiong Wang', 'Mengtian Li', 'Deva Ramanan'] | 2020-05-21 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3553_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123470460.pdf | eccv-2020-8 | ['real-time-multi-object-tracking'] | ['computer-vision'] | [ 6.22098327e-01 -2.37259772e-02 1.92746725e-02 -3.62893850e-01
-5.50829649e-01 -8.57049465e-01 8.06252301e-01 5.55925250e-01
-7.32802033e-01 1.73781261e-01 1.29450038e-01 -1.31930903e-01
-1.18990280e-01 -7.30533063e-01 -7.15090096e-01 -7.26229608e-01
-4.49902475e-01 2.06015423e-01 7.96319664e-01 -2.61142492e-01
2.11339757e-01 3.73979151e-01 -2.15775919e+00 2.85900027e-01
5.18848360e-01 1.32604039e+00 3.83012414e-01 1.19188380e+00
1.52223840e-01 9.34825122e-01 -7.04352975e-01 -2.46376604e-01
2.98297703e-01 -2.59275973e-01 -7.71905899e-01 2.21261308e-01
1.22480683e-01 -3.28379035e-01 1.42338173e-03 9.00275052e-01
3.55762184e-01 1.12438291e-01 2.48078138e-01 -1.54950190e+00
8.94026905e-02 5.02855182e-01 -2.93272793e-01 5.05676150e-01
5.20442963e-01 5.76040506e-01 9.59963679e-01 -3.82165015e-01
7.07399845e-01 1.07073116e+00 3.27949673e-01 1.00700676e-01
-1.08272648e+00 -1.53405532e-01 5.85086942e-01 1.93576589e-01
-9.79141831e-01 -7.27842510e-01 3.02244514e-01 -4.59547132e-01
1.14274037e+00 4.34108257e-01 7.91193426e-01 7.02322960e-01
4.09762084e-01 6.90930605e-01 9.40890431e-01 -8.51080567e-02
7.18928814e-01 -2.22399577e-01 -5.47142811e-02 2.95709014e-01
1.65244397e-02 1.97698116e-01 -9.22124982e-01 7.80049860e-02
2.82539517e-01 -1.87679604e-01 -1.61089554e-01 -2.10522145e-01
-1.55622590e+00 2.88943976e-01 1.32140502e-01 7.94080123e-02
-4.83246624e-01 4.18825656e-01 7.24974334e-01 5.58993042e-01
3.62146020e-01 2.09081441e-01 -3.63083720e-01 -6.52925551e-01
-9.69291031e-01 3.66977602e-01 8.89075458e-01 7.07937002e-01
6.22566044e-01 -9.51917246e-02 -2.31796652e-02 2.87348419e-01
2.93787140e-02 3.67719650e-01 3.02187115e-01 -1.40512943e+00
2.18398049e-01 3.99054408e-01 4.08755481e-01 -9.38066065e-01
-4.52598095e-01 -4.00738806e-01 -4.84080613e-01 6.34748161e-01
4.84295875e-01 -7.73768052e-02 -5.49692392e-01 1.81470168e+00
5.14740884e-01 3.13001931e-01 2.34952152e-01 1.12541318e+00
3.43041658e-01 6.30073667e-01 1.17834203e-01 -3.20373088e-01
1.40022051e+00 -7.84348965e-01 -2.72283196e-01 -4.52506006e-01
3.79927844e-01 -5.97285390e-01 9.93359447e-01 5.52776635e-01
-1.38085794e+00 -5.67241311e-01 -1.28179145e+00 1.50514007e-01
-2.90420681e-01 -7.15531647e-01 5.84209859e-01 3.71127516e-01
-1.21046841e+00 6.68734848e-01 -1.18663526e+00 -3.67933542e-01
-4.44171317e-02 2.52505302e-01 -5.81552982e-02 2.94366091e-01
-6.55640483e-01 5.97018361e-01 3.22560698e-01 -7.93210343e-02
-1.07487273e+00 -6.19903266e-01 -3.74013871e-01 4.36698981e-02
6.48316085e-01 -8.84033561e-01 1.69680250e+00 -1.52705967e+00
-1.54946470e+00 7.75029540e-01 -1.11251317e-01 -7.28411138e-01
7.74966359e-01 -1.04490958e-01 -2.01589257e-01 2.30170175e-01
2.20949240e-02 6.43978953e-01 8.11468303e-01 -1.31110859e+00
-1.24821723e+00 -4.24707800e-01 6.14474177e-01 4.49627966e-01
1.73740871e-02 5.35500422e-02 -6.26604140e-01 -3.37344617e-01
2.29541793e-01 -8.52074265e-01 -3.38555574e-01 4.22582254e-02
1.48128346e-01 -1.42585710e-01 5.56132019e-01 -1.59328923e-01
1.10535777e+00 -2.25616241e+00 8.87635723e-02 -2.33858023e-02
2.98031211e-01 -9.18370262e-02 -5.67313023e-02 5.17418981e-01
2.68701345e-01 8.10210500e-03 -3.26547116e-01 -4.77656990e-01
1.23545788e-01 3.71197551e-01 -5.36832452e-01 5.55526257e-01
6.68163598e-02 6.65280640e-01 -1.21226382e+00 -4.15304869e-01
6.42368197e-02 2.94408619e-01 -5.31011224e-01 1.55690417e-01
-5.49487472e-01 2.83910573e-01 -3.47983897e-01 5.91400385e-01
4.82135385e-01 -3.54666024e-01 1.09795325e-01 1.82715744e-01
-5.08839965e-01 3.94187599e-01 -1.35335481e+00 1.80288231e+00
-3.30990225e-01 7.29230642e-01 4.17759120e-01 -6.85363770e-01
4.58581328e-01 2.38923252e-01 6.32833540e-01 -1.13871622e+00
-4.94621396e-02 1.71434492e-01 -1.35369003e-01 -4.62742418e-01
8.47683966e-01 2.28961155e-01 -5.07579856e-02 4.48221594e-01
-4.98501629e-01 4.67185602e-02 3.52065563e-01 1.61423475e-01
1.55951583e+00 3.67382735e-01 1.26037955e-01 -9.73624587e-02
1.99464306e-01 3.65758061e-01 6.08389080e-01 8.52006435e-01
-4.11793172e-01 4.19001698e-01 6.17537200e-01 -4.87959445e-01
-1.04127181e+00 -1.22556829e+00 2.27404878e-01 1.57073319e+00
6.48456156e-01 -3.25054377e-01 -6.85593963e-01 -1.38039201e-01
-1.54490635e-01 3.76442522e-01 -5.48568368e-01 2.02874780e-01
-5.54823399e-01 -4.77542877e-01 3.19354415e-01 4.52969998e-01
4.15515780e-01 -1.03813469e+00 -1.95157778e+00 5.67924619e-01
-1.67088166e-01 -1.32526457e+00 -1.81601465e-01 6.67958409e-02
-7.23743141e-01 -8.90277088e-01 -2.37308696e-01 -1.02767661e-01
1.41708538e-01 6.77635491e-01 1.45496130e+00 1.93360880e-01
-2.02047713e-02 7.30045795e-01 -5.66894531e-01 -5.01277447e-01
-3.71935040e-01 -1.91059917e-01 -1.35480657e-01 7.78155923e-02
-1.20363921e-01 -6.92381144e-01 -9.83229339e-01 2.06157371e-01
-1.18670177e+00 2.20988035e-01 2.91995645e-01 3.05255264e-01
7.83200145e-01 3.33879367e-02 3.84940267e-01 -5.00963628e-01
3.66560251e-01 -5.88823795e-01 -7.91223884e-01 1.00269355e-01
-4.74467635e-01 -8.34854990e-02 5.48881710e-01 -3.60375255e-01
-9.11673009e-01 1.29888073e-01 1.73584931e-03 -2.19204620e-01
-1.23637035e-01 2.68490404e-01 9.08434615e-02 2.83722758e-01
3.83556455e-01 3.45102876e-01 2.19057817e-02 2.82597467e-02
4.19279546e-01 3.27596903e-01 8.94062102e-01 -5.20449758e-01
2.49996588e-01 1.06946278e+00 -3.66083682e-02 -6.82811379e-01
-6.35573149e-01 -5.25460482e-01 -3.74819070e-01 -5.35041749e-01
8.33908379e-01 -7.74985611e-01 -1.30918860e+00 3.38417590e-01
-1.17823362e+00 -5.45511544e-01 -5.27541816e-01 1.92242727e-01
-9.91231978e-01 1.81839392e-01 -2.59093910e-01 -1.03038490e+00
-1.33517861e-01 -1.16272807e+00 1.19829774e+00 1.64722934e-01
-1.17025413e-01 -5.96966743e-01 4.24145274e-02 1.43678263e-01
4.34860915e-01 4.45078552e-01 3.18073481e-01 -3.83584768e-01
-9.69170332e-01 2.06618950e-01 -2.87129700e-01 -3.48522253e-02
-4.68456268e-01 1.68377325e-01 -1.12742472e+00 -3.19145590e-01
7.48860911e-02 1.45568624e-01 7.66310215e-01 2.84563601e-01
8.68788719e-01 -2.91378587e-01 -1.65965572e-01 5.83299339e-01
1.71053016e+00 3.31038326e-01 4.60527301e-01 5.35614491e-01
1.45753115e-01 7.83463657e-01 8.62369835e-01 7.76119590e-01
7.60758460e-01 6.54025018e-01 8.59759629e-01 1.78737894e-01
-2.60425378e-02 1.11181468e-01 6.53813303e-01 5.95986366e-01
-2.18759045e-01 -5.25653720e-01 -8.61888289e-01 6.14719570e-01
-2.19368410e+00 -1.03304863e+00 1.26570746e-01 2.44128346e+00
3.30817014e-01 5.75633645e-01 3.28087568e-01 2.64429927e-01
2.13741526e-01 3.52200866e-01 -7.57095993e-01 -5.47817111e-01
-4.51978371e-02 -1.63226575e-01 6.32941961e-01 4.59515750e-01
-8.51348579e-01 6.68223917e-01 6.10855103e+00 4.64097708e-01
-1.30956769e+00 3.05189341e-01 5.63540518e-01 -3.26817065e-01
-2.68582016e-01 1.57102555e-01 -4.15147096e-01 5.47506869e-01
1.19176173e+00 -2.42542982e-01 4.82198685e-01 6.85944855e-01
4.67463344e-01 -5.91128588e-01 -1.43478858e+00 8.82740915e-01
-2.08151594e-01 -1.36339319e+00 -3.73180985e-01 5.89588955e-02
2.97850579e-01 3.91959935e-01 -6.45940229e-02 -1.75868273e-01
3.42588365e-01 -8.80649507e-01 1.41240394e+00 6.90078080e-01
4.71912980e-01 -5.73024333e-01 4.07781124e-01 4.75903779e-01
-1.46296155e+00 -2.38654017e-01 5.55793196e-02 -4.04473156e-01
2.97679812e-01 4.01738048e-01 -6.20580852e-01 4.78890568e-01
8.69643509e-01 3.27456087e-01 -3.91060352e-01 1.02015996e+00
3.08815718e-01 3.01900506e-01 -4.86207753e-01 -5.51798788e-04
3.75770152e-01 -4.02788855e-02 7.14264274e-01 1.39826691e+00
2.01303780e-01 1.46793962e-01 2.37180337e-01 6.35316968e-01
3.24519306e-01 -2.73762643e-01 -4.07642245e-01 1.67789429e-01
5.33468783e-01 9.56581593e-01 -1.23119366e+00 -5.51833630e-01
-4.94290739e-01 8.16580951e-01 -2.87661999e-02 3.01305413e-01
-9.08991456e-01 -3.24064791e-02 1.01924086e+00 8.93484429e-02
3.29037905e-01 -5.02165258e-01 -4.83270437e-01 -8.21225822e-01
2.08119452e-01 -7.56567717e-01 3.16403210e-01 -7.72645593e-01
-7.75458574e-01 5.73616803e-01 -1.07314587e-01 -1.17375708e+00
-2.99362451e-01 -2.64423609e-01 -3.18468332e-01 2.05911621e-01
-1.55270851e+00 -6.19770050e-01 -5.00468493e-01 3.07279348e-01
7.55581200e-01 3.94793600e-01 5.60766518e-01 1.69797897e-01
-3.56767565e-01 7.36203790e-02 -1.50242642e-01 -4.14757431e-01
3.26546878e-01 -1.13685763e+00 4.41607386e-01 1.10445833e+00
5.38151227e-02 2.88764331e-02 1.14974368e+00 -2.83519804e-01
-1.71335304e+00 -7.53733754e-01 5.89145839e-01 -5.20978332e-01
7.33292103e-01 -2.05540314e-01 -7.99261570e-01 4.22384918e-01
1.40728191e-01 1.79400798e-02 1.94206461e-01 -2.85876453e-01
-4.23723400e-01 -3.76494735e-01 -8.54837835e-01 6.31297529e-01
1.19904864e+00 -3.20993543e-01 -1.79029822e-01 1.09614834e-01
1.09847474e+00 -4.33715582e-01 -8.33920956e-01 3.39517355e-01
7.79956043e-01 -1.64007223e+00 9.13163364e-01 -1.71769738e-01
3.38910311e-01 -4.44851905e-01 -4.06505436e-01 -8.03342879e-01
3.63157205e-02 -8.38787913e-01 -1.71726122e-01 9.56808627e-01
9.20300633e-02 -5.98984718e-01 6.99512601e-01 7.04937220e-01
-2.46116087e-01 -8.97112608e-01 -1.02731848e+00 -7.01050341e-01
-5.31166196e-01 -9.55025375e-01 7.62942016e-01 4.55144167e-01
-2.21964866e-01 -7.01680854e-02 2.76112854e-01 4.92968380e-01
6.17824972e-01 3.48179698e-01 8.23068023e-01 -9.44079220e-01
-3.22168022e-01 -6.39200211e-01 -5.28485537e-01 -1.39115560e+00
-3.44010621e-01 -3.71664971e-01 2.96411306e-01 -1.39537942e+00
-2.41277888e-02 -5.32289624e-01 -2.32231021e-01 1.92338556e-01
2.03927845e-01 -5.49186319e-02 5.25071859e-01 3.97948891e-01
-1.15644205e+00 1.15271248e-01 1.13254809e+00 2.21450210e-01
-2.31476873e-01 3.93040255e-02 -4.62716371e-01 7.86050022e-01
4.89057541e-01 -2.59101927e-01 -5.62717021e-01 -7.04220712e-01
7.12126255e-01 4.29451138e-01 6.72814488e-01 -1.22388661e+00
6.67155981e-01 -4.11023855e-01 -1.37960821e-01 -4.45848137e-01
5.98776460e-01 -9.27401960e-01 1.54141560e-01 4.43089902e-01
-2.39545316e-01 4.57885802e-01 -3.24540846e-02 7.27149069e-01
-2.03284875e-01 1.09009646e-01 5.83887458e-01 -1.86341375e-01
-1.20778167e+00 2.54234612e-01 -5.55110037e-01 9.73766297e-02
1.22974598e+00 -5.91065466e-01 -2.89819986e-01 -3.57398927e-01
-6.09347105e-01 3.94166589e-01 8.21454704e-01 4.11039799e-01
4.18390542e-01 -7.33627975e-01 -6.17958724e-01 -2.02795733e-02
8.01937655e-02 1.44071907e-01 2.13822410e-01 8.25358391e-01
-5.28464973e-01 8.57228041e-02 -5.02997972e-02 -9.46005940e-01
-1.07758665e+00 4.97672558e-01 3.46425802e-01 -1.54721081e-01
-6.62691653e-01 7.88834870e-01 1.32955089e-01 2.18774319e-01
4.77205187e-01 -5.92937350e-01 1.42473772e-01 1.76703930e-01
5.67820728e-01 5.50081789e-01 1.50807753e-01 -6.02879286e-01
-3.98006976e-01 2.90359646e-01 4.17265922e-01 -4.15266544e-01
1.32677495e+00 -5.81533790e-01 -8.69769603e-02 6.60939097e-01
8.10942948e-01 -1.96464285e-01 -1.80387533e+00 -4.59275618e-02
1.01607323e-01 -4.71712768e-01 -6.40521646e-02 -5.83215237e-01
-7.26617396e-01 6.42499506e-01 7.62941360e-01 8.53558242e-01
1.49625814e+00 -4.53041913e-03 7.96280146e-01 5.59452921e-02
5.67882001e-01 -1.28677189e+00 3.70705202e-02 4.94022012e-01
6.71487808e-01 -1.16515005e+00 -1.43045142e-01 -3.08440149e-01
-6.79072022e-01 8.62397432e-01 3.16127717e-01 -3.56215239e-02
3.40488672e-01 6.14104927e-01 -3.08042690e-02 -1.39038488e-01
-1.34291553e+00 -3.13821524e-01 -2.15883300e-01 4.22579080e-01
3.79297435e-02 1.85591325e-01 3.21153402e-02 3.54770720e-01
-3.49226534e-01 1.80947911e-02 4.51995134e-01 1.07525098e+00
-6.94833100e-01 -6.81386352e-01 -3.08964968e-01 7.89514035e-02
-3.11040312e-01 2.88859367e-01 5.04364201e-04 4.76394296e-01
2.88687170e-01 1.21531713e+00 4.53508347e-01 -2.52803296e-01
3.45328271e-01 -2.75636256e-01 3.04090440e-01 -4.71339345e-01
-7.14301288e-01 -3.60216871e-02 1.48685589e-01 -1.17868221e+00
-7.10752904e-01 -6.70537949e-01 -1.45984411e+00 -4.69781309e-01
2.02591121e-01 -1.92044616e-01 8.78226280e-01 9.79721546e-01
6.16517127e-01 7.44929850e-01 5.40936887e-01 -1.11063552e+00
-3.41564178e-01 -2.79943109e-01 -1.26599684e-01 3.13843101e-01
7.59983003e-01 -3.12780976e-01 -3.17208648e-01 2.53804028e-01] | [8.421772956848145, -0.6509853005409241] |
268d0b47-7340-492c-a155-c3b75090157a | medical-concept-normalization-in-user-1 | null | null | https://aclanthology.org/2020.louhi-1.3 | https://aclanthology.org/2020.louhi-1.3.pdf | Medical Concept Normalization in User-Generated Texts by Learning Target Concept Embeddings | Medical concept normalization helps in discovering standard concepts in free-form text i.e., maps health-related mentions to standard concepts in a clinical knowledge base. It is much beyond simple string matching and requires a deep semantic understanding of concept mentions. Recent research approach concept normalization as either text classification or text similarity. The main drawback in existing a) text classification approach is ignoring valuable target concepts information in learning input concept mention representation b) text similarity approach is the need to separately generate target concept embeddings which is time and resource consuming. Our proposed model overcomes these drawbacks by jointly learning the representations of input concept mention and target concepts. First, we learn input concept mention representation using RoBERTa. Second, we find cosine similarity between embeddings of input concept mention and all the target concepts. Here, embeddings of target concepts are randomly initialized and then updated during training. Finally, the target concept with maximum cosine similarity is assigned to the input concept mention. Our model surpasses all the existing methods across three standard datasets by improving accuracy up to 2.31%. | ['Sivanesan Sangeetha', 'Katikapalli Subramanyam Kalyan'] | null | null | null | null | emnlp-louhi-2020-11 | ['medical-concept-normalization', 'clinical-knowledge'] | ['medical', 'miscellaneous'] | [ 5.76249182e-01 3.17867994e-01 -4.17965263e-01 -3.99551094e-01
-6.62808478e-01 -2.79199809e-01 4.01186287e-01 1.38200307e+00
-9.45485830e-01 5.09086609e-01 6.42981112e-01 -1.67830870e-01
-3.33355248e-01 -1.05162978e+00 -1.47116423e-01 -7.18127549e-01
1.27011999e-01 6.17544532e-01 1.09521382e-01 -4.00430590e-01
2.36063585e-01 1.45314097e-01 -1.39952767e+00 2.89502412e-01
8.77624631e-01 5.91403127e-01 4.39330935e-02 4.10007238e-01
-7.76663065e-01 2.85144687e-01 -8.45049441e-01 -5.41340947e-01
-6.72984496e-02 -3.04397583e-01 -8.28435600e-01 -7.12612987e-01
1.73763618e-01 2.63635576e-01 -3.08116972e-02 1.25097966e+00
5.66822708e-01 3.38996053e-01 9.00621057e-01 -1.07389617e+00
-1.03386903e+00 6.47783220e-01 -4.52417880e-01 3.01221222e-01
2.38547668e-01 -6.06388509e-01 1.07795954e+00 -7.21838176e-01
6.59136951e-01 1.04326057e+00 8.99883926e-01 7.94563532e-01
-9.32449341e-01 -7.32206285e-01 1.83360398e-01 4.01853696e-02
-1.45681036e+00 -6.88086972e-02 2.94506013e-01 -4.11162764e-01
1.26545191e+00 3.13683867e-01 3.34816337e-01 7.12839127e-01
2.54600108e-01 3.26911718e-01 2.71643549e-01 -5.84373653e-01
2.91713893e-01 2.50620723e-01 4.95576024e-01 4.14316416e-01
7.94150233e-01 -2.05109805e-01 -1.49165690e-01 -5.41537404e-01
3.18941772e-01 4.65794355e-01 -1.11818440e-01 -1.31734237e-01
-1.15827048e+00 1.05193806e+00 3.91399562e-01 7.04673111e-01
-2.59833664e-01 3.19131985e-02 7.64035404e-01 1.43260002e-01
1.98230207e-01 6.58857703e-01 -6.61759913e-01 1.50460020e-01
-6.91281736e-01 1.12325005e-01 5.27356148e-01 1.02865851e+00
5.70491314e-01 -4.14340317e-01 -1.33158430e-01 7.87241280e-01
1.45173967e-01 2.77743220e-01 1.11561525e+00 -5.59700690e-02
3.75437737e-01 1.01237190e+00 -3.13468367e-01 -1.14601803e+00
-5.43681681e-01 -3.73356819e-01 -8.35432231e-01 -3.98336977e-01
7.17454478e-02 -1.87269994e-03 -1.08932722e+00 1.66219783e+00
3.87047470e-01 3.56245160e-01 5.68755567e-01 3.91606301e-01
1.26706421e+00 3.01550239e-01 6.60004079e-01 -9.03774947e-02
1.78843617e+00 -6.79483712e-01 -9.85282600e-01 -1.16203584e-01
1.23579741e+00 -9.27687526e-01 7.84900129e-01 -2.93941826e-01
-3.37268502e-01 -2.28065312e-01 -1.13850737e+00 -2.07532063e-01
-9.27022755e-01 -7.74833933e-02 8.03008914e-01 8.66888821e-01
-5.85795462e-01 6.14724278e-01 -7.56646872e-01 -7.74984062e-01
5.04883051e-01 6.51498973e-01 -6.66339040e-01 -2.95288861e-01
-1.42163062e+00 9.03984725e-01 8.94992292e-01 -8.22359145e-01
-4.36180010e-02 -1.34282959e+00 -1.14545238e+00 1.60938531e-01
5.92958331e-02 -8.66033494e-01 8.97007167e-01 -6.02295995e-01
-9.43886995e-01 7.40839422e-01 -1.80495009e-01 -4.01848823e-01
-1.68659851e-01 -1.00283667e-01 -9.38560009e-01 1.89475324e-02
2.64072537e-01 5.17567933e-01 3.41102958e-01 -7.71568716e-01
-7.61355698e-01 -3.20610791e-01 -7.15505853e-02 3.57197136e-01
-7.68798351e-01 -8.39240775e-02 -1.76642373e-01 -8.39334846e-01
4.17964995e-01 -4.06663418e-01 -4.36152726e-01 -1.82613805e-01
-2.18458295e-01 -3.15832108e-01 4.66454685e-01 -3.34452748e-01
1.43557203e+00 -1.99740338e+00 -3.41419250e-01 3.64234805e-01
4.19979692e-01 3.27940434e-01 -1.91135377e-01 4.80979353e-01
-7.15774477e-01 2.23513111e-01 -2.76419014e-01 -1.07071800e-02
-1.56504154e-01 2.86410600e-01 -1.19839139e-01 2.80371755e-01
2.12202549e-01 8.05983186e-01 -1.25349414e+00 -5.16348302e-01
1.43814251e-01 6.03558421e-01 -8.13430369e-01 -4.31627482e-02
1.07650153e-01 -9.87405404e-02 -4.15871441e-01 5.02855182e-01
7.18002379e-01 -8.91990308e-03 3.66462588e-01 -3.56976837e-01
3.08412254e-01 5.17889738e-01 -1.22285044e+00 1.78647590e+00
-2.91143298e-01 1.81357354e-01 -6.60475969e-01 -1.38745034e+00
1.13978553e+00 5.55670798e-01 8.82137835e-01 -4.60243195e-01
2.61617959e-01 1.44139588e-01 3.08125056e-02 -5.30988395e-01
4.93003309e-01 -7.06745565e-01 -1.49833009e-01 4.00132924e-01
2.95055062e-01 3.12404335e-01 1.19636349e-01 3.58143955e-01
1.06010544e+00 -3.46461087e-01 8.65472376e-01 -4.22990948e-01
6.26338363e-01 -5.45112826e-02 7.04446673e-01 5.86091220e-01
1.02529794e-01 5.34104884e-01 3.61811489e-01 -3.74368995e-01
-6.68753624e-01 -1.14220428e+00 -3.81175965e-01 1.12879825e+00
1.12433530e-01 -9.32786405e-01 -4.67153966e-01 -7.47731507e-01
2.63763279e-01 7.24190235e-01 -1.05299819e+00 -6.32658482e-01
-4.84389126e-01 -1.04632175e+00 6.75262034e-01 7.46325552e-01
-2.10625127e-01 -6.59322858e-01 -2.98649251e-01 4.58449244e-01
-1.12368770e-01 -7.20275819e-01 -5.83184004e-01 2.31635079e-01
-1.12571514e+00 -1.39070630e+00 -5.16447544e-01 -9.53386366e-01
1.06457293e+00 6.17850795e-02 9.92773831e-01 2.60769486e-01
-7.97150970e-01 2.76688337e-01 -5.13640940e-01 -7.10620463e-01
-3.38684648e-01 1.72529593e-01 3.98594260e-01 -3.28906894e-01
1.25050843e+00 -3.94531280e-01 -5.18706918e-01 -4.28985711e-03
-1.19885027e+00 -3.74654174e-01 5.57244420e-01 9.00665879e-01
5.88779867e-01 -2.42838636e-02 7.76035905e-01 -1.18545365e+00
8.42727363e-01 -8.04789245e-01 9.98333935e-03 3.65127206e-01
-8.94377172e-01 2.52923280e-01 4.46503043e-01 -6.17850423e-01
-6.92395806e-01 9.70242321e-02 -3.90179992e-01 2.25052938e-01
-2.12833598e-01 6.39401019e-01 -1.02031857e-01 3.70767623e-01
9.61676002e-01 1.14297280e-02 -1.67811453e-01 -5.60573101e-01
5.68495452e-01 6.59025550e-01 5.63924611e-01 -5.33797383e-01
7.24866152e-01 3.38541806e-01 -2.05010206e-01 -5.79604030e-01
-7.20731914e-01 -1.08790326e+00 -8.83670151e-01 6.53462350e-01
9.12030816e-01 -7.99696565e-01 -6.51447415e-01 -2.35078290e-01
-1.00449276e+00 7.32499242e-01 -5.51539779e-01 8.68477941e-01
-9.67896730e-02 4.79357928e-01 -2.63212174e-01 -3.75866681e-01
-6.90721452e-01 -5.90635836e-01 7.92785287e-01 2.83283800e-01
-6.94259226e-01 -1.22798741e+00 3.39420795e-01 -6.77389726e-02
2.85425335e-01 2.08000854e-01 1.43553948e+00 -1.34062350e+00
3.65465671e-01 -7.63983607e-01 -1.36345148e-01 -8.24668482e-02
8.59250605e-01 -3.33631009e-01 -7.71545649e-01 -2.89650291e-01
-4.51235473e-02 2.12744489e-01 8.09125185e-01 2.32491180e-01
8.98485899e-01 -4.18719351e-01 -7.65303969e-01 5.99339306e-01
1.60556412e+00 1.99505478e-01 5.57996571e-01 4.30683523e-01
4.89450127e-01 5.84799886e-01 6.17088556e-01 2.20376521e-01
1.17174737e-01 3.88904542e-01 -2.04033896e-01 -1.96087152e-01
1.49497956e-01 -1.56291425e-01 -1.27779052e-01 8.96681011e-01
2.11210072e-01 -1.64769739e-02 -1.01055396e+00 8.83135557e-01
-1.54939342e+00 -8.31780672e-01 -4.96069528e-02 2.14248538e+00
1.24190390e+00 -3.83716170e-03 -2.96414465e-01 2.87206590e-01
6.97147191e-01 -5.18256664e-01 -4.31243837e-01 -4.06443983e-01
1.25228301e-01 6.25926495e-01 4.60508734e-01 3.44682962e-01
-1.06123948e+00 8.11979473e-01 5.58873606e+00 6.67936146e-01
-8.18767965e-01 2.90185809e-01 -3.41969281e-02 -1.77579328e-01
-3.26251477e-01 -2.65187413e-01 -8.81806731e-01 2.11495802e-01
9.28491473e-01 -8.71706247e-01 -4.80033368e-01 8.60715985e-01
-3.87228429e-01 1.79111317e-01 -1.29173338e+00 1.08804548e+00
3.55937839e-01 -1.46407962e+00 5.33558547e-01 -3.05584788e-01
6.72368646e-01 -3.00140440e-01 -1.94129065e-01 4.53221887e-01
2.46598363e-01 -1.26004565e+00 -1.21973865e-01 3.70349556e-01
9.98831928e-01 -8.52053940e-01 1.35501099e+00 -1.82312161e-01
-1.36585116e+00 -1.37957453e-03 -8.05183887e-01 2.38655195e-01
-1.17990442e-01 7.20902085e-01 -1.04397297e+00 8.03988993e-01
5.25050163e-01 6.52305543e-01 -3.82133991e-01 1.16189849e+00
-1.41485557e-01 2.41522461e-01 -2.53718585e-01 1.75457001e-01
1.00263231e-01 1.62664354e-01 1.68526351e-01 1.53171945e+00
2.39236251e-01 4.59165990e-01 1.32759839e-01 4.21249628e-01
-1.17942005e-01 7.93648958e-01 -4.69805181e-01 -1.49342880e-01
6.39497697e-01 1.11808908e+00 -8.51722956e-01 -6.05990052e-01
-3.40827733e-01 8.45010757e-01 -9.91315395e-02 4.35207821e-02
-6.29756808e-01 -1.09149122e+00 1.21519125e+00 2.08660457e-02
1.41810840e-02 3.06502789e-01 -3.04733992e-01 -9.96060550e-01
-2.11925447e-01 -4.41553503e-01 9.12591934e-01 -2.46870592e-02
-1.42390883e+00 5.93489468e-01 -1.36344299e-01 -1.26465952e+00
-4.41387761e-03 -7.15290129e-01 -5.59771180e-01 9.35545921e-01
-1.53093612e+00 -8.41403842e-01 -1.88647225e-01 7.15886354e-01
1.85076326e-01 -3.06377798e-01 1.66721237e+00 7.40131438e-01
-2.05158725e-01 1.05684721e+00 4.57158625e-01 5.09852707e-01
1.07180595e+00 -1.25890744e+00 5.05274594e-01 3.43660235e-01
-1.32575512e-01 1.35617518e+00 5.91510653e-01 -8.58156621e-01
-6.28963292e-01 -1.22342002e+00 1.47025132e+00 -6.08913124e-01
5.20605862e-01 -2.75815457e-01 -1.23356414e+00 5.43437898e-01
-1.95852697e-01 -4.10135314e-02 1.68404090e+00 2.04371914e-01
-6.54049158e-01 3.66069004e-03 -1.35857117e+00 5.58000386e-01
7.25640714e-01 -4.52829391e-01 -1.29586911e+00 4.16997641e-01
1.03866577e+00 -2.71250516e-01 -1.15904939e+00 1.64150387e-01
3.94373596e-01 -4.74519543e-02 1.29768002e+00 -1.30748260e+00
1.93296537e-01 -3.79423350e-01 -2.10934013e-01 -1.18770838e+00
-2.77468264e-01 -1.37148246e-01 1.35687917e-01 1.13702691e+00
5.75416088e-01 -6.18154347e-01 7.71816373e-01 6.57330751e-01
-2.80072838e-02 -7.09076762e-01 -9.23320889e-01 -7.18347728e-01
4.66919005e-01 -1.42961994e-01 8.85557532e-01 1.55654442e+00
4.97915179e-01 4.16848838e-01 1.98824584e-01 1.52274236e-01
4.57978010e-01 -8.60390738e-02 1.97145283e-01 -1.58210111e+00
2.55378067e-01 -3.95753622e-01 -9.45785165e-01 -4.43125844e-01
-1.38218552e-01 -1.43938982e+00 -2.04391330e-01 -1.71405768e+00
2.17123747e-01 -6.04163766e-01 -7.50944614e-01 7.52798200e-01
-4.32429314e-01 5.89728355e-03 -2.19045296e-01 -4.51227762e-02
-2.67251760e-01 4.08891827e-01 6.73731029e-01 -3.38111371e-01
-2.64614105e-01 -1.60714909e-01 -1.26971304e+00 5.96472681e-01
1.00714457e+00 -1.14598143e+00 -5.72110891e-01 -3.13337207e-01
5.58128767e-02 -5.31438828e-01 -1.60118073e-01 -8.32624137e-01
4.45566624e-01 -1.52092591e-01 3.54345739e-01 -4.29510981e-01
-4.69908342e-02 -8.83444726e-01 -1.76243573e-01 7.45934844e-01
-5.22005796e-01 1.85901180e-01 4.38440919e-01 6.60259724e-01
-1.27305418e-01 -6.38800085e-01 8.03004265e-01 -1.04663976e-01
-6.07178330e-01 9.14188847e-02 -2.26463944e-01 2.74069101e-01
1.04978275e+00 -3.78420323e-01 -2.28813350e-01 2.86462575e-01
-8.01410973e-01 8.22925717e-02 1.18664093e-01 7.04883397e-01
6.43392384e-01 -1.40116549e+00 -6.21874452e-01 1.91873550e-01
6.11010313e-01 -1.26595989e-01 7.90800229e-02 5.22892535e-01
-3.67390931e-01 6.81157053e-01 -1.98265225e-01 -2.17620268e-01
-1.51082873e+00 7.56104112e-01 2.45438263e-01 -1.76338196e-01
-8.62399101e-01 1.02077579e+00 2.47159228e-01 -8.81498694e-01
2.41777763e-01 -4.38152909e-01 -6.83668911e-01 3.23452741e-01
9.79856670e-01 4.16322500e-02 4.47850019e-01 -5.23196876e-01
-7.68014669e-01 7.83288002e-01 -4.83740866e-01 4.11966622e-01
1.55924678e+00 2.81194389e-01 -2.80808508e-01 5.60175627e-02
1.56929243e+00 -7.15320855e-02 -3.63126509e-02 -5.16809642e-01
5.04816234e-01 -3.78828675e-01 -1.26706406e-01 -6.89220965e-01
-8.39359760e-01 7.78424561e-01 8.64335299e-01 -6.31154299e-01
1.03746116e+00 -1.01048984e-02 6.46769285e-01 7.98644722e-01
-2.26584136e-01 -1.06448138e+00 -7.03943521e-02 4.35643494e-01
4.50683475e-01 -1.10600495e+00 2.95543611e-01 -5.18844247e-01
-4.26554173e-01 1.31633532e+00 5.25373816e-01 1.76326469e-01
8.10155153e-01 2.25734711e-01 2.21820503e-01 -4.19188291e-01
-4.79671091e-01 -8.97526592e-02 4.38133299e-01 9.32126939e-01
8.49123955e-01 -3.12002376e-02 -6.45051539e-01 1.07369161e+00
-2.91101426e-01 -7.95199424e-02 3.90333027e-01 8.79953384e-01
-3.99875939e-01 -1.42510676e+00 -2.50956714e-01 8.13574910e-01
-6.25768661e-01 -6.20078683e-01 -1.83952283e-02 6.91897690e-01
3.89524549e-01 7.16282666e-01 2.88886845e-01 -2.64442027e-01
5.86839080e-01 2.84526438e-01 -4.36390936e-02 -1.26090717e+00
-6.33346796e-01 -4.87603575e-01 -3.71821761e-01 -1.94815740e-01
-3.16692382e-01 -4.29666460e-01 -1.90077281e+00 -1.46305114e-01
-5.79031706e-01 6.52457356e-01 6.47171676e-01 8.80088449e-01
2.97077626e-01 7.15258002e-01 -6.22582659e-02 1.03499182e-01
-3.47513467e-01 -8.23563576e-01 -3.71869892e-01 7.39954412e-01
1.94077745e-01 -6.81056499e-01 -1.34744257e-01 1.73074991e-01] | [8.523159980773926, 8.551177978515625] |
51c11f4f-f68a-4ae4-b2a8-4233bf754d10 | a-comprehensive-review-on-the-nilm-algorithms | 2102.12578 | null | https://arxiv.org/abs/2102.12578v2 | https://arxiv.org/pdf/2102.12578v2.pdf | A Comprehensive Review on the NILM Algorithms for Energy Disaggregation | The housing structures have changed with urbanization and the growth due to the construction of high-rise buildings all around the world requires end-use appliance energy conservation and management in real-time. This shift also came along with smart-meters which enabled the estimation of appliance-specific power consumption from the buildings aggregate power consumption reading. Non-intrusive load monitoring (NILM) or energy disaggregation is aimed at separating the household energy measured at the aggregate level into constituent appliances. Over the years, signal processing and machine learning algorithms have been combined to achieve this. Incredible research and publications have been conducted on energy disaggregation, non-intrusive load monitoring, home energy management and appliance classification. There exists an API, NILMTK, a reproducible benchmark algorithm for the same. Many other approaches to perform energy disaggregation has been adapted such as deep neural network architectures and big data approach for household energy disaggregation. This paper provides a survey of the effective NILM system frameworks and reviews the performance of the benchmark algorithms in a comprehensive manner. This paper also summarizes the wide application scope and the effectiveness of the algorithmic performance on three publicly available data sets. | ['Abbas Kouzani', 'Mohiuddin Ahmed', 'M. A. Parvez Mahmud', 'Adnan Anwar', 'Akriti Verma'] | 2021-02-20 | null | null | null | null | ['non-intrusive-load-monitoring', 'non-intrusive-load-monitoring', 'non-intrusive-load-monitoring'] | ['knowledge-base', 'miscellaneous', 'time-series'] | [-3.24716806e-01 -1.93274707e-01 1.95808172e-01 -5.27839661e-01
-4.91345674e-01 -2.87324250e-01 5.69589674e-01 -1.31146550e-01
9.00492668e-02 9.14169729e-01 4.42090243e-01 5.06547466e-02
-6.90948665e-02 -1.24938476e+00 -1.42197043e-01 -1.21419084e+00
2.65436750e-02 4.86371636e-01 -5.84161758e-01 -6.34013787e-02
-5.06527603e-01 3.68575543e-01 -1.60859907e+00 1.25089452e-01
7.93795764e-01 1.23057449e+00 -3.33449692e-02 7.15743601e-01
3.72919947e-01 7.73266554e-01 -5.39431810e-01 1.90433457e-01
2.90553838e-01 -3.64042163e-01 -7.09657669e-01 -2.85208076e-01
-1.01007581e-01 -6.96435571e-01 -4.95012961e-02 8.85751903e-01
1.08731234e+00 2.53812987e-02 6.82880700e-01 -1.59107375e+00
-5.76314867e-01 8.64850879e-01 -3.55105102e-01 1.71944991e-01
4.03094232e-01 1.04988314e-01 6.82768583e-01 -1.65543258e-02
-3.15855384e-01 8.16069663e-01 7.48586178e-01 4.70426120e-03
-1.28308451e+00 -5.48192978e-01 -4.55318063e-01 6.98755920e-01
-1.47130823e+00 -3.29281092e-01 7.89237916e-01 -2.88962960e-01
2.06016016e+00 9.28535759e-01 7.73877621e-01 1.05147064e+00
2.75637060e-01 6.67688191e-01 1.10510659e+00 -3.63447130e-01
6.26441836e-01 -1.41659873e-02 2.87786573e-01 6.25044480e-02
5.20435452e-01 4.46457952e-01 6.82369694e-02 -1.11371592e-01
-2.60138720e-01 1.11620702e-01 9.26314220e-02 2.15827778e-01
-7.11890280e-01 5.57153702e-01 3.25524420e-01 8.68794799e-01
-5.72302878e-01 1.95051640e-01 6.76710784e-01 2.05414042e-01
3.22104216e-01 -1.50964826e-01 -7.25544930e-01 -2.86193192e-01
-1.27943170e+00 -4.50693481e-02 9.54552531e-01 8.13150167e-01
6.12864733e-01 3.49412918e-01 -6.74251243e-02 5.24326563e-01
4.76857215e-01 6.18927777e-01 8.61276686e-01 -1.04436612e+00
2.20176294e-01 7.72731423e-01 -4.06135954e-02 -5.96492648e-01
-1.03661704e+00 1.30246952e-01 -1.74814606e+00 2.64863998e-01
-2.23200768e-01 -4.51852560e-01 -4.86311257e-01 1.42618501e+00
3.36442441e-01 -1.64099023e-01 -8.63058306e-03 2.29912817e-01
7.23431647e-01 1.07453322e+00 3.44205230e-01 -4.64268208e-01
1.46203279e+00 -4.66724962e-01 -1.10951102e+00 4.11378980e-01
3.67917836e-01 -2.89160013e-01 4.03857201e-01 2.93388724e-01
-8.47489893e-01 -5.64504504e-01 -1.19303226e+00 -1.82077810e-01
-1.18127882e+00 -1.94963872e-01 2.52000421e-01 1.11077905e+00
-9.38089907e-01 7.61909902e-01 -1.30837166e+00 -6.49760008e-01
6.25441909e-01 6.08943880e-01 -7.39411935e-02 6.35558248e-01
-1.10112619e+00 1.11731541e+00 7.30969191e-01 1.87841207e-01
-4.15784985e-01 -7.39043951e-01 -6.25058830e-01 2.13503644e-01
-3.90668154e-01 -7.83732831e-01 1.19867349e+00 -6.15769565e-01
-1.62932420e+00 5.21610618e-01 3.53539169e-01 -6.54820919e-01
4.32414353e-01 -1.79417446e-01 -8.55028629e-01 -5.43452263e-01
-1.25349134e-01 2.20948502e-01 4.13528711e-01 -6.28864825e-01
-1.18068290e+00 -5.77808380e-01 -5.40133953e-01 -2.60593176e-01
-1.57806903e-01 -2.54108191e-01 8.77049983e-01 -3.53685528e-01
-5.32714367e-01 -6.18173182e-01 2.31531903e-01 -1.09706450e+00
-6.56489491e-01 -5.92951953e-01 1.39049387e+00 -8.66392851e-01
1.44407237e+00 -1.80385494e+00 -3.57246876e-01 9.46850181e-02
-4.70012762e-02 3.65416519e-02 5.38719535e-01 7.33330369e-01
-4.92156148e-01 -1.67499110e-01 -2.67792016e-01 -4.13691610e-01
7.74969816e-01 3.67908239e-01 3.86371203e-02 6.05959177e-01
-4.82446432e-01 1.31885409e+00 -5.69765627e-01 -2.95970887e-01
1.22579467e+00 6.93462610e-01 1.17889665e-01 1.90707579e-01
7.39574805e-02 1.79864213e-01 -6.48394078e-02 6.63990319e-01
7.79180467e-01 -9.92466323e-03 -1.24177001e-02 -8.17208886e-01
-3.57166380e-01 1.35626495e-01 -1.12269747e+00 1.28446126e+00
-5.66672921e-01 7.39876211e-01 -2.62825694e-02 -1.24357724e+00
3.36058944e-01 8.36313844e-01 1.16156602e+00 -8.84133339e-01
5.49384773e-01 1.61471888e-01 -4.12199587e-01 -5.62191129e-01
2.64000624e-01 4.94554713e-02 -2.31891692e-01 4.01528358e-01
-2.19399799e-02 2.00765088e-01 1.50343195e-01 -5.19620299e-01
1.26450431e+00 5.42312711e-02 9.71815825e-01 -6.35687649e-01
5.43367505e-01 -2.98858911e-01 3.29666078e-01 1.13505341e-01
-4.00727868e-01 8.48828107e-02 -3.82643104e-01 -8.11731994e-01
-9.64565277e-01 -9.57580507e-01 -3.95697355e-01 1.03735459e+00
-5.33088028e-01 -1.74231142e-01 -1.05047107e+00 -1.01624146e-01
6.05362514e-03 1.23899996e+00 -4.37957168e-01 -1.19003624e-01
-7.23883986e-01 -1.54781222e+00 2.34666064e-01 6.49085104e-01
1.17097867e+00 -1.17110515e+00 -1.30527115e+00 5.17626584e-01
-2.09599763e-01 -8.53963614e-01 -2.55760979e-02 8.17177176e-01
-5.13622999e-01 -9.71201658e-01 -8.97754431e-02 -5.87567568e-01
1.12455651e-01 -4.99621183e-01 1.64190364e+00 -4.23305303e-01
-4.81585383e-01 2.17747942e-01 -2.03309432e-02 -4.75433618e-01
-4.03193653e-01 3.64881277e-01 1.77104741e-01 -2.14634299e-01
1.26541901e+00 -1.41801870e+00 -8.14027190e-01 8.09956640e-02
-5.09946942e-01 -3.02295685e-02 3.50308120e-01 1.98350757e-01
3.87345344e-01 9.28348541e-01 6.64302468e-01 -4.19969022e-01
4.37954575e-01 -7.23393619e-01 -8.52291226e-01 -4.61508930e-02
-1.10484433e+00 -9.66582447e-02 7.45259643e-01 1.10020980e-01
-8.26831698e-01 1.32271320e-01 -3.97223115e-01 3.13614994e-01
-8.19608986e-01 -3.88199091e-01 -7.54252791e-01 5.83425701e-01
2.77683586e-01 1.45548031e-01 -8.45888972e-01 -7.02986240e-01
3.32380742e-01 9.35836732e-01 6.73482358e-01 -1.28572673e-01
7.61044145e-01 4.05535877e-01 4.76145409e-02 -8.43015909e-01
-1.46613717e-01 -2.89081603e-01 -7.94598043e-01 -7.73049332e-03
1.35218859e+00 -8.78354788e-01 -1.21256328e+00 9.34557915e-01
-8.94615829e-01 -4.38415408e-01 -8.45285892e-01 1.74043540e-04
-5.37755132e-01 -9.05765500e-03 -4.31485385e-01 -9.47083950e-01
-1.06292629e+00 -8.80117655e-01 1.08251941e+00 4.37629014e-01
-6.41153276e-01 -9.82354999e-01 5.21819830e-01 1.70733541e-01
7.02387333e-01 1.08283651e+00 8.82059693e-01 -5.02601147e-01
-4.31241781e-01 -2.99991220e-01 1.07588798e-01 5.64208210e-01
8.02134871e-01 -2.02039644e-01 -1.35800397e+00 -2.99079090e-01
3.31965417e-01 1.01275273e-01 3.29068631e-01 8.25905681e-01
9.21300352e-01 -6.58880830e-01 -4.80537683e-01 6.03073359e-01
1.84193718e+00 4.76347953e-01 7.22153008e-01 1.79344833e-01
6.58809721e-01 -9.83236134e-02 -4.32129234e-01 5.98114908e-01
5.57534456e-01 3.10992956e-01 3.14144194e-01 1.70691889e-02
1.52155355e-01 1.28349319e-01 3.06076199e-01 8.93173933e-01
-1.34756207e-01 -4.42499071e-01 -5.57880819e-01 4.42211688e-01
-1.82077610e+00 -1.34067643e+00 -1.23289794e-01 1.57338786e+00
7.12658107e-01 -1.78752974e-01 4.57830489e-01 7.82644331e-01
3.13155204e-01 2.04268470e-01 -6.75747752e-01 -5.19910693e-01
8.04737490e-03 2.33340651e-01 7.09743142e-01 3.61334383e-01
-1.10253668e+00 -1.01460762e-01 6.73802519e+00 6.05720937e-01
-5.63093066e-01 4.98646975e-01 6.15667880e-01 -3.08180332e-01
3.09301555e-01 -7.52047777e-01 -6.53401375e-01 1.06279218e+00
1.56307161e+00 -9.37199891e-02 7.05536783e-01 9.48772609e-01
5.58274150e-01 -4.09607440e-01 -1.58225644e+00 1.39421260e+00
-5.40296972e-01 -8.60750675e-01 -5.84125519e-01 3.73747349e-01
8.75772178e-01 4.40862775e-01 -5.06266594e-01 2.72855043e-01
8.14246356e-01 -8.87833178e-01 2.16106653e-01 6.56805694e-01
2.73815066e-01 -9.62453246e-01 9.66178536e-01 2.33569711e-01
-1.76248932e+00 -4.20253307e-01 4.02405001e-02 -2.78619051e-01
2.54738986e-01 9.28045511e-01 -2.45066896e-01 6.62965536e-01
1.29162002e+00 4.55658793e-01 -4.25122380e-01 2.33591840e-01
1.39072090e-01 5.00937879e-01 -1.02384257e+00 6.26902208e-02
-1.32020310e-01 -4.95457560e-01 5.96949868e-02 1.19460654e+00
3.12671125e-01 2.06721365e-01 5.27820271e-03 8.42394769e-01
1.75396770e-01 -3.37783664e-01 -4.50657934e-01 3.45867276e-01
4.18152779e-01 1.59444892e+00 -4.47259843e-01 -5.00602365e-01
-4.47502702e-01 9.56794858e-01 -3.30620021e-01 5.38145825e-02
-8.46963048e-01 -1.90332100e-01 1.06948006e+00 -3.20567004e-02
5.32405749e-02 7.24425390e-02 -4.20126170e-01 -7.19138801e-01
-2.49167904e-01 -3.98688495e-01 6.13783181e-01 -5.63708663e-01
-1.49499965e+00 1.70223236e-01 3.54853451e-01 -5.77184558e-01
-6.64997101e-01 -2.72145271e-01 -9.00778949e-01 6.83932662e-01
-1.07077110e+00 -1.07293499e+00 -6.89366817e-01 7.53535211e-01
6.08819723e-01 -1.05570398e-01 1.15303743e+00 8.05440068e-01
-9.15551007e-01 2.11167946e-01 7.60227561e-01 1.55024201e-01
-8.48276541e-02 -1.85475051e+00 4.69748318e-01 6.74605846e-01
-2.13911340e-01 -4.79108632e-01 8.71139109e-01 -3.00588965e-01
-1.20123577e+00 -1.26763475e+00 6.36124074e-01 -7.20940769e-01
3.04931581e-01 -5.09808898e-01 -3.02321196e-01 8.22121322e-01
1.12592697e+00 -4.32013869e-01 9.00130928e-01 -3.79434258e-01
4.23825800e-01 -7.81936347e-01 -1.65428841e+00 -8.36960971e-02
6.89635396e-01 -3.29747498e-01 -6.00822628e-01 4.24957037e-01
8.81163925e-02 2.43504956e-01 -1.26451051e+00 4.58921671e-01
4.29315418e-01 -1.33297348e+00 8.81732821e-01 3.25059369e-02
-3.39359194e-01 -3.49508345e-01 -6.35244310e-01 -1.20395291e+00
-7.47613549e-01 -6.62392616e-01 -1.00620663e+00 1.65738368e+00
-3.32929194e-01 -8.21284592e-01 4.02319670e-01 9.85741913e-01
1.36022478e-01 -4.11843091e-01 -1.23793793e+00 -5.47531247e-01
-9.02290568e-02 -2.93185294e-01 1.39090860e+00 6.67657197e-01
-5.88347688e-02 4.67558324e-01 -6.91191927e-02 3.50679457e-01
8.51889729e-01 1.55010268e-01 3.03694695e-01 -1.31669676e+00
1.79966673e-01 -5.75436294e-01 -5.08650899e-01 -1.55392841e-01
-1.06068857e-01 -6.22289300e-01 -1.48946866e-01 -1.77363920e+00
-5.01738600e-02 3.04389328e-01 -5.51491320e-01 5.84851444e-01
4.58346814e-01 2.72994399e-01 -1.80566311e-01 -2.18061954e-01
-2.24831998e-01 4.41516489e-01 3.78163196e-02 -4.66486454e-01
-2.54495561e-01 3.49807888e-02 -3.51653308e-01 6.94573224e-01
1.34791100e+00 -1.79081991e-01 -3.64138931e-01 -2.65273824e-02
2.34897047e-01 -6.48439169e-01 2.59998620e-01 -1.51171875e+00
-1.02793761e-01 2.00271904e-01 9.30491984e-01 -1.41361153e+00
-6.29847571e-02 -1.66687250e+00 1.20817363e+00 6.77291155e-01
4.32728589e-01 2.45182991e-01 2.43292022e-02 -4.31095324e-02
2.65054166e-01 3.96499336e-01 1.16758478e+00 -1.02469400e-01
-6.18538976e-01 7.04778880e-02 -5.78540146e-01 -4.35558140e-01
1.34877193e+00 -2.17362404e-01 -2.03843489e-01 1.59289464e-02
-6.46234095e-01 4.14617658e-01 5.12211211e-02 4.08789486e-01
-2.74553806e-01 -1.95100045e+00 -5.92675090e-01 3.18625391e-01
-5.33597171e-01 -6.76881671e-02 3.12787518e-02 6.91305041e-01
-3.56525809e-01 6.58898175e-01 -2.16759607e-01 -6.05981171e-01
-1.06839716e+00 9.14179325e-01 9.23479438e-01 -4.22061682e-01
-6.51857793e-01 -1.17389776e-01 -2.00704843e-01 -3.89606729e-02
9.82433483e-02 -1.14946687e+00 -4.53557782e-02 3.87013346e-01
6.54655516e-01 1.34950161e+00 5.34139216e-01 -6.27349436e-01
-5.52897453e-01 5.77850759e-01 6.79520488e-01 6.81502640e-01
1.69111300e+00 -6.50457203e-01 -3.84227812e-01 7.78489292e-01
1.61254156e+00 -6.54960990e-01 -6.91969097e-01 5.53345919e-01
1.08002491e-01 3.58859509e-01 3.12049419e-01 -1.19773233e+00
-1.37592220e+00 3.00565511e-01 1.51924455e+00 1.06804037e+00
1.83042133e+00 -1.28722876e-01 1.14992762e+00 3.54905725e-01
4.44103628e-01 -1.80086231e+00 -9.30285811e-01 -1.34559870e-01
5.26062489e-01 -1.18753183e+00 9.15289670e-02 4.95342642e-01
5.11214793e-01 7.85247028e-01 1.89569518e-01 -5.20874932e-02
1.33275175e+00 7.81686068e-01 -3.04648161e-01 -8.42272043e-02
-3.19396257e-01 1.82230715e-02 -2.04568356e-01 7.99209654e-01
2.46929333e-01 5.57852328e-01 1.01978056e-01 5.93601465e-01
-6.35206521e-01 3.22503805e-01 -2.62437880e-01 8.40409815e-01
-4.17306304e-01 -7.59453416e-01 -6.02199495e-01 6.55791700e-01
-3.50905716e-01 2.43853375e-01 6.22276999e-02 6.15811884e-01
6.51978433e-01 1.27932072e+00 1.64719641e-01 -2.38004044e-01
5.05429685e-01 4.70885575e-01 2.87289172e-01 2.46097028e-01
-5.75233042e-01 -2.96527296e-01 -1.39673308e-01 -7.42198646e-01
-5.53606927e-01 -5.49683630e-01 -1.11162627e+00 -1.04440105e+00
-8.24538991e-02 -1.50379568e-01 5.43363631e-01 8.94253373e-01
2.10901007e-01 8.35149765e-01 8.36281657e-01 -1.25705564e+00
-2.12013930e-01 -1.30899870e+00 -9.25472319e-01 3.20190698e-01
6.42510295e-01 -1.73120111e-01 -2.72297710e-01 1.95141792e-01] | [16.048173904418945, 7.572981357574463] |
9e0f869b-0f54-424b-adad-37515d922623 | sfharmony-source-free-domain-adaptation-for | 2303.15965 | null | https://arxiv.org/abs/2303.15965v1 | https://arxiv.org/pdf/2303.15965v1.pdf | SFHarmony: Source Free Domain Adaptation for Distributed Neuroimaging Analysis | To represent the biological variability of clinical neuroimaging populations, it is vital to be able to combine data across scanners and studies. However, different MRI scanners produce images with different characteristics, resulting in a domain shift known as the `harmonisation problem'. Additionally, neuroimaging data is inherently personal in nature, leading to data privacy concerns when sharing the data. To overcome these barriers, we propose an Unsupervised Source-Free Domain Adaptation (SFDA) method, SFHarmony. Through modelling the imaging features as a Gaussian Mixture Model and minimising an adapted Bhattacharyya distance between the source and target features, we can create a model that performs well for the target data whilst having a shared feature representation across the data domains, without needing access to the source data for adaptation or target labels. We demonstrate the performance of our method on simulated and real domain shifts, showing that the approach is applicable to classification, segmentation and regression tasks, requiring no changes to the algorithm. Our method outperforms existing SFDA approaches across a range of realistic data scenarios, demonstrating the potential utility of our approach for MRI harmonisation and general SFDA problems. Our code is available at \url{https://github.com/nkdinsdale/SFHarmony}. | ['Ana IL Namburete', 'Mark Jenkinson', 'Nicola K Dinsdale'] | 2023-03-28 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 3.33551228e-01 1.95195328e-03 4.68196580e-04 -7.40443587e-01
-9.27430570e-01 -5.95027089e-01 3.99149179e-01 2.08660156e-01
-7.13993311e-01 6.63296461e-01 1.42556459e-01 -8.19281489e-02
-3.24998140e-01 -2.40210235e-01 -5.24825752e-01 -8.07530701e-01
-2.13736609e-01 5.97908735e-01 3.29704881e-01 3.08464378e-01
6.31300583e-02 6.48977816e-01 -1.02749097e+00 1.06725849e-01
8.80068660e-01 8.71784508e-01 2.81463742e-01 4.44986731e-01
2.64015943e-01 1.33877680e-01 -5.73834002e-01 -2.52111644e-01
5.29140234e-01 -3.75145525e-01 -9.08078074e-01 7.27698728e-02
2.01716080e-01 -1.47287622e-01 -5.04755974e-02 1.07787240e+00
9.02146935e-01 2.22066879e-01 8.57284307e-01 -1.28289950e+00
-4.17240411e-01 1.80006996e-01 -5.83235383e-01 4.19054836e-01
-2.46919738e-03 1.44782633e-01 3.87736589e-01 -3.91416132e-01
8.90502572e-01 7.27165401e-01 7.63302982e-01 6.19316161e-01
-1.75061846e+00 -7.50056744e-01 -2.37627909e-01 2.20397618e-02
-1.38218248e+00 -5.71508706e-01 4.86193717e-01 -7.76836991e-01
5.36514699e-01 2.79864848e-01 4.74597782e-01 9.95435357e-01
3.88153076e-01 4.52193975e-01 1.56899726e+00 -2.48804495e-01
4.50317889e-01 2.62201697e-01 -1.18039390e-02 1.10541552e-01
6.56793416e-02 -9.28302407e-02 -2.22122788e-01 -6.20959282e-01
8.30631077e-01 -1.40206501e-01 -4.19893414e-01 -9.06542122e-01
-1.24245346e+00 7.96306252e-01 2.18417123e-01 2.99506515e-01
-4.94401693e-01 -2.96915084e-01 5.33228397e-01 2.09930331e-01
3.31022739e-01 2.29287818e-01 -2.85538882e-01 1.75254375e-01
-1.01210272e+00 1.81020305e-01 5.30254364e-01 8.70371997e-01
2.53601670e-01 -3.26392144e-01 8.77569392e-02 1.08155572e+00
2.40800023e-01 3.63486618e-01 8.52951348e-01 -1.20102239e+00
2.16054544e-03 1.19208731e-03 6.86478335e-03 -7.26119101e-01
-7.39655435e-01 -3.55885267e-01 -7.80871511e-01 2.03827247e-01
5.94846070e-01 -1.10733867e-01 -9.93683636e-01 1.95924664e+00
6.12113118e-01 8.24431255e-02 -9.76258814e-02 9.38073814e-01
5.43023825e-01 2.33916417e-02 3.44593436e-01 -3.17260444e-01
1.44156849e+00 -3.00134778e-01 -5.97979486e-01 -1.54841959e-01
4.42073673e-01 -6.65775776e-01 7.25043058e-01 3.22779685e-01
-1.05313158e+00 -1.47033244e-01 -8.22899222e-01 2.44198754e-01
-1.96049348e-01 -4.78950709e-01 4.36164200e-01 7.69797981e-01
-1.04129708e+00 4.10563856e-01 -1.10704613e+00 -4.65944380e-01
8.58087540e-01 6.36372864e-01 -7.04579890e-01 -1.70925066e-01
-1.01913917e+00 1.10594940e+00 4.25384581e-01 -1.31949559e-01
-4.71692741e-01 -1.07510412e+00 -6.61871910e-01 -4.02051985e-01
1.05304368e-01 -6.60532176e-01 1.11443186e+00 -8.84915948e-01
-1.11064601e+00 1.03917122e+00 -3.96732055e-03 -4.34001297e-01
7.39998996e-01 3.97906125e-01 -5.43330014e-01 2.14627653e-01
3.33260566e-01 6.97752237e-01 7.48959422e-01 -1.10448480e+00
-2.77857333e-01 -4.20336127e-01 -6.37998343e-01 3.00116897e-01
9.96594727e-02 2.68241912e-01 -1.85566798e-01 -6.46458149e-01
7.45461285e-02 -9.74876642e-01 -4.54343706e-01 1.10309988e-01
1.02451099e-02 3.24508458e-01 4.17648941e-01 -7.61288404e-01
8.54123950e-01 -2.37270522e+00 -2.60337126e-02 5.35458803e-01
4.46171433e-01 2.48476669e-01 2.96280012e-02 8.38524178e-02
-3.40741217e-01 -7.97712356e-02 -7.64557064e-01 -6.19471781e-02
-1.58377096e-01 2.68711090e-01 3.14312875e-01 9.09363627e-01
4.04755622e-02 7.17373371e-01 -9.36074674e-01 -5.21023989e-01
1.14409365e-01 4.08199906e-01 -3.38501632e-01 -2.03094408e-02
3.63545209e-01 1.08622611e+00 -2.73877501e-01 2.56556779e-01
8.86051238e-01 -1.89068571e-01 3.12213361e-01 -7.54895201e-03
2.57639974e-01 -3.55545953e-02 -1.25290167e+00 1.71756876e+00
-1.95774123e-01 2.94198900e-01 3.22709292e-01 -1.19306087e+00
7.71047175e-01 3.98927480e-01 9.38791811e-01 -6.50674880e-01
2.52626777e-01 6.07190430e-01 3.95420015e-01 -3.52436900e-01
-2.65825868e-01 -5.72379291e-01 7.06412422e-04 6.72441423e-01
2.01414675e-01 -3.07108581e-01 -2.05437288e-01 1.26638710e-01
1.02866220e+00 -2.06119701e-01 4.77327645e-01 -4.36670601e-01
2.67127723e-01 -1.64555432e-03 6.50598705e-01 7.35560417e-01
-7.24066377e-01 7.80498683e-01 3.16647381e-01 -1.45211250e-01
-1.17214739e+00 -1.31519318e+00 -8.77073288e-01 5.36390007e-01
-1.90908909e-01 1.68969810e-01 -8.55180800e-01 -6.27724171e-01
2.94337478e-02 5.40331304e-01 -6.44610226e-01 -2.85084665e-01
-2.91045666e-01 -1.05323946e+00 5.12105048e-01 2.49041662e-01
2.29023904e-01 -7.87142754e-01 -1.01632023e+00 3.10513824e-01
-1.31588519e-01 -9.59141970e-01 -6.57529116e-01 2.97721863e-01
-9.05924797e-01 -9.64587688e-01 -1.02967024e+00 -5.36129892e-01
5.67678154e-01 -1.14286266e-01 7.73672163e-01 -4.22049314e-01
-4.99863684e-01 6.19835436e-01 -1.71235800e-01 -3.94608021e-01
-6.19318306e-01 -4.01077755e-02 1.01169325e-01 -2.27309000e-02
3.11908931e-01 -7.63830483e-01 -8.85287762e-01 2.51205087e-01
-1.12428641e+00 -1.01328515e-01 4.40796733e-01 8.98852587e-01
7.56655335e-01 -1.79346457e-01 7.90955245e-01 -1.11230505e+00
7.59116113e-01 -7.72950053e-01 -4.27863836e-01 6.89587444e-02
-6.96004927e-01 -1.65204778e-01 3.22928339e-01 -6.77851379e-01
-8.36701989e-01 1.70982033e-01 2.62771547e-01 -3.45340312e-01
-5.36367059e-01 3.85758400e-01 -6.23267367e-02 -2.50590414e-01
9.21637893e-01 3.30580398e-02 6.49553478e-01 -4.73705143e-01
2.87139565e-01 9.12413895e-01 6.14164889e-01 -3.97147179e-01
2.93483585e-01 6.70071840e-01 5.62692732e-02 -6.58976793e-01
-2.40897343e-01 -5.53176641e-01 -1.01469314e+00 -1.62301943e-01
7.88444459e-01 -6.63290620e-01 -1.14945196e-01 4.99794394e-01
-7.40892410e-01 -2.74258435e-01 -4.53374296e-01 8.48104358e-01
-7.24314630e-01 5.08089185e-01 -1.59649909e-01 -4.30772096e-01
-3.28997493e-01 -1.44983625e+00 6.43483460e-01 -3.05238552e-02
-6.21249557e-01 -1.12578940e+00 1.33210614e-01 2.78891116e-01
4.39029396e-01 5.59575796e-01 9.89913702e-01 -1.10713434e+00
-3.90957147e-02 -2.59315521e-01 -7.83217922e-02 4.23886359e-01
3.78179759e-01 -5.68276465e-01 -8.09435844e-01 -3.76463622e-01
4.92504388e-01 7.69868447e-03 2.59439886e-01 7.32878447e-01
9.50982094e-01 3.78179438e-02 -1.80405140e-01 5.30539513e-01
1.08526587e+00 3.64285171e-01 4.74793106e-01 3.80939901e-01
2.94028193e-01 7.80478776e-01 2.93382704e-01 3.21480334e-01
3.99560750e-01 7.56350875e-01 -3.68127488e-02 -2.33547941e-01
-1.00733079e-01 3.70661527e-01 -2.04897657e-01 6.13377333e-01
2.13034019e-01 2.85390556e-01 -1.34072769e+00 7.72643030e-01
-1.76351976e+00 -5.63340962e-01 -2.60448813e-01 2.50209188e+00
9.18534279e-01 -2.41535902e-01 3.48181814e-01 -2.72901595e-01
8.77150655e-01 -3.30211639e-01 -8.33555460e-01 -4.39284146e-01
8.07693377e-02 4.56536740e-01 8.47364247e-01 2.87088186e-01
-1.07732046e+00 5.03111720e-01 6.71110868e+00 6.31289065e-01
-1.09208691e+00 5.95894277e-01 5.38763642e-01 -3.26799989e-01
-3.31496634e-02 -1.86265171e-01 7.72241801e-02 5.13086438e-01
1.11893249e+00 -4.71181065e-01 5.41738868e-01 3.85799617e-01
4.39803153e-01 -3.72819096e-01 -9.23882067e-01 9.17703450e-01
4.58670594e-02 -9.26314592e-01 -4.88476455e-01 7.56834671e-02
5.26916802e-01 3.54495287e-01 4.11660299e-02 -4.07822549e-01
2.48557761e-01 -9.22976136e-01 4.86303240e-01 5.06479025e-01
9.37659502e-01 -5.10588825e-01 5.85196137e-01 1.32647127e-01
-6.11293912e-01 2.63463408e-01 -1.49269640e-01 4.99960512e-01
1.80372983e-01 4.77060467e-01 -9.16251838e-01 6.09217286e-01
7.59822607e-01 3.95509422e-01 -4.67168957e-01 1.39616370e+00
3.13153148e-01 2.79893637e-01 -4.24349487e-01 6.62376642e-01
-3.46563533e-02 -2.01603159e-01 5.44345260e-01 1.07719350e+00
1.72016472e-01 1.24013178e-01 1.66280828e-02 7.51041651e-01
1.66385040e-01 2.59796679e-01 -4.52654898e-01 1.44253150e-01
5.38608432e-01 1.08130741e+00 -9.84833717e-01 -1.15356326e-01
-3.63001466e-01 9.77246583e-01 7.85370693e-02 2.37561420e-01
-7.01277852e-01 -1.76269829e-01 4.84478623e-01 1.90883517e-01
1.45418227e-01 -7.96654522e-02 -4.83577371e-01 -9.07498121e-01
2.76384540e-02 -1.07207084e+00 6.97807431e-01 -6.26366317e-01
-1.66927218e+00 5.22664726e-01 4.79052424e-01 -1.18974555e+00
-1.69636592e-01 -4.00162131e-01 -1.85703084e-01 1.24054122e+00
-1.12727058e+00 -8.88016164e-01 4.52478044e-02 7.59003460e-01
-2.98234522e-02 -1.74994767e-02 8.01457584e-01 4.33992505e-01
-1.64106801e-01 5.19632876e-01 5.24792433e-01 -8.80138353e-02
1.05614972e+00 -1.09031332e+00 2.57767551e-02 4.88122135e-01
-2.06413984e-01 7.22195446e-01 6.58699512e-01 -6.90247774e-01
-8.72042000e-01 -8.52445245e-01 3.87789369e-01 -3.73757362e-01
7.75580347e-01 -3.12954485e-01 -1.24053466e+00 7.57421851e-01
-1.59163792e-02 1.85265124e-01 1.25949323e+00 -2.34740838e-01
-1.32720292e-01 8.40135589e-02 -1.74154603e+00 3.29998195e-01
9.28031147e-01 -3.93753946e-01 -6.83504045e-01 4.31640238e-01
5.65444827e-02 -4.97601449e-01 -1.51085472e+00 1.49823636e-01
6.37143612e-01 -7.03569472e-01 7.66347110e-01 -5.58506191e-01
-3.94326448e-02 -1.68863669e-01 4.12687771e-02 -1.45683765e+00
-2.95162886e-01 -3.20561975e-01 2.50665218e-01 9.73364294e-01
3.81460637e-01 -1.14070904e+00 4.15700704e-01 1.30892360e+00
9.33448970e-02 -4.76494908e-01 -1.56011117e+00 -9.28849757e-01
5.52043617e-01 -4.93597865e-01 7.37635732e-01 1.19442689e+00
7.42328167e-02 -1.11639403e-01 -8.86989087e-02 2.74068147e-01
8.35708439e-01 -1.58702374e-01 3.72030407e-01 -1.19300401e+00
-2.24740803e-01 -5.04637241e-01 -7.36081898e-01 -2.21677467e-01
7.94322118e-02 -1.26766562e+00 -7.69068971e-02 -1.23324645e+00
2.69502044e-01 -8.15387666e-01 -3.82716507e-01 4.82207626e-01
2.24039052e-02 3.42070162e-01 1.97040170e-01 4.50940609e-01
-2.15612069e-01 1.66235268e-01 1.19610643e+00 2.31565639e-01
-1.09914027e-01 -9.93362889e-02 -7.89632320e-01 5.01073122e-01
8.99988413e-01 -8.35002303e-01 -5.14281869e-01 -3.03171188e-01
-5.56735694e-01 -5.49961105e-02 4.41873938e-01 -8.70943069e-01
1.54218897e-01 -8.00240263e-02 4.50340748e-01 8.90548825e-02
6.12378046e-02 -1.01306725e+00 8.29559803e-01 4.57536221e-01
-1.83041885e-01 -1.76936015e-01 2.90843040e-01 3.84867042e-01
1.93469897e-01 -3.82061064e-01 1.20094860e+00 -1.67279348e-01
-2.80804396e-01 3.48843426e-01 -4.37128365e-01 2.33630076e-01
1.19749343e+00 -3.18091184e-01 -6.88677173e-05 -1.37317285e-01
-1.15496314e+00 2.77226478e-01 6.94955707e-01 1.97923824e-01
3.25382441e-01 -1.17301011e+00 -6.30682051e-01 3.57042074e-01
1.45082906e-01 1.00141071e-01 6.15563750e-01 1.36109209e+00
-3.28038007e-01 1.58619419e-01 -4.09034103e-01 -7.89162695e-01
-1.10343325e+00 4.79575187e-01 6.33749068e-01 -6.46446869e-02
-7.27524102e-01 4.80610937e-01 2.08619609e-01 -6.35331213e-01
-2.17380509e-01 3.29034738e-02 5.09872511e-02 1.25591815e-01
4.44425404e-01 2.44798973e-01 3.19261849e-01 -8.75194073e-01
-5.41624784e-01 1.82832092e-01 -3.56659085e-01 -3.63544792e-01
1.49645805e+00 -2.57514924e-01 -1.78019535e-02 3.84856194e-01
1.12730372e+00 -2.46057406e-01 -1.26563358e+00 -1.65201798e-01
6.92771226e-02 -5.63821495e-01 1.39684349e-01 -9.87668514e-01
-1.03220665e+00 6.00440860e-01 1.05210614e+00 -5.57847228e-03
1.17926443e+00 1.13162979e-01 4.32610601e-01 -4.14245725e-01
4.63541001e-01 -1.01327527e+00 -6.52559042e-01 -9.18361694e-02
7.99871922e-01 -1.15832341e+00 8.00942630e-02 -1.97667047e-01
-1.14167500e+00 6.60676241e-01 1.08119167e-01 -1.02537842e-02
7.72082746e-01 2.40020454e-01 3.26311141e-01 -2.36949652e-01
-2.68328369e-01 1.83851138e-01 2.27408767e-01 1.20408380e+00
1.40329361e-01 2.14981452e-01 -6.67885542e-01 6.52197361e-01
-4.71434705e-02 1.94055438e-01 5.03190458e-01 1.10890651e+00
1.54896125e-01 -1.47527862e+00 -5.57517111e-01 8.54739726e-01
-5.75120509e-01 2.79453881e-02 -1.69479221e-01 7.27877676e-01
1.38861164e-01 5.80105424e-01 9.16641392e-03 2.33949512e-01
3.87665659e-01 3.42206687e-01 5.38269281e-01 -6.78170383e-01
-4.48465735e-01 3.27587783e-01 -2.07386807e-01 -4.13209349e-01
-5.47097862e-01 -1.21747065e+00 -1.23925447e+00 -1.09754063e-01
-2.21523687e-01 8.26837644e-02 6.68461859e-01 7.68501461e-01
6.21086657e-01 3.10670793e-01 4.72619981e-01 -6.99348807e-01
-5.61183453e-01 -7.78128028e-01 -8.53063762e-01 6.60842121e-01
2.12920249e-01 -8.99814427e-01 -9.82357785e-02 1.38027653e-01] | [14.42003345489502, -1.9508531093597412] |
c3eef19f-b880-4251-b4eb-145727eca87c | bertmap-a-bert-based-ontology-alignment | 2112.02682 | null | https://arxiv.org/abs/2112.02682v4 | https://arxiv.org/pdf/2112.02682v4.pdf | BERTMap: A BERT-based Ontology Alignment System | Ontology alignment (a.k.a ontology matching (OM)) plays a critical role in knowledge integration. Owing to the success of machine learning in many domains, it has been applied in OM. However, the existing methods, which often adopt ad-hoc feature engineering or non-contextual word embeddings, have not yet outperformed rule-based systems especially in an unsupervised setting. In this paper, we propose a novel OM system named BERTMap which can support both unsupervised and semi-supervised settings. It first predicts mappings using a classifier based on fine-tuning the contextual embedding model BERT on text semantics corpora extracted from ontologies, and then refines the mappings through extension and repair by utilizing the ontology structure and logic. Our evaluation with three alignment tasks on biomedical ontologies demonstrates that BERTMap can often perform better than the leading OM systems LogMap and AML. | ['Ian Horrocks', 'Denvar Antonyrajah', 'Jiaoyan Chen', 'Yuan He'] | 2021-12-05 | null | null | null | null | ['ontology-matching'] | ['knowledge-base'] | [ 4.17972535e-01 6.57933533e-01 -3.93953979e-01 -3.61490846e-01
-2.85415202e-01 -1.06265821e-01 6.45778894e-01 8.95140886e-01
-3.93158793e-01 5.02963841e-01 4.51487154e-01 -2.24930167e-01
-8.32678795e-01 -8.71234536e-01 -2.65227288e-01 -1.87622622e-01
-8.06687996e-02 1.02762544e+00 2.98760623e-01 -5.00997841e-01
-2.79174596e-02 1.20742977e-01 -1.38959503e+00 2.72308379e-01
1.01949155e+00 7.60730684e-01 -1.27195179e-01 2.36547455e-01
-5.73240578e-01 5.70651770e-01 -4.06317972e-02 -6.58846140e-01
-2.61892714e-02 -1.86664894e-01 -1.12070739e+00 -4.47851181e-01
5.75578436e-02 3.44990522e-01 -4.38855290e-01 1.15540743e+00
4.31130350e-01 -1.00240909e-01 4.93565321e-01 -1.29151893e+00
-8.52082014e-01 9.28591013e-01 2.82279223e-01 9.64273065e-02
6.22187018e-01 -4.44850653e-01 1.36734092e+00 -7.43814707e-01
8.22238028e-01 1.21584308e+00 9.18368518e-01 5.52223980e-01
-1.40948093e+00 -4.43000317e-01 -2.38230944e-01 6.83930695e-01
-1.38198042e+00 -3.79798979e-01 4.21231508e-01 -4.71592754e-01
1.46973741e+00 2.09912762e-01 6.51736140e-01 9.52252567e-01
4.37785625e-01 2.81746447e-01 8.60791147e-01 -1.05462468e+00
2.99985081e-01 3.39013189e-01 2.51243502e-01 6.75904989e-01
4.51774061e-01 -8.65591094e-02 -6.47113800e-01 -5.05321920e-01
-2.97513121e-04 -1.19471394e-01 -6.22487999e-02 -3.23832303e-01
-1.17983711e+00 7.24050105e-01 3.63354117e-01 6.43880367e-01
-2.47413322e-01 -1.65356353e-01 2.98821479e-01 4.19169933e-01
3.57616186e-01 9.05703545e-01 -4.61531430e-01 -2.23021265e-02
-6.07170165e-01 2.05777973e-01 8.83031428e-01 1.16308510e+00
7.35730708e-01 -6.71997309e-01 2.20245272e-01 9.47792053e-01
4.58403170e-01 -4.87676896e-02 9.70654130e-01 -5.54667771e-01
2.27571219e-01 1.21215081e+00 -2.14066789e-01 -9.77411151e-01
-6.08526826e-01 -1.28216594e-01 -3.27596456e-01 -2.45595798e-01
1.50535613e-01 4.32306468e-01 -6.80353582e-01 1.63883889e+00
4.27619666e-01 4.19859052e-01 3.67502034e-01 3.44093889e-01
6.99296355e-01 -8.84748250e-02 2.84486651e-01 1.47341356e-01
1.75353169e+00 -8.37887764e-01 -1.13951206e+00 -2.04409018e-01
1.00355995e+00 -5.59117258e-01 8.10983896e-01 3.07494611e-01
-5.70099294e-01 -3.32826972e-01 -1.27450919e+00 -1.05229422e-01
-1.00847816e+00 -5.44623911e-01 5.89405060e-01 5.89192510e-01
-8.02052438e-01 7.58445799e-01 -7.12128639e-01 -1.05038118e+00
5.80220580e-01 4.62626427e-01 -9.51994240e-01 -2.58076370e-01
-1.65667546e+00 1.41950274e+00 1.01066446e+00 -4.11819726e-01
-6.06916659e-03 -8.60148787e-01 -9.44334626e-01 8.70307237e-02
4.28824931e-01 -8.53642642e-01 8.06055129e-01 -5.99757552e-01
-1.43225944e+00 8.44968557e-01 -3.62499356e-02 -7.35075414e-01
3.57277766e-02 -4.01707590e-01 -1.10265744e+00 1.38811544e-01
1.24893755e-01 5.56344688e-01 1.68485373e-01 -6.37467027e-01
-6.97750747e-01 -3.62912059e-01 8.10177624e-02 -2.76186988e-02
-1.00899279e+00 2.09557954e-02 -1.18533753e-01 -3.91387284e-01
1.20364740e-01 -7.71010280e-01 -2.85277218e-01 -8.53531808e-02
-1.94482267e-01 -4.05692965e-01 3.75855863e-01 -5.61119676e-01
1.66097271e+00 -1.98496366e+00 3.00334424e-01 4.49428916e-01
2.23633394e-01 1.64144158e-01 1.02817737e-01 7.54716575e-01
-2.85236239e-01 5.19732945e-02 -2.90046483e-01 -6.92891539e-04
3.70643765e-01 5.51238954e-01 -2.16927439e-01 2.49080770e-02
2.88547993e-01 9.83110845e-01 -1.16251516e+00 -8.26229036e-01
1.02048628e-01 1.59774944e-01 -8.68188918e-01 7.33053759e-02
-3.36337477e-01 -1.33106530e-01 -2.85914987e-01 6.61671758e-01
9.20740813e-02 -4.73924071e-01 7.98355281e-01 -2.17147052e-01
2.13373721e-01 3.40349495e-01 -9.40788209e-01 2.03057957e+00
-2.06916571e-01 3.62578005e-01 -8.71620238e-01 -1.19395149e+00
9.16338742e-01 6.27683878e-01 8.26293945e-01 -6.10384047e-01
-1.03312306e-01 5.26853621e-01 1.68425649e-01 -8.36533666e-01
3.41521025e-01 -2.36508593e-01 3.46607119e-02 1.10664532e-01
5.20055175e-01 3.88192683e-01 9.40720737e-02 -8.59911367e-03
1.43528867e+00 3.00982356e-01 1.13013518e+00 -4.10328269e-01
6.72162056e-01 3.93777102e-01 6.76894903e-01 2.69030839e-01
4.03297618e-02 6.68896660e-02 3.38352859e-01 -4.36715692e-01
-7.81258345e-01 -9.68672454e-01 -7.45253742e-01 8.61384392e-01
8.58113766e-02 -1.05972397e+00 -6.36256576e-01 -8.39767873e-01
4.09321189e-01 6.77111506e-01 -6.77365124e-01 -5.74280918e-01
-4.07510459e-01 -7.45758772e-01 8.56234550e-01 4.64252561e-01
8.17287341e-02 -9.59046304e-01 -3.27164501e-01 7.76630819e-01
2.33232919e-02 -1.43657839e+00 9.51879323e-02 2.37675950e-01
-8.51705134e-01 -1.26410782e+00 1.50516838e-01 -6.66481256e-01
6.24172807e-01 -4.02608752e-01 9.45836067e-01 6.07699119e-02
-3.53136986e-01 2.34605387e-01 -5.71695626e-01 -6.77400947e-01
-5.39083242e-01 4.17748153e-01 4.16982532e-01 -7.43615404e-02
1.34217036e+00 -7.14986861e-01 -2.41408125e-01 2.56521493e-01
-1.06371784e+00 -6.14313856e-02 6.02829397e-01 1.05410683e+00
4.45939660e-01 1.25233606e-01 9.75593925e-01 -1.22217441e+00
4.64010537e-01 -7.56021798e-01 -1.74269482e-01 5.51549196e-01
-1.35045719e+00 2.25427955e-01 3.33533257e-01 -2.11969286e-01
-6.62122488e-01 -2.02924594e-01 -1.77896723e-01 -2.87441816e-02
-1.67224169e-01 9.69864428e-01 -3.62645596e-01 -2.47307681e-02
8.12413216e-01 -1.64400950e-01 -1.72112091e-03 -5.30883014e-01
3.85162473e-01 1.03382432e+00 4.23413217e-01 -6.55477405e-01
8.25728476e-01 4.13469017e-01 -5.97097585e-03 -5.67353785e-01
-7.71523952e-01 -5.51014781e-01 -8.30933928e-01 2.13111728e-01
1.15597892e+00 -6.06271803e-01 -3.16122174e-01 -2.51254380e-01
-8.85878086e-01 1.57497630e-01 -3.33771914e-01 7.09750772e-01
-5.06732166e-01 3.29905182e-01 -2.05403924e-01 -3.28297496e-01
-2.04840437e-01 -7.74786472e-01 6.50821626e-01 -3.39694843e-02
-7.71819353e-01 -1.27142727e+00 2.60409743e-01 2.42198050e-01
3.31626117e-01 1.12542607e-01 1.57480502e+00 -1.41496789e+00
-9.70356986e-02 -3.62902433e-01 1.51846379e-01 4.01406661e-02
4.65005785e-01 -4.13641542e-01 -1.01345420e+00 6.72756881e-02
-5.04526377e-01 1.08068623e-01 3.29388738e-01 -3.76443475e-01
9.02828157e-01 -2.14645341e-01 -8.26988578e-01 4.98314708e-01
1.43028557e+00 1.76512957e-01 7.42214620e-01 1.06664312e+00
4.20405954e-01 7.63228118e-01 6.97631717e-01 4.78234999e-02
4.89836574e-01 1.00338554e+00 2.81365424e-01 2.93282002e-01
2.26888120e-01 -3.49779844e-01 8.50210860e-02 1.06242204e+00
1.11071172e-03 1.18502252e-01 -1.17792296e+00 4.94337976e-01
-2.17035890e+00 -7.26626039e-01 1.89964071e-01 1.92728794e+00
1.23158145e+00 2.61382788e-01 -2.05247283e-01 3.10694575e-01
3.18564475e-01 -4.17713791e-01 -1.69614613e-01 -3.29738408e-01
-1.40205786e-01 7.35736072e-01 2.92126060e-01 3.88221204e-01
-8.60096216e-01 9.25744534e-01 5.68945360e+00 4.59515065e-01
-4.85922068e-01 3.43773454e-01 -3.69369507e-01 4.03426051e-01
-4.16126072e-01 3.80061567e-01 -7.96999931e-01 4.27238673e-01
1.13018608e+00 -5.13573945e-01 2.19420061e-01 6.95797324e-01
-3.67910862e-01 4.31164712e-01 -1.57919991e+00 7.77939200e-01
1.22298151e-01 -1.52085471e+00 2.39989147e-01 2.51139432e-01
3.34505320e-01 -1.67667568e-01 -4.52112675e-01 3.60963374e-01
3.06553692e-01 -1.22930837e+00 3.97352219e-01 7.44693518e-01
5.19730330e-01 -2.11619362e-01 1.12444592e+00 5.08683659e-02
-1.02363479e+00 -2.63451725e-01 -4.22161460e-01 2.37009138e-01
7.50422180e-02 5.14285326e-01 -9.18565631e-01 1.27671015e+00
7.56040335e-01 9.81780231e-01 -4.99642342e-01 9.46625531e-01
-2.34332711e-01 2.31649980e-01 -1.37991443e-01 2.93003261e-01
-2.91363858e-02 -9.31261182e-02 5.40436566e-01 1.33689952e+00
3.66976976e-01 -1.09173372e-01 6.27750829e-02 4.95024085e-01
9.04119685e-02 5.33346057e-01 -7.07542360e-01 -4.19855088e-01
7.25869298e-01 1.04979968e+00 -2.38368452e-01 -4.53105539e-01
-5.69641054e-01 7.15550661e-01 3.52585644e-01 2.68858802e-02
-6.35598004e-01 -5.90776503e-01 1.04035556e+00 1.20655634e-01
3.48098751e-04 9.05079916e-02 -1.93658993e-02 -1.05735517e+00
2.49006022e-02 -9.45670009e-01 8.73989284e-01 -4.89194423e-01
-1.42059481e+00 6.40747011e-01 1.34367466e-01 -1.31498647e+00
-1.42565280e-01 -9.16928470e-01 -1.54884398e-01 5.26095033e-01
-1.70968413e+00 -1.07944334e+00 -2.26516441e-01 2.56734639e-01
1.39857695e-01 -6.02952003e-01 1.67611551e+00 6.95784569e-01
-3.90837818e-01 6.37064397e-01 -9.55421552e-02 4.49560396e-02
9.83512282e-01 -1.32557762e+00 1.02435790e-01 4.03590053e-01
2.53786623e-01 9.86953199e-01 6.15561903e-01 -5.96359789e-01
-1.29708421e+00 -1.05774057e+00 1.49761963e+00 -5.55975199e-01
9.78298724e-01 -2.02493310e-01 -1.16756690e+00 9.77739036e-01
2.45951146e-01 1.09502882e-01 1.39510643e+00 2.41353363e-01
-5.29425025e-01 -1.78351924e-01 -1.04947412e+00 6.37147307e-01
1.39388084e+00 -6.45285308e-01 -1.30101275e+00 3.26072216e-01
8.23567152e-01 -2.02290684e-01 -1.64258623e+00 6.60055935e-01
6.59528255e-01 -3.45895529e-01 9.84153926e-01 -1.46604168e+00
2.90268123e-01 -3.96560937e-01 -5.10122776e-01 -1.26739442e+00
-4.98642594e-01 -5.35536289e-01 -2.87928253e-01 1.19944429e+00
3.79419684e-01 -1.10326970e+00 3.44898522e-01 3.07956398e-01
-3.17369640e-01 -7.33535588e-01 -1.08076584e+00 -9.30613518e-01
-8.33428428e-02 -4.30136114e-01 1.23140180e+00 1.53166842e+00
8.12386453e-01 1.24643661e-01 1.49790958e-01 4.27206337e-01
3.89448643e-01 -1.78473428e-01 5.65143347e-01 -1.71877110e+00
-2.40027979e-01 -4.91427034e-01 -1.16664624e+00 -2.94263959e-01
3.88967723e-01 -1.63508856e+00 -3.66248012e-01 -1.54136300e+00
3.12598683e-02 -6.28598571e-01 -8.49060237e-01 7.44261265e-01
-2.29447916e-01 3.36206257e-02 -4.64647502e-01 1.76537141e-01
-4.18716073e-01 3.81825924e-01 4.14696306e-01 -1.43268526e-01
1.82185601e-02 -7.60597944e-01 -7.78150678e-01 7.34076977e-01
6.99794352e-01 -8.18888128e-01 -3.15839320e-01 -2.73050010e-01
6.18269384e-01 -5.17885685e-01 1.11974403e-01 -1.02575135e+00
4.32002842e-01 -5.75623177e-02 -1.05451038e-02 8.93527269e-02
2.39179768e-02 -1.13301444e+00 5.70379496e-01 4.33082402e-01
-4.35611993e-01 1.65595319e-02 4.82501760e-02 7.33501673e-01
-3.16653043e-01 -4.32931900e-01 4.55855250e-01 1.78251758e-01
-8.86118531e-01 1.21541090e-01 -6.39453307e-02 -5.98189496e-02
9.89751935e-01 -2.92758167e-01 -4.49315757e-01 4.52242911e-01
-1.03171360e+00 2.54079193e-01 2.98989713e-01 6.40789747e-01
4.40923214e-01 -1.75102663e+00 -4.83527422e-01 2.99329668e-01
9.91153955e-01 -4.14839536e-01 -2.47317553e-01 1.13293958e+00
-3.84623587e-01 5.62242329e-01 -2.97435522e-01 -4.51576263e-01
-1.13158453e+00 4.72461045e-01 3.53568494e-01 -3.52722704e-01
-9.84580636e-01 2.89375901e-01 -1.90303788e-01 -7.10825026e-01
-3.22910235e-03 -9.19659287e-02 -4.89013851e-01 4.28962335e-02
3.85206133e-01 2.38719746e-01 3.87105554e-01 -3.57984006e-01
-6.26513720e-01 4.53004658e-01 -1.04247637e-01 1.19591899e-01
1.42045236e+00 1.32983282e-01 -3.89078975e-01 4.19870585e-01
8.46051037e-01 -6.45777434e-02 -1.82300612e-01 -6.94589913e-01
1.09664237e+00 -4.75900859e-01 -2.31024891e-01 -6.10880017e-01
-4.01881188e-01 5.97394168e-01 4.52171713e-01 6.37950376e-02
7.14650512e-01 6.73550814e-02 6.89369023e-01 6.65246904e-01
3.87893409e-01 -1.22576261e+00 -7.66547024e-02 4.22038853e-01
5.27860463e-01 -9.82681274e-01 -1.79223474e-02 -5.56779861e-01
-2.29302868e-01 1.37872028e+00 4.78514105e-01 1.62431836e-01
8.76453638e-01 2.30598450e-01 1.51270241e-01 -4.22419161e-01
-7.33928919e-01 -2.77238786e-01 4.89936113e-01 5.29395461e-01
4.52705383e-01 -3.01199798e-02 -4.08466250e-01 9.33508933e-01
-3.07301670e-01 3.05012614e-01 6.67755082e-02 1.00263095e+00
-4.83921021e-01 -1.73356664e+00 -2.33488865e-02 5.40818334e-01
-4.42951918e-01 -1.16824746e-01 -2.32018873e-01 8.87560368e-01
4.92207915e-01 6.34063780e-01 -1.76799044e-01 -6.54842377e-01
5.49820364e-01 7.68066168e-01 4.82726425e-01 -9.27050591e-01
-4.50210392e-01 -4.50022489e-01 4.14191842e-01 -6.52843058e-01
-5.14567733e-01 -5.40347219e-01 -1.42144656e+00 3.77294756e-02
-4.69773799e-01 3.73159289e-01 3.81957799e-01 1.27583814e+00
5.13354301e-01 7.01807022e-01 5.19187795e-03 6.48755133e-02
-5.50256789e-01 -1.04742324e+00 -3.78012359e-01 6.40198886e-01
-3.44132990e-01 -1.04037559e+00 -4.73959669e-02 7.54187480e-02] | [9.098376274108887, 8.091619491577148] |
6f8279bf-d214-4af5-9042-81d671d978cd | constructing-a-knowledge-graph-from-textual | 2305.00382 | null | https://arxiv.org/abs/2305.00382v2 | https://arxiv.org/pdf/2305.00382v2.pdf | Constructing a Knowledge Graph from Textual Descriptions of Software Vulnerabilities in the National Vulnerability Database | Knowledge graphs have shown promise for several cybersecurity tasks, such as vulnerability assessment and threat analysis. In this work, we present a new method for constructing a vulnerability knowledge graph from information in the National Vulnerability Database (NVD). Our approach combines named entity recognition (NER), relation extraction (RE), and entity prediction using a combination of neural models, heuristic rules, and knowledge graph embeddings. We demonstrate how our method helps to fix missing entities in knowledge graphs used for cybersecurity and evaluate the performance. | ['Leon Moonen', 'Pierre Lison', 'Anders Mølmen Høst'] | 2023-04-30 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings', 'named-entity-recognition-ner'] | ['graphs', 'methodology', 'natural-language-processing'] | [-2.67368227e-01 3.66711348e-01 -3.84679675e-01 1.07979029e-01
-1.10355057e-01 -9.88462269e-01 4.24279660e-01 8.54491591e-01
-3.53209287e-01 4.99533355e-01 3.63481849e-01 -1.09248400e+00
-4.64062750e-01 -1.54266584e+00 -3.96531940e-01 1.25747561e-01
-5.00006557e-01 5.20499870e-02 5.79290092e-01 -4.18052465e-01
3.22222084e-01 1.01736486e+00 -6.93166614e-01 1.30377442e-01
7.26814032e-01 7.11158693e-01 -6.07407928e-01 3.79534394e-01
-1.13156825e-01 8.17359507e-01 -8.15279901e-01 -8.80563617e-01
2.31149569e-01 3.80485594e-01 -1.17413163e+00 -1.08060801e+00
-8.74964893e-02 -8.05847868e-02 -5.85182846e-01 1.15221739e+00
3.27135980e-01 2.21908927e-01 5.21205783e-01 -1.37395406e+00
-9.09847200e-01 5.94780684e-01 -8.01993832e-02 5.59502482e-01
5.54146469e-01 -2.45854035e-01 8.70015979e-01 -5.30018330e-01
1.03833365e+00 1.08488929e+00 1.00824666e+00 3.82660300e-01
-7.66849995e-01 -4.50514734e-01 8.21627304e-03 6.23225212e-01
-1.38817036e+00 1.58196129e-02 6.76070571e-01 -5.50371408e-01
1.74693704e+00 -2.43561808e-02 4.48447376e-01 8.50519359e-01
2.90933430e-01 2.31046714e-02 6.00607872e-01 -5.64387202e-01
4.32668269e-01 1.86623126e-01 7.10384667e-01 7.54381835e-01
8.40557337e-01 4.53107446e-01 -1.55095279e-01 -7.12486684e-01
4.73383129e-01 -9.51501653e-02 -1.52953088e-01 -2.12983191e-02
-5.04460037e-01 1.01443708e+00 7.47136474e-01 5.22687316e-01
-5.03507316e-01 -1.64875269e-01 5.83509564e-01 3.68107468e-01
3.54371339e-01 1.05844450e+00 -7.18797863e-01 2.34341204e-01
-2.78372526e-01 -2.14160025e-01 1.32917047e+00 4.81078416e-01
7.34968066e-01 2.24517494e-01 1.03870295e-01 8.09770599e-02
6.69273287e-02 3.22722882e-01 -3.98656651e-02 -2.62705594e-01
5.70777476e-01 1.10579300e+00 1.67200357e-01 -1.68772149e+00
-6.37320638e-01 1.52172253e-01 -3.03809822e-01 2.07734823e-01
3.52908224e-02 -4.05586541e-01 -9.89185214e-01 1.36043012e+00
4.50934350e-01 4.04078245e-01 5.15517592e-01 1.68003798e-01
8.09865832e-01 4.24456447e-01 4.68533933e-01 8.57032910e-02
1.08344138e+00 -2.84704238e-01 -5.92403114e-01 1.36553943e-01
7.95067132e-01 5.73149212e-02 1.81607589e-01 6.77161589e-02
-2.43170574e-01 5.74007854e-02 -9.68756616e-01 4.05329615e-01
-1.48254478e+00 -5.06664217e-01 7.77754009e-01 1.18854046e+00
-9.83531356e-01 5.45399070e-01 -7.24492371e-01 -4.44151610e-01
4.52183068e-01 2.54495382e-01 -7.96103954e-01 1.37895986e-01
-1.84015059e+00 1.66714048e+00 1.16571367e+00 -1.00602366e-01
-7.73833752e-01 -6.59816623e-01 -1.12171233e+00 3.41842234e-01
6.57933235e-01 -1.70265093e-01 2.62769341e-01 -5.85392192e-02
-9.17264163e-01 1.52270630e-01 3.93145084e-01 -7.60732889e-01
-5.02161443e-01 -3.02721560e-01 -1.12867641e+00 2.24786609e-01
-3.95147353e-01 -1.37739941e-01 3.69526207e-01 -1.07354248e+00
-2.47641370e-01 -4.70920324e-01 4.61080402e-01 -2.86734730e-01
-7.33581960e-01 4.84804779e-01 1.21390462e-01 -4.30272162e-01
-4.71213132e-01 -5.08414924e-01 -3.83692533e-01 -9.94741321e-01
-6.32853806e-01 -1.61583036e-01 8.50319922e-01 -1.41489947e+00
1.81991899e+00 -1.85488749e+00 2.32295487e-02 9.87575471e-01
3.80865842e-01 8.68571639e-01 -1.78955108e-01 7.87014484e-01
-2.60789543e-01 8.09488893e-01 -1.69513017e-01 6.71506941e-01
-1.16367631e-01 1.47763472e-02 -6.31134093e-01 -8.20865557e-02
3.82730186e-01 1.07396901e+00 -9.85611439e-01 -1.44310102e-01
4.68983166e-02 4.63412017e-01 -1.35194957e-01 -3.19032930e-02
-1.40091911e-01 -1.85012996e-01 -6.64477646e-01 8.29030097e-01
4.56198782e-01 2.18596652e-01 7.17260897e-01 -2.45328367e-01
-8.20497200e-02 1.82252079e-01 -1.02298927e+00 8.61061752e-01
-2.60024786e-01 3.10124516e-01 -4.94335353e-01 -6.51011944e-01
9.70472515e-01 4.28428590e-01 3.84115800e-02 -6.20070040e-01
4.45201471e-02 -7.05569014e-02 -3.84206235e-01 -4.87731278e-01
5.79650521e-01 3.26123089e-01 -3.40213269e-01 5.26071250e-01
3.21056873e-01 3.50630343e-01 4.24149409e-02 3.79174560e-01
1.57146895e+00 -3.07047576e-01 7.25247800e-01 2.14693129e-01
4.23568219e-01 3.09753835e-01 6.75156713e-01 5.58074057e-01
-1.32277384e-01 -4.71139699e-01 8.61806273e-01 -7.63548732e-01
-8.48407269e-01 -9.64470327e-01 1.78214163e-01 4.89814788e-01
-1.46480188e-01 -7.63376296e-01 -7.50023305e-01 -1.42792320e+00
6.53987005e-02 9.02389646e-01 -7.76969671e-01 -6.34689927e-01
-5.47444463e-01 -6.00554466e-01 1.03090680e+00 7.87547112e-01
9.31861773e-02 -1.20586228e+00 -2.79221237e-01 2.49924064e-01
2.20116928e-01 -1.17409945e+00 1.95910797e-01 1.04164325e-01
-5.40653765e-01 -1.47331035e+00 3.51769060e-01 -5.02673864e-01
5.75962424e-01 -5.00380360e-02 7.92559743e-01 3.07506714e-02
-3.02030981e-01 5.79815865e-01 -6.01484120e-01 -4.66539323e-01
-4.38860565e-01 1.76610261e-01 2.70923674e-01 -4.90549624e-01
4.49832171e-01 -5.28867900e-01 -7.30683887e-03 1.30541727e-01
-1.26838970e+00 -1.00527418e+00 3.26242566e-01 3.36370230e-01
3.23165953e-01 6.77293301e-01 7.39937484e-01 -8.48416388e-01
1.02651858e+00 -9.36009407e-01 -8.10832262e-01 9.86390293e-01
-7.62801290e-01 7.98083246e-02 5.04161716e-01 -2.62926996e-01
-1.03322530e+00 -6.59121349e-02 2.92039700e-02 -3.54413122e-01
-1.78447008e-01 1.18291664e+00 -3.92802179e-01 -7.14447260e-01
9.63642836e-01 -3.08839828e-01 -5.44638813e-01 -4.49728698e-01
5.99306166e-01 2.87654698e-01 4.23308522e-01 -2.98131078e-01
1.16479659e+00 -8.48394334e-02 1.46531492e-01 -5.19915462e-01
-6.27963066e-01 -6.37463629e-02 -8.47364306e-01 4.20441367e-02
8.92243803e-01 -4.25164342e-01 -6.99390471e-01 1.81812376e-01
-1.31740892e+00 2.72845104e-02 -1.10759646e-01 3.80064756e-01
3.60416770e-01 4.83993679e-01 -3.94632906e-01 -7.50507236e-01
-5.78955293e-01 -4.27966535e-01 6.91346228e-02 3.88249159e-01
4.70015332e-02 -1.33225441e+00 7.16567695e-01 -9.50379074e-02
4.16555285e-01 8.09249640e-01 1.27727640e+00 -1.32867444e+00
-2.34142303e-01 -6.94961429e-01 -3.14876139e-01 2.54948050e-01
2.73845583e-01 1.17417976e-01 -5.94327092e-01 1.91663709e-02
-6.52488589e-01 9.16250516e-03 5.99655032e-01 -3.03094864e-01
5.61909735e-01 -5.75482070e-01 -5.23185551e-01 2.92680889e-01
1.59642291e+00 5.27262866e-01 9.05076921e-01 6.55235767e-01
8.59783292e-01 7.42668271e-01 1.85836777e-01 1.46434590e-01
5.75058758e-01 4.26118106e-01 4.86221284e-01 3.63537490e-01
3.11386406e-01 -4.36990559e-01 1.90395042e-01 1.87925085e-01
-3.75035048e-01 -3.42545956e-01 -1.62729490e+00 5.95040202e-01
-1.62918866e+00 -1.21063805e+00 2.01796651e-01 2.14196372e+00
5.08648396e-01 1.30883038e-01 -7.00861122e-03 3.55166458e-02
7.90084124e-01 -7.75646046e-03 -2.69966483e-01 -6.76669061e-01
-1.09851405e-01 5.10166824e-01 7.77089536e-01 5.43633461e-01
-1.15873373e+00 1.42104352e+00 6.93811703e+00 4.57413137e-01
-6.00181401e-01 5.30630015e-02 4.46922844e-03 5.61603308e-01
-2.61346042e-01 2.95011938e-01 -7.95523942e-01 -7.02852681e-02
1.47171795e+00 -3.72123301e-01 4.02007490e-01 8.42269242e-01
-4.88886714e-01 2.67331064e-01 -4.08761293e-01 2.16297358e-01
3.36524509e-02 -1.58500922e+00 2.56231010e-01 -4.92897332e-02
4.19246703e-01 -2.31060535e-01 -3.10202837e-01 2.85383582e-01
9.02762949e-01 -9.00178671e-01 2.11028270e-02 7.62261033e-01
4.59103554e-01 -1.10075438e+00 9.74450171e-01 -2.30404079e-01
-1.30912483e+00 -4.81178790e-01 -1.82410955e-01 3.99674088e-01
1.59337819e-01 5.23179889e-01 -1.17965031e+00 1.33571351e+00
7.00010180e-01 4.16010737e-01 -7.96275020e-01 1.09468436e+00
-8.28794897e-01 6.13764286e-01 -1.72652483e-01 1.66812465e-01
1.49319309e-03 1.80081815e-01 6.80223882e-01 1.02983177e+00
-3.83217260e-02 4.02401179e-01 -6.44826367e-02 5.57121992e-01
-4.50074002e-02 -7.00536370e-02 -1.00862849e+00 -6.80003881e-01
8.33426118e-01 1.16764927e+00 -5.90967774e-01 -1.38029009e-01
-2.58445084e-01 3.60240877e-01 6.13540232e-01 4.22885090e-01
-5.05045772e-01 -1.06624317e+00 5.01582205e-01 -3.07699382e-01
4.34043705e-01 -5.14379144e-01 -1.49700671e-01 -1.21219373e+00
-1.98367655e-01 -5.26299000e-01 1.03516710e+00 -4.77098227e-01
-1.28579354e+00 1.10112524e+00 3.19121808e-01 -8.64901841e-01
-5.95263615e-02 -6.61973000e-01 -9.64235008e-01 7.20509529e-01
-1.33052325e+00 -1.07546210e+00 2.40293629e-02 6.29827976e-01
-6.95565701e-01 -3.63117397e-01 1.09902513e+00 2.10862309e-01
-1.04677796e+00 5.00002742e-01 -3.00099403e-01 5.70113480e-01
1.80335760e-01 -1.12457538e+00 1.09064770e+00 1.03284967e+00
1.60892725e-01 7.29655862e-01 1.66081011e-01 -1.46908689e+00
-1.28071880e+00 -1.15954709e+00 1.02649963e+00 -8.91713977e-01
1.33483195e+00 -1.17111392e-01 -1.21038127e+00 7.96828568e-01
-7.31501058e-02 -7.86749795e-02 9.93708432e-01 2.64987022e-01
-1.06868732e+00 2.81844199e-01 -1.39748394e+00 3.19386035e-01
8.20874512e-01 -6.91605031e-01 -9.54855919e-01 7.90219009e-02
1.17994833e+00 4.24344093e-02 -1.33492553e+00 3.82344961e-01
2.22383678e-01 -1.06617093e-01 1.24218035e+00 -1.47777462e+00
-1.45605251e-01 -2.48404562e-01 -7.43586048e-02 -1.33977294e+00
-4.44662094e-01 -4.48667645e-01 -6.78013504e-01 1.26892519e+00
6.83309376e-01 -1.06196582e+00 4.84749258e-01 9.65948701e-01
7.88379312e-02 -2.88026750e-01 -8.66794527e-01 -9.73523319e-01
-2.57064290e-02 -2.79687732e-01 1.01436710e+00 1.53980243e+00
5.06448865e-01 1.13626011e-02 -2.21812189e-01 1.01116264e+00
5.65571308e-01 -2.40789324e-01 1.13919400e-01 -1.53095329e+00
1.97083920e-01 -2.00552925e-01 -9.32360470e-01 5.03851831e-01
4.12236482e-01 -6.88664556e-01 -8.17525208e-01 -1.68382001e+00
-2.78943598e-01 -2.25086629e-01 -7.35368848e-01 1.27033830e+00
-1.43790096e-01 -3.37030292e-01 6.42037913e-02 -2.52996773e-01
-4.23645586e-01 1.44802555e-01 2.22681120e-01 -8.68554041e-02
-5.07535338e-01 -4.00663674e-01 -7.72558093e-01 6.14394844e-01
1.16973269e+00 -7.49004006e-01 -7.06989691e-02 -2.76319385e-01
6.77450955e-01 1.98563650e-01 4.96687621e-01 -7.44539499e-01
2.92957664e-01 -4.06454057e-01 2.37215370e-01 -3.41822594e-01
-1.88816935e-01 -6.88519239e-01 1.73716545e-01 5.55807829e-01
1.46678463e-01 3.01520437e-01 6.74836874e-01 7.29425907e-01
-3.02253544e-01 -2.37351701e-01 1.69131696e-01 1.88486025e-01
-1.14228749e+00 3.72763157e-01 -1.43093690e-01 -2.45247245e-01
1.14455771e+00 5.84217831e-02 -1.08137500e+00 9.75986794e-02
-1.02836716e+00 2.36883014e-01 2.07673877e-01 6.48525178e-01
9.61409092e-01 -1.28351808e+00 -3.54849964e-01 -1.32927999e-01
7.45665655e-02 -7.64021516e-01 2.17148617e-01 3.16208214e-01
-5.06986380e-01 4.88093913e-01 -6.36726439e-01 7.51813948e-01
-9.41792905e-01 1.02876270e+00 1.78771690e-01 -6.88953638e-01
-3.88861716e-01 4.27547127e-01 -6.51086748e-01 -7.23719895e-01
1.36235923e-01 2.62328446e-01 -1.03630161e+00 1.73177466e-01
7.92743683e-01 7.34256089e-01 1.22887656e-01 -4.59650755e-01
-8.58459771e-01 2.55448967e-01 -1.79607302e-01 1.70027122e-01
1.62823987e+00 4.89021987e-01 -2.27628276e-01 -3.44373286e-01
8.56088579e-01 -8.72775540e-03 -3.79241019e-01 -1.63564533e-01
7.10580349e-01 -3.12656254e-01 5.22487648e-02 -1.15947604e+00
-9.15749907e-01 5.65828562e-01 3.57572883e-01 6.24058604e-01
1.02151036e+00 -8.84013772e-02 8.31139684e-01 9.99154627e-01
5.86411238e-01 -8.89437675e-01 -2.62336850e-01 8.32507670e-01
5.52555919e-01 -8.37679744e-01 2.75700074e-02 -5.18473506e-01
-6.40308857e-01 1.23647666e+00 5.59274435e-01 -1.15869664e-01
1.01324141e+00 1.65401280e-01 -3.37566942e-01 -4.45715576e-01
-4.86162126e-01 -2.97068805e-01 4.89895046e-01 1.05510163e+00
-4.57007140e-01 1.52111262e-01 6.85406551e-02 8.48611057e-01
2.39308700e-01 -3.60145360e-01 5.63540459e-01 1.33049750e+00
-6.33147240e-01 -1.19561887e+00 -3.53953600e-01 2.87847042e-01
-3.77078086e-01 -5.57851970e-01 -1.05838156e+00 7.45083213e-01
-1.09377518e-01 9.79646862e-01 -5.19280553e-01 -1.05116570e+00
6.20518386e-01 1.40794441e-01 1.08361058e-01 -5.83313644e-01
-9.68287945e-01 -9.53946173e-01 5.35064161e-01 -5.84813118e-01
-1.65916756e-02 -1.69892147e-01 -1.17056119e+00 -2.73880988e-01
-4.82871681e-01 5.46507120e-01 7.19293833e-01 7.35726774e-01
7.68429101e-01 3.45230162e-01 6.29015207e-01 -1.33869052e-01
-8.73352736e-02 -5.64749360e-01 -3.74147445e-01 9.49480629e-04
-1.96840182e-01 -8.12138855e-01 -3.23927611e-01 -1.87098667e-01] | [6.569391250610352, 7.471474647521973] |
86f38fdb-f1d3-4efe-8084-4e98f686c3df | ffa-ir-towards-an-explainable-and-reliable | null | null | https://openreview.net/pdf?id=FgYTwJbjbf | https://openreview.net/pdf?id=FgYTwJbjbf | FFA-IR: Towards an Explainable and Reliable Medical Report Generation Benchmark | The automatic generation of long and coherent medical reports given medical images (e.g. Chest X-ray and Fundus Fluorescein Angiography (FFA)) has great potential to support clinical practice. Researchers have explored advanced methods from computer vision and natural language processing to incorporate medical domain knowledge for the generation of readable medical reports. However, existing medical report generation (MRG) benchmarks lack both explainable annotations and reliable evaluation tools, hindering the current research advances from two aspects: firstly, existing methods can only predict reports without accurate explanation, undermining the trustworthiness of the diagnostic methods; secondly, the comparison among the predicted reports from different MRG methods is unreliable using the evaluation metrics of natural-language generation (NLG). To address these issues, in this paper, we propose an explainable and reliable MRG benchmark based on FFA Images and Reports (FFA-IR). Specifically, FFA-IR is large, with 10,790 reports along with 1,048,584 FFA images from clinical practice; it includes explainable annotations, based on a schema of 46 categories of lesions; and it is bilingual, providing both English and Chinese reports for each case. Besides using the widely used NLG metrics, we propose a set of nine human evaluation criteria to evaluate the generated reports. We envision FFA-IR as a testbed for explainable and reliable medical report generation. We also hope that it can broadly accelerate medical imaging research and facilitate interaction between the fields of medical imaging, computer vision, and natural language processing. | ['Xiaojun Chang', 'Xiaodan Liang', 'Karin Verspoor', 'Flora D Salim', 'Yizhi Liu', 'Mengke Li', 'Caineng Pan', 'Zhong Liu', 'Xin Chen', 'Cong Wang', 'Xiaoyun Zhao', 'Yuetian Weng', 'Rui Liu', 'Wenjia Cai', 'Mingjie Li'] | 2021-08-19 | null | null | null | thirty-fifth-conference-on-neural-information | ['medical-report-generation'] | ['medical'] | [ 1.77813709e-01 5.18293262e-01 -4.28649008e-01 -4.68881756e-01
-1.25568151e+00 -3.73257399e-01 4.95670706e-01 2.55763412e-01
-5.13951480e-02 1.07043719e+00 6.98920965e-01 -4.66922015e-01
-3.47825401e-02 -5.12681663e-01 -3.19711834e-01 -2.44250327e-01
5.79121150e-02 3.36505592e-01 -3.60052407e-01 3.67403567e-01
1.62287518e-01 2.26040646e-01 -1.05568051e+00 7.50942588e-01
1.16761696e+00 9.27023411e-01 1.49784535e-01 8.01473856e-01
-1.27816871e-01 1.04564035e+00 -6.74410939e-01 -8.07493508e-01
-1.57947600e-01 -8.01862597e-01 -1.02639091e+00 2.86979467e-01
3.60587865e-01 -4.16946709e-01 6.84318040e-03 8.44301701e-01
4.80729073e-01 -3.69363546e-01 7.18265951e-01 -1.08830261e+00
-1.23938704e+00 6.48574948e-01 -4.63581085e-01 2.55345345e-01
4.32640195e-01 3.85523021e-01 7.42587090e-01 -8.19734454e-01
1.16477513e+00 8.44153047e-01 3.51503462e-01 8.40850532e-01
-8.64364088e-01 -5.36926866e-01 -1.26418397e-02 2.18322370e-02
-1.30664384e+00 -2.32293487e-01 5.24693206e-02 -6.17025256e-01
7.00596094e-01 5.95226884e-01 5.09749353e-01 1.08025014e+00
3.13911319e-01 4.24781322e-01 1.23934996e+00 -2.54036307e-01
1.73243470e-02 3.97427350e-01 8.73802975e-03 9.60020363e-01
6.01840019e-01 1.28464803e-01 -4.27117944e-01 -2.87425488e-01
7.86938608e-01 -1.65774345e-01 -6.67276084e-01 4.83164281e-01
-1.71827590e+00 8.84314597e-01 5.03844440e-01 3.68847489e-01
-4.58216846e-01 -2.69949704e-01 2.47371167e-01 7.72301257e-02
5.33658147e-01 7.50803828e-01 -1.05003297e-01 1.70144394e-01
-1.02193022e+00 2.29433209e-01 6.67579174e-01 1.15269673e+00
3.19090128e-01 -2.79389501e-01 -7.47660041e-01 4.40065354e-01
1.88963458e-01 5.50482273e-01 5.52263856e-01 -1.02736473e+00
4.98007774e-01 7.00923920e-01 9.53739434e-02 -1.11713755e+00
-4.23876703e-01 -3.17819983e-01 -1.07626331e+00 -1.75154164e-01
1.00745559e-01 -1.17346086e-01 -8.96008253e-01 1.26573062e+00
5.19496948e-02 3.87962013e-02 2.12346196e-01 1.25482106e+00
1.65326834e+00 4.26246077e-01 3.15717459e-01 -5.48005044e-01
1.62313867e+00 -9.89459515e-01 -8.72568786e-01 4.35269661e-02
7.22710788e-01 -9.65116799e-01 1.12230730e+00 2.92403489e-01
-1.17365885e+00 -4.58539307e-01 -6.75393820e-01 2.14936808e-02
1.51272640e-02 4.71383005e-01 7.23667681e-01 3.06863546e-01
-1.11918247e+00 -8.91188681e-02 -5.75889528e-01 -3.64233494e-01
5.73267937e-01 6.21013250e-03 -6.14575386e-01 -3.88837934e-01
-1.00330591e+00 9.86122489e-01 2.78292239e-01 -3.18957940e-02
-5.94035923e-01 -9.04988408e-01 -7.53542185e-01 -3.06056470e-01
2.45315284e-01 -1.35022628e+00 1.22192407e+00 -6.55181110e-01
-1.01151943e+00 1.29851604e+00 -3.71758401e-01 -5.02712429e-01
4.63632524e-01 1.91837072e-01 -6.69711828e-01 5.76933265e-01
4.54416037e-01 1.11044824e+00 5.34870267e-01 -1.11699152e+00
-5.63981831e-01 -9.70895141e-02 -1.73339799e-01 -7.68847764e-02
6.79660887e-02 1.37224630e-01 -3.86301070e-01 -9.77958918e-01
-1.01608023e-01 -8.71996820e-01 -5.67578137e-01 1.03844553e-01
-7.30438948e-01 -1.76744029e-01 1.52366832e-01 -7.81684577e-01
1.33775830e+00 -1.94097507e+00 -4.54808205e-01 -1.22638121e-01
8.22621107e-01 1.12734623e-01 -2.50038624e-01 -7.39101972e-03
-1.91304430e-01 7.11211264e-01 -1.56126603e-01 -2.05202579e-01
-4.33192283e-01 -8.09430033e-02 -3.36312205e-01 1.67209074e-01
6.64356291e-01 1.15534854e+00 -8.66613865e-01 -8.65445316e-01
-7.99297616e-02 3.59932601e-01 -5.95099747e-01 2.84205526e-01
-5.43642119e-02 8.51472616e-01 -7.12307990e-01 6.35316908e-01
4.21349436e-01 -7.13214159e-01 -5.66220954e-02 -3.56637269e-01
6.76383451e-02 3.22415113e-01 -5.46016812e-01 1.64877284e+00
-4.15984035e-01 5.14480829e-01 -4.10967350e-01 -4.25779909e-01
8.82986128e-01 7.14212954e-01 6.12649798e-01 -5.35212338e-01
-8.99460912e-02 2.70701379e-01 -1.57482579e-01 -8.52527678e-01
3.04006636e-01 -3.31040770e-01 1.05930306e-01 6.49909556e-01
-1.10875167e-01 -2.72183806e-01 3.00551832e-01 5.60653925e-01
1.31820488e+00 -3.35911781e-01 5.04474998e-01 2.31909841e-01
4.08553183e-01 4.98037040e-01 4.76443082e-01 8.44247162e-01
-3.34911495e-01 1.11136365e+00 3.34052503e-01 -7.11243808e-01
-8.27830315e-01 -1.17465138e+00 -3.20745498e-01 1.90701857e-01
-4.71166801e-04 -5.77972233e-01 -5.83818674e-01 -1.00542438e+00
-3.51296991e-01 7.75957644e-01 -6.12850666e-01 1.49584468e-02
-1.51560709e-01 -7.12737858e-01 5.21953940e-01 3.64201337e-01
3.72627735e-01 -1.26053107e+00 -4.69240695e-01 2.89583087e-01
-8.15705836e-01 -1.53240180e+00 -6.68734074e-01 -6.72727764e-01
-8.26090395e-01 -1.25384927e+00 -9.54121113e-01 -4.98984873e-01
1.00594354e+00 1.76582024e-01 1.62674713e+00 4.66234684e-01
-6.48654759e-01 4.86140221e-01 -6.43538296e-01 -6.00357413e-01
-8.57783139e-01 -2.39774764e-01 -1.83981448e-01 -2.99545974e-01
2.18703091e-01 1.53240487e-01 -7.45429933e-01 1.25404432e-01
-9.90597844e-01 5.69025934e-01 9.15051699e-01 7.71320164e-01
8.94750893e-01 -3.11985612e-01 6.96641922e-01 -1.04855990e+00
8.15515101e-01 -4.37784642e-01 -1.25630975e-01 5.43965638e-01
-8.70820582e-01 -6.44614100e-02 3.01156044e-01 -8.06403086e-02
-1.16086900e+00 -2.52812594e-01 1.08148418e-02 -7.15776235e-02
-3.98688346e-01 7.40560770e-01 5.06520033e-01 2.10052118e-01
9.34258401e-01 1.71526819e-01 -9.27963853e-03 1.40696233e-02
4.65502650e-01 7.34343886e-01 6.89923108e-01 -1.67796835e-01
5.37842989e-01 4.36781794e-01 -2.68389374e-01 -2.75302768e-01
-1.14050376e+00 -3.73767942e-01 -1.76854759e-01 -1.44647092e-01
1.02810121e+00 -9.74908412e-01 -4.18794513e-01 -3.46778840e-01
-1.60204089e+00 2.82159418e-01 -2.54621029e-01 9.02892172e-01
-5.31961381e-01 1.63109735e-01 -8.13797891e-01 -4.47602302e-01
-7.82924235e-01 -1.47472715e+00 1.16261935e+00 1.34041473e-01
-6.72542810e-01 -8.47773075e-01 -7.21541867e-02 7.55467713e-01
2.16918454e-01 4.35912490e-01 1.03551400e+00 -3.28413934e-01
-5.34779489e-01 -8.24490860e-02 -6.43718660e-01 9.26372483e-02
3.11296344e-01 8.27265456e-02 -7.14332521e-01 8.50069448e-02
7.83082545e-02 -2.67392904e-01 7.46981144e-01 6.90702021e-01
1.21845031e+00 -7.70075023e-01 -3.05089861e-01 3.60082448e-01
1.23577738e+00 1.47934109e-01 7.00898588e-01 1.90160442e-02
4.64627415e-01 7.30251551e-01 8.10921490e-01 3.13778132e-01
7.38374114e-01 3.64736259e-01 1.62994310e-01 -5.48103809e-01
-4.23236430e-01 -8.24646577e-02 2.88602319e-02 9.35175598e-01
-2.82129437e-01 -2.09023163e-01 -9.87223387e-01 5.38533807e-01
-1.49225628e+00 -6.96529567e-01 -3.59244674e-01 1.81690860e+00
1.15301144e+00 -2.42862538e-01 -1.43121079e-01 -4.39723253e-01
6.95380986e-01 -2.77461857e-01 -2.88890004e-01 -3.17948639e-01
-4.54674140e-02 1.52638435e-01 1.84951395e-01 1.55813202e-01
-8.38319242e-01 6.04056478e-01 6.65892982e+00 5.41910231e-01
-1.16002989e+00 1.93113193e-01 1.12352800e+00 5.98744527e-02
-6.08245909e-01 -3.13760132e-01 -5.66197932e-01 2.79825389e-01
8.56044352e-01 -3.61878097e-01 -1.49727836e-01 6.43445611e-01
4.93342727e-01 -2.67355684e-02 -1.04257119e+00 1.05955219e+00
1.60553813e-01 -2.00225878e+00 6.42928183e-01 1.93149857e-02
9.32739675e-01 -2.61032015e-01 1.07876599e-01 -1.62379533e-01
2.00838372e-01 -1.28473949e+00 4.36002076e-01 7.83947766e-01
1.30024934e+00 -2.54554808e-01 1.03467858e+00 -2.84619126e-02
-8.04664314e-01 3.98105919e-01 -2.97443599e-01 2.98280001e-01
1.45140424e-01 8.31978440e-01 -1.47660124e+00 8.60596895e-01
5.17882288e-01 6.48357987e-01 -5.99548459e-01 1.08651471e+00
-4.11150396e-01 7.10846841e-01 2.75534809e-01 1.37148663e-01
1.39956102e-01 1.30439088e-01 4.58863586e-01 1.23019981e+00
5.07197678e-01 5.28574347e-01 1.30953446e-01 1.32696283e+00
-1.30904138e-01 4.28910434e-01 -6.32441521e-01 -2.38999024e-01
2.82966733e-01 1.35399163e+00 -8.76662374e-01 -5.36103964e-01
-5.66114247e-01 6.37217462e-01 -6.39127940e-02 3.62893373e-01
-8.54312718e-01 5.51827140e-02 1.91039205e-01 3.25568378e-01
-3.19262713e-01 2.16374338e-01 -6.77199483e-01 -1.20763826e+00
6.58423081e-02 -1.06526864e+00 5.31113803e-01 -1.09727061e+00
-1.48645282e+00 1.21153331e+00 -3.08937132e-01 -1.51835239e+00
-4.01451051e-01 -3.96779984e-01 -2.64101863e-01 8.60722840e-01
-1.58187044e+00 -1.29821241e+00 -5.47638357e-01 4.05794621e-01
5.46248972e-01 -2.45570898e-01 1.07632422e+00 2.06214681e-01
-2.66502827e-01 5.75409710e-01 -7.86001563e-01 1.43730685e-01
9.54953909e-01 -1.13294232e+00 3.79004955e-01 5.93202591e-01
2.69423157e-01 8.01355064e-01 3.95115435e-01 -7.64449060e-01
-7.68199444e-01 -1.41829729e+00 1.09421730e+00 -6.51958108e-01
4.37181383e-01 4.52724218e-01 -8.17625344e-01 4.30510432e-01
1.94826245e-01 1.08785465e-01 1.01925409e+00 -3.33467662e-01
-1.97299525e-01 1.39556095e-01 -1.19022036e+00 6.93898320e-01
8.80359411e-01 -2.38368213e-01 -3.76907110e-01 6.61673844e-01
1.06792688e+00 -5.02777815e-01 -1.22634685e+00 4.85648245e-01
3.11929315e-01 -7.76238561e-01 6.29390299e-01 -8.26373637e-01
1.06794631e+00 -3.16894203e-01 2.81900376e-01 -1.03037870e+00
-2.57433504e-01 -3.44397336e-01 2.41801381e-01 1.05055535e+00
9.19158936e-01 -4.67335671e-01 4.66691047e-01 8.17399204e-01
-1.80049166e-01 -9.96695042e-01 -5.82876205e-01 -2.64592677e-01
-1.92346573e-01 -6.30150676e-01 5.65810382e-01 1.03965569e+00
-1.17139548e-01 8.24808106e-02 -1.46420226e-01 1.17431909e-01
3.44036251e-01 2.17785805e-01 5.50740838e-01 -8.65342200e-01
-4.34031673e-02 -2.75503904e-01 -4.89008993e-01 -5.03170550e-01
-9.67510343e-02 -1.00703800e+00 -6.32502045e-03 -1.90637064e+00
7.69754231e-01 -5.41672289e-01 -2.88692507e-04 6.53892040e-01
-5.58699906e-01 5.56079149e-01 1.69833407e-01 5.17448187e-01
-6.55627728e-01 1.30333319e-01 1.73866737e+00 -2.98616439e-01
2.67526451e-02 -7.61383548e-02 -1.19569194e+00 6.79771721e-01
5.01454413e-01 -6.09028518e-01 -2.33917847e-01 -3.48223090e-01
1.17197327e-01 3.23261350e-01 5.75916648e-01 -7.59281933e-01
4.30026650e-02 -3.00271094e-01 2.60850579e-01 -3.40127051e-01
-3.92044127e-01 -3.77435148e-01 1.97021455e-01 5.22605121e-01
-5.50531507e-01 1.13258660e-01 3.77057157e-02 4.01643187e-01
-6.41201556e-01 -5.48057668e-02 4.38880801e-01 -5.56016922e-01
-4.64463919e-01 4.48349744e-01 -3.28930616e-01 2.60827035e-01
1.00808144e+00 -1.03439748e-01 -6.31226122e-01 -4.67261940e-01
-7.97762692e-01 1.03332780e-01 1.77028194e-01 4.39928383e-01
1.07868195e+00 -1.25679827e+00 -1.36997461e+00 -1.69268087e-01
4.61214721e-01 1.28873112e-02 3.46687466e-01 1.09684050e+00
-7.27849364e-01 7.28232861e-01 1.55293629e-01 -5.53380668e-01
-1.17950654e+00 4.32208300e-01 2.13045385e-02 -7.21583962e-01
-6.89585567e-01 5.83347976e-01 5.11170328e-01 -2.35019177e-01
-1.37714595e-01 -6.55740201e-01 -1.70100242e-01 -3.70857686e-01
9.41768885e-01 -6.99936375e-02 1.66188091e-01 -4.79834259e-01
-2.08585620e-01 3.47453296e-01 -2.86789030e-01 -5.43516949e-02
1.16821551e+00 -2.21414268e-01 -1.51900917e-01 2.15977859e-02
7.94399619e-01 6.82568103e-02 -5.51734686e-01 1.56509373e-02
3.11210137e-02 -1.76710635e-01 -4.03711498e-02 -1.28104353e+00
-1.16499317e+00 5.83275676e-01 4.88980144e-01 3.92065495e-02
1.23484647e+00 2.35675722e-01 7.93555617e-01 5.57871675e-03
4.43786711e-01 -4.95598853e-01 1.02591515e-01 -1.80217043e-01
1.25441742e+00 -1.45971763e+00 8.80130157e-02 -7.90309548e-01
-1.15235293e+00 1.11205995e+00 4.35765624e-01 4.63480592e-01
1.74630985e-01 2.13798061e-01 5.23433030e-01 -4.33359683e-01
-8.36428761e-01 -2.01718230e-02 7.44449675e-01 6.28867149e-01
8.90962541e-01 2.91584015e-01 -7.01034367e-01 1.00344527e+00
-2.73033053e-01 4.55566645e-01 7.14063585e-01 4.36365306e-01
-8.44939984e-03 -8.56086791e-01 -4.52587336e-01 7.70804763e-01
-7.93608189e-01 -4.26727355e-01 -3.61952215e-01 6.63126945e-01
2.80590534e-01 1.14665544e+00 -3.59718114e-01 -1.91293284e-01
1.36981070e-01 -4.49386537e-01 2.73651212e-01 -1.05225062e+00
-4.06493932e-01 -2.62786031e-01 1.70089871e-01 -6.65299296e-01
-6.53953075e-01 -2.89433151e-01 -1.38991237e+00 -1.19662486e-01
-2.70013303e-01 2.16625720e-01 4.16113794e-01 1.02938795e+00
6.12195432e-01 5.62125444e-01 3.58628839e-01 2.19147019e-02
-8.35450888e-02 -8.46800983e-01 -4.21734124e-01 5.07489145e-01
2.18120024e-01 -3.35948408e-01 -6.31854683e-02 7.23469555e-01] | [15.04604721069336, -1.3830312490463257] |
3ab2db93-0ebb-4444-9daf-0f815cfe3918 | deep-compressed-pneumonia-detection-for-low | 1911.02007 | null | https://arxiv.org/abs/1911.02007v1 | https://arxiv.org/pdf/1911.02007v1.pdf | Deep Compressed Pneumonia Detection for Low-Power Embedded Devices | Deep neural networks (DNNs) have been expanded into medical fields and triggered the revolution of some medical applications by extracting complex features and achieving high accuracy and performance, etc. On the contrast, the large-scale network brings high requirements of both memory storage and computation resource, especially for portable medical devices and other embedded systems. In this work, we first train a DNN for pneumonia detection using the dataset provided by RSNA Pneumonia Detection Challenge. To overcome hardware limitation for implementing large-scale networks, we develop a systematic structured weight pruning method with filter sparsity, column sparsity and combined sparsity. Experiments show that we can achieve up to 36x compression ratio compared to the original model with 106 layers, while maintaining no accuracy degradation. We evaluate the proposed methods on an embedded low-power device, Jetson TX2, and achieve low power usage and high energy efficiency. | ['Caiwen Ding', 'Sheng Lin', 'Hongjia Li', 'Ning Liu', 'Yanzhi Wang'] | 2019-11-04 | null | null | null | null | ['pneumonia-detection'] | ['medical'] | [ 2.20266446e-01 -1.65052429e-01 -5.09732246e-01 -2.59144604e-01
-1.72668099e-01 1.58617675e-01 -1.90103948e-01 1.08867466e-01
-7.38744080e-01 6.73939645e-01 6.35583475e-02 -2.61040956e-01
-3.54764551e-01 -1.07446802e+00 -1.67928427e-01 -7.73483694e-01
-1.49431545e-02 2.02652082e-01 4.51585144e-01 5.03693759e-01
-3.07204813e-01 4.55754429e-01 -1.22615659e+00 2.57260680e-01
6.50976539e-01 1.35656452e+00 4.78642732e-01 4.22443271e-01
-4.76035581e-04 9.29330349e-01 -3.98598433e-01 -1.22012787e-01
1.75330892e-01 -4.54154871e-02 -3.22608918e-01 -5.06707549e-01
6.45081475e-02 -9.50441658e-01 -8.50179911e-01 1.08102798e+00
9.57323611e-01 -3.35900486e-01 1.29419893e-01 -8.13277483e-01
-1.76492050e-01 5.95395923e-01 -2.86741942e-01 5.74086487e-01
-6.13448679e-01 -1.02584355e-01 5.33847570e-01 -3.17321420e-01
4.46912259e-01 8.12250435e-01 9.98818338e-01 9.52865422e-01
-4.44461703e-01 -9.71929073e-01 -3.35162848e-01 3.30666780e-01
-1.29578841e+00 -6.11941636e-01 5.70456803e-01 2.27891609e-01
1.33341563e+00 3.05952936e-01 8.39457095e-01 1.22516155e+00
5.91226518e-02 6.71404243e-01 5.08508861e-01 -5.21183051e-02
2.59517193e-01 -1.55137181e-01 9.65838060e-02 1.00015306e+00
1.06046414e+00 -1.51727758e-02 -3.96072477e-01 -2.65989095e-01
8.56596470e-01 7.71878123e-01 -2.49559954e-01 2.94581711e-01
-1.04581940e+00 6.55346572e-01 4.50345546e-01 5.43107688e-01
-5.80802858e-01 4.58764404e-01 6.58553958e-01 -2.96032399e-01
9.99443233e-02 -7.53454939e-02 -5.49461603e-01 -2.56655514e-01
-9.46211100e-01 -9.99126658e-02 6.04696631e-01 8.91616642e-01
1.59787685e-02 3.24330986e-01 -2.04219803e-01 1.04377639e+00
1.93155333e-01 6.65334344e-01 9.92534578e-01 -6.84392095e-01
5.69405794e-01 6.68476999e-01 -5.17798364e-01 -8.66938949e-01
-9.29527044e-01 -7.02940226e-01 -1.66855562e+00 -2.41516992e-01
-2.39226684e-01 -5.30331492e-01 -1.05170345e+00 1.49961066e+00
2.49682263e-01 2.67696172e-01 1.97690576e-01 6.97160304e-01
1.21407473e+00 6.16995335e-01 1.54876068e-01 -3.21379751e-01
1.52123392e+00 -8.61023188e-01 -8.98403108e-01 -2.37201691e-01
7.95475900e-01 -1.81160823e-01 4.31516707e-01 4.67735857e-01
-1.07067490e+00 -4.55883950e-01 -1.38660014e+00 -1.90222159e-01
-8.87055472e-02 4.83443111e-01 7.76442945e-01 6.90071523e-01
-8.10528040e-01 8.28207076e-01 -1.47599757e+00 -3.26497078e-01
1.04099548e+00 6.05464816e-01 1.27842188e-01 -6.95179310e-03
-1.14038157e+00 4.85092372e-01 5.42406738e-01 4.20772359e-02
-5.75788260e-01 -8.04520726e-01 -3.55652630e-01 4.99287486e-01
8.20016116e-02 -8.96629870e-01 1.17305553e+00 -2.55118877e-01
-1.24760389e+00 4.46620971e-01 2.18751311e-01 -8.83103490e-01
8.01889077e-02 -2.51226544e-01 -7.48874724e-01 4.62168783e-01
-5.06495655e-01 6.98210359e-01 2.88331270e-01 -3.87834668e-01
-8.16554606e-01 -4.08242136e-01 -2.89667934e-01 3.05996537e-02
-1.20526719e+00 5.43956121e-04 -4.69528198e-01 -6.77434981e-01
1.76911458e-01 -7.25453258e-01 -4.63353395e-01 3.21107149e-01
-4.39943850e-01 -1.14543147e-01 1.15547645e+00 -6.10782921e-01
1.56539035e+00 -2.11737657e+00 -3.07846427e-01 1.39561698e-01
7.01429069e-01 6.92446291e-01 5.66923358e-02 -3.65384430e-01
3.13814163e-01 1.01341315e-01 -6.29417598e-02 -1.82847738e-01
-4.94154871e-01 4.71015751e-01 1.43129379e-01 1.58539742e-01
-1.31645799e-01 8.19197714e-01 -4.85561162e-01 -7.05279410e-01
-3.31967250e-02 8.37351143e-01 -6.41640902e-01 -8.03976730e-02
2.71008939e-01 -3.14476877e-01 -9.00472879e-01 8.09838831e-01
5.64561665e-01 -5.90573907e-01 3.93306196e-01 -6.19097769e-01
3.58015865e-01 3.16258371e-01 -8.58395219e-01 1.58029485e+00
-3.52108449e-01 4.39522743e-01 4.24718320e-01 -1.11560977e+00
7.65692353e-01 5.21011412e-01 8.35202813e-01 -8.15583766e-01
5.67301095e-01 3.72712940e-01 1.59577668e-01 -6.21275425e-01
4.42252196e-02 -1.40369356e-01 3.44619662e-01 5.19077033e-02
1.33170441e-01 5.13134778e-01 2.99436916e-02 -6.76635420e-03
1.46877301e+00 -6.01193309e-01 2.34113961e-01 -3.13636869e-01
1.54047936e-01 -1.55414358e-01 9.45472777e-01 5.01275420e-01
-2.56476402e-01 3.19122612e-01 8.00431445e-02 -6.62335038e-01
-8.78614843e-01 -8.26873064e-01 -3.50157320e-01 3.89939100e-01
-7.85917118e-02 -5.00611782e-01 -7.64548838e-01 -5.19637644e-01
-2.76041031e-01 8.73656347e-02 -8.28113034e-02 -3.23823363e-01
-9.48559821e-01 -1.24835515e+00 1.01372349e+00 9.19638455e-01
1.02728653e+00 -1.02203417e+00 -1.29650795e+00 3.70475054e-01
3.44264507e-02 -1.41468632e+00 -3.53481248e-02 4.13136393e-01
-1.61720324e+00 -1.01356876e+00 -4.40903932e-01 -1.03098142e+00
5.46495259e-01 5.61459549e-02 9.24082339e-01 5.09498656e-01
-8.39156926e-01 -1.55342802e-01 -1.14759728e-01 -6.80739045e-01
8.27490389e-02 2.41163194e-01 3.12401414e-01 -7.10322738e-01
4.99206722e-01 -7.61352956e-01 -1.05363905e+00 -2.52707601e-01
-9.41508830e-01 1.25660315e-01 1.07046390e+00 6.53059721e-01
7.36434221e-01 4.84507233e-01 7.98789620e-01 -8.11056197e-01
4.61238325e-01 -3.09335947e-01 -4.47652102e-01 7.90675506e-02
-7.87108660e-01 -1.51828155e-01 8.35631430e-01 -3.53869438e-01
-8.43177199e-01 1.00320488e-01 -4.92471009e-01 -4.52575594e-01
8.24658424e-02 2.91727781e-01 -8.09893906e-02 4.45209667e-02
4.53626335e-01 1.34005070e-01 -1.06229939e-01 -6.09188735e-01
-2.65457690e-01 8.44406724e-01 5.27238190e-01 -2.15423778e-01
4.83918995e-01 4.64506775e-01 3.23998630e-01 -7.71115959e-01
-5.79484165e-01 -5.96830919e-02 -1.81264415e-01 2.84838110e-01
9.03951526e-01 -1.17369223e+00 -7.00495243e-01 3.47502381e-01
-1.03238463e+00 -8.76072720e-02 -1.53238729e-01 9.05203283e-01
1.47071227e-01 1.38909638e-01 -9.45050061e-01 -3.66561592e-01
-1.38838446e+00 -9.09305990e-01 6.09428346e-01 4.18741435e-01
9.55568627e-02 -6.93383336e-01 -1.49607971e-01 4.95033376e-02
7.10561216e-01 1.01374812e-01 1.09759212e+00 -5.80092311e-01
-5.59950471e-01 -1.96046785e-01 -4.99755979e-01 4.70230848e-01
3.02841991e-01 -3.81825507e-01 -8.41063499e-01 -2.69486606e-01
4.47149813e-01 -3.45277697e-01 1.04598153e+00 5.07269979e-01
1.94745815e+00 -4.17162836e-01 -9.32462335e-01 9.06240284e-01
1.61163616e+00 5.55911601e-01 5.25830865e-01 -2.57318597e-02
8.69634867e-01 -1.60123527e-01 -1.01603165e-01 4.91785198e-01
-8.90728608e-02 1.24267645e-01 3.22568357e-01 -2.14184467e-02
-3.02471131e-01 1.23351678e-01 -1.65636852e-01 1.40110433e+00
-9.31500047e-02 -4.90512103e-01 -8.10157359e-01 5.38962781e-01
-1.53889799e+00 -6.76195979e-01 1.17798172e-01 1.86296165e+00
9.19633865e-01 3.77586067e-01 -3.79508942e-01 2.95384318e-01
4.21885341e-01 1.00818664e-01 -7.74525702e-01 -7.58375600e-02
1.32678017e-01 5.37522554e-01 6.61884069e-01 -1.86757281e-01
-9.76287663e-01 2.51151949e-01 6.39200687e+00 1.11223137e+00
-1.18680441e+00 3.47894460e-01 7.83285916e-01 -8.62134397e-01
1.53814822e-01 -8.69897604e-01 -9.85983491e-01 3.61874849e-01
1.13959050e+00 2.24963784e-01 8.19118023e-02 1.03019428e+00
-1.35972708e-01 2.70447493e-01 -8.54694784e-01 1.24415708e+00
-2.60629207e-01 -1.61635792e+00 7.40520656e-02 -3.35026979e-02
3.13040644e-01 4.93898511e-01 -2.49820292e-01 1.86466023e-01
-1.31716579e-01 -9.54490602e-01 -9.75744501e-02 6.58216001e-03
1.05859673e+00 -9.32900965e-01 8.74036670e-01 4.47951227e-01
-1.20215964e+00 -3.21815580e-01 -7.38651872e-01 8.96412879e-02
8.43894705e-02 1.06939948e+00 -6.02276504e-01 2.43706226e-01
1.09520543e+00 5.85794806e-01 -8.35426375e-02 8.19252789e-01
3.80709499e-01 8.12642038e-01 -6.23721421e-01 -4.55606222e-01
-1.65339373e-02 1.21304698e-01 2.85012811e-01 1.26693714e+00
6.25386119e-01 6.11350954e-01 -7.18748122e-02 3.95869762e-01
-6.20779514e-01 -2.08085731e-01 -4.26611990e-01 -8.28999132e-02
6.07490301e-01 1.56446362e+00 -7.71195531e-01 -5.64854622e-01
-4.57730830e-01 6.46310568e-01 1.80255100e-01 -1.21412411e-01
-9.30976331e-01 -6.45068526e-01 6.32162452e-01 -2.76820548e-02
1.38511866e-01 -1.05403714e-01 -4.77382779e-01 -9.25755978e-01
2.10806772e-01 -5.47640204e-01 1.92114115e-01 -4.33324099e-01
-7.69248903e-01 7.04167068e-01 -3.67249578e-01 -1.13626325e+00
3.59207124e-01 -7.55582690e-01 -5.08982539e-01 1.95347339e-01
-1.50470352e+00 -6.80567026e-01 -5.63924193e-01 4.97443885e-01
3.78408939e-01 -3.34227949e-01 8.94175828e-01 9.54016387e-01
-8.37797523e-01 6.45710886e-01 2.27583438e-01 3.78399283e-01
1.79887131e-01 -5.53233206e-01 1.31235629e-01 6.40516102e-01
-2.54392952e-01 5.41076839e-01 5.48055246e-02 -6.90121651e-01
-1.51067698e+00 -1.53900361e+00 6.69679463e-01 2.30281651e-01
1.73414752e-01 -3.14729005e-01 -8.17035913e-01 3.06533664e-01
1.16683513e-01 3.02042216e-01 7.37750351e-01 -2.83801764e-01
9.94600132e-02 -5.26030004e-01 -1.49598682e+00 4.61203337e-01
1.51959372e+00 -1.63083330e-01 -2.70649970e-01 5.48903525e-01
9.89257991e-01 -4.56987232e-01 -8.74008179e-01 9.90628898e-01
6.09108031e-01 -6.96171463e-01 1.16352212e+00 -3.64647686e-01
3.76033574e-01 9.73417535e-02 -1.63114175e-01 -6.83230102e-01
-5.78600049e-01 -1.55580476e-01 -6.43693268e-01 9.18111503e-01
7.62454122e-02 -3.77741307e-01 1.33881891e+00 2.88792610e-01
-2.18515724e-01 -1.37168968e+00 -1.10079324e+00 -6.14125967e-01
-3.65587145e-01 -3.71087044e-01 5.14801383e-01 6.53327107e-01
-8.80693868e-02 2.50303775e-01 -4.17250454e-01 7.05697984e-02
5.66507876e-01 -2.75740057e-01 -9.31248516e-02 -1.37290943e+00
-2.00966865e-01 -3.70148838e-01 -3.73239011e-01 -9.64110970e-01
-3.30427855e-01 -7.10352480e-01 -1.47455603e-01 -1.74102008e+00
3.53349924e-01 -5.12108386e-01 -7.12916017e-01 6.40422583e-01
1.23580538e-01 3.89429599e-01 -2.51991659e-01 -5.12585081e-02
-6.43730760e-01 4.10695046e-01 9.81494486e-01 -1.37635484e-01
-5.83830848e-02 -2.59301424e-01 -6.23710871e-01 9.73315716e-01
9.26108599e-01 -7.60210037e-01 -6.35958731e-01 -8.38121712e-01
1.51669219e-01 2.09736694e-02 1.44755900e-01 -1.66897583e+00
4.28029299e-01 1.16663560e-01 8.35633755e-01 -6.79015517e-01
2.73647845e-01 -1.15011108e+00 2.85941213e-02 1.34711301e+00
4.69360612e-02 -9.54663381e-02 4.09113735e-01 3.39090735e-01
-2.05260441e-02 -1.51958004e-01 7.38165736e-01 6.15641922e-02
-3.59132349e-01 7.13615000e-01 -3.45426261e-01 -2.75342166e-01
7.73482502e-01 -2.30990462e-02 -5.07736683e-01 3.64093781e-01
-3.00137490e-01 1.89194664e-01 -3.12819630e-01 3.13228339e-01
9.12828088e-01 -1.52741325e+00 -3.47660452e-01 3.46927851e-01
-3.33245933e-01 4.20426369e-01 5.99554121e-01 6.70116305e-01
-8.13321531e-01 7.43628502e-01 -2.81213373e-01 -4.37175095e-01
-1.48714757e+00 3.24521542e-01 3.34117055e-01 -4.94976759e-01
-9.45576131e-01 8.61028910e-01 -3.95167544e-02 -2.57451180e-02
4.93458331e-01 -8.12808573e-01 -1.91367000e-01 -2.52097875e-01
8.22357416e-01 7.35624194e-01 2.55831599e-01 2.49805257e-01
-5.22706091e-01 4.97276574e-01 2.63212677e-02 5.68068564e-01
1.59261715e+00 2.38657787e-01 -1.91132762e-02 -2.65944958e-01
1.27961183e+00 -2.05227077e-01 -7.68637121e-01 -2.07635984e-01
-3.05350810e-01 -1.97140593e-02 4.53777671e-01 -6.81219101e-01
-1.76330006e+00 8.89959991e-01 1.17747533e+00 -9.85089913e-02
1.62052965e+00 -3.34579200e-01 1.65417004e+00 1.01953197e+00
2.72893608e-01 -1.17051530e+00 4.36488492e-03 2.27901593e-01
2.58152485e-01 -1.01501381e+00 3.41030329e-01 -8.61266479e-02
-1.29925340e-01 1.23150229e+00 7.63283372e-01 -8.77182856e-02
1.02881801e+00 7.72785783e-01 -1.22114606e-01 -3.82319778e-01
-7.36386001e-01 3.52389663e-01 -2.14728564e-02 4.34718221e-01
2.01182336e-01 -5.63392136e-03 -3.48719090e-01 1.09005678e+00
4.53576036e-02 5.13825595e-01 7.97498673e-02 1.10063660e+00
-7.64796317e-01 -8.40157509e-01 1.50097251e-01 1.15134180e+00
-1.01213038e+00 -3.02289397e-01 1.85221463e-01 4.37481403e-01
3.66793364e-01 6.79179251e-01 2.49939471e-01 -7.12502778e-01
1.22893520e-01 -1.95962250e-01 2.53538907e-01 -3.25556695e-01
-4.81464505e-01 1.44508928e-01 2.63627857e-01 -5.60173035e-01
-4.00704712e-01 -8.04100111e-02 -1.71594417e+00 -2.18182400e-01
-2.53221005e-01 -8.40410218e-02 7.89776087e-01 7.97083735e-01
6.31682992e-01 1.12034726e+00 2.43487984e-01 -3.48349661e-01
-4.38428164e-01 -9.52259123e-01 -5.31596899e-01 -2.26339445e-01
2.89756805e-01 -1.16568610e-01 3.54915368e-03 -2.88893729e-01] | [8.538002967834473, 3.0167288780212402] |
85a43ebc-c96e-42fc-a724-06d3b37787a1 | covert-communication-based-on-the-poisoning | 2306.01342 | null | https://arxiv.org/abs/2306.01342v1 | https://arxiv.org/pdf/2306.01342v1.pdf | Covert Communication Based on the Poisoning Attack in Federated Learning | Covert communication has become an important area of research in computer security. It involves hiding specific information on a carrier for message transmission and is often used to transmit private data, military secrets, and even malware. In deep learning, many methods have been developed for hiding information in models to achieve covert communication. However, these methods are not applicable to federated learning, where model aggregation invalidates the exact information embedded in the model by the client. To address this problem, we propose a novel method for covert communication in federated learning based on the poisoning attack. Our approach achieves 100% accuracy in covert message transmission between two clients and is shown to be both stealthy and robust through extensive experiments. However, existing defense methods are limited in their effectiveness against our attack scheme, highlighting the urgent need for new protection methods to be developed. Our study emphasizes the necessity of research in covert communication and serves as a foundation for future research in federated learning attacks and defenses. | ['Rong Wang', 'Junchuan Liang'] | 2023-06-02 | null | null | null | null | ['computer-security'] | ['miscellaneous'] | [ 4.86988753e-01 6.42297268e-02 -2.98961788e-01 -1.64737985e-01
-6.10037684e-01 -7.46092200e-01 6.28127217e-01 6.34133965e-02
-5.14841974e-02 7.88641214e-01 -8.75528753e-02 -6.38043046e-01
1.42404824e-01 -1.01867771e+00 -8.89569998e-01 -9.40910637e-01
-4.73858029e-01 1.06333502e-01 7.53208026e-02 -2.84844786e-01
6.87109530e-02 4.78389591e-01 -1.11889410e+00 6.14540815e-01
5.49463212e-01 7.13505387e-01 -5.30986607e-01 3.10127050e-01
-2.89992720e-01 7.33223438e-01 -1.01348269e+00 -7.98203349e-01
3.76305610e-01 -6.79797828e-01 -8.97173464e-01 2.63777617e-02
2.55860031e-01 -5.53220928e-01 -6.19666159e-01 1.31980026e+00
2.11603016e-01 -6.47740781e-01 2.71104634e-01 -1.51683140e+00
-4.50058371e-01 1.24118125e+00 -5.95165014e-01 -1.51822999e-01
5.00232987e-02 1.16971605e-01 7.05255270e-01 -8.87912586e-02
2.62885511e-01 1.38998353e+00 3.97706866e-01 9.63209450e-01
-7.84091651e-01 -1.37911928e+00 1.33302465e-01 6.78392798e-02
-9.87790763e-01 -1.36790231e-01 6.49157047e-01 8.48694965e-02
4.99953508e-01 8.38161767e-01 7.04911470e-01 1.25229526e+00
6.76471233e-01 9.68265533e-01 1.44557583e+00 -3.76328766e-01
1.81559086e-01 5.29254377e-01 -8.38765651e-02 8.71532142e-01
7.31503546e-01 6.33082032e-01 -3.39771271e-01 -6.88739300e-01
4.45554703e-01 3.28904331e-01 -3.53560030e-01 -1.29529297e-01
-7.89842546e-01 1.22078252e+00 4.82055515e-01 4.38155085e-01
3.43928695e-01 4.45959628e-01 4.54577416e-01 8.54227006e-01
5.81775486e-01 4.75631505e-02 -3.08878362e-01 3.69894385e-01
-4.22520816e-01 5.33838384e-02 1.04550445e+00 6.14270151e-01
4.63622659e-01 1.32779926e-01 3.12535137e-01 -4.08893684e-03
4.86879945e-01 5.14324665e-01 3.20770144e-01 -7.64541805e-01
2.65007496e-01 5.34136832e-01 -3.09463590e-01 -1.06395447e+00
2.74503857e-01 -7.74235249e-01 -6.10562563e-01 2.79739410e-01
5.71614280e-02 -1.57572702e-01 -4.03927416e-01 1.44194734e+00
5.45331180e-01 2.39277452e-01 3.81382376e-01 7.19900489e-01
2.46208519e-01 4.92846131e-01 -2.02041075e-01 -1.02105148e-01
1.09937871e+00 -8.32423925e-01 -6.30423546e-01 -2.19908413e-02
1.03152359e+00 -5.99253178e-01 3.42786342e-01 2.27917612e-01
-5.20878017e-01 3.24045867e-01 -1.12801695e+00 6.12850428e-01
-4.30361301e-01 -9.37735200e-01 9.07296121e-01 1.25111461e+00
-6.31020248e-01 4.11986947e-01 -7.62195587e-01 6.46496490e-02
1.00440288e+00 6.38815820e-01 -2.36656442e-01 -2.96504796e-01
-1.53380704e+00 6.70341432e-01 3.56711745e-01 -4.26340669e-01
-1.45764339e+00 -1.90569416e-01 -7.97859132e-01 3.32183875e-02
2.30713233e-01 -3.61657977e-01 1.15468943e+00 -1.12312376e+00
-9.72258449e-01 6.24508083e-01 3.57086211e-01 -1.13016677e+00
5.56299269e-01 8.67771357e-03 -6.12247646e-01 1.85882151e-01
-4.70885873e-01 2.65471965e-01 1.23637342e+00 -1.57801437e+00
-7.53473759e-01 -4.53315198e-01 3.12728494e-01 -2.00326845e-01
-9.57407236e-01 1.36505850e-02 3.38890463e-01 -5.80970824e-01
-1.30523160e-01 -1.00310409e+00 -1.78254887e-01 1.78256765e-01
-5.04934013e-01 1.60147443e-01 2.00915074e+00 -4.03448939e-01
9.02146935e-01 -2.05181646e+00 -1.96412891e-01 3.42240155e-01
4.14415270e-01 5.65834522e-01 4.46126983e-02 6.25900626e-01
3.24026227e-01 3.68338734e-01 -4.59145427e-01 -9.01062936e-02
-5.18057160e-02 2.39195943e-01 -7.15411067e-01 9.35261250e-01
-2.91978180e-01 8.37165773e-01 -8.01244438e-01 -5.63225329e-01
-2.38752179e-03 8.94898236e-01 -2.72079170e-01 -9.33886543e-02
-3.86011481e-01 5.70220768e-01 -8.56293321e-01 7.92845368e-01
7.26873159e-01 -6.68747276e-02 3.35553527e-01 4.60350722e-01
2.49127731e-01 1.05072536e-01 -6.76191092e-01 1.09789050e+00
-2.42077097e-01 5.21056235e-01 3.75023186e-01 -8.18548560e-01
7.54381239e-01 5.36353111e-01 4.56742108e-01 -1.31232545e-01
5.46311855e-01 4.47263539e-01 -1.41274305e-02 -2.72572726e-01
2.75115911e-02 -1.55101836e-01 -8.06409568e-02 1.05791044e+00
-4.98347700e-01 1.28477111e-01 -5.03252625e-01 2.61169612e-01
1.18232584e+00 -3.10223252e-01 -2.65965760e-01 -1.40203477e-03
5.75814843e-01 8.35384578e-02 1.71366304e-01 6.95142984e-01
-2.28246078e-01 -2.70547837e-01 1.71381265e-01 -5.84282100e-01
-6.43060386e-01 -4.32861239e-01 -1.43514611e-02 5.82543254e-01
4.12312150e-01 -2.97172189e-01 -1.05133820e+00 -1.45767117e+00
2.48932019e-01 5.68364799e-01 -3.11214149e-01 -7.26222157e-01
-4.48960960e-01 -6.16356373e-01 1.09646595e+00 -9.11905840e-02
8.73524308e-01 -8.68706107e-01 -6.58065021e-01 1.88460842e-01
-1.36659369e-01 -9.25531924e-01 -4.10181433e-01 6.81754276e-02
-1.05710518e+00 -1.32081079e+00 -3.55248809e-01 -7.53953755e-01
8.14746261e-01 6.40616477e-01 4.09733891e-01 8.39861631e-01
-1.87104344e-01 3.62792581e-01 -3.64782631e-01 -6.20956123e-01
-1.11653244e+00 6.67252466e-02 -2.98176035e-02 3.06617737e-01
3.50966007e-01 -2.38610044e-01 -4.42798018e-01 3.51752102e-01
-1.50574064e+00 -2.63967603e-01 5.40469408e-01 7.38510132e-01
-1.11892512e-02 6.89200342e-01 4.22664255e-01 -1.17256069e+00
5.44049084e-01 -4.98921514e-01 -5.97527683e-01 2.49820635e-01
-6.20793402e-01 1.26747668e-01 6.73607767e-01 -5.83436072e-01
-7.37557173e-01 -2.37176359e-01 -3.07438038e-02 -3.72260630e-01
2.89276727e-02 1.05823249e-01 -2.97919840e-01 -1.04167461e+00
4.11643833e-01 3.95780414e-01 4.09358561e-01 -3.57787848e-01
-9.10164416e-03 8.89674425e-01 -9.63415131e-02 -4.16414976e-01
1.33142090e+00 9.81264651e-01 1.21150360e-01 -4.55553859e-01
-6.09954596e-01 1.72597528e-01 1.02835789e-01 1.16976360e-02
2.76678860e-01 -5.31328797e-01 -6.68624163e-01 6.24339879e-01
-1.07705021e+00 1.15375340e-01 1.35253280e-01 3.01994711e-01
1.84875373e-02 7.34338403e-01 -8.59730661e-01 -9.13489580e-01
-6.62985027e-01 -1.20355463e+00 6.07933760e-01 -1.53026238e-01
6.61561564e-02 -1.21345162e+00 -1.66164950e-01 6.29717767e-01
5.92140377e-01 5.44304252e-01 9.60900724e-01 -8.27171326e-01
-9.38250303e-01 -7.62449503e-01 1.92412496e-01 4.32379067e-01
3.26802611e-01 -3.72715414e-01 -9.65307832e-01 -9.38147962e-01
4.28030014e-01 -3.90666872e-01 9.94845510e-01 -3.73876840e-01
1.31172132e+00 -1.10904157e+00 -5.14225006e-01 5.68198383e-01
1.36843073e+00 1.30237013e-01 5.76864898e-01 3.11168939e-01
6.51784360e-01 7.96470106e-01 2.49199182e-01 2.44027063e-01
1.95840001e-02 2.08967358e-01 1.12041032e+00 -1.11111288e-03
1.37629971e-01 -3.04648042e-01 6.29578650e-01 3.99266064e-01
4.45670515e-01 -2.89903551e-01 -5.59614301e-01 1.55555025e-01
-1.45617223e+00 -1.07202518e+00 -8.72629359e-02 1.99281454e+00
7.57406175e-01 1.36807933e-01 -1.66907191e-01 4.94131476e-01
8.15747082e-01 2.09086686e-01 -5.39881051e-01 -4.94256854e-01
-1.42036662e-01 7.57794129e-03 7.00711489e-01 4.89958674e-01
-1.07500386e+00 9.12359834e-01 6.18802691e+00 8.80755901e-01
-1.29089153e+00 3.95735711e-01 5.93967915e-01 1.51542738e-01
-5.58314919e-01 2.74278522e-01 -7.28510261e-01 5.33357382e-01
9.13663924e-01 -2.95509957e-02 4.75289196e-01 7.18726873e-01
-4.18526381e-01 3.89784157e-01 -7.39443541e-01 4.72296327e-01
-1.12811439e-02 -1.36140990e+00 2.39561543e-01 7.03445911e-01
7.66752064e-01 -3.00224036e-01 3.99877697e-01 9.47612301e-02
2.88982987e-01 -1.15187824e+00 3.54751587e-01 -2.28295818e-01
5.24858713e-01 -1.14975917e+00 6.79134011e-01 6.00072861e-01
-7.39169836e-01 -3.17444712e-01 -3.09896559e-01 -4.24686857e-02
-2.73202658e-01 2.62322664e-01 -6.41262233e-01 4.51342702e-01
4.32994992e-01 3.48170400e-01 -2.00700343e-01 7.11851597e-01
-1.11120909e-01 8.51716280e-01 -1.59644559e-01 -9.78174582e-02
5.49471021e-01 2.72560090e-01 8.19349289e-01 1.00315297e+00
1.08190708e-01 -4.24377143e-01 2.96100169e-01 6.17271841e-01
-2.63416708e-01 -2.10302785e-01 -1.16652286e+00 -2.17939079e-01
5.99314511e-01 9.08624291e-01 -6.76424325e-01 -1.72020689e-01
-3.69015843e-01 8.40809643e-01 -7.71527737e-02 9.64697301e-02
-8.42765749e-01 -2.96324342e-01 7.33882308e-01 1.44328400e-01
1.58107981e-01 -1.75324053e-01 -1.68346748e-01 -1.11351073e+00
-3.71794432e-01 -1.61062312e+00 5.23207366e-01 3.75012040e-01
-1.06683075e+00 7.22896218e-01 -2.84608323e-02 -1.18472564e+00
1.34487212e-01 -2.59491146e-01 -5.30218184e-01 3.19480509e-01
-1.55654633e+00 -1.31838620e+00 1.70038104e-01 8.94360125e-01
2.24044234e-01 -4.60478634e-01 1.08063745e+00 5.05622476e-02
-3.76501381e-01 8.26432049e-01 1.87784895e-01 2.05533832e-01
2.36483052e-01 -4.43682849e-01 2.90004760e-01 7.87282228e-01
4.04939502e-01 5.54867864e-01 7.29687512e-01 -8.61735225e-01
-1.75908995e+00 -1.23130977e+00 5.90019941e-01 -1.97746843e-01
4.62340474e-01 -2.80073255e-01 -8.96033406e-01 8.75578761e-01
3.52882564e-01 -8.92030075e-02 8.24451923e-01 -5.76856434e-01
-7.79709458e-01 -7.00392723e-02 -1.89354312e+00 3.20346445e-01
6.95709407e-01 -4.80826497e-01 -9.11093354e-02 4.80815023e-01
1.15348017e+00 -2.91052788e-01 -4.73956466e-01 1.14200227e-01
5.13815761e-01 -8.64661992e-01 6.93296432e-01 -5.50933063e-01
-4.09756154e-02 -4.65023071e-02 -2.19582573e-01 -9.77069318e-01
3.31738114e-01 -9.46978629e-01 -8.52676392e-01 9.44829941e-01
1.43163115e-01 -1.22566652e+00 9.02039945e-01 2.44978562e-01
3.07936668e-01 -8.88689756e-01 -1.03434348e+00 -9.24247265e-01
3.35675538e-01 -2.75238007e-01 1.15566862e+00 1.15529978e+00
-1.95873201e-01 -2.89393723e-01 -6.71153247e-01 3.31967622e-01
1.44950747e+00 1.36952370e-01 6.00323498e-01 -1.10361910e+00
-1.93314269e-01 -1.16817333e-01 -3.43588918e-01 -4.95012403e-01
4.96771455e-01 -1.01098371e+00 -6.81744218e-01 -9.83941793e-01
1.15593292e-01 -3.63217622e-01 -2.41562858e-01 7.46101975e-01
4.52554852e-01 5.25691926e-01 1.96820453e-01 4.02472675e-01
-2.66916543e-01 4.94212270e-01 1.27426386e+00 -7.58957922e-01
2.42044985e-01 3.10657322e-01 -9.89434123e-01 5.60937583e-01
1.28129292e+00 -1.06122792e+00 -3.56537640e-01 -4.51721281e-01
-2.92090982e-01 -1.16342448e-01 6.32237077e-01 -8.39718163e-01
6.18696511e-02 -2.33234063e-01 1.12364292e-01 -1.60404116e-01
1.90923840e-01 -1.48689485e+00 1.88661456e-01 1.45235813e+00
-3.76799673e-01 -2.45271415e-01 -8.79172385e-02 7.51324296e-01
3.22501734e-02 -1.39765814e-01 8.40648293e-01 -3.51661563e-01
-1.58222005e-01 5.60312510e-01 -3.23714018e-01 -3.29527825e-01
1.30724025e+00 -1.06850475e-01 -5.92488706e-01 -6.39108002e-01
-2.17308775e-01 2.59048313e-01 6.68888271e-01 3.61959010e-01
9.58706737e-01 -1.10257781e+00 -7.36703038e-01 4.45171803e-01
-2.16360763e-01 -4.50869471e-01 -2.95996338e-01 4.75236177e-01
-3.87403667e-01 4.75552529e-01 -6.73017949e-02 -2.81977504e-01
-1.71418810e+00 6.98811650e-01 4.59072649e-01 -2.26341993e-01
-8.17877531e-01 5.28767109e-01 -1.24672957e-01 -3.48436162e-02
7.70499587e-01 3.13844085e-01 1.94564462e-01 -4.92763519e-01
8.06007981e-01 2.52020538e-01 -2.18828246e-01 -7.38588095e-01
-2.19909310e-01 2.03156963e-01 -3.67013067e-01 8.82970616e-02
1.01956332e+00 -4.97835875e-02 -4.92070556e-01 -7.40491077e-02
1.43928468e+00 8.21631998e-02 -9.71727014e-01 -3.52127522e-01
-9.88567770e-02 -8.27710569e-01 2.20664635e-01 -6.01636052e-01
-1.40494859e+00 6.30330026e-01 6.05853736e-01 4.93432850e-01
8.03356886e-01 -1.84091836e-01 1.47575176e+00 4.65885252e-01
1.08910823e+00 -4.21064466e-01 5.58791161e-02 2.34172940e-01
4.28426981e-01 -1.10784292e+00 -1.46551907e-01 -4.19758767e-01
-2.50050545e-01 1.00629592e+00 6.13058627e-01 1.66856311e-02
6.75340474e-01 4.22426403e-01 2.58983999e-01 -2.52442002e-01
-7.56621480e-01 3.47497523e-01 -1.99939162e-01 6.23858750e-01
9.04143527e-02 -3.46533358e-02 -3.43794525e-01 1.64746051e-03
-2.96870731e-02 -4.32706714e-01 3.59468639e-01 1.47623384e+00
-6.84460759e-01 -1.57387328e+00 -7.79703438e-01 1.43988997e-01
-1.05564868e+00 2.19166100e-01 -8.10269773e-01 6.71407461e-01
-3.36329639e-02 1.10933614e+00 -4.95898068e-01 -7.30761886e-01
-4.20611650e-01 -8.89467970e-02 4.09534067e-01 -3.89313847e-01
-8.88428628e-01 -1.67160586e-01 -2.14439735e-01 -1.88201860e-01
-3.77139747e-01 -1.13020003e-01 -1.25900769e+00 -8.60578060e-01
-5.95901787e-01 6.07219338e-01 8.53268862e-01 7.82644808e-01
1.48923174e-01 2.15213552e-01 1.20149553e+00 -2.46899426e-01
-9.24761713e-01 -4.33064431e-01 -6.18551612e-01 1.24345042e-01
7.26506650e-01 -3.29356611e-01 -7.28019416e-01 -4.09768969e-01] | [5.723423004150391, 7.652279376983643] |
34a23611-eebc-4086-b145-f5302e4d8e52 | sketch-tokens-a-learned-mid-level | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Lim_Sketch_Tokens_A_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Lim_Sketch_Tokens_A_2013_CVPR_paper.pdf | Sketch Tokens: A Learned Mid-level Representation for Contour and Object Detection | We propose a novel approach to both learning and detecting local contour-based representations for mid-level features. Our features, called sketch tokens, are learned using supervised mid-level information in the form of hand drawn contours in images. Patches of human generated contours are clustered to form sketch token classes and a random forest classifier is used for efficient detection in novel images. We demonstrate our approach on both topdown and bottom-up tasks. We show state-of-the-art results on the top-down task of contour detection while being over 200x faster than competing methods. We also achieve large improvements in detection accuracy for the bottom-up tasks of pedestrian and object detection as measured on INRIA [5] and PASCAL [10], respectively. These gains are due to the complementary information provided by sketch tokens to low-level features such as gradient histograms. | ['Piotr Dollar', 'C. L. Zitnick', 'Joseph J. Lim'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['contour-detection'] | ['computer-vision'] | [ 2.92481512e-01 -2.78236091e-01 -2.10891113e-01 -3.57537150e-01
-1.15254426e+00 -7.83974409e-01 8.89048696e-01 4.86397177e-01
-5.10792613e-01 3.08282077e-01 -1.60830226e-02 -4.79393378e-02
5.35957456e-01 -9.27512050e-01 -6.26606941e-01 -3.91683012e-01
-4.76659209e-01 1.54943019e-01 9.25190091e-01 -4.72176485e-02
6.78941309e-01 9.87618923e-01 -1.76426792e+00 7.37329423e-01
2.44862407e-01 1.05330551e+00 -7.72702768e-02 1.22214818e+00
-2.20861822e-01 3.26804996e-01 -5.99202394e-01 -4.62743431e-01
4.12662804e-01 -1.44536048e-01 -3.63311261e-01 3.37731659e-01
1.27991426e+00 -1.94648936e-01 -1.01188287e-01 1.05781496e+00
4.58787680e-01 -1.80348083e-02 8.69126499e-01 -1.13528752e+00
-2.92129219e-01 5.32012768e-02 -1.17444444e+00 3.65995497e-01
4.36340779e-01 -1.77032873e-02 9.90182936e-01 -1.60789764e+00
7.83554614e-01 1.60300052e+00 8.72295201e-01 3.81973922e-01
-1.37020302e+00 -5.48125863e-01 2.86109895e-01 -3.91915180e-02
-1.31179023e+00 -3.23052883e-01 7.92545319e-01 -5.73296428e-01
8.53756547e-01 1.10469118e-01 7.04763949e-01 5.31129241e-01
1.99878678e-01 1.26284802e+00 1.13694739e+00 -7.98315763e-01
8.02597180e-02 1.48289278e-01 -2.95410156e-02 1.24630129e+00
3.88711274e-01 4.14775610e-01 -7.02054441e-01 -3.34084898e-01
1.28595638e+00 -1.38681874e-01 2.78948635e-01 -6.57126188e-01
-1.09092128e+00 9.09182549e-01 6.55639231e-01 8.83357972e-02
-2.04514742e-01 3.35666001e-01 4.56403822e-01 4.17433912e-03
5.08831501e-01 8.05773027e-03 -1.77796885e-01 1.51454881e-01
-1.17329168e+00 2.44461343e-01 5.76775789e-01 1.23989356e+00
7.41794407e-01 -8.63687918e-02 -3.60662133e-01 7.46154428e-01
3.99531089e-02 6.13587618e-01 3.47980820e-02 -9.03043449e-01
3.65237117e-01 4.25826937e-01 3.29845071e-01 -9.95442271e-01
-3.90700936e-01 -3.13292623e-01 -5.16868412e-01 9.56625044e-01
8.76841366e-01 -2.90753935e-02 -1.21713185e+00 1.18575168e+00
2.42688552e-01 -1.69024512e-01 -4.29922312e-01 5.33568561e-01
6.03939056e-01 4.82652456e-01 1.79471865e-01 1.94177181e-01
1.67004704e+00 -9.63817775e-01 -1.88183337e-01 -2.70833462e-01
1.98699936e-01 -9.63772833e-01 8.66473675e-01 5.20852506e-01
-1.13905811e+00 -9.09248650e-01 -1.10905552e+00 -8.08885545e-02
-7.53436625e-01 5.87290049e-01 6.75166249e-01 8.23324323e-01
-9.30251658e-01 6.92177534e-01 -4.71328288e-01 -4.51437861e-01
8.48838210e-01 -1.04224198e-01 -3.19430530e-01 -7.95973241e-02
-3.51426750e-01 6.20671868e-01 2.60681892e-03 -3.00790012e-01
-6.12779558e-01 -5.02282560e-01 -9.35991585e-01 1.72431655e-02
2.45046347e-01 -2.11453527e-01 1.08557916e+00 -6.73144996e-01
-1.00301397e+00 1.27050102e+00 -1.71223089e-01 -3.40247422e-01
8.19449484e-01 -1.38814092e-01 1.22780418e-02 5.51049292e-01
3.73364270e-01 1.19749177e+00 1.15294230e+00 -1.27766180e+00
-1.28394675e+00 -1.48233205e-01 -3.38175356e-01 -1.04813211e-01
2.16375533e-02 2.03400359e-01 -5.12912452e-01 -1.14182699e+00
1.49249926e-01 -5.91481805e-01 -1.99467450e-01 8.67860615e-01
-2.81454653e-01 -5.82955480e-01 1.05597603e+00 -5.24059474e-01
5.48311353e-01 -2.02655816e+00 -5.30982852e-01 2.36196622e-01
6.85377494e-02 1.61576763e-01 -3.30982506e-01 2.41079494e-01
9.68179405e-02 2.17004463e-01 -5.40032759e-02 -3.02992076e-01
-1.52415141e-01 -2.49757603e-01 -3.78251523e-01 3.89685154e-01
5.90896547e-01 8.51764798e-01 -1.01701891e+00 -9.21359241e-01
4.83410984e-01 1.49058595e-01 -1.43770576e-01 -5.15863709e-02
-1.43040791e-01 -1.32009789e-01 -1.86349288e-01 1.01298130e+00
8.42515171e-01 8.10671598e-02 -9.01199132e-02 -1.30913526e-01
-3.60136956e-01 -2.52212167e-01 -1.14866507e+00 1.51013005e+00
-7.14084655e-02 9.93330002e-01 1.18751377e-01 -7.14363992e-01
9.69987273e-01 6.32942095e-02 4.76709418e-02 -5.98393798e-01
-2.82623887e-01 7.96938874e-03 -4.92403388e-01 1.21330567e-01
4.86162364e-01 3.55704054e-02 -8.30437690e-02 4.39038962e-01
-4.86572273e-03 -4.34289038e-01 5.05652666e-01 2.28271961e-01
7.62433112e-01 3.43662500e-01 7.09432483e-01 -5.18790245e-01
3.52671087e-01 1.71271488e-01 1.05640173e-01 1.28869891e+00
-4.99688983e-01 7.34963894e-01 4.26280469e-01 -7.69456625e-01
-1.01648521e+00 -1.33709443e+00 -3.00997555e-01 1.64651597e+00
2.10897326e-02 -3.88978064e-01 -5.76127410e-01 -9.62308347e-01
4.28235114e-01 2.63301402e-01 -9.36686158e-01 4.49588746e-01
-7.72410691e-01 -3.33656818e-01 7.06299901e-01 1.20964396e+00
5.81478596e-01 -1.27303410e+00 -1.11868167e+00 1.50583088e-01
3.07550132e-01 -9.54715848e-01 -7.89548516e-01 2.29936436e-01
-9.89004791e-01 -1.07948673e+00 -8.82141650e-01 -9.54719782e-01
7.37717807e-01 6.11412108e-01 1.15448070e+00 7.44699016e-02
-1.25611150e+00 4.50462610e-01 -7.96165541e-02 -6.00627482e-01
-1.20887412e-02 -4.11335796e-01 -4.36526924e-01 -1.08512484e-01
9.02115405e-02 1.55850938e-02 -6.42928839e-01 2.31427222e-01
-4.05272663e-01 -1.35730386e-01 8.38615000e-01 1.01955640e+00
5.46740532e-01 -2.78211653e-01 3.41991961e-01 -8.14985812e-01
3.80712718e-01 4.50735092e-01 -8.30805421e-01 3.50260526e-01
-2.53087103e-01 1.93018556e-01 1.50454670e-01 -4.19536740e-01
-1.06289792e+00 5.84306479e-01 2.11718321e-01 -1.22873619e-01
-4.11131501e-01 -2.48908952e-01 3.26233327e-01 -4.73920405e-01
8.88837874e-01 6.27740473e-02 -2.89085478e-01 -4.39845502e-01
8.45243812e-01 3.07158977e-01 8.41413915e-01 -7.98971295e-01
7.86480010e-01 7.80132949e-01 4.45407853e-02 -1.13645339e+00
-8.24729681e-01 -8.84136081e-01 -1.06257391e+00 -3.27647835e-01
8.24437141e-01 -7.69568384e-01 -3.38271350e-01 4.77327675e-01
-1.15024507e+00 -1.84973791e-01 -5.14829755e-01 -1.33544847e-01
-6.94975674e-01 4.36922729e-01 -7.27713883e-01 -1.21829236e+00
-3.91513497e-01 -5.47357976e-01 1.57613456e+00 5.23866117e-01
1.74517557e-01 -7.87024975e-01 -1.60276189e-01 -1.86112285e-01
3.58088523e-01 6.31837666e-01 8.08371425e-01 -9.88841131e-02
-6.98035955e-01 -5.57052016e-01 -9.32366610e-01 8.89232382e-02
3.28515954e-02 2.50592202e-01 -1.19545710e+00 -4.40493435e-01
-8.46746922e-01 -5.50936818e-01 1.42421412e+00 3.83565277e-01
9.25018311e-01 1.07967682e-01 -7.87480474e-01 1.24970876e-01
1.32545745e+00 -8.42269287e-02 2.83633858e-01 -6.15741499e-02
3.33287716e-01 5.75042307e-01 9.24381375e-01 4.53345478e-01
-7.27132559e-02 6.39595747e-01 -1.54821470e-01 -6.56377494e-01
-6.02709591e-01 -4.27654147e-01 1.04190633e-01 -2.81524628e-01
-1.76592320e-01 2.67097265e-01 -8.36229444e-01 7.07605660e-01
-1.77729321e+00 -1.13009119e+00 3.58981490e-02 2.17725825e+00
8.43319237e-01 6.42759681e-01 5.02656400e-01 5.56064397e-02
9.18454587e-01 2.26910025e-01 -3.45420748e-01 -2.62447655e-01
-4.07096304e-05 5.46723008e-01 5.63536644e-01 3.39863956e-01
-1.77204919e+00 1.23713505e+00 7.27065992e+00 8.84359658e-01
-9.77882981e-01 -1.86712563e-01 6.90662920e-01 2.81563342e-01
3.67524415e-01 -7.75137171e-02 -7.61879385e-01 -1.97708115e-01
2.04918951e-01 1.78871807e-02 -5.27308397e-02 1.09354126e+00
-2.71458775e-01 -3.85008365e-01 -1.05944145e+00 1.00665641e+00
1.53482378e-01 -1.47913909e+00 -3.12227793e-02 -1.95797518e-01
7.15681791e-01 -4.73986156e-02 -1.08270644e-04 2.23782659e-01
4.11416322e-01 -7.99371183e-01 9.38519120e-01 5.17720103e-01
1.17225504e+00 -9.01245117e-01 3.40737730e-01 1.70814499e-01
-1.85150325e+00 1.46346986e-01 -6.36606038e-01 6.24170303e-02
-2.83919811e-01 3.31638217e-01 -9.04737294e-01 7.38074481e-02
6.09114885e-01 5.92007399e-01 -1.12260878e+00 1.27537107e+00
-3.10578674e-01 3.15126032e-01 -3.85763556e-01 -1.72951475e-01
2.05780938e-01 2.71488905e-01 2.37994015e-01 2.17042208e+00
-3.00254256e-01 -1.66164397e-03 6.03091419e-01 8.06068540e-01
-1.08038463e-01 3.35018933e-02 -8.18146646e-01 2.61415571e-01
4.42231089e-01 1.52905202e+00 -1.36047637e+00 -7.09241748e-01
-4.05004442e-01 1.00510728e+00 2.92964250e-01 2.21819714e-01
-4.10082519e-01 -9.36926067e-01 2.86140352e-01 1.15307838e-01
7.78538883e-01 -3.62042010e-01 -3.28505278e-01 -9.34343636e-01
-7.57889450e-02 -3.68578136e-01 5.27981400e-01 -5.14278531e-01
-1.38725007e+00 3.04148614e-01 3.21924384e-03 -1.00281644e+00
-2.32430592e-01 -1.00311434e+00 -1.05556858e+00 5.65584302e-01
-1.27690268e+00 -1.48199153e+00 -4.25419211e-01 3.42202544e-01
8.18731189e-01 3.74201220e-03 9.24696147e-01 -1.07455529e-01
1.38257265e-01 7.21070766e-01 -6.27649054e-02 6.83091283e-01
7.70936787e-01 -1.70449781e+00 1.19279623e+00 9.29502487e-01
7.09256709e-01 2.93553442e-01 2.28109866e-01 -6.80741549e-01
-8.60114455e-01 -9.84206438e-01 6.57683671e-01 -5.81041396e-01
2.46169955e-01 -9.06642258e-01 -4.76957411e-01 3.71694148e-01
2.94792443e-03 6.74780786e-01 1.10144347e-01 2.87564415e-02
-8.02269042e-01 1.27514154e-01 -1.35139573e+00 4.42238241e-01
1.07364571e+00 -6.05460286e-01 -5.19704342e-01 3.20846289e-01
4.35793176e-02 -1.75726891e-01 -4.35919225e-01 1.95961788e-01
1.03117108e+00 -1.01501703e+00 1.37063372e+00 -6.76285863e-01
-2.02324063e-01 -1.73291415e-01 -5.85416378e-03 -8.43227804e-01
-3.87373269e-01 -3.86332840e-01 1.79269791e-01 9.99748468e-01
1.95302293e-01 -2.78311431e-01 1.18872261e+00 3.58475000e-02
4.68104780e-01 -5.67622483e-01 -9.37080145e-01 -6.90401435e-01
1.87463030e-01 -2.40510494e-01 -1.03175253e-01 5.17367303e-01
-4.58466783e-02 2.38894477e-01 -2.28316849e-03 -1.45195559e-01
1.14536917e+00 6.67849898e-01 8.93046379e-01 -1.43913126e+00
8.58005434e-02 -7.18303025e-01 -7.02440560e-01 -1.17791295e+00
-3.25482965e-01 -5.15267253e-01 3.16451848e-01 -1.32395983e+00
3.01691711e-01 -3.65301400e-01 -1.69002146e-01 4.56777960e-01
-3.33255738e-01 6.77545071e-01 5.48688173e-01 -4.05742191e-02
-6.75434828e-01 9.96319926e-04 1.05387115e+00 -2.37522930e-01
2.03870982e-01 1.06255516e-01 -2.88819671e-01 9.56047177e-01
5.98890185e-01 -5.97543001e-01 1.50772333e-01 1.23724602e-01
-5.43237507e-01 2.52914518e-01 4.85868454e-01 -1.26312530e+00
3.97085130e-01 -1.37764126e-01 1.17634463e+00 -1.22920966e+00
4.71349061e-01 -1.45332888e-01 -1.02589238e+00 6.11007273e-01
-2.59514600e-01 6.66293353e-02 5.45534253e-01 8.88144195e-01
-3.95009108e-02 -6.45179488e-03 1.26825035e+00 -3.81873518e-01
-8.67494583e-01 3.71889137e-02 -4.85693723e-01 1.23706296e-01
9.47924078e-01 -2.04630569e-01 -3.45640093e-01 -3.72830063e-01
-8.27544570e-01 -1.21574514e-01 2.75201559e-01 4.36543137e-01
9.36402559e-01 -1.20569193e+00 -8.98719132e-01 4.91695553e-01
2.40180388e-01 -3.66595685e-01 -3.89128149e-01 1.82731450e-01
-5.44350684e-01 4.56891835e-01 -5.11029720e-01 -7.80894578e-01
-1.46787930e+00 6.14319742e-01 2.33872160e-01 -1.30897328e-01
-8.76731455e-01 1.11381614e+00 3.47197950e-01 4.01910357e-02
4.56124276e-01 -5.18843234e-01 1.04858771e-01 1.93191230e-01
7.89135039e-01 4.06280696e-01 -1.07991695e-01 -2.27352530e-01
-5.45480430e-01 8.55183959e-01 -2.23974049e-01 -3.31921607e-01
7.95086384e-01 3.39433551e-01 2.96974301e-01 1.79306552e-01
8.39490950e-01 3.19431454e-01 -1.70198298e+00 -2.62227356e-01
4.72683489e-01 -6.50700688e-01 -1.63959861e-01 -9.22086716e-01
-7.42268443e-01 1.16542935e+00 9.96359885e-01 3.37057896e-02
7.33056307e-01 1.06505178e-01 3.78925443e-01 6.66115463e-01
5.78652799e-01 -1.15058887e+00 5.84921420e-01 2.32818261e-01
9.69128132e-01 -1.40011454e+00 2.05810353e-01 -6.01019800e-01
-4.28906351e-01 1.54362810e+00 3.73348296e-01 -6.04053199e-01
5.06191075e-01 5.53881228e-01 1.77188873e-01 -2.34065741e-01
-5.36869943e-01 -5.88300586e-01 4.57905054e-01 9.16659355e-01
4.50740308e-01 2.51930296e-01 6.26768470e-02 7.30449259e-02
3.46950233e-01 -3.84591103e-01 3.70758235e-01 1.29345226e+00
-1.01576674e+00 -9.99746978e-01 -6.00945234e-01 4.84527379e-01
-1.91908911e-01 -3.73821296e-02 -7.69282937e-01 1.07397616e+00
7.67588289e-03 9.06077623e-01 1.61017746e-01 6.02707900e-02
3.65783602e-01 1.43122822e-01 8.66791964e-01 -7.44382918e-01
-3.96636784e-01 6.47986308e-02 1.44499496e-01 -7.45059669e-01
-1.71243742e-01 -9.18020368e-01 -1.10120940e+00 1.07129082e-01
-3.61125529e-01 -1.29174277e-01 5.85060239e-01 4.38789397e-01
1.03535563e-01 1.65461689e-01 2.49256596e-01 -1.44613409e+00
-5.86444974e-01 -9.13176298e-01 -4.48738694e-01 4.02988106e-01
2.76308179e-01 -8.01182687e-01 2.87566613e-02 3.53914827e-01] | [9.375835418701172, 0.2615113854408264] |
dd682367-4d01-47ec-8fa3-386e87ea3468 | selecting-optimal-context-sentences-for-event | null | null | https://www.aaai.org/AAAI22Papers/AAAI-3912.ManH.pdf | https://www.aaai.org/AAAI22Papers/AAAI-3912.ManH.pdf | Selecting Optimal Context Sentences for Event-Event Relation Extraction | Understanding events entails recognizing the structural and temporal orders between event mentions to build event structures/ graphs for input documents. To achieve this goal, our work addresses the problems of subevent relation extraction
(SRE) and temporal event relation extraction (TRE) that aim to predict subevent and temporal relations between two given event mentions/triggers in texts. Recent state-of-the-art methods for such problems have employed transformer-based language models (e.g., BERT) to induce effective contextual representations for input event mention pairs. However, a major limitation of existing transformer-based models for SRE and TRE is that they can only encode input texts of limited length (i.e., up to 512 sub-tokens in BERT), thus unable to effectively capture important context sentences that are farther away in the documents. In this work, we introduce a novel method to better model document-level context with important context sentences for event-event relation extraction. Our method seeks to identify the most important context sentences for a given entity mention pair in a document and pack them into shorter documents to be consumed entirely by transformer-based language models for representation learning. The REINFORCE algorithm is employed to train models where novel reward functions are presented to capture model performance, and context-based and knowledge-based similarity between sentences for our problem. Extensive experiments demonstrate the effectiveness of the proposed method with state-of-the-art performance on benchmark datasets. | ['and Thien Huu Nguyen', 'Linh Van Ngo', 'Nghia Ngo Trung', 'Hieu Man Duc Trong'] | 2022-04-02 | null | null | null | aaai-2022-4 | ['temporal-relation-extraction', 'temporal-relation-classification', 'relation-classification', 'event-relation-extraction'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 5.10235727e-01 2.64940262e-01 -2.90383250e-01 -5.13251901e-01
-1.08140862e+00 -4.03697848e-01 8.42726767e-01 9.87590671e-01
-5.00042796e-01 9.08573866e-01 6.20819509e-01 -3.94744724e-01
-4.62332487e-01 -1.06451738e+00 -5.97690225e-01 -2.76104152e-01
-5.44280648e-01 4.31049764e-01 4.21560258e-01 -1.86565682e-01
2.84259796e-01 5.62208891e-01 -1.33982825e+00 7.63494134e-01
6.96837246e-01 9.15878356e-01 3.65587234e-01 6.40801847e-01
-4.48891819e-01 1.27613533e+00 -7.65154541e-01 -1.80024356e-01
-2.41692036e-01 -5.85298598e-01 -1.15771937e+00 -2.68465310e-01
-4.55045283e-01 6.46026134e-02 -4.74667996e-01 6.00221634e-01
3.06858301e-01 6.05660737e-01 7.92353213e-01 -7.91320264e-01
-3.17089885e-01 1.16482770e+00 -4.32366222e-01 9.85442281e-01
7.00671554e-01 -4.91824985e-01 1.33066678e+00 -8.13714743e-01
5.80017567e-01 1.18198049e+00 3.52506369e-01 2.09429473e-01
-9.00267899e-01 -3.87794137e-01 5.58281958e-01 6.59806311e-01
-1.29267418e+00 -2.27957487e-01 8.91486883e-01 6.49015680e-02
1.81515551e+00 4.01980668e-01 5.00188053e-01 1.16433537e+00
4.72846895e-01 8.87093008e-01 8.50988209e-01 -6.98837757e-01
3.65042239e-01 -1.23036951e-01 4.07648385e-01 2.57733256e-01
-8.76532272e-02 5.29553778e-02 -7.40874112e-01 -2.51630098e-01
3.83947968e-01 3.84237245e-02 1.17006794e-01 5.76651156e-01
-9.49943125e-01 8.89424384e-01 3.08080614e-01 7.95882940e-01
-7.13094175e-01 -1.22729763e-01 6.54491186e-01 1.32074565e-01
4.74388152e-01 3.15971196e-01 -5.71007729e-01 -1.85780898e-01
-7.46845186e-01 2.65493959e-01 6.99646950e-01 9.90655303e-01
4.63853627e-01 -1.60837129e-01 -6.44813418e-01 6.70595288e-01
1.67686060e-01 -5.45007773e-02 6.46965086e-01 -2.60927659e-02
1.02758396e+00 7.14116096e-01 -2.27168635e-01 -1.08148432e+00
-4.79036897e-01 -2.65250355e-01 -5.57246745e-01 -5.90446591e-01
2.35658437e-02 -1.26715332e-01 -7.60647476e-01 1.71889007e+00
3.42373252e-01 7.49622703e-01 4.18715000e-01 4.22351837e-01
8.69352877e-01 1.13918281e+00 4.53500390e-01 -6.22946739e-01
1.69700289e+00 -5.66731870e-01 -8.11408520e-01 -6.88828945e-01
5.88506103e-01 -6.04998946e-01 8.00546110e-01 -6.09058775e-02
-1.11727488e+00 -2.44286433e-01 -9.71755266e-01 8.19699541e-02
-5.36906540e-01 -8.64390209e-02 6.16116047e-01 1.68842703e-01
-3.57322633e-01 5.42106569e-01 -8.87121916e-01 -3.18826258e-01
2.18959168e-01 2.67110497e-01 2.77204420e-02 1.45848304e-01
-1.81241488e+00 1.12169588e+00 8.67719948e-01 -2.77158488e-02
-8.32032144e-01 -5.14035463e-01 -9.13224459e-01 3.94727349e-01
6.68786407e-01 -2.55997986e-01 1.31942546e+00 -5.08796692e-01
-1.05739498e+00 4.86204147e-01 -4.57737535e-01 -9.43876803e-01
-2.71571398e-01 -2.85610735e-01 -8.02137852e-01 4.24249440e-01
-8.82911384e-02 8.91745836e-02 4.96737570e-01 -8.64931941e-01
-1.02941132e+00 -2.02219129e-01 1.93163216e-01 4.15250093e-01
-4.13505882e-01 6.63618922e-01 -2.21345708e-01 -8.61955047e-01
-5.74253164e-02 -3.45739812e-01 -4.42859828e-01 -1.10580075e+00
-3.80120277e-01 -6.70309842e-01 6.65913045e-01 -6.35037303e-01
1.77888334e+00 -1.76454604e+00 -8.57229754e-02 2.46622890e-01
-1.83077723e-01 8.10799003e-02 -4.31952663e-02 8.68750215e-01
-2.34623998e-01 -7.64870569e-02 -5.77784888e-02 -2.32850686e-01
-1.54042944e-01 5.13194442e-01 -8.22739422e-01 -5.77105433e-02
5.70360661e-01 8.93089831e-01 -1.06812525e+00 -8.56705487e-01
7.42069855e-02 4.33399677e-01 -1.87913805e-01 4.49991286e-01
-3.07812840e-01 2.49783397e-01 -8.00758421e-01 3.34102392e-01
6.09586053e-02 -1.03886686e-01 3.41699392e-01 -9.62193087e-02
1.07173130e-01 1.14189279e+00 -1.18327975e+00 1.35940135e+00
-6.99533284e-01 3.55686367e-01 -7.77686298e-01 -1.33003223e+00
1.02118194e+00 7.45350480e-01 4.90382433e-01 -8.42298985e-01
1.44854859e-01 -5.77304065e-02 -2.44651616e-01 -5.09651065e-01
7.44239748e-01 -3.22665632e-01 -4.39008653e-01 4.01962280e-01
1.23471553e-02 3.32332611e-01 6.53030574e-01 3.73542458e-01
1.29286861e+00 -1.91665471e-01 7.73081720e-01 1.22857645e-01
6.42608821e-01 -2.37319395e-01 6.67991996e-01 6.92631543e-01
2.17971772e-01 3.02102357e-01 4.74572361e-01 -3.66097778e-01
-7.20446527e-01 -9.08599555e-01 6.61995932e-02 1.19645655e+00
-8.02589059e-02 -7.69783735e-01 -4.00045335e-01 -8.47982883e-01
-4.78667259e-01 1.16060901e+00 -6.10862017e-01 -3.11311841e-01
-1.00632501e+00 -9.10680413e-01 6.20286942e-01 9.58947420e-01
1.20449446e-01 -1.41813529e+00 -8.56681347e-01 7.57102072e-01
-4.34129775e-01 -1.20606780e+00 -2.90432334e-01 6.92194760e-01
-9.18840110e-01 -9.34005976e-01 -9.97774079e-02 -8.94247174e-01
4.75490063e-01 -2.23629281e-01 1.26551092e+00 -4.60287541e-01
-1.89643964e-01 1.35941401e-01 -7.92226970e-01 -5.34965098e-01
-2.67563015e-01 5.43144718e-02 -2.32103080e-01 -6.08037747e-02
6.82827294e-01 -6.80394232e-01 -4.01687086e-01 1.84561938e-01
-1.19358122e+00 -1.07423052e-01 6.84853554e-01 8.52664113e-01
8.12152982e-01 4.69219804e-01 9.56909716e-01 -9.53738213e-01
8.44403267e-01 -7.57309318e-01 -2.07513988e-01 5.21043479e-01
-5.44035733e-01 1.50681332e-01 8.00592124e-01 -6.46159530e-01
-1.39928782e+00 -1.43416435e-01 -6.48037791e-02 6.78912923e-02
-2.46805742e-01 1.12937987e+00 7.64869526e-02 7.22947896e-01
7.17928886e-01 4.10323620e-01 -1.02746403e+00 -1.99054733e-01
2.03004152e-01 5.93398154e-01 4.62568223e-01 -8.93027782e-01
3.31702054e-01 3.31145898e-02 -7.70608336e-02 -4.45560694e-01
-1.14639509e+00 -6.18053675e-01 -3.45995814e-01 8.29836577e-02
6.56059325e-01 -7.09293664e-01 -4.45595235e-01 -3.23624820e-01
-1.23098433e+00 4.26088609e-02 -5.03487110e-01 6.50930643e-01
-4.30581421e-01 1.51139975e-01 -7.61256039e-01 -1.14603853e+00
-5.51299989e-01 -6.57612920e-01 9.37632203e-01 3.55739355e-01
-4.85184461e-01 -9.86036062e-01 1.11636601e-01 -1.52648553e-01
9.35642347e-02 3.38011801e-01 1.13581002e+00 -1.08271861e+00
-3.45512092e-01 -2.79129028e-01 9.20308754e-02 -1.84019670e-01
1.62045032e-01 -3.93211544e-01 -7.68543422e-01 -1.77839119e-02
7.47491121e-02 -2.73679048e-01 6.81162179e-01 1.92262471e-01
1.03297544e+00 -5.87666035e-01 -6.24036014e-01 -4.97269034e-02
1.21578634e+00 7.44885564e-01 6.45264447e-01 3.31030816e-01
2.37635016e-01 5.63994884e-01 1.02602756e+00 6.89748347e-01
4.02270377e-01 6.00875258e-01 -1.63544267e-02 2.03393936e-01
1.00894585e-01 -4.90002483e-01 3.48459274e-01 8.07062387e-01
-4.58620451e-02 -4.78497714e-01 -8.60376358e-01 8.54447722e-01
-1.89453042e+00 -1.11120665e+00 6.81340918e-02 1.96371388e+00
1.22564805e+00 5.84253907e-01 -8.26271623e-02 4.88806278e-01
5.85667670e-01 2.24509403e-01 -1.89188078e-01 -5.46674252e-01
-8.41742828e-02 4.96990263e-01 8.46576393e-02 3.60453784e-01
-1.06655622e+00 9.44648504e-01 5.51722574e+00 9.25890863e-01
-6.88450098e-01 -5.66074066e-02 6.12218440e-01 -7.97860399e-02
-2.24527374e-01 1.07211970e-01 -1.12543106e+00 2.82125413e-01
1.50623465e+00 -5.22243500e-01 1.64308906e-01 4.99862731e-01
3.36387306e-01 -1.23426870e-01 -1.42140532e+00 6.16645873e-01
-6.86039105e-02 -1.33164835e+00 8.88526589e-02 -3.49612623e-01
5.80376327e-01 -2.55197376e-01 -2.69138187e-01 5.09440839e-01
1.74356937e-01 -8.64834785e-01 5.89022517e-01 4.13927823e-01
4.02912229e-01 -1.05052316e+00 7.15471268e-01 3.83432209e-01
-1.75744116e+00 -1.56116918e-01 -3.30155462e-01 -1.11651540e-01
4.33235079e-01 7.43585289e-01 -1.26873171e+00 9.78039622e-01
5.19129872e-01 6.65529013e-01 -2.88483709e-01 7.04963505e-01
-4.45676714e-01 8.90818655e-01 -4.67206955e-01 -3.89567703e-01
3.06398690e-01 2.75970131e-01 5.06984591e-01 1.67459261e+00
2.28367344e-01 6.16548240e-01 2.26175353e-01 5.21013498e-01
-4.06782813e-02 2.41151273e-01 -5.60615242e-01 -9.26172063e-02
6.88476324e-01 9.54600394e-01 -9.62105274e-01 -4.95404214e-01
-5.11969864e-01 5.29180586e-01 5.30978739e-01 1.94696084e-01
-8.45000923e-01 -5.86673379e-01 7.03205392e-02 -1.25272259e-01
5.38538516e-01 -7.23251849e-02 -2.25489259e-01 -8.91105235e-01
1.31979123e-01 -7.16369331e-01 1.09035683e+00 -4.52589035e-01
-1.37301505e+00 8.74120355e-01 4.64732885e-01 -1.10206282e+00
-6.78452849e-01 -1.52671531e-01 -8.57655346e-01 7.46340454e-01
-1.50889528e+00 -9.53318477e-01 3.22061121e-01 7.25865960e-01
7.74786830e-01 -1.62664540e-02 9.98425841e-01 2.86361724e-01
-4.32365656e-01 4.71690774e-01 -2.21530423e-01 2.02972755e-01
2.66860604e-01 -1.24690616e+00 3.46761167e-01 1.00564504e+00
5.92240095e-01 5.99120378e-01 6.58776283e-01 -8.36869776e-01
-1.22667634e+00 -1.12988901e+00 1.60715556e+00 -2.80580014e-01
7.12160945e-01 -2.19049916e-01 -8.56248558e-01 8.67046297e-01
1.84322238e-01 -5.96347935e-02 7.55055785e-01 3.18968236e-01
-2.55103827e-01 -2.02356443e-01 -8.76410961e-01 5.77935457e-01
9.05637980e-01 -7.14734018e-01 -1.38180935e+00 3.86436880e-01
7.38790751e-01 -5.09013176e-01 -8.46919596e-01 3.54549378e-01
9.36923698e-02 -3.50578249e-01 1.14103973e+00 -9.25707757e-01
6.10736191e-01 -2.28523895e-01 -6.31326362e-02 -1.17226255e+00
-1.29604369e-01 -5.38291991e-01 -6.96866155e-01 1.51465917e+00
6.11870468e-01 -2.63453543e-01 4.79701310e-01 4.08260405e-01
-1.09407306e-01 -1.03906119e+00 -1.04184663e+00 -6.73676252e-01
-2.59136289e-01 -7.27696419e-01 6.41010940e-01 7.69707084e-01
5.87778389e-01 8.05435956e-01 -1.75897777e-01 4.26218063e-01
1.29161328e-01 4.00846690e-01 -1.30658373e-01 -8.97841215e-01
-3.64487529e-01 -1.67887926e-01 -1.53300583e-01 -8.60953331e-01
2.70961076e-01 -1.01145172e+00 2.06053376e-01 -1.64497721e+00
2.22340509e-01 -3.63276988e-01 -8.07991743e-01 5.97432137e-01
-4.66396987e-01 -4.32360649e-01 -1.02026887e-01 -1.80946276e-01
-6.78889394e-01 5.69911778e-01 7.47868061e-01 -2.11991221e-01
-4.67760146e-01 1.50464103e-01 -5.78157365e-01 5.54313660e-01
7.65493989e-01 -8.77034843e-01 -7.37511456e-01 -1.99356049e-01
2.80936331e-01 4.77032036e-01 1.29645556e-01 -7.35772073e-01
4.36726391e-01 -2.98908025e-01 3.68538171e-01 -9.00662005e-01
1.83925822e-01 -8.36823046e-01 -1.21069558e-01 1.78087413e-01
-8.83688569e-01 2.35911921e-01 2.96211123e-01 7.66090453e-01
-4.79332417e-01 -4.72931236e-01 2.07066104e-01 -3.14386152e-02
-6.23906434e-01 1.31191939e-01 -6.13016427e-01 3.77855927e-01
1.10052001e+00 2.27594003e-01 -1.37153313e-01 -2.27746040e-01
-7.32942462e-01 6.85726404e-02 -6.38101280e-01 3.90135020e-01
9.52051401e-01 -1.31177807e+00 -7.84464777e-01 -2.31813163e-01
1.72366396e-01 7.00151622e-02 1.59734100e-01 3.79915476e-01
1.30483463e-01 6.01466656e-01 1.85986549e-01 -3.30382288e-02
-1.42069125e+00 6.87275171e-01 -9.90161970e-02 -1.18776929e+00
-7.62382507e-01 8.23769093e-01 2.88363956e-02 1.48930952e-01
3.75367582e-01 -8.05259168e-01 -7.31302738e-01 1.51019931e-01
8.30862105e-01 6.59497529e-02 2.57663131e-01 -3.11512768e-01
-5.01227856e-01 7.09514916e-02 -4.00046110e-01 -1.50609776e-01
1.50007880e+00 5.34701918e-04 1.70191199e-01 3.69496167e-01
9.82416391e-01 -1.24652781e-01 -8.17973018e-01 -6.39220834e-01
5.66805005e-01 -2.41916195e-01 6.96617961e-02 -7.76057482e-01
-5.40178776e-01 6.70065582e-01 1.73150703e-01 3.77777576e-01
1.39195883e+00 3.48005891e-01 1.02259290e+00 6.01482034e-01
1.95775062e-01 -1.13541770e+00 1.06804632e-01 7.38865674e-01
8.16258252e-01 -7.62945712e-01 4.71818894e-02 -4.73634124e-01
-7.15252995e-01 1.05179060e+00 4.18491304e-01 7.46523738e-02
5.15395224e-01 4.77373064e-01 -4.69964415e-01 -1.30790770e-01
-1.04828107e+00 -3.52322221e-01 4.29531962e-01 3.38028848e-01
5.61770976e-01 1.22819021e-02 -5.80077887e-01 1.00871491e+00
-2.27987617e-02 -1.63080812e-01 1.85464531e-01 1.42729211e+00
-3.35993439e-01 -1.23838711e+00 -3.62951458e-01 5.08202970e-01
-7.41205037e-01 -2.85323828e-01 -1.67988837e-01 3.25637937e-01
-1.27638564e-01 1.26283026e+00 -2.23179981e-02 -1.83434635e-01
5.47834218e-01 2.65152544e-01 3.23204935e-01 -6.68029666e-01
-8.46111476e-01 4.87038568e-02 5.08076966e-01 -3.35215122e-01
-4.69277769e-01 -7.46727645e-01 -1.72546554e+00 1.56882003e-01
-2.43950754e-01 4.95071977e-01 2.39067316e-01 1.30020511e+00
2.52120852e-01 1.00335228e+00 7.59583950e-01 -3.76968265e-01
-3.09907287e-01 -1.01533210e+00 -3.85803103e-01 3.69926095e-01
3.01243048e-02 -5.49263477e-01 6.02913499e-02 2.42157161e-01] | [9.075793266296387, 9.16012954711914] |
00e46241-8931-4229-afb2-e61c53cbd0f6 | investigating-the-relation-between-chaos-and | 2008.12756 | null | https://arxiv.org/abs/2008.12756v2 | https://arxiv.org/pdf/2008.12756v2.pdf | Investigating the relation between chaos and the three body problem | We review the properties of fractals, the Mandelbrot set and how deterministic chaos ties to the picture. A detailed study on three body systems, one of the major applications of chaos theory was undertaken. Systems belonging to different families produced till date were studied and their properties were analysed. We then segregated them into three classes according to their properties. We suggest that such reviews be carried out in regular intervals of time as there are an infinite number of solutions for three body systems and some of them may prove to be useful in various domains apart from hierarchical systems. | ['Vishak Vikranth', 'T. S. Sachin Venkatesh'] | 2020-08-28 | null | null | null | null | ['physical-simulations'] | ['miscellaneous'] | [-3.14679921e-01 -1.09967068e-01 5.03230155e-01 1.17218927e-01
2.12636605e-01 -6.91969514e-01 9.23649371e-01 7.27434829e-02
-3.03981960e-01 9.57628608e-01 2.17253808e-02 -3.73865098e-01
-6.42185569e-01 -5.92053235e-01 1.52326033e-01 -1.14676034e+00
-6.48338139e-01 6.78581715e-01 6.11902535e-01 -8.39455903e-01
4.23440605e-01 6.96819961e-01 -1.88619137e+00 -1.85089752e-01
7.27716208e-01 5.63856065e-01 2.51130760e-02 8.12717974e-01
1.45003527e-01 6.85440421e-01 -7.21972108e-01 2.67107803e-02
1.87536821e-01 -6.55873895e-01 -7.99714267e-01 -2.33209855e-03
-7.04436004e-01 3.89928401e-01 -4.21937853e-01 7.52274990e-01
4.85914439e-01 7.58894905e-02 1.07705748e+00 -8.44406128e-01
-4.89007831e-01 2.66957849e-01 -2.45233789e-01 5.82128644e-01
2.37982050e-01 7.90264308e-02 3.04304183e-01 -6.64766133e-01
2.40295604e-01 1.02931583e+00 7.63817251e-01 4.04175222e-01
-1.10199630e+00 -1.99306846e-01 -6.55420959e-01 -3.58540118e-02
-1.39427149e+00 -5.83150208e-01 6.29543841e-01 -5.94943345e-01
9.63777065e-01 6.27637923e-01 8.89183879e-01 2.99111813e-01
1.05791461e+00 -3.00993510e-02 1.73794365e+00 -6.96240544e-01
4.18002009e-01 1.18906341e-01 3.25114727e-01 5.59848547e-01
6.61765039e-01 2.66654342e-01 -8.91542621e-03 -2.48218060e-01
6.89552963e-01 -2.48173043e-01 1.85340252e-02 -1.50239080e-01
-1.10588396e+00 5.52851558e-01 3.28986384e-02 1.19775438e+00
-1.21546850e-01 1.61545426e-01 3.57047617e-01 4.63178098e-01
2.27000564e-01 5.72418809e-01 -2.15662923e-02 -2.98230380e-01
-8.50477517e-01 4.22482133e-01 1.01486123e+00 6.73921466e-01
3.01073909e-01 6.94611743e-02 2.27243766e-01 4.55859661e-01
4.99814972e-02 4.84479100e-01 7.22478867e-01 -9.20787752e-01
-1.77861616e-01 4.42288816e-01 9.73813832e-02 -9.26661074e-01
-7.82706678e-01 -1.07715383e-01 -1.06155610e+00 3.46394211e-01
4.56541121e-01 -1.37396619e-01 -5.55082262e-01 1.01050234e+00
8.07213932e-02 -5.51530182e-01 1.96661174e-01 3.80797535e-01
8.49068940e-01 5.85227907e-01 -3.92739832e-01 -5.97093105e-01
1.40829384e+00 -3.72497499e-01 -9.48319316e-01 7.18509018e-01
4.69061941e-01 -8.77404809e-01 5.50799072e-01 4.43776548e-01
-1.07208657e+00 -3.86294544e-01 -9.76821482e-01 5.84917665e-01
-6.84761167e-01 -4.89233553e-01 6.86487734e-01 9.45059597e-01
-1.13726270e+00 8.40070188e-01 -8.63606453e-01 -5.75407803e-01
-2.48858258e-01 5.37302196e-01 5.10220863e-02 7.30580091e-01
-1.13425660e+00 1.05602062e+00 1.17064267e-01 2.14732453e-01
-1.77111536e-01 3.61528873e-01 -4.92299348e-01 -2.69334257e-01
-9.37714651e-02 -5.76888680e-01 8.75935018e-01 -1.69716448e-01
-1.27665865e+00 9.25165057e-01 -2.66144425e-01 -4.68980998e-01
5.11444867e-01 5.00887930e-01 -6.98404968e-01 2.55828351e-01
7.42174014e-02 -4.82176021e-02 7.43087173e-01 -1.24941242e+00
-2.46387929e-01 -3.04633062e-02 -1.28912941e-01 -2.53301203e-01
6.23317920e-02 2.32642785e-01 3.93683910e-01 -4.69768643e-01
2.03443810e-01 -1.02593374e+00 -4.37286377e-01 -8.80888820e-01
-1.37189135e-01 -5.98115206e-01 7.87079930e-01 -1.25097409e-01
1.48969769e+00 -1.71017110e+00 -6.70617744e-02 2.59990275e-01
5.03918469e-01 9.08764228e-02 5.98974288e-01 1.19278669e+00
1.06133625e-01 4.22719538e-01 -3.69010806e-01 7.99858570e-02
-8.01683813e-02 4.26165640e-01 -1.97666287e-01 7.62502611e-01
-1.61661327e-01 7.88839579e-01 -5.45598447e-01 -5.54110527e-01
3.22937280e-01 2.96218753e-01 3.32783605e-03 -2.78495908e-01
4.67384666e-01 6.85276508e-01 -3.91818821e-01 5.95140815e-01
4.61242348e-01 -2.44537383e-01 -3.16727668e-01 5.50558090e-01
-5.93093872e-01 1.44746928e-02 -8.92645717e-01 6.58171952e-01
5.44941202e-02 7.58351624e-01 -7.41450563e-02 -1.01923203e+00
1.08577704e+00 7.05653429e-01 8.10329378e-01 -4.50299501e-01
1.84070691e-01 4.85628605e-01 6.57596231e-01 -6.32393479e-01
5.89048207e-01 -4.98672277e-01 -3.63440484e-01 5.94048619e-01
-1.77229539e-01 -4.45302099e-01 5.45480132e-01 7.88722001e-03
1.03203559e+00 -3.98934335e-01 5.38393557e-01 -1.18301845e+00
8.61901045e-01 -4.24402170e-02 2.05806550e-02 8.24494481e-01
-5.02104700e-01 4.29665476e-01 3.23238909e-01 -8.71931791e-01
-9.98884797e-01 -6.92752481e-01 -6.97978139e-01 4.56160128e-01
4.27220374e-01 -5.74892223e-01 -7.14657307e-01 1.33266196e-01
-1.90325424e-01 -6.75765425e-03 -8.76593947e-01 -4.87453118e-02
-9.01232660e-01 -1.01636887e+00 2.36231118e-01 4.30628471e-02
3.46045643e-01 -1.60377228e+00 -1.19757080e+00 2.48350322e-01
3.53775591e-01 -6.18269801e-01 1.76398903e-01 5.34009576e-01
-9.32427704e-01 -1.09656179e+00 -8.37386608e-01 -6.37388527e-01
4.83791053e-01 1.84574798e-01 1.04783475e+00 5.96037865e-01
-3.22011471e-01 4.20426935e-01 -3.21525037e-01 -3.32678199e-01
-8.60335588e-01 1.67381167e-02 6.95192754e-01 -4.87097234e-01
3.13295573e-01 -8.15863550e-01 -5.48308849e-01 7.70541370e-01
-7.83515811e-01 -6.15918517e-01 9.64509547e-02 5.99612057e-01
6.38320148e-02 5.47450006e-01 7.38437772e-01 -5.73288321e-01
9.95780826e-01 -2.25420281e-01 -4.03553963e-01 -7.89348260e-02
-6.78091288e-01 6.48295432e-02 7.27722287e-01 1.48461964e-02
-5.18806279e-01 -2.45380506e-01 2.67215483e-02 2.83576906e-01
-1.65512398e-01 8.00905675e-02 4.44133788e-01 -5.86434543e-01
6.99624777e-01 5.24686337e-01 4.16516252e-02 -2.83122629e-01
-4.30691391e-02 8.53483498e-01 4.13263679e-01 -5.64544261e-01
8.11989903e-01 5.13998628e-01 4.52763200e-01 -1.34596217e+00
3.62873077e-02 -4.89274204e-01 -7.94458389e-01 -4.39298779e-01
6.06662929e-01 -2.64740914e-01 -9.58694041e-01 3.69921505e-01
-1.00461996e+00 -5.70825636e-02 -7.70960927e-01 1.98628366e-01
-1.01065624e+00 3.41627240e-01 -5.92443466e-01 -1.28894424e+00
-2.15886250e-01 -9.61577058e-01 6.30447030e-01 3.39391798e-01
-5.31912446e-01 -1.24765754e+00 5.16261220e-01 -3.84538472e-01
5.21159708e-01 4.58576292e-01 7.38976121e-01 -4.69901919e-01
1.56403761e-02 -2.04010874e-01 5.34970582e-01 -2.66952783e-01
4.43839222e-01 3.01569313e-01 -8.16242993e-01 -2.80332983e-01
6.31629467e-01 1.02475874e-01 8.67853045e-01 5.48761964e-01
5.40621161e-01 -1.76861838e-01 -5.89143693e-01 7.57991076e-02
1.29386914e+00 7.79000580e-01 5.04360557e-01 5.27683854e-01
-7.66295753e-03 8.42332602e-01 1.23276487e-01 4.23879296e-01
-3.31659731e-03 4.48240101e-01 1.31180525e-01 7.90017918e-02
1.16517529e-01 2.88946867e-01 3.64609733e-02 1.48993742e+00
-6.27397060e-01 -1.50094116e-02 -1.07635784e+00 5.49646020e-01
-1.67849886e+00 -1.20723796e+00 -6.30830765e-01 1.91685534e+00
3.58874321e-01 2.48877689e-01 7.51557052e-01 5.84102154e-01
8.00788760e-01 9.03294608e-03 -7.92034492e-02 -5.77023864e-01
-3.11975658e-01 -3.74666452e-02 2.09535256e-01 5.47682762e-01
-9.07351434e-01 6.91457689e-01 8.82250977e+00 5.32374561e-01
-8.49858284e-01 -9.70049128e-02 3.51546645e-01 2.09312707e-01
-3.79794240e-02 1.94389984e-01 -6.17208719e-01 7.28597820e-01
1.14093328e+00 -5.54970622e-01 2.19540104e-01 8.60452577e-02
3.53172332e-01 -5.01330793e-01 -5.18455684e-01 8.43441546e-01
-4.91294771e-01 -1.01945424e+00 -4.39710081e-01 4.78048295e-01
7.29169965e-01 -3.26450765e-01 9.06829238e-02 4.21935692e-02
1.89571083e-01 -1.23147500e+00 5.35792470e-01 7.49090970e-01
4.67347056e-01 -7.75926292e-01 6.22780979e-01 5.38786471e-01
-1.29483318e+00 -1.07534774e-01 -7.28693306e-01 -6.19931579e-01
3.09643686e-01 5.61234534e-01 -1.82722837e-01 4.61181968e-01
7.00206876e-01 5.10353208e-01 -6.64906502e-01 1.22560656e+00
4.21570122e-01 5.02250254e-01 -5.93176425e-01 -5.35608530e-01
1.79507300e-01 -4.94979382e-01 7.07347989e-01 8.45848203e-01
3.66104424e-01 3.91204327e-01 -3.82176727e-01 6.77273273e-01
8.42130780e-01 -1.01103134e-01 -9.00528669e-01 3.79410200e-02
3.26915324e-01 1.11652601e+00 -1.54204130e+00 -4.54598367e-01
-1.88871026e-02 4.85994607e-01 -2.91526139e-01 -2.00180174e-03
-5.18413723e-01 -6.56048536e-01 1.85768276e-01 3.66490185e-01
2.24063158e-01 -5.24049878e-01 -5.81932247e-01 -8.82004738e-01
-3.10025901e-01 -3.85960907e-01 3.60288881e-02 -3.12051952e-01
-1.12849092e+00 1.10728633e+00 3.32914025e-01 -1.36320114e+00
-5.46780884e-01 -4.23635691e-01 -7.96360433e-01 5.47129035e-01
-5.64049900e-01 -4.12423402e-01 1.92968503e-01 5.07864833e-01
3.35224360e-01 -4.24010575e-01 8.57827544e-01 -1.84317797e-01
-2.92560875e-01 5.43550327e-02 8.17886889e-01 -4.95339960e-01
1.31075934e-01 -1.36792076e+00 5.27646065e-01 3.43928009e-01
-4.52087857e-02 6.46810114e-01 1.29640400e+00 -5.09013116e-01
-8.70222986e-01 -2.41424009e-01 1.01387095e+00 -6.73974276e-01
7.23990679e-01 -4.05832112e-01 -7.56678700e-01 1.78296402e-01
5.45462906e-01 -2.55604833e-01 4.42980289e-01 -1.70526296e-01
6.46543443e-01 2.71558017e-01 -1.04513693e+00 4.91224408e-01
9.98213291e-01 -1.52250975e-01 -8.54055941e-01 3.40624988e-01
2.12871075e-01 4.56737243e-02 -8.90187085e-01 1.71164095e-01
8.22323322e-01 -1.46074963e+00 5.51633000e-01 -2.36735567e-01
1.25944957e-01 -3.19638699e-01 2.11221382e-01 -1.20199132e+00
-6.11102343e-01 -1.19727755e+00 2.75206212e-02 1.05196452e+00
1.87523626e-02 -1.24368978e+00 3.75779063e-01 3.53392690e-01
-1.20266620e-03 -8.50731492e-01 -1.20413685e+00 -1.20712125e+00
5.01509130e-01 2.68936634e-01 6.04512513e-01 6.39146447e-01
6.84461176e-01 2.05191627e-01 -2.28373855e-01 -6.55694902e-01
5.46161473e-01 3.06159258e-01 4.09127086e-01 -1.62008131e+00
-4.21421677e-02 -6.27751946e-01 -6.42172515e-01 -5.48017263e-01
-3.63786250e-01 -6.00192249e-01 -7.15415850e-02 -1.27057469e+00
-7.44841844e-02 -6.69277847e-01 -1.33346349e-01 -1.73317976e-02
1.25641286e-01 4.21705723e-01 1.42290145e-01 9.97913897e-01
-4.75047141e-01 3.54436189e-01 1.11989641e+00 1.87480524e-01
-4.07830238e-01 1.83503047e-01 -6.48256958e-01 7.58215725e-01
9.89443362e-01 -5.59680998e-01 -1.17410071e-01 4.67482865e-01
1.48336530e-01 1.96281299e-01 -9.73411426e-02 -1.51015139e+00
3.46941948e-01 5.92436306e-02 2.15056941e-01 -6.85623407e-01
2.14528203e-01 -9.56719697e-01 4.97457892e-01 9.20928240e-01
1.13643976e-02 4.30700451e-01 -1.63808698e-03 3.38605136e-01
-1.47363707e-01 -5.34401119e-01 9.31696951e-01 -4.30320472e-01
-2.85379529e-01 -1.87123805e-01 -1.08011997e+00 -1.95065618e-01
1.36513507e+00 -6.75741136e-01 -3.01796436e-01 -3.54780972e-01
-9.48402822e-01 -2.47291345e-02 5.58588803e-01 -2.12534983e-02
1.75554812e-01 -1.26507330e+00 -3.55157018e-01 1.25362113e-01
-1.10776387e-01 -3.53102058e-01 8.36786702e-02 1.09858942e+00
-8.43899250e-01 1.00825953e+00 -5.17773509e-01 -7.78436959e-01
-1.12663758e+00 7.62431920e-01 4.97376472e-01 -1.54386833e-01
-8.00843358e-01 2.45789200e-01 1.05905384e-01 -1.34084135e-01
-3.20853204e-01 -8.88265222e-02 -5.23072064e-01 -5.87458611e-02
3.72096062e-01 7.07464993e-01 -1.06660999e-01 -1.22752178e+00
-4.44402456e-01 9.36760008e-01 2.25963458e-01 -1.37481421e-01
1.25904942e+00 -3.89538437e-01 -5.41003942e-01 9.33942318e-01
7.70184338e-01 2.18339503e-01 -8.29903305e-01 3.25655818e-01
1.21262833e-01 -4.69632559e-02 -6.70796394e-01 -2.85461694e-01
-2.95696020e-01 7.07109511e-01 5.10197759e-01 1.46514606e+00
1.06232584e+00 2.88945735e-01 4.17738944e-01 1.42318025e-01
6.27945542e-01 -8.94148469e-01 -3.00394446e-01 6.53523326e-01
9.94814157e-01 -8.22119713e-01 7.75525868e-02 -2.54956067e-01
-1.67760298e-01 1.32591999e+00 1.24217840e-02 -8.42881918e-01
1.05889463e+00 5.19947171e-01 -1.89535260e-01 -3.50570947e-01
-8.44594777e-01 -9.45066512e-02 1.12420321e-01 7.55381286e-01
6.00457788e-01 1.85731500e-01 -1.08865440e+00 4.39667851e-01
-6.83450699e-01 -3.04502666e-01 8.68819118e-01 1.07357180e+00
-1.01915944e+00 -9.38383341e-01 -6.93846464e-01 4.43658888e-01
-4.67480332e-01 2.55823225e-01 -4.28951889e-01 1.07175040e+00
8.02420750e-02 1.10440373e+00 -1.35313973e-01 -4.26465958e-01
1.86101466e-01 -1.96088534e-02 6.87225580e-01 -7.38229975e-02
-8.23568463e-01 4.78789546e-02 -2.02424571e-01 -1.49006369e-02
-6.18352473e-01 -7.65934169e-01 -1.16408837e+00 -6.56835377e-01
-3.20868105e-01 9.00915742e-01 2.68699467e-01 9.68209743e-01
-1.31179214e-01 3.55753571e-01 7.08257079e-01 -9.89006400e-01
-2.81453341e-01 -9.70632792e-01 -1.29853070e+00 -1.61949709e-01
5.29900491e-01 -7.63318002e-01 -5.49429536e-01 1.99198723e-04] | [5.728281021118164, 4.144344329833984] |
f81bb261-5736-4aa2-875d-784109c8a5f2 | few-shot-learning-with-noisy-labels | 2204.05494 | null | https://arxiv.org/abs/2204.05494v2 | https://arxiv.org/pdf/2204.05494v2.pdf | Few-shot Learning with Noisy Labels | Few-shot learning (FSL) methods typically assume clean support sets with accurately labeled samples when training on novel classes. This assumption can often be unrealistic: support sets, no matter how small, can still include mislabeled samples. Robustness to label noise is therefore essential for FSL methods to be practical, but this problem surprisingly remains largely unexplored. To address mislabeled samples in FSL settings, we make several technical contributions. (1) We offer simple, yet effective, feature aggregation methods, improving the prototypes used by ProtoNet, a popular FSL technique. (2) We describe a novel Transformer model for Noisy Few-Shot Learning (TraNFS). TraNFS leverages a transformer's attention mechanism to weigh mislabeled versus correct samples. (3) Finally, we extensively test these methods on noisy versions of MiniImageNet and TieredImageNet. Our results show that TraNFS is on-par with leading FSL methods on clean support sets, yet outperforms them, by far, in the presence of label noise. | ['Tal Hassner', 'Vladan Petrovic', 'Samrudhdhi B. Rangrej', 'Kevin J Liang'] | 2022-04-12 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Liang_Few-Shot_Learning_With_Noisy_Labels_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Liang_Few-Shot_Learning_With_Noisy_Labels_CVPR_2022_paper.pdf | cvpr-2022-1 | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 3.15521896e-01 1.34532571e-01 -1.87517807e-01 -6.36329234e-01
-1.25933921e+00 -4.26105648e-01 7.25930989e-01 1.33798663e-02
-2.99018055e-01 7.31205106e-01 1.94260329e-01 1.16432279e-01
2.91575231e-02 -5.56325138e-01 -7.67157733e-01 -5.56084633e-01
4.26848419e-02 5.33465028e-01 6.32649124e-01 -1.65978417e-01
3.12137716e-02 3.05725902e-01 -2.03164220e+00 4.35793817e-01
6.75336301e-01 1.01432526e+00 -5.54593243e-02 4.77754205e-01
-3.59818786e-01 1.16332424e+00 -6.18757844e-01 -3.73703212e-01
4.12577093e-01 -4.26373452e-01 -8.86548400e-01 3.97143066e-01
9.47495878e-01 -2.09310845e-01 -1.95220515e-01 1.19053805e+00
4.81805086e-01 4.94591624e-01 7.75890946e-01 -1.24728727e+00
-6.58252418e-01 7.94926465e-01 -2.91927248e-01 3.33349168e-01
5.15412055e-02 2.70995051e-01 1.14482665e+00 -1.43607724e+00
9.12600219e-01 1.21354938e+00 9.06188726e-01 7.85403550e-01
-1.27151787e+00 -6.21618092e-01 2.94487834e-01 3.92929584e-01
-1.10663545e+00 -1.01250207e+00 3.56589615e-01 -3.79722625e-01
9.58369076e-01 1.77129567e-01 4.42162395e-01 1.29043281e+00
-2.22610906e-01 1.10085571e+00 1.41180551e+00 -5.37114680e-01
7.19300389e-01 2.04673871e-01 7.66811252e-01 6.16327584e-01
3.19599718e-01 3.64720374e-02 -7.00607181e-01 -2.78210729e-01
1.22601297e-02 6.79131448e-02 -2.61896193e-01 -4.19664234e-01
-8.82959604e-01 8.67893457e-01 2.49052286e-01 2.72223145e-01
-8.65718201e-02 6.68430477e-02 5.87154329e-01 5.33677518e-01
7.17704654e-01 4.48906064e-01 -1.32901385e-01 -2.25477010e-01
-1.17695868e+00 1.40635550e-01 8.77241015e-01 1.20782697e+00
7.36386061e-01 1.16459511e-01 -4.39396471e-01 1.01731586e+00
-1.58997118e-01 3.55660111e-01 5.94039381e-01 -1.00713837e+00
-1.19272716e-01 1.09028853e-02 1.02994867e-01 -5.48400879e-01
-2.69640326e-01 -4.53895599e-01 -5.15208125e-01 1.67216137e-01
1.34866729e-01 1.97688684e-01 -1.33532298e+00 1.48139286e+00
2.16548666e-01 5.54177582e-01 -1.28342614e-01 7.35369861e-01
9.94868994e-01 2.70812780e-01 4.84460928e-02 -3.65359038e-01
1.05302811e+00 -1.02430069e+00 -7.05842555e-01 -2.83659130e-01
6.33465052e-01 -6.59961104e-01 1.26136351e+00 4.59679574e-01
-9.59452629e-01 -2.77302146e-01 -1.07650220e+00 2.11355425e-02
-5.06402493e-01 -3.07631463e-01 6.54911339e-01 7.67160833e-01
-9.90236402e-01 9.77619112e-01 -7.97636330e-01 -6.40600383e-01
8.69133711e-01 -4.15236913e-02 -2.63970315e-01 -5.92891455e-01
-1.02862668e+00 9.37156796e-01 1.67259827e-01 -3.99923891e-01
-1.01309478e+00 -7.33866572e-01 -1.13212264e+00 1.48858488e-01
7.00293422e-01 -3.06073606e-01 1.76208639e+00 -5.61915040e-01
-1.14054453e+00 7.54988313e-01 -3.26665014e-01 -6.85478270e-01
4.63703364e-01 -5.92286848e-02 -3.70288789e-01 1.60890505e-01
4.45203036e-01 5.36047041e-01 1.01530898e+00 -1.31310487e+00
-6.65080726e-01 -1.38616160e-01 -2.16250077e-01 2.75904499e-02
-3.43908489e-01 -1.20837346e-01 -2.09826961e-01 -6.11854553e-01
5.74891344e-02 -6.22846127e-01 -2.18699366e-01 5.09615354e-02
-4.07357633e-01 -3.59319925e-01 7.13846743e-01 -5.53490734e-03
9.55994666e-01 -2.16036344e+00 -3.97488892e-01 1.94092140e-01
4.15101975e-01 4.31947201e-01 -1.82449266e-01 2.22055689e-01
-5.86112496e-03 3.29570509e-02 -2.44986743e-01 -6.66104853e-01
1.86753258e-01 4.87506539e-01 -6.60918117e-01 5.14226973e-01
3.23943555e-01 9.03867185e-01 -1.33434343e+00 -5.11320889e-01
3.05745780e-01 6.75541759e-02 -2.52974838e-01 -2.45888472e-01
-2.19593659e-01 -1.17967598e-01 -1.08925089e-01 9.67045605e-01
5.37636817e-01 -4.86977041e-01 -1.97305486e-01 -2.42259726e-01
4.50530276e-02 6.22270815e-03 -1.10335004e+00 1.37388813e+00
-1.17013939e-01 3.98868054e-01 -2.21629500e-01 -9.32739675e-01
6.14676178e-01 1.86974794e-01 4.20653373e-01 -4.81385291e-01
1.67324081e-01 2.78894067e-01 -3.71242017e-01 -4.75886583e-01
2.51711100e-01 -6.27114654e-01 2.66373288e-02 6.04350090e-01
6.90644026e-01 -1.57660231e-01 5.25063336e-01 5.02016187e-01
1.52895486e+00 -2.15197504e-01 3.26208502e-01 -5.63825443e-02
-2.25965828e-01 8.79748687e-02 7.76331902e-01 1.51100242e+00
-6.21846497e-01 8.09185505e-01 1.60238296e-01 -3.81657869e-01
-1.04436088e+00 -1.04326653e+00 -2.85370409e-01 1.28107238e+00
3.92856859e-02 -4.56575572e-01 -6.55525982e-01 -1.01206505e+00
7.67499283e-02 8.36694121e-01 -6.51118636e-01 -2.43038639e-01
-3.92965861e-02 -7.91211128e-01 4.97201562e-01 5.31664729e-01
2.23554060e-01 -9.38397825e-01 -4.51077789e-01 3.43999892e-01
7.36785233e-02 -1.13864112e+00 -3.75268012e-01 6.36773169e-01
-6.21683240e-01 -1.27569222e+00 -5.50748646e-01 -5.41189551e-01
5.13179958e-01 6.66245520e-01 1.22033405e+00 7.14794397e-02
-6.01416230e-01 4.36080873e-01 -7.93803573e-01 -5.91223121e-01
-2.15306044e-01 -1.47796839e-01 2.91916817e-01 -3.85419726e-02
7.41984367e-01 -4.23460305e-01 -1.87069908e-01 2.40876809e-01
-8.70429158e-01 -4.32090789e-01 3.53048205e-01 1.24678111e+00
7.47423112e-01 -1.69965491e-01 7.23202229e-01 -1.41664028e+00
5.11808991e-01 -5.45749545e-01 -2.23616496e-01 3.54445249e-01
-7.27339923e-01 -1.53713701e-02 5.96309066e-01 -3.71832997e-01
-8.59154105e-01 -9.34091434e-02 -2.15552449e-01 -7.53007412e-01
-2.45489553e-01 1.86538056e-01 1.64175898e-01 -3.39076698e-01
1.21130538e+00 1.30201563e-01 -2.08211386e-05 -6.74072862e-01
4.19485420e-01 6.82808876e-01 6.06736720e-01 -4.98871505e-01
5.66959918e-01 8.51992130e-01 -2.31928915e-01 -9.53737795e-01
-1.55925846e+00 -8.43570054e-01 -4.70616370e-01 -6.45803139e-02
3.38786066e-01 -8.63575757e-01 -9.83645767e-02 3.53733957e-01
-6.29414439e-01 -2.81537920e-01 -9.90101159e-01 3.02981615e-01
-6.98463202e-01 5.09658575e-01 -7.68778443e-01 -7.29935050e-01
-3.37810993e-01 -9.78648186e-01 1.08363283e+00 -3.21541950e-02
-2.45826036e-01 -6.05435491e-01 -3.72021459e-02 7.02548996e-02
3.66431117e-01 9.70276371e-02 6.60904944e-01 -1.16172767e+00
-3.22612286e-01 -3.28672111e-01 -1.58803225e-01 4.18313742e-01
3.53453867e-03 -8.07655603e-02 -1.42808557e+00 -4.51027930e-01
1.19305812e-01 -9.08049107e-01 1.37496197e+00 3.40817422e-01
1.08802402e+00 -7.15309475e-03 -3.27375472e-01 4.15277630e-01
1.11603284e+00 -2.40447059e-01 3.20848435e-01 1.22357570e-01
3.80965710e-01 2.51884490e-01 8.28732193e-01 6.18181348e-01
7.98813477e-02 3.97356927e-01 1.98978424e-01 1.16215623e-03
-5.19750535e-01 -6.31993338e-02 2.49663621e-01 8.54639649e-01
2.81828552e-01 -7.24674761e-02 -5.76301754e-01 6.02344453e-01
-1.87791562e+00 -1.33269858e+00 5.73026948e-02 2.06552100e+00
8.08733284e-01 3.68343651e-01 3.59338406e-03 2.85736918e-01
6.14462674e-01 6.65954202e-02 -8.86276245e-01 -1.02510685e-02
-2.84454316e-01 4.52820122e-01 5.28437734e-01 3.07632089e-01
-1.34876621e+00 1.04642010e+00 6.88360357e+00 1.17750752e+00
-7.17233360e-01 6.84524775e-01 5.87549031e-01 -5.07905364e-01
-2.11018503e-01 -1.13725677e-01 -9.49409604e-01 5.16872823e-01
7.86360085e-01 -2.70953000e-01 2.89506674e-01 1.04860044e+00
-1.32775858e-01 -1.07198499e-01 -1.13526249e+00 1.08532524e+00
4.12019849e-01 -1.47703004e+00 -2.70244658e-01 -2.82555163e-01
9.10495162e-01 5.46298385e-01 4.99916449e-02 8.00739765e-01
7.68410742e-01 -8.72990310e-01 6.66727960e-01 3.62332106e-01
9.02517915e-01 -6.67111158e-01 6.99551880e-01 3.50590318e-01
-8.31968188e-01 -2.27316692e-01 -7.50869155e-01 1.40087783e-01
1.39672130e-01 1.04763544e+00 -7.91093886e-01 9.36162323e-02
8.07377100e-01 1.03246808e+00 -6.53302073e-01 1.32073510e+00
-2.22885132e-01 7.74821222e-01 -4.43399668e-01 1.09209634e-01
3.56324822e-01 3.32112700e-01 4.90114361e-01 1.19841123e+00
2.10176036e-01 1.41552329e-01 3.85417521e-01 6.62013710e-01
-2.68384874e-01 -1.55223589e-02 -8.01488221e-01 2.26744264e-02
7.56405234e-01 1.14112365e+00 -9.98488724e-01 -8.33942354e-01
-5.37496448e-01 8.63368630e-01 3.83335501e-01 2.75318265e-01
-4.05398071e-01 -3.39231819e-01 6.49992645e-01 -1.50502631e-02
4.15289938e-01 3.58962625e-01 -1.97498173e-01 -1.47978878e+00
-4.88418061e-03 -8.05659592e-01 4.92814898e-01 -4.86169636e-01
-1.87160850e+00 5.80525398e-01 -2.48958319e-01 -1.29211974e+00
-6.36086473e-03 -5.47591984e-01 -5.35346746e-01 2.88549781e-01
-1.48584330e+00 -9.12338555e-01 -2.45988533e-01 4.24294531e-01
9.73996699e-01 -1.91463321e-01 8.56015086e-01 2.85820991e-01
-3.22290391e-01 6.73744202e-01 4.89697516e-01 -1.28676608e-01
9.81937170e-01 -1.35889030e+00 6.87531471e-01 8.98807049e-01
3.99285793e-01 4.77122188e-01 7.74547517e-01 -8.54965627e-01
-1.02014196e+00 -1.39802980e+00 7.98916459e-01 -5.36430240e-01
6.52799129e-01 -4.48195070e-01 -1.01212060e+00 7.01923072e-01
-2.35884920e-01 6.57288730e-01 8.80690813e-01 1.75542340e-01
-6.05365634e-01 2.71757483e-01 -1.44374919e+00 4.09682155e-01
1.30808532e+00 -4.87265974e-01 -7.27178633e-01 7.04318881e-01
7.45832384e-01 -1.17959023e-01 -6.53572083e-01 3.98166001e-01
2.28223622e-01 -1.15222645e+00 8.57305110e-01 -7.38346875e-01
1.52893230e-01 -1.49103403e-01 -3.36862057e-01 -1.54679620e+00
-4.45699483e-01 -4.61858869e-01 -3.86407614e-01 9.48841453e-01
1.48902312e-01 -3.70831817e-01 7.71312594e-01 5.96495092e-01
-3.40826273e-01 -8.01176131e-01 -1.04603851e+00 -1.26498139e+00
-2.08397403e-01 -7.55035162e-01 4.59563851e-01 1.01409888e+00
1.35868266e-01 2.09364533e-01 -5.06159186e-01 -3.79784554e-01
1.00949979e+00 -5.54822423e-02 4.58961636e-01 -1.44717216e+00
-3.20991635e-01 -2.36595795e-01 -4.67525512e-01 -5.66173077e-01
2.29922533e-01 -1.01388204e+00 4.38739210e-01 -1.38679242e+00
2.84052044e-01 -5.54944932e-01 -4.06650066e-01 8.54645729e-01
-1.20772325e-01 5.93092263e-01 2.79867411e-01 3.04519892e-01
-1.07418060e+00 5.94893813e-01 8.67137849e-01 -2.38881126e-01
2.56388903e-01 -3.21262404e-02 -6.09008372e-01 7.76312053e-01
5.07846951e-01 -5.62899470e-01 -4.47568506e-01 9.62718204e-02
-6.42102137e-02 -6.38682246e-01 1.97548628e-01 -1.14323616e+00
2.19903514e-01 3.35676931e-02 3.35575163e-01 -4.29691017e-01
3.97363782e-01 -6.01008952e-01 -2.79640496e-01 2.53365636e-01
-4.17704880e-01 -6.40825808e-01 -2.20608979e-01 8.61812174e-01
-8.22213441e-02 -6.06042445e-01 1.05914223e+00 -3.82976919e-01
-1.07751000e+00 4.64615226e-01 -3.79794329e-01 4.63091582e-01
9.61296797e-01 -2.83343554e-01 -4.79449928e-01 -2.00702310e-01
-9.75547016e-01 2.24681526e-01 6.38791382e-01 2.88463473e-01
8.14318776e-01 -1.34629071e+00 -4.78032410e-01 3.27617645e-01
6.19423926e-01 -4.80038002e-02 2.26838887e-01 9.04028416e-01
-7.63752609e-02 1.10727260e-02 1.91025987e-01 -3.89560699e-01
-1.19920969e+00 7.40370393e-01 1.04656003e-01 1.34251162e-01
-8.78598332e-01 1.10391092e+00 -1.75224409e-01 -3.70638072e-01
4.98771727e-01 -1.39362529e-01 4.59523872e-02 2.73798376e-01
9.87924635e-01 4.75665331e-01 4.14524078e-01 -4.20522869e-01
-1.66915566e-01 1.37550205e-01 -5.33361018e-01 9.53857303e-02
1.53168046e+00 -2.90793497e-02 2.56587178e-01 7.71266520e-01
1.00351250e+00 -4.04131740e-01 -1.14659512e+00 -7.49108672e-01
1.28610268e-01 -6.26240253e-01 5.86877093e-02 -7.28807747e-01
-7.63457417e-01 8.39188099e-01 5.60658276e-01 1.24454029e-01
6.65986717e-01 1.61894307e-01 9.12628770e-01 6.91354990e-01
7.31375515e-01 -1.39689612e+00 1.20313913e-01 7.22650766e-01
2.99646050e-01 -1.54629862e+00 -1.10370189e-01 -3.83726746e-01
-5.82079232e-01 9.39175129e-01 4.53215331e-01 -1.02953427e-01
8.19948971e-01 3.05777282e-01 -8.36951006e-03 -4.49154198e-01
-9.68843579e-01 -5.27454674e-01 -2.11940054e-02 7.17605352e-01
4.03365828e-02 -9.73711386e-02 1.38106167e-01 5.47741294e-01
6.05738088e-02 2.31410220e-01 6.55349076e-01 1.28831947e+00
-9.73051488e-01 -8.23599815e-01 -3.84753287e-01 1.07784390e+00
-3.07600558e-01 -2.20666677e-01 -2.22319201e-01 3.94224912e-01
2.02261746e-01 1.06202149e+00 -3.87181714e-02 -3.07537615e-01
3.15683931e-01 3.53981644e-01 4.60522920e-01 -1.17091811e+00
-3.89025420e-01 9.79194138e-03 3.13305692e-03 -6.56814575e-01
-3.36166888e-01 -6.66398108e-01 -9.88886058e-01 -1.42146721e-01
-5.13521254e-01 5.54261655e-02 3.49415720e-01 1.10273027e+00
3.26590121e-01 3.12079161e-01 5.95391095e-01 -9.15809333e-01
-1.03798759e+00 -9.32573080e-01 -9.99641180e-01 6.73287809e-01
2.84769446e-01 -9.79078591e-01 -6.19833887e-01 -4.04405408e-02] | [9.854036331176758, 2.828157901763916] |
2616a886-2ff6-40f0-8fc6-a99dfbbcee85 | learning-to-execute-programs-with-instruction | 2010.12621 | null | https://arxiv.org/abs/2010.12621v1 | https://arxiv.org/pdf/2010.12621v1.pdf | Learning to Execute Programs with Instruction Pointer Attention Graph Neural Networks | Graph neural networks (GNNs) have emerged as a powerful tool for learning software engineering tasks including code completion, bug finding, and program repair. They benefit from leveraging program structure like control flow graphs, but they are not well-suited to tasks like program execution that require far more sequential reasoning steps than number of GNN propagation steps. Recurrent neural networks (RNNs), on the other hand, are well-suited to long sequential chains of reasoning, but they do not naturally incorporate program structure and generally perform worse on the above tasks. Our aim is to achieve the best of both worlds, and we do so by introducing a novel GNN architecture, the Instruction Pointer Attention Graph Neural Networks (IPA-GNN), which achieves improved systematic generalization on the task of learning to execute programs using control flow graphs. The model arises by considering RNNs operating on program traces with branch decisions as latent variables. The IPA-GNN can be seen either as a continuous relaxation of the RNN model or as a GNN variant more tailored to execution. To test the models, we propose evaluating systematic generalization on learning to execute using control flow graphs, which tests sequential reasoning and use of program structure. More practically, we evaluate these models on the task of learning to execute partial programs, as might arise if using the model as a heuristic function in program synthesis. Results show that the IPA-GNN outperforms a variety of RNN and GNN baselines on both tasks. | ['Daniel Tarlow', 'Hugo Larochelle', 'Charles Sutton', 'David Bieber'] | 2020-10-23 | null | http://proceedings.neurips.cc/paper/2020/hash/62326dc7c4f7b849d6f013ba46489d6c-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/62326dc7c4f7b849d6f013ba46489d6c-Paper.pdf | neurips-2020-12 | ['program-repair', 'learning-to-execute', 'systematic-generalization', 'program-repair'] | ['computer-code', 'computer-code', 'reasoning', 'reasoning'] | [ 2.94089943e-01 4.01330292e-01 -5.48470736e-01 -1.62503928e-01
-3.96393687e-01 -3.23310465e-01 2.74676025e-01 4.14846018e-02
-1.73720028e-02 1.12856776e-01 3.09633344e-01 -1.20563900e+00
-1.12599675e-02 -1.00076640e+00 -1.10596836e+00 -2.16559008e-01
-3.33104849e-01 2.25406408e-01 2.17992276e-01 -3.78261179e-01
3.33538294e-01 2.44294465e-01 -1.27250004e+00 2.98257977e-01
6.01242125e-01 3.88924956e-01 5.00184484e-02 9.13888454e-01
-1.38920739e-01 1.51043010e+00 -5.86949825e-01 -2.62822300e-01
2.15525441e-02 -5.23419261e-01 -9.53277290e-01 -4.02843654e-01
3.86097163e-01 -1.31764248e-01 -4.24268007e-01 1.16187966e+00
1.08121790e-01 2.64355123e-01 2.75581926e-01 -1.19575059e+00
-1.00441587e+00 1.32332540e+00 -4.57674444e-01 2.98540682e-01
3.53596598e-01 4.24612671e-01 1.29443085e+00 -2.22810641e-01
4.86480922e-01 1.44375944e+00 9.90514398e-01 8.13717842e-01
-1.40763319e+00 -2.58689702e-01 2.63256818e-01 5.33854254e-02
-6.68248415e-01 -5.89492954e-02 6.89623475e-01 -4.57187384e-01
1.67136252e+00 2.62282584e-02 4.18489039e-01 1.28411078e+00
4.13935035e-01 9.96407330e-01 5.83610475e-01 -4.92908418e-01
1.26134470e-01 -4.39089894e-01 6.76093578e-01 1.09514654e+00
2.06866562e-01 4.11314011e-01 -7.23982751e-02 -2.67939955e-01
5.03879607e-01 1.52147248e-01 -4.87490982e-01 -2.18755215e-01
-9.82458472e-01 1.02297533e+00 8.51074040e-01 3.09257239e-01
-1.57826647e-01 9.24765825e-01 8.99441421e-01 7.08989501e-01
1.52770445e-01 9.62186694e-01 -4.48539972e-01 -2.07178399e-01
-8.96562696e-01 3.05622518e-01 1.03123701e+00 8.58525574e-01
6.68058038e-01 7.37980485e-01 -1.94996461e-01 2.84030467e-01
2.12819442e-01 2.03468606e-01 8.00525546e-01 -6.50296926e-01
5.99514842e-01 7.83150315e-01 -6.21098936e-01 -8.03210914e-01
-6.34780407e-01 -3.34094167e-01 -5.70126951e-01 3.45465213e-01
3.01318824e-01 -1.97956800e-01 -1.02515435e+00 1.97449887e+00
-4.49093401e-01 1.50287330e-01 9.54065472e-02 5.33022702e-01
7.04270899e-01 6.83297396e-01 -6.59773722e-02 1.80080384e-01
9.61466551e-01 -1.22023702e+00 -4.19986159e-01 -5.71743727e-01
1.23638535e+00 9.18841641e-03 1.28765500e+00 4.43784863e-01
-1.03697348e+00 -6.22023582e-01 -1.26824176e+00 -4.07706574e-02
-4.74465519e-01 -2.93007493e-01 1.11594987e+00 8.55057299e-01
-1.63301504e+00 9.33119416e-01 -1.22395933e+00 -1.30215719e-01
2.88444668e-01 5.21361649e-01 -6.54969588e-02 -1.00143015e-01
-8.91044855e-01 7.41863430e-01 4.44146961e-01 2.89283425e-01
-1.15552878e+00 -5.01363337e-01 -1.32704759e+00 4.57095325e-01
6.22644365e-01 -6.97696269e-01 1.45227385e+00 -1.19436836e+00
-1.44502854e+00 5.51117361e-01 4.92695011e-02 -8.14187229e-01
-6.44422695e-02 -1.55795395e-01 -1.59750402e-01 -4.77869928e-01
-5.32272682e-02 2.07408175e-01 6.40005767e-01 -8.74369919e-01
-4.13451008e-02 -2.83557802e-01 4.40500885e-01 -4.59910184e-01
1.43782049e-01 -3.15817958e-03 7.12013990e-02 -4.04189557e-01
-3.36940765e-01 -1.09350026e+00 -4.55326796e-01 -7.85095096e-01
-5.56494176e-01 -3.93360227e-01 8.04990888e-01 -7.56430864e-01
1.41155756e+00 -2.04501891e+00 4.65838253e-01 1.85474336e-01
4.97884005e-01 3.28636974e-01 -3.81209671e-01 2.25717157e-01
-3.78785729e-01 4.45094973e-01 -2.16856912e-01 3.15926038e-02
3.23168218e-01 2.14920402e-01 -4.59816307e-01 1.72160253e-01
3.21575373e-01 1.57840168e+00 -1.02015018e+00 1.79523796e-01
-2.73062855e-01 9.60403979e-02 -1.00353718e+00 1.06828876e-01
-9.49949503e-01 -1.51342124e-01 -3.34167391e-01 4.67773587e-01
-6.07071891e-02 -5.22782207e-01 2.85876840e-01 3.13689321e-01
3.30602050e-01 4.69910443e-01 -5.40918112e-01 1.51870930e+00
-7.76740074e-01 1.04645777e+00 -3.07703435e-01 -1.16890764e+00
6.80406630e-01 1.76996604e-01 -2.46795446e-01 -7.10976362e-01
-5.89134842e-02 6.16102442e-02 5.45098126e-01 -6.27555490e-01
4.46927041e-01 7.77679356e-03 -3.61560613e-01 7.59394646e-01
1.17550708e-01 2.90846527e-02 1.92122862e-01 1.71357721e-01
1.75402403e+00 3.11278313e-01 2.70120919e-01 -6.29331917e-02
3.67180049e-01 -1.17699362e-01 3.84824336e-01 1.07033348e+00
1.02909677e-01 2.01800331e-01 1.35865200e+00 -6.03477895e-01
-8.35907102e-01 -6.15269423e-01 7.20217228e-01 1.35555673e+00
-5.71894288e-01 -6.60314918e-01 -8.52725029e-01 -1.14527392e+00
-1.07332945e-01 1.08402646e+00 -7.74841487e-01 -5.50162971e-01
-1.02654099e+00 -6.28948212e-01 7.62440145e-01 9.53030467e-01
5.86149059e-02 -1.53536427e+00 -7.68303573e-01 1.97433263e-01
3.44521165e-01 -6.53498054e-01 -4.13304806e-01 7.55306423e-01
-9.02054191e-01 -1.20158744e+00 -2.72061050e-01 -7.49150157e-01
5.01844227e-01 -1.99532717e-01 1.73288798e+00 5.78648806e-01
7.37933591e-02 3.62049133e-01 -2.39357606e-01 -2.32337952e-01
-9.16685343e-01 4.44679946e-01 -3.73849928e-01 -7.27634192e-01
4.32149768e-01 -6.48368120e-01 2.55403388e-02 -8.86747465e-02
-1.01876628e+00 -2.86075234e-01 5.39901793e-01 1.16538930e+00
6.37735799e-02 9.10475999e-02 3.71906638e-01 -1.53296781e+00
9.76581275e-01 -5.07248819e-01 -9.25841451e-01 2.37118110e-01
-6.44859970e-01 7.50029445e-01 1.22939777e+00 -4.35036659e-01
-6.58534110e-01 -2.84907430e-01 -2.51411587e-01 -5.71811557e-01
7.52162710e-02 1.04447186e+00 -6.10902719e-02 -1.98098108e-01
1.07430542e+00 3.97558324e-02 -1.22211084e-01 -1.35945752e-01
3.57527673e-01 4.72453199e-02 4.08364087e-01 -8.16224635e-01
6.62909925e-01 -1.59505531e-01 9.86455306e-02 -3.43742281e-01
-5.72749615e-01 3.03933341e-02 -2.97615021e-01 2.53761023e-01
8.32946002e-01 -2.70099282e-01 -9.50167179e-01 9.60774571e-02
-1.18792021e+00 -1.09328854e+00 -2.40565270e-01 -2.48968322e-02
-7.34633803e-01 3.26151699e-01 -9.27825928e-01 -5.47631562e-01
-1.91928133e-01 -1.57407594e+00 8.95345747e-01 4.32965159e-02
-3.09735090e-01 -1.48082471e+00 1.02500029e-01 -3.14309686e-01
6.70780480e-01 4.78538990e-01 1.75387836e+00 -9.30978715e-01
-8.05779219e-01 -2.02533945e-01 -6.51730001e-02 4.10316437e-01
-7.97553882e-02 3.03898193e-03 -7.60222137e-01 -3.53495270e-01
1.15251034e-01 -3.79492432e-01 1.04383326e+00 4.70923632e-01
1.30676579e+00 -4.18517768e-01 -2.44410455e-01 7.79254556e-01
1.59670663e+00 2.82616109e-01 6.35222077e-01 2.46342257e-01
1.05584264e+00 4.09842819e-01 -1.96075156e-01 -1.78679526e-01
2.69317329e-01 3.38350862e-01 8.15407574e-01 2.33030379e-01
-1.55206576e-01 -3.55501384e-01 8.42588127e-01 7.70232141e-01
-9.76583809e-02 -3.01785707e-01 -1.44829631e+00 3.23834389e-01
-1.69801307e+00 -9.32491183e-01 -8.13513324e-02 2.03555346e+00
4.70482409e-01 4.90246445e-01 2.28489652e-01 -3.45784575e-02
2.68682986e-01 3.78969491e-01 -5.32286346e-01 -9.13720131e-01
3.35596323e-01 6.84476733e-01 4.60114807e-01 6.12539887e-01
-7.08602846e-01 9.59003150e-01 6.13184404e+00 4.08545256e-01
-1.25618958e+00 -3.30919921e-02 4.03689027e-01 2.90759385e-01
-3.65172833e-01 2.08972558e-01 -7.18053043e-01 1.64704055e-01
1.66853476e+00 -1.69762090e-01 8.79814863e-01 1.10460913e+00
-2.05056727e-01 2.16716677e-01 -1.79830527e+00 5.55786669e-01
1.42473727e-01 -1.36712384e+00 4.60869744e-02 -2.92415582e-02
6.66422486e-01 2.74384379e-01 2.50893295e-01 1.10999274e+00
8.66182983e-01 -1.47099364e+00 4.33828652e-01 2.24358380e-01
4.64598954e-01 -6.81581318e-01 9.10283327e-01 2.85046458e-01
-9.18970823e-01 -3.83843720e-01 -1.64886445e-01 -1.70315951e-01
-2.07593545e-01 1.07055448e-01 -7.21175730e-01 3.76017720e-01
3.19536358e-01 7.89082229e-01 -1.07809663e+00 7.40515888e-01
-4.58621770e-01 7.94813395e-01 8.78429338e-02 -3.05262148e-01
5.55174232e-01 1.82985589e-01 3.37685794e-01 1.30149102e+00
1.61980748e-01 -5.24202108e-01 1.47156924e-01 1.43636453e+00
-1.90896645e-01 -4.18722421e-01 -1.01372862e+00 -5.38944125e-01
-2.05382675e-01 8.10896814e-01 -8.14297855e-01 -1.50688872e-01
-6.91417158e-01 6.05867147e-01 6.22101843e-01 6.50175273e-01
-8.51797163e-01 -6.16635740e-01 2.33299747e-01 -7.72283599e-03
3.13514233e-01 -1.18487000e-01 -7.25748837e-02 -1.00240171e+00
5.85665405e-02 -1.40900552e+00 2.11688310e-01 -7.04140425e-01
-9.34151053e-01 8.34028542e-01 -1.11377835e-01 -5.87222397e-01
-8.14894974e-01 -1.12125301e+00 -7.89949656e-01 8.01231205e-01
-1.35175633e+00 -9.24373150e-01 -7.67292529e-02 2.02490032e-01
3.56306016e-01 -1.89632937e-01 6.68664813e-01 -4.75754887e-02
-6.78822219e-01 7.55053818e-01 -2.89702326e-01 2.41318420e-01
1.64283887e-01 -1.57339537e+00 1.08535540e+00 1.06077349e+00
1.36526242e-01 1.06564116e+00 6.52644575e-01 -6.72324002e-01
-1.76199925e+00 -1.25312495e+00 4.60528433e-01 -6.91740632e-01
8.98693740e-01 -3.40784788e-01 -1.15244484e+00 1.44959545e+00
1.76663697e-01 1.69408862e-02 3.12686145e-01 5.10712624e-01
-7.36930549e-01 5.23782372e-01 -5.28188229e-01 7.41677344e-01
9.52342212e-01 -9.37453747e-01 -4.80642855e-01 5.17213523e-01
1.16240895e+00 -5.51977217e-01 -5.87460756e-01 3.06964755e-01
1.13976344e-01 -1.08471835e+00 6.70117617e-01 -1.03123891e+00
8.73430550e-01 -5.57476766e-02 -1.60070881e-01 -1.35243750e+00
-3.45157117e-01 -9.63321865e-01 -4.18893725e-01 7.69059420e-01
5.89644909e-01 -5.97465694e-01 1.04925203e+00 3.24637979e-01
-4.90063280e-01 -8.30855429e-01 -2.96106309e-01 -8.41785133e-01
2.33669609e-01 -6.33107543e-01 6.39571905e-01 7.16356397e-01
1.27695635e-01 6.38167679e-01 -3.69581543e-02 1.62325613e-02
1.44087479e-01 7.15817139e-02 6.92713857e-01 -9.99919653e-01
-1.06961238e+00 -1.04612017e+00 -4.06610996e-01 -1.07925940e+00
8.94477725e-01 -1.38470685e+00 9.42920968e-02 -1.27036166e+00
6.11696206e-02 4.09452468e-02 -1.73408151e-01 7.83820450e-01
-2.67640382e-01 -3.98011476e-01 1.28168508e-01 -2.90015548e-01
-5.93682468e-01 1.39374688e-01 7.53174841e-01 -3.84428769e-01
-3.05350244e-01 2.72696257e-01 -6.85948133e-01 6.34119630e-01
4.62029874e-01 -4.06056464e-01 -5.07731199e-01 -5.22606552e-01
9.59814727e-01 5.60876727e-01 4.02046442e-01 -9.76453185e-01
1.50809035e-01 2.23592639e-01 -2.28138104e-01 -1.31079376e-01
-3.54038894e-01 -4.59287286e-01 -5.97879151e-03 7.14525104e-01
-5.45069098e-01 6.46101952e-01 5.25142550e-01 6.31668150e-01
-2.70216823e-01 -5.66464841e-01 3.84788424e-01 -4.91018027e-01
-8.44278514e-01 1.57208353e-01 -4.68136370e-01 1.43280625e-01
4.55104828e-01 -1.54770702e-01 -4.67891186e-01 -4.21327561e-01
-7.28345752e-01 1.46351084e-01 5.63750744e-01 4.98968005e-01
4.98761415e-01 -1.07107139e+00 -3.50989610e-01 3.24219942e-01
3.58086731e-03 -2.90631894e-02 -2.18231492e-02 7.01747715e-01
-7.17678905e-01 6.23729765e-01 -1.60291880e-01 -3.86654526e-01
-8.81556332e-01 1.12032747e+00 5.85650921e-01 -9.27530408e-01
-5.98196745e-01 8.34965467e-01 3.68828654e-01 -8.31734240e-01
1.94224179e-01 -1.01664352e+00 -9.51842293e-02 -4.84659940e-01
2.80128866e-01 1.93180203e-01 1.74097881e-01 3.50253098e-03
-5.27269319e-02 1.87610641e-01 -2.10285127e-01 3.64842951e-01
1.35324502e+00 6.29856706e-01 -4.61547494e-01 5.99785805e-01
1.22401690e+00 -1.33692607e-01 -8.95380199e-01 -1.28244370e-01
6.32132173e-01 1.87195331e-01 -5.48887029e-02 -6.66435122e-01
-1.13134491e+00 1.01759803e+00 7.47269690e-02 4.82475877e-01
1.00498366e+00 -1.85573518e-01 6.37111068e-01 6.59132123e-01
3.06093454e-01 -3.13014746e-01 2.80498564e-01 9.96129274e-01
5.38610816e-01 -1.04957032e+00 -3.25806737e-01 1.59171641e-01
-7.22159669e-02 1.43044841e+00 8.19777906e-01 -4.71741349e-01
2.11632013e-01 5.52081048e-01 -4.35777754e-01 -5.53462744e-01
-9.79664803e-01 2.64100377e-02 3.27259392e-01 6.09991610e-01
7.08902299e-01 9.16287582e-03 4.40404415e-01 5.10703206e-01
-2.85664052e-01 1.45381596e-02 8.69038701e-01 9.65768039e-01
-9.10923630e-02 -9.35924888e-01 -1.23119175e-01 7.78864503e-01
-5.16258180e-01 -4.30303305e-01 -2.28807852e-01 1.06777751e+00
-3.84035408e-01 5.43898463e-01 -3.22403282e-01 -6.08576417e-01
2.51045555e-01 3.38611394e-01 5.48621058e-01 -1.13327253e+00
-1.10592377e+00 -4.91263658e-01 -1.13334088e-02 -9.25713420e-01
4.67132218e-02 -3.60926062e-01 -1.01470459e+00 -3.32729191e-01
-3.02414656e-01 -2.45277081e-02 2.65545458e-01 7.58509278e-01
7.97103420e-02 1.24382019e+00 3.45411822e-02 -8.82687330e-01
-7.63977051e-01 -6.62680328e-01 -2.15351045e-01 1.73967123e-01
8.49879444e-01 -1.54275954e-01 -3.83175164e-01 -2.21933126e-01] | [7.721888542175293, 7.64915657043457] |
44f28cab-9feb-4e39-a4bc-8e316fc4e025 | end-to-end-face-parsing-via-interlinked | 2002.04831 | null | https://arxiv.org/abs/2002.04831v2 | https://arxiv.org/pdf/2002.04831v2.pdf | End-to-End Face Parsing via Interlinked Convolutional Neural Networks | Face parsing is an important computer vision task that requires accurate pixel segmentation of facial parts (such as eyes, nose, mouth, etc.), providing a basis for further face analysis, modification, and other applications. Interlinked Convolutional Neural Networks (iCNN) was proved to be an effective two-stage model for face parsing. However, the original iCNN was trained separately in two stages, limiting its performance. To solve this problem, we introduce a simple, end-to-end face parsing framework: STN-aided iCNN(STN-iCNN), which extends the iCNN by adding a Spatial Transformer Network (STN) between the two isolated stages. The STN-iCNN uses the STN to provide a trainable connection to the original two-stage iCNN pipeline, making end-to-end joint training possible. Moreover, as a by-product, STN also provides more precise cropped parts than the original cropper. Due to these two advantages, our approach significantly improves the accuracy of the original model. Our model achieved competitive performance on the Helen Dataset, the standard face parsing dataset. It also achieved superior performance on CelebAMask-HQ dataset, proving its good generalization. Our code has been released at https://github.com/aod321/STN-iCNN. | ['Xiaolin Hu', 'Valentin Yiu', 'Liang Tang', 'Zi Yin'] | 2020-02-12 | null | null | null | null | ['face-parsing'] | ['computer-vision'] | [ 2.35285401e-01 4.08991247e-01 -7.21870810e-02 -6.50764942e-01
-4.66278523e-01 -4.29317206e-01 3.03678989e-01 -5.25473535e-01
-2.98199832e-01 3.73883963e-01 -2.02799991e-01 -3.46710473e-01
4.35986847e-01 -7.06560254e-01 -9.27144706e-01 -5.32745481e-01
3.55398297e-01 2.61916041e-01 3.50764066e-01 9.85904317e-03
-2.27784798e-01 7.27537990e-01 -1.51009202e+00 3.83879960e-01
5.92240334e-01 1.23594689e+00 2.41741583e-01 5.20586491e-01
-4.32933539e-01 3.48952919e-01 -4.71408457e-01 -8.76040816e-01
1.70740038e-01 -4.40140545e-01 -6.76269770e-01 -2.09969237e-01
5.15805840e-01 -4.56239104e-01 -5.71506806e-02 1.03127694e+00
4.41879153e-01 -3.34106356e-01 4.96018752e-02 -1.29479706e+00
-4.69157517e-01 6.29744112e-01 -7.64471948e-01 -3.39480042e-01
-5.46180606e-02 -1.10890783e-01 8.60697389e-01 -8.67523968e-01
3.66735309e-01 1.45519078e+00 8.25420916e-01 1.01846135e+00
-9.22918022e-01 -1.02606356e+00 2.98512392e-02 -9.81865749e-02
-1.01399720e+00 -5.42839408e-01 7.25903034e-01 -2.45738607e-02
7.62500048e-01 7.61266202e-02 5.77975154e-01 1.03353405e+00
-4.48467582e-02 8.59860361e-01 7.50607073e-01 -3.96769673e-01
5.40970489e-02 -1.22065783e-01 3.59441526e-02 1.04637396e+00
1.93914995e-01 3.03227454e-03 -2.53283501e-01 1.27391860e-01
9.00175214e-01 2.18835939e-02 -2.80615371e-02 -5.83986565e-02
-5.09316981e-01 7.63204098e-01 7.24495649e-01 2.13877171e-01
-6.15916327e-02 2.49346301e-01 2.75156379e-01 -9.73264351e-02
4.98701394e-01 -2.18595669e-01 -6.93631053e-01 1.50394395e-01
-8.45028877e-01 -2.13351265e-01 6.32151246e-01 8.13541591e-01
6.66019142e-01 -6.70192987e-02 -6.36570901e-02 1.06031322e+00
4.16785449e-01 4.05337214e-01 2.64936328e-01 -9.34214473e-01
4.69829559e-01 6.65183127e-01 -5.46233714e-01 -5.03079414e-01
-5.38688362e-01 -3.46594512e-01 -8.49834979e-01 3.22230399e-01
5.58390796e-01 -3.47538590e-01 -1.20975816e+00 1.93140531e+00
5.61536372e-01 2.26813272e-01 -5.90738803e-02 6.99535191e-01
1.25429320e+00 5.34871757e-01 3.95102769e-01 1.03227705e-01
1.67976820e+00 -1.23264742e+00 -6.43499911e-01 -4.48956996e-01
5.38178682e-01 -7.89526045e-01 7.68754661e-01 1.63368940e-01
-9.51896369e-01 -6.49019420e-01 -7.60911763e-01 -3.63870233e-01
-4.07074571e-01 5.15715003e-01 8.51372540e-01 8.20947170e-01
-1.36087894e+00 5.32664478e-01 -8.60357285e-01 -3.02316487e-01
9.24326897e-01 6.36645675e-01 -7.70963788e-01 -6.03935421e-02
-9.31391835e-01 4.30256158e-01 2.67692566e-01 5.93341470e-01
-4.48018253e-01 -5.35581410e-01 -1.09264052e+00 6.29492253e-02
4.13027227e-01 -4.95779216e-01 1.50944412e+00 -1.19506872e+00
-1.76646602e+00 8.80251884e-01 -3.33374798e-01 -7.99971893e-02
2.56449133e-01 -1.74447313e-01 -2.76572645e-01 2.35299721e-01
-3.37169506e-02 1.14986575e+00 9.72742856e-01 -9.67732370e-01
-5.23246706e-01 -5.83167970e-01 -1.17636755e-01 -2.80496985e-01
-1.62785634e-01 1.88077018e-01 -9.81571257e-01 -5.58456719e-01
7.43951052e-02 -9.10276592e-01 7.05135465e-02 2.19390407e-01
-3.94815832e-01 -4.98331338e-01 9.97279644e-01 -1.01071799e+00
7.25201964e-01 -2.36607075e+00 1.35840420e-02 -7.69864470e-02
1.63515940e-01 7.76832700e-01 -5.09816706e-01 -2.65785996e-02
-3.34602386e-01 1.23635851e-01 -4.44073170e-01 -8.96945059e-01
-2.18405768e-01 1.79853335e-01 1.06336474e-01 1.64782643e-01
7.23986566e-01 1.15315616e+00 -3.44642818e-01 -5.25411606e-01
7.06794411e-02 7.51496494e-01 -5.66121161e-01 9.63034034e-02
-3.16245288e-01 3.91467988e-01 -2.90710360e-01 9.72838700e-01
1.02505088e+00 2.17707120e-02 1.56130597e-01 -2.15918139e-01
1.11902654e-01 -1.02298051e-01 -7.81838298e-01 1.67608440e+00
-4.98545945e-01 4.84402388e-01 4.09657329e-01 -8.78781855e-01
8.67378891e-01 2.61392653e-01 1.83692276e-01 -8.57697845e-01
3.70012283e-01 1.79788768e-01 -2.01012082e-02 -3.44641477e-01
8.78536329e-03 -2.23070811e-02 -6.92721317e-03 1.46715686e-01
3.27285111e-01 1.71389908e-01 1.20722316e-01 7.92621076e-02
7.52430379e-01 4.69205350e-01 2.71852780e-02 1.54900812e-02
6.17928267e-01 -3.29434663e-01 8.72504771e-01 1.55079290e-01
-2.71602422e-01 7.98791528e-01 8.94670844e-01 -4.95046675e-01
-7.31621504e-01 -7.37070620e-01 -1.81827173e-01 1.06506681e+00
-1.93236008e-01 -3.33171517e-01 -1.43029237e+00 -1.09465826e+00
-1.62128538e-01 2.22769275e-01 -7.04859316e-01 1.98336765e-01
-7.33986497e-01 -6.70462191e-01 6.96749151e-01 7.87930369e-01
8.11432958e-01 -1.33168936e+00 -2.97949225e-01 6.44877851e-02
-8.47653970e-02 -1.40240836e+00 -4.07209218e-01 1.17397942e-01
-7.44401634e-01 -1.22001743e+00 -6.55248106e-01 -9.92843926e-01
7.01794624e-01 -1.11801224e-02 1.02267492e+00 3.94483119e-01
-3.73990893e-01 -4.75898832e-02 -3.20393801e-01 -4.20130193e-01
-2.96481013e-01 9.63461697e-02 -3.27109247e-01 -1.76294688e-02
3.09531212e-01 -3.27165067e-01 -4.32712197e-01 2.18845814e-01
-8.75881910e-01 1.56146735e-01 6.82862341e-01 7.88396537e-01
4.76552755e-01 -1.16734803e-01 5.30582666e-01 -1.09909642e+00
5.79643354e-04 -3.85341533e-02 -1.00683331e+00 2.91343063e-01
-1.68947741e-01 -2.09142361e-02 6.31215513e-01 -1.42165869e-01
-1.20167577e+00 5.54925382e-01 -8.87707770e-01 -1.29520103e-01
-2.18054235e-01 6.25164136e-02 -8.02986145e-01 -5.58614433e-02
-6.33827895e-02 -1.49638787e-01 4.32366669e-01 -6.61458254e-01
3.26718062e-01 6.92275822e-01 6.27166152e-01 -3.21213037e-01
5.90789020e-01 2.79183149e-01 1.15222231e-01 -8.68669093e-01
-7.53786922e-01 -1.42433876e-02 -8.56411338e-01 1.12690359e-01
1.15969658e+00 -8.77888680e-01 -8.54458153e-01 8.41943681e-01
-1.44896257e+00 -5.87008595e-01 1.26972169e-01 1.12481378e-01
-4.25428487e-02 2.70547539e-01 -8.07067335e-01 -6.30663276e-01
-5.83559871e-01 -1.32810247e+00 1.30682158e+00 7.07428217e-01
1.56261295e-01 -8.15741539e-01 -4.43343461e-01 5.74249208e-01
1.91516533e-01 1.56157866e-01 7.96023905e-01 -6.00358844e-01
-3.29786927e-01 -1.55079812e-01 -5.47587276e-01 6.22868359e-01
-1.76586471e-02 3.47689152e-01 -1.34465063e+00 -1.44799054e-01
-1.88947693e-01 -2.65997380e-01 1.06021249e+00 3.99173558e-01
1.40933681e+00 -3.11367121e-02 -2.97316343e-01 9.41802204e-01
1.20966303e+00 1.97435454e-01 9.00701284e-01 8.72903094e-02
7.76539803e-01 7.70604074e-01 2.52866238e-01 -1.21118926e-01
2.42934525e-01 6.61699593e-01 6.57430530e-01 -5.87886751e-01
-4.72565472e-01 -1.92895547e-01 4.12534207e-01 2.95852870e-01
8.31033438e-02 -1.06335804e-01 -7.86233246e-01 1.28667578e-01
-1.48627067e+00 -5.71408331e-01 -2.74635077e-01 1.78718579e+00
5.22726774e-01 -6.46351045e-03 9.94796082e-02 4.47131507e-02
7.99559057e-01 -4.82554249e-02 -5.17326355e-01 -4.78596479e-01
2.69686803e-02 6.93075895e-01 2.04835489e-01 4.12639052e-01
-1.20933676e+00 1.27098417e+00 5.30927086e+00 8.25329542e-01
-1.15881872e+00 2.47887239e-01 9.16352868e-01 1.97982699e-01
1.67512253e-01 -2.40226880e-01 -1.28255641e+00 4.81510252e-01
7.34762788e-01 8.02818120e-01 3.01604003e-01 8.75977814e-01
-6.56190589e-02 -1.07247733e-01 -8.55827212e-01 8.83922398e-01
-6.07308326e-03 -1.09298563e+00 -1.01578966e-01 1.22281708e-01
1.42379075e-01 -1.22306675e-01 -6.09831400e-02 3.37511510e-01
8.49235728e-02 -1.08821809e+00 5.90673208e-01 -6.45452589e-02
1.10006857e+00 -9.24412012e-01 1.00000072e+00 4.36528809e-02
-1.24370289e+00 -6.88902214e-02 -3.70721102e-01 1.83197781e-01
7.70148337e-02 6.86868608e-01 -6.34866774e-01 5.25120616e-01
8.53085101e-01 4.14936066e-01 -7.98869908e-01 6.38235211e-01
-5.67134678e-01 5.98895371e-01 -3.59342635e-01 3.50078911e-01
8.15794095e-02 -1.92904606e-01 -1.45859355e-02 1.04811502e+00
2.12228000e-01 1.11880697e-01 -2.43688479e-01 6.36945248e-01
-4.91757840e-01 -8.33597779e-02 -2.52331167e-01 -4.30153571e-02
2.93109149e-01 1.76597464e+00 -1.08022487e+00 -1.66345760e-01
-5.94538629e-01 1.06212866e+00 5.50504208e-01 8.61153305e-02
-1.04767835e+00 -4.38711256e-01 7.16531038e-01 -6.22411538e-03
5.67884743e-01 1.79870754e-01 -2.51067728e-01 -8.09094489e-01
1.81677222e-01 -7.92551637e-01 2.94873029e-01 -5.24183393e-01
-9.51570928e-01 1.02261603e+00 -2.64804751e-01 -6.81524932e-01
1.49474200e-02 -1.06475031e+00 -5.94035983e-01 8.80745888e-01
-1.59629869e+00 -1.66823959e+00 -3.38041633e-01 3.95890564e-01
5.21618664e-01 -8.50226879e-02 7.50946283e-01 3.74722600e-01
-1.22016144e+00 9.10499990e-01 -3.75181079e-01 6.73519313e-01
5.49476027e-01 -1.03609288e+00 6.51677072e-01 1.10347378e+00
-3.49506922e-02 3.61868292e-01 3.23249474e-02 -5.71415007e-01
-1.08753467e+00 -1.37332559e+00 8.20522666e-01 -2.54076153e-01
2.13891983e-01 -7.13732123e-01 -8.43909979e-01 7.78502226e-01
1.14833489e-01 2.04056963e-01 6.19259179e-01 -1.44248782e-02
-4.27701682e-01 -2.72341967e-01 -1.26740599e+00 4.09781635e-01
1.09984291e+00 -2.23007202e-01 -2.10839480e-01 1.57313675e-01
9.20617640e-01 -4.84017283e-01 -5.48481941e-01 5.18566012e-01
6.87004089e-01 -1.15292132e+00 9.69273090e-01 -8.30390751e-02
6.32265925e-01 -8.50534439e-02 2.71105431e-02 -9.07199681e-01
-1.13604315e-01 -4.04986680e-01 4.46704067e-02 1.66891670e+00
5.63104391e-01 -7.87118733e-01 9.33045328e-01 5.94118357e-01
-1.34626657e-01 -7.36358404e-01 -9.43440557e-01 -5.60070336e-01
6.44300058e-02 -4.40221488e-01 8.51325929e-01 4.43054706e-01
-5.81601441e-01 4.02878970e-01 -1.43243432e-01 2.49255791e-01
6.01258337e-01 -1.14225656e-01 6.63627505e-01 -1.36836612e+00
-1.41689345e-01 -5.19218564e-01 -1.34471580e-01 -9.31221306e-01
4.13214952e-01 -9.19625819e-01 5.71943149e-02 -1.41759264e+00
1.28686070e-01 -4.49307710e-01 8.72546881e-02 1.04293370e+00
-2.96718657e-01 7.99944580e-01 2.71587729e-01 -1.82836324e-01
-1.38960972e-01 3.14817250e-01 1.27288246e+00 8.12373832e-02
-9.85620171e-02 1.56152993e-01 -8.01420152e-01 7.21597970e-01
1.08764660e+00 -4.36629087e-01 -1.99367568e-01 -6.53686285e-01
-1.71554700e-01 -1.63517535e-01 3.60980362e-01 -8.05254340e-01
1.55482978e-01 3.86228204e-01 7.60973930e-01 -6.68976724e-01
3.63450587e-01 -8.65984738e-01 8.03203359e-02 4.93034065e-01
1.69895321e-01 -1.37920724e-02 6.29258037e-01 2.09542200e-01
-2.04786494e-01 -2.90890902e-01 1.01521409e+00 2.43483838e-02
-7.24333048e-01 4.18069243e-01 3.55633423e-02 -3.83668989e-01
9.49494302e-01 -1.86086521e-01 -3.09654057e-01 9.31036472e-02
-5.96268952e-01 1.21331774e-01 3.39357823e-01 3.75763625e-01
4.29437131e-01 -9.81026232e-01 -5.21347523e-01 7.36939132e-01
-2.88172930e-01 4.99295443e-01 2.42460430e-01 6.79714918e-01
-6.87221527e-01 3.41114759e-01 -1.43239677e-01 -6.61214232e-01
-1.60590076e+00 3.49897653e-01 3.17075700e-01 -1.78494975e-01
-5.35765231e-01 1.13607931e+00 2.96735644e-01 -7.24317551e-01
2.91656792e-01 -3.14548105e-01 -2.59040564e-01 2.13705879e-02
6.97759390e-01 1.28800243e-01 2.66820759e-01 -6.10491157e-01
-4.55628157e-01 7.97506392e-01 -2.45251060e-02 1.01537183e-01
1.49113870e+00 1.04143411e-01 -5.38717687e-01 -3.01565617e-01
1.22698462e+00 -1.59777015e-01 -1.30954051e+00 -6.95150951e-03
-5.60767651e-02 -9.49799791e-02 2.75103301e-02 -8.24157357e-01
-1.62618685e+00 1.03736961e+00 4.62455660e-01 -1.45534813e-01
1.49797845e+00 6.22668639e-02 8.98293614e-01 3.92179824e-02
1.42489150e-01 -7.42633164e-01 -8.33438784e-02 4.53981936e-01
7.77889431e-01 -1.40021229e+00 -2.56078660e-01 -9.50883090e-01
-4.59185988e-01 1.30858886e+00 8.52549434e-01 3.17674398e-01
6.22235715e-01 4.55214560e-01 2.03541145e-01 3.26925330e-02
-2.95305908e-01 -2.16721430e-01 1.68205127e-01 5.59415042e-01
4.74432439e-01 -3.52129452e-02 -1.93650313e-02 8.15548003e-01
-2.49270111e-01 2.90149041e-02 2.09250543e-02 8.13945711e-01
-8.32926631e-02 -1.46366501e+00 -3.20256680e-01 1.35974810e-01
-6.02033377e-01 -1.44341230e-01 -4.62743998e-01 1.03065908e+00
3.46394300e-01 9.34284985e-01 1.98125198e-01 -2.45351136e-01
2.57280022e-01 2.32981607e-01 4.42290336e-01 -5.48358142e-01
-6.12223685e-01 2.79030710e-01 -8.99179131e-02 -7.67629027e-01
-1.95692793e-01 -3.66911411e-01 -1.35565400e+00 -2.72342294e-01
-3.05189878e-01 -1.04123456e-02 9.43210006e-01 8.98938596e-01
4.42074180e-01 6.35141730e-01 4.88444805e-01 -8.45701277e-01
-3.13099809e-02 -9.83052254e-01 -3.75836730e-01 5.76635413e-02
1.78698719e-01 -4.15770113e-01 -7.47673288e-02 -4.56572091e-03] | [13.440844535827637, 0.6436769366264343] |
bc337f42-7340-4787-829f-df91ab509928 | prob-probabilistic-objectness-for-open-world | 2212.01424 | null | https://arxiv.org/abs/2212.01424v1 | https://arxiv.org/pdf/2212.01424v1.pdf | PROB: Probabilistic Objectness for Open World Object Detection | Open World Object Detection (OWOD) is a new and challenging computer vision task that bridges the gap between classic object detection (OD) benchmarks and object detection in the real world. In addition to detecting and classifying seen/labeled objects, OWOD algorithms are expected to detect novel/unknown objects - which can be classified and incrementally learned. In standard OD, object proposals not overlapping with a labeled object are automatically classified as background. Therefore, simply applying OD methods to OWOD fails as unknown objects would be predicted as background. The challenge of detecting unknown objects stems from the lack of supervision in distinguishing unknown objects and background object proposals. Previous OWOD methods have attempted to overcome this issue by generating supervision using pseudo-labeling - however, unknown object detection has remained low. Probabilistic/generative models may provide a solution for this challenge. Herein, we introduce a novel probabilistic framework for objectness estimation, where we alternate between probability distribution estimation and objectness likelihood maximization of known objects in the embedded feature space - ultimately allowing us to estimate the objectness probability of different proposals. The resulting Probabilistic Objectness transformer-based open-world detector, PROB, integrates our framework into traditional object detection models, adapting them for the open-world setting. Comprehensive experiments on OWOD benchmarks show that PROB outperforms all existing OWOD methods in both unknown object detection ($\sim 2\times$ unknown recall) and known object detection ($\sim 10\%$ mAP). Our code will be made available upon publication at https://github.com/orrzohar/PROB. | ['Serena Yeung', 'Kuan-Chieh Wang', 'Orr Zohar'] | 2022-12-02 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zohar_PROB_Probabilistic_Objectness_for_Open_World_Object_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zohar_PROB_Probabilistic_Objectness_for_Open_World_Object_Detection_CVPR_2023_paper.pdf | cvpr-2023-1 | ['open-world-object-detection'] | ['computer-vision'] | [ 1.52927384e-01 2.86863089e-01 -1.72551617e-01 -1.94392487e-01
-8.89534771e-01 -5.68466961e-01 5.43308854e-01 1.66152954e-01
-3.70753527e-01 5.71916938e-01 -2.57721484e-01 -8.63574073e-02
1.38290629e-01 -8.25085700e-01 -7.53996491e-01 -6.37439668e-01
-1.56916101e-02 7.40625262e-01 1.00323284e+00 1.80689827e-01
-6.72519282e-02 4.60301608e-01 -1.75407314e+00 8.95031989e-02
5.39545357e-01 1.11107552e+00 4.71176058e-01 8.31587374e-01
-2.80651093e-01 7.35569179e-01 -5.00613391e-01 -4.14810717e-01
5.87464392e-01 -8.04249868e-02 -6.57079339e-01 2.79906422e-01
8.73497486e-01 -6.70523226e-01 -2.33336642e-01 1.11511278e+00
4.88478184e-01 -6.87939376e-02 8.06581914e-01 -1.66603017e+00
-6.36735380e-01 3.80039722e-01 -7.74784267e-01 6.33836627e-01
1.50908604e-01 1.72020629e-01 1.21330523e+00 -1.28873849e+00
5.87433636e-01 1.31485760e+00 5.69465876e-01 5.27602851e-01
-1.42594612e+00 -7.47760952e-01 3.60318929e-01 3.15473378e-01
-1.60260224e+00 -1.86408564e-01 4.32690859e-01 -6.23481452e-01
5.77968001e-01 3.16907793e-01 3.43206584e-01 8.97395670e-01
-1.41727343e-01 1.21612573e+00 9.83767033e-01 -3.50327313e-01
1.67402968e-01 4.80260432e-01 6.75044060e-01 6.04950428e-01
7.76731610e-01 3.04918736e-01 -2.46504992e-01 -1.30090624e-01
3.65605474e-01 2.33151183e-01 -2.61608213e-01 -6.51809871e-01
-1.16714013e+00 7.23076820e-01 5.91480911e-01 -1.64366454e-01
-1.25028610e-01 1.31683350e-01 1.39598384e-01 2.48346534e-02
2.83198506e-01 5.90911806e-02 -4.20107812e-01 1.72353014e-01
-7.59080172e-01 4.39584106e-01 8.77962530e-01 1.15355313e+00
1.02283692e+00 -2.70125180e-01 -4.75870252e-01 7.40920424e-01
4.53681529e-01 6.64305449e-01 1.25584349e-01 -8.03856730e-01
1.20873027e-01 7.46300697e-01 2.98802078e-01 -7.31329858e-01
-1.97148517e-01 -4.15479988e-01 -3.48118395e-01 4.60829109e-01
5.92819571e-01 1.83005527e-01 -1.13429987e+00 1.51460361e+00
7.99809337e-01 2.79928088e-01 -5.96576743e-02 8.62716734e-01
8.82019043e-01 5.46936274e-01 -2.26256121e-02 -5.57001904e-02
1.65067720e+00 -9.70671773e-01 -5.51746428e-01 -4.56775814e-01
3.81227463e-01 -7.34797120e-01 8.72042358e-01 3.11771154e-01
-6.96281195e-01 -4.67052788e-01 -9.78381515e-01 7.60436431e-02
-4.49886590e-01 -1.31976251e-02 3.99334252e-01 8.06468248e-01
-6.60885096e-01 1.17502227e-01 -7.70439148e-01 -3.53042692e-01
8.79311204e-01 2.67572731e-01 -9.41821784e-02 -2.98940957e-01
-6.57121301e-01 7.65176654e-01 7.74109662e-01 -2.31575131e-01
-1.42179310e+00 -4.96002376e-01 -8.14450741e-01 -1.41701892e-01
8.72241855e-01 -5.83123744e-01 1.45643926e+00 -7.95347035e-01
-7.95133173e-01 9.19838428e-01 -1.13145284e-01 -6.18058324e-01
5.98541737e-01 -3.15478146e-01 -3.06765109e-01 9.40142025e-04
4.53970015e-01 8.76670361e-01 1.04232681e+00 -1.44174469e+00
-1.19492030e+00 -3.80377948e-01 2.36951798e-01 1.88864078e-02
-1.20455846e-01 1.15236454e-01 -4.70025092e-01 -5.94895780e-01
3.57714504e-01 -9.62896168e-01 -1.82250977e-01 5.43561518e-01
-5.42886376e-01 -4.38299388e-01 1.18639946e+00 -1.18060201e-01
8.25407982e-01 -2.17802358e+00 -4.05839503e-01 -1.24020636e-01
6.01378739e-01 2.06768692e-01 -1.34642180e-02 -6.83529228e-02
2.41492078e-01 -2.78393686e-01 -1.29177511e-01 -2.26862252e-01
1.37839451e-01 3.15049529e-01 -5.19934654e-01 4.41007465e-01
4.76222694e-01 7.71550834e-01 -1.02080691e+00 -6.43863857e-01
1.53470248e-01 2.36523926e-01 -3.41316253e-01 1.96455613e-01
-4.23583418e-01 -5.20952642e-02 -2.80333698e-01 1.02441740e+00
1.02193689e+00 -2.42604554e-01 -2.32116267e-01 2.15932764e-02
8.69859084e-02 -8.21479410e-02 -1.89182925e+00 9.22599792e-01
-9.82240885e-02 6.00434184e-01 -1.96509436e-02 -7.40797162e-01
8.84905219e-01 -6.05221801e-02 2.22496346e-01 -2.77328014e-01
1.78713307e-01 2.95873612e-01 9.48417336e-02 -3.15221280e-01
5.75496793e-01 -8.23169798e-02 1.73446134e-01 3.72297138e-01
2.48646900e-01 -1.01148956e-01 4.01564509e-01 3.74614984e-01
1.20725703e+00 9.56261382e-02 4.66115266e-01 -2.07678646e-01
2.76294231e-01 1.84372231e-01 6.70621216e-01 1.23052371e+00
-6.01315141e-01 7.19183505e-01 2.65762955e-01 -3.72447282e-01
-6.38458371e-01 -1.63480961e+00 -5.98431051e-01 1.25561166e+00
5.06742179e-01 -1.11000419e-01 -5.82353950e-01 -9.51265812e-01
2.16041401e-01 5.34382463e-01 -4.81406093e-01 -1.18975742e-02
-3.18513483e-01 -8.44891608e-01 4.16620046e-01 5.84850609e-01
1.26727670e-01 -9.93686676e-01 -7.06624389e-01 3.22204053e-01
-2.69227140e-02 -1.31399930e+00 -4.09642994e-01 4.79510754e-01
-5.88280797e-01 -1.24868953e+00 -7.05577314e-01 -7.71431386e-01
5.88046968e-01 6.89730227e-01 1.06974328e+00 -1.61676586e-01
-8.27527940e-01 4.62879777e-01 -2.52237499e-01 -9.76026952e-01
-4.37263072e-01 -4.22087193e-01 3.17671478e-01 2.36358598e-01
6.18181586e-01 -4.06491533e-02 -7.17372894e-01 7.03330755e-01
-8.12045217e-01 -2.71980941e-01 5.84867358e-01 7.95606077e-01
6.60213888e-01 -1.85196072e-01 3.84591669e-01 -7.10667074e-01
-4.32755165e-02 -6.02897286e-01 -8.37901711e-01 1.12575948e-01
-5.27956188e-01 -7.90454298e-02 1.44999214e-02 -8.13503444e-01
-7.94833601e-01 1.02657162e-01 1.49995863e-01 -6.03678048e-01
-3.13270062e-01 -1.39269367e-01 -2.14126155e-01 1.07877716e-01
8.15931380e-01 9.63846743e-02 -2.76863396e-01 -3.97505641e-01
2.76462555e-01 6.87506497e-01 5.42129278e-01 -5.03463507e-01
9.34910476e-01 9.19787586e-01 -2.48472258e-01 -5.71106136e-01
-1.10171735e+00 -9.45953071e-01 -4.39681381e-01 -6.71572089e-02
7.31085539e-01 -1.09608829e+00 -3.91574532e-01 3.23780179e-01
-1.07198381e+00 -4.58855666e-02 -6.47451639e-01 4.21896130e-01
-3.56148422e-01 3.48678976e-01 -2.88403898e-01 -1.00280070e+00
-1.72000840e-01 -1.14908838e+00 1.30102253e+00 2.62741506e-01
-9.85614210e-03 -5.37534893e-01 -6.53395653e-02 3.37102503e-01
3.30959819e-03 -3.35741080e-02 2.27997810e-01 -9.07322884e-01
-9.36333060e-01 -4.56927985e-01 -7.03758597e-01 3.69015634e-01
6.29075691e-02 -7.67253265e-02 -1.20819998e+00 -2.20371351e-01
-1.78124070e-01 -3.49272758e-01 1.05413568e+00 3.55701834e-01
9.35034990e-01 -1.05925709e-01 -6.49233282e-01 1.84353814e-01
1.33578956e+00 -3.48151214e-02 2.17029721e-01 3.52668256e-01
5.43016315e-01 4.31186974e-01 8.48221183e-01 6.06948793e-01
3.78215909e-01 8.28378916e-01 6.05906308e-01 9.67139564e-03
-3.76284122e-01 -1.03074767e-01 5.20445824e-01 -1.15834750e-01
2.23552525e-01 -3.29903781e-01 -1.11582541e+00 7.48259127e-01
-1.77646518e+00 -9.40873504e-01 -4.59606349e-01 2.15005350e+00
6.13742471e-01 5.63672543e-01 3.78833055e-01 1.34565771e-01
9.29669917e-01 -3.19922954e-01 -5.57710528e-01 1.84165657e-01
-7.52396211e-02 -1.20209694e-01 5.68786740e-01 2.34007403e-01
-1.31953931e+00 7.01190531e-01 5.48707104e+00 8.30764532e-01
-7.32424080e-01 4.98398542e-01 4.84716535e-01 -1.11858323e-01
1.77611515e-01 9.39798057e-02 -1.64995265e+00 3.09537947e-01
4.24336582e-01 -9.33103859e-02 -1.82385325e-01 1.22197545e+00
-5.20734861e-02 -3.66225570e-01 -1.29057944e+00 9.96106744e-01
1.42880931e-01 -1.10295248e+00 4.70121019e-02 6.21558279e-02
6.59274518e-01 3.59798223e-01 -9.83082727e-02 6.34084105e-01
6.30371034e-01 -3.97135228e-01 1.06590760e+00 5.16308807e-02
4.81779337e-01 -7.70160109e-02 6.78097129e-01 5.77971280e-01
-1.24982035e+00 -4.55010980e-01 -4.23790723e-01 -6.13532448e-03
1.21450178e-01 6.52166665e-01 -1.19744086e+00 2.07523078e-01
1.06137371e+00 3.65974516e-01 -5.82271755e-01 1.29475653e+00
-1.30521581e-01 5.10136127e-01 -5.47210157e-01 9.46943611e-02
1.28736511e-01 1.69890642e-01 9.43503857e-01 1.21089125e+00
2.26319022e-02 -5.26126102e-02 6.72063053e-01 1.10367858e+00
3.31256650e-02 -1.16286501e-01 -5.13180196e-01 3.95146757e-01
2.73258120e-01 1.28546214e+00 -1.09728670e+00 -5.20982623e-01
-6.31223083e-01 7.94046700e-01 2.86604792e-01 8.61768126e-02
-1.09403503e+00 -1.97549894e-01 5.52695274e-01 3.33293766e-01
6.94840074e-01 3.89171136e-03 -2.08317172e-02 -1.12420762e+00
2.03109592e-01 -4.89213794e-01 7.66754150e-01 -5.24293065e-01
-1.37816525e+00 3.89058739e-01 2.01655701e-01 -1.38691854e+00
2.94005096e-01 -9.19209719e-01 -4.73306030e-01 3.96579981e-01
-1.64164805e+00 -1.23873484e+00 -4.92326796e-01 3.87024999e-01
9.28371727e-01 1.59790158e-01 3.86051297e-01 3.19049358e-01
-6.03960693e-01 4.15040582e-01 1.30098969e-01 2.09172979e-01
6.03392780e-01 -1.45022130e+00 1.84489071e-01 9.55837309e-01
4.09564883e-01 2.19487488e-01 8.10825467e-01 -5.63086212e-01
-1.11716497e+00 -1.41458106e+00 5.01356602e-01 -8.19736183e-01
8.84288728e-01 -5.97721517e-01 -9.29947257e-01 6.46817029e-01
-4.38220084e-01 9.66782928e-01 3.03904474e-01 -4.45702560e-02
-5.24535179e-01 -1.03418954e-01 -1.15646279e+00 4.32266861e-01
1.17998469e+00 -1.32969245e-01 -7.07821548e-01 3.63718867e-01
7.64568090e-01 -3.79294395e-01 -4.77573782e-01 5.04625976e-01
3.73046428e-01 -8.55363488e-01 1.15590620e+00 -2.85043001e-01
-3.51995081e-02 -7.18035936e-01 -3.82182330e-01 -6.37211740e-01
-1.67900085e-01 -3.53087306e-01 -5.60674369e-01 1.29564762e+00
1.81130871e-01 -5.85967124e-01 7.68808901e-01 4.82190788e-01
-1.31795943e-01 -4.88167316e-01 -9.86724555e-01 -1.38333106e+00
-3.21570396e-01 -7.04093814e-01 2.05682501e-01 6.43709362e-01
-4.50623989e-01 1.78419948e-01 6.54678196e-02 5.31218648e-01
9.04330194e-01 2.05476910e-01 9.58070755e-01 -1.28003716e+00
-4.32788908e-01 -3.74550313e-01 -8.64784181e-01 -1.10063374e+00
-2.29840800e-01 -8.42190266e-01 2.83433259e-01 -1.36611772e+00
6.02302849e-01 -6.28613234e-01 -3.73424292e-01 5.78530788e-01
-4.26720023e-01 6.89470530e-01 2.99428493e-01 3.03851992e-01
-9.77411807e-01 4.41024929e-01 8.56568933e-01 -3.77901703e-01
-1.67171329e-01 2.25848019e-01 -4.46879953e-01 8.66104722e-01
5.22673011e-01 -8.70336473e-01 -4.53937463e-02 -1.09248579e-01
-1.47193819e-01 -3.89273435e-01 9.53542888e-01 -1.09074736e+00
-9.37180826e-04 -4.28947024e-02 1.48414418e-01 -8.78858507e-01
2.97171235e-01 -8.87250960e-01 -1.20068938e-01 6.67262077e-01
2.40727961e-01 -4.41322058e-01 2.29421675e-01 9.65734005e-01
-5.88997379e-02 -3.42032611e-01 9.97200310e-01 -8.04454386e-02
-1.21889222e+00 3.28004926e-01 -3.12215030e-01 1.63648233e-01
1.56314909e+00 -5.92490196e-01 -4.25421745e-01 1.79411918e-01
-8.15444589e-01 2.56064087e-01 2.47393519e-01 5.49135208e-01
5.25709510e-01 -1.02154386e+00 -8.73702943e-01 4.73439991e-02
6.90505385e-01 2.77475774e-01 -6.79694042e-02 8.59817922e-01
-3.20444196e-01 4.28396128e-02 1.77076146e-01 -1.03916276e+00
-1.34771252e+00 8.24123263e-01 4.06939715e-01 -5.84852360e-02
-6.84706748e-01 9.69419062e-01 6.80186272e-01 -4.38251853e-01
3.66986156e-01 -1.86737567e-01 1.18541799e-01 5.01024611e-02
7.32264698e-01 5.66822410e-01 -1.77082010e-02 -4.79889959e-01
-3.70741040e-01 2.08810046e-01 -4.89992917e-01 2.32693762e-01
1.06692553e+00 -1.64759532e-01 1.01511508e-01 5.90080321e-01
7.85837829e-01 -2.44771406e-01 -1.41992021e+00 -6.94680631e-01
2.52228558e-01 -6.05618298e-01 5.09424470e-02 -6.47958159e-01
-6.06031537e-01 6.98831797e-01 1.02792537e+00 2.40932524e-01
7.28059709e-01 6.76531911e-01 5.13607442e-01 3.78308207e-01
4.24708247e-01 -7.78673589e-01 3.07079792e-01 4.38632607e-01
5.88975966e-01 -1.77674830e+00 1.01775318e-01 -6.73629642e-01
-3.27362835e-01 9.16565716e-01 1.07820785e+00 3.25189496e-04
7.29456127e-01 4.38485652e-01 2.46107057e-02 -2.42931217e-01
-5.57268739e-01 -6.06317699e-01 3.32551330e-01 5.34994304e-01
-6.95845410e-02 3.25442888e-02 1.01979427e-01 5.61892271e-01
2.84593254e-01 -1.06217831e-01 5.99931479e-01 1.08888698e+00
-7.38878667e-01 -7.89274335e-01 -8.29587817e-01 6.10412300e-01
-3.58633548e-01 -4.83960770e-02 -1.49813771e-01 6.41719043e-01
4.74447697e-01 7.53494918e-01 2.34759957e-01 1.43171355e-01
2.46480331e-01 -1.54559642e-01 2.54221618e-01 -8.57929468e-01
-3.17298844e-02 -1.39781870e-02 -2.00207442e-01 -3.43190223e-01
-2.15216413e-01 -8.37395549e-01 -1.22203386e+00 2.40642816e-01
-1.03727055e+00 -1.60115451e-01 4.49163586e-01 6.31906688e-01
1.61298737e-01 2.37962201e-01 4.10268128e-01 -9.12812054e-01
-8.14961135e-01 -7.21501112e-01 -6.47025049e-01 3.50739568e-01
5.17248094e-01 -1.03828740e+00 -3.75863522e-01 1.28185794e-01] | [9.34732437133789, 1.46369469165802] |
9aa0be6b-7f56-4273-94cd-489aa60a5c56 | mimo-detection-under-hardware-impairments | 2306.05146 | null | https://arxiv.org/abs/2306.05146v1 | https://arxiv.org/pdf/2306.05146v1.pdf | MIMO Detection under Hardware Impairments: Learning with Noisy Labels | This paper considers a data detection problem in multiple-input multiple-output (MIMO) communication systems with hardware impairments. To address challenges posed by nonlinear and unknown distortion in received signals, two learning-based detection methods, referred to as model-driven and data-driven, are presented. The model-driven method employs a generalized Gaussian distortion model to approximate the conditional distribution of the distorted received signal. By using the outputs of coarse data detection as noisy training data, the model-driven method avoids the need for additional training overhead beyond traditional pilot overhead for channel estimation. An expectation-maximization algorithm is devised to accurately learn the parameters of the distortion model from noisy training data. To resolve a model mismatch problem in the model-driven method, the data-driven method employs a deep neural network (DNN) for approximating a-posteriori probabilities for each received signal. This method uses the outputs of the model-driven method as noisy labels and therefore does not require extra training overhead. To avoid the overfitting problem caused by noisy labels, a robust DNN training algorithm is devised, which involves a warm-up period, sample selection, and loss correction. Simulation results demonstrate that the two proposed methods outperform existing solutions with the same overhead under various hardware impairment scenarios. | ['H. Vincent Poor', 'Yo-Seb Jeon', 'Seunghyeon Jeon', 'Jinman Kwon'] | 2023-06-08 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 4.17489320e-01 -1.35473981e-01 1.75689366e-02 -2.85102099e-01
-1.17151737e+00 2.93826293e-02 1.33349672e-01 -7.08989128e-02
-4.60800231e-01 9.08452928e-01 3.33720818e-02 -5.21102309e-01
-8.64624381e-02 -4.71260190e-01 -4.90420192e-01 -8.67323697e-01
-2.38334417e-01 -3.05311680e-01 3.55791152e-02 8.95337909e-02
4.69918251e-02 2.14156300e-01 -8.39774251e-01 5.40570654e-02
7.75403678e-01 1.27383935e+00 4.70246434e-01 1.02196634e+00
4.93862122e-01 5.89272976e-01 -1.09214342e+00 -7.28984848e-02
6.10392749e-01 -7.35134125e-01 1.30482867e-01 1.44101128e-01
6.58749789e-02 -8.71675253e-01 -9.31623995e-01 1.34372973e+00
1.25163734e+00 -1.45583868e-01 5.23070455e-01 -1.03286350e+00
-3.26892138e-01 2.87410408e-01 -3.22340369e-01 2.61528581e-01
2.60286219e-02 -3.11105609e-01 4.88610774e-01 -1.33206642e+00
-4.37715910e-02 1.21591640e+00 9.71252501e-01 3.47360015e-01
-9.55492616e-01 -9.33238029e-01 -4.08167690e-01 1.93511814e-01
-1.92953908e+00 -7.86598206e-01 5.87135017e-01 -2.34801143e-01
5.77961862e-01 3.48268226e-02 1.27738088e-01 6.70823157e-01
3.79698351e-02 7.61120737e-01 8.85794401e-01 -8.03885698e-01
3.30120236e-01 4.73112613e-01 -1.09137930e-01 5.50581574e-01
3.73499453e-01 5.41860163e-01 -2.86690801e-01 -4.74468768e-01
7.76122093e-01 -2.52668947e-01 -5.97637475e-01 -4.18402590e-02
-1.02215016e+00 5.61454058e-01 1.76011160e-01 9.26201269e-02
-3.22502792e-01 3.61038178e-01 1.94420263e-01 7.64768124e-01
1.15557700e-01 -8.15957114e-02 -1.95427909e-01 6.06003478e-02
-1.25024474e+00 -6.39028102e-03 7.66672552e-01 1.31090248e+00
5.89715362e-01 7.94700503e-01 -4.55501795e-01 7.61967480e-01
7.01196551e-01 7.44113863e-01 3.62384558e-01 -6.19441211e-01
8.76058280e-01 -1.77679077e-01 2.07960129e-01 -9.29968774e-01
-3.63124013e-01 -1.10865593e+00 -1.08311355e+00 8.55797902e-02
3.09921712e-01 -8.41051936e-01 -8.92504215e-01 1.48390388e+00
1.15252636e-01 3.16208661e-01 5.53197086e-01 8.11285257e-01
3.47147107e-01 7.18791783e-01 -5.11277199e-01 -4.09125865e-01
6.78532243e-01 -6.61266387e-01 -9.08954561e-01 -1.54907908e-02
7.73527920e-01 -9.11644518e-01 2.05535144e-01 6.57923400e-01
-8.86612654e-01 -7.06111014e-01 -1.82468164e+00 3.67309153e-01
1.61068663e-01 6.43008053e-01 -3.51106435e-01 1.16117859e+00
-9.53352749e-01 4.26613510e-01 -2.57748634e-01 6.11099079e-02
2.51521736e-01 5.66955328e-01 1.54199809e-01 -2.11758390e-01
-1.34825182e+00 7.41177976e-01 4.66793358e-01 2.43038684e-01
-1.05048847e+00 -3.93145293e-01 -6.51420116e-01 8.24767947e-02
1.05569869e-01 -2.65494496e-01 1.44643474e+00 -1.08117104e+00
-1.36652422e+00 -4.19782251e-02 -1.76053032e-01 -7.81534910e-01
3.43458086e-01 -1.24487618e-04 -9.80824172e-01 6.03985786e-02
-3.83587241e-01 -4.39041816e-02 1.20754015e+00 -1.42256486e+00
-8.53183806e-01 1.60027891e-01 -3.46107543e-01 2.34547868e-01
-1.78337455e-01 -2.47007102e-01 -2.76100636e-01 -1.06488740e+00
3.80802572e-01 -6.02183402e-01 -3.08309197e-01 -1.42361755e-02
-4.95557994e-01 6.32787526e-01 1.08576918e+00 -9.31806386e-01
1.66908407e+00 -2.48882866e+00 -4.78357106e-01 7.71891296e-01
9.49399080e-03 7.21925914e-01 1.03286440e-02 4.46640551e-01
1.36715442e-01 -2.79385030e-01 -2.03450441e-01 -3.32668751e-01
-1.32965118e-01 3.17106396e-01 -1.90966114e-01 6.40784740e-01
7.76859373e-02 1.92074925e-01 -7.71108270e-01 -1.02108821e-01
9.29295644e-02 4.11501437e-01 -4.91576016e-01 3.42544615e-01
2.67792106e-01 3.67681831e-01 -4.97049317e-02 6.28948033e-01
1.00183034e+00 1.95651710e-01 6.48939013e-02 -3.76426220e-01
5.16892932e-02 3.34027648e-01 -1.74177945e+00 1.07107735e+00
-5.23441851e-01 8.18612695e-01 4.61875886e-01 -1.26898742e+00
1.05993986e+00 8.94904137e-01 -5.01278080e-02 -6.93787813e-01
3.68259490e-01 5.40603280e-01 3.33154172e-01 -3.97121459e-01
3.87200624e-01 -2.39359230e-01 -1.49400607e-01 1.60876051e-01
2.71935225e-01 1.83507383e-01 -3.57752472e-01 3.70756648e-02
9.11119342e-01 -4.17423308e-01 5.99539638e-01 3.35208327e-02
6.22975409e-01 -4.51676190e-01 1.04322338e+00 1.10702801e+00
-1.70173436e-01 4.73689288e-01 -1.76112298e-02 6.95025697e-02
-1.14474857e+00 -8.42270613e-01 -2.55965501e-01 5.47686338e-01
2.08254769e-01 1.97744258e-02 -6.27202868e-01 -1.69325411e-01
1.15275607e-01 6.71608508e-01 1.69616759e-01 -2.89031267e-01
-2.66027272e-01 -9.12285447e-01 9.67554033e-01 3.01238477e-01
8.46924007e-01 -7.01335222e-02 1.17860757e-01 5.95399618e-01
1.56897634e-01 -1.15599823e+00 -3.51280093e-01 4.65255976e-01
-6.22169375e-01 -4.24122393e-01 -8.35543931e-01 -8.42767894e-01
9.30393517e-01 5.57319701e-01 2.99297631e-01 5.40903956e-02
1.84492525e-02 3.09326172e-01 -3.00175190e-01 -4.88081396e-01
-6.91117823e-01 -5.33900678e-01 5.47783732e-01 3.42510968e-01
2.36970201e-01 -4.39361095e-01 -3.47820461e-01 3.57563525e-01
-6.18434846e-01 -3.44976038e-01 9.53101397e-01 1.15495837e+00
3.32481384e-01 4.04700696e-01 9.34493423e-01 -4.12490785e-01
8.26480925e-01 -6.77613020e-01 -6.52936637e-01 8.30258802e-03
-5.35168409e-01 -2.18078587e-02 9.78525698e-01 -4.43084210e-01
-9.95604157e-01 1.62193738e-02 -3.15233350e-01 -1.57733127e-01
9.19252783e-02 5.97284496e-01 -3.29651207e-01 -5.25803447e-01
8.37773383e-01 4.51736033e-01 -1.58256292e-01 -4.07015562e-01
-1.30991861e-01 1.67529297e+00 7.25684404e-01 -1.16703138e-01
1.09808779e+00 5.20490203e-03 -3.53192799e-02 -1.14849389e+00
-5.66552103e-01 -5.37531078e-01 -3.91424805e-01 -1.48167044e-01
1.36496231e-01 -1.52576137e+00 -1.68333486e-01 4.84798074e-01
-1.16552544e+00 4.19634543e-02 1.53259188e-01 1.08220482e+00
-3.47623795e-01 5.25570333e-01 -4.32427257e-01 -1.11707950e+00
-3.61727893e-01 -8.97491574e-01 5.12738228e-01 7.11590946e-02
4.91237789e-02 -7.70531535e-01 -4.19094622e-01 -1.01024762e-01
5.46524048e-01 9.42850709e-02 7.93174386e-01 -6.36797667e-01
-6.98992431e-01 -7.47969389e-01 -2.37149253e-01 9.53000009e-01
6.06219508e-02 -6.58327639e-01 -1.01193178e+00 -7.23577678e-01
2.58494914e-01 4.59838621e-02 3.04608226e-01 3.19713354e-01
9.27497447e-01 -4.73747194e-01 -2.49013349e-01 6.22165561e-01
1.61942887e+00 4.98050213e-01 6.06079876e-01 -1.16981372e-01
3.55716467e-01 -2.71188527e-01 6.32625341e-01 7.83065677e-01
2.02613711e-01 6.10488832e-01 1.12527646e-01 -2.71114618e-01
-2.19757497e-01 -1.27025783e-01 5.08743286e-01 9.36761737e-01
5.71573734e-01 -4.38059598e-01 -4.03623253e-01 2.77630895e-01
-1.61305308e+00 -9.38129365e-01 -3.32926661e-01 2.61852694e+00
8.52757037e-01 3.52527916e-01 -1.85414568e-01 7.10311949e-01
6.90265298e-01 -3.54970008e-01 -5.58192670e-01 -3.70283931e-01
-7.89422840e-02 -1.41639069e-01 1.03629601e+00 6.24041617e-01
-1.02320838e+00 9.42483246e-02 6.20948219e+00 1.05260789e+00
-9.90719855e-01 1.25941396e-01 2.82264471e-01 1.13682121e-01
1.98389351e-01 -6.11873269e-02 -1.01768661e+00 6.17465079e-01
1.21856201e+00 -2.58968204e-01 2.66691893e-01 5.87203681e-01
8.04281771e-01 -3.66111994e-01 -1.08772719e+00 1.15505087e+00
3.93083692e-02 -9.42772150e-01 -1.69124857e-01 -1.07411062e-02
6.15455627e-01 -1.87524766e-01 1.17297664e-01 3.43303204e-01
-8.85636136e-02 -5.28083563e-01 7.30063081e-01 7.67516255e-01
7.69333363e-01 -7.11891532e-01 1.03195453e+00 6.56434655e-01
-9.34157908e-01 -5.81542373e-01 -4.94222373e-01 -2.69345492e-01
2.51931220e-01 1.23012495e+00 -1.13914788e+00 4.99813884e-01
1.92963496e-01 1.22310296e-01 4.55150567e-02 1.66055572e+00
-1.36534587e-01 9.20348346e-01 -2.61878401e-01 4.95220833e-02
-1.41365742e-02 1.59256399e-01 7.55473077e-01 1.43717515e+00
8.38554740e-01 3.42401899e-02 3.53452235e-01 2.42117718e-01
-4.19652127e-02 -3.32001567e-01 -4.67853040e-01 3.82234335e-01
9.79108930e-01 7.25689471e-01 1.80069730e-01 -4.94933486e-01
-5.38788617e-01 8.49740386e-01 -3.12031299e-01 5.84665954e-01
-5.89588821e-01 -9.41530347e-01 3.09828937e-01 -1.02464698e-01
3.57235938e-01 -6.82485163e-01 -3.00268561e-01 -5.39194345e-01
-3.06370556e-02 -8.18581820e-01 -1.57526225e-01 -4.10972416e-01
-8.54912996e-01 2.37868711e-01 -2.41953269e-01 -1.94242871e+00
-3.04345042e-01 -3.23544562e-01 -3.38805556e-01 1.25853717e+00
-1.70468020e+00 -4.06740218e-01 -4.79142852e-02 4.74678278e-01
3.02032501e-01 -4.50927973e-01 8.30073059e-01 9.43446159e-01
-4.87118661e-01 1.05914974e+00 8.09707165e-01 2.41795734e-01
6.13876879e-01 -9.84321833e-01 2.90782508e-02 1.16944981e+00
-2.49961361e-01 3.14389616e-01 8.15993130e-01 -4.33345795e-01
-1.13410711e+00 -1.38870168e+00 9.28862274e-01 4.12442416e-01
1.09184824e-01 -3.94146055e-01 -7.13422596e-01 2.12342799e-01
-7.41043710e-04 1.72246873e-01 1.07163107e+00 -7.11526513e-01
1.23478055e-01 -3.11924219e-01 -1.35010898e+00 2.36341670e-01
4.73227322e-01 -4.44487303e-01 -2.51982182e-01 1.59188315e-01
3.17064345e-01 -6.37221634e-01 -7.63291538e-01 1.87241301e-01
3.43658447e-01 -6.95169270e-01 8.02149355e-01 1.90093905e-01
-3.92769724e-01 -7.67626345e-01 -6.46410644e-01 -1.45777941e+00
-3.59264493e-01 -9.04441476e-01 -4.94470209e-01 1.03836393e+00
6.47340834e-01 -3.29847604e-01 4.00429636e-01 1.16580516e-01
-3.20020616e-01 -4.35793996e-01 -1.04375648e+00 -1.02169132e+00
-4.90648419e-01 -3.95445198e-01 1.95342228e-01 5.34078538e-01
-6.55485913e-02 1.80552632e-01 -9.64160740e-01 8.81828070e-01
8.70924711e-01 -7.13494062e-01 7.64595449e-01 -8.15888166e-01
-6.06329262e-01 2.19945818e-01 -5.69377244e-01 -1.72764611e+00
-5.81221342e-01 -6.73604488e-01 2.98794389e-01 -1.19938147e+00
-5.07559061e-01 -5.40481985e-01 -2.90982455e-01 5.82647398e-02
-8.14836919e-02 2.01569438e-01 7.80875012e-02 6.50357306e-02
-3.29408288e-01 5.46603322e-01 7.09380209e-01 -2.39721267e-03
-2.79347926e-01 6.82441354e-01 -6.91791832e-01 5.79898357e-01
9.11966145e-01 -6.76961124e-01 -2.94923574e-01 -4.54574376e-01
-1.72657549e-01 3.13054025e-01 2.82839239e-01 -1.75659978e+00
5.01768768e-01 4.66355950e-01 5.25882423e-01 -5.25148630e-01
4.66352612e-01 -1.19160485e+00 -6.87025934e-02 7.23408997e-01
-1.65724635e-01 -4.92780477e-01 -9.63308476e-03 9.48856115e-01
-3.90033424e-01 -3.72154057e-01 1.02828550e+00 3.96454453e-01
-5.87829232e-01 1.32877594e-02 -8.71292353e-01 -3.60138535e-01
8.79540801e-01 -4.82162535e-01 1.88588426e-01 -8.13137352e-01
-6.40877068e-01 3.02334279e-01 -4.30578798e-01 1.22968920e-01
8.39655459e-01 -1.57370436e+00 -6.13640428e-01 3.99934798e-01
-9.80921388e-02 -2.49604627e-01 1.69349015e-02 7.95684516e-01
-1.01286888e-01 5.75488031e-01 3.10354859e-01 -5.18744230e-01
-1.26883328e+00 8.25428814e-02 5.47860026e-01 2.78249439e-02
-1.26989916e-01 8.45209658e-01 -6.10438406e-01 -1.47193268e-01
7.01860845e-01 -2.89008290e-01 1.65302306e-01 -4.98257279e-01
6.68772340e-01 5.43429613e-01 2.67327219e-01 -6.30851388e-01
5.31970598e-02 1.56965166e-01 2.08741594e-02 -1.18612655e-01
6.79196775e-01 -5.38932085e-01 5.06184995e-01 2.54087567e-01
1.36605394e+00 -1.23377696e-01 -1.31260681e+00 -9.42481279e-01
-6.56504780e-02 -6.23216927e-01 5.45815468e-01 -7.12347269e-01
-6.95112944e-01 8.48324478e-01 1.03328156e+00 9.74446461e-02
1.36641419e+00 -8.77578080e-01 8.44284415e-01 5.53780973e-01
5.40133297e-01 -1.48844469e+00 -1.06897876e-01 5.80132723e-01
5.30909181e-01 -9.64042306e-01 -3.73887420e-02 -9.71371457e-02
-3.03925216e-01 1.20278454e+00 3.17475021e-01 -1.85829304e-05
9.71658945e-01 4.43517834e-01 1.00635782e-01 5.31015694e-01
-4.21970248e-01 -4.56037298e-02 -3.30816098e-02 9.72411156e-01
6.19501173e-02 -2.47043017e-02 -3.20160419e-01 6.65994525e-01
-7.35519677e-02 7.36573637e-02 7.43735075e-01 1.09436059e+00
-9.59367573e-01 -1.11716759e+00 -1.01807976e+00 7.82041967e-01
-4.54933763e-01 -1.71436772e-01 1.80789828e-01 3.51797551e-01
3.98532420e-01 1.31333208e+00 -1.21831715e-01 -7.06093371e-01
5.92438519e-01 -3.99380565e-01 1.63085595e-01 -4.56254900e-01
-8.09960738e-02 2.31153339e-01 2.03626990e-01 2.24314462e-02
-1.73058778e-01 -3.34394574e-01 -1.28375292e+00 -1.57799587e-01
-7.34741926e-01 6.95164725e-02 7.26207435e-01 1.09000623e+00
3.76924306e-01 6.19968593e-01 1.20774961e+00 -7.38799214e-01
-1.08431470e+00 -9.23604071e-01 -7.43909240e-01 -3.63245457e-01
9.66035128e-01 -4.14640725e-01 -2.40326688e-01 2.44254246e-03] | [6.431744575500488, 1.4596154689788818] |
106f0375-58a4-4cda-aa15-4684b3e2d89d | a-systematic-and-analytical-review-of-the | 2003.04452 | null | http://arxiv.org/abs/2003.04452v2 | http://arxiv.org/pdf/2003.04452v2.pdf | A Systematic and Analytical Review of the Socioeconomic and Environmental Impact of the Deployed High-Speed Rail (HSR) Systems on the World | The installation of high-speed rail in the world during the last two decades
resulted in significant socioeconomic and environmental changes. The U.S. has
the longest rail network in the world, but the focus is on carrying a wide
variety of loads including coal, farm crops, industrial products, commercial
goods, and miscellaneous mixed shipments. Freight and passenger services in the
U.S. dates to 1970, with both carried out by private railway companies.
Railways were the main means of transport between cities from the late 19th
century through the middle of the 20th century. However, rapid growth in
production and improvements in technologies changed those dynamics. The fierce
competition for comfortability and pleasantness in passenger travel and the
proliferation of aviation services in the U.S. channeled federal and state
budgets towards motor vehicle infrastructure, which brought demand for
railroads to a halt in the 1950s. Presently, the U.S. has no high-speed trains,
aside from sections of Amtrak s Acela line in the Northeast Corridor that can
reach 150 mph for only 34 miles of its 457-mile span. The average speed between
New York and Boston is about 65 mph. On the other hand, China has the world s
fastest and largest high-speed rail network, with more than 19,000 miles, of
which the vast majority was built in the past decade. Japan s bullet trains can
reach nearly 200 miles per hour and dates to the 1960s. That system moved more
than 9 billion people without a single passenger casualty. In this systematic
review, we studied the effect of High-Speed Rail (HSR) on the U.S. and other
countries including France, Japan, Germany, Italy, and China in terms of energy
consumption, land use, economic development, travel behavior, time use, human
health, and quality of life. | [] | 2020-03-18 | null | null | null | null | ['miscellaneous'] | ['miscellaneous'] | [-5.40723264e-01 -2.11312696e-01 -5.96308708e-01 1.87631845e-01
-4.77833748e-01 -5.72270811e-01 4.89104271e-01 4.23437357e-02
-6.88337564e-01 9.72401917e-01 1.92211345e-01 -1.02393830e+00
-2.46761695e-01 -1.36202633e+00 -2.75226027e-01 -4.78923798e-01
2.03621924e-01 5.05695976e-02 8.15423802e-02 -9.93055820e-01
2.75006384e-01 8.64186943e-01 -1.10094774e+00 -7.82069206e-01
1.09050429e+00 4.99569565e-01 3.80320787e-01 7.93772042e-02
6.09477498e-02 1.61864683e-01 -1.21362567e-01 -2.05279157e-01
-4.58411016e-02 -2.93050677e-01 -7.74439037e-01 -1.29004002e-01
-4.27425951e-01 -2.92169452e-01 -7.76400268e-01 5.45209348e-01
4.37177382e-02 3.17649096e-01 2.58181959e-01 -1.26054192e+00
-3.73181373e-01 5.95181286e-01 -7.45778263e-01 5.31611033e-02
3.07224542e-01 2.01522231e-01 8.52851331e-01 -5.26661575e-01
2.01175854e-01 6.81763828e-01 4.37612891e-01 -6.15216345e-02
-1.11659980e+00 -8.73548806e-01 1.92096949e-01 -9.69572142e-02
-1.51764452e+00 -2.56796569e-01 2.98620373e-01 -3.19903821e-01
9.72814560e-01 4.31315750e-01 1.08919477e+00 1.65681928e-01
4.48876470e-01 9.97160152e-02 8.55551720e-01 -2.42096856e-01
-1.29593119e-01 -9.59006622e-02 -1.81575105e-01 5.20026267e-01
6.18851721e-01 2.68855765e-02 2.93834340e-02 2.78107613e-01
6.79595888e-01 -8.28106105e-02 -2.18664452e-01 4.26546216e-01
-1.21391106e+00 6.70565307e-01 6.54918611e-01 6.38366520e-01
-1.83096290e-01 1.65915430e-01 3.73708397e-01 2.14409992e-01
-2.40856204e-02 2.21973017e-01 -4.12870795e-01 -5.31643689e-01
-5.56837261e-01 1.56786487e-01 4.13901359e-01 7.16768563e-01
5.36330938e-01 7.75258243e-03 8.68588507e-01 6.66440427e-01
1.63768217e-01 1.04123080e+00 -1.28714398e-01 -7.10888803e-01
9.02799904e-01 2.90044099e-01 2.88667649e-01 -8.35450947e-01
-1.03748989e+00 -4.67582226e-01 -8.11164558e-01 -1.23115607e-01
9.60028827e-01 -5.73109806e-01 -6.42113388e-01 1.40475881e+00
-6.55600429e-02 -5.18598497e-01 -2.11190030e-01 8.46283734e-01
-1.12508021e-01 7.71771193e-01 7.78472051e-02 5.78329153e-02
1.29079545e+00 -7.68036842e-01 -5.01226604e-01 3.47019881e-02
7.63880908e-01 -6.83219850e-01 9.13041413e-01 1.50447726e-01
-1.35360622e+00 -4.24150378e-01 -6.54574573e-01 3.01953554e-01
-4.02608961e-01 -4.22464728e-01 4.69675779e-01 9.98800993e-01
-6.65146053e-01 5.41118741e-01 -1.01854277e+00 -5.29368877e-01
3.40486318e-01 1.23096764e-01 -6.12855136e-01 -3.52538258e-01
-1.40486622e+00 1.19570613e+00 -1.11988664e-01 2.51859188e-01
-2.83213049e-01 -4.59504277e-01 -9.28790092e-01 4.72948737e-02
1.32849619e-01 -4.14713860e-01 7.25837529e-01 -1.39194414e-01
-1.30454314e+00 6.35222569e-02 1.32986501e-01 1.50750056e-02
2.26538539e-01 1.01483695e-01 -1.19364178e+00 -7.82325491e-02
3.73684913e-01 1.06879741e-01 -5.27948201e-01 -6.99417293e-01
-1.38997078e+00 -3.38560730e-01 2.58732855e-01 4.78887707e-02
4.17789631e-02 1.35177180e-01 -7.84714296e-02 -1.83430985e-01
3.23093504e-01 -1.33141303e+00 -4.12835598e-01 -7.00046539e-01
-4.17181820e-01 -9.97074842e-02 4.26009268e-01 -5.34534693e-01
1.15633237e+00 -2.05398059e+00 -3.32907647e-01 5.78403234e-01
-5.39473712e-01 -5.21688163e-01 3.09147507e-01 9.33441639e-01
1.59757789e-02 2.11846426e-01 -1.61883637e-01 4.02786583e-01
1.07028864e-01 4.33809996e-01 1.06571577e-01 8.58619213e-01
-1.40285706e-02 4.83037233e-01 -1.22183168e+00 1.09951794e-01
2.25938603e-01 3.01020831e-01 -1.55238956e-01 -4.70481247e-01
8.16632748e-01 6.44753754e-01 -2.35175535e-01 7.32822120e-01
7.02535331e-01 1.38861477e-01 -6.34055957e-02 2.53986448e-01
-1.19043875e+00 5.46903133e-01 -7.29072630e-01 1.74475062e+00
-8.07866812e-01 7.01925814e-01 -1.24588847e-01 -6.39163256e-01
6.21956587e-01 3.39944750e-01 6.16153061e-01 -1.22866607e+00
3.25519294e-02 7.53911078e-01 3.03162724e-01 -6.41737998e-01
8.32410395e-01 -5.65246880e-01 -3.85634571e-01 2.64689684e-01
-9.32437420e-01 -3.55096072e-01 5.69744170e-01 2.22618580e-01
8.35424364e-01 6.74248934e-02 -7.49490410e-02 -5.85828722e-01
2.74261206e-01 4.52971846e-01 5.92435181e-01 -2.05602124e-01
3.06473654e-02 -1.05354019e-01 3.89580399e-01 -1.59673870e-01
-1.06182683e+00 -1.40310001e+00 -6.51201069e-01 8.85284603e-01
4.91621941e-01 1.55764058e-01 -1.74414501e-01 -1.64873198e-01
2.27814183e-01 9.92005587e-01 -6.64927736e-02 9.28303674e-02
-6.49060667e-01 -4.92277056e-01 6.33036435e-01 5.39049804e-01
8.00333023e-01 -2.95261979e-01 -3.83270741e-01 2.56835490e-01
-4.69093323e-01 -7.59162664e-01 -5.05735099e-01 7.81312585e-02
-5.94341993e-01 -7.83794582e-01 -9.90460932e-01 -5.79389751e-01
8.39873254e-01 3.64250183e-01 7.23927498e-01 -1.22594140e-01
-5.73313944e-02 -8.60709921e-02 -4.99610417e-02 -2.86444098e-01
2.43750717e-02 3.53018641e-01 1.19005926e-01 -5.78994930e-01
-1.76351462e-02 -3.59111220e-01 -7.20136940e-01 8.08697283e-01
-3.92963529e-01 -4.34010923e-02 1.71037599e-01 2.94203907e-01
-5.00568189e-02 7.66469300e-01 1.27755654e+00 -3.53483766e-01
-9.30351540e-02 -1.06824493e+00 -4.77482080e-01 -3.23411733e-01
-5.26475370e-01 -5.40842772e-01 5.09780586e-01 3.52918983e-01
-1.15666234e+00 -3.03461879e-01 -6.34516776e-01 9.28774416e-01
-2.79525757e-01 8.25864971e-01 -4.54286784e-02 1.94403172e-01
2.36843273e-01 -1.02785133e-01 -2.25440934e-01 -4.32390690e-01
1.35221645e-01 8.47632885e-01 7.24196851e-01 -4.93370205e-01
1.23514867e+00 6.42312109e-01 2.75942773e-01 -1.10865581e+00
-2.27437243e-01 -4.57316786e-01 -5.59510469e-01 -5.76678932e-01
9.39489126e-01 -9.86085117e-01 -9.00628269e-01 4.62999433e-01
-5.06383479e-01 -3.76915962e-01 -7.00610317e-03 1.02274621e+00
-1.60135016e-01 -7.93996975e-02 -7.40895927e-01 -8.65377724e-01
3.76296848e-01 -9.24657762e-01 1.56609967e-01 3.41748685e-01
-1.81814119e-01 -1.03258526e+00 -7.36207068e-02 3.31355184e-01
6.09530985e-01 5.31147599e-01 1.01400256e+00 2.56005108e-01
-4.29668039e-01 -7.39453807e-02 -4.22669262e-01 1.25294551e-01
7.41622210e-01 -1.52031109e-01 -2.00326279e-01 -4.28411216e-01
-7.38091767e-01 1.20458029e-01 6.03951454e-01 3.80402803e-01
3.65488619e-01 2.70227075e-01 -6.03251159e-01 8.02112520e-02
1.64885819e+00 8.84560347e-01 9.96636868e-01 8.83125246e-01
3.35186005e-01 4.92581636e-01 8.78082514e-01 8.16443712e-02
6.97835267e-01 1.17290832e-01 5.58496296e-01 -4.99826431e-01
3.38094980e-01 -1.68090165e-01 1.05943702e-01 9.61584151e-01
-7.65158832e-01 -1.13417946e-01 -1.09331143e+00 8.49329293e-01
-1.06385911e+00 -8.63021553e-01 -6.47750199e-01 2.46995044e+00
4.53814834e-01 5.90144098e-01 3.04155499e-01 3.48417401e-01
4.45755601e-01 -5.20163476e-01 -2.53595501e-01 -3.66925567e-01
2.77688324e-01 -1.31058469e-01 1.34578621e+00 4.48566288e-01
-3.91930699e-01 1.79128394e-01 6.01234961e+00 2.96747357e-01
-9.08915699e-01 -2.04481289e-01 3.53247672e-01 1.53570488e-01
-7.76095331e-01 2.40945265e-01 -7.14175224e-01 4.44416761e-01
1.15333545e+00 -4.83401448e-01 2.53174663e-01 1.60687461e-01
8.28718364e-01 -5.15235245e-01 -2.94645548e-01 3.28708738e-01
-6.33406103e-01 -1.16759777e+00 -6.54948950e-01 4.93183851e-01
8.79765034e-01 2.80265182e-01 8.08655173e-02 1.48991510e-01
1.29831746e-01 -8.71103466e-01 8.77693713e-01 3.85575145e-01
1.09379768e+00 -1.80884981e+00 5.11136711e-01 3.95617634e-01
-1.45707798e+00 -2.54031450e-01 2.38430858e-01 -5.33521950e-01
8.65550697e-01 5.66102207e-01 -4.48547639e-02 7.49175549e-01
5.89710414e-01 3.76365930e-01 3.05124313e-01 8.93344104e-01
-1.71660542e-01 5.73179543e-01 -4.02647018e-01 1.54138282e-01
8.41962278e-01 -5.17413974e-01 2.21071124e-01 8.16865027e-01
4.57049251e-01 6.77145347e-02 -4.74094339e-02 9.93333086e-02
-3.04339919e-02 4.50318716e-02 -4.42030638e-01 1.07471108e-01
5.54283679e-01 8.60152304e-01 -9.91221607e-01 6.02097921e-02
-9.53095138e-01 1.80761501e-01 -3.41436058e-01 3.24243933e-01
-1.31628895e+00 -1.22854292e+00 9.33455765e-01 8.27799857e-01
-4.53906983e-01 -7.35325992e-01 -5.57461977e-02 -5.23705840e-01
-5.09035707e-01 -2.18195245e-01 2.74060462e-02 -2.64269114e-01
-8.10158551e-01 2.46232226e-01 6.90854564e-02 -9.30594146e-01
2.31148243e-01 -4.60684001e-02 -3.54444832e-01 1.32762182e+00
-1.84255183e+00 -7.54312992e-01 1.85672287e-02 3.31640065e-01
3.74557853e-01 1.05515696e-01 6.13101304e-01 8.99979651e-01
-7.76491642e-01 3.58898550e-01 5.94151020e-01 1.65156990e-01
3.33210796e-01 -7.58891582e-01 9.22749937e-01 5.58457077e-01
-8.70280027e-01 7.02111959e-01 5.01570702e-01 -6.21540606e-01
-1.21774018e+00 -8.72753441e-01 1.17205822e+00 3.15530092e-01
1.15876651e+00 6.26125410e-02 -3.21803778e-01 9.36317623e-01
1.58958226e-01 -7.61706650e-01 5.24734437e-01 2.89570600e-01
1.14447922e-01 -3.52840382e-03 -1.10745704e+00 7.83685982e-01
7.18419552e-01 -4.08871740e-01 5.50328568e-03 2.56631821e-01
5.56695879e-01 -1.16027690e-01 -1.18218374e+00 -5.66983484e-02
1.04435003e+00 -4.45131153e-01 6.53154314e-01 -2.44987831e-01
2.55352527e-01 -2.66969055e-01 -4.42768782e-02 -1.57595241e+00
-7.94396579e-01 -6.39016092e-01 1.12229729e+00 1.16657579e+00
8.73818696e-01 -1.30300534e+00 4.70332474e-01 8.62643898e-01
-7.90000737e-01 -7.21912861e-01 -9.08782244e-01 -1.08668649e+00
4.35659975e-01 -4.46884036e-01 9.86507714e-01 9.34341669e-01
4.22315508e-01 1.20215297e-01 -3.95399928e-02 1.09052062e-01
5.77982724e-01 -1.47251725e-01 5.76395929e-01 -1.03639340e+00
2.51687616e-01 -5.00694931e-01 -2.66174704e-01 -1.03178561e+00
-2.60678738e-01 -8.68987083e-01 -3.97745013e-01 -2.19547868e+00
-1.44727841e-01 -9.78327572e-01 -5.57924770e-02 4.25893635e-01
4.85112995e-01 2.13772818e-01 1.32540166e-01 -4.25437354e-02
6.37973905e-01 3.29103172e-02 1.65478814e+00 -1.19033642e-01
-3.41625899e-01 3.10843587e-01 -7.56744444e-01 6.59399211e-01
1.17315865e+00 -3.48601699e-01 -3.97632957e-01 -4.92208868e-01
7.46155083e-01 2.66725570e-01 -2.03847811e-01 -7.71777809e-01
1.12818561e-01 -7.30139434e-01 8.01492557e-02 -7.16561973e-01
9.09800231e-02 -1.26004410e+00 6.54733002e-01 9.75548446e-01
3.18089455e-01 2.77006835e-01 2.37784758e-01 2.27758095e-01
1.02568030e-01 1.55794665e-01 1.03280568e+00 1.12487696e-01
-2.48664394e-01 -3.18784565e-02 -1.09406328e+00 -2.28133634e-01
1.18251622e+00 -3.85606915e-01 -6.43087447e-01 -9.36645344e-02
-6.12325430e-01 6.05631053e-01 5.73512256e-01 2.90964365e-01
1.25268459e-01 -1.29678917e+00 -7.07809269e-01 5.95219843e-02
-2.02610806e-01 -3.60694557e-01 4.96659756e-01 1.22126758e+00
-8.87325287e-01 6.55563056e-01 -6.08928204e-01 -2.60826703e-02
-5.70444882e-01 2.44852722e-01 2.18316987e-01 6.23130910e-02
-8.85590374e-01 3.76934290e-01 -7.96671882e-02 9.51293707e-02
-2.37822279e-01 -5.68787873e-01 -1.46907168e-02 1.33706689e-01
4.76761937e-01 1.22305310e+00 -1.68169752e-01 -9.58329380e-01
-5.04256546e-01 5.25403082e-01 3.54047298e-01 -1.55016944e-01
1.13815689e+00 -6.33045554e-01 -1.47218019e-01 6.80613041e-01
1.19285774e+00 3.46683860e-01 -8.44212711e-01 5.30428171e-01
-3.29813510e-01 -6.97990894e-01 1.84194610e-01 -6.24029696e-01
-1.22577155e+00 7.62385353e-02 -4.39427793e-02 4.62856680e-01
1.15244448e+00 -1.89407632e-01 1.54376888e+00 -1.05559684e-01
7.70988643e-01 -1.22561264e+00 -5.96082270e-01 3.76119822e-01
3.47314626e-01 -5.51542819e-01 -2.45807394e-02 -2.20003113e-01
-2.21555635e-01 9.83991921e-01 -5.51406927e-02 1.04738593e-01
9.19413745e-01 3.49292904e-01 -2.83330470e-01 -1.13857232e-01
-1.24655485e-01 -3.02137971e-01 -5.53367138e-01 2.18570396e-01
5.04762590e-01 3.59934211e-01 -5.76528966e-01 7.40472926e-03
-2.65043944e-01 -1.11127637e-01 8.50912392e-01 1.10427916e+00
-8.45938027e-01 -7.94521332e-01 -5.71161747e-01 5.72785556e-01
-3.51030439e-01 1.15478255e-01 9.22229052e-01 1.29770684e+00
3.23274046e-01 1.38541043e+00 4.19945866e-01 -8.29204991e-02
9.39030230e-01 -1.38385877e-01 2.43792415e-01 -3.10393702e-02
-4.57446814e-01 -2.24977378e-02 6.16532862e-01 -1.51368886e-01
-2.03013554e-01 -9.26382840e-01 -2.10490012e+00 -1.17043614e+00
-7.30417073e-01 5.86405218e-01 1.21512175e+00 8.33445728e-01
-4.69492882e-01 6.18258953e-01 1.08364618e+00 -6.40312374e-01
9.63479653e-02 -6.07930839e-01 -1.44079483e+00 -4.01487917e-01
2.66493529e-01 -5.78277409e-01 -5.08438349e-01 -1.94366857e-01] | [6.051041126251221, 2.060622215270996] |
ca1621f4-a99f-4327-a1cc-32a507571af2 | regular-decision-processes-for-grid-worlds | 2111.03647 | null | https://arxiv.org/abs/2111.03647v2 | https://arxiv.org/pdf/2111.03647v2.pdf | Regular Decision Processes for Grid Worlds | Markov decision processes are typically used for sequential decision making under uncertainty. For many aspects however, ranging from constrained or safe specifications to various kinds of temporal (non-Markovian) dependencies in task and reward structures, extensions are needed. To that end, in recent years interest has grown into combinations of reinforcement learning and temporal logic, that is, combinations of flexible behavior learning methods with robust verification and guarantees. In this paper we describe an experimental investigation of the recently introduced regular decision processes that support both non-Markovian reward functions as well as transition functions. In particular, we provide a tool chain for regular decision processes, algorithmic extensions relating to online, incremental learning, an empirical evaluation of model-free and model-based solution algorithms, and applications in regular, but non-Markovian, grid worlds. | ['Martijn van Otterlo', 'Nicky Lenaers'] | 2021-11-05 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 9.70112830e-02 2.51668304e-01 -3.17472667e-01 -3.66158158e-01
-4.02608335e-01 -7.42275357e-01 7.24669039e-01 2.44541407e-01
-4.61907238e-01 1.11922431e+00 -1.81051508e-01 -5.93594611e-01
-7.24131763e-01 -6.76234841e-01 -2.91441739e-01 -5.41008174e-01
-6.74705148e-01 7.70390570e-01 6.42389178e-01 -9.43803340e-02
2.15814188e-01 6.19714379e-01 -1.53279769e+00 -5.16340360e-02
5.08851051e-01 9.37338173e-01 1.98746119e-02 7.60050237e-01
1.38310164e-01 8.85744095e-01 -2.07651466e-01 -1.45600781e-01
1.56411856e-01 -1.90027833e-01 -8.48482132e-01 1.41742304e-01
-5.97590864e-01 -3.01263422e-01 8.17765966e-02 9.72513378e-01
1.71631068e-01 3.29161763e-01 6.06648028e-01 -1.53487551e+00
-2.46924497e-02 1.09111679e+00 -1.87012807e-01 -1.53094769e-01
5.06663859e-01 7.25580156e-01 6.82191730e-01 5.60476035e-02
5.02177477e-01 1.41734242e+00 2.97600985e-01 7.56226718e-01
-1.43485928e+00 -3.31442922e-01 5.85813642e-01 3.10683638e-01
-1.02920747e+00 -1.21712953e-01 3.25632453e-01 -4.53305125e-01
1.13016915e+00 8.42312351e-02 7.15917170e-01 1.29673243e+00
7.70170867e-01 8.56072903e-01 1.79004776e+00 -5.09453177e-01
8.20311069e-01 1.68218940e-01 7.38206804e-02 2.74654716e-01
2.89027274e-01 7.85046637e-01 -2.33463809e-01 -1.84079841e-01
6.57450199e-01 -2.33804703e-01 2.54236579e-01 -7.10780203e-01
-1.03675556e+00 6.62088096e-01 -4.96783882e-01 2.16743723e-01
-4.90871698e-01 3.80484551e-01 4.59838301e-01 6.22736275e-01
-2.11671740e-01 4.52531606e-01 -6.50490463e-01 -4.06893492e-01
-6.68502152e-01 6.68372750e-01 1.26269984e+00 9.22082245e-01
5.53013682e-01 1.35722995e-01 -5.68743169e-01 -1.67799979e-01
5.32097816e-01 4.18106079e-01 2.76449144e-01 -1.28109539e+00
1.39680296e-01 1.27101257e-01 7.07515955e-01 -1.97781026e-01
-6.28220737e-01 -6.12934381e-02 -4.92088795e-01 7.74373293e-01
4.86208797e-01 -4.41257060e-01 -7.41840243e-01 1.66371667e+00
3.77437800e-01 2.09218767e-02 3.14023435e-01 5.41171968e-01
-4.63299513e-01 2.61782289e-01 -5.59427254e-02 -7.42555678e-01
1.01158345e+00 -3.40711057e-01 -8.80666852e-01 2.02498566e-02
3.44731092e-01 -2.95189679e-01 8.18405688e-01 8.78350735e-01
-1.06775510e+00 -1.36157453e-01 -8.76924396e-01 8.10254157e-01
9.36063100e-03 -6.53545558e-01 6.97960556e-01 8.20667624e-01
-1.00434613e+00 9.08188224e-01 -1.32109404e+00 -2.87288595e-02
-7.97097012e-03 5.68096399e-01 8.74920115e-02 8.94538164e-02
-1.27045679e+00 1.16880143e+00 7.16788113e-01 -2.08451692e-02
-1.52667689e+00 -6.95532784e-02 -7.32396364e-01 -7.48745725e-02
1.08499873e+00 -3.45248759e-01 1.87712848e+00 -7.78992474e-01
-2.14423633e+00 3.09614837e-01 4.15723443e-01 -9.89369512e-01
9.07626987e-01 -4.74455394e-02 -5.22006869e-01 -1.34763852e-01
-2.58742362e-01 2.05212578e-01 1.09205461e+00 -9.05657172e-01
-6.60655856e-01 -1.66410357e-01 4.25202876e-01 -4.62910719e-03
3.28067869e-01 7.80328512e-02 3.72218907e-01 -8.58848765e-02
-3.00294340e-01 -1.23831964e+00 -7.70608962e-01 -4.81584042e-01
-2.60117769e-01 -3.43198270e-01 3.76486540e-01 -1.35133535e-01
1.02635527e+00 -1.84065437e+00 2.82302946e-01 3.81566823e-01
-3.60725760e-01 -2.24821910e-01 1.22595847e-01 6.09141886e-01
2.39266992e-01 -9.50361416e-02 -2.27412313e-01 -5.39453588e-02
5.26518524e-01 5.82496107e-01 -4.68626738e-01 4.11359996e-01
2.40103260e-01 6.23480976e-01 -1.00214672e+00 -5.15987515e-01
2.78644621e-01 -3.38111699e-01 -3.40379119e-01 1.46874234e-01
-9.55896139e-01 7.42577434e-01 -7.79424191e-01 5.84025383e-01
1.51070863e-01 3.09369504e-01 5.59907913e-01 7.76026964e-01
-3.85763228e-01 1.94396123e-01 -1.70795155e+00 1.42504811e+00
-4.71132815e-01 8.91665220e-02 1.35785967e-01 -7.28753448e-01
6.51687920e-01 4.42129076e-01 5.31593144e-01 -5.40317595e-01
1.18402481e-01 1.69993982e-01 2.08531469e-01 -3.55469346e-01
5.63818693e-01 -4.79263216e-01 -4.02661562e-01 7.34081924e-01
-3.97180803e-02 -4.67581093e-01 3.35135013e-01 -4.08446312e-01
1.23806179e+00 8.66928101e-01 6.48927033e-01 -3.67316276e-01
4.26546037e-01 9.74396430e-03 7.81025469e-01 9.36132371e-01
-3.20315927e-01 -1.04137942e-01 9.57482696e-01 -1.59056231e-01
-7.68254101e-01 -8.64892840e-01 -1.11269526e-01 8.12180340e-01
8.79119486e-02 -2.81351149e-01 -6.28985524e-01 -5.61894238e-01
-3.94878201e-02 1.15268981e+00 -4.69590753e-01 -2.72306412e-01
-4.64380890e-01 -2.14119077e-01 3.18067282e-01 3.88321131e-01
1.78938285e-01 -1.24129462e+00 -1.21859252e+00 6.38693750e-01
4.97796893e-01 -9.00467873e-01 -2.16057915e-02 7.32048154e-01
-1.13195968e+00 -9.40600753e-01 -1.12789087e-01 1.04616322e-01
4.34000008e-02 -4.79919165e-01 9.09652412e-01 -3.88056040e-01
1.45138115e-01 7.77142942e-01 -2.99525321e-01 -5.80789804e-01
-6.47533476e-01 -2.43953004e-01 3.13953459e-01 -6.24205619e-02
-1.11044861e-01 -4.49468404e-01 -1.07894264e-01 4.98770595e-01
-1.04266202e+00 -9.50946733e-02 5.30558348e-01 6.83549404e-01
7.65613616e-01 2.76243150e-01 2.63478935e-01 -6.81087911e-01
9.11574662e-01 -2.43110061e-01 -1.21836376e+00 5.93620837e-01
-1.03227794e+00 7.33240843e-01 4.85739589e-01 -6.85340524e-01
-1.36899078e+00 2.59091944e-01 1.76613718e-01 -3.30666989e-01
-4.50848430e-01 4.28452849e-01 -2.14293584e-01 2.78429121e-01
7.28235722e-01 1.99549988e-01 1.77684754e-01 -1.16352979e-02
2.64652997e-01 3.37126613e-01 -2.00696308e-02 -1.13552022e+00
6.52705729e-01 2.19327450e-01 5.48890829e-01 -2.29503870e-01
-4.03829932e-01 9.16720852e-02 -4.60840344e-01 -2.56299794e-01
7.00596273e-01 -3.96331877e-01 -1.10286975e+00 4.12023276e-01
-7.41910696e-01 -1.02360094e+00 -8.26548100e-01 5.45312107e-01
-1.44043720e+00 1.46523356e-01 -5.38126945e-01 -1.59241045e+00
1.76549405e-01 -1.22627306e+00 6.24360621e-01 2.46413425e-01
-3.46560776e-01 -9.05329943e-01 3.52424800e-01 -3.49039137e-01
2.76479214e-01 3.38956624e-01 7.44363248e-01 -6.69468999e-01
-8.15279484e-01 -1.88915744e-01 5.22855759e-01 1.72606766e-01
-2.66408771e-01 1.21666111e-01 -6.57601595e-01 -3.69947314e-01
9.55003202e-02 -4.72178519e-01 2.54850835e-01 3.43347341e-01
7.10823417e-01 -3.94416720e-01 -2.04009771e-01 -1.47840321e-01
1.24900651e+00 6.61525130e-01 4.84825045e-01 4.41629350e-01
-1.49912432e-01 7.55168140e-01 1.18516636e+00 8.92655373e-01
2.60657430e-01 6.82375848e-01 6.74259365e-01 8.22257519e-01
5.17761111e-01 -2.24369943e-01 7.54974484e-01 -6.37367815e-02
-2.09280953e-01 5.68343513e-02 -8.60274017e-01 2.03404114e-01
-2.23581576e+00 -1.05968845e+00 6.06701300e-02 2.54775786e+00
1.00822389e+00 6.28089070e-01 3.44293773e-01 5.86601011e-02
5.67684710e-01 -2.44570419e-01 -8.91710460e-01 -7.87632287e-01
2.91671067e-01 1.94913834e-01 5.63408017e-01 5.84973931e-01
-9.07188237e-01 8.03967118e-01 6.72474146e+00 7.85555482e-01
-7.59441257e-01 8.04018080e-02 2.49398738e-01 1.81802530e-02
-4.10270870e-01 3.47221553e-01 -9.77001131e-01 2.16221675e-01
1.30784678e+00 -3.35291743e-01 6.53031886e-01 9.81570184e-01
3.32772255e-01 -5.36890566e-01 -1.35155451e+00 5.86764395e-01
-7.27753043e-01 -8.55445027e-01 -5.86576641e-01 1.37183160e-01
7.33797073e-01 -3.73174161e-01 -1.97725937e-01 7.05109298e-01
1.09492683e+00 -8.92219245e-01 1.22180736e+00 6.47315323e-01
5.29981554e-01 -9.35124159e-01 7.70539999e-01 7.54633904e-01
-8.45135927e-01 -5.02373695e-01 5.02937324e-02 -4.52624947e-01
3.09819579e-01 4.02937949e-01 -4.77765024e-01 7.15885282e-01
4.61775899e-01 4.63221878e-01 -7.65119120e-02 9.64360058e-01
-6.16968632e-01 5.02611220e-01 -3.27404469e-01 -3.31500828e-01
3.67624730e-01 -1.33333758e-01 5.87912679e-01 8.45184505e-01
1.73428968e-01 -1.61898881e-01 4.91398096e-01 6.83640063e-01
8.93813968e-01 -4.62530583e-01 -6.50651157e-01 2.12976579e-02
1.70095131e-01 9.99970794e-01 -9.76514280e-01 -1.68635905e-01
-6.29559532e-02 4.92019415e-01 1.16415776e-01 3.62137377e-01
-9.93391812e-01 3.07719447e-02 6.83828771e-01 6.72137365e-02
1.38260603e-01 -5.82842529e-01 -1.54782459e-01 -9.47335541e-01
-1.33867025e-01 -9.76759195e-01 5.15987515e-01 -3.59598607e-01
-9.94121492e-01 2.75607973e-01 7.02486694e-01 -1.12933850e+00
-9.94452775e-01 -4.56601381e-01 -3.94605100e-01 6.38246417e-01
-1.32614815e+00 -7.50689983e-01 4.86623794e-01 7.44414687e-01
3.82255942e-01 -4.45453860e-02 8.03515494e-01 -4.90573376e-01
-3.02858472e-01 -7.82627463e-02 -1.62975192e-02 -6.68342590e-01
3.13626379e-01 -1.39971673e+00 -1.73745248e-02 7.39495099e-01
-1.68321609e-01 3.21948379e-01 1.10264432e+00 -6.86425745e-01
-1.53998053e+00 -7.15944290e-01 2.62297302e-01 -1.94247261e-01
9.94290411e-01 -1.68616667e-01 -6.67640090e-01 8.01210761e-01
2.05462635e-01 -3.55627239e-01 2.02772781e-01 1.29616380e-01
1.21149890e-01 -1.02679856e-01 -1.21572137e+00 8.02773774e-01
8.42977762e-01 -2.45210394e-01 -5.23356020e-01 1.25726387e-01
7.88962781e-01 -5.16893387e-01 -8.36232126e-01 4.08318222e-01
4.08153385e-01 -9.36152220e-01 4.58145976e-01 -5.66003323e-01
-1.95455283e-01 -4.13852125e-01 -1.22729361e-01 -1.15572834e+00
-3.46115589e-01 -1.30026984e+00 -5.81540227e-01 8.76586556e-01
2.88750887e-01 -5.87274075e-01 5.37580252e-01 9.83543694e-01
-1.48837760e-01 -7.15021789e-01 -1.28695095e+00 -1.27420604e+00
5.45981191e-02 -8.82024705e-01 5.37277877e-01 3.46481472e-01
2.22123653e-01 -2.21990850e-02 -4.37190354e-01 -1.07502386e-01
6.16339982e-01 2.94005275e-01 3.54651749e-01 -1.00648654e+00
-8.10772359e-01 -4.53024089e-01 -1.71476051e-01 -6.43107057e-01
2.96759039e-01 -4.10002351e-01 4.23060298e-01 -1.22060156e+00
-2.60152429e-01 -5.22311985e-01 -2.48398229e-01 5.34645379e-01
4.56452549e-01 -7.87177205e-01 9.11554471e-02 -3.90346125e-02
-9.32015777e-01 5.15835524e-01 1.28270769e+00 9.34598818e-02
-5.02995431e-01 7.52845228e-01 -1.90896571e-01 6.04647338e-01
1.10703790e+00 -4.18333739e-01 -6.12392366e-01 2.08843842e-01
6.02401197e-01 7.14080930e-01 1.80502981e-01 -1.06007254e+00
1.18198238e-01 -1.05119169e+00 -2.94140041e-01 -4.39522445e-01
4.17469032e-02 -9.88951385e-01 4.69370604e-01 9.27053750e-01
-6.28935635e-01 -2.02937368e-02 1.75393239e-01 8.82163644e-01
5.26347347e-02 -6.08618736e-01 6.13597393e-01 -2.40647912e-01
-8.53972435e-01 1.90899223e-01 -1.02873945e+00 -1.63361609e-01
1.62626946e+00 -2.75784671e-01 1.81646913e-01 -3.45134169e-01
-1.24603856e+00 4.03851926e-01 2.81090051e-01 1.16679400e-01
4.69637513e-01 -7.98488200e-01 -2.92235404e-01 -2.36448824e-01
-8.92120749e-02 -6.13681190e-02 1.17690213e-01 9.79753077e-01
7.51783550e-02 5.91885030e-01 -5.30454636e-01 -3.15687507e-01
-9.09096062e-01 9.05168772e-01 3.50050002e-01 -8.80750656e-01
-3.55521202e-01 7.47805461e-02 -4.14233446e-01 -2.23661751e-01
3.69450659e-01 -7.31543005e-01 -1.44455761e-01 1.70085020e-02
3.22885066e-01 4.93727446e-01 -1.72798350e-01 3.00594836e-01
-3.05397719e-01 9.70489830e-02 2.80625582e-01 -7.50289261e-01
1.00696635e+00 -1.96961269e-01 5.16745709e-02 1.00188196e+00
-1.03029490e-01 -5.70347250e-01 -1.87513709e+00 -2.30177656e-01
5.97221196e-01 -7.99405277e-02 -2.83650219e-01 -8.91271889e-01
-2.41102427e-01 7.29376256e-01 5.89699566e-01 5.98438859e-01
9.36301410e-01 -1.83191717e-01 5.03786057e-02 7.07060456e-01
1.25917304e+00 -1.44310033e+00 -3.02715693e-02 7.75296748e-01
8.08892727e-01 -7.92188585e-01 -1.22860357e-01 1.11805648e-01
-9.08464611e-01 1.13609743e+00 5.65009952e-01 -8.03019945e-03
5.62198579e-01 7.05901444e-01 -4.14085001e-01 3.52557153e-01
-1.23652685e+00 -5.12332737e-01 -2.90680438e-01 9.61657763e-01
-8.05032030e-02 4.22006309e-01 -3.27421367e-01 5.10478258e-01
2.01691046e-01 3.64335388e-01 6.75300717e-01 1.30447173e+00
-5.73122382e-01 -1.49487507e+00 -5.75987995e-01 3.22668016e-01
-3.47485512e-01 3.15428227e-01 4.21043448e-02 8.07543635e-01
-9.11750495e-02 1.05787897e+00 -2.26303235e-01 -5.09299226e-02
3.28510076e-01 2.22288772e-01 8.64056051e-01 -6.86948299e-01
-6.25941038e-01 1.26806140e-01 3.60486656e-01 -8.19620490e-01
-2.58823156e-01 -1.19414198e+00 -1.36777711e+00 -9.02666152e-02
-1.13876037e-01 6.88595697e-02 6.18638396e-01 1.15475893e+00
-9.56362188e-02 5.66684604e-01 4.89013910e-01 -8.23517144e-01
-1.40529311e+00 -8.43434155e-01 -9.69811320e-01 -2.38147050e-01
1.25983041e-02 -7.05073118e-01 1.24242892e-02 -2.87111819e-01] | [4.415437698364258, 2.1676557064056396] |
7a54a386-c975-474d-8802-76f572204310 | simple-prediction-of-immiscible-metal | 2101.12343 | null | https://arxiv.org/abs/2101.12343v1 | https://arxiv.org/pdf/2101.12343v1.pdf | Simple prediction of immiscible metal alloying based on metastability analysis | It has been known that even though two elemental metals, $X$ and $Y$, are immiscible, they can form alloys on surfaces of other metal $Z$. In order to understand such surface alloying of immiscible metals, we study the energetic stability of binary alloys, $XZ$ and $YZ$, in several structures with various coordination numbers (CNs). By analyzing the formation energy modified to enhance the subtle energy difference between metastable structures, we find that $XZ$ and $YZ$ with B2-type structure (CN$=$8) become energetically stable when the $X$ and $Y$ metals form an alloy on the $Z$ metal surface. This is consistent with the experimental results for Pb-Sn alloys on metal surfaces such as Rh(111) and Ru(0001). Some suitable metal substrates are also predicted to form Pb-Sn alloys. | ['Jun Onoe', 'Junji Yuhara', 'Shota Ono'] | 2021-01-29 | null | null | null | null | ['formation-energy'] | ['miscellaneous'] | [ 1.88766588e-02 4.72055048e-01 -2.41933659e-01 3.71596813e-02
-1.82331368e-01 1.44839302e-01 3.69408697e-01 -2.82009870e-01
-7.04070702e-02 1.06111538e+00 -4.66510862e-01 -1.50460243e-01
1.35758370e-01 -1.39172363e+00 -6.39909565e-01 -1.13218653e+00
1.23802729e-01 7.32114136e-01 6.33928120e-01 -7.91382611e-01
1.44120857e-01 2.97084361e-01 -2.02110338e+00 2.78714269e-01
7.75796175e-01 1.18728149e+00 5.50616443e-01 1.93586666e-02
3.73435803e-02 2.29269922e-01 -1.30378067e-01 -5.09621322e-01
3.05194616e-01 -4.10124630e-01 -7.57034719e-01 -7.57784545e-01
-2.69645244e-01 5.43793440e-01 -3.07810903e-01 1.11129940e+00
-2.40437746e-01 2.57417291e-01 1.35862315e+00 -8.77218783e-01
-6.93310082e-01 1.05856180e+00 -7.63605893e-01 -3.44142824e-01
2.39663661e-01 8.82192627e-02 1.09936190e+00 -8.90629113e-01
7.63231277e-01 1.15399015e+00 5.48107743e-01 1.22510481e+00
-3.00714523e-01 -9.22706902e-01 -1.26435712e-01 2.29418993e-01
-1.69861794e+00 -4.52470928e-01 8.21193874e-01 -3.09514463e-01
1.74966431e+00 6.61091506e-01 2.36424580e-01 1.78179517e-01
8.17268848e-01 1.19109927e-02 8.97711456e-01 -7.41311610e-01
-9.82382074e-02 2.37434078e-02 3.99162531e-01 5.40025771e-01
9.82055426e-01 2.38304898e-01 -1.95846274e-01 -2.69744068e-01
1.24758422e-01 -1.01894595e-01 -2.54626945e-02 1.43080071e-01
-4.47974771e-01 1.21082447e-01 4.54292625e-01 6.77969635e-01
-4.19828087e-01 5.48654497e-01 -4.66361791e-02 1.52690811e-02
7.85001591e-02 4.06775087e-01 -4.16286916e-01 -2.45839551e-01
-1.73289225e-01 4.58292753e-01 4.33827132e-01 8.07491541e-01
9.86468494e-01 -5.56878233e-03 7.49262393e-01 9.57262278e-01
6.02422655e-01 6.70257866e-01 1.64415345e-01 -5.02665699e-01
2.70608336e-01 6.46014750e-01 4.43783641e-01 -3.97385955e-01
-1.05898984e-01 6.47655189e-01 -4.38408047e-01 3.99431139e-01
-7.64921913e-03 2.05469310e-01 -5.28813779e-01 1.54563999e+00
3.40219975e-01 -4.49646741e-01 6.12630069e-01 5.38737833e-01
1.10649884e+00 9.42262948e-01 1.62082747e-01 -6.51957691e-01
1.24434912e+00 -7.20898688e-01 -5.68486214e-01 1.78969383e-01
5.31237662e-01 -5.93618810e-01 1.01713324e+00 1.08750381e-01
-1.82879388e+00 -3.15248758e-01 -1.32307851e+00 2.88992137e-01
-3.03587675e-01 -6.59132779e-01 4.00217026e-01 8.95633161e-01
-6.20718420e-01 9.15673792e-01 -1.10986340e+00 -3.22964758e-01
-3.52180153e-01 5.90548575e-01 -1.58187170e-02 -1.25349179e-01
-1.45922554e+00 9.42201436e-01 3.83275412e-02 6.36321232e-02
-6.27950132e-01 -4.11879063e-01 -3.61098051e-01 -4.81469750e-01
9.22879800e-02 -2.79575169e-01 9.40416515e-01 -4.56635594e-01
-1.57800877e+00 9.75254297e-01 -5.09672642e-01 1.81153849e-01
2.69311015e-02 2.65035123e-01 -8.57288539e-01 9.60039273e-02
2.90254176e-01 1.72620580e-01 1.60483003e-01 -1.46297991e+00
-1.87283561e-01 -2.45526239e-01 -3.02811950e-01 2.51054227e-01
2.04060152e-01 3.56213719e-01 2.87869185e-01 -1.91563770e-01
7.96783268e-01 -9.34293985e-01 -4.00288254e-01 -7.73761809e-01
-6.90879583e-01 -8.65668833e-01 8.72602880e-01 1.27763212e-01
8.18215430e-01 -1.97749746e+00 -1.80989370e-01 7.58544207e-01
1.85283005e-01 -3.47022146e-01 5.83752513e-01 7.98827827e-01
-3.75583887e-01 3.36375952e-01 -3.64410311e-01 3.51505637e-01
1.01411378e-03 -7.75553063e-02 3.74049187e-01 5.96645653e-01
-3.43245238e-01 7.77597249e-01 -4.65776056e-01 -5.48475049e-02
-1.26551628e-01 1.53579608e-01 -5.97520828e-01 -2.86954671e-01
-6.27427578e-01 -2.21014723e-01 -4.73615199e-01 8.96754444e-01
9.85585034e-01 -2.33843908e-01 2.23684728e-01 -2.12877437e-01
-7.11585581e-01 5.36212683e-01 -1.06899440e+00 6.30662680e-01
1.37004420e-01 -3.16223688e-02 4.10136342e-01 -8.36507499e-01
1.07948112e+00 5.08879662e-01 8.73123646e-01 -9.13120449e-01
7.35705420e-02 8.96375120e-01 3.92209858e-01 -2.44529054e-01
6.40593588e-01 -9.95197058e-01 -1.63657874e-01 8.66992056e-01
-6.37151778e-01 -7.14160740e-01 1.42735364e-02 1.31421760e-01
1.05296898e+00 -3.79384428e-01 -2.25844964e-01 -1.44921243e+00
6.99917495e-01 1.31916068e-02 3.63452345e-01 5.73771477e-01
1.13384187e-01 2.94422597e-01 -3.27739418e-02 -5.52803457e-01
-1.16392338e+00 -1.05791795e+00 -6.86403215e-01 6.60435915e-01
1.07566690e+00 -6.11443937e-01 -8.64516437e-01 3.30180258e-01
-2.74252057e-01 6.11666024e-01 -3.07109028e-01 -6.08176351e-01
-5.93538404e-01 -1.49869287e+00 -1.47455916e-01 1.80932090e-01
6.56481564e-01 -1.49086761e+00 -5.27523696e-01 4.01990741e-01
1.12991519e-02 -2.60509461e-01 -3.59154977e-02 3.56021494e-01
-5.56066096e-01 -1.11417139e+00 -6.92253336e-02 -3.60041231e-01
5.16114414e-01 4.09957379e-01 1.09193182e+00 8.56153429e-01
-5.55325210e-01 4.57606554e-01 -3.73560905e-01 -4.81513143e-01
-6.56311214e-01 -4.57868464e-02 8.55305195e-01 -4.82434005e-01
3.67119700e-01 -5.59149206e-01 -8.22995484e-01 6.14308417e-01
-6.72535241e-01 -2.83090800e-01 -3.62427160e-02 2.62268960e-01
6.34995043e-01 7.68486261e-01 6.67818785e-01 -7.42853463e-01
2.32077003e-01 -5.52978516e-01 -1.62736014e-01 8.78514424e-02
-9.38225091e-01 -7.33770207e-02 3.35425466e-01 6.72825947e-02
-1.29290068e+00 -5.53717792e-01 -7.19002247e-01 3.61789703e-01
9.81741771e-02 3.96339417e-01 -5.77913046e-01 -1.53831383e-02
9.10944045e-01 -7.42070451e-02 -1.48545936e-01 -2.01327965e-01
-2.53065616e-01 6.72786891e-01 3.88541698e-01 -1.24743080e+00
5.76079905e-01 6.20525658e-01 1.07907377e-01 -1.17031765e+00
-2.17063248e-01 1.82329133e-01 -3.09477717e-01 -2.13146210e-01
1.08359730e+00 -6.27174199e-01 -9.46247280e-01 5.04654288e-01
-8.17674160e-01 -4.87331629e-01 -3.44436318e-01 2.89144278e-01
-4.58910257e-01 1.57513753e-01 -9.03747618e-01 -8.35422337e-01
-4.39951211e-01 -1.42349112e+00 4.37512010e-01 1.56381577e-01
-5.06909430e-01 -6.50427163e-01 1.43768653e-01 2.72658139e-01
4.63155180e-01 1.96692184e-01 1.61677778e+00 -7.14806700e-03
-6.14978850e-01 -7.48286257e-03 1.69999942e-01 -4.41094488e-01
3.03668171e-01 5.12519121e-01 -4.55854505e-01 -7.45569095e-02
2.97400504e-01 -4.37850058e-02 7.93642938e-01 5.58606565e-01
5.86099625e-01 4.99844663e-02 -9.58535135e-01 -4.47299480e-02
1.51048815e+00 1.25500476e+00 1.03670073e+00 3.99393708e-01
6.25854552e-01 4.60641772e-01 7.37646043e-01 1.15004994e-01
1.55405447e-01 4.15516615e-01 5.32969654e-01 7.48838857e-02
2.51006991e-01 4.05527145e-01 4.26455826e-01 8.27150643e-01
-1.06915581e+00 -4.22540493e-02 -1.23149121e+00 -7.80038610e-02
-1.17292631e+00 -1.03952754e+00 -7.22051203e-01 1.97589982e+00
1.00312698e+00 3.92295003e-01 5.65612651e-02 1.74286783e-01
7.88343728e-01 -6.08103573e-02 -8.44046891e-01 -3.46267343e-01
-2.36654535e-01 7.08870471e-01 3.38622808e-01 7.89934695e-01
-2.60536671e-01 1.17091179e+00 7.85120630e+00 6.69876754e-01
-1.42040634e+00 1.89485252e-01 6.47773564e-01 -1.88753843e-01
-8.59720290e-01 -6.43096492e-02 -9.54323769e-01 5.35830438e-01
9.39185381e-01 -5.37690781e-02 -6.04874408e-03 7.32666552e-01
-2.76199192e-01 -4.58505988e-01 -9.08260107e-01 8.82408440e-01
-6.40237510e-01 -1.66293275e+00 -1.63691506e-01 3.63885835e-02
7.95389891e-01 1.31734744e-01 1.17592335e-01 8.91533215e-03
4.78957176e-01 -1.16452169e+00 9.54530895e-01 3.56498957e-01
1.24507236e+00 -9.80817497e-01 2.39701211e-01 2.42701367e-01
-1.44909954e+00 4.17382151e-01 -8.00732791e-01 -3.02956969e-01
1.77552417e-01 6.51038706e-01 -2.23063882e-02 3.12238902e-01
1.39207935e+00 3.68608564e-01 1.22132070e-01 -2.33943060e-01
3.71703029e-01 -1.11810565e-02 -4.18063015e-01 -3.42823893e-01
8.15133676e-02 -8.73630226e-01 2.07541510e-01 2.51951158e-01
3.35763097e-01 1.19150162e+00 -3.76960993e-01 9.39445257e-01
1.65461794e-01 -7.51584023e-03 -8.01250756e-01 2.95067906e-01
4.60621446e-01 8.50558579e-01 -1.12103307e+00 -4.24591750e-01
-3.14635545e-01 3.89784366e-01 -3.26306894e-02 -8.40886459e-02
-6.03639781e-01 -3.19673628e-01 1.03872001e+00 8.48145425e-01
-3.58446032e-01 -2.54850268e-01 -5.82454950e-02 -8.33372593e-01
-2.00807035e-01 -3.34126711e-01 -3.82128321e-02 -6.41896725e-01
-1.03872764e+00 9.29839671e-01 8.76682773e-02 -4.91195917e-01
3.15540403e-01 -7.19816148e-01 -9.22890186e-01 7.49513924e-01
-8.91439736e-01 -3.54014963e-01 1.35041073e-01 5.65572560e-01
-2.61676386e-02 9.69199371e-03 4.28345323e-01 -1.46222189e-01
-7.29986191e-01 4.99918789e-01 5.43553472e-01 -7.89822638e-01
3.24747652e-01 -6.22893751e-01 2.45763183e-01 3.71250391e-01
-6.16748571e-01 6.33965850e-01 1.01298130e+00 -1.03118491e+00
-1.82002199e+00 -7.37819493e-01 5.73706567e-01 -5.22142410e-01
3.38035822e-02 -4.69268203e-01 -7.81231225e-01 4.33266401e-01
3.34578514e-01 -5.20181239e-01 8.69118452e-01 -1.92511871e-01
1.58017397e-01 1.11633934e-01 -1.51626968e+00 8.31084430e-01
1.47941267e+00 -2.66172618e-01 -1.25939623e-01 5.26259601e-01
8.52050960e-01 -7.01529160e-02 -1.05285835e+00 7.54901111e-01
4.14952278e-01 -1.18192673e+00 1.07604730e+00 -2.10858241e-01
3.81064743e-01 -5.26324920e-02 -6.48168802e-01 -6.76300049e-01
-4.12181050e-01 -3.59280974e-01 4.92773324e-01 5.94051898e-01
8.95293057e-01 -9.79268253e-01 1.13287807e+00 1.53872359e+00
-5.45983553e-01 -6.57908201e-01 -1.21864963e+00 -8.09156656e-01
5.94866097e-01 -1.67720318e-01 1.07528687e+00 9.80670393e-01
6.13974571e-01 -1.18016154e-01 9.17955413e-02 -1.39420569e-01
4.98560578e-01 2.38067254e-01 1.70257017e-02 -1.55378783e+00
9.62492600e-02 -3.93647760e-01 2.90334076e-01 -4.76948470e-01
3.89449447e-01 -1.09046507e+00 6.81256354e-02 -1.14265621e+00
1.87440291e-01 -1.63066924e+00 -2.67661870e-01 2.28972122e-01
1.69244692e-01 2.59204537e-01 -3.32437992e-01 7.71309674e-01
-4.42730993e-01 6.21989191e-01 9.99325693e-01 3.50286327e-02
-2.00035974e-01 -4.70313400e-01 -7.17787445e-01 6.60561681e-01
7.26087093e-01 -5.30838788e-01 -4.51769046e-02 2.73833156e-01
8.47414970e-01 2.71694392e-01 -7.15608299e-01 -1.08938241e+00
-2.91001588e-01 -6.57075226e-01 -2.31786057e-01 -6.73829973e-01
4.31993634e-01 -8.96608829e-01 1.09488726e+00 9.97367442e-01
4.08786923e-01 4.16244194e-02 -3.09027165e-01 -9.11494866e-02
-1.11069024e-01 -7.46793449e-01 1.30732632e+00 -7.99922466e-01
-1.11534126e-01 4.83269215e-01 -6.56823575e-01 -5.68302333e-01
1.54073131e+00 -6.71074629e-01 -3.62867922e-01 7.06408620e-02
-7.46830404e-01 -3.10744166e-01 1.13436854e+00 -8.50450397e-02
3.60248566e-01 -1.47997606e+00 6.82633892e-02 2.51767546e-01
-1.90435752e-01 2.70616591e-01 6.02863371e-01 5.26518524e-01
-1.23444438e+00 3.94857258e-01 -2.17244387e-01 -6.59056902e-01
-7.91166306e-01 5.99561393e-01 8.40483963e-01 6.74061999e-02
-3.03421110e-01 1.09484553e+00 3.87909770e-01 -1.94929555e-01
-6.44640505e-01 -2.76246041e-01 1.23346604e-01 -5.19262314e-01
1.47747055e-01 6.01866424e-01 4.31414157e-01 -7.19488442e-01
-3.69804740e-01 8.66196454e-01 -3.27554911e-01 3.96790542e-02
1.21169317e+00 1.38089761e-01 -9.00116503e-01 2.56161004e-01
8.85978997e-01 2.62852252e-01 -8.23663056e-01 1.20539725e-01
-3.78611535e-02 7.86330272e-03 -9.79318798e-01 1.63160980e-01
-1.09516847e+00 3.13867629e-01 1.26761705e-01 3.75333309e-01
5.44413626e-01 9.87043321e-01 7.36124277e-01 4.53316301e-01
7.44022548e-01 -1.29968917e+00 2.45004892e-01 3.17534596e-01
5.64909697e-01 -8.42494190e-01 -1.72056064e-01 -9.72448170e-01
-7.67781511e-02 8.41525853e-01 1.33002627e+00 -2.67608374e-01
1.23533678e+00 4.08481121e-01 -2.08040521e-01 -9.32505429e-01
-9.88211334e-01 5.03109038e-01 -4.32744861e-01 2.16607094e-01
6.63808405e-01 3.14434618e-01 -7.22046316e-01 9.19649839e-01
-7.22341597e-01 -9.38013136e-01 7.91758835e-01 1.27500522e+00
-8.58938992e-01 -1.43436813e+00 -6.52029753e-01 6.11872852e-01
-4.78165537e-01 -1.36777043e-01 -1.93800136e-01 6.55454397e-01
8.84934068e-02 8.03417981e-01 3.76288623e-01 -3.35886806e-01
2.07538560e-01 2.70601481e-01 6.30756140e-01 -5.15947998e-01
-2.73026884e-01 -8.00882354e-02 4.55248170e-02 -3.31986219e-01
-5.43534100e-01 -7.47007072e-01 -2.04999256e+00 -4.95997518e-01
-5.28977394e-01 7.99509764e-01 3.00923556e-01 8.99093330e-01
2.51432117e-02 -6.67852536e-02 3.92318487e-01 -8.90953302e-01
1.32912204e-01 -7.30468929e-01 -1.39863861e+00 4.83740717e-01
-2.66068220e-01 -1.16040325e+00 -4.53849792e-01 -4.16579008e-01] | [5.377104759216309, 4.909825325012207] |
8176a1cb-5dc9-4b3b-8fff-54f47635747e | multiframe-scene-flow-with-piecewise-rigid | 1710.02124 | null | http://arxiv.org/abs/1710.02124v1 | http://arxiv.org/pdf/1710.02124v1.pdf | Multiframe Scene Flow with Piecewise Rigid Motion | We introduce a novel multiframe scene flow approach that jointly optimizes
the consistency of the patch appearances and their local rigid motions from
RGB-D image sequences. In contrast to the competing methods, we take advantage
of an oversegmentation of the reference frame and robust optimization
techniques. We formulate scene flow recovery as a global non-linear least
squares problem which is iteratively solved by a damped Gauss-Newton approach.
As a result, we obtain a qualitatively new level of accuracy in RGB-D based
scene flow estimation which can potentially run in real-time. Our method can
handle challenging cases with rigid, piecewise rigid, articulated and moderate
non-rigid motion, and does not rely on prior knowledge about the types of
motions and deformations. Extensive experiments on synthetic and real data show
that our method outperforms state-of-the-art. | ['Jan Kautz', 'Matthias Nießner', 'Kihwan Kim', 'Didier Stricker', 'Robert Maier', 'Vladislav Golyanik'] | 2017-10-05 | null | null | null | null | ['scene-flow-estimation'] | ['computer-vision'] | [ 1.21466108e-01 -2.88297057e-01 -5.45881949e-02 -1.57718405e-01
-6.14025831e-01 -7.45963216e-01 5.37346184e-01 -3.52765173e-01
-2.42925480e-01 5.37472367e-01 7.51294494e-02 -2.49027275e-02
4.91862595e-02 -4.68326420e-01 -6.85037494e-01 -5.99063814e-01
2.51954913e-01 4.04800236e-01 4.62819129e-01 -2.58038461e-01
2.50955909e-01 7.40053058e-01 -1.23585570e+00 -2.26071924e-01
6.97897315e-01 6.85499728e-01 3.80611769e-03 9.99471068e-01
2.29281306e-01 9.93609011e-01 -3.52729782e-02 -1.04219101e-01
5.82403600e-01 -4.95087206e-01 -1.06297731e+00 7.14082956e-01
8.87801409e-01 -7.05068886e-01 -6.41113281e-01 8.42581689e-01
3.29977453e-01 4.66492385e-01 5.26377320e-01 -9.75160480e-01
-7.68480524e-02 -4.13286895e-01 -8.93874943e-01 2.01032028e-01
8.54919612e-01 3.02979410e-01 6.86824739e-01 -7.32901633e-01
9.74441528e-01 1.25740087e+00 7.05585063e-01 3.26788604e-01
-1.48868406e+00 -2.12193988e-02 2.18940243e-01 -9.61998552e-02
-1.15499413e+00 -4.85997736e-01 1.03269780e+00 -8.20154965e-01
6.70154989e-01 1.49001405e-01 7.85387456e-01 6.99374199e-01
2.56693922e-02 6.77482188e-01 9.28839028e-01 -2.83635050e-01
1.36118457e-01 -3.58688027e-01 -2.01520085e-01 1.04861474e+00
7.46949315e-02 1.05150580e-01 -4.63712990e-01 -1.49062544e-01
1.31351757e+00 -1.34953946e-01 -5.63803613e-01 -8.55198920e-01
-1.47037172e+00 5.42731583e-01 1.89749286e-01 5.23129851e-02
-3.85219395e-01 4.30248946e-01 -7.34094381e-02 -1.29767492e-01
5.41768253e-01 2.12020744e-02 -2.80365616e-01 -2.06430644e-01
-1.13160932e+00 4.52186316e-01 8.28218877e-01 7.97123969e-01
1.02456284e+00 3.35016578e-01 3.57686430e-01 3.77541542e-01
6.03933692e-01 7.82879651e-01 4.63036858e-02 -1.65400386e+00
3.38103026e-01 3.69170040e-01 3.55577767e-01 -1.22027051e+00
-1.30731344e-01 2.54254062e-02 -6.29731655e-01 3.19842100e-01
7.18644738e-01 7.67806247e-02 -9.04688418e-01 1.40360820e+00
8.90753984e-01 5.58009088e-01 -2.69763052e-01 1.13555932e+00
4.40170288e-01 5.61214507e-01 -4.20008898e-01 -3.28120440e-01
8.51490676e-01 -1.01622903e+00 -6.56972706e-01 -2.44723424e-01
2.55664885e-01 -8.48324597e-01 7.59848833e-01 2.62845844e-01
-1.42490149e+00 -4.73304063e-01 -6.84869528e-01 -3.09169918e-01
2.60408312e-01 -2.01938495e-01 6.06887162e-01 4.63412106e-01
-1.00926077e+00 7.12807477e-01 -1.26253963e+00 -2.49643371e-01
1.53474003e-01 3.67681652e-01 -5.56209981e-01 -1.83498055e-01
-4.42195266e-01 7.51281321e-01 3.80012691e-02 2.91214675e-01
-9.75023806e-01 -8.18077624e-01 -1.08961177e+00 -4.86946374e-01
4.36127156e-01 -1.09313118e+00 1.15895462e+00 -8.22683454e-01
-1.91970694e+00 1.03073776e+00 -5.67321420e-01 5.77984825e-02
8.08161497e-01 -3.44227165e-01 3.26972872e-01 5.27591050e-01
-9.11059305e-02 3.96649927e-01 7.93222249e-01 -1.39040065e+00
-3.49275172e-01 -1.14871405e-01 1.85908616e-01 2.78739333e-01
3.24700654e-01 -1.17495932e-01 -6.28417671e-01 -5.53813636e-01
4.15822864e-01 -1.29766607e+00 -5.28513670e-01 4.26574886e-01
-3.45519125e-01 3.84424448e-01 7.83462048e-01 -7.58659482e-01
8.78453135e-01 -1.83707917e+00 7.22667217e-01 9.11413655e-02
9.59215537e-02 2.78229062e-02 2.47667745e-01 1.73576772e-01
1.79481238e-01 -2.29777172e-01 -6.19818985e-01 -5.20057380e-01
-2.37395614e-01 4.27005678e-01 -1.95161074e-01 9.61207151e-01
7.36246677e-03 7.42896259e-01 -1.08545411e+00 -5.42990088e-01
6.35262787e-01 6.42874062e-01 -6.52979314e-01 3.39312106e-01
-1.35271694e-03 1.05707610e+00 -5.93129575e-01 6.56622112e-01
6.15280509e-01 -2.28174090e-01 -1.21722259e-01 -3.66107047e-01
-3.23513538e-01 2.35895783e-01 -1.67716241e+00 2.31286979e+00
-1.86543390e-01 4.82393682e-01 4.60825175e-01 -9.29679513e-01
4.81742680e-01 2.30188549e-01 9.14103448e-01 -3.03950906e-01
5.99997230e-02 2.82405436e-01 -3.89825523e-01 -5.05494177e-01
4.15051103e-01 -2.51736850e-01 2.59874016e-01 3.63169134e-01
-6.03426341e-03 -6.38962150e-01 1.65710479e-01 1.57428220e-01
9.71095383e-01 7.59908319e-01 1.10208996e-01 -3.59478831e-01
7.08803892e-01 1.92096874e-01 7.88557231e-01 3.94726783e-01
-2.08509803e-01 9.34620917e-01 1.92422301e-01 -5.79207122e-01
-1.07855546e+00 -9.72861052e-01 3.07047162e-02 3.91502291e-01
4.82284009e-01 -1.28569528e-01 -6.39675319e-01 -4.75175470e-01
2.52915975e-02 1.13445237e-01 -4.04648095e-01 3.12458754e-01
-1.02611411e+00 -6.16553187e-01 6.77540451e-02 3.21964175e-01
3.39428812e-01 -6.60686016e-01 -8.54233503e-01 3.01620215e-01
-4.15660858e-01 -1.57726049e+00 -7.52858877e-01 -2.81913131e-01
-1.16320801e+00 -1.11776292e+00 -7.30028689e-01 -4.86346960e-01
8.15736294e-01 4.84943330e-01 1.08772838e+00 -1.15007209e-02
-4.99871612e-01 9.52190161e-01 5.26579581e-02 3.55264336e-01
-2.72799850e-01 -3.06560546e-01 -1.43534869e-01 2.48829052e-01
-5.52401721e-01 -5.63486457e-01 -9.04809058e-01 4.00252134e-01
-8.92979503e-01 1.54392913e-01 9.35474560e-02 5.07223785e-01
8.91308248e-01 -2.66887844e-01 -2.92995632e-01 -6.10948622e-01
-3.75319645e-02 -1.66148305e-01 -7.59373844e-01 8.04652199e-02
-1.19817331e-01 1.35658145e-01 2.10096002e-01 -3.96696925e-01
-1.32732773e+00 7.09310710e-01 5.33600748e-02 -7.60389805e-01
-1.44251436e-01 1.16432942e-01 -7.86533859e-03 -6.62967384e-01
3.15812826e-01 1.14320882e-01 8.80143512e-03 -3.81827414e-01
6.11222863e-01 1.93755061e-03 8.08679938e-01 -7.98503637e-01
1.08826780e+00 1.18231332e+00 4.40137208e-01 -1.03683054e+00
-4.80867803e-01 -7.33617842e-01 -1.15531611e+00 -5.49340725e-01
1.00306416e+00 -8.17113161e-01 -8.09724867e-01 8.48701835e-01
-1.22148097e+00 -7.13485062e-01 -2.68251181e-01 5.99051118e-01
-9.28161979e-01 9.14650977e-01 -8.31040621e-01 -6.77191794e-01
4.69793566e-02 -1.31122637e+00 1.32933819e+00 1.81254819e-01
-6.43673837e-02 -1.33418250e+00 4.82735544e-01 4.54308540e-01
1.29526421e-01 9.18679476e-01 2.16020852e-01 3.90253037e-01
-1.09228861e+00 7.94635341e-02 1.65399268e-01 8.74610767e-02
2.54005700e-01 3.80299866e-01 -8.23845387e-01 -2.55940169e-01
1.24841779e-01 -1.67282641e-01 6.50476277e-01 5.23644567e-01
4.59369212e-01 -3.06669056e-01 -9.43605527e-02 1.02269077e+00
1.61919892e+00 -4.70238514e-02 5.70725918e-01 2.42335826e-01
1.22838390e+00 7.37045407e-01 5.88856936e-01 3.75520200e-01
6.73887610e-01 8.51271689e-01 5.33667803e-01 -2.01089427e-01
-1.76051706e-01 -1.33033216e-01 3.06030780e-01 8.28459978e-01
-5.45956731e-01 1.40920831e-02 -9.79484081e-01 6.87355161e-01
-2.03316355e+00 -9.26528394e-01 -5.19007444e-01 2.36181450e+00
6.87281847e-01 -2.90046066e-01 1.97969869e-01 -3.53357662e-03
3.95550728e-01 2.45548829e-01 -5.14869571e-01 -8.00119489e-02
-1.36039421e-01 1.12847790e-01 5.98181009e-01 1.03679097e+00
-1.03786802e+00 9.44183707e-01 6.74296999e+00 2.40652338e-01
-1.13229620e+00 -1.07456653e-05 3.97771984e-01 -4.32920158e-02
-3.26112300e-01 2.70786971e-01 -5.35075486e-01 2.19919682e-01
4.77028400e-01 8.26486126e-02 5.42022824e-01 4.40992147e-01
3.76437634e-01 -3.13905120e-01 -9.25351679e-01 9.63636041e-01
4.73580696e-02 -1.21622658e+00 -2.90785998e-01 2.01176494e-01
1.01459384e+00 7.22180009e-02 -3.51089329e-01 -5.77304006e-01
5.23607492e-01 -5.59495687e-01 9.28724408e-01 7.04669118e-01
4.21809196e-01 -3.14279377e-01 1.74031064e-01 1.90199107e-01
-1.30357218e+00 3.11358988e-01 -1.07433185e-01 3.10233738e-02
8.23521674e-01 5.54759562e-01 -6.52412772e-02 7.18495369e-01
6.05269194e-01 1.06867707e+00 -1.61118820e-01 1.02049685e+00
-3.49491268e-01 3.77946109e-01 -6.17236316e-01 6.02742493e-01
7.14078322e-02 -5.87531567e-01 7.73751140e-01 1.13809204e+00
2.13727474e-01 4.58421856e-01 4.54024583e-01 8.04728627e-01
3.89213294e-01 -5.71202002e-02 -4.77820277e-01 4.47942704e-01
-7.29831606e-02 1.25677383e+00 -8.40262175e-01 -2.22903579e-01
-5.17614424e-01 1.21997452e+00 1.27448410e-01 6.18695915e-01
-6.43087864e-01 3.16974699e-01 6.70668721e-01 2.84422159e-01
4.08324420e-01 -7.94672608e-01 7.09656328e-02 -1.71836329e+00
1.59147888e-01 -5.94918013e-01 1.70880288e-01 -7.67344117e-01
-9.41743553e-01 3.65609556e-01 1.04765356e-01 -1.24807227e+00
-4.59842265e-01 -5.06286144e-01 -5.92167854e-01 6.59843087e-01
-1.73059690e+00 -1.19760990e+00 -4.27575320e-01 9.23924685e-01
3.83184820e-01 5.72568953e-01 4.20926213e-01 2.39301428e-01
-4.47054654e-01 -1.27394289e-01 -4.03281786e-02 8.41743350e-02
6.34573758e-01 -1.19474483e+00 2.19717234e-01 1.23935556e+00
2.80913915e-02 4.05973464e-01 8.42847466e-01 -5.69977701e-01
-1.77821040e+00 -7.89625168e-01 5.06611347e-01 -3.67200643e-01
6.14446819e-01 -9.97555256e-02 -9.08477247e-01 7.81566679e-01
9.48959775e-03 6.78246677e-01 2.57824987e-01 -5.02669334e-01
-6.47919998e-02 -6.99401125e-02 -1.00272715e+00 2.58156747e-01
1.13290930e+00 -4.98185724e-01 -3.22454482e-01 2.11369500e-01
4.20388699e-01 -9.28684652e-01 -1.00247979e+00 3.80676180e-01
5.28587341e-01 -1.02997148e+00 1.35642850e+00 -4.28481787e-01
3.22149843e-01 -7.23407447e-01 -1.44756451e-01 -1.00017178e+00
-5.26049174e-02 -1.20286131e+00 -3.84412259e-01 9.30608749e-01
-1.56733230e-01 -3.13102216e-01 7.98457444e-01 9.93053854e-01
-1.12257898e-01 -4.06408995e-01 -9.57889438e-01 -6.44101858e-01
-2.64019340e-01 -2.73974895e-01 -6.16156757e-02 1.06054246e+00
-3.92792016e-01 -2.37270235e-03 -6.13519192e-01 2.18489900e-01
1.09021318e+00 2.97184050e-01 1.10884714e+00 -1.01557684e+00
-5.28634131e-01 -1.93377599e-01 -5.47315836e-01 -1.50725567e+00
1.70140192e-01 -5.26459813e-01 2.46598631e-01 -1.46736956e+00
-1.63512364e-01 -2.39470690e-01 1.48909494e-01 3.56556356e-01
-2.46239245e-01 3.43834519e-01 1.47371426e-01 4.33033586e-01
-5.21163285e-01 5.36952138e-01 1.42844760e+00 5.67915030e-02
-4.43649977e-01 -9.22474861e-02 -6.75225407e-02 9.81197953e-01
2.78060496e-01 -4.71524239e-01 -3.53456885e-01 -6.10037923e-01
1.21324942e-01 4.29155588e-01 5.68603635e-01 -7.28903830e-01
3.57996672e-01 -4.51474816e-01 -4.14179824e-02 -3.80008191e-01
4.08234149e-01 -7.68987298e-01 3.34500760e-01 4.02571470e-01
9.25150737e-02 1.74794942e-01 1.56285003e-01 6.63355768e-01
-1.12592854e-01 -2.26604342e-02 8.26536357e-01 -4.01384383e-01
-5.53953886e-01 4.70459968e-01 -2.37577558e-01 1.54248983e-01
9.49969292e-01 -4.37071979e-01 -9.65305567e-02 -3.56505096e-01
-6.83492243e-01 -2.83808336e-02 8.35944295e-01 1.44353703e-01
4.94531065e-01 -1.26910198e+00 -7.19050586e-01 1.52496159e-01
-2.32428923e-01 4.28403378e-01 2.03293771e-01 1.07181251e+00
-1.05828774e+00 2.06318393e-01 -4.31106947e-02 -9.92810667e-01
-1.07885706e+00 3.24619293e-01 5.77917576e-01 -1.47007152e-01
-6.44987285e-01 4.95911509e-01 2.76771873e-01 -4.77421761e-01
-1.15580298e-01 -3.65490526e-01 2.04710633e-01 -3.58938009e-01
1.71699926e-01 6.79528832e-01 -1.37128711e-01 -1.26198566e+00
-5.01136541e-01 1.31500661e+00 4.59984452e-01 -3.39485228e-01
1.41307199e+00 -5.05462170e-01 -1.88628301e-01 4.15752739e-01
1.20182598e+00 1.70390502e-01 -1.92780995e+00 -8.37562978e-02
-3.53808850e-01 -8.05351138e-01 1.04697123e-01 -1.55530483e-01
-1.37615359e+00 8.09085786e-01 1.15660042e-01 -1.85691476e-01
1.09032786e+00 -2.17692569e-01 7.16447294e-01 1.52222246e-01
3.36368114e-01 -7.24047601e-01 1.08017519e-01 3.27044874e-01
6.10830545e-01 -1.11657333e+00 3.55279058e-01 -7.70944655e-01
-2.98693836e-01 1.25290585e+00 3.55150312e-01 -4.72271502e-01
6.22044265e-01 2.44254708e-01 1.00250937e-01 5.37580214e-02
-4.59776849e-01 -1.54165387e-01 5.03999174e-01 4.29109275e-01
2.34360769e-01 -4.31598902e-01 3.47812450e-03 -4.72667456e-01
1.19944245e-01 2.50953231e-02 6.02305233e-01 1.13034511e+00
-2.06679240e-01 -1.18884003e+00 -4.56386983e-01 -3.56680036e-01
-4.81060714e-01 3.25431138e-01 -1.67066187e-01 8.40537012e-01
-1.55858636e-01 7.98370063e-01 -1.71920732e-01 1.43459439e-01
4.98169452e-01 -2.93193221e-01 8.71559799e-01 -2.03285128e-01
-4.80081975e-01 5.58333516e-01 -1.60786390e-01 -1.07945740e+00
-1.10066247e+00 -9.96058643e-01 -1.14276290e+00 -2.85910130e-01
-2.81812996e-01 -2.55261540e-01 4.99514580e-01 9.45024490e-01
2.76843935e-01 2.19348595e-01 5.53309262e-01 -1.28492451e+00
5.43594360e-03 -2.98091680e-01 -2.74382532e-01 7.53486454e-01
7.66838193e-01 -5.72154522e-01 -5.02605021e-01 7.01836526e-01] | [8.615194320678711, -2.018214225769043] |
9b2f0652-ee32-4a2f-805c-f374ca03e62e | long-range-arena-a-benchmark-for-efficient-1 | 2011.04006 | null | https://arxiv.org/abs/2011.04006v1 | https://arxiv.org/pdf/2011.04006v1.pdf | Long Range Arena: A Benchmark for Efficient Transformers | Transformers do not scale very well to long sequence lengths largely because of quadratic self-attention complexity. In the recent months, a wide spectrum of efficient, fast Transformers have been proposed to tackle this problem, more often than not claiming superior or comparable model quality to vanilla Transformer models. To this date, there is no well-established consensus on how to evaluate this class of models. Moreover, inconsistent benchmarking on a wide spectrum of tasks and datasets makes it difficult to assess relative model quality amongst many models. This paper proposes a systematic and unified benchmark, LRA, specifically focused on evaluating model quality under long-context scenarios. Our benchmark is a suite of tasks consisting of sequences ranging from $1K$ to $16K$ tokens, encompassing a wide range of data types and modalities such as text, natural, synthetic images, and mathematical expressions requiring similarity, structural, and visual-spatial reasoning. We systematically evaluate ten well-established long-range Transformer models (Reformers, Linformers, Linear Transformers, Sinkhorn Transformers, Performers, Synthesizers, Sparse Transformers, and Longformers) on our newly proposed benchmark suite. LRA paves the way towards better understanding this class of efficient Transformer models, facilitates more research in this direction, and presents new challenging tasks to tackle. Our benchmark code will be released at https://github.com/google-research/long-range-arena. | ['Donald Metzler', 'Sebastian Ruder', 'Liu Yang', 'Jinfeng Rao', 'Philip Pham', 'Dara Bahri', 'Yikang Shen', 'Samira Abnar', 'Mostafa Dehghani', 'Yi Tay'] | 2020-11-08 | null | null | null | null | ['long-range-modeling'] | ['natural-language-processing'] | [ 4.09386754e-01 -3.94225866e-01 -2.79252883e-03 -3.35400403e-01
-1.00613427e+00 -9.05421793e-01 8.87388110e-01 -7.10066855e-02
-4.25242752e-01 6.01182818e-01 2.24573076e-01 -5.42413235e-01
-2.87678838e-01 -6.58873022e-01 -8.21243525e-01 -4.41623807e-01
1.47181582e-02 7.00675488e-01 -8.82934257e-02 -4.00724173e-01
3.16785544e-01 3.49448860e-01 -1.63507926e+00 5.37006021e-01
9.74925995e-01 1.26632833e+00 4.03184056e-01 5.82344770e-01
-1.22106947e-01 1.00564301e+00 -3.77620876e-01 -9.77007151e-01
1.44958332e-01 -2.68082201e-01 -1.03182554e+00 -2.36652747e-01
6.26414895e-01 -1.25683174e-01 -9.19549763e-02 9.09364462e-01
7.01877713e-01 1.08590394e-01 6.16064370e-01 -1.36099350e+00
-8.71070623e-01 8.39053512e-01 -2.93301463e-01 4.81142938e-01
5.94126821e-01 5.32930374e-01 1.70247030e+00 -1.07855582e+00
6.10490799e-01 1.30892038e+00 9.13330615e-01 3.60826284e-01
-1.01551843e+00 -7.46178389e-01 4.45193127e-02 6.37656510e-01
-1.33522570e+00 -5.77433944e-01 4.46137905e-01 -2.86506027e-01
1.33646059e+00 5.41346073e-01 6.97475433e-01 1.43912423e+00
1.19712882e-01 9.11374211e-01 1.35934007e+00 -1.78039446e-01
-7.34231472e-02 -1.66053623e-01 -1.15910592e-02 6.82523727e-01
-1.10192031e-01 1.06182441e-01 -8.28349829e-01 1.16843738e-01
4.51813102e-01 -2.70905614e-01 -4.25776124e-01 -2.19751358e-01
-1.47960556e+00 5.66368163e-01 1.84498966e-01 5.25318921e-01
-1.97303295e-01 1.11048609e-01 6.83597267e-01 5.39299786e-01
4.06571388e-01 7.38790631e-01 -4.88196164e-01 -6.18311644e-01
-8.72510195e-01 5.67459822e-01 6.26759112e-01 1.00657809e+00
4.35652137e-01 1.50910944e-01 -2.51239121e-01 8.94060969e-01
4.56627421e-02 6.43877864e-01 5.77040255e-01 -9.22769427e-01
7.27891505e-01 3.75871480e-01 -1.38478652e-01 -9.72311676e-01
-3.49838793e-01 -6.27996445e-01 -1.01050675e+00 -1.75831810e-01
6.60292864e-01 3.46964478e-01 -6.30788922e-01 1.76642954e+00
-5.91557287e-02 4.04215939e-02 -1.35109916e-01 8.69544268e-01
9.87485468e-01 7.13089824e-01 -1.24815382e-01 9.47852507e-02
1.41212654e+00 -1.08669353e+00 -5.43709576e-01 -2.51312166e-01
6.88032866e-01 -1.00545204e+00 1.71557486e+00 7.09999740e-01
-1.52040446e+00 -5.22183895e-01 -8.16418350e-01 -4.16093290e-01
-5.26661813e-01 -2.49362126e-01 6.15272105e-01 5.47820389e-01
-1.40226507e+00 6.09655380e-01 -4.62743431e-01 -3.29815775e-01
4.88006055e-01 1.42292202e-01 -3.02047014e-01 -3.19420666e-01
-1.42474496e+00 1.24200797e+00 -2.27054954e-03 1.00597836e-01
-1.05525827e+00 -1.10674751e+00 -8.67536962e-01 1.59075782e-01
3.35853010e-01 -7.82555461e-01 1.57547593e+00 -9.12096858e-01
-1.26345015e+00 1.13271844e+00 -1.08121060e-01 -6.55396760e-01
6.04161799e-01 -2.20513999e-01 -2.53549695e-01 -3.06574553e-01
1.24400616e-01 5.61610162e-01 6.68582261e-01 -7.60060430e-01
-3.64976227e-01 -1.24939524e-01 3.97084624e-01 2.86437184e-01
-2.00835288e-01 1.15919434e-01 -1.11080617e-01 -8.62715542e-01
-5.06501734e-01 -7.66659021e-01 1.43712148e-01 -6.53041229e-02
-3.14785361e-01 -4.11625236e-01 2.82743007e-01 -5.14567733e-01
1.20691109e+00 -2.04321933e+00 4.22986716e-01 -1.65253893e-01
3.02468508e-01 1.30829513e-01 -5.57256222e-01 5.37198722e-01
-3.21793109e-01 1.87619314e-01 -3.97599667e-01 -4.03231263e-01
4.23784703e-01 1.05563529e-01 -5.64407349e-01 2.86993682e-01
-1.76610276e-02 1.25223374e+00 -8.67064834e-01 -4.78879303e-01
1.45093203e-01 3.34099412e-01 -5.79954267e-01 8.23210403e-02
-3.17717344e-01 6.91905990e-02 -1.89676687e-01 8.04274559e-01
5.47778487e-01 -6.16373599e-01 -1.42601624e-01 -3.88740271e-01
-4.42986228e-02 5.03208399e-01 -8.05785477e-01 1.95639932e+00
-6.94847107e-01 7.43868709e-01 -1.94601700e-01 -1.26630807e+00
5.65890551e-01 2.34414101e-01 3.20215821e-01 -1.19360089e+00
7.50074387e-02 3.10125619e-01 9.78096575e-02 -4.72348005e-01
6.48681343e-01 -3.25359762e-01 -7.43183792e-02 3.83520335e-01
2.45418131e-01 -4.23751533e-01 6.22679472e-01 1.81753397e-01
1.16777611e+00 1.32433727e-01 1.83353275e-01 -2.94628024e-01
4.60461557e-01 -1.56249881e-01 1.41029239e-01 7.02916563e-01
-1.04868881e-01 5.95323801e-01 4.52218801e-01 -4.96591628e-01
-1.08868515e+00 -1.16241264e+00 -3.63284387e-02 1.41818070e+00
-3.09166312e-02 -6.99712515e-01 -6.46550298e-01 -3.08718830e-01
-2.52493441e-01 6.59931898e-01 -7.27971315e-01 -9.85903200e-03
-4.10092473e-01 -6.63082123e-01 9.15764570e-01 5.28061271e-01
5.93978286e-01 -1.17184496e+00 -6.39351189e-01 -1.91120151e-02
-6.23200417e-01 -1.19941080e+00 -5.35439193e-01 -2.37059332e-02
-6.80697858e-01 -1.02398562e+00 -8.60122442e-01 -5.27816594e-01
1.11954145e-01 8.11389387e-02 1.82210565e+00 -8.31333548e-02
-9.24263969e-02 4.58799422e-01 -3.76003385e-01 -2.34028235e-01
-1.86582386e-01 1.82824761e-01 -2.44696781e-01 -1.99977621e-01
-1.39433872e-02 -4.83780503e-01 -3.09108943e-01 2.74023652e-01
-8.47770333e-01 2.43553400e-01 6.80186570e-01 9.31525350e-01
5.66369891e-01 -2.05261722e-01 2.76320666e-01 -7.48000085e-01
8.58808756e-01 -4.75340515e-01 -4.57569242e-01 5.51471412e-01
-5.99866390e-01 1.12986483e-01 8.24169874e-01 -3.67971122e-01
-8.53139877e-01 -5.34044921e-01 -3.96927565e-01 -4.80181217e-01
1.43583551e-01 7.77100980e-01 -9.77253243e-02 6.39669225e-02
7.42001116e-01 5.75406611e-01 -1.61135778e-01 -3.25640142e-01
5.60011148e-01 2.15429932e-01 4.17509794e-01 -9.15097535e-01
6.37055874e-01 2.79945433e-01 -1.67060375e-01 -6.45520806e-01
-7.83898115e-01 -1.85099959e-01 1.59393176e-02 -2.01105326e-01
5.84322035e-01 -7.18532145e-01 -6.72968149e-01 7.07488775e-01
-1.10969055e+00 -7.59402990e-01 -3.76907498e-01 1.62811592e-01
-7.30881393e-01 3.32728773e-01 -6.20806575e-01 -3.92646641e-01
-6.71146274e-01 -1.34588003e+00 1.01092494e+00 -1.75714567e-01
-3.59793991e-01 -1.08423734e+00 -3.74566913e-02 5.50193250e-01
8.28731656e-01 7.24469796e-02 8.55203271e-01 -3.33342522e-01
-6.61048591e-01 3.02125692e-01 -2.89846063e-01 3.44472826e-01
-1.02774419e-01 4.91480678e-02 -8.46870959e-01 -2.38719970e-01
-1.36076257e-01 -7.06500411e-01 8.32187235e-01 3.91177118e-01
1.33431005e+00 -2.57861286e-01 -1.70746595e-02 7.57389903e-01
1.15788674e+00 7.82163069e-02 7.20962286e-01 2.89315850e-01
5.74059129e-01 2.60328114e-01 5.01256526e-01 3.48374099e-01
7.09984303e-01 5.89318573e-01 4.69563276e-01 8.06751773e-02
-1.32178754e-01 1.06896227e-02 2.98682272e-01 1.00753081e+00
-2.69750714e-01 -4.79417682e-01 -1.18988526e+00 5.17783463e-01
-1.45886207e+00 -1.18726826e+00 1.47602066e-01 1.90317619e+00
9.14924383e-01 1.70449242e-01 1.06380500e-01 3.67214322e-01
3.06580722e-01 3.85362029e-01 -5.37763357e-01 -5.03290057e-01
-3.82789761e-01 5.40329933e-01 1.38738424e-01 3.16657603e-01
-8.82870257e-01 8.47296119e-01 6.11413383e+00 1.27125549e+00
-1.23541319e+00 1.42098114e-01 6.83157623e-01 -1.15136944e-01
-6.30318105e-01 -2.69621640e-01 -4.22425807e-01 3.20129454e-01
8.72930765e-01 -4.31568265e-01 8.90062571e-01 6.47026479e-01
-1.26163974e-01 1.69194430e-01 -1.30337203e+00 1.21828353e+00
2.09268957e-01 -1.49299347e+00 1.97647735e-01 -2.19033122e-01
3.89344335e-01 3.42106074e-01 5.59257746e-01 4.76970345e-01
3.16756397e-01 -1.38206184e+00 1.13781309e+00 4.14806724e-01
1.02901638e+00 -4.18026775e-01 4.67355937e-01 1.72119513e-01
-1.33317947e+00 1.88513901e-02 -9.17617902e-02 -2.95510918e-01
2.29977146e-02 4.06284392e-01 -4.06110495e-01 6.07742071e-01
9.61595237e-01 1.07448196e+00 -6.43064260e-01 8.87936831e-01
1.52938530e-01 5.72098672e-01 -2.98989058e-01 -5.68465330e-02
4.73110408e-01 1.89945139e-02 3.88007104e-01 1.26341009e+00
5.27239084e-01 -6.70762882e-02 -3.03399920e-01 7.12279916e-01
-2.26345092e-01 1.97758749e-01 -6.80676460e-01 -1.69551998e-01
4.16925848e-01 1.02232790e+00 -5.77237129e-01 -3.28414500e-01
-4.92588669e-01 6.21583104e-01 3.36011529e-01 3.21675360e-01
-1.15161622e+00 -1.51015222e-01 6.05375409e-01 1.60126314e-01
1.45471543e-01 -1.57488137e-02 -3.23278785e-01 -1.27673805e+00
7.63744861e-02 -1.58545649e+00 6.76475346e-01 -1.19101667e+00
-1.38693643e+00 7.79294729e-01 1.04958154e-01 -1.10678124e+00
-1.67353824e-01 -5.41563272e-01 -3.33952218e-01 6.80339813e-01
-1.57325613e+00 -9.57564235e-01 -3.58931571e-01 7.63240576e-01
8.22461009e-01 -6.76787198e-02 7.32880890e-01 5.80821395e-01
-2.74759710e-01 7.65787184e-01 -9.29112881e-02 -4.87199984e-02
5.86118340e-01 -1.18528795e+00 6.73792243e-01 5.53975642e-01
3.10173154e-01 5.16209841e-01 7.49312937e-01 1.03285313e-02
-1.57301199e+00 -8.19372833e-01 9.80528951e-01 -5.41197717e-01
8.61111879e-01 -3.64684194e-01 -8.88125181e-01 7.08368361e-01
4.86394614e-01 -2.12412532e-02 4.38383847e-01 1.08772032e-01
-7.80821681e-01 -3.36837351e-01 -9.00939405e-01 5.75250506e-01
1.38791251e+00 -6.38838828e-01 -4.48788017e-01 3.10045809e-01
5.74331582e-01 -6.34746790e-01 -9.45601583e-01 5.04690528e-01
6.58158004e-01 -1.26801765e+00 1.18894577e+00 -3.85296583e-01
8.89167786e-01 1.18029237e-01 -3.04454505e-01 -1.41288698e+00
-2.15097800e-01 -6.39088213e-01 1.31825045e-01 1.08473420e+00
5.22054195e-01 -7.20700264e-01 4.60866928e-01 2.55264372e-01
-3.17431420e-01 -1.15349495e+00 -1.06355202e+00 -8.78651798e-01
3.85448277e-01 -8.05635989e-01 8.93394768e-01 1.08154619e+00
-6.28214553e-02 4.46445942e-01 -3.07091951e-01 -4.94487166e-01
5.37402868e-01 2.77682811e-01 6.14057422e-01 -9.32907760e-01
-2.42061511e-01 -1.07695055e+00 -2.51810491e-01 -1.11035967e+00
2.05085814e-01 -1.19079292e+00 -3.39389920e-01 -1.55653238e+00
1.60693437e-01 -5.29045880e-01 -1.84054539e-01 5.07217944e-01
4.67039496e-02 2.48031214e-01 3.38563710e-01 -2.49995254e-02
-7.49684095e-01 5.21871448e-01 1.49921250e+00 -3.89256060e-01
5.45815229e-01 -2.79035568e-01 -7.35292375e-01 4.60514128e-01
9.36435997e-01 -1.73191935e-01 -6.71269417e-01 -7.84685552e-01
7.00025082e-01 5.75606897e-02 4.47901100e-01 -9.19547796e-01
-3.37587669e-03 -1.96569294e-01 -1.33513156e-02 -4.53794688e-01
3.52105439e-01 -5.56162119e-01 2.05139816e-01 2.68724620e-01
-5.97424626e-01 6.87204540e-01 2.53959686e-01 8.58351886e-02
-3.10649961e-01 8.79177004e-02 7.04355121e-01 -3.89114946e-01
-7.06341207e-01 3.96200508e-01 -1.30915776e-01 8.55824053e-01
6.83862388e-01 -2.36517012e-01 -3.99258703e-01 -5.30098438e-01
-4.91703153e-01 1.56759769e-01 3.79474461e-01 6.28546655e-01
6.48337543e-01 -1.26623750e+00 -9.59251702e-01 -1.21392198e-02
1.97977439e-01 -1.42632902e-01 4.14936662e-01 9.14067984e-01
-4.47527468e-01 5.64431489e-01 -3.23710263e-01 -5.76478660e-01
-1.20520318e+00 6.88779771e-01 4.02664870e-01 -6.45115912e-01
-3.47854137e-01 1.11827135e+00 3.09434116e-01 -4.97038513e-01
2.33801097e-01 -7.87841618e-01 3.01979408e-02 1.07930943e-01
2.98528671e-01 4.83943611e-01 1.34313107e-01 -7.73865402e-01
-5.65974474e-01 7.54697263e-01 1.27095878e-01 -2.63398383e-02
1.22367239e+00 1.46033004e-01 -8.67819861e-02 4.68163431e-01
1.19771254e+00 -3.63979340e-01 -8.54103625e-01 -3.49147052e-01
-1.65010899e-01 -3.02052915e-01 -3.05349290e-01 -9.37901378e-01
-1.15350831e+00 1.17284060e+00 3.12816828e-01 3.18078637e-01
1.30605686e+00 -4.44880500e-02 6.88696206e-01 3.56062323e-01
3.98179561e-01 -7.18219221e-01 1.97118461e-01 7.64518857e-01
1.25523388e+00 -1.03412485e+00 -2.12332547e-01 5.72157912e-02
-7.72321761e-01 7.40180731e-01 5.08165359e-01 9.57802162e-02
4.04278755e-01 4.55036372e-01 -1.46236435e-01 -1.60636157e-01
-1.15718138e+00 -1.64331764e-01 3.02423328e-01 5.01427531e-01
6.35163546e-01 -5.32473922e-02 -7.08876476e-02 3.67682189e-01
-6.73140287e-01 1.35652006e-01 2.05444202e-01 5.35272300e-01
4.20428365e-02 -9.61120367e-01 -1.44961849e-01 6.23170137e-01
-5.77178538e-01 -5.39505780e-01 -7.22349435e-02 7.58482218e-01
4.91904430e-02 7.96319783e-01 -9.58038196e-02 -1.99826777e-01
5.06373703e-01 -5.59539944e-02 7.88618743e-01 -6.46922439e-02
-8.23086500e-01 -4.04285669e-01 2.71132886e-01 -6.83404207e-01
-5.33154726e-01 -7.03034163e-01 -7.51715541e-01 -6.99523032e-01
-1.09563097e-01 6.76288381e-02 3.63584101e-01 8.64322305e-01
1.96087733e-01 4.38603997e-01 1.57082215e-01 -7.51711309e-01
-6.63645089e-01 -8.90755534e-01 -1.73545897e-01 5.26529193e-01
1.92861706e-01 -7.03900337e-01 -2.73304462e-01 -1.97036080e-02] | [10.453051567077637, 2.092665672302246] |
ef9932b7-689f-46a3-bbe2-c208a93361ca | deepfake-video-forensics-based-on-transfer | 2004.14178 | null | https://arxiv.org/abs/2004.14178v1 | https://arxiv.org/pdf/2004.14178v1.pdf | Deepfake Video Forensics based on Transfer Learning | Deeplearning has been used to solve complex problems in various domains. As it advances, it also creates applications which become a major threat to our privacy, security and even to our Democracy. Such an application which is being developed recently is the "Deepfake". Deepfake models can create fake images and videos that humans cannot differentiate them from the genuine ones. Therefore, the counter application to automatically detect and analyze the digital visual media is necessary in today world. This paper details retraining the image classification models to apprehend the features from each deepfake video frames. After feeding different sets of deepfake clips of video fringes through a pretrained layer of bottleneck in the neural network is made for every video frame, already stated layer contains condense data for all images and exposes artificial manipulations in Deepfake videos. When checking Deepfake videos, this technique received more than 87 per cent accuracy. This technique has been tested on the Face Forensics dataset and obtained good accuracy in detection. | ['Tejeswinee K', 'Ragul M', 'Rahul U', 'Raja Vignesh K'] | 2020-04-29 | null | null | null | null | ['video-forensics'] | ['computer-vision'] | [-2.51081049e-01 2.84794141e-02 -5.78145124e-02 -2.77967602e-01
-9.40878019e-02 -4.57411200e-01 6.11341894e-01 -5.04587710e-01
-3.95478427e-01 5.15067160e-01 -2.08250329e-01 -3.36683005e-01
2.49199480e-01 -6.68128431e-01 -8.45385134e-01 -6.98269486e-01
-3.51548158e-02 3.92094590e-02 2.57964015e-01 -6.45811483e-02
6.40130579e-01 6.91807866e-01 -1.47883821e+00 9.08115327e-01
2.03500941e-01 1.04245782e+00 -1.40383527e-01 7.83788860e-01
-2.11674377e-01 1.03549123e+00 -9.54486430e-01 -9.75342810e-01
4.83773857e-01 -2.35994712e-01 -8.13336670e-01 1.46083131e-01
8.66253555e-01 -9.36837852e-01 -6.06439948e-01 1.38373303e+00
3.63440692e-01 -3.38791013e-01 5.61676145e-01 -1.48860717e+00
-7.27999926e-01 1.78672269e-01 -8.38091731e-01 7.51688778e-01
2.21630767e-01 1.27471432e-01 -1.49045661e-01 -1.00461209e+00
5.15121222e-01 1.47759128e+00 9.25084829e-01 7.70400107e-01
-6.30127847e-01 -1.24464703e+00 -6.34676993e-01 6.56898081e-01
-1.19009793e+00 -6.48483276e-01 6.26882851e-01 -7.11746454e-01
8.31953943e-01 8.46339899e-05 5.12932539e-01 1.03882813e+00
4.60750073e-01 6.89029574e-01 1.03489363e+00 -3.50966454e-01
-2.42047697e-01 4.46073532e-01 9.92667526e-02 9.66267169e-01
3.09921235e-01 3.94304961e-01 -3.57618034e-01 1.08693191e-03
6.54788673e-01 -6.66243955e-02 -3.43283787e-02 2.03228325e-01
-4.94848132e-01 1.01733661e+00 1.70994982e-01 4.43117380e-01
-1.91559702e-01 1.99901357e-01 6.93740606e-01 5.35072088e-01
1.59737587e-01 2.00658783e-01 -1.31312490e-01 3.72075103e-02
-1.14697123e+00 1.36938155e-01 4.37550575e-01 3.38918120e-01
5.86360335e-01 3.35875809e-01 2.16980889e-01 5.59036911e-01
1.74440682e-01 3.91601771e-01 9.01769400e-01 -7.95043349e-01
3.43789428e-01 3.96230072e-01 2.75696237e-02 -1.72448707e+00
-2.60490030e-02 1.34353071e-01 -6.88008249e-01 5.73963702e-01
6.06718361e-01 -2.66199149e-02 -1.09797847e+00 9.68091905e-01
4.35768634e-01 5.71954310e-01 -1.40837237e-01 9.54493701e-01
8.65774691e-01 8.84092391e-01 -4.67868000e-02 1.45200849e-01
1.34162354e+00 -7.21172512e-01 -9.17818248e-01 -1.20641738e-01
3.47689211e-01 -9.99436915e-01 7.59934843e-01 1.08832824e+00
-8.53999019e-01 -7.05612183e-01 -1.18939865e+00 -1.42304927e-01
-7.22015679e-01 1.44075915e-01 6.00921094e-01 1.10656381e+00
-1.01661837e+00 5.68329155e-01 -2.57772893e-01 -2.11945057e-01
9.38460767e-01 6.39464915e-01 -8.33907664e-01 5.25025725e-02
-1.05470836e+00 1.04126620e+00 6.04828715e-01 3.45816135e-01
-1.36132991e+00 -4.56690751e-02 -4.48583335e-01 -2.94512659e-02
2.21221551e-01 1.76820815e-01 1.06615722e+00 -1.64759481e+00
-1.11845911e+00 1.35246539e+00 2.49589860e-01 -6.72881961e-01
7.30334282e-01 -2.20829189e-01 -7.71297157e-01 2.72613615e-01
-1.82837054e-01 4.95863467e-01 1.69395995e+00 -9.84404206e-01
-5.55554390e-01 -3.51687044e-01 -1.75693125e-01 -6.47020102e-01
-3.91805887e-01 3.66287649e-01 -2.55913544e-03 -6.72568500e-01
-2.56966174e-01 -5.59553206e-01 5.89035273e-01 6.76487386e-02
-3.25236440e-01 -1.09376900e-01 1.69783854e+00 -9.81122196e-01
1.22345006e+00 -2.17767930e+00 -5.48729658e-01 1.52576536e-01
2.88236767e-01 1.11169505e+00 -2.13867486e-01 2.30893679e-02
-4.63699341e-01 1.56346485e-01 2.27118492e-01 -5.51213063e-02
-4.66316909e-01 1.33947864e-01 -3.95228356e-01 8.18719149e-01
2.98658282e-01 6.27830148e-01 -6.26760602e-01 -4.34317648e-01
3.22283387e-01 2.73129553e-01 -5.91606617e-01 3.55268240e-01
1.63081124e-01 4.08759378e-02 -1.52330503e-01 6.34030461e-01
1.26409149e+00 9.25888196e-02 -2.11834595e-01 -2.49549344e-01
6.25247061e-02 -2.66983181e-01 -8.08624148e-01 9.66501117e-01
-5.28066419e-02 1.28270721e+00 1.25436157e-01 -1.33142292e+00
7.85198331e-01 3.29195321e-01 3.65623049e-02 -7.97315061e-01
6.22124553e-01 1.63164362e-01 1.37506709e-01 -1.34404099e+00
4.81606662e-01 -2.38620445e-01 2.63664186e-01 4.47421074e-01
3.52639467e-01 3.01609457e-01 -2.52297252e-01 1.80696547e-01
8.29635501e-01 -6.13615625e-02 -5.01522869e-02 -1.85780257e-01
8.19167733e-01 -2.63293505e-01 2.17692092e-01 5.10275304e-01
-5.24922371e-01 1.54578537e-01 5.33061028e-01 -1.13078392e+00
-1.20777297e+00 -6.10806942e-01 1.29767925e-01 8.29644084e-01
-1.30311027e-02 -3.44741740e-04 -1.12445605e+00 -1.04536235e+00
1.86828151e-01 2.42737219e-01 -7.44841039e-01 -4.36248273e-01
-6.41040742e-01 -4.98264045e-01 1.01693809e+00 -4.57172049e-03
1.01205671e+00 -1.25817370e+00 -6.17440164e-01 3.73428836e-02
4.25657965e-02 -1.09047174e+00 -1.47438586e-01 -2.48056144e-01
-5.00167251e-01 -1.62399364e+00 -4.27823931e-01 -8.83992136e-01
4.63351339e-01 3.52449536e-01 6.54046416e-01 5.42237163e-01
-6.42878532e-01 -1.10927530e-01 -1.83886066e-01 -5.57967484e-01
-7.37611473e-01 -4.53486592e-01 -2.47030165e-02 2.42603630e-01
9.60710764e-01 -1.78852081e-01 -4.17114764e-01 1.02235332e-01
-1.03466368e+00 -3.41022283e-01 3.60368162e-01 7.78751016e-01
-1.42990202e-01 2.91113377e-01 6.35749459e-01 -8.52395713e-01
5.47828496e-01 -4.92147326e-01 -6.77495718e-01 5.85136786e-02
-1.78946048e-01 -6.07843921e-02 4.80813950e-01 -5.67318082e-01
-8.60023379e-01 -1.94221392e-01 -4.84081991e-02 -5.98047256e-01
-1.33861199e-01 -1.50126979e-01 -5.54114953e-02 -4.60413218e-01
6.32512212e-01 4.31983061e-02 2.58390903e-01 -3.04633737e-01
-2.34463383e-02 1.16191828e+00 7.42768228e-01 -1.33446306e-01
8.64725173e-01 5.17035544e-01 -1.21570498e-01 -1.09559047e+00
-4.22697842e-01 -1.76374272e-01 -2.56016821e-01 -5.94564974e-01
9.39441979e-01 -5.33383667e-01 -1.01273000e+00 1.16494572e+00
-1.53508878e+00 6.96899295e-02 5.08526981e-01 2.20165715e-01
5.01673184e-02 7.12931991e-01 -6.36188090e-01 -8.05157065e-01
-1.26220182e-01 -1.16306400e+00 6.16173923e-01 8.65655765e-02
1.03452191e-01 -7.20474005e-01 -2.87891984e-01 5.71276784e-01
2.49316558e-01 3.95242870e-01 7.72711277e-01 -5.67709506e-01
-4.91585612e-01 -4.05350983e-01 -4.72585410e-01 9.97293949e-01
5.66862449e-02 1.60344630e-01 -1.33090961e+00 -9.54711735e-02
2.95143545e-01 -3.93460184e-01 8.44641507e-01 4.29733768e-02
1.62241268e+00 -7.35666156e-01 -2.06160620e-01 6.98982596e-01
1.32631660e+00 5.38007557e-01 1.13372970e+00 5.06958842e-01
6.15064681e-01 6.13487363e-01 4.12034690e-01 2.20120087e-01
-1.52349368e-01 3.19193453e-01 6.75837815e-01 -4.65379609e-03
-1.16151921e-01 -1.04742222e-01 7.23241568e-01 2.93974757e-01
8.05323869e-02 -2.92060792e-01 -6.62766039e-01 3.40069473e-01
-1.36832988e+00 -1.50968981e+00 -2.95906663e-01 1.95130014e+00
3.79242748e-01 6.72615096e-02 8.89085233e-02 3.20068926e-01
1.05176234e+00 -7.64250569e-03 -4.02942061e-01 -9.51430142e-01
-2.38658395e-02 4.09089178e-01 6.50404036e-01 2.94148564e-01
-1.38590217e+00 1.32060575e+00 6.30364084e+00 8.83067787e-01
-1.70119846e+00 3.88285428e-01 7.96648562e-01 3.06058884e-01
3.44339281e-01 -3.80238920e-01 -5.62859595e-01 9.00030255e-01
9.03155804e-01 2.78489947e-01 4.73256052e-01 8.76854420e-01
3.60911876e-01 -2.19994679e-01 -8.96199405e-01 1.40918410e+00
3.68496060e-01 -1.45838022e+00 1.86023027e-01 1.84500575e-01
4.37064201e-01 -3.15136909e-01 2.44068235e-01 1.72310203e-01
-8.22723135e-02 -1.47479582e+00 6.75614238e-01 1.17421195e-01
7.84434438e-01 -8.64309132e-01 1.00898647e+00 2.39981502e-01
-4.74075049e-01 -4.56909388e-01 -6.37747824e-01 -4.05432209e-02
-2.58408576e-01 1.66202947e-01 -1.10469079e+00 -2.34303862e-01
8.45233738e-01 4.13654059e-01 -5.15543103e-01 8.96322370e-01
-3.36748213e-02 4.50826079e-01 -4.90491502e-02 2.42614895e-01
3.56323242e-01 4.28295061e-02 2.35353455e-01 1.31115544e+00
5.18021345e-01 -7.99532384e-02 -4.80595976e-01 4.62614775e-01
-1.88638583e-01 -2.25063816e-01 -8.07493627e-01 -3.18802983e-01
3.25588584e-01 1.15118623e+00 -6.18078649e-01 -4.40803200e-01
-3.16940159e-01 1.14885545e+00 -6.97986735e-03 -2.19994243e-02
-1.05458176e+00 -2.62260854e-01 7.01368570e-01 5.81644595e-01
3.41802120e-01 1.56733662e-01 2.30895266e-01 -1.06006217e+00
-3.07124138e-01 -1.22235084e+00 4.50360537e-01 -8.91361833e-01
-1.06502092e+00 6.39846563e-01 -2.20659554e-01 -1.05869555e+00
-1.91292092e-01 -1.20979333e+00 -4.83965009e-01 4.73664016e-01
-1.39183223e+00 -9.87921655e-01 -4.97962981e-01 1.00015557e+00
4.87154752e-01 -6.70103729e-01 5.06990492e-01 7.29896724e-01
-4.86104906e-01 6.67135954e-01 -6.69580922e-02 6.65044427e-01
8.14055502e-01 -5.93361020e-01 1.97524294e-01 9.97283638e-01
6.25474006e-02 4.36971694e-01 5.39818347e-01 -5.59646070e-01
-1.24141979e+00 -7.78719127e-01 5.71410179e-01 -2.86447287e-01
3.73911083e-01 -1.87436655e-01 -9.64173377e-01 6.55115306e-01
3.07029516e-01 1.54843643e-01 5.71116507e-01 -6.43398702e-01
-5.24013221e-01 -2.39399210e-01 -1.76386285e+00 -5.61498590e-02
5.03773749e-01 -5.98521411e-01 -7.45107174e-01 2.18395546e-01
-7.06808493e-02 -2.60176718e-01 -6.73049241e-02 -1.45317569e-01
7.46477962e-01 -1.44495046e+00 1.03127885e+00 -9.72711682e-01
5.20607531e-01 -4.14895773e-01 1.36975423e-01 -8.16172063e-01
1.01758540e-01 -5.76997161e-01 -2.64157772e-01 1.05980456e+00
-2.07583398e-01 -5.59592366e-01 1.07459486e+00 1.86766863e-01
1.29372463e-01 -7.87820965e-02 -9.57224846e-01 -5.55338860e-01
-5.51217943e-02 -2.39241183e-01 5.69211125e-01 1.33968425e+00
-4.34684783e-01 -1.34530455e-01 -9.88431931e-01 2.11097702e-01
5.68676770e-01 -5.20185232e-01 8.60865653e-01 -1.13187611e+00
-2.02985983e-02 -2.07241401e-01 -1.01424527e+00 -5.44035494e-01
9.91213769e-02 -6.85871243e-01 -4.21483934e-01 -7.17256248e-01
1.58941999e-01 8.22989345e-02 -8.16699117e-02 4.90313977e-01
2.59080976e-01 6.79611206e-01 1.35586947e-01 -3.20571251e-02
8.03932268e-03 -1.06932893e-01 1.10644579e+00 -2.51738042e-01
3.57528061e-01 -3.87207657e-01 -3.88965994e-01 9.70476031e-01
7.91711509e-01 -6.66400850e-01 -1.90671906e-01 -6.10458910e-01
-1.27790514e-02 -9.21363086e-02 7.27814734e-01 -1.14429033e+00
3.37274298e-02 3.56249069e-03 9.23528552e-01 -5.60168147e-01
3.24205190e-01 -1.02768338e+00 1.16748601e-01 7.25548506e-01
2.23045773e-03 7.71985054e-02 2.58658737e-01 2.18225777e-01
-3.45940471e-01 -4.99887794e-01 1.17034888e+00 -3.67552578e-01
-1.11658609e+00 1.74409807e-01 -4.37945187e-01 -3.89825791e-01
1.27629328e+00 -6.62654102e-01 -4.81311738e-01 -4.22878921e-01
-4.93391186e-01 -3.14081013e-01 2.05789655e-01 5.15521288e-01
9.40402389e-01 -1.17688799e+00 -4.37212050e-01 5.35929501e-01
-3.42362106e-01 -4.96721119e-01 3.68906617e-01 3.02569062e-01
-1.15313387e+00 9.30624157e-02 -7.55015850e-01 -2.14368552e-01
-1.73973048e+00 8.62725377e-01 4.98530149e-01 3.65130842e-01
-3.63557488e-01 9.76374090e-01 -3.64255905e-02 1.76248953e-01
1.44391671e-01 -1.55591905e-01 -3.68509382e-01 6.35661334e-02
1.03393555e+00 5.13922334e-01 -3.51856425e-02 -9.92835760e-01
-4.12474364e-01 3.56408924e-01 -2.69683868e-01 3.47036690e-01
1.29641676e+00 1.05935223e-01 -3.13636780e-01 -2.13458687e-01
1.54693711e+00 1.49293363e-01 -1.03434122e+00 4.45829123e-01
1.37443438e-01 -9.19167340e-01 2.00817555e-01 -6.84588134e-01
-1.47843409e+00 1.29759419e+00 1.08096898e+00 4.37778831e-01
1.00174880e+00 -4.15245414e-01 1.04110134e+00 3.05224001e-01
2.95427650e-01 -1.12092710e+00 4.56388474e-01 2.70659417e-01
5.88526607e-01 -1.39645195e+00 -1.31324515e-01 -1.01151831e-01
-4.21726823e-01 1.54866660e+00 6.32316768e-01 -4.73454028e-01
5.50946414e-01 3.07559550e-01 2.44733587e-01 -2.82030165e-01
-8.52902830e-02 4.79085743e-01 -2.32537203e-02 8.01129818e-01
1.87896058e-01 -1.35607302e-01 -2.31322020e-01 4.27283943e-01
-1.43828705e-01 2.33035281e-01 8.49923015e-01 4.97330785e-01
-5.91772914e-01 -1.06816411e+00 -6.91620529e-01 3.04406911e-01
-1.13222408e+00 2.04037875e-01 -5.98285198e-01 1.06148005e+00
7.09031641e-01 8.79718125e-01 1.62570611e-01 -5.45915782e-01
-2.55310684e-01 -9.51091647e-02 4.79060709e-01 -9.59765241e-02
-6.82233572e-01 -3.26002955e-01 -6.61943704e-02 -7.77127683e-01
-3.78126621e-01 -2.99841344e-01 -9.20349181e-01 -8.92716765e-01
-2.61769414e-01 2.31953055e-01 8.36970985e-01 8.53039265e-01
1.01532958e-01 -4.47491817e-02 4.91675109e-01 -7.95160055e-01
-3.61655623e-01 -8.07792306e-01 -3.52406442e-01 4.12401170e-01
7.71089494e-01 -6.42294466e-01 -3.46995473e-01 2.34916046e-01] | [12.57412052154541, 1.1897196769714355] |
bea65b2d-0771-45e2-9117-d9acc9cec76c | efficient-subsampling-for-generating-high | 2103.11166 | null | https://arxiv.org/abs/2103.11166v5 | https://arxiv.org/pdf/2103.11166v5.pdf | Efficient Subsampling of Realistic Images From GANs Conditional on a Class or a Continuous Variable | Recently, subsampling or refining images generated from unconditional GANs has been actively studied to improve the overall image quality. Unfortunately, these methods are often observed less effective or inefficient in handling conditional GANs (cGANs) -- conditioning on a class (aka class-conditional GANs) or a continuous variable (aka continuous cGANs or CcGANs). In this work, we introduce an effective and efficient subsampling scheme, named conditional density ratio-guided rejection sampling (cDR-RS), to sample high-quality images from cGANs. Specifically, we first develop a novel conditional density ratio estimation method, termed cDRE-F-cSP, by proposing the conditional Softplus (cSP) loss and an improved feature extraction mechanism. We then derive the error bound of a density ratio model trained with the cSP loss. Finally, we accept or reject a fake image in terms of its estimated conditional density ratio. A filtering scheme is also developed to increase fake images' label consistency without losing diversity when sampling from CcGANs. We extensively test the effectiveness and efficiency of cDR-RS in sampling from both class-conditional GANs and CcGANs on five benchmark datasets. When sampling from class-conditional GANs, cDR-RS outperforms modern state-of-the-art methods by a large margin (except DRE-F-SP+RS) in terms of effectiveness. Although the effectiveness of cDR-RS is often comparable to that of DRE-F-SP+RS, cDR-RS is substantially more efficient. When sampling from CcGANs, the superiority of cDR-RS is even more noticeable in terms of both effectiveness and efficiency. Notably, with the consumption of reasonable computational resources, cDR-RS can substantially reduce Label Score without decreasing the diversity of CcGAN-generated images, while other methods often need to trade much diversity for slightly improved Label Score. | ['William J. Welch', 'Z. Jane Wang', 'Yongwei Wang', 'Xin Ding'] | 2021-03-20 | null | null | null | null | ['density-ratio-estimation'] | ['methodology'] | [ 6.43304527e-01 2.77893960e-01 -2.77261704e-01 -2.05619082e-01
-1.27714109e+00 -4.27962005e-01 8.26396108e-01 -4.25634354e-01
-1.56944409e-01 1.11939573e+00 -1.91869643e-02 -1.08352408e-01
2.40146443e-01 -1.07870150e+00 -8.41253281e-01 -1.06410182e+00
4.44550276e-01 2.13816568e-01 -1.45032793e-01 2.14139774e-01
-1.82838226e-03 2.63799757e-01 -1.45204949e+00 2.91608553e-02
9.91498470e-01 1.17358887e+00 1.70049161e-01 4.33940411e-01
2.12854728e-01 8.08552682e-01 -7.60421753e-01 -7.57666647e-01
4.33146596e-01 -1.23643649e+00 -6.16321564e-01 8.09053555e-02
3.27356756e-01 -5.12142479e-01 1.96834341e-01 1.34218824e+00
5.03786922e-01 -1.47355378e-01 9.65876043e-01 -1.24293160e+00
-5.32268524e-01 2.67960429e-01 -8.02179098e-01 -4.44382638e-01
1.84372708e-01 2.09640548e-01 7.96926260e-01 -5.82792103e-01
5.42908847e-01 1.18647945e+00 6.10323727e-01 7.53244281e-01
-1.47284675e+00 -9.50388372e-01 3.70553955e-02 -2.75441051e-01
-1.41559291e+00 -3.13883454e-01 6.20178998e-01 -3.54152888e-01
2.23976463e-01 4.38704550e-01 5.85111439e-01 1.24951577e+00
-1.81484595e-02 7.99650788e-01 1.53275371e+00 -4.47099149e-01
5.03761709e-01 3.38733107e-01 -4.40719754e-01 6.50031388e-01
1.64767429e-01 3.64160806e-01 -4.55649942e-01 -2.74170637e-01
8.63085687e-01 -2.03187078e-01 -6.16359472e-01 -2.73925830e-02
-9.97613370e-01 1.01130199e+00 4.93884683e-01 3.85541059e-02
-3.65387768e-01 2.46058092e-01 1.70948386e-01 2.90948659e-01
6.70287013e-01 4.04987693e-01 -7.07818791e-02 4.76770364e-02
-1.19960725e+00 2.37763360e-01 4.04695123e-01 1.26043236e+00
8.35013807e-01 1.93364456e-01 -5.45748830e-01 7.85010219e-01
-1.09215073e-01 6.19712591e-01 3.53555620e-01 -8.08862448e-01
1.22963250e-01 1.97726727e-01 1.04585938e-01 -6.88607693e-01
2.57597506e-01 -5.89517534e-01 -1.20114589e+00 3.32531154e-01
3.06267172e-01 -1.74249206e-02 -1.11566126e+00 2.02731466e+00
2.78021753e-01 4.63729613e-02 -1.74304277e-01 7.87670493e-01
3.02532285e-01 6.79325283e-01 7.86829218e-02 -4.75847989e-01
1.30654502e+00 -8.80457580e-01 -7.01855242e-01 -2.60915384e-02
7.05354691e-01 -6.09239638e-01 1.16789508e+00 4.98590648e-01
-9.95898128e-01 -5.22181273e-01 -1.02097237e+00 1.86077416e-01
-6.16027229e-02 4.35872853e-01 3.42894375e-01 1.08754838e+00
-1.09247065e+00 3.77495617e-01 -5.95735490e-01 -1.10962898e-01
8.78427029e-01 1.88704446e-01 -1.98370025e-01 -3.11842501e-01
-9.50510263e-01 4.94390935e-01 9.44101065e-02 -2.16822490e-01
-1.15651762e+00 -9.25806284e-01 -7.26762950e-01 -1.45457596e-01
2.24278465e-01 -7.61851490e-01 9.08492088e-01 -1.43251920e+00
-1.79295564e+00 8.61173451e-01 -1.21110573e-01 -4.95787680e-01
1.05298162e+00 -2.77961902e-02 -2.52843589e-01 1.33205250e-01
2.77664900e-01 9.34750021e-01 1.39762866e+00 -1.40757847e+00
-6.07719481e-01 -1.62725493e-01 -5.32922707e-02 1.31075857e-02
-8.27021077e-02 -1.93388000e-01 -2.22735390e-01 -1.00074613e+00
-1.62488818e-01 -1.13621211e+00 -4.95568737e-02 3.40540595e-02
-7.28282034e-01 1.56255439e-02 8.38479936e-01 -5.89762807e-01
8.40368986e-01 -2.11230564e+00 6.20239116e-02 5.85932992e-02
1.03312746e-01 3.36092860e-01 1.21053271e-02 -3.83084901e-02
-3.80745046e-02 1.56571090e-01 -7.41779149e-01 -5.35293400e-01
-2.38730177e-01 -5.12683131e-02 -3.73390108e-01 5.59659481e-01
6.56449974e-01 8.97167504e-01 -9.18824673e-01 -1.73666924e-01
1.87966034e-01 6.15756333e-01 -6.15191817e-01 2.68667191e-01
-2.53990859e-01 7.81434894e-01 -2.41344556e-01 8.28144133e-01
1.04540288e+00 -1.59811586e-01 6.08246140e-02 -2.37774417e-01
3.20441484e-01 2.34588869e-02 -8.54694724e-01 1.27587891e+00
-4.90885288e-01 4.69208241e-01 -9.04999599e-02 -9.20104027e-01
8.67833495e-01 1.58527613e-01 -4.26958799e-02 -9.04324472e-01
6.83318973e-02 3.16109091e-01 -5.52948773e-01 1.18932180e-01
5.70967495e-01 -5.40130496e-01 -3.06153949e-02 3.63156140e-01
1.69897407e-01 -2.97248691e-01 -1.11872382e-01 2.05664068e-01
9.40302312e-01 2.68477619e-01 3.02359521e-01 -3.37997109e-01
5.33314586e-01 -3.48929316e-01 5.47279119e-01 7.80783534e-01
-8.50573927e-02 9.98683870e-01 5.71963489e-01 1.72037423e-01
-9.51139092e-01 -1.10044813e+00 -2.31069684e-01 6.71819031e-01
1.56436071e-01 -2.18575671e-01 -8.85471225e-01 -1.15263367e+00
-2.84283489e-01 8.40031981e-01 -5.96038043e-01 -3.34194452e-01
-2.92151749e-01 -1.00983584e+00 7.20297694e-01 2.87276864e-01
9.47795749e-01 -8.26799572e-01 -1.81456357e-01 -1.12747490e-01
-3.10046375e-01 -9.40102398e-01 -3.99223089e-01 5.76707125e-02
-6.66366279e-01 -8.13517034e-01 -1.21309507e+00 -5.07794082e-01
1.05670822e+00 -5.83660640e-02 1.05482328e+00 -4.97309342e-02
-5.76855354e-02 7.81109631e-02 -6.14341438e-01 -1.69494867e-01
-6.30833089e-01 -3.81660312e-02 -2.49944150e-01 3.05919111e-01
-7.78530911e-02 -3.31317008e-01 -7.32902944e-01 5.57783544e-01
-1.06461787e+00 1.19204313e-01 7.13177979e-01 1.18253708e+00
8.03389966e-01 1.25491709e-01 8.71329188e-01 -1.31122565e+00
1.49709284e-01 -3.51092279e-01 -6.35813117e-01 1.51290921e-02
-7.95969188e-01 -5.75100966e-02 8.09768915e-01 -4.09283847e-01
-1.19107783e+00 -6.31477311e-02 -2.45072380e-01 -4.53144044e-01
-1.44034863e-01 1.26522884e-01 -3.30376595e-01 -1.64773211e-01
5.69518864e-01 3.87721568e-01 -1.22958414e-01 -2.50120878e-01
4.66522306e-01 6.13599122e-01 5.71380079e-01 -5.52809775e-01
7.86163986e-01 6.64622724e-01 1.08631052e-01 -5.48528552e-01
-7.48523116e-01 -2.90314425e-02 -4.68712039e-02 -1.34466350e-01
7.80982554e-01 -1.10383761e+00 -1.72123417e-01 9.61447120e-01
-6.56235099e-01 -4.30084139e-01 -5.68233430e-01 4.26054776e-01
-6.32304609e-01 1.37548968e-01 -5.81380606e-01 -1.02151847e+00
-1.93447396e-01 -1.27768278e+00 1.23330259e+00 4.28913198e-02
6.14992827e-02 -7.70017147e-01 -3.45628917e-01 5.22465944e-01
4.58236575e-01 5.39053321e-01 9.57428753e-01 -1.21825077e-01
-7.33826816e-01 -1.88778251e-01 -4.73942637e-01 8.40333462e-01
3.11124295e-01 -2.58509845e-01 -1.26941228e+00 -7.21754193e-01
5.20549081e-02 -5.01570463e-01 1.02908933e+00 5.30940831e-01
1.36734295e+00 -4.55685675e-01 -1.79251999e-01 6.73973382e-01
1.64067388e+00 1.32315934e-01 1.06545496e+00 -2.73801712e-03
5.68768442e-01 4.16303426e-01 9.06580329e-01 3.86215597e-01
5.29997470e-03 5.65763593e-01 5.04561841e-01 -1.89118922e-01
-5.46855152e-01 -4.83692259e-01 5.62829673e-01 5.20906568e-01
1.90281067e-02 -5.29290378e-01 -8.06379840e-02 5.72246194e-01
-1.36887336e+00 -7.62247026e-01 4.19766782e-03 2.55573487e+00
1.09663153e+00 -5.22147082e-02 2.06842884e-01 2.08358943e-01
1.00832629e+00 1.71316504e-01 -4.43664223e-01 -1.30373642e-01
-2.88959712e-01 4.43651140e-01 6.06953084e-01 3.83180469e-01
-1.00291634e+00 7.52414882e-01 5.46115828e+00 1.45163321e+00
-1.18668008e+00 2.52482802e-01 1.24455845e+00 6.76966459e-03
-4.01940078e-01 1.37531102e-01 -7.53896534e-01 9.06380177e-01
6.75908327e-01 3.09457332e-01 4.62547421e-01 7.49725163e-01
-2.28668213e-01 -4.62250799e-01 -9.96938109e-01 9.59865153e-01
1.75278917e-01 -1.28310037e+00 1.16484836e-01 3.24845016e-01
1.24468565e+00 -3.97188842e-01 3.14106733e-01 3.01572561e-01
2.58772433e-01 -1.09009993e+00 8.97772849e-01 2.77217537e-01
1.56357265e+00 -9.00777876e-01 8.62263560e-01 1.50233701e-01
-8.05393338e-01 1.99273363e-01 -2.41495982e-01 2.00743839e-01
1.00151464e-01 1.14562559e+00 -4.36211765e-01 1.03423381e+00
5.93455613e-01 5.65314054e-01 -4.93661851e-01 5.48676491e-01
-3.82012993e-01 8.23380291e-01 -1.96482971e-01 3.22075427e-01
1.01356357e-01 -4.98834223e-01 4.53141630e-01 1.05856454e+00
5.55898309e-01 -1.40985280e-01 -2.42427871e-01 1.17407727e+00
-3.47484499e-01 -2.40289584e-01 -4.04390365e-01 9.74873006e-02
4.74830657e-01 1.15351701e+00 -7.44806409e-01 -4.32880998e-01
-9.63394567e-02 1.19295418e+00 7.39724338e-02 2.77427226e-01
-1.06136107e+00 -2.75486529e-01 4.43815023e-01 2.56621480e-01
5.52043557e-01 4.54764664e-01 -1.54270306e-01 -1.04456115e+00
1.55901862e-02 -1.00898659e+00 1.08307466e-01 -7.01392949e-01
-1.40989912e+00 6.55950785e-01 -4.47677970e-02 -1.42617655e+00
-1.60974264e-01 -1.31036922e-01 -3.68517309e-01 7.58688509e-01
-1.68226635e+00 -1.28377616e+00 -4.33657199e-01 6.02115512e-01
3.19499075e-01 -7.80314282e-02 7.61750400e-01 4.08303499e-01
-3.96063805e-01 9.81243432e-01 8.28521326e-02 -6.05070218e-02
8.49310160e-01 -1.15980494e+00 5.62509149e-02 9.06761944e-01
-1.56783909e-01 2.14761272e-01 4.37464356e-01 -7.08380222e-01
-8.85706842e-01 -1.56015992e+00 3.99884462e-01 -1.58762082e-01
-4.23813350e-02 -3.88852686e-01 -5.76074600e-01 5.60680330e-01
1.02615096e-01 1.72844574e-01 3.62953961e-01 -3.88245583e-01
-3.70369852e-01 -1.41408995e-01 -1.70822144e+00 4.30380642e-01
9.35070097e-01 -6.24304831e-01 4.49832045e-02 2.92771280e-01
5.90819836e-01 -1.38965920e-01 -7.49494910e-01 5.31891167e-01
3.13322157e-01 -1.27992725e+00 6.86996222e-01 1.06423952e-01
6.01687968e-01 -5.02796948e-01 -2.54857838e-01 -1.39075434e+00
-1.75200298e-01 -6.64502382e-01 2.19936326e-01 1.65889716e+00
1.87977761e-01 -7.95042276e-01 9.46874976e-01 1.12359107e-01
1.07856980e-02 -6.52036011e-01 -9.79800582e-01 -1.06306887e+00
1.32242933e-01 -2.58160412e-01 6.36696398e-01 1.10930300e+00
-7.12898672e-01 -4.91378792e-02 -7.60249555e-01 -1.17970474e-01
7.66336203e-01 -3.26350108e-02 8.71092975e-01 -6.98370218e-01
-3.74805003e-01 -3.31955403e-01 -3.79797399e-01 -1.08288050e+00
-3.67075130e-02 -7.10949242e-01 1.10889032e-01 -1.06333566e+00
1.37547910e-01 -7.35482335e-01 -7.88834244e-02 1.84642136e-01
-2.98497736e-01 7.95232832e-01 1.96039706e-01 1.96635470e-01
-1.24082416e-01 8.29089701e-01 1.33903718e+00 -1.02306247e-01
7.30903912e-03 3.26027684e-02 -5.70057988e-01 4.06516939e-01
6.92363560e-01 -7.01263249e-01 -4.18514282e-01 8.04128274e-02
1.09184824e-01 -7.61869997e-02 6.64166987e-01 -1.01939130e+00
-2.24850357e-01 1.34448022e-01 4.18114990e-01 -5.41262209e-01
2.87551343e-01 -5.31372845e-01 5.55252314e-01 4.32091236e-01
-1.99858353e-01 -5.16981661e-01 -2.57845998e-01 8.41146767e-01
-3.73122454e-01 -1.73001558e-01 1.19068074e+00 -9.90619659e-02
-1.48119912e-01 1.45943686e-01 -3.00408453e-01 1.18708961e-01
9.88316417e-01 -2.70763308e-01 -3.41154307e-01 -5.20560980e-01
-1.69295013e-01 -3.20466965e-01 7.57251799e-01 -3.31181921e-02
4.63990152e-01 -1.51696682e+00 -6.68099344e-01 5.70600510e-01
1.80315524e-01 2.97870517e-01 2.21522927e-01 9.63182330e-01
-4.65951771e-01 8.96241218e-02 -7.26862773e-02 -6.53818846e-01
-8.91728878e-01 3.61742169e-01 1.33980662e-01 -4.91525829e-01
-3.88981938e-01 1.14999402e+00 6.65785968e-01 -3.84709418e-01
-7.80915543e-02 -1.49949700e-01 3.50457162e-01 -3.99866290e-02
3.26761305e-01 3.98742020e-01 2.24801376e-01 -5.49987495e-01
-1.68510318e-01 4.10220087e-01 2.18501836e-02 -3.67026478e-02
9.18701589e-01 -1.55099444e-02 -1.36094049e-01 -4.29062173e-02
1.20982349e+00 1.76673487e-01 -1.58531940e+00 6.42808080e-02
-4.84801769e-01 -8.72591734e-01 1.84587538e-01 -1.01073170e+00
-1.53124917e+00 5.17014265e-01 6.89871192e-01 1.26031920e-01
1.56643689e+00 -2.68951844e-04 7.00088024e-01 -4.87786233e-01
6.09177053e-01 -8.08035612e-01 1.33176297e-01 -7.89508298e-02
7.02181220e-01 -1.31042087e+00 1.57658309e-01 -5.48052311e-01
-7.95286417e-01 5.89918971e-01 3.85286391e-01 -3.51609766e-01
4.47511166e-01 4.63977903e-02 -1.83123976e-01 1.26711279e-01
-3.73039842e-01 3.82806128e-03 6.45761192e-02 7.58522689e-01
1.51893750e-01 2.42982820e-01 -3.06755275e-01 3.84727627e-01
-3.02793622e-01 -2.81977896e-02 4.89567727e-01 7.02590942e-01
2.22504511e-01 -1.04301345e+00 -2.65291303e-01 7.24509895e-01
-5.24359763e-01 -1.16579920e-01 -1.93440914e-01 6.25804543e-01
3.42650235e-01 9.79343832e-01 3.83843817e-02 -4.02432621e-01
-1.82711229e-01 -1.75057799e-01 5.11365652e-01 -2.88722068e-01
-5.52902400e-01 2.18786314e-01 5.22948205e-02 -4.79626149e-01
-6.96784556e-01 -6.02491915e-01 -6.89052045e-01 -4.18820739e-01
-9.81907725e-01 -1.05992921e-01 5.18710732e-01 7.95626938e-01
3.18845540e-01 4.39344466e-01 9.70834672e-01 -6.10888541e-01
-4.23297107e-01 -7.92192638e-01 -8.34827006e-01 3.52709711e-01
3.36060017e-01 -6.06850863e-01 -7.42325366e-01 1.77170790e-03] | [11.590516090393066, -0.2890143096446991] |
cb61a04d-4c65-46f7-bcf1-00ce8533740e | progressive-sub-graph-clustering-algorithm | 2305.12703 | null | https://arxiv.org/abs/2305.12703v1 | https://arxiv.org/pdf/2305.12703v1.pdf | Progressive Sub-Graph Clustering Algorithm for Semi-Supervised Domain Adaptation Speaker Verification | Utilizing the large-scale unlabeled data from the target domain via pseudo-label clustering algorithms is an important approach for addressing domain adaptation problems in speaker verification tasks. In this paper, we propose a novel progressive subgraph clustering algorithm based on multi-model voting and double-Gaussian based assessment (PGMVG clustering). To fully exploit the relationships among utterances and the complementarity among multiple models, our method constructs multiple k-nearest neighbors graphs based on diverse models and generates high-confidence edges using a voting mechanism. Further, to maximize the intra-class diversity, the connected subgraph is utilized to obtain the initial pseudo-labels. Finally, to prevent disastrous clustering results, we adopt an iterative approach that progressively increases k and employs a double-Gaussian based assessment algorithm to decide whether merging sub-classes. | ['Pengyuan Zhang', 'Wenchao Wang', 'Zhenduo Zhao', 'Jingze Lu', 'Zhuo Li'] | 2023-05-22 | null | null | null | null | ['graph-clustering', 'pseudo-label', 'speaker-verification'] | ['graphs', 'miscellaneous', 'speech'] | [ 2.67672986e-02 1.27168521e-01 -4.16902423e-01 -8.33475828e-01
-1.01450646e+00 -6.69398904e-01 2.80592442e-01 -1.85397286e-02
6.70334324e-02 6.17714047e-01 2.08216935e-01 -3.55096281e-01
-3.03754449e-01 -2.84608364e-01 -8.85125026e-02 -9.04834986e-01
2.41744563e-01 5.51724613e-01 3.11754704e-01 2.29881480e-01
3.05106461e-01 4.83517051e-01 -1.39000785e+00 8.27892944e-02
1.33095098e+00 6.29883766e-01 1.19073559e-02 3.02606314e-01
-2.94883639e-01 3.87424409e-01 -3.19418341e-01 -3.72507006e-01
8.96764621e-02 -7.09521294e-01 -7.46513128e-01 6.12896919e-01
2.81529009e-01 2.40217701e-01 -1.29894003e-01 1.47862434e+00
3.97075623e-01 3.62097532e-01 8.71707022e-01 -1.34974051e+00
-1.97038949e-01 8.42949808e-01 -7.92175710e-01 -1.68590948e-01
1.67300805e-01 -2.31568664e-01 1.01624048e+00 -8.54136765e-01
5.52696168e-01 1.32550788e+00 4.66642618e-01 5.71201384e-01
-1.15035427e+00 -9.63966489e-01 4.57872480e-01 4.39607382e-01
-1.78583109e+00 -5.13154089e-01 1.09119260e+00 -1.90402210e-01
3.16946238e-01 2.18980685e-01 4.17856097e-01 6.54264390e-01
-3.87081385e-01 5.82845449e-01 1.03295410e+00 -6.54048979e-01
4.49365944e-01 2.85242587e-01 4.90991354e-01 7.70421684e-01
7.71605468e-04 -1.87450603e-01 -4.24083740e-01 -6.62383020e-01
1.45724118e-01 -2.67498285e-01 -3.44643593e-01 -6.82564795e-01
-8.24240744e-01 8.51062953e-01 2.02579096e-01 4.62837458e-01
-1.62527114e-01 -4.17939425e-01 5.04861660e-02 1.04437314e-01
2.00399935e-01 1.02663264e-02 -4.01757479e-01 3.05335879e-01
-1.23940718e+00 -2.86461622e-01 6.14759088e-01 1.05810714e+00
1.21124232e+00 -3.73147283e-04 7.79595450e-02 1.01894462e+00
9.41776454e-01 1.63840443e-01 5.10952115e-01 -9.67040718e-01
2.96712995e-01 8.66227686e-01 -2.88511097e-01 -8.78288329e-01
-2.99840510e-01 -4.36168879e-01 -7.63083875e-01 -6.99290857e-02
3.55516225e-01 -1.71852604e-01 -1.12614274e+00 1.84785676e+00
8.30144286e-01 6.35762393e-01 2.53726929e-01 7.01996684e-01
6.68665707e-01 3.13352734e-01 3.83343250e-01 -5.92911899e-01
1.16404474e+00 -7.43755817e-01 -5.72758436e-01 4.89393771e-02
7.55840540e-01 -6.09411895e-01 5.13927102e-01 2.94240654e-01
-4.67211694e-01 -5.57635546e-01 -1.02486074e+00 4.55040723e-01
-8.80360603e-02 7.68232420e-02 1.74738169e-01 9.29753065e-01
-1.09500515e+00 1.41059533e-01 -5.42159200e-01 -3.19626480e-01
2.80588627e-01 4.91285890e-01 -2.85483956e-01 -3.64461869e-01
-1.09128165e+00 5.15663266e-01 7.50768006e-01 -2.04914644e-01
-7.23754764e-01 -2.13707224e-01 -8.88695359e-01 -1.13969743e-01
3.48838717e-01 -2.10912094e-01 8.30369651e-01 -8.41534793e-01
-1.47393799e+00 7.95396388e-01 -5.69278777e-01 -8.74418318e-02
1.64264798e-01 6.95236981e-01 -7.21383870e-01 1.50692657e-01
-1.04059726e-01 5.91144323e-01 1.05856287e+00 -1.52952659e+00
-8.03485632e-01 -7.14592814e-01 -2.00327590e-01 5.20169675e-01
-4.17936057e-01 -5.91541789e-02 -7.49425113e-01 -4.65441555e-01
9.42903578e-01 -1.21297872e+00 -4.68995571e-01 -6.89700961e-01
-6.16999269e-01 -4.96983021e-01 1.09177816e+00 -7.62101829e-01
1.40927064e+00 -2.41827512e+00 2.61355907e-01 9.42092896e-01
3.71720791e-01 8.97115096e-02 -1.08773544e-01 -8.95104855e-02
1.27650857e-01 -1.12781459e-02 -5.92064023e-01 -2.53870219e-01
-1.29430562e-01 2.76449532e-03 3.10501635e-01 4.60044920e-01
-2.05237016e-01 3.67809892e-01 -8.36969018e-01 -9.81247067e-01
1.00130521e-01 2.67540395e-01 -3.27922195e-01 4.00467105e-02
-9.44085419e-02 5.54659307e-01 -5.61037183e-01 7.08647251e-01
1.00764763e+00 -3.03691447e-01 9.06769633e-01 -2.61521995e-01
2.84627318e-01 -6.45739660e-02 -1.56015360e+00 1.61345124e+00
7.48096639e-03 2.58376569e-01 3.08446020e-01 -1.03484488e+00
1.15211570e+00 3.57154906e-01 4.10317659e-01 -2.00611856e-02
1.36592194e-01 3.35928679e-01 -1.12842746e-01 1.10482285e-02
3.31067950e-01 -2.40373649e-02 -1.03193887e-01 4.79261607e-01
1.42572775e-01 -4.17586304e-02 -5.57623990e-02 5.50288200e-01
7.22621381e-01 -2.52176523e-01 4.34382379e-01 -3.26584011e-01
9.03419077e-01 1.12746254e-01 9.14071858e-01 3.41767192e-01
-6.17053032e-01 5.33553481e-01 1.93887293e-01 2.56817400e-01
-5.57228625e-01 -9.36384201e-01 -1.47266999e-01 9.07587588e-01
3.38463575e-01 -2.21824929e-01 -9.29038405e-01 -1.09655797e+00
1.41849369e-03 6.83429480e-01 -3.64190936e-01 -2.81157970e-01
-3.64628583e-01 -6.82192922e-01 4.33509350e-01 2.30873436e-01
5.25892615e-01 -6.47275865e-01 3.50760907e-01 2.59285122e-01
-4.05858636e-01 -9.67986882e-01 -6.37730360e-01 1.83660954e-01
-7.46707559e-01 -1.28183186e+00 -5.93759596e-01 -1.26083887e+00
9.97343063e-01 5.06215394e-01 5.63938379e-01 -1.77320670e-02
1.03534400e-01 3.10570866e-01 -4.41386640e-01 1.98683798e-01
-6.31814718e-01 1.54118016e-01 2.48877242e-01 3.79484624e-01
7.80348778e-01 -4.77286100e-01 -2.71985441e-01 4.92414773e-01
-3.52190256e-01 -4.01306868e-01 3.09361368e-01 7.82778203e-01
6.84520960e-01 3.52545649e-01 7.94151306e-01 -8.04532468e-01
4.72868204e-01 -4.07049894e-01 -7.57417083e-01 6.53302133e-01
-7.34170973e-01 1.34319052e-01 4.15027112e-01 -5.10464787e-01
-1.34728956e+00 4.84694004e-01 1.70589149e-01 -5.40642381e-01
-1.15972042e-01 4.39441383e-01 -6.55347168e-01 -2.94955701e-01
4.49367702e-01 2.51388073e-01 -2.20078290e-01 -2.55995154e-01
6.93366289e-01 1.09662807e+00 2.82709301e-01 -2.60654330e-01
7.54794657e-01 2.79238373e-01 -3.46231371e-01 -6.98917568e-01
-4.47304189e-01 -8.27253997e-01 -9.18397307e-01 -3.56690973e-01
8.86411369e-01 -1.03234148e+00 -5.22132456e-01 4.10238415e-01
-8.08491766e-01 2.45044474e-02 2.82133788e-01 8.78974676e-01
-8.26769918e-02 9.12700415e-01 -5.11867106e-01 -1.04026878e+00
-7.31786117e-02 -1.11726141e+00 5.18697321e-01 4.68428046e-01
-1.35385454e-01 -7.90617883e-01 -4.25118469e-02 5.32162547e-01
-8.23709220e-02 -2.54439145e-01 9.04479206e-01 -1.10948312e+00
-5.13675749e-01 1.48717538e-01 -1.99299231e-01 8.19583535e-02
2.98347145e-01 -5.64689264e-02 -1.09696651e+00 -4.20673877e-01
-1.00494258e-01 -2.45666310e-01 7.04063058e-01 5.08361459e-01
8.81516099e-01 6.82847649e-02 -7.62970805e-01 2.72600174e-01
1.11038852e+00 5.12814701e-01 1.03651673e-01 -8.19828585e-02
8.53845656e-01 8.02666545e-01 7.97039926e-01 4.70821798e-01
6.35361016e-01 5.22037983e-01 -9.12214518e-02 2.21350327e-01
-6.84359074e-02 -2.36401975e-01 1.19460240e-01 1.10174453e+00
4.21142876e-01 -1.46713629e-01 -8.34250808e-01 5.04294991e-01
-1.74673951e+00 -7.93733239e-01 -1.28332570e-01 2.28067827e+00
7.75153458e-01 6.70727575e-03 1.20595090e-01 1.29930928e-01
1.53783631e+00 -1.66320354e-02 -6.88638449e-01 -5.01946025e-02
-1.61653206e-01 -3.02189976e-01 2.40649536e-01 6.45796001e-01
-1.07307971e+00 1.15186357e+00 6.14023447e+00 1.07971537e+00
-8.16270530e-01 8.75962451e-02 7.51376569e-01 4.69262689e-01
-4.37739402e-01 2.14313760e-01 -1.10998464e+00 3.52392763e-01
7.06920326e-01 -2.07791060e-01 3.51845115e-01 9.05247748e-01
1.07743897e-01 6.54372349e-02 -6.43619239e-01 8.98850143e-01
1.83770761e-01 -7.87887931e-01 -1.31931975e-01 7.78399706e-02
9.77707505e-01 -1.40033260e-01 -2.72173852e-01 2.03493953e-01
1.01281488e+00 -3.35838079e-01 3.97580415e-01 1.15976082e-02
6.81178689e-01 -9.09694493e-01 3.18937778e-01 4.48342800e-01
-1.67588830e+00 -1.32886544e-01 -2.22642601e-01 7.54959106e-01
2.37919375e-01 6.14217699e-01 -9.37342405e-01 7.75288582e-01
4.75677103e-01 3.86201739e-01 -4.35213447e-01 9.49677646e-01
-3.03386271e-01 6.99805021e-01 -2.97858059e-01 8.66474807e-02
1.59133255e-01 -4.54465508e-01 5.02865314e-01 1.12430489e+00
2.78200746e-01 2.33145297e-01 4.95291680e-01 2.47989967e-01
-1.59392625e-01 2.55923629e-01 -3.83070201e-01 3.28943253e-01
1.14488149e+00 1.23912179e+00 -8.64108801e-01 -6.66987360e-01
-4.38645750e-01 8.92136872e-01 4.22408909e-01 5.34148097e-01
-7.33691275e-01 -3.08747083e-01 3.65604192e-01 -3.17171276e-01
2.83221751e-01 -1.59740314e-01 -2.11619779e-01 -1.08034658e+00
-8.55538845e-02 -9.69232202e-01 8.67547631e-01 -3.73336971e-01
-1.07798827e+00 7.66450286e-01 -8.01243484e-02 -1.39003384e+00
-2.26730496e-01 2.42655948e-02 -5.32707453e-01 6.65331662e-01
-1.57642961e+00 -1.10415518e+00 -2.22503871e-01 9.73790348e-01
3.59328330e-01 -3.02331597e-01 6.66339576e-01 2.67406732e-01
-5.13383687e-01 9.35336471e-01 1.71013013e-01 1.11511797e-01
8.33825648e-01 -9.63859618e-01 -5.73386736e-02 9.09725010e-01
2.42072389e-01 5.74800730e-01 2.67230481e-01 -8.21914971e-01
-7.71028996e-01 -1.20884728e+00 1.01750481e+00 8.73338357e-02
4.07657743e-01 -2.07471818e-01 -1.14070213e+00 5.95582485e-01
-7.99681898e-03 -3.16371739e-01 1.00847197e+00 2.09511757e-01
-6.05650961e-01 -7.73933828e-02 -1.47148705e+00 4.40304875e-01
9.97969747e-01 -6.36587679e-01 -5.17776489e-01 1.58766016e-01
6.22033715e-01 1.77747924e-02 -7.57040441e-01 5.44810712e-01
2.37762451e-01 -6.42203689e-01 5.33029258e-01 -2.86400139e-01
-5.99559307e-01 -7.52676666e-01 -2.81642407e-01 -1.35257339e+00
-5.98491013e-01 -5.84525228e-01 8.92943963e-02 1.57465684e+00
7.95715988e-01 -6.53830111e-01 1.13230610e+00 6.59832001e-01
-2.21297488e-01 -6.83095083e-02 -1.03976572e+00 -5.98948956e-01
-4.26309764e-01 -3.01560253e-01 6.17927611e-01 1.46595562e+00
4.48373020e-01 2.01085180e-01 -2.81169832e-01 5.96954584e-01
1.06075931e+00 2.54682660e-01 4.95461792e-01 -1.33373070e+00
-2.21964657e-01 -2.78203398e-01 -3.89086127e-01 -1.02141118e+00
6.45041287e-01 -1.00559592e+00 1.84698597e-01 -1.14535320e+00
3.25467736e-01 -6.94569707e-01 -5.19438863e-01 3.59126121e-01
-4.08613831e-01 2.92051196e-01 -3.18274349e-02 3.85641992e-01
-9.06405509e-01 3.80763948e-01 8.42510700e-01 -2.70978242e-01
-5.77154815e-01 1.80528522e-01 -6.83017671e-01 5.78444958e-01
8.52590978e-01 -5.81749856e-01 -4.88629222e-01 7.81682730e-02
-7.43819594e-01 1.87599555e-01 -2.80313373e-01 -8.44892204e-01
5.53659201e-01 -1.43689558e-01 2.09237278e-01 -7.77783394e-01
3.55779529e-01 -8.27120185e-01 2.47486278e-01 4.32462603e-01
-3.34508538e-01 -4.68697697e-01 -1.25643402e-01 9.23031449e-01
-4.77429986e-01 -1.74115822e-01 1.21449649e+00 -1.28897373e-02
-9.79230821e-01 3.76512945e-01 -3.74380320e-01 -2.20622644e-01
1.26465821e+00 -3.40141296e-01 1.27521202e-01 -4.40218151e-01
-1.18037271e+00 5.87904096e-01 4.87796664e-01 3.31401080e-01
6.45091295e-01 -1.49361420e+00 -7.21944213e-01 1.52405307e-01
4.63352740e-01 -3.28411549e-01 4.95117486e-01 4.61633146e-01
5.95781431e-02 9.38395709e-02 3.26353759e-01 -8.25527191e-01
-1.80091202e+00 5.31424642e-01 1.32051632e-01 -2.67724246e-01
-1.29419044e-01 9.22589064e-01 1.60871819e-02 -8.30971301e-01
2.81526893e-01 2.34237149e-01 -4.32196587e-01 1.59982175e-01
2.64614284e-01 2.38922969e-01 2.58475188e-02 -9.79639769e-01
-6.95149422e-01 7.45431483e-01 -1.55670047e-01 -2.36372605e-01
7.40638196e-01 -5.56957662e-01 8.98807198e-02 4.40044180e-02
1.19588518e+00 8.22620988e-02 -1.19580007e+00 -5.59548974e-01
3.36076140e-01 -2.88490862e-01 -5.95859699e-02 -5.28138459e-01
-1.10467947e+00 4.97751504e-01 5.52386463e-01 -1.55696794e-02
1.13637817e+00 1.96435153e-01 3.31862658e-01 1.84318796e-01
4.75557208e-01 -1.26515591e+00 -9.83626917e-02 2.27600008e-01
1.38468564e-01 -1.20326257e+00 -1.03549697e-01 -7.83362746e-01
-8.83271575e-01 8.57741296e-01 6.68672442e-01 4.25726950e-01
9.79616225e-01 -7.88545795e-03 4.60564762e-01 1.46565558e-02
-3.04187894e-01 -3.08301598e-01 3.19819301e-01 7.60705590e-01
1.22174397e-01 3.00178468e-01 -3.46140921e-01 3.11872363e-01
1.86774835e-01 -3.18300933e-01 1.81676477e-01 6.90352440e-01
-7.35140622e-01 -1.43781543e+00 -5.69719374e-01 3.87504637e-01
-2.09306106e-02 -5.91165759e-02 -5.27306795e-01 3.26837182e-01
1.69900781e-03 1.54132760e+00 -3.20612639e-01 -8.95602763e-01
-7.40400553e-02 3.91548991e-01 2.15994701e-01 -5.50706089e-01
-3.00413489e-01 4.85218555e-01 1.03638798e-01 -4.60460633e-02
-6.40657127e-01 -7.72795200e-01 -1.43871462e+00 -1.25095442e-01
-8.76218796e-01 6.74758852e-01 5.92871130e-01 9.36110795e-01
4.91370052e-01 1.16697915e-01 1.15873611e+00 -3.82233322e-01
-4.75574672e-01 -1.02063191e+00 -8.58584940e-01 2.63021350e-01
-1.01555347e-01 -5.35883307e-01 -7.08346844e-01 5.59059232e-02] | [14.859437942504883, 1.196321725845337] |
980c70d4-dd2a-46ae-9d32-cb3ef1f402b8 | which-k-space-sampling-schemes-is-good-for | 2103.08516 | null | https://arxiv.org/abs/2103.08516v1 | https://arxiv.org/pdf/2103.08516v1.pdf | Which K-Space Sampling Schemes is good for Motion Artifact Detection in Magnetic Resonance Imaging? | Motion artifacts are a common occurrence in the Magnetic Resonance Imaging (MRI) exam. Motion during acquisition has a profound impact on workflow efficiency, often requiring a repeat of sequences. Furthermore, motion artifacts may escape notice by technologists, only to be revealed at the time of reading by the radiologists, affecting their diagnostic quality. Designing a computer-aided tool for automatic motion detection and elimination can improve the diagnosis, however, it needs a deep understanding of motion characteristics. Motion artifacts in MRI have a complex nature and it is directly related to the k-space sampling scheme. In this study we investigate the effect of three conventional k-space samplers, including Cartesian, Uniform Spiral and Radial on motion induced image distortion. In this regard, various synthetic motions with different trajectories of displacement and rotation are applied to T1 and T2-weighted MRI images, and a convolutional neural network is trained to show the difficulty of motion classification. The results show that the spiral k-space sampling method get less effect of motion artifact in image space as compared to radial k-space sampled images, and radial k-space sampled images are more robust than Cartesian ones. Cartesian samplers, on the other hand, are the best in terms of deep learning motion detection because they can better reflect motion. | ['Khan A. Wahid', 'Ekta Walia', 'Mohammad Reza Mohebbian'] | 2021-03-15 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 1.85225740e-01 -4.05837297e-01 -8.89418274e-02 -1.73882693e-02
-4.90947485e-01 -5.20209193e-01 2.43976966e-01 -1.91741109e-01
-6.60088778e-01 4.18141603e-01 1.68989390e-01 -5.27002692e-01
-4.47483599e-01 -2.23819777e-01 -3.76459330e-01 -9.52744603e-01
-3.01749229e-01 9.81350988e-02 5.25216818e-01 2.22256333e-01
3.95728141e-01 5.95248461e-01 -8.21891904e-01 2.49410868e-01
7.68434048e-01 6.11589551e-01 5.47857940e-01 6.52657628e-01
5.97916208e-02 9.88815248e-01 -5.04282594e-01 2.30711892e-01
4.17259037e-02 -5.80962896e-01 -1.05219483e+00 -1.04934424e-01
-5.61845526e-02 -4.47742999e-01 -4.43074286e-01 1.01697373e+00
7.63027072e-01 1.35649726e-01 5.83459496e-01 -6.01219535e-01
-4.51304793e-01 4.91928339e-01 -6.15990400e-01 8.36070657e-01
-4.05764207e-02 3.43587667e-01 -2.93446660e-01 -6.49088502e-01
7.53306746e-01 6.87115490e-01 8.31528306e-01 4.86908019e-01
-8.89656603e-01 -4.97284710e-01 -3.65767419e-01 5.79998732e-01
-1.02573657e+00 -1.83331847e-01 9.19199526e-01 -6.64597392e-01
6.52735889e-01 3.84196609e-01 6.15082920e-01 9.08224463e-01
8.88039708e-01 6.61793351e-01 1.03866982e+00 -3.39989215e-01
1.64452970e-01 -2.44137257e-01 1.08151816e-01 3.09678018e-01
2.37558320e-01 -1.50560290e-01 -4.41042706e-02 2.45343857e-02
1.09924936e+00 -7.52810668e-03 -8.13993752e-01 -5.56767106e-01
-1.54368055e+00 4.77040648e-01 4.37217623e-01 8.32908630e-01
-3.31182748e-01 6.52925819e-02 5.97285390e-01 2.11181119e-03
4.66861948e-02 6.11505747e-01 -2.92026028e-02 -3.07699651e-01
-1.11757731e+00 -2.58738473e-02 8.41832682e-02 3.21533203e-01
-2.22970039e-01 4.67728414e-02 -3.47202897e-01 6.78276479e-01
5.69272339e-02 3.80204737e-01 1.36391556e+00 -7.78778553e-01
1.88254178e-01 1.50350422e-01 1.19204231e-01 -1.02539515e+00
-9.61812377e-01 -5.52615643e-01 -1.15911055e+00 1.04261547e-01
6.95083618e-01 2.66003340e-01 -9.33160543e-01 1.15123320e+00
4.30663288e-01 1.68581873e-01 -2.65889376e-01 1.49022591e+00
6.52176023e-01 2.72488177e-01 -1.07965305e-01 -5.48506379e-01
1.40829253e+00 -8.73029828e-01 -1.40884531e+00 8.58397689e-04
9.74986076e-01 -9.99026895e-01 9.97893691e-01 5.79495132e-01
-1.19507718e+00 -5.85557342e-01 -1.07413030e+00 2.33961657e-01
8.72760266e-02 3.20705652e-01 5.10305524e-01 7.66193211e-01
-7.25172102e-01 9.58516419e-01 -1.31787181e+00 2.36587912e-01
2.83550829e-01 1.41966313e-01 -3.24956894e-01 -2.42216945e-01
-1.28549898e+00 1.22001743e+00 7.40537643e-02 4.89565253e-01
-5.90877354e-01 -8.12258244e-01 -5.87868810e-01 -4.40720558e-01
1.96305200e-01 -3.21419775e-01 1.43239093e+00 -6.05963767e-01
-1.27209675e+00 6.21484220e-01 2.72174701e-02 -1.57677591e-01
9.78387117e-01 -1.04495361e-01 -5.75391531e-01 4.69831318e-01
1.56367067e-02 1.89831764e-01 1.05931616e+00 -7.51800537e-01
4.96401861e-02 -3.08528841e-01 -4.05705124e-01 -7.04428405e-02
1.02356464e-01 1.31139010e-02 -1.51560992e-01 -7.70041943e-01
6.36425436e-01 -1.08476806e+00 -3.97834957e-01 -2.10760627e-02
-1.40803456e-01 2.94549167e-01 7.73725271e-01 -9.76705253e-01
1.51633048e+00 -2.01344299e+00 -2.86845773e-01 1.52993528e-02
2.63612419e-01 6.20019615e-01 2.97661901e-01 -1.51932538e-01
-6.17902160e-01 1.48048773e-02 -3.04694355e-01 4.31541741e-01
-6.91789985e-01 -2.10166201e-01 2.30196804e-01 8.52145791e-01
7.20300674e-02 8.44761372e-01 -1.04513216e+00 -3.67949903e-01
3.23597282e-01 3.12516958e-01 -3.79172981e-01 3.03359106e-02
5.74095368e-01 9.29578781e-01 -3.08155149e-01 1.70803517e-01
1.06282365e+00 -3.32436800e-01 2.47008100e-01 -5.47290087e-01
-2.01778635e-01 2.24285454e-01 -1.08288491e+00 1.81844771e+00
-4.00869906e-01 7.05169201e-01 -1.57270357e-01 -7.82848120e-01
3.84065270e-01 5.71960151e-01 6.69682860e-01 -9.67001140e-01
5.31217568e-02 3.33182722e-01 5.39616466e-01 -1.20743966e+00
3.12868595e-01 -1.74894005e-01 5.33460617e-01 5.08377671e-01
-6.68238878e-01 2.51246765e-02 -4.46210094e-02 -7.96527117e-02
1.00174713e+00 -4.72063664e-03 1.11639559e-01 -3.87988657e-01
4.74152982e-01 1.12277679e-02 4.83747005e-01 6.75483465e-01
-4.02374566e-01 9.64509070e-01 3.23067397e-01 -7.45722711e-01
-1.12847590e+00 -8.45126808e-01 -4.59713638e-01 -1.39014702e-02
3.36120427e-01 4.19125259e-01 -6.94110394e-01 -5.65777004e-01
-4.75352883e-01 4.38567579e-01 -4.92140561e-01 -4.27994072e-01
-1.13881314e+00 -9.33875203e-01 4.66420650e-01 4.60970044e-01
4.21249688e-01 -1.06171143e+00 -1.38188148e+00 3.45353782e-01
-6.45855069e-01 -1.03378201e+00 -8.54338169e-01 -3.53495330e-02
-1.20193315e+00 -1.26077211e+00 -1.25513935e+00 -6.62432194e-01
8.27652991e-01 6.62967205e-01 6.86985850e-01 -8.64493698e-02
-6.10244751e-01 -1.44016184e-03 -3.61914158e-01 -7.24945962e-03
-4.95301664e-01 -1.62549913e-01 -1.58380091e-01 -1.56900525e-01
-1.07501119e-01 -2.68723905e-01 -1.24536479e+00 3.72487962e-01
-1.22484756e+00 2.29887813e-01 7.44319737e-01 8.68476152e-01
1.59749448e-01 8.22342485e-02 1.44327641e-01 -6.91065073e-01
5.73625565e-01 -3.55974555e-01 -1.44307733e-01 3.73321660e-02
-3.19875836e-01 1.68575764e-01 4.71768677e-01 -8.74235332e-01
-9.59730864e-01 -1.50345638e-01 -4.54579704e-02 -6.39047682e-01
-1.33737445e-01 2.38521710e-01 3.35360587e-01 -1.49644434e-01
7.95240402e-01 4.72561181e-01 1.27822831e-01 -2.76634216e-01
-7.70622268e-02 3.66722792e-01 5.47161520e-01 1.88575871e-02
3.56433809e-01 6.92750692e-01 7.72299021e-02 -1.03432631e+00
-1.92900687e-01 -4.13635164e-01 -7.32783437e-01 -6.47154927e-01
1.07481086e+00 -3.95498753e-01 -4.05089736e-01 6.65258646e-01
-1.20507336e+00 -8.11132416e-02 4.49149869e-04 1.11542773e+00
-2.37509161e-01 7.32697546e-01 -7.55902171e-01 -6.13464236e-01
-2.48298496e-01 -1.81518447e+00 5.47229946e-01 1.04253918e-01
-4.06972349e-01 -8.98635030e-01 -1.66654885e-01 2.37283513e-01
6.62451923e-01 2.13009536e-01 1.01125872e+00 -1.83067143e-01
-4.60314214e-01 -7.71958902e-02 5.30903116e-02 1.23940416e-01
3.13108116e-01 -4.08680826e-01 -7.68216729e-01 -3.71536672e-01
6.72492266e-01 2.40624771e-01 3.98944259e-01 1.12375832e+00
1.42290342e+00 2.45404560e-02 -2.07311034e-01 7.74731457e-01
1.17223692e+00 8.67411256e-01 9.77600694e-01 6.03929639e-01
6.18278980e-01 4.64286208e-01 5.06397009e-01 7.16355368e-02
-1.42993689e-01 6.75199687e-01 2.27750018e-01 -2.46516198e-01
-3.34232807e-01 1.75666317e-01 -6.35879040e-02 1.17245889e+00
-5.63979484e-02 3.47430259e-01 -1.14975858e+00 6.29311085e-01
-1.40154219e+00 -7.59301484e-01 -7.50927746e-01 2.28762436e+00
8.34221482e-01 -3.02281287e-02 -1.64062709e-01 3.72172534e-01
6.04600787e-01 2.33612433e-02 -5.15624404e-01 -1.05908200e-01
1.44093633e-01 -4.82524000e-02 7.58374214e-01 3.57018173e-01
-9.24376488e-01 1.27213284e-01 6.31630182e+00 7.16794312e-01
-1.68352413e+00 4.40094680e-01 6.20108604e-01 -2.09610194e-01
-2.39426702e-01 -2.28087619e-01 -5.19551821e-02 9.08018112e-01
6.81781769e-01 3.57125759e-01 4.74426188e-02 4.92864817e-01
7.38653541e-01 -4.38644022e-01 -7.80229926e-01 1.19758034e+00
-1.84859693e-01 -1.30220580e+00 -3.07288080e-01 -2.15435848e-01
5.64606905e-01 -3.07396531e-01 4.47053872e-02 -1.91839635e-01
-5.11691093e-01 -9.97295082e-01 6.00922167e-01 6.09910011e-01
8.13376546e-01 -6.46324515e-01 7.98712492e-01 3.81044507e-01
-7.16356993e-01 2.84380466e-01 -2.94286489e-01 1.83947384e-01
1.50753915e-01 6.83751702e-01 -7.90158510e-01 6.61167502e-01
6.54062688e-01 1.58069253e-01 -3.57691199e-01 1.29234457e+00
6.73890263e-02 4.78035539e-01 1.20294645e-01 1.88304231e-01
3.30452561e-01 -7.98851550e-02 6.06131256e-01 1.20573032e+00
4.17340070e-01 2.11589426e-01 -2.58756071e-01 6.04062259e-01
5.35058260e-01 -1.05600603e-01 -4.49450225e-01 4.39736903e-01
1.75631851e-01 1.12204266e+00 -1.14056551e+00 -2.53819883e-01
-2.07786709e-01 8.70851159e-01 -2.53564894e-01 2.89797097e-01
-8.50091100e-01 -2.81100065e-01 1.06262147e-01 5.58639109e-01
7.31226802e-03 -2.94352978e-01 -2.03257203e-01 -1.06395555e+00
1.75578713e-01 -7.62926579e-01 -1.99285392e-02 -8.51084411e-01
-7.32125998e-01 4.81973708e-01 -9.08730626e-02 -1.56586504e+00
-1.89261079e-01 -5.44847488e-01 -6.20906055e-01 8.49589288e-01
-1.37253797e+00 -3.68737996e-01 -4.18433517e-01 3.67040217e-01
5.12522221e-01 3.49554211e-01 4.84313339e-01 5.67313910e-01
-3.26011807e-01 3.16542625e-01 2.04620481e-01 1.48612618e-01
6.82813942e-01 -9.67642426e-01 1.22565746e-01 8.87400270e-01
-3.80043596e-01 7.12549448e-01 6.76927507e-01 -6.06904984e-01
-1.49625731e+00 -8.46153617e-01 4.31497872e-01 -1.79890126e-01
3.64309579e-01 2.81005859e-01 -1.24225390e+00 2.17435569e-01
-1.08844889e-02 2.60833889e-01 3.63983691e-01 -6.87406778e-01
3.04348707e-01 1.78947702e-01 -1.19318926e+00 5.90522885e-01
7.19919860e-01 -3.69526297e-01 -5.54458380e-01 2.92971313e-01
4.67797637e-01 -6.85395241e-01 -9.16398644e-01 4.57781285e-01
7.35248148e-01 -9.84521627e-01 9.84527111e-01 -3.78144801e-01
7.00502038e-01 -4.36209977e-01 6.91701412e-01 -1.44339752e+00
-5.96999288e-01 -4.20970082e-01 1.94672763e-01 4.43369448e-01
5.43622896e-02 -5.58341146e-01 8.61216486e-01 3.06148499e-01
-1.48883849e-01 -7.77053356e-01 -1.03146100e+00 -7.46386170e-01
1.94083810e-01 -4.85776454e-01 2.12947488e-01 1.50911009e+00
-6.75670654e-02 -2.34978020e-01 -3.99324507e-01 2.28619605e-01
5.59359074e-01 -1.92459434e-01 2.14491487e-01 -5.61017275e-01
-2.78945953e-01 -5.60214520e-01 -3.73472393e-01 -9.68505204e-01
-6.47086740e-01 -8.36385131e-01 -1.71228305e-01 -1.40166032e+00
7.41787180e-02 -3.62619638e-01 -1.62069902e-01 -8.04935545e-02
-3.80523413e-01 1.42837465e-01 4.54751253e-02 6.12220764e-01
-1.11584760e-01 1.38176292e-01 2.14801383e+00 -1.32050976e-01
-1.22545607e-01 -2.52804719e-04 2.92625632e-02 7.81460285e-01
7.21870303e-01 -5.48915148e-01 -5.39226770e-01 -6.20725155e-01
-6.55440800e-03 6.14803672e-01 2.00757354e-01 -1.17799008e+00
4.12914515e-01 6.73839301e-02 7.25682676e-01 -8.29894662e-01
-9.84974951e-02 -7.14280188e-01 4.53939855e-01 1.07518649e+00
-2.71634817e-01 4.04286504e-01 2.47931421e-01 4.35281172e-02
-2.20752597e-01 -6.51023328e-01 1.15637171e+00 -4.41088468e-01
-3.88102591e-01 7.29716718e-02 -8.55503798e-01 -7.81722218e-02
8.45806539e-01 -5.68063021e-01 1.31056443e-01 -1.95556372e-01
-9.06722128e-01 -1.82552829e-01 1.08725414e-01 2.67263532e-01
7.55854011e-01 -1.32268381e+00 -3.70674014e-01 3.93499404e-01
-1.56866521e-01 1.03703916e-01 9.90800142e-01 1.70006561e+00
-1.17691553e+00 5.08913040e-01 -1.42917007e-01 -9.88246202e-01
-9.72560108e-01 5.03860354e-01 7.01901197e-01 -4.70877886e-01
-8.90320122e-01 6.02220953e-01 1.26853690e-01 -1.11022353e-01
1.29878167e-02 -7.84781456e-01 -2.53383458e-01 -1.82915509e-01
8.49938512e-01 3.96448672e-01 4.54308569e-01 -2.30766237e-01
-4.30984557e-01 7.08698511e-01 -3.19232821e-01 2.50683036e-02
8.56453419e-01 -1.53241456e-01 1.73923284e-01 4.22666788e-01
9.65796649e-01 -1.34170517e-01 -1.12139606e+00 -4.61884588e-03
1.27438858e-01 -6.67210340e-01 9.68192592e-02 -6.80074930e-01
-1.13262939e+00 1.08893108e+00 1.02557099e+00 2.20307529e-01
1.02123106e+00 -4.68045205e-01 6.95635438e-01 -2.47478202e-01
2.71868765e-01 -8.63905191e-01 1.67899191e-01 1.69104993e-01
7.12536395e-01 -9.98170495e-01 -7.63571709e-02 -2.11796001e-01
-7.32572794e-01 1.23236442e+00 3.49344581e-01 -6.25443785e-03
4.05776083e-01 2.97608435e-01 2.94133365e-01 -9.94327366e-02
-2.39444911e-01 4.27131534e-01 2.46826708e-01 5.56408346e-01
7.50604153e-01 -9.41030085e-02 -6.25672281e-01 3.05656135e-01
4.90831360e-02 1.88227594e-01 7.61861384e-01 9.16985333e-01
-1.78492695e-01 -6.89591169e-01 -8.14282179e-01 3.08084697e-01
-7.88086772e-01 1.80791050e-01 2.87791789e-01 7.19711423e-01
1.93247616e-01 6.25938296e-01 -9.45329219e-02 1.46388412e-01
4.17181373e-01 -3.36235538e-02 6.87482715e-01 -8.50403309e-02
-4.67846483e-01 1.49647042e-01 -2.41565406e-01 -5.35565734e-01
-2.52127975e-01 -4.55861568e-01 -1.35533166e+00 -1.58711791e-01
-2.44671404e-01 1.95342809e-01 9.54885125e-01 7.99711108e-01
-1.25908693e-02 1.09708381e+00 5.25050640e-01 -8.13469648e-01
-5.77716708e-01 -1.02432513e+00 -5.44826329e-01 6.44225240e-01
4.70630318e-01 -5.23207784e-01 -2.53642917e-01 -1.10325068e-01] | [13.740094184875488, -2.440864086151123] |
5628ef86-7514-4a32-bbc8-df523ab21b2c | deep-point-cloud-reconstruction-1 | 2111.11704 | null | https://arxiv.org/abs/2111.11704v2 | https://arxiv.org/pdf/2111.11704v2.pdf | Deep Point Cloud Reconstruction | Point cloud obtained from 3D scanning is often sparse, noisy, and irregular. To cope with these issues, recent studies have been separately conducted to densify, denoise, and complete inaccurate point cloud. In this paper, we advocate that jointly solving these tasks leads to significant improvement for point cloud reconstruction. To this end, we propose a deep point cloud reconstruction network consisting of two stages: 1) a 3D sparse stacked-hourglass network as for the initial densification and denoising, 2) a refinement via transformers converting the discrete voxels into 3D points. In particular, we further improve the performance of transformer by a newly proposed module called amplified positional encoding. This module has been designed to differently amplify the magnitude of positional encoding vectors based on the points' distances for adaptive refinements. Extensive experiments demonstrate that our network achieves state-of-the-art performance among the recent studies in the ScanNet, ICL-NUIM, and ShapeNetPart datasets. Moreover, we underline the ability of our network to generalize toward real-world and unmet scenes. | ['In So Kweon', 'Jaesik Park', 'Francois Rameau', 'Byeongin Joung', 'Jaesung Choe'] | 2021-11-23 | deep-point-cloud-reconstruction | https://openreview.net/forum?id=mKDtUtxIGJ | https://openreview.net/pdf?id=mKDtUtxIGJ | iclr-2022-4 | ['point-cloud-completion', 'point-cloud-reconstruction', 'point-set-upsampling'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 4.76952381e-02 1.09617643e-01 3.30990583e-01 -3.46570730e-01
-5.98243237e-01 -1.21555507e-01 5.78713417e-01 4.67775352e-02
-2.40653098e-01 4.63158965e-01 7.96832591e-02 -1.80714950e-02
-1.61524445e-01 -1.01580667e+00 -9.13454115e-01 -4.71029848e-01
3.20556723e-02 8.58889341e-01 3.02817523e-01 -2.89252609e-01
2.15855688e-01 1.03992414e+00 -1.61290562e+00 -1.03133935e-02
1.05673385e+00 1.17461276e+00 4.50569876e-02 8.78821984e-02
-3.26374233e-01 3.51009727e-01 -9.87515450e-02 -2.08150610e-01
5.64556956e-01 5.95438778e-02 -4.54111218e-01 3.36389840e-01
4.31692123e-01 -3.64437550e-01 -1.23103887e-01 1.12350273e+00
4.34075356e-01 1.27738073e-01 5.31521559e-01 -9.60451603e-01
-4.29195642e-01 4.68534261e-01 -8.53330970e-01 -2.07055584e-01
-3.10358908e-02 2.50573546e-01 6.14739180e-01 -1.33782709e+00
4.59441483e-01 1.28710544e+00 1.04086852e+00 9.60918888e-02
-1.15571141e+00 -7.98561215e-01 1.89069659e-01 4.65756282e-02
-1.60011148e+00 -2.36804441e-01 1.00699067e+00 -3.86358827e-01
7.43651152e-01 1.34394124e-01 8.83533716e-01 8.82616460e-01
-1.48859844e-01 3.22889328e-01 7.98686445e-01 -3.71492021e-02
3.50101173e-01 -3.16109926e-01 -1.96905836e-01 4.36739445e-01
3.35285157e-01 9.27735046e-02 -2.95956880e-01 -2.02727914e-01
1.26382208e+00 1.76564991e-01 -1.90384895e-01 -5.84350884e-01
-9.97685015e-01 7.11304426e-01 7.14374661e-01 3.46846581e-02
-7.59664953e-01 2.56351739e-01 6.51434064e-02 -5.88720106e-02
9.80671704e-01 4.05821919e-01 -2.64020920e-01 1.73506215e-02
-1.21748459e+00 3.30069274e-01 4.67033654e-01 1.06079173e+00
1.01982963e+00 1.25191331e-01 5.70191778e-02 8.45868111e-01
3.80117446e-01 3.44512731e-01 -1.47091776e-01 -1.08150923e+00
5.34866273e-01 7.71426797e-01 3.97647396e-02 -1.20077252e+00
-4.83135402e-01 -8.43930483e-01 -1.32295358e+00 2.47718737e-01
-2.76169628e-01 1.20916076e-01 -1.09283435e+00 1.20729542e+00
5.73945940e-01 8.65627706e-01 -3.17627758e-01 1.06948864e+00
7.82358170e-01 5.65068662e-01 -1.30666465e-01 9.32051465e-02
1.02853799e+00 -8.05945456e-01 -4.93586242e-01 2.61873361e-02
1.67047620e-01 -6.29214823e-01 8.32967997e-01 3.57213736e-01
-1.34880686e+00 -4.48953032e-01 -9.52330410e-01 -2.58160979e-01
9.29925665e-02 -1.82821259e-01 4.76091236e-01 1.62408113e-01
-9.45745111e-01 6.84727728e-01 -1.09656084e+00 5.97505737e-03
6.83763504e-01 1.96743384e-01 -1.26164868e-01 -2.56115526e-01
-7.15811372e-01 4.64549273e-01 1.22578025e-01 4.53417301e-01
-7.15555072e-01 -1.13990676e+00 -7.31926262e-01 2.14802891e-01
2.81506807e-01 -1.12899959e+00 8.85170937e-01 -3.58651519e-01
-1.25607288e+00 6.21786773e-01 -1.57733858e-02 -4.95072454e-01
6.46451473e-01 -1.74994186e-01 -3.87133285e-02 3.79197620e-04
1.22478552e-01 6.97740734e-01 7.30671883e-01 -1.68010521e+00
-5.57385325e-01 -6.24926865e-01 -2.31997654e-01 2.79236466e-01
-1.29520427e-02 -3.96295071e-01 -6.49290740e-01 -7.94881225e-01
7.20669389e-01 -6.66622877e-01 -6.23788178e-01 2.08101898e-01
-3.78622651e-01 -9.71686281e-03 7.89159954e-01 -4.76266384e-01
9.51435685e-01 -2.32156038e+00 3.15517962e-01 6.23423994e-01
4.66205776e-01 -4.41843681e-02 -2.05296129e-01 1.65438041e-01
8.23857859e-02 2.45850787e-01 -7.06805587e-01 -8.42450976e-01
-8.27280506e-02 4.99345392e-01 -1.85673729e-01 5.68272948e-01
3.26985151e-01 6.23029113e-01 -7.09016502e-01 -2.07538277e-01
3.06881309e-01 9.46167231e-01 -8.95470798e-01 2.42042108e-04
-2.84479141e-01 6.74236596e-01 -4.73784268e-01 8.41174066e-01
1.28931057e+00 -3.28409076e-01 -2.75016725e-01 -3.67707044e-01
-4.00355339e-01 1.54544096e-02 -1.41062593e+00 1.92693985e+00
-1.55805528e-01 7.08090141e-02 4.91607666e-01 -6.35220289e-01
1.02950752e+00 7.67248422e-02 6.83622777e-01 -5.84968090e-01
2.69674081e-02 2.48608351e-01 -3.19840282e-01 -8.09257329e-02
7.78653502e-01 -1.77413762e-01 2.72557497e-01 -1.00512862e-01
-3.35206896e-01 -6.33805811e-01 -2.00390384e-01 3.70233390e-03
1.04085457e+00 -8.41329899e-03 -2.29214698e-01 -8.99823010e-02
3.60717982e-01 1.65781230e-02 5.55325508e-01 5.91761589e-01
2.98316330e-01 1.11314952e+00 3.13918442e-01 -4.49108303e-01
-1.17470860e+00 -9.91520166e-01 -2.28277668e-01 4.05997425e-01
3.46459776e-01 -2.72127867e-01 -4.69743848e-01 -2.33219221e-01
2.10758626e-01 5.81300139e-01 -4.04042751e-01 7.03213364e-02
-8.15034389e-01 -5.56182742e-01 3.91259193e-01 5.70498049e-01
6.03163183e-01 -7.08638549e-01 -1.98934004e-01 1.90109029e-01
-6.69391230e-02 -1.15072370e+00 -1.75816238e-01 1.59259468e-01
-1.25101030e+00 -8.73996854e-01 -5.06880939e-01 -5.11880279e-01
7.03638732e-01 3.24044675e-01 1.14503706e+00 2.60440946e-01
2.99390942e-01 5.67592308e-02 -4.44019735e-01 -3.10756147e-01
2.76729502e-02 3.30968678e-01 -9.42106023e-02 1.03606526e-02
-1.17443137e-01 -1.16164994e+00 -6.96486235e-01 3.57505560e-01
-1.21513855e+00 2.41474479e-01 5.97392559e-01 5.67032516e-01
1.19452810e+00 7.41122216e-02 2.21276488e-02 -8.84900093e-01
4.26265538e-01 -7.35518932e-01 -5.81495643e-01 -2.07893297e-01
-5.22170007e-01 6.32045269e-02 3.66446137e-01 -4.46994454e-02
-8.96614671e-01 1.90701842e-01 -6.84286416e-01 -9.30229247e-01
-1.40942931e-01 5.81782758e-01 4.37500933e-03 -3.97591382e-01
3.88718694e-01 1.36324003e-01 -7.92076588e-02 -8.64220679e-01
3.74891490e-01 2.48774096e-01 5.89984357e-01 -5.98327935e-01
1.15812171e+00 8.76590848e-01 2.93252736e-01 -7.83966303e-01
-6.59407318e-01 -4.52053189e-01 -6.09171510e-01 -9.52972472e-02
7.41614759e-01 -1.08250570e+00 -3.55028957e-01 5.11129677e-01
-1.27726138e+00 -2.23410547e-01 -5.03020406e-01 2.35316440e-01
-3.51922095e-01 4.43950981e-01 -6.70194447e-01 -4.34739351e-01
-4.17282104e-01 -1.22120917e+00 1.36116278e+00 -1.39810815e-02
1.87386811e-01 -5.95711350e-01 8.46223980e-02 1.83073997e-01
4.14156586e-01 4.72028166e-01 7.73615420e-01 -9.50843319e-02
-1.01427746e+00 1.21174440e-01 -3.95975590e-01 3.07632029e-01
-1.23894341e-01 -2.44611800e-02 -6.96227551e-01 -1.79593816e-01
2.23711550e-01 3.71853262e-02 8.09462428e-01 4.25899863e-01
1.51887453e+00 -1.84746876e-01 -1.17284626e-01 1.39886487e+00
1.45900989e+00 -2.26178676e-01 8.54625642e-01 3.66167635e-01
9.36114252e-01 1.64594024e-01 4.16943401e-01 7.87822902e-01
4.14921552e-01 5.91853559e-01 1.13238192e+00 -2.78450847e-01
-1.49616301e-01 -1.53118446e-01 -3.51081401e-01 1.04885864e+00
-4.22284096e-01 -4.67782542e-02 -1.04640388e+00 4.39696759e-01
-1.77199376e+00 -6.26726151e-01 -4.60017264e-01 1.76219749e+00
5.26332021e-01 4.94497195e-02 -2.58202046e-01 -3.48303169e-02
4.93075371e-01 3.11743379e-01 -5.62558949e-01 1.11948743e-01
-1.66031301e-01 4.75792736e-01 6.06653750e-01 3.86225313e-01
-7.68564939e-01 8.30876172e-01 5.54561615e+00 8.52822244e-01
-1.14482415e+00 1.51659563e-01 5.26149750e-01 -1.46022812e-01
-6.49985015e-01 -1.53130874e-01 -5.80537438e-01 4.73610103e-01
4.49990362e-01 1.95123896e-01 3.53580892e-01 7.75652885e-01
2.69379675e-01 2.54531562e-01 -7.74270892e-01 1.08384931e+00
-1.29434913e-01 -1.60307741e+00 3.53109688e-01 1.28268287e-01
8.60929132e-01 4.52605873e-01 4.77738343e-02 1.30314454e-01
7.65496567e-02 -9.42336917e-01 9.51826155e-01 6.79351389e-01
7.84670353e-01 -9.22243178e-01 6.43288374e-01 5.25447071e-01
-1.25049925e+00 2.46407717e-01 -5.42290688e-01 -2.05319189e-02
4.11999881e-01 1.03243124e+00 -5.68045139e-01 8.90586674e-01
7.74770439e-01 9.21289921e-01 -4.59167302e-01 1.29982209e+00
-1.40086874e-01 4.24930394e-01 -6.10152662e-01 5.54190874e-01
4.22432363e-01 -5.00085235e-01 7.12778568e-01 7.97293961e-01
6.27735555e-01 4.17882830e-01 1.12575240e-01 1.07540596e+00
-1.64589062e-01 -1.31632298e-01 -4.06819105e-01 4.61928964e-01
5.44204950e-01 1.09117329e+00 -6.35700405e-01 -1.72613159e-01
-2.18792617e-01 6.98491812e-01 3.08102399e-01 3.60664189e-01
-7.57037878e-01 5.01966141e-02 9.12827075e-01 5.57852447e-01
5.15553892e-01 -5.74808657e-01 -7.56894529e-01 -1.04880774e+00
6.24803193e-02 -5.83699822e-01 -5.55417463e-02 -8.42247307e-01
-1.33528066e+00 7.23358870e-01 6.31538779e-02 -1.28159034e+00
1.69642627e-01 -2.32325077e-01 -4.34000403e-01 8.50250244e-01
-1.73258758e+00 -9.58930135e-01 -7.28781939e-01 5.27963579e-01
5.04936039e-01 2.40287915e-01 2.47296035e-01 6.25783026e-01
-4.05977875e-01 9.77176577e-02 3.79667850e-04 -1.81185573e-01
3.04718167e-01 -8.54250014e-01 7.12893367e-01 7.74663210e-01
-2.10943505e-01 3.94738615e-01 5.45540333e-01 -8.57585907e-01
-1.34443641e+00 -1.39082754e+00 3.43441367e-01 -1.95925564e-01
3.59370619e-01 -2.46332511e-01 -1.24566793e+00 6.56019509e-01
-1.64097384e-01 1.96822107e-01 3.96512784e-02 -6.89789876e-02
-1.04597226e-01 -1.09732792e-01 -1.26308298e+00 3.25860083e-01
1.38606405e+00 -6.29633740e-02 -3.56421769e-01 2.06490800e-01
7.83459365e-01 -9.14728642e-01 -8.81023645e-01 7.51558363e-01
1.52091712e-01 -1.09322095e+00 1.23925185e+00 -1.46178931e-01
7.15474486e-01 -4.59811628e-01 -1.40721306e-01 -1.35562551e+00
-6.08110428e-01 -2.36402109e-01 -9.32347849e-02 9.01235282e-01
8.50231498e-02 -4.48684633e-01 9.89323497e-01 3.57052863e-01
-7.36021817e-01 -9.39913988e-01 -1.20732498e+00 -4.53167975e-01
-2.98733972e-02 -7.39202619e-01 1.11659038e+00 9.74046350e-01
-8.53541911e-01 1.06674530e-01 -2.41355404e-01 4.77110416e-01
8.47496569e-01 -1.22136287e-01 8.33896279e-01 -1.49634039e+00
8.72418955e-02 -3.95699561e-01 -3.41911852e-01 -1.41734540e+00
-1.33040607e-01 -7.50881910e-01 2.60409210e-02 -1.58685672e+00
-8.34687576e-02 -9.21737373e-01 1.02138266e-01 3.35839272e-01
4.57687601e-02 3.18420529e-01 1.17151521e-01 4.96914208e-01
-3.90636176e-01 1.03148675e+00 1.45106447e+00 -3.51328254e-02
-1.91129252e-01 -6.31648377e-02 -5.61791420e-01 7.13531613e-01
5.40320098e-01 -4.34812933e-01 -1.90678582e-01 -1.08381355e+00
4.25224632e-01 -2.75082607e-02 5.03478587e-01 -1.17482066e+00
3.49872500e-01 -8.18599239e-02 3.14541787e-01 -1.17679083e+00
6.02531314e-01 -1.17881906e+00 5.70610523e-01 1.05463244e-01
1.69218868e-01 1.91080436e-01 1.76418111e-01 6.01736188e-01
-2.54243970e-01 2.03003623e-02 7.85284936e-01 -2.06833303e-01
-5.06285071e-01 7.48761415e-01 1.46692023e-01 -2.64188170e-01
7.50584245e-01 -4.08832580e-01 -7.74634853e-02 -1.27974644e-01
-5.92923701e-01 4.40882891e-01 7.26891994e-01 3.88105288e-02
9.11687374e-01 -1.40108252e+00 -7.80513346e-01 5.41932821e-01
-1.79623544e-01 9.27902937e-01 4.99194086e-01 8.61745298e-01
-9.33066249e-01 3.03408653e-02 -1.08877540e-01 -9.85770583e-01
-7.93506563e-01 2.04424441e-01 2.44407430e-01 -1.96631491e-01
-9.49153185e-01 8.41164589e-01 5.03682010e-02 -5.84127069e-01
2.73994505e-01 -8.70392919e-01 3.86376157e-02 -9.69474688e-02
3.04917917e-02 5.03672779e-01 4.17214662e-01 -6.20039225e-01
-2.13397816e-01 7.54347265e-01 1.65153116e-01 1.55505568e-01
1.84124970e+00 -6.10969998e-02 -3.57511848e-01 7.65966475e-02
9.03066278e-01 -3.64499129e-02 -1.35692203e+00 -2.07679614e-01
-3.46057951e-01 -5.99353850e-01 2.74351001e-01 -2.50223160e-01
-1.40647304e+00 6.20415688e-01 4.44504976e-01 1.49487153e-01
1.11406791e+00 3.17031033e-02 9.64346051e-01 8.64671096e-02
4.61681217e-01 -6.50148869e-01 -2.69083411e-01 7.01563776e-01
1.04689157e+00 -9.90824223e-01 1.82718486e-01 -7.24210680e-01
-1.19106248e-01 8.53363752e-01 5.63314557e-01 -3.84014010e-01
8.29571843e-01 3.00825894e-01 -3.02639574e-01 -6.07158661e-01
-4.19625700e-01 -1.66169759e-02 2.36917630e-01 3.25096697e-01
-2.70800181e-02 -1.99541207e-02 -1.19619429e-01 4.23900425e-01
-2.72738665e-01 8.56930241e-02 2.92277485e-01 8.12330961e-01
-3.92450541e-01 -8.03715229e-01 -6.30705774e-01 4.64249045e-01
4.52470072e-02 -8.41865167e-02 -1.15727171e-01 7.26090908e-01
2.15013295e-01 3.68664086e-01 4.27828074e-01 -3.32042634e-01
8.14142227e-01 -4.30039257e-01 2.36509085e-01 -6.98703349e-01
-5.59922874e-01 1.21457711e-01 -2.93116182e-01 -7.21409798e-01
-4.00448292e-01 -5.13475895e-01 -1.21173644e+00 -4.59778816e-01
-4.05616909e-02 9.56090093e-02 6.96983874e-01 6.30036831e-01
5.97894490e-01 6.33290768e-01 4.78902221e-01 -1.45628405e+00
-4.98598665e-01 -9.08711076e-01 -5.47751129e-01 3.01738799e-01
1.86310023e-01 -7.55298078e-01 -4.43682551e-01 -3.80182922e-01] | [8.331769943237305, -3.544970989227295] |
c06ef930-aaf4-49aa-9beb-9d98686be94f | interpretable-machine-learning-for-detection | 2210.11235 | null | https://arxiv.org/abs/2210.11235v3 | https://arxiv.org/pdf/2210.11235v3.pdf | Application of Explainable Machine Learning in Detecting and Classifying Ransomware Families Based on API Call Analysis | Ransomware has appeared as one of the major global threats in recent days. The alarming increasing rate of ransomware attacks and new ransomware variants intrigue the researchers to constantly examine the distinguishing traits of ransomware and refine their detection strategies. Application Programming Interface (API) is a way for one program to collaborate with another; API calls are the medium by which they communicate. Ransomware uses this strategy to interact with the OS and makes a significantly higher number of calls in different sequences to ask for taking action. This research work utilizes the frequencies of different API calls to detect and classify ransomware families. First, a Web-Crawler is developed to automate collecting the Windows Portable Executable (PE) files of 15 different ransomware families. By extracting different frequencies of 68 API calls, we develop our dataset in the first phase of the two-phase feature engineering process. After selecting the most significant features in the second phase of the feature engineering process, we deploy six Supervised Machine Learning models: Na"ive Bayes, Logistic Regression, Random Forest, Stochastic Gradient Descent, K-Nearest Neighbor, and Support Vector Machine. Then, the performances of all the classifiers are compared to select the best model. The results reveal that Logistic Regression can efficiently classify ransomware into their corresponding families securing 99.15% overall accuracy. Finally, instead of relying on the 'Black box' characteristic of the Machine Learning models, we present the post-hoc analysis of our best-performing model using 'SHapley Additive exPlanations' or SHAP values to ascertain the transparency and trustworthiness of the model's prediction. | ['Kaushik Roy', 'Madhuri Siddula', 'Rawshan Ara Mowri'] | 2022-10-16 | null | null | null | null | ['interpretable-machine-learning'] | ['methodology'] | [-6.72249421e-02 -6.40550733e-01 -5.44828951e-01 -1.75544575e-01
-8.77186134e-02 -1.04675782e+00 5.15563011e-01 2.91908951e-03
-1.76987812e-01 2.75153011e-01 3.46289352e-02 -1.02213931e+00
-3.92932147e-01 -6.21528566e-01 -6.28687665e-02 -4.31412756e-01
-2.07536831e-03 2.97113568e-01 3.51471275e-01 -1.69860363e-01
9.30621266e-01 5.78587711e-01 -1.24699211e+00 3.84976983e-01
6.57855690e-01 1.03240299e+00 -3.76435250e-01 7.93409050e-01
1.48598716e-01 7.82526135e-01 -6.33986354e-01 -5.39756238e-01
4.62818146e-01 -4.74546909e-01 -7.47016370e-01 -3.73865873e-01
-2.60694206e-01 -4.41614985e-01 1.00215105e-03 8.81754458e-01
7.52602816e-02 -2.64654070e-01 4.45314467e-01 -1.51814699e+00
-3.19851160e-01 7.38154352e-01 -4.96519029e-01 4.64537501e-01
6.89546227e-01 1.37458667e-01 7.79848695e-01 -5.27228117e-01
5.76303780e-01 8.74918878e-01 8.72095168e-01 1.24510042e-01
-8.64449859e-01 -1.08693802e+00 -4.69299138e-01 3.11987370e-01
-1.25889754e+00 1.16443587e-02 7.06153214e-01 -8.56905878e-01
8.72435153e-01 6.57014191e-01 4.69629705e-01 1.08673286e+00
6.44052386e-01 3.06929857e-03 1.18993819e+00 -2.74778992e-01
1.47540286e-01 5.32623410e-01 6.58459663e-01 5.31045556e-01
7.55652905e-01 2.46773183e-01 -2.87083477e-01 -1.01283073e+00
1.98228046e-01 3.23990554e-01 -1.62540138e-01 7.10490122e-02
-7.85167575e-01 9.44630206e-01 8.16175416e-02 6.60278976e-01
-1.94552153e-01 -3.39137852e-01 4.69269127e-01 3.89154345e-01
-5.18220477e-02 7.15462267e-01 -6.78152740e-01 -4.46045607e-01
-7.60123253e-01 -1.83576196e-01 1.06961131e+00 4.08010989e-01
5.22583544e-01 -2.46129334e-01 4.20774281e-01 2.49964342e-01
5.54485142e-01 1.56590462e-01 7.37441123e-01 -3.43539178e-01
2.03036353e-01 9.07934427e-01 -3.64044160e-02 -1.46644568e+00
-1.46022305e-01 -1.58940911e-01 -4.03944910e-01 1.69517065e-04
3.29431206e-01 -2.50736624e-01 -3.23152304e-01 8.19806099e-01
3.38833958e-01 4.01120782e-02 -4.12789941e-01 5.64856529e-01
3.21313173e-01 5.18205523e-01 -2.43543074e-01 -1.65537462e-01
1.46566451e+00 -3.48175287e-01 -2.53047258e-01 4.47626054e-01
7.67905354e-01 -1.06640279e+00 7.51159668e-01 6.25784278e-01
-6.91313297e-02 -3.07013929e-01 -1.17375314e+00 7.25346148e-01
-4.64665115e-01 5.19057438e-02 6.20600939e-01 1.36465931e+00
-3.81022155e-01 6.89380288e-01 -7.75251091e-01 -2.71959245e-01
1.60245538e-01 1.85705274e-01 -3.91132355e-01 2.19072208e-01
-7.50737488e-01 9.20760453e-01 4.38207209e-01 -9.24514532e-02
-7.80047715e-01 -4.59895521e-01 -4.15241838e-01 1.22830113e-02
4.19102609e-01 -1.97841823e-01 5.69821417e-01 -7.51777768e-01
-1.23584712e+00 5.34833908e-01 1.21496134e-01 -1.30213499e-01
4.06174034e-01 -1.65975749e-01 -6.07803643e-01 -2.64008623e-02
-2.07009196e-01 -6.04723096e-01 1.05688834e+00 -1.09670079e+00
-4.14402574e-01 -3.43985289e-01 -3.24711837e-02 -5.82846999e-01
-3.16572338e-01 7.47294426e-01 3.49570602e-01 -2.97956735e-01
2.04093382e-01 -9.75122392e-01 1.50502518e-01 -9.71444666e-01
-6.57185733e-01 -3.88600901e-02 1.28421545e+00 -6.83272839e-01
1.66962397e+00 -2.27364707e+00 -3.67789298e-01 6.50527477e-01
4.24054116e-01 2.91779667e-01 2.23947659e-01 7.20487297e-01
-2.27926865e-01 5.61215103e-01 4.50484753e-02 3.43606383e-01
-8.15035254e-02 -1.57314152e-01 -6.23156130e-01 5.30845463e-01
-3.04535985e-01 3.05461913e-01 -6.22705996e-01 -2.11929649e-01
1.09012067e-01 2.87744612e-01 -6.16162181e-01 2.94663578e-01
4.31291491e-01 3.72567713e-01 -4.97530043e-01 6.51598155e-01
4.41620350e-01 -2.54286319e-01 4.90728974e-01 -1.19578928e-01
-1.38292819e-01 4.23268110e-01 -1.02844858e+00 4.06104892e-01
-2.37309575e-01 5.07682800e-01 -3.26971501e-01 -5.95701098e-01
1.09695232e+00 1.29786596e-01 3.20062101e-01 1.68328688e-01
5.69316626e-01 4.49602187e-01 3.55690062e-01 -6.24801517e-01
3.20517391e-01 1.53392881e-01 -4.76846769e-02 9.15738165e-01
-5.99201731e-02 2.47509554e-01 -1.36272117e-01 -3.21473437e-03
1.37915945e+00 -1.01645350e-01 7.38366187e-01 -1.62777409e-01
6.54813230e-01 2.01714680e-01 3.80420089e-01 6.88355327e-01
-2.02453375e-01 1.08426005e-01 5.88163078e-01 -7.66675055e-01
-4.92976576e-01 -8.43496561e-01 -5.75122386e-02 1.17031920e+00
-1.32609578e-02 -6.46404207e-01 -6.62916064e-01 -1.16021633e+00
-1.83721319e-01 8.67749870e-01 -5.17978728e-01 -3.48808676e-01
-4.42460209e-01 -9.47671652e-01 8.80973279e-01 1.94368109e-01
4.53284204e-01 -7.16566324e-01 -1.14570808e+00 -1.94534048e-01
3.83977070e-02 -7.67805040e-01 -3.56557906e-01 2.76221007e-01
-7.02458799e-01 -1.74806738e+00 2.12353498e-01 -1.13254711e-01
5.20543516e-01 2.91830331e-01 3.99517417e-01 4.89106059e-01
-2.45381221e-01 7.83992708e-02 -8.22610617e-01 -1.36672691e-01
-6.00780427e-01 -5.09578325e-02 3.04531693e-01 1.68636441e-01
7.65388310e-01 -4.12984163e-01 -5.52795567e-02 7.52608061e-01
-7.26909995e-01 -7.40649164e-01 4.53481674e-01 6.66260302e-01
-1.78043187e-01 5.91096759e-01 -4.04759981e-02 -8.86523545e-01
6.99159741e-01 -8.57260942e-01 -6.83029890e-01 1.31951421e-01
-7.85020769e-01 -2.46524200e-01 7.62822926e-01 -7.73215294e-01
-8.18843842e-01 -2.46164039e-01 1.13219090e-01 -1.31244451e-01
-3.83228779e-01 6.17533922e-01 4.50736508e-02 -1.74024612e-01
8.98846865e-01 1.73338637e-01 -1.54309481e-01 -5.30315578e-01
-1.44086331e-01 1.02375257e+00 8.21763575e-02 -3.24494094e-01
1.30596435e+00 3.03161144e-01 -3.83558869e-01 -7.15685189e-01
-3.83438170e-01 -5.83029032e-01 -6.17390156e-01 -2.21299574e-01
8.55807602e-01 -2.48135701e-02 -1.11403668e+00 4.88410681e-01
-1.20165706e+00 1.89147606e-01 4.67404723e-01 7.25876987e-01
2.70272344e-02 6.30296767e-01 -7.86849022e-01 -8.92045557e-01
-2.15181604e-01 -1.24139082e+00 3.42727542e-01 4.84406017e-02
-8.47342908e-01 -7.11746097e-01 2.97845602e-01 8.00597191e-01
4.89986002e-01 6.42618686e-02 1.15325785e+00 -1.63095462e+00
-2.32503101e-01 -7.42857099e-01 -2.68329769e-01 1.90528333e-01
1.46266624e-01 5.56294560e-01 -8.49323094e-01 -2.98463851e-01
5.97404420e-01 1.58387661e-01 2.37786457e-01 -5.55044152e-02
8.83720577e-01 -5.87726593e-01 -3.34912926e-01 6.45283580e-01
1.21859932e+00 9.38618243e-01 5.10480464e-01 5.64351976e-01
6.49097860e-01 7.11981475e-01 5.20875394e-01 4.77472246e-01
1.31923854e-01 4.35116619e-01 3.97294074e-01 5.26842177e-01
7.75603235e-01 -4.14748251e-01 6.70993447e-01 8.37890625e-01
-3.91759366e-01 3.28160435e-01 -1.22081089e+00 -2.47166708e-01
-1.40647662e+00 -1.07466292e+00 -3.12864065e-01 2.33967710e+00
2.33735532e-01 2.27598369e-01 3.23789775e-01 4.51606184e-01
7.28799462e-01 -4.84708734e-02 4.15340159e-03 -6.11389339e-01
4.29417878e-01 1.04937568e-01 4.38083827e-01 3.25425297e-01
-1.05359423e+00 4.71833944e-01 6.27415800e+00 5.92753530e-01
-1.37880349e+00 -2.45930310e-02 4.34825659e-01 4.77283210e-01
3.65602374e-02 5.48937857e-01 -7.55796552e-01 6.18662953e-01
1.13427997e+00 -7.43317679e-02 5.76895475e-01 1.06480908e+00
1.65207297e-01 -2.04582214e-01 -9.31862056e-01 8.76654088e-01
1.89896792e-01 -1.25063920e+00 -2.32298877e-02 3.94878954e-01
2.31511414e-01 -5.23007624e-02 -2.80689806e-01 9.67969224e-02
3.33165318e-01 -1.04301882e+00 4.00479674e-01 1.19894095e-01
2.44484410e-01 -5.75496018e-01 1.15078640e+00 5.14614940e-01
-1.12072647e+00 -4.36114252e-01 -1.49477022e-02 -3.79693002e-01
-1.51351690e-01 4.65888143e-01 -1.07539511e+00 6.43847108e-01
8.29910278e-01 2.52670974e-01 -5.53736925e-01 8.14802051e-01
-4.64946292e-02 8.78719151e-01 -2.18687519e-01 -2.71182358e-01
-1.51011884e-01 -2.51850128e-01 5.65494478e-01 1.04376328e+00
3.83299321e-01 4.61761542e-02 -3.69786955e-02 7.81516671e-01
4.32509154e-01 2.07436278e-01 -7.66984165e-01 -3.49535048e-01
7.75732815e-01 1.26514637e+00 -9.99453664e-01 -4.75035310e-02
-1.86729684e-01 5.00928819e-01 -4.26079273e-01 8.63156170e-02
-9.94292557e-01 -7.40513980e-01 3.76473427e-01 3.07014704e-01
1.00399107e-01 -3.33922952e-01 -4.64209110e-01 -1.15224600e+00
-1.72711089e-01 -1.24845362e+00 5.16379774e-01 -3.23866308e-01
-1.25485468e+00 8.16315711e-01 1.69645976e-02 -1.16124988e+00
-2.65173554e-01 -6.47695482e-01 -9.77671385e-01 5.54266393e-01
-4.82307345e-01 -7.47527957e-01 -2.10725278e-01 4.58599836e-01
1.37553997e-02 -4.04193878e-01 5.76569796e-01 1.79852080e-02
-6.63263977e-01 3.54939103e-01 -3.13299328e-01 5.57701349e-01
4.68393803e-01 -7.98489809e-01 3.85931646e-03 9.66890275e-01
3.12678479e-02 1.41268992e+00 5.71153402e-01 -1.00272036e+00
-1.50922835e+00 -3.77917945e-01 9.11695540e-01 -8.73827100e-01
1.08074820e+00 -1.23571090e-01 -7.28715181e-01 6.83601677e-01
-1.92820325e-01 -3.62888634e-01 1.28697789e+00 -6.14284305e-03
-8.29568148e-01 1.52770057e-01 -1.42449045e+00 2.69487351e-01
3.74938518e-01 -5.50004900e-01 -7.43238986e-01 -1.97622806e-01
3.06854993e-01 9.32274684e-02 -8.84061217e-01 1.92196772e-01
7.95829058e-01 -1.37084639e+00 6.72150493e-01 -7.96190977e-01
3.50013524e-01 -4.29291606e-01 -4.02062863e-01 -7.67231345e-01
-6.94384724e-02 -9.84385133e-01 8.33185688e-02 1.20003104e+00
2.52738118e-01 -1.19670272e+00 5.34661710e-01 3.07116151e-01
3.66928697e-01 -6.91209733e-01 -7.55692542e-01 -7.08072722e-01
-2.78016120e-01 -5.68955958e-01 7.12281406e-01 1.55566239e+00
4.22735631e-01 1.93880320e-01 -3.00377667e-01 1.71790645e-01
5.81130087e-01 2.29472350e-02 1.01353097e+00 -1.24436522e+00
-7.45122373e-01 -5.01854420e-01 -4.98525977e-01 -4.91668046e-01
-2.11175427e-01 -7.49885440e-01 -3.86014849e-01 -4.73343998e-01
3.73804152e-01 -4.40900147e-01 -3.83139923e-02 4.74118173e-01
5.01026101e-02 2.13670358e-01 3.34719390e-01 3.83938104e-01
-7.50735998e-02 -2.47594699e-01 4.30540651e-01 2.10921317e-01
-3.69616479e-01 2.25910276e-01 -6.18232250e-01 1.02989793e+00
8.56651604e-01 -6.85263872e-01 -1.54982984e-01 3.33030641e-01
3.33985597e-01 -3.97180766e-02 2.07194030e-01 -7.14795291e-01
9.82207805e-03 -4.75231647e-01 2.80310273e-01 -4.49616969e-01
-2.65030175e-01 -1.02244520e+00 4.96152729e-01 1.00617981e+00
1.19768783e-01 4.42763388e-01 -1.06722318e-01 1.95284382e-01
6.71845973e-02 -5.09667814e-01 6.46800637e-01 2.55726546e-01
-1.92942739e-01 -7.78404698e-02 -9.06860888e-01 -3.18592072e-01
1.27916896e+00 -4.61122602e-01 -5.58477581e-01 -1.12414256e-01
-2.82994181e-01 -4.14586514e-01 6.45028293e-01 4.31686580e-01
5.74854434e-01 -7.73529708e-01 -2.79557586e-01 4.38163906e-01
-9.20187905e-02 -9.55376208e-01 -2.10829526e-01 1.28308535e+00
-8.34754944e-01 1.86541051e-01 -2.67239988e-01 -4.24817979e-01
-1.58636773e+00 7.21814036e-01 2.16059104e-01 -2.66575396e-01
-2.08350554e-01 3.40529084e-01 -5.03259957e-01 -4.80051219e-01
-1.91137776e-01 -5.93501665e-02 -2.47388750e-01 3.59143987e-02
6.85988188e-01 9.72612679e-01 -1.21554233e-01 -7.28643119e-01
-6.83532655e-01 4.14954484e-01 -1.03760511e-01 1.41770855e-01
1.14165747e+00 3.56750757e-01 -6.92680657e-01 3.57634753e-01
1.09475005e+00 6.31746590e-01 -2.88446158e-01 5.30932665e-01
4.44106996e-01 -9.21830058e-01 -3.81355643e-01 -7.52503812e-01
-7.26482689e-01 5.07644832e-01 2.76473939e-01 7.03016162e-01
8.98489058e-01 -1.35637075e-01 6.02276981e-01 2.21400261e-01
7.33920932e-01 -4.82977986e-01 -8.01020116e-02 5.77106595e-01
3.40073436e-01 -1.00248134e+00 -6.37825066e-03 -5.67976356e-01
-3.99492532e-01 1.37982583e+00 5.02266586e-01 -1.06504135e-01
8.85598838e-01 4.61926669e-01 1.81163698e-01 -6.25055283e-02
-4.87347990e-01 5.36624372e-01 2.09629610e-01 6.14681065e-01
4.36563581e-01 2.47831002e-01 -6.18194759e-01 8.64089668e-01
-4.86789346e-01 -2.08733141e-01 7.10445166e-01 8.23279858e-01
-4.22805130e-01 -1.12901378e+00 -8.88105512e-01 7.30563819e-01
-7.14779735e-01 2.62070429e-02 -9.20607388e-01 5.80605447e-01
4.75648828e-02 1.26734471e+00 -3.14215809e-01 -1.36960578e+00
-1.04751147e-01 -4.19755727e-02 -3.86096835e-02 -4.25845832e-01
-1.15019572e+00 -3.04226279e-01 1.26136169e-01 -6.74901247e-01
1.62192747e-01 -4.35776561e-01 -8.79422069e-01 -9.40901339e-01
-5.19790769e-01 4.21802521e-01 6.94451511e-01 1.03568339e+00
2.26890385e-01 -1.00073136e-01 9.57034111e-01 -7.99443245e-01
-8.97814572e-01 -8.11405361e-01 -2.76792616e-01 3.22630197e-01
1.21149004e-01 -6.90552711e-01 -8.43506575e-01 -1.16740875e-01] | [14.399721145629883, 9.671430587768555] |
45307256-b613-42c2-ad28-ee116c8ee57f | sex-detection-in-the-early-stage-of | 2305.02325 | null | https://arxiv.org/abs/2305.02325v1 | https://arxiv.org/pdf/2305.02325v1.pdf | Sex Detection in the Early Stage of Fertilized Chicken Eggs via Image Recognition | Culling newly hatched male chicks in industrial hatcheries poses a serious ethical problem. Both laying and broiler breeders need males, but it is a problem because they are produced more than needed. Being able to determine the sex of chicks in the egg at the beginning or early stage of incubation can eliminate ethical problems as well as many additional costs. When we look at the literature, the methods used are very costly, low in applicability, invasive, inadequate in accuracy, or too late to eliminate ethical problems. Considering the embryo's development, the earliest observed candidate feature for sex determination is blood vessels. Detection from blood vessels can eliminate ethical issues, and these vessels can be seen when light is shined into the egg until the first seven days. In this study, sex determination was made by morphological analysis from embryonic vascular images obtained in the first week when the light was shined into the egg using a standard camera without any invasive procedure to the egg. | ['Efendi Nasibov', 'Ufuk Asil'] | 2023-05-03 | null | null | null | null | ['morphological-analysis'] | ['natural-language-processing'] | [-1.49205521e-01 2.07078122e-02 2.09544569e-01 -5.46970181e-02
8.24517727e-01 -8.06780696e-01 -1.56015888e-01 4.96542424e-01
-6.85238123e-01 5.79283774e-01 -5.14077187e-01 -3.15195441e-01
2.20825542e-02 -7.74102449e-01 -4.65047359e-01 -7.13022232e-01
3.35631073e-02 -4.30675521e-02 4.33467597e-01 4.27068211e-03
4.01926994e-01 7.06225872e-01 -1.51922703e+00 -3.94760996e-01
4.80698794e-01 3.22859794e-01 1.65528655e-01 6.42082155e-01
7.12916255e-02 2.20171645e-01 -6.19868517e-01 -6.72935665e-01
5.03393531e-01 -1.03635514e+00 -1.36719435e-01 -1.63929254e-01
-1.46125704e-01 -9.64903235e-01 6.16315067e-01 8.46428812e-01
3.01486939e-01 -1.72406629e-01 7.07852066e-01 -8.54237258e-01
-3.16419989e-01 4.14083391e-01 -8.54934871e-01 2.52409488e-01
1.09795168e-01 2.71682031e-02 2.21168473e-01 -7.29192793e-01
6.29729569e-01 7.13118494e-01 3.75077248e-01 8.21254015e-01
-1.18705225e+00 -5.75880051e-01 -5.29651761e-01 -2.28609487e-01
-1.01118350e+00 -5.66157162e-01 2.11755604e-01 -9.80278611e-01
1.46548018e-01 -5.22608832e-02 1.14784145e+00 3.16846997e-01
7.85943985e-01 6.55303104e-03 1.01135421e+00 -4.61677343e-01
2.90449888e-01 3.54993641e-01 -2.83318490e-01 7.73515284e-01
1.00418878e+00 5.74031413e-01 -2.39012033e-01 1.55556768e-01
1.15720594e+00 -1.93994194e-02 -2.49749780e-01 -3.82401764e-01
-1.00779724e+00 7.59638369e-01 -2.33897477e-01 3.66721570e-01
-1.52357683e-01 -1.04788616e-01 4.29388583e-01 3.80494297e-01
1.76439747e-01 8.40777874e-01 -4.95938927e-01 -1.56013593e-01
-8.45799387e-01 -2.92123616e-01 7.08527684e-01 3.70366663e-01
2.14608237e-01 -1.36243030e-01 5.45545518e-01 6.58571064e-01
4.72134680e-01 3.84943813e-01 1.72709655e-02 -9.67314422e-01
-4.95831460e-01 5.71347773e-01 -2.63504475e-01 -8.03298533e-01
-3.64099145e-02 -5.28585091e-02 -4.03135478e-01 9.25039530e-01
8.66712749e-01 -7.02041864e-01 -5.07226229e-01 1.26118755e+00
8.31566453e-01 -5.33037186e-01 9.20792669e-02 9.91264224e-01
1.10583687e+00 6.40418112e-01 1.73071586e-02 -6.06830120e-01
1.51371348e+00 -4.74087864e-01 -7.67455578e-01 -1.21119790e-01
2.67143369e-01 -1.03259218e+00 7.58952945e-02 3.22231621e-01
-1.02980030e+00 -1.19621001e-01 -1.36646760e+00 2.62458771e-01
-2.87659854e-01 3.52369428e-01 4.31792885e-01 7.98431218e-01
-7.61177659e-01 6.38308287e-01 -9.60688949e-01 -7.68206179e-01
-3.29916388e-01 4.13537562e-01 -7.93577015e-01 4.78737473e-01
-4.28275555e-01 1.21396124e+00 -4.13953327e-02 4.71106708e-01
-7.34652936e-01 -7.26546764e-01 -9.13852632e-01 -1.77042708e-01
3.20371926e-01 -3.64124596e-01 9.25337493e-01 -5.08526206e-01
-1.84709811e+00 1.11381197e+00 1.54438421e-01 -2.22781688e-01
3.64293575e-01 -1.19462758e-01 1.40504735e-02 4.26694065e-01
1.24814101e-01 4.40367639e-01 6.25236690e-01 -9.66315806e-01
-6.46814883e-01 -3.43581945e-01 -4.82427068e-02 -1.99980736e-01
4.84786987e-01 6.14877462e-01 1.32961750e-01 -4.04231697e-01
3.15740407e-01 -6.72806323e-01 -2.36624345e-01 2.64506578e-01
3.79881948e-01 1.75482944e-01 4.46339339e-01 -6.49038315e-01
6.55228972e-01 -2.35113454e+00 -2.90651381e-01 -1.66407809e-01
1.98262379e-01 2.60204285e-01 2.53071368e-01 5.75640976e-01
-8.13563261e-03 9.46080871e-03 2.88315624e-01 5.98011315e-01
-1.84797615e-01 1.09358961e-02 4.70252216e-01 8.65917206e-01
-8.79093334e-02 2.58500397e-01 -8.81844521e-01 -7.66368091e-01
2.90226996e-01 3.48090529e-01 -2.55255431e-01 3.15076202e-01
6.00625396e-01 1.82554349e-01 -1.92722261e-01 6.75171554e-01
5.13304591e-01 4.06327069e-01 1.17296398e-01 9.57157165e-02
-4.55431819e-01 -2.12265208e-01 -7.24333525e-01 9.00624454e-01
-1.61879629e-01 7.71760464e-01 4.81391877e-01 -6.18008912e-01
8.27411890e-01 7.04853415e-01 3.85415256e-01 -2.42739588e-01
4.75595713e-01 2.75428921e-01 3.74926686e-01 -6.68519557e-01
2.78599679e-01 -5.48654616e-01 4.36676055e-01 4.76681918e-01
4.56303433e-02 -2.52126485e-01 7.31070042e-01 -4.13087904e-02
6.13777459e-01 1.73791125e-01 5.99127591e-01 -6.47451460e-01
1.98769733e-01 -1.80541933e-01 8.84068549e-01 -1.37380153e-01
-5.03376983e-02 4.92299348e-01 8.43725204e-01 -6.62895501e-01
-7.96901584e-01 -7.16136038e-01 -4.05440509e-01 8.84361088e-01
2.89032966e-01 5.92029579e-02 -1.05585933e+00 -4.69938904e-01
1.49645478e-01 4.78929818e-01 -6.53854787e-01 -1.11629896e-01
-2.18017384e-01 -4.51448292e-01 1.32677570e-01 1.96153402e-01
1.00552723e-01 -8.14421356e-01 -1.48545372e+00 2.91638374e-01
4.67471868e-01 -5.14194548e-01 -1.28059864e-01 2.78232008e-01
-9.03453052e-01 -1.21665144e+00 -1.07860589e+00 -7.57724524e-01
1.33671510e+00 2.28339911e-01 5.49124837e-01 5.22975564e-01
-4.93397355e-01 -3.55830789e-02 -3.30520511e-01 -7.21117795e-01
-7.03280866e-01 -5.74701548e-01 9.02430043e-02 -4.33674514e-01
4.82613266e-01 -2.03599483e-01 -5.80509663e-01 5.01096725e-01
-4.61283743e-01 -1.54519811e-01 4.76544827e-01 7.72554457e-01
1.32464215e-01 1.06778190e-01 3.48545223e-01 -8.44640315e-01
9.13523585e-02 -2.85838425e-01 -9.33834195e-01 1.37646273e-01
-4.47193712e-01 -3.14688422e-02 5.12362301e-01 -2.01941818e-01
-9.17825580e-01 -2.27853075e-01 1.03460848e-01 1.76322386e-01
-6.13453388e-01 2.18279019e-01 2.83863157e-01 -4.02819902e-01
4.70731169e-01 -3.36226314e-01 4.07014698e-01 -5.17993152e-01
-5.06621480e-01 8.00330639e-02 4.50201690e-01 -3.47074002e-01
9.52305853e-01 6.90738037e-02 4.69602138e-01 -9.22147095e-01
-1.76313385e-01 -2.39100456e-01 -7.37015009e-01 -7.22691298e-01
8.98466766e-01 -4.69752967e-01 -8.05503130e-01 3.84102225e-01
-9.09213781e-01 -6.91526979e-02 -1.07178181e-01 1.17094064e+00
1.41559646e-01 5.09591103e-01 -8.48026454e-01 -6.30693853e-01
-1.77395418e-01 -1.07906222e+00 3.56032342e-01 5.59525371e-01
-1.33680865e-01 -8.76573145e-01 9.52895451e-03 2.17150137e-01
1.33145317e-01 7.18987882e-01 1.03440881e+00 -2.14781106e-01
-1.66745484e-01 -4.54248607e-01 8.86116549e-02 2.85651177e-01
2.36890644e-01 1.03176093e+00 -4.35180426e-01 -2.60100871e-01
1.60764039e-01 5.60842007e-02 3.48478466e-01 7.04216301e-01
3.28254104e-02 -1.94935463e-02 -2.91448981e-01 5.25558293e-01
1.39067698e+00 8.56960654e-01 1.32163331e-01 1.41521379e-01
-7.20267668e-02 1.30559313e+00 9.78868484e-01 2.52329737e-01
-4.79264967e-02 6.16706582e-03 2.65960902e-01 -4.60906893e-01
2.93176360e-02 1.45995505e-02 2.98202753e-01 7.74432361e-01
-7.02048302e-01 2.14797650e-02 -6.39942706e-01 6.86937094e-01
-9.42936599e-01 -8.24441433e-01 -5.91615975e-01 2.45691347e+00
7.82811224e-01 -1.38426706e-01 8.90184939e-02 2.87184399e-02
8.70659709e-01 -6.86831415e-01 2.39684567e-01 -9.37306345e-01
4.13811594e-01 3.75479497e-02 5.82180083e-01 2.18267515e-01
-5.20936012e-01 2.11214796e-01 7.10441351e+00 -4.75216731e-02
-1.25328922e+00 -3.95320863e-01 4.16938901e-01 -2.50899225e-01
3.01428348e-01 2.02517852e-01 -6.10449314e-01 4.12885696e-01
3.85590881e-01 -3.32399189e-01 -3.82340141e-02 6.50308132e-01
-4.45269160e-02 -1.10280430e+00 -1.44226861e+00 4.43265975e-01
-1.23315312e-01 -7.47790039e-01 -6.28136754e-01 2.99712032e-01
3.65658075e-01 -7.97423780e-01 -2.05185369e-01 -3.71221989e-01
-3.23413312e-01 -7.31594086e-01 5.75899184e-01 6.00796714e-02
8.50173235e-01 -7.74554133e-01 1.24599206e+00 3.03391963e-01
-8.74776542e-01 4.82052565e-01 -5.81308305e-01 -3.26105207e-01
-6.51321039e-02 2.78590739e-01 -9.88196611e-01 -3.06931026e-02
4.52378333e-01 5.46668135e-02 -4.89593476e-01 1.41266310e+00
-2.79220611e-01 5.65407753e-01 -4.03197140e-01 -2.88508594e-01
-1.02595776e-01 -6.22537136e-01 4.19582188e-01 6.81560814e-01
4.99603122e-01 3.81959260e-01 -8.81399691e-01 8.82750750e-01
5.86757600e-01 3.85916829e-01 -7.37418532e-01 -8.26018006e-02
9.78430137e-02 1.25873792e+00 -1.59826374e+00 1.76609725e-01
-6.79700732e-01 5.94560385e-01 -6.30531907e-01 1.66003689e-01
-3.59758973e-01 -7.67616391e-01 2.39785030e-01 3.20035487e-01
8.79359320e-02 3.22883338e-01 5.72479144e-02 -5.05331159e-01
-1.63484707e-01 -3.93800408e-01 1.25714496e-01 -1.70266062e-01
-5.75778008e-01 1.24925682e-02 4.20821518e-01 -1.13525677e+00
-1.40733108e-01 -6.34601057e-01 -7.63602674e-01 4.74679947e-01
-7.20009387e-01 -4.47652012e-01 1.43567637e-01 -4.34268981e-01
3.71103287e-01 -2.19836887e-02 6.18895948e-01 2.34574884e-01
-5.83199501e-01 3.39671910e-01 7.76977763e-02 1.76898822e-01
7.30889201e-01 -1.22666073e+00 -1.55392200e-01 1.07664204e+00
-4.98043507e-01 8.49228323e-01 9.43299353e-01 -6.61562502e-01
-1.05730617e+00 -1.66285545e-01 1.11774981e+00 4.27392386e-02
1.40903175e-01 -2.89764732e-01 -4.86453235e-01 3.22224230e-01
4.99499947e-01 -2.62423873e-01 1.11793053e+00 -2.27739990e-01
4.51894432e-01 -1.44323707e-01 -1.16991723e+00 5.04049659e-01
3.11976578e-02 1.53718621e-01 -7.13168442e-01 -1.61897197e-01
3.41186188e-02 -3.43799263e-01 -1.02277792e+00 3.55885595e-01
1.02675653e+00 -1.04455757e+00 5.55162251e-01 -5.11303358e-02
6.63014591e-01 -4.58451301e-01 8.38973820e-01 -1.33654845e+00
-2.36119211e-01 -5.12057185e-01 9.92489100e-01 1.38188231e+00
7.99356639e-01 -5.86026490e-01 5.45315981e-01 7.62766182e-01
2.48222142e-01 -4.52511519e-01 -5.28412700e-01 -5.04680514e-01
-8.63482952e-02 7.61868298e-01 4.39425349e-01 8.92346203e-01
5.62034011e-01 1.54885575e-01 -2.71695107e-01 1.47422403e-01
8.09815705e-01 1.36244550e-01 5.58876038e-01 -1.27516210e+00
-9.76832435e-02 -1.81147143e-01 -5.59238672e-01 -2.59854406e-01
-5.24463236e-01 -1.76287085e-01 3.46578121e-01 -1.41068292e+00
1.17898881e-02 2.16585428e-01 2.79672444e-01 2.28106275e-01
1.17973275e-02 4.77084937e-03 -6.20373711e-02 -5.12511849e-01
2.64046073e-01 -2.98376381e-01 1.39977193e+00 3.58084291e-01
-1.20530836e-01 2.22000107e-01 -2.62622297e-01 9.73345876e-01
6.89409316e-01 -7.63497353e-01 -2.97428668e-01 -3.36468190e-01
4.32926893e-01 6.87261879e-01 1.09834932e-01 -6.92859828e-01
3.91479015e-01 -1.95402414e-01 4.92421478e-01 -5.63637853e-01
-8.59448686e-02 -1.10840309e+00 7.71570206e-01 8.55996609e-01
2.33391989e-02 -1.42248288e-01 2.83753574e-02 -1.50599390e-01
-8.32012109e-03 -9.72253859e-01 1.10521019e+00 -2.73367256e-01
-2.35793054e-01 -3.84681821e-01 -1.09496093e+00 -2.99801052e-01
1.53708839e+00 -5.93829453e-01 -2.82051325e-01 1.16382539e-01
-5.04419744e-01 1.53306678e-01 1.07163191e+00 -2.15853766e-01
5.23400724e-01 -7.26825953e-01 -6.44938827e-01 4.04716372e-01
-3.88825312e-02 3.30287479e-02 1.40644625e-01 1.27709532e+00
-1.75164354e+00 7.48326778e-02 -6.62198365e-01 -3.44099373e-01
-1.69217622e+00 7.31972873e-01 7.10683316e-02 5.62112451e-01
-5.23223937e-01 1.18493235e+00 5.67493618e-01 2.41336718e-01
-5.85801527e-02 -5.39920807e-01 -4.16997284e-01 9.14738178e-02
3.54251266e-01 7.69059002e-01 -2.36452743e-01 -4.93595541e-01
-2.64580727e-01 9.29199755e-01 1.58512428e-01 2.52625465e-01
1.15726519e+00 -9.64849442e-02 -5.77682912e-01 4.64137226e-01
7.92364359e-01 4.09654051e-01 -9.53105450e-01 3.66667181e-01
-4.22638088e-01 -9.77609515e-01 -1.26972720e-01 -5.57456136e-01
-1.32117784e+00 1.06571198e+00 1.73975229e-01 5.25314748e-01
9.91790295e-01 -1.47498339e-01 3.96051347e-01 2.11764015e-02
4.95841950e-01 -1.20608115e+00 -4.96342421e-01 -2.12292761e-01
6.05042160e-01 -1.12018371e+00 3.66875440e-01 -5.84402084e-01
-5.07305801e-01 1.33791161e+00 7.00314224e-01 -5.43215573e-02
5.81773877e-01 5.66768229e-01 4.80789274e-01 -3.47430527e-01
-8.51571321e-01 2.01388076e-01 -3.12628560e-02 5.97854733e-01
8.36587608e-01 -8.39512348e-02 -1.24277282e+00 2.03673050e-01
-2.58113712e-01 -1.44100580e-02 9.63576853e-01 1.04151154e+00
-5.84314525e-01 -9.30795908e-01 -4.62532014e-01 4.68052268e-01
-1.16397667e+00 2.03185663e-01 -3.73265892e-01 7.18683183e-01
3.67723048e-01 9.41013098e-01 -1.01060688e-01 2.06774652e-01
2.08135396e-01 -2.45142132e-01 5.46555519e-01 -7.26777434e-01
-5.84443688e-01 3.81839901e-01 4.04954925e-02 -1.29923478e-01
-4.68210936e-01 -7.51660168e-01 -1.12369168e+00 -4.62919861e-01
-7.81761885e-01 8.02884996e-01 9.57839966e-01 5.58752298e-01
-4.53297019e-01 -1.02763779e-01 5.79652071e-01 -2.27920696e-01
-3.43423262e-02 -5.37050784e-01 -1.40608442e+00 -4.69194263e-01
2.99712151e-01 -6.02189422e-01 -9.03909206e-01 6.66498125e-01] | [14.443156242370605, -2.940850019454956] |
1ba8a4e1-7bc3-46b1-9442-59c560210c62 | adaptive-graph-convolutional-network-with | 2003.06167 | null | https://arxiv.org/abs/2003.06167v1 | https://arxiv.org/pdf/2003.06167v1.pdf | Adaptive Graph Convolutional Network with Attention Graph Clustering for Co-saliency Detection | Co-saliency detection aims to discover the common and salient foregrounds from a group of relevant images. For this task, we present a novel adaptive graph convolutional network with attention graph clustering (GCAGC). Three major contributions have been made, and are experimentally shown to have substantial practical merits. First, we propose a graph convolutional network design to extract information cues to characterize the intra- and interimage correspondence. Second, we develop an attention graph clustering algorithm to discriminate the common objects from all the salient foreground objects in an unsupervised fashion. Third, we present a unified framework with encoder-decoder structure to jointly train and optimize the graph convolutional network, attention graph cluster, and co-saliency detection decoder in an end-to-end manner. We evaluate our proposed GCAGC method on three cosaliency detection benchmark datasets (iCoseg, Cosal2015 and COCO-SEG). Our GCAGC method obtains significant improvements over the state-of-the-arts on most of them. | ['Jin Chen', 'Shiwen Shen', 'Tengpeng Li', 'Kaihua Zhang', 'Bo Liu', 'Qingshan Liu'] | 2020-03-13 | adaptive-graph-convolutional-network-with-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Zhang_Adaptive_Graph_Convolutional_Network_With_Attention_Graph_Clustering_for_Co-Saliency_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zhang_Adaptive_Graph_Convolutional_Network_With_Attention_Graph_Clustering_for_Co-Saliency_CVPR_2020_paper.pdf | cvpr-2020-6 | ['co-saliency-detection'] | ['computer-vision'] | [ 3.94783884e-01 1.75733238e-01 -1.41571268e-01 -1.69306815e-01
-7.73106694e-01 -1.96114972e-01 5.01483679e-01 5.23686856e-02
6.76383525e-02 1.74877211e-01 2.63610959e-01 -1.16013721e-01
1.52042389e-01 -3.60953003e-01 -9.05367315e-01 -5.29366314e-01
-4.70417738e-02 1.28109872e-01 6.87318861e-01 1.32284775e-01
3.47443789e-01 2.41127908e-01 -1.57585359e+00 3.51148397e-01
1.12536585e+00 8.82410765e-01 6.38376713e-01 7.71230817e-01
2.13704869e-01 1.10213780e+00 -1.92508966e-01 -2.68164366e-01
6.76792189e-02 -6.32485867e-01 -7.94382274e-01 4.78655726e-01
7.60605752e-01 1.47570996e-02 -4.62049633e-01 1.13994014e+00
4.62141037e-01 8.12925771e-02 5.58070004e-01 -1.30233276e+00
-9.05377150e-01 7.20274985e-01 -9.50239897e-01 7.41205871e-01
-1.74425602e-01 6.71871752e-02 1.20824099e+00 -1.08683658e+00
6.51109695e-01 1.29026616e+00 3.04425895e-01 1.70714542e-01
-9.97175992e-01 -5.76130033e-01 5.82307637e-01 5.21879256e-01
-1.39594984e+00 -2.61993080e-01 1.07200217e+00 -4.56399858e-01
8.22745085e-01 6.02842011e-02 7.79797137e-01 6.71070337e-01
1.86927497e-01 1.35552764e+00 5.96026599e-01 -2.84610033e-01
1.54277682e-01 -4.49537426e-01 1.67636827e-01 9.49900448e-01
4.01152879e-01 -3.05686921e-01 -6.74069226e-01 1.35720253e-01
5.94651461e-01 -1.16595291e-01 -2.19954535e-01 -5.94884813e-01
-1.14051330e+00 7.07335949e-01 8.68365705e-01 2.81116515e-01
-4.43958789e-01 4.58161771e-01 2.71358281e-01 -2.87680149e-01
7.60798573e-01 1.66664958e-01 -1.31212220e-01 3.64227831e-01
-9.25751328e-01 -3.50023471e-02 2.81458616e-01 1.20206094e+00
7.59636104e-01 1.47537336e-01 -5.52163422e-01 5.15395522e-01
4.97426420e-01 4.83450741e-01 1.17510557e-01 -6.49898648e-01
4.45151031e-01 7.41538942e-01 -3.38868141e-01 -1.33914566e+00
-2.48256683e-01 -8.24759722e-01 -7.04559684e-01 -2.42850721e-01
-6.31884560e-02 -1.01282917e-01 -1.10547698e+00 1.45184565e+00
3.62800270e-01 8.15279245e-01 -8.43864754e-02 1.15102589e+00
1.04894197e+00 5.84557831e-01 2.58192956e-01 6.93288147e-02
1.22321069e+00 -1.42646968e+00 -7.64972270e-01 -6.55875921e-01
3.17226112e-01 -6.84144914e-01 9.15762782e-01 5.87553764e-03
-1.16482496e+00 -6.99333072e-01 -1.04781425e+00 -2.80976593e-01
-2.40823299e-01 3.52881938e-01 7.94923484e-01 2.25086361e-01
-1.21519876e+00 3.67133558e-01 -9.83199835e-01 -3.56034875e-01
9.26908910e-01 1.70135304e-01 2.33534813e-01 8.31729099e-02
-7.64886022e-01 3.51755887e-01 5.45386016e-01 1.97654262e-01
-1.14849305e+00 -4.82456654e-01 -1.02502942e+00 3.69816959e-01
6.07974589e-01 -7.42238641e-01 9.65039611e-01 -1.28936374e+00
-1.10421908e+00 1.05798268e+00 -2.39354417e-01 -6.17941201e-01
1.40574932e-01 -3.29251707e-01 -2.99654901e-01 3.15283448e-01
5.52578092e-01 7.15457618e-01 9.01208043e-01 -1.26005423e+00
-7.82891095e-01 -1.53851226e-01 -3.36973578e-01 2.54952908e-01
-5.79943359e-02 3.88243258e-01 -1.18865645e+00 -9.04404879e-01
1.78719521e-01 -7.67804980e-01 -2.02531204e-01 -3.33524942e-01
-8.01595807e-01 -2.31907085e-01 1.08894444e+00 -5.51131606e-01
1.14134920e+00 -2.14557028e+00 3.79526258e-01 -1.89031631e-01
7.39269674e-01 2.98980534e-01 -1.26814591e-02 -6.01109001e-04
-9.13907066e-02 4.30239029e-02 -4.32302445e-01 -5.64082146e-01
-1.55760601e-01 -1.35600030e-01 -2.09108874e-01 5.90386212e-01
6.92962468e-01 1.32318687e+00 -1.16535878e+00 -7.63845444e-01
3.73124272e-01 4.42054629e-01 -4.96811658e-01 3.34185392e-01
-2.55319595e-01 2.07393289e-01 -6.33304477e-01 8.15660894e-01
5.34328759e-01 -9.92916405e-01 2.01823071e-01 -2.64410377e-01
1.41015174e-02 1.43277228e-01 -8.99898529e-01 1.58888292e+00
1.71790794e-01 1.09623587e+00 1.03472680e-01 -9.36958373e-01
6.16791785e-01 -1.49386331e-01 3.17573994e-01 -6.14081562e-01
4.04165059e-01 6.68512424e-03 -1.26802161e-01 -2.87636518e-01
5.52749872e-01 4.54071283e-01 1.47071093e-01 1.40010297e-01
3.18013668e-01 2.82072604e-01 2.77545422e-01 6.29303157e-01
8.97351325e-01 5.08327298e-02 1.47621781e-01 -7.09328115e-01
5.54604053e-01 -1.62449047e-01 6.93815947e-01 6.90545380e-01
-4.52620089e-01 9.97482836e-01 6.93245947e-01 -1.31842613e-01
-5.19720793e-01 -9.80198741e-01 5.12629092e-01 1.26761150e+00
7.77292430e-01 -4.46182936e-01 -7.71581471e-01 -9.22137380e-01
-2.07067505e-01 5.66550195e-01 -8.06155205e-01 -1.07020818e-01
-6.19794488e-01 -7.08282828e-01 1.31760910e-01 7.37574339e-01
6.86874449e-01 -1.06511879e+00 -5.76496184e-01 9.14104506e-02
-2.91755080e-01 -1.45738649e+00 -9.59596872e-01 5.87060526e-02
-5.54300487e-01 -1.30925965e+00 -5.64657390e-01 -1.28137934e+00
6.72250986e-01 9.20931935e-01 1.39260697e+00 3.46628785e-01
-2.66614288e-01 2.24781811e-01 -3.78562748e-01 -4.47952956e-01
1.36908859e-01 1.75718069e-01 -6.11942589e-01 4.04509753e-01
3.03252667e-01 -2.71436632e-01 -8.11526597e-01 1.83845595e-01
-5.67030907e-01 5.64402163e-01 5.68944871e-01 5.01914144e-01
8.67727995e-01 -2.51402229e-01 5.14035404e-01 -1.01733994e+00
2.32957408e-01 -5.11545718e-01 -7.62907207e-01 3.07953477e-01
-3.87712061e-01 -1.17027149e-01 3.00635219e-01 8.53748918e-02
-7.36056983e-01 2.68802941e-01 2.56316811e-01 -7.74636805e-01
9.12703499e-02 2.65585989e-01 -4.58215445e-01 -1.49510562e-01
2.06474051e-01 2.53719717e-01 -5.05785525e-01 -3.35943580e-01
5.86971045e-01 3.51338625e-01 9.58516896e-01 -1.98944435e-01
6.19659126e-01 5.56601584e-01 -1.82097405e-01 -8.22696805e-01
-1.04077387e+00 -7.35705912e-01 -7.22650468e-01 -4.71626550e-01
1.28873003e+00 -1.24646831e+00 -2.25918710e-01 4.68992025e-01
-1.07912111e+00 -6.16461694e-01 1.44058233e-02 1.88776538e-01
-4.83744770e-01 5.34652114e-01 -5.09624183e-01 -6.26423180e-01
-4.85886872e-01 -1.27294052e+00 1.57519162e+00 4.73340422e-01
2.65844077e-01 -8.86348605e-01 -3.75997007e-01 2.69696385e-01
1.18359856e-01 6.73290730e-01 5.52712262e-01 -4.72810060e-01
-1.06240356e+00 2.53221512e-01 -8.90249312e-01 -1.22622885e-01
1.10139363e-01 7.39910081e-02 -9.21790123e-01 -3.19540083e-01
-5.14069140e-01 -2.24498753e-02 1.17658138e+00 6.91788614e-01
1.32481229e+00 3.62925231e-02 -6.35747969e-01 9.02301490e-01
1.36104298e+00 1.24531440e-01 5.11197209e-01 1.10541426e-01
1.35816765e+00 2.71911651e-01 6.52776718e-01 1.82460863e-02
5.40191650e-01 5.69337010e-01 6.40727460e-01 -5.31710744e-01
-6.89132512e-01 -1.80792123e-01 2.85705984e-01 7.87701190e-01
1.12996362e-01 -6.10711277e-01 -8.73935044e-01 1.12982333e+00
-2.15178466e+00 -6.44323349e-01 -4.82678235e-01 1.72379708e+00
3.50075573e-01 2.02802300e-01 1.37883633e-01 -3.01449329e-01
1.28797376e+00 5.34650624e-01 -6.02105796e-01 1.68715373e-01
-4.24259573e-01 6.77559301e-02 6.60707533e-01 4.53433335e-01
-1.66008544e+00 1.40569437e+00 5.58211851e+00 6.40329540e-01
-9.88209009e-01 1.79888278e-01 9.47499275e-01 1.68570980e-01
-5.94502948e-02 4.19473415e-03 -5.01472712e-01 5.63982964e-01
6.38204992e-01 -2.76716113e-01 1.49183050e-01 9.32869852e-01
8.76189619e-02 1.33182630e-01 -8.03775489e-01 1.00411093e+00
3.98967445e-01 -1.56217873e+00 -3.68772484e-02 -2.00850874e-01
1.13427806e+00 3.70699912e-01 7.46416375e-02 6.58410639e-02
3.99170429e-01 -7.62433589e-01 9.89127934e-01 1.96356133e-01
5.52535236e-01 -7.27669239e-01 5.68722725e-01 -2.09163994e-01
-1.66892707e+00 1.84937209e-01 -2.82062918e-01 2.13169739e-01
1.78660512e-01 5.59483767e-01 -6.32721782e-01 7.85146236e-01
7.53968477e-01 1.27792037e+00 -9.73635793e-01 1.16099167e+00
-4.37469035e-01 9.14596498e-01 1.58599511e-01 -2.58303620e-02
5.01175106e-01 -3.23627237e-03 6.89990163e-01 1.48565912e+00
-1.18357576e-02 -3.63735855e-02 3.30714107e-01 1.20436168e+00
-2.65293866e-01 -1.39373122e-02 -2.59252876e-01 -2.29657233e-01
2.17572227e-01 1.23355615e+00 -1.43292665e+00 -5.86568296e-01
-4.41650271e-01 1.09414101e+00 4.17415231e-01 4.59709823e-01
-1.11568999e+00 -4.86569077e-01 4.54386055e-01 5.91218136e-02
8.91576827e-01 -9.51351300e-02 -2.56853282e-01 -1.15675473e+00
-2.38424733e-01 -6.18722618e-01 5.08504093e-01 -9.30218518e-01
-1.09760499e+00 5.99438548e-01 -1.79608658e-01 -9.13546026e-01
2.04621151e-01 -2.63688058e-01 -8.99042964e-01 6.26370430e-01
-1.73886144e+00 -1.40126693e+00 -6.10360205e-01 4.63435680e-01
8.28148663e-01 -1.06751665e-01 -7.91536793e-02 3.17343831e-01
-7.98414826e-01 3.63012135e-01 -6.28812164e-02 3.62132072e-01
3.79422188e-01 -1.43243074e+00 9.52813447e-01 1.54125702e+00
2.54023880e-01 2.71137983e-01 4.21139151e-01 -9.35336292e-01
-1.33485615e+00 -1.85983288e+00 7.92163432e-01 -2.51806200e-01
5.49490929e-01 -5.83699346e-01 -9.60141838e-01 7.87999749e-01
5.07928252e-01 3.05695415e-01 1.20276235e-01 -1.18305482e-01
-4.31133583e-02 2.68808156e-01 -5.62806368e-01 6.57919228e-01
1.28833640e+00 -4.62319762e-01 -4.07464206e-01 4.34302479e-01
1.17910886e+00 -5.00896871e-01 -3.53896797e-01 3.46057087e-01
-2.59865765e-02 -9.57785070e-01 9.19325054e-01 -3.35879505e-01
5.92751861e-01 -6.66666687e-01 -6.97953906e-03 -1.00171447e+00
-7.04017758e-01 -9.36095297e-01 -4.18190181e-01 1.17899823e+00
2.40904212e-01 -3.75378519e-01 9.06378269e-01 -5.96661270e-02
-6.55531526e-01 -7.05643177e-01 -4.93822157e-01 -4.10453796e-01
-3.76482189e-01 -3.08375537e-01 5.09214044e-01 8.05064797e-01
-4.60196853e-01 5.86712003e-01 -3.21849585e-01 5.28231442e-01
9.26008701e-01 5.11993408e-01 7.21417785e-01 -1.07690811e+00
-2.09133923e-01 -5.87988257e-01 -5.10665774e-01 -1.18255734e+00
1.18182801e-01 -1.04825127e+00 2.18083054e-01 -1.79622638e+00
5.08419752e-01 9.91787612e-02 -5.58082044e-01 3.33709419e-01
-8.92543972e-01 3.39057118e-01 4.02761936e-01 -2.24814960e-03
-1.50157869e+00 6.46398127e-01 1.25078142e+00 -3.39915305e-01
-1.49897143e-01 -3.57157499e-01 -1.01630569e+00 6.60865664e-01
6.11173391e-01 -3.81557643e-01 -4.59484190e-01 -5.84821761e-01
-7.88347945e-02 -3.64811897e-01 6.43730223e-01 -1.20401108e+00
2.83847779e-01 -8.27392656e-03 2.90110052e-01 -1.03022158e+00
-9.35343057e-02 -4.05989438e-01 -2.44582728e-01 2.49354497e-01
-2.26608366e-01 -1.05901159e-01 3.42310309e-01 1.01148307e+00
-4.92555648e-02 4.55009282e-01 8.25365424e-01 2.54924804e-01
-1.16408312e+00 3.71069074e-01 -3.12491298e-01 4.70752478e-01
1.09397161e+00 -7.84810185e-02 -4.70212638e-01 -2.89663196e-01
-4.60606813e-01 4.97378528e-01 3.51632446e-01 6.40385091e-01
7.21633017e-01 -1.26046014e+00 -8.04770470e-01 2.56441057e-01
2.80224383e-01 1.51474565e-01 2.32194096e-01 9.08750176e-01
-6.20029449e-01 3.32525402e-01 6.89157844e-02 -8.44891012e-01
-1.34664869e+00 7.20913708e-01 2.52638370e-01 -9.69066396e-02
-8.23144436e-01 1.10301793e+00 9.01427984e-01 1.92492545e-01
2.46059015e-01 -6.97637916e-01 -2.26042449e-01 -3.59437019e-01
3.49219739e-01 1.08367331e-01 -1.35260209e-01 -8.95213425e-01
-5.20887017e-01 5.51438868e-01 -3.83877195e-02 5.21314919e-01
1.20618701e+00 -3.14998448e-01 1.16592851e-02 9.59243178e-02
1.19531035e+00 -2.94471622e-01 -1.60792780e+00 -2.83687651e-01
8.62755403e-02 -5.91352999e-01 2.92096198e-01 -3.45886618e-01
-1.62345207e+00 8.53663206e-01 5.69700181e-01 2.23393172e-01
1.23544610e+00 3.81309927e-01 7.66648173e-01 -1.85632691e-01
-7.58242607e-02 -9.82318938e-01 2.72461861e-01 4.05623585e-01
8.38946223e-01 -1.38446474e+00 3.85369062e-02 -7.34564185e-01
-7.96092391e-01 6.40177727e-01 6.73674345e-01 -4.38129127e-01
5.52209377e-01 2.85706352e-02 -2.79567629e-01 -7.72269607e-01
-6.71473265e-01 -8.47012043e-01 7.58473694e-01 6.19278014e-01
3.79146814e-01 5.79215772e-02 -1.06239421e-02 5.63360810e-01
2.86543846e-01 -3.06649953e-01 2.81159997e-01 6.89895093e-01
-5.70826292e-01 -4.17453587e-01 -1.44125208e-01 3.79338741e-01
-6.13357484e-01 -2.28622288e-01 -7.79990733e-01 7.21010149e-01
8.78852233e-02 1.12298203e+00 2.98511744e-01 -4.06562299e-01
1.34784028e-01 -4.16718543e-01 8.22279975e-02 -7.40136623e-01
-4.89900440e-01 5.15483201e-01 -3.37051637e-02 -6.41216755e-01
-5.43991685e-01 -5.80519140e-01 -1.22526729e+00 1.69789746e-01
-5.02246916e-01 1.51065309e-02 1.56923085e-01 8.56402040e-01
7.67365992e-01 1.21965265e+00 4.27335739e-01 -1.23352432e+00
4.37502503e-01 -6.60615385e-01 -4.45415378e-01 4.16480422e-01
3.70712191e-01 -7.40791500e-01 -1.11073062e-01 2.95680463e-01] | [9.722726821899414, -0.2636004090309143] |
81ac36f2-9a6f-429e-b2b4-ae8efeeeaf4c | inr-v-a-continuous-representation-space-for | 2210.16579 | null | https://arxiv.org/abs/2210.16579v2 | https://arxiv.org/pdf/2210.16579v2.pdf | INR-V: A Continuous Representation Space for Video-based Generative Tasks | Generating videos is a complex task that is accomplished by generating a set of temporally coherent images frame-by-frame. This limits the expressivity of videos to only image-based operations on the individual video frames needing network designs to obtain temporally coherent trajectories in the underlying image space. We propose INR-V, a video representation network that learns a continuous space for video-based generative tasks. INR-V parameterizes videos using implicit neural representations (INRs), a multi-layered perceptron that predicts an RGB value for each input pixel location of the video. The INR is predicted using a meta-network which is a hypernetwork trained on neural representations of multiple video instances. Later, the meta-network can be sampled to generate diverse novel videos enabling many downstream video-based generative tasks. Interestingly, we find that conditional regularization and progressive weight initialization play a crucial role in obtaining INR-V. The representation space learned by INR-V is more expressive than an image space showcasing many interesting properties not possible with the existing works. For instance, INR-V can smoothly interpolate intermediate videos between known video instances (such as intermediate identities, expressions, and poses in face videos). It can also in-paint missing portions in videos to recover temporally coherent full videos. In this work, we evaluate the space learned by INR-V on diverse generative tasks such as video interpolation, novel video generation, video inversion, and video inpainting against the existing baselines. INR-V significantly outperforms the baselines on several of these demonstrated tasks, clearly showcasing the potential of the proposed representation space. | ['C. V. Jawahar', 'Vinay P Namboodiri', 'Aditya Agarwal', 'Bipasha Sen'] | 2022-10-29 | null | null | null | null | ['video-generation', 'video-inpainting'] | ['computer-vision', 'computer-vision'] | [ 4.67242509e-01 2.81437278e-01 -1.93092510e-01 -2.35312730e-01
-7.36288548e-01 -4.53660131e-01 6.36162758e-01 -9.51647997e-01
1.48937525e-02 1.00849140e+00 4.40928370e-01 8.31058398e-02
1.58189908e-01 -6.13180578e-01 -1.55775464e+00 -8.10591638e-01
-8.49703550e-02 1.53508157e-01 -3.06863517e-01 -1.32348984e-01
-1.52679145e-01 6.45615995e-01 -1.72918510e+00 7.35582530e-01
3.74241829e-01 9.08880532e-01 1.76020399e-01 8.30221176e-01
7.45136365e-02 1.09070933e+00 -4.86677855e-01 -2.74662971e-01
4.69269931e-01 -7.08627403e-01 -4.85955328e-01 3.78720790e-01
6.43221259e-01 -6.08320117e-01 -7.41240323e-01 6.56719685e-01
4.50530723e-02 3.81487638e-01 6.54636323e-01 -1.58516431e+00
-1.04344285e+00 5.64567626e-01 -4.45783794e-01 -3.53471458e-01
7.58534372e-01 4.27020311e-01 7.01977015e-01 -1.13397443e+00
1.15290201e+00 1.38002801e+00 6.33108437e-01 8.53868902e-01
-1.34400308e+00 -6.37837112e-01 7.69325048e-02 1.64184943e-01
-1.38657045e+00 -5.74740529e-01 8.42048109e-01 -4.15098697e-01
8.61250222e-01 3.54340583e-01 9.83599067e-01 1.64982152e+00
-4.41164756e-03 8.31041753e-01 6.99916184e-01 -2.34497651e-01
1.31139547e-01 -2.78066784e-01 -6.16213441e-01 7.67367899e-01
-1.94190785e-01 2.85539240e-01 -7.55499244e-01 1.50670439e-01
1.37527490e+00 3.84948999e-01 -6.74711704e-01 -2.05870688e-01
-1.36469579e+00 7.75033295e-01 4.67792451e-01 1.01030104e-01
-5.05805194e-01 6.21429741e-01 2.22307727e-01 3.23570579e-01
2.02941477e-01 4.05238628e-01 -1.96017563e-01 8.21894631e-02
-1.15830946e+00 3.78793925e-01 5.15561461e-01 1.19404387e+00
8.43348861e-01 6.31636918e-01 -4.72369671e-01 3.82693022e-01
3.80179062e-02 4.02267188e-01 4.62577879e-01 -1.59216416e+00
3.83590192e-01 9.91130397e-02 5.38068078e-02 -7.81221569e-01
7.48476461e-02 -6.33985549e-03 -9.74540114e-01 1.24963984e-01
3.27955753e-01 -2.73997575e-01 -9.84763324e-01 1.95525301e+00
1.81549415e-01 9.75758493e-01 2.15833038e-01 7.61440039e-01
9.44993258e-01 1.14160097e+00 -1.71057284e-01 -3.87497365e-01
7.76654482e-01 -9.83973801e-01 -6.86502814e-01 1.36666715e-01
7.83914700e-02 -3.79730821e-01 8.87863994e-01 2.51899093e-01
-1.46680522e+00 -8.43473017e-01 -7.48911679e-01 -1.97815239e-01
-3.57321426e-02 -9.26805567e-03 4.89968598e-01 9.66200307e-02
-1.44329906e+00 7.82172740e-01 -6.74369931e-01 3.64034399e-02
4.74192888e-01 2.96622396e-01 -8.73878896e-01 -1.04089141e-01
-1.00447273e+00 2.88687289e-01 4.74285990e-01 3.18692029e-01
-1.23892295e+00 -9.02808070e-01 -1.25727236e+00 -7.65266716e-02
2.18481690e-01 -1.10717905e+00 9.44735646e-01 -1.57643831e+00
-1.66695809e+00 6.83803082e-01 -2.84324914e-01 -5.84982932e-01
6.44613206e-01 -1.68552592e-01 -2.71817625e-01 4.90744084e-01
-6.25082757e-03 1.41343248e+00 1.49438214e+00 -1.27617729e+00
-2.51481920e-01 8.83022025e-02 9.85101834e-02 6.70226961e-02
3.04690134e-02 -2.89349735e-01 -7.15898573e-01 -9.97793555e-01
-1.71781406e-01 -1.06742573e+00 -1.35934263e-01 1.99485153e-01
-3.54088724e-01 -9.28736851e-03 1.12125576e+00 -8.51929247e-01
6.97760940e-01 -2.18073034e+00 6.35875285e-01 -8.18102583e-02
4.80457284e-02 1.10384166e-01 -4.98629153e-01 2.84247518e-01
-5.22451222e-01 4.05514389e-02 -1.06612422e-01 -4.04138565e-01
-1.63769722e-01 5.29118478e-01 -6.93818986e-01 3.41248870e-01
4.97119129e-01 1.29633939e+00 -8.17165375e-01 -3.17574918e-01
3.09909880e-01 1.06875598e+00 -9.42834377e-01 6.35402918e-01
-6.04923844e-01 8.98035526e-01 2.76059527e-02 6.02648616e-01
4.21176612e-01 -2.62881160e-01 5.98146990e-02 -4.97206807e-01
2.11887568e-01 -3.45592797e-01 -8.75722885e-01 2.07717228e+00
-4.18509930e-01 8.06934655e-01 -9.57979560e-02 -1.05778825e+00
6.52805269e-01 5.03727436e-01 5.78960121e-01 -3.54139835e-01
-6.32472849e-03 -1.32853016e-01 -5.60258746e-01 -7.27854788e-01
4.88304079e-01 -1.12997726e-01 1.34212643e-01 1.70190737e-01
3.35779548e-01 -6.37378916e-02 1.97983146e-01 2.80967295e-01
8.46738279e-01 9.60199356e-01 -8.82496014e-02 3.74590576e-01
2.46049091e-01 -3.60210270e-01 5.91912687e-01 5.65957487e-01
2.96510339e-01 9.48325336e-01 4.13910836e-01 -3.80860716e-01
-1.27544403e+00 -1.30913901e+00 2.20683858e-01 8.68814111e-01
-1.38026297e-01 -3.68207276e-01 -7.28086710e-01 -5.28879881e-01
-1.52453095e-01 6.79620385e-01 -8.42862189e-01 -2.02341750e-01
-7.80216992e-01 -1.45280659e-01 4.78442997e-01 6.75531566e-01
3.62655908e-01 -1.37761116e+00 -3.31426084e-01 1.69294059e-01
-4.75321770e-01 -1.27953553e+00 -5.85730970e-01 -1.96673438e-01
-8.35110724e-01 -1.00594234e+00 -9.36533332e-01 -9.18200076e-01
8.17925870e-01 2.09450796e-01 1.05763006e+00 -7.94539303e-02
-3.37856680e-01 6.41431868e-01 -2.92360365e-01 3.30654532e-01
-5.89065254e-01 -5.65793812e-01 8.80682189e-03 3.19499254e-01
-2.01723024e-01 -9.09132242e-01 -6.24135792e-01 1.79978721e-02
-1.23934352e+00 4.47042346e-01 4.64155704e-01 1.04852033e+00
7.51771569e-01 -3.97634715e-01 4.78126377e-01 -7.41271913e-01
1.63992599e-01 -6.15440667e-01 -3.80306363e-01 2.44689971e-01
3.33703846e-01 2.03463450e-01 7.53050745e-01 -7.05693901e-01
-9.96766806e-01 2.92400509e-01 -1.41671941e-01 -1.17357862e+00
-8.51098895e-02 1.23275407e-01 -7.22184703e-02 1.15065157e-01
4.91291344e-01 3.89887482e-01 6.34152368e-02 -4.19803709e-02
8.65310788e-01 1.16176568e-01 1.01528430e+00 -6.77068770e-01
9.04281735e-01 5.23827493e-01 1.87642798e-02 -9.72801507e-01
-5.91188669e-01 1.20163172e-01 -5.34251750e-01 -3.11653405e-01
1.05214250e+00 -1.25083745e+00 -6.21995270e-01 1.96601957e-01
-1.36939859e+00 -5.00952423e-01 -6.69233084e-01 3.64454538e-01
-9.87641692e-01 2.35730171e-01 -7.85279095e-01 -4.44606483e-01
-7.34562799e-02 -1.25567198e+00 1.26438928e+00 9.89579260e-02
-2.54102498e-01 -9.10903454e-01 -1.25277355e-01 3.25693130e-01
5.76624013e-02 7.70442247e-01 5.90217352e-01 1.10439241e-01
-9.84288692e-01 1.53793767e-01 1.20307572e-01 4.80875552e-01
6.49886578e-03 3.85566711e-01 -9.39661860e-01 -3.79618764e-01
-2.57434398e-01 -5.30260324e-01 1.00186753e+00 5.82223475e-01
1.60437179e+00 -7.78661728e-01 -1.29180565e-01 1.19784212e+00
1.22670102e+00 1.94810688e-01 1.00313783e+00 -1.11689880e-01
8.89675319e-01 4.10500079e-01 1.77041769e-01 3.27675462e-01
4.35776375e-02 6.26801789e-01 5.55796504e-01 -1.19582571e-01
-3.90353233e-01 -5.89037776e-01 8.65584493e-01 5.90176702e-01
-4.70008433e-01 -1.55264497e-01 -1.63639367e-01 3.28492463e-01
-1.92086041e+00 -1.50913882e+00 3.45906287e-01 1.89601350e+00
8.00987780e-01 -3.12030822e-01 1.38246894e-01 -1.84201345e-01
8.00464928e-01 1.61842689e-01 -5.93654335e-01 -9.11582187e-02
-2.09989905e-01 4.08623338e-01 4.64279726e-02 4.11918402e-01
-8.86015534e-01 9.10933793e-01 6.44548559e+00 6.00160360e-01
-1.14090645e+00 7.93708861e-03 7.29542851e-01 -4.03223336e-01
-5.56118309e-01 -1.31113723e-01 -5.22571743e-01 3.64864409e-01
7.44333744e-01 3.05228941e-02 7.80212283e-01 7.99132526e-01
2.98406571e-01 4.10233706e-01 -1.44316876e+00 1.33927977e+00
3.52597624e-01 -2.01659322e+00 5.55701554e-01 -1.99694857e-01
1.12100124e+00 -3.39759767e-01 3.51955384e-01 4.26533878e-01
2.29582846e-01 -1.29707336e+00 9.07725990e-01 7.80535579e-01
1.27579284e+00 -6.30045295e-01 9.36945528e-02 5.16814552e-02
-9.55642581e-01 -4.94420566e-02 -2.93066442e-01 1.66952819e-01
4.12884772e-01 1.63769662e-01 -7.09165871e-01 4.06370670e-01
6.28687859e-01 1.06909144e+00 -2.32651979e-01 6.52385056e-01
-3.24533373e-01 4.28778529e-01 -4.27047126e-02 5.62431872e-01
2.69604027e-01 -4.06049281e-01 4.87625688e-01 1.12185836e+00
7.08037436e-01 3.85079905e-02 9.87230316e-02 1.06730199e+00
-5.10983706e-01 -3.76776814e-01 -1.08466029e+00 -1.18285231e-01
2.51812607e-01 1.18409419e+00 -3.01537722e-01 -3.82521451e-01
-3.22878957e-01 1.26543558e+00 2.09673643e-01 8.38582635e-01
-1.30117762e+00 -3.58765051e-02 7.03214586e-01 2.64797688e-01
5.20372272e-01 -3.04863781e-01 4.95199144e-01 -1.36101067e+00
-2.06158366e-02 -9.62681055e-01 7.24287331e-02 -1.34613955e+00
-1.00456548e+00 7.75360763e-01 1.13662437e-01 -1.37909091e+00
-9.28472400e-01 -5.01091421e-01 -5.45568466e-01 4.16372836e-01
-1.06273961e+00 -1.36581481e+00 -4.60833967e-01 1.12388492e+00
8.01637888e-01 -3.15433770e-01 6.93447411e-01 9.50396582e-02
-3.80607009e-01 4.70289886e-01 -1.20809101e-01 2.15089604e-01
5.47393024e-01 -8.10377240e-01 3.72397482e-01 9.16146517e-01
4.94934410e-01 5.14891684e-01 5.62938154e-01 -4.92085278e-01
-1.63853955e+00 -1.48799872e+00 1.46049961e-01 -2.30743691e-01
4.30350840e-01 -3.12655926e-01 -7.80990541e-01 1.22520423e+00
3.66475612e-01 3.85377795e-01 4.75251406e-01 -5.55744231e-01
-4.88678545e-01 -6.61809519e-02 -9.65373516e-01 8.42629731e-01
1.35325205e+00 -6.85842335e-01 -3.86626869e-01 4.90143955e-01
1.07038426e+00 -6.13032997e-01 -8.19738209e-01 2.51417607e-01
5.28647184e-01 -1.08707297e+00 1.36951995e+00 -8.73287082e-01
8.71464431e-01 -3.55948925e-01 -2.20788121e-01 -1.22278762e+00
-1.37385339e-01 -1.08747458e+00 -3.93577635e-01 1.13116515e+00
1.68895259e-01 -1.98955581e-01 8.58937263e-01 5.24609029e-01
-2.08129317e-01 -6.08392239e-01 -7.18578875e-01 -6.31456792e-01
-1.73387140e-01 -4.28551972e-01 6.42254591e-01 9.35341001e-01
-4.65684891e-01 5.42485304e-02 -8.36325824e-01 -5.77327348e-02
4.88546073e-01 7.00305179e-02 8.72828305e-01 -5.63824892e-01
-6.87899232e-01 -1.04880430e-01 -3.59102726e-01 -1.17707336e+00
7.21838951e-01 -9.49106038e-01 -9.83829871e-02 -1.19803703e+00
-1.36934351e-02 -6.29672036e-02 1.74561501e-01 5.81446350e-01
1.34904072e-01 5.70934832e-01 3.74339849e-01 1.21558696e-01
-4.09401923e-01 7.56902397e-01 1.55095387e+00 -2.67949283e-01
-1.63498372e-01 -3.35222244e-01 -4.86349225e-01 6.93305373e-01
4.94261384e-01 -8.17216486e-02 -6.74753785e-01 -3.82671416e-01
2.13386208e-01 6.96086526e-01 7.13281155e-01 -9.56763148e-01
-5.53006493e-02 -4.45212811e-01 9.17318344e-01 -3.09643835e-01
7.95812130e-01 -6.77601397e-01 8.54564250e-01 2.15537250e-01
-3.76202106e-01 1.23787522e-01 -2.42257938e-02 6.87806666e-01
-3.39761168e-01 5.94841316e-02 5.95118761e-01 -3.64018619e-01
-9.07317460e-01 6.43900514e-01 -3.85094471e-02 7.64628127e-02
1.21202433e+00 -5.33416152e-01 -6.61491528e-02 -6.88349247e-01
-9.61877942e-01 -2.56692290e-01 3.95999849e-01 6.04362726e-01
1.02738631e+00 -1.69988775e+00 -6.82656050e-01 6.02523685e-01
-2.16330215e-01 1.87834591e-01 3.36589605e-01 5.84491968e-01
-6.73561871e-01 1.28615769e-02 -5.10670483e-01 -8.02121580e-01
-1.08253133e+00 6.79755509e-01 1.83999404e-01 1.13829352e-01
-9.45531905e-01 8.76243293e-01 4.22930539e-01 -1.05806887e-02
1.06375173e-01 -3.27770859e-01 3.26208435e-02 -2.67527997e-01
5.71432650e-01 2.68471204e-02 -4.72803116e-01 -8.34227741e-01
1.43423721e-01 5.00780344e-01 1.52369276e-01 -1.96091488e-01
1.50152636e+00 1.43174067e-01 -1.18951842e-01 1.96404532e-01
1.47756624e+00 -1.78828061e-01 -1.97473514e+00 1.59588471e-01
-6.25591278e-01 -4.94358748e-01 -3.80763382e-01 -2.34201297e-01
-1.33253777e+00 6.33283913e-01 4.63690609e-02 -3.34962279e-01
1.18089736e+00 -1.06950633e-01 8.29495847e-01 3.01190227e-01
3.35510790e-01 -8.00376654e-01 6.18781924e-01 3.27013791e-01
1.34378445e+00 -8.74692380e-01 -2.51005024e-01 -2.77730703e-01
-7.69736826e-01 1.21894658e+00 4.71838355e-01 -3.71317774e-01
3.59977752e-01 2.25921631e-01 -2.76622087e-01 2.26668522e-01
-8.26356828e-01 1.92987517e-01 3.36794525e-01 8.28096569e-01
3.78399372e-01 -1.90516844e-01 3.04257005e-01 2.10340634e-01
-1.87892914e-01 1.59634009e-01 5.31575322e-01 5.73403955e-01
-2.27837395e-02 -1.00000954e+00 -4.52009141e-01 1.69207022e-01
-1.41797617e-01 1.35806613e-02 -3.85370441e-02 8.46491814e-01
2.33327240e-01 5.65167189e-01 2.04617575e-01 -3.04719031e-01
-1.79507628e-01 4.51709144e-02 9.30108964e-01 -4.57902819e-01
-2.51117587e-01 1.15929887e-01 -1.48597226e-01 -8.83381844e-01
-6.77536786e-01 -5.72325706e-01 -1.18328488e+00 -2.67076790e-01
1.72228575e-01 -2.27661505e-02 2.89849639e-01 7.09613800e-01
3.84091347e-01 5.61529100e-01 4.71382946e-01 -1.50293398e+00
-1.51361734e-01 -5.21883607e-01 -4.65604901e-01 8.61663163e-01
4.82817471e-01 -5.65753281e-01 -2.86992788e-01 7.19422817e-01] | [10.839298248291016, -0.6124202013015747] |
2cee712a-a18e-4c71-ba34-dbbf3caa4877 | artificial-life-sustainable-self-replicating | 2105.13971 | null | https://arxiv.org/abs/2105.13971v2 | https://arxiv.org/pdf/2105.13971v2.pdf | Artificial life: sustainable self-replicating systems | Nature has found one method of organizing living matter, but maybe other options exist -- not yet discovered -- on how to create life. To study the life "as it could be" is the objective of an interdisciplinary field called Artificial Life (commonly abbreviated as ALife). The word "artificial" refers to the fact that humans are involved in the creation process. The artificial life forms might be completely unlike natural forms of life, with different chemical compositions, and even computer programs exhibiting life-like behaviours. | ['Jitka Cejkova', 'Carlos Gershenson'] | 2021-05-27 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 7.83925131e-02 2.56007165e-01 1.77358299e-01 2.43144289e-01
7.53682256e-01 -8.01762223e-01 8.76054645e-01 -2.64723450e-01
-1.58587113e-01 1.08231640e+00 1.51956543e-01 -4.13731545e-01
2.48619169e-01 -1.06661475e+00 -4.70138401e-01 -1.16397488e+00
1.63162380e-01 3.18058759e-01 4.48050201e-02 -6.83106601e-01
5.71009576e-01 4.16037351e-01 -1.70987546e+00 -2.57822961e-01
9.00502861e-01 1.18974954e-01 3.39034170e-01 7.14495540e-01
-5.09147346e-01 7.81211138e-01 -6.11888647e-01 -1.99706748e-01
1.45034358e-01 -1.11914861e+00 -8.47887695e-01 -7.71362260e-02
-8.15230727e-01 2.80291706e-01 -2.51124233e-01 8.27806175e-01
6.45688474e-02 -7.71243647e-02 6.71124578e-01 -7.88986087e-01
-1.04612517e+00 4.69151527e-01 9.63744819e-02 -2.57674545e-01
6.67954504e-01 3.95827770e-01 3.08405161e-01 -5.18009663e-01
7.41986930e-01 1.16548538e+00 5.54381907e-01 9.97816026e-01
-1.28693056e+00 -5.63505553e-02 -6.20870471e-01 -2.58720726e-01
-1.14266539e+00 -4.44993079e-01 3.84589732e-01 -7.51091003e-01
9.42844510e-01 4.16689217e-01 1.21289432e+00 7.35402107e-01
1.06562781e+00 2.16935560e-01 1.19756782e+00 -6.48160398e-01
4.84214425e-01 3.69547993e-01 -3.84425297e-02 4.70702827e-01
9.60568190e-01 3.33623886e-01 -4.89545673e-01 -1.95839003e-01
7.18017876e-01 -3.42899710e-01 -1.96301654e-01 4.19016518e-02
-1.49989104e+00 4.39788789e-01 5.32464832e-02 1.03677690e+00
-3.28376383e-01 5.92048764e-01 1.22922480e-01 3.14142644e-01
1.69931471e-01 1.11342394e+00 -4.96002674e-01 -3.74158919e-01
-2.73086727e-01 -2.72942279e-02 1.35259962e+00 5.82563162e-01
7.23937869e-01 2.40248874e-01 5.71344078e-01 4.16825891e-01
2.54986703e-01 2.53181666e-01 8.42519581e-01 -8.97328675e-01
-6.66820109e-01 8.26090813e-01 1.88845187e-01 -4.45660830e-01
-2.96584517e-01 -1.85994297e-01 -6.88496113e-01 3.71235549e-01
3.64093065e-01 -2.22486824e-01 -7.32273579e-01 1.23001182e+00
3.86017203e-01 -3.18809360e-01 2.61920929e-01 4.36077774e-01
8.06887507e-01 9.44007576e-01 1.02738149e-01 -4.32581842e-01
1.18471539e+00 -7.36078143e-01 -8.99390101e-01 8.47937465e-02
5.48998177e-01 -4.88481015e-01 1.00959206e+00 3.66239667e-01
-9.36864436e-01 -1.08533286e-01 -1.26786089e+00 1.55194536e-01
-9.34487760e-01 -6.80037320e-01 1.15842474e+00 9.74272370e-01
-1.14553642e+00 6.25375271e-01 -4.26818043e-01 -7.81221628e-01
-2.40208223e-01 5.39165437e-02 -2.45137811e-01 4.91032988e-01
-8.88940632e-01 1.25828946e+00 6.13054812e-01 -1.95053935e-01
-9.45625782e-01 -3.98816854e-01 -2.94228703e-01 -3.16630602e-01
2.41678208e-01 -1.22832704e+00 9.42791462e-01 -1.09740520e+00
-1.47810376e+00 1.23964846e+00 -3.45533013e-01 -1.74922451e-01
2.16129541e-01 2.08704099e-01 -4.75669891e-01 -8.47876221e-02
-1.18728973e-01 4.06060755e-01 4.19630378e-01 -1.67585433e+00
-1.65231168e-01 -2.70152688e-01 1.71546444e-01 -5.37070110e-02
-2.22572535e-01 2.39517063e-01 2.37264395e-01 -7.95729935e-01
3.55323792e-01 -9.92892861e-01 -6.71573222e-01 7.84738660e-02
-1.02937192e-01 -3.96464974e-01 5.94079733e-01 -2.63793319e-01
9.79623020e-01 -2.00654984e+00 1.61492661e-01 6.98675448e-03
4.08207089e-01 3.31156626e-02 3.59662324e-01 7.56289780e-01
8.53204653e-02 6.30293965e-01 -4.62144285e-01 5.40168464e-01
7.23576471e-02 3.66490275e-01 1.81769297e-01 3.07876289e-01
-1.26533747e-01 9.41039681e-01 -1.02170932e+00 -4.15528774e-01
3.69736440e-02 3.56915087e-01 1.21517584e-01 -8.81969258e-02
-5.73503733e-01 6.87127888e-01 -4.46640998e-01 8.77971292e-01
1.70352578e-01 -2.18537927e-01 2.76542187e-01 5.31080544e-01
-5.74571192e-01 -4.83481735e-02 -4.93952662e-01 1.37472618e+00
-5.79637848e-02 7.63349116e-01 1.29792243e-01 -7.43213892e-01
9.81806636e-01 7.13908672e-01 5.34291148e-01 -6.12044811e-01
1.60846889e-01 5.70673466e-01 3.54522556e-01 -9.94568467e-01
3.24985296e-01 -3.94780546e-01 6.97628269e-03 4.51705068e-01
-3.67495626e-01 -6.52868867e-01 4.13472593e-01 -1.86396673e-01
1.26713395e+00 4.76506293e-01 6.91972852e-01 -8.35523844e-01
5.97552776e-01 3.41031998e-01 6.31397486e-01 4.84128565e-01
-2.54431754e-01 1.71275944e-01 1.84988022e-01 -8.25472713e-01
-1.19643068e+00 -9.69165027e-01 -3.58252943e-01 9.30321157e-01
3.58323216e-01 -2.30934083e-01 -8.86593938e-01 1.00857809e-01
-1.41950861e-01 7.69255102e-01 -7.24341035e-01 6.66524842e-02
-4.65483129e-01 -8.56656194e-01 3.33290726e-01 -1.89637586e-01
5.27899742e-01 -1.58524883e+00 -8.71618509e-01 7.04977095e-01
2.00500205e-01 -4.88597661e-01 3.66870731e-01 4.25172925e-01
-8.49303067e-01 -5.52377939e-01 -5.13394654e-01 -5.38869679e-01
7.46602476e-01 8.92917514e-02 1.37138212e+00 7.33229160e-01
-5.00473499e-01 1.73921794e-01 -5.72546601e-01 -5.91224730e-01
-8.47506166e-01 -2.31343374e-01 8.99201855e-02 -5.00696361e-01
2.13901520e-01 -8.90805423e-01 -4.90959674e-01 -4.43740785e-02
-1.13072944e+00 1.52243719e-01 5.25030971e-01 6.17689192e-01
1.20979816e-01 2.36645862e-01 9.58314121e-01 -9.13394988e-01
4.94037211e-01 -4.83433455e-01 7.48518035e-02 5.14646590e-01
-6.12755895e-01 1.42349079e-01 7.08743274e-01 -1.72664285e-01
-7.27370381e-01 -2.85808425e-02 -8.54468420e-02 6.61400616e-01
-4.46999669e-01 2.25947797e-01 -6.03762627e-01 -3.55067700e-01
6.33572996e-01 4.61755455e-01 3.91164012e-02 -4.00921583e-01
1.78311676e-01 5.45526564e-01 4.88309979e-01 -8.22831154e-01
6.63109124e-01 4.86848295e-01 8.77747685e-02 -1.24989939e+00
-1.16812579e-01 5.66404089e-02 -3.60829979e-01 -5.94693065e-01
1.11446738e+00 -2.03102291e-01 -3.72522175e-01 3.91927749e-01
-1.21607113e+00 -1.90094262e-01 -6.55515254e-01 4.09565896e-01
-6.16625786e-01 -5.30411210e-03 -4.24003810e-01 -1.00364745e+00
-3.46060097e-01 -7.64099479e-01 3.44393253e-01 7.13526666e-01
-4.49603885e-01 -1.11929941e+00 4.94106144e-01 2.73097217e-01
6.81383014e-01 5.26541650e-01 1.01735115e+00 -1.45291850e-01
-4.57207918e-01 -2.80887038e-02 3.37788850e-01 -5.79711162e-02
3.92982632e-01 6.35501206e-01 -6.92885578e-01 2.56978363e-01
3.32818359e-01 3.62144977e-01 3.62010896e-01 4.42809798e-02
6.04900122e-01 -4.66805369e-01 -4.32640791e-01 1.60134807e-01
1.81522357e+00 1.00918508e+00 1.16911197e+00 6.20012343e-01
3.54243875e-01 1.13992095e+00 3.20855618e-01 5.50347149e-01
-7.76244029e-02 7.72329345e-02 8.66646692e-02 9.72962826e-02
3.46391648e-02 3.35321240e-02 3.51800531e-01 1.19406760e+00
-3.64926398e-01 -3.01718771e-01 -1.05130804e+00 5.28406143e-01
-1.58504152e+00 -9.66675222e-01 -5.74705839e-01 1.86483252e+00
9.03515816e-01 7.28515387e-02 9.48728472e-02 8.85225534e-02
6.69298828e-01 -4.20188248e-01 -7.06677496e-01 -5.84339261e-01
-4.36025441e-01 3.22566986e-01 3.01778078e-01 8.33719000e-02
-6.43700242e-01 8.13339651e-01 8.20426464e+00 4.13506657e-01
-9.74602580e-01 -5.63412160e-02 2.29089200e-01 3.49273473e-01
-7.26121306e-01 6.80594027e-01 -2.28649855e-01 7.05597639e-01
1.10078239e+00 -6.46007061e-01 3.43592733e-01 4.95217174e-01
5.69089830e-01 -3.79723877e-01 -8.04009616e-01 3.67238879e-01
-2.32121810e-01 -1.34264648e+00 6.47542849e-02 3.09416443e-01
7.53534734e-01 -6.12659812e-01 -7.10812986e-01 -2.26022407e-01
3.28063607e-01 -1.12918866e+00 9.85269964e-01 1.04745603e+00
4.67964649e-01 -4.55449760e-01 3.13267380e-01 5.32531798e-01
-1.03216577e+00 -7.14656105e-03 -2.53643513e-01 -4.39178050e-01
2.94067621e-01 7.20978200e-01 -2.74918526e-01 3.28138024e-01
7.61288047e-01 2.10828528e-01 -6.67960107e-01 1.31536591e+00
-4.47608568e-02 5.09462655e-01 -8.50468874e-02 -6.98661864e-01
-1.29322559e-01 -7.54019856e-01 7.10978985e-01 7.56651402e-01
3.66554558e-01 4.57392722e-01 -3.33992809e-01 9.18915093e-01
4.89618123e-01 1.42483518e-01 -1.06284106e+00 -8.47834885e-01
3.73341113e-01 9.69433308e-01 -1.17644083e+00 -3.75442505e-01
-1.26881480e-01 5.89378357e-01 -7.57202029e-01 7.11348355e-02
-6.12099349e-01 -5.53515553e-01 5.48559070e-01 5.79495430e-01
-3.79248887e-01 -5.62280536e-01 -6.37809336e-01 -9.00023401e-01
-4.62805986e-01 -9.64691281e-01 -3.92696053e-01 -8.06572556e-01
-1.03030956e+00 4.21377838e-01 -3.35862726e-01 -8.18305552e-01
-1.31822086e-03 -4.60373163e-01 -7.55604148e-01 3.55468810e-01
-3.47632498e-01 -9.65215087e-01 -2.31869712e-01 -5.50583750e-02
2.38296479e-01 -1.96292087e-01 8.31586421e-01 -1.97826937e-01
-4.55639035e-01 -9.91989598e-02 5.64203024e-01 -5.69185913e-01
2.23548606e-01 -1.03619182e+00 1.77428514e-01 6.03969097e-01
-3.89812976e-01 7.33564615e-01 1.20138073e+00 -7.35522211e-01
-1.68779564e+00 -2.17356682e-01 1.00593352e+00 -3.46269757e-01
4.04567003e-01 -2.48732284e-01 -7.05064178e-01 1.30940333e-01
6.92532897e-01 -9.19182122e-01 8.51962805e-01 -3.94707590e-01
1.67628825e-01 1.62510499e-01 -1.51583087e+00 9.31293786e-01
1.24717653e+00 -1.82507679e-01 -6.43602848e-01 4.47699577e-01
1.25674188e+00 5.66218853e-01 -8.90916884e-01 3.04046065e-01
9.21535254e-01 -8.79119158e-01 7.59872556e-01 -3.92018795e-01
3.94807488e-01 -7.14426160e-01 1.40405938e-01 -7.86628127e-01
-4.66386139e-01 -9.43595469e-01 7.28130862e-02 1.24662709e+00
4.59976763e-01 -1.11759675e+00 6.78212166e-01 8.06430161e-01
-2.34603956e-01 -5.98000526e-01 -7.21269369e-01 -8.40762079e-01
2.30219066e-01 3.12536091e-01 9.52279747e-01 1.23708534e+00
2.68753767e-01 2.18614653e-01 1.17362596e-01 -6.68747365e-01
6.24317944e-01 -7.26617724e-02 5.50862551e-01 -1.51677632e+00
-3.41528475e-01 -5.63484609e-01 -3.11928719e-01 -2.37996355e-01
-5.92914894e-02 -4.76940632e-01 1.92279145e-01 -1.51438642e+00
-1.19807143e-02 -3.08060050e-01 1.89058840e-01 4.80880290e-02
1.43033519e-01 -7.34603405e-02 4.42895219e-02 4.57713872e-01
-4.14278824e-03 3.83136272e-01 1.36501503e+00 1.49319530e-01
-2.84431159e-01 -5.37863255e-01 -8.73669982e-01 7.75753677e-01
1.02198923e+00 -3.63738447e-01 -2.57659584e-01 1.44222602e-01
7.20794678e-01 1.28593609e-01 5.02615459e-02 -1.38296449e+00
9.95831564e-02 -8.78981948e-01 -9.98552144e-02 -3.22191745e-01
-4.38408069e-02 -7.41324544e-01 1.35584795e+00 8.59842598e-01
3.97510976e-02 -8.46006498e-02 -4.40172344e-01 2.39407256e-01
2.20409930e-01 -8.80339742e-01 7.61291742e-01 -8.07401717e-01
-6.42205417e-01 -3.77932668e-01 -1.22440970e+00 -3.67299527e-01
1.49091065e+00 -9.29179728e-01 -5.98571777e-01 -1.87452808e-02
-5.65852165e-01 -1.64959595e-01 1.37485147e+00 2.46560369e-02
2.78071254e-01 -1.06093895e+00 -4.99943316e-01 -2.72919506e-01
9.49768629e-03 -3.38117361e-01 -3.20772290e-01 3.18120480e-01
-1.45361805e+00 2.74860919e-01 -4.25704300e-01 9.11376700e-02
-6.40522659e-01 7.03124285e-01 5.32167673e-01 2.27898106e-01
-5.16490281e-01 6.63142025e-01 3.46518457e-01 -4.35896248e-01
-2.93163091e-01 3.23683470e-01 -2.02598155e-01 -2.03547239e-01
2.57311791e-01 4.48601454e-01 -3.86162609e-01 -4.51435268e-01
-3.60447645e-01 5.90077221e-01 6.43326938e-01 -2.86760122e-01
1.44668603e+00 -1.88452810e-01 -1.28785574e+00 9.58441436e-01
4.77589935e-01 -2.62564421e-01 -4.30376559e-01 5.74685991e-01
1.95531130e-01 -2.52497017e-01 -7.00115681e-01 -6.16521418e-01
-4.61746454e-01 3.60086560e-01 4.01132911e-01 7.66538680e-01
8.93657625e-01 -1.71216056e-02 8.07165623e-01 4.67223883e-01
8.62596333e-01 -1.33075213e+00 2.40031239e-02 3.83272558e-01
7.80709743e-01 -5.93885124e-01 -1.04378805e-01 -2.25267738e-01
-2.56951749e-02 9.31485176e-01 5.66032648e-01 -1.82413295e-01
9.22291279e-01 5.22371292e-01 -2.71293253e-01 -2.60075539e-01
-1.01027024e+00 -2.01636195e-01 -3.79031420e-01 7.67966747e-01
9.91171658e-01 2.28375018e-01 -1.47383142e+00 2.62513638e-01
-4.29383487e-01 1.90721020e-01 7.99619257e-01 1.45631003e+00
-9.61140633e-01 -1.43179595e+00 -7.85690725e-01 4.18600827e-01
-4.53787774e-01 3.26233834e-01 -1.06442046e+00 3.80082905e-01
7.77193606e-01 1.18999040e+00 -2.74155498e-01 -5.03461659e-01
-1.45729870e-01 1.73056141e-01 3.41480911e-01 -3.56959611e-01
-9.92721975e-01 -4.88670886e-01 3.22673321e-01 -1.41295671e-01
-8.12297940e-01 -5.00571966e-01 -1.53371823e+00 -1.00622404e+00
-2.45016962e-01 5.28526783e-01 1.00694573e+00 9.87743139e-01
3.19669433e-02 3.04524839e-01 5.09883881e-01 -8.28477085e-01
1.96906000e-01 -5.74962497e-01 -7.91988134e-01 -2.36098319e-02
-1.89663112e-01 -3.42761338e-01 -5.67330480e-01 4.23710585e-01] | [5.593796253204346, 4.181454658508301] |
e7b1b019-34a2-4ddb-ae54-d25ba2d93068 | robust-one-shot-segmentation-of-brain-tissues | 2211.14521 | null | https://arxiv.org/abs/2211.14521v3 | https://arxiv.org/pdf/2211.14521v3.pdf | Robust One-shot Segmentation of Brain Tissues via Image-aligned Style Transformation | One-shot segmentation of brain tissues is typically a dual-model iterative learning: a registration model (reg-model) warps a carefully-labeled atlas onto unlabeled images to initialize their pseudo masks for training a segmentation model (seg-model); the seg-model revises the pseudo masks to enhance the reg-model for a better warping in the next iteration. However, there is a key weakness in such dual-model iteration that the spatial misalignment inevitably caused by the reg-model could misguide the seg-model, which makes it converge on an inferior segmentation performance eventually. In this paper, we propose a novel image-aligned style transformation to reinforce the dual-model iterative learning for robust one-shot segmentation of brain tissues. Specifically, we first utilize the reg-model to warp the atlas onto an unlabeled image, and then employ the Fourier-based amplitude exchange with perturbation to transplant the style of the unlabeled image into the aligned atlas. This allows the subsequent seg-model to learn on the aligned and style-transferred copies of the atlas instead of unlabeled images, which naturally guarantees the correct spatial correspondence of an image-mask training pair, without sacrificing the diversity of intensity patterns carried by the unlabeled images. Furthermore, we introduce a feature-aware content consistency in addition to the image-level similarity to constrain the reg-model for a promising initialization, which avoids the collapse of image-aligned style transformation in the first iteration. Experimental results on two public datasets demonstrate 1) a competitive segmentation performance of our method compared to the fully-supervised method, and 2) a superior performance over other state-of-the-art with an increase of average Dice by up to 4.67%. The source code is available at: https://github.com/JinxLv/One-shot-segmentation-via-IST. | ['Qiang Li', 'Zhiwei Wang', 'Ran Duan', 'Sheng Wang', 'Xiaoyu Zeng', 'Jinxin Lv'] | 2022-11-26 | null | null | null | null | ['one-shot-segmentation'] | ['computer-vision'] | [ 3.09902877e-01 2.68019050e-01 -6.25376264e-03 -4.14334893e-01
-8.99869025e-01 -5.56685269e-01 3.86658996e-01 -2.94595003e-01
-4.32909757e-01 5.21389127e-01 -2.91420016e-02 1.78900547e-02
5.03225103e-02 -4.93670046e-01 -7.09318995e-01 -1.14096534e+00
2.78900564e-01 4.22035038e-01 5.11038363e-01 4.75317519e-03
6.88240230e-02 4.74222541e-01 -9.34219956e-01 3.03794127e-02
1.05582607e+00 5.54202974e-01 4.70899671e-01 3.17760408e-01
-2.58270353e-01 3.52739811e-01 -1.57692492e-01 -3.87004882e-01
6.15999520e-01 -6.96640134e-01 -9.96259034e-01 3.38877410e-01
2.65442193e-01 -2.21120089e-01 -1.71476662e-01 1.43936455e+00
2.80421019e-01 -2.14355160e-03 6.48638546e-01 -1.06882560e+00
-4.54835266e-01 4.70679373e-01 -1.08357561e+00 1.17992595e-01
-1.42533481e-01 3.17125678e-01 5.23343444e-01 -7.60735273e-01
8.07524502e-01 8.36518049e-01 4.40949440e-01 6.86586916e-01
-1.44949400e+00 -7.50047505e-01 8.23203474e-02 -1.40540734e-01
-1.44546115e+00 -2.96207875e-01 8.60482335e-01 -5.50726295e-01
1.91243395e-01 1.15172416e-01 6.63156569e-01 7.30262637e-01
2.85604388e-01 5.63486040e-01 1.39999092e+00 -2.52145618e-01
1.83259219e-01 -4.03772704e-02 1.98733896e-01 8.69570851e-01
-4.73424532e-02 1.59865335e-01 -4.58060689e-02 -1.69035885e-02
1.07659554e+00 5.10761626e-02 -4.03828472e-01 -4.45728928e-01
-1.37450981e+00 4.96413618e-01 7.08774388e-01 6.25906169e-01
-2.79348075e-01 -3.02413791e-01 7.64918253e-02 6.82959482e-02
4.21530217e-01 3.37050796e-01 -9.22992900e-02 2.64645249e-01
-1.39478493e+00 -2.21064582e-01 4.25603867e-01 7.95155525e-01
1.02334630e+00 -6.24457337e-02 -1.65247679e-01 8.07078660e-01
1.81622833e-01 2.26479203e-01 6.23212337e-01 -9.92802858e-01
1.02454588e-01 5.29607177e-01 -2.90080339e-01 -5.12693405e-01
-3.99182111e-01 -5.22356272e-01 -1.01951742e+00 2.64063567e-01
8.03803623e-01 -1.48138061e-01 -1.28510702e+00 1.82462442e+00
5.37520409e-01 6.48418128e-01 -2.50213027e-01 1.05583239e+00
6.05906785e-01 4.37043399e-01 -1.45423278e-01 -4.86898452e-01
1.27317595e+00 -1.06424022e+00 -5.43631196e-01 -2.40095913e-01
7.51935840e-01 -8.10312867e-01 1.06611812e+00 8.31406489e-02
-1.11246061e+00 -4.88446474e-01 -1.04564333e+00 2.17036784e-01
2.76166461e-02 -1.56105310e-01 1.22684546e-01 5.02436876e-01
-1.06736636e+00 5.50236821e-01 -1.14921427e+00 -3.19133431e-01
5.87647200e-01 3.58399779e-01 -4.13433313e-01 -4.90173213e-02
-7.23092675e-01 7.42097259e-01 4.36002076e-01 -2.31439136e-02
-6.85165465e-01 -1.08329356e+00 -6.33899868e-01 -3.49173278e-01
2.40276098e-01 -5.14747679e-01 8.07771385e-01 -1.19180882e+00
-1.59567666e+00 1.21573400e+00 -1.45533979e-01 -5.96280098e-02
7.22557366e-01 2.90258616e-01 -1.74904332e-01 2.91243434e-01
2.97833800e-01 9.86682594e-01 8.02163064e-01 -1.53259706e+00
-1.59928441e-01 -3.87879908e-01 -3.11334372e-01 1.94526494e-01
-4.13347110e-02 -3.97934318e-02 -7.29621112e-01 -8.35273623e-01
4.59102690e-01 -1.19729578e+00 -3.84592652e-01 1.54674768e-01
-5.29475152e-01 4.95329410e-01 7.33073771e-01 -5.81708252e-01
9.11458671e-01 -2.31745005e+00 1.85579687e-01 3.64451855e-01
2.58126169e-01 1.73555598e-01 -3.95268828e-01 -7.31272772e-02
-4.00974691e-01 -5.82245290e-02 -9.01907325e-01 -3.53097081e-01
-6.87174618e-01 3.12449485e-01 -9.03911367e-02 7.63358712e-01
-5.39289974e-03 1.05942357e+00 -9.50242817e-01 -7.06993878e-01
1.47037581e-01 3.24174106e-01 -4.80774760e-01 3.17302525e-01
1.10222243e-01 1.21971703e+00 -4.16581571e-01 4.38114941e-01
8.56064677e-01 -3.46451253e-01 2.01873317e-01 -4.53652054e-01
-1.71122402e-01 -2.64423698e-01 -1.03430033e+00 1.95483530e+00
3.78114805e-02 3.43310386e-02 1.03691868e-01 -1.03530955e+00
7.60537922e-01 3.34389746e-01 8.77380073e-01 -5.78272700e-01
3.11215788e-01 4.07165676e-01 2.03022704e-01 -2.96500772e-01
-2.21097499e-01 -3.16460758e-01 2.44834319e-01 6.35985792e-01
2.35492334e-01 -3.97651404e-01 1.21879153e-01 1.92804307e-01
6.57965124e-01 2.41491720e-01 -4.62562926e-02 -6.43074691e-01
7.16395378e-01 -2.65795469e-01 6.88871264e-01 4.48073208e-01
-3.84346426e-01 1.02607775e+00 2.25338414e-01 -3.86966318e-01
-1.03166211e+00 -1.12772536e+00 -3.66379380e-01 5.63144922e-01
3.77848208e-01 8.95268917e-02 -1.36405206e+00 -8.31152260e-01
-4.38391864e-01 5.21283448e-01 -6.42305374e-01 -3.58678885e-02
-6.88880384e-01 -9.90758002e-01 5.01263916e-01 1.75497577e-01
6.58668339e-01 -7.97024310e-01 -3.51688921e-01 1.86653271e-01
-3.23541701e-01 -1.01644099e+00 -9.24508035e-01 3.69124264e-02
-9.92174447e-01 -9.65697110e-01 -1.12272024e+00 -1.02423334e+00
1.20149410e+00 1.59789369e-01 5.82241952e-01 3.48152429e-01
-1.37593523e-01 1.03381358e-01 -1.73792884e-01 6.49807006e-02
-5.55795729e-01 -1.17777146e-01 -1.45962518e-02 3.17934394e-01
-1.71508044e-01 -7.38639414e-01 -6.92315817e-01 5.60514510e-01
-1.11830258e+00 2.99633771e-01 4.23432291e-01 9.51320529e-01
9.47268605e-01 -3.30375731e-01 3.35517108e-01 -8.63639176e-01
8.64110291e-02 -2.82616407e-01 -5.52607536e-01 4.39669371e-01
-6.00684345e-01 1.27922297e-02 4.64968801e-01 -6.37870669e-01
-9.55036104e-01 3.79968822e-01 -1.00913569e-01 -5.91490209e-01
-1.79431617e-01 8.88858065e-02 -7.53797367e-02 -5.05568266e-01
4.17829454e-01 3.74251306e-01 4.36869711e-01 -2.24526152e-01
5.19822240e-01 2.07225189e-01 8.45246553e-01 -5.56123912e-01
1.05457091e+00 7.63579667e-01 -1.10585354e-01 -6.57012284e-01
-7.86280096e-01 -2.97129035e-01 -1.08475459e+00 -3.06814343e-01
1.15646422e+00 -5.15290618e-01 -1.95928700e-02 8.00860465e-01
-9.85211134e-01 -5.54455936e-01 -3.33175242e-01 6.33994639e-01
-5.88072538e-01 6.18146658e-01 -7.43958354e-01 -2.58571863e-01
-3.79038066e-01 -1.63335502e+00 6.79885983e-01 4.12325084e-01
-8.39726627e-02 -9.57405210e-01 1.09889954e-01 4.20328617e-01
2.22511902e-01 1.73673972e-01 9.22823608e-01 -6.42901123e-01
-5.09420931e-01 2.67273068e-01 -1.95127383e-01 2.52920061e-01
3.76124412e-01 -1.16835281e-01 -9.68681395e-01 -4.90659475e-01
3.20386112e-01 -1.56928077e-01 5.77519059e-01 6.72632277e-01
1.12912273e+00 8.96206498e-03 -2.30221465e-01 1.05361676e+00
1.32557034e+00 3.27546805e-01 7.06488550e-01 1.80447116e-01
7.68247247e-01 6.02053940e-01 4.22868848e-01 -1.14965476e-02
4.08242531e-02 5.93096018e-01 1.54457852e-01 -7.21906126e-01
-3.11916053e-01 -2.65821666e-02 8.99346992e-02 1.08822036e+00
-5.15634231e-02 1.62769854e-01 -7.86767781e-01 4.27051276e-01
-1.60011590e+00 -6.59025192e-01 -1.69332936e-01 2.34746504e+00
1.15037096e+00 -4.53472696e-02 6.69072792e-02 -1.91287175e-01
7.65914261e-01 2.25286260e-01 -5.60828865e-01 3.51390205e-02
-2.52643377e-02 2.17900828e-01 4.46346730e-01 7.31946945e-01
-1.03054821e+00 1.06011021e+00 5.60102797e+00 8.85522187e-01
-1.45782483e+00 3.86074543e-01 1.09443605e+00 -1.90075729e-02
-2.43659288e-01 1.22688636e-02 -4.21961308e-01 5.98248720e-01
4.27989632e-01 -1.71436518e-01 5.34668505e-01 3.62401158e-01
2.42133021e-01 -2.91224327e-02 -1.00183558e+00 8.46795619e-01
5.23659810e-02 -1.19904470e+00 2.11979505e-02 2.26900369e-01
1.00619340e+00 1.25159636e-01 -7.72270709e-02 -1.47823438e-01
2.73296535e-02 -7.52988935e-01 6.68555200e-01 5.64742565e-01
9.38827634e-01 -5.35886467e-01 4.91497517e-01 2.38171414e-01
-1.10832739e+00 4.11630690e-01 -2.30090439e-01 4.73254532e-01
3.39588374e-01 6.45600677e-01 -4.54283148e-01 6.80637479e-01
5.43404818e-01 4.78887796e-01 -4.42285419e-01 9.56303835e-01
-1.00373521e-01 5.32005787e-01 -1.75044179e-01 6.49804175e-01
4.01662380e-01 -7.95474708e-01 7.49263704e-01 9.92375851e-01
1.84217349e-01 3.59885633e-01 3.74070972e-01 1.17895079e+00
-2.01183613e-02 2.41046757e-01 -3.16443086e-01 2.47181937e-01
2.26407915e-01 1.50561106e+00 -1.26119244e+00 -4.17843342e-01
-4.42841709e-01 1.13977063e+00 1.74338102e-01 5.18297195e-01
-8.29752326e-01 3.45042944e-02 3.82862300e-01 2.13615865e-01
5.97217642e-02 2.46066544e-02 -4.31841433e-01 -1.10610282e+00
-7.32815117e-02 -7.35029936e-01 2.68227100e-01 -4.88667905e-01
-1.33122301e+00 8.55432570e-01 6.64030761e-02 -1.34852076e+00
2.19980422e-02 -1.43320367e-01 -8.77942383e-01 9.04375494e-01
-1.32396674e+00 -1.23121250e+00 -2.37700909e-01 6.87349200e-01
4.19846267e-01 5.68102524e-02 5.99063993e-01 2.22283736e-01
-6.32559001e-01 5.96755326e-01 6.75466657e-02 2.50572085e-01
9.31319535e-01 -1.01141262e+00 1.64759368e-01 1.11980796e+00
9.14341658e-02 6.56321824e-01 3.56886744e-01 -7.69861221e-01
-9.00134087e-01 -1.08181000e+00 4.46923733e-01 -1.11877963e-01
7.07614839e-01 -1.50303677e-01 -1.34214306e+00 7.48679280e-01
2.53857106e-01 2.71715820e-01 6.59604728e-01 -5.83685279e-01
-1.52043238e-01 -7.62351230e-02 -1.30554163e+00 6.51631236e-01
7.92656004e-01 -2.78274298e-01 -6.31870627e-01 2.20204294e-01
5.73579073e-01 -5.28088272e-01 -8.75308216e-01 3.30474257e-01
5.32197237e-01 -8.72683942e-01 8.83870542e-01 -2.01347351e-01
2.89805621e-01 -5.34313202e-01 2.29243457e-01 -1.26286244e+00
-4.10637081e-01 -7.26197898e-01 4.29821581e-01 1.26333284e+00
4.18619961e-01 -7.48747170e-01 6.56241953e-01 6.65221095e-01
-2.25321069e-01 -9.17367339e-01 -1.10570538e+00 -7.20623672e-01
4.65017319e-01 -9.96152461e-02 5.18740952e-01 1.24501872e+00
-2.16512471e-01 -1.52781844e-01 -1.10313334e-01 1.89569965e-01
8.75944853e-01 1.70612223e-02 4.39154476e-01 -9.25403953e-01
-2.50662208e-01 -5.29658496e-01 -1.96609333e-01 -9.33766961e-01
6.71268180e-02 -1.17640281e+00 1.30872384e-01 -1.10053575e+00
4.21800047e-01 -4.72506493e-01 -4.29582894e-01 6.31334126e-01
-2.91753978e-01 4.79946613e-01 1.21045880e-01 4.88982737e-01
-1.53256252e-01 4.51062113e-01 1.86596370e+00 -4.40579318e-02
-3.36457700e-01 -2.00614393e-01 -5.76302588e-01 7.70017087e-01
6.90489411e-01 -6.19361103e-01 -5.01644313e-01 -3.79945695e-01
-5.43820024e-01 4.72291782e-02 3.28559905e-01 -9.45704937e-01
2.49715120e-01 -8.74840096e-02 3.08256298e-01 -2.48208418e-01
1.58789158e-02 -6.64940596e-01 2.56939888e-01 5.70348442e-01
-1.80991888e-01 -2.44021863e-01 1.19671807e-01 2.28507534e-01
-5.46961837e-03 -2.92656720e-01 1.37608314e+00 -1.30526349e-01
-2.56593674e-01 5.88927209e-01 -7.11331144e-02 1.59294114e-01
1.01073134e+00 -4.20094401e-01 -5.54404445e-02 -6.81420490e-02
-6.66221499e-01 1.79724798e-01 8.78286421e-01 1.36264756e-01
3.49589378e-01 -1.11974883e+00 -6.89910233e-01 4.01032209e-01
-1.34627998e-01 1.01689965e-01 2.74139881e-01 1.45812821e+00
-4.32486594e-01 -1.62150189e-01 -3.78542691e-01 -8.62425447e-01
-1.04694343e+00 4.32422727e-01 6.62518919e-01 -3.28334600e-01
-9.24912930e-01 8.16682816e-01 8.00207555e-01 -4.82566327e-01
-1.25557929e-01 -2.08201200e-01 1.19460337e-01 -2.46186912e-01
4.71043169e-01 1.89252421e-02 3.55253965e-02 -8.76873791e-01
-4.26443696e-01 9.56114769e-01 -2.70865113e-01 -3.18414927e-01
1.21010709e+00 -1.82295814e-01 -2.38772526e-01 2.08441064e-01
1.23102343e+00 -7.53338486e-02 -1.46861374e+00 -3.72345567e-01
-1.69701889e-01 -3.36685032e-01 1.46315977e-01 -7.32865214e-01
-1.60075665e+00 6.81080937e-01 7.32929409e-01 -3.28437299e-01
1.14553034e+00 -1.79557577e-02 9.43251967e-01 -3.01727414e-01
3.71549249e-01 -7.93132365e-01 1.26607390e-02 1.84521556e-01
7.41600990e-01 -1.06014037e+00 -1.91749334e-01 -4.97817785e-01
-8.04125488e-01 9.71363723e-01 5.09715378e-01 -1.73537821e-01
6.75741076e-01 4.47435975e-01 4.04229879e-01 -2.19559535e-01
-3.85176130e-02 -9.13748220e-02 3.98876339e-01 4.61791635e-01
2.95735657e-01 -9.49653462e-02 -3.57190043e-01 4.72968966e-01
-4.93507646e-02 7.59392977e-02 2.15939403e-01 7.84305930e-01
-9.48770270e-02 -1.25826001e+00 -3.99383664e-01 2.50988156e-01
-2.36170545e-01 -3.75170782e-02 -2.30268076e-01 5.88383555e-01
1.43713027e-01 6.81681931e-01 1.07945003e-01 -2.41207287e-01
6.15738556e-02 1.05043374e-01 6.74914241e-01 -5.91246784e-01
-6.35056078e-01 5.96356809e-01 -5.33105195e-01 -6.32503569e-01
-4.17881370e-01 -5.61538041e-01 -1.61123633e+00 -8.09251815e-02
-1.81751356e-01 8.24896805e-03 3.23254496e-01 1.17945087e+00
1.93744019e-01 3.50742787e-01 8.49284887e-01 -8.35349381e-01
-1.72242895e-01 -7.46874809e-01 -6.37781620e-01 5.48408031e-01
6.13193214e-02 -6.52288198e-01 -4.34421152e-01 3.41287494e-01] | [14.392355918884277, -2.2677383422851562] |
3ecf118f-f0c4-4a4d-82fd-01383ab5106c | multi-resolution-beta-divergence-nmf-for | 2007.03893 | null | https://arxiv.org/abs/2007.03893v3 | https://arxiv.org/pdf/2007.03893v3.pdf | Multi-Resolution Beta-Divergence NMF for Blind Spectral Unmixing | Many datasets are obtained as a resolution trade-off between two adversarial dimensions; for example between the frequency and the temporal resolutions for the spectrogram of an audio signal, and between the number of wavelengths and the spatial resolution for a hyper/multi-spectral image. To perform blind source separation using observations with different resolutions, a standard approach is to use coupled nonnegative matrix factorizations (NMF). Most previous works have focused on the least squares error measure, which is the $\beta$-divergence for $\beta = 2$. In this paper, we formulate this multi-resolution NMF problem for any $\beta$-divergence, and propose an algorithm based on multiplicative updates (MU). We show on numerical experiments that the MU are able to obtain high resolutions in both dimensions on two applications: (1) blind unmixing of audio spectrograms: to the best of our knowledge, this is the first time a coupled NMF model is used in this context, and (2) the fusion of hyperspectral and multispectral images: we show that the MU compete favorable with state-of-the-art algorithms in particular in the presence of non-Gaussian noise. | ['Cédric Févotte', 'Nicolas Gillis', 'Valentin Leplat'] | 2020-07-08 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 5.39893031e-01 -5.44230819e-01 3.18932772e-01 2.14087531e-01
-1.01409280e+00 -6.07736111e-01 2.96927124e-01 -1.92296356e-01
-4.62704003e-01 6.87463164e-01 -1.15950666e-01 -1.32079750e-01
-7.45440602e-01 -6.51556849e-01 -3.82756561e-01 -1.02292049e+00
-1.47879168e-01 -1.99710056e-02 -2.02619061e-01 -3.26706946e-01
5.90259768e-03 5.11419177e-01 -1.90546286e+00 4.57679033e-02
1.08483005e+00 1.10770917e+00 1.61259636e-01 8.26010466e-01
2.05731988e-01 3.99939805e-01 -4.96383011e-01 -3.31686795e-01
5.93026936e-01 -5.75183511e-01 -5.20278752e-01 2.12175548e-01
4.63192612e-01 1.82033613e-01 1.22419924e-01 1.61076319e+00
5.79106987e-01 3.76998007e-01 7.48669982e-01 -1.10620391e+00
-5.13327956e-01 4.19888377e-01 -8.68293345e-01 2.21961707e-01
1.69315502e-01 -2.32845128e-01 8.25638890e-01 -6.58275902e-01
1.74531475e-01 1.05795240e+00 4.77308601e-01 1.49725139e-01
-1.46700168e+00 -7.16421723e-01 -1.53920680e-01 1.17970824e-01
-1.49809754e+00 -3.32954735e-01 7.73138106e-01 -7.56113350e-01
5.51406860e-01 5.51632702e-01 2.10432589e-01 6.49385333e-01
3.75046358e-02 1.38270915e-01 1.47981286e+00 -6.44455850e-01
2.03454703e-01 1.87217817e-02 -2.04016536e-01 3.01522583e-01
2.34156206e-01 2.09722012e-01 -6.98966980e-01 -3.86194944e-01
6.60278082e-01 -2.43566439e-01 -7.18605757e-01 -1.56351030e-01
-1.08689690e+00 8.89499187e-01 1.08743303e-01 5.52567303e-01
-3.44438851e-01 -2.11805493e-01 -3.49720642e-02 4.93502468e-01
6.78289592e-01 6.41690671e-01 -1.55922130e-01 2.35857338e-01
-1.08136344e+00 3.73994112e-02 5.27031004e-01 2.38629222e-01
8.40136588e-01 1.68307558e-01 2.98881143e-01 1.06151867e+00
2.77683586e-01 8.79121244e-01 5.31272888e-01 -1.11369109e+00
3.38954449e-01 4.75979485e-02 2.99634248e-01 -9.77445364e-01
-4.04790640e-01 -4.72696543e-01 -1.23467577e+00 4.41214561e-01
3.71510059e-01 -2.88924158e-01 -5.54288089e-01 1.88139725e+00
2.33029827e-01 3.48327219e-01 3.14416230e-01 1.09669316e+00
3.84092897e-01 6.78164423e-01 -4.25899863e-01 -8.63044143e-01
1.19253373e+00 -4.92451012e-01 -9.20556426e-01 -2.36559123e-01
-1.53638169e-01 -1.29327476e+00 6.49964511e-01 7.88444340e-01
-1.10655928e+00 -5.30440152e-01 -1.08240795e+00 4.22959447e-01
-2.91154116e-01 2.17321008e-01 2.95284301e-01 9.53952730e-01
-9.31091130e-01 5.85269749e-01 -5.92703223e-01 -2.67614890e-02
-2.30065554e-01 1.66801870e-01 -4.18699533e-01 5.68063855e-02
-1.14492011e+00 6.20529413e-01 5.31062260e-02 1.17431991e-01
-3.10177624e-01 -7.26184011e-01 -5.08549333e-01 -2.37772450e-01
3.42516989e-01 -5.57876229e-01 7.13726997e-01 -1.15319884e+00
-1.39760768e+00 6.34927213e-01 -4.25808504e-02 -3.91538441e-01
2.76961774e-01 -2.36894593e-01 -7.03929663e-01 2.59320587e-01
1.14622481e-01 1.29457623e-01 1.53216708e+00 -1.27869582e+00
-3.60564679e-01 -6.42884970e-01 -3.71659594e-03 1.15809701e-01
-4.29538071e-01 2.44939085e-02 1.05875000e-01 -9.25041318e-01
4.67311710e-01 -1.09221327e+00 -1.61115572e-01 -3.46902430e-01
-2.21933886e-01 4.76652443e-01 5.07234931e-01 -7.55840242e-01
1.02870643e+00 -2.36639714e+00 5.98306596e-01 2.50564724e-01
-7.92628247e-03 2.62211055e-01 -7.16923475e-02 2.71056116e-01
-4.77534086e-01 -8.57025683e-02 -7.01816738e-01 -3.05782855e-01
-2.30852515e-01 -1.05326742e-01 -4.80262071e-01 7.99392819e-01
-1.70684218e-01 1.11362338e-01 -6.88894331e-01 -4.42307070e-02
1.04141913e-01 6.90128207e-01 -2.37717882e-01 2.48401359e-01
1.52404904e-01 4.25473541e-01 1.65286586e-01 3.19277614e-01
9.59039509e-01 -3.71200293e-02 8.79631191e-02 -5.36598802e-01
-3.57466131e-01 -3.85244697e-01 -1.85043716e+00 1.75034022e+00
-4.23093677e-01 4.06064123e-01 8.37407529e-01 -1.01409721e+00
7.95101404e-01 5.27913749e-01 6.93680823e-01 -3.88370812e-01
-7.84545466e-02 4.26606476e-01 -6.40467033e-02 -3.32058400e-01
5.50616860e-01 -6.07895494e-01 2.14246094e-01 5.61711073e-01
1.08181618e-01 -4.44244355e-01 3.36274952e-01 -1.19630128e-01
5.40650070e-01 -3.92749548e-01 1.96649402e-01 -3.84644866e-01
7.44932830e-01 -5.57455778e-01 3.25383455e-01 5.59170961e-01
5.89025952e-02 7.42959023e-01 2.44343430e-01 2.45312408e-01
-6.96807206e-01 -9.70440090e-01 -3.39377493e-01 9.05591786e-01
-5.72085306e-02 -1.41168803e-01 -7.81586349e-01 -5.30111752e-02
1.26319006e-02 3.69919062e-01 -4.18663144e-01 6.98429272e-02
-1.45600632e-01 -1.20944691e+00 4.58813310e-01 1.17758416e-01
4.18466955e-01 -4.09366399e-01 -4.81680602e-01 3.82365473e-02
-2.66389042e-01 -1.18704581e+00 -3.34650338e-01 3.37826431e-01
-6.80656731e-01 -1.03692091e+00 -6.56756818e-01 -1.92377791e-01
2.92577267e-01 4.57012147e-01 7.32306778e-01 -5.58264434e-01
-2.83875346e-01 5.75739980e-01 -3.21140915e-01 -2.53021121e-01
-3.25789005e-01 -3.95528823e-01 5.13692379e-01 6.20781541e-01
-3.64003301e-01 -1.09267199e+00 -4.19793040e-01 3.62945229e-01
-1.40151048e+00 -1.01028174e-01 3.61048937e-01 6.34552240e-01
6.98127806e-01 6.16428196e-01 4.88278985e-01 -4.89238083e-01
7.34047949e-01 -2.87989736e-01 -7.79474437e-01 3.06170881e-01
-5.88434875e-01 7.43622705e-02 6.94611490e-01 -3.68607730e-01
-8.02326798e-01 6.59683496e-02 3.84617224e-02 -5.73206544e-01
-4.60611880e-02 3.83103192e-01 -1.04134776e-01 -3.07094425e-01
8.67544115e-01 1.76021263e-01 -4.09820862e-02 -6.03629649e-01
6.65942788e-01 6.74682975e-01 6.63545609e-01 -5.10561883e-01
9.42500949e-01 6.43910110e-01 3.13018590e-01 -1.16184938e+00
-7.09722996e-01 -6.28860652e-01 -4.23139155e-01 -2.02440336e-01
8.68854225e-01 -9.20930862e-01 -4.32926089e-01 5.82590759e-01
-7.54832685e-01 -4.79982048e-02 -2.52390236e-01 7.76157081e-01
-6.01492643e-01 5.61325192e-01 -3.46924990e-01 -1.12918711e+00
-2.40949020e-01 -1.23623693e+00 8.36528659e-01 8.60465840e-02
3.26629430e-01 -8.23670447e-01 2.82144606e-01 3.89263928e-01
4.17905957e-01 2.91251689e-01 5.42766809e-01 -1.02187052e-01
-9.11006108e-02 1.61980540e-01 -1.61321253e-01 7.08769500e-01
3.17477018e-01 -1.10390067e-01 -1.26925683e+00 -5.90043366e-01
4.76033449e-01 -1.11834250e-01 1.02409160e+00 6.13648891e-01
8.60291421e-01 -2.44920149e-01 2.24299148e-01 7.26995647e-01
1.53510475e+00 6.07164949e-02 4.11760151e-01 -6.88827829e-03
4.57665145e-01 5.47809839e-01 4.60567385e-01 6.18231833e-01
-2.41702810e-01 7.49527752e-01 6.11614883e-01 1.74461678e-02
9.38118100e-02 4.10669804e-01 2.91726679e-01 7.67314315e-01
-3.17434251e-01 -1.60815850e-01 -6.69308245e-01 3.91780198e-01
-1.73921824e+00 -1.01926506e+00 -6.00502789e-02 2.50532174e+00
6.51765168e-01 -2.83182472e-01 1.00967877e-01 5.72007477e-01
7.85659134e-01 5.58742285e-01 -2.63424933e-01 8.70656397e-04
-5.16588509e-01 5.09336233e-01 6.56830013e-01 8.15083683e-01
-1.29821551e+00 1.77874982e-01 5.51063538e+00 9.37740207e-01
-1.21580970e+00 3.03948909e-01 1.93396196e-01 -3.27639490e-01
-2.30825201e-01 -2.82642007e-01 -3.39484006e-01 3.37504417e-01
8.87045383e-01 3.34908776e-02 9.02087569e-01 3.14407110e-01
-2.22781952e-02 7.30576506e-03 -5.30889988e-01 1.36725354e+00
2.57916719e-01 -8.32110763e-01 -2.56239116e-01 1.30222082e-01
8.53190243e-01 -1.07265159e-01 4.76111412e-01 -3.41734707e-01
-8.68693441e-02 -1.04932237e+00 5.04731059e-01 6.10771358e-01
8.48801315e-01 -8.72671664e-01 4.29180175e-01 3.12305242e-01
-1.32465756e+00 -2.50935376e-01 -2.10673347e-01 2.25130796e-01
1.78384423e-01 1.02801573e+00 -3.41919065e-02 9.11373317e-01
6.67829454e-01 4.65669036e-01 -2.67380714e-01 8.27781618e-01
-1.66977532e-02 2.54024833e-01 -3.65281671e-01 5.68452120e-01
6.91201240e-02 -5.99001408e-01 8.97972286e-01 1.07889831e+00
8.48957360e-01 1.82992175e-01 5.84501699e-02 5.71771920e-01
7.34482035e-02 1.58362359e-01 -4.75324720e-01 -8.86281729e-02
1.62527844e-01 1.17648315e+00 -5.45430720e-01 2.19639823e-01
-2.96625435e-01 9.40675378e-01 -1.21650457e-01 3.91547650e-01
-5.68778932e-01 -1.33480310e-01 9.47499216e-01 4.83232997e-02
3.86280537e-01 -2.40198359e-01 -1.79021098e-02 -1.27341855e+00
-3.25469412e-02 -1.12230015e+00 5.00480890e-01 -7.04373717e-01
-1.35927868e+00 7.06862986e-01 -1.30022988e-01 -1.42781115e+00
-2.12384820e-01 -5.95962882e-01 1.34530682e-02 1.13235128e+00
-1.59892642e+00 -7.26356924e-01 -4.89797816e-02 9.27929461e-01
1.62721694e-01 -2.80195475e-01 1.05021381e+00 3.75624776e-01
-4.27844137e-01 2.50758529e-01 5.30730188e-01 -2.37814337e-01
8.16217840e-01 -1.23148108e+00 -1.95251197e-01 1.19556618e+00
3.83806646e-01 4.61366862e-01 1.02876544e+00 -2.57127643e-01
-1.27909160e+00 -7.86236346e-01 2.56941676e-01 8.67782906e-02
7.20559835e-01 3.20617184e-02 -7.95627296e-01 1.99266255e-01
2.21237630e-01 9.22401920e-02 1.15112841e+00 -1.65580705e-01
-5.54230750e-01 -3.74446154e-01 -1.14810514e+00 7.37178512e-03
4.95638013e-01 -8.68073642e-01 -1.64318070e-01 4.27900195e-01
5.08574605e-01 -2.72984028e-01 -1.04323864e+00 5.94583809e-01
4.58048314e-01 -1.37772965e+00 1.27835059e+00 -2.34982058e-01
1.24892421e-01 -6.59362793e-01 -7.81583428e-01 -1.57003152e+00
-2.59502202e-01 -7.71303892e-01 -2.40182597e-02 1.05017674e+00
3.48042250e-01 -5.65720260e-01 1.27629057e-01 2.00463012e-01
2.86351353e-01 -3.01211327e-01 -1.28010619e+00 -8.18714261e-01
-9.62415934e-02 -6.57177985e-01 4.66598630e-01 9.94535685e-01
-2.19272494e-01 1.77089125e-01 -6.81004405e-01 7.87221670e-01
9.31326509e-01 3.40891153e-01 3.35772544e-01 -1.36073101e+00
-7.08795190e-01 -4.79175031e-01 -2.12673888e-01 -6.13342106e-01
1.25567690e-01 -5.66084146e-01 -9.22760740e-02 -1.02283621e+00
-1.37237430e-01 -1.37770697e-01 -3.01744759e-01 -2.54176860e-03
-6.45246729e-02 4.02339637e-01 4.30773765e-01 1.89810634e-01
-9.97082070e-02 2.04024032e-01 8.92714202e-01 -2.00290322e-01
-3.37020725e-01 8.49490464e-02 -6.87557876e-01 7.15095878e-01
6.51863933e-01 -4.00390953e-01 -4.37395066e-01 -4.05424386e-01
4.95187253e-01 4.11979049e-01 3.50575835e-01 -1.27153933e+00
2.04967141e-01 -1.89360783e-01 -3.12807150e-02 -1.37954876e-01
7.22126961e-01 -9.38671172e-01 5.69013357e-01 9.65244249e-02
-1.42812178e-01 -1.11768819e-01 1.92102507e-01 6.80556178e-01
-3.80471587e-01 -1.95389494e-01 1.13262117e+00 -2.41989885e-02
-2.39725038e-01 9.66452286e-02 -2.51767755e-01 -1.69210255e-01
6.45606399e-01 2.32621685e-01 -4.18269753e-01 -4.66271460e-01
-7.71723568e-01 -4.29169029e-01 2.80885935e-01 1.35319129e-01
3.89844298e-01 -1.19583488e+00 -8.93726945e-01 4.25691515e-01
-1.56755969e-01 -3.97542328e-01 5.01786828e-01 1.10360360e+00
-2.02333748e-01 9.83241498e-02 -1.02435812e-01 -4.20758843e-01
-1.37825298e+00 6.31378293e-01 5.76900840e-01 -2.19190612e-01
5.38056269e-02 8.75670969e-01 1.30207777e-01 -1.72056466e-01
-1.41944336e-02 5.19612022e-02 -2.16568470e-01 5.50861537e-01
8.75566721e-01 5.84661305e-01 3.30702037e-01 -1.04970837e+00
-3.39696527e-01 8.84524405e-01 4.78069663e-01 -6.33813918e-01
1.20162630e+00 -1.52919590e-01 -5.94918251e-01 5.24836898e-01
1.26005864e+00 3.92187387e-01 -1.02906370e+00 -1.95499077e-01
-6.73127294e-01 -6.63338125e-01 3.30179989e-01 -6.01414979e-01
-1.25065470e+00 8.39263082e-01 9.77797747e-01 8.22628796e-01
1.64160800e+00 -4.26258683e-01 3.76271605e-01 -1.93082150e-02
2.15583920e-01 -9.20175374e-01 -9.02749598e-02 1.75015599e-01
9.35350001e-01 -8.13211918e-01 3.64516377e-02 -4.18850929e-01
-4.18593228e-01 1.02624583e+00 1.26580726e-02 1.07533634e-01
8.18384647e-01 3.25592786e-01 1.80923700e-01 1.44740149e-01
-1.63393006e-01 -4.48023528e-01 5.05075037e-01 5.48901796e-01
3.42728406e-01 3.30518097e-01 -2.97028422e-01 5.28626502e-01
-2.95416087e-01 -2.56245077e-01 3.44701648e-01 4.54123944e-01
-4.27422285e-01 -1.13695812e+00 -1.05597353e+00 3.17450225e-01
-5.78455567e-01 -1.19430669e-01 -1.77435637e-01 1.93472505e-01
3.61564279e-01 1.37298715e+00 -3.51514399e-01 -4.22489047e-01
2.61310399e-01 1.74262673e-01 4.40238178e-01 -1.98060900e-01
-3.83435577e-01 3.42046708e-01 -2.46091038e-01 -4.08950716e-01
-1.01244247e+00 -7.08346963e-01 -6.96718216e-01 -2.49949872e-01
-4.96313900e-01 4.76015002e-01 6.41846716e-01 7.18722939e-01
2.61729006e-02 3.80948782e-01 7.77459085e-01 -8.69665921e-01
-5.08976579e-01 -8.25386822e-01 -1.27247334e+00 3.10661703e-01
5.82019031e-01 -5.79123795e-01 -8.28727305e-01 5.13381809e-02] | [10.060131072998047, -2.0127081871032715] |
16ea58b4-abfc-4e00-bbf6-45e739e82a5c | multilingual-complex-word-identification | null | null | https://aclanthology.org/R19-2013 | https://aclanthology.org/R19-2013.pdf | Multilingual Complex Word Identification: Convolutional Neural Networks with Morphological and Linguistic Features | The paper is about our experiments with Complex Word Identification system using deep learning approach with word embeddings and engineered features. | ['Kim Cheng SHEANG'] | 2019-09-01 | null | null | null | ranlp-2019-9 | ['complex-word-identification'] | ['natural-language-processing'] | [-4.13829416e-01 -1.68103769e-01 -3.16771448e-01 -2.45273232e-01
1.48351684e-01 -5.30521452e-01 6.20462656e-01 1.21866755e-01
-1.03456223e+00 3.03277314e-01 4.79363531e-01 -8.42634201e-01
1.13157310e-01 -7.84303248e-01 3.32275748e-01 1.18945040e-01
-3.47530812e-01 6.42205536e-01 -1.35679737e-01 -7.81832397e-01
5.02737343e-01 6.07753515e-01 -8.79023075e-01 1.54936640e-02
1.86386108e-01 6.39467180e-01 1.48487493e-01 1.62403309e+00
-5.11834025e-01 4.24487025e-01 -9.35151398e-01 -2.64955431e-01
4.12783414e-01 4.02111530e-01 -1.12039697e+00 -5.28049350e-01
1.95652559e-01 -5.45343339e-01 -6.77064419e-01 1.10569978e+00
1.26373327e+00 6.54017329e-02 6.34537220e-01 -6.58947051e-01
-1.28690159e+00 7.60596752e-01 -4.03290950e-02 6.35488451e-01
6.63101852e-01 4.14348513e-01 1.42805231e+00 -1.20632422e+00
1.07914589e-01 1.72153080e+00 1.09080946e+00 6.50627553e-01
-9.40278351e-01 -7.70249367e-01 -2.17840835e-01 2.48187110e-01
-1.49498701e+00 -1.54547378e-01 1.18992530e-01 -5.81765890e-01
2.40627742e+00 -1.91346467e-01 5.19285381e-01 1.12342107e+00
3.07056814e-01 4.54038173e-01 5.11629164e-01 -7.15145886e-01
-3.30961913e-01 1.10771894e-01 1.48863065e+00 1.27051842e+00
8.78394365e-01 5.80232918e-01 -4.42733057e-02 -3.11705679e-01
3.24803770e-01 -2.55551666e-01 2.83894897e-01 4.98202562e-01
-8.25721443e-01 1.23945105e+00 -2.37480819e-01 7.79034793e-01
-1.25110894e-01 3.40230733e-01 1.02941096e+00 4.86657619e-01
7.34289661e-02 8.77525568e-01 -1.03632700e+00 -1.96053818e-01
-2.37509370e-01 4.43436205e-04 1.03572679e+00 1.04905736e+00
5.30101359e-01 1.00371754e+00 -2.26029158e-01 1.08043253e+00
4.69246954e-01 4.94812012e-01 1.67671967e+00 1.84610859e-01
-2.51329869e-01 5.20257235e-01 -3.22242558e-01 -7.41738141e-01
-6.65610611e-01 -1.36383727e-01 -3.69532108e-01 3.61324757e-01
-3.64288926e-01 -7.01350451e-01 -1.14822662e+00 1.02111590e+00
-4.40312505e-01 -3.59064750e-02 3.62968296e-01 2.18526348e-01
1.34887123e+00 6.05937064e-01 3.75413775e-01 6.46331072e-01
2.00327682e+00 -9.12765026e-01 -1.11830759e+00 -1.41865015e-01
1.00581706e+00 -9.67251599e-01 8.72436762e-01 3.79740149e-01
-5.73592424e-01 -9.28157628e-01 -1.44331849e+00 -4.21497941e-01
-1.29091501e+00 3.94794270e-02 8.63103569e-01 1.41191924e+00
-9.24937010e-01 3.84366155e-01 -2.37023428e-01 -7.92520940e-01
3.18953246e-01 8.40123475e-01 -6.82792425e-01 3.92622918e-01
-1.62727368e+00 1.32442880e+00 1.03984976e+00 -6.51302636e-01
-6.93132699e-01 -3.60746235e-01 -1.21746624e+00 -4.70826365e-02
-5.30442536e-01 -1.75856054e-01 1.01757860e+00 -2.95914263e-01
-1.38406622e+00 8.59128058e-01 6.00598045e-02 -4.76130158e-01
-4.03491408e-01 -4.09777969e-01 -1.03969204e+00 -5.56488514e-01
-3.53115439e-01 3.96800667e-01 8.96034718e-01 -5.39171755e-01
-6.03351593e-01 -2.85153717e-01 -8.63335729e-02 -3.64742041e-01
-1.04214311e+00 1.32921919e-01 2.92424232e-01 -7.85985231e-01
-8.42763722e-01 -3.03858846e-01 9.58472118e-02 -5.55145562e-01
-3.66466850e-01 -1.16121340e+00 7.37027705e-01 -7.42157459e-01
1.67069495e+00 -1.87044346e+00 -3.45609993e-01 -2.45754179e-02
5.58294952e-01 1.21819854e+00 -5.59639335e-01 6.97808981e-01
-6.07201397e-01 5.21595299e-01 5.04724741e-01 -2.62713909e-01
5.96198320e-01 3.29469234e-01 -3.73002887e-03 3.79582405e-01
4.26476210e-01 1.12558591e+00 -7.41025686e-01 -4.51149374e-01
4.32351679e-01 2.92694181e-01 -3.06458563e-01 4.43663597e-01
3.32018465e-01 -9.36486304e-01 -2.97505111e-01 6.15912855e-01
7.43028462e-01 6.26249015e-01 -9.66009125e-02 -4.79160458e-01
-1.60416842e-01 1.29558906e-01 -1.09402883e+00 1.22751439e+00
-8.05665255e-01 9.96629179e-01 -4.42019939e-01 -8.86646748e-01
1.05258572e+00 5.37125707e-01 -1.86688021e-01 -4.29025620e-01
5.78486383e-01 2.05209419e-01 4.95598763e-01 -8.52229834e-01
6.38487816e-01 -1.93650231e-01 -1.75362960e-01 6.54034555e-01
1.17506897e+00 3.15675557e-01 -2.38232538e-01 -3.82328406e-02
1.28040111e+00 -5.87299764e-01 6.75637901e-01 -7.00427055e-01
8.97261381e-01 -7.73449317e-02 -7.85352811e-02 6.47905290e-01
-5.11575043e-01 2.07402527e-01 -9.95164085e-03 -1.14267206e+00
-1.24466038e+00 -9.04359639e-01 -2.79725701e-01 1.46303141e+00
-4.63890165e-01 -6.40667140e-01 -5.93444645e-01 -6.95004165e-01
2.33520254e-01 4.83826160e-01 -6.32482350e-01 -5.53662956e-01
-6.95619881e-01 -3.58192682e-01 1.43106401e+00 8.89699280e-01
1.39187381e-01 -1.16568434e+00 -1.96868017e-01 3.00916463e-01
9.50041711e-01 -1.02960742e+00 -9.76916134e-01 6.09306633e-01
-2.43802324e-01 -1.04968619e+00 -3.00457865e-01 -1.62822866e+00
2.18729720e-01 -6.26916671e-03 1.31820714e+00 4.16603953e-01
-1.22363365e+00 2.43361697e-01 -5.87275207e-01 -5.45974314e-01
-2.43520096e-01 4.56417084e-01 8.19364727e-01 -5.87303936e-01
1.56528330e+00 9.95418057e-02 -4.62786168e-01 -4.65786785e-01
-9.53026772e-01 -8.46444845e-01 3.66341740e-01 1.26418102e+00
-8.23890030e-01 5.01168035e-02 4.79804128e-01 -6.64199889e-01
1.63842356e+00 -2.70142496e-01 -4.45647687e-01 1.18177272e-01
-8.04759502e-01 6.96242034e-01 3.01160455e-01 -7.39611149e-01
-3.18151623e-01 7.34971687e-02 -9.71598625e-01 4.26718801e-01
-2.06458926e-01 1.32493168e-01 -1.44999353e-02 -2.63384581e-01
7.13044167e-01 1.31489620e-01 5.48624210e-02 -6.86704636e-01
4.40974802e-01 1.40192127e+00 1.14647381e-01 -2.73737580e-01
8.59218776e-01 -3.79574150e-01 -7.90015399e-01 -1.02661908e+00
-2.35626996e-02 -7.22909749e-01 -9.44730818e-01 2.35190868e-01
1.11191463e+00 -9.30231810e-01 -1.07686627e+00 7.63587654e-01
-1.57020330e+00 8.57392922e-02 -1.15410388e-01 5.12477577e-01
1.94339216e-01 4.31676477e-01 -6.38157308e-01 -8.41903150e-01
-9.82736111e-01 -1.02522993e+00 9.71440077e-01 -1.05298243e-01
-1.00720310e+00 -1.71166992e+00 5.46922982e-01 -2.16824651e-01
6.02943480e-01 -3.85917276e-01 9.84193027e-01 -1.42899895e+00
6.71633005e-01 -8.61859381e-01 -2.15819851e-01 8.45898211e-01
4.22797352e-01 -9.90881026e-02 -1.11236417e+00 -1.82414949e-01
-3.00715089e-01 -4.42554951e-01 7.78396428e-01 -1.63551882e-01
9.32161570e-01 -9.97692645e-02 -2.07552269e-01 6.03000581e-01
1.73905551e+00 1.03209771e-01 4.37972397e-01 2.58942336e-01
9.86706495e-01 1.14351392e-01 -2.56625205e-01 5.17463505e-01
1.48711294e-01 1.48793980e-01 -2.28810832e-01 1.21125661e-01
-2.46928871e-01 7.76146948e-02 4.41429466e-01 1.01124907e+00
5.76012433e-01 -5.61705053e-01 -1.22425091e+00 8.05386066e-01
-1.13274705e+00 -6.12626135e-01 -2.32540518e-02 1.19313335e+00
6.37777269e-01 -9.75498036e-02 1.45376608e-01 4.18068767e-01
8.27886283e-01 3.78855616e-01 -3.27976435e-01 -1.56211042e+00
-1.18230857e-01 1.32880735e+00 8.86751652e-01 1.15614283e+00
-1.14340818e+00 1.53133571e+00 8.18894196e+00 7.85024226e-01
-8.94906223e-01 2.18338355e-01 -1.06946252e-01 5.50793231e-01
-1.03396542e-01 -7.19256699e-01 -1.06802070e+00 -3.00803576e-02
1.19695377e+00 -4.80421871e-01 9.02467743e-02 7.27512777e-01
-4.38503057e-01 5.65734208e-01 -1.16237915e+00 1.14433706e+00
2.35308066e-01 -9.09258902e-01 3.01507920e-01 9.32254791e-02
4.09191139e-02 3.01535398e-01 1.58196285e-01 3.93131524e-01
1.16517329e+00 -1.63214302e+00 -2.57077757e-02 1.68020576e-02
1.01459062e+00 -8.52312744e-01 9.90114689e-01 -3.01228315e-01
-1.19412005e+00 -2.34261230e-01 -4.85113382e-01 -3.08111966e-01
-3.01883370e-01 2.58584172e-01 -7.74471700e-01 -2.94358462e-01
1.93258524e-01 7.04405785e-01 -4.76056963e-01 2.74260521e-01
1.97208822e-01 3.60302240e-01 1.04679517e-01 -8.83797348e-01
4.18099403e-01 3.30540299e-01 1.31060883e-01 2.08799243e+00
1.05421722e-01 5.34791313e-02 -1.14557125e-01 3.25400829e-01
4.21316139e-02 1.70597598e-01 -1.10636330e+00 -7.62564659e-01
2.34303713e-01 1.30725801e+00 -1.21520661e-01 -6.55664861e-01
-2.91463047e-01 1.05459917e+00 3.48757893e-01 -1.62607819e-01
-2.03285679e-01 -1.16638136e+00 1.62012100e+00 -3.47877175e-01
1.68560743e-01 -7.32541323e-01 -5.32257020e-01 -9.33695138e-01
-6.22601926e-01 -6.55608356e-01 1.82734057e-01 -7.61530325e-02
-1.66053355e+00 1.02070320e+00 -4.43800837e-01 -6.36403918e-01
5.64951412e-02 -1.77311456e+00 -8.64238679e-01 1.23895323e+00
-1.23860276e+00 -6.88946545e-01 6.72997162e-02 5.65025985e-01
7.42673934e-01 -1.38070691e+00 1.59670603e+00 5.33541381e-01
-5.83461106e-01 1.18133008e+00 4.17454541e-02 7.87912905e-01
5.58739483e-01 -1.68292046e+00 1.41322291e+00 1.39813527e-01
1.68509260e-01 9.89135921e-01 5.50852835e-01 -3.60410899e-01
-1.80053949e+00 -9.20298874e-01 1.09237516e+00 -7.32271194e-01
1.25117099e+00 -8.49262774e-01 -5.17584085e-01 4.39307809e-01
1.10150611e+00 1.14719057e-02 1.23336136e+00 -8.83323550e-02
-5.92338443e-01 1.42521739e-01 -1.22713745e+00 2.90970683e-01
9.64565456e-01 -1.00992167e+00 -8.35232198e-01 6.35418057e-01
1.05453897e+00 4.14817840e-01 -1.06464362e+00 1.69934079e-01
7.98612356e-01 -6.43835887e-02 1.30874205e+00 -1.36813831e+00
-3.10068280e-01 2.52861708e-01 -1.59625486e-01 -1.54329967e+00
-8.08189690e-01 -5.46343267e-01 9.71401259e-02 9.73465979e-01
4.26411033e-01 -1.06035948e+00 4.36581999e-01 7.89890066e-02
9.89497900e-02 -1.62780121e-01 -5.65443456e-01 -9.09193814e-01
6.61537647e-01 -4.44967300e-01 6.54404342e-01 1.06572008e+00
2.01301634e-01 1.01662934e+00 -2.22245082e-01 -8.54450837e-02
-1.78492218e-02 -8.01018775e-01 2.12400794e-01 -1.38178349e+00
-7.51060620e-02 -6.01172626e-01 -1.09855938e+00 -6.73046052e-01
7.65432596e-01 -6.50556386e-01 -1.34954333e-01 -1.05033267e+00
-1.65558517e-01 2.53091663e-01 -4.66384321e-01 2.69411206e-01
-2.22645834e-01 3.46550137e-01 -2.07049787e-01 -6.15616798e-01
-9.69713330e-02 2.97661096e-01 3.85155708e-01 -5.75378299e-01
4.66990083e-01 -6.78566933e-01 -5.50073802e-01 5.74065208e-01
1.04391313e+00 -1.99873209e-01 -2.24015102e-01 -8.93532693e-01
1.27430761e-03 -9.60812032e-01 -3.78994197e-01 -7.84102857e-01
1.82034168e-02 4.77615118e-01 1.64060563e-01 -4.93241429e-01
1.20997570e-01 -7.98028469e-01 -7.52881765e-01 1.20209849e+00
-3.15444738e-01 8.71607959e-01 3.78700763e-01 1.60890985e-02
-5.22923283e-02 -5.76127827e-01 8.94861758e-01 -2.67328113e-01
-1.26417816e+00 2.41893336e-01 -1.13830471e+00 -1.73165038e-01
8.48142207e-01 -1.80529594e-01 -1.58366427e-01 1.47598371e-01
-8.04949284e-01 1.89366937e-01 -1.81646362e-01 1.00550902e+00
8.91664326e-01 -1.63950431e+00 -7.31213748e-01 7.05047131e-01
5.20020545e-01 -9.30485070e-01 -6.71876431e-01 -5.42728722e-01
-1.19915307e+00 9.41137850e-01 -3.46017152e-01 2.06532598e-01
-1.60842764e+00 6.45162523e-01 5.73355138e-01 1.03924341e-01
-2.97319949e-01 1.28904879e+00 -3.25074315e-01 -7.81948090e-01
4.87317115e-01 -3.26761037e-01 -8.25985253e-01 9.96116102e-02
8.11698973e-01 4.01592880e-01 5.59274666e-02 -8.03518057e-01
-5.44380009e-01 7.71640956e-01 -3.01620632e-01 -1.65699124e-02
1.05551898e+00 1.90099671e-01 -3.14109713e-01 2.51189232e-01
2.11656642e+00 -4.86821532e-01 3.32669318e-01 -3.04101169e-01
5.12270510e-01 -5.12856282e-02 2.46289298e-01 -4.85831410e-01
-7.37646878e-01 1.22324193e+00 1.58324587e+00 2.34257296e-01
4.47247952e-01 -3.48975509e-01 1.35270441e+00 9.67634201e-01
-1.58637092e-01 -1.37309170e+00 2.28774726e-01 1.29497516e+00
5.15219688e-01 -1.59301770e+00 -1.78252444e-01 3.54691505e-01
-1.68364272e-01 1.60241413e+00 7.87641346e-01 -9.23289359e-01
1.64652634e+00 8.08819056e-01 2.45032430e-01 -5.55794477e-01
-7.83565521e-01 -6.12055063e-01 4.94419076e-02 1.06709480e+00
7.32266426e-01 3.24395120e-01 -5.61407447e-01 6.22224271e-01
-1.94647029e-01 -4.87857133e-01 2.90483564e-01 6.32861555e-01
-9.16072190e-01 -1.44963515e+00 -1.37923911e-01 2.87715882e-01
-5.78875363e-01 -8.92939568e-01 -9.39060986e-01 5.12179852e-01
2.06352577e-01 8.30695808e-01 3.33239794e-01 -9.79219675e-01
4.77654219e-01 3.10965240e-01 1.02954991e-01 -1.24978220e+00
-1.08648026e+00 -6.31781816e-01 3.24167699e-01 -1.54361844e-01
2.32527182e-01 -9.15993527e-02 -1.22056472e+00 -5.47909439e-01
-5.32207429e-01 1.53711155e-01 8.05371225e-01 8.80707324e-01
4.06866930e-02 5.12457550e-01 6.54913068e-01 -4.18941528e-01
-5.78450143e-01 -1.66793025e+00 -6.61576867e-01 2.71503329e-01
6.71393454e-01 -3.23096603e-01 -1.13582313e-01 -3.26377936e-02] | [10.456299781799316, 10.156827926635742] |
6d158000-9792-4006-8cde-bd02cd1f4605 | realheponet-a-robust-single-stage-convnet-for | 2011.01890 | null | https://arxiv.org/abs/2011.01890v1 | https://arxiv.org/pdf/2011.01890v1.pdf | RealHePoNet: a robust single-stage ConvNet for head pose estimation in the wild | Human head pose estimation in images has applications in many fields such as human-computer interaction or video surveillance tasks. In this work, we address this problem, defined here as the estimation of both vertical (tilt/pitch) and horizontal (pan/yaw) angles, through the use of a single Convolutional Neural Network (ConvNet) model, trying to balance precision and inference speed in order to maximize its usability in real-world applications. Our model is trained over the combination of two datasets: 'Pointing'04' (aiming at covering a wide range of poses) and 'Annotated Facial Landmarks in the Wild' (in order to improve robustness of our model for its use on real-world images). Three different partitions of the combined dataset are defined and used for training, validation and testing purposes. As a result of this work, we have obtained a trained ConvNet model, coined RealHePoNet, that given a low-resolution grayscale input image, and without the need of using facial landmarks, is able to estimate with low error both tilt and pan angles (~4.4{\deg} average error on the test partition). Also, given its low inference time (~6 ms per head), we consider our model usable even when paired with medium-spec hardware (i.e. GTX 1060 GPU). * Code available at: https://github.com/rafabs97/headpose_final * Demo video at: https://www.youtube.com/watch?v=2UeuXh5DjAE | ['Manuel J. Marín-Jiménez', 'Rafael Muñoz-Salinas', 'Francisco J. Madrid-Cuevas', 'Rafael Berral-Soler'] | 2020-11-03 | null | null | null | null | ['head-pose-estimation'] | ['computer-vision'] | [-2.05005065e-01 2.13111833e-01 2.79507548e-01 -4.45546925e-01
-4.89255667e-01 -1.96656391e-01 3.31220478e-01 -1.29289478e-01
-7.83582509e-01 6.48191154e-01 -1.02838442e-01 -3.06834072e-01
1.50166795e-01 -6.71354711e-01 -7.33144343e-01 -6.54351890e-01
-2.06773430e-01 3.82366538e-01 2.87772864e-01 -1.69218391e-01
6.60437066e-03 7.02752113e-01 -1.94938219e+00 -1.62221119e-01
4.05193061e-01 1.19283390e+00 -1.12666544e-02 7.39154100e-01
6.02334559e-01 5.06225765e-01 -5.75936377e-01 -5.00240684e-01
3.08899343e-01 1.19360983e-01 -5.54318011e-01 -1.67517468e-01
6.96084261e-01 -4.77793992e-01 -1.45379469e-01 9.84782815e-01
9.62648571e-01 -7.87955076e-02 2.45953515e-01 -1.27260685e+00
2.56877124e-01 -1.32406756e-01 -4.53864783e-01 5.41867539e-02
5.29814661e-01 3.89742583e-01 4.55881715e-01 -8.99023712e-01
5.76826334e-01 1.03975570e+00 8.20422351e-01 6.10893965e-01
-1.00687301e+00 -9.64371145e-01 -3.43168616e-01 2.93197274e-01
-1.76836169e+00 -7.23999083e-01 4.47100639e-01 -1.71855047e-01
8.87604952e-01 3.60102803e-01 7.53566444e-01 1.34127557e+00
3.32933635e-01 4.47188586e-01 1.10724950e+00 -2.84635484e-01
3.65269601e-01 4.19353843e-02 -1.26653776e-01 7.40699768e-01
2.90402651e-01 -4.17063832e-02 -4.59234416e-01 -7.76665658e-02
7.41920352e-01 -3.07976633e-01 -5.62815607e-01 -1.96880892e-01
-8.31220508e-01 7.58356392e-01 5.08166194e-01 2.20043406e-01
-4.06832278e-01 1.67404518e-01 4.16747391e-01 1.17164396e-03
3.12341869e-01 -4.33937758e-02 -5.01320779e-01 -2.50995725e-01
-9.33107555e-01 3.03045064e-01 9.09225225e-01 9.22789574e-01
7.01532423e-01 -1.97304916e-02 3.37018788e-01 6.85952127e-01
4.65585738e-01 6.28466964e-01 4.34508890e-01 -6.98783875e-01
2.28727624e-01 1.80561751e-01 -7.81329498e-02 -1.09588420e+00
-8.88420761e-01 -1.77083269e-01 -9.04788136e-01 5.37097156e-01
6.72248960e-01 -4.49764132e-01 -8.34747970e-01 1.68130994e+00
5.81624448e-01 1.00559555e-01 -3.06253284e-01 9.91945326e-01
7.99295962e-01 2.43986726e-01 -3.20807137e-02 -2.67514102e-02
1.61501873e+00 -4.82383430e-01 -3.45534712e-01 -2.17772961e-01
5.28608084e-01 -8.53906155e-01 9.09078658e-01 6.52207494e-01
-1.06078768e+00 -5.82658231e-01 -1.08599162e+00 -1.09835498e-01
-5.68458974e-01 3.79678756e-01 2.28687137e-01 8.86747360e-01
-1.50152564e+00 4.77866113e-01 -9.24196959e-01 -6.07626855e-01
2.92835802e-01 9.09153938e-01 -4.98529375e-01 1.16292797e-01
-9.90915716e-01 9.43095505e-01 2.16693133e-01 4.90890920e-01
-4.25747067e-01 -2.96800524e-01 -8.13710928e-01 -1.32103309e-01
2.02201605e-01 -5.67866325e-01 1.26750636e+00 -9.29549634e-01
-1.68093526e+00 9.13812160e-01 -7.26107061e-02 -4.03933913e-01
8.43009770e-01 -4.34652656e-01 -2.52268225e-01 1.09749049e-01
-1.85118154e-01 9.64880645e-01 7.17725337e-01 -9.39772010e-01
-4.60215777e-01 -6.07451618e-01 -7.94409961e-02 5.61542399e-02
-9.87971500e-02 3.97783443e-02 -8.40818048e-01 -2.31398568e-01
-4.04369272e-02 -1.25235820e+00 9.64789689e-02 1.28942952e-01
-3.73932689e-01 6.56533837e-02 8.23520601e-01 -7.98092127e-01
8.94234359e-01 -1.92319620e+00 -3.18043858e-01 4.60436732e-01
1.11141555e-01 4.96871948e-01 3.03829908e-01 1.62451267e-02
-1.86561286e-01 -2.09187582e-01 4.37832363e-02 -5.51204622e-01
-1.18205838e-01 4.27287780e-02 2.58150429e-01 1.00340617e+00
-2.25429550e-01 6.66098237e-01 -4.72059190e-01 -5.77780783e-01
4.54236627e-01 1.07231951e+00 -4.74885106e-01 2.79136710e-02
1.51362687e-01 4.93038625e-01 -2.54244238e-01 6.17791116e-01
9.55093026e-01 2.05168873e-01 3.03558677e-01 -3.27535063e-01
-1.49917796e-01 -2.03451011e-02 -1.38710332e+00 1.39218831e+00
-5.22770107e-01 8.55771482e-01 1.86314449e-01 -4.24467027e-01
8.50774467e-01 4.49746102e-01 1.93252861e-01 -8.78013194e-01
8.04198444e-01 2.57065713e-01 -3.78822833e-02 -4.05723870e-01
3.62451553e-01 2.61760354e-01 2.00475469e-01 5.44520691e-02
6.19194284e-02 5.21911606e-02 -4.94930930e-02 -2.89702207e-01
7.62505829e-01 1.03847541e-01 2.73522913e-01 -4.09175187e-01
5.39722741e-01 -4.48038459e-01 2.40049437e-01 2.49929458e-01
-2.98378587e-01 7.85324156e-01 4.96438086e-01 -5.41843534e-01
-1.03274596e+00 -5.82138836e-01 -3.56742322e-01 9.11308050e-01
-1.50287524e-01 -2.22879410e-01 -1.25547063e+00 -2.92690605e-01
-2.68563002e-01 4.87620860e-01 -6.61898077e-01 1.52396098e-01
-8.34777474e-01 -7.75566578e-01 7.51105070e-01 5.19425750e-01
6.38231456e-01 -1.04394555e+00 -1.36305559e+00 -1.70599222e-02
1.16649263e-01 -1.06927466e+00 -1.65111780e-01 7.91707188e-02
-4.30208445e-01 -1.22421074e+00 -6.48005068e-01 -4.78146583e-01
6.44682348e-01 -1.20327175e-01 1.06472266e+00 1.76206976e-01
-2.37354696e-01 4.11394894e-01 -9.73907411e-02 -3.55952799e-01
1.48176581e-01 1.35848254e-01 2.59003341e-01 7.58180395e-02
4.33027238e-01 -5.93541622e-01 -9.00742590e-01 3.02199841e-01
-7.23900080e-01 5.45895621e-02 3.58799756e-01 6.50511742e-01
1.04964554e-01 -4.29206014e-01 -8.85910019e-02 -3.99620563e-01
4.13731635e-01 -2.70694762e-01 -8.75292480e-01 4.48468886e-03
-3.90648335e-01 -2.96753824e-01 4.83610034e-01 -2.54868269e-01
-6.06890142e-01 1.13519378e-01 -7.39522696e-01 -4.67514098e-01
-5.76251984e-01 -1.83671713e-02 -2.26370752e-01 -3.15073162e-01
6.40876949e-01 -1.92269776e-02 2.59995013e-01 -2.81348854e-01
-8.78766552e-02 7.45889962e-01 5.59546292e-01 -3.85562927e-01
3.46859396e-01 4.78791833e-01 1.50645494e-01 -1.02574539e+00
-1.86653733e-01 -2.04224110e-01 -8.16028893e-01 -4.90940183e-01
1.01621032e+00 -8.57726932e-01 -1.25336373e+00 5.76293945e-01
-1.17112505e+00 -4.23049212e-01 2.27389544e-01 6.01199806e-01
-4.55799937e-01 1.59525484e-01 -2.78069168e-01 -6.66164339e-01
-6.33499861e-01 -1.37961507e+00 1.16226470e+00 4.19147849e-01
-2.81322151e-01 -8.01308692e-01 -1.99817598e-01 1.54947722e-02
2.78938800e-01 5.41698158e-01 1.29848644e-01 -5.38001895e-01
-3.57215196e-01 -3.85117620e-01 -1.78174689e-01 2.60780931e-01
-3.35274547e-01 1.28753915e-01 -1.22289371e+00 -4.99598622e-01
-7.94167165e-03 -2.70036161e-01 2.82960773e-01 3.66081953e-01
1.08909976e+00 -2.31399834e-01 -1.32980764e-01 8.06489229e-01
1.30404949e+00 1.47962630e-01 9.07776833e-01 4.79496270e-01
5.74392736e-01 3.82969737e-01 4.65843707e-01 5.33040702e-01
3.63095671e-01 1.11026251e+00 5.73109865e-01 -9.32782367e-02
-2.25417893e-02 1.19051315e-01 1.92180380e-01 3.25763941e-01
-3.93532068e-01 -2.13729292e-01 -1.18907583e+00 1.27328292e-01
-1.51174486e+00 -6.06626987e-01 -1.34427354e-01 2.52811265e+00
4.61057514e-01 2.49523789e-01 2.16186434e-01 1.78938091e-01
5.71478724e-01 5.11931675e-03 -2.22076043e-01 -5.86199999e-01
2.23677650e-01 5.42343676e-01 6.87732697e-01 5.07787645e-01
-9.95196044e-01 6.78655386e-01 4.66107798e+00 5.78664958e-01
-1.73977351e+00 1.14935890e-01 8.44399989e-01 -1.70075759e-01
4.43976194e-01 -4.36171353e-01 -1.01045144e+00 6.00646913e-01
1.20470560e+00 3.85755658e-01 2.57505834e-01 8.80860686e-01
2.50504494e-01 -4.19163346e-01 -7.79919922e-01 1.11329305e+00
2.99179226e-01 -8.52435231e-01 -5.35281003e-01 3.12872171e-01
7.31878430e-02 3.11958045e-01 -2.75448430e-02 1.33494243e-01
-8.31238851e-02 -1.06502235e+00 8.00637126e-01 2.64579952e-01
1.08104718e+00 -8.19333494e-01 9.51018810e-01 3.47400784e-01
-1.12202907e+00 1.12013340e-01 -1.79622009e-01 1.22601129e-01
-7.34270513e-02 2.19143704e-01 -1.12328720e+00 2.68886894e-01
9.66045320e-01 5.26932105e-02 -6.89651728e-01 9.42383468e-01
-1.85567200e-01 2.56419659e-01 -7.91523039e-01 -6.58947080e-02
8.96719173e-02 3.39228325e-02 3.13517183e-01 1.28780782e+00
4.96972382e-01 9.95861813e-02 -2.15194285e-01 3.30516636e-01
1.05023630e-01 1.04819685e-01 -4.54848170e-01 8.97706211e-01
3.35362464e-01 1.27952349e+00 -8.17122340e-01 -1.63115934e-01
-1.68721601e-01 8.14091861e-01 2.61023611e-01 1.02504857e-01
-1.26104510e+00 -2.79065937e-01 6.61726117e-01 3.55354369e-01
3.81101310e-01 -2.85488099e-01 -2.07856074e-01 -1.00766194e+00
1.21889487e-01 -8.81662488e-01 3.16640325e-02 -8.68217766e-01
-3.21534812e-01 9.69586849e-01 1.75690845e-01 -1.22243023e+00
-5.79960644e-01 -9.03945327e-01 -5.43785751e-01 8.63316953e-01
-1.11434078e+00 -1.16762483e+00 -6.77273810e-01 8.06698859e-01
3.41745049e-01 9.31224748e-02 8.05064440e-01 4.77171242e-01
-5.20756721e-01 9.29941416e-01 -1.66201115e-01 1.91858724e-01
5.75557649e-01 -9.50851798e-01 4.61840779e-01 5.97046375e-01
-2.11593747e-01 5.33473074e-01 8.85777295e-01 -3.29787403e-01
-1.28958678e+00 -9.56160247e-01 7.97865093e-01 -2.52031714e-01
1.80512950e-01 -5.53512633e-01 -7.00655639e-01 6.72524035e-01
1.64926261e-01 2.95889139e-01 3.57908219e-01 -1.12536594e-01
-1.39922932e-01 -1.70084104e-01 -1.41525424e+00 6.47737503e-01
6.74084961e-01 -2.42315978e-01 -4.77006771e-02 1.28391206e-01
1.20521607e-02 -9.30355012e-01 -7.34806061e-01 4.84416276e-01
9.26249743e-01 -1.42783296e+00 9.31263328e-01 8.21500123e-02
4.26461883e-02 -1.86598226e-01 -6.11110404e-02 -9.35244262e-01
4.86442037e-02 -5.07675588e-01 1.58241484e-02 8.28694582e-01
1.42922342e-01 -8.97247434e-01 8.23206484e-01 5.34265816e-01
9.77215320e-02 -1.07022107e+00 -1.37310195e+00 -3.07083070e-01
-2.14074731e-01 -5.67267179e-01 5.30166268e-01 4.58716393e-01
-2.52094269e-01 7.88019970e-02 -2.64196068e-01 4.04885620e-01
4.50185269e-01 -4.52876359e-01 1.03722739e+00 -9.93103802e-01
-4.24393639e-03 -3.76896173e-01 -8.14408779e-01 -7.61217535e-01
-9.78965908e-02 -2.75043190e-01 -7.30538666e-02 -8.04079294e-01
-2.94196159e-01 -2.56548494e-01 2.24904358e-01 5.83003998e-01
3.20765615e-01 7.14678884e-01 2.78109014e-01 -8.53182077e-02
-1.64932936e-01 1.22145347e-01 7.08300769e-01 3.42460126e-01
1.20869568e-02 1.38109252e-01 -2.30112016e-01 9.07552779e-01
1.02629268e+00 -2.10530162e-01 -5.68683334e-02 -3.64002734e-01
4.71325368e-02 1.03138007e-01 6.83662832e-01 -1.46433973e+00
3.37338835e-01 3.24697137e-01 7.10692883e-01 -4.09098417e-01
6.75705969e-01 -8.62321317e-01 3.73365760e-01 5.22155523e-01
2.55875021e-01 3.90060127e-01 4.17667508e-01 -1.71900988e-01
5.76016195e-02 -7.34539479e-02 1.03628361e+00 -4.79910988e-03
-7.03564405e-01 2.23578796e-01 -4.90900204e-02 -3.60728204e-01
1.01242435e+00 -4.61638659e-01 -2.40119815e-01 -4.61618930e-01
-7.02575147e-01 -6.27174452e-02 5.60737133e-01 2.73818821e-01
5.16851783e-01 -9.28556740e-01 -4.99812663e-01 5.07470310e-01
-1.23734154e-01 5.98281287e-02 3.39773595e-01 1.07907188e+00
-1.04064512e+00 3.93328041e-01 -3.41656685e-01 -7.54680336e-01
-1.60435569e+00 1.78361148e-01 5.46141863e-01 3.93500179e-02
-5.41527927e-01 7.54130483e-01 -7.96268284e-02 -3.00845891e-01
4.03213501e-01 -3.77282232e-01 -2.28771836e-01 1.11099079e-01
5.43992639e-01 2.82682300e-01 5.30955911e-01 -9.87566650e-01
-5.02979338e-01 6.20962083e-01 1.38245001e-01 -1.44554839e-01
1.22776496e+00 2.78115720e-02 6.82876706e-02 -2.66052894e-02
1.46887851e+00 7.56954551e-02 -1.15553784e+00 3.65393877e-01
-2.24568650e-01 -4.34208900e-01 6.22571334e-02 -4.70712602e-01
-1.14661217e+00 8.47374737e-01 1.19220102e+00 3.04349959e-02
1.25403512e+00 -3.52965355e-01 5.08119822e-01 1.50854394e-01
4.78953540e-01 -9.11601305e-01 -3.84652287e-01 4.47574466e-01
8.22854936e-01 -1.12308991e+00 2.00165763e-01 -2.46792421e-01
-4.05592084e-01 1.27201962e+00 6.93223894e-01 -7.08666295e-02
6.90150797e-01 3.64079773e-01 1.75084725e-01 -2.25787044e-01
-4.20342654e-01 -2.36331910e-01 1.33952618e-01 5.07466316e-01
5.32087564e-01 6.01578951e-02 -2.69832462e-01 -3.91464233e-02
-6.48916781e-01 1.54141575e-01 3.31207663e-01 9.60658252e-01
-1.97607845e-01 -7.70883441e-01 -6.38487995e-01 2.12650985e-01
-5.51501632e-01 1.13571636e-01 8.64420459e-02 1.17622387e+00
5.74205220e-01 5.51979244e-01 3.37561369e-01 -4.03790891e-01
3.19698572e-01 -1.31784126e-01 3.70649785e-01 -1.03084154e-01
-6.26070678e-01 1.16808787e-01 6.72664493e-02 -7.70540416e-01
-2.72680759e-01 -6.10626578e-01 -8.10188711e-01 -6.60555542e-01
-3.23578715e-02 -1.02770284e-01 9.73515689e-01 6.29788280e-01
2.38835022e-01 3.32451522e-01 3.51614177e-01 -1.44446492e+00
-2.80912966e-01 -1.05477679e+00 -4.07741934e-01 1.00758173e-01
2.21533388e-01 -6.69413865e-01 -2.61875629e-01 -8.19485486e-02] | [13.668924331665039, 0.2891647517681122] |
357d7d01-782b-42d9-87f9-6431206470a0 | spatially-regularized-parametric-map | 1911.03786 | null | https://arxiv.org/abs/1911.03786v2 | https://arxiv.org/pdf/1911.03786v2.pdf | Spatially Regularized Parametric Map Reconstruction for Fast Magnetic Resonance Fingerprinting | Magnetic resonance fingerprinting (MRF) provides a unique concept for simultaneous and fast acquisition of multiple quantitative MR parameters. Despite acquisition efficiency, adoption of MRF into the clinics is hindered by its dictionary matching-based reconstruction, which is computationally demanding and lacks scalability. Here, we propose a convolutional neural network-based reconstruction, which enables both accurate and fast reconstruction of parametric maps, and is adaptable based on the needs of spatial regularization and the capacity for the reconstruction. We evaluated the method using MRF T1-FF, an MRF sequence for T1 relaxation time of water (T1H2O) and fat fraction (FF) mapping. We demonstrate the method's performance on a highly heterogeneous dataset consisting of 164 patients with various neuromuscular diseases imaged at thighs and legs. We empirically show the benefit of incorporating spatial regularization during the reconstruction and demonstrate that the method learns meaningful features from MR physics perspective. Further, we investigate the ability of the method to handle highly heterogeneous morphometric variations and its generalization to anatomical regions unseen during training. The obtained results outperform the state-of-the-art in deep learning-based MRF reconstruction. The method achieved normalized root mean squared errors of 0.048 $\pm$ 0.011 for T1H2O maps and 0.027 $\pm$ 0.004 for FF maps when compared to the dictionary matching in a test set of 50 patients. Coupled with fast MRF sequences, the proposed method has the potential of enabling multiparametric MR imaging in clinically feasible time. | ['Benjamin Marty', 'Pierre G. Carlier', 'Olivier Scheidegger', 'Fabian Balsiger', 'Mauricio Reyes', 'Alain Jungo'] | 2019-11-09 | null | null | null | null | ['magnetic-resonance-fingerprinting'] | ['medical'] | [ 3.19669634e-01 5.12701645e-02 -1.32001206e-01 -4.31815654e-01
-9.15870249e-01 -2.58620441e-01 1.19416453e-01 1.45449057e-01
-6.44596815e-01 8.51007342e-01 7.88432285e-02 -7.94890150e-02
-6.37880147e-01 -5.25406480e-01 -7.10020542e-01 -8.23717058e-01
-7.55499959e-01 7.82066584e-01 2.26890460e-01 2.50046421e-02
-7.30796158e-02 6.14504278e-01 -9.21171725e-01 1.00852899e-01
9.45079744e-01 1.04741013e+00 7.30064631e-01 2.39849597e-01
2.68676311e-01 6.64153397e-01 -1.57490358e-01 8.68596956e-02
2.11228877e-01 -2.06665650e-01 -7.58388937e-01 -2.22957179e-01
6.18359506e-01 -5.30119777e-01 -6.26306355e-01 8.04535508e-01
9.20445919e-01 3.26211810e-01 4.09377068e-01 -4.48654473e-01
-4.29900676e-01 5.83915949e-01 -5.54839432e-01 7.18509436e-01
-4.38047089e-02 2.30014533e-01 4.15884405e-01 -7.05426812e-01
8.76142144e-01 4.12678033e-01 9.82027471e-01 4.65946555e-01
-1.33820975e+00 -4.11224574e-01 -4.26885277e-01 1.10155456e-01
-1.16688633e+00 -3.23374599e-01 5.52385628e-01 -5.85278690e-01
9.04748201e-01 8.95720199e-02 5.21293044e-01 7.15636015e-01
5.88088989e-01 4.30051565e-01 1.40679073e+00 -1.61920473e-01
-1.03439735e-02 -3.44115436e-01 -1.15519516e-01 8.23763907e-01
4.83433604e-02 3.29602987e-01 -3.33840787e-01 -6.51267841e-02
1.25491571e+00 -2.87843555e-01 -6.62638485e-01 -5.35878778e-01
-1.66793799e+00 6.97806776e-01 4.35504973e-01 4.90178406e-01
-5.54549217e-01 2.54539698e-01 5.73074758e-01 1.24215856e-02
2.10263580e-01 4.70331669e-01 -1.50613293e-01 -5.58351912e-02
-1.10767031e+00 1.17780112e-01 2.01455474e-01 4.29764479e-01
2.60616224e-02 1.63455993e-01 -1.71131238e-01 1.01223874e+00
4.57755551e-02 5.21335065e-01 7.10881054e-01 -1.00163162e+00
2.17596948e-01 -2.26361081e-02 -6.00145608e-02 -8.63956511e-01
-1.06828511e+00 -6.82616591e-01 -9.48947906e-01 -6.57457625e-03
5.46230376e-01 1.12221226e-01 -8.73870194e-01 1.69376445e+00
4.77092564e-01 3.47517128e-03 -3.54410112e-01 1.50495338e+00
8.65411043e-01 7.25487024e-02 9.10572931e-02 -3.75803471e-01
1.07313454e+00 -7.01547682e-01 -6.86560750e-01 1.74963742e-01
6.38498902e-01 -5.98983645e-01 9.23436284e-01 3.70435238e-01
-1.23549175e+00 -4.83823240e-01 -8.85678411e-01 2.58973807e-01
3.57073992e-01 1.52470589e-01 8.23686540e-01 6.12762630e-01
-9.27064419e-01 1.02501512e+00 -1.11761391e+00 3.64626274e-02
3.72965485e-01 7.29737461e-01 -5.97599745e-01 -2.46086732e-01
-1.36620235e+00 1.08665049e+00 3.49399000e-01 1.87865824e-01
-8.99397075e-01 -1.33279753e+00 -6.60424888e-01 -3.88108552e-01
1.99869379e-01 -6.54698014e-01 8.37891042e-01 -4.23064977e-01
-1.50539804e+00 8.59606802e-01 2.51745105e-01 -5.63578069e-01
7.84631789e-01 5.67234196e-02 -7.35223472e-01 7.83493996e-01
2.05720216e-01 4.66980517e-01 6.84509575e-01 -7.81863987e-01
-5.58394417e-02 -5.28154492e-01 -2.31373727e-01 3.90795320e-02
1.93341509e-01 -7.70601779e-02 1.01485297e-01 -6.66624546e-01
3.85570884e-01 -8.51970434e-01 -3.57809961e-01 1.77991152e-01
-1.21661253e-01 5.04902542e-01 3.13649714e-01 -1.03636003e+00
8.77412438e-01 -1.81097412e+00 1.26241669e-01 3.57110173e-01
6.00145698e-01 1.01926282e-01 1.09225782e-02 6.06162064e-02
-2.88002074e-01 -2.08518028e-01 -4.83121485e-01 2.55422413e-01
-1.56264260e-01 7.61043280e-02 1.75610214e-01 1.02073896e+00
-1.85748696e-01 1.02382863e+00 -9.96731520e-01 -4.33612794e-01
4.96999800e-01 7.08405733e-01 -3.41708452e-01 7.97464848e-02
4.45229888e-01 1.13379896e+00 -4.03871298e-01 4.48739499e-01
7.72697628e-01 -2.44597465e-01 6.24583662e-01 -6.71664655e-01
-2.30521843e-01 1.66838020e-02 -8.59708369e-01 2.18062711e+00
-5.08878171e-01 2.81486958e-01 3.47086728e-01 -1.13194740e+00
8.08224320e-01 4.29950684e-01 1.24010897e+00 -1.23485565e+00
7.49246627e-02 6.45932078e-01 4.64202195e-01 -6.27894104e-01
1.34638473e-01 -7.60451972e-01 2.65068859e-01 4.31958497e-01
2.33527556e-01 7.32458830e-02 1.05623454e-01 -1.83121711e-01
1.15527260e+00 1.90287054e-01 -5.20766266e-02 -8.34454000e-01
6.18231833e-01 -1.42374367e-01 4.17403966e-01 8.12018514e-01
-4.66985315e-01 6.65852845e-01 -4.04688828e-02 -7.15678096e-01
-1.22913456e+00 -1.32210696e+00 -9.20944333e-01 6.00251615e-01
9.69647765e-02 1.04996465e-01 -4.85058159e-01 -4.66983676e-01
1.81129500e-01 2.80167460e-01 -6.55915916e-01 -6.76367152e-03
-1.01449859e+00 -1.08861279e+00 4.85366195e-01 5.34579158e-01
2.97148705e-01 -8.90433729e-01 -6.83834851e-01 5.01862347e-01
-3.49571913e-01 -1.21488750e+00 -3.37467998e-01 2.46911287e-01
-1.28187275e+00 -9.04407084e-01 -1.10735428e+00 -5.32805920e-01
4.07365412e-01 -9.54992175e-02 1.18373585e+00 -9.59481746e-02
-5.09298205e-01 3.71183097e-01 -8.80662799e-02 3.94522130e-01
-4.89850730e-01 1.24142103e-01 3.19256693e-01 -2.48731434e-01
-2.40818709e-01 -9.99334276e-01 -1.01960409e+00 4.08664048e-01
-8.31517696e-01 -6.76623266e-03 6.90186799e-01 9.97847259e-01
9.69449878e-01 -2.00077385e-01 6.42792702e-01 -8.76006305e-01
4.39053774e-01 -4.27105218e-01 -3.18900406e-01 2.10016236e-01
-6.34223104e-01 6.24997392e-02 5.85209966e-01 -4.76128638e-01
-7.70966530e-01 -7.50974864e-02 -1.86530411e-01 -3.89310807e-01
-3.58636864e-02 5.27838349e-01 4.39448416e-01 -7.09167182e-01
6.81069374e-01 4.63280261e-01 2.67025650e-01 -4.87947404e-01
1.16894111e-01 7.15507716e-02 8.34543824e-01 -8.54708910e-01
3.23288709e-01 5.93716204e-01 4.42068070e-01 -7.04529047e-01
-2.85368502e-01 -2.82742888e-01 -9.32779014e-01 -4.72988844e-01
6.95739150e-01 -6.14031136e-01 -6.89509392e-01 1.60267383e-01
-5.57962418e-01 -3.65277261e-01 -3.44537109e-01 8.58693898e-01
-8.37526739e-01 7.26150692e-01 -8.52682829e-01 -1.76165611e-01
-6.50922537e-01 -1.49912000e+00 6.94485784e-01 -1.89414874e-01
-5.15048765e-02 -1.17984617e+00 8.79682899e-02 5.11893213e-01
9.27097082e-01 7.01650739e-01 1.09915125e+00 -3.95299524e-01
-3.74961138e-01 5.87945245e-02 -1.24607682e-01 1.83306083e-01
8.72202590e-02 -9.13085639e-01 -5.74734151e-01 -5.37297606e-01
2.21669555e-01 -3.58368546e-01 6.35116458e-01 8.86765301e-01
1.22593498e+00 3.26405764e-01 3.48496102e-02 8.68488550e-01
1.52162778e+00 2.13553369e-01 7.36827075e-01 3.69771630e-01
7.96088755e-01 3.58063936e-01 3.66015583e-01 4.19165492e-01
1.64633870e-01 7.94504166e-01 2.74872065e-01 -2.20952168e-01
-4.04817134e-01 1.98746100e-02 -1.36877373e-01 1.09987950e+00
-3.72851729e-01 4.33697164e-01 -9.91422415e-01 4.86061931e-01
-1.37646198e+00 -5.83349824e-01 -1.75425291e-01 2.39338422e+00
7.84657955e-01 -5.84813878e-02 8.90230462e-02 -1.28385633e-01
6.65582478e-01 1.24914937e-01 -7.29753852e-01 -8.05614591e-02
4.22196351e-02 7.33730733e-01 7.19224989e-01 3.81196439e-01
-9.23201561e-01 2.51796126e-01 6.27187347e+00 7.34207809e-01
-1.41693521e+00 5.72770119e-01 4.80524391e-01 -1.51213706e-01
-3.30925971e-01 -4.00864273e-01 -9.07827243e-02 4.23617601e-01
9.73648250e-01 1.81879416e-01 6.62288666e-01 2.45261073e-01
2.88589239e-01 -1.39994621e-01 -7.66559005e-01 8.96922290e-01
-2.68621117e-01 -1.40470862e+00 -3.51012409e-01 -7.14047030e-02
4.28080946e-01 3.60524178e-01 1.74211502e-01 6.94134310e-02
-4.35806662e-01 -1.25983346e+00 4.71999586e-01 6.44008696e-01
1.35561502e+00 -6.04198217e-01 6.25813067e-01 3.44522074e-02
-9.44597185e-01 1.83224753e-01 -3.13204736e-01 3.73732269e-01
3.36962283e-01 7.29495406e-01 -8.30959320e-01 7.46447980e-01
6.18800759e-01 4.52131927e-01 -7.44437799e-02 8.92051935e-01
1.58708751e-01 3.10529590e-01 -2.89909869e-01 6.39335752e-01
1.72484949e-01 -3.25114995e-01 4.74123776e-01 1.08548152e+00
2.70934075e-01 1.63098082e-01 1.70030579e-01 8.85196447e-01
9.67672914e-02 3.08967263e-01 -2.38230810e-01 6.33782446e-02
1.06829770e-01 1.28141153e+00 -7.58252442e-01 -1.28732085e-01
-2.95578212e-01 6.50576591e-01 1.98205262e-01 2.12549239e-01
-8.95399034e-01 -1.63082480e-01 -2.57859100e-03 6.82759285e-01
2.29872689e-01 -2.96942830e-01 5.41520538e-03 -1.10338414e+00
1.52347594e-01 -8.23254645e-01 2.90717483e-01 -5.20880461e-01
-1.31388164e+00 6.98588967e-01 -4.45114896e-02 -1.00757122e+00
-9.14114043e-02 -5.06399333e-01 -1.14068806e-01 1.03403413e+00
-1.53598154e+00 -9.82639313e-01 -2.87228078e-01 4.41408038e-01
-1.24179311e-02 -2.96380054e-02 7.51624823e-01 8.58195722e-01
-2.13456124e-01 4.46622729e-01 3.40775788e-01 -1.39542043e-01
6.79838538e-01 -1.24170268e+00 -2.17827875e-02 3.62656236e-01
-3.61607045e-01 7.17410743e-01 4.43269372e-01 -6.61294520e-01
-1.54841387e+00 -7.96291113e-01 4.47867423e-01 -1.14904372e-02
6.88679099e-01 1.16834655e-01 -9.24770653e-01 3.93146604e-01
-1.86245784e-01 7.40945518e-01 6.89907432e-01 -3.28554630e-01
4.21272740e-02 -7.90771917e-02 -1.68433571e+00 -4.67056483e-02
9.30998206e-01 -5.44278562e-01 -4.50224131e-01 3.91277045e-01
4.11617845e-01 -9.59682584e-01 -1.92007053e+00 6.73496246e-01
8.62259805e-01 -9.48477983e-01 1.24268234e+00 -4.54839915e-01
3.16436440e-01 -1.35180444e-01 -2.08809748e-01 -1.05262530e+00
-4.67923015e-01 -3.61068517e-01 -6.76310435e-02 4.26275253e-01
7.01814666e-02 -6.87480927e-01 7.78080225e-01 5.53884029e-01
-5.34107089e-01 -9.27856565e-01 -1.38132739e+00 -6.46549225e-01
4.40678000e-01 -3.31690997e-01 2.64176965e-01 1.17513859e+00
-2.56054014e-01 -3.83125752e-01 -4.30974841e-01 1.08271264e-01
9.49970424e-01 1.40508518e-01 4.28575324e-03 -9.68357921e-01
-5.27248800e-01 -2.18741015e-01 -3.27020049e-01 -7.91168094e-01
5.69529459e-02 -1.28298247e+00 -2.18133792e-01 -1.29969132e+00
2.66547680e-01 -8.46961141e-01 -7.05935359e-01 9.37780961e-02
3.06048632e-01 2.96856523e-01 7.74666816e-02 3.62097442e-01
-2.63053566e-01 3.40379566e-01 1.97535884e+00 -1.05226628e-01
9.33374837e-03 -2.54317135e-01 -9.12436470e-02 3.17822486e-01
6.32439315e-01 -5.17868698e-01 -3.83459389e-01 -3.66242945e-01
-2.01710179e-01 7.12899983e-01 3.59110743e-01 -1.09175956e+00
-6.81066364e-02 1.90748185e-01 6.19459093e-01 -2.37675563e-01
1.87122032e-01 -7.07105994e-01 4.64953184e-01 7.37328947e-01
-2.74202585e-01 9.55569372e-02 2.75914043e-01 1.56266868e-01
-4.49513756e-02 -2.19179705e-01 9.73284006e-01 -4.39855874e-01
-5.74249148e-01 5.85493803e-01 -2.37176687e-01 2.39529237e-01
5.70974946e-01 -1.79678127e-01 -3.36619988e-02 6.63569868e-02
-1.16633534e+00 -1.47387117e-01 2.41841599e-01 2.91173048e-02
7.41713047e-01 -1.34054327e+00 -5.39977252e-01 1.90184802e-01
-2.85831720e-01 -2.88636506e-01 9.55348611e-01 1.75118530e+00
-8.51803720e-01 4.91382092e-01 -6.90863907e-01 -8.95660400e-01
-5.40304124e-01 3.88459176e-01 8.87765586e-01 -5.44364512e-01
-1.09700882e+00 5.09212852e-01 1.48988500e-01 -7.38743007e-01
-1.95695117e-01 -3.44934255e-01 3.14477384e-02 -4.43211555e-01
3.06826174e-01 4.19052035e-01 3.55409920e-01 -8.17825973e-01
-5.25650382e-01 7.28264093e-01 -9.34127495e-02 9.64550450e-02
1.48757970e+00 -1.86527252e-01 -7.51441270e-02 2.02209130e-01
1.12655759e+00 -2.67854631e-01 -1.23967409e+00 -2.70454943e-01
-2.25948896e-02 -5.03433108e-01 5.25493085e-01 -1.05469751e+00
-1.51148582e+00 9.00559068e-01 1.21570408e+00 -3.20092082e-01
9.51391339e-01 -2.05121174e-01 1.16955137e+00 -2.02901214e-01
7.15185285e-01 -7.46156156e-01 -3.12639356e-01 5.54109700e-02
7.75567055e-01 -1.16934681e+00 1.74036682e-01 -3.24525893e-01
-4.75099415e-01 1.32277763e+00 2.36835748e-01 -4.86563556e-02
3.23059142e-01 2.86859423e-01 9.08869877e-02 -5.25971174e-01
-9.43125635e-02 2.93271869e-01 4.40044433e-01 8.00619841e-01
6.16745710e-01 3.02674055e-01 -5.65159559e-01 3.79370034e-01
1.04060873e-01 1.86708897e-01 2.32944176e-01 8.44847083e-01
-1.19533718e-01 -1.01145697e+00 -1.58086762e-01 5.90432584e-01
-6.70218527e-01 7.68971890e-02 5.22420287e-01 1.01820672e+00
1.26634568e-01 1.92087650e-01 -3.00774992e-01 -5.06585389e-02
3.56550872e-01 -2.07531333e-01 1.09614050e+00 -2.09006682e-01
-7.90156782e-01 1.13361008e-01 1.07273376e-02 -8.76885653e-01
-5.41049659e-01 -6.40397012e-01 -1.52206528e+00 -2.06635788e-01
6.55588880e-02 1.44013297e-02 6.83481097e-01 8.26773465e-01
2.12433934e-01 6.43132985e-01 4.21473473e-01 -8.13995838e-01
-5.76727629e-01 -7.69345760e-01 -1.04017138e+00 5.36384165e-01
1.96063191e-01 -1.02420998e+00 2.35145632e-02 -4.53035384e-01] | [13.515033721923828, -2.37788724899292] |
4b94dded-b1af-4cb6-8519-353c7489fa90 | advancing-referring-expression-segmentation | 2305.12452 | null | https://arxiv.org/abs/2305.12452v1 | https://arxiv.org/pdf/2305.12452v1.pdf | Advancing Referring Expression Segmentation Beyond Single Image | Referring Expression Segmentation (RES) is a widely explored multi-modal task, which endeavors to segment the pre-existing object within a single image with a given linguistic expression. However, in broader real-world scenarios, it is not always possible to determine if the described object exists in a specific image. Typically, we have a collection of images, some of which may contain the described objects. The current RES setting curbs its practicality in such situations. To overcome this limitation, we propose a more realistic and general setting, named Group-wise Referring Expression Segmentation (GRES), which expands RES to a collection of related images, allowing the described objects to be present in a subset of input images. To support this new setting, we introduce an elaborately compiled dataset named Grouped Referring Dataset (GRD), containing complete group-wise annotations of target objects described by given expressions. We also present a baseline method named Grouped Referring Segmenter (GRSer), which explicitly captures the language-vision and intra-group vision-vision interactions to achieve state-of-the-art results on the proposed GRES and related tasks, such as Co-Salient Object Detection and RES. Our dataset and codes will be publicly released in https://github.com/yixuan730/group-res. | ['Rui Zhao', 'Feng Zhu', 'Xie Chi', 'Zhao Zhang', 'Yixuan Wu'] | 2023-05-21 | null | null | null | null | ['co-saliency-detection', 'referring-expression', 'referring-expression-segmentation', 'salient-object-detection-1'] | ['computer-vision', 'computer-vision', 'computer-vision', 'computer-vision'] | [ 0.26821464 -0.0370755 -0.16865744 -0.48081124 -0.86522514 -0.48586568
0.56338686 0.15766022 -0.4188552 0.28413746 0.14787894 0.09176537
0.36138752 -0.43165997 -0.72085875 -0.4931409 0.38306946 0.33414233
0.33845982 -0.29932952 0.10673436 0.32803497 -1.5586662 0.4072142
0.64067274 1.066105 0.60532784 0.3340344 -0.06007988 0.88266325
-0.50633496 -0.38228455 0.11654918 -0.47724637 -1.1103791 0.4938707
0.5646938 -0.23807022 0.0410067 1.3964826 0.32982537 0.42074454
0.46812117 -1.3121892 -0.8380925 0.6104829 -1.082208 0.24537472
0.35002664 0.10406183 1.1876924 -0.9803695 0.6913931 1.3093963
0.19707789 0.5636492 -1.0710843 -0.57529646 0.51069117 0.24437799
-1.5863518 -0.42250332 0.9636007 -0.3726913 0.5640846 0.32977465
0.6124 0.9929629 -0.44977757 1.2897867 1.1161938 -0.38541764
0.09940857 0.18306649 0.369591 0.531533 -0.20133317 -0.3594248
-0.4429258 -0.01594704 0.61834526 -0.01915658 -0.6038923 -0.3402191
-1.3435122 0.8056243 0.6069984 0.49543858 -0.46380565 0.11123592
0.44298032 -0.24609561 0.54388946 0.27698705 -0.24987052 0.00552307
-0.85380614 0.2796176 0.405704 1.1887263 0.9557345 -0.36978686
-0.42456332 1.0356632 0.14176556 0.17298627 0.2941288 -1.0726548
0.2531672 0.7526373 0.16810063 -1.0237538 -0.26636082 -0.2723815
-0.748303 -0.27604005 0.3438289 0.03782833 -0.77154666 2.0283504
0.5345888 0.13040704 0.02008948 1.3404741 1.1420611 0.6281537
0.36163816 -0.14898327 1.663655 -1.3889208 -0.67386156 -0.6009749
0.49131748 -0.6453229 1.3413588 0.04797448 -1.1228464 -0.5486564
-0.51303685 -0.38935298 -0.435666 0.25809467 0.6320489 0.10361257
-0.99833304 0.01227723 -0.57593775 -0.5509423 0.56163275 0.13549735
-0.31687158 -0.03344575 -1.104923 0.8059244 0.540674 0.21063101
-0.8225586 -0.62091327 -0.9945898 -0.2926401 0.8141614 -0.7460123
1.4058508 -1.279161 -0.98106307 1.550496 -0.43075454 -0.2525936
0.57838166 -0.24481034 -0.16197817 0.25282595 0.32663867 1.0855923
0.72100127 -1.572483 -0.6214048 -0.4299793 0.50248504 0.38759354
-0.08197211 0.5923936 -1.0494794 -0.57852906 -0.11225295 -0.8857842
-0.18801062 0.05820877 -0.79791236 -0.5029462 0.92333984 -0.5206233
1.0152704 -2.273144 0.27829245 -0.31709686 0.26656666 0.307335
-0.18629304 0.20037216 -0.06106988 0.15579505 -0.49674907 -0.7978009
-0.08632023 0.26107258 -0.29038674 0.39816993 0.2100348 1.0983261
-0.9823093 -0.7485995 0.20024197 0.3167519 -0.2269197 0.278119
-0.39843994 0.7905354 -0.6967656 0.83862233 0.60433584 -0.42283353
-0.37794274 -0.5057005 -0.03459717 -0.3046774 -0.99372876 1.8308705
-0.46078354 0.59416914 0.13634154 -1.0545973 0.8072733 -0.11186419
0.549118 -0.5588781 0.30261597 0.01829718 -0.227324 -0.6609547
0.7140247 0.08276922 -0.22205175 0.3130127 -0.03858379 -0.1193803
0.40706658 0.29705465 0.6437798 0.17717457 0.5878272 -0.14172575
0.6786991 0.16363053 0.5665568 0.6695101 -0.5399459 0.8502674
0.34643543 -0.12843566 -0.7599264 -0.75959915 -0.16895574 1.3243101
0.7360702 -0.2998868 -0.6710888 -0.58918685 -0.22800584 0.76137036
-0.7177278 0.27061436 -0.45180959 -0.623466 0.3711706 0.3685869
0.9787967 -1.2883376 -0.8779389 -0.11652165 -0.5683614 -1.5651052
-0.70448375 -0.08017851 -0.17878991 -1.0839076 -0.88771486 -0.83705467
0.912535 0.45644206 1.1817794 0.01311069 -0.14285155 0.598944
-0.5356026 -0.33390367 -0.32090777 -0.24872191 -0.2894031 0.3989455
0.3740335 -0.16844657 -0.6913264 0.4089078 -0.9121999 0.6581831
0.41262236 0.6546734 1.0815214 -0.28238958 0.29000053 -0.7275816
0.3950807 -0.41201156 -0.21879187 0.38932323 0.22021918 -0.46535954
0.4247634 -0.5926581 -0.79348266 0.19712614 0.07979333 -0.7575054
-0.40674773 0.5306243 -0.26272792 0.02373381 0.36552045 0.3126687
-0.12842812 -0.30736038 0.65700275 0.7594498 0.8603614 -0.6226126
0.40603143 0.58578575 -0.28989476 -0.7522115 -1.3509166 -0.8651612
-0.6446152 -0.29634443 1.016933 -1.0220479 -0.52044237 0.43637633
-1.4185678 -0.5316563 -0.16851237 0.0466441 -0.7441171 0.3937308
-0.32652864 -0.77799207 -0.3991147 -1.3924742 1.4947449 0.344048
-0.26167658 -0.86392576 -0.43325746 0.5781484 0.16183104 0.5858531
0.53490555 -0.8154477 -0.3222798 0.14298965 -0.5586789 0.33241668
0.232417 -0.12243096 -0.88023996 -0.14575113 -0.06248742 -0.4502802
0.79010755 0.26046148 1.406927 -0.12231332 -0.39162955 0.4753556
1.2580701 0.08549452 0.287644 0.30368024 0.9643886 0.92511463
1.0165681 0.2819957 0.6018577 0.985702 0.47577593 -0.5425061
-0.23074993 -0.11619698 0.12087063 0.32669267 0.00993259 -0.36326706
-0.9765021 0.78357404 -2.0753124 -0.8211367 -0.14450751 1.7498146
0.8767838 -0.4439271 0.2344693 -0.31608227 1.1566612 0.28151634
-0.7927515 0.00830171 -0.28489736 -0.28943792 0.17959225 0.15381402
-1.363562 1.1730806 4.9432926 0.9550153 -1.1187739 0.202878
0.9141596 0.19138114 -0.06250534 -0.1587787 -0.8261425 0.38024822
0.46571293 -0.39947715 0.2327118 0.8308043 0.45618248 -0.41393483
-1.0888481 1.2113382 0.36348078 -0.9349613 0.18494943 -0.2852341
0.64122 -0.12333337 0.07882316 0.12187234 0.06833749 -0.7578667
0.9798308 0.25813106 0.7197112 -0.54455256 0.54007524 0.41711375
-1.157051 0.08020014 -0.1461257 0.19737913 0.40477777 0.29276466
-0.5244808 0.566155 0.88874686 0.9746033 -0.7004051 0.8443974
-0.28160274 0.3754805 -0.28430474 0.12671018 0.45156533 -0.33218786
0.60414 1.1776276 0.08959926 0.29667765 0.35700056 1.3565699
-0.0559644 0.32260975 -0.35825467 0.07159774 0.390409 1.4973737
-0.8558667 -0.40802833 -0.56180876 1.175091 0.34380114 0.49321577
-1.0228851 -0.09401939 0.5365969 0.04572494 0.32066435 -0.06380387
0.1058915 -1.1540983 0.09856366 -0.7938784 0.46844822 -1.2161963
-1.2777097 0.6190922 0.38171574 -1.1825361 -0.07766744 -0.27471828
-0.45392883 0.76770324 -1.6081121 -1.6612822 -0.75713885 0.83026373
0.7999384 0.21607278 0.538326 0.24118257 -0.69384634 0.40127152
-0.45767486 0.22692266 0.7225508 -1.0439934 0.06241197 0.9788902
0.3024406 0.66753185 0.83353984 -0.45316038 -1.0324668 -1.3392183
0.5631565 -0.20640959 0.8435767 -0.38332707 -1.0449144 0.84487396
0.13715933 0.41762373 0.34968528 -0.33982438 -0.16540061 0.34048632
-0.9950157 0.8544426 1.2594873 -0.59253937 -0.63735723 0.5200676
0.816424 -0.84441334 -0.6160537 0.55733305 -0.07462236 -1.0098158
0.9096639 -0.35698813 0.5565576 -0.62518233 -0.21469258 -1.0706602
0.14828965 -0.56730515 0.09695281 1.5598762 0.03444153 -0.3677939
0.33614323 0.7148343 -0.3189663 -0.9178034 -0.8111552 -0.63540727
-0.1857327 -0.51402986 0.4322083 1.0668763 -0.3000064 0.3447361
-0.26721203 0.23102851 0.409582 0.49401048 0.7630686 -0.655535
-0.22501664 -0.41403303 -0.5651969 -1.3394766 0.51581997 -0.86886704
0.1024731 -1.5741785 0.582939 -0.3437443 -0.15782012 0.67528987
-0.4219682 0.49966404 0.38150847 0.3751768 -1.0991815 0.5710178
1.5403094 -0.3130984 -0.15012205 -0.10784565 -0.88734436 0.8717561
0.6227354 -0.08271115 -0.3392366 -0.33908004 -0.23475775 0.02185099
0.77824074 -0.56572264 0.25547215 -0.19837697 -0.13762261 -0.6634669
0.3623929 -0.6640301 0.12196374 -0.07919931 -0.5045938 -0.0917199
0.2831104 0.4266773 -0.5207195 -0.21946801 0.8485322 -0.12974845
-1.358383 0.41145298 0.07556021 0.2097468 1.3726163 -0.3554951
-0.34517744 -0.35472962 -0.72613573 0.5283535 0.56749153 0.534796
0.5612229 -1.1364121 -0.68596816 -0.38024753 0.64192027 0.3809082
0.53976506 1.026155 -0.103917 0.01051781 -0.07775138 -0.77950746
-1.575112 0.6409447 0.50157243 0.03016076 -0.79521453 0.89455897
0.81327444 -0.08455718 -0.04839035 -0.19935744 -0.45349497 0.04356033
0.5457226 -0.1573393 -0.34645125 -1.3333037 -0.2975537 0.77101314
-0.1463911 0.08076777 0.97907215 -0.49515414 -0.3744993 0.698203
1.2416992 -0.22474523 -1.226249 -0.5215236 -0.17418078 -0.45235953
0.02427073 -0.59021705 -1.196785 0.6203642 0.39223865 0.0484861
1.3689204 0.57863015 0.6445151 0.04621111 0.5222717 -0.85845137
0.21484096 0.33045286 1.2020947 -1.5288174 -0.17492169 -0.713872
-1.0599507 0.8807274 0.72963846 0.0693975 0.19991581 -0.0970208
0.19460967 -0.34253854 -0.44245893 -0.71466 0.26089615 0.56930894
0.28298265 -0.00833585 -0.20100138 0.59860504 -0.09823401 -0.20021415
0.4245319 0.72329676 -0.08762287 -0.76406705 -0.3772426 0.21190757
-0.41139764 0.0110325 -0.45842156 0.8149289 0.19936548 0.996464
0.21422331 0.0129615 0.36541146 -0.25981733 0.22652073 -0.6433981
-0.48269737 0.16255528 0.06941076 -0.58061355 -1.0845135 -0.78140545
-1.5066077 0.06911943 -0.19631022 -0.10883951 0.3614382 0.99266
0.2843684 0.51636666 0.3465826 -0.9937217 -0.01270752 -0.7717759
-0.5550232 0.8774767 0.24390449 -0.79722166 -0.16504493 0.35409483] | [10.1730375289917, 1.21395742893219] |
e56a505c-4658-4038-96b3-874c83708a77 | acgnet-action-complement-graph-network-for | 2112.10977 | null | https://arxiv.org/abs/2112.10977v1 | https://arxiv.org/pdf/2112.10977v1.pdf | ACGNet: Action Complement Graph Network for Weakly-supervised Temporal Action Localization | Weakly-supervised temporal action localization (WTAL) in untrimmed videos has emerged as a practical but challenging task since only video-level labels are available. Existing approaches typically leverage off-the-shelf segment-level features, which suffer from spatial incompleteness and temporal incoherence, thus limiting their performance. In this paper, we tackle this problem from a new perspective by enhancing segment-level representations with a simple yet effective graph convolutional network, namely action complement graph network (ACGNet). It facilitates the current video segment to perceive spatial-temporal dependencies from others that potentially convey complementary clues, implicitly mitigating the negative effects caused by the two issues above. By this means, the segment-level features are more discriminative and robust to spatial-temporal variations, contributing to higher localization accuracies. More importantly, the proposed ACGNet works as a universal module that can be flexibly plugged into different WTAL frameworks, while maintaining the end-to-end training fashion. Extensive experiments are conducted on the THUMOS'14 and ActivityNet1.2 benchmarks, where the state-of-the-art results clearly demonstrate the superiority of the proposed approach. | ['Di Huang', 'Jie Qin', 'Zichen Yang'] | 2021-12-21 | null | null | null | null | ['weakly-supervised-temporal-action', 'action-localization'] | ['computer-vision', 'computer-vision'] | [ 1.80044889e-01 -2.74380147e-01 -6.23524249e-01 -1.39782175e-01
-5.32182097e-01 -3.42236698e-01 5.77068388e-01 -5.64964488e-03
-4.18775350e-01 5.04940808e-01 4.28596944e-01 3.21792811e-02
-5.37748188e-02 -2.48417631e-01 -5.92430830e-01 -8.68901670e-01
-3.01354468e-01 -4.08708125e-01 6.65248334e-01 -2.64020283e-02
7.62739778e-02 1.15768529e-01 -1.25811958e+00 3.27816665e-01
9.54813361e-01 1.20898330e+00 2.48221591e-01 1.64051056e-01
2.54782736e-01 1.22660983e+00 -2.64886200e-01 -1.40207916e-01
1.60781190e-01 -3.90880823e-01 -4.37995523e-01 1.58302054e-01
4.53615218e-01 -3.57118070e-01 -7.54917204e-01 9.44350839e-01
2.85263330e-01 3.28215331e-01 2.23564833e-01 -1.42747331e+00
-5.02385259e-01 3.87716919e-01 -7.35691369e-01 3.17149639e-01
3.52173865e-01 4.24799234e-01 1.23063970e+00 -7.90482700e-01
4.74662304e-01 1.21049798e+00 4.36552763e-01 2.90364116e-01
-1.06989956e+00 -6.21649623e-01 6.63040459e-01 6.09270155e-01
-1.36889291e+00 -3.84035051e-01 9.99454677e-01 -4.09123123e-01
8.37664425e-01 -1.03445865e-01 6.13325715e-01 1.29342115e+00
1.15414903e-01 1.22488105e+00 8.68999541e-01 -6.63019568e-02
9.47514996e-02 -3.47777486e-01 -2.84173340e-01 7.65507698e-01
-5.24649099e-02 -5.77882901e-02 -7.38325000e-01 1.69843331e-01
9.51860845e-01 3.19459379e-01 -6.51999474e-01 -6.95956111e-01
-1.42348576e+00 4.75280821e-01 8.07348967e-01 5.75652480e-01
-3.86527896e-01 3.94225359e-01 7.08125651e-01 1.18497640e-01
4.59097445e-01 1.92841999e-02 -3.32266837e-01 -3.03394973e-01
-7.65275896e-01 -1.17236458e-01 2.23035663e-01 8.80884290e-01
6.55711353e-01 5.19861318e-02 -5.01726925e-01 6.10463798e-01
2.61692494e-01 -5.66896051e-03 5.43319106e-01 -7.75919616e-01
8.18119764e-01 7.62589812e-01 5.46947196e-02 -1.19878566e+00
-3.26647788e-01 -6.75128460e-01 -7.21376240e-01 -2.89261267e-02
4.04601038e-01 9.62311253e-02 -7.89664686e-01 2.04342532e+00
2.99174994e-01 7.85144031e-01 -1.42941877e-01 1.23923147e+00
6.58289909e-01 4.29542929e-01 3.68218422e-01 -1.55260935e-01
1.19152939e+00 -1.49687326e+00 -7.36702263e-01 -3.57857645e-01
7.88320601e-01 -5.35225034e-01 1.08190477e+00 2.71427482e-01
-6.82154298e-01 -8.19890559e-01 -1.11097372e+00 -4.16773511e-03
-1.93054989e-01 3.84108543e-01 6.14834547e-01 1.89908564e-01
-9.10732508e-01 6.55664802e-01 -1.07653701e+00 -2.61821687e-01
6.34140790e-01 1.52174294e-01 -6.16950274e-01 -3.49063575e-01
-1.08500433e+00 5.03142715e-01 4.99120027e-01 5.29971004e-01
-8.77157569e-01 -3.60930949e-01 -9.68037844e-01 -4.17289361e-02
9.23300862e-01 -3.42005998e-01 9.63577509e-01 -1.09852815e+00
-1.32218385e+00 2.93981194e-01 -1.31697774e-01 -3.63183230e-01
7.34102309e-01 -3.87033015e-01 -5.13396561e-01 4.70826954e-01
2.53602684e-01 4.84540343e-01 9.04027462e-01 -9.74772632e-01
-7.67478883e-01 -3.60066861e-01 3.60501617e-01 2.92179495e-01
-5.32137215e-01 -2.62836963e-01 -7.91779995e-01 -9.01176393e-01
7.94781595e-02 -8.35612476e-01 -9.69167277e-02 2.06489056e-01
-1.48268208e-01 -4.17586952e-01 1.05829787e+00 -6.32342935e-01
1.29263818e+00 -2.32550383e+00 2.77308106e-01 -1.95373371e-01
2.63010085e-01 3.65539044e-01 -3.91360372e-01 5.01808763e-01
3.22642922e-02 -1.76955059e-01 4.48639952e-02 -4.75269794e-01
-1.60814136e-01 1.67063624e-01 1.38008118e-01 6.92725420e-01
3.83675337e-01 1.02949297e+00 -1.29500294e+00 -6.32210970e-01
4.77235198e-01 4.85740632e-01 -3.43943954e-01 9.76872817e-02
-2.44012579e-01 8.21482360e-01 -6.86497152e-01 7.86212325e-01
4.36542541e-01 -3.19879621e-01 1.47197708e-01 -3.42577100e-01
-8.00983310e-02 1.33734211e-01 -9.88334119e-01 2.24931884e+00
-4.09027636e-01 6.15710557e-01 -4.48509902e-02 -1.19728875e+00
6.12853825e-01 3.41808856e-01 7.10804224e-01 -9.39293921e-01
-3.16095799e-02 1.51475459e-01 4.85668108e-02 -7.61603594e-01
1.84877351e-01 2.23720521e-01 -1.35795204e-02 5.00024334e-02
2.41061881e-01 7.58158982e-01 1.84545293e-01 2.08958447e-01
1.28175581e+00 6.97371125e-01 3.26619893e-01 -1.82999875e-02
7.22810328e-01 -2.89739519e-01 7.62448251e-01 4.64634627e-01
-5.45218587e-01 3.95135343e-01 5.48627734e-01 -2.51829803e-01
-4.50901777e-01 -7.77268767e-01 2.58745342e-01 1.23407662e+00
4.97322261e-01 -6.77440166e-01 -4.46861774e-01 -9.62608516e-01
-1.80747852e-01 2.90274680e-01 -7.59438515e-01 -1.80666268e-01
-7.05200970e-01 -1.54654205e-01 4.12670434e-01 8.29977751e-01
7.61017382e-01 -9.72197890e-01 -6.69546187e-01 3.25357258e-01
-5.78709424e-01 -1.57106352e+00 -6.82501376e-01 -4.42021303e-02
-8.54269147e-01 -1.05104935e+00 -8.25962961e-01 -6.29614890e-01
5.00924170e-01 7.33502328e-01 7.57811904e-01 1.43520772e-01
8.58036503e-02 1.27701685e-01 -7.72489429e-01 2.16833726e-01
1.86021879e-01 -4.93488908e-02 -1.46085590e-01 5.78071773e-01
2.18744799e-01 -7.89551675e-01 -1.10215104e+00 5.18418431e-01
-8.35093439e-01 7.11432695e-02 8.21707487e-01 9.11906660e-01
5.16054332e-01 9.06530768e-02 5.85080326e-01 -4.54468787e-01
1.03108123e-01 -4.50698853e-01 -3.02039206e-01 2.15299353e-01
-2.19873920e-01 -1.66833460e-01 8.34588110e-01 -5.19743204e-01
-9.79068100e-01 2.23389938e-01 9.59583744e-02 -8.09368610e-01
-1.08586699e-01 6.56677365e-01 -4.28957373e-01 -1.87529787e-01
3.22224975e-01 3.61815035e-01 -1.22916453e-01 -4.64270592e-01
3.91451925e-01 3.07906270e-01 6.65875971e-01 -3.30691844e-01
6.35699213e-01 5.57737350e-01 -6.92437887e-02 -6.25091672e-01
-9.23752785e-01 -8.24063361e-01 -8.26298892e-01 -5.53852618e-01
8.54008257e-01 -1.20137715e+00 -4.79820907e-01 4.78608489e-01
-9.38114703e-01 -3.88656437e-01 -5.66307306e-02 6.12390995e-01
-4.33929801e-01 6.79349780e-01 -3.68501604e-01 -6.29020095e-01
3.69188488e-02 -1.16938674e+00 1.16352499e+00 2.85074532e-01
9.71748829e-02 -9.16385174e-01 -2.30258003e-01 4.28027451e-01
1.18618108e-01 4.68063474e-01 5.20376682e-01 -4.76685673e-01
-7.86242843e-01 -2.25138634e-01 -3.71578574e-01 4.01825398e-01
2.35478386e-01 -2.95698851e-01 -9.52205956e-01 -4.64419633e-01
-1.56192720e-01 -3.55403155e-01 1.06476533e+00 1.46892935e-01
1.21288598e+00 -2.55263031e-01 -3.24072331e-01 5.15560269e-01
1.29218221e+00 -9.72028449e-02 5.77618659e-01 2.75872171e-01
1.11824596e+00 3.95467609e-01 1.02228892e+00 3.96177262e-01
4.26928341e-01 1.00834751e+00 6.49662971e-01 -1.81273103e-01
-2.43630424e-01 -4.85133767e-01 5.12898982e-01 6.05743527e-01
-3.06262881e-01 -3.27472627e-01 -5.37340999e-01 5.26883662e-01
-2.47768021e+00 -8.96393061e-01 -8.57888386e-02 1.91249657e+00
4.97389644e-01 2.36504212e-01 2.35110551e-01 1.57806918e-01
5.39152026e-01 7.53513753e-01 -5.41716397e-01 4.02115136e-01
-6.39321879e-02 -1.86493352e-01 3.46417397e-01 -9.92345065e-02
-1.33584356e+00 9.39492762e-01 3.93964505e+00 1.04344952e+00
-1.20539057e+00 2.51856208e-01 3.73932898e-01 -6.54786006e-02
1.90083280e-01 6.97947666e-02 -2.50253618e-01 6.65725172e-01
4.27185744e-01 1.72471017e-01 2.75995314e-01 7.64649570e-01
6.17401659e-01 -1.92390904e-01 -1.22452748e+00 1.06546712e+00
-2.27126945e-03 -9.94290113e-01 -2.46042699e-01 8.44475441e-03
6.31173730e-01 9.87584610e-03 -2.10250597e-02 4.28469747e-01
-2.94224977e-01 -7.85569906e-01 8.85210395e-01 3.40222418e-01
6.97844684e-01 -6.19632483e-01 8.64497185e-01 3.20015669e-01
-1.85732985e+00 -2.82337904e-01 -8.80509093e-02 -1.07830353e-01
2.80849367e-01 3.13329011e-01 -3.75286013e-01 8.69224489e-01
6.24450386e-01 1.31985700e+00 -7.54793286e-01 1.14329159e+00
-4.62037414e-01 5.13083041e-01 -3.08227018e-02 1.87256023e-01
6.69390678e-01 -1.63379125e-02 3.64672035e-01 1.14475572e+00
1.60470307e-01 6.67060465e-02 5.44552445e-01 4.67217773e-01
-8.39863867e-02 4.11778465e-02 -4.88620847e-01 -1.41460478e-01
2.77063847e-01 1.28762698e+00 -7.48266160e-01 -1.36197537e-01
-8.17604959e-01 1.20452809e+00 4.07977968e-01 5.24050057e-01
-1.05187273e+00 -1.47521183e-01 5.10342836e-01 7.35642537e-02
5.64287126e-01 -3.75824809e-01 1.50197610e-01 -1.39369440e+00
4.16808069e-01 -8.95576477e-01 5.29040635e-01 -5.65624833e-01
-1.12462199e+00 3.88167232e-01 -9.88185033e-02 -1.73492742e+00
3.51388124e-03 -4.07216161e-01 -5.68239629e-01 4.60263014e-01
-1.58814192e+00 -1.43447542e+00 -5.05592823e-01 7.94126928e-01
7.71424234e-01 1.13327138e-01 3.77938926e-01 5.11739075e-01
-7.72670388e-01 7.28441060e-01 -2.17912883e-01 2.59011090e-01
6.93541050e-01 -9.27606523e-01 2.92581078e-02 1.16494310e+00
3.04446131e-01 4.30637598e-01 3.51284862e-01 -4.88129050e-01
-1.46349943e+00 -1.34012341e+00 5.18511593e-01 -2.69149601e-01
8.48170161e-01 -3.70098680e-01 -9.25071716e-01 5.42048335e-01
8.62646699e-02 5.04956543e-01 3.37486088e-01 -6.48584887e-02
-4.33366537e-01 -1.91470876e-01 -6.47026896e-01 7.54231393e-01
1.50068545e+00 -6.36749387e-01 -2.42616653e-01 2.60011137e-01
7.02790856e-01 -3.10399860e-01 -7.33379662e-01 5.17692387e-01
5.38774610e-01 -1.04886663e+00 8.73136520e-01 -3.27795565e-01
4.79266167e-01 -4.89212334e-01 -9.24648996e-03 -1.06562769e+00
-3.94727767e-01 -6.63700342e-01 -4.35896605e-01 1.26845419e+00
8.87472481e-02 -5.32294393e-01 5.41466951e-01 1.13982275e-01
-2.33487234e-01 -1.01235986e+00 -1.16170251e+00 -9.36403692e-01
-5.02763212e-01 -4.82823938e-01 2.78250545e-01 9.45035875e-01
1.45000175e-01 4.19271737e-01 -7.39457965e-01 1.14708446e-01
4.22317237e-01 2.42970549e-02 6.95416629e-01 -8.39007735e-01
-3.55323911e-01 -4.29664195e-01 -7.82666624e-01 -1.53925097e+00
1.53358683e-01 -6.19482994e-01 2.41047964e-01 -1.27897918e+00
9.00534913e-02 -2.38131166e-01 -8.36797893e-01 5.83624065e-01
-3.08381140e-01 3.05855304e-01 3.19038719e-01 3.09412211e-01
-1.11334193e+00 8.72934937e-01 1.45378339e+00 -1.11883327e-01
-6.38815090e-02 -2.04846099e-01 -4.14077401e-01 7.96931684e-01
6.35800660e-01 -3.53999346e-01 -8.00778687e-01 -5.17630816e-01
-2.96353698e-01 1.59363583e-01 6.46651447e-01 -1.13630116e+00
2.73236901e-01 -1.59839705e-01 2.84202129e-01 -5.11410475e-01
3.75516683e-01 -8.52267861e-01 -2.36653432e-01 2.11445853e-01
-1.82032570e-01 -1.09826677e-01 5.03822640e-02 1.18018496e+00
-5.15627325e-01 2.28811860e-01 5.50022840e-01 4.35432680e-02
-1.24281859e+00 5.85696995e-01 -3.48113738e-02 6.76166732e-03
1.22969770e+00 -2.15278298e-01 -2.65082985e-01 -3.52338463e-01
-4.65918154e-01 3.55132908e-01 4.74756986e-01 6.69682741e-01
5.98861158e-01 -1.44483209e+00 -4.37947065e-01 8.84884298e-02
4.59691346e-01 -1.64466336e-01 6.63627505e-01 1.43877530e+00
-1.03724331e-01 3.82593900e-01 -1.29363522e-01 -8.83885503e-01
-1.02033687e+00 6.12465680e-01 1.36930898e-01 -3.06841254e-01
-7.74419606e-01 7.76898324e-01 4.52199072e-01 2.17369780e-01
4.21120167e-01 -4.21607852e-01 -2.99389750e-01 9.48166698e-02
3.61543208e-01 1.81820691e-01 -8.28619972e-02 -9.03265536e-01
-6.16879880e-01 5.04254818e-01 3.03972717e-02 7.51237944e-02
1.13108075e+00 -2.47157171e-01 2.67643720e-01 3.18266809e-01
1.20602381e+00 -2.28240997e-01 -1.73428297e+00 -6.00182354e-01
-7.72135481e-02 -7.23815858e-01 3.43189873e-02 -5.31242549e-01
-1.32018137e+00 9.57764387e-01 5.10509610e-01 8.69474337e-02
1.36225104e+00 -2.29925308e-02 8.42145085e-01 7.41392449e-02
4.91272628e-01 -1.00023472e+00 5.67936718e-01 1.40138417e-01
7.63026774e-01 -1.38121402e+00 -7.59918317e-02 -4.20633793e-01
-7.15432167e-01 1.00006032e+00 7.00171709e-01 -1.73119921e-02
3.49604756e-01 -2.39079535e-01 -1.33667037e-01 -1.12369575e-01
-6.58545673e-01 -3.55108231e-01 4.41967815e-01 5.27070045e-01
4.20778483e-01 -1.22885197e-01 -4.06794757e-01 6.12123370e-01
5.94126701e-01 1.34803683e-01 5.82713522e-02 1.05578494e+00
-5.34871966e-02 -9.35833097e-01 1.19287521e-01 1.27631024e-01
-3.93340349e-01 1.57084137e-01 -1.95352569e-01 8.87898445e-01
3.45631331e-01 1.07327127e+00 -3.75932157e-01 -6.13347828e-01
3.61114770e-01 -3.21225375e-01 3.17510515e-01 -3.69186729e-01
-1.89321548e-01 3.58593732e-01 1.40773907e-01 -1.00814855e+00
-8.58980119e-01 -6.09656334e-01 -1.18793201e+00 5.65435216e-02
-3.34182471e-01 -1.25868589e-01 7.66678676e-02 9.67100143e-01
4.63575959e-01 7.12880611e-01 5.86639166e-01 -1.06005156e+00
-4.28877860e-01 -9.63278413e-01 -4.63808537e-01 4.77613360e-01
4.65726078e-01 -1.02711785e+00 -1.37277588e-01 7.46565461e-02] | [8.48635482788086, 0.6669294834136963] |
f288b0fc-5bb5-4870-92a1-219372d88f24 | neural-enquirer-learning-to-query-tables-in | null | null | https://aclanthology.org/W16-0105 | https://aclanthology.org/W16-0105.pdf | Neural Enquirer: Learning to Query Tables in Natural Language | null | ['Zhengdong Lu', 'Pengcheng Yin', 'Hang Li', 'Kao Ben'] | 2016-06-01 | null | null | null | ws-2016-6 | ['learning-to-execute'] | ['computer-code'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.420971870422363, 3.9701995849609375] |
5992616c-6fe6-4314-a48d-6ee059056a6f | learning-multi-dimensional-edge-feature-based | 2205.01782 | null | https://arxiv.org/abs/2205.01782v2 | https://arxiv.org/pdf/2205.01782v2.pdf | Learning Multi-dimensional Edge Feature-based AU Relation Graph for Facial Action Unit Recognition | The activations of Facial Action Units (AUs) mutually influence one another. While the relationship between a pair of AUs can be complex and unique, existing approaches fail to specifically and explicitly represent such cues for each pair of AUs in each facial display. This paper proposes an AU relationship modelling approach that deep learns a unique graph to explicitly describe the relationship between each pair of AUs of the target facial display. Our approach first encodes each AU's activation status and its association with other AUs into a node feature. Then, it learns a pair of multi-dimensional edge features to describe multiple task-specific relationship cues between each pair of AUs. During both node and edge feature learning, our approach also considers the influence of the unique facial display on AUs' relationship by taking the full face representation as an input. Experimental results on BP4D and DISFA datasets show that both node and edge feature learning modules provide large performance improvements for CNN and transformer-based backbones, with our best systems achieving the state-of-the-art AU recognition results. Our approach not only has a strong capability in modelling relationship cues for AU recognition but also can be easily incorporated into various backbones. Our PyTorch code is made available. | ['Hatice Gunes', 'Linlin Shen', 'Weicheng Xie', 'Siyang Song', 'Cheng Luo'] | 2022-05-02 | null | null | null | null | ['facial-action-unit-detection'] | ['computer-vision'] | [ 1.55943632e-01 1.78630292e-01 -3.03841513e-02 -5.84225178e-01
-1.87200040e-01 -1.87813073e-01 5.50172508e-01 -2.22567990e-01
1.33121148e-01 6.07153736e-02 3.57794911e-02 3.64512026e-01
-1.77150462e-02 -1.01397872e+00 -7.27018833e-01 -7.75652468e-01
-3.46158832e-01 1.88279822e-01 2.88987253e-02 -4.57094610e-01
-3.63984585e-01 7.70876229e-01 -1.82272387e+00 5.77004969e-01
7.70587614e-03 1.67941558e+00 -1.38032064e-01 4.74979967e-01
1.60236567e-01 7.97238469e-01 -3.16563696e-01 -4.86673892e-01
3.21724683e-01 -4.12672222e-01 -4.38562751e-01 5.67422472e-02
9.35368657e-01 -4.66666937e-01 -6.33764207e-01 7.18708515e-01
6.68159842e-01 -2.40915194e-01 8.30524981e-01 -1.72292864e+00
-7.56415129e-01 3.42350870e-01 -8.21134210e-01 7.75076598e-02
3.65663320e-01 -2.49850787e-02 1.29016590e+00 -8.31972420e-01
4.98982936e-01 1.33343101e+00 4.80164021e-01 8.63727689e-01
-1.07746685e+00 -1.02141511e+00 2.64594108e-01 2.59868592e-01
-1.39077687e+00 -6.03657901e-01 9.93802547e-01 -3.03670079e-01
1.12286532e+00 4.34466116e-02 8.64352822e-01 1.07626343e+00
6.44206926e-02 8.62654030e-01 6.67678475e-01 -1.32805377e-01
-2.75236368e-01 -2.47867435e-01 1.91777587e-01 1.20511365e+00
-3.06316227e-01 3.39254998e-02 -8.81846845e-01 -2.95178831e-01
9.27504003e-01 -4.41871248e-02 -9.76815745e-02 -2.18504146e-01
-4.89693761e-01 5.07712126e-01 7.98062921e-01 2.62473941e-01
-3.69353235e-01 6.39061928e-01 8.27028900e-02 4.34988052e-01
1.80848196e-01 -1.70828402e-01 -2.01108843e-01 4.18512486e-02
-3.28030080e-01 -6.01139218e-02 7.08303273e-01 8.09384942e-01
1.21685541e+00 2.34609470e-02 -2.53226280e-01 9.09213960e-01
3.53569388e-01 3.66783261e-01 6.16032258e-02 -6.57649398e-01
1.22980237e-01 9.99413311e-01 -4.64337260e-01 -1.53875244e+00
-3.27066571e-01 -7.48297051e-02 -6.80365384e-01 3.98090780e-01
6.45259470e-02 -6.61689416e-02 -9.06904638e-01 2.06829000e+00
3.21167827e-01 6.46456420e-01 -7.94472471e-02 7.16471553e-01
1.41876113e+00 3.50206673e-01 -6.38584569e-02 1.79692194e-01
1.41425598e+00 -8.01917553e-01 -4.28442210e-01 -2.68518180e-01
6.35133445e-01 -4.77058560e-01 4.83001769e-01 1.82988290e-02
-9.80461001e-01 -5.13878405e-01 -9.13561583e-01 2.19211951e-02
-3.31932992e-01 3.42135549e-01 9.87301171e-01 2.83034533e-01
-1.36048734e+00 4.52935874e-01 -6.05800390e-01 -3.85065556e-01
6.90546453e-01 9.39101398e-01 -1.01289403e+00 -4.48836721e-02
-1.15468013e+00 6.37981951e-01 -3.28996271e-01 3.74282598e-01
-1.02582622e+00 -4.72771078e-01 -1.12660193e+00 9.90726650e-02
1.61240324e-01 -5.08711874e-01 9.01965141e-01 -1.25179517e+00
-1.11754668e+00 1.03975546e+00 -1.74329638e-01 1.80092290e-01
-5.84999770e-02 1.27713352e-01 -5.43285429e-01 1.69306681e-01
-1.86499894e-01 9.98406708e-01 1.12376988e+00 -1.30486107e+00
-4.33831394e-01 -6.59690440e-01 3.58723283e-01 3.54648352e-01
-6.87192678e-01 3.16058815e-01 -7.01474607e-01 -1.16524495e-01
-2.19819564e-02 -8.34340394e-01 3.02003324e-01 6.90409184e-01
-1.12026483e-01 -6.77904248e-01 1.08718014e+00 -1.18670210e-01
1.12460530e+00 -2.34720087e+00 1.56311616e-01 2.69271970e-01
3.67332399e-01 1.62602469e-01 -6.13666534e-01 3.66425455e-01
-2.47178346e-01 -2.34710887e-01 -1.19834460e-01 -4.45994318e-01
-2.28050083e-01 5.59735537e-01 1.92543656e-01 4.60268229e-01
6.40056074e-01 1.00494695e+00 -7.65783846e-01 -3.14241052e-01
-1.48159310e-01 9.02035475e-01 -4.28847253e-01 5.66800237e-01
1.84728816e-01 -2.98251864e-04 -4.41768140e-01 9.29560184e-01
5.80604732e-01 -8.79380405e-02 1.72384918e-01 -5.04311979e-01
3.94618928e-01 -1.08970240e-01 -9.68479991e-01 1.45222783e+00
-2.59822607e-01 5.95720232e-01 4.56711054e-01 -7.10641742e-01
1.35338521e+00 3.95492911e-01 6.69840932e-01 -5.86992323e-01
3.74762267e-01 -4.91581038e-02 2.35194221e-01 -3.64088953e-01
-1.24719463e-01 2.10212514e-01 2.09818199e-01 5.93277395e-01
2.79749125e-01 2.82141328e-01 -1.43129423e-01 1.22363903e-01
1.30718243e+00 2.60839593e-02 2.19916984e-01 -1.35246277e-01
5.72721541e-01 -8.32301974e-01 5.17880201e-01 3.19599271e-01
-2.64152676e-01 5.34039974e-01 8.46662939e-01 -5.35084784e-01
-3.83578718e-01 -8.66933703e-01 -1.60550922e-01 1.24981034e+00
2.07343455e-02 -7.22640991e-01 -6.26470327e-01 -8.85187924e-01
3.93031955e-01 -4.72971573e-02 -1.44462132e+00 -3.85989398e-01
-3.96424025e-01 -4.72951472e-01 7.04841375e-01 8.15471828e-01
5.39263964e-01 -1.20309460e+00 -4.63889241e-01 -1.44567862e-01
2.88078606e-01 -1.04209578e+00 -3.63828033e-01 2.95654684e-01
-3.79345745e-01 -1.43260038e+00 -2.71946132e-01 -9.83528018e-01
1.04748368e+00 3.44665319e-01 8.97931218e-01 5.69107533e-01
-4.67813432e-01 8.51096988e-01 -6.51461408e-02 -3.73102844e-01
5.93795814e-03 -3.37971419e-01 1.64706469e-01 7.12604821e-01
6.75400257e-01 -8.05889547e-01 -5.81755042e-01 6.03699863e-01
-7.70400107e-01 -3.32411110e-01 3.40894163e-01 8.58096600e-01
4.62791175e-01 -4.21034276e-01 2.78449863e-01 -6.69585586e-01
6.15233123e-01 -5.43291926e-01 1.02376351e-02 3.23852718e-01
-2.75884420e-01 -2.02953190e-01 1.43565133e-01 -3.74787092e-01
-7.96473503e-01 3.89212519e-01 1.24260224e-02 -7.63035059e-01
-2.02333078e-01 3.48802000e-01 -3.86860579e-01 -3.22467685e-01
5.58229983e-01 -1.34387344e-01 4.12953585e-01 -1.85476020e-01
1.48205116e-01 5.75493217e-01 3.60120684e-01 -6.63011611e-01
5.13562024e-01 5.57191432e-01 2.55517900e-01 -8.88956785e-01
-5.57711005e-01 -3.14413875e-01 -9.19584453e-01 -5.94307840e-01
8.54803503e-01 -9.93018687e-01 -9.74250793e-01 7.37950563e-01
-1.15745485e+00 -4.11490649e-01 2.35862270e-01 -1.47875041e-01
-3.13151419e-01 2.16004908e-01 -6.07376695e-01 -5.76907158e-01
-3.17927182e-01 -1.05987096e+00 1.35996497e+00 2.71249354e-01
-2.34112293e-01 -8.82946968e-01 -2.71950942e-02 1.15715995e-01
3.14807519e-02 3.36484939e-01 6.65821552e-01 -3.21606189e-01
-2.97271430e-01 -3.97079080e-01 -4.90917206e-01 2.76628017e-01
5.54859936e-01 4.01327103e-01 -1.14140797e+00 -2.49494359e-01
-3.85393769e-01 -5.80868483e-01 7.92883039e-01 -6.53414521e-03
9.02198553e-01 -6.88721538e-02 -5.90889871e-01 6.86984301e-01
1.00396085e+00 1.10228151e-01 7.13033140e-01 -1.27841130e-01
9.38063800e-01 7.95001447e-01 4.15764213e-01 4.20195967e-01
2.90770948e-01 9.55085993e-01 9.91013706e-01 -1.83990046e-01
-2.85865784e-01 -3.20603549e-02 6.79435253e-01 2.97120392e-01
-4.47924286e-01 -4.56563011e-02 -6.95830405e-01 3.22518259e-01
-1.90220916e+00 -8.42699409e-01 -3.52873071e-03 1.61473656e+00
5.82262814e-01 -2.81008869e-01 1.33904964e-01 -2.28817478e-01
5.10607779e-01 2.99562991e-01 -5.53652823e-01 -5.81493676e-01
-2.18759909e-01 1.82276979e-01 -3.82410474e-02 2.88475186e-01
-9.26258028e-01 1.09227085e+00 6.41363859e+00 3.37374717e-01
-1.15069532e+00 -1.62861377e-01 6.09609485e-01 -1.29329845e-01
-1.87214851e-01 -3.94703537e-01 -1.02372372e+00 4.54271026e-02
5.75906813e-01 2.31726885e-01 4.39633042e-01 7.93746352e-01
-2.62093633e-01 1.28799811e-01 -1.50034189e+00 1.22757280e+00
4.67225373e-01 -1.01782274e+00 1.95887312e-01 2.36119658e-01
4.66045856e-01 -6.08967505e-02 1.09931082e-01 -6.39197305e-02
2.62664378e-01 -1.38889205e+00 4.55631047e-01 4.81957674e-01
1.01652050e+00 -6.85331583e-01 4.66530353e-01 -1.32278636e-01
-1.58575010e+00 -1.39548853e-01 -4.09699112e-01 -3.87508363e-01
-4.13171411e-01 -2.20209941e-01 -7.44657397e-01 2.27210656e-01
9.36108828e-01 1.29067874e+00 -6.49971247e-01 3.42460871e-01
-5.11844754e-01 2.67625958e-01 -4.76658046e-01 -3.59608158e-02
3.08976293e-01 -2.14538887e-01 2.68571287e-01 1.03360248e+00
1.92208558e-01 3.84029329e-01 -1.45710111e-01 8.46220911e-01
-2.43489891e-01 3.24853063e-02 -1.03209066e+00 6.38754517e-02
3.05317849e-01 1.56848383e+00 -2.73454547e-01 4.47528176e-02
-7.46982932e-01 1.06937814e+00 9.62652743e-01 2.18809903e-01
-7.39269674e-01 1.62692085e-01 1.44436097e+00 7.08670989e-02
4.03158098e-01 -5.71509451e-02 2.76674688e-01 -7.75024474e-01
1.26855983e-03 -7.43302882e-01 6.10418200e-01 -8.27540636e-01
-1.45884466e+00 9.21986878e-01 -1.99640825e-01 -1.06462610e+00
-1.12794302e-01 -1.02666628e+00 -9.22988653e-01 7.77442694e-01
-1.27921057e+00 -1.72991490e+00 -5.56731820e-01 1.09495294e+00
1.20065464e-02 -3.52158010e-01 1.24202013e+00 2.85954028e-01
-7.96466947e-01 1.10593855e+00 -5.85379660e-01 5.92264056e-01
5.76439917e-01 -8.23514223e-01 1.29085287e-01 4.82779384e-01
4.58225071e-01 4.92798656e-01 8.55761170e-02 -5.02560556e-01
-1.45835602e+00 -7.43671358e-01 6.28092229e-01 -4.80771899e-01
8.36530924e-01 -5.92114449e-01 -8.55064631e-01 9.01330888e-01
1.36178434e-01 5.49094915e-01 9.44540083e-01 2.54068643e-01
-7.93092787e-01 -4.66750294e-01 -6.92630291e-01 6.22583449e-01
1.63767815e+00 -7.91274548e-01 -1.43257722e-01 1.50339305e-03
2.72602849e-02 -2.78657198e-01 -9.74222720e-01 5.00567794e-01
7.72877514e-01 -1.17667782e+00 8.19964290e-01 -8.61434042e-01
5.18896580e-01 -1.28681451e-01 -2.47360349e-01 -1.23278999e+00
-4.26598132e-01 -5.32250822e-01 -2.59621471e-01 1.22442269e+00
4.48820800e-01 -5.97596049e-01 9.04279232e-01 8.29358637e-01
-4.77393791e-02 -1.18563700e+00 -1.06206453e+00 -3.88441116e-01
-4.15036798e-01 -4.47549820e-01 7.43685961e-01 1.00591397e+00
1.21906452e-01 4.50019598e-01 -4.03437495e-01 2.06549183e-01
3.39422137e-01 1.32108971e-01 8.44747484e-01 -1.25669503e+00
-2.51775503e-01 -5.77539265e-01 -1.03084600e+00 -8.66032183e-01
4.57478821e-01 -1.02524340e+00 -2.02906027e-01 -1.49375570e+00
1.28883436e-01 -5.40660381e-01 -4.63972032e-01 1.25847745e+00
1.77265704e-02 5.05126476e-01 6.85813501e-02 -2.99784429e-02
-1.91138804e-01 6.39047086e-01 1.28608334e+00 -3.88196379e-01
2.26164144e-02 -2.09711075e-01 -6.26972973e-01 5.80306828e-01
4.18149650e-01 -1.96944058e-01 -4.55637634e-01 -5.64084768e-01
1.24081843e-01 -1.24914311e-01 4.80434418e-01 -9.31022823e-01
2.91660249e-01 1.51239648e-01 4.56308007e-01 -4.97613251e-02
7.56684065e-01 -1.14758003e+00 9.59245488e-02 1.76217467e-01
-3.57395783e-02 -3.27295214e-02 3.26864511e-01 6.04447961e-01
-3.22302431e-01 1.22682087e-01 5.18782139e-01 -7.44799599e-02
-8.52121174e-01 9.10540104e-01 -8.52007866e-02 -5.54033220e-01
1.18918550e+00 -4.93753225e-01 -1.43947482e-01 -6.56270385e-01
-7.80713499e-01 2.09090590e-01 2.66792923e-01 7.52544463e-01
8.88751686e-01 -1.56228030e+00 -7.14531183e-01 5.48327684e-01
3.81897777e-01 -1.93651795e-01 7.45989531e-02 6.79806590e-01
-4.19327465e-04 -1.23695042e-02 -6.82986498e-01 -6.39288723e-01
-1.89336383e+00 1.00631610e-01 7.32071221e-01 2.27068409e-01
-3.29449266e-01 1.35246038e+00 5.79382956e-01 -1.57187462e-01
2.85732210e-01 1.93899810e-01 -5.86473644e-01 3.95657122e-01
5.17399192e-01 -8.76090452e-02 -1.01760708e-01 -1.29842985e+00
-4.60636318e-01 8.04832876e-01 -5.76756820e-02 4.42886800e-01
1.39534700e+00 2.76732177e-01 -5.22503138e-01 2.09385231e-01
1.54025269e+00 -3.33083302e-01 -1.41282904e+00 -3.39320511e-01
-3.89810622e-01 -4.58973110e-01 -4.38813008e-02 -5.55512488e-01
-1.64053226e+00 9.50959384e-01 3.87664318e-01 -1.53448328e-01
1.29664183e+00 3.71840984e-01 5.32310605e-01 2.19959170e-01
3.53267819e-01 -6.52336836e-01 4.42370027e-01 3.16683859e-01
1.24398732e+00 -1.22257376e+00 -2.20532402e-01 -7.40742028e-01
-6.52327776e-01 1.32503974e+00 1.24928629e+00 -9.30889919e-02
1.00206792e+00 5.08350313e-01 3.02413315e-01 -8.12256575e-01
-8.94487202e-01 -2.87417561e-01 6.35995746e-01 7.43592322e-01
5.69594979e-01 -6.59893006e-02 1.93788335e-01 4.97626454e-01
1.26346737e-01 -3.70826542e-01 -5.16665056e-02 8.65899026e-01
-1.03029817e-01 -1.22388446e+00 4.70352732e-02 6.21069252e-01
3.72971334e-02 -6.09993637e-02 -1.09285295e+00 8.56273532e-01
1.67132184e-01 6.71760201e-01 5.54479003e-01 -9.06080306e-01
4.06400949e-01 2.76878905e-02 7.16664970e-01 -7.85346985e-01
-7.38661408e-01 -1.23562269e-01 2.02587873e-01 -9.54064310e-01
-5.16744137e-01 -7.81685054e-01 -1.40509081e+00 -2.27924287e-01
-2.25491356e-02 -3.15471202e-01 2.30185956e-01 6.74198985e-01
5.61640382e-01 2.56117702e-01 6.69100881e-01 -7.52607226e-01
9.92296413e-02 -8.91776562e-01 -7.55649090e-01 5.07427514e-01
9.52501073e-02 -1.11863482e+00 -2.54973948e-01 -2.85297811e-01] | [13.658355712890625, 1.5728684663772583] |
78626ec7-93a0-4baf-ae54-c53f1dc470ed | triple-consistency-loss-for-pairing | 1811.03492 | null | http://arxiv.org/abs/1811.03492v1 | http://arxiv.org/pdf/1811.03492v1.pdf | Triple consistency loss for pairing distributions in GAN-based face synthesis | Generative Adversarial Networks have shown impressive results for the task of
object translation, including face-to-face translation. A key component behind
the success of recent approaches is the self-consistency loss, which encourages
a network to recover the original input image when the output generated for a
desired attribute is itself passed through the same network, but with the
target attribute inverted. While the self-consistency loss yields
photo-realistic results, it can be shown that the input and target domains,
supposed to be close, differ substantially. This is empirically found by
observing that a network recovers the input image even if attributes other than
the inversion of the original goal are set as target. This stops one combining
networks for different tasks, or using a network to do progressive forward
passes. In this paper, we show empirical evidence of this effect, and propose a
new loss to bridge the gap between the distributions of the input and target
domains. This "triple consistency loss", aims to minimise the distance between
the outputs generated by the network for different routes to the target,
independent of any intermediate steps. To show this is effective, we
incorporate the triple consistency loss into the training of a new
landmark-guided face to face synthesis, where, contrary to previous works, the
generated images can simultaneously undergo a large transformation in both
expression and pose. To the best of our knowledge, we are the first to tackle
the problem of mismatching distributions in self-domain synthesis, and to
propose "in-the-wild" landmark-guided synthesis. Code will be available at
https://github.com/ESanchezLozano/GANnotation | ['Enrique Sanchez', 'Michel Valstar'] | 2018-11-08 | null | null | null | null | ['face-to-face-translation'] | ['computer-vision'] | [ 5.05133510e-01 6.11385167e-01 1.05456956e-01 -3.43805611e-01
-8.87545228e-01 -8.28666270e-01 7.21777797e-01 -5.93702018e-01
-1.56207800e-01 8.09573352e-01 -3.26649994e-02 -8.23644176e-03
1.17143698e-01 -7.93349624e-01 -9.97527421e-01 -9.09949481e-01
2.57946610e-01 6.13739312e-01 -4.15980034e-02 -3.05511624e-01
-1.15962319e-01 7.05153883e-01 -1.41184819e+00 2.80519724e-01
4.40909147e-01 7.47795522e-01 -1.61199182e-01 6.62773550e-01
1.85154572e-01 4.16358382e-01 -7.00063944e-01 -6.78083539e-01
6.41047716e-01 -9.36226785e-01 -7.76654482e-01 1.75181419e-01
7.84106493e-01 -1.83310285e-01 -2.25535378e-01 1.09609461e+00
6.16491079e-01 -7.28436187e-02 7.34333634e-01 -1.54007506e+00
-7.12640345e-01 3.42046618e-01 -3.92333746e-01 -4.27489311e-01
4.17853922e-01 2.72769541e-01 8.17160487e-01 -8.04544270e-01
9.49199438e-01 1.32681584e+00 5.37541091e-01 9.63306189e-01
-1.50633168e+00 -7.81233966e-01 -6.03206195e-02 -2.27806970e-01
-1.29727530e+00 -6.63504422e-01 6.95512712e-01 -3.24419141e-01
3.82006556e-01 2.75500923e-01 4.97783065e-01 1.38619673e+00
1.45112751e-02 4.17693079e-01 1.25595355e+00 -6.00493014e-01
6.37415275e-02 4.35154647e-01 -7.49187231e-01 6.06598198e-01
-1.26612380e-01 3.70515019e-01 -4.03586656e-01 1.11397253e-02
7.89204717e-01 -3.58030826e-01 -3.85415673e-01 -6.82780564e-01
-9.60098207e-01 7.68795609e-01 5.62048554e-01 2.83816338e-01
-2.16272607e-01 3.00498486e-01 8.57330635e-02 5.75379074e-01
4.36604053e-01 5.11144876e-01 -3.67654473e-01 2.51149476e-01
-9.63896513e-01 4.64980841e-01 7.29965150e-01 9.72606719e-01
7.68296063e-01 1.24530047e-01 -1.94139212e-01 6.16796434e-01
1.32254690e-01 5.01880050e-01 3.01319808e-01 -1.18725491e+00
3.92649502e-01 2.62080789e-01 9.07970518e-02 -8.41526926e-01
-3.03900205e-02 -3.29051971e-01 -6.36368573e-01 7.66990900e-01
9.32711780e-01 -2.46073499e-01 -1.07300711e+00 2.24434900e+00
4.24540460e-01 2.82617472e-02 1.14391036e-01 9.33255911e-01
4.18807387e-01 4.98003244e-01 -1.58688709e-01 -5.05961897e-03
1.07454431e+00 -7.11706102e-01 -4.14961845e-01 -2.94686705e-01
3.24576467e-01 -8.54433775e-01 1.07798052e+00 2.56136090e-01
-1.34322536e+00 -4.31785226e-01 -1.04545271e+00 -7.15307891e-03
-3.03551316e-01 1.73560437e-02 2.26982921e-01 6.26613736e-01
-1.30937254e+00 8.17692995e-01 -4.54678714e-01 -3.91521692e-01
7.00520456e-01 3.76964301e-01 -6.69353724e-01 -1.38954699e-01
-1.08719850e+00 9.24714088e-01 1.35339215e-01 3.25313509e-02
-9.47582304e-01 -7.43383646e-01 -6.16605163e-01 -7.71212950e-02
3.36191386e-01 -8.20400894e-01 1.14526927e+00 -1.81495917e+00
-1.59184563e+00 1.20454109e+00 -1.09280668e-01 -2.40038469e-01
1.06741822e+00 1.56428069e-01 -9.63890627e-02 -7.94822425e-02
1.30416170e-01 1.10826707e+00 1.24413157e+00 -1.52292669e+00
-2.67234355e-01 -4.70392883e-01 1.24227457e-01 1.30034909e-01
1.69389397e-02 -2.46644095e-02 -4.33520645e-01 -7.24065363e-01
-7.34484866e-02 -1.21992159e+00 5.63107524e-03 3.75130296e-01
-5.35896719e-01 4.11188118e-02 6.91705406e-01 -5.30728161e-01
4.24504489e-01 -2.25657797e+00 3.56344014e-01 3.77518773e-01
9.57562681e-03 2.15336576e-01 -4.74043816e-01 3.01983446e-01
-4.19567943e-01 1.11223228e-01 -5.77735126e-01 -5.00302851e-01
4.69699316e-02 2.30907455e-01 -5.60956120e-01 6.28585935e-01
3.79811466e-01 9.35264707e-01 -7.61473894e-01 -2.32615143e-01
-1.40289649e-01 7.24161983e-01 -4.53270525e-01 2.21192658e-01
-2.38090798e-01 6.73982084e-01 -2.24381331e-02 2.59380996e-01
8.30160975e-01 7.99434483e-02 1.07771367e-01 -1.51520491e-01
2.01108545e-01 7.82653876e-03 -1.13043582e+00 1.56926215e+00
-3.81377846e-01 6.51972711e-01 2.00383916e-01 -8.60105932e-01
7.97540009e-01 4.03620213e-01 2.32778370e-01 -7.83765435e-01
8.62357840e-02 4.52032626e-01 1.32962376e-01 -5.88139109e-02
1.76358148e-01 -5.87491810e-01 1.17997169e-01 4.34644341e-01
1.38407439e-01 -5.13734043e-01 2.41704676e-02 8.55162740e-02
7.67442167e-01 4.30331051e-01 -8.40137899e-02 -6.15940280e-02
4.27222669e-01 -8.42245519e-02 3.04381639e-01 4.65702415e-01
3.20163332e-02 1.15930033e+00 8.89870524e-01 -2.01943249e-01
-1.34793842e+00 -1.19168937e+00 4.66326959e-02 7.84468651e-01
-1.47192806e-01 1.11352369e-01 -1.08921480e+00 -8.70979488e-01
3.33496258e-02 7.35702217e-01 -8.76773238e-01 -2.86719978e-01
-5.41903317e-01 -2.61597753e-01 8.13215077e-01 2.70961672e-01
2.53868490e-01 -1.19425261e+00 -2.59206325e-01 -3.97128016e-02
-2.78138444e-02 -9.73088384e-01 -5.34290731e-01 3.76797654e-02
-6.45294726e-01 -1.02474809e+00 -9.43086267e-01 -5.98500490e-01
1.07207191e+00 -2.50586152e-01 9.87859964e-01 2.00378925e-01
-1.33407176e-01 2.12051153e-01 -7.04263374e-02 -3.94893646e-01
-7.54781306e-01 6.66250801e-03 -1.02539361e-01 1.93106100e-01
-2.07964942e-01 -7.42315292e-01 -4.24545795e-01 4.62203205e-01
-1.16383314e+00 2.21811458e-02 4.37115043e-01 9.04338419e-01
4.25891459e-01 -1.01190932e-01 4.21797782e-01 -9.54666913e-01
4.87519056e-01 -1.85848132e-01 -5.48554301e-01 6.87501114e-03
-5.19464672e-01 1.64634630e-01 8.48356605e-01 -5.73456466e-01
-7.65524626e-01 1.89350218e-01 -2.66963214e-01 -5.92590988e-01
-2.21936703e-01 -2.16238558e-01 -4.12887424e-01 -1.54097930e-01
7.32922316e-01 9.10719186e-02 3.36443573e-01 -2.25517973e-01
5.40474176e-01 3.07946682e-01 5.08781135e-01 -5.95518470e-01
1.18832481e+00 6.93965077e-01 1.67174667e-01 -3.97732764e-01
-6.81225240e-01 1.67358369e-01 -5.90144753e-01 -9.33917165e-02
6.37233853e-01 -6.37716115e-01 -5.17790258e-01 5.60702324e-01
-1.28865361e+00 -6.26409352e-01 -8.02106798e-01 1.00004412e-01
-9.20464694e-01 -8.16358030e-02 -1.58198863e-01 -5.15103638e-01
4.93123420e-02 -1.14117622e+00 9.82762873e-01 8.73626024e-02
-2.07790151e-01 -8.91511381e-01 6.25762390e-03 9.19382647e-02
4.18677241e-01 3.44942093e-01 7.79007733e-01 -5.50280333e-01
-5.56639135e-01 -2.48120755e-01 -2.09188193e-01 6.43204212e-01
1.41068205e-01 3.63570079e-03 -1.03432047e+00 -4.15874213e-01
-9.88653745e-04 -3.33332330e-01 6.87430620e-01 1.54623553e-01
8.06972861e-01 -4.82269764e-01 -9.36741456e-02 7.94860482e-01
1.36242843e+00 -3.84475687e-03 1.06119883e+00 2.25884616e-01
7.18752205e-01 8.99608970e-01 3.02357644e-01 1.95753220e-02
5.93933053e-02 8.98698986e-01 5.39419651e-01 -4.12596762e-01
-5.36292195e-01 -3.90702367e-01 4.59069520e-01 5.54111153e-02
4.00700271e-02 -3.68193239e-01 -6.22455478e-01 5.50140500e-01
-1.52232122e+00 -1.01071203e+00 2.01096922e-01 2.34665108e+00
9.03495729e-01 -3.79218087e-02 9.92778093e-02 -4.21103351e-02
6.77937448e-01 1.22626238e-01 -5.65685451e-01 -5.44616878e-01
-1.80886149e-01 4.68198746e-01 5.93159497e-01 8.12264562e-01
-7.55504966e-01 9.63229179e-01 5.97402334e+00 9.33301210e-01
-1.24750006e+00 2.36953497e-01 7.21014440e-01 -2.07801998e-01
-4.43855673e-01 2.73873180e-01 -5.39341629e-01 4.80618000e-01
6.71962738e-01 -1.98718589e-02 6.26697183e-01 5.60173213e-01
9.72469300e-02 8.60823691e-02 -1.23460567e+00 4.85688567e-01
1.70713156e-01 -1.15644383e+00 6.05404787e-02 1.35769501e-01
7.64156997e-01 -3.22838098e-01 3.87445569e-01 -4.85508293e-02
3.42142701e-01 -1.25923526e+00 9.09028471e-01 5.08982480e-01
1.08860111e+00 -7.65723944e-01 4.19297785e-01 2.01701641e-01
-6.11676872e-01 1.79699004e-01 -1.10480756e-01 2.25701004e-01
5.11587411e-02 4.26900655e-01 -8.78717303e-01 3.86539340e-01
3.49743247e-01 2.11178765e-01 -3.60540986e-01 5.69019139e-01
-4.85028118e-01 2.42295235e-01 -2.64821380e-01 4.15211588e-01
1.38655081e-01 -2.25482702e-01 7.92589068e-01 8.11115324e-01
5.25330722e-01 -3.61557096e-01 -2.03827739e-01 1.23287141e+00
-3.41553241e-01 -1.62536204e-02 -8.85823429e-01 7.53499717e-02
1.45795718e-01 1.10015571e+00 -5.20471334e-01 -1.10692136e-01
-2.04616226e-02 1.35129750e+00 2.46657148e-01 5.03311753e-01
-9.10575211e-01 -3.73860061e-01 6.68959498e-01 4.89384443e-01
3.61046135e-01 -8.49507563e-03 -2.41117209e-01 -7.23415375e-01
2.01449543e-01 -1.13521695e+00 -4.80869561e-02 -9.15619373e-01
-1.12412465e+00 6.21731758e-01 -6.47622794e-02 -1.14476252e+00
-4.52919453e-01 -4.43267643e-01 -5.69167852e-01 1.18611813e+00
-1.33082831e+00 -1.37233078e+00 2.85487082e-02 6.26412690e-01
2.72540480e-01 -1.05617650e-01 6.90587223e-01 3.03621411e-01
-2.32390821e-01 9.08337414e-01 -4.93020155e-02 1.21091641e-01
9.31046605e-01 -1.09173393e+00 4.76943552e-01 8.27413142e-01
2.52478957e-01 5.27688622e-01 8.92182410e-01 -3.90473813e-01
-1.11366820e+00 -1.05695879e+00 8.65846932e-01 -5.79958856e-01
3.72910231e-01 -4.47170794e-01 -7.61141539e-01 7.62556732e-01
2.98283964e-01 1.00338899e-01 1.20437676e-02 -3.48886371e-01
-4.65517640e-01 -1.35152474e-01 -1.42703223e+00 7.19420195e-01
9.94930863e-01 -4.83246416e-01 -1.08070709e-02 4.10643131e-01
5.57917058e-01 -5.53757429e-01 -5.46679676e-01 3.41692924e-01
6.59961104e-01 -1.13617241e+00 9.99987781e-01 -6.47642672e-01
6.09710038e-01 -3.52044970e-01 -7.93579668e-02 -1.53952551e+00
6.20679930e-02 -7.53682077e-01 2.59417683e-01 1.38216221e+00
6.17674410e-01 -8.41660321e-01 8.59285235e-01 4.25616056e-01
7.75378048e-02 -7.26798773e-01 -1.13600338e+00 -8.38879526e-01
5.08449554e-01 -1.61064222e-01 6.29770696e-01 7.23438323e-01
-5.65016866e-01 1.82916373e-01 -4.92052674e-01 9.41685885e-02
4.44200695e-01 -1.15838416e-01 9.61666226e-01 -8.21601868e-01
-3.72021765e-01 -5.13835073e-01 -3.33996475e-01 -7.86357105e-01
3.11041623e-01 -9.40106153e-01 1.68108150e-01 -1.23365593e+00
-1.50957733e-01 -5.84973097e-01 1.87787652e-01 6.23676896e-01
7.38347545e-02 5.74403405e-01 4.88250792e-01 1.37012199e-01
1.66372284e-01 4.00765449e-01 1.61132407e+00 -4.41317596e-02
-5.11222929e-02 1.03091314e-01 -8.18988442e-01 6.01832449e-01
9.61064219e-01 -7.97877371e-01 -4.32051778e-01 -3.13817948e-01
2.02127337e-01 -1.23594627e-01 6.24159098e-01 -8.34316850e-01
-2.33746264e-02 -9.16492715e-02 4.24153566e-01 9.21965092e-02
5.38928449e-01 -9.48200047e-01 5.21242738e-01 3.25445533e-01
-4.33175325e-01 -5.01058400e-02 1.30087763e-01 2.54100591e-01
-8.01578443e-03 -3.34195822e-01 1.12894118e+00 -6.35801852e-02
-9.08923745e-02 3.03844601e-01 3.03014647e-02 1.18830532e-01
1.06099761e+00 -3.14650834e-01 -2.11859196e-01 -6.44300461e-01
-7.14879751e-01 -2.13821620e-01 8.72896791e-01 3.46328795e-01
4.51105446e-01 -1.49517155e+00 -9.58346486e-01 5.24711907e-01
-1.13252312e-01 2.02456512e-03 5.14511131e-02 8.20964754e-01
-4.37641740e-01 -7.26608261e-02 -2.34457552e-01 -4.64976996e-01
-1.17931509e+00 4.23066527e-01 6.93829238e-01 -1.39468580e-01
-4.18165118e-01 8.62787426e-01 3.50870520e-01 -6.33199751e-01
-3.58884297e-02 2.20863387e-01 3.43384236e-01 2.99201775e-02
2.64248729e-01 1.08123526e-01 8.13420862e-02 -7.36232877e-01
-1.90570354e-01 6.67114735e-01 3.07407063e-02 -5.42366207e-01
1.08775461e+00 1.11276969e-01 -1.70445636e-01 1.22790560e-01
1.36416411e+00 2.42931023e-01 -1.50139761e+00 1.41899276e-03
-5.24599195e-01 -7.08028376e-01 -4.35197085e-01 -8.69742990e-01
-1.38939643e+00 8.84731293e-01 6.11433387e-01 7.29887560e-02
1.14352024e+00 7.89950788e-02 6.11606836e-01 -7.55377635e-02
1.19943857e-01 -8.25521171e-01 1.30388767e-01 3.04793626e-01
1.21193266e+00 -8.58448327e-01 -1.55078366e-01 -3.31091613e-01
-6.47949040e-01 8.76923800e-01 5.13711274e-01 -3.80322546e-01
2.16864914e-01 2.72934735e-01 2.35450819e-01 -3.36246677e-02
-4.36669797e-01 -1.04885206e-01 1.96040049e-01 8.28667462e-01
2.75965482e-01 -9.14832875e-02 5.59742749e-03 -1.89844355e-01
-3.54473442e-01 -3.81645560e-02 3.99640799e-01 5.96214294e-01
1.06026679e-02 -1.54571903e+00 -5.18873990e-01 1.10194534e-01
-4.75977272e-01 -7.83593208e-02 -6.67119741e-01 9.07641351e-01
2.98576713e-01 5.83795547e-01 1.13983899e-01 -3.67854163e-02
4.59708899e-01 1.92088231e-01 8.53895724e-01 -4.62815583e-01
-4.56846625e-01 -1.64981812e-01 -8.01517665e-02 -5.62170386e-01
-3.51034790e-01 -4.88373160e-01 -1.08509576e+00 -4.08845067e-01
-5.63015155e-02 -5.77021837e-02 7.09554076e-01 7.03758597e-01
3.66117299e-01 3.58872235e-01 8.45739007e-01 -1.06141579e+00
-5.57352722e-01 -7.17417061e-01 -3.50107700e-01 6.67091906e-01
4.24760312e-01 -4.65081364e-01 -6.40550613e-01 1.63166642e-01] | [11.79932975769043, -0.39382246136665344] |
be484c7b-7f2e-4c88-95c7-3dcd6f58f691 | few-shot-model-adaptation-for-customized | 2104.09457 | null | https://arxiv.org/abs/2104.09457v1 | https://arxiv.org/pdf/2104.09457v1.pdf | Few-Shot Model Adaptation for Customized Facial Landmark Detection, Segmentation, Stylization and Shadow Removal | Despite excellent progress has been made, the performance of deep learning based algorithms still heavily rely on specific datasets, which are difficult to extend due to labor-intensive labeling. Moreover, because of the advancement of new applications, initial definition of data annotations might not always meet the requirements of new functionalities. Thus, there is always a great demand in customized data annotations. To address the above issues, we propose the Few-Shot Model Adaptation (FSMA) framework and demonstrate its potential on several important tasks on Faces. The FSMA first acquires robust facial image embeddings by training an adversarial auto-encoder using large-scale unlabeled data. Then the model is equipped with feature adaptation and fusion layers, and adapts to the target task efficiently using a minimal amount of annotated images. The FSMA framework is prominent in its versatility across a wide range of facial image applications. The FSMA achieves state-of-the-art few-shot landmark detection performance and it offers satisfying solutions for few-shot face segmentation, stylization and facial shadow removal tasks for the first time. | ['Yu-Wing Tai', 'Weinong Wang', 'Bingkun Liu', 'Zhen Wei'] | 2021-04-19 | null | null | null | null | ['shadow-removal'] | ['computer-vision'] | [ 2.14782685e-01 3.07084993e-02 -1.58446670e-01 -6.26280725e-01
-7.42573917e-01 -1.99880913e-01 4.42801505e-01 -3.80270720e-01
-3.02431673e-01 3.28530133e-01 -2.66449153e-01 1.70871958e-01
9.14312303e-02 -4.78274018e-01 -4.43928927e-01 -6.46855116e-01
3.13838124e-01 5.27545631e-01 2.63626456e-01 -2.91960210e-01
-7.67644867e-02 6.41760945e-01 -1.79588830e+00 -6.76653087e-02
7.42232740e-01 1.15342879e+00 9.92911980e-02 2.42665827e-01
-3.79771829e-01 3.37863654e-01 -4.38295424e-01 -7.55933285e-01
2.69852877e-01 -3.97818387e-01 -5.87225974e-01 3.86727750e-01
3.90590400e-01 -3.71979743e-01 -1.68127045e-01 1.17539954e+00
6.44578457e-01 1.78360403e-01 5.60396969e-01 -1.49252379e+00
-7.35496640e-01 7.42447898e-02 -8.02171409e-01 4.90768999e-02
1.26908505e-02 7.36484081e-02 7.17163920e-01 -1.21288919e+00
7.03258097e-01 1.16798878e+00 5.63156545e-01 1.21435881e+00
-1.09292364e+00 -6.48512542e-01 1.32236481e-01 5.61378039e-02
-1.45038712e+00 -8.26711595e-01 8.54775310e-01 -3.69613022e-01
4.50184941e-01 -6.73824456e-03 4.94993627e-01 1.10493326e+00
-3.45748186e-01 8.16214263e-01 6.88852310e-01 -2.94683427e-01
4.33516681e-01 1.64819375e-01 -2.26512238e-01 9.86986816e-01
-8.63716900e-02 -3.06888163e-01 -5.61080396e-01 1.39548436e-01
5.26060760e-01 1.79287076e-01 -1.44368201e-01 -4.71265376e-01
-3.97480071e-01 7.50069797e-01 1.88387886e-01 1.99560836e-01
-6.23579696e-02 2.77710278e-02 5.12237489e-01 9.75017473e-02
5.41237414e-01 1.60366446e-01 -4.44099873e-01 -8.80106725e-03
-1.23232603e+00 -5.54540940e-02 5.15572309e-01 9.03358459e-01
8.57571304e-01 2.44814202e-01 1.90204103e-02 1.09199536e+00
2.04932600e-01 3.02315533e-01 6.10947728e-01 -8.48831713e-01
1.35153532e-03 5.81113517e-01 -1.72945112e-01 -7.84773707e-01
-2.67516315e-01 -9.79514271e-02 -7.16952085e-01 3.57944489e-01
3.28474402e-01 -4.25814912e-02 -1.26185787e+00 1.80018592e+00
6.46685719e-01 3.09080482e-01 -1.78739831e-01 8.44521999e-01
8.96719396e-01 4.51780945e-01 3.30075383e-01 -2.23283380e-01
1.30766833e+00 -9.64364707e-01 -7.98200548e-01 -3.64216000e-01
4.49666351e-01 -5.99119067e-01 1.13648665e+00 1.00729145e-01
-7.62938678e-01 -6.10410929e-01 -9.43068326e-01 -1.22703783e-01
-5.28688192e-01 6.95475638e-02 5.57162046e-01 8.59733045e-01
-8.48945498e-01 3.75529796e-01 -8.87607038e-01 -4.96367872e-01
1.00006771e+00 5.16244531e-01 -6.63619280e-01 -2.79819846e-01
-8.19524348e-01 5.43157816e-01 4.19551395e-02 1.73208237e-01
-9.78931725e-01 -7.18968213e-01 -1.17967129e+00 1.31992415e-01
6.80690885e-01 -3.51612866e-01 1.13786304e+00 -1.13250291e+00
-1.63978803e+00 1.09491909e+00 -9.51247290e-02 -1.19126141e-01
5.59976697e-01 -1.51130840e-01 -4.05780226e-01 1.42274007e-01
1.71453014e-01 7.70515323e-01 1.29651797e+00 -1.15592682e+00
-3.79381686e-01 -6.03696287e-01 -1.90165251e-01 -1.57348841e-01
-6.53846741e-01 2.24068642e-01 -8.48091841e-01 -5.02824008e-01
-1.88922107e-01 -7.01058149e-01 -2.12987050e-01 3.47620070e-01
-1.41657919e-01 -2.06867695e-01 1.26576948e+00 -2.89269269e-01
1.01698053e+00 -2.36681533e+00 3.55208144e-02 -1.65479288e-01
1.06292725e-01 8.04689646e-01 -3.74344021e-01 1.44826949e-01
2.84308940e-02 -9.19333380e-03 -5.06183386e-01 -7.89218962e-01
7.32731298e-02 2.47024730e-01 -2.20047385e-01 5.20666540e-01
5.39974332e-01 9.77434218e-01 -7.44200945e-01 -6.80217445e-01
3.14120829e-01 6.11015618e-01 -5.22832274e-01 2.84010708e-01
-3.98449183e-01 4.26397443e-01 -4.39740032e-01 1.05497539e+00
7.33049333e-01 -1.83534980e-01 -1.13655582e-01 -9.26647633e-02
1.29841208e-01 -6.94666147e-01 -9.65044320e-01 1.87705243e+00
-3.72940987e-01 5.18699110e-01 1.88348278e-01 -9.87182617e-01
8.03231120e-01 2.74859399e-01 5.72731376e-01 -4.92554903e-01
3.87396693e-01 1.52735844e-01 -2.31997803e-01 -7.92564690e-01
1.64339438e-01 -2.15352252e-01 3.10821254e-02 4.59906250e-01
5.73992014e-01 -1.34887388e-02 2.12525636e-01 1.37422919e-01
8.33134234e-01 2.64590949e-01 2.77437836e-01 3.22708525e-02
5.70151091e-01 -3.66461813e-01 8.38563800e-01 5.90532348e-02
-5.97193837e-01 7.87818193e-01 4.24255878e-01 -5.03949344e-01
-8.17376196e-01 -7.13154554e-01 -1.51000559e-01 1.52804112e+00
3.79667357e-02 -3.11995894e-01 -1.19891214e+00 -1.00036335e+00
-1.32280448e-02 2.92865217e-01 -9.78772879e-01 -1.11266330e-01
-2.21422553e-01 -5.90798676e-01 5.45896411e-01 4.94463593e-01
4.71145064e-01 -1.10308945e+00 -6.14786923e-01 -3.59290168e-02
3.37955117e-01 -1.36921477e+00 -5.17653942e-01 -7.03821033e-02
-3.75230342e-01 -1.23052537e+00 -8.05659235e-01 -8.10891807e-01
8.60327005e-01 2.67196238e-01 7.11344838e-01 1.07715733e-01
-7.49371290e-01 4.04625714e-01 -3.36479545e-01 -4.59636241e-01
-1.37547165e-01 -3.14843841e-02 7.43199959e-02 5.05000532e-01
5.65555811e-01 -3.07202011e-01 -5.29782832e-01 1.88887492e-01
-1.10430241e+00 -3.91942859e-01 4.02414113e-01 8.88591349e-01
5.61019659e-01 -1.37625366e-01 8.88448238e-01 -1.08761704e+00
2.52348095e-01 -2.73876786e-01 -5.90798080e-01 3.60227644e-01
-3.63726199e-01 -1.94607601e-01 5.93244493e-01 -2.90614575e-01
-1.27587831e+00 3.92584085e-01 -3.41699570e-01 -7.14747310e-01
-2.00625673e-01 5.77324629e-02 -5.72037399e-01 -3.16118598e-01
5.09783983e-01 -7.08500445e-02 2.51182467e-01 -4.21332568e-01
6.51402652e-01 6.96117282e-01 4.99273270e-01 -3.06177646e-01
7.65821517e-01 7.07810998e-01 -6.32800981e-02 -1.08829772e+00
-9.75691736e-01 -2.95746922e-01 -9.84671295e-01 -3.36743027e-01
1.01404083e+00 -5.54507494e-01 -3.50366652e-01 4.99635130e-01
-9.61538970e-01 -2.63706446e-01 -3.83546382e-01 -7.30386078e-02
-4.85873967e-01 3.43851715e-01 -3.35640579e-01 -8.31581593e-01
-3.61120641e-01 -1.35007679e+00 1.12979674e+00 4.77190256e-01
-6.44117221e-02 -9.35405433e-01 -3.01875174e-01 3.68194848e-01
2.44210184e-01 2.32330054e-01 7.98979998e-01 -6.76607668e-01
-4.46113855e-01 -4.42735910e-01 -1.18918628e-01 4.31473196e-01
2.76035517e-01 2.17342198e-01 -1.31737769e+00 -2.60084361e-01
-8.45039412e-02 -6.51898801e-01 8.18199933e-01 2.58335441e-01
1.42724526e+00 1.24666862e-01 -2.81394362e-01 8.73494208e-01
1.23609936e+00 3.07603572e-02 4.33771133e-01 -9.46880355e-02
7.15508938e-01 7.62947559e-01 6.61216617e-01 4.20110554e-01
7.94114247e-02 6.14300668e-01 5.70898473e-01 -1.29168078e-01
-2.14589849e-01 -2.28331208e-01 7.70138800e-02 3.38696152e-01
3.51677276e-02 -3.25864814e-02 -7.21706390e-01 4.81864542e-01
-1.73069620e+00 -8.85891259e-01 5.37832081e-01 2.02169204e+00
7.12125540e-01 -1.66755170e-01 1.46806628e-01 8.95299017e-02
7.27288485e-01 3.15182269e-01 -7.80606151e-01 -2.88975298e-01
1.77627847e-01 4.09440994e-01 1.09698281e-01 9.03466120e-02
-1.18487370e+00 1.42723227e+00 5.98635674e+00 9.91073072e-01
-1.23625267e+00 3.84611189e-01 8.23743403e-01 -1.56414285e-02
1.33373782e-01 -3.66243154e-01 -8.07099819e-01 4.89072293e-01
5.45961201e-01 -4.66644689e-02 3.30091238e-01 1.14711559e+00
-1.34199932e-01 7.51712993e-02 -1.03996658e+00 1.20517874e+00
5.20081282e-01 -1.15291584e+00 -5.16008809e-02 -6.50905743e-02
8.58933330e-01 -2.53954619e-01 3.73344094e-01 2.66199619e-01
-2.91473418e-02 -1.15183556e+00 4.30827916e-01 1.96860313e-01
1.21127808e+00 -8.19319248e-01 5.87446511e-01 1.44815624e-01
-1.04956448e+00 -1.52865186e-01 -4.39105153e-01 3.43633443e-01
1.94146231e-01 2.93942600e-01 -5.84311724e-01 1.30446330e-01
5.26713371e-01 4.42279816e-01 -4.18330938e-01 8.16953242e-01
-2.24171251e-01 2.77071655e-01 -1.07279949e-01 3.23527515e-01
1.77379712e-01 -1.18294150e-01 1.94271758e-01 9.42289829e-01
1.72768533e-01 1.51970789e-01 2.26554256e-02 6.19708955e-01
-4.92142797e-01 3.10572684e-01 -6.40462220e-01 -2.22312465e-01
4.55816686e-01 1.55363941e+00 -8.99232745e-01 -2.02423587e-01
-6.80284262e-01 1.15855348e+00 4.45924491e-01 3.13732594e-01
-7.43701041e-01 -3.63228679e-01 8.04303050e-01 1.24947608e-01
3.60421717e-01 5.63208945e-03 -4.13459577e-02 -9.71067369e-01
-2.52097577e-01 -7.61327207e-01 3.27825159e-01 -3.68038327e-01
-1.17282856e+00 6.88049197e-01 -3.53088111e-01 -9.40206945e-01
-1.88698769e-01 -7.08098650e-01 -6.15143478e-01 3.49446058e-01
-1.51172185e+00 -1.40324938e+00 -4.30665970e-01 5.87007761e-01
9.38954651e-01 -5.83333373e-01 9.79924381e-01 4.79433477e-01
-1.02177954e+00 8.62806976e-01 -6.51732981e-02 3.02824646e-01
8.57574403e-01 -8.56224060e-01 2.87055701e-01 8.39746535e-01
2.25823954e-01 4.30077791e-01 4.67911005e-01 -3.42602700e-01
-1.28724825e+00 -1.26815116e+00 4.41737354e-01 -2.57017404e-01
5.49799085e-01 -4.00440365e-01 -1.01003432e+00 6.82294905e-01
-8.08314681e-02 7.30161130e-01 9.46150720e-01 -1.22757576e-01
-4.61507916e-01 -2.98563778e-01 -1.38533843e+00 5.37999988e-01
9.91800010e-01 -5.71014464e-01 -2.77456850e-01 2.60951877e-01
4.90033746e-01 -2.20752433e-01 -5.31295538e-01 3.44074398e-01
4.97541428e-01 -9.10589695e-01 7.55934238e-01 -7.72209644e-01
3.72958899e-01 -9.83751267e-02 -9.47645307e-02 -1.05197847e+00
-1.92186385e-02 -7.10214913e-01 -1.82771191e-01 1.48255718e+00
9.56456736e-02 -4.12096292e-01 9.43575978e-01 9.78366971e-01
-6.91806599e-02 -9.03446198e-01 -1.02483284e+00 -5.94770491e-01
-2.92140484e-01 -2.49229819e-01 8.50077510e-01 9.01204586e-01
-3.00017506e-01 2.27013722e-01 -4.69236583e-01 4.67742421e-02
6.88758433e-01 6.33524433e-02 9.12141740e-01 -1.25966871e+00
-8.98740534e-03 -3.91549230e-01 -6.42403364e-01 -6.33986771e-01
4.99795407e-01 -6.63091242e-01 1.51838586e-01 -1.20106792e+00
1.26487955e-01 -2.48256549e-01 -2.62727410e-01 6.72815084e-01
-2.10133672e-01 7.31059194e-01 1.22637205e-01 1.31390672e-02
-7.60297477e-01 7.65992880e-01 9.73109663e-01 -1.23087019e-01
-1.11782841e-01 1.18180355e-02 -5.28636098e-01 9.68913198e-01
6.43274844e-01 -2.36412078e-01 -5.69864571e-01 -3.98960948e-01
-2.37925828e-01 -3.28712821e-01 8.48283172e-02 -8.94220054e-01
1.97843269e-01 -1.57253131e-01 4.34376121e-01 -1.10956624e-01
5.57038426e-01 -8.10793579e-01 -1.51609138e-01 1.30794689e-01
-2.58916393e-02 -3.92908335e-01 1.15810521e-01 6.75111115e-01
-2.49680623e-01 -2.98657984e-01 1.29949808e+00 -9.50785056e-02
-1.12824142e+00 8.71263564e-01 4.88687530e-02 5.34098782e-02
1.43768930e+00 -3.17784280e-01 6.64547607e-02 -8.42809007e-02
-7.50366151e-01 1.52992949e-01 6.08511269e-01 5.12516916e-01
6.16116822e-01 -1.18279278e+00 -2.19922066e-01 6.36779249e-01
2.64637262e-01 9.74299237e-02 5.06473303e-01 5.82138360e-01
-2.84679443e-01 4.88238633e-02 -3.08008403e-01 -4.38536048e-01
-1.40379083e+00 6.88298643e-01 2.90921152e-01 2.97297239e-01
-5.16299903e-01 1.03682315e+00 3.24489057e-01 -1.00089252e-01
2.35148370e-01 2.69217461e-01 -1.95603088e-01 3.45648646e-01
7.70057142e-01 2.74093002e-01 4.09501083e-02 -9.10569489e-01
-2.10976109e-01 7.07506418e-01 -1.85189992e-01 1.18432298e-01
1.39893901e+00 -1.17214791e-01 5.70760816e-02 1.54316425e-01
1.20270586e+00 -1.16295949e-01 -1.58011425e+00 -2.53009468e-01
-7.11074919e-02 -6.46827936e-01 -4.76958193e-02 -4.03843284e-01
-1.52293241e+00 1.18359995e+00 5.24838746e-01 -1.09041207e-01
1.16199064e+00 6.55055642e-02 9.52702582e-01 5.24495393e-02
4.08911884e-01 -1.28560030e+00 3.76240343e-01 2.04042524e-01
5.83273947e-01 -1.44212449e+00 -1.53312042e-01 -4.50353354e-01
-7.55036294e-01 9.50561881e-01 7.22591341e-01 1.32878855e-01
7.59555101e-01 2.26827264e-01 2.69821525e-01 -2.31137022e-01
-3.29777181e-01 -4.49698478e-01 1.09993950e-01 7.26463616e-01
1.70126408e-01 -2.35997587e-01 5.02444543e-02 7.07552373e-01
2.02027246e-01 2.75264066e-02 1.03785522e-01 6.75004661e-01
-5.15855432e-01 -1.12917531e+00 -1.09428972e-01 3.02676946e-01
-6.29117548e-01 2.53558427e-01 -2.53846228e-01 8.09785426e-01
2.40943208e-01 6.83511078e-01 -2.58119944e-02 -1.73352271e-01
8.38813186e-02 3.24847192e-01 4.90330458e-01 -8.62252355e-01
8.39489140e-03 -2.29559187e-03 -4.84807432e-01 -6.11968994e-01
-3.57702196e-01 -5.87489545e-01 -1.34652233e+00 -5.89354336e-02
-2.71336228e-01 -6.43220022e-02 7.40520060e-01 1.05212677e+00
5.08273959e-01 3.67487222e-01 5.56069255e-01 -9.02817190e-01
-4.69993562e-01 -8.50095987e-01 -6.76052153e-01 5.40262163e-01
1.76402465e-01 -9.50097799e-01 -1.05815127e-01 2.28200123e-01] | [13.479117393493652, 1.330198049545288] |
ba947728-56c3-4d8f-ac43-20b39bbe5ccf | comospeech-one-step-speech-and-singing-voice | 2305.06908 | null | https://arxiv.org/abs/2305.06908v1 | https://arxiv.org/pdf/2305.06908v1.pdf | CoMoSpeech: One-Step Speech and Singing Voice Synthesis via Consistency Model | Denoising diffusion probabilistic models (DDPMs) have shown promising performance for speech synthesis. However, a large number of iterative steps are required to achieve high sample quality, which restricts the inference speed. Maintaining sample quality while increasing sampling speed has become a challenging task. In this paper, we propose a "Co"nsistency "Mo"del-based "Speech" synthesis method, CoMoSpeech, which achieve speech synthesis through a single diffusion sampling step while achieving high audio quality. The consistency constraint is applied to distill a consistency model from a well-designed diffusion-based teacher model, which ultimately yields superior performances in the distilled CoMoSpeech. Our experiments show that by generating audio recordings by a single sampling step, the CoMoSpeech achieves an inference speed more than 150 times faster than real-time on a single NVIDIA A100 GPU, which is comparable to FastSpeech2, making diffusion-sampling based speech synthesis truly practical. Meanwhile, objective and subjective evaluations on text-to-speech and singing voice synthesis show that the proposed teacher models yield the best audio quality, and the one-step sampling based CoMoSpeech achieves the best inference speed with better or comparable audio quality to other conventional multi-step diffusion model baselines. Audio samples are available at https://comospeech.github.io/. | ['Yike Guo', 'Qifeng Liu', 'Jie Chen', 'Xu Tan', 'Wei Xue', 'Zhen Ye'] | 2023-05-11 | null | null | null | null | ['speech-synthesis', 'singing-voice-synthesis'] | ['speech', 'speech'] | [-2.82647014e-01 -3.81793343e-02 -8.38847086e-02 -1.19860418e-01
-1.32148516e+00 -3.36480409e-01 7.33673453e-01 -3.03015321e-01
-2.17047185e-01 4.38035458e-01 3.74398351e-01 -4.26403493e-01
2.31022164e-01 -6.42908275e-01 -5.32119513e-01 -1.01978195e+00
4.65186536e-01 4.10663486e-01 3.85984480e-01 1.70206770e-01
4.16494627e-03 2.67339170e-01 -1.37817013e+00 1.89188629e-01
1.08620620e+00 7.02582836e-01 6.96733117e-01 1.15045369e+00
-4.53155711e-02 7.83416867e-01 -9.79116976e-01 -2.95207858e-01
-1.96645692e-01 -6.65850639e-01 -3.66224766e-01 -2.46001687e-02
3.97531301e-01 -7.29475319e-01 -5.03184676e-01 1.30290341e+00
9.92964089e-01 4.41559404e-01 5.66206276e-01 -9.67515886e-01
-5.60530901e-01 8.90286148e-01 -3.31860274e-01 2.60387093e-01
1.69106409e-01 2.58223057e-01 7.81916618e-01 -9.91542995e-01
2.43034363e-01 1.66766560e+00 3.94618928e-01 6.69594049e-01
-1.03652525e+00 -9.70068157e-01 4.71927933e-02 2.48372689e-01
-1.49363005e+00 -8.91091645e-01 5.91523409e-01 -2.21859589e-01
9.91347671e-01 3.34628761e-01 5.29907882e-01 1.23671412e+00
1.83543220e-01 1.13376951e+00 1.00543058e+00 -2.32700333e-01
4.63126391e-01 4.27201986e-02 -2.34302133e-01 5.17442226e-01
-4.14281666e-01 8.90297219e-02 -8.51558506e-01 -1.67203352e-01
8.23752463e-01 -4.72862631e-01 -2.83303618e-01 4.32651460e-01
-1.14366806e+00 8.20992708e-01 7.39971697e-02 1.92498282e-01
-4.85198647e-01 3.25159907e-01 2.92041659e-01 2.10856825e-01
8.28345597e-01 -1.20525457e-01 6.51998669e-02 -7.83185363e-01
-1.42071962e+00 2.91245431e-01 8.36411893e-01 1.00256121e+00
1.06826156e-01 8.72291267e-01 -2.99082011e-01 1.09016490e+00
4.65838999e-01 1.12900782e+00 5.36767781e-01 -1.29748321e+00
4.40692574e-01 -4.41936374e-01 6.25198102e-03 -7.78148472e-01
1.56300932e-01 -3.44305396e-01 -1.23736572e+00 1.20132022e-01
2.29416162e-01 -3.44421476e-01 -7.40565300e-01 1.53805137e+00
6.79910243e-01 5.25885761e-01 1.68111145e-01 9.66618299e-01
8.58843088e-01 1.36134815e+00 -2.82852173e-01 -4.26540762e-01
1.27316880e+00 -1.30479503e+00 -1.17460883e+00 7.64045939e-02
3.48511338e-01 -1.23336363e+00 1.20852160e+00 8.60786557e-01
-1.47546589e+00 -6.61460996e-01 -9.97401655e-01 -2.63998304e-02
3.83542538e-01 2.18391150e-01 3.89007270e-01 8.84041607e-01
-1.13891900e+00 5.20719945e-01 -1.19810367e+00 1.53459355e-01
1.55451566e-01 1.95361495e-01 1.41013071e-01 2.45242193e-02
-9.09994125e-01 4.74113822e-01 -9.23279002e-02 -1.58838984e-02
-1.19830072e+00 -7.69785941e-01 -5.65600514e-01 1.42252773e-01
1.49896964e-01 -5.39688766e-01 1.72188246e+00 -4.61333632e-01
-2.46441817e+00 -8.22522736e-04 -5.62684357e-01 -5.16505539e-01
6.57208383e-01 -2.31301144e-01 -4.08366501e-01 4.05517489e-01
-1.84030041e-01 7.82044709e-01 1.39531541e+00 -7.59431899e-01
-6.34723127e-01 3.89179327e-02 -3.72721225e-01 3.24291289e-01
-5.70386887e-01 1.15725346e-01 -6.41450882e-01 -9.95749295e-01
1.72897857e-02 -9.15715635e-01 -2.75998786e-02 -5.13037704e-02
-2.47168809e-01 -4.42546397e-01 1.00932610e+00 -7.83323646e-01
1.42867041e+00 -2.18550062e+00 7.09151626e-02 -1.05243497e-01
2.12137282e-01 4.38590616e-01 -1.23643048e-01 2.76658833e-01
3.37820321e-01 -1.16521247e-01 -6.05329275e-02 -6.42224789e-01
2.75951643e-02 1.68296963e-01 -4.66524333e-01 3.63849968e-01
-1.70218125e-01 5.09564638e-01 -9.26676869e-01 -4.54290748e-01
3.45535874e-01 8.85708094e-01 -8.49309325e-01 3.00120324e-01
-2.23851502e-01 4.00703192e-01 -1.54937983e-01 3.26720506e-01
7.68834114e-01 2.74227886e-03 1.55443281e-01 -9.93579030e-02
-2.32790112e-02 6.48441970e-01 -1.39182699e+00 1.52864528e+00
-7.55400777e-01 8.91713977e-01 5.32572985e-01 -4.90510523e-01
1.05830503e+00 7.02491105e-01 5.75434789e-02 -5.19725502e-01
2.08923910e-02 3.02461356e-01 1.25839785e-02 -2.60958999e-01
6.06890976e-01 -6.70889094e-02 4.69210535e-01 4.48866487e-01
-2.66240630e-02 -9.65348423e-01 9.00058225e-02 3.62887263e-01
8.15056860e-01 -2.30633453e-01 -2.26557747e-01 -2.80048728e-01
3.39251518e-01 -4.49519336e-01 5.17995000e-01 7.44054019e-01
-1.63471505e-01 6.95076883e-01 2.36762375e-01 1.96504056e-01
-1.12104905e+00 -1.23601103e+00 -2.18333416e-02 7.06354260e-01
-1.71418458e-01 -5.52527785e-01 -1.14735067e+00 -2.57533729e-01
-4.14849728e-01 8.32370698e-01 1.13177925e-01 1.75260022e-01
-5.27284861e-01 -5.79647541e-01 8.22454870e-01 4.07182693e-01
6.36963904e-01 -8.13336492e-01 2.62088962e-02 4.10325587e-01
-3.66305918e-01 -1.11150968e+00 -9.76127744e-01 -2.99729288e-01
-9.13650155e-01 -2.53001988e-01 -1.11933613e+00 -7.23003864e-01
2.43883610e-01 2.91817665e-01 7.05125809e-01 -2.52628058e-01
2.61244595e-01 5.79915904e-02 -9.75304320e-02 -2.16146037e-01
-8.25157225e-01 4.50557284e-02 4.48760748e-01 -1.75206050e-01
-2.19109608e-03 -6.36336625e-01 -6.44349158e-01 7.09452704e-02
-8.55440259e-01 1.25062048e-01 4.08288538e-01 9.91295636e-01
6.36629701e-01 3.47310692e-01 6.05382800e-01 -3.13854605e-01
9.87713039e-01 -2.37765208e-01 -7.91774809e-01 -1.29775003e-01
-6.63033068e-01 -5.76057062e-02 7.43133545e-01 -9.14623320e-01
-1.30959928e+00 -2.68718868e-01 -8.58714461e-01 -7.33405948e-01
1.30764902e-01 2.98839509e-01 -5.74104115e-02 2.92096823e-01
4.61388052e-01 4.77274001e-01 1.59875765e-01 -5.01375854e-01
4.68859047e-01 1.15782201e+00 4.51908857e-01 -7.26288617e-01
4.10886794e-01 2.32279360e-01 -4.38420206e-01 -1.39478910e+00
-4.51616615e-01 -1.75858811e-01 -6.31525517e-02 -1.88515395e-01
7.28876829e-01 -1.09343672e+00 -7.88356066e-01 8.42115343e-01
-1.21499133e+00 -7.41394520e-01 -1.59410283e-01 9.46141124e-01
-4.04521048e-01 4.98142928e-01 -1.15578008e+00 -1.02707636e+00
-7.66718745e-01 -1.56697333e+00 1.11513758e+00 1.57519951e-01
-2.41370812e-01 -9.37406003e-01 -1.30839765e-01 4.05373365e-01
4.21215624e-01 -6.86840832e-01 5.45063853e-01 -2.42215633e-01
-5.38315594e-01 2.14058116e-01 -5.29340841e-02 6.83177531e-01
5.41402102e-02 2.54464865e-01 -9.69290435e-01 -4.00859892e-01
2.45679855e-01 -5.65828569e-02 4.29596126e-01 6.08497858e-01
1.04941201e+00 -5.06810427e-01 -7.93167483e-03 5.95382929e-01
7.96875656e-01 4.26693380e-01 4.79737133e-01 -3.40424895e-01
5.18216670e-01 1.80976555e-01 5.10824263e-01 5.72120786e-01
3.02803457e-01 6.28955185e-01 -1.96816057e-01 1.67361736e-01
-7.15243161e-01 -4.30039614e-01 7.04715073e-01 2.04080892e+00
3.85503709e-01 -4.57830101e-01 -7.13176250e-01 5.75340509e-01
-1.49765110e+00 -9.24535334e-01 -3.83705080e-01 2.07386851e+00
1.26179850e+00 8.03323388e-02 3.59801203e-03 2.09788531e-01
5.86372852e-01 3.32279265e-01 -4.19443011e-01 -5.21879733e-01
1.58342883e-01 4.76426750e-01 1.62949458e-01 1.01632440e+00
-6.30004764e-01 1.13188791e+00 6.19847775e+00 1.75294614e+00
-1.29298341e+00 4.86301988e-01 4.48279351e-01 -3.62928569e-01
-4.84298825e-01 -2.28205338e-01 -1.08705246e+00 6.85583115e-01
1.38133121e+00 -2.42593855e-01 5.70109129e-01 7.68872619e-01
7.78642774e-01 -1.59913208e-03 -6.64258063e-01 1.17234957e+00
-8.67466107e-02 -1.33250630e+00 2.24237293e-02 -7.33351260e-02
8.01571190e-01 1.18218958e-01 2.46963561e-01 2.56189823e-01
6.18860066e-01 -6.56370938e-01 7.82811761e-01 2.87650257e-01
7.81588078e-01 -8.91995370e-01 2.74085104e-01 6.35530651e-01
-1.23482382e+00 3.57272655e-01 -4.00429159e-01 1.01599336e-01
3.30805928e-01 9.89285827e-01 -9.97680426e-01 1.33053616e-01
7.26009309e-01 4.63944852e-01 1.06632642e-01 7.98867762e-01
-5.13176799e-01 1.29008353e+00 -5.52111745e-01 -3.61427754e-01
3.18141878e-01 -3.13766152e-01 7.46397555e-01 1.31255245e+00
7.43502915e-01 2.68940717e-01 -8.64791423e-02 7.61316955e-01
-1.16983503e-01 -4.01757471e-02 -2.76855171e-01 -1.64334401e-01
9.58984613e-01 9.93828952e-01 -4.00481910e-01 -7.02337623e-01
-1.15675770e-01 1.12379730e+00 -7.61916861e-02 3.86260033e-01
-1.08789575e+00 -4.86782014e-01 8.05848658e-01 -2.34502628e-01
3.32937956e-01 -6.29055202e-01 -2.11620450e-01 -1.10397363e+00
-1.99776709e-01 -1.33353007e+00 -1.90320253e-01 -9.09315228e-01
-9.43323076e-01 7.01658309e-01 -2.38744050e-01 -9.88259137e-01
-4.50072169e-01 -1.41372113e-02 -4.38684732e-01 9.99632299e-01
-1.23040426e+00 -6.72576010e-01 -1.32568568e-01 4.28689748e-01
1.23847866e+00 -1.22770809e-01 7.01051056e-01 4.74141419e-01
-7.34577715e-01 7.65773594e-01 4.53554422e-01 -3.53407472e-01
6.03380740e-01 -1.10462034e+00 7.11373746e-01 9.07218814e-01
3.11072916e-01 5.41704237e-01 7.32633114e-01 -7.56060719e-01
-1.54771292e+00 -9.44423199e-01 8.96369576e-01 2.27605570e-02
6.97730780e-01 -2.85028815e-01 -1.08387852e+00 1.11325167e-01
5.57194710e-01 -3.80775869e-01 2.76636451e-01 -2.74335533e-01
-3.66257057e-02 -1.49612486e-01 -9.50309277e-01 9.08410013e-01
7.68439591e-01 -7.88577974e-01 -2.57493705e-01 4.81750876e-01
9.12732363e-01 -7.11443901e-01 -1.03637624e+00 4.26745005e-02
3.56156230e-01 -7.43401766e-01 9.22043979e-01 2.64102548e-01
2.49276370e-01 -2.22950608e-01 -1.11460716e-01 -1.43722880e+00
4.35131714e-02 -1.28934693e+00 -5.02517104e-01 1.47436988e+00
2.15283975e-01 -5.86610138e-01 7.83379734e-01 2.14757904e-01
-1.53298244e-01 -5.77064395e-01 -9.86836195e-01 -9.66116011e-01
2.93638349e-01 -7.85992622e-01 5.54833233e-01 6.96935773e-01
-2.66048938e-01 4.21784312e-01 -3.91089052e-01 4.45956409e-01
7.63690233e-01 -2.94651270e-01 8.22263658e-01 -4.90359336e-01
-5.49045742e-01 -4.75233495e-01 2.63217777e-01 -1.99885857e+00
1.48654416e-01 -6.90987110e-01 2.56717116e-01 -1.45111454e+00
-2.43431970e-01 -5.36057651e-01 4.09394115e-01 -1.21091940e-02
-2.65099257e-01 -1.36188110e-02 1.17721431e-01 1.82421416e-01
-4.14636694e-02 1.04113913e+00 1.58222187e+00 -2.87083626e-01
-3.29824030e-01 2.08839744e-01 -1.38674140e-01 6.57420039e-01
7.71267176e-01 -5.34230113e-01 -7.38583505e-01 -7.62643397e-01
-3.11164647e-01 5.31832695e-01 -1.06292702e-01 -1.02887905e+00
5.98128855e-01 -3.02440803e-02 -1.03302009e-01 -7.13148832e-01
8.07896197e-01 -3.48070472e-01 5.40430332e-03 4.36279505e-01
-2.29388639e-01 4.60414626e-02 1.58221379e-01 3.76621574e-01
-4.23264503e-01 -3.21320444e-01 8.41089189e-01 2.28534937e-01
-2.36441001e-01 2.84073740e-01 -7.82317162e-01 4.97912951e-02
5.58011413e-01 2.22916707e-01 -7.57991970e-02 -8.64282429e-01
-6.55098557e-01 -1.20535240e-01 -1.02414619e-02 3.28096539e-01
8.02947342e-01 -1.24528050e+00 -8.56366456e-01 2.71722883e-01
-7.54058897e-01 1.24145932e-01 4.39692467e-01 7.44762003e-01
-5.64511657e-01 2.16654077e-01 5.60154557e-01 -9.28078592e-01
-1.50192976e+00 1.71804547e-01 1.51131272e-01 8.72346088e-02
-7.45388687e-01 1.13288927e+00 1.51442988e-02 -3.95773262e-01
4.80681419e-01 -4.86051351e-01 2.58807391e-01 -1.83379382e-01
7.26709485e-01 6.93412483e-01 1.06668815e-01 -2.83583105e-01
1.25623062e-01 2.00726703e-01 5.79981990e-02 -7.54971862e-01
1.00028956e+00 -1.97774723e-01 1.35619432e-01 1.91104010e-01
1.03342164e+00 3.57102841e-01 -1.45768762e+00 -1.98995844e-01
-4.95084435e-01 -5.60788155e-01 5.91969073e-01 -4.16305929e-01
-1.12361431e+00 1.30604851e+00 4.39438790e-01 9.85137448e-02
1.02936244e+00 -3.29351634e-01 1.29285991e+00 1.71148285e-01
1.12550803e-01 -1.11660409e+00 2.57484764e-01 6.52350724e-01
7.96389580e-01 -9.61090982e-01 -1.29054964e-01 -4.01418298e-01
-6.76209092e-01 1.03117919e+00 4.26022798e-01 7.40847960e-02
6.55864537e-01 8.75264049e-01 1.42818853e-01 3.26913625e-01
-1.00892735e+00 1.23369046e-01 2.22115085e-01 6.46641731e-01
4.01418477e-01 2.09813848e-01 -1.73402905e-01 4.06476706e-01
-5.92268527e-01 -1.88064411e-01 4.25853193e-01 4.47982341e-01
-4.82452631e-01 -1.18407702e+00 -5.87213635e-01 1.62991762e-01
-4.76851791e-01 -4.26265419e-01 3.24656069e-02 2.61875272e-01
-4.86137062e-01 1.49672461e+00 1.59132168e-01 -3.13082546e-01
9.49379802e-02 -1.57404125e-01 4.49222118e-01 -3.97519022e-01
-6.23518944e-01 7.00778842e-01 2.22865447e-01 -2.70052731e-01
-2.51441486e-02 -5.13558090e-01 -1.31829035e+00 -9.47978675e-01
-5.10295033e-01 3.32682461e-01 8.88863087e-01 7.40427434e-01
4.13977951e-01 6.63350940e-01 7.01917470e-01 -7.30082393e-01
-8.58066261e-01 -1.17816460e+00 -5.09777606e-01 -2.03909293e-01
1.89537957e-01 -2.10823759e-01 -4.26833719e-01 -3.55763845e-02] | [15.088391304016113, 6.222421646118164] |
236847f1-8f78-4e3f-ba5b-fdeb40f28da5 | 2110-06794 | 2110.06794 | null | https://arxiv.org/abs/2110.06794v2 | https://arxiv.org/pdf/2110.06794v2.pdf | The Layout Generation Algorithm of Graphic Design Based on Transformer-CVAE | Graphic design is ubiquitous in people's daily lives. For graphic design, the most time-consuming task is laying out various components in the interface. Repetitive manual layout design will waste a lot of time for professional graphic designers. Existing templates are usually rudimentary and not suitable for most designs, reducing efficiency and limiting creativity. This paper implemented the Transformer model and conditional variational autoencoder (CVAE) to the graphic design layout generation task. It proposed an end-to-end graphic design layout generation model named LayoutT-CVAE. We also proposed element disentanglement and feature-based disentanglement strategies and introduce new graphic design principles and similarity metrics into the model, which significantly increased the controllability and interpretability of the deep model. Compared with the existing state-of-art models, the layout generated by ours performs better on many metrics. | ['Xiaodong Xie', 'Dangqing Huang', 'Mengxi Guo'] | 2021-10-08 | null | null | null | null | ['layout-design'] | ['computer-vision'] | [-4.06033009e-01 7.64278928e-03 3.16707343e-01 -1.60005778e-01
1.45210713e-01 -6.64573133e-01 4.87556279e-01 -3.97955686e-01
1.57314599e-01 3.76956880e-01 4.01005715e-01 -4.42216009e-01
-2.69577950e-01 -9.06551063e-01 -3.08687210e-01 -3.00445735e-01
5.09170115e-01 5.42038023e-01 -4.44313139e-01 -6.66299090e-02
7.59181157e-02 6.75939202e-01 -1.51875901e+00 4.67201203e-01
9.55063581e-01 7.69570589e-01 3.49547118e-01 4.70608890e-01
-4.66686845e-01 5.28899908e-01 -6.16082907e-01 -7.00578511e-01
2.17902083e-02 -6.19031250e-01 -1.80330798e-01 1.65626198e-01
3.33604962e-01 -3.31482738e-01 -4.44927096e-01 8.21098864e-01
4.61076885e-01 3.17490339e-01 1.13839638e+00 -1.20127809e+00
-1.59218824e+00 8.65307152e-01 -5.18739522e-01 -3.93149167e-01
1.42399356e-01 2.30723396e-02 1.23031509e+00 -9.65195715e-01
6.00087762e-01 1.25576866e+00 5.85477948e-01 6.18782520e-01
-1.43044448e+00 -5.48813701e-01 1.40924662e-01 -2.13905368e-02
-1.17649019e+00 -7.29943365e-02 1.13198805e+00 -6.75684512e-01
9.04561639e-01 6.24423504e-01 8.95248055e-01 1.38010097e+00
9.36389565e-02 8.85313630e-01 7.43014574e-01 -3.07593167e-01
1.66544452e-01 2.81772763e-01 -1.06863163e-01 7.42590785e-01
4.91438627e-01 5.14195636e-02 4.70547937e-02 1.92037016e-01
1.46893930e+00 1.00431152e-01 -5.57969473e-02 -4.19864714e-01
-1.08794308e+00 9.25233066e-01 5.69596946e-01 5.71321845e-01
-1.78914428e-01 9.83840153e-02 2.04664171e-01 -5.32186925e-02
2.06161831e-02 1.06276286e+00 -1.87985286e-01 -1.07837051e-01
-8.93749833e-01 2.25369811e-01 6.64034724e-01 1.21010041e+00
1.97159111e-01 3.01316112e-01 -5.01463890e-01 7.41847515e-01
3.31045449e-01 4.90840405e-01 2.36562401e-01 -7.09658146e-01
2.71602750e-01 8.41850758e-01 5.11534140e-02 -1.51658070e+00
-6.08237505e-01 -6.78560257e-01 -1.30186367e+00 2.15716124e-01
3.12516689e-02 -2.43351698e-01 -7.78979957e-01 1.35905480e+00
-2.00949594e-01 -4.13583010e-01 -5.01400590e-01 9.16436672e-01
1.15404081e+00 6.66265428e-01 -1.44199384e-02 3.36126804e-01
1.53846347e+00 -1.15560544e+00 -1.12970150e+00 -1.67365730e-01
3.77271563e-01 -7.46078551e-01 1.46012855e+00 3.77360016e-01
-9.78082180e-01 -9.56719697e-01 -1.13095367e+00 -4.21480715e-01
-3.74426365e-01 8.02373528e-01 9.86972451e-01 8.20986390e-01
-7.00523734e-01 6.15854025e-01 -4.32368904e-01 2.45641638e-02
5.83301544e-01 1.66783392e-01 -2.76255548e-01 4.33168828e-01
-8.69425297e-01 8.13825428e-01 9.09271017e-02 3.01366895e-01
-7.16341957e-02 -7.81403124e-01 -8.00354421e-01 4.07184809e-01
1.64303124e-01 -1.31932771e+00 1.09835088e+00 -4.28940564e-01
-1.74388826e+00 2.89770007e-01 2.80462742e-01 1.53135002e-01
6.49874866e-01 -1.85053632e-01 -3.24295133e-01 -6.24971449e-01
-2.43018284e-01 6.97972596e-01 7.97792077e-01 -1.40898156e+00
-1.27089456e-01 1.28841266e-01 -6.66235387e-02 2.21851803e-02
-2.95347571e-01 -5.33281207e-01 -5.49478233e-01 -9.51581895e-01
-1.46779314e-01 -7.17112482e-01 -1.03829980e-01 1.32746130e-01
-6.99235260e-01 -4.66559194e-02 7.07026958e-01 -7.81895161e-01
1.74984753e+00 -1.96861601e+00 4.01129931e-01 8.51524547e-02
4.70281750e-01 1.44081518e-01 -6.48264214e-02 4.96859193e-01
-1.15037210e-01 4.14681137e-01 1.58751726e-01 -4.97209102e-01
5.45435667e-01 4.14649174e-02 -9.91230756e-02 -3.67258079e-02
8.22457892e-04 1.33231878e+00 -6.67463601e-01 -2.93580949e-01
5.79527974e-01 6.40230656e-01 -1.03625834e+00 3.02834630e-01
-3.93460304e-01 4.05070662e-01 -4.24490631e-01 2.50510484e-01
5.80157280e-01 -5.48142552e-01 1.97395846e-01 -7.08417773e-01
-1.13534831e-01 -6.95239380e-02 -1.09038758e+00 1.71228909e+00
-9.69557345e-01 1.08467090e+00 -6.05875671e-01 -4.49376583e-01
1.02234864e+00 -4.67024297e-02 1.21469520e-01 -1.05356944e+00
6.00115776e-01 -2.64517188e-01 5.11607639e-02 -8.27055693e-01
5.31602144e-01 1.37745991e-01 -3.08154404e-01 5.14549017e-01
-3.23132396e-01 1.20782647e-02 -3.21575962e-02 -1.99498132e-01
7.38599420e-01 3.57703745e-01 2.17503265e-01 -7.87074342e-02
8.76843333e-02 -4.54387248e-01 1.13424294e-01 5.47795594e-01
4.12997991e-01 7.76671529e-01 5.64022660e-01 -5.47306836e-01
-1.47740614e+00 -1.06232727e+00 2.02156588e-01 7.63930321e-01
-1.86570019e-01 -6.58009946e-01 -8.01688313e-01 -3.09013158e-01
-1.06380060e-01 1.04673433e+00 -8.40805888e-01 -1.82074234e-01
-4.77874726e-01 -2.25442126e-01 4.66811866e-01 7.27476120e-01
4.75123733e-01 -1.22366178e+00 -8.41566682e-01 2.57197134e-02
-1.56709850e-02 -6.07530177e-01 -5.67460060e-01 -2.50409096e-01
-5.64086676e-01 -6.11682534e-01 -8.77512336e-01 -8.77352953e-01
9.02525902e-01 -5.33661395e-02 1.16940522e+00 -1.45839751e-01
-4.46570098e-01 -1.71142325e-01 -2.01682478e-01 -1.64480746e-01
-1.43339723e-01 1.82416111e-01 -2.26077512e-01 -3.61849628e-02
-1.80555284e-01 -5.73090255e-01 -7.85543680e-01 1.42757431e-01
-7.34921992e-01 7.76416659e-01 8.59275520e-01 9.50016320e-01
3.18283200e-01 2.65350670e-01 1.34484559e-01 -9.81082439e-01
1.04132569e+00 -8.29529762e-02 -4.36347365e-01 2.87414372e-01
-3.59217107e-01 5.23887217e-01 1.00755227e+00 -5.51089883e-01
-1.17135417e+00 -2.63603717e-01 -9.55153033e-02 -6.22172117e-01
-7.67779797e-02 4.92480248e-01 -6.26168728e-01 3.71042401e-01
5.44543922e-01 -1.66335776e-01 -2.59426177e-01 -7.91059673e-01
8.46538544e-01 4.24871892e-01 3.57241154e-01 -3.73792320e-01
8.06466460e-01 1.76655147e-02 -6.33099526e-02 -8.26959133e-01
-2.41240337e-01 4.33540374e-01 -5.39194465e-01 -1.22911312e-01
8.55168819e-01 -3.43672633e-01 -1.24105167e+00 8.60554352e-02
-1.34227836e+00 -3.10389280e-01 -1.75321683e-01 9.31082442e-02
-3.22738230e-01 1.52791619e-01 -3.15876991e-01 -6.79373980e-01
-1.96836635e-01 -1.26937330e+00 1.12928009e+00 2.65091300e-01
-7.42992699e-01 -1.12791860e+00 -8.72251112e-03 9.81700271e-02
5.40143907e-01 4.76764798e-01 1.35276711e+00 3.86738293e-02
-4.91415590e-01 -2.50175893e-01 -5.24619401e-01 4.81099077e-03
2.62836576e-01 2.34960258e-01 -8.71699691e-01 3.05926770e-01
-2.90044487e-01 3.58141661e-01 6.28820360e-01 5.64182520e-01
1.32584190e+00 -6.14218831e-01 -3.49053532e-01 7.45688379e-01
1.30531228e+00 5.55208385e-01 7.23149121e-01 1.65407330e-01
1.35088301e+00 5.22013187e-01 -8.31108019e-02 4.85281140e-01
2.80110776e-01 7.49714196e-01 -1.29726321e-01 -1.70320049e-01
-2.94532120e-01 -7.33843029e-01 -1.94321260e-01 9.88793850e-01
1.02625914e-01 -6.58848286e-01 -6.96022570e-01 8.38549286e-02
-1.92088842e+00 -8.88060153e-01 -1.78348094e-01 1.57875383e+00
3.77732515e-01 3.68985310e-02 7.08535174e-03 9.44632962e-02
4.79210883e-01 -4.48965356e-02 -3.53411496e-01 -6.35823965e-01
9.87192765e-02 6.80891275e-02 3.38240415e-02 1.74687892e-01
-9.20470417e-01 7.69198835e-01 5.96800947e+00 7.19586015e-01
-7.60159671e-01 -2.69808322e-01 4.97515321e-01 2.48415396e-02
-7.87761092e-01 -1.53724149e-01 -2.51198620e-01 5.89284480e-01
4.97244716e-01 1.15559818e-02 5.37585735e-01 1.01311803e+00
2.86443681e-01 2.17101365e-01 -1.40602744e+00 1.60362029e+00
-4.18059453e-02 -1.53017116e+00 4.13216442e-01 1.28469169e-01
8.15122783e-01 -1.07923806e+00 5.58868706e-01 2.06112310e-01
1.02586195e-01 -1.34188926e+00 7.98795521e-01 8.72417390e-01
8.21305633e-01 -8.26834202e-01 3.55832100e-01 4.62681092e-02
-1.15475786e+00 -1.00606337e-01 -2.04366565e-01 -2.38883287e-01
2.88785636e-01 5.18571258e-01 -3.84757698e-01 4.47128028e-01
3.43225807e-01 4.90857154e-01 -7.62728512e-01 1.05235207e+00
-2.88932264e-01 3.62476893e-02 2.67565288e-02 -6.37792885e-01
1.34101138e-01 -4.62392688e-01 3.13588202e-01 1.11027813e+00
4.72932845e-01 -2.58892655e-01 -2.81916767e-01 1.58533680e+00
7.00851763e-03 -9.30436105e-02 -6.25168741e-01 -3.17850500e-01
5.01074135e-01 1.17693043e+00 -9.66806948e-01 -1.79421350e-01
-3.04511130e-01 1.14024878e+00 1.58284381e-01 4.00334060e-01
-1.23152876e+00 -7.71810591e-01 4.98780489e-01 -7.95313194e-02
3.57007444e-01 -3.73935431e-01 -1.02715230e+00 -1.15268004e+00
-1.89694818e-02 -8.51501882e-01 -1.66203514e-01 -1.14159310e+00
-1.29773712e+00 8.65229189e-01 -3.82862911e-02 -1.27029097e+00
-3.93818989e-02 -8.43939841e-01 -5.70693552e-01 8.23530376e-01
-3.18611681e-01 -1.27076936e+00 -3.91335666e-01 -1.21789306e-01
6.02694929e-01 -1.71510041e-01 8.19131255e-01 4.74101692e-01
-7.98850000e-01 9.46923733e-01 9.82609317e-02 1.75069988e-01
1.79453120e-01 -1.13769615e+00 1.04854202e+00 5.53969026e-01
4.62693512e-01 9.20826137e-01 8.22842717e-01 -6.36699200e-01
-1.48053992e+00 -8.08905125e-01 6.84032261e-01 -5.46450496e-01
3.93531412e-01 -7.73181975e-01 -6.09707177e-01 4.13151026e-01
2.56893665e-01 -8.40833426e-01 6.68406487e-01 2.47869700e-01
-3.51181537e-01 3.60004306e-01 -6.32783294e-01 1.37473547e+00
1.32859874e+00 -5.37974298e-01 -5.85353971e-01 -1.51996419e-01
6.57676458e-01 -2.85428613e-01 -5.35104930e-01 -1.02781102e-01
1.18391287e+00 -8.87738407e-01 8.70849609e-01 -4.50447559e-01
6.08215094e-01 -3.03216994e-01 2.00296670e-01 -1.46762621e+00
-1.07717037e+00 -5.43459833e-01 -1.49398789e-01 1.22446001e+00
5.36773920e-01 -2.04547748e-01 6.62111998e-01 8.29709828e-01
-1.86558634e-01 -5.82736075e-01 -2.92860538e-01 -6.89419746e-01
-1.71565086e-01 -5.70316315e-01 1.15999055e+00 7.83136070e-01
-7.91272223e-02 7.45006740e-01 -5.38749635e-01 -2.79816717e-01
4.15172428e-01 3.30510765e-01 7.61730373e-01 -1.38240182e+00
-5.85981190e-01 -1.09159529e+00 -2.28105187e-01 -9.48034406e-01
-3.73770073e-02 -8.96242142e-01 -2.92563409e-01 -2.01230288e+00
-7.28669241e-02 -1.49069279e-01 2.17139035e-01 2.70184249e-01
1.48352325e-01 -6.74261302e-02 3.75675321e-01 -2.74945498e-01
-1.94371998e-01 7.65787244e-01 1.68960059e+00 -3.46897274e-01
-4.70896482e-01 -2.13454396e-01 -9.86661434e-01 4.68928874e-01
5.71911693e-01 -3.71777527e-02 -6.63178563e-01 -8.01175892e-01
8.23244929e-01 -7.32017010e-02 2.75638372e-01 -9.04009283e-01
1.32126585e-01 2.11878102e-02 9.03436363e-01 -6.24482393e-01
1.36672601e-01 -9.91398036e-01 8.08712542e-01 2.42507353e-01
-4.20233071e-01 1.69941828e-01 4.15485978e-01 2.58119792e-01
3.15579236e-01 3.31178494e-02 4.48180944e-01 2.45712902e-02
-3.83821517e-01 1.39929503e-01 -5.26454389e-01 -4.22563702e-01
5.83577275e-01 -4.84910399e-01 -3.03153574e-01 -3.60441834e-01
-8.54484618e-01 -1.61391526e-01 3.53126794e-01 7.91575789e-01
6.99428082e-01 -1.79166663e+00 -2.30272979e-01 5.25553584e-01
-4.60349470e-02 -2.11929917e-01 6.97826087e-01 2.25142434e-01
-8.07285607e-01 6.47538006e-01 -5.76594472e-01 1.74278561e-02
-8.55334103e-01 9.04161870e-01 9.21339765e-02 -9.36163962e-02
-9.07356143e-01 7.79540420e-01 4.87544000e-01 -2.07146570e-01
2.20154569e-01 -8.10861290e-01 -2.75991350e-01 2.26927713e-01
4.41546172e-01 6.45144820e-01 -2.01695323e-01 -3.02359372e-01
-8.04358050e-02 7.82510877e-01 9.76116061e-02 -2.94596016e-01
1.20914972e+00 1.20979592e-01 1.59917027e-01 4.49278623e-01
1.32618511e+00 -3.23751807e-01 -1.05968714e+00 4.83860433e-01
-1.02549709e-01 -4.15252328e-01 1.67412207e-01 -8.68308008e-01
-1.10768652e+00 1.01375914e+00 4.77373987e-01 3.20343047e-01
8.52505982e-01 -3.49868804e-01 6.76033556e-01 4.80331153e-01
3.05400789e-01 -1.01761627e+00 2.19744056e-01 2.36459628e-01
1.31127584e+00 -7.00957179e-01 -1.06029294e-01 -1.67560667e-01
-8.68182361e-01 1.19434237e+00 5.91624141e-01 -1.90418914e-01
5.93261778e-01 3.40246439e-01 -1.80529624e-01 -3.87325644e-01
-5.15238285e-01 4.91578467e-02 8.88002872e-01 6.36487007e-01
5.93735099e-01 2.60752231e-01 -5.33544458e-02 1.18113911e+00
-6.39807045e-01 -1.87803790e-01 1.60306826e-01 5.76232731e-01
3.54133956e-02 -1.02836537e+00 -3.75433378e-02 5.30004621e-01
1.95544139e-01 3.11515126e-02 -5.14051259e-01 7.99710989e-01
1.83864683e-01 6.63134098e-01 2.84569144e-01 -7.07418144e-01
5.39900124e-01 -9.55772623e-02 5.74881673e-01 -2.99794436e-01
-4.32699591e-01 2.20519543e-01 1.04402296e-01 -3.37191194e-01
1.78257480e-01 -2.80522227e-01 -8.23467374e-01 -4.80200469e-01
-2.73512840e-01 -1.62540093e-01 5.74991286e-01 4.81141984e-01
6.66031957e-01 1.16985500e+00 3.68643850e-01 -1.06925488e+00
1.07812814e-01 -8.63188326e-01 -4.79929954e-01 3.16695422e-01
-2.99539883e-02 -9.85327661e-01 1.62254319e-01 8.96718279e-02] | [11.312917709350586, -0.15699486434459686] |
39b93ace-34b1-4b56-9bd0-1b8a51c760c7 | exploring-the-grounding-issues-in-image | 2305.14616 | null | https://arxiv.org/abs/2305.14616v1 | https://arxiv.org/pdf/2305.14616v1.pdf | Exploring the Grounding Issues in Image Caption | This paper explores the grounding issue concerning multimodal semantic representation from a computational cognitive-linguistic view. Five perceptual properties of groundedness are annotated and analyzed: Affordance, Perceptual salience, Object number, Gaze cueing, and Ecological Niche Association (ENA). We annotated selected images from the Flickr30k dataset with exploratory analyses and statistical modeling of their captions. Our findings suggest that a comprehensive understanding of an object or event requires cognitive attention, semantic distinctions in linguistic expression, and multimodal construction. During this construction process, viewers integrate situated meaning and affordance into multimodal semantics, which is consolidated into image captions used in the image-text dataset incorporating visual and textual elements. Our findings suggest that situated meaning and affordance grounding are critical for grounded natural language understanding systems to generate appropriate responses and show the potential to advance the understanding of human construal in diverse situations. | ['Shu-Kai Hsieh', 'Yu-Hsiang Tseng', 'Po-Ya Angela Wang', 'Hsin-Yu Chou', 'Pin-Er Chen'] | 2023-05-24 | null | null | null | null | ['image-captioning'] | ['computer-vision'] | [ 4.86483395e-01 2.62145638e-01 -1.98901474e-01 -6.00055218e-01
-1.71449661e-01 -8.48820686e-01 8.11758697e-01 9.62927461e-01
-4.28195477e-01 2.22968161e-01 1.18054509e+00 -3.13026339e-01
-1.41680524e-01 -4.47525620e-01 -4.92109597e-01 -3.57352555e-01
-5.49311042e-02 -2.29635268e-01 -3.02392781e-01 -5.24359345e-01
6.54951870e-01 1.95144057e-01 -1.84437704e+00 6.71754301e-01
7.26019979e-01 1.00307167e+00 4.34738964e-01 5.96371174e-01
-2.97355533e-01 1.10368180e+00 -2.12896958e-01 -3.38983804e-01
-2.41273791e-01 -4.75679368e-01 -8.51750910e-01 2.05709621e-01
5.22503555e-01 -1.81074977e-01 2.04639077e-01 1.02940559e+00
3.93663198e-01 4.45670575e-01 4.00406659e-01 -1.45888722e+00
-1.37736166e+00 6.28584385e-01 -2.35276241e-02 3.23195547e-01
7.95522094e-01 2.80611902e-01 1.39002264e+00 -8.62957835e-01
8.60856533e-01 1.65384030e+00 1.33946374e-01 3.82987678e-01
-1.27787173e+00 -2.31167227e-02 3.11615407e-01 -3.21071222e-02
-1.07662451e+00 -4.64102328e-01 6.21006727e-01 -7.52683818e-01
8.82889032e-01 5.82628250e-01 1.07663071e+00 9.62767184e-01
-2.59255439e-01 5.71811855e-01 1.18315566e+00 -5.53198695e-01
2.92460054e-01 1.82445422e-01 2.22095340e-01 6.82848692e-01
2.51958482e-02 -1.08717255e-01 -1.11844206e+00 -2.01189816e-01
8.08680177e-01 -5.54909781e-02 -1.62254825e-01 -9.50743854e-02
-1.85121810e+00 8.64731789e-01 9.28475142e-01 3.02973896e-01
-5.12747526e-01 4.12198871e-01 3.33168685e-01 -1.50382832e-01
4.21103716e-01 7.67413437e-01 2.05179125e-01 -5.80974296e-02
-2.65388250e-01 2.37013936e-01 2.74736017e-01 8.28478277e-01
6.60082579e-01 -3.27432692e-01 -3.48278463e-01 8.41543555e-01
6.57681108e-01 8.05635571e-01 1.16075844e-01 -1.34309924e+00
2.19393790e-01 9.14628685e-01 2.40245208e-01 -1.52999485e+00
-3.76289934e-01 2.28799582e-01 5.68595864e-02 -1.55885279e-01
1.16690248e-01 2.16939270e-01 -5.80483139e-01 1.94593477e+00
4.74095941e-01 -4.45754081e-01 1.12198323e-01 1.30410039e+00
1.00959170e+00 5.94413698e-01 1.17958891e+00 1.14917926e-01
2.02185225e+00 -3.25031102e-01 -8.24923217e-01 -3.83837432e-01
7.75273561e-01 -5.71523845e-01 1.70307827e+00 -5.35059385e-02
-9.58369613e-01 -4.18367684e-01 -9.44110572e-01 -7.39060402e-01
-6.51523948e-01 -7.24707097e-02 8.32800388e-01 2.65228719e-01
-1.06240904e+00 -7.67058879e-02 -2.13700235e-01 -7.79832304e-01
4.30719554e-01 -3.93847585e-01 -3.47613066e-01 1.49515361e-01
-1.10115707e+00 9.83086288e-01 3.50909770e-01 1.40443578e-01
-7.07657933e-01 -1.91585734e-01 -1.32591200e+00 -2.42378786e-01
2.76012599e-01 -6.81630075e-01 1.15322435e+00 -1.43024409e+00
-8.23420167e-01 1.25803363e+00 -2.80418277e-01 -5.15685901e-02
-2.83394396e-01 -3.15074176e-01 -3.83262604e-01 5.01466095e-01
3.80890310e-01 1.25682533e+00 5.74212551e-01 -1.51085746e+00
-3.18641752e-01 -5.49943149e-01 4.33879495e-01 7.17627347e-01
-1.83449894e-01 2.76764244e-01 2.55063176e-01 -7.27783024e-01
3.37874949e-01 -6.13905907e-01 6.28250986e-02 2.32258946e-01
-2.94148505e-01 -2.73138821e-01 4.60047543e-01 -5.43846309e-01
1.21121919e+00 -2.39401698e+00 2.21093029e-01 5.05646020e-02
4.35011446e-01 -5.99357724e-01 -3.18168670e-01 8.56684446e-01
2.04850003e-01 5.25182486e-01 1.34941876e-01 2.53455788e-02
3.34322810e-01 1.34430930e-01 -4.67464119e-01 3.67719591e-01
7.58166313e-02 1.17913151e+00 -1.30602169e+00 -7.23419130e-01
3.30554247e-01 5.61386406e-01 -6.08743310e-01 9.08018574e-02
-6.20257795e-01 2.83373415e-01 -5.75103819e-01 7.46269882e-01
1.47306308e-01 -2.27983952e-01 2.32513487e-01 -3.67132425e-01
-4.02158767e-01 3.90412837e-01 -5.17011940e-01 1.88358009e+00
-2.61732936e-01 1.01648867e+00 -2.38098457e-01 -2.39950120e-01
7.28813648e-01 3.10559832e-02 2.46731769e-02 -9.92830694e-01
1.39534339e-01 -2.41815403e-01 -5.20910621e-02 -1.06278455e+00
9.00095105e-01 -2.53764868e-01 -4.07882065e-01 5.56650519e-01
-3.08361679e-01 -2.30711013e-01 -1.05345637e-01 5.47161520e-01
2.09657848e-01 3.68686855e-01 3.23088825e-01 -4.91638303e-01
-1.12182967e-01 2.76946753e-01 -8.34701508e-02 3.24821800e-01
-1.18492812e-01 3.22582334e-01 5.58305204e-01 -4.37063724e-01
-1.02199888e+00 -1.09511948e+00 -1.82105765e-01 1.72079027e+00
6.47880912e-01 -5.41750968e-01 -7.51208007e-01 2.07064584e-01
-2.88364947e-01 1.45425928e+00 -9.46596265e-01 -6.18274510e-02
-1.02416776e-01 -3.39781761e-01 1.64955586e-01 4.47315156e-01
1.94656327e-01 -1.37938416e+00 -1.68845725e+00 -7.52101317e-02
-6.78975880e-01 -1.13907635e+00 -2.62048572e-01 -2.78192997e-01
-4.03080136e-01 -9.62729335e-01 -6.44427165e-02 -5.22689879e-01
7.39242554e-01 5.48499346e-01 1.12187970e+00 3.66838515e-01
-2.86623538e-01 1.01804507e+00 -5.71165740e-01 -6.10773683e-01
-3.26736271e-01 -5.11969447e-01 -3.22180241e-01 -6.77376166e-02
3.38270843e-01 3.50484513e-02 -7.83457816e-01 1.37087196e-01
-1.23320711e+00 6.02603137e-01 -1.62304156e-02 1.52970642e-01
4.79992777e-01 -7.31421292e-01 1.91549361e-01 -1.90119311e-01
9.42318201e-01 -8.19851518e-01 -8.17438290e-02 4.51073170e-01
-1.42156025e-02 -9.50644165e-02 -4.39543575e-01 -2.86175609e-01
-1.26810420e+00 -3.12604785e-01 5.13862669e-01 1.50676817e-01
-4.01068628e-01 5.91516018e-01 -8.60219002e-02 3.44746113e-01
8.72791409e-01 -2.09870815e-01 9.79837775e-02 9.00660008e-02
1.08050227e+00 5.18100917e-01 4.15004998e-01 -9.03094411e-01
8.31935555e-02 8.03399026e-01 -1.13931462e-01 -1.20170844e+00
-1.05825377e+00 -3.91638547e-01 -6.89266205e-01 -6.75893605e-01
1.42444909e+00 -1.14537561e+00 -9.73344743e-01 -2.23341063e-01
-1.07476914e+00 -2.45167673e-01 -4.16446269e-01 3.50100875e-01
-6.92353904e-01 6.87589943e-02 -1.49975136e-01 -9.36140597e-01
-4.45124917e-02 -9.00577843e-01 1.45367324e+00 1.56922132e-01
-8.85272980e-01 -9.61958826e-01 -1.73567206e-01 5.71122766e-01
1.81647748e-01 8.53554249e-01 1.10829926e+00 -4.79413599e-01
-5.28940678e-01 2.77043879e-01 -3.94082695e-01 -3.57479006e-01
8.27588141e-02 2.09533013e-02 -8.98450196e-01 3.59107941e-01
-2.36576378e-01 -8.20147693e-01 4.80573386e-01 1.90843374e-01
7.08928227e-01 -6.19609296e-01 -1.20155364e-01 -9.76226875e-04
1.48111904e+00 1.31041393e-01 4.72179979e-01 3.30935985e-01
6.46722794e-01 1.17389596e+00 9.38461959e-01 3.05128038e-01
8.01251531e-01 2.66163021e-01 5.23924351e-01 -5.42041957e-02
3.09680998e-01 -4.82335776e-01 2.26255327e-01 1.91669315e-01
1.12073207e-02 -3.38803291e-01 -1.18268561e+00 5.01677632e-01
-1.76386261e+00 -1.08008873e+00 -1.94567382e-01 1.78625250e+00
6.13109767e-01 -4.64809954e-01 9.62246284e-02 -3.43031675e-01
6.05729878e-01 4.46093082e-01 -2.76481569e-01 -6.07726812e-01
-4.18263346e-01 -6.53105855e-01 -7.53572807e-02 5.46342492e-01
-6.76843405e-01 1.24226272e+00 6.57662678e+00 1.94323659e-01
-8.33829403e-01 9.70607102e-02 6.73084676e-01 -1.06432118e-01
-1.18068671e+00 7.50728622e-02 -2.36560285e-01 -8.94850641e-02
6.58776879e-01 2.32821018e-01 6.40833497e-01 5.10882258e-01
4.63859081e-01 -7.72177577e-01 -1.19222069e+00 8.71055305e-01
4.82911766e-01 -1.23378897e+00 3.65127474e-01 -5.39043136e-02
1.88829646e-01 -4.06876981e-01 2.51763701e-01 -4.62417483e-01
1.18201941e-01 -1.28660810e+00 1.37460017e+00 7.56995797e-01
7.91826308e-01 -3.14068854e-01 6.33775517e-02 -1.18184127e-01
-9.09017920e-01 -1.24573305e-01 -1.51024163e-01 -3.04698020e-01
4.07109529e-01 -2.09173277e-01 -6.50901735e-01 -8.20442382e-03
5.34213841e-01 5.10709226e-01 -6.13631845e-01 2.89903969e-01
-1.77819654e-01 1.04497321e-01 1.28474319e-04 -4.25389647e-01
3.73421222e-01 -6.41299635e-02 8.46578598e-01 1.00662231e+00
1.34540156e-01 4.27124351e-01 2.89132260e-02 1.23147106e+00
3.29788595e-01 2.68204361e-01 -6.82003617e-01 -6.57891452e-01
6.93621576e-01 1.05288899e+00 -1.06239653e+00 -3.33167702e-01
-2.61547744e-01 7.04669356e-01 2.33960077e-01 6.28107071e-01
-6.97359502e-01 3.30123127e-01 7.10821986e-01 2.77529061e-01
-2.78366536e-01 -4.51755643e-01 -4.22739118e-01 -9.00548816e-01
-3.55695724e-01 -6.92336380e-01 3.51269394e-01 -1.62575352e+00
-9.30650413e-01 4.01469618e-01 4.74094838e-01 -6.35064721e-01
1.28073126e-01 -6.24583721e-01 -7.27398992e-02 5.32790124e-01
-9.32543218e-01 -1.55981851e+00 -6.16800129e-01 2.75695324e-01
4.89407808e-01 5.21798790e-01 8.17583323e-01 -4.04881358e-01
-1.02341369e-01 -3.99850249e-01 -7.26900220e-01 -2.04080299e-01
4.35005814e-01 -1.16415036e+00 1.32919848e-01 4.40028965e-01
2.22640902e-01 9.41299856e-01 8.31743300e-01 -7.90102899e-01
-1.35971272e+00 -4.80832577e-01 8.98866296e-01 -9.42807317e-01
7.15336442e-01 -3.41809750e-01 -5.33443749e-01 6.10307515e-01
7.99167037e-01 -4.05062169e-01 1.43491983e+00 5.50277419e-02
-4.57035542e-01 5.18402338e-01 -9.96213436e-01 1.14869201e+00
1.26713157e+00 -9.33005333e-01 -9.66533959e-01 7.65084326e-02
1.13717401e+00 -1.59284905e-01 -6.34278476e-01 -5.48083708e-02
6.19101286e-01 -8.82073760e-01 9.96476233e-01 -7.68336654e-01
4.62572575e-01 -4.11400825e-01 -7.38332927e-01 -9.95932102e-01
-2.72374660e-01 -3.06361973e-01 4.61954296e-01 1.16032994e+00
2.28488192e-01 -2.20638514e-01 -2.72292614e-01 1.18490553e+00
-1.25508308e-02 -8.17781091e-02 -6.04437113e-01 1.80852622e-01
-3.19777876e-01 -7.93035090e-01 7.52278686e-01 1.14360440e+00
5.10336637e-01 4.91905242e-01 3.02806292e-02 2.79818270e-02
3.48955154e-01 1.63456932e-01 4.16217953e-01 -1.07644868e+00
4.74134892e-01 -3.48471731e-01 -9.10496339e-02 -4.83232260e-01
1.27944559e-01 -8.20972204e-01 -1.98958561e-01 -1.66083086e+00
3.53908986e-01 -1.79985747e-01 -5.63402958e-02 5.15502393e-01
-6.41824603e-02 6.02532178e-02 5.29710233e-01 8.26428980e-02
-9.06200707e-01 4.14497256e-01 1.43823636e+00 6.12516813e-02
-3.33745092e-01 -1.19121706e+00 -1.18532765e+00 8.06263685e-01
6.47888422e-01 3.71644609e-02 -6.29931986e-01 -6.07313633e-01
9.68238950e-01 -7.86748901e-02 1.26174557e+00 -3.92629355e-01
-1.59963921e-01 -5.81664443e-01 4.36375380e-01 -2.71681756e-01
4.49818134e-01 -9.58731949e-01 -2.51379013e-02 2.27619067e-01
-9.38893914e-01 3.27396482e-01 3.19630086e-01 6.23808682e-01
3.43005545e-02 1.04419254e-01 2.54179448e-01 -2.19965890e-01
-1.11877728e+00 -3.89181644e-01 -5.86467385e-01 1.88605458e-01
9.63758171e-01 -4.74187762e-01 -5.12361109e-01 -4.67152297e-01
-9.42630529e-01 1.80504754e-01 6.64058685e-01 7.49580741e-01
7.53895164e-01 -1.42532885e+00 -3.99747312e-01 -7.67105445e-02
5.18837452e-01 -2.73572087e-01 4.66635257e-01 5.26865542e-01
-4.27377105e-01 2.14810327e-01 -3.28280985e-01 -7.30019510e-01
-9.11634028e-01 6.49522424e-01 4.93656509e-02 1.04818416e+00
-2.56189108e-01 7.18966663e-01 4.91455495e-01 -1.00996508e-03
-7.57772624e-02 -6.76941812e-01 -4.84014660e-01 4.54736710e-01
4.49853212e-01 2.02224880e-01 -9.19677496e-01 -1.44569302e+00
-3.42883348e-01 5.52670598e-01 8.54835749e-01 -6.60065114e-01
8.10517728e-01 -8.33580673e-01 -3.97158325e-01 8.67901862e-01
7.99219072e-01 -7.02141300e-02 -1.07859313e+00 -6.60435855e-02
1.61466673e-01 -5.57668269e-01 -1.82893217e-01 -8.78119946e-01
-8.41792151e-02 9.59457994e-01 4.20822173e-01 4.40853089e-01
7.86940873e-01 5.35804391e-01 2.38241777e-01 2.93403864e-01
1.50135001e-02 -1.13589728e+00 5.13558269e-01 1.93795830e-01
1.50730240e+00 -1.26103199e+00 -5.65012060e-02 -4.31729198e-01
-1.27479529e+00 9.10289049e-01 7.40507841e-01 5.01682758e-01
1.71967626e-01 -1.92554787e-01 1.94085762e-01 -6.69050455e-01
-7.33827710e-01 -5.38829207e-01 5.24482965e-01 5.46605051e-01
5.99075973e-01 2.51464635e-01 -8.99717957e-02 3.96038800e-01
-3.92402440e-01 -5.39258897e-01 8.75419378e-02 7.06271648e-01
-7.82738209e-01 -1.67073056e-01 -4.25761729e-01 -1.26636535e-01
-1.41428873e-01 -4.03472632e-01 -7.44654655e-01 7.05582201e-01
3.04696500e-01 1.01315022e+00 5.59295058e-01 6.25170469e-02
-3.59823890e-02 1.11172520e-01 4.39399809e-01 -6.76032126e-01
-4.78362948e-01 1.11244127e-01 4.54058647e-01 -7.12092876e-01
-9.15763140e-01 -7.19756067e-01 -1.61380005e+00 1.28567770e-01
3.67131829e-02 -1.86801013e-02 7.75706947e-01 8.50899696e-01
5.74607015e-01 1.96694911e-01 -1.76777959e-01 -6.62367880e-01
4.03289318e-01 -7.12197483e-01 -1.68135643e-01 6.56781256e-01
4.64301944e-01 -5.18781126e-01 -1.14949383e-01 4.05863851e-01] | [10.775086402893066, 1.6524487733840942] |
3423fd5c-1df9-42eb-ae13-12b67518a536 | car-pose-in-context-accurate-pose-estimation | 1912.04363 | null | https://arxiv.org/abs/1912.04363v1 | https://arxiv.org/pdf/1912.04363v1.pdf | Car Pose in Context: Accurate Pose Estimation with Ground Plane Constraints | Scene context is a powerful constraint on the geometry of objects within the scene in cases, such as surveillance, where the camera geometry is unknown and image quality may be poor. In this paper, we describe a method for estimating the pose of cars in a scene jointly with the ground plane that supports them. We formulate this as a joint optimization that accounts for varying car shape using a statistical atlas, and which simultaneously computes geometry and internal camera parameters. We demonstrate that this method produces significant improvements for car pose estimation, and we show that the resulting 3D geometry, when computed over a video sequence, makes it possible to improve on state of the art classification of car behavior. We also show that introducing the planar constraint allows us to estimate camera focal length in a reliable manner. | ['Gregory D. Hager', 'Weichao Qiu', 'Michael Peven', 'Alan L. Yuille', 'Pengfei Li'] | 2019-12-09 | null | null | null | null | ['car-pose-estimation'] | ['computer-vision'] | [ 2.31087375e-02 -1.51879236e-01 3.83632518e-02 -3.00001085e-01
-9.72811222e-01 -9.69996572e-01 6.69902384e-01 -6.48937970e-02
-3.65330279e-01 2.04097301e-01 -3.10622919e-02 -2.96041191e-01
2.52435148e-01 -4.55752671e-01 -1.04303265e+00 -6.62057996e-01
1.53329164e-01 5.79018652e-01 6.99237525e-01 1.04514509e-01
2.16519222e-01 1.06528461e+00 -1.35746574e+00 -3.01583320e-01
1.82398930e-01 7.00128973e-01 3.47433567e-01 1.13515496e+00
3.55311841e-01 7.23985314e-01 -3.51701319e-01 -4.19162244e-01
3.45070392e-01 -4.74388385e-03 -5.73600590e-01 9.32658315e-01
7.46952951e-01 -5.83263338e-01 -3.29237074e-01 8.15662861e-01
-1.77294493e-01 1.88101888e-01 8.45390975e-01 -1.09233212e+00
3.39540899e-01 -3.54435384e-01 -5.59650421e-01 2.35795736e-01
2.59520978e-01 4.05141711e-02 8.41465712e-01 -7.79086053e-01
6.45952880e-01 9.68309283e-01 7.07234919e-01 -2.39241440e-02
-1.16602445e+00 -1.25850827e-01 3.21362793e-01 5.47861047e-02
-1.59462321e+00 -7.10452974e-01 9.04331326e-01 -5.98542929e-01
7.40853190e-01 1.54752165e-01 3.87543291e-01 3.02132100e-01
2.23773390e-01 5.97632647e-01 4.29484516e-01 -4.49066699e-01
1.55948058e-01 1.35025904e-01 1.12459734e-01 7.96662271e-01
3.72043580e-01 -1.01042449e-01 -8.54060128e-02 -1.39548719e-01
8.81385446e-01 1.19661838e-01 -7.89274126e-02 -8.84554863e-01
-1.14071023e+00 8.29546928e-01 2.46471971e-01 -1.23066857e-01
-1.63616002e-01 2.70431161e-01 -1.31776452e-01 -3.62218052e-01
3.16294670e-01 3.51172030e-01 -2.71736681e-01 -6.27847090e-02
-7.23401845e-01 5.13946652e-01 8.33263874e-01 1.31474698e+00
8.49096894e-01 -2.11457580e-01 3.82189244e-01 4.95055914e-01
1.70308366e-01 8.65952611e-01 -4.41589296e-01 -1.51613224e+00
3.63428146e-01 1.74210131e-01 4.67877924e-01 -1.18443704e+00
-2.98052132e-01 -1.74592473e-02 -3.58186588e-02 2.29201525e-01
6.38928294e-01 -1.80048764e-01 -6.22503161e-01 1.48012090e+00
5.05849600e-01 2.94380128e-01 -1.36647761e-01 7.63218880e-01
2.52183706e-01 4.55440164e-01 -3.15663874e-01 -2.01820910e-01
1.42820203e+00 -7.85706818e-01 -4.57360387e-01 -4.49773669e-01
6.14510536e-01 -9.69777167e-01 4.06449586e-01 4.45204198e-01
-1.02843821e+00 -1.81080252e-01 -9.06536281e-01 8.21830034e-02
2.58397404e-02 2.83222198e-01 5.01971185e-01 6.74453139e-01
-9.47205365e-01 5.17027937e-02 -1.11329007e+00 -3.54623437e-01
4.23827581e-02 3.62776130e-01 -5.03471613e-01 -1.79572344e-01
-2.64593869e-01 1.27328575e+00 -1.09542385e-02 -8.73896927e-02
-9.16597128e-01 -2.50549823e-01 -1.11744308e+00 -2.13942841e-01
7.74201393e-01 -3.09439063e-01 1.53934193e+00 -5.25686920e-01
-1.26172638e+00 7.39586890e-01 -4.66990024e-01 -1.45152509e-01
4.45071697e-01 -1.49645820e-01 -1.33819804e-01 4.53316629e-01
-3.54153402e-02 5.10080457e-01 7.54227281e-01 -1.54282105e+00
-7.06172705e-01 -4.91062135e-01 2.57715136e-01 4.02518719e-01
3.64298046e-01 1.51427701e-01 -1.05655420e+00 -1.68601960e-01
1.12452239e-01 -1.36548758e+00 -5.14750302e-01 1.79015234e-01
-2.76029527e-01 2.33531386e-01 9.98124540e-01 -7.29059994e-01
6.66300178e-01 -2.13960338e+00 2.40072813e-02 3.99065763e-01
1.50951385e-01 -8.71089697e-02 2.10347593e-01 2.06925154e-01
2.09463701e-01 -1.82346389e-01 -1.71380863e-01 -5.14023721e-01
-4.66136515e-01 4.26947176e-01 -1.88414212e-02 1.02135909e+00
2.54472256e-01 4.95880276e-01 -6.16681099e-01 -5.55359125e-01
5.15280664e-01 4.33267117e-01 -5.81992090e-01 3.77236694e-01
-1.79686081e-02 2.54942417e-01 -4.93706673e-01 4.43504959e-01
8.37067425e-01 -3.33453156e-02 2.15039119e-01 -1.25630409e-01
-2.56765097e-01 -3.19391116e-02 -1.35644126e+00 1.21523046e+00
-4.14600104e-01 7.67485201e-01 5.31047702e-01 -6.93750024e-01
5.69107294e-01 1.49347216e-01 5.59168279e-01 2.89474451e-03
2.16800570e-01 -1.48600340e-01 -2.13849783e-01 -4.04927403e-01
6.54254735e-01 -5.75795695e-02 1.11419698e-02 1.49724707e-01
-8.59946683e-02 -6.96170628e-01 3.18944633e-01 1.61783203e-01
9.52336371e-01 -9.25882682e-02 2.72126913e-01 -2.80702323e-01
2.97915488e-01 1.50798038e-01 3.34430873e-01 4.60270643e-01
-1.11652650e-02 8.53565037e-01 5.30883610e-01 -1.51732057e-01
-1.33000100e+00 -1.17760170e+00 -1.70289263e-01 4.73780662e-01
4.97729987e-01 -1.71845853e-01 -8.84865224e-01 -4.88055825e-01
-3.80134024e-02 5.03364623e-01 -3.31740558e-01 1.24613531e-01
-8.29994619e-01 -5.94483674e-01 2.53402651e-03 7.02772200e-01
1.32635027e-01 -2.04350755e-01 -6.75287247e-01 4.82516959e-02
-2.46728867e-01 -1.93736410e+00 -6.84898734e-01 -6.68154806e-02
-7.53079891e-01 -1.50644004e+00 -2.73933023e-01 -4.57325041e-01
9.68459368e-01 8.44976723e-01 1.04115200e+00 1.52142465e-01
-2.07872525e-01 8.19594681e-01 -2.72472709e-01 -3.27784926e-01
-3.32627445e-01 -3.42203796e-01 -8.53037164e-02 7.96722770e-02
1.50826097e-01 -7.77531788e-02 -3.99778485e-01 6.53930783e-01
-6.34911418e-01 -5.80011792e-02 2.08955228e-01 3.38072270e-01
6.40699565e-01 3.09188038e-01 -3.14901382e-01 -8.09471250e-01
-1.58859640e-01 -2.81919450e-01 -1.13382685e+00 -2.62952559e-02
1.35130417e-02 -1.42223731e-01 3.05804580e-01 -1.75784647e-01
-9.03885663e-01 6.52824163e-01 1.56618550e-01 -6.10820889e-01
-3.26324970e-01 8.92420486e-02 -2.56065220e-01 -4.43059117e-01
3.52720290e-01 -1.18774980e-01 2.27527201e-01 -3.75630677e-01
2.53108472e-01 2.73124903e-01 7.75116682e-01 -4.16041017e-01
8.49034488e-01 8.31714869e-01 3.51213455e-01 -1.24046147e+00
-6.44479573e-01 -8.66632879e-01 -9.12512422e-01 -3.84092122e-01
1.07777929e+00 -1.11631536e+00 -6.81112289e-01 2.56090760e-01
-1.40506887e+00 -2.98892826e-01 2.17118129e-01 9.47385430e-01
-6.79753780e-01 4.98968452e-01 -3.02531570e-01 -1.03249717e+00
5.43311656e-01 -1.45541441e+00 1.35936987e+00 -2.01744676e-01
6.08279705e-02 -1.03708613e+00 -1.33584842e-01 5.00731707e-01
-1.65756285e-01 3.38525355e-01 3.56323987e-01 -1.01184502e-01
-1.17369962e+00 -4.67939913e-01 -4.41234261e-02 1.91047788e-01
-1.08896315e-01 4.39220220e-01 -9.18222904e-01 -2.09521979e-01
5.48284464e-02 2.66356796e-01 6.45079315e-01 6.91423178e-01
8.98631334e-01 4.53402922e-02 -5.78655720e-01 6.46248162e-01
1.50472653e+00 9.76305977e-02 3.55903149e-01 1.69018116e-02
8.87165070e-01 7.18723238e-01 7.73413002e-01 2.56871551e-01
5.57144582e-01 1.00555789e+00 5.61107635e-01 -4.44293655e-02
1.53281301e-01 4.90912534e-02 2.18670562e-01 4.96628702e-01
-2.19809115e-01 -3.92871052e-01 -8.56493771e-01 6.21764362e-01
-1.69947410e+00 -8.64940405e-01 -3.60214531e-01 2.33480167e+00
6.37066662e-02 1.08525477e-01 3.82814437e-01 -1.64064541e-01
7.46830106e-01 -6.60993606e-02 -3.81260991e-01 -7.91347623e-02
5.50064892e-02 -3.85876358e-01 1.37373507e+00 1.02833283e+00
-1.22862613e+00 6.18790746e-01 7.87457466e+00 2.68580973e-01
-7.36390650e-01 -5.88335749e-03 5.39230168e-01 1.28142223e-01
-2.11759180e-01 1.94021389e-01 -9.94230688e-01 2.13179380e-01
6.37202382e-01 1.26857370e-01 3.87201905e-01 8.07499051e-01
1.35720104e-01 -5.49796879e-01 -1.15227628e+00 8.65258813e-01
3.86291862e-01 -1.08730805e+00 -3.29457641e-01 4.27478611e-01
5.87742448e-01 1.17414564e-01 -1.90502614e-01 -2.88223326e-01
4.22750264e-01 -6.51294947e-01 9.66436744e-01 2.51371413e-01
6.18935764e-01 -8.73791873e-01 6.29482508e-01 5.91157675e-01
-1.26980984e+00 1.86921237e-03 -2.91145325e-01 2.58654002e-02
4.33402568e-01 3.68715435e-01 -1.06818831e+00 2.37407193e-01
3.06765497e-01 5.02053559e-01 -5.62029958e-01 1.06590950e+00
-1.63046010e-02 3.69801432e-01 -6.88501477e-01 2.52941638e-01
1.45805478e-01 -3.66464287e-01 7.50401676e-01 1.16646993e+00
3.66034687e-01 3.99206936e-01 4.99881774e-01 4.96748537e-01
1.99953586e-01 -2.21631899e-01 -9.24200118e-01 3.97753507e-01
2.86985874e-01 1.20573020e+00 -1.03379047e+00 -1.00416139e-01
-6.37163818e-01 5.89229167e-01 2.22522765e-01 3.76580477e-01
-7.72197902e-01 1.47422045e-01 7.33426690e-01 3.42350245e-01
6.25748098e-01 -8.06775093e-01 1.02695618e-02 -1.35678279e+00
2.49270976e-01 -4.59425658e-01 6.15716949e-02 -7.62435496e-01
-6.48671269e-01 3.42355788e-01 5.09266436e-01 -1.27437699e+00
-5.60943246e-01 -9.63172495e-01 -5.67422748e-01 4.49689806e-01
-1.15315378e+00 -1.27164817e+00 -1.79205477e-01 4.63705063e-01
5.91723263e-01 6.68386742e-02 3.63979191e-01 -3.74207809e-03
-3.42563421e-01 1.96287543e-01 1.14677303e-01 7.35234320e-02
2.39454851e-01 -1.24636757e+00 2.38805443e-01 1.10130167e+00
1.89507172e-01 3.57942253e-01 9.92398560e-01 -4.82679784e-01
-1.68070602e+00 -9.47659373e-01 5.56478620e-01 -1.11911249e+00
4.70781773e-01 -5.80991507e-01 -5.51013887e-01 9.87821341e-01
-1.32894814e-01 2.71618485e-01 3.10477912e-01 1.04846843e-01
-1.68615744e-01 -4.23577875e-02 -9.66954947e-01 3.83509666e-01
7.11783111e-01 -5.94423473e-01 -2.27859080e-01 4.57264751e-01
4.60797846e-01 -6.35369241e-01 -4.63147670e-01 1.85727969e-01
5.06987631e-01 -8.73994112e-01 1.15392303e+00 -1.57085955e-01
-7.76327401e-02 -5.21508038e-01 -4.32604343e-01 -1.04214621e+00
-2.04606190e-01 -3.11909348e-01 5.57430573e-02 8.60631526e-01
2.77659357e-01 -2.41705298e-01 9.71039832e-01 8.00280631e-01
-1.36248112e-01 -3.98439705e-01 -8.35111380e-01 -7.12377846e-01
-1.55077383e-01 -6.59683347e-01 2.66203195e-01 4.21743423e-01
-4.04369175e-01 1.73453197e-01 -2.90958524e-01 7.02964664e-01
7.53787220e-01 -8.24526176e-02 1.13676250e+00 -1.08807611e+00
-3.21421087e-01 -1.40446588e-01 -8.70138824e-01 -1.33079445e+00
3.45915169e-01 -2.36195818e-01 1.89482540e-01 -9.35268581e-01
4.03978318e-01 -2.92112827e-01 4.23718661e-01 -1.64903179e-01
-1.94109213e-02 1.57328010e-01 2.23810360e-01 6.28707856e-02
-5.78178406e-01 -3.61479563e-03 9.81895387e-01 1.19732566e-01
1.83576331e-01 2.27287978e-01 -3.89906615e-01 1.13261151e+00
4.42279041e-01 -2.83243090e-01 -3.50117087e-01 -6.64372921e-01
-1.42181203e-01 2.79331148e-01 6.17865384e-01 -9.58543122e-01
2.18500689e-01 -2.96986997e-01 3.54236186e-01 -8.04068565e-01
8.81776869e-01 -1.31458902e+00 1.16926573e-01 5.80636561e-02
-2.62340512e-02 2.82348335e-01 9.00412276e-02 7.00026870e-01
-8.05572495e-02 -4.03671622e-01 9.33076918e-01 -1.68362632e-01
-5.70519865e-01 2.37994686e-01 -4.30378437e-01 -4.41206209e-02
1.22597098e+00 -2.57204711e-01 1.36162341e-01 -6.87817991e-01
-5.09399295e-01 2.36974254e-01 8.40164900e-01 1.20917313e-01
3.62068951e-01 -1.13418782e+00 -5.96856773e-01 1.31784767e-01
1.32294163e-01 1.24316335e-01 3.02670393e-02 7.73056328e-01
-9.85222220e-01 3.68678212e-01 2.85337389e-01 -9.77300346e-01
-1.67009401e+00 6.28002107e-01 4.52868640e-01 4.75416556e-02
-3.92190486e-01 5.82661211e-01 4.87616450e-01 -2.11016893e-01
4.21121120e-02 -3.75405192e-01 -1.60217527e-02 -3.25457603e-01
5.74278235e-01 2.68872947e-01 3.87148075e-02 -1.35932064e+00
-4.23078090e-01 8.92101645e-01 6.22059777e-02 -3.73041570e-01
1.14656222e+00 -4.59299386e-01 2.62859017e-01 3.58891785e-01
1.26279032e+00 2.91787386e-01 -1.77990448e+00 -1.93125993e-01
-2.68633842e-01 -8.18209231e-01 2.15927348e-01 -9.40030161e-03
-9.85139132e-01 7.53987670e-01 2.39785761e-01 1.77417174e-02
7.10041046e-01 2.85196394e-01 4.76188600e-01 5.86423814e-01
4.61796910e-01 -9.42601919e-01 -1.78804830e-01 4.29715693e-01
5.22213876e-01 -1.35701489e+00 2.77151823e-01 -9.20942962e-01
-5.89024305e-01 1.16209459e+00 4.52605069e-01 -1.63335547e-01
7.30740726e-01 4.24334019e-01 -4.72301245e-02 -2.01448843e-01
-6.09546959e-01 -2.21213266e-01 2.30069295e-01 6.20965540e-01
1.23655163e-01 1.88252315e-01 4.20958579e-01 -1.67656124e-01
-8.54769871e-02 -5.47903717e-01 7.84984350e-01 9.50091541e-01
-6.23281658e-01 -8.17945242e-01 -7.93838203e-01 2.18098477e-01
-4.40858215e-01 2.88608909e-01 -1.64913446e-01 9.18420792e-01
9.43260267e-02 1.05504096e+00 3.48682612e-01 -6.99997917e-02
4.60128099e-01 -4.49614882e-01 7.18617678e-01 -6.07101262e-01
1.59882218e-01 4.45780218e-01 3.76766980e-01 -4.78297085e-01
-4.90887702e-01 -1.10942662e+00 -1.00990736e+00 -1.87368318e-01
-7.48056233e-01 4.53820489e-02 7.61814713e-01 1.03705204e+00
-1.15128942e-01 3.36865410e-02 7.79019892e-01 -1.19314671e+00
-2.00232804e-01 -4.10866678e-01 -7.82142520e-01 2.78425068e-01
7.21314490e-01 -9.08015490e-01 -3.95908564e-01 5.71002007e-01] | [7.751688003540039, -2.416954755783081] |
d30aa701-17d3-4455-94ec-24dc69ff1ce6 | tunaoil-a-tuning-algorithm-strategy-for | 2208.02606 | null | https://arxiv.org/abs/2208.02606v1 | https://arxiv.org/pdf/2208.02606v1.pdf | TunaOil: A Tuning Algorithm Strategy for Reservoir Simulation Workloads | Reservoir simulations for petroleum fields and seismic imaging are known as the most demanding workloads for high-performance computing (HPC) in the oil and gas (O&G) industry. The optimization of the simulator numerical parameters plays a vital role as it could save considerable computational efforts. State-of-the-art optimization techniques are based on running numerous simulations, specific for that purpose, to find good parameter candidates. However, using such an approach is highly costly in terms of time and computing resources. This work presents TunaOil, a new methodology to enhance the search for optimal numerical parameters of reservoir flow simulations using a performance model. In the O&G industry, it is common to use ensembles of models in different workflows to reduce the uncertainty associated with forecasting O&G production. We leverage the runs of those ensembles in such workflows to extract information from each simulation and optimize the numerical parameters in their subsequent runs. To validate the methodology, we implemented it in a history matching (HM) process that uses a Kalman filter algorithm to adjust an ensemble of reservoir models to match the observed data from the real field. We mine past execution logs from many simulations with different numerical configurations and build a machine learning model based on extracted features from the data. These features include properties of the reservoir models themselves, such as the number of active cells, to statistics of the simulation's behavior, such as the number of iterations of the linear solver. A sampling technique is used to query the oracle to find the numerical parameters that can reduce the elapsed time without significantly impacting the quality of the results. Our experiments show that the predictions can improve the overall HM workflow runtime on average by 31%. | ['Josep Lluís Berral', 'José Roberto Pereira Rodrigues', 'David Buchaca Prats', 'Felipe Albuquerque Portella'] | 2022-08-04 | null | null | null | null | ['seismic-imaging'] | ['miscellaneous'] | [-3.72578710e-01 -5.54598093e-01 4.57712650e-01 2.03966144e-02
-6.62036479e-01 -6.89587533e-01 5.53495407e-01 5.49460351e-01
-4.26904798e-01 5.66002250e-01 5.60108759e-02 -5.37161112e-01
-1.99819952e-01 -9.67922151e-01 -6.15539193e-01 -9.23911452e-01
-5.38916767e-01 6.63028836e-01 4.02044594e-01 -1.91455901e-01
6.98106289e-01 9.35184896e-01 -1.59271693e+00 3.13083292e-03
6.81027830e-01 8.88999224e-01 1.20494403e-01 8.77161503e-01
1.81480378e-01 6.43506050e-01 -4.76844400e-01 3.44894022e-01
4.30707067e-01 -2.11980805e-01 -4.33917314e-01 -2.79242963e-01
-4.15798783e-01 -2.72353142e-01 -1.23059452e-01 6.90339088e-01
6.05816007e-01 4.73108768e-01 4.16603327e-01 -9.20488596e-01
8.45003128e-01 5.75671792e-01 -2.45097354e-01 4.89278466e-01
1.13592553e-03 8.55722845e-01 3.88954818e-01 -7.86257803e-01
3.51450622e-01 7.59847283e-01 5.21765649e-01 -4.59913492e-01
-1.25669730e+00 -5.30524075e-01 -3.25642347e-01 2.21310984e-02
-1.55391228e+00 -4.87260222e-01 2.93127954e-01 -9.69565570e-01
1.12074578e+00 2.39843264e-01 9.35345948e-01 1.91140369e-01
5.19475937e-01 -1.30101115e-01 1.03217518e+00 -2.96407878e-01
8.22725296e-01 -5.51078767e-02 -2.25329578e-01 4.81212795e-01
2.33570457e-01 4.93619114e-01 -7.02993512e-01 -5.09584963e-01
5.45009911e-01 -2.27688849e-01 -2.21301928e-01 6.49667084e-02
-9.56534088e-01 7.34211087e-01 5.18155657e-02 1.83350533e-01
-5.30331314e-01 1.98087022e-01 4.96606380e-01 1.22690149e-01
3.14605922e-01 1.07406223e+00 -4.57833976e-01 -4.46094036e-01
-1.29212570e+00 6.18775487e-01 1.15013885e+00 2.96937853e-01
9.65040267e-01 1.41277105e-01 -2.37591341e-01 2.03749090e-01
2.28536367e-01 5.71312904e-01 3.26175511e-01 -1.03541648e+00
1.55853242e-01 4.89157379e-01 2.29308873e-01 -7.22641647e-01
-4.30579066e-01 -3.36717606e-01 -5.37535846e-01 5.58247209e-01
3.68699938e-01 -2.56562591e-01 -8.70566070e-01 9.07025158e-01
3.53468537e-01 6.41271770e-01 4.12789397e-02 6.92303061e-01
9.99114066e-02 8.95116448e-01 -1.66666098e-02 -1.44201368e-01
1.02845705e+00 -6.45472109e-01 -1.76834717e-01 -6.88371733e-02
7.35147655e-01 -8.98746490e-01 6.84977710e-01 2.14650869e-01
-1.00319815e+00 -2.64660686e-01 -1.12820601e+00 7.72790134e-01
-4.47182506e-01 -2.26034731e-01 4.49425131e-01 4.64825988e-01
-6.91757023e-01 1.45810521e+00 -1.29172373e+00 1.10201150e-01
-9.81491432e-03 2.21448898e-01 2.62833595e-01 1.81913525e-01
-1.01490581e+00 1.06755042e+00 4.62719917e-01 3.35051507e-01
-1.21961904e+00 -1.16437876e+00 -5.63988745e-01 5.28420389e-01
4.13235188e-01 -4.36665416e-01 1.15145743e+00 -1.07222334e-01
-1.58311069e+00 5.79964668e-02 4.03249152e-02 -3.97183210e-01
7.28474975e-01 1.52447522e-01 -1.74327374e-01 -1.06418416e-01
-1.88873693e-01 -3.00317973e-01 5.74847102e-01 -9.18957114e-01
-4.74803090e-01 -1.40151575e-01 -4.86630052e-01 1.75082441e-02
2.12368429e-01 -5.92875145e-02 -4.71164882e-01 -2.63183504e-01
3.52620035e-02 -1.08448553e+00 -7.33869612e-01 -5.89364052e-01
1.31813586e-02 3.58501017e-01 4.15762782e-01 -9.07505453e-01
1.21631205e+00 -1.85520613e+00 -9.06051472e-02 8.85514140e-01
-1.63932070e-01 -4.82343920e-02 3.82919669e-01 6.85680330e-01
1.67055830e-01 1.84681460e-01 -4.00495231e-01 -8.86576921e-02
-2.50133604e-01 1.41136602e-01 -3.69459599e-01 5.17630100e-01
1.33554980e-01 2.74819106e-01 -8.28335881e-01 -2.40451843e-01
4.18280393e-01 7.10405409e-02 -3.72312874e-01 7.47453153e-01
-3.56182277e-01 8.77034247e-01 -4.17684078e-01 3.52772206e-01
6.00975871e-01 -3.35259736e-01 2.56340533e-01 -4.49505858e-02
-5.43923676e-01 4.94761989e-02 -1.45690334e+00 1.12513292e+00
-7.40451157e-01 2.95567513e-01 6.44147489e-03 -7.72174954e-01
9.13482964e-01 1.27269328e-01 7.20731080e-01 -4.79463220e-01
8.99204761e-02 5.35951555e-01 9.99927819e-02 -6.36516929e-01
6.86956882e-01 1.51158243e-01 7.56291896e-02 5.69873452e-01
-4.14764404e-01 -7.25948989e-01 5.17244101e-01 -1.84386503e-02
1.29491162e+00 1.74476445e-01 1.39851958e-01 -4.15311009e-01
4.47494417e-01 2.90404141e-01 4.07113731e-01 7.16047049e-01
4.32513952e-01 1.59677252e-01 6.27963543e-01 -4.70005840e-01
-1.35960543e+00 -5.65493166e-01 -1.95216704e-02 6.86293840e-01
-5.51664531e-02 -3.14494729e-01 -3.33335608e-01 1.26715183e-01
1.83365449e-01 7.86270440e-01 -4.11088675e-01 -4.04589884e-02
-7.96783566e-01 -1.15973639e+00 2.67177731e-01 2.83496380e-01
1.09633960e-01 -9.09721017e-01 -1.16563082e+00 5.62339008e-01
6.13682568e-01 -8.93846631e-01 3.26050594e-02 3.51489276e-01
-9.47107494e-01 -9.86604333e-01 -2.71686882e-01 2.70828962e-01
5.32386959e-01 -5.04653871e-01 1.19601703e+00 4.10175681e-01
-1.82199061e-01 -2.59201944e-01 -1.93759605e-01 -2.69535452e-01
-7.89563477e-01 1.40052080e-01 -2.42167503e-01 -3.07725877e-01
-3.54900181e-01 -6.43442631e-01 -5.12804329e-01 3.35930258e-01
-8.30111444e-01 5.23032621e-03 3.13169301e-01 8.41055572e-01
5.32336116e-01 3.69720399e-01 4.94114794e-02 -7.39642859e-01
5.21760225e-01 -7.53073692e-01 -1.52613795e+00 2.88154036e-01
-1.03790009e+00 4.81945455e-01 8.55296254e-01 -2.02512383e-01
-8.72736812e-01 2.03983888e-01 1.07912883e-01 -5.31932414e-01
3.30527753e-01 1.11173403e+00 2.67646283e-01 -2.11199999e-01
4.34709936e-01 8.13750327e-02 4.23886143e-02 -5.21041751e-01
-2.30001643e-01 3.52735877e-01 4.77550477e-01 -8.45860302e-01
7.56576478e-01 1.54312789e-01 5.60061157e-01 -7.07136333e-01
-2.92271405e-01 -4.35155749e-01 -3.13908249e-01 -5.19863844e-01
2.55710036e-01 -6.47629201e-01 -1.08827603e+00 5.90503275e-01
-7.51307845e-01 -5.72393656e-01 -4.55520535e-03 6.37863994e-01
-2.58676142e-01 1.08226366e-01 -3.06160122e-01 -9.85123277e-01
-4.03545260e-01 -1.54362333e+00 7.77185917e-01 5.15701890e-01
4.82370667e-02 -7.97526479e-01 3.79196316e-01 -1.19955853e-01
9.03978109e-01 5.70503592e-01 7.28321135e-01 -5.12932301e-01
-9.69659269e-01 -1.63406640e-01 2.10687265e-01 -3.09278164e-02
-4.83119845e-01 4.47383165e-01 -9.00961518e-01 -3.73630613e-01
3.37252021e-02 2.74953574e-01 7.09025681e-01 2.97888249e-01
1.34489930e+00 -1.80861071e-01 -3.14259708e-01 8.79190445e-01
1.73547363e+00 2.10749894e-01 5.97780883e-01 5.00790000e-01
3.66886139e-01 3.62291813e-01 7.43772507e-01 1.04739130e+00
-2.84467578e-01 6.38065398e-01 4.11369801e-01 9.11825150e-02
5.88188887e-01 1.12855621e-01 2.60723144e-01 7.34408140e-01
-3.74023497e-01 6.88841715e-02 -1.46766996e+00 5.13460219e-01
-1.72480786e+00 -7.83757627e-01 6.58540949e-02 2.67929745e+00
7.30030477e-01 2.42999777e-01 -1.84791267e-01 -3.01733892e-02
4.14837569e-01 -1.65714286e-02 -7.16835618e-01 -4.22356158e-01
3.60987604e-01 5.68984330e-01 1.19559383e+00 4.04868752e-01
-6.69894099e-01 5.46514094e-01 5.49489737e+00 5.41416347e-01
-1.36499798e+00 -2.07780287e-01 5.18840015e-01 -1.40359104e-01
-9.16613489e-02 4.94541615e-01 -6.92013323e-01 6.73786104e-01
1.64902484e+00 -5.99448383e-01 8.36715937e-01 8.71284544e-01
6.64755642e-01 -7.49178767e-01 -9.63778913e-01 5.43299258e-01
-4.74139243e-01 -1.91174567e+00 -6.23241186e-01 6.83271363e-02
6.03082955e-01 3.38071525e-01 -4.63114947e-01 1.99303687e-01
4.70723242e-01 -1.20464659e+00 7.78270662e-01 9.93154109e-01
4.25779343e-01 -8.60810697e-01 1.08328378e+00 3.18298995e-01
-1.25151396e+00 -1.36612151e-02 -4.68904339e-02 -1.39917612e-01
4.81498390e-01 8.39916110e-01 -1.25112748e+00 6.31658792e-01
8.99724245e-01 1.51444793e-01 -5.71404815e-01 1.25161231e+00
2.98131317e-01 8.20030868e-01 -8.31214964e-01 4.68172878e-02
9.06319547e-05 -4.68949169e-01 5.27858317e-01 1.00883615e+00
6.98978662e-01 2.90551577e-02 2.73924559e-01 8.53912830e-01
3.45714062e-01 -4.60542887e-02 -2.60747224e-01 -9.38219950e-02
7.15590537e-01 1.07642031e+00 -7.38803744e-01 -3.93577248e-01
9.66655836e-02 9.40324068e-02 -1.07420988e-01 2.03066796e-01
-6.91345930e-01 -7.69274011e-02 6.64641857e-01 2.18131766e-01
3.50768715e-01 -5.21444261e-01 -5.05532205e-01 -6.86251342e-01
-1.04015589e-01 -8.07124853e-01 3.07224780e-01 -4.65500832e-01
-9.77727771e-01 4.21720117e-01 1.24982916e-01 -1.18060243e+00
-6.49721980e-01 -3.89196575e-01 -7.23128259e-01 1.37101996e+00
-1.42788231e+00 -3.79381835e-01 -5.85734725e-01 5.30707010e-04
6.19328767e-02 -8.73660892e-02 6.37994468e-01 4.44202945e-02
-6.24141157e-01 1.16737774e-02 6.19881749e-01 -1.72216535e-01
4.04481560e-01 -1.01186490e+00 3.51692080e-01 1.01176870e+00
-5.64265192e-01 4.55105841e-01 1.18085396e+00 -8.33973765e-01
-1.58915627e+00 -8.35296333e-01 3.14220667e-01 2.46830042e-02
9.78182554e-01 8.63981172e-02 -1.26624227e+00 1.89934462e-01
-6.86922446e-02 2.26063266e-01 2.41583571e-01 -2.44040668e-01
3.37201983e-01 -2.07283616e-01 -9.48667824e-01 3.15740466e-01
2.81518668e-01 -2.48197541e-01 3.73612531e-02 3.32243145e-01
1.42032653e-01 -8.98443580e-01 -1.51304543e+00 4.31113392e-01
4.81036723e-01 -8.54568601e-01 6.54133320e-01 -4.44606155e-01
4.23733234e-01 -6.25371814e-01 9.23851281e-02 -1.53652799e+00
2.15300158e-01 -7.05978274e-01 -1.50435075e-01 9.17237222e-01
4.25990850e-01 -7.53675401e-01 4.24885362e-01 8.71877074e-01
-9.89532936e-03 -8.20602298e-01 -8.90247405e-01 -6.78663015e-01
-1.03584774e-01 -3.08753937e-01 1.07971442e+00 5.94483852e-01
-3.68839532e-01 -1.86692476e-01 -7.86199346e-02 6.90945446e-01
5.83706439e-01 3.31531197e-01 9.50317025e-01 -1.04559398e+00
-6.26325488e-01 -2.58738458e-01 -1.25691697e-01 -1.14545658e-01
-6.89813644e-02 -5.94088137e-01 1.60393402e-01 -9.48492110e-01
-1.64282784e-01 -8.03980052e-01 -8.71887058e-02 2.01108053e-01
-1.66586295e-01 -4.36177135e-01 2.86963999e-01 4.28491056e-01
1.23444915e-01 3.77608895e-01 7.52950311e-01 1.84522495e-01
-4.92100477e-01 1.56115636e-01 2.03877032e-01 4.36066031e-01
7.67761946e-01 -7.61492074e-01 1.44752458e-01 -1.91020235e-01
5.07709742e-01 5.00622869e-01 2.55575269e-01 -1.28351974e+00
3.97748649e-01 -3.62739503e-01 3.18978906e-01 -7.11397290e-01
-6.61370605e-02 -7.35142827e-01 8.46413016e-01 6.79718494e-01
-1.72628313e-02 4.22618151e-01 3.92895073e-01 3.54198724e-01
-1.49127230e-01 -5.08425236e-01 8.99919748e-01 -2.49737799e-01
-7.66578734e-01 1.54540241e-01 -5.46692193e-01 -1.75489157e-01
9.18108642e-01 1.58371508e-01 -1.96127802e-01 -1.11529477e-01
-5.77674031e-01 6.51936352e-01 6.53124630e-01 -1.70388430e-01
6.76876605e-02 -5.40074468e-01 -8.14427495e-01 2.03612328e-01
-1.32459551e-01 2.15075344e-01 2.09102646e-01 8.25028718e-01
-1.00532019e+00 -2.68406458e-02 -4.28893715e-02 -6.42903328e-01
-8.46513093e-01 2.86219746e-01 7.41646230e-01 -8.82698238e-01
-4.11878645e-01 2.85012990e-01 -8.12679470e-01 -1.35273814e-01
-3.56393933e-01 -2.51019448e-01 -6.92070946e-02 3.72938886e-02
4.35357243e-01 6.76620901e-01 6.95292652e-01 -1.46997213e-01
-2.62842745e-01 2.71688491e-01 4.52947676e-01 -3.69821757e-01
1.77720523e+00 3.04436356e-01 -2.85578072e-01 3.28435838e-01
8.44603658e-01 3.27819958e-02 -1.30682087e+00 4.94692810e-02
3.16606164e-01 -7.35115469e-01 4.68275130e-01 -6.17199004e-01
-1.20703292e+00 4.63443100e-01 6.25327528e-01 4.59265001e-02
1.01724553e+00 -3.13307077e-01 4.14567500e-01 2.44894579e-01
2.51124829e-01 -1.21875978e+00 -4.95688766e-01 5.86515605e-01
7.73231030e-01 -8.69313300e-01 4.97610539e-01 2.20151339e-02
-3.85881871e-01 1.36888981e+00 3.79148364e-01 -4.55624145e-03
8.77641737e-01 8.63432467e-01 -9.78008211e-02 -2.69911826e-01
-7.46758282e-01 8.74404386e-02 -3.41847539e-01 -2.00143903e-01
-4.84321900e-02 1.29490361e-01 -9.71681699e-02 1.07051767e-01
-2.58735776e-01 1.71846926e-01 6.10334456e-01 1.20575500e+00
-2.81127840e-01 -9.22742188e-01 -7.70528555e-01 7.45062947e-01
-1.40209362e-01 6.30411282e-02 6.51832819e-01 3.52006286e-01
-2.31770411e-01 6.23897851e-01 1.90186143e-01 -1.27878562e-01
3.94444048e-01 -3.37898284e-02 4.86261733e-02 -4.01544958e-01
-1.07754493e+00 -1.30975127e-01 1.86699972e-01 -5.59332967e-01
-1.82933714e-02 -8.73242915e-01 -1.29514194e+00 -6.30140603e-01
-1.46265447e-01 3.89890939e-01 1.07065332e+00 1.02357376e+00
3.64049971e-01 5.43281972e-01 8.96734655e-01 -1.34186518e+00
-8.96474779e-01 -9.23188269e-01 -6.54446423e-01 3.32468972e-02
-6.19226620e-02 -6.59180582e-01 -6.73033655e-01 -2.13472359e-02] | [6.370100498199463, 3.353407382965088] |
8d4259d4-720a-4539-921d-c2102a0ff2f1 | noveld-a-simple-yet-effective-exploration | null | null | http://proceedings.neurips.cc/paper/2021/hash/d428d070622e0f4363fceae11f4a3576-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/d428d070622e0f4363fceae11f4a3576-Paper.pdf | NovelD: A Simple yet Effective Exploration Criterion | Efficient exploration under sparse rewards remains a key challenge in deep reinforcement learning. Previous exploration methods (e.g., RND) have achieved strong results in multiple hard tasks. However, if there are multiple novel areas to explore, these methods often focus quickly on one without sufficiently trying others (like a depth-wise first search manner). In some scenarios (e.g., four corridor environment in Sec 4.2), we observe they explore in one corridor for long and fail to cover all the states. On the other hand, in theoretical RL, with optimistic initialization and the inverse square root of visitation count as a bonus, it won't suffer from this and explores different novel regions alternatively (like a breadth-first search manner). In this paper, inspired by this, we propose a simple but effective criterion called NovelD by weighting every novel area approximately equally. Our algorithm is very simple but yet shows comparable performance or even outperforms multiple SOTA exploration methods in many hard exploration tasks. Specifically, NovelD solves all the static procedurally-generated tasks in Mini-Grid with just 120M environment steps, without any curriculum learning. In comparison, the previous SOTA only solves 50% of them. NovelD also achieves SOTA on multiple tasks in NetHack, a rogue-like game that contains more challenging procedurally-generated environments. In multiple Atari games (e.g., MonteZuma's Revenge, Venture, Gravitar), NovelD outperforms RND. We analyze NovelD thoroughly in MiniGrid and found that empirically it helps the agent explore the environment more uniformly with a focus on exploring beyond the boundary. | ['Yuandong Tian', 'Joseph E. Gonzalez', 'Kurt Keutzer', 'Yi Wu', 'Xiaolong Wang', 'Huazhe Xu', 'Tianjun Zhang'] | 2021-12-01 | null | https://openreview.net/forum?id=CYUzpnOkFJp | https://openreview.net/pdf?id=CYUzpnOkFJp | neurips-2021-12 | ['montezumas-revenge', 'nethack'] | ['playing-games', 'playing-games'] | [-1.07310705e-01 2.05216885e-01 -8.72456729e-02 2.31226385e-01
-7.96996117e-01 -8.03876758e-01 3.54307443e-01 6.93474114e-02
-8.07731748e-01 1.30580699e+00 -4.01993617e-02 -6.67414546e-01
-5.75577140e-01 -8.71301651e-01 -8.25610161e-01 -9.28969443e-01
-4.74438220e-01 8.34634244e-01 1.72182664e-01 -6.08653486e-01
4.73371178e-01 2.00540647e-01 -1.30563939e+00 -2.66968787e-01
1.11755502e+00 4.95870948e-01 7.36691952e-01 3.70638609e-01
1.23525113e-02 4.42880601e-01 -7.06750929e-01 1.82502285e-01
4.66527343e-01 -3.77957344e-01 -1.06122291e+00 -1.04044192e-01
-1.45133153e-01 -2.37191007e-01 4.23510885e-03 8.88208866e-01
6.23181105e-01 5.82170427e-01 2.40652323e-01 -1.09599042e+00
8.40374082e-02 1.09954810e+00 -1.10785568e+00 1.10697404e-01
2.20558971e-01 4.48427647e-01 7.96878040e-01 -4.84761924e-01
6.36935234e-01 1.16314745e+00 2.35494733e-01 4.02009964e-01
-1.28999949e+00 -7.79653609e-01 5.25927424e-01 -8.17531198e-02
-1.01079452e+00 1.36289492e-01 5.24822116e-01 3.80136371e-02
9.22182322e-01 3.60845208e-01 9.21939492e-01 1.09914112e+00
4.61011201e-01 9.39666688e-01 1.57446063e+00 -1.91303626e-01
8.12631726e-01 -1.65670902e-01 -3.65030527e-01 4.74598318e-01
2.95233488e-01 4.76155102e-01 -5.01395345e-01 -1.22819908e-01
7.26922333e-01 -1.46780983e-01 -1.58642724e-01 -7.03867257e-01
-1.12510073e+00 1.03419876e+00 3.74610603e-01 2.16324970e-01
-3.69308382e-01 4.38117594e-01 3.20459634e-01 4.69783932e-01
-6.83068810e-03 1.14049184e+00 -4.08239394e-01 -6.93634629e-01
-9.48352993e-01 7.83485830e-01 7.28406489e-01 6.91349566e-01
8.23562682e-01 1.79763302e-01 6.31626397e-02 5.25833666e-01
-1.71087697e-01 1.56595200e-01 4.38536316e-01 -1.13816869e+00
6.73998356e-01 2.01454431e-01 3.30057144e-01 -6.86692774e-01
-8.70880306e-01 -8.76774728e-01 -6.80739045e-01 9.24784064e-01
6.21264637e-01 -6.49820805e-01 -9.24279451e-01 1.82398355e+00
4.10802156e-01 -2.30226710e-01 1.08112246e-01 1.05325127e+00
2.58210570e-01 4.36086923e-01 -2.85196096e-01 -1.85990110e-01
1.09440565e+00 -1.20094419e+00 -4.88989949e-01 -6.17810011e-01
6.37698710e-01 -2.88351685e-01 1.26165009e+00 8.88321877e-01
-1.33073044e+00 -3.10409307e-01 -8.62454593e-01 4.66805339e-01
-3.67401779e-01 -4.36085135e-01 1.02460980e+00 6.20429873e-01
-1.04065335e+00 6.99590921e-01 -8.28340650e-01 -1.72561705e-01
2.36274809e-01 3.33666563e-01 -8.02215189e-02 -2.98195500e-02
-1.15420330e+00 9.62091029e-01 6.05982900e-01 -2.10868623e-02
-1.47606611e+00 -3.35820466e-01 -7.44713128e-01 1.59326375e-01
1.28701735e+00 -3.53253156e-01 1.30773878e+00 -5.92441559e-01
-1.65028358e+00 2.27704361e-01 1.45227283e-01 -6.41430318e-01
8.04076374e-01 -3.47208560e-01 3.60188037e-01 -3.07332963e-01
1.41327351e-01 9.79272485e-01 4.24873143e-01 -1.41269422e+00
-5.92764556e-01 -1.95142493e-01 7.77103186e-01 7.78772712e-01
6.27041981e-03 -5.29804468e-01 -9.86866653e-02 -3.63689989e-01
1.19929932e-01 -1.15082788e+00 -9.95258987e-01 -5.19332469e-01
-4.25362736e-01 -1.54653355e-01 2.63557762e-01 -3.68708298e-02
1.10111558e+00 -1.81226230e+00 5.47254324e-01 4.25887674e-01
2.86631614e-01 -1.88510895e-01 -1.88044369e-01 5.99531412e-01
1.99052870e-01 1.69774041e-01 -4.03155178e-01 -2.21602052e-01
1.19462810e-01 5.73280990e-01 -2.33958289e-01 1.21269792e-01
-3.84409904e-01 8.67615104e-01 -1.34332013e+00 -2.63356268e-01
6.06391430e-02 -1.32942528e-01 -7.77374506e-01 -1.47885874e-01
-3.81471694e-01 5.47931075e-01 -6.24975204e-01 4.66427088e-01
6.40030742e-01 -8.11785981e-02 3.04298222e-01 7.63369739e-01
-3.83814126e-01 2.71446913e-01 -1.46354163e+00 2.09006834e+00
-6.94837689e-01 4.10454005e-01 2.15363503e-01 -9.44652081e-01
7.76238143e-01 -2.22645894e-01 3.87610734e-01 -1.03713226e+00
-1.62878726e-02 3.00147384e-01 2.05518529e-01 -9.96730328e-02
8.78599465e-01 9.87317488e-02 -2.85865456e-01 5.86845160e-01
-4.45781142e-01 -5.07477701e-01 3.44215035e-01 1.41037926e-01
1.47518265e+00 5.12953579e-01 2.96285361e-01 -5.01985669e-01
-5.79267517e-02 4.03353602e-01 7.06605852e-01 1.44237959e+00
-2.01013759e-01 4.06950712e-01 8.50273967e-01 -4.21617925e-01
-7.00677693e-01 -1.04143131e+00 3.64884734e-02 1.12031591e+00
5.33992767e-01 -4.06440437e-01 -6.70675933e-01 -6.58424139e-01
-2.36366447e-02 9.08110499e-01 -8.01717997e-01 4.34361398e-02
-7.37166405e-01 -6.98654354e-01 3.18257838e-01 1.59610704e-01
6.50783539e-01 -1.51811063e+00 -1.46157444e+00 3.34214479e-01
-4.61542606e-02 -4.18535382e-01 -1.43584982e-01 1.07596087e+00
-8.68384659e-01 -8.59209180e-01 -7.81918705e-01 -3.67409021e-01
4.08908427e-01 3.45077515e-01 1.18389833e+00 -2.05594495e-01
-2.00346470e-01 9.51913521e-02 -3.65442842e-01 -2.32126355e-01
1.41738251e-01 4.07944530e-01 -7.78706521e-02 -8.92810643e-01
-5.67427389e-02 -5.92625976e-01 -6.59778535e-01 4.36123908e-01
-5.49843848e-01 1.14574723e-01 6.86071098e-01 1.00855803e+00
6.51558518e-01 5.24108589e-01 4.73079741e-01 -7.23756552e-01
9.38368976e-01 -7.20295072e-01 -7.61760712e-01 -5.21593578e-02
-5.37182868e-01 2.73978889e-01 4.97269690e-01 -6.63763165e-01
-8.91613901e-01 -2.05984637e-01 9.19915140e-02 -3.19238931e-01
-1.64987997e-03 6.34720206e-01 2.62652695e-01 -6.02484569e-02
8.74190509e-01 2.45470747e-01 -1.72040686e-01 -3.43827814e-01
1.07550770e-01 -4.43666130e-02 8.91816914e-02 -1.04650998e+00
8.73969793e-01 2.18850479e-01 -1.96818803e-02 -4.06499952e-01
-5.48387885e-01 -5.10572009e-02 -2.91305538e-02 -1.88598543e-01
5.56547046e-01 -7.23661363e-01 -1.02715027e+00 1.59237668e-01
-6.30306065e-01 -1.16368759e+00 -6.30717933e-01 4.22018230e-01
-7.83185124e-01 -2.07150728e-02 -2.63032377e-01 -1.05882955e+00
1.76083744e-01 -1.54135311e+00 7.79281855e-01 5.67048609e-01
-2.43238881e-01 -6.80055797e-01 3.01631033e-01 9.16067213e-02
5.53031504e-01 4.60962623e-01 5.09418666e-01 -3.11850011e-01
-6.68467402e-01 5.80766916e-01 1.57236457e-01 -4.83826667e-01
-6.05152212e-02 -6.04033291e-01 -4.02836949e-01 -6.48079634e-01
-1.25425950e-01 -7.72026002e-01 9.69603240e-01 4.63373363e-01
1.40327358e+00 -1.14310257e-01 -3.43482852e-01 6.06955826e-01
1.22000968e+00 6.58175886e-01 6.35628521e-01 1.07144630e+00
1.73275679e-01 6.09188080e-01 1.10671747e+00 7.47113943e-01
2.84160614e-01 7.04358280e-01 1.02573431e+00 -1.92526966e-01
4.25461560e-01 -5.06486632e-02 4.52650279e-01 5.29993810e-02
-1.00404263e-01 -2.77845085e-01 -8.45603704e-01 6.22003317e-01
-1.85482311e+00 -8.34674835e-01 2.55390674e-01 2.05781960e+00
1.03960180e+00 6.05436265e-01 2.71301180e-01 -6.83893040e-02
2.94478714e-01 3.58538926e-01 -1.02210152e+00 -8.04111600e-01
-7.87307844e-02 4.84886140e-01 6.48447156e-01 6.45611942e-01
-6.35079741e-01 1.15089929e+00 5.59292173e+00 1.10753036e+00
-7.45845318e-01 -6.53403113e-03 5.97162187e-01 -6.85206473e-01
-5.02916276e-01 1.91467389e-01 -6.80826485e-01 3.11837763e-01
4.10224795e-01 -6.93215430e-02 1.00367367e+00 9.35537338e-01
1.68752462e-01 -8.73504996e-01 -7.66510963e-01 6.37023747e-01
-4.51937765e-01 -1.09919083e+00 -5.52850127e-01 4.36663628e-01
9.53874528e-01 2.53286008e-02 3.41055363e-01 7.23440945e-01
1.07215631e+00 -1.34156382e+00 7.76827335e-01 5.12239002e-02
5.49202085e-01 -1.18585408e+00 5.68378806e-01 6.95599735e-01
-9.38078940e-01 -3.01132590e-01 -3.81508946e-01 -3.26224267e-01
1.17113411e-01 4.18409914e-01 -6.53026223e-01 5.20060956e-01
9.54551458e-01 1.22823119e-02 -2.39614785e-01 1.10644102e+00
-3.35586429e-01 2.62192518e-01 -4.87862378e-01 -2.37444565e-01
1.09794152e+00 -3.56687546e-01 6.91415966e-01 6.39051318e-01
3.12990725e-01 9.33352709e-02 4.06560987e-01 9.81560171e-01
3.80614758e-01 -2.67241359e-01 -6.43893242e-01 1.19292118e-01
3.74990821e-01 1.14056134e+00 -7.86623657e-01 1.84877887e-02
2.25213006e-01 6.63460732e-01 4.92277622e-01 4.88760710e-01
-9.51582730e-01 -3.94083112e-01 5.35027683e-01 4.22603488e-02
2.71217734e-01 -3.51927012e-01 -3.99662048e-01 -7.87460744e-01
-1.50526345e-01 -1.14264417e+00 2.76104391e-01 -5.62162101e-01
-5.48314869e-01 5.73921204e-01 2.21919373e-01 -9.20206726e-01
-2.62871712e-01 -2.18025878e-01 -7.23770320e-01 7.52072871e-01
-1.60749471e+00 -4.53010768e-01 -2.54481316e-01 5.19564688e-01
9.02087629e-01 -1.16826355e-01 4.66263741e-01 -2.51137972e-01
-5.14864385e-01 4.75940108e-01 3.17987621e-01 -4.62406635e-01
4.58229989e-01 -1.56519949e+00 4.12020773e-01 5.30626416e-01
-1.31984413e-01 6.31235600e-01 9.52520013e-01 -8.08882535e-01
-1.33764565e+00 -5.30467808e-01 6.48437068e-02 8.58597457e-02
5.69818795e-01 -3.67600441e-01 -6.86589003e-01 4.77730721e-01
3.76659423e-01 -4.97603297e-01 1.20076150e-01 5.04447639e-01
2.42711201e-01 2.18500197e-01 -9.75009739e-01 9.08530354e-01
1.34640813e+00 1.47538930e-01 -4.11459953e-01 3.32695186e-01
5.81310868e-01 -9.96741951e-01 -3.40861529e-01 3.71511787e-01
4.52482432e-01 -1.29355931e+00 7.87675083e-01 -4.22360420e-01
4.21956211e-01 -2.19175979e-01 1.16136499e-01 -1.78128457e+00
-1.57454222e-01 -1.06390595e+00 -3.23049128e-02 5.41232288e-01
4.41084564e-01 -7.64640927e-01 1.12775910e+00 1.17943123e-01
-2.20809206e-01 -1.27641857e+00 -1.02971196e+00 -1.05881739e+00
3.69851053e-01 -1.66833535e-01 6.32490277e-01 6.17802143e-01
2.99836881e-02 -1.40909582e-01 -4.50810134e-01 -1.36791408e-01
7.42925107e-01 3.30888301e-01 8.70116293e-01 -8.35758865e-01
-7.46702909e-01 -5.81528187e-01 5.02804518e-01 -1.30049717e+00
-9.65760201e-02 -5.63336432e-01 2.18973666e-01 -1.58073771e+00
1.25779703e-01 -1.15439463e+00 -2.13518560e-01 6.92290187e-01
-2.60588881e-02 -2.50193328e-01 3.37189555e-01 -6.38862774e-02
-7.35160232e-01 5.55249572e-01 1.83124065e+00 -1.02137662e-01
-7.02978253e-01 -4.42446321e-02 -8.78409028e-01 6.22537255e-01
1.10480392e+00 -5.28999805e-01 -6.05000377e-01 -3.38285178e-01
6.89947903e-01 5.17502367e-01 1.43461019e-01 -1.13173378e+00
1.48505941e-01 -7.85584986e-01 1.23055086e-01 -6.24170423e-01
4.28452909e-01 -5.23044586e-01 2.03721568e-01 8.38835418e-01
-3.46304566e-01 2.45660529e-01 3.88565063e-01 5.24288535e-01
6.27680123e-02 -4.52408850e-01 5.45056760e-01 -5.09382486e-01
-6.86347365e-01 2.05003135e-02 -5.51051557e-01 3.83699656e-01
1.25153124e+00 -5.60198784e-01 -4.68653530e-01 -4.96865660e-01
-7.29203939e-01 1.02059579e+00 5.03479898e-01 2.61490107e-01
5.17514408e-01 -9.54264879e-01 -4.78762984e-01 -1.30896628e-01
-3.04222226e-01 6.46199465e-01 3.65761697e-01 8.00866663e-01
-3.77954096e-01 1.85861334e-01 -4.88084048e-01 -3.55158389e-01
-8.85842860e-01 4.05428976e-01 3.98899734e-01 -1.03511608e+00
-6.58411562e-01 8.61840487e-01 5.01551747e-01 -4.70089912e-01
4.15971845e-01 -2.44961515e-01 -2.52465665e-01 1.23966217e-01
2.41704345e-01 3.87748748e-01 -2.98976630e-01 3.14349204e-01
-2.76516289e-01 4.55942065e-01 -2.24363312e-01 -4.45018560e-01
1.37499905e+00 1.59419142e-02 3.12243551e-01 3.94234002e-01
5.52131414e-01 2.63669621e-02 -1.60802650e+00 1.49828255e-01
-2.18071207e-01 -5.49659848e-01 1.12303514e-02 -1.05913186e+00
-9.48876798e-01 8.11140895e-01 3.15830022e-01 2.75043577e-01
1.04178143e+00 -2.35460967e-01 5.62074482e-01 6.52761519e-01
9.89079177e-01 -1.36991322e+00 3.59893113e-01 7.35410094e-01
9.00614440e-01 -1.09384239e+00 1.43892393e-01 2.39849642e-01
-9.83966768e-01 9.11769211e-01 1.04698229e+00 -1.19517893e-01
-2.88742054e-02 3.23372602e-01 -4.08372313e-01 -3.21664184e-01
-9.49366868e-01 -3.66796702e-01 -4.80084151e-01 4.78661954e-01
-2.43558511e-01 1.54494599e-01 -4.62262869e-01 1.02500357e-01
-6.17548883e-01 -3.93870384e-01 7.66699612e-01 1.13741088e+00
-8.03485334e-01 -1.05667937e+00 -4.43444133e-01 3.10839772e-01
-1.17823638e-01 -1.10910311e-02 -3.19275968e-02 1.11341012e+00
4.75753359e-02 8.23079586e-01 3.71328481e-02 -1.69238493e-01
3.64125781e-02 -4.96715873e-01 4.54016358e-01 -4.76572096e-01
-9.37496603e-01 1.21590920e-01 1.21273473e-01 -8.89455080e-01
7.70270005e-02 -7.51662791e-01 -1.39680970e+00 -4.17718410e-01
-1.84923232e-01 4.39876080e-01 6.14607334e-01 7.41075277e-01
6.05754294e-02 8.02613258e-01 5.49037218e-01 -9.18237686e-01
-7.52051532e-01 -6.10128820e-01 -6.48724318e-01 -2.29378060e-01
2.29210943e-01 -1.09083915e+00 -3.68267357e-01 -9.98361886e-01] | [3.8701369762420654, 1.7086776494979858] |
cf106dc6-f3a5-4fff-bed2-3ee861856d2f | all-weather-deep-outdoor-lighting-estimation-1 | 1906.04909 | null | https://arxiv.org/abs/1906.04909v1 | https://arxiv.org/pdf/1906.04909v1.pdf | All-Weather Deep Outdoor Lighting Estimation | We present a neural network that predicts HDR outdoor illumination from a single LDR image. At the heart of our work is a method to accurately learn HDR lighting from LDR panoramas under any weather condition. We achieve this by training another CNN (on a combination of synthetic and real images) to take as input an LDR panorama, and regress the parameters of the Lalonde-Matthews outdoor illumination model. This model is trained such that it a) reconstructs the appearance of the sky, and b) renders the appearance of objects lit by this illumination. We use this network to label a large-scale dataset of LDR panoramas with lighting parameters and use them to train our single image outdoor lighting estimation network. We demonstrate, via extensive experiments, that both our panorama and single image networks outperform the state of the art, and unlike prior work, are able to handle weather conditions ranging from fully sunny to overcast skies. | ['Jean-François Lalonde', 'Jonathan Eisenmann', 'Yannick Hold-Geoffroy', 'Sunil Hadap', 'Jinsong Zhang', 'Kalyan Sunkavalli'] | 2019-06-12 | all-weather-deep-outdoor-lighting-estimation | http://openaccess.thecvf.com/content_CVPR_2019/html/Zhang_All-Weather_Deep_Outdoor_Lighting_Estimation_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Zhang_All-Weather_Deep_Outdoor_Lighting_Estimation_CVPR_2019_paper.pdf | cvpr-2019-6 | ['lighting-estimation'] | ['computer-vision'] | [ 4.80257899e-01 -1.98028609e-01 1.72404960e-01 -5.89164019e-01
-5.45837402e-01 -7.60860741e-01 6.78971589e-01 -8.16974103e-01
-6.56899363e-02 6.45081162e-01 1.64553121e-01 -3.33342969e-01
6.08773470e-01 -8.39668095e-01 -1.12412202e+00 -7.30230868e-01
3.73693645e-01 3.34708720e-01 -1.22857057e-01 -1.84429094e-01
-2.04432309e-01 6.37396395e-01 -1.77051926e+00 -6.83255401e-03
5.97359180e-01 9.85459089e-01 2.37186611e-01 1.32761621e+00
4.59784418e-01 1.21593225e+00 -6.45175040e-01 -7.65501559e-02
7.09820211e-01 -4.81023520e-01 -5.14138699e-01 4.09835845e-01
1.22729874e+00 -9.14733112e-01 -7.58855164e-01 7.14732707e-01
3.19029868e-01 2.17068434e-01 4.33829069e-01 -9.09378946e-01
-7.47057259e-01 -1.52187079e-01 -4.10638630e-01 -7.51001537e-02
1.89689100e-01 4.87840980e-01 8.38280320e-01 -5.14999688e-01
6.62919164e-01 1.22626281e+00 6.34285331e-01 3.08039010e-01
-1.20866966e+00 -4.88915443e-01 -8.56571868e-02 -2.03472286e-01
-9.69796658e-01 -6.20746374e-01 5.35457075e-01 -8.99707749e-02
9.00710762e-01 3.41552883e-01 8.12268138e-01 1.07202888e+00
1.10023759e-01 5.23477197e-01 1.72995055e+00 -5.09823620e-01
1.38903454e-01 -8.50260332e-02 -3.09636801e-01 5.83914220e-01
-2.16735885e-01 2.75218278e-01 -2.87615269e-01 2.24370137e-02
8.76434028e-01 -2.77646706e-02 -4.90767598e-01 -1.26376346e-01
-1.05589235e+00 5.59183478e-01 7.33931065e-01 -2.83102244e-01
-1.44287825e-01 3.90311480e-01 -2.45305389e-01 4.67039436e-01
6.67909622e-01 4.38256204e-01 -4.33668256e-01 3.05130631e-01
-9.83256102e-01 1.33788735e-01 8.07049334e-01 4.74388421e-01
1.18460679e+00 1.89742729e-01 3.21243793e-01 8.70493352e-01
2.65140682e-01 1.25094700e+00 -3.67170647e-02 -1.46430254e+00
1.28198132e-01 1.26522779e-01 2.51822770e-01 -5.96036315e-01
-2.74332523e-01 6.76114336e-02 -6.58531547e-01 5.95430017e-01
3.78795296e-01 -2.75584042e-01 -1.46840894e+00 1.57525241e+00
2.43838578e-01 2.57166564e-01 1.86190531e-01 1.02365720e+00
5.47530830e-01 1.00455594e+00 -3.76326352e-01 8.18253681e-02
9.15789187e-01 -1.08963680e+00 -4.52259302e-01 -6.43924713e-01
2.23894641e-02 -9.74648714e-01 1.26991510e+00 4.69332278e-01
-1.00057387e+00 -5.77422738e-01 -1.17647886e+00 -4.44894522e-01
-4.11460191e-01 4.08956170e-01 5.51520944e-01 6.42559350e-01
-1.53627682e+00 4.16198999e-01 -5.93383074e-01 -6.55286789e-01
8.96353554e-03 5.01073599e-02 -1.82131857e-01 -4.32751387e-01
-1.13897896e+00 1.18326569e+00 -4.61127907e-02 3.14442128e-01
-1.63024998e+00 -5.44838011e-01 -9.17445958e-01 -1.90806329e-01
-1.48541927e-01 -5.01710117e-01 1.18740141e+00 -1.56665218e+00
-1.78743875e+00 1.01850474e+00 -1.10995568e-01 -3.69023383e-01
4.81178105e-01 -2.89974630e-01 -6.20267570e-01 1.72237992e-01
-3.01361203e-01 9.04633880e-01 1.00327110e+00 -1.69049895e+00
-4.30725187e-01 -1.66765422e-01 2.69374400e-01 4.95652348e-01
8.16558599e-02 -2.92823417e-03 -5.40845931e-01 -2.36104444e-01
-2.41969407e-01 -1.23474216e+00 -1.44754857e-01 1.16353288e-01
-5.36179483e-01 6.05691731e-01 9.56583858e-01 -9.30840492e-01
3.37819546e-01 -2.05536985e+00 -1.59982845e-01 1.72938958e-01
-4.44340929e-02 9.56625212e-03 -2.71063000e-01 7.91443288e-02
-1.35115147e-01 -4.16618764e-01 -2.73654997e-01 -4.43165302e-01
-5.05800880e-02 2.71341711e-01 -7.69305050e-01 8.51041555e-01
-4.53674421e-02 7.44764924e-01 -5.77622294e-01 -7.53811002e-02
7.28414476e-01 9.69458520e-01 -1.14697546e-01 6.99212015e-01
-6.37987316e-01 6.65586233e-01 3.55000496e-01 5.07173955e-01
8.73890340e-01 -2.39757597e-01 1.78911984e-01 -1.66172430e-01
-2.16727674e-01 1.81149021e-01 -9.03391182e-01 1.54445124e+00
-8.37265253e-01 1.25954604e+00 2.34271586e-02 -1.50143191e-01
1.03197300e+00 1.31679714e-01 4.84716743e-02 -9.25200820e-01
-4.17921320e-02 1.48063347e-01 -5.46178281e-01 -4.36005324e-01
6.19859099e-01 -5.85787110e-02 1.92893788e-01 5.81165075e-01
-1.15904182e-01 -6.53742969e-01 -1.21599577e-01 -1.36282220e-02
9.30614710e-01 5.12431264e-01 -9.78050008e-02 -6.41555898e-03
3.09833169e-01 -1.98552713e-01 2.45343730e-01 6.73744142e-01
2.14148551e-01 1.07809937e+00 8.55159089e-02 -9.85192835e-01
-1.58724988e+00 -1.21335685e+00 -1.22725300e-01 1.17957246e+00
5.96853867e-02 3.22346300e-01 -6.70774758e-01 -3.20289552e-01
-2.24683717e-01 8.64249051e-01 -8.10148299e-01 3.01046222e-01
-6.69889808e-01 -9.53348339e-01 4.52959061e-01 2.21226856e-01
7.76036561e-01 -1.14556992e+00 -5.95362008e-01 -3.83779377e-01
-3.08419466e-01 -1.17579901e+00 -1.94324374e-01 2.93702841e-01
-2.97319204e-01 -9.70990419e-01 -7.76114702e-01 -5.68183720e-01
6.45879090e-01 3.55971515e-01 1.60978174e+00 -9.78520662e-02
-5.44308543e-01 3.75515044e-01 -1.34970192e-02 -2.05289558e-01
-5.94084740e-01 -2.60980695e-01 -1.35640815e-01 4.16393466e-02
7.81809613e-02 -5.19956648e-01 -8.87216032e-01 3.46904188e-01
-1.09291577e+00 2.30926245e-01 3.20576668e-01 2.82401502e-01
6.44869864e-01 3.29311155e-02 -1.33007631e-01 -7.89704263e-01
-1.73053026e-01 -2.63049453e-01 -1.05725098e+00 4.29370522e-01
-3.32788318e-01 -1.90598533e-01 7.33020902e-01 -2.85652056e-02
-1.49074900e+00 3.28626931e-01 1.44698676e-02 -5.10073662e-01
-5.34323156e-01 -5.02768636e-01 -1.84838295e-01 -2.69462436e-01
7.93497086e-01 2.92366773e-01 -3.89009327e-01 -2.44895741e-01
8.28483105e-01 6.88513756e-01 8.85673463e-01 -4.76874769e-01
1.09363174e+00 1.07903242e+00 7.13109449e-02 -8.48896742e-01
-1.13546968e+00 -2.32254028e-01 -4.83831584e-01 -2.94990033e-01
1.09979606e+00 -1.33632708e+00 -2.97158539e-01 8.83952439e-01
-7.81655788e-01 -1.29465914e+00 -1.71823815e-01 1.70019984e-01
-6.81947768e-01 2.86593996e-02 -6.19794130e-01 -7.13852465e-01
-1.58128306e-01 -8.95983279e-01 1.35236824e+00 3.84935319e-01
3.04549038e-01 -1.00155330e+00 6.25789642e-01 2.69084364e-01
4.49579448e-01 4.47557449e-01 6.21961951e-01 2.26758569e-01
-8.94627631e-01 7.43640363e-02 -5.02620399e-01 5.17388761e-01
1.43414825e-01 3.39819670e-01 -1.60490322e+00 -3.40046078e-01
6.84787631e-02 -8.27780128e-01 1.06980777e+00 5.27375400e-01
1.08856690e+00 -2.22747415e-01 2.26097420e-01 1.32642615e+00
1.94211495e+00 -2.32239455e-01 1.02315891e+00 5.22119224e-01
9.25933242e-01 4.21551794e-01 3.46493095e-01 1.11343160e-01
5.30311882e-01 4.44456935e-01 7.14055955e-01 -7.90729642e-01
-5.04512846e-01 -1.75279170e-01 4.90745723e-01 2.47710258e-01
-6.89650029e-02 -4.90265071e-01 -6.17214680e-01 3.99186403e-01
-1.42432058e+00 -6.35709405e-01 9.25886035e-02 2.26251197e+00
7.82218575e-01 -3.59616011e-01 -1.70643121e-01 -4.10438359e-01
4.79814231e-01 8.05085778e-01 -7.45540321e-01 -5.38451374e-01
-6.66026890e-01 1.21657982e-01 1.00657535e+00 6.79308116e-01
-1.23373806e+00 1.08813262e+00 7.19643307e+00 4.33273204e-02
-1.31529093e+00 -6.84245378e-02 1.05169988e+00 -6.72447905e-02
-3.63621891e-01 -2.18197666e-02 -5.85420787e-01 2.35297754e-01
1.10418606e+00 7.13097095e-01 1.30914831e+00 6.90076649e-01
1.62428990e-01 -2.54888743e-01 -9.13387656e-01 7.75435865e-01
5.74054837e-01 -8.61790180e-01 -1.86458364e-01 -4.79405038e-02
1.43737376e+00 6.95971012e-01 3.95091921e-01 -1.02531470e-01
1.16560543e+00 -1.14270508e+00 4.59475458e-01 5.99607468e-01
1.13140917e+00 -3.81583780e-01 2.82960832e-01 -1.35225430e-01
-8.75424385e-01 5.13998531e-02 -4.98902529e-01 1.44066727e-02
1.38148531e-01 4.79041815e-01 -5.64353108e-01 9.64884087e-02
9.80535746e-01 7.31312633e-01 -7.48156190e-01 6.24902010e-01
-7.47781098e-01 4.67879087e-01 -4.54372883e-01 6.75059199e-01
6.67985380e-02 -3.03322017e-01 8.82921070e-02 1.06644714e+00
2.76294380e-01 -1.82880461e-01 -1.30351171e-01 7.39676476e-01
-3.90066206e-01 -3.31704557e-01 -7.24794626e-01 4.07873034e-01
7.98370391e-02 1.54626024e+00 -4.78058636e-01 -3.99491042e-01
-4.00401115e-01 1.20350742e+00 1.83978036e-01 8.44322503e-01
-9.40462887e-01 -9.22708288e-02 6.56094015e-01 -1.50535226e-01
3.79123122e-01 -1.30571663e-01 -1.07431591e-01 -1.43225002e+00
-4.67034042e-01 -1.01625025e+00 2.91795675e-02 -1.72838867e+00
-1.01700366e+00 5.33670485e-01 -5.26362419e-01 -9.35480475e-01
4.06027660e-02 -6.97128057e-01 -7.72823930e-01 7.86385417e-01
-2.01674604e+00 -1.41211855e+00 -7.87268162e-01 6.14652932e-01
4.41517830e-01 3.01223576e-01 8.07007730e-01 1.18554495e-01
-4.65736538e-01 -2.96252798e-02 5.96044064e-01 7.06048906e-02
1.09004891e+00 -1.66198027e+00 8.22885633e-01 1.16456521e+00
2.58261412e-01 2.05131069e-01 9.23338711e-01 -2.81125993e-01
-1.41942811e+00 -1.40736365e+00 4.37684059e-01 -7.05403149e-01
5.21151662e-01 -4.29741800e-01 -4.83602464e-01 1.16214299e+00
5.05332649e-01 2.41210461e-01 2.29842186e-01 -1.62206084e-01
-7.80368805e-01 -4.89469975e-01 -1.08337295e+00 6.37536645e-01
5.31716287e-01 -6.80827260e-01 -2.96557158e-01 6.74574256e-01
7.59837627e-01 -6.57595158e-01 -7.15760112e-01 2.60446012e-01
6.49989605e-01 -1.15223873e+00 1.12905622e+00 -1.04182206e-01
6.98240519e-01 -5.29239058e-01 -4.77159739e-01 -1.48156488e+00
9.43985358e-02 -5.57847083e-01 -5.27302809e-02 8.08693886e-01
2.74057180e-01 -5.18214524e-01 4.47285384e-01 5.63474119e-01
5.73455393e-02 -1.45472154e-01 -3.39217603e-01 -4.08408761e-01
-3.49634215e-02 -9.87359360e-02 7.20750868e-01 7.09074616e-01
-1.03985381e+00 1.16759650e-02 -9.70496237e-01 4.31758791e-01
9.29673254e-01 5.18223405e-01 1.20590866e+00 -7.80568182e-01
-4.57161248e-01 1.82526976e-01 1.30600706e-01 -9.32137012e-01
1.96349785e-01 -2.58030295e-01 5.91531217e-01 -1.54768014e+00
2.54022330e-01 -3.66086602e-01 -2.18793079e-01 4.37725276e-01
-8.19125324e-02 9.84405696e-01 2.08119769e-03 3.36599767e-01
-7.79180467e-01 2.67042071e-01 1.04858553e+00 1.11694913e-02
-6.95620403e-02 -3.71286899e-01 -1.43080160e-01 7.69028544e-01
8.50549221e-01 -1.64787143e-01 -2.19178781e-01 -8.90239298e-01
5.41154087e-01 -1.93797320e-01 5.88873625e-01 -1.09235203e+00
-1.51528269e-01 -1.46801814e-01 1.03328013e+00 -3.92618060e-01
6.05896473e-01 -7.49162734e-01 3.30455571e-01 1.98662587e-04
-4.06842828e-01 -2.93449372e-01 -9.83676985e-02 4.39867556e-01
2.28050694e-01 3.89946163e-01 1.04979420e+00 -2.11283609e-01
-7.03781724e-01 3.22259188e-01 -7.31312335e-02 -5.16137555e-02
4.91457760e-01 1.43741757e-01 -9.68372881e-01 -5.74510574e-01
-3.14828813e-01 -2.14896686e-02 1.35393357e+00 3.16902101e-01
2.86799878e-01 -1.07205057e+00 -4.84177530e-01 4.42649126e-01
1.42425317e-02 4.99792807e-02 2.56253272e-01 3.87592733e-01
-1.17442393e+00 2.23045766e-01 -2.71930486e-01 -5.06712496e-01
-9.78416562e-01 4.03685600e-01 7.16449201e-01 -4.41192165e-02
-6.72243655e-01 5.55335462e-01 4.87359732e-01 -7.44401157e-01
5.94334342e-02 -1.81869164e-01 1.38431534e-01 -5.31137228e-01
5.50638914e-01 1.13968298e-01 -8.25244039e-02 -8.05480957e-01
3.83147560e-02 7.34678268e-01 4.16571259e-01 -2.64660299e-01
1.43939149e+00 -5.33139944e-01 -3.77775252e-01 5.96269786e-01
1.31641066e+00 -4.11743559e-02 -2.13025045e+00 -1.06381655e-01
-9.50543761e-01 -6.54750407e-01 2.26433799e-01 -1.05950248e+00
-1.26538873e+00 7.79924929e-01 8.80394280e-01 -1.84893422e-02
1.42794907e+00 -1.40144348e-01 8.62694144e-01 5.37813008e-01
7.53516406e-02 -1.04082060e+00 -4.11526673e-02 4.67905432e-01
4.02876645e-01 -1.42991018e+00 9.71341133e-02 3.46674919e-01
-5.92619300e-01 1.20050120e+00 5.27183115e-01 -4.46553588e-01
3.59563470e-01 3.07681829e-01 7.22328782e-01 -1.00361332e-01
-5.38906097e-01 -1.82739571e-01 1.88409492e-01 4.66315955e-01
9.35750380e-02 7.97339082e-02 7.70843506e-01 -6.95951045e-01
-2.78062612e-01 -3.91148664e-02 6.03593469e-01 3.55037212e-01
-5.67907870e-01 -5.28770387e-01 -6.14428639e-01 1.26146153e-01
-4.21983331e-01 -1.76223010e-01 -4.37758327e-01 5.51555336e-01
2.26312488e-01 9.25726950e-01 2.31197774e-01 -6.75252229e-02
-2.02795953e-01 -2.17152908e-01 5.40927112e-01 -1.40063614e-01
-2.41291910e-01 -5.34497350e-02 3.38329747e-02 -7.99543858e-01
-6.21338189e-01 -3.80823106e-01 -4.79718238e-01 -3.17182183e-01
-8.09352845e-02 -4.64567035e-01 1.09622443e+00 7.94678807e-01
-1.94297910e-01 3.66367400e-01 1.11476505e+00 -1.20282137e+00
4.78356704e-02 -7.35411763e-01 -7.29859650e-01 4.18577790e-01
8.32129240e-01 -3.95728014e-02 -8.54398727e-01 3.24151605e-01] | [9.827959060668945, -2.939762830734253] |
3897ddc8-cc6d-4ab9-bcb6-fbd5449cc268 | sentence-ambiguity-grammaticality-and | 2210.06928 | null | https://arxiv.org/abs/2210.06928v2 | https://arxiv.org/pdf/2210.06928v2.pdf | Sentence Ambiguity, Grammaticality and Complexity Probes | It is unclear whether, how and where large pre-trained language models capture subtle linguistic traits like ambiguity, grammaticality and sentence complexity. We present results of automatic classification of these traits and compare their viability and patterns across representation types. We demonstrate that template-based datasets with surface-level artifacts should not be used for probing, careful comparisons with baselines should be done and that t-SNE plots should not be used to determine the presence of a feature among dense vectors representations. We also show how features might be highly localized in the layers for these models and get lost in the upper layers. | ['Ondřej Bojar', 'Vilém Zouhar', 'Sunit Bhattacharya'] | 2022-10-13 | null | null | null | null | ['sentence-ambiguity'] | ['miscellaneous'] | [ 2.03075260e-01 1.90512672e-01 1.76232204e-01 -6.17443621e-01
-5.39669573e-01 -7.60708153e-01 7.37076938e-01 3.52598071e-01
-4.82658803e-01 2.81200111e-01 6.76812768e-01 -3.67464274e-01
-1.60772637e-01 -5.87973475e-01 -4.55603331e-01 -4.26751345e-01
-2.37303838e-01 4.07901108e-01 -5.98071739e-02 -3.55046302e-01
4.29340005e-01 3.41108203e-01 -1.39591467e+00 5.60467720e-01
6.99924409e-01 5.53562105e-01 -5.15893437e-02 4.08639550e-01
-3.57590765e-01 4.86084372e-01 -8.58419776e-01 -5.27492464e-01
-2.86838003e-02 -3.03880185e-01 -6.25085652e-01 -6.81690648e-02
1.02901447e+00 -2.58344293e-01 -1.04758851e-01 1.03722370e+00
4.16056573e-01 -7.21921623e-02 9.41807032e-01 -5.43985307e-01
-7.26551175e-01 8.87872338e-01 -4.70919639e-01 7.21578419e-01
2.54556209e-01 2.22316548e-01 1.25371861e+00 -1.05279994e+00
7.99418926e-01 1.50407875e+00 9.18917239e-01 3.32044542e-01
-1.59802675e+00 -5.14213383e-01 4.98631269e-01 -1.78408340e-01
-1.25824475e+00 -8.39960098e-01 7.17992544e-01 -5.80697536e-01
1.34837818e+00 2.69405693e-01 4.40395266e-01 1.36393535e+00
2.94837624e-01 5.37542939e-01 1.33780193e+00 -3.20270330e-01
3.58990878e-02 1.77714601e-01 5.49214303e-01 6.62045240e-01
6.93830132e-01 -3.93056124e-03 -5.64267397e-01 -3.24423850e-01
6.85283601e-01 -5.74937582e-01 -1.40131429e-01 -2.76345432e-01
-1.14382100e+00 7.62248933e-01 3.53142470e-01 7.91523933e-01
-3.36505473e-01 1.08284168e-01 4.80344296e-01 3.39837700e-01
4.00954902e-01 8.56154144e-01 -6.76238775e-01 -3.98447782e-01
-1.06420946e+00 2.26279363e-01 4.05133516e-01 9.44357932e-01
6.57777548e-01 3.25417131e-01 -2.29552478e-01 1.03972936e+00
2.20125774e-03 8.91941637e-02 4.64965075e-01 -8.57821405e-01
4.91135031e-01 6.70529723e-01 -1.24085963e-01 -1.30721080e+00
-7.50482380e-01 -5.34033954e-01 -3.77764463e-01 -1.25514772e-02
6.19401395e-01 -9.85192358e-02 -5.78586280e-01 1.82907760e+00
-3.16456854e-01 -4.79403228e-01 -1.50657430e-01 6.52327478e-01
7.50973165e-01 2.82392472e-01 3.41611296e-01 3.93081531e-02
1.27702832e+00 -2.97289312e-01 -5.93655109e-01 -7.91299224e-01
8.67611468e-01 -7.95024991e-01 1.38221323e+00 1.50778592e-01
-1.32339239e+00 -6.13188922e-01 -9.12103057e-01 -3.74481082e-01
-4.50796634e-01 1.44129455e-01 8.84569526e-01 8.07629108e-01
-1.16137457e+00 8.87106895e-01 -6.09772623e-01 -6.13907814e-01
4.11296546e-01 3.27827036e-01 -4.93032396e-01 1.85649902e-01
-8.01196396e-01 1.20251858e+00 -4.72728312e-02 1.87944561e-01
-3.75425160e-01 -4.51869130e-01 -9.96859848e-01 1.56318128e-01
-3.29492122e-01 -2.29122981e-01 7.51385272e-01 -9.44715381e-01
-9.43006456e-01 1.16521382e+00 -3.85248154e-01 1.05360799e-01
-6.76396787e-02 -2.57720277e-02 -2.92008638e-01 -1.89669922e-01
3.19193490e-02 5.51317692e-01 4.58599478e-01 -1.26448178e+00
-1.02868952e-01 -5.45161963e-01 2.10544020e-01 8.90738145e-02
-2.95089692e-01 2.37696156e-01 1.13226697e-01 -7.96683311e-01
4.95518595e-01 -5.94201863e-01 1.18277576e-02 -2.10904721e-02
-3.21974248e-01 -2.70316750e-01 1.87585711e-01 -6.55186117e-01
1.05996418e+00 -2.09248257e+00 -5.07802963e-02 1.87354684e-01
1.06567256e-01 -5.63120171e-02 -5.16225755e-01 4.73436654e-01
1.47389714e-02 8.02451193e-01 -1.42443404e-01 -1.86129719e-01
2.39979953e-01 4.87609506e-02 -2.66666621e-01 7.03388929e-01
5.54570675e-01 9.15157497e-01 -6.63452864e-01 -3.50763500e-01
-1.79661512e-01 3.15089196e-01 -6.68703556e-01 -1.95660114e-01
5.24737090e-02 -8.60395879e-02 -1.07665323e-01 6.69563115e-01
6.61538482e-01 1.90129131e-02 2.94279397e-01 -2.24399552e-01
-3.13602746e-01 1.02898383e+00 -9.11193848e-01 1.53483462e+00
-3.57052684e-01 9.21420395e-01 2.21240431e-01 -7.62349665e-01
9.01650786e-01 -4.50489894e-02 -2.55048364e-01 -9.22502398e-01
1.59139350e-01 3.08016866e-01 8.27487409e-01 -2.77581304e-01
5.73527694e-01 -1.39366910e-01 -1.99614540e-01 4.81616825e-01
1.20708831e-01 4.08962183e-02 1.56402126e-01 -1.59504309e-01
1.07319450e+00 2.08929032e-02 -1.01136416e-01 -8.76917064e-01
5.09708514e-03 -1.10076919e-01 5.46715021e-01 9.02077973e-01
-1.34771513e-02 7.66455770e-01 8.03648591e-01 -3.54195684e-01
-1.11608422e+00 -9.19625282e-01 -7.78706074e-01 1.31213474e+00
-3.75391632e-01 -6.08335316e-01 -3.46376091e-01 -4.81181115e-01
-3.74006294e-02 8.77628446e-01 -8.18693697e-01 -2.91026652e-01
-5.51334620e-01 -8.54514003e-01 8.00636232e-01 6.36118352e-01
-3.98554690e-02 -9.48651850e-01 -5.10559857e-01 1.67009428e-01
2.77598709e-01 -8.82221937e-01 -9.19658169e-02 5.30112207e-01
-9.69004095e-01 -8.64578724e-01 -3.81461650e-01 -7.89620459e-01
8.70385289e-01 8.22759792e-02 1.28720927e+00 3.01285923e-01
-3.04788761e-02 2.70816177e-01 -1.50349706e-01 -2.03961819e-01
-3.42904538e-01 1.53770134e-01 5.86163737e-02 -3.82653981e-01
6.01381361e-01 -5.84289610e-01 -4.21770036e-01 1.38509229e-01
-3.27999979e-01 -2.79572546e-01 5.11041045e-01 8.45041811e-01
5.30924723e-02 -3.82723302e-01 3.91405672e-01 -1.04999208e+00
9.38317597e-01 -2.50986069e-01 -1.81146130e-01 2.35206455e-01
-3.49682838e-01 1.13312103e-01 3.09499562e-01 -4.21490341e-01
-5.87302864e-01 -4.18470562e-01 -2.18740076e-01 9.60333943e-02
-2.32464314e-01 3.87810081e-01 -5.13708964e-02 9.93025377e-02
9.47458804e-01 1.04119562e-01 -9.66439024e-02 -6.06337190e-01
9.04951841e-02 6.37925446e-01 1.02675691e-01 -9.10382152e-01
6.69638157e-01 1.56124815e-01 -5.31931579e-01 -9.94908214e-01
-6.35375082e-01 4.48855013e-02 -5.58262110e-01 3.74440581e-01
4.11935478e-01 -8.97962093e-01 -2.58019835e-01 8.79385248e-02
-1.26836026e+00 -2.95546502e-01 -1.90483570e-01 2.05548748e-01
-3.94072026e-01 2.02160731e-01 -6.96240485e-01 -7.84682810e-01
-5.87898456e-02 -1.08961654e+00 1.04257357e+00 6.75301999e-02
-1.03087568e+00 -1.04765868e+00 -2.13948324e-01 1.56969309e-01
6.98006749e-01 -1.48620293e-01 1.40791762e+00 -9.42479730e-01
-1.11549474e-01 -1.74178272e-01 -2.96519458e-01 3.27817281e-03
2.44777754e-01 2.35649809e-01 -1.34483004e+00 -2.84636348e-01
-1.46703273e-01 -4.57884431e-01 8.42476189e-01 3.14702183e-01
1.05247211e+00 -4.27515447e-01 -1.41025528e-01 5.64194083e-01
1.03813386e+00 -4.29133773e-02 5.64313889e-01 2.43685111e-01
4.32444245e-01 9.36626136e-01 -2.23754883e-01 -1.06479235e-01
3.29558969e-01 7.61964023e-01 -2.58643627e-01 4.10342664e-02
-2.13903666e-01 -2.56226629e-01 6.25606298e-01 8.90796363e-01
2.36484408e-01 -1.01167947e-01 -9.77415860e-01 7.17941165e-01
-1.37426829e+00 -9.26863849e-01 -1.98856354e-01 2.00106335e+00
9.66670275e-01 4.24575984e-01 -8.60475004e-02 6.99361339e-02
4.61554229e-01 2.60356873e-01 -3.14770073e-01 -9.89414692e-01
-5.08830726e-01 2.35114470e-01 1.61079526e-01 5.48002481e-01
-7.58513629e-01 1.21946883e+00 8.05649471e+00 3.17896515e-01
-1.11593854e+00 -1.89659558e-02 7.34708965e-01 -2.63764914e-02
-7.72701025e-01 -1.16737699e-02 -7.72248685e-01 2.07696214e-01
9.72097218e-01 6.82708323e-02 2.79235274e-01 5.81683934e-01
8.12523961e-02 1.00767296e-02 -1.24073136e+00 5.28400898e-01
9.67879519e-02 -1.10866582e+00 -4.11283635e-02 1.10008620e-01
3.21416676e-01 1.92567915e-01 2.49035307e-03 5.05600393e-01
2.74982154e-01 -1.35565913e+00 8.49425793e-01 3.08945209e-01
7.04049587e-01 -3.46081555e-01 4.06013429e-01 1.86005071e-01
-7.53397763e-01 3.12401708e-02 -8.21461082e-01 -3.74410957e-01
-1.56203523e-01 2.68055379e-01 -5.13812244e-01 -1.27345428e-01
5.74065030e-01 4.44826931e-01 -1.07046509e+00 9.03793514e-01
1.86625272e-02 7.77476847e-01 -4.59986210e-01 -2.04825088e-01
1.77799776e-01 7.50724971e-02 2.64062405e-01 1.42556381e+00
2.64766723e-01 -4.30392534e-01 -1.81970611e-01 1.26198244e+00
2.11337015e-01 2.51261324e-01 -9.70006227e-01 -2.94592053e-01
5.44576287e-01 1.16656923e+00 -7.51475871e-01 -9.27932188e-02
-6.26687944e-01 8.67032647e-01 7.71228313e-01 2.81226873e-01
-1.90428898e-01 -1.35206372e-01 9.89861310e-01 3.74947846e-01
2.62631893e-01 -4.49474663e-01 -9.54312682e-01 -1.19453216e+00
-4.69916034e-03 -8.78439367e-01 2.11203769e-01 -7.18305349e-01
-1.47413719e+00 6.14797175e-01 -1.07683077e-01 -6.21181607e-01
-3.37622702e-01 -8.65764320e-01 -7.29888678e-01 1.03350794e+00
-1.02963293e+00 -7.81551242e-01 1.05478048e-01 2.37422928e-01
3.61550838e-01 -2.82359600e-01 1.16531706e+00 -1.59764975e-01
-5.17543018e-01 8.64989340e-01 -1.00992925e-01 4.81230408e-01
3.82704973e-01 -1.26826715e+00 8.17257822e-01 6.43559456e-01
5.34023821e-01 1.34107304e+00 8.18661034e-01 -4.98336643e-01
-1.11479020e+00 -5.34450591e-01 1.21929014e+00 -9.37600136e-01
6.57052457e-01 -8.78783822e-01 -1.17318022e+00 7.63379574e-01
3.10018748e-01 -1.59219489e-01 7.12080717e-01 1.01781321e+00
-7.20196903e-01 1.21994682e-01 -9.35623050e-01 6.48721635e-01
1.30238390e+00 -6.60481215e-01 -9.35848713e-01 2.22497791e-01
3.35401028e-01 -5.57474978e-02 -6.66554511e-01 3.12747598e-01
4.18292403e-01 -1.01254880e+00 6.96128547e-01 -7.16850936e-01
3.34964603e-01 -9.03165434e-03 -2.60200083e-01 -1.41388369e+00
-7.08148777e-01 -1.75060689e-01 3.27186972e-01 1.54663765e+00
9.76859331e-01 -3.80810708e-01 6.14693105e-01 5.86909533e-01
-1.30329862e-01 -4.84306127e-01 -1.07337356e+00 -6.98711038e-01
8.51435423e-01 -4.25115854e-01 4.62180465e-01 1.07590330e+00
2.38916069e-01 5.95637560e-01 1.71216354e-01 -2.15065088e-02
1.47254869e-01 -1.67630389e-01 2.36500099e-01 -1.28396785e+00
-1.05096661e-01 -8.28173816e-01 -4.47059125e-01 -6.60136938e-01
2.19776481e-01 -1.04981375e+00 -2.40502745e-01 -1.58192074e+00
5.18278480e-02 -5.10329723e-01 -1.24386169e-01 5.62168300e-01
-3.78812067e-02 1.41303435e-01 1.75092205e-01 1.42724246e-01
-1.91756293e-01 4.48525578e-01 9.86633062e-01 -2.04459861e-01
3.26260366e-02 -4.95537400e-01 -1.21553981e+00 8.29425931e-01
8.37103963e-01 -3.81782591e-01 -3.46287191e-01 -8.16375136e-01
5.08716941e-01 -3.75901788e-01 2.04735920e-01 -1.09598088e+00
-1.48498580e-01 -1.54636493e-02 7.07327962e-01 -3.91156822e-02
3.80140930e-01 -6.14308774e-01 -4.09279913e-01 1.66225314e-01
-6.37504160e-01 3.80351305e-01 4.93013352e-01 1.06770499e-02
8.78480971e-02 -3.51335555e-01 6.18194222e-01 -4.29459035e-01
-2.20705792e-01 -3.29716653e-01 -6.35744274e-01 2.28134960e-01
4.28966910e-01 -2.72724032e-01 -4.52897757e-01 -1.64171934e-01
-5.07977724e-01 -3.56848575e-02 6.68292761e-01 6.49869442e-01
3.57094437e-01 -1.24106359e+00 -7.73827255e-01 2.88646042e-01
2.79139698e-01 -4.19600427e-01 -8.78396630e-02 5.08942544e-01
-2.74618596e-01 2.69251913e-01 2.61439141e-02 -3.61484706e-01
-9.34023619e-01 2.63450831e-01 4.99825239e-01 1.40781403e-01
-7.41645992e-01 9.58154321e-01 1.81529433e-01 -6.00428045e-01
1.47503629e-01 -3.22381705e-01 -2.74086684e-01 3.88042778e-01
4.53897566e-01 -2.19480708e-01 1.24433957e-01 -8.65533888e-01
-6.10075176e-01 5.26867509e-01 -2.73414552e-01 1.66471917e-02
1.42874420e+00 2.17156652e-02 -1.08142346e-01 9.18751895e-01
1.09426093e+00 2.25192398e-01 -1.01681662e+00 8.81754048e-03
3.75975817e-01 -3.03390056e-01 -5.83407432e-02 -9.43112671e-01
-8.44419599e-01 1.15843391e+00 4.98707592e-01 1.51188523e-01
4.60869163e-01 1.85014993e-01 1.62016496e-01 2.93607295e-01
3.56606126e-01 -1.06939971e+00 -4.03001755e-01 7.22310245e-01
1.21246636e+00 -7.96561539e-01 -4.35977541e-02 -2.63860971e-01
-6.81596637e-01 1.09697759e+00 7.20095575e-01 -4.00922567e-01
4.08760548e-01 5.71406364e-01 1.73545584e-01 -2.78594494e-01
-1.07872868e+00 -3.20619315e-01 3.58030885e-01 7.16795087e-01
1.22331023e+00 -7.27778673e-02 -4.89738226e-01 8.60392034e-01
-8.33229661e-01 -5.84356010e-01 5.33175528e-01 8.56380403e-01
-3.36468279e-01 -1.05820477e+00 -2.74501652e-01 7.19902992e-01
-4.08066809e-01 -4.81886268e-01 -8.02382410e-01 7.49284208e-01
1.09840982e-01 7.29231119e-01 4.70478714e-01 -4.06660527e-01
4.41568732e-01 6.79621935e-01 5.59391141e-01 -9.26145971e-01
-1.00723946e+00 -2.54426133e-02 6.17174029e-01 -4.70871806e-01
1.26001975e-02 -9.91475403e-01 -9.62566257e-01 -1.48074999e-01
-1.82066232e-01 -7.32014701e-03 4.53396171e-01 8.15647304e-01
4.46449161e-01 1.51440710e-01 5.04335612e-02 -6.61828756e-01
-5.35292208e-01 -1.32604861e+00 -4.23759371e-01 3.82520825e-01
2.41973639e-01 -6.44435287e-01 -5.73410749e-01 -3.23520184e-01] | [10.811280250549316, 9.420488357543945] |
da5a8380-b88b-467e-8a05-7642eed2ac48 | simple-yet-effective-bridge-reasoning-for | 1909.07597 | null | https://arxiv.org/abs/1909.07597v2 | https://arxiv.org/pdf/1909.07597v2.pdf | Simple yet Effective Bridge Reasoning for Open-Domain Multi-Hop Question Answering | A key challenge of multi-hop question answering (QA) in the open-domain setting is to accurately retrieve the supporting passages from a large corpus. Existing work on open-domain QA typically relies on off-the-shelf information retrieval (IR) techniques to retrieve \textbf{answer passages}, i.e., the passages containing the groundtruth answers. However, IR-based approaches are insufficient for multi-hop questions, as the topic of the second or further hops is not explicitly covered by the question. To resolve this issue, we introduce a new sub-problem of open-domain multi-hop QA, which aims to recognize the bridge (\emph{i.e.}, the anchor that links to the answer passage) from the context of a set of start passages with a reading comprehension model. This model, the \textbf{bridge reasoner}, is trained with a weakly supervised signal and produces the candidate answer passages for the \textbf{passage reader} to extract the answer. On the full-wiki HotpotQA benchmark, we significantly improve the baseline method by 14 point F1. Without using any memory-inefficient contextual embeddings, our result is also competitive with the state-of-the-art that applies BERT in multiple modules. | ['William Yang Wang', 'Hong Wang', 'Wenhan Xiong', 'Shiyu Chang', 'Xiaoxiao Guo', 'Murray Campbell', 'Mo Yu'] | 2019-09-17 | simple-yet-effective-bridge-reasoning-for-1 | https://aclanthology.org/D19-5806 | https://aclanthology.org/D19-5806.pdf | ws-2019-11 | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 3.74370553e-02 5.05013287e-01 8.12670663e-02 -3.20974179e-02
-2.20365810e+00 -1.03071010e+00 3.05631727e-01 5.73912442e-01
-4.91674423e-01 7.20074117e-01 6.05344594e-01 -5.13272107e-01
-2.75552511e-01 -9.37871516e-01 -1.06641626e+00 -3.19602668e-01
4.60480332e-01 9.64978755e-01 6.99030519e-01 -8.11536372e-01
3.39225680e-01 -4.49390918e-01 -1.24223495e+00 7.46227205e-01
1.23153436e+00 1.06523430e+00 2.93509722e-01 9.20910060e-01
-6.10265970e-01 8.00573826e-01 -7.79300034e-01 -6.44297242e-01
-3.18659991e-01 -5.54713964e-01 -1.78709865e+00 -6.96657479e-01
8.08626592e-01 -2.97247440e-01 -2.27687046e-01 7.69552946e-01
4.60461557e-01 3.17240298e-01 5.39153576e-01 -5.30570209e-01
-1.11146259e+00 5.64651370e-01 2.70185322e-02 5.79220176e-01
8.59096408e-01 -2.83809483e-01 1.77499247e+00 -9.46513653e-01
7.36279368e-01 1.12392175e+00 4.08482581e-01 5.60475826e-01
-8.48819256e-01 3.88837010e-02 9.53752771e-02 5.68408787e-01
-1.15498972e+00 -3.35000485e-01 5.22126317e-01 -2.01398060e-02
9.49752688e-01 4.00899202e-01 -1.20498903e-01 1.03832006e+00
-1.50536463e-01 9.06884789e-01 6.66397095e-01 -7.05450654e-01
1.67997807e-01 -7.05089122e-02 7.25115716e-01 5.57002962e-01
-2.45635435e-01 -5.44955432e-01 -4.12622511e-01 -3.23038220e-01
1.66914538e-02 -5.67300141e-01 -6.74601436e-01 1.27779916e-01
-9.09145772e-01 1.20550370e+00 4.12282169e-01 3.29397798e-01
-1.99259534e-01 -2.46168151e-01 3.64167243e-01 5.93191028e-01
4.38400626e-01 7.56298661e-01 -8.29012930e-01 -2.37132832e-01
-7.51368523e-01 6.00942731e-01 1.34407353e+00 7.70558655e-01
7.53201544e-01 -9.76962149e-01 -5.51526964e-01 1.00482070e+00
2.35072300e-01 5.42275727e-01 2.81474501e-01 -1.05426455e+00
1.20260596e+00 7.82935083e-01 3.48679453e-01 -6.17016435e-01
-1.36820391e-01 -3.08372557e-01 -8.79897773e-02 -7.45999277e-01
7.85473287e-01 -1.32771820e-01 -5.35106659e-01 1.49579978e+00
6.68257535e-01 -2.90128775e-02 4.49901611e-01 9.07703221e-01
1.46313548e+00 1.15036035e+00 -1.95670594e-02 2.39026457e-01
1.79292870e+00 -1.43710446e+00 -5.84693730e-01 -4.98064220e-01
8.45970988e-01 -8.63392591e-01 1.41397512e+00 -1.74052287e-02
-1.13393843e+00 -4.17850643e-01 -8.49375129e-01 -7.16003478e-01
-4.47611928e-01 -3.21410224e-02 -2.25769356e-01 6.38761520e-02
-8.83625209e-01 1.40146866e-01 -1.05505519e-01 -2.28925630e-01
8.59379545e-02 -2.65334100e-01 -1.59335330e-01 -5.40133595e-01
-1.72878718e+00 1.10792601e+00 5.67052700e-02 -1.39119774e-01
-7.73851216e-01 -1.00680542e+00 -1.03558123e+00 2.02923357e-01
5.51878989e-01 -8.89046490e-01 1.69116282e+00 -5.48833728e-01
-1.30087137e+00 9.06838596e-01 -5.24828374e-01 -5.38763523e-01
1.84570812e-02 -6.75907910e-01 -4.30801213e-01 7.63868392e-01
5.12604952e-01 4.02642667e-01 7.70579398e-01 -1.10126126e+00
-6.99952245e-01 -3.92490178e-01 5.70041299e-01 3.85086656e-01
-9.28969458e-02 5.46999648e-02 -4.55817074e-01 -1.33418292e-01
-6.62628040e-02 -5.67348361e-01 2.19044149e-01 -4.43999290e-01
-3.36649686e-01 -8.51733983e-01 5.75170040e-01 -1.28698540e+00
1.43292797e+00 -2.00720501e+00 2.13474974e-01 -2.40292966e-01
1.64078370e-01 2.73840725e-01 -4.44060534e-01 7.59368300e-01
4.71712291e-01 -1.55281708e-01 -3.00513625e-01 -1.19539499e-01
2.50967383e-01 1.80468440e-01 -1.01630449e+00 1.99913718e-02
2.10634992e-01 1.12546849e+00 -1.03129864e+00 -5.14734983e-01
-3.35122138e-01 1.98655456e-01 -4.77882802e-01 5.81959248e-01
-8.72753918e-01 4.75307494e-01 -7.10633039e-01 3.41513485e-01
3.34155828e-01 -4.35011595e-01 -4.53092575e-01 1.01223953e-01
1.94622442e-01 1.08351684e+00 -7.51202941e-01 2.14617586e+00
-7.79026687e-01 4.80391622e-01 1.22001424e-01 -8.34360421e-01
8.66020381e-01 5.01382053e-01 -8.58699307e-02 -1.01816952e+00
-1.59676239e-01 5.27336895e-01 -3.82619739e-01 -7.84198046e-01
7.55268037e-01 1.31122947e-01 -4.62805361e-01 4.44894612e-01
2.46173888e-01 -1.09730057e-01 9.92162153e-02 4.66016710e-01
1.48865497e+00 1.76197290e-01 1.84058677e-02 -1.11514598e-01
8.84294808e-01 4.46149945e-01 -4.37268838e-02 8.88853371e-01
-1.20374314e-01 6.98441565e-01 3.07379156e-01 -7.56873786e-02
-7.37356961e-01 -1.27743757e+00 -4.22842540e-02 1.48176086e+00
3.22956532e-01 -2.64379263e-01 -1.04583216e+00 -9.39402997e-01
-1.54647037e-01 1.00210488e+00 -5.01260102e-01 -5.43054305e-02
-1.03505480e+00 8.24265182e-02 6.61493778e-01 2.51912326e-01
3.61461639e-01 -1.06497562e+00 -4.29261535e-01 3.96448076e-01
-1.10364306e+00 -1.16362822e+00 -4.30313826e-01 2.01490999e-04
-5.38345814e-01 -1.12787080e+00 -7.51027703e-01 -1.13943410e+00
1.90911770e-01 5.62624894e-02 1.81792355e+00 2.45114923e-01
1.45974606e-01 5.53469002e-01 -7.65851021e-01 -2.98678875e-01
-4.13954407e-01 4.82184380e-01 -7.95745432e-01 -1.71234533e-01
6.52873099e-01 -4.07069772e-02 -8.33346725e-01 2.21390024e-01
-8.50327015e-01 -4.00918514e-01 2.09178314e-01 8.47720802e-01
7.98996329e-01 -3.92625183e-01 1.08596265e+00 -7.44903982e-01
9.02821064e-01 -8.06962550e-01 -1.95586890e-01 6.79020524e-01
1.05953477e-01 2.20029473e-01 8.60185385e-01 -1.62354469e-01
-1.21898901e+00 -7.02789426e-01 -6.66455507e-01 1.96372971e-01
-2.16196790e-01 6.40961528e-01 -1.65566102e-01 4.36136872e-01
7.77326643e-01 1.69517875e-01 -4.03310984e-01 -7.29400158e-01
8.08260441e-01 8.66545379e-01 6.38112366e-01 -7.65406251e-01
6.34608626e-01 3.09942752e-01 -6.12420917e-01 -6.89873755e-01
-1.91703963e+00 -9.63879168e-01 -3.55918080e-01 8.92472044e-02
1.04153299e+00 -8.36306393e-01 -3.86910945e-01 -2.10189357e-01
-1.51896167e+00 -1.89092800e-01 -4.37490910e-01 2.32343506e-02
-5.52531600e-01 3.06597173e-01 -6.81499541e-01 -4.68841106e-01
-7.27064967e-01 -6.24184787e-01 1.39587486e+00 2.91282535e-01
-4.57924753e-01 -1.09047580e+00 3.81455094e-01 1.24642646e+00
2.83061326e-01 -1.99938625e-01 1.34762001e+00 -1.06611669e+00
-6.29764080e-01 -2.72110254e-01 -7.65134096e-02 3.19985151e-01
-3.48805040e-01 -6.52590811e-01 -9.66200709e-01 2.64989655e-03
1.58833414e-01 -8.05842519e-01 8.95858824e-01 4.37626392e-02
7.03014374e-01 -4.25095171e-01 -7.21723810e-02 -1.72059275e-02
1.36294556e+00 -2.41118923e-01 6.84870660e-01 5.74090958e-01
3.34059149e-01 7.48884916e-01 8.16038311e-01 -7.84344450e-02
9.64743674e-01 4.53055918e-01 3.71397227e-01 2.01816797e-01
-2.88925231e-01 -5.82477331e-01 3.75450820e-01 1.04528439e+00
6.05093360e-01 -4.72574651e-01 -8.80654156e-01 1.15928423e+00
-1.65413260e+00 -7.44886577e-01 -3.17820728e-01 1.90976715e+00
1.14884686e+00 -7.60087818e-02 -1.89241096e-01 -7.80616105e-02
3.79982561e-01 1.92787752e-01 -3.37380230e-01 -4.94473070e-01
-2.07977998e-03 6.99505568e-01 -1.65663302e-01 7.13213623e-01
-1.00648701e+00 8.22451591e-01 4.97928810e+00 9.44885373e-01
-3.05118889e-01 5.22401392e-01 2.61441737e-01 2.19029143e-01
-4.85147089e-01 1.79188982e-01 -1.27739573e+00 2.49293551e-01
1.32285690e+00 1.64002910e-01 2.81667173e-01 5.06492972e-01
-2.85657465e-01 -2.23619029e-01 -1.30951464e+00 4.21239585e-01
4.55373049e-01 -1.34938025e+00 1.58607975e-01 -5.32556355e-01
6.46683455e-01 5.96843883e-02 -2.61824168e-02 8.25082779e-01
2.27651954e-01 -1.00437069e+00 5.87936342e-01 4.49133873e-01
4.89375919e-01 -5.12170315e-01 8.91699493e-01 5.98442256e-01
-9.72168982e-01 -1.15756415e-01 -4.52627689e-01 -3.05324942e-02
3.84573013e-01 3.55327994e-01 -5.76183438e-01 8.31245840e-01
7.31105268e-01 1.28629729e-01 -5.47586262e-01 9.09864366e-01
-6.64220810e-01 9.96328115e-01 -1.75750136e-01 -3.07293087e-01
6.90744638e-01 4.81504761e-02 6.06106639e-01 9.03266728e-01
2.69208372e-01 2.23934844e-01 -1.36607140e-01 8.28419805e-01
-3.94883007e-01 2.99061209e-01 -2.48376369e-01 1.55476198e-01
5.70071399e-01 8.16484809e-01 3.70707624e-02 -2.59220779e-01
-7.10596204e-01 1.05020511e+00 7.66754150e-01 3.73251438e-01
-6.08945549e-01 -7.64465868e-01 3.87103468e-01 -1.14837140e-01
7.46099412e-01 8.87610912e-02 -7.09435120e-02 -1.08401370e+00
6.10036731e-01 -1.11858523e+00 1.12953424e+00 -7.56932974e-01
-1.59837854e+00 6.12558603e-01 -4.52281624e-01 -9.17204857e-01
-1.55246899e-01 -3.42040390e-01 -6.46582663e-01 1.09721804e+00
-2.00849128e+00 -1.02371049e+00 -6.96195289e-03 6.92438066e-01
7.81280518e-01 3.23823631e-01 1.10193920e+00 2.92958915e-01
-5.63488193e-02 4.96437192e-01 3.28211725e-01 2.81243771e-01
7.29021966e-01 -1.31269312e+00 2.89615244e-01 7.06700385e-01
4.30705309e-01 5.22696555e-01 6.19927645e-01 -3.46510470e-01
-1.39424181e+00 -9.63845909e-01 1.65138555e+00 -1.26486301e+00
9.03855503e-01 -2.12364987e-01 -1.28009081e+00 5.03264308e-01
4.78064805e-01 -2.13731170e-01 6.76651835e-01 2.87279546e-01
-5.16289949e-01 -5.09311296e-02 -8.89996588e-01 4.60259825e-01
6.16696775e-01 -8.78397942e-01 -1.68933058e+00 4.09554154e-01
1.44379938e+00 -5.77530801e-01 -8.93456042e-01 8.42460468e-02
-1.66412160e-01 -3.73485059e-01 1.16015375e+00 -7.81208158e-01
7.62510955e-01 -2.15789974e-01 -2.58139938e-01 -1.26059723e+00
3.48201215e-01 -4.43513155e-01 -4.62216139e-01 1.33756256e+00
8.16037476e-01 -3.68759781e-01 2.80393422e-01 4.45718825e-01
-4.13727432e-01 -7.60150492e-01 -1.47453892e+00 -4.90454286e-01
6.45022511e-01 -2.95442939e-01 5.89570701e-01 3.85489851e-01
8.56839940e-02 9.42707360e-01 1.63461730e-01 3.91246706e-01
3.12239677e-01 5.58493316e-01 5.14078259e-01 -1.04755414e+00
-2.32163563e-01 -4.61442024e-02 2.32316867e-01 -1.76282287e+00
3.01473618e-01 -8.33285213e-01 3.71692002e-01 -2.05790067e+00
-9.53557491e-02 -3.28948289e-01 -1.44571900e-01 -1.22934528e-01
-4.43954587e-01 -1.59869701e-01 -1.26584560e-01 5.96318431e-02
-1.22144926e+00 5.86832941e-01 1.40000546e+00 -3.58116895e-01
3.20507139e-02 -6.94961324e-02 -8.97500098e-01 5.00081420e-01
5.48138678e-01 -5.50734103e-01 -3.82955730e-01 -9.74929035e-01
4.73198235e-01 4.87874568e-01 3.51541519e-01 -6.65596783e-01
4.37550336e-01 3.29102188e-01 -2.25784034e-01 -5.95811427e-01
3.65173429e-01 -6.12440646e-01 -8.16485703e-01 -1.47194669e-01
-7.71515131e-01 -1.46196410e-02 2.20409245e-03 6.86676025e-01
-5.73758125e-01 -8.43774319e-01 3.06953579e-01 -2.48209596e-01
-4.91398752e-01 -5.49970455e-02 -1.80392981e-01 1.20399559e+00
4.98477608e-01 3.78672600e-01 -9.35158372e-01 -4.63123173e-01
-4.42130059e-01 4.90377992e-01 -6.09500892e-02 5.01873374e-01
5.83114445e-01 -9.35524046e-01 -9.61147487e-01 -3.63512039e-01
3.76018047e-01 4.05309439e-01 3.94459516e-01 5.67890286e-01
-2.80145407e-01 8.22295070e-01 5.77120483e-01 -2.87372559e-01
-9.07533050e-01 3.12419057e-01 2.05179572e-01 -8.07812870e-01
-6.99370921e-01 1.11546016e+00 -4.09751646e-02 -7.21754789e-01
2.67592460e-01 -2.78365344e-01 -5.71436524e-01 9.61565077e-02
8.41112316e-01 3.03805232e-01 4.39102083e-01 -5.54446399e-01
-1.80501223e-01 4.29031849e-01 -8.20680633e-02 -2.08518893e-01
1.16253459e+00 -5.05670071e-01 -3.26593488e-01 4.01771218e-01
1.40334320e+00 1.67972706e-02 -7.90419102e-01 -6.83648407e-01
3.91884923e-01 -1.85510352e-01 -1.46374911e-01 -1.08393538e+00
-2.23246962e-01 1.03157473e+00 1.78250417e-01 2.54648805e-01
7.67673612e-01 6.42863989e-01 1.70410717e+00 8.37048829e-01
3.32137138e-01 -1.26476884e+00 2.16624573e-01 1.09160054e+00
1.01709032e+00 -1.22699952e+00 -6.70346916e-01 -1.29555985e-01
-4.55511719e-01 1.03257215e+00 5.88282645e-01 6.87221810e-02
3.29865485e-01 -4.26267326e-01 3.76468003e-01 -6.03866577e-01
-7.14296579e-01 -3.27609628e-01 6.12503171e-01 3.70350808e-01
1.91976905e-01 -1.15012877e-01 -1.60820082e-01 9.28401709e-01
-3.69298816e-01 -3.92817885e-01 1.86157376e-01 9.45600092e-01
-8.61773968e-01 -8.16245854e-01 -3.48247379e-01 4.21494514e-01
-6.67329729e-01 -4.49137658e-01 -3.18553746e-01 3.81910652e-01
-1.94748864e-01 1.63839281e+00 1.69670880e-02 2.03423470e-01
5.43186545e-01 4.85681593e-01 2.07513094e-01 -7.62815893e-01
-7.75350988e-01 -6.91109717e-01 4.89572644e-01 -4.17974055e-01
-2.29139179e-01 -3.48653734e-01 -1.36974204e+00 1.14574030e-01
-3.70499998e-01 7.52715826e-01 3.47746044e-01 1.33854151e+00
4.46502030e-01 2.57293403e-01 3.35992724e-01 1.94263831e-01
-8.10392737e-01 -9.85936046e-01 1.49119664e-02 5.60520828e-01
6.26727998e-01 -2.30324060e-01 -4.62782711e-01 -1.58659860e-01] | [11.318596839904785, 7.96478271484375] |
f1eea86b-4528-40d1-8d66-676e16a026c0 | supervised-contrastive-learning-approach-for | 2207.03153 | null | https://arxiv.org/abs/2207.03153v1 | https://arxiv.org/pdf/2207.03153v1.pdf | Supervised Contrastive Learning Approach for Contextual Ranking | Contextual ranking models have delivered impressive performance improvements over classical models in the document ranking task. However, these highly over-parameterized models tend to be data-hungry and require large amounts of data even for fine tuning. This paper proposes a simple yet effective method to improve ranking performance on smaller datasets using supervised contrastive learning for the document ranking problem. We perform data augmentation by creating training data using parts of the relevant documents in the query-document pairs. We then use a supervised contrastive learning objective to learn an effective ranking model from the augmented dataset. Our experiments on subsets of the TREC-DL dataset show that, although data augmentation leads to an increasing the training data sizes, it does not necessarily improve the performance using existing pointwise or pairwise training objectives. However, our proposed supervised contrastive loss objective leads to performance improvements over the standard non-augmented setting showcasing the utility of data augmentation using contrastive losses. Finally, we show the real benefit of using supervised contrastive learning objectives by showing marked improvements in smaller ranking datasets relating to news (Robust04), finance (FiQA), and scientific fact checking (SciFact). | ['Avishek Anand', 'Koustav Rudra', 'Jurek Leonhardt', 'Abhijit Anand'] | 2022-07-07 | null | null | null | null | ['document-ranking'] | ['natural-language-processing'] | [ 2.97291428e-01 6.25715032e-02 -2.88879275e-01 -7.25946248e-01
-1.81013870e+00 -5.24424970e-01 1.19560266e+00 5.30231655e-01
-7.37669706e-01 9.57570314e-01 5.96356988e-01 -7.56174996e-02
-6.92823589e-01 -5.38255036e-01 -8.66371870e-01 -6.62104726e-01
-3.85071844e-01 1.04350579e+00 1.02159023e-01 -4.61685210e-01
5.58450520e-01 1.81404904e-01 -1.60389256e+00 6.76153779e-01
7.23366320e-01 9.25417781e-01 -2.13845164e-01 7.86509812e-01
-2.28992589e-02 8.30324709e-01 -6.37016177e-01 -6.29412413e-01
4.40905660e-01 -4.85002734e-02 -1.12323713e+00 -4.21608746e-01
9.05993164e-01 -2.20884785e-01 1.29052818e-01 7.33625770e-01
6.41316831e-01 2.19939336e-01 6.59732282e-01 -9.88841653e-01
-3.44599366e-01 8.01770091e-01 -4.40084070e-01 7.79748484e-02
3.56889516e-01 -6.81295753e-01 1.73794520e+00 -8.54395270e-01
6.02308750e-01 1.28627098e+00 5.31177461e-01 2.87005991e-01
-1.21128380e+00 -5.15666246e-01 1.10087367e-02 -6.93601891e-02
-8.60103846e-01 -2.85709769e-01 7.81826496e-01 -1.37136251e-01
5.70968866e-01 6.88085675e-01 8.18654895e-02 9.17209506e-01
-3.72632444e-01 7.53527403e-01 1.45262539e+00 -7.51961112e-01
-1.34671614e-01 3.01242650e-01 5.01352787e-01 4.46707219e-01
2.01471001e-01 3.14949304e-01 -4.92248237e-01 -5.47138631e-01
9.94418934e-02 -5.65717556e-02 -5.10232747e-02 -3.87329370e-01
-1.16471350e+00 7.56941199e-01 4.70373541e-01 3.33358049e-01
-1.01389833e-01 2.50593722e-01 7.30100274e-01 7.54773259e-01
9.70550418e-01 9.34626162e-01 -8.53385568e-01 8.54394399e-03
-1.07382810e+00 6.39154077e-01 6.39988303e-01 6.06136739e-01
5.17197013e-01 -5.93933821e-01 -7.48032749e-01 1.17716873e+00
-2.61157993e-02 2.64744520e-01 2.66032189e-01 -1.05946767e+00
7.23169565e-01 5.09323776e-01 3.18668425e-01 -8.15437138e-01
-3.72877836e-01 -7.91061521e-01 -5.51855683e-01 1.08891487e-01
4.20731723e-01 3.01718324e-01 -7.45479584e-01 1.77316570e+00
-7.03476444e-02 -4.28878546e-01 -3.03524397e-02 6.88928783e-01
6.93965673e-01 5.26715457e-01 -1.53718470e-02 -1.58821300e-01
1.07551074e+00 -9.26769912e-01 -4.89381135e-01 2.16034591e-01
1.07635796e+00 -7.61920929e-01 1.52432132e+00 5.54221451e-01
-1.21047068e+00 -4.05175120e-01 -9.80525553e-01 -3.84963006e-01
-5.34667313e-01 1.85848340e-01 7.69608855e-01 5.36135972e-01
-1.01060009e+00 6.24504626e-01 -1.12984292e-01 -6.96280785e-03
3.76950026e-01 6.15401685e-01 -3.34354639e-01 -3.27969044e-01
-1.39814711e+00 8.18508565e-01 3.84537935e-01 -3.56973559e-01
-6.26086414e-01 -8.60589504e-01 -4.93264467e-01 2.24648446e-01
4.28628951e-01 -5.10159612e-01 1.09143698e+00 -7.94723928e-01
-1.12522960e+00 9.79914308e-01 1.56692788e-01 -5.89275837e-01
8.47054064e-01 -6.14330828e-01 -2.46777549e-01 -1.13309748e-01
1.00909285e-01 5.98903894e-01 3.58753562e-01 -1.40376735e+00
-6.93370283e-01 -2.94313431e-01 3.21736693e-01 3.09705436e-01
-4.88997966e-01 1.97650090e-01 -3.77499342e-01 -7.01815248e-01
-2.81253874e-01 -8.21164012e-01 -1.88472018e-01 -4.89176124e-01
-2.88677663e-01 -3.89167130e-01 5.55470526e-01 -6.33789420e-01
1.21946061e+00 -1.96721339e+00 -8.52070078e-02 4.92518872e-01
-3.93606462e-02 1.96549311e-01 -4.80257601e-01 3.83819342e-01
-1.24553397e-01 3.11386555e-01 -1.40952632e-01 -3.20554823e-01
-7.09903659e-03 5.55640757e-02 -4.68032598e-01 1.66749209e-01
3.52203958e-02 8.84942830e-01 -9.28668916e-01 -5.31739354e-01
-2.63304293e-01 2.94188619e-01 -7.96578825e-01 -1.87269151e-01
-2.88124472e-01 1.18300267e-01 -2.86015451e-01 3.91641825e-01
3.77127469e-01 -2.77012199e-01 1.83080986e-01 -8.26614257e-03
2.14100510e-01 7.91010261e-01 -9.19540405e-01 1.46888661e+00
-6.20640039e-01 5.80514193e-01 -3.26896161e-01 -1.10124552e+00
7.33787537e-01 1.21562183e-01 5.30983210e-01 -9.50303555e-01
-2.16840968e-01 4.98933345e-01 -1.52328819e-01 -6.91003650e-02
8.48928392e-01 -1.03482738e-01 -2.41950557e-01 5.12750626e-01
-1.66753173e-01 -1.45100459e-01 4.44109887e-01 5.47959566e-01
1.07351625e+00 -1.29579023e-01 -1.33818179e-01 -4.61533189e-01
5.34478545e-01 5.67724966e-02 1.12262197e-01 1.06018603e+00
5.08254886e-01 6.52406037e-01 7.42069781e-01 -3.28219116e-01
-1.11712515e+00 -7.08286405e-01 -3.91162992e-01 1.55807972e+00
-2.62383491e-01 -6.65042281e-01 -1.96210831e-01 -1.15414214e+00
6.84445798e-02 5.76996744e-01 -6.93646491e-01 -7.67490864e-02
-5.74650288e-01 -1.15890706e+00 3.30347776e-01 4.66842711e-01
3.66058677e-01 -7.98749566e-01 1.14144823e-02 2.46039145e-02
-2.37875417e-01 -7.26206303e-01 -4.15086359e-01 4.42867547e-01
-1.11862612e+00 -9.72134411e-01 -8.21024477e-01 -6.24226630e-01
8.22403312e-01 3.62667702e-02 1.55289888e+00 2.44080737e-01
-7.48031065e-02 3.23048234e-01 -4.01142716e-01 -3.61147791e-01
-2.70627260e-01 2.51266420e-01 -2.70190164e-02 -4.90177423e-01
4.88832705e-02 -1.00608647e-01 -3.95200998e-01 2.50982910e-01
-9.87845182e-01 -3.83280635e-01 6.24669909e-01 1.30799627e+00
5.32212853e-01 1.38010010e-01 7.61048377e-01 -1.62739968e+00
8.69904995e-01 -1.37547284e-01 -6.58426225e-01 4.24805552e-01
-1.12665343e+00 5.23395121e-01 4.39186007e-01 -9.64174271e-02
-1.00220549e+00 -3.50421309e-01 1.18820816e-01 3.12027305e-01
5.00056148e-01 8.26972842e-01 1.95364654e-01 1.50255054e-01
8.87704253e-01 -3.46718997e-01 -1.71308249e-01 -6.34775043e-01
3.74173671e-01 5.34841061e-01 2.81080037e-01 -8.10521483e-01
9.07128692e-01 3.52411151e-01 2.35459015e-01 -6.13956377e-02
-1.27540898e+00 -6.15263104e-01 -4.36778367e-01 2.11053640e-01
1.60084531e-01 -8.53243709e-01 -4.19953912e-01 -8.75138640e-02
-7.63059020e-01 -1.36446536e-01 -5.10118246e-01 4.28203672e-01
-3.79247248e-01 2.61680394e-01 -5.10653794e-01 -6.23158693e-01
-3.22978646e-01 -8.03516865e-01 1.19505715e+00 -4.73423004e-01
-5.58619329e-04 -1.06350648e+00 4.41866130e-01 6.36352301e-01
4.13622320e-01 1.33211657e-01 1.25049412e+00 -1.04746079e+00
-3.90571475e-01 -8.04000616e-01 -2.48838753e-01 6.79771602e-01
-8.21466446e-02 -8.49250928e-02 -9.68207240e-01 -3.36013854e-01
-6.22081220e-01 -7.00805128e-01 1.37399077e+00 2.46875763e-01
1.48887527e+00 -5.16996324e-01 -9.88886580e-02 1.58247128e-01
1.31502318e+00 -1.62729010e-01 5.82914174e-01 5.33556521e-01
5.22068083e-01 9.97498453e-01 1.01536453e+00 1.76949769e-01
6.15062639e-02 9.18811083e-01 1.83117270e-01 -2.58327752e-01
4.43382002e-02 -8.13771933e-02 2.65185326e-01 5.44390559e-01
-3.18860233e-01 -1.86506927e-01 -6.99166119e-01 3.94821823e-01
-1.76902318e+00 -1.01654446e+00 -1.94554925e-01 2.70336843e+00
1.14795661e+00 3.57698977e-01 2.17955396e-01 2.97823876e-01
4.78555948e-01 9.08650830e-02 2.09941063e-02 -3.94056976e-01
-4.61586773e-01 3.11683923e-01 5.03488123e-01 6.75455987e-01
-1.20789886e+00 6.76917374e-01 6.54236507e+00 9.86707687e-01
-6.98815286e-01 6.31761402e-02 9.08815324e-01 -5.81811786e-01
-5.53597212e-01 1.70990787e-02 -6.76905513e-01 3.75446081e-01
9.96358335e-01 -9.07327905e-02 1.73804656e-01 7.39400685e-01
-1.51208028e-01 8.28422140e-03 -1.33417284e+00 6.91039979e-01
-3.42552401e-02 -1.34403515e+00 2.87171721e-01 3.36118162e-01
1.05294597e+00 -5.98063432e-02 3.43785793e-01 6.72078907e-01
4.84941036e-01 -1.02446699e+00 3.55172217e-01 4.56382036e-01
7.75810659e-01 -9.41200137e-01 1.20748401e+00 2.07934808e-02
-3.66797358e-01 -1.53855056e-01 -4.31686491e-01 6.28918558e-02
-3.51727456e-01 9.77011621e-01 -8.00940394e-01 6.26127183e-01
6.80599451e-01 4.86651629e-01 -9.34137940e-01 1.03648973e+00
4.83108126e-02 5.68045259e-01 -1.16568305e-01 4.37013768e-02
2.57861495e-01 8.35970044e-02 3.95791054e-01 1.17321420e+00
3.93490642e-02 -2.54767299e-01 -8.56817067e-02 6.84747398e-02
-7.10364461e-01 3.30077738e-01 -4.88518894e-01 9.97623429e-02
6.75686896e-02 1.15747070e+00 -1.80240825e-01 -5.12241006e-01
-1.96410239e-01 6.36924624e-01 5.03813982e-01 1.29907429e-01
-5.50216198e-01 -3.75537425e-01 -1.36748876e-03 2.63415009e-01
-6.93214834e-02 2.43386731e-01 -2.01454937e-01 -8.42279673e-01
2.14120537e-01 -9.65906620e-01 8.12352002e-01 -4.00421023e-01
-1.43563366e+00 5.04871249e-01 2.48916596e-01 -1.27627027e+00
-4.65203702e-01 -5.98323584e-01 -2.80365199e-01 6.85680389e-01
-1.72472489e+00 -9.46621597e-01 7.72751123e-02 3.07038993e-01
1.61401391e-01 -3.17375571e-01 6.68135166e-01 5.04196763e-01
8.42037350e-02 8.09904099e-01 6.02476180e-01 -1.10950083e-01
1.24222898e+00 -1.69937062e+00 -1.17852472e-01 3.14398199e-01
2.88387090e-01 7.26951361e-01 5.88016510e-01 -3.13733190e-01
-7.47796655e-01 -9.35787439e-01 1.25816107e+00 -9.66747165e-01
3.38722438e-01 -3.78635466e-01 -6.90973878e-01 4.80382413e-01
-4.23693098e-02 8.86663720e-02 6.33631587e-01 8.01119566e-01
-6.69879258e-01 -4.49921727e-01 -1.20899487e+00 4.09538478e-01
8.07478964e-01 -5.99681616e-01 -6.38474941e-01 7.63984203e-01
6.14339769e-01 -2.22275332e-02 -7.01024532e-01 6.70406580e-01
7.14556575e-01 -5.98304927e-01 1.19764030e+00 -1.09744096e+00
9.48436797e-01 6.20563999e-02 -3.48938197e-01 -1.24339426e+00
-2.21055374e-01 -3.56532037e-01 -5.78321377e-03 1.47588873e+00
7.86431015e-01 -3.85597020e-01 8.11313450e-01 5.96355021e-01
-2.12014057e-02 -6.92003012e-01 -6.40289545e-01 -9.30061877e-01
4.17579591e-01 -1.81379512e-01 3.58797580e-01 9.95762050e-01
-1.47942394e-01 4.34837878e-01 -4.48420227e-01 -2.94285685e-01
6.20366573e-01 -2.52201483e-02 6.94118381e-01 -1.57753003e+00
-3.75220120e-01 -4.57997859e-01 -9.86180827e-02 -6.98800802e-01
1.24042146e-01 -1.15048897e+00 -8.93424973e-02 -1.57859087e+00
6.86926246e-01 -8.23853731e-01 -8.89279842e-01 4.20969218e-01
-4.00881201e-01 3.51951808e-01 -4.57783639e-02 2.40669772e-01
-1.01316881e+00 3.64021569e-01 1.03861427e+00 -3.20145428e-01
-1.19021699e-01 2.49157518e-01 -9.39248562e-01 3.07277560e-01
5.63758135e-01 -6.83699727e-01 -4.19996649e-01 -3.87970626e-01
8.30808163e-01 -1.91957533e-01 3.68055016e-01 -5.45127988e-01
4.29740250e-02 2.60833055e-01 3.78417104e-01 -6.66143894e-01
1.74905449e-01 -7.78694272e-01 -4.03430820e-01 3.34293455e-01
-1.19530463e+00 7.24292696e-02 1.54881021e-02 6.21008039e-01
-5.08673787e-01 -2.64668971e-01 5.53734601e-01 -5.42323925e-02
-2.33907536e-01 1.86918527e-02 2.99196899e-01 2.92694777e-01
3.50078970e-01 1.99801087e-01 -5.11574268e-01 -4.29132253e-01
-4.80524808e-01 2.98796654e-01 1.67601004e-01 3.05519640e-01
2.03656688e-01 -1.25496936e+00 -1.12418127e+00 -2.80080557e-01
4.48223054e-01 -1.56404659e-01 -5.45139052e-02 6.93046391e-01
-2.54228085e-01 8.81407976e-01 9.08650681e-02 -2.96584308e-01
-1.59080839e+00 3.78150821e-01 -2.53954008e-02 -1.19522071e+00
-8.41512084e-02 8.52192104e-01 1.18202999e-01 -8.24058414e-01
4.70207810e-01 -1.14543021e-01 -2.28983536e-01 2.89126605e-01
3.55202317e-01 5.79425573e-01 5.17307818e-01 -5.46731204e-02
-3.27276826e-01 3.44146013e-01 -7.54454017e-01 -4.22867656e-01
1.50716162e+00 1.98937669e-01 -1.73815504e-01 1.73330113e-01
1.62202990e+00 4.63935524e-01 -8.46871495e-01 -5.15158534e-01
5.00558138e-01 -5.09715617e-01 3.14480245e-01 -1.41633904e+00
-7.85183728e-01 5.96354783e-01 5.02616107e-01 1.77723855e-01
1.05716658e+00 1.18772440e-01 2.84499735e-01 8.23195815e-01
3.50311220e-01 -1.18792498e+00 2.90104836e-01 4.64106649e-01
1.12028992e+00 -1.55312061e+00 3.88963103e-01 -5.40718548e-02
-5.98566413e-01 7.83005834e-01 2.20711440e-01 -3.19439024e-02
4.58489507e-01 -1.05172068e-01 -9.15866494e-02 -4.45571691e-01
-9.22646403e-01 -5.30470535e-02 7.86745489e-01 1.68681160e-01
7.77224183e-01 -1.28000483e-01 -7.47111022e-01 3.74491066e-01
-3.61241847e-01 -3.26941043e-01 2.16679871e-01 7.56368816e-01
-1.47222534e-01 -1.54381895e+00 -1.67539924e-01 9.23500061e-01
-9.77245867e-01 -4.39482182e-01 -6.46246970e-01 9.72543538e-01
-3.05010259e-01 7.63384223e-01 -5.93272932e-02 -1.61891922e-01
2.90298909e-01 8.58273506e-02 4.64971125e-01 -6.60855472e-01
-8.32752228e-01 -5.94469048e-02 5.09585023e-01 -2.58149266e-01
-5.83452940e-01 -6.65342510e-01 -8.33206713e-01 -8.01257417e-02
-4.72001672e-01 5.59737444e-01 7.25752234e-01 7.11068749e-01
2.28122011e-01 5.33040285e-01 9.07513857e-01 -3.43393356e-01
-9.09242690e-01 -9.71396089e-01 -5.09729147e-01 6.84396684e-01
4.98605400e-01 -5.90156436e-01 -5.88294446e-01 -3.12065542e-01] | [11.411486625671387, 7.571452617645264] |
b4733a01-3398-419a-89df-d477caf51bfe | vocal-tract-length-perturbation-for-text | 2011.12536 | null | https://arxiv.org/abs/2011.12536v2 | https://arxiv.org/pdf/2011.12536v2.pdf | Vocal Tract Length Perturbation for Text-Dependent Speaker Verification with Autoregressive Prediction Coding | In this letter, we propose a vocal tract length (VTL) perturbation method for text-dependent speaker verification (TD-SV), in which a set of TD-SV systems are trained, one for each VTL factor, and score-level fusion is applied to make a final decision. Next, we explore the bottleneck (BN) feature extracted by training deep neural networks with a self-supervised objective, autoregressive predictive coding (APC), for TD-SV and compare it with the well-studied speaker-discriminant BN feature. The proposed VTL method is then applied to APC and speaker-discriminant BN features. In the end, we combine the VTL perturbation systems trained on MFCC and the two BN features in the score domain. Experiments are performed on the RedDots challenge 2016 database of TD-SV using short utterances with Gaussian mixture model-universal background model and i-vector techniques. Results show the proposed methods significantly outperform the baselines. | ['Zheng-Hua Tan', 'Achintya kr. Sarkar'] | 2020-11-25 | null | null | null | null | ['text-dependent-speaker-verification'] | ['speech'] | [ 2.87199497e-01 -2.61070222e-01 2.05915779e-01 -5.17866254e-01
-1.42880440e+00 -4.54123050e-01 4.54819471e-01 -2.01499864e-01
-3.19908708e-01 8.68946612e-02 5.65306425e-01 -4.73178655e-01
2.96719730e-01 8.54784921e-02 -2.97731191e-01 -1.01590765e+00
9.57969725e-02 1.42168537e-01 -1.92342371e-01 -9.04597640e-02
-1.06010415e-01 3.38935316e-01 -1.42112827e+00 2.82580972e-01
6.40642583e-01 1.30036259e+00 6.18946999e-02 1.02930129e+00
1.75023347e-01 5.50630748e-01 -9.37979162e-01 -4.77084011e-01
1.88867256e-01 -4.77281064e-01 -4.22717959e-01 -2.14249223e-01
4.29662496e-01 -2.05951840e-01 -3.23679894e-01 1.08494401e+00
1.17004693e+00 3.78212392e-01 8.89735162e-01 -1.14104962e+00
-4.01244104e-01 8.87974620e-01 -4.50278610e-01 3.61803561e-01
1.97039619e-01 2.53776401e-01 1.00767195e+00 -1.15282500e+00
5.66801354e-02 1.75151277e+00 6.41014338e-01 8.43887925e-01
-1.28630173e+00 -8.94104958e-01 1.32420540e-01 4.69748437e-01
-1.31351972e+00 -1.08495677e+00 1.10699344e+00 -4.41413879e-01
1.00824451e+00 3.95859957e-01 8.02795291e-02 1.43731773e+00
-1.40860498e-01 1.11228096e+00 8.78706455e-01 -5.96123159e-01
3.09105128e-01 9.60311890e-02 2.41420478e-01 3.59919727e-01
-4.79877591e-01 4.37476873e-01 -9.22598541e-01 -2.86058158e-01
-2.38233253e-01 -5.89661658e-01 -3.74318480e-01 8.05017576e-02
-7.69145548e-01 7.80014336e-01 -1.59097522e-01 3.35049033e-01
-2.40649521e-01 -1.62986994e-01 7.66648412e-01 4.83294070e-01
6.32922292e-01 -2.56319255e-01 -6.10196471e-01 -2.41624564e-01
-1.16865063e+00 2.47180983e-01 7.22255290e-01 5.55908203e-01
1.74673349e-01 7.04210401e-01 -7.51212299e-01 1.26252556e+00
8.01935136e-01 7.58424044e-01 9.58438337e-01 -7.15925276e-01
4.87338245e-01 -2.39497587e-01 -2.92935997e-01 -5.48309743e-01
-1.72391385e-01 -5.30429900e-01 -8.83645356e-01 2.20329136e-01
2.06831679e-01 -1.85333639e-01 -8.30904484e-01 1.99677944e+00
4.03575599e-01 1.72876045e-01 2.60588437e-01 6.16358221e-01
1.00281918e+00 6.95745826e-01 -2.61728372e-03 -3.40320736e-01
1.11760604e+00 -1.04552877e+00 -9.42067385e-01 7.13732466e-02
5.39909422e-01 -8.78706098e-01 9.60161328e-01 4.84645128e-01
-9.40162182e-01 -8.25202465e-01 -1.02194011e+00 1.48806244e-01
-1.68477774e-01 3.35709751e-01 -4.36659902e-01 1.13430560e+00
-1.12391412e+00 4.96568412e-01 -6.39614522e-01 1.29841343e-02
1.18392497e-01 3.80136579e-01 -8.79566185e-03 1.76784247e-01
-1.31356931e+00 8.12233567e-01 4.50240597e-02 2.19067380e-01
-1.14227951e+00 -6.16200089e-01 -1.00925124e+00 1.01560012e-01
-2.64718980e-01 -1.51844591e-01 1.70752418e+00 -5.98964691e-01
-2.28455305e+00 7.17803478e-01 -4.69941586e-01 -6.02276564e-01
3.83921653e-01 -1.37163922e-01 -7.61832416e-01 -1.43737420e-01
-3.52599233e-01 2.50362784e-01 1.47827923e+00 -9.77776170e-01
-3.59605581e-01 -5.49825191e-01 -7.01373339e-01 2.34042913e-01
-4.60383832e-01 4.12096918e-01 6.51294589e-02 -7.27732360e-01
1.57073140e-02 -6.88039839e-01 1.86745539e-01 -5.30810952e-01
-6.87966943e-01 -5.77462673e-01 1.13233733e+00 -1.40324616e+00
1.37544560e+00 -2.61930752e+00 4.02981520e-01 1.74958989e-01
-1.10140190e-01 6.50053442e-01 -2.63784021e-01 9.38505828e-02
-8.74586329e-02 -1.78898513e-01 -2.85871267e-01 -1.11516035e+00
3.50968182e-01 -6.64397255e-02 -2.79768288e-01 6.03396833e-01
1.49756029e-01 5.52279234e-01 -3.81161749e-01 -5.62064826e-01
2.11180389e-01 5.99206269e-01 -5.56087017e-01 2.68610477e-01
6.96604326e-02 4.42417622e-01 1.77087903e-01 5.54806173e-01
1.02399325e+00 7.78669953e-01 -5.34463190e-02 -1.42438889e-01
8.83345958e-03 6.66500926e-01 -8.72500658e-01 1.45454037e+00
-3.58093113e-01 7.74693251e-01 4.92418945e-01 -7.48333693e-01
1.02899337e+00 5.75559735e-01 1.33995518e-01 -2.28144214e-01
3.13088477e-01 1.27718017e-01 3.13385934e-01 -3.19440752e-01
1.66445091e-01 -3.06788623e-01 6.51711449e-02 2.24054456e-01
5.32976866e-01 -2.17485860e-01 -3.11024070e-01 -2.47382242e-02
1.01746750e+00 -1.49046645e-01 7.61448070e-02 -1.04023293e-02
9.77116168e-01 -7.95372009e-01 6.20698035e-01 4.66392100e-01
-9.08475578e-01 4.99495149e-01 2.73586184e-01 3.14335912e-01
-7.21725762e-01 -1.10495770e+00 -2.05427170e-01 1.39264512e+00
-5.77799916e-01 -2.32582524e-01 -9.58790064e-01 -7.49581337e-01
1.22020647e-01 1.10270452e+00 -3.43408674e-01 -3.70504469e-01
-5.46103835e-01 -4.68073249e-01 1.02803743e+00 3.87658477e-01
1.54282495e-01 -8.96099150e-01 2.58311570e-01 2.68705964e-01
-2.56445050e-01 -9.18299019e-01 -8.62829149e-01 3.37869883e-01
-4.86009002e-01 -2.75041729e-01 -8.58889997e-01 -7.39067852e-01
-1.52026508e-02 -6.19088449e-02 4.30158705e-01 -6.26946330e-01
2.08488330e-01 1.88344911e-01 -2.34484032e-01 -4.77614254e-01
-1.02788162e+00 -1.42931342e-01 5.20216465e-01 3.72375011e-01
6.21024311e-01 -5.09262145e-01 -2.95207739e-01 2.23163396e-01
-4.62496102e-01 -4.42887217e-01 4.83681500e-01 1.15472770e+00
1.67406291e-01 -2.60883242e-01 7.69161463e-01 -2.96117831e-03
7.20410943e-01 -9.66793746e-02 -4.94273424e-01 -2.09443159e-02
-3.19860488e-01 8.26554596e-02 3.98903102e-01 -7.79885650e-01
-1.13551402e+00 -1.38982683e-02 -7.54545152e-01 -7.90348589e-01
-1.90139309e-01 4.10825610e-01 -6.10054672e-01 4.96218354e-02
6.16541743e-01 4.56825763e-01 1.88935027e-01 -7.59101033e-01
3.31076324e-01 1.38176179e+00 6.26949370e-01 -2.32534558e-01
7.74717391e-01 -9.92528647e-02 -4.71068799e-01 -9.61762786e-01
-6.01856232e-01 -7.72787929e-01 -6.39168739e-01 -1.74225613e-01
6.39508247e-01 -7.60409594e-01 -8.65242302e-01 7.99294472e-01
-1.35537839e+00 -1.59807831e-01 -2.35042930e-01 6.28271878e-01
-5.29774785e-01 5.90354919e-01 -7.22577333e-01 -1.38309288e+00
-6.20110750e-01 -1.28929710e+00 1.18165076e+00 -1.44459233e-01
-2.08184615e-01 -6.48404598e-01 3.19110453e-01 5.53983450e-01
5.87010384e-01 -2.94476300e-01 7.41972268e-01 -1.28479183e+00
2.66493201e-01 -2.73708701e-01 1.14904255e-01 1.21940351e+00
1.13822632e-02 -9.06018466e-02 -1.73672879e+00 -5.02864361e-01
4.92253065e-01 -1.24169275e-01 1.02635419e+00 6.47053063e-01
1.10140014e+00 -4.22042221e-01 -8.19235444e-02 5.51611423e-01
6.45425916e-01 2.57450521e-01 -1.55912954e-02 -1.95883751e-01
5.29932678e-01 4.36756432e-01 2.45182440e-01 3.85532171e-01
2.57886052e-01 9.31646705e-01 8.86611566e-02 1.80789173e-01
-5.19256115e-01 -7.91096166e-02 1.06546926e+00 1.62519884e+00
1.88483164e-01 -2.65117556e-01 -7.88624644e-01 4.41571325e-01
-1.38058412e+00 -9.68498647e-01 1.12909533e-01 2.24052310e+00
7.94676006e-01 1.17287152e-02 3.61942887e-01 6.23314977e-01
9.48574483e-01 3.27947825e-01 -5.90762377e-01 -6.20140851e-01
-2.71582574e-01 1.31619126e-01 3.93853560e-02 6.58466578e-01
-1.11574423e+00 7.31067598e-01 5.71744823e+00 1.08519411e+00
-1.29302382e+00 5.88309765e-01 4.91284728e-01 -1.09525524e-01
9.34404582e-02 -5.05151153e-01 -1.04550731e+00 3.82523358e-01
1.52369428e+00 -2.29424775e-01 4.02523816e-01 8.57400715e-01
2.41840035e-01 4.41705734e-01 -1.09768236e+00 1.00922084e+00
4.36278909e-01 -4.58234936e-01 -2.61101127e-01 -1.19913453e-02
4.23321664e-01 2.05709219e-01 6.11240208e-01 7.18498647e-01
2.22299963e-01 -7.85923839e-01 8.88065755e-01 8.13174322e-02
8.89794230e-01 -6.99445248e-01 6.05486095e-01 2.14482412e-01
-1.06189930e+00 -3.22049350e-01 -2.37353578e-01 5.73654175e-01
5.07249832e-02 2.74474740e-01 -1.30277050e+00 3.97234827e-01
5.58996379e-01 4.26595330e-01 -1.62409008e-01 8.75599742e-01
5.25793731e-02 1.24800205e+00 -2.75145173e-01 -5.50594181e-02
-9.57702696e-02 2.60223359e-01 1.17350733e+00 1.51464272e+00
3.47018957e-01 -4.28304583e-01 -3.42533886e-01 7.50091851e-01
-8.23934451e-02 1.94657058e-01 -3.51111740e-01 5.48290238e-02
3.99150312e-01 9.88990605e-01 1.04803905e-01 -2.92149574e-01
-8.35171640e-02 9.92618620e-01 1.52410686e-01 4.84697223e-01
-7.52963006e-01 -2.84468710e-01 8.81859303e-01 -4.28010732e-01
6.11149132e-01 3.44394781e-02 5.13187870e-02 -1.11812913e+00
-1.63512602e-01 -1.15822232e+00 2.20245361e-01 -4.23553884e-01
-1.35125208e+00 8.46783698e-01 -2.53766596e-01 -1.26035547e+00
-5.64188659e-01 -5.74281216e-01 -6.83338881e-01 1.26707816e+00
-1.61069787e+00 -1.07400572e+00 3.03607315e-01 6.61666036e-01
8.55565429e-01 -5.69532692e-01 1.02375853e+00 3.60702217e-01
-8.33197474e-01 1.27338088e+00 5.73387086e-01 2.88407892e-01
8.69959176e-01 -1.15038085e+00 5.56098163e-01 8.50302815e-01
7.50479177e-02 2.56256342e-01 7.16899335e-01 -3.03668946e-01
-1.03109550e+00 -1.11831343e+00 9.31970894e-01 -2.03595996e-01
3.52540702e-01 -6.42588079e-01 -8.55461597e-01 4.13813680e-01
1.58013076e-01 -5.46669178e-02 8.79601598e-01 6.96435571e-02
-5.75081944e-01 -2.38235325e-01 -1.28762531e+00 2.50845850e-01
5.89454532e-01 -8.21778536e-01 -7.60724366e-01 1.16372019e-01
9.83495772e-01 -3.50274056e-01 -6.05648637e-01 3.57689381e-01
5.47856450e-01 -7.29476452e-01 9.42542017e-01 -5.51975191e-01
-2.10062921e-01 -1.30299672e-01 -6.73698843e-01 -1.65460324e+00
-2.44931147e-01 -9.27375793e-01 -3.99448276e-01 1.61359143e+00
5.03580272e-01 -4.08778757e-01 3.79093051e-01 9.45127085e-02
-4.16566968e-01 -3.35329860e-01 -1.67547798e+00 -9.20227885e-01
1.55975014e-01 -9.87120032e-01 7.15224206e-01 5.92213988e-01
6.31367117e-02 3.70069534e-01 -3.78875464e-01 3.31143588e-01
7.29448378e-01 -4.27421480e-01 6.17598653e-01 -1.10936189e+00
-4.72700894e-01 -7.02935815e-01 -5.48469663e-01 -1.02407157e+00
4.49919045e-01 -9.48761165e-01 3.82520318e-01 -7.69192815e-01
-9.10988823e-02 3.44221145e-02 -6.68298662e-01 2.16859490e-01
-3.02901685e-01 -1.02702305e-01 3.61470819e-01 -1.02223292e-01
-1.80201516e-01 9.48078752e-01 7.53182471e-01 -4.20915842e-01
-3.10089141e-01 4.98093635e-01 -2.60489911e-01 6.11763477e-01
6.88519955e-01 -4.11555707e-01 9.17853713e-02 9.71595421e-02
-8.26811016e-01 3.17138821e-01 5.42066060e-02 -1.04718685e+00
2.40062982e-01 4.82070386e-01 2.17821881e-01 -8.99262071e-01
8.76542866e-01 -3.64618659e-01 -5.76937020e-01 4.32524920e-01
-5.82120776e-01 -5.11727571e-01 3.09709638e-01 5.27060866e-01
-4.12064970e-01 -6.10848069e-02 1.07436204e+00 6.33264422e-01
-3.79039422e-02 6.11368679e-02 -6.82161748e-01 -1.39246002e-01
5.35658181e-01 1.01658471e-01 3.29777300e-02 -5.70556462e-01
-8.66677582e-01 -2.85512596e-01 -3.74277622e-01 4.72779632e-01
6.86681032e-01 -1.42534113e+00 -1.16313183e+00 5.99991858e-01
5.93744777e-02 -4.07011002e-01 5.49041212e-01 8.90460730e-01
2.57962078e-01 4.82102931e-01 2.09152550e-01 -9.29387629e-01
-1.73771477e+00 4.74166632e-01 4.62010682e-01 -2.28549689e-01
-1.33474052e-01 1.33001947e+00 1.72813967e-01 -8.06681871e-01
7.81755328e-01 -4.94777679e-01 -1.82053968e-01 -7.84052387e-02
5.83051741e-01 4.11654711e-01 3.65928799e-01 -1.24985778e+00
-4.03028518e-01 1.88300386e-01 -1.18491523e-01 -6.48655772e-01
1.08294761e+00 -3.73778403e-01 2.80253172e-01 5.90292156e-01
1.38814735e+00 2.32400507e-01 -9.67458606e-01 -6.24128520e-01
-8.25789571e-02 -1.45775095e-01 4.09130961e-01 -7.95543432e-01
-8.27326953e-01 1.27393436e+00 9.80645835e-01 8.61780047e-02
1.12448800e+00 -1.74603477e-01 8.68289232e-01 3.99680324e-02
-3.53545129e-01 -9.58468556e-01 -1.51211098e-01 6.38513088e-01
1.16310453e+00 -1.19969296e+00 -6.00433469e-01 6.64096652e-03
-7.12377131e-01 9.55301583e-01 4.69639063e-01 4.78560537e-01
9.27367866e-01 2.60916680e-01 3.63242328e-01 4.59666848e-01
-7.78497517e-01 -5.51028214e-02 5.81944346e-01 8.16347241e-01
3.94181341e-01 2.15794384e-01 2.04614595e-01 8.40412736e-01
-5.68150163e-01 -4.78917181e-01 -4.15042192e-02 4.29627091e-01
-1.91085622e-01 -1.04290152e+00 -8.17773938e-01 3.10201883e-01
-6.02653921e-01 -2.05999255e-01 -2.77917057e-01 8.40108246e-02
-1.03206657e-01 1.41924334e+00 -2.13517785e-01 -8.01416099e-01
2.17508778e-01 6.76586390e-01 1.99221507e-01 -3.14735532e-01
-7.94666529e-01 4.11456645e-01 -1.45553499e-02 -1.86887279e-01
-1.76699143e-02 -1.03485656e+00 -6.58531368e-01 -1.67076230e-01
-6.68549538e-01 7.63819646e-03 1.15785980e+00 9.26522076e-01
2.54181385e-01 5.12185991e-01 9.79847431e-01 -9.70672905e-01
-1.26733553e+00 -1.71216476e+00 -6.47675812e-01 2.64702499e-01
9.58922863e-01 -4.70278889e-01 -8.78072023e-01 -6.55994862e-02] | [14.378482818603516, 6.087107181549072] |
d3fb5c5c-3ba0-4f93-a2fb-13c79a70e4f6 | how-can-objects-help-action-recognition-1 | 2306.11726 | null | https://arxiv.org/abs/2306.11726v1 | https://arxiv.org/pdf/2306.11726v1.pdf | How can objects help action recognition? | Current state-of-the-art video models process a video clip as a long sequence of spatio-temporal tokens. However, they do not explicitly model objects, their interactions across the video, and instead process all the tokens in the video. In this paper, we investigate how we can use knowledge of objects to design better video models, namely to process fewer tokens and to improve recognition accuracy. This is in contrast to prior works which either drop tokens at the cost of accuracy, or increase accuracy whilst also increasing the computation required. First, we propose an object-guided token sampling strategy that enables us to retain a small fraction of the input tokens with minimal impact on accuracy. And second, we propose an object-aware attention module that enriches our feature representation with object information and improves overall accuracy. Our resulting framework achieves better performance when using fewer tokens than strong baselines. In particular, we match our baseline with 30%, 40%, and 60% of the input tokens on SomethingElse, Something-something v2, and Epic-Kitchens, respectively. When we use our model to process the same number of tokens as our baseline, we improve by 0.6 to 4.2 points on these datasets. | ['Cordelia Schmid', 'Chen Sun', 'Anurag Arnab', 'Xingyi Zhou'] | 2023-06-20 | how-can-objects-help-action-recognition | http://openaccess.thecvf.com//content/CVPR2023/html/Zhou_How_Can_Objects_Help_Action_Recognition_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhou_How_Can_Objects_Help_Action_Recognition_CVPR_2023_paper.pdf | cvpr-2023-1 | ['action-recognition-in-videos'] | ['computer-vision'] | [ 1.00581832e-01 -1.84438646e-01 -2.96746671e-01 -1.01055823e-01
-6.46832645e-01 -6.29997492e-01 6.77959442e-01 1.89744428e-01
-8.01566839e-01 3.73279065e-01 2.70020992e-01 1.26098007e-01
2.69048065e-01 -6.62150681e-01 -1.06243372e+00 -3.67337078e-01
9.50722620e-02 1.39850184e-01 7.60441124e-01 2.57511020e-01
2.74694264e-01 1.55333608e-01 -1.65531969e+00 6.58857405e-01
6.47462189e-01 1.06721961e+00 2.13265032e-01 7.99447834e-01
-6.22596964e-02 1.13793898e+00 -5.08431196e-01 -4.81251895e-01
2.84852028e-01 -1.12380259e-01 -9.62913990e-01 2.97437906e-01
7.68976092e-01 -6.77155554e-01 -6.59005463e-01 7.22253144e-01
1.60106599e-01 1.70046881e-01 5.72984338e-01 -1.12895143e+00
-6.34121358e-01 6.75713956e-01 -6.81675851e-01 2.56963581e-01
2.66202182e-01 3.62686306e-01 9.42754686e-01 -1.00153768e+00
5.85155368e-01 1.08921063e+00 5.37847996e-01 5.30223191e-01
-1.07120037e+00 -4.17573005e-01 5.72217643e-01 4.85109448e-01
-1.37489843e+00 -7.12547064e-01 3.99051845e-01 -4.06881750e-01
1.06292319e+00 1.59659013e-01 7.92041898e-01 9.45302188e-01
-1.41282395e-01 1.28037119e+00 4.42800403e-01 -3.47873449e-01
9.71275866e-02 -1.12671681e-01 1.89957485e-01 6.10020041e-01
2.08493844e-01 -3.29158425e-01 -7.09722340e-01 -9.96445678e-03
8.47149253e-01 3.91665250e-01 -2.20972911e-01 -3.97662818e-01
-1.40872681e+00 4.50904280e-01 3.13633531e-01 1.89646900e-01
-5.50504804e-01 7.03862786e-01 5.01017988e-01 -9.47090089e-02
4.46265489e-01 1.86617076e-01 -2.52314925e-01 -6.23319387e-01
-1.21057081e+00 3.07956815e-01 4.90042448e-01 1.04490459e+00
7.30578005e-01 -2.85242830e-04 -4.88557518e-01 6.41976416e-01
-4.24281098e-02 3.28196526e-01 2.53551275e-01 -1.09874356e+00
5.65299451e-01 3.68807286e-01 3.73486847e-01 -6.85314476e-01
8.14422816e-02 -9.01689380e-02 -3.38349700e-01 -7.00524598e-02
4.29851592e-01 -2.49878559e-02 -1.21831167e+00 1.59395254e+00
9.91701037e-02 6.36086881e-01 -3.26772183e-01 9.09590304e-01
5.56524634e-01 6.25104368e-01 3.40758979e-01 2.30288535e-01
1.42598665e+00 -1.24152672e+00 -5.02472937e-01 -3.52832496e-01
7.31092632e-01 -6.67204320e-01 1.11110175e+00 2.63922870e-01
-1.45251441e+00 -4.76767570e-01 -7.68985331e-01 -1.51773870e-01
-6.15238510e-02 1.83323398e-01 5.79799533e-01 4.24144417e-01
-1.09262252e+00 7.58007526e-01 -1.21057451e+00 -2.71814227e-01
7.80556858e-01 3.16395491e-01 -3.20425242e-01 -3.23317617e-01
-7.11980522e-01 7.40391910e-01 2.15499297e-01 -1.66908368e-01
-9.66940582e-01 -8.66144240e-01 -9.43990231e-01 4.07714486e-01
6.76805615e-01 -6.32831573e-01 1.41168928e+00 -1.14321458e+00
-1.10787249e+00 5.25223196e-01 -5.84854364e-01 -5.34284115e-01
6.11435354e-01 -7.37264693e-01 2.23321021e-01 4.29024726e-01
-7.08617717e-02 1.01524985e+00 7.15147913e-01 -1.06102896e+00
-6.87109172e-01 5.51032415e-03 4.27971065e-01 1.22874647e-01
-5.04678249e-01 3.43202017e-02 -1.04308808e+00 -8.19160998e-01
-1.54957563e-01 -8.77892375e-01 -1.25819832e-01 2.64451295e-01
-1.52093142e-01 -1.89513609e-01 7.10969388e-01 -5.45185149e-01
1.00761139e+00 -2.25135660e+00 8.24770257e-02 -1.11071497e-01
3.54050845e-01 3.57885569e-01 -3.76852840e-01 4.47843909e-01
2.82726020e-01 3.62924248e-01 -2.35100128e-02 -6.81371272e-01
-6.10993914e-02 2.10305423e-01 -2.04822883e-01 3.48361820e-01
3.96082759e-01 1.01991093e+00 -9.85835612e-01 -4.49034274e-01
2.51030177e-01 4.61024821e-01 -1.04747212e+00 -9.73779336e-02
-3.13109547e-01 -3.50268744e-02 -2.84727931e-01 5.89902937e-01
5.49102545e-01 -2.61793256e-01 1.34027138e-01 -2.13767782e-01
3.32978964e-02 4.66676205e-01 -1.12790501e+00 1.62443745e+00
-3.48630548e-01 8.39169681e-01 -1.37736633e-01 -9.19405043e-01
3.63449663e-01 4.10323590e-01 6.73758566e-01 -4.78556842e-01
1.81217082e-02 -1.16454564e-01 -1.43470570e-01 -5.17672420e-01
8.33895385e-01 2.12859914e-01 1.87508151e-01 4.50016439e-01
-2.66839638e-02 4.07772303e-01 4.02336240e-01 3.68881494e-01
1.33838105e+00 2.67635584e-01 3.37254666e-02 1.13884471e-02
1.39618739e-01 -1.02988943e-01 4.92883474e-01 1.01453948e+00
-1.76299244e-01 8.13248515e-01 5.56790888e-01 -4.27252054e-01
-1.11755252e+00 -9.52339768e-01 2.68444955e-01 1.06858134e+00
2.89734244e-01 -6.73522830e-01 -8.09326530e-01 -7.54383743e-01
1.63229147e-03 4.04526979e-01 -5.83895266e-01 -1.90001369e-01
-7.66600490e-01 -2.96606064e-01 4.65566039e-01 9.86322343e-01
4.47028220e-01 -1.08789539e+00 -6.95805967e-01 1.14497885e-01
-3.02664906e-01 -1.27096045e+00 -8.22639048e-01 -2.29111090e-01
-7.17057168e-01 -1.00167632e+00 -6.87152147e-01 -6.62086844e-01
6.14436746e-01 5.52441895e-01 1.11414218e+00 3.17289382e-01
-8.05471465e-02 3.98920059e-01 -6.29631519e-01 -2.26353064e-01
1.36418343e-01 1.96961120e-01 -1.09518796e-01 1.10792927e-01
4.37967241e-01 -3.60025287e-01 -6.29543364e-01 1.49189770e-01
-1.03630340e+00 1.77381128e-01 5.94283402e-01 7.43173063e-01
3.43322366e-01 -2.86987364e-01 3.12447608e-01 -7.21584618e-01
-2.37202691e-03 -3.83705676e-01 -2.09588259e-01 1.75258651e-01
-1.54365757e-02 7.48068318e-02 6.02314293e-01 -8.77173007e-01
-6.48938954e-01 1.07646652e-01 1.72949910e-01 -1.02102053e+00
-6.73997700e-02 2.73064643e-01 5.95566444e-02 1.73105121e-01
1.95427760e-01 2.52670884e-01 -3.45614254e-01 -4.72542495e-01
3.51105481e-01 3.59357417e-01 3.68928671e-01 -6.53337300e-01
5.81786275e-01 7.08060741e-01 -4.52669472e-01 -6.58793688e-01
-6.17438376e-01 -5.77145636e-01 -5.57640433e-01 -1.10110521e-01
7.78943479e-01 -9.81970727e-01 -8.63956690e-01 5.12322962e-01
-1.18511629e+00 -5.65890551e-01 -3.60303015e-01 5.64633191e-01
-4.43213373e-01 4.33868349e-01 -7.16689050e-01 -7.46076405e-01
-6.99146241e-02 -1.13884318e+00 1.25424564e+00 -1.09221227e-01
-2.15025723e-01 -5.79526246e-01 -4.25865978e-01 1.88680187e-01
3.37309778e-01 -3.91402990e-02 4.92103517e-01 -3.73228014e-01
-1.05165112e+00 -8.66319612e-02 -3.36675346e-01 1.32315949e-01
1.45295322e-01 -1.66774448e-02 -9.44879472e-01 -2.61012733e-01
-4.25665945e-01 -2.25285441e-01 1.25362825e+00 3.49785596e-01
1.36896622e+00 -4.74295825e-01 -4.10843223e-01 4.43221748e-01
1.27378261e+00 5.15597165e-02 7.22263098e-01 2.74993092e-01
8.91753435e-01 4.02732641e-01 5.79545677e-01 5.84904194e-01
5.29171348e-01 7.98177719e-01 3.74569863e-01 8.03435780e-03
-2.29748979e-01 -2.36602083e-01 5.14040291e-01 5.03534257e-01
-1.85314521e-01 -3.60041916e-01 -8.26594532e-01 1.12951958e+00
-2.11557722e+00 -1.30981851e+00 1.79713890e-01 2.15244126e+00
7.42431521e-01 3.20677936e-01 3.55632365e-01 1.98974103e-01
5.89035153e-01 2.73962945e-01 -3.68447930e-01 -1.07538633e-01
2.80524105e-01 1.96955502e-01 7.49806821e-01 3.22619110e-01
-1.13063693e+00 1.10377252e+00 6.32786989e+00 6.91691518e-01
-1.16590607e+00 8.21258407e-03 5.71754098e-01 -8.87980998e-01
-1.40871853e-01 9.22487155e-02 -7.01861620e-01 6.10429645e-01
7.93876410e-01 6.12023547e-02 5.31320393e-01 7.47375369e-01
3.10604244e-01 -2.21697003e-01 -1.47188365e+00 9.70395386e-01
2.08089367e-01 -1.50320208e+00 2.71869630e-01 8.23321715e-02
6.88384593e-01 -2.12485585e-02 -6.35911971e-02 3.61364186e-01
2.48680666e-01 -9.18064117e-01 1.39747310e+00 5.92257440e-01
6.11536562e-01 -6.30425870e-01 7.08907843e-01 2.43671164e-01
-1.38841963e+00 -1.19101509e-01 -2.51404375e-01 -2.04206184e-01
3.11621517e-01 4.58978415e-01 -6.30944908e-01 4.13212210e-01
8.12946200e-01 8.01317871e-01 -5.53697526e-01 1.27803528e+00
6.73642382e-02 7.99605250e-01 -4.47628438e-01 -4.47182208e-02
4.29224014e-01 2.72708088e-01 3.07442427e-01 1.18238115e+00
1.99673086e-01 7.18387812e-02 5.17496765e-01 5.71224153e-01
-3.68376464e-01 -1.55625030e-01 -3.24503243e-01 6.42069876e-02
7.02750087e-01 8.55229437e-01 -6.72326386e-01 -8.38576794e-01
-6.64411485e-01 9.77752149e-01 5.50024569e-01 4.13110346e-01
-1.04444325e+00 -2.70532727e-01 7.57400155e-01 4.88619149e-01
8.91538858e-01 -2.99976647e-01 -2.08208069e-01 -1.28627062e+00
4.58235025e-01 -6.30809903e-01 4.56059910e-02 -5.82222104e-01
-9.44691181e-01 2.99218208e-01 7.56026804e-02 -1.10385060e+00
-1.70801044e-01 -3.99940789e-01 -6.39815629e-01 5.76618552e-01
-1.46169448e+00 -1.13546336e+00 -3.22454333e-01 3.10856551e-01
8.03740323e-01 4.53864634e-01 3.05667371e-01 5.68087339e-01
-4.08691227e-01 6.08394921e-01 -3.48585844e-03 5.03621936e-01
6.75312817e-01 -9.57598329e-01 6.93747640e-01 8.92416656e-01
2.39003837e-01 6.82006717e-01 4.38203007e-01 -5.87177455e-01
-1.44238257e+00 -1.22977364e+00 8.27574193e-01 -6.90881968e-01
4.20058221e-01 -5.41325331e-01 -8.73948932e-01 9.57886219e-01
4.46851999e-02 4.05632317e-01 2.35065252e-01 1.12864546e-01
-4.44734991e-01 6.74022585e-02 -9.59220648e-01 8.05377483e-01
1.25855243e+00 -4.69600856e-01 -4.42748785e-01 -6.45545796e-02
7.49267459e-01 -2.58071840e-01 -6.77283108e-01 1.69119850e-01
8.19541514e-01 -6.83831871e-01 1.00109410e+00 -6.51522458e-01
4.76551116e-01 -3.06995630e-01 -1.14700668e-01 -1.02257442e+00
-3.58823657e-01 -2.85154462e-01 -4.54069644e-01 1.18872988e+00
1.01555780e-01 -2.34583780e-01 1.05848396e+00 5.78418314e-01
-1.29216954e-01 -8.41149986e-01 -8.47355902e-01 -8.68089616e-01
-9.08231139e-02 -5.58971703e-01 6.08649194e-01 6.80888355e-01
7.85148367e-02 -1.48582608e-01 -5.32791138e-01 -1.45046804e-02
4.75947022e-01 -5.84518574e-02 8.36965740e-01 -7.37678945e-01
-2.32786611e-01 -5.19366741e-01 -4.58657205e-01 -1.39064825e+00
3.24876271e-02 -6.61800086e-01 7.13595077e-02 -1.55648398e+00
4.75189269e-01 -3.24339181e-01 -3.32638472e-01 7.25826502e-01
-4.81473118e-01 5.16851008e-01 6.78328216e-01 3.32551420e-01
-1.05878592e+00 4.83287811e-01 9.92381871e-01 -1.68913350e-01
-9.77396891e-02 -3.77199769e-01 -5.26052833e-01 6.55576169e-01
5.80848455e-01 -4.65223432e-01 -1.96759164e-01 -9.35335457e-01
-2.79385149e-01 -1.58232942e-01 6.08979166e-01 -1.03268468e+00
1.27393588e-01 -1.50385082e-01 5.42755961e-01 -5.62663436e-01
6.27431273e-01 -7.07420349e-01 -1.34089351e-01 3.35466743e-01
-4.33895528e-01 1.72694698e-02 1.97051883e-01 4.80992973e-01
-1.32110074e-01 -2.14309379e-01 5.43725789e-01 -2.11255997e-01
-7.28909791e-01 3.81426662e-01 -6.29890919e-01 -8.40099081e-02
1.04402137e+00 -4.07056928e-01 -2.40588039e-01 -3.43483329e-01
-5.64247012e-01 2.51586109e-01 7.38145053e-01 6.11298323e-01
4.84578073e-01 -1.34157217e+00 -6.53397441e-01 -4.88209836e-02
4.69487868e-02 3.89171839e-02 2.93454528e-01 7.43076086e-01
-4.28406686e-01 4.30232704e-01 -1.30186640e-02 -6.94967091e-01
-1.32440603e+00 6.86515033e-01 2.86872359e-03 -1.48937717e-01
-6.40800953e-01 6.21817410e-01 3.78575772e-01 1.37675390e-01
4.01871771e-01 -6.20751441e-01 -2.23183185e-02 -3.71834412e-02
6.51797235e-01 4.07544672e-01 -1.56901076e-01 -4.70790863e-01
-4.69567686e-01 6.45333827e-01 -3.61834347e-01 -1.06880791e-01
1.25147867e+00 7.90480152e-02 3.72953296e-01 2.19330847e-01
1.01246095e+00 1.61123797e-01 -1.76117659e+00 -3.44550580e-01
-1.77655071e-02 -8.19171309e-01 -9.14086178e-02 -4.25649077e-01
-1.15369201e+00 7.08298981e-01 1.96865067e-01 2.59370115e-02
8.86258245e-01 1.40602753e-01 9.30723608e-01 2.84448057e-01
3.11808974e-01 -8.37708890e-01 2.21674174e-01 4.46232170e-01
4.14740324e-01 -1.16651714e+00 -5.07435724e-02 -4.44579154e-01
-6.76240861e-01 7.45946944e-01 7.07933724e-01 -3.63033235e-01
1.80408478e-01 3.56190503e-01 -3.19550425e-01 -1.05157401e-03
-1.11684811e+00 -2.83747345e-01 1.68458909e-01 4.24097806e-01
4.10590738e-01 -2.46154830e-01 -6.20143116e-02 5.06542146e-01
1.44111827e-01 2.48860210e-01 5.50382197e-01 1.24873710e+00
-5.74949801e-01 -9.30995345e-01 -3.01558435e-01 6.04068756e-01
-6.70821249e-01 -2.81348318e-01 -1.14661001e-01 6.37901187e-01
8.30631107e-02 8.30454767e-01 3.98160249e-01 -2.35332593e-01
2.93760478e-01 1.20957427e-01 5.04116654e-01 -6.21418834e-01
-5.67499399e-01 8.03590119e-02 5.52699082e-02 -6.29729569e-01
-5.05871177e-01 -8.26963842e-01 -1.17863786e+00 -5.61691284e-01
-3.46129775e-01 -2.28551496e-02 3.39472324e-01 9.02586043e-01
5.20066500e-01 5.41164756e-01 2.58657962e-01 -1.08582330e+00
-2.64977396e-01 -9.41157460e-01 -2.09585443e-01 5.45725167e-01
3.65924716e-01 -7.16553748e-01 -9.55401585e-02 4.56128031e-01] | [9.136466979980469, 0.427143931388855] |
4867e9aa-698c-488b-a6a6-1c9c0cb5c555 | semeval-2016-task-3-community-question-1 | 1912.01972 | null | https://arxiv.org/abs/1912.01972v1 | https://arxiv.org/pdf/1912.01972v1.pdf | SemEval-2016 Task 3: Community Question Answering | This paper describes the SemEval--2016 Task 3 on Community Question Answering, which we offered in English and Arabic. For English, we had three subtasks: Question--Comment Similarity (subtask A), Question--Question Similarity (B), and Question--External Comment Similarity (C). For Arabic, we had another subtask: Rerank the correct answers for a new question (D). Eighteen teams participated in the task, submitting a total of 95 runs (38 primary and 57 contrastive) for the four subtasks. A variety of approaches and features were used by the participating systems to address the different subtasks, which are summarized in this paper. The best systems achieved an official score (MAP) of 79.19, 76.70, 55.41, and 45.83 in subtasks A, B, C, and D, respectively. These scores are significantly better than those for the baselines that we provided. For subtask A, the best system improved over the 2015 winner by 3 points absolute in terms of Accuracy. | ['Walid Magdy', 'Lluís Màrquez', 'Abed Alhakim Freihat', 'James Glass', 'Preslav Nakov', 'Hamdy Mubarak', 'Bilal Randeree', 'Alessandro Moschitti'] | 2019-12-03 | semeval-2016-task-3-community-question | https://aclanthology.org/S16-1083 | https://aclanthology.org/S16-1083.pdf | semeval-2016-6 | ['question-similarity'] | ['natural-language-processing'] | [-1.07253335e-01 -7.96975195e-02 5.39743960e-01 -3.31094325e-01
-1.50950658e+00 -9.12066996e-01 8.86271894e-01 6.44457996e-01
-7.09023118e-01 7.76302993e-01 5.13490379e-01 -4.01109993e-01
2.08193958e-02 -1.89521506e-01 -3.58147949e-01 -2.61117250e-01
9.41252783e-02 5.77359915e-01 6.47573888e-01 -7.05065548e-01
7.02685177e-01 -3.28069448e-01 -1.34804678e+00 9.14469182e-01
1.26567554e+00 1.03057814e+00 -1.05524451e-01 1.19236851e+00
-2.45249853e-01 9.80342746e-01 -8.79247725e-01 -9.33918893e-01
-2.56113678e-01 -5.42257726e-01 -1.52167380e+00 -4.21541780e-01
7.57398248e-01 -4.90673855e-02 -1.30359465e-02 5.45580566e-01
6.92083716e-01 2.23449841e-01 7.64361024e-01 -1.06647432e+00
-1.06171644e+00 4.39626426e-01 -1.75094634e-01 3.01764995e-01
1.25275850e+00 -5.84951602e-02 1.55554354e+00 -1.38654983e+00
4.66060579e-01 1.19952309e+00 6.76921308e-01 5.31092703e-01
-7.89226532e-01 -4.74041551e-01 1.79245528e-02 5.99164307e-01
-1.17640698e+00 -3.93486261e-01 6.54299408e-02 -5.49649000e-01
1.23426425e+00 6.50063932e-01 1.30670264e-01 8.10372591e-01
-1.51994368e-02 6.86097860e-01 1.33695054e+00 -4.40756857e-01
1.38999462e-01 1.38582796e-01 5.62955856e-01 2.38550216e-01
-5.58392555e-02 -6.04387999e-01 -4.36696440e-01 -4.51066762e-01
-2.28926510e-01 -4.15276319e-01 -3.55232626e-01 5.14595151e-01
-1.29410410e+00 8.68111134e-01 5.78647256e-01 4.64175254e-01
-2.71160632e-01 -2.92267948e-01 4.61218029e-01 8.09406519e-01
3.66407156e-01 8.41635287e-01 -7.09094822e-01 -1.41389668e-01
-5.69417536e-01 9.26287949e-01 1.37714267e+00 8.36379111e-01
5.69768727e-01 -6.07220292e-01 -7.94339955e-01 9.12582219e-01
1.36211261e-01 5.78320801e-01 4.12857950e-01 -1.18746173e+00
9.76787746e-01 5.56809247e-01 4.72449183e-01 -1.01774561e+00
-5.78097761e-01 -2.04055369e-01 -7.11652815e-01 -5.63627362e-01
7.99045265e-01 -3.45486820e-01 -5.37221014e-01 1.27095437e+00
9.77455080e-02 -3.92275929e-01 2.28399396e-01 7.95426607e-01
1.70966458e+00 6.96068108e-01 6.25055656e-02 1.27295539e-01
1.72056866e+00 -1.49815178e+00 -7.86113024e-01 -1.99336782e-01
7.45581567e-01 -1.29454923e+00 1.37575257e+00 1.17286302e-01
-1.13750160e+00 -5.10610402e-01 -6.79600358e-01 -1.95271730e-01
-3.54524821e-01 9.40190852e-02 -1.45287782e-01 3.84727657e-01
-1.41072881e+00 9.15868431e-02 -6.29740134e-02 -7.10490525e-01
-2.97760367e-01 -2.75618657e-02 -2.19687656e-01 -9.28056091e-02
-1.64036429e+00 1.17688930e+00 7.36418217e-02 -1.55483201e-01
-5.36188602e-01 -4.91139084e-01 -5.16022325e-01 -7.32333511e-02
1.90128893e-01 -5.96210897e-01 1.75554204e+00 -7.43151784e-01
-1.28694975e+00 1.18954670e+00 -3.06476653e-01 -3.73137683e-01
2.88671285e-01 -3.74291122e-01 -5.11421859e-01 3.34654093e-01
4.82895941e-01 4.57342565e-01 5.45872092e-01 -9.71628964e-01
-7.39233434e-01 -2.41463557e-01 5.62140167e-01 3.97024840e-01
-9.62046087e-02 7.84750819e-01 -3.30441028e-01 -6.35787785e-01
1.38734207e-01 -9.34274912e-01 -5.33897150e-03 -5.53247094e-01
-1.87265158e-01 -8.44805658e-01 2.51619011e-01 -1.48853374e+00
1.51931059e+00 -1.73329258e+00 6.66913539e-02 -2.89158762e-01
3.12887549e-01 2.41861522e-01 -3.99497598e-01 8.62927973e-01
-2.20621023e-02 2.09901392e-01 -2.97121078e-01 -2.87184060e-01
6.59707785e-02 -1.71015948e-01 -2.28459254e-01 -6.08424358e-02
3.17038476e-01 1.04008913e+00 -9.59438801e-01 -2.88924545e-01
-6.98180854e-01 -9.84855518e-02 -5.42739570e-01 3.44329208e-01
-3.39719057e-01 3.80620062e-01 -2.64119148e-01 6.14945412e-01
4.59099114e-01 -5.08396268e-01 -1.91267818e-01 3.01935107e-01
4.20599543e-02 7.89658487e-01 -8.65308821e-01 1.48661947e+00
-2.86608279e-01 6.22973561e-01 2.28133217e-01 -3.95827562e-01
8.93121958e-01 6.18634999e-01 -2.60011666e-02 -7.57539928e-01
-2.47786164e-01 6.20570540e-01 1.78795725e-01 -8.86113644e-01
7.77274132e-01 3.27180535e-01 -2.64119685e-01 8.58136892e-01
-6.02006689e-02 -3.47577542e-01 5.58369040e-01 6.27375185e-01
1.26130915e+00 -4.02083784e-01 6.94575384e-02 -4.45341736e-01
1.01038146e+00 4.21701334e-02 -1.95818841e-01 8.94362092e-01
-4.45090234e-01 8.95890892e-01 5.71982563e-01 -3.42853606e-01
-6.77281678e-01 -8.30276012e-01 2.51662463e-01 1.58272982e+00
-1.15777969e-01 -8.15837920e-01 -7.76699245e-01 -9.31385398e-01
-4.11522761e-02 4.74695057e-01 -7.02046931e-01 2.04071879e-01
-7.03962743e-01 -4.66983974e-01 7.72745252e-01 3.32365572e-01
5.15047491e-01 -9.97755527e-01 -1.87848911e-01 1.25778466e-01
-1.10883534e+00 -1.07279110e+00 -7.82083750e-01 -1.23338155e-01
-5.69427013e-01 -1.23788297e+00 -1.15888357e+00 -1.07436121e+00
1.09746508e-01 2.83261329e-01 1.59437287e+00 4.71099287e-01
3.26188177e-01 4.70643401e-01 -9.24803734e-01 -2.20889956e-01
-2.07470343e-01 4.86278564e-01 -3.28318268e-01 -2.75049329e-01
2.72265077e-01 -5.53040095e-02 -5.76944351e-01 4.88660514e-01
-6.15598261e-01 -4.75190043e-01 3.51240337e-01 8.64446163e-01
1.63238328e-02 -9.77159381e-01 1.09641349e+00 -5.79944432e-01
1.15789938e+00 -7.30351210e-01 -1.71150989e-03 5.94069719e-01
-3.97841722e-01 -1.88872427e-01 4.83353287e-01 -5.17494828e-02
-8.18317533e-01 -4.85809386e-01 -5.06402850e-01 4.97101188e-01
-2.91558690e-02 6.84506834e-01 2.65938699e-01 9.88438055e-02
1.14514244e+00 -5.13552576e-02 -2.53010482e-01 -5.77817142e-01
2.90466577e-01 1.05665398e+00 4.28938925e-01 -6.28931522e-01
6.30922735e-01 -2.17098996e-01 -7.77782738e-01 -6.39733493e-01
-1.30196989e+00 -8.37564349e-01 -4.42966074e-01 -1.41888559e-01
1.03662193e+00 -9.77996945e-01 -6.15268707e-01 7.10802197e-01
-1.50203931e+00 -1.09432995e-01 -2.96622235e-03 7.80575573e-02
-3.75479124e-02 5.53797603e-01 -7.68629253e-01 -4.07470942e-01
-7.02030361e-01 -9.07068014e-01 8.43012810e-01 2.52176732e-01
-5.55871308e-01 -9.59851444e-01 1.57834217e-01 1.00848985e+00
7.52151906e-01 1.56103177e-02 6.40475571e-01 -1.30877745e+00
1.06219009e-01 -6.47672936e-02 -3.70331287e-01 2.45956406e-01
-1.05886608e-01 -9.26953331e-02 -8.74883115e-01 -3.40272516e-01
-6.64001405e-02 -8.88111949e-01 7.33700275e-01 -2.22407654e-01
7.43144631e-01 -4.47639465e-01 1.85924575e-01 -2.90152162e-01
6.82018697e-01 -2.16389686e-01 4.41974759e-01 5.45470238e-01
2.68424392e-01 9.27907169e-01 5.76369762e-01 2.40382344e-01
1.19576502e+00 7.66329944e-01 2.36072138e-01 3.76698703e-01
-9.50737894e-02 1.03736870e-01 4.47349936e-01 1.37480032e+00
-2.73352358e-02 -1.82132572e-01 -1.34234846e+00 8.08088601e-01
-1.80436778e+00 -7.14345455e-01 -8.69916320e-01 1.81580973e+00
1.06071126e+00 -2.08906695e-01 3.81470978e-01 5.40023148e-02
6.36060178e-01 1.40227526e-01 -1.39600977e-01 -5.91794968e-01
-2.80529350e-01 2.76152045e-01 -3.01995933e-01 6.37223482e-01
-1.06772900e+00 6.62911534e-01 6.56632900e+00 5.01158237e-01
-3.85818124e-01 4.12144601e-01 4.28449869e-01 3.40658873e-01
-3.55707079e-01 1.06853328e-03 -6.92333221e-01 4.10817564e-01
1.31221557e+00 -1.80506811e-01 1.80136427e-01 3.12605679e-01
-3.39665681e-01 -2.99412131e-01 -9.28107083e-01 6.92068517e-01
5.80580473e-01 -1.03146029e+00 7.12489560e-02 -7.07702458e-01
1.11244893e+00 1.50013804e-01 -2.02627167e-01 7.40873992e-01
3.88062149e-01 -1.10585904e+00 6.58508182e-01 4.54289913e-01
3.57564300e-01 -4.48379725e-01 1.17337978e+00 4.32923049e-01
-9.86187279e-01 -6.61495551e-02 -1.11023255e-01 -4.88642037e-01
9.30683762e-02 2.48374760e-01 -5.05611897e-01 7.63147891e-01
1.16024268e+00 4.58337694e-01 -1.16087925e+00 1.11023581e+00
-6.38540983e-01 8.17271113e-01 -5.94356358e-02 -5.18653154e-01
3.89391601e-01 1.07673146e-01 4.76457447e-01 1.34556472e+00
-1.98876467e-02 5.42184040e-02 9.34385210e-02 2.96990037e-01
-2.96226799e-01 4.09300059e-01 2.01980457e-01 2.01216966e-01
6.15336299e-01 1.20418465e+00 -2.68222451e-01 -4.80681211e-01
-3.17322642e-01 1.02194893e+00 4.83928800e-01 4.28498268e-01
-6.43908262e-01 -9.96378720e-01 1.97910756e-01 -5.33222482e-02
1.04123913e-01 -6.42476976e-02 -1.99021518e-01 -1.17751968e+00
3.30831558e-01 -1.44928253e+00 8.58198762e-01 -7.22602546e-01
-1.71291161e+00 8.54923010e-01 -2.88639754e-01 -8.03406894e-01
-3.05548698e-01 -3.98959041e-01 -6.48728430e-01 1.22248411e+00
-1.49557400e+00 -8.68406594e-01 -5.38117826e-01 6.95189536e-01
5.35957098e-01 -1.23697445e-01 1.01783133e+00 4.94889677e-01
-1.70548528e-01 7.97997952e-01 1.69680655e-01 2.26364106e-01
1.37451673e+00 -1.43790877e+00 6.36869967e-01 6.02274358e-01
9.28173289e-02 5.24062097e-01 5.35913110e-01 -2.87607878e-01
-8.99021745e-01 -7.83826470e-01 2.06081200e+00 -9.98436809e-01
1.05503809e+00 -7.08148032e-02 -1.08468831e+00 3.95992905e-01
6.50671065e-01 -4.49544787e-01 7.65427828e-01 2.06709653e-01
-5.32114387e-01 2.09634840e-01 -8.72943699e-01 4.18762654e-01
4.66400087e-01 -8.85161817e-01 -1.05314684e+00 6.48904383e-01
9.28119063e-01 -6.01384342e-01 -1.05718553e+00 1.71987906e-01
4.77663219e-01 -8.83129120e-01 8.33957851e-01 -8.37772071e-01
6.15311980e-01 -2.24485978e-01 -2.95516729e-01 -1.09605968e+00
-3.10978740e-01 -5.85661948e-01 -1.26979217e-01 1.17874300e+00
8.71050477e-01 -6.51662350e-01 2.49057963e-01 3.20111156e-01
-2.58106112e-01 -8.29262912e-01 -9.48114514e-01 -4.98935401e-01
6.56201661e-01 -7.74418041e-02 4.77282524e-01 8.87145042e-01
1.97281986e-02 9.09896016e-01 5.96763901e-02 -7.29128495e-02
-7.82964379e-02 8.50482062e-02 6.01511121e-01 -1.17897868e+00
-8.73839781e-02 -5.96645713e-01 2.18546599e-01 -1.18523884e+00
-9.15274676e-03 -1.03634787e+00 -1.62350424e-02 -2.08102751e+00
2.74899572e-01 -1.11018285e-01 -6.95288107e-02 3.28205377e-01
-9.01607454e-01 3.92917693e-01 3.30814421e-01 3.99266988e-01
-1.19277251e+00 4.00821120e-01 1.06558359e+00 -1.42758101e-01
9.74487960e-02 2.10195944e-01 -1.02596676e+00 5.07966757e-01
9.30567980e-01 -3.89708310e-01 2.47691810e-01 -7.41107762e-01
5.35437822e-01 -1.81298256e-01 2.99427837e-01 -7.41736352e-01
4.68761057e-01 2.53738821e-01 -8.10947642e-02 -5.63584507e-01
1.03666969e-01 -2.50630885e-01 -4.91565078e-01 5.59041083e-01
-5.57361722e-01 6.34635210e-01 -7.29039758e-02 2.10200861e-01
-5.11760235e-01 -4.71048474e-01 4.70161855e-01 5.30700944e-02
-3.81205231e-01 -1.87465027e-01 -5.84616601e-01 9.99066651e-01
6.47919655e-01 1.73156634e-01 -7.05491841e-01 -8.56826484e-01
-7.51551986e-01 7.67409086e-01 -6.73057958e-02 7.28760719e-01
5.58597386e-01 -1.30154884e+00 -1.51200533e+00 -4.55861509e-01
5.11340857e-01 -2.91044027e-01 2.76224911e-01 1.13556349e+00
-5.33429682e-01 6.90511584e-01 1.34142771e-01 -4.57013845e-01
-1.28945029e+00 -1.17405087e-01 2.56269366e-01 -8.25622439e-01
9.34092179e-02 1.15585029e+00 -5.26857555e-01 -1.01324058e+00
3.60043049e-01 -1.56560764e-01 -7.28040338e-01 5.80365360e-01
7.76212275e-01 7.57121384e-01 5.03631353e-01 -7.02570081e-01
-4.06570703e-01 6.34527862e-01 -2.40198746e-01 -9.97480825e-02
9.65482116e-01 -2.34934852e-01 -5.71534634e-01 4.02342051e-01
9.97417629e-01 -1.37679977e-02 -4.86193568e-01 -5.48435688e-01
6.34560406e-01 -1.04209848e-01 -6.25716150e-01 -1.40815318e+00
-3.14962387e-01 8.48022997e-01 2.08036721e-01 7.14412510e-01
8.34122837e-01 2.55688518e-01 9.63007808e-01 7.30616629e-01
-8.74644495e-04 -9.73593831e-01 4.50112671e-01 1.55417490e+00
1.51312363e+00 -1.33358335e+00 -3.54506344e-01 1.72431264e-02
-8.04710567e-01 8.98736238e-01 6.24555647e-01 1.35221988e-01
3.56505662e-01 -3.82732898e-01 2.09739715e-01 -1.89940169e-01
-9.50137019e-01 -2.45797411e-01 8.65599334e-01 3.44027042e-01
7.93459773e-01 -8.62983689e-02 -6.65281296e-01 1.01117802e+00
-3.63693386e-01 -5.12800753e-01 3.77016038e-01 9.58562255e-01
-6.26466632e-01 -8.31303656e-01 -4.66308534e-01 3.86285990e-01
-7.55517900e-01 -2.98057377e-01 -9.72211719e-01 5.88499129e-01
-2.64105052e-01 1.73373425e+00 -1.27239019e-01 -4.55010265e-01
6.67842686e-01 1.50409371e-01 1.59972414e-01 -7.26753533e-01
-1.49822366e+00 -5.73228538e-01 5.68467796e-01 -2.85764068e-01
-4.19893622e-01 -5.93936622e-01 -1.06748998e+00 -3.71307999e-01
-4.49091524e-01 6.81934714e-01 3.42790633e-01 9.53833699e-01
3.71409357e-01 2.24951968e-01 4.76768732e-01 -3.00651640e-01
-8.61337006e-01 -1.53047478e+00 -1.10673510e-01 3.97061646e-01
3.39584857e-01 -1.40284657e-01 -5.83146513e-01 -4.92865331e-02] | [11.369216918945312, 8.019886016845703] |
6b0e0544-8c5e-4ff8-b9ff-930af9335d1d | semantic-aware-scene-recognition | 1909.02410 | null | https://arxiv.org/abs/1909.02410v3 | https://arxiv.org/pdf/1909.02410v3.pdf | Semantic-Aware Scene Recognition | Scene recognition is currently one of the top-challenging research fields in computer vision. This may be due to the ambiguity between classes: images of several scene classes may share similar objects, which causes confusion among them. The problem is aggravated when images of a particular scene class are notably different. Convolutional Neural Networks (CNNs) have significantly boosted performance in scene recognition, albeit it is still far below from other recognition tasks (e.g., object or image recognition). In this paper, we describe a novel approach for scene recognition based on an end-to-end multi-modal CNN that combines image and context information by means of an attention module. Context information, in the shape of semantic segmentation, is used to gate features extracted from the RGB image by leveraging on information encoded in the semantic representation: the set of scene objects and stuff, and their relative locations. This gating process reinforces the learning of indicative scene content and enhances scene disambiguation by refocusing the receptive fields of the CNN towards them. Experimental results on four publicly available datasets show that the proposed approach outperforms every other state-of-the-art method while significantly reducing the number of network parameters. All the code and data used along this paper is available at https://github.com/vpulab/Semantic-Aware-Scene-Recognition | ['Álvaro García-Martín', 'Jesús Bescós', 'Marcos Escudero-Viñolo', 'Alejandro López-Cifuentes'] | 2019-09-05 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [ 7.78828979e-01 -3.46679479e-01 1.35708362e-01 -6.62275553e-01
-5.03181159e-01 -5.56290567e-01 6.41506076e-01 3.88597697e-01
-6.25442922e-01 3.80119652e-01 5.76506630e-02 5.12214974e-02
-1.45488814e-01 -8.02931249e-01 -7.77108669e-01 -1.06187272e+00
5.31824887e-01 1.06360376e-01 3.39678675e-01 -1.95133090e-02
6.43181682e-01 5.67047358e-01 -1.77919281e+00 5.88351667e-01
4.83270854e-01 1.24953234e+00 7.76677251e-01 3.84093523e-01
-3.47481608e-01 7.87090123e-01 -4.00488794e-01 -1.02319214e-02
7.98956156e-02 -3.43245953e-01 -9.31692958e-01 2.99198061e-01
6.29888058e-01 1.94922241e-03 -1.13258563e-01 1.31925356e+00
2.91957259e-01 3.31112325e-01 3.40159476e-01 -1.05266321e+00
-4.76600140e-01 2.23682612e-01 -3.39124888e-01 4.47513521e-01
4.78543676e-02 9.86670181e-02 8.69407654e-01 -8.42796743e-01
3.44476759e-01 1.13787580e+00 2.06756204e-01 3.62102211e-01
-9.95792925e-01 -3.40301067e-01 3.48065734e-01 4.49740529e-01
-1.49763060e+00 -3.43051285e-01 1.00028801e+00 -3.61592829e-01
7.85898268e-01 2.41453469e-01 4.28208679e-01 8.41825068e-01
-1.56840950e-01 7.37805247e-01 1.16123831e+00 -3.53326589e-01
3.33272994e-01 8.94417986e-02 1.70647204e-01 4.82574344e-01
2.16967866e-01 -1.09642223e-01 -4.25153524e-01 2.72458434e-01
7.40458786e-01 3.99165571e-01 -2.64108241e-01 -3.86794716e-01
-1.19501936e+00 6.31002426e-01 1.03202569e+00 5.72278619e-01
-5.44221878e-01 1.29537880e-01 1.45884022e-01 -2.79206872e-01
1.76407561e-01 4.17883307e-01 -3.88276726e-01 1.71396270e-01
-6.69286847e-01 8.23842213e-02 3.84976655e-01 5.86376965e-01
1.05889392e+00 -1.39758334e-01 -1.79206312e-01 8.49567533e-01
2.80270696e-01 3.51919144e-01 4.89633054e-01 -5.65674722e-01
2.96513587e-01 8.87431562e-01 -2.74193764e-01 -1.06327486e+00
-4.74154294e-01 -5.29007614e-01 -7.09746778e-01 6.18264563e-02
5.11259317e-01 2.96379775e-01 -1.16658330e+00 1.50700879e+00
4.28650230e-01 4.08176243e-01 9.78325307e-02 1.30295849e+00
1.13028336e+00 5.77929616e-01 4.04399604e-01 3.89979333e-01
1.54387784e+00 -1.08724320e+00 -2.56806880e-01 -7.16324210e-01
3.22009414e-01 -7.81257451e-01 8.76567841e-01 1.45758688e-01
-6.22128367e-01 -7.71756649e-01 -7.78231561e-01 -1.59486890e-01
-8.52188349e-01 1.26341969e-01 7.23541498e-01 3.25793535e-01
-9.01754200e-01 3.78197163e-01 -4.09644157e-01 -7.32739031e-01
8.72524917e-01 2.28821591e-01 -3.26749593e-01 -3.41797113e-01
-7.65441239e-01 6.98034942e-01 9.11177635e-01 3.35738212e-01
-9.26702201e-01 -4.05113935e-01 -9.78948236e-01 5.53437062e-02
4.35756773e-01 -5.94031692e-01 9.13237274e-01 -1.32649398e+00
-1.16631722e+00 1.02862573e+00 -1.36359837e-02 -1.40821382e-01
-3.66935544e-02 -2.62881368e-01 -2.61052728e-01 2.29424760e-01
2.40281492e-01 8.22713494e-01 8.36215138e-01 -1.37282991e+00
-7.59480953e-01 -5.75375736e-01 1.31952956e-01 4.42920804e-01
-8.80694836e-02 2.13995688e-02 -5.71778357e-01 -3.49713326e-01
2.92372614e-01 -6.99757814e-01 -3.00898790e-01 -1.32773921e-01
-5.99739671e-01 -1.86882704e-01 9.02080119e-01 -2.53835201e-01
6.54906929e-01 -2.37663841e+00 4.39622924e-02 5.06088436e-02
-1.99465647e-01 3.30938905e-01 -1.88404471e-01 2.20914975e-01
-1.67259306e-01 -9.14694071e-02 -5.31344533e-01 -1.25336587e-01
-2.85775453e-01 3.34711254e-01 -2.98636407e-01 5.67930877e-01
4.82125729e-01 8.60720754e-01 -9.94758546e-01 -2.91246623e-01
7.30075896e-01 5.61257660e-01 -2.42769733e-01 1.71882436e-01
-2.66974032e-01 8.01078737e-01 -4.07280505e-01 8.17449152e-01
6.56608522e-01 -3.77428621e-01 9.68040824e-02 -4.09159124e-01
-8.57347399e-02 2.30796486e-01 -1.26082850e+00 1.89875031e+00
-3.60734373e-01 5.17455041e-01 -2.21470594e-01 -1.39813828e+00
8.31015110e-01 -1.30357534e-01 1.82596520e-01 -9.00413811e-01
4.27170247e-01 1.81063414e-01 -7.48633891e-02 -7.01518714e-01
4.44069684e-01 1.41258791e-01 -1.31175950e-01 4.28408049e-02
1.72625288e-01 -2.07132995e-01 1.58581764e-01 -1.02696948e-01
5.99379599e-01 1.15458965e-01 2.81695545e-01 -2.36514688e-01
6.54875934e-01 9.52103883e-02 3.75316471e-01 7.81204522e-01
-2.48220459e-01 6.29216433e-01 6.24141917e-02 -4.21767145e-01
-7.41820037e-01 -9.80046630e-01 -3.15647721e-01 1.12283719e+00
6.34262443e-01 -6.47656769e-02 -6.12236202e-01 -4.80024338e-01
-5.40546291e-02 5.91468811e-01 -7.56210446e-01 -7.34252855e-02
-3.61012489e-01 -7.61962950e-01 2.69205093e-01 5.77730179e-01
8.67517292e-01 -1.35289764e+00 -9.94302690e-01 9.67017934e-02
-1.38845548e-01 -1.41015840e+00 -6.09325245e-03 3.84670675e-01
-7.58105695e-01 -1.23757803e+00 -3.50741148e-01 -9.36425447e-01
8.63284826e-01 6.45947576e-01 9.20870543e-01 1.23104542e-01
-5.00110626e-01 4.21699524e-01 -3.84971738e-01 -4.03221726e-01
9.38721150e-02 -2.75395904e-02 -3.96358639e-01 5.37700772e-01
5.05113304e-01 -3.63540798e-01 -7.53872931e-01 1.03329346e-01
-1.16237688e+00 2.63874978e-01 6.82517171e-01 7.15689778e-01
7.77572572e-01 1.30868062e-01 1.88893855e-01 -9.17633414e-01
-2.25572977e-02 -5.43112338e-01 -5.02660394e-01 9.94681194e-02
-1.35980397e-01 -9.58965644e-02 4.04876441e-01 -1.46187633e-01
-1.02875662e+00 3.18556875e-01 5.99136762e-02 -3.03985536e-01
-8.83611977e-01 4.33983058e-01 -3.25121105e-01 -1.43828169e-01
3.75164866e-01 3.89601558e-01 -4.56298739e-01 -3.87314111e-01
3.44523132e-01 6.26672566e-01 6.87982619e-01 -2.71272302e-01
4.69003171e-01 7.13283241e-01 -1.75346565e-02 -9.71422851e-01
-1.23403478e+00 -8.96895885e-01 -8.30027401e-01 -3.87632638e-01
1.07586694e+00 -8.69497180e-01 -4.67796594e-01 7.15692997e-01
-1.03311670e+00 -3.18623513e-01 -1.25667676e-01 3.33259404e-01
-3.67970586e-01 -8.66722874e-03 -1.46318734e-01 -7.04893708e-01
-2.34577507e-02 -1.11926854e+00 1.27903533e+00 7.40974903e-01
1.20194316e-01 -8.82793665e-01 -3.32861543e-01 5.82227588e-01
4.78480428e-01 2.64586687e-01 8.20163727e-01 -7.53062606e-01
-8.07221770e-01 3.03982440e-02 -6.79228544e-01 1.75332561e-01
2.25249082e-01 -1.35472357e-01 -1.46120477e+00 1.80604104e-02
-1.85013801e-01 -3.06302577e-01 1.10151970e+00 4.03387427e-01
1.52602208e+00 -5.20351231e-02 -3.41766834e-01 6.73319936e-01
1.85931063e+00 2.33299449e-01 6.75834179e-01 5.14733911e-01
9.69670653e-01 8.28222334e-01 5.27382851e-01 3.36054951e-01
2.78817624e-01 6.33708477e-01 8.26658010e-01 -2.42317706e-01
-2.72451699e-01 5.70259504e-02 -2.08654050e-02 1.53031588e-01
1.93907350e-01 -1.35415718e-01 -9.90008712e-01 7.43707597e-01
-1.85915887e+00 -9.19539511e-01 -1.47302732e-01 1.93326497e+00
5.95991671e-01 -7.81104667e-03 -2.46264264e-01 7.41499215e-02
7.74288297e-01 2.96239883e-01 -6.39524758e-01 -2.76346982e-01
-3.15115392e-01 3.01424414e-01 4.55870777e-01 2.07216144e-01
-1.41563535e+00 1.12577391e+00 4.61619663e+00 6.73560500e-01
-1.44938576e+00 -1.87949743e-02 7.36198008e-01 1.41433597e-01
2.10939482e-01 -7.70792067e-02 -6.89602911e-01 4.01562423e-01
5.34695506e-01 3.04116428e-01 3.79363775e-01 9.12499070e-01
-5.70580773e-02 -5.12143552e-01 -8.84475887e-01 1.09583080e+00
3.35537642e-01 -1.24272454e+00 4.78135385e-02 -3.58332068e-01
6.63886964e-01 3.15536708e-01 4.29086387e-02 5.95098622e-02
-1.20383576e-01 -9.48114216e-01 7.89555192e-01 4.56163764e-01
3.87495041e-01 -4.81698692e-01 7.11771250e-01 7.70973116e-02
-1.29589128e+00 -2.66537786e-01 -4.03815508e-01 -1.07576914e-01
-1.53266773e-01 5.16782284e-01 -7.21321940e-01 5.20024300e-01
9.52784419e-01 8.19038570e-01 -8.88988614e-01 1.28401756e+00
-3.70000273e-01 4.31247085e-01 -2.26989031e-01 -9.24999267e-02
4.62717921e-01 -8.11500847e-02 2.72926182e-01 1.35221505e+00
7.05448762e-02 9.01032463e-02 2.77847826e-01 1.08745241e+00
-1.62256937e-02 4.75271083e-02 -6.73405230e-01 2.18218267e-02
2.55069882e-01 1.65659368e+00 -1.12678897e+00 -3.79096776e-01
-3.30429673e-01 1.11972296e+00 2.66587675e-01 4.31252122e-01
-6.46533608e-01 -3.94536585e-01 5.40405273e-01 -2.43818462e-01
4.83362913e-01 -1.39682636e-01 -4.09089476e-01 -9.29322481e-01
3.70044522e-02 -3.37430745e-01 4.30220634e-01 -9.22549069e-01
-1.15162146e+00 6.76461756e-01 -8.08540136e-02 -9.41910446e-01
1.79019615e-01 -8.15651238e-01 -5.68905056e-01 8.33279252e-01
-1.90513849e+00 -1.22718477e+00 -7.04204381e-01 7.31402099e-01
6.09190881e-01 1.92484558e-01 7.63276279e-01 3.06318581e-01
-5.86952209e-01 7.53969327e-02 -5.91355897e-02 3.64678025e-01
3.74784380e-01 -1.06853652e+00 6.18723370e-02 1.02237582e+00
2.94351369e-01 3.60966802e-01 3.85745972e-01 -2.73001552e-01
-1.19227231e+00 -1.37370288e+00 7.34915197e-01 -3.29383522e-01
4.16138202e-01 -4.21683758e-01 -9.07512784e-01 2.79482424e-01
2.51026869e-01 2.48766243e-01 6.35623813e-01 -9.95045304e-02
-4.50651348e-01 -1.59905508e-01 -9.88691449e-01 4.26922739e-01
9.45705891e-01 -6.52740061e-01 -4.61227536e-01 3.66743863e-01
4.59949285e-01 -3.21009547e-01 -4.24437910e-01 4.91905540e-01
1.72017142e-01 -9.75598872e-01 1.00688672e+00 -4.41684037e-01
3.69679421e-01 -5.94108284e-01 -5.11923373e-01 -1.10218620e+00
-3.54367614e-01 1.54718712e-01 3.79963845e-01 1.20340121e+00
6.45665750e-02 -5.42282581e-01 5.78465044e-01 4.16122735e-01
-2.20298409e-01 -6.16701126e-01 -8.31985593e-01 -5.38508594e-01
-3.61085534e-01 -5.10714591e-01 5.44352412e-01 9.50321376e-01
-4.72963095e-01 4.32029665e-01 1.81972861e-01 4.64171648e-01
6.16238236e-01 5.92283547e-01 5.17276824e-01 -1.04717839e+00
-4.10329439e-02 -7.41248369e-01 -6.20627403e-01 -7.26312220e-01
1.02328323e-01 -1.11353743e+00 2.52123743e-01 -1.65912938e+00
3.10581148e-01 -4.55030978e-01 -5.59713483e-01 7.91040480e-01
-2.72295058e-01 5.15842438e-01 4.10762221e-01 4.82166745e-02
-8.91604364e-01 4.51347947e-01 1.06922305e+00 -2.92635441e-01
-2.38892855e-03 -2.94733137e-01 -7.61290789e-01 6.61897421e-01
1.01509666e+00 -3.47408146e-01 -1.30871549e-01 -5.35503089e-01
-1.92614168e-01 -4.21284288e-01 9.99642670e-01 -1.34261215e+00
3.50804567e-01 -1.74437970e-01 5.22459090e-01 -4.97857213e-01
3.45126688e-01 -9.83932853e-01 -6.57884255e-02 4.37732577e-01
-2.10882828e-01 -3.40888590e-01 5.37646353e-01 4.55362976e-01
-3.10206980e-01 -2.86516011e-01 9.54614103e-01 -2.17464402e-01
-1.50602126e+00 2.49053001e-01 -2.29791507e-01 -1.49804965e-01
1.01542985e+00 -2.85303503e-01 -4.25849974e-01 7.64163733e-02
-4.78794396e-01 -3.68524827e-02 2.74679601e-01 7.00120270e-01
7.05938399e-01 -1.12977135e+00 -4.96927083e-01 3.86383563e-01
4.40606505e-01 2.39510730e-01 5.09735584e-01 6.55070841e-01
-3.10055375e-01 5.86692393e-01 -3.48266393e-01 -9.29125488e-01
-1.22049177e+00 4.62252527e-01 5.51939547e-01 1.95534885e-01
-3.02581638e-01 1.06254506e+00 5.56370020e-01 -3.49792302e-01
1.46649376e-01 -2.96879888e-01 -3.73285502e-01 -4.74315695e-02
4.70305681e-01 -8.29414353e-02 1.06903188e-01 -9.81502533e-01
-6.60696685e-01 7.64843106e-01 2.18144655e-02 3.52235436e-01
1.31667650e+00 -2.10315928e-01 -3.02877933e-01 4.42327619e-01
1.20398712e+00 -5.03108084e-01 -1.20553052e+00 -5.11005938e-01
4.93834401e-03 -5.98601937e-01 1.35374174e-01 -1.03088546e+00
-1.27216196e+00 9.55007315e-01 7.61314094e-01 9.76209063e-03
1.45687914e+00 3.17366630e-01 4.75086182e-01 2.74485052e-01
3.97836864e-01 -8.94592643e-01 1.02001779e-01 5.86360812e-01
6.91363037e-01 -1.55990684e+00 -1.42455727e-01 -4.08665121e-01
-5.28510809e-01 1.12292194e+00 7.40384936e-01 -1.74180657e-01
6.06512010e-01 -1.54362664e-01 1.44536927e-01 -4.91682798e-01
-2.36517698e-01 -7.64712632e-01 3.85847896e-01 4.68791991e-01
3.23015213e-01 1.29634514e-01 2.48206675e-01 4.54405248e-01
1.93378240e-01 -3.36600006e-01 1.18966669e-01 9.63665843e-01
-6.21836782e-01 -8.89002860e-01 -4.09261495e-01 3.07792932e-01
-3.47566932e-01 -2.51361132e-01 -3.64300638e-01 5.11668324e-01
4.98906612e-01 1.09826100e+00 3.17906171e-01 -1.77550986e-01
2.26211488e-01 -1.75319165e-01 5.16840160e-01 -7.05261767e-01
-5.97659886e-01 -7.68998414e-02 -1.75273761e-01 -8.14336181e-01
-7.97300339e-01 -7.18877971e-01 -1.32377410e+00 8.87966305e-02
-1.82172984e-01 -2.30092525e-01 8.99107039e-01 9.71927941e-01
3.43867421e-01 7.41495907e-01 5.17713785e-01 -9.92645621e-01
1.68629065e-01 -7.20498264e-01 -4.24101144e-01 7.24867225e-01
2.65528888e-01 -7.55128026e-01 -9.86925140e-03 1.31902859e-01] | [9.492964744567871, -0.45026350021362305] |
ee921906-1562-4cfd-b249-699ff12ebd60 | bayesian-inference-on-brain-computer | 2304.07401 | null | https://arxiv.org/abs/2304.07401v1 | https://arxiv.org/pdf/2304.07401v1.pdf | Bayesian inference on Brain-Computer Interface using the GLASS Model | The brain-computer interface (BCI) enables individuals with severe physical impairments to communicate with the world. BCIs offer computational neuroscience opportunities and challenges in converting real-time brain activities to computer commands and are typically framed as a classification problem. This article focuses on the P300 BCI that uses the event-related potential (ERP) BCI design, where the primary challenge is classifying target/non-target stimuli. We develop a novel Gaussian latent group model with sparse time-varying effects (GLASS) for making Bayesian inferences on the P300 BCI. GLASS adopts a multinomial regression framework that directly addresses the dataset imbalance in BCI applications. The prior specifications facilitate i) feature selection and noise reduction using soft-thresholding, ii) smoothing of the time-varying effects using global shrinkage, and iii) clustering of latent groups to alleviate high spatial correlations of EEG data. We develop an efficient gradient-based variational inference (GBVI) algorithm for posterior computation and provide a user-friendly Python module available at https://github.com/BangyaoZhao/GLASS. The application of GLASS identifies important EEG channels (PO8, Oz, PO7, Pz, C3) that align with existing literature. GLASS further reveals a group effect from channels in the parieto-occipital region (PO8, Oz, PO7), which is validated in cross-participant analysis. | ['Jian Kang', 'Jane E. Huggins', 'Bangyao Zhao'] | 2023-04-14 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 2.21360013e-01 -3.37895662e-01 1.60653725e-01 -7.29543790e-02
-9.58270669e-01 -1.56595990e-01 4.45453048e-01 -4.22160715e-01
-3.21466684e-01 9.97469425e-01 5.94479501e-01 -3.51244599e-01
-6.26057267e-01 -2.05982268e-01 -5.92715740e-01 -9.65194464e-01
-4.53290254e-01 -1.28829882e-01 -2.65497386e-01 3.75466138e-01
6.18173838e-01 1.39231786e-01 -1.43354797e+00 4.25772935e-01
1.17240810e+00 8.62731397e-01 5.44878840e-01 2.30364725e-01
5.14166176e-01 2.37062071e-02 -6.40159845e-01 -1.02899991e-01
-1.97579771e-01 -3.69235754e-01 -3.91274214e-01 -5.03834069e-01
-2.44242579e-01 -2.82248855e-01 -3.25953156e-01 1.03732646e+00
1.00054622e+00 9.85791013e-02 9.84778821e-01 -1.45926404e+00
-4.96913701e-01 2.72891313e-01 -7.76067853e-01 4.43178385e-01
3.55398774e-01 1.72126338e-01 5.87800086e-01 -1.08949661e+00
2.13666543e-01 9.87232506e-01 5.56529164e-01 1.97439685e-01
-1.28661025e+00 -1.04591823e+00 -5.07956706e-02 9.71035361e-01
-1.87347710e+00 -4.53852952e-01 6.30888700e-01 -6.92192912e-01
1.21274579e+00 2.79311240e-01 8.92332792e-01 1.51093400e+00
8.84699225e-01 4.75060374e-01 1.20049226e+00 -2.45314941e-01
3.76939356e-01 -2.39343166e-01 4.48000073e-01 -3.27931702e-01
1.14650778e-01 9.35959592e-02 -1.11760986e+00 -4.78527337e-01
6.03920102e-01 4.98299114e-02 -7.74240494e-01 1.96471557e-01
-1.20916057e+00 6.65163577e-01 -1.22865327e-01 2.05411077e-01
-1.01335371e+00 9.44270715e-02 9.92851257e-02 -8.97467658e-02
5.58051169e-01 -4.66877408e-02 -5.31063914e-01 -6.08612776e-01
-1.18238783e+00 9.56141353e-02 3.07130963e-01 7.42859662e-01
2.59368092e-01 -6.97337165e-02 -7.17107296e-01 1.09962225e+00
5.23845971e-01 5.80849767e-01 6.18155658e-01 -7.37221837e-01
3.61168325e-01 -7.50289187e-02 -6.89067319e-02 -6.45414770e-01
-3.60201508e-01 -6.39481619e-02 -7.39735782e-01 8.79840702e-02
1.26550034e-01 -4.71769691e-01 -4.78292435e-01 1.84645700e+00
-1.64754927e-01 3.59167278e-01 -4.66528505e-01 6.88678265e-01
4.17340875e-01 4.31609482e-01 1.99262276e-01 -5.70800662e-01
1.53757370e+00 -1.89947769e-01 -8.84515643e-01 -1.37315109e-01
1.28542185e-01 -3.36468548e-01 9.78650689e-01 7.25304365e-01
-1.02160811e+00 -2.21164301e-01 -7.23256826e-01 1.94040000e-01
1.76024009e-02 2.48571426e-01 5.90053082e-01 6.91266119e-01
-1.19024706e+00 3.02317053e-01 -1.04653323e+00 -6.71121180e-02
7.64219403e-01 5.50011218e-01 -4.34454471e-01 1.05444759e-01
-1.11521673e+00 9.56041157e-01 -3.48534659e-02 2.23081589e-01
-9.29656148e-01 -1.00509703e+00 -3.34208459e-01 4.04429287e-02
-1.21139042e-01 -4.38706547e-01 5.52927792e-01 -6.22396111e-01
-1.68896806e+00 4.37093109e-01 -7.10847974e-01 1.25410959e-01
-5.37700318e-02 -2.63253063e-01 -3.12462181e-01 1.08373806e-01
2.12992877e-01 5.45303702e-01 9.51452792e-01 -5.57246268e-01
-3.61523360e-01 -8.96827400e-01 -7.52572596e-01 1.59942627e-01
-2.44427100e-01 5.94535649e-01 -1.88610315e-01 -7.14251220e-01
1.45331427e-01 -6.01920664e-01 3.30541581e-01 -5.23148119e-01
-3.96531701e-01 -5.03423452e-01 1.69446066e-01 -1.42341375e+00
1.13198173e+00 -2.24771857e+00 1.22707598e-01 3.32375497e-01
3.40814412e-01 -5.03151476e-01 5.54609448e-02 1.77955657e-01
-6.29900098e-01 -1.55909821e-01 -3.05128068e-01 -1.40221477e-01
7.48013109e-02 -3.83755773e-01 -1.25304416e-01 8.58863175e-01
6.03694059e-02 1.06497109e+00 -3.38032603e-01 7.71848112e-02
-7.44051337e-02 7.53653526e-01 -6.94843233e-01 -2.18524062e-03
6.26859725e-01 7.07831383e-01 2.61945277e-01 4.66958702e-01
9.01938856e-01 1.31503388e-01 -1.33603767e-01 -5.79024702e-02
-5.16227424e-01 5.55334747e-01 -1.04547477e+00 1.53986371e+00
1.10863999e-01 8.56877208e-01 2.38390341e-01 -1.34041262e+00
5.62438428e-01 6.26452267e-01 4.94776547e-01 -6.16587162e-01
2.54580051e-01 1.29885629e-01 4.24654037e-01 -5.62176287e-01
-2.80886859e-01 1.16650969e-01 3.12247455e-01 3.88110697e-01
1.91996142e-01 4.69609350e-02 -1.92323998e-01 -1.51734706e-02
1.15920842e+00 -1.03664277e-02 2.13596389e-01 -7.06650674e-01
-3.49539109e-02 -6.01513922e-01 6.10198259e-01 7.01219559e-01
-1.13482058e-01 6.76209867e-01 7.16855109e-01 5.65956950e-01
-3.68874431e-01 -1.30080879e+00 -6.09980047e-01 7.39052057e-01
-2.94011056e-01 -2.05155313e-01 -6.75152004e-01 2.84191072e-01
-1.19047582e-01 9.56894159e-01 -4.21828806e-01 -3.74221653e-01
1.06500639e-02 -1.08206499e+00 1.89351872e-01 4.74544138e-01
2.28245914e-01 -1.19465506e+00 -6.27358913e-01 2.54319191e-01
-4.42258358e-01 -6.05526865e-01 -1.70876920e-01 3.17377776e-01
-8.04209530e-01 -7.58913815e-01 -9.00302410e-01 -5.05935431e-01
4.67490882e-01 -7.41977431e-03 4.45291817e-01 -8.23053062e-01
-4.50914204e-01 6.70821309e-01 -1.89398959e-01 -7.05909967e-01
6.22686267e-01 -3.93437833e-01 2.99044043e-01 -4.74515632e-02
8.55036974e-01 -1.14340472e+00 -7.54775167e-01 4.81673405e-02
-3.05282116e-01 8.63433927e-02 5.11548996e-01 6.92551613e-01
3.87633324e-01 7.16723800e-02 7.22460449e-01 1.06381625e-02
1.06947494e+00 -8.68101537e-01 -3.59729916e-01 6.73263744e-02
-3.96944731e-01 -5.48007607e-01 -9.21638384e-02 -8.07182610e-01
-1.22682023e+00 -5.25050700e-01 1.06042437e-02 -1.53836366e-02
-3.53581667e-01 6.76281214e-01 -5.93511343e-01 2.06292629e-01
4.87847298e-01 2.67706096e-01 -1.09955557e-01 -2.93617517e-01
-8.15948024e-02 1.08189929e+00 5.81829906e-01 -4.58445281e-01
1.63896824e-03 3.24651152e-01 -5.54832220e-01 -1.08989084e+00
1.16693802e-01 -3.21990639e-01 -6.55505180e-01 -2.35505864e-01
8.87853682e-01 -1.03834081e+00 -8.21877360e-01 8.90986323e-01
-1.21369374e+00 -5.69524825e-01 2.32422039e-01 1.23445451e+00
-6.14747643e-01 1.08699717e-01 -4.52298284e-01 -9.25475776e-01
-5.61230004e-01 -1.13785326e+00 7.46627629e-01 6.02748282e-02
-5.70868313e-01 -4.24519062e-01 -5.82408160e-03 1.65298983e-01
2.89888740e-01 -1.17677137e-01 9.16135669e-01 -3.72673720e-01
-7.24493489e-02 -5.47130853e-02 -2.44104683e-01 2.73947686e-01
4.63127252e-03 -2.56399244e-01 -9.45277691e-01 -1.61329508e-01
5.05984366e-01 2.48014778e-01 2.50894397e-01 1.32960296e+00
1.34056497e+00 5.88022768e-02 -4.53136653e-01 6.56070352e-01
8.02303791e-01 4.91323888e-01 1.16078663e+00 -1.94129512e-01
4.71214503e-01 6.50549769e-01 -1.20217957e-01 6.29859507e-01
3.36129963e-01 7.02117443e-01 -5.19689247e-02 3.95599037e-01
1.34040907e-01 2.40642816e-01 6.04133070e-01 5.76779544e-01
-1.74828604e-01 2.37322465e-01 -9.18339074e-01 7.13596463e-01
-1.55909193e+00 -8.33997428e-01 -5.41499794e-01 2.51384234e+00
7.87244618e-01 -2.46230736e-01 -2.05339104e-01 1.81100592e-01
7.25455403e-01 -7.35675752e-01 -5.66605687e-01 1.02992296e-01
-9.18293446e-02 6.91905200e-01 2.82541633e-01 3.12307060e-01
-4.64387476e-01 6.06130242e-01 5.87436438e+00 8.49947691e-01
-1.02089167e+00 6.80964768e-01 4.24050748e-01 -5.41064262e-01
-6.89571947e-02 -1.36908919e-01 -5.57450652e-01 9.24078166e-01
1.33991194e+00 -3.25846344e-01 8.86984408e-01 1.82205677e-01
9.22627211e-01 -6.69463456e-01 -9.68656301e-01 1.60036492e+00
1.29131060e-02 -8.17693293e-01 -5.87743521e-01 3.83355916e-01
2.67475933e-01 3.14263731e-01 -1.48661006e-02 1.40875638e-01
-2.95862496e-01 -1.13391054e+00 9.05369401e-01 9.05699313e-01
1.17230475e+00 -6.24149382e-01 3.57182354e-01 2.37969682e-01
-7.50900686e-01 -1.69511229e-01 -3.06142449e-01 -2.65675873e-01
2.52522022e-01 7.51738846e-01 -9.04300585e-02 7.22729191e-02
1.28325868e+00 6.88217998e-01 -1.74425259e-01 1.28395903e+00
-6.47208750e-01 1.02772844e+00 -3.50963622e-01 2.30728984e-02
-6.06171846e-01 -4.18152779e-01 7.09361970e-01 8.79861951e-01
8.47822368e-01 5.18351018e-01 -9.15339708e-01 1.41278827e+00
5.04622936e-01 -9.73455161e-02 -3.69365096e-01 1.85785398e-01
6.91981852e-01 1.05346668e+00 -4.90069538e-01 1.23612307e-01
-4.88683134e-01 1.01572478e+00 1.00711986e-01 8.97036314e-01
-6.53940558e-01 -6.50941968e-01 4.27861273e-01 -1.16213396e-01
-3.93227004e-02 -2.97528267e-01 -7.71083355e-01 -1.10678720e+00
3.14232379e-01 -8.07880938e-01 1.02157079e-01 -1.21784735e+00
-1.24692917e+00 1.61836669e-01 5.52264035e-01 -7.73448110e-01
-1.31527424e-01 -7.86314905e-01 -8.62365484e-01 1.54033017e+00
-9.25575495e-01 -5.39594412e-01 -2.29980469e-01 1.18758440e+00
5.48750162e-02 -2.23644208e-02 8.23639750e-01 4.67314005e-01
-8.08042288e-01 4.20917392e-01 2.21488997e-01 -3.32614958e-01
7.20297277e-01 -8.11476111e-01 -2.06138909e-01 9.58002865e-01
-3.79011661e-01 1.06513739e+00 3.57928395e-01 -7.55379021e-01
-1.02597690e+00 -3.17709655e-01 7.42951095e-01 -2.60403156e-01
5.66937745e-01 -7.05305040e-01 -8.11544895e-01 4.98709887e-01
1.78586960e-01 -5.88509560e-01 9.85225081e-01 9.18394551e-02
2.58343279e-01 6.60929084e-02 -1.00598621e+00 7.29414403e-01
1.03612459e+00 -7.15924025e-01 -6.79459810e-01 2.93077052e-01
1.29091842e-02 2.29886800e-01 -8.72958541e-01 2.24982187e-01
7.11725771e-01 -5.75810075e-01 7.44470656e-01 -8.90939832e-02
4.24017794e-02 8.19694623e-02 5.88309877e-02 -1.54751325e+00
-5.67647994e-01 -8.38089645e-01 -2.90490454e-03 1.14739895e+00
1.83886945e-01 -1.03339112e+00 2.62124866e-01 1.18978298e+00
-2.90469646e-01 -4.22935009e-01 -1.30339241e+00 -4.47258979e-01
1.92048714e-01 -1.03433681e+00 2.34583378e-01 4.01947916e-01
9.04455066e-01 1.22501869e-02 -1.05258569e-01 2.24815130e-01
5.42742550e-01 -5.28764307e-01 1.10851824e-01 -1.27976561e+00
-3.83659899e-02 -6.31312490e-01 -2.41477966e-01 -8.07380676e-01
2.07831413e-01 -9.47641373e-01 2.05732629e-01 -1.59393573e+00
6.15485430e-01 7.32773915e-02 -2.94426918e-01 5.75535774e-01
-2.97112793e-01 5.91321662e-02 -4.87815112e-01 2.60250688e-01
1.44776404e-01 7.72922993e-01 4.73599404e-01 -2.81243064e-02
-4.15708244e-01 7.84074068e-02 -9.21238780e-01 5.03027380e-01
9.77567017e-01 -6.29469216e-01 -3.77214283e-01 -2.18493387e-01
-4.55255844e-02 -1.91823058e-02 4.05689687e-01 -8.28239143e-01
3.11914831e-01 6.80032372e-02 6.51119232e-01 -5.67277849e-01
3.84847552e-01 -4.13854003e-01 2.01827422e-01 1.94614246e-01
1.72018558e-01 -1.94896862e-01 4.04803962e-01 2.22515076e-01
2.36290954e-02 -2.36824840e-01 5.09044826e-01 3.69215727e-01
-1.56060636e-01 2.14975417e-01 -1.23148406e+00 -1.09351330e-01
8.30597520e-01 -4.59758073e-01 -1.97247684e-01 -3.91727865e-01
-8.05087566e-01 2.06185672e-02 -3.87519270e-01 3.12494397e-01
7.31702864e-01 -1.25461197e+00 -8.52499843e-01 5.38220584e-01
-8.88974369e-02 -4.70878661e-01 6.53797328e-01 1.60331166e+00
8.86382535e-02 5.11365175e-01 -5.78883767e-01 -5.28657913e-01
-1.11741364e+00 -4.12933640e-02 1.70576513e-01 4.11563486e-01
-6.27240062e-01 1.26990843e+00 3.57123971e-01 1.24434121e-01
3.51694018e-01 -2.14655668e-01 -3.57425153e-01 3.98035765e-01
7.10639954e-01 7.66995490e-01 2.01390192e-01 -4.32041824e-01
-6.48631454e-01 1.48163298e-02 3.89713585e-01 -7.02985644e-01
1.68402421e+00 -1.65916383e-01 -7.28698492e-01 4.72193867e-01
1.06568134e+00 -1.24617882e-01 -1.32214761e+00 3.35709840e-01
-1.50581047e-01 -2.62536258e-01 4.35036838e-01 -8.23143303e-01
-7.82159388e-01 1.04706717e+00 1.01462793e+00 -4.09675062e-01
1.13796616e+00 -8.53353888e-02 1.24866419e-01 -1.01202279e-01
5.09391785e-01 -1.22016525e+00 -2.94142485e-01 2.86928087e-01
1.35704422e+00 -5.15146792e-01 -1.04597948e-01 4.18213680e-02
-5.11954963e-01 8.12772393e-01 1.92014411e-01 -1.06287785e-01
1.15205681e+00 3.33416939e-01 -6.27475321e-01 -2.84890682e-01
-4.96087462e-01 1.96592316e-01 6.28615141e-01 7.47208476e-01
4.79056239e-01 3.05760115e-01 -9.56539333e-01 1.71260142e+00
-3.37612301e-01 1.52535364e-01 2.52546132e-01 5.90124607e-01
7.95243531e-02 -6.85901403e-01 -6.79934323e-01 9.99438643e-01
-5.17956793e-01 -6.30787551e-01 1.00927435e-01 4.50938225e-01
1.15186721e-01 1.43379343e+00 1.80158809e-01 -1.29099369e-01
1.51872948e-01 5.04147291e-01 4.16970402e-01 -5.63583374e-01
-1.68729395e-01 5.15184641e-01 -3.26030314e-01 -6.73009932e-01
-3.85021162e-03 -1.26628006e+00 -1.07154834e+00 1.38830706e-01
-2.92984098e-01 6.79947957e-02 9.55980539e-01 9.24077213e-01
7.70042002e-01 6.24928474e-01 -3.85594442e-02 -1.22952461e+00
-1.23160057e-01 -1.64300597e+00 -9.33685541e-01 2.59477999e-02
-7.67638758e-02 -1.05458236e+00 -6.08399868e-01 4.34056595e-02] | [13.096738815307617, 3.3481109142303467] |
4affc42d-c30e-4d40-aa72-7c945e917e34 | dsvo-direct-stereo-visual-odometry | 1810.03963 | null | https://arxiv.org/abs/1810.03963v2 | https://arxiv.org/pdf/1810.03963v2.pdf | DSVO: Direct Stereo Visual Odometry | This paper proposes a novel approach to stereo visual odometry without stereo matching. It is particularly robust in scenes of repetitive high-frequency textures. Referred to as DSVO (Direct Stereo Visual Odometry), it operates directly on pixel intensities, without any explicit feature matching, and is thus efficient and more accurate than the state-of-the-art stereo-matching-based methods. It applies a semi-direct monocular visual odometry running on one camera of the stereo pair, tracking the camera pose and mapping the environment simultaneously; the other camera is used to optimize the scale of monocular visual odometry. We evaluate DSVO in a number of challenging scenes to evaluate its performance and present comparisons with the state-of-the-art stereo visual odometry algorithms. | ['Junaed Sattar', 'Jiawei Mo'] | 2018-09-19 | null | null | null | null | ['stereo-matching', 'monocular-visual-odometry'] | ['computer-vision', 'robots'] | [-2.44516926e-03 -1.61905289e-01 -5.27481325e-02 -3.39035764e-02
-5.91243319e-02 -3.87296289e-01 8.57675731e-01 -2.95358688e-01
-4.83384818e-01 5.74689686e-01 7.24058002e-02 1.37189473e-03
2.72517860e-01 -5.30493796e-01 -6.35598958e-01 -6.22500837e-01
3.39092195e-01 1.11095953e+00 7.80243278e-01 -3.39915842e-01
6.10115111e-01 5.26714385e-01 -1.83097887e+00 -4.93837059e-01
6.20744407e-01 7.17830777e-01 6.62781358e-01 8.14107656e-01
2.01643616e-01 8.10536504e-01 -1.31776705e-01 1.40051514e-01
4.73589182e-01 -3.33578110e-01 -7.19138741e-01 3.71817321e-01
1.01775885e+00 -3.47768307e-01 -7.15162933e-01 1.26879346e+00
6.01232171e-01 -6.11362280e-03 4.57006365e-01 -8.13452244e-01
-3.87793422e-01 -5.91672421e-01 -4.53126252e-01 2.49802649e-01
1.11396551e+00 2.02733293e-01 7.45313525e-01 -8.35333169e-01
1.34998798e+00 1.34369981e+00 6.08108640e-01 2.07153812e-01
-1.63762033e+00 -2.81386793e-01 -3.73938888e-01 1.81785941e-01
-1.38400614e+00 -3.97346050e-01 4.98793036e-01 -7.45365798e-01
1.00854254e+00 -8.24766979e-03 8.79402518e-01 5.07155418e-01
5.71550965e-01 1.43646389e-01 1.35337853e+00 -6.23247206e-01
1.60210669e-01 -4.36940640e-02 -2.22730309e-01 9.98701751e-01
3.74260724e-01 8.08221459e-01 -8.91051769e-01 -1.18878044e-01
1.10354650e+00 -5.20346733e-03 -3.54279041e-01 -1.31144702e+00
-1.37356389e+00 6.99626684e-01 5.54158032e-01 -1.66465551e-01
-1.83131024e-01 1.07552230e-01 3.88551317e-02 4.72026020e-01
2.53568411e-01 3.83856803e-01 6.94112033e-02 -1.03253417e-01
-8.43181074e-01 4.52573448e-01 6.46521688e-01 9.81942356e-01
1.28769565e+00 1.87226713e-01 3.49512726e-01 5.95923960e-01
5.89509606e-01 8.36953104e-01 5.05235493e-01 -1.29853904e+00
2.61280030e-01 4.80078340e-01 1.64520219e-01 -7.99694300e-01
-4.86138523e-01 1.21011525e-01 -4.38144952e-01 9.01283026e-01
2.95027614e-01 1.70361161e-01 -1.10614204e+00 1.08396065e+00
5.50038517e-01 7.36832991e-02 2.68453211e-02 1.19147778e+00
6.93589926e-01 3.55416954e-01 -7.22222865e-01 2.14481391e-02
1.15653276e+00 -9.24126863e-01 -6.97656155e-01 -7.01041043e-01
2.44916946e-01 -1.19140220e+00 6.12471759e-01 2.36006677e-01
-9.11803424e-01 -6.14883602e-01 -1.45437551e+00 -2.99396187e-01
-2.31086969e-01 -3.81398290e-01 3.65025818e-01 5.55048943e-01
-1.30562186e+00 4.00769234e-01 -6.12990439e-01 -6.54385090e-01
-4.52531904e-01 4.53353673e-01 -4.41060364e-01 -1.47421569e-01
-8.86176825e-01 1.31132972e+00 2.44789734e-01 -4.52314258e-01
-1.04900551e+00 -2.72509634e-01 -1.31896639e+00 -4.69880283e-01
3.75407711e-02 -1.05137980e+00 1.24434197e+00 -6.87817693e-01
-1.87388849e+00 1.67103124e+00 -6.78524792e-01 -2.59799600e-01
7.39261210e-01 -1.21301822e-01 -3.11058253e-01 4.17379141e-02
3.62022847e-01 8.09423327e-01 8.79679680e-01 -1.41105187e+00
-6.97035849e-01 -6.35097325e-01 -1.05869835e-02 6.13227129e-01
6.27229691e-01 -3.18310678e-01 -6.47690594e-01 -1.13055436e-03
8.98115516e-01 -1.30133665e+00 -1.51158720e-01 1.36319011e-01
-2.55088478e-01 5.95687568e-01 8.13332319e-01 -2.55899608e-01
7.21621931e-01 -2.01321149e+00 3.11844170e-01 -2.62208807e-04
8.45457092e-02 -6.25568777e-02 2.91025877e-01 3.08475018e-01
7.80666247e-02 -1.01148963e+00 1.76131800e-01 -4.19421345e-01
-1.40122220e-01 3.84401977e-01 -1.90892756e-01 1.04825795e+00
-3.97485882e-01 6.26834452e-01 -9.58242714e-01 -3.94577771e-01
7.39632666e-01 3.35312277e-01 -4.22262698e-01 3.18096429e-01
1.48114517e-01 7.60766864e-01 1.20624952e-01 6.55802369e-01
8.98876190e-01 -1.67464372e-02 2.71205842e-01 4.19453904e-02
-6.18884027e-01 4.92133260e-01 -1.72959137e+00 1.75185144e+00
-3.97432268e-01 1.02389956e+00 -7.16887563e-02 -3.46401453e-01
1.20652759e+00 -1.44013632e-02 1.92181692e-01 -1.00840139e+00
3.89129817e-02 5.24337769e-01 -4.59275305e-01 -2.10406840e-01
9.01125491e-01 -3.07674378e-01 1.10980645e-01 1.30215317e-01
2.35049635e-01 -1.05608821e+00 1.88433334e-01 -1.04932398e-01
6.95364475e-01 4.00574595e-01 8.83034885e-01 -7.02072978e-01
8.66328955e-01 3.72603238e-01 3.21307421e-01 4.54640687e-01
-2.70974100e-01 8.05355251e-01 -1.21267803e-01 -7.01499164e-01
-1.29848850e+00 -1.59864295e+00 -3.42587948e-01 4.54591393e-01
1.04826593e+00 -9.84532088e-02 -3.88307989e-01 1.57443792e-01
3.62378448e-01 -6.97254241e-02 -4.79540676e-01 2.40501598e-01
-3.06129664e-01 -2.56228030e-01 8.76201391e-02 1.81203172e-01
5.76370358e-01 -5.34274995e-01 -6.95503950e-01 2.60125577e-01
-3.14823277e-02 -1.13462579e+00 -4.72198486e-01 2.87077606e-01
-1.08611119e+00 -1.24130273e+00 -7.62982011e-01 -8.11307251e-01
5.81901371e-01 8.63687336e-01 1.16029584e+00 -4.27176356e-01
-2.06340373e-01 1.15853801e-01 1.75292090e-01 -2.42739290e-01
-3.28114569e-01 -1.68521076e-01 2.43642986e-01 -1.93608344e-01
5.21107197e-01 -5.34519970e-01 -4.10951078e-01 8.11169207e-01
-4.01734114e-01 2.58441549e-02 -5.92986569e-02 8.92055631e-01
6.87252045e-01 -6.32196307e-01 -8.05348873e-01 -5.65160513e-01
7.84553215e-02 1.08730465e-01 -1.48210347e+00 -4.64228302e-01
-6.44474149e-01 3.91246915e-01 2.01483950e-01 3.39468047e-02
-8.99173617e-01 4.74257737e-01 2.67317951e-01 -4.14096177e-01
3.91671620e-02 -6.75873756e-02 2.79364824e-01 -6.22844458e-01
9.57976818e-01 3.36032867e-01 1.80434376e-01 -2.54694521e-01
-5.52433170e-02 5.91603160e-01 9.75887418e-01 4.56997678e-02
1.12768090e+00 1.35855460e+00 4.83544230e-01 -1.06399846e+00
-6.08747661e-01 -1.19834769e+00 -1.24052620e+00 -3.58633250e-01
9.61082697e-01 -1.28888643e+00 -6.83895350e-01 7.24751532e-01
-1.13369596e+00 -2.94363230e-01 -9.19701234e-02 7.36154377e-01
-1.04289472e+00 4.28817034e-01 -4.02213693e-01 -5.18148541e-01
2.08288357e-01 -1.36933899e+00 1.51768446e+00 1.57773405e-01
-3.12966645e-01 -1.29482853e+00 1.06750202e+00 6.48182631e-02
-1.41727716e-01 1.22796491e-01 -1.21760638e-02 5.57651877e-01
-9.87644672e-01 5.49141094e-02 -1.80957377e-01 -2.14575902e-01
8.26831348e-03 -1.91600308e-01 -1.19925642e+00 -2.98246264e-01
-1.96765363e-01 -2.79267058e-02 6.71541870e-01 4.25477773e-01
-3.10083833e-02 5.16184151e-01 -2.60915428e-01 1.11007106e+00
1.77489650e+00 9.36181471e-02 9.98560011e-01 1.04594338e+00
7.47216344e-01 4.22677934e-01 8.35314691e-01 4.51233298e-01
5.18592179e-01 1.44040465e+00 6.58104479e-01 -2.08180398e-01
-3.10338736e-02 -4.60132182e-01 4.35855359e-01 6.65374100e-01
-3.01927358e-01 4.11584347e-01 -8.61920953e-01 4.96534616e-01
-1.74874532e+00 -8.70620131e-01 -4.73776609e-01 2.34067297e+00
3.28962445e-01 -1.35997266e-01 3.91960330e-02 2.21163467e-01
7.35472500e-01 5.86438000e-01 -3.41982663e-01 -6.88481033e-01
-3.00951272e-01 -1.07338458e-01 1.15755820e+00 9.77788091e-01
-1.03742874e+00 1.05932522e+00 7.37330055e+00 -4.15592752e-02
-1.20211637e+00 2.07292303e-01 -4.65543717e-01 2.34476417e-01
-4.70764283e-03 2.86951989e-01 -1.08277392e+00 2.30412021e-01
3.94063324e-01 -2.27333922e-02 7.78033435e-01 9.29643989e-01
2.15914637e-01 -5.57998300e-01 -9.06430364e-01 1.38457263e+00
3.39635223e-01 -1.28971636e+00 -2.33614340e-01 3.72699201e-01
1.21323574e+00 5.95174253e-01 -2.64136106e-01 -3.78958166e-01
7.10605025e-01 -5.42787969e-01 1.01954305e+00 2.58323580e-01
7.19458640e-01 -2.93701231e-01 6.34194851e-01 2.57227421e-01
-1.49033356e+00 1.79400891e-01 -6.02802038e-01 -5.36472082e-01
5.26912510e-01 4.39737886e-01 -6.39591038e-01 5.33524632e-01
8.81680071e-01 9.19099450e-01 -4.00645077e-01 1.02919137e+00
-4.64155316e-01 -5.88858366e-01 -1.67119578e-01 4.15548056e-01
4.29220319e-01 -4.41125870e-01 6.80774570e-01 9.06073809e-01
4.27605301e-01 -4.45096731e-01 2.30721056e-01 4.46758300e-01
4.22615051e-01 -1.75903946e-01 -1.10475242e+00 7.17955828e-01
1.89784542e-01 5.84940553e-01 -8.41223747e-02 -4.83642191e-01
-3.69183451e-01 1.35169840e+00 2.20904112e-01 5.96333034e-02
-3.30267072e-01 -1.33125573e-01 1.17592251e+00 3.13646048e-01
3.40047657e-01 -4.94344682e-01 -2.32902765e-01 -1.39415944e+00
-2.71857232e-01 -4.32319880e-01 3.76816034e-01 -1.34429479e+00
-7.52880454e-01 2.97129333e-01 -2.19316661e-01 -1.77100408e+00
-6.32496834e-01 -9.52168345e-01 -2.64423162e-01 1.06129587e+00
-1.79772830e+00 -5.65690875e-01 -9.12704587e-01 9.26882923e-01
2.13899359e-01 2.20000017e-02 6.67187452e-01 1.58396363e-01
3.12023610e-01 -2.25783333e-01 4.51650351e-01 -2.91264147e-01
9.15145993e-01 -1.44877446e+00 7.81448781e-01 9.01219845e-01
7.43982894e-03 2.43718088e-01 1.11970782e+00 -4.58997309e-01
-1.54231811e+00 -6.49185836e-01 1.28972685e+00 -6.69996977e-01
5.78448355e-01 -2.86795169e-01 -4.21294242e-01 8.48032296e-01
-5.06145880e-02 1.03594735e-01 -2.23722085e-01 -2.79265195e-01
-1.68671355e-01 -8.71661976e-02 -1.01844609e+00 6.93217397e-01
1.25961363e+00 -1.04353046e+00 -8.37388754e-01 9.02947187e-02
1.68411896e-01 -1.14767230e+00 -5.30313551e-01 8.28453451e-02
7.68209040e-01 -1.56106043e+00 1.21868443e+00 2.04011023e-01
-3.72457504e-03 -7.86908209e-01 -2.61744738e-01 -1.33179045e+00
-4.37486380e-01 -6.65912986e-01 2.77040154e-01 4.01426435e-01
-1.79923654e-01 -9.84043777e-01 8.32046807e-01 -4.19904828e-01
-9.53866094e-02 3.61784607e-01 -1.12318659e+00 -1.13467968e+00
-7.14523911e-01 -1.33775711e-01 2.94170052e-01 9.02554095e-01
2.08100289e-01 2.42198452e-01 -5.90753317e-01 4.38776582e-01
9.12506878e-01 2.38941610e-01 1.49217701e+00 -1.41481149e+00
-3.82512927e-01 -8.07642180e-04 -1.28567386e+00 -1.53387320e+00
-7.37791462e-03 -5.07708192e-01 2.30411604e-01 -1.21352875e+00
-3.18796225e-02 2.03209281e-01 4.13236290e-01 -3.30179006e-01
1.41804561e-01 6.18359089e-01 1.09873101e-01 5.23156285e-01
-1.66480690e-01 2.47228459e-01 1.50897837e+00 -2.65636574e-02
-2.20873207e-01 -3.04924726e-01 3.01156074e-01 7.93018997e-01
2.15368435e-01 -5.84854126e-01 -1.06107771e-01 -2.58101553e-01
7.81460851e-02 6.92247367e-03 4.97660398e-01 -1.30498779e+00
3.10179144e-01 -1.54386535e-01 6.35446310e-02 -7.43789613e-01
5.92206895e-01 -7.31740415e-01 6.18256211e-01 9.78959799e-01
4.21997547e-01 3.05869639e-01 -1.55827150e-01 3.39352816e-01
-5.60011804e-01 -1.81632802e-01 1.32634163e+00 -2.79388875e-01
-1.44717848e+00 4.53166515e-02 -4.15023953e-01 -3.63130644e-02
8.62331927e-01 -8.58334959e-01 -3.90238643e-01 -3.96090746e-01
-6.43596888e-01 -1.51157185e-01 1.46994472e+00 2.99320430e-01
3.72957021e-01 -1.48122358e+00 -1.25487864e-01 7.75251567e-01
6.87043667e-01 -2.59174526e-01 4.55322443e-03 8.71060789e-01
-1.37240422e+00 8.48011255e-01 -6.02817714e-01 -1.30481112e+00
-1.44360936e+00 5.24211884e-01 7.13203490e-01 -2.39058092e-01
-5.80747366e-01 2.15756118e-01 4.44338381e-01 -7.47888267e-01
-1.31913394e-01 -1.60689071e-01 2.13407520e-02 -2.70624101e-01
5.00580251e-01 6.23543203e-01 7.29446411e-02 -1.03920209e+00
-4.71684366e-01 1.29206014e+00 6.61963582e-01 -5.46618581e-01
7.51745284e-01 -7.25898206e-01 3.44007611e-02 4.99706209e-01
1.02242279e+00 -6.91269785e-02 -1.52923560e+00 -3.53792429e-01
-1.90042078e-01 -1.07495224e+00 2.20691800e-01 2.65321676e-02
-6.17691696e-01 9.31500852e-01 8.36021304e-01 -1.20223925e-01
7.32216716e-01 -1.42323881e-01 3.29807103e-01 2.92116821e-01
9.56752539e-01 -1.15127814e+00 -1.02254907e-02 1.11142111e+00
4.89677817e-01 -1.37441623e+00 3.29110950e-01 -5.91433048e-01
-4.66587782e-01 1.09164083e+00 3.49255890e-01 -3.22408199e-01
5.35673678e-01 2.90145218e-01 5.77529848e-01 -2.91628808e-01
-3.09784383e-01 -6.06388092e-01 1.81845471e-01 9.27060544e-01
9.01288986e-02 -1.82010666e-01 2.38830060e-01 -1.18936968e+00
-1.18562907e-01 1.62135869e-01 5.89311063e-01 1.07261598e+00
-7.68169820e-01 -1.01195896e+00 -9.12122011e-01 -3.37511867e-01
1.00804612e-01 2.13246532e-02 -3.96965951e-01 8.29008460e-01
-4.35835496e-02 5.11651397e-01 3.05416167e-01 -2.83767194e-01
6.68107808e-01 -3.05515260e-01 7.68087387e-01 -5.99017382e-01
-2.32096314e-01 -6.38854429e-02 1.64258555e-01 -9.51389551e-01
-7.17570364e-01 -9.33525741e-01 -9.93279815e-01 -6.62962317e-01
-6.94982708e-02 -3.87125105e-01 8.25541019e-01 7.74479747e-01
-1.27742393e-02 -1.14202246e-01 6.00352407e-01 -1.63995755e+00
-1.74992710e-01 -4.82080877e-01 -1.07879078e+00 4.52042490e-01
6.47531569e-01 -1.14236367e+00 -5.28234899e-01 1.50844902e-01] | [7.809682369232178, -2.219496011734009] |
47a9ae1b-de4d-4eb7-9b9b-08e60045ec6d | a-lightweight-hybrid-cnn-lstm-model-for-ecg | 2209.00988 | null | https://arxiv.org/abs/2209.00988v1 | https://arxiv.org/pdf/2209.00988v1.pdf | A lightweight hybrid CNN-LSTM model for ECG-based arrhythmia detection | Electrocardiogram (ECG) is the most frequent and routine diagnostic tool used for monitoring heart electrical signals and evaluating its functionality. The human heart can suffer from a variety of diseases, including cardiac arrhythmias. Arrhythmia is an irregular heart rhythm that in severe cases can lead to heart stroke and can be diagnosed via ECG recordings. Since early detection of cardiac arrhythmias is of great importance, computerized and automated classification and identification of these abnormal heart signals have received much attention for the past decades. Methods: This paper introduces a light deep learning approach for high accuracy detection of 8 different cardiac arrhythmias and normal rhythm. To leverage deep learning method, resampling and baseline wander removal techniques are applied to ECG signals. In this study, 500 sample ECG segments were used as model inputs. The rhythm classification was done by an 11-layer network in an end-to-end manner without the need for hand-crafted manual feature extraction. Results: In order to evaluate the proposed technique, ECG signals are chosen from the two physionet databases, the MIT-BIH arrhythmia database and the long-term AF database. The proposed deep learning framework based on the combination of Convolutional Neural Network(CNN) and Long Short Term Memory (LSTM) showed promising results than most of the state-of-the-art methods. The proposed method reaches the mean diagnostic accuracy of 98.24%. Conclusion: A trained model for arrhythmia classification using diverse ECG signals were successfully developed and tested. Significance: Since the present work uses a light classification technique with high diagnostic accuracy compared to other notable methods, it could successfully be implemented in holter monitor devices for arrhythmia detection. | ['Fahimeh Nasimi', 'Nima Alamatsaz', 'Hamidreza Payan', 'Mohammadreza Yazdchi', 'Leyla s Tabatabaei', 'Negin Alamatsaz'] | 2022-08-29 | null | null | null | null | ['arrhythmia-detection'] | ['medical'] | [ 4.15744781e-01 -3.97810072e-01 1.64757431e-01 -1.40070423e-01
-6.61004543e-01 -4.22948778e-01 1.09878760e-02 2.39724621e-01
-4.09248561e-01 8.78845096e-01 -3.96244735e-01 -4.81322318e-01
-3.78791034e-01 -4.92814183e-01 -1.56638607e-01 -7.52148449e-01
-4.65784043e-01 3.49494606e-01 -4.36506450e-01 1.74887791e-01
1.02809295e-01 7.65430331e-01 -1.11411393e+00 3.73600841e-01
6.04496062e-01 1.29214764e+00 -2.57818937e-01 1.08313870e+00
3.96098137e-01 5.39479136e-01 -9.81689155e-01 1.06023908e-01
9.93028060e-02 -8.31470311e-01 -4.41674978e-01 -2.73497760e-01
-5.06758280e-02 -1.33870766e-01 1.35980129e-01 4.36853290e-01
1.54561794e+00 -3.86418879e-01 4.66552466e-01 -7.15984643e-01
6.74660727e-02 3.73963058e-01 -3.54388535e-01 7.71279752e-01
1.79524124e-01 1.34041220e-01 3.65329593e-01 -8.45426142e-01
1.91284448e-01 2.73312837e-01 1.36049867e+00 1.40148059e-01
-1.10366035e+00 -8.02206933e-01 -8.00740063e-01 2.44185284e-01
-1.38095438e+00 -2.49779001e-01 9.70340908e-01 -6.09554946e-01
9.90071416e-01 2.96336770e-01 1.13698900e+00 1.03915823e+00
6.16939902e-01 2.38361850e-01 9.89978909e-01 -4.47122127e-01
1.54266590e-02 -7.20993653e-02 3.06191266e-01 5.55219769e-01
3.47378880e-01 1.95760712e-01 -2.89863706e-01 -5.30873120e-01
6.84742868e-01 1.94081023e-01 -3.28199685e-01 2.55569071e-01
-1.35122454e+00 4.91614997e-01 6.05882965e-02 6.67645156e-01
-8.35305095e-01 9.19453576e-02 9.49138999e-01 6.14711821e-01
3.02520424e-01 6.16817474e-01 -5.44385552e-01 -5.49926400e-01
-1.12286723e+00 4.27077226e-02 6.45830691e-01 1.17225394e-01
1.11149214e-01 3.75635326e-01 -4.30968553e-01 8.09436023e-01
4.44461852e-02 3.73603016e-01 8.66895854e-01 -4.28716183e-01
1.16763771e-01 5.18891513e-01 -2.86757611e-02 -1.04510391e+00
-7.51734316e-01 -1.13474786e+00 -1.41008925e+00 1.58273071e-01
2.84738660e-01 -4.53947783e-01 -8.07877481e-01 1.24848151e+00
3.09637450e-02 3.79072905e-01 1.69913158e-01 6.73329711e-01
8.47244382e-01 4.67879653e-01 -2.75244061e-02 -5.43008387e-01
1.37187195e+00 -2.05585748e-01 -8.71964574e-01 3.45362157e-01
5.06230474e-01 -5.56871951e-01 6.98711336e-01 7.85652220e-01
-9.29075122e-01 -6.86701179e-01 -1.21819174e+00 3.38554949e-01
-1.10528931e-01 4.91350085e-01 3.83482575e-01 8.54629099e-01
-7.73635507e-01 8.05193126e-01 -8.40166271e-01 -4.63317782e-01
6.35031819e-01 4.29524511e-01 -1.01055585e-01 4.77418721e-01
-1.30308855e+00 7.46755838e-01 3.18610936e-01 6.84704304e-01
-8.25688899e-01 -5.60471117e-01 -3.90603006e-01 -2.21357290e-02
-2.86937565e-01 -8.75674009e-01 8.29750896e-01 -9.90841210e-01
-1.43489790e+00 9.71107543e-01 8.83331224e-02 -8.11220646e-01
6.20582461e-01 -3.56863320e-01 -5.12497485e-01 1.01512432e-01
-2.64701545e-01 -2.53546357e-01 9.54250634e-01 -4.71667260e-01
-2.66235411e-01 -5.31522810e-01 -5.40351927e-01 -4.85152379e-02
-1.84239775e-01 1.05344847e-01 1.27962753e-01 -7.85712361e-01
1.34665191e-01 -7.83327579e-01 -9.72125120e-03 -4.44932759e-01
-3.88878047e-01 1.53272524e-01 6.38721228e-01 -8.69540036e-01
1.60468996e+00 -2.18688703e+00 -6.16625249e-02 4.14238840e-01
3.91976207e-01 7.20871210e-01 4.36783075e-01 4.57947254e-01
-3.32293719e-01 1.05275810e-01 -4.29807723e-01 2.91620735e-02
-4.63241637e-01 -6.03381805e-02 -6.76892400e-02 7.20605612e-01
-4.47506607e-02 9.73537087e-01 -3.73474687e-01 -1.47689998e-01
4.44943041e-01 8.45770121e-01 1.71427175e-01 2.76012987e-01
4.24666196e-01 7.54393816e-01 -1.85690120e-01 5.14908791e-01
3.05312365e-01 -5.63104749e-02 9.91377234e-02 -3.40957701e-01
5.25775664e-02 2.82235071e-02 -9.38742757e-01 1.80863607e+00
-4.43748683e-01 6.53058827e-01 -3.77625287e-01 -1.17606366e+00
1.13166595e+00 1.01873815e+00 5.76679289e-01 -3.01679790e-01
4.43239450e-01 4.72999454e-01 5.10600328e-01 -7.69491315e-01
-6.47398233e-01 -2.53707558e-01 1.87831149e-01 4.12118942e-01
-1.09091111e-01 2.46111006e-01 -8.37737918e-02 -4.17494744e-01
1.21190107e+00 -6.20769635e-02 6.21332884e-01 -2.96659499e-01
7.27264583e-01 -3.63963127e-01 6.71674192e-01 8.81362677e-01
-2.70972162e-01 7.06906676e-01 2.75782704e-01 -1.32234561e+00
-5.89568555e-01 -7.35894024e-01 -3.44576627e-01 2.36675546e-01
-5.57416379e-01 -1.77075878e-01 -5.92782319e-01 -3.46716285e-01
-2.62688488e-01 6.06046282e-02 -4.99624103e-01 -2.83524454e-01
-7.37937927e-01 -1.10018635e+00 1.34515011e+00 5.34800589e-01
6.98376596e-01 -1.51876819e+00 -1.50258946e+00 6.50893450e-01
-9.66830831e-03 -6.40764177e-01 1.30352885e-01 4.07326728e-01
-1.14410174e+00 -1.11433494e+00 -8.56740177e-01 -6.65263236e-01
5.76177798e-02 -7.03168392e-01 1.07062221e+00 8.38874802e-02
-1.00774348e+00 -8.90515968e-02 -5.98176531e-02 -8.11767697e-01
-3.83129954e-01 2.37652108e-01 1.51579245e-03 2.96931177e-01
3.77534211e-01 -9.88855064e-01 -9.42488730e-01 -1.51760280e-01
-3.34284872e-01 -1.81679517e-01 8.35819900e-01 8.43385100e-01
5.86015701e-01 -4.85343300e-02 1.22850466e+00 -9.16381538e-01
8.97648335e-01 -2.88875043e-01 -2.61075050e-01 -3.33850645e-02
-9.29107547e-01 -4.70395178e-01 7.32633531e-01 -2.31648982e-01
-3.95941824e-01 6.24664202e-02 -3.92301559e-01 -4.36997801e-01
-4.11156893e-01 6.69011354e-01 2.61316206e-02 1.68819815e-01
7.39683509e-01 2.87348241e-01 9.94457584e-03 -2.63213217e-01
-4.50447142e-01 7.50385642e-01 5.20144939e-01 -1.96058854e-01
2.69368052e-01 1.38822496e-01 2.96644360e-01 -9.45143580e-01
-4.59070653e-01 -4.23010141e-01 -6.17165685e-01 -1.76804632e-01
9.58122671e-01 -7.25001514e-01 -6.19253814e-01 8.44715059e-01
-1.14172506e+00 -1.83275773e-03 -3.33089858e-01 5.50087810e-01
-2.94229627e-01 1.62111282e-01 -7.19105542e-01 -1.02199757e+00
-1.32235420e+00 -7.78868675e-01 6.96942985e-01 -1.54983863e-01
-5.91964066e-01 -8.67353141e-01 2.50748366e-01 -1.24637105e-01
7.41391718e-01 1.05592263e+00 8.07593644e-01 -7.83174157e-01
1.31809369e-01 -6.58834875e-01 2.24956110e-01 5.39062381e-01
4.55353349e-01 -1.26381204e-01 -1.26131976e+00 -3.60858113e-01
2.84142315e-01 -6.27610460e-02 7.00109899e-01 6.97303712e-01
1.14382756e+00 5.92621565e-02 -5.77911623e-02 8.47037733e-01
1.40484822e+00 6.52436852e-01 7.49811351e-01 1.38941199e-01
7.15660870e-01 -1.32082611e-01 3.40191722e-02 5.37640810e-01
-2.33414948e-01 2.25763530e-01 -2.85187978e-02 -5.56162536e-01
5.96603528e-02 4.31448162e-01 5.88169508e-02 7.78908193e-01
-5.74650466e-01 4.86505516e-02 -1.16275716e+00 3.82244736e-01
-1.61663389e+00 -1.15991735e+00 -4.09535229e-01 2.41540051e+00
7.65670955e-01 1.82463646e-01 7.96175525e-02 9.90854919e-01
5.97803116e-01 -1.16970651e-01 -6.32973075e-01 -6.32851779e-01
-1.27694085e-01 7.53605962e-01 -9.33061838e-02 -3.16575821e-03
-1.10705435e+00 2.90336348e-02 5.48113823e+00 1.83780178e-01
-1.72055614e+00 1.80290818e-01 8.21911097e-01 1.85611658e-02
5.49623489e-01 -4.81052190e-01 -3.33753079e-01 3.83219481e-01
1.39065826e+00 6.71088547e-02 -2.75858846e-02 4.34419066e-01
5.31183302e-01 2.87025988e-01 -9.81969953e-01 1.58978736e+00
6.51525036e-02 -1.26733017e+00 -2.63570994e-01 -2.31200829e-01
3.15070957e-01 -2.32529398e-02 -8.45041201e-02 2.43974537e-01
-1.02252507e+00 -1.23344481e+00 1.83335647e-01 8.41329277e-01
1.27345908e+00 -9.71232891e-01 1.20265412e+00 2.71672487e-01
-1.10270929e+00 -3.98592114e-01 1.33317754e-01 -2.85827816e-01
-1.63369570e-02 1.17114091e+00 -8.25667799e-01 4.97665256e-01
6.52370811e-01 9.54669356e-01 -4.59782928e-01 1.16868746e+00
5.15814424e-02 1.24487102e+00 -2.80165434e-01 3.36677104e-01
-2.95218378e-01 1.42627098e-02 6.09182775e-01 1.23466849e+00
4.18176860e-01 -6.36830926e-02 1.40139773e-01 6.37638152e-01
7.46867657e-02 1.41865224e-01 -8.85607600e-01 9.21860561e-02
1.31528035e-01 1.27450669e+00 -7.28726745e-01 -4.98648882e-01
-8.16961229e-02 7.91153550e-01 -2.15438709e-01 2.69993514e-01
-8.39407384e-01 -8.59668851e-01 3.45898308e-02 2.56646276e-01
-2.49428004e-01 2.64878511e-01 -7.01650739e-01 -9.39255178e-01
1.59996241e-01 -1.07547474e+00 3.99915040e-01 -2.49243811e-01
-1.00909030e+00 1.04134250e+00 -3.79303545e-01 -1.22829211e+00
-4.34164554e-01 -1.83817104e-01 -9.51660156e-01 1.19150889e+00
-1.16624165e+00 -7.64424264e-01 -4.33396548e-01 5.55751979e-01
5.22972226e-01 -3.60701740e-01 1.46034455e+00 6.89196527e-01
-5.66308320e-01 4.51863259e-01 -2.33903602e-01 2.26921663e-01
5.65145075e-01 -1.23302853e+00 6.30484223e-02 8.41398954e-01
-1.65182550e-03 6.72993183e-01 4.47369009e-01 -4.41394657e-01
-9.74804997e-01 -1.16090298e+00 1.01435566e+00 -2.25000679e-01
-8.82656276e-02 -1.36278510e-01 -7.40018129e-01 2.60868847e-01
3.76656115e-01 4.26135689e-01 8.73081923e-01 -1.86822996e-01
3.33205074e-01 -5.67451775e-01 -1.02076232e+00 -4.78933156e-02
4.20985669e-01 -5.21895826e-01 -5.23543239e-01 4.20726798e-02
-2.81794131e-01 -2.41541758e-01 -1.01218522e+00 8.13952148e-01
1.06157172e+00 -1.16719770e+00 7.64265597e-01 -3.35815370e-01
9.07640904e-02 -1.95095092e-01 4.24878150e-01 -1.15105081e+00
-1.41768068e-01 -1.06527269e+00 -4.02042061e-01 9.04655397e-01
4.26605403e-01 -7.42219031e-01 5.27204633e-01 1.06321741e-02
-1.35776371e-01 -1.27959394e+00 -8.16332877e-01 -4.26280767e-01
-2.61353552e-01 -3.90925914e-01 2.47701347e-01 8.40281546e-01
-3.90906006e-01 5.46238005e-01 -4.28563625e-01 -8.06095526e-02
5.16060054e-01 1.23168506e-01 3.72206211e-01 -1.64247668e+00
-3.47135752e-01 -2.67763376e-01 -7.08364666e-01 1.33565024e-01
-3.43591273e-01 -9.02640462e-01 -3.53155822e-01 -1.46820903e+00
-8.42301473e-02 -3.02507937e-01 -1.03663969e+00 3.43454689e-01
-1.15585051e-01 5.97317219e-01 -3.76308769e-01 -1.68665759e-02
4.20452580e-02 4.88846004e-03 6.57288313e-01 1.07265841e-02
-5.41786790e-01 5.14333487e-01 -2.30447873e-01 8.02646637e-01
1.17103279e+00 -5.03599882e-01 -5.31028986e-01 6.94888681e-02
2.32654288e-01 3.42702389e-01 3.99015754e-01 -1.53250146e+00
-7.05166981e-02 7.54988015e-01 8.90148401e-01 -5.34358501e-01
8.12950507e-02 -7.10837066e-01 5.04176378e-01 9.34348404e-01
-3.11311662e-01 5.36264598e-01 2.44707748e-01 4.14445132e-01
-3.59581023e-01 2.59061456e-01 7.58954704e-01 -1.64102271e-01
-4.15120795e-02 4.07809794e-01 -6.77959800e-01 2.86736246e-02
1.01501024e+00 -3.25196326e-01 3.49444807e-01 -9.34296101e-02
-1.03352153e+00 -4.23647165e-01 -4.10495609e-01 8.00403506e-02
8.92466247e-01 -1.16324592e+00 -9.47130024e-01 4.66431409e-01
3.97854038e-02 -1.88877508e-01 1.83958426e-01 1.43921244e+00
-8.69401097e-01 2.32995048e-01 -4.21027660e-01 -8.21510017e-01
-1.34916949e+00 1.13236912e-01 8.49411607e-01 -3.42794389e-01
-1.10646141e+00 6.07675135e-01 -4.80692148e-01 1.17841482e-01
4.32751238e-01 -5.37570536e-01 -5.12791395e-01 1.20312452e-01
5.71546376e-01 4.55213934e-01 5.65864921e-01 -1.52041703e-01
-5.44455647e-01 6.40523612e-01 2.08943442e-01 1.49885997e-01
1.27528489e+00 1.42033532e-01 -1.33085594e-01 9.41400945e-01
9.70910788e-01 -2.12219074e-01 -3.62658739e-01 3.46228570e-01
-3.00233229e-03 -5.44794314e-02 8.20440426e-02 -1.23266566e+00
-1.21039557e+00 1.33460498e+00 1.43366289e+00 1.75534412e-01
1.42270648e+00 -8.66176367e-01 9.31577921e-01 5.00524461e-01
3.00920546e-01 -6.76088452e-01 -3.22779238e-01 2.76885126e-02
8.14817846e-01 -9.66605783e-01 -1.09023228e-01 3.42949390e-01
-4.01770830e-01 1.51192677e+00 7.31717721e-02 -3.64594519e-01
9.54707026e-01 2.85588592e-01 5.13751328e-01 -3.56319249e-01
-5.36332369e-01 2.49070808e-01 1.05089925e-01 4.37983841e-01
9.23290312e-01 -9.77904722e-02 -6.04701936e-01 7.86391318e-01
2.01504782e-01 6.05727673e-01 3.69019210e-01 9.10838425e-01
1.18766641e-02 -8.48136544e-01 -1.32498533e-01 8.31879675e-01
-1.18884456e+00 -1.44741386e-01 -2.16587290e-01 3.83303553e-01
3.98580641e-01 8.33832979e-01 -4.13778394e-01 -8.45817924e-02
2.13299289e-01 5.64458907e-01 3.27132434e-01 -4.76602107e-01
-1.28948379e+00 -1.88010149e-02 1.26362398e-01 -2.83443242e-01
-4.62672144e-01 -4.42231148e-01 -1.17434978e+00 4.35246885e-01
-2.68093199e-01 1.14137203e-01 6.25633299e-01 9.23843503e-01
6.90127134e-01 1.00900495e+00 3.63192618e-01 -5.57307839e-01
-4.01371330e-01 -1.33381283e+00 -7.82693505e-01 3.62208158e-01
7.10715950e-01 -3.04493070e-01 -1.91678867e-01 3.01765442e-01] | [14.282060623168945, 3.2794275283813477] |
76a71b2d-e719-49ec-94b5-3af1917a6c86 | convolutional-neural-networks-deceived-by | 1811.10565 | null | http://arxiv.org/abs/1811.10565v1 | http://arxiv.org/pdf/1811.10565v1.pdf | Convolutional Neural Networks Deceived by Visual Illusions | Visual illusions teach us that what we see is not always what it is
represented in the physical world. Its special nature make them a fascinating
tool to test and validate any new vision model proposed. In general, current
vision models are based on the concatenation of linear convolutions and
non-linear operations. In this paper we get inspiration from the similarity of
this structure with the operations present in Convolutional Neural Networks
(CNNs). This motivated us to study if CNNs trained for low-level visual tasks
are deceived by visual illusions. In particular, we show that CNNs trained for
image denoising, image deblurring, and computational color constancy are able
to replicate the human response to visual illusions, and that the extent of
this replication varies with respect to variation in architecture and spatial
pattern size. We believe that this CNNs behaviour appears as a by-product of
the training for the low level vision tasks of denoising, color constancy or
deblurring. Our work opens a new bridge between human perception and CNNs: in
order to obtain CNNs that better replicate human behaviour, we may need to
start aiming for them to better replicate visual illusions. | ['Marcelo Bertalmío', 'Javier Vazquez-Corral', 'Adrián Martín', 'Alexander Gomez-Villa'] | 2018-11-26 | null | null | null | null | ['color-constancy'] | ['computer-vision'] | [ 1.05796747e-01 -6.42896220e-02 5.15471339e-01 -2.60480583e-01
5.90117753e-01 -6.94367945e-01 9.26968873e-01 -6.54669926e-02
-3.12022984e-01 4.10356581e-01 1.72435477e-01 -2.72831529e-01
2.09056854e-01 -6.35763168e-01 -8.42174411e-01 -7.83705771e-01
2.27480158e-01 -2.13639036e-01 3.13728780e-01 -6.90685689e-01
4.72367585e-01 8.00572693e-01 -1.74165165e+00 4.73751187e-01
6.31486773e-01 7.45181024e-01 9.75247398e-02 9.12212014e-01
1.32493839e-01 1.01804638e+00 -5.05114496e-01 -4.06033516e-01
2.74842650e-01 -6.93054020e-01 -6.45264924e-01 1.80422366e-01
9.41639960e-01 -2.11278513e-01 -5.11599720e-01 1.25352168e+00
3.43035072e-01 -7.98658431e-02 8.56670022e-01 -1.30439842e+00
-1.44125342e+00 3.33477464e-03 -3.70167226e-01 5.21636844e-01
2.15935916e-01 5.71319461e-01 5.46647072e-01 -6.54061556e-01
5.27846575e-01 1.36692822e+00 8.36843848e-01 6.51153743e-01
-1.51311219e+00 -1.45265862e-01 -1.15511797e-01 3.03288013e-01
-1.03374267e+00 -5.30839384e-01 5.45121014e-01 -4.59144324e-01
8.96619022e-01 2.86603302e-01 8.38972330e-01 1.16704512e+00
5.52046835e-01 4.34046805e-01 1.73847258e+00 -4.84718740e-01
2.25575909e-01 2.89396107e-01 5.89793883e-02 5.90206444e-01
6.20595396e-01 5.53755760e-01 -3.10800552e-01 3.95422548e-01
1.12925291e+00 9.17747244e-02 -6.62101865e-01 -2.04814270e-01
-1.13254368e+00 5.11060953e-01 8.55801404e-01 7.36620963e-01
-3.18757623e-01 4.50043947e-01 1.32750347e-01 6.34890378e-01
1.47932991e-01 8.00661743e-01 -3.33303422e-01 2.27278888e-01
-8.63222122e-01 1.35837540e-01 7.02651322e-01 3.30512345e-01
7.18934119e-01 3.76351565e-01 -9.84350145e-02 5.51405966e-01
5.35450177e-03 4.11486745e-01 7.70136774e-01 -1.05169976e+00
-3.84878755e-01 4.92671281e-01 -3.69216315e-02 -1.32798803e+00
-5.37149966e-01 -3.69976372e-01 -1.08156610e+00 1.00367141e+00
7.65674651e-01 2.52871484e-01 -1.08891165e+00 1.72852254e+00
-3.49611491e-01 -3.44372168e-02 1.81015119e-01 1.24263239e+00
6.69670343e-01 3.05858046e-01 -2.24524081e-01 1.14182815e-01
1.48995924e+00 -5.96424103e-01 -5.87840259e-01 -3.44096601e-01
2.44541168e-01 -9.11641717e-01 1.28776932e+00 5.10498881e-01
-1.18924963e+00 -9.36419129e-01 -1.21548033e+00 -2.66370356e-01
-6.33310556e-01 -1.97603360e-01 5.68839252e-01 8.94423783e-01
-1.43103385e+00 8.97820055e-01 -3.85840446e-01 -6.86739981e-01
3.31803709e-01 -1.35295913e-01 -4.41093773e-01 -1.53735101e-01
-8.15244973e-01 1.29992247e+00 2.01215103e-01 3.50371957e-01
-7.68016517e-01 -4.17345107e-01 -6.63257778e-01 1.24561422e-01
-1.41489193e-01 -7.32770026e-01 8.80287886e-01 -1.55851090e+00
-1.10248053e+00 1.31804013e+00 -1.49574563e-01 -5.26171684e-01
4.92301852e-01 2.00282767e-01 -4.18490797e-01 4.07794975e-02
-3.33643407e-01 5.76270759e-01 1.18793869e+00 -1.80718756e+00
-1.69585105e-02 -4.18302953e-01 2.02388633e-02 -2.50281900e-01
6.12170175e-02 -2.27485940e-01 1.80643395e-01 -6.84864998e-01
2.91418105e-01 -8.68476510e-01 7.59231597e-02 2.46876135e-01
-3.48240584e-01 2.51766890e-01 6.96951509e-01 -5.86042047e-01
4.84796196e-01 -2.17787957e+00 3.14264372e-02 7.42220730e-02
5.72675049e-01 5.06091058e-01 -4.37759578e-01 4.99741048e-01
-5.07122755e-01 2.39380747e-01 -1.62512094e-01 -2.57667869e-01
1.13490425e-01 3.69368285e-01 -4.09838706e-01 8.14517736e-01
3.22154433e-01 1.26188552e+00 -6.37561500e-01 -1.23460619e-02
1.94404826e-01 6.96335912e-01 -3.39944661e-01 7.94671550e-02
-1.12144552e-01 5.58034599e-01 3.52453560e-01 4.45998818e-01
1.06562400e+00 -6.45026043e-02 -2.46428624e-01 -3.00684780e-01
-3.42350572e-01 -2.54801691e-01 -7.81449318e-01 1.30239320e+00
-1.20778605e-01 1.33818972e+00 -3.59254628e-02 -1.18215489e+00
7.67208815e-01 8.61607864e-02 -2.29587972e-01 -1.37789226e+00
3.03406745e-01 2.36985043e-01 4.42936778e-01 -7.56411970e-01
4.06720221e-01 -5.39246678e-01 4.22418684e-01 3.99577260e-01
-1.71771757e-02 -5.56017160e-01 2.96200309e-02 -1.27684390e-02
8.59125078e-01 -3.14237177e-01 2.26953868e-02 -3.15069646e-01
4.31141526e-01 -2.10588172e-01 -4.34773453e-02 1.03146601e+00
-1.76632404e-01 9.54739332e-01 5.89597821e-01 -7.91876018e-01
-1.25702214e+00 -1.20931506e+00 -8.95186439e-02 7.97024965e-01
1.06529362e-01 2.84865886e-01 -8.15944672e-01 -5.01405410e-02
-1.77542999e-01 8.01603556e-01 -1.06294119e+00 -5.00973463e-01
-3.09658676e-01 -5.80210567e-01 6.21440768e-01 7.12028742e-02
8.01533759e-01 -1.39369965e+00 -9.99202967e-01 -2.49404415e-01
1.46931678e-01 -1.11048615e+00 -1.61768869e-01 2.64330238e-01
-6.83882654e-01 -1.14107680e+00 -9.92355108e-01 -7.77086973e-01
6.64216638e-01 5.51604986e-01 1.26078844e+00 5.55764258e-01
-4.44174767e-01 4.08133090e-01 -3.36189270e-01 -4.41149384e-01
-7.03734934e-01 -4.01398599e-01 -7.20007643e-02 5.62578719e-03
8.68593454e-02 -8.46160352e-01 -8.70193303e-01 1.14554726e-01
-1.51429236e+00 -1.39798731e-01 7.15993345e-01 7.05566168e-01
-1.27573371e-01 -3.35934833e-02 -6.03505783e-03 -5.37926972e-01
9.80176568e-01 -7.19427392e-02 -1.92012340e-01 1.20719686e-01
-4.82599825e-01 3.01319093e-01 6.15892529e-01 -5.60463727e-01
-8.47373962e-01 -2.41644919e-01 1.10136950e-02 -2.39263833e-01
-6.00233614e-01 2.71045893e-01 2.27765158e-01 -4.41011429e-01
1.10896075e+00 5.85120261e-01 2.20813468e-01 -2.63623893e-01
5.37780523e-01 3.29008937e-01 6.17398143e-01 -4.77938168e-02
9.86233652e-01 8.64416718e-01 5.56553192e-02 -1.05765593e+00
-6.01036370e-01 1.44947441e-02 -7.66184211e-01 -2.54580557e-01
1.04158401e+00 -3.77128482e-01 -1.03319132e+00 8.35228860e-01
-1.53047514e+00 -5.34712732e-01 -5.23177326e-01 1.50623381e-01
-5.38821101e-01 4.44450051e-01 -4.62207764e-01 -7.17642844e-01
1.62962437e-01 -9.09164310e-01 6.31830275e-01 3.05355638e-01
-6.12539845e-03 -1.08696771e+00 1.61581963e-01 2.90266536e-02
8.88177991e-01 2.46631280e-01 1.07941949e+00 -1.43250659e-01
-5.19079745e-01 -4.75249365e-02 -7.89102197e-01 7.26733267e-01
-1.23066707e-02 -9.58690941e-02 -1.29007208e+00 -2.87078440e-01
2.81304687e-01 -3.02240878e-01 1.29887640e+00 4.59614605e-01
8.02130401e-01 -3.69061530e-01 2.99319893e-01 6.88893616e-01
1.65295482e+00 -1.25195786e-01 1.37156498e+00 3.31234604e-01
4.74516302e-01 8.43055367e-01 -3.22437793e-01 -1.87095646e-02
-1.17411345e-01 5.59100449e-01 7.25582004e-01 -7.25296795e-01
-5.07080495e-01 3.67017090e-03 2.38939345e-01 3.27965796e-01
-4.71466422e-01 -6.39851168e-02 -7.55500257e-01 4.35917199e-01
-1.57965803e+00 -1.27883101e+00 -5.92637002e-01 2.13917470e+00
7.18015909e-01 1.21131293e-01 -3.20580862e-02 2.10524529e-01
3.53138983e-01 6.03630394e-02 -3.52190018e-01 -8.94840121e-01
-6.37134850e-01 2.25338772e-01 4.83789295e-01 3.73237163e-01
-5.58694124e-01 8.45788717e-01 6.77809954e+00 3.42255265e-01
-1.50271714e+00 -2.34518070e-02 6.34712160e-01 4.07395780e-01
-1.87976420e-01 -6.87787235e-02 8.94669071e-02 2.01385304e-01
6.11087799e-01 3.02840024e-01 7.79208124e-01 2.05579847e-01
2.89040238e-01 -3.08507740e-01 -1.13826573e+00 1.05612552e+00
3.75742525e-01 -1.26252115e+00 2.36766368e-01 -2.42994353e-02
7.26343274e-01 -1.11542664e-01 6.25040412e-01 -4.14469419e-03
2.29548156e-01 -1.68131673e+00 8.20947528e-01 8.88136208e-01
3.33983988e-01 -2.35380307e-01 5.90495288e-01 2.40387693e-01
-5.37888348e-01 -6.88951910e-02 -6.29293799e-01 -4.33768660e-01
-1.20200634e-01 4.07970607e-01 -6.20351732e-01 -1.01286635e-01
7.19887137e-01 4.25648421e-01 -9.67486322e-01 1.13887715e+00
-1.88862309e-01 2.59773463e-01 2.74761766e-01 5.00525609e-02
2.07605809e-01 -7.13508502e-02 5.15333176e-01 1.29934621e+00
1.34055957e-01 -2.41332158e-01 -5.57803392e-01 1.34176612e+00
2.89175123e-01 -2.81494141e-01 -7.17053592e-01 8.75674188e-02
-1.85825557e-01 1.01536357e+00 -8.45245242e-01 -1.93025813e-01
-3.99381995e-01 1.65587425e+00 1.22726180e-01 7.20004618e-01
-6.13486946e-01 -2.63295382e-01 7.52919793e-01 3.05856884e-01
4.15111870e-01 -3.54361713e-01 -4.99910802e-01 -1.12048769e+00
-1.13844156e-01 -9.16906893e-01 -4.28361982e-01 -1.40926468e+00
-1.47401762e+00 7.12423980e-01 -3.79744202e-01 -8.02887738e-01
6.92191198e-02 -9.12924886e-01 -8.25362861e-01 1.01604688e+00
-1.61549389e+00 -1.14432776e+00 -7.26273835e-01 8.05455804e-01
7.27015361e-02 6.56110868e-02 6.47722483e-01 -1.92856997e-01
-9.07417014e-02 3.03844571e-01 -9.57204327e-02 1.58931419e-01
5.94199121e-01 -1.35775077e+00 5.59009910e-01 8.54741096e-01
3.15882087e-01 6.64547563e-01 1.31597722e+00 -1.03341326e-01
-1.08194935e+00 -5.72309673e-01 6.55855179e-01 -4.33742583e-01
5.42431116e-01 -3.19234520e-01 -1.15302658e+00 4.20873851e-01
6.11680686e-01 1.48712724e-01 1.12924881e-01 -2.55207509e-01
-8.77932370e-01 5.16786017e-02 -1.13135219e+00 7.23770142e-01
8.42406452e-01 -6.60208642e-01 -7.57317483e-01 1.00304142e-01
4.41925883e-01 1.16164334e-01 -1.68271482e-01 2.21200865e-02
5.56635439e-01 -1.67828023e+00 1.18732989e+00 -6.92374945e-01
6.65934563e-01 -2.06967920e-01 1.41975153e-02 -1.57436061e+00
-3.73871654e-01 -3.20814461e-01 4.79414195e-01 6.62431121e-01
2.82250255e-01 -8.61752331e-01 3.75370950e-01 3.92281532e-01
-2.56626587e-02 -7.89987594e-02 -7.61612892e-01 -7.83773124e-01
3.02849084e-01 -2.24935532e-01 1.64454039e-02 1.08240533e+00
-3.01911443e-01 1.32342115e-01 -3.15460354e-01 -2.63968818e-02
6.28064096e-01 -8.57852548e-02 6.69674456e-01 -1.34128785e+00
-2.47460306e-01 -7.33375549e-01 -5.70961773e-01 -8.35522234e-01
-6.50827661e-02 -6.54855072e-01 -5.92040382e-02 -1.35240149e+00
1.02777310e-01 -5.91029786e-02 -1.83814749e-01 3.28691840e-01
2.24752456e-01 7.80196488e-01 4.10652518e-01 3.22005659e-01
-2.31833667e-01 1.64256483e-01 1.63293338e+00 1.85664501e-02
-1.47291958e-01 -2.21023604e-01 -9.25722361e-01 7.99061716e-01
9.25563097e-01 -2.10192218e-01 -1.07897379e-01 -3.84782404e-01
5.49263000e-01 -3.51823628e-01 1.10933983e+00 -1.13780951e+00
2.72288591e-01 -6.33375794e-02 6.40378058e-01 -4.65903096e-02
3.72485995e-01 -8.52584779e-01 -1.34753257e-01 8.24784398e-01
-2.08571732e-01 5.70417084e-02 3.90884697e-01 4.26484227e-01
-2.03833446e-01 -2.98866212e-01 1.06095910e+00 -5.67508161e-01
-8.01083505e-01 -2.30121836e-01 -7.46990085e-01 4.81783859e-02
6.37064636e-01 -6.28079772e-01 -6.06356323e-01 -5.48003495e-01
-7.84329236e-01 -6.18299246e-01 7.44217277e-01 2.92025000e-01
7.35966623e-01 -9.39388037e-01 -7.52898335e-01 2.71417022e-01
3.67735066e-02 -6.80029035e-01 3.06697398e-01 9.76502478e-01
-7.28861392e-01 2.99516618e-01 -9.73502994e-01 -3.82240653e-01
-1.17112148e+00 7.91483879e-01 7.22445250e-01 3.46039742e-01
-3.83264154e-01 8.46857309e-01 3.61914366e-01 4.65922840e-02
-1.12259493e-03 -6.68652833e-01 -3.31964865e-02 -1.28941476e-01
5.82895279e-01 7.12358505e-02 -1.04282267e-01 -5.08684695e-01
-7.54368976e-02 5.69872260e-01 1.13981701e-01 -1.35632511e-02
1.40937722e+00 -2.00011790e-01 -5.19927740e-01 5.01942158e-01
1.11804211e+00 2.33645551e-02 -1.23241699e+00 -1.07634086e-02
-2.60092556e-01 -4.86189187e-01 5.64612933e-02 -9.56218958e-01
-1.16185236e+00 1.14116788e+00 8.38807821e-01 8.53189230e-01
1.25646102e+00 1.14810020e-01 3.16840798e-01 2.71172196e-01
1.94876455e-02 -7.38404989e-01 6.21688128e-01 5.46365738e-01
1.34540677e+00 -1.23717630e+00 -2.53240287e-01 -1.27746016e-01
-5.05045891e-01 1.24533439e+00 4.14824307e-01 -6.54855609e-01
4.59379971e-01 1.46202400e-01 1.77518457e-01 -4.42910671e-01
-4.24598396e-01 -6.77134275e-01 3.41748595e-01 1.04195428e+00
4.10452396e-01 -1.78956717e-01 4.13407199e-02 8.84684399e-02
-2.85679817e-01 -8.63507539e-02 8.29842031e-01 3.99445921e-01
-7.18290567e-01 -7.31973350e-01 -5.98608553e-01 -1.98835626e-01
-2.22033367e-01 -3.12055856e-01 -8.99234772e-01 8.62942278e-01
4.97792065e-01 8.00935864e-01 2.39633560e-01 -3.54302138e-01
4.27275717e-01 -6.04696684e-02 8.88225317e-01 -2.56479084e-01
-5.98107159e-01 -5.34694254e-01 -3.75501901e-01 -4.60603774e-01
-4.29452270e-01 -3.29682946e-01 -7.62851179e-01 -7.13800967e-01
1.38617650e-01 -5.40988564e-01 7.38644660e-01 8.08793008e-01
-5.74579388e-02 5.12218058e-01 1.59116328e-01 -1.08973157e+00
-1.99558884e-01 -1.10855079e+00 -7.70006716e-01 9.78757679e-01
8.24568689e-01 -3.84311676e-01 -7.08521008e-01 3.17992419e-01] | [10.079373359680176, 2.314777374267578] |
9dfaa168-449f-4695-83c5-fc4d14cf093d | sentence-level-recurrent-topic-model-letting | 1604.02038 | null | http://arxiv.org/abs/1604.02038v2 | http://arxiv.org/pdf/1604.02038v2.pdf | Sentence Level Recurrent Topic Model: Letting Topics Speak for Themselves | We propose Sentence Level Recurrent Topic Model (SLRTM), a new topic model
that assumes the generation of each word within a sentence to depend on both
the topic of the sentence and the whole history of its preceding words in the
sentence. Different from conventional topic models that largely ignore the
sequential order of words or their topic coherence, SLRTM gives full
characterization to them by using a Recurrent Neural Networks (RNN) based
framework. Experimental results have shown that SLRTM outperforms several
strong baselines on various tasks. Furthermore, SLRTM can automatically
generate sentences given a topic (i.e., topics to sentences), which is a key
technology for real world applications such as personalized short text
conversation. | ['Tie-Yan Liu', 'Fei Tian', 'Bin Gao', 'Di He'] | 2016-04-07 | null | null | null | null | ['short-text-conversation'] | ['natural-language-processing'] | [ 1.08432092e-01 4.87438977e-01 -3.49277169e-01 -4.48555410e-01
-6.94994748e-01 -2.07632169e-01 1.15720546e+00 3.97569835e-02
1.55513495e-01 9.59740818e-01 1.02631843e+00 -3.15246165e-01
2.97288924e-01 -9.27893102e-01 -5.12479901e-01 -6.63992763e-01
7.38437707e-03 4.59430218e-01 3.53539139e-01 -5.34286737e-01
3.57692242e-01 -3.99801880e-01 -1.07468700e+00 3.35353583e-01
5.84745765e-01 2.92209804e-01 7.32516885e-01 6.69494629e-01
-6.38651252e-01 7.42554009e-01 -1.06791222e+00 -1.08230017e-01
-5.84999919e-01 -6.10452414e-01 -9.94155228e-01 1.10886000e-01
5.33787012e-02 -1.31246552e-01 -6.00858271e-01 6.99855208e-01
1.32322893e-01 4.93788898e-01 6.89165473e-01 -1.08976662e+00
-4.22383696e-01 1.24368215e+00 -4.49653506e-01 1.19349621e-01
3.67205590e-01 -3.34001988e-01 1.09105253e+00 -7.81548679e-01
8.80909562e-01 1.62675619e+00 4.61954445e-01 6.08722687e-01
-8.83380651e-01 -5.84185064e-01 7.80075431e-01 -2.26037398e-01
-8.61574650e-01 3.78801599e-02 8.36374521e-01 -2.03319207e-01
1.14655709e+00 6.15327246e-02 5.89437783e-01 1.49162090e+00
9.11439061e-01 9.86505985e-01 8.04639220e-01 -2.45432809e-01
1.80607617e-01 1.07461214e-01 8.82291198e-01 -1.27685934e-01
5.82392514e-02 -6.74657762e-01 -8.46611202e-01 -4.94492412e-01
4.25118566e-01 3.79839130e-02 -4.17817533e-02 2.79772729e-01
-1.22231007e+00 1.17853177e+00 -2.34123543e-02 4.27124679e-01
-7.17069387e-01 2.14999139e-01 6.28303707e-01 2.49250303e-03
1.04161572e+00 2.18131781e-01 -4.05727625e-01 -2.46634811e-01
-1.04760182e+00 5.30324697e-01 9.94304061e-01 1.23390114e+00
5.81512570e-01 -1.08657449e-01 -5.23535192e-01 7.34857976e-01
4.46887493e-01 3.36188257e-01 1.10044277e+00 -5.14651179e-01
6.21378183e-01 3.72905940e-01 5.07552475e-02 -9.88070548e-01
-2.16148317e-01 -3.89165461e-01 -8.46316755e-01 -7.60452211e-01
-4.14200604e-01 -4.38081592e-01 -9.18067634e-01 1.66512203e+00
6.58407882e-02 2.40981266e-01 3.09598684e-01 2.31624186e-01
1.13968444e+00 1.71032739e+00 3.22665751e-01 -6.92559958e-01
1.24750745e+00 -9.83533561e-01 -1.11493742e+00 -2.92395145e-01
2.60856301e-01 -5.54755270e-01 9.15420234e-01 2.24431440e-01
-9.54500079e-01 -3.36934894e-01 -8.75136912e-01 -6.30023926e-02
-2.58420140e-01 -2.26252511e-01 7.14006007e-01 3.92869025e-01
-1.26862013e+00 2.31956318e-01 -7.51148462e-01 -6.73678696e-01
-1.29253879e-01 7.42093027e-02 2.30498686e-01 2.19314203e-01
-1.65366375e+00 5.63189507e-01 5.20401001e-01 -7.40641579e-02
-9.70478415e-01 -6.09443545e-01 -7.90456057e-01 1.18562564e-01
2.37816229e-01 -6.76093400e-01 1.69269717e+00 -3.43479246e-01
-1.57604444e+00 3.63609463e-01 -1.02353787e+00 -9.01068270e-01
-1.69672355e-01 -5.63539565e-01 -2.55523652e-01 6.55172318e-02
3.24480176e-01 7.19280481e-01 8.10221553e-01 -1.04942143e+00
-6.34550512e-01 1.08159795e-01 -1.26298070e-01 4.99833465e-01
-4.35483605e-01 2.78386891e-01 -3.02612394e-01 -7.26234496e-01
1.16881594e-01 -7.44757950e-01 -4.27711993e-01 -9.56693769e-01
-7.03394890e-01 -1.08274567e+00 1.26810575e+00 -5.46953797e-01
1.44236231e+00 -1.88913798e+00 7.78656676e-02 -3.57946813e-01
6.97724074e-02 -1.14800110e-01 9.58920196e-02 1.07518804e+00
-5.29896989e-02 2.92912632e-01 -9.72780283e-04 -5.68450809e-01
7.86805805e-03 1.66457057e-01 -1.41485488e+00 -6.53063506e-02
2.61383932e-02 8.84299159e-01 -9.57151234e-01 -4.70159560e-01
-5.54780252e-02 5.12218833e-01 -7.25551024e-02 2.32524112e-01
-7.52647579e-01 1.32821715e-02 -6.18177056e-01 -1.99561447e-01
1.18920363e-01 -4.97321248e-01 1.65764958e-01 4.57964867e-01
7.93846138e-03 1.26405549e+00 -5.54899156e-01 1.61740112e+00
-5.63154936e-01 1.02863014e+00 -5.70816755e-01 -4.85856891e-01
1.12978792e+00 9.97610867e-01 2.80473679e-01 6.10736292e-03
-1.79394349e-01 -3.54307383e-01 -4.27257210e-01 2.25882325e-02
1.43020797e+00 -1.09040178e-01 -4.20289457e-01 1.14616668e+00
1.09301277e-01 -1.95604637e-01 2.31256515e-01 9.78708744e-01
6.68808341e-01 -1.62909865e-01 3.18093061e-01 -3.10650826e-01
2.12787967e-02 -2.12446451e-01 4.18088257e-01 9.62819874e-01
1.36882529e-01 6.52182221e-01 6.92168593e-01 -3.07968050e-01
-9.68782187e-01 -8.65892231e-01 -3.75788137e-02 1.33348548e+00
3.14014405e-02 -8.63825560e-01 -6.51144564e-01 -4.75756347e-01
-5.61999798e-01 1.26740754e+00 -7.22784042e-01 4.12390083e-02
-6.25523090e-01 -9.76544738e-01 7.84496963e-02 2.65762866e-01
2.98886538e-01 -1.45016170e+00 -2.88016558e-01 5.40869892e-01
-7.02384651e-01 -9.01575506e-01 -7.56107926e-01 -1.32007107e-01
-1.05054641e+00 -3.53779048e-01 -8.26960146e-01 -6.30520940e-01
3.21803480e-01 6.98334932e-01 1.24366105e+00 -4.02534723e-01
1.87852442e-01 2.88238674e-01 -5.84353387e-01 -5.52419364e-01
-5.62112987e-01 6.29072964e-01 -9.76259038e-02 -2.03473985e-01
2.86199063e-01 -6.21165872e-01 -5.15913844e-01 3.04880142e-02
-8.48002374e-01 2.77743131e-01 2.02674791e-01 6.71628714e-01
9.11680758e-02 3.24715644e-01 8.92226756e-01 -1.16807473e+00
1.21799874e+00 -8.85197282e-01 1.55978305e-02 2.61030644e-01
-3.91419530e-01 -1.92009076e-01 6.87293634e-02 -5.56881428e-01
-1.41214430e+00 -6.42712295e-01 -6.74699768e-02 2.81583130e-01
-6.91672266e-02 7.51441360e-01 1.82414249e-01 1.18886364e+00
2.09492847e-01 8.75784576e-01 -2.69491702e-01 -2.25816593e-01
2.97746629e-01 6.72680616e-01 2.31685579e-01 -4.42476779e-01
3.00255030e-01 3.35186362e-01 -5.57468653e-01 -1.19661760e+00
-9.98643219e-01 -6.83234274e-01 -3.07160735e-01 -2.76194930e-01
7.10399866e-01 -1.11876202e+00 -2.91396320e-01 3.69479835e-01
-1.73313177e+00 -1.64591938e-01 -2.09876627e-01 3.22011232e-01
-3.68205726e-01 3.41602057e-01 -8.10081601e-01 -1.15738750e+00
-8.37228835e-01 -6.69558108e-01 1.41566920e+00 4.10807818e-01
-7.81765938e-01 -1.36329532e+00 1.13121502e-01 -1.35251686e-01
5.15003443e-01 -1.49603248e-01 8.46638799e-01 -9.03057575e-01
-3.94527644e-01 -1.77611411e-01 1.73393145e-01 -1.37751639e-01
3.26633811e-01 1.41008161e-02 -8.18236232e-01 -2.18447879e-01
4.39000696e-01 -4.89001796e-02 9.72234011e-01 7.47954786e-01
5.24440169e-01 -6.38232112e-01 -7.03534782e-01 -1.91081449e-01
1.00338614e+00 3.36931944e-01 7.85329223e-01 1.30992606e-01
2.65800744e-01 7.14583099e-01 4.67069745e-01 4.23596114e-01
6.33852541e-01 5.09751856e-01 -6.01113252e-02 1.00810669e-01
-5.98824508e-02 -5.38141966e-01 8.14628899e-01 1.35784101e+00
3.84461731e-01 -6.72402501e-01 -6.65840030e-01 7.68973589e-01
-2.00682616e+00 -1.06267178e+00 -3.52091372e-01 1.66240740e+00
1.01151562e+00 3.53302538e-01 1.16618276e-01 -3.90253514e-01
1.02032971e+00 6.92597568e-01 -4.20524538e-01 -4.92615491e-01
-2.58559227e-01 -1.84163630e-01 -1.81601837e-01 5.62995672e-01
-9.45717573e-01 1.25637817e+00 7.39512253e+00 8.27337205e-01
-9.05894876e-01 2.29868010e-01 7.16056943e-01 9.11205485e-02
-5.57722270e-01 1.90678071e-02 -1.49229646e+00 4.72857028e-01
1.19469762e+00 -9.94651914e-01 -2.57959664e-01 9.39151704e-01
5.58531404e-01 -1.94820195e-01 -6.64439797e-01 1.56931609e-01
3.29081804e-01 -1.36214018e+00 4.41637903e-01 8.31215382e-02
1.07519782e+00 1.01186998e-01 8.02021846e-02 4.63266075e-01
8.16188872e-01 -6.44280434e-01 3.14436257e-01 3.99145871e-01
3.74573953e-02 -5.65142930e-01 8.60037982e-01 5.18270016e-01
-1.01405120e+00 4.09759939e-01 -6.14862263e-01 -9.83556435e-02
7.25537419e-01 7.81646013e-01 -1.17114556e+00 5.24322331e-01
5.53610861e-01 9.43751514e-01 -9.98996794e-02 5.65724194e-01
-4.24868107e-01 1.22293663e+00 -5.61443828e-02 -4.17073697e-01
4.09614921e-01 5.99633483e-03 9.70361352e-01 1.42162037e+00
2.39518046e-01 -1.09614335e-01 1.07209861e-01 6.25899971e-01
-3.17455977e-02 2.40640670e-01 -6.37993813e-01 -2.29756776e-02
4.64108199e-01 1.16775048e+00 -8.43758404e-01 -6.34670675e-01
-1.11970119e-01 8.03400815e-01 -1.35404408e-01 5.71103394e-01
-4.99513596e-01 -3.76448542e-01 5.10230064e-01 -9.68949497e-02
3.16867411e-01 -3.96859765e-01 -2.39312366e-01 -1.02188504e+00
3.42623820e-03 -3.05714518e-01 2.22653642e-01 -8.98005843e-01
-1.32583153e+00 1.02427256e+00 3.13737780e-01 -1.03793192e+00
-7.66885400e-01 1.52938977e-01 -1.41598117e+00 1.00338030e+00
-1.29903758e+00 -1.22162545e+00 2.85905808e-01 1.65105239e-01
1.39395010e+00 -2.48512954e-01 7.75655687e-01 -6.93241715e-01
-4.51793522e-01 -3.12215537e-02 -9.79181193e-03 -2.41156727e-01
6.66111410e-01 -1.22938144e+00 1.10468674e+00 7.45685279e-01
4.10138555e-02 1.18724859e+00 1.10084403e+00 -1.07645535e+00
-8.90764415e-01 -1.31341934e+00 1.60744190e+00 -3.09157759e-01
7.49458730e-01 -6.90475583e-01 -9.78961766e-01 8.95367265e-01
1.04917479e+00 -1.21813023e+00 7.62512326e-01 6.00711763e-01
-2.36685225e-03 2.37462819e-01 -2.96144873e-01 6.99066877e-01
4.07343537e-01 -4.45642978e-01 -1.08181858e+00 8.12668502e-01
1.85040236e+00 -1.94827870e-01 -3.89192075e-01 7.43064145e-03
2.64365703e-01 -5.07792652e-01 4.81562257e-01 -6.64441884e-01
8.20731163e-01 5.73354214e-03 -5.50618023e-02 -1.35132718e+00
-1.62201703e-01 -1.04726112e+00 -2.88755745e-01 1.61842763e+00
5.91746211e-01 -6.17296934e-01 7.26208806e-01 5.40776610e-01
-3.49768430e-01 -6.04489982e-01 -8.02651107e-01 -5.39245188e-01
1.90386549e-01 -5.38258672e-01 6.02692068e-01 7.00351477e-01
4.73708987e-01 8.34657013e-01 -6.42740548e-01 1.41918808e-02
2.21529692e-01 2.95453310e-01 5.80204725e-01 -1.13209867e+00
-7.65811950e-02 -3.88569713e-01 3.04261371e-02 -1.42245018e+00
3.47438872e-01 -4.54237729e-01 5.09585381e-01 -1.94846225e+00
5.60160160e-01 6.35627434e-02 -5.71628474e-03 6.00038804e-02
-4.52773839e-01 -2.32874930e-01 -4.46249247e-02 2.76614308e-01
-8.84458542e-01 1.06025565e+00 9.21150088e-01 -8.69172141e-02
-4.66033816e-01 4.06334430e-01 -8.87021542e-01 5.94330192e-01
8.20828199e-01 -4.91161704e-01 -6.36152983e-01 -1.19385056e-01
3.18075001e-01 2.44793117e-01 -2.01789632e-01 -4.28630978e-01
5.13245106e-01 -1.99268907e-01 -2.69726783e-01 -1.32918000e+00
3.99228722e-01 -1.30515039e-01 -2.28375077e-01 3.02592874e-01
-7.93724775e-01 8.44140872e-02 6.60273135e-02 8.10663402e-01
-3.68834645e-01 -2.17261776e-01 -2.35539265e-02 -1.75828099e-01
-3.42792839e-01 2.74351805e-01 -1.06752181e+00 -2.52580434e-01
8.26334417e-01 1.49259150e-01 -4.06311274e-01 -1.18078971e+00
-4.59559441e-01 3.56283039e-01 -1.19555868e-01 8.88826966e-01
6.58015609e-01 -1.03187549e+00 -7.96348810e-01 -3.91436815e-01
-1.98939130e-01 2.13600948e-01 3.35817546e-01 1.57164976e-01
2.78758377e-01 1.07289231e+00 5.78061938e-01 -6.27795458e-01
-1.09333742e+00 3.49490345e-01 -2.94515997e-01 -6.80423141e-01
-9.07099068e-01 6.40052140e-01 7.45219290e-01 -1.57243490e-01
3.47101539e-01 -1.85397968e-01 -5.88429987e-01 6.04857355e-02
7.71585941e-01 7.67639428e-02 -3.44052881e-01 -4.86699551e-01
1.40507489e-01 1.10256456e-01 -7.28995025e-01 -4.77337003e-01
1.35218310e+00 -4.72559065e-01 -4.72369820e-01 1.30812991e+00
9.61273432e-01 -4.77249920e-01 -8.52257013e-01 -4.62851107e-01
9.04867798e-02 1.93276212e-01 -4.27007005e-02 -3.62239331e-01
-3.09907228e-01 8.39528203e-01 -2.59935677e-01 6.03153408e-01
7.12656379e-01 4.15971041e-01 1.17400897e+00 4.27659512e-01
2.38259539e-01 -9.73028958e-01 2.92828679e-01 9.99945879e-01
1.15535700e+00 -8.21442068e-01 1.47028297e-01 -5.33624649e-01
-9.19928849e-01 1.02548265e+00 3.37339640e-01 -3.52993794e-02
8.06358874e-01 -1.61871299e-01 -1.47200942e-01 -1.69670761e-01
-1.54724371e+00 3.35506380e-01 1.72070071e-01 2.81338364e-01
8.22372794e-01 2.60600005e-03 -5.15753329e-01 7.13936210e-01
-6.46331549e-01 -4.28587019e-01 8.82326305e-01 7.45419502e-01
-8.06874931e-01 -1.08186579e+00 -1.31512597e-01 4.29466009e-01
-3.85429651e-01 -2.19940811e-01 -3.35781634e-01 4.96514201e-01
-5.78425288e-01 1.15994871e+00 3.42755973e-01 -2.69843161e-01
-2.71125525e-01 3.53893131e-01 -3.89122337e-01 -1.19599950e+00
-6.08084142e-01 4.07581806e-01 1.34561270e-01 1.18887508e-02
-3.33713144e-01 -1.07467985e+00 -1.04481864e+00 1.73096135e-02
-3.92861187e-01 6.81939363e-01 8.29250932e-01 1.19514740e+00
3.94890577e-01 7.11995780e-01 8.36396039e-01 -5.73344946e-01
-1.17435955e-01 -1.67793465e+00 -6.09452963e-01 -4.85346988e-02
2.27968872e-01 -2.44638681e-01 4.50537633e-03 3.02155167e-01] | [10.40390396118164, 7.000944137573242] |
b7bec9fa-ca54-4d0b-8c48-d19e6bc9c566 | conservative-offline-distributional | 2107.06106 | null | https://arxiv.org/abs/2107.06106v2 | https://arxiv.org/pdf/2107.06106v2.pdf | Conservative Offline Distributional Reinforcement Learning | Many reinforcement learning (RL) problems in practice are offline, learning purely from observational data. A key challenge is how to ensure the learned policy is safe, which requires quantifying the risk associated with different actions. In the online setting, distributional RL algorithms do so by learning the distribution over returns (i.e., cumulative rewards) instead of the expected return; beyond quantifying risk, they have also been shown to learn better representations for planning. We propose Conservative Offline Distributional Actor Critic (CODAC), an offline RL algorithm suitable for both risk-neutral and risk-averse domains. CODAC adapts distributional RL to the offline setting by penalizing the predicted quantiles of the return for out-of-distribution actions. We prove that CODAC learns a conservative return distribution -- in particular, for finite MDPs, CODAC converges to an uniform lower bound on the quantiles of the return distribution; our proof relies on a novel analysis of the distributional Bellman operator. In our experiments, on two challenging robot navigation tasks, CODAC successfully learns risk-averse policies using offline data collected purely from risk-neutral agents. Furthermore, CODAC is state-of-the-art on the D4RL MuJoCo benchmark in terms of both expected and risk-sensitive performance. | ['Osbert Bastani', 'Dinesh Jayaraman', 'Yecheng Jason Ma'] | 2021-07-12 | null | http://proceedings.neurips.cc/paper/2021/hash/a05d886123a54de3ca4b0985b718fb9b-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/a05d886123a54de3ca4b0985b718fb9b-Paper.pdf | neurips-2021-12 | ['distributional-reinforcement-learning', 'd4rl'] | ['methodology', 'robots'] | [-3.99847507e-01 7.26447284e-01 -3.40111524e-01 -7.46070594e-02
-1.22615612e+00 -5.89389741e-01 4.52846438e-01 4.89463657e-01
-9.63317335e-01 9.57094908e-01 1.31832466e-01 -5.00034153e-01
-4.34456617e-01 -7.31077731e-01 -1.08985329e+00 -9.34620380e-01
-7.53013849e-01 7.12637722e-01 -1.36816546e-01 -1.51774511e-01
-5.01575060e-02 2.39200890e-01 -9.56807911e-01 -4.04857785e-01
8.18181634e-01 1.21070611e+00 1.10246902e-02 5.26110768e-01
4.88066018e-01 9.21318114e-01 -5.38065970e-01 -2.19997898e-01
5.74815750e-01 -3.28043848e-01 -3.67928445e-01 -6.00572586e-01
-2.84539878e-01 -7.09038079e-01 -2.12063000e-01 1.30890334e+00
4.43380237e-01 4.19549584e-01 8.75785649e-01 -1.33698940e+00
-1.85172603e-01 9.21340883e-01 -5.91141522e-01 -1.46944642e-01
-1.28834769e-01 3.99046391e-01 1.10169578e+00 -2.16113016e-01
2.55339146e-01 1.49791312e+00 3.68825793e-01 6.57363474e-01
-1.24010253e+00 -3.75255316e-01 3.93915474e-01 -4.25797820e-01
-7.44132280e-01 1.94593102e-01 5.19121401e-02 -5.74851811e-01
6.57460093e-01 -2.07465008e-01 4.59464073e-01 1.18194282e+00
5.85796058e-01 8.39429140e-01 1.17379701e+00 4.93134968e-02
9.01451111e-01 -1.31516397e-01 -4.10592228e-01 5.80906987e-01
4.01008517e-01 9.85681534e-01 -4.69894707e-02 -3.66605341e-01
5.28682053e-01 -1.42640173e-01 1.37556959e-02 -7.00555801e-01
-8.53339255e-01 1.01747537e+00 2.56116360e-01 -4.09387082e-01
-5.38419664e-01 6.39952242e-01 5.81295788e-01 5.28699636e-01
2.80608833e-01 4.79993552e-01 -5.12685776e-01 -3.26195419e-01
-1.93235725e-01 6.90603673e-01 9.19242084e-01 8.55714977e-01
2.09458828e-01 1.30038366e-01 -4.87155408e-01 3.22574317e-01
3.75052154e-01 8.49235177e-01 7.61481747e-02 -1.19059646e+00
8.23916376e-01 -1.08415997e-02 8.31080616e-01 -4.08785224e-01
-4.91379559e-01 -3.63592982e-01 -3.97752255e-01 8.38618398e-01
8.31088066e-01 -6.92918479e-01 -5.65487623e-01 2.27575803e+00
1.95851594e-01 -2.74786532e-01 3.80668163e-01 8.52008164e-01
-3.68798047e-01 5.80043554e-01 3.51540267e-01 -3.75518233e-01
7.81272233e-01 -5.28679013e-01 -4.18535799e-01 -1.17306516e-01
7.29451418e-01 2.18030587e-01 1.22773087e+00 6.07977808e-01
-1.01220202e+00 2.09001318e-01 -9.03129578e-01 2.84445047e-01
1.18308611e-01 -1.39575779e-01 1.09084837e-01 4.08532083e-01
-5.72358847e-01 8.36992562e-01 -1.05932498e+00 9.06255022e-02
5.85862458e-01 1.09599508e-01 2.57579423e-02 2.44915545e-01
-1.04682934e+00 1.02595627e+00 3.90219569e-01 -1.93596661e-01
-1.96934378e+00 -8.71727169e-01 -7.59970367e-01 1.92945123e-01
1.04704547e+00 -2.06464440e-01 1.73294628e+00 -7.37373531e-01
-2.02330995e+00 2.26449296e-01 8.32349300e-01 -1.19669783e+00
1.15095377e+00 -4.61189657e-01 2.65390277e-01 -6.52440405e-03
4.35700491e-02 2.63109475e-01 9.51577902e-01 -1.00766182e+00
-6.95878804e-01 -2.53496319e-01 4.04969841e-01 3.36210877e-01
8.26141983e-02 -4.18581665e-01 4.77005631e-01 -4.04689074e-01
-8.11319411e-01 -1.10440564e+00 -6.72294378e-01 2.68475227e-02
-8.42492282e-02 -3.07480305e-01 1.81188032e-01 -4.65268612e-01
7.44786978e-01 -2.07733107e+00 2.17370242e-01 1.84372604e-01
-2.48132139e-01 -7.77769983e-02 -6.40935227e-02 4.15191323e-01
3.09931129e-01 -1.63674459e-01 -4.85949665e-01 -2.49609113e-01
8.07445049e-01 3.18704188e-01 -8.34275782e-01 9.21213686e-01
-2.45468616e-02 5.15515685e-01 -1.30947912e+00 1.14943698e-01
-8.38384926e-02 -3.22565585e-02 -6.85913503e-01 4.33772206e-01
-7.49415874e-01 3.63402128e-01 -5.73844194e-01 1.16963476e-01
2.76671082e-01 5.73589385e-01 1.66525394e-01 6.51046574e-01
-1.22648031e-01 -1.00743718e-01 -8.62836897e-01 1.27993929e+00
-7.89974689e-01 2.07886964e-01 2.87082762e-01 -9.82946634e-01
6.31368041e-01 1.02184229e-01 4.04529721e-01 -5.00418544e-01
4.10622180e-01 3.94862980e-01 -1.88589394e-01 -1.68343410e-01
2.85558909e-01 -5.07324934e-01 -6.57804847e-01 6.97627544e-01
-4.33282927e-03 -3.45032990e-01 -1.46174565e-01 4.09660265e-02
1.01991713e+00 3.99073750e-01 4.40602094e-01 -5.99144697e-01
1.30967395e-02 -1.38411105e-01 5.73728561e-01 9.28586662e-01
-4.40784961e-01 -1.04878545e-01 1.40571392e+00 4.34820801e-02
-9.17576015e-01 -1.36192191e+00 2.39451855e-01 1.01210761e+00
4.32541370e-02 1.03996098e-02 -6.72997475e-01 -1.13487601e+00
5.22377312e-01 1.25515521e+00 -7.97077715e-01 -5.12251914e-01
-2.27793694e-01 -4.44258511e-01 4.92456615e-01 5.23580968e-01
8.21262971e-02 -7.97517419e-01 -1.36231899e+00 1.03068367e-01
4.11596179e-01 -5.64533472e-01 -5.61394691e-01 4.96760100e-01
-5.35622299e-01 -1.15026736e+00 -8.42526376e-01 -3.82739417e-02
5.51245213e-01 -3.29108894e-01 7.67317593e-01 -7.12661922e-01
6.80305213e-02 9.12321150e-01 -1.70332491e-02 -8.44891548e-01
-4.14953709e-01 -2.98355162e-01 3.36880326e-01 -1.39438450e-01
-2.41983876e-01 -2.97894299e-01 -6.33856714e-01 2.05396652e-01
-6.91870928e-01 -5.95301569e-01 3.51743102e-01 8.26678813e-01
6.54867649e-01 1.34330988e-01 7.25644886e-01 -7.00220168e-01
9.62871194e-01 -8.10354829e-01 -1.55545306e+00 1.95615813e-01
-6.02021873e-01 6.12946510e-01 9.97634470e-01 -4.76360559e-01
-1.14003062e+00 -1.74007878e-01 2.52259374e-01 -2.49758378e-01
3.43298972e-01 4.20174807e-01 -2.80904789e-02 4.24361199e-01
6.19074762e-01 -1.82302520e-01 2.62655020e-01 -3.30313444e-01
4.67254937e-01 3.18309814e-01 4.32691097e-01 -1.31259131e+00
6.69027984e-01 4.00823206e-01 4.37780559e-01 -3.90265971e-01
-9.65389192e-01 7.43303970e-02 1.22735411e-01 -3.31334919e-01
7.56451368e-01 -8.85753512e-01 -1.40170693e+00 2.57505178e-01
-5.64023972e-01 -1.14014256e+00 -9.34944212e-01 6.21287942e-01
-1.45879257e+00 2.38581430e-02 -2.66974568e-01 -1.46121013e+00
-8.05053264e-02 -9.05487955e-01 6.76880658e-01 1.01695649e-01
3.86076063e-01 -9.17914808e-01 3.87337357e-01 -4.49922055e-01
3.39129031e-01 5.96443534e-01 9.12407219e-01 -4.60854352e-01
-1.34171069e-01 1.38387010e-01 1.35260388e-01 4.69049037e-01
-3.87981087e-01 -3.83316636e-01 -6.32462084e-01 -5.72858214e-01
1.79172814e-01 -9.26926672e-01 8.64671946e-01 5.31970978e-01
1.05014396e+00 -7.92496026e-01 8.86235461e-02 3.23079884e-01
1.36938262e+00 4.05529857e-01 2.36543268e-01 5.53989768e-01
1.01616085e-01 7.95294881e-01 1.19467866e+00 9.76239145e-01
3.53595078e-01 4.40885127e-01 9.27319407e-01 7.78900623e-01
6.95825100e-01 -7.43280053e-01 1.05376768e+00 -2.32989579e-01
2.32725725e-01 -1.19566716e-01 -6.37767673e-01 3.63466233e-01
-2.18733549e+00 -9.18963194e-01 6.05015874e-01 2.68803048e+00
9.43504333e-01 2.69183934e-01 7.00147450e-01 -4.36841190e-01
2.95075774e-01 -1.56230256e-01 -1.22569978e+00 -5.97178519e-01
1.39359668e-01 -7.25405067e-02 1.22035742e+00 7.24480510e-01
-9.71248806e-01 5.21241009e-01 5.38721371e+00 7.82088459e-01
-8.31576765e-01 7.26315007e-02 7.09654570e-01 -4.06848669e-01
-1.32470518e-01 -1.40681550e-01 -5.69576442e-01 4.57730412e-01
1.09224916e+00 -5.50990283e-01 6.97946489e-01 1.45526922e+00
2.03474030e-01 -3.94915730e-01 -1.29679108e+00 5.09176612e-01
-6.56275988e-01 -8.44581127e-01 -6.83845997e-01 2.80312061e-01
6.77406847e-01 1.17618904e-01 4.06194091e-01 8.33637118e-01
1.14388001e+00 -1.05835867e+00 1.25868022e+00 6.08926833e-01
6.52703881e-01 -1.25566268e+00 7.39832520e-01 6.34247065e-01
-6.58158839e-01 -7.06044495e-01 -4.14796621e-01 -1.04714534e-03
1.08375430e-01 3.83762449e-01 -6.88166261e-01 2.82752365e-01
4.60179627e-01 3.70470464e-01 1.99849114e-01 7.80858278e-01
-4.72643226e-01 4.63247985e-01 -5.26379347e-01 -2.03235373e-01
6.11463785e-01 -3.67045492e-01 5.41372001e-01 8.13026905e-01
3.50095659e-01 -2.38686185e-02 4.30008829e-01 7.42570162e-01
8.27850401e-02 -2.01082170e-01 -6.67390466e-01 -2.68604279e-01
2.00548068e-01 8.82767737e-01 -2.99481332e-01 1.02254130e-01
1.21492647e-01 2.75460929e-01 5.98694086e-01 1.87704086e-01
-9.69614446e-01 -2.59454399e-01 8.19208741e-01 -3.85831803e-01
3.02654386e-01 -3.33740801e-01 1.83041897e-02 -7.88786471e-01
-3.64740752e-02 -7.44667053e-01 7.32890129e-01 -1.27736285e-01
-1.51195562e+00 8.95826221e-02 3.48410726e-01 -1.03472435e+00
-5.26497245e-01 -7.65996158e-01 -4.12726343e-01 6.25119507e-01
-1.37401080e+00 -3.47530216e-01 5.13040841e-01 4.08281237e-01
1.52654782e-01 -6.26648515e-02 4.80626017e-01 -2.39244565e-01
-4.98110890e-01 4.55755532e-01 4.90915477e-01 -2.31473520e-01
4.90750730e-01 -1.66841888e+00 1.52570372e-02 4.78686452e-01
-3.71735722e-01 1.22599769e-02 9.40205336e-01 -5.38903236e-01
-1.66829264e+00 -1.07706702e+00 -3.16514641e-01 -2.17928842e-01
9.73457336e-01 -2.11719349e-01 -3.51301879e-01 7.06979811e-01
-1.87352046e-01 9.52710435e-02 -1.36946165e-03 -1.83572203e-01
-2.96690106e-01 -1.23284236e-01 -1.49896491e+00 7.78298080e-01
7.40583003e-01 -2.14819640e-01 -4.53196615e-01 1.73011750e-01
9.03437793e-01 -4.99303460e-01 -6.63708448e-01 1.50011286e-01
4.34709817e-01 -7.36473680e-01 5.28112948e-01 -9.80395436e-01
1.77495137e-01 -1.07412226e-01 -3.04542065e-01 -1.80715311e+00
2.71235764e-01 -1.04792356e+00 -3.41533899e-01 7.56045640e-01
1.55656889e-01 -9.05011535e-01 3.62771034e-01 4.93537396e-01
-6.62302598e-02 -8.86235952e-01 -1.29603529e+00 -1.37261975e+00
6.17975235e-01 -3.94994497e-01 3.37004155e-01 1.96714342e-01
2.34757662e-01 -1.58986822e-01 -4.41864789e-01 1.98517084e-01
1.10244596e+00 -1.37495741e-01 6.49494946e-01 -7.44056702e-01
-6.21695518e-01 -5.85794508e-01 1.49349287e-01 -7.77322471e-01
5.28947711e-01 -6.37623608e-01 5.82570493e-01 -1.08670127e+00
-8.81987587e-02 -6.02495968e-01 -2.39712253e-01 2.77803600e-01
2.30094835e-01 -7.86912441e-01 4.44152027e-01 -3.46492678e-01
-6.58355355e-01 1.11429095e+00 1.15084779e+00 -1.37255907e-01
-2.82656580e-01 3.05597216e-01 -5.07288337e-01 7.85256207e-01
9.45234656e-01 -6.21525586e-01 -7.54116237e-01 -1.48966908e-01
4.85383838e-01 4.74208117e-01 2.40884617e-01 -6.83008552e-01
-1.75950468e-01 -7.50743628e-01 -2.81012714e-01 -3.67875844e-01
2.51620123e-03 -7.73173332e-01 -2.43822038e-01 9.64829206e-01
-9.43851471e-01 -4.48873378e-02 -3.04170270e-02 1.24038374e+00
3.91184241e-01 -2.55412668e-01 9.52172101e-01 -3.83053459e-02
-1.58507243e-01 3.90788287e-01 -5.06955087e-01 7.58341372e-01
1.24130893e+00 8.10865760e-01 -2.40362570e-01 -6.75668418e-01
-4.36011344e-01 7.39097476e-01 1.71905294e-01 9.20184422e-03
3.64232570e-01 -9.03636217e-01 -5.10538757e-01 -1.37840360e-01
-5.50413691e-02 1.83676004e-01 1.02839254e-01 7.78167486e-01
-1.36295557e-01 2.13091210e-01 -1.12568855e-01 -1.55397356e-01
-3.01860958e-01 7.31436670e-01 6.74796879e-01 -5.12711346e-01
-5.63697815e-01 6.17626727e-01 4.44044143e-01 -3.83930653e-01
7.36279249e-01 -7.19235957e-01 2.79844373e-01 1.16388090e-01
4.79015559e-01 6.44925416e-01 -2.92018682e-01 1.09557934e-01
-4.02104892e-02 -3.86839062e-02 1.94986194e-01 -7.78847277e-01
1.45902848e+00 3.68955322e-02 3.85423124e-01 5.59073150e-01
8.21795821e-01 -1.73398033e-01 -2.37486339e+00 -3.22944149e-02
3.81731272e-01 -2.97356933e-01 -1.32644296e-01 -1.02228713e+00
-8.73228669e-01 8.50099802e-01 4.42258000e-01 1.65691033e-01
7.02226877e-01 -1.77070469e-01 4.30776238e-01 5.96747160e-01
9.18298364e-01 -1.52061033e+00 1.43247530e-01 7.09822834e-01
1.03209651e+00 -1.00541854e+00 -1.20247267e-01 6.11193061e-01
-1.21709788e+00 8.90482605e-01 4.16827291e-01 -5.23516059e-01
6.77654862e-01 3.02040696e-01 -2.90820330e-01 2.43300512e-01
-1.04979789e+00 -4.95766811e-02 -2.24471956e-01 6.22003555e-01
-3.25836301e-01 5.79600692e-01 -1.57224223e-01 9.32810724e-01
-1.00163862e-01 -2.36513257e-01 7.24927068e-01 9.59038317e-01
-4.83255208e-01 -6.89553320e-01 -1.64740816e-01 1.94309980e-01
-4.50773627e-01 3.93664002e-01 6.92442060e-02 7.49531925e-01
-4.72180068e-01 6.98116779e-01 -1.27344817e-01 1.24899969e-01
3.23521942e-01 -2.75678158e-01 5.44792593e-01 -4.51244175e-01
-2.47845590e-01 -2.06531972e-01 -2.28873212e-02 -7.74626613e-01
1.77653521e-01 -7.46578515e-01 -1.39027238e+00 -1.66724056e-01
1.71886072e-01 2.89215893e-01 6.16284490e-01 8.97303700e-01
5.21082170e-02 2.17082411e-01 8.78475368e-01 -7.87527323e-01
-2.02436161e+00 -6.93556845e-01 -7.59202838e-01 -6.72282353e-02
6.24380171e-01 -9.30410743e-01 -6.60275161e-01 -7.51956582e-01] | [4.243551731109619, 2.471790075302124] |
f44e3f7a-575c-489c-90e5-f43a580a9934 | dynint-dynamic-interaction-modeling-for-large | 2301.08139 | null | https://arxiv.org/abs/2301.08139v1 | https://arxiv.org/pdf/2301.08139v1.pdf | DynInt: Dynamic Interaction Modeling for Large-scale Click-Through Rate Prediction | Learning feature interactions is the key to success for the large-scale CTR prediction in Ads ranking and recommender systems. In industry, deep neural network-based models are widely adopted for modeling such problems. Researchers proposed various neural network architectures for searching and modeling the feature interactions in an end-to-end fashion. However, most methods only learn static feature interactions and have not fully leveraged deep CTR models' representation capacity. In this paper, we propose a new model: DynInt. By extending Polynomial-Interaction-Network (PIN), which learns higher-order interactions recursively to be dynamic and data-dependent, DynInt further derived two modes for modeling dynamic higher-order interactions: dynamic activation and dynamic parameter. In dynamic activation mode, we adaptively adjust the strength of learned interactions by instance-aware activation gating networks. In dynamic parameter mode, we re-parameterize the parameters by different formulations and dynamically generate the parameters by instance-aware parameter generation networks. Through instance-aware gating mechanism and dynamic parameter generation, we enable the PIN to model dynamic interaction for potential industry applications. We implement the proposed model and evaluate the model performance on real-world datasets. Extensive experiment results demonstrate the efficiency and effectiveness of DynInt over state-of-the-art models. | ['Liubo Li', 'YaChen Yan'] | 2023-01-03 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [-3.26960415e-01 -4.67281252e-01 -4.72605705e-01 -7.02660620e-01
-3.12737703e-01 -7.12471962e-01 4.93995398e-01 -3.19041789e-01
-2.30529726e-01 3.20186704e-01 9.59631056e-02 -3.70989054e-01
-5.04684210e-01 -8.23588073e-01 -7.16465950e-01 -4.88608241e-01
-3.01446974e-01 7.17910051e-01 4.33853328e-01 -5.05252004e-01
2.13087544e-01 2.58814722e-01 -1.45738697e+00 6.32450938e-01
7.67617166e-01 1.08035779e+00 1.68160692e-01 4.32570308e-01
-2.82364309e-01 4.28124875e-01 -1.79542616e-01 -2.50217080e-01
4.47050452e-01 1.04460932e-01 -4.74316835e-01 -5.71990550e-01
-1.03834085e-02 -3.68583322e-01 -4.78973240e-01 5.66738367e-01
4.74188775e-01 3.71571273e-01 3.99065942e-01 -1.07247543e+00
-9.90562260e-01 1.15463185e+00 -5.51426888e-01 3.88613552e-01
3.69840749e-02 2.82178223e-01 1.44363678e+00 -8.09037924e-01
2.86774695e-01 1.49628639e+00 2.69703895e-01 5.02272129e-01
-1.15276182e+00 -1.03892350e+00 9.95461285e-01 2.70044267e-01
-1.04568398e+00 4.71355021e-02 8.60751271e-01 -3.27309251e-01
9.58410561e-01 1.86465427e-01 7.23381102e-01 1.06446910e+00
4.15915281e-01 9.00318563e-01 5.99443793e-01 -6.92492584e-03
-1.05887905e-01 -5.98751307e-02 6.51527643e-01 3.90160263e-01
-1.12011686e-01 3.29759181e-01 -2.50456721e-01 -4.55674887e-01
1.08630085e+00 3.89878273e-01 2.92005762e-02 -1.67729929e-01
-9.70939517e-01 1.13847423e+00 7.20775545e-01 1.31189182e-01
-1.97269112e-01 4.81584877e-01 4.20051396e-01 5.39794087e-01
2.12763906e-01 4.69380587e-01 -9.83592212e-01 -4.52882573e-02
-8.95225108e-02 4.95553583e-01 6.00760162e-01 7.65014887e-01
4.27181154e-01 -1.87264353e-01 -3.56604874e-01 1.12148404e+00
5.07698894e-01 1.62595227e-01 7.09441543e-01 -4.53440934e-01
2.47801885e-01 8.45514476e-01 -2.07547229e-02 -9.57697093e-01
-3.94267023e-01 -8.05158317e-01 -6.89569294e-01 -4.74037766e-01
-5.18429466e-02 -1.52047262e-01 -9.89972234e-01 1.90243244e+00
3.79499286e-01 4.88461941e-01 -3.08290124e-01 8.53984475e-01
8.58679354e-01 6.25403821e-01 2.03000605e-01 5.79235740e-02
1.22863317e+00 -1.39899528e+00 -5.17656982e-01 -1.70704387e-02
6.12023294e-01 -4.06240135e-01 1.40181494e+00 6.00387633e-01
-7.88749695e-01 -8.18248570e-01 -1.05817854e+00 2.15102509e-01
-4.04982775e-01 1.09889947e-01 1.28830802e+00 1.49778202e-01
-8.53426218e-01 6.87846720e-01 -8.18543613e-01 1.16801284e-01
2.34304368e-01 1.14198101e+00 1.99000195e-01 1.53201401e-01
-1.61163592e+00 1.69767991e-01 9.79612619e-02 6.01396441e-01
-8.27359200e-01 -7.43608117e-01 -3.22503299e-01 3.09649557e-01
3.15806121e-01 -6.87603831e-01 1.46834075e+00 -7.80464709e-01
-1.66291833e+00 2.65349261e-02 2.39079401e-01 -3.85265261e-01
9.75443348e-02 -4.75957304e-01 -7.85948396e-01 -4.62578773e-01
-5.11943042e-01 5.35748422e-01 8.43712151e-01 -1.27186179e+00
-4.22504634e-01 -8.65567997e-02 6.01968706e-01 1.82756320e-01
-4.82609898e-01 -6.00410588e-02 -7.65087008e-01 -3.59731287e-01
-1.40677392e-01 -1.01525235e+00 -6.22605801e-01 -2.58136153e-01
-2.87214398e-01 -5.03970385e-01 8.44494462e-01 -4.31591757e-02
1.69398749e+00 -2.18133402e+00 -8.80639032e-02 4.74161297e-01
4.73908901e-01 2.95339882e-01 -6.31520092e-01 4.17668402e-01
-5.58059551e-02 5.72954230e-02 4.67876852e-01 -9.09687877e-02
1.01657197e-01 3.92679513e-01 -4.71164227e-01 -2.62169719e-01
-1.84607670e-01 1.20961094e+00 -8.24694633e-01 -9.75511074e-02
1.95886642e-01 7.41472661e-01 -1.04239643e+00 1.85254037e-01
-7.06918538e-01 4.22310412e-01 -1.08772123e+00 4.62819993e-01
6.62588298e-01 -6.96710885e-01 4.02992845e-01 -4.64384943e-01
1.19157426e-01 1.18575945e-01 -1.24152839e+00 1.15109515e+00
-7.52190650e-01 8.38270262e-02 -1.91997826e-01 -6.89684451e-01
9.16417420e-01 -1.87696606e-01 4.23508584e-01 -7.88751721e-01
4.47619885e-01 1.27704054e-01 3.86389881e-01 -2.81859010e-01
5.58505297e-01 5.07404864e-01 -9.67458412e-02 2.98030704e-01
-3.56383771e-01 7.23116159e-01 -5.08274473e-02 1.14134192e-01
1.02711558e+00 -1.35294020e-01 -3.56419593e-01 -8.85568932e-03
6.78064108e-01 -4.72278953e-01 4.29149300e-01 9.66859818e-01
2.19707102e-01 1.04597844e-01 3.70587081e-01 -7.47224748e-01
-3.77758235e-01 -8.07723343e-01 -1.02705456e-01 1.72491992e+00
2.67134488e-01 -5.33334613e-01 -2.88225085e-01 -6.92077637e-01
2.07608014e-01 4.15305078e-01 -8.24473917e-01 -3.50370795e-01
-9.60649252e-01 -9.12257433e-01 -6.36502281e-02 6.42075181e-01
4.16699588e-01 -1.21745110e+00 4.68000188e-04 4.59962368e-01
2.70112634e-01 -7.58848965e-01 -8.03876936e-01 2.98038691e-01
-8.38338077e-01 -8.74600530e-01 -9.88558084e-02 -8.32409322e-01
4.43389237e-01 1.32523820e-01 1.06086636e+00 4.61236894e-01
-6.45108745e-02 4.86022457e-02 -4.74941701e-01 -1.24877147e-01
2.68035293e-01 3.96243602e-01 -1.09344982e-01 9.18233022e-02
3.18878472e-01 -8.71114194e-01 -1.07032263e+00 6.60587072e-01
-1.04064059e+00 -3.30692142e-01 7.70973861e-01 9.98851836e-01
6.59643948e-01 -1.39733180e-01 7.46136248e-01 -1.35733771e+00
1.07436490e+00 -5.84768593e-01 -6.13053203e-01 1.15692891e-01
-8.78582478e-01 1.15228437e-01 9.39705551e-01 -1.20570087e+00
-8.38679016e-01 6.55989945e-02 -2.16467187e-01 -4.78729784e-01
3.66867363e-01 9.26412344e-01 -4.53045331e-02 -1.32863417e-01
4.45705444e-01 -1.23421261e-02 -4.26368117e-01 -6.56153381e-01
5.64094722e-01 4.62371260e-01 1.02349870e-01 -9.13703024e-01
5.99072456e-01 2.74892934e-02 -4.36704785e-01 -8.64498690e-02
-9.97464120e-01 -2.35893071e-01 -2.78651327e-01 6.75121471e-02
4.96763706e-01 -9.20345664e-01 -1.10103166e+00 3.49685580e-01
-7.85699308e-01 -5.05710661e-01 -4.38468903e-02 2.80498356e-01
-1.26837203e-02 -2.43989918e-02 -6.96666181e-01 -4.22861069e-01
-5.87532103e-01 -1.28538001e+00 8.72054815e-01 3.88113797e-01
3.32740009e-01 -9.80360329e-01 1.80826262e-01 7.86146298e-02
7.86756277e-01 -9.84023884e-02 1.18065763e+00 -9.89857376e-01
-8.11255991e-01 -3.05190772e-01 -1.32637233e-01 5.39725088e-02
-1.57603711e-01 4.42630565e-03 -5.82605898e-01 -1.97169051e-01
-6.44826472e-01 -1.39336690e-01 7.29527533e-01 5.07130921e-01
1.81998348e+00 -3.99821341e-01 -6.62970245e-01 5.97664118e-01
1.20910895e+00 4.22913790e-01 5.39569318e-01 1.83540612e-01
7.93752372e-01 -3.41813639e-02 6.20588303e-01 3.96153003e-01
4.73900139e-01 9.35751796e-01 5.22203803e-01 -5.77882454e-02
4.47680920e-01 -4.35686976e-01 1.74157247e-01 8.75484407e-01
7.12462068e-02 -5.05232155e-01 -3.87475610e-01 1.61270797e-01
-2.24745297e+00 -7.03183591e-01 -9.41251311e-03 2.01306057e+00
7.57674694e-01 6.29179001e-01 2.19380215e-01 -5.46402335e-01
4.96825099e-01 9.23029408e-02 -9.64433789e-01 -4.08542901e-01
2.57282704e-01 2.76708752e-01 2.99381554e-01 4.49634194e-01
-8.15467417e-01 1.24880111e+00 6.03586292e+00 8.82070720e-01
-1.32801974e+00 -2.79718749e-02 5.53725898e-01 -2.01889366e-01
-5.13436019e-01 3.58036458e-02 -1.15465879e+00 4.49417293e-01
9.06017005e-01 -2.40552407e-02 6.08476877e-01 1.06309330e+00
1.86973423e-01 6.60508752e-01 -1.19028330e+00 8.94465148e-01
-5.00519335e-01 -1.28456950e+00 5.12921333e-01 3.23450565e-01
4.86013383e-01 1.26989260e-01 3.52325171e-01 1.01156211e+00
7.31175721e-01 -8.59417021e-01 2.51054406e-01 4.98336315e-01
4.28570777e-01 -8.22114468e-01 7.28959084e-01 5.64686917e-02
-1.53420830e+00 -7.73673475e-01 -3.57599437e-01 -9.71026067e-03
2.62461752e-01 4.30339158e-01 -3.37563902e-01 3.68841618e-01
5.42947233e-01 8.73823166e-01 -5.66049993e-01 8.50263000e-01
-1.80092603e-02 7.49735832e-01 -2.99760222e-01 -3.16708714e-01
3.33102047e-01 -2.10892200e-01 9.27650854e-02 9.03328598e-01
4.68060412e-02 2.75108844e-01 5.82295835e-01 5.02025783e-01
-1.55258864e-01 1.32448375e-01 5.32573694e-03 -2.73902148e-01
7.58027196e-01 1.52634001e+00 -3.28660786e-01 1.66947208e-02
-2.08445445e-01 5.40056944e-01 4.21277821e-01 5.00723362e-01
-1.16875303e+00 -1.55158445e-01 6.65956140e-01 4.08825427e-01
6.38129413e-01 -2.03297406e-01 2.52975225e-01 -9.24919963e-01
-2.26108462e-01 -8.37780237e-01 5.33713937e-01 -5.01663744e-01
-1.65448010e+00 9.46478784e-01 2.24544317e-01 -1.14900839e+00
-1.47131011e-01 -6.94700837e-01 -9.33796465e-01 4.72295135e-01
-1.20104384e+00 -1.54424119e+00 -1.83676988e-01 7.20045567e-01
5.05290329e-01 -1.31178454e-01 5.92115760e-01 5.40745080e-01
-8.01175117e-01 1.09457147e+00 2.16143027e-01 -1.10595331e-01
5.97222149e-01 -9.12683666e-01 4.70850259e-01 5.38152270e-02
-4.87526208e-02 1.26390159e+00 4.45241541e-01 -4.18291599e-01
-1.54874229e+00 -9.98382211e-01 3.82100642e-01 -1.38957247e-01
8.23394835e-01 -7.36385405e-01 -6.70792341e-01 8.03355038e-01
-2.68603247e-02 1.74566999e-01 6.82683289e-01 6.84212208e-01
-3.76636922e-01 -4.49099034e-01 -8.48706841e-01 5.71299374e-01
1.29043376e+00 -2.04979241e-01 5.44866845e-02 4.36614931e-01
1.26191115e+00 -4.78119224e-01 -8.05539370e-01 4.97481048e-01
9.67944860e-01 -4.78654593e-01 1.15610754e+00 -9.54083085e-01
3.43405128e-01 -1.16864435e-01 -1.42707601e-01 -1.06073248e+00
-8.58099818e-01 -5.26997745e-01 -5.17759502e-01 9.45989251e-01
5.68649530e-01 -8.43164444e-01 7.21934497e-01 6.50785089e-01
-2.38882482e-01 -1.53739119e+00 -3.17388743e-01 -3.45168591e-01
-1.03339039e-01 -1.30533233e-01 9.34206843e-01 8.29994857e-01
-2.94479966e-01 6.71930254e-01 -5.00847399e-01 -4.02154997e-02
3.70565765e-02 2.99992770e-01 7.19482958e-01 -1.50181341e+00
-9.72681642e-01 -2.81501949e-01 -5.96879646e-02 -1.71068954e+00
1.70911908e-01 -7.09316730e-01 -3.23761106e-01 -1.27298844e+00
3.58463734e-01 -8.59144151e-01 -9.85059023e-01 5.39936364e-01
-3.44105512e-02 -1.80986226e-01 -6.19752780e-02 2.46990383e-01
-8.31377149e-01 5.71647048e-01 1.40112615e+00 -4.28323895e-02
-7.30368555e-01 2.06681237e-01 -1.00573516e+00 3.93933237e-01
6.75865114e-01 -3.00429016e-01 -1.02425265e+00 -5.68989933e-01
4.59956318e-01 -8.71732533e-02 2.31925547e-01 -4.51464176e-01
1.29473567e-01 -2.37308741e-01 3.46125931e-01 -8.38649154e-01
3.30874622e-01 -7.66687691e-01 2.44005337e-01 2.00150505e-01
-7.83264279e-01 3.70340288e-01 2.04856530e-01 9.62929964e-01
9.28638652e-02 1.79419264e-01 3.14933658e-01 -4.98471670e-02
-5.22886455e-01 8.27606678e-01 -8.61409605e-02 -1.73943534e-01
7.16970742e-01 1.02074973e-01 -3.24942052e-01 -3.95637751e-01
-9.32891369e-01 6.30749166e-01 -2.40492985e-01 8.61242354e-01
5.05682349e-01 -1.39642620e+00 -1.51831955e-01 2.04147935e-01
1.53189257e-01 -1.01842247e-01 4.91372287e-01 5.25195241e-01
4.89677452e-02 3.38218629e-01 1.90033227e-01 -4.69510406e-01
-1.09105837e+00 6.20527864e-01 3.28335732e-01 -8.59831512e-01
-3.82687122e-01 9.80532587e-01 5.24607360e-01 -6.86794758e-01
3.51361930e-01 -4.02135372e-01 -5.55179715e-01 -4.37751889e-01
5.58245540e-01 -2.27704242e-01 -1.02913342e-01 -4.65415381e-02
-1.94359750e-01 3.60908866e-01 -1.03556085e+00 1.38917685e-01
1.44799590e+00 1.23011218e-02 1.82511404e-01 1.93298593e-01
1.10454845e+00 -1.78985611e-01 -9.75510180e-01 -3.16899866e-01
-4.02196020e-01 -4.15661454e-01 2.66937673e-01 -1.07114124e+00
-1.45726621e+00 4.56775457e-01 8.75841141e-01 -6.67915270e-02
1.04523098e+00 -9.20630917e-02 1.16838360e+00 7.47958362e-01
2.52926797e-01 -9.55946088e-01 3.87629122e-01 6.92698836e-01
6.93479002e-01 -8.65519702e-01 -2.38530099e-01 -4.31157768e-01
-4.77208912e-01 8.95746648e-01 1.22631121e+00 -3.70803505e-01
1.38119090e+00 1.82650477e-01 -1.13583356e-01 -4.28872675e-01
-1.30875552e+00 3.21961999e-01 4.20949787e-01 2.38537699e-01
5.26550770e-01 -3.69754918e-02 -5.73003113e-01 1.18782771e+00
7.17476606e-02 8.25957507e-02 -3.81268468e-03 8.30525815e-01
-3.56478781e-01 -1.61277807e+00 3.09652001e-01 9.42092836e-01
-1.74273461e-01 -3.68228257e-01 -1.95130304e-01 6.87829316e-01
2.97577735e-02 6.98579788e-01 -3.10962368e-02 -1.06212652e+00
5.50747395e-01 -3.62027586e-01 2.56204247e-01 -4.24879670e-01
-9.94903207e-01 2.74635375e-01 3.82031575e-02 -6.43819690e-01
6.13855086e-02 -3.60439926e-01 -1.33907437e+00 -1.38538748e-01
-7.32906997e-01 2.97483683e-01 4.58372056e-01 7.22963393e-01
8.33342552e-01 7.55700290e-01 9.49839234e-01 -7.33461618e-01
-9.78534579e-01 -9.40180779e-01 -3.77283126e-01 1.67595071e-03
9.36816558e-02 -9.98223126e-01 -2.50919580e-01 -3.42719495e-01] | [10.136177062988281, 5.4773359298706055] |
121ed9c9-3da5-460c-bd8b-77370638cc26 | x-cal-explicit-calibration-for-survival-1 | 2101.05346 | null | https://arxiv.org/abs/2101.05346v1 | https://arxiv.org/pdf/2101.05346v1.pdf | X-CAL: Explicit Calibration for Survival Analysis | Survival analysis models the distribution of time until an event of interest, such as discharge from the hospital or admission to the ICU. When a model's predicted number of events within any time interval is similar to the observed number, it is called well-calibrated. A survival model's calibration can be measured using, for instance, distributional calibration (D-CALIBRATION) [Haider et al., 2020] which computes the squared difference between the observed and predicted number of events within different time intervals. Classically, calibration is addressed in post-training analysis. We develop explicit calibration (X-CAL), which turns D-CALIBRATION into a differentiable objective that can be used in survival modeling alongside maximum likelihood estimation and other objectives. X-CAL allows practitioners to directly optimize calibration and strike a desired balance between predictive power and calibration. In our experiments, we fit a variety of shallow and deep models on simulated data, a survival dataset based on MNIST, on length-of-stay prediction using MIMIC-III data, and on brain cancer data from The Cancer Genome Atlas. We show that the models we study can be miscalibrated. We give experimental evidence on these datasets that X-CAL improves D-CALIBRATION without a large decrease in concordance or likelihood. | ['Rajesh Ranganath', 'Adler J. Perotte', 'Aahlad Puli', 'Xintian Han', 'Mark Goldstein'] | 2021-01-13 | x-cal-explicit-calibration-for-survival | http://proceedings.neurips.cc/paper/2020/hash/d4a93297083a23cc099f7bd6a8621131-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/d4a93297083a23cc099f7bd6a8621131-Paper.pdf | neurips-2020-12 | ['length-of-stay-prediction'] | ['medical'] | [-3.62573341e-02 1.10107400e-01 -3.68893147e-01 -9.31675732e-01
-1.20437789e+00 -4.17185903e-01 3.46512169e-01 6.04022682e-01
-6.16352439e-01 9.91204083e-01 4.50418770e-01 -5.81409931e-01
-2.58899450e-01 -8.04704070e-01 -6.14415407e-01 -7.33728707e-01
-4.83232200e-01 7.70950496e-01 -1.92984611e-01 1.66067481e-01
-1.00842342e-01 4.02326703e-01 -6.60103381e-01 -8.15207362e-02
6.23724878e-01 7.23509431e-01 -2.69445777e-01 6.85821116e-01
2.89779484e-01 4.58834112e-01 -4.00041938e-01 -2.91514963e-01
1.02769453e-02 -3.61760408e-01 -6.54450238e-01 -4.70405996e-01
-5.11467457e-02 -1.61656648e-01 -4.35010493e-01 7.24443913e-01
5.16919613e-01 -2.19540566e-01 1.15124214e+00 -1.50158215e+00
-1.95422605e-01 7.86459088e-01 -4.50512230e-01 3.30516756e-01
-8.27839002e-02 6.04579337e-02 7.23595858e-01 -3.64563197e-01
2.64261782e-01 1.08001876e+00 1.23814368e+00 4.71510321e-01
-1.56657577e+00 -5.61131835e-01 -4.03279662e-01 -2.69720733e-01
-1.37593210e+00 -1.71582401e-01 8.45471919e-02 -7.61534750e-01
7.01363564e-01 1.48580164e-01 5.39067626e-01 1.13654804e+00
9.82812405e-01 4.15659636e-01 9.50234175e-01 -2.13827133e-01
5.52111149e-01 -1.32541567e-01 4.18893218e-01 3.13094676e-01
1.49174258e-01 8.82226527e-01 -1.88890621e-01 -5.44501185e-01
7.61688888e-01 1.63495913e-01 -2.86322355e-01 -2.67489552e-01
-1.22200251e+00 1.09605646e+00 4.45994914e-01 -3.97597663e-02
-2.96022207e-01 3.83660853e-01 5.31155586e-01 1.94021791e-01
5.00220895e-01 2.20494211e-01 -7.47958302e-01 -2.57625952e-02
-1.00362778e+00 6.22210987e-02 7.69012392e-01 7.88913846e-01
3.11543584e-01 -4.22934204e-01 -4.05918032e-01 5.14334857e-01
1.24989785e-02 3.00794303e-01 4.08926517e-01 -8.66685092e-01
4.74691913e-02 3.42987299e-01 5.98518066e-02 -2.31565565e-01
-1.10979652e+00 -8.24827552e-01 -1.21350443e+00 -4.25078198e-02
6.77193165e-01 -3.76929402e-01 -8.75852406e-01 2.26208258e+00
-4.09931429e-02 2.50967801e-01 2.10475266e-01 5.95893085e-01
4.93486345e-01 4.31114197e-01 4.21436012e-01 -4.32713330e-01
1.22840905e+00 -4.24286932e-01 -5.36934495e-01 -2.16076195e-01
1.16850340e+00 -2.82800585e-01 9.13289845e-01 1.93986639e-01
-9.04801726e-01 -4.48010080e-02 -6.81312263e-01 1.72822118e-01
-7.26593733e-02 -8.49347413e-02 4.99882638e-01 7.21849322e-01
-9.98656452e-01 6.88789845e-01 -1.16269124e+00 -5.40225744e-01
6.15750790e-01 3.62138122e-01 -3.96583200e-01 3.76130193e-02
-1.18244171e+00 1.06050038e+00 5.51732123e-01 -4.46659088e-01
-1.11672521e+00 -1.06959701e+00 -5.95718384e-01 2.16697872e-01
4.31906655e-02 -1.17993498e+00 1.21747696e+00 -8.65977407e-01
-8.44824791e-01 9.88514781e-01 -1.80764310e-03 -8.55217040e-01
6.72543526e-01 -3.73217389e-02 -3.54951829e-01 -3.89390379e-01
2.84688510e-02 5.29576302e-01 2.18239799e-01 -6.90985799e-01
-4.58962172e-01 -6.94099009e-01 -2.57084370e-01 -9.41606089e-02
-7.98626915e-02 -1.20573808e-02 -1.37646139e-01 -6.70785666e-01
-1.24849379e-01 -9.67918396e-01 -5.12700677e-01 -1.04060292e-01
-4.12937164e-01 -2.72113122e-02 8.53113234e-02 -7.86044955e-01
1.23283529e+00 -2.10994339e+00 9.52218324e-02 2.16784835e-01
2.44562194e-01 -4.64385629e-01 -1.83235705e-02 2.66582429e-01
-6.19510651e-01 9.18019563e-02 -6.74966693e-01 -4.98281717e-01
-2.92348862e-01 3.31992298e-01 -1.98789146e-02 9.20411408e-01
2.66175158e-02 8.95033121e-01 -7.92524397e-01 -3.97841096e-01
1.44952107e-02 5.30648887e-01 -6.25574052e-01 2.98222840e-01
1.50525585e-01 5.95196486e-01 1.24268141e-02 2.89086401e-01
3.48681033e-01 -5.45113623e-01 3.04591730e-02 2.37103954e-01
1.86800033e-01 -4.37231734e-02 -4.72624362e-01 1.44899225e+00
-3.65122080e-01 5.43364048e-01 -5.11695504e-01 -1.13032055e+00
7.20866323e-01 2.95445591e-01 5.17524242e-01 -3.49754959e-01
4.88823473e-01 5.29165864e-02 -4.23547700e-02 -5.49221896e-02
-7.21645504e-02 -6.32267833e-01 -3.37959051e-01 2.33278781e-01
1.86684579e-02 1.56802833e-01 -2.76256859e-01 1.90170705e-01
1.27254343e+00 -2.66072989e-01 8.54988098e-01 -7.00025499e-01
1.16707146e-01 1.64375708e-01 6.00149810e-01 7.46699870e-01
-9.37922671e-02 5.47441125e-01 9.53054190e-01 -3.15950751e-01
-1.22243321e+00 -1.39885139e+00 -8.95031631e-01 8.26079488e-01
-5.44309974e-01 -1.59851089e-01 -8.05689692e-01 -6.16362095e-01
1.81246281e-01 1.04160976e+00 -1.15502501e+00 -5.33985376e-01
-2.01978713e-01 -1.20691752e+00 7.66620278e-01 7.82245517e-01
-2.37819135e-01 -5.89870691e-01 -6.64294839e-01 1.36135101e-01
-9.80270803e-02 -3.76402140e-01 -5.95735431e-01 7.18147278e-01
-1.23369479e+00 -1.17859602e+00 -8.60804081e-01 -4.46614325e-01
7.36336708e-01 -4.89924610e-01 1.46428537e+00 -1.64683610e-01
-1.15131751e-01 1.51544511e-01 -4.05645519e-02 -4.08090472e-01
-6.54102325e-01 7.60907531e-02 1.78738832e-01 -4.09090936e-01
3.67944956e-01 -4.29278374e-01 -6.81169391e-01 3.08964133e-01
-8.27022195e-01 -8.29706192e-02 4.01830256e-01 1.28199220e+00
8.13615680e-01 -3.29637289e-01 8.69291723e-01 -1.05986953e+00
4.40842420e-01 -8.71518850e-01 -7.07127094e-01 2.71809518e-01
-9.09958243e-01 2.10277170e-01 5.35654545e-01 -3.24920118e-01
-3.67908686e-01 9.42187235e-02 7.83345476e-03 -3.82551253e-01
-3.10751535e-02 9.36840951e-01 -4.39649858e-02 4.67927396e-01
7.88791418e-01 1.38285058e-02 -4.14341912e-02 -1.83559567e-01
1.32350788e-01 5.37890732e-01 8.44246864e-01 -3.28862160e-01
1.23832196e-01 2.74847746e-01 3.54406118e-01 -3.44450861e-01
-9.37072873e-01 -3.50636572e-01 -7.10208893e-01 1.70508623e-02
7.64160454e-01 -9.64222968e-01 -7.69822359e-01 2.82856792e-01
-7.30740607e-01 -6.48246467e-01 -4.25803691e-01 6.48589551e-01
-7.23289847e-01 1.23158112e-01 -3.73786628e-01 -5.59062243e-01
-3.29010844e-01 -8.92520845e-01 9.00900543e-01 7.01525509e-02
-5.32129526e-01 -1.55058849e+00 4.04018402e-01 -2.48407319e-01
2.57365227e-01 7.28567004e-01 1.28580713e+00 -8.91009212e-01
1.69322312e-01 -4.39393759e-01 -1.19107552e-01 2.13493466e-01
7.17033297e-02 -1.68627650e-01 -7.49256849e-01 -7.19560742e-01
-1.65620819e-01 -1.35398865e-01 9.11800385e-01 1.15597141e+00
1.48889971e+00 -2.57709354e-01 -6.26522362e-01 9.80332851e-01
1.42699671e+00 2.32248038e-01 6.74813449e-01 2.60665357e-01
1.55512899e-01 4.31175888e-01 5.02455354e-01 5.68139732e-01
4.92092371e-01 5.29177189e-01 3.12002242e-01 -1.06295429e-01
3.40261996e-01 -2.69775689e-01 2.42565945e-03 3.48283529e-01
3.56743991e-01 -3.27707708e-01 -1.23854446e+00 5.98458886e-01
-1.76027656e+00 -6.80149555e-01 -3.52491766e-01 2.79276276e+00
9.49517131e-01 2.40228891e-01 2.63535380e-01 -2.72092726e-02
5.53874373e-01 -5.46133697e-01 -8.86296511e-01 -8.39369223e-02
-2.59204388e-01 7.79936910e-02 8.78951192e-01 6.57472253e-01
-1.00616312e+00 3.54185522e-01 7.50244045e+00 5.13027072e-01
-9.28056121e-01 3.25892389e-01 1.39958620e+00 -3.08701545e-01
-1.07172400e-01 8.08810666e-02 -4.84863102e-01 5.30247808e-01
1.61884570e+00 -6.21388495e-01 1.52089179e-01 6.65610790e-01
1.16971761e-01 -1.43120870e-01 -1.82822943e+00 8.97231996e-01
-2.74197638e-01 -1.20001018e+00 -4.78462905e-01 1.79538190e-01
5.25807977e-01 1.56774700e-01 1.27312228e-01 4.83051658e-01
6.27429962e-01 -1.44711661e+00 6.18678212e-01 8.07759762e-01
1.25290930e+00 -1.01228583e+00 1.10780692e+00 4.86784697e-01
-4.65759456e-01 -8.67206156e-02 -1.46754295e-01 2.42771000e-01
3.18682015e-01 7.89011478e-01 -1.02190757e+00 3.33125532e-01
5.74054837e-01 6.65454030e-01 -4.92574006e-01 1.16808319e+00
2.83595860e-01 1.01486921e+00 -3.91595066e-01 3.15164149e-01
-8.68938938e-02 4.61315922e-02 2.20293999e-01 1.14026213e+00
5.04457355e-01 2.79386878e-01 -1.06431134e-01 6.91139221e-01
1.15335584e-01 -9.94903147e-02 -3.51974040e-01 2.62553543e-01
3.63904685e-01 8.89358103e-01 -5.58924973e-01 -3.66240323e-01
-1.95995599e-01 5.49503684e-01 2.23409951e-01 1.02336243e-01
-9.44036603e-01 4.97435480e-02 3.89696717e-01 2.75442094e-01
-2.75142968e-01 1.08664878e-01 -7.11804926e-01 -8.66165161e-01
-7.08141983e-01 -4.53312635e-01 9.34710026e-01 -7.92890608e-01
-1.65108550e+00 5.46522677e-01 1.43380761e-01 -1.08890796e+00
-2.40887105e-01 -4.48762089e-01 -4.85964060e-01 1.09708023e+00
-1.13737810e+00 -6.11228108e-01 -2.13061631e-01 4.98983383e-01
1.26606487e-02 -3.28065604e-02 9.38748360e-01 1.83534801e-01
-7.00853229e-01 9.32903528e-01 4.27436560e-01 8.73673111e-02
9.78011608e-01 -1.38653398e+00 4.63283837e-01 4.24143583e-01
-3.72154981e-01 4.53003228e-01 8.50941896e-01 -7.10907221e-01
-7.56940007e-01 -1.37593508e+00 7.64600754e-01 -7.55815089e-01
5.53004861e-01 8.16176459e-02 -9.55664039e-01 9.86888051e-01
-2.55767524e-01 -1.04669586e-01 1.06954265e+00 1.52160808e-01
-1.33650392e-01 -4.58993018e-02 -1.26337314e+00 2.90485442e-01
7.90246248e-01 -2.33844817e-01 -5.50524652e-01 4.78722155e-01
6.18063331e-01 -4.22084749e-01 -1.33370352e+00 5.35489976e-01
5.02421141e-01 -8.78963232e-01 9.23962951e-01 -1.05235958e+00
5.29273987e-01 1.31757244e-01 -2.86187291e-01 -1.61709249e+00
-4.48820859e-01 -2.86389112e-01 6.62602261e-02 8.19082677e-01
4.73794818e-01 -7.46089399e-01 5.91105103e-01 7.48169482e-01
-1.23952940e-01 -8.83578777e-01 -1.27581716e+00 -7.88318276e-01
7.95212984e-01 -3.05076361e-01 7.41818309e-01 1.11694705e+00
8.29617307e-02 1.69633865e-01 -1.90400794e-01 3.34760278e-01
7.60819674e-01 -3.49013567e-01 4.93273556e-01 -1.52933538e+00
-2.14105442e-01 -3.93448234e-01 -3.19044292e-01 -5.89553833e-01
8.25658217e-02 -1.06545663e+00 -6.18140884e-02 -1.36470246e+00
6.63307667e-01 -5.67775369e-01 -4.67255205e-01 5.45674205e-01
-2.31917411e-01 -2.97961593e-01 -4.19109464e-01 1.35282829e-01
-1.24530151e-01 5.14328241e-01 7.02866077e-01 2.11835533e-01
-1.00725591e-01 5.28214350e-02 -5.85894287e-01 6.68876290e-01
8.30615282e-01 -9.21896577e-01 -2.01226324e-01 -5.43751717e-02
1.12703912e-01 1.06749892e+00 4.50300843e-01 -7.55630910e-01
1.03707016e-01 -1.53849900e-01 5.06017029e-01 -6.32198691e-01
-1.59477070e-02 -7.96367288e-01 6.46043897e-01 8.77340376e-01
-6.64583802e-01 4.27683860e-01 2.07019776e-01 6.28878534e-01
8.00089836e-02 5.11975624e-02 1.03127861e+00 2.60848820e-01
-4.62467819e-02 4.35340405e-01 -2.29994744e-01 2.49735236e-01
1.13731432e+00 9.83396769e-02 -4.06154811e-01 -4.31080461e-01
-9.90296960e-01 4.92424905e-01 6.68313801e-01 1.21965095e-01
4.26471949e-01 -1.36241817e+00 -1.04214478e+00 7.26695657e-02
3.87789667e-01 -1.50492162e-01 2.33229727e-01 1.13689649e+00
-4.15140897e-01 4.03900176e-01 -5.17166965e-02 -7.91629434e-01
-1.06931198e+00 6.54270411e-01 6.85853124e-01 -4.03761685e-01
-5.81823289e-01 6.84184313e-01 5.40403724e-01 -5.10517955e-01
2.11578727e-01 -5.54805219e-01 3.51685397e-02 -2.01132745e-01
4.09154415e-01 2.83523142e-01 -9.93204117e-02 -4.71788883e-01
-3.60436946e-01 1.93035245e-01 1.57767341e-01 -8.46993029e-02
1.23405015e+00 -4.90719154e-02 -8.22615847e-02 6.92912698e-01
1.27697802e+00 -6.79158747e-01 -1.38124275e+00 -1.81882575e-01
8.69402885e-02 -2.77637899e-01 2.55143762e-01 -1.15294492e+00
-1.06240320e+00 8.07453513e-01 8.36895704e-01 -5.51989190e-02
1.23100090e+00 1.13089591e-01 3.43062520e-01 9.07939374e-02
4.22814965e-01 -4.05855417e-01 -3.02851200e-01 1.96222782e-01
7.87707031e-01 -1.16795421e+00 -1.55268535e-01 1.33726954e-01
-6.27931595e-01 9.85683680e-01 2.62107909e-01 -4.97058593e-02
9.52394485e-01 5.43545246e-01 -1.45972610e-01 -2.96227187e-02
-9.41300154e-01 4.71346438e-01 1.81270868e-01 4.60410863e-01
4.89942491e-01 5.33319473e-01 -3.35313171e-01 1.09547877e+00
-4.13116843e-01 3.58764559e-01 6.24111176e-01 4.22625065e-01
-1.79731429e-01 -7.64421821e-01 -3.80738258e-01 8.47532392e-01
-5.41731834e-01 -1.21921286e-01 8.75579789e-02 5.97250223e-01
-3.11305404e-01 6.39708281e-01 2.97502190e-01 -1.70023099e-01
2.32079118e-01 6.67371079e-02 2.54714251e-01 -2.90885180e-01
-3.60664487e-01 -1.54627621e-01 -4.38381471e-02 -3.99620831e-01
-8.54834169e-02 -8.55279207e-01 -1.28933930e+00 -5.51223397e-01
-2.99191415e-01 2.16896802e-01 3.77436996e-01 8.79334927e-01
7.88205117e-02 8.28878999e-01 4.45410490e-01 -3.14774364e-01
-8.24058533e-01 -9.57677960e-01 -6.76242232e-01 1.52722120e-01
5.41398048e-01 -4.88329053e-01 -5.08636475e-01 1.30258590e-01] | [7.8232502937316895, 5.630885601043701] |
cb330354-dcf6-47b8-98db-fcf22d340d35 | evolutionary-multitasking-for-semantic-web | 1902.06370 | null | http://arxiv.org/abs/1902.06370v1 | http://arxiv.org/pdf/1902.06370v1.pdf | Evolutionary Multitasking for Semantic Web Service Composition | Web services are basic functions of a software system to support the concept
of service-oriented architecture. They are often composed together to provide
added values, known as web service composition. Researchers often employ
Evolutionary Computation techniques to efficiently construct composite services
with near-optimized functional quality (i.e., Quality of Semantic Matchmaking)
or non-functional quality (i.e., Quality of Service) or both due to the
complexity of this problem. With a significant increase in service composition
requests, many composition requests have similar input and output requirements
but may vary due to different preferences from different user segments. This
problem is often treated as a multi-objective service composition so as to cope
with different preferences from different user segments simultaneously. Without
taking a multi-objective approach that gives rise to a solution selection
challenge, we perceive multiple similar service composition requests as jointly
forming an evolutionary multi-tasking problem in this work. We propose an
effective permutation-based evolutionary multi-tasking approach that can
simultaneously generate a set of solutions, with one for each service request.
We also introduce a neighborhood structure over multiple tasks to allow newly
evolved solutions to be evaluated on related tasks. Our proposed method can
perform better at the cost of only a fraction of time, compared to one
state-of-art single-tasking EC-based method. We also found that the use of the
proper neighborhood structure can enhance the effectiveness of our approach. | ['Hui Ma', 'Sven Hartmann', 'Gang Chen', 'Chen Wang'] | 2019-02-18 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [ 3.59965175e-01 -3.26180607e-01 2.90348470e-01 -4.44671243e-01
-6.56008124e-01 -5.09216726e-01 2.47928873e-01 -2.69184351e-01
-2.16873601e-01 4.57341611e-01 -1.72835160e-02 8.69318098e-02
-5.47856569e-01 -7.35061288e-01 -4.67189103e-01 -7.17042327e-01
1.72470361e-01 6.79080248e-01 4.62563187e-01 -4.75777060e-01
3.73114228e-01 1.56121001e-01 -2.44188118e+00 4.83574837e-01
1.42969131e+00 9.28014517e-01 6.87481999e-01 3.52134347e-01
-5.82829893e-01 -7.45856538e-02 -7.98104286e-01 -3.70138794e-01
2.31074572e-01 -5.17330050e-01 -6.67119265e-01 4.28786874e-01
-2.49506310e-01 3.74108285e-01 5.69601893e-01 1.21195245e+00
7.00503170e-01 3.94782394e-01 2.64988780e-01 -1.79543400e+00
-4.21129167e-01 6.12218559e-01 -2.55128711e-01 -1.44600376e-01
4.83458936e-01 -1.07274361e-01 1.04008424e+00 -6.24275923e-01
5.30344188e-01 1.06741214e+00 2.21692234e-01 6.29510283e-01
-1.20581758e+00 -3.47021431e-01 3.60842943e-01 5.47014117e-01
-1.38423645e+00 -4.75620210e-01 6.80226445e-01 3.44625115e-03
1.03627419e+00 7.22933829e-01 6.39527142e-01 6.92367077e-01
1.89182654e-01 5.05864441e-01 8.47499132e-01 -2.54492909e-01
5.89599848e-01 1.15502745e-01 -2.23174810e-01 2.09960014e-01
3.35011750e-01 -3.17555159e-01 -3.48699749e-01 -4.40365821e-01
-5.42078018e-02 8.56950507e-02 -2.59094656e-01 -5.15413284e-01
-9.94061589e-01 5.62770605e-01 -5.84000722e-02 4.61984634e-01
-5.76802909e-01 -1.40550256e-01 3.08223099e-01 6.45496905e-01
1.00010157e-01 6.31316185e-01 -8.20684493e-01 -3.54686886e-01
-4.43567872e-01 3.94996375e-01 1.16393340e+00 1.00738287e+00
8.06624651e-01 7.70854875e-02 -6.02097437e-02 1.05253708e+00
2.45425090e-01 1.58650771e-01 7.04084754e-01 -9.47524786e-01
2.39445865e-02 8.28387558e-01 2.62956887e-01 -7.64382601e-01
-2.44375288e-01 -5.45804024e-01 -3.03210497e-01 2.15930834e-01
2.24995855e-02 8.82701483e-03 -7.17778683e-01 1.75668168e+00
6.00342155e-01 -1.58353150e-01 5.42892106e-02 9.35464442e-01
2.21526667e-01 4.54756558e-01 -4.06568408e-01 -5.59191585e-01
1.50265348e+00 -1.16960979e+00 -8.40885460e-01 -6.68139085e-02
1.93327904e-01 -8.96676660e-01 1.25808764e+00 4.70762610e-01
-9.93092775e-01 -5.44652790e-02 -9.67548788e-01 6.70326948e-01
-2.43994296e-01 -6.19897544e-01 4.32638198e-01 1.05330658e+00
-1.05021334e+00 3.48551482e-01 -3.45451593e-01 -3.41967851e-01
-2.68751442e-01 3.52617174e-01 4.28485945e-02 -2.52778351e-01
-9.53112483e-01 8.45206380e-01 1.04435913e-01 -1.27075806e-01
-7.48754680e-01 -4.96333778e-01 -4.08936679e-01 3.73251498e-01
1.08437431e+00 -7.74260461e-01 1.31937218e+00 -1.42129517e+00
-1.48581970e+00 1.33265465e-01 -3.66385490e-01 2.96665847e-01
3.06371689e-01 4.64029104e-01 -8.12354147e-01 -2.85954744e-01
-1.13450950e-02 7.03352019e-02 8.48267078e-01 -1.44350135e+00
-1.16432202e+00 -2.86808491e-01 2.03915939e-01 5.28095543e-01
-4.98912126e-01 4.67395991e-01 -5.23826003e-01 -3.79676610e-01
3.00520152e-01 -9.49344993e-01 -3.90260249e-01 -3.45936447e-01
1.07862897e-01 -2.99233407e-01 6.99243188e-01 -3.28272015e-01
1.19446695e+00 -1.81077087e+00 6.47076607e-01 2.66017169e-01
3.10249832e-02 -5.02832904e-02 -4.55513895e-01 5.30549526e-01
2.04360232e-01 2.51468688e-01 -4.59280938e-01 -8.43938366e-02
1.96367741e-01 4.91137922e-01 4.74428952e-01 -9.47784409e-02
8.84464681e-02 5.11913419e-01 -8.59543443e-01 -2.15331122e-01
-4.63273913e-01 2.48396009e-01 -5.87477624e-01 2.47415215e-01
-6.25592232e-01 1.58213064e-01 -6.17907763e-01 9.48060989e-01
5.19229472e-01 -1.69224203e-01 5.01580596e-01 -4.70236577e-02
-7.46189207e-02 -5.63458074e-03 -1.61074150e+00 1.63290787e+00
-7.11866617e-01 -2.69782878e-02 3.12083274e-01 -1.13143814e+00
1.01073158e+00 5.80083370e-01 8.27995360e-01 -8.07497144e-01
1.01302467e-01 6.08750641e-01 5.54321289e-01 -5.58996618e-01
4.30370748e-01 1.56445876e-02 -7.92996436e-02 8.22714627e-01
-2.94411331e-01 -3.05316988e-02 3.90691429e-01 -3.56677145e-01
1.10358500e+00 2.08441705e-01 6.96286112e-02 -3.87777865e-01
5.81095994e-01 -6.38367459e-02 1.12382293e+00 4.48445857e-01
-6.20870218e-02 4.06168103e-01 3.30637485e-01 -4.13797647e-01
-1.08661568e+00 -6.56684518e-01 1.28790796e-01 1.41398585e+00
2.51015574e-01 -3.13127816e-01 -6.49780810e-01 -5.67168117e-01
-6.08871989e-02 8.14673007e-01 -2.33291954e-01 -3.01391214e-01
-5.34341156e-01 -9.16173041e-01 1.58530757e-01 -1.71529189e-01
5.94202541e-02 -1.06980968e+00 -9.61042881e-01 5.77870607e-01
-4.23065305e-01 -9.29072559e-01 -5.75541615e-01 1.79238781e-01
-5.31702638e-01 -8.78246307e-01 -7.57141352e-01 -6.19410396e-01
7.01685131e-01 4.82935756e-01 1.13939750e+00 3.56517315e-01
-1.63449392e-01 1.72578514e-01 -8.30273449e-01 -1.57499909e-01
-4.14011210e-01 -1.82848405e-02 1.06318936e-01 3.16311955e-01
2.29938447e-01 -8.81536543e-01 -3.17336768e-01 9.58761632e-01
-1.25377786e+00 -1.60470754e-02 4.98066425e-01 8.68358016e-01
5.19293129e-01 5.67353666e-01 7.73522973e-01 -3.79132867e-01
1.16926610e+00 -5.23788154e-01 -4.38896149e-01 7.65246153e-01
-1.14679849e+00 2.87526727e-01 8.22576642e-01 -8.30494881e-01
-1.03150153e+00 -2.65828788e-01 1.60361737e-01 -2.27826834e-01
2.26647630e-01 5.87658167e-01 -5.36680996e-01 -1.80179700e-01
3.71771395e-01 3.91518295e-01 1.35283038e-01 -5.23398936e-01
1.65830597e-01 7.79994607e-01 -1.22064613e-01 -9.82547820e-01
4.41439331e-01 6.17669779e-04 -1.52805641e-01 -1.40306950e-01
-7.93233886e-02 -2.59804428e-01 1.56827420e-02 -3.86197984e-01
5.28743386e-01 -2.12134555e-01 -8.74426067e-01 4.08592880e-01
-9.47883248e-01 1.32731631e-01 -1.49955060e-02 -2.55698320e-02
-3.68027151e-01 2.57687598e-01 2.69647930e-02 -9.92944360e-01
-4.11418140e-01 -1.78566337e+00 7.44447827e-01 2.21536636e-01
1.40295727e-02 -3.75388682e-01 -2.46681243e-01 3.06491137e-01
8.82168591e-01 5.43251485e-02 1.03824198e+00 -7.96744704e-01
-4.41629499e-01 8.81707296e-02 4.96324487e-02 4.69009764e-02
4.51404929e-01 -2.72145737e-02 -2.78392434e-01 -3.51282716e-01
2.24940121e-01 -2.94292509e-03 9.75036621e-02 -1.42440107e-02
8.76063287e-01 -3.76885593e-01 -1.71637405e-02 3.56101066e-01
1.37750256e+00 8.84251654e-01 5.22354305e-01 6.09549224e-01
3.40021700e-01 1.17686987e+00 7.63392508e-01 7.62973130e-01
5.00583887e-01 1.22673237e+00 4.99707341e-01 3.59673589e-01
1.73351571e-01 4.38755244e-01 3.50443423e-01 9.11913872e-01
3.83360907e-02 -4.58365142e-01 -9.45773661e-01 4.20657516e-01
-2.17435813e+00 -7.00967789e-01 -6.20571189e-02 2.18559647e+00
7.25787699e-01 -1.48054853e-01 1.89171821e-01 1.85717210e-01
8.73022497e-01 -3.56534988e-01 -7.98509359e-01 -7.81639695e-01
-4.51987199e-02 -2.53312681e-02 4.62013111e-02 1.14488058e-01
-2.79049635e-01 5.69944382e-01 5.28988123e+00 7.74041176e-01
-8.91449690e-01 2.39240408e-01 2.09740728e-01 -2.64696360e-01
-9.97001112e-01 1.14794523e-01 -3.88720095e-01 8.38282347e-01
7.04141915e-01 -7.91467190e-01 1.14534509e+00 5.99373400e-01
1.67239711e-01 -1.52690588e-02 -8.19736183e-01 7.32470870e-01
1.54731944e-01 -8.47289562e-01 -6.82353154e-02 1.72564864e-01
8.77409697e-01 -3.50850999e-01 -1.06898636e-01 2.51782108e-02
2.45230526e-01 -6.44275784e-01 9.12885249e-01 2.62045741e-01
4.56686944e-01 -9.13871229e-01 7.11424053e-01 4.50703025e-01
-1.30392277e+00 -4.02177513e-01 -1.74373031e-01 1.39907971e-01
5.68378747e-01 5.64345300e-01 -3.60349536e-01 1.02903271e+00
1.02322912e+00 -1.78899586e-01 -1.77586958e-01 1.34085584e+00
1.84520915e-01 -1.62937850e-01 -2.12894022e-01 -4.80762959e-01
4.87126447e-02 -4.58106756e-01 8.17914486e-01 5.20092666e-01
8.72543871e-01 -5.82306460e-02 3.95736307e-01 6.63936675e-01
3.00366193e-01 3.43150049e-01 -1.34299085e-01 -1.50788203e-01
6.19579196e-01 1.18281174e+00 -6.25639379e-01 -1.89459920e-01
-3.18399101e-01 1.02230501e+00 -9.02503729e-02 3.27560276e-01
-8.44845295e-01 -4.24975544e-01 1.17938149e+00 -2.02789500e-01
2.59235680e-01 -2.25498509e-02 -2.55252570e-01 -1.06742013e+00
3.98438096e-01 -1.38992047e+00 1.76662087e-01 -5.37593901e-01
-1.14320183e+00 1.04316902e+00 -2.47903064e-01 -1.31111276e+00
-3.48278701e-01 7.28959264e-03 -4.42959100e-01 8.70508075e-01
-1.44079208e+00 -8.08000326e-01 -2.72322983e-01 4.78849351e-01
7.21010149e-01 -4.17602867e-01 8.13907683e-01 5.31876385e-01
-5.79613805e-01 4.25974637e-01 6.29189834e-02 -1.07046533e+00
6.13860130e-01 -8.81461740e-01 2.07493529e-01 9.38036561e-01
-3.25856388e-01 3.92157912e-01 9.52962220e-01 -5.81701994e-01
-1.76461911e+00 -6.82646453e-01 9.72018063e-01 1.85232267e-01
3.89412850e-01 -2.57756442e-01 -8.23119819e-01 -7.83956945e-02
5.16486801e-02 -4.18993443e-01 7.84899771e-01 7.86667392e-02
-6.43014610e-02 -3.26135099e-01 -1.51612437e+00 7.85399973e-01
1.32729280e+00 4.69708554e-02 -3.32925856e-01 1.62488148e-01
1.13679910e+00 -2.48719426e-03 -7.10632443e-01 4.68888402e-01
6.76329255e-01 -8.67918313e-01 6.25313044e-01 -5.23888290e-01
1.18735127e-01 -6.37625158e-01 -5.09174049e-01 -1.55741799e+00
-4.22802687e-01 -6.75441384e-01 1.69703588e-01 1.17759085e+00
5.95195651e-01 -9.74193275e-01 5.71517944e-01 8.96075428e-01
-4.25066739e-01 -7.27997661e-01 -1.04973662e+00 -1.08751667e+00
-5.31691730e-01 -2.82776922e-01 1.36826348e+00 8.92287731e-01
-1.35821059e-01 -4.35646661e-02 -3.18373680e-01 3.47562693e-02
4.58815664e-01 3.39856237e-01 4.13310111e-01 -1.40378129e+00
-5.83478034e-01 -7.88686454e-01 -7.62854284e-03 -4.03905332e-01
2.09442377e-02 -7.45145142e-01 2.83721864e-01 -1.63367486e+00
1.32460911e-02 -7.69849479e-01 -4.20304537e-01 4.25509840e-01
-1.56587020e-01 -2.92355597e-01 2.82810330e-01 1.48359030e-01
-4.13101047e-01 5.11126935e-01 1.27717018e+00 -8.78718942e-02
-4.01176125e-01 2.76301801e-01 -9.48825955e-01 1.95805490e-01
8.22564244e-01 -8.40473354e-01 -6.55454576e-01 -6.24018848e-01
4.11790341e-01 2.39955485e-01 -4.10925239e-01 -6.57394171e-01
2.94845968e-01 -5.81823707e-01 -3.47286820e-01 8.51681605e-02
3.03025573e-01 -1.19818556e+00 9.17157292e-01 5.11854649e-01
-4.94718254e-02 6.24405801e-01 -2.39393428e-01 3.74207735e-01
-2.13715032e-01 -5.02962351e-01 3.56539518e-01 -1.96726009e-01
-8.14982295e-01 1.27065077e-01 -5.78602314e-01 -3.55389833e-01
1.36252201e+00 -5.25511086e-01 -1.78802773e-01 -2.70948470e-01
-3.83140087e-01 5.29174089e-01 8.10889363e-01 8.79698694e-01
5.48344135e-01 -1.23901057e+00 -7.96892464e-01 3.16328034e-02
2.89295882e-01 -4.08813357e-01 1.57226428e-01 6.33104205e-01
-8.93037301e-03 -9.05058696e-04 -5.09276092e-01 -3.50779951e-01
-1.32792783e+00 5.07825375e-01 2.31583819e-01 -1.62622690e-01
-2.29130343e-01 6.14794493e-01 -1.91727117e-01 -6.91507220e-01
2.26605818e-01 7.67245367e-02 -1.75560504e-01 3.24677169e-01
3.86915028e-01 4.72835451e-01 3.42190742e-01 -4.09109175e-01
-6.07735991e-01 4.11740929e-01 4.14014965e-01 -4.17441696e-01
1.29376709e+00 -1.06135555e-01 -4.03418630e-01 1.06935449e-01
8.18841994e-01 -2.55102098e-01 -7.52001822e-01 -1.48246109e-01
3.21407408e-01 -8.04918706e-01 -1.74318597e-01 -1.09455061e+00
-1.27763760e+00 2.01055914e-01 4.77934599e-01 4.45928037e-01
1.64887464e+00 -2.62421399e-01 8.86626422e-01 2.14444056e-01
7.26648688e-01 -1.26765025e+00 1.44601703e-01 3.11559051e-01
7.83711195e-01 -8.82025182e-01 -5.16076565e-01 -4.25610363e-01
-8.38089228e-01 8.76881719e-01 6.26861632e-01 4.53785986e-01
2.38951191e-01 5.31911075e-01 -9.64653939e-02 1.00838348e-01
-1.22205091e+00 -4.33518827e-01 3.28118265e-01 6.92961335e-01
1.30057484e-01 1.62622020e-01 -9.58966374e-01 7.89377809e-01
8.33282322e-02 -2.61134207e-01 4.42522794e-01 1.15533423e+00
-6.62404001e-01 -1.88536656e+00 -4.35139865e-01 4.59834427e-01
-2.53254622e-01 3.09104286e-02 -1.59739286e-01 5.20981103e-02
2.06392169e-01 1.36311841e+00 -4.39827070e-02 -6.57640517e-01
6.59780800e-01 2.11879492e-01 2.55524963e-01 -4.15367246e-01
-1.00688124e+00 3.31412077e-01 4.11822855e-01 -6.31524086e-01
-2.00857058e-01 -7.34807134e-01 -1.10274744e+00 -2.62605995e-01
-2.73766071e-01 4.57435846e-01 1.10716999e+00 8.31808388e-01
6.41071618e-01 8.77003074e-01 1.00423026e+00 -6.26159668e-01
-7.49269307e-01 -4.74774927e-01 -4.10435855e-01 4.05577421e-01
-2.87516803e-01 -8.44188273e-01 -1.06458589e-01 -2.33471140e-01] | [8.594792366027832, 6.931784152984619] |
31e2cea6-c566-4d46-887e-f739900234fb | are-pretrained-transformers-robust-in-intent | 2106.04564 | null | https://arxiv.org/abs/2106.04564v3 | https://arxiv.org/pdf/2106.04564v3.pdf | Are Pretrained Transformers Robust in Intent Classification? A Missing Ingredient in Evaluation of Out-of-Scope Intent Detection | Pre-trained Transformer-based models were reported to be robust in intent classification. In this work, we first point out the importance of in-domain out-of-scope detection in few-shot intent recognition tasks and then illustrate the vulnerability of pre-trained Transformer-based models against samples that are in-domain but out-of-scope (ID-OOS). We construct two new datasets, and empirically show that pre-trained models do not perform well on both ID-OOS examples and general out-of-scope examples, especially on fine-grained few-shot intent detection tasks. To figure out how the models mistakenly classify ID-OOS intents as in-scope intents, we further conduct analysis on confidence scores and the overlapping keywords, as well as point out several prospective directions for future work. Resources are available on https://github.com/jianguoz/Few-Shot-Intent-Detection. | ['Caiming Xiong', 'Ye Liu', 'Zhiwei Liu', 'JianGuo Zhang', 'Philip S. Yu', 'Yao Wan', 'Kazuma Hashimoto'] | 2021-06-08 | null | null | null | null | ['intent-recognition'] | ['natural-language-processing'] | [ 2.01476678e-01 -1.75640061e-01 -3.98348987e-01 -4.15684640e-01
-1.10034752e+00 -3.84515077e-01 6.80245578e-01 1.42057657e-01
-2.48307079e-01 2.89111406e-01 4.42447931e-01 -2.66127378e-01
6.57767281e-02 -6.18272483e-01 -1.92467332e-01 -1.51823774e-01
2.33146213e-02 3.48846525e-01 5.23938298e-01 -3.99193197e-01
6.57294214e-01 -1.44013330e-01 -1.44984877e+00 3.96777749e-01
8.59026849e-01 6.90228105e-01 3.46891843e-02 6.46129787e-01
-7.89880231e-02 1.05246699e+00 -1.02346075e+00 -3.80249321e-01
5.50421700e-02 -3.45171481e-01 -5.57439744e-01 3.28631029e-02
4.13988262e-01 -6.12391949e-01 -4.69113946e-01 9.76907134e-01
2.56536454e-01 3.20962459e-01 1.03012025e+00 -1.43444979e+00
-9.19387043e-01 3.09730530e-01 -2.03058302e-01 8.56208086e-01
7.85162389e-01 4.03441638e-01 1.39907348e+00 -1.18044031e+00
6.40622854e-01 9.64129806e-01 8.30206752e-01 4.10419762e-01
-8.41685534e-01 -7.59691238e-01 1.09544337e-01 4.48133826e-01
-1.44078052e+00 -6.11518741e-01 7.87844598e-01 -4.90740448e-01
1.52338064e+00 3.09251040e-01 2.52021283e-01 1.61473572e+00
3.33711177e-01 1.04415774e+00 7.50066638e-01 -5.33104539e-01
4.23044384e-01 8.60051438e-02 9.07669246e-01 6.03611887e-01
3.43821645e-01 8.31323955e-03 -6.24449432e-01 -4.37042207e-01
6.82752654e-02 5.12845039e-01 -1.95511863e-01 -2.09551286e-02
-1.07410765e+00 1.20788515e+00 1.85780540e-01 6.74229920e-01
-1.94041476e-01 -1.50497243e-01 6.26328409e-01 3.92317921e-01
7.12675691e-01 6.44120216e-01 -3.84829283e-01 -4.66312766e-01
-1.24815011e+00 1.06412172e-01 7.00730979e-01 1.16831696e+00
6.55190349e-01 -5.68082593e-02 -7.36208498e-01 1.01797819e+00
-7.94550776e-02 4.11436595e-02 8.57617259e-01 -4.63328511e-01
4.44504738e-01 7.98530638e-01 -3.06325732e-03 -6.39726758e-01
-1.06655315e-01 -2.99565971e-01 -4.46041524e-01 -3.38128567e-01
9.66724157e-02 6.50058538e-02 -1.04358053e+00 1.39684963e+00
-2.13072538e-01 -3.29832733e-02 -1.14849722e-02 5.72986841e-01
9.05323207e-01 5.40591776e-01 2.36959502e-01 -9.67162848e-02
1.91461301e+00 -1.07700098e+00 -9.00655031e-01 -6.74721062e-01
1.01621950e+00 -6.64848685e-01 1.58094597e+00 -1.18526123e-01
-5.87616980e-01 -4.32701945e-01 -1.07250416e+00 -3.03154178e-02
-7.39228964e-01 1.69977859e-01 5.00462830e-01 5.88601172e-01
-7.16049373e-01 2.27357447e-01 -5.35713136e-01 -8.67935538e-01
5.64729035e-01 -2.45129704e-01 5.24966493e-02 -2.05283776e-01
-1.64145422e+00 8.81189466e-01 3.58687133e-01 -8.39450717e-01
-1.19597435e+00 -9.99322593e-01 -9.68048036e-01 2.37510756e-01
7.31421053e-01 -4.97711718e-01 1.63597369e+00 -2.10685343e-01
-6.22436643e-01 9.42823946e-01 -3.84352803e-01 -4.43507791e-01
1.12630285e-01 -5.76069094e-02 -7.70362496e-01 6.74058101e-04
6.38743758e-01 4.07886863e-01 9.49827552e-01 -7.57309973e-01
-7.30248988e-01 -2.26297379e-01 2.39467099e-01 4.63184491e-02
-7.28283644e-01 2.31511682e-01 -2.05668867e-01 -9.73475218e-01
-1.94979116e-01 -5.99671662e-01 1.67350575e-01 -3.81455749e-01
-6.00851536e-01 -6.20790422e-01 1.16705942e+00 -3.68374497e-01
1.68498194e+00 -2.09426022e+00 -8.24333906e-01 -5.35252094e-01
2.42980439e-02 2.36014709e-01 -3.27202454e-02 7.61094093e-01
4.61201891e-02 2.97855407e-01 -8.80840197e-02 -3.42940509e-01
1.87463894e-01 -1.60718542e-02 -6.60381675e-01 1.41389966e-01
1.80145413e-01 9.46741998e-01 -9.12451982e-01 -5.62380552e-01
2.10877240e-01 -6.27565617e-03 -2.20393091e-01 3.18466187e-01
-3.48659813e-01 -1.81603089e-01 -5.54953456e-01 1.12536573e+00
1.66702822e-01 -5.12949765e-01 -2.83123106e-01 -2.70980805e-01
9.00371596e-02 9.10989165e-01 -4.33656633e-01 1.45251310e+00
-5.46032250e-01 7.17280865e-01 -5.42810440e-01 -7.10618556e-01
6.86898887e-01 4.07731116e-01 -1.61085315e-02 -7.33195007e-01
1.04916349e-01 1.67071432e-01 -1.00660458e-01 -4.26024586e-01
8.19308043e-01 -5.45437932e-01 -6.16486967e-01 6.94657087e-01
3.80664825e-01 -1.01016350e-01 4.60254252e-01 4.73526925e-01
1.45264041e+00 -5.32938957e-01 8.48808825e-01 -9.67350230e-02
3.29003483e-01 4.05745536e-01 2.47058988e-01 1.04536080e+00
-6.18685067e-01 6.88024640e-01 2.85207331e-01 -2.21460134e-01
-7.82584608e-01 -7.42257535e-01 -1.13495499e-01 1.48197246e+00
4.68308628e-02 -7.68498182e-01 -2.83914536e-01 -1.26082516e+00
-1.51159227e-01 1.58063948e+00 -5.87694943e-01 -4.23330933e-01
-9.37247351e-02 -5.95975339e-01 8.26938808e-01 4.51963931e-01
2.13456705e-01 -1.03161418e+00 -8.04804146e-01 8.46927688e-02
-5.13973594e-01 -1.05950344e+00 -8.60278368e-01 6.00159943e-01
-7.07886875e-01 -9.32822943e-01 -8.22146118e-01 -5.51972508e-01
3.81191820e-01 6.28876686e-01 1.11344075e+00 8.31772014e-02
-4.91983473e-01 3.93831164e-01 -7.86279559e-01 -5.01739264e-01
-3.97328347e-01 5.07462658e-02 -5.14304191e-02 -5.02436876e-01
1.12178206e+00 -4.68019158e-01 -2.87369370e-01 5.37208378e-01
-7.61733174e-01 -3.26783180e-01 4.18119878e-01 8.88630390e-01
-4.05383073e-02 1.58189759e-01 4.95865792e-01 -7.40862131e-01
1.02704024e+00 -6.93719864e-01 -7.26143569e-02 2.20658481e-01
-6.82325244e-01 -1.27904639e-01 6.67953551e-01 -5.65842152e-01
-9.82194304e-01 -5.90563297e-01 -1.78095892e-01 -7.53936946e-01
-4.75820452e-01 3.96887422e-01 2.88830310e-01 3.36766899e-01
8.28748465e-01 5.54199219e-01 -5.20261109e-01 -3.45948607e-01
-9.97374207e-02 1.03433704e+00 8.10809433e-03 -1.65002182e-01
7.42575228e-01 3.71759206e-01 -6.88237607e-01 -1.04913473e+00
-1.28363323e+00 -1.08566296e+00 -2.76537657e-01 -4.20057811e-02
5.77155828e-01 -9.72729981e-01 5.69314174e-02 4.23196852e-01
-1.02706659e+00 -2.76039004e-01 -4.32007730e-01 1.58539191e-01
-4.10013378e-01 3.81826729e-01 -7.89255023e-01 -9.43256497e-01
-5.36314547e-01 -9.75451410e-01 1.29146850e+00 2.02886507e-01
-9.71561611e-01 -1.14964616e+00 1.47472918e-01 5.50675333e-01
3.68797034e-01 -5.95090687e-01 9.61174846e-01 -1.47557735e+00
-2.08939865e-01 -6.10417545e-01 -3.16727757e-02 -4.85394336e-03
3.00597459e-01 -3.25032204e-01 -1.04956377e+00 -1.45031646e-01
4.26392376e-01 -6.89609230e-01 1.02525270e+00 2.28329316e-01
7.26275921e-01 -3.09250176e-01 -5.86277485e-01 2.62771994e-01
1.17131960e+00 4.55989003e-01 6.64744973e-01 3.52184653e-01
3.52683067e-01 3.88216585e-01 1.28290915e+00 5.34089804e-01
2.16433302e-01 8.89392912e-01 2.44309157e-02 3.59855682e-01
-1.71264261e-01 -4.48211163e-01 5.12226641e-01 5.24497032e-01
5.14308274e-01 -5.02362788e-01 -9.25595522e-01 1.00172198e+00
-1.52714169e+00 -1.24882412e+00 6.83452114e-02 1.82771575e+00
6.14935577e-01 4.87318665e-01 2.39244699e-01 8.46639276e-02
8.52811277e-01 6.28257215e-01 -4.62103188e-01 -3.42608720e-01
2.45734930e-01 -2.76011266e-02 2.08842039e-01 1.29936293e-01
-1.26460779e+00 9.93326068e-01 6.39939022e+00 1.31428719e+00
-6.62832499e-01 4.72277939e-01 6.39187336e-01 -7.59815574e-02
-2.69864470e-01 7.22671598e-02 -1.06189787e+00 5.93559802e-01
9.14418519e-01 -2.35689178e-01 -8.33535343e-02 1.28408587e+00
-8.34159851e-02 -1.24888055e-01 -1.05561149e+00 6.47793114e-01
3.85018617e-01 -1.15464962e+00 -9.82087851e-02 -1.84932083e-03
5.85166216e-01 4.17243838e-02 -1.00302458e-01 1.09226847e+00
2.13828251e-01 -5.78014672e-01 5.67962706e-01 -2.86083370e-02
8.87236297e-01 -3.13724816e-01 7.59349942e-01 7.23067284e-01
-1.19762039e+00 -8.12899172e-02 -4.73976374e-01 -1.54427320e-01
1.13532506e-01 6.51129663e-01 -9.99185681e-01 2.87981272e-01
6.65597916e-01 8.87867033e-01 -7.57014334e-01 7.20711112e-01
-2.13712692e-01 7.68311083e-01 -8.41011107e-02 -3.89348656e-01
2.69283593e-01 3.40925425e-01 8.06902587e-01 1.39252603e+00
4.58140135e-01 6.97665811e-02 4.29574549e-02 1.11758840e+00
6.44917712e-02 -2.91661888e-01 -7.97305405e-01 -5.04767418e-01
4.53803867e-01 1.20584297e+00 -8.00500512e-01 -7.64886260e-01
-5.52973330e-01 8.62800300e-01 1.27975672e-01 1.25757203e-01
-9.65251446e-01 -5.97225785e-01 8.98872554e-01 1.64003029e-01
4.83910680e-01 2.13629365e-01 -2.51818061e-01 -1.61956942e+00
-1.29814669e-01 -7.60112643e-01 7.78632581e-01 -8.59868765e-01
-1.59109020e+00 3.23943287e-01 1.79987088e-01 -1.59832406e+00
-4.42057073e-01 -2.84913033e-01 -1.20359588e+00 4.92084771e-01
-1.31321979e+00 -8.72166634e-01 -2.57240474e-01 3.70360017e-01
1.39933765e+00 -2.06173599e-01 8.54573429e-01 1.32583857e-01
-5.07764339e-01 7.48001337e-01 -2.73047119e-01 1.81071132e-01
8.64293814e-01 -7.94699013e-01 8.29628289e-01 8.95097196e-01
1.11003682e-01 9.05577600e-01 8.85368049e-01 -9.52685118e-01
-1.04820919e+00 -1.11743283e+00 1.08401501e+00 -6.04849219e-01
9.63219285e-01 -4.08881783e-01 -9.75166500e-01 9.42345500e-01
8.92868936e-02 -1.81519940e-01 7.70704508e-01 3.12446356e-01
-6.06200218e-01 5.27684450e-01 -1.15500140e+00 6.97608709e-01
1.23830557e+00 -7.54964888e-01 -1.36244512e+00 5.69067001e-01
8.27325165e-01 -1.02172546e-01 -4.56324995e-01 4.00672913e-01
3.17167938e-01 -1.23274040e+00 1.05671966e+00 -4.84402120e-01
4.97249365e-01 6.89912438e-02 -2.90804923e-01 -1.03795600e+00
-3.60676587e-01 -1.54233202e-01 -4.16954279e-01 1.18630731e+00
2.02027902e-01 -5.39141953e-01 7.30524719e-01 2.48593256e-01
-3.66948456e-01 -6.76601589e-01 -7.19502151e-01 -1.21570909e+00
-3.47128749e-01 -6.86516941e-01 2.22133979e-01 9.60807085e-01
5.63437104e-01 7.26351559e-01 -3.38771939e-01 -1.48293689e-01
4.88033950e-01 1.42481327e-01 5.25458336e-01 -1.02000940e+00
-1.64472312e-01 -2.91560709e-01 -3.37892711e-01 -1.00912046e+00
1.51343882e-01 -5.84626377e-01 1.20501243e-01 -1.55264699e+00
4.67394620e-01 -1.03324503e-01 -2.22562999e-01 8.11961174e-01
-4.07337546e-01 3.89265567e-01 3.68402638e-02 6.46275401e-01
-1.02588820e+00 5.35833657e-01 7.46078491e-01 -4.44407463e-01
2.10322272e-02 1.76989987e-01 -7.34858274e-01 7.12769151e-01
8.00857842e-01 -8.06526899e-01 -5.21506786e-01 4.45781082e-01
-3.73354793e-01 -3.45671847e-02 1.74489364e-01 -1.11464202e+00
8.41564611e-02 -2.11046889e-01 2.50728969e-02 -9.29376900e-01
5.68647802e-01 -5.81796408e-01 -2.48710155e-01 7.20333397e-01
-4.22892034e-01 -3.47610652e-01 -2.77560041e-03 5.62532425e-01
-2.21307218e-01 -7.58070648e-01 5.79665124e-01 -4.05853301e-01
-1.23327935e+00 -1.17294483e-01 -8.04725468e-01 5.77224612e-01
1.15329719e+00 -4.09147084e-01 -6.91943288e-01 -6.80646122e-01
-3.37435454e-01 -7.91015700e-02 6.47259772e-01 5.77090621e-01
6.72035456e-01 -1.01594388e+00 -2.44780570e-01 2.29325652e-01
1.04115784e+00 -7.13892460e-01 4.89778697e-01 9.10922527e-01
6.07912354e-02 1.05642390e+00 1.61957055e-01 -4.43381488e-01
-1.35147309e+00 8.21834922e-01 1.10975187e-02 -6.11872017e-01
-5.27189314e-01 8.62302244e-01 4.68617916e-01 -4.68565732e-01
2.11426854e-01 -2.90812641e-01 -3.11093386e-02 1.02460369e-01
7.04161823e-01 2.95042515e-01 1.22900397e-01 -4.71328616e-01
-7.91837156e-01 1.91826105e-01 -4.60487247e-01 1.57493398e-01
8.58620703e-01 -2.67822482e-02 5.09494305e-01 6.60738111e-01
1.42736351e+00 -2.08348125e-01 -6.67300284e-01 -1.07205492e-02
1.08241446e-01 -5.90498030e-01 -7.40777627e-02 -8.16660583e-01
-3.58529836e-01 8.15418959e-01 1.91867277e-01 4.36080694e-01
9.31616902e-01 2.85247236e-01 9.39938188e-01 4.90844190e-01
5.05598724e-01 -9.52710390e-01 6.49681151e-01 9.00841296e-01
5.91262519e-01 -1.39419365e+00 3.86037193e-02 -3.16404551e-01
-9.86183107e-01 8.23871732e-01 6.02816820e-01 -1.54498264e-01
4.02457505e-01 2.05857679e-01 -1.16316631e-01 -5.30782521e-01
-1.14200866e+00 -4.61877227e-01 2.96303421e-01 5.40316701e-01
5.07507384e-01 -1.69177935e-01 -1.39712796e-01 5.69566011e-01
-6.42738342e-02 1.38598353e-01 5.56623757e-01 1.23781013e+00
-6.15947127e-01 -6.68284118e-01 -4.72737014e-01 1.00792575e+00
-3.58805269e-01 -4.98920798e-01 -5.11292040e-01 7.45399117e-01
-2.94774413e-01 1.16249490e+00 4.82462831e-02 -7.69188762e-01
2.20611036e-01 6.51202738e-01 -3.44157480e-02 -1.05084062e+00
-4.85462934e-01 -1.29113346e-01 3.39670330e-01 -3.63548785e-01
-2.59615630e-02 -5.16972721e-01 -8.01829040e-01 -2.56383568e-02
-5.45238853e-01 4.89387028e-02 1.64399385e-01 1.00320554e+00
4.10512298e-01 5.93132377e-01 2.49646112e-01 -6.05236709e-01
-7.78443754e-01 -1.27470255e+00 -5.40064514e-01 4.29609269e-01
3.52074146e-01 -6.97931170e-01 -1.09814000e+00 -2.72593915e-01] | [11.98349666595459, 7.592231750488281] |
cf6e5a2f-bab4-421c-b6ed-23469bd86f12 | leveraging-factored-action-spaces-for | 2305.01738 | null | https://arxiv.org/abs/2305.01738v1 | https://arxiv.org/pdf/2305.01738v1.pdf | Leveraging Factored Action Spaces for Efficient Offline Reinforcement Learning in Healthcare | Many reinforcement learning (RL) applications have combinatorial action spaces, where each action is a composition of sub-actions. A standard RL approach ignores this inherent factorization structure, resulting in a potential failure to make meaningful inferences about rarely observed sub-action combinations; this is particularly problematic for offline settings, where data may be limited. In this work, we propose a form of linear Q-function decomposition induced by factored action spaces. We study the theoretical properties of our approach, identifying scenarios where it is guaranteed to lead to zero bias when used to approximate the Q-function. Outside the regimes with theoretical guarantees, we show that our approach can still be useful because it leads to better sample efficiency without necessarily sacrificing policy optimality, allowing us to achieve a better bias-variance trade-off. Across several offline RL problems using simulators and real-world datasets motivated by healthcare, we demonstrate that incorporating factored action spaces into value-based RL can result in better-performing policies. Our approach can help an agent make more accurate inferences within underexplored regions of the state-action space when applying RL to observational datasets. | ['Jenna Wiens', 'Finale Doshi-Velez', 'Michael W. Sjoding', 'Maggie Makar', 'Shengpu Tang'] | 2023-05-02 | null | null | null | null | ['offline-rl'] | ['playing-games'] | [ 6.85035661e-02 2.99822509e-01 -6.79920495e-01 -5.22736683e-02
-8.75340760e-01 -6.30309999e-01 3.61106306e-01 2.17119768e-01
-6.59708977e-01 1.29687083e+00 2.38986582e-01 -7.57033408e-01
-4.76569891e-01 -8.02625895e-01 -8.63790810e-01 -6.55912399e-01
-5.24688125e-01 4.61403191e-01 -5.41642345e-02 -2.54980605e-02
4.68964912e-02 2.47800633e-01 -1.26844156e+00 -1.16964877e-01
9.69361842e-01 6.79660857e-01 -4.15447950e-02 6.51544869e-01
2.52735257e-01 9.54343259e-01 -5.82624495e-01 -6.43339306e-02
7.34825850e-01 -6.29136562e-01 -4.74893600e-01 1.25109345e-01
-1.98090868e-03 -8.50165069e-01 -1.73860848e-01 1.10144806e+00
3.98407727e-01 3.58448416e-01 3.57592463e-01 -1.44547522e+00
-3.87284271e-02 6.29606664e-01 -5.15034318e-01 -8.80809054e-02
2.82488853e-01 6.23443305e-01 1.05194807e+00 2.90081561e-01
4.29962695e-01 1.58498347e+00 4.00929421e-01 6.01330757e-01
-1.74938262e+00 -6.16758645e-01 4.58139390e-01 -1.48462519e-01
-7.17592895e-01 -1.94189414e-01 2.69086689e-01 -1.57867730e-01
8.30892444e-01 1.15852550e-01 9.25367177e-01 9.94391024e-01
5.84348023e-01 8.62951756e-01 1.42959261e+00 -1.71977490e-01
7.95613408e-01 -1.10229932e-01 -2.13018388e-01 4.42165226e-01
5.95778644e-01 6.55687928e-01 -3.53607148e-01 -5.11046410e-01
8.84744048e-01 2.44819492e-01 -1.52683273e-01 -6.26110494e-01
-1.04854965e+00 1.03897274e+00 8.36335570e-02 -3.44032586e-01
-8.47420931e-01 5.43787003e-01 2.75816798e-01 5.83860397e-01
2.77542174e-01 7.18589246e-01 -4.50838894e-01 -4.48231429e-01
-5.65688670e-01 8.21775973e-01 8.13096464e-01 6.00841284e-01
4.83751684e-01 1.12558588e-01 -3.06623667e-01 3.44425589e-01
2.73684319e-02 6.84257090e-01 2.22567335e-01 -1.56824839e+00
4.83661294e-01 2.51527160e-01 9.93843019e-01 -5.10442019e-01
-5.29275835e-01 -4.28337425e-01 -1.86092049e-01 4.05099273e-01
7.40163326e-01 -7.59901524e-01 -6.66161954e-01 2.19742060e+00
4.53392267e-01 1.35386623e-02 2.80458212e-01 7.67187119e-01
-3.95010322e-01 4.30581778e-01 8.76230448e-02 -6.01821363e-01
1.07699835e+00 -5.02953410e-01 -7.84970403e-01 -3.77743930e-01
6.34364545e-01 -1.59512669e-01 1.27256894e+00 4.82757986e-01
-1.08562684e+00 4.84827608e-02 -9.06647861e-01 6.51779950e-01
1.55045643e-01 -3.37079197e-01 8.21982443e-01 6.63351834e-01
-8.27454269e-01 7.67527997e-01 -1.09864938e+00 -3.05549771e-01
4.85846311e-01 4.32986051e-01 1.10771738e-01 1.11329183e-02
-1.11162877e+00 8.99392605e-01 2.44011506e-01 -3.20848674e-01
-1.22339058e+00 -8.54631484e-01 -5.14663517e-01 1.08898774e-01
1.22086847e+00 -7.16712415e-01 1.77236271e+00 -1.01195180e+00
-1.69402885e+00 -2.48993754e-01 1.50122687e-01 -9.47186768e-01
7.77867854e-01 -2.03951046e-01 -6.89857751e-02 1.34271845e-01
1.51305258e-01 4.14246321e-01 7.69218802e-01 -9.31232512e-01
-7.81782210e-01 -3.61739159e-01 6.46278918e-01 4.43439960e-01
-1.16870627e-01 -3.44228745e-01 2.15826929e-01 -3.86956304e-01
-4.47454095e-01 -1.17058158e+00 -8.98729384e-01 -1.89065173e-01
-7.91070312e-02 -5.00727668e-02 2.19675303e-01 -3.61427128e-01
1.12478936e+00 -1.67980075e+00 -1.07132040e-01 2.31131524e-01
1.00588687e-01 -1.04323089e-01 -1.21681631e-01 7.44483531e-01
4.43891734e-01 5.22587933e-02 -1.14111252e-01 5.88627048e-02
6.15536124e-02 4.99872923e-01 -4.38102245e-01 6.27360702e-01
2.91697569e-02 8.00565720e-01 -1.20068026e+00 -6.90272376e-02
1.60774350e-01 -1.37016907e-01 -9.06497121e-01 1.83385327e-01
-6.51711583e-01 4.92762655e-01 -9.00593758e-01 2.14996502e-01
2.94972271e-01 -4.53752875e-02 6.83712780e-01 4.65892851e-01
1.14381932e-01 2.88643926e-01 -1.23214245e+00 1.11938417e+00
-7.36367702e-01 1.41863704e-01 2.18038395e-01 -1.16898167e+00
4.16990548e-01 1.38819907e-02 8.04224491e-01 -7.51629651e-01
8.46977755e-02 -9.10883918e-02 2.73311555e-01 -4.04532582e-01
2.53860384e-01 -4.52983648e-01 -1.31592318e-01 7.13577449e-01
-2.74228990e-01 -4.96669710e-02 7.08364099e-02 7.58817941e-02
1.23873746e+00 2.89301485e-01 5.73030889e-01 -4.99323666e-01
-7.26924762e-02 1.33223087e-01 7.66778827e-01 1.10021007e+00
-3.01459372e-01 -2.79810548e-01 1.07737672e+00 -1.64077669e-01
-8.71839583e-01 -1.01279163e+00 7.56397024e-02 8.70058298e-01
-3.68295684e-02 -2.50011057e-01 -6.32248282e-01 -7.83867300e-01
5.86626291e-01 9.94444966e-01 -7.63239741e-01 -3.36830735e-01
-2.20470175e-01 -7.94911146e-01 2.69306302e-01 5.05123019e-01
2.64304698e-01 -6.67622864e-01 -1.23348045e+00 5.17843246e-01
7.50723332e-02 -8.04577529e-01 -4.99743283e-01 2.45772541e-01
-1.10075641e+00 -1.08863461e+00 -5.09231329e-01 2.76319742e-01
6.21870279e-01 2.07260624e-01 9.09601450e-01 -4.63605493e-01
1.99327454e-01 7.92138755e-01 -9.15107727e-02 -5.24542928e-01
-4.82894301e-01 -3.44257087e-01 3.47237617e-01 -9.76766273e-02
1.73532948e-01 -3.85131240e-01 -8.10497940e-01 2.05149636e-01
-9.06525612e-01 -1.93924665e-01 4.16346490e-01 9.78193104e-01
3.72236580e-01 -1.20409094e-02 1.00609386e+00 -8.88072073e-01
9.88110244e-01 -5.66989362e-01 -1.14922810e+00 2.69310534e-01
-8.80953670e-01 5.43210626e-01 9.26549673e-01 -7.61670053e-01
-8.92511964e-01 -3.63143571e-02 3.95610005e-01 -4.48702455e-01
2.61869103e-01 3.37089449e-01 4.38913852e-02 1.39641911e-01
6.56366646e-01 1.01149201e-01 6.18359566e-01 -2.94167399e-01
3.84824127e-01 4.75671232e-01 7.65846204e-03 -1.02066529e+00
3.29657376e-01 4.39881533e-01 3.02332491e-01 -5.23402512e-01
-6.82990730e-01 -5.66865467e-02 1.36517406e-01 -3.12010288e-01
3.82291675e-01 -9.09777403e-01 -1.29019070e+00 -3.15849692e-01
-2.89673001e-01 -7.50830352e-01 -6.72463894e-01 7.37204731e-01
-1.18517339e+00 2.41857365e-01 -3.32087338e-01 -1.35371637e+00
1.90593705e-01 -1.28281033e+00 6.16328061e-01 1.60747886e-01
-2.89240219e-02 -7.36041129e-01 3.24928373e-01 1.53413326e-01
1.92686424e-01 1.13630839e-01 6.82630897e-01 -4.90464747e-01
-3.86734575e-01 2.05718398e-01 2.85380781e-01 1.22035280e-01
1.95016548e-01 -3.01570237e-01 -6.17797792e-01 -7.80586302e-01
1.35970367e-02 -5.37089229e-01 6.44430995e-01 6.48898602e-01
1.10773981e+00 -9.10482168e-01 -1.54850394e-01 2.07057342e-01
1.30022562e+00 6.23961568e-01 2.07406208e-01 3.39276046e-01
6.53907508e-02 5.59230387e-01 1.16603458e+00 9.59943652e-01
3.96005124e-01 6.02678716e-01 4.57774460e-01 3.28313857e-01
4.99196023e-01 -6.96828663e-01 7.50688672e-01 -1.66021287e-01
6.32090345e-02 -1.27535686e-01 -7.04184949e-01 4.66895878e-01
-2.17925286e+00 -9.33201253e-01 5.41063786e-01 2.80220985e+00
9.18401062e-01 2.09897757e-01 8.06416035e-01 -3.06678176e-01
2.61131674e-01 -1.22842103e-01 -1.32260978e+00 -5.76598823e-01
4.40036595e-01 7.76459947e-02 1.12434399e+00 5.29375851e-01
-7.72861302e-01 6.47895157e-01 6.96053505e+00 7.63106883e-01
-8.72894406e-01 1.20173506e-01 7.09363997e-01 -5.04576802e-01
-2.95788109e-01 8.81363079e-02 -5.63176632e-01 4.59796339e-01
1.38552523e+00 -4.50402617e-01 9.60156143e-01 7.42587328e-01
7.66090214e-01 -5.92291653e-01 -1.15928876e+00 5.09614050e-01
-5.90911031e-01 -1.05742407e+00 -2.91200221e-01 4.61093128e-01
7.92831063e-01 -1.33345827e-01 -8.06890354e-02 5.46622694e-01
1.16288030e+00 -9.56419766e-01 5.33971190e-01 2.59338737e-01
6.43717110e-01 -1.18197238e+00 4.00668472e-01 6.31776273e-01
-7.28675604e-01 -6.70035422e-01 -4.04441386e-01 -3.29401046e-01
-2.04080939e-01 3.00235718e-01 -9.93901074e-01 3.84503037e-01
2.17029914e-01 4.27942514e-01 1.83816969e-01 8.08385968e-01
-4.42402847e-02 8.36122990e-01 -4.70743775e-01 -2.70994574e-01
4.22329128e-01 -4.06192392e-01 4.45274591e-01 6.25306785e-01
7.94808120e-02 9.88849699e-02 6.49984121e-01 7.08639443e-01
2.03432113e-01 2.34909318e-02 -8.01900089e-01 -4.33318287e-01
4.73865092e-01 7.42408693e-01 -6.62929356e-01 -2.85871625e-01
-3.81309092e-01 4.49394792e-01 2.86955714e-01 5.56308806e-01
-8.00440788e-01 5.10446616e-02 1.07150900e+00 -1.18968762e-01
1.51711836e-01 -3.12880903e-01 -3.89577374e-02 -1.14727008e+00
-2.79324502e-01 -1.33873940e+00 5.94932437e-01 -1.08899400e-01
-1.10519326e+00 -1.99202612e-01 2.83448786e-01 -1.21959555e+00
-7.38797426e-01 -2.99152255e-01 -1.75837040e-01 5.75471342e-01
-1.05272424e+00 -4.07998443e-01 6.09286308e-01 3.68640006e-01
5.32508194e-01 6.56922832e-02 4.37132150e-01 -1.68941781e-01
-5.35464764e-01 4.47025359e-01 4.90748316e-01 -4.77626473e-01
4.69623357e-01 -1.36903763e+00 -4.54775654e-02 6.39864802e-01
-1.25470936e-01 3.88853341e-01 1.02278292e+00 -8.49452972e-01
-1.77324569e+00 -9.77217674e-01 -1.91218495e-01 -2.59271443e-01
8.37622643e-01 -2.31831938e-01 -3.74408156e-01 7.91510224e-01
-8.79664198e-02 -2.05604136e-01 4.05615240e-01 2.32611120e-01
3.78217325e-02 -3.34255159e-01 -1.40155756e+00 9.57513690e-01
9.70358610e-01 -3.72575581e-01 -3.45807970e-01 3.40002030e-01
7.85945654e-01 -1.72445133e-01 -8.64257395e-01 1.51827633e-01
5.72361052e-01 -5.91145277e-01 8.45474064e-01 -1.24903345e+00
1.64620638e-01 -1.28053769e-01 -1.91491902e-01 -1.71788001e+00
-5.76510513e-03 -1.02989578e+00 -3.34613174e-01 5.76187193e-01
2.01127112e-01 -1.00067508e+00 6.90225005e-01 7.49692082e-01
5.01449406e-01 -8.71160746e-01 -1.00057721e+00 -1.26121533e+00
4.02328819e-01 -4.56495523e-01 5.92920780e-01 5.45473695e-01
1.74663022e-01 1.02942660e-02 -5.24638116e-01 1.21492870e-01
9.24850941e-01 2.13615835e-01 8.59687090e-01 -6.22926116e-01
-8.46006453e-01 -3.46599221e-01 4.84955162e-02 -1.18062603e+00
1.04443021e-01 -3.92404199e-01 1.24800757e-01 -1.30185008e+00
3.26166660e-01 -5.00113368e-01 -3.15677077e-01 5.94647527e-01
-1.44201308e-01 -3.69320959e-01 4.80383098e-01 -9.11978483e-02
-5.85078955e-01 7.08222568e-01 1.43356919e+00 8.19040164e-02
-5.94073713e-01 3.69977295e-01 -7.67551839e-01 4.41222399e-01
9.43225086e-01 -4.81116951e-01 -9.39417362e-01 1.16793603e-01
2.29735792e-01 7.34441519e-01 2.06801534e-01 -6.67712390e-01
-1.92371964e-01 -1.01645362e+00 5.56403100e-02 -2.46233121e-01
3.18677276e-02 -8.02703023e-01 1.42647624e-01 1.01875830e+00
-6.39697373e-01 6.71343431e-02 3.22785914e-01 1.01855743e+00
3.50412130e-01 -6.98654056e-02 6.80854023e-01 -1.91343412e-01
-1.29797816e-01 2.06019104e-01 -6.44629836e-01 3.44564795e-01
1.17451608e+00 2.26459160e-01 -1.76249474e-01 -9.18224216e-01
-4.36213881e-01 7.48158097e-01 3.22308898e-01 1.90969154e-01
2.56396592e-01 -9.70882177e-01 -4.78278428e-01 -1.17728695e-01
-2.68910736e-01 -3.80146414e-01 1.85748175e-01 8.55034232e-01
-3.49562615e-02 4.89322871e-01 -1.63357615e-01 -3.66459906e-01
-8.62242758e-01 8.09941292e-01 6.25926256e-01 -4.85465020e-01
-6.83862209e-01 1.22690327e-01 3.01965803e-01 -3.21537644e-01
2.29395807e-01 -6.73142850e-01 2.58686692e-01 -9.71206464e-03
5.16896665e-01 4.92703587e-01 -3.80194664e-01 8.12756717e-02
-1.97793826e-01 3.46561372e-02 4.27348465e-02 -5.69734693e-01
1.27207386e+00 -1.61688000e-01 5.95668316e-01 3.82324934e-01
6.83979571e-01 2.82481238e-02 -2.03537679e+00 -1.11209758e-01
1.63236726e-03 -6.55832648e-01 2.45840758e-01 -9.69328761e-01
-9.39027727e-01 4.43681985e-01 5.21680653e-01 2.82696486e-01
1.04690313e+00 -3.79048049e-01 5.10233819e-01 6.51310027e-01
8.66910040e-01 -1.24487627e+00 -1.71534851e-01 -2.37768814e-02
5.63352108e-01 -1.13365424e+00 2.04110309e-01 2.82339126e-01
-9.48749065e-01 7.77292907e-01 4.41262752e-01 -1.71877086e-01
4.02779460e-01 1.84403524e-01 -2.14548290e-01 1.65916368e-01
-1.25162625e+00 -4.40555274e-01 -2.30837941e-01 4.91206259e-01
-1.95385367e-01 5.66673160e-01 -4.70219254e-01 4.77592140e-01
1.31811015e-02 1.36359945e-01 8.80883217e-01 1.14365959e+00
-4.43371117e-01 -1.24931014e+00 -5.00145793e-01 9.10187960e-01
-5.80565512e-01 3.26557815e-01 8.93965922e-03 9.17878211e-01
-3.28995645e-01 1.01632690e+00 1.15571238e-01 -1.10827714e-01
2.41409719e-01 -1.24332659e-01 7.12589800e-01 -3.45845729e-01
-2.90640205e-01 2.46417299e-01 3.31415385e-01 -1.18776453e+00
-2.67595023e-01 -9.23243880e-01 -1.21399105e+00 -4.06100273e-01
2.81602219e-02 2.14933813e-01 3.08961779e-01 8.55643928e-01
3.86194766e-01 4.71362680e-01 8.89058173e-01 -3.36597234e-01
-1.57054651e+00 -5.49301565e-01 -6.56707585e-01 1.11900888e-01
7.50708699e-01 -1.02677512e+00 -3.53761017e-01 -6.68862581e-01] | [4.148134708404541, 2.561164617538452] |
e11811e7-3cd6-48c1-a351-cc2445c23cc3 | improving-the-domain-adaptation-of-retrieval | 2210.02627 | null | https://arxiv.org/abs/2210.02627v1 | https://arxiv.org/pdf/2210.02627v1.pdf | Improving the Domain Adaptation of Retrieval Augmented Generation (RAG) Models for Open Domain Question Answering | Retrieval Augment Generation (RAG) is a recent advancement in Open-Domain Question Answering (ODQA). RAG has only been trained and explored with a Wikipedia-based external knowledge base and is not optimized for use in other specialized domains such as healthcare and news. In this paper, we evaluate the impact of joint training of the retriever and generator components of RAG for the task of domain adaptation in ODQA. We propose \textit{RAG-end2end}, an extension to RAG, that can adapt to a domain-specific knowledge base by updating all components of the external knowledge base during training. In addition, we introduce an auxiliary training signal to inject more domain-specific knowledge. This auxiliary signal forces \textit{RAG-end2end} to reconstruct a given sentence by accessing the relevant information from the external knowledge base. Our novel contribution is unlike RAG, RAG-end2end does joint training of the retriever and generator for the end QA task and domain adaptation. We evaluate our approach with datasets from three domains: COVID-19, News, and Conversations, and achieve significant performance improvements compared to the original RAG model. Our work has been open-sourced through the Huggingface Transformers library, attesting to our work's credibility and technical consistency. | ['Suranga Nanayakkara', 'Rajib Rana', 'Tharindu Kaluarachchi', 'Elliott Wen', 'Rivindu Weerasekera', 'Shamane Siriwardhana'] | 2022-10-06 | null | null | null | null | ['open-domain-question-answering'] | ['natural-language-processing'] | [-8.86869952e-02 6.10680163e-01 8.00805092e-02 -3.93255830e-01
-1.43755054e+00 -9.09821570e-01 4.39438403e-01 1.86027229e-01
-5.61142802e-01 1.05950379e+00 4.67811286e-01 -1.87248051e-01
-1.49913535e-01 -7.39194453e-01 -9.02924180e-01 -6.70031756e-02
2.73397207e-01 1.26126194e+00 5.70153534e-01 -9.05251563e-01
-2.67446250e-01 -5.35561964e-02 -1.11564946e+00 4.91777211e-01
1.32797897e+00 8.30238998e-01 2.33869717e-01 6.97769523e-01
-1.48691311e-01 1.03497040e+00 -9.12697852e-01 -8.90591502e-01
-4.89375787e-03 -2.99568862e-01 -1.35374570e+00 -6.33991361e-01
3.52647305e-01 -4.35179204e-01 -3.91782850e-01 5.48899949e-01
1.06966865e+00 2.28480756e-01 5.52684844e-01 -7.41818130e-01
-9.51711118e-01 6.62844539e-01 1.98880941e-01 2.28124499e-01
6.27436638e-01 8.62288252e-02 1.13939595e+00 -7.18580008e-01
1.09405613e+00 1.16428316e+00 5.81099093e-01 8.88398528e-01
-9.11424577e-01 -3.23731393e-01 -1.80561263e-02 3.07623714e-01
-1.18966663e+00 -5.71060061e-01 4.86162484e-01 -8.22759122e-02
1.19251370e+00 1.52570173e-01 -2.44860910e-02 1.25305951e+00
-3.61060351e-01 1.06440723e+00 5.56631625e-01 -5.32546759e-01
2.99532056e-01 3.16981554e-01 2.98920304e-01 5.87621629e-01
-5.45318797e-03 -4.57279198e-02 -5.06231964e-01 -5.08148015e-01
2.64680356e-01 -7.20927000e-01 -4.80029017e-01 -2.07487702e-01
-9.49958980e-01 8.83210838e-01 3.61793309e-01 9.43995267e-03
-4.61591512e-01 -4.69189435e-02 4.51122731e-01 6.47983015e-01
4.34644818e-01 8.49792600e-01 -8.07659388e-01 -2.66571224e-01
-5.02963603e-01 6.44167602e-01 1.36291969e+00 1.15855539e+00
7.25080550e-01 -4.98834610e-01 -7.23133564e-01 1.12046039e+00
1.81299955e-01 7.38701999e-01 5.34561932e-01 -1.03761780e+00
8.44860137e-01 5.96266866e-01 2.71648675e-01 -4.69743460e-01
-4.79541719e-01 -5.13687789e-01 -2.54069149e-01 -6.36629283e-01
3.67557615e-01 -4.25657719e-01 -8.93600464e-01 1.75592268e+00
3.22050542e-01 -1.10973559e-01 6.67058229e-01 7.86934257e-01
1.48415506e+00 5.91527879e-01 1.45041957e-01 2.60903209e-01
1.48324239e+00 -1.10119605e+00 -7.52581298e-01 -4.23539281e-01
7.29264677e-01 -5.22346437e-01 1.05347753e+00 1.23938672e-01
-1.09601033e+00 -3.83650601e-01 -7.94932902e-01 -4.50546384e-01
-3.79379779e-01 -3.11469417e-02 2.96429306e-01 4.77477312e-01
-9.62430656e-01 5.24997041e-02 -4.28894669e-01 -4.04153824e-01
1.14725068e-01 7.52268210e-02 -1.64021477e-01 -4.41149950e-01
-1.89591920e+00 1.11385155e+00 4.41633612e-01 -3.49241346e-01
-8.85336876e-01 -7.38212168e-01 -8.04732203e-01 -1.82738900e-01
4.22775090e-01 -1.29143643e+00 1.79677773e+00 -5.70257843e-01
-1.71811962e+00 7.05671251e-01 -8.75090584e-02 -7.07609117e-01
3.92236650e-01 -3.06979299e-01 -6.79949880e-01 3.40845466e-01
3.05951834e-01 5.75426579e-01 5.54094195e-01 -1.00876296e+00
-2.70726144e-01 -3.37890923e-01 4.28404927e-01 5.22544086e-01
-9.29746628e-02 -2.10193083e-01 -8.96292746e-01 -5.97624481e-01
-2.68921077e-01 -7.29220748e-01 -4.35529463e-02 -6.53193057e-01
-2.83903658e-01 -3.55159342e-01 2.35310301e-01 -1.18924248e+00
1.38094282e+00 -1.87504923e+00 1.52908608e-01 6.90238327e-02
-1.23828173e-01 4.96486753e-01 -4.27892774e-01 7.22858608e-01
4.78563368e-01 -1.73651367e-01 -3.57812524e-01 -2.68305570e-01
1.62858337e-01 3.51057887e-01 -4.91615474e-01 -1.88484207e-01
2.98163086e-01 1.19911230e+00 -1.13351679e+00 -3.37919742e-01
-4.62141842e-01 2.93654352e-01 -8.18257868e-01 3.11934441e-01
-9.64513779e-01 3.96322340e-01 -7.79339671e-01 5.03395975e-01
3.74632269e-01 -4.06699032e-01 1.96466729e-01 -1.55712768e-01
5.14150977e-01 8.05478990e-01 -8.52755904e-01 2.10266447e+00
-5.43462992e-01 9.51486081e-02 -5.58039844e-02 -6.87221169e-01
9.41489756e-01 5.87434053e-01 1.16494503e-02 -1.11223578e+00
-5.93813322e-02 4.67028499e-01 -3.61182213e-01 -7.38003552e-01
8.45221519e-01 -7.26258531e-02 -3.34517479e-01 6.39363408e-01
5.37212551e-01 -4.36268806e-01 2.74879873e-01 6.15698338e-01
1.32465029e+00 3.32934298e-02 -1.10773639e-04 2.07329497e-01
5.35957158e-01 2.85999745e-01 8.00535530e-02 1.01060903e+00
1.35746017e-01 6.10198200e-01 2.43718237e-01 7.05416426e-02
-6.99182212e-01 -1.11684525e+00 -2.35654246e-02 1.38057208e+00
-5.45699596e-02 -1.53837278e-01 -8.07371795e-01 -1.22335017e+00
8.32312927e-02 1.11288202e+00 -4.74488139e-01 -4.09296185e-01
-6.31013811e-01 -4.36312139e-01 1.01495886e+00 2.39230826e-01
5.51583648e-01 -1.25305462e+00 -1.28534675e-01 5.16871929e-01
-7.93075085e-01 -1.19259715e+00 -4.32912648e-01 -3.70161310e-02
-6.59371972e-01 -1.02488208e+00 -8.02914619e-01 -6.97314858e-01
3.33447069e-01 -1.32199436e-01 1.76289034e+00 -1.54769167e-01
1.30339414e-01 8.44581068e-01 -7.17764258e-01 -6.04211152e-01
-7.93547392e-01 7.29714453e-01 -4.47768360e-01 -3.22823256e-01
4.51335996e-01 -2.06240073e-01 -4.17992115e-01 2.83679068e-01
-1.10664177e+00 -3.23574394e-01 4.95318800e-01 7.98936307e-01
4.56203789e-01 -6.05682254e-01 1.22509634e+00 -1.14297545e+00
1.17593336e+00 -7.79119313e-01 -3.48168701e-01 5.56629002e-01
-3.24284852e-01 4.41192567e-01 2.42240027e-01 -6.04347177e-02
-1.37443423e+00 -3.58301908e-01 -6.86087728e-01 4.60521318e-02
1.22216322e-01 9.33412611e-01 -1.23565368e-01 4.42697495e-01
1.24874687e+00 7.51031190e-02 8.58830884e-02 -7.85503685e-01
6.83124304e-01 9.30234373e-01 7.08355963e-01 -7.71191478e-01
5.90050161e-01 2.09110558e-01 -7.92405725e-01 -3.21224928e-01
-1.21523798e+00 -6.33999169e-01 -1.82080880e-01 2.78106451e-01
6.63914144e-01 -1.21611381e+00 -3.78062606e-01 2.26985916e-01
-1.31129396e+00 -4.12345707e-01 -6.30382299e-01 2.28417352e-01
-5.04060686e-01 1.83862209e-01 -3.54723305e-01 -3.52805108e-01
-8.46474409e-01 -6.61802113e-01 1.08413672e+00 1.32603675e-01
-4.52780314e-02 -1.18278635e+00 5.20531595e-01 9.23610866e-01
5.28405070e-01 -1.49506927e-01 9.00897443e-01 -1.12964177e+00
-3.04286152e-01 -2.44087487e-01 -5.32979183e-02 5.60620248e-01
-1.47308871e-01 -6.84434116e-01 -9.46728051e-01 -2.54425198e-01
-2.06750587e-01 -7.26752818e-01 9.00738418e-01 -1.57642052e-01
5.95163226e-01 -3.32540274e-01 -1.68414131e-01 2.24018127e-01
1.04580176e+00 -1.41461352e-02 5.87705910e-01 6.06507599e-01
2.01342016e-01 6.01733625e-01 7.12135494e-01 2.07292363e-01
1.00720716e+00 9.23422456e-01 2.18562827e-01 1.59088328e-01
-4.57619697e-01 -3.64881068e-01 4.49987829e-01 9.50015664e-01
1.97409123e-01 -5.34640551e-01 -9.77822244e-01 8.73725593e-01
-1.68784058e+00 -7.55051017e-01 2.67037719e-01 1.96458054e+00
1.41557372e+00 -1.63857847e-01 7.61241606e-03 -5.30323744e-01
3.17395031e-01 -1.10183291e-01 -7.64006197e-01 -3.36608112e-01
-2.97209769e-01 6.08553171e-01 2.15213776e-01 6.95453584e-01
-7.77200401e-01 9.84653831e-01 5.98061037e+00 7.01133430e-01
-6.39530122e-01 4.94655579e-01 7.59147555e-02 6.20187819e-02
-6.89930558e-01 -1.09810382e-01 -1.01260424e+00 1.85016081e-01
1.05373728e+00 -2.79033214e-01 3.64251733e-01 6.52995110e-01
-3.29448789e-01 1.61868840e-01 -9.92372453e-01 4.03649420e-01
3.15588295e-01 -1.29843962e+00 2.06342369e-01 -4.64801937e-01
6.02453887e-01 4.87851441e-01 -1.21618407e-02 1.00888634e+00
6.99984789e-01 -7.56718636e-01 5.57784915e-01 5.67832232e-01
6.80229485e-01 -5.12399077e-01 1.05755317e+00 4.60759848e-01
-4.05176759e-01 -5.26805893e-02 -2.35431746e-01 3.64808053e-01
4.76984680e-01 4.90595043e-01 -1.51358569e+00 1.11631739e+00
5.36975205e-01 2.24923253e-01 -6.36296511e-01 6.94742262e-01
-5.97513199e-01 6.93259060e-01 -2.36519784e-01 8.23907182e-02
1.23803847e-01 3.11393380e-01 6.69214427e-01 1.15474546e+00
6.62958026e-02 1.59029603e-01 -1.80406079e-01 6.15671396e-01
-4.83996660e-01 1.46273345e-01 -2.90718287e-01 -1.04378633e-01
6.56616151e-01 7.18917727e-01 4.44569618e-01 -4.79848713e-01
-3.83543700e-01 1.09604573e+00 6.21296704e-01 4.06939536e-01
-6.52832985e-01 -6.27781391e-01 4.04292226e-01 6.35771677e-02
4.51684982e-01 1.18787110e-01 2.13531882e-01 -1.22015214e+00
2.26875931e-01 -1.32616210e+00 9.64795709e-01 -7.98969746e-01
-1.56639898e+00 8.38329077e-01 1.84154266e-03 -7.71819293e-01
-7.42257416e-01 -2.80412704e-01 2.48284843e-02 1.03893340e+00
-1.97254205e+00 -1.04848027e+00 -1.51020408e-01 9.33029413e-01
2.69396156e-01 -3.60024452e-01 1.18895531e+00 5.82972825e-01
-1.75255224e-01 8.49429905e-01 4.02415395e-01 2.68110842e-01
1.05202436e+00 -1.21506894e+00 6.00747526e-01 3.96791130e-01
-1.41463960e-02 7.08876312e-01 6.78405643e-01 -6.53924167e-01
-1.35653341e+00 -1.11662781e+00 1.09653974e+00 -1.05308104e+00
5.22592783e-01 -3.77920777e-01 -1.03538454e+00 8.53331208e-01
2.00610355e-01 -2.19948441e-01 6.31988347e-01 2.29532436e-01
-5.56772530e-01 -1.06051803e-01 -1.27285135e+00 1.83800817e-01
9.47068036e-01 -6.74861789e-01 -1.17416751e+00 4.74750906e-01
1.19965708e+00 -7.96579897e-01 -1.02929115e+00 4.78750408e-01
2.32847348e-01 -5.44408381e-01 1.17871141e+00 -6.69229031e-01
2.02765167e-01 -2.05020100e-01 -1.60058439e-01 -1.48636591e+00
8.18723440e-02 -5.53092599e-01 -4.12921697e-01 1.20886087e+00
7.72128046e-01 -8.94499719e-01 4.95094359e-01 5.01276016e-01
-4.32292461e-01 -2.89213419e-01 -1.08749330e+00 -8.08354616e-01
2.52333999e-01 -2.74410963e-01 8.17745924e-01 8.11859429e-01
-6.72428310e-02 8.50734890e-01 -7.67175294e-03 2.18831152e-01
1.19592696e-01 -6.56103157e-03 9.94020283e-01 -1.00087857e+00
-7.22292423e-01 1.74201116e-01 2.23226920e-01 -1.36799359e+00
4.11771722e-02 -1.06765819e+00 2.71278583e-02 -1.96948969e+00
-1.15997426e-01 -6.85316980e-01 -3.60459425e-02 5.75744987e-01
-3.85963500e-01 -7.14718178e-02 3.73689495e-02 1.65071398e-01
-8.64942312e-01 6.84740543e-01 1.33150363e+00 -2.62029558e-01
-2.83659101e-01 -6.14830256e-02 -1.09506571e+00 1.07480839e-01
5.76325715e-01 -4.82244641e-01 -6.15010142e-01 -8.82517576e-01
4.69629377e-01 2.40015671e-01 2.42975771e-01 -6.74898922e-01
2.76237905e-01 4.14226532e-01 -2.32657805e-01 -3.90157461e-01
3.82066518e-01 -4.30766463e-01 -3.41391340e-02 1.07307434e-01
-3.39151382e-01 -1.15062676e-01 5.22349596e-01 4.69024599e-01
-4.88731503e-01 -4.88038868e-01 3.24822515e-01 -2.79949695e-01
-7.22432375e-01 -4.82767336e-02 -7.73859844e-02 8.14958513e-01
4.74026352e-01 2.39394814e-01 -8.69325817e-01 -7.14764655e-01
-7.08883345e-01 5.77966452e-01 9.08608884e-02 5.74804604e-01
2.66851187e-01 -1.01881611e+00 -1.14235163e+00 -1.15200924e-02
5.17630517e-01 2.75326371e-01 4.21164721e-01 6.17541015e-01
-3.41717005e-01 7.20673859e-01 1.66593075e-01 -1.85897097e-01
-7.26867199e-01 3.28379065e-01 3.21912080e-01 -8.80675793e-01
-2.34165132e-01 9.20855701e-01 -1.64596379e-01 -1.19902325e+00
2.06466734e-01 5.54332659e-02 -4.24313694e-01 8.97509009e-02
5.48669994e-01 2.72839010e-01 6.13163710e-01 -2.36397058e-01
-2.93896705e-01 -3.71197797e-02 -3.85114610e-01 -4.27072436e-01
1.30528951e+00 -1.67111963e-01 -5.75330146e-02 -5.12254573e-02
9.94918346e-01 9.44748819e-02 -6.05356514e-01 -6.83104396e-01
5.26763313e-02 3.30030844e-02 -1.95860356e-01 -1.65659940e+00
-5.56773305e-01 5.61464489e-01 4.27973866e-01 -1.20162807e-01
1.03205299e+00 3.66877049e-01 1.00112581e+00 8.57332826e-01
4.24261183e-01 -1.08137965e+00 3.12593073e-01 1.06305373e+00
1.19595313e+00 -1.06411743e+00 -1.62085712e-01 -4.64368910e-02
-9.58934844e-01 5.31642914e-01 5.85030913e-01 4.18019146e-01
2.75627941e-01 -3.47291052e-01 4.10575897e-01 -4.72157419e-01
-7.95651138e-01 -3.19690049e-01 4.76329416e-01 7.73525476e-01
3.40415776e-01 -1.63462028e-01 -2.94825882e-01 9.56621826e-01
-3.05917859e-01 2.56068021e-01 2.25946158e-01 9.96435046e-01
-2.97611624e-01 -1.32755733e+00 -1.18308425e-01 2.91703016e-01
-4.19913083e-01 -4.82571185e-01 -3.67769778e-01 6.33872211e-01
-6.57412857e-02 1.03048515e+00 -1.63226962e-01 4.40692231e-02
8.11065912e-01 4.02519464e-01 3.44975144e-01 -9.27144468e-01
-8.70825946e-01 -6.52978063e-01 8.01721334e-01 -3.55156183e-01
-2.07413435e-01 -5.72794259e-01 -1.21973670e+00 1.46318570e-01
-3.32896322e-01 7.04998732e-01 6.49066687e-01 7.36391187e-01
8.94010425e-01 4.47605938e-01 5.23580089e-02 1.17284907e-02
-8.61660779e-01 -1.14574015e+00 -7.83860907e-02 4.27367687e-01
3.33208799e-01 -4.81373459e-01 -4.83920872e-02 -2.78633311e-02] | [11.303743362426758, 7.987534999847412] |
cca61e03-02b2-46b2-af56-a1a19cf07359 | the-ubuntu-dialogue-corpus-a-large-dataset-1 | 1506.08909 | null | http://arxiv.org/abs/1506.08909v3 | http://arxiv.org/pdf/1506.08909v3.pdf | The Ubuntu Dialogue Corpus: A Large Dataset for Research in Unstructured Multi-Turn Dialogue Systems | This paper introduces the Ubuntu Dialogue Corpus, a dataset containing almost
1 million multi-turn dialogues, with a total of over 7 million utterances and
100 million words. This provides a unique resource for research into building
dialogue managers based on neural language models that can make use of large
amounts of unlabeled data. The dataset has both the multi-turn property of
conversations in the Dialog State Tracking Challenge datasets, and the
unstructured nature of interactions from microblog services such as Twitter. We
also describe two neural learning architectures suitable for analyzing this
dataset, and provide benchmark performance on the task of selecting the best
next response. | ['Ryan Lowe', 'Iulian Serban', 'Nissan Pow', 'Joelle Pineau'] | 2015-06-30 | the-ubuntu-dialogue-corpus-a-large-dataset | https://aclanthology.org/W15-4640 | https://aclanthology.org/W15-4640.pdf | ws-2015-9 | ['conversational-response-selection'] | ['natural-language-processing'] | [-3.39760870e-01 6.40305161e-01 -2.68608451e-01 -8.35556746e-01
-8.37513387e-01 -7.26247013e-01 9.80239034e-01 8.89010057e-02
-5.53210557e-01 1.32395637e+00 7.56911516e-01 -5.46274781e-01
4.85572815e-01 -5.24860859e-01 6.56719282e-02 -1.08948581e-01
-2.06609562e-01 1.15824294e+00 1.79947227e-01 -9.80796754e-01
1.78648904e-01 1.62408248e-01 -8.66018116e-01 6.25237346e-01
4.09585148e-01 7.89559007e-01 -3.22564058e-02 1.17679870e+00
-4.93092090e-01 1.51344013e+00 -9.37092364e-01 -3.87472123e-01
-1.95332617e-01 -4.22433794e-01 -1.61199141e+00 3.47249955e-02
1.37924537e-01 -9.45845366e-01 -5.91236472e-01 3.17386538e-01
6.93949163e-01 4.37054366e-01 5.12387633e-01 -1.08935297e+00
-3.25210035e-01 8.60692084e-01 1.64699778e-01 6.09846890e-01
7.09854603e-01 2.86110878e-01 1.16216373e+00 -4.31403130e-01
8.51724744e-01 1.61613011e+00 3.90714765e-01 1.02145410e+00
-1.05222201e+00 -8.59522223e-02 -4.88256104e-02 -1.98011339e-01
-5.13555110e-01 -1.01648188e+00 4.24678177e-01 -4.07820970e-01
1.53433681e+00 1.84356049e-01 1.98709056e-01 1.69302452e+00
-5.88488728e-02 1.17018771e+00 9.97517943e-01 -3.58304620e-01
5.17188162e-02 5.27975142e-01 6.68985188e-01 6.78183317e-01
-8.37566257e-01 -4.84873861e-01 -4.44980860e-01 -7.05303550e-01
3.56311589e-01 -5.14837563e-01 -1.09094650e-01 2.70402789e-01
-1.06384528e+00 1.23058414e+00 1.93600040e-02 2.20713690e-01
-1.31301537e-01 -3.46873105e-01 8.96876216e-01 6.66061223e-01
6.45186245e-01 6.72239482e-01 -8.71654868e-01 -8.32499146e-01
-1.43368512e-01 3.86718869e-01 1.83746028e+00 9.42192674e-01
6.30814850e-01 1.09431341e-01 -1.87335983e-01 1.36423767e+00
1.08028471e-01 2.07266033e-01 6.67003155e-01 -1.31823671e+00
7.47641861e-01 4.66113150e-01 3.19667786e-01 -5.69049597e-01
-7.74195015e-01 4.76408333e-01 -4.63186383e-01 -4.01329964e-01
7.88724840e-01 -9.77315545e-01 -5.35395779e-02 1.61469173e+00
3.67810220e-01 -6.63758337e-01 4.03401673e-01 4.47060794e-01
1.38186896e+00 8.69610369e-01 -2.01088302e-02 -3.79716128e-01
1.31318331e+00 -1.31123471e+00 -9.27928329e-01 -2.91788489e-01
1.01295972e+00 -6.31385326e-01 9.32558477e-01 8.68567303e-02
-1.15263879e+00 -8.78223032e-02 -6.17832720e-01 -3.41532767e-01
-5.89082301e-01 -3.33742321e-01 7.48698533e-01 5.74841022e-01
-1.18676591e+00 2.17698500e-01 -4.40593094e-01 -6.29486799e-01
-2.14532420e-01 1.36666104e-01 -1.45649061e-01 4.50756431e-01
-1.57347405e+00 1.35254657e+00 2.84893185e-01 -1.51538998e-01
-5.23505569e-01 -8.43851045e-02 -8.51080656e-01 -7.10348552e-03
3.82690996e-01 -1.94652341e-02 2.14732933e+00 -5.91449320e-01
-1.95565474e+00 7.72507787e-01 -8.47883746e-02 -5.96886933e-01
3.90348345e-01 -4.91139852e-02 -2.83410609e-01 1.90197825e-01
-8.91306102e-02 6.53528988e-01 6.25788346e-02 -7.95482755e-01
-5.70087314e-01 -1.88576177e-01 4.18894202e-01 5.77717185e-01
-4.82950717e-01 1.61541209e-01 -1.44250497e-01 -1.00515950e-02
-6.44106686e-01 -8.96082222e-01 -2.28892580e-01 -7.72505701e-01
-7.06995845e-01 -6.39529109e-01 8.31946611e-01 -7.61392891e-01
1.23179126e+00 -1.54229283e+00 -3.03929802e-02 -3.54807079e-01
3.18718195e-01 2.39653662e-01 -1.83402404e-01 1.04420376e+00
3.85336339e-01 1.83375269e-01 1.83427513e-01 -3.83334696e-01
1.11302666e-01 2.35415146e-01 -2.30167896e-01 -4.45357728e-04
3.29947732e-02 8.31193268e-01 -8.57501388e-01 -3.84287000e-01
3.93231306e-03 -1.08905226e-01 -2.35598937e-01 6.48310602e-01
-6.33274376e-01 5.21448910e-01 -5.61013937e-01 3.10889393e-01
-6.48337454e-02 -5.58403254e-01 3.88604939e-01 2.89025545e-01
-1.23926565e-01 1.00691712e+00 -3.24596226e-01 1.25889361e+00
-4.33484018e-01 9.72366989e-01 6.38430953e-01 -5.18848062e-01
8.56146336e-01 5.89437544e-01 3.47240090e-01 -4.82863247e-01
3.79785895e-01 -9.84271616e-03 1.34320080e-01 -6.36227787e-01
9.85056221e-01 2.02595577e-01 -4.11436319e-01 1.26383722e+00
3.02436650e-01 -3.97643149e-02 4.62211579e-01 6.77171052e-01
1.23282647e+00 -7.37045109e-01 3.75976980e-01 -1.42001987e-01
4.02875841e-01 1.42783046e-01 1.92920014e-01 8.01312983e-01
-5.75898767e-01 9.13378149e-02 9.81160700e-01 -7.27885485e-01
-1.06302416e+00 -4.80528057e-01 -1.45775676e-02 1.60052216e+00
-3.95874470e-01 -3.04906815e-01 -7.74418890e-01 -6.70113385e-01
-1.23603739e-01 6.12517476e-01 -4.53509927e-01 3.02972823e-01
-5.30483484e-01 -6.59111142e-01 9.12768781e-01 2.99558729e-01
6.99543118e-01 -1.49302280e+00 -2.07929268e-01 4.27812964e-01
-7.59058893e-01 -1.28425169e+00 -3.25097084e-01 2.93559879e-01
-5.68775237e-01 -1.11575615e+00 -3.67984325e-01 -7.76563883e-01
-5.91872111e-02 -8.05494934e-02 1.57597911e+00 3.45552415e-02
5.61689772e-02 5.07401764e-01 -3.38240951e-01 -2.13308886e-01
-1.15771008e+00 6.78903878e-01 2.17745695e-02 -3.43411475e-01
5.55369258e-01 -1.22556448e-01 -7.09813368e-03 2.79847592e-01
-3.27565104e-01 -5.00721000e-02 -4.12666090e-02 1.13416266e+00
-4.82338130e-01 -6.36095166e-01 1.14798808e+00 -1.15829420e+00
1.65653408e+00 -7.87538350e-01 -2.93027997e-01 3.34935337e-01
-1.96191370e-01 -2.11119533e-01 5.21235943e-01 -1.95434138e-01
-1.35269678e+00 -2.47495875e-01 -5.24164617e-01 5.34211040e-01
-3.06725472e-01 4.14082855e-01 3.32645237e-01 2.40730599e-01
7.47664690e-01 1.03393942e-01 4.83107656e-01 -5.70824921e-01
4.47994351e-01 1.43886828e+00 1.55060992e-01 -5.34999609e-01
-2.80628711e-01 -9.98526588e-02 -8.95610213e-01 -1.42061460e+00
-6.62974775e-01 -5.74624240e-01 -6.23299003e-01 -3.40623617e-01
8.09152186e-01 -7.23502398e-01 -1.24641001e+00 5.69118857e-01
-1.47628820e+00 -8.64584029e-01 1.82836562e-01 9.64626819e-02
-7.17267573e-01 3.49130005e-01 -1.61519480e+00 -1.05571795e+00
-6.90416634e-01 -1.12000823e+00 5.42169631e-01 2.11777419e-01
-7.90772498e-01 -1.36197221e+00 3.97923321e-01 6.58027112e-01
5.37569225e-01 -1.84711054e-01 1.05794191e+00 -1.52916765e+00
-1.21114552e-01 -1.40192032e-01 -7.22021908e-02 2.39975274e-01
2.02404946e-01 -3.75955482e-03 -1.00027966e+00 -2.53195822e-01
2.01569982e-02 -1.47452056e+00 3.58547717e-01 2.16393638e-02
4.22427863e-01 -7.47724235e-01 6.89474540e-03 -4.83667731e-01
5.99388063e-01 4.79674250e-01 9.67177972e-02 1.87783197e-01
3.98489565e-01 1.05693889e+00 2.55346060e-01 6.16026759e-01
8.60288560e-01 4.38385963e-01 2.38762394e-01 2.16937497e-01
4.54885691e-01 1.56841055e-01 1.50737435e-01 1.20859504e+00
5.47498882e-01 -5.70845783e-01 -1.23584580e+00 5.31955004e-01
-1.91480327e+00 -9.58431602e-01 5.09351678e-03 1.68869686e+00
1.33667529e+00 9.45236757e-02 5.31702220e-01 -5.65247595e-01
6.32592738e-01 6.01839185e-01 -5.82738578e-01 -1.08159637e+00
-4.58350964e-02 -3.22607249e-01 6.55458048e-02 9.99215782e-01
-1.21839821e+00 1.02297556e+00 7.79314613e+00 2.66601801e-01
-8.41319442e-01 -5.44217974e-02 9.63710546e-01 -6.59446940e-02
6.48403242e-02 -5.29055297e-01 -1.02171314e+00 4.51697797e-01
1.57364070e+00 -2.88231581e-01 5.45928240e-01 7.44940102e-01
1.50347918e-01 -2.17497453e-01 -1.01953816e+00 3.52890223e-01
-2.39590295e-02 -1.42246163e+00 -2.17018217e-01 -2.70394962e-02
4.46888477e-01 5.34814537e-01 -3.20094883e-01 7.42658675e-01
1.08374012e+00 -9.32682216e-01 -1.72835700e-02 8.16174597e-02
3.75937074e-01 -5.69658577e-01 7.15079725e-01 8.20712566e-01
-4.04849797e-01 -1.04275882e-01 -2.45287493e-01 -3.67619604e-01
3.09614241e-01 -7.22691640e-02 -1.58379412e+00 -2.24658430e-01
5.02843261e-01 5.39033473e-01 -2.46121123e-01 3.46408814e-01
1.58068836e-01 5.74605763e-01 -3.69690746e-01 -6.77605331e-01
6.30192161e-01 -1.50086567e-01 4.67712313e-01 1.32762837e+00
-6.55269742e-01 2.74248153e-01 3.93663466e-01 3.38569283e-01
-3.83447081e-01 1.76974609e-01 -9.71321821e-01 -4.32451487e-01
7.95026600e-01 1.25458086e+00 -3.70512992e-01 -5.66534102e-01
-5.49732327e-01 5.37420690e-01 6.83324873e-01 3.18155795e-01
-2.04178676e-01 -3.08279097e-01 6.48790956e-01 -4.73587036e-01
-2.46171921e-01 -2.79681295e-01 7.72370026e-02 -1.01384878e+00
-2.76570201e-01 -1.29480124e+00 3.32659394e-01 -4.47200030e-01
-1.46978903e+00 1.02307832e+00 -1.65733829e-01 -4.72177863e-01
-1.26474619e+00 -4.75452214e-01 -5.58351696e-01 8.41313839e-01
-9.66021657e-01 -7.53815949e-01 1.61411956e-01 4.32051212e-01
1.03141725e+00 -4.31729615e-01 1.26323438e+00 1.36077121e-01
-7.93426871e-01 3.54080200e-01 3.32333326e-01 7.98862636e-01
9.25626755e-01 -1.38832808e+00 7.62486279e-01 -2.44705394e-01
-4.38046992e-01 6.16732776e-01 5.63298643e-01 -3.91737103e-01
-1.13189697e+00 -5.75704634e-01 1.15332460e+00 -8.23157072e-01
9.53271806e-01 -5.05296886e-01 -9.39307153e-01 1.02049315e+00
8.51979196e-01 -6.91008627e-01 9.37222719e-01 5.22454321e-01
-1.11364990e-01 4.68586802e-01 -1.12096524e+00 6.71703815e-01
3.90961885e-01 -8.15550745e-01 -7.43163526e-01 7.68473327e-01
8.23567450e-01 -6.74761355e-01 -9.21592414e-01 4.85617928e-02
5.63597023e-01 -1.05092037e+00 5.83498359e-01 -1.24637580e+00
7.06528902e-01 9.32739139e-01 2.82406788e-02 -1.40838170e+00
2.02725619e-01 -8.90316486e-01 -3.64618301e-01 1.32885969e+00
7.90141702e-01 -5.38560808e-01 8.13788652e-01 1.31024384e+00
8.47215950e-02 -8.15385222e-01 -8.25300753e-01 -2.38104984e-01
3.18409264e-01 2.31527641e-01 2.62959272e-01 1.04506743e+00
6.15629554e-01 1.26481795e+00 -6.31910980e-01 -7.48237014e-01
4.27936129e-02 -3.48190963e-02 9.48352098e-01 -1.04902995e+00
-1.37411579e-01 -3.97168934e-01 2.60108024e-01 -1.61787570e+00
4.68944550e-01 -3.75762820e-01 1.56361237e-01 -1.26249909e+00
-1.32151231e-01 -3.17853808e-01 5.12993634e-01 4.16395098e-01
3.06077786e-02 -2.78148353e-01 5.66923898e-03 2.02931657e-01
-8.77416909e-01 7.62451768e-01 1.03762913e+00 -3.29209238e-01
-4.23567772e-01 1.60599858e-01 -4.52620953e-01 7.17785358e-01
8.28653753e-01 -9.29111540e-02 -3.45229357e-01 -4.49341774e-01
-1.35966644e-01 9.18140352e-01 -3.59176397e-01 -2.61900932e-01
2.12187648e-01 -7.00093359e-02 -2.17700616e-01 -5.37319124e-01
8.43280673e-01 -3.29469442e-01 -7.74096191e-01 1.42794952e-01
-1.18944943e+00 1.86062470e-01 1.88478723e-01 4.50908065e-01
-2.58939922e-01 -2.32178286e-01 6.30605519e-01 -6.28702581e-01
-6.78094685e-01 -1.59079015e-01 -1.13252330e+00 5.29000700e-01
6.62827253e-01 2.80024916e-01 -7.83468902e-01 -1.21729898e+00
-7.39134848e-01 8.17039251e-01 1.35251343e-01 7.53099740e-01
1.64362818e-01 -9.00706947e-01 -6.57107294e-01 -2.87248701e-01
-1.03247516e-01 -2.85479814e-01 1.38840079e-01 3.76381904e-01
-4.92828131e-01 7.45566487e-01 -3.50236535e-01 -3.19252789e-01
-1.32142210e+00 9.92603451e-02 4.59619522e-01 -4.56882536e-01
-4.32484508e-01 9.29679394e-01 -3.31897110e-01 -1.13902974e+00
4.48538274e-01 3.18669170e-01 -6.67087197e-01 5.05695879e-01
7.58652568e-01 4.28075582e-01 1.40429512e-02 -6.25471115e-01
5.08306772e-02 -7.64915407e-01 -5.48071563e-01 -5.08556604e-01
1.04591787e+00 -4.27300960e-01 -3.70402247e-01 9.52478290e-01
1.52408040e+00 -3.31254452e-01 -9.44297016e-01 -5.49154639e-01
2.64561385e-01 4.37067449e-02 -4.30747300e-01 -8.67935061e-01
-5.05655468e-01 7.03422606e-01 -7.56607503e-02 9.55241144e-01
1.36380687e-01 -1.19326688e-01 1.08717704e+00 1.28083730e+00
2.54520714e-01 -1.33312452e+00 3.82317096e-01 1.44155920e+00
8.68438244e-01 -1.58544278e+00 -2.86788434e-01 2.55592968e-02
-1.00419319e+00 1.13779950e+00 9.82061565e-01 3.20435673e-01
6.16613805e-01 2.16766387e-01 6.63382351e-01 -2.73166955e-01
-1.46720123e+00 5.38066700e-02 -3.51103693e-01 3.77617389e-01
7.68708646e-01 -1.27947062e-01 -1.69175729e-01 3.94662857e-01
-3.39617461e-01 -5.30452132e-01 8.81726027e-01 1.07622278e+00
-5.84201753e-01 -1.04147291e+00 9.53795239e-02 6.80446744e-01
-4.68220502e-01 -1.72621056e-01 -1.26756287e+00 6.69222534e-01
-1.17485750e+00 1.54709184e+00 8.52162167e-02 -4.79987830e-01
1.35615557e-01 6.42343283e-01 -2.28872299e-01 -8.31598580e-01
-1.11904049e+00 -3.41145664e-01 1.25998175e+00 -4.55295503e-01
-2.60872245e-01 -3.23135644e-01 -9.98946786e-01 -8.18351507e-01
-2.13240460e-01 5.64935386e-01 5.00847161e-01 9.15629447e-01
2.27654025e-01 1.17372960e-01 8.82225692e-01 -8.27447116e-01
-9.71882999e-01 -1.60125959e+00 -4.15346205e-01 2.79890418e-01
4.32163656e-01 -2.96834648e-01 -2.51999766e-01 -2.49679700e-01] | [12.75489330291748, 7.967925548553467] |
82568499-547c-4803-b941-3bef93e77bc1 | enhancing-low-resource-ner-using-assisting | 2306.06477 | null | https://arxiv.org/abs/2306.06477v1 | https://arxiv.org/pdf/2306.06477v1.pdf | Enhancing Low Resource NER Using Assisting Language And Transfer Learning | Named Entity Recognition (NER) is a fundamental task in NLP that is used to locate the key information in text and is primarily applied in conversational and search systems. In commercial applications, NER or comparable slot-filling methods have been widely deployed for popular languages. NER is used in applications such as human resources, customer service, search engines, content classification, and academia. In this paper, we draw focus on identifying name entities for low-resource Indian languages that are closely related, like Hindi and Marathi. We use various adaptations of BERT such as baseBERT, AlBERT, and RoBERTa to train a supervised NER model. We also compare multilingual models with monolingual models and establish a baseline. In this work, we show the assisting capabilities of the Hindi and Marathi languages for the NER task. We show that models trained using multiple languages perform better than a single language. However, we also observe that blind mixing of all datasets doesn't necessarily provide improvements and data selection methods may be required. | ['Dipali Kadam', 'Raviraj Joshi', 'Parth Patil', 'Onkar Litake', 'Aparna Ranade', 'Maithili Sabane'] | 2023-06-10 | null | null | null | null | ['named-entity-recognition-ner', 'slot-filling'] | ['natural-language-processing', 'natural-language-processing'] | [-5.98108053e-01 -2.03633547e-01 -3.13640833e-01 -2.81081468e-01
-1.07027388e+00 -9.85199213e-01 7.75088966e-01 4.01798755e-01
-1.03520322e+00 1.26993573e+00 4.55809593e-01 -4.25462902e-01
5.56770116e-02 -5.50965905e-01 -2.59388030e-01 -2.37713024e-01
1.83143869e-01 9.18666184e-01 2.50408679e-01 -5.25473773e-01
1.48442134e-01 7.46704042e-01 -7.97462583e-01 1.11972809e-01
1.02872658e+00 1.67067856e-01 3.26200604e-01 3.16945881e-01
-8.13410103e-01 9.03322041e-01 -8.09936464e-01 -6.77361667e-01
-1.36216916e-02 -2.05013812e-01 -1.28246725e+00 -6.14562273e-01
1.01797402e-01 2.01069519e-01 -3.22395593e-01 8.39714289e-01
7.24565506e-01 3.88874114e-01 4.50837970e-01 -9.48899090e-01
-2.54985780e-01 1.07470858e+00 -4.21286196e-01 2.96933323e-01
3.91142190e-01 -6.86290681e-01 1.03756070e+00 -8.52402985e-01
1.12786794e+00 1.02237630e+00 6.88961387e-01 6.10239029e-01
-7.24642694e-01 -5.82514048e-01 -1.45318866e-01 1.60456643e-01
-1.76343358e+00 -5.82625628e-01 4.83718067e-01 -4.53363918e-02
8.27353299e-01 2.83111691e-01 -1.27967641e-01 8.57594788e-01
-1.81883603e-01 9.29434478e-01 1.01784515e+00 -6.73970759e-01
4.72291559e-02 5.08853376e-01 9.59767997e-02 3.29812020e-01
7.47817457e-02 -5.83803773e-01 -6.61151350e-01 -4.12069559e-01
5.21164238e-01 -1.37837887e-01 -1.80761382e-01 4.11415920e-02
-1.39001596e+00 8.77703309e-01 9.81058255e-02 8.52420330e-01
-4.73603308e-01 -4.43785906e-01 4.10180479e-01 2.66156673e-01
4.07461882e-01 8.33100080e-01 -8.76598537e-01 -3.96779358e-01
-9.41945434e-01 -3.12757269e-02 1.42974377e+00 1.12124288e+00
6.95091188e-01 -2.82389939e-01 -5.48456833e-02 1.40281165e+00
1.48697877e-02 5.12200117e-01 6.30696774e-01 -6.54298186e-01
5.22800386e-01 4.99247789e-01 4.71454442e-01 -5.59107065e-01
-6.65462673e-01 -1.80846274e-01 -5.54951787e-01 -6.35403216e-01
4.85352874e-01 -3.01660925e-01 -8.24771881e-01 1.58055544e+00
4.04872447e-01 -1.80299938e-01 5.42872727e-01 6.06304049e-01
9.35673594e-01 6.48500144e-01 3.74204785e-01 -1.10131845e-01
1.64380991e+00 -9.86416161e-01 -7.37745225e-01 -2.64467299e-01
8.60652506e-01 -1.31765676e+00 5.75591505e-01 -2.05742314e-01
-7.48023629e-01 -2.03952074e-01 -4.40497667e-01 -1.98572949e-01
-7.63214707e-01 4.10423100e-01 6.00848973e-01 6.36829019e-01
-8.77726138e-01 5.41202545e-01 -6.81097031e-01 -1.10085535e+00
-2.67692894e-01 2.08144754e-01 -7.73788691e-01 -9.47947353e-02
-1.49930763e+00 1.14185977e+00 5.18458903e-01 -1.69302255e-01
-4.76074070e-01 -3.34467202e-01 -8.56271744e-01 -1.54020011e-01
7.60105327e-02 -1.12795211e-01 1.24525559e+00 -3.07722986e-01
-1.10543644e+00 1.08120048e+00 -4.29968864e-01 -3.90076399e-01
2.18961000e-01 -3.04031640e-01 -9.05465424e-01 -7.45472685e-02
6.93722248e-01 3.60634178e-01 1.19956605e-01 -8.27248096e-01
-9.84294891e-01 -2.12894082e-01 3.66372690e-02 2.86641568e-01
-2.02329203e-01 8.05260181e-01 -4.99052703e-01 -5.88102877e-01
1.99599564e-01 -9.29184318e-01 -2.84347624e-01 -7.34031439e-01
-5.36890209e-01 -4.57184047e-01 3.86761725e-01 -1.01993549e+00
1.28638005e+00 -2.06956887e+00 -3.32797706e-01 1.74414188e-01
-7.31970891e-02 2.59315699e-01 -1.36876747e-01 9.73088920e-01
1.51229516e-01 3.55961055e-01 8.46653059e-02 -1.67546019e-01
2.11249441e-02 3.05543512e-01 -1.55252323e-01 1.93528578e-01
2.21543629e-02 7.17827737e-01 -7.48374820e-01 -7.22995758e-01
-1.39357597e-01 4.47797626e-01 -9.30334851e-02 1.48080736e-01
1.55901715e-01 5.23282170e-01 -4.32515621e-01 6.19873464e-01
3.44696224e-01 -1.38897166e-01 4.30155933e-01 2.20267405e-03
-5.34118772e-01 9.09069598e-01 -1.23268962e+00 1.73125041e+00
-7.11240053e-01 5.82558453e-01 1.19203523e-01 -6.96282625e-01
9.19036806e-01 5.97374499e-01 2.82078147e-01 -7.14330792e-01
-1.13713220e-01 6.97230935e-01 -8.27468783e-02 -2.81791151e-01
9.19128537e-01 8.94962773e-02 -4.34077054e-01 4.89335626e-01
2.20564261e-01 2.66229481e-01 6.27936721e-01 1.51830971e-01
9.54167545e-01 -1.61561444e-02 6.69528723e-01 -3.13780069e-01
6.94278538e-01 3.85451138e-01 8.27101409e-01 9.53390896e-01
-2.54758775e-01 3.51842344e-01 1.45488992e-01 -2.34491691e-01
-1.03888583e+00 -7.34787107e-01 -3.25113535e-01 1.51186037e+00
-2.57118996e-02 -4.98567015e-01 -5.43410659e-01 -7.89553583e-01
-4.09410805e-01 8.85779798e-01 -1.80770308e-01 4.46448296e-01
-8.97691429e-01 -4.59582120e-01 9.56368923e-01 4.01845604e-01
5.61613977e-01 -1.16158116e+00 5.49497157e-02 4.81287390e-01
-6.12413526e-01 -1.42979276e+00 -5.68770468e-01 3.91490012e-01
-4.83352214e-01 -8.42324853e-01 -1.16944373e+00 -1.15679610e+00
3.35066885e-01 1.48475468e-01 1.14659262e+00 -3.36141676e-01
1.11126021e-01 3.21594745e-01 -5.95545709e-01 -3.40303570e-01
-4.51705933e-01 8.40447903e-01 1.51663750e-01 -3.68962228e-01
6.96593523e-01 -3.30522358e-01 -2.34451428e-01 6.55181766e-01
-6.12790704e-01 -3.45267624e-01 5.56773841e-01 8.00929904e-01
1.93249136e-01 -1.92961812e-01 5.10423005e-01 -1.18353474e+00
4.61435676e-01 -5.23534894e-01 -2.71106780e-01 6.64267063e-01
-3.60484183e-01 2.89885223e-01 4.62838382e-01 -2.97698051e-01
-1.24243546e+00 -1.64906308e-02 -5.62121391e-01 3.84105474e-01
-4.81373936e-01 7.24855125e-01 -1.62685290e-01 -1.59596533e-01
8.20585728e-01 1.58549145e-01 -7.98008025e-01 -9.96637642e-01
2.94881880e-01 1.09995770e+00 3.85056943e-01 -7.08885670e-01
5.63813567e-01 1.74280271e-01 -5.27396023e-01 -1.16120183e+00
-8.29581380e-01 -1.12516558e+00 -6.74836099e-01 1.59335926e-01
6.99458122e-01 -1.19836473e+00 -3.38801622e-01 1.35403588e-01
-1.50943065e+00 -7.36277699e-02 -1.33920386e-01 8.38582337e-01
-4.93098162e-02 2.42721170e-01 -9.66326416e-01 -7.11056054e-01
-3.43524337e-01 -6.61630392e-01 8.99747252e-01 6.98954821e-01
-3.36238474e-01 -1.15157688e+00 3.79230708e-01 3.43230605e-01
5.02770185e-01 -3.80779058e-01 8.75824332e-01 -1.70059681e+00
8.56447145e-02 -1.71391562e-01 -1.53102130e-01 -8.94256681e-02
1.66613862e-01 -4.38471407e-01 -7.29580700e-01 -1.09324433e-01
-3.60330909e-01 -1.54527575e-01 4.49062794e-01 -2.08061069e-01
1.12069525e-01 -2.15737090e-01 -3.90172571e-01 2.16513708e-01
1.14755774e+00 4.46349293e-01 5.22980332e-01 6.91815078e-01
6.89806759e-01 8.69745255e-01 5.21032333e-01 2.80674398e-01
6.93632007e-01 4.50069457e-01 -4.81185228e-01 -2.58981943e-01
1.64406985e-01 -4.66881126e-01 2.33088657e-01 1.11064422e+00
-1.03718258e-01 -1.21316098e-01 -1.36388135e+00 9.05054092e-01
-1.69755530e+00 -7.74005353e-01 -8.52806419e-02 2.15591764e+00
1.23546410e+00 -2.89928794e-01 -7.56370947e-02 -4.65658993e-01
1.04989171e+00 -6.15139268e-02 -1.38373017e-01 -2.38688439e-01
-4.29662615e-01 5.17966509e-01 6.91951394e-01 2.81260937e-01
-1.18357432e+00 1.36576414e+00 5.90139294e+00 9.07254517e-01
-1.04073858e+00 3.74199927e-01 3.23735297e-01 6.14494920e-01
-6.61081672e-02 3.03212255e-01 -1.30856204e+00 1.11727849e-01
1.13049889e+00 -2.93830365e-01 2.83197373e-01 9.33792233e-01
-1.58133097e-02 1.92027669e-02 -7.94778407e-01 9.82827783e-01
9.43935476e-03 -9.92430747e-01 -3.26663196e-01 -1.03682429e-01
8.53119791e-01 5.20269275e-01 -6.57511950e-01 5.87605417e-01
5.57254791e-01 -6.83624148e-01 4.26087826e-01 9.85253975e-02
4.38308239e-01 -7.31489420e-01 1.01378429e+00 3.75812709e-01
-1.15455031e+00 3.29038799e-01 -4.63420808e-01 4.28831398e-01
3.53620827e-01 3.49187285e-01 -9.14597511e-01 7.00692952e-01
6.36753201e-01 1.18717238e-01 -3.88178378e-01 1.30421615e+00
-5.51227629e-01 7.18377531e-01 -5.12739420e-01 -1.86258703e-01
1.30807698e-01 4.72319771e-05 4.19724047e-01 1.39195359e+00
3.00970644e-01 -9.53774452e-02 3.00520211e-01 1.63203269e-01
-3.44839364e-01 8.99057806e-01 -4.24327940e-01 -4.70340788e-01
7.07915306e-01 1.32081044e+00 -8.12238634e-01 -3.32831979e-01
-5.69701612e-01 1.22303951e+00 3.25802952e-01 4.31289047e-01
-4.95433509e-01 -8.80567908e-01 5.68872869e-01 -7.28610978e-02
2.89868921e-01 -5.15985072e-01 1.36391461e-01 -1.35270119e+00
-2.77234793e-01 -8.77102375e-01 6.76878691e-01 -4.92270589e-01
-1.28838181e+00 9.64062870e-01 -2.49884635e-01 -9.64990079e-01
-5.08300424e-01 -4.89551306e-01 -3.71629477e-01 1.00836682e+00
-1.61241448e+00 -1.17197347e+00 1.95928395e-01 7.28908122e-01
3.42404664e-01 -2.72263885e-01 1.14866853e+00 8.57002079e-01
-5.94091415e-01 5.43173730e-01 4.38052684e-01 8.97413254e-01
1.34105659e+00 -1.14075398e+00 6.28716409e-01 8.45824003e-01
6.73432827e-01 1.06403148e+00 6.42362893e-01 -6.98530436e-01
-8.48173440e-01 -7.97316909e-01 2.01934814e+00 -2.58978128e-01
9.10038829e-01 -2.91298062e-01 -7.23223507e-01 6.58949137e-01
4.20536458e-01 -3.08650672e-01 9.25932944e-01 5.50816178e-01
-3.13214451e-01 1.02087498e-01 -1.02221525e+00 6.12019956e-01
7.30164587e-01 -8.02511692e-01 -7.37388253e-01 4.85913843e-01
5.58120966e-01 -3.48470300e-01 -9.84720170e-01 -2.98932642e-02
2.49762937e-01 -5.11587203e-01 6.72708631e-01 -8.63870263e-01
-2.16717422e-01 -3.04446906e-01 -8.57468322e-02 -1.19941592e+00
-1.51393339e-01 -7.05302417e-01 4.20241356e-01 1.91348279e+00
8.97809923e-01 -7.71201074e-01 6.47122443e-01 8.05444241e-01
2.20126644e-01 2.11026922e-01 -1.00654221e+00 -7.99410164e-01
8.13582912e-02 -2.54170507e-01 5.11087418e-01 1.36569345e+00
1.35577753e-01 7.84255624e-01 -3.96636158e-01 2.24417105e-01
3.49197835e-01 4.06064205e-02 5.76392949e-01 -1.31195116e+00
1.55317426e-01 2.97256541e-02 -2.81590611e-01 -9.73097563e-01
3.41082573e-01 -8.37223887e-01 1.64821371e-01 -1.60878217e+00
8.23978800e-03 -9.14260268e-01 -4.22871143e-01 8.28127801e-01
-3.94097604e-02 1.42039910e-01 1.02558084e-01 6.43282354e-01
-6.78067565e-01 2.05952153e-01 5.86245954e-01 4.69331210e-03
-3.61268520e-01 2.84166541e-03 -6.80784523e-01 5.30165970e-01
8.56600523e-01 -8.51572871e-01 2.37178087e-01 -4.33310300e-01
3.96829545e-01 -2.94383503e-02 -2.89220035e-01 -8.09662461e-01
6.79729462e-01 -2.10548699e-01 7.82801583e-02 -3.08292508e-01
-5.07997610e-02 -5.22438049e-01 -5.83362542e-02 1.20358445e-01
-2.40295112e-01 1.30699828e-01 -1.50134610e-02 3.62150967e-02
-5.25242150e-01 -5.79622507e-01 4.90075171e-01 -3.39721620e-01
-1.10709321e+00 1.44726038e-01 -3.84015054e-01 5.29394567e-01
5.56182265e-01 3.61411572e-01 -2.16329768e-01 -3.67906272e-01
-5.04477620e-01 2.21397713e-01 3.77399653e-01 4.54863310e-01
-1.26549480e-02 -1.18003798e+00 -8.43161464e-01 -1.73362717e-01
3.68529081e-01 -5.29578984e-01 2.66503766e-02 8.61210942e-01
-7.58274376e-01 7.65751839e-01 -6.56052157e-02 -7.59485662e-02
-9.63523924e-01 1.21458940e-01 -1.32436976e-01 -5.41778266e-01
-1.25144213e-01 6.75609589e-01 -1.19401850e-01 -1.00756192e+00
8.45929012e-02 3.05557698e-01 -7.01134562e-01 2.72174567e-01
5.44235945e-01 2.93889016e-01 1.60565466e-01 -1.00135791e+00
-7.30173230e-01 1.90478459e-01 -3.12970638e-01 -4.70784664e-01
1.18831217e+00 -3.68177414e-01 -3.59625638e-01 4.47250813e-01
9.29959834e-01 6.07226968e-01 -1.32624984e-01 -5.67434371e-01
7.21121728e-01 8.34569409e-02 -3.07543367e-01 -7.67748356e-01
-5.80534041e-01 5.56269169e-01 3.25884521e-01 1.81561351e-01
6.92092478e-01 1.96016192e-01 8.83233786e-01 8.05731714e-01
8.18371475e-01 -1.43115640e+00 -8.61158907e-01 1.03596210e+00
2.73706138e-01 -1.36365724e+00 -3.65062416e-01 -2.21578747e-01
-7.45758355e-01 9.94130015e-01 2.99135834e-01 4.69742686e-01
5.46350241e-01 7.86036476e-02 5.54654121e-01 9.84683633e-02
-3.58076394e-01 -7.20766306e-01 2.28580475e-01 4.94494945e-01
8.95987034e-01 -1.13714309e-02 -9.06864941e-01 6.47562861e-01
-2.35211387e-01 -4.11265880e-01 3.83030325e-01 1.20353687e+00
-3.12624454e-01 -1.56408370e+00 -4.13412809e-01 1.87415332e-01
-9.93837357e-01 -6.42818332e-01 -3.97028863e-01 8.37482095e-01
-1.10620230e-01 1.10436642e+00 -1.71437800e-01 -4.82711606e-02
3.83205444e-01 3.93775314e-01 -4.00676802e-02 -7.07200348e-01
-8.24864388e-01 5.31644821e-02 5.90327501e-01 -1.38206333e-01
-5.95783412e-01 -5.18514037e-01 -1.19870651e+00 -3.25473607e-01
-4.82899070e-01 1.13868761e+00 1.03122818e+00 9.88262713e-01
3.27119976e-01 -2.09463820e-01 4.98368323e-01 -2.18012035e-01
-1.99702561e-01 -9.95146334e-01 -7.70410657e-01 2.42017433e-01
-2.05909297e-01 -3.58421862e-01 -2.61206955e-01 -2.84407157e-02] | [9.877628326416016, 9.75691032409668] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.