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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
28f94215-1d6e-44e6-8376-6ae88c465137 | planarrecon-real-time-3d-plane-detection-and-1 | 2206.07710 | null | https://arxiv.org/abs/2206.07710v1 | https://arxiv.org/pdf/2206.07710v1.pdf | PlanarRecon: Real-time 3D Plane Detection and Reconstruction from Posed Monocular Videos | We present PlanarRecon -- a novel framework for globally coherent detection and reconstruction of 3D planes from a posed monocular video. Unlike previous works that detect planes in 2D from a single image, PlanarRecon incrementally detects planes in 3D for each video fragment, which consists of a set of key frames, from a volumetric representation of the scene using neural networks. A learning-based tracking and fusion module is designed to merge planes from previous fragments to form a coherent global plane reconstruction. Such design allows PlanarRecon to integrate observations from multiple views within each fragment and temporal information across different ones, resulting in an accurate and coherent reconstruction of the scene abstraction with low-polygonal geometry. Experiments show that the proposed approach achieves state-of-the-art performances on the ScanNet dataset while being real-time. | ['Huaizu Jiang', 'Xiaowei Zhou', 'Fengting Yang', 'Matheus Gadelha', 'Yiming Xie'] | 2022-06-15 | planarrecon-real-time-3d-plane-detection-and | http://openaccess.thecvf.com//content/CVPR2022/html/Xie_PlanarRecon_Real-Time_3D_Plane_Detection_and_Reconstruction_From_Posed_Monocular_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Xie_PlanarRecon_Real-Time_3D_Plane_Detection_and_Reconstruction_From_Posed_Monocular_CVPR_2022_paper.pdf | cvpr-2022-1 | ['3d-plane-detection'] | ['computer-vision'] | [ 3.65877114e-02 -1.34913847e-01 7.53880814e-02 -3.26247633e-01
-5.78609645e-01 -5.60069084e-01 3.73540699e-01 -7.52598420e-02
-2.97927618e-01 2.65376896e-01 -3.71202886e-01 1.46942690e-01
-1.35673761e-01 -9.51057971e-01 -1.02823985e+00 -3.42349917e-01
-2.91167110e-01 5.91338933e-01 9.80143964e-01 1.65989891e-01
6.61434233e-02 1.06372416e+00 -1.54278457e+00 4.12939340e-01
1.43836677e-01 1.20107114e+00 3.95843416e-01 7.38760412e-01
1.11767031e-01 6.88810647e-01 -9.85332206e-02 1.44356107e-02
6.53205395e-01 1.04121700e-01 -4.25327569e-01 4.37974691e-01
1.17017305e+00 -7.76971698e-01 -6.31773174e-01 8.98885727e-01
1.02437086e-01 -1.40984923e-01 3.71278435e-01 -8.04984152e-01
3.22024256e-01 6.71524853e-02 -7.06427097e-01 8.52629244e-02
8.05461884e-01 -8.60292241e-02 6.48638129e-01 -1.30335748e+00
1.24154699e+00 1.20904565e+00 6.95653796e-01 -8.79087895e-02
-9.75923419e-01 -4.16855186e-01 3.56074780e-01 1.80799216e-01
-1.37959349e+00 -2.72597939e-01 8.98874819e-01 -4.51864421e-01
1.05899024e+00 1.21853575e-01 1.05374467e+00 3.22737843e-01
4.50654626e-01 5.19361556e-01 5.11143804e-01 -3.26436430e-01
1.09535649e-01 -4.13234115e-01 -2.19284967e-01 1.00638771e+00
2.97387213e-01 3.15117329e-01 -8.15316319e-01 1.21268906e-01
1.55581760e+00 3.11405003e-01 -4.72268432e-01 -1.13476419e+00
-1.49916542e+00 4.06459749e-01 4.85299051e-01 1.21242590e-02
-8.18628430e-01 3.44603032e-01 6.38210103e-02 2.06965059e-01
5.40357471e-01 2.10567191e-02 -3.93673092e-01 2.91949928e-01
-1.01511598e+00 3.41752231e-01 6.71087742e-01 1.08045661e+00
9.92038071e-01 3.04293305e-01 3.52884561e-01 1.92893416e-01
2.79758513e-01 6.42882824e-01 -8.27485248e-02 -1.43402934e+00
3.90853792e-01 6.51544452e-01 1.30403072e-01 -1.19607818e+00
-7.71484911e-01 -2.66680777e-01 -5.18170238e-01 6.02013707e-01
1.89423904e-01 -1.45344296e-03 -7.94086099e-01 1.09343088e+00
8.38892281e-01 1.73957944e-01 -2.72437751e-01 1.04049158e+00
9.57559705e-01 5.64761043e-01 -7.53620803e-01 -2.29796097e-01
1.06185091e+00 -8.49804461e-01 -2.59087056e-01 -2.60500640e-01
-1.71920508e-02 -6.56850159e-01 -8.22984129e-02 7.42724121e-01
-1.45422065e+00 -6.58111393e-01 -1.08657157e+00 -1.03705842e-03
8.50761086e-02 -9.26373601e-02 2.89263129e-01 9.01507512e-02
-1.09034252e+00 3.05769563e-01 -9.97093201e-01 -2.97244668e-01
1.39084741e-01 1.38619125e-01 -9.17708635e-01 -2.87503839e-01
-3.22217792e-01 4.56910074e-01 4.96276677e-01 -4.80111176e-03
-9.75333631e-01 -7.98579633e-01 -1.12861729e+00 -1.24283925e-01
7.63878405e-01 -1.01421177e+00 1.24716711e+00 -6.81935430e-01
-1.17648041e+00 8.06841791e-01 -4.28316206e-01 -5.63056052e-01
6.84239924e-01 -6.12874627e-01 5.32041602e-02 7.78897345e-01
-2.77477619e-03 7.20347524e-01 8.34521174e-01 -1.46644425e+00
-9.94648337e-01 -5.80081522e-01 3.15618925e-02 6.07066989e-01
5.77173948e-01 -1.76750734e-01 -1.01645732e+00 -1.45342276e-01
1.09363008e+00 -6.43054307e-01 -2.88744926e-01 4.55577612e-01
-3.00095737e-01 2.22994938e-01 1.02979410e+00 -5.49169660e-01
5.71961284e-01 -1.79680872e+00 3.51709187e-01 1.62399948e-01
5.01510739e-01 -3.60883065e-02 2.72780657e-01 1.64759204e-01
-1.61405638e-01 -4.18143123e-01 5.77429235e-02 -5.88903606e-01
-5.84423780e-01 1.01248629e-01 -7.78569430e-02 1.02297533e+00
-1.52993843e-01 5.95018685e-01 -8.36216450e-01 -3.89599025e-01
1.05376434e+00 4.16515172e-01 -6.19876564e-01 2.59123534e-01
-3.24558496e-01 3.90419364e-01 -2.45885804e-01 9.08196867e-01
9.79919791e-01 -1.25904921e-02 -2.84359735e-02 -4.21833575e-01
-5.14446139e-01 -1.48213819e-01 -1.54041958e+00 1.96258020e+00
-1.26091301e-01 6.97102904e-01 2.69397527e-01 -8.25869083e-01
7.76594281e-01 4.19317573e-01 9.32079792e-01 -6.38542831e-01
1.71617299e-01 1.98748320e-01 -5.10287642e-01 -4.08259839e-01
5.67232430e-01 1.06975578e-01 8.84690508e-02 1.52882457e-01
2.61351943e-01 -4.97015715e-01 1.73954770e-01 -8.60935152e-02
9.43457127e-01 4.52706426e-01 4.93438900e-01 2.25627393e-01
6.63409650e-01 8.57491940e-02 5.78067601e-01 6.23402894e-01
1.45321861e-02 1.14831376e+00 -2.38930620e-03 -9.98521090e-01
-1.14765668e+00 -1.21289730e+00 -3.17614414e-02 2.39659578e-01
6.67314053e-01 -1.80489510e-01 -3.58768553e-01 -3.77785593e-01
-5.98676987e-02 2.22403869e-01 -3.82435560e-01 4.69056070e-01
-8.45673621e-01 3.96308340e-02 -1.73336044e-01 3.66724432e-01
5.02829671e-01 -7.31516421e-01 -1.49144387e+00 2.71466792e-01
-2.71518469e-01 -1.56926763e+00 -1.81616902e-01 2.88877666e-01
-1.11727309e+00 -1.38641536e+00 -6.06765211e-01 -6.83376193e-01
4.50897664e-01 8.35889280e-01 1.21415150e+00 -1.97536781e-01
-2.27637932e-01 5.92748165e-01 -4.78900112e-02 -5.03569096e-02
-5.31837046e-02 -5.00162601e-01 -5.38722277e-02 4.50929031e-02
-1.78107806e-02 -7.26141334e-01 -5.81317008e-01 3.87306035e-01
-6.06119454e-01 4.51764107e-01 2.66945481e-01 3.71041566e-01
9.97788191e-01 -3.76816802e-02 -3.33250552e-01 -4.13552791e-01
-3.47996831e-01 -3.14366281e-01 -1.10555363e+00 -2.10595597e-02
2.26582438e-02 -3.69875908e-01 3.66165578e-01 -6.82924986e-02
-9.60103095e-01 5.30749798e-01 2.75775492e-01 -1.10253167e+00
-4.00654525e-01 3.11050028e-01 2.02687129e-01 -1.77698150e-01
6.43732071e-01 4.16267246e-01 -1.94418907e-01 -3.08250070e-01
4.30896223e-01 -6.65706024e-02 9.35501277e-01 -2.95523375e-01
8.02656531e-01 1.15225852e+00 1.84730440e-01 -1.02393496e+00
-6.29918933e-01 -8.90691936e-01 -1.37200141e+00 -6.93264365e-01
1.03297055e+00 -1.30143869e+00 -4.58394587e-01 4.78281289e-01
-1.64021552e+00 3.75716761e-02 -2.84747571e-01 7.97900081e-01
-8.72196555e-01 3.42263907e-01 -4.30726290e-01 -6.84736669e-01
-2.62720257e-01 -1.11552644e+00 1.37292290e+00 2.32301414e-01
-3.31349601e-03 -7.81988621e-01 2.77028173e-01 6.69188201e-02
-2.51236767e-01 6.47359014e-01 2.02633426e-01 -2.18767929e-03
-1.42591608e+00 -1.88454226e-01 -7.28718564e-02 5.35633117e-02
-1.55388027e-01 1.07448965e-01 -9.05801654e-01 -6.01704195e-02
1.35212541e-01 1.36631519e-01 6.00219846e-01 7.81481624e-01
6.29485369e-01 8.93528759e-02 -5.16604364e-01 1.15275264e+00
1.77805424e+00 5.00602007e-01 4.06930149e-01 5.49784124e-01
8.19097042e-01 4.38329518e-01 4.96131718e-01 4.45617944e-01
3.80328000e-01 8.30654681e-01 8.29604745e-01 -2.31360778e-01
-2.72365957e-01 -1.18146434e-01 2.04405695e-01 3.73748928e-01
-1.78550065e-01 -9.34734195e-02 -8.21508706e-01 5.46867669e-01
-1.84686005e+00 -9.82943952e-01 -2.47349098e-01 2.22561598e+00
1.13450371e-01 1.68390200e-01 2.80364864e-02 -7.09497854e-02
5.30272126e-01 2.51691490e-01 -6.06479168e-01 -3.62481661e-02
-1.74488232e-01 -8.16309750e-02 6.62974417e-01 5.09730160e-01
-1.18854368e+00 8.49717617e-01 6.78725481e+00 2.43507653e-01
-1.24307895e+00 -6.19362779e-02 1.64038107e-01 -2.35260919e-01
2.78790686e-02 -6.32703230e-02 -5.18811524e-01 -6.96278512e-02
2.84699857e-01 -5.93201667e-02 1.21663079e-01 9.36323285e-01
1.17639765e-01 -5.79019010e-01 -1.29031563e+00 1.24285340e+00
3.17318827e-01 -1.63681281e+00 -1.48697393e-02 -4.20161448e-02
1.00687158e+00 6.13405347e-01 -3.58069837e-01 -3.48901749e-01
2.15290517e-01 -6.76785350e-01 1.33806837e+00 7.51013279e-01
6.88302457e-01 -7.60390103e-01 5.23881316e-01 6.33028984e-01
-1.60873759e+00 3.51830982e-02 -3.29503685e-01 -4.15829420e-02
5.60170770e-01 7.50905812e-01 -8.58582377e-01 1.13994288e+00
9.83453870e-01 9.18852329e-01 -3.32201958e-01 1.41119695e+00
3.27220298e-02 -7.81162828e-02 -5.97569108e-01 6.69320166e-01
1.73887223e-01 -2.83872545e-01 8.92706275e-01 9.12982225e-01
5.58876693e-01 7.56118074e-02 5.61946869e-01 8.28564346e-01
2.57774293e-01 -3.18040431e-01 -9.64108944e-01 7.29983926e-01
4.40259993e-01 1.20925736e+00 -8.76045763e-01 -4.98907685e-01
-5.85816979e-01 8.43765438e-01 3.47375244e-01 2.80914813e-01
-5.97036660e-01 1.75979793e-01 3.00668001e-01 3.78588349e-01
6.99614406e-01 -5.66635013e-01 -2.72733063e-01 -1.12140000e+00
2.86413819e-01 -3.65994424e-01 3.26710522e-01 -1.20445728e+00
-6.11889005e-01 7.50388205e-01 3.60413492e-01 -1.70996451e+00
-4.50497627e-01 -4.42516774e-01 -4.23517376e-01 6.88774288e-01
-1.42644477e+00 -1.09274769e+00 -7.54500806e-01 7.16480792e-01
9.13017392e-01 -1.17238298e-01 3.48785162e-01 -1.99128198e-03
4.59439401e-03 -2.70620972e-01 -1.51023507e-01 -4.58131395e-02
3.30946505e-01 -9.90327835e-01 4.38272417e-01 1.19931006e+00
3.95288646e-01 2.58925170e-01 4.71261442e-01 -7.88511276e-01
-1.46400309e+00 -1.01159120e+00 3.46557766e-01 -4.34167951e-01
4.87491209e-03 -1.99108168e-01 -7.61896789e-01 8.46078396e-01
4.50677089e-02 3.43551397e-01 -1.61885545e-01 -5.04971683e-01
-2.55649716e-01 -1.19919866e-01 -1.10673463e+00 3.85166019e-01
1.12240398e+00 -3.70248258e-01 -6.31209850e-01 6.22782372e-02
7.00808167e-01 -1.11727750e+00 -6.48175836e-01 4.62795705e-01
6.77428722e-01 -1.62318528e+00 1.36440980e+00 2.68108815e-01
3.52898389e-01 -6.84503198e-01 -2.42425054e-01 -1.00408912e+00
-2.24007607e-01 -4.78010237e-01 -3.19234312e-01 2.12337628e-01
-2.88484722e-01 -3.28597516e-01 1.02875018e+00 2.45922916e-02
-2.77230680e-01 -6.16291344e-01 -1.23543489e+00 -6.56880677e-01
-4.86059278e-01 -7.96411812e-01 2.63229132e-01 4.92051721e-01
-4.67643291e-01 -3.54635902e-02 -2.46630490e-01 6.86491668e-01
1.09494233e+00 2.34702617e-01 1.19148004e+00 -1.39750004e+00
-6.91437870e-02 -1.91057026e-01 -8.34039211e-01 -1.33924663e+00
-2.38307908e-01 -6.28797531e-01 8.74919072e-02 -1.59429789e+00
-1.43381596e-01 -3.84235792e-02 2.45541126e-01 3.57239023e-02
3.75529528e-01 3.99262339e-01 3.24666381e-01 2.66778231e-01
-7.39867628e-01 3.26590598e-01 1.23571026e+00 1.76942065e-01
-1.07993215e-01 -2.06744924e-01 6.69095442e-02 1.25975013e+00
1.38797760e-01 -4.23141718e-01 -2.83664852e-01 -6.01532459e-01
7.90922437e-03 6.63614094e-01 4.80875283e-01 -1.43642139e+00
5.94674647e-01 -1.04192145e-01 7.99829304e-01 -1.58893967e+00
8.85845006e-01 -1.12668550e+00 5.80980599e-01 3.34482700e-01
3.47974420e-01 2.79076904e-01 1.40395984e-01 7.14006782e-01
-2.70833820e-01 3.27974670e-02 6.61817253e-01 -5.09945393e-01
-1.05212939e+00 5.75749576e-01 -1.53795719e-01 -3.55419070e-01
1.23095393e+00 -7.55971253e-01 8.86196643e-02 -3.30238968e-01
-7.12082744e-01 1.59959614e-01 8.03637922e-01 3.75825673e-01
1.22236001e+00 -1.20847070e+00 -6.99309170e-01 6.38545454e-01
4.06956002e-02 8.78672361e-01 5.04662633e-01 5.92761695e-01
-1.28666139e+00 6.05914831e-01 -2.14003921e-01 -1.46112227e+00
-1.41420603e+00 3.35990995e-01 7.11596012e-01 -2.02384423e-02
-1.34585500e+00 6.81527197e-01 7.11300254e-01 -3.74937803e-01
2.27053568e-01 -6.29643679e-01 -9.62958336e-02 -2.23381311e-01
5.73410869e-01 2.17722759e-01 2.17207596e-01 -8.30033600e-01
-3.71057004e-01 1.09347582e+00 1.30309701e-01 -4.16862220e-01
1.46262109e+00 -2.96699941e-01 8.84819180e-02 3.78135353e-01
1.08044660e+00 -1.42038781e-02 -1.74884140e+00 -5.41513920e-01
-4.28842187e-01 -7.43926883e-01 1.98797777e-01 -2.75013596e-01
-1.18513274e+00 5.03821671e-01 5.16548634e-01 -1.98758692e-01
8.49036813e-01 6.86432943e-02 4.73644316e-01 4.10170645e-01
9.26401615e-01 -5.70327044e-01 6.32868931e-02 6.28574193e-01
9.89867985e-01 -9.37542617e-01 4.48677301e-01 -5.61450481e-01
-8.27266127e-02 1.65902579e+00 6.17980480e-01 -3.85325283e-01
6.21824622e-01 2.71869123e-01 -4.31158133e-02 -6.51290119e-01
-6.23240948e-01 1.08478092e-01 4.29545641e-01 7.56932676e-01
-1.78475156e-01 -2.31890544e-01 3.86424661e-01 -1.11792929e-01
-9.42154303e-02 -5.55232018e-02 6.70009136e-01 1.07548106e+00
-5.29835105e-01 -4.81851131e-01 -9.00103152e-01 -4.98787276e-02
7.35170171e-02 3.98655444e-01 1.37482602e-02 1.03013480e+00
3.04882646e-01 5.89218795e-01 5.76524615e-01 -3.17433417e-01
5.03952265e-01 -2.53572464e-01 8.56065273e-01 -7.00543642e-01
-1.51310265e-01 6.00208759e-01 5.05633242e-02 -1.21971035e+00
-5.84203243e-01 -8.88573229e-01 -1.19338441e+00 -2.18721610e-02
-5.93967736e-02 -3.31803679e-01 6.95460081e-01 8.49280834e-01
1.47684053e-01 4.63229686e-01 6.17738008e-01 -1.62034249e+00
1.44170761e-01 -4.85949516e-01 -5.76846957e-01 -3.83341610e-02
6.46230221e-01 -8.27243984e-01 -7.34421089e-02 2.38208249e-01] | [8.117650032043457, -2.3018736839294434] |
715c8be5-3e00-4dc2-aaab-5e463f6cc8d6 | experimental-demonstration-of-bandwidth | 2301.04573 | null | https://arxiv.org/abs/2301.04573v2 | https://arxiv.org/pdf/2301.04573v2.pdf | Experimental demonstration of bandwidth enhancement in photonic time delay reservoir computing | Time delay reservoir computing (TDRC) using semiconductor lasers (SLs) has proven to be a promising photonic analog approach for information processing. One appealing property is that SLs subject to delayed optical feedback and external optical injection, allow tuning the response bandwidth by changing the level of optical injection. Here we use strong optical injection, thereby expanding the SL's modulation response up to tens of GHz. Performing a nonlinear time series prediction task, we demonstrate experimentally that for appropriate operating conditions, our TDRC system can operate with sampling times as small as 11.72 ps, without sacrificing computational performance. | ['Ingo Fischer', 'Apostolos Argyris', 'Irene Estebanez'] | 2023-01-11 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [ 2.76122749e-01 -3.15254718e-01 -4.34178188e-02 1.44549981e-01
-2.20416799e-01 -7.46003389e-01 5.15609443e-01 -3.02173018e-01
-7.85456896e-01 8.73330891e-01 -3.38066876e-01 -3.59768450e-01
2.48917658e-03 -5.05920887e-01 -1.72055468e-01 -1.04464924e+00
-2.26902202e-01 1.40960336e-01 4.11906213e-01 -1.11363560e-01
6.87286258e-01 7.26366639e-01 -1.32554626e+00 -3.99829447e-03
6.98263586e-01 1.29944563e+00 5.31173907e-02 4.90993828e-01
2.83782512e-01 3.11278403e-01 -7.11200535e-01 3.21439832e-01
3.41528088e-01 -4.02898163e-01 -1.74751699e-01 -7.57681787e-01
-5.14656790e-02 -4.17742044e-01 -1.09088469e+00 8.30014288e-01
4.57707226e-01 7.94402510e-02 4.05151367e-01 -6.28712416e-01
-6.42531157e-01 6.37027860e-01 -2.88052499e-01 8.15815508e-01
1.48642510e-01 8.82674873e-01 5.89985430e-01 -4.23656255e-01
7.95463622e-01 5.61968565e-01 1.98969811e-01 6.68846309e-01
-1.80333364e+00 -9.69372690e-01 -6.88968837e-01 8.94232690e-02
-1.11659110e+00 -9.69135463e-01 8.56117427e-01 -3.30433041e-01
1.23910952e+00 -8.29938352e-02 8.50646377e-01 5.18755019e-01
9.00431395e-01 -2.79862314e-01 1.77880883e+00 -4.21740443e-01
3.66335958e-01 2.79020518e-01 8.36929083e-02 4.61589336e-01
2.06957199e-02 8.03780198e-01 -8.98431063e-01 -2.07686782e-01
6.66074753e-01 -3.40509772e-01 -4.19820011e-01 4.19017643e-01
-1.23717380e+00 4.55477685e-01 6.02680266e-01 3.16240937e-01
-2.63640910e-01 6.88769400e-01 9.48922783e-02 7.95004129e-01
2.76515752e-01 1.20344269e+00 2.91317075e-01 -9.44074094e-02
-9.43665922e-01 -9.67112333e-02 4.23013717e-01 5.15732229e-01
6.13020599e-01 3.62764925e-01 -4.27315384e-01 1.32088155e-01
8.45438316e-02 7.50325799e-01 5.47532499e-01 -1.03894925e+00
-1.55987799e-01 -2.43075788e-01 3.29300523e-01 -2.53080785e-01
-4.43260312e-01 -3.92501354e-01 -4.86673713e-01 1.93496257e-01
3.11329752e-01 -2.66054332e-01 -9.21562493e-01 1.28572142e+00
-2.99308598e-01 -5.06369062e-02 2.57692188e-01 8.00650358e-01
5.08996606e-01 7.94364214e-01 -3.11246693e-01 -6.98260844e-01
1.12827063e+00 -2.96695471e-01 -8.77607703e-01 -1.22290276e-01
-8.60796198e-02 -6.08078659e-01 8.34751725e-01 2.58929223e-01
-9.78767991e-01 6.44156113e-02 -1.18276560e+00 4.06920500e-02
1.43737439e-03 -1.36879638e-01 7.21453428e-01 6.08974993e-01
-1.12633371e+00 8.95769596e-01 -8.45446706e-01 1.87359989e-01
5.67856692e-02 7.11004436e-01 5.34729436e-02 4.30317558e-02
-9.68712926e-01 7.30564713e-01 2.98078984e-01 2.34827101e-01
-4.49140370e-01 -9.62511539e-01 4.08988409e-02 -1.27594620e-01
-1.45102918e-01 -8.06868315e-01 1.07929754e+00 -1.24533828e-02
-2.43759847e+00 6.28817320e-01 -4.84359682e-01 -1.02264595e+00
1.27865598e-01 4.85956520e-01 -5.48510253e-01 6.52088583e-01
-2.51241446e-01 2.69657850e-01 1.15032947e+00 -4.94353622e-01
2.30278403e-01 -4.50024158e-01 -4.51481789e-01 -5.91338873e-01
-1.60753444e-01 3.37172896e-01 3.97063076e-01 1.25970185e-01
5.65577209e-01 -1.15008318e+00 -1.67176843e-01 -2.34320849e-01
-1.70461938e-01 1.45478606e-01 9.48448896e-01 1.18425451e-01
8.33270729e-01 -2.14262152e+00 -2.73744643e-01 1.77808806e-01
3.31320077e-01 1.01287216e-01 -1.14448652e-01 8.06293428e-01
2.90127963e-01 1.11587338e-01 -2.02756673e-02 1.57818884e-01
-3.70026559e-01 -6.05509877e-01 -6.45095587e-01 5.74904323e-01
2.81731606e-01 9.32173133e-01 -6.64877892e-01 2.47425392e-01
-1.21112414e-01 2.28126630e-01 -6.62769675e-01 1.88679874e-01
-3.93970966e-01 1.00218070e+00 -2.07525477e-01 6.68863952e-01
4.90765840e-01 -3.62509519e-01 -3.48437764e-02 -2.72151232e-01
-1.04042578e+00 8.03140938e-01 -3.74494433e-01 1.16910017e+00
-1.58380225e-01 1.33341599e+00 -7.33435526e-02 -3.59823942e-01
1.01525760e+00 2.24994138e-01 4.88648355e-01 -1.27814293e+00
2.68447176e-02 4.45105076e-01 5.37135720e-01 -4.04866427e-01
5.60620368e-01 -8.17828715e-01 1.94407739e-02 8.34576666e-01
-1.05210179e-02 -5.30416846e-01 -5.67216519e-03 5.94670624e-02
1.26683187e+00 -4.99525249e-01 -1.57147214e-01 -3.43570799e-01
-1.26021668e-01 -3.78592573e-02 4.60202321e-02 8.58563662e-01
-3.39073718e-01 -9.36219022e-02 4.79814649e-01 -1.78083077e-01
-1.20228720e+00 -1.07717240e+00 -4.24090981e-01 4.43772763e-01
3.10004950e-01 -3.04163724e-01 -7.42730871e-02 6.05076969e-01
2.31147513e-01 2.30155721e-01 7.77309202e-03 -3.41340870e-01
-5.97514629e-01 -7.30594218e-01 5.92811644e-01 -3.97286937e-02
6.17754340e-01 -8.13321531e-01 -5.25928676e-01 6.43394887e-01
4.00459111e-01 -1.41038024e+00 -1.26961797e-01 1.24685191e-01
-9.77090120e-01 -2.83058673e-01 -2.98671663e-01 -1.64401665e-01
2.42958054e-01 1.68226197e-01 5.90566814e-01 -3.66295934e-01
-7.56368339e-01 2.37904429e-01 4.08954144e-01 -6.79009557e-02
-3.71513933e-01 -1.27570614e-01 4.61084396e-01 -3.97070318e-01
1.10805370e-01 -1.23631740e+00 -1.05466843e+00 -2.23405287e-01
-3.34080696e-01 -1.20944612e-01 7.20260143e-01 4.99615014e-01
2.80096322e-01 -3.33019674e-01 7.88874567e-01 -5.88326097e-01
7.67255902e-01 -9.43115056e-02 -1.29604697e+00 -3.49999517e-01
-6.19845450e-01 4.06486452e-01 7.33911157e-01 -4.71909434e-01
-6.31733537e-01 -5.07844448e-01 1.76950157e-01 3.60499732e-02
2.62959272e-01 7.09375918e-01 8.54320228e-01 -8.90679657e-01
6.94997966e-01 5.66560090e-01 2.33465403e-01 1.21732332e-01
1.58818617e-01 6.31269574e-01 3.02563071e-01 -3.41852635e-01
9.10826445e-01 5.78134060e-01 3.39027524e-01 -1.21998036e+00
-2.33268991e-01 -1.72240660e-01 -2.98168659e-01 -2.29949683e-01
3.64647567e-01 -8.47350180e-01 -1.65993333e+00 5.20165145e-01
-1.01322591e+00 -4.86967742e-01 2.98305862e-02 6.17202818e-01
-4.56183314e-01 -2.85860658e-01 -1.00848329e+00 -1.22547984e+00
-4.91354138e-01 -6.83231711e-01 5.23230851e-01 4.77613896e-01
2.52464980e-01 -2.54519790e-01 -2.18331262e-01 -7.11259781e-04
1.20547867e+00 2.51315147e-01 1.00499988e+00 2.71937251e-01
-1.51547313e+00 -2.71792531e-01 -2.22800568e-01 -1.79058611e-01
-1.10842846e-01 -2.73168981e-02 -1.11943173e+00 -4.36981440e-01
2.17502534e-01 -2.52121747e-01 1.15518713e+00 3.29781771e-01
9.98093128e-01 -5.91502227e-02 -4.73822027e-01 4.35654432e-01
1.60946560e+00 4.64213341e-01 6.96856141e-01 -1.68271780e-01
-9.93508101e-02 9.09975395e-02 7.81993121e-02 4.37858820e-01
-3.26019615e-01 7.07189023e-01 -7.81429633e-02 6.87988758e-01
1.00223385e-01 5.93023524e-02 5.73160112e-01 1.10595858e+00
-2.19108105e-01 -7.37862736e-02 -7.08354115e-01 5.28628863e-02
-1.09088588e+00 -9.87752497e-01 1.42280420e-03 2.58553767e+00
9.03036594e-01 2.30866715e-01 1.21913642e-01 -3.80002916e-01
4.33045268e-01 3.08853567e-01 -8.47865105e-01 -3.71958822e-01
1.89033821e-01 8.13230515e-01 8.83320212e-01 3.75300258e-01
-4.41357553e-01 9.63916600e-01 7.60865307e+00 1.38970941e-01
-1.88000894e+00 2.24113554e-01 -1.83138952e-01 -4.38622713e-01
-4.37918037e-01 2.60094970e-01 -7.08273351e-01 9.32589233e-01
1.45108593e+00 -6.20903075e-01 9.84652698e-01 -4.11161687e-03
3.79119396e-01 -4.68869120e-01 -8.35413516e-01 8.67274880e-01
-7.33323276e-01 -1.77434909e+00 -2.91379839e-01 3.15570921e-01
3.53279531e-01 2.02735618e-01 5.67084432e-01 8.16181302e-03
-4.58464175e-01 -9.31523621e-01 2.91379869e-01 7.35422671e-01
1.42257249e+00 -4.98811096e-01 5.94879091e-02 4.16627616e-01
-8.09597313e-01 -3.21935803e-01 -5.30388832e-01 -4.03500080e-01
1.60442501e-01 8.25202346e-01 -7.02772915e-01 -2.15406373e-01
3.15138429e-01 2.77243316e-01 -1.53817423e-02 1.01844645e+00
1.87230095e-01 7.65316069e-01 -5.34879088e-01 -5.39734244e-01
2.18349434e-02 -6.12118423e-01 9.04984176e-01 6.86682582e-01
5.22620857e-01 5.96867442e-01 -3.04772675e-01 1.59649920e+00
-2.04828411e-01 -6.75182402e-01 -7.32393742e-01 -9.40762103e-01
8.82537127e-01 9.00783837e-01 -5.28891146e-01 1.71807915e-01
2.23291084e-01 9.41973865e-01 1.25264212e-01 4.68284458e-01
-2.79479623e-01 -9.34597731e-01 8.79057705e-01 3.17279845e-01
8.31380114e-02 -1.07979453e+00 4.97901030e-02 -1.08454299e+00
-9.63729396e-02 7.08316341e-02 -3.71474832e-01 -7.19002545e-01
-7.89680421e-01 4.71351802e-01 -5.06829143e-01 -1.06251895e+00
-3.72922458e-02 -6.39385939e-01 -5.91999471e-01 1.09988427e+00
-1.86315501e+00 -2.19750434e-01 5.45333363e-02 1.70578629e-01
-3.42531174e-01 7.40680331e-03 8.47386897e-01 -4.59657945e-02
-3.10619771e-01 3.48972410e-01 4.51116920e-01 -3.31304073e-01
6.18738592e-01 -8.16552401e-01 3.03589910e-01 5.04018843e-01
-2.24008292e-01 1.04568136e+00 1.00984204e+00 -4.04983401e-01
-2.19138145e+00 -6.22779191e-01 5.24287522e-01 1.09046988e-01
8.73272717e-01 -6.06780887e-01 -6.76597953e-01 2.72282779e-01
3.49541396e-01 2.55009472e-01 7.14453638e-01 -3.90106380e-01
-4.66684848e-01 -3.16455185e-01 -1.55988276e+00 5.58495879e-01
1.00738561e+00 -1.16917992e+00 2.16638558e-02 6.36734128e-01
9.09115255e-01 -6.00533426e-01 -8.78164649e-01 2.50486583e-01
8.84766519e-01 -7.83453405e-01 6.41891778e-01 -4.84580666e-01
3.39099407e-01 -2.96839893e-01 2.40795806e-01 -1.24719846e+00
-2.69896746e-01 -1.45266223e+00 -7.70805702e-02 4.98856932e-01
3.53965521e-01 -1.65162838e+00 6.34577453e-01 4.73886102e-01
-1.51472136e-01 -1.94576427e-01 -1.18651593e+00 -1.02300036e+00
-2.51966178e-01 5.03853373e-02 7.65961111e-02 3.09566289e-01
6.07079625e-01 1.97225496e-01 -1.35029331e-01 2.93826371e-01
6.07666433e-01 3.97648335e-01 1.65886637e-02 -1.05635691e+00
-3.40713173e-01 -4.64184731e-01 -5.05138040e-01 -1.20358407e+00
1.47178233e-01 -1.03761780e+00 1.69825733e-01 -9.56165075e-01
-1.85251251e-01 -7.00118780e-01 -3.90177608e-01 -1.64982885e-01
5.11349320e-01 7.32111633e-01 2.55806595e-01 3.83993477e-01
7.64185563e-02 6.88470185e-01 1.15115690e+00 1.32280618e-01
-3.57082963e-01 1.45760119e-01 -2.60127306e-01 -2.35483259e-01
8.19150925e-01 -5.26660979e-01 -2.21477263e-02 -1.03140466e-01
3.31770957e-01 3.09888601e-01 3.55044246e-01 -1.30446506e+00
7.84737289e-01 -1.84947491e-01 2.25613043e-02 -9.17410478e-02
8.56992066e-01 -1.01285480e-01 1.31191179e-01 8.82426381e-01
-4.00911808e-01 -3.10472339e-01 3.58791590e-01 6.40746295e-01
-3.39170881e-02 -2.68314779e-01 1.00427639e+00 -1.28865406e-01
-3.07840317e-01 2.30170190e-01 -8.84600341e-01 -2.93974847e-01
7.43277669e-01 -2.37455056e-03 -9.94599521e-01 -1.96566075e-01
-4.61309820e-01 -3.09868217e-01 3.98759723e-01 -4.04518694e-01
5.77954650e-01 -1.14302683e+00 -8.82835910e-02 4.46140885e-01
-1.62496462e-01 -8.06692779e-01 2.32571512e-01 9.53823984e-01
-4.24157649e-01 8.88025522e-01 -2.15508983e-01 -6.83934629e-01
-7.63290048e-01 3.13718110e-01 4.58785534e-01 2.55333036e-01
-7.22041130e-01 7.45834291e-01 -8.82448912e-01 3.38383585e-01
-5.26182413e-01 -3.53230566e-01 2.25792825e-01 -2.59600908e-01
5.22716820e-01 1.72632206e-02 9.05083269e-02 2.47893408e-01
-2.84925550e-01 5.03860474e-01 -9.42329466e-02 -4.70979780e-01
1.22448230e+00 1.58536136e-01 -6.70098305e-01 5.60624957e-01
9.93222952e-01 2.57927179e-01 -9.91397679e-01 -4.58821267e-01
1.27150817e-02 -6.36473715e-01 3.52493256e-01 -5.51145017e-01
-5.71351469e-01 9.47168231e-01 6.29484177e-01 6.01153731e-01
8.22205603e-01 1.05166063e-01 8.67346406e-01 6.87734663e-01
8.56403291e-01 -8.12415659e-01 2.49910504e-01 6.80896103e-01
5.32912970e-01 -8.20342898e-01 -1.26164332e-01 -3.73187810e-01
2.19531506e-01 1.43625033e+00 1.78796351e-01 -3.87634099e-01
5.84500670e-01 4.03074741e-01 -4.30454463e-01 -2.88059324e-01
-1.12138760e+00 -5.65152429e-02 9.55111533e-03 3.70075434e-01
3.81138265e-01 2.63959140e-01 -3.81738067e-01 8.07202309e-02
-2.47572437e-01 9.81525779e-02 1.15520942e+00 7.11262226e-01
-5.99457502e-01 -8.07872832e-01 2.36645669e-01 6.78097963e-01
-1.99421912e-01 -2.07525074e-01 -8.24919790e-02 1.43368840e-01
-6.63102567e-01 6.77785456e-01 4.27541584e-01 -3.70741129e-01
8.01155791e-02 2.52319127e-01 9.67104256e-01 -6.59626842e-01
-1.73546463e-01 -3.06547374e-01 -2.71729022e-01 -4.39379364e-01
-2.19047040e-01 -4.21047866e-01 -1.26926744e+00 -4.67587709e-01
-4.06490684e-01 9.44575667e-02 7.95619130e-01 9.42219138e-01
8.66849542e-01 1.89519867e-01 9.11181808e-01 -6.48812473e-01
-6.47911966e-01 -7.30454624e-01 -8.23076963e-01 -7.87141263e-01
8.35586905e-01 -3.54342580e-01 -5.98441482e-01 -6.77311778e-01] | [5.634557723999023, 4.864107608795166] |
640f7323-d288-4429-bf5a-014806c6bb04 | movie-recommendation-system-using-sentiment | 1811.10804 | null | https://arxiv.org/abs/1811.10804v1 | https://arxiv.org/pdf/1811.10804v1.pdf | Movie Recommendation System using Sentiment Analysis from Microblogging Data | Recommendation systems are important intelligent systems that play a vital role in providing selective information to users. Traditional approaches in recommendation systems include collaborative filtering and content-based filtering. However, these approaches have certain limitations like the necessity of prior user history and habits for performing the task of recommendation. In order to reduce the effect of such dependencies, this paper proposes a hybrid recommendation system which combines the collaborative filtering, content-based filtering with sentiment analysis of movie tweets. The movie tweets have been collected from microblogging websites to understand the current trends and user response of the movie. Experiments conducted on public database produce promising results. | ['Sudhanshu Kumar', 'Kanjar De', 'Shirsendu Sukanta Halder', 'Partha Pratim Roy'] | 2018-11-27 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [-3.50608259e-01 -5.84600747e-01 -4.20885593e-01 -7.27779329e-01
4.65630693e-03 -4.71882939e-01 5.80945134e-01 4.35069770e-01
-6.14109337e-01 3.74627888e-01 7.30758131e-01 -4.23612654e-01
-3.85669857e-01 -1.01630926e+00 5.05865067e-02 -2.98887044e-01
3.69402081e-01 -1.35036772e-02 7.14065969e-01 -8.09652507e-01
1.05883563e+00 2.49913096e-01 -1.61133528e+00 7.51806617e-01
6.43844426e-01 8.46409976e-01 4.20105040e-01 5.48622668e-01
-4.95151013e-01 7.77095139e-01 -2.61266679e-01 -1.92653432e-01
7.81456903e-02 -3.42832148e-01 -4.82057095e-01 1.58986542e-02
-1.20797664e-01 -4.52015102e-01 -2.54344165e-01 9.13681149e-01
3.11707020e-01 8.01705837e-01 6.38663948e-01 -6.36730313e-01
-7.26611018e-01 7.07690597e-01 -1.15817510e-01 6.64155066e-01
5.72373927e-01 -7.52267241e-01 7.75493562e-01 -7.75064230e-01
2.89975286e-01 8.64506960e-01 4.64593410e-01 3.04341555e-01
-3.61812264e-01 -4.91545111e-01 4.70236301e-01 2.51149863e-01
-9.26963091e-01 -1.53095767e-01 5.33545375e-01 -5.96945882e-01
8.90130520e-01 4.28865731e-01 3.03660244e-01 4.35045540e-01
3.23335797e-01 3.19782853e-01 8.99706304e-01 -4.15807009e-01
2.38175899e-01 7.81347215e-01 7.83004284e-01 2.48374879e-01
-1.37639612e-01 -3.93758446e-01 -3.59442741e-01 -4.07932699e-01
4.19000655e-01 7.51886606e-01 2.42528860e-02 5.43597162e-01
-3.81504029e-01 9.90682960e-01 1.13929249e-01 7.15360582e-01
-4.40221846e-01 -8.08260798e-01 4.63334113e-01 6.11621380e-01
7.64404297e-01 3.67140472e-01 -4.62682039e-01 2.99464226e-01
-6.38402343e-01 7.93493763e-02 7.30302274e-01 5.28250098e-01
3.34558457e-01 -3.71928275e-01 -3.37404571e-02 7.58690715e-01
7.83190489e-01 1.47772133e-01 8.97167921e-01 -3.36955488e-01
1.15339980e-01 7.31200218e-01 4.79278892e-01 -1.56362689e+00
-2.44326398e-01 -3.52573872e-01 -2.82805413e-01 -2.84249604e-01
2.17653990e-01 -5.02864122e-01 -6.43407762e-01 8.29169273e-01
4.37510014e-01 -1.68770179e-01 -9.03475136e-02 9.89742577e-01
1.11950099e+00 8.80235314e-01 8.63916352e-02 -5.27426481e-01
1.14526713e+00 -7.35342503e-01 -7.90413320e-01 4.74734783e-01
2.04279929e-01 -1.24949241e+00 8.47119987e-01 8.29833508e-01
-6.74004138e-01 -8.39753091e-01 -6.70365870e-01 4.40886438e-01
-6.03521109e-01 9.35471803e-02 7.41059959e-01 8.47389519e-01
-6.41162574e-01 6.93764746e-01 -1.79430589e-01 -6.56153202e-01
-2.40341678e-01 4.04391229e-01 1.68597862e-01 1.69610098e-01
-1.27040780e+00 4.67099696e-01 -3.92066725e-02 -1.78641841e-01
-2.85505533e-01 -1.02040641e-01 6.90695713e-04 -1.41210482e-01
2.03571290e-01 -3.15696443e-03 9.88797307e-01 -1.07483852e+00
-1.49435532e+00 -2.03608181e-02 -9.27649364e-02 -1.43342704e-01
3.05595323e-02 -2.48717621e-01 -8.63463044e-01 -2.77006954e-01
-3.13027233e-01 -5.81187248e-01 5.60303330e-01 -6.36153579e-01
-1.30785167e+00 -3.77012581e-01 2.47349843e-01 2.91636586e-01
-9.35810745e-01 6.98793590e-01 -4.34951663e-01 -5.26060045e-01
2.07131848e-01 -6.51717544e-01 -4.90573347e-01 -1.00019121e+00
2.15933308e-01 -5.89693129e-01 6.05478168e-01 -4.06544834e-01
1.63763571e+00 -1.87621629e+00 -4.37420726e-01 3.97464275e-01
-1.17752381e-01 2.32693955e-01 3.09201032e-01 8.86142373e-01
5.73423088e-01 1.41634777e-01 9.09623325e-01 1.96448416e-01
-3.34622890e-01 -1.33290604e-01 -4.18373078e-01 2.47931570e-01
-9.10930336e-01 -7.30451792e-02 -7.31434703e-01 -3.90212178e-01
1.26608938e-01 3.19633842e-01 -8.29652727e-01 4.19419557e-01
-2.48331219e-01 5.65912426e-01 -1.11059964e+00 1.65803030e-01
2.24844396e-01 -7.47640803e-02 3.05004656e-01 -2.85807997e-01
-6.32237196e-01 5.93416035e-01 -1.15558302e+00 1.06414020e+00
-4.18941081e-01 5.29273935e-02 -7.62572363e-02 -8.82264555e-01
9.52820420e-01 4.39728290e-01 5.52531421e-01 -4.33207214e-01
6.36205435e-01 -1.30361095e-01 -4.86126468e-02 -9.18955505e-01
9.04732406e-01 1.77156776e-01 4.75842096e-02 6.65517151e-01
-2.50236690e-01 4.95611608e-01 3.63422304e-01 5.05686045e-01
7.10705280e-01 -3.75354029e-02 5.93018569e-02 -2.15633303e-01
7.04792082e-01 2.49542758e-01 4.69087273e-01 7.80291378e-01
1.90795451e-01 4.12041619e-02 -2.70272195e-01 -4.14016366e-01
-6.02630973e-01 -1.51695311e-01 -4.06908728e-02 1.61933672e+00
2.35097900e-01 -7.39939570e-01 -3.20751667e-01 -6.81889057e-01
-1.55196786e-01 4.44056183e-01 -4.96619940e-01 1.78360119e-01
-1.47517949e-01 -7.77006567e-01 -3.62264812e-01 2.09919810e-01
1.42521232e-01 -7.34968483e-01 1.29924402e-01 5.47365129e-01
-4.05497998e-02 -4.43542063e-01 -5.28281033e-01 -3.18639517e-01
-9.34284747e-01 -8.80691469e-01 -1.36561573e-01 -6.17355406e-01
7.51374662e-01 8.09121609e-01 5.13301492e-01 5.29277384e-01
5.23080111e-01 1.76609188e-01 -1.17005897e+00 -3.33763003e-01
-2.63164770e-02 -8.06205198e-02 2.78456211e-01 2.15283036e-01
6.45669103e-01 -5.10371149e-01 -7.85342515e-01 8.39628756e-01
-6.92666411e-01 -3.64592791e-01 3.43408585e-02 2.80714363e-01
1.92955613e-01 7.45669365e-01 7.05882847e-01 -1.58573580e+00
9.83183801e-01 -1.09849167e+00 -3.43905747e-01 -3.68512869e-02
-1.05973268e+00 -6.52259767e-01 1.01997554e+00 -3.58309269e-01
-1.44363177e+00 -2.37619653e-01 -3.84464383e-01 6.36920333e-01
-2.93726057e-01 8.64134252e-01 1.78258315e-01 -1.77213456e-02
6.14808381e-01 -5.23159616e-02 -5.22223592e-01 -1.08170760e+00
3.63346636e-02 1.34665418e+00 -4.95622121e-02 3.07730045e-02
1.86604097e-01 3.26031566e-01 -6.55326962e-01 -5.30475140e-01
-1.10396552e+00 -1.15547466e+00 -2.89113104e-01 -6.56007767e-01
5.21209240e-01 -6.00085020e-01 -4.94132072e-01 2.23201942e-02
-5.51619351e-01 3.37929100e-01 4.67187822e-01 9.88049507e-01
4.22333032e-01 3.19727898e-01 -8.63874257e-01 -8.85553241e-01
-4.20351684e-01 -5.58186769e-01 -2.54542708e-01 5.77340782e-01
-2.52522707e-01 -8.94493759e-01 1.86661959e-01 6.68782175e-01
7.26013064e-01 -4.64574069e-01 5.47503948e-01 -1.09586394e+00
-2.53652576e-02 -7.35908687e-01 1.26797095e-01 2.85373926e-01
3.95362258e-01 2.85882711e-01 -5.15118301e-01 -1.85293674e-01
2.37788618e-01 5.83215617e-02 4.46258336e-01 2.27851719e-01
8.17172647e-01 -4.95513767e-01 -3.36419344e-01 -1.77043304e-02
1.51392794e+00 8.00407469e-01 3.70925397e-01 2.87766695e-01
4.68043953e-01 5.38177609e-01 1.05899692e+00 8.55935276e-01
3.79440039e-01 2.60940850e-01 1.35181144e-01 3.71542901e-01
3.44402432e-01 -2.51814127e-01 2.88559854e-01 1.27318907e+00
-3.57324690e-01 -5.58875918e-01 -4.71738756e-01 3.31958055e-01
-1.99958181e+00 -9.40466881e-01 -5.32906055e-01 2.17475367e+00
3.90613675e-01 3.19552511e-01 4.58597511e-01 3.26729178e-01
6.04443967e-01 -2.83695579e-01 -5.61114065e-02 -6.34320378e-01
2.99732119e-01 -4.44359779e-02 4.59699750e-01 4.05356318e-01
-9.56439614e-01 7.75104403e-01 6.04175711e+00 5.78996599e-01
-1.25296879e+00 3.53492171e-01 2.58971006e-01 1.31129641e-02
-3.69398832e-01 -3.44131440e-02 -1.15482223e+00 7.10755885e-01
1.03482711e+00 -1.11152269e-01 2.01133221e-01 7.72567332e-01
6.56529486e-01 -4.49975610e-01 -2.23055422e-01 3.23793530e-01
2.30598167e-01 -1.23229551e+00 -1.92010745e-01 -1.65022805e-01
9.23963726e-01 6.90399334e-02 -1.40894741e-01 1.37713745e-01
5.04687071e-01 -3.13927561e-01 3.01792920e-01 8.90683532e-01
-2.09106266e-01 -7.18085945e-01 1.00776446e+00 5.66129506e-01
-7.04933345e-01 -1.74176008e-01 -7.63898671e-01 -7.65599787e-01
1.76291123e-01 6.89068019e-01 -5.44325650e-01 2.70393372e-01
9.14199650e-01 7.94455230e-01 -1.17590882e-01 1.19190300e+00
5.98480962e-02 9.81182396e-01 -1.68691531e-01 -5.78528941e-01
2.30375789e-02 -5.31367838e-01 1.34693310e-01 1.14138174e+00
1.92296475e-01 6.54268444e-01 3.81642163e-01 -3.31067026e-01
2.60857612e-01 1.10269439e+00 -2.05177680e-01 -8.42197612e-02
4.36367601e-01 1.34770787e+00 -1.06138146e+00 -3.54543000e-01
-7.68219531e-01 5.46925426e-01 -1.70493498e-01 4.30618087e-03
-5.35315335e-01 -3.33864152e-01 1.50220454e-01 5.39458513e-01
3.68503988e-01 -1.33352265e-01 -8.06024745e-02 -9.24333155e-01
-6.64604843e-01 -8.17344964e-01 6.56568408e-01 -2.59733438e-01
-1.24526179e+00 6.41995192e-01 -2.97012448e-01 -1.13347673e+00
1.56607494e-01 -3.05276901e-01 -7.11286604e-01 7.03044057e-01
-7.70310760e-01 -6.55116558e-01 -1.04590252e-01 8.23926508e-01
7.47000575e-01 -4.64818120e-01 7.34164596e-01 8.40778828e-01
-5.51902354e-01 1.91071287e-01 4.64800507e-01 -2.98221797e-01
9.30093229e-01 -8.04778337e-01 -2.84221977e-01 8.24382365e-01
-1.54175267e-01 9.82116103e-01 9.02898550e-01 -9.32172120e-01
-1.36066043e+00 -6.50109410e-01 1.02704501e+00 -4.70647961e-01
6.16804302e-01 1.58827245e-01 -4.47598666e-01 3.67069215e-01
1.97967291e-01 -6.07346058e-01 1.31844807e+00 4.34037447e-01
2.62280926e-02 -2.47556210e-01 -1.18550241e+00 3.52594048e-01
4.58416641e-01 -2.36792207e-01 -6.01968944e-01 6.07327700e-01
1.71646804e-01 2.44902074e-01 -1.06686258e+00 1.98496394e-02
8.24557781e-01 -9.13484752e-01 6.15981340e-01 -8.50557387e-01
3.41997027e-01 -5.26092768e-01 -2.36849084e-01 -1.11772168e+00
-7.61605859e-01 -2.48243868e-01 4.14040238e-01 1.53339410e+00
5.56005657e-01 -4.49704736e-01 6.72343433e-01 8.66835296e-01
6.53617084e-02 -4.40138787e-01 2.52829581e-01 -1.70243252e-02
-4.34633851e-01 -3.74668151e-01 4.14715171e-01 1.01086140e+00
5.58797896e-01 5.27427077e-01 -8.09587955e-01 -3.43381539e-02
1.12103499e-01 2.24331021e-01 7.49278009e-01 -1.24839699e+00
-2.89306998e-01 -2.88332961e-02 1.80400684e-01 -1.05257201e+00
-5.30649602e-01 -5.27139544e-01 -2.56758988e-01 -1.52118444e+00
1.45055875e-01 -6.42072916e-01 -6.84047639e-01 -3.01530939e-02
4.73843403e-02 4.64683861e-01 -2.83593923e-01 7.57182956e-01
-8.74917030e-01 -2.68628359e-01 1.02126980e+00 4.07112181e-01
-4.03535962e-01 5.62995434e-01 -7.90698111e-01 9.28378284e-01
9.30890858e-01 -6.81023359e-01 -6.21747911e-01 -1.77982688e-01
7.97304332e-01 -1.22653931e-01 -8.10475290e-01 -4.22990382e-01
5.65086901e-01 -6.35062516e-01 2.84922332e-01 -7.56982684e-01
-9.59709287e-02 -7.95156956e-01 3.54790509e-01 1.41247526e-01
-5.54905295e-01 5.57258315e-02 -3.77395272e-01 6.40470386e-01
-2.15505615e-01 -4.86195028e-01 3.27481866e-01 -3.10264260e-01
-8.87746215e-01 2.44720474e-01 -1.08196688e+00 -7.63996184e-01
5.61599731e-01 -7.70417079e-02 -1.35740951e-01 -6.83939278e-01
-7.86447942e-01 -7.17686815e-03 2.48492837e-01 5.31381547e-01
4.54498798e-01 -8.63452673e-01 -4.39089090e-01 -4.55119796e-02
-3.21016684e-02 -8.28092873e-01 3.11191797e-01 7.36885428e-01
-3.69108170e-01 4.29747880e-01 -2.45551720e-01 4.68153596e-01
-1.45541334e+00 5.64161539e-01 -1.90204397e-01 2.96725370e-02
-1.61236614e-01 9.51334596e-01 -2.74899244e-01 -1.00981705e-01
3.33156049e-01 1.58138588e-01 -1.49072123e+00 3.42822492e-01
9.74985063e-01 4.47998762e-01 1.42857254e-01 -5.91817617e-01
-1.59201294e-01 2.46051043e-01 -5.57212532e-01 7.94074498e-03
1.32883215e+00 -7.35923827e-01 -2.70081282e-01 4.46172833e-01
6.69836044e-01 6.05801463e-01 -3.05033118e-01 -4.14855778e-01
-7.95366988e-02 -7.99128115e-01 6.23303473e-01 -7.29923308e-01
-1.09016740e+00 1.37047440e-01 4.04919088e-01 8.25926363e-01
9.43584561e-01 -4.19077098e-01 8.71046722e-01 3.91048223e-01
3.48861933e-01 -1.45121455e+00 -2.92885840e-01 4.43300068e-01
4.21278358e-01 -1.30557394e+00 1.27545580e-01 -4.15334493e-01
-4.19019341e-01 1.06525826e+00 5.43301046e-01 -3.18174005e-01
1.54611409e+00 -4.37697209e-02 3.17474842e-01 -2.43675765e-02
-7.96965957e-01 -1.51206508e-01 4.67615157e-01 3.16350460e-02
1.16348660e+00 -6.15746807e-03 -1.38158607e+00 1.13680232e+00
1.31652892e-01 2.29753852e-01 4.30502087e-01 1.04360676e+00
-9.80825424e-01 -1.35497606e+00 -3.20039898e-01 8.79041016e-01
-1.19363272e+00 -1.41441086e-02 -3.98751497e-01 -1.08402334e-01
2.51655132e-01 1.67675102e+00 -2.80937165e-01 -9.50508058e-01
2.15618238e-01 -2.83831358e-01 -3.63937169e-02 -8.90611827e-01
-1.05210984e+00 3.83714616e-01 4.53608006e-01 -1.50877938e-01
-6.05965257e-01 -6.43946230e-01 -8.87400389e-01 -4.54135805e-01
-8.24481606e-01 9.67367947e-01 9.93800104e-01 9.86265123e-01
1.61031082e-01 1.72896504e-01 1.09955060e+00 -4.36373651e-01
-1.87608078e-01 -1.17760301e+00 -6.67522788e-01 5.83554268e-01
-1.63318112e-01 -4.16798621e-01 -3.84862691e-01 4.03952412e-03] | [10.07238483428955, 5.83654260635376] |
287d816c-a1f5-4903-9387-6e7b22d5ca9d | cross-domain-activity-recognition-via | 2102.03353 | null | https://arxiv.org/abs/2102.03353v3 | https://arxiv.org/pdf/2102.03353v3.pdf | Cross-domain Activity Recognition via Substructural Optimal Transport | It is expensive and time-consuming to collect sufficient labeled data for human activity recognition (HAR). Domain adaptation is a promising approach for cross-domain activity recognition. Existing methods mainly focus on adapting cross-domain representations via domain-level, class-level, or sample-level distribution matching. However, they might fail to capture the fine-grained locality information in activity data. The domain- and class-level matching are too coarse that may result in under-adaptation, while sample-level matching may be affected by the noise seriously and eventually cause over-adaptation. In this paper, we propose substructure-level matching for domain adaptation (SSDA) to better utilize the locality information of activity data for accurate and efficient knowledge transfer. Based on SSDA, we propose an optimal transport-based implementation, Substructural Optimal Transport (SOT), for cross-domain HAR. We obtain the substructures of activities via clustering methods and seeks the coupling of the weighted substructures between different domains. We conduct comprehensive experiments on four public activity recognition datasets (i.e. UCI-DSADS, UCI-HAR, USC-HAD, PAMAP2), which demonstrates that SOT significantly outperforms other state-of-the-art methods w.r.t classification accuracy (9%+ improvement). In addition, our mehtod is 5x faster than traditional OT-based DA methods with the same hyper-parameters. | ['Xin Qin', 'Jindong Wang', 'Yiqiang Chen', 'Wang Lu'] | 2021-01-29 | null | null | null | null | ['cross-domain-activity-recognition'] | ['computer-vision'] | [ 1.95964023e-01 -7.31716692e-01 -5.39496005e-01 -3.72484297e-01
-1.03309214e+00 -5.00307918e-01 4.80131418e-01 1.56363130e-01
-3.66307586e-01 9.40757275e-01 4.23019469e-01 1.04563370e-01
-3.25682074e-01 -7.68033862e-01 -7.03087747e-01 -7.73276091e-01
-1.00656167e-01 4.42342699e-01 7.39929199e-01 8.54890049e-02
1.29345775e-01 3.97778481e-01 -1.35082912e+00 6.29561841e-01
1.11463821e+00 1.00575280e+00 4.03403081e-02 2.33378246e-01
-3.74033093e-01 5.65833092e-01 -7.20420122e-01 -7.08568143e-03
-1.21531049e-02 -6.22761548e-01 -8.87631595e-01 2.03002483e-01
2.46098831e-01 -2.95628700e-02 -2.43828431e-01 7.16019511e-01
4.96837169e-01 3.64673048e-01 9.77456987e-01 -1.28974819e+00
-6.09122872e-01 2.28125066e-01 -8.44820738e-01 4.40669924e-01
4.39357877e-01 -2.03752965e-02 4.57033247e-01 -9.06782985e-01
4.28242117e-01 1.02998734e+00 7.65459180e-01 5.56321979e-01
-1.18954539e+00 -7.43718743e-01 3.25739533e-01 5.91918111e-01
-1.66666281e+00 -2.64259398e-01 7.69108891e-01 -3.33776623e-01
1.01945913e+00 6.88111633e-02 5.64939857e-01 1.51599717e+00
-2.57043587e-03 1.11742020e+00 1.16793334e+00 -3.12421948e-01
4.65422034e-01 -1.17650218e-01 1.11257993e-01 4.40100372e-01
2.23413691e-01 -2.55194098e-01 -9.84190047e-01 -4.01668638e-01
7.56051242e-01 1.30392283e-01 -1.69797704e-01 -6.35282755e-01
-1.46388185e+00 4.79206741e-01 2.31035069e-01 5.58374345e-01
-3.47378671e-01 -2.53665149e-01 4.31005925e-01 1.75229296e-01
4.45577383e-01 1.73640866e-02 -6.63164318e-01 -5.70167899e-01
-7.80951321e-01 9.75065231e-02 5.67173302e-01 1.10436809e+00
7.46103287e-01 -8.25939253e-02 -3.35094869e-01 1.33181942e+00
7.61400759e-02 4.71042216e-01 8.22799325e-01 -7.93740511e-01
7.41435111e-01 8.22282970e-01 -4.29989211e-02 -6.69675469e-01
-2.35290781e-01 -4.04143453e-01 -1.01815367e+00 -3.50992888e-01
6.62063241e-01 4.80524935e-02 -8.22898030e-01 1.73606098e+00
5.46349287e-01 6.38535321e-01 1.02276012e-01 6.24964476e-01
3.05431783e-01 6.55331314e-01 2.83147424e-01 -2.20475912e-01
1.36476767e+00 -1.11326814e+00 -5.93876779e-01 -3.61564904e-01
9.30903077e-01 -4.00156260e-01 1.15702248e+00 2.54982322e-01
-7.81496167e-01 -6.59823895e-01 -1.15242124e+00 2.89279610e-01
-4.93500680e-01 -7.41097257e-02 3.82846504e-01 6.55498922e-01
-4.21631277e-01 4.98871505e-01 -8.85778546e-01 -7.21698403e-01
7.44192243e-01 1.07051358e-01 -5.39264917e-01 -3.09465855e-01
-1.20308340e+00 6.06795371e-01 4.15539831e-01 -4.21407074e-01
-7.49789059e-01 -8.87848318e-01 -6.67168319e-01 -1.28151864e-01
3.36639374e-01 -4.22938675e-01 1.01597381e+00 -8.60517263e-01
-1.37990904e+00 5.59000254e-01 -3.82534921e-01 -3.49173456e-01
4.02668864e-01 -1.74348440e-03 -9.18993175e-01 -7.58743752e-03
3.22951972e-01 3.73950690e-01 4.55013096e-01 -9.28567827e-01
-9.22332466e-01 -5.96718132e-01 -5.10462165e-01 3.47668678e-01
-7.08678246e-01 -1.19890012e-01 -5.85879385e-01 -9.51512992e-01
2.02390086e-03 -7.71742105e-01 1.07763425e-01 -7.96964988e-02
5.54015823e-02 -3.73629749e-01 9.18325305e-01 -5.15268445e-01
1.52511835e+00 -2.20051050e+00 -7.54616037e-02 2.21472025e-01
-2.33286873e-01 3.61539662e-01 -1.23174995e-01 3.41309071e-01
1.00277856e-01 -2.77107686e-01 -5.42657495e-01 2.37681884e-02
-5.21126948e-02 4.54752773e-01 1.35956630e-01 4.91859704e-01
3.65041755e-02 6.46371722e-01 -8.98218334e-01 -7.31696725e-01
8.25724825e-02 2.65000314e-01 -5.65843761e-01 1.68114781e-01
4.96140160e-02 4.39108282e-01 -5.13570964e-01 7.48468757e-01
6.94976985e-01 -3.15585196e-01 4.03340638e-01 -1.76851183e-01
1.66110948e-01 2.95213163e-01 -1.40751052e+00 2.08094978e+00
-3.31486464e-01 2.66518295e-01 -3.88033718e-01 -1.20581162e+00
8.82616580e-01 8.87103975e-02 8.79911900e-01 -8.96494865e-01
-2.44462967e-01 3.30008090e-01 -3.11325490e-01 -1.80704206e-01
1.75991058e-01 1.61233097e-01 -5.12616992e-01 3.98599327e-01
1.24516867e-01 6.35865510e-01 6.29034340e-02 3.42015736e-02
1.34594131e+00 2.93736637e-01 6.36214197e-01 -2.90508956e-01
6.19605541e-01 1.19010307e-01 8.90239894e-01 5.55128694e-01
-5.36043823e-01 5.19459963e-01 7.66017511e-02 -1.04449570e-01
-7.68916368e-01 -1.28332853e+00 -1.62403658e-01 1.43811405e+00
2.24903271e-01 -3.04393381e-01 -9.93175566e-01 -1.09406328e+00
1.23576205e-02 4.56902593e-01 -5.21701515e-01 -4.71596509e-01
-7.30041206e-01 -7.90728033e-01 9.57948625e-01 9.40523803e-01
1.22599661e+00 -9.36537385e-01 -2.55826022e-02 4.91839170e-01
-5.98916292e-01 -1.04931009e+00 -1.03208041e+00 2.02300519e-01
-9.72471893e-01 -9.25034642e-01 -9.15041745e-01 -7.95088053e-01
5.40481329e-01 4.13970858e-01 8.92619133e-01 -3.32108796e-01
-2.66143307e-02 2.65378177e-01 -4.27523524e-01 1.67034436e-02
-5.31258807e-02 4.22133118e-01 2.57547706e-01 3.09736282e-01
9.82490540e-01 -7.01395273e-01 -5.82437932e-01 9.51262474e-01
-8.14399004e-01 -4.11463857e-01 7.14521229e-01 8.65510702e-01
8.03527951e-01 6.47037104e-02 6.53518617e-01 -7.38606036e-01
5.46877563e-01 -8.08180630e-01 -1.09538540e-01 4.34231132e-01
-6.53102100e-01 1.15667388e-01 5.83341837e-01 -9.39128816e-01
-1.41931117e+00 2.27000639e-01 1.57468081e-01 -3.78467977e-01
-4.70671892e-01 2.53069222e-01 -6.47597492e-01 1.55474648e-01
1.01554322e+00 6.87749028e-01 -2.68937856e-01 -7.38946378e-01
-1.03567969e-02 9.94030774e-01 6.02082968e-01 -9.65439737e-01
5.37241876e-01 5.28160393e-01 -3.94858599e-01 -5.90110898e-01
-8.07858586e-01 -8.75646949e-01 -8.89593959e-01 -5.41443229e-02
7.90322363e-01 -1.00444937e+00 -1.85007185e-01 7.55294859e-01
-7.58458495e-01 -4.38938856e-01 -1.74274743e-01 4.39766765e-01
-5.67360282e-01 5.77283204e-01 -3.38143647e-01 -4.40462261e-01
-8.24133083e-02 -7.33033657e-01 9.22816813e-01 2.07887396e-01
-3.10295790e-01 -1.00850308e+00 3.88290316e-01 6.75041139e-01
2.33152911e-01 3.32635678e-02 6.14577830e-01 -7.76582062e-01
-3.26849729e-01 1.67250276e-01 -1.16744742e-01 2.10520208e-01
4.20487463e-01 -4.57438141e-01 -7.84663558e-01 -2.50363588e-01
-5.53669333e-01 -2.03213707e-01 5.39359093e-01 2.26807624e-01
1.31604791e+00 -2.24976614e-01 -7.31693447e-01 6.05819702e-01
1.06284320e+00 3.75224978e-01 7.56799221e-01 6.02560878e-01
6.33981109e-01 3.60030562e-01 9.92308140e-01 6.78103387e-01
5.53098798e-01 1.15711021e+00 -2.54267275e-01 2.50636395e-02
-3.07379782e-01 -4.64999110e-01 6.41115308e-01 6.89852536e-01
-1.64721149e-03 -1.42233204e-02 -9.30189133e-01 9.70753074e-01
-2.12247157e+00 -1.12129438e+00 -1.12652585e-01 2.13540220e+00
1.05351281e+00 5.08009121e-02 5.52787721e-01 1.30539104e-01
6.92779601e-01 -2.59558801e-02 -7.05754399e-01 -5.09441495e-02
-1.28543511e-01 6.86554313e-02 5.72574615e-01 9.41584259e-02
-1.14854968e+00 7.38511264e-01 5.57511282e+00 1.31327450e+00
-4.77280021e-01 4.31737423e-01 1.02237523e-01 -1.37445241e-01
1.24769188e-01 -3.30912292e-01 -9.88492489e-01 9.47787821e-01
1.03706729e+00 -9.05545577e-02 3.81125361e-01 1.02801824e+00
-6.80742692e-03 -1.28303945e-01 -9.99052405e-01 1.02990615e+00
7.70270601e-02 -1.08590055e+00 1.83866862e-02 1.19473018e-01
7.61221647e-01 -7.34880269e-02 -2.82662272e-01 5.75687826e-01
4.23822939e-01 -4.78059858e-01 4.23548639e-01 3.32790494e-01
6.92750394e-01 -8.13510954e-01 5.41100323e-01 5.17454743e-01
-1.73764741e+00 -5.42486049e-02 -3.07711065e-01 2.98829466e-01
5.95824905e-02 4.88079250e-01 -5.08902013e-01 6.92476749e-01
1.11920190e+00 1.01251924e+00 -5.34508824e-01 8.93514812e-01
1.42294556e-01 6.41193092e-01 -3.23352933e-01 5.12628965e-02
9.01087821e-02 -1.13994971e-01 2.07358822e-01 1.46423817e+00
4.72293586e-01 1.01709716e-01 2.35651657e-01 3.65506828e-01
-1.32247806e-01 -4.47837822e-02 -3.74998122e-01 2.18953297e-01
7.94283807e-01 7.29271591e-01 -3.51066858e-01 -6.04436815e-01
-4.57838386e-01 1.32321119e+00 4.28350985e-01 2.57524282e-01
-1.13338029e+00 -4.81335878e-01 1.06085992e+00 2.67219216e-01
3.88545901e-01 -1.64407015e-01 -1.92860335e-01 -1.35364389e+00
1.78445548e-01 -1.06083620e+00 1.16858244e+00 -4.84525621e-01
-1.57440841e+00 2.49426976e-01 3.16274911e-01 -1.64495301e+00
-7.53734857e-02 -3.33190292e-01 -2.12656200e-01 5.55413783e-01
-1.44128251e+00 -9.58738804e-01 -5.01021862e-01 1.25929260e+00
8.02751601e-01 -2.86122680e-01 6.90816581e-01 8.15385520e-01
-6.62980795e-01 9.26527858e-01 2.83636391e-01 2.64383852e-01
1.00207901e+00 -9.65342104e-01 3.04692626e-01 7.94420660e-01
1.34618223e-01 4.77945298e-01 1.57436505e-01 -6.33269966e-01
-1.08211899e+00 -1.37196481e+00 7.32942402e-01 -6.03454113e-01
5.57676375e-01 -3.45159143e-01 -1.30197096e+00 5.73016703e-01
-1.41729459e-01 1.78962275e-01 1.01919544e+00 7.73814842e-02
-5.47406316e-01 -4.87926126e-01 -1.24269104e+00 4.62966651e-01
1.72141433e+00 -4.59280401e-01 -7.47780919e-01 1.65118665e-01
2.84308165e-01 -9.90769267e-02 -1.26262045e+00 3.28205973e-01
5.08400500e-01 -7.97578335e-01 1.17657948e+00 -6.65648758e-01
-1.69399202e-01 -5.24421334e-01 -3.03518474e-01 -1.35982573e+00
-6.43419147e-01 -2.62496680e-01 -3.58580202e-01 1.45759618e+00
3.41136545e-01 -6.06631935e-01 9.06427681e-01 2.56842673e-01
-1.58304974e-01 -3.91929388e-01 -1.11486185e+00 -1.34449315e+00
-5.12340805e-03 -4.82603759e-01 9.42443252e-01 1.24497104e+00
1.36347264e-01 2.02314034e-01 -4.11197126e-01 2.25756347e-01
7.84301281e-01 -9.07887053e-03 8.02328169e-01 -1.15910363e+00
-3.39840293e-01 -3.33544761e-01 -2.63099492e-01 -1.34598303e+00
3.73778045e-02 -6.51914060e-01 -5.63821755e-02 -1.38495922e+00
3.05008292e-01 -4.12951291e-01 -6.47744596e-01 7.20159054e-01
-6.36203098e-04 1.12614617e-01 -2.90484607e-01 4.84209359e-01
-9.98660386e-01 5.70325613e-01 1.03994739e+00 -2.88559169e-01
-3.87520850e-01 -4.83714715e-02 -5.10572314e-01 5.25178254e-01
9.68361199e-01 -7.05501318e-01 -4.47785348e-01 -2.72626221e-01
-5.57500362e-01 -1.74391672e-01 7.77978450e-02 -1.38587022e+00
3.76763195e-01 -5.53386509e-01 6.16335213e-01 -6.03432596e-01
3.37092042e-01 -9.22151387e-01 2.44680002e-01 4.00529444e-01
-2.13957638e-01 -2.47224808e-01 1.61800966e-01 1.02570701e+00
-1.56414151e-01 1.53290570e-01 9.82711613e-01 8.44523758e-02
-1.03006279e+00 3.94058913e-01 -5.83774447e-01 2.74300098e-01
1.23003197e+00 -6.79456174e-01 -2.84364551e-01 1.06891565e-01
-5.36078036e-01 1.42813712e-01 3.54304433e-01 5.27810633e-01
3.73242050e-01 -1.80805719e+00 -4.48586911e-01 2.68114686e-01
7.91038752e-01 -1.98744684e-01 4.33266789e-01 7.99444973e-01
-5.28842658e-02 2.83324212e-01 -3.78399849e-01 -6.78030908e-01
-1.31900978e+00 3.67830932e-01 3.93730313e-01 -5.63307762e-01
-4.02994961e-01 5.72198808e-01 4.28901315e-02 -5.86372137e-01
2.12168038e-01 -2.94247568e-01 -3.09303906e-02 -4.21263836e-02
5.78087747e-01 6.36005402e-01 1.01388909e-01 -4.35827076e-01
-8.49258780e-01 6.69313788e-01 -4.46740091e-02 8.18344653e-02
1.03856397e+00 -6.90714791e-02 3.96291107e-01 2.74413735e-01
1.11819077e+00 -2.77168721e-01 -1.48915851e+00 -7.05040336e-01
3.21018308e-01 -6.67955160e-01 -2.12850928e-01 -8.65375340e-01
-7.40166783e-01 7.73682177e-01 7.05666602e-01 -1.23020172e-01
1.29156983e+00 1.36273965e-01 9.79641497e-01 3.51386756e-01
5.46570003e-01 -1.66000152e+00 4.35064077e-01 2.49123976e-01
4.48729813e-01 -1.00485635e+00 -5.94005063e-02 -2.32419416e-01
-8.87552679e-01 6.64242983e-01 9.60287869e-01 2.60411114e-01
4.18720394e-01 1.21017560e-01 -2.16594502e-01 9.15432200e-02
-6.01511955e-01 -2.87670493e-01 2.54619092e-01 1.13348937e+00
1.16230227e-01 9.20692384e-02 -9.81121883e-02 7.16012061e-01
4.30510700e-01 1.45981342e-01 2.64032464e-02 1.14207840e+00
-6.19536757e-01 -1.34474719e+00 -4.81508493e-01 3.26862723e-01
-1.95067480e-01 3.34342301e-01 -3.27155799e-01 7.43046045e-01
1.76247254e-01 9.46551263e-01 1.38575491e-02 -3.78168166e-01
7.93768346e-01 2.06129342e-01 5.03996789e-01 -4.63667244e-01
-3.70734721e-01 -5.43189198e-02 2.14433298e-01 -6.80447400e-01
-6.21970952e-01 -8.95286977e-01 -1.36199892e+00 -5.56198895e-01
3.39806490e-02 1.40887260e-01 1.29813075e-01 8.84422302e-01
7.84245253e-01 4.12943721e-01 5.05444646e-01 -2.78320342e-01
-3.82312000e-01 -1.04646635e+00 -7.05659568e-01 6.39266491e-01
-1.73612684e-01 -9.29228604e-01 -1.54429123e-01 1.89263880e-01] | [8.04428482055664, 1.0286751985549927] |
cf70d098-5fc2-4474-a890-9836369bcec7 | adaptively-learning-facial-expression-1 | null | null | https://doi.org/10.1109/TIP.2021.3049955 | https://doi.org/10.1109/TIP.2021.3049955 | Adaptively Learning Facial Expression Representation via C-F Labels and Distillation | Facial expression recognition is of significant importance in criminal investigation and digital entertainment. Under unconstrained conditions, existing expression datasets are highly class-imbalanced, and the similarity between expressions is high. Previous methods tend to improve the performance of facial expression recognition through deeper or wider network structures, resulting in increased storage and computing costs. In this paper, we propose a new adaptive supervised objective named AdaReg loss, re-weighting category importance coefficients to address this class imbalance and increasing the discrimination power of expression representations. Inspired by human beings’ cognitive mode, an innovative coarse-fine (C-F) labels strategy is designed to guide the model from easy to difficult to classify highly similar representations. On this basis, we propose a novel training framework named the emotional education mechanism (EEM) to transfer knowledge, composed of a knowledgeable teacher network (KTN) and a self-taught student network (STSN). Specifically, KTN integrates the outputs of coarse and fine streams, learning expression representations from easy to difficult. Under the supervision of the pre-trained KTN and existing learning experience, STSN can maximize the potential performance and compress the original KTN. Extensive experiments on public benchmarks demonstrate that the proposed method achieves superior performance compared to current state-of-the-art frameworks with 88.07% on RAF-DB, 63.97% on AffectNet and 90.49% on FERPlus. | ['Hangyu Li; Nannan Wang; Xinpeng Ding; Xi Yang; Xinbo Gao'] | 2021-01-13 | null | null | null | ieee-transactions-on-image-processing-2021-1 | ['facial-expression-recognition'] | ['computer-vision'] | [ 1.91924259e-01 -9.36975703e-02 -3.36674899e-01 -8.25826526e-01
-3.80620658e-02 1.49258614e-01 2.21206099e-01 -3.02008241e-01
-4.97766793e-01 7.07448483e-01 -4.55137268e-02 1.98568434e-01
-3.66756953e-02 -8.12365115e-01 -2.80991256e-01 -7.73973465e-01
4.04080451e-02 7.91949853e-02 -2.61515468e-01 -4.01436210e-01
1.31465733e-01 5.53839505e-01 -1.59431529e+00 5.71133554e-01
9.42885399e-01 1.72937584e+00 -2.81074643e-01 4.97190021e-02
-4.05990064e-01 1.14628613e+00 -6.46080315e-01 -7.35410452e-01
-8.35439116e-02 -3.26748997e-01 -7.62382030e-01 2.41643861e-02
4.20748107e-02 -3.57806802e-01 -1.61125019e-01 1.16166437e+00
6.81220949e-01 1.95347190e-01 5.61974406e-01 -1.45859432e+00
-6.23766243e-01 3.10816497e-01 -8.22502017e-01 1.60297945e-01
4.51217964e-02 -6.92443997e-02 5.10823131e-01 -9.07886684e-01
3.53463292e-01 1.31122327e+00 6.42880619e-01 8.56307268e-01
-8.00350845e-01 -1.27784657e+00 3.01388502e-01 5.39623439e-01
-1.46938252e+00 -4.68562990e-01 9.58091140e-01 -2.84167051e-01
6.99723721e-01 2.94975825e-02 6.33796275e-01 1.20530617e+00
-1.39134124e-01 9.00987983e-01 1.27400458e+00 -1.27967805e-01
9.41113383e-02 4.17909831e-01 1.05610080e-01 7.04211891e-01
-2.10822657e-01 -7.61788040e-02 -7.30776072e-01 -7.74448738e-02
4.85279173e-01 2.02837765e-01 -1.62024781e-01 -1.25271780e-02
-2.78283417e-01 7.71701157e-01 6.64153218e-01 2.41574198e-01
-5.69606364e-01 -7.30905458e-02 6.95072472e-01 3.82713050e-01
6.39280915e-01 4.40812260e-02 -3.24849099e-01 -3.51336569e-01
-6.61246359e-01 -1.08329341e-01 4.01552707e-01 4.32732552e-01
8.61748755e-01 1.62422627e-01 -2.69528538e-01 1.13034117e+00
1.94556817e-01 3.22436929e-01 7.11760700e-01 -7.37493336e-01
2.71187961e-01 8.34842384e-01 -5.17612457e-01 -1.35699654e+00
-2.05892578e-01 -6.19160295e-01 -1.23039079e+00 1.43841833e-01
-1.34982765e-01 -2.86648005e-01 -7.31810391e-01 1.90006614e+00
2.90234894e-01 3.53075325e-01 8.04714113e-02 8.72430921e-01
8.35261106e-01 8.92845929e-01 4.21676666e-01 -3.57537538e-01
1.12374032e+00 -8.84843171e-01 -8.34080160e-01 -7.30571225e-02
4.46230531e-01 -3.51358861e-01 8.77451956e-01 6.83721662e-01
-9.73612487e-01 -6.90905273e-01 -8.59394372e-01 1.89283267e-01
-3.41058016e-01 2.24200845e-01 8.51029634e-01 6.21284544e-01
-7.59431720e-01 5.35968602e-01 -3.07759255e-01 1.03480130e-01
1.09113932e+00 5.04922450e-01 -1.80505618e-01 -6.45643249e-02
-1.46644866e+00 7.65893161e-01 3.33341509e-01 2.81727791e-01
-6.56525433e-01 -6.40565991e-01 -7.05949306e-01 2.99509048e-01
1.29510820e-01 -2.63339430e-01 1.07085896e+00 -1.86723685e+00
-1.88106930e+00 7.74999201e-01 6.59991428e-02 -1.08124457e-01
2.96233624e-01 -1.76223710e-01 -7.03134954e-01 2.71991968e-01
-3.93193990e-01 7.86273956e-01 8.44378114e-01 -9.71292794e-01
-4.21209842e-01 -5.34353673e-01 -6.80207089e-02 2.75226325e-01
-1.01689219e+00 2.33883277e-01 -2.80416876e-01 -6.37790501e-01
-2.98492163e-01 -4.96719837e-01 -4.27972302e-02 3.14284831e-01
9.06448215e-02 -5.86774707e-01 9.18507874e-01 -6.56457305e-01
1.31746912e+00 -2.29093266e+00 4.31799181e-02 5.37928700e-01
1.05082706e-01 6.92209542e-01 -1.81732848e-01 -2.97352701e-01
-2.68509895e-01 -6.05438240e-02 -1.40644878e-01 -1.62407354e-01
7.17439223e-03 3.25593948e-01 -1.50792986e-01 1.49983838e-01
4.51628327e-01 7.15953708e-01 -7.64026940e-01 -6.59891307e-01
-1.62463143e-01 5.97318113e-01 -6.48338854e-01 4.05358106e-01
1.65680006e-01 1.87778994e-01 -6.26737654e-01 9.29129064e-01
8.21015775e-01 -1.81239888e-01 8.91180784e-02 -1.95264608e-01
2.03096777e-01 -2.09667280e-01 -9.79330063e-01 1.46966922e+00
-5.61229646e-01 3.20014060e-01 1.62288964e-01 -1.33682477e+00
1.40273142e+00 1.55444011e-01 3.69817734e-01 -1.15734673e+00
4.49038297e-01 7.70547912e-02 -7.49781728e-02 -6.47497118e-01
2.11209193e-01 -2.70062685e-01 1.56436171e-02 1.89410672e-01
2.17154816e-01 4.65680927e-01 -1.24386281e-01 -4.45041098e-02
7.47913241e-01 -3.01207714e-02 6.84119985e-02 4.69488986e-02
7.01629400e-01 -6.12620473e-01 9.67155755e-01 3.92766930e-02
-3.76393914e-01 -1.06211573e-01 6.05020404e-01 -5.62896073e-01
-4.27011728e-01 -8.06094766e-01 -1.82989463e-01 1.57856095e+00
7.20168427e-02 -2.06033483e-01 -8.37187827e-01 -6.76387906e-01
-9.78913009e-02 3.11898738e-01 -7.29583263e-01 -7.26418912e-01
-3.53153735e-01 -7.90682375e-01 6.76403403e-01 5.66118956e-01
9.66455758e-01 -1.37755990e+00 -3.10277581e-01 5.15665580e-03
-1.30285844e-01 -7.43327558e-01 -3.18484828e-02 3.52222919e-02
-6.27828717e-01 -6.28745198e-01 -7.64848292e-01 -6.90754533e-01
7.74161935e-01 -2.73279548e-01 1.05701482e+00 2.06097156e-01
-2.08732754e-01 -9.05682147e-02 -4.44982231e-01 -5.39360881e-01
5.81572913e-02 -8.81911516e-02 4.48437855e-02 6.26886904e-01
6.59893394e-01 -7.08032668e-01 -5.29621780e-01 2.20074818e-01
-8.86983812e-01 -9.52996090e-02 8.17307711e-01 9.97829556e-01
4.31699365e-01 -3.82102793e-03 6.96037233e-01 -6.85969770e-01
7.56600261e-01 -7.05133736e-01 -1.42588049e-01 1.89458430e-01
-7.18316436e-01 -2.42728055e-01 7.51350582e-01 -6.54565096e-01
-1.37440479e+00 -1.90796867e-01 -2.76995122e-01 -7.05222487e-01
-4.35704626e-02 4.34927881e-01 -3.64880055e-01 -2.61944860e-01
3.72242212e-01 3.47048789e-01 1.88124940e-01 -1.43757448e-01
-7.46546760e-02 9.05951560e-01 4.66876835e-01 -7.23835170e-01
3.34390342e-01 1.49388492e-01 -2.23658666e-01 -4.93885607e-01
-8.77290010e-01 -9.68788937e-02 -2.44830295e-01 -4.93026853e-01
6.76777780e-01 -9.71639335e-01 -1.02590322e+00 7.42826104e-01
-1.03537488e+00 -2.20877230e-01 -1.43796071e-01 2.31574088e-01
-2.52946109e-01 7.94584379e-02 -7.05017805e-01 -8.53285551e-01
-5.89044869e-01 -9.04127896e-01 8.87697816e-01 7.80696511e-01
-1.51100814e-01 -6.61337733e-01 -2.55691528e-01 4.13666427e-01
6.86709583e-01 2.78571069e-01 7.10284948e-01 -3.60696912e-01
2.68710945e-02 2.00356524e-02 -5.43604136e-01 7.21281230e-01
-1.47174746e-01 -4.08136286e-02 -1.13893247e+00 4.53719608e-02
-4.99564148e-02 -1.02392900e+00 8.98888052e-01 -4.91189435e-02
2.00681496e+00 -3.91008258e-01 -1.60947353e-01 8.38372529e-01
1.14613640e+00 3.37465823e-01 9.15292084e-01 2.56204665e-01
4.71542686e-01 7.46041834e-01 6.01484776e-01 8.20304036e-01
4.29211020e-01 5.33816040e-01 3.36927921e-01 -3.40099901e-01
2.09829554e-01 -1.47902146e-01 5.60296357e-01 7.66215026e-01
-2.16553837e-01 7.06988648e-02 -5.55566728e-01 2.15701699e-01
-1.60409129e+00 -1.05998695e+00 5.14940619e-01 1.57224333e+00
1.22626138e+00 -2.61552855e-02 -1.97673179e-02 1.47280246e-01
5.80153942e-01 1.22655019e-01 -7.71072686e-01 -7.92528272e-01
-1.15483060e-01 5.52454710e-01 -2.36424315e-03 9.42451581e-02
-9.38718319e-01 8.28469753e-01 5.00038862e+00 1.35976017e+00
-1.56840551e+00 1.28635034e-01 1.16991031e+00 -2.07521424e-01
-1.07627560e-03 -6.60462499e-01 -6.98304176e-01 8.06032896e-01
9.01079178e-01 -5.41328192e-02 2.93296695e-01 1.06100702e+00
-6.40397146e-03 3.04171771e-01 -6.20275021e-01 1.33174813e+00
2.27742940e-01 -9.66473937e-01 2.07028016e-01 -1.81809038e-01
4.73795027e-01 -4.60480601e-01 2.59939790e-01 8.30536246e-01
-4.51886542e-02 -1.24872565e+00 3.06708634e-01 8.99800181e-01
8.85177910e-01 -1.16315055e+00 9.24415767e-01 1.92416921e-01
-9.22875464e-01 -2.84753293e-01 -5.78678489e-01 -1.26906648e-01
-1.54074892e-01 5.55958986e-01 -2.15827942e-01 3.87432069e-01
9.13150191e-01 8.51956785e-01 -4.25696105e-01 4.65580285e-01
-2.27114633e-01 5.98442674e-01 -1.16000801e-01 -2.49573156e-01
2.42315874e-01 -1.78474531e-01 -1.43232182e-01 1.25238729e+00
2.38709301e-01 5.95998108e-01 1.44733638e-01 5.77848256e-01
-3.72035235e-01 3.17853451e-01 -1.71959504e-01 8.08873698e-02
2.54308105e-01 1.50976276e+00 -2.60721475e-01 -4.54036862e-01
-1.70060858e-01 9.87090170e-01 7.42263436e-01 1.67419165e-01
-8.29542696e-01 -5.57654321e-01 7.45319963e-01 -2.00968266e-01
1.45156592e-01 4.03766274e-01 -3.13212397e-03 -9.22267854e-01
6.82025477e-02 -1.08118224e+00 6.47677362e-01 -7.93680906e-01
-1.41270304e+00 7.86213696e-01 -3.52163076e-01 -1.03737211e+00
1.01106763e-01 -7.84174681e-01 -7.08585501e-01 6.28588080e-01
-1.86840320e+00 -9.72226322e-01 -6.26559198e-01 8.87577891e-01
2.58925140e-01 -3.58842820e-01 9.36819971e-01 7.01260507e-01
-8.54931355e-01 1.04371548e+00 -1.60862729e-01 3.76503915e-01
6.16951227e-01 -8.35618138e-01 -5.65882087e-01 3.14244717e-01
-3.19102883e-01 3.06553066e-01 5.47495373e-02 -1.67569891e-01
-1.01708281e+00 -1.25130928e+00 6.65792048e-01 3.55518788e-01
3.28599572e-01 -2.01231152e-01 -1.08490288e+00 1.18584059e-01
9.16849300e-02 2.90259451e-01 1.02728593e+00 1.76864475e-01
-5.06449759e-01 -7.20695674e-01 -1.32059956e+00 3.06222200e-01
1.04360962e+00 -4.58141685e-01 -3.60648304e-01 1.47522077e-01
2.64767915e-01 -1.99119493e-01 -9.16944444e-01 6.67688787e-01
7.24653423e-01 -9.88828063e-01 7.79569268e-01 -7.96291947e-01
8.44783187e-01 9.82979611e-02 -1.03046440e-01 -1.25710511e+00
-3.41910660e-01 -2.74232686e-01 -2.30003685e-01 1.29855406e+00
1.65576227e-02 -5.81913173e-01 9.10541594e-01 4.67162758e-01
6.60722330e-02 -1.39971638e+00 -7.63072789e-01 -5.44172585e-01
-1.29574075e-01 -9.05336887e-02 7.67982066e-01 1.37374270e+00
-5.75391622e-03 3.63647312e-01 -4.93993878e-01 -1.33720651e-01
3.01507771e-01 4.03082557e-02 5.22124827e-01 -1.26998127e+00
-1.19153641e-01 -7.39490449e-01 -6.01540446e-01 -8.94302905e-01
6.27148032e-01 -9.30412591e-01 -2.50198841e-01 -8.43022704e-01
3.76686722e-01 -4.35105979e-01 -7.87220836e-01 7.98778713e-01
-2.20853269e-01 3.57773155e-01 1.14057720e-01 -8.17273185e-02
-8.96217346e-01 1.09477413e+00 1.25905335e+00 -2.58034497e-01
8.86482075e-02 -1.21698529e-01 -9.03322279e-01 9.21849787e-01
8.22524905e-01 -3.41063440e-01 -4.60618466e-01 -3.57591301e-01
2.93933582e-02 -1.16676100e-01 1.71277881e-01 -8.63321304e-01
2.25739822e-01 -1.41263276e-01 8.22539628e-01 -2.87961215e-01
3.88646036e-01 -7.68898606e-01 -2.61197686e-01 3.75297755e-01
-4.48845536e-01 -1.32937759e-01 2.41364062e-01 2.20002711e-01
-5.32403469e-01 -6.55451417e-02 8.93531442e-01 1.25969827e-01
-1.00783753e+00 7.60490954e-01 -1.27764896e-01 -7.39204511e-02
1.05570281e+00 -2.21902505e-01 -2.24505767e-01 -4.13265795e-01
-6.36380911e-01 4.04078037e-01 -1.29011527e-01 4.86124039e-01
8.45403433e-01 -1.66243279e+00 -6.09562576e-01 2.06768543e-01
9.63088125e-02 -1.57339975e-01 5.73461711e-01 6.85974538e-01
-1.58861458e-01 -2.44851723e-01 -7.21794069e-01 -4.29588526e-01
-1.44807816e+00 1.26373440e-01 5.08198380e-01 -4.24384594e-01
4.41904813e-02 1.14782476e+00 1.34032726e-01 -6.09367728e-01
3.70236725e-01 2.24674284e-01 -5.99767268e-01 3.45098078e-01
8.10600102e-01 3.17794204e-01 -1.32428348e-01 -6.46951556e-01
-2.71851718e-01 5.92701733e-01 -2.18440145e-01 4.35475558e-01
1.44924831e+00 1.62057728e-01 -3.14078212e-01 1.03828393e-01
1.45158935e+00 -3.99207324e-01 -1.03023398e+00 -3.88601422e-01
-2.91396022e-01 -3.43181670e-01 8.80794153e-02 -9.22618330e-01
-1.48830295e+00 1.11055863e+00 8.82052779e-01 -2.81044602e-01
1.83562911e+00 -2.19156325e-01 8.06413770e-01 4.99285102e-01
1.03084780e-01 -1.35509479e+00 4.40605164e-01 4.59211856e-01
7.82524347e-01 -1.07402492e+00 -1.26830712e-01 -3.11429143e-01
-7.95063198e-01 1.07507563e+00 1.16290927e+00 9.52651631e-03
6.71021938e-01 2.09555879e-01 1.30795136e-01 -2.43622482e-01
-8.36777747e-01 1.63054630e-01 1.00832954e-01 4.75117773e-01
3.15480113e-01 -4.67406958e-02 -2.48076960e-01 1.13800967e+00
-9.75651443e-02 3.72202933e-01 -1.07572310e-01 7.34976590e-01
-3.69134575e-01 -8.93315971e-01 -8.30302536e-02 5.39043546e-01
-6.25740230e-01 7.84613043e-02 -2.98382074e-01 4.29929048e-01
6.72464788e-01 6.13240123e-01 2.81781465e-01 -5.49186707e-01
3.66401464e-01 9.67094451e-02 2.94088811e-01 -9.47854817e-02
-6.08259797e-01 -3.25941592e-01 -1.58415791e-02 -6.87715113e-01
-6.28663957e-01 -3.74634802e-01 -1.27721047e+00 -2.78368205e-01
-7.45322555e-02 2.64065683e-01 3.78463954e-01 9.03790116e-01
5.48985183e-01 6.16635144e-01 1.10799170e+00 -5.96986175e-01
-6.75421178e-01 -9.09218192e-01 -5.56849182e-01 5.88324606e-01
-2.47040167e-02 -8.15181792e-01 -9.21906084e-02 -1.77277066e-02] | [13.601231575012207, 1.6808793544769287] |
1f5057b7-9479-4d69-9bcb-6cad86613de5 | interpretability-driven-sample-selection | 2104.06087 | null | https://arxiv.org/abs/2104.06087v1 | https://arxiv.org/pdf/2104.06087v1.pdf | Interpretability-Driven Sample Selection Using Self Supervised Learning For Disease Classification And Segmentation | In supervised learning for medical image analysis, sample selection methodologies are fundamental to attain optimum system performance promptly and with minimal expert interactions (e.g. label querying in an active learning setup). In this paper we propose a novel sample selection methodology based on deep features leveraging information contained in interpretability saliency maps. In the absence of ground truth labels for informative samples, we use a novel self supervised learning based approach for training a classifier that learns to identify the most informative sample in a given batch of images. We demonstrate the benefits of the proposed approach, termed Interpretability-Driven Sample Selection (IDEAL), in an active learning setup aimed at lung disease classification and histopathology image segmentation. We analyze three different approaches to determine sample informativeness from interpretability saliency maps: (i) an observational model stemming from findings on previous uncertainty-based sample selection approaches, (ii) a radiomics-based model, and (iii) a novel data-driven self-supervised approach. We compare IDEAL to other baselines using the publicly available NIH chest X-ray dataset for lung disease classification, and a public histopathology segmentation dataset (GLaS), demonstrating the potential of using interpretability information for sample selection in active learning systems. Results show our proposed self supervised approach outperforms other approaches in selecting informative samples leading to state of the art performance with fewer samples. | ['Dwarikanath Mahapatra'] | 2021-04-13 | null | null | null | null | ['lung-disease-classification'] | ['medical'] | [ 8.96436810e-01 5.92650712e-01 -6.22154593e-01 -6.37517512e-01
-1.42432678e+00 -3.24251443e-01 5.73566258e-01 7.70627737e-01
-5.78363597e-01 7.06664503e-01 1.65199697e-01 -1.96280047e-01
-6.95388079e-01 -5.75318038e-01 -5.90603054e-01 -9.31762815e-01
1.80526108e-01 9.51992810e-01 3.65817606e-01 4.90843147e-01
3.26336503e-01 6.31514549e-01 -1.48630571e+00 5.13394952e-01
1.02282917e+00 1.18275654e+00 3.59059185e-01 7.58401453e-01
-7.10267052e-02 1.16578794e+00 -2.19454646e-01 -2.92659663e-02
1.08299419e-01 -5.75569093e-01 -1.30288303e+00 3.59411269e-01
3.62847328e-01 -1.40033767e-01 3.93186688e-01 6.10033214e-01
5.78851044e-01 -9.62524638e-02 8.81137609e-01 -9.12801921e-01
-1.39594287e-01 5.78615725e-01 -1.31718427e-01 3.95301372e-01
1.74446648e-03 4.91786391e-01 1.22809458e+00 -8.61638784e-01
8.29085350e-01 8.72326851e-01 5.62106609e-01 5.68755567e-01
-1.40069652e+00 -1.87877938e-01 -5.85189499e-02 7.95309767e-02
-9.23039615e-01 -6.16596639e-01 7.15854764e-01 -5.39253235e-01
5.13127327e-01 8.28549981e-01 7.97671139e-01 1.08564675e+00
8.88125449e-02 1.22914732e+00 1.40309954e+00 -7.76190042e-01
7.67733991e-01 6.22362971e-01 4.27160680e-01 9.17467356e-01
3.11356813e-01 1.95138589e-01 -5.78787625e-01 -5.45100331e-01
3.68907541e-01 9.84482393e-02 -7.86691755e-02 -6.04547620e-01
-1.40412295e+00 8.83960485e-01 6.73192918e-01 1.12185858e-01
-5.30241847e-01 1.11100256e-01 2.06970304e-01 -3.24176513e-02
8.19390535e-01 7.93032169e-01 -3.30046147e-01 5.20795882e-01
-1.36358309e+00 5.66062257e-02 5.35289347e-01 4.78664219e-01
6.76262558e-01 -4.71834689e-01 -7.76494026e-01 6.84826910e-01
7.66565561e-01 3.40448648e-01 5.44777632e-01 -6.44665956e-01
-2.57520992e-02 9.65350628e-01 -2.70794809e-01 -2.85185188e-01
-4.87581760e-01 -5.31237364e-01 -3.45380545e-01 2.99233019e-01
2.14534551e-01 1.26410469e-01 -1.20573413e+00 1.19886756e+00
5.73215902e-01 -1.50214761e-01 -1.28577843e-01 7.04096854e-01
9.44427609e-01 -8.39974284e-02 3.37088048e-01 -3.07716966e-01
1.23964703e+00 -7.79016495e-01 -4.73652512e-01 -4.36614417e-02
8.24152291e-01 -3.98082107e-01 1.15403652e+00 3.90003115e-01
-7.30345190e-01 -7.42939487e-02 -9.66152430e-01 2.18740538e-01
-2.96204597e-01 3.59375000e-01 7.13791490e-01 7.27438748e-01
-9.58344519e-01 5.12671590e-01 -9.21989202e-01 -4.93430048e-01
1.24847472e+00 6.17829144e-01 -1.58043746e-02 2.06556097e-01
-7.23437250e-01 5.74466169e-01 4.65403497e-01 -2.93980658e-01
-1.10376048e+00 -1.01566410e+00 -5.11083961e-01 -1.83887750e-01
5.78232408e-01 -7.88471162e-01 1.25715292e+00 -1.33777475e+00
-1.29279590e+00 1.13379872e+00 -2.11699922e-02 -7.44180679e-01
7.26395369e-01 -1.10308602e-02 1.36563808e-01 5.91305435e-01
6.95044026e-02 1.02968240e+00 9.51426148e-01 -1.32991636e+00
-5.66134930e-01 -3.00462395e-01 -2.60715038e-01 1.78493902e-01
-2.38543600e-01 -2.06314072e-01 1.42582551e-01 -5.68783581e-01
4.11412157e-02 -1.04482949e+00 -4.49996382e-01 4.56614137e-01
-7.73971975e-01 -2.33269781e-01 8.18682134e-01 -2.25422934e-01
9.64580595e-01 -1.68789339e+00 -1.14528887e-01 4.73144978e-01
5.98998129e-01 1.21043749e-01 2.58903235e-01 1.18328957e-02
-1.51658103e-01 2.66974628e-01 -5.81273139e-01 -1.54509664e-01
-2.36749679e-01 -6.93884864e-02 -4.93702479e-02 5.11656165e-01
8.19640040e-01 1.09183609e+00 -1.08388031e+00 -1.13356495e+00
4.35481995e-01 1.54825017e-01 -5.39917707e-01 3.67961586e-01
-5.01351893e-01 6.63212180e-01 -5.32889068e-01 9.39599633e-01
1.84681714e-01 -5.42673469e-01 -1.58723760e-02 -2.32441872e-01
1.39222756e-01 2.74837852e-01 -7.48081744e-01 1.42551744e+00
-4.24943984e-01 4.20270294e-01 -3.98157507e-01 -8.47537756e-01
6.32718265e-01 5.01739025e-01 7.98682630e-01 -3.73723090e-01
5.33141121e-02 3.80020708e-01 3.82291004e-02 -6.19116247e-01
-6.14474043e-02 -2.00899199e-01 3.07864636e-01 6.84578359e-01
3.09353024e-01 -1.66468546e-01 5.55231469e-03 2.62077391e-01
1.37341738e+00 2.66088936e-02 7.00875282e-01 -7.43634105e-01
4.17286724e-01 3.14929932e-01 2.36230180e-01 9.44452763e-01
-3.88476402e-01 9.12056804e-01 5.37180543e-01 -2.36597210e-01
-7.61202633e-01 -1.08092690e+00 -4.64942425e-01 7.98675954e-01
-2.44195297e-01 4.58270758e-02 -7.89869070e-01 -1.35891080e+00
-1.83725923e-01 7.98782587e-01 -9.51815903e-01 -1.45142198e-01
-3.26473504e-01 -1.01554835e+00 2.35366404e-01 2.32425362e-01
2.18613312e-01 -1.18292236e+00 -1.04657233e+00 -1.53300911e-01
1.14989087e-01 -4.38608259e-01 -7.37492889e-02 8.05341065e-01
-1.16912937e+00 -1.27039814e+00 -5.76396167e-01 -4.34176832e-01
1.14896750e+00 -1.71619609e-01 1.28361845e+00 1.53403595e-01
-7.24103630e-01 7.02200115e-01 -4.43263024e-01 -8.53897870e-01
-5.87908208e-01 3.68316472e-01 -4.60116088e-01 5.39294481e-02
1.45606861e-01 1.09825432e-01 -8.31574619e-01 5.90602681e-02
-1.17218745e+00 1.57216817e-01 1.16852999e+00 1.03933907e+00
1.00224745e+00 -1.99176893e-01 7.94147015e-01 -1.67272174e+00
3.00623149e-01 -6.54907823e-01 -2.57978380e-01 4.83713239e-01
-1.03741372e+00 2.43709385e-01 2.75972933e-01 -1.07666202e-01
-1.14064395e+00 6.00914598e-01 -8.87367725e-02 -8.34509358e-02
-3.20758283e-01 2.06741273e-01 1.03903778e-01 -2.41613045e-01
1.23289096e+00 -1.04703076e-01 2.69517154e-01 -1.03208326e-01
2.80153099e-02 7.69393086e-01 -3.73679101e-02 -3.68291199e-01
4.77560341e-01 7.19996631e-01 1.83196560e-01 -4.71302927e-01
-1.15889454e+00 -7.50157177e-01 -8.57509136e-01 -4.03111041e-01
6.56480193e-01 -5.55790365e-01 -1.69079989e-01 1.14801385e-01
-5.21702826e-01 -2.57766455e-01 -1.05939627e+00 5.02089381e-01
-7.88601756e-01 4.98256758e-02 -1.85492396e-01 -8.76769125e-01
-6.64889872e-01 -1.36675858e+00 1.54673040e+00 1.84100613e-01
-5.34038782e-01 -1.22345388e+00 2.37641394e-01 6.44293547e-01
4.11292404e-01 3.68007004e-01 8.63997459e-01 -1.28962016e+00
-7.98912346e-01 -2.45936826e-01 1.45998644e-02 1.17722213e-01
3.86241227e-01 5.34142926e-02 -1.35445213e+00 -1.26623660e-01
1.39004171e-01 -6.33469105e-01 1.06312299e+00 5.84232032e-01
1.25700617e+00 -4.24841166e-01 -4.53212053e-01 1.70767948e-01
1.44723499e+00 -6.01831684e-03 4.00080323e-01 2.54211932e-01
4.59561974e-01 7.76724160e-01 8.40103984e-01 2.19528183e-01
-1.14875980e-01 3.44305336e-01 4.89573240e-01 -6.05707943e-01
-3.25777948e-01 7.78843537e-02 -2.01537266e-01 2.39925876e-01
3.63252938e-01 -3.51941437e-01 -1.16529644e+00 5.91282666e-01
-1.70098543e+00 -6.63142264e-01 -9.49436706e-03 2.32836890e+00
1.14106977e+00 2.57206142e-01 1.40854090e-01 3.23283076e-01
4.35252160e-01 -1.55454487e-01 -7.43425012e-01 5.61391972e-02
8.70082676e-02 4.47909743e-01 5.09871006e-01 2.44443551e-01
-1.40476298e+00 4.18790758e-01 6.00522327e+00 9.95066762e-01
-1.18359613e+00 1.96763098e-01 1.38634729e+00 -1.80690125e-01
-4.97965127e-01 1.21072322e-01 -5.38876832e-01 2.55350590e-01
8.30421448e-01 2.86284387e-01 -2.30962411e-01 7.88407683e-01
2.94246852e-01 -6.21073127e-01 -1.43162036e+00 5.59445739e-01
7.54977539e-02 -1.65248609e+00 1.24749750e-01 -1.75650660e-02
6.91032290e-01 -4.38070297e-02 2.38438845e-01 -2.32768297e-01
1.27184540e-01 -1.02941084e+00 5.14077187e-01 8.19622338e-01
6.60614967e-01 -3.05969357e-01 7.03907371e-01 1.92921802e-01
-5.26055992e-01 3.69866416e-02 9.62547064e-02 6.37429297e-01
-2.96842396e-01 1.02821112e+00 -1.64087307e+00 4.70543623e-01
4.51902419e-01 4.60656911e-01 -1.08057094e+00 1.25869358e+00
2.04542965e-01 1.16659296e+00 -2.60776043e-01 -1.46284267e-01
4.76805270e-02 3.02943170e-01 7.00195968e-01 1.17848313e+00
-2.18852282e-01 -3.85397464e-01 -4.61698957e-02 1.12600291e+00
9.15849283e-02 2.73068815e-01 -2.50406802e-01 -7.59004280e-02
2.55158275e-01 1.52433085e+00 -1.45721126e+00 -3.66123676e-01
6.31623790e-02 7.07139134e-01 8.69832933e-02 -7.79424608e-02
-6.02131188e-01 -1.71486158e-02 -3.93954456e-01 3.41354549e-01
5.05926572e-02 6.31571472e-01 -5.20252228e-01 -4.99772877e-01
-3.03571224e-01 -7.74032772e-01 6.71546102e-01 -5.62135279e-01
-1.30008125e+00 4.46438611e-01 2.15539798e-01 -1.31222415e+00
-3.25333029e-01 -4.26641732e-01 -5.45181751e-01 5.34784973e-01
-1.32679522e+00 -1.14089787e+00 -3.33073586e-01 -3.83596271e-02
6.45209908e-01 -1.47492051e-01 7.61731505e-01 -5.47226891e-02
-4.06189352e-01 3.74233246e-01 3.58069055e-02 -2.09854513e-01
6.64404750e-01 -1.73873329e+00 -1.21232033e-01 4.87655520e-01
2.45394036e-01 3.87286097e-01 4.48590219e-01 -7.30861783e-01
-1.09858084e+00 -1.27709329e+00 4.15443867e-01 -6.35305703e-01
1.67376757e-01 -2.34062940e-01 -7.30861783e-01 3.86186659e-01
-5.35033420e-02 3.30703735e-01 1.09791732e+00 -2.87648231e-01
3.00114483e-01 -9.62577611e-02 -1.75231099e+00 2.52608776e-01
6.50118232e-01 -1.99641690e-01 -4.43672270e-01 7.15760410e-01
6.11898839e-01 -2.38000765e-01 -7.29677081e-01 6.89732194e-01
3.22503895e-01 -9.90115345e-01 9.03549790e-01 -3.70328546e-01
2.66878515e-01 -1.44512385e-01 1.46967188e-01 -9.88089919e-01
-2.70746499e-01 -1.99508414e-01 4.43711062e-04 1.02922952e+00
7.15806782e-01 -5.78574419e-01 1.12959850e+00 4.67020005e-01
-1.61734521e-01 -1.44613540e+00 -9.44849670e-01 -2.85918415e-01
-2.35814825e-01 -1.44939750e-01 1.79110616e-01 7.52705634e-01
-4.43671823e-01 1.43485516e-01 2.87010401e-01 -1.85454085e-01
7.19848752e-01 -8.52650031e-02 1.33132651e-01 -1.36806500e+00
-3.86306316e-01 -2.91527182e-01 -4.62396890e-01 -2.34863192e-01
-8.12205523e-02 -1.14182365e+00 1.46405816e-01 -1.65242255e+00
6.76558018e-01 -8.01869988e-01 -4.86300528e-01 4.41299498e-01
-4.30411309e-01 2.06115767e-01 -2.70335943e-01 4.66615200e-01
-8.12790751e-01 2.10530654e-01 9.94694889e-01 -2.57211149e-01
-9.00061280e-02 2.24292263e-01 -8.05413127e-01 6.51957989e-01
6.15032494e-01 -7.35392511e-01 -7.29715168e-01 1.98762760e-01
5.15153520e-02 -1.96514085e-01 7.11760998e-01 -1.20101178e+00
5.87436818e-02 -2.19951764e-01 6.95879221e-01 -6.77903652e-01
-9.68339965e-02 -8.58787298e-01 1.21633604e-01 9.16836381e-01
-9.43331718e-01 -7.42899001e-01 -1.69812784e-01 7.93877184e-01
1.68004893e-02 -5.18252313e-01 1.03934836e+00 -2.68861651e-01
-4.57863063e-01 2.07444087e-01 -2.38147438e-01 4.29618508e-02
1.23119199e+00 -3.67261380e-01 -3.69448036e-01 1.85260370e-01
-6.13542616e-01 4.49280720e-03 2.73051888e-01 -5.62776849e-02
5.18966496e-01 -1.05245090e+00 -6.82110190e-01 4.74530831e-02
4.21333820e-01 3.11264187e-01 9.53130424e-02 1.26376367e+00
-6.58168972e-01 4.33954209e-01 2.00162143e-01 -1.15124440e+00
-1.23198593e+00 1.86875284e-01 5.07113099e-01 -5.95179617e-01
-2.84225583e-01 1.00676072e+00 2.57212281e-01 -6.05614781e-01
7.16792867e-02 -4.55169231e-01 -2.53257483e-01 1.43840358e-01
2.48232596e-02 3.62800807e-01 5.50981700e-01 -2.76726931e-01
-4.78841394e-01 3.93203162e-02 -2.13882208e-01 2.15537529e-02
1.25698411e+00 2.50248700e-01 5.23109511e-02 8.03757250e-01
1.09091485e+00 -3.77221592e-02 -1.07682717e+00 -2.74298310e-01
4.52287942e-01 -3.78396362e-01 3.93675715e-01 -1.18901885e+00
-8.83753598e-01 4.81580615e-01 1.18584704e+00 2.24473596e-01
1.16081607e+00 2.62300611e-01 2.42948323e-01 4.12012488e-01
-3.86695052e-03 -1.13572681e+00 3.68301421e-01 -3.25168371e-01
7.54037917e-01 -1.65451288e+00 5.25494277e-01 -5.69869697e-01
-7.58959532e-01 1.05848765e+00 4.22429711e-01 2.10183877e-02
6.20921910e-01 1.60486668e-01 6.58065379e-02 -6.16205513e-01
-1.04992747e+00 -3.48013729e-01 6.29140258e-01 5.02750516e-01
6.49630010e-01 -1.45407533e-03 -1.01995938e-01 2.72417933e-01
2.03512043e-01 3.51869538e-02 1.47975877e-01 1.18533182e+00
-6.04154229e-01 -9.90682542e-01 -3.74319643e-01 1.36568832e+00
-5.29679000e-01 -1.84796806e-02 -7.92451441e-01 6.63569391e-01
9.59027931e-02 6.06144905e-01 4.17851321e-02 4.11663530e-03
-8.34764019e-02 9.77798626e-02 3.53026956e-01 -1.07321167e+00
-8.35741997e-01 -1.22489758e-01 3.67407948e-02 -2.99890161e-01
-7.09802449e-01 -8.27886283e-01 -1.25819218e+00 6.21850252e-01
-7.63732612e-01 -1.68545514e-01 4.74986792e-01 1.02456391e+00
3.14875424e-01 6.90715611e-01 7.72800505e-01 -5.33639014e-01
-4.38150465e-01 -7.30297029e-01 -3.27269912e-01 3.40181530e-01
3.53637636e-01 -5.34972847e-01 -4.46356446e-01 3.75927575e-02] | [14.88312816619873, -2.271135091781616] |
c7fb593b-cc94-4c03-8b83-1c0ac40b3b32 | hierarchical-resnext-models-for-breast-cancer | 1810.09025 | null | http://arxiv.org/abs/1810.09025v1 | http://arxiv.org/pdf/1810.09025v1.pdf | Hierarchical ResNeXt Models for Breast Cancer Histology Image Classification | Microscopic histology image analysis is a cornerstone in early detection of
breast cancer. However these images are very large and manual analysis is error
prone and very time consuming. Thus automating this process is in high demand.
We proposed a hierarchical system of convolutional neural networks (CNN) that
classifies automatically patches of these images into four pathologies: normal,
benign, in situ carcinoma and invasive carcinoma. We evaluated our system on
the BACH challenge dataset of image-wise classification and a small dataset
that we used to extend it. Using a train/test split of 75%/25%, we achieved an
accuracy rate of 0.99 on the test split for the BACH dataset and 0.96 on that
of the extension. On the test of the BACH challenge, we've reached an accuracy
of 0.81 which rank us to the 8th out of 51 teams. | ['Ismaël Koné', 'Lahsen Boulmane'] | 2018-10-21 | null | null | null | null | ['breast-cancer-histology-image-classification'] | ['medical'] | [ 1.00752905e-01 3.82221103e-01 5.05313650e-02 -3.02539229e-01
-8.72938335e-01 -3.12692702e-01 3.37747842e-01 5.43003440e-01
-9.14863765e-01 5.97614229e-01 -2.32065499e-01 -5.71674347e-01
1.58349108e-02 -6.71454966e-01 -4.96590972e-01 -8.35056663e-01
-1.05834939e-01 7.95801878e-01 5.70746481e-01 6.75383732e-02
-6.55294955e-02 5.50095379e-01 -1.25706863e+00 8.45946550e-01
4.66165990e-01 1.20304000e+00 -5.69974296e-02 1.26226687e+00
1.01690747e-01 9.42267179e-01 -5.53555846e-01 -4.09440994e-01
1.62530586e-01 -2.39356264e-01 -1.11095214e+00 -4.35668826e-02
5.27894437e-01 -1.03040911e-01 -7.46494308e-02 8.36795270e-01
4.99251097e-01 -5.65540195e-01 6.86448276e-01 -1.03742909e+00
-1.37118902e-02 4.59905505e-01 -4.62383777e-01 3.92623127e-01
-2.32476577e-01 -2.48288829e-02 5.46663642e-01 -6.24173164e-01
9.16371286e-01 5.71966946e-01 1.07489884e+00 5.33830583e-01
-1.24607038e+00 -6.00158453e-01 -6.67506576e-01 1.81785047e-01
-1.45186412e+00 -4.77241427e-01 4.29137982e-02 -5.97718418e-01
7.12100089e-01 2.89219499e-01 7.09219038e-01 8.18708956e-01
5.74191272e-01 6.21797025e-01 1.15796494e+00 -5.05572557e-01
2.57223397e-01 4.48748231e-01 8.02395418e-02 8.38360071e-01
3.13369632e-01 -1.52720809e-01 4.81463559e-02 -8.71371552e-02
6.04784548e-01 -1.10940069e-01 -6.40059263e-02 -1.49873480e-01
-1.03642845e+00 7.54111826e-01 6.65048599e-01 5.86399496e-01
-3.10143948e-01 -2.41339337e-02 5.34449399e-01 3.80713910e-01
2.26657003e-01 3.03516895e-01 -2.57643700e-01 2.04007477e-01
-1.00149298e+00 -3.60973477e-02 8.22840869e-01 4.00262564e-01
3.06831360e-01 -6.46223247e-01 -8.00912455e-02 9.06756043e-01
-3.04796416e-02 -3.02808695e-02 7.41634190e-01 -6.27141178e-01
-2.10315421e-01 9.26189780e-01 -1.71242535e-01 -6.00429595e-01
-7.98995912e-01 -7.21138179e-01 -1.38929141e+00 3.82378072e-01
6.38182223e-01 5.03969612e-03 -1.26683247e+00 1.21172225e+00
1.26669034e-01 -3.21882069e-01 -6.88288286e-02 5.69564998e-01
9.22565639e-01 2.50534564e-01 1.18678071e-01 -2.30272841e-02
1.45720220e+00 -7.94355631e-01 -3.09928030e-01 1.77752823e-01
9.96928811e-01 -7.20979750e-01 6.58080637e-01 6.57320261e-01
-9.32557523e-01 -4.04404849e-01 -1.12731516e+00 1.67766154e-01
-4.26913351e-01 5.26018858e-01 4.63240445e-01 4.91650581e-01
-1.51724970e+00 5.09815812e-01 -1.11506045e+00 -6.74809396e-01
7.31492996e-01 6.58099592e-01 -9.35749412e-01 -5.21453172e-02
-6.20707273e-01 7.44906783e-01 5.14452875e-01 -7.78531656e-02
-9.15591836e-01 -6.08280838e-01 -3.25206786e-01 -4.51341011e-02
-2.24548250e-01 -4.39463496e-01 1.32418442e+00 -1.10849369e+00
-8.86405587e-01 1.47473538e+00 2.91535735e-01 -7.42136717e-01
7.51343668e-01 3.63131136e-01 -3.44044060e-01 2.56767005e-01
1.07803844e-01 8.88486624e-01 2.26034880e-01 -9.46137607e-01
-1.19454730e+00 -2.11974710e-01 -1.74180165e-01 -3.61067235e-01
-3.15997422e-01 -2.04107031e-01 -4.64699924e-01 -1.30949169e-01
5.86815327e-02 -1.03969109e+00 -2.89328277e-01 1.01481549e-01
-2.75347859e-01 -1.84685454e-01 5.73678732e-01 -7.59163857e-01
8.40730488e-01 -2.16279721e+00 -1.63860902e-01 2.49316022e-01
4.96540934e-01 2.55250961e-01 1.00049615e-01 1.47669464e-01
-3.00440073e-01 2.93405026e-01 7.70344138e-02 -3.87809485e-01
-3.67674619e-01 -8.35142346e-05 5.07751286e-01 5.73386490e-01
2.82672673e-01 5.73723733e-01 -6.43542051e-01 -6.78129613e-01
-8.46405923e-02 4.39942867e-01 -4.17820752e-01 3.17220390e-01
2.05589667e-01 2.04049736e-01 -3.99561822e-02 7.22536445e-01
4.84596848e-01 -5.08844316e-01 1.24431923e-01 -3.89787585e-01
8.34487155e-02 -2.99595654e-01 -8.34881008e-01 1.21404028e+00
-2.73923367e-01 9.07603443e-01 2.53201038e-01 -8.78835440e-01
6.78298533e-01 4.17247593e-01 3.81639093e-01 -5.58853567e-01
3.80393624e-01 3.92198533e-01 5.27198732e-01 -5.08736432e-01
1.82522967e-01 -1.93934873e-01 1.44868746e-01 1.55440867e-01
2.85307765e-01 2.74870664e-01 4.34685022e-01 1.70113295e-01
1.66084588e+00 -5.01172781e-01 3.98443699e-01 -4.90498036e-01
5.43171704e-01 2.00938269e-01 3.48960131e-01 6.44515753e-01
-5.17075360e-01 8.17868233e-01 9.68679249e-01 -1.02472830e+00
-1.19626939e+00 -8.10298026e-01 -3.44963610e-01 6.82844102e-01
-4.56050664e-01 -3.42533410e-01 -7.51914918e-01 -9.07163620e-01
-1.95493877e-01 1.34501293e-01 -1.19391215e+00 -8.38897750e-02
-2.01209679e-01 -9.30643022e-01 6.17130160e-01 4.33262795e-01
4.71989721e-01 -1.02344596e+00 -5.39740622e-01 1.11418419e-01
1.88591648e-02 -9.97704446e-01 1.18361928e-01 5.02066016e-01
-5.96648574e-01 -1.38471711e+00 -8.37794542e-01 -1.13106525e+00
1.04739428e+00 -2.42401823e-01 1.27595687e+00 3.39368463e-01
-7.35208511e-01 -2.23217830e-01 -4.26342875e-01 -6.27930701e-01
-7.56773174e-01 3.92777801e-01 -3.42066735e-01 9.00530145e-02
2.58504003e-01 -2.04446748e-01 -8.47036183e-01 3.18768710e-01
-1.01876259e+00 8.98265392e-02 1.05732810e+00 8.05809081e-01
5.71104944e-01 5.69224693e-02 1.93558440e-01 -1.13369286e+00
2.21453190e-01 -2.61574149e-01 -3.41774464e-01 7.88717493e-02
-2.89761275e-01 -3.48467201e-01 5.63149214e-01 -3.43707293e-01
-4.21301246e-01 3.70350659e-01 -5.19273698e-01 -5.04950620e-02
-5.05830586e-01 3.45914751e-01 3.94600123e-01 -2.76976049e-01
8.45779061e-01 -1.84941188e-01 2.35022843e-01 -2.51718372e-01
-5.06008506e-01 8.36678743e-01 5.07463634e-01 8.96984115e-02
2.90662467e-01 5.25994360e-01 2.27261886e-01 -7.66671121e-01
-6.52154326e-01 -6.45550013e-01 -5.18985569e-01 -3.01604778e-01
9.50611115e-01 -7.36125171e-01 -5.88702619e-01 6.17719650e-01
-8.32849503e-01 -5.86590827e-01 -1.34086579e-01 3.16472024e-01
-1.94192901e-01 -1.70851387e-02 -9.12171781e-01 -2.48879999e-01
-4.81762707e-01 -1.26825643e+00 1.14651060e+00 1.79299936e-01
-4.37565297e-01 -9.01323497e-01 1.59815729e-01 3.38953793e-01
6.91726625e-01 5.90810716e-01 8.87801886e-01 -7.72482097e-01
-1.30347982e-01 -8.53509009e-01 -3.75994235e-01 3.30740035e-01
-6.51160777e-02 2.17843339e-01 -1.06681168e+00 -4.60246950e-01
-3.16248924e-01 -5.76978326e-01 1.08969831e+00 3.73893440e-01
1.29564142e+00 1.49371341e-01 -6.83694422e-01 5.21566927e-01
1.58566904e+00 1.83728352e-01 8.28220546e-01 5.22104502e-01
2.20117047e-01 5.84477484e-01 3.71848106e-01 -6.60206750e-02
8.58597681e-02 3.79677266e-01 4.38599288e-01 -5.72235882e-01
-1.42687663e-01 2.39952609e-01 -1.52001679e-01 3.54214668e-01
-1.77800119e-01 -2.45028451e-01 -1.39000297e+00 6.84865236e-01
-1.44649136e+00 -5.13174057e-01 -3.66374671e-01 2.00516295e+00
6.55460417e-01 4.07051772e-01 7.25552738e-02 3.18088979e-01
5.87033808e-01 -5.83206534e-01 1.55992001e-01 -4.18619514e-01
2.50906974e-01 3.73436749e-01 5.03733099e-01 2.44575024e-01
-1.40669811e+00 4.39836055e-01 6.19604063e+00 8.56100798e-01
-1.45373964e+00 2.61234883e-02 1.18707693e+00 1.70210928e-01
5.90564966e-01 -4.09234732e-01 -6.80343688e-01 3.31967294e-01
1.28578675e+00 2.89051354e-01 -4.55099225e-01 6.86360776e-01
-1.83651865e-01 -5.00190973e-01 -1.09661293e+00 6.89653158e-01
-4.88568582e-02 -1.39271486e+00 -9.67763737e-02 3.50959003e-01
6.49174154e-01 2.08078086e-01 -2.17278913e-01 3.82294297e-01
1.46411881e-01 -1.36752844e+00 2.55749762e-01 4.62710202e-01
9.44167554e-01 -6.58391058e-01 1.71341860e+00 4.54012364e-01
-7.69395292e-01 3.83773223e-02 -2.52918512e-01 2.56272912e-01
-4.55542058e-01 5.34274817e-01 -1.45346820e+00 1.57411039e-01
9.36479926e-01 3.34983617e-01 -9.88228977e-01 1.31354260e+00
3.26340705e-01 5.68120062e-01 -3.00195575e-01 6.95450380e-02
2.47202441e-01 5.18720984e-01 -3.90513637e-03 1.49658549e+00
2.84240484e-01 -2.09155992e-01 -7.60381343e-03 8.74533951e-02
-1.85233742e-01 1.19909093e-01 -3.18781734e-01 1.58669472e-01
-8.82641003e-02 1.75655019e+00 -1.31457949e+00 -3.69376183e-01
2.13351324e-02 6.82574987e-01 3.98107082e-01 -1.74981982e-01
-5.60690999e-01 -4.87276196e-01 -1.13519453e-01 2.70911098e-01
2.58897126e-01 3.74402165e-01 1.22237559e-02 -6.32847846e-01
-2.96673626e-01 -8.02592695e-01 5.93144655e-01 -5.45678198e-01
-1.12348127e+00 9.64439392e-01 -3.45109940e-01 -1.11121011e+00
-6.59191683e-02 -9.60770071e-01 -6.13636613e-01 5.50096095e-01
-1.01679087e+00 -1.13575375e+00 -7.44511664e-01 1.91696361e-01
2.12743223e-01 -1.89625084e-01 1.03350914e+00 4.77527529e-01
-4.43019301e-01 7.12711573e-01 1.58933401e-01 5.11800349e-01
7.45911241e-01 -1.39372981e+00 -1.50952995e-01 3.12157750e-01
-3.87430787e-01 1.85729682e-01 4.54785705e-01 -1.50331452e-01
-7.60661900e-01 -1.28389311e+00 1.04152727e+00 -2.73027599e-01
5.98536611e-01 -2.71317393e-01 -8.18116009e-01 4.88867760e-01
2.38464102e-01 1.78644955e-01 1.00565255e+00 -1.68352544e-01
-3.00714094e-03 -2.13064060e-01 -1.31933773e+00 2.50054389e-01
3.13346088e-01 -1.06181905e-01 -1.05910329e-02 4.62812245e-01
6.80624545e-02 -4.81069207e-01 -1.17022908e+00 6.38789117e-01
7.32512891e-01 -1.19227982e+00 5.88570178e-01 -3.28103095e-01
6.78394556e-01 -3.44490670e-02 6.02758676e-02 -1.07593811e+00
-4.54576045e-01 1.59368813e-01 5.12069285e-01 9.44735765e-01
6.65443361e-01 -3.97254497e-01 1.15596473e+00 9.53943431e-02
-5.46936169e-02 -1.05914068e+00 -9.14927900e-01 -4.97149438e-01
2.82404840e-01 -7.45073855e-02 1.78278852e-02 7.85519063e-01
-2.72445172e-01 1.15395933e-01 3.19925070e-01 -1.80920854e-01
4.45455730e-01 -3.29851061e-01 7.65859246e-01 -1.27099240e+00
-1.43751055e-01 -7.10684359e-01 -9.75878656e-01 1.53245017e-01
-3.70804787e-01 -9.07720387e-01 5.88977672e-02 -1.57549322e+00
6.93073928e-01 -2.89833009e-01 -4.73591775e-01 6.43324852e-01
1.02179162e-01 9.27003324e-01 -3.34430963e-01 3.24362576e-01
-5.47173619e-01 -4.35508907e-01 8.57231081e-01 -2.39323676e-01
1.83882967e-01 1.84173603e-02 -5.66249847e-01 6.92780137e-01
9.90682662e-01 -3.55022550e-01 1.25966668e-01 -3.97293968e-03
1.56925008e-01 -9.32822227e-02 4.10792440e-01 -1.32171679e+00
2.07771972e-01 2.88079917e-01 8.80600870e-01 -5.72814524e-01
5.32290600e-02 -7.70665348e-01 4.23838228e-01 9.82573569e-01
-5.10080874e-01 -2.01133639e-01 3.59757304e-01 2.35964432e-01
-3.81574452e-01 -1.58601254e-01 1.04181170e+00 -2.78800040e-01
-4.11170453e-01 2.51700133e-01 -4.18344498e-01 -4.26611453e-01
1.13370943e+00 -1.22071698e-01 -2.39959344e-01 -1.04589798e-01
-8.83795977e-01 1.51310265e-01 5.12908340e-01 3.02833989e-02
3.72508794e-01 -1.08795571e+00 -1.02682185e+00 1.35821804e-01
4.17344093e-01 6.87051117e-02 2.59213418e-01 1.29901576e+00
-1.16881561e+00 3.83901417e-01 -3.99867326e-01 -1.01960683e+00
-1.80834615e+00 2.93141119e-02 7.45304286e-01 -8.01847517e-01
-3.17855477e-01 1.01578438e+00 1.14349440e-01 -3.59347045e-01
2.92121172e-01 -2.73489445e-01 -5.34830272e-01 7.11428840e-03
5.89467049e-01 4.64740731e-02 5.52646220e-01 -4.80591804e-01
-2.43162170e-01 2.79932290e-01 -4.80751336e-01 1.45614803e-01
1.34994936e+00 5.82637548e-01 -4.27723974e-01 2.04660460e-01
1.54321706e+00 -3.32722366e-01 -5.99367142e-01 2.15280578e-01
4.10438403e-02 -7.94656128e-02 2.55845279e-01 -8.28977585e-01
-9.90076065e-01 8.68972242e-01 1.04727852e+00 5.64110339e-01
1.17456770e+00 5.62471375e-02 4.49390888e-01 4.64161038e-01
1.25070021e-01 -8.29140425e-01 -2.61373520e-01 3.55161160e-01
8.42206240e-01 -1.44548380e+00 9.82570276e-03 -2.94396013e-01
-1.71853527e-01 1.34510148e+00 5.55412710e-01 -2.82744944e-01
5.87824583e-01 4.67395812e-01 2.50937968e-01 -3.44108254e-01
-1.05265701e+00 3.41590419e-02 2.40601629e-01 5.57153881e-01
7.45687425e-01 1.14958882e-01 -3.40248704e-01 3.17871422e-01
-2.34829515e-01 3.28680933e-01 4.57589239e-01 1.11133158e+00
-4.64511484e-01 -8.07430267e-01 -2.67157763e-01 8.41872811e-01
-9.70728278e-01 3.10211688e-01 -7.08297670e-01 1.23896945e+00
3.22577566e-01 5.47601938e-01 2.02213198e-01 -5.40740013e-01
2.49519825e-01 -7.26124272e-02 3.13113749e-01 -4.52487439e-01
-8.42935562e-01 7.34224841e-02 2.66559482e-01 -4.04533654e-01
-2.59097934e-01 -4.09445077e-01 -9.55395758e-01 -1.85218722e-01
-2.09079564e-01 2.52667785e-01 7.29874253e-01 5.98917365e-01
1.30608380e-02 7.59381890e-01 4.23295856e-01 -6.07767284e-01
-2.65767962e-01 -1.21997857e+00 -4.61961091e-01 4.70533371e-01
4.77141649e-01 -2.41243184e-01 -3.95947844e-01 3.20865065e-01] | [15.138184547424316, -2.9679644107818604] |
9e9dd344-6366-483e-9ede-0d55ca168d68 | generating-distractors-for-reading | 1809.02768 | null | http://arxiv.org/abs/1809.02768v2 | http://arxiv.org/pdf/1809.02768v2.pdf | Generating Distractors for Reading Comprehension Questions from Real Examinations | We investigate the task of distractor generation for multiple choice reading
comprehension questions from examinations. In contrast to all previous works,
we do not aim at preparing words or short phrases distractors, instead, we
endeavor to generate longer and semantic-rich distractors which are closer to
distractors in real reading comprehension from examinations. Taking a reading
comprehension article, a pair of question and its correct option as input, our
goal is to generate several distractors which are somehow related to the
answer, consistent with the semantic context of the question and have some
trace in the article. We propose a hierarchical encoder-decoder framework with
static and dynamic attention mechanisms to tackle this task. Specifically, the
dynamic attention can combine sentence-level and word-level attention varying
at each recurrent time step to generate a more readable sequence. The static
attention is to modulate the dynamic attention not to focus on question
irrelevant sentences or sentences which contribute to the correct option. Our
proposed framework outperforms several strong baselines on the first prepared
distractor generation dataset of real reading comprehension questions. For
human evaluation, compared with those distractors generated by baselines, our
generated distractors are more functional to confuse the annotators. | ['Yifan Gao', 'Piji Li', 'Michael R. Lyu', 'Irwin King', 'Lidong Bing'] | 2018-09-08 | null | null | null | null | ['distractor-generation'] | ['natural-language-processing'] | [ 7.06064820e-01 6.48409545e-01 3.52198094e-01 -1.72385126e-01
-1.10500693e+00 -5.36458075e-01 4.95848715e-01 2.40205526e-01
-5.87416828e-01 8.08464348e-01 8.53329360e-01 -6.10294402e-01
1.53524727e-01 -6.75027132e-01 -5.99201381e-01 -3.43641639e-01
9.35793042e-01 4.00200307e-01 6.09193802e-01 -6.43662333e-01
6.57434165e-01 -2.43031859e-01 -1.48596406e+00 5.50932288e-01
1.33871579e+00 3.99184465e-01 8.19907665e-01 8.96012783e-01
-4.45840746e-01 1.05757046e+00 -1.14360392e+00 -3.99286300e-01
-3.11796039e-01 -1.28195119e+00 -1.25495613e+00 -8.27451423e-03
7.08234251e-01 -2.87864923e-01 -2.69093156e-01 1.26480305e+00
8.97632301e-01 4.27478313e-01 8.49596739e-01 -3.95381570e-01
-1.27265525e+00 9.48921502e-01 -3.58747035e-01 9.40784872e-01
9.80705857e-01 3.46593261e-01 8.06222916e-01 -8.50427508e-01
3.56516808e-01 1.59248722e+00 -1.90686658e-01 1.10118377e+00
-8.38125527e-01 -2.65658408e-01 4.94733185e-01 7.04696596e-01
-1.08661544e+00 -4.97457683e-01 5.52415311e-01 -2.29941741e-01
9.70326126e-01 3.82886410e-01 2.13404089e-01 1.69466686e+00
5.55234730e-01 7.80071437e-01 8.58063221e-01 -6.92558885e-01
4.93891314e-02 -1.62271023e-01 6.43547177e-01 3.16822052e-01
8.43074545e-02 -1.99466571e-01 -2.88871706e-01 3.54352355e-01
3.26377034e-01 -3.55195433e-01 -8.38539720e-01 5.90527058e-01
-1.38189709e+00 7.67864048e-01 8.88276473e-02 1.27635285e-01
-4.22830522e-01 -3.20726871e-01 9.56764668e-02 3.93630207e-01
-9.74213108e-02 6.23160660e-01 -3.19555789e-01 -3.42982709e-01
-5.40821373e-01 3.86289388e-01 7.48120248e-01 1.15039289e+00
1.97969615e-01 -8.56911540e-02 -1.33202612e+00 7.62219071e-01
7.65071213e-02 6.79096401e-01 9.11187291e-01 -7.10854113e-01
9.27666187e-01 4.07210290e-01 1.72084734e-01 -9.32046413e-01
-3.00390989e-01 -5.75527072e-01 -7.32535362e-01 -3.44413698e-01
5.31058371e-01 -1.39676899e-01 -7.74483442e-01 1.62369156e+00
5.87052554e-02 -2.21998259e-01 2.39747852e-01 1.00241244e+00
1.36763644e+00 7.09493339e-01 -7.61759207e-02 -2.26966530e-01
1.95063150e+00 -1.52651620e+00 -1.35156155e+00 -5.43163717e-01
4.50438172e-01 -1.01563334e+00 1.71320093e+00 4.04714257e-01
-1.42355275e+00 -9.05657232e-01 -1.00895691e+00 -5.36299527e-01
1.00151360e-01 -1.49425760e-01 -5.63815057e-01 3.93933952e-02
-7.06195354e-01 1.03760641e-02 -2.06207305e-01 8.55358914e-02
1.91451147e-01 -3.40389490e-01 2.17395529e-01 -2.22398520e-01
-1.61305451e+00 1.23002505e+00 1.25289664e-01 1.97953835e-01
-9.32610691e-01 -3.26605201e-01 -1.00771511e+00 3.44700217e-01
6.81062460e-01 -9.15378451e-01 1.79560840e+00 -1.02759421e+00
-1.50897789e+00 6.61433876e-01 -4.40776378e-01 -1.79238394e-01
2.56746799e-01 -4.14688170e-01 -4.32853550e-01 2.20036939e-01
3.78141224e-01 5.07865250e-01 8.99030983e-01 -7.56547093e-01
-2.82014877e-01 4.70716022e-02 1.76225066e-01 5.89773953e-01
1.78228900e-01 -1.34429887e-01 -1.55545086e-01 -8.76764894e-01
-1.57291777e-02 -5.81141889e-01 -5.38994297e-02 -5.94201624e-01
-6.14983678e-01 -5.46941817e-01 1.11512318e-01 -9.00488138e-01
1.58519161e+00 -1.94187784e+00 4.70819175e-01 -4.27684635e-01
2.08143845e-01 2.70607263e-01 -5.33089101e-01 3.59746665e-01
-1.18720964e-01 1.26238674e-01 -1.47570312e-01 1.19662456e-01
-2.38378104e-02 8.26288834e-02 -3.69591296e-01 -8.47210139e-02
4.67174053e-01 1.08836675e+00 -1.20657372e+00 -5.81558466e-01
-1.90990105e-01 -2.02694103e-01 -4.13698554e-01 5.66036165e-01
-5.34983814e-01 4.84203190e-01 -4.75123018e-01 2.56921291e-01
5.86358607e-01 -3.88072222e-01 -3.83460730e-01 1.24742687e-01
3.50404829e-01 9.89556670e-01 -6.83345258e-01 1.75438631e+00
-3.13860208e-01 2.28083283e-01 -2.80677408e-01 -5.96433640e-01
8.41017604e-01 1.66959807e-01 -6.54145300e-01 -1.22930729e+00
1.75039113e-01 -8.11746567e-02 6.35017335e-01 -1.21411192e+00
6.44813359e-01 -1.50375292e-01 -9.92387161e-02 4.81366187e-01
9.18236077e-02 7.22866878e-03 3.49338502e-01 4.61877376e-01
1.14755297e+00 1.90413982e-01 5.23465455e-01 -2.74457812e-01
1.17497575e+00 -1.86676264e-01 1.47467911e-01 1.05621135e+00
-2.06900164e-01 1.00533259e+00 7.71204352e-01 1.88818619e-01
-6.46452069e-01 -1.07887149e+00 2.14854807e-01 1.13553059e+00
3.96841258e-01 -2.92886674e-01 -8.68491888e-01 -8.98041904e-01
-6.16280377e-01 1.38962412e+00 -6.19165361e-01 -4.48138505e-01
-8.33856404e-01 -2.52724618e-01 4.32003260e-01 4.00342911e-01
3.86328340e-01 -1.41546142e+00 -4.05774921e-01 4.10307139e-01
-7.27745533e-01 -8.35038066e-01 -1.25580907e+00 -1.86930045e-01
-3.91222358e-01 -1.16864932e+00 -7.68662453e-01 -1.01999319e+00
7.39058673e-01 2.87169695e-01 1.61879516e+00 2.17970237e-01
4.33221981e-02 1.50791630e-01 -6.46971226e-01 -5.18698156e-01
-6.66661203e-01 2.69217610e-01 -7.22307742e-01 -3.14663976e-01
6.97729945e-01 -1.91129923e-01 -7.73446679e-01 7.18864053e-02
-9.90577340e-01 2.05580086e-01 8.01903605e-01 9.95809078e-01
4.12910968e-01 -7.80765951e-01 8.69625449e-01 -8.46351147e-01
1.46044898e+00 -8.24767590e-01 -1.41915485e-01 4.85837609e-01
-2.41324499e-01 3.86895180e-01 8.39241326e-01 -6.07402146e-01
-1.08203483e+00 -5.46384931e-01 -5.20393133e-01 -6.98482478e-03
-2.44372860e-01 3.87434393e-01 -3.03348243e-01 8.02822709e-01
9.15247262e-01 7.63283372e-01 1.98275238e-01 -3.77462298e-01
4.57413614e-01 4.87807453e-01 3.89704674e-01 -5.81968307e-01
4.75564152e-01 -5.39143503e-01 -3.53560239e-01 -4.11370218e-01
-1.39973474e+00 -1.36397585e-01 -3.21139842e-01 -1.92764089e-01
1.07555330e+00 -6.15031779e-01 -6.94143891e-01 1.60450161e-01
-1.60367477e+00 7.24964291e-02 -3.63475710e-01 1.69705972e-01
-3.27529132e-01 5.65199792e-01 -5.09393990e-01 -6.12324297e-01
-4.51401204e-01 -1.30994725e+00 1.08032918e+00 4.64544415e-01
-4.60312426e-01 -8.28772902e-01 -7.88651928e-02 6.11647129e-01
4.49377179e-01 -1.36714742e-01 1.18443716e+00 -1.05006218e+00
-4.29033756e-01 2.50168204e-01 -1.29571501e-02 5.08865356e-01
2.79892117e-01 -4.53494936e-01 -5.14687896e-01 1.03823684e-01
4.42961216e-01 -5.22968590e-01 7.23219454e-01 -1.30848903e-02
1.29573119e+00 -8.41514528e-01 7.13421851e-02 -1.17183998e-01
9.04427767e-01 2.14086652e-01 9.59527373e-01 -2.44269427e-02
5.05964577e-01 7.91262567e-01 6.48870111e-01 -1.18037134e-01
7.40437865e-01 3.91082674e-01 2.13504270e-01 2.03423187e-01
-5.28025985e-01 -3.80040199e-01 4.24827993e-01 1.46529675e+00
3.66413265e-01 -5.96168280e-01 -7.37296462e-01 6.57584190e-01
-1.53668618e+00 -1.23229706e+00 -6.78493202e-01 1.95523536e+00
1.20126212e+00 3.35918844e-01 -3.04122031e-01 -2.39227355e-01
6.88803673e-01 9.98515636e-02 -4.72139686e-01 -6.04023039e-01
-2.76574463e-01 5.51899254e-01 -2.18087316e-01 8.58565390e-01
-3.11904788e-01 7.27978468e-01 5.76425409e+00 9.32465971e-01
-6.64668143e-01 2.02629253e-01 4.65422630e-01 -2.39202648e-01
-7.42023349e-01 -2.31704459e-01 -9.09916461e-01 7.94977367e-01
1.02428377e+00 -4.29176986e-01 2.50374287e-01 2.77915508e-01
4.96207863e-01 -2.82646269e-01 -1.10152793e+00 5.23996890e-01
6.68735027e-01 -9.28003073e-01 4.92224604e-01 -4.97473806e-01
5.50918818e-01 -6.30009174e-01 -1.15160951e-02 6.64492965e-01
-5.42909466e-02 -1.39029670e+00 7.14536071e-01 1.06312048e+00
3.79556417e-01 -5.97660780e-01 8.45919609e-01 7.76675582e-01
-5.37229359e-01 1.75320968e-01 -3.96458119e-01 -2.33116910e-01
2.35654637e-01 2.52194405e-01 -6.09577358e-01 5.56111038e-01
2.21594006e-01 3.28676492e-01 -8.99157226e-01 5.41267693e-01
-8.28844726e-01 6.82175159e-01 3.51954579e-01 -6.07787728e-01
2.88234055e-01 -7.70518333e-02 5.12676597e-01 1.03048050e+00
3.20883095e-01 4.57306862e-01 -1.30846635e-01 9.57353890e-01
-1.87986329e-01 3.61430585e-01 -3.69606763e-01 2.66252428e-01
4.90842968e-01 7.89644122e-01 -5.53216226e-02 -5.49145758e-01
-5.39774299e-01 1.19107521e+00 4.53079402e-01 3.43963265e-01
-9.36894715e-01 -7.51782596e-01 2.25688428e-01 9.94411260e-02
2.17520833e-01 2.10640550e-01 -1.04600161e-01 -1.28950405e+00
3.64317030e-01 -1.56221652e+00 3.33216071e-01 -1.05290461e+00
-1.54150379e+00 8.72240067e-01 -2.14007437e-01 -1.19952166e+00
-3.06534052e-01 -2.70000935e-01 -7.84473240e-01 1.48328388e+00
-1.64878750e+00 -5.29887557e-01 -3.96339744e-01 5.05651832e-01
1.29507840e+00 1.11205839e-01 6.71227455e-01 5.18333837e-02
-3.43619734e-01 6.71038449e-01 -4.62499082e-01 -7.16966912e-02
7.72326469e-01 -1.26500797e+00 2.01198354e-01 1.10266423e+00
-2.11760581e-01 8.03033829e-01 8.51270556e-01 -7.37746119e-01
-1.03621137e+00 -8.57236683e-01 1.33518815e+00 -7.87427485e-01
4.62050229e-01 -2.26221651e-01 -1.43123662e+00 4.28455532e-01
9.73694026e-01 -6.37009978e-01 6.05163336e-01 -3.49063337e-01
4.46651429e-02 2.36025825e-01 -7.09671497e-01 9.12847757e-01
1.03441048e+00 -2.53682107e-01 -1.68292856e+00 4.08559918e-01
1.14552832e+00 -7.47960389e-01 -2.29992256e-01 1.07302316e-01
-7.66657144e-02 -7.20459104e-01 5.57388902e-01 -8.28469753e-01
1.09550846e+00 -3.82916182e-01 3.63769352e-01 -1.35780895e+00
-5.53849995e-01 -5.23047924e-01 -2.50221372e-01 1.13547027e+00
5.50927162e-01 -2.92322874e-01 6.88999221e-02 2.01592684e-01
-6.37439251e-01 -7.02699244e-01 -9.28270221e-01 -4.98030186e-01
5.65457106e-01 1.66188538e-01 5.29996514e-01 3.55074108e-01
9.94564816e-02 1.15416670e+00 -1.56665877e-01 -1.32671952e-01
1.85773551e-01 -4.21951748e-02 3.60558897e-01 -5.89723349e-01
-2.50907868e-01 -5.86367428e-01 3.17606956e-01 -1.88515198e+00
2.05479175e-01 -8.91869068e-01 4.81900066e-01 -1.65491927e+00
1.39575809e-01 4.02970761e-01 3.92434299e-02 1.13949239e-01
-1.13900197e+00 -3.45296472e-01 4.23310772e-02 -1.48828521e-01
-6.34335697e-01 8.44697952e-01 2.20942235e+00 -2.91821450e-01
7.12679550e-02 -8.87068287e-02 -1.08592725e+00 2.66831756e-01
4.35792863e-01 -3.35596740e-01 -7.38438189e-01 -6.51813924e-01
2.19716430e-01 5.44977784e-01 1.28272727e-01 -5.85907102e-01
3.28670263e-01 -3.03016156e-01 2.07473069e-01 -7.33522296e-01
-1.75327376e-01 -3.14688027e-01 -5.38432121e-01 4.07115579e-01
-1.02132225e+00 6.21061325e-01 1.41051576e-01 5.27559221e-01
-2.91659057e-01 -9.08180475e-01 5.70017874e-01 -5.02149999e-01
-2.24017173e-01 -2.22899467e-01 -9.11169291e-01 7.67519414e-01
7.73913026e-01 -5.34011517e-04 -7.59455681e-01 -6.49195433e-01
-8.07565629e-01 6.16175294e-01 -3.75985168e-02 8.07871461e-01
8.88313770e-01 -1.10347784e+00 -1.30754995e+00 -5.40197594e-03
3.30235213e-01 2.67515451e-01 4.64969546e-01 5.97083330e-01
-4.12160665e-01 5.61529756e-01 3.06997150e-02 -2.87858784e-01
-1.03190148e+00 8.73071551e-01 4.07232970e-01 -2.05700651e-01
-5.83270431e-01 1.01680934e+00 3.19790781e-01 -2.33179834e-02
2.13276565e-01 -4.46051091e-01 -7.97673047e-01 9.84567851e-02
9.60704923e-01 1.01958267e-01 9.97126922e-02 -2.30674028e-01
-6.19987175e-02 1.40815407e-01 -3.52049053e-01 -1.08953088e-03
6.22244418e-01 -5.58008671e-01 -1.17573954e-01 5.77286541e-01
6.49444282e-01 2.10913002e-01 -7.82497764e-01 -1.28188476e-01
8.50118101e-02 -2.02108338e-01 -4.93067831e-01 -1.20526743e+00
-5.37509263e-01 1.03470039e+00 2.17751097e-02 2.00376034e-01
1.10104954e+00 1.55984953e-01 8.81409228e-01 3.25741172e-01
-2.31274515e-01 -9.14302647e-01 6.44333661e-01 9.28825378e-01
1.37475395e+00 -1.00584006e+00 -4.35719341e-01 -3.81608874e-01
-8.05855751e-01 1.23705029e+00 1.40201843e+00 1.87399182e-02
-8.86163786e-02 -1.47007123e-01 1.78108513e-01 -8.15613344e-02
-1.19840455e+00 -2.11705089e-01 4.71155882e-01 4.38046545e-01
7.51059830e-01 -4.08160657e-01 -9.70142663e-01 8.88194978e-01
-5.60824692e-01 -3.73817384e-01 1.01054966e+00 6.71408772e-01
-6.13535881e-01 -1.00709176e+00 -3.35558981e-01 6.25780404e-01
-3.99506509e-01 -7.39714801e-01 -3.17253679e-01 2.67380089e-01
-2.46820971e-02 1.16780365e+00 1.22242406e-01 1.91686109e-01
6.93347335e-01 2.58005172e-01 5.67098439e-01 -1.09988415e+00
-9.46818054e-01 -9.75470394e-02 1.65185168e-01 -2.47341767e-01
-8.33420977e-02 -6.60159469e-01 -1.11583805e+00 1.33285820e-01
-3.94479692e-01 3.06880593e-01 -1.15836104e-02 1.21687055e+00
2.79414386e-01 1.22492766e+00 4.23499227e-01 7.20153078e-02
-1.09238422e+00 -1.55282903e+00 8.16561431e-02 5.72769403e-01
5.99640012e-01 -2.55746692e-01 -3.75205308e-01 8.85589607e-03] | [11.546570777893066, 8.222095489501953] |
c39772f9-31a6-4969-addd-18e58f652103 | iterative-geometry-encoding-volume-for-stereo | 2303.06615 | null | https://arxiv.org/abs/2303.06615v2 | https://arxiv.org/pdf/2303.06615v2.pdf | Iterative Geometry Encoding Volume for Stereo Matching | Recurrent All-Pairs Field Transforms (RAFT) has shown great potentials in matching tasks. However, all-pairs correlations lack non-local geometry knowledge and have difficulties tackling local ambiguities in ill-posed regions. In this paper, we propose Iterative Geometry Encoding Volume (IGEV-Stereo), a new deep network architecture for stereo matching. The proposed IGEV-Stereo builds a combined geometry encoding volume that encodes geometry and context information as well as local matching details, and iteratively indexes it to update the disparity map. To speed up the convergence, we exploit GEV to regress an accurate starting point for ConvGRUs iterations. Our IGEV-Stereo ranks $1^{st}$ on KITTI 2015 and 2012 (Reflective) among all published methods and is the fastest among the top 10 methods. In addition, IGEV-Stereo has strong cross-dataset generalization as well as high inference efficiency. We also extend our IGEV to multi-view stereo (MVS), i.e. IGEV-MVS, which achieves competitive accuracy on DTU benchmark. Code is available at https://github.com/gangweiX/IGEV. | ['Xin Yang', 'Xiaohuan Ding', 'Xianqi Wang', 'Gangwei Xu'] | 2023-03-12 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xu_Iterative_Geometry_Encoding_Volume_for_Stereo_Matching_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xu_Iterative_Geometry_Encoding_Volume_for_Stereo_Matching_CVPR_2023_paper.pdf | cvpr-2023-1 | ['stereo-matching-1'] | ['computer-vision'] | [-1.94926992e-01 -2.61696905e-01 -1.39475977e-02 -6.08932376e-01
-9.54840958e-01 -5.14528394e-01 5.07060170e-01 -1.04827024e-01
-1.84365332e-01 4.69590575e-01 4.59832132e-01 -1.23249799e-01
-2.69809127e-01 -1.01260138e+00 -9.77176547e-01 -3.46050411e-01
1.56758919e-01 6.19434237e-01 2.55972654e-01 -3.48620713e-01
2.71372467e-01 3.66545677e-01 -1.40024126e+00 4.58208054e-01
9.96627927e-01 1.11975741e+00 3.46477777e-01 2.81231433e-01
7.65598267e-02 4.69279855e-01 1.77891746e-01 -7.06985116e-01
6.31128430e-01 -1.89232994e-02 -7.55742788e-01 -4.09248173e-01
1.33893359e+00 -5.91289520e-01 -8.08817327e-01 9.19672310e-01
7.39027381e-01 1.53966680e-01 4.43356633e-01 -9.18870568e-01
-4.73805696e-01 4.54569250e-01 -8.28718066e-01 3.59540433e-01
2.21938699e-01 3.84871751e-01 1.22348452e+00 -1.29704618e+00
8.68658066e-01 1.30908024e+00 9.26547647e-01 1.99933723e-01
-1.07523930e+00 -8.85929763e-01 7.86029473e-02 3.86666805e-01
-1.40258825e+00 -2.81997383e-01 8.54586899e-01 -4.96481895e-01
1.09495032e+00 1.95381373e-01 7.83757329e-01 9.28500175e-01
2.30894417e-01 8.47727597e-01 1.04095411e+00 1.53648434e-02
-1.40797704e-01 -6.45451963e-01 9.12059173e-02 7.32730508e-01
3.58109251e-02 5.82354128e-01 -7.32957780e-01 5.93392551e-03
1.05201256e+00 -8.30989098e-04 -2.64547646e-01 -4.46262807e-01
-1.42451668e+00 8.28096509e-01 1.04499352e+00 7.76492385e-03
-2.84790605e-01 3.67370218e-01 3.53758991e-01 2.05461740e-01
5.59626341e-01 4.73001093e-01 -2.70865023e-01 -5.34638949e-02
-9.14212644e-01 5.20115852e-01 1.40394717e-01 7.45640755e-01
1.02092433e+00 -1.13954015e-01 -3.28328043e-01 1.11598909e+00
-2.55321041e-02 6.98917389e-01 6.59721997e-03 -1.16911018e+00
1.02982771e+00 5.11004448e-01 -2.56654680e-01 -1.22213411e+00
-5.03224254e-01 -6.31539047e-01 -1.06714785e+00 -8.19915012e-02
3.17688107e-01 2.15900421e-01 -1.02110517e+00 1.75892639e+00
3.34585220e-01 3.62799197e-01 -3.43018889e-01 1.24790895e+00
1.35731876e+00 3.87454480e-01 -3.26282620e-01 5.07846773e-01
9.96633172e-01 -1.00456882e+00 -1.23536743e-01 -4.32925045e-01
4.36662853e-01 -8.25786591e-01 8.86434317e-01 1.08140282e-01
-1.11480260e+00 -6.46624148e-01 -8.76525640e-01 -5.71327031e-01
-9.40764844e-02 1.14823096e-02 9.74489212e-01 -9.72542679e-04
-1.32457149e+00 7.41725206e-01 -6.57095253e-01 -9.10947472e-02
6.13065839e-01 4.10908908e-01 -4.58624840e-01 -4.72548485e-01
-1.35552096e+00 4.82269257e-01 1.34400874e-01 2.89447457e-01
-4.53057647e-01 -1.01091266e+00 -1.16725504e+00 -1.32771537e-01
3.72877061e-01 -1.06834042e+00 1.08030975e+00 -4.08208787e-01
-1.07520056e+00 1.05726624e+00 -3.34373564e-01 -3.25127602e-01
7.86892951e-01 -2.09903523e-01 1.08105121e-02 3.03654280e-02
3.39831233e-01 1.17250574e+00 3.64941329e-01 -9.77215886e-01
-5.12226462e-01 -5.28507292e-01 1.10360362e-01 2.98479199e-01
1.63195282e-01 -5.22146821e-01 -7.04313278e-01 -7.41956115e-01
6.76154315e-01 -7.73088992e-01 -1.81432918e-01 9.67348814e-02
-4.82311696e-01 -1.69544458e-01 2.75123864e-01 -7.80888617e-01
1.05821669e+00 -2.07577634e+00 2.50949442e-01 4.09442097e-01
5.15759647e-01 3.42241451e-02 -3.37914139e-01 3.19547921e-01
-4.53629829e-02 -2.67937094e-01 -2.18136653e-01 -3.73217702e-01
-1.08792791e-02 9.54139531e-02 -2.08831877e-01 5.23265839e-01
8.69761314e-03 1.34779000e+00 -8.44326973e-01 -4.40900296e-01
4.94605005e-01 4.78213578e-01 -9.89700496e-01 -7.35182017e-02
-4.93596867e-02 5.85478723e-01 -4.56554919e-01 6.96183622e-01
1.12566710e+00 -4.18582112e-01 -2.51855880e-01 -6.61984861e-01
-2.13806614e-01 3.26156497e-01 -1.13793838e+00 2.16312838e+00
-4.08443749e-01 7.10155368e-01 -2.18267247e-01 -7.98711419e-01
9.00030673e-01 -2.88099885e-01 5.75774491e-01 -1.19513500e+00
8.62924606e-02 4.22301918e-01 -1.23649217e-01 -8.08759406e-02
5.58807611e-01 2.18428165e-01 1.39708593e-01 2.37075780e-02
3.67758274e-02 -2.48022616e-01 1.25383705e-01 2.26026878e-01
8.44371021e-01 2.76843935e-01 2.06508934e-01 -3.37311715e-01
4.65653777e-01 -1.29044041e-01 7.24290311e-01 6.62366986e-01
6.45489292e-03 1.09463573e+00 2.52039015e-01 -7.43343532e-01
-1.13154352e+00 -1.38871729e+00 -3.43764275e-01 4.87255216e-01
5.32225490e-01 -4.00371552e-01 -4.92480010e-01 -2.11101159e-01
3.82812858e-01 3.02785873e-01 -5.83775699e-01 2.52019633e-02
-8.56117964e-01 -3.80254626e-01 3.15351099e-01 7.48195767e-01
9.51555371e-01 -7.96400964e-01 -1.14845723e-01 1.36784971e-01
-4.88443762e-01 -1.04242337e+00 -8.87190819e-01 -3.48380685e-01
-8.96693230e-01 -1.12826693e+00 -7.12422788e-01 -6.29454672e-01
4.22830641e-01 4.13517684e-01 1.49140620e+00 1.95465628e-02
-3.02656859e-01 -3.50972638e-02 -3.11299451e-02 5.21010906e-02
2.10497409e-01 2.54573196e-01 -1.74056798e-01 -1.33162752e-01
3.30720216e-01 -7.52020419e-01 -1.15812075e+00 5.25852621e-01
-5.01718342e-01 4.85834420e-01 4.84030902e-01 9.56861377e-01
9.99371827e-01 -5.53277671e-01 1.35207281e-01 -7.62741864e-01
8.86820480e-02 -2.15021804e-01 -7.65397668e-01 7.90929273e-02
-4.42800641e-01 7.89490119e-02 2.72088170e-01 1.46931008e-01
-8.21382403e-01 -4.99002226e-02 -4.79180604e-01 -8.23516905e-01
1.79245055e-01 3.73527199e-01 -1.24893701e-02 -2.97730267e-01
4.61106688e-01 1.29719749e-01 -1.25786215e-01 -5.74553847e-01
2.24851742e-01 1.76573470e-01 6.70800865e-01 -6.35970831e-01
7.35182822e-01 5.46571612e-01 1.46026298e-01 -3.73715222e-01
-9.76503432e-01 -5.07128775e-01 -5.86813807e-01 -1.78437322e-01
6.94987953e-01 -1.23607838e+00 -7.44749725e-01 7.37669647e-01
-1.10205770e+00 -4.24372405e-01 -5.23769334e-02 4.83795941e-01
-4.98514503e-01 3.75742584e-01 -7.63737381e-01 -2.22752392e-01
-4.79801744e-01 -1.28624070e+00 1.45209062e+00 1.24462627e-01
-8.10289383e-02 -9.04650509e-01 8.54339972e-02 6.36099041e-01
3.11787933e-01 3.47030044e-01 5.16054630e-01 -5.94074056e-02
-1.05889678e+00 1.47074059e-01 -6.36088908e-01 -5.46549931e-02
-8.34734589e-02 -2.36478269e-01 -8.79819036e-01 -4.05526370e-01
-4.19402480e-01 -3.55307549e-01 1.38248980e+00 7.51465082e-01
1.36672318e+00 9.93962437e-02 -3.08766037e-01 1.46070504e+00
1.46605265e+00 -5.64785488e-02 7.41557717e-01 5.23941040e-01
1.08181334e+00 4.07467455e-01 6.03754163e-01 2.69974232e-01
6.91347718e-01 9.95930076e-01 4.12857354e-01 -3.41505349e-01
-3.55432361e-01 -5.12161374e-01 -1.11478187e-01 8.29996824e-01
-2.23077878e-01 -8.02526921e-02 -1.11940837e+00 4.32252884e-01
-1.77345955e+00 -9.87305224e-01 -4.72422987e-01 2.06227112e+00
4.87219691e-01 4.36374620e-02 -1.67222172e-01 -2.68838525e-01
6.92486405e-01 4.52299088e-01 -7.43512511e-01 -2.34075333e-03
-4.77897257e-01 4.42296326e-01 5.92668653e-01 7.78709173e-01
-1.23816144e+00 1.01137078e+00 4.93431520e+00 1.00903296e+00
-1.06413245e+00 5.86434714e-02 8.50704849e-01 -8.51034969e-02
-5.91492057e-01 -2.56875921e-02 -6.82204664e-01 2.49515221e-01
1.82553735e-02 2.38630235e-01 3.59086603e-01 6.01749480e-01
7.62504414e-02 -9.06802788e-02 -1.07724035e+00 1.45202184e+00
1.49529455e-02 -1.63615143e+00 1.32257625e-01 -4.34723236e-02
1.00945771e+00 5.31164408e-01 1.80868894e-01 1.75770283e-01
1.61970451e-01 -1.07088614e+00 6.32588029e-01 4.73763615e-01
1.00046396e+00 -7.13482022e-01 7.82150090e-01 -3.22773606e-02
-1.54592741e+00 2.40155831e-01 -4.39161897e-01 2.85323747e-02
2.48232394e-01 6.97205603e-01 -2.46241510e-01 1.01083529e+00
8.00840914e-01 1.23439622e+00 -6.71594501e-01 1.08772147e+00
9.33537725e-03 1.01953506e-01 -4.64170277e-01 4.43427056e-01
4.56752688e-01 -4.13231194e-01 4.85427082e-01 8.77548218e-01
4.05392021e-01 7.19792694e-02 1.96259499e-01 9.79225159e-01
-8.76355246e-02 6.04394935e-02 -6.05109036e-01 5.05663931e-01
4.04130101e-01 8.98417711e-01 -5.36918521e-01 -2.30886042e-01
-4.21657950e-01 9.21013236e-01 6.64960325e-01 2.86670119e-01
-6.82094455e-01 -5.24780788e-02 8.54828835e-01 2.08628237e-01
2.95466006e-01 -2.03749929e-02 -5.16109288e-01 -1.44571030e+00
2.57855505e-01 -7.73899078e-01 5.18893838e-01 -9.29227293e-01
-1.38138282e+00 5.74065924e-01 -1.83107015e-02 -1.44302881e+00
-1.36152029e-01 -4.37971443e-01 -4.00540918e-01 9.96779263e-01
-1.53300989e+00 -1.12110889e+00 -7.09612012e-01 6.38161957e-01
5.19698739e-01 -1.35214910e-01 2.51927406e-01 5.22619307e-01
-2.05545813e-01 8.32997501e-01 8.77402127e-02 3.08385223e-01
8.35726798e-01 -9.36969817e-01 1.09351707e+00 6.24950409e-01
1.53161868e-01 5.01368701e-01 2.87364364e-01 -6.91670597e-01
-1.38163757e+00 -1.14667058e+00 7.94649124e-01 -4.52775300e-01
3.80546957e-01 -2.99547762e-01 -8.74084592e-01 6.01483047e-01
-8.76366645e-02 1.57701641e-01 5.59517480e-02 6.29054606e-02
-6.51879907e-01 -3.14548224e-01 -8.91692579e-01 5.81822455e-01
1.66433167e+00 -6.28452301e-01 -3.38353992e-01 2.00182185e-01
5.83130658e-01 -1.16768861e+00 -7.51631021e-01 7.88509130e-01
6.11111760e-01 -1.48031616e+00 1.30489886e+00 -1.91092968e-01
7.38818884e-01 -1.45404905e-01 -4.37202394e-01 -1.20489836e+00
-4.50893521e-01 -4.45650756e-01 1.86089218e-01 9.59752440e-01
3.24683368e-01 -8.51169586e-01 9.30732310e-01 4.06264693e-01
-4.43618864e-01 -1.10485387e+00 -1.08230948e+00 -6.42908692e-01
1.48878366e-01 -5.84615052e-01 7.74029613e-01 9.53432977e-01
-6.03645086e-01 1.33722156e-01 -3.18568647e-01 1.04178809e-01
9.49020445e-01 5.87090194e-01 8.34538043e-01 -1.06682849e+00
-4.56204295e-01 -5.79997063e-01 -6.17980957e-01 -1.40593004e+00
2.23393477e-02 -1.27371514e+00 -3.13078761e-01 -1.65164673e+00
3.36201131e-01 -5.07820129e-01 -1.82893630e-02 2.63099402e-01
-2.98900187e-01 4.52019036e-01 1.61911860e-01 1.71116918e-01
-2.83808023e-01 6.40700221e-01 1.67308402e+00 -2.20238507e-01
-6.09383956e-02 -1.39697552e-01 -3.77662599e-01 4.49239403e-01
7.80904591e-01 -6.94039091e-02 -3.30063939e-01 -8.42493713e-01
3.13032985e-01 1.97951004e-01 6.99747801e-01 -9.69255328e-01
2.02499703e-01 -2.47340500e-02 4.49339539e-01 -1.07010674e+00
5.05253971e-01 -3.60880405e-01 4.13942248e-01 3.97669762e-01
-1.74107864e-01 3.19810450e-01 9.75667015e-02 3.69476825e-01
-4.43173856e-01 2.86980063e-01 6.79379582e-01 -1.28462493e-01
-1.02867639e+00 9.48381305e-01 5.76987863e-01 4.70025897e-01
5.80940604e-01 -4.17248100e-01 -4.01286751e-01 -3.96037608e-01
-3.75044376e-01 6.45579755e-01 6.31648362e-01 4.41890121e-01
7.56082892e-01 -1.66833603e+00 -8.82043421e-01 2.58048505e-01
2.64124870e-01 4.38425511e-01 7.37160504e-01 8.22505772e-01
-8.68627191e-01 4.87577885e-01 -1.42725185e-01 -9.40651476e-01
-1.00445318e+00 1.98974833e-01 5.79821229e-01 -2.80906677e-01
-9.96443689e-01 1.12561619e+00 7.49731243e-01 -6.00058019e-01
8.20424706e-02 -2.40801111e-01 -3.50993080e-03 -1.99626222e-01
2.59666115e-01 2.52264380e-01 1.60341069e-01 -7.25813806e-01
-4.48829114e-01 1.17101777e+00 -1.04725629e-01 5.04843965e-02
1.32467926e+00 4.11150604e-02 -1.97213173e-01 4.53302786e-02
1.47672343e+00 -2.14884192e-01 -1.38916123e+00 -5.19196928e-01
-4.15278345e-01 -8.37672055e-01 1.71067908e-01 -4.72789824e-01
-1.56127095e+00 9.58483279e-01 6.20309174e-01 -7.30851173e-01
9.61393535e-01 1.68430638e-02 1.12682605e+00 2.52678722e-01
6.52902961e-01 -8.18381727e-01 2.78738029e-02 7.56320775e-01
1.11160111e+00 -1.42151761e+00 1.00775048e-01 -6.64572477e-01
-4.30784553e-01 1.03445566e+00 9.84522700e-01 -3.23022336e-01
6.69612288e-01 5.44404564e-03 -2.37118863e-02 -4.49906349e-01
-4.97390836e-01 -2.57879764e-01 7.09476411e-01 3.91604155e-01
3.73702884e-01 3.42718102e-02 -1.85796484e-01 7.67541081e-02
-5.36666989e-01 -1.29755139e-01 -6.98370263e-02 4.92775470e-01
-1.07395999e-01 -8.84983361e-01 -2.20873401e-01 7.47171402e-01
-7.32657015e-02 -4.36389089e-01 -2.81681389e-01 8.28976452e-01
1.42532721e-01 4.79818702e-01 3.72726738e-01 -6.53141558e-01
5.98518610e-01 -4.50062305e-01 5.46440184e-01 -1.56136230e-01
-6.14292681e-01 1.69017930e-02 7.95395821e-02 -1.11421287e+00
-2.96692848e-01 -7.70467639e-01 -9.45629537e-01 -7.04111099e-01
4.84838150e-02 -2.64831424e-01 2.35908717e-01 7.55659878e-01
4.34843272e-01 2.31915444e-01 6.83301628e-01 -1.01474857e+00
-2.38443941e-01 -7.51169145e-01 -1.79192081e-01 5.72103858e-01
2.81772196e-01 -9.48764384e-01 -2.17540354e-01 -5.28015435e-01] | [8.851652145385742, -2.3508188724517822] |
8cbeb63c-772f-41fa-94cc-d84f96081ac7 | did-you-miss-the-sign-a-false-negative-alarm | 1903.06391 | null | http://arxiv.org/abs/1903.06391v1 | http://arxiv.org/pdf/1903.06391v1.pdf | Did You Miss the Sign? A False Negative Alarm System for Traffic Sign Detectors | Object detection is an integral part of an autonomous vehicle for its
safety-critical and navigational purposes. Traffic signs as objects play a
vital role in guiding such systems. However, if the vehicle fails to locate any
critical sign, it might make a catastrophic failure. In this paper, we propose
an approach to identify traffic signs that have been mistakenly discarded by
the object detector. The proposed method raises an alarm when it discovers a
failure by the object detector to detect a traffic sign. This approach can be
useful to evaluate the performance of the detector during the deployment phase.
We trained a single shot multi-box object detector to detect traffic signs and
used its internal features to train a separate false negative detector (FND).
During deployment, FND decides whether the traffic sign detector (TSD) has
missed a sign or not. We are using precision and recall to measure the accuracy
of FND in two different datasets. For 80% recall, FND has achieved 89.9%
precision in Belgium Traffic Sign Detection dataset and 90.8% precision in
German Traffic Sign Recognition Benchmark dataset respectively. To the best of
our knowledge, our method is the first to tackle this critical aspect of false
negative detection in robotic vision. Such a fail-safe mechanism for object
detection can improve the engagement of robotic vision systems in our daily
life. | ['Niko Sünderhauf', 'Quazi Marufur Rahman', 'Feras Dayoub'] | 2019-03-15 | null | null | null | null | ['traffic-sign-recognition', 'traffic-sign-detection'] | ['computer-vision', 'computer-vision'] | [ 7.91563243e-02 3.83289494e-02 3.67436931e-02 -1.77929431e-01
-4.35711443e-01 -5.09640157e-01 7.19290495e-01 -2.62505680e-01
-5.35704553e-01 5.94455004e-01 -6.26260877e-01 -6.62676096e-01
-1.28333256e-01 -6.55679941e-01 -6.99694157e-01 -7.91112781e-01
2.32702926e-01 5.72859526e-01 1.11864007e+00 -1.35647040e-02
4.41620767e-01 8.51623714e-01 -2.02636909e+00 3.02817374e-01
4.65620667e-01 9.96761739e-01 3.37034881e-01 9.75753427e-01
5.09805838e-03 9.40715194e-01 -8.71586442e-01 -1.01237886e-01
4.35363203e-01 -8.39460865e-02 -5.54796994e-01 -4.79843989e-02
6.80816293e-01 -6.41505599e-01 3.27083394e-02 9.15285349e-01
3.06866974e-01 -2.76903510e-01 8.94180954e-01 -1.81989920e+00
1.66622445e-01 -1.36646360e-01 -3.50273550e-01 4.44239020e-01
1.61129981e-02 7.65077591e-01 4.69980985e-01 -9.27840948e-01
6.67875528e-01 9.13805246e-01 4.65913564e-01 7.11733580e-01
-5.82938612e-01 -7.86594212e-01 -1.90365702e-01 6.63441777e-01
-1.02594578e+00 -6.45405352e-01 4.93156433e-01 -6.93665206e-01
8.30573499e-01 -7.39525585e-03 4.59525973e-01 7.13330090e-01
3.29982862e-02 6.77337348e-01 8.61464918e-01 -4.26660240e-01
2.60344297e-01 2.28181899e-01 4.46671844e-01 7.60564148e-01
8.03050995e-01 6.40408576e-01 -7.33417720e-02 8.50635394e-02
1.03150301e-01 -2.51619309e-01 2.35237285e-01 -2.35299781e-01
-8.95326316e-01 5.61514914e-01 2.57741123e-01 3.07978034e-01
-6.76009595e-01 3.16908985e-01 4.73875701e-01 4.82153088e-01
-3.58537674e-01 -1.54171074e-02 -1.27758503e-01 -4.07261342e-01
-6.10141873e-01 -3.88919376e-02 7.16922760e-01 6.99966967e-01
2.47709379e-01 1.61476120e-01 -4.33126748e-01 4.70651656e-01
2.77112693e-01 9.61380124e-01 1.50325909e-01 -7.71443665e-01
5.94156608e-02 7.93342054e-01 4.34886962e-01 -8.19369793e-01
-4.85548049e-01 -4.20181639e-02 -1.91769555e-01 1.05502439e+00
7.47943997e-01 4.25404608e-02 -1.13460851e+00 1.02363074e+00
4.50986683e-01 1.85176343e-01 1.11906432e-01 1.03985715e+00
8.63742590e-01 3.43875170e-01 9.66559649e-02 7.66668990e-02
1.44074464e+00 -5.71134627e-01 -5.45144200e-01 -1.70868188e-01
7.39905655e-01 -9.67370808e-01 3.95009726e-01 7.01363862e-01
-4.75052893e-01 -5.99741638e-01 -1.21633875e+00 5.55028439e-01
-5.04706323e-01 5.98605931e-01 2.89726764e-01 7.81181455e-01
-7.25160837e-01 -5.56388013e-02 -5.42318583e-01 -8.05280805e-01
4.26951259e-01 3.75023603e-01 -2.77592599e-01 -9.60998088e-02
-7.53050268e-01 1.36099148e+00 2.70497233e-01 1.60448283e-01
-1.08715439e+00 -3.52436192e-02 -2.20992684e-01 -3.53134781e-01
6.01018488e-01 -8.17731842e-02 1.40478790e+00 -7.99813330e-01
-9.42902207e-01 1.11294985e+00 -1.98532864e-01 -7.69186854e-01
9.44301486e-01 2.72043273e-02 -6.65553033e-01 1.78247228e-01
2.44666100e-01 5.16610920e-01 1.17976665e+00 -1.24909103e+00
-1.25327182e+00 -1.05262741e-01 -1.60478383e-01 -4.89956439e-01
4.34449196e-01 1.62160248e-01 -5.87010495e-02 6.61933795e-02
4.50785831e-02 -9.03312445e-01 3.43639135e-01 3.25052083e-01
-1.85380504e-01 -5.77183008e-01 1.57541001e+00 -2.31862187e-01
8.53666902e-01 -2.01230264e+00 -8.53158832e-01 1.92546934e-01
2.89958678e-02 1.13181424e+00 -8.97840187e-02 8.43410790e-02
1.65386841e-01 -2.77215838e-01 5.33248894e-02 2.43659630e-01
-2.82243967e-01 4.55548227e-01 -5.12076616e-01 5.85265636e-01
5.72265506e-01 8.10515463e-01 -8.45279574e-01 -5.12165606e-01
6.76341772e-01 2.41466641e-01 5.47047369e-02 -6.75559565e-02
1.52731404e-01 1.62341371e-01 -4.75760102e-01 9.38893259e-01
6.70886338e-01 3.03978980e-01 -2.93599784e-01 -2.61596173e-01
-3.22418153e-01 -6.33693114e-02 -1.38636637e+00 3.35588574e-01
-1.46308675e-01 1.10095906e+00 -9.00633931e-02 -1.04773271e+00
1.11809897e+00 2.92616069e-01 1.83844596e-01 -1.08471823e+00
2.98125684e-01 7.80347347e-01 3.28436077e-01 -9.18200433e-01
5.65532148e-01 -3.64632718e-02 1.70204937e-01 2.80936092e-01
-2.54965425e-01 1.63638473e-01 3.59283656e-01 6.88278750e-02
1.37675428e+00 -5.01072928e-02 1.42240614e-01 1.70247883e-01
9.15709734e-01 4.01640326e-01 3.23384404e-01 1.12631142e+00
-9.58505273e-01 2.90738076e-01 4.59457666e-01 -5.33187330e-01
-9.91137266e-01 -8.26544106e-01 -1.92704797e-01 6.18613660e-01
2.20283136e-01 4.48645294e-01 -3.55791718e-01 -1.02186155e+00
2.89463073e-01 8.16084683e-01 -4.02897656e-01 -4.74538177e-01
-7.06933200e-01 -2.32444420e-01 8.17564130e-01 3.37536126e-01
4.99490142e-01 -1.27742660e+00 -1.40186214e+00 9.34841856e-02
2.36425698e-01 -1.09004271e+00 2.87744761e-01 2.03862652e-01
-4.67780799e-01 -2.01038980e+00 -4.54992265e-01 -6.70760453e-01
7.33684182e-01 6.58737957e-01 6.35062516e-01 4.95283216e-01
-4.88721043e-01 5.00399292e-01 -5.20239949e-01 -7.53530681e-01
-9.99318302e-01 -4.28995311e-01 4.18018885e-02 1.49385974e-01
6.42041028e-01 3.37868541e-01 -5.40604591e-01 9.13602352e-01
-4.87621278e-01 -4.98220325e-01 6.40498459e-01 4.69515383e-01
1.94342345e-01 -1.74644634e-01 6.99688315e-01 -3.09720963e-01
5.05651712e-01 -2.09256291e-01 -9.73348081e-01 2.46059671e-01
-3.01876247e-01 1.22731486e-02 2.48030201e-01 -4.73858327e-01
-5.61083972e-01 3.25605035e-01 1.82430148e-01 -3.79114777e-01
-4.72069919e-01 -3.04578990e-01 4.05307502e-01 -3.07442844e-01
7.65472174e-01 1.12474822e-01 3.07259977e-01 -2.67364740e-01
1.03970245e-02 1.01628971e+00 4.81362641e-01 1.73226907e-03
6.77163780e-01 5.52053511e-01 4.62561846e-01 -8.59825373e-01
-3.42277139e-01 -9.58482206e-01 -4.55895394e-01 -9.55610037e-01
5.77646196e-01 -4.56005663e-01 -1.11888075e+00 5.00037432e-01
-1.35402632e+00 8.27123970e-02 -2.37193272e-01 4.60342407e-01
-3.48570794e-01 2.00687721e-01 3.44079912e-01 -1.45513332e+00
-2.16355979e-01 -1.07474399e+00 1.11784053e+00 8.71440470e-02
2.71968227e-02 -1.65951103e-01 -2.04097390e-01 7.11719319e-02
5.59310853e-01 5.56640290e-02 1.73248947e-01 -7.99608946e-01
-6.30796552e-01 -8.35565507e-01 -5.19929111e-01 4.55007017e-01
-6.19952306e-02 3.07265669e-01 -1.10228837e+00 2.18296662e-01
-2.55629480e-01 7.12427944e-02 1.07303476e+00 2.54133046e-01
1.29006833e-01 9.43945423e-02 -6.03996098e-01 -2.20591143e-01
1.22373140e+00 8.29144061e-01 6.35073721e-01 5.42035639e-01
1.57376841e-01 4.34183657e-01 1.16485167e+00 7.69902859e-03
-7.54018500e-02 7.94043541e-01 6.78750396e-01 2.38527060e-01
-5.86625099e-01 1.83662191e-01 5.36163390e-01 -2.30656683e-01
7.36139193e-02 -2.52646953e-01 -1.14918387e+00 4.58870441e-01
-1.93767607e+00 -1.34538662e+00 -8.96721184e-01 2.32270908e+00
-3.55879255e-02 5.40417135e-01 4.21313971e-01 6.04285896e-01
8.39105189e-01 -4.31411207e-01 -3.74706656e-01 -5.16064107e-01
4.83294576e-02 -4.25125450e-01 8.77916515e-01 3.33378762e-01
-1.17354131e+00 8.97078097e-01 5.28676414e+00 5.31460226e-01
-1.27128685e+00 2.31665611e-01 8.75920504e-02 1.22997947e-01
6.66315496e-01 -7.05617294e-02 -1.11610758e+00 6.21347547e-01
7.76544213e-01 2.89665967e-01 -1.95874929e-01 1.00076890e+00
4.41971749e-01 -8.26371193e-01 -8.47113729e-01 8.02868187e-01
1.56768486e-01 -8.73442173e-01 -2.43701026e-01 -6.88560754e-02
9.79126766e-02 1.28997430e-01 -2.66477674e-01 3.67525309e-01
-3.30419876e-02 -5.96260667e-01 9.47997510e-01 9.04910386e-01
4.75333989e-01 -5.38577378e-01 9.88068163e-01 5.23699641e-01
-1.03482878e+00 -2.46324435e-01 -3.08675498e-01 -4.13056724e-02
4.66389477e-01 3.08223903e-01 -1.49863231e+00 -2.62057967e-02
4.32051122e-01 2.06729889e-01 -7.14918017e-01 1.68946373e+00
-3.27491969e-01 6.02309346e-01 -5.21491706e-01 -4.33349669e-01
3.59320909e-01 2.65978426e-01 1.09423065e+00 9.66973424e-01
2.00446099e-01 -2.66127735e-01 -9.48150977e-02 5.98013282e-01
6.37295783e-01 -3.23436379e-01 -9.56948042e-01 1.95485100e-01
2.28555039e-01 9.74880874e-01 -9.64448154e-01 -5.05079567e-01
-1.47414178e-01 7.08097339e-01 -1.68609023e-01 7.24022612e-02
-1.05919051e+00 -4.88736063e-01 4.10264909e-01 2.40214646e-01
5.92192531e-01 -6.96061477e-02 -1.33018658e-01 -3.82236630e-01
3.00933152e-01 -2.84625202e-01 2.62860179e-01 -9.01473522e-01
-7.16749310e-01 5.67546666e-01 -2.16847003e-01 -1.71592319e+00
-2.33835980e-01 -1.05670142e+00 -6.73164845e-01 3.92090648e-01
-1.40214646e+00 -9.43124652e-01 -4.26883101e-01 1.14065118e-01
4.95820105e-01 -4.97250378e-01 1.83299184e-01 3.27491641e-01
-3.18926215e-01 3.72062743e-01 -3.08967590e-01 1.73923105e-01
6.25701427e-01 -5.43584943e-01 -5.13610989e-02 1.09174514e+00
-1.75567225e-01 2.09296588e-02 8.05636168e-01 -9.02754664e-01
-1.27855933e+00 -9.26143706e-01 9.67641175e-01 -7.24790871e-01
4.04447168e-01 2.40010247e-01 -6.53483093e-01 1.27207354e-01
-4.28618431e-01 1.69029683e-01 -1.86464310e-01 -5.69375038e-01
-1.56452343e-01 -3.55033338e-01 -1.46367788e+00 2.78797865e-01
6.93797350e-01 -8.15347210e-03 -7.30199754e-01 2.75831878e-01
1.30672991e-01 9.00905877e-02 5.50491326e-02 5.46306133e-01
8.97592247e-01 -1.01468730e+00 8.39680910e-01 -4.99373943e-01
-2.78466105e-01 -7.74536610e-01 1.08100800e-02 -4.77893293e-01
2.08098844e-01 -1.36099532e-01 -1.96850628e-01 1.01453519e+00
2.70514995e-01 -7.44366348e-01 5.97838759e-01 4.52332377e-01
-1.85376793e-01 -2.91543573e-01 -1.31613863e+00 -1.14119506e+00
-5.41854203e-01 -8.85002792e-01 1.98361650e-01 2.26726159e-01
-4.76420701e-01 -9.36722234e-02 -2.44736880e-01 2.21741453e-01
5.84698081e-01 -3.33808243e-01 1.04290676e+00 -1.42416072e+00
4.19373661e-01 -7.27144241e-01 -1.12668955e+00 -4.25593555e-01
-3.07204396e-01 -5.92166901e-01 3.08214605e-01 -1.50564539e+00
-1.74675077e-01 -3.90362203e-01 -1.56230643e-01 4.54699576e-01
2.78278887e-01 3.69419962e-01 2.46929735e-01 2.17812121e-01
-7.46467412e-01 -4.41968068e-02 8.65430593e-01 -1.15479775e-01
-5.27398288e-02 3.20861697e-01 3.68341394e-02 7.04711795e-01
9.68070447e-01 -7.46322393e-01 2.11655825e-01 3.65193672e-02
1.07919075e-01 -3.22997570e-01 1.04470754e+00 -1.23423016e+00
6.16591513e-01 -4.47265804e-03 3.09057739e-02 -1.01036155e+00
1.16568103e-01 -1.30654800e+00 -1.48689821e-01 1.20957625e+00
3.07144791e-01 -3.17849338e-01 3.02457273e-01 5.76298833e-01
1.87342800e-02 -6.09012783e-01 8.57038558e-01 9.17452425e-02
-1.32009935e+00 -3.29589725e-01 -7.87370563e-01 -4.28746015e-01
1.52145791e+00 -8.87780428e-01 -4.19827163e-01 4.09875363e-02
-2.84772485e-01 3.09149683e-01 1.04091518e-01 6.80633724e-01
7.82215834e-01 -1.06280899e+00 -5.06092668e-01 4.53263491e-01
3.67628187e-01 -5.65771401e-01 -1.04336925e-01 1.18997014e+00
-6.11912310e-01 6.66658521e-01 -1.26629159e-01 -9.57394540e-01
-1.75947702e+00 4.41495895e-01 5.94891667e-01 3.64535987e-01
-5.77165365e-01 4.42278713e-01 -4.51387644e-01 1.97574012e-02
4.52647060e-01 -3.85036498e-01 -3.79620403e-01 2.76415497e-01
6.40055478e-01 8.52317989e-01 2.77008444e-01 -1.02831590e+00
-7.21466362e-01 8.38053465e-01 7.99029171e-02 -6.31989017e-02
7.54938900e-01 1.45915031e-01 2.79523373e-01 3.07314217e-01
7.14612067e-01 -3.95355284e-01 -1.04786742e+00 1.20669767e-01
2.30031416e-01 -4.23253298e-01 -5.20142354e-03 -1.11067295e+00
-7.66123652e-01 5.63790441e-01 1.09786856e+00 2.02203169e-01
6.88584983e-01 1.39877543e-01 3.84091765e-01 8.15512002e-01
6.70132339e-01 -1.14020753e+00 -1.03273027e-01 6.27955854e-01
9.32026565e-01 -1.72670674e+00 -2.36832842e-01 -1.97832689e-01
-6.45441413e-01 1.29415059e+00 7.45593190e-01 -1.80374086e-01
4.73830491e-01 1.16434708e-01 1.30834967e-01 -4.03745979e-01
-5.48697591e-01 -7.75751352e-01 4.16744649e-01 7.02485383e-01
-2.67087340e-01 6.49618059e-02 -5.62211096e-01 8.48836526e-02
2.68971831e-01 4.15383279e-01 4.41450834e-01 1.10627151e+00
-1.19050825e+00 -5.62738717e-01 -6.66433692e-01 5.26136637e-01
6.41198307e-02 4.66492206e-01 -6.14181161e-01 1.07065916e+00
4.25896585e-01 1.17202783e+00 1.67577446e-01 -3.22946966e-01
6.87925637e-01 2.68251985e-01 2.32266977e-01 -1.32665083e-01
-3.60001355e-01 -4.43830222e-01 3.42250109e-01 -5.13527334e-01
-3.42698604e-01 -7.17269063e-01 -1.45266569e+00 1.42019764e-01
-4.37692106e-01 -2.24490091e-01 1.00252664e+00 9.33254361e-01
2.23410055e-01 3.71920496e-01 3.67396235e-01 -4.91009176e-01
-4.53816950e-01 -8.36573303e-01 -5.51142953e-02 2.45812878e-01
5.71864009e-01 -1.18598139e+00 -4.08109099e-01 -1.77142456e-01] | [7.988853454589844, -0.8304251432418823] |
4ebe861f-688e-4a4e-83b0-1d21b28bdec2 | dynamic-mlp-for-fine-grained-image | 2203.03253 | null | https://arxiv.org/abs/2203.03253v1 | https://arxiv.org/pdf/2203.03253v1.pdf | Dynamic MLP for Fine-Grained Image Classification by Leveraging Geographical and Temporal Information | Fine-grained image classification is a challenging computer vision task where various species share similar visual appearances, resulting in misclassification if merely based on visual clues. Therefore, it is helpful to leverage additional information, e.g., the locations and dates for data shooting, which can be easily accessible but rarely exploited. In this paper, we first demonstrate that existing multimodal methods fuse multiple features only on a single dimension, which essentially has insufficient help in feature discrimination. To fully explore the potential of multimodal information, we propose a dynamic MLP on top of the image representation, which interacts with multimodal features at a higher and broader dimension. The dynamic MLP is an efficient structure parameterized by the learned embeddings of variable locations and dates. It can be regarded as an adaptive nonlinear projection for generating more discriminative image representations in visual tasks. To our best knowledge, it is the first attempt to explore the idea of dynamic networks to exploit multimodal information in fine-grained image classification tasks. Extensive experiments demonstrate the effectiveness of our method. The t-SNE algorithm visually indicates that our technique improves the recognizability of image representations that are visually similar but with different categories. Furthermore, among published works across multiple fine-grained datasets, dynamic MLP consistently achieves SOTA results https://paperswithcode.com/dataset/inaturalist and takes third place in the iNaturalist challenge at FGVC8 https://www.kaggle.com/c/inaturalist-2021/leaderboard. Code is available at https://github.com/ylingfeng/DynamicMLP.git | ['Jian Yang', 'Jiajun Liang', 'Shihao Zhou', 'Juntian Tao', 'Borui Zhao', 'RenJie Song', 'Xiang Li', 'Lingfeng Yang'] | 2022-03-07 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Yang_Dynamic_MLP_for_Fine-Grained_Image_Classification_by_Leveraging_Geographical_and_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Yang_Dynamic_MLP_for_Fine-Grained_Image_Classification_by_Leveraging_Geographical_and_CVPR_2022_paper.pdf | cvpr-2022-1 | ['fine-grained-image-classification'] | ['computer-vision'] | [-2.87199169e-02 -3.64420652e-01 -3.54543000e-01 -2.51073569e-01
-4.82511133e-01 -9.80442464e-01 7.54807830e-01 -5.83574059e-04
-3.92499506e-01 5.01524508e-01 2.16930538e-01 -2.23459050e-01
-2.86286503e-01 -5.56904495e-01 -8.11317265e-01 -9.23416436e-01
1.72222350e-02 1.40996069e-01 -2.17777923e-01 7.00678080e-02
3.59186292e-01 5.56888521e-01 -1.70204401e+00 5.37494659e-01
5.63912332e-01 1.16376579e+00 3.67755473e-01 6.22964919e-01
-7.95233920e-02 3.87906879e-01 -3.41667473e-01 -3.35403264e-01
3.78824085e-01 -2.27489062e-02 -6.44984901e-01 1.25860974e-01
8.72069418e-01 -3.33943039e-01 -4.27779555e-01 1.03637791e+00
3.83519202e-01 -1.50744794e-02 7.40339875e-01 -1.51936173e+00
-1.04200423e+00 5.14541745e-01 -9.51429367e-01 1.93050668e-01
7.54243881e-02 3.06381017e-01 1.02075517e+00 -9.82160985e-01
3.80743980e-01 1.34925020e+00 5.22298932e-01 4.01885748e-01
-1.35233271e+00 -7.54860520e-01 5.23682535e-01 3.18883061e-01
-1.40559745e+00 -3.06517184e-01 8.21346521e-01 -5.90402126e-01
5.31802237e-01 4.96378750e-01 4.39368814e-01 1.34912121e+00
-7.39281103e-02 8.40935767e-01 1.38619375e+00 -1.82386860e-01
-1.02808543e-01 1.36813000e-01 1.10967368e-01 9.09678400e-01
-2.18945257e-02 2.18592733e-01 -6.60686851e-01 -8.44208747e-02
7.41551042e-01 4.73437339e-01 -3.80443275e-01 -3.77534598e-01
-1.50862324e+00 8.46566856e-01 7.06395149e-01 1.98426157e-01
-9.04016495e-02 1.54278085e-01 2.30703726e-01 3.21733803e-01
3.27378035e-01 4.78652596e-01 -4.10110384e-01 -1.97754666e-01
-6.98628902e-01 -4.86053191e-02 5.12792706e-01 6.02692485e-01
8.91292453e-01 6.96806004e-03 -2.53591597e-01 1.13944435e+00
2.99390107e-01 6.80045843e-01 4.96974409e-01 -9.68176484e-01
3.50979209e-01 6.92074776e-01 -8.57107341e-02 -1.10389519e+00
-4.47254181e-01 -2.17012197e-01 -9.23052430e-01 3.48446101e-01
6.27087533e-01 -8.20621625e-02 -1.05525196e+00 1.81568229e+00
8.29077810e-02 8.02097470e-02 -2.72409946e-01 1.01507747e+00
1.00769484e+00 5.60393393e-01 8.01506937e-02 3.35583210e-01
1.36338294e+00 -1.10474801e+00 -1.33190975e-01 -2.25482836e-01
2.66165733e-01 -7.30737746e-01 1.31869698e+00 3.16324681e-01
-6.64187074e-01 -5.27259767e-01 -9.42460537e-01 -5.45062944e-02
-8.17183316e-01 4.56157148e-01 8.47774088e-01 4.50332969e-01
-1.16822696e+00 5.53856373e-01 -5.49923301e-01 -4.86280560e-01
5.24870157e-01 2.68168211e-01 -7.30623901e-01 -2.49189228e-01
-7.66460001e-01 6.51248395e-01 2.76415527e-01 2.22265482e-01
-8.46899986e-01 -6.18013263e-01 -8.21207941e-01 -8.32937062e-02
2.63438642e-01 -6.98712528e-01 7.63517976e-01 -9.59031045e-01
-1.06800187e+00 9.25425112e-01 -2.03565747e-01 -1.41667604e-01
3.39674056e-01 1.01544879e-01 -2.01973438e-01 1.67575508e-01
1.35329530e-01 1.02528715e+00 1.12926388e+00 -1.41461349e+00
-6.32313848e-01 -5.54455757e-01 3.63815129e-01 5.68761788e-02
-4.88857150e-01 -2.75674731e-01 -5.29922187e-01 -6.99209809e-01
-1.95090860e-01 -1.13067985e+00 2.50835204e-03 2.83324450e-01
-5.13641834e-01 -2.91840255e-01 7.52238393e-01 -4.78554219e-01
9.74071085e-01 -2.25125694e+00 4.02885556e-01 -4.48047817e-02
4.04008389e-01 -5.75970300e-02 -4.87854838e-01 6.04594290e-01
-1.75187185e-01 4.32045937e-01 -1.33765534e-01 -4.63143229e-01
2.90878713e-01 1.61709830e-01 -3.90136987e-01 4.97919291e-01
3.23923737e-01 1.08544314e+00 -6.97911322e-01 -3.37482572e-01
3.87375027e-01 5.38248837e-01 -2.74355501e-01 2.91763563e-02
7.68864974e-02 3.14801008e-01 -3.33441049e-01 1.12864840e+00
6.64185047e-01 -4.37637269e-01 -1.21554919e-01 -6.00046813e-01
-2.44362146e-01 -4.08859760e-01 -8.83355796e-01 1.67199445e+00
-3.42280507e-01 7.09204733e-01 -7.92806223e-02 -8.21048021e-01
5.90682149e-01 -9.35296789e-02 1.44513309e-01 -7.52571166e-01
7.40818530e-02 9.97268874e-03 -1.18370742e-01 -3.26765418e-01
5.08437097e-01 2.16549486e-02 -2.96511203e-01 2.81643689e-01
6.63985312e-02 1.59578532e-01 2.39333808e-01 1.65539458e-01
8.78069520e-01 2.50556678e-01 1.96395621e-01 2.81264614e-02
8.64158496e-02 -9.53790098e-02 2.44360521e-01 9.52536643e-01
-3.35501313e-01 7.97935486e-01 4.58825290e-01 -4.34548587e-01
-9.32545722e-01 -1.21123552e+00 -3.00868183e-01 1.40910470e+00
2.53521442e-01 -3.99409652e-01 -2.99138427e-01 -7.65357077e-01
2.81916976e-01 3.07028055e-01 -1.11379135e+00 -6.09913841e-02
-7.72288665e-02 -6.98469758e-01 6.28482103e-01 5.58396757e-01
3.10137182e-01 -9.88441825e-01 -3.67713839e-01 -4.30671930e-01
-4.98716980e-02 -8.32900405e-01 -4.51452404e-01 4.22593683e-01
-6.48661852e-01 -1.01217604e+00 -8.72548342e-01 -5.40036976e-01
7.46839523e-01 5.32327831e-01 9.02804554e-01 5.26911765e-02
-4.34794754e-01 5.18436551e-01 -4.37216818e-01 -2.47027397e-01
1.81035265e-01 7.78216645e-02 -3.20793013e-03 1.91544354e-01
2.68844873e-01 -5.67381203e-01 -7.74720430e-01 4.05711323e-01
-8.93458843e-01 1.28315687e-01 7.69543231e-01 1.08713675e+00
6.14474893e-01 -1.65989771e-01 3.18739653e-01 -6.97329938e-01
5.10767221e-01 -6.14826381e-01 -3.95797729e-01 3.71929705e-01
-2.61938781e-01 1.21603727e-01 7.18351543e-01 -5.63428164e-01
-7.78447509e-01 2.33446993e-02 6.19973764e-02 -7.97794640e-01
-4.83898461e-01 3.65235209e-01 -1.50970342e-02 -2.88927406e-01
4.83598769e-01 2.79900759e-01 -1.71169758e-01 -6.19328022e-01
5.90491593e-01 5.98114848e-01 3.42058361e-01 -6.23714805e-01
8.23057234e-01 6.47293985e-01 -1.60341665e-01 -7.26913750e-01
-7.22155511e-01 -4.73012716e-01 -6.26429200e-01 -6.77109659e-02
6.26067698e-01 -8.79733086e-01 -8.93984675e-01 4.94563937e-01
-8.11115086e-01 -5.04412293e-01 -1.62822641e-02 2.11486489e-01
-3.70150536e-01 3.52804661e-01 -4.57732886e-01 -5.31293333e-01
1.38746845e-02 -1.10525870e+00 1.36487961e+00 4.56213742e-01
-1.29765980e-02 -1.01835203e+00 -7.24785030e-02 4.14039135e-01
2.81687587e-01 2.07791150e-01 7.53895044e-01 -2.97131091e-01
-6.29157722e-01 5.67267591e-04 -6.93897605e-01 1.53509766e-01
2.54879504e-01 9.29885060e-02 -1.22508478e+00 -4.07898277e-01
-5.17437518e-01 -5.85887551e-01 1.33472443e+00 2.82434136e-01
1.64153159e+00 -2.68528074e-01 -3.48526180e-01 9.87466693e-01
1.40235507e+00 -5.79190068e-02 3.66385370e-01 4.52838421e-01
1.10217547e+00 5.89385092e-01 5.49094558e-01 5.36667943e-01
5.58263063e-01 6.60459578e-01 6.62949204e-01 -2.14420050e-01
-2.66010255e-01 -2.90136814e-01 2.48493731e-01 3.28546733e-01
-1.60195857e-01 -1.23409949e-01 -7.65431881e-01 5.33914030e-01
-1.87052405e+00 -9.50069487e-01 3.07007730e-01 1.93724382e+00
6.37955844e-01 -3.33467603e-01 2.41860505e-02 -1.94503322e-01
5.51575780e-01 5.20583034e-01 -7.36316383e-01 -3.31536800e-01
-3.76579940e-01 -1.60881639e-01 7.21894205e-01 2.17017367e-01
-1.29882538e+00 9.50550318e-01 4.86611032e+00 1.11056292e+00
-1.43793309e+00 -4.73587550e-02 6.70313120e-01 -3.37639540e-01
-3.41096550e-01 -1.65534198e-01 -7.53470302e-01 7.16080606e-01
6.06405973e-01 1.43628955e-01 6.59518123e-01 6.42774045e-01
1.31754264e-01 -3.93506140e-02 -1.00538027e+00 1.29784501e+00
2.75984585e-01 -1.35941315e+00 1.94085792e-01 2.05791503e-01
5.61370850e-01 1.91884398e-01 5.56421220e-01 1.55734017e-01
1.56668738e-01 -1.25788283e+00 7.66315937e-01 6.35281384e-01
9.15859342e-01 -6.03424966e-01 4.45358008e-01 1.09025218e-01
-1.20641065e+00 -2.89553225e-01 -4.47522193e-01 8.38803723e-02
-4.44462858e-02 2.52957404e-01 -7.25395679e-01 4.35722560e-01
1.08538342e+00 9.49842930e-01 -1.09367549e+00 1.16691303e+00
-1.78535208e-01 4.39716041e-01 -1.83263734e-01 7.16064349e-02
3.46206307e-01 -3.08229029e-02 4.13632721e-01 1.24553561e+00
3.80999267e-01 -3.39498997e-01 2.72191375e-01 6.34429634e-01
-1.68059632e-01 -1.37200162e-01 -8.48612726e-01 -3.19807023e-01
3.04897964e-01 1.59899211e+00 -6.55450523e-01 9.53155234e-02
-4.49628949e-01 1.04842269e+00 6.08249128e-01 5.62386394e-01
-7.95267165e-01 -1.20481923e-01 7.01840758e-01 -1.07506797e-01
4.95277375e-01 -3.15396339e-01 -2.71135807e-01 -1.33944035e+00
-2.39216108e-02 -8.74791861e-01 5.78442633e-01 -8.18336666e-01
-1.73026788e+00 7.10409045e-01 -6.79158494e-02 -1.28876126e+00
-7.24237263e-02 -8.81735861e-01 -3.35016340e-01 8.10403705e-01
-1.56242371e+00 -1.74140847e+00 -4.76569235e-01 8.41041386e-01
4.69919324e-01 -1.96923509e-01 7.95827389e-01 2.01429218e-01
-6.38287306e-01 7.20479846e-01 1.47215292e-01 2.81689111e-02
9.69978034e-01 -1.40851665e+00 -3.47976224e-03 7.01419115e-01
3.21020544e-01 7.54636586e-01 4.37975645e-01 -4.14650679e-01
-1.53660929e+00 -1.11466193e+00 3.62515986e-01 -5.89545250e-01
9.60504830e-01 -3.97536576e-01 -7.85161495e-01 5.12134075e-01
2.70537376e-01 1.39621403e-02 7.84663022e-01 2.78435886e-01
-7.54812717e-01 -1.88956499e-01 -8.56064320e-01 7.31308341e-01
9.46691036e-01 -7.69088447e-01 -3.61127794e-01 2.38415644e-01
5.59108853e-01 -1.52111620e-01 -8.61031353e-01 1.82494134e-01
7.98821688e-01 -8.66301835e-01 1.06083167e+00 -4.42508399e-01
4.95125800e-01 -5.00328481e-01 -4.57907051e-01 -1.41386259e+00
-4.20952141e-01 -1.73886821e-01 3.77404578e-02 1.26589227e+00
3.95668924e-01 -7.27446496e-01 6.60978317e-01 3.92950028e-01
1.68148596e-02 -7.84653962e-01 -8.96046758e-01 -6.46323204e-01
9.10945609e-02 -3.76600355e-01 3.32902789e-01 8.86682570e-01
-3.36176515e-01 1.70549437e-01 -6.29058480e-01 2.87070334e-01
6.79331660e-01 5.82280219e-01 7.48533607e-01 -1.07652485e+00
-3.32640260e-01 -6.93406701e-01 -3.41857851e-01 -9.45908725e-01
2.22619802e-01 -9.38052237e-01 -1.16201706e-01 -1.38747096e+00
5.32176554e-01 -4.38101172e-01 -4.45349753e-01 8.74517202e-01
-8.30880031e-02 7.62011170e-01 5.70240021e-01 3.70548069e-01
-7.09809661e-01 5.00967264e-01 1.13237405e+00 -4.72675502e-01
2.00373046e-02 -2.80233324e-01 -1.02753472e+00 6.35639787e-01
9.50665534e-01 -2.56180584e-01 -2.72114605e-01 -5.80604553e-01
-3.10198171e-03 -2.52979338e-01 8.04430962e-01 -5.95941663e-01
2.31873885e-01 -2.68430829e-01 6.27281785e-01 -5.38580179e-01
6.93634510e-01 -6.33143604e-01 9.17738974e-02 6.23027980e-02
-6.40330836e-02 1.56573474e-01 4.51576710e-01 6.08579278e-01
-3.45816493e-01 -8.63144398e-02 5.75699925e-01 -1.94131389e-01
-1.11199975e+00 5.45367718e-01 -7.95377344e-02 -1.98016331e-01
8.55712056e-01 -3.37000519e-01 -8.06632638e-01 -2.01802343e-01
-4.72905248e-01 2.77582198e-01 7.70135999e-01 7.55560756e-01
6.82015836e-01 -1.39862263e+00 -5.23854733e-01 1.32350251e-01
5.91799259e-01 -4.55536544e-01 6.63500547e-01 8.94736648e-01
-2.00728685e-01 3.87256980e-01 -5.01719415e-01 -7.84958005e-01
-1.42193782e+00 6.11426711e-01 1.92049563e-01 -3.70701328e-02
-3.51712257e-01 8.54346812e-01 7.08642662e-01 -4.73012835e-01
1.43710598e-01 -1.16439477e-01 -3.31278145e-01 3.49366814e-01
6.09647870e-01 3.59340049e-02 -2.41419271e-01 -7.69212127e-01
-4.30083364e-01 8.49290609e-01 -1.34407416e-01 1.00522175e-01
1.37683177e+00 -3.39873672e-01 -1.34462252e-01 6.10698521e-01
1.28495669e+00 1.50750697e-01 -1.52276397e+00 -6.05090484e-02
-2.24851862e-01 -7.55652249e-01 -5.42590618e-02 -1.07119262e+00
-1.11666858e+00 1.05551541e+00 8.42758417e-01 2.97038257e-01
1.22345543e+00 3.47754270e-01 3.85823846e-01 3.05858940e-01
3.27372819e-01 -6.10876501e-01 2.26764098e-01 3.60835433e-01
1.19702423e+00 -1.66018486e+00 -1.48230940e-01 -6.43292218e-02
-7.25997627e-01 1.11653388e+00 5.38341999e-01 5.21121025e-02
4.86016154e-01 6.41372129e-02 1.72727704e-01 -1.28564030e-01
-7.06051767e-01 -4.09467667e-01 4.84005213e-01 6.63934767e-01
2.30035678e-01 3.76399904e-01 3.65190096e-02 5.32384038e-01
-1.77060172e-01 -3.59791368e-01 3.44917476e-01 5.88610649e-01
-1.89919785e-01 -1.08719504e+00 -4.22283262e-01 5.67186475e-01
-3.18356335e-01 -2.99617410e-01 -6.18127763e-01 7.88319826e-01
3.34668487e-01 8.40113580e-01 -2.82804463e-02 -5.91985762e-01
1.71895787e-01 -1.90454960e-01 5.14400661e-01 -4.37032521e-01
-4.28406358e-01 -1.08225435e-01 -8.57231691e-02 -7.19411373e-01
-3.57363552e-01 -6.04763329e-01 -7.12022364e-01 -3.48189175e-01
1.43804580e-01 -2.36147612e-01 6.35314763e-01 5.69501817e-01
4.44745511e-01 3.18788469e-01 5.22663355e-01 -1.15991032e+00
-1.26140669e-01 -7.04224527e-01 -6.93037748e-01 4.23207909e-01
5.43626547e-01 -9.19287264e-01 -5.42270720e-01 -6.43090606e-02] | [9.675597190856934, 2.040534019470215] |
a10d35fc-7e24-4fe3-82a5-4bec4a02d7c7 | time-associated-meta-learning-for-clinical | 2303.02570 | null | https://arxiv.org/abs/2303.02570v1 | https://arxiv.org/pdf/2303.02570v1.pdf | Time Associated Meta Learning for Clinical Prediction | Rich Electronic Health Records (EHR), have created opportunities to improve clinical processes using machine learning methods. Prediction of the same patient events at different time horizons can have very different applications and interpretations; however, limited number of events in each potential time window hurts the effectiveness of conventional machine learning algorithms. We propose a novel time associated meta learning (TAML) method to make effective predictions at multiple future time points. We view time-associated disease prediction as classification tasks at multiple time points. Such closely-related classification tasks are an excellent candidate for model-based meta learning. To address the sparsity problem after task splitting, TAML employs a temporal information sharing strategy to augment the number of positive samples and include the prediction of related phenotypes or events in the meta-training phase. We demonstrate the effectiveness of TAML on multiple clinical datasets, where it consistently outperforms a range of strong baselines. We also develop a MetaEHR package for implementing both time-associated and time-independent few-shot prediction on EHR data. | ['Christopher King', 'DaCheng Tao', 'Bradley Fritz', 'Yixin Chen', 'Lecheng Kong', 'Zehao Dong', 'Muhan Zhang', 'Hao liu'] | 2023-03-05 | null | null | null | null | ['disease-prediction'] | ['medical'] | [ 4.02901888e-01 -7.30714202e-02 -6.80056393e-01 -6.05893612e-01
-1.25318217e+00 4.75297011e-02 4.24781024e-01 7.79556036e-01
-1.76940471e-01 7.01331079e-01 5.74138165e-01 -1.47830412e-01
-3.83969963e-01 -6.10842466e-01 -4.29347634e-01 -5.82154930e-01
-4.26598310e-01 5.64598322e-01 -4.84598540e-02 8.81361812e-02
-1.11844338e-01 -1.52276963e-01 -1.15306282e+00 1.13163865e+00
6.35334373e-01 7.14717507e-01 -6.04974292e-02 5.75622022e-01
5.30349649e-02 1.04455125e+00 -4.11219031e-01 -3.49918485e-01
1.64128706e-01 -4.74781245e-01 -5.49727798e-01 -2.29640976e-01
-3.63056734e-02 5.92571683e-03 -3.60357791e-01 5.30046701e-01
6.00663781e-01 -1.62639504e-03 4.77319032e-01 -1.33629537e+00
-1.61805615e-01 7.87995696e-01 -4.10002619e-01 3.52732331e-01
6.70225397e-02 7.27617294e-02 7.52792954e-01 -6.46510720e-01
7.51181304e-01 9.21914160e-01 1.11409962e+00 5.66941142e-01
-1.15226603e+00 -5.34493625e-01 -8.86436924e-03 3.18754911e-01
-1.04127836e+00 -4.16309625e-01 4.39300954e-01 -3.86817455e-01
1.04683781e+00 4.39576715e-01 5.53599834e-01 1.45557225e+00
6.92131996e-01 8.23105216e-01 7.04176903e-01 -1.82653740e-01
5.19838691e-01 -2.89096802e-01 3.73590678e-01 3.76182407e-01
-8.98293853e-02 6.20004982e-02 -6.35962307e-01 -9.54256237e-01
1.89412937e-01 1.05016232e+00 9.23128054e-02 2.00485304e-01
-1.55141568e+00 6.06727123e-01 -4.14942019e-02 1.17718562e-01
-5.20861149e-01 -1.27595598e-02 9.51197624e-01 3.61098051e-01
8.97996724e-01 4.30486739e-01 -9.99980032e-01 -1.90884739e-01
-8.50752652e-01 2.38386393e-01 6.74906075e-01 6.99707985e-01
2.75552690e-01 -3.90855044e-01 -7.72719502e-01 8.76226366e-01
-1.44943386e-01 -3.19884941e-02 8.48030508e-01 -8.43227208e-01
4.50351745e-01 6.15909517e-01 1.66896850e-01 -5.68020105e-01
-6.16470277e-01 -3.58541250e-01 -1.01659453e+00 -4.29917753e-01
2.95729697e-01 -3.29983890e-01 -9.83198941e-01 1.72382450e+00
5.92844665e-01 9.52936769e-01 2.01024950e-01 3.01053464e-01
5.78036964e-01 7.39941537e-01 5.68370819e-01 -7.81576097e-01
1.40110993e+00 -8.26910257e-01 -8.96997869e-01 3.12737003e-02
1.31902957e+00 -3.99933904e-01 4.78594184e-01 3.43638152e-01
-1.03828037e+00 -2.43758306e-01 -5.44515073e-01 4.21881936e-02
-1.91588551e-01 -1.70803726e-01 7.97838748e-01 1.36436298e-01
-3.40005189e-01 1.00836825e+00 -1.35986865e+00 -1.77083641e-01
5.31449378e-01 1.06118597e-01 -2.71894217e-01 -1.97082520e-01
-1.19858050e+00 6.42587662e-01 3.69227707e-01 -3.10132474e-01
-7.27015257e-01 -1.56277025e+00 -6.50859833e-01 -1.80181246e-02
1.41270146e-01 -1.03419518e+00 1.25647712e+00 -5.33149302e-01
-8.45187128e-01 7.37189174e-01 -4.96438622e-01 -8.25560331e-01
5.14061332e-01 -1.84016079e-01 -7.72929847e-01 -2.29974478e-01
8.72483328e-02 2.66863346e-01 6.54461980e-01 -3.60478759e-01
-8.74384582e-01 -6.41067147e-01 -7.52408683e-01 -3.96173261e-02
-1.56332284e-01 -1.68345422e-01 -2.29493454e-01 -9.95740175e-01
-9.74434242e-02 -8.73746872e-01 -7.44749069e-01 -1.64776474e-01
-3.38795692e-01 -2.22904295e-01 7.02437222e-01 -7.00219154e-01
1.55888081e+00 -2.11085916e+00 8.11573491e-03 -4.23858434e-01
5.14418662e-01 1.97105229e-01 -1.28979206e-01 7.32978761e-01
-3.85611415e-01 -2.96455380e-02 3.31088006e-02 -4.86867696e-01
-6.05428517e-01 2.30737820e-01 -3.21028262e-01 2.77066737e-01
5.55978045e-02 1.16747165e+00 -1.14094472e+00 -4.64026183e-01
1.72964856e-01 3.97903800e-01 -4.34086949e-01 1.26642793e-01
-3.51745307e-01 4.78076994e-01 -5.12844920e-01 7.37369657e-01
2.35315457e-01 -8.73156309e-01 4.33171332e-01 -3.04037154e-01
2.12863073e-01 2.55841821e-01 -5.84164858e-01 1.70773673e+00
-2.90370554e-01 1.52203113e-01 -6.63912594e-01 -1.15257633e+00
4.44602549e-01 8.03087711e-01 1.31965864e+00 -5.18245995e-01
-3.06942046e-01 -1.25808250e-02 -7.13533238e-02 -6.20664716e-01
2.52131112e-02 -2.97685266e-01 -1.11092411e-01 4.65945840e-01
-7.49452189e-02 8.17031801e-01 -1.39135122e-01 3.86148728e-02
1.77990162e+00 -1.82562217e-01 8.25452566e-01 1.39292195e-01
-2.52678931e-01 2.25049809e-01 1.30778754e+00 8.17937136e-01
-4.28131163e-01 4.13509399e-01 2.03506023e-01 -1.17160976e+00
-1.15665221e+00 -9.27437246e-01 -4.07979310e-01 1.18796635e+00
-2.93447405e-01 -7.23998785e-01 -7.20791519e-02 -7.84925044e-01
1.01470411e-01 5.46869874e-01 -8.44453812e-01 -3.82331669e-01
-4.08338308e-01 -1.54627979e+00 4.42476749e-01 5.81022263e-01
-2.03464448e-01 -7.75008142e-01 -7.56824672e-01 7.38618672e-01
-3.64426911e-01 -8.10006618e-01 -4.12923008e-01 3.78591180e-01
-1.24594843e+00 -1.08393610e+00 -6.57430649e-01 -5.22555590e-01
2.06410035e-01 2.02978671e-01 1.17309725e+00 -1.98556855e-01
-6.95082784e-01 3.30178775e-02 -4.56876218e-01 -8.45854223e-01
-4.50869590e-01 -1.54285148e-01 7.23099932e-02 -8.15617442e-02
6.15889847e-01 -5.76850772e-01 -7.77800202e-01 6.74829036e-02
-6.42656624e-01 4.51910317e-01 4.17812675e-01 1.19822848e+00
8.90021265e-01 -1.54570147e-01 1.16429090e+00 -1.32549381e+00
1.25400022e-01 -9.82214212e-01 -2.76271421e-02 5.14174461e-01
-7.98951387e-01 -1.77283157e-02 4.80387181e-01 -7.01408863e-01
-8.91776323e-01 6.88458011e-02 1.35199681e-01 -4.80767667e-01
5.27383061e-03 7.07811952e-01 2.53359348e-01 6.36744380e-01
6.59534037e-01 2.15133876e-01 7.20347315e-02 -4.77335870e-01
8.69775340e-02 6.08118773e-01 8.92562494e-02 -5.94111010e-02
3.32826860e-02 6.21892095e-01 1.52231097e-01 -4.92083341e-01
-1.35489440e+00 -8.01426113e-01 -2.62404144e-01 2.18066037e-01
7.13249862e-01 -1.30142164e+00 -5.25911689e-01 2.42653668e-01
-8.93919289e-01 -3.03504318e-01 -5.44682622e-01 6.27754748e-01
-5.02662182e-01 9.35246944e-02 -9.38770771e-01 -5.84852517e-01
-6.09470129e-01 -7.59037614e-01 1.29848528e+00 -2.27128565e-01
-6.68497980e-01 -1.25624907e+00 5.42175174e-01 2.95381993e-01
8.99790600e-02 6.22709990e-01 1.25421286e+00 -1.17565632e+00
-9.51465592e-02 -3.65663469e-01 1.32781774e-01 -3.80129993e-01
4.91896033e-01 -3.00096363e-01 -9.49199557e-01 -2.90117294e-01
6.84948545e-03 -1.99109942e-01 1.14274883e+00 7.58970082e-01
1.46611869e+00 -4.28491801e-01 -9.90425050e-01 6.53670073e-01
1.20213032e+00 3.37088019e-01 3.30649287e-01 -9.33784693e-02
5.20591974e-01 4.61262226e-01 8.00972641e-01 1.11104071e+00
4.43510503e-01 6.80621207e-01 -1.89275052e-02 -2.12459400e-01
1.79583967e-01 -2.81114113e-02 3.10085062e-02 8.38596463e-01
1.14899287e-02 -1.02478750e-01 -9.19100940e-01 6.51125848e-01
-2.51811981e+00 -1.38550007e+00 -7.12189227e-02 2.44504213e+00
1.07371187e+00 -1.84518650e-01 1.38478428e-01 -8.71107131e-02
6.15068853e-01 -1.77833922e-02 -9.07066166e-01 5.59262522e-02
-4.51614789e-04 1.42666459e-01 2.56039888e-01 -1.91592816e-02
-1.25154924e+00 3.58507454e-01 6.71162701e+00 7.42852509e-01
-8.69821787e-01 5.70197165e-01 1.18634999e+00 -5.40859461e-01
1.47263454e-02 -1.38389587e-01 -7.84695089e-01 6.04102850e-01
1.54299283e+00 -4.60864663e-01 -1.44715831e-01 7.86006749e-01
5.24475217e-01 3.56646955e-01 -1.55235302e+00 1.01815963e+00
-3.39082897e-01 -1.85289109e+00 -2.25383006e-02 7.75013939e-02
9.15329814e-01 3.85338873e-01 -6.56298772e-02 4.94149268e-01
4.03163046e-01 -7.95808077e-01 -2.84945704e-02 7.38221109e-01
5.56482434e-01 -5.07952750e-01 5.53614974e-01 3.86365801e-01
-1.15133905e+00 -3.46017450e-01 -3.06489199e-01 -7.22128479e-03
3.89952749e-01 9.98264670e-01 -1.07931292e+00 7.19336152e-01
4.31375355e-01 1.27841306e+00 -3.21124464e-01 1.08989537e+00
7.57772386e-01 7.66896784e-01 -6.56205341e-02 3.90964419e-01
-2.80620337e-01 3.32643360e-01 3.38398188e-01 1.11990035e+00
4.62089121e-01 2.33263120e-01 5.03323972e-01 3.28898877e-01
6.14658976e-03 6.33843094e-02 -6.50285006e-01 -1.10212296e-01
6.29465699e-01 1.04332900e+00 -3.56709629e-01 -6.76576018e-01
-6.03909135e-01 8.42794955e-01 3.02552640e-01 7.91655257e-02
-9.61505055e-01 1.48156080e-02 7.51669586e-01 -2.77833622e-02
-7.91841075e-02 3.46693128e-01 -2.73835272e-01 -1.49570227e+00
-2.65590042e-01 -8.71841669e-01 1.19551063e+00 -3.21596235e-01
-1.74452341e+00 1.84014022e-01 -2.16059312e-01 -1.65829146e+00
-4.24467504e-01 -4.65276018e-02 -6.07786298e-01 5.34734130e-01
-1.19145060e+00 -1.06998229e+00 2.10528821e-02 5.27043700e-01
8.90394807e-01 -2.26021502e-02 1.05604172e+00 3.32717896e-01
-6.89703643e-01 4.20927167e-01 2.47447431e-01 -8.92220959e-02
1.11386430e+00 -1.00657260e+00 3.48012328e-01 2.85779297e-01
-3.30275595e-02 5.75372696e-01 4.39725041e-01 -9.88285482e-01
-1.43146539e+00 -1.85019803e+00 1.06396997e+00 -7.79483080e-01
3.63183230e-01 8.00270066e-02 -1.15439010e+00 9.41280007e-01
-2.70240635e-01 2.33011529e-01 1.44610727e+00 4.84772295e-01
-3.08170289e-01 -1.92884117e-01 -1.05379307e+00 4.89088893e-01
8.69488597e-01 -4.64774758e-01 -6.10565901e-01 7.89647400e-01
9.43481326e-01 -1.52693540e-01 -1.29798043e+00 7.74284422e-01
6.46423221e-01 -5.57640433e-01 1.08422565e+00 -1.33816302e+00
6.17347777e-01 1.41023293e-01 -3.17836069e-02 -1.13623607e+00
-4.39378083e-01 -6.22317076e-01 -6.12287045e-01 7.28808582e-01
5.56111872e-01 -4.26788658e-01 8.22538674e-01 7.89361954e-01
-1.20106926e-02 -9.00115728e-01 -8.78005445e-01 -6.62700653e-01
-2.73260862e-01 -4.78303045e-01 4.69445646e-01 1.37752628e+00
6.03912890e-01 2.32133403e-01 -8.23962986e-01 1.64103568e-01
6.74811065e-01 3.58298481e-01 4.01447803e-01 -1.39601481e+00
-6.83362663e-01 5.75081818e-02 -2.84151137e-01 -2.03338534e-01
-1.52451992e-01 -9.98438239e-01 -2.93864280e-01 -1.17515635e+00
7.85210729e-01 -4.34908003e-01 -8.41600537e-01 7.35808372e-01
-6.53772295e-01 -5.95094524e-02 -2.40813509e-01 5.01507580e-01
-8.83169949e-01 3.23571861e-01 8.80229831e-01 -1.04626425e-01
-4.97837871e-01 3.29455614e-01 -3.31993729e-01 5.65982103e-01
6.83342040e-01 -6.70938730e-01 -4.10370767e-01 -2.89756730e-02
8.51423144e-02 8.03394020e-01 1.51595727e-01 -8.30966413e-01
1.81426823e-01 -4.11333770e-01 4.68213797e-01 -6.14947140e-01
3.74082714e-01 -4.36644495e-01 7.45530069e-01 8.73048782e-01
-7.84838557e-01 1.26002654e-01 4.98641422e-03 1.16826975e+00
-1.04635805e-01 3.09211314e-01 5.81938922e-01 -4.26516503e-01
-5.04396856e-01 7.53777325e-01 -2.08357170e-01 9.47008431e-02
1.18572319e+00 1.18652798e-01 -1.87907562e-01 -7.07187653e-02
-1.15282178e+00 2.62428850e-01 8.47263783e-02 3.88590962e-01
4.56722587e-01 -1.27393007e+00 -7.76111007e-01 -1.00183889e-01
6.13886416e-01 -3.39245796e-01 6.64110422e-01 1.10121834e+00
9.49346125e-02 2.67326385e-01 9.26595703e-02 -6.54549658e-01
-1.40956974e+00 8.93815339e-01 7.43035525e-02 -6.82325125e-01
-8.95163238e-01 3.67437273e-01 2.39116639e-01 -6.49678856e-02
1.47128671e-01 -3.45408708e-01 -2.72799120e-03 1.50049835e-01
1.02852678e+00 5.17352283e-01 1.57424718e-01 7.58737326e-02
-1.63975075e-01 -1.15355052e-01 -5.06744027e-01 1.76818073e-01
1.64306808e+00 1.27365738e-01 4.05852757e-02 9.43753719e-01
1.10746086e+00 -5.48431933e-01 -1.04519713e+00 -3.97517741e-01
3.02127659e-01 -2.54838407e-01 -4.85654399e-02 -8.47387552e-01
-5.85978985e-01 6.57900929e-01 5.69942832e-01 -6.81698620e-02
1.09587562e+00 -1.35128424e-01 1.05058849e+00 3.43749940e-01
3.73317778e-01 -7.70436049e-01 -2.86103114e-02 1.75869212e-01
1.82948112e-01 -1.50359368e+00 -1.12681553e-01 -2.77155757e-01
-7.38304615e-01 8.82220984e-01 3.56920332e-01 3.15851003e-01
8.65200877e-01 3.28093380e-01 -2.70444099e-02 -6.82784542e-02
-1.88219178e+00 2.79674619e-01 1.46458507e-01 4.36727077e-01
7.15399742e-01 3.17504942e-01 -3.16344053e-01 8.14608395e-01
5.28938234e-01 5.70260882e-01 2.71102965e-01 1.11938798e+00
-2.16332391e-01 -1.23255610e+00 7.11151659e-02 1.04199541e+00
-9.18924749e-01 -2.56678313e-01 2.29437485e-01 2.39900023e-01
-2.04804987e-02 7.54171550e-01 5.28271459e-02 -3.78124714e-01
1.48320496e-02 5.83096802e-01 1.42244652e-01 -8.53603899e-01
-7.51255751e-01 2.37917677e-01 1.16603069e-01 -8.00166130e-01
-2.49201447e-01 -8.04565132e-01 -1.13341546e+00 1.50626001e-03
-3.21337320e-02 6.82781339e-02 7.35645592e-02 1.10384655e+00
1.12306893e+00 6.20921731e-01 6.03860676e-01 -3.77841294e-01
-7.45636106e-01 -8.15112770e-01 -5.24886250e-01 7.20926285e-01
4.32703167e-01 -2.51180112e-01 8.61398950e-02 3.64450485e-01] | [7.909491539001465, 6.117120265960693] |
7e18906c-99a2-40d1-8575-b6fc17605f4b | embodied-artificial-intelligence-through | 1704.01407 | null | http://arxiv.org/abs/1704.01407v3 | http://arxiv.org/pdf/1704.01407v3.pdf | Embodied Artificial Intelligence through Distributed Adaptive Control: An Integrated Framework | In this paper, we argue that the future of Artificial Intelligence research
resides in two keywords: integration and embodiment. We support this claim by
analyzing the recent advances of the field. Regarding integration, we note that
the most impactful recent contributions have been made possible through the
integration of recent Machine Learning methods (based in particular on Deep
Learning and Recurrent Neural Networks) with more traditional ones (e.g.
Monte-Carlo tree search, goal babbling exploration or addressable memory
systems). Regarding embodiment, we note that the traditional benchmark tasks
(e.g. visual classification or board games) are becoming obsolete as
state-of-the-art learning algorithms approach or even surpass human performance
in most of them, having recently encouraged the development of first-person 3D
game platforms embedding realistic physics. Building upon this analysis, we
first propose an embodied cognitive architecture integrating heterogenous
sub-fields of Artificial Intelligence into a unified framework. We demonstrate
the utility of our approach by showing how major contributions of the field can
be expressed within the proposed framework. We then claim that benchmarking
environments need to reproduce ecologically-valid conditions for bootstrapping
the acquisition of increasingly complex cognitive skills through the concept of
a cognitive arms race between embodied agents. | ['Martì Sanchez-Fibla', 'Jordi-Ysard Puigbò', 'Clément Moulin-Frier', 'Paul F. M. J. Verschure', 'Xerxes D. Arsiwalla'] | 2017-04-05 | null | null | null | null | ['board-games'] | ['playing-games'] | [ 9.68187675e-03 4.79917914e-01 4.31529105e-01 3.14561903e-01
4.10423502e-02 -6.23762965e-01 8.73179197e-01 1.53153362e-02
-5.31352341e-01 7.02066541e-01 1.12805642e-01 -3.11796159e-01
-7.33886242e-01 -8.83518338e-01 -6.33348823e-01 -4.35991615e-01
-3.24296504e-01 3.98403466e-01 1.42917931e-01 -7.80827105e-01
3.68322760e-01 5.34888744e-01 -2.12305069e+00 1.02562578e-02
8.02412748e-01 8.30693424e-01 2.45808139e-01 5.49388409e-01
6.20278269e-02 9.98378336e-01 -4.99534905e-01 -3.60989898e-01
1.41833559e-01 -4.82968152e-01 -8.05241048e-01 -4.74344462e-01
-2.43785352e-01 -1.05328925e-01 -5.08626774e-02 9.63680565e-01
5.21165907e-01 3.40620846e-01 5.29399216e-01 -1.00057745e+00
-8.62511694e-01 4.79265600e-01 8.10873583e-02 7.42166266e-02
4.60966140e-01 3.02957684e-01 6.70275986e-01 -5.63891351e-01
6.55570924e-01 1.03870261e+00 8.08106482e-01 6.61154509e-01
-9.89581525e-01 -1.87149987e-01 2.79105455e-01 4.21919614e-01
-1.05670810e+00 -3.57124925e-01 8.56643200e-01 -5.76678693e-01
1.23471582e+00 3.67959827e-01 1.22560561e+00 1.34140229e+00
1.17852166e-01 5.84228516e-01 1.51460493e+00 -7.36402631e-01
5.91611624e-01 2.76899666e-01 2.90942937e-02 7.70322263e-01
3.06409627e-01 6.27802908e-01 -6.00118458e-01 2.70668626e-01
8.32668126e-01 -4.82942075e-01 5.46440072e-02 -5.57397783e-01
-1.08419120e+00 7.54722714e-01 3.44814003e-01 7.41688609e-01
-5.70544779e-01 2.43027106e-01 3.72823566e-01 1.43707335e-01
2.41577387e-01 6.56265974e-01 8.83913040e-03 -4.24337476e-01
-8.13922048e-01 4.64834571e-01 8.93648803e-01 4.92195100e-01
3.76340538e-01 3.16300899e-01 9.56561491e-02 3.41372281e-01
4.84774828e-01 1.29072517e-01 7.36095130e-01 -9.32888687e-01
-2.14178994e-01 6.48838103e-01 2.57344306e-01 -1.06766009e+00
-6.82337403e-01 -7.81686962e-01 -4.40369397e-01 8.84550452e-01
2.34991521e-01 -5.83700016e-02 -4.83110875e-01 1.86400402e+00
3.34511340e-01 1.03812754e-01 2.22450674e-01 8.12839389e-01
6.50343478e-01 2.01930404e-01 2.67064691e-01 1.17580026e-01
1.46048522e+00 -8.00219953e-01 -5.71266592e-01 -1.99012890e-01
4.33645457e-01 -1.10883087e-01 1.02214944e+00 6.03019178e-01
-1.20396030e+00 -6.49498940e-01 -1.34727418e+00 6.18418418e-02
-9.31605935e-01 -2.37164333e-01 1.01616454e+00 1.14455843e+00
-1.40817094e+00 7.07697451e-01 -8.30941796e-01 -6.72971308e-01
2.01541454e-01 2.64533073e-01 -1.96032330e-01 6.55652940e-01
-1.22430205e+00 1.48819768e+00 6.13418579e-01 2.40348548e-01
-1.01207185e+00 -4.46440756e-01 -5.89530587e-01 1.65726580e-02
4.57448512e-01 -1.09004676e+00 1.10548425e+00 -1.14515853e+00
-1.93259346e+00 1.22313964e+00 5.74321330e-01 -7.03254521e-01
6.43328905e-01 -4.14197326e-01 -2.88417846e-01 1.52046187e-02
-4.49924856e-01 5.49143016e-01 4.14423317e-01 -1.31852627e+00
-2.54758477e-01 -2.61052161e-01 4.71167833e-01 1.21634297e-01
-3.25711131e-01 9.09507461e-03 2.66001076e-01 -5.91922939e-01
-2.25059092e-01 -9.20271158e-01 -1.50252327e-01 -1.72817439e-01
-1.15510775e-02 -1.73404977e-01 4.38490629e-01 -4.83211517e-01
1.12265062e+00 -1.85971475e+00 6.88459456e-01 -1.42195448e-01
4.10003394e-01 3.86242419e-01 1.66137479e-02 7.15031922e-01
-1.02956124e-01 7.53756464e-02 2.77558509e-02 -2.30917141e-01
7.10714221e-01 1.40414352e-03 -2.49066398e-01 2.17778549e-01
-2.61662990e-01 1.16339445e+00 -8.89672160e-01 -2.12378457e-01
5.42312622e-01 6.05871320e-01 -5.08727551e-01 -4.27399613e-02
-2.93161064e-01 5.43192089e-01 -3.69649202e-01 3.00982386e-01
1.45001277e-01 5.63639728e-03 7.85410255e-02 2.08900213e-01
-5.41606009e-01 2.12742418e-01 -1.07940030e+00 2.17704725e+00
-5.14696300e-01 3.73723596e-01 -1.79038584e-01 -1.11181521e+00
8.67416263e-01 2.96260238e-01 3.00316840e-01 -7.93147981e-01
4.13189203e-01 2.47377440e-01 1.61148816e-01 -5.67199707e-01
5.83308637e-01 -3.26871067e-01 8.99346992e-02 4.86456424e-01
1.05887488e-01 -1.88253298e-01 -9.60024148e-02 -3.16369116e-01
8.19812894e-01 9.15487409e-01 6.09978795e-01 -5.03977418e-01
5.49002349e-01 -3.68349850e-02 -7.73301050e-02 8.45086098e-01
-3.84179413e-01 9.59289148e-02 2.28896201e-01 -4.76764798e-01
-9.56637800e-01 -1.02213717e+00 1.26759499e-01 1.17804313e+00
-5.05389236e-02 -2.86015451e-01 -1.28064084e+00 -9.68113840e-02
-4.03865635e-01 9.54987824e-01 -9.42374587e-01 -2.86005437e-01
-4.13794428e-01 -7.02945888e-01 7.81156123e-01 4.07505572e-01
6.85390413e-01 -1.52305508e+00 -1.78633118e+00 3.23601037e-01
3.42560887e-01 -7.30192780e-01 9.12595212e-01 2.90289134e-01
-8.20015907e-01 -7.91004479e-01 -5.74396610e-01 -3.92140210e-01
-1.05844758e-01 -6.56429455e-02 1.21380353e+00 3.57664436e-01
-2.72729009e-01 7.66237497e-01 -5.79938352e-01 -5.66780686e-01
-4.37936008e-01 8.14835355e-02 1.77396283e-01 -4.71385598e-01
3.13744187e-01 -1.19730604e+00 -7.22278595e-01 -1.52960658e-01
-9.30093646e-01 4.16532487e-01 4.47382987e-01 6.05532527e-01
3.58345807e-02 -1.64268583e-01 5.30518055e-01 -3.61582190e-01
8.76586437e-01 -4.59622711e-01 -3.98975730e-01 3.00379097e-01
-4.87456262e-01 1.17764264e-01 3.34410965e-01 -3.12227637e-01
-1.13683474e+00 -5.20478368e-01 -2.69119501e-01 -6.56488985e-02
-5.08769095e-01 6.28909528e-01 3.76853757e-02 -2.89357811e-01
7.27020919e-01 4.46937531e-01 2.24694535e-02 -2.60714114e-01
6.47809923e-01 3.44972730e-01 5.86871684e-01 -9.69747603e-01
6.69594526e-01 3.54661196e-01 5.88085167e-02 -6.48977220e-01
-2.95498967e-01 3.88448030e-01 -6.13648713e-01 -6.53197050e-01
1.10480976e+00 -5.32366931e-01 -1.16488338e+00 4.35721457e-01
-1.10131836e+00 -5.21055281e-01 -6.31277323e-01 5.48623681e-01
-1.05411649e+00 2.03424811e-01 -4.29790467e-01 -1.27053499e+00
-7.40402713e-02 -1.05609977e+00 8.33898783e-01 3.87096256e-01
-6.28443897e-01 -1.11636782e+00 5.27160943e-01 2.96629041e-01
6.59150302e-01 5.92628241e-01 1.09015548e+00 -6.23554766e-01
-4.84645426e-01 2.58819796e-02 3.17581832e-01 -7.00305626e-02
-4.00811613e-01 -5.40461913e-02 -1.17038214e+00 -2.12793890e-02
2.97488928e-01 -4.88884747e-01 5.93827248e-01 9.80274081e-02
5.21004319e-01 2.43123949e-01 -1.34475708e-01 3.75334144e-01
1.47617018e+00 4.92811024e-01 8.55949283e-01 9.13761199e-01
1.00159131e-01 1.08342052e+00 2.46965110e-01 4.07440603e-01
4.15734529e-01 8.68942380e-01 6.70182407e-01 1.43445790e-01
-6.48775548e-02 -2.78185189e-01 3.26390803e-01 6.61319971e-01
-8.00392866e-01 -4.61903252e-02 -1.08692992e+00 2.23204300e-01
-1.92950797e+00 -1.11534417e+00 2.90927202e-01 2.10196328e+00
2.98595190e-01 1.70279190e-01 3.53579432e-01 2.60421336e-01
3.67920190e-01 9.32777077e-02 -4.22937423e-01 -6.46547079e-01
-4.87951264e-02 4.59016532e-01 -3.48097652e-01 3.76118600e-01
-8.98593247e-01 1.03488231e+00 6.48019934e+00 4.90459085e-01
-8.92722845e-01 2.87353694e-01 2.82260716e-01 -9.06273872e-02
-2.66836554e-01 -3.14486653e-01 -2.44982824e-01 2.04674974e-01
9.33028340e-01 -5.86633496e-02 7.71991432e-01 8.34797561e-01
1.75433874e-01 -2.55791128e-01 -1.05672729e+00 8.44639003e-01
3.72366533e-02 -1.17317688e+00 -1.39554277e-01 1.65808588e-01
3.24014515e-01 -2.53586322e-01 4.35884416e-01 6.28080487e-01
3.94153655e-01 -1.44997752e+00 1.17375088e+00 8.61488223e-01
2.21673414e-01 -5.42256534e-01 3.80611479e-01 3.67968112e-01
-9.60133553e-01 -2.51956135e-01 -1.17636316e-01 -7.61324465e-01
3.06714792e-02 8.01028460e-02 -7.39700533e-03 7.35399425e-01
6.59434736e-01 3.06995034e-01 -4.92031187e-01 9.65662599e-01
-1.80407420e-01 1.37457952e-01 -1.28616974e-01 -4.11794573e-01
4.19858783e-01 -3.38313192e-01 7.12764442e-01 8.39512527e-01
5.13333619e-01 1.15311168e-01 -3.80333990e-01 1.30672252e+00
6.08247161e-01 4.45712842e-02 -8.72550249e-01 -1.32036120e-01
2.66873389e-01 1.18505800e+00 -1.07318699e+00 -5.69251478e-02
-2.77450204e-01 8.20551455e-01 4.55839664e-01 2.99554467e-01
-9.87126708e-01 5.14238067e-02 6.22301340e-01 -2.95768112e-01
2.78827429e-01 -3.67581338e-01 -4.10013348e-01 -9.42544103e-01
-1.95464894e-01 -1.08193231e+00 -1.38456702e-01 -1.01272643e+00
-8.54562521e-01 7.86607027e-01 4.26011235e-02 -8.58956337e-01
-5.15782058e-01 -9.30301964e-01 -4.03195024e-01 6.18999660e-01
-1.10160017e+00 -1.20201600e+00 -3.47952038e-01 4.05438095e-01
1.11125387e-01 -3.28214794e-01 1.28119862e+00 -9.79687646e-02
-4.14666057e-01 9.37768668e-02 -2.07346454e-01 -3.05712074e-01
-1.11509867e-01 -1.17816484e+00 3.69640976e-01 5.80771744e-01
2.70431042e-01 6.62998438e-01 9.27507281e-01 -3.83144855e-01
-1.23881507e+00 -1.81996256e-01 3.56661379e-01 -5.76358557e-01
7.01181710e-01 -5.64619601e-01 -5.66695273e-01 6.18445337e-01
5.96978188e-01 -5.44964135e-01 7.25788176e-01 2.42052078e-01
-1.26717150e-01 3.40943962e-01 -9.82740343e-01 9.68493164e-01
1.54287779e+00 -7.25314856e-01 -1.16535842e+00 -1.44223109e-01
5.21791160e-01 -2.27422610e-01 -6.49940729e-01 2.72603393e-01
8.58118594e-01 -1.70485675e+00 1.14895988e+00 -8.03546190e-01
3.94428194e-01 -1.62486851e-01 -1.50060803e-01 -1.18909955e+00
-2.74529159e-01 -7.41600752e-01 -2.42046490e-01 9.31431472e-01
-6.00175336e-02 -6.85374558e-01 6.72153652e-01 5.53585410e-01
-2.12531611e-01 -7.12051153e-01 -1.14529753e+00 -7.22082436e-01
3.53624910e-01 -6.43740237e-01 5.83335996e-01 8.02433789e-01
3.86615872e-01 8.09496269e-02 -1.65250286e-01 -2.25411490e-01
2.80075848e-01 -9.73154232e-02 6.30073428e-01 -1.52793765e+00
-5.42914271e-01 -8.59860778e-01 -6.81771874e-01 -5.17485082e-01
3.42828363e-01 -7.61598468e-01 -2.86938906e-01 -1.41851389e+00
7.76497647e-02 -1.50641188e-01 -4.25290525e-01 2.27959707e-01
1.78425357e-01 2.00843900e-01 4.78766263e-01 -1.13974601e-01
-6.11596107e-01 5.98193467e-01 9.50636029e-01 3.16643924e-01
-1.14295065e-01 -4.78935361e-01 -9.40085888e-01 1.03615522e+00
8.60450506e-01 -1.20213889e-01 -5.29546201e-01 -2.17681736e-01
8.51262391e-01 -1.14454716e-01 9.17381227e-01 -1.63782823e+00
2.95846909e-01 1.39183402e-01 1.95597723e-01 -4.22216840e-02
3.85281444e-01 -9.37811613e-01 5.35111487e-01 8.16367745e-01
-9.42487270e-02 1.78321198e-01 5.01531780e-01 3.66182894e-01
7.29769170e-02 -3.97785485e-01 3.38983506e-01 -3.84556085e-01
-9.41940546e-01 -5.20437062e-01 -7.26850867e-01 -2.27101311e-01
1.27763557e+00 -7.91449249e-01 -3.18946004e-01 -2.00855568e-01
-9.21151519e-01 -3.23547542e-01 3.72180641e-01 4.27308708e-01
3.62956911e-01 -1.00466275e+00 -4.52551991e-01 -1.86526522e-01
-1.66046754e-01 -6.68421865e-01 4.39550072e-01 7.27839708e-01
-6.57101333e-01 6.60075605e-01 -8.08344960e-01 -1.25097826e-01
-9.32409525e-01 6.56086504e-01 7.19620526e-01 -3.18443537e-01
-5.43619215e-01 6.34254813e-01 1.77425995e-01 -3.37663502e-01
1.58755377e-01 -2.30542779e-01 -4.58774686e-01 1.16942890e-01
3.97047013e-01 5.59401870e-01 -4.82365601e-02 -3.47774982e-01
-4.10793126e-01 6.14741504e-01 4.12015170e-01 -5.02035618e-01
1.51893902e+00 6.21288009e-02 -8.40637237e-02 6.80053294e-01
3.07214767e-01 -5.42985648e-02 -1.12944651e+00 3.99843663e-01
3.18659134e-02 2.41233289e-01 -1.14370603e-02 -1.15416312e+00
-5.53420663e-01 9.64739501e-01 8.84169996e-01 4.42379028e-01
1.17046022e+00 -1.58332065e-01 1.02258302e-01 5.73878765e-01
7.69162655e-01 -1.01912653e+00 1.01601876e-01 3.84159148e-01
1.08471847e+00 -7.08748579e-01 -9.44042429e-02 -7.38871796e-03
-3.12466025e-01 1.16530180e+00 3.91602278e-01 -2.26698369e-01
3.60848367e-01 3.05912673e-01 -1.90331325e-01 -4.56112355e-01
-5.90159118e-01 -5.34513533e-01 1.44833222e-01 9.92560267e-01
4.70071405e-01 1.12062125e-02 -4.44495142e-01 9.07649696e-01
-6.00871861e-01 3.46807390e-01 2.53238469e-01 8.89300644e-01
-5.05696952e-01 -9.19262826e-01 -4.22697514e-01 -2.91474730e-01
-8.57567713e-02 -1.00968830e-01 -4.50036317e-01 1.32895112e+00
4.68523353e-01 7.47219503e-01 -1.99338585e-01 -4.12725955e-01
2.63662189e-01 3.95540565e-01 9.59834874e-01 -3.23930293e-01
-1.06343710e+00 -4.71285611e-01 7.59261698e-02 -5.04795790e-01
-7.52081156e-01 -4.55645025e-01 -8.69646847e-01 -4.50230926e-01
1.26413271e-01 2.12333146e-02 7.13086545e-01 9.02888775e-01
2.25272581e-01 8.85371387e-01 -3.48492444e-01 -1.19215024e+00
-4.24151033e-01 -8.84681582e-01 -5.54445386e-01 -8.49386230e-02
-1.69334084e-01 -1.11647427e+00 -1.24795653e-01 -1.89190239e-01] | [4.283015727996826, 1.4476736783981323] |
fd6bcfb2-0722-407c-896c-35476aca43a4 | query-reformulation-using-query-history-for | 2005.02230 | null | https://arxiv.org/abs/2005.02230v2 | https://arxiv.org/pdf/2005.02230v2.pdf | Multi-Stage Conversational Passage Retrieval: An Approach to Fusing Term Importance Estimation and Neural Query Rewriting | Conversational search plays a vital role in conversational information seeking. As queries in information seeking dialogues are ambiguous for traditional ad-hoc information retrieval (IR) systems due to the coreference and omission resolution problems inherent in natural language dialogue, resolving these ambiguities is crucial. In this paper, we tackle conversational passage retrieval (ConvPR), an important component of conversational search, by addressing query ambiguities with query reformulation integrated into a multi-stage ad-hoc IR system. Specifically, we propose two conversational query reformulation (CQR) methods: (1) term importance estimation and (2) neural query rewriting. For the former, we expand conversational queries using important terms extracted from the conversational context with frequency-based signals. For the latter, we reformulate conversational queries into natural, standalone, human-understandable queries with a pretrained sequence-tosequence model. Detailed analyses of the two CQR methods are provided quantitatively and qualitatively, explaining their advantages, disadvantages, and distinct behaviors. Moreover, to leverage the strengths of both CQR methods, we propose combining their output with reciprocal rank fusion, yielding state-of-the-art retrieval effectiveness, 30% improvement in terms of NDCG@3 compared to the best submission of TREC CAsT 2019. | ['Chuan-Ju Wang', 'Ming-Feng Tsai', 'Jheng-Hong Yang', 'Sheng-Chieh Lin', 'Rodrigo Nogueira', 'Jimmy Lin'] | 2020-05-05 | null | null | null | null | ['ad-hoc-information-retrieval', 'conversational-search'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.52172875e-01 1.99320480e-01 -3.37950438e-01 -2.01576531e-01
-1.62726164e+00 -8.95734429e-01 1.12177992e+00 2.89790839e-01
-8.13818812e-01 7.28056014e-01 9.15736377e-01 -5.17227590e-01
-5.57540536e-01 -1.54355377e-01 -1.46369770e-01 -3.39267761e-01
-6.66007400e-02 9.52265203e-01 1.55453205e-01 -1.07795763e+00
5.95813096e-01 3.30575287e-01 -1.48110712e+00 6.40797853e-01
1.09301937e+00 8.60524535e-01 1.73205703e-01 1.05677962e+00
-2.53960401e-01 8.51993799e-01 -7.31608152e-01 -4.41434860e-01
-2.61535078e-01 -3.78756136e-01 -1.51205575e+00 -5.97709119e-01
2.76850853e-02 -2.84373641e-01 -3.91160339e-01 7.00106382e-01
7.90290594e-01 4.84349102e-01 5.78536928e-01 -8.43910277e-01
-4.57324892e-01 5.93915582e-01 -1.26073092e-01 3.71489316e-01
1.08539116e+00 -1.71107799e-01 1.58202064e+00 -8.56970131e-01
5.98362982e-01 1.63856018e+00 9.30596441e-02 6.54972434e-01
-1.06201696e+00 -1.99028477e-01 8.44703522e-03 4.32291031e-01
-1.37093329e+00 -6.77350283e-01 6.05837524e-01 -1.17787449e-02
1.29979885e+00 9.00546193e-01 2.89912164e-01 1.04154050e+00
-1.37056664e-01 1.11951256e+00 4.85109270e-01 -5.43554127e-01
2.40463056e-02 4.11360115e-02 2.81213194e-01 1.22007020e-01
-4.63265806e-01 7.09141791e-03 -6.77052438e-01 -7.18226612e-01
6.80259690e-02 -2.11358041e-01 -4.50386763e-01 1.90027446e-01
-1.02280867e+00 9.37299252e-01 4.64585125e-01 4.47369665e-01
-5.07348835e-01 -1.05165862e-01 4.26988751e-01 6.44282639e-01
2.46029839e-01 9.23402131e-01 -2.25376084e-01 -3.11192393e-01
-6.05721533e-01 7.17661500e-01 1.12479722e+00 8.03290606e-01
4.71066564e-01 -9.01582539e-01 -9.22470510e-01 1.34770119e+00
4.26379293e-01 4.88500863e-01 5.75266480e-01 -1.09422731e+00
4.93078262e-01 5.39209902e-01 2.18049139e-01 -9.81298804e-01
-4.03457850e-01 -2.46128291e-01 -5.79810798e-01 -8.63348901e-01
1.27837822e-01 1.60648692e-02 -4.84210551e-01 1.60489237e+00
1.54485568e-01 -5.27393401e-01 4.09976393e-01 8.96792114e-01
1.01238835e+00 6.24225080e-01 -1.00970101e-02 -4.86456305e-01
1.89015722e+00 -1.02591658e+00 -8.82796645e-01 -1.10997379e-01
5.16825974e-01 -1.05894387e+00 8.69799912e-01 7.16934949e-02
-1.12347603e+00 -1.51281551e-01 -6.59795225e-01 -5.72165728e-01
-2.48211384e-01 -2.61482924e-01 4.26059455e-01 3.17193985e-01
-1.34923160e+00 3.40651162e-02 -3.05382907e-01 -3.93753290e-01
-4.04138356e-01 2.52836853e-01 1.62816539e-01 -1.14805825e-01
-1.96292627e+00 1.04195392e+00 1.00392289e-02 1.04185320e-01
-5.44224024e-01 -5.14828920e-01 -6.06687903e-01 3.30671698e-01
6.39015079e-01 -7.37089157e-01 1.95526862e+00 -1.37295455e-01
-1.38500488e+00 6.53617084e-01 -6.04926229e-01 -4.19465601e-01
2.75861293e-01 -3.63139719e-01 -3.17517340e-01 3.86990786e-01
5.48236109e-02 7.31839001e-01 3.80380183e-01 -1.10770774e+00
-6.74710929e-01 -2.71324128e-01 3.58736008e-01 8.95236433e-01
2.37405058e-02 2.34762639e-01 -1.00786567e+00 -4.33981657e-01
3.94844748e-02 -9.48640764e-01 -1.26807183e-01 -4.77073252e-01
-3.49529326e-01 -9.96662617e-01 5.53147435e-01 -7.28115678e-01
1.64318585e+00 -1.89599562e+00 2.12857202e-01 1.86882138e-01
2.21383497e-01 3.71865332e-01 -4.27349448e-01 1.02085149e+00
3.88843149e-01 1.04909651e-01 -1.10559329e-01 -2.92068243e-01
2.23911211e-01 1.36786085e-02 -4.15810525e-01 -2.42666915e-01
9.87606272e-02 1.24997032e+00 -1.12776363e+00 -6.55397832e-01
-2.01053336e-01 4.29030865e-01 -5.70525348e-01 3.61535013e-01
-3.27597350e-01 4.06455964e-01 -7.75729775e-01 6.07023597e-01
1.20685309e-01 -1.57657295e-01 2.56453902e-01 -1.21899530e-01
1.09993510e-01 9.29174960e-01 -5.66703856e-01 1.90823781e+00
-4.65138555e-01 6.39619410e-01 1.41180307e-01 -8.04254532e-01
5.11137724e-01 6.31122351e-01 5.50653756e-01 -1.20585251e+00
-1.58573300e-01 2.67426103e-01 -2.89644361e-01 -5.25570035e-01
1.24553192e+00 1.91903487e-01 -3.74051124e-01 7.96522737e-01
-1.16158873e-01 -4.29848433e-01 3.33114058e-01 6.03210270e-01
1.17100000e+00 -4.34963226e-01 2.82685846e-01 4.68645729e-02
8.67578745e-01 1.26546035e-02 3.19661163e-02 1.21958923e+00
-1.87230423e-01 5.49517334e-01 3.11322778e-01 -1.11563681e-02
-4.99698281e-01 -7.11770654e-01 1.94385070e-02 1.58123112e+00
1.60172209e-01 -3.81595343e-01 -5.49055398e-01 -4.71498281e-01
-6.62079901e-02 7.46081412e-01 -2.67472923e-01 -2.54714519e-01
-8.04553211e-01 -5.01379192e-01 7.99669564e-01 -2.24965382e-02
3.96334618e-01 -1.28849757e+00 -2.38089487e-01 2.30841309e-01
-1.02212703e+00 -1.02416217e+00 -9.45712805e-01 -1.25476927e-01
-4.41884696e-01 -1.11835337e+00 -1.13770103e+00 -6.60401642e-01
-6.83704466e-02 7.49403238e-01 1.35380840e+00 3.61738861e-01
-2.20485672e-01 8.46379280e-01 -5.36158919e-01 -2.30573758e-01
-5.25026262e-01 4.14572239e-01 -1.09548792e-01 -2.90833592e-01
5.47187269e-01 -3.29839736e-01 -9.69258308e-01 3.63800287e-01
-1.02319860e+00 -3.53325784e-01 7.11558700e-01 9.79348361e-01
-2.99708173e-03 -5.48015535e-01 7.67842472e-01 -4.91982400e-01
1.77811921e+00 -3.74620855e-01 -2.98381090e-01 7.89629519e-01
-8.63086522e-01 2.59338975e-01 -1.69532523e-01 -1.87483072e-01
-1.29704428e+00 -5.35829008e-01 -1.98677331e-01 2.32302532e-01
2.34879181e-01 8.91227841e-01 3.06220680e-01 2.31300399e-01
9.05128241e-01 3.92126590e-01 -1.69228204e-02 -3.45286608e-01
6.94291890e-01 1.14091516e+00 4.39939499e-01 -6.66568339e-01
3.43694478e-01 -9.23723131e-02 -5.59267581e-01 -9.52659249e-01
-6.69446051e-01 -1.28915024e+00 -1.70198798e-01 -2.30035305e-01
7.75245905e-01 -7.25681841e-01 -1.31605101e+00 7.29810223e-02
-1.52752388e+00 1.49594039e-01 -1.40398201e-02 4.03337598e-01
-5.39539456e-01 6.88939631e-01 -7.47227967e-01 -1.23876190e+00
-8.61239791e-01 -1.12627614e+00 1.36923158e+00 2.88171321e-01
-5.40438771e-01 -7.61043131e-01 4.13566977e-01 9.34450805e-01
6.88584805e-01 -5.55364430e-01 1.01880753e+00 -1.02539754e+00
-3.53953987e-01 -1.06394306e-01 -3.86058718e-01 -6.78404719e-02
9.58663002e-02 -6.38216555e-01 -1.01262236e+00 -2.07438260e-01
-5.35463877e-02 -5.42615414e-01 9.01106179e-01 7.45824799e-02
5.49097240e-01 -4.92437631e-01 -3.29753846e-01 -2.49089167e-01
7.36685216e-01 4.49148327e-01 5.61697245e-01 7.49118328e-02
8.05594102e-02 9.95638609e-01 5.79606652e-01 2.51937270e-01
5.77811062e-01 9.61875260e-01 -7.00192750e-02 1.96430996e-01
2.88391337e-02 -1.78151891e-01 1.35364980e-01 8.20325196e-01
5.35815842e-02 -5.50202727e-01 -8.22056234e-01 4.06477034e-01
-2.01940417e+00 -9.84642863e-01 4.52444226e-01 2.00827956e+00
1.29206407e+00 -1.69500872e-01 -7.79350773e-02 -3.06923240e-01
5.67751229e-01 2.58298218e-01 -6.59983516e-01 -3.75175804e-01
-5.78792170e-02 1.17807798e-02 -1.09944731e-01 9.36718643e-01
-7.50901043e-01 8.38584900e-01 6.32665825e+00 8.14555287e-01
-6.41017199e-01 -1.78229198e-01 3.56252313e-01 4.28189039e-02
-6.11210048e-01 -5.94753399e-02 -6.45548344e-01 -7.28159100e-02
9.17696536e-01 -4.79468852e-01 8.96040559e-01 3.77752870e-01
8.97978321e-02 -1.76915780e-01 -1.20669866e+00 9.56962585e-01
1.25243053e-01 -1.25081706e+00 2.73262948e-01 -8.38414058e-02
2.72897512e-01 1.13756254e-01 -2.43172809e-01 8.88410985e-01
3.97250712e-01 -9.12416756e-01 -3.27995582e-03 7.17150331e-01
3.72457981e-01 -4.97013539e-01 7.60384738e-01 5.15783489e-01
-9.49442148e-01 1.64746563e-03 8.55429471e-02 3.78855407e-01
3.16348612e-01 1.81432903e-01 -9.77534473e-01 8.49325836e-01
5.05563319e-01 1.86213851e-01 -3.80704820e-01 7.73830295e-01
4.03514616e-02 2.09175944e-01 -3.06015104e-01 -3.76117408e-01
3.33563596e-01 -1.32121826e-02 9.26246107e-01 1.30963755e+00
-6.15709685e-02 4.50449854e-01 -2.88299099e-02 6.60085261e-01
-3.75249654e-01 1.09884225e-01 -3.64827693e-01 -3.06262136e-01
7.86783516e-01 9.89380479e-01 -7.64515698e-02 -3.14740956e-01
-1.44387305e-01 9.54681575e-01 2.05234602e-01 6.43036783e-01
-8.90879035e-02 -6.03325844e-01 6.08736336e-01 -6.03541553e-01
-2.62964666e-01 1.45064697e-01 3.71985972e-01 -1.08555651e+00
2.34762609e-01 -1.34423566e+00 6.93814933e-01 -6.05170250e-01
-1.23558426e+00 6.05249166e-01 2.47347787e-01 -9.33847487e-01
-1.12855911e+00 -1.83912106e-02 -2.83729821e-01 1.09888184e+00
-1.74315274e+00 -6.89618409e-01 3.54303000e-03 4.58784312e-01
8.98653209e-01 -8.72335117e-03 1.03334272e+00 3.54798317e-01
-5.22567704e-02 7.84095764e-01 3.92893143e-03 1.99456047e-02
8.60939384e-01 -1.10060561e+00 3.06050718e-01 2.85671771e-01
2.95453388e-02 1.09201515e+00 5.91478467e-01 -3.35375011e-01
-1.69777465e+00 -4.53103304e-01 1.45115840e+00 -3.99350613e-01
4.81036484e-01 -1.72896937e-01 -1.02373481e+00 1.36133879e-01
4.78330851e-01 -8.00709069e-01 6.90678120e-01 4.68366444e-01
-2.43875638e-01 1.15574606e-01 -8.22432935e-01 9.18467402e-01
7.49307096e-01 -1.08073294e+00 -1.04516566e+00 6.01263821e-01
1.30642629e+00 -3.52960795e-01 -8.16922188e-01 4.51617271e-01
6.89902723e-01 -6.42218709e-01 1.36657608e+00 -6.16989672e-01
-9.45959538e-02 1.48689911e-01 -1.90535083e-01 -1.05534136e+00
1.45085854e-02 -1.17384374e+00 -7.63627663e-02 8.95448864e-01
6.89245462e-01 -5.26125908e-01 3.63371223e-01 8.47353101e-01
-6.05688579e-02 -5.20540297e-01 -1.05583906e+00 -2.65288770e-01
-1.44422064e-02 -2.10317075e-01 5.65383077e-01 7.49869227e-01
6.23504341e-01 1.07189703e+00 -1.54659539e-01 -2.14327604e-01
1.33890316e-01 3.65629718e-02 5.45835555e-01 -1.21787846e+00
-2.16406211e-01 -8.50856662e-01 3.35297346e-01 -1.73318911e+00
-2.03916021e-02 -7.16063321e-01 4.48082387e-01 -1.54291213e+00
2.34484181e-01 -2.65031606e-01 -2.20119238e-01 1.79951355e-01
-4.55069602e-01 -2.99047381e-01 7.73859695e-02 5.69941819e-01
-9.25799310e-01 8.55033457e-01 1.20107865e+00 -4.66646165e-01
-6.28312290e-01 1.68937236e-01 -9.42517638e-01 9.28627476e-02
3.06550682e-01 -1.60915583e-01 -6.19657099e-01 -3.00196648e-01
4.46041286e-01 8.14193547e-01 -2.72495747e-02 -3.10112625e-01
7.87401795e-01 -3.03386394e-02 -4.13245678e-01 -8.15640926e-01
5.90930700e-01 -3.70135427e-01 -5.20075798e-01 3.54086578e-01
-1.20487106e+00 1.62285581e-01 3.50796357e-02 7.22572505e-01
-5.85741758e-01 -2.68011451e-01 1.90258734e-02 -7.60847479e-02
-3.69641006e-01 -8.90074223e-02 -6.17853343e-01 4.13295865e-01
6.56838492e-02 3.88393819e-01 -4.23471212e-01 -1.05330622e+00
-5.57267785e-01 8.81029725e-01 -3.33596468e-01 8.44893754e-01
6.73869669e-01 -1.03481197e+00 -9.98078644e-01 -2.73405373e-01
5.21884203e-01 -6.30117580e-02 1.34653255e-01 7.85729289e-01
4.75778840e-02 1.39369214e+00 6.13022208e-01 -6.63955152e-01
-1.32129741e+00 3.01844597e-01 2.06595361e-01 -7.40936875e-01
-3.19668986e-02 6.05095744e-01 6.82251230e-02 -8.19080651e-01
6.25493944e-01 -4.07121480e-02 -6.95677996e-01 3.17460477e-01
7.83336222e-01 1.09493412e-01 4.16074753e-01 -3.71652901e-01
-2.81505883e-01 1.74662784e-01 -4.66169834e-01 -7.15839505e-01
7.68794894e-01 -5.38478673e-01 -2.46773198e-01 1.11408979e-01
1.31041646e+00 -2.11339876e-01 -2.03837708e-01 -8.89876366e-01
4.78368968e-01 -1.02869486e-02 8.19560722e-04 -1.07757890e+00
-3.05294871e-01 5.33007264e-01 3.96171272e-01 4.70475793e-01
1.04187775e+00 1.33978963e-01 9.05154288e-01 1.27876687e+00
1.42329827e-01 -1.04456854e+00 1.50847062e-01 9.04012144e-01
1.38514197e+00 -1.35778093e+00 -7.34722540e-02 -3.74975875e-02
-5.78202069e-01 8.40839446e-01 7.39937276e-02 6.69698000e-01
3.97686422e-01 -6.07520878e-01 2.64865279e-01 -4.82259810e-01
-9.97550428e-01 -4.06236112e-01 8.53840709e-01 1.75523683e-01
5.48111618e-01 -2.96045840e-01 -7.56557524e-01 1.86459810e-01
-1.91836357e-01 -3.10575873e-01 -1.32630661e-01 1.00419641e+00
-3.59963655e-01 -1.12095952e+00 -1.47121385e-01 1.79643810e-01
-3.16995651e-01 -4.90917683e-01 -8.05243373e-01 5.01620531e-01
-1.02558589e+00 1.70249164e+00 -1.70399174e-01 -3.23568970e-01
4.74579155e-01 2.61552453e-01 1.35281691e-02 -4.12199587e-01
-9.50829983e-01 1.62520826e-01 5.70340991e-01 -6.03489161e-01
-4.75838989e-01 -4.34483826e-01 -9.64984715e-01 -1.02878963e-04
-5.06498992e-01 9.59775209e-01 7.12834656e-01 1.02945054e+00
7.28723347e-01 2.51705915e-01 8.00397635e-01 -5.31257629e-01
-8.88023734e-01 -1.34316194e+00 2.57740207e-02 3.67495179e-01
6.60765350e-01 -2.89087594e-01 -4.26040351e-01 -3.95065606e-01] | [12.091672897338867, 7.800274848937988] |
be1ffbe0-48f2-4409-bfb8-1a7884b3942a | point-voxel-adaptive-feature-abstraction-for | 2210.15514 | null | https://arxiv.org/abs/2210.15514v2 | https://arxiv.org/pdf/2210.15514v2.pdf | Point-Voxel Adaptive Feature Abstraction for Robust Point Cloud Classification | Great progress has been made in point cloud classification with learning-based methods. However, complex scene and sensor inaccuracy in real-world application make point cloud data suffer from corruptions, such as occlusion, noise and outliers. In this work, we propose Point-Voxel based Adaptive (PV-Ada) feature abstraction for robust point cloud classification under various corruptions. Specifically, the proposed framework iteratively voxelize the point cloud and extract point-voxel feature with shared local encoding and Transformer. Then, adaptive max-pooling is proposed to robustly aggregate the point cloud feature for classification. Experiments on ModelNet-C dataset demonstrate that PV-Ada outperforms the state-of-the-art methods. In particular, we rank the $2^{nd}$ place in ModelNet-C classification track of PointCloud-C Challenge 2022, with Overall Accuracy (OA) being 0.865. Code will be available at https://github.com/zhulf0804/PV-Ada. | ['Chen Zheng', 'Ninghua Yang', 'Changwei Lin', 'Lifa Zhu'] | 2022-10-27 | null | null | null | null | ['point-cloud-classification'] | ['computer-vision'] | [-2.33938649e-01 -6.05713725e-01 5.45689538e-02 -6.43919706e-01
-1.00877655e+00 -4.47205991e-01 3.83001417e-01 5.35005271e-01
-1.28739208e-01 3.20920080e-01 -4.85096514e-01 1.10173710e-01
-6.63136765e-02 -8.75846505e-01 -9.93670642e-01 -5.91063440e-01
-3.83242041e-01 3.16119492e-01 2.96752632e-01 9.38185006e-02
5.27231574e-01 1.15141916e+00 -1.88543558e+00 4.16914284e-01
8.16832483e-01 1.52145553e+00 1.69472575e-01 3.62521470e-01
-3.36299717e-01 7.14326128e-02 -4.60521817e-01 -2.31962819e-02
5.59142649e-01 6.56688809e-01 -3.99851412e-01 -1.34763764e-02
1.00163031e+00 -1.46788463e-01 -1.89907089e-01 1.08045399e+00
4.36839193e-01 9.37509090e-02 5.06930649e-01 -1.60310888e+00
-3.74318004e-01 -1.02823541e-01 -7.06647336e-01 -4.58184769e-03
-1.33993709e-02 3.48785400e-01 8.09609354e-01 -1.49552441e+00
3.87782812e-01 1.24918973e+00 8.45190227e-01 1.04247397e-02
-1.00143087e+00 -1.04446387e+00 1.54484481e-01 3.13899308e-01
-1.74740267e+00 -1.42626017e-01 8.57375681e-01 -5.40453553e-01
1.17964721e+00 4.57561165e-01 6.98561966e-01 4.84968811e-01
2.66603798e-01 6.75582945e-01 7.78038383e-01 1.15713544e-01
3.47306281e-01 -2.65268654e-01 1.05791993e-01 5.29074192e-01
4.02998984e-01 4.47268188e-02 -5.64312041e-01 -5.12001216e-01
6.61925375e-01 4.83334720e-01 -4.16066349e-02 -4.79023099e-01
-1.19250751e+00 5.30494392e-01 1.08956587e+00 -1.11169375e-01
-6.16181970e-01 4.62474763e-01 3.26696426e-01 2.69729823e-01
6.42274439e-01 2.35630706e-01 -7.15601206e-01 1.83308646e-02
-9.94629502e-01 6.74008429e-01 1.90954402e-01 1.26814985e+00
1.05663145e+00 -4.80269566e-02 5.56751713e-02 7.74766922e-01
4.88416672e-01 8.86096001e-01 -3.27316523e-02 -9.59289968e-01
6.40497327e-01 8.23299408e-01 -1.36193976e-01 -1.09333479e+00
-3.30341876e-01 -4.83946770e-01 -9.23050582e-01 6.24401212e-01
-2.20743552e-01 3.28192711e-01 -9.14811552e-01 9.35227394e-01
5.03117144e-01 7.82294989e-01 -2.76806682e-01 7.95814693e-01
1.07174432e+00 6.45000339e-01 -5.60194552e-02 3.29568803e-01
1.06093812e+00 -5.69466591e-01 -2.88375616e-02 2.04389729e-02
3.15060377e-01 -7.47712970e-01 9.45983589e-01 4.04601455e-01
-8.19066584e-01 -6.07693732e-01 -1.14938784e+00 -1.51536642e-02
-3.24789196e-01 -1.38353691e-01 5.13628304e-01 3.09102058e-01
-7.25236893e-01 7.77379036e-01 -1.04377794e+00 -1.18270859e-01
1.10632658e+00 5.23231149e-01 -3.96213979e-01 -2.80475050e-01
-3.86226863e-01 3.47539723e-01 8.16097632e-02 1.80856124e-01
-6.98150933e-01 -1.26540577e+00 -6.18094206e-01 -5.97781166e-02
-1.26065373e-01 -6.27672255e-01 9.71050382e-01 -3.15609634e-01
-1.03819299e+00 6.85848117e-01 -2.55406916e-01 -2.07176536e-01
3.75634283e-01 -3.84061038e-01 -2.24598274e-01 7.28073418e-02
2.05154270e-01 8.63433003e-01 8.79321277e-01 -1.51071775e+00
-9.25495088e-01 -6.47045791e-01 -2.85756528e-01 1.56949721e-02
1.38722971e-01 -1.11523211e-01 -4.23424453e-01 -2.56235480e-01
8.81199360e-01 -9.30917621e-01 -2.55221665e-01 5.47655404e-01
-1.70755610e-01 -5.27420759e-01 1.10516000e+00 -9.39490795e-02
4.63909119e-01 -2.42028666e+00 -3.68667245e-01 4.69367206e-01
3.11387688e-01 8.12349096e-02 -1.24473296e-01 2.09194347e-01
-6.57119527e-02 1.47193745e-01 -3.45447898e-01 -6.68049097e-01
-3.36212777e-02 1.74168468e-01 -4.76459146e-01 8.88141632e-01
5.05153239e-01 7.33618855e-01 -6.97658300e-01 -4.26022172e-01
7.29721010e-01 6.65974617e-01 -5.73053598e-01 -3.31576429e-02
-2.94732511e-01 3.83590192e-01 -6.07421398e-01 1.31931746e+00
1.43175948e+00 -5.37136085e-02 -6.49914443e-01 -1.91574901e-01
-2.62489974e-01 2.18916626e-04 -1.09699023e+00 1.74727035e+00
-2.62671471e-01 4.67224270e-01 2.11907998e-02 -5.24555564e-01
1.09068990e+00 3.40202753e-03 8.49468946e-01 -4.01322007e-01
-1.62015241e-02 3.59488606e-01 -2.48617068e-01 -2.06380352e-01
5.46942353e-01 3.22004765e-01 4.05544117e-02 -3.60928595e-01
-1.17792763e-01 -7.28914738e-01 -4.42681789e-01 -8.84126276e-02
1.06463814e+00 6.56088144e-02 -1.20433271e-01 -2.27726117e-01
4.20317143e-01 1.96825743e-01 6.19306922e-01 6.18862271e-01
-3.24009776e-01 1.09679258e+00 -1.05768204e-01 -4.77719516e-01
-8.04438174e-01 -1.26741862e+00 -4.51475173e-01 4.79275763e-01
2.54159361e-01 -4.73743021e-01 -2.10887730e-01 -2.95839667e-01
6.49414957e-01 5.43668270e-01 -1.05199337e-01 -3.82828228e-02
-6.43574119e-01 -3.72761130e-01 3.89173210e-01 5.47776759e-01
4.75840628e-01 -8.44268918e-01 -5.23962796e-01 1.73491746e-01
1.29643485e-01 -1.14003015e+00 7.64068142e-02 -8.73019621e-02
-1.33779299e+00 -8.45157981e-01 -8.32171813e-02 -4.73159939e-01
6.24480188e-01 5.91161728e-01 1.04904795e+00 2.48943508e-01
-3.42915118e-01 3.16199183e-01 -3.86654139e-01 -6.99247479e-01
3.94045472e-01 5.33674918e-02 9.16023254e-02 -2.03367501e-01
6.22973502e-01 -6.79134369e-01 -7.46357620e-01 1.16771922e-01
-6.83996797e-01 -3.65614742e-01 3.39631230e-01 3.69909227e-01
1.03758419e+00 -4.52686250e-02 3.79250832e-02 -2.90012449e-01
1.10060520e-01 -6.15296185e-01 -8.69201362e-01 -3.38625252e-01
-4.22658324e-01 -5.47792554e-01 2.80384511e-01 6.98766187e-02
-3.54231507e-01 3.36697042e-01 -1.54049948e-01 -1.09981620e+00
-3.70845675e-01 3.88652116e-01 -1.84848487e-01 -5.11458337e-01
4.84222353e-01 1.61159217e-01 -2.75818706e-01 -6.49744570e-01
2.59932488e-01 7.34546781e-01 5.71776748e-01 -7.75607467e-01
1.17941368e+00 8.60449731e-01 2.23310605e-01 -1.03011596e+00
-2.60212988e-01 -7.36441493e-01 -7.60121822e-01 -2.43232206e-01
5.06259143e-01 -1.24898779e+00 -7.99680769e-01 5.56435287e-01
-1.37363720e+00 7.65249878e-02 -2.86928087e-01 4.40129757e-01
-5.01951516e-01 1.94247812e-01 -2.70034492e-01 -6.02943420e-01
-5.08762062e-01 -1.34794319e+00 1.49371731e+00 7.87130743e-02
1.57157227e-01 -3.14897925e-01 -5.17120212e-03 1.98871598e-01
3.31640273e-01 5.47609389e-01 4.42768335e-01 -3.50170106e-01
-1.22409880e+00 -4.90817219e-01 -4.11412776e-01 4.14365828e-01
-1.03406170e-02 3.16843957e-01 -1.18578362e+00 -4.67694938e-01
-1.49462983e-01 -5.88115305e-02 7.52600014e-01 3.68421167e-01
1.81509829e+00 1.10050149e-01 -3.68522376e-01 1.08139241e+00
1.69927740e+00 -1.76243037e-01 5.73000610e-01 2.90929705e-01
9.35133338e-01 1.81249484e-01 7.97999382e-01 6.04903996e-01
4.07258332e-01 5.63781261e-01 1.08685279e+00 1.31978929e-01
6.20500408e-02 6.04628585e-02 6.63982378e-03 6.43667936e-01
-1.53440386e-01 1.86702117e-01 -1.31749320e+00 6.24065042e-01
-1.76547718e+00 -6.43345177e-01 -5.99055469e-01 2.11556363e+00
2.95795560e-01 1.36682302e-01 -4.01166350e-01 1.76501106e-02
5.67905664e-01 2.61206627e-01 -5.17985046e-01 -2.09175050e-02
-1.02582313e-01 5.12807727e-01 8.44637930e-01 2.83408999e-01
-1.29313982e+00 9.46586192e-01 4.92939854e+00 7.80597925e-01
-1.38257194e+00 2.11777493e-01 1.94909304e-01 -3.86335433e-01
2.99584121e-03 -2.54107993e-02 -6.65156722e-01 5.00222087e-01
5.87328017e-01 2.08451729e-02 2.22838353e-02 1.12470961e+00
1.31794840e-01 1.90256864e-01 -9.61270094e-01 1.49683058e+00
-1.17979832e-01 -1.47988725e+00 -5.17480150e-02 1.74685031e-01
6.63151979e-01 9.45886850e-01 1.44859299e-01 2.26398453e-01
-1.12317838e-02 -8.72962713e-01 8.38749230e-01 4.99713093e-01
7.24720776e-01 -6.52101099e-01 7.46441364e-01 1.77656502e-01
-1.52100086e+00 8.16897526e-02 -7.90701985e-01 -1.07573956e-01
-1.30118191e-01 8.69765401e-01 -6.59441948e-01 7.71307588e-01
1.29009771e+00 9.38813746e-01 -6.95047617e-01 1.61197829e+00
1.55542076e-01 4.61801857e-01 -8.61241221e-01 2.72242546e-01
2.59770632e-01 -5.07262535e-02 8.23660612e-01 1.07455945e+00
5.09067535e-01 1.84296757e-01 4.00937557e-01 7.70505905e-01
-1.83196813e-01 7.89056644e-02 -5.77812850e-01 5.04147947e-01
8.45380843e-01 1.23534262e+00 -6.19167328e-01 -2.96512693e-01
-3.11318994e-01 6.00465953e-01 2.82356113e-01 8.34080800e-02
-6.47789717e-01 -2.96720713e-01 1.37928748e+00 8.50563943e-02
4.90634114e-01 -6.30555570e-01 -7.13511586e-01 -1.02251780e+00
3.80439788e-01 -4.51601177e-01 -1.26547530e-01 -8.12589884e-01
-1.41219962e+00 5.50100744e-01 1.16901666e-01 -1.84906304e+00
4.68792826e-01 -6.55707598e-01 -6.89531088e-01 7.29198337e-01
-1.62415361e+00 -1.17592084e+00 -8.13637137e-01 4.96313274e-01
6.24514520e-01 -1.68376043e-01 5.73233008e-01 4.90822107e-01
-2.76318729e-01 2.96976626e-01 2.46595114e-01 -7.47283176e-02
4.86526370e-01 -9.81067002e-01 8.53720844e-01 6.49258256e-01
-3.99948359e-02 4.45322663e-01 4.90535349e-01 -7.80534089e-01
-1.55205441e+00 -1.57721710e+00 4.82169628e-01 -5.04724264e-01
4.58214849e-01 -5.11898220e-01 -1.11973846e+00 4.90878850e-01
-3.02267462e-01 6.85353875e-01 3.08313251e-01 -2.72839427e-01
-5.08541644e-01 -4.73967373e-01 -1.43576097e+00 2.29310498e-01
9.69772995e-01 -3.86719704e-01 -2.95258403e-01 6.41675770e-01
8.11174691e-01 -6.52105510e-01 -1.13681901e+00 7.70804703e-01
3.06429893e-01 -6.68755889e-01 1.27517748e+00 -2.83685446e-01
2.27100939e-01 -6.72010124e-01 -7.20806599e-01 -1.03842330e+00
-4.76078123e-01 -9.00878534e-02 1.34415925e-01 1.01021707e+00
1.38154462e-01 -8.15506876e-01 8.86210918e-01 3.68363976e-01
-5.86134374e-01 -7.55021274e-01 -1.34633899e+00 -9.20130491e-01
2.14452550e-01 -8.98595989e-01 1.06418133e+00 9.75144744e-01
-5.29013216e-01 -4.43208247e-01 2.83888727e-01 7.49855399e-01
9.00893211e-01 1.37543768e-01 1.09215915e+00 -1.54876697e+00
3.92652512e-01 -4.33104366e-01 -1.00185263e+00 -7.56802917e-01
1.55055607e-02 -1.00352633e+00 -8.63868464e-03 -1.48163795e+00
-3.06307018e-01 -1.00173593e+00 -4.03916031e-01 4.87114877e-01
4.44179624e-02 3.99244130e-01 5.48361599e-01 6.59979761e-01
-4.54961985e-01 6.93327546e-01 7.19277859e-01 -3.79400432e-01
-4.83698025e-02 9.18863192e-02 -8.67031291e-02 6.05246067e-01
1.18897736e+00 -6.68426931e-01 1.13082998e-01 -6.89810693e-01
-4.11065035e-02 -3.86585921e-01 7.45112777e-01 -1.57441628e+00
3.73847634e-01 -1.72570273e-01 5.61594963e-01 -1.38307297e+00
7.08983362e-01 -1.17871988e+00 3.22356284e-01 3.92885685e-01
3.11076283e-01 2.74395555e-01 3.51253778e-01 3.74345124e-01
-2.36700669e-01 4.04245742e-02 6.72095895e-01 -1.14477962e-01
-6.54341638e-01 8.42289567e-01 3.31140846e-01 -5.22550881e-01
8.77882600e-01 -3.66852224e-01 -4.43382412e-01 2.05088928e-01
-3.56191784e-01 3.59285504e-01 6.41045809e-01 5.55678666e-01
1.02481174e+00 -1.41516757e+00 -1.02664781e+00 3.87004822e-01
6.15556419e-01 6.54819191e-01 3.09994608e-01 5.68344057e-01
-9.53511536e-01 1.55277744e-01 -1.14720881e-01 -1.38823211e+00
-1.28082156e+00 1.28487155e-01 1.63504377e-01 4.49448913e-01
-8.55481207e-01 9.67321873e-01 -1.96271226e-01 -6.66768968e-01
1.01780802e-01 -7.28878736e-01 4.07040298e-01 -2.62144804e-01
2.58334577e-01 4.24953848e-01 5.97649693e-01 -8.24790180e-01
-7.96755195e-01 9.24314022e-01 7.46857300e-02 2.78776616e-01
1.45737648e+00 3.01783264e-01 -3.63998860e-01 4.07671839e-01
1.38714600e+00 -2.73244053e-01 -1.21812713e+00 -3.73744428e-01
-5.29635474e-02 -9.94268715e-01 1.82597026e-01 -2.71758825e-01
-1.18586648e+00 8.62042606e-01 1.05764902e+00 -1.65710554e-01
7.82060921e-01 2.11871620e-02 6.74800754e-01 4.78998661e-01
6.93896472e-01 -6.62228584e-01 -3.70861799e-01 4.46160823e-01
1.12539876e+00 -1.37335908e+00 3.54328096e-01 -5.65535307e-01
-1.16518371e-01 8.48259807e-01 6.11461103e-01 -7.80180514e-01
9.76924002e-01 1.38178736e-01 7.00108036e-02 -5.30746698e-01
-6.55797720e-01 3.98960076e-02 1.80544034e-01 6.27073109e-01
-1.98212173e-02 2.85725981e-01 3.17453533e-01 1.78640243e-02
-4.51983541e-01 -6.89531937e-02 3.63589376e-02 1.18940914e+00
-4.92206037e-01 -6.69295847e-01 -7.89867520e-01 6.68621719e-01
-1.37695417e-01 -5.91736212e-02 -1.11225478e-01 5.63677669e-01
2.96561688e-01 7.96132207e-01 5.31897008e-01 -4.37284827e-01
6.02019250e-01 -2.01880857e-01 3.11425984e-01 -6.04440928e-01
-6.28811240e-01 -1.17407404e-01 -5.20988107e-01 -9.63067830e-01
-2.06858829e-01 -9.12177503e-01 -1.26599658e+00 -4.12438810e-01
-2.24245012e-01 -1.33029029e-01 1.24119711e+00 4.83703256e-01
6.79017007e-01 3.68963718e-01 6.28331184e-01 -1.46396303e+00
-3.54957610e-01 -7.84019887e-01 -3.56781572e-01 2.86433667e-01
5.73513091e-01 -7.61984706e-01 -3.72327864e-01 -3.28304350e-01] | [7.78437614440918, -3.1042640209198] |
d736fdf9-0a91-459b-91bd-f95d7b41ad9f | automatic-detection-of-aerial-survey-ground | 2303.03041 | null | https://arxiv.org/abs/2303.03041v1 | https://arxiv.org/pdf/2303.03041v1.pdf | Automatic detection of aerial survey ground control points based on Yolov5-OBB | The use of ground control points (GCPs) for georeferencing is the most common strategy in unmanned aerial vehicle (UAV) photogrammetry, but at the same time their collection represents the most time-consuming and expensive part of UAV campaigns. Recently, deep learning has been rapidly developed in the field of small object detection. In this letter, to automatically extract coordinates information of ground control points (GCPs) by detecting GCP-markers in UAV images, we propose a solution that uses a deep learning-based architecture, YOLOv5-OBB, combined with a confidence threshold filtering algorithm and an optimal ranking algorithm. We applied our proposed method to a dataset collected by DJI Phantom 4 Pro drone and obtained good detection performance with the mean Average Precision (AP) of 0.832 and the highest AP of 0.982 for the cross-type GCP-markers. The proposed method can be a promising tool for future implementation of the end-to-end aerial triangulation process. | ['Zhuang Zhiheng', 'Chang Mengxia', 'Dong Di', 'Li Xiaopeng', 'Zheng Zhi', 'Wang Chao', 'Yang Jia', 'Cheng Chuanxiang'] | 2023-03-06 | null | null | null | null | ['small-object-detection'] | ['computer-vision'] | [-1.12419076e-01 -2.88642436e-01 3.32469344e-01 -9.76425633e-02
-5.37418008e-01 -4.99911219e-01 4.21936154e-01 1.69858322e-01
-6.19711339e-01 4.85920727e-01 -5.80983400e-01 -6.57588392e-02
-3.96006525e-01 -1.04583275e+00 -7.55514205e-01 -5.32979369e-01
-5.43167233e-01 6.66974723e-01 3.25426430e-01 -4.09268945e-01
1.53895751e-01 8.81375670e-01 -1.65171623e+00 -3.14678907e-01
8.98366272e-01 1.30312335e+00 3.52459997e-01 3.52806985e-01
5.39071083e-01 4.43782248e-02 -4.08460528e-01 -2.70200700e-01
8.96102965e-01 5.63588962e-02 -3.80071759e-01 -1.51580602e-01
8.26844275e-01 -4.84351814e-01 1.33567572e-01 1.20024776e+00
5.11944056e-01 1.40043423e-01 5.73352396e-01 -1.07582593e+00
-9.04433802e-02 2.09336594e-01 -9.61360812e-01 -3.21074165e-02
8.90939832e-02 -1.60958730e-02 8.52191985e-01 -8.83755386e-01
3.63637507e-01 8.21177781e-01 9.52819824e-01 -1.31651863e-01
-8.22552145e-01 -5.94283581e-01 -2.34366819e-01 1.92314297e-01
-1.83272958e+00 -3.99889983e-02 4.11014587e-01 -8.21574807e-01
7.18899786e-01 1.65632382e-01 9.95429337e-01 4.15520906e-01
2.35008031e-01 3.74679357e-01 8.34338188e-01 -4.39145088e-01
6.93785399e-02 -2.72785306e-01 -3.04290056e-01 1.10539138e+00
6.64309800e-01 3.71739388e-01 4.19559292e-02 2.58414075e-02
8.19830298e-01 -7.48010576e-02 -2.75924563e-01 -4.33984578e-01
-1.20315099e+00 6.80692971e-01 9.96281028e-01 1.17060706e-01
-6.52635276e-01 2.69542962e-01 6.13551475e-02 8.50675106e-02
4.10409242e-01 6.61135793e-01 -3.59305412e-01 3.10096174e-01
-1.26941693e+00 3.58668357e-01 3.23338062e-01 1.09086847e+00
7.78814375e-01 5.64410202e-02 4.80336510e-03 4.77713466e-01
3.31866592e-01 7.67713547e-01 -2.83178259e-02 -5.94967782e-01
3.01550716e-01 7.13596821e-01 5.52829564e-01 -1.50599694e+00
-5.64281166e-01 -6.16975307e-01 -8.15546691e-01 4.41852599e-01
2.69816071e-01 -4.76740360e-01 -7.95254290e-01 1.07466924e+00
3.20097774e-01 8.59721005e-02 -2.95125246e-01 1.32932758e+00
5.08456826e-01 6.00016296e-01 -1.32971242e-01 1.78138301e-01
1.19104362e+00 -4.48254198e-01 -2.35947415e-01 -3.37780565e-01
5.88405252e-01 -7.90100694e-01 5.12848556e-01 3.89649510e-01
-4.76810127e-01 -6.64025187e-01 -1.38674128e+00 2.60987490e-01
-3.56276661e-01 1.09175348e+00 4.01754141e-01 2.88902909e-01
-9.92198169e-01 7.89216876e-01 -7.08629370e-01 -5.37736297e-01
3.77950937e-01 6.33488059e-01 -4.22555685e-01 2.59225070e-01
-8.52413893e-01 9.20086086e-01 7.40832627e-01 5.04211128e-01
-9.13048446e-01 -2.97384560e-01 -5.86161971e-01 5.43997111e-03
4.06724423e-01 -6.93126857e-01 7.25850463e-01 -8.71734858e-01
-1.15027177e+00 9.63707268e-01 6.00889385e-01 -7.73782969e-01
6.92387283e-01 -6.24633431e-01 -1.36140794e-01 1.66505314e-02
1.13340594e-01 9.20280695e-01 9.66770530e-01 -9.86117005e-01
-1.15074658e+00 -5.98745763e-01 4.53268364e-02 1.19379073e-01
2.02562008e-02 -4.80525643e-02 -2.61420906e-02 -4.00234282e-01
4.37797755e-01 -1.23234975e+00 -1.23124262e-02 7.26074874e-02
-1.39426470e-01 -5.69097213e-02 6.22019112e-01 -6.93667233e-01
9.48464274e-01 -1.83419287e+00 1.36463523e-01 2.88113862e-01
4.91868192e-03 6.76324069e-01 4.47669402e-02 3.02726567e-01
1.21206148e-02 -2.27533787e-01 -2.03626305e-01 8.07225034e-02
-3.14430118e-01 -4.47956264e-01 -1.48493960e-01 7.09811151e-01
2.55748689e-01 4.91004020e-01 -1.08313012e+00 -3.11300218e-01
6.19116783e-01 3.56412143e-01 -2.37530902e-01 1.09184675e-01
-2.92261600e-01 2.08979517e-01 -3.44231665e-01 7.92225063e-01
9.21613693e-01 3.62514973e-01 -1.49607524e-01 -4.31340903e-01
-6.79997742e-01 -4.03858602e-01 -1.15145063e+00 1.40791631e+00
-2.59195209e-01 8.74284446e-01 1.05951808e-01 -5.21846414e-01
1.31033838e+00 -4.99144793e-02 4.53720808e-01 -9.84368008e-03
5.06634176e-01 2.87225306e-01 -2.90144514e-02 -3.36584777e-01
8.33804369e-01 2.07080394e-01 1.48029670e-01 -4.26274598e-01
1.46148652e-01 -3.58802229e-01 1.47890627e-01 -4.21034575e-01
6.02823913e-01 2.56678879e-01 4.67951000e-01 -3.75435621e-01
4.95619178e-01 5.14714777e-01 4.40985590e-01 4.79748398e-01
3.22974548e-02 6.88758194e-01 2.62120855e-03 -8.46336007e-01
-8.89317572e-01 -3.68426114e-01 -2.24335790e-01 4.08911884e-01
3.08570087e-01 -1.99490964e-01 -6.74655080e-01 -2.94264555e-01
1.32321015e-01 3.61479163e-01 -2.87245810e-01 5.13462070e-03
-2.13279665e-01 -7.06312299e-01 5.03535688e-01 2.61461169e-01
9.77912486e-01 -5.78800917e-01 -1.09692144e+00 2.30185345e-01
-4.83730845e-02 -9.94098246e-01 2.80530035e-01 -2.99645169e-03
-8.02881479e-01 -1.32048845e+00 -8.32404077e-01 -6.42723739e-01
4.67752576e-01 4.79807377e-01 7.24712312e-01 8.97302553e-02
-1.65713564e-01 4.76576686e-02 -5.14110506e-01 -5.66433609e-01
1.65668711e-01 2.08541721e-01 1.80893466e-01 9.06114206e-02
2.73913145e-01 -1.51039138e-01 -5.27460098e-01 5.06238818e-01
-3.46441239e-01 -1.65058643e-01 7.25909710e-01 6.70366228e-01
5.89011610e-01 -5.92503976e-03 -1.64113827e-02 -1.85323045e-01
2.40586832e-01 -1.52820349e-01 -1.69065738e+00 1.12019956e-01
-3.94861907e-01 -4.98300135e-01 4.71593231e-01 1.49920151e-01
-3.09748888e-01 6.37291014e-01 -2.53710262e-02 -6.77185476e-01
-3.17768991e-01 7.09728658e-01 3.35032567e-02 -7.67081320e-01
7.58679450e-01 -2.19435975e-01 -1.76347837e-01 -1.75154343e-01
-2.14745272e-02 6.41147196e-01 4.43087518e-01 -1.79818496e-01
8.80780995e-01 3.45427811e-01 4.42350924e-01 -1.14105165e+00
-6.46493793e-01 -5.41705728e-01 -9.77694154e-01 -4.39457148e-01
9.66825843e-01 -1.35197699e+00 -6.15489304e-01 4.86152023e-01
-1.29756105e+00 -9.44105014e-02 1.35176525e-01 7.14300573e-01
-2.76682466e-01 3.69347066e-01 4.60127220e-02 -8.47915649e-01
-4.43368405e-01 -1.13916731e+00 1.30169380e+00 3.29974055e-01
8.95585492e-02 -4.61599708e-01 1.97299179e-02 9.19717923e-02
9.34547931e-02 7.49389887e-01 1.96618676e-01 -3.46351326e-01
-9.86831427e-01 -8.11463416e-01 -3.09858173e-01 2.66965538e-01
-9.70331877e-02 3.37511450e-01 -8.78156841e-01 -4.55226183e-01
-3.76797944e-01 3.58005315e-02 8.77864003e-01 5.27971268e-01
6.99208379e-01 9.79737192e-02 -4.02771652e-01 8.34431767e-01
1.67295241e+00 8.60495269e-02 4.73758429e-01 5.22995591e-01
7.33699024e-01 3.12210202e-01 1.26276731e+00 6.66913033e-01
-3.27857062e-02 9.16069388e-01 1.08723342e+00 -6.93171024e-02
2.85902828e-01 -1.10910945e-01 3.53662372e-01 7.16045499e-02
-8.35237384e-01 -2.52764523e-01 -1.13716853e+00 5.79815149e-01
-1.74694264e+00 -8.14307332e-01 -4.50460225e-01 2.45797205e+00
-8.46222565e-02 -1.02450937e-01 -9.22909975e-02 1.42607361e-01
9.69481111e-01 1.29007801e-01 3.83623838e-02 2.08737716e-01
-4.51637134e-02 -1.63070530e-01 1.01983702e+00 3.87881905e-01
-1.70309067e+00 1.25009632e+00 4.47383070e+00 6.33049846e-01
-1.15603960e+00 -2.95302927e-01 -1.00135908e-01 3.16798210e-01
8.00426126e-01 -1.11772381e-02 -9.40075636e-01 3.78424019e-01
4.37564522e-01 2.57708877e-01 2.60520607e-01 1.16899109e+00
4.66236733e-02 -3.18791300e-01 -6.14961147e-01 1.10722327e+00
6.25841990e-02 -1.07483804e+00 -9.78974625e-02 8.30107033e-02
6.86517954e-01 3.05124551e-01 -1.22056700e-01 -2.38961533e-01
8.33781511e-02 -5.21897614e-01 7.75539160e-01 4.19258714e-01
9.22828138e-01 -8.17427278e-01 1.27549922e+00 3.00458938e-01
-1.33109939e+00 -2.48291641e-01 -8.80061924e-01 -1.16547078e-01
-7.06925467e-02 5.57197630e-01 -1.24847615e+00 8.34214270e-01
7.61523664e-01 7.61997879e-01 -5.96188843e-01 1.73388910e+00
-6.04247212e-01 2.23815590e-01 -6.41762733e-01 -4.88592274e-02
5.25160193e-01 -5.25141776e-01 8.19127202e-01 7.81325042e-01
1.01669335e+00 5.15031181e-02 4.46735799e-01 5.41925848e-01
5.22454679e-02 7.92939365e-02 -9.02063608e-01 1.74850315e-01
4.88718271e-01 1.60567641e+00 -9.46359277e-01 -8.08389187e-02
1.31573692e-01 5.45334339e-01 -1.14468768e-01 -2.03284442e-01
-9.43014443e-01 -6.07915461e-01 4.24540251e-01 3.69315296e-01
2.75422096e-01 -6.14870548e-01 1.89920440e-01 -9.70170677e-01
-1.17056914e-01 -5.96685350e-01 1.72694445e-01 -1.02759814e+00
-7.67137945e-01 8.47194016e-01 -7.86337927e-02 -1.94963002e+00
4.40653563e-02 -9.19528544e-01 -4.22178030e-01 7.39269316e-01
-1.31671619e+00 -1.37072158e+00 -1.03783607e+00 2.61719227e-01
2.94020444e-01 -4.63257849e-01 7.45181322e-01 3.03022027e-01
-3.77192885e-01 1.19084127e-01 3.59816141e-02 3.49128515e-01
6.33569837e-01 -9.71702337e-01 2.43721843e-01 1.17696035e+00
3.16809833e-01 3.55092138e-01 6.04542971e-01 -9.13077056e-01
-1.20936608e+00 -1.39394701e+00 4.54962194e-01 -1.86034977e-01
4.10405636e-01 1.64859705e-02 -3.42111081e-01 7.41111815e-01
1.12630185e-02 -1.46264791e-01 1.39711618e-01 -6.66937903e-02
4.10421193e-01 -4.65528518e-01 -1.09253752e+00 2.70255208e-01
6.75633430e-01 7.10882097e-02 -4.46083367e-01 5.78027904e-01
2.64242649e-01 -6.39247060e-01 -6.43153965e-01 6.79238200e-01
5.53442717e-01 -1.02584577e+00 1.00803351e+00 -1.06064416e-03
3.00549287e-02 -8.75289023e-01 -1.23526238e-01 -1.50528848e+00
-5.06876111e-01 -3.52941930e-01 5.46944976e-01 9.48233545e-01
6.06641062e-02 -3.76161933e-01 6.56742215e-01 1.03083448e-02
-4.60733294e-01 -2.18642801e-01 -8.55370998e-01 -7.93855309e-01
-6.44484580e-01 -4.87853289e-02 4.61971760e-01 8.05106163e-01
-6.01037085e-01 1.33873090e-01 -4.66507524e-01 8.68635774e-01
6.29717648e-01 5.71068600e-02 9.43569601e-01 -1.90263391e+00
2.62659758e-01 -1.75386101e-01 -8.47621143e-01 -7.11674988e-01
-8.75004288e-03 -8.34703267e-01 1.90224394e-01 -1.28980732e+00
-7.39345729e-01 -5.36111355e-01 1.36783391e-01 3.86108249e-01
2.27740094e-01 2.37338647e-01 1.16643094e-01 1.85852587e-01
-3.14050049e-01 5.28628945e-01 7.97907710e-01 -6.26036227e-02
-1.52025849e-01 2.83024639e-01 1.33445799e-01 1.03401291e+00
7.50917315e-01 -3.81830245e-01 -5.47370873e-03 -7.54850686e-01
3.63426656e-01 -6.18006513e-02 7.12488949e-01 -1.76849735e+00
3.57936829e-01 2.04869211e-01 6.84993804e-01 -1.07673788e+00
4.88020509e-01 -1.32432210e+00 2.26616621e-01 6.13963664e-01
3.33726615e-01 2.26458892e-01 3.20113540e-01 4.47682202e-01
-3.19160163e-01 -3.21528763e-01 9.15370464e-01 -2.23566025e-01
-1.04066396e+00 3.57236385e-01 -2.83049196e-01 -4.53041106e-01
1.17634857e+00 -1.66423663e-01 -9.97039676e-02 6.87275678e-02
-3.32147628e-01 2.00754762e-01 6.44383967e-01 9.04966295e-02
8.16636920e-01 -9.57369387e-01 -6.22962058e-01 3.80881608e-01
1.91040218e-01 2.03534156e-01 -8.32575113e-02 8.74269187e-01
-1.28855383e+00 3.70166898e-01 -6.73039436e-01 -7.58285522e-01
-1.27197409e+00 2.30244428e-01 4.70569551e-01 2.45446011e-01
-2.71961629e-01 9.33984280e-01 -5.51837265e-01 -1.43378988e-01
6.32101595e-02 -4.82606530e-01 -3.85836393e-01 5.35685480e-01
2.44938821e-01 5.58057845e-01 3.77483398e-01 -8.12917233e-01
-4.47693884e-01 1.20658088e+00 3.66343826e-01 2.27857977e-01
1.38842046e+00 2.15700164e-01 -2.26545244e-01 -6.61673397e-02
4.76809829e-01 -1.07201234e-01 -9.66823339e-01 2.94677258e-01
2.18443245e-01 -7.89812326e-01 3.96627843e-01 -5.98726273e-01
-1.00644851e+00 9.72029626e-01 9.76440132e-01 4.26761024e-02
8.56118679e-01 -5.37015915e-01 4.96138930e-01 7.63470232e-01
1.03598082e+00 -9.79341090e-01 -5.34618497e-01 5.56624532e-01
9.72830296e-01 -1.50156665e+00 3.33620787e-01 -2.68292814e-01
-4.79288489e-01 1.40913653e+00 4.93177474e-01 -6.43832028e-01
3.11011642e-01 -1.94724098e-01 -4.29899544e-02 -3.62616837e-01
2.53243819e-02 -5.88077664e-01 4.01127338e-01 5.73580444e-01
2.37711862e-01 1.87872157e-01 -1.80512384e-01 -4.84093418e-03
-2.24889904e-01 1.59775987e-01 4.01564449e-01 6.41888082e-01
-6.63723230e-01 -6.79874659e-01 -6.39779270e-01 2.56496459e-01
-2.60956079e-01 -8.41543451e-02 -4.83796567e-01 1.11135209e+00
4.31132257e-01 7.20959961e-01 1.20518003e-02 -7.68475235e-01
4.48397398e-01 -2.27698892e-01 4.77848470e-01 -4.52581078e-01
-5.24976850e-01 7.39740431e-02 -4.66658846e-02 -4.23523545e-01
-5.46767354e-01 -4.28533435e-01 -7.44333684e-01 -1.34122238e-01
-7.43516326e-01 9.21195149e-02 8.05839956e-01 5.85257649e-01
4.30015504e-01 1.33934557e-01 5.02929628e-01 -1.16436565e+00
-3.18311661e-01 -9.40529525e-01 -5.75303435e-01 -2.48061091e-01
2.16475040e-01 -1.06903398e+00 -1.39358729e-01 -8.32177848e-02] | [7.838160514831543, -1.5912714004516602] |
b065867e-7f5f-4572-b0d7-182daeb1c62c | revisiting-global-statistics-aggregation-for | 2112.04491 | null | https://arxiv.org/abs/2112.04491v4 | https://arxiv.org/pdf/2112.04491v4.pdf | Improving Image Restoration by Revisiting Global Information Aggregation | Global operations, such as global average pooling, are widely used in top-performance image restorers. They aggregate global information from input features along entire spatial dimensions but behave differently during training and inference in image restoration tasks: they are based on different regions, namely the cropped patches (from images) and the full-resolution images. This paper revisits global information aggregation and finds that the image-based features during inference have a different distribution than the patch-based features during training. This train-test inconsistency negatively impacts the performance of models, which is severely overlooked by previous works. To reduce the inconsistency and improve test-time performance, we propose a simple method called Test-time Local Converter (TLC). Our TLC converts global operations to local ones only during inference so that they aggregate features within local spatial regions rather than the entire large images. The proposed method can be applied to various global modules (e.g., normalization, channel and spatial attention) with negligible costs. Without the need for any fine-tuning, TLC improves state-of-the-art results on several image restoration tasks, including single-image motion deblurring, video deblurring, defocus deblurring, and image denoising. In particular, with TLC, our Restormer-Local improves the state-of-the-art result in single image deblurring from 32.92 dB to 33.57 dB on GoPro dataset. The code is available at https://github.com/megvii-research/tlc. | ['Xin Lu', 'Chengpeng Chen', 'Liangyu Chen', 'Xiaojie Chu'] | 2021-12-08 | null | null | null | null | ['color-image-denoising', 'image-dehazing', 'grayscale-image-denoising'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 3.52261513e-01 -6.23667836e-01 -2.05524508e-02 -3.20312947e-01
-9.53649879e-01 -3.64583999e-01 3.17062676e-01 -3.72946084e-01
-5.70499957e-01 6.65395975e-01 4.16474164e-01 -1.88324422e-01
-2.02882141e-01 -6.39822066e-01 -8.82580638e-01 -1.29342520e+00
3.23398948e-01 -2.98098445e-01 2.55683213e-01 -1.51927605e-01
4.81004030e-01 5.39054096e-01 -1.50969481e+00 4.44294363e-01
1.15987706e+00 8.88750851e-01 5.31697512e-01 5.94434261e-01
3.55781227e-01 7.15456009e-01 -8.04133177e-01 -4.61608991e-02
1.21060856e-01 -5.30510485e-01 -5.40994048e-01 -9.21864137e-02
7.40599930e-01 -7.44223833e-01 -5.27813137e-01 1.45516264e+00
9.41744149e-01 1.58508599e-01 3.63735467e-01 -6.81279898e-01
-1.02690411e+00 2.12276936e-01 -6.55993581e-01 5.61917543e-01
-7.75944144e-02 3.27063471e-01 3.34669262e-01 -9.16249871e-01
4.64951873e-01 1.22184646e+00 5.69497943e-01 3.03207070e-01
-1.04790819e+00 -7.82851636e-01 -5.60009927e-02 6.94556534e-01
-1.28533697e+00 -6.04933739e-01 5.61045468e-01 -2.98139900e-01
9.39087689e-01 3.06313783e-01 2.15820894e-01 7.22383976e-01
4.61805016e-01 4.68743891e-01 1.34917009e+00 -1.58606723e-01
-9.35196131e-02 -3.54291707e-01 1.35871589e-01 3.76245588e-01
2.07343400e-01 1.27560943e-01 -5.43727160e-01 3.05297911e-01
1.09037912e+00 -9.62278694e-02 -9.34776485e-01 2.84312695e-01
-1.22565496e+00 4.35057491e-01 6.41543806e-01 4.17188376e-01
-2.72016048e-01 6.69357106e-02 1.90819040e-01 4.69315886e-01
6.48645699e-01 1.74565881e-01 -4.16288584e-01 6.70416057e-02
-8.19881976e-01 4.30476926e-02 1.74095735e-01 4.61454511e-01
8.77397597e-01 3.75804640e-02 -4.35805827e-01 1.20977664e+00
-1.41048312e-01 7.34863400e-01 6.57592356e-01 -8.06882262e-01
5.01269937e-01 1.15912251e-01 1.82474628e-02 -1.08178508e+00
-3.87546808e-01 -7.22071826e-01 -1.36521459e+00 1.99368954e-01
2.41890565e-01 9.27254707e-02 -1.24966800e+00 1.72699654e+00
1.00288205e-01 3.70355219e-01 -1.19592845e-01 1.39050269e+00
8.89627099e-01 7.05405474e-01 -2.61880875e-01 -3.14602494e-01
1.25853753e+00 -9.83993292e-01 -8.70565057e-01 -3.42532098e-01
3.08141083e-01 -8.49182725e-01 8.99410784e-01 4.77855861e-01
-1.06222558e+00 -7.64849544e-01 -1.10038340e+00 -4.48974997e-01
-1.12853497e-01 2.95565665e-01 4.60203707e-01 5.36516249e-01
-1.30019724e+00 7.51355767e-01 -7.29853511e-01 -1.62669659e-01
5.04344285e-01 3.15695047e-01 -5.14124513e-01 -5.23235798e-01
-9.88662004e-01 8.62891793e-01 1.55798802e-02 4.03257817e-01
-1.04070342e+00 -6.92154169e-01 -6.67415857e-01 1.72607768e-02
1.37203887e-01 -6.83623016e-01 7.42311776e-01 -6.65293157e-01
-1.43816376e+00 6.24310732e-01 -3.74992251e-01 -2.03487113e-01
2.04149827e-01 -2.39236593e-01 -3.69559675e-01 1.89773351e-01
-3.24857906e-02 4.43904191e-01 1.18107498e+00 -1.03794813e+00
-3.56753379e-01 -4.84950662e-01 -1.26665294e-01 2.84542173e-01
-1.30759105e-01 9.15362760e-02 -6.13331914e-01 -9.53954220e-01
4.50993091e-01 -5.31528413e-01 5.03581017e-02 -1.77229181e-01
-2.84183621e-01 2.52249688e-01 6.19047225e-01 -1.21240866e+00
1.03722489e+00 -2.28645205e+00 3.26626629e-01 -2.89766878e-01
3.08882575e-02 3.04484636e-01 -3.67056578e-01 -1.27038479e-01
-2.68102735e-01 1.19032405e-01 -4.97463852e-01 -2.42274940e-01
-4.04956549e-01 -5.29343970e-02 -2.94321060e-01 8.48560333e-01
8.15342888e-02 8.98687065e-01 -5.84506691e-01 -7.33779594e-02
4.60468590e-01 7.02311873e-01 -5.42943180e-01 1.17059663e-01
3.38414609e-01 7.39512146e-01 -9.90354419e-02 6.50501311e-01
1.32452703e+00 -7.61334598e-02 -2.39158511e-01 -7.16023743e-01
-2.02422664e-01 1.97401196e-01 -8.40533078e-01 1.86628103e+00
-5.40456593e-01 7.54249573e-01 1.26423329e-01 -1.12783992e+00
7.21958697e-01 1.54466510e-01 2.06190169e-01 -9.94343579e-01
-2.90092058e-03 3.04596335e-01 -3.88493985e-02 -6.22308314e-01
3.16494018e-01 -1.79486006e-01 2.32449129e-01 1.91094518e-01
3.01695406e-01 2.17838567e-02 -5.03632165e-02 -1.19699799e-01
1.04219913e+00 -6.80014566e-02 8.46584439e-02 -3.46036255e-01
4.48499978e-01 -4.16711897e-01 6.06518865e-01 8.65111828e-01
-3.42160352e-02 1.16313219e+00 2.51181334e-01 -2.08887681e-01
-8.42740893e-01 -9.80458260e-01 -4.41460878e-01 7.49965310e-01
4.99958247e-01 -3.57813910e-02 -7.44834781e-01 -2.74787396e-01
-2.73632437e-01 3.17881495e-01 -4.15251225e-01 -2.59119123e-01
-6.61076307e-01 -1.35302138e+00 3.61453861e-01 2.10322306e-01
9.08241212e-01 -1.07380414e+00 -1.70747831e-01 1.00584447e-01
-6.81160510e-01 -1.09225857e+00 -7.53444195e-01 1.86402574e-02
-1.02531481e+00 -8.28559399e-01 -9.49871540e-01 -7.52646148e-01
7.37220228e-01 6.70563817e-01 8.17204773e-01 9.53940600e-02
-3.30983132e-01 -1.45718366e-01 -2.45056763e-01 2.68538266e-01
1.22597711e-02 -2.43723258e-01 -1.74853891e-01 1.29344225e-01
-1.00813426e-01 -8.77620816e-01 -9.32891071e-01 6.31402314e-01
-1.10802007e+00 2.42645945e-02 8.61245275e-01 1.07454813e+00
5.76445758e-01 2.44349375e-01 6.01713061e-01 -4.38928306e-01
5.88348091e-01 -2.62578577e-01 -3.68449628e-01 1.79194078e-01
-3.81578386e-01 9.36194416e-03 5.11596560e-01 -6.04962289e-01
-1.10616112e+00 -6.19543850e-01 -8.34936202e-02 -4.68150884e-01
-2.17303902e-01 4.35640365e-01 -3.91917199e-01 -4.34534699e-01
6.54634118e-01 4.23409402e-01 -4.98626940e-02 -8.49728942e-01
8.03796872e-02 6.29782557e-01 8.76515329e-01 -3.28768969e-01
7.16563880e-01 4.64781791e-01 -1.61851242e-01 -8.16990554e-01
-4.64896053e-01 -2.91638702e-01 -2.78071374e-01 -5.81844933e-02
7.83965945e-01 -1.21388888e+00 -5.61311007e-01 1.25007904e+00
-1.06045604e+00 -4.10033017e-01 5.27827144e-02 4.80819613e-01
-1.92273378e-01 4.61978674e-01 -8.64076674e-01 -2.36021057e-01
-3.25266421e-01 -1.44641697e+00 8.65485191e-01 4.27971244e-01
4.53563541e-01 -5.98767161e-01 -3.31071615e-01 4.44339901e-01
8.67006361e-01 -1.07649967e-01 8.21096957e-01 1.21425122e-01
-7.66093075e-01 -1.83686502e-02 -5.61253250e-01 7.18872964e-01
3.03288013e-01 -5.37061512e-01 -1.13456035e+00 -6.23015583e-01
3.30981046e-01 -1.75104104e-02 1.38638473e+00 8.56952965e-01
1.42813432e+00 -3.30340683e-01 -2.22154707e-02 1.12528658e+00
1.45166373e+00 1.65363148e-01 1.26051378e+00 3.78493160e-01
7.39235222e-01 2.92715639e-01 3.82421851e-01 2.06585765e-01
2.27070644e-01 8.54762375e-01 4.19950068e-01 -2.19622687e-01
-7.14760840e-01 1.41396642e-01 6.21027887e-01 7.93665290e-01
-1.53581306e-01 -2.76664793e-01 -5.10756433e-01 7.06025541e-01
-1.60196006e+00 -8.41144979e-01 3.34896259e-02 2.43330574e+00
1.12032390e+00 -3.79126340e-01 -6.15017831e-01 -4.48222160e-02
9.34238374e-01 3.77859890e-01 -6.55079782e-01 -4.34481464e-02
-6.06038630e-01 3.45439643e-01 6.42984807e-01 9.50500906e-01
-1.11232424e+00 8.58181417e-01 4.97340965e+00 1.28865576e+00
-1.28730118e+00 4.01228458e-01 9.95708525e-01 -2.12779254e-01
-1.06405787e-01 -7.59718269e-02 -5.50315678e-01 4.91598308e-01
6.92169070e-01 2.63047427e-01 9.34616804e-01 1.75607294e-01
3.47699225e-01 -4.13607389e-01 -6.64482117e-01 1.22555363e+00
2.97733188e-01 -1.12622881e+00 -3.76107208e-02 -7.24266693e-02
8.49604905e-01 9.67893898e-02 2.88195759e-01 7.65675679e-02
-2.17166036e-01 -1.26401913e+00 4.60495979e-01 6.73488438e-01
9.55313325e-01 -6.09338105e-01 9.91886020e-01 1.64287865e-01
-6.86100841e-01 -1.60641465e-02 -5.53124785e-01 1.21437669e-01
2.32555479e-01 1.06571281e+00 -1.52813485e-02 7.44739056e-01
1.10012162e+00 8.45761657e-01 -5.74207485e-01 1.20053697e+00
-3.65086734e-01 4.65845197e-01 -2.28498265e-01 6.77085757e-01
-2.88589150e-01 -2.59562105e-01 6.21281803e-01 1.04089200e+00
5.88900447e-01 2.07155213e-01 -3.58775616e-01 6.50131643e-01
-1.26395568e-01 -2.91226715e-01 -1.05005279e-01 3.80858481e-01
2.55325109e-01 1.15762973e+00 -3.01395923e-01 -2.39813432e-01
-1.84975326e-01 1.20459569e+00 -7.76722878e-02 8.11937392e-01
-7.92869389e-01 -6.04908466e-01 8.30339313e-01 4.85111363e-02
3.61049980e-01 -9.03767496e-02 -3.32189798e-01 -1.39499044e+00
2.96875566e-01 -1.00980330e+00 -2.44930238e-02 -9.65428829e-01
-1.15040267e+00 6.19515479e-01 -9.13755596e-02 -1.23097312e+00
1.96775138e-01 -4.73714530e-01 -4.40509140e-01 1.14395010e+00
-1.74728942e+00 -8.62741530e-01 -6.04419887e-01 7.47798920e-01
2.73630977e-01 1.31997228e-01 3.74447912e-01 6.87836051e-01
-6.48174584e-01 6.69223189e-01 1.34681210e-01 3.82555053e-02
1.16109943e+00 -6.76970840e-01 2.37835541e-01 1.24674189e+00
-3.37080061e-01 6.79345727e-01 5.37217915e-01 -5.86735308e-01
-1.32665837e+00 -9.39737201e-01 5.50215781e-01 4.71999198e-02
2.88544625e-01 1.93791259e-02 -1.14982080e+00 2.59158283e-01
2.55233467e-01 4.46939230e-01 3.56632099e-02 -2.76429981e-01
-3.49282146e-01 -4.73624289e-01 -1.28186142e+00 4.24121469e-01
1.03003597e+00 -6.96375728e-01 -2.64424905e-02 1.88283116e-01
5.86203277e-01 -6.68910801e-01 -7.59287119e-01 5.99829972e-01
3.12451959e-01 -1.17906570e+00 1.21954441e+00 -3.59682478e-02
4.99780864e-01 -5.93539298e-01 -2.40521356e-01 -1.36988103e+00
-4.84033048e-01 -5.51721156e-01 1.27253875e-01 1.12847304e+00
1.05248451e-01 -1.04122162e+00 1.10171482e-01 1.36680994e-02
-4.54135478e-01 -5.23858368e-01 -1.16029882e+00 -5.79347432e-01
5.69444522e-02 -8.36222917e-02 3.95963162e-01 9.26555157e-01
-4.54684734e-01 -4.16977964e-02 -5.98371029e-01 4.72679853e-01
7.71000624e-01 3.89193259e-02 4.30043727e-01 -7.11246550e-01
-2.69491196e-01 -3.78999799e-01 -4.00741071e-01 -1.43905354e+00
-1.65562779e-01 -7.65635848e-01 2.27321014e-01 -1.46005166e+00
3.81123126e-01 -2.29905173e-01 -5.19796312e-01 6.25470579e-01
-4.08289790e-01 6.30048394e-01 1.22525290e-01 3.28072488e-01
-1.77241310e-01 5.95950246e-01 1.71132600e+00 -2.70776749e-01
-9.80791003e-02 -2.82590061e-01 -8.15940857e-01 3.00133198e-01
1.02168739e+00 -5.07560492e-01 -1.92625314e-01 -9.64132845e-01
-8.75533000e-02 1.29087403e-01 7.51203835e-01 -1.10784161e+00
3.98219675e-01 1.24931578e-02 5.13566256e-01 -3.53428423e-01
3.51582915e-01 -3.72654349e-01 1.61246657e-01 1.87295452e-01
4.14227434e-02 -8.24150220e-02 3.58199179e-01 3.67712885e-01
-5.87608576e-01 -7.26879090e-02 1.00468957e+00 4.60693948e-02
-5.50480545e-01 2.44975790e-01 -1.06927276e-01 -1.51734143e-01
3.98740143e-01 -7.24155232e-02 -1.06716597e+00 -3.97228330e-01
-5.60483456e-01 -2.73490995e-01 4.82277572e-01 3.50165367e-01
7.37038672e-01 -1.18732584e+00 -8.94697309e-01 3.76292050e-01
-4.03131247e-01 -5.39688170e-02 9.20689583e-01 1.35245824e+00
-4.79744166e-01 2.40142673e-01 -3.98994625e-01 -6.82292879e-01
-1.06604207e+00 2.22810313e-01 5.25830686e-01 -4.27312329e-02
-7.17704654e-01 9.76351798e-01 5.99874437e-01 -1.13233037e-01
-5.71389757e-02 -4.90710735e-01 -1.41410083e-01 -8.05604607e-02
8.75503600e-01 3.26476336e-01 3.68298113e-01 -4.78416115e-01
-3.04838777e-01 1.07933962e+00 -1.56393632e-01 1.51868863e-02
1.37998819e+00 -2.88108408e-01 -6.12145603e-01 -5.02752932e-03
1.29503703e+00 6.25107512e-02 -1.41448486e+00 -2.57948607e-01
-6.23320878e-01 -8.54006886e-01 6.15205526e-01 -9.47353184e-01
-1.60529625e+00 1.03141320e+00 1.06330538e+00 -2.08001807e-01
1.80947649e+00 -1.77131578e-01 5.89202642e-01 -5.43084294e-02
3.39450210e-01 -5.81842721e-01 3.68562080e-02 4.56546217e-01
1.23091686e+00 -1.17235553e+00 1.71696499e-01 -3.47365469e-01
-3.66336673e-01 9.42412853e-01 3.96125495e-01 -1.29206613e-01
4.80781496e-01 2.49990389e-01 -2.08353065e-02 1.61120266e-01
-4.21345115e-01 4.57886197e-02 3.15751493e-01 5.38242996e-01
2.45663822e-01 -4.31879684e-02 -2.48126462e-01 6.22469604e-01
1.42124817e-01 -6.18736558e-02 5.01590014e-01 4.18251932e-01
-4.27084744e-01 -8.19273949e-01 -5.16052365e-01 2.83072978e-01
-5.38133025e-01 -5.79313517e-01 2.49663025e-01 3.53656381e-01
2.02990890e-01 1.13953686e+00 4.07896936e-02 -4.23931628e-01
1.42107964e-01 -3.80347192e-01 6.02168977e-01 -1.34629503e-01
-4.44843799e-01 1.71500802e-01 -1.72347903e-01 -7.13902175e-01
-4.17393327e-01 -4.71109986e-01 -9.16704834e-01 -5.06909788e-01
-3.77896339e-01 -2.24795625e-01 5.63720822e-01 7.43984640e-01
5.27778745e-01 7.29917109e-01 4.64546531e-01 -1.21718442e+00
-2.19707325e-01 -1.33512104e+00 -4.64445174e-01 2.50683874e-01
8.47783625e-01 -4.93243694e-01 -6.77779436e-01 2.00669318e-02] | [11.258380889892578, -2.2924392223358154] |
0335e57d-d400-4ffb-8c2d-ec3d6bec69b3 | look-listen-and-attend-co-attention-network | 2008.05789 | null | https://arxiv.org/abs/2008.05789v1 | https://arxiv.org/pdf/2008.05789v1.pdf | Look, Listen, and Attend: Co-Attention Network for Self-Supervised Audio-Visual Representation Learning | When watching videos, the occurrence of a visual event is often accompanied by an audio event, e.g., the voice of lip motion, the music of playing instruments. There is an underlying correlation between audio and visual events, which can be utilized as free supervised information to train a neural network by solving the pretext task of audio-visual synchronization. In this paper, we propose a novel self-supervised framework with co-attention mechanism to learn generic cross-modal representations from unlabelled videos in the wild, and further benefit downstream tasks. Specifically, we explore three different co-attention modules to focus on discriminative visual regions correlated to the sounds and introduce the interactions between them. Experiments show that our model achieves state-of-the-art performance on the pretext task while having fewer parameters compared with existing methods. To further evaluate the generalizability and transferability of our approach, we apply the pre-trained model on two downstream tasks, i.e., sound source localization and action recognition. Extensive experiments demonstrate that our model provides competitive results with other self-supervised methods, and also indicate that our approach can tackle the challenging scenes which contain multiple sound sources. | ['Ruize Wang', 'Ying Cheng', 'Zhihao Pan', 'Yuejie Zhang', 'Rui Feng'] | 2020-08-13 | null | null | null | null | ['audio-visual-synchronization', 'audio-visual-synchronization'] | ['audio', 'computer-vision'] | [ 2.51691014e-01 -4.07959670e-01 -1.12639643e-01 -1.56549945e-01
-9.83704388e-01 -3.80838126e-01 4.48065788e-01 -1.97788015e-01
-2.38465726e-01 2.60380864e-01 5.52567899e-01 4.29508746e-01
2.83151984e-01 -1.03959166e-01 -8.48777354e-01 -7.19919860e-01
1.80347695e-03 -2.26231381e-01 3.84397835e-01 1.45650938e-01
1.83674514e-01 6.70056269e-02 -1.89802647e+00 6.79119408e-01
3.39734137e-01 1.18122399e+00 4.69804287e-01 4.95734572e-01
1.47030428e-01 1.09142625e+00 -5.42352200e-01 2.06573028e-02
-1.71042383e-01 -6.72950804e-01 -7.36131728e-01 1.91748247e-01
5.23817360e-01 -1.89526364e-01 -4.56159294e-01 1.04543865e+00
6.58717036e-01 2.70460218e-01 4.87811774e-01 -1.47430551e+00
-4.03493941e-01 5.54792643e-01 -6.61747992e-01 4.84640270e-01
6.28991365e-01 3.10176224e-01 1.20407832e+00 -1.01902676e+00
3.04798573e-01 1.20137990e+00 5.20464659e-01 4.29237098e-01
-9.78282392e-01 -8.45211923e-01 2.88555413e-01 6.06144011e-01
-1.36356270e+00 -9.01987731e-01 1.10398543e+00 -3.62949401e-01
6.67824030e-01 8.02254453e-02 5.51624775e-01 1.48242831e+00
-1.36465028e-01 1.01516712e+00 7.18762696e-01 -4.32185799e-01
9.27768797e-02 -1.36599541e-01 -3.63101989e-01 4.10787284e-01
-3.84463549e-01 -1.48557443e-02 -1.01585877e+00 5.02786189e-02
6.41715109e-01 -1.98114980e-02 -5.47660232e-01 -1.16822779e-01
-1.43494785e+00 5.76266944e-01 3.33004117e-01 5.11120737e-01
-1.92790449e-01 4.08940792e-01 5.70245922e-01 -3.48539948e-02
5.01588166e-01 1.20910153e-01 -3.00901800e-01 -2.62119442e-01
-8.45316947e-01 -2.61907995e-01 4.31115448e-01 6.90748751e-01
4.73862469e-01 1.35786384e-01 -2.91481078e-01 9.20513213e-01
4.71468598e-01 3.95769417e-01 7.50182331e-01 -1.04244411e+00
4.26557779e-01 1.02616310e-01 -9.60209444e-02 -1.03334641e+00
-2.33151972e-01 -3.34974915e-01 -7.47765660e-01 -1.30422622e-01
1.64794981e-01 3.50322551e-03 -5.44571757e-01 2.08628416e+00
3.00415814e-01 9.43837464e-01 -4.81285639e-02 1.07143199e+00
1.11533546e+00 8.11022758e-01 1.39976442e-01 -4.71876115e-01
1.26779008e+00 -1.22689271e+00 -8.78114402e-01 -2.73088038e-01
1.65396050e-01 -8.31949532e-01 1.03603876e+00 1.92182809e-01
-9.69202697e-01 -1.04492342e+00 -6.05996966e-01 5.75850941e-02
1.20861880e-01 3.59060526e-01 3.87650520e-01 -7.54684806e-02
-8.20792735e-01 5.37008941e-01 -9.23570096e-01 -4.13521588e-01
4.55319822e-01 -4.64410223e-02 -3.58667970e-01 2.26927683e-01
-1.07351267e+00 3.19853842e-01 1.99192882e-01 1.09608404e-01
-1.34608483e+00 -6.06950939e-01 -9.33958411e-01 6.73901141e-02
4.42727208e-01 -4.00099218e-01 1.25096774e+00 -1.36402452e+00
-1.58173954e+00 8.63322496e-01 -3.71130675e-01 -2.06574589e-01
1.00127280e-01 -4.81193960e-01 -6.11101568e-01 5.47885656e-01
3.95077974e-01 6.64133132e-01 1.19697928e+00 -1.18879879e+00
-6.47946060e-01 -2.66402457e-02 4.57658172e-02 3.00872028e-01
-5.14782012e-01 4.27491665e-01 -7.18487620e-01 -7.45958745e-01
-1.66436166e-01 -7.62597442e-01 2.43002698e-01 -7.36966357e-02
-3.39230210e-01 -3.87820423e-01 7.56494045e-01 -3.32706004e-01
1.16881263e+00 -2.74540830e+00 1.70952022e-01 -2.62614101e-01
-1.00995190e-01 1.20306820e-01 -4.18702006e-01 4.20660615e-01
-3.44479710e-01 -2.37708017e-01 -2.73360242e-03 -4.43818808e-01
-1.51041910e-01 4.53341790e-02 -5.60033619e-01 5.23288667e-01
2.48506874e-01 7.28185773e-01 -1.14203894e+00 -6.42852008e-01
-5.38465939e-02 5.55911779e-01 -3.88796151e-01 6.19721115e-01
-3.65067944e-02 7.88418531e-01 -2.98110247e-01 5.87000847e-01
1.24379791e-01 -3.91813040e-01 -6.97983103e-03 -5.54307044e-01
7.67755322e-04 4.34324175e-01 -1.11424851e+00 2.17135048e+00
-4.71937180e-01 8.63590419e-01 4.37492952e-02 -1.16472971e+00
5.53464055e-01 5.22867084e-01 5.96654475e-01 -5.83833039e-01
8.60737637e-02 -1.00596815e-01 -3.25395092e-02 -9.77168143e-01
-8.89223889e-02 -5.68065792e-02 6.12141639e-02 4.99438763e-01
5.02109885e-01 2.06473574e-01 3.18142064e-02 1.02527976e-01
1.02084768e+00 3.22547972e-01 2.09799767e-01 1.28698960e-01
6.17778599e-01 -5.14341116e-01 5.33879578e-01 4.64647740e-01
-1.91546112e-01 8.51598680e-01 3.89353096e-01 3.16291489e-02
-4.86147523e-01 -9.62936640e-01 3.48467901e-02 1.48728967e+00
3.54491204e-01 -6.45684540e-01 -5.30497611e-01 -6.68682516e-01
-2.32293889e-01 2.62508243e-01 -5.48005342e-01 -2.39970714e-01
-4.13920850e-01 -4.24653471e-01 6.04483545e-01 7.72738755e-01
4.27560776e-01 -1.34829807e+00 -6.27947927e-01 1.15642203e-02
-6.80894256e-01 -1.32232714e+00 -6.68394327e-01 9.03494582e-02
-5.94909549e-01 -1.02240610e+00 -6.45623446e-01 -8.97560179e-01
3.47217441e-01 5.34140110e-01 1.00570786e+00 -1.10965356e-01
-2.78045118e-01 8.16291869e-01 -4.69603002e-01 -2.36055210e-01
-2.91995350e-02 -2.68901378e-01 1.11822911e-01 7.08382308e-01
1.32306561e-01 -8.06610227e-01 -7.86609948e-01 3.25816453e-01
-8.14359367e-01 -1.47883430e-01 3.63755256e-01 8.15170407e-01
6.37450755e-01 5.39888702e-02 6.67630792e-01 -3.05906326e-01
2.95805782e-01 -6.07310176e-01 -3.51042263e-02 -3.92993428e-02
1.27102122e-01 -2.35824600e-01 4.60594803e-01 -8.31688344e-01
-9.86763775e-01 4.31022763e-01 1.59532558e-02 -9.40104723e-01
-4.64828163e-01 3.12893689e-01 -4.18897331e-01 1.50902882e-01
3.74068171e-01 1.90536737e-01 -1.03199571e-01 -5.18384039e-01
3.10374707e-01 6.76594555e-01 7.40533233e-01 -2.72983998e-01
5.72321951e-01 6.77955508e-01 -2.51524955e-01 -7.96625555e-01
-1.26143026e+00 -7.67999887e-01 -5.19274294e-01 -4.91826952e-01
9.85008359e-01 -1.25597835e+00 -6.47037506e-01 4.68323350e-01
-1.22797275e+00 -3.57953995e-01 -1.92822769e-01 7.71013796e-01
-7.33151317e-01 4.42186683e-01 -4.35024261e-01 -7.81631470e-01
1.32318169e-01 -9.22340930e-01 1.31624353e+00 1.99283704e-01
-1.90462291e-01 -8.32300067e-01 4.45093155e-01 3.79535347e-01
6.30182251e-02 4.92727337e-03 4.33086455e-01 -6.96770251e-01
-3.31286371e-01 8.09260160e-02 -5.90963848e-02 4.38909441e-01
3.31658661e-01 -1.04455456e-01 -1.45507455e+00 -1.38135493e-01
2.38452423e-02 -6.40773892e-01 1.06439424e+00 2.42420524e-01
1.43193507e+00 -1.81948066e-01 -2.57665634e-01 5.76928616e-01
9.72611129e-01 9.91871431e-02 5.46277523e-01 -8.41892045e-03
7.52079606e-01 6.24856055e-01 6.71010673e-01 4.53719884e-01
2.99222082e-01 9.03939188e-01 6.03708804e-01 -2.70200133e-01
-3.67657810e-01 -4.25244480e-01 6.85610175e-01 1.02163148e+00
-7.86809772e-02 -1.65258929e-01 -5.36504865e-01 6.60638571e-01
-2.04411697e+00 -1.20060766e+00 7.05267861e-02 1.97457910e+00
9.59166706e-01 -1.48877293e-01 2.41871208e-01 1.97662503e-01
8.38699460e-01 4.54407185e-01 -4.23928231e-01 2.86280304e-01
1.24851037e-02 3.44182462e-01 -2.53292561e-01 1.25640765e-01
-1.50080097e+00 9.50189114e-01 6.09591198e+00 9.18352425e-01
-1.31062031e+00 3.95647764e-01 3.51031721e-01 -2.91969031e-01
1.94282427e-01 -2.19674870e-01 -4.78631377e-01 6.60933435e-01
8.87684703e-01 9.89213958e-02 3.68676931e-01 6.26271665e-01
2.84386277e-01 1.46469725e-02 -1.40026212e+00 1.27699578e+00
5.48109174e-01 -9.88363445e-01 -1.84445366e-01 -3.43715966e-01
5.66601574e-01 -6.69202283e-02 5.02325073e-02 2.29945809e-01
-8.21875706e-02 -9.36538815e-01 8.36222887e-01 4.84909177e-01
7.84405470e-01 -5.90218127e-01 4.49400991e-01 1.62563682e-01
-1.49957669e+00 -9.46748853e-02 -1.52115792e-01 -1.88352750e-03
2.54740208e-01 2.61014044e-01 -3.34867448e-01 5.20009935e-01
1.11365521e+00 1.18621016e+00 -3.57601225e-01 1.10244048e+00
-5.37234902e-01 8.62846971e-01 -2.12022483e-01 3.36703330e-01
1.30965794e-02 3.04455012e-01 5.71892977e-01 1.31988621e+00
2.76018113e-01 -4.22868803e-02 2.34421998e-01 7.72182167e-01
1.52308103e-02 1.98702633e-01 -6.36641204e-01 -5.78143522e-02
3.26886326e-01 1.27131522e+00 -4.92755502e-01 -1.74045756e-01
-5.89508295e-01 1.07837522e+00 2.68968433e-01 5.20421386e-01
-1.14681304e+00 -3.25639129e-01 5.48546493e-01 -1.33870304e-01
5.83811879e-01 3.36409062e-02 2.84739196e-01 -1.29713643e+00
8.15601870e-02 -8.93261850e-01 3.76704574e-01 -1.14611685e+00
-1.35639048e+00 6.00896001e-01 -1.20923094e-01 -1.69570684e+00
-2.45367765e-01 -3.13999206e-01 -1.00903535e+00 3.80507678e-01
-1.49547374e+00 -9.79528725e-01 -4.35655951e-01 1.01734865e+00
6.28090203e-01 -2.41939396e-01 6.66449189e-01 4.57245201e-01
-6.28288090e-01 5.97459614e-01 -1.85929447e-01 3.12260181e-01
1.04426289e+00 -8.04186881e-01 -2.46234760e-01 8.43822658e-01
8.65355134e-01 2.42765754e-01 4.01589811e-01 -3.58646065e-01
-1.30286193e+00 -1.15323973e+00 6.04313970e-01 -2.37439469e-01
9.15298223e-01 -3.16480160e-01 -9.02583122e-01 6.06551409e-01
3.30146074e-01 3.58670712e-01 9.36053932e-01 -7.53607005e-02
-4.92666304e-01 -3.54647934e-01 -4.42263305e-01 2.41881803e-01
1.17406821e+00 -9.53340948e-01 -7.73409069e-01 2.73214966e-01
7.07451940e-01 -2.19056398e-01 -4.18056011e-01 4.14739966e-01
4.67832655e-01 -9.74351943e-01 1.11469579e+00 -6.45714819e-01
5.97960591e-01 -3.32508683e-01 -2.27185622e-01 -1.27079868e+00
-3.00506175e-01 -7.79959738e-01 -3.12263221e-01 1.66039455e+00
1.37854576e-01 -1.41312137e-01 1.15612045e-01 -1.85705751e-01
-6.89533874e-02 -3.84794712e-01 -1.14626491e+00 -7.21905351e-01
-5.62157333e-01 -6.82899237e-01 9.55826044e-02 1.02075040e+00
1.89131036e-01 5.83396912e-01 -7.04664052e-01 2.80743241e-01
4.82078582e-01 2.54322767e-01 7.53150702e-01 -1.09668565e+00
-5.50217867e-01 -3.18257362e-01 -4.13139313e-01 -1.24076176e+00
6.30740345e-01 -7.83060133e-01 4.43152219e-01 -1.12539363e+00
3.87400120e-01 8.59027654e-02 -7.64011204e-01 5.95934212e-01
-1.85095489e-01 5.21113873e-01 2.06231847e-01 3.95779938e-01
-1.14205968e+00 8.20922792e-01 1.06927061e+00 -1.66233599e-01
-2.03312889e-01 -1.03670441e-01 -5.53553462e-01 9.77453291e-01
4.87003833e-01 -5.54406941e-01 -4.66876656e-01 -4.49487448e-01
-2.15775184e-02 1.64004922e-01 6.82883263e-01 -1.04288578e+00
3.21160853e-01 -4.79413681e-02 2.13407382e-01 -3.73154074e-01
4.49997663e-01 -7.55877614e-01 -1.97441295e-01 4.40310128e-02
-5.90043426e-01 -3.40712130e-01 2.20901921e-01 8.91389728e-01
-6.89036489e-01 -9.68677253e-02 7.50963569e-01 2.97047645e-02
-6.13126993e-01 1.61473706e-01 -2.92335480e-01 2.47684330e-01
8.80648196e-01 1.45215347e-01 -9.91793871e-02 -6.63123608e-01
-8.51257622e-01 7.47216418e-02 -8.33112150e-02 6.23838663e-01
7.00438261e-01 -1.67586792e+00 -5.52735806e-01 1.96466282e-01
2.08223790e-01 -2.12638468e-01 3.49526435e-01 1.07880878e+00
1.86990544e-01 1.77639917e-01 -2.25613490e-01 -9.94291723e-01
-1.35200942e+00 6.59624636e-01 1.53710768e-01 1.25832736e-01
-7.43520260e-01 9.70107794e-01 7.10330486e-01 2.57326633e-01
7.20420599e-01 -2.68162757e-01 -5.19407034e-01 2.95273602e-01
7.33712912e-01 2.25546271e-01 -3.78774166e-01 -8.15438151e-01
-5.74732959e-01 6.93609416e-01 3.06957185e-01 -1.64702222e-01
1.32200253e+00 -2.66371816e-01 6.20142519e-02 9.94411945e-01
1.33776057e+00 1.12576738e-01 -1.49447918e+00 -3.71735394e-01
-2.98582047e-01 -4.44586903e-01 8.52996483e-02 -5.05910277e-01
-1.48052609e+00 1.18331635e+00 6.61070228e-01 1.25916898e-01
1.40647840e+00 5.40000618e-01 5.91620147e-01 1.37076601e-01
9.51251686e-02 -9.42354918e-01 7.32599318e-01 2.91660041e-01
1.09349191e+00 -1.25517821e+00 -3.58619422e-01 -2.50546992e-01
-8.32973540e-01 1.03237939e+00 6.45366967e-01 -8.64091516e-02
5.89302540e-01 1.52230337e-01 8.66275057e-02 -9.13621709e-02
-8.73405576e-01 -6.13711715e-01 5.62564313e-01 6.06596589e-01
4.68182594e-01 -3.97459000e-01 2.11067244e-01 7.62969136e-01
2.00058743e-01 -3.52388062e-02 6.33841008e-02 6.82631493e-01
-2.65604496e-01 -6.77061677e-01 -2.66223669e-01 -1.94474518e-01
-6.81592643e-01 -9.85213667e-02 -4.16198343e-01 5.01168549e-01
1.41629577e-01 1.14633036e+00 2.58945793e-01 -4.87803578e-01
2.99535245e-01 1.75231025e-01 4.79359806e-01 -6.34090722e-01
-4.19314653e-01 7.10718334e-01 -1.34954438e-01 -8.98328722e-01
-1.00790143e+00 -6.56081438e-01 -1.11103356e+00 5.47923982e-01
-2.29632139e-01 4.94667748e-03 1.11779317e-01 1.04879546e+00
4.66932416e-01 7.17503667e-01 9.51828897e-01 -1.10292721e+00
-2.50091195e-01 -1.07710838e+00 -4.42355603e-01 7.24976778e-01
5.22980392e-01 -9.16677237e-01 -6.69631958e-01 5.23722112e-01] | [14.709012985229492, 4.927501201629639] |
b10d36be-f41a-42d4-b421-afbdabf088dd | neural-architecture-search-for-parameter | 2305.16597 | null | https://arxiv.org/abs/2305.16597v1 | https://arxiv.org/pdf/2305.16597v1.pdf | Neural Architecture Search for Parameter-Efficient Fine-tuning of Large Pre-trained Language Models | Parameter-efficient tuning (PET) methods fit pre-trained language models (PLMs) to downstream tasks by either computing a small compressed update for a subset of model parameters, or appending and fine-tuning a small number of new model parameters to the pre-trained network. Hand-designed PET architectures from the literature perform well in practice, but have the potential to be improved via automated neural architecture search (NAS). We propose an efficient NAS method for learning PET architectures via structured and unstructured pruning. We present experiments on GLUE demonstrating the effectiveness of our algorithm and discuss how PET architectural design choices affect performance in practice. | ['Greg Ver Steeg', 'Aram Galstyan', 'Govind Thattai', 'Anoop Kumar', 'Neal Lawton'] | 2023-05-26 | null | null | null | null | ['architecture-search'] | ['methodology'] | [ 6.60098568e-02 2.53401130e-01 -4.79095966e-01 -6.96756542e-01
-8.06423843e-01 -5.57106972e-01 5.08083880e-01 -1.54020861e-01
-8.63708436e-01 6.81533873e-01 2.63157934e-01 -5.51970422e-01
-8.39689225e-02 -2.90338486e-01 -7.73706913e-01 -4.42854494e-01
-4.89234217e-02 1.14483690e+00 2.24431902e-01 1.19313132e-02
1.42854825e-01 5.45095384e-01 -1.17139876e+00 2.13043213e-01
4.06285852e-01 8.04405689e-01 5.64825952e-01 6.08476281e-01
-1.84746325e-01 4.88961786e-01 -4.58394974e-01 -4.24755156e-01
3.53418171e-01 -6.03160709e-02 -7.90068865e-01 -3.91701043e-01
9.60392281e-02 -4.37365830e-01 -2.65414059e-01 6.23005271e-01
4.54384774e-01 2.38312319e-01 5.63684225e-01 -7.76000619e-01
-2.21095458e-01 1.28803527e+00 -3.00100744e-01 5.54096043e-01
-7.59825110e-01 1.51582479e-01 1.06280255e+00 -1.03115237e+00
3.60798329e-01 1.07701528e+00 8.54314983e-01 7.21725166e-01
-1.55829406e+00 -7.65054643e-01 4.32403326e-01 -7.76667073e-02
-1.42238414e+00 -1.18198454e+00 3.43426585e-01 -1.73580289e-01
1.91773391e+00 -8.30509067e-02 6.46203160e-01 1.04682767e+00
3.31837311e-02 4.74046946e-01 4.82357323e-01 -6.12007856e-01
3.64948809e-01 2.41414219e-01 9.52431336e-02 8.00956249e-01
2.56204188e-01 4.06009674e-01 -5.61059713e-01 -5.11833310e-01
8.03728461e-01 -4.45256412e-01 9.58589911e-02 -3.45558226e-01
-9.14215982e-01 8.65598798e-01 3.40618789e-01 3.36437285e-01
-3.16587925e-01 7.60991395e-01 4.77226824e-01 3.56685549e-01
1.97781727e-01 8.06118011e-01 -1.14831567e+00 -2.00257733e-01
-1.35231888e+00 7.42356032e-02 7.68870652e-01 9.86413479e-01
7.85219371e-01 3.00079614e-01 -1.08619340e-01 1.07834375e+00
4.14561212e-01 2.29494236e-02 9.62973237e-01 -1.02099609e+00
6.16016805e-01 1.66347802e-01 -1.25830218e-01 -1.79881379e-01
-5.11879563e-01 -6.55313015e-01 -5.98902464e-01 -4.26198989e-01
5.85107468e-02 -2.96596587e-01 -1.18180406e+00 2.03286457e+00
9.14907753e-02 3.73776746e-03 -2.22568110e-01 3.70707989e-01
2.91949391e-01 7.00265765e-01 4.86867934e-01 -2.35603794e-01
7.72529840e-01 -1.37053192e+00 -2.29622245e-01 -8.63214791e-01
1.04451180e+00 -4.33819532e-01 1.11801052e+00 1.44237608e-01
-1.40217125e+00 -3.27088475e-01 -9.71840858e-01 -2.11631864e-01
8.55882477e-04 4.85242933e-01 7.40344584e-01 6.87089801e-01
-1.50753748e+00 7.85913706e-01 -1.20674062e+00 -1.95012704e-01
6.12722993e-01 1.05310893e+00 8.68103206e-02 1.38026893e-01
-9.28742528e-01 9.17675793e-01 8.14534009e-01 -7.14475438e-02
-1.16604793e+00 -9.66722369e-01 -3.71353686e-01 4.67539757e-01
2.69434988e-01 -1.19849432e+00 1.74358130e+00 -7.32393861e-01
-1.83266890e+00 6.58646703e-01 -3.26989710e-01 -9.74326134e-01
4.83858176e-02 -2.64919937e-01 -1.22466184e-01 -1.25631643e-03
-3.82241637e-01 1.24263990e+00 9.70148563e-01 -7.20963240e-01
-5.27905047e-01 -4.71121967e-02 -6.23776801e-02 1.33880183e-01
-6.81534469e-01 2.09385499e-01 -7.16461658e-01 -6.15158975e-01
1.44081950e-01 -9.15070772e-01 -6.49910212e-01 -1.19680651e-01
-2.80444473e-01 -9.61493850e-02 2.29842782e-01 -6.17903709e-01
1.72292948e+00 -1.99413359e+00 1.43291593e-01 4.16184604e-01
2.04837471e-01 4.88487661e-01 -3.26228559e-01 1.61284789e-01
1.64420813e-01 3.76334846e-01 -1.47410795e-01 -6.90888286e-01
1.45561052e-02 3.40035290e-01 -4.65045869e-01 -6.04866967e-02
6.14668690e-02 1.02205598e+00 -3.94470662e-01 -5.14655590e-01
-2.15784490e-01 1.97645277e-01 -9.92156029e-01 6.20040623e-03
-5.73002517e-01 3.56943272e-02 -4.47068781e-01 3.87605458e-01
5.69025129e-02 -4.92660463e-01 3.22955102e-01 -2.09776033e-02
1.40660182e-01 8.59171093e-01 -8.10463667e-01 1.58252311e+00
-5.35488844e-01 4.47712660e-01 1.86955646e-01 -9.79129553e-01
8.27540219e-01 2.70086288e-01 6.38485625e-02 -2.88467884e-01
2.31866971e-01 3.66825223e-01 1.76450551e-01 7.73614272e-02
2.27487013e-01 -1.30552351e-01 2.24605501e-01 8.43316197e-01
3.33678931e-01 -2.60621458e-02 1.69091210e-01 1.71327926e-02
1.23800981e+00 -6.09013364e-02 4.60035354e-01 -1.64854273e-01
1.92948699e-01 1.12865977e-01 4.42482352e-01 1.00040638e+00
-7.43037537e-02 1.33068964e-01 1.94345206e-01 -4.38597471e-01
-1.50098073e+00 -8.04348052e-01 6.35079145e-02 1.66069520e+00
-7.57074714e-01 -7.08149731e-01 -8.46500993e-01 -5.30118287e-01
-1.96038127e-01 7.02789903e-01 -3.47378016e-01 -3.12152922e-01
-9.45122957e-01 -6.88305020e-01 8.00916135e-01 8.50987196e-01
3.43312860e-01 -9.66576397e-01 -4.95472491e-01 5.79193950e-01
2.94415325e-01 -8.11210036e-01 -5.96634150e-01 9.13799167e-01
-1.43761575e+00 -2.14782655e-01 -3.97488981e-01 -8.62228513e-01
6.07116222e-01 7.12086782e-02 1.30965877e+00 3.46817613e-01
5.26775643e-02 -1.78370520e-01 3.14800441e-01 -2.90321141e-01
-5.83422601e-01 8.11890483e-01 3.19345862e-01 -5.70878267e-01
3.75322044e-01 -7.10342884e-01 -3.74579072e-01 1.88734785e-01
-5.09555101e-01 2.82753438e-01 8.84480774e-01 9.98139501e-01
8.23093832e-01 -2.65015483e-01 5.56745470e-01 -1.11565101e+00
6.76348865e-01 -2.42206737e-01 -8.58876765e-01 3.50285411e-01
-1.00005066e+00 5.39792478e-01 6.56770289e-01 -5.82709968e-01
-1.05258787e+00 3.28788936e-01 -1.13207325e-01 -5.83839238e-01
6.02232385e-03 6.11929536e-01 2.18154863e-01 -9.87260938e-02
8.41838837e-01 -7.54226372e-02 -2.64986545e-01 -8.85210574e-01
4.99834925e-01 2.86734492e-01 3.98891479e-01 -7.26880848e-01
7.13582575e-01 1.19344711e-01 -2.39394754e-01 -3.86460006e-01
-8.49326849e-01 -2.06960037e-01 -6.39370859e-01 5.23881614e-01
3.51226211e-01 -1.08059597e+00 1.31472936e-02 -1.22077338e-01
-8.99831116e-01 -8.52925718e-01 -2.98719734e-01 4.63694781e-01
-6.56427145e-01 -1.17395781e-01 -7.12123215e-01 -3.42379123e-01
-7.18743742e-01 -1.03188527e+00 7.06625462e-01 -1.89783107e-02
-6.12031341e-01 -8.66135716e-01 1.51340812e-01 1.29579857e-01
6.97817266e-01 -9.30399060e-01 1.45731521e+00 -9.90932882e-01
-4.87045109e-01 2.54225954e-02 -2.00437773e-02 2.24445313e-01
-5.27390718e-01 -1.18058793e-01 -7.96744525e-01 -3.77183825e-01
-1.62323773e-01 -4.46944505e-01 1.12940145e+00 7.44426668e-01
1.38154256e+00 -4.22611833e-01 -9.18660760e-01 9.24054325e-01
1.19592738e+00 1.23581506e-01 6.84891492e-02 4.27441955e-01
3.60469162e-01 1.44315258e-01 1.91711619e-01 2.73827851e-01
-7.00390562e-02 5.77430964e-01 -1.47866502e-01 3.92265260e-01
-3.22918594e-02 -4.29946661e-01 2.85268426e-01 1.14513040e+00
2.41326243e-01 -9.30895805e-02 -1.05968177e+00 5.55113792e-01
-1.74849153e+00 -4.64419872e-01 6.16172493e-01 1.88916719e+00
1.30290198e+00 6.40632868e-01 -4.50146161e-02 -2.11350426e-01
5.65378249e-01 -2.03014657e-01 -9.20546651e-01 -6.78182304e-01
6.77146614e-02 5.89312255e-01 7.00808287e-01 4.49884087e-01
-7.34082758e-01 1.46495259e+00 7.92750645e+00 9.44125772e-01
-9.63279366e-01 3.76059890e-01 8.37490976e-01 -6.94476068e-01
-4.11725968e-01 3.61725271e-01 -1.36802995e+00 2.52264831e-03
1.76447284e+00 -2.28981867e-01 7.53435850e-01 1.11697924e+00
1.39247596e-01 1.88770965e-01 -1.33341634e+00 8.00988376e-01
-3.52808267e-01 -1.71087444e+00 2.03207225e-01 -1.10340975e-01
7.74470091e-01 7.36916423e-01 4.99841310e-02 6.22897387e-01
6.15557194e-01 -1.00321257e+00 6.65429652e-01 2.37036809e-01
7.45036721e-01 -5.46449125e-01 9.24516246e-02 5.34874678e-01
-8.62761080e-01 -5.47410905e-01 -7.17299819e-01 3.91722098e-02
-1.41991684e-02 2.30324075e-01 -1.31053901e+00 -4.17132348e-01
7.38557220e-01 1.90596715e-01 -6.57538056e-01 1.11212397e+00
-1.49483860e-01 1.14044356e+00 -8.36540222e-01 -1.28573462e-01
4.37496603e-01 8.55258703e-02 3.24289024e-01 1.06177318e+00
2.53007174e-01 -8.37780312e-02 -3.57815117e-01 9.74404693e-01
-2.62240589e-01 1.95107190e-03 -1.87765360e-01 -5.74543238e-01
1.04719281e+00 1.05050480e+00 -4.97440338e-01 -3.81603360e-01
-1.44370586e-01 4.62964326e-01 9.33967948e-01 4.49931651e-01
-4.95363861e-01 1.59140048e-03 5.24346590e-01 -6.77940622e-02
6.45252526e-01 -3.45200241e-01 -5.41227460e-01 -8.10416341e-01
-3.49924594e-01 -7.77650058e-01 4.09027457e-01 -6.24882638e-01
-9.17498171e-01 6.96410656e-01 7.11677363e-03 -5.29589832e-01
-7.12200403e-01 -5.19715369e-01 -4.76213813e-01 7.82402217e-01
-1.27754068e+00 -8.86811435e-01 4.88004237e-01 3.04145277e-01
8.12319100e-01 -5.62396526e-01 1.03600824e+00 5.52444458e-02
-8.32463145e-01 9.94789898e-01 3.26065302e-01 -2.94389397e-01
5.81816792e-01 -7.18094945e-01 7.44058847e-01 5.48484802e-01
4.29906607e-01 9.68934298e-01 6.17091835e-01 -4.26331967e-01
-1.06641662e+00 -1.05138242e+00 1.11376536e+00 -7.08325207e-02
6.48378432e-01 -4.20496911e-01 -7.89027631e-01 1.21270514e+00
-7.70917581e-03 -2.63322115e-01 6.61292493e-01 3.45323116e-01
-4.02111709e-01 -7.27708042e-02 -8.38335931e-01 7.84704506e-01
1.14844322e+00 -3.48866284e-01 -4.96038586e-01 2.79429883e-01
1.06628942e+00 -2.76195347e-01 -5.75304091e-01 2.84017622e-01
3.75746340e-01 -4.98459339e-01 8.93833041e-01 -1.00794911e+00
4.55344506e-02 1.47036597e-01 -7.54491091e-02 -1.40397906e+00
-5.43566048e-01 -1.02613640e+00 -5.89649618e-01 9.78173494e-01
9.70923424e-01 -4.57316041e-01 9.46059227e-01 8.40972662e-01
-2.73595721e-01 -1.03863513e+00 -9.49833572e-01 -5.18012583e-01
3.89412522e-01 -5.29390872e-01 8.49815130e-01 4.38096672e-01
-2.62530625e-01 6.99702561e-01 -1.08490974e-01 1.88104308e-03
1.63289905e-01 -2.12355152e-01 4.90687668e-01 -1.16661775e+00
-8.47412586e-01 -7.90668547e-01 3.70626241e-01 -1.31582475e+00
2.53480822e-01 -9.59552586e-01 -1.34384841e-01 -1.13852370e+00
1.73210531e-01 -8.66876245e-01 -3.92678201e-01 7.77755380e-01
1.48091063e-01 -1.02436818e-01 8.03662278e-03 5.28882325e-01
-4.84704703e-01 4.52720702e-01 5.97529292e-01 -5.78511097e-02
-3.61705691e-01 1.17552295e-01 -7.49890625e-01 8.38598728e-01
9.49340463e-01 -7.34689534e-01 -6.87694550e-01 -1.04075110e+00
4.78802800e-01 -2.32884139e-01 -1.87614173e-01 -7.91629314e-01
5.25545001e-01 -2.14061681e-02 4.84110445e-01 -5.05831957e-01
4.27063406e-01 -5.99595189e-01 4.17941026e-02 4.23109680e-01
-8.00414085e-01 2.67570078e-01 4.79350060e-01 5.34266055e-01
2.46375903e-01 -7.54680991e-01 8.29867899e-01 -2.89518803e-01
-6.97042823e-01 3.82624090e-01 -5.01835167e-01 1.42599293e-03
5.27209818e-01 -2.70893034e-02 -1.12636000e-01 -1.66646287e-01
-8.81409407e-01 2.10690781e-01 5.46277948e-02 1.55900866e-01
4.57516372e-01 -9.10751939e-01 -2.85495758e-01 3.20010781e-01
-1.55741826e-01 -9.05933604e-02 -1.95393518e-01 6.23829722e-01
-3.64793271e-01 8.93719018e-01 5.52369617e-02 -3.44180882e-01
-9.64041710e-01 3.38447750e-01 5.05845964e-01 -5.54988325e-01
-5.58851719e-01 1.37192965e+00 1.01806246e-01 -6.03331387e-01
4.87315297e-01 -3.50103974e-01 1.26304924e-01 -3.41014326e-01
3.83921295e-01 9.61482748e-02 2.58988023e-01 -1.43291324e-01
-1.37572974e-01 1.57258138e-02 -6.17732227e-01 -3.37887555e-01
1.41731215e+00 -1.99343577e-01 2.83583477e-02 3.20943981e-01
1.04407978e+00 -5.81006825e-01 -1.32421994e+00 -7.25309193e-01
3.68575692e-01 1.31814465e-01 5.45227170e-01 -8.95704865e-01
-9.49192226e-01 7.62213767e-01 3.02418739e-01 -5.09895444e-01
1.09406757e+00 1.52723685e-01 8.60284209e-01 1.14701867e+00
3.33888590e-01 -1.36644733e+00 -7.05299154e-02 7.01893508e-01
5.77917516e-01 -6.53617740e-01 -1.61175169e-02 4.40066531e-02
-4.71182883e-01 1.08240497e+00 7.44980097e-01 -5.00772521e-02
9.77570415e-01 7.31042624e-01 -2.05899253e-01 -3.93950418e-02
-1.61349607e+00 1.61059693e-01 3.56250629e-02 3.75923693e-01
2.93604732e-01 -1.32926553e-01 -7.77480751e-02 9.00780439e-01
-5.38260758e-01 -3.94896753e-02 -1.49702290e-02 8.69890630e-01
-6.99258566e-01 -1.20599174e+00 1.26848534e-01 8.79388273e-01
-3.08434188e-01 -5.88460147e-01 -2.88520306e-01 5.13103426e-01
-2.12208375e-01 5.14040172e-01 2.07929283e-01 -2.23178223e-01
-7.26576895e-02 5.02909303e-01 4.57937658e-01 -8.27152967e-01
-9.83331442e-01 3.92634481e-01 2.83283025e-01 -5.13500810e-01
1.00348741e-01 -5.61274350e-01 -1.29421341e+00 -2.19864443e-01
-4.27369863e-01 2.76159123e-02 7.53415287e-01 1.18352568e+00
7.15857863e-01 2.57004261e-01 2.50456452e-01 -9.28855240e-01
-1.18262398e+00 -1.08261931e+00 -2.33379096e-01 -4.46468204e-01
5.83983436e-02 -5.43478251e-01 -2.99435407e-01 -8.99726711e-03] | [8.709086418151855, 3.6283881664276123] |
79a9ee22-be25-48cf-8725-afc4b6c292c1 | time-series-counterfactual-inference-with | null | null | https://openreview.net/forum?id=JVs1OrQgR3A | https://openreview.net/pdf?id=JVs1OrQgR3A | Time Series Counterfactual Inference with Hidden Confounders | We present augmented counterfactual ordinary differential equations (ACODEs), a new approach to counterfactual inference on time series data with a focus on healthcare applications. ACODEs model interventions in continuous time with differential equations, augmented by auxiliary confounding variables to reduce inference bias. Experiments on tumor growth simulation and sepsis patient treatment response show that ACODEs outperform other methods like counterfactual Gaussian processes, recurrent marginal structural networks, and time series deconfounders in the accuracy of counterfactual inference. The learned auxiliary variables also reveal new insights into causal interventions and hidden confounders. | ['Yan Liu', 'Samuel A Assefa', 'Jiahao Chen', 'Guangyu Li'] | 2021-01-01 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 4.99545127e-01 5.72788000e-01 -4.90798831e-01 2.44243219e-02
-4.54787940e-01 3.44900452e-02 6.94824040e-01 -2.43497223e-01
-3.32797438e-01 1.78715718e+00 9.53916788e-01 -9.95918989e-01
-4.48010087e-01 -8.13011289e-01 -6.76463902e-01 -7.10170984e-01
-7.38937616e-01 6.38448834e-01 -7.17985809e-01 1.76046982e-01
1.11963421e-01 1.65941060e-01 -6.52544677e-01 -1.35943815e-01
1.20615244e+00 1.39802963e-01 -3.66853744e-01 5.60147583e-01
1.01898015e-02 8.31940532e-01 -4.94330883e-01 -4.56218690e-01
-4.27756086e-02 -3.12532216e-01 -1.62983641e-01 -6.39959514e-01
-4.74664509e-01 -1.87716097e-01 -5.61292171e-01 7.43026078e-01
8.12888086e-01 6.63297921e-02 9.82312024e-01 -1.50140345e+00
-9.02013004e-01 1.20200562e+00 -5.43345094e-01 3.37089032e-01
7.75197372e-02 5.55093408e-01 1.27222329e-01 -2.52758235e-01
2.26486757e-01 1.71996760e+00 1.17968059e+00 8.71319711e-01
-1.84132457e+00 -8.11424077e-01 3.56204957e-01 -4.60022181e-01
-8.16415310e-01 -1.22464009e-01 3.78728628e-01 -4.09718841e-01
7.98477709e-01 3.94971490e-01 6.35632992e-01 1.96655738e+00
1.30988729e+00 5.39601922e-01 1.53655648e+00 5.05281277e-02
6.13516331e-01 -4.61173475e-01 1.54412970e-01 9.30494666e-02
8.41520607e-01 1.06393707e+00 -2.34121665e-01 -1.10666943e+00
1.10226822e+00 7.40174532e-01 -4.25335467e-01 -1.45959809e-01
-1.43559396e+00 1.00253737e+00 -1.95540830e-01 -1.46858901e-01
-1.31621873e+00 5.58914185e-01 3.08661759e-01 1.66248500e-01
7.66250074e-01 3.91979784e-01 -8.88575554e-01 3.48660536e-02
-3.09341669e-01 3.31815958e-01 1.03791201e+00 6.30563796e-01
-2.24680349e-01 4.43045110e-01 -1.00423479e+00 1.65588409e-02
-4.24366854e-02 1.26127827e+00 3.42804492e-01 -1.05566752e+00
1.64174467e-01 -1.69139400e-01 4.76062268e-01 -2.66568780e-01
-6.52655721e-01 -4.47694719e-01 -1.39647055e+00 -2.09260695e-02
4.48420376e-01 -9.83470559e-01 -1.03857112e+00 2.12387896e+00
7.26775974e-02 8.51299167e-01 2.05495209e-01 1.82527870e-01
4.07843515e-02 2.45685443e-01 7.19481349e-01 -1.03709209e+00
1.11075628e+00 5.02715483e-02 -1.33602035e+00 4.02662009e-01
5.13118207e-01 -2.65792787e-01 3.57336760e-01 4.90178801e-02
-1.11312151e+00 6.82013482e-03 -2.24832147e-01 7.57715106e-01
-1.76156729e-01 -8.25370073e-01 1.06500125e+00 7.58643925e-01
-8.12879622e-01 1.02814627e+00 -9.66417789e-01 2.10820600e-01
7.80614614e-01 2.49229193e-01 2.99911916e-01 1.45264015e-01
-1.66546130e+00 7.90970922e-01 -5.68255708e-02 -2.10085586e-01
-1.27279496e+00 -2.13964367e+00 -7.40525305e-01 2.84793153e-02
3.80603135e-01 -1.83278787e+00 1.36720622e+00 -5.03641307e-01
-1.63493681e+00 2.07153723e-01 -1.02492839e-01 -1.24382162e+00
9.00066316e-01 9.67519507e-02 -5.74679255e-01 -3.14375103e-01
2.35164285e-01 5.86769804e-02 6.72085702e-01 -7.42729783e-01
-3.67515683e-01 -5.32317877e-01 -4.37673599e-01 2.40597110e-02
6.12095714e-01 -3.07695419e-01 8.15176487e-01 -1.01330841e+00
-6.94849849e-01 -9.23201442e-01 -1.05674720e+00 -3.38048249e-01
-6.32049203e-01 -2.74431854e-01 3.56149346e-01 -8.22524488e-01
1.15064895e+00 -1.56903183e+00 -3.99786323e-01 -8.25522020e-02
3.19112331e-01 -4.51051533e-01 4.17115465e-02 2.26414084e-01
-7.44870245e-01 3.42000157e-01 -3.48989069e-01 3.52509543e-02
-6.26837164e-02 1.81630805e-01 -4.77965057e-01 5.17636240e-01
9.39777028e-03 1.36368072e+00 -1.14747417e+00 -2.03021944e-01
2.42851123e-01 4.27451730e-01 -2.38054290e-01 -5.25514372e-02
-1.93119809e-01 4.44214821e-01 -6.28161967e-01 4.01401632e-02
8.70003998e-01 -2.15016291e-01 1.79921746e-01 4.02001083e-01
-1.65379718e-01 2.40765840e-01 -7.61409938e-01 1.22416782e+00
-5.88423491e-01 7.44690821e-02 -4.68859375e-01 -9.76909518e-01
8.43442008e-02 1.00359714e+00 8.97869349e-01 -4.09841001e-01
1.13094509e-01 -2.11530298e-01 2.69357115e-01 -5.87766886e-01
-1.87329650e-01 -1.10141885e+00 -1.63542315e-01 8.27410877e-01
-3.16081285e-01 1.81604072e-01 -3.21329147e-01 2.31848016e-01
1.25451577e+00 -3.94967496e-02 6.47311926e-01 -5.31102061e-01
-3.35726768e-01 -6.10149130e-02 7.82473087e-01 1.13890541e+00
-1.41040310e-01 1.13700636e-01 8.38656723e-01 -3.80189925e-01
-9.46830392e-01 -1.64972353e+00 -2.60170817e-01 1.35871828e-01
-5.99078655e-01 5.78093290e-01 -4.10714447e-01 -7.49816895e-01
5.22612274e-01 1.52642632e+00 -1.35687065e+00 -3.94479632e-01
-4.56714958e-01 -1.61786628e+00 5.18164575e-01 6.42192543e-01
-5.86226024e-02 -8.30692172e-01 -5.31137645e-01 5.13175428e-01
-1.57438427e-01 -1.23438925e-01 -7.05298126e-01 -6.61713034e-02
-1.24772418e+00 -1.27148390e+00 -1.28410029e+00 4.43081856e-01
4.35243666e-01 -1.16499148e-01 1.27595937e+00 -8.52359891e-01
-2.42817178e-01 4.87053841e-01 6.18540823e-01 -1.24370980e+00
-5.13858020e-01 -8.40727448e-01 3.10166031e-01 -3.14220607e-01
4.29065287e-01 -8.72994423e-01 -1.16909862e+00 -2.27031171e-01
-7.26458848e-01 4.58543822e-02 5.19171536e-01 1.16864645e+00
4.66775477e-01 -3.42772812e-01 1.03686833e+00 -1.35863590e+00
9.21330810e-01 -9.74174440e-01 -4.43936020e-01 1.63988933e-01
-1.12122250e+00 3.56461525e-01 2.98163712e-01 -9.46811914e-01
-1.73236334e+00 -6.83435202e-01 4.54416037e-01 -4.33234900e-01
-1.24715857e-01 6.60925508e-01 5.23349270e-02 8.99116874e-01
5.51165044e-01 -1.07341986e-02 5.66616714e-01 -2.01685771e-01
5.03299654e-01 1.74852028e-01 3.54724407e-01 -2.82596350e-01
1.14537179e-01 1.06194043e+00 4.72221851e-01 -8.67206901e-02
-6.12741649e-01 1.21965349e-01 1.40650021e-02 3.17715406e-01
8.22365582e-01 -1.03408480e+00 -1.22127664e+00 2.34914631e-01
-1.05861759e+00 -6.81385577e-01 -9.91916478e-01 1.10151601e+00
-1.14472806e+00 -2.76160389e-01 -5.58472157e-01 -1.24989212e+00
-2.51581311e-01 -3.18705887e-01 6.61918581e-01 -9.98833915e-04
-5.73031008e-01 -1.80176997e+00 6.02079034e-01 -4.52750593e-01
4.10846829e-01 9.01918232e-01 9.44405377e-01 -3.84245843e-01
-3.35878640e-01 5.97128086e-02 1.25035718e-01 -1.99425712e-01
4.47903961e-01 -1.96562394e-01 -6.08207107e-01 3.74129713e-02
3.56560349e-01 7.41410077e-01 6.68522060e-01 1.70769548e+00
1.08031785e+00 -7.47346401e-01 -1.09823191e+00 4.07416224e-01
1.28242409e+00 6.02060139e-01 7.17048407e-01 -3.17233622e-01
1.69900611e-01 4.69412237e-01 2.68533081e-01 8.07294071e-01
2.88576990e-01 -2.87173986e-01 1.75612181e-01 -3.07966024e-01
3.19634736e-01 -6.63951337e-01 1.30253404e-01 -5.40054329e-02
-2.57704288e-01 -6.45424202e-02 -7.62962520e-01 8.53113770e-01
-1.75597477e+00 -1.54122770e+00 -7.27209330e-01 2.30795479e+00
9.10592318e-01 2.27005407e-02 2.40955070e-01 -4.11089152e-01
8.49049985e-01 -3.48748416e-01 -9.78037179e-01 -3.59781295e-01
-3.59297723e-01 4.94017035e-01 1.18942368e+00 2.96086729e-01
-5.97330928e-01 1.80787351e-02 7.57968760e+00 5.43004513e-01
-6.27620280e-01 4.89504486e-01 1.04570341e+00 -8.46289918e-02
-7.07202137e-01 7.55069330e-02 -1.94485113e-01 5.82140863e-01
1.51332259e+00 -1.01265121e+00 -7.15506822e-02 1.64929435e-01
1.10274661e+00 -4.11521085e-02 -1.33823013e+00 5.21496892e-01
-5.75210571e-01 -1.54908478e+00 -4.30203229e-02 2.87025064e-01
1.15846372e+00 -1.50640085e-01 5.63374400e-01 3.36033970e-01
1.51803637e+00 -1.06015062e+00 1.22203298e-01 1.32929873e+00
8.78814816e-01 -7.25556374e-01 8.45381856e-01 2.55198807e-01
-1.10882699e-01 4.98828851e-02 -7.50450715e-02 -4.08386946e-01
6.03953600e-01 8.19034457e-01 -9.03318465e-01 5.29068649e-01
5.91000989e-02 4.02549267e-01 5.68496168e-01 1.00569069e+00
-2.16610432e-01 8.42231631e-01 -1.82331383e-01 1.06375389e-01
-9.70756710e-02 9.96242985e-02 7.36690998e-01 9.28930938e-01
4.66972619e-01 3.55756938e-01 -4.72021282e-01 1.17388833e+00
2.59147942e-01 -3.95240843e-01 -9.93393481e-01 1.33351997e-01
2.07571164e-01 3.34742844e-01 -3.41073275e-01 -5.68987489e-01
-3.10872227e-01 5.69916010e-01 -5.66259921e-01 9.32658613e-01
-1.01966751e+00 1.52211413e-01 1.11618161e+00 6.66144937e-02
-9.77300256e-02 5.18164158e-01 -5.81448436e-01 -1.14120770e+00
-8.34487915e-01 -6.49613500e-01 9.80740130e-01 -6.41150594e-01
-1.74829924e+00 -2.01700211e-01 1.98276564e-01 -8.95883203e-01
-5.08483410e-01 -4.80186999e-01 -8.51554692e-01 1.42287552e+00
-1.14424169e+00 -4.07283098e-01 7.18596220e-01 6.38037264e-01
2.85161883e-01 5.59090018e-01 8.78227234e-01 -3.69822681e-01
-6.36548519e-01 1.28048837e-01 3.26998502e-01 -2.20217869e-01
4.17914122e-01 -1.49420917e+00 5.38068712e-01 4.89750892e-01
-6.34167194e-01 6.98256850e-01 1.17568970e+00 -1.26587605e+00
-1.03767705e+00 -1.20161557e+00 5.41724503e-01 -7.76975036e-01
7.93706894e-01 1.70091957e-01 -7.16374516e-01 7.37774253e-01
4.78160590e-01 -4.82992738e-01 8.89045656e-01 1.91644967e-01
1.03239030e-01 3.28107506e-01 -1.29358733e+00 8.94544184e-01
1.05251336e+00 -2.11239904e-01 -9.74858582e-01 4.71446216e-01
1.23350751e+00 -1.07769132e-01 -8.53460133e-01 6.36351109e-01
3.78720194e-01 -2.44324848e-01 9.90098357e-01 -1.56623185e+00
4.34734881e-01 4.47725177e-01 5.40393353e-01 -1.82985604e+00
-4.25111324e-01 -1.40536237e+00 -3.50126594e-01 5.56558371e-01
5.72140753e-01 -1.31612444e+00 4.65375423e-01 9.99020755e-01
2.78320700e-01 -4.92448181e-01 -9.48962927e-01 -8.28937352e-01
4.57732022e-01 -4.47208226e-01 1.08576572e+00 1.27180433e+00
1.54297411e-01 1.44790649e-01 -2.95411497e-01 1.88475177e-01
1.17252457e+00 3.82542573e-02 2.22657815e-01 -1.16113412e+00
-2.25993648e-01 -4.02608842e-01 3.88250202e-01 -2.05089003e-01
4.27124858e-01 -5.37486553e-01 -2.52288491e-01 -1.37199891e+00
6.36357844e-01 -2.62048185e-01 -5.59018493e-01 -1.15172550e-01
-8.20790410e-01 -5.40630698e-01 -3.39934319e-01 -4.46184427e-01
1.01318695e-01 7.75314271e-01 1.18967891e+00 2.59864721e-02
-2.81597227e-01 6.52971685e-01 -8.05797338e-01 6.17172658e-01
8.09038103e-01 -8.06038499e-01 -5.51778138e-01 2.66361266e-01
-7.55881239e-03 1.07713175e+00 6.97628140e-01 -4.45948094e-02
-2.98731327e-02 -8.21085513e-01 2.51792789e-01 -6.47569537e-01
-1.81820750e-01 -3.22980464e-01 9.76832092e-01 1.30056393e+00
-4.10325825e-01 2.26253510e-01 4.46907580e-01 1.15857387e+00
3.24227065e-01 4.75028813e-01 2.59792566e-01 -2.82482386e-01
2.28381053e-01 6.22335434e-01 -8.11877310e-01 2.23726690e-01
9.75257039e-01 1.08928449e-01 -5.21833122e-01 -7.55928397e-01
-1.28385293e+00 4.71405119e-01 -1.90124497e-01 -1.85882077e-02
6.43766463e-01 -1.12830019e+00 -1.11288702e+00 -4.27717417e-01
-4.55038577e-01 -3.71992558e-01 8.22527409e-01 1.26047003e+00
2.60620803e-01 8.18222523e-01 9.32327136e-02 -1.06096350e-01
-6.78157389e-01 1.43153346e+00 5.10647774e-01 -7.90853143e-01
-3.75024408e-01 3.30852002e-01 8.47153187e-01 -2.37200573e-01
-3.24908882e-01 -2.87381023e-01 1.54384270e-01 -2.21343964e-01
6.15140975e-01 8.09770823e-01 -6.99644923e-01 3.16435814e-01
8.05685222e-02 -4.29585308e-01 2.28048176e-01 -3.70026797e-01
1.42958736e+00 -4.69131351e-01 -8.79930556e-02 8.98414433e-01
6.26823664e-01 -5.61751485e-01 -1.40248156e+00 -1.73301503e-01
5.89629784e-02 -6.61553741e-02 -1.38114586e-01 -1.03214240e+00
-3.67459297e-01 6.21044934e-01 6.42287135e-01 3.82283539e-01
1.02211261e+00 -2.33653441e-01 1.21245883e-01 -3.10547858e-01
2.69409299e-01 -6.21314764e-01 -4.89188582e-01 -3.90796997e-02
8.78505409e-01 -7.53663719e-01 -7.67021328e-02 -1.77170336e-01
-3.03859830e-01 3.81996840e-01 -1.84099823e-02 -4.03383493e-01
1.34071791e+00 6.85794473e-01 -5.50249517e-02 -1.48739172e-02
-1.48998356e+00 4.27456722e-02 -1.50911868e-01 1.04681933e+00
4.42641884e-01 7.46930659e-01 -7.85724282e-01 8.36569607e-01
-6.82200119e-02 7.16324270e-01 6.78610802e-01 4.43985313e-01
7.47423530e-01 -4.90870208e-01 -5.83552897e-01 8.69671404e-01
-1.06383491e+00 -5.29451668e-01 3.09421271e-01 1.13256085e+00
-1.39474243e-01 5.76774657e-01 3.45810920e-01 6.64846539e-01
3.73166919e-01 2.10281134e-01 3.09341848e-01 -3.48030746e-01
-2.01231673e-01 4.58444431e-02 -8.90239477e-02 -5.51562607e-01
-4.05978888e-01 -1.07315552e+00 -9.68100131e-01 -6.72618449e-01
-2.41021544e-01 2.60518402e-01 1.78833261e-01 7.98583627e-01
6.18322730e-01 1.17071295e+00 5.85007787e-01 -2.36554086e-01
-9.99389648e-01 -1.17395627e+00 -5.31465292e-01 1.90539643e-01
6.66911542e-01 -4.79475677e-01 -6.34400487e-01 -1.05810829e-01] | [8.02066707611084, 5.388671398162842] |
004e500b-a9ae-4d22-8f85-9e88620806f7 | document-layout-analysis-with-aesthetic | 2111.13809 | null | https://arxiv.org/abs/2111.13809v1 | https://arxiv.org/pdf/2111.13809v1.pdf | Document Layout Analysis with Aesthetic-Guided Image Augmentation | Document layout analysis (DLA) plays an important role in information extraction and document understanding. At present, document layout analysis has reached a milestone achievement, however, document layout analysis of non-Manhattan is still a challenge. In this paper, we propose an image layer modeling method to tackle this challenge. To measure the proposed image layer modeling method, we propose a manually-labeled non-Manhattan layout fine-grained segmentation dataset named FPD. As far as we know, FPD is the first manually-labeled non-Manhattan layout fine-grained segmentation dataset. To effectively extract fine-grained features of documents, we propose an edge embedding network named L-E^3Net. Experimental results prove that our proposed image layer modeling method can better deal with the fine-grained segmented document of the non-Manhattan layout. | ['Cheng Jin', 'Liang Xue', 'Zhao Zhou', 'Xiangcheng Du', 'Xin Li', 'Xingjiao Wu', 'Tianlong Ma'] | 2021-11-27 | null | null | null | null | ['document-layout-analysis', 'image-augmentation'] | ['computer-vision', 'computer-vision'] | [ 3.88017222e-02 -3.86677012e-02 -2.95753062e-01 -1.85221389e-01
-1.59426093e-01 -6.90511405e-01 4.18804646e-01 2.08917797e-01
-7.13025033e-02 1.42533571e-01 5.88154122e-02 -6.20129049e-01
-4.86469626e-01 -9.98316228e-01 -5.50269544e-01 -4.21365499e-01
-3.78199331e-02 3.22972208e-01 1.47848293e-01 1.06399357e-01
6.09629571e-01 6.93161249e-01 -7.45133996e-01 3.86169553e-01
9.88912404e-01 8.22896719e-01 2.66701341e-01 9.48519409e-01
-6.87011480e-01 5.46342552e-01 -5.86918056e-01 -2.30814010e-01
-1.12171862e-02 -2.26259708e-01 -9.45078909e-01 6.69526756e-01
5.41738749e-01 -2.51926839e-01 -3.87722582e-01 1.12784338e+00
1.92842320e-01 -2.98557073e-01 8.20190609e-01 -1.31201577e+00
-6.92646384e-01 7.77194142e-01 -1.02484381e+00 2.90523712e-02
4.03655507e-02 -1.93966493e-01 9.85712826e-01 -6.92190528e-01
8.39528918e-01 1.43696940e+00 6.31934762e-01 -5.93993254e-03
-7.39442468e-01 -3.25068593e-01 4.16093200e-01 1.86063930e-01
-1.48054671e+00 3.95272464e-01 1.10517669e+00 -5.26376247e-01
3.72248739e-01 2.41123304e-01 5.97494960e-01 4.68505651e-01
2.97897726e-01 1.32645476e+00 1.13482702e+00 -5.37696481e-01
6.08722903e-02 4.76148538e-02 5.63479602e-01 1.19424641e+00
4.41224158e-01 -7.64794648e-01 -3.48246098e-02 5.13960600e-01
9.91382182e-01 -6.86517879e-02 -2.61943370e-01 -4.75377470e-01
-9.56885397e-01 4.80455875e-01 6.18492305e-01 6.91176534e-01
-5.33096492e-02 9.20908749e-02 3.29100490e-01 -5.04832864e-02
2.81333804e-01 4.62664425e-01 -2.96110749e-01 -1.29973635e-01
-1.19269383e+00 1.00268751e-01 6.39388978e-01 1.17084920e+00
6.50134265e-01 -4.33741659e-02 -6.97187185e-02 6.40015483e-01
2.15071499e-01 3.01638097e-01 1.96617573e-01 -3.78337920e-01
6.42477989e-01 1.26860034e+00 -1.84462607e-01 -1.36989677e+00
-5.62085211e-01 -3.08976829e-01 -1.08566689e+00 -1.30210416e-02
5.14277697e-01 7.25473417e-03 -9.66814578e-01 8.06048095e-01
-3.03656776e-02 -2.81954169e-01 -2.62559503e-01 7.98807979e-01
5.83271980e-01 8.57546985e-01 -2.14816496e-01 1.89573199e-01
1.58168960e+00 -1.23362541e+00 -7.44747996e-01 -4.53365892e-02
5.40204406e-01 -9.56002831e-01 1.23573792e+00 2.74886191e-01
-7.70918250e-01 -6.82856798e-01 -1.32047629e+00 -1.92976832e-01
-6.90719664e-01 3.25522661e-01 6.91780448e-01 7.57385612e-01
-7.16764569e-01 3.01725119e-01 -4.54976350e-01 -5.02613008e-01
6.98363543e-01 2.20672175e-01 -4.48102176e-01 -5.95100522e-01
-6.15021706e-01 4.35176909e-01 8.02641451e-01 3.09364200e-01
-3.43355626e-01 -4.60355371e-01 -7.94439793e-01 2.28943080e-01
6.66259110e-01 -3.09147656e-01 7.89399564e-01 -4.55257386e-01
-9.47027981e-01 6.13042712e-01 2.27721874e-02 -1.58740819e-01
4.28244025e-01 1.69793572e-02 -4.05925632e-01 2.04065800e-01
-1.05203576e-01 5.06584585e-01 6.52960181e-01 -1.56126988e+00
-6.07792079e-01 -3.62284601e-01 2.17490837e-01 1.93188172e-02
-4.78672862e-01 -3.75073314e-01 -9.33151841e-01 -5.80603182e-01
2.85647571e-01 -6.55659080e-01 -1.59310430e-01 -1.93786383e-01
-9.71749783e-01 -9.49683785e-02 1.38594115e+00 -7.14912057e-01
1.62186372e+00 -2.00754023e+00 -2.87049055e-01 6.51679873e-01
3.24409842e-01 4.27809745e-01 -3.19188029e-01 4.97808099e-01
1.84157044e-01 5.80126762e-01 -2.80311018e-01 -5.31962095e-03
2.15595528e-01 -6.04616106e-02 3.63170058e-02 1.01659745e-01
2.62051940e-01 1.14802861e+00 -7.17518210e-01 -9.24504399e-01
3.45495880e-01 2.52134562e-01 -2.72185802e-01 -1.30426949e-02
-3.30282152e-01 -8.71300027e-02 -7.04050481e-01 9.13754582e-01
1.12684286e+00 -1.95378989e-01 1.91420510e-01 -6.38108730e-01
-2.45617911e-01 -6.07868552e-01 -1.20449984e+00 1.68826377e+00
-3.98512602e-01 9.71123099e-01 -2.26847813e-01 -6.70023441e-01
9.47401822e-01 -2.68217713e-01 3.52377832e-01 -7.79900730e-01
4.33654368e-01 1.06020987e-01 1.38891727e-01 -5.74746490e-01
8.66310358e-01 4.84042406e-01 -2.64018297e-01 2.63058960e-01
-2.32029647e-01 -9.54406783e-02 4.23137009e-01 2.20680729e-01
1.21110070e+00 -1.29108699e-02 1.16165057e-01 -2.85565853e-01
5.52255571e-01 1.10257417e-01 2.49453217e-01 6.86173022e-01
-1.65123984e-01 6.42076313e-01 8.48528802e-01 -3.46591085e-01
-1.36145365e+00 -8.04554701e-01 -9.42310914e-02 5.85693598e-01
2.84449965e-01 -8.33008170e-01 -1.49029374e+00 -8.98417830e-01
5.38367219e-02 2.78479159e-01 -4.37367618e-01 1.54394150e-01
-6.63993716e-01 -5.91661274e-01 6.33260190e-01 7.14815080e-01
1.14960229e+00 -1.02865875e+00 -3.06356281e-01 -8.12852941e-03
1.04348145e-01 -1.19808280e+00 -8.39959621e-01 2.07176292e-03
-4.79270369e-01 -1.25284064e+00 -8.22745025e-01 -1.20315361e+00
1.09467864e+00 3.94064248e-01 8.62427592e-01 1.03723086e-01
-6.65022194e-01 4.29810256e-01 -3.58282238e-01 -2.86761701e-01
1.62363365e-01 3.52389306e-01 -6.70587480e-01 1.63973168e-01
4.35249448e-01 1.09964006e-01 -6.29517496e-01 1.24780409e-01
-1.24388492e+00 1.11232072e-01 8.94259334e-01 5.08329451e-01
6.86090171e-01 1.02222919e+00 -4.29323278e-02 -1.31433845e+00
1.10286570e+00 -6.22325726e-02 -6.42775834e-01 4.79988575e-01
-7.23782480e-01 8.76368210e-02 7.53426075e-01 -1.14109457e-01
-9.84848619e-01 -1.21305339e-01 1.91687629e-01 -1.03395320e-01
-3.44078034e-01 8.12457860e-01 -7.50000358e-01 -1.15783922e-01
-9.93462130e-02 1.17767826e-01 -6.43346190e-01 -5.59028327e-01
6.54555976e-01 1.00538409e+00 4.92152035e-01 -4.10789371e-01
1.09573853e+00 4.11731094e-01 2.70467579e-01 -1.30334640e+00
-6.15633488e-01 -5.45798838e-01 -1.03674436e+00 -2.31952980e-01
1.11232603e+00 -1.87607810e-01 -8.70884836e-01 4.93273854e-01
-1.23543429e+00 -2.78645247e-01 2.49369293e-02 -5.28774736e-03
-2.10271060e-01 5.58625519e-01 -4.51860517e-01 -7.86576569e-01
-1.21104725e-01 -9.55948651e-01 1.17697310e+00 3.31022769e-01
-8.03079680e-02 -1.26673472e+00 -1.89469963e-01 4.04801875e-01
-7.77930394e-02 4.13913101e-01 1.42613077e+00 -5.20604011e-03
-1.09666276e+00 -3.08237731e-01 -9.46060598e-01 3.13821673e-01
1.35962561e-01 2.67103583e-01 -6.94499195e-01 -9.16061178e-02
-3.90751034e-01 2.65706807e-01 9.14471567e-01 4.22893584e-01
1.55037689e+00 -1.60382643e-01 -4.22815174e-01 6.45367742e-01
1.61329043e+00 4.90054250e-01 7.41797447e-01 6.48707211e-01
1.28864586e+00 4.89645332e-01 7.54322231e-01 4.09692168e-01
5.32321393e-01 1.95097044e-01 2.43463144e-01 -3.71355951e-01
-2.79474020e-01 -5.28278530e-01 -1.64563924e-01 9.91375864e-01
7.84760639e-02 -7.58038580e-01 -9.76296544e-01 5.46340942e-01
-1.69221163e+00 -7.37819731e-01 -4.50740635e-01 1.52716887e+00
2.78435946e-01 5.15694916e-01 -1.05802357e-01 5.02912998e-01
7.50230312e-01 4.44274217e-01 -4.08996075e-01 -7.97857702e-01
-3.06530327e-01 4.87290286e-02 7.29745090e-01 3.33221644e-01
-1.23099446e+00 1.23346496e+00 4.85768557e+00 1.01915669e+00
-8.18622947e-01 -3.61320317e-01 4.16485578e-01 8.06251884e-01
-4.76523072e-01 1.66045412e-01 -8.13321233e-01 5.73819995e-01
4.24870402e-01 1.33348942e-01 5.12096360e-02 7.44075656e-01
1.15048751e-01 -1.00100808e-01 -8.68571639e-01 1.06766331e+00
2.17474312e-01 -1.51503813e+00 2.17738748e-01 2.12320641e-01
9.11791742e-01 -9.40756083e-01 1.28404021e-01 6.78763390e-02
-1.80402443e-01 -1.02965355e+00 4.09233093e-01 4.82138515e-01
8.30725908e-01 -1.11983860e+00 8.03353786e-01 2.07240447e-01
-1.62620378e+00 3.02725900e-02 -3.18912387e-01 3.79405797e-01
-4.54803072e-02 7.24574924e-01 -1.03647673e+00 6.57582283e-01
3.73092443e-01 6.10891521e-01 -9.11314189e-01 1.25481534e+00
-1.25178844e-01 4.94251698e-01 2.56233126e-01 -2.13264823e-01
6.90366805e-01 -4.11196589e-01 1.61122009e-01 1.48772168e+00
1.99137345e-01 -2.98622996e-01 2.80201703e-01 8.01683903e-01
-3.65881771e-01 1.51160389e-01 -5.90164721e-01 -6.04950011e-01
3.05209935e-01 1.27528191e+00 -1.41198170e+00 -2.89370194e-02
-2.71705836e-01 1.30624187e+00 -4.11409624e-02 3.21716219e-01
-5.98614514e-01 -1.13117456e+00 1.45296440e-01 1.58830732e-01
2.18645036e-01 -6.76738560e-01 -5.06498218e-01 -8.29592168e-01
2.23021246e-02 -6.64983451e-01 4.78580743e-02 -8.36919188e-01
-8.77744853e-01 3.80151808e-01 -1.85170472e-01 -1.02837253e+00
3.51527244e-01 -9.86079395e-01 -6.33110881e-01 6.44051433e-01
-1.49949706e+00 -1.59968030e+00 -6.15429819e-01 1.99189514e-01
7.14638531e-01 5.51798241e-03 6.47782147e-01 3.39355737e-01
-8.07095230e-01 5.44525564e-01 3.19310099e-01 5.08647740e-01
3.94062161e-01 -1.62960851e+00 6.18021011e-01 8.59092355e-01
1.70769915e-01 6.84845626e-01 2.58915842e-01 -7.05367267e-01
-1.48454142e+00 -1.29051161e+00 6.26955867e-01 -1.74118459e-01
4.77323741e-01 -7.08830118e-01 -8.55255127e-01 5.24952173e-01
3.25915515e-01 -4.33752447e-01 4.35236752e-01 4.39666696e-02
-2.15058357e-01 -1.62115067e-01 -1.01912582e+00 8.78611326e-01
9.46709514e-01 -6.23782814e-01 -2.94036567e-01 1.75923124e-01
5.34108400e-01 -2.47340873e-01 -7.47906744e-01 1.84302732e-01
6.34692371e-01 -6.16879880e-01 8.85246754e-01 -2.18873605e-01
4.82166946e-01 -5.04547954e-01 4.60328460e-02 -1.18572628e+00
-3.56740326e-01 -4.34631616e-01 -3.68309431e-02 1.52904844e+00
-4.38978756e-03 -8.36173538e-03 1.23402596e+00 4.87591736e-02
-1.92286372e-01 -7.88335025e-01 -1.53690472e-01 -8.91077816e-01
2.29733530e-02 -3.84648591e-01 8.67507279e-01 7.24327922e-01
-2.42698371e-01 2.51109451e-01 -9.85717252e-02 2.33734637e-01
7.00712979e-01 3.38750392e-01 7.43484914e-01 -1.26577830e+00
1.81900963e-01 -6.17756963e-01 -5.97148716e-01 -1.09890401e+00
1.49613112e-01 -8.66082907e-01 -2.27788493e-01 -2.12032652e+00
1.83829620e-01 -3.48545849e-01 -2.44179264e-01 6.20862581e-02
1.23507433e-01 1.88605219e-01 1.12592772e-01 -1.41413122e-01
-6.28194690e-01 1.28963754e-01 1.76505542e+00 -6.78653240e-01
-1.04487374e-01 -3.88701290e-01 -6.37890279e-01 8.66261303e-01
9.29318249e-01 -4.45550382e-02 -6.34426892e-01 -4.32065398e-01
1.00167662e-01 -3.00415367e-01 3.51984799e-02 -9.15682316e-01
2.50413001e-01 -1.58110365e-01 6.33108497e-01 -1.22590768e+00
-5.47441021e-02 -9.56698120e-01 -3.94996464e-01 1.66679189e-01
-1.83134317e-01 3.31373722e-03 2.82542586e-01 3.54435980e-01
-2.80363470e-01 -4.97313857e-01 5.02815604e-01 1.58444941e-02
-1.13980067e+00 3.43286991e-01 -4.15669322e-01 -3.68051529e-02
9.80139852e-01 -5.42922020e-01 -5.88915408e-01 1.21628456e-01
-2.10024193e-01 1.94000751e-01 4.25156921e-01 4.71040875e-01
8.48250389e-01 -1.23029602e+00 -1.96663588e-01 2.49184415e-01
3.30552161e-01 4.32064235e-02 4.76610243e-01 4.08735722e-01
-1.12355888e+00 9.48505521e-01 -3.50544304e-01 -2.93796808e-01
-1.30466068e+00 5.10828018e-01 -2.80564725e-01 -3.97906214e-01
-4.96672511e-01 6.31300628e-01 1.52139500e-01 -3.54337454e-01
1.74213842e-01 -7.46632695e-01 -3.28132451e-01 8.36750418e-02
3.46344471e-01 5.65213382e-01 -1.09644808e-01 -4.71276015e-01
-1.85342655e-01 1.03526723e+00 -3.60412866e-01 1.86599717e-01
9.51014698e-01 -1.36848882e-01 -2.76205719e-01 3.46889049e-01
1.36571443e+00 1.89264506e-01 -1.06491292e+00 2.20539674e-01
3.08857620e-01 -6.19183242e-01 2.36418918e-01 -7.34286785e-01
-7.50119507e-01 1.23444557e+00 7.05169678e-01 4.03117210e-01
1.04538751e+00 -4.01358068e-01 1.02011299e+00 5.30078411e-01
1.45225197e-01 -1.42051649e+00 2.96240985e-01 4.99380201e-01
6.48629546e-01 -1.06792307e+00 -3.17096747e-02 -5.70033431e-01
-3.20795327e-01 1.44067037e+00 7.33944178e-01 -7.89597705e-02
6.02342427e-01 2.83589065e-01 -9.88382921e-02 -2.93681890e-01
1.18489295e-01 -3.25837106e-01 5.36257863e-01 5.15575647e-01
4.08962578e-01 1.84721574e-01 -1.23545021e-01 5.63919127e-01
-1.49057776e-01 -6.49449900e-02 5.16719699e-01 1.06745887e+00
-4.96652097e-01 -1.35542583e+00 -4.46256548e-01 4.52783346e-01
-6.92417547e-02 2.77688988e-02 -6.36736393e-01 1.03572214e+00
4.03908402e-01 6.85821056e-01 2.10956827e-01 -4.12811875e-01
5.08060277e-01 -1.40879363e-01 6.39427841e-01 -3.24863493e-01
-1.73624426e-01 1.40428022e-01 -2.14966223e-01 -1.80849239e-01
2.01227609e-02 -1.99141651e-01 -1.46694720e+00 -4.25702095e-01
-1.51476279e-01 -3.61211114e-02 8.98643613e-01 6.32593632e-01
1.58850670e-01 1.04595864e+00 5.79602778e-01 -4.88653779e-01
1.41943857e-01 -5.98472536e-01 -1.02848125e+00 3.30520451e-01
1.47340804e-01 -3.64892960e-01 1.60528690e-01 9.89893079e-02] | [11.673186302185059, 2.5230350494384766] |
584fa3e0-33c8-4d89-84a6-8e1fa84958b8 | guir-at-semeval-2017-task-12-a-framework-for | null | null | https://aclanthology.org/S17-2180 | https://aclanthology.org/S17-2180.pdf | GUIR at SemEval-2017 Task 12: A Framework for Cross-Domain Clinical Temporal Information Extraction | Clinical TempEval 2017 (SemEval 2017 Task 12) addresses the task of cross-domain temporal extraction from clinical text. We present a system for this task that uses supervised learning for the extraction of temporal expression and event spans with corresponding attributes and narrative container relations. Approaches include conditional random fields and decision tree ensembles, using lexical, syntactic, semantic, distributional, and rule-based features. Our system received best or second best scores in TIMEX3 span, EVENT span, and CONTAINS relation extraction. | ['Nazli Goharian', 'Sean MacAvaney', 'Arman Cohan'] | 2017-08-01 | null | null | null | semeval-2017-8 | ['temporal-information-extraction'] | ['natural-language-processing'] | [ 2.41173118e-01 4.45294142e-01 -8.65501046e-01 -4.06649917e-01
-9.69822168e-01 -5.08910179e-01 7.65588105e-01 1.13459611e+00
-6.93466187e-01 1.24979103e+00 7.16368318e-01 -4.61179107e-01
-6.74756825e-01 -3.88042808e-01 -9.06376690e-02 -2.61272728e-01
-6.74398005e-01 5.44078887e-01 1.88015014e-01 -5.91759719e-02
2.81243734e-02 1.84154958e-01 -6.82753205e-01 9.88170445e-01
3.72458637e-01 1.23515391e+00 -5.02489686e-01 5.96409440e-01
-6.35205209e-03 1.67976725e+00 -3.27847242e-01 -4.76263165e-01
-5.11922359e-01 -3.70595276e-01 -1.16354930e+00 -5.88906348e-01
-5.32773793e-01 2.68424809e-01 -8.77687782e-02 2.86575615e-01
5.33862710e-01 -9.49099064e-02 9.45513248e-01 -9.56086516e-01
-6.16771728e-02 1.05288982e+00 -9.31046382e-02 6.40580714e-01
7.76373863e-01 -5.96119687e-02 8.38784158e-01 -4.91171002e-01
1.51693416e+00 6.82006121e-01 9.58536267e-01 4.58450973e-01
-1.27571726e+00 -7.15499043e-01 -2.50369072e-01 1.78557038e-01
-1.19989431e+00 -2.72585869e-01 2.09851656e-02 -8.00693989e-01
1.92744434e+00 7.94776678e-02 5.61813891e-01 1.48795331e+00
7.54037678e-01 4.99135882e-01 1.33061409e+00 -5.20825505e-01
3.20913583e-01 -1.89171270e-01 3.25311154e-01 4.41695362e-01
-3.32610130e-01 4.13112968e-01 -8.48060668e-01 -7.85022616e-01
5.72808739e-03 -2.00212449e-01 -1.33421227e-01 7.30229795e-01
-1.28962147e+00 7.13132620e-01 -1.23878336e-02 3.15897226e-01
-6.93781495e-01 -2.79847682e-01 1.02856207e+00 -6.56892508e-02
5.99966705e-01 5.65659523e-01 -1.22574306e+00 -2.65682280e-01
-1.09085822e+00 3.29936683e-01 1.02177870e+00 8.60641479e-01
-2.68351436e-01 -7.67279327e-01 -4.77921307e-01 5.37955284e-01
-6.39302889e-03 -2.04964988e-02 6.96414411e-01 -5.32107592e-01
2.23937660e-01 4.81231570e-01 -1.95743486e-01 -3.03868383e-01
-1.00281537e+00 2.24884465e-01 -4.09433752e-01 -4.26865280e-01
5.07625658e-03 -3.18364143e-01 -9.76597786e-01 1.48790181e+00
2.02441245e-01 3.25978994e-02 4.75668222e-01 1.07944652e-01
1.25653458e+00 2.00261816e-01 1.15636313e+00 -9.61560726e-01
1.68333864e+00 -2.43784025e-01 -1.54239190e+00 9.85972509e-02
1.06714904e+00 -7.32977986e-01 -7.61522800e-02 3.25193971e-01
-9.43981946e-01 4.26786780e-01 -8.10885906e-01 3.14368755e-02
-8.55613828e-01 -1.75585568e-01 7.60841668e-01 -1.36674643e-01
-4.22245651e-01 7.84372628e-01 -9.98210549e-01 -4.55814332e-01
4.19485509e-01 1.81550816e-01 -7.55232155e-01 3.06419462e-01
-1.75023615e+00 1.47460258e+00 9.26470339e-01 -4.57935691e-01
-3.99635881e-01 -1.18013632e+00 -1.05343318e+00 -4.77167606e-01
3.34797353e-01 -6.62287176e-01 1.24144340e+00 9.36525464e-02
-9.84001219e-01 1.11470377e+00 -7.55603015e-02 -8.65022302e-01
2.24190041e-01 -1.44279584e-01 -1.07858598e+00 1.14593744e-01
2.50169039e-01 2.87362695e-01 6.96841031e-02 9.22159553e-02
-5.02735913e-01 -4.78259027e-01 -5.88276327e-01 -2.28155836e-01
3.24693434e-02 6.74951732e-01 9.93239209e-02 -7.63601303e-01
-3.15423250e-01 -6.31175220e-01 -6.53839707e-01 -4.43849146e-01
-5.75744569e-01 -7.85603642e-01 2.05464378e-01 -9.35267627e-01
1.81774366e+00 -2.18455458e+00 -2.14011461e-01 -1.52782095e-03
2.38143921e-01 -3.23881567e-01 5.54870129e-01 7.36391544e-01
-7.51418948e-01 5.43350354e-02 -8.63871872e-02 2.66298980e-01
-3.58451843e-01 1.59239337e-01 -2.43974611e-01 2.34545767e-01
3.68599445e-01 9.86242652e-01 -9.46858287e-01 -1.19982684e+00
-1.17440984e-01 2.16514617e-01 -3.28054249e-01 -9.51558352e-02
-4.67837960e-01 1.76208794e-01 -4.72936302e-01 5.55265844e-01
-2.33131617e-01 -2.14677021e-01 6.28586113e-01 -2.29371890e-01
-1.19517483e-01 1.02336526e+00 -4.35208291e-01 1.69144702e+00
-1.87495887e-01 2.51826465e-01 -6.29985988e-01 -3.68425012e-01
6.43551290e-01 9.57387924e-01 1.03476882e+00 -4.94652987e-01
1.92556694e-01 1.78844452e-01 -3.37939203e-01 -9.18174505e-01
-6.65853098e-02 -5.73390841e-01 -6.07277989e-01 1.35414153e-01
3.57324064e-01 2.37645671e-01 1.74091354e-01 3.00805211e-01
1.53430223e+00 4.73960727e-01 1.39701498e+00 -1.76230788e-01
2.24942654e-01 2.68273681e-01 8.40936661e-01 -5.52970432e-02
-1.43456101e-01 1.46338567e-01 9.85712886e-01 -5.11192620e-01
-4.30971712e-01 -8.00873101e-01 -8.12608898e-01 5.95949233e-01
-7.83061624e-01 -1.17316735e+00 -8.29573348e-02 -9.82709587e-01
-3.15319262e-02 9.81194735e-01 -1.05399382e+00 -1.39179418e-03
-5.11253357e-01 -9.48253214e-01 9.10564065e-01 6.68328524e-01
-5.18709958e-01 -1.26071250e+00 -7.65002251e-01 6.85899258e-01
-3.31659526e-01 -1.60551703e+00 -1.89633340e-01 7.52889156e-01
-6.79394603e-01 -1.39464617e+00 -1.81772262e-01 -3.94597024e-01
-6.37914389e-02 -1.42049134e+00 1.12125015e+00 -8.67852867e-01
-6.22761548e-01 -9.15924758e-02 -3.03289592e-01 -8.18497419e-01
-3.78662854e-01 -1.52310416e-01 -1.77782565e-01 -6.88461304e-01
8.68937433e-01 -4.67606783e-01 -2.93484092e-01 -2.02337280e-01
-7.32830286e-01 -3.12577188e-02 3.08109254e-01 7.51233876e-01
9.00311589e-01 -4.57113981e-01 5.94211698e-01 -1.10869944e+00
7.02958643e-01 -9.25454378e-01 -1.17851518e-01 2.45371953e-01
-9.64583039e-01 1.97958127e-01 9.75956693e-02 -4.72532392e-01
-7.95346439e-01 1.95726380e-01 -1.32752046e-01 9.29968208e-02
-3.23518038e-01 1.08539534e+00 2.49519646e-01 8.85107815e-01
1.11834645e+00 -3.26160528e-02 -2.19715521e-01 -2.74329513e-01
3.74977738e-01 4.69673902e-01 4.80078608e-01 -3.00046057e-01
-2.42233425e-01 3.36688221e-01 2.18943551e-01 -1.29834086e-01
-9.28927422e-01 -4.14151669e-01 -6.01081848e-01 2.54881501e-01
1.39965630e+00 -8.86725187e-01 -6.75778687e-01 -1.22643426e-01
-1.08948100e+00 -2.66776383e-01 -5.70513725e-01 9.08477426e-01
-6.05647027e-01 -2.76995212e-01 -8.47824812e-01 -4.31784153e-01
-5.34945905e-01 -5.44809461e-01 8.52887213e-01 -2.10820064e-01
-1.26292419e+00 -1.13766885e+00 5.61162233e-01 -2.70244539e-01
-1.35713890e-01 1.07689023e+00 1.19381034e+00 -1.25744259e+00
3.16849411e-01 -2.97806412e-01 6.61513954e-02 -6.51292026e-01
1.42856002e-01 3.25117111e-02 -4.74035293e-01 5.45194149e-01
-4.33769017e-01 -3.01657677e-01 8.38248193e-01 6.37738883e-01
8.15800965e-01 -6.02927446e-01 -1.02371359e+00 3.73392165e-01
1.27037776e+00 4.94240552e-01 4.07730043e-01 2.26140052e-01
2.43263692e-02 7.79487848e-01 6.66512489e-01 6.88763082e-01
3.07978809e-01 5.55923998e-01 -2.81426281e-01 1.71522751e-01
-3.56402807e-02 -1.03169486e-01 2.26281378e-02 1.81021363e-01
-1.84464201e-01 1.34750336e-01 -1.32404256e+00 9.37956750e-01
-1.67734301e+00 -7.92301357e-01 -1.85665518e-01 1.46043861e+00
1.67723894e+00 4.49568897e-01 9.45403054e-02 4.46641669e-02
7.25195482e-02 -2.34649360e-01 -1.55550674e-01 -6.39698386e-01
-4.50110510e-02 6.79006934e-01 5.50523162e-01 8.47172439e-02
-1.26174879e+00 7.16253161e-01 6.85197258e+00 8.56849611e-01
-6.82303250e-01 4.40770090e-01 4.81812179e-01 -3.28833342e-01
9.74998623e-02 9.57711134e-03 -9.58795369e-01 3.84547412e-01
1.62896132e+00 -2.24168062e-01 -3.00268918e-01 4.98085380e-01
2.69488513e-01 -2.63351470e-01 -1.44860268e+00 5.37736952e-01
-1.92741781e-01 -1.73292100e+00 -5.30192375e-01 3.68261486e-02
4.47147518e-01 -2.30187718e-02 -3.35846096e-01 1.69765025e-01
6.22512519e-01 -1.11835277e+00 4.54702497e-01 7.83123732e-01
1.27025926e+00 -3.83837104e-01 8.34984481e-01 -1.74341425e-02
-9.33620095e-01 1.17908016e-01 7.05132961e-01 3.82129520e-01
4.30819452e-01 7.79450357e-01 -1.29945111e+00 8.28495860e-01
7.44217038e-01 1.15276730e+00 -2.59704798e-01 6.98627412e-01
-3.99519712e-01 7.53244638e-01 -2.41543680e-01 1.29976854e-01
2.91775484e-02 6.92905843e-01 5.28080106e-01 1.82057703e+00
-8.73999521e-02 7.30256736e-01 2.23616436e-02 3.79258722e-01
6.81309849e-02 4.53112692e-01 -6.42713606e-01 -2.27774426e-01
1.76421762e-01 9.40042734e-01 -7.38469958e-01 -4.71290946e-01
-1.83153778e-01 3.31308246e-01 4.73706238e-03 -6.54294491e-02
-8.21158111e-01 -3.18891138e-01 1.96624368e-01 2.34779730e-01
1.21319667e-01 3.26048017e-01 -5.71190715e-01 -7.06031263e-01
-4.61184800e-01 -6.06779754e-01 1.59152031e+00 -5.38365662e-01
-1.34304011e+00 1.05132651e+00 4.61262017e-01 -1.17268586e+00
-8.01928639e-01 -4.53807861e-01 -6.27852082e-02 6.46000028e-01
-8.99907172e-01 -1.17607856e+00 5.00716209e-01 5.97030401e-01
3.59674692e-01 -5.54208830e-02 1.39944971e+00 2.55529761e-01
-2.51128435e-01 4.97787774e-01 -5.79478562e-01 2.53581405e-01
1.22047687e+00 -1.24779594e+00 -2.93278471e-02 2.24763472e-02
-3.64305079e-01 3.36390883e-01 5.39000034e-01 -1.14313316e+00
-4.58887935e-01 -1.08356833e+00 2.07939577e+00 -7.55145490e-01
1.04118598e+00 7.39386305e-02 -4.61309344e-01 9.06663597e-01
2.52944887e-01 -1.69586297e-02 1.31147075e+00 4.14554328e-01
-6.07324302e-01 3.50191593e-01 -1.19885123e+00 1.79932863e-01
7.45072603e-01 -6.62637651e-01 -9.61366236e-01 6.49329305e-01
7.07938433e-01 -5.51349521e-01 -1.69179702e+00 8.05470645e-01
5.47509789e-01 7.77879963e-04 8.59078169e-01 -1.40838850e+00
8.49277437e-01 -3.57418275e-03 8.43843911e-03 -7.77218282e-01
2.89016799e-03 -7.77853727e-01 -3.37908834e-01 8.91814172e-01
1.29500854e+00 -3.22252750e-01 6.84009418e-02 7.02278256e-01
9.21427980e-02 -1.05779183e+00 -1.00733447e+00 -3.66824329e-01
8.97535086e-02 -6.45146132e-01 2.08051443e-01 1.54620242e+00
1.08643329e+00 5.40708959e-01 9.86218303e-02 -1.61200970e-01
6.41378313e-02 6.62569180e-02 -4.68672186e-01 -1.02704632e+00
-2.04874665e-01 -2.51777261e-01 -3.25661987e-01 3.67417276e-01
1.45744428e-01 -1.04478621e+00 -2.32354686e-01 -1.71012092e+00
3.80673468e-01 -1.57108858e-01 -5.09577572e-01 1.13986051e+00
-5.40608214e-03 -1.37449056e-01 -5.11919081e-01 -3.22876535e-02
-6.48787260e-01 -5.80320731e-02 5.37396431e-01 -7.49797523e-02
-3.93712729e-01 -1.74763098e-01 -3.06918979e-01 6.17172122e-01
5.28396904e-01 -1.04763007e+00 8.22072551e-02 3.45080286e-01
5.69006920e-01 6.47932231e-01 -3.49286646e-02 -4.25588340e-01
3.45319718e-01 -2.80512333e-01 3.48674655e-01 -8.46638083e-01
1.11682783e-03 -4.35728937e-01 4.56713766e-01 6.84006929e-01
-9.09662902e-01 3.97933759e-02 6.02433085e-01 4.67768580e-01
-3.63286585e-01 1.42850786e-01 5.19598365e-01 -1.89592138e-01
-4.18653965e-01 4.74330932e-02 -7.98529625e-01 2.10844979e-01
1.25953162e+00 3.22053701e-01 -1.85142279e-01 1.84486642e-01
-1.74543822e+00 2.31963485e-01 -4.67973948e-01 3.40832025e-01
4.98663187e-01 -1.09536135e+00 -1.01013136e+00 -4.43273634e-01
4.58499879e-01 -3.16204578e-01 1.61610469e-01 1.38063776e+00
-7.38780349e-02 6.98912263e-01 3.36962529e-02 -1.80686176e-01
-1.39816082e+00 9.67009246e-01 1.89993143e-01 -1.24678028e+00
-6.08502805e-01 7.07547963e-01 -2.15669498e-01 2.77846828e-02
7.71526620e-02 -2.60943085e-01 -6.32738888e-01 5.75094402e-01
5.32592416e-01 6.41562510e-03 2.75430858e-01 -2.02988997e-01
-9.43702340e-01 2.98008882e-02 -2.86908060e-01 -4.31053519e-01
1.45153999e+00 5.89541137e-01 -2.38180622e-01 4.23911393e-01
1.11196578e+00 -2.32502535e-01 -2.64880478e-01 -4.02812243e-01
8.16474617e-01 6.19450390e-01 -4.62905467e-02 -1.46239603e+00
-2.03128412e-01 3.56240571e-01 3.11042488e-01 -1.42689854e-01
1.22495747e+00 5.54826319e-01 4.21292901e-01 -1.54813439e-01
-2.38251500e-02 -9.76336122e-01 -4.89721656e-01 5.54512680e-01
7.73452640e-01 -7.29814947e-01 5.43835163e-01 -5.98757625e-01
-7.05982864e-01 1.15901506e+00 2.48810261e-01 2.47924328e-01
1.20401073e+00 8.73915613e-01 -1.89019948e-01 -5.36778986e-01
-1.48411584e+00 -2.62370944e-01 5.11476636e-01 3.35675240e-01
7.58335292e-01 3.75972182e-01 -9.60790515e-01 1.30295599e+00
-1.84070505e-02 5.30650616e-01 -5.10679372e-02 1.04221940e+00
5.20350695e-01 -1.18445539e+00 4.67300676e-02 4.22856092e-01
-1.24164784e+00 -5.65780282e-01 -4.03533906e-01 8.42191815e-01
4.39009339e-01 5.75095236e-01 -8.07807893e-02 -2.37462714e-01
4.82212067e-01 5.55685401e-01 3.98328751e-01 -4.86690789e-01
-9.87564683e-01 3.83599043e-01 9.10777807e-01 -8.83370876e-01
-7.30856419e-01 -9.47460711e-01 -1.73191667e+00 3.84407997e-01
-1.73900604e-01 2.30436534e-01 4.46765631e-01 1.13696694e+00
3.98408860e-01 9.36892927e-01 6.50319755e-02 4.35961246e-01
-3.92104685e-02 -1.13171256e+00 -6.36702701e-02 1.52404249e-01
8.73369947e-02 -4.21258956e-01 1.03470698e-01 5.87767422e-01] | [8.51390266418457, 9.005548477172852] |
d426974d-f0f4-4f8f-937a-ccd7b2c41ec7 | learning-with-fuzzy-hypergraphs-a-topical | 1906.09445 | null | https://arxiv.org/abs/1906.09445v1 | https://arxiv.org/pdf/1906.09445v1.pdf | Learning with fuzzy hypergraphs: a topical approach to query-oriented text summarization | Existing graph-based methods for extractive document summarization represent sentences of a corpus as the nodes of a graph or a hypergraph in which edges depict relationships of lexical similarity between sentences. Such approaches fail to capture semantic similarities between sentences when they express a similar information but have few words in common and are thus lexically dissimilar. To overcome this issue, we propose to extract semantic similarities based on topical representations of sentences. Inspired by the Hierarchical Dirichlet Process, we propose a probabilistic topic model in order to infer topic distributions of sentences. As each topic defines a semantic connection among a group of sentences with a certain degree of membership for each sentence, we propose a fuzzy hypergraph model in which nodes are sentences and fuzzy hyperedges are topics. To produce an informative summary, we extract a set of sentences from the corpus by simultaneously maximizing their relevance to a user-defined query, their centrality in the fuzzy hypergraph and their coverage of topics present in the corpus. We formulate a polynomial time algorithm building on the theory of submodular functions to solve the associated optimization problem. A thorough comparative analysis with other graph-based summarization systems is included in the paper. Our obtained results show the superiority of our method in terms of content coverage of the summaries. | ['Tommy W. S. Chow', 'Hadrien Van Lierde'] | 2019-06-22 | null | null | null | null | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 3.44699383e-01 6.94429457e-01 -2.72476375e-01 -3.70059073e-01
-5.31705737e-01 -4.98289555e-01 5.91889024e-01 9.91263628e-01
8.45411886e-03 8.35286558e-01 6.81271911e-01 3.12667251e-01
-4.90725219e-01 -1.17604291e+00 -3.64334822e-01 -5.58301330e-01
8.08561966e-03 7.16776431e-01 3.20158899e-01 -3.55327189e-01
7.22684085e-01 1.31882176e-01 -1.48546040e+00 3.22613060e-01
1.27959704e+00 5.04898608e-01 3.90025526e-01 5.61220407e-01
-6.08621657e-01 3.05706203e-01 -9.23849225e-01 -4.23300594e-01
-1.20784543e-01 -5.39270282e-01 -1.03443849e+00 6.06117427e-01
2.13868096e-01 1.86851785e-01 -1.13126308e-01 1.49069905e+00
2.20542535e-01 4.42189068e-01 1.03354573e+00 -1.16032386e+00
-4.66102332e-01 8.53213429e-01 -6.34004056e-01 1.30610555e-01
6.82788134e-01 -6.99168265e-01 1.48242128e+00 -6.06792331e-01
7.71862984e-01 1.30864859e+00 3.10446084e-01 1.67951375e-01
-1.26914680e+00 1.42079934e-01 3.48903745e-01 1.25654504e-01
-1.34496653e+00 -2.81586796e-01 1.02857232e+00 -1.55674666e-01
9.76575255e-01 6.67281985e-01 7.70536065e-01 3.16068858e-01
5.04183412e-01 6.39227629e-01 4.20300186e-01 -5.20948648e-01
6.32614791e-01 2.75795996e-01 5.74448168e-01 5.79240143e-01
6.96834683e-01 -1.10476387e+00 -5.58575273e-01 -5.55819809e-01
-5.69323339e-02 -6.62811026e-02 -5.14292300e-01 -3.33400548e-01
-7.76950777e-01 1.11675274e+00 1.75245702e-01 5.87523818e-01
-7.74894834e-01 -1.39108866e-01 3.34328622e-01 1.37059361e-01
8.18767965e-01 5.72281659e-01 2.21445695e-01 5.88811219e-01
-1.13571882e+00 3.66658717e-01 1.28421831e+00 9.58103001e-01
9.66289997e-01 -3.23302358e-01 -2.49966562e-01 6.32051408e-01
3.72360617e-01 2.08002120e-01 2.80925810e-01 -8.32624078e-01
3.80349368e-01 1.03370774e+00 -1.82649001e-01 -1.81727350e+00
-1.89702317e-01 -1.77208751e-01 -7.90680468e-01 -5.43919742e-01
-3.09353858e-01 4.46788147e-02 -4.84566003e-01 1.43476188e+00
3.16302240e-01 -1.01404004e-01 3.68826807e-01 6.47930086e-01
1.00705671e+00 1.11398697e+00 -1.68298289e-01 -8.69008303e-01
1.46724629e+00 -5.58221340e-01 -1.06199312e+00 9.65160280e-02
3.14021856e-01 -5.97410083e-01 5.50868928e-01 2.06607714e-01
-1.30260253e+00 -1.62566826e-02 -1.02177131e+00 5.63388243e-02
-2.17154205e-01 -2.63513684e-01 3.22718352e-01 5.97059727e-01
-1.39855778e+00 4.98302251e-01 -4.15009826e-01 -7.74018407e-01
1.02390625e-01 3.30544531e-01 -7.45268613e-02 1.62111610e-01
-1.22758341e+00 7.28349805e-01 9.46626723e-01 -2.78354466e-01
-2.84727722e-01 -4.20142114e-01 -8.96129012e-01 4.79347438e-01
5.75880229e-01 -1.10028636e+00 8.05136740e-01 -6.99640870e-01
-1.04279387e+00 8.16753030e-01 -4.39909428e-01 -5.82774222e-01
-1.72150105e-01 4.44483221e-01 -1.42207950e-01 8.83869350e-01
2.82558680e-01 4.82458562e-01 6.89257026e-01 -1.50932693e+00
-7.55075991e-01 -4.32473481e-01 2.02488959e-01 5.81860304e-01
-6.60972714e-01 -1.35638326e-01 -4.55246836e-01 -4.98737574e-01
3.22950780e-01 -3.86818022e-01 -3.04987282e-01 -8.08240175e-01
-8.84658039e-01 -7.04161644e-01 7.43350923e-01 -5.65892339e-01
1.75017762e+00 -1.74538839e+00 4.37245965e-01 4.66426939e-01
5.72047114e-01 -2.40351647e-01 1.76226094e-01 1.02805340e+00
4.06167448e-01 2.88598865e-01 -6.03654623e-01 -1.97121412e-01
1.13764361e-01 2.46108800e-01 -4.71164048e-01 3.46950263e-01
-1.42551020e-01 4.31136370e-01 -1.04792869e+00 -9.35693920e-01
-1.13183156e-01 7.71431848e-02 -5.67367971e-01 -1.18744172e-01
-5.94631672e-01 -2.55291283e-01 -8.90094399e-01 1.03964560e-01
6.93058431e-01 -4.84798066e-02 3.93689752e-01 -1.30602226e-01
4.78578620e-02 2.13294864e-01 -1.07293510e+00 1.67725348e+00
9.67132673e-03 5.15822530e-01 4.30394569e-03 -1.42833638e+00
1.17013204e+00 3.19172025e-01 6.13919020e-01 3.69704105e-02
8.10956508e-02 -1.05840370e-01 -5.24497747e-01 -4.19334412e-01
1.23556352e+00 -2.60071367e-01 -3.75715107e-01 6.32636547e-01
6.99508712e-02 -5.91259718e-01 7.41502523e-01 1.00104475e+00
9.18111026e-01 -7.75488496e-01 6.20677292e-01 -7.82372415e-01
7.33594656e-01 2.22843856e-01 3.02271336e-01 6.79436207e-01
2.78870285e-01 4.23403829e-01 1.03390539e+00 1.13295745e-02
-9.12643611e-01 -9.58205402e-01 9.92334485e-02 6.97780430e-01
4.50479567e-01 -8.46000493e-01 -1.14257801e+00 -5.04574537e-01
-1.58611715e-01 1.09717500e+00 -4.62407142e-01 -1.75518155e-01
-1.94913059e-01 -6.44602478e-01 6.60491362e-02 -1.17431432e-01
4.10177946e-01 -8.65257740e-01 -4.63130653e-01 1.03144884e-01
-5.69409370e-01 -8.01534534e-01 -5.59253633e-01 -2.79842168e-01
-8.50638330e-01 -9.70318794e-01 -5.79069793e-01 -9.40538883e-01
9.16842163e-01 4.64414001e-01 1.07255602e+00 -6.31665513e-02
7.20664859e-02 7.36947834e-01 -4.85089481e-01 -4.78264004e-01
-5.84372282e-01 1.28996968e-01 -2.58279219e-02 1.21945582e-01
3.18730414e-01 -5.00783622e-01 -3.75704080e-01 -2.30281368e-01
-1.32776558e+00 -1.68056995e-01 1.35145679e-01 4.96747494e-01
5.57760298e-01 6.07967854e-01 8.31275582e-01 -9.82262850e-01
1.23619449e+00 -7.70780623e-01 -2.34066218e-01 5.21676064e-01
-5.01495600e-01 1.74619958e-01 4.72275078e-01 1.44475773e-01
-1.26917470e+00 -3.02727491e-01 4.02097553e-01 6.61870316e-02
1.69649571e-01 9.56655204e-01 -2.38412812e-01 5.10765493e-01
4.73251879e-01 3.98130089e-01 -3.57243195e-02 -5.58087565e-02
5.30358315e-01 7.89811254e-01 4.23478812e-01 -4.01647002e-01
3.38621140e-01 7.20197618e-01 1.25225738e-01 -1.37964237e+00
-8.21588397e-01 -8.81539047e-01 -4.30267751e-01 -3.64157259e-01
1.00541294e+00 -5.74606478e-01 -4.52187628e-01 -1.25429720e-01
-1.44131625e+00 7.83205211e-01 -5.27062774e-01 3.99627864e-01
-6.13247335e-01 8.75022173e-01 -6.69857636e-02 -8.28763664e-01
-5.86857378e-01 -6.05033696e-01 1.03762555e+00 4.05494690e-01
-4.53307122e-01 -1.30768430e+00 2.53511995e-01 2.14706540e-01
3.95200588e-02 3.23668510e-01 1.18890941e+00 -1.14079988e+00
-3.67610753e-01 -2.49340311e-01 1.89698711e-02 1.89549327e-01
3.44185919e-01 -3.50181758e-02 -3.62838119e-01 -2.49797165e-01
4.57525522e-01 2.80458063e-01 1.08186007e+00 7.53127217e-01
5.78590035e-01 -7.49221385e-01 -5.91454089e-01 -3.17340940e-02
1.46678090e+00 8.42722319e-03 4.42933947e-01 8.60046670e-02
4.32901680e-01 1.03645766e+00 3.30853999e-01 6.71492815e-01
4.63670522e-01 3.38868946e-01 2.68668681e-01 2.64067233e-01
2.73673713e-01 -1.61744907e-01 1.71771780e-01 1.10891831e+00
1.95555419e-01 -7.65263855e-01 -5.96711040e-01 7.79534161e-01
-1.94825554e+00 -1.10551214e+00 -4.36610639e-01 2.05582070e+00
7.19713926e-01 3.34372185e-03 1.41066775e-01 5.98243661e-02
1.37257218e+00 3.83372575e-01 -4.74421903e-02 -7.40369380e-01
-2.75029153e-01 -1.66858390e-01 1.60644382e-01 7.39596367e-01
-7.89617121e-01 8.38443220e-01 5.61861467e+00 9.56873655e-01
-4.03510571e-01 -2.23169789e-01 3.78691822e-01 4.20909598e-02
-1.03608942e+00 2.87458390e-01 -6.09492838e-01 2.56049544e-01
7.18890548e-01 -1.12425864e+00 -1.09815806e-01 4.88801509e-01
4.16396797e-01 -5.69902897e-01 -7.65726268e-01 4.20125932e-01
6.93948567e-01 -1.51056540e+00 6.15330279e-01 -3.95561382e-02
1.13172591e+00 -5.83176672e-01 -2.89508432e-01 -3.36973518e-01
1.05088376e-01 -5.35407901e-01 4.18804020e-01 7.54532039e-01
2.05257341e-01 -1.07901692e+00 6.77817285e-01 4.87627923e-01
-1.05997109e+00 3.70601386e-01 -7.99079955e-01 -9.89489164e-03
3.24132681e-01 8.85027170e-01 -9.89638984e-01 1.14752984e+00
1.82004109e-01 5.70940912e-01 -2.41634712e-01 9.92845774e-01
-1.32229969e-01 4.51882780e-01 -2.15982452e-01 -5.17235398e-01
2.14527369e-01 -6.37356460e-01 1.16790164e+00 1.24722743e+00
3.73915344e-01 3.20632517e-01 3.67107749e-01 8.93383741e-01
-2.68120289e-01 5.24467826e-01 -7.82028317e-01 -3.88062000e-02
5.28525114e-01 1.13488841e+00 -1.22867286e+00 -7.60874391e-01
-1.39476299e-01 8.73011827e-01 -1.22226989e-02 2.45163396e-01
-2.34571442e-01 -7.10203767e-01 2.68594146e-01 2.83246934e-02
8.11075494e-02 8.30476210e-02 -4.30418015e-01 -1.01354969e+00
5.43824323e-02 -4.17398214e-01 5.83547473e-01 -5.67342937e-01
-1.14374518e+00 5.21858335e-01 3.55215311e-01 -9.09625530e-01
-1.98216081e-01 2.38012001e-01 -9.41573501e-01 7.03301668e-01
-1.07924116e+00 -6.52825594e-01 1.10680927e-02 4.82901841e-01
7.44155347e-01 -1.17346227e-01 7.37839341e-01 -5.80264509e-01
-1.88464105e-01 -1.72804296e-01 1.61691844e-01 -4.37134951e-01
2.47968152e-01 -1.45990551e+00 1.76199168e-01 8.42335463e-01
1.22276405e-02 5.79223514e-01 1.15778732e+00 -1.07197714e+00
-1.08751619e+00 -9.04897630e-01 1.61118507e+00 8.95753801e-02
5.76026559e-01 -9.38362703e-02 -9.99336839e-01 3.98838341e-01
7.96120763e-01 -1.02364123e+00 7.55298734e-01 -4.63359207e-02
2.35094175e-01 1.33017480e-01 -1.15614641e+00 5.41864514e-01
6.38285279e-01 -1.72090292e-01 -1.29228592e+00 6.90388858e-01
8.59606922e-01 2.07770243e-01 -7.05722272e-01 -1.65450405e-02
-1.11034892e-01 -9.30466413e-01 5.78339040e-01 -4.81508583e-01
5.34396470e-01 -2.14173317e-01 -9.09595117e-02 -1.68307388e+00
-2.57725030e-01 -7.47741818e-01 1.55370340e-01 1.39071929e+00
2.64569402e-01 -4.70716655e-01 8.12135994e-01 3.14502358e-01
-2.67256498e-01 -4.87450957e-01 -7.52721310e-01 -4.67461258e-01
-1.56630218e-01 4.20786329e-02 3.54750395e-01 7.75671601e-01
7.11022854e-01 6.90328538e-01 3.42753492e-02 1.02812432e-01
7.67310619e-01 3.93458784e-01 3.87212306e-01 -1.63821924e+00
1.86951622e-01 -5.36947906e-01 -4.72940356e-01 -6.19690180e-01
5.19077003e-01 -1.16627896e+00 2.12837886e-02 -2.26233554e+00
6.91464365e-01 2.68397927e-01 3.01439285e-01 -2.57183671e-01
-1.41624779e-01 -3.80603373e-01 1.64846137e-01 3.14507842e-01
-8.99661660e-01 6.95993185e-01 1.04864061e+00 -3.31456393e-01
-4.26818788e-01 6.41877651e-02 -1.00112629e+00 9.04213786e-01
6.94832981e-01 -5.59721053e-01 -6.80372357e-01 -9.49009359e-02
3.79678130e-01 3.80945385e-01 -1.58657968e-01 -5.48722804e-01
5.44407189e-01 -2.10760118e-04 -2.65104562e-01 -8.82497013e-01
1.10608317e-01 -5.77402234e-01 -2.98937373e-02 4.12592113e-01
-6.02913558e-01 -6.38523474e-02 -1.97726026e-01 9.63921905e-01
-4.73305583e-01 -7.25888610e-01 4.01267201e-01 -1.62167624e-01
-4.51960772e-01 -5.60418256e-02 -7.50579119e-01 9.40235183e-02
1.07443464e+00 -4.35101569e-01 -2.99577296e-01 -6.65675700e-01
-6.26941204e-01 5.01541913e-01 4.70682323e-01 1.57765392e-03
8.61177504e-01 -1.04708660e+00 -1.03951383e+00 -4.18226093e-01
-1.54052665e-02 1.25539690e-01 4.35885906e-01 5.50746381e-01
-4.27605301e-01 6.56248808e-01 -6.16773404e-02 -4.12799060e-01
-1.28319705e+00 5.28127730e-01 -1.94515571e-01 -3.21358919e-01
-5.44709861e-01 5.72555780e-01 3.64381075e-01 9.91877988e-02
-3.20198387e-02 -2.47656077e-01 -7.63906658e-01 6.03771329e-01
3.27327251e-01 6.50383890e-01 -2.53768750e-02 -7.51825094e-01
-3.01687717e-01 4.19107735e-01 -1.03632495e-01 -1.63208976e-01
1.26922953e+00 -4.72464889e-01 -8.24943483e-01 4.42557842e-01
1.15746582e+00 2.03882828e-01 -4.68395203e-01 -2.15977818e-01
2.99325764e-01 -2.02359021e-01 7.67614022e-02 -9.39452127e-02
-6.87020600e-01 2.79636592e-01 -4.59519595e-01 1.08174074e+00
1.29288423e+00 5.04534066e-01 6.54348671e-01 4.91297752e-01
-6.07467517e-02 -1.26618719e+00 -1.29720774e-02 4.23013508e-01
1.12244189e+00 -6.23722136e-01 3.04576933e-01 -8.78001690e-01
-7.74463058e-01 1.18849576e+00 -6.39190525e-02 -3.07077229e-01
4.50999826e-01 -1.44379422e-01 -5.80499411e-01 -5.83022952e-01
-7.58402050e-01 -2.08179131e-01 4.91646081e-01 4.14388359e-01
2.31115818e-01 1.14071973e-01 -1.19484806e+00 5.66797018e-01
-4.00614053e-01 -5.78804910e-01 1.09135604e+00 7.51010835e-01
-1.21926451e+00 -6.46752059e-01 -4.45913047e-01 5.47906518e-01
-4.11146790e-01 -3.45692709e-02 -8.25753629e-01 2.36304939e-01
-5.00450730e-01 1.38395512e+00 2.24150792e-01 -1.24141432e-01
3.35510582e-01 1.23588461e-02 1.79264545e-01 -9.60631609e-01
-3.64209980e-01 1.29173309e-01 2.32763931e-01 4.45902087e-02
-5.29462755e-01 -8.14015865e-01 -1.47424948e+00 -3.09187353e-01
-4.46479708e-01 8.92330348e-01 7.37089992e-01 9.52143252e-01
2.08232060e-01 5.99886358e-01 7.18133330e-01 -6.11589551e-01
-5.43957174e-01 -8.70757937e-01 -1.05581284e+00 2.52625167e-01
-5.52273393e-02 -2.28791803e-01 -3.80956233e-01 1.58874556e-01] | [12.555212020874023, 9.576050758361816] |
d9ac49f6-856e-4341-afda-1d2fb708004b | point-cloud-learning-with-transformer | 2104.13636 | null | https://arxiv.org/abs/2104.13636v2 | https://arxiv.org/pdf/2104.13636v2.pdf | Point Cloud Learning with Transformer | Remarkable performance from Transformer networks in Natural Language Processing promote the development of these models in dealing with computer vision tasks such as image recognition and segmentation. In this paper, we introduce a novel framework, called Multi-level Multi-scale Point Transformer (MLMSPT) that works directly on the irregular point clouds for representation learning. Specifically, a point pyramid transformer is investigated to model features with diverse resolutions or scales we defined, followed by a multi-level transformer module to aggregate contextual information from different levels of each scale and enhance their interactions. While a multi-scale transformer module is designed to capture the dependencies among representations across different scales. Extensive evaluation on public benchmark datasets demonstrate the effectiveness and the competitive performance of our methods on 3D shape classification, segmentation tasks. | ['Xian-Feng Han', 'Qi Zhong'] | 2021-04-28 | null | null | null | null | ['3d-shape-retrieval'] | ['computer-vision'] | [ 1.44840613e-01 -2.53301710e-01 -4.20798175e-02 -4.96106654e-01
-8.58013868e-01 -4.15664434e-01 7.56511390e-01 2.65701890e-01
4.72415164e-02 -6.45066872e-02 1.42322689e-01 -3.58890891e-02
-2.23914325e-01 -1.22491312e+00 -8.89708042e-01 -6.64671540e-01
6.75376058e-02 3.96589369e-01 6.37098014e-01 -2.24647820e-01
2.18392178e-01 1.10655534e+00 -1.43902647e+00 7.62353837e-01
8.23980629e-01 1.22837007e+00 3.55675191e-01 1.76189497e-01
-6.75853431e-01 4.61996645e-01 -2.70019382e-01 -2.25226730e-01
2.71899641e-01 3.36867452e-01 -5.41690350e-01 2.92989343e-01
5.00992417e-01 2.27778181e-01 -1.38133720e-01 1.10159469e+00
2.35536203e-01 -3.16517800e-02 7.07054257e-01 -1.06758046e+00
-8.41818750e-01 4.88302439e-01 -9.54856277e-01 1.80108920e-01
6.62125051e-02 -6.18671291e-02 1.08649039e+00 -1.24574399e+00
4.39839214e-01 1.81798255e+00 8.53253901e-01 1.71117224e-02
-1.27042675e+00 -5.91757417e-01 2.94358104e-01 8.62996504e-02
-1.21231306e+00 1.01974227e-01 1.13918138e+00 -3.04792106e-01
1.00712383e+00 8.07993412e-02 6.59176886e-01 7.64866650e-01
1.67892098e-01 9.57143664e-01 1.13959730e+00 -1.35912791e-01
-9.13172364e-02 -1.31100178e-01 3.91198546e-01 8.03103864e-01
-1.32516026e-03 -2.19448522e-01 -4.97982830e-01 -2.51287580e-01
1.27596903e+00 2.03477487e-01 2.22492218e-01 -7.03669369e-01
-1.07295966e+00 7.52519250e-01 9.92134273e-01 6.45374775e-01
-6.85114980e-01 -3.83440070e-02 3.94793898e-01 1.98707476e-01
7.17095435e-01 8.30222070e-02 -4.67524767e-01 4.82236117e-01
-6.76653624e-01 1.02106594e-01 1.57322168e-01 1.18986249e+00
9.13904667e-01 -5.46300737e-03 -5.02864718e-01 1.20162523e+00
3.53510469e-01 3.55823517e-01 2.70575911e-01 -6.96286023e-01
7.15471983e-01 1.35707045e+00 -4.97590214e-01 -1.20815933e+00
-4.39247370e-01 -5.38719058e-01 -1.25325048e+00 1.93894833e-01
-6.86166435e-02 6.40245616e-01 -1.12116337e+00 1.44342768e+00
5.13541639e-01 5.61213672e-01 -6.90029413e-02 6.24653339e-01
1.05039978e+00 9.04146016e-01 2.69483060e-01 1.42308265e-01
1.56558728e+00 -7.48332739e-01 -1.65932268e-01 6.29311800e-02
2.33210206e-01 -7.28242457e-01 1.01458645e+00 2.89429754e-01
-1.24082911e+00 -1.07727087e+00 -8.17977011e-01 -2.74428487e-01
-5.92555225e-01 1.89947173e-01 6.08615935e-01 2.04005197e-01
-1.10107219e+00 6.84645772e-01 -9.79122937e-01 -2.69874781e-01
8.07697773e-01 4.42827016e-01 -3.63528013e-01 -8.47820193e-02
-6.91380799e-01 6.51315689e-01 2.38872975e-01 -3.50141264e-02
-5.03030658e-01 -8.40031683e-01 -9.89088535e-01 1.48736775e-01
-2.50247717e-02 -7.14375973e-01 7.18597293e-01 -6.15587354e-01
-1.24791622e+00 1.05057812e+00 -2.25511357e-01 -3.90040547e-01
1.27365261e-01 -1.38308406e-01 1.96175743e-02 2.11630166e-01
3.26961815e-01 7.65865028e-01 9.94525313e-01 -1.40014124e+00
-7.29019046e-01 -8.21408153e-01 6.27083555e-02 2.55220622e-01
-2.17065305e-01 5.14693856e-02 -7.63995349e-01 -7.41029322e-01
6.47887707e-01 -6.63496614e-01 -4.75711673e-01 -5.14495075e-02
-1.68778226e-01 -8.65796566e-01 1.01163864e+00 -3.29166085e-01
6.92296863e-01 -2.32643175e+00 3.97257864e-01 3.80134046e-01
2.53630847e-01 2.32747436e-01 -4.61469799e-01 2.00976089e-01
-2.50567961e-02 1.03898197e-01 -3.06836188e-01 -4.52294558e-01
-4.63746041e-02 4.60244030e-01 -4.86703485e-01 1.81674197e-01
6.53334260e-01 1.22958279e+00 -5.69512606e-01 -6.94182575e-01
6.51982546e-01 6.97688818e-01 -4.62270170e-01 1.20350584e-01
-1.34699442e-03 3.64611059e-01 -9.40782249e-01 9.41116512e-01
9.12821412e-01 -5.21041155e-01 -4.68715817e-01 -5.93624294e-01
-1.19626701e-01 1.31230667e-01 -1.07307577e+00 1.77343130e+00
-4.18665111e-01 -1.18645318e-01 1.04220837e-01 -1.07647848e+00
1.23125279e+00 1.21744357e-01 7.02231348e-01 -7.22401559e-01
-6.18847683e-02 7.24893576e-03 -3.88834000e-01 -1.53297603e-01
3.01496118e-01 -1.04972601e-01 -1.47796631e-01 -1.01242490e-01
7.37729743e-02 -6.61630034e-01 -3.36214378e-02 4.56890576e-02
8.50254476e-01 1.81551412e-01 3.38224769e-01 -3.16164374e-01
8.65481436e-01 -1.93767264e-01 5.94332457e-01 6.32295787e-01
-8.85289609e-02 5.18943608e-01 2.74399519e-01 -7.16813326e-01
-9.41858172e-01 -1.34623873e+00 -3.96276027e-01 1.24763227e+00
1.91902280e-01 -4.33542043e-01 -3.24197859e-01 -5.64770520e-01
3.16680461e-01 2.90666342e-01 -5.88973701e-01 2.52965242e-02
-7.47121513e-01 -5.61990082e-01 5.18369496e-01 8.73121381e-01
7.02549160e-01 -1.20152259e+00 -4.04191971e-01 1.58148229e-01
9.65012014e-02 -1.25620306e+00 -2.28073522e-01 2.29521424e-01
-1.11094129e+00 -8.66680324e-01 -4.21541810e-01 -1.07954109e+00
5.13430655e-01 2.49815345e-01 1.18301582e+00 -1.86103299e-01
-1.35625526e-01 4.47519809e-01 -3.37163955e-01 -3.69422883e-01
-1.50841057e-01 9.73396078e-02 -2.84099191e-01 1.36886895e-01
2.64522761e-01 -8.37640584e-01 -2.26973787e-01 1.92636207e-01
-1.10782909e+00 7.69423023e-02 9.94625688e-01 7.33805716e-01
1.06304348e+00 -1.78460643e-01 4.21104342e-01 -7.73201406e-01
5.65126717e-01 -2.66372800e-01 -5.82506001e-01 3.57307225e-01
-7.71475956e-02 -3.23317088e-02 6.24465644e-01 -3.40453327e-01
-9.81353939e-01 2.36443326e-01 -3.11264008e-01 -8.18299651e-01
-4.03382748e-01 4.15242285e-01 -3.28563869e-01 -3.03963304e-01
3.03361505e-01 5.14614344e-01 -2.93758839e-01 -6.01135075e-01
5.17994821e-01 3.51083249e-01 4.56505209e-01 -8.51273596e-01
1.00685465e+00 5.81240594e-01 2.56924063e-01 -9.28585052e-01
-8.49576473e-01 -4.48174775e-01 -1.01643860e+00 1.04668379e-01
9.69012439e-01 -8.16255987e-01 -3.43691587e-01 5.30237377e-01
-1.36387873e+00 2.07886338e-01 -4.35892910e-01 3.97757404e-02
-5.43263257e-01 3.00982773e-01 -7.92119801e-01 -4.01099324e-01
-4.64093566e-01 -1.22692084e+00 1.69442403e+00 3.64188731e-01
4.55337912e-01 -9.21593606e-01 -1.22563891e-01 3.30538690e-01
2.59430885e-01 3.81663620e-01 1.31538498e+00 -5.40072620e-01
-8.25007677e-01 7.72325769e-02 -5.99379301e-01 5.25855601e-01
3.02692354e-01 -1.68662712e-01 -7.07809865e-01 -2.86688328e-01
-2.12604436e-03 -3.12865049e-01 8.62088561e-01 4.28836524e-01
1.78657627e+00 8.92065838e-02 -4.03197348e-01 8.24840903e-01
1.33240402e+00 -4.62036431e-02 4.39341486e-01 8.52729902e-02
9.86854732e-01 4.18437451e-01 3.65645349e-01 2.06535921e-01
3.71607214e-01 5.66864610e-01 2.98190325e-01 -3.32428932e-01
-5.42272143e-02 -2.85447717e-01 -1.53797083e-02 1.00954330e+00
-1.64842203e-01 2.90109396e-01 -9.26485777e-01 3.98218542e-01
-1.71456873e+00 -6.23813808e-01 -1.37537876e-02 1.72218347e+00
4.69631940e-01 2.63137430e-01 -2.20348805e-01 -1.52711913e-01
6.59902155e-01 4.16478544e-01 -5.86108565e-01 -2.37697527e-01
-3.60409528e-01 5.83030939e-01 2.02394798e-01 1.49938524e-01
-1.21750009e+00 1.26564479e+00 5.74094534e+00 1.09523296e+00
-1.06333649e+00 -1.38762593e-01 6.09394550e-01 6.15940869e-01
-2.08371505e-01 -3.02817971e-01 -7.14436710e-01 -5.76552227e-02
2.70694792e-01 -8.01831409e-02 6.41554967e-02 7.50893056e-01
-3.10767884e-03 6.15433514e-01 -1.00435686e+00 1.07110584e+00
-1.06100209e-01 -1.40609920e+00 8.24059725e-01 -7.47514963e-02
4.96745855e-01 3.68443429e-01 1.46706134e-01 4.41136807e-01
1.98077798e-01 -8.77376676e-01 3.30773175e-01 5.28125286e-01
5.21696806e-01 -7.71306932e-01 3.68167967e-01 3.53632569e-01
-1.96486306e+00 4.21224795e-02 -7.21988976e-01 3.69446725e-01
-5.27652502e-02 3.02692354e-01 -6.46963239e-01 9.56473827e-01
9.33530331e-01 1.24442911e+00 -7.89612949e-01 8.51565301e-01
-8.59230980e-02 3.54751647e-01 -4.79994059e-01 2.78105438e-01
4.97211844e-01 -3.83102924e-01 5.90944707e-01 1.38991129e+00
3.45404536e-01 2.54678965e-01 3.80579710e-01 1.02639413e+00
-1.94319021e-02 3.07274669e-01 -8.95775259e-01 2.79972285e-01
4.64683235e-01 1.46228778e+00 -7.99374580e-01 -4.36265737e-01
-6.02986515e-01 5.70109546e-01 6.15351498e-01 2.22004235e-01
-6.95093632e-01 6.03550486e-02 7.28750527e-01 -1.02958620e-01
6.29464984e-01 -4.27671224e-01 -5.46051502e-01 -9.63399410e-01
-3.02201957e-02 -6.58789933e-01 5.33716917e-01 -6.94249213e-01
-1.81945860e+00 7.22290814e-01 2.04781562e-01 -1.31019843e+00
3.58799249e-01 -7.77487934e-01 -6.08002841e-01 7.92481840e-01
-1.72461879e+00 -1.76412439e+00 -1.26116514e-01 1.06994712e+00
7.45355129e-01 -1.53201535e-01 4.89272267e-01 7.18695596e-02
-2.34310105e-01 2.37377122e-01 -2.36249417e-01 1.92056403e-01
4.15825918e-02 -1.41472220e+00 5.04962504e-01 5.40907741e-01
4.62380320e-01 6.79421723e-01 6.78365678e-02 -4.51753825e-01
-1.31513429e+00 -1.35385120e+00 5.47873974e-01 -3.12762737e-01
6.10310256e-01 -3.41477513e-01 -1.27202308e+00 5.12905121e-01
-4.89844158e-02 3.57720971e-01 4.24004704e-01 1.12260178e-01
-6.79393888e-01 -3.09093028e-01 -1.09253740e+00 3.52581680e-01
1.20677662e+00 -5.52717984e-01 -1.08944690e+00 2.41960630e-01
8.93053353e-01 -4.86871928e-01 -1.11673641e+00 9.79493499e-01
8.61312076e-02 -7.71768093e-01 1.60754931e+00 -5.60620487e-01
3.59137595e-01 -1.95706561e-01 -3.56913626e-01 -1.18530238e+00
-7.89151073e-01 9.36848447e-02 2.20590636e-01 1.24123394e+00
1.04462624e-01 -6.21790767e-01 6.95722699e-01 1.48739219e-01
-4.09342319e-01 -9.49188650e-01 -1.24524403e+00 -4.90199834e-01
3.47370535e-01 -6.00342989e-01 8.11025500e-01 7.95412004e-01
-5.14323473e-01 5.27045667e-01 1.68389112e-01 5.75618207e-01
7.88218260e-01 6.66561723e-01 5.60772717e-01 -1.45213795e+00
-2.40901560e-02 -6.90874398e-01 -6.68526709e-01 -1.30529118e+00
2.06955940e-01 -1.15956640e+00 -2.46196747e-01 -1.62894487e+00
1.47106349e-01 -6.07404053e-01 -7.50170231e-01 5.29979050e-01
-9.96904969e-02 2.50321746e-01 3.69156688e-01 2.66849279e-01
-6.08856857e-01 8.82080495e-01 1.56346774e+00 -3.94425541e-01
-2.09926680e-01 2.08403319e-01 -5.15632033e-01 1.12610042e+00
4.80638266e-01 -2.29073167e-01 -3.99866581e-01 -6.54791057e-01
-1.90819070e-01 -2.31047673e-03 4.29494649e-01 -1.09657705e+00
2.63173550e-01 -1.80262133e-01 5.81196547e-01 -1.15730119e+00
5.78126788e-01 -9.38369334e-01 -1.89675406e-01 1.58477187e-01
-3.00327867e-01 1.49725959e-01 2.94021815e-01 5.09319007e-01
-4.93540406e-01 1.96899846e-01 9.22438681e-01 -3.55304360e-01
-6.85544193e-01 7.68255353e-01 2.38310695e-01 -2.24522889e-01
1.00755930e+00 -1.57834932e-01 -6.69233948e-02 2.27108806e-01
-8.32690477e-01 2.48854265e-01 1.79815471e-01 7.32876241e-01
1.04446673e+00 -1.57912457e+00 -7.25575805e-01 4.55774903e-01
2.51342714e-01 5.08045077e-01 7.21620694e-02 5.17131031e-01
-2.57510573e-01 4.33747351e-01 -4.36344206e-01 -1.16351867e+00
-1.34219587e+00 4.74762768e-01 3.28363657e-01 -6.29270434e-01
-9.10017133e-01 9.64148760e-01 7.70018697e-01 -7.57137477e-01
9.50497687e-02 -8.36696923e-01 -6.20872915e-01 -1.26126530e-02
1.61816850e-01 8.93371701e-02 5.55765517e-02 -1.01415336e+00
-3.46771717e-01 1.29680014e+00 -1.24549925e-01 2.93598503e-01
1.71630383e+00 9.92874503e-02 -5.02148092e-01 5.19350946e-01
1.01618075e+00 -1.82234988e-01 -1.19677484e+00 -7.92863727e-01
1.81753382e-01 -3.31438631e-01 -9.02299806e-02 -3.23013514e-01
-1.17496991e+00 8.45188618e-01 5.12829304e-01 3.22583228e-01
1.30267334e+00 4.01260316e-01 6.33318663e-01 4.63254333e-01
5.99221528e-01 -7.01533020e-01 1.89526841e-01 7.75381863e-01
1.27943277e+00 -1.04388297e+00 -1.83959544e-01 -7.34975994e-01
-5.30511856e-01 1.02890873e+00 6.87452793e-01 -6.17570102e-01
9.03035700e-01 2.23119885e-01 -1.53144762e-01 -4.22919422e-01
-6.62751913e-01 -3.20830256e-01 6.68628573e-01 5.74485600e-01
1.49733394e-01 1.37268856e-01 2.69340212e-03 5.91947377e-01
-4.63641509e-02 -2.32073590e-01 -1.82157353e-01 7.58977830e-01
-4.68322158e-01 -1.25440192e+00 -4.19312775e-01 5.67494929e-01
-2.20863912e-02 8.70409608e-02 -3.75529438e-01 6.16553366e-01
4.06911105e-01 5.57659030e-01 1.38713211e-01 -3.28637987e-01
6.01831555e-01 2.13131309e-04 3.75233263e-01 -7.85913289e-01
-8.62912536e-01 1.62147656e-01 -4.61837560e-01 -7.35217273e-01
-6.09106839e-01 -8.31845820e-01 -1.34564686e+00 2.79229075e-01
1.39232010e-01 -1.40999211e-02 4.38114166e-01 9.04678524e-01
4.16596323e-01 5.51734209e-01 6.95619106e-01 -1.02900410e+00
-5.69912314e-01 -9.63851750e-01 -5.21153510e-01 6.05408728e-01
3.64649564e-01 -6.58301473e-01 1.17953832e-03 -1.25234231e-01] | [7.986091613769531, -3.566007137298584] |
623edd58-9056-400b-9fd4-a48f1a960663 | satellite-galaxy-abundance-dependency-on | 2110.05498 | null | https://arxiv.org/abs/2110.05498v2 | https://arxiv.org/pdf/2110.05498v2.pdf | Satellite galaxy abundance dependency on cosmology in Magneticum simulations | Context: Modelling satellite galaxy abundance $N_s$ in Galaxy Clusters (GCs) is a key element in modelling the Halo Occupation Distribution (HOD), which itself is a powerful tool to connect observational studies with numerical simulations. Aims: To study the impact of cosmological parameters on satellite abundance both in cosmological simulations and in mock observations. Methods: We build an emulator (HODEmu, \url{https://github.com/aragagnin/HODEmu/}) of satellite abundance based on cosmological parameters $\Omega_m, \Omega_b, \sigma_8, h_0$ and redshift $z.$ We train our emulator using \magneticum hydrodynamic simulations that span 15 different cosmologies, each over $4$ redshift slices between $0<z<0.5,$ and for each setup we fit normalisation $A$, log-slope $\beta$ and Gaussian fractional-scatter $\sigma$ of the $N_s-M$ relation. The emulator is based on multi-variate output Gaussian Process Regression (GPR). Results: We find that $A$ and $\beta$ depend on cosmological parameters, even if weakly, especially on $\Omega_m,$ $\Omega_b.$ This dependency can explain some discrepancies found in literature between satellite HOD of different cosmological simulations (Magneticum, Illustris, BAHAMAS). We also show that satellite abundance cosmology dependency differs between full-physics (FP) simulations, dark-matter only (DMO), and non-radiative simulations. Conclusions: This work provides a preliminary calibration of the cosmological dependency of the satellite abundance of high mass halos, and we showed that modelling HOD with cosmological parameters is necessary to interpret satellite abundance, and we showed the importance of using FP simulations in modelling this dependency. | ['Sebastian Bocquet', 'Matteo Costanzi', 'Alexandro Saro', 'Klaus Dolag', 'Tiago Castro', 'Alessandra Fumagalli', 'Antonio Ragagnin'] | 2021-10-11 | null | null | null | null | ['gpr', 'gpr'] | ['computer-vision', 'miscellaneous'] | [-6.32308722e-01 -2.57767260e-01 5.05203545e-01 -7.93470666e-02
-2.32974470e-01 -5.20190239e-01 1.11653912e+00 -3.02605122e-01
-6.26900136e-01 7.54890800e-01 -2.14061603e-01 -6.55193746e-01
-3.85385081e-02 -1.23450172e+00 -5.86903334e-01 -1.45704782e+00
-4.16128755e-01 9.52681601e-01 6.14097893e-01 -4.73279834e-01
2.08887726e-01 1.63930357e-01 -1.76903868e+00 -6.75914407e-01
7.79867232e-01 7.12897480e-01 4.47102606e-01 7.88157523e-01
1.50836647e-01 2.46926591e-01 -2.66848147e-01 -2.52102107e-01
4.97329205e-01 -5.78584373e-01 -5.33709705e-01 -2.53074855e-01
-3.25348407e-01 2.28080824e-01 -2.86758095e-01 1.30389524e+00
5.85663795e-01 5.69191158e-01 1.10801244e+00 -6.06834769e-01
-1.43614545e-01 5.00111938e-01 -4.65972900e-01 7.24802852e-01
-5.59746087e-01 7.66302288e-01 4.37254339e-01 -5.12155354e-01
4.42969531e-01 1.16794336e+00 5.04424989e-01 1.73240706e-01
-1.36572134e+00 -8.65506411e-01 -6.40636504e-01 -2.54610240e-01
-1.68835199e+00 -5.12578078e-02 4.63449061e-02 -8.69066358e-01
1.31902802e+00 4.90297794e-01 7.38032639e-01 5.91178238e-01
2.16753498e-01 -5.89419723e-01 1.43257117e+00 -5.38877130e-01
5.02352238e-01 3.79082680e-01 -1.32471234e-01 2.34739453e-01
3.34671974e-01 5.47892690e-01 -1.06822491e-01 -3.32337439e-01
1.05384696e+00 -8.42196345e-01 -5.81070967e-02 2.20932633e-01
-9.67146933e-01 1.18959486e+00 9.49919373e-02 4.85172689e-01
-5.72390929e-02 5.06904781e-01 1.74490020e-01 6.63343519e-02
8.00617456e-01 3.17013979e-01 -4.50464934e-01 -2.27543980e-01
-8.30438197e-01 4.95237261e-01 5.99270344e-01 5.75508893e-01
9.21611428e-01 3.86784375e-01 7.01546967e-01 1.04970837e+00
7.63464510e-01 1.18973494e+00 8.40408862e-01 -1.03253484e+00
-1.22336961e-01 1.00270294e-01 1.20696925e-01 -5.86517274e-01
-4.48505580e-01 -5.73270023e-01 -9.01399314e-01 4.48699117e-01
3.85469317e-01 -1.50083005e-01 -4.20386761e-01 1.62680519e+00
3.71026337e-01 9.72294733e-02 -1.12425260e-01 7.93059170e-01
6.56679571e-01 7.23025322e-01 3.42251837e-01 -2.78146625e-01
1.32213855e+00 -4.37432170e-01 2.89086074e-01 -3.32215756e-01
5.26584089e-01 -7.01977611e-01 7.80658484e-01 8.97500813e-02
-1.22235215e+00 -6.05503082e-01 -7.21396327e-01 5.08401215e-01
-3.53547812e-01 -4.35533851e-01 7.64204025e-01 1.21126568e+00
-1.40485740e+00 8.46827030e-01 -1.02456319e+00 -7.07805037e-01
-6.04712665e-01 4.70327765e-01 2.27938130e-01 7.27403581e-01
-1.24557662e+00 7.72565246e-01 5.93050838e-01 -3.89103472e-01
-6.61872387e-01 -3.26834768e-01 -3.22454929e-01 3.23689431e-02
-2.55132288e-01 -9.92520928e-01 1.04172730e+00 -5.68518698e-01
-1.14372432e+00 1.22112012e+00 1.73008338e-01 -6.43723071e-01
3.44021395e-02 7.66451120e-01 -4.86160159e-01 -2.49379445e-02
-1.65251985e-01 5.86435497e-01 2.64713317e-01 -1.12642670e+00
-4.65606779e-01 -4.31236595e-01 -3.92172247e-01 -1.06880330e-02
3.26757312e-01 5.80115616e-01 -1.39673963e-01 -2.34888330e-01
3.75366777e-01 -1.01405084e+00 -4.88143653e-01 -1.04364049e+00
1.97605029e-01 -1.42433122e-01 6.84412718e-02 -8.05740535e-01
1.16715407e+00 -1.84482884e+00 3.08325291e-01 4.73904043e-01
-7.99075067e-02 -1.35815978e-01 7.10411370e-01 2.77733386e-01
-1.57957196e-01 4.95742083e-01 -5.06318271e-01 -3.11399579e-01
3.95078093e-01 1.75130710e-01 7.45671168e-02 6.29193306e-01
-4.09734994e-01 2.01559097e-01 -5.91537893e-01 -2.95413762e-01
4.93301392e-01 5.14185667e-01 -4.29949462e-01 4.03180420e-02
-4.75582272e-01 8.24660063e-01 1.83144212e-01 1.41860828e-01
1.03200948e+00 -2.38502577e-01 6.38481528e-02 4.64132816e-01
-8.94504249e-01 3.10639054e-01 -1.17959380e+00 1.09925473e+00
-1.71021044e-01 3.45775813e-01 4.14823890e-01 -8.29246938e-01
1.16868365e+00 -2.20562629e-02 1.86213434e-01 -7.03125060e-01
5.96887648e-01 5.85118234e-01 5.57592332e-01 -1.31598860e-01
6.73780322e-01 -8.54809999e-01 1.31791294e-01 4.46574152e-01
2.69377977e-01 -7.68242180e-01 3.85806292e-01 2.38601938e-01
6.31607592e-01 7.25173876e-02 3.50497589e-02 -1.10898006e+00
5.85630238e-01 -2.55053014e-01 9.20859426e-02 6.71800315e-01
1.08220264e-01 5.93728840e-01 7.79246926e-01 8.74265134e-02
-1.57700086e+00 -9.80581164e-01 -6.73317790e-01 1.01781976e+00
-4.98816520e-02 -3.87654155e-01 -8.39227498e-01 2.31681943e-01
-2.12081105e-01 1.06237042e+00 -4.89697754e-01 -1.08597159e-01
-4.72737491e-01 -2.18259454e+00 3.38091224e-01 -1.83018029e-01
4.92418766e-01 -8.50106895e-01 -3.84906948e-01 -1.47845432e-01
8.22039843e-02 -4.52112764e-01 2.07879588e-01 3.04224312e-01
-9.67387736e-01 -4.91075099e-01 -4.37741011e-01 -2.53968656e-01
1.98802888e-01 -2.05130175e-01 1.56229579e+00 3.72723043e-01
4.76743206e-02 1.57405853e-01 -2.14532822e-01 -2.39496022e-01
-8.06052506e-01 -8.14251453e-02 2.90304929e-01 -5.25099754e-01
4.62614059e-01 -1.09709930e+00 -8.87378454e-01 6.29694819e-01
-9.80336905e-01 -3.30632508e-01 4.00621295e-02 3.36762428e-01
4.36721802e-01 3.99898261e-01 1.00846766e-02 -2.21289605e-01
-1.17619291e-01 -9.91829336e-01 -1.15018249e+00 -5.56141555e-01
-7.59658337e-01 -8.69121850e-02 2.70828396e-01 9.08010527e-02
-8.64968359e-01 -7.61698365e-01 -7.26146042e-01 -6.97036926e-03
-3.00453067e-01 3.73384714e-01 9.90574211e-02 -2.13866513e-02
8.58999670e-01 3.04020733e-01 1.40202418e-01 -7.27750838e-01
-2.18726099e-01 4.69289482e-01 5.05149662e-01 -8.88045847e-01
7.44050801e-01 6.43420637e-01 2.95715928e-01 -1.10215068e+00
2.21419170e-01 -6.44687116e-01 -5.68206966e-01 -3.01458061e-01
9.63651896e-01 -1.00524890e+00 -8.21266234e-01 3.47521484e-01
-5.56752861e-01 -6.79614007e-01 -6.12132996e-02 7.82129169e-01
-5.02503812e-01 5.73207319e-01 -4.47460860e-01 -1.16989648e+00
-3.09078306e-01 -1.25450289e+00 6.92609966e-01 4.09098953e-01
1.93206251e-01 -1.42173827e+00 5.07886052e-01 1.97280958e-01
5.96769512e-01 1.22832067e-01 8.33378196e-01 -2.54218042e-01
-5.30830979e-01 1.88866675e-01 -2.74839222e-01 3.04894000e-01
-5.66720128e-01 3.67593020e-01 -9.28556740e-01 -4.08339888e-01
4.23400253e-01 2.68721163e-01 9.24628615e-01 6.92621469e-01
6.26939774e-01 7.45204613e-02 -1.15145981e-01 8.37463915e-01
1.72968948e+00 6.73636496e-02 9.77964938e-01 8.08174670e-01
3.57009917e-02 5.65700531e-01 1.59275740e-01 6.35588586e-01
6.85533732e-02 6.04020417e-01 6.27732337e-01 2.26517782e-01
1.64093912e-01 6.04433417e-01 5.30921161e-01 1.05859804e+00
-9.15960670e-01 7.66746104e-02 -1.08279216e+00 3.46784025e-01
-1.40457022e+00 -9.66508329e-01 -8.80737126e-01 2.75986600e+00
5.70319116e-01 2.23365068e-01 5.72538018e-01 -5.72403014e-01
5.41258574e-01 -3.07994355e-02 8.02984610e-02 -2.89687693e-01
-4.17887181e-01 7.41774797e-01 9.96001899e-01 7.55238175e-01
-7.10652292e-01 6.27966106e-01 5.36198378e+00 9.78641987e-01
-1.09226632e+00 6.37061000e-01 6.57898486e-01 -2.63244390e-01
-3.06373090e-01 2.40976959e-01 -9.32222486e-01 9.67163384e-01
1.94644308e+00 -2.25419968e-01 4.53577518e-01 7.63713658e-01
5.16200542e-01 -8.27616692e-01 -2.57672966e-01 6.85448229e-01
-2.26177052e-01 -9.35038924e-01 -6.15481615e-01 5.78323781e-01
6.86508715e-01 7.54372537e-01 4.84254807e-02 5.54858148e-01
6.01776838e-01 -9.41067040e-01 6.12931073e-01 4.41228867e-01
8.31433654e-01 -9.15395141e-01 8.99530947e-01 4.10470545e-01
-9.72267687e-01 3.97556961e-01 -1.04987228e+00 -2.19765067e-01
3.86821806e-01 6.96705937e-01 -7.26427913e-01 5.52926600e-01
1.19232249e+00 -1.58475041e-01 -7.72919953e-01 8.10627162e-01
3.32339168e-01 8.35241497e-01 -7.22116828e-01 8.93865824e-02
3.39512408e-01 -1.19413018e+00 4.96097654e-01 1.36492491e+00
9.09113407e-01 3.65555435e-01 -5.82349539e-01 1.25223672e+00
5.74450970e-01 2.16760203e-01 -2.03398734e-01 3.95527966e-02
7.70113468e-02 1.15300691e+00 -1.26421022e+00 -3.26337099e-01
-3.79474819e-01 3.08680594e-01 -3.04046959e-01 -2.52056587e-02
-9.31036830e-01 9.76077542e-02 9.22738433e-01 5.70702970e-01
4.42342639e-01 -6.09999835e-01 -2.53687352e-01 -1.03406572e+00
-8.66873384e-01 -4.71004546e-01 3.23683679e-01 -9.65871453e-01
-9.97663379e-01 3.93064171e-01 3.81077319e-01 -8.55212092e-01
-4.12403554e-01 -9.18024302e-01 -7.23199487e-01 1.57402611e+00
-7.29736149e-01 -6.20281279e-01 2.05220059e-02 1.78052988e-02
1.02948971e-01 -5.99726029e-02 6.97432458e-01 1.63188383e-01
-5.86761117e-01 -7.65081346e-02 8.86058450e-01 -4.91626680e-01
4.19178188e-01 -1.56259108e+00 6.18242860e-01 5.67993343e-01
-4.79662150e-01 7.94956326e-01 1.54559898e+00 -9.70925689e-01
-7.29507983e-01 -7.13009894e-01 4.23268497e-01 -5.42100132e-01
1.02815282e+00 -4.01620030e-01 -8.80086839e-01 5.17175913e-01
5.88997066e-01 -3.99810553e-01 7.54103661e-01 1.35227488e-02
1.52702957e-01 4.60911542e-01 -1.20901620e+00 5.02058089e-01
6.49164498e-01 -2.74951547e-01 -1.94471136e-01 4.86301631e-01
5.88481307e-01 7.12898150e-02 -1.25943840e+00 2.66129196e-01
2.53093690e-01 -1.63262522e+00 1.26572108e+00 -1.64481491e-01
6.41018823e-02 -3.30264598e-01 -4.26411897e-01 -1.18841422e+00
-4.65531349e-01 -3.00525486e-01 4.31373268e-01 1.01245224e+00
4.56386507e-01 -7.36165941e-01 6.08823955e-01 5.76039672e-01
-3.02849263e-01 5.29097207e-02 -1.18154609e+00 -7.94567466e-01
9.85864758e-01 -5.63188136e-01 6.09841943e-01 8.19046080e-01
-3.22423697e-01 3.66611816e-02 -5.48391379e-02 5.43662548e-01
8.53351533e-01 -5.23363233e-01 6.30107701e-01 -1.36539292e+00
-1.00077450e+00 -8.22622657e-01 -1.31794393e-01 -3.98928225e-01
-1.03826307e-01 -8.95786405e-01 -1.87752008e-01 -1.02704430e+00
2.69904166e-01 -5.60451388e-01 2.06803963e-01 -5.83947599e-01
1.83008939e-01 4.29337859e-01 -2.18369335e-01 3.99234354e-01
-1.48509070e-01 3.69016558e-01 5.84010243e-01 3.97476584e-01
1.92141384e-01 -2.12886691e-01 5.97150475e-02 8.73335600e-01
9.07904983e-01 -3.09802145e-01 2.38325506e-01 9.52342376e-02
6.59551203e-01 -4.29229438e-02 5.12983441e-01 -1.23852968e+00
-3.12596917e-01 -2.18509957e-01 1.45787850e-01 -5.03333151e-01
3.77034426e-01 -2.29323328e-01 8.07843685e-01 3.52518022e-01
6.95303321e-01 -7.33899400e-02 1.94490477e-01 5.27488701e-02
2.21914917e-01 -9.91210878e-01 1.27206171e+00 -7.91784585e-01
-3.64977360e-01 -2.97921468e-02 -6.91008389e-01 -4.20799136e-01
4.05850917e-01 1.41474694e-01 -3.52864116e-01 -3.08727652e-01
-8.20113957e-01 -1.24902353e-01 1.19040918e+00 -4.86790776e-01
-4.74987805e-01 -9.11725998e-01 -5.09485245e-01 -2.86669750e-03
-3.15717131e-01 8.82321596e-03 5.28793752e-01 1.27293408e+00
-1.01714063e+00 4.24970388e-01 1.04814358e-01 -6.37217522e-01
-9.02980924e-01 5.48176885e-01 8.46199095e-01 -1.77570935e-02
5.29373065e-02 1.17382097e+00 3.19287270e-01 -7.65462101e-01
-4.29954797e-01 -7.90045038e-02 -1.48043362e-02 -8.00998583e-02
3.87738168e-01 6.30459666e-01 1.07919455e-01 -1.05268133e+00
1.01455502e-01 4.56102192e-01 6.04188681e-01 -5.73189855e-01
1.19139445e+00 -5.85638344e-01 -7.95184851e-01 7.41725206e-01
7.99754500e-01 8.94580316e-03 -9.01173830e-01 1.17856950e-01
-7.03913122e-02 -3.48553479e-01 1.14399772e-02 -3.47298354e-01
-7.17497170e-01 6.93056762e-01 7.14407086e-01 7.47646093e-01
6.03855133e-01 5.39767265e-01 2.13147908e-01 -1.55953467e-01
5.44866264e-01 -1.23318624e+00 -4.12774771e-01 6.39386654e-01
7.87474394e-01 -1.12288940e+00 -7.06486627e-02 -2.18706936e-01
-1.75266787e-01 8.72163475e-01 5.90276003e-01 -1.00859506e-02
8.84865344e-01 -8.42395425e-02 -2.77521491e-01 -4.41334188e-01
-5.04770935e-01 -4.67715830e-01 -2.42612705e-01 1.70348987e-01
6.19002998e-01 2.47142315e-01 -8.17589343e-01 5.58258295e-01
-9.38234091e-01 -8.21391463e-01 6.68175995e-01 5.25560796e-01
-8.98847640e-01 -1.26707959e+00 -1.15606415e+00 4.87736732e-01
-3.34689826e-01 -3.38332504e-01 4.12182920e-02 8.86025608e-01
4.68723446e-01 7.01707661e-01 5.74746609e-01 -3.71747673e-01
-2.86728978e-01 4.92842019e-01 3.20538551e-01 -5.01381874e-01
-6.93181515e-01 6.19606435e-01 1.77326620e-01 2.09431916e-01
-4.24718738e-01 -1.18064499e+00 -1.26274097e+00 -9.57583129e-01
-3.31288069e-01 6.65777743e-01 1.06112969e+00 6.41315401e-01
-3.53871405e-01 2.12201267e-01 5.51305830e-01 -1.25000191e+00
-2.70954698e-01 -1.56474710e+00 -1.39006639e+00 1.76007062e-01
-5.08619845e-01 -9.16748047e-01 -1.46824956e+00 -1.52744368e-01] | [7.0327372550964355, 3.4071221351623535] |
b2179e1d-f5e5-4f7c-ae58-83f65a1880a8 | indoor-localization-for-personalized-ambient | 2207.09025 | null | https://arxiv.org/abs/2207.09025v1 | https://arxiv.org/pdf/2207.09025v1.pdf | Indoor Localization for Personalized Ambient Assisted Living of Multiple Users in Multi-Floor Smart Environments | This paper presents a multifunctional interdisciplinary framework that makes four scientific contributions towards the development of personalized ambient assisted living, with a specific focus to address the different and dynamic needs of the diverse aging population in the future of smart living environments. First, it presents a probabilistic reasoning-based mathematical approach to model all possible forms of user interactions for any activity arising from the user diversity of multiple users in such environments. Second, it presents a system that uses this approach with a machine learning method to model individual user profiles and user-specific user interactions for detecting the dynamic indoor location of each specific user. Third, to address the need to develop highly accurate indoor localization systems for increased trust, reliance, and seamless user acceptance, the framework introduces a novel methodology where two boosting approaches Gradient Boosting and the AdaBoost algorithm are integrated and used on a decision tree-based learning model to perform indoor localization. Fourth, the framework introduces two novel functionalities to provide semantic context to indoor localization in terms of detecting each user's floor-specific location as well as tracking whether a specific user was located inside or outside a given spatial region in a multi-floor-based indoor setting. These novel functionalities of the proposed framework were tested on a dataset of localization-related Big Data collected from 18 different users who navigated in 3 buildings consisting of 5 floors and 254 indoor spatial regions. The results show that this approach of indoor localization for personalized AAL that models each specific user always achieves higher accuracy as compared to the traditional approach of modeling an average user. | ['Chia Y. Han', 'Nirmalya Thakur'] | 2022-07-19 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-1.17450975e-01 -8.61027539e-02 2.50735730e-01 -5.71829021e-01
-7.13005304e-01 -1.69977948e-01 3.40570182e-01 5.51212311e-01
-4.09889430e-01 1.13118374e+00 4.24910992e-01 -4.64185119e-01
-3.61246198e-01 -1.14087319e+00 -4.41820621e-01 -6.14444137e-01
-1.78989619e-01 3.81272703e-01 2.54229248e-01 -1.34069487e-01
-1.33093163e-01 2.96214759e-01 -1.89778566e+00 1.32768616e-01
1.13917387e+00 9.27806556e-01 4.77321535e-01 7.97419846e-01
1.85751885e-01 4.80492175e-01 -6.30623817e-01 3.25148970e-01
-1.98489875e-01 1.12184197e-01 -3.29236001e-01 -4.13670301e-01
2.52979308e-01 -1.40413627e-01 2.82579452e-01 3.94788653e-01
9.87909555e-01 3.78906518e-01 3.99916798e-01 -1.19886780e+00
-1.29502460e-01 1.71480492e-01 2.02474385e-01 2.84299850e-01
1.21470630e+00 -3.93255472e-01 4.81100857e-01 -4.87942576e-01
-2.13449910e-01 9.19643521e-01 1.06828237e+00 1.87494949e-01
-1.04203701e+00 -2.92564631e-01 3.83564651e-01 3.61824363e-01
-1.69417036e+00 -4.07190293e-01 5.53582489e-01 -2.93538272e-01
7.83821166e-01 8.45278561e-01 9.48863328e-01 9.92185116e-01
3.55537057e-01 4.07512873e-01 1.22976029e+00 -7.96456099e-01
1.02921009e+00 5.10110557e-01 6.02480114e-01 4.51613247e-01
3.04869950e-01 -3.35312039e-01 -3.50368470e-01 -5.18636584e-01
-1.34903550e-01 3.93707752e-01 -2.54584127e-03 -3.60956013e-01
-7.71360397e-01 3.21739435e-01 6.86698139e-01 9.27162826e-01
-6.54118776e-01 -4.37457860e-02 -1.53484374e-01 -5.61036825e-01
3.03299189e-01 -1.64497569e-01 -7.47253895e-01 -2.25671008e-01
-9.45389867e-01 3.72026145e-01 1.01303554e+00 9.23758984e-01
8.39442015e-01 -5.54821253e-01 -2.58583069e-01 5.82167864e-01
9.63231623e-01 7.48463154e-01 5.54942310e-01 -5.61873913e-01
1.82752877e-01 6.58865154e-01 5.53911567e-01 -1.06345427e+00
-7.90499628e-01 -5.72633207e-01 -5.04293978e-01 -1.25657201e-01
2.18728155e-01 -4.75757569e-01 -6.10691726e-01 1.56352615e+00
7.45828390e-01 -1.48536026e-01 -1.19036958e-01 -5.82888955e-03
6.34041190e-01 4.46931422e-01 3.95290971e-01 -4.14704122e-02
1.59427464e+00 -7.21736372e-01 -8.98741663e-01 -3.35013002e-01
5.49933255e-01 -2.21436739e-01 8.34337294e-01 1.40304983e-01
-4.85684663e-01 -7.45886028e-01 -1.17428946e+00 3.31238329e-01
-1.15694618e+00 1.63447633e-01 4.11125004e-01 1.59665132e+00
-1.31304598e+00 -5.96183948e-02 -8.28909039e-01 -1.05426729e+00
1.99750066e-01 5.61842263e-01 7.66356438e-02 -2.76096016e-02
-1.07931554e+00 8.10880303e-01 -2.97984071e-02 -2.44675018e-02
-2.73013264e-01 -7.44000852e-01 -7.30289340e-01 -1.46029577e-01
-2.91075557e-01 -8.60634506e-01 9.72196221e-01 -3.74312371e-01
-9.73394871e-01 2.03001678e-01 -7.39813387e-01 -3.94872963e-01
5.22739112e-01 -4.64086980e-01 -9.16635871e-01 -6.56558156e-01
7.15660632e-01 -2.49442533e-02 1.12648308e-01 -1.52204275e+00
-1.08176088e+00 -9.95163679e-01 1.07285567e-01 2.09424630e-01
-2.82500237e-01 -4.62637007e-01 1.09378681e-01 -3.50497067e-02
2.76360154e-01 -7.13045359e-01 -4.44968820e-01 -3.91608298e-01
-1.54514108e-02 -7.28307143e-02 8.05019021e-01 -9.05341566e-01
1.51376414e+00 -1.83443201e+00 -7.01032579e-01 6.23877108e-01
-2.71126837e-01 -1.92641333e-01 8.96156132e-01 4.57353443e-01
4.02626723e-01 -1.85398445e-01 5.01752570e-02 -5.73392510e-01
1.33111328e-01 1.89141721e-01 2.99312204e-01 2.70152271e-01
-1.02267730e+00 2.39727676e-01 -1.20053113e+00 -4.60542262e-01
9.05621767e-01 8.77360821e-01 -2.59661496e-01 1.72291636e-01
3.08898836e-01 4.79243696e-01 -5.70629954e-01 7.59532750e-01
7.10603952e-01 2.43938178e-01 3.36625695e-01 4.01123315e-02
-3.98998648e-01 1.86917447e-02 -1.38706696e+00 1.55464292e+00
-1.07275152e+00 3.82563807e-02 -5.39359972e-02 -3.79658401e-01
8.99235368e-01 3.56091708e-01 6.45434618e-01 -7.59554565e-01
9.14914161e-02 2.94574857e-01 -6.58397377e-01 -4.46500331e-01
2.96982408e-01 3.38765323e-01 -1.39802054e-01 1.78163752e-01
-3.92276347e-01 5.29326200e-01 -1.98404044e-01 -2.82101959e-01
1.47444308e+00 3.52072030e-01 6.89971209e-01 -6.07830167e-01
1.10811377e+00 -4.75499272e-01 2.95801193e-01 1.11392915e+00
-7.49798417e-01 9.14688557e-02 -7.71920085e-01 -5.33167481e-01
-2.03740925e-01 -1.25143099e+00 -3.75655115e-01 1.49845195e+00
1.11602224e-01 -5.75237691e-01 -9.99899566e-01 -6.25220358e-01
-1.02424368e-01 1.02461374e+00 -6.06829822e-01 2.55189508e-01
-1.23757429e-01 -9.45309103e-01 1.45433187e-01 2.05997407e-01
1.07409108e+00 -6.38825357e-01 -9.99622226e-01 3.05650860e-01
-5.76270878e-01 -8.06599438e-01 2.69847542e-01 4.10332561e-01
-6.46089673e-01 -7.30349958e-01 -5.73667526e-01 -6.79038703e-01
7.31776536e-01 2.80950397e-01 9.68751609e-01 -1.11755349e-01
-2.20245704e-01 1.14746010e+00 -4.28680658e-01 -6.49609625e-01
-5.77395316e-03 2.23549038e-01 2.88846940e-01 6.04679137e-02
7.42561579e-01 -1.02255845e+00 -9.00451064e-01 4.90738213e-01
5.31870350e-02 -3.81716520e-01 1.22237235e-01 2.34874394e-02
2.24810153e-01 4.12966996e-01 4.55154181e-01 -3.47059250e-01
3.80813897e-01 -9.24113572e-01 -9.79951099e-02 5.61345220e-01
-6.31046951e-01 -3.08804274e-01 4.38688189e-01 2.44150888e-02
-1.03789401e+00 3.58772814e-01 -5.60069680e-01 9.29007947e-01
-8.07662487e-01 -2.24891365e-01 -8.46855700e-01 -1.42557383e-01
9.12228823e-01 2.74709821e-01 -6.68701589e-01 -4.70599294e-01
1.61888629e-01 1.08272076e+00 2.37982884e-01 -6.78695023e-01
4.50063825e-01 5.64707220e-01 -2.69010276e-01 -1.01345789e+00
-5.31165600e-01 -8.06876063e-01 -9.32852149e-01 -7.26889491e-01
8.53115320e-01 -8.89780462e-01 -8.24956656e-01 3.45085591e-01
-8.09879780e-01 -2.01443568e-01 6.61755130e-02 1.17845759e-01
-4.65557009e-01 1.85092106e-01 2.87728637e-01 -1.62101221e+00
-3.32068413e-01 -8.13452363e-01 9.81656969e-01 4.74121690e-01
-6.79126203e-01 -1.27596128e+00 2.56770045e-01 7.10887551e-01
6.47220731e-01 3.70581597e-01 3.84955496e-01 -4.54822004e-01
-2.44448051e-01 -6.15699947e-01 2.96666741e-01 -6.03479631e-02
7.86360204e-01 -5.80461860e-01 -1.27606177e+00 -1.05207391e-01
3.09769493e-02 4.67822850e-01 -2.21017718e-01 5.06040692e-01
6.55061603e-01 -2.52724797e-01 -8.38993251e-01 3.59997973e-02
1.58060956e+00 5.73587358e-01 6.88218057e-01 7.79106915e-01
5.00274956e-01 2.96945989e-01 6.20663822e-01 6.67822719e-01
1.01622534e+00 8.25157940e-01 3.84267032e-01 -5.78536354e-02
2.24942476e-01 -2.12353498e-01 1.72863603e-01 -3.41629162e-02
-3.64919960e-01 -1.28827378e-01 -8.94466341e-01 4.74633366e-01
-2.05905485e+00 -8.89211416e-01 -1.80596188e-01 2.46329665e+00
5.52512780e-02 -4.78740409e-02 2.85718530e-01 7.30554223e-01
6.45318389e-01 3.59962620e-02 -3.91553938e-02 -2.14417279e-01
4.13127810e-01 1.22076094e-01 7.23412693e-01 7.63617575e-01
-1.32206011e+00 3.34824353e-01 5.90263987e+00 3.47895235e-01
-4.00958896e-01 5.08104324e-01 4.59998399e-01 4.57673758e-01
2.09763944e-01 -3.69763494e-01 -9.15152550e-01 8.49194407e-01
1.29408705e+00 4.91217613e-01 2.59854525e-01 1.28836560e+00
6.54020250e-01 -7.57735908e-01 -7.50026405e-01 9.51354563e-01
-1.45732472e-02 -8.31551731e-01 -5.66753149e-01 2.37734646e-01
6.70230627e-01 -2.62704581e-01 -2.88151771e-01 1.82895273e-01
2.00767264e-01 -4.28919584e-01 8.78505230e-01 9.01085436e-01
-1.25894725e-01 -6.85857713e-01 7.79875159e-01 5.12992799e-01
-1.56778252e+00 -4.49901491e-01 2.46468619e-01 -4.17337656e-01
2.77486686e-02 4.83056456e-01 -9.74045515e-01 6.32991612e-01
1.28540003e+00 1.51366159e-01 -8.41937304e-01 1.28356135e+00
-5.14460653e-02 2.38086656e-01 -6.77868247e-01 -5.09429455e-01
-1.66994527e-01 7.88628682e-02 1.80622146e-01 1.18972874e+00
6.93281114e-01 -3.28887254e-01 7.58712739e-02 4.46886599e-01
9.11826849e-01 2.29230464e-01 -6.99818134e-01 1.13186109e+00
7.03305662e-01 1.23492038e+00 -6.64660335e-01 -9.39473733e-02
-1.53302193e-01 9.49214578e-01 -2.51426727e-01 3.74267936e-01
-8.25124979e-01 -3.66478264e-02 3.42528552e-01 7.85327911e-01
1.32404556e-02 -3.21944058e-01 -6.26716793e-01 -3.86060804e-01
1.43977165e-01 -3.01963121e-01 2.86857843e-01 -6.97371066e-01
-7.31527865e-01 2.80646324e-01 -2.03069076e-02 -7.39973664e-01
8.83541827e-04 -3.36529613e-01 -5.79118192e-01 7.82513559e-01
-7.90538371e-01 -1.40031624e+00 -8.55799556e-01 7.56156445e-01
2.87988842e-01 1.10187307e-01 1.46623981e+00 7.82862544e-01
-3.37336212e-01 4.73659039e-01 3.00290674e-01 -4.95368570e-01
2.76867330e-01 -1.50099349e+00 -1.21527128e-01 8.46295536e-01
-3.91611576e-01 8.81394804e-01 8.84186387e-01 -6.40705705e-01
-9.25054193e-01 -1.10722661e+00 1.32232559e+00 -8.50730956e-01
4.88758646e-02 -5.85140765e-01 -3.82672101e-02 6.70144975e-01
-2.58754253e-01 -2.39836574e-01 1.23141682e+00 5.75513721e-01
3.86294335e-01 -6.92810476e-01 -1.86840820e+00 6.00252628e-01
9.80001807e-01 -2.59826422e-01 -4.41685259e-01 4.29513663e-01
1.42484814e-01 1.07430451e-01 -7.79446244e-01 4.02324557e-01
8.34178686e-01 -1.43399382e+00 1.24140346e+00 5.76729655e-01
-6.85861945e-01 -6.23582125e-01 -8.69925678e-01 -7.58699477e-01
-4.42439944e-01 -3.05019945e-01 -2.78960079e-01 1.36179185e+00
2.44241431e-01 -8.45520616e-01 7.55615115e-01 1.21293414e+00
-7.65589997e-02 -5.24921715e-01 -1.21360767e+00 -3.54173839e-01
-7.00918496e-01 -7.62147069e-01 1.01039898e+00 2.26304993e-01
-1.28507033e-01 -1.80919562e-02 -1.91676334e-01 5.43430984e-01
6.11125290e-01 -8.56048167e-01 9.06333148e-01 -1.33019900e+00
9.90949497e-02 1.10987559e-01 -7.69621611e-01 -6.93093181e-01
-4.56746131e-01 -2.80185014e-01 3.01063448e-01 -2.23380613e+00
-1.73657134e-01 -9.08185363e-01 -5.52218854e-01 2.17741072e-01
6.12015203e-02 -1.98319107e-02 -5.07851124e-01 -2.93762356e-01
-8.09662759e-01 1.81132823e-01 1.67326689e-01 -1.44597992e-01
-6.49828196e-01 8.77427518e-01 -6.74840629e-01 7.16458142e-01
6.69264436e-01 -3.75029147e-02 -4.93619233e-01 1.08886778e-01
7.47709870e-02 -4.69757169e-01 6.20496273e-01 -2.18300986e+00
3.78462404e-01 2.08189189e-01 8.88546884e-01 -8.94423544e-01
4.49367136e-01 -1.62625027e+00 4.31411833e-01 6.59220457e-01
3.92316490e-01 -1.63408473e-01 -1.67065226e-02 6.32237196e-01
5.49075484e-01 4.16710488e-02 4.88050580e-01 -1.82697564e-01
-7.72456646e-01 -1.36188924e-01 -7.86205530e-01 -8.57384562e-01
1.44751155e+00 -6.00762486e-01 1.10006683e-01 -3.15168917e-01
-1.07004309e+00 3.37824002e-02 3.97441983e-01 3.96651119e-01
2.08330363e-01 -1.32317817e+00 2.15237781e-01 1.43404946e-01
9.37005803e-02 -2.18911365e-01 4.19604719e-01 4.49594945e-01
-4.34584618e-01 5.04995346e-01 -1.16712257e-01 -5.04769623e-01
-1.51455855e+00 2.51886398e-01 6.87107563e-01 -2.07924992e-01
-3.39906693e-01 5.58130741e-01 -3.69748622e-01 -7.40860105e-01
6.66311443e-01 -3.43053788e-01 -3.94455910e-01 -5.84688261e-02
6.10700786e-01 1.03953552e+00 2.00229123e-01 -8.64908636e-01
-9.01023686e-01 4.22030717e-01 6.35601640e-01 -8.60515535e-02
9.55837071e-01 -7.17375278e-01 -4.28989232e-02 9.56422269e-01
6.91312015e-01 2.74238050e-01 -5.63307405e-01 1.02057688e-01
1.19663216e-01 -1.69641376e-01 -1.39001414e-01 -1.22583961e+00
8.32862128e-03 1.85364276e-01 1.70122910e+00 4.13333505e-01
9.89080608e-01 -2.22682536e-01 4.81200665e-01 2.86805004e-01
1.14938176e+00 -1.22367525e+00 -5.21729887e-01 1.92478985e-01
4.22908694e-01 -1.13514125e+00 9.92158800e-02 -1.31373286e-01
2.56149083e-01 7.61817098e-01 4.97920603e-01 3.02526921e-01
1.29728556e+00 7.92171359e-02 -5.24610616e-02 1.55237108e-01
2.79276609e-01 -2.94763505e-01 1.03781074e-01 1.23209012e+00
4.98027831e-01 5.32748163e-01 -3.44864100e-01 9.11703885e-01
-5.83016157e-01 1.13363780e-01 -1.84034839e-01 1.45680285e+00
-7.97871113e-01 -1.12047362e+00 -9.66957748e-01 2.78047413e-01
-1.81040302e-01 2.10899010e-01 1.44986855e-02 4.31502461e-01
1.15720499e+00 1.59890723e+00 -2.70051122e-01 -4.52277601e-01
3.72387052e-01 2.44228736e-01 3.28660190e-01 -3.47224444e-01
-4.93284345e-01 -5.00187099e-01 7.95964301e-02 -4.13982630e-01
-4.67056364e-01 -7.98647106e-01 -1.15138853e+00 -1.17733978e-01
1.65006548e-01 2.74375021e-01 1.24608970e+00 1.09082127e+00
2.81641036e-01 7.62381911e-01 5.94842136e-01 -1.08260167e+00
3.96378160e-01 -8.85103524e-01 -5.51234603e-01 -1.29283026e-01
3.81692976e-01 -7.27809429e-01 -2.52456874e-01 -2.96443850e-01] | [6.571868419647217, 0.9442386627197266] |
05675967-884d-4574-8a27-a117c8ab7225 | imitation-from-arbitrary-experience-a-dual | 2302.08560 | null | https://arxiv.org/abs/2302.08560v2 | https://arxiv.org/pdf/2302.08560v2.pdf | Dual RL: Unification and New Methods for Reinforcement and Imitation Learning | The goal of reinforcement learning (RL) is to maximize the expected cumulative return. It has been shown that this objective can be represented by an optimization problem of the state-action visitation distribution under linear constraints. The dual problem of this formulation, which we refer to as dual RL, is unconstrained and easier to optimize. We show that several state-of-the-art off-policy deep reinforcement learning (RL) algorithms, under both online and offline, RL and imitation learning (IL) settings, can be viewed as dual RL approaches in a unified framework. This unification provides a common ground to study and identify the components that contribute to the success of these methods and also reveals the common shortcomings across methods with new insights for improvement. Our analysis shows that prior off-policy imitation learning methods are based on an unrealistic coverage assumption and are minimizing a particular f-divergence between the visitation distributions of the learned policy and the expert policy. We propose a new method using a simple modification to the dual RL framework that allows for performant imitation learning with arbitrary off-policy data to obtain near-expert performance, without learning a discriminator. Further, by framing a recent SOTA offline RL method XQL in the dual RL framework, we propose alternative choices to replace the Gumbel regression loss, which achieve improved performance and resolve the training instability issue of XQL. Project code and details can be found at this https://hari-sikchi.github.io/dual-rl. | ['Amy Zhang', 'Qinqing Zheng', 'Scott Niekum', 'Harshit Sikchi'] | 2023-02-16 | null | null | null | null | ['offline-rl'] | ['playing-games'] | [-2.12999940e-01 1.72892541e-01 -7.61880338e-01 9.02838930e-02
-1.17769206e+00 -8.46679866e-01 6.18354082e-01 -3.93384784e-01
-5.92153370e-01 1.13747680e+00 1.17364405e-02 -7.31191039e-01
-4.23173755e-01 -4.28947479e-01 -1.02289081e+00 -8.20932329e-01
-1.82120651e-01 5.07193387e-01 -2.75493979e-01 -1.84684843e-01
1.96007848e-01 3.98686856e-01 -1.30088329e+00 -1.01157732e-01
7.61029959e-01 9.23070014e-01 7.02916756e-02 7.29793131e-01
2.13740796e-01 9.87045467e-01 -5.76030970e-01 -6.33003563e-02
5.74685097e-01 -5.25927663e-01 -7.83755064e-01 -1.29025415e-01
9.85882282e-02 -7.37395465e-01 -4.46493626e-01 7.28646576e-01
6.60574496e-01 2.23211065e-01 5.56861341e-01 -1.40826786e+00
-6.39004469e-01 3.53129447e-01 -5.20662606e-01 1.62476167e-01
4.02883470e-01 4.51093733e-01 1.14364326e+00 -2.41864666e-01
3.27523291e-01 1.32003498e+00 6.79276109e-01 6.79202616e-01
-1.41499090e+00 -4.98110265e-01 4.41463470e-01 -1.23733589e-02
-7.71161318e-01 -2.53583461e-01 2.20673785e-01 -1.63271517e-01
9.57163215e-01 -5.42305373e-02 6.63360894e-01 1.50022829e+00
1.91100895e-01 1.35321832e+00 1.59611905e+00 -4.19743776e-01
3.36629599e-01 1.13387063e-01 -5.20549715e-01 8.46398711e-01
-5.22164553e-02 7.42720723e-01 -1.33403718e-01 -3.36650550e-01
9.53671515e-01 1.00146197e-01 -1.63323030e-01 -5.81209362e-01
-8.74996006e-01 1.17449522e+00 1.49275094e-01 2.77222246e-02
-4.54386741e-01 6.81402743e-01 3.51139009e-01 7.02483535e-01
3.90474170e-01 4.16132808e-01 -6.11423671e-01 -4.61081386e-01
-5.97158194e-01 6.00002825e-01 1.03296673e+00 7.27480531e-01
5.63241899e-01 2.33057797e-01 -3.71781290e-01 5.58012187e-01
3.10416013e-01 6.79456770e-01 4.26772267e-01 -1.52647841e+00
3.83483648e-01 -5.53933084e-02 5.70449471e-01 -2.41612121e-01
-2.61523426e-01 -5.87173283e-01 -5.89601733e-02 6.50169194e-01
5.12614548e-01 -6.38734460e-01 -5.23272932e-01 1.97627473e+00
1.94802985e-01 1.33229092e-01 1.80955559e-01 5.41406751e-01
-7.55870417e-02 4.88571972e-01 -1.62712693e-01 -3.95132244e-01
7.28888571e-01 -1.03644967e+00 -4.96566415e-01 -9.17312130e-02
6.59243524e-01 -4.12972689e-01 1.31045318e+00 4.41514015e-01
-1.29334962e+00 -1.68309286e-01 -9.11127627e-01 3.79606754e-01
-1.45073310e-01 1.25893235e-01 7.05044925e-01 4.77507442e-01
-1.09772575e+00 1.14954448e+00 -9.13974762e-01 -7.33449012e-02
3.77822280e-01 4.10207838e-01 1.20453581e-01 2.05777511e-01
-1.20442796e+00 1.20590222e+00 8.38207155e-02 -3.47258002e-01
-1.56150627e+00 -7.18200862e-01 -5.84035158e-01 -1.00025102e-01
9.87249255e-01 -4.99596030e-01 2.07053852e+00 -1.26415634e+00
-2.02788067e+00 4.25777555e-01 2.68481910e-01 -5.67798495e-01
8.89778912e-01 -4.35086906e-01 3.91664878e-02 5.27761988e-02
2.06957608e-01 3.79651397e-01 1.05075097e+00 -1.26535809e+00
-8.27520669e-01 -2.27694526e-01 4.52860236e-01 4.28035498e-01
1.67480316e-02 -3.61659706e-01 -1.97075218e-01 -5.21163762e-01
-8.27147543e-01 -1.12713611e+00 -2.33205229e-01 -2.50115216e-01
4.92096879e-02 -5.20712972e-01 6.97694361e-01 -5.59184730e-01
1.14919639e+00 -1.88761544e+00 2.10175261e-01 3.09734102e-02
-9.39146802e-02 2.11204752e-01 -3.36137623e-01 6.55448675e-01
1.48968682e-01 -1.50396852e-02 -1.66104496e-01 -4.14733887e-01
4.64985996e-01 4.98552263e-01 -6.71759307e-01 8.44350755e-01
-1.49095222e-01 9.72736180e-01 -1.31344998e+00 -1.55092344e-01
2.07810968e-01 8.28919262e-02 -5.93049586e-01 3.90172273e-01
-5.28414547e-01 4.83989477e-01 -5.80905437e-01 5.39907575e-01
7.06940293e-02 -1.93981588e-01 2.76342779e-01 4.59470749e-01
-2.12270543e-01 2.02732325e-01 -9.71791923e-01 1.55411541e+00
-7.12448061e-01 2.71351010e-01 2.14437291e-01 -1.24288058e+00
4.27203506e-01 3.84093374e-01 8.85326922e-01 -7.40012944e-01
4.78436192e-03 2.57619053e-01 -2.94389933e-01 -3.62574607e-01
1.52940765e-01 -1.77622482e-01 -1.31365746e-01 8.53770792e-01
1.36025280e-01 -9.68030468e-02 5.42780533e-02 4.39915899e-03
9.70504820e-01 9.63044524e-01 3.80955040e-01 -1.15053847e-01
8.26343894e-02 7.59372115e-02 4.71593946e-01 1.21195364e+00
-4.50476676e-01 -1.19984336e-01 8.35210681e-01 -1.01367712e-01
-9.71986055e-01 -1.17763126e+00 9.38342046e-03 1.20436573e+00
-2.16476217e-01 -1.80015564e-02 -5.55443943e-01 -1.04734802e+00
5.94299138e-01 8.52611005e-01 -7.61618078e-01 -1.50837034e-01
-4.72376078e-01 -4.70727175e-01 7.85628498e-01 4.49200600e-01
2.96804368e-01 -1.00341976e+00 -7.66692221e-01 2.01287374e-01
1.80315264e-02 -6.97767615e-01 -6.03494763e-01 2.68748641e-01
-7.80076325e-01 -1.06316948e+00 -8.64175737e-01 -2.50781089e-01
2.22072557e-01 -5.82836270e-02 1.05543268e+00 -2.37328529e-01
-3.55531089e-02 1.20570898e+00 -2.89453387e-01 -4.14450049e-01
-5.39377153e-01 -6.05014861e-02 4.29802805e-01 -2.95695901e-01
6.85122162e-02 -4.90871996e-01 -7.79840648e-01 2.24344879e-01
-6.37527585e-01 -3.57817650e-01 5.38959265e-01 9.79559898e-01
5.62604666e-01 -5.43739736e-01 9.82573867e-01 -6.28404081e-01
1.02734637e+00 -6.46523178e-01 -9.95745420e-01 3.20888191e-01
-1.12507927e+00 4.57211047e-01 8.67552638e-01 -5.72356880e-01
-8.83106053e-01 -2.34929174e-01 -8.49693418e-02 -8.05514693e-01
1.17942408e-01 2.41265848e-01 4.75356698e-01 -1.60260901e-01
4.34274346e-01 3.79368484e-01 5.33706307e-01 -3.75657380e-01
5.15073478e-01 5.29846609e-01 6.19980469e-02 -9.74249780e-01
3.78866792e-01 3.92649382e-01 2.13031527e-02 -3.98005456e-01
-9.89379108e-01 -1.33181170e-01 -2.34641582e-02 -3.06648880e-01
4.39267814e-01 -8.44878435e-01 -1.34770226e+00 1.44861311e-01
-5.98691523e-01 -1.23551106e+00 -7.45625615e-01 5.31582773e-01
-1.46541893e+00 4.26390588e-01 -5.80593646e-01 -1.15252447e+00
-5.51442429e-02 -1.21651030e+00 9.13892925e-01 2.33857602e-01
2.01871738e-01 -1.36726034e+00 3.82510334e-01 -8.35964642e-03
3.86865199e-01 2.16878116e-01 6.38100564e-01 -4.99024928e-01
-3.23392779e-01 1.28160939e-01 1.40465736e-01 6.35962009e-01
-6.15496049e-03 -2.47414634e-01 -6.74400151e-01 -8.19546759e-01
-5.55079430e-03 -8.89098704e-01 7.60519266e-01 6.44476831e-01
1.13573599e+00 -5.51154494e-01 -1.22404441e-01 6.42641544e-01
1.58740449e+00 2.66095906e-01 2.46150732e-01 5.60633361e-01
2.66207844e-01 2.99406916e-01 8.39434624e-01 8.68360698e-01
2.30979487e-01 6.53837800e-01 4.31415200e-01 1.47528708e-01
2.58204788e-01 -6.86175704e-01 1.02109540e+00 2.44709358e-01
4.08633389e-02 -2.06931625e-02 -4.50109124e-01 3.71730685e-01
-2.16777849e+00 -1.17452693e+00 6.22945607e-01 2.32508016e+00
9.05716717e-01 -1.24223277e-01 5.94560206e-01 -5.42477012e-01
2.06600145e-01 1.44905090e-01 -1.20739245e+00 -6.33996964e-01
2.06137940e-01 2.75671422e-01 1.05114341e+00 7.03141212e-01
-8.89159679e-01 9.25302088e-01 7.36100006e+00 1.06768215e+00
-9.67452765e-01 3.28992099e-01 4.61383253e-01 -5.29892743e-01
-1.88078210e-01 -7.96308368e-03 -8.46622765e-01 3.66407752e-01
1.18001187e+00 -2.04985425e-01 1.17052865e+00 8.42788041e-01
4.96272117e-01 -1.17621697e-01 -1.11809897e+00 7.54493237e-01
-1.98643386e-01 -1.19345367e+00 -3.86485457e-01 3.51415455e-01
8.51279616e-01 3.31260771e-01 3.61438334e-01 1.03237009e+00
1.02554166e+00 -8.86207581e-01 6.73588216e-01 5.25024295e-01
8.39063168e-01 -8.86134624e-01 1.78475395e-01 5.10953426e-01
-8.39451730e-01 -5.52840650e-01 -9.53192636e-02 -3.79690938e-02
1.37085710e-02 -6.64944649e-02 -4.24675226e-01 4.24729645e-01
4.88519698e-01 7.79907763e-01 1.11449070e-01 5.42480946e-01
-4.44252491e-01 5.32902539e-01 -2.03712523e-01 -7.49862567e-02
8.03036749e-01 -4.22359467e-01 6.03468060e-01 8.99709761e-01
1.42556340e-01 -2.92093396e-01 5.87664783e-01 1.00380063e+00
4.69776541e-02 -2.18850732e-01 -8.19459140e-01 -2.20057651e-01
1.40220448e-01 9.02700543e-01 -1.91628560e-01 -1.00700557e-01
-4.01996076e-01 9.08728004e-01 6.76398098e-01 7.02961206e-01
-1.02967346e+00 -7.97055066e-02 8.48289311e-01 -2.94902116e-01
3.97236168e-01 -2.74858296e-01 1.52379543e-01 -9.35568392e-01
-1.17604792e-01 -1.10184264e+00 4.56073284e-01 -2.68566072e-01
-1.14631867e+00 -4.52031866e-02 3.13222229e-01 -1.14545333e+00
-7.12294221e-01 -7.21179664e-01 -4.66906399e-01 6.33363783e-01
-1.67732608e+00 -6.80520356e-01 4.53419358e-01 6.95993721e-01
4.96724218e-01 -2.03046650e-01 5.60618281e-01 9.22952965e-02
-5.91203213e-01 7.64587104e-01 8.85408223e-01 -3.66372555e-01
5.57883620e-01 -1.66606021e+00 -9.67295915e-02 3.50676894e-01
-2.09564790e-01 2.06463516e-01 6.34627759e-01 -4.65818524e-01
-1.68941748e+00 -8.76551390e-01 1.07443407e-02 -4.57177490e-01
9.08342123e-01 4.16557025e-03 -2.91302651e-01 1.06427956e+00
2.42197007e-01 -2.32808754e-01 4.54104573e-01 3.27814296e-02
6.41297027e-02 2.86838040e-02 -1.07233346e+00 7.33299434e-01
8.63559186e-01 -5.08195281e-01 -4.29925531e-01 4.82674032e-01
6.20822430e-01 -2.98627317e-01 -8.30587864e-01 1.74231261e-01
6.20540559e-01 -8.73733699e-01 9.10698354e-01 -9.98479724e-01
3.03480506e-01 1.22204468e-01 -4.56482135e-02 -1.70283604e+00
-8.38629529e-02 -1.13556182e+00 -5.99618673e-01 5.54433525e-01
1.63681090e-01 -1.08633149e+00 5.16267300e-01 2.04761699e-01
-8.40634704e-02 -1.27839398e+00 -1.03549898e+00 -1.28233922e+00
4.83663887e-01 -1.66746914e-01 1.18257314e-01 5.31490445e-01
6.65366501e-02 2.82559786e-02 -6.59101605e-01 -1.76514000e-01
7.15039551e-01 2.67722607e-01 5.91441631e-01 -6.00803256e-01
-8.12601566e-01 -7.05153227e-01 4.80425835e-01 -1.42210853e+00
6.39386475e-01 -8.92763257e-01 2.95349155e-02 -1.51210964e+00
-1.61020681e-01 -5.96095026e-01 -4.28060174e-01 5.55234313e-01
1.15326762e-01 -3.87715518e-01 2.39647463e-01 1.79212123e-01
-7.11481631e-01 9.34885681e-01 1.41674089e+00 1.78310752e-01
-3.04057926e-01 3.58887970e-01 -5.98894536e-01 5.37146330e-01
9.99796212e-01 -5.42493820e-01 -5.28662682e-01 -1.83329299e-01
1.22869574e-01 4.44613606e-01 3.19413513e-01 -5.15896678e-01
-1.93582535e-01 -5.75515509e-01 -2.69887932e-02 -2.01546073e-01
2.62399435e-01 -5.74578285e-01 -2.48633862e-01 8.55678678e-01
-5.49848616e-01 2.99836606e-01 2.85677701e-01 7.11455464e-01
2.46304378e-01 -3.74569297e-01 8.27558100e-01 -3.83641630e-01
-4.37901527e-01 4.01530653e-01 -4.55429584e-01 4.51431632e-01
1.17845488e+00 1.12811863e-01 -2.18568325e-01 -7.21909642e-01
-5.64039290e-01 6.95150018e-01 3.15095454e-01 2.62454569e-01
4.08496082e-01 -1.22050488e+00 -6.42442822e-01 -2.10357070e-01
-3.43872756e-01 -6.12162411e-01 4.80181910e-02 1.06531942e+00
-3.31992358e-02 5.15520215e-01 -1.43370137e-01 -1.28026649e-01
-6.87288165e-01 5.92530608e-01 7.58449972e-01 -7.60377109e-01
-5.78274071e-01 3.97703588e-01 -1.13854095e-01 -5.04711032e-01
4.97093230e-01 -1.06368721e-01 1.44796833e-01 -8.11497197e-02
2.80522019e-01 5.12397945e-01 -5.00807464e-01 -4.80418801e-02
1.28215943e-02 2.45361283e-01 7.69380257e-02 -4.85228598e-01
1.29315722e+00 -2.70305760e-02 3.35576802e-01 5.32945752e-01
1.07361114e+00 -2.05119506e-01 -1.92101586e+00 -1.05296776e-01
-1.67145357e-01 -3.57093185e-01 1.24365181e-01 -8.55793417e-01
-7.94853151e-01 6.12124085e-01 7.31691003e-01 1.82085708e-01
7.59834886e-01 -1.63921773e-01 5.76593518e-01 4.38782394e-01
4.14010257e-01 -1.49784577e+00 3.48741204e-01 6.50772393e-01
8.54928315e-01 -1.17874539e+00 3.41819897e-02 6.99686944e-01
-7.96999753e-01 9.58273590e-01 5.69878757e-01 -4.75102454e-01
4.76814747e-01 1.36926651e-01 -2.45089874e-01 1.26276091e-02
-9.49236274e-01 -4.87087309e-01 -2.40281999e-01 5.05199134e-01
9.02332962e-02 2.95077384e-01 -2.59522140e-01 2.43349165e-01
2.84215808e-01 1.78466901e-01 4.16387379e-01 1.04129922e+00
-3.35180730e-01 -1.33090198e+00 -9.42043886e-02 5.20037115e-01
-6.14482820e-01 2.50829279e-01 -1.00655213e-01 8.95675123e-01
-3.86644214e-01 7.82639027e-01 -1.24888562e-01 -1.65243924e-01
3.25281501e-01 -3.48613691e-03 8.22743654e-01 -3.40225428e-01
-6.30484581e-01 -2.62964126e-02 -3.46573368e-02 -9.10306752e-01
-2.91673869e-01 -6.85756266e-01 -1.11170578e+00 -4.11388695e-01
-4.57243174e-02 2.26020783e-01 5.12294769e-01 1.11147237e+00
3.40110004e-01 4.82448876e-01 9.38208401e-01 -9.44617987e-01
-1.78553164e+00 -6.53746784e-01 -6.80808067e-01 2.32580498e-01
7.06904173e-01 -9.69203353e-01 -5.37113607e-01 -7.25911498e-01] | [4.1233344078063965, 2.2865819931030273] |
637f8417-1752-4d5d-bf05-5aa0866b91df | leat-towards-robust-deepfake-disruption-in | 2307.01520 | null | https://arxiv.org/abs/2307.01520v1 | https://arxiv.org/pdf/2307.01520v1.pdf | LEAT: Towards Robust Deepfake Disruption in Real-World Scenarios via Latent Ensemble Attack | Deepfakes, malicious visual contents created by generative models, pose an increasingly harmful threat to society. To proactively mitigate deepfake damages, recent studies have employed adversarial perturbation to disrupt deepfake model outputs. However, previous approaches primarily focus on generating distorted outputs based on only predetermined target attributes, leading to a lack of robustness in real-world scenarios where target attributes are unknown. Additionally, the transferability of perturbations between two prominent generative models, Generative Adversarial Networks (GANs) and Diffusion Models, remains unexplored. In this paper, we emphasize the importance of target attribute-transferability and model-transferability for achieving robust deepfake disruption. To address this challenge, we propose a simple yet effective disruption method called Latent Ensemble ATtack (LEAT), which attacks the independent latent encoding process. By disrupting the latent encoding process, it generates distorted output images in subsequent generation processes, regardless of the given target attributes. This target attribute-agnostic attack ensures robust disruption even when the target attributes are unknown. Additionally, we introduce a Normalized Gradient Ensemble strategy that effectively aggregates gradients for iterative gradient attacks, enabling simultaneous attacks on various types of deepfake models, involving both GAN-based and Diffusion-based models. Moreover, we demonstrate the insufficiency of evaluating disruption quality solely based on pixel-level differences. As a result, we propose an alternative protocol for comprehensively evaluating the success of defense. Extensive experiments confirm the efficacy of our method in disrupting deepfakes in real-world scenarios, reporting a higher defense success rate compared to previous methods. | ['Hyunsoo Yoon', 'Joonkyo Shim'] | 2023-07-04 | null | null | null | null | ['face-swapping'] | ['computer-vision'] | [ 2.20997393e-01 -1.44501448e-01 2.63440818e-01 2.72337496e-01
-7.39731252e-01 -1.21203458e+00 8.50046992e-01 -4.17200625e-01
-2.34601740e-02 7.22217143e-01 1.86251715e-01 -2.45374799e-01
-7.70039810e-03 -1.10520482e+00 -7.82316267e-01 -1.04625893e+00
1.07634321e-01 -2.29391471e-01 -6.27964959e-02 -2.00069591e-01
1.36378795e-01 7.24994957e-01 -1.10281742e+00 1.96132902e-03
1.21413422e+00 7.44946539e-01 -4.22540218e-01 7.70310342e-01
2.37017706e-01 7.43108511e-01 -1.29260993e+00 -8.85731995e-01
4.91030484e-01 -5.52767515e-01 -3.40311080e-01 -1.66386604e-01
2.77896672e-01 -7.43269324e-01 -6.25606418e-01 1.21499360e+00
8.40650678e-01 -2.24319577e-01 7.74409354e-01 -1.58626890e+00
-9.64667797e-01 3.75616223e-01 -3.67264777e-01 8.78263935e-02
1.18633598e-01 6.94811225e-01 5.30879200e-01 -3.85899723e-01
1.73153996e-01 1.25188899e+00 5.33329308e-01 8.71524274e-01
-1.35735929e+00 -1.01991355e+00 1.81105897e-01 1.58815861e-01
-1.24290550e+00 -3.85392159e-01 1.04208052e+00 -2.90786296e-01
6.34323418e-01 4.64808702e-01 3.89600724e-01 1.66511202e+00
4.79702532e-01 5.43147981e-01 1.31012905e+00 -7.68271759e-02
4.14224148e-01 -7.11185560e-02 -4.03563678e-01 3.39761198e-01
2.83475608e-01 6.35917425e-01 -2.62383699e-01 -4.80554491e-01
8.16820800e-01 -1.74578905e-01 -5.43937445e-01 -1.54122919e-01
-8.63841355e-01 8.22257519e-01 4.07879621e-01 -2.05001025e-03
-4.73713309e-01 4.74128515e-01 2.83922553e-01 1.66123778e-01
3.25852036e-01 4.08658385e-01 1.19362272e-01 -3.12012695e-02
-6.10193312e-01 8.53760540e-02 6.36344969e-01 5.77012718e-01
5.64217329e-01 5.97565353e-01 -4.30731803e-01 4.30541694e-01
8.88563544e-02 8.87010276e-01 2.70912528e-01 -7.65483797e-01
3.58850837e-01 3.72377694e-01 -7.27751479e-02 -1.41397154e+00
7.30663836e-02 -6.87287390e-01 -1.06943524e+00 5.88224947e-01
2.80281365e-01 -4.45976675e-01 -1.06733251e+00 2.07753849e+00
4.31187928e-01 4.98172402e-01 2.64388978e-01 6.07330501e-01
4.59866434e-01 5.20675361e-01 3.83540504e-02 6.00302741e-02
9.09198940e-01 -4.85990405e-01 -6.31010532e-01 1.51039243e-01
2.52751768e-01 -6.53277755e-01 1.04932463e+00 4.30300981e-01
-8.79853070e-01 -3.28519911e-01 -1.24005938e+00 5.59043586e-01
-3.32880706e-01 -2.61256665e-01 2.26900652e-01 1.32892358e+00
-9.34617281e-01 3.85594994e-01 -6.52142465e-01 3.89292128e-02
4.78144318e-01 3.41368198e-01 -1.25995606e-01 7.62311667e-02
-1.44045842e+00 7.46420801e-01 -2.73744483e-02 1.66565433e-01
-1.64329433e+00 -7.56296575e-01 -6.22876167e-01 8.25414211e-02
2.28923336e-01 -8.03437650e-01 8.32769334e-01 -7.96422243e-01
-1.73036671e+00 3.91717762e-01 2.88624555e-01 -5.16901612e-01
9.51819122e-01 -1.04326703e-01 -5.68733096e-01 1.63363561e-01
-3.10348600e-01 2.49494553e-01 1.23924112e+00 -1.65520775e+00
-1.08835518e-01 -2.00715005e-01 4.26584035e-01 7.19877034e-02
-8.18973184e-01 -1.27374664e-01 7.83201307e-02 -9.75461006e-01
-4.99663055e-01 -9.27027285e-01 -1.29900798e-01 -3.28978151e-01
-9.25142765e-01 3.21801305e-01 1.06053650e+00 -5.61056077e-01
1.19939089e+00 -2.21185899e+00 -1.61535461e-02 3.05604219e-01
4.73171204e-01 6.27063274e-01 -3.22048306e-01 6.39494956e-01
1.83160707e-01 5.18887699e-01 -3.68425459e-01 -9.54923332e-02
1.39550164e-01 1.10165812e-01 -9.63582873e-01 4.29182321e-01
1.27404273e-01 1.06942654e+00 -7.12723970e-01 1.48136005e-01
1.26526386e-01 6.52261913e-01 -3.58787715e-01 2.40150481e-01
4.74791378e-02 4.39017892e-01 -4.36567247e-01 6.80030525e-01
9.23929214e-01 2.69115478e-01 -2.00034469e-01 -2.73611277e-01
2.18454883e-01 -1.97749138e-01 -8.38571846e-01 9.18398380e-01
-3.72100651e-01 3.88970524e-01 -4.42484245e-02 -5.42630315e-01
8.79738212e-01 2.37610683e-01 1.20617397e-01 -5.77968061e-01
1.54660016e-01 1.87484145e-01 1.31499469e-01 -8.15183446e-02
2.49277368e-01 -5.70070483e-02 -2.98763663e-01 5.42259991e-01
-2.19322056e-01 -5.18466569e-02 -3.41320723e-01 3.85277659e-01
1.42472768e+00 -1.85326874e-01 -3.28664407e-02 -2.83499639e-02
3.74815166e-01 -3.36820394e-01 5.50027192e-01 8.88610721e-01
-3.98071826e-01 5.93615890e-01 6.05903506e-01 -1.11173794e-01
-1.00858974e+00 -1.42381442e+00 3.05785745e-01 4.73874390e-01
4.31686789e-01 -2.97169358e-01 -1.09513700e+00 -1.09688008e+00
1.08037375e-01 7.61836469e-01 -8.08131933e-01 -7.73761034e-01
-4.32569236e-01 -9.76476848e-01 1.41734743e+00 4.42593992e-01
9.33950841e-01 -7.91275740e-01 -3.04367810e-01 1.49935126e-01
-2.57214475e-02 -9.56402421e-01 -4.99410063e-01 -3.46069872e-01
-3.87258112e-01 -1.02286625e+00 -6.96545005e-01 -1.17971212e-01
6.27934992e-01 8.13158229e-02 6.91655397e-01 8.70921612e-02
-1.09482341e-01 6.03040278e-01 -2.55186409e-01 -3.68631363e-01
-7.82479167e-01 -2.30725303e-01 1.34458050e-01 3.66866171e-01
-6.48692548e-02 -8.41957629e-01 -6.97111487e-01 4.91993666e-01
-1.23220658e+00 -3.32818538e-01 6.31729722e-01 6.47372067e-01
2.35001266e-01 2.86308616e-01 8.89736533e-01 -6.68486476e-01
1.05893373e+00 -3.37858260e-01 -3.93849641e-01 2.82290548e-01
-4.52189296e-01 -2.33937070e-01 9.52899516e-01 -8.10938478e-01
-9.93263304e-01 -4.27570134e-01 -1.82961524e-01 -5.71797788e-01
-2.26750702e-01 2.73950875e-01 -7.13306606e-01 -3.71672004e-01
7.81491756e-01 6.01102829e-01 -1.20096356e-01 -1.42514467e-01
4.84640896e-01 3.89793277e-01 6.95228100e-01 -5.86946964e-01
1.43902719e+00 6.66833043e-01 -3.45515311e-02 -4.20945704e-01
-1.29733503e-01 3.77172947e-01 1.48635497e-02 -4.20915008e-01
6.87853575e-01 -7.39927948e-01 -6.68856263e-01 1.29790509e+00
-1.16193593e+00 -4.45812047e-01 -1.69116128e-02 1.27853483e-01
-4.01562989e-01 5.47624469e-01 -6.01336539e-01 -7.07332671e-01
-4.72142518e-01 -1.12739968e+00 6.52396500e-01 1.66153580e-01
9.96501669e-02 -8.84096920e-01 1.07999325e-01 1.82300985e-01
6.47262573e-01 7.88824141e-01 9.20174778e-01 -6.94860339e-01
-7.23202169e-01 -2.35486120e-01 5.82699329e-02 6.58444583e-01
3.96664917e-01 1.21398792e-01 -1.03009725e+00 -5.70366621e-01
1.06451578e-01 -2.34122291e-01 6.81672812e-01 3.81361321e-02
1.08663380e+00 -6.71719670e-01 -6.65853620e-02 8.07006121e-01
1.31629264e+00 3.87150824e-01 1.10766733e+00 3.26459438e-01
8.62983763e-01 3.22077513e-01 6.10656440e-02 4.56883997e-01
9.13681760e-02 5.83170772e-01 8.04697156e-01 -2.36658856e-01
-2.58173615e-01 -3.34134430e-01 6.75264955e-01 3.13054949e-01
1.83682218e-02 -7.29475439e-01 -6.46123767e-01 4.13523585e-01
-1.44125533e+00 -1.10726559e+00 6.86575472e-02 2.25581384e+00
6.33935094e-01 2.14898288e-01 3.63770276e-02 1.32148936e-01
9.12082613e-01 1.94261327e-01 -9.10774887e-01 -3.93099666e-01
-3.57728183e-01 5.28832972e-02 5.63214898e-01 2.71079421e-01
-9.45900500e-01 8.18542778e-01 6.01007414e+00 1.12351322e+00
-1.32119763e+00 1.10595942e-01 6.97752535e-01 -1.04558468e-01
-7.02700853e-01 -1.54803157e-01 -6.11004472e-01 8.65590692e-01
6.59582019e-01 -4.49633300e-01 4.21275824e-01 5.24201035e-01
2.90621877e-01 3.80255163e-01 -7.61461139e-01 6.51256502e-01
1.73431993e-01 -1.20017505e+00 3.25808316e-01 5.76779366e-01
8.04018199e-01 -4.05490428e-01 5.99293947e-01 2.26896793e-01
6.76322937e-01 -1.04648530e+00 5.86368442e-01 3.71314317e-01
7.83389986e-01 -1.23027790e+00 4.50657785e-01 1.95252180e-01
-7.93559372e-01 -1.12060316e-01 -7.90011585e-02 1.71237692e-01
2.01793760e-01 6.88177526e-01 -5.40805936e-01 6.17745221e-01
5.19529462e-01 6.34388030e-02 -4.49887633e-01 8.03237319e-01
-6.84443533e-01 8.67734492e-01 -1.17837012e-01 3.47386509e-01
1.84277579e-01 1.34760961e-01 8.69174540e-01 9.73849535e-01
4.46739793e-01 -2.39068922e-02 -8.85319039e-02 9.89926517e-01
-1.54666886e-01 -2.87710726e-01 -8.25887501e-01 4.29820009e-02
7.28462100e-01 1.14107502e+00 -4.97144550e-01 -1.17000639e-01
1.47058219e-01 1.17111123e+00 -9.76071600e-03 6.60430312e-01
-1.34930384e+00 -4.70413536e-01 1.11106873e+00 -1.35342807e-01
2.52261132e-01 -2.52034783e-01 -3.34741384e-01 -1.06945789e+00
2.30181813e-02 -1.23451114e+00 1.86909512e-01 -4.28958654e-01
-1.50002325e+00 6.45007730e-01 1.13187106e-02 -1.32987678e+00
-1.55364826e-01 -2.72666514e-01 -1.03734791e+00 9.33070600e-01
-1.39624560e+00 -1.35234773e+00 -2.91444659e-01 8.91970634e-01
1.56907097e-01 -2.75145292e-01 6.36275351e-01 7.77399614e-02
-7.36244023e-01 1.12508261e+00 1.87583074e-01 1.96415171e-01
6.07869506e-01 -1.00640428e+00 5.74083507e-01 1.48405194e+00
-8.26159567e-02 6.35257006e-01 6.32594049e-01 -7.89499521e-01
-1.22651207e+00 -1.32711148e+00 5.00140041e-02 -3.92705560e-01
5.58801234e-01 -5.52829087e-01 -8.77807736e-01 3.68480086e-01
3.12729359e-01 -2.21802533e-01 7.91037738e-01 -4.94331062e-01
-5.47935009e-01 -1.53320029e-01 -1.49169445e+00 8.36107016e-01
8.60833287e-01 -6.35744870e-01 -1.24871239e-01 -9.91680399e-02
7.42733538e-01 -2.33789444e-01 -6.72139049e-01 6.34986103e-01
3.11865568e-01 -1.05999184e+00 1.10487998e+00 -4.10548776e-01
4.59341049e-01 -4.72812116e-01 -1.90261915e-01 -1.70564115e+00
-3.33660275e-01 -8.02587986e-01 -4.26352322e-01 1.71067274e+00
4.33789976e-02 -1.03350425e+00 5.92709601e-01 4.26149189e-01
-8.69794264e-02 -5.28072119e-01 -8.94589007e-01 -1.14538109e+00
3.24159205e-01 -2.67703474e-01 9.99958098e-01 9.32519734e-01
-5.56672573e-01 -2.30313495e-01 -8.08178067e-01 6.87279940e-01
7.76755333e-01 -2.00115576e-01 1.06995749e+00 -6.08885765e-01
-3.76064956e-01 -5.81345677e-01 -6.54935300e-01 -5.83366394e-01
1.24238603e-01 -4.99743551e-01 -1.34955198e-01 -1.23405695e+00
1.63001027e-02 -3.18938166e-01 -3.63201946e-01 5.26494861e-01
-4.45914954e-01 6.36176348e-01 3.54607552e-01 3.17565978e-01
4.95678969e-02 8.22732151e-01 1.07105458e+00 -3.34576160e-01
3.33604068e-02 -1.69880003e-01 -8.89345407e-01 5.36508262e-01
9.36883509e-01 -5.21355450e-01 -7.30797887e-01 -4.49245065e-01
-1.21648647e-01 -3.97795141e-01 8.23700309e-01 -1.15890133e+00
5.18703684e-02 -3.87696117e-01 3.50591660e-01 -1.21259324e-01
1.56108603e-01 -5.49523294e-01 4.42370683e-01 6.47341847e-01
-8.10389966e-02 1.24590844e-02 2.85655349e-01 7.11993098e-01
-4.87782210e-02 1.66356221e-01 7.39684999e-01 1.75951943e-01
-5.38365304e-01 4.06584978e-01 -5.62427759e-01 -1.23711415e-01
1.40225923e+00 -3.12981695e-01 -7.19484150e-01 -6.00241363e-01
-4.38502461e-01 -7.94437751e-02 7.12037623e-01 2.81395316e-01
7.55846560e-01 -1.47816885e+00 -9.12944019e-01 2.62019992e-01
-2.06275389e-01 -4.30530727e-01 6.34632409e-01 4.80071932e-01
-3.12756419e-01 -1.68988749e-01 -3.05035144e-01 -3.20869178e-01
-1.14057803e+00 7.70286322e-01 4.97307390e-01 -2.11964309e-01
-3.57232302e-01 7.31535554e-01 7.86303639e-01 -2.26861209e-01
1.78940538e-02 2.78871894e-01 -1.16424793e-02 -2.33643010e-01
5.46370447e-01 4.53990370e-01 6.98734596e-02 -5.33012629e-01
-3.30613494e-01 2.47910723e-01 -5.41749150e-02 -2.19734330e-02
8.98298502e-01 -1.32353261e-01 9.47881415e-02 -2.90318400e-01
9.13114905e-01 5.12159288e-01 -1.59131885e+00 1.04224458e-01
-6.39393389e-01 -6.51990950e-01 -6.68532476e-02 -1.15239942e+00
-1.23432899e+00 7.51172543e-01 5.63245118e-01 3.63925636e-01
1.48110235e+00 -3.78641009e-01 9.71573710e-01 -3.57315391e-02
3.11404705e-01 -6.14284396e-01 3.82723749e-01 1.65278360e-01
8.72549951e-01 -6.39701009e-01 -3.51710975e-01 -3.39041233e-01
-4.90572959e-01 7.11174786e-01 8.10264111e-01 -1.73516527e-01
2.32706115e-01 4.34316248e-01 3.04513305e-01 1.11769944e-01
-5.14136672e-01 1.50977179e-01 1.12234458e-01 9.77090061e-01
-2.72738248e-01 7.84593448e-02 -9.63819772e-02 5.02313673e-01
-2.10131049e-01 -4.83932644e-01 8.21251154e-01 7.18832970e-01
-5.24115842e-03 -1.26796150e+00 -7.49073625e-01 1.81872010e-01
-5.92884839e-01 -1.68526739e-01 -6.83251441e-01 5.59408009e-01
1.21258803e-01 1.20522642e+00 -3.49818319e-01 -7.49064147e-01
2.82456130e-01 -3.27568352e-01 3.44884098e-01 -7.97322094e-02
-6.44359291e-01 -6.94169253e-02 -1.90382421e-01 -4.80396658e-01
-8.27093981e-03 -4.00956184e-01 -8.14006627e-01 -7.84210682e-01
-2.55584478e-01 5.35519347e-02 4.99438941e-01 8.30098987e-01
6.34519637e-01 5.19872725e-01 1.13760424e+00 -7.12582231e-01
-9.45865095e-01 -5.50771236e-01 -4.84813124e-01 4.22395229e-01
3.27011049e-01 -6.26348555e-01 -6.15912080e-01 -9.10217240e-02] | [5.5091352462768555, 7.899404048919678] |
d57624de-f8da-4efc-811a-c05eef17b349 | a-benchmark-study-on-time-series-clustering | 2004.09546 | null | https://arxiv.org/abs/2004.09546v2 | https://arxiv.org/pdf/2004.09546v2.pdf | A Benchmark Study on Time Series Clustering | This paper presents the first time series clustering benchmark utilizing all time series datasets currently available in the University of California Riverside (UCR) archive -- the state of the art repository of time series data. Specifically, the benchmark examines eight popular clustering methods representing three categories of clustering algorithms (partitional, hierarchical and density-based) and three types of distance measures (Euclidean, dynamic time warping, and shape-based). We lay out six restrictions with special attention to making the benchmark as unbiased as possible. A phased evaluation approach was then designed for summarizing dataset-level assessment metrics and discussing the results. The benchmark study presented can be a useful reference for the research community on its own; and the dataset-level assessment metrics reported may be used for designing evaluation frameworks to answer different research questions. | ['Dona M. Rizzo', 'Byung Suk Lee', 'Ali Javed'] | 2020-04-20 | null | null | null | null | ['time-series-clustering'] | ['time-series'] | [-3.03171158e-01 -6.61172688e-01 -2.71822214e-01 -2.25714520e-01
-6.26243651e-01 -8.74559999e-01 6.41128123e-01 6.39257073e-01
-4.09704953e-01 1.91472054e-01 3.15299064e-01 -4.71598566e-01
-9.03125763e-01 -6.43912852e-01 2.76796579e-01 -6.99568510e-01
-9.39883530e-01 3.72218549e-01 1.01047188e-01 -1.78530261e-01
5.71953058e-01 6.86488152e-01 -1.91350985e+00 -4.34002690e-02
6.47544980e-01 1.02658141e+00 -2.31996179e-01 3.48032206e-01
1.24283753e-01 4.08036411e-01 -9.45856571e-01 2.40267023e-01
3.97994727e-01 -3.49294931e-01 -6.15731001e-01 -6.13954104e-02
-4.60442513e-01 2.23054424e-01 -2.05947742e-01 7.05897033e-01
4.66850013e-01 5.43987036e-01 6.13669157e-01 -1.59443188e+00
-5.32615960e-01 5.36808908e-01 -3.93156826e-01 7.96949804e-01
3.93211752e-01 -4.22300287e-02 8.49782348e-01 -2.24651784e-01
7.74536788e-01 8.63399684e-01 7.55158424e-01 2.01536104e-01
-1.23055279e+00 -5.51784933e-01 -7.70005286e-02 4.35822755e-01
-1.44092298e+00 -1.18554048e-01 9.81263399e-01 -7.40624428e-01
1.09208965e+00 5.89702308e-01 7.30296791e-01 7.90422022e-01
2.13495418e-02 2.94327140e-01 1.37608397e+00 -2.37782970e-01
6.65423870e-01 -3.40622425e-01 4.67086315e-01 -3.00398946e-01
3.05903167e-01 3.15806605e-02 1.73978180e-01 -5.16844571e-01
5.99966586e-01 -6.54726997e-02 -1.66235045e-01 -4.05951291e-01
-1.33286726e+00 6.28608942e-01 5.55041246e-03 9.73041773e-01
-4.98012811e-01 -3.68953854e-01 9.13566530e-01 6.54126704e-01
7.17554688e-01 5.11858046e-01 -5.80454469e-01 -7.66029537e-01
-1.15438390e+00 2.36759111e-01 7.87563503e-01 5.33229172e-01
2.05567390e-01 2.57015109e-01 1.50972558e-02 7.78916359e-01
1.63301647e-01 2.10227922e-01 9.36523080e-01 -1.04504347e+00
3.94253790e-01 5.68883717e-01 7.71886408e-02 -1.18563259e+00
-6.22396648e-01 -2.77325977e-02 -6.66193485e-01 -2.40311652e-01
2.00585589e-01 -2.59046823e-01 -5.13391972e-01 1.37026370e+00
5.65321855e-02 1.94058463e-01 -1.55382559e-01 5.80869615e-01
5.82787573e-01 7.61419296e-01 -6.59428760e-02 -9.25088346e-01
9.57893610e-01 -3.01317066e-01 -8.37897658e-01 5.55312693e-01
7.08418906e-01 -7.57715166e-01 7.53441215e-01 3.34200650e-01
-8.00707281e-01 -5.00625610e-01 -8.63175392e-01 6.33709311e-01
-7.27662683e-01 -3.99982750e-01 3.16486657e-01 7.49626040e-01
-1.33453822e+00 7.69164801e-01 -1.26456642e+00 -8.90256941e-01
-2.79408634e-01 1.18554942e-01 -2.49688607e-02 3.80655199e-01
-9.12741423e-01 8.15345049e-01 2.76475042e-01 -3.12795848e-01
-4.82410312e-01 -6.94348395e-01 -6.48694217e-01 -1.77902415e-01
-2.52660692e-01 9.94863361e-02 9.23232198e-01 -5.55036902e-01
-1.01454842e+00 6.46380186e-01 7.16512799e-02 -4.32411611e-01
3.46737266e-01 2.93667644e-01 -1.24441409e+00 2.85660625e-01
2.34898105e-01 1.17763974e-01 3.04797113e-01 -1.16488469e+00
-4.15230274e-01 -5.65997422e-01 -5.40469050e-01 -1.52139708e-01
-5.31458199e-01 3.37259918e-01 -9.30265188e-02 -1.10797238e+00
1.41188875e-01 -9.26089525e-01 -2.68631786e-01 -7.46157348e-01
1.05517194e-01 -4.31168914e-01 1.18356514e+00 -5.34222007e-01
1.91888559e+00 -2.29482603e+00 -4.39644046e-02 4.45986986e-01
-2.43030667e-01 6.31189123e-02 -1.82280764e-01 1.11023605e+00
-6.41967893e-01 3.78222406e-01 -3.32131207e-01 -2.26437479e-01
-4.80943620e-02 4.46223579e-02 -4.55639511e-01 7.22397327e-01
-1.44357279e-01 3.36499512e-01 -9.66931462e-01 -2.02507317e-01
6.67066038e-01 3.52674901e-01 -5.81620336e-02 -1.73387453e-01
2.58850604e-01 4.41566408e-01 -2.01907054e-01 5.04526317e-01
3.73475015e-01 1.91705808e-01 8.45690817e-02 -6.84407726e-02
-8.21692169e-01 7.09783509e-02 -1.05929589e+00 1.29066455e+00
3.90095152e-02 8.52430463e-01 -4.62708980e-01 -1.00883770e+00
1.16284573e+00 4.97526050e-01 1.41555488e+00 -5.63548386e-01
1.84877828e-01 1.41875699e-01 1.77881584e-01 -4.12977815e-01
4.28293556e-01 3.23720604e-01 -2.54312575e-01 6.63300514e-01
-2.08339244e-01 -2.51358211e-01 6.86966062e-01 -1.71006039e-01
1.22535765e+00 -1.04909882e-01 1.25002801e-01 -6.90442324e-01
3.58063668e-01 3.90594423e-01 4.83156353e-01 1.63073048e-01
-7.56504476e-01 5.29414892e-01 3.24877888e-01 -4.53645110e-01
-1.24882054e+00 -1.00015283e+00 -3.58408868e-01 7.25795627e-01
-1.83305889e-01 -7.03975260e-01 -5.56507230e-01 2.17958409e-02
-8.38687569e-02 8.25349092e-01 -9.00997996e-01 8.09598491e-02
-2.96182454e-01 -1.00851727e+00 5.09651721e-01 3.53666395e-01
2.61389524e-01 -1.05669343e+00 -8.49343956e-01 9.99059975e-02
-2.60685116e-01 -6.34850562e-01 -2.81449497e-01 1.27311140e-01
-1.32652760e+00 -1.10099196e+00 -4.93935466e-01 -5.50732553e-01
3.01028728e-01 4.77309138e-01 9.76823330e-01 -2.57371873e-01
-3.75051111e-01 9.34414089e-01 -8.79228294e-01 -3.13126981e-01
-1.03224747e-01 -4.81202789e-02 1.19275019e-01 -8.92011374e-02
8.71460497e-01 -8.17191839e-01 -7.22457349e-01 5.95504999e-01
-9.33732927e-01 -9.10621881e-01 -2.68106580e-01 1.35313660e-01
5.01583815e-01 5.51364183e-01 6.04854465e-01 -9.97665972e-02
1.11293995e+00 -8.24310303e-01 -5.45585036e-01 1.29885271e-01
-8.10810328e-01 -5.47123969e-01 5.63115478e-01 -4.12983328e-01
-5.12268126e-01 -3.29800010e-01 3.58743221e-01 -6.62710309e-01
-3.08164954e-01 7.50029981e-01 5.00109136e-01 2.41658807e-01
6.51486754e-01 3.61365020e-01 6.03344813e-02 -4.40296233e-01
1.16380610e-01 8.12748432e-01 4.89533544e-01 -7.28877068e-01
5.94705701e-01 7.18701601e-01 -2.41651848e-01 -1.10317290e+00
-7.37341642e-02 -1.09518957e+00 -9.43922520e-01 -6.50251150e-01
7.65082121e-01 -6.55276477e-01 -3.30986828e-01 4.23832536e-01
-3.98808450e-01 -3.33451003e-01 -4.43214804e-01 5.20437419e-01
-6.21920109e-01 3.41748744e-01 -5.02726376e-01 -8.56177747e-01
-8.92420933e-02 -9.04424369e-01 7.22982705e-01 -8.20148364e-02
-7.20941842e-01 -1.40545094e+00 6.35294199e-01 -4.12414335e-02
4.86988693e-01 9.92246091e-01 7.60843992e-01 -9.19884443e-01
4.04887021e-01 -3.60904068e-01 3.25161636e-01 9.17782560e-02
3.11598480e-01 8.14479053e-01 -6.85219407e-01 -4.88889545e-01
2.61283308e-01 2.35815972e-01 2.26385087e-01 6.22356236e-01
1.17287385e+00 7.14435428e-02 -2.74475038e-01 7.48903379e-02
1.52861238e+00 8.37201059e-01 5.99553287e-01 6.89157844e-01
1.30264208e-01 7.47112811e-01 7.75705516e-01 9.33834732e-01
4.55358505e-01 4.95711535e-01 -2.05581803e-02 1.93267569e-01
3.30532461e-01 3.91380429e-01 1.52495816e-01 1.39167213e+00
-4.48746383e-01 -1.23352826e-01 -1.46727884e+00 1.05828488e+00
-1.96954572e+00 -1.43924320e+00 -4.42306906e-01 2.37513924e+00
3.34084451e-01 -1.91408813e-01 7.73575366e-01 7.64469743e-01
8.98635209e-01 2.21249163e-01 -1.90629035e-01 -3.37752879e-01
5.08520082e-02 5.73806688e-02 2.58619428e-01 -1.74966112e-01
-1.32572365e+00 1.22197926e-01 7.35508251e+00 4.12712812e-01
-1.10120082e+00 -2.92133093e-02 3.48436803e-01 -5.33686802e-02
-6.13257810e-02 -3.04007828e-02 1.01462379e-01 7.45203912e-01
1.58663356e+00 -8.98861766e-01 2.81336486e-01 5.47594428e-01
8.26624215e-01 5.57403043e-02 -9.17348027e-01 9.86038089e-01
-3.15954238e-01 -9.25510347e-01 -2.65031815e-01 1.32479101e-01
7.07543433e-01 3.09031308e-01 8.21175948e-02 2.58150309e-01
3.36671650e-01 -5.46361566e-01 5.67996323e-01 3.87214392e-01
4.36905086e-01 -9.37008440e-01 6.17498219e-01 -1.22849025e-01
-1.67750823e+00 -2.94637710e-01 1.43450508e-02 -3.72195691e-01
2.39433005e-01 4.57271039e-01 -2.79788096e-02 9.88875329e-01
1.12220347e+00 9.99956310e-01 -4.72403407e-01 1.31348443e+00
6.50994241e-01 8.54783535e-01 -2.72141308e-01 2.23828971e-01
2.80305892e-01 -3.95644665e-01 6.11682951e-01 1.25068247e+00
5.66128671e-01 1.72326416e-01 3.80950511e-01 3.37247968e-01
6.75283909e-01 1.63751677e-01 -7.24762440e-01 -2.69582987e-01
1.13967299e+00 1.15418720e+00 -1.39778805e+00 -1.07707992e-01
-5.18007636e-01 3.75798523e-01 -2.30305091e-01 2.48077109e-01
-7.08880126e-01 -5.21045268e-01 6.98855579e-01 2.98190843e-02
5.88964410e-02 -6.01619899e-01 -3.88139725e-01 -1.00260890e+00
-2.35810146e-01 -7.71293521e-01 8.71857464e-01 -6.45189285e-01
-1.45148718e+00 4.94532019e-01 6.84845448e-01 -1.72150600e+00
-2.54545659e-01 -2.25738242e-01 -7.25157917e-01 3.51489305e-01
-7.17586994e-01 -3.61950845e-01 -2.73616403e-01 5.86574972e-01
3.87728542e-01 -1.32825792e-01 8.78644526e-01 5.60064673e-01
-8.60711694e-01 5.72072603e-02 4.94602323e-01 2.17035681e-01
5.00406325e-01 -1.33562279e+00 3.32374245e-01 7.50725687e-01
-2.51336873e-01 8.48455966e-01 7.87017822e-01 -5.32642484e-01
-1.27931666e+00 -1.14585984e+00 5.57968676e-01 -5.81076145e-01
1.25740683e+00 3.22176628e-02 -9.55345690e-01 5.68183303e-01
3.27800065e-01 -4.00561661e-01 1.11404741e+00 3.04860249e-02
-3.97322960e-02 -8.47756118e-02 -1.17881274e+00 4.30159390e-01
6.22785330e-01 -3.80461633e-01 -8.44333231e-01 1.36373192e-01
4.33308929e-01 1.17394805e-01 -1.81246483e+00 3.78597558e-01
2.77845323e-01 -1.00383699e+00 7.45910168e-01 -2.23932892e-01
7.79430866e-02 -2.59204000e-01 -1.57565251e-01 -1.31621063e+00
-4.81291592e-01 -6.77833498e-01 1.21202506e-01 1.42819023e+00
1.02968968e-01 -7.70333588e-01 1.57035366e-01 4.62921470e-01
-1.63183168e-01 -4.04717624e-01 -9.16992307e-01 -1.19664574e+00
2.02599064e-01 -6.44168198e-01 8.45362544e-01 1.57831132e+00
4.73659873e-01 -1.70294374e-01 2.90859163e-01 -1.58717409e-01
6.32129252e-01 1.74784392e-01 6.46247327e-01 -1.51708460e+00
5.34311235e-01 -9.68439221e-01 -7.15877056e-01 -4.82873172e-02
-1.14145875e-01 -6.14938438e-01 -3.50206077e-01 -1.39584017e+00
-2.46919826e-01 -3.13399106e-01 -5.45316339e-01 3.40793759e-01
2.60665447e-01 7.35069439e-02 9.60779712e-02 7.30303168e-01
-3.11017662e-01 2.46506259e-01 7.30110228e-01 2.00729445e-01
-3.77523929e-01 -6.04785718e-02 -2.34269142e-01 1.99629799e-01
9.57545757e-01 -3.51646662e-01 -7.07260013e-01 -5.63113913e-02
-3.79220456e-01 1.41500220e-01 7.24460185e-02 -1.25843060e+00
3.10278267e-01 -2.19364524e-01 -1.04341216e-01 -8.92443955e-01
-1.39074281e-01 -1.09469056e+00 7.01285243e-01 5.83443642e-01
-1.88406393e-01 9.10482109e-01 4.18745399e-01 4.33313578e-01
-5.07698298e-01 2.19390422e-01 7.37517536e-01 4.73407544e-02
-8.49886775e-01 2.20494792e-01 -5.42658985e-01 1.11311331e-01
1.52484655e+00 -4.80246186e-01 -2.17152178e-01 -3.89491111e-01
-8.14370751e-01 4.10607129e-01 5.79545915e-01 8.23411703e-01
1.79519787e-01 -1.53435993e+00 -6.76564157e-01 -1.77188069e-01
1.73382089e-01 -7.05850601e-01 2.78188497e-01 8.67526829e-01
-5.67321420e-01 5.44683039e-01 -4.85396355e-01 -7.09363699e-01
-1.04023814e+00 8.87451172e-01 1.92787632e-01 -5.53011410e-02
-6.69951081e-01 -3.51885036e-02 -5.22857726e-01 -3.41020256e-01
2.24591359e-01 -3.14703017e-01 -3.51894826e-01 6.18065417e-01
3.71387362e-01 9.07760143e-01 1.63860023e-01 -7.92058110e-01
-5.47329426e-01 6.38656676e-01 4.35945958e-01 -3.37402582e-01
1.67337167e+00 -2.77244866e-01 -4.02164102e-01 1.27613974e+00
1.19034493e+00 -5.40944576e-01 -7.01960206e-01 1.97143719e-01
6.25747979e-01 -3.08477372e-01 -9.82966721e-02 -4.16256338e-01
-7.79043496e-01 5.04779458e-01 8.00282657e-01 1.29863107e+00
1.43732142e+00 -3.74309510e-01 2.82893777e-01 -2.65628286e-03
5.01137257e-01 -1.41826272e+00 -2.34778136e-01 2.77542114e-01
7.17954397e-01 -7.56569624e-01 1.58832187e-03 8.44293833e-02
-4.58995849e-01 1.22853720e+00 2.52684087e-01 -2.85133958e-01
1.21249425e+00 3.29524845e-01 1.38569772e-01 -3.54060113e-01
-8.23543906e-01 -2.02681139e-01 2.84802198e-01 7.80861795e-01
8.05763006e-01 1.52365118e-01 -7.43258476e-01 2.79816836e-01
-2.67843693e-01 -1.39576703e-01 4.49445516e-01 1.14555562e+00
-2.60743111e-01 -9.57376301e-01 -5.81805766e-01 6.05548322e-01
-5.38986742e-01 4.33643699e-01 -4.99290138e-01 1.08684576e+00
-2.37408459e-01 1.48751235e+00 4.61209357e-01 -5.09724021e-01
4.82067257e-01 3.34771909e-02 1.00194559e-01 -3.02981943e-01
-9.95728076e-01 1.72847703e-01 -1.19810745e-01 -5.16317844e-01
-9.91136074e-01 -1.18838882e+00 -9.72162843e-01 -5.55656493e-01
-1.84695870e-02 5.16236782e-01 5.86785078e-01 5.49976647e-01
4.48709488e-01 5.30849516e-01 1.00157773e+00 -8.56871426e-01
-6.86622411e-02 -9.55182612e-01 -6.01931691e-01 6.11183584e-01
-1.58586241e-02 -5.11176169e-01 -5.39260209e-01 1.70129299e-01] | [7.232631683349609, 3.3616225719451904] |
9a447d9e-88b9-4029-abe8-fb83a7cfd1b1 | few-shot-knowledge-graph-to-text-generation | 2106.01623 | null | https://arxiv.org/abs/2106.01623v1 | https://arxiv.org/pdf/2106.01623v1.pdf | Few-shot Knowledge Graph-to-Text Generation with Pretrained Language Models | This paper studies how to automatically generate a natural language text that describes the facts in knowledge graph (KG). Considering the few-shot setting, we leverage the excellent capacities of pretrained language models (PLMs) in language understanding and generation. We make three major technical contributions, namely representation alignment for bridging the semantic gap between KG encodings and PLMs, relation-biased KG linearization for deriving better input representations, and multi-task learning for learning the correspondence between KG and text. Extensive experiments on three benchmark datasets have demonstrated the effectiveness of our model on KG-to-text generation task. In particular, our model outperforms all comparison methods on both fully-supervised and few-shot settings. Our code and datasets are available at https://github.com/RUCAIBox/Few-Shot-KG2Text. | ['Ji-Rong Wen', 'Nicholas Jing Yuan', 'Zhicheng Wei', 'Wayne Xin Zhao', 'Tianyi Tang', 'Junyi Li'] | 2021-06-03 | null | https://aclanthology.org/2021.findings-acl.136 | https://aclanthology.org/2021.findings-acl.136.pdf | findings-acl-2021-8 | ['kg-to-text'] | ['natural-language-processing'] | [ 5.11656366e-02 7.76591063e-01 -8.18154395e-01 -9.56967697e-02
-9.45461094e-01 -2.82063454e-01 8.60419214e-01 3.01380754e-01
9.91386250e-02 9.80402827e-01 7.89655566e-01 -2.26756230e-01
-1.79493558e-02 -1.29557431e+00 -8.44847381e-01 -2.00994343e-01
2.59028882e-01 7.03747869e-01 -1.49773329e-01 -4.14878279e-01
1.09028198e-01 -1.83939740e-01 -1.19958985e+00 4.48432386e-01
1.18442941e+00 5.91035962e-01 1.85598135e-01 7.17300415e-01
-7.87634611e-01 1.24054217e+00 -3.35704505e-01 -9.63139832e-01
4.58953623e-03 -4.76718664e-01 -1.38092625e+00 -2.06351683e-01
3.12994570e-01 -1.07052051e-01 -6.40722156e-01 8.59311640e-01
7.00415313e-01 1.78596541e-01 9.24781680e-01 -1.39695787e+00
-1.23154807e+00 1.57991147e+00 -4.01022047e-01 6.43650219e-02
5.19465387e-01 6.53716084e-03 1.35310066e+00 -8.90955985e-01
9.52976346e-01 1.24835086e+00 3.69821340e-01 6.16216958e-01
-1.04754031e+00 -5.80566585e-01 7.15248752e-03 4.64205921e-01
-1.55140424e+00 -5.43863595e-01 7.28133798e-01 -4.10441071e-01
1.28558135e+00 -2.05874667e-01 6.13181591e-01 1.47303724e+00
-8.84208605e-02 1.10611057e+00 6.84106767e-01 -7.15644777e-01
-1.49829969e-01 1.48291394e-01 2.22936064e-01 1.09507763e+00
8.01207900e-01 -1.65298730e-01 -9.07388866e-01 -2.79137433e-01
7.27687061e-01 -4.69500363e-01 -3.39494973e-01 -2.64427990e-01
-1.13772750e+00 8.57278049e-01 2.47713774e-01 2.59032995e-01
-2.30598256e-01 4.68305618e-01 4.28237379e-01 1.96174756e-01
6.01545930e-01 6.15114629e-01 -3.28727454e-01 9.84166749e-04
-6.26878500e-01 3.71652693e-01 9.37187135e-01 1.62733793e+00
6.99182451e-01 1.91596627e-01 -6.79541171e-01 7.63893604e-01
6.86306059e-02 5.22043169e-01 5.29309034e-01 -4.36008066e-01
9.37065959e-01 5.70993900e-01 -6.60496578e-02 -9.62569058e-01
-1.46615282e-01 -4.61504310e-01 -8.93215775e-01 -7.51607239e-01
1.52141720e-01 -2.04578891e-01 -8.09215724e-01 1.70563602e+00
4.64644358e-02 2.65096396e-01 5.74545085e-01 3.94104272e-01
1.41073108e+00 6.28081679e-01 4.06578511e-01 3.68153304e-02
1.33791244e+00 -1.18740737e+00 -7.42239654e-01 -4.12290335e-01
1.05585945e+00 -4.10850495e-01 1.26702988e+00 -1.55605376e-01
-1.01961374e+00 -4.05972004e-01 -8.63929987e-01 -4.01472181e-01
-6.63592517e-01 3.24590445e-01 7.49961615e-01 1.53549775e-01
-9.45359766e-01 5.62223077e-01 -3.99373561e-01 -5.44661939e-01
6.27178311e-01 -2.63914406e-01 -1.46709815e-01 -2.27536559e-01
-1.79913402e+00 7.61637449e-01 9.56220746e-01 -3.61431450e-01
-8.35064292e-01 -9.41204429e-01 -1.10403919e+00 1.11372113e-01
7.59577870e-01 -1.57219660e+00 1.26006925e+00 -2.98126847e-01
-1.26165652e+00 9.17809904e-01 -1.52493820e-01 -6.64374352e-01
2.95198441e-01 -2.50598103e-01 -4.43912745e-01 7.03814551e-02
2.02657714e-01 6.55423999e-01 6.32853150e-01 -1.19071805e+00
-3.00943971e-01 2.86465860e-03 1.04339831e-01 3.81022751e-01
-2.67981976e-01 -2.65500128e-01 -5.56972206e-01 -8.07694554e-01
-4.83728915e-01 -3.88391107e-01 -1.36525780e-01 -5.55741727e-01
-9.46152151e-01 -6.40521049e-01 8.17380920e-02 -6.24710321e-01
1.39366877e+00 -1.54485261e+00 2.07970440e-01 -1.62368774e-01
1.54447362e-01 1.93706527e-01 -3.26233596e-01 1.01159847e+00
5.68774994e-03 4.35471416e-01 7.96542987e-02 -3.24900895e-01
1.49654821e-01 9.69374031e-02 -9.20756996e-01 -3.16411734e-01
2.80524492e-01 1.84387493e+00 -1.38949668e+00 -7.00961292e-01
-1.36159033e-01 1.10701643e-01 -8.82762447e-02 2.77220905e-01
-7.27778554e-01 5.31058013e-03 -6.29134476e-01 5.67348301e-01
2.09547698e-01 -6.16990387e-01 2.01798379e-01 -3.65949005e-01
6.99989736e-01 6.17002368e-01 -8.00619841e-01 1.99832487e+00
-5.39483368e-01 1.61876544e-01 -7.33797550e-01 -7.88578212e-01
9.57300723e-01 3.11860770e-01 -1.41305909e-01 -4.24599439e-01
1.13593183e-01 7.55860284e-03 -5.09220004e-01 -4.12371725e-01
8.67784619e-01 -1.86314940e-01 -2.90269583e-01 8.18386853e-01
5.61477542e-01 -4.64139789e-01 5.74727893e-01 9.96799767e-01
8.22625577e-01 2.05240279e-01 6.86356544e-01 -1.42568544e-01
9.91232321e-02 5.39877824e-02 1.76234022e-01 1.00739729e+00
3.76792550e-01 2.61712402e-01 6.56233191e-01 -1.38906956e-01
-8.18903625e-01 -9.61062431e-01 3.79777253e-01 1.02087688e+00
3.73291075e-02 -1.00595057e+00 -4.01226103e-01 -8.65514219e-01
2.17118829e-01 1.42229879e+00 -6.14780366e-01 -5.74065804e-01
-2.63916552e-01 -6.04518712e-01 1.04482281e+00 7.28822649e-01
2.82359391e-01 -1.05867696e+00 1.03870437e-01 1.75086394e-01
-4.90552098e-01 -1.23682070e+00 -4.34191346e-01 -1.99261576e-01
-6.82634592e-01 -1.04257321e+00 -3.96233529e-01 -6.03463531e-01
7.72031546e-01 2.54497916e-01 1.62802982e+00 1.23743415e-02
-4.27920073e-01 5.21926880e-01 -5.87540269e-01 -5.73526800e-01
-4.41909343e-01 3.67063344e-01 -2.31117234e-01 -4.16671097e-01
3.17171901e-01 -4.25092995e-01 -1.85965583e-01 -3.24634254e-01
-7.55010843e-01 7.61974692e-01 5.34107864e-01 6.36185646e-01
5.39902270e-01 -1.63178034e-02 9.00027990e-01 -1.38646734e+00
1.12698555e+00 -8.04313719e-01 -1.61352172e-01 7.46298075e-01
-8.24862659e-01 4.89137799e-01 5.84061027e-01 -3.99910212e-02
-1.22069669e+00 -4.25961226e-01 1.48675531e-01 -3.08247656e-01
2.36277714e-01 8.68357539e-01 -8.90954807e-02 3.72754872e-01
8.01986635e-01 4.54329908e-01 -3.54621917e-01 -3.09788108e-01
1.14487290e+00 4.33640897e-01 4.78644937e-01 -1.11119676e+00
9.33951497e-01 2.53743142e-01 -1.74368605e-01 -3.18985373e-01
-1.39440632e+00 -3.08147907e-01 -6.48893952e-01 5.33642131e-04
5.26768088e-01 -1.20795155e+00 3.40794586e-02 2.70910174e-01
-1.17685807e+00 -5.82180858e-01 -7.97844410e-01 -1.04803094e-05
-6.46188378e-01 2.77503580e-01 -5.69801688e-01 -5.68267107e-01
-1.03851163e+00 -3.28936905e-01 1.02665651e+00 1.58837304e-01
-1.99960828e-01 -1.36283147e+00 7.74772912e-02 5.15556931e-01
2.34521881e-01 6.04439490e-02 1.23107290e+00 -6.24143243e-01
-5.93416750e-01 5.35428971e-02 -2.57177174e-01 -7.28275478e-02
2.30787039e-01 -2.80250814e-02 -7.44846463e-01 -5.45385480e-02
-8.56483340e-01 -9.61526811e-01 1.25380373e+00 4.01531719e-02
1.18860018e+00 -6.03442550e-01 -5.39784849e-01 5.87585807e-01
1.51841223e+00 -3.43154460e-01 6.59502149e-01 -4.16614339e-02
9.93378460e-01 4.80862498e-01 6.15848362e-01 4.07525837e-01
9.42351937e-01 5.11306345e-01 -2.56884787e-02 1.22790992e-01
-5.74505985e-01 -1.14595056e+00 3.08912396e-01 1.01815367e+00
-9.02183503e-02 -5.53061724e-01 -9.68591988e-01 9.53601837e-01
-2.04428649e+00 -1.10534298e+00 -2.70234630e-03 1.76031888e+00
1.29469001e+00 1.00275083e-02 -8.95634443e-02 -3.18408817e-01
5.69321930e-01 3.75295639e-01 -4.40983295e-01 -2.66814619e-01
-3.18143666e-01 3.18397224e-01 4.58492517e-01 5.99468231e-01
-8.70625377e-01 1.52085173e+00 5.52257204e+00 1.26778626e+00
-5.52109063e-01 9.44552645e-02 4.75568265e-01 -1.23991951e-01
-7.04258323e-01 4.74574827e-02 -1.21418107e+00 1.87099606e-01
8.70787144e-01 -1.34528708e+00 3.19714963e-01 8.15750539e-01
-1.63727880e-01 3.06985140e-01 -1.06211770e+00 8.21473777e-01
3.40363830e-01 -1.94167471e+00 9.40396130e-01 -3.21667403e-01
1.03670371e+00 -1.07221417e-01 -2.83708215e-01 6.90395355e-01
1.13225985e+00 -1.06401920e+00 5.43517768e-01 6.66282535e-01
1.00606072e+00 -6.41746342e-01 5.55215895e-01 2.52747208e-01
-1.39766264e+00 2.73878366e-01 -6.27461612e-01 8.83128718e-02
2.92458147e-01 8.37780714e-01 -1.26278234e+00 1.36099362e+00
2.11505994e-01 1.03069758e+00 -6.16415739e-01 4.04509872e-01
-1.00561357e+00 4.34576839e-01 2.46865302e-01 1.60558857e-02
-3.03100012e-02 8.87758061e-02 3.66748512e-01 1.40654588e+00
3.32365096e-01 2.11025640e-01 1.60200417e-01 1.24922335e+00
-5.37632525e-01 3.79391581e-01 -1.04368782e+00 -7.03104913e-01
7.24722505e-01 1.35889041e+00 -3.33612442e-01 -8.47998917e-01
-2.11325556e-01 7.80340612e-01 8.49161804e-01 4.17356521e-01
-5.35696149e-01 -5.16119480e-01 4.05838907e-01 4.56451774e-02
2.66372323e-01 -5.33876382e-03 -2.58939594e-01 -1.46465921e+00
-1.71351165e-01 -5.30888438e-01 8.42232168e-01 -1.03928483e+00
-1.55605137e+00 4.17601019e-01 2.46502057e-01 -7.80094087e-01
-6.93243623e-01 -1.58903405e-01 -7.09527016e-01 8.82464647e-01
-1.68736041e+00 -1.52830994e+00 -2.66674250e-01 6.14472985e-01
4.50933039e-01 -1.43835977e-01 8.46232653e-01 -1.23969071e-01
-5.31101048e-01 6.79442227e-01 -3.15857291e-01 2.96481520e-01
6.26079619e-01 -1.30384350e+00 9.26153541e-01 8.36447001e-01
4.48901534e-01 7.06092358e-01 3.07468206e-01 -1.04086781e+00
-1.42120397e+00 -1.64820886e+00 1.15488994e+00 -5.92361510e-01
9.44507778e-01 -4.70613569e-01 -8.17888319e-01 9.47612822e-01
3.64711523e-01 -2.50284851e-01 8.57455611e-01 2.76738882e-01
-6.92249477e-01 1.62500709e-01 -6.78532302e-01 8.07720244e-01
1.48474240e+00 -6.57432377e-01 -7.36209452e-01 6.69890285e-01
1.13185465e+00 -3.55983615e-01 -9.29631650e-01 2.25772947e-01
9.80476886e-02 -2.58840024e-01 9.21615005e-01 -1.11452389e+00
9.56040561e-01 -1.94172058e-02 2.08973080e-01 -1.77359557e+00
-3.77988458e-01 -6.52468204e-01 -6.93098485e-01 1.38076317e+00
6.68907106e-01 -4.36220586e-01 6.67941093e-01 1.75989032e-01
-1.15479417e-01 -9.95326877e-01 -4.01860744e-01 -8.53625536e-01
7.96905980e-02 -4.12175149e-01 7.46077657e-01 1.17630041e+00
4.64013696e-01 8.06290627e-01 -4.84182090e-01 -8.59558806e-02
6.88154399e-01 4.93844509e-01 8.82201970e-01 -9.17387962e-01
-3.85855556e-01 -2.64066279e-01 3.40111204e-03 -8.56422544e-01
5.72606862e-01 -1.69738090e+00 -3.21657151e-01 -2.23476601e+00
6.08620524e-01 -2.70392388e-01 -1.42869100e-01 9.67064202e-01
-6.31124318e-01 -2.78657079e-01 3.30896266e-02 4.25833352e-02
-5.94861925e-01 7.81095624e-01 1.33049774e+00 -2.99674958e-01
6.21563103e-03 -4.84208316e-01 -1.31165934e+00 3.73677224e-01
1.00312495e+00 -3.66630316e-01 -9.18536961e-01 -5.90029895e-01
5.47817588e-01 3.72903496e-02 2.86572784e-01 -4.92774755e-01
2.88396835e-01 -3.42496753e-01 6.43414408e-02 -3.66827011e-01
1.17918111e-01 6.29271716e-02 1.52389854e-01 2.62757897e-01
-6.62432075e-01 -4.10855383e-01 2.64620364e-01 5.95050097e-01
-8.97111297e-02 -2.82126456e-01 3.71094823e-01 -3.74291450e-01
-9.79269266e-01 5.76737940e-01 1.60130098e-01 7.02439010e-01
7.85954237e-01 1.83491856e-01 -1.07037652e+00 -7.09770262e-01
-5.18691242e-01 4.51968729e-01 2.16457546e-01 6.00609004e-01
7.31087923e-01 -1.43633592e+00 -1.00962889e+00 -5.89749739e-02
6.04642093e-01 4.70246337e-02 2.07738593e-01 5.22916436e-01
-1.43317863e-01 7.21101820e-01 7.72535279e-02 2.63156474e-01
-9.69114184e-01 5.33821285e-01 8.15608054e-02 -7.99323201e-01
-6.55627847e-01 1.06421816e+00 5.89668527e-02 -3.67397279e-01
-4.11662459e-02 -1.99579269e-01 -1.53483272e-01 2.38763466e-02
5.22685111e-01 2.79983312e-01 -1.77283332e-01 -2.38974720e-01
4.46112715e-02 2.03009441e-01 -4.66355830e-01 1.73825175e-01
1.20590782e+00 6.65151328e-02 -2.10509524e-01 4.65080827e-01
7.78698683e-01 -1.10040829e-01 -6.05238795e-01 -5.97522914e-01
1.40559494e-01 -2.86583573e-01 -1.26060709e-01 -8.66294444e-01
-9.06322300e-01 6.68603122e-01 -5.15015543e-01 -9.42511633e-02
7.68553615e-01 5.04713655e-01 9.06665206e-01 7.22553194e-01
6.20317817e-01 -9.36063647e-01 2.90939808e-01 6.21761978e-01
1.15182936e+00 -8.85518670e-01 1.19446315e-01 -8.71736586e-01
-1.00749612e+00 8.06478858e-01 7.23718882e-01 7.36112297e-02
3.47539037e-01 1.18523970e-01 -2.71723956e-01 -3.91351968e-01
-1.41298640e+00 -4.77623731e-01 2.99062371e-01 7.30770171e-01
4.40178633e-01 1.62218884e-01 -1.72800854e-01 8.90968621e-01
-5.37393928e-01 3.24877918e-01 6.85505092e-01 7.60562301e-01
-2.73917317e-01 -1.18933535e+00 3.79018247e-01 7.67663896e-01
-7.22874328e-02 -6.79504752e-01 -7.07697034e-01 5.45918763e-01
-2.82134354e-01 7.65380859e-01 -3.19377929e-01 -3.74939859e-01
2.19090283e-01 4.92735267e-01 6.43830359e-01 -1.13381267e+00
-2.39876509e-01 -5.34560800e-01 7.10639536e-01 -4.32569444e-01
-5.49057201e-02 -3.03037912e-01 -1.29504371e+00 -3.81478012e-01
-1.77556634e-01 2.80938923e-01 1.30687743e-01 7.45652497e-01
7.72643983e-01 5.93628705e-01 1.57229900e-01 -2.03719199e-01
-4.87159103e-01 -1.10382915e+00 -7.61933804e-01 5.23368061e-01
-3.43272209e-01 -5.24765790e-01 -9.70622599e-02 2.66632169e-01] | [9.94914436340332, 8.214590072631836] |
0b23f93e-4e85-4cd0-a083-c60c03a4c56b | adaptive-support-driven-bayesian-reweighted | 2008.03877 | null | https://arxiv.org/abs/2008.03877v1 | https://arxiv.org/pdf/2008.03877v1.pdf | Adaptive support driven Bayesian reweighted algorithm for sparse signal recovery | Sparse learning has been widely studied to capture critical information from enormous data sources in the filed of system identification. Often, it is essential to understand internal working mechanisms of unknown systems (e.g. biological networks) in addition to input-output relationships. For this purpose, various feature selection techniques have been developed. For example, sparse Bayesian learning (SBL) was proposed to learn major features from a dictionary of basis functions, which makes identified models interpretable. Reweighted L1-regularization algorithms are often applied in SBL to solve optimization problems. However, they are expensive in both computation and memory aspects, thus not suitable for large-scale problems. This paper proposes an adaptive support driven Bayesian reweighted (ASDBR) algorithm for sparse signal recovery. A restart strategy based on shrinkage-thresholding is developed to conduct adaptive support estimate, which can effectively reduce computation burden and memory demands. Moreover, ASDBR accurately extracts major features and excludes redundant information from large datasets. Numerical experiments demonstrate the proposed algorithm outperforms state-of-the-art methods. | ['Cheng Cheng', 'Wei Zhou', 'Junlin Li'] | 2020-08-10 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 3.35394144e-01 -5.28002977e-01 -2.10219100e-01 -3.10104519e-01
-5.37670135e-01 -6.77128360e-02 1.75215472e-02 8.46973136e-02
5.79390824e-02 8.41165721e-01 1.06980123e-01 1.63076892e-01
-4.47565943e-01 -4.47742730e-01 -4.45939153e-01 -1.03253186e+00
1.25801682e-01 -2.41525006e-02 1.56966135e-01 9.52181593e-02
4.62144315e-01 4.51527774e-01 -1.37153041e+00 6.95385560e-02
8.52265894e-01 1.08244920e+00 4.67255563e-01 6.42103478e-02
-1.33082807e-01 8.35084140e-01 -2.86551684e-01 2.02594340e-01
4.40881774e-02 -4.84752536e-01 -1.29295975e-01 3.49677503e-01
-2.66427308e-01 1.99230872e-02 -4.96677160e-01 1.08074129e+00
5.21387637e-01 1.30604833e-01 5.29955864e-01 -1.11847150e+00
-3.34839284e-01 3.85145366e-01 -9.39411223e-01 3.36088955e-01
1.32110834e-01 2.69356743e-02 6.70849442e-01 -1.43038774e+00
2.50737697e-01 1.20674598e+00 6.46694422e-01 3.31061427e-03
-1.30342579e+00 -7.98113227e-01 1.08524017e-01 2.27543429e-01
-1.62316406e+00 -8.04095626e-01 1.09189510e+00 -4.45103616e-01
4.63816553e-01 3.46628129e-01 4.88936722e-01 8.40011835e-01
1.96642309e-01 7.84209907e-01 1.10211384e+00 -2.04587877e-01
3.61937910e-01 1.40348241e-01 3.21559250e-01 6.59078121e-01
8.28580320e-01 -2.69441426e-01 -7.29205608e-01 -4.46184158e-01
8.67883742e-01 4.67939615e-01 -4.56332415e-01 -2.78131962e-01
-1.11297894e+00 6.18789554e-01 -5.22466935e-02 3.58855844e-01
-5.30480921e-01 -7.59245679e-02 3.07664782e-01 1.03911281e-01
2.48089314e-01 -1.81843787e-02 -3.27178717e-01 7.72096366e-02
-9.96176839e-01 -1.05440192e-01 7.04465330e-01 7.95955300e-01
7.40303576e-01 5.72459519e-01 1.82185918e-02 9.33465719e-01
6.46445572e-01 7.25806355e-01 5.16278028e-01 -8.10488284e-01
2.86136389e-01 6.42007053e-01 -6.99598566e-02 -1.52582860e+00
-3.49517465e-01 -7.13078916e-01 -1.51064682e+00 -4.00918722e-01
9.28084403e-02 3.90216213e-04 -4.78017241e-01 1.25457883e+00
3.72036755e-01 6.67298317e-01 -1.09328166e-01 9.78115916e-01
8.11161935e-01 9.77833629e-01 -1.39942557e-01 -7.50553489e-01
1.15265203e+00 -2.78378993e-01 -8.01555812e-01 -2.59907931e-01
2.70342499e-01 -8.12705219e-01 7.76608706e-01 6.44139886e-01
-8.03212762e-01 -6.04465365e-01 -9.21307564e-01 3.79093796e-01
3.64236534e-01 5.07431507e-01 6.67564929e-01 3.42344999e-01
-3.05316865e-01 3.05070579e-01 -8.88767898e-01 7.82470554e-02
5.09282053e-01 3.22193354e-01 -2.84691691e-01 -3.14277261e-01
-1.11149395e+00 5.22971451e-01 1.98254108e-01 3.97615105e-01
-9.46430683e-01 -4.42869157e-01 -8.22228134e-01 2.22339153e-01
4.98857826e-01 -3.25067371e-01 7.07733989e-01 -5.94285369e-01
-1.26918256e+00 3.21032107e-01 -3.72644365e-01 -2.52835214e-01
-2.20588401e-01 -2.85827946e-02 -4.21249628e-01 1.30951956e-01
-1.39328569e-01 -3.21899712e-01 1.40826428e+00 -1.12420726e+00
-1.15139611e-01 -5.50962448e-01 -5.68774700e-01 -8.77835602e-02
-6.10971034e-01 1.33992722e-02 -1.82505533e-01 -8.94339383e-01
7.29049265e-01 -7.05294192e-01 -4.84095126e-01 -1.61064759e-01
-3.93433154e-01 1.71551555e-01 1.04556191e+00 -7.31921196e-01
1.53884125e+00 -2.19854617e+00 2.57814556e-01 5.59747338e-01
2.42376968e-01 2.72142738e-01 1.18056498e-01 3.00464123e-01
-1.19930036e-01 -3.42070609e-01 -4.38583881e-01 1.07922666e-02
-4.65616435e-01 -4.82225828e-02 -4.35418993e-01 7.88119793e-01
1.53955489e-01 3.28446269e-01 -4.85942572e-01 -6.88686073e-01
1.10025726e-01 5.27048111e-01 -2.19608605e-01 3.20048153e-01
2.20457733e-01 5.80470085e-01 -8.62062573e-01 8.86554837e-01
5.94165504e-01 -5.66650808e-01 1.23293780e-01 -6.33492827e-01
-8.66474733e-02 -2.32895240e-01 -1.76382959e+00 1.56882441e+00
-3.98162633e-01 2.63002843e-01 5.06545067e-01 -1.72553718e+00
1.27020597e+00 3.83104265e-01 6.28407061e-01 -2.52716273e-01
3.19663763e-01 2.66891450e-01 -3.99911217e-02 -6.49496675e-01
-9.95468646e-02 -1.32435754e-01 6.71157092e-02 2.47399643e-01
-8.57599750e-02 -8.47482979e-02 1.92847494e-02 2.18935877e-01
1.01867044e+00 -2.07160592e-01 7.40891993e-01 -3.74271244e-01
1.05481935e+00 -1.66910231e-01 1.19760406e+00 2.58512259e-01
4.32201624e-02 4.04454738e-01 1.69657141e-01 -3.48222226e-01
-6.64793611e-01 -5.51882267e-01 -1.30558342e-01 5.84019601e-01
3.15297812e-01 -3.42983752e-01 -3.52101028e-01 -1.62453383e-01
1.59806862e-01 3.00022542e-01 -1.24258660e-01 -4.67773646e-01
-4.41549391e-01 -9.39190388e-01 1.15056574e-01 3.23874772e-01
3.55031461e-01 -7.31312215e-01 -4.90792274e-01 4.25410837e-01
1.15564160e-01 -8.53970528e-01 -3.62715185e-01 1.21017851e-01
-1.12174261e+00 -1.09949112e+00 -6.61641181e-01 -7.03961432e-01
9.49010491e-01 7.34610021e-01 4.92327094e-01 6.86445013e-02
-5.08583009e-01 2.72963524e-01 -3.20455700e-01 -2.14340895e-01
6.52520806e-02 -2.36259252e-01 5.09311140e-01 5.62854052e-01
1.32916361e-01 -7.40949452e-01 -3.43848318e-01 4.48344350e-01
-8.50786805e-01 -7.83220753e-02 1.03433096e+00 1.10217702e+00
1.00945520e+00 3.43049943e-01 9.51267660e-01 -9.14796054e-01
6.67481601e-01 -7.24763274e-01 -8.43002439e-01 3.17421019e-01
-6.19067490e-01 2.71436460e-02 9.21992481e-01 -5.39217114e-01
-1.17925572e+00 2.57360309e-01 2.57430524e-01 -7.31648326e-01
1.18914880e-01 8.74806225e-01 -4.12031651e-01 -1.75964206e-01
3.29017520e-01 7.94089615e-01 4.18508276e-02 -5.88329196e-01
-1.07844017e-01 8.41638029e-01 2.58286476e-01 -7.59406865e-01
6.91654086e-01 2.60002494e-01 3.00866574e-01 -1.12995970e+00
-8.05663943e-01 -5.83340049e-01 -3.07898879e-01 -9.47747752e-02
1.83481172e-01 -1.08975422e+00 -5.92252433e-01 4.24348712e-01
-7.67050862e-01 3.48477274e-01 -1.99582532e-01 7.31035948e-01
-3.12853038e-01 5.14641583e-01 -3.72294575e-01 -1.07215440e+00
-4.45279121e-01 -9.92820978e-01 7.83244431e-01 3.62436295e-01
-1.08174622e-01 -5.35542250e-01 -6.90445080e-02 2.41142660e-01
4.17028606e-01 -2.63483953e-02 6.07547581e-01 -5.23047566e-01
-4.23127443e-01 -3.30414116e-01 -1.10378750e-01 3.05033088e-01
3.12337160e-01 -1.20354526e-01 -6.82386875e-01 -5.50623119e-01
6.29355073e-01 -2.58987427e-01 8.74492764e-01 6.07343912e-01
1.40301287e+00 -4.29352939e-01 -4.16245759e-01 4.98426080e-01
1.08900034e+00 1.77944541e-01 4.05412257e-01 -1.64650381e-01
7.97423244e-01 4.60618913e-01 8.47422898e-01 1.06961131e+00
-7.87651539e-02 2.31463730e-01 2.02270254e-01 7.84221366e-02
2.97963142e-01 -1.93817783e-02 3.57993960e-01 1.32436693e+00
1.21704243e-01 1.39912993e-01 -7.19371438e-01 4.22090113e-01
-1.94779813e+00 -9.08206403e-01 -2.42840409e-01 2.10463715e+00
9.12778318e-01 1.62047341e-01 -1.97485089e-01 4.77034867e-01
9.66124773e-01 6.92968145e-02 -9.59682167e-01 3.15546453e-01
-1.40238315e-01 4.31315377e-02 2.28840590e-01 5.43379039e-02
-8.63172531e-01 3.97277087e-01 5.15725851e+00 1.20484459e+00
-1.01981318e+00 1.54700428e-01 6.57442927e-01 2.08661035e-02
-1.75584242e-01 1.83072835e-01 -9.28247452e-01 6.67957425e-01
7.53179252e-01 -3.09879422e-01 2.51897722e-01 9.26239312e-01
5.92862904e-01 -1.83874145e-01 -6.85068190e-01 1.49771905e+00
5.09368181e-02 -1.15687525e+00 -1.59715936e-01 -1.72850043e-01
7.22893059e-01 -4.87056106e-01 -8.18251222e-02 -1.35518638e-02
-1.89236164e-01 -7.51438439e-01 2.32431248e-01 7.86489546e-01
7.48073757e-01 -5.52098811e-01 7.29894221e-01 5.10105252e-01
-1.40772200e+00 -3.22658837e-01 -5.87134302e-01 -8.28916579e-02
8.20382237e-02 1.50391662e+00 -3.67117435e-01 3.14851791e-01
5.40088475e-01 1.18110812e+00 -1.41673014e-01 1.12221038e+00
-2.94439383e-02 1.00186896e+00 -3.70919496e-01 1.42901346e-01
-2.99856603e-01 -4.19855982e-01 7.38739967e-01 7.52061725e-01
5.76710999e-01 4.60980386e-01 4.82794672e-01 8.49870861e-01
1.19737145e-02 2.45545611e-01 -5.78849256e-01 -2.76401430e-01
7.86267221e-01 1.43367624e+00 -6.19453132e-01 -2.25603759e-01
-4.42275703e-01 4.43797231e-01 -1.52987525e-01 1.74702078e-01
-6.21118486e-01 -3.90379012e-01 3.82938862e-01 2.98538744e-01
2.45632589e-01 -3.87256563e-01 -1.85189724e-01 -1.38199413e+00
-1.46738380e-01 -9.93234098e-01 2.91556358e-01 -5.51456928e-01
-1.38528740e+00 2.40962684e-01 -1.87389433e-01 -1.53801584e+00
2.07764387e-01 -1.34775728e-01 -4.96546119e-01 7.56758928e-01
-1.39902496e+00 -7.86822081e-01 -3.88711423e-01 7.99880505e-01
7.16460407e-01 -4.97088730e-01 6.39659345e-01 2.92765915e-01
-1.07042027e+00 1.93463355e-01 5.15573084e-01 -1.83499902e-01
6.64856970e-01 -5.51855624e-01 -4.05150443e-01 8.73725116e-01
5.58636896e-03 9.60916638e-01 5.99553704e-01 -7.59007394e-01
-1.99450910e+00 -1.03429139e+00 2.18789980e-01 4.08695638e-01
5.69538474e-01 3.86117920e-02 -1.09566081e+00 2.76828527e-01
-3.56385589e-01 4.57042933e-01 9.24416542e-01 -1.34875387e-01
-9.00356472e-02 -6.28390372e-01 -1.03879786e+00 1.20422862e-01
6.23050690e-01 -3.46261799e-01 -5.69356084e-01 2.02642694e-01
2.67636180e-01 -8.35335776e-02 -9.50265169e-01 4.49517518e-01
4.31178927e-01 -5.85813999e-01 1.05348575e+00 -3.30508977e-01
8.53037015e-02 -6.89423561e-01 -2.85498887e-01 -1.07481086e+00
-4.81138289e-01 -6.14763021e-01 -7.82534838e-01 1.29425526e+00
-8.98758546e-02 -6.31111681e-01 6.11534178e-01 4.67934221e-01
-4.81715463e-02 -8.33719671e-01 -8.68728340e-01 -7.22559690e-01
-7.93402731e-01 -1.31155938e-01 2.13903621e-01 9.67558265e-01
-7.67872809e-03 5.51244020e-01 -6.76893115e-01 1.18289769e-01
8.99173796e-01 5.06280005e-01 5.80965579e-01 -1.53876960e+00
-3.45127881e-01 -3.01411673e-02 -3.01498801e-01 -9.37916934e-01
1.94541052e-01 -7.33451188e-01 -6.18298687e-02 -1.17982352e+00
3.82513911e-01 -6.64581180e-01 -6.21531308e-01 4.35732067e-01
-2.90914953e-01 3.76482606e-02 -1.43954754e-01 7.64429748e-01
-6.52487159e-01 8.61826718e-01 1.05857563e+00 -2.88922578e-01
-2.07497001e-01 7.19100535e-02 -7.94282317e-01 7.54573464e-01
8.59432459e-01 -6.32781506e-01 -5.43079793e-01 -1.39671117e-01
1.08009502e-01 3.60569119e-01 1.43384144e-01 -1.03987432e+00
3.91641676e-01 -4.01162386e-01 5.24403095e-01 -4.86759156e-01
3.03159833e-01 -1.12052548e+00 3.59313190e-01 5.49806356e-01
-1.10820323e-01 -2.79292196e-01 -6.33673146e-02 9.13308859e-01
-4.82603043e-01 -3.58191401e-01 6.97346807e-01 -1.27296612e-01
-5.08619070e-01 2.80287772e-01 -4.09220845e-01 -2.19287295e-02
9.57720757e-01 -1.54169559e-01 6.89289868e-02 -2.62905180e-01
-5.41542470e-01 2.82864183e-01 5.60819767e-02 -1.11479372e-01
1.05310512e+00 -1.25817418e+00 -7.65673697e-01 4.63384867e-01
-9.36257839e-02 1.44506425e-01 4.74681586e-01 1.19971836e+00
-1.26916766e-01 4.54428077e-01 3.77743447e-04 -5.36893725e-01
-1.25368989e+00 4.16831374e-01 -2.42911920e-01 -1.15516536e-01
-5.08316338e-01 7.87606537e-01 -6.02740608e-02 1.01090319e-01
6.68864995e-02 -4.18137759e-02 -3.20087820e-01 8.46872926e-02
6.16258860e-01 5.16543031e-01 -7.09821656e-02 -6.51760280e-01
-3.93945843e-01 7.66703129e-01 2.60855956e-03 4.32396948e-01
1.70058918e+00 -1.86137959e-01 -4.00475502e-01 5.36413610e-01
9.30333257e-01 6.77231997e-02 -1.04005349e+00 -5.27408600e-01
2.53777998e-03 -4.82515782e-01 4.07417327e-01 -1.52406991e-01
-1.23466527e+00 9.13202822e-01 3.99227589e-01 3.17565165e-02
1.27175176e+00 -4.29230094e-01 8.92548919e-01 6.23526275e-01
5.94873071e-01 -1.00585306e+00 3.93328965e-02 1.66107878e-01
8.89769673e-01 -1.22573280e+00 5.70460200e-01 -5.46977878e-01
-4.45735097e-01 1.12240148e+00 5.19551635e-01 -2.54535735e-01
9.02836919e-01 3.73693049e-01 -6.01311922e-01 -3.43413316e-02
-6.45064712e-01 2.54758745e-01 1.84179097e-01 2.14902803e-01
3.72669309e-01 -2.88459301e-01 -5.39913535e-01 1.25530732e+00
5.27760327e-01 1.27241984e-01 3.00399423e-01 1.03228235e+00
-6.06642306e-01 -8.88516307e-01 -8.22291970e-01 9.08580005e-01
-2.57479370e-01 -3.00235767e-03 -6.52084884e-04 1.28949493e-01
-1.43623635e-01 9.40233231e-01 -4.46229935e-01 -3.73191059e-01
8.31749961e-02 -2.32428193e-01 1.86585203e-01 -6.36290193e-01
-2.21305946e-03 3.52502674e-01 -3.41147065e-01 -4.80117023e-01
-5.04906714e-01 -8.30977201e-01 -1.32323420e+00 1.84353590e-01
-8.05034101e-01 4.14399415e-01 4.64604020e-01 7.90783823e-01
5.08752108e-01 5.58945596e-01 8.76178026e-01 -7.25366533e-01
-6.73633456e-01 -9.41690087e-01 -9.44198847e-01 1.06318586e-01
1.16385378e-01 -8.93687367e-01 -4.57827508e-01 2.90013969e-01] | [12.435935020446777, 0.41179534792900085] |
6d0e0d70-44ad-4e74-a7c0-11b8341c7b5d | risk-aware-meta-level-decision-making-for | 2209.05580 | null | https://arxiv.org/abs/2209.05580v1 | https://arxiv.org/pdf/2209.05580v1.pdf | Risk-aware Meta-level Decision Making for Exploration Under Uncertainty | Robotic exploration of unknown environments is fundamentally a problem of decision making under uncertainty where the robot must account for uncertainty in sensor measurements, localization, action execution, as well as many other factors. For large-scale exploration applications, autonomous systems must overcome the challenges of sequentially deciding which areas of the environment are valuable to explore while safely evaluating the risks associated with obstacles and hazardous terrain. In this work, we propose a risk-aware meta-level decision making framework to balance the tradeoffs associated with local and global exploration. Meta-level decision making builds upon classical hierarchical coverage planners by switching between local and global policies with the overall objective of selecting the policy that is most likely to maximize reward in a stochastic environment. We use information about the environment history, traversability risk, and kinodynamic constraints to reason about the probability of successful policy execution to switch between local and global policies. We have validated our solution in both simulation and on a variety of large-scale real world hardware tests. Our results show that by balancing local and global exploration we are able to significantly explore large-scale environments more efficiently. | ['Ali-akbar Agha-mohammadi', 'Joel Burdick', 'Mykel J. Kochenderfer', 'Harrison Delecki', 'Mamoru Sobue', 'Oriana Peltzer', 'Amanda Bouman', 'Sung-Kyun Kim', 'Joshua Ott'] | 2022-09-12 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 2.56207556e-01 3.60446006e-01 -2.22896650e-01 -3.06173772e-01
-7.36663699e-01 -5.53832948e-01 3.63707244e-01 5.29793561e-01
-8.37537587e-01 1.12022746e+00 9.43575054e-02 -6.47883058e-01
-5.84479690e-01 -1.12162364e+00 -7.12255299e-01 -4.19158667e-01
-6.30959690e-01 8.14461589e-01 5.30434668e-01 -2.37547576e-01
4.05599892e-01 6.26150250e-01 -1.40974653e+00 -5.62984109e-01
9.11818922e-01 9.00056601e-01 3.97016436e-01 6.52785420e-01
2.83916086e-01 3.94645512e-01 -3.86824936e-01 4.99045640e-01
3.45279962e-01 -1.81914549e-02 -8.58464301e-01 -1.73190966e-01
-5.74149668e-01 -4.15048659e-01 7.42452964e-02 9.85986888e-01
4.02057916e-01 3.57707381e-01 4.91817057e-01 -1.25750375e+00
6.82766557e-01 7.06145108e-01 -4.47483510e-01 1.95406318e-01
2.16838032e-01 4.14389580e-01 3.97442728e-01 3.41420993e-03
5.76467812e-01 1.39318705e+00 3.37952599e-02 -8.14163014e-02
-1.03699172e+00 -4.01377976e-01 5.46977520e-01 1.02809682e-01
-1.24691176e+00 -2.05449149e-01 3.52922589e-01 -2.48468816e-01
9.37730849e-01 5.66358306e-02 5.22037685e-01 6.84448779e-01
9.07138944e-01 3.37082058e-01 1.00664675e+00 -3.55968297e-01
9.42248106e-01 -2.73304790e-01 -4.39579606e-01 4.17832404e-01
5.67622840e-01 5.16773939e-01 -3.67616385e-01 -4.32809263e-01
4.58331704e-01 -3.22062880e-01 1.12071209e-01 -7.69455791e-01
-1.17164195e+00 5.79014421e-01 3.86717439e-01 -1.27348259e-01
-6.04117632e-01 4.15986538e-01 2.35343069e-01 1.23398632e-01
-3.45445544e-01 8.08992743e-01 -2.89365500e-01 -2.66967505e-01
-3.07988048e-01 5.03835320e-01 8.84689033e-01 9.52415526e-01
8.25172544e-01 -8.87791961e-02 2.19502643e-01 2.32308716e-01
2.73001075e-01 4.20554072e-01 5.82342856e-02 -1.29891789e+00
6.56857312e-01 3.17839950e-01 7.18840420e-01 -7.05803037e-01
-6.09079838e-01 -2.29486957e-01 -8.65582004e-02 9.01465356e-01
1.38934284e-01 -5.70257366e-01 -8.80476356e-01 1.57472789e+00
3.63659263e-01 -4.28951889e-01 3.47372860e-01 6.63511097e-01
-5.75181961e-01 4.49052542e-01 2.11693943e-01 -6.78436309e-02
9.35654223e-01 -2.07878098e-01 -3.72078300e-01 -6.07130826e-01
4.00077999e-01 -2.43090332e-01 4.89123732e-01 4.64637786e-01
-9.51790869e-01 9.67311338e-02 -1.42352104e+00 5.10025799e-01
-2.01451704e-01 -3.15107048e-01 4.64873850e-01 3.15697610e-01
-8.98665786e-01 6.17150307e-01 -1.18265498e+00 -4.86325413e-01
5.33992648e-02 3.69664580e-01 -7.62109235e-02 -8.53288397e-02
-1.00294137e+00 1.33757806e+00 7.11493075e-01 2.43191533e-02
-1.39341295e+00 1.67807490e-01 -7.98737705e-01 1.00395061e-01
1.01140392e+00 -2.36526623e-01 1.40136552e+00 8.43633860e-02
-1.43836725e+00 -1.28033519e-01 3.94478291e-02 -6.18002653e-01
6.22625887e-01 -1.32497385e-01 9.94524732e-02 -3.04373235e-01
3.33914876e-01 5.02782226e-01 1.10424615e-01 -1.11786699e+00
-1.05065930e+00 -3.82889211e-01 1.87179074e-01 8.59475493e-01
2.36050695e-01 -5.88492513e-01 -1.36292279e-01 1.81273162e-01
4.45119500e-01 -1.18321061e+00 -1.08545995e+00 -2.70479888e-01
-3.46871018e-01 2.23660007e-01 4.13377166e-01 -1.09616295e-01
6.38666213e-01 -1.64115345e+00 2.29190692e-01 6.77329123e-01
-4.68225062e-01 -6.11424029e-01 6.81823045e-02 6.51841462e-01
8.64473462e-01 9.29995477e-02 -1.46060348e-01 8.62236097e-02
-1.45822903e-02 4.53527689e-01 -2.88379103e-01 3.61990273e-01
-1.75635248e-01 2.55639046e-01 -1.01183212e+00 -3.11610937e-01
3.47218961e-01 -1.30301520e-01 -3.10968697e-01 2.97483075e-02
-4.07088429e-01 4.74814951e-01 -1.10288942e+00 6.96459532e-01
2.78769553e-01 3.63895863e-01 6.78835392e-01 5.38921475e-01
-6.22739851e-01 4.08869237e-01 -1.33444965e+00 1.53805852e+00
-6.52049541e-01 2.63820887e-01 5.82025290e-01 -5.48725247e-01
7.91717887e-01 -1.57646775e-01 4.19280946e-01 -6.94853067e-01
3.03874403e-01 5.25240839e-01 -9.46129858e-02 -1.39211655e-01
8.22843492e-01 1.09353736e-01 -5.83804905e-01 3.86654139e-01
-4.64701563e-01 -4.91646618e-01 -2.55494248e-02 -7.03086108e-02
1.56677186e+00 2.93752044e-01 5.70236206e-01 -6.50070012e-01
-7.56226853e-02 5.48980415e-01 6.78464115e-01 9.85421240e-01
-2.84126133e-01 -2.37632409e-01 5.58714807e-01 -3.78906488e-01
-5.69683731e-01 -9.51289117e-01 1.28138393e-01 7.18786538e-01
1.12268627e+00 1.40279252e-02 -3.32288772e-01 -2.11371228e-01
1.12032264e-01 1.10162044e+00 -4.76307154e-01 -5.45688998e-03
-5.42480528e-01 -5.77118456e-01 1.84513032e-02 3.41088980e-01
2.69799918e-01 -8.16656351e-01 -1.80354965e+00 5.26892066e-01
9.13674235e-02 -9.33560848e-01 5.01392633e-02 7.93050528e-01
-7.89624929e-01 -9.42309201e-01 1.20463282e-01 -6.83672875e-02
6.23327196e-01 1.41458794e-01 4.87966716e-01 -3.13471645e-01
-4.48630810e-01 2.11983114e-01 -2.87573725e-01 -5.32168567e-01
-2.79000938e-01 5.86833172e-02 1.13652565e-01 -7.39442229e-01
-2.67100930e-01 -6.03423119e-01 -3.64450634e-01 5.14420807e-01
-5.32537878e-01 -2.48596258e-02 8.03539991e-01 3.70211035e-01
7.63589501e-01 7.76545167e-01 1.31863028e-01 -3.60573322e-01
6.76892161e-01 -7.51818180e-01 -1.10137796e+00 2.83274710e-01
-5.88013649e-01 5.43640494e-01 3.57058607e-02 -2.82893836e-01
-1.03969765e+00 2.96828419e-01 3.49066287e-01 2.11391270e-01
-1.45726204e-02 6.02247477e-01 -4.83559906e-01 -1.07480094e-01
5.00487506e-01 -2.03427851e-01 -2.61699319e-01 -1.45369351e-01
1.81622460e-01 2.11211771e-01 3.40450972e-01 -1.14584661e+00
4.98147786e-01 3.97232085e-01 2.95101374e-01 -4.47089255e-01
-3.61135527e-02 -6.21051975e-02 -2.89153010e-01 -3.47933680e-01
5.78931093e-01 -7.55169630e-01 -8.03298950e-01 4.32197750e-02
-5.99600971e-01 -6.55341685e-01 -3.08143824e-01 3.60294640e-01
-8.13407600e-01 -1.53799996e-01 1.73293129e-01 -1.24571145e+00
2.17912018e-01 -1.60159099e+00 7.73943961e-01 3.92701954e-01
-2.16902435e-01 -4.27969813e-01 3.42118181e-02 -3.77689511e-01
6.31893337e-01 8.22354138e-01 6.80109739e-01 -1.58989042e-01
-1.07492626e+00 -1.61443260e-02 1.94664240e-01 -7.40886509e-01
8.71387310e-03 -4.69784766e-01 -3.63393545e-01 -4.19715285e-01
-1.63647234e-01 -2.56385297e-01 4.76259053e-01 2.95396298e-01
6.28138065e-01 -3.13782394e-01 -8.77839804e-01 2.32770041e-01
1.59380436e+00 6.95320725e-01 4.47936118e-01 9.48803544e-01
-7.13616386e-02 7.58851469e-01 1.38866079e+00 8.05895209e-01
5.75594485e-01 8.13180149e-01 1.08290327e+00 5.22906959e-01
6.24832094e-01 -3.84186476e-01 3.58415216e-01 -5.68192005e-01
2.14865357e-01 -5.80313027e-01 -1.14759982e+00 7.82054007e-01
-2.04511118e+00 -4.98575032e-01 7.08623409e-01 2.47537041e+00
4.00948495e-01 6.12213969e-01 5.67140058e-03 -1.29622132e-01
4.92123634e-01 -1.80775180e-01 -1.04890680e+00 -4.02825445e-01
3.12759131e-01 -4.83651936e-01 1.56167543e+00 9.42774534e-01
-8.83327663e-01 7.76208639e-01 6.32539463e+00 4.75347698e-01
-9.21440005e-01 -2.57190943e-01 3.93436670e-01 -3.08115244e-01
-3.69832069e-01 4.80540454e-01 -8.15018415e-01 1.46265939e-01
9.28772986e-01 -3.47809821e-01 5.26397765e-01 9.54876006e-01
1.11435167e-01 -1.08278728e+00 -9.02386189e-01 8.87863189e-02
-6.70286536e-01 -1.05777299e+00 -5.56170881e-01 4.33749110e-01
4.76887137e-01 3.83910775e-01 -2.08451837e-01 1.60839692e-01
1.29779053e+00 -7.92739749e-01 1.24309111e+00 2.23651275e-01
5.21749318e-01 -8.89851213e-01 6.81030393e-01 7.37430513e-01
-1.07673132e+00 -5.57864666e-01 -7.40314797e-02 -2.55481750e-01
5.75388014e-01 2.96042353e-01 -1.21644580e+00 6.03986025e-01
6.83088541e-01 -2.60132223e-01 6.14166521e-02 1.06884992e+00
-3.01521152e-01 -5.93480580e-02 -9.07004058e-01 -3.13785493e-01
3.62693578e-01 3.86948697e-02 7.12247968e-01 4.38800812e-01
5.03921807e-01 2.93237239e-01 6.66774690e-01 6.85085595e-01
7.80368865e-01 -4.41625297e-01 -5.71527839e-01 2.38596767e-01
9.49083507e-01 5.88629007e-01 -8.79337668e-01 7.31396452e-02
4.35614288e-01 3.18372697e-01 2.82406151e-01 2.97234356e-01
-4.54643250e-01 -3.81639719e-01 6.96714640e-01 -1.33371549e-02
-3.48682143e-02 -7.03628242e-01 -4.67619568e-01 -4.35757875e-01
5.20250276e-02 -2.99366981e-01 2.61621505e-01 -3.93424004e-01
-3.08613896e-01 6.09925568e-01 5.36518216e-01 -9.85484302e-01
-5.04442513e-01 -3.01292628e-01 -5.57233393e-01 8.45363438e-01
-1.60484624e+00 -4.77574646e-01 -1.48076043e-01 -7.12909037e-03
4.40384328e-01 1.72525734e-01 5.76048136e-01 -5.33874154e-01
-1.70811221e-01 -1.68942332e-01 5.08252233e-02 -6.34501159e-01
1.50941402e-01 -1.02541721e+00 4.01255369e-01 9.06316459e-01
-8.32551360e-01 2.37903073e-01 1.16603374e+00 -1.18179882e+00
-1.45562303e+00 -6.50779724e-01 2.33432278e-01 6.32738182e-03
4.25380975e-01 -3.55113968e-02 -3.29547048e-01 5.63229501e-01
-2.70650685e-01 -4.98038948e-01 -1.09726503e-01 6.95417002e-02
1.33648962e-01 1.13088503e-01 -1.54138279e+00 7.21615314e-01
8.78629804e-01 1.62655085e-01 -3.87707464e-02 -3.05399597e-02
6.37747467e-01 -6.37081027e-01 -5.58867931e-01 8.89342844e-01
5.85207045e-01 -7.08082497e-01 5.51334202e-01 6.28885701e-02
-2.74284810e-01 -3.95277679e-01 -5.62306166e-01 -1.25516272e+00
-5.40960059e-02 -7.17785358e-01 3.26920956e-01 5.36654174e-01
4.95477498e-01 -6.85325980e-01 8.64549339e-01 9.07008767e-01
-6.88015446e-02 -6.25583053e-01 -1.31213522e+00 -8.85202706e-01
-1.22892790e-01 -2.66751111e-01 8.06502759e-01 9.23085138e-02
1.74613982e-01 -2.28979126e-01 -1.58729002e-01 7.27063179e-01
8.47428262e-01 1.81642547e-02 6.38970137e-01 -9.85877633e-01
2.52644042e-03 -3.10397863e-01 -2.41903156e-01 -5.78511178e-01
-1.00715704e-01 -9.10064802e-02 9.29202497e-01 -1.72066402e+00
-2.00478733e-01 -1.07811213e+00 -1.30749807e-01 5.26156843e-01
2.05317721e-01 -7.84841657e-01 8.92615225e-03 1.66787848e-01
-6.25826597e-01 5.77102184e-01 8.41108561e-01 -8.67255628e-02
-6.55666828e-01 2.38479692e-02 -4.70952958e-01 5.45677483e-01
8.87135983e-01 -3.63097578e-01 -5.81268370e-01 -5.52813292e-01
2.11202472e-01 7.68138289e-01 -2.38068700e-01 -1.39901125e+00
4.77618545e-01 -1.04738545e+00 -9.50715616e-02 -5.37176013e-01
4.50267583e-01 -9.49109674e-01 2.91437745e-01 8.53897810e-01
-5.80478191e-01 6.85490146e-02 4.00782645e-01 1.00872159e+00
1.65609077e-01 -1.48607180e-01 5.51904142e-01 -1.68701366e-01
-1.12841868e+00 1.10095285e-01 -7.68516004e-01 -2.91132122e-01
1.49860036e+00 -4.84974645e-02 -1.43712834e-01 -4.91738439e-01
-6.22850597e-01 1.03389084e+00 8.83418024e-01 3.98317277e-01
3.95676106e-01 -6.25391722e-01 -3.12963754e-01 3.03531885e-02
1.03948854e-01 1.78908572e-01 2.25860849e-02 2.99637735e-01
-5.34096062e-01 3.04512650e-01 -5.23589551e-01 -3.19917232e-01
-7.38294423e-01 2.01605886e-01 3.13068181e-01 -3.93025756e-01
-3.11137676e-01 6.21400177e-01 -1.92459524e-01 -4.06068146e-01
4.09988552e-01 -3.87711972e-01 -7.58414268e-02 -4.46620017e-01
3.69756848e-01 4.86626238e-01 -7.01338798e-02 -1.50336906e-01
-5.15448749e-01 3.56056273e-01 2.62687325e-01 -8.49222481e-01
1.00540209e+00 -3.89809251e-01 1.18950017e-01 1.50478005e-01
2.90314347e-01 -1.68169141e-02 -1.65555584e+00 -1.12167738e-01
3.57612133e-01 -4.66243327e-01 2.31343597e-01 -9.73065317e-01
-2.19353363e-01 3.33989382e-01 3.79591167e-01 -1.76345289e-01
8.79665434e-01 -1.92452863e-01 3.15630972e-01 7.19743490e-01
1.52117836e+00 -1.29711723e+00 -1.67853311e-01 5.97058773e-01
6.36028767e-01 -8.88554811e-01 3.35077912e-01 -1.92620993e-01
-6.95825338e-01 7.55969107e-01 7.36711621e-01 -1.06631825e-02
3.59691530e-01 6.05858803e-01 -1.94168806e-01 -4.99896035e-02
-9.16649818e-01 -1.03168279e-01 -6.04654670e-01 4.85093355e-01
-6.70386553e-01 4.44952250e-01 -4.08145674e-02 3.72344293e-02
-4.28860225e-02 -2.14401290e-01 7.56690979e-01 1.64643633e+00
-1.24497724e+00 -1.14492238e+00 -6.88337684e-01 3.40269953e-01
1.12146690e-01 5.43462932e-01 -9.64167193e-02 7.80612171e-01
-8.23538378e-03 9.93955493e-01 -9.34122950e-02 -4.02523398e-01
2.76491612e-01 -3.69453430e-01 3.24050277e-01 -6.18275225e-01
4.98544909e-02 -7.73329139e-02 5.20973027e-01 -9.40194547e-01
1.95897982e-01 -9.33657706e-01 -1.51168668e+00 1.63967341e-01
-1.77544415e-01 2.97414601e-01 1.22059047e+00 9.72112060e-01
4.60109442e-01 6.45435810e-01 3.62612039e-01 -1.14709508e+00
-9.50539351e-01 -4.28767800e-01 -3.99633795e-01 -6.17114127e-01
3.83481264e-01 -9.62053597e-01 -8.28785971e-02 -8.25540423e-01] | [4.7596211433410645, 1.879514455795288] |
860ca1a4-ce25-4f06-8427-03c0911e8123 | self-supervised-predictive-convolutional | 2111.09099 | null | https://arxiv.org/abs/2111.09099v6 | https://arxiv.org/pdf/2111.09099v6.pdf | Self-Supervised Predictive Convolutional Attentive Block for Anomaly Detection | Anomaly detection is commonly pursued as a one-class classification problem, where models can only learn from normal training samples, while being evaluated on both normal and abnormal test samples. Among the successful approaches for anomaly detection, a distinguished category of methods relies on predicting masked information (e.g. patches, future frames, etc.) and leveraging the reconstruction error with respect to the masked information as an abnormality score. Different from related methods, we propose to integrate the reconstruction-based functionality into a novel self-supervised predictive architectural building block. The proposed self-supervised block is generic and can easily be incorporated into various state-of-the-art anomaly detection methods. Our block starts with a convolutional layer with dilated filters, where the center area of the receptive field is masked. The resulting activation maps are passed through a channel attention module. Our block is equipped with a loss that minimizes the reconstruction error with respect to the masked area in the receptive field. We demonstrate the generality of our block by integrating it into several state-of-the-art frameworks for anomaly detection on image and video, providing empirical evidence that shows considerable performance improvements on MVTec AD, Avenue, and ShanghaiTech. We release our code as open source at https://github.com/ristea/sspcab. | ['Mubarak Shah', 'Thomas B. Moeslund', 'Fahad Shahbaz Khan', 'Kamal Nasrollahi', 'Radu Tudor Ionescu', 'Neelu Madan', 'Nicolae-Catalin Ristea'] | 2021-11-17 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Ristea_Self-Supervised_Predictive_Convolutional_Attentive_Block_for_Anomaly_Detection_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Ristea_Self-Supervised_Predictive_Convolutional_Attentive_Block_for_Anomaly_Detection_CVPR_2022_paper.pdf | cvpr-2022-1 | ['one-class-classification'] | ['miscellaneous'] | [ 1.44311890e-01 8.52707177e-02 9.97444540e-02 -4.47928190e-01
-4.47348654e-01 -1.98682383e-01 5.35695493e-01 2.59657592e-01
-2.52864957e-01 2.52780616e-01 -2.25507632e-01 -5.84071577e-02
1.29634380e-01 -6.96968675e-01 -8.31112742e-01 -7.80995846e-01
-4.38887656e-01 7.13668242e-02 4.82540935e-01 -1.61774665e-01
3.30147177e-01 6.38832390e-01 -1.77290940e+00 5.52027404e-01
8.06677639e-01 1.48259211e+00 -2.51598269e-01 5.55107594e-01
2.88543589e-02 9.58554149e-01 -4.79520679e-01 -1.91191986e-01
3.30286801e-01 -3.74380767e-01 -6.12675071e-01 3.12906921e-01
5.57895422e-01 -4.62695658e-01 -4.46678370e-01 1.08556056e+00
1.92231715e-01 5.16011007e-03 5.60609400e-01 -1.25120759e+00
-3.08832467e-01 1.48712039e-01 -3.87830049e-01 6.92849576e-01
2.03873128e-01 2.78008878e-01 1.01417303e+00 -1.06931078e+00
3.25586140e-01 8.99830222e-01 5.84984243e-01 4.95845795e-01
-1.16803920e+00 -4.27247256e-01 4.81011212e-01 6.59994245e-01
-1.17518175e+00 -4.42321211e-01 8.18749905e-01 -5.67800343e-01
8.64732146e-01 2.14656532e-01 5.46008050e-01 1.09165907e+00
2.48543397e-01 9.72783387e-01 6.00349903e-01 -3.03668231e-01
4.17067558e-01 4.72989194e-02 9.70261693e-02 8.33159387e-01
1.52232125e-01 8.63902271e-02 -2.94879407e-01 -2.41761096e-02
4.94614273e-01 4.37364578e-01 -2.46932179e-01 -4.60690081e-01
-8.55379879e-01 7.36034334e-01 5.83429337e-01 4.21964377e-01
-4.81825858e-01 -7.14073097e-03 3.79920363e-01 4.44320023e-01
6.58443093e-01 2.23230362e-01 -2.80433536e-01 1.72724769e-01
-1.00671899e+00 1.23486876e-01 4.76238161e-01 4.95233893e-01
7.04088151e-01 2.68179297e-01 -2.63312995e-01 6.50530696e-01
4.55227107e-01 1.55383563e-02 6.04441047e-01 -5.74608207e-01
2.08507746e-01 9.17136848e-01 -2.02674717e-01 -1.00263560e+00
-3.54584724e-01 -6.17752075e-01 -7.66364932e-01 6.12684011e-01
3.63198042e-01 -4.61844262e-03 -9.10458863e-01 1.47135377e+00
2.16767102e-01 5.38305223e-01 -9.07585993e-02 7.75402367e-01
3.35725814e-01 5.14481664e-01 -1.31298289e-01 3.40884961e-02
9.38798487e-01 -9.35920179e-01 -3.53020996e-01 -3.37469488e-01
7.00112283e-01 -4.64281261e-01 7.75394857e-01 6.59109354e-01
-1.13162816e+00 -5.09854794e-01 -1.10421085e+00 2.66071439e-01
-4.17880923e-01 1.84623882e-01 1.87246755e-01 3.72959763e-01
-1.09608066e+00 7.97334135e-01 -1.15155602e+00 -3.91491205e-01
6.88337922e-01 1.93995252e-01 -4.23902690e-01 6.00725003e-02
-8.21147203e-01 7.42762685e-01 2.00693056e-01 2.24448055e-01
-1.07263851e+00 -4.19374794e-01 -8.61357808e-01 1.03903197e-01
2.67530531e-01 -2.64806539e-01 9.55542386e-01 -1.39448202e+00
-1.10212374e+00 8.83346021e-01 -1.13217905e-01 -8.41746211e-01
6.32035136e-01 -4.25940931e-01 -5.87955773e-01 2.84809202e-01
-5.63958809e-02 3.78388613e-01 1.38795292e+00 -1.10651612e+00
-7.65415609e-01 -4.28753883e-01 -1.82734236e-01 -1.09235018e-01
-4.40942496e-01 -3.92449498e-02 -2.30126947e-01 -9.40493405e-01
4.12959367e-01 -5.76541185e-01 -3.13781679e-01 2.15599939e-01
-4.32518482e-01 -6.68130741e-02 1.00771320e+00 -7.33011007e-01
1.39882231e+00 -2.38146424e+00 3.83582450e-02 4.24975872e-01
1.67543396e-01 3.82658601e-01 3.99936214e-02 9.59489271e-02
-5.08295059e-01 -2.19962567e-01 -5.84304392e-01 -3.93427879e-01
-1.26354262e-01 1.80933312e-01 -4.42676872e-01 7.24197805e-01
7.05567718e-01 5.19905329e-01 -6.40482128e-01 -3.82688418e-02
4.57204491e-01 2.78079391e-01 -7.40112782e-01 3.92559350e-01
-1.60621136e-01 7.04654932e-01 -3.63401920e-01 9.23597813e-01
5.94138801e-01 -4.46562432e-02 -3.32703382e-01 1.41795084e-01
-1.03298463e-01 1.43828064e-01 -1.15130126e+00 1.59043539e+00
-9.49906185e-02 5.08241296e-01 -2.29231361e-02 -1.57094383e+00
9.23696280e-01 3.19715708e-01 5.54081380e-01 -5.16381860e-01
9.88022387e-02 3.27811152e-01 1.29906535e-01 -5.28835058e-01
1.07122064e-01 2.65493244e-01 2.30107322e-01 2.31752664e-01
2.95146674e-01 5.00674367e-01 9.71391872e-02 6.46130890e-02
1.48476481e+00 4.85109575e-02 2.75114864e-01 -2.26701617e-01
9.84752834e-01 -2.05359593e-01 4.72974837e-01 7.30085790e-01
-3.74776602e-01 7.59992957e-01 7.89514005e-01 -7.41354823e-01
-9.48146462e-01 -1.09409368e+00 -3.74506086e-01 8.45992625e-01
-1.53823972e-01 -2.65544862e-01 -8.44665766e-01 -1.02946436e+00
2.95333788e-02 5.37953138e-01 -7.39371359e-01 -4.11765963e-01
-5.86809039e-01 -7.64814913e-01 4.48643684e-01 6.02777302e-01
4.69776273e-01 -1.18523550e+00 -6.34177148e-01 1.02179907e-01
1.58062920e-01 -8.99599910e-01 -4.02226634e-02 2.97526449e-01
-1.04083431e+00 -1.05681658e+00 -3.15854847e-01 -5.59305549e-01
8.93789589e-01 -1.53305307e-01 1.13027108e+00 3.07166964e-01
-5.12054503e-01 4.71987575e-01 -4.50531811e-01 -3.67971808e-01
-2.48799220e-01 -2.96340525e-01 1.45382918e-02 7.14947641e-01
4.13504720e-01 -7.29048729e-01 -7.71290183e-01 3.58069301e-01
-9.45672512e-01 -4.41243619e-01 5.65051973e-01 9.90598738e-01
7.58144319e-01 -1.57429338e-01 4.27392513e-01 -7.26299584e-01
-8.01938325e-02 -7.86850095e-01 -7.20649838e-01 -1.11671098e-01
-5.19261837e-01 -3.18726264e-02 6.14574134e-01 -1.38623446e-01
-6.95496559e-01 1.49020508e-01 -5.24132133e-01 -6.95729733e-01
-5.73751509e-01 2.42970325e-02 -2.48259440e-01 -5.16098440e-02
6.23565674e-01 3.19184005e-01 6.13525733e-02 -5.05951226e-01
-7.66473189e-02 4.65760320e-01 5.87949932e-01 -2.58471847e-01
7.25052953e-01 7.21934319e-01 -2.75222715e-02 -7.25287199e-01
-7.49931037e-01 -5.97484350e-01 -7.04517305e-01 -2.76838183e-01
7.06708908e-01 -8.39018047e-01 -3.01885933e-01 6.61554515e-01
-8.57511878e-01 -1.97090000e-01 -5.88547170e-01 3.09400350e-01
-5.08195639e-01 2.85465986e-01 -4.41120952e-01 -7.64505506e-01
-2.58722335e-01 -1.04927504e+00 9.71827805e-01 6.30792379e-02
-2.73979064e-02 -9.12231743e-01 3.81950252e-02 5.81694096e-02
4.48529631e-01 3.84206414e-01 7.36411273e-01 -1.24334192e+00
-6.86677873e-01 -6.25223756e-01 2.27594096e-02 9.13393676e-01
-1.41397417e-01 -7.33982027e-02 -1.26965451e+00 -3.17197949e-01
2.33968571e-01 -5.62072955e-02 1.22706735e+00 2.83119172e-01
1.72401142e+00 -2.93036371e-01 -2.26842120e-01 8.57915640e-01
1.33417392e+00 9.48109329e-02 9.65517879e-01 4.35524374e-01
6.75850093e-01 3.86885285e-01 4.61364627e-01 4.96451676e-01
-1.55143561e-02 5.93678355e-01 9.23405886e-01 -3.00932471e-02
6.26304373e-02 2.07365721e-01 6.06865168e-01 3.88640374e-01
-1.19500041e-01 -1.56014726e-01 -9.09378707e-01 5.96040905e-01
-1.94541204e+00 -1.04146767e+00 -6.81171776e-04 2.38572001e+00
2.58003384e-01 3.19187105e-01 5.82638085e-02 4.52488512e-01
6.44898355e-01 1.91143572e-01 -6.09386861e-01 -3.85512739e-01
-1.05991982e-01 3.02775264e-01 9.37897712e-02 2.06559300e-01
-1.56175387e+00 5.96047640e-01 5.63988972e+00 6.20111525e-01
-1.17243791e+00 1.39242820e-02 8.37635100e-01 -1.95885584e-01
1.67257428e-01 -1.92518502e-01 -5.69980621e-01 6.25693083e-01
9.18718457e-01 4.02870923e-01 2.26894036e-01 1.06756711e+00
5.11770602e-03 -8.18286017e-02 -1.34843302e+00 6.95451915e-01
2.45775074e-01 -1.09974802e+00 -7.09505677e-02 -7.97427967e-02
4.64989424e-01 2.21766785e-01 1.55365109e-01 3.34525377e-01
-3.24403286e-01 -1.02452230e+00 6.68802738e-01 5.54628313e-01
4.81138527e-01 -5.89079142e-01 8.26352000e-01 1.73840970e-01
-9.46828008e-01 -4.34684098e-01 -2.94605732e-01 -6.53821230e-02
-2.03721479e-01 6.94244504e-01 -5.51262498e-01 4.43305135e-01
8.88991773e-01 1.05387032e+00 -8.03207994e-01 1.13371861e+00
-2.85240322e-01 8.48316073e-01 -2.03933999e-01 5.93577266e-01
2.74361104e-01 -6.66894317e-02 9.62718248e-01 1.22464502e+00
3.76474142e-01 -3.23048592e-01 1.55946925e-01 8.30142736e-01
9.57569182e-02 1.24898084e-01 -6.22261226e-01 3.93262714e-01
-1.60120219e-01 1.16787767e+00 -6.21236920e-01 -3.96683186e-01
-7.38956988e-01 1.05917323e+00 3.10545504e-01 1.83640510e-01
-7.44671047e-01 -2.52692312e-01 7.46238947e-01 3.89351070e-01
5.63678145e-01 2.72139370e-01 -4.68789428e-01 -1.39991522e+00
3.96946162e-01 -7.81521976e-01 6.37499630e-01 -4.06483948e-01
-1.45970356e+00 7.59868503e-01 -3.64714831e-01 -1.65508354e+00
-2.04106629e-01 -8.14577818e-01 -9.13633049e-01 6.87673807e-01
-1.43962979e+00 -7.85522521e-01 -4.86150503e-01 8.20286870e-01
5.48768103e-01 -5.71951151e-01 9.12348807e-01 4.49096382e-01
-9.20018554e-01 6.83182418e-01 -2.73381434e-02 3.64472330e-01
5.61725140e-01 -1.39811730e+00 5.14404535e-01 1.41374290e+00
1.16364084e-01 1.56783164e-01 4.17996585e-01 -4.41587865e-01
-8.37825358e-01 -1.32623804e+00 4.13159192e-01 -4.63161051e-01
7.55587161e-01 -2.00347677e-01 -1.40641904e+00 8.14333975e-01
-4.65143658e-02 7.64038801e-01 4.14357901e-01 -2.06557110e-01
-4.13763553e-01 -2.41274416e-01 -1.24162507e+00 3.00697982e-01
8.91325593e-01 -2.07093641e-01 -5.93006611e-01 1.77975431e-01
1.36571467e-01 -3.63290876e-01 -5.16866326e-01 5.24269640e-01
2.17462823e-01 -1.35445321e+00 8.59151840e-01 -7.25222766e-01
4.85339075e-01 -4.30012554e-01 -2.07559124e-01 -1.19722092e+00
-1.39606223e-01 -3.40058088e-01 -6.38461709e-01 7.85074174e-01
3.50980431e-01 -8.37911487e-01 7.97331274e-01 2.37280756e-01
-5.40243387e-01 -1.00809383e+00 -1.09249759e+00 -6.86609387e-01
-1.62620485e-01 -5.95156491e-01 2.61195183e-01 7.42846847e-01
-7.29386136e-02 -1.84751824e-01 -3.01623583e-01 4.37627345e-01
5.38547873e-01 -1.89693600e-01 4.83573437e-01 -1.33001852e+00
-3.24249893e-01 -4.62536901e-01 -1.13147175e+00 -7.46820271e-01
7.81960506e-03 -8.81111026e-01 1.68856774e-02 -1.05689991e+00
-1.42422184e-01 -2.17827052e-01 -6.67843938e-01 6.70798421e-01
-8.38365033e-02 4.89629328e-01 -1.76439881e-01 2.11438730e-01
-6.22732282e-01 4.46693629e-01 6.86607540e-01 6.95695430e-02
-1.17123388e-01 3.31523955e-01 -3.32217813e-01 1.16512954e+00
9.80070055e-01 -4.42821681e-01 -2.60662660e-02 -2.62970150e-01
-1.08909853e-01 -3.82680118e-01 7.63146281e-01 -1.49831200e+00
1.18017100e-01 2.69273460e-01 6.57585382e-01 -4.54715371e-01
1.55754715e-01 -9.86638248e-01 -3.83223653e-01 5.91635823e-01
-2.07646117e-01 1.58032805e-01 1.53694525e-01 6.48414671e-01
-4.59259629e-01 -9.89378020e-02 9.94039059e-01 -4.32951152e-02
-8.81941319e-01 3.85093093e-01 -3.80482912e-01 -7.45984539e-02
1.33578658e+00 -1.90767288e-01 -1.58074960e-01 -1.17399357e-01
-9.90597367e-01 6.63966462e-02 3.70696574e-01 3.95071179e-01
7.46585429e-01 -1.19044566e+00 -7.37358034e-01 8.62627625e-01
3.14779431e-01 -3.86033789e-03 3.71731371e-01 1.09277105e+00
-5.62260747e-01 2.08212510e-01 -4.52299774e-01 -1.00042033e+00
-9.43100691e-01 6.28753662e-01 6.38595045e-01 -1.86081588e-01
-8.44718158e-01 7.21531570e-01 2.01408342e-01 1.41691631e-02
5.29712439e-01 -2.82193899e-01 -2.95376241e-01 -1.09774888e-01
7.79763937e-01 3.80569339e-01 3.56714725e-01 -6.36061966e-01
-4.38151389e-01 2.57578105e-01 -1.31751150e-01 2.35201582e-01
1.33638012e+00 -1.28898080e-02 -2.26604417e-01 4.99991596e-01
9.78559613e-01 -1.67945176e-01 -1.40687025e+00 -2.42545769e-01
1.48198366e-01 -5.25877595e-01 1.60587602e-03 -6.01056695e-01
-1.33287024e+00 8.73636484e-01 9.32632864e-01 2.89774776e-01
1.50470567e+00 -3.52225527e-02 5.15676379e-01 3.29752386e-01
4.77257408e-02 -8.37072074e-01 2.44328082e-01 4.14296448e-01
8.51219475e-01 -1.38838208e+00 -1.81856632e-01 -1.86328292e-01
-5.12754023e-01 1.03551912e+00 7.90803492e-01 -6.35350168e-01
8.62532735e-01 1.91014722e-01 -1.13213481e-02 -1.49438038e-01
-7.35496521e-01 -1.35176867e-01 5.68361223e-01 4.46029097e-01
3.21519077e-01 -1.25473991e-01 -2.71817427e-02 6.87915742e-01
2.35782117e-01 -3.82811695e-01 4.30346340e-01 8.93708050e-01
-5.29301345e-01 -8.64010215e-01 -3.88223857e-01 7.23579705e-01
-7.68438816e-01 1.22735156e-02 -3.89182605e-02 4.12267298e-01
3.67247969e-01 7.26601541e-01 5.57538390e-01 -3.12347054e-01
4.27908510e-01 3.19615036e-01 4.21517119e-02 -6.49740100e-01
-5.43031752e-01 8.18016194e-03 -3.05842519e-01 -1.01364946e+00
-8.78555104e-02 -8.24869871e-01 -1.02952194e+00 1.00470185e-01
-2.18810420e-02 -1.06467322e-01 4.14191782e-01 9.74341750e-01
3.28929901e-01 7.35686541e-01 6.78561687e-01 -8.48685443e-01
-3.80757183e-01 -9.39098299e-01 -4.98282671e-01 5.30171990e-01
7.66105354e-01 -6.26756728e-01 -5.01124918e-01 6.25178069e-02] | [7.696124076843262, 2.07959246635437] |
54c30fb4-d31c-4079-bafd-4e20bb83c642 | strategies-for-improving-low-resource-speech | 2306.00208 | null | https://arxiv.org/abs/2306.00208v1 | https://arxiv.org/pdf/2306.00208v1.pdf | Strategies for improving low resource speech to text translation relying on pre-trained ASR models | This paper presents techniques and findings for improving the performance of low-resource speech to text translation (ST). We conducted experiments on both simulated and real-low resource setups, on language pairs English - Portuguese, and Tamasheq - French respectively. Using the encoder-decoder framework for ST, our results show that a multilingual automatic speech recognition system acts as a good initialization under low-resource scenarios. Furthermore, using the CTC as an additional objective for translation during training and decoding helps to reorder the internal representations and improves the final translation. Through our experiments, we try to identify various factors (initializations, objectives, and hyper-parameters) that contribute the most for improvements in low-resource setups. With only 300 hours of pre-training data, our model achieved 7.3 BLEU score on Tamasheq - French data, outperforming prior published works from IWSLT 2022 by 1.6 points. | ['Alejandro Ciuba', 'Cecile Macaire', 'Tomas Pavlicek', 'Marek Sarvas', 'Santosh Kesiraju'] | 2023-05-31 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 9.94543210e-02 1.70535259e-02 -7.20205083e-02 -3.62083912e-01
-1.61531746e+00 -7.17575788e-01 7.23085880e-01 -3.30938876e-01
-5.98191619e-01 8.64053905e-01 5.72161078e-01 -8.12493443e-01
4.72697526e-01 3.38749923e-02 -7.66950965e-01 -3.59100580e-01
3.37618530e-01 8.20693195e-01 -2.17857122e-01 -4.57874566e-01
-1.20973289e-01 2.01217785e-01 -8.35091710e-01 5.33795714e-01
8.83789062e-01 3.56981695e-01 5.86995363e-01 8.37059081e-01
8.13079253e-02 6.03565753e-01 -7.49846339e-01 -7.30233908e-01
2.91429371e-01 -4.56524342e-01 -9.10056591e-01 -1.22596227e-01
1.92724094e-01 -1.15012363e-01 -4.34873313e-01 6.13046288e-01
1.07064736e+00 6.73338100e-02 4.68322039e-01 -6.63178205e-01
-5.75412452e-01 1.08143032e+00 -1.11402169e-01 4.59565401e-01
5.03499389e-01 1.03993282e-01 7.90260315e-01 -1.29863846e+00
5.47633290e-01 1.31261957e+00 4.40718591e-01 6.11871600e-01
-9.03666258e-01 -5.39012372e-01 -2.40458459e-01 5.95338009e-02
-1.51804793e+00 -1.16402221e+00 2.94256657e-02 1.66721269e-01
1.72111332e+00 2.93270439e-01 2.67070860e-01 1.47184277e+00
2.12470010e-01 7.87549317e-01 1.12386990e+00 -8.70705783e-01
-5.95997050e-02 4.54769582e-01 -4.30920005e-01 3.74708235e-01
-9.42119285e-02 9.30021554e-02 -7.41783679e-01 1.36474743e-01
3.37364763e-01 -7.84667492e-01 -1.70991704e-01 4.53929514e-01
-1.41058218e+00 5.77113748e-01 1.68499369e-02 5.80614388e-01
-2.50351012e-01 1.21750860e-02 5.18867552e-01 7.28046119e-01
4.66835350e-01 5.28034270e-01 -8.19409490e-01 -7.32685745e-01
-9.68348026e-01 -5.45865536e-01 7.94873178e-01 1.35170162e+00
3.92532259e-01 3.52303267e-01 -4.45779502e-01 1.11056566e+00
2.21104890e-01 1.03490424e+00 8.26502979e-01 -5.27761459e-01
1.25286400e+00 -5.20469248e-02 1.71026103e-02 -2.48622730e-01
1.82270426e-02 -7.89578259e-01 -4.13820684e-01 -7.49259889e-01
1.69452533e-01 -7.05745578e-01 -9.64180052e-01 1.47213316e+00
-2.74153352e-01 -5.02234064e-02 5.69775462e-01 6.55986190e-01
5.15701592e-01 9.30065453e-01 -2.49011591e-01 -3.33375841e-01
1.18352556e+00 -1.42235076e+00 -8.87798429e-01 -5.61652362e-01
9.63677883e-01 -1.43608940e+00 1.31760299e+00 -7.51096848e-03
-1.41151559e+00 -6.04739845e-01 -8.81222427e-01 1.54664084e-01
-1.21637546e-01 8.31251919e-01 1.26861691e-01 8.89605045e-01
-1.50531793e+00 3.00111681e-01 -9.28123713e-01 -6.78093016e-01
-2.01313287e-01 5.18618762e-01 -2.24806949e-01 -1.54854253e-01
-1.26659238e+00 1.27897429e+00 1.16160139e-01 1.38433740e-01
-1.07834637e+00 -2.13694766e-01 -6.38537288e-01 7.94544742e-02
1.33753687e-01 -4.38817650e-01 1.57258463e+00 -9.11435843e-01
-2.01953864e+00 5.79353034e-01 -5.82479119e-01 -5.62920272e-01
5.06525040e-01 -3.53919834e-01 -7.13217258e-01 -3.78563814e-02
-2.67273858e-02 6.03344738e-01 6.75800681e-01 -9.38653231e-01
-5.53773344e-01 9.19887722e-02 -2.08199307e-01 7.11792052e-01
-3.42370242e-01 6.10364079e-01 -7.41300941e-01 -5.59052825e-01
-1.31735310e-01 -1.11457956e+00 -8.78076181e-02 -1.05710030e+00
-2.52649426e-01 2.72672534e-01 4.54193592e-01 -1.14369702e+00
1.15239596e+00 -1.96554708e+00 1.92326292e-01 -8.49814340e-02
-5.51246881e-01 5.02153814e-01 -3.62511724e-01 7.29471803e-01
7.81309530e-02 3.13635200e-01 -4.81900349e-02 -6.92750871e-01
-1.59544155e-01 2.63688385e-01 -9.31764916e-02 1.87571943e-01
3.07691604e-01 1.08432686e+00 -5.81886590e-01 -3.22693229e-01
1.26705348e-01 6.78200841e-01 -2.27070212e-01 2.98272103e-01
1.82645455e-01 5.66577375e-01 -1.39614120e-01 4.41916615e-01
3.05058360e-01 9.10487697e-02 3.19001526e-01 1.52788609e-01
-2.15959713e-01 1.02298915e+00 -5.69333017e-01 1.87135458e+00
-9.66998875e-01 9.23973858e-01 -4.27948013e-02 -5.52170098e-01
1.04225945e+00 7.75314033e-01 1.18917458e-01 -1.11040771e+00
3.58814120e-01 6.69708371e-01 1.14750795e-01 -3.09475422e-01
6.50173068e-01 1.39917314e-01 -6.07655570e-02 5.28790057e-01
2.43136182e-01 -1.46468893e-01 1.25893727e-01 9.96620953e-02
9.22228694e-01 7.21753910e-02 -4.02230099e-02 -2.58196712e-01
6.48479760e-01 -1.78310350e-02 1.72611400e-01 4.72300291e-01
-8.62824023e-02 7.82300770e-01 -2.63323281e-02 8.24719958e-04
-1.27659416e+00 -7.73106456e-01 1.83795273e-01 9.92522418e-01
-4.07039255e-01 -5.03435552e-01 -1.11063683e+00 -6.19291365e-01
-6.86950386e-01 1.00685465e+00 -3.82501706e-02 -8.09838772e-02
-9.61101353e-01 -7.21550405e-01 1.08170605e+00 2.85552204e-01
5.14582872e-01 -7.35612273e-01 -4.57078516e-02 4.04546916e-01
-7.97254682e-01 -1.65452313e+00 -8.84250343e-01 4.18183833e-01
-7.84487188e-01 -3.32914978e-01 -9.44794416e-01 -9.62457359e-01
5.00940740e-01 2.91950583e-01 1.17931449e+00 -2.25646332e-01
2.39641458e-01 3.84519845e-02 -5.30536830e-01 -1.29092485e-01
-9.24917281e-01 5.84351778e-01 2.27174819e-01 -1.74456760e-01
3.22777539e-01 -4.58701178e-02 -1.23741843e-01 5.70784628e-01
-3.20062846e-01 9.93998498e-02 9.47389603e-01 8.41791749e-01
1.60860822e-01 -3.86334717e-01 4.61098194e-01 -4.39609200e-01
7.70927846e-01 -2.51888812e-01 -3.30420226e-01 6.58722222e-01
-6.06049895e-01 3.67667466e-01 7.29502439e-01 -3.88034135e-01
-1.15196598e+00 -9.38948616e-02 -4.16388273e-01 -2.84573406e-01
2.12071344e-01 5.59091508e-01 -1.76123008e-01 9.97264832e-02
7.14173794e-01 5.19409001e-01 -2.26234317e-01 -5.73226690e-01
1.70696482e-01 1.29326785e+00 1.82338253e-01 -6.54512107e-01
6.53490901e-01 -1.94779441e-01 -6.49733603e-01 -7.06146240e-01
-3.40381145e-01 -3.69638860e-01 -8.24831367e-01 2.54889037e-02
5.35179496e-01 -1.38878524e+00 -1.69493765e-01 1.77219927e-01
-1.19704187e+00 -5.05850077e-01 2.28293315e-02 8.39384913e-01
-5.23177981e-01 1.89683437e-01 -6.51759982e-01 -7.60971844e-01
-6.26792014e-01 -1.59223151e+00 1.20114875e+00 -2.39786372e-01
-1.22039571e-01 -9.45067048e-01 -1.46133145e-02 7.37085819e-01
6.58435345e-01 -6.93580866e-01 5.51843405e-01 -6.32435679e-01
-2.45261490e-01 7.62294605e-02 -1.71126705e-02 5.00684023e-01
1.63655937e-01 -1.47587240e-01 -9.67671752e-01 -7.32373714e-01
-1.72938138e-01 -2.62599558e-01 2.88474500e-01 2.21534874e-02
3.33270192e-01 -4.64046031e-01 3.92682515e-02 5.02979457e-01
1.09350741e+00 2.72514790e-01 6.40090287e-01 2.02895477e-01
4.39163923e-01 2.91114777e-01 6.88714147e-01 1.77712455e-01
4.58395392e-01 9.59692359e-01 -1.51044577e-01 -1.43607467e-01
-5.12051880e-01 -3.14670235e-01 1.18919361e+00 1.93646109e+00
8.79750121e-04 -6.37665689e-01 -1.05945945e+00 4.85617816e-01
-1.45738780e+00 -4.84644562e-01 1.73100166e-03 2.35171604e+00
9.45906281e-01 3.68161984e-02 -6.26998721e-03 -1.58628225e-01
7.62497127e-01 -2.06403032e-01 -3.57490331e-02 -9.13275063e-01
-3.16051692e-01 3.58752012e-01 8.38158965e-01 9.45037365e-01
-6.82008684e-01 1.62697852e+00 7.01295614e+00 8.62736106e-01
-1.29492927e+00 4.65795189e-01 5.65386355e-01 -1.73615381e-01
-1.92883298e-01 2.28998866e-02 -9.72768784e-01 3.27422440e-01
2.06100440e+00 -2.53925085e-01 8.34962249e-01 3.86896729e-01
5.18107951e-01 3.35295320e-01 -1.03467882e+00 8.52733254e-01
3.17921460e-01 -1.02925432e+00 1.39785647e-01 -9.56817046e-02
8.41081977e-01 6.97500110e-01 6.14731237e-02 5.60731053e-01
4.70044017e-01 -1.03063321e+00 9.36995804e-01 -3.18685062e-02
1.27120864e+00 -6.12004399e-01 1.02508771e+00 2.88845539e-01
-1.12137568e+00 1.19377479e-01 -4.07879233e-01 1.52890801e-01
1.36283010e-01 6.20789975e-02 -1.52474713e+00 8.64764392e-01
3.40434581e-01 5.72707653e-01 -6.46819532e-01 5.85543036e-01
-4.01821822e-01 1.08921134e+00 -4.25220132e-01 -1.03306949e-01
4.73225772e-01 8.34329203e-02 3.78314614e-01 1.69223332e+00
7.07261086e-01 -4.46177274e-01 -5.92672266e-02 2.12463617e-01
-2.49401778e-01 4.43498999e-01 -4.31382924e-01 -1.79665416e-01
6.87384188e-01 8.21934581e-01 -3.68864596e-01 -3.15126657e-01
-3.20531994e-01 1.23116374e+00 2.18979150e-01 5.19190431e-01
-9.12631214e-01 -9.49610025e-02 5.76095104e-01 -1.68877512e-01
2.05527157e-01 -6.16632164e-01 -7.97190741e-02 -1.41706800e+00
4.93205637e-02 -1.60442638e+00 -9.03385654e-02 -7.83472180e-01
-4.52783734e-01 1.15015507e+00 -2.83504575e-01 -1.25505972e+00
-6.11984432e-01 -3.96959007e-01 -1.20412238e-01 1.18406975e+00
-1.56017542e+00 -1.16634655e+00 3.14804226e-01 5.60266018e-01
1.06708658e+00 -5.64022720e-01 9.69080865e-01 6.12519860e-01
-6.91348910e-01 1.03279543e+00 4.90411371e-01 1.12287119e-01
8.67988586e-01 -7.91580200e-01 9.76914406e-01 1.12274480e+00
6.20310187e-01 6.46527648e-01 6.38989091e-01 -5.27048230e-01
-1.77730846e+00 -9.40926969e-01 1.49132919e+00 -5.00398636e-01
5.52946866e-01 -6.49279594e-01 -3.62312287e-01 6.72149599e-01
8.21849287e-01 -5.58633208e-01 4.07719463e-01 -1.01031289e-02
-6.52921274e-02 1.56556559e-03 -8.59814286e-01 7.13792801e-01
9.57128882e-01 -8.49958181e-01 -3.54122370e-01 4.19504821e-01
8.26022208e-01 -6.45861506e-01 -8.85655284e-01 1.31935984e-01
3.37700784e-01 -2.98086107e-01 7.29142547e-01 -4.67838496e-01
7.77909160e-02 -4.47827886e-04 -5.99669337e-01 -1.68488836e+00
1.20254443e-03 -1.24016261e+00 2.93808043e-01 1.06108451e+00
1.17906141e+00 -5.81539929e-01 4.18032020e-01 1.54638052e-01
-4.66919631e-01 -4.14412677e-01 -1.10442114e+00 -8.64399135e-01
1.90589145e-01 -4.31323975e-01 6.51851773e-01 6.19124055e-01
-1.72506332e-01 8.32579911e-01 -5.07206619e-01 1.23710804e-01
6.99847788e-02 -6.11006320e-01 8.29085410e-01 -4.49450761e-01
-3.20265740e-01 -1.36574447e-01 3.80861647e-02 -1.27065337e+00
8.15333501e-02 -1.03671813e+00 1.25917330e-01 -1.41660964e+00
-1.48458272e-01 -4.02605087e-01 -1.32765293e-01 5.48097849e-01
-4.96413372e-02 4.00383212e-02 3.74713659e-01 3.39955896e-01
-2.58061469e-01 5.81833422e-01 1.07736683e+00 -1.81464732e-01
-3.01515181e-02 -6.67238757e-02 -4.84075993e-01 -1.41273560e-02
8.92891645e-01 -6.17965877e-01 -4.52400863e-01 -1.21880102e+00
-3.75900269e-02 2.18733862e-01 -5.00356078e-01 -9.59867597e-01
2.10340545e-01 1.18306883e-01 -7.35460892e-02 -3.99960876e-01
6.00264907e-01 -8.20017278e-01 7.34071732e-02 3.91130298e-01
-4.28523928e-01 6.13533914e-01 4.16985840e-01 -1.04645036e-01
-3.22529286e-01 -2.25929365e-01 5.22764742e-01 -1.00909788e-02
-2.50266582e-01 -1.34604350e-01 -6.84562922e-01 2.12913528e-01
5.46803951e-01 -1.09172976e-02 -2.56570280e-01 -6.46776855e-01
-6.30314350e-01 8.23303238e-02 2.56758988e-01 7.23008513e-01
4.16748524e-01 -1.15319443e+00 -1.21888423e+00 4.11237121e-01
-1.22652955e-01 -7.79091418e-01 -3.44494194e-01 9.08415437e-01
-6.04686618e-01 1.02753353e+00 -4.04454172e-02 -5.81330895e-01
-1.33042824e+00 -7.18216524e-02 3.36754233e-01 -4.26428944e-01
-3.55741233e-01 7.79444814e-01 -5.29175222e-01 -7.79951870e-01
3.07448387e-01 -2.43539080e-01 1.17189633e-02 -1.93718433e-01
2.33242571e-01 2.11436778e-01 7.34109998e-01 -1.03381991e+00
-4.00275797e-01 3.13008964e-01 -2.01003104e-01 -8.11906517e-01
1.05863070e+00 -4.70142335e-01 3.48850220e-01 3.21807981e-01
1.28471291e+00 2.07166344e-01 -7.95980036e-01 -3.23443741e-01
1.57983184e-01 -3.17837834e-01 1.56200364e-01 -1.16164196e+00
-9.08739567e-01 1.05555141e+00 5.18018484e-01 -3.71998370e-01
9.10129130e-01 -3.07743639e-01 8.65708053e-01 7.00510263e-01
5.91627419e-01 -1.24578679e+00 -1.77488059e-01 1.04626751e+00
8.99778903e-01 -1.13976884e+00 -4.64947760e-01 -1.93561032e-01
-9.60588157e-01 9.71063972e-01 2.48243198e-01 4.64346260e-01
1.08926147e-01 5.85806906e-01 5.13821006e-01 4.68293369e-01
-1.02327764e+00 -9.70714763e-02 1.82825476e-01 4.51917470e-01
8.77060592e-01 2.70217746e-01 -1.99484810e-01 1.58348620e-01
-6.06842637e-01 -2.37852886e-01 4.90776420e-01 7.31684327e-01
-2.20809981e-01 -1.48778975e+00 -3.78014892e-01 4.03075553e-02
-6.89597547e-01 -6.69061482e-01 -5.86021066e-01 5.60248911e-01
-4.01261181e-01 1.44898760e+00 -1.91121593e-01 -6.54907942e-01
3.68874490e-01 2.97765553e-01 3.83384705e-01 -6.83331847e-01
-1.14768231e+00 3.77579063e-01 7.63804555e-01 -3.65123540e-01
-7.96919912e-02 -7.77867913e-01 -1.03329897e+00 -3.71647239e-01
-5.31831980e-01 5.27872324e-01 1.22282767e+00 9.44187224e-01
5.52754641e-01 5.56334496e-01 9.83472466e-01 -4.45710123e-01
-7.18903244e-01 -1.49140429e+00 2.24236131e-01 -1.47161245e-01
9.79383141e-02 -1.48802355e-01 -3.17760497e-01 5.61384484e-02] | [14.48227596282959, 7.160282611846924] |
5c483df3-da6f-4024-b758-6dd106cf89b2 | speechgen-unlocking-the-generative-power-of | 2306.02207 | null | https://arxiv.org/abs/2306.02207v2 | https://arxiv.org/pdf/2306.02207v2.pdf | SpeechGen: Unlocking the Generative Power of Speech Language Models with Prompts | Large language models (LLMs) have gained considerable attention for Artificial Intelligence Generated Content (AIGC), particularly with the emergence of ChatGPT. However, the direct adaptation of continuous speech to LLMs that process discrete tokens remains an unsolved challenge, hindering the application of LLMs for speech generation. The advanced speech LMs are in the corner, as that speech signals encapsulate a wealth of information, including speaker and emotion, beyond textual data alone. Prompt tuning has demonstrated notable gains in parameter efficiency and competitive performance on some speech classification tasks. However, the extent to which prompts can effectively elicit generation tasks from speech LMs remains an open question. In this paper, we present pioneering research that explores the application of prompt tuning to stimulate speech LMs for various generation tasks, within a unified framework called SpeechGen, with around 10M trainable parameters. The proposed unified framework holds great promise for efficiency and effectiveness, particularly with the imminent arrival of advanced speech LMs, which will significantly enhance the capabilities of the framework. The code and demos of SpeechGen will be available on the project website: \url{https://ga642381.github.io/SpeechPrompt/speechgen} | ['Hung-Yi Lee', 'Yuan-Kuei Wu', 'Kai-Wei Chang', 'Haibin Wu'] | 2023-06-03 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 1.29889414e-01 4.59648252e-01 2.30015181e-02 -2.56451726e-01
-1.05867147e+00 -5.45438826e-01 8.38501155e-01 -1.96837053e-01
-1.01364583e-01 6.94551945e-01 3.65784228e-01 -4.29960072e-01
8.42967182e-02 -5.54371357e-01 -3.56591105e-01 -6.99014246e-01
1.06419079e-01 4.79618847e-01 -8.15082118e-02 -5.67130446e-01
3.04867607e-02 1.57844290e-01 -1.54457927e+00 4.61113781e-01
9.30295706e-01 8.74392807e-01 6.54365599e-01 7.67863691e-01
-4.69467044e-01 8.40036213e-01 -1.05305779e+00 -2.36721084e-01
-1.97735578e-01 -5.47857583e-01 -8.79581094e-01 -1.10196114e-01
-1.56114265e-01 -1.10695966e-01 -2.50454783e-01 7.67205894e-01
8.52463663e-01 4.40949410e-01 4.04271185e-01 -1.28248286e+00
-5.60363531e-01 1.10547698e+00 1.86943620e-01 1.92369670e-01
4.67228800e-01 2.93863803e-01 1.02100146e+00 -8.91348541e-01
3.36958557e-01 1.26770115e+00 7.00992048e-02 9.57168937e-01
-8.18371356e-01 -6.47398233e-01 8.15300792e-02 8.66575018e-02
-1.24258351e+00 -9.35314476e-01 6.27318740e-01 -6.96711242e-02
1.28690195e+00 4.51200545e-01 2.89805919e-01 1.51691103e+00
-3.19929987e-01 1.26048350e+00 7.76407242e-01 -5.78020573e-01
3.38608861e-01 4.70673680e-01 -2.19592959e-01 1.91520020e-01
-4.56664205e-01 7.06026256e-02 -8.92197669e-01 -2.12128963e-02
6.17748499e-01 -6.54549360e-01 -3.31039846e-01 5.24553895e-01
-1.15604782e+00 7.99218595e-01 -1.37442872e-02 5.49847543e-01
-2.95735955e-01 -2.24085338e-02 3.04949254e-01 4.31385726e-01
6.55371368e-01 4.69288498e-01 -2.76611596e-01 -7.67286837e-01
-7.59990633e-01 4.06352103e-01 8.99685502e-01 1.21263516e+00
3.44650030e-01 5.48282981e-01 -1.38922110e-01 1.16143179e+00
2.45488420e-01 6.41176164e-01 8.70067000e-01 -7.61427581e-01
7.12928057e-01 3.63826007e-01 -3.03275920e-02 -4.03223813e-01
-2.67702490e-01 -5.20522952e-01 -7.56766498e-01 -1.76699117e-01
1.30179554e-01 -5.65208912e-01 -4.73818332e-01 1.77357209e+00
3.13076615e-01 1.43709570e-01 4.26270783e-01 7.60721266e-01
1.13818955e+00 1.22642446e+00 1.57495007e-01 -2.01312929e-01
9.96061742e-01 -8.86202753e-01 -8.17803442e-01 -4.03390884e-01
6.69881582e-01 -8.19520712e-01 1.24144399e+00 3.16476136e-01
-1.33121169e+00 -5.89521945e-01 -7.59334803e-01 1.68907121e-01
-3.00057322e-01 9.22541618e-02 5.11171162e-01 7.09359229e-01
-1.34728587e+00 1.19786963e-01 -6.97607458e-01 -3.07938486e-01
1.54288579e-02 4.38635081e-01 9.99889448e-02 3.19723010e-01
-1.59029973e+00 6.39740825e-01 3.70655507e-01 2.05313310e-01
-7.23194897e-01 -3.87380242e-01 -8.06300282e-01 9.28438678e-02
3.29667628e-01 -3.78537744e-01 1.73213053e+00 -8.43388975e-01
-2.04391170e+00 3.36751789e-01 -1.54608235e-01 -5.28265536e-01
3.20432305e-01 -4.90406416e-02 -5.41440845e-01 -6.28835410e-02
-2.12515935e-01 8.42696905e-01 7.49988794e-01 -1.04501069e+00
-7.70562887e-01 3.13904472e-02 4.01900895e-02 5.36540389e-01
-6.92491472e-01 3.16001177e-01 -1.51866883e-01 -7.39828169e-01
-4.18729275e-01 -8.61435652e-01 -2.34272823e-01 -7.41104960e-01
-2.42883146e-01 -5.80111980e-01 6.56071842e-01 -4.15931344e-01
1.44335568e+00 -2.15247917e+00 1.56043880e-02 -1.05850868e-01
-2.16513172e-01 5.50808549e-01 -3.87450606e-01 9.24988151e-01
2.11466864e-01 1.39558271e-01 -3.05307824e-02 -4.88422215e-01
3.12680900e-01 8.08971897e-02 -5.24149001e-01 -2.46859610e-01
2.66670048e-01 1.04275799e+00 -7.67226696e-01 -1.90330699e-01
2.97423154e-01 4.59224284e-01 -3.38378489e-01 4.28791821e-01
-4.99752194e-01 4.87828910e-01 -5.67395866e-01 3.31213504e-01
2.53739178e-01 -1.80308893e-01 -3.12035214e-02 4.77088511e-01
-3.00340027e-01 6.86000466e-01 -1.10612273e+00 1.56147695e+00
-5.44910014e-01 4.93129849e-01 2.80874074e-01 -8.74175727e-01
1.18677020e+00 9.09864783e-01 3.40821534e-01 -7.38037109e-01
6.99730590e-02 3.53046387e-01 6.41146526e-02 -4.24791515e-01
7.30726540e-01 -1.13955885e-01 -1.42623127e-01 5.05707085e-01
1.71604306e-01 -4.45020318e-01 2.24675164e-01 2.91837394e-01
8.78965318e-01 -1.36332467e-01 1.33289620e-01 2.63613425e-02
5.51248372e-01 -2.34112099e-01 -1.19550824e-02 6.68146551e-01
-8.70075170e-03 3.09367061e-01 2.21237749e-01 1.88196093e-01
-7.00835168e-01 -8.84028614e-01 6.02808632e-02 1.29515564e+00
-2.48773292e-01 -4.66854453e-01 -1.02442420e+00 -2.37264693e-01
-4.48306352e-01 1.03838885e+00 5.36371283e-02 -1.25487894e-01
-5.58841586e-01 -6.76568687e-01 7.67405748e-01 4.02939945e-01
2.55560577e-01 -1.82287049e+00 -1.97474822e-01 4.68125075e-01
-4.64646786e-01 -1.29274428e+00 -3.75403553e-01 1.52544335e-01
-7.28751540e-01 -2.49987766e-01 -8.88579786e-01 -9.02677238e-01
3.49307150e-01 2.24798657e-02 8.12381625e-01 -5.38177639e-02
1.50090694e-01 3.74316365e-01 -6.79561198e-01 -4.99457896e-01
-1.06761444e+00 5.31426966e-01 1.05755866e-01 -6.96134567e-02
7.90963024e-02 -6.06080413e-01 -2.96237022e-01 1.64436162e-01
-9.81463015e-01 3.17870736e-01 4.93451834e-01 5.97457588e-01
-2.47760024e-02 -1.56350769e-02 1.26713765e+00 -6.82148337e-01
1.12466288e+00 -5.68179548e-01 -4.09247398e-01 7.66405649e-03
-2.45904237e-01 -1.00840591e-01 8.58487844e-01 -3.45490396e-01
-1.41443408e+00 -1.79414853e-01 -7.56368399e-01 -7.02096000e-02
-4.14975137e-01 7.29316711e-01 -3.03550005e-01 3.39819759e-01
5.77498555e-01 4.49260712e-01 -5.94112612e-02 -3.85171205e-01
4.47072297e-01 1.27470994e+00 4.71709341e-01 -7.36118436e-01
5.59013486e-01 -1.97189316e-01 -6.94361508e-01 -1.23273587e+00
-3.49043131e-01 -4.52551156e-01 -8.90183225e-02 -2.77956128e-01
4.06172693e-01 -8.86477411e-01 -5.15990794e-01 5.41384518e-01
-1.19797909e+00 -7.89979517e-01 -2.32775673e-01 3.44940156e-01
-5.64983845e-01 1.69226587e-01 -8.11689913e-01 -1.11249220e+00
-6.57396793e-01 -1.10825002e+00 1.00780797e+00 3.65196198e-01
-4.87477511e-01 -1.10659075e+00 -1.62946671e-01 5.60939431e-01
6.53814197e-01 -4.64495987e-01 7.30957270e-01 -9.21518862e-01
-5.46619058e-01 -1.89550608e-01 3.01539510e-01 5.17612696e-01
1.83929756e-01 -1.22290336e-01 -1.17842913e+00 -2.23822773e-01
1.40728597e-02 -5.04034042e-01 2.38967076e-01 3.17660362e-01
8.51977706e-01 -4.62034374e-01 -3.56054679e-02 2.20705032e-01
7.05412805e-01 5.16455770e-01 4.39435184e-01 8.47020596e-02
2.95436502e-01 7.49911666e-01 4.49702889e-01 6.82134867e-01
3.55597883e-01 7.13179827e-01 1.17675796e-01 2.77590714e-02
-1.90030009e-01 -2.20595896e-01 7.41778612e-01 1.51082814e+00
3.16629224e-02 -7.54400373e-01 -8.81792545e-01 4.08426970e-01
-1.62164330e+00 -8.57124567e-01 -1.08584762e-03 2.09169650e+00
1.00805366e+00 5.16504794e-02 -1.33014210e-02 6.21089935e-02
5.35714924e-01 2.94149697e-01 -3.60708803e-01 -4.91155207e-01
-1.69667646e-01 3.51593822e-01 -2.40481034e-01 6.10061646e-01
-6.77480936e-01 1.18280935e+00 5.62959909e+00 1.21852994e+00
-1.15433598e+00 6.95152441e-03 5.04167199e-01 -2.25974824e-02
-4.16859120e-01 -2.09976658e-01 -1.11272788e+00 5.14484942e-01
1.54476285e+00 -6.26745582e-01 6.13548398e-01 7.49562383e-01
7.39472926e-01 2.19849512e-01 -8.88357341e-01 8.93138468e-01
-1.00021057e-01 -1.17157614e+00 5.41525632e-02 -7.08935857e-02
5.85403085e-01 1.25189736e-01 3.00005287e-01 6.47854447e-01
2.76870906e-01 -8.93912554e-01 7.64993072e-01 2.52270307e-02
7.87980378e-01 -7.50440240e-01 5.40695786e-01 7.63702810e-01
-1.12871110e+00 8.76993034e-03 -2.01002643e-01 -9.34547633e-02
4.69610006e-01 1.61052436e-01 -1.40113842e+00 4.47123736e-01
3.02475661e-01 3.29819918e-01 -3.27007413e-01 7.69415975e-01
-2.71609187e-01 1.04403937e+00 -2.09050149e-01 -4.12528276e-01
3.21512818e-01 -5.66918589e-02 6.49772465e-01 1.33356583e+00
5.30119061e-01 1.86282456e-01 8.16092342e-02 7.37523496e-01
-2.04458963e-02 4.46888864e-01 -4.43194389e-01 -4.62998837e-01
6.69995666e-01 1.15040410e+00 -4.69436049e-01 -3.66334796e-01
-3.29505503e-01 6.70396984e-01 9.93302613e-02 5.68055570e-01
-6.69528842e-01 -2.61062533e-01 4.99690145e-01 -1.50294071e-02
-5.11523224e-02 -1.87428162e-01 -1.29855275e-01 -9.52510297e-01
-3.85412667e-03 -1.33572757e+00 1.38802007e-01 -7.80290544e-01
-1.02051568e+00 1.00554097e+00 1.83956139e-02 -1.07162356e+00
-8.83878767e-01 -4.39241588e-01 -5.91249943e-01 9.39173043e-01
-9.72909689e-01 -9.83778417e-01 -5.88449538e-02 3.99502218e-01
1.23202395e+00 -4.40248609e-01 1.00691235e+00 2.65594989e-01
-4.76620466e-01 7.08140075e-01 1.33819774e-01 -9.20016989e-02
5.21898925e-01 -1.09833407e+00 7.66434193e-01 4.92062032e-01
4.10402000e-01 4.71906930e-01 7.26323962e-01 -3.60198379e-01
-1.31168592e+00 -9.81463850e-01 1.04509068e+00 -4.00962681e-01
7.44189382e-01 -6.54524744e-01 -9.24769759e-01 4.77247596e-01
4.75272477e-01 -6.60993278e-01 6.61063671e-01 -7.55350366e-02
3.47594172e-01 6.91213161e-02 -6.02048218e-01 8.65704417e-01
8.38721097e-01 -5.49007297e-01 -2.76909113e-01 2.59581000e-01
9.65534687e-01 -4.94606316e-01 -6.79229140e-01 2.02330276e-01
1.10637933e-01 -5.31473458e-01 7.00603604e-01 -2.42888927e-01
2.79761076e-01 5.09480666e-03 -4.03938740e-02 -1.58850825e+00
1.05771050e-01 -1.15243232e+00 -1.15165852e-01 1.65329587e+00
8.15091908e-01 -7.63467252e-01 6.07509553e-01 5.86723089e-01
-4.94969070e-01 -6.68869138e-01 -8.30846310e-01 -6.74350500e-01
1.93163007e-01 -7.29173005e-01 6.58539653e-01 6.36065841e-01
4.57386434e-01 5.39172232e-01 -3.52100730e-01 1.21684568e-02
1.30978376e-01 -2.18123287e-01 7.03945994e-01 -6.78360701e-01
-4.75478828e-01 -5.33182085e-01 1.21731274e-02 -1.36784673e+00
2.92442381e-01 -1.07076538e+00 2.52445191e-01 -1.45021284e+00
-2.51108378e-01 -4.45229858e-01 -6.27436442e-03 4.34322655e-01
-2.03642815e-01 -2.72437811e-01 3.78893793e-01 -8.21759030e-02
-3.99946421e-01 9.59794700e-01 1.19882154e+00 -2.26438232e-03
-4.58130002e-01 4.04592335e-01 -6.61919951e-01 3.66491258e-01
1.17406321e+00 -2.51426309e-01 -8.37640226e-01 -3.31718475e-01
-7.96930566e-02 5.28617263e-01 -2.69882008e-02 -9.90879893e-01
4.24194574e-01 -1.46933302e-01 -2.29535386e-01 -2.68683076e-01
6.36917889e-01 -2.03414425e-01 1.62931800e-01 2.18292791e-02
-6.26111984e-01 2.67429743e-02 4.44559902e-01 7.99702555e-02
-4.24280524e-01 -3.77885044e-01 2.83357710e-01 -8.51851851e-02
-7.19633698e-01 2.13296548e-01 -7.83346713e-01 1.38973415e-01
5.20923138e-01 1.96593925e-02 -2.77701586e-01 -8.20068300e-01
-6.73005223e-01 1.51917681e-01 -9.65283588e-02 8.14980388e-01
7.01263070e-01 -1.09163547e+00 -8.52039158e-01 2.31364533e-01
-7.81779960e-02 -2.20798030e-02 3.63011032e-01 5.54701209e-01
-3.15708923e-03 6.48812413e-01 2.68010855e-01 -3.53918970e-01
-1.16985655e+00 1.44660696e-01 1.46432027e-01 -9.18406844e-02
-5.22642672e-01 9.13242698e-01 3.33629906e-01 -5.10542750e-01
3.41535419e-01 -3.40548664e-01 -1.74970880e-01 -1.41978636e-01
6.48323059e-01 1.53779209e-01 1.46137729e-01 -5.66084266e-01
2.31438763e-02 -2.01526433e-01 -1.00709777e-02 -5.60192108e-01
1.16658044e+00 -2.23500848e-01 2.01627180e-01 5.36685944e-01
7.84871280e-01 -4.79883142e-02 -1.03468299e+00 -1.91516787e-01
1.27958640e-01 3.80873904e-02 -1.00257769e-01 -8.58595550e-01
-6.89529479e-01 8.48646700e-01 1.54562980e-01 4.61737454e-01
1.00195110e+00 2.41922572e-01 9.34312522e-01 3.24389547e-01
4.35434282e-01 -1.09956527e+00 1.09620385e-01 7.80054390e-01
1.12512112e+00 -1.03617656e+00 -7.80973494e-01 -1.95962325e-01
-9.01010990e-01 9.00494993e-01 6.74350142e-01 4.24159497e-01
2.86109686e-01 5.81789136e-01 3.58736098e-01 1.55195549e-01
-1.21982801e+00 -1.80629671e-01 1.18035525e-01 5.87677717e-01
8.40966642e-01 2.88239926e-01 -1.55849397e-01 7.73260832e-01
-8.04012895e-01 -1.14373468e-01 3.94460559e-01 6.36967540e-01
-5.58368206e-01 -1.37281132e+00 -3.18227082e-01 2.36240581e-01
-2.92231172e-01 -3.60431492e-01 -3.31919521e-01 4.43515152e-01
-3.06996584e-01 1.36289048e+00 -1.70741916e-01 -2.58184314e-01
3.60542059e-01 2.76496947e-01 -2.39339098e-02 -8.07400346e-01
-6.11513078e-01 2.85121232e-01 3.64678323e-01 -1.67524934e-01
-5.53041622e-02 -6.34408653e-01 -1.41991210e+00 -2.03011513e-01
-2.94324189e-01 5.54758847e-01 7.63792455e-01 8.34934652e-01
2.63916999e-01 4.77088809e-01 6.77296937e-01 -7.65637040e-01
-7.81706452e-01 -1.27739251e+00 -3.98938388e-01 2.38212775e-02
7.04143047e-02 -1.11213595e-01 -3.72381836e-01 1.31133273e-01] | [14.5322847366333, 6.887794494628906] |
7dd9b6dd-ec4f-4087-a910-787121bdcb6d | pct-net-full-resolution-image-harmonization | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Guerreiro_PCT-Net_Full_Resolution_Image_Harmonization_Using_Pixel-Wise_Color_Transformations_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Guerreiro_PCT-Net_Full_Resolution_Image_Harmonization_Using_Pixel-Wise_Color_Transformations_CVPR_2023_paper.pdf | PCT-Net: Full Resolution Image Harmonization Using Pixel-Wise Color Transformations | In this paper, we present PCT-Net, a simple and general image harmonization method that can be easily applied to images at full resolution. The key idea is to learn a parameter network that uses downsampled input images to predict the parameters for pixel-wise color transforms (PCTs) which are applied to each pixel in the full-resolution image. We show that affine color transforms are both efficient and effective, resulting in state-of-the-art harmonization results. Moreover, we explore both CNNs and Transformers as the parameter network and show that Transformers lead to better results. We evaluate the proposed method on the public full-resolution iHarmony4 dataset, which is comprised of four datasets, and show a reduction of the foreground MSE (fMSE) and MSE values by more than 20% and an increase of the PSNR value by 1.4dB while keeping the architecture light-weight. In a user study with 20 people, we show that the method achieves a higher B-T score than two other recent methods. | ['Björn Stenger', 'Mitsuru Nakazawa', 'Julian Jorge Andrade Guerreiro'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['image-harmonization'] | ['computer-vision'] | [ 3.44479591e-01 -2.76927084e-01 2.38586396e-01 -2.29195148e-01
-7.78269708e-01 -2.03377530e-01 3.81326526e-01 -3.42181742e-01
-5.70132494e-01 5.33907712e-01 1.40622780e-01 5.52249029e-02
2.08935086e-02 -8.28734696e-01 -8.96466315e-01 -7.23537982e-01
1.58203229e-01 -1.46526499e-02 3.18834484e-01 -4.62066442e-01
5.14604375e-02 3.92654628e-01 -1.42136371e+00 3.06622833e-01
9.66693461e-01 1.19263101e+00 7.24496767e-02 5.35554290e-01
3.09418082e-01 5.21204233e-01 -4.84169394e-01 -8.31007242e-01
5.43757796e-01 -3.30414772e-01 -6.61545038e-01 1.02144256e-01
7.97159374e-01 -3.26575786e-01 -3.09143215e-01 1.14275944e+00
1.04637873e+00 2.05519665e-02 2.70691574e-01 -9.08435285e-01
-7.60597825e-01 5.76385200e-01 -9.95704889e-01 5.26639558e-02
-4.24674116e-02 1.02895387e-01 8.30650806e-01 -7.73662686e-01
5.95260680e-01 1.26157820e+00 8.69461834e-01 2.75646776e-01
-1.55399013e+00 -8.35328162e-01 -2.94464976e-01 6.51948512e-01
-1.62335360e+00 -3.09886843e-01 6.79524004e-01 3.27190645e-02
7.73293734e-01 2.61579841e-01 7.15813935e-01 8.87596369e-01
2.60139257e-01 4.45504993e-01 1.34101498e+00 -4.11997408e-01
1.43173914e-02 -1.83903918e-01 -5.11355758e-01 5.88268399e-01
7.75654316e-02 9.87921655e-02 -5.79652846e-01 3.03138196e-01
9.57034230e-01 -2.92895615e-01 -4.15153891e-01 -3.09796095e-01
-1.28230906e+00 6.41934574e-01 8.84803355e-01 2.82835007e-01
-5.01921713e-01 4.09766316e-01 3.24375898e-01 1.65177554e-01
4.13295686e-01 3.76285791e-01 -1.28864869e-01 1.59695044e-01
-8.84064078e-01 2.53647208e-01 3.09850186e-01 6.78016961e-01
5.85316181e-01 6.74047694e-02 -4.38177168e-01 1.11003244e+00
-2.86493212e-01 5.37940860e-01 2.57581234e-01 -1.34790492e+00
4.03091818e-01 3.48325819e-01 1.17703855e-01 -1.21471965e+00
-3.98889840e-01 -6.59506500e-01 -1.25474846e+00 3.86494845e-01
2.30532154e-01 5.33590280e-02 -8.38214040e-01 1.71504748e+00
9.07967314e-02 2.85721064e-01 -5.39886132e-02 1.05825686e+00
6.73969209e-01 5.71134925e-01 -1.12039700e-01 -1.51517376e-01
1.54118264e+00 -1.12900162e+00 -8.35998476e-01 -8.57082382e-02
-1.06274195e-01 -8.68306398e-01 1.19450152e+00 6.10867381e-01
-1.42675245e+00 -8.40969980e-01 -1.24660110e+00 -2.09864825e-01
-1.37877941e-01 1.25595063e-01 2.70038575e-01 7.66410947e-01
-1.33976567e+00 8.77009094e-01 -6.13102853e-01 -3.07368249e-01
4.42425638e-01 3.69519204e-01 -2.66371369e-01 -5.88386245e-02
-1.13668168e+00 9.01568711e-01 2.65992671e-01 1.11329392e-01
-5.06075442e-01 -8.32479954e-01 -5.49248636e-01 1.68631718e-01
3.56216133e-01 -8.19613934e-01 1.00357687e+00 -9.74087656e-01
-1.56200278e+00 8.36784959e-01 -2.28944551e-02 -7.42130399e-01
8.15329552e-01 -2.78658658e-01 -4.71577346e-01 2.39833847e-01
-1.23359263e-01 9.13081229e-01 9.58562255e-01 -1.20324850e+00
-8.12959194e-01 -1.58079714e-01 2.48787589e-02 3.11933339e-01
-5.45732200e-01 2.01048888e-02 -1.02903509e+00 -9.14510846e-01
-8.06749836e-02 -9.00963783e-01 -2.65591592e-01 1.21179320e-01
-4.61921364e-01 4.12750572e-01 5.95109940e-01 -1.06528056e+00
1.08430791e+00 -2.08378315e+00 2.01503947e-01 1.56127393e-01
2.60943830e-01 2.76389688e-01 -1.79550529e-01 -1.86199501e-01
-1.44171298e-01 1.17545791e-01 -2.65560597e-01 -3.26057762e-01
-6.71429634e-02 8.12882334e-02 -1.38159022e-01 4.22132254e-01
9.37297121e-02 1.00636947e+00 -3.65111083e-01 -4.74066049e-01
3.49478483e-01 1.06202030e+00 -5.85763276e-01 5.30565083e-02
2.24518329e-01 3.56963307e-01 4.89780456e-02 3.53635728e-01
9.73514080e-01 -2.20921814e-01 1.87221467e-01 -7.74360478e-01
-1.23332255e-01 -1.97230726e-01 -1.24818909e+00 1.66288364e+00
-5.22046208e-01 8.46786976e-01 -9.51322019e-02 -5.84606469e-01
8.56434584e-01 9.11890641e-02 3.52316737e-01 -1.39986324e+00
1.73118621e-01 7.07305297e-02 -1.39559880e-01 -1.01161540e-01
6.20973647e-01 1.51893288e-01 9.10898075e-02 2.23890185e-01
-1.92495838e-01 2.42861547e-02 3.09086949e-01 -1.99472129e-01
6.80565357e-01 -6.86657280e-02 8.44613537e-02 -1.54133037e-01
5.90772152e-01 -3.84481311e-01 4.40564841e-01 7.04649866e-01
1.03451081e-01 9.88680720e-01 4.44041312e-01 -6.00428760e-01
-1.43397987e+00 -1.06109166e+00 -6.02587648e-02 9.31749940e-01
2.15356857e-01 -3.08618605e-01 -9.33829665e-01 -1.96005702e-02
-3.87347817e-01 5.19496024e-01 -7.48798192e-01 4.29889485e-02
-8.09924304e-01 -8.63830447e-01 4.36111897e-01 5.11629701e-01
1.24089766e+00 -9.03663218e-01 -8.76262069e-01 -5.03981626e-03
-5.71834683e-01 -1.50239050e+00 -6.01393580e-01 -1.87609047e-01
-6.83946669e-01 -6.74528301e-01 -1.03831673e+00 -6.59541845e-01
4.55212116e-01 1.07068509e-01 1.05106556e+00 7.53800198e-02
-3.99218529e-01 -1.70600787e-01 -2.07627356e-01 -6.44730553e-02
-2.79830098e-01 1.10096775e-01 -2.33665794e-01 1.48676679e-01
-3.06459993e-01 -6.29896104e-01 -9.52227533e-01 4.95703101e-01
-8.60218048e-01 4.00925010e-01 6.55996144e-01 7.83010125e-01
8.48651230e-01 3.25213611e-01 2.04182230e-02 -6.26506805e-01
5.41795850e-01 2.94428527e-01 -7.10064590e-01 3.19861829e-01
-8.55671406e-01 -2.82073356e-02 6.52852952e-01 -4.25566435e-01
-1.10331643e+00 1.27835609e-02 -8.09733495e-02 -4.87675995e-01
2.87827671e-01 1.13373706e-02 7.19803870e-02 -4.08395827e-01
6.90647244e-01 2.11461112e-01 -1.56350911e-01 -4.19497967e-01
5.87348342e-01 2.34832197e-01 1.05993021e+00 -2.66600847e-01
9.52172756e-01 6.97241247e-01 1.32145017e-01 -5.38197577e-01
-5.34240484e-01 6.07171729e-02 -4.50293690e-01 -1.43827274e-01
1.00185835e+00 -1.06369257e+00 -7.86749005e-01 5.56308687e-01
-8.79316986e-01 -4.92281139e-01 -1.79508656e-01 1.70690984e-01
-5.45080960e-01 3.33980650e-01 -6.61498129e-01 -2.47707441e-01
-7.53468692e-01 -1.30017149e+00 7.22624302e-01 4.04006004e-01
1.17174454e-01 -5.78162849e-01 -3.45444441e-01 3.06332409e-01
8.95424962e-01 3.63223612e-01 8.06443989e-01 1.82653628e-02
-5.35135210e-01 2.82924175e-01 -5.73415875e-01 3.89425784e-01
-8.44580680e-02 1.31031191e-02 -8.37103605e-01 -4.73491728e-01
-2.71023154e-01 -1.78727265e-02 8.83639812e-01 6.06578171e-01
1.36888087e+00 -2.63691872e-01 7.71516338e-02 1.13717496e+00
1.67861891e+00 2.88184397e-02 1.18645215e+00 6.64598048e-01
6.23848736e-01 2.83139974e-01 3.87777567e-01 4.08757031e-01
3.01496416e-01 1.17724013e+00 4.38420743e-01 -7.88730145e-01
-5.90775549e-01 2.01967042e-02 1.66024089e-01 4.63387072e-01
-4.44199562e-01 -2.12571081e-02 -5.68552256e-01 4.22144741e-01
-1.62406218e+00 -9.51694846e-01 -5.11076562e-02 2.09266210e+00
9.88730311e-01 -6.66525736e-02 1.78902090e-01 5.37965372e-02
8.96433651e-01 2.07001895e-01 -5.57084382e-01 -2.54777730e-01
-4.23383862e-01 6.81717396e-01 7.57534385e-01 2.91171581e-01
-1.12034452e+00 9.10803616e-01 6.72781277e+00 8.63802671e-01
-1.46323228e+00 1.07973412e-01 1.02454650e+00 -3.08295310e-01
1.18122660e-01 -3.92097592e-01 -2.37630650e-01 3.11037570e-01
7.48539507e-01 -1.69049263e-01 8.02139938e-01 5.14217615e-01
2.37492055e-01 6.30919412e-02 -6.00528598e-01 1.18263733e+00
1.00225732e-01 -1.53077352e+00 7.64965415e-02 -4.65455428e-02
9.05745804e-01 -1.33924022e-01 4.91825640e-01 -6.78665116e-02
2.75883973e-02 -1.22593176e+00 8.52108002e-01 3.14769864e-01
1.09923995e+00 -1.11189246e+00 7.39255011e-01 -3.73110950e-01
-1.20339000e+00 -1.21567495e-01 -4.61711943e-01 4.51241225e-01
2.16657743e-01 3.78466278e-01 -3.92769039e-01 6.02261007e-01
1.40070879e+00 4.53048915e-01 -8.12901497e-01 1.06578827e+00
-3.42794120e-01 2.50104666e-01 -8.64466056e-02 4.53935564e-01
-8.86363536e-02 -2.15566412e-01 2.07813114e-01 1.18408549e+00
5.06260514e-01 3.34131829e-02 -3.42598617e-01 9.80923116e-01
-2.97605336e-01 1.86206073e-01 1.17271237e-01 5.04343748e-01
3.25646073e-01 1.44674122e+00 -7.04590261e-01 -2.43499249e-01
-1.89186022e-01 1.33449578e+00 6.44649863e-02 3.87888879e-01
-1.11192179e+00 -4.31471407e-01 5.86205602e-01 4.40000668e-02
8.05193543e-01 1.20647140e-01 -2.87380010e-01 -7.59988666e-01
1.42863244e-01 -1.14411294e+00 1.75141245e-01 -1.07036650e+00
-9.74374115e-01 7.46382952e-01 -9.68051925e-02 -1.18717778e+00
-1.13634184e-01 -4.84574944e-01 -3.39505225e-01 9.13596630e-01
-1.70924699e+00 -1.17663181e+00 -8.22703123e-01 8.10652018e-01
2.43208975e-01 -2.40156930e-02 5.81152558e-01 6.41264975e-01
-5.98214984e-01 8.28083634e-01 3.89031991e-02 2.22158656e-01
9.12971437e-01 -1.02563608e+00 6.91745698e-01 1.00916290e+00
2.04522759e-02 4.21195149e-01 8.84416759e-01 -1.83456674e-01
-1.19613636e+00 -1.14180708e+00 6.01960719e-01 1.58071324e-01
3.50367069e-01 -1.82232887e-01 -7.64496088e-01 4.15047079e-01
6.01625860e-01 1.80751219e-01 1.29371017e-01 -3.80161881e-01
-5.23427427e-01 -4.70733762e-01 -1.25786829e+00 8.53224277e-01
9.87544358e-01 -2.06848413e-01 -8.91842395e-02 -1.35078996e-01
8.45627785e-01 -7.19540715e-01 -9.46912289e-01 4.08264399e-01
8.18909824e-01 -1.48380685e+00 1.37403750e+00 5.51270135e-02
6.68075860e-01 -4.58985835e-01 -2.07563385e-01 -1.34645140e+00
-5.84887028e-01 -5.95976770e-01 2.47853518e-01 1.02235043e+00
3.56907845e-01 -5.67286849e-01 4.23277348e-01 3.36770415e-01
1.89917609e-01 -7.02120066e-01 -9.83296037e-01 -5.38154423e-01
-1.33334428e-01 -1.92754447e-01 6.13700807e-01 7.17310190e-01
-6.87254548e-01 1.67535469e-01 -6.27063274e-01 6.05289713e-02
8.58578503e-01 -1.02598928e-02 7.21162379e-01 -7.74668753e-01
-2.33414799e-01 -5.43144047e-01 -2.79634029e-01 -7.49723852e-01
-1.95015892e-01 -4.34023827e-01 -1.42179504e-01 -1.38049698e+00
4.99098063e-01 -9.51954052e-02 -3.39903414e-01 5.71502209e-01
-1.15703657e-01 1.00186980e+00 4.35588092e-01 -2.73565929e-02
-3.67095262e-01 4.22261029e-01 1.27490485e+00 -2.51170605e-01
-1.30271345e-01 -3.47352237e-01 -8.44243944e-01 5.54687917e-01
7.78116405e-01 -1.63998723e-01 -2.69472599e-01 -5.08683324e-01
2.32276589e-01 -1.06475629e-01 4.94209170e-01 -1.32774377e+00
1.72054991e-01 1.00862369e-01 7.80703843e-01 -5.43732405e-01
4.81448859e-01 -6.94167793e-01 5.51367581e-01 6.22110665e-01
-4.15597409e-01 2.91832924e-01 3.44184339e-01 2.08512887e-01
-8.46525282e-02 3.18052739e-01 1.23612988e+00 5.81472144e-02
-6.57994688e-01 2.01804757e-01 1.95001483e-01 -1.56558052e-01
8.44901741e-01 -2.26982221e-01 -4.54415321e-01 -6.54861093e-01
-4.05202925e-01 -1.24475256e-01 5.82298458e-01 3.48498344e-01
6.11073256e-01 -1.43902493e+00 -1.05421090e+00 1.51529625e-01
-1.28478006e-01 -3.35533857e-01 5.57508588e-01 9.63720441e-01
-7.79545069e-01 7.26221502e-02 -7.28787303e-01 -5.34637213e-01
-1.49967837e+00 3.83250594e-01 5.50947428e-01 -2.89715469e-01
-8.69529307e-01 6.34768784e-01 8.29820111e-02 -9.73431394e-02
1.85586944e-01 -1.43463537e-01 -2.35626116e-01 -1.17597796e-01
6.34700835e-01 3.94010723e-01 1.26801297e-01 -7.37960339e-01
-1.95914686e-01 1.04162765e+00 2.67399708e-03 -3.09506327e-01
1.56867993e+00 -2.57115632e-01 -2.62465447e-01 -1.46533519e-01
1.18450058e+00 -9.51597020e-02 -1.38151336e+00 -4.43361580e-01
-4.79106665e-01 -6.93867683e-01 1.90285563e-01 -9.06398475e-01
-1.76052785e+00 6.83155298e-01 1.26430976e+00 -6.41147494e-02
1.75255561e+00 -2.75430262e-01 1.00039291e+00 2.14635916e-02
1.80249900e-01 -1.10416257e+00 2.64312893e-01 1.48922339e-01
1.00234687e+00 -8.59026432e-01 2.37083718e-01 -3.63832533e-01
-6.66470110e-01 9.68510032e-01 4.35972124e-01 -8.77707899e-02
1.85988441e-01 2.44180366e-01 6.61166459e-02 9.46226493e-02
-4.32887793e-01 -1.36685342e-01 4.26057994e-01 5.56518257e-01
2.59438753e-01 -1.26533508e-02 -2.64841497e-01 2.76526034e-01
-5.35372257e-01 -7.95625150e-02 3.74596626e-01 3.11358184e-01
-1.86284050e-01 -9.60239291e-01 -5.11985123e-01 4.51977067e-02
-6.59266114e-01 -1.55547157e-01 -7.43166730e-02 8.59250247e-01
2.20510095e-01 8.18161786e-01 1.87731370e-01 -5.33019423e-01
6.50428653e-01 -4.29479748e-01 6.50441408e-01 2.54309833e-01
-6.69182658e-01 1.61426142e-01 -1.02632768e-01 -9.97971594e-01
-6.18836582e-01 -2.59784847e-01 -1.05058253e+00 -6.78535700e-01
9.06970650e-02 -4.65951920e-01 6.77931011e-01 5.13272524e-01
3.19344550e-01 7.97354758e-01 5.55387259e-01 -9.95078087e-01
-2.51314580e-01 -6.65388644e-01 -5.35405815e-01 5.78280509e-01
2.49731988e-01 -4.19094563e-01 -6.56163916e-02 2.30689764e-01] | [11.029041290283203, -1.8954800367355347] |
b888c6c1-6dd0-48ec-ae82-893f6949cec8 | pixel-level-cycle-association-a-new | 2011.00147 | null | https://arxiv.org/abs/2011.00147v1 | https://arxiv.org/pdf/2011.00147v1.pdf | Pixel-Level Cycle Association: A New Perspective for Domain Adaptive Semantic Segmentation | Domain adaptive semantic segmentation aims to train a model performing satisfactory pixel-level predictions on the target with only out-of-domain (source) annotations. The conventional solution to this task is to minimize the discrepancy between source and target to enable effective knowledge transfer. Previous domain discrepancy minimization methods are mainly based on the adversarial training. They tend to consider the domain discrepancy globally, which ignore the pixel-wise relationships and are less discriminative. In this paper, we propose to build the pixel-level cycle association between source and target pixel pairs and contrastively strengthen their connections to diminish the domain gap and make the features more discriminative. To the best of our knowledge, this is a new perspective for tackling such a challenging task. Experiment results on two representative domain adaptation benchmarks, i.e. GTAV $\rightarrow$ Cityscapes and SYNTHIA $\rightarrow$ Cityscapes, verify the effectiveness of our proposed method and demonstrate that our method performs favorably against previous state-of-the-arts. Our method can be trained end-to-end in one stage and introduces no additional parameters, which is expected to serve as a general framework and help ease future research in domain adaptive semantic segmentation. Code is available at https://github.com/kgl-prml/Pixel- Level-Cycle-Association. | ['Alexander G. Hauptmann', 'Yueting Zhuang', 'Yi Yang', 'Yunchao Wei', 'Guoliang Kang'] | 2020-10-31 | null | http://proceedings.neurips.cc/paper/2020/hash/243be2818a23c980ad664f30f48e5d19-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/243be2818a23c980ad664f30f48e5d19-Paper.pdf | neurips-2020-12 | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 3.40923518e-01 2.30819017e-01 -4.14844990e-01 -5.04409134e-01
-9.42757726e-01 -5.25683045e-01 4.01975363e-01 -1.88749701e-01
-4.47699070e-01 7.26534188e-01 -3.17038149e-01 -1.60207152e-01
2.48400196e-01 -7.85621464e-01 -7.88206637e-01 -7.12804317e-01
4.44363177e-01 3.43999326e-01 5.66366374e-01 -3.95780683e-01
6.58973902e-02 1.16876408e-01 -1.02372658e+00 1.28923565e-01
1.23689926e+00 1.04918814e+00 4.93198812e-01 5.52213490e-02
-1.35637512e-02 3.81161541e-01 -3.73655111e-01 -4.58253890e-01
5.03672719e-01 -4.74308819e-01 -6.84539020e-01 1.19892284e-01
5.04695535e-01 -1.86046124e-01 -2.85251051e-01 1.26046515e+00
5.15884817e-01 1.84077665e-01 5.37519217e-01 -1.11528015e+00
-6.30667448e-01 3.00360411e-01 -7.71992683e-01 1.65026203e-01
-8.76950994e-02 2.36392722e-01 9.83486414e-01 -7.21772671e-01
6.47713304e-01 9.80132163e-01 5.29755056e-01 7.16117799e-01
-1.20112145e+00 -9.45446372e-01 7.10314572e-01 3.25734317e-01
-1.37116253e+00 -1.42036170e-01 1.15963626e+00 -3.34663361e-01
5.91090620e-01 -8.36324468e-02 5.25801957e-01 1.21592629e+00
-2.59399742e-01 1.03678131e+00 1.10476577e+00 -3.89780819e-01
3.13574672e-01 1.15145095e-01 -1.53146103e-01 6.34713411e-01
1.84824166e-03 2.22569928e-02 -3.74310464e-01 2.57537216e-01
8.86224329e-01 -1.77609265e-01 -3.00182998e-01 -6.74839854e-01
-1.00426996e+00 8.65455210e-01 6.29655957e-01 3.46061051e-01
-4.11980636e-02 6.07097782e-02 3.88419449e-01 9.63801518e-02
7.17365384e-01 1.71739951e-01 -6.63763225e-01 4.48245518e-02
-9.48377311e-01 2.53624320e-01 3.29481393e-01 1.11032867e+00
8.63196373e-01 1.85816750e-01 2.54714191e-02 1.13927770e+00
9.59992781e-02 3.87048364e-01 4.76055682e-01 -1.05935180e+00
6.44338548e-01 4.96144921e-01 3.85528430e-02 -7.40840852e-01
-1.99199483e-01 -6.34342015e-01 -6.21316731e-01 3.01634461e-01
6.45573437e-01 -1.40666991e-01 -1.24385238e+00 1.92957056e+00
4.55361307e-01 5.22710443e-01 -1.77243631e-02 1.14489985e+00
5.54170966e-01 5.46139479e-01 2.70714819e-01 1.48245379e-01
1.17546928e+00 -1.24331260e+00 -4.01470661e-01 -7.12565422e-01
5.89651763e-01 -6.60316288e-01 1.40287995e+00 2.64544696e-01
-1.01833212e+00 -7.53131151e-01 -1.02053499e+00 -7.41007999e-02
-3.86429489e-01 7.09639043e-02 4.15007651e-01 5.65646887e-01
-7.22083330e-01 5.55932879e-01 -9.81118739e-01 -1.91393510e-01
7.04574943e-01 2.42018312e-01 -6.49126023e-02 -9.03784484e-02
-1.34292781e+00 7.44196177e-01 6.16761267e-01 -1.42155021e-01
-7.72817731e-01 -9.34433103e-01 -9.16138470e-01 -2.40200967e-01
6.03919268e-01 -6.44051552e-01 1.36399782e+00 -1.43367064e+00
-1.53363025e+00 1.12327349e+00 -7.22916201e-02 -4.47271973e-01
7.38147795e-01 -2.05215588e-01 -3.05164039e-01 1.04669124e-01
4.04292554e-01 1.02295852e+00 7.82476604e-01 -1.36726820e+00
-9.29828525e-01 -4.20017391e-01 -5.43225789e-03 3.46959144e-01
-3.99619401e-01 -3.78190517e-01 -9.55891311e-01 -1.02039218e+00
1.91155896e-01 -9.77295339e-01 -4.47207212e-01 1.38448298e-01
-3.05413365e-01 -1.90436505e-02 8.10559750e-01 -7.64123738e-01
8.90526772e-01 -2.12139368e+00 -1.22630103e-02 3.73826176e-02
-1.16070069e-01 4.55953538e-01 -1.01056404e-01 -1.12031605e-02
-1.46434009e-01 -4.85797673e-02 -8.16757739e-01 -3.13100159e-01
-8.31499025e-02 1.44749358e-01 -8.81756246e-02 5.21354854e-01
2.79935688e-01 8.93202245e-01 -9.16870892e-01 -5.76540887e-01
4.33629155e-01 4.06820148e-01 -5.32753050e-01 -1.39375422e-02
-5.52591562e-01 7.90051222e-01 -7.28272557e-01 5.87592006e-01
8.57820928e-01 -1.62152365e-01 5.33903949e-02 -1.42521679e-01
7.92222843e-02 2.10068285e-01 -1.07239294e+00 2.24076247e+00
-4.58674550e-01 4.75737303e-01 1.28158197e-01 -1.41234612e+00
9.53135014e-01 -5.76354451e-02 5.42846918e-01 -1.22436094e+00
4.56053875e-02 3.47622007e-01 -1.35219485e-01 -1.03954367e-01
1.94005787e-01 -2.34024107e-01 -2.18621239e-01 -2.21709818e-01
-7.65388757e-02 -3.00702512e-01 1.33707643e-01 2.26409943e-03
6.71635866e-01 5.06212473e-01 2.44510844e-01 -2.73224533e-01
6.01969481e-01 1.68714359e-01 1.01625121e+00 4.05886412e-01
-3.95986855e-01 8.40772808e-01 3.34789187e-01 -1.31565914e-01
-9.79830027e-01 -1.20158207e+00 -5.27535193e-02 1.03221345e+00
6.53896689e-01 1.21960171e-01 -1.00333929e+00 -1.02564466e+00
-1.28002703e-01 9.96954381e-01 -4.45568949e-01 -1.50356382e-01
-6.84690237e-01 -4.79669333e-01 4.42680985e-01 6.90005362e-01
8.69270325e-01 -8.80989254e-01 -3.26079696e-01 1.65271237e-01
-3.36539060e-01 -1.29628825e+00 -6.29943550e-01 1.57404616e-02
-9.85968590e-01 -7.65590131e-01 -1.14502668e+00 -1.08480906e+00
7.08157539e-01 1.71440884e-01 1.09787726e+00 -3.51829290e-01
-1.96425449e-02 9.04377028e-02 -3.82548660e-01 -3.46849918e-01
-1.68316022e-01 2.30542138e-01 -4.23509538e-01 -1.31005526e-01
2.94783145e-01 -6.02638543e-01 -9.62834477e-01 6.39534652e-01
-9.09255326e-01 1.79828957e-01 4.83620852e-01 8.51776958e-01
1.05602944e+00 3.46097946e-02 6.50858343e-01 -1.12016881e+00
1.33401304e-01 -4.26217258e-01 -6.38790131e-01 1.68805480e-01
-6.00560069e-01 -1.30826965e-01 8.74763727e-01 -4.09560770e-01
-1.33402956e+00 4.09822881e-01 -3.42374593e-01 -5.20028234e-01
-3.58682543e-01 1.72611922e-01 -5.99637389e-01 1.05642721e-01
6.67959154e-01 3.90296578e-01 -2.57366121e-01 -4.59032327e-01
4.64433640e-01 3.97650868e-01 5.34542084e-01 -6.18557334e-01
7.48200595e-01 7.27989018e-01 -2.44120017e-01 -6.34783506e-01
-7.96480298e-01 -6.50686026e-01 -4.94322121e-01 -3.90758179e-02
8.18227589e-01 -1.21119702e+00 -9.28547457e-02 5.69860756e-01
-8.20501029e-01 -7.35434890e-01 -3.69042009e-01 2.64652550e-01
-6.81559682e-01 4.85776901e-01 -2.57024139e-01 -4.26071614e-01
-1.52518451e-01 -1.06259084e+00 1.11103094e+00 3.81152630e-01
1.86554883e-02 -1.16617668e+00 -1.20376125e-01 5.56375742e-01
1.69880867e-01 3.16578776e-01 6.49356246e-01 -5.03058612e-01
-3.80872577e-01 1.35228425e-01 -3.83958519e-01 6.29162848e-01
8.68468285e-02 -3.55423480e-01 -9.14318740e-01 -1.95671558e-01
-8.62032324e-02 -1.63871214e-01 1.08982420e+00 5.24478316e-01
1.38916481e+00 -2.25328305e-03 -3.88253540e-01 7.12583184e-01
1.57278419e+00 2.76963174e-01 7.49388874e-01 4.73013997e-01
7.43328631e-01 5.65732777e-01 1.13884854e+00 3.29090506e-01
4.82923031e-01 9.07496870e-01 5.44922173e-01 -5.43948114e-01
-3.84268373e-01 -3.48961294e-01 2.51468509e-01 2.64790654e-01
2.14282155e-01 -3.87854308e-01 -9.19069290e-01 8.71461868e-01
-1.82763720e+00 -5.62982917e-01 -1.03304580e-01 1.90809929e+00
1.04315913e+00 3.45156074e-01 1.85390517e-01 -1.31527647e-01
7.92140961e-01 2.21774384e-01 -8.56074274e-01 -1.53844416e-01
-5.64845875e-02 4.78679895e-01 7.21710265e-01 3.78233910e-01
-1.21050954e+00 1.45011246e+00 4.35740089e+00 1.40085578e+00
-1.22597671e+00 2.70267516e-01 9.22914803e-01 1.91561524e-02
-2.71841675e-01 -2.19704807e-02 -7.66813159e-01 7.23087370e-01
4.39266682e-01 1.13492779e-01 1.02024913e-01 1.13849509e+00
2.79092103e-01 -8.42440128e-02 -9.16752279e-01 8.12286258e-01
-2.02766225e-01 -1.07913291e+00 -1.78110063e-01 -1.96170494e-01
9.72354054e-01 -7.07497373e-02 2.90162534e-01 4.24473673e-01
3.65741968e-01 -7.40165651e-01 8.39791596e-01 7.20241740e-02
8.23598623e-01 -7.35500038e-01 4.83541995e-01 3.75612050e-01
-1.16587949e+00 1.07271045e-01 -3.27174604e-01 1.82453886e-01
2.46487826e-01 6.36732578e-01 -7.13408172e-01 6.68954909e-01
8.17427337e-01 8.42973709e-01 -3.64010066e-01 7.67534018e-01
-4.33398038e-01 6.34260714e-01 -3.03751022e-01 3.16779196e-01
4.16531324e-01 -2.18281373e-01 4.11288112e-01 1.18044448e+00
2.87103057e-01 5.13008758e-02 2.72449881e-01 9.10255492e-01
-1.33195639e-01 2.27701008e-01 -3.68344665e-01 2.58933276e-01
2.64787316e-01 9.23891664e-01 -9.02122438e-01 -2.12721616e-01
-5.31474173e-01 1.33018434e+00 2.18832538e-01 4.23596561e-01
-1.19806314e+00 -1.72503442e-01 7.93524325e-01 3.01369071e-01
6.15605712e-01 -2.19453067e-01 -7.73995697e-01 -1.05516624e+00
1.62632748e-01 -8.08625281e-01 4.43402261e-01 -5.41419029e-01
-1.21810210e+00 3.58568043e-01 -5.75876385e-02 -1.44140863e+00
3.45587842e-02 -5.22836924e-01 -4.10708249e-01 7.62696743e-01
-1.92074943e+00 -1.31194985e+00 -3.52800846e-01 8.14075589e-01
8.95725667e-01 3.07829920e-02 5.15338480e-01 4.70473170e-01
-5.74035406e-01 8.26458037e-01 2.71205604e-01 1.14410475e-01
8.59097540e-01 -1.19237125e+00 4.74337041e-01 9.78812993e-01
1.23679847e-01 1.34135336e-01 6.79842949e-01 -6.58453524e-01
-6.29236758e-01 -1.28420353e+00 3.95903438e-01 -1.80489630e-01
4.82966959e-01 -3.56925040e-01 -9.20845449e-01 4.44671839e-01
9.76218656e-02 1.08705200e-01 3.68416846e-01 -2.49595195e-01
-3.17466080e-01 -3.27436954e-01 -1.31250727e+00 7.04630494e-01
1.28930724e+00 -1.05754524e-01 -3.17822099e-01 2.51543850e-01
7.42610991e-01 -5.65091968e-01 -6.71631753e-01 5.11466444e-01
2.35496998e-01 -1.04445219e+00 1.07390141e+00 -2.11969778e-01
5.47799170e-01 -3.35160017e-01 -5.45101203e-02 -1.33918548e+00
-1.33037746e-01 -2.76257068e-01 3.13970983e-01 1.35391390e+00
5.07140756e-01 -6.70365810e-01 1.21744573e+00 4.32683349e-01
-3.59695405e-01 -6.90747023e-01 -9.76177812e-01 -1.01168954e+00
4.84074563e-01 -5.28306663e-01 2.95564502e-01 1.04333019e+00
-4.82736081e-01 1.08085662e-01 -1.35756880e-01 2.21772745e-01
6.07384920e-01 1.27345368e-01 6.72549069e-01 -9.65568244e-01
-3.09150070e-01 -5.67439556e-01 -2.07043394e-01 -1.44960856e+00
3.26246440e-01 -8.05550337e-01 1.24940105e-01 -1.48165452e+00
-1.54746503e-01 -6.69063032e-01 -4.03321803e-01 5.65077305e-01
-2.29636401e-01 3.71216506e-01 5.48394807e-02 5.06567284e-02
-5.11067986e-01 6.03195429e-01 1.56914759e+00 -2.76402324e-01
-2.95149595e-01 6.07942864e-02 -8.98362756e-01 8.45454216e-01
1.13829529e+00 -5.53303957e-01 -7.05933928e-01 -5.04408240e-01
-3.49473059e-01 -2.81721890e-01 3.92589033e-01 -9.59337473e-01
-1.94127429e-02 -3.64285767e-01 2.09105551e-01 -3.26207370e-01
4.06707436e-01 -9.61050689e-01 -1.85715675e-01 2.94018298e-01
-2.05561906e-01 -5.67326128e-01 3.47574621e-01 6.14341259e-01
-3.97921979e-01 -1.92787200e-01 1.18448067e+00 -1.91985771e-01
-1.29026878e+00 3.56174946e-01 9.10277590e-02 4.52162415e-01
1.33071184e+00 -5.03494561e-01 -5.94161637e-02 -2.47071013e-01
-6.54594600e-01 3.48539144e-01 6.95077062e-01 3.80606592e-01
4.94075179e-01 -1.12875450e+00 -6.92165971e-01 3.34588848e-02
2.45065227e-01 4.41684395e-01 5.03338873e-01 6.58278763e-01
-5.36081970e-01 2.85309970e-01 -2.17400327e-01 -6.56183124e-01
-1.07316899e+00 4.02059704e-01 4.24100548e-01 -4.13794965e-01
-5.50777376e-01 1.15282190e+00 7.22928286e-01 -4.40430284e-01
1.48629636e-01 -1.97375983e-01 1.09635890e-01 -1.86950818e-01
7.06432685e-02 1.60589486e-01 2.26636808e-02 -5.62048495e-01
-4.67839718e-01 6.69789255e-01 -1.66845009e-01 -1.56794295e-01
1.18285573e+00 -1.91273898e-01 4.24210876e-01 3.37954313e-02
1.22575903e+00 -1.19297601e-01 -1.82229722e+00 -3.47107470e-01
-1.96027040e-01 -5.91788292e-01 -7.95730669e-03 -9.06539917e-01
-1.49881506e+00 9.49064851e-01 8.25285792e-01 -1.58775494e-01
1.41779768e+00 9.62568521e-02 1.08126473e+00 -1.34757683e-01
2.62960345e-01 -1.44683862e+00 2.09215999e-01 2.44010448e-01
6.76893830e-01 -1.37496030e+00 -1.74932361e-01 -7.83009589e-01
-9.37329710e-01 6.46369040e-01 9.35061038e-01 -2.29927376e-01
4.94261414e-01 1.03945136e-01 9.49451849e-02 2.38763183e-01
-1.72282115e-01 -5.33202231e-01 2.06631929e-01 9.38917041e-01
3.24217677e-01 -1.30915847e-02 -3.67386311e-01 6.91417634e-01
6.30020872e-02 -1.62714422e-02 4.33635712e-02 7.99456179e-01
-3.41839820e-01 -1.41603839e+00 -2.37241149e-01 5.51753379e-02
-4.95237470e-01 -1.10410079e-01 -2.69478858e-01 9.74604487e-01
4.86873418e-01 8.21662605e-01 -1.31743550e-01 -1.05645411e-01
5.89080453e-01 -4.51745056e-02 3.66329551e-01 -5.45601904e-01
-2.85431325e-01 2.15062395e-01 -5.63125797e-02 -5.12570262e-01
-3.41492027e-01 -6.93429112e-01 -1.54560077e+00 -3.16673832e-04
-1.75479278e-01 -2.47604147e-01 5.43496490e-01 6.80877686e-01
3.93089682e-01 5.51812410e-01 4.99846071e-01 -6.66798055e-01
-3.83320153e-01 -6.47938609e-01 -4.74541783e-01 5.41859627e-01
9.27041750e-03 -7.20248938e-01 -5.24064191e-02 3.33858460e-01] | [9.634251594543457, 1.3009954690933228] |
a4015b33-7f63-4d66-a155-e8118f603612 | slam-based-quasi-dense-reconstruction-for | 1705.09107 | null | http://arxiv.org/abs/1705.09107v1 | http://arxiv.org/pdf/1705.09107v1.pdf | SLAM based Quasi Dense Reconstruction For Minimally Invasive Surgery Scenes | Recovering surgical scene structure in laparoscope surgery is crucial step
for surgical guidance and augmented reality applications. In this paper, a
quasi dense reconstruction algorithm of surgical scene is proposed. This is
based on a state-of-the-art SLAM system, and is exploiting the initial
exploration phase that is typically performed by the surgeon at the beginning
of the surgery. We show how to convert the sparse SLAM map to a quasi dense
scene reconstruction, using pairs of keyframe images and correlation-based
featureless patch matching. We have validated the approach with a live porcine
experiment using Computed Tomography as ground truth, yielding a Root Mean
Squared Error of 4.9mm. | ['J. M. M. Montiel', 'Alexandre Hostettler', 'Toby Collins', 'Nader Mahmoud', 'Luc Soler', 'Christophe Doignon'] | 2017-05-25 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 3.29865992e-01 3.65603119e-01 2.26151302e-01 -1.21247321e-01
-5.87862492e-01 -3.53592366e-01 2.37883672e-01 3.27080339e-01
-6.19036496e-01 6.38409615e-01 -9.71877128e-02 -3.80833596e-01
-4.20368642e-01 -4.42423850e-01 -6.98707461e-01 -5.74224830e-01
-2.96885893e-02 7.92330742e-01 1.43922195e-01 -1.61612689e-01
2.66953886e-01 7.07019150e-01 -1.00345445e+00 1.53806433e-01
5.76548696e-01 7.95387924e-01 7.29180753e-01 5.96698999e-01
1.96223274e-01 5.43892205e-01 9.37532783e-02 1.40524730e-02
6.58208609e-01 -4.40843612e-01 -7.42592990e-01 1.55101433e-01
3.51323187e-01 -3.05891693e-01 -4.85267818e-01 1.07508540e+00
4.36082512e-01 -6.17683455e-02 6.26170859e-02 -4.55308378e-01
5.03289998e-01 2.28626043e-01 -2.01080754e-01 -1.68484867e-01
6.98736489e-01 -2.83553183e-01 2.11827114e-01 -9.87478912e-01
1.29529405e+00 4.13773715e-01 8.39268863e-01 1.54135838e-01
-1.12948954e+00 -3.87649089e-01 -7.11717725e-01 -2.22167611e-01
-1.35691571e+00 -3.35949689e-01 6.47240341e-01 -5.32776177e-01
5.81062138e-01 2.27605730e-01 1.28732026e+00 5.11198461e-01
1.15979123e+00 1.01656429e-01 1.36491847e+00 -7.22363830e-01
2.57333398e-01 2.57474452e-01 -2.91912764e-01 1.05840611e+00
5.89560807e-01 7.50123918e-01 -5.31671882e-01 -1.94604799e-01
1.39716554e+00 4.46036786e-01 -5.67504287e-01 -1.42992234e+00
-1.56011558e+00 6.77617431e-01 7.91706324e-01 3.14936042e-01
-7.69214690e-01 1.49585918e-01 -2.35992726e-02 1.22153156e-01
-1.12203993e-01 6.52912915e-01 8.71957242e-02 -3.28869522e-02
-8.85994375e-01 -2.11755887e-01 9.67140079e-01 1.07610488e+00
6.93255186e-01 -4.75271374e-01 3.38648647e-01 7.86671564e-02
4.29639697e-01 1.55346945e-01 4.12208676e-01 -9.22561824e-01
-1.25246450e-01 4.96768892e-01 4.90570366e-02 -8.83137524e-01
-5.85996568e-01 -6.45883262e-01 -7.85704970e-01 1.30191922e-01
-4.78726327e-02 3.48013192e-02 -9.24170673e-01 8.48809481e-01
4.04887497e-01 5.80900013e-01 2.48177890e-02 9.60086882e-01
6.58241570e-01 -1.65616363e-01 -5.14798045e-01 -4.92341667e-01
1.04581571e+00 -7.37596631e-01 -5.97259760e-01 -2.37488627e-01
6.49035692e-01 -1.01858842e+00 3.21331978e-01 4.90232438e-01
-9.11035001e-01 -8.87378231e-02 -8.97759259e-01 1.43150926e-01
2.56679654e-01 -9.59339458e-03 7.00269043e-01 2.76630878e-01
-1.02115417e+00 6.40010655e-01 -1.24280465e+00 -7.54446030e-01
7.30517209e-02 4.72159624e-01 -1.01825988e+00 -3.76026124e-01
-5.63547492e-01 1.06844509e+00 1.65905595e-01 3.67950022e-01
-8.68781030e-01 -7.54001200e-01 -1.11050701e+00 -4.19603914e-01
1.91746444e-01 -1.04423499e+00 1.12533271e+00 -1.61127657e-01
-1.70996463e+00 1.26150918e+00 -1.01904206e-01 -3.45505804e-01
4.85917687e-01 -1.93944469e-01 6.73674094e-03 3.12005728e-01
-5.69947436e-02 -6.97392225e-03 4.39217448e-01 -1.45203781e+00
-1.99935064e-01 -5.24113536e-01 -7.19106197e-02 2.62093335e-01
6.52668297e-01 -3.66372228e-01 -3.93609405e-01 -1.26818031e-01
1.14303839e+00 -1.32292604e+00 -8.94908249e-01 3.10735583e-01
-7.32208639e-02 1.09670615e+00 2.71912873e-01 -4.74540651e-01
6.15724802e-01 -1.92515385e+00 2.83305377e-01 5.37873685e-01
2.45746076e-01 -3.15418720e-01 2.20278561e-01 5.93450665e-01
1.56759262e-01 -7.73015440e-01 -1.25693679e-01 -4.18388128e-01
-6.67855680e-01 3.95369232e-01 4.80880626e-02 1.24488783e+00
-7.58046985e-01 7.04961658e-01 -1.14398885e+00 -6.16494656e-01
6.31458759e-01 5.21390915e-01 -5.55037200e-01 4.29290444e-01
5.10399759e-01 1.16800785e+00 -3.96591365e-01 6.48033798e-01
6.72555506e-01 -3.95160206e-02 3.53116840e-01 -4.70484436e-01
-5.26905656e-01 9.68108475e-02 -1.01624501e+00 3.04428244e+00
-8.37284267e-01 3.17833900e-01 6.56161189e-01 -5.35366476e-01
8.17822337e-01 4.47506368e-01 9.09122765e-01 -4.74567711e-01
4.95429188e-01 6.93396628e-01 -1.37701005e-01 -1.69709370e-01
6.13028944e-01 -5.26394188e-01 2.94294320e-02 8.80350992e-02
1.18468352e-01 -9.14675236e-01 -4.52926785e-01 1.89513534e-01
1.23681092e+00 2.41326436e-01 7.48337448e-01 -6.72980011e-01
2.19631478e-01 5.90057611e-01 2.24640056e-01 5.63902557e-01
5.09991162e-02 7.52929151e-01 2.27189474e-02 -6.28189147e-01
-9.42608416e-01 -1.06331217e+00 -5.46746731e-01 -3.24729681e-01
7.13936448e-01 -2.89571762e-01 -4.04913902e-01 -2.89328039e-01
-4.68396209e-02 2.78843403e-01 -7.43732810e-01 -3.37634957e-03
-5.84038436e-01 -1.28295243e-01 -4.14909691e-01 -4.92444308e-03
-8.80291238e-02 -6.17535949e-01 -1.13583028e+00 3.49407911e-01
1.37499616e-01 -1.14090550e+00 1.08568355e-01 5.11435628e-01
-1.25491834e+00 -1.31843185e+00 -4.24692839e-01 -7.19437659e-01
1.17777848e+00 4.78649378e-01 1.14957285e+00 3.00121810e-02
-7.20900118e-01 4.95726913e-01 -3.61682564e-01 -1.35603517e-01
-2.96740144e-01 -3.37643981e-01 7.84131326e-03 -3.40702862e-01
-3.64512950e-01 -5.70396602e-01 -7.66537428e-01 3.47242832e-01
-5.11668444e-01 4.22419280e-01 1.00189757e+00 9.43943620e-01
1.05717123e+00 -3.03725272e-01 -6.03267610e-01 -1.15120161e+00
-1.51005574e-02 -2.99930334e-01 -8.20031047e-01 -2.50892580e-01
-3.61639500e-01 -3.38612974e-01 2.38721848e-01 1.48582935e-01
-5.96425414e-01 8.43843520e-01 -1.73894793e-01 -6.86912715e-01
-1.07697941e-01 6.25472665e-01 2.99839616e-01 -8.15158665e-01
7.00494647e-01 1.49237394e-01 4.15444374e-01 -4.50281620e-01
4.85308282e-02 2.01096922e-01 7.35713124e-01 -8.47837105e-02
6.96540117e-01 8.90236497e-01 5.19070446e-01 -7.85577714e-01
-6.33528471e-01 -1.00445235e+00 -1.02418840e+00 -1.70703530e-01
5.03159642e-01 -8.31375182e-01 -6.25897944e-01 -1.26062436e-02
-1.04797673e+00 -3.37657630e-01 -4.84358877e-01 1.32390153e+00
-9.52199996e-01 1.06146865e-01 -4.18785781e-01 -1.58541411e-01
-2.36983448e-01 -1.23157191e+00 1.22784710e+00 -1.08402997e-01
-8.17216560e-02 -1.02834105e+00 6.04573607e-01 -8.36778879e-02
3.48472267e-01 6.83626354e-01 1.57625422e-01 1.09305132e-04
-9.91147518e-01 -5.55609822e-01 1.47055969e-01 -3.44509572e-01
1.57127008e-01 -5.97661316e-01 -6.66194797e-01 -5.78444302e-01
4.53320503e-01 1.73929676e-01 4.01392817e-01 6.02987587e-01
5.55334032e-01 3.26254308e-01 -6.70606196e-01 1.27817225e+00
1.86345994e+00 -1.33522347e-01 6.85448825e-01 4.58757073e-01
6.26103461e-01 3.16980898e-01 1.12193000e+00 3.90538692e-01
7.21521601e-02 8.07578981e-01 7.99669385e-01 -1.28374010e-01
-6.57283366e-02 -2.07014576e-01 -3.39254260e-01 7.95714319e-01
7.56910071e-02 4.53635991e-01 -1.06573939e+00 4.36730087e-01
-1.77759504e+00 -3.91908228e-01 -2.41281509e-01 2.54268408e+00
4.50028092e-01 -2.38841191e-01 -7.90196717e-01 -1.19627297e-01
5.18395863e-02 -2.45334506e-01 9.55964345e-03 -2.02463984e-01
5.54412127e-01 3.48354667e-01 8.94031107e-01 9.45202291e-01
-7.51081824e-01 7.26783991e-01 6.17226076e+00 7.32147321e-03
-1.38212848e+00 1.46195039e-01 -5.65006733e-01 1.28791049e-01
-1.46980569e-01 3.45330596e-01 -3.19018751e-01 -8.70899558e-02
5.86642385e-01 -1.11477457e-01 1.89092353e-01 8.19399953e-01
2.27765571e-02 -8.08118701e-01 -1.10690439e+00 1.15170884e+00
2.17170745e-01 -1.54953635e+00 -3.29200000e-01 2.96086103e-01
6.37549579e-01 1.14570126e-01 -3.09681892e-01 -2.20817804e-01
-5.65859443e-03 -8.94441187e-01 2.92768627e-01 7.09865272e-01
8.89773726e-01 -3.19252253e-01 9.70683753e-01 5.13087630e-01
-9.50374424e-01 2.86772639e-01 -3.82468998e-01 7.90406540e-02
4.15355533e-01 7.62274683e-01 -1.30310309e+00 9.34784830e-01
3.75445426e-01 6.86228275e-01 -9.11369547e-02 1.53111899e+00
-1.24298953e-01 -2.22844169e-01 -3.80819261e-01 4.70790684e-01
8.59976262e-02 -2.49817982e-01 6.98839545e-01 7.33067751e-01
7.10915029e-01 3.92044425e-01 2.34750539e-01 3.69561344e-01
3.95517617e-01 8.81814957e-02 -8.75018895e-01 6.14056289e-01
-1.09148808e-01 1.41483867e+00 -7.50683606e-01 8.90357271e-02
-5.88161610e-02 1.11466217e+00 -6.11395761e-02 -1.89541191e-01
-2.66986728e-01 -9.83827338e-02 2.90519387e-01 4.36546385e-01
-3.63925397e-02 -4.72816050e-01 -1.80580452e-01 -1.13735509e+00
-1.53036863e-01 -3.41123968e-01 -6.40025958e-02 -7.84531772e-01
-4.12279665e-01 9.40513730e-01 -1.32057726e-01 -1.63294160e+00
-3.84572655e-01 -4.62901741e-01 -2.92431623e-01 9.41795230e-01
-1.51720572e+00 -1.08626473e+00 -9.99020040e-01 6.46604836e-01
7.13084564e-02 1.68737113e-01 1.22279429e+00 -1.03886304e-02
2.36265644e-01 -2.56682873e-01 1.60266414e-01 -2.27449730e-01
7.12587118e-01 -1.10518682e+00 -2.75925416e-02 6.73806846e-01
-3.10581587e-02 8.83518577e-01 9.92033184e-01 -7.46950626e-01
-1.99946117e+00 -7.20273674e-01 7.78154194e-01 -4.10247296e-01
4.28559989e-01 -2.24905819e-01 -5.00093579e-01 9.07724380e-01
-1.01630852e-01 6.05280757e-01 6.18969738e-01 -8.73446465e-02
4.51504141e-01 -6.10817736e-03 -1.37058997e+00 1.81532875e-01
9.62540686e-01 -3.65625471e-01 -6.01228833e-01 4.49118078e-01
2.38229468e-01 -1.48465693e+00 -1.10374486e+00 6.66357517e-01
8.94107759e-01 -1.10367167e+00 8.22084486e-01 6.31957278e-02
-8.47488455e-03 -2.95320123e-01 -1.74048264e-02 -1.50160837e+00
-1.46947712e-01 -7.21227944e-01 2.57712692e-01 3.99291776e-02
-5.00987433e-02 -6.72268867e-01 1.11013186e+00 2.87256092e-01
-5.81008434e-01 -7.11633623e-01 -1.14748096e+00 -4.92791474e-01
-5.27406394e-01 -2.27515519e-01 -1.43866941e-01 6.96820319e-01
2.34139919e-01 -2.42561743e-01 -2.13872828e-02 5.99426031e-01
9.38645899e-01 5.49329877e-01 8.40283215e-01 -1.30927229e+00
-2.53615588e-01 3.04322839e-01 -8.74178529e-01 -8.19296062e-01
-6.28306195e-02 -7.39629507e-01 4.62720603e-01 -1.70900786e+00
8.49958360e-02 -9.21041608e-01 -1.63579062e-01 -8.81318599e-02
4.53268707e-01 4.05225784e-01 -1.54520944e-01 4.42746162e-01
-3.10697407e-01 8.08289796e-02 1.44011569e+00 6.35029197e-01
-3.42026621e-01 1.74622580e-01 -1.02439024e-01 7.79249072e-01
3.98482114e-01 -6.95047259e-01 -3.09920967e-01 -3.06818843e-01
2.53364563e-01 5.82410932e-01 5.15455425e-01 -1.10344934e+00
7.15504646e-01 1.69143360e-02 3.66047360e-02 -6.69634402e-01
6.95128798e-01 -1.58730912e+00 8.99131298e-01 1.02787685e+00
1.84496343e-01 -2.24805042e-01 2.28773296e-01 5.40759623e-01
-3.19878131e-01 -2.45081887e-01 8.04335296e-01 -4.61123735e-01
-6.38783813e-01 3.27849537e-01 1.45653173e-01 -5.52142143e-01
1.20005453e+00 -3.03715616e-01 1.46131456e-01 9.44977105e-02
-1.07922900e+00 -3.75273794e-01 1.16350281e+00 -3.00156295e-01
1.07914257e+00 -1.06410134e+00 -6.68996453e-01 6.93504870e-01
2.34816730e-01 2.87207007e-01 3.73764366e-01 1.58135092e+00
-1.39797199e+00 6.11295700e-01 -4.98903871e-01 -1.01673937e+00
-1.22858787e+00 5.94788909e-01 6.44279242e-01 -2.80374736e-01
-9.36863363e-01 5.15027761e-01 2.61455029e-01 -4.80858296e-01
-1.05989061e-01 -3.97312820e-01 1.71764493e-01 -6.92763686e-01
2.61135221e-01 -8.32060501e-02 4.62705940e-01 -6.31245315e-01
-5.54543495e-01 1.00334990e+00 1.66381583e-01 -1.54043004e-01
1.37768352e+00 -1.27117321e-01 -2.66731441e-01 3.79034668e-01
1.00310695e+00 3.91934574e-01 -8.90234590e-01 -2.24537462e-01
-3.93036008e-01 -1.08043206e+00 5.52702248e-01 -4.58945721e-01
-7.61295855e-01 6.02660835e-01 5.83536685e-01 -4.40341592e-01
8.73259485e-01 9.28900857e-03 5.24008691e-01 2.76324153e-01
1.34379828e+00 -2.55210906e-01 -4.38262016e-01 3.37195069e-01
1.06706107e+00 -1.14862537e+00 5.83136380e-01 -9.18977559e-01
-2.60106087e-01 1.02467978e+00 -5.80874681e-02 -3.56741279e-01
9.92883682e-01 5.20741940e-01 1.56913280e-01 -5.20569861e-01
-3.32357228e-01 -3.68154384e-02 9.54549462e-02 3.67647588e-01
3.54593247e-01 -8.96058325e-03 -4.06936198e-01 -2.24233508e-01
-1.93748146e-01 3.48619342e-01 5.48771322e-01 1.52348781e+00
-2.72458047e-01 -1.01808715e+00 -2.51216948e-01 1.19214639e-01
-3.11711997e-01 -1.20693602e-01 2.20922887e-01 8.97645772e-01
-1.80454999e-01 3.95000398e-01 -4.24131677e-02 -4.98833507e-02
5.97141802e-01 -5.32380402e-01 1.06750441e+00 -9.76489723e-01
-6.83286667e-01 6.08598776e-02 -1.06737196e-01 -1.16915393e+00
-3.47724974e-01 -6.14798486e-01 -1.26861525e+00 6.60117492e-02
-2.71000773e-01 3.30835015e-01 1.36102378e+00 4.14608181e-01
2.56906837e-01 4.50751990e-01 5.64745307e-01 -1.20410836e+00
-3.55287082e-02 -5.99187493e-01 -9.07647789e-01 1.40902698e-01
5.94287932e-01 -6.68180466e-01 -2.56637871e-01 -1.73019484e-01] | [13.824315071105957, -3.1116580963134766] |
83619106-1619-4616-92bc-8e49c53e8d1e | difface-blind-face-restoration-with-diffused | 2212.06512 | null | https://arxiv.org/abs/2212.06512v2 | https://arxiv.org/pdf/2212.06512v2.pdf | DifFace: Blind Face Restoration with Diffused Error Contraction | While deep learning-based methods for blind face restoration have achieved unprecedented success, they still suffer from two major limitations. First, most of them deteriorate when facing complex degradations out of their training data. Second, these methods require multiple constraints, e.g., fidelity, perceptual, and adversarial losses, which require laborious hyper-parameter tuning to stabilize and balance their influences. In this work, we propose a novel method named DifFace that is capable of coping with unseen and complex degradations more gracefully without complicated loss designs. The key of our method is to establish a posterior distribution from the observed low-quality (LQ) image to its high-quality (HQ) counterpart. In particular, we design a transition distribution from the LQ image to the intermediate state of a pre-trained diffusion model and then gradually transmit from this intermediate state to the HQ target by recursively applying a pre-trained diffusion model. The transition distribution only relies on a restoration backbone that is trained with $L_2$ loss on some synthetic data, which favorably avoids the cumbersome training process in existing methods. Moreover, the transition distribution can contract the error of the restoration backbone and thus makes our method more robust to unknown degradations. Comprehensive experiments show that DifFace is superior to current state-of-the-art methods, especially in cases with severe degradations. Code and model are available at https://github.com/zsyOAOA/DifFace. | ['Chen Change Loy', 'Zongsheng Yue'] | 2022-12-13 | null | null | null | null | ['blind-face-restoration'] | ['computer-vision'] | [ 7.99509734e-02 -2.88401008e-01 7.97109306e-02 -1.32937461e-01
-6.92099035e-01 -1.91322625e-01 2.56164998e-01 -3.44328880e-01
-1.46286517e-01 6.33756220e-01 1.61761224e-01 -1.27098992e-01
-9.86010879e-02 -7.29925692e-01 -6.55661583e-01 -9.72242117e-01
2.40386859e-01 -3.57062370e-02 1.99831709e-01 -1.91742241e-01
-1.21273726e-01 4.20664012e-01 -1.37221563e+00 2.78718770e-02
1.26794028e+00 1.11658108e+00 2.92941272e-01 3.52733433e-01
1.38690129e-01 5.45107186e-01 -6.62193477e-01 -5.71424425e-01
4.17246729e-01 -5.93578398e-01 -3.55492711e-01 2.47578353e-01
4.56670791e-01 -6.26254439e-01 -6.39844298e-01 1.37580454e+00
8.29306006e-01 1.24708377e-01 5.71771204e-01 -1.18849754e+00
-8.92639577e-01 8.63421825e-04 -6.24807477e-01 1.43654365e-02
1.19362764e-01 5.64825475e-01 6.44055009e-01 -9.64980602e-01
4.02262092e-01 1.35638630e+00 7.58746445e-01 7.13496625e-01
-1.43754709e+00 -8.14355433e-01 1.46194771e-01 3.06830257e-01
-1.33558810e+00 -8.70249152e-01 8.25344980e-01 -3.51492912e-01
4.28618878e-01 -4.02774885e-02 4.64620739e-01 1.03064096e+00
1.44599639e-02 7.05383003e-01 1.15273881e+00 -1.83148250e-01
2.46204630e-01 -8.37170929e-02 -4.07071292e-01 6.22930825e-01
1.24240573e-02 2.59869039e-01 -3.51773888e-01 -5.63680893e-03
8.11179221e-01 -1.20296203e-01 -8.57157409e-01 -2.47964472e-01
-6.75959706e-01 5.38332403e-01 7.23052621e-01 1.32980764e-01
-4.41634923e-01 -4.63743024e-02 1.20729096e-01 4.84172493e-01
5.30458689e-01 -5.47302328e-02 -2.71613717e-01 2.00839028e-01
-1.06727254e+00 1.75344071e-03 3.55653644e-01 6.98606372e-01
7.04897046e-01 1.57172278e-01 -2.85257578e-01 1.23731053e+00
4.77866560e-01 5.82429230e-01 3.36426795e-01 -1.06849742e+00
4.83261615e-01 1.14162013e-01 1.70145139e-01 -1.01206601e+00
-2.34850924e-02 -7.22022653e-01 -1.14466691e+00 5.42211592e-01
3.94327670e-01 -3.71422991e-02 -1.00426948e+00 1.92531562e+00
3.53627890e-01 1.17438562e-01 -2.33033597e-02 1.01125920e+00
4.44592714e-01 8.75200808e-01 -7.22098723e-02 -5.53634584e-01
8.97147655e-01 -1.03298795e+00 -7.38259435e-01 -1.73531458e-01
-1.62739053e-01 -8.52290988e-01 1.43196356e+00 5.33327103e-01
-1.22313559e+00 -6.69509292e-01 -1.08310497e+00 1.08067552e-02
3.19511116e-01 1.45288989e-01 3.26915011e-02 5.13517499e-01
-1.28435981e+00 6.62380815e-01 -8.15768838e-01 -1.34278253e-01
6.73764288e-01 1.72739193e-01 -1.27105251e-01 -4.98434305e-01
-1.03241873e+00 5.60521364e-01 -4.36616614e-02 2.93832481e-01
-1.22382009e+00 -6.23836398e-01 -6.07948780e-01 7.56839244e-03
2.78156817e-01 -7.62663484e-01 1.19839096e+00 -1.09216392e+00
-1.82582724e+00 4.35077846e-01 -1.67601094e-01 2.76621450e-02
8.24394703e-01 -2.74335384e-01 -4.70306247e-01 1.86153784e-01
-4.58279252e-02 5.10185301e-01 1.41613364e+00 -1.57814538e+00
-2.84794420e-01 -2.34127805e-01 -8.36569369e-02 3.29440869e-02
-6.24433994e-01 -6.49576113e-02 -9.33996320e-01 -9.38248992e-01
2.07066014e-01 -6.92224562e-01 3.75700323e-03 6.28353775e-01
-4.64219451e-01 1.11049220e-01 7.49281347e-01 -8.92483115e-01
1.18670154e+00 -2.40395284e+00 1.23031363e-01 -1.33651167e-01
1.66517764e-01 6.15902603e-01 -4.51438904e-01 3.58079910e-01
1.31142229e-01 -5.59805669e-02 -6.29299164e-01 -6.36628091e-01
-7.06327185e-02 9.12265405e-02 -3.00740153e-01 6.54505968e-01
1.82533517e-01 4.72463042e-01 -8.28666925e-01 -3.72808397e-01
6.09740466e-02 8.92336488e-01 -6.18930638e-01 4.48925108e-01
-2.30883688e-01 7.59409487e-01 -2.11026475e-01 6.74634337e-01
1.05767560e+00 -1.13314569e-01 -6.71210513e-02 -3.99269670e-01
1.20057113e-01 9.08208564e-02 -1.07997966e+00 1.49270523e+00
-5.49691200e-01 4.64504331e-01 4.85813051e-01 -8.34302664e-01
8.98640633e-01 2.79484540e-01 1.40561566e-01 -8.58142674e-01
-1.01707928e-01 2.66032845e-01 -1.87123120e-01 -5.68998754e-01
4.49601701e-03 -4.55127954e-01 6.05778694e-01 1.76192120e-01
-2.43406519e-02 -5.31266257e-02 -2.08841227e-02 -5.02291247e-02
1.00664115e+00 8.32886174e-02 -1.82316840e-01 4.95959297e-02
5.85597157e-01 -6.18692994e-01 1.08425736e+00 3.47820312e-01
-4.64526385e-01 8.71461093e-01 4.26994890e-01 -1.22450395e-02
-1.10418260e+00 -1.36589503e+00 -2.29897931e-01 6.00733161e-01
3.10925096e-01 -6.21574931e-02 -7.80222535e-01 -5.89111626e-01
-5.19972332e-02 4.59737003e-01 -1.64343297e-01 -4.34124380e-01
-4.22388494e-01 -8.09564829e-01 3.79329771e-01 1.94634601e-01
8.71780097e-01 -9.70996499e-01 9.67464894e-02 1.85805008e-01
-2.70212084e-01 -9.58051085e-01 -7.51057088e-01 -5.22850573e-01
-9.31246340e-01 -8.65512550e-01 -1.00497293e+00 -8.70840847e-01
9.54362273e-01 2.72654235e-01 8.84569108e-01 3.30873966e-01
-1.01821668e-01 8.39014500e-02 -2.19190791e-01 8.11073333e-02
-5.38621426e-01 -4.00850654e-01 1.45942360e-01 3.73204678e-01
-2.75288224e-01 -8.28152716e-01 -9.99732018e-01 5.02889574e-01
-1.13142753e+00 -1.21637128e-01 6.54617786e-01 1.12657368e+00
6.10195041e-01 5.04802585e-01 7.42078006e-01 -3.34472179e-01
6.32918835e-01 -3.82298559e-01 -6.93458378e-01 5.75709939e-02
-8.02336931e-01 -1.25623584e-01 8.03808331e-01 -5.45233786e-01
-1.24591327e+00 -1.69341922e-01 -3.19121510e-01 -6.21426284e-01
1.34966090e-01 2.64881611e-01 -6.75839067e-01 -8.57008770e-02
5.41433692e-01 2.84005225e-01 2.44397849e-01 -7.17514932e-01
2.59221077e-01 7.20924258e-01 6.75310254e-01 -5.19674301e-01
1.10584056e+00 4.27446872e-01 -2.24191710e-01 -6.25459909e-01
-7.60801435e-01 -1.43503705e-02 -9.79289338e-02 -3.78267616e-01
3.77235860e-01 -9.51670945e-01 -6.36564255e-01 1.16474020e+00
-9.71853077e-01 -6.14280462e-01 -1.55996338e-01 3.60505819e-01
-4.32287365e-01 6.10531330e-01 -9.89602387e-01 -6.84729457e-01
-3.16772401e-01 -1.17100191e+00 7.07520366e-01 4.27279294e-01
4.46711719e-01 -7.08393693e-01 -7.45032132e-02 4.99901772e-01
6.35264397e-01 1.04939766e-01 8.41073871e-01 3.04408669e-01
-6.06951058e-01 -4.02198508e-02 -3.26955318e-01 1.04923046e+00
3.49634439e-01 4.03879471e-02 -9.33757663e-01 -9.07961607e-01
2.46097237e-01 -4.29147929e-01 9.89010215e-01 3.17538261e-01
1.18348718e+00 -5.63476324e-01 -6.01956584e-02 8.78486454e-01
1.34769011e+00 1.16209410e-01 9.46729839e-01 3.11779648e-01
3.95116627e-01 5.37985742e-01 4.65337902e-01 2.70716041e-01
1.90417439e-01 7.76007175e-01 5.99097908e-01 -2.53386348e-01
-5.66207588e-01 -3.76637250e-01 7.00976968e-01 7.99340844e-01
1.91780895e-01 -4.06565011e-01 -5.84560394e-01 4.66742724e-01
-1.51206243e+00 -8.28742027e-01 3.18595409e-01 2.34920335e+00
1.21803415e+00 1.22868024e-01 -1.14333637e-01 2.44094804e-01
7.54873872e-01 2.99663097e-01 -8.86799216e-01 1.50848284e-01
-1.62375852e-01 3.55072841e-02 1.06497630e-01 6.00165486e-01
-9.17980850e-01 8.22799623e-01 5.47503376e+00 9.62628305e-01
-1.16339743e+00 2.32193246e-01 7.73345768e-01 -1.54773831e-01
-3.11336935e-01 -9.69762728e-02 -4.45467472e-01 7.65624642e-01
5.38172722e-01 5.92923202e-02 6.50920331e-01 4.44042295e-01
4.17798787e-01 4.70067821e-02 -6.95883155e-01 8.99127066e-01
-5.28075397e-02 -8.92692327e-01 -2.23114751e-02 -2.65880506e-02
7.89758742e-01 -1.26360804e-01 3.69406849e-01 4.71815765e-02
1.23453900e-01 -8.97080302e-01 6.92380488e-01 6.78520739e-01
9.41249311e-01 -6.19595647e-01 4.62345243e-01 2.76036888e-01
-7.94210672e-01 -1.84439987e-01 -5.01654088e-01 3.13227534e-01
2.72876531e-01 1.10030615e+00 -2.88743436e-01 3.77158761e-01
7.45086730e-01 6.19956493e-01 -2.43507341e-01 1.28454161e+00
-6.94327235e-01 7.37559736e-01 -1.15144819e-01 6.24155700e-01
-1.73626930e-01 -3.08412552e-01 7.67190397e-01 7.87089646e-01
5.03234267e-01 -1.52182765e-02 9.97091979e-02 9.17093873e-01
-3.73895615e-01 -1.19661726e-01 -1.72071844e-01 2.06026748e-01
6.01384103e-01 1.01763833e+00 -2.22697482e-01 -1.28176257e-01
-2.83906609e-01 1.19277430e+00 2.03975841e-01 7.02533424e-01
-7.24628270e-01 -3.57587039e-01 9.53439951e-01 1.96346596e-01
3.15236270e-01 -2.16443747e-01 4.25262563e-02 -1.18136501e+00
3.88607651e-01 -1.02340472e+00 7.10701197e-02 -7.70925581e-01
-1.55591416e+00 8.78325582e-01 -4.62434024e-01 -1.43488419e+00
1.42422512e-01 -2.00361133e-01 -6.65427029e-01 1.03483284e+00
-1.76237965e+00 -9.29075301e-01 -4.18798119e-01 7.21541226e-01
4.12029833e-01 -1.19892843e-01 4.20181185e-01 6.59131765e-01
-9.01889622e-01 9.04332459e-01 4.07652467e-01 -7.45254308e-02
9.67039824e-01 -8.23364079e-01 2.35726967e-01 1.13717949e+00
-1.95789889e-01 2.91851819e-01 7.57022202e-01 -5.65195918e-01
-1.08318603e+00 -1.20885134e+00 4.96024132e-01 9.93453115e-02
4.49455321e-01 -1.40721530e-01 -1.22713184e+00 1.97191522e-01
2.67377943e-02 2.62328029e-01 2.60517538e-01 -3.56544733e-01
-5.41998982e-01 -5.88724971e-01 -1.34537637e+00 6.40856624e-01
1.06127465e+00 -6.35612309e-01 -2.81697195e-02 2.69966692e-01
7.35385954e-01 -3.43187302e-01 -7.36950099e-01 5.58661282e-01
3.29949886e-01 -1.15229106e+00 9.49312150e-01 -1.28841296e-01
4.39406902e-01 -6.48350775e-01 -7.88368732e-02 -1.52311456e+00
-2.32326731e-01 -8.08601916e-01 -2.47063845e-01 1.44714880e+00
2.69945264e-01 -7.72695661e-01 5.17662406e-01 2.70270556e-01
-2.55465358e-01 -7.71299541e-01 -1.04438388e+00 -1.09950590e+00
6.71604201e-02 -1.59041703e-01 5.62394798e-01 7.70390928e-01
-6.09447300e-01 1.09960232e-03 -5.78212321e-01 4.42766517e-01
9.30713177e-01 -1.45251617e-01 5.90953469e-01 -8.28437865e-01
-4.23546046e-01 -3.53707343e-01 -2.32906967e-01 -1.23555958e+00
-4.16003838e-02 -5.45764685e-01 3.19637269e-01 -1.57688594e+00
6.58358037e-02 -6.58643782e-01 -2.53715634e-01 4.44473177e-01
-3.90939713e-01 3.37369412e-01 2.77535170e-01 6.17059410e-01
-1.74543694e-01 1.16470432e+00 1.62144864e+00 -2.78834730e-01
-2.60891944e-01 9.82005298e-02 -6.95521832e-01 4.85632747e-01
8.22740912e-01 -5.59836507e-01 -4.89381403e-01 -7.32102811e-01
-5.59826791e-02 1.16071008e-01 4.14136767e-01 -1.14076221e+00
1.85101688e-01 -3.88964452e-02 2.38332585e-01 3.62141728e-02
4.18880224e-01 -5.58093011e-01 1.31414652e-01 4.73283261e-01
-5.78748956e-02 -3.22347581e-01 2.08873644e-01 6.91705227e-01
-3.78254801e-01 -7.23444670e-02 1.29908931e+00 2.62965739e-01
-1.57551274e-01 5.72522342e-01 -2.10220758e-02 2.49961969e-02
6.86601937e-01 2.70255581e-02 -4.63434696e-01 -6.51936710e-01
-7.11248636e-01 1.94882646e-01 8.33276391e-01 3.74784112e-01
6.93351090e-01 -1.39281225e+00 -8.42593133e-01 3.62372100e-01
-1.94449216e-01 9.97056663e-02 5.54585934e-01 7.68366218e-01
-3.65433395e-01 -3.09382826e-01 -9.52766240e-02 -4.62304235e-01
-9.49892282e-01 6.23888850e-01 5.21819711e-01 1.47530921e-02
-8.30567658e-01 8.53587389e-01 2.26929009e-01 -2.09141821e-01
5.03480196e-01 1.43834889e-01 9.48608816e-02 -2.07003608e-01
6.75059497e-01 3.18828434e-01 8.13794881e-03 -6.12178802e-01
-1.72007129e-01 5.39796233e-01 -2.65601203e-02 -9.41484943e-02
1.32254899e+00 -4.11882579e-01 -2.34639809e-01 -7.59403184e-02
1.20725346e+00 9.31320563e-02 -1.89477682e+00 -4.31293905e-01
-4.34222490e-01 -7.96421051e-01 3.14415336e-01 -8.39673877e-01
-1.71632278e+00 7.87192345e-01 9.41874504e-01 -1.19260103e-02
1.75420511e+00 -1.45564526e-01 1.09782159e+00 -4.77243438e-02
1.73789755e-01 -8.62151265e-01 3.81771922e-01 1.11760348e-01
1.20807064e+00 -1.05100846e+00 -1.28382757e-01 -3.18552762e-01
-2.91946322e-01 8.41025531e-01 6.10488951e-01 1.70558155e-01
6.70216799e-01 3.68467383e-02 3.34266305e-01 1.41340405e-01
-4.98602360e-01 4.60399687e-02 7.15557709e-02 7.66609251e-01
5.43400459e-02 -1.69008255e-01 -1.61492988e-01 2.78925776e-01
7.46617764e-02 7.91335106e-02 3.68414462e-01 6.48076773e-01
-3.17851424e-01 -1.19568348e+00 -5.09720862e-01 1.65546402e-01
-3.02305222e-01 -1.32695690e-01 1.39639288e-01 3.07578534e-01
2.41582498e-01 1.19346464e+00 -2.28015319e-01 -3.56976479e-01
4.71659273e-01 -4.01210219e-01 2.17552215e-01 -3.02746743e-01
-1.14025928e-01 1.56334653e-01 -2.68705159e-01 -6.28899455e-01
-2.81014204e-01 -6.23518050e-01 -9.10225451e-01 -4.37596858e-01
-2.45201409e-01 -6.34464920e-02 4.74329054e-01 6.43486202e-01
4.71244574e-01 4.63678867e-01 1.06045413e+00 -8.39364648e-01
-8.05475473e-01 -9.10635054e-01 -4.71650839e-01 3.74333322e-01
7.51392186e-01 -6.64886296e-01 -6.50047421e-01 5.91721572e-02] | [11.588679313659668, -1.9850542545318604] |
cbf3bf45-619d-444f-9c30-c249f0045ef4 | orthogonal-nonnegative-tucker-decomposition | 1910.09979 | null | https://arxiv.org/abs/1910.09979v2 | https://arxiv.org/pdf/1910.09979v2.pdf | Orthogonal Nonnegative Tucker Decomposition | In this paper, we study the nonnegative tensor data and propose an orthogonal nonnegative Tucker decomposition (ONTD). We discuss some properties of ONTD and develop a convex relaxation algorithm of the augmented Lagrangian function to solve the optimization problem. The convergence of the algorithm is given. We employ ONTD on the image data sets from the real world applications including face recognition, image representation, hyperspectral unmixing. Numerical results are shown to illustrate the effectiveness of the proposed algorithm. | ['Ye Liu', 'Xiongjun Zhang', 'Hong Yan', 'Michael K. Ng', 'Junjun Pan'] | 2019-10-21 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 7.23682493e-02 -4.16446924e-01 -5.44785023e-01 -8.28646868e-02
-3.36536407e-01 -2.60556579e-01 -4.07992341e-02 -6.73027039e-01
-2.23639414e-01 6.04487658e-01 1.62798971e-01 -4.33274299e-01
-4.07738358e-01 -2.03550234e-01 -7.86703378e-02 -9.59424317e-01
-3.01871896e-01 8.37882161e-02 -7.12754786e-01 -1.31008849e-02
1.52111351e-01 4.94132280e-01 -1.09553146e+00 1.59480527e-01
9.08598661e-01 1.17862129e+00 5.05745877e-03 5.15342474e-01
1.18652686e-01 5.50105751e-01 -1.34641469e-01 -1.14587493e-01
7.10446477e-01 -2.75784964e-03 -8.46953094e-01 1.11708546e+00
2.11926922e-01 -3.14027071e-01 -1.74241126e-01 1.43825674e+00
3.15659165e-01 2.85485625e-01 5.68736613e-01 -1.64972901e+00
-8.94197166e-01 1.65937066e-01 -1.46754658e+00 2.39137903e-01
3.35146463e-03 -5.32496750e-01 1.03655350e+00 -1.46350968e+00
4.40823227e-01 1.30062568e+00 4.43276256e-01 3.37396532e-01
-1.07766938e+00 -3.67678434e-01 3.79824974e-02 1.79117844e-01
-1.21280277e+00 -3.22181195e-01 8.88781786e-01 -5.58422565e-01
5.62102973e-01 5.74215651e-01 5.19778132e-01 1.81622565e-01
-4.48304228e-02 9.45885479e-01 1.44075823e+00 -4.57480371e-01
-4.36163694e-02 -3.05009544e-01 4.79192734e-01 9.80295897e-01
6.72440171e-01 1.56742427e-02 -3.53755414e-01 -6.20168269e-01
6.48897171e-01 7.00724646e-02 -3.34703207e-01 -2.72359967e-01
-1.30895388e+00 8.83233368e-01 -3.67686190e-02 -1.61726009e-02
-6.45633996e-01 -1.19488373e-01 3.54744345e-01 4.44182962e-01
8.52013111e-01 -1.63135901e-01 -1.39225379e-01 2.08764091e-01
-5.28990507e-01 -1.22696057e-01 6.60422981e-01 9.91319716e-01
5.09955108e-01 4.23847049e-01 1.12601317e-01 1.23657238e+00
5.96310914e-01 8.93093169e-01 2.98018545e-01 -1.13930011e+00
7.14822650e-01 3.77228826e-01 2.58312136e-01 -1.24334168e+00
-3.40254337e-01 -1.48334906e-01 -1.17787588e+00 -7.55315423e-02
-1.14094548e-01 -2.85943568e-01 -7.94169605e-01 1.16862428e+00
5.79168499e-01 4.36194122e-01 4.34237659e-01 1.25591469e+00
6.29772365e-01 8.20350409e-01 -4.64426339e-01 -8.19979131e-01
1.24627268e+00 -8.63551259e-01 -1.10171497e+00 1.11135736e-01
4.09807086e-01 -1.05709374e+00 6.14833832e-01 5.97310603e-01
-1.00315607e+00 -1.57991514e-01 -9.30015802e-01 -1.32428799e-02
2.67444402e-01 6.39550447e-01 1.09280515e+00 6.70822859e-01
-7.51515210e-01 1.33835852e-01 -8.88717175e-01 -2.12906629e-01
3.37452255e-02 6.69047475e-01 -5.69449067e-01 -3.23824912e-01
-7.11993992e-01 3.65774035e-01 3.87638569e-01 1.01056719e+00
-6.01606846e-01 -2.45743126e-01 -7.66443849e-01 -5.12734890e-01
9.19388533e-02 -5.43103456e-01 9.62162852e-01 -7.28509188e-01
-1.38376606e+00 9.74390328e-01 -4.48387623e-01 2.81076938e-01
8.34136680e-02 1.37696207e-01 -3.99050176e-01 2.19917729e-01
2.33973831e-01 -1.85338706e-01 1.01494789e+00 -1.13503802e+00
-2.09976315e-01 -5.63181460e-01 1.87358540e-02 1.56688020e-01
-6.31951511e-01 2.06966072e-01 -4.00180429e-01 -8.98476720e-01
8.48887980e-01 -1.15846848e+00 -5.02266407e-01 -3.73231113e-01
-5.97475767e-01 -5.29027823e-03 1.13055539e+00 -7.10745573e-01
1.03276873e+00 -2.22455478e+00 6.58640504e-01 3.55750889e-01
2.71970749e-01 -5.12966849e-02 -3.12101364e-01 2.49152541e-01
-5.57282686e-01 6.92626834e-02 -5.06024599e-01 -4.62959260e-01
-1.91378534e-01 4.23099369e-01 -3.71535242e-01 9.24599171e-01
-1.49275333e-01 2.36687213e-01 -8.66374671e-01 -3.14690292e-01
-6.84586028e-03 1.81336790e-01 -5.78578651e-01 2.01404750e-01
2.21617237e-01 1.03104457e-01 -6.07932806e-01 1.08231223e+00
1.18007660e+00 -6.60876483e-02 3.37975740e-01 -6.37087107e-01
-2.78615236e-01 -3.40509444e-01 -1.57532346e+00 1.27359986e+00
-1.66540742e-01 3.30563068e-01 8.33881140e-01 -1.18683231e+00
6.26469254e-01 4.09321040e-01 9.50981677e-01 -2.01702744e-01
2.74523169e-01 2.90764451e-01 -1.28412098e-01 -9.56135631e-01
6.57341480e-01 -6.70993865e-01 4.12940353e-01 8.04008663e-01
-2.50262111e-01 2.44345739e-02 4.04987961e-01 2.39531636e-01
5.04336596e-01 -4.26613003e-01 2.11213917e-01 -5.44023454e-01
8.77630115e-01 -9.38655622e-03 7.58814692e-01 7.30525553e-02
-1.79382831e-01 3.17758590e-01 5.02832711e-01 -3.00363183e-01
-7.33475804e-01 -5.76899946e-01 -3.83655190e-01 8.38117182e-01
4.10605967e-02 -1.83493212e-01 -3.04303110e-01 -5.37713587e-01
-4.21925373e-02 -1.10451631e-01 -3.38603973e-01 3.22481245e-01
-3.14060867e-01 -1.62074077e+00 2.47758366e-02 2.54904330e-01
6.67730391e-01 -2.39955202e-01 2.62621075e-01 -9.50910598e-02
-4.46706653e-01 -1.36148572e+00 -7.33087659e-01 -3.09260428e-01
-1.28718424e+00 -1.02655363e+00 -7.41958439e-01 -9.20076251e-01
1.14080596e+00 1.03027880e+00 5.09566247e-01 1.49018645e-01
-1.40217364e-01 7.28865504e-01 -1.00671045e-01 8.63949768e-03
1.88072771e-02 -3.75731021e-01 6.51041925e-01 6.02237403e-01
-5.03037423e-02 -3.82863611e-01 -2.41544127e-01 4.45300937e-01
-1.09924126e+00 7.69749358e-02 3.48914713e-01 1.04348695e+00
7.00124323e-01 4.36404943e-01 1.44113809e-01 -7.20874608e-01
9.73544896e-01 -4.82509762e-01 -8.31461251e-01 2.50193000e-01
-4.24860239e-01 5.27466275e-02 1.99567840e-01 -3.35803777e-01
-9.80944574e-01 1.72113031e-01 5.02370656e-01 -6.82592571e-01
7.85682976e-01 1.11061811e+00 -2.16694146e-01 -5.03346801e-01
2.11527199e-01 6.17932528e-02 1.35076925e-01 -6.44992232e-01
3.43883723e-01 9.11933899e-01 4.58269149e-01 -1.16550052e+00
9.56304908e-01 9.39059019e-01 2.17545897e-01 -1.19478917e+00
-9.47201192e-01 -7.82288015e-01 -4.99482334e-01 -2.58963197e-01
5.09558797e-01 -7.35389113e-01 -1.00608504e+00 3.72737885e-01
-1.14485264e+00 2.13943914e-01 2.45872512e-01 9.70940113e-01
-4.42664951e-01 8.59072208e-01 -7.45332241e-01 -1.09079099e+00
-3.00812513e-01 -1.26428378e+00 7.57775664e-01 -3.00293982e-01
6.46371901e-01 -1.25105333e+00 2.65412122e-01 7.94222832e-01
-2.80372262e-01 2.76926666e-01 7.39448607e-01 2.41966441e-01
-3.83372009e-01 -1.97615519e-01 -5.20642757e-01 8.12582195e-01
3.44341904e-01 2.09553227e-01 -7.05969691e-01 -5.70728600e-01
5.79236090e-01 -5.82582988e-02 7.16100752e-01 2.87750483e-01
1.16459417e+00 -5.25909722e-01 -2.48604603e-02 8.69080067e-01
1.49642885e+00 -2.19525713e-02 2.75064230e-01 7.49233365e-02
9.88470793e-01 6.04631126e-01 8.35755169e-01 6.44472718e-01
1.89068392e-01 3.06173414e-01 5.50238252e-01 -2.52046943e-01
4.62498188e-01 3.84174645e-01 5.49161315e-01 1.66480553e+00
-7.67676890e-01 1.04331456e-01 -7.59131968e-01 4.27808166e-01
-2.01009297e+00 -9.57887709e-01 -6.92547917e-01 2.06814957e+00
5.07288337e-01 -7.01139927e-01 -4.12182026e-02 3.05781424e-01
8.86988044e-01 2.95071185e-01 -3.56616467e-01 -5.67253053e-01
-2.15757757e-01 -1.64498851e-01 8.17976832e-01 7.92037129e-01
-1.12292826e+00 5.32816112e-01 7.53564835e+00 3.89761716e-01
-1.09488845e+00 2.21853614e-01 3.52037966e-01 1.92703933e-01
-3.51652503e-01 -8.46608058e-02 -3.47339004e-01 -6.02191780e-04
2.42329404e-01 -3.58936608e-01 8.10819805e-01 5.45947552e-01
5.66374242e-01 8.25374201e-02 -6.19603038e-01 1.52682662e+00
1.72080278e-01 -9.71005201e-01 1.15755506e-01 3.42239469e-01
1.06490266e+00 -1.06070094e-01 3.50703895e-01 -3.14369440e-01
1.55896783e-01 -6.88333929e-01 3.45231146e-01 1.72255948e-01
6.27146184e-01 -5.91869652e-01 4.28814143e-01 1.52450889e-01
-1.18645620e+00 -8.31492320e-02 -7.57234514e-01 -2.94309348e-01
1.07610136e-01 1.01490462e+00 -6.39930010e-01 7.49717712e-01
2.70354301e-01 1.03764403e+00 -1.29188359e-01 9.53958452e-01
-6.62240982e-02 5.93593061e-01 -3.49374533e-01 4.54080075e-01
3.40898603e-01 -1.17531383e+00 8.09638143e-01 7.68450499e-01
3.49570215e-01 8.00340056e-01 3.30944359e-01 4.27417666e-01
-1.57265142e-01 4.25423026e-01 -4.79965925e-01 -4.97442603e-01
-1.03393421e-01 1.33562708e+00 -5.18728316e-01 -6.07427135e-02
-5.37113488e-01 8.85636330e-01 7.01169595e-02 6.93771601e-01
-6.76233351e-01 8.59043002e-02 7.26592720e-01 -4.44074273e-01
-6.91614673e-02 -6.41402960e-01 -2.25288510e-01 -1.72288346e+00
1.91019624e-01 -8.39612067e-01 6.09194100e-01 -5.03759682e-01
-1.40017283e+00 3.01076591e-01 5.29498793e-02 -1.45357060e+00
5.32086134e-01 -9.66820240e-01 -3.71657163e-01 7.85080492e-01
-1.45329070e+00 -8.99466515e-01 -1.52504876e-01 7.44819582e-01
-4.26653698e-02 -2.81042039e-01 4.20404553e-01 5.61491728e-01
-1.25393140e+00 4.59725969e-02 4.97491717e-01 8.73639360e-02
2.01020062e-01 -1.06653249e+00 -3.84409457e-01 9.99201119e-01
-2.17989460e-01 8.64825904e-01 6.39398873e-01 -5.48833787e-01
-1.92027628e+00 -8.01569700e-01 5.44530988e-01 2.47297019e-01
9.92653191e-01 1.07637942e-01 -4.74494934e-01 8.17203462e-01
2.67692834e-01 4.73779410e-01 1.02958524e+00 1.46749347e-01
-4.11198825e-01 -1.78285599e-01 -1.17331016e+00 3.71851474e-01
8.21576953e-01 -5.10337114e-01 -2.80634582e-01 9.88182068e-01
3.54942232e-01 -4.08879966e-01 -1.08300519e+00 4.03425008e-01
5.82540929e-01 -5.81237972e-01 1.06032610e+00 -1.02025938e+00
2.01244593e-01 -6.03772283e-01 -6.87722445e-01 -1.13001466e+00
-3.79667193e-01 -8.96603823e-01 -3.76678109e-02 8.65047216e-01
1.43749952e-01 -8.57894599e-01 5.67859709e-01 7.11451948e-01
-9.26576182e-02 -6.22098625e-01 -1.11846364e+00 -9.72957551e-01
-3.87245327e-01 -5.37298620e-01 3.41620624e-01 1.37607288e+00
3.04244637e-01 2.98548251e-01 -7.87075877e-01 4.70072389e-01
1.13771558e+00 2.23748088e-01 3.34659189e-01 -8.90219092e-01
-2.59048253e-01 8.57850686e-02 -3.18612695e-01 -9.40419316e-01
4.06477123e-01 -1.07840526e+00 -3.80578250e-01 -1.39170671e+00
5.75546980e-01 -4.65364814e-01 -4.60783005e-01 4.72930044e-01
2.64323317e-02 3.94858301e-01 2.83187389e-01 5.35961211e-01
-3.19958329e-01 5.91394246e-01 1.66508794e+00 -4.88750428e-01
-1.23710811e-01 -2.31277540e-01 -5.47409356e-01 7.33211339e-01
6.90096736e-01 -3.79631758e-01 -5.38593829e-01 -9.27281857e-01
2.60172188e-01 2.93394327e-01 1.12692758e-01 -3.22285354e-01
-1.14655122e-01 -7.51925826e-01 -1.46969244e-01 -7.18437135e-01
5.06664813e-01 -9.96739149e-01 8.02995563e-02 5.40221393e-01
8.22041780e-02 1.81119740e-01 -6.43671155e-02 4.14613932e-01
-2.93064743e-01 -3.03083301e-01 7.76965261e-01 2.49045163e-01
-4.64373916e-01 6.39868736e-01 -7.91467577e-02 -1.96256742e-01
8.63594234e-01 1.60436090e-02 -3.53444964e-01 -1.14278421e-01
-9.64658558e-01 2.36960858e-01 1.11514792e-01 1.17947133e-02
8.01289141e-01 -1.82856452e+00 -6.95625722e-01 2.26156116e-01
-1.27416536e-01 -4.16391492e-01 2.01808423e-01 1.38044393e+00
-6.73579156e-01 4.04649764e-01 1.01152867e-01 -4.84665394e-01
-1.44743538e+00 6.33999169e-01 3.18198204e-01 1.05102658e-01
-1.86339960e-01 5.91644585e-01 1.88478716e-02 -4.59464103e-01
-1.23603433e-01 -3.14716876e-01 -1.49590641e-01 2.28877947e-01
5.26674926e-01 8.27336311e-01 -2.71011726e-03 -1.00865483e+00
-3.58858913e-01 6.99869275e-01 1.84915215e-01 -3.45146924e-01
1.49164689e+00 -1.99438915e-01 -1.20203054e+00 2.88735747e-01
1.54381979e+00 1.23646602e-01 -6.10502183e-01 -1.53207377e-01
-2.37927884e-01 -8.12403142e-01 4.05948102e-01 -3.20864804e-02
-1.55320370e+00 6.83284223e-01 3.92473251e-01 3.42440844e-01
1.24870384e+00 -5.52168429e-01 6.90790892e-01 5.40879369e-01
3.49685043e-01 -9.28773463e-01 -9.52078179e-02 4.68411744e-01
1.06710446e+00 -1.13320482e+00 5.74993968e-01 -8.52553070e-01
-4.21660990e-01 1.20971119e+00 3.23472440e-01 1.10067114e-01
9.16910827e-01 7.02265948e-02 9.78612676e-02 -2.22458661e-01
-6.04800105e-01 -2.11161509e-01 2.10648209e-01 3.14627111e-01
4.76053268e-01 2.45015189e-01 -8.09889555e-01 7.57891834e-02
1.26131266e-01 -6.00716434e-02 8.21171999e-01 8.82592678e-01
-1.64354593e-01 -1.15745938e+00 -1.14910281e+00 2.94402957e-01
-5.36178291e-01 -6.77063689e-03 -3.04309368e-01 2.04341248e-01
-1.30871907e-01 1.07679081e+00 -3.65457386e-01 -2.70210564e-01
5.12337424e-02 -2.83255666e-01 5.92688322e-01 -4.96656537e-01
1.68512166e-01 2.82030523e-01 7.33072460e-02 -1.04956821e-01
-9.64600801e-01 -7.87383199e-01 -1.26385653e+00 -5.19830525e-01
-5.39096534e-01 4.79118347e-01 6.38351858e-01 6.83204651e-01
1.04847074e-01 -5.65470643e-02 1.20949793e+00 -7.12472081e-01
-5.68957031e-01 -6.14247561e-01 -1.02994943e+00 1.62769452e-01
4.67248321e-01 -7.23293602e-01 -4.98166233e-01 1.54336214e-01] | [7.438353061676025, 4.47137451171875] |
5b538c79-32e7-4a15-93df-9f525966d563 | on-the-equivalence-between-node-embeddings-1 | 1910.00452 | null | https://arxiv.org/abs/1910.00452v3 | https://arxiv.org/pdf/1910.00452v3.pdf | On the Equivalence between Positional Node Embeddings and Structural Graph Representations | This work provides the first unifying theoretical framework for node (positional) embeddings and structural graph representations, bridging methods like matrix factorization and graph neural networks. Using invariant theory, we show that the relationship between structural representations and node embeddings is analogous to that of a distribution and its samples. We prove that all tasks that can be performed by node embeddings can also be performed by structural representations and vice-versa. We also show that the concept of transductive and inductive learning is unrelated to node embeddings and graph representations, clearing another source of confusion in the literature. Finally, we introduce new practical guidelines to generating and using node embeddings, which fixes significant shortcomings of standard operating procedures used today. | ['Bruno Ribeiro', 'Balasubramaniam Srinivasan'] | 2019-10-01 | on-the-equivalence-between-positional-node | https://openreview.net/forum?id=SJxzFySKwH | https://openreview.net/pdf?id=SJxzFySKwH | iclr-2020-1 | ['triad-prediction'] | ['graphs'] | [ 1.36185125e-01 6.61673844e-01 -5.79036415e-01 -1.20350227e-01
2.04691306e-01 -7.76533544e-01 8.69244993e-01 3.67069483e-01
-6.49252832e-02 6.00340903e-01 5.23790359e-01 -8.31861734e-01
-4.01497871e-01 -1.21607745e+00 -3.66773605e-01 -6.96041644e-01
-5.79311371e-01 4.22098011e-01 -3.75535488e-02 -3.04450452e-01
-6.67648623e-03 7.17206419e-01 -1.21597517e+00 -3.09608638e-01
4.36611414e-01 3.87076467e-01 -3.86142910e-01 8.98702145e-01
-4.42843765e-01 8.11688840e-01 -5.08159399e-01 -4.39704835e-01
-1.75098889e-02 -3.03541124e-01 -9.47874248e-01 -1.54131278e-01
2.84328282e-01 -1.37226209e-01 -1.02835572e+00 9.79967058e-01
1.45869777e-01 3.28027725e-01 1.08332038e+00 -1.85205448e+00
-1.65238464e+00 8.05968165e-01 -2.79940039e-01 2.22044796e-01
5.47188818e-01 -6.27510667e-01 1.47423196e+00 -7.55854249e-01
6.63350403e-01 1.42183113e+00 8.96496475e-01 5.79037786e-01
-1.46410084e+00 -7.36254528e-02 2.34866470e-01 -6.24813177e-02
-1.19784820e+00 -4.04579751e-02 7.64866292e-01 -5.46374559e-01
8.00127745e-01 2.56632566e-01 7.51705468e-01 1.04589772e+00
1.70506880e-01 6.82277083e-01 5.70088208e-01 -6.61840320e-01
8.95192176e-02 1.83047995e-01 5.44312119e-01 9.42347467e-01
8.04269493e-01 1.47264645e-01 -4.44344580e-02 -4.88927394e-01
9.20914352e-01 3.30937177e-01 -1.78259790e-01 -9.42925334e-01
-9.74474013e-01 1.32040250e+00 6.56673431e-01 7.71263063e-01
-1.11553110e-01 6.02875710e-01 5.59069097e-01 4.89419609e-01
4.04401869e-01 5.21259427e-01 -3.28022659e-01 2.86788404e-01
-2.75702447e-01 -1.85829878e-01 1.04496443e+00 8.73442054e-01
8.67849290e-01 3.54003757e-01 2.36176476e-01 5.01374245e-01
4.53038663e-01 4.22665536e-01 2.81178653e-01 -5.90681314e-01
-1.46694019e-01 7.33391762e-01 -3.39777410e-01 -1.31790602e+00
-3.58256191e-01 -3.16452682e-01 -7.08484530e-01 -1.86780363e-01
1.23043060e-01 -5.96479662e-02 -7.44346917e-01 1.78235328e+00
9.88322496e-02 1.61329538e-01 -7.15547299e-04 3.06793183e-01
9.75171506e-01 6.04188740e-01 3.46666239e-02 -1.02417268e-01
1.15661752e+00 -6.43356740e-01 -8.61647546e-01 7.56011009e-02
9.23030734e-01 -2.47820377e-01 6.52729392e-01 -2.16676578e-01
-6.35119319e-01 -3.69930029e-01 -9.82116044e-01 -1.92727089e-01
-8.37988675e-01 -2.42136240e-01 1.38010764e+00 9.32192683e-01
-1.53409517e+00 5.78579068e-01 -7.34052002e-01 -8.58017385e-01
1.58532947e-01 4.32180136e-01 -8.25956464e-01 -2.82396916e-02
-1.16339779e+00 1.00509274e+00 4.07867759e-01 -1.16849720e-01
-6.74780309e-01 -7.05169022e-01 -1.33139801e+00 6.72810897e-02
3.61809656e-02 -6.34744465e-01 6.47030056e-01 -6.40782952e-01
-9.42266464e-01 8.09702456e-01 -8.17843303e-02 -4.14345413e-01
-3.22237074e-01 -4.17181589e-02 -5.20585120e-01 2.76031673e-01
-3.89529914e-02 4.04486060e-01 7.66638517e-01 -1.21481955e+00
-3.02093662e-02 -3.64667714e-01 4.28660214e-01 -1.42767742e-01
-8.93989444e-01 -4.07463580e-01 6.57716393e-02 -4.74408656e-01
1.07092001e-01 -7.64996946e-01 -3.75391752e-01 8.12737867e-02
-4.08549786e-01 -6.01118267e-01 9.89856660e-01 -3.31793368e-01
1.23851049e+00 -2.01842070e+00 5.06823719e-01 4.19683903e-01
9.56344485e-01 1.20311104e-01 -5.26255846e-01 1.09463429e+00
-5.93700409e-01 6.99995637e-01 1.22257642e-01 -1.05008297e-01
3.58923376e-01 6.01314723e-01 -3.00722837e-01 8.53907406e-01
1.25083402e-01 1.25885344e+00 -1.21901560e+00 -3.41938287e-01
3.88257921e-01 7.79793143e-01 -5.90500712e-01 -3.78417224e-02
2.72389829e-01 -2.65348375e-01 -3.62244636e-01 6.43192530e-01
4.79780018e-01 -2.68925816e-01 7.26858675e-01 -2.07287624e-01
3.58377367e-01 2.54879862e-01 -9.21434820e-01 1.46615541e+00
-3.65469456e-01 9.34563279e-01 8.69503319e-02 -1.52513707e+00
9.74767983e-01 3.08568269e-01 4.49170470e-01 -1.47498503e-01
1.91992491e-01 -1.84027001e-01 -1.08513184e-01 -2.24529505e-01
6.33710563e-01 -2.74435192e-01 -2.10919559e-01 9.12920475e-01
4.89634782e-01 -4.57511395e-02 1.78565785e-01 9.12684917e-01
1.35435307e+00 -1.82527870e-01 5.06213844e-01 -4.98197258e-01
1.86141178e-01 -2.96527982e-01 1.43048599e-01 6.28202915e-01
-2.06452787e-01 6.33298904e-02 9.79102075e-01 -5.05075932e-01
-9.53994215e-01 -1.66492736e+00 -1.38439238e-01 1.32620895e+00
2.62594316e-02 -9.63042200e-01 -3.76300901e-01 -8.71180534e-01
4.04463708e-01 4.03983802e-01 -1.33256555e+00 -3.91835660e-01
-1.00682996e-01 -4.17679042e-01 6.12342238e-01 8.84651423e-01
-5.24821639e-01 -7.30258942e-01 1.92047507e-01 -1.44829452e-01
2.96353579e-01 -7.39832103e-01 -7.46032968e-02 3.33509177e-01
-1.09179950e+00 -1.21414876e+00 -2.42670789e-01 -1.08951199e+00
1.06483424e+00 5.68781734e-01 1.37735868e+00 5.12919247e-01
-3.72075468e-01 1.04361486e+00 -6.21633828e-01 9.23489109e-02
-4.97275859e-01 2.12722868e-01 4.01953816e-01 -3.61774474e-01
4.84310448e-01 -9.90478098e-01 -2.26472735e-01 -4.83044945e-02
-1.18855906e+00 -6.37761474e-01 3.21588665e-01 9.18328166e-01
1.02234066e-01 -1.27131328e-01 5.83792090e-01 -1.17272556e+00
1.10956502e+00 -7.48629332e-01 -2.17683837e-01 2.75572866e-01
-5.58040142e-01 3.23281556e-01 5.00772297e-01 -3.58134747e-01
-3.13374162e-01 -3.04157227e-01 6.49167001e-02 -3.82268131e-01
2.08418190e-01 6.65803492e-01 6.68895096e-02 -2.46346891e-01
8.49439144e-01 -3.48132849e-02 4.49323714e-01 -2.71460980e-01
1.17113268e+00 3.44984263e-01 1.74463779e-01 -6.07177854e-01
1.34159482e+00 4.41050798e-01 1.78816974e-01 -1.15551960e+00
-4.67217416e-01 -3.10394317e-01 -1.00032949e+00 8.35777372e-02
7.78490365e-01 -5.42456567e-01 -7.21029162e-01 -3.94985497e-01
-1.00094080e+00 6.88178092e-02 -7.22710669e-01 2.71179378e-01
-2.75120378e-01 7.05313802e-01 -7.50166714e-01 -8.47184241e-01
5.94212674e-03 -5.67241967e-01 7.19165623e-01 -5.15825525e-02
-3.96097690e-01 -2.01220036e+00 4.06570941e-01 -1.04610570e-01
5.52009225e-01 2.90649682e-01 1.27077174e+00 -8.03101659e-01
-1.53904870e-01 -5.17705858e-01 -3.56163353e-01 4.01683182e-01
2.48903155e-01 2.84827143e-01 -8.10117662e-01 -4.62837726e-01
-4.90815073e-01 -1.69423133e-01 8.06513727e-01 2.28303537e-01
6.05557978e-01 -1.73460782e-01 -6.31202698e-01 4.18035775e-01
1.43150294e+00 -1.97210848e-01 6.57580912e-01 -2.02999473e-01
9.81147289e-01 5.40040374e-01 -1.82229429e-01 1.59525976e-01
3.20506364e-01 3.67824808e-02 3.08534980e-01 -1.29250318e-01
-6.33040816e-02 -6.40356719e-01 3.03359598e-01 1.11193299e+00
-4.03131321e-02 -1.29014686e-01 -6.45876527e-01 7.95640230e-01
-1.63442373e+00 -1.14486206e+00 -2.19279632e-01 1.84779286e+00
2.19187170e-01 -2.12151885e-01 1.91706404e-01 2.73988694e-01
7.77117372e-01 3.72871101e-01 -4.81104739e-02 -7.75498986e-01
6.67162389e-02 5.91053426e-01 4.69514102e-01 6.75177991e-01
-1.15269375e+00 9.49005365e-01 8.79580212e+00 3.97855461e-01
-5.40446758e-01 9.38991979e-02 2.93243706e-04 4.55867559e-01
-9.91729498e-01 2.13047653e-01 -1.74627587e-01 -2.50106633e-01
9.88016784e-01 -4.74544168e-01 3.61948967e-01 1.04234028e+00
-3.85991037e-01 4.46604460e-01 -1.54255140e+00 7.18637943e-01
2.75288135e-01 -1.38592207e+00 1.40580222e-01 3.44379246e-01
6.21461153e-01 -1.12820782e-01 1.30145252e-01 4.12444562e-01
7.75862455e-01 -1.35790181e+00 -4.76046167e-02 1.64555252e-01
6.78006291e-01 -4.78957355e-01 7.22530246e-01 -3.23562831e-01
-1.51285088e+00 -4.13098857e-02 -6.60511851e-01 -3.40713561e-01
-7.09544495e-02 6.63832605e-01 -8.70125115e-01 6.86710477e-01
3.56162563e-02 1.12561798e+00 -6.29652917e-01 6.74582064e-01
-4.80852187e-01 5.49484730e-01 -1.31564766e-01 -3.18157494e-01
2.33416352e-02 -4.29341108e-01 4.73622918e-01 1.21112800e+00
-6.73171878e-03 -2.85268873e-01 8.63255709e-02 8.47619534e-01
-1.88134640e-01 -2.49513742e-02 -1.48802614e+00 -6.50761008e-01
6.54903293e-01 1.43841779e+00 -8.15527380e-01 -3.21089000e-01
-4.60291356e-01 8.25769961e-01 4.40699816e-01 5.10728359e-01
-4.92823482e-01 -6.06103897e-01 9.60229874e-01 7.25945011e-02
3.69940281e-01 -5.19295752e-01 1.05305448e-01 -1.21727252e+00
-3.52946401e-01 -2.61016190e-01 4.31381166e-01 -4.98266757e-01
-1.71069527e+00 4.02644098e-01 -3.97326387e-02 -8.22197258e-01
-2.41261601e-01 -1.08564067e+00 -6.04964197e-01 4.82861847e-01
-1.08262503e+00 -9.95258808e-01 8.71599242e-02 6.10624135e-01
-2.99217373e-01 6.34253994e-02 1.31551540e+00 1.53060615e-01
-3.92033905e-01 5.36369622e-01 1.99598745e-01 6.10579729e-01
8.85846764e-02 -1.60637808e+00 4.72726554e-01 5.78300118e-01
5.84125638e-01 1.09780622e+00 5.23572266e-01 -3.91255230e-01
-2.04428315e+00 -7.98654556e-01 8.45686793e-01 -6.31460309e-01
1.30935872e+00 -6.16126657e-01 -6.73944712e-01 1.17658794e+00
2.41591096e-01 5.11868954e-01 1.05924666e+00 8.04250598e-01
-8.28273773e-01 1.36206985e-01 -9.27199006e-01 4.37099040e-01
1.34865785e+00 -1.21370077e+00 -7.46023059e-01 4.17636842e-01
9.72546995e-01 3.53453368e-01 -1.13190866e+00 4.46409732e-02
5.26048958e-01 -5.49633980e-01 1.22041392e+00 -1.08727920e+00
2.66213007e-02 -7.17799217e-02 -3.31433058e-01 -1.43041790e+00
-8.04463744e-01 -5.20717740e-01 -3.37030649e-01 1.00077689e+00
3.98913741e-01 -1.03539622e+00 7.34817743e-01 3.41481894e-01
1.28529727e-01 -4.11562353e-01 -6.05065405e-01 -8.43732893e-01
3.13460588e-01 -3.70274454e-01 3.64757448e-01 1.29760444e+00
5.25312245e-01 6.44397736e-01 -5.13657443e-02 -9.58607867e-02
5.21472275e-01 -2.95250025e-03 7.17553318e-01 -1.77045989e+00
-1.22675121e-01 -4.60094750e-01 -1.19633961e+00 -9.57527876e-01
6.28650248e-01 -1.50370145e+00 -5.58269083e-01 -2.00720906e+00
-8.83798487e-03 -4.66636270e-01 -3.75709623e-01 3.98629397e-01
2.02037618e-01 3.68684858e-01 -4.51638550e-03 -1.71717748e-01
-5.79950154e-01 5.44209898e-01 9.36282158e-01 -1.03819370e-01
1.99877638e-02 -4.85863060e-01 -1.02436817e+00 5.48716307e-01
5.94761491e-01 -3.43340039e-01 -7.65356243e-01 -1.80648535e-01
5.47607243e-01 -1.84355900e-01 3.13776702e-01 -4.37791139e-01
4.37338538e-02 1.02868706e-01 1.08690128e-01 -2.54598826e-01
1.21093340e-01 -8.39146554e-01 -1.65723741e-01 4.87732440e-01
-3.48720491e-01 4.48228240e-01 -2.73677167e-02 9.01987851e-01
3.28697711e-02 -2.84189701e-01 1.30901128e-01 7.81909749e-02
-6.64980114e-01 3.83714616e-01 -5.41264474e-01 1.52984068e-01
1.02222121e+00 -2.03068405e-01 -5.98064303e-01 -4.38230217e-01
-9.27911580e-01 -3.39186303e-02 4.46494937e-01 4.71278906e-01
6.84361517e-01 -1.72322941e+00 -4.49423581e-01 1.59490407e-01
-4.14939038e-02 -6.94171667e-01 -1.12556837e-01 7.88722277e-01
-4.64725256e-01 2.09429234e-01 -9.50645357e-02 -2.58184940e-01
-1.05031466e+00 1.06629384e+00 6.09045662e-02 3.15060243e-02
-6.81375623e-01 6.29427016e-01 8.26063603e-02 -7.45913982e-01
1.51244238e-01 -3.09334725e-01 -1.58263505e-01 3.67990017e-01
3.75208288e-01 3.59433055e-01 -2.79621601e-01 -5.50400734e-01
-6.55438304e-01 3.95769060e-01 5.77265099e-02 1.43351048e-01
1.34288740e+00 1.44466817e-01 -7.14576244e-01 5.94792962e-01
1.58922720e+00 7.47851357e-02 -3.03584486e-01 -1.67816132e-01
-9.44639519e-02 -3.58886212e-01 -1.03306897e-01 -3.99776995e-02
-1.17786407e+00 1.00949347e+00 1.58569947e-01 1.15064979e+00
5.76347709e-01 4.10819173e-01 3.36620837e-01 6.07060671e-01
2.80695409e-01 -7.49028504e-01 1.46834850e-01 3.59634519e-01
5.84107459e-01 -7.23696291e-01 1.78253859e-01 -6.34769738e-01
-1.56552449e-01 1.30045223e+00 9.05719697e-02 -6.23904526e-01
1.15762305e+00 2.66607046e-01 -3.72838527e-01 -5.10906816e-01
-6.29429221e-01 -2.84058243e-01 2.32265726e-01 1.23798239e+00
6.76913381e-01 5.10941386e-01 -2.01321334e-01 1.56595513e-01
-2.35943094e-01 -3.78846020e-01 6.10937536e-01 8.96254301e-01
-3.68086189e-01 -1.35612690e+00 -1.41625270e-01 5.96305013e-01
-4.79393750e-02 -1.26094684e-01 -5.66449761e-01 1.04901779e+00
-3.17601591e-01 9.10226643e-01 3.95419389e-01 -8.34663749e-01
-1.09010696e-01 1.73087910e-01 5.74207306e-01 -7.89288282e-01
-7.02038631e-02 -6.41397834e-01 1.51462719e-01 -3.11742097e-01
-5.44019878e-01 -2.24819437e-01 -1.14521313e+00 -8.83255363e-01
-4.33448017e-01 4.85956103e-01 3.74198198e-01 6.40569508e-01
3.31410348e-01 5.90051711e-01 4.38619614e-01 -6.76358879e-01
-2.14425758e-01 -8.11002791e-01 -9.53397930e-01 3.83151710e-01
1.60249382e-01 -8.68140042e-01 -7.24964082e-01 -2.30036810e-01] | [7.091310977935791, 6.124690532684326] |
cdd48b7a-05c8-4926-a58d-25de3b0d5678 | spoof-detection-using-x-vector-and-feature | 1904.07453 | null | https://arxiv.org/abs/1904.07453v2 | https://arxiv.org/pdf/1904.07453v2.pdf | Spoof detection using time-delay shallow neural network and feature switching | Detecting spoofed utterances is a fundamental problem in voice-based biometrics. Spoofing can be performed either by logical accesses like speech synthesis, voice conversion or by physical accesses such as replaying the pre-recorded utterance. Inspired by the state-of-the-art \emph{x}-vector based speaker verification approach, this paper proposes a time-delay shallow neural network (TD-SNN) for spoof detection for both logical and physical access. The novelty of the proposed TD-SNN system vis-a-vis conventional DNN systems is that it can handle variable length utterances during testing. Performance of the proposed TD-SNN systems and the baseline Gaussian mixture models (GMMs) is analyzed on the ASV-spoof-2019 dataset. The performance of the systems is measured in terms of the minimum normalized tandem detection cost function (min-t-DCF). When studied with individual features, the TD-SNN system consistently outperforms the GMM system for physical access. For logical access, GMM surpasses TD-SNN systems for certain individual features. When combined with the decision-level feature switching (DLFS) paradigm, the best TD-SNN system outperforms the best baseline GMM system on evaluation data with a relative improvement of 48.03\% and 49.47\% for both logical and physical access, respectively. | ['B. Bharathi', 'Suvidha Rupesh Kumar', 'Hema A. Murthy', 'Saranya M', 'Mari Ganesh Kumar'] | 2019-04-16 | null | null | null | null | ['voice-anti-spoofing'] | ['audio'] | [ 9.52485278e-02 -2.75349557e-01 -1.46414652e-01 -2.98771590e-01
-5.14421225e-01 -5.61139941e-01 6.86018944e-01 -1.58508033e-01
-4.26124096e-01 2.54273057e-01 -1.68816876e-02 -7.47447550e-01
-3.09527684e-02 -3.13679487e-01 -1.54243171e-01 -7.75438011e-01
1.45165950e-01 2.60545403e-01 1.65794000e-01 -2.11227641e-01
-4.20720363e-03 8.47728133e-01 -1.41900957e+00 1.72234833e-01
3.77443433e-01 1.05541110e+00 -1.00031123e-01 9.42832112e-01
-1.98471218e-01 2.06570402e-01 -1.03360999e+00 -4.44461554e-01
4.17054556e-02 -3.61925334e-01 -5.60729980e-01 -2.64838427e-01
3.38459730e-01 -3.20918739e-01 -5.09868741e-01 1.17794037e+00
1.00067544e+00 2.44303137e-01 3.90392125e-01 -1.12608063e+00
-2.93161452e-01 6.49474978e-01 -2.00330868e-01 5.02482474e-01
5.53900421e-01 8.44001584e-03 6.20228887e-01 -8.97758901e-01
2.09958106e-01 1.57826090e+00 5.78099847e-01 9.48743582e-01
-1.10178828e+00 -8.80576909e-01 -4.33861226e-01 2.00226471e-01
-1.47599566e+00 -1.16098595e+00 8.07780385e-01 -2.11275294e-01
1.35410190e+00 4.82364833e-01 -7.61004910e-02 1.27430427e+00
1.56751528e-01 6.46223009e-01 8.11223328e-01 -5.94615519e-01
1.23232119e-01 4.59079176e-01 5.65617681e-01 5.02458274e-01
1.06086619e-01 4.38976437e-01 -7.72238195e-01 -3.61809105e-01
3.35488081e-01 -1.85671717e-01 -1.39796481e-01 3.77839386e-01
-8.80955517e-01 6.89080417e-01 -3.29582691e-01 5.05896449e-01
-4.83963877e-01 -1.65727198e-01 7.09251881e-01 4.46158558e-01
1.10172234e-01 1.48016289e-01 -4.07860667e-01 -3.65589768e-01
-1.28533638e+00 -1.06912740e-02 1.00420964e+00 5.36701560e-01
-4.64498773e-02 7.50685871e-01 -2.62251613e-03 1.02403378e+00
5.05419970e-01 8.49114299e-01 8.28062654e-01 -4.27301466e-01
3.26424569e-01 -2.38669813e-01 -7.15402737e-02 -8.80655706e-01
-3.61603528e-01 -3.56291562e-01 -7.46007383e-01 -1.21464901e-01
3.56163979e-01 -1.99288279e-01 -7.97701895e-01 1.51373816e+00
2.32590050e-01 1.44692138e-01 1.90013692e-01 3.92215312e-01
1.04659259e+00 5.72027981e-01 -3.81353438e-01 -4.13334161e-01
1.38726616e+00 -7.18603551e-01 -1.28629208e+00 1.93643391e-01
2.59786397e-01 -1.01673126e+00 7.54105389e-01 7.27237165e-01
-7.76959121e-01 -7.40954995e-01 -9.37321663e-01 4.45561737e-01
-4.72694457e-01 -1.51234418e-01 1.09932259e-01 1.86672342e+00
-1.26502323e+00 3.63170087e-01 -7.35244274e-01 -4.62440431e-01
5.12062907e-02 8.24756563e-01 -3.26728612e-01 3.07976484e-01
-1.47964942e+00 3.98416460e-01 2.07541913e-01 3.52976024e-02
-1.00347769e+00 -2.43894696e-01 -8.29651415e-01 3.87618281e-02
-9.33839232e-02 -1.61892436e-02 1.34346402e+00 -2.10118935e-01
-2.03864574e+00 6.34348392e-01 -5.82144797e-01 -6.82275534e-01
1.92895308e-01 -9.58966464e-02 -1.31801891e+00 1.97070703e-01
-4.38051045e-01 2.54512280e-01 1.26287198e+00 -8.55649769e-01
-3.59137297e-01 -2.40860716e-01 -4.60991889e-01 -4.22831863e-01
-6.72039807e-01 6.59764171e-01 1.37729449e-02 -3.95319432e-01
9.68232527e-02 -8.20112705e-01 3.25158745e-01 -6.43256843e-01
-8.91228855e-01 -3.19602579e-01 1.44756496e+00 -1.11399627e+00
1.32094109e+00 -2.39219451e+00 -4.43460375e-01 1.96555346e-01
-2.20733911e-01 1.20955539e+00 4.09274325e-02 3.46383274e-01
-1.89441845e-01 8.58645700e-03 9.88998786e-02 -7.13686764e-01
1.05616555e-01 -1.21562824e-01 -2.50635654e-01 7.06367195e-01
-2.19734788e-01 3.38242143e-01 -5.65482497e-01 -3.57629478e-01
5.82990050e-01 8.42024565e-01 -4.45628434e-01 2.69814819e-01
3.81966323e-01 2.52623141e-01 1.57017738e-01 8.40291679e-01
9.15764570e-01 3.91775668e-01 9.55185946e-03 4.01575118e-02
3.89207415e-02 7.54436374e-01 -1.24963355e+00 1.21122611e+00
-4.99064237e-01 8.06193352e-01 3.89120877e-01 -8.66955757e-01
1.16898060e+00 1.02795541e+00 1.60971135e-01 -2.96585798e-01
3.02920729e-01 4.53653097e-01 1.57513425e-01 -5.84633172e-01
4.59708780e-01 -1.38468266e-01 1.56887919e-01 2.46126860e-01
3.73506337e-01 1.06624968e-01 -3.77752930e-01 -1.84715778e-01
9.10300612e-01 -5.89734554e-01 1.72020361e-01 -3.14633548e-01
1.03920448e+00 -7.57504284e-01 3.36966693e-01 8.60726595e-01
-6.77371085e-01 1.31229371e-01 9.52632874e-02 6.91935793e-02
-4.86996114e-01 -1.01371753e+00 -5.27837038e-01 1.00326729e+00
-1.36569411e-01 -1.31589502e-01 -9.55457330e-01 -7.17131734e-01
7.51487911e-02 8.74848425e-01 -6.03670552e-02 -3.03174672e-03
-5.29309630e-01 -4.62169677e-01 1.32060993e+00 6.64672554e-02
6.16886318e-01 -1.14565051e+00 -2.58429736e-01 3.07204187e-01
-1.31093130e-01 -1.30200601e+00 -5.18916428e-01 2.51127273e-01
-5.75655043e-01 -4.97380793e-01 -5.21230757e-01 -7.24060476e-01
-1.01364963e-02 2.66930789e-01 4.34693515e-01 -2.39618450e-01
-1.47888614e-02 3.28471884e-02 -1.97742328e-01 -4.92973715e-01
-8.64079893e-01 1.12818470e-02 8.60692143e-01 4.18370903e-01
9.01406884e-01 -3.88425589e-01 -2.44107246e-01 3.85276020e-01
-7.82626331e-01 -8.54063332e-01 -7.58527592e-03 1.03705382e+00
-1.23429537e-01 3.96329045e-01 6.73656642e-01 -4.04502809e-01
7.83581376e-01 -2.60861844e-01 -3.42160374e-01 -8.94183517e-02
-6.60261214e-01 -1.38296545e-01 4.79678482e-01 -6.11994863e-01
-1.01605523e+00 -3.73530030e-01 -7.84008026e-01 -5.37366509e-01
-5.84633231e-01 1.04186893e-01 -6.34788632e-01 -1.74995586e-01
5.04134059e-01 4.80232567e-01 1.65460795e-01 -8.58739138e-01
-2.80258227e-02 1.69620180e+00 7.21525371e-01 -7.82984123e-02
5.81034124e-01 3.34000975e-01 -3.31113815e-01 -1.62823379e+00
1.17133848e-01 -7.82782078e-01 -4.08647776e-01 1.27327174e-01
6.75335228e-01 -3.54796290e-01 -1.18581855e+00 1.28400302e+00
-1.33703721e+00 1.11167639e-01 8.11140577e-04 6.83878124e-01
-1.07074626e-01 7.59258866e-01 -9.71771657e-01 -1.47010398e+00
-5.48449636e-01 -1.46322930e+00 8.73867393e-01 7.99368694e-02
-3.61968577e-01 -1.04266334e+00 -2.31736794e-01 4.48166490e-01
7.61963546e-01 -3.31215650e-01 5.51785707e-01 -1.57405531e+00
4.26407337e-01 -6.23297751e-01 1.10794999e-01 9.34764266e-01
5.04134476e-01 -9.69019532e-02 -1.61397099e+00 -5.87076008e-01
5.21413088e-01 1.14419453e-01 7.41777837e-01 5.85538924e-01
8.36728275e-01 -2.66752541e-01 -2.08999515e-01 6.56879663e-01
9.76259708e-01 7.44445384e-01 3.33034486e-01 6.67675165e-04
5.07006586e-01 5.69281638e-01 9.98481438e-02 3.77750814e-01
-1.80874899e-01 9.48226690e-01 3.27924222e-01 1.90048382e-01
-2.36715108e-01 -5.17511517e-02 8.03454697e-01 9.46761966e-01
4.65677083e-01 -5.89760005e-01 -7.77537704e-01 3.80899072e-01
-1.01796949e+00 -9.73840296e-01 1.07510888e-03 2.42925882e+00
4.38067615e-01 1.87279508e-01 3.65586400e-01 9.75996435e-01
1.17446613e+00 4.60923851e-01 -2.07807168e-01 -9.53594863e-01
-4.56534103e-02 3.65955323e-01 3.91549557e-01 9.32797492e-01
-1.22872519e+00 9.02470350e-01 5.87429047e+00 1.02053511e+00
-1.49674582e+00 5.10904312e-01 2.28462026e-01 -8.64731148e-02
3.46862346e-01 -6.13940597e-01 -1.33264291e+00 6.05141222e-01
1.65721154e+00 3.53001684e-01 5.53989410e-01 6.33825839e-01
3.41232508e-01 3.18535924e-01 -9.59596217e-01 1.24020207e+00
2.92211533e-01 -9.02927995e-01 -1.39465213e-01 4.06420082e-01
9.75898579e-02 -7.28259459e-02 3.60316128e-01 1.09399915e-01
-6.16739430e-02 -1.00377536e+00 5.21519661e-01 -5.17872460e-02
1.11190593e+00 -7.14179277e-01 1.12204504e+00 1.95215642e-01
-1.02110672e+00 -1.30784467e-01 -7.68777058e-02 3.17515373e-01
3.16052765e-01 5.77308595e-01 -1.10651028e+00 4.72633749e-01
4.07006621e-01 1.91190261e-02 2.91834865e-02 7.10512638e-01
8.71715248e-02 1.08304942e+00 -5.52599311e-01 -1.78903416e-01
2.79393010e-02 3.28523129e-01 9.04291570e-01 1.74815011e+00
3.65362614e-01 -3.88772845e-01 -2.82853514e-01 4.53881204e-01
1.66104689e-01 -1.05878487e-01 -4.57130045e-01 -2.05679953e-01
8.58413458e-01 7.30727255e-01 -2.42650777e-01 -3.18460464e-01
3.29499207e-02 1.07356191e+00 -5.38683057e-01 3.52655411e-01
-5.60666680e-01 -8.62663150e-01 8.42399776e-01 -3.29211466e-02
3.92205030e-01 -1.04426458e-01 3.84297483e-02 -8.14107358e-01
-2.05980182e-01 -1.13955331e+00 1.17573731e-01 2.83645876e-02
-8.48964334e-01 9.69782829e-01 -1.90616131e-01 -9.01002169e-01
-4.56030548e-01 -7.12602675e-01 -4.93173152e-01 1.12976170e+00
-1.34746766e+00 -8.24739933e-01 1.80258662e-01 7.98197448e-01
6.43556118e-01 -7.78962076e-01 1.08955443e+00 5.49645603e-01
-7.50853121e-01 1.35836995e+00 1.94735005e-01 3.35415304e-01
5.12664497e-01 -8.67943704e-01 7.71401465e-01 1.04764438e+00
1.38360754e-01 9.09880102e-01 9.30524766e-01 -3.27599794e-01
-1.36697364e+00 -6.45819068e-01 1.37827444e+00 -8.28601494e-02
2.89021790e-01 -3.80117655e-01 -8.29792321e-01 1.31358251e-01
1.51909843e-01 5.81912063e-02 8.39716136e-01 -2.26426110e-01
-5.83375752e-01 -1.70682684e-01 -1.81380010e+00 -1.04328729e-02
5.76484263e-01 -1.12853837e+00 -4.43570375e-01 1.93201497e-01
8.35754812e-01 -1.40367299e-01 -6.24009728e-01 1.79760933e-01
7.43539691e-01 -1.10337925e+00 1.04928601e+00 -5.55960476e-01
-4.63828027e-01 6.86789453e-02 -6.16756499e-01 -1.03938901e+00
4.34334800e-02 -1.29252267e+00 -4.52385724e-01 1.62035763e+00
2.11003542e-01 -9.26229715e-01 6.91953778e-01 8.07658955e-02
8.27009454e-02 -2.49605551e-01 -1.47130549e+00 -1.35155845e+00
-1.71654671e-01 -8.51009309e-01 9.06331003e-01 8.33066702e-01
1.16863184e-01 -9.01558399e-02 -6.53081059e-01 4.44837570e-01
6.26154661e-01 -7.59513676e-01 4.95881200e-01 -1.07788730e+00
-5.17717779e-01 -3.72594208e-01 -6.17084265e-01 -1.13243628e+00
2.34021559e-01 -6.58129334e-01 2.28457320e-02 -7.04354107e-01
-5.05566537e-01 -8.95036608e-02 -7.28290319e-01 7.40569457e-03
2.23828495e-01 -4.62154150e-02 1.88694447e-01 -1.05571523e-02
1.72233969e-01 8.23468268e-02 4.93234903e-01 -1.71242103e-01
-4.87875551e-01 4.55753356e-01 -2.08740477e-02 5.02771437e-01
7.92943776e-01 -3.95852417e-01 -1.43254980e-01 9.50369164e-02
-7.59974718e-01 5.70265234e-01 5.30608892e-02 -1.18556130e+00
1.53516874e-01 2.71944761e-01 -7.78048933e-02 -7.02180982e-01
7.27774262e-01 -7.00815678e-01 -7.86342919e-02 8.59604239e-01
-2.93931901e-01 -1.06259815e-01 1.08886473e-01 3.16064328e-01
-1.27133608e-01 -2.31942311e-01 9.82012510e-01 4.53374654e-01
-1.23684697e-01 -7.30520338e-02 -6.92580879e-01 -4.80846226e-01
4.79858518e-01 -3.96320105e-01 -3.16841453e-01 -5.16502798e-01
-5.99574149e-01 -6.77443922e-01 -7.05125853e-02 4.72388953e-01
8.31861496e-01 -1.00109243e+00 -4.98375416e-01 9.05932844e-01
-2.70119816e-01 -6.85462236e-01 1.97978452e-01 7.91164100e-01
-3.11753780e-01 1.17788708e+00 -1.98099297e-03 -6.27787709e-01
-1.78678811e+00 5.53910315e-01 4.69991475e-01 -4.35362533e-02
-7.64963776e-02 1.34065533e+00 -2.72775799e-01 -6.43681467e-01
7.32501745e-01 -1.47185773e-01 -6.19679727e-02 -1.04558550e-01
8.50750506e-01 7.32843816e-01 4.69127685e-01 -1.02126944e+00
-7.33734429e-01 -3.95479333e-03 -8.06392506e-02 -5.30349195e-01
8.59351218e-01 -9.77404192e-02 -5.18495515e-02 3.23433995e-01
1.44266915e+00 1.86108336e-01 -3.81292462e-01 -4.32580978e-01
-1.38699383e-01 -4.23904359e-01 1.49238020e-01 -6.41383827e-01
-8.89566123e-01 1.23289931e+00 1.13511419e+00 5.30251861e-01
7.73111761e-01 -2.52028525e-01 1.07986033e+00 3.77593458e-01
2.90876180e-01 -7.30382919e-01 -2.71296859e-01 5.31557441e-01
4.57693815e-01 -9.95063424e-01 -6.25657916e-01 -2.97925204e-01
-2.45066330e-01 1.01950228e+00 1.39864624e-01 3.84770334e-01
1.13518381e+00 2.00701237e-01 2.18634561e-01 -3.27115990e-02
-2.17431128e-01 4.14696068e-01 1.04949959e-01 8.40086520e-01
5.40931165e-01 3.43444705e-01 -5.63158700e-03 4.69829738e-01
-4.62221891e-01 -2.67941445e-01 3.19757521e-01 6.51677608e-01
-3.15036565e-01 -1.18700874e+00 -8.17129552e-01 4.00247514e-01
-9.49447751e-01 -1.92859635e-01 -3.21302235e-01 3.38554978e-01
7.17971101e-02 1.78903091e+00 -1.46599233e-01 -1.12687862e+00
6.15435913e-02 5.38548410e-01 -8.94461498e-02 -4.06770110e-01
-9.12295520e-01 2.41218194e-01 2.46950790e-01 -3.24537903e-01
-1.74768433e-01 -9.55001950e-01 -8.87163460e-01 -6.40029013e-01
-8.36617291e-01 3.23045075e-01 1.08769751e+00 1.07271266e+00
1.37914747e-01 3.90697122e-01 9.21645999e-01 -7.23886669e-01
-1.05427325e+00 -1.28127420e+00 -9.01609778e-01 4.07042988e-02
9.64455545e-01 -3.98523957e-01 -6.32569790e-01 -2.50568360e-01] | [14.112353324890137, 5.901759147644043] |
ffa29693-bdbc-45ee-aa3b-129790eb7b24 | sril-selective-regularization-for-class | 2305.05175 | null | https://arxiv.org/abs/2305.05175v1 | https://arxiv.org/pdf/2305.05175v1.pdf | SRIL: Selective Regularization for Class-Incremental Learning | Human intelligence gradually accepts new information and accumulates knowledge throughout the lifespan. However, deep learning models suffer from a catastrophic forgetting phenomenon, where they forget previous knowledge when acquiring new information. Class-Incremental Learning aims to create an integrated model that balances plasticity and stability to overcome this challenge. In this paper, we propose a selective regularization method that accepts new knowledge while maintaining previous knowledge. We first introduce an asymmetric feature distillation method for old and new classes inspired by cognitive science, using the gradient of classification and knowledge distillation losses to determine whether to perform pattern completion or pattern separation. We also propose a method to selectively interpolate the weight of the previous model for a balance between stability and plasticity, and we adjust whether to transfer through model confidence to ensure the performance of the previous class and enable exploratory learning. We validate the effectiveness of the proposed method, which surpasses the performance of existing methods through extensive experimental protocols using CIFAR-100, ImageNet-Subset, and ImageNet-Full. | ['Wonjun Hwang', 'Jaemin Na', 'Jisu Han'] | 2023-05-09 | null | null | null | null | ['class-incremental-learning', 'incremental-learning'] | ['computer-vision', 'methodology'] | [ 1.46864697e-01 1.45654175e-02 8.45205933e-02 -2.43850917e-01
1.62238851e-01 -2.63103098e-01 4.14316297e-01 2.46812984e-01
-8.19853544e-01 9.98048842e-01 -8.18077922e-02 2.26994187e-01
-4.43826377e-01 -8.93901527e-01 -6.41886115e-01 -7.52815127e-01
1.06330447e-01 3.00093740e-01 5.57998300e-01 5.19038662e-02
4.32460189e-01 4.89338756e-01 -1.72840703e+00 1.92123532e-01
1.43054879e+00 9.51736987e-01 5.50949275e-01 1.85009986e-01
-1.26574665e-01 9.12132204e-01 -2.77404517e-01 -3.68748337e-01
2.74924189e-01 -4.02487993e-01 -8.10502529e-01 -1.44572973e-01
2.81005472e-01 -6.27136007e-02 -2.85174727e-01 1.05260122e+00
4.31529790e-01 3.12204182e-01 5.03969610e-01 -1.08579183e+00
-9.89088476e-01 6.27227247e-01 -3.03404778e-01 6.43425107e-01
-1.60155118e-01 2.20333919e-01 4.74593729e-01 -1.05137813e+00
4.66546357e-01 9.05865371e-01 7.70513296e-01 6.87666893e-01
-1.17967463e+00 -7.34284341e-01 6.05190694e-01 3.39137554e-01
-1.28890228e+00 -4.79726493e-01 7.74565876e-01 -4.67678666e-01
6.77715003e-01 -2.73974121e-01 1.01913786e+00 7.74410129e-01
2.47235373e-01 8.19723070e-01 1.21608043e+00 -2.96256334e-01
4.37124550e-01 3.76861244e-01 3.13250303e-01 8.89998734e-01
3.86070132e-01 7.46127516e-02 -7.67885804e-01 1.10744573e-01
6.61500275e-01 1.70422181e-01 -3.92295510e-01 -4.85164195e-01
-1.08177686e+00 3.50777239e-01 6.89002514e-01 4.88189399e-01
-5.24148047e-01 -2.53120154e-01 4.73290272e-02 4.84389126e-01
1.40996054e-01 5.80797195e-01 -4.06423658e-01 9.88239720e-02
-1.08876050e+00 -2.39306819e-02 3.17307115e-01 6.18908465e-01
1.02513492e+00 3.07945549e-01 -4.15513426e-01 9.04688358e-01
6.48241863e-02 1.09464779e-01 1.09909332e+00 -7.36988664e-01
1.37526989e-01 9.43664253e-01 -1.81253627e-01 -8.64632010e-01
-4.44305837e-01 -1.05933952e+00 -7.83797324e-01 3.59919876e-01
1.42827675e-01 -9.20223966e-02 -9.87990141e-01 1.95843554e+00
8.80503803e-02 3.47557366e-01 1.14506379e-01 6.23738468e-01
5.27381718e-01 3.74031931e-01 2.61023283e-01 -3.93171042e-01
7.84257114e-01 -1.04400361e+00 -4.34703648e-01 -4.72003132e-01
3.63189101e-01 7.20439404e-02 1.14533150e+00 5.71640611e-01
-1.19445562e+00 -9.69270289e-01 -1.34553683e+00 2.21820325e-01
-2.96791524e-01 1.79055482e-01 5.53097129e-01 4.25623089e-01
-1.02012312e+00 9.70515668e-01 -7.74721324e-01 -2.26775751e-01
7.75776744e-01 2.98834205e-01 -1.38901547e-01 1.31052792e-01
-1.18069470e+00 8.71317685e-01 9.13838327e-01 -1.19699985e-02
-8.04386079e-01 -1.00673366e+00 -4.12021130e-01 3.35623980e-01
-4.43838425e-02 -9.02959824e-01 7.86207378e-01 -1.39674282e+00
-1.51533377e+00 6.90858364e-01 1.05061725e-01 -6.29883707e-01
4.22777295e-01 -3.30604702e-01 -3.32754880e-01 -1.09097786e-01
-1.72873318e-01 8.69940996e-01 8.94071281e-01 -1.14186835e+00
-7.66263187e-01 -5.19405723e-01 -1.46080360e-01 4.12438065e-01
-7.68989742e-01 -6.47815287e-01 -2.61462390e-01 -7.55867898e-01
4.94351923e-01 -7.18136787e-01 -1.45454137e-02 1.27153605e-01
1.27510920e-01 -5.51907159e-02 4.11040425e-01 -5.30845284e-01
1.30043364e+00 -2.16500044e+00 1.79332569e-01 1.31344825e-01
2.67070740e-01 4.96779501e-01 -1.18991248e-01 -1.23994507e-01
2.40311697e-02 -1.89943120e-01 -4.01747614e-01 -1.11716390e-01
-3.96549672e-01 1.77715719e-01 -1.90608889e-01 2.04011887e-01
2.39331961e-01 7.18077242e-01 -9.10035968e-01 -2.43527740e-01
-1.94491699e-01 3.41864526e-01 -6.82400465e-01 -7.06672594e-02
-4.83778976e-02 4.58042830e-01 -1.75227016e-01 4.13207024e-01
7.68249512e-01 -2.34721079e-01 7.16485456e-02 -1.81805387e-01
-2.01532066e-01 -1.31928638e-01 -1.10563791e+00 1.72414088e+00
-3.00304472e-01 2.04694554e-01 -7.66908526e-02 -1.08552635e+00
1.22123039e+00 -8.56557935e-02 1.77926138e-01 -8.67907465e-01
1.47704452e-01 1.91652045e-01 2.31466413e-01 -1.97241083e-01
2.64934361e-01 -1.97317064e-01 5.00283182e-01 2.68719971e-01
3.40925306e-01 2.45849818e-01 2.24520326e-01 6.60515130e-02
9.37511623e-01 8.18155333e-02 2.21285373e-02 -4.57596570e-01
6.83410645e-01 -2.94656962e-01 9.80151534e-01 8.68894935e-01
-5.06540239e-01 5.36818683e-01 1.56294823e-01 -4.95959401e-01
-5.94957054e-01 -1.40033138e+00 4.91671823e-02 1.10153449e+00
1.30083980e-02 1.67355806e-01 -5.34899652e-01 -7.75304794e-01
9.73486230e-02 8.63699794e-01 -6.03993952e-01 -8.41955781e-01
-6.08165741e-01 -9.95164454e-01 1.79769129e-01 5.62705278e-01
1.06596386e+00 -1.21268702e+00 -5.44721067e-01 3.48314404e-01
7.20039457e-02 -5.55936038e-01 -1.66953251e-01 2.28669897e-01
-1.23337042e+00 -9.38027322e-01 -7.21493423e-01 -1.10143363e+00
9.56756055e-01 9.72922668e-02 8.74039054e-01 2.48536721e-01
-1.93299904e-01 2.70343155e-01 -1.21809751e-01 -1.43861815e-01
-9.41548944e-02 2.70271957e-01 4.03920531e-01 1.29747659e-01
1.20182924e-01 -1.03325832e+00 -7.80412197e-01 1.04376785e-01
-8.22228014e-01 1.56611100e-01 8.03902566e-01 8.73115718e-01
6.26584649e-01 3.48861158e-01 1.14041793e+00 -7.28892505e-01
6.43806815e-01 -4.89932895e-01 -3.31417829e-01 4.88204181e-01
-1.13988960e+00 3.07828039e-01 7.10134566e-01 -6.98034704e-01
-1.46878743e+00 3.91449146e-02 1.10978134e-01 -3.80944908e-01
2.80380338e-01 6.31826937e-01 -1.17565252e-01 -1.64559498e-01
6.53560340e-01 5.57296574e-01 -1.06564298e-01 -5.49000561e-01
3.40053827e-01 1.25758290e-01 7.74409235e-01 -6.85717344e-01
5.61424077e-01 4.67745751e-01 -5.49578905e-01 -3.78136158e-01
-8.81900370e-01 1.52243152e-01 -9.44534898e-01 -2.54130661e-01
4.34363902e-01 -8.02989066e-01 -5.42253017e-01 7.66885340e-01
-7.83628583e-01 -2.87399769e-01 -6.00867093e-01 5.89498460e-01
-3.41987193e-01 2.17236072e-01 -4.61103648e-01 -4.45874810e-01
-4.89317358e-01 -5.98703384e-01 1.17690310e-01 7.41075337e-01
9.70682502e-02 -8.71825218e-01 2.65717536e-01 1.93878710e-02
5.61592281e-01 -1.27991170e-01 9.65036988e-01 -5.24215043e-01
-3.55546355e-01 1.80984437e-01 -2.95687318e-01 3.76742870e-01
6.11418448e-02 -3.50820690e-01 -8.66085589e-01 -6.48538470e-01
3.39904800e-02 -4.42321599e-01 1.62307012e+00 1.83631722e-02
1.14218593e+00 -3.47956985e-01 -2.83037722e-01 5.65902531e-01
1.25957465e+00 3.02211493e-01 6.94813550e-01 4.97964859e-01
3.91223878e-01 3.89978498e-01 1.27651200e-01 3.97539437e-01
3.43930215e-01 1.45832196e-01 2.87542008e-02 3.21201116e-01
-3.47878337e-01 -4.54879582e-01 2.39685848e-01 9.27062809e-01
3.62756103e-03 1.95379242e-01 -8.15855742e-01 7.57288873e-01
-1.76257312e+00 -9.65907931e-01 7.08852530e-01 2.42629075e+00
1.26928937e+00 6.57413900e-01 -2.29868934e-01 1.00007027e-01
5.88633418e-01 -1.34656549e-01 -9.79961514e-01 -3.59274708e-02
-2.79312670e-01 2.09293053e-01 1.01227313e-01 3.23105991e-01
-7.65283167e-01 9.95399237e-01 5.91010094e+00 7.33347297e-01
-1.20428824e+00 4.87265922e-02 5.60253263e-01 -1.56957909e-01
-2.55368173e-01 -2.39321794e-02 -7.92220235e-01 5.06287634e-01
6.83983743e-01 -2.54450262e-01 4.50656563e-01 7.25720048e-01
-1.68736756e-01 -1.73810855e-01 -9.05808032e-01 8.29568565e-01
-4.61935885e-02 -1.24484122e+00 2.56515533e-01 -5.75039566e-01
8.72773290e-01 -1.76437601e-01 2.82112598e-01 7.40318120e-01
1.66605338e-01 -6.65407181e-01 7.14675665e-01 1.16742969e+00
2.35158980e-01 -5.57343900e-01 1.98909849e-01 5.17408013e-01
-9.39806879e-01 -5.22329092e-01 -4.37464207e-01 -2.14585274e-01
-2.61957705e-01 7.24199593e-01 -5.95699310e-01 1.65628880e-01
7.33991325e-01 6.73386276e-01 -1.09818769e+00 1.36397958e+00
-1.96904793e-01 5.03266633e-01 -7.45455772e-02 1.91644013e-01
-1.90568149e-01 -3.79727073e-02 3.32007349e-01 8.32870126e-01
4.01135594e-01 3.93897854e-02 1.51692569e-01 1.00520825e+00
-2.20857248e-01 5.82543612e-02 -1.33823633e-01 7.73843154e-02
7.28723109e-01 8.88586462e-01 -8.09359550e-01 -3.64888161e-01
-8.06850120e-02 1.11002386e+00 8.84412110e-01 4.03633714e-01
-4.94243741e-01 -3.94044131e-01 8.25994164e-02 7.37858564e-02
3.97319108e-01 -1.75848335e-01 -3.27276081e-01 -1.20885980e+00
4.83812243e-02 -4.09050107e-01 4.86451924e-01 -5.57811797e-01
-1.24112749e+00 6.72155678e-01 -1.84750974e-01 -9.55332935e-01
1.76399752e-01 -3.33166420e-01 -7.53790915e-01 5.52048624e-01
-1.55672109e+00 -8.00538480e-01 -3.82766336e-01 5.27051508e-01
4.63692486e-01 -4.37307209e-01 3.97624195e-01 3.56240183e-01
-6.66075766e-01 7.40120411e-01 1.42907545e-01 -7.24747628e-02
4.76952344e-01 -9.45571244e-01 3.09978910e-02 7.40443468e-01
-1.15510188e-01 7.16221273e-01 4.01304036e-01 -8.71221542e-01
-8.99270475e-01 -1.11090600e+00 6.78018630e-01 -4.58873771e-02
3.39962214e-01 -1.98314503e-01 -1.35644197e+00 4.59410638e-01
-1.23842411e-01 -1.01571798e-01 4.68721688e-01 3.47619355e-02
-4.53405708e-01 -6.43168390e-01 -1.34943330e+00 4.90237087e-01
1.15519238e+00 -8.82709473e-02 -8.90536249e-01 -1.35437235e-01
6.65927827e-01 1.48452386e-01 -5.50578713e-01 6.99607670e-01
6.85685635e-01 -1.01284742e+00 7.39212513e-01 -3.82705003e-01
-1.20753065e-01 -4.37389135e-01 1.49643347e-01 -1.47773671e+00
-9.01722252e-01 -2.90515631e-01 -2.86729664e-01 1.36455595e+00
4.92468208e-01 -6.83782995e-01 8.57628882e-01 3.71319562e-01
-1.99584961e-01 -6.88929439e-01 -9.09147799e-01 -9.11882043e-01
2.68333197e-01 1.74720854e-01 4.30427313e-01 9.58045840e-01
-5.19610122e-02 2.22948834e-01 -1.47179782e-01 -9.11592692e-02
6.62589371e-01 -5.61264828e-02 7.78407604e-02 -1.55178368e+00
-2.29091808e-01 -6.01931572e-01 -2.95936108e-01 -7.55998194e-01
1.48613647e-01 -9.92760777e-01 -8.12488720e-02 -1.45184612e+00
3.59413117e-01 -5.12829125e-01 -9.87927496e-01 7.06894457e-01
-3.31197023e-01 6.67974055e-02 2.14300185e-01 3.54868531e-01
-7.29105234e-01 9.69493628e-01 1.11246264e+00 -1.32189184e-01
-6.94169283e-01 -4.00099829e-02 -9.19023633e-01 7.52826571e-01
9.95466173e-01 -2.83853590e-01 -7.82845318e-01 -4.32770252e-01
3.24568957e-01 -4.50960279e-01 3.42376709e-01 -1.73632002e+00
5.30870080e-01 -1.23864263e-02 8.43268335e-01 -3.18395495e-01
2.74319723e-02 -5.81803858e-01 -4.35326481e-04 8.01194310e-01
-3.94633919e-01 -9.92712975e-02 3.94718379e-01 7.28790283e-01
8.91850889e-02 -3.32078755e-01 1.17739952e+00 -1.95318177e-01
-9.52487350e-01 3.65942001e-01 -3.03498298e-01 1.53203398e-01
8.42151821e-01 -2.97617823e-01 -3.43314260e-01 1.03795074e-01
-1.00015986e+00 3.92455906e-01 3.21696877e-01 6.36724412e-01
1.03015292e+00 -1.39410865e+00 -7.06329525e-01 5.16582489e-01
-1.96091216e-02 -5.69049180e-01 4.64746922e-01 7.85442054e-01
-2.58177966e-01 1.06422398e-02 -6.52373195e-01 -2.06434116e-01
-7.38838851e-01 8.16248536e-01 5.23777187e-01 -1.67398587e-01
-5.31592429e-01 9.67982650e-01 2.54715979e-01 -3.20377737e-01
4.31971580e-01 -2.40158349e-01 -5.18573403e-01 5.61726168e-02
7.18600869e-01 3.42241526e-01 2.25255623e-01 -1.22505359e-01
-2.85769641e-01 3.25511485e-01 -5.97547174e-01 3.53812836e-02
1.30808747e+00 -3.42480153e-01 -4.74275500e-02 5.29025495e-01
7.72576571e-01 -3.86134863e-01 -1.59279716e+00 -4.59362149e-01
4.39726487e-02 -2.00845316e-01 1.42109217e-02 -1.07372820e+00
-1.19693458e+00 7.32637405e-01 1.02526748e+00 -3.57697576e-01
1.29445791e+00 -2.23800674e-01 6.59556210e-01 5.77333331e-01
3.49216938e-01 -1.30688536e+00 4.67867851e-01 6.66816473e-01
7.68423140e-01 -8.79066646e-01 8.69038701e-02 2.37483814e-01
-5.43894708e-01 8.34238827e-01 1.14915442e+00 -2.21811846e-01
9.27538276e-01 -2.34556943e-01 -3.74591559e-01 -1.04268678e-01
-7.94602871e-01 -1.36334881e-01 1.29774630e-01 5.89506030e-01
1.43747553e-01 -3.50183755e-01 -3.78764421e-01 1.06012666e+00
-1.22188061e-01 3.82999361e-01 3.19968581e-01 1.07253301e+00
-1.07724524e+00 -7.12897241e-01 -1.08646452e-01 5.06817877e-01
-1.20796271e-01 -1.77179947e-01 -1.34079814e-01 3.98134291e-01
5.59329331e-01 5.68873167e-01 2.47149635e-02 -3.25821102e-01
2.72188604e-01 3.75707507e-01 7.18085825e-01 -4.35605198e-01
-4.36121345e-01 -4.57584918e-01 -4.76699561e-01 -8.41021240e-02
-3.34422857e-01 -4.77582127e-01 -1.38913059e+00 8.52741003e-02
-2.59420484e-01 3.35641801e-01 2.06987247e-01 8.51139843e-01
5.44344962e-01 5.92423320e-01 5.43084979e-01 -3.95061016e-01
-5.03741086e-01 -7.31204689e-01 -6.40537977e-01 2.34872624e-01
2.16535479e-01 -1.04927266e+00 -3.96017522e-01 1.74956113e-01] | [9.81713581085205, 3.3640546798706055] |
a69b05f0-dd48-49b2-9c9a-c707ef8e4fb6 | conditional-random-fields-as-recurrent-neural | 1807.07464 | null | http://arxiv.org/abs/1807.07464v1 | http://arxiv.org/pdf/1807.07464v1.pdf | Conditional Random Fields as Recurrent Neural Networks for 3D Medical Imaging Segmentation | The Conditional Random Field as a Recurrent Neural Network layer is a
recently proposed algorithm meant to be placed on top of an existing
Fully-Convolutional Neural Network to improve the quality of semantic
segmentation. In this paper, we test whether this algorithm, which was shown to
improve semantic segmentation for 2D RGB images, is able to improve
segmentation quality for 3D multi-modal medical images. We developed an
implementation of the algorithm which works for any number of spatial
dimensions, input/output image channels, and reference image channels. As far
as we know this is the first publicly available implementation of this sort. We
tested the algorithm with two distinct 3D medical imaging datasets, we
concluded that the performance differences observed were not statistically
significant. Finally, in the discussion section of the paper, we go into the
reasons as to why this technique transfers poorly from natural images to
medical images. | ['Mário A. T. Figueiredo', 'Arlindo L. Oliveira', 'Miguel Monteiro'] | 2018-07-19 | null | null | null | null | ['3d-medical-imaging-segmentation', 'volumetric-medical-image-segmentation'] | ['medical', 'medical'] | [ 6.41936362e-01 4.98283744e-01 1.65452972e-01 -4.38981801e-01
-7.17620075e-01 -1.93260193e-01 3.07194114e-01 -5.23150526e-02
-7.33874261e-01 4.69904333e-01 2.63044909e-02 -5.97231388e-01
-2.77973711e-01 -7.41042197e-01 -5.75368941e-01 -6.29567742e-01
-3.11461482e-02 4.10386086e-01 5.43124676e-01 4.47799489e-02
2.47362122e-01 7.39731252e-01 -1.56204224e+00 4.95612532e-01
3.05173635e-01 9.28277791e-01 2.69042432e-01 7.74632752e-01
-1.87537029e-01 6.99234903e-01 -5.57818949e-01 1.33450925e-01
9.73969400e-02 -4.06061977e-01 -1.25914085e+00 1.68131337e-01
2.44086742e-01 -2.76278824e-01 9.82288495e-02 6.69420600e-01
6.91743493e-01 -1.25279859e-01 6.24105513e-01 -4.66269732e-01
-4.35019672e-01 6.51905954e-01 -1.40535101e-01 2.80457765e-01
2.09531456e-01 -6.57010898e-02 4.49519366e-01 -5.03044546e-01
8.31616819e-01 1.00246501e+00 7.36059368e-01 5.83777010e-01
-1.00146627e+00 -3.01063776e-01 -2.31214747e-01 -1.05088659e-01
-1.03367841e+00 -9.19179991e-02 7.29295373e-01 -4.42987978e-01
1.25357318e+00 3.13427486e-02 5.73532164e-01 8.09689224e-01
2.31367812e-01 8.43284786e-01 1.53988051e+00 -6.45783186e-01
2.02781290e-01 1.75535098e-01 1.37058541e-01 6.51744008e-01
1.43425494e-01 2.90092438e-01 -1.66492999e-01 2.56725699e-01
8.68245244e-01 -4.07183856e-01 -5.66017777e-02 -1.24115780e-01
-1.07439947e+00 8.69250059e-01 6.42399251e-01 1.00841844e+00
-4.28573072e-01 1.31080911e-01 3.37752640e-01 -3.55455503e-02
5.99873364e-01 5.12735069e-01 -4.34967309e-01 -4.76898532e-03
-1.16446257e+00 -1.73952341e-01 4.29885000e-01 2.05554530e-01
6.07491612e-01 -1.35488316e-01 3.15903835e-02 5.88638544e-01
3.26788723e-01 2.88903892e-01 6.30167484e-01 -1.08818400e+00
-5.18332198e-02 5.76799393e-01 -3.00868452e-01 -8.00941110e-01
-9.99160230e-01 -3.57194334e-01 -6.66364074e-01 5.23065925e-01
5.86845577e-01 -1.46145314e-01 -1.28088272e+00 1.24657524e+00
1.10427201e-01 -5.94389848e-02 8.15593898e-02 9.56127644e-01
1.16858244e+00 1.52827218e-01 7.39216655e-02 7.27844536e-02
9.26677287e-01 -4.72146869e-01 -6.28763855e-01 5.08761555e-02
6.72641516e-01 -9.68386054e-01 7.99747705e-01 4.17143494e-01
-1.12875545e+00 -5.85765719e-01 -1.06334043e+00 -1.15304021e-02
-7.27526367e-01 3.08106076e-02 9.32450056e-01 9.85204399e-01
-1.36858022e+00 9.54796076e-01 -1.09205925e+00 -5.02318919e-01
3.93107891e-01 4.44877088e-01 -2.60973960e-01 1.63523495e-01
-1.24512494e+00 1.10477233e+00 5.00392497e-01 3.50345850e-01
-3.69409263e-01 -2.36552879e-01 -6.94547474e-01 -4.89144683e-01
2.01195311e-02 -5.88976085e-01 1.03562486e+00 -1.03112566e+00
-1.67993832e+00 1.43199098e+00 2.20430434e-01 -4.26607072e-01
4.41830248e-01 4.93107401e-02 -2.28539169e-01 3.48040581e-01
2.20977347e-02 9.24625456e-01 4.73771811e-01 -1.23514247e+00
-3.91540885e-01 -6.39654279e-01 -1.39868379e-01 -5.40024647e-03
3.80552083e-01 -2.80272383e-02 -3.59666377e-01 -5.48560739e-01
4.79126930e-01 -9.73928690e-01 -5.02525151e-01 -2.71804690e-01
-3.19639713e-01 -1.25377163e-01 5.24153352e-01 -5.73445380e-01
5.42781174e-01 -2.25030160e+00 7.49951154e-02 4.17711139e-01
-1.30603701e-01 1.58380657e-01 1.66637346e-01 -1.02182016e-01
-3.82904232e-01 2.19449356e-01 -5.67057014e-01 -1.23682559e-01
-1.60729140e-01 3.87449682e-01 2.28740171e-01 5.97858310e-01
1.44779652e-01 1.10235345e+00 -4.66399401e-01 -4.52419728e-01
5.93508065e-01 8.06140780e-01 -2.15900972e-01 -1.76681772e-01
1.17282579e-02 5.53805709e-01 -3.01510125e-01 5.00794232e-01
6.68206990e-01 -3.24649155e-01 -2.66408682e-01 -7.20734298e-02
-3.57522555e-02 1.50698185e-01 -1.04352725e+00 2.06784797e+00
-2.37117991e-01 5.75385332e-01 -3.30343217e-01 -1.07581937e+00
7.56631434e-01 3.47446442e-01 7.45487452e-01 -9.56499755e-01
4.87639338e-01 4.83124614e-01 -1.20498046e-01 -5.11141956e-01
3.71571124e-01 -3.84257704e-01 1.48934245e-01 3.82870346e-01
9.21413377e-02 -2.97203690e-01 -7.20384866e-02 -2.42602557e-01
8.07185888e-01 3.52964133e-01 3.16400453e-02 -1.88887104e-01
3.92353684e-01 1.44112617e-01 -8.39521587e-02 7.57406414e-01
-1.87348425e-01 1.04900646e+00 3.62042934e-01 -4.30524588e-01
-7.60202169e-01 -9.11463916e-01 -4.53289479e-01 5.90107799e-01
1.21690799e-02 2.03571305e-01 -9.94971693e-01 -4.31802332e-01
-4.36085880e-01 5.72056592e-01 -8.76128972e-01 1.22833319e-01
-4.79576021e-01 -9.56567526e-01 6.94868445e-01 3.65813673e-01
5.46491683e-01 -1.47017157e+00 -1.23019528e+00 2.74354488e-01
9.91521496e-03 -1.09681487e+00 3.89427960e-01 6.96171999e-01
-1.23571301e+00 -1.24214256e+00 -1.06201136e+00 -6.87698424e-01
4.82081711e-01 -1.89846918e-01 1.11042345e+00 6.98274672e-02
-4.45501924e-01 4.49455708e-01 -5.26920319e-01 -3.68341327e-01
-4.86214429e-01 1.84591576e-01 -6.39453650e-01 -5.34867644e-01
2.71513373e-01 -2.79524714e-01 -6.85039461e-01 8.02272931e-02
-1.26029921e+00 1.63811207e-01 7.24364519e-01 7.18234956e-01
6.22914374e-01 9.91885290e-02 -3.77362370e-02 -1.16365087e+00
3.43397528e-01 5.28642051e-02 -1.76803008e-01 -7.00060278e-02
-5.56372643e-01 1.43387645e-01 -3.63519124e-04 1.61305875e-01
-8.05761814e-01 4.08375263e-01 -7.75209725e-01 -5.62482774e-02
-8.01734686e-01 5.81023216e-01 3.98146540e-01 -1.49702877e-01
7.83106327e-01 2.64153704e-02 2.04131335e-01 -4.10919547e-01
3.12329948e-01 5.84013045e-01 5.10247350e-01 -3.47376205e-02
2.30491534e-01 7.29543328e-01 1.53217077e-01 -7.64932930e-01
-5.30442297e-01 -5.78590989e-01 -9.58166361e-01 -1.75552338e-01
1.41117215e+00 -5.55705488e-01 -3.45421940e-01 6.93728507e-01
-1.09491527e+00 -7.09966719e-01 -3.29645365e-01 5.42833269e-01
-7.26745546e-01 1.10430628e-01 -7.10771382e-01 -4.73730624e-01
-1.28995806e-01 -1.28038514e+00 1.15557563e+00 1.52296707e-01
-1.00099839e-01 -1.14060318e+00 -5.91138192e-02 3.72557014e-01
3.78832579e-01 5.95843852e-01 8.36606979e-01 -3.82758528e-01
-2.50069201e-01 -1.00260310e-01 -2.56092161e-01 4.08341169e-01
1.10224135e-01 -1.04378462e-01 -1.15846598e+00 -5.38735976e-03
3.01683098e-01 -1.22889154e-01 1.08355045e+00 9.44860637e-01
1.29473162e+00 3.92325044e-01 -1.86908036e-01 4.88310307e-01
1.45327449e+00 1.84254661e-01 1.02062845e+00 6.27477109e-01
4.40386504e-01 5.98345518e-01 2.53109127e-01 -8.37735385e-02
4.70780544e-02 4.57102627e-01 4.22845304e-01 -7.71552861e-01
-5.15003800e-01 1.91237390e-01 -1.88799240e-02 5.47185242e-01
-1.20579146e-01 4.30559739e-02 -9.60104227e-01 4.62741882e-01
-1.37608743e+00 -5.71636617e-01 -3.41194093e-01 1.79582477e+00
4.15650129e-01 3.76268774e-01 7.02720955e-02 4.72224921e-01
4.58684593e-01 -1.42217815e-01 -4.30482894e-01 -6.00434363e-01
-1.91460267e-01 6.27818584e-01 8.49317431e-01 4.11211729e-01
-1.09959280e+00 7.66807616e-01 7.32104206e+00 4.63505358e-01
-1.54292452e+00 1.97861105e-01 8.71169686e-01 3.17288667e-01
-1.71385437e-01 -3.01785707e-01 -2.38040879e-01 2.13289902e-01
1.12319362e+00 8.47025037e-01 7.47089088e-02 6.27956986e-01
1.98418766e-01 -6.29539073e-01 -7.24431336e-01 8.74978065e-01
1.78127438e-01 -1.20618427e+00 -2.27042213e-01 -5.76925762e-02
5.09868920e-01 3.04524839e-01 1.72071040e-01 -1.50714144e-01
5.37041724e-02 -1.40999722e+00 4.64881688e-01 4.73725975e-01
6.18067861e-01 -7.96900332e-01 1.12195432e+00 -2.92549785e-02
-5.37912071e-01 3.83246273e-01 -1.27037674e-01 7.12312534e-02
2.38896683e-01 6.06384575e-01 -8.60089004e-01 5.16238391e-01
9.94297087e-01 5.45321643e-01 -9.16217387e-01 1.03482783e+00
-2.13480979e-01 4.14760768e-01 -2.92116612e-01 1.68336213e-01
5.92281520e-01 2.76578367e-01 2.59424984e-01 1.17151403e+00
1.48710340e-01 -4.10601310e-02 -2.68193334e-01 7.24335253e-01
4.91885096e-01 5.64230233e-02 -6.10867560e-01 8.06575119e-02
-5.45088768e-01 8.48620355e-01 -1.59800375e+00 -1.40003607e-01
-2.67818451e-01 1.16177619e+00 -2.65695274e-01 3.95348728e-01
-6.71141148e-01 -2.69488215e-01 -8.28636736e-02 5.85239902e-02
4.15341526e-01 -4.27121967e-02 -8.01474214e-01 -6.01576209e-01
-2.42234588e-01 -4.88813370e-01 4.76737857e-01 -1.05369866e+00
-1.02274776e+00 7.31808066e-01 3.28602716e-02 -6.87270761e-01
-2.36063555e-01 -1.02297688e+00 2.31076451e-03 8.91372681e-01
-1.48586488e+00 -9.77828085e-01 -6.95454180e-02 6.25657678e-01
1.75642192e-01 2.38358676e-02 1.03150463e+00 1.10447347e-01
-2.09194750e-01 1.39887929e-01 -1.09395210e-03 1.86565921e-01
4.16279018e-01 -1.33834159e+00 2.09755465e-01 6.52386129e-01
1.21780649e-01 3.53274018e-01 6.63200438e-01 -5.22774816e-01
-7.26323009e-01 -6.82542145e-01 7.50628233e-01 -2.48365164e-01
1.60247162e-01 1.01988725e-01 -7.80983686e-01 4.65574205e-01
3.52785796e-01 -1.06922992e-01 7.84989178e-01 -9.45610628e-02
1.55934423e-01 4.25095171e-01 -1.40502238e+00 7.37836584e-02
5.66260397e-01 -4.55976814e-01 -7.33768046e-01 2.94344962e-01
6.16216660e-01 -8.73296082e-01 -9.95728850e-01 6.33692443e-01
2.99148589e-01 -1.46426129e+00 1.07952893e+00 -2.81086564e-02
5.30288696e-01 -2.92001128e-01 -1.61028560e-02 -1.07415128e+00
4.83695157e-02 -9.22980681e-02 5.01100540e-01 6.00005746e-01
6.32619560e-01 -6.19694471e-01 1.03266263e+00 4.13266748e-01
-2.47011334e-01 -7.25522280e-01 -1.18777978e+00 -2.55839646e-01
4.31156904e-01 -9.35507238e-01 3.34226906e-01 5.85117996e-01
-4.51497227e-01 1.57836519e-04 1.14494830e-01 -8.13488811e-02
3.98821265e-01 2.79432926e-02 2.38189146e-01 -9.74292457e-01
-1.75087914e-01 -7.06657350e-01 -5.86598814e-01 -6.02372944e-01
-9.36164241e-03 -1.04517996e+00 2.65654083e-02 -1.89481044e+00
-1.81319192e-02 -4.90217716e-01 -4.59377468e-01 4.35148567e-01
1.07337840e-01 6.31006479e-01 1.58474725e-02 3.96268256e-02
-2.83243895e-01 -1.34708390e-01 1.49283397e+00 -9.56426784e-02
-2.38586321e-01 1.10123031e-01 -6.11935914e-01 5.93570232e-01
8.69646788e-01 -5.10657012e-01 -2.81795025e-01 -3.58720630e-01
1.64238900e-01 -5.75284064e-02 4.89625782e-01 -1.05582166e+00
1.39018027e-02 3.36757690e-01 7.71090090e-01 -8.43087375e-01
1.09839715e-01 -9.58756089e-01 1.98306382e-01 6.98032439e-01
-4.59358603e-01 -1.97381452e-01 5.28091431e-01 8.24205130e-02
-2.19674870e-01 -3.05574268e-01 9.82225060e-01 -6.51389420e-01
-9.33328390e-01 -7.81167671e-02 -6.16227150e-01 -2.41030067e-01
8.40708435e-01 -5.17406344e-01 1.62072286e-01 -1.22982450e-01
-1.08215022e+00 -2.77355433e-01 3.89554948e-01 2.04693854e-01
6.69799924e-01 -1.02195990e+00 -4.16154921e-01 7.79498443e-02
-1.65046543e-01 1.17330000e-01 5.14277995e-01 8.85036170e-01
-8.96199882e-01 8.60378742e-01 -4.38902855e-01 -9.67239976e-01
-1.08544874e+00 4.20334339e-01 6.71527743e-01 -2.05294803e-01
-7.08550394e-01 7.93148696e-01 -3.67026240e-01 -7.07437456e-01
2.43192300e-01 -6.96607947e-01 -4.62750494e-01 -2.73062419e-02
2.62078106e-01 1.43911451e-01 5.47557294e-01 -8.47438693e-01
-4.01959181e-01 7.92953670e-01 2.80295551e-01 -2.67754614e-01
1.48728526e+00 -2.11833715e-01 -1.92838162e-01 7.20512986e-01
1.39244807e+00 -5.40625751e-01 -8.95812035e-01 1.33062094e-01
1.04257867e-01 -1.30007535e-01 4.07558411e-01 -1.02985156e+00
-1.32090878e+00 9.81887579e-01 1.36559629e+00 2.07156822e-01
1.34849203e+00 2.33762056e-01 5.24040639e-01 2.80763935e-02
2.22884327e-01 -1.09434724e+00 -8.84631053e-02 5.47201514e-01
4.87871975e-01 -1.34744966e+00 7.00217532e-03 -2.50157982e-01
-5.05091369e-01 1.30688560e+00 1.94604367e-01 -1.92744061e-01
6.57423794e-01 2.42741704e-01 3.45221072e-01 -5.52361250e-01
1.16739504e-03 -5.16440392e-01 3.65037680e-01 7.42557585e-01
7.85474956e-01 6.78194016e-02 -3.58172417e-01 4.52795532e-03
-2.73546636e-01 3.78557056e-01 5.10009348e-01 8.25244784e-01
-2.23070011e-01 -1.10231268e+00 -4.90264446e-01 3.36396962e-01
-7.02097058e-01 9.56111103e-02 -3.99855077e-01 1.00698507e+00
3.88309509e-01 7.73522675e-01 7.44506121e-02 -2.58289695e-01
3.41245204e-01 1.10897087e-01 6.03308320e-01 -5.20031095e-01
-7.50723481e-01 1.32464796e-01 -1.49200171e-01 -6.83942437e-01
-9.93072867e-01 -5.91828287e-01 -1.60802937e+00 2.01173678e-01
-1.62491009e-01 -1.27483934e-01 1.09414113e+00 1.08670926e+00
-1.03664376e-01 9.56038177e-01 1.64166480e-01 -9.32469666e-01
1.66937441e-01 -8.88162374e-01 -6.39230728e-01 3.77442449e-01
3.03710580e-01 -6.24397218e-01 -2.12093085e-01 -5.17680719e-02] | [14.433799743652344, -2.499025821685791] |
a76de40b-d9e9-465b-bb4c-1320740d3d51 | convolutive-audio-source-separation-using | 1708.03989 | null | http://arxiv.org/abs/1708.03989v1 | http://arxiv.org/pdf/1708.03989v1.pdf | Convolutive Audio Source Separation using Robust ICA and an intelligent evolving permutation ambiguity solution | Audio source separation is the task of isolating sound sources that are
active simultaneously in a room captured by a set of microphones. Convolutive
audio source separation of equal number of sources and microphones has a number
of shortcomings including the complexity of frequency-domain ICA, the
permutation ambiguity and the problem's scalabity with increasing number of
sensors. In this paper, the authors propose a multiple-microphone audio source
separation algorithm based on a previous work of Mitianoudis and Davies (2003).
Complex FastICA is substituted by Robust ICA increasing robustness and
performance. Permutation ambiguity is solved using two methodologies. The first
is using the Likelihood Ration Jump solution, which is now modified to decrease
computational complexity in the case of multiple microphones. The application
of the MuSIC algorithm, as a preprocessing step to the previous solution, forms
a second methodology with promising results. | ['Thomas Sgouros', 'Nikolaos Mitianoudis', 'Dimitrios Mallis'] | 2017-08-14 | null | null | null | null | ['audio-source-separation'] | ['audio'] | [ 4.35059905e-01 -4.46923137e-01 6.23381793e-01 2.08320573e-01
-9.48199391e-01 -8.04077089e-01 3.16214770e-01 -6.04041517e-02
-4.28972542e-01 1.02632260e+00 3.05881858e-01 -7.35388324e-02
-7.22985566e-01 -5.67812100e-02 -3.17749113e-01 -8.87434840e-01
1.91021673e-02 1.69777721e-01 1.63140953e-01 1.41985357e-01
3.04706514e-01 5.19844115e-01 -1.65353537e+00 -1.01990148e-01
8.85907114e-01 7.70507634e-01 2.81072736e-01 1.10983539e+00
-9.07226354e-02 4.93068308e-01 -9.48570251e-01 1.12113193e-01
3.74601603e-01 -6.62380874e-01 -2.03872532e-01 1.34001859e-02
1.35476962e-01 1.80307347e-02 4.02784586e-01 1.07338023e+00
7.98988581e-01 4.22039241e-01 6.04165077e-01 -1.17453825e+00
1.37858048e-01 4.43978131e-01 -6.30721629e-01 4.80117321e-01
5.09638369e-01 -4.57017601e-01 4.52975035e-01 -9.69818532e-01
-1.51766628e-01 1.19111586e+00 6.92163467e-01 -1.03902772e-01
-1.16379642e+00 -7.15548813e-01 -2.70101011e-01 1.37854278e-01
-1.52443647e+00 -6.46166682e-01 1.12510514e+00 -4.17758465e-01
5.84080338e-01 7.21251845e-01 4.01551813e-01 6.76527202e-01
-2.55371958e-01 2.04400212e-01 1.20652425e+00 -7.63944983e-01
4.50616509e-01 1.75206900e-01 5.06383032e-02 1.36573343e-02
3.97464722e-01 -2.86658406e-01 -4.60404187e-01 -4.79222268e-01
6.78480327e-01 -3.22939128e-01 -2.99231946e-01 -1.53187275e-01
-1.07425344e+00 7.01044321e-01 -2.88206428e-01 7.48205245e-01
-4.70175833e-01 -1.57042339e-01 1.91736281e-01 2.34554559e-01
3.20120573e-01 6.89062774e-01 -3.63143951e-01 -2.14695334e-01
-1.38467968e+00 1.60898015e-01 1.03478372e+00 4.64859545e-01
1.17211021e-01 4.89726454e-01 2.66525090e-01 8.19371819e-01
6.22285008e-01 6.38316631e-01 6.95858598e-01 -8.49183619e-01
2.59115666e-01 1.33604482e-01 2.29965061e-01 -9.27341878e-01
-3.37211162e-01 -7.12791622e-01 -6.32202208e-01 4.24808592e-01
8.12808871e-01 -5.80856085e-01 -7.48645127e-01 1.42245710e+00
3.67784679e-01 5.99089324e-01 2.21335620e-01 8.11319649e-01
5.92592001e-01 7.54262626e-01 -3.24694604e-01 -7.52186239e-01
1.23255146e+00 -9.04246688e-01 -1.05677485e+00 -2.60359887e-02
-4.63724732e-01 -1.44226956e+00 4.10176158e-01 9.47166860e-01
-1.19799769e+00 -7.15883255e-01 -1.02281964e+00 3.17039192e-01
-3.76744628e-01 6.32424951e-02 1.55526921e-01 1.02079487e+00
-9.18893397e-01 -5.98920286e-02 -5.06674409e-01 -1.50611967e-01
-2.91733652e-01 4.47614789e-01 -4.18472253e-02 4.56207633e-01
-7.72587597e-01 5.78747272e-01 6.58424497e-02 2.67446727e-01
-6.41048372e-01 -5.68600297e-01 -4.60109740e-01 1.63789377e-01
3.55080843e-01 -4.54379469e-01 1.01422012e+00 -1.20265794e+00
-1.96379972e+00 8.67847204e-02 -5.24712265e-01 -6.13989979e-02
5.41361988e-01 -4.36566174e-01 -6.74460173e-01 2.02749267e-01
9.57085490e-02 -4.52595837e-02 1.39204621e+00 -1.34202123e+00
-5.99106014e-01 -3.51699561e-01 -5.13502538e-01 3.60030740e-01
1.93196014e-02 2.13160709e-01 -1.67462170e-01 -7.64907300e-01
5.48892498e-01 -7.88460135e-01 -1.96892709e-01 -8.17628980e-01
-2.74339169e-01 2.08770290e-01 5.42543590e-01 -9.11072314e-01
1.16741300e+00 -2.39375782e+00 3.03291112e-01 5.38003683e-01
-1.75803021e-01 3.46735626e-01 8.79567489e-02 4.45954829e-01
-5.63913345e-01 -2.80294806e-01 -4.21252072e-01 -4.58763391e-01
-3.10315102e-01 -1.18008539e-01 -2.86977470e-01 4.43906844e-01
-6.22596852e-02 6.81155920e-02 -5.69250882e-01 -4.19007450e-01
2.71922559e-01 6.39188886e-01 -3.07219714e-01 1.32655308e-01
5.28954864e-01 8.47807407e-01 1.09466523e-01 5.06195247e-01
8.55157793e-01 4.46895868e-01 -6.88767731e-02 -4.96537164e-02
-5.96020758e-01 1.32608190e-01 -2.29339910e+00 1.42342257e+00
-2.75688291e-01 4.64800030e-01 7.27147937e-01 -9.48487580e-01
1.04730856e+00 9.42169905e-01 5.47303319e-01 -3.44869122e-02
2.71097481e-01 7.01629877e-01 1.71517521e-01 -6.22280717e-01
3.38295132e-01 -2.09292352e-01 1.90380424e-01 1.57868326e-01
1.12403594e-01 -2.71140426e-01 4.07183588e-01 -2.99218506e-01
6.13566816e-01 -6.03611544e-02 5.94069183e-01 -4.29333389e-01
9.65487063e-01 -4.67298746e-01 4.64926451e-01 7.64025986e-01
-1.83130577e-01 6.72429562e-01 1.34374753e-01 2.52236694e-01
-4.90911186e-01 -1.08744323e+00 -1.38592795e-01 9.05003607e-01
-3.38700444e-01 5.55112213e-02 -6.04503274e-01 5.99817783e-02
-1.44356236e-01 4.56882864e-01 -1.90811664e-01 2.97066271e-01
-7.22715020e-01 -7.52998292e-01 5.47629416e-01 1.48979053e-01
3.64270955e-01 -6.64875388e-01 -7.43680418e-01 3.00148666e-01
-3.12883407e-01 -6.96335733e-01 -8.49373564e-02 5.46161473e-01
-8.77938628e-01 -7.52151728e-01 -1.05272460e+00 -7.07444489e-01
4.87333089e-01 2.75463969e-01 4.89008486e-01 -5.05285740e-01
-5.65375313e-02 6.23570383e-01 -3.45263064e-01 -1.00031757e+00
-2.50744790e-01 -2.33090386e-01 2.90310770e-01 4.04262364e-01
6.86946586e-02 -1.00638950e+00 -2.28614330e-01 4.75050062e-02
-7.95617819e-01 -7.03606784e-01 5.03088593e-01 4.54558283e-01
2.26813883e-01 6.07133865e-01 8.07557046e-01 -2.10567802e-01
9.47830737e-01 -4.72604781e-01 -6.84952736e-01 -1.44813061e-01
-1.74083650e-01 -3.27815592e-01 7.32394755e-01 -5.37634373e-01
-1.22278011e+00 2.08304793e-01 8.65873939e-04 -1.41497582e-01
-4.23411101e-01 4.01145577e-01 -3.88402402e-01 -8.39560255e-02
4.64921355e-01 2.81085014e-01 -7.24055171e-02 -8.27783346e-01
-2.73467284e-02 7.12843001e-01 7.25685656e-01 -5.86458147e-02
8.78790140e-01 2.03188583e-01 1.20551296e-01 -1.11738598e+00
-2.25034714e-01 -9.22566533e-01 -6.75957918e-01 -1.57810822e-01
6.49999857e-01 -8.09697747e-01 -5.55641651e-01 5.50603688e-01
-1.27042770e+00 4.27452534e-01 -1.99109197e-01 1.19333041e+00
-1.93952829e-01 3.97232056e-01 -1.61135793e-02 -1.44630814e+00
-1.98358390e-02 -1.00522077e+00 6.13568246e-01 4.41871017e-01
-3.34890276e-01 -8.50829363e-01 2.64451563e-01 2.31441453e-01
3.07831585e-01 1.72838956e-01 3.61749053e-01 -6.92362368e-01
-1.90778360e-01 -2.10964099e-01 3.35556656e-01 5.17450869e-01
3.27620238e-01 -7.66491890e-02 -1.27343106e+00 -1.78138047e-01
9.20581520e-01 4.87238914e-01 4.29143012e-01 7.98419833e-01
3.84346634e-01 -1.40202582e-01 2.81467456e-02 3.56862754e-01
1.42680812e+00 8.47330570e-01 3.81826907e-01 1.93490252e-01
3.75295848e-01 5.48507392e-01 3.44988555e-01 4.34386820e-01
-2.60254949e-01 3.93289059e-01 1.96864948e-01 -7.84752518e-02
-1.15210690e-01 2.53050655e-01 4.09827560e-01 1.11809456e+00
-2.04394415e-01 -3.61282647e-01 -6.67955339e-01 6.78091705e-01
-1.64424837e+00 -9.06241596e-01 -4.79471505e-01 2.40137744e+00
4.24539626e-01 -1.02326117e-01 4.49373692e-01 9.81877446e-01
7.62765646e-01 -5.16624302e-02 -2.55850479e-02 -5.08699417e-01
-1.35704175e-01 4.25316811e-01 4.70520735e-01 9.95758414e-01
-1.03101254e+00 1.31594762e-01 6.48615408e+00 6.16858482e-01
-1.08660078e+00 2.19221026e-01 -1.27391025e-01 -1.90662280e-01
1.12450182e-01 -1.33300619e-02 -4.83639687e-01 6.86438739e-01
8.64939749e-01 -1.96615960e-02 6.43035591e-01 2.85285324e-01
2.48450071e-01 -6.72590852e-01 -7.43529201e-01 1.16787398e+00
5.23363411e-01 -3.91860574e-01 -4.17660236e-01 -1.58953980e-01
5.32176495e-01 -3.97537231e-01 8.06993544e-02 -1.66138992e-01
-2.14643896e-01 -8.22377503e-01 7.80228376e-01 5.20773113e-01
4.38362248e-02 -7.88607180e-01 6.22383535e-01 3.48528564e-01
-1.20424831e+00 -3.84638995e-01 1.94248408e-02 -5.23639619e-01
4.10729825e-01 7.18389332e-01 -9.73283589e-01 7.25312233e-01
6.37603700e-01 6.70404732e-02 -3.62598270e-01 1.56170201e+00
-6.48656040e-02 9.56611395e-01 -8.24337065e-01 1.64402321e-01
-1.36952922e-02 -4.54126775e-01 1.23673797e+00 1.26646793e+00
7.36071646e-01 -1.00265667e-01 -5.05160354e-02 5.06166518e-01
5.32070100e-01 1.91759378e-01 -4.59912807e-01 3.78244370e-01
5.75129747e-01 1.23267722e+00 -9.68100071e-01 -1.35346204e-01
-1.86151683e-01 6.75278485e-01 -5.88290989e-01 5.36479533e-01
-6.41032338e-01 -5.17307043e-01 1.51226372e-01 -1.41858697e-01
4.15146589e-01 -4.48452473e-01 -4.22964245e-01 -6.04579031e-01
3.82363982e-02 -9.68053520e-01 2.94023633e-01 -5.57546616e-01
-9.20722067e-01 3.40750039e-01 1.63279295e-01 -1.28500736e+00
-3.44136566e-01 -2.64256090e-01 -6.50187671e-01 1.13546193e+00
-1.18958187e+00 -7.27596581e-01 5.30350171e-02 7.51770198e-01
6.68687642e-01 -3.22396606e-01 8.95175099e-01 7.17300653e-01
-3.92201334e-01 2.52284974e-01 4.98067051e-01 -2.79139102e-01
8.04591894e-01 -1.29938567e+00 -4.12259936e-01 1.24087477e+00
1.46790445e-01 5.10705709e-01 1.15383577e+00 -4.96320009e-01
-1.03093112e+00 -3.57915431e-01 9.71931219e-01 -2.81526357e-01
5.47354400e-01 -3.85728449e-01 -7.22717106e-01 4.23318714e-01
4.37645614e-01 -4.07456458e-01 1.05367458e+00 -6.49266839e-02
1.36601672e-01 -3.37662220e-01 -9.43334818e-01 2.35258251e-01
2.33596623e-01 -3.48258704e-01 -1.01841855e+00 -5.45401610e-02
1.42557099e-01 -1.05316415e-01 -5.28772473e-01 8.00552294e-02
4.99332696e-01 -1.07990360e+00 9.94822025e-01 1.10401288e-01
-2.76857823e-01 -7.74729908e-01 -1.08893909e-01 -1.29938412e+00
-3.84212464e-01 -1.02673483e+00 2.84460485e-01 1.67656088e+00
2.66717523e-01 -9.50088322e-01 1.19636275e-01 1.46138355e-01
5.95776513e-02 2.14401603e-01 -1.32497823e+00 -8.63587976e-01
-3.99109840e-01 -3.14107507e-01 3.40944439e-01 7.52989173e-01
7.17358217e-02 5.54384470e-01 -3.93620908e-01 4.74076718e-01
6.44248068e-01 -2.22943366e-01 7.71504700e-01 -1.61773467e+00
-6.27020717e-01 -3.26131970e-01 -9.99453068e-02 -7.13037133e-01
-1.59940839e-01 -4.17650163e-01 1.43940523e-01 -1.09776092e+00
-4.21873719e-01 -1.97429970e-01 -4.65654910e-01 -1.67411029e-01
-2.05558404e-01 1.91020489e-01 4.91065443e-01 1.09912753e-01
-1.44238293e-01 1.02568902e-01 7.35885978e-01 2.00633347e-01
-7.20556080e-01 4.04706508e-01 -5.60034215e-01 1.00439513e+00
6.55105531e-01 -6.59062564e-01 -6.57602668e-01 -2.93545216e-01
1.58072934e-01 2.41440207e-01 2.85677731e-01 -1.46734405e+00
4.33726966e-01 4.08696830e-01 3.98289025e-01 -6.43787324e-01
7.05909908e-01 -1.17757511e+00 6.47741199e-01 3.53080183e-01
3.80706824e-02 1.71593741e-01 4.05915618e-01 5.91401041e-01
-4.84681040e-01 -6.62754118e-01 6.55575097e-01 -7.79427066e-02
-2.10266918e-01 -7.04322755e-01 -8.36206198e-01 -3.20214152e-01
9.56039488e-01 -4.25043225e-01 2.72869587e-01 -5.86191654e-01
-8.04907024e-01 -3.45774859e-01 -3.07711661e-01 5.25113903e-02
2.31635571e-01 -9.56735373e-01 -7.43425012e-01 3.94837052e-01
-7.33837724e-01 -7.72370100e-02 1.63456604e-01 1.21763003e+00
-2.60502607e-01 4.28842396e-01 -2.68170267e-01 -4.41951632e-01
-1.68178058e+00 3.72225493e-01 1.55525967e-01 -1.03457443e-01
4.43692580e-02 9.59153116e-01 -1.13605365e-01 4.13395651e-02
3.05229485e-01 -2.25475401e-01 -6.58488810e-01 5.52976072e-01
4.90930140e-01 1.06075263e+00 -5.59458695e-02 -7.55437672e-01
-4.68571275e-01 8.96331429e-01 4.93171394e-01 -8.21965814e-01
1.18297911e+00 -4.29490447e-01 -5.09119332e-01 9.50158238e-01
9.21192944e-01 9.82819676e-01 -7.20315814e-01 1.35378212e-01
-9.59570706e-03 -5.28361499e-01 -1.64642669e-02 -8.78790140e-01
-5.70187330e-01 7.82926023e-01 8.14326644e-01 5.51333129e-01
1.45504463e+00 -4.97936070e-01 2.58807272e-01 3.79660204e-02
-1.23775844e-02 -9.78470027e-01 -2.08605140e-01 4.47834134e-01
8.70968163e-01 -8.60915601e-01 -1.52416393e-01 -4.73702520e-01
-2.95122355e-01 1.11375022e+00 2.09904462e-02 -8.58947933e-02
7.96669245e-01 4.22437251e-01 2.74905026e-01 2.74942100e-01
7.14182779e-02 -2.03448325e-01 2.89087713e-01 5.84416449e-01
4.67732042e-01 -3.67390923e-02 -6.24261379e-01 7.51466334e-01
-3.24625820e-01 9.09544826e-02 4.57203954e-01 8.74844432e-01
-4.63918835e-01 -1.08785796e+00 -1.47201276e+00 -8.20138957e-03
-7.95778215e-01 -1.59676865e-01 -2.75216937e-01 4.31531519e-01
4.08248335e-01 1.71020579e+00 -1.78780913e-01 6.99263345e-03
3.95057291e-01 3.98803532e-01 3.44066978e-01 -3.24621260e-01
-7.55882323e-01 9.59010005e-01 -1.74108982e-01 1.00329302e-01
-7.24646747e-01 -9.68474209e-01 -9.03152823e-01 2.25700110e-01
-4.25176650e-01 6.85571313e-01 1.01234889e+00 8.59157264e-01
8.04818273e-02 8.95292640e-01 5.76019943e-01 -8.59806895e-01
-3.94713074e-01 -1.08949590e+00 -8.67235363e-01 -4.14816737e-02
7.24394441e-01 -4.64374244e-01 -7.70773530e-01 2.33131826e-01] | [15.163986206054688, 5.664031028747559] |
78a20a95-1fd8-45e4-aa85-78cd159d177f | video-salient-object-detection-via | 2111.02368 | null | https://arxiv.org/abs/2111.02368v1 | https://arxiv.org/pdf/2111.02368v1.pdf | Video Salient Object Detection via Contrastive Features and Attention Modules | Video salient object detection aims to find the most visually distinctive objects in a video. To explore the temporal dependencies, existing methods usually resort to recurrent neural networks or optical flow. However, these approaches require high computational cost, and tend to accumulate inaccuracies over time. In this paper, we propose a network with attention modules to learn contrastive features for video salient object detection without the high computational temporal modeling techniques. We develop a non-local self-attention scheme to capture the global information in the video frame. A co-attention formulation is utilized to combine the low-level and high-level features. We further apply the contrastive learning to improve the feature representations, where foreground region pairs from the same video are pulled together, and foreground-background region pairs are pushed away in the latent space. The intra-frame contrastive loss helps separate the foreground and background features, and the inter-frame contrastive loss improves the temporal consistency. We conduct extensive experiments on several benchmark datasets for video salient object detection and unsupervised video object segmentation, and show that the proposed method requires less computation, and performs favorably against the state-of-the-art approaches. | ['Ming-Hsuan Yang', 'Xiaohui Shen', 'Xiaojie Jin', 'Yi-Wen Chen'] | 2021-11-03 | null | null | null | null | ['video-salient-object-detection', 'unsupervised-video-object-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.48697823e-01 -5.36963642e-01 -3.93201381e-01 -1.73458546e-01
-4.53730732e-01 -3.79587002e-02 2.77403265e-01 -2.82515376e-03
-4.57738906e-01 5.24261951e-01 2.61501938e-01 2.99913913e-01
1.92985982e-02 -4.09598380e-01 -7.21156538e-01 -1.05805826e+00
-1.91382557e-01 -4.87467587e-01 9.00636852e-01 7.09622428e-02
3.11033458e-01 2.28250965e-01 -1.50278461e+00 2.64153838e-01
7.76062071e-01 1.13926649e+00 6.11672461e-01 4.62062091e-01
-1.11832716e-01 1.37916470e+00 -1.60481185e-01 1.58759937e-01
1.31231248e-01 -5.82152426e-01 -7.05806494e-01 3.48429084e-01
4.68500674e-01 -6.02156460e-01 -5.89297175e-01 1.29696441e+00
2.81568199e-01 4.96296227e-01 1.78508773e-01 -1.29691577e+00
-7.01115429e-01 3.24176669e-01 -1.01930869e+00 9.40783978e-01
1.02395430e-01 2.12500870e-01 8.92722666e-01 -9.91235912e-01
6.23246491e-01 1.34279466e+00 2.57495582e-01 4.34440643e-01
-9.26009953e-01 -5.58512449e-01 8.34814191e-01 5.92288196e-01
-1.16841602e+00 -4.27637577e-01 1.23524964e+00 -4.60630476e-01
4.94166642e-01 2.24940583e-01 8.87026429e-01 7.60500610e-01
1.98534891e-01 1.37627423e+00 5.95141768e-01 -1.31274700e-01
1.32961199e-02 -2.14998037e-01 1.85000449e-01 9.78887856e-01
1.40139207e-01 1.00700952e-01 -6.29583776e-01 5.24282269e-02
1.02848768e+00 6.82300329e-01 -5.29930890e-01 -4.98105228e-01
-1.07073081e+00 8.36295128e-01 6.25444531e-01 4.61897701e-01
-4.83822495e-01 2.34546289e-01 4.68302459e-01 -1.15480155e-01
5.76210916e-01 -1.10615835e-01 -2.96772093e-01 1.64534543e-02
-1.21708596e+00 1.23442523e-01 8.04846436e-02 7.98030674e-01
8.72771442e-01 2.47280076e-01 -5.42546272e-01 6.22079790e-01
5.27115703e-01 1.21651538e-01 5.37001133e-01 -9.78890359e-01
3.23432684e-01 5.45789003e-01 1.82628229e-01 -1.52699506e+00
-1.91242784e-01 -3.92606407e-01 -7.15352714e-01 1.85721237e-02
1.47364601e-01 -1.03357963e-01 -9.69821274e-01 1.65735650e+00
4.25389171e-01 6.22307062e-01 -4.10580859e-02 1.42993081e+00
8.92340541e-01 1.07530665e+00 3.39319110e-01 -6.93320334e-01
1.02745879e+00 -1.46515799e+00 -8.77065718e-01 -2.70743042e-01
9.17961895e-02 -7.79943764e-01 7.34727859e-01 -2.05420554e-01
-1.50873566e+00 -8.17557395e-01 -8.07357669e-01 -2.71080822e-01
1.33010134e-01 -3.49303968e-02 5.59549212e-01 -1.11254454e-01
-7.67050445e-01 5.00870585e-01 -1.05006528e+00 -7.85382688e-02
7.02354848e-01 1.73425063e-01 9.09487084e-02 -2.98063904e-02
-1.12416565e+00 3.58714640e-01 4.94777769e-01 4.23555791e-01
-1.14519823e+00 -4.81085867e-01 -1.00313175e+00 2.15381503e-01
5.40968478e-01 -5.09284735e-01 9.51970339e-01 -1.50994778e+00
-1.24804926e+00 5.02723634e-01 -4.76918042e-01 -3.20802122e-01
4.99744385e-01 -4.07782048e-01 -3.03041726e-01 5.04818857e-01
1.94666758e-01 8.29590619e-01 1.19488025e+00 -1.17798734e+00
-1.07231283e+00 5.04063554e-02 7.66605437e-02 2.87254661e-01
-3.73089522e-01 4.55259591e-01 -7.48612106e-01 -9.78510201e-01
2.61484087e-01 -5.56738079e-01 -2.65547007e-01 1.77886277e-01
-8.01318809e-02 -1.30200386e-01 1.33340204e+00 -7.86903977e-01
1.34414852e+00 -2.24990749e+00 3.13306481e-01 -2.67116427e-01
2.91899920e-01 2.55815774e-01 5.35466336e-03 -3.17693591e-01
1.40193272e-02 -2.08803624e-01 -2.44534224e-01 -1.22571081e-01
-4.62560922e-01 6.18097931e-02 -2.22420856e-01 5.51787436e-01
4.58325893e-01 9.49421763e-01 -1.17211998e+00 -1.00384343e+00
3.84945601e-01 5.55006146e-01 -5.85422397e-01 3.19838107e-01
-1.36911690e-01 3.66032273e-01 -7.13390112e-01 7.46007740e-01
6.28911138e-01 -4.36696649e-01 -2.20506459e-01 -1.82879686e-01
-2.79926121e-01 6.77147657e-02 -1.11493158e+00 1.66752183e+00
9.45117623e-02 7.65342236e-01 1.03139102e-01 -1.05504358e+00
6.40446484e-01 -7.65910279e-03 6.85232639e-01 -6.35072649e-01
2.34003916e-01 -1.22131705e-02 1.92485824e-02 -7.41283000e-01
5.38421333e-01 1.84675619e-01 4.08518881e-01 1.05746448e-01
-1.94601645e-03 6.87321067e-01 3.46790165e-01 1.35077149e-01
5.38280368e-01 3.33737165e-01 8.66861120e-02 -4.55445260e-01
8.80190372e-01 -2.88198739e-01 1.11149704e+00 4.41155493e-01
-6.63868785e-01 6.19064212e-01 2.69616455e-01 -7.18576074e-01
-6.37140274e-01 -8.24282646e-01 2.14962810e-01 1.43287051e+00
9.02937055e-01 -1.24462686e-01 -6.67811990e-01 -6.25932634e-01
-2.01906443e-01 1.37453258e-01 -6.74503922e-01 -3.29324752e-01
-8.48220766e-01 -6.99519515e-01 -2.24031597e-01 6.26856983e-01
6.83643103e-01 -1.36426747e+00 -9.73688483e-01 4.05851096e-01
-3.52909863e-01 -9.50559735e-01 -9.43572342e-01 -2.21185058e-01
-9.65656281e-01 -8.07792425e-01 -1.09249747e+00 -1.29496396e+00
7.13259995e-01 7.09789336e-01 8.36412966e-01 3.46304268e-01
-3.28111768e-01 7.41331950e-02 -2.75725603e-01 6.26182854e-02
2.29949266e-01 -1.41267434e-01 -7.48968273e-02 4.31588620e-01
2.04921246e-01 -3.10301304e-01 -9.42866623e-01 3.05572748e-01
-9.30476189e-01 1.92751944e-01 4.44518447e-01 9.35649574e-01
6.96071625e-01 -9.47160348e-02 3.06412339e-01 -3.49299252e-01
5.85296229e-02 -4.36636269e-01 -5.26860535e-01 1.46355897e-01
-1.15883976e-01 -1.11163862e-01 3.32713395e-01 -6.37792766e-01
-1.11620378e+00 1.03428468e-01 3.75915855e-01 -8.92705321e-01
1.60647824e-01 2.66879171e-01 -2.07144156e-01 -6.00151382e-02
2.69217137e-03 5.21786392e-01 -3.27744544e-01 -2.62144446e-01
1.33405089e-01 1.62170857e-01 5.35731614e-01 -3.93020183e-01
5.94666898e-01 6.32129192e-01 -2.97465980e-01 -6.90495849e-01
-1.07426620e+00 -5.67536175e-01 -4.97415394e-01 -4.01243210e-01
1.02000475e+00 -1.04511285e+00 -4.81610030e-01 5.49130142e-01
-1.17206609e+00 -2.36190274e-01 -2.08754405e-01 5.40066957e-01
-4.69192892e-01 5.00178933e-01 -8.55140567e-01 -8.43685806e-01
-4.08894986e-01 -1.22055352e+00 1.02017879e+00 7.85188019e-01
1.30412593e-01 -7.92828679e-01 -2.66126543e-01 3.69341522e-02
2.36929461e-01 2.09237188e-01 4.22698617e-01 -9.36554000e-02
-1.03331804e+00 2.07785994e-01 -5.49760163e-01 1.39943346e-01
3.22777390e-01 2.66308039e-01 -7.92842329e-01 -3.89538527e-01
1.54721439e-01 4.28690430e-04 1.19094706e+00 8.55198920e-01
1.34134340e+00 -4.13122624e-01 -4.15245533e-01 6.39388204e-01
1.23427868e+00 3.41662467e-01 4.04169381e-01 2.98332125e-01
1.13579571e+00 5.87291777e-01 9.13191259e-01 3.57369155e-01
2.59121358e-01 5.28414011e-01 3.41804117e-01 -4.26148176e-01
-2.97001414e-02 -1.13498807e-01 4.99104649e-01 8.82370174e-01
-2.57702708e-01 -5.86258881e-02 -5.32824934e-01 8.12268436e-01
-2.26805329e+00 -1.37138772e+00 6.21538758e-02 1.85698974e+00
7.47005939e-01 3.29739779e-01 2.61519164e-01 -2.63718247e-01
1.11421394e+00 5.64674556e-01 -6.94010198e-01 8.32213387e-02
-3.28548968e-01 -3.57906699e-01 1.85509220e-01 4.99536872e-01
-1.46229887e+00 1.01294088e+00 5.83459711e+00 7.96481609e-01
-1.27651048e+00 1.25777334e-01 9.51807737e-01 -4.52782571e-01
-4.51946110e-02 -8.63214396e-03 -5.94348907e-01 8.79123271e-01
2.86396861e-01 -1.25688598e-01 1.19260110e-01 8.73966873e-01
3.66869986e-01 -1.41712949e-01 -8.59916747e-01 1.15819418e+00
2.67277628e-01 -1.47682178e+00 3.70475128e-02 -2.95744240e-01
1.03276014e+00 -6.69335872e-02 9.83144864e-02 8.70576799e-02
-2.23913819e-01 -5.91429114e-01 9.80572104e-01 8.03699136e-01
1.13884449e-01 -8.69779289e-01 5.93723297e-01 2.40904726e-02
-1.65413415e+00 -3.03467512e-01 -4.17984664e-01 -5.80982007e-02
2.65759289e-01 4.21878874e-01 1.52926728e-01 3.35719436e-01
1.04663718e+00 1.23682868e+00 -4.36636418e-01 1.11961567e+00
2.12850813e-02 4.28734422e-01 -7.21997172e-02 -1.26352146e-01
5.65447152e-01 -2.51702785e-01 7.09273338e-01 1.16140711e+00
-8.03206116e-02 1.85758218e-01 6.81013763e-01 9.89728093e-01
-1.12428074e-03 1.16551802e-01 -2.47916393e-02 1.52640507e-01
1.12816580e-01 1.28357160e+00 -9.12985027e-01 -6.48696423e-01
-5.68264484e-01 1.17991173e+00 1.65142551e-01 4.99512643e-01
-1.03394604e+00 -3.42529953e-01 6.50360644e-01 -6.04217965e-03
6.71060324e-01 -7.93050826e-02 1.95612535e-01 -1.32501066e+00
2.14937419e-01 -5.32389522e-01 5.21107137e-01 -7.58056462e-01
-1.13304389e+00 4.03344542e-01 -1.83394089e-01 -1.41839111e+00
-1.10769376e-01 -9.63939503e-02 -7.95885205e-01 7.11700559e-01
-1.77503729e+00 -8.53283823e-01 -3.86489213e-01 6.19981229e-01
9.52838898e-01 6.50783777e-02 -6.26904815e-02 4.01937842e-01
-8.20116043e-01 3.24847400e-01 5.31128906e-02 3.83477062e-01
4.42557395e-01 -7.54077971e-01 1.26280367e-01 1.26255620e+00
-9.09797549e-02 5.56873381e-01 5.56804359e-01 -6.21407986e-01
-1.01968420e+00 -1.32298291e+00 5.75064123e-01 -2.60942470e-04
5.78269422e-01 -1.41876876e-01 -1.21203971e+00 4.97124195e-01
2.17216596e-01 6.15018368e-01 2.47059524e-01 -3.42233270e-01
-4.56949556e-03 -7.13145062e-02 -7.98304737e-01 6.13240957e-01
1.02673411e+00 -4.19135690e-01 -6.81234777e-01 6.71369657e-02
9.93053913e-01 -2.94097185e-01 -4.74167109e-01 3.71192038e-01
4.48134333e-01 -8.40782881e-01 9.14242566e-01 -6.50096595e-01
6.77343547e-01 -7.33299196e-01 5.76513559e-02 -6.76520467e-01
-6.30404055e-01 -8.01354825e-01 -4.70207900e-01 1.35084689e+00
-6.24563396e-02 -1.96765527e-01 7.65519261e-01 4.54182178e-01
-8.58339071e-02 -8.10440600e-01 -6.95635617e-01 -5.35005212e-01
-3.74374926e-01 4.64945547e-02 1.66054308e-01 1.00353038e+00
-1.99628070e-01 2.19838127e-01 -6.10453308e-01 1.38357386e-01
6.05274260e-01 5.95069587e-01 3.19753170e-01 -1.09034860e+00
-1.84060365e-01 -6.46399438e-01 -5.16930699e-01 -1.24489450e+00
9.40350890e-02 -5.05482137e-01 2.23944053e-01 -1.23859513e+00
6.80282414e-01 -8.49760883e-03 -9.06181276e-01 3.95425171e-01
-7.94095993e-01 1.16410099e-01 3.17420393e-01 3.12559903e-01
-1.12411356e+00 8.51494610e-01 1.32713556e+00 -3.06276441e-01
-4.34855372e-01 -3.91755730e-01 -4.71915334e-01 1.04363513e+00
5.45775652e-01 -4.93229359e-01 -2.91764885e-01 -5.09417176e-01
-4.14772391e-01 -2.11871648e-03 5.43239057e-01 -9.93265569e-01
3.21303457e-01 -4.29317087e-01 7.89793730e-01 -6.49070680e-01
2.29186743e-01 -6.41582429e-01 -3.71550739e-01 5.40136695e-01
-3.74702603e-01 -4.29160185e-02 1.81767821e-01 6.61512733e-01
-4.17671651e-01 -5.57839759e-02 9.96170938e-01 -4.80229706e-02
-1.03908384e+00 7.33351886e-01 -2.87551194e-01 1.60935655e-01
1.05829287e+00 -2.37517849e-01 -8.12892988e-02 -1.38319984e-01
-5.83526731e-01 3.84089828e-01 3.83782744e-01 6.45049453e-01
8.02984297e-01 -1.44878280e+00 -6.08296931e-01 2.35460013e-01
-8.33129212e-02 1.50978537e-02 7.17352331e-01 9.11744058e-01
-4.00764436e-01 3.75709049e-02 -4.45844978e-01 -8.93437266e-01
-1.28221476e+00 8.59270096e-01 3.50241691e-01 1.24106236e-01
-7.39336908e-01 1.10068727e+00 6.62463367e-01 5.00513613e-01
2.43462786e-01 -3.54393154e-01 -3.17507446e-01 1.45978078e-01
7.83273458e-01 4.21285987e-01 -6.24001920e-01 -9.67134595e-01
-4.66587991e-01 8.63322556e-01 -3.76628906e-01 2.16750279e-01
1.22650003e+00 -4.23454642e-01 -1.23606279e-01 4.90792215e-01
1.39010370e+00 -2.45241299e-01 -1.86220908e+00 -4.72424656e-01
-1.27398357e-01 -8.18758309e-01 3.30161244e-01 -6.93451539e-02
-1.57030141e+00 8.90544057e-01 7.95895100e-01 1.84012443e-01
1.36558437e+00 -8.72981176e-02 1.02418232e+00 7.61206076e-03
-2.96562538e-02 -1.11178935e+00 2.79847085e-01 3.26994658e-01
6.39144123e-01 -1.42507088e+00 1.42393932e-01 -4.03728127e-01
-6.57205462e-01 9.18600202e-01 7.87208736e-01 -4.87620205e-01
4.98706222e-01 -6.29526079e-02 -9.49802473e-02 -2.35010851e-02
-6.69572711e-01 -3.23209584e-01 4.01497304e-01 2.16206238e-01
2.67159164e-01 -3.69964570e-01 -2.35220715e-01 5.08335769e-01
5.89641988e-01 -1.99252535e-02 2.97350585e-01 1.06844974e+00
-6.29101932e-01 -4.70108747e-01 -2.53460407e-01 2.69587785e-01
-6.14108622e-01 -1.42458037e-01 -1.63236912e-03 5.48743665e-01
2.32215032e-01 7.77064085e-01 3.15575600e-01 -1.12031877e-01
-1.06171174e-02 -3.26569319e-01 2.19361752e-01 -2.04117373e-01
-2.64020801e-01 6.06018603e-01 -5.30037701e-01 -6.51382923e-01
-9.12943721e-01 -9.26552594e-01 -1.45275640e+00 1.90393552e-02
-3.54564518e-01 1.01575926e-01 -6.54749870e-02 8.02471101e-01
2.22614184e-01 7.79239953e-01 8.13880026e-01 -1.08659482e+00
6.50872737e-02 -6.13545418e-01 -4.07962739e-01 5.19478977e-01
8.39223325e-01 -8.77227068e-01 -2.79241204e-01 4.44021851e-01] | [9.546661376953125, -0.2995413541793823] |
7ec098c0-3395-4e03-a3ea-45a1828b2808 | multi-modal-fusion-for-diabetes-mellitus-and | 1604.03443 | null | http://arxiv.org/abs/1604.03443v1 | http://arxiv.org/pdf/1604.03443v1.pdf | Multi-modal Fusion for Diabetes Mellitus and Impaired Glucose Regulation Detection | Effective and accurate diagnosis of Diabetes Mellitus (DM), as well as its
early stage Impaired Glucose Regulation (IGR), has attracted much attention
recently. Traditional Chinese Medicine (TCM) [3], [5] etc. has proved that
tongue, face and sublingual diagnosis as a noninvasive method is a reasonable
way for disease detection. However, most previous works only focus on a single
modality (tongue, face or sublingual) for diagnosis, although different
modalities may provide complementary information for the diagnosis of DM and
IGR. In this paper, we propose a novel multi-modal classification method to
discriminate between DM (or IGR) and healthy controls. Specially, the tongue,
facial and sublingual images are first collected by using a non-invasive
capture device. The color, texture and geometry features of these three types
of images are then extracted, respectively. Finally, our so-called multi-modal
similar and specific learning (MMSSL) approach is proposed to combine features
of tongue, face and sublingual, which not only exploits the correlation but
also extracts individual components among them. Experimental results on a
dataset consisting of 192 Healthy, 198 DM and 114 IGR samples (all samples were
obtained from Guangdong Provincial Hospital of Traditional Chinese Medicine)
substantiate the effectiveness and superiority of our proposed method for the
diagnosis of DM and IGR, compared to the case of using a single modality. | ['Jinxing Li', 'Jian Wu', 'David Zhang', 'Yongcheng Li'] | 2016-04-12 | null | null | null | null | ['multi-modal-classification'] | ['miscellaneous'] | [ 1.02562435e-01 -6.16568744e-01 -6.45999730e-01 -2.05828279e-01
-6.51482701e-01 -2.59280533e-01 2.86733508e-01 -4.02034670e-02
6.57484308e-02 6.73108518e-01 7.53901452e-02 -7.30756372e-02
-3.38730007e-01 -6.65870547e-01 1.69497997e-01 -1.06114197e+00
-1.00682259e-01 5.10546565e-01 -2.82436222e-01 1.17210068e-01
2.38869965e-01 3.75306636e-01 -1.49338055e+00 1.85811713e-01
1.44984865e+00 1.16278434e+00 2.20587179e-01 2.84205437e-01
-2.94286191e-01 2.37366036e-01 -2.18703344e-01 -3.90875004e-02
1.20984219e-01 -7.83755243e-01 -3.83537889e-01 5.14828265e-01
3.31921548e-01 -4.29476202e-01 7.72425532e-02 1.13656390e+00
7.73069322e-01 -2.21945703e-01 8.79980028e-01 -9.59353566e-01
-7.83247650e-01 1.99010018e-02 -8.65130365e-01 -5.92431426e-02
2.74622470e-01 -2.40641329e-02 2.01447248e-01 -4.82656837e-01
3.38817924e-01 1.30065143e+00 3.94771188e-01 6.69553697e-01
-9.51755106e-01 -7.61376321e-01 4.63584960e-02 4.65467125e-01
-1.33593822e+00 -3.43363881e-01 7.37134457e-01 -4.08072025e-01
5.74198132e-03 1.92463428e-01 8.31388891e-01 6.02640510e-01
3.70798290e-01 6.72708452e-01 1.72360063e+00 -3.66074026e-01
1.28408084e-02 1.94288716e-01 -3.34817655e-02 6.99761868e-01
3.23355496e-01 1.55469462e-01 1.96650401e-02 -1.31960943e-01
7.85803258e-01 3.42306137e-01 -3.27842385e-01 -1.22036129e-01
-8.93629014e-01 6.97160840e-01 1.69714056e-02 5.28707325e-01
-4.10974085e-01 -6.27853453e-01 2.50681996e-01 2.53901362e-01
4.84239429e-01 -3.76387179e-01 -4.53325272e-01 2.43146241e-01
-5.07208467e-01 -2.88649559e-01 5.52463233e-01 7.75541604e-01
5.31842649e-01 -6.16502352e-02 -7.49837458e-02 9.60754812e-01
5.88857055e-01 8.71124506e-01 7.41539478e-01 -6.39379442e-01
3.74686159e-02 8.80460739e-01 -2.32777208e-01 -8.98761213e-01
-5.16099215e-01 -1.02304883e-01 -1.05413568e+00 8.54816064e-02
3.31622541e-01 -1.91928178e-01 -1.01385379e+00 1.14121592e+00
5.11612177e-01 1.81998819e-01 1.60808504e-01 1.10730004e+00
1.05514693e+00 2.89821893e-01 2.59231269e-01 -5.87782681e-01
1.72974288e+00 -4.74978775e-01 -8.53691816e-01 1.46257445e-01
4.50350821e-01 -9.74851251e-01 6.46902859e-01 4.69197899e-01
-7.70076931e-01 -4.18005943e-01 -7.39102304e-01 2.03825563e-01
-2.99284190e-01 5.25145233e-01 6.74124539e-01 8.20207477e-01
-5.19244969e-01 3.98778349e-01 -5.71469069e-01 -6.48522019e-01
3.76915991e-01 3.32520038e-01 -4.00423050e-01 -6.52717948e-01
-1.00130606e+00 1.06148422e+00 6.64826632e-02 1.26756251e-01
-3.33955497e-01 -6.16850972e-01 -7.56449640e-01 -5.71152270e-01
1.70571953e-01 -6.85385525e-01 5.73789239e-01 -9.76018548e-01
-1.62359774e+00 9.15200174e-01 -3.16377968e-01 2.57293582e-01
2.70614326e-01 9.00290459e-02 -8.62504780e-01 6.24625027e-01
-1.39790783e-02 8.70272219e-02 6.52829587e-01 -1.10748267e+00
-7.43832171e-01 -1.03800106e+00 -2.79692382e-01 5.81203494e-03
1.75390374e-02 1.14105090e-01 -1.29991144e-01 -4.05251116e-01
4.84477550e-01 -7.57455766e-01 3.82552221e-02 3.02307338e-01
-4.08611417e-01 -3.89142454e-01 9.58720386e-01 -1.00645661e+00
8.48389089e-01 -2.06063700e+00 7.78234825e-02 3.69039804e-01
1.59874782e-01 4.12111282e-01 -1.94180071e-01 2.31699599e-03
9.60239917e-02 3.17525327e-01 3.62606049e-02 1.87490553e-01
-1.41481489e-01 3.13371837e-01 7.68687010e-01 7.67731488e-01
2.27963209e-01 4.86821413e-01 -7.16765106e-01 -8.51270556e-01
6.07098460e-01 6.73826337e-01 -8.15395862e-02 -1.25413373e-01
1.47811040e-01 8.13682079e-01 -6.72954738e-01 1.44154191e+00
1.11936188e+00 1.51320100e-02 5.04579425e-01 -6.01366937e-01
-2.60214686e-01 -4.27192509e-01 -1.00792837e+00 1.32538486e+00
-2.02672943e-01 -4.38486896e-02 1.18109867e-01 -1.05780113e+00
9.10704076e-01 7.12477684e-01 7.70450592e-01 -1.12544417e+00
4.57080722e-01 5.09111226e-01 1.20016873e-01 -1.17643583e+00
-5.43353021e-01 -4.81542945e-01 4.13695484e-01 -1.12763703e-01
-1.43078327e-01 6.01079702e-01 3.76586378e-01 -4.50402588e-01
1.75958872e-01 1.94199622e-01 3.86597067e-01 -2.19713479e-01
8.63845944e-01 -2.41325237e-02 7.15089679e-01 4.61899154e-02
-4.88408864e-01 3.28088313e-01 4.05512899e-01 -1.63449407e-01
-4.96283591e-01 -1.00682616e+00 -7.26354718e-01 3.24710757e-01
3.89408886e-01 9.98704582e-02 -1.83029443e-01 -5.90254426e-01
3.17938834e-01 1.22167319e-01 -4.39357042e-01 -2.05591783e-01
-3.71783674e-01 -1.23141778e+00 2.03025252e-01 3.22340280e-01
9.22861576e-01 -3.00713122e-01 -2.63714697e-02 4.34539951e-02
-2.14669645e-01 -8.09552073e-01 -3.05690616e-01 -3.70318711e-01
-1.22663271e+00 -1.42227995e+00 -1.15615666e+00 -9.27419901e-01
7.39669621e-01 2.93890268e-01 7.99328983e-01 1.78834260e-01
-4.58909631e-01 2.70105332e-01 -4.02818859e-01 -2.68751621e-01
-2.19570354e-01 -5.51529884e-01 1.81510169e-02 3.38482291e-01
6.22712374e-01 -6.09308541e-01 -9.38011885e-01 4.02465343e-01
-4.95437413e-01 -3.18647653e-01 1.13739562e+00 4.34261233e-01
7.01383829e-01 2.64803380e-01 7.30562091e-01 -8.58365476e-01
1.81732446e-01 -4.80812401e-01 -2.94367820e-01 2.79943675e-01
-7.66886294e-01 -4.81821984e-01 2.54656255e-01 -4.12461013e-01
-8.96036088e-01 -1.16999969e-01 -6.82452917e-02 -3.66045117e-01
-5.35765350e-01 6.37375176e-01 -3.57262462e-01 -3.36626977e-01
6.78106993e-02 4.00232583e-01 5.44968486e-01 -8.41682792e-01
-9.96873751e-02 8.91812503e-01 1.32098779e-01 -4.93826181e-01
5.08785784e-01 4.98398960e-01 3.72499585e-01 -9.94420171e-01
-5.15530765e-01 -5.87864935e-01 -6.34108841e-01 -2.97500610e-01
1.02575815e+00 -9.27228153e-01 -1.05112863e+00 6.60549104e-01
-5.29922247e-01 2.70852715e-01 4.68709707e-01 1.14405572e+00
-1.79628402e-01 4.49993104e-01 -5.46233475e-01 -9.12783861e-01
-3.55553716e-01 -8.23361516e-01 8.70875299e-01 4.63299453e-01
2.55454153e-01 -1.24347782e+00 -1.53970614e-01 4.45199192e-01
4.97288674e-01 7.36356854e-01 1.52586114e+00 -2.81065077e-01
-2.85618395e-01 -1.80009171e-01 -4.85964328e-01 3.10103774e-01
7.40016580e-01 7.04132989e-02 -5.45363307e-01 -1.30791068e-01
3.06200739e-02 1.15297571e-01 6.51870608e-01 7.44846582e-01
6.46285474e-01 -9.26557258e-02 -3.44779640e-01 4.89094138e-01
1.82038355e+00 7.73796737e-01 8.20730090e-01 8.74444619e-02
6.42005801e-01 4.48817223e-01 6.13985896e-01 4.50573593e-01
4.50744838e-01 4.22783852e-01 3.37352127e-01 -5.27779222e-01
-1.12496048e-01 5.05224705e-01 2.87736624e-01 7.25438714e-01
-6.98635101e-01 2.88843691e-01 -4.62160647e-01 3.08438867e-01
-1.33483112e+00 -8.72978389e-01 -5.04517853e-01 2.17032123e+00
7.54922092e-01 -4.83625501e-01 4.99885410e-01 4.27659638e-02
9.18125629e-01 -5.04610896e-01 -5.60411274e-01 -4.43865173e-02
-1.81031078e-01 3.54656398e-01 2.38926649e-01 -8.24266672e-02
-9.98294294e-01 3.46462019e-02 5.73676395e+00 6.06245279e-01
-1.49780428e+00 -2.12776177e-02 4.08526391e-01 1.00948706e-01
6.29252046e-02 -2.55466521e-01 -5.29345453e-01 6.33732498e-01
5.73430598e-01 2.59110313e-02 3.68828773e-01 4.43533957e-01
4.69885021e-01 -2.92342126e-01 -6.52511239e-01 1.01301312e+00
4.53403369e-02 -7.98506737e-01 1.10377170e-01 1.47309959e-01
4.83969003e-01 -4.99103367e-01 -1.12102311e-02 1.83993176e-01
-4.97148007e-01 -7.31256127e-01 -2.49097105e-02 8.80489886e-01
8.62348139e-01 -8.41785908e-01 1.06085801e+00 2.22683381e-02
-1.29291880e+00 1.09194204e-01 -1.32884160e-01 2.73757935e-01
-1.07023910e-01 6.92946970e-01 -2.05170631e-01 1.06950843e+00
4.00269002e-01 1.01952374e+00 -5.09142280e-01 1.25057936e+00
1.48900941e-01 2.06054211e-01 8.29143375e-02 1.20186560e-01
-9.45742801e-02 -6.86760187e-01 2.93306053e-01 8.08566213e-01
5.88307440e-01 3.45866323e-01 4.93025452e-01 6.67423368e-01
4.51079994e-01 4.65825975e-01 -2.10757673e-01 -1.26947343e-01
6.94174841e-02 1.32273567e+00 -5.01331627e-01 -4.65758801e-01
-7.29783058e-01 5.42647481e-01 -6.32664740e-01 2.01782286e-01
-6.07934356e-01 -2.61801898e-01 5.02812862e-01 3.02808899e-02
2.03642137e-02 1.79416195e-01 -6.45371452e-02 -1.28196502e+00
-2.13952929e-01 -9.98623729e-01 7.26132691e-01 -3.14533114e-01
-1.59859014e+00 1.19603455e-01 -1.37753144e-01 -1.60690224e+00
3.01099569e-01 -6.94612741e-01 -5.71405411e-01 1.06196141e+00
-2.00207853e+00 -1.36244380e+00 -3.88404518e-01 8.12982678e-01
1.88050270e-01 -3.47963534e-02 9.89408672e-01 7.64889061e-01
-9.77709591e-01 2.98545837e-01 2.74734110e-01 5.16356155e-02
7.19326675e-01 -1.12474322e+00 -1.03050148e+00 9.20009315e-02
-8.95918429e-01 6.82529569e-01 1.83651090e-01 -6.80608988e-01
-1.58525038e+00 -1.06060851e+00 6.91314340e-01 3.00048947e-01
7.30303973e-02 3.15620840e-01 -5.10450363e-01 2.60646433e-01
8.74397233e-02 -2.18216985e-01 1.07378769e+00 -1.38437003e-03
5.75840808e-02 -3.46004575e-01 -1.64110053e+00 3.64958882e-01
6.46551073e-01 -9.36783701e-02 -2.52669960e-01 3.93525302e-01
-1.28733844e-01 -2.93370515e-01 -1.50535738e+00 5.84181964e-01
1.02425504e+00 -8.35408807e-01 9.46466506e-01 -4.06623989e-01
2.66878188e-01 -5.55328786e-01 -2.34134808e-01 -1.05493522e+00
-2.68785208e-01 4.33080532e-02 5.21401465e-02 1.17410195e+00
-1.26254797e-01 -8.94274354e-01 3.54668051e-01 2.41421565e-01
-6.90154731e-02 -6.53442383e-01 -6.04685426e-01 -7.92825341e-01
1.90230943e-02 3.13899010e-01 2.96961546e-01 1.15490294e+00
-4.40281592e-02 5.61372973e-02 -3.31776321e-01 1.36376888e-01
9.65773344e-01 5.40555418e-01 2.55090147e-01 -1.65840435e+00
1.93249151e-01 -2.48624742e-01 -5.37267685e-01 -2.86171228e-01
-1.86388567e-01 -9.71242070e-01 -5.84808528e-01 -1.86487710e+00
3.96054894e-01 -3.91371936e-01 -4.18838948e-01 3.24802309e-01
-1.86324678e-02 4.51482058e-01 -2.87524275e-02 8.71044770e-02
-5.21585941e-02 1.92138568e-01 1.89652538e+00 -1.96712047e-01
-2.53099680e-01 1.70391798e-01 -9.00272906e-01 6.88746214e-01
8.47541213e-01 -2.04186663e-01 -2.01364964e-01 -5.75232692e-03
-5.72926641e-01 2.79425740e-01 5.24429917e-01 -7.86655068e-01
-2.58138806e-01 -4.06390250e-01 8.62168491e-01 -5.98260939e-01
1.40752003e-01 -8.68380606e-01 1.83201283e-01 8.71545672e-01
1.60552874e-01 -4.45503026e-01 7.84459263e-02 2.91658878e-01
-6.40307590e-02 2.27279905e-02 1.18050814e+00 -4.85017896e-01
-7.73095667e-01 3.52975845e-01 -5.82548201e-01 -2.73602307e-01
1.17517257e+00 -1.66080847e-01 -4.93175209e-01 3.89028564e-02
-1.16739070e+00 3.16031128e-01 1.33096233e-01 1.32504344e-01
6.31645977e-01 -1.60373926e+00 -9.02097702e-01 2.64558613e-01
3.67837176e-02 -3.63825381e-01 5.94774485e-01 1.79984629e+00
-2.26525024e-01 2.44327992e-01 -5.60165167e-01 -7.62969315e-01
-1.42372322e+00 5.38107038e-01 5.24277925e-01 1.01992860e-01
-6.14525557e-01 -5.55702345e-03 -6.84857070e-02 -1.08431898e-01
-5.31254597e-02 -2.99492508e-01 -5.72244465e-01 3.96347314e-01
5.69247246e-01 5.56805074e-01 -5.76240681e-02 -7.75178969e-01
-4.69700217e-01 1.31151175e+00 7.42800236e-02 5.65996826e-01
1.05405128e+00 -4.83029544e-01 -4.51858193e-01 3.77017885e-01
1.20265555e+00 -7.73885474e-02 -3.07147801e-01 -1.95426941e-01
-3.08272868e-01 -5.14882088e-01 1.41556740e-01 -9.44322467e-01
-1.44345534e+00 8.17055345e-01 1.28215432e+00 1.14802994e-01
1.45954740e+00 -9.37889218e-02 8.92769516e-01 -2.42334217e-01
3.16596508e-01 -8.23959589e-01 -2.12907538e-01 -1.58364892e-01
6.28692210e-01 -1.29234159e+00 9.16817319e-03 -5.98812640e-01
-4.48772073e-01 1.44348061e+00 3.85779381e-01 -6.05516508e-02
7.85934210e-01 -9.86037403e-02 4.95332718e-01 -6.51647523e-02
-3.53505254e-01 -6.10134363e-01 3.16345066e-01 9.01268899e-01
4.92682189e-01 1.64790705e-01 -7.69683599e-01 4.39552039e-01
5.04565716e-01 3.58948827e-01 4.79099303e-02 1.02902877e+00
-5.69039702e-01 -1.15979576e+00 -5.23976147e-01 7.40820229e-01
-5.81366181e-01 1.64765686e-01 -2.57122219e-01 1.02321470e+00
6.49203658e-01 1.06166053e+00 -1.54803917e-01 -3.07249367e-01
4.19626862e-01 8.41253102e-02 7.85369277e-01 -2.39381701e-01
-1.48686156e-01 7.46691942e-01 7.81590194e-02 -3.96891624e-01
-9.95425761e-01 -7.29839385e-01 -1.31181324e+00 -3.81620735e-01
-2.76894599e-01 -1.27733991e-01 5.53045213e-01 1.23931408e+00
6.51724786e-02 4.45442796e-01 8.70209992e-01 -5.43489039e-01
-2.19563678e-01 -8.88690531e-01 -1.27992451e+00 3.88333678e-01
3.61800760e-01 -8.13958347e-01 -4.12042618e-01 1.18362255e-01] | [15.829055786132812, -3.9758546352386475] |
21cdddbc-1226-4be9-9379-75a3db21f8b8 | conformal-link-prediction-to-control-the | 2306.14693 | null | https://arxiv.org/abs/2306.14693v1 | https://arxiv.org/pdf/2306.14693v1.pdf | Conformal link prediction to control the error rate | Most link prediction methods return estimates of the connection probability of missing edges in a graph. Such output can be used to rank the missing edges, from most to least likely to be a true edge, but it does not directly provide a classification into true and non-existent. In this work, we consider the problem of identifying a set of true edges with a control of the false discovery rate (FDR). We propose a novel method based on high-level ideas from the literature on conformal inference. The graph structure induces intricate dependence in the data, which we carefully take into account, as this makes the setup different from the usual setup in conformal inference, where exchangeability is assumed. The FDR control is empirically demonstrated for both simulated and real data. | ['Ariane Marandon'] | 2023-06-26 | null | null | null | null | ['link-prediction'] | ['graphs'] | [ 1.31846428e-01 5.16912341e-01 -2.60889828e-01 -3.71752322e-01
-2.06172109e-01 -5.48217297e-01 5.19160628e-01 3.91061425e-01
-8.19622576e-02 9.49886739e-01 -7.28590507e-03 -4.63989168e-01
-5.56807876e-01 -1.25742459e+00 -9.83612657e-01 -7.47424960e-01
-5.08671522e-01 6.75122142e-01 3.12551796e-01 7.07198381e-02
1.80682674e-01 3.76464337e-01 -1.28877068e+00 -9.20700133e-02
7.46223629e-01 4.58229065e-01 -1.88420061e-02 4.95774716e-01
1.16334133e-01 3.11427653e-01 -2.36230493e-01 -8.99228156e-01
6.62586279e-03 -3.97765040e-01 -8.12845111e-01 -2.04544693e-01
5.06763935e-01 7.61859268e-02 -1.35282651e-01 1.37819660e+00
1.94678888e-01 -2.63777047e-01 9.62217152e-01 -1.49233735e+00
-2.18103096e-01 7.84099579e-01 -5.49067676e-01 3.18158478e-01
2.20219955e-01 -4.39058781e-01 1.37414336e+00 -7.10879624e-01
6.62000120e-01 1.10540009e+00 9.12897229e-01 6.34669513e-02
-1.69867623e+00 -5.06769598e-01 -2.22830102e-01 1.04334451e-01
-1.62970495e+00 -1.85090378e-01 5.22663355e-01 -6.33618116e-01
1.38356820e-01 4.08853650e-01 5.29070735e-01 1.16761994e+00
3.85272741e-01 8.84589553e-02 1.14997339e+00 -5.65846205e-01
3.73621315e-01 2.16242075e-01 4.17065382e-01 7.55460382e-01
7.95323014e-01 2.72900075e-01 -4.20659482e-01 -5.41837990e-01
7.94920862e-01 -3.07246625e-01 -4.49941605e-01 -6.23456955e-01
-9.01723146e-01 9.42612410e-01 3.09083134e-01 2.01188162e-01
-1.23568818e-01 2.81245261e-02 8.59821588e-02 4.68508214e-01
4.90870357e-01 2.10369140e-01 -3.93326193e-01 3.13660413e-01
-6.20088875e-01 1.51391933e-02 1.03144169e+00 9.45371449e-01
8.13624084e-01 -5.89479387e-01 -3.44210356e-01 4.61810589e-01
3.87466758e-01 1.80350631e-01 -2.53046483e-01 -5.42516649e-01
3.16515446e-01 3.44772130e-01 1.14809014e-01 -1.20930994e+00
-3.85650694e-01 -6.26921833e-01 -8.94383132e-01 1.67968258e-01
7.37239778e-01 -1.69683650e-01 -3.83023590e-01 2.07192254e+00
3.68546069e-01 6.24020398e-01 -3.63895744e-01 7.25193918e-01
3.81721288e-01 2.40221605e-01 -2.43253391e-02 -3.64165068e-01
1.28686178e+00 -2.46775702e-01 -5.79416037e-01 1.97438851e-01
4.00279433e-01 -4.39351887e-01 6.71274245e-01 2.58310318e-01
-8.25433016e-01 -2.25798741e-01 -9.58288670e-01 3.45650285e-01
-1.73381239e-01 9.89664346e-02 8.26467276e-01 7.22803891e-01
-1.01178277e+00 8.50730836e-01 -5.26752770e-01 -2.95328081e-01
8.15502182e-02 3.07628810e-01 -4.64769065e-01 -6.17545992e-02
-1.31122780e+00 7.25277245e-01 2.33021855e-01 3.15579288e-02
-5.23876607e-01 -7.22962201e-01 -4.65605497e-01 3.15975249e-01
5.46858370e-01 -8.00673842e-01 6.25087440e-01 -7.94500589e-01
-8.64868581e-01 7.83762574e-01 -1.27095833e-01 -4.01212901e-01
8.02721441e-01 3.41056079e-01 -3.87495548e-01 -8.15374479e-02
1.17144592e-01 1.22177079e-02 6.58596814e-01 -1.04901838e+00
-3.66314918e-01 -4.78879035e-01 2.25033462e-01 -2.63928741e-01
-2.09616665e-02 -3.49930853e-01 -2.99258679e-01 -5.41928470e-01
2.32790798e-01 -9.22455907e-01 -1.07233915e-02 5.79703078e-02
-7.24272788e-01 -1.82726249e-01 4.26248536e-02 -2.69933999e-01
1.13374794e+00 -2.03794074e+00 8.80073011e-02 7.13850319e-01
3.05451959e-01 -4.83613402e-01 1.13179751e-01 8.62888336e-01
-2.87319720e-01 1.62020981e-01 -2.52425343e-01 -2.76064090e-02
-1.27528980e-01 1.88344032e-01 -1.75560623e-01 7.10284352e-01
2.44094878e-01 5.80059052e-01 -7.27548957e-01 -3.99055183e-01
-3.28283876e-01 4.38267231e-01 -4.74243671e-01 -3.67175564e-02
1.04191922e-01 3.31189513e-01 -4.34470087e-01 1.66343316e-01
7.76507437e-01 -4.84380275e-01 4.94581342e-01 -1.60637230e-01
2.72055604e-02 2.34496444e-01 -1.37172782e+00 1.00128770e+00
-2.60106530e-02 3.81771475e-01 2.67139226e-02 -1.19234693e+00
7.30035603e-01 2.50831127e-01 1.47184327e-01 -2.80323565e-01
-6.30259812e-02 -1.58680640e-02 2.46574968e-01 -2.37083882e-01
1.48582727e-01 -3.26651335e-01 2.34685373e-02 1.98422149e-01
6.49123192e-02 4.80857790e-01 1.45027980e-01 4.29292619e-01
1.34498870e+00 -2.29725555e-01 2.86374807e-01 -9.06745076e-01
2.42039829e-01 -2.15445220e-01 7.45678544e-01 9.76695538e-01
3.56174648e-01 5.39267778e-01 1.18586004e+00 6.38553947e-02
-1.06135666e+00 -1.35316718e+00 -7.20918715e-01 4.68029857e-01
2.37886921e-01 -4.00498897e-01 -6.63477421e-01 -7.28735626e-01
3.26973200e-01 5.20279050e-01 -9.31770563e-01 -2.57124871e-01
1.00092657e-01 -1.05804980e+00 2.74055570e-01 1.70050308e-01
-1.56423911e-01 -5.72401822e-01 -2.11366341e-02 -1.23378530e-01
-2.01735646e-02 -9.65556502e-01 -1.74484953e-01 1.69641674e-01
-9.74530876e-01 -1.58759105e+00 -1.95272684e-01 -4.45989013e-01
1.00750279e+00 -9.66383889e-02 1.37989497e+00 2.32612759e-01
-2.60444403e-01 2.68158197e-01 -2.58358777e-01 -1.06678218e-01
-3.87615085e-01 -1.08522527e-01 2.49320567e-02 2.31489450e-01
1.39944702e-01 -8.90589476e-01 -2.84890324e-01 3.85635138e-01
-7.29411602e-01 -3.82436961e-02 5.71616948e-01 9.74375725e-01
4.75713700e-01 5.07844687e-01 6.86058700e-01 -1.26618814e+00
4.35020685e-01 -7.54879296e-01 -8.27995777e-01 4.66982126e-01
-8.16077292e-01 3.12716067e-01 5.09720862e-01 -3.21680158e-01
-6.77046657e-01 -1.67211190e-01 1.74015567e-01 -1.56287849e-02
2.91653238e-02 6.19381726e-01 -3.73949289e-01 -4.15992513e-02
3.53022337e-01 -3.75898063e-01 -2.15461612e-01 -5.04736483e-01
1.10588148e-01 1.77762389e-01 3.88689786e-01 -6.05317175e-01
7.82317638e-01 4.96452689e-01 6.60057962e-01 -5.99466264e-01
-7.03263521e-01 -1.48159519e-01 -9.16556776e-01 -3.37368771e-02
4.67856020e-01 -5.70019960e-01 -8.25973332e-01 -1.37754992e-01
-9.55646873e-01 -5.73647395e-02 -2.59339929e-01 5.96381545e-01
-3.35437894e-01 4.19442415e-01 -5.19936681e-01 -8.20126176e-01
1.72425538e-01 -7.01369107e-01 7.22347140e-01 -1.25714213e-01
-9.52776894e-02 -1.29034889e+00 2.53454894e-01 -1.46848395e-01
-4.09758203e-02 2.70428270e-01 1.28689432e+00 -5.85992932e-01
-4.71588343e-01 -3.26583713e-01 -2.10305810e-01 -1.59896359e-01
-2.14945301e-02 1.03712186e-01 -7.68529832e-01 -2.51795202e-01
-1.40820578e-01 3.26393932e-01 8.60108793e-01 5.42137802e-01
1.03509319e+00 -3.18118215e-01 -6.65921867e-01 2.71002352e-01
1.54673088e+00 -4.34768289e-01 6.83874667e-01 -2.66454577e-01
4.78471905e-01 7.31004417e-01 3.47152531e-01 4.10134047e-01
1.85597092e-01 9.47469652e-01 5.02782047e-01 -7.26128370e-02
1.58088937e-01 -5.37598133e-01 1.44090587e-02 4.91674781e-01
3.83954383e-02 -4.03926373e-01 -6.78126395e-01 2.55635470e-01
-1.83486581e+00 -1.01247954e+00 -8.82773578e-01 2.84681821e+00
7.89803684e-01 1.82531744e-01 1.68628976e-01 1.77770451e-01
1.22463214e+00 -4.86573517e-01 -1.11119017e-01 3.03153209e-02
-7.31324926e-02 7.64656663e-02 5.92468262e-01 7.68950224e-01
-7.43745208e-01 2.30001807e-01 6.97676992e+00 6.44829631e-01
-6.49874151e-01 7.47243538e-02 6.66850686e-01 3.13505709e-01
-6.08692765e-01 4.87773120e-01 -7.60081291e-01 6.86650038e-01
9.39253688e-01 -2.10610360e-01 1.91332400e-01 4.16598439e-01
-3.20431218e-03 -2.33738720e-01 -1.44589496e+00 5.13090670e-01
-2.08267257e-01 -9.15568650e-01 -1.05157726e-01 4.65490103e-01
4.92534727e-01 -1.81803629e-01 -3.10035765e-01 -1.65904433e-01
4.56791341e-01 -9.51470673e-01 4.09267604e-01 9.12293971e-01
6.18306160e-01 -8.32059383e-01 8.42846811e-01 4.13997531e-01
-9.02227879e-01 1.32272914e-01 -4.41007227e-01 -1.89496651e-01
1.05222851e-01 1.42072487e+00 -6.70486569e-01 7.86975861e-01
4.04338747e-01 5.23361683e-01 -5.20735979e-01 1.22055697e+00
-1.74681470e-01 8.42885017e-01 -4.09031659e-01 6.33134395e-02
-3.81780386e-01 -4.79350954e-01 6.40029073e-01 1.03996730e+00
4.49162304e-01 -2.75201742e-02 -4.30649072e-02 9.89004970e-01
-1.62231140e-02 1.02533154e-01 -8.04608941e-01 2.73304969e-01
4.29196388e-01 1.26799595e+00 -1.05324304e+00 -6.33913279e-02
-3.50936085e-01 8.41770351e-01 4.44646984e-01 1.79634973e-01
-7.14178801e-01 -5.97019196e-02 3.25093687e-01 5.90812206e-01
2.35291272e-01 -2.99011953e-02 -1.73068177e-02 -1.12495387e+00
7.84162283e-02 -2.50928253e-01 5.68059564e-01 -3.84619445e-01
-1.83432806e+00 4.04798239e-02 1.61470965e-01 -9.92184281e-01
3.37436423e-02 -6.16593301e-01 -8.01454186e-01 8.31111908e-01
-1.11938810e+00 -6.48184478e-01 -1.80376321e-01 1.98530257e-01
-2.69896656e-01 2.72378206e-01 7.12497413e-01 4.39624369e-01
-4.37973708e-01 5.71747541e-01 8.57545957e-02 6.77209273e-02
6.18334353e-01 -1.42690647e+00 3.11632901e-01 6.77769303e-01
3.18987399e-01 6.70041978e-01 9.33997035e-01 -8.65146637e-01
-1.30967069e+00 -8.57663214e-01 9.06303346e-01 -4.46640104e-01
8.45049202e-01 -5.99452078e-01 -8.53633106e-01 7.54983842e-01
-2.95522690e-01 2.88542211e-01 6.50722444e-01 7.08586395e-01
-2.88967729e-01 -6.44539371e-02 -1.07802486e+00 4.38176364e-01
1.27878416e+00 -2.79446423e-01 -2.49030545e-01 2.38496870e-01
2.43133441e-01 1.77634746e-01 -9.51687157e-01 5.22217572e-01
5.04160821e-01 -1.05623543e+00 8.74592423e-01 -5.51095724e-01
2.94105768e-01 -2.52190620e-01 3.17669362e-02 -1.21432471e+00
-4.55710530e-01 -3.53898227e-01 1.48751244e-01 1.52323866e+00
5.90983093e-01 -7.07709312e-01 6.62628174e-01 4.14600104e-01
2.79150605e-01 -5.56289673e-01 -1.16694963e+00 -7.83654332e-01
2.48636119e-02 -2.60716617e-01 2.58757979e-01 1.29849398e+00
1.42334908e-01 6.84663475e-01 -4.12832469e-01 4.76039499e-01
9.53523338e-01 9.59611386e-02 5.03971577e-01 -2.03436804e+00
-7.12083399e-01 -3.39382350e-01 -8.43015492e-01 -6.90006554e-01
2.56065160e-01 -1.18361545e+00 -1.76357672e-01 -1.22036207e+00
5.87156534e-01 -7.58857667e-01 -1.18589513e-02 9.54595879e-02
-8.12449977e-02 1.10004947e-01 -4.82852012e-01 -6.62100613e-02
-2.63818562e-01 3.68428648e-01 8.23119223e-01 3.85619067e-02
1.41567379e-01 3.51205438e-01 -7.04305708e-01 6.05348647e-01
5.46824217e-01 -7.87974358e-01 -4.48722780e-01 8.54303241e-02
7.51407027e-01 4.22611773e-01 6.48221612e-01 -6.22411072e-01
1.39574632e-01 -4.92545823e-03 2.31317610e-01 -2.80039936e-01
3.38809229e-02 -9.28821266e-01 6.05342686e-01 4.10175681e-01
-4.01593506e-01 8.75557065e-02 -6.40995950e-02 9.61753666e-01
1.15511149e-01 -6.76977932e-01 3.80941689e-01 1.81784704e-01
-4.26408313e-02 1.76982313e-01 7.48688951e-02 1.38548026e-02
8.91542733e-01 2.38231093e-01 -2.54407197e-01 -4.34226781e-01
-9.25548553e-01 5.17584682e-02 4.33151633e-01 6.11044206e-02
1.99482635e-01 -1.28233349e+00 -8.14757228e-01 6.23063967e-02
1.77633137e-01 -5.74596345e-01 8.74754936e-02 1.08329153e+00
3.97612862e-02 8.89286585e-03 1.29915729e-01 -5.63757002e-01
-1.30177093e+00 6.25578344e-01 1.62230566e-01 -6.91865236e-02
-7.26825595e-01 4.69567895e-01 3.57206136e-01 -2.08021224e-01
4.79888991e-02 -1.38504086e-02 -1.81757256e-01 2.53142007e-02
2.47900575e-01 3.95804316e-01 1.76908243e-02 -3.53730559e-01
-2.04735801e-01 5.05462587e-01 1.95424587e-01 1.17588118e-02
1.10601580e+00 -1.44303918e-01 -4.60373133e-01 7.89498866e-01
9.75950539e-01 4.37355697e-01 -8.93767536e-01 -7.78620094e-02
1.78028956e-01 -5.30039847e-01 -9.90199074e-02 -6.48174584e-01
-9.65816021e-01 6.70822322e-01 4.75188524e-01 6.93606079e-01
5.96141458e-01 2.77070105e-01 -1.92935601e-01 1.06128566e-01
6.06957912e-01 -7.25158215e-01 -4.30793703e-01 9.74560380e-02
6.28952742e-01 -1.16514313e+00 1.95866913e-01 -1.00578296e+00
-7.47086927e-02 1.05947471e+00 3.04082572e-01 -2.81865597e-01
9.68107462e-01 3.94137740e-01 -6.31579697e-01 -4.13565338e-01
-7.37632811e-01 -2.40285378e-02 3.75391543e-01 4.36520487e-01
5.08764327e-01 1.26458868e-01 -8.08566809e-01 5.43980598e-01
-1.16262451e-01 -2.48666719e-01 6.89645588e-01 2.65710682e-01
-1.46467343e-01 -1.16417253e+00 -1.33453071e-01 6.65963948e-01
-6.08323991e-01 -2.02807397e-01 -4.23382133e-01 7.66140938e-01
-1.28255725e-01 8.52878332e-01 3.84727865e-01 -1.67702720e-01
1.51262075e-01 -8.54418874e-02 5.33942163e-01 -4.74193543e-01
-7.35580847e-02 -2.02144757e-01 3.50251138e-01 -7.87632167e-02
-3.26684505e-01 -6.69290245e-01 -9.23730016e-01 -4.86846179e-01
-7.46612012e-01 5.82107306e-01 3.96004826e-01 9.47516739e-01
2.79845506e-01 4.12557453e-01 5.87089777e-01 -3.54986459e-01
-3.85864019e-01 -8.26685309e-01 -9.60323632e-01 3.11975718e-01
1.04715034e-01 -9.75425780e-01 -7.09758878e-01 -2.92142928e-01] | [7.212117671966553, 5.178137302398682] |
432e8a7f-f8fc-43a5-8920-ea80dada2eb2 | srg-snippet-relatedness-based-temporal-action | 1911.11306 | null | https://arxiv.org/abs/1911.11306v2 | https://arxiv.org/pdf/1911.11306v2.pdf | SRG: Snippet Relatedness-based Temporal Action Proposal Generator | Recent temporal action proposal generation approaches have suggested integrating segment- and snippet score-based methodologies to produce proposals with high recall and accurate boundaries. In this paper, different from such a hybrid strategy, we focus on the potential of the snippet score-based approach. Specifically, we propose a new snippet score-based method, named Snippet Relatedness-based Generator (SRG), with a novel concept of "snippet relatedness". Snippet relatedness represents which snippets are related to a specific action instance. To effectively learn this snippet relatedness, we present "pyramid non-local operations" for locally and globally capturing long-range dependencies among snippets. By employing these components, SRG first produces a 2D relatedness score map that enables the generation of various temporal intervals reliably covering most action instances with high overlap. Then, SRG evaluates the action confidence scores of these temporal intervals and refines their boundaries to obtain temporal action proposals. On THUMOS-14 and ActivityNet-1.3 datasets, SRG outperforms state-of-the-art methods for temporal action proposal generation. Furthermore, compared to competing proposal generators, SRG leads to significant improvements in temporal action detection. | ['Jinyoung Moon', 'Hyunjun Eun', 'Sumin Lee', 'Changick Kim', 'Jongyoul Park', 'Chanho Jung'] | 2019-11-26 | null | null | null | null | ['temporal-action-proposal-generation'] | ['computer-vision'] | [ 4.40684050e-01 -1.58390313e-01 -5.13403893e-01 -1.51207030e-01
-1.10268652e+00 -2.63338059e-01 8.43290150e-01 3.89431655e-01
-2.48000324e-01 7.88913310e-01 5.85732341e-01 2.40874514e-01
-5.20393610e-01 -7.38552034e-01 -2.40070686e-01 -4.54236001e-01
-3.06949794e-01 4.05664146e-01 9.10804093e-01 -1.74799457e-01
6.44633532e-01 3.76976073e-01 -1.72679102e+00 5.56567132e-01
9.44363415e-01 8.84067714e-01 2.35993192e-01 4.19988573e-01
-9.45094153e-02 6.83155239e-01 -9.54524577e-01 1.31387055e-01
8.77606496e-02 -8.00469756e-01 -7.77257085e-01 -7.43347928e-02
1.14797398e-01 -2.22931877e-01 2.34009065e-02 5.60843468e-01
5.08702934e-01 5.31761885e-01 6.79352701e-01 -1.19457114e+00
7.48760030e-02 7.94375718e-01 -6.59210801e-01 5.31233609e-01
8.44575286e-01 1.35619298e-01 1.14922810e+00 -7.61627853e-01
8.02751184e-01 1.07469606e+00 6.41883075e-01 3.83735746e-01
-1.05803430e+00 -4.98407811e-01 3.58488232e-01 4.15608317e-01
-1.47679853e+00 -2.63826195e-02 5.80756783e-01 -3.80359828e-01
1.25570393e+00 1.78698108e-01 9.12853599e-01 1.19637871e+00
2.41470784e-01 1.13404310e+00 8.44849467e-01 -3.34115982e-01
3.85349661e-01 -5.33499718e-01 -2.53076494e-01 1.55766174e-01
-6.53713346e-02 2.28627458e-01 -9.50985789e-01 -6.38616681e-02
8.39504004e-01 -1.29942428e-02 7.36119375e-02 -6.52485266e-02
-1.50891256e+00 5.73021650e-01 2.18589649e-01 6.41667545e-01
-7.39199579e-01 3.62252325e-01 4.53913152e-01 -2.82338738e-01
3.58354449e-01 4.19579089e-01 -1.38515800e-01 -5.02477586e-01
-1.26164663e+00 7.66100824e-01 4.14699018e-01 8.88054371e-01
5.18407285e-01 -1.76472396e-01 -1.21327841e+00 6.17433012e-01
1.50494918e-01 1.18264286e-02 5.02118230e-01 -8.54549110e-01
7.59244978e-01 9.58046556e-01 2.26362079e-01 -8.17659795e-01
-3.23608577e-01 -2.98990160e-01 -4.28210378e-01 6.11637272e-02
1.75982341e-01 2.15221494e-01 -9.07998919e-01 1.53985262e+00
3.36045772e-01 4.76222992e-01 -1.68459788e-01 5.87227345e-01
4.50957954e-01 6.74958825e-01 4.01701480e-01 -3.35120797e-01
1.10760081e+00 -9.24288809e-01 -7.07292080e-01 -1.05182528e-01
6.42978072e-01 -5.26120603e-01 7.16443956e-01 3.93579125e-01
-1.03410828e+00 -7.24159479e-01 -9.47830021e-01 5.91874480e-01
-2.33752072e-01 1.92370504e-01 4.51389849e-01 2.46990204e-01
-7.93686211e-01 8.32078755e-01 -8.87266695e-01 -4.14065540e-01
4.52888250e-01 1.02337763e-01 -7.39176199e-02 2.33623788e-01
-1.12641060e+00 9.77210164e-01 9.66827035e-01 -1.86960995e-01
-1.08756959e+00 -5.35684168e-01 -7.50549674e-01 -9.26350355e-02
7.40619481e-01 -4.26111370e-01 1.44829226e+00 -5.52170038e-01
-1.23603117e+00 1.61973119e-01 -1.86604694e-01 -9.03534174e-01
4.12750781e-01 -2.06754535e-01 -5.33587217e-01 3.28806996e-01
4.36756283e-01 8.96925509e-01 6.54215634e-01 -6.68931067e-01
-1.14923751e+00 -4.42108326e-02 -1.80794612e-01 4.90538776e-01
-8.14052448e-02 3.00636172e-01 -5.22293091e-01 -9.35524285e-01
2.31546938e-01 -6.20058119e-01 -4.51845020e-01 -3.61908019e-01
-2.99646169e-01 -7.68567383e-01 6.28025472e-01 -3.35674554e-01
1.94110835e+00 -1.91095638e+00 -3.56685445e-02 2.97450513e-01
-2.53914893e-02 4.35169250e-01 -3.00523967e-01 6.27385378e-01
7.22501501e-02 -1.05514847e-01 -2.58892447e-01 5.73722348e-02
-8.56293738e-03 1.06420785e-01 -7.60043338e-02 -6.28668889e-02
3.35289359e-01 9.52564061e-01 -1.30578327e+00 -6.01938426e-01
4.64947104e-01 2.76806235e-01 -4.26865816e-01 7.66082704e-02
-4.02019441e-01 5.01240253e-01 -7.23271608e-01 6.35165632e-01
1.31615594e-01 1.13749854e-01 3.83464694e-02 -2.04485998e-01
-1.65834248e-01 5.16635001e-01 -1.39380181e+00 1.83092189e+00
-5.63477799e-02 3.68603528e-01 -8.18087161e-01 -8.06811512e-01
1.31449091e+00 3.54013830e-01 9.64777648e-01 -8.79270256e-01
-2.17702970e-01 2.17686385e-01 -2.51311034e-01 -3.94684613e-01
7.41423607e-01 1.46122454e-02 -2.89948791e-01 4.52299148e-01
-4.13055941e-02 -2.91397423e-02 7.85246611e-01 2.81513572e-01
1.54628921e+00 9.17459905e-01 8.11726689e-01 1.93408445e-01
6.12004340e-01 6.79562939e-03 6.57559335e-01 8.25929284e-01
-3.86140674e-01 7.68797994e-01 4.54043865e-01 -3.74632180e-01
-7.61831582e-01 -1.22451341e+00 1.96713433e-01 1.01121914e+00
2.06992865e-01 -9.05731082e-01 -5.78357637e-01 -1.10261965e+00
-2.39940912e-01 8.12593818e-01 -7.07655787e-01 -1.11066565e-01
-7.32787490e-01 -4.54930335e-01 5.33320427e-01 8.85010183e-01
5.84097207e-01 -1.58668399e+00 -9.13482487e-01 6.97111070e-01
-4.00084257e-01 -7.82405376e-01 -6.85627639e-01 1.24879868e-03
-1.08600271e+00 -1.02593029e+00 -6.72291815e-01 -1.64586499e-01
4.27410364e-01 2.81866007e-02 1.16124988e+00 -3.21378559e-01
-3.01808685e-01 2.26706192e-01 -9.76088464e-01 -2.61624426e-01
-2.92277575e-01 3.87026765e-03 -2.38570645e-01 -6.67858683e-03
5.41774809e-01 -6.23826921e-01 -7.67410457e-01 6.74845934e-01
-8.50862801e-01 3.47857587e-02 8.33542645e-01 3.63955259e-01
1.01125383e+00 7.96016678e-02 8.99115503e-01 -3.24279159e-01
7.46320665e-01 -4.22686905e-01 -3.28173459e-01 2.83689976e-01
-6.25294983e-01 1.08204827e-01 1.79430708e-01 -5.68327963e-01
-1.10898519e+00 1.03134930e-01 -1.09040076e-02 -2.82431394e-01
-3.19663346e-01 5.03341496e-01 2.18526900e-01 6.23553932e-01
9.88755107e-01 3.73944223e-01 -3.29193205e-01 -4.68428284e-01
3.91574323e-01 2.17785493e-01 6.10124469e-01 -5.34599781e-01
4.05143470e-01 3.23363960e-01 -2.32701749e-01 -3.16719532e-01
-6.85742557e-01 -8.24419677e-01 -5.89955389e-01 -6.43466532e-01
1.01726639e+00 -6.70822918e-01 -3.85492593e-01 2.10072309e-01
-1.12308955e+00 -6.77564964e-02 -7.97018826e-01 6.89133286e-01
-7.72238553e-01 2.85433501e-01 -2.29991421e-01 -9.12027657e-01
-2.74516702e-01 -8.82475376e-01 1.29305804e+00 3.10164273e-01
-7.10574210e-01 -5.42354167e-01 3.50663662e-01 1.97105318e-01
1.84896618e-01 5.45448899e-01 2.16046318e-01 -9.25291419e-01
-6.74598694e-01 -4.20591295e-01 -1.02699064e-02 2.16161698e-01
1.68446168e-01 -7.48191997e-02 -3.69309992e-01 9.77367256e-03
-5.30425787e-01 -6.64003333e-03 7.79978931e-01 6.60786092e-01
9.87779975e-01 -1.94870546e-01 -5.08108854e-01 4.55698110e-02
1.07102621e+00 5.55293441e-01 9.24887180e-01 3.69779557e-01
3.43782008e-01 5.89297831e-01 1.49893928e+00 8.66787493e-01
1.82081074e-01 1.16413128e+00 4.41700101e-01 2.09278718e-01
-8.96137208e-02 -4.38120693e-01 4.89656866e-01 1.09719232e-01
-4.71342623e-01 -2.64849722e-01 -6.51440680e-01 7.53499091e-01
-2.26003885e+00 -1.39952266e+00 -1.00796714e-01 2.21508265e+00
6.04048252e-01 3.87318164e-01 6.36263311e-01 2.24378973e-01
8.50680470e-01 3.39381456e-01 -3.26102972e-01 -1.33374408e-01
1.66626379e-01 2.73359209e-01 2.47311771e-01 -6.41864836e-02
-1.02210712e+00 1.02066720e+00 5.72468424e+00 1.34213400e+00
-4.61061865e-01 8.14791322e-02 2.07002610e-01 -1.23803832e-01
-1.20928206e-01 6.35715872e-02 -1.12483943e+00 5.18675864e-01
7.38834918e-01 -2.73789495e-01 -1.58647239e-01 7.67654955e-01
7.16068625e-01 -5.25594831e-01 -1.03788686e+00 6.35794044e-01
-2.07228400e-02 -1.33514631e+00 9.04185250e-02 -1.58960700e-01
8.58147204e-01 -2.36578643e-01 -3.42528552e-01 3.97955686e-01
2.59149045e-01 -6.78366661e-01 8.09599578e-01 6.32401288e-01
5.37849903e-01 -7.78972208e-01 6.28509700e-01 1.68195665e-01
-1.82675302e+00 -1.47602901e-01 1.13162398e-01 2.35077992e-01
6.90556467e-01 4.57665890e-01 -1.09103155e+00 9.09726083e-01
5.42273283e-01 1.03296232e+00 -5.37820995e-01 1.51702571e+00
-6.13153934e-01 5.26402891e-01 -2.21742511e-01 -1.29426017e-01
4.42832351e-01 -4.38998155e-02 6.87236905e-01 1.24720204e+00
7.12292969e-01 1.04599498e-01 1.98959857e-01 8.42804968e-01
5.49442470e-01 1.29837859e-02 -4.94501412e-01 -7.50649273e-02
7.06340194e-01 1.04419065e+00 -1.02847099e+00 -5.72837174e-01
-5.24366610e-02 7.68169820e-01 -1.24111958e-01 1.15156747e-01
-1.19460845e+00 -2.97354579e-01 3.03264558e-01 2.28796557e-01
5.45880795e-01 -8.25870186e-02 1.19893877e-02 -5.75721323e-01
1.32712498e-01 -7.15118289e-01 7.92462766e-01 -7.79860497e-01
-8.14793289e-01 5.99908054e-01 6.43623829e-01 -1.99260187e+00
-6.32919550e-01 7.59619325e-02 -8.75320315e-01 6.18215263e-01
-8.54609311e-01 -1.06011975e+00 -3.18365425e-01 5.18938959e-01
1.05095148e+00 3.94142349e-04 5.45141578e-01 3.72212417e-02
-4.13541377e-01 3.21392536e-01 -4.81572330e-01 -2.56313860e-01
4.49899852e-01 -1.07268667e+00 6.17147207e-01 9.15399194e-01
3.36711586e-01 2.78210670e-01 5.74962556e-01 -1.15471005e+00
-3.60933930e-01 -1.26375759e+00 1.16630483e+00 -6.50247037e-01
4.79186773e-01 1.69209465e-01 -7.41309524e-01 4.33745801e-01
-2.20954090e-01 -2.78924674e-01 3.18640620e-01 -1.98422328e-01
-1.14119940e-01 -2.39643425e-01 -9.05458868e-01 7.02688575e-01
1.40896630e+00 -8.00807625e-02 -9.90315139e-01 3.44756812e-01
5.67970634e-01 -2.14278087e-01 -7.63736308e-01 6.87235475e-01
5.23422241e-01 -1.29561055e+00 1.10003495e+00 -1.40050977e-01
3.59812856e-01 -6.56633556e-01 1.27706960e-01 -9.63664830e-01
-2.56339252e-01 -7.82066703e-01 -3.70871395e-01 1.17025149e+00
4.68140632e-01 -1.57434881e-01 7.35118389e-01 8.38251971e-03
-4.86358881e-01 -8.19957852e-01 -9.86665308e-01 -1.15423501e+00
-8.49293113e-01 -7.53519833e-01 6.93196476e-01 4.11707193e-01
1.46656469e-01 6.89169765e-02 -4.24060673e-01 -3.41579378e-01
3.27553540e-01 -1.97694242e-01 5.62248588e-01 -8.60758841e-01
-3.66622806e-01 -7.64641762e-01 -5.21346450e-01 -9.36000288e-01
-3.38480204e-01 -6.62158906e-01 3.92415136e-01 -1.86689007e+00
3.26476023e-02 -2.26613998e-01 -5.61985373e-01 6.89597964e-01
-2.27508947e-01 3.32816303e-01 1.06471740e-01 1.10569045e-01
-1.03660762e+00 5.18908858e-01 1.12703240e+00 4.89201769e-02
-7.85248101e-01 3.91692579e-01 -2.26851687e-01 5.67996442e-01
6.23375714e-01 -5.57248354e-01 -5.80270648e-01 2.58559823e-01
5.01791947e-02 1.95645988e-01 2.43023500e-01 -1.43793285e+00
1.51396722e-01 -4.30287063e-01 8.93211365e-02 -1.27335691e+00
1.74242839e-01 -4.79499459e-01 4.17581946e-01 4.56641376e-01
-5.44667363e-01 -3.52591984e-02 2.35741455e-02 7.88185954e-01
-2.98237234e-01 -2.77303994e-01 4.00780439e-01 -7.38561600e-02
-1.10942268e+00 2.37716645e-01 -5.55743992e-01 -1.15900479e-01
1.37657797e+00 -6.52043819e-01 -4.83273342e-02 -2.42810659e-02
-8.06197584e-01 1.31800741e-01 -8.87912512e-02 5.09293914e-01
8.18532526e-01 -1.50786781e+00 -7.25611150e-01 -1.26518548e-01
4.77816939e-01 -1.04762644e-01 3.32242757e-01 1.10987222e+00
-1.01005189e-01 3.61207217e-01 -2.65346646e-01 -6.71919584e-01
-1.24761355e+00 2.46102899e-01 -1.20237982e-02 -7.42244065e-01
-7.83400416e-01 7.97521412e-01 -3.68661657e-02 1.69931620e-01
3.70110720e-01 -5.00919163e-01 -6.12082005e-01 2.64933825e-01
6.15190804e-01 7.00374305e-01 -5.08648083e-02 -4.33841586e-01
-5.26910484e-01 4.92466092e-01 7.72060901e-02 -4.59841788e-01
1.03633380e+00 2.45150834e-01 3.94121796e-01 1.02150552e-01
4.76655394e-01 -2.86043912e-01 -1.44137061e+00 -9.96822938e-02
4.86758113e-01 -4.27697688e-01 -3.58684391e-01 -9.55042243e-01
-6.86838567e-01 2.96760708e-01 4.44213450e-01 1.15606628e-01
1.46906340e+00 2.03948215e-01 7.49578416e-01 -5.91573715e-02
4.90676582e-01 -1.38924563e+00 6.90149426e-01 3.39092016e-01
9.73866761e-01 -8.38678539e-01 1.80633202e-01 -2.75475800e-01
-8.99022281e-01 7.38452494e-01 8.39770138e-01 5.43509014e-02
3.65442000e-02 -1.37567222e-01 -5.28693914e-01 -2.58630186e-01
-6.85744524e-01 -5.90235054e-01 7.22567797e-01 6.70625150e-01
3.28282744e-01 1.15365483e-01 -9.72471893e-01 4.73438799e-01
2.64052242e-01 2.98575014e-01 1.59690052e-01 1.12523222e+00
-6.59402788e-01 -1.34314227e+00 -2.10154921e-01 4.98652697e-01
-7.14824349e-02 4.69093844e-02 -2.37497970e-01 7.81298518e-01
3.88863057e-01 7.91327715e-01 2.87688263e-02 -7.13063776e-01
5.89225054e-01 -1.20311163e-01 4.51543719e-01 -8.84654522e-01
-8.01713228e-01 3.79323065e-01 3.86554658e-01 -9.96333361e-01
-7.32019901e-01 -8.23221505e-01 -1.37309539e+00 2.59150535e-01
-2.77145147e-01 2.02309296e-01 4.19361591e-01 1.10017002e+00
3.96575332e-01 7.52599895e-01 4.86996055e-01 -1.01443911e+00
-4.35603559e-01 -1.24585390e+00 -3.58388990e-01 4.02129591e-01
-3.30635667e-01 -7.11183786e-01 -1.21794216e-01 -2.57711462e-03] | [8.338515281677246, 0.44218024611473083] |
6923d9ff-824c-4b30-87a9-3adc33fae96b | cic-fbk-approach-to-native-language | null | null | https://aclanthology.org/W17-5042 | https://aclanthology.org/W17-5042.pdf | CIC-FBK Approach to Native Language Identification | We present the CIC-FBK system, which took part in the Native Language Identification (NLI) Shared Task 2017. Our approach combines features commonly used in previous NLI research, i.e., word n-grams, lemma n-grams, part-of-speech n-grams, and function words, with recently introduced character n-grams from misspelled words, and features that are novel in this task, such as typed character n-grams, and syntactic n-grams of words and of syntactic relation tags. We use log-entropy weighting scheme and perform classification using the Support Vector Machines (SVM) algorithm. Our system achieved 0.8808 macro-averaged F1-score and shared the 1st rank in the NLI Shared Task 2017 scoring. | ['Carlo Strapparava', 'Lingzhen Chen', 'Ilia Markov', 'Grigori Sidorov'] | 2017-09-01 | null | null | null | ws-2017-9 | ['native-language-identification'] | ['natural-language-processing'] | [ 1.36224300e-01 2.42091957e-02 -4.18325663e-01 -3.05684179e-01
-9.08671856e-01 -9.06317353e-01 7.11000085e-01 4.72213805e-01
-8.71987879e-01 8.75247598e-01 7.32177734e-01 -5.73380291e-01
-1.17885545e-01 -5.56456208e-01 -3.83516669e-01 -2.76127040e-01
-2.40355149e-01 4.92211968e-01 1.06417410e-01 -8.68041888e-02
4.24428821e-01 2.50874490e-01 -1.19939041e+00 6.56211734e-01
7.99505055e-01 9.92503405e-01 5.44724520e-03 7.80212402e-01
-4.84370530e-01 4.79148448e-01 -4.51067954e-01 -7.10099399e-01
-1.48763761e-01 7.32584111e-03 -9.41209316e-01 -9.75807071e-01
3.25547934e-01 1.26416218e-02 -2.70543754e-01 1.00239134e+00
1.94701627e-01 1.45535454e-01 1.03548157e+00 -7.23788619e-01
-5.59878945e-01 1.11084402e+00 -8.00331011e-02 4.68808621e-01
7.16907978e-01 -1.05782636e-01 1.51677799e+00 -1.11319661e+00
6.26628876e-01 1.37666750e+00 9.08140659e-01 6.44902050e-01
-1.12486458e+00 -8.64331126e-01 -3.24692696e-01 6.84801815e-03
-1.14695692e+00 -2.85726130e-01 2.62762040e-01 -7.07165241e-01
1.67085910e+00 2.32243091e-01 2.75464714e-01 1.33499920e+00
2.72183686e-01 9.69575882e-01 1.04085839e+00 -9.17621613e-01
-4.88217883e-02 2.70765901e-01 7.84814417e-01 5.13772666e-01
4.16111887e-01 2.83842891e-01 -6.95886374e-01 -4.49903250e-01
4.32770178e-02 -3.18527997e-01 1.09271280e-01 2.84439892e-01
-1.55083394e+00 9.11030412e-01 -3.54725718e-01 8.10469329e-01
-8.01512673e-02 -6.35599792e-02 6.86434269e-01 1.45922452e-01
6.07905388e-01 5.99684894e-01 -1.19151211e+00 -5.07878542e-01
-7.21504569e-01 6.98697194e-02 1.16047108e+00 8.26605916e-01
6.72677696e-01 -1.77133933e-01 -5.27691722e-01 1.07470393e+00
1.15399130e-01 4.48815823e-01 1.34046757e+00 -3.94677013e-01
5.34661353e-01 6.38933003e-01 -3.86178225e-01 -4.85854775e-01
-3.10268790e-01 -7.32518435e-02 -6.93051696e-01 -2.59395003e-01
4.40798581e-01 -2.27749377e-01 -7.38113999e-01 1.56407428e+00
-1.26038015e-01 -2.55908161e-01 3.26283753e-01 2.42412060e-01
9.38981354e-01 6.23191774e-01 4.63591486e-01 -4.62728217e-02
1.51920664e+00 -9.48535740e-01 -4.99069691e-01 -1.44150004e-01
1.08155274e+00 -9.04138744e-01 8.80423665e-01 1.77122459e-01
-8.28728437e-01 -5.16300917e-01 -6.04922771e-01 1.11084305e-01
-1.06879008e+00 3.62229407e-01 7.14718103e-01 9.01000619e-01
-9.23333883e-01 7.60534465e-01 -1.77216843e-01 -5.10769308e-01
3.64410467e-02 4.24355716e-01 -7.35024095e-01 2.56484985e-01
-1.52684689e+00 8.99242580e-01 8.55046570e-01 -6.79661155e-01
-4.79748577e-01 -8.79860818e-01 -1.03236187e+00 -2.00876314e-02
1.59133762e-01 4.27435562e-02 1.17593789e+00 -7.25714028e-01
-1.50102937e+00 1.05092084e+00 -3.37436199e-01 -4.81466085e-01
-1.48322314e-01 -2.43514732e-01 -7.20094264e-01 -1.69635311e-01
2.91086175e-02 2.96132505e-01 3.56242448e-01 -5.95518768e-01
-8.72858703e-01 -2.29636446e-01 -2.73829848e-01 -1.17228284e-01
-6.97331011e-01 5.56606293e-01 3.18655133e-01 -5.76679349e-01
-3.11852068e-01 -8.65459204e-01 2.05770776e-01 -7.32137024e-01
-4.01081145e-01 -7.03614235e-01 3.89561504e-01 -1.23038411e+00
1.56404519e+00 -2.44720602e+00 -1.98834971e-01 3.48194182e-01
3.97367366e-02 7.43233323e-01 -1.36675566e-01 4.15441453e-01
-1.78262919e-01 5.42553961e-01 1.82993948e-01 -1.32465228e-01
5.82805835e-02 7.70527944e-02 -2.03642651e-01 -5.29474057e-02
-6.98454902e-02 1.10163856e+00 -1.01821828e+00 -3.06502491e-01
5.10123633e-02 1.21396549e-01 -1.51080251e-01 1.32745013e-01
-3.37343439e-02 -5.71312681e-02 -1.01161659e-01 5.05694151e-01
5.07503301e-02 2.15305209e-01 5.47773913e-02 1.34494016e-02
-3.10516536e-01 1.13534319e+00 -7.45601237e-01 1.12105227e+00
-6.22247040e-01 3.86507273e-01 -2.72264451e-01 -9.15510178e-01
8.29052687e-01 4.55982417e-01 1.87980626e-02 -3.05574000e-01
1.68275371e-01 5.80364287e-01 2.97311425e-01 -1.79836854e-01
3.76738250e-01 1.34610236e-01 -4.02268231e-01 1.50980413e-01
4.04914290e-01 2.45130077e-01 2.42756784e-01 1.48901835e-01
1.22805512e+00 -1.68936640e-01 1.18268263e+00 -5.37912309e-01
8.35444152e-01 -4.35595587e-02 2.27927297e-01 6.86651826e-01
-1.58757210e-01 2.60609865e-01 3.77863944e-01 -2.61843771e-01
-1.11310291e+00 -7.83031702e-01 -1.01604439e-01 1.48101354e+00
-7.22077489e-01 -6.16901815e-01 -5.87849379e-01 -9.75834906e-01
2.59246647e-01 9.20506775e-01 -5.24009168e-01 -1.36665031e-01
-5.92164516e-01 -2.55408198e-01 1.00707090e+00 4.63638395e-01
-9.07193571e-02 -1.32796156e+00 1.33657604e-01 3.87652785e-01
8.59516934e-02 -1.01029921e+00 -6.72487199e-01 5.51511943e-01
-5.61606944e-01 -7.71723211e-01 -7.08404064e-01 -1.20848680e+00
1.82315841e-01 -1.29227161e-01 1.13273108e+00 -2.80234247e-01
-4.89581883e-01 1.58157200e-01 -7.10456491e-01 -2.62955278e-01
-5.97851217e-01 6.37409031e-01 3.20094496e-01 -3.09522986e-01
8.99666846e-01 -2.49637336e-01 4.54803519e-02 -7.32778236e-02
-4.45247591e-01 -3.04875016e-01 7.26879299e-01 1.03310740e+00
2.55387634e-01 -2.25673363e-01 3.84639710e-01 -8.70953083e-01
5.35914123e-01 -2.04490662e-01 -1.39872700e-01 4.75638419e-01
-6.41146004e-01 1.02647938e-01 8.84050369e-01 -9.45899069e-01
-8.31816673e-01 -7.84131065e-02 -3.78387600e-01 3.58736783e-01
-3.68005723e-01 4.87232715e-01 -4.88943219e-01 -3.06198657e-01
3.52689922e-01 4.99479562e-01 -4.29568589e-01 -6.69103801e-01
3.84156674e-01 1.17390645e+00 3.82879764e-01 -6.93415880e-01
6.72137976e-01 -3.86770099e-01 -2.70009249e-01 -8.63479495e-01
-1.19349253e+00 -1.03020096e+00 -8.20805848e-01 2.04941630e-01
7.89379358e-01 -7.67566502e-01 -8.02854121e-01 5.79933405e-01
-1.17134273e+00 -7.23888204e-02 -1.31229579e-01 6.48585021e-01
-7.95183927e-02 6.49684668e-01 -7.00793505e-01 -1.02227783e+00
-7.34996974e-01 -5.96726537e-01 7.21295893e-01 2.14212000e-01
-8.98502767e-01 -1.05435836e+00 1.69566110e-01 2.50557512e-01
5.50841212e-01 -3.11012268e-01 1.34038341e+00 -1.67289376e+00
3.55144233e-01 -3.49636763e-01 -1.83706179e-01 8.20274472e-01
-1.37012869e-01 -2.72564828e-01 -7.40230560e-01 -1.80879869e-02
-6.06371105e-01 -3.94599408e-01 1.06381488e+00 9.83177796e-02
9.28657115e-01 -8.01293731e-01 -5.22324562e-01 5.28311372e-01
1.25230730e+00 2.42655203e-01 4.55490977e-01 2.20310330e-01
6.29668713e-01 6.99728251e-01 2.80189484e-01 2.67080873e-01
2.70869642e-01 4.96711016e-01 -2.97244459e-01 6.52334511e-01
-6.96869269e-02 -5.68432212e-01 6.27732158e-01 1.11678517e+00
1.46416917e-01 -2.61428505e-01 -1.23959446e+00 9.03543651e-01
-1.43604541e+00 -7.84785807e-01 -2.65756965e-01 2.10575390e+00
1.25695050e+00 3.86881381e-01 -8.13785288e-03 2.69354731e-01
6.52624071e-01 -7.00005190e-03 1.07270256e-01 -8.31298351e-01
-2.77365059e-01 7.72110164e-01 9.07348275e-01 7.61103213e-01
-1.24293196e+00 1.41136682e+00 5.87385702e+00 1.49064803e+00
-6.80355549e-01 2.63257474e-01 2.50755548e-01 2.69443244e-01
-2.90794522e-01 8.66605565e-02 -1.52107954e+00 6.97804213e-01
1.33769202e+00 -4.40422595e-01 5.27968049e-01 1.03220856e+00
-6.59929395e-01 -9.99521092e-02 -1.04253495e+00 7.50549376e-01
2.09371313e-01 -9.01042521e-01 1.17016681e-01 1.88500956e-01
5.29912889e-01 8.70877057e-02 -6.51195832e-03 6.60176873e-01
5.13396084e-01 -9.99984622e-01 5.03635406e-01 2.64330029e-01
1.19367719e+00 -6.08625472e-01 9.75941956e-01 3.97502095e-01
-1.14554727e+00 -1.12022795e-02 -2.49528229e-01 -2.30251774e-01
-1.63897440e-01 9.21961725e-01 -8.83288622e-01 3.81397128e-01
3.35329801e-01 4.41724807e-01 -3.08142811e-01 8.42400908e-01
-3.10463190e-01 1.22453249e+00 -2.66638875e-01 -7.76587546e-01
2.77213007e-01 1.82522580e-01 7.05655873e-01 1.85045695e+00
3.16983551e-01 -8.85953456e-02 -2.58203447e-02 1.15761213e-01
5.00004031e-02 5.46496511e-01 -5.03187180e-01 -3.28907698e-01
4.49775368e-01 1.31722796e+00 -1.85823113e-01 -6.64812028e-01
-3.69096488e-01 1.13541508e+00 4.84987646e-01 -2.68349554e-02
-2.20147952e-01 -8.78139496e-01 1.02661669e+00 -7.93863609e-02
2.35518262e-01 -4.23617810e-01 -3.40180665e-01 -1.12597692e+00
-1.40796587e-01 -6.12025619e-01 4.39499229e-01 -2.69673526e-01
-1.76473868e+00 7.25830376e-01 -8.55834261e-02 -5.39674699e-01
-5.14875948e-01 -1.31469262e+00 -3.59792262e-01 1.20488000e+00
-1.36673474e+00 -1.04926753e+00 2.79584587e-01 9.27881077e-02
3.25023085e-01 -7.71736801e-01 1.20487261e+00 2.01171234e-01
-3.56880128e-01 1.11775720e+00 5.37448049e-01 6.83731735e-01
9.19309437e-01 -1.34412527e+00 7.39900708e-01 2.67247915e-01
3.35381359e-01 7.90853560e-01 2.99862444e-01 -8.14786136e-01
-1.04291177e+00 -9.17238414e-01 2.34087205e+00 -3.03503484e-01
1.06152070e+00 -6.00463927e-01 -6.41933441e-01 5.65187395e-01
-2.68532783e-01 -1.44598708e-01 1.00056756e+00 3.12446386e-01
-6.14257216e-01 8.44053626e-02 -1.16946661e+00 2.46226192e-01
1.13712609e+00 -7.89978981e-01 -1.00755918e+00 3.12125385e-01
9.06881630e-01 -3.48919928e-02 -6.76312864e-01 3.34613234e-01
9.16829050e-01 -2.05012843e-01 1.02360368e+00 -9.39882457e-01
3.84074956e-01 3.40100586e-01 -2.90192217e-01 -1.43719208e+00
-6.49880111e-01 -5.44252992e-01 -7.23076165e-02 1.60057998e+00
9.87449765e-01 -6.66888952e-01 3.74717325e-01 4.52994436e-01
8.79273787e-02 -5.24519622e-01 -1.13814175e+00 -1.20912492e+00
2.63603836e-01 -4.86724794e-01 3.16631436e-01 7.74562120e-01
4.75444853e-01 4.37233895e-01 -2.90564716e-01 -4.75736707e-01
3.80167842e-01 -6.12934411e-01 2.58907527e-01 -1.46663713e+00
-2.90248513e-01 -6.03529513e-01 -5.96391380e-01 -6.89712346e-01
8.38125169e-01 -1.37541878e+00 -9.62948352e-02 -9.73849595e-01
3.20314288e-01 -1.31413847e-01 -3.67163062e-01 8.49970996e-01
-2.03657210e-01 1.08760662e-01 -4.36267257e-02 1.81743933e-03
-5.08743882e-01 1.36040390e-01 3.87957722e-01 -1.01606689e-01
8.27860907e-02 -1.06487095e-01 -5.46596050e-01 8.01855028e-01
9.01337743e-01 -5.86851120e-01 3.73169273e-01 -3.77972573e-02
1.62770569e-01 -5.34720242e-01 -4.19260338e-02 -8.75582695e-01
7.53912926e-02 -1.37057245e-01 4.18607116e-01 -4.60366338e-01
9.98863056e-02 -5.04700541e-01 -3.44138682e-01 6.09162569e-01
-5.20121634e-01 -1.54716402e-01 1.16128549e-01 1.70963407e-01
-1.10293932e-01 -7.60718465e-01 3.92249316e-01 -2.73938268e-01
-5.21153867e-01 6.74050720e-03 -5.71354508e-01 2.10713461e-01
7.65465140e-01 -1.28928274e-01 -1.63078979e-01 1.21533208e-01
-3.57416272e-01 -2.46183261e-01 1.27657428e-01 4.82501149e-01
2.11975962e-01 -1.25933266e+00 -8.10174525e-01 2.44976521e-01
4.32557970e-01 -1.00067699e+00 -3.09393346e-01 2.29661793e-01
-1.73136219e-01 9.95060563e-01 2.48819906e-02 3.18192065e-01
-1.50097489e+00 3.58841509e-01 -3.53577942e-01 -8.86787474e-01
-3.26801062e-01 9.93882358e-01 -1.16141364e-01 -6.03838086e-01
2.77770579e-01 -3.80374342e-01 -4.07668740e-01 2.80423105e-01
9.32149231e-01 4.30487871e-01 1.75434388e-02 -8.28388155e-01
-8.81339431e-01 4.74175513e-01 -1.66635230e-01 -3.52866352e-01
1.02724195e+00 1.54870704e-01 -4.56118137e-01 6.69999242e-01
1.59582567e+00 2.39726499e-01 -5.79825640e-02 -5.70200622e-01
5.16426086e-01 -1.38855562e-01 -2.47475296e-01 -1.20967591e+00
-2.22027227e-01 7.92198539e-01 6.14427701e-02 6.35798201e-02
2.48939753e-01 1.20377123e-01 1.28867698e+00 4.28379714e-01
4.42542344e-01 -1.26926422e+00 -4.25852567e-01 1.37969410e+00
4.04090971e-01 -1.01331925e+00 -3.53533596e-01 -4.20134157e-01
-3.57980490e-01 1.15217352e+00 3.03721845e-01 1.49695337e-01
7.06399202e-01 3.93497676e-01 -2.52669424e-01 3.89181316e-01
-6.66685164e-01 -3.82287860e-01 7.35295415e-01 5.90343952e-01
8.46630752e-01 7.05902576e-01 -8.62337708e-01 1.27850556e+00
-5.71622312e-01 -2.41882041e-01 -2.01780170e-01 4.13378358e-01
-5.36964536e-01 -1.17993641e+00 -6.32360056e-02 9.16703343e-01
-8.44375908e-01 -7.34431386e-01 -5.20877182e-01 3.06690574e-01
2.75275946e-01 8.19052160e-01 -1.40963495e-01 -7.23047912e-01
4.55207415e-02 9.59968030e-01 2.61186182e-01 -8.38158190e-01
-1.13355505e+00 -7.21885383e-01 5.11205733e-01 -2.80623466e-01
2.02964738e-01 -8.27464342e-01 -9.41393554e-01 -1.62384003e-01
-2.23671392e-01 1.56475812e-01 9.98595595e-01 1.00824571e+00
1.08173385e-01 9.13752466e-02 3.49605799e-01 -4.27697271e-01
-8.25266421e-01 -1.45283961e+00 -5.41960478e-01 3.54072630e-01
6.60574660e-02 -2.80755103e-01 -7.75547028e-01 -1.81198224e-01] | [10.437905311584473, 10.461492538452148] |
0f704dac-eef6-4ba1-a66a-ea42a0dc896c | learning-populations-of-parameters | 1709.02707 | null | http://arxiv.org/abs/1709.02707v2 | http://arxiv.org/pdf/1709.02707v2.pdf | Learning Populations of Parameters | Consider the following estimation problem: there are $n$ entities, each with
an unknown parameter $p_i \in [0,1]$, and we observe $n$ independent random
variables, $X_1,\ldots,X_n$, with $X_i \sim $ Binomial$(t, p_i)$. How
accurately can one recover the "histogram" (i.e. cumulative density function)
of the $p_i$'s? While the empirical estimates would recover the histogram to
earth mover distance $\Theta(\frac{1}{\sqrt{t}})$ (equivalently, $\ell_1$
distance between the CDFs), we show that, provided $n$ is sufficiently large,
we can achieve error $O(\frac{1}{t})$ which is information theoretically
optimal. We also extend our results to the multi-dimensional parameter case,
capturing settings where each member of the population has multiple associated
parameters. Beyond the theoretical results, we demonstrate that the recovery
algorithm performs well in practice on a variety of datasets, providing
illuminating insights into several domains, including politics, sports
analytics, and variation in the gender ratio of offspring. | ['Kevin Tian', 'Gregory Valiant', 'Weihao Kong'] | 2017-09-08 | learning-populations-of-parameters-1 | http://papers.nips.cc/paper/7160-learning-populations-of-parameters | http://papers.nips.cc/paper/7160-learning-populations-of-parameters.pdf | neurips-2017-12 | ['sports-analytics'] | ['computer-vision'] | [-1.89578682e-01 1.51203498e-01 -4.61430073e-01 -1.99867412e-01
-1.05822122e+00 -7.91145682e-01 -1.71013772e-01 4.09318328e-01
-6.87358260e-01 9.10734057e-01 -2.37861469e-01 -4.89133686e-01
-7.08654583e-01 -1.27888286e+00 -9.62034106e-01 -7.19910741e-01
-7.33266890e-01 7.78849959e-01 -1.57807246e-01 9.74128619e-02
2.22259283e-01 -3.85911129e-02 -1.24057961e+00 -8.43346179e-01
9.84501064e-01 1.13832664e+00 -4.78687137e-01 7.41685688e-01
1.85681671e-01 2.60847598e-01 -8.66770506e-01 -8.67513478e-01
3.93311948e-01 -1.99683771e-01 -4.66369152e-01 -3.74884129e-01
1.40096039e-01 -2.65012652e-01 -1.92435592e-01 1.39778936e+00
1.67739019e-01 2.30211794e-01 9.79863703e-01 -1.25025833e+00
-4.35779244e-01 9.29272354e-01 -1.34262896e+00 3.42537880e-01
3.80632311e-01 -1.90805152e-01 9.87402439e-01 -3.65586162e-01
4.55933839e-01 1.03375590e+00 8.07273865e-01 -1.90523058e-01
-1.17064750e+00 -1.39688468e+00 -7.83172101e-02 -3.55025321e-01
-1.94157493e+00 -9.76846442e-02 3.23148638e-01 -4.94410932e-01
2.04601064e-01 1.92552969e-01 3.71845067e-01 3.27416837e-01
1.47029851e-02 3.73935819e-01 6.32226765e-01 -4.82412338e-01
2.07700387e-01 3.15992944e-02 6.46569803e-02 6.52755320e-01
8.00672948e-01 -1.23943105e-01 -5.96607804e-01 -4.18634772e-01
7.29103386e-01 7.44632855e-02 -6.95128292e-02 -1.82077046e-02
-6.40726924e-01 8.69609237e-01 -2.98369974e-02 8.59625116e-02
-1.64089963e-01 7.50942528e-01 -1.77077334e-02 5.61261252e-02
4.96295899e-01 -9.12083536e-02 -5.67710936e-01 -3.56740475e-01
-8.22709143e-01 3.92219037e-01 8.68941069e-01 1.14591086e+00
9.69354689e-01 -8.57763663e-02 2.23881871e-01 5.96306324e-01
2.00700998e-01 1.16630340e+00 -1.95152342e-01 -1.35288644e+00
1.15083039e+00 3.53093117e-01 5.68717003e-01 -1.20303845e+00
-1.54761776e-01 -3.04454058e-01 -9.05406952e-01 -5.32717466e-01
1.04644179e+00 -7.92272389e-01 -5.08054554e-01 2.30639863e+00
2.39365637e-01 1.07827239e-01 -3.48130167e-01 4.21777755e-01
7.74068087e-02 5.70093572e-01 3.13971311e-01 -3.18417281e-01
1.36845815e+00 1.75146148e-01 -3.23665679e-01 -3.21442068e-01
4.68986839e-01 -4.76838917e-01 6.01856887e-01 1.64160833e-01
-1.42036641e+00 -6.96992800e-02 -5.90736508e-01 4.74216968e-01
-2.48811081e-01 -6.15610592e-02 7.68970251e-01 1.16228402e+00
-8.01559389e-01 3.59960496e-01 -8.25623393e-01 -3.37647386e-02
5.17342925e-01 5.23061872e-01 -2.54644990e-01 -1.93765372e-01
-1.12352312e+00 1.77053362e-01 -1.41171247e-01 -2.13774726e-01
-3.12438816e-01 -7.68699646e-01 -6.35929167e-01 3.62328619e-01
5.31237960e-01 -3.89899701e-01 8.18020046e-01 -2.52036780e-01
-6.29670739e-01 6.76469386e-01 -1.25577882e-01 -4.17690784e-01
2.27200672e-01 2.48778790e-01 -2.19342023e-01 1.25775546e-01
5.55991530e-01 9.37118828e-02 2.95106053e-01 -9.81019974e-01
-8.79343867e-01 -9.90989208e-01 -6.75411597e-02 5.29837459e-02
-4.00433272e-01 3.82591598e-02 -4.84017730e-01 -3.26799244e-01
1.87680081e-01 -8.68590117e-01 -2.53206372e-01 -1.72073454e-01
-4.60328221e-01 6.68894267e-03 -4.96772639e-02 -4.66483444e-01
1.57930541e+00 -2.10573411e+00 -2.23145351e-01 9.18942273e-01
2.45917961e-01 -2.00218052e-01 4.07110453e-01 3.82286578e-01
1.91546962e-01 5.00144780e-01 -2.20878690e-01 -2.32325584e-01
2.32990339e-01 -1.09353453e-01 -3.06559429e-02 8.63816321e-01
-6.38195872e-01 4.15809125e-01 -5.95417857e-01 -3.24124962e-01
-2.88378954e-01 2.24529922e-01 -4.88362551e-01 -3.14026356e-01
6.52185827e-02 -3.40581015e-02 -6.58709586e-01 4.92654264e-01
9.55862582e-01 -5.91247916e-01 4.97035265e-01 4.28773642e-01
-2.52367575e-02 -2.60742903e-01 -1.68926775e+00 1.10024834e+00
-1.32488285e-03 2.51726151e-01 4.86343503e-01 -1.21820331e+00
7.06645966e-01 -3.69636342e-02 7.38277078e-01 -6.82926834e-01
2.20282212e-01 1.70162886e-01 -1.37347251e-01 -1.02409206e-01
5.33415914e-01 -4.13536310e-01 -6.68821871e-01 8.10810924e-01
-2.98445076e-01 2.68477678e-01 2.50616431e-01 2.01393187e-01
1.39520633e+00 -5.05896747e-01 3.12686116e-01 -2.80720115e-01
-1.41027525e-01 -2.22201660e-01 8.17264140e-01 1.06788445e+00
-2.54054248e-01 2.94934243e-01 1.11786878e+00 6.05232492e-02
-8.60057890e-01 -1.50352466e+00 -3.48761648e-01 1.13210857e+00
4.57595736e-01 -2.61388540e-01 -7.65498698e-01 -2.92056143e-01
4.80152637e-01 5.49930036e-01 -8.87652755e-01 2.72774577e-01
-4.09288138e-01 -1.18167341e+00 6.38864100e-01 4.90703851e-01
5.95846593e-01 -3.32519233e-01 -4.37587649e-01 3.64896506e-02
-3.30909282e-01 -7.96020389e-01 -5.39070666e-01 5.17919660e-02
-6.68504417e-01 -1.01002645e+00 -5.50869524e-01 -2.31396109e-01
7.55501091e-01 2.61847787e-02 1.10532796e+00 3.47227007e-02
-2.33588099e-01 6.72296226e-01 7.40433671e-03 -6.05921268e-01
1.95255980e-01 1.15133464e-01 -4.48645689e-02 -3.48926753e-01
4.41976577e-01 -7.19765306e-01 -1.18316126e+00 1.96630687e-01
-8.74338150e-01 -8.13085020e-01 2.32819051e-01 4.03887898e-01
5.98781466e-01 5.26210725e-01 4.82158989e-01 -1.10596979e+00
4.11109865e-01 -9.81181443e-01 -8.81357670e-01 2.77240813e-01
-6.07273698e-01 -1.52836189e-01 5.32509208e-01 -3.16803217e-01
-6.07419074e-01 -4.50393260e-01 1.85874730e-01 -2.04871863e-01
5.13143539e-02 5.81624210e-01 1.40511170e-02 3.45301330e-01
5.57512343e-01 -3.17708664e-02 -6.39142096e-02 -3.71369600e-01
1.59410119e-01 4.07044411e-01 8.47458661e-01 -8.05360913e-01
7.34168589e-01 6.08198106e-01 2.24114031e-01 -4.91349071e-01
-6.94333136e-01 -1.40609220e-01 -4.73024063e-02 2.53115982e-01
5.11896789e-01 -1.06137073e+00 -1.48022747e+00 3.97881977e-02
-4.81824070e-01 -3.81143898e-01 -3.32059711e-01 5.87773144e-01
-2.28843287e-01 2.66753197e-01 -4.42768186e-01 -1.37918234e+00
-1.60934597e-01 -6.46049976e-01 7.68099308e-01 5.20874321e-01
-1.24348186e-01 -1.06329739e+00 -7.67733380e-02 4.31914270e-01
2.23975167e-01 2.51767367e-01 1.12121308e+00 -4.69432473e-01
-6.75935745e-01 -6.25892878e-01 -3.86454076e-01 -1.53143957e-01
-1.61192760e-01 -8.64691064e-02 -2.92819470e-01 -3.22096139e-01
-3.57029229e-01 8.43885168e-02 3.70143712e-01 9.34610605e-01
1.39556968e+00 -6.21610940e-01 -6.67541981e-01 4.01714385e-01
1.42787158e+00 1.69176474e-01 6.20111704e-01 -2.41742041e-02
1.02922544e-01 1.53145492e-01 6.14970744e-01 1.13494039e+00
8.71829212e-01 3.02599788e-01 2.04420552e-01 1.76755697e-01
6.05356455e-01 -2.63813376e-01 1.82637069e-02 1.48035586e-01
-1.93957407e-02 -5.42321265e-01 -9.66704905e-01 7.28515029e-01
-1.47873378e+00 -1.01007712e+00 2.25592181e-01 2.73424816e+00
7.35907733e-01 1.37767330e-01 4.55770314e-01 -1.63788334e-01
9.25815880e-01 -2.97992136e-02 -7.04168618e-01 3.28310356e-02
1.39509574e-01 9.47955787e-01 1.19173825e+00 3.69724125e-01
-6.41485393e-01 3.42097759e-01 6.10978413e+00 1.03125167e+00
-5.85560918e-01 -1.31220501e-02 1.01984715e+00 -2.85873264e-01
-4.59437698e-01 2.16253921e-01 -8.85655642e-01 1.05233681e+00
1.19066370e+00 -5.05713880e-01 4.44451034e-01 6.83897257e-01
-8.40494931e-02 -8.08567107e-01 -8.10238838e-01 1.02419353e+00
-1.67745173e-01 -1.11836600e+00 -5.69009125e-01 6.53743565e-01
6.93548858e-01 -2.97766209e-01 7.04419374e-01 3.36767435e-01
9.76212502e-01 -1.14010537e+00 4.26464617e-01 2.94781536e-01
1.24240637e+00 -1.18962550e+00 5.50024331e-01 5.96786261e-01
-1.42511570e+00 -3.91539842e-01 -3.10951084e-01 -3.14164683e-02
1.69300765e-01 8.54003310e-01 -6.54133439e-01 4.69679981e-01
9.20914114e-01 3.71972024e-02 -3.06843594e-02 7.88693726e-01
3.95445943e-01 5.91377378e-01 -9.11653757e-01 2.69279648e-02
-1.98828965e-01 -5.05948424e-01 1.79670900e-01 7.13596642e-01
9.00015235e-01 7.34203160e-01 2.34065368e-03 6.20654821e-01
-6.94411457e-01 1.49874091e-01 -3.29226196e-01 5.50577901e-02
1.32905543e+00 8.10057640e-01 -9.01794136e-01 -1.41325742e-01
-1.12035245e-01 1.68352053e-01 2.58864582e-01 4.39733863e-01
-9.87314224e-01 -7.67205358e-01 7.69876242e-01 4.60536659e-01
4.98243064e-01 -2.18351975e-01 -4.35475796e-01 -8.15913022e-01
9.63325426e-02 -4.27394152e-01 7.73819149e-01 -2.55391002e-01
-1.33087742e+00 -2.14840412e-01 2.15228990e-01 -9.42700326e-01
-3.72482002e-01 -2.90204614e-01 -3.76714766e-01 8.86476457e-01
-9.17452514e-01 -4.83014405e-01 2.54734278e-01 4.54570472e-01
-4.20310527e-01 4.33492027e-02 5.77256501e-01 4.48852479e-01
-6.51335657e-01 1.00211644e+00 4.93280530e-01 3.52716416e-01
4.53449816e-01 -1.13445878e+00 4.19157557e-03 6.17750525e-01
-1.87480733e-01 7.08654583e-01 8.16603720e-01 -7.07128942e-01
-1.51873398e+00 -6.58774495e-01 5.32490492e-01 -3.86635542e-01
5.00255823e-01 -1.89275101e-01 -3.39858055e-01 1.04012263e+00
-2.29704425e-01 8.16085339e-02 1.05599463e+00 3.90788317e-01
-1.48686260e-01 -3.66069317e-01 -1.69973719e+00 4.68249470e-01
8.79937470e-01 -1.18264392e-01 3.96981627e-01 1.43298909e-01
2.17309251e-01 -6.38968647e-01 -1.40290785e+00 2.19500437e-01
7.10087478e-01 -1.03945398e+00 1.15223145e+00 -3.19747925e-01
6.59211501e-02 7.26753175e-02 -4.47512567e-01 -8.12585413e-01
1.28887981e-01 -6.30379498e-01 1.92625374e-02 1.22868776e+00
6.23663008e-01 -7.71279871e-01 1.11806417e+00 1.38402808e+00
6.94516361e-01 -5.46424210e-01 -1.41428757e+00 -4.37563747e-01
6.24531031e-01 -4.93357420e-01 9.79154170e-01 7.05249846e-01
-6.08479418e-02 -1.12583846e-01 -1.80588827e-01 3.22743386e-01
9.06799257e-01 1.70157865e-01 9.41228390e-01 -1.31743729e+00
-4.80665296e-01 -2.02164322e-01 -2.09836811e-01 -1.08432424e+00
-6.70372844e-02 -4.06410277e-01 -2.40276065e-02 -9.09033358e-01
5.70698857e-01 -1.20198095e+00 -5.61747253e-02 2.35246345e-01
-1.71822295e-01 8.79810471e-03 -1.11384273e-01 -1.41762182e-01
-8.49133015e-01 1.76577881e-01 8.30128372e-01 2.46595204e-01
4.25545499e-02 1.23050436e-01 -1.15565097e+00 7.15515196e-01
4.22845215e-01 -6.11809373e-01 -3.09634447e-01 -4.83377218e-01
5.72605908e-01 1.02760863e+00 -5.64506883e-03 -6.74523532e-01
2.59454548e-01 -4.16112453e-01 5.74447036e-01 -5.57283044e-01
3.95599157e-01 -4.04549927e-01 4.08206105e-01 1.21609293e-01
-2.04265282e-01 2.36528099e-01 -1.20611534e-01 8.43262434e-01
2.44050786e-01 -1.11351043e-01 6.36367500e-01 -1.96358725e-01
2.91766077e-01 6.93998754e-01 -7.66255930e-02 4.66142684e-01
9.28342283e-01 5.74730374e-02 -4.90331590e-01 -7.52358079e-01
-3.51782471e-01 4.04189736e-01 5.29121399e-01 -5.01426220e-01
4.28600945e-02 -9.80510831e-01 -4.15284097e-01 -1.40080288e-01
-2.02020556e-01 2.47891456e-01 6.06030762e-01 6.58269942e-01
-2.62039810e-01 -5.39925769e-02 4.23897415e-01 -4.08295751e-01
-7.63653278e-01 1.30373046e-01 1.54576376e-01 -4.93232697e-01
1.34685084e-01 1.03325593e+00 -1.47475764e-01 -3.13907593e-01
1.89969838e-01 -6.51135892e-02 3.86009485e-01 -8.54468942e-02
3.68616402e-01 7.75039256e-01 -3.27100217e-01 -3.66146713e-01
-3.13532442e-01 5.60010254e-01 3.23316222e-03 -5.51185727e-01
1.32468140e+00 -4.97236758e-01 -1.01880856e-01 1.34405434e-01
1.17793381e+00 4.81417030e-01 -1.10587883e+00 -1.30837455e-01
-2.29295880e-01 -7.65291870e-01 -6.59443021e-01 -4.32191223e-01
-1.36146307e+00 5.94665229e-01 4.71253842e-01 5.58797419e-01
9.54835713e-01 1.75550476e-01 6.54242158e-01 -1.06240049e-01
6.83356345e-01 -1.13033009e+00 -1.52631313e-01 1.53327614e-01
-1.73424765e-01 -9.97507215e-01 2.27278993e-01 -3.13741028e-01
-2.42906049e-01 5.68392873e-01 3.42547268e-01 -5.91788292e-02
9.70528185e-01 3.75313491e-01 -4.27452207e-01 -3.35560024e-01
-4.02092606e-01 1.68238744e-01 -4.22543347e-01 3.58086914e-01
3.46661597e-01 3.68316770e-01 -3.84986848e-01 1.03242683e+00
-6.03240550e-01 -8.72662812e-02 4.99181330e-01 8.12086880e-01
-4.48047161e-01 -1.13895905e+00 -5.37659407e-01 9.04458880e-01
-9.39868927e-01 2.14792311e-01 3.29893947e-01 1.07498157e+00
2.24515021e-01 9.07476306e-01 4.62421954e-01 7.64655927e-03
1.30222932e-01 -9.34983343e-02 4.87185508e-01 -3.13639820e-01
-1.85894504e-01 -8.45637396e-02 -1.52236432e-01 -1.89781547e-01
9.29474384e-02 -1.05051339e+00 -1.42371976e+00 -1.09887052e+00
-1.17828526e-01 4.18599010e-01 5.85748076e-01 7.49387145e-01
2.60847002e-01 -3.51994634e-01 8.88755143e-01 -1.58706427e-01
-4.07722563e-01 -7.33981311e-01 -1.41680527e+00 2.83657402e-01
-8.16170126e-02 -8.07252526e-01 -5.87953985e-01 -3.36643010e-01] | [6.748971939086914, 4.673530578613281] |
d10ae1bb-5f2a-4c91-bb3e-7507bce5720a | federated-sparse-training-lottery-aware-model | 2208.13092 | null | https://arxiv.org/abs/2208.13092v2 | https://arxiv.org/pdf/2208.13092v2.pdf | Lottery Aware Sparsity Hunting: Enabling Federated Learning on Resource-Limited Edge | Limited computation and communication capabilities of clients pose significant challenges in federated learning (FL) over resource-limited edge nodes. A potential solution to this problem is to deploy off-the-shelf sparse learning algorithms that train a binary sparse mask on each client with the expectation of training a consistent sparse server mask yielding sparse weight tensors. However, as we investigate in this paper, such naive deployments result in a significant drop in accuracy compared to FL with dense models, especially for clients with limited resource budgets. In particular, our investigations reveal a serious lack of consensus among the trained sparsity masks on clients, which prevents convergence for the server mask and potentially leads to a substantial drop in model performance. Based on such key observations, we propose \textit{federated lottery aware sparsity hunting} (FLASH), a unified sparse learning framework to make the server win a lottery in terms of yielding a sparse sub-model, able to maintain classification performance under highly resource-limited client settings. Moreover, to support FL on different devices requiring different parameter density, we leverage our findings to present \textit{hetero-FLASH}, where clients can have different target sparsity budgets based on their device resource limits. Experimental evaluations with multiple models on various datasets (both IID and non-IID) show superiority of our models in closing the gap with unpruned baseline while yielding up to $\mathord{\sim}10.1\%$ improved accuracy with $\mathord{\sim}10.26\times$ fewer communication costs, compared to existing alternatives, at similar hyperparameter settings. | ['Salman Avestimehr', 'Yue Niu', 'Saurav Prakash', 'Souvik Kundu', 'Sara Babakniya'] | 2022-08-27 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 8.29305202e-02 1.36452228e-01 -5.25890589e-01 -1.49015471e-01
-9.07058418e-01 -3.09453785e-01 5.10176681e-02 -3.52807313e-01
5.55587523e-02 6.50798082e-01 1.42686874e-01 -3.34619850e-01
-3.49711239e-01 -5.38915277e-01 -7.56744504e-01 -8.01170886e-01
-8.34292024e-02 6.02315545e-01 -6.05085269e-02 2.63296306e-01
-8.68254676e-02 3.15147728e-01 -1.34143591e+00 4.72156286e-01
4.69426721e-01 1.43051016e+00 1.48208007e-01 1.00364685e-01
-2.68714100e-01 9.82117355e-01 -2.66746908e-01 -4.98402357e-01
6.48776889e-01 -4.66984278e-03 -5.83352625e-01 3.64542812e-01
4.73467082e-01 -3.81693006e-01 -5.44044673e-01 7.25445211e-01
3.59806627e-01 -3.81525397e-01 3.99376720e-01 -1.36959624e+00
-1.71733811e-01 9.11180079e-01 -7.86532164e-01 1.67090386e-01
3.75994220e-02 2.84948796e-01 1.02152944e+00 -9.03409481e-01
4.66047078e-01 6.29261792e-01 9.26960349e-01 2.12736443e-01
-1.44739687e+00 -1.04186296e+00 4.35127914e-01 5.44765331e-02
-1.64974070e+00 -8.61163199e-01 6.37680411e-01 2.74416655e-02
8.94225538e-01 5.22629857e-01 2.97704488e-01 9.29223895e-01
-2.50721693e-01 6.79801285e-01 9.83949184e-01 -1.75631240e-01
4.85488504e-01 4.44414824e-01 1.09268492e-02 5.46587050e-01
5.97158849e-01 -3.94300550e-01 -9.47341919e-01 -5.46974957e-01
6.69438958e-01 2.76260644e-01 -2.68742949e-01 -5.00408053e-01
-6.97600663e-01 6.64060831e-01 1.30179673e-01 2.38856420e-01
-6.18814945e-01 1.02992989e-01 2.80016452e-01 2.37422138e-01
3.18631589e-01 -8.28490928e-02 -5.40724576e-01 -1.32815331e-01
-1.36941338e+00 -8.11328441e-02 1.20109951e+00 1.16158605e+00
9.94279444e-01 2.91751981e-01 -2.85044014e-01 6.17735982e-01
-9.02144425e-03 5.32109976e-01 8.27636942e-02 -1.13319480e+00
7.70488799e-01 5.71331561e-01 3.68592292e-02 -9.15058553e-01
-2.88089275e-01 -8.39237809e-01 -1.05815673e+00 -4.23441455e-02
2.81049520e-01 -2.74915457e-01 -4.24984932e-01 1.67373323e+00
3.10695827e-01 7.91388452e-01 1.29616847e-02 1.02889645e+00
5.42065084e-01 4.07669574e-01 -1.37647673e-01 -5.05515873e-01
1.08618248e+00 -9.35670853e-01 -2.71321535e-01 -1.18467979e-01
7.36877203e-01 -5.72975159e-01 1.17579973e+00 5.34515738e-01
-1.10987365e+00 1.32712796e-01 -7.04125345e-01 4.63133544e-01
2.80695796e-01 9.76346135e-02 1.09227514e+00 9.15915906e-01
-1.08406174e+00 3.43819141e-01 -7.97540903e-01 -4.07399863e-01
8.53401721e-01 6.38765812e-01 -1.09267317e-01 -4.10625726e-01
-4.08952087e-01 2.36963913e-01 -3.25108886e-01 -2.37006247e-01
-8.69939327e-01 -1.19416165e+00 -3.84701818e-01 6.17428660e-01
6.22751951e-01 -7.39071250e-01 9.14045095e-01 -7.83284605e-01
-1.22415268e+00 6.87364221e-01 -3.11258078e-01 -6.01861477e-01
5.51131964e-01 1.72273546e-01 -2.74841785e-01 6.49043098e-02
-5.60544729e-02 1.19917668e-01 8.32803488e-01 -1.38772452e+00
-5.83143532e-01 -4.42060828e-01 -3.18403281e-02 2.11941283e-02
-9.88705754e-01 -1.33144140e-01 -6.77191019e-01 -4.43286330e-01
1.22005396e-01 -9.12437737e-01 -3.63197029e-01 5.17884232e-02
-3.30807477e-01 9.07508377e-03 1.07321596e+00 -1.38797477e-01
1.34847534e+00 -2.07720041e+00 -1.07477941e-01 5.28445959e-01
5.99385500e-01 9.55345929e-02 -8.92444327e-02 3.34344476e-01
4.45693493e-01 -1.63241867e-02 8.78634527e-02 -8.41588438e-01
-2.08096534e-01 3.92698348e-01 -2.87567735e-01 5.79952657e-01
-2.18236104e-01 4.99112159e-01 -3.77791852e-01 -4.66119647e-01
1.09994495e-02 4.42247182e-01 -8.18160236e-01 2.00229689e-01
-1.26155660e-01 2.37789601e-01 -6.35840714e-01 1.02226591e+00
8.82967174e-01 -9.71364498e-01 7.42124975e-01 -3.52875799e-01
3.01078856e-01 1.34921744e-01 -1.60529280e+00 1.56130254e+00
-7.23586917e-01 1.03207693e-01 9.64927375e-01 -1.23663330e+00
5.77165782e-01 2.66413003e-01 9.32532489e-01 -5.08242786e-01
1.99192807e-01 3.68414521e-01 -5.04602134e-01 -2.36708328e-01
1.73327804e-01 5.11946380e-02 1.46648720e-01 8.10229182e-01
-2.12802663e-02 4.12086993e-01 -1.76052183e-01 5.11728406e-01
1.64317000e+00 -5.91164410e-01 -3.79634053e-01 -3.38714093e-01
1.71196029e-01 -2.39319742e-01 5.97469389e-01 9.09982741e-01
9.01249349e-02 6.40223026e-01 5.35628259e-01 -4.40760523e-01
-8.50824118e-01 -8.73220921e-01 -1.02136545e-01 1.15850806e+00
2.29727104e-01 -5.59071779e-01 -4.89914745e-01 -6.81801975e-01
3.08961660e-01 4.37756419e-01 -2.89834350e-01 2.20113963e-01
-4.96979833e-01 -9.09279168e-01 3.68657708e-01 2.76245326e-01
2.21551329e-01 -6.17482126e-01 -4.88119513e-01 2.55708396e-02
4.34144326e-02 -1.49782586e+00 -4.18052107e-01 3.42148095e-01
-7.64628410e-01 -8.55214894e-01 -5.31345367e-01 -4.64127302e-01
7.19787240e-01 7.09887266e-01 1.12242818e+00 2.29581699e-01
-2.92720854e-01 5.38800716e-01 -2.75505126e-01 -4.43782322e-02
7.81618953e-02 4.35344934e-01 2.80763268e-01 3.35892707e-01
2.15552881e-01 -1.15289402e+00 -8.41680646e-01 4.46247995e-01
-7.44459152e-01 -3.10319029e-02 6.90800428e-01 5.69967270e-01
6.28504634e-01 1.07977159e-01 5.68372250e-01 -1.10647166e+00
1.86463729e-01 -1.06944764e+00 -4.09071356e-01 2.26690605e-01
-1.04971564e+00 -5.60144372e-02 8.66181254e-01 -4.86250758e-01
-8.37198734e-01 -8.46667588e-02 4.19108778e-01 -9.99364197e-01
2.75252223e-01 2.10232064e-01 -8.56809989e-02 -3.73548687e-01
5.28545022e-01 2.74134040e-01 9.30161104e-02 -6.75056219e-01
1.72522277e-01 7.17001915e-01 1.52970731e-01 -8.31200302e-01
8.09503436e-01 8.51486564e-01 1.26930192e-01 -5.37870467e-01
-7.34808922e-01 -3.70662391e-01 6.96359426e-02 -4.17045951e-02
-6.51210034e-03 -1.30889237e+00 -1.03351879e+00 1.20956488e-01
-4.16150212e-01 -4.26806927e-01 -4.08727199e-01 1.22496620e-01
-2.25278243e-01 2.91538596e-01 -4.80838358e-01 -8.64834845e-01
-6.82767808e-01 -9.79093373e-01 9.91128802e-01 1.58475846e-01
-1.89698085e-01 -7.32177615e-01 -4.94630098e-01 8.44220936e-01
8.96710038e-01 -1.50090322e-01 7.10133493e-01 -6.64875746e-01
-9.48944271e-01 -1.44605681e-01 -5.03593564e-01 1.49824783e-01
1.15239128e-01 -5.10737181e-01 -8.40153813e-01 -7.87028491e-01
-4.95430501e-03 -6.05534852e-01 6.98895216e-01 2.02251241e-01
1.48170209e+00 -6.93150163e-01 -6.08174622e-01 9.06865060e-01
1.55748773e+00 -5.28767586e-01 2.97050655e-01 -1.20229401e-01
6.56188786e-01 2.41747350e-01 1.86690778e-01 1.13563180e+00
4.63235348e-01 7.98061788e-01 4.55022454e-01 -1.19121760e-01
-2.91774511e-01 -1.84685260e-01 2.51429796e-01 5.05532205e-01
2.70474702e-01 -2.49796361e-01 -6.53207839e-01 6.19256973e-01
-1.79748929e+00 -7.96578705e-01 5.99289127e-02 2.33408809e+00
6.82678342e-01 1.06189467e-01 3.22685301e-01 4.00266051e-02
3.31143707e-01 1.62873745e-01 -8.52108061e-01 1.90087352e-02
-2.52283603e-01 3.83177340e-01 8.78817737e-01 4.37556468e-02
-4.41294253e-01 6.60861433e-01 5.38190746e+00 1.18895030e+00
-1.44068229e+00 4.63758200e-01 9.98092115e-01 -8.60438704e-01
-5.01022875e-01 1.18144445e-01 -9.53374803e-01 7.14878201e-01
1.01146448e+00 -1.13545761e-01 7.64641404e-01 1.06984735e+00
3.41210216e-01 -1.42285436e-01 -1.04498208e+00 1.34583163e+00
3.30472477e-02 -1.82747769e+00 -1.62311599e-01 3.19426537e-01
9.46591437e-01 3.99996728e-01 2.13182986e-01 1.61164701e-01
3.41367394e-01 -1.02195740e+00 5.69990814e-01 1.39784992e-01
1.02109671e+00 -4.03388828e-01 4.67610061e-01 3.58108014e-01
-1.12773907e+00 -4.08569574e-01 -3.09740424e-01 -1.31300449e-01
1.84499905e-01 1.02366447e+00 -7.18303621e-01 3.74348909e-01
1.04012704e+00 3.13326955e-01 -9.26966220e-02 6.73324108e-01
5.95985115e-01 9.85322356e-01 -7.99365401e-01 2.56249815e-01
-9.25676525e-02 -4.68681827e-02 3.78601074e-01 9.79204893e-01
5.19712210e-01 1.68919206e-01 3.93308669e-01 7.72247672e-01
-5.02849162e-01 2.29384482e-01 -2.76568234e-01 2.22446248e-01
8.33133638e-01 1.36480200e+00 -5.67940593e-01 -1.73619643e-01
-6.04388773e-01 7.68421292e-01 5.15131593e-01 4.57245708e-01
-8.46347690e-01 3.25900584e-01 9.16697204e-01 7.09807336e-01
6.27011001e-01 4.66661379e-02 -7.58436680e-01 -1.36470723e+00
3.77292275e-01 -8.10072720e-01 4.75661755e-01 -1.04939424e-01
-1.43565893e+00 3.51645231e-01 -2.79439330e-01 -9.46421146e-01
2.25848496e-01 -5.40555902e-02 -6.85735464e-01 6.51025772e-01
-1.42607749e+00 -1.32135689e+00 -4.50203031e-01 8.24506760e-01
2.87120968e-01 -3.68953824e-01 5.99923313e-01 6.90723956e-01
-8.53737175e-01 1.28325891e+00 2.51458615e-01 -4.07523453e-01
3.86433452e-01 -4.59640652e-01 -2.90688515e-01 4.56463695e-01
2.61954635e-01 3.73426825e-01 6.14427567e-01 -3.84627640e-01
-1.97879052e+00 -1.04572368e+00 4.41047579e-01 -2.33531818e-01
5.77024877e-01 -5.20712078e-01 -8.64575028e-01 4.41910267e-01
-1.73549861e-01 5.21733820e-01 8.68072867e-01 2.84134299e-01
-3.49101454e-01 -5.86569607e-01 -1.48899329e+00 3.80254835e-01
1.33580732e+00 -2.86160797e-01 4.78769928e-01 6.25629365e-01
5.72921515e-01 -1.85885370e-01 -9.42939878e-01 4.77672070e-01
4.25881475e-01 -1.14948034e+00 7.84650266e-01 -4.36928540e-01
-1.50045278e-02 3.02754510e-02 -4.66191739e-01 -5.44370532e-01
-2.45984718e-01 -1.02403724e+00 -6.39268577e-01 1.47307539e+00
4.51067805e-01 -5.61172307e-01 1.64174891e+00 1.03949928e+00
5.63359857e-02 -1.18768632e+00 -1.27188396e+00 -7.85152853e-01
-2.57000446e-01 -5.39403439e-01 6.09572351e-01 8.65919590e-01
-1.15426242e-01 -9.72372666e-02 -6.87651217e-01 3.04608792e-01
7.66032636e-01 2.17176795e-01 9.05174732e-01 -9.74314392e-01
-6.66204453e-01 -2.64257550e-01 -3.15122083e-02 -1.04045749e+00
2.26824909e-01 -9.15214241e-01 -6.56930506e-01 -1.00379443e+00
5.28955936e-01 -1.35263932e+00 -3.49682897e-01 7.76934564e-01
1.51778415e-01 4.30225432e-01 2.78350770e-01 6.93156004e-01
-9.28529143e-01 4.52181220e-01 3.99112523e-01 5.35571724e-02
-2.20287353e-01 2.60625422e-01 -1.04588878e+00 2.98951149e-01
5.30287445e-01 -5.05215049e-01 -3.94046545e-01 -5.55607855e-01
2.02231154e-01 6.10519871e-02 2.83469975e-01 -1.04403019e+00
3.64550978e-01 -3.02661717e-01 6.49577007e-02 3.13655995e-02
3.79405856e-01 -1.24416327e+00 4.41615820e-01 2.18784258e-01
1.04688220e-01 -2.81665176e-01 -4.09646163e-04 5.09822667e-01
2.35440060e-01 1.55864403e-01 5.79540908e-01 1.16631001e-01
-1.58510745e-01 4.83265162e-01 -3.89277600e-02 7.53065944e-02
1.17928469e+00 -3.53472680e-01 -2.42609054e-01 -6.29892886e-01
-3.92106414e-01 4.05994087e-01 6.73797250e-01 -2.07064468e-02
1.27837867e-01 -1.02237546e+00 -5.92450619e-01 2.33836651e-01
-2.11420178e-01 1.41136593e-03 6.80971980e-01 1.20959210e+00
-1.89319625e-01 4.60386068e-01 2.66924083e-01 -5.05376756e-01
-1.07417500e+00 4.54797447e-01 1.10823423e-01 -4.43125367e-01
-6.15636349e-01 8.74085903e-01 -2.71185003e-02 1.80738643e-02
6.08544767e-01 8.69966522e-02 5.53326428e-01 -2.07923561e-01
3.38330239e-01 5.72110355e-01 3.91139030e-01 -4.22437936e-01
-3.91169786e-01 2.14783698e-01 -1.63900942e-01 4.29962784e-01
1.60847890e+00 -1.97731018e-01 4.29540575e-02 -1.86691061e-01
1.15333712e+00 2.67118603e-01 -1.48854983e+00 -5.92048287e-01
-4.25099403e-01 -6.13778353e-01 2.30828270e-01 -6.66390300e-01
-1.74830687e+00 2.52511263e-01 6.57455087e-01 1.39714494e-01
1.22218621e+00 3.83950233e-01 1.14066148e+00 1.32674247e-01
9.35323656e-01 -7.20804930e-01 -6.61508739e-03 -7.84419626e-02
3.42971921e-01 -1.31502497e+00 1.50580108e-01 -6.10136449e-01
-6.35900319e-01 6.02417171e-01 6.34189487e-01 2.98161954e-02
7.78163970e-01 6.68126225e-01 -2.25492507e-01 -3.17836046e-01
-9.76834893e-01 1.31841332e-01 -3.47127706e-01 2.81876266e-01
-7.73404166e-03 1.62577461e-02 -1.19023323e-01 1.04078746e+00
-2.38916203e-02 1.88581288e-01 2.41933048e-01 8.89873922e-01
-1.06693007e-01 -1.15389943e+00 -3.73871118e-01 1.05416381e+00
-5.94175458e-01 -3.62295695e-02 1.21767007e-01 2.37471685e-01
1.45903826e-02 1.00533497e+00 5.34683056e-02 -3.69137377e-01
7.37437755e-02 -1.28477097e-01 8.47059339e-02 -5.65566242e-01
-6.87620640e-01 -3.04727312e-02 8.24474618e-02 -8.61770988e-01
-3.92803140e-02 -6.41076982e-01 -1.08931041e+00 -6.29477262e-01
-2.08476365e-01 1.22882418e-01 4.48599458e-01 7.64966846e-01
1.02174664e+00 2.04517823e-02 9.96588409e-01 -8.49230528e-01
-8.95879865e-01 -5.77886581e-01 -8.10779929e-01 3.60794872e-01
-1.19104594e-01 -5.85281253e-01 -7.01484442e-01 -3.52611482e-01] | [5.900199890136719, 6.108227252960205] |
f5969316-f25c-4375-bd6b-f3f284b8fc57 | would-you-ask-it-that-way-measuring-and | 2205.12768 | null | https://arxiv.org/abs/2205.12768v1 | https://arxiv.org/pdf/2205.12768v1.pdf | Would You Ask it that Way? Measuring and Improving Question Naturalness for Knowledge Graph Question Answering | Knowledge graph question answering (KGQA) facilitates information access by leveraging structured data without requiring formal query language expertise from the user. Instead, users can express their information needs by simply asking their questions in natural language (NL). Datasets used to train KGQA models that would provide such a service are expensive to construct, both in terms of expert and crowdsourced labor. Typically, crowdsourced labor is used to improve template-based pseudo-natural questions generated from formal queries. However, the resulting datasets often fall short of representing genuinely natural and fluent language. In the present work, we investigate ways to characterize and remedy these shortcomings. We create the IQN-KGQA test collection by sampling questions from existing KGQA datasets and evaluating them with regards to five different aspects of naturalness. Then, the questions are rewritten to improve their fluency. Finally, the performance of existing KGQA models is compared on the original and rewritten versions of the NL questions. We find that some KGQA systems fare worse when presented with more realistic formulations of NL questions. The IQN-KGQA test collection is a resource to help evaluate KGQA systems in a more realistic setting. The construction of this test collection also sheds light on the challenges of constructing large-scale KGQA datasets with genuinely NL questions. | ['Krisztian Balog', 'Trond Linjordet'] | 2022-05-25 | null | null | null | null | ['graph-question-answering', 'natural-questions'] | ['graphs', 'miscellaneous'] | [-2.03572318e-01 4.35806841e-01 1.46711916e-01 -5.60305893e-01
-1.30489171e+00 -1.14168620e+00 5.89812338e-01 1.83064535e-01
-5.50186932e-01 8.62186670e-01 3.34546298e-01 -6.52997673e-01
1.99732166e-02 -1.12042785e+00 -6.71465695e-01 1.59706146e-01
5.99395275e-01 1.04083645e+00 4.29519564e-01 -8.26504409e-01
-3.87186147e-02 1.66954607e-01 -1.09431732e+00 5.49441814e-01
1.47711515e+00 7.29713261e-01 5.15437499e-03 7.88854837e-01
-6.26514912e-01 1.10665500e+00 -7.55025446e-01 -1.10325634e+00
2.77143151e-01 -5.09046257e-01 -1.40212679e+00 -2.05505192e-01
7.84751713e-01 -4.11650062e-01 -1.62283525e-01 6.92592323e-01
3.93390059e-01 2.03040853e-01 3.72141391e-01 -1.42450106e+00
-1.21516597e+00 4.29841727e-01 2.84105867e-01 5.56411408e-02
1.05606997e+00 3.78351361e-01 1.41576278e+00 -8.48314345e-01
7.24703848e-01 1.25048733e+00 4.57105100e-01 7.14521170e-01
-1.15913558e+00 -1.69894651e-01 -1.15611538e-01 3.13774228e-01
-1.22846711e+00 -3.41134250e-01 6.82844222e-01 -1.79968461e-01
1.07307994e+00 4.68866289e-01 3.67223203e-01 9.01801705e-01
-5.81584036e-01 9.04112279e-01 9.10402417e-01 -4.49748397e-01
2.01145425e-01 4.68510181e-01 3.58408540e-01 8.04676950e-01
3.28461379e-01 -3.23497474e-01 -4.89436150e-01 -4.85428244e-01
4.22637135e-01 -5.12513697e-01 -4.40134883e-01 -3.01661372e-01
-8.80424917e-01 1.00172210e+00 4.04575676e-01 1.50024906e-01
-3.08685690e-01 -5.29514328e-02 5.80468588e-02 7.11688876e-01
1.68098137e-01 1.21374905e+00 -8.13739061e-01 -9.88045707e-02
-4.82816070e-01 7.69157350e-01 1.46235180e+00 1.24091542e+00
1.10620189e+00 -3.26721817e-01 -4.40810025e-01 6.29927695e-01
3.48708600e-01 7.03470290e-01 2.89951861e-01 -1.10924053e+00
6.85320497e-01 1.04524863e+00 5.04815459e-01 -1.03914511e+00
-8.85269195e-02 3.54733760e-03 -1.48254648e-01 -3.83546680e-01
9.42708194e-01 -1.65372983e-01 -7.04899251e-01 1.57026041e+00
4.28381324e-01 -5.89476168e-01 3.74062121e-01 9.41923201e-01
1.29133666e+00 6.93699002e-01 2.06084743e-01 2.28133157e-01
1.53071070e+00 -1.00122583e+00 -6.64921224e-01 -3.51175547e-01
8.93405437e-01 -5.06837308e-01 1.91875029e+00 4.58878726e-02
-9.44923639e-01 -2.84515500e-01 -4.35137212e-01 -5.91372788e-01
-6.77253783e-01 -1.21838905e-01 5.65957904e-01 8.71238887e-01
-1.35233212e+00 -1.02488892e-02 -2.76705801e-01 -3.52709264e-01
2.32982680e-01 7.62643153e-03 -1.82290897e-01 -6.14219487e-01
-1.70450222e+00 1.02556908e+00 5.81171662e-02 -5.14560640e-02
-8.50158691e-01 -9.60525095e-01 -1.05720341e+00 -4.66107838e-02
8.32693338e-01 -9.29108202e-01 1.65831125e+00 -6.43791020e-01
-1.20565593e+00 7.64164567e-01 -3.38451296e-01 -2.95962632e-01
5.41194141e-01 -8.94291922e-02 -4.07778293e-01 2.28926808e-01
3.43449116e-01 5.26319623e-01 4.91645575e-01 -1.14883935e+00
-1.46376818e-01 -2.68518448e-01 9.56128418e-01 2.47479483e-01
-3.41481678e-02 -1.64667606e-01 -3.69862348e-01 -2.60806084e-01
-4.20240074e-01 -6.36309505e-01 -2.06853211e-01 -1.02824360e-01
-2.87041008e-01 -5.94877422e-01 5.09458363e-01 -7.64674604e-01
9.59089816e-01 -1.66316795e+00 -3.53161961e-01 9.77041498e-02
4.27404106e-01 4.76638615e-01 -3.61910641e-01 8.78943324e-01
5.48939586e-01 5.01183391e-01 -1.32947624e-01 6.73291460e-02
3.08400989e-01 4.57821101e-01 -4.86369759e-01 -3.51589173e-01
4.73265797e-01 1.59806359e+00 -1.25487387e+00 -5.95525026e-01
-3.51225883e-01 -8.27145576e-02 -6.75630033e-01 4.56126839e-01
-9.28119898e-01 2.31087342e-01 -6.59008980e-01 6.68583930e-01
3.25480878e-01 -4.65291083e-01 -1.69766590e-01 -1.31222665e-01
4.61357504e-01 6.14874005e-01 -7.55298138e-01 1.61222517e+00
-6.11964583e-01 4.43988174e-01 -1.69998378e-01 -4.74734187e-01
7.49051690e-01 2.89671004e-01 7.42086535e-03 -8.95097852e-01
-3.16863656e-01 1.73885196e-01 -2.31930241e-01 -8.32182765e-01
6.20131075e-01 -3.52550715e-01 -2.26878062e-01 7.81672359e-01
-3.21797431e-02 -7.89597690e-01 4.80839252e-01 6.82009697e-01
1.21996081e+00 -3.18692535e-01 6.92189559e-02 -1.13135576e-01
5.41383803e-01 5.68091929e-01 5.85591011e-02 1.09860861e+00
-2.90029943e-01 2.09154397e-01 4.49300021e-01 -4.51565385e-01
-7.28494346e-01 -1.12870800e+00 3.70847732e-01 1.17949641e+00
2.42795539e-03 -5.03116846e-01 -6.78902149e-01 -1.21197033e+00
1.18593901e-01 1.14316380e+00 -3.91437322e-01 -4.67463695e-02
-3.79010320e-01 -8.22873637e-02 9.17075157e-01 4.46869522e-01
5.62704921e-01 -1.27192593e+00 -4.62718546e-01 2.00009108e-01
-7.25791276e-01 -1.31417572e+00 -4.54997182e-01 -5.83098888e-01
-4.20548677e-01 -1.49361873e+00 -4.63447511e-01 -5.66110015e-01
5.29256582e-01 4.00509894e-01 1.93217897e+00 3.48309696e-01
1.04189120e-01 9.22197700e-01 -7.23606408e-01 -5.84925234e-01
-5.40265739e-01 9.81893316e-02 -5.14846861e-01 -3.95811647e-01
7.53793657e-01 -1.97147056e-01 -5.34461319e-01 4.44819450e-01
-1.34580541e+00 -6.19781613e-02 4.43685025e-01 4.87631649e-01
2.47455597e-01 -3.13438237e-01 9.44563568e-01 -1.24571252e+00
1.26467955e+00 -4.88929570e-01 -5.62367797e-01 8.05205524e-01
-5.71502745e-01 3.23461831e-01 7.90520489e-01 -1.70175508e-01
-1.18614769e+00 -3.31299514e-01 -1.97393149e-01 1.49375051e-01
-7.46229365e-02 8.07235122e-01 -3.49164188e-01 -1.77499101e-01
1.06996739e+00 4.57600281e-02 -2.03492180e-01 -2.85955340e-01
1.01384854e+00 5.24391830e-01 3.22158635e-01 -9.86879528e-01
1.09620476e+00 2.04321235e-01 -5.85085630e-01 -7.81839311e-01
-1.04048073e+00 -5.43689311e-01 -1.88510865e-01 -3.58596519e-02
6.47736967e-01 -5.95408142e-01 -6.43772006e-01 7.93439671e-02
-1.18193090e+00 -6.11580729e-01 -5.71961224e-01 -1.92954801e-02
-4.92717326e-01 3.45348716e-01 -4.04973477e-01 -6.95693910e-01
-3.89111906e-01 -8.88233960e-01 1.05695796e+00 1.23799488e-01
-3.26762378e-01 -1.06639099e+00 2.43459120e-01 1.05200589e+00
6.29833043e-01 3.94853130e-02 1.28313923e+00 -9.04663563e-01
-5.82954824e-01 -2.57053912e-01 -2.78087765e-01 2.46139303e-01
2.43704781e-01 -3.37390512e-01 -8.73115003e-01 7.59097487e-02
-1.06891662e-01 -8.82443190e-01 3.33390266e-01 -4.06481981e-01
7.75486112e-01 -7.59132564e-01 3.14041555e-01 -8.93352851e-02
1.28782201e+00 -2.57376973e-02 5.87756336e-01 4.17997539e-02
6.25472963e-01 8.72880518e-01 4.93768334e-01 -1.04825459e-02
1.04553866e+00 3.42820108e-01 2.19081566e-01 1.43690169e-01
-5.20386919e-02 -6.03071690e-01 4.65320796e-02 5.95016181e-01
3.60812962e-01 -4.69281822e-01 -1.31404805e+00 9.76564467e-01
-1.47899723e+00 -6.52160287e-01 -1.32920399e-01 1.72078001e+00
1.31862783e+00 -3.24085057e-01 2.06287980e-01 -1.94257975e-01
1.87297955e-01 -2.49377768e-02 -5.68642914e-01 -3.17745626e-01
-1.42048061e-01 4.85299945e-01 4.35455367e-02 7.90384054e-01
-5.22625566e-01 1.19340324e+00 5.86579943e+00 3.94714057e-01
-5.41535199e-01 -7.14980885e-02 2.75906444e-01 2.60999739e-01
-1.12273920e+00 2.48506501e-01 -5.37414789e-01 1.09405950e-01
9.88471448e-01 -5.94564497e-01 6.38780832e-01 6.72865927e-01
-6.93589896e-02 3.03704571e-02 -1.25015008e+00 6.57183886e-01
1.13324083e-01 -1.29366028e+00 6.04211390e-01 -3.68983686e-01
5.47305167e-01 -1.96314842e-01 -2.45303705e-01 7.19806254e-01
7.85748005e-01 -1.14032555e+00 4.37905520e-01 4.54212576e-01
5.74454069e-01 -2.90053278e-01 7.69475818e-01 6.56847239e-01
-9.16842937e-01 1.57256112e-01 -3.33232254e-01 -1.92469865e-01
1.85349956e-01 3.24172616e-01 -1.45484507e+00 7.77425230e-01
4.73480552e-01 -1.23487473e-01 -1.06397629e+00 6.91733837e-01
-5.87168872e-01 6.37490630e-01 -3.31712097e-01 -5.35830855e-01
4.19493943e-01 -7.74107641e-03 1.62659034e-01 8.79000127e-01
-9.44047142e-03 4.86155123e-01 1.17358312e-01 1.03895211e+00
-4.34179783e-01 3.33196878e-01 -8.54606926e-01 -5.71862400e-01
4.16839689e-01 1.10702884e+00 1.71815678e-02 -3.97150755e-01
-7.23583162e-01 7.15042770e-01 6.37228251e-01 6.78620100e-01
-3.23462397e-01 -4.46174651e-01 2.68462807e-01 3.28620225e-01
-1.90892354e-01 -1.54891714e-01 6.59619272e-02 -1.30032384e+00
4.07797903e-01 -1.40505779e+00 6.97195411e-01 -1.15832615e+00
-1.68924665e+00 6.33358240e-01 -2.73582507e-02 -5.28571308e-01
-6.92329884e-01 -5.33851445e-01 -2.64364868e-01 1.09740007e+00
-1.60951126e+00 -1.05007553e+00 -6.85934007e-01 8.81543636e-01
3.21953684e-01 2.21479222e-01 8.73747766e-01 2.18281314e-01
5.95651157e-02 6.33305550e-01 -6.30105376e-01 3.23461890e-01
7.66274452e-01 -1.52435875e+00 6.85951710e-01 6.38319433e-01
4.93249297e-01 8.23164642e-01 5.92885137e-01 -7.98397183e-01
-1.58423936e+00 -1.16683435e+00 1.17713332e+00 -1.20217991e+00
7.49977410e-01 -2.99043387e-01 -1.34277415e+00 6.78348005e-01
1.74931526e-01 -3.88741195e-02 8.82672250e-01 -3.81786400e-03
-6.10903502e-01 5.47291972e-02 -1.35506356e+00 5.93058705e-01
9.50669765e-01 -1.05790114e+00 -1.01813042e+00 5.54791033e-01
1.20197999e+00 -1.53480157e-01 -7.91374981e-01 1.77925438e-01
1.89547181e-01 -5.63629925e-01 6.86704159e-01 -1.20859098e+00
1.93121433e-01 -3.52898121e-01 -1.61944017e-01 -1.32572484e+00
1.20315507e-01 -6.50273740e-01 -6.02371395e-02 1.07881415e+00
8.18400502e-01 -6.88968241e-01 9.85001743e-01 1.40488243e+00
6.15561791e-02 -6.50511980e-01 -6.52210116e-01 -7.07124531e-01
1.64456859e-01 -5.05198419e-01 9.20510769e-01 1.06621337e+00
-4.64106724e-03 5.88607073e-01 2.53595769e-01 1.57645017e-01
2.41131023e-01 1.40007049e-01 1.12732053e+00 -9.49058175e-01
-2.06432343e-01 -2.54592896e-02 -5.61556667e-02 -1.00070918e+00
1.55823618e-01 -8.53328943e-01 -1.15746856e-02 -1.94932103e+00
-1.82675749e-01 -4.28135186e-01 3.77176434e-01 6.17875159e-01
-5.71573615e-01 -6.32434562e-02 8.29910040e-02 -3.36993150e-02
-7.31989324e-01 5.56810737e-01 1.52091956e+00 -2.38399222e-01
-1.59245819e-01 -3.77216265e-02 -1.13361967e+00 4.00346160e-01
7.49688923e-01 -1.98748425e-01 -1.00640583e+00 -8.17729533e-01
7.57203043e-01 1.46356776e-01 5.58136404e-01 -4.60912138e-01
4.20655400e-01 -3.06527227e-01 -2.27084860e-01 -3.04573923e-01
1.11035109e-01 -7.09740877e-01 -2.13572502e-01 8.25658292e-02
-4.85127628e-01 2.68000424e-01 1.46779150e-01 4.13180292e-01
-3.93385172e-01 -3.63548458e-01 1.61459774e-01 -4.03525084e-01
-6.88326359e-01 3.28834444e-01 -2.37588182e-01 9.74949419e-01
7.21231699e-01 -3.42001813e-03 -7.49936461e-01 -1.07189894e+00
-3.43613058e-01 7.24368572e-01 5.01849353e-01 2.63818830e-01
6.82484686e-01 -1.16642570e+00 -8.06644201e-01 1.22241152e-03
6.53418660e-01 2.96355069e-01 6.05120175e-02 2.72132009e-01
-7.45567739e-01 7.49999881e-01 1.29155949e-01 1.71421077e-02
-6.33437991e-01 4.89164591e-01 4.12889451e-01 -4.83644009e-01
7.09812641e-02 8.90981674e-01 -4.32303920e-02 -9.45182383e-01
-9.15999860e-02 -3.98517281e-01 -1.85909674e-01 -1.01155661e-01
5.02293229e-01 2.54129291e-01 1.97335437e-01 -2.73345381e-01
-2.54190922e-01 -1.42054493e-02 2.60645733e-03 -3.66726875e-01
8.68257046e-01 -1.36547638e-02 -1.69050947e-01 -1.10479789e-02
9.97085929e-01 3.22349906e-01 -5.41695595e-01 -6.02781117e-01
3.45578194e-01 -4.99525130e-01 -5.78785360e-01 -1.18548012e+00
-6.05829656e-01 6.94651306e-01 -1.45467490e-01 5.84385097e-01
9.69989300e-01 3.35751265e-01 9.63396430e-01 1.22863650e+00
6.21236384e-01 -6.45569563e-01 3.10794920e-01 5.33847868e-01
1.17109048e+00 -1.34588170e+00 -4.38628763e-01 -4.59657133e-01
-7.88271666e-01 7.71164238e-01 9.16738451e-01 2.87636042e-01
2.50207424e-01 -3.71590048e-01 5.95786154e-01 -6.32906675e-01
-7.22720921e-01 -3.86080652e-01 4.91267622e-01 8.16682279e-01
3.11567187e-01 -1.00926138e-01 2.17144517e-03 8.10294390e-01
-6.96163297e-01 -1.95350498e-02 5.65014541e-01 9.53004420e-01
-2.51493096e-01 -1.20119524e+00 -7.25216940e-02 5.57779849e-01
-2.65665442e-01 -3.70947510e-01 -1.11858666e+00 8.45399976e-01
-4.68301296e-01 1.49732840e+00 -5.62543809e-01 -1.24631211e-01
7.25775599e-01 2.38605693e-01 3.56491834e-01 -9.22335625e-01
-6.93261981e-01 -9.81822789e-01 6.39952481e-01 -6.18315101e-01
-1.86118245e-01 -1.02919869e-01 -1.05076218e+00 -1.74366534e-01
-2.07808122e-01 8.29507470e-01 3.27655375e-01 1.03926170e+00
3.78561884e-01 -9.90281850e-02 2.79021591e-01 2.12231636e-01
-8.28445077e-01 -8.99071157e-01 -2.13117033e-01 7.49184430e-01
2.19127819e-01 -1.36636436e-01 -2.55830586e-01 -3.28609273e-02] | [10.86252212524414, 7.929474830627441] |
86c7f048-7ed6-4b49-b392-e6389a9b4b21 | deep-structured-implicit-functions | 1912.06126 | null | https://arxiv.org/abs/1912.06126v2 | https://arxiv.org/pdf/1912.06126v2.pdf | Local Deep Implicit Functions for 3D Shape | The goal of this project is to learn a 3D shape representation that enables accurate surface reconstruction, compact storage, efficient computation, consistency for similar shapes, generalization across diverse shape categories, and inference from depth camera observations. Towards this end, we introduce Local Deep Implicit Functions (LDIF), a 3D shape representation that decomposes space into a structured set of learned implicit functions. We provide networks that infer the space decomposition and local deep implicit functions from a 3D mesh or posed depth image. During experiments, we find that it provides 10.3 points higher surface reconstruction accuracy (F-Score) than the state-of-the-art (OccNet), while requiring fewer than 1 percent of the network parameters. Experiments on posed depth image completion and generalization to unseen classes show 15.8 and 17.8 point improvements over the state-of-the-art, while producing a structured 3D representation for each input with consistency across diverse shape collections. | ['Thomas Funkhouser', 'Avneesh Sud', 'Aaron Sarna', 'Kyle Genova', 'Forrester Cole'] | 2019-12-12 | local-deep-implicit-functions-for-3d-shape | http://openaccess.thecvf.com/content_CVPR_2020/html/Genova_Local_Deep_Implicit_Functions_for_3D_Shape_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Genova_Local_Deep_Implicit_Functions_for_3D_Shape_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-shape-representation'] | ['computer-vision'] | [-1.38452798e-01 2.88274854e-01 2.15841740e-01 -5.19344449e-01
-9.07373130e-01 -6.75923824e-01 4.97725546e-01 -7.44296834e-02
9.71331894e-02 3.49015683e-01 1.83939099e-01 -1.51481032e-01
-8.46011490e-02 -1.02554977e+00 -1.03708339e+00 -3.23286533e-01
-2.30111957e-01 1.01264405e+00 1.60081148e-01 6.40755445e-02
2.30775595e-01 1.26290047e+00 -1.46020448e+00 5.57064652e-01
3.80091190e-01 1.55336499e+00 -7.00935125e-02 4.41804916e-01
-3.56360674e-01 1.19981885e-01 -1.73319057e-01 -5.32836700e-03
5.21214128e-01 5.16372740e-01 -8.31308961e-01 1.33560747e-01
1.14928246e+00 -7.12111473e-01 -1.74093261e-01 6.64384365e-01
4.76542622e-01 -2.86017098e-02 1.02814317e+00 -8.80774677e-01
-8.38387489e-01 -3.03980231e-01 -2.30585948e-01 -4.26649809e-01
2.04824716e-01 1.41321942e-01 8.13419342e-01 -1.46430123e+00
8.32252264e-01 1.62070060e+00 1.15319979e+00 5.88836610e-01
-1.45137227e+00 -3.96182626e-01 -2.29946431e-02 -6.47464097e-01
-1.52405500e+00 -5.46590507e-01 6.30676806e-01 -4.61505979e-01
1.45359349e+00 6.91477731e-02 5.56064010e-01 6.66691065e-01
1.33777410e-01 7.39708424e-01 4.95319456e-01 3.16685773e-02
4.63179678e-01 -4.20307875e-01 -2.06401050e-01 9.70864773e-01
4.37440753e-01 -1.55005500e-01 -3.66876096e-01 -3.43596429e-01
1.60017359e+00 9.72352400e-02 -7.91753381e-02 -8.06379437e-01
-8.37582409e-01 4.75167334e-01 5.91542721e-01 -1.40269071e-01
-4.12483066e-01 6.13861024e-01 1.87880263e-01 2.87858158e-01
9.50407922e-01 2.87491739e-01 -8.05801213e-01 -9.44390241e-03
-5.59811831e-01 6.06334329e-01 8.61565471e-01 1.14903092e+00
9.46234882e-01 3.49498034e-01 4.18381274e-01 7.90776134e-01
4.35499549e-01 9.02107239e-01 -1.87485993e-01 -1.36583257e+00
4.20408607e-01 9.86131489e-01 3.54855508e-02 -1.11775410e+00
-4.20287222e-01 -2.38838330e-01 -8.65779161e-01 7.12126672e-01
1.37452900e-01 2.43692204e-01 -1.21985769e+00 1.55571663e+00
3.25138807e-01 2.02038124e-01 -6.95448518e-02 7.51587093e-01
1.09752834e+00 5.47627568e-01 -2.93564439e-01 6.52765632e-01
8.35688829e-01 -3.84265423e-01 -7.73085579e-02 -5.46429604e-02
4.49982166e-01 -3.62584174e-01 1.01920235e+00 3.79057735e-01
-1.51406538e+00 -4.79642749e-01 -9.47765291e-01 -5.17771184e-01
-2.07947984e-01 -1.48883108e-02 7.24501848e-01 1.42528996e-01
-1.55266988e+00 7.77846515e-01 -1.16237044e+00 -1.26540571e-01
8.18796098e-01 4.97938514e-01 -5.24324715e-01 -1.33279651e-01
-2.99388528e-01 4.98281777e-01 -1.89608514e-01 -4.92285967e-01
-1.12689304e+00 -1.24701583e+00 -1.09513581e+00 -3.92356887e-02
-1.55941337e-01 -1.08549809e+00 1.27413464e+00 -5.53016722e-01
-1.34337676e+00 1.20044386e+00 -2.70304382e-01 -2.24842459e-01
2.24945724e-01 -3.09229702e-01 3.35369587e-01 3.58339906e-01
-1.08866319e-02 9.78406608e-01 8.25036943e-01 -1.31834447e+00
-1.37507945e-01 -7.24915028e-01 -4.53749532e-03 2.11019531e-01
4.44334969e-02 -6.50908947e-01 -3.37094098e-01 -6.19838655e-01
7.78024793e-01 -6.97938979e-01 -4.31817165e-03 1.02202344e+00
6.22670837e-02 -4.79211956e-01 8.79068732e-01 -5.83164692e-01
2.91530520e-01 -2.13301826e+00 2.10347101e-01 1.98119462e-01
4.57755923e-01 -1.44941881e-01 -4.56444919e-01 1.45765409e-01
8.67626220e-02 2.37298444e-01 -5.30858397e-01 -8.99163067e-01
7.72994012e-02 4.46351111e-01 -4.99597818e-01 6.27837837e-01
4.76633102e-01 1.13379455e+00 -4.47147280e-01 4.93636131e-02
4.50843498e-02 8.21886361e-01 -8.71013224e-01 2.39407435e-01
-4.38438714e-01 9.03305486e-02 -3.52783114e-01 1.08753550e+00
9.75912690e-01 -4.27057773e-01 -7.28954598e-02 -2.77335346e-01
2.05946818e-01 5.81075370e-01 -1.12534010e+00 2.19944072e+00
-3.74446243e-01 3.99494797e-01 5.31915545e-01 -8.04586768e-01
1.36706829e+00 2.30545118e-01 5.80114484e-01 -5.48376501e-01
-4.26987112e-02 4.70711619e-01 -6.99066997e-01 -1.10667124e-01
2.49377698e-01 -4.99377064e-02 2.21493527e-01 5.32741427e-01
3.55413333e-02 -7.41505861e-01 -6.80340111e-01 -1.65964052e-01
1.11936986e+00 2.96439737e-01 -2.46050939e-01 -6.01016223e-01
1.35137752e-01 -1.72063172e-01 2.85457909e-01 3.21361810e-01
3.05003762e-01 9.78322685e-01 5.44541515e-02 -1.09436071e+00
-1.29217720e+00 -1.72085977e+00 -2.13701233e-01 6.16136074e-01
1.39790609e-01 -1.46398365e-01 -5.08107424e-01 -2.64852762e-01
6.73170567e-01 3.92733127e-01 -6.66571021e-01 1.95376277e-01
-7.65301466e-01 1.01556994e-01 5.91119468e-01 1.00708318e+00
4.23745841e-01 -7.99494505e-01 -4.63459551e-01 -8.10807124e-02
2.95297086e-01 -9.34974849e-01 -3.01320165e-01 -1.61255404e-01
-1.72101986e+00 -1.08128846e+00 -5.69200456e-01 -1.03251767e+00
9.46417689e-01 2.06533566e-01 1.49255431e+00 2.11281657e-01
-2.14892268e-01 6.58938646e-01 9.18061361e-02 -2.35100359e-01
-1.18248262e-01 -1.36897098e-02 1.41628787e-01 -4.45277870e-01
1.24663673e-01 -1.16238511e+00 -5.50155878e-01 1.45592362e-01
-8.51563454e-01 9.72466990e-02 3.11995655e-01 6.87284648e-01
1.04982746e+00 -3.83837640e-01 3.94172132e-01 -4.96799231e-01
2.10139677e-01 -2.23221824e-01 -6.44176424e-01 -1.26914904e-01
-3.37655365e-01 2.90601522e-01 3.63184363e-01 -1.89801171e-01
-8.99645567e-01 1.67870417e-01 -3.29042792e-01 -9.15063679e-01
-3.88308048e-01 1.25192434e-01 -3.49909663e-02 -2.13189453e-01
8.06713343e-01 3.28976154e-01 4.33217198e-01 -9.74005699e-01
3.35692197e-01 3.60269755e-01 5.55184603e-01 -1.10343516e+00
6.61979854e-01 9.78772163e-01 1.77150175e-01 -8.66984308e-01
-6.48815572e-01 -1.56238198e-01 -7.99867511e-01 2.83879042e-01
6.58027947e-01 -1.15513706e+00 -7.50617266e-01 5.85090458e-01
-1.51908505e+00 -7.78282464e-01 -4.65000004e-01 -1.40357062e-01
-8.64403248e-01 3.51183340e-02 -7.29704261e-01 -6.85368061e-01
-7.25641847e-01 -9.63454485e-01 1.80945575e+00 -2.99741507e-01
-3.15624237e-01 -1.03951931e+00 -6.62603080e-02 1.46722514e-02
4.55279768e-01 5.71559489e-01 9.99319255e-01 -1.38999641e-01
-1.03726614e+00 -2.22857241e-02 -4.87773538e-01 2.62466669e-01
2.08979547e-01 -3.07148337e-01 -9.80378687e-01 -4.43282157e-01
-1.32010221e-01 -5.12056649e-01 7.62262940e-01 4.64162946e-01
1.53714550e+00 -4.29610193e-01 -1.70598388e-01 1.25414777e+00
1.47241116e+00 -1.82019055e-01 5.66266000e-01 -4.22770530e-02
5.63192129e-01 3.45889956e-01 9.54941288e-03 3.85009050e-01
5.08062720e-01 2.66864479e-01 8.18120360e-01 2.53170319e-02
-4.74493921e-01 -4.18603987e-01 5.63469119e-02 6.66456580e-01
-1.43180981e-01 2.00079814e-01 -1.04917741e+00 7.46381819e-01
-1.53059638e+00 -4.94183064e-01 1.20683707e-01 2.10449314e+00
7.72723615e-01 1.26476049e-01 -1.11493148e-01 -2.07308367e-01
1.59523338e-01 3.00705209e-02 -8.96429718e-01 -3.95990521e-01
-1.37257367e-01 6.48968816e-01 2.99871802e-01 6.92897558e-01
-8.30479085e-01 7.27038622e-01 6.85166025e+00 5.82565069e-01
-1.11854589e+00 -1.82089031e-01 6.71809316e-01 -1.56473629e-02
-8.42776239e-01 -5.09760797e-01 -5.88782966e-01 -1.62279963e-01
4.74480271e-01 1.56303450e-01 5.30282855e-01 1.00983405e+00
-2.89864093e-01 4.02584165e-01 -1.41988266e+00 1.36509454e+00
7.35957697e-02 -1.86115348e+00 4.00194049e-01 1.71212316e-01
9.58531976e-01 5.28424919e-01 -1.44500686e-02 5.97787201e-02
2.33228385e-01 -1.35787868e+00 8.48175466e-01 6.72178328e-01
1.46648240e+00 -5.59997380e-01 3.07021320e-01 5.23599803e-01
-1.39026034e+00 2.79654026e-01 -6.69648409e-01 -2.99019396e-01
-1.25043541e-01 5.80804884e-01 -7.81596959e-01 1.66480362e-01
8.62454772e-01 9.64980483e-01 -1.26398236e-01 6.44300342e-01
8.01950395e-02 2.31664211e-01 -7.99291253e-01 1.78242818e-01
3.84166650e-03 -1.73810683e-02 5.86088598e-01 8.89403224e-01
4.24058020e-01 5.02711177e-01 1.09613895e-01 1.39505506e+00
-5.33218920e-01 -3.67109954e-01 -1.05949366e+00 1.68184936e-01
5.39924204e-01 7.37306178e-01 -3.45651716e-01 -2.61086822e-01
-1.74331684e-02 8.56011987e-01 6.22947454e-01 3.56543839e-01
-1.57606930e-01 -1.60418212e-01 1.22031116e+00 4.43951517e-01
4.69359428e-01 -7.30130732e-01 -1.15432549e+00 -9.07091618e-01
9.19130743e-02 -3.44266087e-01 -1.48092778e-02 -9.57972467e-01
-1.60433996e+00 3.82385105e-01 -5.05911000e-02 -1.03344309e+00
2.82479338e-02 -1.06420624e+00 -3.46491635e-01 9.50613737e-01
-1.32138014e+00 -1.15182447e+00 -4.08142120e-01 3.42573076e-01
6.21248126e-01 -1.26746103e-01 1.06279719e+00 1.27846822e-01
3.19746077e-01 5.50768971e-01 -1.61906451e-01 2.61545181e-01
6.36026934e-02 -1.17980528e+00 1.11226368e+00 1.77531168e-01
-1.07396595e-01 6.53626919e-01 2.04492304e-02 -7.83402085e-01
-1.98816311e+00 -1.12776542e+00 5.35100996e-01 -7.71622002e-01
1.19778223e-01 -3.91711801e-01 -1.00505841e+00 7.42158413e-01
-3.27117622e-01 3.45551282e-01 2.34928891e-01 -1.77568775e-02
-7.51877606e-01 -1.26185685e-01 -1.56751800e+00 4.71126795e-01
1.68859291e+00 -8.04583311e-01 -6.89436018e-01 1.66923143e-02
9.48247790e-01 -9.55338478e-01 -1.15319586e+00 7.71678507e-01
7.28194416e-01 -7.65742242e-01 1.47149861e+00 -6.39720142e-01
6.01031125e-01 -1.40115663e-01 -6.81924164e-01 -1.00470722e+00
-4.67964709e-01 -2.04736352e-01 -5.38129091e-01 4.95133221e-01
2.44742140e-01 -6.29519701e-01 1.18900502e+00 9.27509367e-01
-6.61310613e-01 -1.28843069e+00 -1.00953174e+00 -7.74792194e-01
4.56530362e-01 -5.88335693e-01 8.60235333e-01 8.45993817e-01
-6.19434536e-01 3.18519920e-02 3.62605184e-01 2.81544477e-01
8.82791936e-01 4.55056489e-01 1.00650930e+00 -1.57458484e+00
6.42175898e-02 -5.46791792e-01 -5.30977070e-01 -1.75844872e+00
1.85320705e-01 -1.17348826e+00 -9.88980681e-02 -1.91227007e+00
-2.23490074e-01 -8.22089314e-01 2.68196315e-01 8.25932264e-01
6.32124901e-01 4.24060524e-01 -8.33918154e-02 3.73823643e-02
-2.99181968e-01 8.44261467e-01 1.48881876e+00 -2.89246142e-01
-1.24000475e-01 -3.56201530e-01 -4.76537675e-01 8.35306585e-01
7.19126225e-01 -2.12371588e-01 -4.56739008e-01 -1.25560260e+00
2.47552887e-01 -1.37057770e-02 5.84639490e-01 -8.27995121e-01
9.43947434e-02 -1.83058530e-01 6.70164168e-01 -1.02700210e+00
9.00811672e-01 -8.02265108e-01 1.65704295e-01 2.47716412e-01
-5.95854260e-02 6.54973239e-02 8.05200160e-01 3.43554199e-01
1.22282669e-01 2.25046903e-01 6.66089296e-01 -3.04683894e-01
-4.39465195e-01 7.39064872e-01 1.81161180e-01 2.29921997e-01
4.65762079e-01 -5.91485858e-01 -9.64700356e-02 -2.11584389e-01
-6.67096615e-01 -1.58554316e-01 8.05267632e-01 3.26180458e-01
1.27336037e+00 -1.75706697e+00 -8.67076695e-01 7.16705799e-01
-1.55898169e-01 1.02889144e+00 3.58394012e-02 1.27162248e-01
-9.97466803e-01 2.28767753e-01 -1.03639498e-01 -1.03518450e+00
-7.50490129e-01 -2.44087666e-01 5.94134271e-01 2.00864553e-01
-1.04465055e+00 1.11984289e+00 4.31171119e-01 -1.05702794e+00
3.86032254e-01 -8.59741509e-01 6.61620080e-01 -5.69958389e-01
3.76936823e-01 3.93908352e-01 3.33897650e-01 -3.11481476e-01
-3.57341796e-01 1.02243650e+00 4.42568958e-01 7.41675794e-02
1.77751243e+00 2.64176428e-01 -2.52556711e-01 2.99739808e-01
1.51639009e+00 -2.16550201e-01 -1.67491865e+00 -2.55527228e-01
-4.24478889e-01 -5.73326051e-01 9.54872146e-02 -7.13577449e-01
-1.09122074e+00 9.34208930e-01 4.28020507e-01 -9.41504762e-02
7.69165635e-01 1.89995393e-01 1.00814116e+00 7.58666992e-01
7.51246095e-01 -5.61326027e-01 1.87753275e-01 9.41707373e-01
1.59003031e+00 -9.45580840e-01 6.93525001e-02 -4.21984375e-01
5.93939424e-02 1.26699579e+00 6.43198907e-01 -7.10152447e-01
1.07172072e+00 5.01533329e-01 -2.10294977e-01 -6.99404657e-01
-7.67776251e-01 3.48080188e-01 5.95598161e-01 7.53943682e-01
1.07107691e-01 6.70196638e-02 6.55698895e-01 3.96607071e-01
-1.44777507e-01 -7.58801103e-02 -7.30136186e-02 7.49033093e-01
-4.88596171e-01 -6.80377007e-01 -2.77845472e-01 4.79128867e-01
9.02917236e-02 9.58823413e-02 -4.60434496e-01 6.39006853e-01
-1.72697648e-01 1.49468601e-01 6.03771925e-01 -1.42962664e-01
5.59553206e-01 -1.20717529e-02 7.58699894e-01 -7.62878895e-01
-2.53247797e-01 -2.13513196e-01 2.83539426e-02 -9.14955735e-01
-5.67545369e-02 -5.25625348e-01 -1.75646687e+00 -5.64815938e-01
2.67983496e-01 -4.88512069e-01 7.15307713e-01 5.89730918e-01
9.18124318e-01 1.12621091e-01 3.59766603e-01 -1.43977594e+00
-5.95853508e-01 -7.34985352e-01 -4.52427715e-01 5.69469929e-01
5.36228836e-01 -7.96586037e-01 -3.01883578e-01 -9.54437163e-03] | [8.587191581726074, -3.657292127609253] |
d90c634f-b752-4fb8-88d2-c2b637a3217b | procc-progressive-cross-primitive-consistency | 2211.12417 | null | https://arxiv.org/abs/2211.12417v3 | https://arxiv.org/pdf/2211.12417v3.pdf | ProCC: Progressive Cross-primitive Compatibility for Open-World Compositional Zero-Shot Learning | Open-World Compositional Zero-shot Learning (OW-CZSL) aims to recognize novel compositions of state and object primitives in images with no priors on the compositional space, which induces a tremendously large output space containing all possible state-object compositions. Existing works either learn the joint compositional state-object embedding or predict simple primitives with separate classifiers. However, the former heavily relies on external word embedding methods, and the latter ignores the interactions of interdependent primitives, respectively. In this paper, we revisit the primitive prediction approach and propose a novel method, termed Progressive Cross-primitive Compatibility (ProCC), to mimic the human learning process for OW-CZSL tasks. Specifically, the cross-primitive compatibility module explicitly learns to model the interactions of state and object features with the trainable memory units, which efficiently acquires cross-primitive visual attention to reason high-feasibility compositions, without the aid of external knowledge. Moreover, considering the partial-supervision setting (pCZSL) as well as the imbalance issue of multiple task prediction, we design a progressive training paradigm to enable the primitive classifiers to interact to obtain discriminative information in an easy-to-hard manner. Extensive experiments on three widely used benchmark datasets demonstrate that our method outperforms other representative methods on both OW-CZSL and pCZSL settings by large margins. | ['Ziming Liu', 'Haozhao Wang', 'Jingcai Guo', 'Song Guo', 'Wenchao Xu', 'Fushuo Huo'] | 2022-11-19 | null | null | null | null | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 3.73754889e-01 -4.12082709e-02 -2.81046420e-01 -1.38835430e-01
-4.97406632e-01 -2.23908067e-01 8.86118054e-01 -1.57947317e-01
-3.29605162e-01 2.01696187e-01 1.97847918e-01 7.26595595e-02
2.23454952e-01 -6.50033474e-01 -8.74574542e-01 -9.69326437e-01
3.26216787e-01 4.77683574e-01 6.64456487e-01 -1.23036131e-01
7.87820965e-02 1.52222797e-01 -1.67572403e+00 3.43940109e-01
7.02203751e-01 1.11697924e+00 4.75670993e-01 4.06996161e-01
-1.59276247e-01 9.46277142e-01 -1.56428292e-01 -5.50188959e-01
4.28601742e-01 -2.14240462e-01 -5.20493925e-01 2.61812270e-01
4.94718283e-01 -4.32410389e-01 -4.23047841e-01 1.05375147e+00
3.93574893e-01 2.50146478e-01 7.02042878e-01 -1.30862331e+00
-1.03831768e+00 4.23037291e-01 -4.27878648e-01 -3.32801677e-02
1.52344912e-01 7.84781694e-01 1.38780963e+00 -9.75293040e-01
7.08121717e-01 1.27430439e+00 2.43889317e-01 7.04571843e-01
-1.49106216e+00 -7.50611007e-01 6.96835756e-01 6.67484999e-01
-1.33764553e+00 -3.87131810e-01 1.01239157e+00 -6.41349435e-01
9.60791171e-01 7.23962784e-02 7.63591051e-01 1.37946558e+00
-3.34546156e-02 1.21365690e+00 9.51267958e-01 -4.36634608e-02
3.27073097e-01 -4.59382199e-02 1.65283740e-01 9.71628010e-01
-6.42273203e-02 7.82541409e-02 -6.62123382e-01 -1.08786009e-01
6.66521072e-01 3.03245783e-01 -2.83654630e-01 -7.06347167e-01
-1.36522114e+00 4.73822623e-01 4.31160688e-01 4.89123203e-02
-2.02508867e-01 2.13732585e-01 3.95554125e-01 2.92528898e-01
1.60680920e-01 2.82141447e-01 -4.54971194e-01 2.35608831e-01
-4.27892029e-01 1.48196831e-01 5.42426586e-01 1.12220919e+00
9.71257687e-01 1.43313035e-01 -4.02648956e-01 5.52795053e-01
2.91989744e-01 3.82119715e-01 7.74788380e-01 -6.14352226e-01
5.26322782e-01 7.80969083e-01 -1.93063483e-01 -5.47190547e-01
7.87190273e-02 -2.38647938e-01 -8.80434632e-01 1.38152540e-01
2.26134006e-02 2.39613205e-01 -1.00538826e+00 1.96639061e+00
4.13199425e-01 6.59915030e-01 -4.01876234e-02 7.84710586e-01
6.42417908e-01 8.87382567e-01 2.84505248e-01 -1.78899601e-01
1.34350443e+00 -1.34241652e+00 -4.64068741e-01 -2.85182714e-01
3.12735349e-01 -3.91387612e-01 1.54260993e+00 1.17585607e-01
-7.61105001e-01 -7.65770137e-01 -1.15043175e+00 -2.71784365e-01
-2.95731992e-01 -1.15418687e-01 4.13399935e-01 3.67809907e-02
-6.42671824e-01 5.59133053e-01 -9.86427188e-01 1.26036033e-01
6.10461712e-01 3.23770672e-01 -3.36258441e-01 -1.20914802e-01
-8.28492701e-01 7.95415878e-01 5.48313558e-01 -2.06822649e-01
-1.26649320e+00 -7.91330159e-01 -1.10558128e+00 2.58501738e-01
8.35040510e-01 -8.32945287e-01 1.15775931e+00 -1.01275837e+00
-1.71479595e+00 4.56211716e-01 -9.99497548e-02 -1.74541011e-01
4.09214586e-01 -1.35062203e-01 -6.55527934e-02 4.17365171e-02
-5.26318476e-02 7.08348930e-01 1.25420940e+00 -1.31782150e+00
-5.71864784e-01 -2.65418530e-01 6.44005910e-02 3.32309425e-01
-2.79106379e-01 -1.71606466e-01 -5.90893447e-01 -7.20581532e-01
2.70934254e-02 -7.95757890e-01 -1.37380794e-01 4.46272463e-01
-2.78721243e-01 -6.55284047e-01 8.17480206e-01 -4.05417562e-01
8.19902778e-01 -2.14376354e+00 5.72377920e-01 -2.67488271e-01
2.49581471e-01 1.96243793e-01 -3.53520632e-01 2.97855705e-01
4.06148359e-02 -3.41017425e-01 -3.53166401e-01 -6.42321646e-01
2.08421707e-01 5.74388862e-01 -6.07352853e-01 3.78464758e-01
5.39399803e-01 1.21406710e+00 -1.06462276e+00 -6.20995402e-01
2.83260375e-01 2.25410044e-01 -6.01065278e-01 3.89785916e-01
-6.04320586e-01 3.41234744e-01 -2.49614522e-01 6.17497265e-01
3.13083261e-01 -4.15268660e-01 8.23570862e-02 -3.43941540e-01
-1.86737943e-02 1.56134456e-01 -9.55037355e-01 1.91575241e+00
-3.95217419e-01 4.28588502e-02 -3.71636331e-01 -1.04841709e+00
5.36237717e-01 2.94817299e-01 2.63446003e-01 -5.77174485e-01
1.43341790e-03 -1.62501670e-02 1.40253186e-01 -4.79798108e-01
-3.91914807e-02 -4.29823250e-01 -7.03051984e-02 5.61896324e-01
6.10074341e-01 1.73719898e-02 7.82582164e-03 1.83031276e-01
8.62323940e-01 4.71142411e-01 4.36085701e-01 -1.24393836e-01
5.33240557e-01 -4.90236878e-01 9.06455636e-01 4.66548055e-01
-3.44823420e-01 4.77198422e-01 4.83879894e-01 -4.60169315e-01
-8.93245399e-01 -1.26584494e+00 2.87671268e-01 1.36671770e+00
6.26841545e-01 -5.27373612e-01 -3.07824075e-01 -8.70973408e-01
-1.45039797e-01 5.82452774e-01 -6.85864270e-01 -4.58219171e-01
-6.89823985e-01 -4.43417728e-01 1.05497487e-01 5.95467567e-01
2.21368864e-01 -1.27250206e+00 -6.96759999e-01 1.27905309e-01
-3.44831012e-02 -1.09211481e+00 -7.42051482e-01 1.79263726e-01
-6.27621412e-01 -1.00800550e+00 -6.77216947e-01 -1.12128019e+00
7.47578979e-01 1.62948236e-01 6.53887987e-01 -9.44876447e-02
-3.69680047e-01 1.83083773e-01 -2.14280233e-01 3.13612516e-03
-4.21961099e-02 -2.08976761e-01 1.11229658e-01 5.69278300e-01
1.56711310e-01 -7.58209467e-01 -5.82939506e-01 4.43603061e-02
-9.15320992e-01 6.83003187e-01 8.68457019e-01 1.12182951e+00
7.26604581e-01 -2.20724329e-01 2.42941752e-01 -6.93168283e-01
1.72414869e-01 -3.42681378e-01 -4.61799204e-01 5.44394195e-01
-5.32507002e-01 5.02834022e-01 8.80191624e-01 -1.00734544e+00
-1.03798473e+00 3.73340786e-01 6.78391233e-02 -1.08086038e+00
-8.30655023e-02 1.42150715e-01 -5.85042953e-01 9.64725539e-02
2.37165362e-01 7.67149448e-01 -1.52487859e-01 -6.31480336e-01
5.96477091e-01 1.80360913e-01 6.76221848e-01 -8.47025394e-01
1.08969593e+00 5.20684958e-01 -2.51833677e-01 -5.68544686e-01
-8.70145798e-01 -4.14397597e-01 -7.17140853e-01 1.37936115e-01
9.53806877e-01 -1.00417686e+00 -5.73026478e-01 6.28585637e-01
-1.19734037e+00 -5.31903148e-01 -5.25299847e-01 2.03916177e-01
-6.40790939e-01 5.62628865e-01 -8.46400261e-01 -6.05096638e-01
-3.12473923e-01 -1.46658790e+00 1.16207969e+00 5.72007224e-02
-6.66149184e-02 -5.57546377e-01 3.60522307e-02 2.06724137e-01
2.82432791e-02 -2.78083384e-02 1.34336746e+00 -4.95054662e-01
-9.28538203e-01 6.96034804e-02 -2.96599418e-01 3.41312945e-01
1.15013778e-01 -1.48977056e-01 -8.04765046e-01 -2.40168780e-01
1.60004854e-01 -5.58926284e-01 8.73972058e-01 -2.19228268e-01
1.34984684e+00 -5.17050564e-01 -5.82348742e-02 5.68698943e-01
1.21236622e+00 6.21884614e-02 3.71289134e-01 -5.74714541e-02
1.02492213e+00 6.02083683e-01 3.50720227e-01 3.98327351e-01
5.86746097e-01 7.37975836e-01 4.40827996e-01 2.71598637e-01
-3.15753490e-01 -6.13752365e-01 5.61225891e-01 1.10984671e+00
-3.43988463e-02 2.05396175e-01 -6.27542973e-01 4.38039154e-01
-1.97554588e+00 -8.31592739e-01 5.16310871e-01 1.91090870e+00
1.07774079e+00 1.07897431e-01 -1.71275333e-01 -1.61651954e-01
5.08142292e-01 4.13908273e-01 -8.39783490e-01 2.32635960e-01
1.08223766e-01 3.57206762e-01 2.69784778e-02 3.98397803e-01
-1.01003623e+00 1.07474852e+00 4.69754171e+00 9.73930240e-01
-1.12349701e+00 4.15082157e-01 3.74566942e-01 -2.74995893e-01
-3.16538543e-01 2.70735741e-01 -6.71581089e-01 6.37285352e-01
3.26709598e-01 -6.37865290e-02 5.07124722e-01 8.59882176e-01
-4.62226927e-01 2.19123855e-01 -1.44507456e+00 1.03172898e+00
2.25191474e-01 -1.14267766e+00 4.78937626e-01 -1.32787645e-01
7.29797184e-01 -1.74369477e-02 1.38090923e-01 6.78002179e-01
3.03356290e-01 -4.99862969e-01 1.02562320e+00 4.04929876e-01
7.76009202e-01 -3.09681743e-01 2.12359071e-01 5.87432086e-01
-1.52068543e+00 -2.88123488e-01 -3.47218871e-01 -1.05816290e-01
3.04443538e-01 -6.75292686e-02 -2.81867862e-01 4.41803694e-01
4.49787438e-01 9.09154952e-01 -6.03487968e-01 7.54080892e-01
-5.70857167e-01 2.92222440e-01 -2.04181015e-01 2.24536024e-02
2.13287696e-01 -1.34509876e-01 4.66583490e-01 7.01200485e-01
5.29900007e-02 1.54083639e-01 4.22702193e-01 1.04565465e+00
-6.22170009e-02 1.68757122e-02 -2.71943778e-01 8.66645575e-03
2.03311980e-01 1.01115370e+00 -4.32949394e-01 -7.42324769e-01
-6.36472464e-01 1.12819123e+00 7.20079422e-01 3.50011617e-01
-8.37924004e-01 -7.63299409e-03 7.31645405e-01 -1.17875412e-01
5.67251205e-01 -1.70674056e-01 -1.48227260e-01 -1.72676122e+00
2.01243162e-01 -9.28381741e-01 3.42316419e-01 -5.21027863e-01
-1.48645365e+00 4.13541466e-01 9.72169191e-02 -1.32232893e+00
1.27131298e-01 -6.12535298e-01 -8.77510607e-01 5.18127918e-01
-1.46945941e+00 -1.62630332e+00 -6.74377754e-02 7.49211073e-01
9.99669552e-01 -1.38100296e-01 6.75351858e-01 1.70322224e-01
-8.72118950e-01 5.18828690e-01 -2.82496274e-01 -8.80073830e-02
5.91322243e-01 -1.12406933e+00 4.46312129e-01 8.84127557e-01
3.60935956e-01 6.63819373e-01 4.20109302e-01 -6.93523943e-01
-1.61641860e+00 -1.21454084e+00 6.32610857e-01 -4.60462093e-01
9.39998746e-01 -8.09225738e-01 -1.01920629e+00 7.87201583e-01
6.30694553e-02 3.59630227e-01 6.08860672e-01 -7.07171112e-02
-9.83166158e-01 -1.12746611e-01 -4.10231352e-01 1.04658246e+00
1.35850346e+00 -8.75241280e-01 -9.77582574e-01 2.24152833e-01
1.05150580e+00 -9.64700989e-03 -5.31120062e-01 5.45857608e-01
5.32365024e-01 -5.87830365e-01 9.72182751e-01 -6.95536852e-01
4.63962734e-01 -4.12197500e-01 -1.84496850e-01 -9.59190488e-01
-4.65699196e-01 -4.06346053e-01 -8.22549284e-01 9.02418017e-01
1.12383634e-01 -5.90915918e-01 5.34646869e-01 6.75195038e-01
-2.72058338e-01 -1.15910542e+00 -9.31210220e-01 -8.48993242e-01
-1.09654680e-01 -2.47934431e-01 5.51277936e-01 8.52481544e-01
-1.45182431e-01 7.16142237e-01 -5.40046871e-01 1.80095524e-01
6.44405663e-01 6.41298115e-01 7.86155105e-01 -9.98302042e-01
-8.98681521e-01 -6.14730477e-01 -5.30884147e-01 -1.30968094e+00
5.00207722e-01 -8.93883586e-01 2.91636467e-01 -1.12008286e+00
4.31335598e-01 -4.06985998e-01 -5.44690192e-01 7.73333430e-01
-5.77645361e-01 -1.43821696e-02 5.06038487e-01 5.46321869e-01
-9.22384679e-01 1.28185964e+00 1.37908888e+00 -5.64311683e-01
2.15075128e-02 -2.19244331e-01 -5.51580369e-01 6.85190737e-01
2.70331651e-01 -3.77333194e-01 -7.16479361e-01 -5.48335195e-01
-1.59027398e-01 -2.69240469e-01 4.41756129e-01 -9.73234057e-01
2.41783291e-01 -3.59971642e-01 1.22797154e-01 -4.33227479e-01
7.17022359e-01 -8.47955108e-01 3.37079726e-02 6.59201205e-01
-2.74919957e-01 -2.81614453e-01 -2.83076793e-01 1.01634300e+00
-7.44364038e-02 -9.98653471e-02 7.49459803e-01 -1.77300319e-01
-1.00736845e+00 8.70090842e-01 -4.23731320e-02 -1.85450345e-01
1.31758046e+00 2.07085703e-02 -4.35309291e-01 1.10178389e-01
-5.84488571e-01 3.57706755e-01 6.08762920e-01 6.20581925e-01
6.21334255e-01 -1.42515719e+00 -2.30260387e-01 4.59029824e-01
5.80249608e-01 3.71332675e-01 4.55417395e-01 8.34414780e-01
-1.73958287e-01 2.41377633e-02 -2.09485397e-01 -5.27837276e-01
-9.76220667e-01 1.13316405e+00 7.54915699e-02 -2.44345486e-01
-1.06126523e+00 1.07816207e+00 9.92302060e-01 -3.28746766e-01
3.40225220e-01 -3.55812252e-01 1.01966701e-01 -1.95522443e-03
5.12305498e-01 -2.52169725e-02 -4.92658287e-01 -7.10771382e-01
-9.15777609e-02 5.42309582e-01 -3.24415535e-01 1.63058303e-02
1.21124876e+00 6.91145286e-02 -1.06458038e-01 8.20976079e-01
1.22130132e+00 -4.87855822e-01 -1.83752036e+00 -7.95078218e-01
7.99285546e-02 -3.05907190e-01 -3.33870381e-01 -2.82864332e-01
-8.93592596e-01 1.16539025e+00 3.57606351e-01 -3.87163788e-01
8.87938976e-01 1.76747650e-01 9.37208176e-01 5.61620116e-01
4.12845373e-01 -8.46481800e-01 6.17851019e-01 4.39965069e-01
8.22034001e-01 -1.24736345e+00 -1.57851174e-01 -2.97975183e-01
-8.70887697e-01 7.50229359e-01 1.13950598e+00 -2.14880958e-01
6.29052699e-01 -4.41229865e-02 -2.84200281e-01 -7.87130892e-02
-1.24148333e+00 -2.54678339e-01 3.11710864e-01 3.40338707e-01
-2.38364056e-01 4.90501011e-03 5.29297180e-02 7.82575905e-01
2.27730334e-01 -2.01223999e-01 -7.34961703e-02 9.98970270e-01
-2.71160454e-01 -1.16914976e+00 2.27735043e-01 4.71315145e-01
8.40563253e-02 -1.81767628e-01 -2.40315393e-01 3.33170980e-01
4.01748836e-01 1.53508365e-01 1.78667568e-02 -3.98062140e-01
5.51903956e-02 3.76341343e-01 5.38723826e-01 -9.53943551e-01
-3.40610266e-01 4.52040359e-02 -4.14715797e-01 -6.38797879e-01
-2.76741207e-01 -4.59657878e-01 -1.10955215e+00 2.11629987e-01
-3.26063633e-01 -3.44115078e-01 1.58552140e-01 1.20575941e+00
3.13127130e-01 5.10078669e-01 4.88404900e-01 -1.16880226e+00
-8.01236629e-01 -8.43839645e-01 -3.67793471e-01 7.16189623e-01
3.88804018e-01 -8.88605714e-01 -1.60012409e-01 2.04660982e-01] | [10.233449935913086, 2.2483973503112793] |
e8b754fe-18eb-4945-853d-937cb182b12e | prepositional-phrase-attachment-over-word | null | null | https://aclanthology.org/W17-6305 | https://aclanthology.org/W17-6305.pdf | Prepositional Phrase Attachment over Word Embedding Products | We present a low-rank multi-linear model for the task of solving prepositional phrase attachment ambiguity (PP task). Our model exploits tensor products of word embeddings, capturing all possible conjunctions of latent embeddings. Our results on a wide range of datasets and task settings show that tensor products are the best compositional operation and that a relatively simple multi-linear model that uses only word embeddings of lexical features can outperform more complex non-linear architectures that exploit the same information. Our proposed model gives the current best reported performance on an out-of-domain evaluation and performs competively on out-of-domain dependency parsing datasets. | ['Pranava Swaroop Madhyastha', 'Xavier Carreras', 'Ariadna Quattoni'] | 2017-09-01 | null | null | null | ws-2017-9 | ['prepositional-phrase-attachment'] | ['natural-language-processing'] | [-2.10105851e-01 7.52873719e-02 -4.55047131e-01 -5.76492906e-01
-9.58504200e-01 -7.77404785e-01 3.48186195e-01 3.42537940e-01
-8.94572020e-01 4.70102668e-01 5.39677858e-01 -5.29711246e-01
-1.25131279e-01 -3.10623497e-01 -4.93099719e-01 -3.74976605e-01
-3.72099608e-01 9.41983640e-01 3.12492430e-01 -3.80490601e-01
1.80295870e-01 2.57255256e-01 -7.72641659e-01 6.63580298e-01
2.86490560e-01 7.23707378e-01 2.05739856e-01 4.78651762e-01
-5.49623549e-01 2.78644919e-01 -2.95452416e-01 -9.15697694e-01
1.38723016e-01 5.50853372e-01 -9.92420316e-01 -6.72499537e-01
4.24550086e-01 -1.88809767e-01 5.31954728e-02 6.76825225e-01
4.76401746e-01 2.21155994e-02 6.15842283e-01 -9.17482734e-01
-1.28159976e+00 8.39841723e-01 -3.12903523e-01 9.17122126e-01
4.68971521e-01 -3.41865689e-01 2.02557397e+00 -1.49428654e+00
5.02681494e-01 1.53766489e+00 8.62766385e-01 4.16746736e-01
-1.53766990e+00 -3.58417302e-01 4.60384369e-01 1.86493874e-01
-9.32764530e-01 -3.15092087e-01 6.70041621e-01 -1.87142104e-01
1.84306669e+00 -1.48456782e-01 5.65714873e-02 1.43827701e+00
6.08980596e-01 7.00945973e-01 1.21056771e+00 -5.10572076e-01
-7.47569278e-03 -4.24262255e-01 7.85908937e-01 7.92440712e-01
3.55712116e-01 -7.34296441e-02 -1.06232631e+00 -6.53551340e-01
4.08855915e-01 -3.70579004e-01 2.76086062e-01 -4.08993304e-01
-1.07015967e+00 1.21225786e+00 1.92089736e-01 5.78834414e-01
-4.13536459e-01 9.76555645e-02 2.77703822e-01 2.83693552e-01
5.56405067e-01 5.86521447e-01 -1.07115686e+00 -1.90962285e-01
-3.47411036e-01 2.53180057e-01 8.78629684e-01 7.55701184e-01
5.19296825e-01 -1.57963187e-01 1.13695510e-01 1.16310048e+00
6.00324214e-01 6.75369650e-02 5.95822096e-01 -8.76480401e-01
1.00373983e+00 3.09130788e-01 -7.76280463e-02 -6.16150141e-01
-6.28831625e-01 -1.25222206e-01 -1.62378818e-01 -1.49036080e-01
6.14811182e-01 -2.39624470e-01 -5.89507878e-01 1.91926312e+00
1.84450835e-01 -1.31584406e-01 3.51126790e-01 6.72638297e-01
7.57537484e-01 7.19723880e-01 6.06684983e-01 -1.62441283e-01
1.95823765e+00 -9.32996631e-01 -7.21460819e-01 -9.34709013e-01
5.24294496e-01 -8.74947250e-01 1.04118156e+00 1.97105378e-01
-1.17341840e+00 -3.67486745e-01 -1.04962921e+00 -6.65399790e-01
-4.22479212e-01 -2.27243066e-01 1.22264600e+00 4.62337375e-01
-9.89413977e-01 4.15921807e-01 -8.80667567e-01 -1.88473403e-01
2.58096755e-02 4.61249650e-01 -8.21288288e-01 -3.87781620e-01
-1.22061670e+00 1.32531548e+00 4.52648848e-01 1.32427633e-01
-2.68225759e-01 -8.20429862e-01 -1.01709700e+00 2.42631659e-02
2.86227521e-02 -7.09202230e-01 1.22325039e+00 -4.30936292e-02
-1.36216593e+00 8.75748634e-01 -3.69479537e-01 -1.99387282e-01
-2.76130140e-01 -7.56038070e-01 -2.42138624e-01 -1.79883316e-02
3.65822136e-01 6.26291633e-01 7.21960485e-01 -8.69909406e-01
-4.01778609e-01 -4.48968977e-01 2.81382263e-01 2.49844283e-01
-5.49430490e-01 3.94283891e-01 -1.60647333e-01 -5.85150123e-01
6.22226417e-01 -9.62227583e-01 -2.86145091e-01 -3.55288565e-01
-4.18647006e-02 -8.61975431e-01 3.43138695e-01 -6.89005852e-01
1.11745584e+00 -2.18308592e+00 5.38592517e-01 -3.90606940e-01
6.77247625e-03 8.19609314e-02 -5.90564907e-01 6.54443145e-01
-2.93004602e-01 2.25898027e-01 -1.81200773e-01 -8.34252894e-01
2.31092706e-01 7.06561685e-01 -4.03262705e-01 9.95031744e-02
8.18873048e-01 9.85545039e-01 -7.68969417e-01 -5.78128934e-01
-2.58941531e-01 3.07967842e-01 -7.86734998e-01 2.04732060e-01
-1.40789419e-01 -2.01828286e-01 -3.51748645e-01 6.10295892e-01
5.39983988e-01 1.59561895e-02 6.15207434e-01 -1.94468081e-01
-2.98406761e-02 8.23397756e-01 -8.02030802e-01 2.09867001e+00
-7.29522109e-01 2.76639223e-01 -8.35286230e-02 -7.80233920e-01
6.48396313e-01 4.00001049e-01 3.13146859e-01 -3.61079454e-01
-1.19453907e-01 4.56939220e-01 3.05696130e-01 -3.46859455e-01
4.07133132e-01 -2.81509936e-01 -6.04184270e-01 5.34434915e-01
6.21108532e-01 5.54220788e-02 2.36912489e-01 1.46822426e-02
1.36006141e+00 2.50196397e-01 4.43002343e-01 -4.85341966e-01
2.25707471e-01 -5.10077849e-02 7.72252440e-01 3.83190840e-01
-7.99711570e-02 3.51570278e-01 8.13364744e-01 -9.27061260e-01
-9.85641062e-01 -1.57291007e+00 -4.16954339e-01 1.77546227e+00
-3.76773655e-01 -5.49120307e-01 -9.28544030e-02 -7.15917349e-01
4.80126478e-02 6.87582493e-01 -4.94261682e-01 3.95664394e-01
-1.31469738e+00 -1.36586833e+00 6.11189842e-01 8.17301095e-01
-1.60840645e-01 -1.01668191e+00 -1.76266283e-01 5.84464669e-01
-1.21449403e-01 -1.70207095e+00 -3.10660720e-01 6.95852637e-01
-1.05229235e+00 -8.54846716e-01 -1.54149607e-01 -1.35140944e+00
4.10979390e-01 -1.40074819e-01 1.27953994e+00 -3.55745673e-01
6.61304146e-02 1.55762643e-01 -3.82439852e-01 -4.26298305e-02
-5.32458946e-02 2.11140841e-01 2.28757650e-01 -4.61767435e-01
7.72177279e-01 -8.32869828e-01 4.65731472e-02 -1.43870264e-01
-6.19332850e-01 -5.69010019e-01 7.68931866e-01 1.20582294e+00
4.47080553e-01 -5.75201988e-01 3.76239240e-01 -9.02469695e-01
1.02948523e+00 -6.27771616e-01 -3.72777313e-01 2.35302687e-01
-4.49734271e-01 6.60538554e-01 4.26681280e-01 -5.96918166e-01
-9.88082170e-01 -1.00887120e-02 -3.02940667e-01 -1.08619198e-01
1.62190467e-01 5.36619544e-01 -7.79690146e-02 3.37054133e-01
3.07839215e-01 -2.35870436e-01 -5.07379293e-01 -8.12040925e-01
6.76495612e-01 1.57026172e-01 3.56321573e-01 -1.07122540e+00
5.81838965e-01 -7.63645917e-02 1.12406850e-01 -7.86816120e-01
-9.35484827e-01 -4.05531228e-01 -1.02851701e+00 7.01709867e-01
1.24763417e+00 -7.85291374e-01 -3.58860075e-01 -3.20871323e-02
-1.89859521e+00 1.58930317e-01 8.96235928e-02 6.07588530e-01
-1.98345959e-01 5.17293096e-01 -1.06375158e+00 -4.57631379e-01
-2.95982599e-01 -9.12517190e-01 9.96022284e-01 -2.34206721e-01
-3.45141709e-01 -1.14690673e+00 4.57566321e-01 5.17436028e-01
2.67299354e-01 -3.60514939e-01 1.62385356e+00 -1.04393637e+00
-3.96407723e-01 1.48948729e-01 -1.79900721e-01 1.53513849e-01
-2.55499892e-02 -4.37019616e-01 -8.50117028e-01 -1.22870348e-01
-8.19605589e-02 -2.99639225e-01 9.89855409e-01 2.11002678e-01
4.36248422e-01 -2.50087172e-01 -1.39157325e-01 5.70067227e-01
1.41931438e+00 -1.66400477e-01 1.19381726e-01 3.97243619e-01
7.55642414e-01 5.27209163e-01 4.76151794e-01 -3.59978452e-02
8.52605343e-01 6.97401762e-01 3.22596937e-01 3.94490033e-01
6.05135113e-02 -2.00407282e-01 5.71069837e-01 1.23649859e+00
9.09066200e-03 -1.22970648e-01 -1.07240999e+00 8.51599097e-01
-1.67215860e+00 -5.01792073e-01 -5.60220927e-02 1.51640952e+00
7.69364357e-01 4.37864304e-01 -1.49696097e-01 -2.41723716e-01
3.85145128e-01 4.82493520e-01 -7.08192140e-02 -1.35335815e+00
-1.68677956e-01 9.44987655e-01 3.91666263e-01 8.36459994e-01
-1.40643489e+00 1.48240304e+00 7.58213377e+00 2.91405141e-01
-4.99251157e-01 6.35659397e-01 3.13535966e-02 1.40340567e-01
-4.11338687e-01 3.63421887e-02 -1.05103540e+00 4.54764776e-02
1.08327031e+00 2.23536238e-01 2.34888673e-01 8.52845609e-01
-5.95229387e-01 2.33310223e-01 -1.44406843e+00 7.04009891e-01
2.36903995e-01 -1.10141897e+00 2.64048781e-02 -5.47508784e-02
1.72448695e-01 3.12133074e-01 7.44845346e-02 3.63667786e-01
5.43931365e-01 -8.15424323e-01 3.02857846e-01 -1.04711778e-01
2.90225923e-01 -4.16450530e-01 7.66428471e-01 2.34371945e-02
-1.14953732e+00 -3.97644252e-01 -5.42129040e-01 -2.06677467e-01
6.53015375e-01 3.85315120e-01 -8.32230091e-01 2.63786554e-01
6.63795710e-01 3.39947701e-01 -4.92402166e-01 3.33467960e-01
-5.71428537e-01 5.03307879e-01 -3.60636264e-01 -2.02881247e-02
4.33234721e-01 -7.57321715e-02 4.81184870e-01 1.55840480e+00
5.14145568e-02 2.23754600e-01 2.22905260e-03 4.26091462e-01
-1.99866164e-02 3.27922970e-01 -6.11455798e-01 -1.36400476e-01
5.17760992e-01 1.08110332e+00 -3.14672559e-01 9.37262475e-02
-7.13248849e-01 8.89464498e-01 1.16972923e+00 -3.39152627e-02
-4.89275426e-01 4.12489697e-02 1.20959032e+00 -4.37047660e-01
7.28348792e-01 -1.06308806e+00 -4.02843535e-01 -1.11477900e+00
2.15543002e-01 -4.94972169e-01 7.99431622e-01 -4.30840939e-01
-1.94672096e+00 8.02195489e-01 3.13774258e-01 -5.62366605e-01
-5.10463476e-01 -1.31748748e+00 -4.04982269e-01 8.63755703e-01
-1.77709532e+00 -1.41616607e+00 6.16298795e-01 1.44491553e-01
6.66991711e-01 -2.46976107e-01 1.60419273e+00 2.40732387e-01
-5.03432333e-01 4.99881744e-01 -4.21103597e-01 2.40314990e-01
5.23695946e-01 -1.40749776e+00 8.56209278e-01 9.25866842e-01
6.43763423e-01 8.67347419e-01 4.76776242e-01 -3.11198115e-01
-1.55495512e+00 -4.29362983e-01 1.83013010e+00 -8.37865949e-01
1.15894592e+00 -5.86165488e-01 -8.69237006e-01 1.09902382e+00
3.73929083e-01 2.14536667e-01 1.08787429e+00 7.12611973e-01
-9.22702193e-01 1.66395400e-02 -8.52816939e-01 3.43326211e-01
1.23865879e+00 -6.52466118e-01 -1.52393651e+00 3.62570107e-01
1.09191167e+00 -2.80535132e-01 -1.20487940e+00 2.61765599e-01
5.78279138e-01 -4.73250449e-01 1.37554538e+00 -1.16351461e+00
6.06775641e-01 2.28301525e-01 -6.77333951e-01 -1.19840455e+00
-7.97305703e-01 -3.53426933e-01 -1.59197614e-01 1.15074849e+00
9.81417418e-01 -7.36889482e-01 4.53172207e-01 7.43979573e-01
-2.82013774e-01 -7.76773095e-01 -1.49368572e+00 -6.42382741e-01
5.06742418e-01 -5.44679999e-01 2.74216056e-01 8.62282515e-01
1.44808888e-01 9.36849415e-01 -1.16507560e-01 6.06347024e-01
5.93694746e-01 1.32826060e-01 -1.99005567e-02 -1.29097509e+00
-3.79995167e-01 -7.60908276e-02 -3.45243990e-01 -1.07176805e+00
7.52393842e-01 -1.00029552e+00 -2.60094643e-01 -1.46948290e+00
-9.65968966e-02 -6.73741341e-01 -6.02235794e-01 6.99797213e-01
-2.38053873e-01 2.00446814e-01 2.95381069e-01 4.45650294e-02
-3.13939512e-01 1.88467860e-01 6.78040028e-01 -5.25036976e-02
-1.95888150e-02 -4.75817651e-01 -7.96571910e-01 7.34250009e-01
6.84015632e-01 -8.44962239e-01 -1.26415640e-01 -1.37546837e+00
4.67966169e-01 -7.91677982e-02 -2.07587123e-01 -3.60674471e-01
1.27738670e-01 -1.66715056e-01 1.34269878e-01 -2.84582615e-01
9.25661206e-01 -5.25479078e-01 -3.68654817e-01 2.55099505e-01
-2.46688038e-01 1.06234050e+00 3.12888622e-01 3.57221723e-01
-8.43917131e-02 -4.08321708e-01 2.11557776e-01 -8.91854763e-02
-8.82363200e-01 9.84843671e-02 -2.44488478e-01 3.58415902e-01
6.32229090e-01 3.05459827e-01 -4.19415325e-01 4.69578475e-01
-8.97238135e-01 -2.37554740e-02 -1.54389352e-01 9.05936122e-01
6.63851738e-01 -1.30833304e+00 -7.86108375e-01 2.34140381e-01
6.39295429e-02 -2.63632268e-01 -3.73321921e-01 4.11474675e-01
-4.14931238e-01 5.05574584e-01 -3.58774990e-01 -3.06613386e-01
-1.14136362e+00 5.03612757e-01 -2.73332000e-01 -7.31907785e-01
-5.24670959e-01 1.26909268e+00 -6.86925277e-02 -6.48017764e-01
-1.56762600e-01 -5.77939034e-01 -3.83754849e-01 3.42221558e-01
2.50427008e-01 -1.09357238e-02 1.01122975e-01 -7.52430856e-01
-8.61155152e-01 5.73213041e-01 -3.42314780e-01 -2.89272755e-01
1.91641724e+00 1.82626188e-01 -3.68316233e-01 6.04402244e-01
1.17472768e+00 6.35977685e-02 -7.14667499e-01 -3.82957637e-01
6.20311320e-01 -2.81170428e-01 -1.47892788e-01 -5.62118590e-01
-5.16113997e-01 1.10785246e+00 1.51976109e-01 -1.67477965e-01
5.55087864e-01 3.30053210e-01 8.44449997e-01 6.14498734e-01
5.05854011e-01 -8.83497715e-01 2.75834054e-01 8.98967028e-01
8.15339863e-01 -1.00437379e+00 8.91264826e-02 -6.29645288e-01
-7.66182542e-01 1.23784089e+00 5.61951280e-01 -4.51616347e-01
8.52103472e-01 4.72574830e-01 8.43747556e-02 -1.50593460e-01
-1.12355721e+00 -1.04742974e-01 8.53414983e-02 6.73702478e-01
6.14666283e-01 1.72691762e-01 -8.57527971e-01 9.24810052e-01
-3.35370898e-01 -7.03619361e-01 2.49423087e-01 9.20143366e-01
-7.46685863e-02 -1.80536425e+00 -1.39099658e-01 -1.90887544e-02
-8.17859292e-01 -5.10120630e-01 -1.65887550e-01 7.52399087e-01
9.54068750e-02 9.19233024e-01 2.03050613e-01 -9.66598243e-02
3.75166297e-01 6.97394729e-01 7.03762174e-01 -9.26793337e-01
-5.72813213e-01 -3.29450846e-01 5.95830441e-01 -7.11241245e-01
-3.17748040e-01 -8.73804569e-01 -1.15637255e+00 9.18536261e-02
-3.39405537e-02 2.99401022e-03 9.42884862e-01 1.08395600e+00
1.98219776e-01 1.49374634e-01 2.28892267e-01 -7.43561685e-01
-8.03120375e-01 -1.03896046e+00 -1.93370193e-01 2.32504085e-01
2.40569159e-01 -8.26806724e-01 -2.86305428e-01 -2.42149800e-01] | [10.494112968444824, 9.463590621948242] |
40e697f0-7a59-4415-be6c-7db6b865c6bf | a-study-of-low-resource-speech-commands | 2110.03894 | null | https://arxiv.org/abs/2110.03894v3 | https://arxiv.org/pdf/2110.03894v3.pdf | Neural Model Reprogramming with Similarity Based Mapping for Low-Resource Spoken Command Classification | In this study, we propose a novel adversarial reprogramming (AR) approach for low-resource spoken command recognition (SCR), and build an AR-SCR system. The AR procedure aims to modify the acoustic signals (from the target domain) to repurpose a pretrained SCR model (from the source domain). To solve the label mismatches between source and target domains, and further improve the stability of AR, we propose a novel similarity-based label mapping technique to align classes. In addition, the transfer learning (TL) technique is combined with the original AR process to improve the model adaptation capability. We evaluate the proposed AR-SCR system on three low-resource SCR datasets, including Arabic, Lithuanian, and dysarthric Mandarin speech. Experimental results show that with a pretrained AM trained on a large-scale English dataset, the proposed AR-SCR system outperforms the current state-of-the-art results on Arabic and Lithuanian speech commands datasets, with only a limited amount of training data. | ['Yu Tsao', 'Pin-Yu Chen', 'Sabato Marco Siniscalchi', 'Hu Hu', 'Chao-Han Huck Yang', 'Pin-Jui Ku', 'Hao Yen'] | 2021-10-08 | null | null | null | null | ['spoken-command-recognition'] | ['speech'] | [ 5.06517529e-01 -3.43596607e-01 5.11186600e-01 -5.94459176e-01
-1.15169513e+00 -6.80540740e-01 5.20230412e-01 -6.64352655e-01
-7.05219388e-01 5.64152896e-01 1.83416009e-01 -1.34794042e-01
4.15134639e-01 -2.98734158e-01 -6.47016764e-01 -6.71951652e-01
3.83108705e-01 4.81695801e-01 1.16697684e-01 -4.65443105e-01
5.29201217e-02 3.28709245e-01 -1.31078255e+00 1.89629674e-01
1.20642102e+00 7.56119490e-01 4.47074503e-01 7.01360941e-01
6.29510432e-02 8.51252377e-01 -1.02871192e+00 -2.57823199e-01
3.40626091e-02 -8.99682999e-01 -6.08371258e-01 -2.54390001e-01
6.22169748e-02 -1.53755262e-01 -4.44804370e-01 9.84379113e-01
8.49113226e-01 5.34004450e-01 7.20804155e-01 -8.25028837e-01
-9.79992568e-01 8.09459984e-01 -9.01632532e-02 -1.03323152e-02
4.03340161e-01 -1.21405371e-01 4.82404083e-01 -1.19344199e+00
3.19479913e-01 1.23495471e+00 3.72063518e-01 1.00820339e+00
-8.64493489e-01 -9.96282339e-01 1.20776400e-01 4.11453754e-01
-1.75053203e+00 -7.86707342e-01 9.28500354e-01 -4.76674847e-02
6.72029138e-01 4.50042874e-01 1.26340359e-01 1.26830494e+00
-2.21209452e-01 6.04743838e-01 1.26551080e+00 -6.37018740e-01
4.30521101e-01 7.74312913e-02 -1.69810057e-01 5.07569253e-01
-7.84325242e-01 -1.12763070e-01 -6.81258202e-01 4.51411344e-02
4.36338097e-01 -4.67315853e-01 -4.61635500e-01 1.35118946e-01
-1.22188091e+00 8.09700191e-01 -1.76124815e-02 4.52105135e-01
-1.33771300e-01 -4.19319987e-01 4.36617315e-01 6.32318377e-01
4.32345718e-01 3.08626592e-01 -3.93612295e-01 -3.96282583e-01
-6.42137170e-01 -1.70619294e-01 6.89230740e-01 1.03746271e+00
3.29336286e-01 8.96612644e-01 -5.26558124e-02 1.56451225e+00
2.77310669e-01 8.97702277e-01 1.09205091e+00 -5.27599275e-01
5.10236502e-01 4.08656001e-02 -1.90703943e-01 -4.48668420e-01
-3.82657535e-02 -1.98878527e-01 -7.82363534e-01 8.43281969e-02
9.76289436e-02 -1.82843804e-01 -1.23616409e+00 1.74774039e+00
7.68021867e-02 4.12529856e-01 7.93095171e-01 1.00578463e+00
1.00344026e+00 1.01195681e+00 -1.58329919e-01 -2.43109033e-01
9.60536301e-01 -1.16395879e+00 -9.51442897e-01 -2.50388652e-01
4.35265034e-01 -8.33936989e-01 1.70077181e+00 6.56783998e-01
-8.06520760e-01 -8.13052475e-01 -1.12762785e+00 1.34525210e-01
-3.36993814e-01 4.04797196e-01 -1.31002545e-01 6.74889088e-01
-7.85882294e-01 3.39604728e-02 -5.06571412e-01 -1.94233611e-01
-1.71741024e-01 1.73250467e-01 -4.87909079e-01 -3.23075056e-01
-1.51325572e+00 9.43603814e-01 2.76796073e-01 2.69998312e-01
-1.17822540e+00 -2.96358854e-01 -8.99637997e-01 -3.23278755e-01
1.17933527e-01 1.33711457e-01 1.17137313e+00 -1.09192276e+00
-2.39636111e+00 7.31569886e-01 9.03376192e-03 -2.42949158e-01
3.85267049e-01 -3.70732307e-01 -8.67597938e-01 -4.66919914e-02
-4.54517573e-01 2.34774753e-01 1.14576352e+00 -1.34043419e+00
-3.25916827e-01 -1.90446764e-01 -4.35792118e-01 5.82067072e-01
-2.25392431e-01 5.09521484e-01 -3.00078094e-01 -1.04414737e+00
9.05691274e-03 -9.20429707e-01 1.83506414e-01 -7.73437858e-01
-1.68834656e-01 -6.53681532e-02 8.03834736e-01 -1.21285677e+00
1.11602342e+00 -2.51540041e+00 6.82453215e-01 1.35108218e-01
-5.40811479e-01 5.89262307e-01 -3.89616996e-01 1.84726253e-01
6.45289868e-02 -1.47511616e-01 -6.78549349e-01 -4.71491396e-01
-4.29986306e-02 4.06319797e-01 -5.26477635e-01 3.69632125e-01
1.89307407e-01 4.95810688e-01 -6.71369910e-01 -1.66092619e-01
1.03487313e-01 5.42312026e-01 -1.48069963e-01 7.60353982e-01
1.86536834e-02 9.25423563e-01 5.68691036e-03 5.88102579e-01
5.81376791e-01 6.95041418e-01 -7.90185258e-02 1.51255161e-01
4.86228801e-02 2.32978269e-01 -1.00166607e+00 1.78596401e+00
-6.95447564e-01 2.97683269e-01 1.62690654e-01 -9.24903691e-01
1.41665518e+00 3.21788728e-01 -1.19530767e-01 -6.83542132e-01
1.34771496e-01 4.17333722e-01 2.74022847e-01 -3.56044412e-01
4.83007938e-01 -1.99964345e-01 -4.19016570e-01 3.85701865e-01
8.05031657e-02 -5.54178357e-01 -5.60622811e-01 -2.19510660e-01
1.02917981e+00 5.96028268e-02 -4.36244644e-02 1.43566683e-01
1.00771785e+00 -3.03953886e-01 6.09675527e-01 4.48628902e-01
-1.68102175e-01 9.29429352e-01 -1.01414219e-01 1.17355727e-01
-8.35302889e-01 -1.16192091e+00 5.27538098e-02 1.36005831e+00
-1.18645370e-01 6.85887933e-02 -9.61305916e-01 -7.62508631e-01
-4.41478461e-01 1.05065513e+00 -3.34164381e-01 -4.85592604e-01
-9.28157985e-01 -3.64751279e-01 1.19662130e+00 3.68574142e-01
7.34061658e-01 -1.37703693e+00 -8.81661847e-03 1.80009887e-01
-3.28823388e-01 -1.20557892e+00 -6.95899785e-01 2.27650795e-02
-1.76326737e-01 -4.23164189e-01 -9.78482425e-01 -1.12387133e+00
3.71228993e-01 -2.38655970e-01 4.67643380e-01 -3.08422983e-01
3.03986371e-01 2.29969427e-01 -9.16152298e-01 -3.20439011e-01
-1.10493565e+00 6.19471408e-02 5.00701308e-01 2.91361988e-01
2.28431776e-01 -2.13080600e-01 1.54509783e-01 4.38731790e-01
-7.59922504e-01 -2.33958736e-01 5.31511724e-01 1.07176530e+00
6.28148735e-01 -1.73156172e-01 1.07089889e+00 -8.05092871e-01
6.31483555e-01 -2.49440357e-01 -4.32024717e-01 3.82836163e-01
-3.73456776e-01 -1.36781558e-01 1.05691111e+00 -9.61011529e-01
-1.27980936e+00 1.02571584e-01 -5.70175588e-01 -6.96628690e-01
-1.97293773e-01 5.55664182e-01 -5.97322583e-01 -6.01031482e-02
4.80352193e-01 8.84380996e-01 4.77584861e-02 -6.70070350e-01
4.95397687e-01 1.42010438e+00 1.06554484e+00 -4.58554000e-01
8.53724897e-01 -1.76806256e-01 -7.11029589e-01 -9.19448853e-01
-2.66929924e-01 -8.40726420e-02 -7.48504996e-01 -2.26947248e-01
8.46903384e-01 -9.47940826e-01 -5.51146269e-02 1.17863619e+00
-1.11233985e+00 -5.85704923e-01 -2.13031456e-01 7.27911532e-01
-7.21376896e-01 2.71776646e-01 -7.37305939e-01 -8.47545505e-01
-3.97777855e-01 -1.19307768e+00 7.89584875e-01 6.69235513e-02
1.97038054e-02 -6.45622015e-01 2.31912822e-01 3.95468026e-01
5.38727343e-01 -1.88562587e-01 9.63320017e-01 -1.16986716e+00
3.17802876e-02 -6.97879791e-02 3.19686055e-01 1.00407648e+00
3.39148521e-01 -2.11982489e-01 -1.13422096e+00 -3.50188941e-01
2.61180490e-01 -5.19953251e-01 3.61818731e-01 -3.39458525e-01
8.70637953e-01 -1.36284038e-01 3.92190099e-01 5.52814245e-01
7.06851482e-01 7.00067103e-01 7.79344380e-01 1.01028070e-01
9.20263767e-01 3.48923266e-01 1.04083490e+00 1.36654034e-01
3.67006958e-01 8.24863315e-01 -1.47823080e-01 -5.05966647e-03
-2.73458391e-01 -2.41110444e-01 9.74317431e-01 2.00043392e+00
2.18589306e-01 -1.81417376e-01 -9.02164161e-01 2.90919542e-01
-1.48320556e+00 -4.80035305e-01 4.49228346e-01 2.33750844e+00
1.34875774e+00 -1.91942409e-01 -1.43030301e-01 1.63014695e-01
6.85240686e-01 1.18516691e-01 -4.92963880e-01 -7.22350776e-01
-3.62305641e-01 5.30206621e-01 -1.09788496e-02 6.62248671e-01
-9.12293911e-01 1.38582456e+00 5.87459040e+00 9.74273086e-01
-1.37909055e+00 3.78669471e-01 1.07502580e-01 2.90514261e-01
-2.20769361e-01 -4.29110587e-01 -5.66446126e-01 5.40256679e-01
1.29345441e+00 9.00945719e-03 8.79613280e-01 7.14608312e-01
7.75070563e-02 4.37076122e-01 -1.06592453e+00 9.84711289e-01
7.54470587e-01 -5.21503985e-01 4.14999649e-02 -4.81385052e-01
4.32746381e-01 4.94288206e-02 2.15902299e-01 6.52088642e-01
3.07737142e-01 -1.15633976e+00 7.11690485e-01 7.15155959e-01
1.37020540e+00 -9.62415934e-01 8.45571578e-01 3.64244878e-01
-8.75411630e-01 8.57839808e-02 -3.43424678e-01 3.32121313e-01
-6.77573904e-02 -1.80712774e-01 -7.58490920e-01 5.51206708e-01
5.76619685e-01 3.58178198e-01 -4.30286467e-01 4.07786131e-01
-3.58161122e-01 9.94430780e-01 -1.49711028e-01 1.05859473e-01
-1.73140243e-01 -2.98450112e-01 6.25752985e-01 1.17572618e+00
2.63908982e-01 1.95598453e-01 -1.14242062e-02 6.05710566e-01
-1.28667846e-01 3.47968787e-01 -4.63885874e-01 -1.05284750e-01
7.96922088e-01 8.15399110e-01 -3.86008471e-02 -1.33077070e-01
-3.20509762e-01 1.45212448e+00 2.39577815e-01 4.42309171e-01
-8.86152029e-01 -8.05911422e-01 4.86339509e-01 -5.54346025e-01
1.91212296e-01 -4.28795367e-01 6.18803948e-02 -1.18214393e+00
-5.34806773e-02 -1.56232882e+00 6.29277751e-02 -8.50010991e-01
-1.24189961e+00 1.10852671e+00 -3.18399310e-01 -1.34413135e+00
-7.31444657e-01 -1.58659175e-01 -4.87717003e-01 1.19394243e+00
-1.51926649e+00 -1.30220199e+00 -2.04698667e-01 8.97914469e-01
9.96566117e-01 -8.94032061e-01 1.27411556e+00 3.50173563e-01
-5.49051404e-01 1.05344021e+00 2.96971858e-01 2.75202841e-01
1.02204275e+00 -9.38964248e-01 3.63407850e-01 1.09200799e+00
1.91890880e-01 2.70652175e-01 4.55554396e-01 -5.29877365e-01
-1.29616511e+00 -1.13164771e+00 4.74604130e-01 -2.00214699e-01
4.40148890e-01 -7.76545942e-01 -1.28111017e+00 6.29133523e-01
1.97659984e-01 -1.31143302e-01 7.82922208e-01 -4.27188575e-01
-4.56026942e-01 -1.22763559e-01 -1.26442742e+00 4.97349828e-01
8.15169334e-01 -8.57490957e-01 -9.54668403e-01 -3.78692150e-02
9.78674352e-01 -5.35115421e-01 -9.99851644e-01 4.63522732e-01
4.50922966e-01 -2.30794460e-01 6.24793828e-01 -4.17656332e-01
-4.76920716e-02 -4.46277738e-01 -5.96476912e-01 -1.83917236e+00
3.01575311e-03 -7.31569529e-01 1.12286381e-01 1.67488027e+00
3.19936424e-01 -7.40707397e-01 1.69832021e-01 3.03792596e-01
-5.53576231e-01 -1.61499426e-01 -1.24142528e+00 -9.70275760e-01
2.93352991e-01 -3.34972173e-01 5.46071708e-01 1.03316116e+00
-1.33523688e-01 4.33655381e-01 -5.15583396e-01 2.69493163e-01
1.65340275e-01 -4.30347055e-01 7.44480550e-01 -7.90930688e-01
-3.17138880e-01 9.67131555e-02 -2.55764604e-01 -6.87001288e-01
7.53964067e-01 -8.25431764e-01 5.70540130e-01 -9.32503462e-01
-4.65608686e-01 -6.74044728e-01 -4.08830404e-01 5.31313658e-01
-1.81468099e-01 2.11037919e-01 3.09441000e-01 2.89682388e-01
-1.71488449e-01 1.08609200e+00 7.69596636e-01 -1.00335702e-01
-4.11844075e-01 4.19933572e-02 -1.71410441e-01 4.97040629e-01
9.12961066e-01 -4.23954934e-01 -4.58373964e-01 -4.16596711e-01
-4.93228972e-01 -9.48370341e-03 -1.64061844e-01 -1.16239417e+00
1.24261588e-01 1.73599906e-02 1.47677958e-01 -3.81393403e-01
6.32314801e-01 -6.34611428e-01 -2.92948596e-02 2.52230346e-01
-4.00635630e-01 7.20358863e-02 2.86396950e-01 2.81392068e-01
-4.89452809e-01 -2.56921113e-01 1.10105300e+00 2.53695607e-01
-7.98498690e-01 -2.13669956e-01 -4.92276877e-01 1.50750592e-01
9.51120019e-01 1.24949403e-01 -4.51418996e-01 -5.31693995e-01
-6.08112991e-01 -3.89329274e-04 3.17669928e-01 8.55419517e-01
1.00992167e+00 -1.34262455e+00 -1.08098340e+00 7.35186875e-01
8.03719461e-02 -4.41449918e-02 2.59443104e-01 4.01667684e-01
-3.54482472e-01 3.44184972e-02 -3.07852238e-01 -3.20563346e-01
-1.42653072e+00 3.83549750e-01 3.89795244e-01 2.21312970e-01
-4.53472227e-01 8.90461147e-01 -6.98695779e-02 -1.08553529e+00
4.62884724e-01 2.78149564e-02 -4.53824818e-01 -2.58395195e-01
3.76473099e-01 2.26328477e-01 7.04030320e-02 -1.03994727e+00
-4.73077148e-01 5.77250302e-01 -1.84193224e-01 -7.33834326e-01
1.01504552e+00 -1.60308987e-01 7.94178173e-02 8.11110198e-01
8.51127565e-01 4.81230050e-01 -8.44248354e-01 -3.59894246e-01
-3.26840490e-01 -1.94649175e-01 -1.44447237e-01 -1.07457983e+00
-8.29073846e-01 8.81056964e-01 8.10739160e-01 -2.80566752e-01
1.29936957e+00 -8.81494880e-02 7.92237878e-01 3.76648784e-01
3.69003505e-01 -1.28719723e+00 4.45587672e-02 9.27090704e-01
1.36900151e+00 -8.31827104e-01 -5.74456215e-01 -2.15592667e-01
-1.30426288e+00 8.09050262e-01 7.10564435e-01 -8.26514605e-03
4.15807188e-01 3.37076277e-01 5.82257450e-01 5.87394834e-01
-3.97593826e-01 -5.69475442e-02 1.24354750e-01 9.53085721e-01
1.34388477e-01 1.04993217e-01 -1.31044403e-01 9.89829957e-01
-5.28658807e-01 -2.62370974e-01 6.93536520e-01 6.91776097e-01
-2.27922156e-01 -1.26311433e+00 -3.79031926e-01 1.75163776e-01
-3.94274563e-01 -3.26960087e-01 -5.51980257e-01 5.12955964e-01
-2.12232083e-01 1.15771282e+00 1.04980171e-01 -1.08752120e+00
5.48119783e-01 3.88137370e-01 1.88854083e-01 -7.07580984e-01
-5.78781009e-01 1.04618967e-01 1.85736492e-01 -2.30715439e-01
-2.40052998e-01 -6.21669769e-01 -1.44392383e+00 2.47963518e-01
-2.17670172e-01 2.55860895e-01 8.71013224e-01 8.63399506e-01
2.42546275e-01 5.81348240e-01 1.05414581e+00 -4.88330960e-01
-9.21246827e-01 -1.52410054e+00 -6.02799296e-01 1.89082965e-01
2.53990948e-01 -3.80094677e-01 -5.66205382e-01 9.92037058e-02] | [14.63351058959961, 6.599139213562012] |
9bfd8d5d-bde4-42e9-940a-152801476052 | solving-visual-analogies-using-neural | 2111.10361 | null | https://arxiv.org/abs/2111.10361v1 | https://arxiv.org/pdf/2111.10361v1.pdf | Solving Visual Analogies Using Neural Algorithmic Reasoning | We consider a class of visual analogical reasoning problems that involve discovering the sequence of transformations by which pairs of input/output images are related, so as to analogously transform future inputs. This program synthesis task can be easily solved via symbolic search. Using a variation of the `neural analogical reasoning' approach of (Velickovic and Blundell 2021), we instead search for a sequence of elementary neural network transformations that manipulate distributed representations derived from a symbolic space, to which input images are directly encoded. We evaluate the extent to which our `neural reasoning' approach generalizes for images with unseen shapes and positions. | ['Tirtharaj Dash', 'Ashwin Srinivasan', 'Lovekesh Vig', 'Gautam Shroff', 'Atharv Sonwane'] | 2021-11-19 | null | null | null | null | ['visual-analogies'] | ['computer-vision'] | [ 5.72782159e-01 2.88164526e-01 4.64291833e-02 -5.09608805e-01
9.32432413e-02 -1.07631707e+00 1.08197665e+00 2.89195538e-01
-2.09927067e-01 6.13073289e-01 -1.04014210e-01 -7.19888389e-01
-1.30306333e-01 -1.10864687e+00 -1.07818627e+00 -2.25642800e-01
1.68234244e-01 2.36828208e-01 3.31636280e-01 -2.82604396e-01
5.68108499e-01 9.01537895e-01 -1.77742863e+00 6.23969376e-01
5.94569266e-01 6.68933153e-01 1.48852974e-01 1.07667923e+00
-2.32937708e-01 1.12552285e+00 -4.11777824e-01 -3.78306419e-01
7.52953216e-02 -7.58325279e-01 -1.20235026e+00 -3.74586910e-01
6.48734927e-01 -1.78059414e-01 -3.26526076e-01 1.42614388e+00
-1.31229579e-01 4.70640361e-01 7.72182405e-01 -1.41193235e+00
-1.61663318e+00 2.75152445e-01 -1.75785929e-01 3.49748105e-01
7.91906893e-01 5.10648489e-01 8.21579218e-01 -5.68962753e-01
8.03094208e-01 1.45320356e+00 5.00348568e-01 6.71485364e-01
-1.88004971e+00 -2.75800318e-01 3.64032425e-02 4.18010116e-01
-1.26266479e+00 -2.60713905e-01 5.38029492e-01 -4.33535904e-01
1.20067859e+00 4.28734422e-01 7.18464971e-01 6.75071001e-01
2.76863933e-01 5.13732851e-01 1.12865019e+00 -8.42019379e-01
4.05813456e-01 1.14029385e-01 9.93679240e-02 7.72105038e-01
2.42784657e-02 5.19553483e-01 -2.57450730e-01 -2.05477893e-01
1.23820221e+00 -7.66456202e-02 -1.65882707e-01 -4.77372885e-01
-1.11105156e+00 7.94388115e-01 8.50020051e-01 4.77082789e-01
-4.13018703e-01 5.57571650e-01 1.51424050e-01 7.31005907e-01
-4.68115330e-01 1.02097583e+00 -2.46082664e-01 6.36367381e-01
-4.75304425e-01 6.11402750e-01 8.09331357e-01 9.99205828e-01
6.47004902e-01 1.62351266e-01 -1.16314985e-01 1.81768045e-01
3.86838049e-01 4.82036769e-01 4.89366859e-01 -1.53758931e+00
1.83138385e-01 3.78437668e-01 -5.48567884e-02 -1.13744843e+00
4.75833304e-02 2.18699604e-01 -4.39406037e-01 6.76580489e-01
5.67033529e-01 1.36202544e-01 -7.45419383e-01 2.00282598e+00
-1.64739922e-01 6.49234131e-02 4.71886724e-01 6.11951888e-01
6.55778110e-01 9.32293177e-01 3.23848099e-01 -2.85852309e-02
1.07157147e+00 -5.21629512e-01 -3.05428803e-01 -2.62018770e-01
4.37549531e-01 -4.57744449e-01 1.36074805e+00 2.28476375e-01
-1.39916182e+00 -7.84767628e-01 -1.14912045e+00 -3.13535839e-01
-6.72147930e-01 -3.35011393e-01 7.59138465e-01 5.18113017e-01
-1.45166647e+00 1.18516648e+00 -4.96493071e-01 -4.56044167e-01
4.10062701e-01 3.80596757e-01 -1.09506376e-01 2.77793646e-01
-8.99564207e-01 1.18473136e+00 7.77790368e-01 -4.09165717e-04
-6.55451238e-01 -6.35499358e-01 -7.92288840e-01 1.97652653e-01
1.01016186e-01 -9.68121886e-01 1.59660304e+00 -1.58532619e+00
-1.39225781e+00 1.18014407e+00 -1.46314308e-01 -5.38470089e-01
7.16011003e-02 2.99177408e-01 -3.47279400e-01 1.50264934e-01
-8.57542679e-02 1.11808705e+00 6.50514007e-01 -1.40386391e+00
-5.07694006e-01 -4.74146008e-01 5.13054192e-01 2.35582247e-01
1.85833886e-01 3.35831754e-02 7.04077333e-02 -6.21438622e-01
9.94094014e-02 -9.28584635e-01 -2.22984985e-01 6.81884527e-01
-1.57976046e-01 -9.45502743e-02 6.25384033e-01 -5.41565239e-01
6.29643142e-01 -2.02583337e+00 5.41971862e-01 5.38173914e-01
6.42306805e-02 6.07416481e-02 -1.94583118e-01 2.78137147e-01
-5.19816816e-01 1.65668592e-01 -2.17488959e-01 4.58054185e-01
2.62486249e-01 4.47778791e-01 -9.29242730e-01 1.78635091e-01
7.80271515e-02 1.47961760e+00 -1.02357805e+00 -3.60885262e-01
1.57595664e-01 6.89502014e-03 -6.00993156e-01 2.38507181e-01
-6.34915471e-01 3.88137817e-01 -2.24404112e-01 2.54614502e-01
9.55976769e-02 -1.93988889e-01 2.66760558e-01 1.41149210e-02
1.42287210e-01 -2.91581973e-02 -9.12483692e-01 1.76697767e+00
-4.05122787e-01 1.02190483e+00 -4.48589027e-01 -9.90310073e-01
9.41615701e-01 3.40050995e-01 -5.96043885e-01 -8.57786894e-01
2.88350165e-01 5.53876646e-02 1.66009441e-01 -6.23175383e-01
3.06134552e-01 -7.02884376e-01 -1.41989574e-01 6.61399126e-01
-1.32774591e-01 -6.75277829e-01 -1.99420288e-01 4.54658978e-02
7.16685891e-01 5.08471727e-01 6.38022482e-01 -3.40781510e-01
4.88185972e-01 3.30926955e-01 -1.64566815e-01 1.06953025e+00
-7.23906904e-02 2.48988554e-01 6.14651680e-01 -7.29143143e-01
-1.36882925e+00 -1.61836326e+00 3.59575391e-01 9.77309942e-01
1.14200532e-01 1.67758942e-01 -6.08825266e-01 -2.26906613e-02
-4.14791852e-02 1.46869504e+00 -8.08915496e-01 -4.35759246e-01
-6.72288835e-01 3.79008241e-02 6.98857248e-01 9.04374123e-01
1.64334401e-01 -1.61226046e+00 -1.06123078e+00 -1.59966141e-01
4.20452297e-01 -4.16222274e-01 7.92435631e-02 2.05331430e-01
-1.26111627e+00 -9.86642718e-01 -3.86345267e-01 -9.74124551e-01
8.83772731e-01 -1.47363707e-01 1.18071043e+00 2.40448013e-01
-3.27627599e-01 4.70108360e-01 1.22124799e-01 -3.58321190e-01
-8.84291053e-01 -5.34224391e-01 -2.46375024e-01 -3.61389607e-01
1.30512148e-01 -8.61297250e-01 -1.71677500e-01 -2.20433325e-01
-1.03888738e+00 2.44279325e-01 2.76850849e-01 6.77763224e-01
4.82273340e-01 -1.39147535e-01 2.52316743e-01 -7.69727468e-01
9.09640610e-01 -2.44714484e-01 -8.88259172e-01 6.62301898e-01
-2.95358986e-01 6.35741711e-01 1.07374990e+00 -7.18618333e-01
-1.08059812e+00 1.26647383e-01 3.02165002e-01 -6.49989545e-01
-3.74833912e-01 3.36491197e-01 -4.98218238e-02 -3.21289599e-02
1.05832839e+00 3.63063246e-01 -6.43901229e-02 -2.66576782e-02
9.32128906e-01 6.07973635e-02 1.22449875e+00 -1.06485832e+00
7.71809936e-01 3.32251728e-01 4.41981584e-01 -2.17535689e-01
-2.60075122e-01 4.65658903e-01 -8.00460696e-01 1.18381478e-01
1.02356732e+00 -2.01602757e-01 -1.07047212e+00 -9.36417654e-02
-1.76972699e+00 -3.69605720e-01 -7.50056803e-01 -1.68382470e-03
-1.16680491e+00 4.73913066e-02 -3.24953288e-01 -5.00294566e-01
1.19490668e-01 -9.59545434e-01 4.72968102e-01 2.46201932e-01
-9.63094175e-01 -1.11307144e+00 1.18391566e-01 -4.81636763e-01
1.72345310e-01 2.89904594e-01 1.85647786e+00 -5.93548059e-01
-5.46923995e-01 -9.31871086e-02 -4.17569995e-01 -4.72344831e-02
-4.10563610e-02 4.32901867e-02 -8.11472833e-01 4.10087630e-02
4.87027802e-02 -2.54815757e-01 9.97246429e-02 1.94633156e-01
1.34953523e+00 -6.25344098e-01 -3.17714781e-01 4.62627023e-01
1.51779544e+00 8.70687544e-01 8.52565229e-01 2.09764987e-01
4.27157760e-01 7.30440259e-01 5.91427879e-03 -2.48877034e-01
-2.43042689e-02 4.89725798e-01 9.39796269e-02 1.76492155e-01
1.14193624e-02 -5.04004419e-01 -2.34197512e-01 5.99255078e-02
-3.10362518e-01 1.67333260e-01 -1.01418233e+00 5.09521842e-01
-1.85082304e+00 -1.34410989e+00 9.80862230e-02 1.98615241e+00
8.51389468e-01 -9.37403291e-02 -5.17784320e-02 -3.32251564e-02
7.93502092e-01 -2.70724773e-01 -5.69950402e-01 -1.23006964e+00
1.51648879e-01 6.39305294e-01 6.18161634e-02 6.40062213e-01
-5.16339302e-01 8.50144804e-01 7.47183561e+00 4.25295770e-01
-7.59284079e-01 -2.38739476e-01 5.41674972e-01 6.10937849e-02
-5.52043319e-01 3.12321782e-01 -1.27082631e-01 3.17322053e-02
1.14974046e+00 -6.40905678e-01 1.13930714e+00 6.25781357e-01
-2.01481462e-01 -2.79338509e-02 -1.97592080e+00 8.20857346e-01
-7.91675877e-03 -1.68535495e+00 5.58241427e-01 -4.41619515e-01
8.11020255e-01 -7.66151130e-01 1.78934291e-01 2.21853524e-01
7.75855482e-01 -1.43401313e+00 8.76370311e-01 7.96203673e-01
6.19837463e-01 -6.88108802e-01 -3.91231813e-02 2.18571454e-01
-8.95971298e-01 -2.53443956e-01 -2.79307038e-01 -3.55388552e-01
-1.86616346e-01 -5.00436246e-01 -6.04449391e-01 1.32369146e-01
5.55489898e-01 3.06739092e-01 -5.43245554e-01 5.96999824e-01
-7.03611732e-01 -1.18600346e-01 -1.69409737e-02 -1.58427298e-01
8.41056705e-02 3.63967079e-03 2.19455183e-01 7.89689004e-01
3.27755541e-01 5.35023630e-01 -4.93928075e-01 1.80208290e+00
4.17221785e-02 -3.52875888e-01 -1.24396789e+00 7.02837706e-02
5.52735090e-01 6.43800855e-01 -9.38012064e-01 -5.84275246e-01
-1.63642928e-01 1.13497710e+00 2.64613360e-01 6.40541255e-01
-8.24494958e-01 -5.40415168e-01 3.58654946e-01 -3.02684993e-01
3.22115183e-01 -4.03053185e-04 -5.62695980e-01 -7.04088211e-01
-1.74741596e-01 -9.22778189e-01 1.44521683e-01 -1.58985281e+00
-1.02976573e+00 4.86861646e-01 3.73945773e-01 -8.67508769e-01
-6.53573632e-01 -7.89733291e-01 -1.01642060e+00 1.06422937e+00
-7.96140015e-01 -9.29108500e-01 6.25031739e-02 6.41337931e-01
1.35247648e-01 -8.87765810e-02 1.11489511e+00 -4.62309390e-01
2.84846760e-02 3.08345348e-01 -2.73785591e-01 -3.93207818e-02
-2.10856065e-01 -1.15361702e+00 6.78498864e-01 6.07173979e-01
3.06008637e-01 8.75834167e-01 8.75172317e-01 -4.01896983e-01
-1.51414526e+00 -7.91664779e-01 8.34834039e-01 -5.65138340e-01
8.39655459e-01 8.57808888e-02 -1.09000039e+00 1.24930191e+00
3.47059131e-01 -7.72588828e-04 3.00753087e-01 -4.41577107e-01
-6.65506840e-01 2.23068759e-01 -1.15801764e+00 1.26022029e+00
1.03972173e+00 -1.05232954e+00 -1.31916738e+00 4.90400568e-02
7.65860617e-01 -2.80249029e-01 -6.61874831e-01 -6.28159344e-02
5.36504984e-01 -8.33996415e-01 1.24034297e+00 -1.25875592e+00
8.84100854e-01 -3.59251052e-01 -4.60198760e-01 -1.21517539e+00
-3.64467710e-01 -3.84269714e-01 1.31270185e-01 6.80493951e-01
2.70773888e-01 -7.52859890e-01 4.10044760e-01 9.25072193e-01
2.12435618e-01 -3.23703319e-01 -9.67545509e-01 -8.65855813e-01
3.29065621e-01 -2.08532289e-01 8.71103823e-01 9.43168581e-01
3.01353276e-01 4.22532022e-01 3.72421592e-01 9.06534865e-02
2.43216336e-01 4.86064404e-01 3.61797214e-01 -1.14177871e+00
-7.60724902e-01 -1.00039852e+00 -4.90991294e-01 -6.72943115e-01
4.80027258e-01 -1.28005040e+00 8.22666064e-02 -1.21790850e+00
7.80925453e-02 4.55135815e-02 -6.26294091e-02 6.39944494e-01
3.34289163e-01 1.27194509e-01 4.81692940e-01 3.39523435e-01
-8.95238444e-02 1.71526507e-01 1.17721772e+00 -2.24242330e-01
-1.86592609e-01 -4.32679564e-01 -7.36470044e-01 7.76505947e-01
6.14647746e-01 -4.39770848e-01 -6.90776169e-01 -6.17201447e-01
5.50909579e-01 6.08975112e-01 1.16707337e+00 -7.07751870e-01
4.14441794e-01 -6.85192823e-01 6.90456688e-01 -1.32611156e-01
1.10619433e-01 -8.80899727e-01 5.25347829e-01 7.81744361e-01
-8.31007719e-01 6.26540422e-01 7.38931954e-01 2.87964821e-01
5.10742851e-02 -8.15290332e-01 7.64857233e-01 -4.71826434e-01
-9.18349206e-01 -2.78632849e-01 -5.60291946e-01 -2.14819074e-01
1.22889054e+00 -5.62269092e-01 -3.87027204e-01 -3.34559560e-01
-1.03522158e+00 -2.29899362e-01 7.31090605e-01 1.12621956e-01
7.05021501e-01 -1.48777950e+00 -1.10212184e-01 4.01070155e-02
-5.59686311e-02 -3.48786891e-01 -1.67736605e-01 3.73143584e-01
-7.31960475e-01 4.78957921e-01 -7.99851120e-01 -2.14624822e-01
-9.87036705e-01 1.10772491e+00 6.30913794e-01 3.56584072e-01
-6.72379851e-01 7.16567993e-01 3.91073197e-01 -3.68064255e-01
-1.74798638e-01 -5.49489319e-01 8.28085244e-02 -5.82677484e-01
4.48957115e-01 2.97335267e-01 -4.92371857e-01 -2.22525045e-01
-1.77429885e-01 4.13393170e-01 2.14956179e-01 -4.35703933e-01
1.10847449e+00 2.54797488e-01 -5.09088099e-01 6.16684675e-01
1.15349603e+00 -6.63641989e-01 -8.89879584e-01 -1.11445062e-01
-4.14405465e-02 -4.67904150e-01 -4.03588384e-01 -6.95203245e-01
-6.49145067e-01 1.07699454e+00 3.50548357e-01 1.98449045e-01
1.12680101e+00 1.38916552e-01 8.90356153e-02 9.34975088e-01
1.30073249e-01 -5.83549380e-01 1.64536655e-01 3.27838153e-01
1.27211678e+00 -6.52997971e-01 2.66221003e-03 -1.57785609e-01
-4.25505638e-01 1.44549060e+00 5.33571243e-01 -4.57347304e-01
2.33358383e-01 1.55040577e-01 -4.28657830e-01 -4.06327039e-01
-6.32901430e-01 2.40906939e-01 3.36045057e-01 8.93552184e-01
1.27058104e-01 -9.80299264e-02 2.31665760e-01 3.32331330e-01
-4.01142806e-01 2.07804412e-01 4.61824745e-01 1.12646425e+00
-3.08377117e-01 -7.84800470e-01 -5.80401123e-01 2.04378873e-01
3.95372301e-01 -2.19926700e-01 -5.23033202e-01 8.33554447e-01
1.98311005e-02 2.98294693e-01 4.48926985e-01 1.55966207e-02
3.81721675e-01 2.53186226e-01 1.15491080e+00 -6.54339314e-01
-5.32592773e-01 -5.69505155e-01 -3.71591389e-01 -4.89258677e-01
-2.79150426e-01 -5.61702669e-01 -1.61787844e+00 -4.98530954e-01
4.16872472e-01 -2.17320025e-01 3.95962209e-01 1.08220530e+00
3.78155224e-02 5.65281570e-01 2.95654181e-02 -9.07327235e-01
-5.45809388e-01 -2.01753929e-01 -1.85741603e-01 5.70825100e-01
2.08701372e-01 -2.25133359e-01 5.24193654e-03 5.64407468e-01] | [10.580831527709961, 2.309551477432251] |
8d7eafbb-e81d-471f-a533-2738a879961f | face-inverse-rendering-via-hierarchical | 2301.06733 | null | https://arxiv.org/abs/2301.06733v2 | https://arxiv.org/pdf/2301.06733v2.pdf | Face Inverse Rendering via Hierarchical Decoupling | Previous face inverse rendering methods often require synthetic data with ground truth and/or professional equipment like a lighting stage. However, a model trained on synthetic data or using pre-defined lighting priors is typically unable to generalize well for real-world situations, due to the gap between synthetic data/lighting priors and real data. Furthermore, for common users, the professional equipment and skill make the task expensive and complex. In this paper, we propose a deep learning framework to disentangle face images in the wild into their corresponding albedo, normal, and lighting components. Specifically, a decomposition network is built with a hierarchical subdivision strategy, which takes image pairs captured from arbitrary viewpoints as input. In this way, our approach can greatly mitigate the pressure from data preparation, and significantly broaden the applicability of face inverse rendering. Extensive experiments are conducted to demonstrate the efficacy of our design, and show its superior performance in face relighting over other state-of-the-art alternatives. {Our code is available at \url{https://github.com/AutoHDR/HD-Net.git}} | ['Jiawan Zhang', 'Wenjing Dai', 'Xiaojie Guo', 'Meng Wang'] | 2023-01-17 | null | null | null | null | ['inverse-rendering'] | ['computer-vision'] | [ 1.61112875e-01 2.88436515e-03 2.34988227e-01 -6.46171093e-01
-4.44015294e-01 -2.99715638e-01 4.83169794e-01 -6.95397139e-01
1.14028446e-01 6.06199563e-01 5.48022240e-02 -1.25308514e-01
2.31905654e-01 -8.57496500e-01 -5.85904479e-01 -6.61463916e-01
4.40579414e-01 3.00841451e-01 -2.86624044e-01 -2.43805200e-01
-5.84109761e-02 5.04245877e-01 -1.68888414e+00 4.98487987e-02
1.01523530e+00 9.77966845e-01 7.34236762e-02 2.93212235e-01
5.21652177e-02 3.53499502e-01 -4.24840957e-01 -5.40703893e-01
5.12602448e-01 -5.17307878e-01 -3.47406834e-01 3.69112968e-01
5.60961485e-01 -7.75543869e-01 -3.61428559e-01 9.19334471e-01
6.64462149e-01 8.23304281e-02 4.31385010e-01 -1.33614540e+00
-7.86892414e-01 -1.48387447e-01 -7.80173600e-01 -3.25363159e-01
5.23804903e-01 3.55089575e-01 4.89879191e-01 -1.13192964e+00
2.65637130e-01 1.26688564e+00 5.99422932e-01 7.78693080e-01
-1.43308246e+00 -1.06194186e+00 2.49761939e-02 -1.05683900e-01
-1.43319798e+00 -8.78185451e-01 1.07339740e+00 -3.39970320e-01
2.00448766e-01 2.32228264e-01 7.58511901e-01 1.34167051e+00
-2.15149209e-01 4.23627287e-01 1.30411184e+00 -1.95414335e-01
9.07971933e-02 8.15189779e-02 -5.12454033e-01 8.82751107e-01
2.09961697e-01 9.68045071e-02 -5.64981759e-01 -3.28745805e-02
1.12503123e+00 1.22733623e-01 -6.31967187e-01 -4.80138183e-01
-7.55349696e-01 5.88851452e-01 3.47763747e-01 -1.91240087e-01
-2.04902261e-01 -1.10018387e-01 -1.23377349e-02 2.15742458e-02
7.16361642e-01 -5.17628528e-02 -1.84205949e-01 1.99911669e-01
-9.65314984e-01 1.84141174e-01 7.23295212e-01 8.62016976e-01
9.39824939e-01 2.31477559e-01 1.96415424e-01 9.69118476e-01
5.74423015e-01 5.20609975e-01 4.75151762e-02 -1.27746391e+00
3.75407457e-01 4.48903590e-01 1.78324103e-01 -1.01192105e+00
-1.22639358e-01 -2.04588026e-01 -1.00410998e+00 4.15238589e-01
5.03227293e-01 -1.51601344e-01 -9.78303194e-01 1.88105309e+00
6.89916909e-01 4.52635556e-01 -2.15356350e-01 1.09469521e+00
9.49333668e-01 4.85772669e-01 -2.53983378e-01 -1.82341397e-01
1.36056530e+00 -6.88356221e-01 -6.03189707e-01 -3.29782158e-01
6.11842871e-02 -9.92629468e-01 1.48163795e+00 3.48621547e-01
-1.06331027e+00 -5.91831803e-01 -1.00062180e+00 -3.49455178e-01
6.48549348e-02 3.01139534e-01 6.55229986e-01 7.31958807e-01
-1.04759622e+00 5.40396988e-01 -9.31817353e-01 -1.32454425e-01
6.13976359e-01 3.51523668e-01 -2.41451383e-01 -3.04565310e-01
-1.02235591e+00 4.10670877e-01 -2.24273652e-01 6.05981410e-01
-8.44850957e-01 -8.15399051e-01 -8.56015563e-01 -6.45736838e-03
3.15048099e-01 -6.75766110e-01 1.10201645e+00 -8.91128302e-01
-1.78022468e+00 8.86965573e-01 -1.07731871e-01 3.41118455e-01
7.56739736e-01 -2.08603784e-01 -2.49762982e-01 7.81593546e-02
-3.31883840e-02 5.47582388e-01 1.03067505e+00 -1.64974904e+00
6.27911389e-02 -5.97341001e-01 2.34154701e-01 1.12681627e-01
-4.45058465e-01 -2.07226843e-01 -6.65804923e-01 -5.57226658e-01
1.15458831e-01 -8.63672316e-01 -7.62441941e-03 5.08679807e-01
-2.70193130e-01 8.57085288e-02 8.52237999e-01 -8.39749336e-01
7.73478389e-01 -2.30256963e+00 -1.32538930e-01 8.84853527e-02
2.31791839e-01 1.67819604e-01 -1.41902640e-01 1.89058155e-01
-1.73095331e-01 1.88553594e-02 -1.97595611e-01 -5.76063454e-01
9.40877572e-03 1.95070624e-01 -3.49500507e-01 6.55818820e-01
1.95114911e-02 4.98606980e-01 -6.70886159e-01 -5.04286587e-01
2.27007672e-01 1.12028062e+00 -5.76738179e-01 4.69324023e-01
-2.09400132e-01 9.86241698e-01 -3.49075794e-01 6.37074947e-01
1.05242229e+00 -2.48068780e-01 2.07174405e-01 -3.07697952e-01
1.36800274e-01 1.72048882e-01 -1.22976422e+00 1.74842811e+00
-7.46438980e-01 5.72468400e-01 4.27249491e-01 -6.87200546e-01
9.62606251e-01 3.76133889e-01 3.95761222e-01 -7.09307253e-01
2.05952033e-01 7.37817958e-02 -3.02702338e-01 -5.34979403e-01
1.86499476e-01 -1.20175399e-01 4.63448793e-01 5.50421298e-01
-2.70917028e-01 -4.32300478e-01 -5.73638566e-02 -1.33754611e-01
5.00593901e-01 5.71106672e-01 -8.92981142e-02 -1.43703848e-01
5.25914252e-01 -5.56315184e-01 8.78100395e-01 2.25990769e-02
6.32538795e-02 9.25222993e-01 5.76897264e-01 -5.21939099e-01
-9.96220291e-01 -1.08551061e+00 -1.54049739e-01 9.61456299e-01
1.62570387e-01 -2.68719316e-01 -1.07308376e+00 -3.05165678e-01
-2.03858793e-01 5.51802278e-01 -5.24135768e-01 2.63854931e-03
-7.02138901e-01 -6.77181602e-01 2.30202451e-01 3.32591325e-01
6.89224958e-01 -8.42184603e-01 -3.92413199e-01 -3.32418174e-01
-3.61025423e-01 -1.10072923e+00 -5.71264088e-01 -6.65502131e-01
-7.61202991e-01 -9.82499719e-01 -6.81334138e-01 -6.12023473e-01
1.03231621e+00 4.09723014e-01 1.10215187e+00 4.00471836e-01
-3.30829322e-01 1.65506214e-01 -4.26593842e-03 -2.54710257e-01
-1.80913344e-01 -2.77666807e-01 9.13049430e-02 3.15425009e-01
-1.52924024e-02 -9.86148238e-01 -1.16153371e+00 5.07640600e-01
-9.06103253e-01 3.69989842e-01 2.13993177e-01 7.35539079e-01
4.42370147e-01 -1.35681614e-01 1.10997751e-01 -7.82199502e-01
4.50225502e-01 -2.89503068e-01 -8.60032141e-01 9.26909447e-02
-4.62956250e-01 -3.04071307e-01 6.84292972e-01 -4.26615387e-01
-1.39430928e+00 5.94075955e-02 -1.44640163e-01 -5.48427939e-01
-1.48684159e-01 -3.76624912e-02 -6.00712001e-01 -1.67002529e-01
6.01538777e-01 1.39520660e-01 2.42503256e-01 -5.13095796e-01
2.22944245e-01 6.34301841e-01 3.84974122e-01 -8.02725434e-01
9.69424665e-01 7.96511173e-01 -3.52255218e-02 -8.57573271e-01
-8.38165998e-01 3.09725404e-01 -5.80666304e-01 -3.13705713e-01
6.01387143e-01 -9.91717875e-01 -8.62852633e-01 5.30646324e-01
-1.08741832e+00 -4.65592086e-01 -1.19244615e-02 3.25496703e-01
-5.40086865e-01 3.91380280e-01 -4.91319805e-01 -7.03122735e-01
-2.37660199e-01 -1.29837632e+00 1.14016962e+00 5.09190202e-01
1.14347264e-01 -7.42421448e-01 -2.57839948e-01 7.31511652e-01
3.46935689e-01 4.52646911e-01 6.50415897e-01 1.42838120e-01
-8.04035187e-01 -4.99497261e-03 -3.65646243e-01 4.24835533e-01
3.23740929e-01 1.78434014e-01 -1.25705969e+00 -4.77219015e-01
1.78372249e-01 -4.71230656e-01 3.92805666e-01 1.79523602e-01
1.60187066e+00 -3.23567957e-01 -8.11790079e-02 1.09001076e+00
1.25136614e+00 -1.11261562e-01 5.80617070e-01 -1.10223286e-01
8.45902264e-01 8.01147819e-01 4.30150598e-01 6.87693417e-01
5.55196941e-01 6.30162776e-01 5.59444487e-01 -4.09452438e-01
-2.08042085e-01 -3.49524856e-01 8.38381350e-02 5.67898333e-01
-3.81602138e-01 -1.34224087e-01 -8.69116008e-01 1.12346038e-01
-1.34544313e+00 -6.80284202e-01 1.69237912e-01 2.32286167e+00
8.74522984e-01 -2.54078001e-01 -1.39766112e-01 -8.66877362e-02
6.35216355e-01 2.89988041e-01 -6.45175993e-01 2.31536273e-02
1.97054803e-01 2.02778384e-01 1.16454232e-02 4.72241640e-01
-7.39881098e-01 7.30101585e-01 5.26534796e+00 6.21151149e-01
-1.33465350e+00 1.16102427e-01 9.00891185e-01 -1.83488995e-01
-5.65865457e-01 -7.52311721e-02 -4.63902116e-01 5.07540584e-01
5.40153384e-01 1.12963254e-02 8.40642333e-01 6.20041072e-01
4.06010985e-01 1.31992534e-01 -1.13758016e+00 1.17519248e+00
1.43534750e-01 -9.64188755e-01 -9.99839902e-02 2.66863465e-01
6.35365486e-01 -2.56631315e-01 3.45449567e-01 5.59537672e-03
2.15715066e-01 -1.10359180e+00 5.86934030e-01 4.60244983e-01
1.17858589e+00 -4.88042235e-01 1.85468599e-01 2.40582332e-01
-1.09682500e+00 1.94223166e-01 -3.06337327e-01 -1.35240080e-02
7.14689195e-02 5.63461781e-01 -6.21498883e-01 4.45376068e-01
6.92347229e-01 4.41672921e-01 -3.85326147e-01 5.85094988e-01
-4.77983743e-01 4.34414834e-01 -3.36131603e-01 4.02760625e-01
-4.01536167e-01 -6.56765163e-01 2.14059189e-01 7.22794890e-01
4.57637012e-01 4.37939495e-01 1.01172343e-01 9.71288383e-01
-2.61095792e-01 -2.03737095e-02 -6.09399080e-01 2.08511516e-01
3.77444595e-01 1.38181782e+00 -6.43085480e-01 3.29384916e-02
-4.93529141e-01 8.76048803e-01 2.68074840e-01 7.40637243e-01
-1.07408547e+00 -1.76918045e-01 8.62882674e-01 4.27589059e-01
1.23650730e-02 -2.72640169e-01 -2.54820883e-01 -1.32550752e+00
3.43088806e-01 -8.93031478e-01 -2.66558304e-02 -9.60660100e-01
-1.09153378e+00 8.16460729e-01 -8.33837837e-02 -1.33063626e+00
-6.49447441e-02 -5.22844255e-01 -5.62025368e-01 9.46149826e-01
-1.46341348e+00 -1.27769685e+00 -8.33056808e-01 6.27116442e-01
4.51852411e-01 3.57849374e-02 7.70237446e-01 4.62577045e-01
-7.71409154e-01 6.61322355e-01 -8.58815908e-02 2.08274409e-01
7.41633415e-01 -7.99065888e-01 3.64082247e-01 7.95865476e-01
-7.64557049e-02 6.78781152e-01 6.42587245e-01 -2.91727036e-01
-1.51030266e+00 -9.05741751e-01 2.49158964e-01 -2.30180591e-01
3.30614060e-01 -7.44246542e-01 -9.97567177e-01 5.79271257e-01
1.45298317e-01 3.69953513e-01 6.68695688e-01 -4.65239026e-02
-4.61054295e-01 -5.24151504e-01 -1.28169179e+00 8.71778905e-01
1.13368142e+00 -5.04486561e-01 -7.88097307e-02 5.02060711e-01
5.03577709e-01 -7.10400462e-01 -8.15319002e-01 4.43191797e-01
8.15482199e-01 -1.38475478e+00 1.08774269e+00 -1.27824739e-01
6.75988078e-01 -2.93949097e-01 -5.63917719e-02 -1.16749537e+00
-1.83529295e-02 -7.59554148e-01 -6.95903227e-02 1.43662846e+00
1.22858889e-01 -9.16293204e-01 9.10552919e-01 8.76837730e-01
2.08428651e-01 -8.85624886e-01 -7.69355476e-01 -4.39901561e-01
-2.05983460e-01 -2.68487960e-01 1.00065660e+00 9.00754929e-01
-6.77783787e-01 1.72878787e-01 -4.05947566e-01 4.13176119e-01
8.19664299e-01 3.72718364e-01 1.04226375e+00 -1.14347184e+00
-2.00883836e-01 -1.53195009e-01 3.10059432e-02 -1.01814914e+00
2.82135099e-01 -4.63423431e-01 5.12453802e-02 -1.35453713e+00
2.53696530e-03 -6.97121739e-01 1.74008563e-01 4.53816324e-01
-4.49169874e-02 6.16549551e-01 1.38478652e-01 2.43986368e-01
3.65522988e-02 8.36010158e-01 1.64185679e+00 8.43664631e-02
-8.74369815e-02 9.34721231e-02 -8.38847220e-01 9.38588500e-01
9.03492749e-01 -2.76640028e-01 -7.15181947e-01 -8.46451759e-01
5.94658218e-02 1.88795909e-01 4.39427972e-01 -7.19558716e-01
-5.17044477e-02 -3.84760946e-01 6.84654117e-01 -1.74125671e-01
8.58479977e-01 -8.32454562e-01 4.03179914e-01 2.25031152e-01
1.46204785e-01 -1.06423758e-01 4.42526937e-02 3.30194801e-01
2.95262896e-02 1.21213399e-01 8.38121116e-01 -1.14909291e-01
-7.46924877e-02 6.56281114e-01 3.32868963e-01 -2.52382290e-02
7.90438473e-01 -2.72018582e-01 -3.19824398e-01 -5.55938542e-01
-4.58106697e-01 7.19011500e-02 8.27803195e-01 2.79875755e-01
6.98426247e-01 -1.30269492e+00 -6.87578619e-01 7.89558113e-01
-1.76007330e-01 3.58224064e-01 3.46075207e-01 7.52735376e-01
-6.90017939e-01 -1.74727380e-01 -1.97967917e-01 -5.83125889e-01
-1.17000556e+00 3.76245677e-01 3.71970892e-01 3.01649928e-01
-6.59624636e-01 7.22623587e-01 5.89196503e-01 -6.19873524e-01
1.26515448e-01 -5.29119708e-02 1.26603812e-01 -1.36680976e-01
4.34146971e-01 2.75272131e-01 -8.21475238e-02 -6.52822971e-01
-1.89223781e-01 7.98410356e-01 1.97850093e-01 -1.18719250e-01
1.40185630e+00 -1.92066789e-01 -2.46060312e-01 8.51781480e-03
1.04146576e+00 3.88990343e-02 -1.70286024e+00 -1.05885521e-01
-7.20920086e-01 -9.90686953e-01 -1.19848661e-01 -3.66051823e-01
-1.58238733e+00 9.79701400e-01 4.89412546e-01 -1.38755823e-02
1.52224910e+00 -2.61541903e-01 7.44483531e-01 6.63868114e-02
3.04596841e-01 -7.15379357e-01 1.47284582e-01 4.19586301e-02
1.12938440e+00 -1.25935936e+00 1.87566534e-01 -8.16088200e-01
-2.98725784e-01 1.01566803e+00 8.15269232e-01 -3.17307003e-02
7.07683384e-01 3.20498407e-01 3.00610065e-01 -1.39371201e-01
-4.07476425e-01 1.41942292e-01 1.01807840e-01 4.78174359e-01
4.72985685e-01 -5.95558695e-02 -2.67547872e-02 2.17362642e-01
-4.81629550e-01 8.70619044e-02 5.01186728e-01 6.02510989e-01
3.10608018e-02 -1.07148027e+00 -5.00850141e-01 1.94410935e-01
-3.12176108e-01 1.19414009e-01 -9.03188214e-02 7.15271771e-01
1.99600786e-01 8.77811372e-01 4.05156054e-02 -1.33569673e-01
3.29257786e-01 -2.35075116e-01 5.79791069e-01 -5.22816002e-01
-1.45918038e-02 2.28371486e-01 -2.49396816e-01 -6.79819405e-01
-3.49481970e-01 -5.10935128e-01 -9.75631177e-01 -6.33302510e-01
-1.31477401e-01 -1.46837354e-01 6.54777408e-01 6.23740375e-01
4.56895232e-01 2.89700300e-01 9.36902821e-01 -1.21496344e+00
-3.19320112e-01 -8.51891398e-01 -4.49319899e-01 3.50538135e-01
4.52977717e-01 -8.51391017e-01 -3.89517367e-01 2.67698437e-01] | [12.87390422821045, -0.131914883852005] |
376a86d7-29e3-497e-bb51-a4f632419233 | a-unified-one-shot-prosody-and-speaker | 2211.06535 | null | https://arxiv.org/abs/2211.06535v1 | https://arxiv.org/pdf/2211.06535v1.pdf | A unified one-shot prosody and speaker conversion system with self-supervised discrete speech units | We present a unified system to realize one-shot voice conversion (VC) on the pitch, rhythm, and speaker attributes. Existing works generally ignore the correlation between prosody and language content, leading to the degradation of naturalness in converted speech. Additionally, the lack of proper language features prevents these systems from accurately preserving language content after conversion. To address these issues, we devise a cascaded modular system leveraging self-supervised discrete speech units as language representation. These discrete units provide duration information essential for rhythm modeling. Our system first extracts utterance-level prosody and speaker representations from the raw waveform. Given the prosody representation, a prosody predictor estimates pitch, energy, and duration for each discrete unit in the utterance. A synthesizer further reconstructs speech based on the predicted prosody, speaker representation, and discrete units. Experiments show that our system outperforms previous approaches in naturalness, intelligibility, speaker transferability, and prosody transferability. Code and samples are publicly available. | ['Alexander Rudnicky', 'Shinji Watanabe', 'Li-Wei Chen'] | 2022-11-12 | null | null | null | null | ['voice-conversion', 'voice-conversion'] | ['audio', 'speech'] | [ 6.01684228e-02 1.24737872e-02 -4.39210802e-01 -3.16797853e-01
-9.19302702e-01 -6.37946427e-01 7.83976912e-02 -6.01381820e-04
1.59851819e-01 4.63309526e-01 8.09156179e-01 -1.05908751e-01
4.47834313e-01 -5.78659654e-01 -3.36594015e-01 -2.97872573e-01
2.35235408e-01 -1.64727315e-01 -5.24571538e-02 -3.31167907e-01
-7.35295564e-02 2.23603323e-01 -1.47804010e+00 2.88832337e-01
7.96066821e-01 8.76075089e-01 3.33017260e-01 9.88616109e-01
-2.76973009e-01 7.19600558e-01 -7.50748515e-01 6.15994185e-02
6.30363524e-02 -9.75900650e-01 -4.48854178e-01 9.86360312e-02
-1.73649415e-01 -2.42037758e-01 -3.05424631e-01 1.02174008e+00
6.29183471e-01 1.99969143e-01 4.99634326e-01 -8.27402532e-01
-6.49125636e-01 1.08475387e+00 1.83969513e-02 -1.69696417e-02
6.11896276e-01 -2.44668648e-02 1.18900716e+00 -9.92884994e-01
2.73867846e-01 1.33964467e+00 5.01751900e-01 6.64702237e-01
-1.22420609e+00 -7.37396002e-01 7.21115470e-02 -1.63973972e-01
-1.31902099e+00 -1.02743459e+00 1.08649158e+00 -2.82402426e-01
9.40474391e-01 3.76156628e-01 6.17645502e-01 9.84740674e-01
2.11560614e-02 4.79295790e-01 5.81194758e-01 -8.44398677e-01
1.65020406e-01 3.07541430e-01 7.07679242e-02 4.35977012e-01
-6.39340997e-01 3.41740608e-01 -8.42447817e-01 9.39543545e-02
5.40057480e-01 -4.00946081e-01 -3.64398241e-01 9.95543748e-02
-8.85813951e-01 5.08353055e-01 6.05665706e-02 2.70395726e-01
-2.49550626e-01 -1.31845087e-01 5.29045582e-01 5.11214018e-01
5.23429096e-01 2.19783500e-01 -2.17986733e-01 -3.59541774e-01
-1.01121628e+00 -1.64027035e-01 8.10372293e-01 9.64859009e-01
3.40951413e-01 8.46336603e-01 -1.04355000e-01 1.19883847e+00
1.48568630e-01 5.02898395e-01 8.60250235e-01 -8.50249648e-01
1.76899731e-01 8.66984874e-02 -2.37447466e-03 -5.87366760e-01
-1.04174703e-01 -1.16531745e-01 -4.69953090e-01 -6.77894503e-02
-1.62094176e-01 -9.44386199e-02 -6.17636025e-01 1.94535983e+00
9.56901163e-02 -1.88995432e-02 3.60683262e-01 7.16091156e-01
8.46267879e-01 1.16527939e+00 -4.75313514e-02 -7.69639075e-01
1.20607102e+00 -1.12691891e+00 -1.34499681e+00 -1.02993123e-01
-1.03647605e-01 -1.03815520e+00 1.53952694e+00 3.23816597e-01
-1.22174585e+00 -9.35121775e-01 -1.23363936e+00 -1.31805956e-01
8.15676823e-02 3.64052385e-01 7.33829364e-02 9.91147101e-01
-7.77731657e-01 4.57590371e-01 -5.77936947e-01 8.94415006e-02
-3.40539187e-01 1.73062146e-01 1.30769625e-01 5.53039193e-01
-1.19791305e+00 4.82614160e-01 2.31738493e-01 -4.08224374e-01
-7.10438669e-01 -7.28983879e-01 -9.95906115e-01 3.40595692e-01
-7.87373260e-02 -9.25875902e-02 1.63104141e+00 -9.89760816e-01
-2.31146693e+00 3.13158095e-01 -3.09655309e-01 -5.98905265e-01
9.68852043e-02 -1.32575303e-01 -9.59516585e-01 2.64032949e-02
-3.04614514e-01 6.39360368e-01 1.10416472e+00 -9.14180338e-01
-4.59374279e-01 1.53070286e-01 -5.24509966e-01 3.85142922e-01
-7.36512899e-01 9.32262391e-02 -2.50725091e-01 -9.13856864e-01
5.53853065e-02 -6.33463740e-01 2.80507833e-01 -3.27376336e-01
-4.44287270e-01 -2.42392682e-02 6.16832078e-01 -9.74254012e-01
1.53765416e+00 -2.58766079e+00 1.18686773e-01 -2.02402636e-01
-3.20376426e-01 3.54818590e-02 -2.38032922e-01 3.35438728e-01
-8.41904711e-03 -7.65847936e-02 -6.02527149e-03 -4.92601991e-01
1.10031739e-01 -1.21081673e-01 -9.15830016e-01 4.03520763e-02
5.13335280e-02 4.61475730e-01 -6.12100482e-01 -3.55885327e-01
2.59324610e-01 6.41382158e-01 -7.29659855e-01 4.88766491e-01
-2.42802620e-01 4.07710224e-01 1.22361653e-01 5.04714429e-01
1.95315808e-01 6.96902394e-01 3.12494755e-01 -1.68063223e-01
-2.78881997e-01 1.13472748e+00 -8.49464178e-01 1.49688613e+00
-8.69162261e-01 6.58490181e-01 6.74518272e-02 -4.98606503e-01
1.37314904e+00 8.61186922e-01 3.79304826e-01 -4.54447806e-01
1.21514201e-01 2.37259135e-01 6.94196895e-02 -1.97012901e-01
6.24820828e-01 -4.29540426e-01 -1.67793691e-01 3.77498239e-01
2.80859590e-01 -6.28148258e-01 -1.52926564e-01 -3.58199149e-01
5.78769684e-01 -2.58553680e-02 5.19303560e-01 3.61814420e-03
3.81635755e-01 -3.52428526e-01 6.18118346e-01 9.12161320e-02
-2.92475462e-01 8.15067947e-01 1.78260610e-01 2.11181678e-03
-1.04786277e+00 -1.36922848e+00 1.19659744e-01 1.52331030e+00
-7.98651129e-02 -5.28071463e-01 -9.04237807e-01 1.72674417e-01
-3.54830921e-01 1.07810259e+00 -1.05692454e-01 -4.55959022e-01
-6.68769240e-01 -1.18008800e-01 7.96424985e-01 3.76348078e-01
-1.36063978e-01 -1.24352872e+00 -2.59977095e-02 4.97529566e-01
-5.22752106e-01 -1.12187016e+00 -1.10541713e+00 1.50397539e-01
-6.65283501e-01 -2.71535933e-01 -3.56449723e-01 -1.02488244e+00
4.28154686e-04 1.27325028e-01 6.53965414e-01 -3.70823115e-01
-1.10858463e-01 1.11099198e-01 -2.86664635e-01 -4.77492630e-01
-1.14357615e+00 1.36701586e-02 5.94160974e-01 -1.64247176e-03
9.34864953e-02 -7.80762076e-01 -1.58134326e-01 8.69385824e-02
-5.45940518e-01 4.24050204e-02 1.11579120e-01 5.20812094e-01
6.50691032e-01 -4.55910191e-02 1.08093143e+00 -3.40275198e-01
9.36497569e-01 -1.67074248e-01 -4.41591114e-01 1.12545282e-01
-5.16420364e-01 -9.39399675e-02 1.00505722e+00 -7.93240249e-01
-1.17770648e+00 1.68368608e-01 -2.84933537e-01 -4.78369266e-01
-1.47335799e-02 2.06163362e-01 -4.83451307e-01 5.80225348e-01
8.08665276e-01 3.80555332e-01 1.77803650e-01 -5.98963499e-01
5.79217494e-01 1.17074466e+00 8.90234828e-01 -4.26369667e-01
4.78216380e-01 -1.33458108e-01 -8.72025669e-01 -1.20392442e+00
-6.10556364e-01 -2.29881793e-01 -4.74368572e-01 8.27828422e-03
5.84526837e-01 -1.18445694e+00 -4.30440605e-01 9.16562378e-02
-1.13060212e+00 -3.82002890e-02 -6.36025310e-01 8.48345220e-01
-7.72466540e-01 2.38826349e-01 -8.52748811e-01 -1.16583586e+00
-6.51001036e-01 -1.02103972e+00 7.41784573e-01 1.28903002e-01
-5.65076470e-01 -6.60652339e-01 2.06102848e-01 1.66486710e-01
4.59331155e-01 -1.82645321e-01 1.06281364e+00 -4.09627646e-01
-5.82744293e-02 -7.63408691e-02 5.33773601e-01 6.67545438e-01
7.14297950e-01 2.01253176e-01 -1.34582078e+00 -1.73356995e-01
3.40028405e-01 -2.70235091e-01 5.08737087e-01 4.52645689e-01
8.44315708e-01 -5.16708732e-01 1.62376627e-01 5.01241744e-01
7.58526027e-01 3.92547309e-01 3.30158174e-01 -5.12824595e-01
3.61576885e-01 6.29214823e-01 4.15332466e-01 4.65364218e-01
2.03951254e-01 5.43474615e-01 -1.87135339e-01 8.36665463e-03
-6.65867090e-01 -5.86864591e-01 1.16260648e+00 1.76392162e+00
2.28277639e-01 -1.29524484e-01 -4.92826670e-01 5.90407431e-01
-1.13965631e+00 -9.63274658e-01 5.83001733e-01 2.17332935e+00
1.37234318e+00 2.59956867e-01 3.35586011e-01 3.65828097e-01
9.39719915e-01 2.53665537e-01 -4.96282846e-01 -7.93518066e-01
2.54865028e-02 2.35090211e-01 -8.66027772e-02 8.07012916e-01
-7.00910270e-01 1.17812741e+00 6.80807304e+00 8.14765334e-01
-1.41783047e+00 3.38640660e-02 3.59738261e-01 -4.40745711e-01
-4.49273407e-01 -2.21696422e-01 -7.40548372e-01 1.82450026e-01
1.40713704e+00 -7.50736058e-01 9.40236270e-01 7.73403108e-01
5.22498071e-01 5.17929554e-01 -1.32073188e+00 7.89202690e-01
2.00630814e-01 -9.76831138e-01 4.91404571e-02 -3.53721708e-01
4.75171000e-01 -2.56745130e-01 2.13600084e-01 5.34388423e-01
-8.75407085e-02 -1.01081204e+00 1.16257453e+00 1.93445235e-01
1.24153614e+00 -8.41887772e-01 -6.27932176e-02 3.94949287e-01
-1.48027861e+00 -3.88312303e-02 -2.70706028e-01 -1.01755135e-01
1.62121028e-01 2.26020932e-01 -1.16512012e+00 2.35974029e-01
1.13663957e-01 2.62219131e-01 8.44194815e-02 4.74675804e-01
-2.90514469e-01 1.11260796e+00 -2.35412344e-01 4.10379767e-02
-4.18861628e-01 -2.12363508e-02 6.42717898e-01 1.33833551e+00
4.11245167e-01 9.17284191e-02 2.55816102e-01 9.78576660e-01
-3.07308018e-01 4.02356744e-01 -3.58345598e-01 -4.83331561e-01
1.04558730e+00 9.15108323e-01 -4.14795101e-01 -1.17368169e-01
-2.69667566e-01 7.99013078e-01 -3.87675054e-02 3.51109147e-01
-7.20882595e-01 -3.70273262e-01 9.14566576e-01 -8.25198889e-02
1.38854772e-01 -2.75723010e-01 -4.15359795e-01 -1.03088641e+00
-1.17594860e-02 -9.11722183e-01 -3.14155500e-03 -5.27954459e-01
-1.00337982e+00 9.49637949e-01 -3.31951737e-01 -1.46784317e+00
-8.67758214e-01 -1.51409388e-01 -6.20176792e-01 9.17002559e-01
-1.32365775e+00 -7.94391990e-01 1.44022226e-01 2.88359582e-01
1.18762529e+00 -4.80868667e-01 1.33370042e+00 5.10268472e-02
-4.44537193e-01 7.44592786e-01 -1.88301995e-01 3.02818641e-02
7.87056088e-01 -9.59583342e-01 7.04923093e-01 7.08338797e-01
3.57449591e-01 4.52475518e-01 7.44085371e-01 -3.86275113e-01
-1.22413778e+00 -1.00722706e+00 8.85372818e-01 3.84943783e-02
4.82241243e-01 -6.65457070e-01 -8.65879118e-01 3.84507447e-01
2.63298333e-01 -1.47281796e-01 1.11858702e+00 4.44779769e-02
-3.97899121e-01 -2.86140531e-01 -7.74786711e-01 6.40520155e-01
6.07841432e-01 -1.08425701e+00 -9.80199873e-01 3.42248101e-03
1.42242813e+00 -3.66372555e-01 -6.04325771e-01 3.43179069e-02
5.97016394e-01 -6.85134292e-01 9.09951806e-01 -3.04181814e-01
2.29966715e-01 -9.46491957e-02 -4.43994462e-01 -1.42793059e+00
-4.53302152e-02 -1.02439284e+00 -1.29934192e-01 1.53138781e+00
5.43797731e-01 -1.62107900e-01 5.42994812e-02 2.45269626e-01
-1.91460535e-01 -1.26676589e-01 -7.62536347e-01 -9.81668711e-01
8.38400796e-02 -5.96361935e-01 7.37487316e-01 7.74308383e-01
5.00701666e-01 7.10240960e-01 -5.95346570e-01 2.49447197e-01
1.92347169e-01 3.71976554e-01 4.71483409e-01 -8.84814739e-01
-4.64592129e-01 -3.41540575e-01 1.92718863e-01 -8.91946375e-01
2.65395790e-01 -8.32263649e-01 4.70665276e-01 -1.05713987e+00
-3.39985013e-01 4.78271358e-02 -3.87291223e-01 3.90374929e-01
-5.26970252e-02 -3.66252512e-02 4.22609597e-01 1.89009845e-01
5.49765006e-02 9.09967959e-01 9.08198774e-01 -2.00096238e-02
-9.18316305e-01 1.46368191e-01 -4.42661643e-01 7.45249569e-01
1.01997399e+00 -3.43307763e-01 -5.60242951e-01 -1.76471919e-01
-6.13350689e-01 6.04094148e-01 -2.85438508e-01 -1.16284359e+00
1.54232293e-01 -2.55952448e-01 1.23467118e-01 -4.34721261e-01
7.66245365e-01 -3.89897883e-01 -2.67291516e-02 3.51929009e-01
-6.16068125e-01 -2.69593179e-01 3.24297667e-01 1.92913175e-01
-5.50199389e-01 -5.36566786e-02 8.66034031e-01 1.97983369e-01
-2.38666117e-01 -5.98845705e-02 -5.24270952e-01 -1.11784503e-01
4.19800937e-01 -7.09557012e-02 7.21571222e-02 -6.44371271e-01
-7.69047201e-01 -3.58868122e-01 1.76763594e-01 8.06705177e-01
6.55052364e-01 -1.36562002e+00 -7.51324415e-01 6.15172148e-01
-8.03687796e-02 -5.10183513e-01 1.44019544e-01 1.81334838e-01
-1.71853855e-01 2.92460173e-01 -2.61223644e-01 -2.67135650e-01
-1.21009254e+00 5.94981253e-01 2.96755165e-01 3.48796964e-01
-5.74147403e-01 5.79482377e-01 1.63391113e-01 -3.90301377e-01
5.38991272e-01 -6.68070853e-01 -1.31379470e-01 4.19385619e-02
6.58909678e-01 1.41527876e-01 -5.98124973e-02 -6.49858296e-01
-1.95148975e-01 2.52506554e-01 1.83741048e-01 -6.52764082e-01
1.24696863e+00 -4.43937927e-01 1.70116901e-01 9.67985094e-01
1.02696800e+00 7.08282173e-01 -1.22770870e+00 -2.55648077e-01
-1.90276533e-01 -4.66703176e-02 1.25825614e-01 -5.83548188e-01
-5.50545156e-01 9.84349787e-01 3.82440329e-01 2.87386626e-01
1.30299222e+00 -3.74468803e-01 1.07219756e+00 1.92404166e-01
5.80343083e-02 -1.28036177e+00 1.39432445e-01 7.37034261e-01
1.03244448e+00 -7.71874189e-01 -3.08837593e-01 -6.05159879e-01
-7.99472511e-01 1.18220305e+00 3.93218338e-01 1.02383606e-01
5.57016194e-01 5.83370209e-01 4.00136620e-01 5.44163704e-01
-8.45122397e-01 -3.83369513e-02 3.37577909e-01 4.88513917e-01
7.42172182e-01 3.26997072e-01 -5.08677922e-02 1.09300160e+00
-1.10229337e+00 -2.81837642e-01 5.07913172e-01 3.65305632e-01
-7.93826282e-01 -1.10964930e+00 -5.40613532e-01 -3.76586825e-01
-3.19846243e-01 -2.53013968e-01 -5.06769300e-01 1.85440496e-01
-2.07174703e-01 1.17439139e+00 1.10216878e-01 -6.23129368e-01
3.35217297e-01 5.14315724e-01 2.58870393e-01 -8.61002624e-01
-4.43799764e-01 7.57047355e-01 -6.08220175e-02 -1.12021603e-01
1.01721331e-01 -3.52463633e-01 -1.67533398e+00 -1.95454480e-03
-2.67220229e-01 1.43062904e-01 7.87859201e-01 6.33282900e-01
2.39270791e-01 8.53414834e-01 1.19494712e+00 -5.45876205e-01
-8.79016459e-01 -1.12910759e+00 -6.76476002e-01 9.18292254e-02
7.02303529e-01 -7.38467723e-02 -2.41069049e-01 4.68372017e-01] | [14.955738067626953, 6.554706573486328] |
be4ce4ea-ce4a-4ac5-9172-a316e378f67b | knns-of-semantic-encodings-for-rating | 2302.00412 | null | https://arxiv.org/abs/2302.00412v2 | https://arxiv.org/pdf/2302.00412v2.pdf | KNNs of Semantic Encodings for Rating Prediction | This paper explores a novel application of textual semantic similarity to user-preference representation for rating prediction. The approach represents a user's preferences as a graph of textual snippets from review text, where the edges are defined by semantic similarity. This textual, memory-based approach to rating prediction enables review-based explanations for recommendations. The method is evaluated quantitatively, highlighting that leveraging text in this way outperforms both strong memory-based and model-based collaborative filtering baselines. | ['Lucas Dixon', 'Thomas Bonald', 'Raghuram Vadapalli', 'Léo Laugier'] | 2023-02-01 | null | null | null | null | ['collaborative-filtering', 'semantic-textual-similarity'] | ['miscellaneous', 'natural-language-processing'] | [ 2.14825094e-01 2.88183868e-01 -1.06053519e+00 -8.57432723e-01
-5.84958076e-01 -3.82227868e-01 8.63069355e-01 7.88415194e-01
-8.42111707e-02 3.46830100e-01 1.14521599e+00 -5.57748079e-01
-8.64373207e-01 -7.26668298e-01 -1.01673774e-01 1.35940194e-01
-9.88807436e-03 7.74138153e-01 4.92348552e-01 -9.02335823e-01
1.39761841e+00 -1.07805997e-01 -1.85834658e+00 1.13482511e+00
7.95596421e-01 1.27869666e+00 2.32482746e-01 5.22163332e-01
-4.71564144e-01 6.79466665e-01 -2.91273266e-01 -8.69767666e-01
-1.23959750e-01 -1.63419932e-01 -7.78032839e-01 -3.42137396e-01
4.97190267e-01 1.64934024e-01 -2.74591893e-01 6.93625212e-01
1.89814836e-01 9.22537267e-01 9.09308612e-01 -6.69092059e-01
-1.33682799e+00 9.23240602e-01 -2.01096430e-01 3.84361476e-01
1.11144471e+00 -9.26912665e-01 1.87933457e+00 -1.25451469e+00
6.94965541e-01 1.24666941e+00 5.61421812e-01 4.00131762e-01
-1.01197708e+00 -3.34322937e-02 6.14156783e-01 3.48478496e-01
-6.38141811e-01 -1.04182832e-01 4.94935572e-01 -1.97805896e-01
1.66157329e+00 4.67994839e-01 7.72317946e-01 7.24322915e-01
2.03369305e-01 6.63406909e-01 6.80108070e-01 -3.39528084e-01
4.85334426e-01 4.52107459e-01 6.44114077e-01 5.51290452e-01
2.02890843e-01 5.73308067e-03 -1.10008478e+00 -6.60251915e-01
2.27294475e-01 5.84578812e-01 4.87742983e-02 -5.37954867e-01
-5.62669456e-01 1.22523451e+00 2.74319470e-01 1.85748376e-02
-1.29293203e-01 -2.64454991e-01 3.63124609e-01 7.14721382e-01
1.06568182e+00 9.61042106e-01 -8.61838281e-01 2.17488363e-01
-8.44906807e-01 1.32811069e-01 8.66353929e-01 7.23711610e-01
5.23545742e-01 -3.85078639e-01 -1.74128726e-01 1.06834793e+00
1.01094925e+00 2.43029013e-01 1.02487028e+00 -4.37073946e-01
2.24283636e-01 5.87557137e-01 1.52403906e-01 -1.46883523e+00
-1.77310213e-01 -4.00870502e-01 1.85955122e-01 -2.29552314e-01
-3.59106839e-01 8.53131115e-02 -8.17394197e-01 8.90571058e-01
-7.01118633e-02 -7.20707104e-02 1.88444525e-01 7.96411276e-01
1.05881727e+00 4.85682130e-01 1.91542268e-01 -2.54844874e-01
9.03902531e-01 -1.23664951e+00 -4.28368360e-01 -4.09011066e-01
7.34046698e-01 -8.57733965e-01 1.11008048e+00 4.81130868e-01
-9.58829999e-01 -6.87317073e-01 -1.03573942e+00 2.34463885e-01
-6.56851292e-01 -1.13722749e-01 8.36551368e-01 7.78502941e-01
-1.18658733e+00 1.11123872e+00 -3.36122401e-02 -6.18826926e-01
8.76289830e-02 3.74631792e-01 1.70131847e-01 5.69949746e-02
-1.45835865e+00 1.21590304e+00 -2.01623470e-01 -5.91292441e-01
-2.08637670e-01 -8.00155818e-01 -7.32977509e-01 1.41227901e-01
2.94336332e-05 -8.16576123e-01 1.23676741e+00 -7.77380407e-01
-1.30492342e+00 4.02566731e-01 -4.74730432e-01 -3.64647686e-01
-4.10648853e-01 -2.37682119e-01 -1.16309190e+00 1.46278401e-03
-9.07264501e-02 1.11728974e-01 8.96181405e-01 -1.28516483e+00
-8.61902773e-01 -4.43266004e-01 1.88937470e-01 4.29395378e-01
-8.79163265e-01 2.82966048e-01 -5.85547462e-02 -7.66192555e-01
-7.25506395e-02 -5.77826738e-01 -6.46621406e-01 -5.55944920e-01
2.89099693e-01 -5.14318883e-01 5.89442551e-01 -2.59566516e-01
1.93141270e+00 -1.44905865e+00 -1.89124316e-01 8.37867677e-01
1.86496109e-01 -7.02317134e-02 -3.43050152e-01 1.00077152e+00
1.04318030e-01 4.04481918e-01 6.30232215e-01 -3.98417294e-01
1.66563168e-01 -3.32242101e-02 -8.00045550e-01 -2.07990766e-01
-5.10906816e-01 1.02154946e+00 -1.26333427e+00 -2.68206269e-01
4.07900997e-02 5.04255474e-01 -7.12309539e-01 -2.55382527e-02
-4.33440149e-01 -3.84767592e-01 -8.23172510e-01 4.11363810e-01
-2.13778898e-01 -5.70116758e-01 5.76950729e-01 -1.73130751e-01
3.58026326e-01 1.12634122e+00 -4.89481181e-01 1.71747017e+00
-7.74595141e-01 2.38902792e-01 -8.00317168e-01 -6.13763094e-01
1.43157172e+00 1.22487433e-01 2.15029761e-01 -7.15281248e-01
-1.01995111e-01 1.69445723e-01 -4.40405935e-01 -2.33808100e-01
1.06587851e+00 -2.12021358e-02 2.62339953e-02 9.13945794e-01
-1.64448619e-01 -2.00366035e-01 1.57818958e-01 1.02211380e+00
1.19322157e+00 2.68630497e-02 3.48538935e-01 -3.14658582e-01
3.25916141e-01 1.65878654e-01 -3.22791457e-01 1.00528097e+00
5.48208356e-01 3.29203963e-01 -2.55202740e-01 -3.06380481e-01
-5.47267377e-01 -8.98648381e-01 3.36830676e-01 1.70514667e+00
2.16724783e-01 -1.29449904e+00 1.85147256e-01 -1.11776054e+00
3.34112465e-01 1.03829336e+00 -8.59721005e-01 -5.58625996e-01
1.31068423e-01 -3.37498158e-01 -5.62612474e-01 8.89662743e-01
-4.78582770e-01 -7.35408902e-01 7.10538998e-02 1.99394658e-01
2.78983593e-01 -1.60165742e-01 -7.50136018e-01 1.76263705e-01
-1.40742159e+00 -5.96272945e-01 -1.75185889e-01 -5.19902170e-01
5.47998011e-01 8.43181670e-01 1.83812749e+00 6.18718088e-01
2.94548273e-01 1.04721916e+00 -1.02682471e+00 -1.62530616e-01
1.09491020e-01 -6.90767616e-02 1.12296835e-01 -2.95073807e-01
8.83794308e-01 -6.26852036e-01 -9.60585117e-01 6.17163837e-01
-4.75325167e-01 -5.35953701e-01 2.58404166e-01 4.87313956e-01
4.85867202e-01 -3.80971432e-01 9.98649061e-01 -1.28952873e+00
1.43945348e+00 -8.43079627e-01 1.93388805e-01 5.71213961e-01
-1.75908375e+00 1.59090921e-01 5.04217863e-01 -4.03084517e-01
-1.23233128e+00 -1.24172881e-01 7.09958151e-02 3.14469635e-01
2.68081129e-01 1.03424919e+00 3.46043140e-01 -3.53881046e-02
1.13513660e+00 -1.94350019e-01 -9.70457643e-02 -6.57473922e-01
7.27377653e-01 5.91382921e-01 3.11253294e-02 -2.41792738e-01
2.97685146e-01 2.37212941e-01 -4.28478509e-01 -1.46769613e-01
-1.45828247e+00 -1.19979370e+00 -3.87540370e-01 -3.75401676e-01
2.93675810e-01 -4.85155076e-01 -2.56085575e-01 -8.97375941e-01
-6.94211841e-01 3.36719126e-01 -4.33280557e-01 5.35012424e-01
-3.39777708e-01 1.68866262e-01 -5.20448744e-01 -6.90967381e-01
-3.70860040e-01 -3.76403272e-01 7.12606668e-01 1.13435768e-01
-9.49835718e-01 -1.57955086e+00 5.77250004e-01 7.70998776e-01
7.13811994e-01 -7.02286959e-01 9.55436647e-01 -1.44402254e+00
2.26493165e-01 -6.82121992e-01 -2.94861104e-03 -2.03955352e-01
-1.43533945e-01 -1.87082171e-01 -6.08729720e-01 -9.70355347e-02
-2.52682269e-01 -4.22604114e-01 1.11687648e+00 1.39159948e-01
9.65148509e-01 -2.69954145e-01 -7.86043227e-01 -2.44815081e-01
1.31336713e+00 3.35693896e-01 5.18118560e-01 1.77602112e-01
3.77200544e-01 7.16264963e-01 1.08327770e+00 7.55547881e-01
4.48818743e-01 5.43650925e-01 3.66407365e-01 2.04750389e-01
1.51105836e-01 -7.05577016e-01 3.82177562e-01 8.84411216e-01
-2.32384518e-01 -6.02255821e-01 -2.81398952e-01 3.01347435e-01
-2.21143937e+00 -1.20440733e+00 -2.69815505e-01 2.29761243e+00
4.42602605e-01 1.71566755e-01 2.65353292e-01 -4.59361938e-04
6.80330753e-01 2.63255805e-01 -4.22954589e-01 -1.06071734e+00
1.24323577e-01 5.62764287e-01 3.15330029e-01 7.70172596e-01
-4.93436873e-01 7.86769271e-01 7.59156799e+00 8.52513492e-01
-4.34816927e-01 -2.06939713e-03 3.08224201e-01 -2.87865520e-01
-1.13422823e+00 1.40038222e-01 -9.27386284e-01 3.19625698e-02
1.15365028e+00 -5.09771109e-01 1.67252809e-01 8.82363260e-01
1.21756665e-01 9.82016549e-02 -1.05827439e+00 4.87720132e-01
6.65265739e-01 -1.75939012e+00 6.65413141e-01 -2.24304244e-01
9.04488027e-01 -1.06883429e-01 3.75289410e-01 2.54710704e-01
9.60055768e-01 -9.17508304e-01 2.97153085e-01 6.54689729e-01
1.40302628e-01 -6.74897671e-01 6.06940329e-01 -1.89448684e-01
-1.31015301e+00 -3.85323554e-01 -7.63614297e-01 -3.37514430e-01
2.96427220e-01 6.45893991e-01 -6.03729844e-01 5.05781710e-01
5.07091641e-01 1.56907105e+00 -8.44245732e-01 8.67608070e-01
-2.22550660e-01 5.34216046e-01 3.24063808e-01 -5.73933244e-01
5.39342463e-02 -1.04439788e-01 1.58500254e-01 1.25179505e+00
5.48343360e-01 1.40320212e-01 1.51846364e-01 3.60938930e-03
5.68835363e-02 5.69913268e-01 -6.14077032e-01 -9.23460796e-02
5.47073305e-01 1.28241646e+00 -7.70376682e-01 -5.26556432e-01
-6.34159684e-01 8.86301517e-01 2.18162119e-01 9.28418785e-02
-3.62045884e-01 -2.68503010e-01 5.88703811e-01 2.88765013e-01
5.06372929e-01 1.99708223e-01 -4.47760671e-01 -9.18386161e-01
-5.73422670e-01 -6.94424331e-01 9.53811705e-01 -9.61747944e-01
-1.65680242e+00 6.26044214e-01 -3.69792849e-01 -1.18827951e+00
-2.74254292e-01 -4.64417934e-01 -8.54224265e-01 5.23570478e-01
-1.05368447e+00 -7.23490715e-01 2.05853254e-01 4.68704313e-01
6.77896976e-01 -4.33281898e-01 1.04001904e+00 -2.08784430e-03
4.67177510e-01 3.79327953e-01 2.46894911e-01 -8.50719988e-01
8.31436932e-01 -1.49247134e+00 3.85218710e-01 1.54243588e-01
8.01587582e-01 1.03788006e+00 8.75585973e-01 -9.74447012e-01
-1.12980080e+00 -7.41946816e-01 1.44198561e+00 -9.52998042e-01
9.56656516e-01 1.74401969e-01 -6.79516375e-01 3.61526102e-01
3.47548723e-01 -4.16356355e-01 1.44516647e+00 8.43309522e-01
-7.46284723e-01 4.91861207e-03 -9.93821204e-01 3.98614466e-01
1.24036002e+00 -7.07524538e-01 -1.21508265e+00 5.10089457e-01
6.75439596e-01 5.55999398e-01 -1.02615154e+00 2.08438575e-01
7.75671601e-01 -9.72416759e-01 1.37998974e+00 -1.13421237e+00
8.67824435e-01 -4.57431078e-02 -1.98738322e-01 -1.90349245e+00
-6.05365098e-01 -3.45686644e-01 -3.52263212e-01 6.36773467e-01
1.24511826e+00 -4.04419035e-01 1.02605939e+00 8.15222979e-01
-3.00249070e-01 -8.75207186e-01 2.96640601e-02 -3.87929469e-01
-2.60744840e-01 -2.84666955e-01 5.77232659e-01 8.22158337e-01
1.11560512e+00 7.77282417e-01 -3.65363359e-01 -4.70746428e-01
1.47504315e-01 3.77919555e-01 1.43039003e-02 -1.81757522e+00
-2.99980074e-01 -6.14382029e-01 -4.17428687e-02 -1.36622274e+00
1.19648082e-02 -1.29034770e+00 -3.13712567e-01 -2.11373830e+00
9.37146023e-02 -3.52557033e-01 -8.68532479e-01 -6.39365166e-02
-6.55944571e-02 4.49971557e-01 -1.54021308e-01 2.52446115e-01
-1.44275033e+00 2.19776019e-01 1.01819432e+00 4.30867076e-02
-1.42976269e-01 2.05929950e-01 -1.34002686e+00 5.48335373e-01
7.87909567e-01 -5.65850496e-01 -1.24251890e+00 -2.57353842e-01
1.34897995e+00 2.17605278e-01 -3.28667581e-01 -2.08847910e-01
5.28921008e-01 -3.36197734e-01 3.96234930e-01 -4.78961498e-01
3.33937615e-01 -6.28820658e-01 -2.18229294e-02 6.05428927e-02
-1.44337797e+00 5.28580785e-01 -3.86499703e-01 1.17669129e+00
-2.05570444e-01 -4.14962918e-01 -9.46316309e-03 -1.93111241e-01
-7.00032890e-01 3.14609796e-01 -7.42672145e-01 -1.70983940e-01
1.32060647e-01 -3.32635969e-01 -5.00165462e-01 -9.19614673e-01
-1.02657390e+00 2.72934455e-02 2.22194105e-01 8.29294443e-01
1.17257607e+00 -1.35028648e+00 -2.34250471e-01 -3.21261972e-01
6.15431309e-01 -1.25047708e+00 -7.49523863e-02 4.20071959e-01
5.59852123e-01 6.57819688e-01 1.44515023e-01 3.32387239e-01
-1.62163329e+00 7.59882212e-01 -2.04320028e-01 -3.33230257e-01
-2.87502944e-01 1.12636447e+00 -4.60388541e-01 -3.97341549e-01
1.22981317e-01 9.67711583e-02 -1.29966879e+00 3.90299916e-01
8.75557959e-01 5.63898802e-01 3.21177304e-01 -2.11332336e-01
-2.43011862e-01 3.80154490e-01 -3.03743571e-01 -2.58766681e-01
1.30779147e+00 -6.97706640e-01 -5.81312366e-02 5.42185068e-01
8.36483121e-01 3.86615098e-01 -3.46018404e-01 -4.83853221e-01
7.38713682e-01 -9.31887984e-01 3.56195360e-01 -1.43208110e+00
-8.70820880e-01 5.03427625e-01 1.28688127e-01 6.00170553e-01
6.15308285e-01 1.70218065e-01 7.68854916e-01 7.24736631e-01
1.21421695e-01 -1.40195584e+00 4.26637232e-01 5.05584419e-01
7.80079484e-01 -1.03097403e+00 1.90666243e-01 -2.35119089e-01
-9.84134316e-01 1.24982774e+00 4.64065462e-01 -1.49298593e-01
1.06039643e+00 -2.44304419e-01 1.86973736e-01 -5.64698577e-01
-1.53836942e+00 -1.27528667e-01 1.18740070e+00 6.77822471e-01
1.22029960e+00 -1.01370476e-01 -7.73896754e-01 1.04853559e+00
-1.49291307e-01 -2.91837245e-01 2.75165319e-01 7.19627082e-01
-1.02323604e+00 -1.39831924e+00 2.69899696e-01 1.31116784e+00
-2.85880536e-01 -7.46224761e-01 -8.83547425e-01 -2.28923529e-01
-4.73012239e-01 1.56866050e+00 -8.67390558e-02 -9.78673697e-01
5.49940653e-02 1.19101062e-01 1.64663151e-01 -1.31560802e+00
-1.19453776e+00 -2.52310693e-01 7.02357352e-01 -7.06962585e-01
-5.30154943e-01 -4.80403394e-01 -1.29072607e+00 -1.42166078e-01
-6.00390434e-01 1.01752591e+00 6.69260085e-01 8.33035290e-01
7.12517321e-01 9.83431116e-02 7.40738094e-01 -4.90736306e-01
-6.45303428e-01 -7.74917126e-01 -7.05087900e-01 4.97724086e-01
-2.65614003e-01 -7.31515884e-01 -5.20388901e-01 -2.59602070e-01] | [10.033125877380371, 5.759647846221924] |
72582cec-f07f-4fe9-b533-70a036604e70 | content-and-style-aware-generation-of-text | 2204.05539 | null | https://arxiv.org/abs/2204.05539v1 | https://arxiv.org/pdf/2204.05539v1.pdf | Content and Style Aware Generation of Text-line Images for Handwriting Recognition | Handwritten Text Recognition has achieved an impressive performance in public benchmarks. However, due to the high inter- and intra-class variability between handwriting styles, such recognizers need to be trained using huge volumes of manually labeled training data. To alleviate this labor-consuming problem, synthetic data produced with TrueType fonts has been often used in the training loop to gain volume and augment the handwriting style variability. However, there is a significant style bias between synthetic and real data which hinders the improvement of recognition performance. To deal with such limitations, we propose a generative method for handwritten text-line images, which is conditioned on both visual appearance and textual content. Our method is able to produce long text-line samples with diverse handwriting styles. Once properly trained, our method can also be adapted to new target data by only accessing unlabeled text-line images to mimic handwritten styles and produce images with any textual content. Extensive experiments have been done on making use of the generated samples to boost Handwritten Text Recognition performance. Both qualitative and quantitative results demonstrate that the proposed approach outperforms the current state of the art. | ['Mauricio Villegas', 'Alicia Fornés', 'Marçal Rusiñol', 'Pau Riba', 'Lei Kang'] | 2022-04-12 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 5.10474682e-01 -3.79637837e-01 -1.83810666e-02 -4.92672026e-01
-4.51954126e-01 -7.94160664e-01 6.81143403e-01 -3.95039380e-01
-1.77550524e-01 6.34783745e-01 -3.68863642e-01 -1.82075649e-02
3.75617951e-01 -6.33990943e-01 -5.68861187e-01 -7.46549368e-01
6.59520388e-01 7.29202569e-01 1.84827790e-01 -2.89834924e-02
4.44436044e-01 7.69283056e-01 -1.47714233e+00 4.58343744e-01
1.05783355e+00 8.34825516e-01 3.07234287e-01 6.72037363e-01
-5.32959580e-01 6.03896081e-01 -1.07024336e+00 -3.93645495e-01
3.63870203e-01 -6.74654961e-01 -2.79263973e-01 1.09502912e+00
6.17779374e-01 -3.22882116e-01 -3.14606041e-01 1.04907370e+00
4.25166845e-01 -5.58114387e-02 7.84665763e-01 -1.00016999e+00
-8.26533198e-01 4.78309184e-01 -7.72256076e-01 -3.56837273e-01
1.45375624e-01 2.80770063e-01 4.38225120e-01 -1.06104803e+00
7.01508760e-01 9.98839378e-01 2.77147502e-01 6.63426101e-01
-1.29655969e+00 -5.47869682e-01 -7.61972070e-02 -2.38967061e-01
-1.18454194e+00 -3.26084495e-01 1.01604378e+00 -3.66243780e-01
4.29366499e-01 3.43309224e-01 4.44715619e-01 1.45437527e+00
-8.94528776e-02 1.08417273e+00 1.43443096e+00 -7.95782149e-01
2.38740236e-01 6.02853000e-01 2.06100345e-02 6.11326218e-01
5.16211875e-02 -1.64256260e-01 -3.98649603e-01 1.20944321e-01
9.55611229e-01 -1.37566760e-01 -3.76974285e-01 -5.30699611e-01
-1.21684694e+00 5.86954236e-01 7.54738674e-02 3.12169433e-01
-2.83136088e-02 -3.63021016e-01 3.65969181e-01 2.37121582e-01
2.82191038e-01 4.38101381e-01 3.93545069e-02 -1.58467337e-01
-1.27722073e+00 6.90353010e-03 9.51647639e-01 1.14939690e+00
5.51095724e-01 2.84959167e-01 -1.04775786e-01 1.48781657e+00
2.84672454e-02 9.50924039e-01 8.15511405e-01 -2.96222270e-01
8.37361634e-01 8.48084927e-01 1.63715124e-01 -9.34246898e-01
1.15084574e-01 -1.68239325e-01 -9.84888017e-01 4.43505764e-01
6.77734852e-01 1.13813719e-02 -1.32612491e+00 8.80733669e-01
-3.51807177e-02 -6.31645560e-01 4.57254425e-02 9.44762051e-01
3.71113330e-01 7.46733665e-01 -3.56514961e-01 -9.99684930e-02
8.88211548e-01 -1.25300419e+00 -6.81992233e-01 -3.30373555e-01
3.15326422e-01 -1.18022263e+00 1.56847727e+00 6.12109900e-01
-7.74621129e-01 -6.52620137e-01 -1.24740875e+00 3.31416398e-01
-2.16212586e-01 9.32346165e-01 2.06265733e-01 9.89480376e-01
-5.03452837e-01 2.32763410e-01 -7.06678808e-01 -3.54866356e-01
3.73649538e-01 1.54458016e-01 -3.80288929e-01 -2.81948060e-01
-5.48368454e-01 5.27079642e-01 4.21966791e-01 3.38719100e-01
-5.26448071e-01 -1.49076030e-01 -4.50887144e-01 -1.80129230e-01
3.69388640e-01 1.16611734e-01 9.84144688e-01 -1.47315836e+00
-1.84799361e+00 6.61322594e-01 1.52579531e-01 1.13419183e-02
1.25593996e+00 4.74155508e-02 -5.85429549e-01 7.54788145e-02
-2.56058961e-01 5.46607256e-01 1.28699303e+00 -1.59943199e+00
-4.64655250e-01 -2.07294688e-01 -5.98339498e-01 1.62023365e-01
-5.29919982e-01 -1.28916308e-01 -6.84234083e-01 -1.00551903e+00
1.72437146e-01 -1.07341552e+00 9.40698013e-02 7.87736326e-02
-5.29725254e-01 2.16625512e-01 1.28709912e+00 -5.50478518e-01
6.82004392e-01 -2.12482977e+00 -1.52912915e-01 4.38491583e-01
-2.18538299e-01 5.25229096e-01 -2.44631648e-01 5.34730196e-01
3.71374249e-01 -4.30295989e-02 -2.01499596e-01 -2.55805790e-01
-6.27370998e-02 2.13230744e-01 -5.29026330e-01 8.42625797e-02
3.14665198e-01 6.43417239e-01 -5.56027651e-01 -4.51925099e-01
1.80434018e-01 2.10767567e-01 -1.93856340e-02 3.42530996e-01
-5.02070606e-01 3.83074552e-01 -4.34200197e-01 8.21746290e-01
6.98920190e-01 -2.06271887e-01 1.47093415e-01 -8.84049162e-02
5.39502650e-02 -5.47400594e-01 -1.18490803e+00 1.24343252e+00
-4.61878955e-01 7.86517262e-01 -3.10732812e-01 -6.01136148e-01
1.58323848e+00 2.65354440e-02 -6.52352348e-02 -6.28378272e-01
1.43701091e-01 4.88872617e-01 -5.27808815e-03 -4.70792323e-01
7.10129201e-01 1.90828562e-01 2.09664136e-01 6.13266289e-01
-1.95310548e-01 -5.01184642e-01 4.16629255e-01 -1.02145754e-01
5.24213135e-01 3.38065684e-01 -1.62805274e-01 -7.05293864e-02
6.71018362e-01 1.45213678e-01 1.12434894e-01 7.93822825e-01
2.15644345e-01 8.58287871e-01 4.07001346e-01 -5.94703019e-01
-1.50131202e+00 -7.08270848e-01 2.22342238e-02 8.11809778e-01
1.43227549e-02 2.12037861e-01 -8.69126976e-01 -8.89304638e-01
-1.80815160e-01 5.61105847e-01 -5.70996761e-01 2.00566322e-01
-5.45677900e-01 -7.72420645e-01 7.27111518e-01 6.28008425e-01
8.35642993e-01 -1.22563171e+00 -4.22777414e-01 2.69376785e-01
1.12974502e-01 -1.08943391e+00 -6.44852102e-01 1.11070946e-01
-8.09959292e-01 -8.40036988e-01 -1.24903846e+00 -8.78632963e-01
1.19834507e+00 -3.17446664e-02 6.21091306e-01 -5.35640791e-02
-4.75381464e-01 1.72278479e-01 -5.30284166e-01 -1.84837714e-01
-9.23404813e-01 8.60394537e-02 -2.43593320e-01 5.07849932e-01
5.68067543e-02 6.62331469e-03 -3.16974998e-01 8.19368601e-01
-1.22079360e+00 2.88550407e-01 8.82018387e-01 1.35711968e+00
4.73174781e-01 -2.06347182e-02 3.77881140e-01 -1.09215069e+00
7.45358944e-01 2.99310386e-01 -8.69455397e-01 6.35914326e-01
-6.36068165e-01 2.24383727e-01 1.02541554e+00 -9.83707547e-01
-1.35239434e+00 2.47248217e-01 2.98965454e-01 -3.64152551e-01
-3.44500184e-01 2.38934040e-01 -1.37888476e-01 -1.87456921e-01
7.76910186e-01 8.13105464e-01 5.70066348e-02 -4.04008210e-01
9.16132331e-02 1.15479374e+00 5.02354860e-01 -6.89992130e-01
7.56968379e-01 3.09920102e-01 -2.46029645e-01 -1.13304472e+00
-4.67775762e-01 8.24316125e-03 -8.14614356e-01 -2.83869326e-01
3.89041871e-01 -4.15107220e-01 -6.78023249e-02 1.15079594e+00
-8.22572231e-01 -5.61488509e-01 -7.38166347e-02 1.94025829e-01
-3.80645961e-01 6.31111741e-01 -5.47490120e-01 -8.69857788e-01
-2.05750629e-01 -1.28854692e+00 1.01892495e+00 1.98562890e-01
1.28282551e-02 -8.90658975e-01 -1.45077661e-01 4.08231378e-01
4.17067528e-01 5.62717021e-02 9.19935524e-01 -6.90816462e-01
-5.32210708e-01 -7.03922212e-01 -3.91767383e-01 7.06124902e-01
4.69504982e-01 3.67658734e-01 -8.71932566e-01 -3.42947036e-01
-2.30980694e-01 -6.69823706e-01 4.64097291e-01 -3.16138744e-01
1.15667915e+00 -2.41124958e-01 -7.02031255e-02 2.96572685e-01
1.22297525e+00 3.97821546e-01 5.65127552e-01 3.30098122e-01
7.10244358e-01 4.31555718e-01 6.56399250e-01 4.83953893e-01
-2.28401914e-01 5.33663750e-01 -1.87024176e-01 -1.71353832e-01
-1.57439813e-01 -3.38312119e-01 2.73396283e-01 6.49363518e-01
-5.07281236e-02 -6.55664027e-01 -9.29043531e-01 8.79801884e-02
-1.47126794e+00 -7.58908033e-01 -6.94735348e-02 2.22954464e+00
1.00216663e+00 2.61469692e-01 5.98057657e-02 4.44771796e-01
8.33448410e-01 2.10993942e-02 -6.38958693e-01 -2.81501114e-01
-3.83563399e-01 -1.56571209e-01 5.01242042e-01 1.25491783e-01
-8.01143765e-01 9.99056399e-01 5.92828417e+00 9.43255305e-01
-1.81014609e+00 -6.32852316e-01 8.94054472e-01 2.40290880e-01
1.37131624e-02 -4.07602906e-01 -8.69818985e-01 6.21389627e-01
3.93129319e-01 7.52025545e-02 5.18420041e-01 7.74615169e-01
3.99121307e-02 -1.42183214e-01 -1.06216526e+00 1.13992250e+00
3.87061924e-01 -1.03119802e+00 3.88609648e-01 -8.60399231e-02
1.01695716e+00 -4.60906476e-01 1.90457746e-01 6.50188625e-02
2.19373358e-03 -9.65466917e-01 7.33374119e-01 4.26811427e-01
1.10352302e+00 -5.48908055e-01 7.50713527e-01 4.72217500e-01
-7.18326926e-01 1.79710016e-01 -4.72476155e-01 3.68044764e-01
-3.33787739e-01 4.78823245e-01 -1.17146206e+00 3.32768798e-01
1.72422498e-01 3.98081481e-01 -8.35535228e-01 8.89040589e-01
-7.33870640e-02 6.01788342e-01 -2.92018861e-01 -4.74461257e-01
2.67388403e-01 -3.69687498e-01 5.92741072e-02 1.20853293e+00
3.96703392e-01 -3.40727419e-01 2.36245781e-01 8.95177126e-01
-7.31012821e-02 3.29179078e-01 -4.92553532e-01 -5.63395560e-01
2.85838783e-01 1.13740146e+00 -1.07031298e+00 -4.81860816e-01
-3.06025118e-01 1.38813341e+00 5.49264476e-02 4.58963573e-01
-6.65334284e-01 -5.54991007e-01 -3.20384294e-01 8.38720985e-03
2.10036516e-01 -2.77061194e-01 -4.86709237e-01 -1.15563464e+00
3.23667169e-01 -1.28394604e+00 -1.78151727e-01 -6.67811990e-01
-1.31845081e+00 9.14200306e-01 -4.57933307e-01 -1.38924742e+00
-2.23533735e-01 -8.37583542e-01 -3.84206325e-01 8.91746104e-01
-8.65397930e-01 -1.29811609e+00 -6.14180326e-01 4.53866929e-01
9.53956544e-01 -7.23838091e-01 7.09367037e-01 -2.75294855e-03
-5.13587594e-01 8.65468681e-01 5.70586324e-01 3.56869400e-01
8.69921744e-01 -1.16371870e+00 3.12725425e-01 6.34589255e-01
3.25231582e-01 1.85201779e-01 5.67602456e-01 -6.78301096e-01
-1.44089520e+00 -1.04069078e+00 3.73512805e-01 -2.07898974e-01
5.12812078e-01 -5.52443981e-01 -1.00548732e+00 4.05642241e-01
6.76307231e-02 -2.14303192e-02 5.49613476e-01 -4.13451731e-01
-3.51294518e-01 -1.95619762e-01 -1.02128208e+00 7.36201823e-01
5.16589284e-01 -1.62769496e-01 -4.17210788e-01 2.64060378e-01
-2.34719828e-01 -4.80974644e-01 -5.43642461e-01 7.64640942e-02
6.10787034e-01 -8.64929795e-01 3.57721388e-01 -1.44745350e-01
5.64931810e-01 -2.80401468e-01 6.04173020e-02 -1.25989246e+00
1.86301649e-01 -4.45177317e-01 3.25208873e-01 1.45236385e+00
5.82196414e-01 -5.64706564e-01 9.54271138e-01 7.52397180e-01
1.90183967e-01 -5.84123731e-01 -3.18280995e-01 -1.05186200e+00
-4.00795043e-02 1.91190541e-01 4.80137438e-01 8.25252235e-01
-1.73884973e-01 2.03529015e-01 -5.73791027e-01 -1.04762286e-01
6.38106585e-01 4.38180298e-01 1.10830545e+00 -8.76044035e-01
-3.95738780e-01 -3.49088758e-01 -3.72634530e-01 -1.14807427e+00
-1.00478001e-01 -5.90685487e-01 2.32508838e-01 -1.07620311e+00
1.28094733e-01 -6.83194339e-01 2.84609646e-01 4.09907103e-01
-5.02820872e-02 5.84740222e-01 2.53097922e-01 3.74303639e-01
-2.40647852e-01 5.13656914e-01 1.62235451e+00 -4.24189806e-01
-2.81541914e-01 5.29416166e-02 4.87920037e-03 4.56276417e-01
8.06346416e-01 -1.78776130e-01 -3.75753433e-01 -4.83032644e-01
-2.19182089e-01 1.37588456e-01 1.25111770e-02 -1.10692668e+00
1.01819754e-01 -2.60852516e-01 9.20992315e-01 -6.49981499e-01
1.89363003e-01 -8.20621252e-01 9.93271638e-03 3.70568663e-01
-5.13442874e-01 -2.46361680e-02 5.44675551e-02 4.15165961e-01
-4.36731517e-01 -4.71210480e-01 9.60660458e-01 -7.61005189e-03
-4.23400283e-01 8.19370970e-02 -4.01758730e-01 -1.76352412e-01
9.18387175e-01 -5.31218767e-01 -2.32853338e-01 -2.06598237e-01
-4.41987872e-01 -1.36656180e-01 8.97931993e-01 4.98520911e-01
6.37157202e-01 -1.19778359e+00 -6.92267299e-01 6.08280957e-01
3.34041595e-01 -2.01649949e-01 2.43000705e-02 3.21888685e-01
-9.82635498e-01 2.20867008e-01 -4.87748265e-01 -7.23037243e-01
-1.23163462e+00 5.30178785e-01 2.71446764e-01 -1.30278558e-01
-5.37131786e-01 4.29618895e-01 -3.21399793e-02 -2.89780259e-01
3.81057739e-01 -1.87750474e-01 1.72861859e-01 -2.40711510e-01
4.50208604e-01 2.62322187e-01 1.32915214e-01 -3.15842211e-01
2.73450315e-01 5.37635267e-01 -4.08145249e-01 -2.37828478e-01
1.04544973e+00 1.22330844e-01 1.85144439e-01 5.43702245e-01
8.05982053e-01 3.04926276e-01 -1.54401410e+00 -1.21085182e-01
7.26844966e-02 -8.58861208e-01 -4.46475476e-01 -1.13533962e+00
-1.07887185e+00 7.97059298e-01 5.99460185e-01 1.74403757e-01
9.73166585e-01 -5.22705495e-01 4.77604210e-01 7.63642251e-01
4.94352430e-01 -1.30221057e+00 3.71094078e-01 2.67208755e-01
1.03046715e+00 -1.43936479e+00 -2.04870597e-01 -3.38126838e-01
-8.78837943e-01 1.66596758e+00 7.06055284e-01 -1.46750718e-01
1.56554338e-02 4.90227222e-01 6.42919302e-01 3.37370157e-01
-2.61098206e-01 4.98707473e-01 2.66425818e-01 4.33762044e-01
4.88064647e-01 -1.34249581e-02 -1.49680987e-01 7.44483769e-02
-7.15472177e-02 4.67095599e-02 7.02663779e-01 9.56614196e-01
-1.95938274e-01 -1.61916411e+00 -6.45562828e-01 5.56091309e-01
-1.64082348e-01 2.05912799e-01 -8.22684765e-01 8.01900387e-01
-3.33917081e-01 7.05093384e-01 -1.64986134e-01 -2.08081216e-01
2.63797343e-01 1.93656638e-01 7.15276599e-01 -3.62230957e-01
-4.06027198e-01 1.76543906e-01 -1.98040232e-01 3.31841521e-02
-1.14926569e-01 -4.04929280e-01 -9.06653345e-01 -3.61306779e-02
-4.60171610e-01 7.62972236e-02 8.19384098e-01 5.71162283e-01
-3.30264233e-02 2.62552828e-01 8.90210867e-01 -8.63413155e-01
-1.01255941e+00 -1.11924088e+00 -7.72526443e-01 7.30188012e-01
3.56305903e-03 -2.96967775e-01 -2.45474488e-01 5.22105157e-01] | [11.842345237731934, 2.2734744548797607] |
bac51659-2717-4b1a-b6ce-e2c7fcccf991 | in-situ-monitoring-additive-manufacturing | 2301.00554 | null | https://arxiv.org/abs/2301.00554v1 | https://arxiv.org/pdf/2301.00554v1.pdf | In-situ monitoring additive manufacturing process with AI edge computing | In-situ monitoring system can be used to monitor the quality of additive manufacturing (AM) processes. In the case of digital image correlation (DIC) based in-situ monitoring systems, high-speed cameras were used to capture images of high resolutions. This paper proposed a novel in-situ monitoring system to accelerate the process of digital images using artificial intelligence (AI) edge computing board. It built a visual transformer based video super resolution (ViTSR) network to reconstruct high resolution (HR) videos frames. Fully convolutional network (FCN) was used to simultaneously extract the geometric characteristics of molten pool and plasma arc during the AM processes. Compared with 6 state-of-the-art super resolution methods, ViTSR ranks first in terms of peak signal to noise ratio (PSNR). The PSNR of ViTSR for 4x super resolution reached 38.16 dB on test data with input size of 75 pixels x 75 pixels. Inference time of ViTSR and FCN was optimized to 50.97 ms and 67.86 ms on AI edge board after operator fusion and model pruning. The total inference time of the proposed system was 118.83 ms, which meets the requirement of real-time quality monitoring with low cost in-situ monitoring equipment during AM processes. The proposed system achieved an accuracy of 96.34% on the multi-objects extraction task and can be applied to different AM processes. | ['Liwei Chen', 'Yuqing Hou', 'Yikai Zhang', 'Hui Li', 'Wenkang Zhu'] | 2023-01-02 | null | null | null | null | ['video-super-resolution'] | ['computer-vision'] | [ 5.32470942e-01 -1.89651385e-01 3.78992945e-01 8.92852992e-02
-3.80864292e-01 2.02696826e-02 -5.92892952e-02 -1.34642452e-01
-1.75903693e-01 3.62652242e-01 -5.35482764e-01 -2.23845392e-02
-5.15862167e-01 -1.06925821e+00 -6.04251266e-01 -4.98051465e-01
2.29618564e-01 4.10696715e-01 3.43848854e-01 -1.93573654e-01
5.35440028e-01 9.89330292e-01 -1.87712717e+00 8.54156375e-01
6.05803490e-01 1.50578129e+00 8.37529182e-01 9.14568722e-01
3.83277014e-02 1.04017818e+00 -6.23565257e-01 -1.63970888e-01
3.23947489e-01 -2.12338880e-01 -4.50897962e-01 3.04808378e-01
3.20721298e-01 -3.35160941e-01 -3.43385994e-01 1.22959924e+00
4.69421983e-01 -3.80059285e-03 2.79455364e-01 -6.95457518e-01
-6.33921623e-01 3.74888361e-01 -8.22114646e-01 6.42997265e-01
3.64250779e-01 1.38320595e-01 1.82895064e-01 -1.16054010e+00
8.47108841e-01 1.12261188e+00 4.35002685e-01 -1.61692455e-01
-8.32341969e-01 -7.92490065e-01 -6.37361228e-01 7.00198531e-01
-1.40834081e+00 -3.01333398e-01 8.48782241e-01 -3.17721218e-01
9.24591303e-01 2.09959313e-01 4.55239207e-01 2.83111572e-01
9.36816514e-01 7.87173286e-02 1.26517940e+00 -3.16251338e-01
1.65738761e-01 -1.76686987e-01 -1.13114104e-01 9.22462642e-01
2.05148682e-01 2.50471383e-01 -7.17314780e-01 4.67405438e-01
1.74782181e+00 -1.10985912e-01 -6.93642646e-02 5.39578140e-01
-1.06708455e+00 4.00614262e-01 1.99829429e-01 4.68266904e-01
-7.11059570e-01 -1.33279815e-01 3.17034900e-01 4.58431274e-01
2.53930807e-01 4.95802104e-01 -3.86539817e-01 -1.38235539e-01
-7.57643998e-01 -3.46450984e-01 4.02061969e-01 1.06202888e+00
4.29577619e-01 5.75553894e-01 9.08280760e-02 7.79670000e-01
-4.49934565e-02 2.69991815e-01 2.90757507e-01 -1.53642929e+00
2.20702201e-01 5.47926664e-01 -1.04052253e-01 -1.08276331e+00
-1.26433074e-01 -2.66413093e-01 -1.09098589e+00 7.28726327e-01
3.33112595e-03 1.79497927e-01 -1.00663960e+00 3.90189886e-01
1.83666259e-01 -4.63324934e-02 -4.92126569e-02 9.72963154e-01
6.99390054e-01 9.64423656e-01 -5.59962153e-01 -5.98707676e-01
1.57138085e+00 -7.94735491e-01 -1.15136564e+00 1.98349375e-02
-2.44428843e-01 -1.22745860e+00 5.43539941e-01 9.43453431e-01
-1.69607639e+00 -1.12800336e+00 -1.43995273e+00 -2.73169801e-02
-5.59549741e-02 3.73534024e-01 4.76839334e-01 -3.02950647e-02
-7.12396264e-01 9.11716223e-01 -7.14749038e-01 1.68831006e-01
3.34561735e-01 5.57211936e-01 -3.28099310e-01 -2.73521632e-01
-7.08940089e-01 8.07311714e-01 2.72861928e-01 4.86137122e-01
-8.86225581e-01 -9.19234931e-01 -5.32425642e-01 1.24479257e-01
5.90855360e-01 -4.69291955e-01 9.72371876e-01 -9.75405753e-01
-1.73315513e+00 5.82258165e-01 1.59270436e-01 -2.54806370e-01
2.30713055e-01 -2.52533972e-01 -8.27803612e-01 7.67871261e-01
-2.19461694e-01 3.15429294e-03 9.57535207e-01 -1.12862980e+00
-9.21695590e-01 -4.41517889e-01 -2.17879757e-01 -1.03503451e-01
3.30259800e-01 5.04215717e-01 -3.68557483e-01 -5.50880432e-01
3.89363825e-01 -4.06914413e-01 -1.90836161e-01 9.36128795e-02
7.66166672e-02 2.55325049e-01 1.15986300e+00 -1.15382826e+00
1.00467861e+00 -1.97452927e+00 2.88964845e-02 -9.65318084e-02
-2.56419461e-02 3.93043995e-01 3.16657990e-01 -9.60551426e-02
-2.43165895e-01 -5.03882468e-01 5.54355532e-02 1.07016154e-01
-7.55328357e-01 6.08972134e-03 3.06414932e-01 3.69573385e-01
1.64119184e-01 4.55126107e-01 -3.68112653e-01 -7.02451289e-01
7.91460335e-01 6.75921798e-01 7.38185178e-03 2.17545450e-01
1.25482798e-01 3.50188583e-01 -1.84031129e-01 1.11573160e+00
9.69570637e-01 -4.69163917e-02 -1.25609469e-02 -9.18861449e-01
-3.55776459e-01 -4.50040430e-01 -1.76087379e+00 1.47509682e+00
-5.92398942e-01 7.75487065e-01 7.21855819e-01 -5.65630555e-01
1.17984807e+00 4.80424106e-01 7.04234660e-01 -9.87194836e-01
4.45741624e-01 5.87359034e-02 -3.37850749e-01 -7.90660262e-01
6.51594579e-01 -9.89118218e-02 3.41258764e-01 -2.62275428e-01
-8.37162882e-02 -1.60008579e-01 5.25497533e-02 -1.85258016e-01
8.73605430e-01 1.12873808e-01 1.40832618e-01 -8.22911188e-02
8.93633485e-01 1.14953995e-01 7.47527659e-01 1.24446586e-01
2.80916214e-01 6.24945581e-01 -1.23928912e-01 -5.22611499e-01
-1.47024035e+00 -8.98394346e-01 -1.64723232e-01 1.88466892e-01
4.65469569e-01 -7.79286325e-02 -5.67504048e-01 3.11332792e-01
-3.10004443e-01 5.14820218e-01 -8.88576359e-02 8.93479735e-02
-9.29950893e-01 -2.78054029e-01 -2.58489847e-01 7.51643836e-01
8.70217860e-01 -1.07245874e+00 -7.75169194e-01 4.86311883e-01
3.30330044e-01 -1.60785282e+00 5.00745848e-02 1.13211116e-02
-1.13353741e+00 -1.08977664e+00 -2.13064268e-01 -1.00696993e+00
6.50749505e-01 2.21158177e-01 1.02215397e+00 -1.07111029e-01
-7.84033000e-01 1.11572593e-01 -6.58288971e-02 -2.44703412e-01
-5.46150744e-01 -6.85738862e-01 -3.29837166e-02 -7.49454498e-02
-1.62883595e-01 -4.36793119e-01 -5.68390429e-01 4.76592273e-01
-7.82880247e-01 3.80777299e-01 8.83166075e-01 6.64359033e-01
1.01316023e+00 9.36082125e-01 5.46898693e-02 -5.87555110e-01
2.32329652e-01 1.36991426e-01 -1.17347562e+00 1.31298751e-02
-7.34926462e-01 -2.80289352e-01 7.82556355e-01 -3.44134808e-01
-1.60241139e+00 1.53183207e-01 2.26526722e-01 -8.62476766e-01
-7.18083903e-02 3.29840124e-01 -1.61815450e-01 -1.45099938e-01
8.44325721e-02 -1.89736471e-01 1.46393225e-01 -6.53091788e-01
-2.28181228e-01 9.03294683e-01 1.05929220e+00 -1.93166643e-01
5.26750445e-01 4.00641799e-01 4.05313998e-01 -1.07028747e+00
-4.16559517e-01 -2.51772821e-01 -5.64251602e-01 -6.61427617e-01
1.26614356e+00 -1.09287453e+00 -9.90928948e-01 6.36588335e-01
-9.68644261e-01 3.51601094e-02 -2.32044220e-01 7.33927071e-01
-4.57533777e-01 3.87620747e-01 -1.41733670e+00 -8.22324634e-01
-6.41357303e-01 -1.12139523e+00 9.61473823e-01 3.28173846e-01
3.54874820e-01 -2.88298786e-01 -8.36173892e-01 9.10002589e-01
3.80922377e-01 3.41496468e-01 6.45898163e-01 3.54448467e-01
-1.05450439e+00 -1.93824112e-01 -3.90510619e-01 6.55906320e-01
1.40987396e-01 2.50522912e-01 -6.60308778e-01 1.32018011e-02
5.58888018e-01 3.96797299e-01 3.30277413e-01 5.87042809e-01
1.22833145e+00 4.44516204e-02 -1.05060391e-01 5.86865067e-01
2.06791353e+00 6.89156413e-01 1.03103817e+00 3.38753372e-01
7.18000293e-01 3.38128000e-01 1.26380277e+00 4.55753416e-01
-3.79339993e-01 8.38791490e-01 6.77268267e-01 -7.09239095e-02
-3.45662057e-01 3.72044444e-01 4.19582009e-01 9.94833648e-01
-6.45328164e-01 3.75707522e-02 -4.83866096e-01 2.76408166e-01
-1.33994055e+00 -1.07504833e+00 -6.32590950e-01 1.95233202e+00
4.63462979e-01 4.14348781e-01 -5.07115543e-01 2.55131841e-01
1.14696181e+00 -3.09751153e-01 -3.20134908e-01 -7.37058163e-01
-1.02783374e-01 6.54492319e-01 7.93911219e-01 4.49589372e-01
-7.24984586e-01 5.38823426e-01 4.89351845e+00 9.46581662e-01
-9.79955137e-01 1.34455398e-01 5.03246665e-01 -8.95240754e-02
5.25891125e-01 -3.25566441e-01 -7.37269104e-01 5.21800637e-01
1.09117436e+00 3.72871906e-02 5.46642780e-01 8.19724917e-01
3.02590162e-01 -3.02239716e-01 -7.80101061e-01 1.28605580e+00
-2.24468932e-02 -1.55628264e+00 -2.02520698e-01 7.21768141e-02
7.47138262e-01 -4.01018888e-01 -9.76139829e-02 -3.38648915e-01
-2.04770975e-02 -9.56276953e-01 6.00754678e-01 6.63294554e-01
9.96968687e-01 -1.12226224e+00 9.38556850e-01 -3.30669843e-02
-1.54818785e+00 -4.84616876e-01 -6.39879823e-01 1.25408068e-01
6.23523116e-01 8.65623772e-01 -4.84580666e-01 9.82066691e-01
9.49222863e-01 3.07960391e-01 -2.51821756e-01 5.76685786e-01
1.96760640e-01 1.40121564e-01 -3.35179985e-01 6.88430250e-01
-7.23878816e-02 -4.83872116e-01 4.75779533e-01 7.58753836e-01
7.67119348e-01 4.51139212e-01 -1.16942368e-01 7.39271820e-01
1.63897499e-01 -3.27291399e-01 -2.01329872e-01 7.81677067e-02
3.92644942e-01 1.40405154e+00 -8.32937181e-01 -3.66319358e-01
-2.73756385e-01 1.04666054e+00 -3.83228123e-01 -1.68917358e-01
-1.05760539e+00 -3.91814172e-01 4.59439933e-01 3.52181345e-01
6.93673313e-01 -4.68091369e-01 -4.52569902e-01 -5.16287923e-01
-1.81575753e-02 -7.72204459e-01 2.67044246e-01 -1.44573641e+00
-9.94597077e-01 6.06495023e-01 -1.96915403e-01 -1.31517816e+00
1.42441422e-01 -9.41046774e-01 -2.32986480e-01 6.79435968e-01
-1.30079424e+00 -1.03364456e+00 -6.89658105e-01 4.93870527e-01
1.24889934e+00 -1.66169018e-01 3.95075798e-01 4.96383190e-01
-7.02427626e-01 -5.05816080e-02 1.05115220e-01 -2.38092035e-01
2.69026279e-01 -8.76088738e-01 -1.41111687e-01 1.09243643e+00
-4.60887671e-01 -1.35440044e-02 7.34626472e-01 -8.78619313e-01
-1.86584425e+00 -9.08213615e-01 1.07694380e-01 4.66324650e-02
2.61166751e-01 2.46683210e-01 -8.84663045e-01 3.92045468e-01
2.69713342e-01 3.53457183e-02 1.87494513e-02 -9.09680724e-01
5.13972223e-01 -3.88961762e-01 -1.45492303e+00 1.07022382e-01
7.99175262e-01 -2.60815024e-01 -2.28668153e-01 2.39090368e-01
7.36100614e-01 -7.25642562e-01 -1.64915931e+00 6.83160484e-01
4.48886186e-01 -8.54122519e-01 1.23068023e+00 1.70953929e-01
7.41974950e-01 -7.06101775e-01 -2.11982593e-01 -5.40418386e-01
-5.72173536e-01 -4.53368962e-01 -3.65358293e-01 1.21356499e+00
6.54979348e-02 -2.08870500e-01 6.50594056e-01 3.73171151e-01
-6.31933451e-01 -6.63655579e-01 -8.33207250e-01 -6.53306007e-01
-8.99047852e-01 -2.68598288e-01 3.55989307e-01 6.75912857e-01
-6.11900806e-01 2.84880131e-01 -1.90369621e-01 6.93192184e-01
7.10095167e-01 1.42268270e-01 2.15665892e-01 -1.30526447e+00
-2.06663996e-01 5.24513908e-02 -7.46736348e-01 -2.97250152e-01
-4.61068451e-01 -2.79149979e-01 -2.75126398e-01 -1.68382442e+00
-9.21520442e-02 6.63504675e-02 -2.47636139e-01 -1.89094812e-01
2.95676112e-01 3.15315992e-01 1.14153586e-01 1.14637189e-01
1.71818994e-02 7.72032738e-02 1.60824406e+00 -6.02319241e-02
8.42715241e-03 -1.30603030e-01 -1.38697475e-01 6.49828076e-01
6.88533306e-01 -5.96263073e-02 -1.29403472e-01 -3.53312910e-01
2.52025902e-01 5.48881590e-01 2.58934081e-01 -1.35388923e+00
5.85191131e-01 -3.42970528e-02 9.70092356e-01 -1.02126729e+00
6.77602589e-01 -1.26182389e+00 9.46194351e-01 5.07894635e-01
1.88922375e-01 3.07404339e-01 8.28436017e-02 3.16555470e-01
-2.78844804e-01 -3.10456008e-01 9.99725103e-01 -5.43145895e-01
-1.04412639e+00 4.60885087e-04 -3.05671781e-01 -8.65648627e-01
1.43353450e+00 -8.35995138e-01 -3.55245471e-01 1.08090624e-01
-6.93242192e-01 -2.51662135e-01 5.54071724e-01 9.35381576e-02
1.25495636e+00 -1.08198750e+00 -6.37337327e-01 4.61166501e-01
-5.27238965e-01 2.49816298e-01 8.12743068e-01 7.24524081e-01
-1.27165341e+00 1.05484046e-01 -6.36140347e-01 -5.76701403e-01
-1.57506132e+00 8.65326941e-01 3.06261688e-01 -1.31390631e-01
-9.94676411e-01 6.50911093e-01 -4.92248416e-01 4.88514900e-01
-2.05686942e-01 -1.85056746e-01 -2.07715765e-01 -5.45717217e-02
6.24354362e-01 1.05176330e+00 3.32210988e-01 -5.65371811e-01
-5.10058887e-02 9.03846681e-01 -2.51485705e-02 1.17348097e-01
1.65353918e+00 -1.13226734e-01 -2.68149942e-01 1.15041964e-01
9.74038184e-01 -1.00985296e-01 -1.15708935e+00 7.52168596e-02
-3.24239463e-01 -6.84893012e-01 6.82268739e-01 -6.72277868e-01
-1.62360668e+00 5.65089881e-01 1.15556037e+00 -1.38286233e-01
1.59066010e+00 -1.49354458e-01 9.72432911e-01 -5.68101779e-02
5.47351420e-01 -1.53880548e+00 -2.08784137e-02 2.29786783e-02
8.47509086e-01 -9.13731158e-01 6.70510709e-01 -8.81121159e-01
-6.03157938e-01 1.61952877e+00 8.02524567e-01 -1.93527535e-01
3.33709508e-01 8.86583269e-01 -3.63918900e-01 -6.91892207e-01
-6.27093792e-01 2.60751486e-01 -1.31049529e-02 3.59512538e-01
1.71848305e-03 -1.85473964e-01 7.85638243e-02 1.68401003e-01
5.57868034e-02 3.62414062e-01 8.04967225e-01 1.12170506e+00
-5.19497752e-01 -4.46117848e-01 -7.40316451e-01 4.24584359e-01
-6.98389351e-01 3.14059645e-01 3.46761107e-01 6.10940099e-01
2.72716761e-01 1.02049088e+00 4.33007181e-01 -4.62919950e-01
5.25278926e-01 -4.34736669e-01 6.70279562e-01 -2.03620553e-01
-5.66168368e-01 1.52737394e-01 1.94486320e-01 -1.00665390e+00
-4.39844429e-01 -4.77899760e-01 -1.36644614e+00 -3.72947246e-01
-5.11312127e-01 -2.24466935e-01 1.02863741e+00 4.28159893e-01
2.97524899e-01 1.25839841e+00 6.18677437e-01 -8.12946141e-01
1.15526311e-01 -8.52532506e-01 -9.51445639e-01 -6.68021291e-02
-3.51354033e-02 -5.12797594e-01 -1.73862487e-01 4.69231963e-01] | [10.972984313964844, -2.1345794200897217] |
acf67b5d-c4f5-440b-977f-d461b29a994d | generalist-equivariant-transformer-towards-3d | 2306.01474 | null | https://arxiv.org/abs/2306.01474v2 | https://arxiv.org/pdf/2306.01474v2.pdf | Generalist Equivariant Transformer Towards 3D Molecular Interaction Learning | Many processes in biology and drug discovery involve various 3D interactions between different molecules, such as protein and protein, protein and small molecule, etc. Designing a generalist model to learn universal molecular interactions is valuable yet challenging, given that different molecules are usually represented in different granularity. In this paper, we first propose to universally represent a 3D molecule as a geometric graph of sets, in contrast to conventional single-level representations. Upon the proposed unified representation, we then propose a Generalist Equivariant Transformer (GET) to effectively capture both sparse block-level and dense atom-level interactions. To be specific, GET consists of a bilevel attention module, a feed-forward module and a layer normalization module, where, notably, each module is E(3) equivariant to meet the symmetry of 3D world. Extensive experiments on the prediction of protein-protein affinity, ligand binding affinity, and ligand efficacy prediction verify the effectiveness of our proposed method against existing methods, and reveal its potential to learn transferable knowledge across different domains and different tasks. | ['Yang Liu', 'Wenbing Huang', 'Xiangzhe Kong'] | 2023-06-02 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 2.6293585e-01 -6.1612416e-02 -2.1707812e-01 -4.2545781e-01
-4.8952329e-01 -5.3120017e-01 5.8588916e-01 3.3808339e-01
-2.0394111e-02 1.0838958e+00 3.0083555e-01 -4.0232402e-01
-2.1186908e-01 -7.4542814e-01 -1.1744449e+00 -1.0283538e+00
1.5948053e-02 6.8639100e-01 -7.8188907e-03 -8.2664385e-02
9.4277561e-02 7.5026327e-01 -9.3540961e-01 3.7488556e-01
8.9942449e-01 7.5806886e-01 -6.1561972e-02 2.7936929e-01
-1.2758221e-01 5.5328304e-01 -1.2381869e-01 -5.2661365e-01
-3.6415037e-02 -3.2696670e-01 -7.7966940e-01 -3.8901180e-02
4.4065720e-01 1.9662909e-01 -3.3621365e-01 8.5604870e-01
5.5817443e-01 -2.8317438e-02 1.1352358e+00 -9.0976644e-01
-9.3276840e-01 3.9891154e-01 -5.6666523e-01 -6.2046934e-02
3.1019014e-01 2.6050106e-01 1.2903843e+00 -9.8916447e-01
6.4325058e-01 1.5152996e+00 3.9124328e-01 4.0071309e-01
-1.3670827e+00 -7.3984540e-01 3.6405787e-01 9.0718763e-03
-1.2699504e+00 -1.0874943e-01 5.9597403e-01 -6.1990327e-01
1.0589174e+00 7.4995659e-02 5.9090966e-01 1.2711155e+00
5.8595198e-01 4.8136923e-01 8.4630746e-01 3.6684074e-02
1.0599648e-01 -2.1104515e-01 2.3577784e-01 7.0952982e-01
4.3188673e-01 -4.2558685e-01 -5.9587234e-01 -4.3256417e-01
8.4897280e-01 4.8517379e-01 -2.5725585e-01 -7.1271873e-01
-1.2690781e+00 7.3683488e-01 6.4730084e-01 2.6366201e-01
-4.1758308e-01 2.2463034e-01 2.9484025e-01 9.8403081e-02
2.7145815e-01 4.2687151e-01 -5.9014559e-01 5.3186738e-01
-1.9693245e-01 3.2304195e-01 5.8710289e-01 7.7357525e-01
8.6887163e-01 -3.1175491e-01 -2.1728198e-01 5.0699341e-01
4.2923334e-01 8.5533261e-02 3.5168529e-01 -8.7680131e-02
2.3504913e-01 8.2981324e-01 1.1559846e-01 -9.5677590e-01
-5.6761545e-01 -6.3334960e-01 -1.2503955e+00 -1.8520960e-01
8.2121879e-02 1.7164190e-01 -9.2565870e-01 1.8886232e+00
5.1639962e-01 5.7171565e-01 5.0539646e-02 7.8800851e-01
9.6477026e-01 6.5801859e-01 5.0491333e-01 -1.7136461e-01
1.3534741e+00 -6.8784130e-01 -5.5390006e-01 1.3590911e-01
7.4519265e-01 -4.5479232e-01 7.7720499e-01 -8.6893931e-02
-1.1082541e+00 -3.2744679e-01 -9.9703664e-01 -4.5891002e-01
-5.8751565e-01 -2.1969797e-01 1.1780949e+00 1.2871200e-01
-5.1175332e-01 5.3702950e-01 -6.1819410e-01 -1.6468818e-01
5.4198086e-01 7.2839224e-01 -6.6377211e-01 -6.9765635e-03
-1.4534482e+00 8.5156149e-01 3.4483239e-01 3.8449816e-02
-8.9655071e-01 -1.0489254e+00 -7.7296615e-01 1.7354451e-01
-1.3768086e-02 -1.1936464e+00 6.1441702e-01 -4.7910488e-01
-1.4785665e+00 8.7592661e-01 -3.3197144e-01 -3.0803955e-01
7.9291336e-02 -8.5510440e-02 -9.3187764e-02 -3.1024763e-01
-1.3361178e-01 3.8007584e-01 4.5943421e-01 -1.0193328e+00
-2.1366742e-01 -7.0140916e-01 4.3318939e-01 4.6823660e-01
-9.9970847e-02 -3.3084467e-01 -1.9458041e-01 -5.7631940e-01
1.3117725e-01 -7.8438383e-01 -4.9431929e-01 -8.8880986e-02
-3.9495814e-01 -2.8481919e-01 4.7392806e-01 -2.4888730e-01
7.7747113e-01 -1.9736149e+00 8.6316884e-01 3.8562745e-01
5.3488433e-01 7.6782800e-02 -1.5245053e-01 4.9492192e-01
-5.0549638e-01 1.1577464e-01 -2.9779586e-01 6.0495205e-02
-3.0452378e-02 5.2393634e-02 -1.6354367e-01 6.2737095e-01
3.4719044e-01 1.3134634e+00 -9.2999208e-01 -8.3534189e-02
2.1684580e-01 8.9215666e-01 -6.0286629e-01 8.6130962e-02
-5.6679028e-01 6.8408370e-01 -9.8317933e-01 6.5218586e-01
7.7067500e-01 -7.5428843e-01 4.0280253e-01 -5.1501101e-01
1.7846777e-01 3.3166230e-01 -9.0097117e-01 1.7061204e+00
-2.3813489e-01 -7.2590448e-02 -2.0674214e-01 -1.2279745e+00
8.2063085e-01 3.9303929e-01 6.3732350e-01 -4.0879437e-01
1.2730643e-01 1.6358723e-01 1.6715023e-01 1.9790216e-03
-2.4535251e-01 -2.4487107e-01 4.3208409e-02 2.0443402e-01
2.0704035e-01 8.3760835e-02 -3.6576602e-01 2.0471947e-01
1.0058693e+00 -1.0561713e-01 5.0793886e-01 -3.4068072e-01
8.8537240e-01 -4.3687814e-01 5.0214237e-01 4.5014918e-01
2.4452105e-01 1.9907686e-01 6.9682038e-01 -7.3809820e-01
-8.8514680e-01 -8.7285715e-01 -3.1691173e-01 9.2258686e-01
3.5816616e-01 -3.8178807e-01 -5.7951516e-01 -6.8551797e-01
2.4730743e-01 1.9135205e-03 -7.0108664e-01 -3.8170385e-01
-3.6171189e-01 -1.2537594e+00 1.9255149e-01 1.9280653e-01
2.1197009e-01 -8.5704434e-01 2.5929824e-01 3.5533875e-01
3.0054268e-01 -8.8790119e-01 -5.7952040e-01 5.0584781e-01
-7.4256295e-01 -1.2333659e+00 -5.0856018e-01 -8.5516912e-01
7.3724252e-01 3.5358644e-01 9.7355056e-01 -3.3098976e-03
-3.0536810e-01 -1.3929161e-01 -1.7861550e-01 -3.2622254e-01
5.5196905e-03 3.8484842e-02 2.5855762e-01 2.4956620e-01
4.3073606e-01 -9.6469796e-01 -8.0375022e-01 1.7984843e-01
-9.5425540e-01 1.4014535e-01 7.8182733e-01 1.0713565e+00
9.2169666e-01 -3.8333720e-01 6.2631810e-01 -1.2231333e+00
6.4666635e-01 -5.7410020e-01 -4.1263509e-01 2.4340627e-01
-2.0019826e-01 2.5240609e-01 6.7584962e-01 -3.3617488e-01
-7.1180105e-01 3.5352722e-01 -2.1794479e-01 -2.6589456e-01
-9.7593628e-03 5.8512539e-01 -7.5237375e-01 -4.2997792e-01
4.7060081e-01 3.0596608e-01 -3.2737911e-01 -4.6161395e-01
4.4845235e-01 3.1404692e-01 6.7870438e-02 -6.9423747e-01
7.2741723e-01 4.0422201e-01 5.3827518e-01 -6.4802283e-01
-8.9246494e-01 -1.8509008e-01 -8.8515079e-01 4.5850244e-01
9.3566376e-01 -1.1174848e+00 -1.1183027e+00 4.0307611e-01
-1.2358692e+00 -6.8583712e-02 1.8245630e-01 5.0340170e-01
-3.8664368e-01 2.9337680e-01 -7.2833169e-01 -1.7449127e-01
-3.3826858e-01 -1.3524966e+00 1.2602962e+00 5.6525916e-02
5.4751530e-02 -1.1878954e+00 2.2429495e-01 3.0722916e-01
2.9458707e-02 4.9991947e-01 1.3964776e+00 -6.5557230e-01
-7.3379302e-01 2.8053850e-02 -3.5926607e-01 -1.2524396e-01
5.1307750e-01 -2.2388746e-01 -7.0194477e-01 -4.2008799e-01
-3.7629485e-01 -3.4538141e-01 9.1314453e-01 4.4046170e-01
1.4098599e+00 -3.3654457e-01 -6.4524001e-01 8.6398154e-01
1.1090138e+00 2.5915512e-01 5.0471032e-01 -5.7556018e-02
1.1591914e+00 3.0775645e-01 2.0721820e-01 3.9882419e-01
2.4795920e-01 7.5531262e-01 4.5516884e-01 -4.2562118e-01
1.9843562e-01 -2.7921146e-01 1.9899324e-01 7.3204261e-01
-2.3310420e-01 -2.6586226e-01 -5.5505526e-01 2.3353133e-02
-1.9463052e+00 -9.0443408e-01 -1.4906734e-03 2.1156254e+00
1.0080013e+00 1.0234142e-01 -2.6980575e-02 -4.2283997e-01
5.1070827e-01 1.8552864e-01 -9.7614646e-01 -2.8059265e-01
-2.4326396e-01 4.2897168e-01 4.4955954e-01 6.4222294e-01
-1.0319583e+00 1.0958586e+00 6.3231168e+00 8.5899013e-01
-1.1809900e+00 -1.3721399e-01 7.5674623e-01 1.6645597e-01
-6.2223399e-01 -9.9689305e-02 -7.9694784e-01 3.6384907e-01
5.1753777e-01 -2.1779738e-01 1.5038827e-01 5.0383770e-01
1.8445227e-01 7.5195557e-01 -1.3793753e+00 9.2875320e-01
-2.0786837e-01 -1.7676648e+00 8.3623195e-01 2.1682724e-01
6.8131387e-01 -2.0748855e-01 1.7647083e-01 1.3073087e-01
3.3706477e-01 -1.5537978e+00 5.8983695e-02 5.5312353e-01
6.1666495e-01 -7.7753788e-01 5.2495778e-01 1.1719626e-01
-1.3130312e+00 4.6168339e-01 -4.9346861e-01 1.6636241e-02
-1.7254613e-01 5.7907850e-01 -7.2952688e-01 7.8494805e-01
2.7869070e-01 1.2178359e+00 -1.2355938e-01 5.9667718e-01
-2.0120392e-02 1.1749325e-01 2.2653099e-02 -7.0926458e-02
4.5870525e-01 -4.7207987e-01 3.1826776e-01 1.2102479e+00
-2.7290788e-02 5.1073307e-01 2.7702361e-01 7.4530715e-01
-5.6375480e-01 3.9412549e-01 -6.6067982e-01 -1.0908322e-01
2.9271778e-01 1.1335489e+00 -3.4404999e-01 -1.3176391e-01
-6.1082226e-01 9.5554096e-01 5.4199815e-01 5.3261530e-01
-9.7011054e-01 -1.3921380e-01 1.2043108e+00 -2.4314793e-02
1.2874863e-01 -1.4607184e-01 2.8764725e-02 -1.2853861e+00
-1.9058663e-01 -9.8781312e-01 2.5564194e-01 -4.7797862e-01
-1.5123473e+00 3.1580088e-01 -3.5870823e-01 -9.8609811e-01
4.5835280e-01 -1.0174854e+00 -4.1889280e-01 9.7333211e-01
-1.4827311e+00 -1.3855642e+00 -3.6271282e-03 6.2954563e-01
2.4501567e-01 -1.4175266e-01 9.0787381e-01 4.0065533e-01
-8.0558628e-01 6.5360421e-01 3.3179474e-01 -2.7578485e-01
7.5508505e-01 -1.1999576e+00 5.4703474e-01 3.0960277e-02
6.4067333e-03 1.0820941e+00 4.4643277e-01 -5.6956017e-01
-1.8266158e+00 -1.4044375e+00 4.8451957e-01 -6.6858280e-01
6.1315733e-01 -6.0763001e-01 -1.0882155e+00 7.7252281e-01
-2.2848791e-01 9.2873752e-02 9.4348872e-01 2.5446749e-01
-6.0429311e-01 -1.4367492e-02 -9.8404014e-01 7.8197092e-01
1.3490220e+00 -5.4268259e-01 -3.6151040e-01 8.7264854e-01
9.6329957e-01 -4.8815185e-01 -1.2412572e+00 6.5485060e-01
6.0016346e-01 -5.4245710e-01 1.4115075e+00 -1.3169602e+00
2.6276749e-01 -4.6235967e-01 -1.6630788e-01 -1.1459841e+00
-7.6405650e-01 -4.5707631e-01 -1.9352572e-01 6.6887909e-01
5.0255400e-01 -6.5896666e-01 7.7375114e-01 3.8723525e-01
-8.8672906e-02 -1.0469248e+00 -9.5430213e-01 -3.1434083e-01
2.8850517e-01 2.1363065e-01 7.3006600e-01 1.2411561e+00
2.6322320e-02 9.9734336e-01 -5.4136503e-01 3.3693543e-01
5.1355737e-01 3.0861196e-01 8.5676473e-01 -1.4911388e+00
-4.1201448e-01 -4.4625172e-01 -6.1646867e-01 -1.3478040e+00
3.2336003e-01 -1.1104804e+00 -5.1628417e-01 -1.3491758e+00
7.1450019e-01 -3.4106961e-01 -7.3555207e-01 3.8179183e-01
-2.1533367e-01 8.7790340e-03 -3.1263712e-01 7.9426736e-02
-6.2984651e-01 8.2733530e-01 1.6367321e+00 -4.8916575e-01
-8.5608974e-02 -2.3200628e-01 -1.0207938e+00 5.3520346e-01
4.4846624e-01 -2.1669023e-01 -4.5177847e-01 -1.5738733e-01
2.9453006e-01 -9.5306270e-02 2.4699438e-01 -4.9488968e-01
-4.8095219e-02 -4.7279316e-01 5.9703761e-01 -3.0144781e-01
4.1964260e-01 -7.3664987e-01 4.4819611e-01 4.7413468e-01
-2.8727481e-01 -3.3259195e-01 1.0262055e-01 8.0417913e-01
-8.8866554e-02 3.1776184e-01 8.4556156e-01 -2.5153732e-01
-2.4225804e-01 1.1042347e+00 4.1178841e-02 -2.5667214e-01
1.0299356e+00 8.8827915e-02 -1.9620089e-01 1.8708250e-02
-6.0112041e-01 2.3240194e-01 5.1957011e-01 2.6545817e-01
5.5659950e-01 -1.3320481e+00 -6.4944601e-01 1.3975878e-01
2.0488599e-01 2.0473988e-01 2.6398075e-01 5.9708321e-01
-4.1311234e-01 7.5638312e-01 -2.8258568e-01 -4.8114431e-01
-1.1204998e+00 6.3256025e-01 5.1374340e-01 -3.3797026e-01
-4.2271489e-01 8.3921814e-01 1.0873408e+00 -4.9983683e-01
1.4323837e-01 -2.5738293e-01 -2.5063306e-01 -2.0261797e-01
3.5040748e-01 -2.0928594e-01 -8.1047662e-02 -6.8524557e-01
-5.5080402e-01 9.5555431e-01 -4.2455277e-01 5.5469280e-01
1.4090731e+00 3.3818936e-01 -3.5602033e-01 3.6383250e-01
1.2557338e+00 -1.8917969e-01 -1.0213258e+00 -3.6604366e-01
-2.0047517e-01 -1.6138227e-01 -2.7972755e-01 -4.9948364e-01
-7.2887707e-01 8.0521768e-01 3.0703300e-01 -2.6214471e-01
6.2538803e-01 1.1575663e-01 6.1303741e-01 7.2395855e-01
3.0773854e-01 -2.1064340e-01 4.7033545e-02 5.3041661e-01
9.2355347e-01 -1.1808016e+00 2.1468693e-01 -5.2566236e-01
-2.6142418e-01 8.3066034e-01 6.6363102e-01 -1.4151751e-01
6.5816486e-01 -2.1660148e-01 -4.6533450e-01 -5.0967312e-01
-8.2902455e-01 3.2455385e-02 4.8008400e-01 3.0155480e-01
9.6888375e-01 1.1358303e-01 -4.2517766e-01 6.5839618e-01
3.4562340e-01 -2.1715477e-01 1.8195404e-02 5.3293896e-01
-3.5634771e-01 -1.4365717e+00 7.8444995e-02 3.2609135e-01
-4.2153150e-01 -2.6610073e-01 -7.7829742e-01 6.3070458e-01
9.0569094e-02 3.6409190e-01 -1.5836576e-01 -2.9696190e-01
3.5648015e-01 -6.7750402e-03 5.2412218e-01 -6.8988556e-01
-3.8734108e-01 -3.9555959e-02 -4.1993433e-01 -3.8868138e-01
-5.7398582e-01 -3.3229831e-01 -1.3582094e+00 -2.5154948e-01
-2.4231990e-01 3.4942475e-01 2.0515886e-01 8.7321883e-01
7.5641382e-01 6.5652686e-01 6.1537075e-01 -7.3375446e-01
-1.7564303e-01 -6.9225228e-01 -6.1545378e-01 6.0498697e-01
3.8489798e-01 -9.8363161e-01 -6.3104577e-02 4.0542088e-03] | [5.103739261627197, 5.838719844818115] |
46fcd26f-45b9-4785-b6a2-ab107ec9f69e | tacoformer-token-channel-compounded-cross | 2306.13592 | null | https://arxiv.org/abs/2306.13592v1 | https://arxiv.org/pdf/2306.13592v1.pdf | TACOformer:Token-channel compounded Cross Attention for Multimodal Emotion Recognition | Recently, emotion recognition based on physiological signals has emerged as a field with intensive research. The utilization of multi-modal, multi-channel physiological signals has significantly improved the performance of emotion recognition systems, due to their complementarity. However, effectively integrating emotion-related semantic information from different modalities and capturing inter-modal dependencies remains a challenging issue. Many existing multimodal fusion methods ignore either token-to-token or channel-to-channel correlations of multichannel signals from different modalities, which limits the classification capability of the models to some extent. In this paper, we propose a comprehensive perspective of multimodal fusion that integrates channel-level and token-level cross-modal interactions. Specifically, we introduce a unified cross attention module called Token-chAnnel COmpound (TACO) Cross Attention to perform multimodal fusion, which simultaneously models channel-level and token-level dependencies between modalities. Additionally, we propose a 2D position encoding method to preserve information about the spatial distribution of EEG signal channels, then we use two transformer encoders ahead of the fusion module to capture long-term temporal dependencies from the EEG signal and the peripheral physiological signal, respectively. Subject-independent experiments on emotional dataset DEAP and Dreamer demonstrate that the proposed model achieves state-of-the-art performance. | ['Xinda Li'] | 2023-06-23 | null | null | null | null | ['multimodal-emotion-recognition', 'multimodal-emotion-recognition'] | ['computer-vision', 'speech'] | [ 0.01101531 -0.696022 0.21215975 -0.5246184 -0.9581585 -0.3156082
0.3022957 0.14945237 -0.47645628 0.8281086 0.30202046 0.36713526
-0.2612656 -0.26818335 -0.5410425 -0.83340985 -0.02241809 -0.46930593
-0.36676037 -0.06498421 0.17977136 0.07277316 -1.5773698 0.75479984
1.1143876 1.5178552 0.16523539 0.29593664 -0.172898 0.514674
-0.4369249 -0.27978554 -0.38006502 -0.49829143 -0.44781694 -0.12354165
-0.231162 0.21861973 -0.4054498 1.2354602 0.8851598 0.01816692
0.3528289 -1.5141943 -0.5017829 0.58231807 -0.69622934 0.22816725
0.39708152 -0.15129273 0.7540049 -0.98340684 0.13278316 0.91625184
0.5658016 0.17013636 -1.1799399 -0.90839595 0.29364505 0.6782057
-1.5754474 -0.3736464 1.0569419 -0.23376256 1.0464487 0.1785871
0.7549322 1.3239589 0.67476195 0.9483743 1.541745 -0.04657433
-0.04469676 0.15128908 0.12690102 0.23488705 -0.43354052 -0.16422093
-0.97774655 -0.0572193 0.46605653 0.14841388 -0.48977426 0.17562607
-1.3946548 0.35582507 0.43592095 0.74783266 -0.82089764 -0.01554095
0.72342664 0.17311211 0.3779764 0.15182962 -0.5716608 -0.39231694
-0.79394305 -0.2257462 0.3305491 0.7381653 0.70495135 -0.03229288
-0.30836937 0.8975043 0.2612554 0.4450066 0.7267992 -0.519893
0.27070895 0.5237767 -0.26656637 -1.084277 -0.6111944 -0.34699637
-1.1194531 -0.35276705 -0.16853093 -0.3318989 -0.47782147 2.1168957
-0.04452389 0.50888044 0.20055868 0.9719338 0.7070502 0.5840291
0.38646343 -0.46287313 1.6117705 -0.62659615 -1.2165539 0.0412115
0.24987914 -0.55098933 0.4725243 0.45472014 -1.0357579 -0.58070695
-0.82224053 0.12416863 -0.70009834 0.14476229 0.6832335 0.35959312
-0.71407694 0.15462178 -0.6458403 -0.03258794 0.4099982 0.502922
-0.80297095 0.03723354 -1.6848615 0.92846006 0.49672446 0.61470425
-0.47891682 -0.55725795 -0.777727 0.2784396 0.11656793 -0.4098209
0.6353463 -1.0638306 -1.3662546 0.24729995 -0.49548143 0.10546834
-0.23015477 -0.13749872 -0.93771505 0.19350116 -0.13233635 0.6257798
0.6459308 -1.1052396 -0.55032253 -0.72074604 -0.3830993 0.5836528
-0.5655997 0.16935511 -0.36610004 -0.45480582 0.15828663 -0.4209641
0.14981082 -0.6193449 -0.256585 -0.03635171 0.6002565 -0.7078992
1.2368027 -2.4453075 0.4880791 0.24659541 0.21728258 -0.13917772
-0.27243656 0.35656095 -0.31535923 -0.11392206 -0.1402527 -0.44723493
0.17602302 -0.00987325 -0.00801877 0.35943827 0.33688816 1.075766
-0.6302882 -0.4795367 0.36722147 0.94090915 -0.1975598 0.18527304
0.4896952 0.7867415 -0.2681623 0.7845489 0.84850407 0.09379359
-0.06260736 -0.7389302 -0.15464583 -0.09366537 -0.97713 2.2979329
-0.50625765 0.51043403 0.22490728 -1.1084037 0.81433946 0.72235084
0.8594866 -1.1029251 0.5136436 0.17236984 0.01044713 -0.60669196
0.36229038 -0.39019206 -0.3613706 0.02112254 0.52435076 0.5089358
-0.16564333 -0.03839982 0.8279067 -0.13855132 0.10495427 -0.05283156
0.6184056 -0.5981249 0.6517502 0.28117448 -0.22926351 0.27140424
0.3489265 0.25700203 -0.38577342 -0.81058115 -0.3422708 0.87642926
0.42293796 -0.42538914 -0.6297325 -0.25443798 -0.12967949 0.33051437
-0.72695047 -0.5330244 0.13711202 -1.1061927 0.757697 0.72983015
0.6806854 -0.90155953 -0.53310114 0.4082906 -0.6541398 -1.2389059
-0.29773033 0.5416122 -0.6189342 -0.7518694 -0.60514116 -0.42294917
0.29794276 -0.07231787 0.5495734 -0.43412462 -0.32471484 0.54089504
-0.4463304 -0.26056612 0.5487819 -0.18289492 0.04338462 0.66812277
0.64866984 -0.7572058 -0.40494016 0.14771101 -0.952477 -0.04107535
0.8172636 1.0312418 0.52264524 0.2369478 0.9187592 -0.14715053
0.9300714 -0.64448893 0.01886329 0.40962368 -0.07533977 0.03431078
0.5537648 -0.5376775 -1.3221234 0.00624271 -0.13682768 -0.657627
-0.38446856 0.88459694 -0.7322942 -0.06970577 -0.07640778 0.4917838
-0.366171 -0.27948815 0.39289522 0.9491028 0.67413205 -0.6781562
-0.06711522 0.23817138 -0.3398667 -0.6778012 -0.19233222 -0.47202066
-0.67213595 -0.39254424 1.0579375 -1.2181867 -1.1242665 0.8159973
-1.3169736 0.22181022 0.13206406 0.9846771 -0.31989294 0.07477865
-0.61628646 -0.88580644 -0.269903 -1.2742536 1.1570653 0.35835212
0.12373889 -0.88348657 -0.14923063 0.17804638 0.48668477 0.07748136
0.6567661 -0.42729276 -0.04940086 -0.22290424 -0.39329195 0.37604913
0.05803179 -0.50311035 -1.2133727 0.10746577 0.19133256 -0.3436028
0.76208436 0.44759026 1.289825 0.26402912 -0.33950356 0.42556998
1.2515233 0.36870176 0.8040005 -0.16658464 0.7501794 0.5875662
0.18198495 0.7153994 0.6400714 0.48472136 0.28030884 -0.18720104
0.30781844 0.08158736 0.41661564 1.1727269 -0.01480469 -0.04394037
-0.57634425 0.50118464 -1.9648563 -1.0518839 -0.17197844 2.0179965
0.62869275 -0.43263137 -0.24102789 0.12359919 0.6734004 -0.16213104
-0.45719856 -0.18815473 -0.5680305 0.03556413 0.08698378 0.21164568
-1.0592747 0.5510618 5.558675 0.9794233 -1.3348595 0.42714012
0.37585977 -0.16376723 -0.2796077 -0.2760322 -0.3617412 0.79064214
1.0443856 -0.03478094 0.6388466 0.07339772 0.11805831 -0.32165554
-1.0080436 1.5445936 0.36742502 -0.6796982 -0.39594352 -0.00697721
0.39403847 -0.15164308 0.0788766 0.39125687 -0.6128958 -1.0443815
0.67072976 1.025376 0.6131306 -0.94465864 1.0168331 0.15677446
-1.5795187 -0.18327907 -0.03249959 0.03780419 0.3220866 0.7360222
0.2185404 1.1114882 0.86709654 0.9970481 -0.47293946 0.8959749
0.18041441 0.17061894 -0.32207593 0.19552967 0.01494418 -0.1222159
0.22207826 1.3082994 0.5499784 0.43758976 -0.03898641 0.8967706
-0.07576025 0.1846828 -0.34715047 -0.17783485 0.42511234 1.4461809
-0.44478962 -0.2625248 -0.6152948 1.3006397 0.12683575 0.53737086
-1.1167082 -0.59211224 0.6068858 -0.85643935 0.00791849 -0.05400874
-0.4415444 -1.5149647 0.18310021 -0.706455 0.15522985 -0.94937724
-1.4276617 0.73277247 -0.11422274 -1.0474662 0.30428514 -0.42434195
-0.40000105 1.1379827 -1.5086137 -1.1189506 -0.35838553 1.1209899
0.08129516 0.12761107 0.94456685 0.9242963 -0.75068015 0.5383714
0.12166512 -0.02287324 0.87155634 -0.8443502 -0.7584767 0.5451682
-0.10259337 0.56926036 0.15083964 -0.3401989 -1.5966327 -0.7009013
0.66595817 -0.08353715 0.50076455 -0.46305963 -1.0714847 0.3400435
0.6446445 -0.03879032 0.9337556 0.29332092 -0.3142117 -0.26545703
-0.8639977 0.2240974 0.50690854 -1.0319134 -0.5123495 -0.19104224
0.34631127 -0.16086715 -1.2980094 0.49613848 0.8505473 -0.8880085
0.6114815 -0.35239112 0.19591029 -0.34116894 -0.56405133 -1.5923411
-0.18763955 -0.17842922 0.07311018 1.4667917 0.1521004 -0.54916847
-0.06468752 0.6776302 -0.23818822 -0.7877612 -1.1086729 -0.37343243
-0.19379544 -0.7291476 0.55089396 1.0404723 0.92663866 0.478617
-0.68853927 0.16477326 0.2581701 -0.05412622 0.05197066 -1.0285609
-0.01131569 -0.6380799 -0.5440532 -0.75458956 0.51912206 -0.8888311
0.24267441 -1.2540269 0.44609022 0.07927167 -1.0015016 0.55085
-0.28458318 0.27267823 0.08385219 -0.19676434 -0.79471254 1.1599665
0.8939699 -0.09551539 -0.13441278 -0.75281173 -0.6934255 0.35598442
0.48884612 -0.23244163 -0.21359025 -0.41273475 0.09834508 0.38341823
0.39420635 -1.1754543 0.515633 0.21077926 0.6722948 -0.41336137
0.7847038 -1.0569879 0.16073073 -0.12170245 -0.35547498 0.08496059
0.5474252 0.6215584 -0.75227654 0.41029423 0.5126472 0.21276003
-0.73560065 0.3336498 -0.5633615 -0.30114838 0.93100095 -0.05107754
-0.3130602 -0.14779222 -0.89107656 0.36058956 -0.18492106 0.5219031
0.77813137 -1.7025291 -0.46081915 0.4167714 0.2953553 -0.69518757
1.0612541 1.6457744 0.57005805 0.46411285 -0.66477597 -0.5125058
-1.1377858 0.41450763 0.44071016 0.01339234 -0.10618811 0.7762062
0.26044974 -0.26307622 0.11363945 -0.17917645 -0.36775225 0.5552807
0.52920175 0.20540926 0.18788134 -1.0314385 -0.7032678 0.58603305
0.22073175 -0.4251718 1.0590773 -0.42599007 -0.38043746 0.927093
1.4077679 -0.40584007 -0.8798511 -0.2482364 -0.36922085 -0.28666303
0.412944 -1.0932279 -1.2102793 1.471269 0.71819985 -0.22663441
1.7253877 -0.21082675 0.7840918 0.02732918 0.35438615 -1.3219
-0.03428691 0.343205 0.660718 -1.1643057 -0.4489209 -0.03675187
-0.9065394 0.9433649 0.5165385 0.27882788 0.8609393 0.3271215
-0.175919 -0.20308822 -0.85026944 -0.18044582 0.5109597 0.3224643
0.6484327 0.20496546 -0.36195534 1.4285036 0.41313142 0.14918181
-0.07059243 0.6615871 -0.0430162 -0.9620565 -0.55347306 0.40606725
-0.41395587 -0.34551468 -0.10995854 0.35937896 0.48156282 1.1426555
-0.02042109 -0.89714485 0.3644068 0.5888136 0.57094294 0.05010041
-0.7129636 0.5621677 -0.26385325 -0.7461844 -0.80385596 -0.62950873
-1.3765228 0.05073624 -0.45631385 0.32020664 0.8556999 1.1751786
0.83375937 0.9589815 0.4854838 -0.9499017 0.11928587 -1.1597406
-1.013797 0.374358 0.23184815 -0.87630767 -0.05114207 -0.03919558] | [13.233315467834473, 4.882299423217773] |
601d4675-3421-4328-b8a7-8a814bce4aa0 | continual-predictive-learning-from-videos | 2204.05624 | null | https://arxiv.org/abs/2204.05624v1 | https://arxiv.org/pdf/2204.05624v1.pdf | Continual Predictive Learning from Videos | Predictive learning ideally builds the world model of physical processes in one or more given environments. Typical setups assume that we can collect data from all environments at all times. In practice, however, different prediction tasks may arrive sequentially so that the environments may change persistently throughout the training procedure. Can we develop predictive learning algorithms that can deal with more realistic, non-stationary physical environments? In this paper, we study a new continual learning problem in the context of video prediction, and observe that most existing methods suffer from severe catastrophic forgetting in this setup. To tackle this problem, we propose the continual predictive learning (CPL) approach, which learns a mixture world model via predictive experience replay and performs test-time adaptation with non-parametric task inference. We construct two new benchmarks based on RoboNet and KTH, in which different tasks correspond to different physical robotic environments or human actions. Our approach is shown to effectively mitigate forgetting and remarkably outperform the na\"ive combinations of previous art in video prediction and continual learning. | ['Xiaokang Yang', 'Mingsheng Long', 'Yunbo Wang', 'Siyu Gao', 'Han Lu', 'Wendong Zhang', 'Geng Chen'] | 2022-04-12 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Chen_Continual_Predictive_Learning_From_Videos_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Chen_Continual_Predictive_Learning_From_Videos_CVPR_2022_paper.pdf | cvpr-2022-1 | ['video-prediction'] | ['computer-vision'] | [ 2.28800476e-01 -1.98784530e-01 -3.07265986e-02 1.86662357e-02
-4.95625079e-01 -2.66186327e-01 7.70706534e-01 -1.34756416e-02
-3.40532571e-01 1.04669487e+00 1.88160446e-02 -2.61000893e-03
-4.19665515e-01 -4.72168118e-01 -1.28365636e+00 -6.74648762e-01
-3.19903970e-01 8.54514837e-01 6.24648392e-01 9.45865214e-02
1.44710064e-01 3.14585686e-01 -1.86860609e+00 3.50316137e-01
5.95166445e-01 7.85527110e-01 5.99246621e-01 1.18164444e+00
3.04543287e-01 1.32177830e+00 -4.06369597e-01 -2.67589271e-01
2.39466950e-01 -1.02280304e-01 -5.70078790e-01 1.80043444e-01
2.35006034e-01 -2.19139040e-01 -8.82533789e-01 5.65952897e-01
3.82216841e-01 6.31239593e-01 6.04657710e-01 -1.39843369e+00
-5.46543598e-01 3.16908658e-01 -3.08818966e-01 2.64436483e-01
4.35907513e-01 5.31485021e-01 4.01899368e-01 -7.18331993e-01
5.28518260e-01 1.09285295e+00 9.85604286e-01 5.86251736e-01
-1.22490156e+00 -4.05713469e-01 4.37920511e-01 7.47732759e-01
-1.21296847e+00 -4.85918730e-01 5.55744886e-01 -4.36769664e-01
1.00612032e+00 -7.57006695e-03 8.13360572e-01 1.73106325e+00
7.53591239e-01 1.11057830e+00 9.37163293e-01 -3.25116605e-01
5.47478735e-01 -2.36304812e-02 -1.12421982e-01 4.78930295e-01
2.12532550e-01 1.31804943e-01 -1.07323289e+00 -2.09545478e-01
5.62740386e-01 4.46760058e-01 -3.12204719e-01 -7.58115768e-01
-1.19034445e+00 5.94386272e-02 -4.34373140e-01 -1.45861611e-01
-5.35459638e-01 3.78708005e-01 4.24126148e-01 6.34800434e-01
4.30589855e-01 2.92826384e-01 -7.27932751e-01 -4.82764363e-01
-7.36875653e-01 5.04364431e-01 1.08003533e+00 1.16796720e+00
4.64428574e-01 -8.71682763e-02 -1.82733074e-01 5.83101511e-01
-2.49599218e-01 4.91829306e-01 8.14532340e-01 -1.16915751e+00
2.32027680e-01 -1.44739091e-01 7.69459903e-01 -5.38857281e-01
-3.10646445e-01 -1.73114464e-01 -5.73082030e-01 1.17862396e-01
2.58366853e-01 -1.02533437e-01 -9.20552552e-01 1.53366137e+00
1.55238435e-01 1.15449142e+00 1.20701142e-01 4.46787447e-01
5.14079444e-02 7.32012570e-01 4.12307978e-01 -5.18244386e-01
5.34098625e-01 -1.16848421e+00 -5.45411468e-01 -3.00727665e-01
3.36241513e-01 -4.31847453e-01 1.34882855e+00 9.53567863e-01
-1.00563443e+00 -8.94382417e-01 -9.48288321e-01 3.65230292e-01
-1.01661466e-01 -7.14954659e-02 5.34688652e-01 3.29277873e-01
-1.08591497e+00 9.93420660e-01 -1.40174901e+00 -4.98712450e-01
1.12175286e-01 3.75748575e-01 -2.31321812e-01 -4.83277261e-01
-8.39143872e-01 9.66845155e-01 5.53039253e-01 -2.42400855e-01
-1.54941654e+00 -6.80766344e-01 -5.38442016e-01 -9.75016281e-02
7.58644402e-01 -8.97977710e-01 1.47761369e+00 -8.86426806e-01
-1.70801616e+00 1.84086472e-01 -1.34417161e-01 -6.84225917e-01
6.65888131e-01 -1.01241517e+00 -5.54637492e-01 -2.09954038e-01
-2.69972831e-01 3.34467500e-01 1.11661637e+00 -1.27692664e+00
-5.32095253e-01 -2.69611236e-02 -2.68527307e-02 3.89986873e-01
-1.45744175e-01 -5.38513482e-01 -6.72273338e-01 -3.63552809e-01
-2.36932144e-01 -1.33745706e+00 -1.91224322e-01 -2.72178501e-01
-8.13962445e-02 -2.22498067e-02 1.05148625e+00 -4.48133409e-01
7.69483328e-01 -2.23996925e+00 4.07961071e-01 -2.95913011e-01
-3.47782895e-02 2.67243564e-01 -1.59755334e-01 4.02101099e-01
1.91380799e-01 -3.43247682e-01 1.93281204e-01 -7.64381289e-01
-1.39117643e-01 7.14623809e-01 -5.33085406e-01 3.18636745e-01
-2.76477665e-01 6.13169491e-01 -1.23491216e+00 -4.28537875e-01
3.30171824e-01 3.55003744e-01 -5.36492884e-01 2.45130092e-01
-7.15972602e-01 5.47102153e-01 -2.10335851e-01 3.95548195e-01
3.22909236e-01 -2.56476998e-01 3.07194889e-01 3.10468405e-01
1.62276313e-01 -1.94113508e-01 -1.07090628e+00 1.85724223e+00
-5.87400615e-01 5.79188764e-01 -6.05952740e-01 -7.72697270e-01
4.73778665e-01 3.82318795e-01 5.73393703e-01 -3.73013973e-01
-4.27479029e-01 -1.20680735e-01 -3.28455150e-01 -7.15352714e-01
6.87492430e-01 -2.48447433e-01 7.62400478e-02 2.58304328e-01
3.37378711e-01 -1.70112461e-01 -1.30775914e-01 -3.90540101e-02
1.74460196e+00 7.09205508e-01 1.89763546e-01 2.04492331e-01
2.25706641e-02 -1.70627348e-02 7.77686357e-01 1.44194484e+00
-4.49031949e-01 5.36720753e-01 2.19258547e-01 -9.15163159e-01
-1.14211857e+00 -1.44307184e+00 4.15058076e-01 1.34001267e+00
5.09693801e-01 -3.76936913e-01 -3.23804200e-01 -4.29949045e-01
-2.12694369e-02 1.15530849e+00 -6.48830593e-01 -5.50865591e-01
-4.79555458e-01 -8.27819109e-01 2.04419866e-02 3.68016630e-01
2.82048613e-01 -1.30699134e+00 -8.29969525e-01 3.68659347e-01
-1.26535669e-01 -9.61007774e-01 -7.91310966e-02 3.99163187e-01
-8.25955510e-01 -8.90232325e-01 -2.23117307e-01 -4.72928137e-01
4.73816283e-02 4.09132957e-01 1.30407715e+00 -2.86890388e-01
-1.52120784e-01 1.08127213e+00 -3.99046034e-01 -4.47376192e-01
-4.23430741e-01 3.34603190e-02 7.45883882e-01 -1.52286544e-01
2.67911971e-01 -8.06619227e-01 -4.97276068e-01 1.72715425e-01
-8.44321907e-01 2.14899763e-01 6.89524949e-01 8.72951567e-01
9.30208027e-01 4.76889610e-01 3.47047567e-01 -8.72196734e-01
3.38429451e-01 -8.07898343e-01 -2.04964951e-01 6.71109140e-01
-3.78388017e-01 8.24780911e-02 8.63313079e-01 -1.02372706e+00
-1.42811072e+00 2.17899993e-01 3.40366304e-01 -9.46797371e-01
-2.76131630e-01 1.70478061e-01 -1.55562192e-01 1.85106337e-01
6.30860269e-01 6.02403104e-01 -3.53397250e-01 -3.43827099e-01
2.07382232e-01 1.40386000e-01 7.48923957e-01 -8.10741544e-01
6.35073245e-01 4.08570647e-01 -1.00347534e-01 -7.91823745e-01
-8.63937020e-01 -2.95951039e-01 -7.43143380e-01 -5.43724835e-01
4.04663742e-01 -1.05375004e+00 -5.13068140e-01 7.64679968e-01
-9.09678459e-01 -1.15966916e+00 -6.83216333e-01 5.87749362e-01
-1.35731018e+00 3.13498050e-01 -7.98208833e-01 -8.45988631e-01
1.39433280e-01 -9.06563342e-01 9.69010651e-01 1.06449127e-01
-9.52592790e-02 -9.13773656e-01 4.73706901e-01 -2.32815128e-02
3.02906305e-01 -4.14282978e-02 7.51582146e-01 -5.71446955e-01
-7.04170644e-01 -2.97823548e-01 4.59853828e-01 2.10465953e-01
-9.79226604e-02 -2.56629199e-01 -8.86675775e-01 -4.16436881e-01
1.96620777e-01 -5.28175414e-01 7.38176227e-01 2.52470255e-01
1.37077034e+00 -1.60690650e-01 -4.85482007e-01 3.09299499e-01
1.37769091e+00 2.54417211e-01 8.20578098e-01 3.35158229e-01
5.76702118e-01 2.00181171e-01 8.46333265e-01 7.81565607e-01
1.42986789e-01 5.12443841e-01 3.45406622e-01 6.51609302e-01
-5.62709868e-02 -5.37971377e-01 6.08904958e-01 8.43833327e-01
-2.26309687e-01 -4.42806184e-01 -8.71214330e-01 5.75804770e-01
-2.30535722e+00 -1.28663802e+00 2.68328577e-01 2.42543125e+00
6.56140387e-01 3.51787716e-01 -1.80626199e-01 -2.07995117e-01
3.83301437e-01 -6.71394691e-02 -1.02040184e+00 1.78891942e-01
-7.27436990e-02 -5.06136976e-02 4.87214595e-01 2.45016560e-01
-1.25958884e+00 9.30899024e-01 6.75314093e+00 7.66224027e-01
-8.66804361e-01 5.57322502e-01 7.46989548e-01 -6.56817377e-01
1.88272581e-01 -1.36919588e-01 -6.34699166e-01 4.85162675e-01
1.31676292e+00 -6.67737350e-02 4.77441579e-01 1.07906568e+00
9.03028399e-02 -3.94457042e-01 -1.33349228e+00 1.07060778e+00
1.45188779e-01 -1.02838433e+00 -1.69048026e-01 -3.94169152e-01
9.82213020e-01 3.75526458e-01 2.98833251e-01 8.52151155e-01
7.69748330e-01 -7.49046564e-01 7.38305867e-01 1.29853165e+00
3.42694104e-01 -4.59678173e-01 5.03551662e-01 7.96369195e-01
-7.49633729e-01 -3.86738390e-01 -6.08247876e-01 -2.13205829e-01
2.22913697e-01 3.57321531e-01 -6.14286184e-01 3.25571060e-01
9.00741577e-01 6.74678981e-01 -5.20706952e-01 1.47526944e+00
1.14076138e-01 9.04433072e-01 -2.89602131e-01 1.93579674e-01
-1.29845396e-01 2.51542330e-01 5.82629085e-01 8.71270120e-01
7.49480784e-01 -1.45565778e-01 4.16547716e-01 1.28514215e-01
1.07045613e-01 -2.41812319e-01 -8.54048431e-01 5.52675605e-01
4.34308171e-01 5.90600312e-01 -6.42507255e-01 -4.98604655e-01
-4.38946992e-01 1.42495775e+00 5.69195867e-01 5.16290843e-01
-1.10410607e+00 3.89548957e-01 5.72626114e-01 2.48046294e-01
3.14718485e-01 -5.10124803e-01 1.55985609e-01 -1.28869557e+00
3.09005170e-03 -5.79517126e-01 2.37130553e-01 -9.95285511e-01
-1.36911142e+00 3.04053456e-01 7.51111880e-02 -1.19848835e+00
-2.20716432e-01 -4.92101729e-01 -4.04412031e-01 2.70513892e-01
-1.31694722e+00 -9.63118553e-01 -3.51652831e-01 7.04075158e-01
1.13203108e+00 -1.89164147e-01 7.32248306e-01 -9.14810412e-03
-4.00114298e-01 1.56289235e-01 6.67726755e-01 -8.01537812e-01
1.08716965e+00 -1.21164262e+00 2.20573932e-01 6.72365606e-01
1.70398131e-01 3.63377780e-01 1.27994883e+00 -9.46232498e-01
-1.69226098e+00 -1.39918864e+00 4.04211313e-01 -8.54484081e-01
8.03454995e-01 -2.23050594e-01 -1.13627994e+00 1.13508868e+00
-1.00705512e-01 2.94184506e-01 5.92617214e-01 1.27061442e-01
-9.38718617e-02 -1.90430097e-02 -6.92971468e-01 6.91079676e-01
1.36929047e+00 -4.00370270e-01 -3.91614079e-01 7.66931474e-01
8.61293972e-01 -4.45069164e-01 -5.94484687e-01 2.75794446e-01
5.02930760e-01 -9.60757852e-01 9.82979000e-01 -7.83573091e-01
2.87746102e-01 -2.43161350e-01 -1.36161461e-01 -1.42900264e+00
-3.71615291e-01 -8.64151418e-01 -8.11435044e-01 7.50812352e-01
1.55978531e-01 -1.46032393e-01 8.18989635e-01 8.25784028e-01
-1.83814526e-01 -3.73021245e-01 -9.89710808e-01 -1.25970590e+00
-1.17080942e-01 -9.09863830e-01 3.32646012e-01 4.11053121e-01
1.10891469e-01 1.58360913e-01 -1.25732505e+00 2.69484401e-01
5.59204400e-01 -2.04646990e-01 9.37763512e-01 -1.11483169e+00
-9.01721537e-01 1.63303629e-01 -3.81075889e-01 -1.15139401e+00
3.49212646e-01 -8.93140882e-02 4.74619716e-01 -9.70009923e-01
4.67666268e-01 -5.29112041e-01 -5.07753253e-01 1.99344441e-01
-3.13434184e-01 -2.25787953e-01 1.17556088e-01 7.30885088e-01
-1.62317729e+00 8.34786236e-01 8.92948508e-01 -6.67005852e-02
-4.21118230e-01 2.08909810e-01 1.25102699e-02 7.80791819e-01
7.80600786e-01 -3.10171276e-01 -9.47850287e-01 -3.77585471e-01
2.67818749e-01 3.18131208e-01 4.74941552e-01 -1.84788942e+00
6.07845128e-01 -4.51589286e-01 4.74164099e-01 -3.46225321e-01
7.07149804e-01 -8.93489480e-01 7.36803651e-01 3.89876157e-01
-3.25559646e-01 1.23386206e-02 2.32204020e-01 1.42895710e+00
1.63377941e-01 -8.97289291e-02 3.90204012e-01 -2.95105428e-01
-1.20520854e+00 4.34110671e-01 -6.10193074e-01 -2.38108113e-01
1.37818861e+00 -1.54471934e-01 -2.17004493e-01 -5.87611079e-01
-1.34366345e+00 1.04192868e-01 7.50043750e-01 4.00177985e-01
5.90363324e-01 -9.77484107e-01 -2.65972584e-01 -2.22699389e-01
8.79651755e-02 -2.06529483e-01 7.15770006e-01 7.81658590e-01
-2.55257308e-01 9.49123055e-02 -2.42013693e-01 -4.14631307e-01
-8.73696625e-01 1.11552930e+00 2.15425536e-01 -4.43422377e-01
-8.67815375e-01 6.49836957e-01 7.46431947e-02 -1.67551056e-01
2.89488882e-01 7.34458640e-02 5.64924888e-02 -4.86819118e-01
5.72881341e-01 4.84500110e-01 -1.74586371e-01 -2.37902313e-01
2.18039587e-01 -8.36487338e-02 -2.52032518e-01 -4.75127250e-02
1.45705652e+00 -4.17614043e-01 3.70999873e-01 1.24597597e+00
4.63726521e-01 -4.32367265e-01 -2.20564747e+00 -4.49561894e-01
8.21054354e-02 -3.61196339e-01 -4.59018350e-01 -7.19972610e-01
-5.09029150e-01 5.30109882e-01 6.54565156e-01 -9.62386578e-02
8.97576928e-01 8.77492875e-03 6.59818649e-01 7.95240223e-01
1.02144074e+00 -1.45900083e+00 3.57812196e-01 6.85806334e-01
5.28502822e-01 -1.13373101e+00 -5.40848970e-02 -9.90994498e-02
-7.44776487e-01 8.89453053e-01 7.84069419e-01 -1.58884630e-01
7.44658113e-01 3.47635895e-01 -5.34848273e-01 1.67887568e-01
-1.38290417e+00 2.10795030e-01 -2.90112376e-01 8.49748790e-01
-6.31818920e-02 1.53118953e-01 4.64460313e-01 6.25953734e-01
2.58463882e-02 4.72815007e-01 6.69461668e-01 1.15225446e+00
-5.06239533e-01 -7.49092698e-01 -3.36970329e-01 6.09687090e-01
-4.62892167e-02 1.24333873e-01 1.17509626e-01 6.37303412e-01
2.57118165e-01 6.98436916e-01 2.87352055e-02 -7.17640281e-01
1.74273536e-01 4.92450625e-01 7.13196874e-01 -5.01192451e-01
7.14485720e-02 -2.44201437e-01 -1.09742895e-01 -7.21626163e-01
-3.40675384e-01 -1.20597661e+00 -7.17850387e-01 -5.56570232e-01
7.19960183e-02 -7.11117089e-02 1.31706238e-01 1.01018798e+00
4.40064698e-01 5.62324345e-01 -1.97639801e-02 -1.16131103e+00
-5.05634069e-01 -9.19849694e-01 -7.83976197e-01 4.43780631e-01
2.83401877e-01 -1.03723609e+00 -1.46522596e-01 5.94206989e-01] | [8.426531791687012, 0.5204747319221497] |
671f8389-72af-4055-b363-ec314753f7ad | alzheimers-disease-diagnosis-based-on | 1810.10941 | null | http://arxiv.org/abs/1810.10941v1 | http://arxiv.org/pdf/1810.10941v1.pdf | Alzheimer's Disease Diagnosis Based on Cognitive Methods in Virtual Environments and Emotions Analysis | Dementia is a syndrome characterised by the decline of different cognitive
abilities. Alzheimer's Disease (AD) is the most common dementia affecting
cognitive domains such as memory and learning, perceptual-motion or executive
function. High rate of deaths and high cost for detection, treatments and
patient's care count amongst its consequences. Early detection of AD is
considered of high importance for improving the quality of life of patients and
their families. The aim of this thesis is to introduce novel non-invasive early
diagnosis methods in order to speed the diagnosis, reduce the associated costs
and make them widely accessible. Novel AD's screening tests based on virtual
environments using new immersive technologies combined with advanced Human
Computer Interaction (HCI) systems are introduced. Four tests demonstrate the
wide range of screening mechanisms based on cognitive domain impairments that
can be designed using virtual environments. The use of emotion recognition to
analyse AD symptoms has been also proposed. A novel multimodal dataset was
specifically created to remark the autobiographical memory deficits of AD
patients. Data from this dataset is used to introduce novel descriptors for
Electroencephalogram (EEG) and facial images data. EEG features are based on
quaternions in order to keep the correlation information between sensors,
whereas, for facial expression recognition, a preprocessing method for motion
magnification and descriptors based on origami crease pattern algorithm are
proposed to enhance facial micro-expressions. These features have been proved
on classifiers such as SVM and Adaboost for the classification of reactions to
autobiographical stimuli such as long and short term memories. | ['Juan Manuel Fernández Montenegro'] | 2018-10-25 | null | null | null | null | ['motion-magnification'] | ['computer-vision'] | [-1.18751310e-01 -1.41163856e-01 3.61593783e-01 -3.78087491e-01
1.03980556e-01 -3.02506536e-01 5.33281922e-01 2.46049553e-01
-9.13998246e-01 1.02066803e+00 9.13048387e-02 1.85953081e-01
-4.31138784e-01 -7.05606461e-01 -1.07166126e-01 -5.91991007e-01
-5.74328423e-01 2.22853780e-01 7.97323138e-02 -5.37169695e-01
2.83539563e-01 7.42600799e-01 -2.07294655e+00 4.16911334e-01
5.82771301e-01 1.01719880e+00 3.68178010e-01 5.73510528e-01
7.62661472e-02 2.46180430e-01 -6.71520531e-01 -1.58217356e-01
-3.00348908e-01 -2.22353026e-01 -4.20090109e-01 -1.11169815e-01
-3.38530391e-01 -3.91130656e-01 3.92255485e-02 8.98543835e-01
9.65726078e-01 5.90451760e-03 8.35148573e-01 -1.11743283e+00
-4.92483586e-01 -3.11663181e-01 1.16158865e-01 3.50585729e-01
9.37856793e-01 -1.97664067e-01 1.06190868e-01 -8.80069911e-01
8.30300629e-01 1.26572287e+00 5.00003934e-01 7.32584178e-01
-1.00251150e+00 -5.10842562e-01 -2.71765649e-01 1.19894946e+00
-1.20064151e+00 -2.14421272e-01 6.67496502e-01 -5.60911477e-01
1.02782643e+00 5.39468825e-01 1.37517905e+00 1.36671913e+00
6.93509221e-01 3.85309644e-02 1.36457253e+00 -3.27412963e-01
6.29962802e-01 5.94396651e-01 4.07632113e-01 4.19137806e-01
3.22778553e-01 9.78986695e-02 -3.09991032e-01 -2.97226876e-01
5.03870606e-01 8.69964510e-02 -3.80569249e-01 -2.29394734e-01
-8.57781649e-01 7.50749290e-01 3.62576544e-01 7.06995904e-01
-5.48786044e-01 -3.03811312e-01 6.49949133e-01 4.87147570e-01
3.03509176e-01 8.54435563e-02 -3.86386633e-01 -3.77684891e-01
-3.72377574e-01 1.05324142e-01 7.69372344e-01 2.35176846e-01
2.58812845e-01 -1.56776726e-01 9.85737368e-02 5.62260270e-01
4.83240038e-01 8.03259075e-01 9.64460373e-01 -4.82767403e-01
-1.60967097e-01 9.78233755e-01 -1.59340620e-01 -1.02239573e+00
-7.80739546e-01 -1.55459121e-01 -8.66504490e-01 8.61624777e-01
-2.39909098e-01 4.50661220e-02 -5.89229167e-01 1.38215363e+00
2.22271815e-01 -2.19314456e-01 1.70462370e-01 1.00950491e+00
9.53230619e-01 4.77729738e-01 2.63811022e-01 -5.96017182e-01
1.71195066e+00 8.81848782e-02 -1.12624705e+00 5.23650758e-02
5.87447643e-01 -2.44796798e-01 8.91734660e-01 6.05112851e-01
-7.25729764e-01 -3.84861380e-01 -1.29159391e+00 2.84626722e-01
-8.61980915e-01 1.05199426e-01 3.92085373e-01 1.11178267e+00
-1.08111191e+00 7.13957310e-01 -8.17068517e-01 -8.85679066e-01
3.08174074e-01 6.32242382e-01 -6.86902165e-01 2.62651294e-01
-1.20281482e+00 1.44876444e+00 1.48191482e-01 -2.04206966e-02
2.68146559e-03 -2.93951243e-01 -4.90832120e-01 -2.17672154e-01
-5.16283333e-01 -7.96029866e-01 4.25620317e-01 -1.35559249e+00
-1.47445560e+00 1.06920159e+00 2.20499411e-02 -2.76499748e-01
6.04505956e-01 -1.41878441e-01 -8.29525352e-01 4.66007173e-01
-1.50074452e-01 4.59402084e-01 5.95565915e-01 -6.23313427e-01
-1.45213604e-01 -8.36527526e-01 -5.70250869e-01 7.87406042e-02
-5.98274112e-01 2.00547367e-01 5.81004083e-01 -4.16607916e-01
-1.22331018e-02 -7.61124790e-01 1.23043224e-01 2.47866958e-01
1.18281640e-01 -1.10009536e-02 1.05214787e+00 -1.09962463e+00
9.19270337e-01 -2.00442266e+00 3.37127537e-01 2.79357314e-01
1.55699581e-01 3.35934222e-01 3.66394252e-01 1.63347647e-01
-3.26687366e-01 -1.34594068e-01 -1.35879919e-01 5.90435751e-02
7.98795149e-02 1.35942891e-01 2.16812029e-01 5.38238108e-01
1.47321850e-01 5.38207889e-01 -4.15415615e-01 -3.14547747e-01
4.76627558e-01 9.20337856e-01 -1.57542914e-01 4.25584354e-02
3.31638604e-01 3.89965951e-01 -3.05315107e-01 4.48054075e-01
6.58384442e-01 4.78374869e-01 -1.85405966e-02 -2.23298267e-01
-1.38335422e-01 -2.57991999e-01 -9.47111011e-01 1.01423931e+00
-3.21790427e-02 6.63007081e-01 -1.20880254e-01 -1.17880368e+00
1.14033711e+00 6.49854064e-01 2.45604143e-01 -1.31284523e+00
4.72094893e-01 2.69905239e-01 -1.50356174e-01 -1.01133227e+00
-1.84319720e-01 -2.04422161e-01 3.30139875e-01 -3.85840349e-02
-2.27129355e-01 4.29412454e-01 6.66587204e-02 -3.77143800e-01
9.51365650e-01 -1.90954264e-02 5.65771103e-01 -3.57718974e-01
9.01108563e-01 -2.62952954e-01 1.56368479e-01 -7.34845623e-02
-3.12068880e-01 7.63473213e-02 3.93674582e-01 -3.60225677e-01
-9.36184824e-01 -9.62602735e-01 -4.24049973e-01 4.97961283e-01
-2.40250394e-01 6.04138561e-02 -6.68645978e-01 4.21301201e-02
-1.32650971e-01 6.97108328e-01 -5.12832046e-01 -5.01199961e-01
-1.92628786e-01 -9.11422133e-01 2.51368910e-01 2.68927664e-01
7.66216457e-01 -1.03761578e+00 -1.22301710e+00 1.21235467e-01
2.67529279e-01 -6.43579662e-01 7.95416594e-01 8.50147232e-02
-1.28508794e+00 -1.03599584e+00 -9.87453878e-01 -8.53803515e-01
3.54652405e-01 -3.70689601e-01 6.65591598e-01 -1.89094767e-01
-7.74243116e-01 6.91032410e-01 -5.45231223e-01 -4.56171632e-01
-3.19683015e-01 -4.89376158e-01 3.01595390e-01 -1.39974803e-01
4.94192064e-01 -9.70469713e-01 -7.36698568e-01 8.87385756e-02
-5.42014599e-01 -4.22349483e-01 6.61676824e-01 3.93894106e-01
2.96169132e-01 -4.26214308e-01 7.53246129e-01 -8.46950486e-02
9.51744497e-01 -4.19063240e-01 -1.51402634e-02 1.72642544e-01
-5.00046790e-01 -2.07659602e-01 1.85363963e-01 -5.50910413e-01
-8.61684799e-01 -2.42712066e-01 -2.24805474e-01 1.83892362e-02
-5.38669765e-01 1.42841086e-01 -4.01007026e-01 -2.48084173e-01
6.74676418e-01 2.81024873e-01 2.23015636e-01 -3.63943487e-01
-2.46127322e-01 8.74455094e-01 3.10352445e-01 1.37446914e-02
1.04635656e-01 3.00616831e-01 2.81978816e-01 -1.59905720e+00
2.31876493e-01 -5.92884086e-02 -5.68593204e-01 -6.24966681e-01
1.26277578e+00 -6.67449057e-01 -9.30457056e-01 4.14426863e-01
-1.17162740e+00 2.25362629e-01 1.42745495e-01 7.78455436e-01
-4.04248327e-01 2.86736488e-01 -2.98957825e-01 -1.00172901e+00
-6.40601814e-01 -9.09862101e-01 5.34100056e-01 1.80599988e-01
-5.67034900e-01 -9.31476712e-01 3.21390957e-01 7.31882825e-02
2.51563162e-01 5.60329854e-01 9.85238910e-01 -4.85219657e-01
-1.43905044e-01 -3.38046879e-01 -9.05803069e-02 4.01581913e-01
-3.09770610e-02 -2.49759957e-01 -8.22593451e-01 -7.30385259e-02
3.77232432e-01 -1.09121181e-01 4.52589482e-01 3.07965040e-01
2.54007101e-01 5.08204708e-03 -3.57033014e-01 2.62677759e-01
1.26692200e+00 9.08652842e-01 1.28684556e+00 8.82022858e-01
7.62805864e-02 6.45457387e-01 5.40256619e-01 5.24030387e-01
-2.91964877e-02 6.89195812e-01 4.97000635e-01 2.15131089e-01
2.88270861e-01 8.02285135e-01 6.59316123e-01 4.73710656e-01
-6.29716575e-01 1.81382984e-01 -6.05011761e-01 4.21712510e-02
-1.42560756e+00 -1.13464665e+00 -6.84999883e-01 2.33634973e+00
1.99462250e-01 6.66783191e-03 1.37741879e-01 7.19874322e-01
7.39584148e-01 -5.36164045e-01 -2.57321656e-01 -6.73336506e-01
-4.44360942e-01 6.55363977e-01 1.45860268e-02 1.73622295e-01
-8.72399092e-01 9.44391713e-02 4.93014288e+00 6.56168908e-02
-1.39499640e+00 2.16398865e-01 1.03954315e-01 -9.68267843e-02
1.60263509e-01 -5.45326233e-01 -3.46666247e-01 6.89119220e-01
1.10213888e+00 -1.92298423e-02 2.02021003e-01 6.65814638e-01
4.33020145e-01 -4.30295706e-01 -8.82020056e-01 1.45602822e+00
2.83933401e-01 -6.93217099e-01 -3.46121825e-02 1.88550614e-02
-7.42925182e-02 -4.53406125e-01 7.73255667e-03 5.87635487e-02
-1.06833303e+00 -9.27121401e-01 6.48452044e-01 1.24157834e+00
5.91960788e-01 -9.95194495e-01 1.00954580e+00 -1.42041355e-01
-8.61871600e-01 -2.33128905e-01 -2.72233456e-01 -4.42847997e-01
1.50620118e-01 2.95438468e-01 -6.61096871e-01 2.48399928e-01
7.02559114e-01 4.54251647e-01 -7.55089939e-01 1.12843394e+00
3.49811092e-02 1.44957080e-01 -2.49968886e-01 -7.06333637e-01
-2.80571073e-01 -3.43740225e-01 7.77893245e-01 1.02913570e+00
7.48270512e-01 2.29625896e-01 -6.72712624e-01 8.27752292e-01
8.26724887e-01 3.62076521e-01 -5.67207932e-01 -1.65110692e-01
1.06638342e-01 1.09666264e+00 -8.86820734e-01 -1.10202968e-01
-1.75207794e-01 1.47454691e+00 -4.22429651e-01 -1.92528572e-02
-6.50079250e-01 -4.28485930e-01 6.07717693e-01 2.39984661e-01
-1.29320920e-01 -1.98272973e-01 -2.03886062e-01 -8.64939868e-01
1.16550758e-01 -4.59086329e-01 2.17932403e-01 -9.89504457e-01
-7.96755314e-01 5.62287509e-01 -6.42352328e-02 -1.06531668e+00
-1.89097717e-01 -1.03085291e+00 -5.33674955e-01 8.25244725e-01
-6.42291367e-01 -9.02306318e-01 -5.29824913e-01 8.41901004e-01
6.94046393e-02 -3.43584955e-01 1.45484376e+00 4.15880144e-01
-4.49673057e-01 -8.19162726e-02 1.38584107e-01 -3.49695116e-01
7.20711172e-01 -1.02267969e+00 -5.20281971e-01 1.17639564e-01
-6.89992189e-01 1.15280703e-01 7.97110736e-01 -7.75457740e-01
-1.10724783e+00 -5.07596731e-01 8.95873427e-01 -7.77737722e-02
2.83158958e-01 -2.69508958e-01 -7.37165987e-01 2.28410244e-01
5.89162037e-02 -4.89309520e-01 6.63493276e-01 -4.47244048e-01
3.30763102e-01 -1.71310097e-01 -1.49097204e+00 5.48381150e-01
7.46685207e-01 -5.26060283e-01 -9.71025944e-01 1.66316539e-01
-8.54535680e-03 3.18545133e-01 -8.93441439e-01 1.19015470e-01
8.14249039e-01 -1.48205185e+00 1.08098745e+00 -3.83006662e-01
1.19246081e-01 1.61481738e-01 -1.58460125e-01 -1.18857527e+00
-1.93772465e-01 -1.38926255e-02 1.92401558e-01 1.07014441e+00
-1.25797078e-01 -1.05921936e+00 4.89928335e-01 4.96478975e-01
1.23971462e-01 -5.53894460e-01 -1.22852302e+00 -6.82658970e-01
-3.45448911e-01 -1.96352363e-01 -6.25951122e-03 7.88889945e-01
4.13872808e-01 3.54969740e-01 3.38412851e-01 3.55887897e-02
4.76337045e-01 -6.47489727e-01 1.24234073e-01 -1.87865222e+00
2.11033449e-01 -3.04961801e-01 -1.55886924e+00 4.99939412e-01
-4.67637554e-02 -7.70329297e-01 -9.56698656e-01 -1.53084517e+00
6.99173808e-02 2.90114760e-01 -6.18271716e-02 -2.37334520e-02
5.11622310e-01 2.51534373e-01 -9.35009494e-02 -2.51852244e-01
4.10534665e-02 6.71841800e-01 8.95476103e-01 1.00099459e-01
-4.66711909e-01 -2.21460521e-01 -8.33196416e-02 6.79332256e-01
7.96563804e-01 -9.58777443e-02 -4.56681699e-01 3.86222720e-01
4.62026834e-01 1.15927808e-01 9.59539890e-01 -1.46168554e+00
-6.54610544e-02 4.85083342e-01 1.02335060e+00 -4.17735606e-01
8.31453145e-01 -9.86339569e-01 4.65135485e-01 9.84743297e-01
3.49028081e-01 1.80549517e-01 2.33914047e-01 2.96662837e-01
-1.28808301e-02 -1.82105124e-01 8.98269832e-01 -2.10839864e-02
-8.34325135e-01 -3.25626850e-01 -1.05260849e+00 -5.12581766e-01
1.56085610e+00 -8.17299366e-01 -8.21937546e-02 -3.21547478e-01
-1.38013685e+00 -2.86892235e-01 2.50904083e-01 1.31792665e-01
8.48234653e-01 -1.31331968e+00 -4.12707895e-01 2.12239936e-01
-7.79597610e-02 -8.23770344e-01 5.71989179e-01 1.21448934e+00
-8.58974874e-01 4.33407217e-01 -1.21477985e+00 -3.51855457e-01
-2.04148412e+00 4.82016861e-01 3.76551181e-01 2.22403124e-01
-6.56945467e-01 5.14522552e-01 -2.05195561e-01 3.18975091e-01
9.35765281e-02 -2.94806629e-01 -9.42122817e-01 7.61044681e-01
6.74748600e-01 9.52335358e-01 3.44272614e-01 -6.98092937e-01
-7.77897656e-01 6.61355674e-01 2.64493138e-01 -3.91277850e-01
1.63976324e+00 -1.83718875e-01 -3.69221687e-01 6.23618662e-01
1.13759041e+00 -4.09841329e-01 -4.55680549e-01 8.07543337e-01
6.15025386e-02 1.05577871e-01 1.14906728e-01 -1.09838212e+00
-5.36601841e-01 1.12033045e+00 1.81156194e+00 2.01561645e-01
1.31860113e+00 -3.65560919e-01 3.77571106e-01 6.84642911e-01
3.43997210e-01 -1.25121629e+00 7.66729191e-03 8.05717483e-02
1.36386347e+00 -7.85561383e-01 3.05685773e-02 -1.58671677e-01
-5.36602497e-01 1.77167094e+00 6.38474450e-02 -2.02432722e-01
7.35768080e-01 1.93600535e-01 -1.10254236e-01 -1.32536709e-01
-3.21489453e-01 -2.28729948e-01 1.51514724e-01 9.63585913e-01
3.51276487e-01 2.83835202e-01 -8.97682667e-01 1.24718940e+00
-1.81922078e-01 2.66701251e-01 3.24164718e-01 1.10777831e+00
-7.64395654e-01 -8.36311817e-01 -9.37341809e-01 4.30776805e-01
-1.91036388e-01 4.49794680e-01 -7.62831211e-01 9.18226540e-01
5.67119420e-01 6.40523434e-01 2.26395458e-01 -4.09305304e-01
5.94988227e-01 5.33125818e-01 8.68014097e-01 1.26793459e-01
-3.23073655e-01 -2.44443312e-01 3.77895296e-01 -5.94120204e-01
-5.67341447e-01 -8.20510507e-01 -1.30004787e+00 -2.04390243e-01
3.18098903e-01 -2.22590640e-01 1.08499563e+00 1.07723105e+00
2.66606152e-01 6.31567240e-01 1.39082387e-01 -1.15256429e+00
-9.83514115e-02 -9.91189659e-01 -9.21265364e-01 2.29294881e-01
1.00184023e-01 -1.03333914e+00 -4.09306437e-01 -1.30157880e-02] | [13.333396911621094, 3.2887208461761475] |
44bc46d0-ac07-415c-a076-e266bdb7182c | st-sql-semi-supervised-self-training-for-text | null | null | https://openreview.net/forum?id=CPdvNQGrOr0 | https://openreview.net/pdf?id=CPdvNQGrOr0 | ST-SQL: Semi-Supervised Self-Training for Text-to-SQL via Column Specificity Meta-Learning | The few-shot problem is an urgent challenge for the generalization capability of the single-table text-to-SQL task. Current few-shot methods neglect the potential information of unlabeled data and have a domain bias due to the same weight of samples. Motivated by this, this paper proposes a Self-Training text-to-SQL (ST-SQL) method which handles the problem from both views of data and algorithms. At the data level, ST-SQL performs data expansion by using an iterative framework to attach pseudo-labels to unlabeled data. The expanded data are sampled to reversely train the model. At the algorithm level, ST-SQL defines a column specificity to perform a more fine-grained gradient update during meta-training. The common samples are attached more weight to eliminate the domain bias. ST-SQL achieves state-of-the-art results on both open-domain and domain-specific benchmarks and brings more significant improvements on few-shot tests. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['text-to-sql'] | ['computer-code'] | [ 8.70153755e-02 -1.70162350e-01 -6.98004067e-01 -7.48977482e-01
-9.15103674e-01 -2.71794558e-01 4.69055027e-01 2.85135478e-01
-3.33722174e-01 7.57020593e-01 1.50136098e-01 -1.66527539e-01
2.22746328e-01 -1.01153100e+00 -6.75262630e-01 -3.75982583e-01
3.77469480e-01 8.90202582e-01 5.67405999e-01 -3.81816953e-01
2.07323223e-01 3.80548239e-02 -1.77400732e+00 6.13950253e-01
1.21252656e+00 1.06806850e+00 1.25725731e-01 3.25994462e-01
-6.05727851e-01 9.42044854e-01 -2.99234658e-01 -5.31052291e-01
3.58153045e-01 -3.99507314e-01 -5.34426153e-01 1.26255766e-01
4.33309466e-01 -3.77058834e-01 -1.44305632e-01 9.76621032e-01
6.82460368e-01 5.65576911e-01 5.40935636e-01 -1.16638851e+00
-6.70280039e-01 6.07058883e-01 -5.21538734e-01 1.50602952e-01
1.51173264e-01 2.51452535e-01 1.10994959e+00 -1.51116884e+00
9.71048236e-01 9.95318949e-01 4.14035380e-01 7.18244433e-01
-1.15590584e+00 -4.04370755e-01 2.64543176e-01 3.51839572e-01
-1.15036023e+00 -3.14855695e-01 5.86805761e-01 -2.25199655e-01
1.12667549e+00 1.62277326e-01 4.20304120e-01 1.11275065e+00
-1.04967318e-01 1.07004535e+00 6.26771331e-01 -4.11269665e-01
8.86473179e-01 4.75477874e-01 5.56390285e-01 5.69657445e-01
4.09605235e-01 -2.53625870e-01 -7.73662031e-01 -6.41977191e-02
1.87474310e-01 5.57281137e-01 2.01087520e-01 -8.86548817e-01
-6.34840906e-01 1.14267421e+00 1.87418312e-01 7.16702417e-02
-1.59473121e-01 -3.91067356e-01 8.27608049e-01 3.96928579e-01
6.01021111e-01 6.18658543e-01 -6.25651479e-01 -2.54211783e-01
-9.44160461e-01 2.33574599e-01 9.01104033e-01 1.33960021e+00
1.01452529e+00 4.15156409e-02 -5.80383420e-01 1.15892136e+00
-1.13969937e-01 2.80208498e-01 7.85619915e-01 -4.21030432e-01
9.03258801e-01 9.90535557e-01 -1.60682157e-01 -2.32319459e-01
1.55046592e-02 -4.57920164e-01 -5.57982326e-01 4.97858366e-03
-4.28707786e-02 -1.78995267e-01 -1.33931935e+00 1.39020145e+00
4.66283530e-01 1.02661408e-01 1.39214009e-01 6.35868907e-01
8.90970051e-01 6.30059540e-01 1.25171505e-02 -3.07945997e-01
1.13368356e+00 -1.37017941e+00 -6.69403911e-01 -5.85205376e-01
9.40948904e-01 -1.75403342e-01 1.63485503e+00 2.19493821e-01
-9.61398065e-01 -5.40099263e-01 -1.19637918e+00 -1.89527780e-01
-9.34663653e-01 -1.84908926e-01 2.55597144e-01 6.46768630e-01
-3.33889395e-01 6.75936222e-01 -7.14517534e-01 -5.70557714e-01
5.01984596e-01 -1.09613882e-02 -1.13838553e-01 -6.46799564e-01
-1.30977774e+00 7.26380169e-01 5.93092263e-01 -6.12794280e-01
-7.55295336e-01 -1.09148264e+00 -1.06632352e+00 2.56047398e-01
9.27985847e-01 -4.30809796e-01 1.31483519e+00 -3.17357928e-01
-1.35397220e+00 8.04906726e-01 -2.62332231e-01 -5.25911093e-01
5.00771284e-01 -4.05468076e-01 -5.48930109e-01 -1.31281167e-01
2.90991366e-01 3.62503946e-01 7.19079018e-01 -9.49415267e-01
-6.81809247e-01 -5.03051937e-01 -4.23481107e-01 3.29932064e-01
-7.38051295e-01 -3.88011515e-01 -7.89923549e-01 -6.01895094e-01
-2.25120373e-02 -5.91780186e-01 -2.30575189e-01 -2.94008046e-01
-3.21413279e-01 -1.61883935e-01 9.21020687e-01 -5.14314324e-02
1.68867838e+00 -2.13853168e+00 -5.36648594e-02 -1.21436298e-01
1.91289455e-01 4.55749512e-01 -2.29031876e-01 5.76621771e-01
-1.32780680e-02 -5.89501997e-03 -2.93207377e-01 -2.57918775e-01
-3.33652576e-03 2.67847151e-01 -4.13837463e-01 -1.43553214e-02
4.30509955e-01 1.03387403e+00 -9.73454714e-01 -6.42513394e-01
1.77258670e-01 -2.46981025e-01 -7.97244251e-01 2.29334325e-01
-6.45865977e-01 -1.59511045e-01 -5.35577118e-01 7.76105523e-01
7.20515132e-01 -4.80509996e-01 8.95389542e-02 1.32264674e-01
7.84395263e-03 7.08208829e-02 -1.23700857e+00 1.88340604e+00
-2.86535561e-01 7.41586685e-02 -5.67355514e-01 -9.73882198e-01
1.05544925e+00 -1.09561592e-01 3.03252965e-01 -9.14541602e-01
-6.57833219e-02 3.70552450e-01 -2.92171896e-01 -4.95266885e-01
6.25817835e-01 -1.71053514e-01 -2.25731090e-01 3.39582711e-01
4.10006315e-01 -1.13587052e-01 5.63737392e-01 4.76378739e-01
1.19948518e+00 1.85887933e-01 4.69834507e-01 -1.08362682e-01
3.96602124e-01 2.98731416e-01 6.82037055e-01 9.48317230e-01
-1.33517951e-01 6.06300950e-01 2.69356996e-01 -5.87074995e-01
-1.09645557e+00 -1.12944853e+00 -1.44613430e-01 1.75839543e+00
1.36699736e-01 -5.06716549e-01 -4.13260251e-01 -1.01364636e+00
2.94884145e-01 1.25648952e+00 -7.44248927e-01 -6.67470753e-01
-1.79499418e-01 -6.58689916e-01 2.78278422e-02 7.48327613e-01
3.89558941e-01 -8.85136068e-01 -6.02234185e-01 3.81034136e-01
7.65858293e-02 -8.78435850e-01 -3.80019993e-01 8.26342106e-01
-9.88326728e-01 -7.35382259e-01 -6.87901318e-01 -8.00596833e-01
4.97111887e-01 1.35850236e-01 1.26242077e+00 -4.40684110e-01
-4.73795980e-01 -8.35981369e-02 -7.04478443e-01 -6.13688111e-01
-1.83111936e-01 3.35219562e-01 -2.42292166e-01 -7.42784142e-02
9.30192530e-01 -3.18338633e-01 -3.11352253e-01 3.32310826e-01
-7.16904104e-01 -2.99581498e-01 4.51455563e-01 1.19554007e+00
7.90787995e-01 -2.53892541e-01 7.45669544e-01 -1.34400880e+00
7.14983404e-01 -7.22197711e-01 -2.93562263e-01 6.34938955e-01
-9.33356285e-01 3.05029511e-01 8.71630907e-01 -6.35881782e-01
-1.20747113e+00 4.80023026e-02 2.91221619e-01 -7.10116029e-01
6.50318190e-02 4.87793475e-01 -2.01333955e-01 3.88306051e-01
1.06455064e+00 4.56138611e-01 -1.84368446e-01 -5.69534779e-01
6.02865338e-01 6.27740860e-01 3.06552023e-01 -5.31695724e-01
5.54477274e-01 3.34275693e-01 -2.30789110e-01 -5.88063300e-01
-1.20831037e+00 -8.00010979e-01 -4.37316984e-01 6.63659051e-02
6.35183692e-01 -8.83502007e-01 -3.22187915e-02 1.27982572e-01
-4.66444761e-01 -3.19020748e-01 -1.00076771e+00 3.22178602e-01
-5.46523094e-01 1.00884087e-01 -5.38170397e-01 -7.21645772e-01
-4.82781887e-01 -8.42529297e-01 9.78755236e-01 2.98585743e-01
6.57156035e-02 -7.20125020e-01 2.75696695e-01 1.13274287e-02
2.75121391e-01 -2.25934625e-01 1.08030534e+00 -1.42840493e+00
-1.77105337e-01 -5.29193044e-01 -3.41023952e-02 3.02289039e-01
-2.85550989e-02 -2.32817709e-01 -9.27411556e-01 -2.96287268e-01
2.34754741e-01 -1.01524830e+00 1.08549905e+00 1.85980290e-01
1.15481532e+00 1.03470512e-01 -3.15680712e-01 6.45238936e-01
1.59897852e+00 2.92966485e-01 5.54471314e-01 1.80076018e-01
5.24420679e-01 3.45718622e-01 1.17169607e+00 8.74928236e-01
1.10627815e-01 2.82053411e-01 1.15324214e-01 1.71164393e-01
1.57139450e-01 -4.95186061e-01 2.47949138e-02 6.41531765e-01
5.31842887e-01 -2.85730481e-01 -1.01958930e+00 4.50098395e-01
-1.99427176e+00 -9.98609066e-01 1.79366365e-01 2.40460777e+00
9.07359779e-01 4.57393050e-01 8.11325759e-02 -9.82538052e-03
8.39951098e-01 2.42971778e-01 -1.15548980e+00 -3.07815760e-01
-9.03796405e-03 1.25510588e-01 3.15943480e-01 -3.48194465e-02
-1.08817291e+00 1.23265362e+00 6.17751026e+00 1.20836234e+00
-9.30591345e-01 2.82288883e-02 7.36222744e-01 -4.61816698e-01
-1.67274147e-01 -3.86460312e-02 -1.34178722e+00 4.05520648e-01
9.04817939e-01 -4.53625262e-01 4.01530832e-01 1.44001758e+00
-3.24181139e-01 -3.28165703e-02 -1.30869246e+00 8.05549800e-01
2.47913107e-01 -1.45960915e+00 2.16239691e-01 -2.86929429e-01
1.00914812e+00 2.16552302e-01 -9.80651453e-02 1.12169206e+00
4.07840848e-01 -5.81482649e-01 3.96722943e-01 3.39553744e-01
1.08174253e+00 -6.79783046e-01 5.69615364e-01 6.17989421e-01
-1.26920891e+00 -3.65923285e-01 -7.99208999e-01 1.06663451e-01
-4.69870605e-02 5.25017381e-01 -8.98926735e-01 5.73384345e-01
5.77840567e-01 8.61856043e-01 -7.72360981e-01 8.07999134e-01
1.53892756e-01 3.94907415e-01 -7.30537176e-02 -3.88776422e-01
2.06784248e-01 -1.33685678e-01 2.93323398e-01 1.00408995e+00
2.02868775e-01 -8.10163468e-02 4.24958915e-01 8.52714241e-01
-2.76732504e-01 3.52615654e-01 -7.24007726e-01 -3.01352054e-01
7.41352320e-01 1.07117152e+00 -5.26021481e-01 -9.58805323e-01
-7.09130108e-01 8.88687432e-01 5.99573791e-01 3.17563176e-01
-6.26234114e-01 -6.76144958e-01 2.32569262e-01 2.26478457e-01
6.87251031e-01 3.85266393e-01 -5.03745735e-01 -1.40832150e+00
1.10506795e-01 -9.04451966e-01 7.04875708e-01 -4.87026572e-01
-1.56606376e+00 3.54955286e-01 -1.31444752e-01 -1.64628196e+00
-3.83091897e-01 -4.07283574e-01 -7.12890923e-01 7.01622009e-01
-1.30726779e+00 -6.13574326e-01 -3.91981095e-01 6.19213045e-01
1.10483456e+00 -6.40167952e-01 8.50906074e-01 2.41199359e-01
-7.16300011e-01 7.33325958e-01 5.43916821e-01 1.03475358e-02
9.75746095e-01 -1.29588544e+00 7.11614907e-01 8.07074904e-01
-9.15166438e-02 7.48870671e-01 5.26355386e-01 -1.01143503e+00
-1.56355178e+00 -1.25595891e+00 6.65357411e-01 -3.82333338e-01
6.58074498e-01 -5.40243626e-01 -1.07232499e+00 5.37357390e-01
-2.16115206e-01 4.05628502e-01 9.38152552e-01 2.27658913e-01
-6.07073963e-01 -2.71138161e-01 -1.30273473e+00 4.82232511e-01
9.43655074e-01 -4.72970665e-01 -9.06450748e-01 3.01690191e-01
1.10022414e+00 -1.25086725e-01 -6.82771027e-01 3.84733319e-01
3.04623634e-01 -8.99392486e-01 6.81748867e-01 -1.06050873e+00
6.78686798e-01 1.81879312e-01 -3.75360608e-01 -1.25401914e+00
-2.33392254e-01 -3.06034327e-01 -4.20189440e-01 1.13268387e+00
4.78642493e-01 -4.94336873e-01 1.17188895e+00 6.65240824e-01
-2.45362341e-01 -1.01513219e+00 -6.93587661e-01 -1.10366905e+00
7.26464242e-02 -2.08992824e-01 4.56258774e-01 9.51700509e-01
3.58724862e-01 7.16572046e-01 -4.59390491e-01 -6.06754005e-01
5.65008759e-01 1.68344051e-01 8.24567974e-01 -1.36458123e+00
-3.06722671e-01 -1.43310994e-01 -1.02204651e-01 -1.04712915e+00
-1.83160156e-01 -9.66850460e-01 7.72827864e-02 -1.44578028e+00
4.20971930e-01 3.87168564e-02 -3.21346641e-01 2.54018307e-01
-5.01729131e-01 -2.41209716e-01 1.87502056e-01 2.65296623e-02
-1.11283600e+00 7.46557534e-01 1.10894740e+00 -7.40239471e-02
-4.74992126e-01 -5.30183800e-02 -5.60807347e-01 3.46950501e-01
6.09213710e-01 -5.60676277e-01 -7.95448601e-01 -9.53725129e-02
-1.78584587e-02 7.57295638e-02 -3.40109915e-01 -1.06440878e+00
3.12943995e-01 -2.55615830e-01 3.54476571e-01 -9.30028617e-01
3.00078064e-01 -6.81306958e-01 -5.23168802e-01 6.54996037e-01
-5.55459082e-01 -1.67496249e-01 2.11460590e-02 8.55583668e-01
-2.62957364e-01 -5.76168001e-01 9.29873109e-01 -3.01417053e-01
-1.01637423e+00 4.57875967e-01 -4.86673415e-02 7.80482054e-01
1.10931230e+00 -3.58264118e-01 -4.16850179e-01 -9.19471532e-02
-6.56334221e-01 6.26037657e-01 3.69180143e-01 3.90093058e-01
6.06956959e-01 -1.37814903e+00 -4.37529653e-01 5.49936831e-01
7.84810364e-01 6.32344484e-02 3.69899660e-01 4.36048388e-01
-1.67948276e-01 3.36403668e-01 -1.74123943e-01 -5.11352420e-01
-8.31556022e-01 1.08301520e+00 -1.60952285e-02 -5.94367027e-01
-6.04571342e-01 9.35256362e-01 -6.72440156e-02 -5.83537638e-01
5.65418065e-01 -1.34583488e-01 -1.48292705e-02 3.95906538e-01
6.96466863e-01 6.23888612e-01 2.10949183e-01 2.38205478e-01
-2.61247247e-01 1.78025961e-01 -6.26800716e-01 -1.45671070e-01
1.46461689e+00 1.37251899e-01 4.69088107e-01 8.82030725e-01
1.25549018e+00 -2.42697120e-01 -1.19147658e+00 -6.44745052e-01
2.66757160e-01 -4.92239296e-01 -3.17660183e-01 -8.98978710e-01
-6.73509777e-01 1.07022095e+00 3.74859840e-01 1.02562673e-01
8.48820925e-01 -1.55950263e-01 6.34809732e-01 8.63853276e-01
2.79950678e-01 -1.64876103e+00 5.54446876e-01 8.25810492e-01
4.56656396e-01 -1.56141829e+00 -4.48948219e-02 -2.88175315e-01
-1.13611555e+00 8.79455566e-01 1.03237212e+00 -9.37662125e-02
4.96391445e-01 2.14548662e-01 -3.01799625e-01 -6.34203404e-02
-1.07858229e+00 -2.22742662e-01 1.63379267e-01 4.89387751e-01
2.70928562e-01 -2.85988212e-01 -3.26127470e-01 7.91357696e-01
3.13555270e-01 2.76464462e-01 1.94238231e-01 1.38369584e+00
-7.94144928e-01 -9.61036146e-01 7.42358491e-02 9.18876171e-01
6.84822127e-02 -3.86402667e-01 -3.17111433e-01 6.75919354e-01
-1.60598963e-01 6.65109277e-01 4.64932658e-02 -5.23956299e-01
4.96432394e-01 8.17360520e-01 3.96285318e-02 -1.21542978e+00
-4.65114892e-01 8.58393312e-02 -2.61881351e-02 -4.72665906e-01
3.79223406e-01 -4.48668420e-01 -1.30429041e+00 -2.28921827e-02
-2.92139471e-01 2.78634399e-01 3.92328382e-01 6.80034637e-01
6.25013292e-01 5.93930304e-01 7.29540765e-01 -4.19982791e-01
-1.28295207e+00 -1.05640781e+00 -7.63890088e-01 6.93688512e-01
-1.26783445e-01 -5.58726311e-01 -3.06393802e-01 -3.46240878e-01] | [10.081921577453613, 3.5285208225250244] |
49519d5e-d530-4c47-b39f-356f9422f921 | handling-rare-items-in-data-to-text | null | null | https://aclanthology.org/W18-6543 | https://aclanthology.org/W18-6543.pdf | Handling Rare Items in Data-to-Text Generation | Neural approaches to data-to-text generation generally handle rare input items using either delexicalisation or a copy mechanism. We investigate the relative impact of these two methods on two datasets (E2E and WebNLG) and using two evaluation settings. We show (i) that rare items strongly impact performance; (ii) that combining delexicalisation and copying yields the strongest improvement; (iii) that copying underperforms for rare and unseen items and (iv) that the impact of these two mechanisms greatly varies depending on how the dataset is constructed and on how it is split into train, dev and test. | ['Claire Gardent', 'Anastasia Shimorina'] | 2018-11-01 | null | null | null | ws-2018-11 | ['kg-to-text'] | ['natural-language-processing'] | [ 3.03396881e-01 -3.23507786e-02 -2.17245504e-01 -1.20413356e-01
-8.51602137e-01 -7.68780351e-01 1.30480385e+00 3.95262450e-01
-8.65386486e-01 9.64056790e-01 6.94986641e-01 -4.68031824e-01
7.97971711e-02 -7.65725434e-01 -7.28827298e-01 -2.97992796e-01
1.67108908e-01 5.57133973e-01 3.10846537e-01 -2.45578289e-01
4.44332808e-01 1.75434560e-01 -1.85098982e+00 3.56074572e-01
8.25566709e-01 6.23939037e-01 8.63695424e-03 7.06319094e-01
-3.85712147e-01 5.13460517e-01 -9.54675019e-01 -5.45499206e-01
4.11541075e-01 -7.01482594e-01 -9.15891349e-01 -1.51658013e-01
4.11116838e-01 -1.34049281e-01 -4.55216616e-02 6.46015048e-01
7.75943339e-01 1.68869048e-01 1.01734710e+00 -9.24503624e-01
-8.69818866e-01 9.02281940e-01 -5.85490167e-01 2.85490870e-01
3.64290714e-01 3.73105139e-01 1.00921381e+00 -9.64715064e-01
9.28620994e-01 1.07254219e+00 6.13997042e-01 5.60950935e-01
-1.33023739e+00 -4.86610413e-01 1.10122487e-01 -3.09030622e-01
-1.13725722e+00 -7.77796805e-01 2.51902640e-01 -4.42216098e-01
1.18427277e+00 1.19255379e-01 4.32950616e-01 1.36192286e+00
1.17907040e-01 6.23829246e-01 1.06483078e+00 -6.48080289e-01
4.86507475e-01 1.13020130e-02 1.41099587e-01 6.82389587e-02
5.54863393e-01 4.96376976e-02 -5.41098297e-01 -3.07832271e-01
3.04675967e-01 -4.39264387e-01 -3.22838277e-01 2.23918617e-01
-1.02150273e+00 9.24986720e-01 -4.10646982e-02 4.44736391e-01
-4.67876881e-01 2.57649869e-01 5.63677669e-01 4.13993210e-01
2.16697991e-01 7.52721131e-01 -3.16644549e-01 -4.38776881e-01
-9.58669245e-01 6.82981372e-01 9.34693813e-01 1.02262461e+00
2.86051303e-01 2.04365417e-01 -4.85397935e-01 1.11883140e+00
-1.04198746e-01 2.61530876e-01 1.08642018e+00 -3.23013783e-01
8.65179062e-01 1.52742222e-01 2.23416716e-01 -4.61337835e-01
-2.41038382e-01 -2.99839407e-01 -3.35281461e-01 1.68188617e-01
6.57108366e-01 -6.28346562e-01 -1.28766370e+00 2.24387717e+00
-1.84226200e-01 -4.38665539e-01 3.27254459e-03 4.60732281e-01
9.80656087e-01 7.45064318e-01 1.97724506e-01 -4.49901558e-02
1.15824103e+00 -5.41781127e-01 -5.81616402e-01 -5.71765184e-01
8.36445272e-01 -7.43488967e-01 1.32338274e+00 2.64204055e-01
-1.45772457e+00 -4.38332289e-01 -1.06397784e+00 -1.94989115e-01
-4.95839179e-01 -1.13037400e-01 2.19595775e-01 8.54006648e-01
-1.08878088e+00 6.99854434e-01 -4.80795205e-01 -3.76804978e-01
3.24566334e-01 6.93850443e-02 -3.30258042e-01 -1.54498130e-01
-1.12187135e+00 8.62394214e-01 6.34925187e-01 -4.81133550e-01
-6.20816350e-01 -5.45866251e-01 -5.94154000e-01 1.27736345e-01
1.77837342e-01 -7.41345048e-01 1.18644202e+00 -1.00128663e+00
-1.31597388e+00 7.15716898e-01 1.61951501e-02 -4.41579014e-01
7.97412813e-01 -3.36315185e-01 -2.05165654e-01 -3.21388602e-01
-9.74240676e-02 8.81852269e-01 6.90864742e-01 -1.26198459e+00
-6.17921948e-01 -2.02260017e-01 -2.71507919e-01 3.66419584e-01
-2.01102242e-01 1.38959482e-01 -4.73624974e-01 -1.07857907e+00
-1.25079080e-01 -1.02009761e+00 1.66165113e-01 -4.35226023e-01
-5.35914481e-01 -3.59896868e-01 5.35386682e-01 -6.28156126e-01
1.24211872e+00 -1.94816196e+00 4.32113670e-02 7.15513229e-02
1.22924354e-02 2.39758924e-01 -4.38681155e-01 6.44786179e-01
-2.13931069e-01 4.78888005e-01 -7.57179707e-02 -3.09909105e-01
6.70166388e-02 6.11113533e-02 -3.06363374e-01 -5.19687310e-02
2.05432057e-01 8.99057627e-01 -6.19641066e-01 -1.49937049e-01
-3.32514495e-01 2.45164573e-01 -5.80171108e-01 8.10755417e-02
-2.00987667e-01 -7.85037577e-02 -1.55347869e-01 3.18472743e-01
3.94584239e-01 -1.22126348e-01 7.21205249e-02 3.19290638e-01
-8.09392482e-02 7.28987873e-01 -1.10406137e+00 1.26236510e+00
-3.34787786e-01 6.47728801e-01 -3.09161097e-01 -3.22803557e-01
8.10354471e-01 4.06669796e-01 -1.95876449e-01 -7.70884752e-01
1.65326931e-02 2.54200697e-01 1.39309257e-01 -2.19451487e-01
8.60128701e-01 -4.52683866e-01 -1.13362290e-01 9.40620422e-01
4.94968146e-02 -6.04788773e-02 5.28374135e-01 3.48777473e-01
1.40966463e+00 6.09007999e-02 1.64123699e-01 -3.22097868e-01
-7.84465373e-02 -3.02325696e-01 4.23305601e-01 1.01188827e+00
1.58974349e-01 7.49072611e-01 7.95890749e-01 -4.29869331e-02
-1.19851398e+00 -9.65177178e-01 -2.21900120e-02 1.31261313e+00
-1.62385628e-01 -4.29785311e-01 -7.85661697e-01 -7.91621268e-01
9.03399587e-02 1.47231293e+00 -7.56235659e-01 -2.59016812e-01
-3.85176986e-01 -8.24398279e-01 7.94538379e-01 7.15037167e-01
1.10828467e-01 -1.31133485e+00 -6.05188131e-01 1.00539222e-01
-1.79575309e-01 -6.53874516e-01 -5.42490304e-01 5.79326510e-01
-7.20138967e-01 -6.12310886e-01 -6.40392184e-01 -4.14614141e-01
5.13680995e-01 1.98740721e-01 1.37987638e+00 1.49167523e-01
2.39474013e-01 1.17615208e-01 -6.13424003e-01 -6.15491033e-01
-6.98300540e-01 4.79093730e-01 -9.87465158e-02 -4.99864578e-01
2.30491951e-01 -4.97272104e-01 -2.12330222e-01 1.43252388e-01
-1.11027300e+00 1.63866859e-02 6.58319652e-01 7.17755079e-01
2.68793076e-01 -4.94248159e-02 7.76317000e-01 -1.33117890e+00
1.07328188e+00 -8.17282975e-01 -2.37348810e-01 2.03490585e-01
-8.52516115e-01 2.49196544e-01 6.42687082e-01 -4.84724879e-01
-1.09882891e+00 -3.70108724e-01 -2.03842729e-01 -7.59237409e-02
-2.67871439e-01 5.01394033e-01 -5.72192669e-02 5.27231634e-01
8.57661784e-01 6.70159981e-02 -2.42568389e-01 -5.65897763e-01
4.70702469e-01 6.07010484e-01 5.70930839e-01 -7.42080867e-01
7.17303216e-01 -4.47915681e-02 -5.18705726e-01 -6.56200409e-01
-2.50169843e-01 2.94106841e-01 -4.37341213e-01 1.30170211e-01
9.63637650e-01 -6.75388336e-01 3.66496332e-02 4.00185883e-01
-1.11697650e+00 -6.79815233e-01 -6.61345840e-01 3.55072409e-01
-3.73938560e-01 8.97134021e-02 -6.92800581e-01 -4.83034909e-01
-3.26678276e-01 -1.06586969e+00 7.73156941e-01 1.07755281e-01
-6.43597424e-01 -8.11899126e-01 2.32139945e-01 1.15695909e-01
4.84249264e-01 1.82698563e-01 1.35805964e+00 -8.83762836e-01
-1.68241680e-01 -4.04531509e-01 -1.74814895e-01 2.13135183e-01
9.24334973e-02 1.46813855e-01 -8.11974287e-01 -3.44766259e-01
-2.83434719e-01 -6.38999283e-01 9.46207464e-01 1.64442047e-01
7.90008307e-01 -3.74855548e-01 -1.61388800e-01 3.49900961e-01
1.28691781e+00 2.00171679e-01 9.38800037e-01 2.76471704e-01
5.63866317e-01 7.42362678e-01 9.77215767e-02 2.74521261e-01
2.33606711e-01 7.28263199e-01 1.43544242e-01 8.27849954e-02
-3.74285132e-01 -4.44712788e-01 3.03675234e-01 6.91515267e-01
1.62824944e-01 -9.42552030e-01 -8.34656179e-01 7.20241547e-01
-1.67789149e+00 -1.10269308e+00 1.68186743e-02 2.37656045e+00
1.16333723e+00 4.44519341e-01 2.93552130e-01 -7.15306476e-02
6.48025334e-01 1.70051917e-01 -3.22229683e-01 -6.86860442e-01
-3.22319508e-01 3.38611573e-01 4.11951512e-01 2.41455972e-01
-5.86043000e-01 7.26143301e-01 7.48173380e+00 9.53314602e-01
-9.46011603e-01 1.15055561e-01 7.52524734e-01 -3.56100500e-01
-6.26495838e-01 -2.44176582e-01 -8.70037436e-01 8.34295273e-01
1.03347731e+00 -2.17223525e-01 3.42731953e-01 5.37921369e-01
-2.05196783e-01 -1.51155353e-01 -1.29167140e+00 5.05194604e-01
1.22927977e-02 -9.85149145e-01 2.72595316e-01 9.00284871e-02
7.46605694e-01 2.12408096e-01 4.97197136e-02 3.65348935e-01
7.62827575e-01 -1.14367318e+00 1.08945179e+00 1.90594420e-01
7.50493348e-01 -7.03050435e-01 6.76632702e-01 3.83062482e-01
-6.78817332e-01 -7.55722495e-03 -8.29727501e-02 1.34856135e-01
1.44991189e-01 4.00040716e-01 -7.68765092e-01 3.70816678e-01
3.29806358e-01 -3.47836805e-03 -8.61747146e-01 7.94231176e-01
-4.37453032e-01 8.43184590e-01 -2.59798288e-01 -3.19092115e-03
2.20620796e-01 3.80924717e-02 5.51547229e-01 1.53403234e+00
3.39559406e-01 -1.63223460e-01 -2.80495822e-01 8.33266616e-01
-2.38645703e-01 2.10056994e-02 -6.07493520e-01 -2.40020946e-01
7.08646297e-01 9.29546714e-01 -8.08374286e-01 -3.94882321e-01
-5.12506545e-01 7.78489292e-01 3.81127715e-01 3.53454232e-01
-5.47469676e-01 -6.97220087e-01 4.77905929e-01 3.53129059e-01
7.07441509e-01 1.18186824e-01 -3.46626580e-01 -8.31899345e-01
2.03950126e-02 -1.00694656e+00 4.47973549e-01 -8.65815401e-01
-1.09188974e+00 6.87214255e-01 2.67815828e-01 -7.50681818e-01
-7.09445953e-01 -2.24385425e-01 -7.23401785e-01 1.15995491e+00
-1.02809846e+00 -5.82389057e-01 -1.56306222e-01 1.05287217e-01
4.58790243e-01 -1.32916585e-01 5.42318583e-01 2.02461839e-01
-5.04961491e-01 9.00081992e-01 1.16744682e-01 3.99784409e-02
6.89116001e-01 -1.43250930e+00 9.63533044e-01 9.49712694e-01
1.52226582e-01 6.41587019e-01 7.98821151e-01 -7.41108835e-01
-1.03346884e+00 -1.10699987e+00 9.84431982e-01 -7.30395317e-01
1.35094300e-01 -4.25785989e-01 -1.04754448e+00 7.35940456e-01
3.85537714e-01 -4.30718839e-01 5.97487688e-01 2.31761307e-01
-6.09548628e-01 3.70922476e-01 -1.13420570e+00 7.08084881e-01
9.46420491e-01 -2.33514041e-01 -6.39454246e-01 2.40332231e-01
7.42755294e-01 -4.06799585e-01 -6.22197509e-01 3.55367884e-02
5.09501576e-01 -1.03795922e+00 4.37080026e-01 -7.01539636e-01
8.46384883e-01 5.77006526e-02 -2.03261182e-01 -1.66050255e+00
-4.20963079e-01 -4.58120316e-01 2.02535875e-02 1.39251482e+00
8.95666063e-01 -7.50441909e-01 5.74460387e-01 6.97010458e-01
-6.71879202e-02 -9.85374868e-01 -6.51158810e-01 -7.94777572e-01
5.04325211e-01 -2.28748024e-01 8.27284217e-01 9.23200250e-01
-3.02836508e-01 7.44549692e-01 -7.97403306e-02 -7.30820537e-01
-5.54986857e-02 -3.61144722e-01 8.47662032e-01 -9.15340424e-01
-4.89293575e-01 -6.86021805e-01 4.13181484e-02 -8.62433374e-01
-2.11743638e-01 -7.88519561e-01 1.82273135e-01 -1.38533723e+00
1.87454432e-01 -5.67862451e-01 -1.48508713e-01 6.25871539e-01
-5.40641665e-01 1.05297551e-01 2.68340439e-01 2.35426709e-01
-2.49033704e-01 4.46562588e-01 5.97358882e-01 3.96383792e-01
-3.78674418e-01 -2.29555622e-01 -1.20234239e+00 4.00791109e-01
6.78277194e-01 -5.13659954e-01 -6.04476333e-01 -7.34683633e-01
3.67261499e-01 -8.19652155e-02 -4.82576042e-02 -7.99693644e-01
-7.33243003e-02 -7.70464092e-02 4.98215079e-01 -3.12240303e-01
8.79740715e-03 -2.74840653e-01 1.35386497e-01 4.18288589e-01
-8.37927520e-01 5.66294491e-01 4.30246919e-01 5.60137570e-01
1.77439198e-01 -6.38511419e-01 8.73641372e-01 -1.43305184e-02
-1.70882300e-01 -1.54000089e-01 -4.37778443e-01 6.71396375e-01
6.37712657e-01 -3.80621105e-01 -4.50292826e-01 -5.49275696e-01
-6.35305271e-02 -4.57911380e-02 6.10295057e-01 5.75490654e-01
1.54816568e-01 -1.32493353e+00 -9.45995331e-01 1.97409108e-01
1.70269608e-01 -2.70663947e-01 -2.55806804e-01 3.84255409e-01
-2.93075651e-01 2.98867285e-01 -5.78148523e-03 -1.27207473e-01
-1.18711078e+00 2.39330173e-01 3.21889762e-03 -3.61728042e-01
-4.71497953e-01 9.87805128e-01 1.14464767e-01 -2.98925608e-01
2.47332513e-01 -1.35075882e-01 -2.94681415e-02 3.14895272e-01
4.44936991e-01 4.07138675e-01 3.59670192e-01 -5.89996338e-01
-3.29906717e-02 9.08307582e-02 -6.76154792e-01 -4.88218039e-01
1.13909137e+00 2.67713964e-01 3.28688473e-01 4.23511744e-01
9.78095949e-01 1.63226157e-01 -1.01503873e+00 -2.86509544e-01
-1.41317204e-01 -3.37504417e-01 4.60552275e-02 -9.41776931e-01
-7.56239951e-01 5.56575894e-01 3.60348374e-01 3.13960195e-01
9.65915740e-01 -2.43370339e-01 8.55016351e-01 2.04942390e-01
1.49021149e-01 -1.21176112e+00 -5.28736487e-02 3.62415701e-01
8.95805538e-01 -6.58876538e-01 2.91307926e-01 -5.79684041e-02
-7.98319817e-01 9.10089374e-01 6.14783227e-01 1.20545991e-01
1.38971195e-01 3.37906599e-01 -1.54830933e-01 -1.25173628e-01
-1.24264705e+00 9.46253538e-02 1.93099201e-01 5.38702369e-01
6.72857821e-01 -1.30879223e-01 -5.66966236e-01 5.63548207e-01
-5.89079022e-01 -6.90277889e-02 5.48892856e-01 1.16692817e+00
-2.91922808e-01 -1.15220201e+00 -2.28154540e-01 8.95491004e-01
-5.75193405e-01 -2.90378302e-01 -8.74293149e-01 1.13375080e+00
1.54550835e-01 7.45237887e-01 2.86523640e-01 -3.08660656e-01
3.11645508e-01 3.41294348e-01 5.03773153e-01 -8.61931860e-01
-9.16244388e-01 -1.20450325e-01 3.32676083e-01 -2.14081764e-01
3.51883858e-01 -8.11491430e-01 -1.04297078e+00 -3.83655161e-01
-6.14476860e-01 1.18730970e-01 7.64260828e-01 8.57891560e-01
5.33794522e-01 3.94639313e-01 2.84447640e-01 -6.03592634e-01
-8.79942179e-01 -1.23581922e+00 -5.74628353e-01 7.14500546e-01
1.19144961e-01 -4.74423379e-01 -4.55460399e-01 2.01368984e-03] | [11.544432640075684, 9.06477165222168] |
4ff20331-1dfa-48ae-b427-88262273f332 | an-intelligent-safety-system-for-human | 1812.03953 | null | http://arxiv.org/abs/1812.03953v2 | http://arxiv.org/pdf/1812.03953v2.pdf | An Intelligent Safety System for Human-Centered Semi-Autonomous Vehicles | Nowadays, automobile manufacturers make efforts to develop ways to make cars
fully safe. Monitoring driver's actions by computer vision techniques to detect
driving mistakes in real-time and then planning for autonomous driving to avoid
vehicle collisions is one of the most important issues that has been
investigated in the machine vision and Intelligent Transportation Systems
(ITS). The main goal of this study is to prevent accidents caused by fatigue,
drowsiness, and driver distraction. To avoid these incidents, this paper
proposes an integrated safety system that continuously monitors the driver's
attention and vehicle surroundings, and finally decides whether the actual
steering control status is safe or not. For this purpose, we equipped an
ordinary car called FARAZ with a vision system consisting of four mounted
cameras along with a universal car tool for communicating with surrounding
factory-installed sensors and other car systems, and sending commands to
actuators. The proposed system leverages a scene understanding pipeline using
deep convolutional encoder-decoder networks and a driver state detection
pipeline. We have been identifying and assessing domestic capabilities for the
development of technologies specifically of the ordinary vehicles in order to
manufacture smart cars and eke providing an intelligent system to increase
safety and to assist the driver in various conditions/situations. | ['Hadi Abdi Khojasteh', 'Ebrahim Ansari', 'Parvin Razzaghi', 'Alireza Abbas Alipour'] | 2018-12-10 | null | null | null | null | ['steering-control'] | ['computer-vision'] | [-3.83980870e-02 2.35053077e-01 3.97102982e-02 -5.62871039e-01
-1.65078595e-01 -2.70642102e-01 5.85295737e-01 -5.05194187e-01
-4.25841421e-01 7.89652094e-02 -1.34083077e-01 -6.06573045e-01
4.41730544e-02 -5.75013101e-01 -4.03431714e-01 -6.07696474e-01
5.66661716e-01 7.64854178e-02 5.08432865e-01 -7.27431595e-01
4.06186789e-01 8.25626850e-01 -2.20491862e+00 1.80254597e-02
4.54189241e-01 1.01061821e+00 5.01148582e-01 9.35435891e-01
2.68464535e-01 9.30960298e-01 -6.56947494e-02 -1.26931136e-02
1.24016784e-01 -1.75520275e-02 -2.53040701e-01 8.77578277e-03
6.51648939e-02 -6.67343318e-01 -5.15486300e-01 9.53407049e-01
1.97398096e-01 2.58192588e-02 2.49773458e-01 -1.68512690e+00
-1.20461481e-02 -1.48388728e-01 1.13714658e-01 2.34750912e-01
-1.72431216e-01 9.73230362e-01 1.18624330e-01 -3.38115603e-01
1.92365691e-01 1.11388075e+00 2.60183752e-01 6.19263351e-01
-4.13854420e-01 -7.07202673e-01 -1.60381660e-01 8.43699396e-01
-1.10605907e+00 -9.81364608e-01 9.32004213e-01 -5.32046914e-01
8.83050382e-01 5.61871789e-02 3.70900065e-01 9.93650079e-01
6.58084095e-01 3.38696808e-01 6.54272497e-01 -1.74122319e-01
2.68431932e-01 7.72658050e-01 5.71618497e-01 6.30771458e-01
2.76239276e-01 5.33361554e-01 -1.53646037e-01 6.98055506e-01
5.07051498e-02 1.04802802e-01 2.28781074e-01 2.86695212e-02
-8.97777140e-01 6.44231856e-01 5.27405785e-03 1.11345336e-01
-5.67575693e-01 1.61677197e-01 4.35298622e-01 2.55778849e-01
-1.79328829e-01 -6.45668358e-02 -1.84426263e-01 -3.13036799e-01
-2.76190370e-01 3.44098383e-03 3.92330945e-01 1.13889790e+00
7.90150344e-01 1.35690629e-01 -1.35658637e-01 2.89407402e-01
5.25081277e-01 8.02548409e-01 1.58725932e-01 -1.26187611e+00
1.75071254e-01 7.33045697e-01 2.47853741e-01 -1.01055443e+00
-5.41053593e-01 -9.14556012e-02 -3.26730251e-01 6.43154025e-01
-5.31738363e-02 -1.32134289e-01 -4.25328642e-01 1.18406594e+00
2.65434146e-01 1.83140680e-01 2.01919764e-01 1.05171525e+00
5.74479818e-01 4.60095614e-01 5.52476496e-02 1.29148319e-01
1.61570394e+00 -7.29590178e-01 -1.07364392e+00 -5.65243185e-01
7.82767057e-01 -6.01225019e-01 7.33860254e-01 5.71969211e-01
-6.86589420e-01 -1.18633258e+00 -1.26419199e+00 -2.05586329e-01
-7.02248573e-01 4.66556162e-01 1.60884529e-01 7.86115408e-01
-9.57660794e-01 1.37230068e-01 -6.29686356e-01 -3.65321100e-01
1.84805810e-01 2.73569524e-01 -3.95003438e-01 -1.08977482e-01
-1.22445548e+00 1.56690347e+00 2.30421141e-01 4.59425807e-01
-1.09019232e+00 -2.90954292e-01 -9.48524356e-01 -1.44291937e-01
4.37753767e-01 -4.23690915e-01 1.41741431e+00 -6.14286900e-01
-1.51029348e+00 5.96285403e-01 -2.24411875e-01 -7.06589878e-01
1.89723521e-01 -3.13545167e-01 -1.07806218e+00 -1.01390801e-01
2.73944531e-02 5.50668240e-01 6.62081361e-01 -8.26738775e-01
-1.17628348e+00 -4.37610388e-01 -2.29940400e-01 -1.17092524e-02
-8.35131854e-03 1.41968638e-01 -4.37463403e-01 5.77887654e-01
-5.67178249e-01 -1.11342251e+00 -2.44102068e-02 -2.09672853e-01
-2.79372334e-01 -3.89225096e-01 1.61268759e+00 -4.92957473e-01
1.06656086e+00 -2.26009321e+00 -7.13720739e-01 -2.23819874e-02
1.95959844e-02 7.60300457e-01 -1.29553586e-01 1.98065955e-02
1.71340346e-01 -6.96330965e-01 4.31354314e-01 -2.60403633e-01
2.78116502e-02 4.11893040e-01 -2.28248715e-01 6.31624460e-01
4.01761651e-01 6.57324910e-01 -5.13864875e-01 -1.92716256e-01
1.03785372e+00 4.57653642e-01 -1.99605778e-01 4.79044646e-01
4.07579318e-02 4.85334665e-01 -4.71329689e-01 3.31536204e-01
5.73482275e-01 6.34302557e-01 -3.85196775e-01 -1.46265537e-01
-9.47406232e-01 3.71032268e-01 -1.16621900e+00 8.47348213e-01
-5.34362435e-01 1.10232198e+00 3.61081660e-01 -9.69233274e-01
9.35511708e-01 2.52920717e-01 1.30511358e-01 -1.16076064e+00
7.64240146e-01 4.71009538e-02 -2.24831887e-02 -1.25621748e+00
4.26824450e-01 3.82747144e-01 -7.88173378e-02 -1.43256918e-01
-3.18947971e-01 -5.29936748e-03 2.22580694e-02 -8.25698674e-02
1.15277755e+00 -4.52980883e-02 -2.17255428e-01 -6.72863349e-02
9.54067886e-01 2.43914902e-01 4.51167583e-01 2.56479502e-01
-8.56393993e-01 -2.65099015e-02 2.69966811e-01 -5.58245838e-01
-9.54650939e-01 -6.49255693e-01 3.01021278e-01 7.45104730e-01
4.08719540e-01 3.10138483e-02 -8.96765411e-01 -1.36074841e-01
-2.08682850e-01 1.45583713e+00 -2.07031950e-01 -8.74397039e-01
-3.94756347e-01 2.90969849e-01 2.87457556e-01 4.24692631e-01
7.27626681e-01 -8.87690485e-01 -1.25252044e+00 1.04551829e-01
1.64611205e-01 -1.38759315e+00 -2.86879718e-01 -7.48579949e-02
-3.59988422e-04 -1.23082995e+00 2.63254166e-01 -6.72306538e-01
2.94063330e-01 9.03330982e-01 4.44713742e-01 -1.27051786e-01
-3.69670272e-01 3.23107421e-01 -1.43174408e-02 -9.41717207e-01
-8.02265525e-01 -1.35158330e-01 2.26431817e-01 2.86303043e-01
9.00913954e-01 1.99295253e-01 -8.21592212e-01 5.45733631e-01
-4.23032343e-01 1.98703602e-01 6.67668343e-01 2.82248352e-02
1.21263713e-01 4.48682874e-01 6.16666675e-01 -5.80972791e-01
3.83725286e-01 -5.41074455e-01 -1.06858838e+00 -1.07197173e-01
-7.56341755e-01 -1.80427969e-01 5.51027477e-01 2.63893366e-01
-1.22300982e+00 3.41276944e-01 -3.22825521e-01 -3.86083126e-01
-9.19061184e-01 -2.61073112e-01 -4.49822903e-01 1.56824499e-01
2.13923022e-01 1.02514885e-01 5.77108204e-01 5.06780632e-02
4.45732623e-01 1.36089504e+00 8.09424162e-01 4.70393151e-01
5.08797109e-01 4.04999465e-01 1.37322530e-01 -9.88103569e-01
-3.78594756e-01 -7.87903666e-01 -5.24504185e-01 -1.00868988e+00
1.21629119e+00 -1.02960432e+00 -1.33092761e+00 4.81073350e-01
-1.33343577e+00 1.57845989e-02 1.20568238e-01 5.48469722e-01
-4.76504505e-01 2.66717803e-02 -2.14453131e-01 -9.51256752e-01
-1.36243463e-01 -1.51639473e+00 1.07948124e+00 4.31955069e-01
-3.38338464e-02 -4.20330912e-01 -1.13259545e-02 9.30405617e-01
7.37269223e-01 -1.84622124e-01 3.16515774e-01 -1.66272461e-01
-7.65061319e-01 -3.88019681e-01 -1.78477660e-01 7.00332999e-01
1.11761816e-01 2.05678776e-01 -1.27633393e+00 1.40537038e-01
1.06548563e-01 3.24139178e-01 5.35034239e-01 3.71257126e-01
9.19892013e-01 -1.08971030e-01 -5.18284798e-01 1.90245852e-01
1.09573114e+00 8.38832259e-01 1.02998769e+00 2.62031227e-01
3.96087825e-01 1.08202481e+00 9.81991410e-01 8.35066736e-02
7.30366468e-01 6.40318692e-01 8.58310640e-01 -7.41135925e-02
-1.94831967e-01 1.88241497e-01 7.38674939e-01 4.28368062e-01
2.52411962e-01 -1.26593471e-01 -7.00641394e-01 5.10592878e-01
-1.77870667e+00 -1.11423683e+00 -6.78672373e-01 1.89018178e+00
1.11455865e-01 3.73070836e-01 -1.11537911e-01 1.85430512e-01
7.43663192e-01 -4.77442175e-01 -5.27262747e-01 -9.18087184e-01
4.34755445e-01 -2.71543205e-01 9.78203297e-01 3.95192981e-01
-1.02330840e+00 8.24812055e-01 5.34849310e+00 3.53971869e-01
-1.25997710e+00 1.01210035e-01 4.30284798e-01 -7.19548315e-02
1.39659181e-01 -1.08879909e-01 -1.07706273e+00 6.66975021e-01
1.78140545e+00 2.33213499e-01 2.79751748e-01 1.24506605e+00
1.04188824e+00 -3.16527814e-01 -7.92407215e-01 7.83134282e-01
-2.61785034e-02 -1.16071939e+00 -6.04297042e-01 -1.05336271e-01
7.38896355e-02 9.80430692e-02 2.55383216e-02 4.47192430e-01
-3.29582766e-02 -7.48385966e-01 6.93676054e-01 9.22031522e-01
4.92671758e-01 -1.02404845e+00 8.19102287e-01 6.31486118e-01
-1.02816057e+00 -4.58812147e-01 -1.41732588e-01 -3.96092862e-01
3.40835631e-01 2.90529639e-01 -1.10118401e+00 -8.28611106e-02
5.48364758e-01 4.79432970e-01 -4.93583202e-01 5.56714535e-01
-1.00308329e-01 5.13171017e-01 7.63115734e-02 -1.83059782e-01
2.89991081e-01 -2.67198861e-01 4.74285096e-01 1.02326095e+00
2.48381734e-01 3.14228348e-02 -2.02843681e-01 8.37341249e-01
5.24375796e-01 -5.81165552e-01 -1.16291523e+00 3.06707978e-01
2.80316323e-01 1.47630346e+00 -2.82582015e-01 -1.65405393e-01
-6.52906954e-01 5.89834273e-01 -3.21745038e-01 3.60825211e-02
-1.18923831e+00 -7.59595394e-01 1.28541493e+00 4.51174527e-01
1.23280557e-02 -4.61160451e-01 -2.79385746e-01 -2.01991171e-01
-9.64227021e-02 -1.38667837e-01 -1.79136276e-01 -1.18860638e+00
-3.78179342e-01 4.23701584e-01 -1.26735240e-01 -1.15959227e+00
-6.60256222e-02 -9.97570217e-01 -8.58831227e-01 7.54765987e-01
-2.03432441e+00 -9.83857453e-01 -6.86822534e-01 5.54013968e-01
8.43690097e-01 -5.42654395e-01 1.88773677e-01 5.01284957e-01
-1.08471346e+00 9.58204493e-02 -3.51562440e-01 -1.31693035e-01
6.84761643e-01 -5.72990894e-01 2.39666358e-01 1.19940746e+00
-8.20903838e-01 2.09812820e-01 8.74674618e-01 -4.52799886e-01
-1.96099043e+00 -1.38885558e+00 8.92960846e-01 -4.77718353e-01
2.18496427e-01 -1.72201201e-01 -5.68600357e-01 5.52594602e-01
5.65020084e-01 -2.47706130e-01 3.83790247e-02 -7.43835270e-01
3.01676482e-01 -6.09739304e-01 -1.01951039e+00 7.37454772e-01
6.16454184e-01 -4.51851457e-01 -5.36287546e-01 1.16469800e-01
6.29204094e-01 1.42348735e-02 -1.62987411e-01 1.28737122e-01
2.78284818e-01 -1.03916907e+00 5.08875012e-01 -3.22619051e-01
9.37206000e-02 -7.93622613e-01 1.15581051e-01 -9.26280797e-01
-3.77507031e-01 -5.47182977e-01 1.48723945e-01 1.02207351e+00
7.17529561e-04 -8.75175178e-01 5.66030562e-01 9.65039074e-01
-8.26222837e-01 -1.85430139e-01 -9.77804780e-01 -3.55489939e-01
-6.24419212e-01 -1.04097831e+00 4.66542393e-01 1.94917560e-01
-1.86971620e-01 4.17286277e-01 -2.34267950e-01 6.88250601e-01
4.01621372e-01 -5.49403548e-01 1.03341329e+00 -9.71029401e-01
5.89264572e-01 -4.46815610e-01 -8.09265137e-01 -6.42035007e-01
2.12363780e-01 -3.03961247e-01 4.86306489e-01 -1.38253915e+00
-2.91601360e-01 -5.63402772e-02 -7.15854764e-02 1.97258279e-01
3.56135547e-01 -2.60660619e-01 -3.07661116e-01 -2.42190838e-01
-5.43079436e-01 3.11052918e-01 9.12669182e-01 -4.68480289e-02
1.84250131e-01 2.28418678e-01 -6.72401488e-01 5.84927976e-01
8.95943999e-01 -8.17566216e-02 -4.81014907e-01 -2.54308879e-01
6.69935271e-02 -1.82176456e-02 8.01001072e-01 -1.44826281e+00
8.64228785e-01 -2.44003743e-01 -7.52242729e-02 -9.84361112e-01
3.08489054e-01 -1.47952747e+00 1.74852639e-01 7.35334098e-01
-2.30235204e-01 -3.96056958e-02 2.82899082e-01 4.20425981e-01
-7.58021325e-02 2.45531313e-02 9.32954967e-01 2.78494030e-01
-1.51148546e+00 1.34206032e-02 -1.13916540e+00 -7.92145967e-01
1.84426510e+00 -4.06288207e-01 -5.66456854e-01 -2.57918268e-01
-9.09009427e-02 5.84168077e-01 6.66987672e-02 9.94511127e-01
7.36128628e-01 -1.05103695e+00 -4.55269367e-01 8.43765914e-01
2.70394385e-01 -3.25331092e-01 5.59101045e-01 8.03684175e-01
-6.22999310e-01 1.08939910e+00 -3.70587915e-01 -5.69380820e-01
-1.26524925e+00 7.60812938e-01 5.06634533e-01 5.95015228e-01
-4.08325344e-01 2.36571789e-01 -1.05040014e-01 -3.14597460e-03
1.86695039e-01 -3.73849750e-01 -6.33547544e-01 -2.47928247e-01
7.54836261e-01 8.18829477e-01 4.65926886e-01 -9.72392321e-01
-4.96346533e-01 3.11880499e-01 1.67699173e-01 3.05670291e-01
7.16269553e-01 -5.39127469e-01 3.95846486e-01 1.94507018e-01
1.16865540e+00 -5.25606930e-01 -1.40526164e+00 3.00295800e-01
-1.29771888e-01 -1.75903469e-01 6.10509098e-01 -6.45167410e-01
-1.14923751e+00 9.11956668e-01 1.04830050e+00 1.78085148e-01
9.51615393e-01 -1.54671282e-01 1.05753410e+00 4.08972234e-01
3.83373469e-01 -1.59311414e+00 -3.60184163e-01 4.98731881e-01
5.23678064e-01 -1.59629965e+00 -5.66643894e-01 -1.66658014e-01
-9.48264778e-01 1.17530239e+00 8.39118302e-01 2.10397124e-01
7.19155848e-01 4.35803205e-01 3.15145314e-01 -3.81918401e-01
-1.03218472e+00 -4.70786005e-01 1.65733874e-01 6.96788907e-01
-1.47003666e-01 1.59985438e-01 2.81554729e-01 5.92046320e-01
1.71486735e-02 2.19979942e-01 4.22531575e-01 7.60593116e-01
-1.03745830e+00 -5.17646492e-01 -4.15261567e-01 2.40258068e-01
2.40280345e-01 5.95776498e-01 2.94249989e-02 5.13272285e-01
7.39595413e-01 1.70935297e+00 3.55689377e-01 -7.42664218e-01
8.90199065e-01 7.87372142e-02 -3.21876943e-01 -2.64435709e-01
-2.59471536e-01 -5.65576375e-01 3.46786380e-01 -8.28679681e-01
-1.41578212e-01 -6.30248368e-01 -1.25241387e+00 -4.02469724e-01
-1.41131878e-01 -3.64312053e-01 1.19446850e+00 1.22421622e+00
6.87804520e-01 8.80455196e-01 9.00798321e-01 -5.90274394e-01
-3.46567452e-01 -5.65811872e-01 -3.31930220e-01 1.12696193e-01
4.97940272e-01 -7.20580339e-01 -1.04723640e-01 9.43818968e-03] | [7.7683587074279785, -0.776201069355011] |
938f2f12-010c-48e8-89ee-14b5ab43ab0f | continuous-and-interactive-factual-knowledge | null | null | https://openreview.net/forum?id=GxT3-eeWLNx | https://openreview.net/pdf?id=GxT3-eeWLNx | Continuous and Interactive Factual Knowledge Learning in Verification Dialogues | Knowledge bases (KBs) used in applications such as dialogue systems need to be continuously expanded in order to serve the users well. This process is known as knowledge base completion (KBC). A piece of knowledge or a fact is often represented as a triple (s, r, t), meaning that the entity s and the entity t have the relation r or are linked by r. KBC builds a model to infer missing facts from the existing ones in a given KB. Existing KBC research typically makes the closed-world assumption that to infer a new fact (s, r, t), it assumes that s, r and t are already in the KB, but are not linked. Clearly, this assumption is a serious limitation. In this paper, we eliminate this assumption and allow s, r and/or t to be unknown to the KB, which we call open-world knowledge base completion (OKBC). We focus on solving OKBC via user interactions, which enables the proposed system to potentially serve as an engine for learning new knowledge during dialogue. Experimental results show the effectiveness of the proposed approach. | ['Anonymous'] | 2020-10-15 | null | null | null | neurips-workshop-hamlets-2020-12 | ['knowledge-base-completion', 'knowledge-base-completion'] | ['graphs', 'knowledge-base'] | [-2.29431883e-01 6.64328456e-01 -2.21418068e-01 -3.43288958e-01
-3.42768162e-01 -6.79652512e-01 5.19720376e-01 4.27278012e-01
-2.42209449e-01 1.50914800e+00 1.81942895e-01 -2.93128431e-01
-1.40508011e-01 -9.43542957e-01 -6.42326176e-01 -3.26368362e-01
5.29680736e-02 5.56134880e-01 5.93698144e-01 -6.38059855e-01
1.40561521e-01 1.16099440e-01 -1.41122651e+00 4.01584238e-01
1.05372429e+00 4.34657454e-01 2.96106100e-01 3.66583198e-01
-7.26070642e-01 9.82779324e-01 -7.68899977e-01 -8.13565731e-01
-1.65797234e-01 -3.51482332e-01 -1.64205778e+00 -2.19521239e-01
-2.50700265e-01 -4.98694390e-01 -6.13860153e-02 1.15670931e+00
1.22215502e-01 1.91845939e-01 3.91729265e-01 -1.37851405e+00
-4.59939182e-01 8.14882636e-01 -2.08109945e-01 3.22875939e-02
8.16577733e-01 -5.12906849e-01 8.24266255e-01 -5.88903308e-01
7.40741014e-01 1.11982536e+00 5.44493139e-01 5.26004672e-01
-5.01221061e-01 -3.76597285e-01 2.83432215e-01 7.48775423e-01
-1.59493232e+00 -3.19033712e-01 5.24807334e-01 -1.55349225e-01
8.55835021e-01 4.97029275e-01 7.27699220e-01 4.32320982e-01
-2.35747799e-01 6.78935289e-01 1.00892413e+00 -8.13083887e-01
3.34781110e-01 7.74294138e-01 5.37029743e-01 5.97656190e-01
4.73858207e-01 -6.50376856e-01 -5.67565501e-01 -1.61968857e-01
7.61269689e-01 -4.12062407e-01 -5.77938676e-01 -1.68765023e-01
-1.06204176e+00 4.66690809e-01 1.27028629e-01 3.53071481e-01
-4.97857481e-01 -3.89992237e-01 3.48642588e-01 3.45690012e-01
2.33152043e-02 3.52616072e-01 -6.70568824e-01 -2.84874707e-01
-2.06426397e-01 3.28198999e-01 1.38939571e+00 1.29242206e+00
1.00379539e+00 -4.25240844e-01 3.11031014e-01 6.74812734e-01
2.61618733e-01 1.04340419e-01 5.55725455e-01 -7.94977307e-01
4.70381498e-01 1.00978017e+00 8.50117207e-01 -1.06172621e+00
-3.02837729e-01 -4.12349626e-02 -5.33453703e-01 -4.29475307e-01
3.25564802e-01 -1.53019264e-01 -4.72529769e-01 1.62924671e+00
8.29111457e-01 5.78092515e-01 7.64766335e-01 8.30636203e-01
1.28722572e+00 7.96408415e-01 -4.11146358e-02 -7.52154231e-01
1.37947965e+00 -5.73423147e-01 -1.27597845e+00 9.01859719e-03
6.79283857e-01 -5.84107935e-01 6.91605508e-01 3.42700660e-01
-8.51252913e-01 -4.00289625e-01 -9.05649960e-01 -1.07662886e-01
-6.93382204e-01 -2.13351265e-01 6.64309561e-01 6.70611620e-01
-7.92748809e-01 2.19088197e-01 -4.29437518e-01 -4.32891667e-01
-2.54964739e-01 1.34341910e-01 -5.25172174e-01 -3.46602619e-01
-1.90463364e+00 1.26969361e+00 1.12484658e+00 4.56443906e-01
-6.63845599e-01 -1.90573201e-01 -7.57262945e-01 3.76655757e-02
1.11470401e+00 -6.25838041e-01 1.29237819e+00 -5.76815426e-01
-1.28651667e+00 5.65252364e-01 -2.81745076e-01 -4.14637208e-01
2.04505250e-01 -3.31174314e-01 -7.97100604e-01 8.95827077e-03
1.13301031e-01 1.32578894e-01 3.81864578e-01 -1.68985820e+00
-1.10371101e+00 -4.70724553e-01 8.71434689e-01 7.16919303e-01
-1.18456610e-01 7.40990043e-02 -8.30336273e-01 -7.47531727e-02
2.63861746e-01 -5.35406351e-01 1.47686332e-01 -5.84433734e-01
-4.87595469e-01 -3.90964538e-01 7.83206224e-01 -1.01354861e+00
1.54314733e+00 -1.58362591e+00 6.27420051e-03 2.85692126e-01
1.80222735e-01 7.14600384e-01 3.42995763e-01 7.78831482e-01
4.40219119e-02 2.12909415e-01 -9.55221429e-02 2.30161577e-01
-2.39471257e-01 5.95229566e-01 -2.97086626e-01 -2.05536842e-01
-2.14404300e-01 5.79217136e-01 -1.05393934e+00 -7.23081172e-01
1.71767026e-02 1.60693824e-01 -1.03187300e-01 9.45787728e-02
-6.00284576e-01 2.73350269e-01 -6.80990815e-01 3.89076740e-01
6.80156350e-01 -8.39920864e-02 6.89855218e-01 -3.46201569e-01
-1.58480152e-01 1.88618034e-01 -1.57833314e+00 1.43863666e+00
-2.54125237e-01 4.61540893e-02 4.23998274e-02 -9.81707275e-01
1.01495004e+00 8.93548310e-01 2.07441822e-02 -1.65300757e-01
-1.26857877e-01 1.75086379e-01 -7.00026378e-02 -1.01253724e+00
6.85995460e-01 -2.52747685e-01 2.34552786e-01 2.20007539e-01
-1.02155685e-01 -3.94725129e-02 2.56369978e-01 4.07875061e-01
7.50129044e-01 5.73567525e-02 7.62039185e-01 8.31097215e-02
9.57786024e-01 2.13451415e-01 7.86569238e-01 6.73356175e-01
6.19413331e-03 -3.96806955e-01 6.46057785e-01 -2.62627274e-01
-7.27856398e-01 -5.52251637e-01 1.49690330e-01 6.03545964e-01
5.36930680e-01 -6.94391012e-01 -5.63005030e-01 -8.23424578e-01
-2.55068153e-01 8.19857657e-01 -3.45078379e-01 -8.68926570e-02
-4.60459560e-01 -1.14633083e-01 5.15060604e-01 2.04439119e-01
7.57369399e-01 -1.01161265e+00 -3.83399606e-01 4.43830043e-01
-7.80977190e-01 -1.05324435e+00 1.40121907e-01 -1.97704613e-01
-7.18389988e-01 -1.38930190e+00 -3.49148065e-01 -7.42110431e-01
6.82650805e-01 2.53566146e-01 8.39705169e-01 3.97509456e-01
2.83954531e-01 4.68484014e-01 -8.43878090e-01 -5.51430106e-01
-4.03713107e-01 -8.27597901e-02 4.30812016e-02 3.94959375e-02
3.29125375e-01 -3.71311486e-01 -1.33475691e-01 5.18953741e-01
-1.02327418e+00 2.85694718e-01 2.75349170e-01 7.43653834e-01
3.93498927e-01 8.21157932e-01 9.79146302e-01 -1.18566656e+00
7.96151757e-01 -6.14406645e-01 -3.12018067e-01 9.36674595e-01
-3.47830266e-01 5.06806634e-02 5.03796399e-01 -9.11399126e-02
-1.68399000e+00 -2.55294323e-01 3.28559093e-02 1.42350331e-01
-1.83495194e-01 1.35064256e+00 -6.86447918e-01 2.55800694e-01
4.11637425e-01 2.89593756e-01 -2.46280029e-01 -5.07637620e-01
5.10359406e-01 7.99667358e-01 5.99887669e-01 -8.29125464e-01
5.05406260e-01 1.70124725e-01 -1.75623104e-01 -6.80164993e-01
-9.60680485e-01 -6.99894726e-01 -8.49296689e-01 -3.96853119e-01
5.06020188e-01 -9.70213234e-01 -8.23387027e-01 2.87066132e-01
-1.38988316e+00 1.39058471e-01 -9.80597734e-02 5.70397437e-01
-1.70645833e-01 5.64603448e-01 -2.19179258e-01 -1.18540406e+00
-1.21037148e-01 -6.33352399e-01 2.58168876e-01 4.96178985e-01
-5.23994304e-02 -1.16119862e+00 -1.49935216e-01 5.50351739e-01
2.85772458e-02 4.33116369e-02 1.15122962e+00 -9.14727032e-01
-4.65404004e-01 -3.20385516e-01 -1.61748379e-01 1.88482195e-01
3.89436543e-01 -1.61177278e-01 -6.03756607e-01 -1.45600094e-02
-1.40525386e-01 -2.49597982e-01 2.76006788e-01 -3.11668932e-01
6.15855277e-01 -7.33280897e-01 -3.98608774e-01 -2.14910746e-01
1.22756720e+00 5.19504786e-01 9.62199271e-01 3.04218173e-01
4.71164435e-01 7.63990700e-01 9.70392346e-01 4.41143662e-01
7.36672163e-01 7.35291302e-01 5.85994311e-02 3.18073720e-01
2.61480421e-01 -2.18958274e-01 7.03164842e-03 9.97294366e-01
-4.90336448e-01 -9.51854289e-02 -9.87832546e-01 5.99903464e-01
-2.22542882e+00 -8.46894801e-01 -2.64805466e-01 2.22458887e+00
1.49116647e+00 1.19937487e-01 -3.16495925e-01 2.01643154e-01
9.91512418e-01 -3.39518726e-01 -5.68366885e-01 -1.71205848e-01
1.18926771e-01 -1.16781458e-01 -6.79779127e-02 8.41429949e-01
-4.13270682e-01 1.17013252e+00 5.56093788e+00 7.67041683e-01
-5.83547294e-01 -9.58836824e-02 -3.16530056e-02 5.33260524e-01
-4.10711855e-01 5.06063759e-01 -1.03459525e+00 1.50983155e-01
4.88021314e-01 -7.14500785e-01 4.44740772e-01 6.78464651e-01
1.09105296e-01 -5.14301717e-01 -1.02154684e+00 7.94470310e-01
1.35203123e-01 -1.02531934e+00 2.79179186e-01 -2.30417192e-01
4.65216130e-01 -8.91519725e-01 -6.98488593e-01 6.57406747e-01
4.12671745e-01 -5.96081495e-01 3.88488770e-01 7.62568712e-01
5.62142611e-01 -8.29145730e-01 1.14196968e+00 8.34101379e-01
-1.17642033e+00 2.72493094e-01 -4.97566134e-01 -8.26083422e-02
2.09779575e-01 5.87076128e-01 -1.25449884e+00 1.38003695e+00
4.90229756e-01 4.15461898e-01 -3.77481073e-01 9.84801590e-01
-4.79838163e-01 3.07024568e-01 -2.02848971e-01 -3.07185892e-02
-1.24316119e-01 -2.47398958e-01 4.60029989e-01 8.14458728e-01
1.48303807e-01 8.54993343e-01 2.49642521e-01 4.40080941e-01
-1.02091342e-01 2.96042651e-01 -5.05283475e-01 1.95612803e-01
9.57220912e-01 1.04272461e+00 -5.03399074e-01 -7.75282502e-01
-3.38993341e-01 8.85951877e-01 4.07395750e-01 3.85292262e-01
-5.91386735e-01 -4.83180732e-01 9.98866633e-02 -2.03769848e-01
1.21754641e-02 9.86086801e-02 3.13166440e-01 -1.21774912e+00
2.50647098e-01 -8.70056689e-01 5.25288522e-01 -1.16962337e+00
-9.26757872e-01 4.13885295e-01 3.33209246e-01 -9.48644340e-01
-1.56452909e-01 -2.32956842e-01 -3.84667009e-01 8.56112301e-01
-1.60125387e+00 -1.06796825e+00 -4.29639310e-01 8.87936354e-01
1.91177636e-01 1.99615136e-02 1.06292951e+00 8.61361399e-02
-4.08749372e-01 7.92587548e-02 -2.68654227e-01 4.54390943e-02
6.76917613e-01 -1.15078831e+00 -1.39315471e-01 6.55360937e-01
-1.23986401e-01 1.17721021e+00 8.68155241e-01 -1.22080302e+00
-1.24985814e+00 -7.56002128e-01 1.12170041e+00 -3.22046518e-01
4.64223146e-01 -1.23145501e-03 -1.36757755e+00 1.00783765e+00
1.06970869e-01 -3.75247240e-01 8.49399090e-01 4.32292193e-01
-2.56325573e-01 -1.95241243e-01 -1.14793444e+00 5.66225708e-01
8.05823028e-01 -5.44343650e-01 -1.32100928e+00 3.05702597e-01
7.94369340e-01 -4.48762447e-01 -1.04420233e+00 4.15599704e-01
4.42343861e-01 -8.05554628e-01 6.17285371e-01 -9.78454351e-01
3.63305546e-02 -6.18629873e-01 -3.24154422e-02 -1.31756806e+00
1.24168791e-01 -4.10471827e-01 -4.75063115e-01 1.54385281e+00
5.31323254e-01 -5.07856190e-01 5.72382689e-01 1.15008795e+00
-1.39854521e-01 -3.89231205e-01 -9.75156665e-01 -6.72918499e-01
-3.10790092e-01 -2.72794634e-01 1.05893242e+00 1.58917105e+00
8.15432191e-01 5.50784886e-01 -5.84188104e-01 5.04003644e-01
7.26734027e-02 2.87415087e-02 8.56208682e-01 -1.42686641e+00
2.46056542e-03 1.99064583e-01 -1.08408526e-01 -1.07064414e+00
-8.52960069e-03 -4.72038209e-01 -7.76502118e-02 -2.08438063e+00
1.63172200e-01 -6.70687735e-01 7.08021922e-03 6.72469318e-01
-4.43807632e-01 -5.52214384e-01 -1.45233288e-01 4.16507244e-01
-7.85629690e-01 5.25253415e-01 1.30004513e+00 1.32587746e-01
-4.32650268e-01 -5.74217476e-02 -8.13479602e-01 8.66225421e-01
6.56649947e-01 -1.51944950e-01 -6.45650923e-01 -9.67661217e-02
6.24295413e-01 7.08998859e-01 8.31455365e-02 -5.31228185e-01
6.58122420e-01 -3.52945268e-01 -1.06096752e-01 -4.26294535e-01
4.47627783e-01 -1.13320386e+00 3.78581762e-01 2.08371148e-01
-1.38103321e-01 -3.27638596e-01 1.79691553e-01 5.32627583e-01
-4.89645302e-01 -7.23502040e-01 4.73513734e-03 -3.71690124e-01
-1.09202337e+00 -1.15775377e-01 -1.49920270e-01 -1.36424899e-01
1.16915441e+00 -2.35054910e-01 -6.06505334e-01 -5.18248498e-01
-9.66115177e-01 5.81265926e-01 4.48896661e-02 2.66100258e-01
7.70555854e-01 -1.20483541e+00 -5.05632281e-01 -3.34312230e-01
3.85952532e-01 2.81366616e-01 3.84704798e-01 5.12909651e-01
-4.32462782e-01 4.61279154e-01 -1.59334511e-01 1.19930886e-01
-1.34924519e+00 6.40298426e-01 1.00107387e-01 -3.31007928e-01
-4.86484140e-01 4.91692156e-01 -2.20821813e-01 -4.79593396e-01
2.92329967e-01 -1.49927467e-01 -8.57647955e-01 4.47186083e-02
6.02543890e-01 2.28422895e-01 5.33552952e-02 -5.44283807e-01
-1.85946181e-01 1.41887099e-01 -3.45799267e-01 -1.60693079e-01
9.37919617e-01 -4.98970330e-01 -5.07824838e-01 7.76217520e-01
5.24491131e-01 1.09456174e-01 -3.89792204e-01 -5.72677493e-01
9.81172025e-02 -4.90172267e-01 -2.81563073e-01 -1.02491558e+00
-5.37109792e-01 2.59930819e-01 4.72362526e-02 4.78527606e-01
7.24405468e-01 7.00115338e-02 6.33142412e-01 9.68724191e-01
8.90632510e-01 -1.20656705e+00 -3.67835879e-01 6.32065773e-01
9.75678265e-01 -1.00869751e+00 1.11977644e-01 -8.99241924e-01
-9.15028930e-01 9.51536000e-01 8.01056027e-01 7.48715878e-01
5.24657190e-01 -7.92372674e-02 -1.19465411e-01 -3.08751315e-01
-8.15133214e-01 -3.15071225e-01 -3.36658619e-02 6.87116802e-01
1.70732498e-01 -4.67488468e-02 -5.71068764e-01 8.07472765e-01
-2.12360665e-01 3.95653784e-01 7.88366497e-01 1.25516415e+00
-7.63484895e-01 -1.22440159e+00 -4.70479757e-01 2.59115458e-01
-1.42469659e-01 2.62651686e-02 -5.51269650e-01 8.12274992e-01
2.31981277e-01 1.17417169e+00 -3.03035021e-01 -1.81696877e-01
5.57361722e-01 4.05886143e-01 3.00394773e-01 -7.92586327e-01
-2.27089629e-01 -3.05419832e-01 6.32070422e-01 7.36713111e-02
-6.68702245e-01 -3.09742182e-01 -1.61213291e+00 -3.70613039e-01
-9.07503247e-01 7.96560168e-01 5.29338062e-01 1.26441050e+00
2.07521673e-02 2.44852781e-01 2.75045186e-01 1.91594899e-01
-1.56915247e-01 -9.26318586e-01 -7.66233504e-01 3.88458163e-01
5.90223111e-02 -7.83401847e-01 2.94463001e-02 4.31685358e-01] | [9.487417221069336, 8.315291404724121] |
55db41ec-d5eb-49a6-9058-d2b3c18025fe | cross-lingual-cross-age-group-adaptation-for | 2306.14517 | null | https://arxiv.org/abs/2306.14517v1 | https://arxiv.org/pdf/2306.14517v1.pdf | Cross-Lingual Cross-Age Group Adaptation for Low-Resource Elderly Speech Emotion Recognition | Speech emotion recognition plays a crucial role in human-computer interactions. However, most speech emotion recognition research is biased toward English-speaking adults, which hinders its applicability to other demographic groups in different languages and age groups. In this work, we analyze the transferability of emotion recognition across three different languages--English, Mandarin Chinese, and Cantonese; and 2 different age groups--adults and the elderly. To conduct the experiment, we develop an English-Mandarin speech emotion benchmark for adults and the elderly, BiMotion, and a Cantonese speech emotion dataset, YueMotion. This study concludes that different language and age groups require specific speech features, thus making cross-lingual inference an unsuitable method. However, cross-group data augmentation is still beneficial to regularize the model, with linguistic distance being a significant influence on cross-lingual transferability. We release publicly release our code at https://github.com/HLTCHKUST/elderly_ser. | ['Pascale Fung', 'Zihan Liu', 'Rita Frieske', 'Willy Chung', 'Holy Lovenia', 'Samuel Cahyawijaya'] | 2023-06-26 | null | null | null | null | ['emotion-recognition', 'speech-emotion-recognition'] | ['computer-vision', 'speech'] | [-6.30591750e-01 -8.60810503e-02 -5.27274907e-02 -6.79938734e-01
-4.07774478e-01 -2.27129668e-01 3.00224513e-01 -9.39679742e-02
-8.82976055e-01 7.29805052e-01 4.90878612e-01 -1.77573234e-01
5.75169861e-01 -2.68296182e-01 -3.66129018e-02 -3.74878436e-01
-7.38854036e-02 2.54669264e-02 -3.97570193e-01 -2.94457823e-01
-3.06707650e-01 1.74547106e-01 -1.48782766e+00 1.77914485e-01
1.13510215e+00 9.24437881e-01 -1.10961376e-02 4.59786505e-01
-1.23143665e-01 6.56618416e-01 -7.65973747e-01 -8.12161505e-01
-2.10880458e-01 -3.33009899e-01 -7.08328843e-01 -1.81030020e-01
6.27796426e-02 -3.79947796e-02 -3.38142902e-01 9.22012389e-01
1.12064958e+00 2.33971506e-01 6.51312470e-01 -1.33765435e+00
-9.27283585e-01 7.57983625e-01 -4.26065028e-02 1.04053296e-01
4.85653341e-01 -1.08406380e-01 5.99690855e-01 -1.05769145e+00
3.20102036e-01 1.42729938e+00 7.76111186e-01 1.14340961e+00
-6.81262434e-01 -8.74888301e-01 4.79330629e-01 3.64895582e-01
-1.51496661e+00 -8.43362331e-01 9.73052979e-01 -3.58642519e-01
1.04016387e+00 1.88377082e-01 7.46879637e-01 1.87061119e+00
-7.15350285e-02 9.85592663e-01 1.47505009e+00 -2.96644658e-01
3.33299726e-01 4.75220621e-01 1.97684109e-01 3.56722116e-01
-1.52944550e-01 8.38599354e-02 -9.40439105e-01 3.10042854e-02
2.58994326e-02 -3.29012781e-01 -4.18410927e-01 2.66957074e-01
-1.11158431e+00 7.15516031e-01 -1.05340801e-01 6.37639046e-01
-4.23786789e-01 -4.11853969e-01 8.80663395e-01 7.31570423e-01
9.21366572e-01 -1.13773502e-01 -6.86595678e-01 -8.64327312e-01
-5.55291116e-01 -2.91320622e-01 1.02000821e+00 7.57575333e-01
2.67335653e-01 3.02629948e-01 3.57891530e-01 1.47988594e+00
1.86057642e-01 7.17135072e-01 7.22488284e-01 -5.62659323e-01
5.25320232e-01 2.46753633e-01 -2.92786866e-01 -6.55499220e-01
-3.57118040e-01 -7.94294011e-03 -9.33201849e-01 -5.55185638e-02
3.40325058e-01 -6.83336198e-01 -4.06599730e-01 2.09035420e+00
2.22102642e-01 -4.48371440e-01 5.31612337e-01 7.27241755e-01
1.00476360e+00 4.92007762e-01 4.63595092e-01 -4.38163728e-01
1.54536879e+00 -9.85812187e-01 -9.23815846e-01 -6.04960859e-01
6.92036152e-01 -6.58474803e-01 1.36121821e+00 4.45036769e-01
-1.05470836e+00 -5.07182658e-01 -7.28707552e-01 -1.11884885e-01
-6.97267234e-01 3.21647614e-01 7.12270558e-01 1.09661031e+00
-9.38015521e-01 4.46539000e-03 -8.38336408e-01 -5.12379229e-01
1.20739058e-01 2.45609209e-01 -5.90912819e-01 1.78963304e-01
-1.61628032e+00 1.06012142e+00 2.58681595e-01 1.06213975e-03
-2.06629306e-01 -4.67812836e-01 -9.97073114e-01 -1.83696300e-01
-1.75333709e-01 -2.22549215e-01 1.13133442e+00 -1.34867656e+00
-1.72974491e+00 9.88785982e-01 -3.01289946e-01 -8.24157968e-02
3.76886308e-01 -3.66977513e-01 -9.38911974e-01 -9.56275538e-02
-1.27595574e-01 6.46869421e-01 5.53218365e-01 -7.56953537e-01
-5.98031521e-01 -6.83025599e-01 -2.47958362e-01 3.03977996e-01
-8.34140837e-01 6.00188792e-01 -2.02324986e-01 -9.58204985e-01
-2.30685189e-01 -1.00181806e+00 2.51847953e-01 -2.10779473e-01
4.56913002e-02 -6.28441751e-01 3.22444737e-01 -1.16886675e+00
1.44950938e+00 -2.50059700e+00 9.15340334e-02 -8.40278193e-02
-1.67051598e-01 1.53627535e-02 1.69758424e-02 2.68811971e-01
-1.05557971e-01 1.72628939e-01 8.76408722e-03 -8.38788807e-01
2.12189004e-01 9.07761827e-02 4.01813030e-01 3.08999836e-01
1.76533744e-01 7.40455210e-01 -5.78502476e-01 -4.63133663e-01
-5.73318638e-02 7.78478384e-01 -4.19544876e-01 1.83911443e-01
3.71674240e-01 5.14021277e-01 -3.82442057e-01 7.12987423e-01
5.18484473e-01 4.49326187e-01 7.52827749e-02 8.65496323e-02
-2.28528962e-01 5.21244705e-01 -7.58307099e-01 1.60478210e+00
-7.81471193e-01 7.07982063e-01 4.36421245e-01 -9.47879553e-01
1.06975162e+00 5.50117552e-01 4.03263509e-01 -7.09168673e-01
5.21010518e-01 3.93618673e-01 2.53012866e-01 -6.83702052e-01
4.54469293e-01 -3.77199411e-01 -3.70733678e-01 2.20129400e-01
-8.86334926e-02 1.05139375e-01 -4.99147363e-02 -2.70842046e-01
5.56590199e-01 -3.07167202e-01 1.98510215e-01 -3.57604176e-01
4.48788971e-01 -6.66003346e-01 7.51962304e-01 2.43704304e-01
-7.28036582e-01 2.37319946e-01 2.04786211e-01 -9.98032913e-02
-3.72442722e-01 -9.73154068e-01 -6.87417090e-02 1.35345900e+00
-3.40898722e-01 -4.83140290e-01 -9.39993262e-01 -7.38133907e-01
-3.44569117e-01 8.24550390e-01 -5.19038320e-01 -3.93074900e-01
-2.62253016e-01 -5.94792008e-01 7.93392003e-01 6.63036823e-01
6.50910616e-01 -1.29872918e+00 -4.15175945e-01 3.58362049e-02
-6.58132136e-01 -1.25168967e+00 -7.13488996e-01 1.64516836e-01
-3.14010561e-01 -5.57373106e-01 -1.02322030e+00 -1.10394716e+00
2.32895941e-01 -3.14021558e-01 1.14101827e+00 -1.55739680e-01
1.25143766e-01 7.61898756e-01 -7.00546265e-01 -8.22979808e-01
-4.58294779e-01 3.28436941e-01 5.53188145e-01 -1.70436744e-02
9.84411001e-01 -5.28576553e-01 -2.65293330e-01 2.80001968e-01
-2.81741500e-01 -1.78301558e-01 1.52184844e-01 6.40158892e-01
-1.73040822e-01 -1.02323696e-01 7.10509121e-01 -2.73618877e-01
6.62937939e-01 -4.66195941e-01 1.57004669e-01 3.49603981e-01
-2.32138321e-01 -2.59458095e-01 4.67233688e-01 -8.92877698e-01
-1.20056188e+00 -2.53151417e-01 -5.92406511e-01 6.91683441e-02
-3.67402285e-01 5.34028828e-01 -6.63974643e-01 5.30119002e-01
2.93065608e-01 6.31195530e-02 1.35224387e-01 -4.26102489e-01
1.47160487e-02 1.43046260e+00 3.17241997e-01 -7.66236067e-01
7.80809000e-02 -3.93845178e-02 -9.62337017e-01 -1.46281445e+00
-1.94324419e-01 -1.10363007e-01 -5.56028545e-01 -3.72031212e-01
1.21887457e+00 -1.20340312e+00 -6.63569450e-01 1.09434307e+00
-1.20967054e+00 -6.12647831e-01 7.12178722e-02 9.77913320e-01
-2.76064336e-01 2.60483176e-01 -7.68496037e-01 -1.05514455e+00
-4.93280202e-01 -1.14974439e+00 9.31394696e-01 2.38116570e-02
-6.32818282e-01 -1.15593183e+00 -2.06173271e-01 4.26468939e-01
3.31571788e-01 -1.50395796e-01 5.99221408e-01 -7.97919512e-01
5.28404057e-01 5.02895974e-02 2.94746727e-01 9.25354242e-01
2.99925029e-01 -9.14619341e-02 -1.01326418e+00 -1.34378955e-01
8.32904577e-02 -6.71899199e-01 4.09461349e-01 3.11228633e-01
7.92797267e-01 1.02810971e-02 4.73756753e-02 2.88500786e-01
4.54574406e-01 4.10315812e-01 5.74482203e-01 7.50475675e-02
5.38308740e-01 1.04107797e+00 5.50823271e-01 4.30725098e-01
9.88016129e-01 4.80921477e-01 -4.45796400e-01 -5.35341091e-02
1.17605664e-01 5.30072823e-02 1.13317800e+00 1.54884684e+00
-1.02924988e-01 -2.48148486e-01 -9.29939151e-01 4.67863351e-01
-1.53903997e+00 -8.50689650e-01 2.02416793e-01 2.08295298e+00
1.01109064e+00 -2.72093236e-01 3.13275635e-01 1.61464334e-01
6.88782454e-01 1.01294599e-01 -3.54158223e-01 -5.91210127e-01
-3.98100466e-01 1.32570624e-01 -2.66227603e-01 3.93594176e-01
-9.85973001e-01 1.09174943e+00 5.12743807e+00 6.87931836e-01
-1.57560182e+00 2.21943676e-01 8.58105242e-01 -7.36724213e-02
-1.04274623e-01 -6.04508877e-01 -6.10410154e-01 5.94212353e-01
1.29319501e+00 -1.47028908e-01 4.56985623e-01 8.77823889e-01
3.33629191e-01 4.09323126e-02 -1.12038743e+00 1.47868836e+00
3.64169359e-01 -1.77456573e-01 -6.18176460e-01 -1.40117496e-01
1.71865225e-01 1.18162788e-01 -1.48845926e-01 6.82192147e-01
-1.16436705e-01 -7.47051716e-01 1.04075146e+00 2.70346910e-01
7.53040493e-01 -7.68171906e-01 5.30927420e-01 1.83251828e-01
-1.03365362e+00 2.59677432e-02 -6.35319948e-02 -6.54411539e-02
1.75079867e-01 2.07318172e-01 -1.35083437e-01 1.95135966e-01
1.16951311e+00 6.34943545e-01 -4.73408818e-01 7.05649257e-02
-1.87537521e-01 6.79603040e-01 -3.17089260e-01 -3.65804791e-01
-3.13459069e-01 -2.35131368e-01 2.03761801e-01 1.22129202e+00
6.28751874e-01 1.44539654e-01 8.17131177e-02 1.90875486e-01
-1.53021086e-02 4.42491055e-01 -5.42415559e-01 -3.79393786e-01
6.36042714e-01 9.07433689e-01 -4.01164383e-01 -3.47569525e-01
-9.72550988e-01 1.31739020e+00 4.64318991e-01 4.30712223e-01
-9.71480787e-01 -3.15301776e-01 1.02809680e+00 -3.38019967e-01
-2.55089439e-02 -4.38228279e-01 -1.99247897e-01 -1.49139333e+00
2.96598852e-01 -1.35482991e+00 2.97014385e-01 -5.76137424e-01
-1.56374824e+00 8.50149810e-01 -1.46618098e-01 -8.52843940e-01
-3.40215206e-01 -8.45940471e-01 -3.41710061e-01 7.82189250e-01
-1.08492875e+00 -1.07983637e+00 -1.53574511e-01 6.53980732e-01
6.98803663e-01 -3.25844467e-01 8.35521042e-01 7.58925676e-01
-7.88889229e-01 1.03949583e+00 -3.41957659e-01 3.29456031e-01
9.98965502e-01 -9.73082542e-01 2.25446552e-01 5.36993623e-01
-1.48253202e-01 7.14475453e-01 4.84682709e-01 -4.09453481e-01
-9.55732048e-01 -6.58052146e-01 1.31398928e+00 -2.95944542e-01
6.84938312e-01 -7.50724673e-01 -8.53460968e-01 5.12925744e-01
6.98584855e-01 -3.93023849e-01 9.74960327e-01 5.71424663e-01
-3.25538844e-01 5.77789955e-02 -9.23193336e-01 9.74713206e-01
1.29038572e+00 -9.44461286e-01 -4.69567090e-01 -2.98433363e-01
6.39386535e-01 1.25121057e-01 -1.32026434e+00 3.26704979e-01
7.76174068e-01 -9.95503426e-01 6.41178370e-01 -4.61959809e-01
1.10415354e-01 3.06218535e-01 -2.99920022e-01 -1.61413980e+00
7.44730756e-02 -5.31109810e-01 1.35840625e-01 1.75132251e+00
5.27237296e-01 -1.03077233e+00 2.98760682e-01 1.07210171e+00
-2.69190371e-01 -5.74603736e-01 -1.09173799e+00 -8.88407171e-01
4.55510139e-01 -8.21721613e-01 5.60471535e-01 1.19109607e+00
3.47553134e-01 4.82693732e-01 -3.22698861e-01 7.55320787e-02
-6.93849847e-02 -6.21160805e-01 7.19702184e-01 -1.20205855e+00
6.26202971e-02 -6.02973402e-01 -2.63065010e-01 -7.45029509e-01
9.61016893e-01 -8.03644836e-01 6.38475791e-02 -1.14163888e+00
-2.52848893e-01 -2.21636146e-01 -9.92658660e-02 6.48747981e-01
-3.31698358e-01 -2.51567125e-01 3.22036333e-02 -5.21148443e-01
-2.17380613e-01 1.17026746e+00 5.63441098e-01 -8.60777646e-02
-2.87794739e-01 4.18525822e-02 -6.81923270e-01 7.48489916e-01
1.22100830e+00 -1.80262521e-01 -3.93423617e-01 -3.73274535e-01
-1.84689045e-01 -1.82984516e-01 8.65244642e-02 -9.39592361e-01
-7.68870115e-02 9.46491733e-02 1.84998408e-01 9.42732673e-03
5.97712457e-01 -6.72932923e-01 -2.21143335e-01 2.18702957e-01
-1.27709076e-01 4.79940087e-01 3.24241936e-01 -1.83755964e-01
-4.01093751e-01 -7.15128556e-02 5.67853272e-01 1.55526906e-01
-7.07511663e-01 3.71767521e-01 -8.08660626e-01 6.23424590e-01
5.56014955e-01 -2.36047104e-01 -2.16542810e-01 -5.65730929e-01
-8.43936205e-01 1.11765295e-01 3.77196640e-01 1.04506171e+00
3.83432209e-01 -1.40611374e+00 -7.56314456e-01 3.56273025e-01
3.43346089e-01 -6.36086047e-01 4.14474130e-01 1.06526852e+00
7.97795057e-02 1.12194093e-02 -2.46978886e-02 -6.62325835e-03
-1.59014535e+00 3.16490203e-01 3.33805233e-01 2.68143445e-01
-1.07613645e-01 7.64627278e-01 2.10870486e-02 -7.49496758e-01
6.19952679e-01 -4.12099361e-01 -8.44655633e-02 3.65286142e-01
4.33511466e-01 4.32653069e-01 -1.29297420e-01 -9.77678001e-01
-6.57885194e-01 3.26736003e-01 1.53505728e-01 -3.97309780e-01
1.17910767e+00 -4.50869262e-01 -1.23513676e-01 9.15272176e-01
1.34796762e+00 4.33250576e-01 -6.27159476e-01 1.79138556e-02
1.10953040e-01 1.47354752e-01 -9.07288864e-02 -6.45544112e-01
-9.61673021e-01 1.03443229e+00 7.92735696e-01 1.55875877e-01
1.28029919e+00 1.17744870e-01 7.60004997e-01 2.99193740e-01
4.32768583e-01 -1.68706703e+00 -1.83494568e-01 8.17084730e-01
1.09870887e+00 -1.36020410e+00 -6.18914306e-01 -4.18391109e-01
-1.19231462e+00 6.89902842e-01 7.50381351e-01 5.81861615e-01
1.02688193e+00 2.53850967e-01 6.15253568e-01 1.50043726e-01
-5.57925344e-01 -3.29716057e-01 2.01110661e-01 7.43203163e-01
1.10819042e+00 2.85240054e-01 -4.94852901e-01 1.14252877e+00
-7.63123453e-01 -1.06563695e-01 -8.48194510e-02 8.04955959e-01
1.55620664e-01 -1.34973705e+00 -2.85594612e-01 9.64255929e-02
-6.43288314e-01 -3.69940549e-01 -5.83490252e-01 9.20241714e-01
2.24976063e-01 1.46356618e+00 6.77272305e-02 -5.33266187e-01
4.40678418e-01 7.16234744e-01 2.49953523e-01 -3.74048918e-01
-6.64632916e-01 -6.37112334e-02 6.61336660e-01 -2.10115865e-01
-5.26157618e-01 -9.29700255e-01 -1.24081969e+00 -2.88259417e-01
-2.09239945e-02 4.35722411e-01 7.08208978e-01 8.11073184e-01
4.36029792e-01 2.11938322e-01 6.21132731e-01 -5.09356081e-01
-1.83084115e-01 -1.38251114e+00 -6.10646069e-01 3.08313847e-01
2.04802111e-01 -5.76751351e-01 -6.22669399e-01 -3.40049304e-02] | [13.539072036743164, 5.833860874176025] |
74fb0b79-1bf8-4c1f-8b5e-a78f0136c1ad | revisiting-random-channel-pruning-for-neural | 2205.05676 | null | https://arxiv.org/abs/2205.05676v1 | https://arxiv.org/pdf/2205.05676v1.pdf | Revisiting Random Channel Pruning for Neural Network Compression | Channel (or 3D filter) pruning serves as an effective way to accelerate the inference of neural networks. There has been a flurry of algorithms that try to solve this practical problem, each being claimed effective in some ways. Yet, a benchmark to compare those algorithms directly is lacking, mainly due to the complexity of the algorithms and some custom settings such as the particular network configuration or training procedure. A fair benchmark is important for the further development of channel pruning. Meanwhile, recent investigations reveal that the channel configurations discovered by pruning algorithms are at least as important as the pre-trained weights. This gives channel pruning a new role, namely searching the optimal channel configuration. In this paper, we try to determine the channel configuration of the pruned models by random search. The proposed approach provides a new way to compare different methods, namely how well they behave compared with random pruning. We show that this simple strategy works quite well compared with other channel pruning methods. We also show that under this setting, there are surprisingly no clear winners among different channel importance evaluation methods, which then may tilt the research efforts into advanced channel configuration searching methods. | ['Luc van Gool', 'Radu Timofte', 'Shuhang Gu', 'Wen Li', 'Kamil Adamczewski', 'Yawei Li'] | 2022-05-11 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Revisiting_Random_Channel_Pruning_for_Neural_Network_Compression_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Revisiting_Random_Channel_Pruning_for_Neural_Network_Compression_CVPR_2022_paper.pdf | cvpr-2022-1 | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [ 3.84079725e-01 -1.85213700e-01 -5.19044757e-01 -3.00321966e-01
-3.55360895e-01 -2.12323949e-01 5.91783404e-01 2.04949722e-01
-6.32851422e-01 1.05181694e+00 -1.58640444e-01 -6.22914016e-01
-2.47368991e-01 -7.98274696e-01 -6.25794947e-01 -9.66101885e-01
-3.02335650e-01 2.65661001e-01 4.65056062e-01 -3.26090962e-01
4.77641284e-01 6.61112130e-01 -1.88231730e+00 -4.84699830e-02
4.78854507e-01 1.25014389e+00 2.92933434e-01 6.43296242e-01
-1.36607394e-01 3.25812846e-01 -7.22545266e-01 -5.34555256e-01
1.96006015e-01 -6.89821959e-01 -6.26920879e-01 -4.99095529e-01
3.67565006e-02 8.71227980e-02 -1.42008513e-01 1.26150155e+00
6.98097944e-01 -1.66340038e-01 3.97142023e-01 -8.67174029e-01
2.30697572e-01 1.07064319e+00 -2.40134329e-01 5.06898165e-01
-4.94725828e-04 -3.04069966e-01 1.03047252e+00 -2.89830893e-01
6.08204961e-01 9.58335638e-01 7.33109713e-01 6.09781742e-01
-1.11250508e+00 -7.83984005e-01 1.61331564e-01 4.44317669e-01
-1.36345184e+00 -5.66997051e-01 6.64887846e-01 9.35865846e-03
8.20857763e-01 3.98129731e-01 9.19764638e-01 1.09821510e+00
-1.97392069e-02 6.26601577e-01 1.19129968e+00 -8.67723346e-01
3.49224031e-01 4.68782783e-01 3.40907335e-01 6.01448953e-01
4.37561572e-01 2.80299544e-01 -4.90550607e-01 -1.82670131e-01
5.01319051e-01 -3.51552784e-01 -4.43047106e-01 -2.84286499e-01
-7.14544892e-01 8.18737864e-01 1.43802818e-02 6.11349821e-01
7.86344856e-02 1.79368600e-01 5.12289882e-01 6.66261256e-01
1.44068092e-01 5.94656050e-01 -6.34539068e-01 -2.85998195e-01
-8.20068538e-01 5.11458635e-01 1.00331771e+00 5.76999784e-01
5.27010083e-01 2.31649540e-02 -8.66046026e-02 8.22099864e-01
1.62192300e-01 2.83889502e-01 2.38568306e-01 -5.82721531e-01
4.31052357e-01 2.12260053e-01 -3.48056048e-01 -9.98388052e-01
-4.41908181e-01 -8.13714862e-01 -8.77747178e-01 3.42421323e-01
8.07605207e-01 -1.95823476e-01 -4.78849173e-01 1.69964623e+00
-5.29478267e-02 -5.83078749e-02 -4.17298712e-02 6.48239195e-01
3.66965383e-01 3.17121118e-01 -1.15938678e-01 -4.19633120e-01
1.07575107e+00 -4.13592488e-01 -6.33717358e-01 2.14225635e-01
4.98833477e-01 -7.73352802e-01 6.57751620e-01 7.39279628e-01
-1.07311809e+00 -3.45284194e-01 -1.31693339e+00 6.99364305e-01
-4.98073936e-01 -1.27595961e-01 1.06559241e+00 1.25454390e+00
-8.37132812e-01 9.67221439e-01 -5.07246375e-01 -1.86402261e-01
5.57978332e-01 5.00103474e-01 1.52129501e-01 9.41732675e-02
-1.43592560e+00 8.84400427e-01 5.12111306e-01 6.41780570e-02
-5.66519499e-01 -1.93938643e-01 -2.84612119e-01 -3.25179733e-02
4.69441652e-01 -4.54776019e-01 1.11365044e+00 -7.99629688e-01
-1.43887019e+00 4.69885886e-01 -3.72731686e-02 -9.37686026e-01
5.22708058e-01 1.88516408e-01 -3.88890058e-01 8.82080942e-03
-4.25595671e-01 4.57996160e-01 8.75730395e-01 -1.08113337e+00
-8.06433201e-01 -3.69341761e-01 1.25221953e-01 1.70995459e-01
-3.31628114e-01 -6.09919690e-02 -6.87790453e-01 -6.27305090e-01
9.86656472e-02 -6.85528159e-01 -6.36473536e-01 -2.21968874e-01
-6.81045473e-01 1.39977457e-02 3.58731478e-01 -1.31050567e-03
1.63112378e+00 -1.81735253e+00 -6.85888827e-02 7.91314542e-01
1.84940249e-01 3.36023420e-01 1.98379695e-01 1.81390449e-01
-1.72679289e-03 2.83620834e-01 2.54158918e-02 -9.35907066e-02
-2.41115659e-01 2.12327152e-01 -1.69766307e-01 4.52876985e-01
-2.72674263e-01 5.20264506e-01 -5.29077947e-01 -3.63870919e-01
4.91621122e-02 5.57920516e-01 -5.20845175e-01 -2.78128326e-01
-2.82476358e-02 3.55978757e-01 -5.74231386e-01 6.34850800e-01
5.51622987e-01 5.09346724e-02 1.56642005e-01 6.98041078e-03
-2.04392806e-01 3.25487196e-01 -1.26060653e+00 1.26490247e+00
-1.72548786e-01 7.34670162e-01 -1.22093990e-01 -1.43484604e+00
8.50782216e-01 1.71764046e-01 3.55413020e-01 -1.03576601e+00
4.64437872e-01 3.93460661e-01 3.29683870e-01 -2.69103676e-01
3.70146841e-01 -1.31422535e-01 1.48539498e-01 3.31024140e-01
-7.46944323e-02 2.48503610e-01 2.60995269e-01 -1.02897130e-01
1.17016137e+00 -3.11556250e-01 3.47659379e-01 -3.09715629e-01
4.16831315e-01 -1.76032186e-01 3.46425563e-01 1.27992427e+00
-2.59877909e-02 2.86943048e-01 7.33682275e-01 -3.97257566e-01
-9.24719930e-01 -6.39688671e-01 -5.65539002e-01 6.95075929e-01
3.58071059e-01 -4.61231619e-01 -7.70004213e-01 -3.92419934e-01
-1.28892243e-01 2.78399408e-01 -6.06952071e-01 -8.20761472e-02
-3.24990153e-01 -1.13010883e+00 9.74274158e-01 2.79580325e-01
5.08853495e-01 -8.89382124e-01 -9.75850284e-01 2.80007243e-01
2.04762593e-01 -8.89447331e-01 4.35988635e-01 9.03208792e-01
-1.33925343e+00 -1.13916159e+00 -8.15667808e-01 -3.80665690e-01
4.17329073e-01 9.11848247e-02 8.86166751e-01 4.50648010e-01
-2.44158268e-01 -2.31202334e-01 -6.07173622e-01 -6.02204561e-01
-4.00980592e-01 3.22173506e-01 -1.29641637e-01 -2.66995519e-01
3.24159652e-01 -6.99760854e-01 -4.33855802e-01 2.94744253e-01
-5.58751345e-01 -2.87917584e-01 9.55035627e-01 6.89848006e-01
6.04665220e-01 5.96798539e-01 3.79669517e-01 -1.25453556e+00
1.00463092e+00 -2.14813605e-01 -7.18325138e-01 8.95040259e-02
-1.03784263e+00 4.75175530e-01 7.12204039e-01 -2.82047957e-01
-6.38301075e-01 8.29720125e-03 -7.00866818e-01 -6.46607727e-02
-3.54825348e-01 3.56150389e-01 -1.57408580e-01 -5.03008246e-01
7.15786338e-01 1.76315606e-01 -5.64207137e-02 -6.98572636e-01
4.19985838e-02 3.43314528e-01 1.91394553e-01 -4.65504020e-01
6.03837848e-01 3.45820308e-01 1.78648099e-01 -1.11687171e+00
-4.26716030e-01 -3.65080595e-01 -3.96693945e-01 -3.21516171e-02
5.44505179e-01 -2.55747944e-01 -6.20303094e-01 3.62760633e-01
-1.11211443e+00 -2.18763664e-01 -1.62095308e-01 4.22101021e-01
-2.38787755e-01 4.70778972e-01 -3.29831481e-01 -1.03293347e+00
-9.65770259e-02 -1.24157441e+00 5.12122095e-01 1.36568367e-01
-4.04502265e-02 -8.33006740e-01 -1.25033990e-01 -8.68806764e-02
5.81343055e-01 -1.20388456e-02 1.25768518e+00 -5.16758502e-01
-4.97760475e-01 -2.58866966e-01 -1.20813802e-01 3.15063208e-01
-4.69369769e-01 -2.04736337e-01 -1.08151925e+00 8.55439305e-02
-2.89669801e-02 1.29599690e-01 1.26323462e+00 7.05246568e-01
1.35441351e+00 9.53811556e-02 -6.37603939e-01 5.86815298e-01
1.50456429e+00 4.81173038e-01 9.01551366e-01 3.95698190e-01
1.46355182e-01 4.10228491e-01 3.44880164e-01 4.10727352e-01
-2.84277201e-01 8.78144264e-01 5.27086437e-01 -8.43363330e-02
-4.55489904e-02 -2.32950691e-02 -9.48126540e-02 4.40875649e-01
-5.16299069e-01 -2.07950041e-01 -5.04231751e-01 -7.78049454e-02
-1.49806261e+00 -1.12822676e+00 -2.87700683e-01 2.57721829e+00
6.53989017e-01 8.52377534e-01 9.11347494e-02 9.47243631e-01
5.73433936e-01 2.25160364e-02 -2.23393530e-01 -4.83055741e-01
-3.69574815e-01 5.10296285e-01 7.37023890e-01 3.67113531e-01
-1.11935234e+00 4.80970681e-01 7.04042721e+00 1.04151654e+00
-1.26820564e+00 1.63675211e-02 6.05697691e-01 2.05978110e-01
-1.73849210e-01 2.94755958e-02 -1.20380914e+00 3.55119258e-01
8.26185286e-01 1.05833434e-01 3.24691176e-01 6.70381308e-01
3.64807174e-02 -4.35002834e-01 -8.94373715e-01 1.17218626e+00
3.78903635e-02 -1.36643851e+00 -6.91766962e-02 3.21128577e-01
4.09517020e-01 -1.32824898e-01 4.55292016e-02 1.73594862e-01
-4.06411409e-01 -9.76962984e-01 7.50163257e-01 3.31519723e-01
5.67719281e-01 -8.52014244e-01 8.54082584e-01 3.71524870e-01
-1.01556158e+00 -2.62893379e-01 -5.79307199e-01 -3.42987329e-01
-3.35198231e-02 9.06997025e-01 -6.39668584e-01 3.61457348e-01
6.30716622e-01 3.14690262e-01 -6.55317605e-01 1.58793485e+00
-1.02581307e-01 6.77583933e-01 -3.70138168e-01 -5.32452881e-01
9.15820375e-02 -1.51507719e-03 5.75318336e-01 1.15606821e+00
2.68562078e-01 -2.16020629e-01 -3.43236476e-01 5.33951163e-01
1.02899045e-01 3.12111676e-01 -7.65871644e-01 8.50838721e-02
5.08359551e-01 8.25533390e-01 -1.11957693e+00 -8.93156379e-02
-4.30406958e-01 5.43029606e-01 -1.37497904e-02 2.80952334e-01
-7.37267792e-01 -5.76248407e-01 5.08996606e-01 1.59482300e-01
3.43586683e-01 -3.53144929e-02 -4.32131946e-01 -7.22470224e-01
-1.00745447e-01 -8.35296512e-01 1.88171133e-01 -9.08121988e-02
-8.34880710e-01 6.81913555e-01 4.58719768e-02 -1.07901573e+00
3.87697667e-02 -7.17830896e-01 -3.96542400e-01 4.96340960e-01
-1.62601471e+00 -2.00026155e-01 -3.23941931e-03 5.82746089e-01
4.58405197e-01 -2.36395612e-01 8.57815564e-01 6.35023117e-01
-4.42147523e-01 9.62674081e-01 7.93314204e-02 -7.67399743e-02
3.69560540e-01 -9.40610051e-01 -8.77560601e-02 7.20634520e-01
3.28268498e-01 6.03039622e-01 9.91409481e-01 -2.78490305e-01
-1.27530515e+00 -5.18755257e-01 9.29829240e-01 -6.45075366e-02
3.98978353e-01 -2.60725975e-01 -6.37566984e-01 -1.45947039e-01
1.09015694e-02 -4.13355976e-01 6.17247581e-01 6.20954514e-01
-5.88967814e-04 -2.30597436e-01 -7.62761474e-01 6.03660226e-01
1.18170738e+00 -1.12838326e-02 -2.30637759e-01 -8.69590715e-02
1.49084732e-01 -3.44993770e-01 -4.14178640e-01 4.06157494e-01
7.93444395e-01 -1.44067585e+00 8.07016551e-01 -2.89212763e-01
1.44659266e-01 5.14695831e-02 -2.90373296e-01 -1.00131989e+00
2.36015290e-01 -7.02656269e-01 -3.48218419e-02 8.57166529e-01
1.00983906e+00 -7.20611751e-01 1.03662443e+00 2.57790107e-02
-2.00269558e-02 -1.24687171e+00 -9.94877219e-01 -8.56890798e-01
-1.77005917e-01 -1.05165017e+00 4.60537136e-01 6.40745759e-01
-1.01429343e-01 4.25092459e-01 -2.29037419e-01 -1.81963205e-01
5.24535954e-01 4.93630916e-02 5.24345517e-01 -1.59894657e+00
-5.92688441e-01 -1.06164443e+00 -6.44250751e-01 -1.29363966e+00
-1.19764216e-01 -7.35712290e-01 -1.58563942e-01 -8.42580974e-01
1.16000682e-01 -1.06775951e+00 -4.58348632e-01 2.09332928e-01
1.20817095e-01 2.73247510e-01 -8.50821137e-02 7.42398500e-02
-5.08498251e-01 6.36228770e-02 1.12767541e+00 -6.44642487e-02
-3.61481071e-01 7.67428279e-01 -9.09051538e-01 6.40404582e-01
9.73163247e-01 -5.28895080e-01 -4.34846342e-01 -1.03978202e-01
6.01341307e-01 -2.38514487e-02 2.54658788e-01 -1.28352618e+00
3.31965238e-01 3.19286466e-01 3.57826203e-01 -4.21406895e-01
2.50393897e-01 -9.08758283e-01 -6.39633089e-02 6.76782072e-01
-4.37770903e-01 -2.25823253e-01 -2.24951049e-03 6.70663476e-01
-2.20715001e-01 -6.88869417e-01 7.13444710e-01 -1.65122017e-01
-8.64858091e-01 9.42904353e-02 -4.29077357e-01 7.26092933e-03
7.34099686e-01 -7.47176766e-01 1.06811054e-01 -3.77349317e-01
-7.74410307e-01 -1.23109326e-01 1.10618010e-01 1.01469509e-01
4.14031506e-01 -8.73740196e-01 -2.18263477e-01 2.90219903e-01
-2.00362634e-02 -3.77535522e-01 -5.29547185e-02 1.04298890e+00
-3.85487914e-01 6.34774268e-01 -2.18304425e-01 -4.40948874e-01
-1.49841690e+00 2.16094315e-01 3.09612751e-01 -3.29694927e-01
-6.86101258e-01 1.02500367e+00 -4.86768842e-01 1.86560512e-01
9.49547112e-01 -4.31808263e-01 -4.33466136e-01 1.56071723e-01
5.67166805e-01 3.71281058e-01 3.15623581e-01 -2.34072089e-01
-1.63214833e-01 5.79548776e-01 6.91021979e-02 -2.23803520e-02
1.10523701e+00 7.80830532e-02 2.02705264e-01 2.92191565e-01
1.04704714e+00 9.88560617e-02 -7.63484836e-01 1.38381585e-01
-2.82823406e-02 -3.52146596e-01 9.85319093e-02 -5.50146997e-01
-1.22890747e+00 1.10577679e+00 8.37943316e-01 6.47772729e-01
1.34761322e+00 -7.96290562e-02 4.77603883e-01 6.62643611e-01
6.42998159e-01 -1.14495420e+00 -3.99912059e-01 4.96223658e-01
2.15553388e-01 -9.97344017e-01 1.39902622e-01 -5.04075050e-01
-2.70041406e-01 1.20129657e+00 9.31196362e-02 8.10729787e-02
8.90926182e-01 5.60773909e-01 -9.59765539e-02 -1.23850010e-01
-5.80558896e-01 -5.76382816e-01 -1.53889358e-01 6.69055283e-01
4.34327692e-01 -6.71996251e-02 -8.33208501e-01 4.53921467e-01
-2.04095215e-01 -2.43536025e-01 2.31819019e-01 6.50373757e-01
-8.09134185e-01 -1.59941113e+00 -2.22681537e-01 8.00985634e-01
-7.00123012e-01 -2.01975778e-01 -4.89361167e-01 1.08140945e+00
3.01609576e-01 8.03463340e-01 -4.18052197e-01 -5.85048616e-01
3.42977822e-01 -1.84733756e-02 7.14388013e-01 -2.62442172e-01
-4.49286401e-01 1.17640235e-01 2.71843910e-01 -5.53813040e-01
-3.94099534e-01 -4.92824733e-01 -1.09782982e+00 -1.56620085e-01
-7.71091938e-01 4.50552225e-01 8.25791657e-01 1.03901255e+00
-6.21133335e-02 5.76265097e-01 4.58528668e-01 -6.79415345e-01
-4.93766129e-01 -4.62061465e-01 -8.49938273e-01 -2.02221908e-02
1.68175660e-02 -8.50209534e-01 -3.12570751e-01 -4.84779805e-01] | [8.525535583496094, 3.239650249481201] |
1a593d0e-69ac-4367-a13b-64aad5a465f6 | bayesian-convolutional-neural-network-based | 2005.08460 | null | https://arxiv.org/abs/2005.08460v1 | https://arxiv.org/pdf/2005.08460v1.pdf | Bayesian convolutional neural network based MRI brain extraction on nonhuman primates | Brain extraction or skull stripping of magnetic resonance images (MRI) is an essential step in neuroimaging studies, the accuracy of which can severely affect subsequent image processing procedures. Current automatic brain extraction methods demonstrate good results on human brains, but are often far from satisfactory on nonhuman primates, which are a necessary part of neuroscience research. To overcome the challenges of brain extraction in nonhuman primates, we propose a fully-automated brain extraction pipeline combining deep Bayesian convolutional neural network (CNN) and fully connected three-dimensional (3D) conditional random field (CRF). The deep Bayesian CNN, Bayesian SegNet, is used as the core segmentation engine. As a probabilistic network, it is not only able to perform accurate high-resolution pixel-wise brain segmentation, but also capable of measuring the model uncertainty by Monte Carlo sampling with dropout in the testing stage. Then, fully connected 3D CRF is used to refine the probability result from Bayesian SegNet in the whole 3D context of the brain volume. The proposed method was evaluated with a manually brain-extracted dataset comprising T1w images of 100 nonhuman primates. Our method outperforms six popular publicly available brain extraction packages and three well-established deep learning based methods with a mean Dice coefficient of 0.985 and a mean average symmetric surface distance of 0.220 mm. A better performance against all the compared methods was verified by statistical tests (all p-values<10-4, two-sided, Bonferroni corrected). The maximum uncertainty of the model on nonhuman primate brain extraction has a mean value of 0.116 across all the 100 subjects... | ['Rasmus M. Birn', 'Mary E. Meyerand', 'Jonathan A. Oler', 'Gengyan Zhao', 'Fang Liu', 'Ned H. Kalin'] | 2020-05-18 | null | null | null | null | ['skull-stripping'] | ['medical'] | [ 6.47947341e-02 1.30182996e-01 2.98632950e-01 -6.89425230e-01
-5.17762005e-01 -1.93942249e-01 4.68865901e-01 3.22785340e-02
-1.01415420e+00 9.83382583e-01 -2.64798880e-01 -1.22200653e-01
-2.53676295e-01 -4.41273272e-01 -5.78206778e-01 -8.59369397e-01
-3.67684931e-01 9.20358062e-01 4.30661410e-01 5.91496527e-01
3.10368210e-01 8.89949679e-01 -1.04942846e+00 -2.41381377e-01
9.49856699e-01 9.25059080e-01 3.14298451e-01 3.31221789e-01
1.83388099e-01 1.88109186e-02 -3.84295613e-01 -3.73096675e-01
3.93848903e-02 3.08368169e-02 -7.18002439e-01 -1.61247104e-01
2.11961806e-01 -5.50065994e-01 8.83737355e-02 1.27331865e+00
5.97400069e-01 1.04326636e-01 1.07338083e+00 -8.83099854e-01
-1.49882004e-01 6.63306355e-01 -7.69908488e-01 4.50064749e-01
-1.92349479e-01 4.53888327e-01 3.54206711e-01 -7.37472951e-01
4.58867431e-01 1.08680010e+00 5.81642389e-01 4.10031378e-01
-1.37004697e+00 -1.00948346e+00 -4.68022197e-01 1.75707027e-01
-1.58438528e+00 -3.32983851e-01 4.64013934e-01 -1.01470470e+00
8.77109051e-01 -2.89279222e-01 8.03871989e-01 6.90701723e-01
7.64743388e-01 2.16419458e-01 1.65385246e+00 7.64139965e-02
3.82123917e-01 -1.97522998e-01 1.40238419e-01 5.40325582e-01
4.61955607e-01 7.87200704e-02 -3.86360399e-02 -2.47194022e-01
9.05265331e-01 -3.79394919e-01 8.91029164e-02 -1.64822876e-01
-1.17099023e+00 7.88603365e-01 4.81352031e-01 3.81715119e-01
-6.99512422e-01 3.36208232e-02 3.11196178e-01 -6.63293123e-01
4.17836845e-01 -3.93379405e-02 -3.90464246e-01 2.82578796e-01
-1.52033365e+00 3.36843342e-01 4.79963481e-01 5.53675354e-01
2.94982344e-01 5.88512346e-02 -2.25961015e-01 6.85882866e-01
5.89785755e-01 5.37217557e-01 5.09485364e-01 -7.25935280e-01
-7.86109939e-02 2.07612813e-01 -2.56374806e-01 -7.90150404e-01
-7.52831936e-01 -4.28888828e-01 -1.09927082e+00 4.06900853e-01
4.98455822e-01 -2.52600163e-01 -1.25518644e+00 1.57089877e+00
4.89435405e-01 -1.56092495e-01 -6.24547064e-01 9.83896613e-01
7.94863462e-01 9.48711112e-02 5.64277411e-01 -2.47254446e-01
1.56052911e+00 -3.64258230e-01 -4.69558567e-01 -1.59067750e-01
-1.29516320e-02 -5.07960439e-01 4.03679341e-01 4.61611211e-01
-9.78469074e-01 -2.80168355e-01 -1.10975134e+00 1.00613795e-01
-1.50494382e-01 4.46013063e-02 6.33997083e-01 8.20983589e-01
-1.00301719e+00 5.44164121e-01 -1.21312892e+00 6.39994591e-02
1.26975989e+00 5.85503340e-01 -6.21484995e-01 -1.01193991e-02
-8.90405238e-01 1.30297458e+00 6.44045115e-01 3.40342820e-01
-1.09476399e+00 -7.78174698e-01 -5.71395993e-01 -5.41133136e-02
1.80898886e-02 -6.39821529e-01 8.18102717e-01 -3.95086318e-01
-1.28786910e+00 1.12674069e+00 -4.61432822e-02 -6.27244115e-01
7.02460706e-01 -5.40188327e-02 3.58612463e-02 4.27739978e-01
1.27980277e-01 1.11450815e+00 8.66665423e-01 -8.82413507e-01
-4.53195758e-02 -7.71655142e-01 -6.34585500e-01 -1.79683074e-01
3.63267303e-01 4.56294805e-01 3.01164053e-02 -3.37227464e-01
3.56292665e-01 -7.06979990e-01 -3.89476448e-01 -3.36862206e-02
-4.42332089e-01 -2.61210889e-01 2.56055862e-01 -1.24241853e+00
5.57491779e-01 -1.85569072e+00 -1.46889225e-01 4.51154083e-01
5.81166029e-01 8.57815966e-02 2.90372789e-01 -6.05484128e-01
-1.28394127e-01 1.11058913e-01 -8.98463726e-01 -1.72811180e-01
-4.52275462e-02 -3.17211688e-01 3.79745215e-01 1.10467768e+00
2.40416601e-01 7.86809504e-01 -6.83106422e-01 -8.50558937e-01
2.12107316e-01 8.92095923e-01 -5.14326632e-01 -4.58746739e-02
1.22442422e-02 8.60616028e-01 -2.82604098e-01 3.13429981e-01
1.05384851e+00 1.66173041e-01 4.69843186e-02 -1.81495741e-01
-6.27755225e-02 -9.24281869e-03 -9.02826726e-01 1.49676192e+00
4.45724092e-02 5.31769753e-01 1.49743408e-01 -7.68321574e-01
8.88747633e-01 2.96290487e-01 5.29487610e-01 -4.62413073e-01
5.26088297e-01 2.65356600e-01 4.93720382e-01 -1.46875501e-01
-1.19649388e-01 -3.49827468e-01 4.34313208e-01 3.52107555e-01
5.05395889e-01 -4.31994885e-01 7.73386285e-02 5.63165024e-02
7.86016464e-01 2.81290084e-01 4.14317936e-01 -7.76605189e-01
3.48461002e-01 -2.59797394e-01 7.03072667e-01 2.97812611e-01
-5.41260421e-01 5.63835561e-01 6.92121148e-01 -1.64677352e-01
-9.98588622e-01 -1.15636957e+00 -8.24550271e-01 3.24240923e-01
-4.68234986e-01 3.05978298e-01 -1.27765131e+00 -4.39337552e-01
-1.73819661e-01 8.20601821e-01 -5.30502617e-01 2.99273252e-01
-3.67192596e-01 -1.22590458e+00 5.20814478e-01 2.45145425e-01
6.33255124e-01 -1.24614394e+00 -9.35062528e-01 1.68882370e-01
-1.45838141e-01 -1.27311087e+00 -5.45784533e-02 1.68639749e-01
-1.05900645e+00 -1.10479069e+00 -8.56254160e-01 -4.11638081e-01
7.06343830e-01 -4.35003996e-01 8.12539935e-01 -4.85762168e-04
-5.49020529e-01 -1.54224560e-01 2.21205309e-01 -3.62392753e-01
-3.45950872e-02 -2.38415431e-02 1.96448952e-01 -3.79498273e-01
6.01904333e-01 -8.24343979e-01 -7.91901648e-01 2.14730084e-01
-7.71602035e-01 -2.20595747e-02 6.14103198e-01 5.53123891e-01
7.90343106e-01 2.09902629e-01 5.26943266e-01 -5.44098616e-01
3.65368724e-01 -5.09458661e-01 -9.15360928e-01 -1.66013584e-01
-4.95888621e-01 -2.47285753e-01 1.00299962e-01 -3.89257520e-01
-1.01028252e+00 1.42001808e-01 -3.50866675e-01 -6.58427849e-02
-6.80326343e-01 3.71996105e-01 -1.01879507e-01 -1.26068546e-02
4.50674742e-01 -1.18100323e-01 1.71744041e-02 -3.32608730e-01
6.80884495e-02 5.01108825e-01 6.82152033e-01 -5.10129035e-01
4.30377185e-01 4.76779103e-01 2.83213496e-01 -8.66616070e-01
-3.51160347e-01 -8.00802335e-02 -1.34105122e+00 -2.88323253e-01
1.41438937e+00 -5.39054573e-01 -6.83939457e-01 6.49531662e-01
-1.30061305e+00 -4.59316045e-01 3.52817297e-01 9.91616726e-01
-5.82629263e-01 3.30776989e-01 -4.53772634e-01 -7.69656241e-01
-7.11092412e-01 -1.59655797e+00 5.79269409e-01 3.58896554e-01
-3.91114503e-01 -6.60171986e-01 -2.09649816e-01 5.09489596e-01
3.06431353e-01 4.26631123e-01 1.10684133e+00 -9.96512592e-01
-3.14264387e-01 -1.86306685e-01 -4.55033988e-01 3.84331256e-01
-2.74414390e-01 4.48016077e-02 -1.10283113e+00 2.29212735e-02
2.70935565e-01 -3.33666094e-02 6.90270841e-01 1.11676931e+00
1.18255723e+00 1.20725185e-01 -1.71304122e-01 4.30178493e-01
1.23700571e+00 1.42123863e-01 7.18016386e-01 2.75229453e-04
4.62487698e-01 8.74052346e-01 8.88470858e-02 2.41977140e-01
1.51771784e-01 3.22152734e-01 3.99654269e-01 2.79239565e-01
-2.08792806e-01 2.99262553e-01 -2.65368372e-02 5.31741798e-01
-1.12898238e-01 4.63832617e-01 -1.20848906e+00 5.93824863e-01
-1.20249808e+00 -8.04348707e-01 -5.26700377e-01 2.26672482e+00
9.97058094e-01 2.74524391e-01 1.64456367e-01 -1.23219796e-01
8.73540580e-01 -3.14845979e-01 -6.62007272e-01 -2.09507138e-01
1.40272439e-01 6.30403638e-01 4.47595686e-01 3.65058035e-01
-1.00854039e+00 8.17890108e-01 6.16101789e+00 7.63071537e-01
-8.91174197e-01 3.78526837e-01 8.49482536e-01 6.07118458e-02
2.34065130e-01 -2.15102449e-01 -6.76763237e-01 5.98144650e-01
9.86657500e-01 6.21305443e-02 3.67109150e-01 6.87643170e-01
2.89097279e-01 -8.18235517e-01 -9.39664602e-01 8.91218901e-01
2.18771715e-02 -8.95132065e-01 -4.70544368e-01 2.29652748e-01
5.32076895e-01 3.88282627e-01 -3.02867055e-01 -1.83841903e-02
6.08589910e-02 -1.37030852e+00 6.49398685e-01 8.22481692e-01
8.47194374e-01 -8.48719418e-01 8.76418769e-01 4.68361050e-01
-7.29584277e-01 4.14128840e-01 -3.96114230e-01 2.96697736e-01
3.14549655e-01 1.20674586e+00 -7.46593773e-01 4.17320244e-02
9.64828074e-01 -4.73904647e-02 -4.36179072e-01 1.49606776e+00
-3.86708111e-01 5.32940388e-01 -6.54995859e-01 1.92412272e-01
-3.55994217e-02 -2.74250209e-01 5.27397215e-01 1.21491408e+00
2.93770749e-02 2.87042141e-01 -3.09101760e-01 1.45758533e+00
6.57531917e-02 1.27463952e-01 -1.05920985e-01 1.83260456e-01
2.91438401e-01 1.60339785e+00 -1.50434375e+00 -2.25475118e-01
-1.17769903e-02 4.32833761e-01 1.51553974e-01 -3.30509208e-02
-8.05260718e-01 -2.74131894e-01 1.43246561e-01 1.15463816e-01
1.69479415e-01 -4.38031197e-01 -7.66717553e-01 -6.36203349e-01
-1.38136566e-01 -4.52798873e-01 -5.78098297e-02 -7.83046901e-01
-1.27817023e+00 6.76177621e-01 5.87846577e-01 -3.50576341e-01
-1.44886777e-01 -6.44445837e-01 -4.66268569e-01 1.32018375e+00
-1.24789536e+00 -7.71272123e-01 -5.17444536e-02 3.89254093e-01
1.43539727e-01 3.92776355e-02 7.60561585e-01 3.95493031e-01
-5.49684465e-01 1.17412545e-01 -4.00662750e-01 2.50887662e-01
4.55184042e-01 -9.18212712e-01 1.39240250e-01 9.12179947e-01
-4.26520765e-01 9.10114646e-01 4.69740361e-01 -1.13504279e+00
-6.13294244e-01 -7.82721460e-01 8.38804543e-01 -1.25272483e-01
5.68806052e-01 -1.95363343e-01 -1.03015757e+00 5.31551361e-01
9.52108428e-02 1.05174705e-01 6.31891549e-01 -3.23882490e-01
-1.70556456e-02 3.63073587e-01 -1.77845454e+00 2.90420383e-01
6.20424211e-01 -1.85181841e-01 -7.45820284e-01 2.50921696e-01
3.01006585e-01 -1.89169884e-01 -1.20149577e+00 6.29201651e-01
7.96925783e-01 -9.28038597e-01 9.17066395e-01 -2.64186859e-01
3.24497163e-01 -1.91680297e-01 1.55817196e-01 -1.03079927e+00
-2.03624159e-01 -1.46465093e-01 9.29726586e-02 1.07061231e+00
1.99180484e-01 -6.24678671e-01 6.13285840e-01 8.99228454e-01
-5.09298816e-02 -6.61367834e-01 -1.32741570e+00 -5.03750205e-01
5.45006216e-01 -6.40499175e-01 4.88884449e-01 6.28495276e-01
-4.50709343e-01 -6.03804998e-02 2.48796493e-01 1.03660539e-01
1.30776298e+00 -4.62480515e-01 2.79801637e-01 -1.51230431e+00
2.51553565e-01 -7.34975576e-01 -5.61716914e-01 -3.88060719e-01
4.12119091e-01 -1.03015411e+00 2.78386384e-01 -1.66766036e+00
6.39314413e-01 -2.54793048e-01 -6.62506521e-02 3.56791943e-01
1.49267256e-01 2.64584959e-01 -2.77950048e-01 7.77311027e-02
8.38960037e-02 3.42428893e-01 1.16803622e+00 -7.02469284e-03
6.76233992e-02 -1.40258282e-01 -2.65419990e-01 1.11175668e+00
9.56078291e-01 -8.75333369e-01 -5.25372773e-02 -1.66656986e-01
-2.26190269e-01 -1.69510856e-01 7.06775129e-01 -1.13132060e+00
2.10879758e-01 2.15012655e-02 9.22764659e-01 -8.90778542e-01
-3.99008542e-02 -7.52617121e-01 2.29335323e-01 4.60002661e-01
1.05881602e-01 -2.45130911e-01 1.75906390e-01 1.30913057e-03
3.04775923e-01 -5.00665307e-01 1.38590014e+00 -1.14383660e-01
-1.49766758e-01 6.04303837e-01 -5.66678762e-01 -1.20209910e-01
9.13337052e-01 -2.13231504e-01 8.00745115e-02 1.38749272e-01
-8.46959412e-01 -1.74057409e-02 1.51283117e-02 -2.34009728e-01
6.22804046e-01 -1.05101478e+00 -7.81136751e-01 6.41193762e-02
-4.41166133e-01 2.08314672e-01 4.56556499e-01 1.45788336e+00
-6.28577232e-01 4.26442742e-01 -6.75282896e-01 -7.28398144e-01
-9.83765602e-01 1.16105035e-01 4.86041069e-01 -8.78885984e-02
-6.12080216e-01 7.13460743e-01 -6.28604814e-02 -3.04703891e-01
1.26397371e-01 -4.11179155e-01 -3.30486655e-01 6.74136728e-02
4.89240646e-01 5.11048436e-01 2.57715464e-01 -1.07932055e+00
-6.11635208e-01 2.62027055e-01 -8.04258808e-02 -4.17449862e-01
1.62326670e+00 4.84872684e-02 -5.99491775e-01 4.13422883e-02
9.88073528e-01 -3.33653420e-01 -1.15691423e+00 5.09541892e-02
2.17726026e-02 -1.31390214e-01 4.89721745e-01 -8.68758917e-01
-1.20960963e+00 1.01397252e+00 9.33650851e-01 -4.11334962e-01
7.65798867e-01 -1.15599208e-01 5.84323227e-01 -1.03593282e-01
4.79464293e-01 -9.61880386e-01 -5.66259325e-01 3.96443367e-01
9.19248462e-01 -1.14949381e+00 4.48036492e-01 -1.81603417e-01
-4.95768934e-01 9.23760951e-01 5.96304655e-01 -2.81141937e-01
1.04160035e+00 2.64786601e-01 -3.62416029e-01 -4.53224868e-01
-3.70673500e-02 -4.14819308e-02 4.88125592e-01 7.92178273e-01
4.78280246e-01 2.37574160e-01 -5.86391747e-01 9.97803092e-01
-4.20096904e-01 1.76905990e-01 1.03463620e-01 5.53784966e-01
-4.25440341e-01 -5.94251931e-01 -5.27611494e-01 7.68734813e-01
-7.62935936e-01 -2.02324331e-01 6.20978102e-02 7.14184046e-01
2.29878142e-01 6.83564723e-01 1.01302296e-01 1.77346855e-01
-1.72323972e-01 2.60681152e-01 7.70070195e-01 -5.14946699e-01
-4.61721450e-01 2.03311771e-01 -2.82329142e-01 -4.53527331e-01
-4.75385189e-01 -9.04601097e-01 -1.57312679e+00 -2.71505088e-01
-2.93897957e-01 -1.44823924e-01 1.17625403e+00 1.27638805e+00
1.24336760e-02 3.72510135e-01 -1.98890623e-02 -1.17447686e+00
-1.46920443e-01 -1.42270482e+00 -6.97478354e-01 -1.24518424e-02
-6.49984628e-02 -1.09731674e+00 -3.00605386e-01 3.33064087e-02] | [14.18805980682373, -2.298013687133789] |
0aaa416c-85d1-4a6f-a7ed-bc127072b153 | bayesian-optimization-for-radio-resource | 2012.08469 | null | https://arxiv.org/abs/2012.08469v2 | https://arxiv.org/pdf/2012.08469v2.pdf | Bayesian Optimization for Radio Resource Management: Open Loop Power Control | We provide the reader with an accessible yet rigorous introduction to Bayesian optimisation with Gaussian processes (BOGP) for the purpose of solving a wide variety of radio resource management (RRM) problems. We believe that BOGP is a powerful tool that has been somewhat overlooked in RRM research, although it elegantly addresses pressing requirements for fast convergence, safe exploration, and interpretability. BOGP also provides a natural way to exploit prior knowledge during optimization. After explaining the nuts and bolts of BOGP, we delve into more advanced topics, such as the choice of the acquisition function and the optimization of dynamic performance functions. Finally, we put the theory into practice for the RRM problem of uplink open-loop power control (OLPC) in 5G cellular networks, for which BOGP is able to converge to almost optimal solutions in tens of iterations without significant performance drops during exploration. | ['Jakob Hoydis', 'Alvaro Valcarce Rial', 'Lorenzo Maggi'] | 2020-12-15 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 7.97909200e-02 1.51850462e-01 -4.02396992e-02 -9.07820165e-02
-6.17495120e-01 -3.04776728e-01 4.01595294e-01 -3.54116201e-01
-1.41292885e-01 1.21586525e+00 -8.19661394e-02 -9.14950728e-01
-8.24244261e-01 -2.98892707e-01 -9.20782760e-02 -1.19169784e+00
-5.21441638e-01 5.23297369e-01 -3.88809204e-01 -1.01604261e-01
3.03552508e-01 6.53665364e-01 -7.79164553e-01 -9.51133370e-01
4.71909076e-01 1.09163010e+00 2.18510017e-01 8.34795117e-01
3.46915573e-01 3.10145289e-01 -7.09167302e-01 -2.06200913e-01
1.51613325e-01 -4.66342032e-01 -5.98129094e-01 1.50003627e-01
-6.00849152e-01 1.48336262e-01 -2.15843335e-01 8.17663550e-01
1.03279865e+00 1.55161917e-01 6.08597696e-01 -1.14503551e+00
1.77927613e-01 5.09714365e-01 -6.23074412e-01 5.08390725e-01
-3.07921022e-01 1.75206602e-01 1.01406419e+00 -5.93820691e-01
-2.72897501e-02 1.35534108e+00 5.02805412e-01 2.50694156e-01
-1.33413804e+00 -5.19355834e-01 4.72249873e-02 -2.44447500e-01
-1.75947607e+00 -7.54847407e-01 1.44042328e-01 -8.23847353e-02
4.94650155e-01 5.64632297e-01 5.47088206e-01 4.00443256e-01
4.39836472e-01 4.31857556e-01 7.32766569e-01 -6.27581537e-01
5.70166290e-01 1.00487366e-01 9.80251729e-02 4.87016410e-01
3.56104702e-01 2.61449218e-01 -4.00471032e-01 -7.68639028e-01
1.01946592e+00 -7.81103432e-01 -6.50714040e-01 -4.17313546e-01
-1.02711689e+00 9.16177034e-01 -2.00829670e-01 -3.76633741e-02
-4.10037398e-01 5.96089363e-01 -1.63035706e-01 3.71653110e-01
3.18305075e-01 7.57818937e-01 -4.88582432e-01 -4.78551328e-01
-9.65788960e-01 2.63785154e-01 9.88351107e-01 1.11336470e+00
2.70059258e-01 2.84735650e-01 -3.17980468e-01 6.56865001e-01
1.04980397e+00 7.29431689e-01 -4.95916247e-01 -1.29048669e+00
2.42623940e-01 -6.93439722e-01 6.39474452e-01 -5.77636957e-01
-6.58567846e-01 -1.19870305e+00 -8.35126221e-01 4.27895859e-02
4.25606787e-01 -8.33573937e-01 -4.55506891e-01 1.55081415e+00
6.99909776e-02 4.75470424e-02 -3.31514440e-02 7.43087590e-01
-4.78099175e-02 6.98503852e-01 -2.25291267e-01 -1.00892878e+00
1.32146764e+00 -4.24081177e-01 -8.97273958e-01 -2.20479175e-01
2.65813619e-01 -7.59545326e-01 3.56077909e-01 5.57822824e-01
-1.37484050e+00 1.65270925e-01 -1.14006412e+00 7.68653870e-01
3.71972054e-01 -2.06666455e-01 8.20332408e-01 1.40016818e+00
-1.11768901e+00 5.37508667e-01 -5.82699835e-01 -3.61710072e-01
5.73010921e-01 5.85053146e-01 6.09225154e-01 2.60031428e-02
-9.06693578e-01 8.41177046e-01 1.48041964e-01 4.66883689e-01
-6.78137183e-01 -7.27848232e-01 -3.63975793e-01 1.85997322e-01
6.86845124e-01 -8.79711390e-01 1.44801843e+00 -2.11292818e-01
-1.98424697e+00 1.80601120e-01 -2.15919182e-01 -4.86175954e-01
3.13901424e-01 -7.91580081e-02 -3.23688924e-01 -6.19533435e-02
-2.39196137e-01 8.04256946e-02 1.02592719e+00 -1.01424873e+00
-4.61875618e-01 -1.55680969e-01 -3.40787858e-01 3.32954109e-01
2.65005738e-01 3.69392276e-01 -5.56924462e-01 -6.20022953e-01
3.13301057e-01 -9.91742432e-01 -7.05174685e-01 -5.56607962e-01
-4.52438086e-01 1.80157065e-01 3.21797639e-01 -2.86178261e-01
1.37314272e+00 -1.79852426e+00 2.45247051e-01 9.45534110e-01
1.67139992e-01 -2.21427500e-01 1.70994192e-01 3.52851659e-01
4.84792404e-02 4.43377614e-01 7.90008577e-04 -2.01119572e-01
1.78605869e-01 2.82416075e-01 -2.68337995e-01 7.41329253e-01
-2.92317808e-01 8.31868470e-01 -8.26712489e-01 -5.04525118e-02
7.12956414e-02 1.06544763e-01 -3.66574824e-01 -1.05366632e-01
-7.18706474e-02 7.30047762e-01 -6.15343630e-01 6.23592138e-01
4.74535108e-01 -4.62454319e-01 4.23601657e-01 1.81586027e-01
-1.89193383e-01 2.14549080e-01 -1.33691657e+00 1.03344452e+00
-5.73101580e-01 6.28202975e-01 5.60907483e-01 -7.21927464e-01
5.47709644e-01 5.74471653e-01 6.53675258e-01 -1.49553567e-01
4.32601959e-01 1.79964885e-01 1.32247582e-01 2.01624468e-01
3.46387625e-01 -4.77949023e-01 1.01634879e-02 5.90401053e-01
-4.16984968e-02 -3.73337805e-01 -8.18450227e-02 -8.64358768e-02
1.00840652e+00 -3.50445062e-01 5.71737766e-01 -9.02028739e-01
3.25978398e-01 -5.90571582e-01 6.09378934e-01 1.20565844e+00
-2.17192248e-01 1.78809509e-01 6.20898485e-01 -1.10715047e-01
-6.59181952e-01 -1.07116413e+00 -5.40629923e-01 8.10093164e-01
1.84246991e-02 -3.28886122e-01 -1.95540622e-01 2.37681065e-03
-2.37211630e-01 7.86044717e-01 -1.14439540e-01 2.10254669e-01
-2.55546033e-01 -1.96945560e+00 2.02174544e-01 2.77128606e-03
2.02188507e-01 -3.54834735e-01 -3.70744318e-01 5.82870126e-01
-4.63757776e-02 -1.12622774e+00 -1.98436022e-01 5.82181811e-01
-9.00120735e-01 -5.90302467e-01 -7.26060927e-01 -1.51798194e-02
2.54208595e-01 1.72305003e-01 1.26380110e+00 -2.09564358e-01
-1.27355114e-01 5.99427879e-01 1.04497842e-01 -5.95429540e-01
-3.95294577e-01 4.07925881e-02 2.40601882e-01 -1.55791789e-01
-2.59076774e-01 -7.95791268e-01 -3.13515753e-01 7.53365815e-01
7.05127642e-02 -2.65584409e-01 4.17414188e-01 6.36219442e-01
3.79354089e-01 5.74058652e-01 5.93105376e-01 -4.80765969e-01
8.17945600e-01 -5.27032793e-01 -1.09678316e+00 4.24542986e-02
-7.17829287e-01 1.52048513e-01 -4.50757444e-02 1.84333742e-01
-8.64217997e-01 -5.15316606e-01 -1.12298310e-01 -3.88233457e-03
3.78273934e-01 2.90803820e-01 -3.39906603e-01 -5.19867778e-01
4.68776107e-01 -1.53244913e-01 -9.11822617e-02 -3.30920368e-01
3.76708716e-01 4.28991795e-01 2.38808572e-01 -8.58504236e-01
8.24911356e-01 3.50878358e-01 5.94552934e-01 -1.13697410e+00
-8.41537058e-01 -4.14441139e-01 -9.84785706e-02 -1.62334397e-01
4.20238078e-01 -7.35471785e-01 -8.96382213e-01 -4.15374786e-02
-8.82112265e-01 -5.13923705e-01 -2.13467330e-01 5.71337163e-01
-1.04046297e+00 2.78875947e-01 -3.50572437e-01 -1.40022755e+00
-1.30927041e-01 -9.45047438e-01 7.11158514e-01 2.23340288e-01
-2.64968455e-01 -1.15564919e+00 -2.67651975e-01 3.91263030e-02
6.62300944e-01 -1.39142692e-01 1.06413114e+00 -1.72800258e-01
-8.38501930e-01 1.12677194e-01 -1.16313379e-02 -1.76057413e-01
-9.04556885e-02 -1.60460413e-01 -7.83495903e-01 -5.11423469e-01
3.68501127e-01 5.25718749e-01 2.69533604e-01 1.16498256e+00
1.17987013e+00 -1.56406850e-01 -6.25854552e-01 7.69760787e-01
1.39482224e+00 1.88023880e-01 8.72817576e-01 2.39057750e-01
3.04346401e-02 2.67609566e-01 7.16840386e-01 9.62434530e-01
-1.36414841e-01 8.03033710e-01 5.16299963e-01 1.75732210e-01
6.37493014e-01 5.29043555e-01 5.58272041e-02 4.14224744e-01
-2.85639197e-01 -6.76125348e-01 -7.04903722e-01 -6.31260574e-02
-1.82482135e+00 -6.41158044e-01 -1.57850102e-01 2.46296263e+00
4.21818763e-01 3.38962555e-01 4.50170264e-02 9.34592783e-02
7.14521468e-01 1.46244951e-02 -1.72265202e-01 -2.21149027e-01
5.15076332e-02 2.50007719e-01 1.20464349e+00 6.99323654e-01
-9.77133036e-01 5.01039863e-01 7.69058847e+00 9.75185812e-01
-5.43867230e-01 1.79628938e-01 8.31102014e-01 -6.86606690e-02
-5.35930209e-02 1.98835969e-01 -1.01057553e+00 3.59220713e-01
1.43927550e+00 -1.62922695e-01 6.68713391e-01 3.68453205e-01
1.06138623e+00 -5.23012698e-01 -8.18258643e-01 1.15733945e+00
-5.26891410e-01 -1.33799601e+00 -5.88220894e-01 5.20939529e-01
6.06140077e-01 9.43598673e-02 -9.12750363e-02 1.60829276e-01
3.53769779e-01 -1.27909017e+00 3.57129157e-01 6.47926271e-01
4.20965075e-01 -9.23444629e-01 8.31375778e-01 1.60373956e-01
-7.15771496e-01 -3.23795289e-01 -3.93667519e-01 -5.87736443e-02
7.66648412e-01 1.09646058e+00 -8.19979668e-01 6.83658421e-01
3.42069387e-01 1.79482356e-01 7.24620521e-02 1.65094888e+00
-7.33926222e-02 7.43277907e-01 -6.38445258e-01 3.66236828e-02
2.57059276e-01 -4.75914925e-01 1.07370627e+00 8.96109462e-01
4.93532658e-01 8.98703784e-02 -1.12217233e-01 7.35460937e-01
3.34760994e-01 -2.26836726e-02 -1.36362985e-01 2.58764606e-02
6.61710083e-01 1.18660522e+00 -6.51435494e-01 1.26299009e-01
-6.58518076e-02 2.55562067e-01 -6.51328683e-01 7.31137753e-01
-8.76245737e-01 -1.47700563e-01 9.41236377e-01 7.87249431e-02
1.80844247e-01 -6.09047115e-01 -6.44005835e-01 -5.17568290e-01
-4.79564160e-01 -6.46713197e-01 2.38108069e-01 -7.02346146e-01
-8.40266109e-01 1.56026641e-02 1.44652203e-01 -8.25409889e-01
-2.32816055e-01 -5.05311668e-01 -5.62721252e-01 1.19451153e+00
-1.44035888e+00 -1.79925963e-01 1.10218026e-01 9.92769608e-04
3.27694267e-01 -2.68989950e-01 6.38492346e-01 1.64377555e-01
-6.89309716e-01 2.14922845e-01 5.84989607e-01 -4.92261589e-01
2.75901616e-01 -1.22502935e+00 2.85191715e-01 6.80515826e-01
-1.37296811e-01 6.05091929e-01 1.26360202e+00 -1.61283478e-01
-1.35744023e+00 -7.97919214e-01 2.43493944e-01 -3.52006197e-01
7.30730534e-01 -1.04921684e-01 -2.18883410e-01 3.73043150e-01
-8.38473532e-03 -3.76107872e-01 8.33026230e-01 3.39221656e-01
6.94717944e-01 9.92428437e-02 -7.86636591e-01 8.57858956e-01
7.26116121e-01 9.50872824e-02 1.08103633e-01 4.72506106e-01
3.50954145e-01 -3.90679210e-01 -1.00828016e+00 2.79908597e-01
3.77309024e-01 -6.33110404e-01 9.66046333e-01 -1.88363329e-01
-8.32082391e-01 -4.14631158e-01 -4.12746847e-01 -1.15206182e+00
-4.45000470e-01 -2.03456426e+00 -3.39144021e-01 1.00109673e+00
4.93233562e-01 -7.92330265e-01 6.68412030e-01 5.08408964e-01
-2.57196417e-03 -8.91989470e-01 -1.01648581e+00 -8.14801335e-01
-4.99475375e-02 -7.73950040e-01 2.93911785e-01 2.92720586e-01
-7.60368109e-02 4.73056972e-01 -4.59046632e-01 6.09866858e-01
9.36551332e-01 -3.38224828e-01 6.06691837e-01 -1.27148628e+00
-6.19429708e-01 -5.41763961e-01 -1.56762660e-01 -1.48308516e+00
-2.98971474e-01 -3.91345143e-01 2.33769313e-01 -1.21250403e+00
-1.30122751e-01 -9.35539603e-01 -1.63125053e-01 -2.33697638e-01
5.90613782e-02 6.19734079e-02 5.54118194e-02 7.05637410e-02
-5.53948998e-01 5.01872480e-01 1.03752327e+00 3.72120321e-01
-3.90453041e-01 9.47506487e-01 -1.15853894e+00 6.03687227e-01
1.00622427e+00 -4.52551097e-01 -4.22228664e-01 1.44110054e-01
5.01127124e-01 3.84543598e-01 3.29863429e-01 -8.15264404e-01
-1.10770278e-01 -3.25159878e-01 1.56670138e-01 -6.09058321e-01
4.84416366e-01 -6.75815642e-01 2.06509829e-01 1.26521930e-01
1.85008734e-01 -2.86103457e-01 2.57639557e-01 8.63778532e-01
4.36021388e-01 -4.69789505e-01 1.01627409e+00 1.91921797e-02
-4.42447662e-01 2.96560943e-01 -8.25244665e-01 1.73937708e-01
7.05209076e-01 -3.62420976e-02 6.10650182e-02 -9.83919501e-01
-1.11320651e+00 7.70097852e-01 4.34249602e-02 -9.12572145e-02
1.11989506e-01 -8.80757332e-01 -5.93154609e-01 -2.14122012e-01
-5.06663620e-01 -2.68278927e-01 -6.02602325e-02 1.28436601e+00
-2.49476939e-01 8.35409701e-01 4.83407259e-01 -8.51737022e-01
-9.83798027e-01 6.10144734e-02 8.56929839e-01 -3.31656426e-01
-4.23365861e-01 8.26011002e-01 7.91095372e-04 1.82441622e-01
3.02164555e-01 -1.55185265e-02 1.52446955e-01 -3.17553341e-01
3.85250777e-01 6.07785106e-01 9.04045105e-02 -4.80754264e-02
-2.79417306e-01 3.70687813e-01 4.29168314e-01 -4.79574412e-01
1.02875924e+00 -9.12425816e-01 -2.59406388e-01 3.29185903e-01
6.31515563e-01 1.24310665e-01 -1.24129891e+00 7.41480244e-03
2.43679762e-01 -3.86463434e-01 5.87990046e-01 -6.30710900e-01
-7.33832479e-01 5.11643291e-01 5.15435100e-01 1.55593678e-01
9.65015948e-01 -6.84415847e-02 1.59709901e-01 5.45334399e-01
7.26072192e-01 -1.04909062e+00 -2.73787439e-01 6.39728725e-01
6.19549155e-01 -6.77942932e-01 4.25280929e-01 -5.86869538e-01
-1.61332026e-01 1.06342804e+00 -1.84234872e-01 4.63261425e-01
9.50131476e-01 3.06389689e-01 -3.88079762e-01 -1.76276669e-01
-8.35466504e-01 -1.92097977e-01 -5.28359152e-02 6.40067279e-01
1.51736811e-01 1.12607047e-01 -2.41167754e-01 2.50115365e-01
-1.69916242e-01 -4.26645786e-01 7.19469011e-01 6.63577080e-01
-8.44940841e-01 -1.17056441e+00 -8.04817140e-01 6.22089207e-01
-5.85466087e-01 -3.36726278e-01 4.57888454e-01 7.10785747e-01
-5.22694230e-01 1.12655056e+00 -2.11502120e-01 3.73276174e-01
-1.25354216e-01 3.07600610e-02 6.34076774e-01 -5.98877013e-01
1.53081685e-01 5.39737046e-01 5.25616765e-01 -3.97652596e-01
-9.26549137e-02 -8.96032393e-01 -8.58088911e-01 -3.68442237e-01
-5.76370478e-01 5.88895738e-01 6.80318594e-01 1.16619468e+00
1.27176225e-01 8.56984437e-01 7.33670890e-01 -7.27001011e-01
-7.21975565e-01 -6.94437563e-01 -1.02510059e+00 -1.06128812e+00
2.96623290e-01 -7.47336984e-01 -2.75736272e-01 -4.84575808e-01] | [6.351013660430908, 3.716660737991333] |
1c5931be-c089-4c2c-b36e-ab045df7e47f | rocketqav2-a-joint-training-method-for-dense | 2110.07367 | null | https://arxiv.org/abs/2110.07367v2 | https://arxiv.org/pdf/2110.07367v2.pdf | RocketQAv2: A Joint Training Method for Dense Passage Retrieval and Passage Re-ranking | In various natural language processing tasks, passage retrieval and passage re-ranking are two key procedures in finding and ranking relevant information. Since both the two procedures contribute to the final performance, it is important to jointly optimize them in order to achieve mutual improvement. In this paper, we propose a novel joint training approach for dense passage retrieval and passage re-ranking. A major contribution is that we introduce the dynamic listwise distillation, where we design a unified listwise training approach for both the retriever and the re-ranker. During the dynamic distillation, the retriever and the re-ranker can be adaptively improved according to each other's relevance information. We also propose a hybrid data augmentation strategy to construct diverse training instances for listwise training approach. Extensive experiments show the effectiveness of our approach on both MSMARCO and Natural Questions datasets. Our code is available at https://github.com/PaddlePaddle/RocketQA. | ['Ji-Rong Wen', 'Haifeng Wang', 'Hua Wu', 'Qiaoqiao She', 'Wayne Xin Zhao', 'Jing Liu', 'Yingqi Qu', 'Ruiyang Ren'] | 2021-10-14 | null | https://aclanthology.org/2021.emnlp-main.224 | https://aclanthology.org/2021.emnlp-main.224.pdf | emnlp-2021-11 | ['passage-re-ranking'] | ['natural-language-processing'] | [-7.12634549e-02 -4.34173971e-01 -1.83807239e-01 -2.66744196e-01
-1.50976527e+00 -6.23960674e-01 5.90001166e-01 5.84631085e-01
-7.82689214e-01 7.89973974e-01 7.02512324e-01 -1.71698496e-01
-3.63847166e-01 -6.45489931e-01 -5.87164938e-01 -4.52958852e-01
7.57535547e-02 6.28124297e-01 3.29546511e-01 -3.54892224e-01
4.83173996e-01 1.23566516e-01 -1.45050383e+00 4.29025710e-01
1.20418680e+00 6.49935484e-01 4.50170547e-01 9.32283163e-01
-2.27035135e-01 8.28577697e-01 -3.87215048e-01 -3.77985954e-01
1.39852241e-01 -4.23464984e-01 -1.07466614e+00 -4.66050267e-01
2.88300872e-01 -4.48298365e-01 -3.75868917e-01 7.98357010e-01
7.76188135e-01 8.30465496e-01 4.96039003e-01 -7.09531605e-01
-5.44705629e-01 8.74455512e-01 -3.30757558e-01 5.94949901e-01
5.54031253e-01 -2.18956336e-01 1.43625057e+00 -1.08842719e+00
3.35951298e-01 1.05086625e+00 2.08701551e-01 3.27484787e-01
-7.41867781e-01 -5.25186956e-01 3.13415796e-01 4.95505512e-01
-1.21662509e+00 -5.60774207e-01 7.96215415e-01 -1.27776310e-01
7.09097564e-01 3.13517272e-01 3.71591240e-01 6.30599439e-01
-2.94094950e-01 1.30375433e+00 7.69275606e-01 -5.64125299e-01
-1.32811978e-01 8.08479339e-02 4.93764669e-01 5.86999238e-01
1.29464669e-02 -3.11206520e-01 -3.99250209e-01 -3.98277014e-01
4.90499228e-01 1.23877838e-01 -4.59956020e-01 -5.52317910e-02
-1.10212803e+00 7.46317267e-01 2.72295207e-01 3.97258192e-01
-2.88653880e-01 1.64969917e-02 4.49388176e-01 4.66081321e-01
4.28426236e-01 7.57994354e-01 -4.71348673e-01 -2.77282596e-01
-8.00164759e-01 3.82692426e-01 7.90950358e-01 7.29110658e-01
6.76582694e-01 -6.84518278e-01 -7.71939635e-01 1.42701602e+00
3.20969522e-01 2.21032456e-01 6.72670245e-01 -7.43391514e-01
7.29344070e-01 5.38126409e-01 2.65127212e-01 -9.50224996e-01
-1.60428822e-01 -4.70899194e-01 -7.02115953e-01 -6.39217257e-01
1.29623428e-01 -4.61808033e-02 -6.69440627e-01 1.73275161e+00
3.39094281e-01 2.52127618e-01 8.36413130e-02 8.40411603e-01
1.10108209e+00 9.63465989e-01 1.22950807e-01 -1.78949326e-01
1.37013555e+00 -1.36506963e+00 -6.96902812e-01 -1.57387510e-01
8.00468862e-01 -1.06958985e+00 1.25142622e+00 5.70025034e-02
-1.42535305e+00 -6.56585157e-01 -7.85622597e-01 -4.93039757e-01
-7.83920512e-02 3.14534485e-01 4.11345214e-01 -1.06362164e-01
-9.44054544e-01 4.91482466e-01 -5.76602042e-01 -1.12141751e-01
1.07664533e-01 3.08528960e-01 -7.00305775e-02 -3.12706828e-01
-1.64558482e+00 6.72614038e-01 4.87828791e-01 1.42878875e-01
-7.45539963e-01 -6.15439475e-01 -6.16580486e-01 2.40046442e-01
3.80231589e-01 -7.84576416e-01 1.53884780e+00 -3.59130621e-01
-1.31190789e+00 6.33338749e-01 -3.15424591e-01 -2.14786932e-01
3.14285725e-01 -5.40255368e-01 -1.53183699e-01 1.28794476e-01
1.84545934e-01 3.85563225e-01 3.18783432e-01 -1.10599637e+00
-8.35116267e-01 -7.37311989e-02 2.77286559e-01 8.65456462e-01
-4.49496984e-01 -2.02206559e-02 -1.08821392e+00 -7.20275879e-01
1.46607041e-01 -6.50797486e-01 -2.67361641e-01 -4.79825705e-01
-2.96353638e-01 -6.60520554e-01 2.36995518e-01 -7.57311463e-01
1.61490667e+00 -1.98208714e+00 2.14574650e-01 1.26099244e-01
2.39173815e-01 5.42130530e-01 -6.11832500e-01 7.50368536e-01
2.66635448e-01 1.56834126e-01 -1.09594926e-01 -5.57173431e-01
-1.02972925e-01 1.34349950e-02 -3.21060896e-01 4.63299192e-02
-4.29215562e-03 1.00813210e+00 -1.23102355e+00 -6.27766311e-01
-1.69530008e-02 3.68561864e-01 -5.59808731e-01 5.26872396e-01
-1.07587904e-01 4.76290643e-01 -8.01291525e-01 2.60483265e-01
5.44801176e-01 -4.21117246e-01 -1.49516180e-01 -1.50730863e-01
5.81342503e-02 8.77410710e-01 -1.08284760e+00 1.81441832e+00
-5.13462365e-01 2.19988182e-01 -1.89912811e-01 -7.65151918e-01
7.10915089e-01 2.87217140e-01 3.30058336e-01 -1.03978682e+00
8.69513154e-02 2.34341085e-01 -2.54016250e-01 -5.56907058e-01
1.05338526e+00 3.13951261e-03 -7.48981908e-02 7.17398465e-01
-6.07071966e-02 1.10062875e-01 5.87014258e-01 6.73325121e-01
9.40738440e-01 -1.13440871e-01 3.01349878e-01 1.34316375e-02
9.36274529e-01 -1.23879656e-01 5.51837027e-01 1.03630841e+00
-3.36672105e-02 6.16582811e-01 3.18321943e-01 -4.42680456e-02
-8.45276475e-01 -8.61070514e-01 1.96833313e-01 1.50973129e+00
3.42108458e-01 -4.04320478e-01 -2.26687297e-01 -7.58806527e-01
-2.22314462e-01 6.48155808e-01 -4.25305992e-01 -3.26567441e-01
-8.18956971e-01 -4.62804049e-01 3.62508953e-01 3.75625074e-01
4.66739297e-01 -9.72300172e-01 1.66903362e-01 2.86798060e-01
-7.48468578e-01 -9.74775612e-01 -7.89525807e-01 -1.79794103e-01
-9.46886241e-01 -9.43389297e-01 -8.43469501e-01 -1.02231538e+00
6.19793594e-01 6.78584576e-01 1.27607250e+00 7.23841369e-01
2.35142514e-01 3.96900505e-01 -8.10217977e-01 7.45500475e-02
-3.72754276e-01 4.66189116e-01 -1.53833032e-01 -3.25096011e-01
2.07292169e-01 -4.88954782e-01 -9.69563842e-01 1.85503259e-01
-9.77757275e-01 8.29199627e-02 6.27963424e-01 8.08721244e-01
7.56897569e-01 -1.67764679e-01 7.31623828e-01 -9.47796524e-01
1.21235645e+00 -5.59883356e-01 -4.16708320e-01 5.85931242e-01
-4.55816537e-01 3.19692045e-01 5.21868587e-01 -4.08651501e-01
-1.00555253e+00 -3.23144346e-01 -4.28320497e-01 -2.88923550e-02
1.86645046e-01 8.79209995e-01 6.90903813e-02 3.43597263e-01
3.37867856e-01 2.83399165e-01 -3.85976315e-01 -7.88293958e-01
4.90181357e-01 6.01635098e-01 2.48543546e-01 -7.51580000e-01
7.84213066e-01 1.24331512e-01 -5.47210872e-01 -4.00599569e-01
-1.16700149e+00 -9.99439180e-01 -5.09186089e-01 -3.96585315e-02
3.79109681e-01 -9.96397376e-01 -5.79688668e-01 2.91675806e-01
-1.15881181e+00 -1.10656030e-01 -2.38231838e-01 5.61873615e-01
-1.04211167e-01 5.54087579e-01 -5.78426123e-01 -5.50328732e-01
-7.72908628e-01 -9.67462301e-01 9.26847816e-01 5.32627761e-01
1.24299064e-01 -1.02049577e+00 5.29832006e-01 5.56008160e-01
3.31619024e-01 -3.70083302e-01 7.47342050e-01 -1.14988971e+00
-4.75605100e-01 -3.78703892e-01 -3.01720440e-01 2.92883158e-01
1.50196388e-01 -2.93975741e-01 -5.42683482e-01 -4.59929973e-01
-1.37018412e-01 -4.81342256e-01 1.14081037e+00 3.30300480e-02
1.04522204e+00 -4.11193460e-01 -2.61199594e-01 3.14530194e-01
1.20146871e+00 -2.48394627e-02 4.87000197e-01 3.65946710e-01
6.48890138e-01 3.26664120e-01 8.50330353e-01 4.17172462e-01
6.31854475e-01 4.45473075e-01 -2.88513135e-02 5.87544180e-02
-1.23493470e-01 -3.09253007e-01 1.39138341e-01 1.51327288e+00
1.07070081e-01 -4.42767799e-01 -8.18768442e-01 7.35772073e-01
-1.85935140e+00 -9.99954402e-01 1.32232279e-01 2.41766119e+00
1.07635379e+00 -8.84355232e-02 -5.33778183e-02 -1.69782281e-01
6.59657061e-01 2.22969592e-01 -3.04868400e-01 -1.04973413e-01
1.17678449e-01 1.70853794e-01 -3.31273512e-03 9.36870635e-01
-1.02859724e+00 1.04199946e+00 5.34338856e+00 9.19660091e-01
-7.66324520e-01 1.10449634e-01 5.35036504e-01 -2.06518278e-01
-5.15810728e-01 -1.81245301e-02 -8.65704179e-01 2.69947916e-01
6.77627206e-01 -4.15187687e-01 3.59673768e-01 4.42723632e-01
1.89722300e-01 -1.38620615e-01 -8.63616884e-01 7.31868744e-01
1.76296487e-01 -1.09943032e+00 3.00658911e-01 -4.72637415e-01
6.92576408e-01 9.13567469e-02 -5.02668694e-02 6.87746108e-01
2.81246543e-01 -5.92914701e-01 2.48986423e-01 5.54353237e-01
2.94765353e-01 -8.69612932e-01 7.36058593e-01 4.55417871e-01
-1.32368410e+00 1.77727584e-02 -4.03648287e-01 1.32190630e-01
3.13323975e-01 5.75477242e-01 -7.47528970e-01 7.39984572e-01
4.25250649e-01 6.77321553e-01 -5.79320490e-01 1.33912170e+00
-5.38292587e-01 6.44622028e-01 -2.28673160e-01 -2.77634442e-01
2.27183849e-01 -1.24050811e-01 6.98111534e-01 1.22698617e+00
6.91929683e-02 4.60801005e-01 2.72066236e-01 2.73065805e-01
-3.65031153e-01 4.60615635e-01 -4.44463938e-02 -1.10377304e-01
7.91735709e-01 1.21864426e+00 -3.15650791e-01 -5.60110629e-01
-3.10468882e-01 9.74253297e-01 7.83667684e-01 3.80919725e-01
-6.36217594e-01 -5.18765152e-01 4.04725432e-01 -1.66417122e-01
1.54267281e-01 -1.75978422e-01 2.45180845e-01 -1.42493331e+00
1.47459686e-01 -9.08536971e-01 7.74019241e-01 -3.93001616e-01
-1.42498147e+00 6.63914502e-01 -5.68689518e-02 -1.25738335e+00
-2.17716396e-01 -6.36655241e-02 -5.85468650e-01 8.90487611e-01
-1.91537797e+00 -8.78043175e-01 -2.13960350e-01 5.47810376e-01
6.64585948e-01 7.75080919e-02 5.29726148e-01 6.82581902e-01
-4.75716352e-01 6.57278717e-01 3.43689203e-01 2.12737441e-01
8.26446712e-01 -1.07043779e+00 2.37348244e-01 8.98363888e-01
2.60677099e-01 9.25793588e-01 5.71696937e-01 -5.01028538e-01
-1.17543018e+00 -8.59675884e-01 1.17946792e+00 -2.64647275e-01
5.40200293e-01 -1.13844313e-01 -1.25335658e+00 4.74855691e-01
1.89938113e-01 -4.02153522e-01 8.96158040e-01 5.23969233e-01
-1.84635177e-01 -1.45501792e-01 -7.19283044e-01 6.91662133e-01
7.43347466e-01 -5.74855685e-01 -9.39271033e-01 6.27532601e-01
9.13786471e-01 -5.22158742e-01 -6.83485329e-01 5.46271801e-01
3.95419210e-01 -5.48362792e-01 1.00830138e+00 -5.99427760e-01
4.79227722e-01 -3.13572288e-01 1.48162581e-02 -1.20831287e+00
-4.05606359e-01 -4.43450093e-01 -3.90463352e-01 1.37963605e+00
5.12293756e-01 -5.29324949e-01 5.20313740e-01 5.53324640e-01
-1.03074804e-01 -1.00920272e+00 -7.08240986e-01 -4.21594054e-01
1.40650138e-01 -2.14662910e-01 5.89191377e-01 7.81484962e-01
-5.93218841e-02 6.12312376e-01 -5.05923629e-01 1.98823571e-01
2.15619743e-01 2.74075270e-01 6.97268665e-01 -1.00248849e+00
-4.63772535e-01 -4.47729677e-01 1.71301767e-01 -1.72191131e+00
-6.11025654e-02 -1.06133592e+00 3.17760676e-01 -2.00043058e+00
6.15028501e-01 -5.97994566e-01 -8.21188927e-01 3.09299946e-01
-8.55602264e-01 -1.50031924e-01 1.22544393e-01 4.88703012e-01
-1.25891805e+00 8.06926310e-01 1.38241541e+00 6.69341087e-02
-5.57877839e-01 1.35980308e-01 -1.02265203e+00 2.41480157e-01
9.27656949e-01 -6.71840668e-01 -5.85324883e-01 -9.17339087e-01
3.86620820e-01 1.61739022e-01 1.57381505e-01 -8.27246726e-01
5.55449903e-01 6.75827563e-02 -3.89871397e-03 -7.86073148e-01
3.02330732e-01 -3.98952633e-01 -4.87973273e-01 2.27109149e-01
-8.47825110e-01 3.55063498e-01 2.17072248e-01 5.63609004e-01
-3.85078877e-01 -5.07039607e-01 5.95366657e-01 -1.68093115e-01
-5.80957055e-01 5.11309922e-01 -1.43087700e-01 3.39360684e-01
4.95264024e-01 3.56918097e-01 -3.86561185e-01 -6.39962673e-01
-5.18575370e-01 7.78921306e-01 1.80635795e-01 5.53546309e-01
7.16538250e-01 -1.31247318e+00 -1.06660903e+00 -1.99675798e-01
1.02873854e-01 5.43457977e-02 4.73253995e-01 7.22983062e-01
-2.91356802e-01 5.44331551e-01 3.30622315e-01 -1.45644069e-01
-1.39023364e+00 3.46297294e-01 1.67890847e-01 -8.94893348e-01
-2.73661554e-01 1.15720356e+00 -3.46773565e-02 -5.92413068e-01
3.68696392e-01 -1.25915974e-01 -6.59296393e-01 9.60917920e-02
8.68864417e-01 2.06487551e-01 2.51161028e-02 -4.73419815e-01
-2.18394652e-01 3.72808188e-01 -7.79333532e-01 -2.94885159e-01
1.22970819e+00 -4.81468558e-01 -3.96133691e-01 3.67776603e-01
1.43535399e+00 2.22098574e-01 -5.37128031e-01 -6.75481737e-01
-2.57981252e-02 -5.41455388e-01 8.59718397e-02 -7.94423878e-01
-7.43140757e-01 6.39694035e-01 3.04726928e-01 9.35095176e-03
1.31160164e+00 6.30918220e-02 1.03824413e+00 7.59947360e-01
3.63689102e-02 -9.67305779e-01 2.74796486e-01 8.81446421e-01
8.86291265e-01 -1.31004345e+00 1.78627014e-01 -2.13542536e-01
-5.63250542e-01 7.14032650e-01 7.11845875e-01 -4.68067192e-02
4.44000542e-01 -3.61771464e-01 7.46983439e-02 -8.54284018e-02
-9.41078067e-01 -3.53608668e-01 7.57968903e-01 -6.63197562e-02
8.49195838e-01 -2.90503949e-01 -8.06085587e-01 5.45908511e-01
-2.83426046e-02 -9.01606604e-02 1.35749459e-01 1.06491423e+00
-6.70350969e-01 -1.48941457e+00 -8.57873559e-02 5.09119511e-01
-4.54122126e-01 -5.34285128e-01 -1.57457128e-01 4.87477452e-01
-3.75225008e-01 1.09184468e+00 -6.11397550e-02 -3.21551621e-01
2.44375989e-01 -3.12792510e-02 2.29662508e-01 -7.21353948e-01
-7.18256712e-01 -5.70851676e-02 1.38495669e-01 -2.50042558e-01
-1.88392967e-01 -4.98267710e-01 -1.29437315e+00 -1.21456705e-01
-5.25843203e-01 8.53025675e-01 3.30214918e-01 9.60561216e-01
5.29581964e-01 4.45153177e-01 8.31895292e-01 -3.37768197e-01
-5.10550976e-01 -1.04632950e+00 -1.65193394e-01 4.35351342e-01
3.94407213e-01 -3.09146911e-01 -3.46827567e-01 -3.98097128e-01] | [11.454859733581543, 7.708746433258057] |
91ad9afd-60dd-4c75-a260-0b03d0cd84a4 | pre-training-co-evolutionary-protein | 2110.15527 | null | https://arxiv.org/abs/2110.15527v1 | https://arxiv.org/pdf/2110.15527v1.pdf | Pre-training Co-evolutionary Protein Representation via A Pairwise Masked Language Model | Understanding protein sequences is vital and urgent for biology, healthcare, and medicine. Labeling approaches are expensive yet time-consuming, while the amount of unlabeled data is increasing quite faster than that of the labeled data due to low-cost, high-throughput sequencing methods. In order to extract knowledge from these unlabeled data, representation learning is of significant value for protein-related tasks and has great potential for helping us learn more about protein functions and structures. The key problem in the protein sequence representation learning is to capture the co-evolutionary information reflected by the inter-residue co-variation in the sequences. Instead of leveraging multiple sequence alignment as is usually done, we propose a novel method to capture this information directly by pre-training via a dedicated language model, i.e., Pairwise Masked Language Model (PMLM). In a conventional masked language model, the masked tokens are modeled by conditioning on the unmasked tokens only, but processed independently to each other. However, our proposed PMLM takes the dependency among masked tokens into consideration, i.e., the probability of a token pair is not equal to the product of the probability of the two tokens. By applying this model, the pre-trained encoder is able to generate a better representation for protein sequences. Our result shows that the proposed method can effectively capture the inter-residue correlations and improves the performance of contact prediction by up to 9% compared to the MLM baseline under the same setting. The proposed model also significantly outperforms the MSA baseline by more than 7% on the TAPE contact prediction benchmark when pre-trained on a subset of the sequence database which the MSA is generated from, revealing the potential of the sequence pre-training method to surpass MSA based methods in general. | ['Tie-Yan Liu', 'Tao Qin', 'Bin Shao', 'Pan Deng', 'Jianwei Zhu', 'Yingce Xia', 'Siyuan Liu', 'He Zhang', 'Fusong Ju', 'Huanhuan Xia', 'Lijun Wu', 'Shizhuo Zhang', 'Liang He'] | 2021-10-29 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 7.22338438e-01 1.42293513e-01 -3.93774390e-01 -4.35149878e-01
-7.94905007e-01 -5.50539434e-01 1.95077688e-01 4.44090962e-01
-5.69007695e-01 9.43510592e-01 -1.33769121e-02 -4.74807620e-01
4.32736903e-01 -5.54229319e-01 -1.12593079e+00 -9.69874024e-01
9.57689956e-02 5.95731020e-01 1.13825187e-01 -1.10156924e-01
-6.21582568e-02 1.61662653e-01 -1.32311547e+00 4.31832254e-01
1.15964949e+00 5.75064898e-01 8.05147767e-01 2.40483195e-01
-4.01075661e-01 5.29494822e-01 -3.81986231e-01 -2.66697168e-01
6.47768974e-02 -7.06639707e-01 -7.74917841e-01 1.52247092e-02
7.60916695e-02 1.12393163e-01 -6.19456619e-02 1.04653049e+00
2.25868940e-01 -1.17420852e-01 5.26103377e-01 -6.91149473e-01
-4.66524482e-01 6.63305044e-01 -6.86993659e-01 -6.60655722e-02
1.56935990e-01 1.26965627e-01 1.21784735e+00 -8.20985973e-01
7.49406338e-01 1.01482880e+00 3.05019826e-01 5.20794094e-01
-1.46379137e+00 -4.88833904e-01 1.87102377e-01 3.33504051e-01
-1.22913313e+00 -1.07611351e-01 4.45077717e-01 -4.63238001e-01
1.14749646e+00 -7.60781392e-02 2.46114939e-01 9.73671734e-01
1.84090078e-01 7.40528584e-01 9.05426264e-01 -4.84630436e-01
1.74685359e-01 -6.21455982e-02 2.61365831e-01 6.48079336e-01
9.45293903e-02 -1.47592813e-01 -4.80425090e-01 -3.53624493e-01
2.33799800e-01 2.95356721e-01 -3.05827856e-01 -5.40749490e-01
-1.21903467e+00 8.14875126e-01 2.59436399e-01 2.06785083e-01
-5.44510782e-01 -1.03687219e-01 3.32687020e-01 9.30927172e-02
3.12320173e-01 3.98324162e-01 -8.52602899e-01 -1.23633645e-01
-9.23235655e-01 3.21838893e-02 5.25637805e-01 7.31300652e-01
1.00926983e+00 -2.66017437e-01 1.03500597e-01 7.69655943e-01
1.45458400e-01 3.40741754e-01 6.14583254e-01 -1.86999395e-01
4.09676671e-01 7.61105120e-01 1.79683611e-01 -5.31871378e-01
-2.11341798e-01 -3.08833033e-01 -6.54053509e-01 -1.41937367e-03
5.80544591e-01 4.21744250e-02 -1.12592149e+00 2.08466697e+00
2.28664652e-01 2.92377830e-01 2.83786833e-01 6.85648441e-01
3.80979806e-01 8.37045431e-01 1.93955362e-01 -3.96313787e-01
1.45400989e+00 -8.26859415e-01 -6.09217703e-01 -3.04931372e-01
9.00083184e-01 -6.96703911e-01 8.99980843e-01 1.61677480e-01
-6.15121305e-01 -4.05175894e-01 -1.29553461e+00 2.03310978e-02
-9.21159312e-02 1.51487365e-01 5.34888268e-01 2.98787236e-01
-3.97793859e-01 6.99112177e-01 -9.93144214e-01 -1.90768614e-01
3.68374377e-01 5.14816225e-01 -4.90669310e-01 -4.21006829e-01
-1.18514264e+00 9.96583402e-01 7.48045981e-01 -1.80488676e-01
-5.57494104e-01 -6.54697001e-01 -8.15129042e-01 2.01821089e-01
4.61675107e-01 -1.42849192e-01 1.06545138e+00 -9.57734227e-01
-1.21500397e+00 7.17244685e-01 -7.73049533e-01 -6.19563162e-01
2.09473029e-01 -3.10589314e-01 -2.51030236e-01 9.98271182e-02
2.33218316e-02 5.66555738e-01 4.79259014e-01 -1.02207422e+00
-4.53619331e-01 -2.89924085e-01 -1.11649580e-01 1.81557424e-02
1.43962473e-01 -8.18590224e-02 -3.07696253e-01 -4.24079567e-01
-1.64903626e-02 -1.07944417e+00 -6.09284818e-01 -2.12013230e-01
-1.74236372e-01 1.43558010e-02 6.28753066e-01 -7.62711108e-01
9.33896840e-01 -2.09620595e+00 1.90244064e-01 1.64126754e-01
1.27611384e-01 6.69055939e-01 -2.97886312e-01 7.40857244e-01
-2.58113712e-01 -7.39007667e-02 -7.52192378e-01 2.02757921e-02
-3.52951735e-01 3.97331029e-01 -2.35081553e-01 3.71956229e-01
5.29883087e-01 8.63658965e-01 -1.02977788e+00 -2.36450657e-01
2.83430703e-02 5.66283762e-01 -5.00948370e-01 2.45256662e-01
-6.39014184e-01 6.67889655e-01 -2.60326713e-01 2.93444276e-01
7.70394981e-01 -7.70239651e-01 9.29396749e-01 -1.14390396e-01
1.38377279e-01 5.52584589e-01 -8.35969508e-01 1.71023321e+00
-2.86717385e-01 1.04251459e-01 -3.90559107e-01 -1.16325569e+00
1.06831944e+00 3.51831645e-01 6.28916144e-01 -4.14671212e-01
-2.58903623e-01 3.48069012e-01 4.64608967e-01 -2.19989955e-01
1.60841703e-01 -2.97567785e-01 4.28801812e-02 4.18230027e-01
7.95960501e-02 6.14600003e-01 1.73973590e-01 1.92099854e-01
1.09564865e+00 4.31548297e-01 5.31771123e-01 -2.61703674e-02
6.58937991e-01 1.12560883e-01 1.12006855e+00 2.58825928e-01
1.69158891e-01 3.33988279e-01 5.41721046e-01 -3.24613124e-01
-1.18491125e+00 -7.23711729e-01 -4.68363315e-02 9.20818806e-01
1.31496459e-01 -5.53266406e-01 -7.25888550e-01 -8.28599274e-01
9.93142091e-03 4.30081457e-01 -3.64507675e-01 -4.34187800e-01
-7.36763358e-01 -1.08807623e+00 3.48903120e-01 5.08718073e-01
8.80884454e-02 -1.00975764e+00 -5.04433155e-01 5.25259137e-01
-2.69810915e-01 -1.12274992e+00 -5.83754063e-01 7.26873398e-01
-8.16489041e-01 -1.14025247e+00 -6.21821582e-01 -9.38056111e-01
6.77602053e-01 2.35412464e-01 9.75509584e-01 -1.47307487e-02
-1.79998502e-01 -5.49811482e-01 -4.07763451e-01 -1.83408320e-01
-5.36750615e-01 1.82100609e-01 2.12174635e-02 1.86015114e-01
8.55749249e-01 -5.71885407e-01 -5.20442545e-01 2.94237018e-01
-9.93930042e-01 1.73558980e-01 9.47554827e-01 1.34375834e+00
8.57589006e-01 -2.31453761e-01 8.32945108e-01 -1.22332644e+00
1.18766434e-01 -4.68216330e-01 -5.43270469e-01 5.01988590e-01
-5.94010472e-01 6.25200689e-01 7.97714174e-01 -6.09931290e-01
-9.09831107e-01 6.83175802e-01 -1.61175266e-01 -1.47485867e-01
-4.49274369e-02 7.33588159e-01 -5.49974680e-01 3.15442383e-01
3.29198867e-01 4.05040115e-01 1.83168858e-01 -6.82047784e-01
2.72005111e-01 6.56700552e-01 3.68410081e-01 -5.28888762e-01
4.81054544e-01 2.10916936e-01 2.59724669e-02 -6.17220283e-01
-6.32387638e-01 -6.05045795e-01 -8.40876281e-01 4.72085357e-01
7.15990305e-01 -8.55744958e-01 -6.31787717e-01 2.86680311e-01
-1.12354660e+00 -9.86293033e-02 1.71971340e-02 6.40031338e-01
-3.97566020e-01 8.42162728e-01 -6.49047852e-01 -6.31535828e-01
-1.74551249e-01 -1.52831376e+00 8.38762403e-01 -1.60954833e-01
-1.85365915e-01 -6.26564860e-01 5.95106147e-02 4.14242834e-01
-1.17703313e-02 1.20474562e-01 1.56887412e+00 -1.06712210e+00
-5.09759605e-01 -6.98861107e-02 -4.80826795e-02 3.73051584e-01
4.47977453e-01 -3.60935479e-01 -7.09634602e-01 -2.98187286e-01
-1.44646913e-01 -4.07105356e-01 1.07072735e+00 6.54029027e-02
8.41408312e-01 -9.16850865e-02 -4.78011489e-01 2.11873978e-01
1.34874308e+00 3.09077948e-01 5.02580583e-01 4.17658081e-03
6.87571824e-01 5.39264619e-01 8.06432366e-01 2.58024096e-01
4.18810640e-03 8.59547317e-01 4.23327506e-01 -1.06996924e-01
3.03649604e-01 -4.56954926e-01 4.15788263e-01 7.81305194e-01
1.37519941e-01 -2.27807805e-01 -8.49565387e-01 4.32747126e-01
-2.00810575e+00 -8.83876085e-01 -2.22529341e-02 2.47894120e+00
1.24651206e+00 2.00639050e-02 -2.55916119e-02 1.56516228e-02
8.78662109e-01 1.99859850e-02 -9.06031966e-01 -2.76935339e-01
-1.86064616e-01 4.95962113e-01 6.91018939e-01 5.03904760e-01
-8.51452589e-01 9.97588873e-01 5.76655674e+00 7.00522661e-01
-1.19945085e+00 -1.57868147e-01 4.41767514e-01 1.59172341e-01
-1.91560403e-01 1.87271073e-01 -8.75895202e-01 6.78560674e-01
1.23296726e+00 -1.04519151e-01 2.71872014e-01 8.21128547e-01
1.35149613e-01 4.28681821e-02 -1.29807031e+00 6.88920379e-01
-1.67606249e-01 -1.35541201e+00 1.12298615e-01 3.60592693e-01
4.86038595e-01 1.40725940e-01 -2.24992752e-01 1.59380987e-01
4.55473453e-01 -1.15593600e+00 2.08684981e-01 1.78898454e-01
8.13783646e-01 -9.40338910e-01 9.82675433e-01 6.31085694e-01
-1.09423459e+00 3.90978485e-01 -5.81496418e-01 7.30835497e-02
2.25959405e-01 7.88633347e-01 -1.23296905e+00 6.70982659e-01
5.83462119e-02 7.32613862e-01 -1.04895577e-01 7.69886076e-01
-2.67316908e-01 8.21575284e-01 -1.78887308e-01 1.69445097e-01
1.88937694e-01 -4.24238414e-01 1.06324717e-01 1.16444397e+00
3.69484685e-02 -5.42315356e-02 5.24085701e-01 6.38341725e-01
-2.50020325e-01 2.61305809e-01 -2.59304911e-01 -4.14908111e-01
3.60412478e-01 9.86972392e-01 -5.33252776e-01 -4.27328557e-01
-6.90637231e-01 1.11229360e+00 5.26435792e-01 2.37750411e-01
-7.75713205e-01 -2.99682230e-01 8.79307508e-01 -6.08834103e-02
6.80916309e-01 -1.37532085e-01 1.55835435e-01 -1.09907842e+00
9.21167433e-02 -1.26578975e+00 1.46690995e-01 -2.02029794e-01
-1.29325390e+00 6.31508350e-01 -3.86236280e-01 -1.05393541e+00
-4.28280652e-01 -7.06717193e-01 -1.22397527e-01 1.24212730e+00
-1.63279617e+00 -9.29373085e-01 2.48614058e-01 1.46006748e-01
4.58998501e-01 -9.77569539e-03 1.10453951e+00 3.47623706e-01
-5.14562905e-01 6.31349146e-01 4.30320203e-01 4.06894684e-02
7.99405634e-01 -1.09179664e+00 5.99419594e-01 7.38941848e-01
2.11958557e-01 8.62389922e-01 6.12324655e-01 -7.22716153e-01
-1.18956423e+00 -1.14962256e+00 1.11513162e+00 -1.58475742e-01
3.21177572e-01 -5.88796258e-01 -1.47409689e+00 5.62831104e-01
-1.11591220e-01 -1.75515506e-02 1.21575189e+00 -1.70792509e-02
-5.86335778e-01 2.76455522e-01 -9.71129179e-01 3.61961544e-01
8.42052519e-01 -5.83634853e-01 -6.62656188e-01 1.86237738e-01
9.18505013e-01 -1.52066827e-01 -7.18956590e-01 5.43116152e-01
4.62473214e-01 -5.50060570e-01 7.89712787e-01 -9.69930172e-01
4.86210406e-01 -4.65170324e-01 -3.39224517e-01 -1.17863631e+00
-2.34739780e-01 -4.80987012e-01 -1.44512922e-01 9.69937027e-01
7.48564005e-01 -6.29553974e-01 9.51380193e-01 2.78389573e-01
-5.67921139e-02 -9.53005791e-01 -6.83971107e-01 -8.57940078e-01
9.92968306e-02 -2.41621789e-02 7.05159605e-01 9.03727591e-01
2.33029142e-01 5.96849382e-01 -6.09317660e-01 2.14079723e-01
2.73124069e-01 3.32360417e-01 5.57989717e-01 -1.19407797e+00
-7.54944623e-01 1.57624871e-01 -3.57592374e-01 -1.48626626e+00
3.20145786e-01 -1.14539564e+00 2.48992071e-01 -1.30777645e+00
4.62046057e-01 -3.31300527e-01 -6.24768019e-01 5.62102735e-01
-6.19335890e-01 -3.05189937e-02 1.14993583e-02 3.40433329e-01
-3.09180826e-01 4.75435257e-01 1.02792192e+00 -1.91487536e-01
-1.70520276e-01 -1.09263450e-01 -5.33133507e-01 5.24889052e-01
8.37034583e-01 -6.17369354e-01 -3.65186632e-01 -1.35161192e-03
-3.92056517e-02 8.70435014e-02 -1.09739028e-01 -6.32204831e-01
-7.92319551e-02 -2.58675456e-01 9.27961916e-02 -7.25133836e-01
3.90091687e-01 -7.92676032e-01 1.77693531e-01 6.35001004e-01
-4.49775755e-01 3.60115953e-02 -5.31607717e-02 8.16296697e-01
-2.65079170e-01 -2.83757389e-01 7.00242341e-01 -1.53580561e-01
-4.83906895e-01 2.27972224e-01 -4.04281229e-01 -6.18531108e-02
8.80890846e-01 -1.66788138e-02 -3.15501317e-02 -1.17107034e-01
-5.45317292e-01 2.90131737e-02 6.77870274e-01 2.23939374e-01
3.97694737e-01 -9.32028413e-01 -6.27846837e-01 2.58974761e-01
3.57717752e-01 -1.04930429e-02 4.03264686e-02 6.73115492e-01
-2.63737589e-01 5.80188930e-01 -1.46230713e-01 -6.26265407e-01
-1.47882152e+00 8.59981537e-01 -2.83192173e-02 -6.17820978e-01
-5.57652593e-01 5.75194895e-01 4.92658705e-01 -4.38107103e-01
3.13032344e-02 -2.10974336e-01 2.05033235e-02 -3.06706876e-01
5.28218985e-01 -7.76482299e-02 1.70457244e-01 -7.57157564e-01
-4.21665668e-01 3.56619805e-01 -6.91940486e-01 2.23736793e-01
1.40342414e+00 2.22047210e-01 -8.39364007e-02 4.06913131e-01
1.20920682e+00 8.26493800e-02 -1.28747725e+00 -6.30372643e-01
4.82509524e-01 -1.86186701e-01 -4.88124222e-01 -8.19311142e-01
-8.28604043e-01 9.30888712e-01 3.62091154e-01 -6.62144363e-01
7.38954306e-01 -3.85219209e-05 9.99746442e-01 5.41751981e-01
5.10183096e-01 -6.91475034e-01 -1.19413033e-01 4.63707119e-01
1.94157302e-01 -1.20947933e+00 -1.85124502e-01 -6.12254977e-01
-8.36566269e-01 7.99001932e-01 3.95280808e-01 8.32183361e-02
2.19239175e-01 2.35433906e-01 1.57574043e-01 6.54786080e-02
-7.89983511e-01 -1.86881170e-01 9.16692168e-02 5.56586742e-01
7.67594635e-01 8.39089975e-02 -4.92690742e-01 4.84385222e-01
1.58457443e-01 -3.61833870e-02 2.42827326e-01 1.00512314e+00
-5.67004502e-01 -1.90975106e+00 -9.91964564e-02 4.31375921e-01
-7.40624845e-01 -3.32840145e-01 -4.68121082e-01 2.26310015e-01
4.69885059e-02 9.11155760e-01 -2.08342806e-01 -3.27121973e-01
4.30478752e-02 5.28268754e-01 3.46463084e-01 -9.04279828e-01
-3.76971602e-01 7.05994144e-02 -2.91556772e-02 -2.30631530e-01
-4.14517999e-01 -5.16295910e-01 -1.68123770e+00 9.32390839e-02
-5.14167190e-01 5.24860620e-01 4.18117672e-01 9.30214882e-01
6.46158397e-01 3.83784533e-01 5.74933171e-01 -3.80574882e-01
-8.03245783e-01 -9.16681468e-01 -3.08189034e-01 5.44077575e-01
2.11622283e-01 -6.42104924e-01 7.87926316e-02 1.56746730e-01] | [4.673201084136963, 5.638831615447998] |
35d41c86-be20-49ae-8b08-41cf42d245d3 | when-cnns-meet-random-rnns-towards-multi | 2004.12349 | null | https://arxiv.org/abs/2004.12349v2 | https://arxiv.org/pdf/2004.12349v2.pdf | When CNNs Meet Random RNNs: Towards Multi-Level Analysis for RGB-D Object and Scene Recognition | Recognizing objects and scenes are two challenging but essential tasks in image understanding. In particular, the use of RGB-D sensors in handling these tasks has emerged as an important area of focus for better visual understanding. Meanwhile, deep neural networks, specifically convolutional neural networks (CNNs), have become widespread and have been applied to many visual tasks by replacing hand-crafted features with effective deep features. However, it is an open problem how to exploit deep features from a multi-layer CNN model effectively. In this paper, we propose a novel two-stage framework that extracts discriminative feature representations from multi-modal RGB-D images for object and scene recognition tasks. In the first stage, a pretrained CNN model has been employed as a backbone to extract visual features at multiple levels. The second stage maps these features into high level representations with a fully randomized structure of recursive neural networks (RNNs) efficiently. To cope with the high dimensionality of CNN activations, a random weighted pooling scheme has been proposed by extending the idea of randomness in RNNs. Multi-modal fusion has been performed through a soft voting approach by computing weights based on individual recognition confidences (i.e. SVM scores) of RGB and depth streams separately. This produces consistent class label estimation in final RGB-D classification performance. Extensive experiments verify that fully randomized structure in RNN stage encodes CNN activations to discriminative solid features successfully. Comparative experimental results on the popular Washington RGB-D Object and SUN RGB-D Scene datasets show that the proposed approach achieves superior or on-par performance compared to state-of-the-art methods both in object and scene recognition tasks. Code is available at https://github.com/acaglayan/CNN_randRNN. | ['Nevrez Imamoglu', 'Ali Caglayan', 'Ryosuke Nakamura', 'Ahmet Burak Can'] | 2020-04-26 | null | null | null | null | ['scene-recognition'] | ['computer-vision'] | [ 3.83578539e-01 -3.60554516e-01 -1.34210750e-01 -5.98388076e-01
-7.65723884e-01 -4.18952197e-01 4.79841411e-01 -2.88655668e-01
-5.54886460e-01 5.39044023e-01 -3.69693860e-02 -1.68521628e-02
-1.95353463e-01 -9.51557398e-01 -6.41136706e-01 -1.04437184e+00
4.76172507e-01 -1.20261438e-01 2.27668092e-01 1.68886945e-01
4.46901828e-01 1.10051012e+00 -2.12192798e+00 5.53614378e-01
3.51597816e-01 1.58461130e+00 2.64738500e-01 7.30352104e-01
-4.00702178e-01 9.47584093e-01 -4.16769266e-01 -4.10055891e-02
5.16179621e-01 -1.18614800e-01 -5.33114791e-01 2.02646747e-01
3.27353477e-01 -2.28883415e-01 -5.28731942e-01 9.52376902e-01
5.17459452e-01 1.06020138e-01 5.31546295e-01 -1.15151024e+00
-5.21472633e-01 6.86049163e-02 -5.74736536e-01 1.42536134e-01
7.65311792e-02 1.84281707e-01 8.52416277e-01 -9.84581888e-01
2.28206918e-01 1.02755082e+00 4.19947386e-01 4.24242020e-01
-9.40904677e-01 -5.16055703e-01 -1.89496189e-01 2.92939156e-01
-1.44075990e+00 -2.28480190e-01 9.75545585e-01 -2.88739085e-01
1.15994763e+00 2.77146578e-01 5.59474468e-01 8.18038702e-01
-2.92917136e-02 9.44264412e-01 1.30180168e+00 -3.55671942e-01
1.57697022e-01 5.06099276e-02 1.60505325e-01 7.99847782e-01
2.22409695e-01 1.36821806e-01 -4.76336896e-01 2.05012053e-01
8.51032495e-01 6.36690974e-01 -2.80221581e-01 -3.70693654e-01
-1.00610745e+00 7.52990544e-01 1.00486708e+00 4.75083083e-01
-5.82567692e-01 3.28575641e-01 2.22112328e-01 1.95915177e-02
8.81063193e-02 -2.92802807e-02 -4.61340189e-01 8.04549754e-02
-7.52938151e-01 -7.65306726e-02 4.01098222e-01 5.75803995e-01
1.01605988e+00 8.43262821e-02 -1.76520035e-01 8.83394122e-01
3.46846014e-01 6.71245873e-01 5.40280342e-01 -8.07641149e-01
3.18956971e-01 1.02586520e+00 -3.76744837e-01 -1.02896881e+00
-4.41399872e-01 -2.46873364e-01 -1.31598067e+00 5.62477350e-01
2.34006032e-01 4.00338113e-01 -1.29801023e+00 1.37916386e+00
3.10189068e-01 2.30838791e-01 2.69370168e-01 9.22587931e-01
1.21441698e+00 7.29789972e-01 -1.67818204e-01 2.53925979e-01
1.34104192e+00 -7.21234977e-01 -1.93061680e-01 2.41253735e-03
1.86526284e-01 -7.38427639e-01 7.14634836e-01 4.11923170e-01
-7.51593649e-01 -9.96571541e-01 -1.00542521e+00 -2.40344077e-01
-7.97185242e-01 6.14405096e-01 6.71171308e-01 8.05299759e-01
-9.01889086e-01 4.27153885e-01 -7.87315071e-01 -1.95737004e-01
8.51924121e-01 5.39968908e-01 -6.10115767e-01 -4.05305296e-01
-7.37800062e-01 7.00913489e-01 7.11556375e-01 5.04244864e-01
-8.69808018e-01 -3.47852185e-02 -8.98255587e-01 -1.44354165e-01
7.15196133e-02 -3.73683184e-01 7.73355544e-01 -8.60610008e-01
-1.69791007e+00 9.35973942e-01 -7.18633905e-02 -1.07391737e-01
1.01875708e-01 -3.81137617e-02 -9.77392718e-02 2.18682095e-01
-2.53584623e-01 6.84822023e-01 1.01392686e+00 -1.33862448e+00
-7.41454899e-01 -4.94637340e-01 -1.98078882e-02 1.01092055e-01
-3.02226007e-01 -2.55541652e-01 -4.35068458e-01 -3.98201674e-01
6.68171048e-01 -7.95991004e-01 -1.93443432e-01 1.42030969e-01
-5.03861248e-01 -3.37505847e-01 7.83836603e-01 -2.22877637e-01
7.12188661e-01 -2.12215853e+00 1.81992412e-01 3.77662838e-01
1.17547654e-01 4.31146562e-01 -1.28480852e-01 -1.08773440e-01
-1.71700884e-02 -9.81631204e-02 -3.84383112e-01 -4.25052106e-01
-1.75285250e-01 4.93600667e-01 -2.85957843e-01 6.04921460e-01
4.77260649e-01 1.00293446e+00 -4.16210979e-01 -3.23395640e-01
7.76439071e-01 6.72263920e-01 -2.80680984e-01 3.32305074e-01
1.50455274e-02 3.09129655e-01 -5.35216868e-01 1.09140968e+00
9.07643616e-01 -3.25743258e-01 -1.79657355e-01 -4.80660677e-01
-1.91366985e-01 -4.11166325e-02 -1.29892802e+00 1.76173651e+00
-4.27779973e-01 4.72244233e-01 -3.79642814e-01 -1.35743010e+00
1.07201397e+00 4.16980544e-03 2.32212663e-01 -7.30108917e-01
3.56433809e-01 3.02070975e-01 -3.34696501e-01 -5.97868383e-01
5.05245030e-01 -4.07270417e-02 -1.48594007e-01 2.36213103e-01
2.61073232e-01 -3.96174669e-01 -2.03663349e-01 -3.31387758e-01
9.01145697e-01 2.74800450e-01 3.56523842e-01 2.50537813e-01
7.56294668e-01 -5.79021350e-02 4.58232552e-01 7.49949455e-01
-1.63346991e-01 9.24141765e-01 1.18883207e-01 -6.53944075e-01
-8.95551801e-01 -9.73128736e-01 -1.80379674e-01 6.21964037e-01
1.21646322e-01 1.27025560e-01 -2.83322990e-01 -5.73488653e-01
1.45523638e-01 2.84895420e-01 -7.58221984e-01 -1.53160258e-03
-4.40387845e-01 -7.29617953e-01 6.97165728e-01 5.98128557e-01
1.02448463e+00 -1.18899453e+00 -1.08552408e+00 2.06131339e-02
2.24760935e-01 -1.17134213e+00 4.57321852e-01 7.40161538e-01
-9.73361075e-01 -1.07984245e+00 -7.80488968e-01 -6.28199100e-01
6.34662509e-01 5.74864686e-01 6.27451003e-01 -9.47565213e-02
-7.01328874e-01 3.36893499e-01 -5.27650535e-01 -1.22333482e-01
6.33003935e-02 -2.18874053e-03 -2.43789718e-01 1.81644931e-01
4.58029568e-01 -4.31453288e-01 -6.39841378e-01 1.42268106e-01
-1.29558718e+00 -2.09667850e-02 9.92075086e-01 8.40322375e-01
6.80330276e-01 -3.56776342e-02 1.73407674e-01 -3.45865935e-01
1.72313768e-02 -2.95084089e-01 -5.91562271e-01 3.57262433e-01
7.29358122e-02 2.34556258e-01 4.72982198e-01 -1.68849558e-01
-9.13065672e-01 4.31910723e-01 -2.86751986e-01 -6.23115957e-01
-6.30023479e-01 2.67882138e-01 -1.10414818e-01 -3.26531351e-01
3.65030229e-01 4.87300754e-01 -7.18953982e-02 -3.30761224e-01
4.51964051e-01 9.48120117e-01 4.54809248e-01 -3.33954632e-01
8.22732031e-01 6.09528184e-01 1.24701269e-01 -9.15487528e-01
-7.26037681e-01 -5.99940956e-01 -6.53820395e-01 -2.52247423e-01
1.11027980e+00 -9.77811038e-01 -8.52551281e-01 9.24956918e-01
-1.21888936e+00 -1.20353527e-01 -2.75237918e-01 4.82763559e-01
-4.02710527e-01 1.58171058e-01 -4.23666626e-01 -8.50469887e-01
-1.38272718e-01 -1.30553031e+00 1.19367576e+00 6.82915628e-01
3.52648914e-01 -5.36489308e-01 -8.66647959e-02 3.66777718e-01
5.31397045e-01 4.08354759e-01 7.57173121e-01 -4.49679434e-01
-9.92546797e-01 -3.42134237e-01 -7.17481911e-01 8.49871099e-01
2.59556741e-01 -3.11307106e-02 -1.40850914e+00 1.39915094e-01
9.48107690e-02 -7.43107498e-01 1.23607790e+00 4.22334820e-01
1.44396579e+00 1.17464766e-01 -3.45797874e-02 7.87398577e-01
1.91605592e+00 -6.31430447e-02 9.11336362e-01 2.53896207e-01
8.02671432e-01 3.45460325e-01 2.80437708e-01 6.45751357e-01
1.76056087e-01 3.07759225e-01 6.97790563e-01 -1.18279569e-01
-9.69424248e-02 1.04243025e-01 1.72925934e-01 6.61339700e-01
-2.16375247e-01 -2.08814025e-01 -7.55712152e-01 2.32719436e-01
-1.71803069e+00 -8.10456455e-01 6.00544550e-02 1.99347031e+00
5.66419542e-01 1.10326134e-01 -3.59950423e-01 4.61259305e-01
6.65037334e-01 2.42578238e-01 -4.85658973e-01 -2.15205476e-01
-5.51678121e-01 5.60054958e-01 6.36373937e-01 2.53989287e-02
-1.15671730e+00 7.91307688e-01 4.43855619e+00 9.50929880e-01
-1.36294079e+00 -7.68964216e-02 6.85699522e-01 4.80735302e-02
-1.79535721e-03 -2.50817418e-01 -8.23014796e-01 1.51575059e-01
4.05075163e-01 6.09461725e-01 1.67269886e-01 8.79675984e-01
-2.28831515e-01 -3.79865646e-01 -7.50297248e-01 1.42124343e+00
2.26301298e-01 -1.34729099e+00 8.63874704e-02 -8.17799866e-02
8.07460606e-01 3.13546419e-01 1.28830403e-01 9.80735719e-02
1.28541186e-01 -1.10220945e+00 6.60667658e-01 7.23805249e-01
7.63161719e-01 -6.61956608e-01 1.00326037e+00 1.41800210e-01
-1.37512755e+00 -2.83533037e-01 -6.30419075e-01 5.58106154e-02
-1.32427871e-01 7.97573090e-01 -4.27413046e-01 6.86831653e-01
9.83095169e-01 8.03943098e-01 -5.61396897e-01 1.03248310e+00
-3.01077455e-01 1.47490114e-01 -5.35038948e-01 -1.27085879e-01
3.71571600e-01 7.52806067e-02 7.09582865e-02 1.07021856e+00
4.97724771e-01 3.07848066e-01 -1.35292768e-01 8.49239528e-01
-2.31382683e-01 -1.82080314e-01 -7.89631069e-01 3.57243582e-03
1.10541329e-01 1.60103750e+00 -9.36040163e-01 -2.69999534e-01
-4.06484276e-01 1.04785573e+00 2.98078030e-01 2.58558929e-01
-6.93129361e-01 -4.91449505e-01 6.45398140e-01 -5.29395163e-01
7.16355681e-01 -3.64526480e-01 -3.94568652e-01 -1.06369126e+00
1.41143920e-02 -5.94541669e-01 2.58670568e-01 -7.65563905e-01
-1.13222563e+00 7.82437921e-01 -3.52541268e-01 -1.41402352e+00
6.65729120e-02 -1.07923722e+00 -3.55250716e-01 7.82040179e-01
-1.86053765e+00 -1.32337391e+00 -8.69308233e-01 9.38872933e-01
3.08756799e-01 -1.17768101e-01 7.57207513e-01 1.47183716e-01
-5.79182804e-01 2.48915598e-01 1.02543913e-01 2.95393139e-01
1.87291920e-01 -8.86679351e-01 -4.36126031e-02 7.48325765e-01
2.21991196e-01 2.84424335e-01 7.00243264e-02 -2.30112016e-01
-1.52529824e+00 -1.06953955e+00 4.95638967e-01 -1.02788776e-01
1.70230404e-01 -2.32324153e-01 -6.81695700e-01 3.20103705e-01
-2.95751058e-02 5.35080194e-01 7.00281858e-01 -5.52006006e-01
-4.95581210e-01 -3.10836524e-01 -1.23613572e+00 5.05671352e-02
7.82453954e-01 -7.47756779e-01 -5.17738640e-01 -1.14931576e-02
4.35084790e-01 -3.30169559e-01 -6.59086049e-01 6.77311182e-01
5.68876863e-01 -1.28307569e+00 1.10756540e+00 -2.04663351e-01
4.87585902e-01 -6.53482199e-01 -7.78032541e-01 -7.81416535e-01
2.80954987e-02 5.83917648e-02 7.29134008e-02 9.96157706e-01
9.15314853e-02 -5.89975893e-01 8.23370457e-01 3.98213863e-01
-4.54577468e-02 -7.93594003e-01 -1.12765610e+00 -4.51603264e-01
-3.72183800e-01 -7.67720342e-01 2.88964510e-01 5.89255035e-01
-7.56373286e-01 -2.09473416e-01 -1.52145728e-01 2.52601385e-01
6.98949099e-01 3.63935322e-01 7.42364168e-01 -1.08943188e+00
-9.97162387e-02 -5.82393050e-01 -9.09376740e-01 -1.00521541e+00
1.36010990e-01 -9.59893346e-01 1.26500472e-01 -1.46743977e+00
2.13174179e-01 -5.89286327e-01 -6.40255272e-01 7.39823401e-01
1.29278824e-01 7.63762414e-01 2.27937132e-01 2.06850991e-01
-5.56277275e-01 7.67629147e-01 9.11144197e-01 -2.23905548e-01
-1.34834766e-01 2.73601133e-02 -3.39846939e-01 6.27736509e-01
8.88077021e-01 -4.60610777e-01 -1.76102028e-03 -3.64920110e-01
4.56703529e-02 -2.61312872e-01 8.37832332e-01 -1.32543910e+00
4.69231546e-01 1.19909393e-02 8.47317636e-01 -7.89569736e-01
5.08401215e-01 -9.12576973e-01 2.97663044e-02 4.98794198e-01
-1.23560719e-01 -1.77583769e-01 3.50943238e-01 4.29907084e-01
-4.48545992e-01 -1.45794824e-01 8.71699274e-01 -2.74471253e-01
-9.84436750e-01 2.72583097e-01 -8.67435783e-02 -4.19968694e-01
1.09069061e+00 -6.27872288e-01 -2.10021436e-01 1.56966105e-01
-4.70228493e-01 -2.80249149e-01 1.37664542e-01 3.85543704e-01
1.20725441e+00 -1.35598648e+00 -4.60206330e-01 4.86493915e-01
2.08299577e-01 3.22627395e-01 4.83968794e-01 5.71556687e-01
-5.76977730e-01 4.66296047e-01 -5.11788428e-01 -1.00460351e+00
-1.17158651e+00 1.87098831e-01 2.93162197e-01 -9.41306502e-02
-2.04369619e-01 1.13753009e+00 -1.20036386e-01 -4.98870879e-01
3.12912703e-01 -7.05376565e-01 -1.78766608e-01 1.55275360e-01
4.14948553e-01 1.90210208e-01 1.39787778e-01 -7.89565086e-01
-4.76318359e-01 1.04815745e+00 1.66301474e-01 2.13222563e-01
1.62332928e+00 -3.88157479e-02 -3.14152956e-01 3.50290060e-01
1.58192801e+00 -5.24724483e-01 -1.27447200e+00 -4.24975812e-01
-8.95172358e-02 -4.74543154e-01 1.97807610e-01 -5.31979084e-01
-1.44173503e+00 1.08961654e+00 8.54559183e-01 6.31926954e-03
1.49360764e+00 -1.96643709e-03 5.69443107e-01 6.13822222e-01
3.76991868e-01 -7.70245492e-01 3.57759863e-01 4.35657084e-01
7.05801427e-01 -1.51387239e+00 4.32551540e-02 4.66691069e-02
-4.12453264e-01 1.51274240e+00 4.94827092e-01 -3.61970901e-01
7.39767909e-01 1.54878005e-01 7.02494159e-02 -1.40192598e-01
-2.39281043e-01 -7.14636683e-01 2.95570493e-01 3.07989269e-01
9.27023068e-02 -1.85916275e-02 4.17515993e-01 3.51727843e-01
2.74593830e-01 1.73373669e-01 1.07929513e-01 1.16676033e+00
-5.58381617e-01 -1.00794959e+00 -3.94071877e-01 4.16519642e-01
-2.67326713e-01 -3.23192887e-02 1.13993930e-02 6.57739699e-01
3.87894332e-01 7.34902143e-01 -5.31358197e-02 -6.82628930e-01
1.97059065e-01 -2.23841649e-02 5.61010480e-01 -3.86086017e-01
-5.47413170e-01 -2.74394065e-01 -4.46664661e-01 -8.00187349e-01
-9.27508712e-01 -4.31837618e-01 -1.20594954e+00 -5.53163365e-02
-4.32752758e-01 -3.34309429e-01 1.06551230e+00 8.89572084e-01
1.59854561e-01 5.72943807e-01 7.90627718e-01 -1.28625107e+00
-3.20783019e-01 -7.41294026e-01 -4.69155043e-01 2.07464322e-01
4.52530414e-01 -6.27892733e-01 -4.17479813e-01 -1.59610063e-01] | [9.633014678955078, -0.8843486905097961] |
dae8d5a9-072e-482f-ad33-f17cfd0c7779 | image-edge-restoring-filter | 2112.13540 | null | https://arxiv.org/abs/2112.13540v1 | https://arxiv.org/pdf/2112.13540v1.pdf | Image Edge Restoring Filter | In computer vision, image processing and computer graphics, image smoothing filtering is a very basic and important task and to be expected possessing good edge-preserving smoothing property. Here we address the problem that the edge-preserving ability of many popular local smoothing filters needs to be improved. In this paper, we propose the image Edge Restoring Filter (ERF) to restore the blur edge pixels in the output of local smoothing filters to be clear. The proposed filter can been implemented after many local smoothing filter (such as Box filter, Gaussian filter, Bilateral Filter, Guided Filter and so on). The combinations of "original local smoothing filters + ERF" have better edge-preserving smoothing property than the original local smoothing filters. Experiments on image smoothing, image denoising and image enhancement demonstrate the excellent edges restoring ability of the proposed filter and good edgepreserving smoothing property of the combination "original local smoothing filters + ERF". The proposed filter would benefit a great variety of applications given that smoothing filtering is a high frequently used and fundamental operation. | ['Zhihang Wang', 'Yongpeng Li', 'Qian Liu'] | 2021-12-27 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 1.55256823e-01 -3.57595742e-01 4.11687940e-01 -2.54170120e-01
6.18585013e-02 -1.35356128e-01 5.05807638e-01 -1.29039630e-01
-5.25105119e-01 6.94487751e-01 5.88100433e-01 -2.14193821e-01
-1.12717681e-01 -7.76303172e-01 -2.98969328e-01 -7.86920428e-01
1.07268706e-01 -7.47796655e-01 8.90483260e-01 -2.69251555e-01
5.81961155e-01 6.85520887e-01 -1.58686638e+00 2.26736471e-01
1.17786372e+00 6.89433873e-01 5.85008383e-01 7.85281181e-01
-2.95634568e-01 3.00616890e-01 -2.58773535e-01 1.96117818e-01
2.83756971e-01 -4.51136023e-01 -4.60262567e-01 4.84317124e-01
4.12184238e-01 -4.19139326e-01 -2.81393584e-02 1.63385427e+00
5.37552357e-01 4.60863322e-01 6.38844073e-01 -7.05301881e-01
-8.51794600e-01 7.04781264e-02 -8.97287607e-01 5.19111991e-01
5.48871718e-02 -6.97734579e-02 9.98549238e-02 -8.47709179e-01
6.43673003e-01 1.51215947e+00 7.93975830e-01 1.54488087e-01
-1.06858182e+00 -3.86321425e-01 -9.88166779e-02 2.63714254e-01
-1.26011276e+00 -1.80889174e-01 5.88997006e-01 -2.90812552e-01
5.05822301e-01 3.71098816e-01 2.60017395e-01 -1.15676843e-01
8.39457631e-01 4.54798996e-01 1.34120250e+00 -5.71361125e-01
-5.48154712e-02 2.92063177e-01 3.54179114e-01 5.90839565e-01
7.22039118e-02 1.87233537e-01 -2.19376191e-01 3.06057986e-02
1.23707652e+00 1.54004246e-01 -6.06295645e-01 1.99932247e-01
-1.08994627e+00 4.25082207e-01 6.41740978e-01 6.78520143e-01
-6.52436793e-01 7.03977188e-03 2.91549116e-01 3.68904144e-01
5.22729814e-01 -4.05603856e-01 -2.07178801e-01 1.70007303e-01
-8.00601125e-01 8.78887996e-02 3.03815931e-01 6.65935040e-01
7.20468760e-01 1.66362852e-01 -2.23443896e-01 6.04103267e-01
3.24963689e-01 1.96324334e-01 4.05322433e-01 -6.98119223e-01
-1.67613477e-01 3.53956670e-01 4.40420508e-01 -1.17347240e+00
-3.31122875e-01 -2.83537567e-01 -1.15635848e+00 1.10866284e+00
3.54931533e-01 -8.63090679e-02 -9.02845740e-01 1.00755179e+00
3.02033007e-01 4.82166380e-01 -1.19536951e-01 7.39757359e-01
9.10225391e-01 8.70100021e-01 3.06860238e-01 -3.23326528e-01
1.45395124e+00 -7.71503389e-01 -1.31842566e+00 -5.54151200e-02
1.44516736e-01 -1.57956278e+00 7.98679531e-01 4.05479342e-01
-1.26759446e+00 -1.01267934e+00 -8.16310704e-01 -1.68406352e-01
-4.45542514e-01 3.63555163e-01 5.85758388e-01 5.73628783e-01
-1.28502607e+00 8.25723588e-01 -5.66051543e-01 -6.07435346e-01
6.98253512e-02 7.68576264e-02 -4.73955661e-01 8.01667497e-02
-6.89090371e-01 9.60169196e-01 3.73496294e-01 3.71093363e-01
-7.94470403e-03 -5.80581725e-01 -5.68816602e-01 1.82899892e-01
-6.80674911e-02 -5.74995935e-01 8.18685889e-01 -9.74287570e-01
-1.32818043e+00 5.77781916e-01 -6.39492393e-01 -1.47624120e-01
5.73123813e-01 -1.83049187e-01 -5.61250031e-01 -1.28217891e-01
-3.65676284e-01 1.49503067e-01 7.90800512e-01 -1.07761133e+00
-9.90356266e-01 -3.08082998e-01 -4.71433103e-01 1.92420527e-01
8.21625441e-02 2.22015500e-01 2.61322781e-02 -1.00231910e+00
4.43730444e-01 -2.82495230e-01 -3.50867957e-01 2.74434835e-01
2.83048712e-02 -2.39202559e-01 1.31957102e+00 -1.16438389e+00
1.24741054e+00 -2.51634073e+00 -3.76235813e-01 2.86155373e-01
-1.50804773e-01 3.26294392e-01 -8.17029271e-03 2.56828964e-01
-2.69072473e-01 2.75656618e-02 -1.25760466e-01 8.45518187e-02
-3.96348774e-01 -7.47527555e-02 2.95070056e-02 6.38016582e-01
-1.92453772e-01 5.11886597e-01 -4.45489138e-01 -4.47826594e-01
6.35087192e-01 9.77775037e-01 -2.76838452e-01 -2.98552494e-02
5.45560777e-01 6.01551294e-01 -2.23108336e-01 4.01896209e-01
1.41412401e+00 3.63962412e-01 -5.20468712e-01 -6.27810955e-01
-6.54641986e-01 -6.75798059e-01 -1.88286817e+00 1.07480764e+00
-1.54001221e-01 6.31840169e-01 5.96581995e-01 -6.86361670e-01
1.40555573e+00 4.23918486e-01 5.85622294e-03 -4.49645847e-01
3.23061794e-01 1.99531466e-01 -3.54773611e-01 -6.34244263e-01
6.50492251e-01 -3.38658363e-01 8.24243724e-01 -6.05837665e-02
-3.16247493e-01 1.38436705e-01 1.61421001e-01 4.39307792e-03
5.68014503e-01 -8.89492501e-03 7.08928227e-01 -9.81592119e-01
1.31343448e+00 -3.54382902e-01 7.08071768e-01 5.38287103e-01
-3.46226543e-01 5.87311029e-01 -1.37638226e-01 -2.65418112e-01
-1.01417160e+00 -9.17607784e-01 -3.89627099e-01 7.46140361e-01
5.16287446e-01 6.97545037e-02 -6.93764925e-01 2.50926260e-02
-5.09269424e-02 4.19724107e-01 -3.52972835e-01 -4.38922308e-02
-6.03215098e-01 -4.80164617e-01 -8.72728303e-02 5.24173141e-01
1.24826539e+00 -1.11772287e+00 -1.35589153e-01 3.35409462e-01
1.96934998e-01 -5.69769323e-01 -7.67619193e-01 -2.82989055e-01
-1.27250087e+00 -6.93483829e-01 -9.98007476e-01 -1.45335495e+00
9.49883759e-01 6.59463823e-01 5.31206846e-01 5.68094730e-01
-2.63010114e-01 -6.12493791e-02 -2.92025030e-01 -1.53307900e-01
-3.11633945e-01 -9.16123092e-01 -3.74849111e-01 1.48598030e-01
3.10739785e-01 -4.53508735e-01 -8.77449155e-01 4.47296619e-01
-9.66574728e-01 -2.76154339e-01 6.76745951e-01 8.64897788e-01
4.89539564e-01 1.00175166e+00 4.24381346e-01 -9.44287002e-01
1.06050348e+00 1.17392890e-01 -6.76902115e-01 1.63665786e-01
-3.93703222e-01 -1.64690569e-01 6.22511029e-01 -1.75080851e-01
-1.92540431e+00 -2.27373257e-01 -1.72537610e-01 2.91790783e-01
-5.01473546e-01 3.45666528e-01 -1.75026953e-01 -5.07106781e-01
7.02832103e-01 3.08849841e-01 1.57906130e-01 -1.06508803e+00
1.70624524e-01 6.16921723e-01 1.01634359e+00 -1.19627118e-01
6.85361207e-01 7.27192461e-01 2.16988772e-01 -1.15106273e+00
3.04622620e-01 -8.70079935e-01 -5.64562023e-01 -1.78489685e-01
1.02272403e+00 -6.02027416e-01 -9.54548895e-01 8.12479854e-01
-1.18875778e+00 1.45824343e-01 -4.70372923e-02 6.07029498e-01
-2.35138863e-01 8.01166117e-01 -6.78630531e-01 -1.04955304e+00
-2.61259586e-01 -1.16358674e+00 5.19017756e-01 9.85862494e-01
2.82922745e-01 -1.43835390e+00 -4.13620442e-01 -1.89601377e-01
9.57387984e-01 3.04136425e-01 5.24603903e-01 3.25879693e-01
-4.94071484e-01 -1.92401931e-01 -5.87289035e-01 5.22873521e-01
5.27342558e-01 2.04532981e-01 -5.85404634e-01 -2.81815708e-01
2.00306490e-01 6.93084180e-01 1.01047194e+00 9.77729321e-01
9.25568402e-01 -3.34120579e-02 -3.43011826e-01 6.09652936e-01
1.97198045e+00 4.62777734e-01 1.24282801e+00 3.77554864e-01
5.11902794e-02 3.87995005e-01 6.01593494e-01 2.45923385e-01
-3.99679184e-01 1.46383464e-01 -6.97583556e-02 -6.92699254e-01
-4.07586604e-01 3.21712978e-02 1.94545954e-01 5.81350267e-01
-4.54490095e-01 1.32744610e-01 -1.84892461e-01 6.01700366e-01
-1.88183630e+00 -1.18249404e+00 -7.63614893e-01 2.25752258e+00
5.83647251e-01 3.28222848e-02 -2.66545981e-01 2.73198485e-01
1.19451857e+00 1.46093778e-02 1.58559188e-01 -8.22856307e-01
-2.20807508e-01 2.16334209e-01 8.15019965e-01 1.05414021e+00
-1.22637773e+00 6.84511364e-01 5.71830463e+00 1.04511154e+00
-1.01739347e+00 -3.26870270e-02 2.76638508e-01 1.09884703e+00
-1.56001091e-01 2.56168485e-01 -7.70266950e-01 4.98557925e-01
1.42246678e-01 -5.31503677e-01 3.40278894e-01 6.41617775e-01
9.65894520e-01 -7.97327220e-01 -2.06680000e-01 8.48872006e-01
-4.26292807e-01 -1.09699297e+00 -1.01488575e-01 -1.66531160e-01
8.39514017e-01 -7.25165904e-01 -1.62889063e-01 -2.80012339e-01
4.25419472e-02 -8.45092952e-01 6.60949498e-02 1.01595175e+00
1.06119633e-01 -8.70902002e-01 1.13682973e+00 4.75093216e-01
-1.60305786e+00 1.48917079e-01 -8.26854169e-01 -1.24554694e-01
6.52791917e-01 9.02574956e-01 -3.13313991e-01 5.53177297e-01
8.61762881e-01 5.89124918e-01 -4.99263972e-01 1.77770698e+00
-4.03431728e-02 1.62776366e-01 -1.92088529e-01 3.91526759e-01
2.07268283e-01 -8.70687008e-01 6.40150249e-01 1.57199407e+00
4.71005559e-01 4.71673012e-01 9.35008749e-03 5.76701283e-01
4.51739281e-01 4.81047988e-01 -4.22816098e-01 3.29292774e-01
1.91588789e-01 1.19844377e+00 -8.83910835e-01 -5.49158573e-01
-6.03708863e-01 9.31079745e-01 -6.34767532e-01 4.84564424e-01
-3.32749158e-01 -6.95863843e-01 4.37837809e-01 8.28784481e-02
2.32181355e-01 -2.31201038e-01 -6.76705956e-01 -7.63709307e-01
-2.65029490e-01 -4.55088466e-01 4.23527390e-01 -9.12259996e-01
-1.15770960e+00 3.84435117e-01 -3.23513389e-01 -1.14188409e+00
5.83328187e-01 -5.63882589e-01 -1.21246886e+00 1.45166230e+00
-1.47710371e+00 -1.05175078e+00 -6.53099239e-01 8.62490654e-01
6.88791215e-01 1.17659010e-01 1.39466003e-01 3.27615261e-01
1.79653376e-01 9.01370123e-02 5.25738597e-01 -3.23817253e-01
9.15399432e-01 -1.27417910e+00 2.61276457e-02 1.36380315e+00
-7.46352971e-01 9.19530988e-01 1.13586414e+00 -1.07748115e+00
-6.32027626e-01 -5.37791729e-01 1.05975294e+00 5.49765468e-01
2.29549319e-01 4.70234007e-01 -1.40735495e+00 5.81132948e-01
6.88400865e-01 -8.77413452e-02 -9.60548595e-02 -4.17989612e-01
3.89662653e-01 -1.94066733e-01 -1.57219410e+00 6.19936168e-01
5.88712633e-01 -1.50991276e-01 -7.33093202e-01 1.54827476e-01
8.07878375e-02 -2.40375131e-01 -7.15350389e-01 4.16293889e-01
4.04412568e-01 -1.23874545e+00 1.18448806e+00 -9.12332609e-02
-1.66223556e-01 -9.16402757e-01 2.82738358e-01 -1.27894175e+00
-8.60572040e-01 -6.24086380e-01 8.15803766e-01 1.47921503e+00
-1.50020318e-02 -8.43094230e-01 5.67837775e-01 4.70042199e-01
-1.00604303e-01 -6.26790822e-02 -6.18622601e-01 -6.10746682e-01
-2.50847578e-01 4.56358880e-01 1.88790664e-01 7.46579230e-01
-7.31093138e-02 -2.82490015e-01 -5.69398999e-01 3.73381078e-01
7.56366611e-01 -3.73184741e-01 6.45816267e-01 -1.23675609e+00
8.11696723e-02 -4.96358037e-01 -4.39700186e-01 -1.15745175e+00
-5.18216908e-01 -4.00004297e-01 2.41421908e-02 -1.90287411e+00
-8.16473439e-02 -1.87973492e-02 -1.21809594e-01 -4.12683338e-02
-5.59431434e-01 2.16932133e-01 -6.31081685e-02 -7.97409862e-02
3.59463692e-01 2.55058020e-01 1.91539359e+00 1.31545216e-01
-4.70855951e-01 1.42873973e-01 -3.20911288e-01 8.67284596e-01
6.52136564e-01 1.32604450e-01 -1.79241046e-01 -2.27399305e-01
-2.50442713e-01 -2.01217219e-01 3.22662830e-01 -7.67229080e-01
7.49611855e-01 -3.15716743e-01 4.50470805e-01 -6.81146860e-01
6.22288324e-02 -1.02345824e+00 3.75753313e-01 5.52586973e-01
9.29708257e-02 6.90026656e-02 2.01444149e-01 5.63315690e-01
-5.75527370e-01 -4.42544281e-01 1.34352148e+00 -3.23927343e-01
-1.42382061e+00 -2.16138825e-01 -4.97824371e-01 -8.05947602e-01
1.02533901e+00 -6.97272360e-01 -5.80236733e-01 -4.41470742e-01
-1.05310869e+00 -1.82807118e-01 3.62221092e-01 -2.78613239e-04
6.92722738e-01 -1.16214585e+00 -7.48364806e-01 4.60352719e-01
-5.22323847e-01 -3.55521619e-01 6.82545483e-01 1.22881663e+00
-1.09704256e+00 -1.01130351e-03 -7.16536343e-01 -1.64825812e-01
-1.75088668e+00 5.87493360e-01 1.70088515e-01 3.18142384e-01
-9.75692153e-01 8.53890240e-01 3.29273462e-01 1.83240712e-01
6.57386035e-02 -3.34742665e-01 -5.08329570e-01 -4.49796557e-01
8.61658990e-01 9.75150287e-01 -6.65682927e-02 -7.11287796e-01
-1.37106642e-01 1.00183761e+00 7.04713240e-02 2.09559813e-01
1.16326833e+00 -6.15576446e-01 -5.31220078e-01 -2.66444355e-01
9.71241236e-01 5.60081363e-01 -1.04982388e+00 -1.18367784e-01
-7.00525194e-02 -1.14333212e+00 4.45091367e-01 -6.55006647e-01
-9.19862509e-01 8.33603203e-01 9.81787920e-01 3.09777737e-01
1.37724292e+00 -5.07645845e-01 7.62918115e-01 -1.70215398e-01
-4.82027419e-02 -1.17790294e+00 -5.17863095e-01 1.46062598e-01
9.67448771e-01 -1.08329070e+00 1.61305398e-01 -9.28109169e-01
-2.95387447e-01 1.43012011e+00 3.18093061e-01 -6.23403609e-01
1.02146804e+00 3.82756412e-01 2.14797154e-01 1.87819138e-01
-9.26527753e-02 -3.37692142e-01 4.63325888e-01 6.37180507e-01
6.69474363e-01 -1.16076857e-01 -1.52612555e+00 2.11236134e-01
3.60424101e-01 5.46104789e-01 7.06850708e-01 8.49367797e-01
-1.29117405e+00 -9.57727313e-01 -1.00961900e+00 1.91707611e-01
-4.17830139e-01 1.28989011e-01 2.35665649e-01 5.05198181e-01
3.09775352e-01 1.09988129e+00 -3.15089785e-02 1.32020906e-01
3.17909509e-01 -1.55111000e-01 7.68342912e-02 1.61991507e-01
-5.14281631e-01 5.32142282e-01 -1.90151498e-01 -2.54478276e-01
-5.12375474e-01 -3.10822994e-01 -1.26178312e+00 -6.91336930e-01
-5.25518775e-01 1.45605028e-01 7.68703878e-01 8.23387384e-01
1.43509852e-02 5.01085997e-01 2.14075834e-01 -7.73762763e-01
-2.05996484e-01 -8.62904072e-01 -1.20369625e+00 6.47103012e-01
3.55901748e-01 -5.28840482e-01 -5.78944266e-01 6.30429506e-01] | [11.09510612487793, -2.575838565826416] |
08a3744d-6636-46c0-a74c-fe8c5f060eef | rethinking-domain-generalization-for-face | 2303.13662 | null | https://arxiv.org/abs/2303.13662v1 | https://arxiv.org/pdf/2303.13662v1.pdf | Rethinking Domain Generalization for Face Anti-spoofing: Separability and Alignment | This work studies the generalization issue of face anti-spoofing (FAS) models on domain gaps, such as image resolution, blurriness and sensor variations. Most prior works regard domain-specific signals as a negative impact, and apply metric learning or adversarial losses to remove them from feature representation. Though learning a domain-invariant feature space is viable for the training data, we show that the feature shift still exists in an unseen test domain, which backfires on the generalizability of the classifier. In this work, instead of constructing a domain-invariant feature space, we encourage domain separability while aligning the live-to-spoof transition (i.e., the trajectory from live to spoof) to be the same for all domains. We formulate this FAS strategy of separability and alignment (SA-FAS) as a problem of invariant risk minimization (IRM), and learn domain-variant feature representation but domain-invariant classifier. We demonstrate the effectiveness of SA-FAS on challenging cross-domain FAS datasets and establish state-of-the-art performance. | ['Wen-Sheng Chu', 'Yixuan Li', 'Xiaoming Liu', 'Yaojie Liu', 'Yiyou Sun'] | 2023-03-23 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Sun_Rethinking_Domain_Generalization_for_Face_Anti-Spoofing_Separability_and_Alignment_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Sun_Rethinking_Domain_Generalization_for_Face_Anti-Spoofing_Separability_and_Alignment_CVPR_2023_paper.pdf | cvpr-2023-1 | ['metric-learning', 'face-anti-spoofing', 'metric-learning'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 6.35523498e-01 -5.88425610e-04 -2.83773720e-01 -3.21667373e-01
-6.39604390e-01 -7.92398989e-01 6.74608469e-01 -4.39623088e-01
9.75021571e-02 7.17199028e-01 1.03240833e-01 -1.59311339e-01
-2.60742128e-01 -4.98077124e-01 -7.88470089e-01 -9.19468701e-01
-1.51834235e-01 -4.62823026e-02 1.40379919e-02 -2.34304383e-01
4.00147475e-02 7.56614625e-01 -1.25231719e+00 3.28912377e-01
8.02450299e-01 1.05449581e+00 -5.54011583e-01 3.11809450e-01
3.69713068e-01 5.42539239e-01 -9.70256090e-01 -6.13180816e-01
5.86314499e-01 -5.51094353e-01 -7.90814340e-01 1.93077162e-01
7.26275086e-01 -4.13074225e-01 -6.31185889e-01 1.33891273e+00
4.25345540e-01 -2.27344871e-01 9.69881356e-01 -1.87543929e+00
-9.32161868e-01 -7.10942075e-02 -6.89582527e-01 8.36299360e-02
3.93772513e-01 2.10959353e-02 6.08985484e-01 -7.19517708e-01
7.84842610e-01 1.46439052e+00 7.54577458e-01 7.79924750e-01
-1.45592356e+00 -1.06687677e+00 -1.46010742e-02 1.79874182e-01
-1.35800207e+00 -8.59948814e-01 1.04010415e+00 -6.73765063e-01
4.06401426e-01 1.61288664e-01 1.95516720e-01 1.78006101e+00
2.78896868e-01 3.84350479e-01 1.20907617e+00 -2.66210496e-01
2.08749264e-01 2.18375131e-01 -2.03381971e-01 4.57713455e-01
4.39654291e-01 6.84608340e-01 -7.06493556e-01 -4.84546483e-01
6.69410110e-01 -2.36609280e-01 -4.13456321e-01 -7.39237905e-01
-7.50580609e-01 9.13798928e-01 1.34037182e-01 -9.14867073e-02
-2.30615824e-01 -3.71558279e-01 3.32366467e-01 6.50768876e-01
5.23558855e-01 3.92148972e-01 -4.65056598e-01 2.92551190e-01
-8.58068109e-01 3.24707508e-01 7.56656647e-01 8.20331633e-01
7.38449156e-01 -6.00684248e-02 -7.98488483e-02 6.85541928e-01
8.54873434e-02 7.90909171e-01 1.97405934e-01 -8.82518411e-01
3.30432713e-01 2.01970026e-01 1.06704101e-01 -1.16342700e+00
-6.95152730e-02 -1.44786879e-01 -6.52319372e-01 1.73767433e-01
6.70926750e-01 -4.41439122e-01 -9.08223748e-01 2.19539833e+00
5.09708166e-01 5.70352674e-01 8.26353729e-02 7.82841146e-01
2.86821842e-01 3.21273386e-01 -3.92508954e-02 -3.96384954e-01
1.05433321e+00 -4.23072428e-01 -7.35612988e-01 -3.41205329e-01
3.81658107e-01 -6.68496311e-01 7.83202589e-01 3.84017766e-01
-6.08631074e-01 -1.92297697e-01 -1.18380761e+00 2.36807227e-01
-2.12776840e-01 -4.25051928e-01 2.28858918e-01 1.02429497e+00
-8.48855734e-01 6.15987122e-01 -5.21747887e-01 -3.87974560e-01
7.22721756e-01 3.79399210e-01 -7.61363626e-01 -3.13587904e-01
-1.37836051e+00 7.87069142e-01 4.75996397e-02 -4.69279617e-01
-1.14947069e+00 -9.04520810e-01 -8.38449121e-01 -3.29310179e-01
2.42875338e-01 -4.39749539e-01 7.17999160e-01 -1.20411909e+00
-1.39631903e+00 1.13079333e+00 -2.62263954e-01 -3.63716096e-01
5.61903477e-01 -1.69457048e-01 -9.52522814e-01 2.65606791e-01
1.30972937e-01 3.11808765e-01 1.66974461e+00 -1.45685935e+00
-4.19710368e-01 -5.29033780e-01 -1.56552643e-01 -2.85484821e-01
-5.60905874e-01 2.44350389e-01 1.85795799e-01 -7.76904345e-01
-1.67789888e-02 -9.51601744e-01 4.18906063e-01 2.06393242e-01
-3.36193621e-01 5.75961247e-02 1.36903155e+00 -9.05108094e-01
1.05897665e+00 -2.59913588e+00 2.05525011e-02 1.65502861e-01
1.61467970e-03 4.29634511e-01 -4.46076274e-01 9.40543190e-02
-2.78827578e-01 2.34933779e-01 -4.91835147e-01 -1.54949173e-01
-2.37514198e-01 2.79815942e-01 -4.93859112e-01 1.13291717e+00
8.18180442e-01 3.68418902e-01 -8.32022607e-01 -5.38051188e-01
-5.07081822e-02 5.70953667e-01 -8.39685678e-01 2.03596547e-01
2.60879230e-02 7.89626598e-01 -3.22576076e-01 8.49794388e-01
1.46931696e+00 1.84898615e-01 -2.06167251e-02 -1.34329230e-01
3.02360356e-01 2.40140811e-01 -1.00215590e+00 1.30495965e+00
-2.45067000e-01 5.91004789e-01 4.90590602e-01 -1.17687428e+00
8.96125913e-01 2.45617211e-01 7.04399765e-01 -6.71387196e-01
-7.17136413e-02 1.08840890e-01 -1.07340991e-01 -4.61689562e-01
-2.70842444e-02 -3.89949650e-01 6.22089207e-02 2.53013402e-01
4.34402734e-01 4.09147777e-02 -5.70245266e-01 -2.37010583e-01
1.15122008e+00 1.89352676e-03 1.22965038e-01 -5.28121412e-01
4.21574533e-01 -2.05195472e-01 9.09285188e-01 4.62888390e-01
-8.28263521e-01 6.16859972e-01 5.70903897e-01 -4.82807793e-02
-1.03382266e+00 -1.31888092e+00 -6.16750538e-01 9.97978806e-01
4.52514231e-01 -1.25292595e-03 -7.96022952e-01 -1.10442066e+00
3.70270252e-01 4.99679267e-01 -6.53335214e-01 -6.18037641e-01
-6.28212631e-01 -4.76020753e-01 9.03240860e-01 -8.72900039e-02
5.87033868e-01 -4.85294521e-01 -2.06282154e-01 -1.29068285e-01
-2.01792181e-01 -1.14732599e+00 -4.49560821e-01 1.16191553e-02
-4.52787936e-01 -1.23991966e+00 -4.02113080e-01 -7.68032670e-01
4.93403554e-01 2.06950128e-01 8.41408491e-01 -2.65236288e-01
-1.58865139e-01 3.63139689e-01 -1.85528368e-01 -4.04668212e-01
-4.94975775e-01 -4.00811017e-01 4.05358911e-01 3.93156737e-01
4.75030750e-01 -5.15076578e-01 -5.40066957e-01 5.03750563e-01
-7.79020786e-01 -7.38666236e-01 4.68691327e-02 1.04661775e+00
1.93061143e-01 1.59206718e-01 7.33591139e-01 -6.69532180e-01
4.52350020e-01 -7.42479682e-01 -3.74987900e-01 2.07801685e-01
-5.96398175e-01 1.25590675e-02 4.68369544e-01 -8.05709362e-01
-1.02307308e+00 -1.23415865e-01 1.49383038e-01 -7.69680321e-01
-2.58666247e-01 -2.11315617e-01 -5.54185092e-01 -4.15494978e-01
9.45855439e-01 8.81296173e-02 3.45469326e-01 -2.52854228e-01
3.17890137e-01 9.02299941e-01 5.24198174e-01 -6.59238935e-01
1.21996391e+00 7.78430641e-01 6.63684160e-02 -1.10582006e+00
-5.60088634e-01 -3.80572081e-01 -4.65417027e-01 -1.20507851e-02
5.27464807e-01 -9.09633696e-01 -3.01709563e-01 5.91143668e-01
-1.25696814e+00 5.63460588e-02 -3.39025944e-01 2.24270836e-01
-5.70703268e-01 5.36418676e-01 -2.47736186e-01 -7.84635782e-01
2.31227726e-01 -1.01717925e+00 1.11574268e+00 -7.06726313e-02
-1.35411710e-01 -9.45268154e-01 -1.34890433e-03 2.28965685e-01
1.82178855e-01 6.31270528e-01 7.91895151e-01 -9.14341152e-01
-7.43847117e-02 -5.53739071e-02 -1.78240240e-01 7.56141841e-01
3.84493709e-01 -1.77827671e-01 -1.37621164e+00 -7.44156957e-01
5.44429004e-01 -3.40810150e-01 9.07086968e-01 2.32356116e-01
1.06745899e+00 -7.49126613e-01 -4.16917503e-01 9.43940520e-01
1.11382854e+00 1.85765654e-01 5.84994733e-01 1.70680270e-01
4.81328100e-01 9.58227754e-01 6.55875146e-01 5.23598850e-01
-7.45250359e-02 6.62035584e-01 4.51745957e-01 9.22324229e-03
-2.06967860e-01 -4.22059625e-01 7.60376394e-01 -1.18897274e-01
5.92477441e-01 -2.00443745e-01 -6.67877376e-01 6.67537272e-01
-1.28729141e+00 -1.08928084e+00 4.53643471e-01 2.31918216e+00
7.89673805e-01 -2.16027856e-01 2.89563507e-01 2.58506924e-01
1.10290408e+00 2.95526147e-01 -6.85981214e-01 -4.62892026e-01
-3.17502499e-01 1.95338205e-01 5.22143722e-01 4.00352448e-01
-1.41907489e+00 8.44991446e-01 6.27261209e+00 8.07792068e-01
-1.44254196e+00 1.92314252e-01 4.46251005e-01 2.89657097e-02
-1.18379518e-01 -1.55773863e-01 -6.06862307e-01 6.24551058e-01
8.77408445e-01 -1.42678276e-01 7.54024923e-01 7.28282452e-01
-7.61119947e-02 5.09705365e-01 -1.28324211e+00 9.53651309e-01
2.15154454e-01 -8.88240397e-01 -2.67376192e-02 3.77682626e-01
7.79944181e-01 -2.31336072e-01 4.57153738e-01 1.19128257e-01
2.33990729e-01 -1.07460940e+00 6.96460724e-01 -1.20897762e-01
1.26577759e+00 -6.54311359e-01 3.27848494e-01 9.52293351e-02
-7.58920193e-01 -3.51624429e-01 -3.76007468e-01 2.91632742e-01
-2.29573473e-01 5.34315705e-01 -8.35499167e-01 5.02812564e-01
5.97976387e-01 6.20107412e-01 -3.18998843e-01 5.65992713e-01
-1.90143716e-02 6.45310879e-01 -2.47146010e-01 7.52365947e-01
-2.45658770e-01 4.76532020e-02 9.31408644e-01 1.10733676e+00
3.50149900e-01 -1.53857797e-01 -4.33906280e-02 8.48041415e-01
-2.16001838e-01 -4.71083164e-01 -1.09450603e+00 -1.84259698e-01
9.43841100e-01 5.88487446e-01 1.16152421e-01 1.88031793e-01
-3.57051909e-01 1.15115142e+00 1.07651269e-02 4.68670160e-01
-8.00036252e-01 -2.09266827e-01 1.46924591e+00 2.98033416e-01
3.29985678e-01 1.33732110e-01 -3.53557914e-01 -1.16876924e+00
6.51011840e-02 -1.15178764e+00 5.16379356e-01 -1.62462126e-02
-1.65474844e+00 2.40494281e-01 -9.40794349e-02 -1.28107345e+00
-2.35312849e-01 -6.08399689e-01 -3.43801379e-01 7.35803962e-01
-1.76114428e+00 -1.21722567e+00 9.52363014e-02 9.84869719e-01
1.17686421e-01 -2.90774494e-01 7.06784844e-01 2.32789725e-01
-4.66615081e-01 1.19659019e+00 2.67756164e-01 2.75422484e-01
1.09098876e+00 -7.18376935e-01 1.52267903e-01 1.00858986e+00
-2.07932279e-01 6.32901847e-01 7.43127882e-01 -6.20757461e-01
-1.45089984e+00 -1.31669748e+00 6.86598420e-01 -5.44964373e-01
6.97561204e-01 -4.87457216e-01 -9.07976687e-01 5.28479099e-01
-1.57332793e-01 5.12718856e-01 5.93073964e-01 -1.86815575e-01
-1.02600539e+00 -1.02558292e-01 -1.96380281e+00 3.16879481e-01
1.36474478e+00 -9.46802735e-01 -5.05774975e-01 2.07690850e-01
7.45052159e-01 -8.86549503e-02 -7.71133065e-01 6.60052836e-01
5.68422377e-01 -9.67756867e-01 1.26612318e+00 -8.18160534e-01
2.91871250e-01 -4.10274528e-02 -4.17851269e-01 -1.31327784e+00
-2.87900090e-01 -9.57419157e-01 -3.08669746e-01 1.50673950e+00
-1.66435558e-02 -8.23912203e-01 6.74049437e-01 1.74083382e-01
2.50597775e-01 -2.44363979e-01 -1.44560552e+00 -1.31461930e+00
6.07361794e-01 -1.19464751e-02 9.98133361e-01 1.37371933e+00
-6.22267947e-02 -1.51860058e-01 -4.47915524e-01 7.69261837e-01
9.27170575e-01 -2.62799203e-01 5.41611016e-01 -1.34097791e+00
-4.19992507e-02 -1.69577166e-01 -5.21318972e-01 -7.31059968e-01
5.83167434e-01 -7.08621979e-01 -1.05833970e-01 -5.05349696e-01
-1.20134698e-02 -3.49445224e-01 -4.84973192e-01 1.30748942e-01
1.29552126e-01 4.83738072e-02 1.38337925e-01 2.10489199e-01
-1.61012858e-01 3.69596213e-01 1.27709663e+00 -2.94713765e-01
-5.31367678e-03 -1.44639894e-01 -7.63972282e-01 4.16230083e-01
6.32514775e-01 -6.77526653e-01 -4.09743726e-01 -2.24435166e-01
-5.09957135e-01 1.17408730e-01 5.32836974e-01 -9.36096847e-01
-1.39715418e-01 -5.75972080e-01 4.32362378e-01 1.66012928e-01
4.47335541e-01 -1.00739026e+00 -1.24607973e-01 5.27296305e-01
-3.05480540e-01 -5.32498419e-01 1.77858964e-01 6.88799679e-01
-1.22217119e-01 2.07329303e-01 1.29926920e+00 3.82954270e-01
-4.09928709e-01 4.29872870e-01 1.84442624e-01 4.40176845e-01
1.04528248e+00 -3.96213204e-01 -7.84364939e-01 -1.39220178e-01
-3.07838827e-01 -1.79444849e-01 7.99887002e-01 6.46609306e-01
5.81368208e-01 -1.35146046e+00 -8.22919786e-01 8.40757012e-01
3.30585927e-01 -2.54481167e-01 6.44760355e-02 3.45308006e-01
-1.29589196e-02 1.68230951e-01 -3.12875986e-01 -6.50096416e-01
-1.18434668e+00 7.41015255e-01 3.55649352e-01 8.89807418e-02
-2.23797619e-01 1.01568055e+00 5.11911035e-01 -3.56482148e-01
1.33815199e-01 2.53919393e-01 1.22264452e-01 2.83611845e-02
4.28404599e-01 3.71960819e-01 1.81705318e-03 -8.90027583e-01
-7.54056156e-01 5.12670159e-01 5.55192307e-02 -9.38737392e-02
1.08131766e+00 -1.54596299e-01 6.99886531e-02 -2.01098457e-01
1.55541968e+00 1.15271591e-01 -1.73411274e+00 -2.42142230e-01
8.86076838e-02 -8.62775564e-01 -7.50760809e-02 -7.34823644e-01
-8.43137324e-01 8.40233147e-01 9.12799358e-01 1.65452316e-01
1.28191018e+00 7.38936588e-02 6.27804339e-01 -1.40190959e-01
3.44742417e-01 -9.64415431e-01 2.07539573e-01 3.98839831e-01
9.50585306e-01 -1.26737022e+00 -2.84171462e-01 -5.51158369e-01
-5.54374039e-01 6.36659682e-01 4.92487758e-01 -1.44669071e-01
7.77008355e-01 2.95623511e-01 -1.25020325e-01 3.67045477e-02
-3.50278407e-01 1.77252442e-01 3.22345644e-02 1.50748622e+00
-6.11862391e-02 1.46827817e-01 5.25624417e-02 3.91708463e-01
-2.64343709e-01 -1.24973990e-01 2.16655850e-01 9.44192410e-01
-6.02931418e-02 -1.17609596e+00 -6.50744081e-01 2.32317612e-01
-5.72221279e-01 2.99802572e-01 -6.56059206e-01 6.67871177e-01
4.02405947e-01 1.29177701e+00 6.20449036e-02 -8.43255699e-01
1.33326173e-01 1.37125120e-01 5.77367783e-01 -4.31728482e-01
-9.28285718e-02 -3.40836257e-01 -7.91987684e-03 -5.75714707e-01
-9.66371149e-02 -8.07079315e-01 -6.60344779e-01 -7.28486776e-01
-3.73027503e-01 -1.55436799e-01 5.92402160e-01 8.21945608e-01
5.87537527e-01 -1.68012397e-03 1.34827209e+00 -5.05816281e-01
-1.07115602e+00 -6.47002220e-01 -6.29983962e-01 9.10472512e-01
1.26019967e+00 -1.02510226e+00 -7.54059494e-01 1.00521602e-01] | [13.067424774169922, 1.186003565788269] |
e2bdc041-b2e2-4ec2-8b94-3469912f636f | ai-predicts-independent-construction-safety | 1908.05972 | null | https://arxiv.org/abs/1908.05972v2 | https://arxiv.org/pdf/1908.05972v2.pdf | AI-based Prediction of Independent Construction Safety Outcomes from Universal Attributes | This paper significantly improves on, and finishes to validate, an approach proposed in previous research in which safety outcomes were predicted from attributes with machine learning. Like in the original study, we use Natural Language Processing (NLP) to extract fundamental attributes from raw incident reports and machine learning models are trained to predict safety outcomes. The outcomes predicted here are injury severity, injury type, body part impacted, and incident type. However, unlike in the original study, safety outcomes were not extracted via NLP but were provided by independent human annotations, eliminating any potential source of artificial correlation between predictors and predictands. Results show that attributes are still highly predictive, confirming the validity of the original approach. Other improvements brought by the current study include the use of (1) a much larger dataset featuring more than 90,000 reports, (2) two new models, XGBoost and linear SVM (Support Vector Machines), (3) model stacking, (4) a more straightforward experimental setup with more appropriate performance metrics, and (5) an analysis of per-category attribute importance scores. Finally, the injury severity outcome is well predicted, which was not the case in the original study. This is a significant advancement. | ['Antoine J. -P. Tixier', 'Matthew R. Hallowell', 'Henrietta Baker'] | 2019-08-16 | null | null | null | null | ['injury-prediction'] | ['playing-games'] | [ 2.32075527e-01 3.45493495e-01 -5.58783472e-01 -3.89324635e-01
-8.37286294e-01 -4.45850164e-01 5.15147090e-01 1.14072526e+00
-5.67035675e-01 1.22772288e+00 5.94738007e-01 -3.15884024e-01
-4.13641751e-01 -9.44338679e-01 -4.93986219e-01 -4.62684363e-01
-2.28098392e-01 4.15088207e-01 8.55447426e-02 6.31863400e-02
4.47810590e-01 4.52792406e-01 -1.48713708e+00 4.32492107e-01
6.36938870e-01 1.18579090e+00 -3.28277141e-01 4.76853400e-01
6.38400540e-02 1.16657627e+00 -4.51022744e-01 -5.06032884e-01
2.60857433e-01 -1.68084294e-01 -8.84224474e-01 -2.25409731e-01
2.26435736e-01 -2.31549039e-01 1.51207298e-01 2.95571536e-01
1.99475557e-01 -3.61865982e-02 8.16898882e-01 -1.33223355e+00
-5.35168827e-01 5.06060004e-01 -3.23797435e-01 -8.78932700e-02
6.21430457e-01 1.28275052e-01 9.98093069e-01 -5.55903077e-01
4.64220017e-01 9.34757173e-01 9.19819117e-01 3.54783535e-01
-1.13815701e+00 -8.05930376e-01 1.16471216e-01 4.15171981e-01
-1.18028092e+00 -7.56646395e-02 5.34357548e-01 -7.90766418e-01
1.03186345e+00 3.97763044e-01 6.02889180e-01 1.15073121e+00
2.42372602e-01 1.69749245e-01 1.26512253e+00 -6.28803253e-01
4.69505757e-01 5.64489543e-01 5.61142862e-01 3.57814312e-01
4.24881488e-01 4.93315428e-01 -3.49432766e-01 -5.04731536e-01
5.79770356e-02 3.21978599e-01 -3.17543782e-02 -1.76683173e-01
-1.28169668e+00 1.05788934e+00 2.08611965e-01 1.19713880e-01
-7.85061061e-01 -3.10838282e-01 6.52201712e-01 3.26098800e-02
5.38655043e-01 7.07625508e-01 -7.70851612e-01 -1.77623272e-01
-8.55189979e-01 1.98310524e-01 9.02439952e-01 6.41856849e-01
4.10966516e-01 -8.78413171e-02 -3.31684768e-01 6.08290613e-01
2.40969043e-02 2.50395417e-01 4.20444727e-01 -7.91486800e-01
4.58983451e-01 8.19051206e-01 4.33732942e-02 -1.07851648e+00
-7.66968846e-01 -3.51170450e-01 -6.63405597e-01 1.80728495e-01
4.13940549e-01 -3.04163843e-01 -7.03978240e-01 1.54853117e+00
8.57058465e-02 -1.34833172e-01 -2.90565174e-02 7.70989597e-01
5.81952631e-01 3.51488769e-01 7.88087904e-01 -4.57796812e-01
1.29254472e+00 -6.24005020e-01 -7.12234080e-01 7.84295574e-02
7.97795951e-01 -4.95911539e-01 7.68679142e-01 6.49465382e-01
-9.41383481e-01 -4.20963526e-01 -7.09505141e-01 4.19244051e-01
-7.29535639e-01 -3.94135155e-02 1.01094019e+00 7.08785355e-01
-6.38090551e-01 6.82551682e-01 -4.69242573e-01 -6.29122913e-01
3.74209404e-01 3.82172257e-01 -7.03102171e-01 1.29840374e-01
-1.33749962e+00 1.29677606e+00 5.76300144e-01 -5.06438971e-01
-4.14311081e-01 -9.54125345e-01 -1.03119445e+00 9.14489105e-03
5.15130341e-01 -6.94118083e-01 8.31629276e-01 -6.36185884e-01
-9.55973029e-01 6.36927903e-01 -8.68413523e-02 -4.97327089e-01
2.74702370e-01 -2.88835704e-01 -4.70370263e-01 5.10131456e-02
1.64652690e-01 1.91152275e-01 8.62049405e-03 -1.08965337e+00
-8.58510196e-01 -5.50220191e-01 -1.79411590e-01 -2.50254840e-01
-3.99649084e-01 3.99703175e-01 1.75389752e-01 -7.11066782e-01
-1.64968625e-01 -7.78720379e-01 -4.23318535e-01 -3.36323291e-01
-3.05913866e-01 -2.21063226e-01 1.84281338e-02 -9.27871883e-01
1.52190495e+00 -1.93289614e+00 -3.10214043e-01 3.01460415e-01
4.41518910e-02 2.17688859e-01 1.69783264e-01 7.26633489e-01
-6.33801877e-01 2.97452122e-01 -1.63293585e-01 -4.30790633e-02
-1.71423584e-01 2.21103672e-02 -3.09861988e-01 2.68184006e-01
3.50665957e-01 5.68642139e-01 -7.09062636e-01 -5.64817071e-01
4.19144303e-01 3.06292921e-02 -3.45696628e-01 6.69189766e-02
3.75890434e-01 2.28650615e-01 -2.75147289e-01 5.23505628e-01
3.16695392e-01 2.09933728e-01 -2.63825595e-01 -6.04176968e-02
-3.63386035e-01 1.50330260e-01 -1.02002299e+00 1.07749569e+00
-3.69228184e-01 1.93129256e-01 -3.49667162e-01 -1.22622049e+00
1.08088541e+00 6.83060765e-01 9.43069994e-01 -5.78053176e-01
4.62372229e-02 -6.85554743e-02 -2.65301019e-01 -7.12368190e-01
1.26291648e-01 -4.06828076e-01 -3.55022490e-01 4.17975038e-01
5.68783656e-02 9.90026221e-02 2.00582951e-01 -3.39194238e-02
1.12412214e+00 -1.82484180e-01 7.67876148e-01 -9.35588777e-02
3.82849932e-01 5.14786720e-01 7.39975810e-01 8.32615674e-01
-7.01961145e-02 7.72096217e-01 4.47058767e-01 -6.70117497e-01
-9.58062470e-01 -9.72431242e-01 -3.26223254e-01 9.32631612e-01
-4.62559164e-01 -5.63431084e-01 -5.36977232e-01 -7.84899056e-01
1.19478077e-01 1.20435596e+00 -7.20031500e-01 -2.56777942e-01
-3.25074792e-01 -9.86284018e-01 4.32849050e-01 7.79924035e-01
-2.29982644e-01 -9.38555598e-01 -4.67828631e-01 3.41731906e-01
-1.43893719e-01 -1.02234018e+00 2.39992529e-01 2.82313406e-01
-9.27119792e-01 -1.29241323e+00 -1.97309986e-01 -3.20980012e-01
4.15223807e-01 -4.07873839e-01 1.01395977e+00 2.64113490e-02
-3.18168670e-01 3.50870907e-01 -6.07319832e-01 -1.16590214e+00
-3.19534719e-01 1.28183775e-02 2.50027180e-01 -9.86078829e-02
6.07330918e-01 -4.72668827e-01 -3.86114061e-01 -1.44429887e-02
-7.31822252e-01 -9.69298929e-02 7.43777394e-01 6.44732893e-01
2.96945006e-01 1.54267894e-02 1.05954468e+00 -9.15143311e-01
5.56319535e-01 -8.47047031e-01 -1.66953832e-01 1.48416325e-01
-1.24995470e+00 -2.24694908e-01 7.90597677e-01 -1.60555944e-01
-9.28684533e-01 1.54065369e-02 -2.86922663e-01 6.22813627e-02
-8.77381563e-01 4.81554598e-01 1.84360951e-01 4.18947905e-01
7.51870036e-01 -1.79641694e-01 1.14104182e-01 -4.65226144e-01
-1.32430285e-01 7.12010503e-01 3.36562663e-01 -3.13707739e-01
6.51008427e-01 2.44632274e-01 1.98227689e-01 -2.91857243e-01
-9.35871363e-01 -4.93525803e-01 -8.13172281e-01 -5.58152348e-02
8.34212124e-01 -6.90985501e-01 -8.38747978e-01 1.39094546e-01
-1.05934441e+00 4.95458655e-02 -4.71266031e-01 1.21102369e+00
-3.79268467e-01 3.30472551e-02 -5.94497919e-01 -1.16836762e+00
-4.19226140e-01 -6.23177409e-01 7.76800156e-01 -2.71220896e-02
-8.44421327e-01 -9.08865392e-01 1.73727542e-01 5.96085191e-01
3.48288625e-01 7.25162923e-01 1.19754326e+00 -1.09982133e+00
6.21838644e-02 -7.31443167e-01 -1.97566718e-01 2.43499786e-01
2.28151351e-01 2.21084371e-01 -6.44211233e-01 3.65283005e-02
4.19951528e-02 -2.12255046e-01 5.57789743e-01 2.48371422e-01
1.01357853e+00 -4.77614015e-01 -4.04456705e-01 6.15504533e-02
1.25975120e+00 5.47485113e-01 5.30932844e-01 7.19414711e-01
4.24212605e-01 1.12140715e+00 9.02078032e-01 4.59611505e-01
6.01377070e-01 8.37422729e-01 2.55029708e-01 -3.81291062e-01
7.78248608e-02 -2.03098021e-02 1.77757934e-01 3.55941623e-01
-4.37839448e-01 2.54137218e-01 -1.17729497e+00 5.30401826e-01
-1.82180536e+00 -9.54820335e-01 -7.28858888e-01 2.29970312e+00
6.66569173e-01 1.53512985e-01 3.97339821e-01 4.69341159e-01
4.36611563e-01 -3.64398688e-01 -1.05506487e-01 -8.54038298e-01
1.40144646e-01 3.60295326e-01 5.95881820e-01 2.06691265e-01
-9.87459004e-01 2.98763633e-01 7.28161097e+00 4.20806646e-01
-8.40772927e-01 7.24893957e-02 5.96711695e-01 -4.35164690e-01
1.93919808e-01 3.14668678e-02 -4.58733469e-01 6.29442096e-01
1.18595850e+00 -2.70815611e-01 4.13047597e-02 8.24246526e-01
4.29855227e-01 -2.24185675e-01 -1.06220007e+00 4.08583581e-01
1.29576564e-01 -1.09423256e+00 -1.24253534e-01 -6.24999143e-02
2.42784426e-01 -4.12848771e-01 -4.61584151e-01 3.66710186e-01
2.39406168e-01 -9.88845885e-01 7.34554589e-01 8.80906224e-01
6.08164907e-01 -6.78549767e-01 1.22482669e+00 2.40070984e-01
-5.80615699e-01 -5.74227989e-01 -1.19287977e-02 -4.80290294e-01
2.67483056e-01 8.26371551e-01 -1.00031483e+00 9.62215126e-01
9.90830898e-01 4.90675330e-01 -6.77915335e-01 9.73092437e-01
-2.04787031e-01 1.14695144e+00 6.54491130e-04 -1.60878841e-02
-2.68751904e-02 2.68205643e-01 4.07498389e-01 1.33293951e+00
2.11887926e-01 1.87246054e-01 7.28946626e-02 5.21749496e-01
6.44967735e-01 3.43265980e-01 -6.47070050e-01 4.31664407e-01
2.59446919e-01 1.13413334e+00 -4.50785488e-01 -2.82206893e-01
-6.40743077e-01 2.40520269e-01 1.13425307e-01 1.42478660e-01
-7.73400664e-01 -2.73756951e-01 2.54814982e-01 5.72005570e-01
-1.59531444e-01 7.79717267e-02 -8.43415320e-01 -6.21598303e-01
2.13208105e-02 -5.44503212e-01 8.34890068e-01 -6.46682560e-01
-1.56958377e+00 3.96610111e-01 3.45412016e-01 -1.33974910e+00
-3.11906397e-01 -5.10567188e-01 -6.02047682e-01 9.76493120e-01
-1.07876539e+00 -9.07820821e-01 -1.46784723e-01 3.85384053e-01
2.39986554e-01 -2.84493059e-01 1.10386217e+00 3.69206071e-01
-6.74881518e-01 4.43226337e-01 -1.79727107e-01 1.50933832e-01
8.76984119e-01 -1.20025265e+00 -1.68198913e-01 3.64085227e-01
-3.44690740e-01 4.86597925e-01 6.19454086e-01 -8.05138767e-01
-7.29329526e-01 -1.07032824e+00 1.24609303e+00 -8.16787422e-01
7.19021320e-01 -2.92056024e-01 -7.92601287e-01 5.08599818e-01
5.05453385e-02 -5.10263264e-01 1.31186867e+00 3.40695798e-01
4.37195972e-02 -2.56481409e-01 -1.35045016e+00 3.18464756e-01
7.54501998e-01 -4.59366888e-02 -9.66660976e-01 3.18470180e-01
5.52368760e-01 -1.84863228e-02 -1.17985761e+00 6.95775747e-01
6.65087759e-01 -6.82084918e-01 8.15916955e-01 -1.22962618e+00
6.92708015e-01 -1.49890017e-02 9.23052952e-02 -1.17273343e+00
-6.07049823e-01 5.33835366e-02 1.98486377e-03 1.54693556e+00
7.30992556e-01 -7.46359885e-01 4.58443463e-01 1.35669219e+00
-1.49764046e-01 -1.00695884e+00 -6.55924320e-01 -7.15230346e-01
1.10186689e-01 -6.01492107e-01 6.88343227e-01 1.15209436e+00
3.43787998e-01 2.64199853e-01 -3.74620467e-01 -1.79586560e-03
4.61344182e-01 -3.51806842e-02 5.16607165e-01 -1.60860562e+00
-6.66226074e-02 -2.45240971e-01 -5.90491474e-01 2.06844911e-01
7.84207210e-02 -7.90107012e-01 -3.26810092e-01 -1.71931767e+00
3.15183848e-01 -4.86833423e-01 -6.15961671e-01 9.82767045e-01
-2.47144029e-01 1.55037016e-01 1.45124212e-01 3.26212309e-02
-1.90444097e-01 1.49875775e-01 7.04184413e-01 9.55404639e-02
-1.22227922e-01 1.28979623e-01 -1.10849905e+00 7.47447133e-01
9.42121625e-01 -9.08242881e-01 -1.76127911e-01 8.41094274e-03
8.84141251e-02 3.02177727e-01 5.68191350e-01 -9.50980902e-01
1.37967065e-01 -4.76451308e-01 6.74483597e-01 -2.35903278e-01
3.72328274e-02 -1.08693397e+00 3.17951530e-01 3.41265351e-01
-5.73632836e-01 1.98127255e-01 3.02942067e-01 2.66915977e-01
-2.20951319e-01 -3.59923840e-01 2.19548777e-01 2.77490113e-02
-4.08533156e-01 -8.05538446e-02 -4.66508776e-01 -4.63477999e-01
1.54560435e+00 -4.53936964e-01 -4.95698094e-01 -1.00520104e-01
-1.05227685e+00 9.88273025e-02 2.03973725e-01 5.43028116e-01
2.44504720e-01 -1.27635086e+00 -8.86770070e-01 -6.91229850e-02
3.42406988e-01 -3.20672244e-01 1.06421441e-01 1.17932308e+00
-1.08140983e-01 5.79879403e-01 -3.38625222e-01 -1.69033095e-01
-1.06843388e+00 9.30474460e-01 -2.93313265e-01 -4.28091139e-01
-5.50410688e-01 1.74776167e-01 -5.89058809e-02 -2.04933271e-01
5.17920824e-04 -1.40145300e-02 -5.72976351e-01 3.17753762e-01
5.62050462e-01 6.50018513e-01 3.11084688e-01 -5.78985095e-01
-4.33766574e-01 3.21604580e-01 1.28699318e-02 6.20731488e-02
1.69023526e+00 1.54845476e-01 -1.53342769e-01 6.14004552e-01
8.38880777e-01 1.15127467e-01 -5.78091800e-01 1.65182501e-01
1.80350035e-01 -3.17776442e-01 -1.90986097e-01 -1.12650526e+00
-9.05404985e-01 5.48914313e-01 4.28190738e-01 3.96154910e-01
1.27964175e+00 2.56464891e-02 5.41295469e-01 9.17344261e-03
4.37180758e-01 -1.11182809e+00 -4.89739388e-01 2.11948007e-01
9.07826126e-01 -1.10652697e+00 1.65730789e-01 -3.80464971e-01
-8.13103259e-01 1.04168081e+00 5.51323831e-01 9.79749262e-02
5.73633194e-01 3.03527385e-01 -6.67427182e-02 -1.68088138e-01
-7.76391625e-01 -5.82668558e-03 2.78657436e-01 6.15991414e-01
4.99520123e-01 9.36043859e-02 -9.49605405e-01 1.18294561e+00
-3.83643210e-01 2.02937990e-01 3.72753710e-01 8.63990366e-01
-5.01457751e-02 -9.77940261e-01 -6.56787992e-01 1.10897887e+00
-7.32972264e-01 -4.65265363e-02 -4.65199441e-01 8.76755714e-01
3.83946568e-01 1.38334262e+00 -7.06276074e-02 -6.10494316e-01
7.98797607e-01 3.54547054e-01 -1.36688262e-01 -8.28966200e-01
-8.89158666e-01 -6.96030617e-01 2.26197779e-01 -6.34041667e-01
-1.57264173e-01 -7.47433662e-01 -1.06444442e+00 -1.21020302e-01
-2.57720262e-01 3.62595469e-01 6.55130446e-01 9.49734569e-01
2.89217323e-01 6.51984394e-01 7.04831600e-01 -5.32148838e-01
-3.28987539e-01 -9.72896516e-01 -5.71309686e-01 7.50665665e-01
6.98793381e-02 -8.49318147e-01 -4.06094611e-01 1.68716878e-01] | [8.139104843139648, 5.198251247406006] |
c8ca79e4-134d-4ac4-b787-9e24dd441e43 | robust-reflection-removal-with-reflection | 2103.04273 | null | https://arxiv.org/abs/2103.04273v2 | https://arxiv.org/pdf/2103.04273v2.pdf | Robust Reflection Removal with Reflection-free Flash-only Cues | We propose a simple yet effective reflection-free cue for robust reflection removal from a pair of flash and ambient (no-flash) images. The reflection-free cue exploits a flash-only image obtained by subtracting the ambient image from the corresponding flash image in raw data space. The flash-only image is equivalent to an image taken in a dark environment with only a flash on. We observe that this flash-only image is visually reflection-free, and thus it can provide robust cues to infer the reflection in the ambient image. Since the flash-only image usually has artifacts, we further propose a dedicated model that not only utilizes the reflection-free cue but also avoids introducing artifacts, which helps accurately estimate reflection and transmission. Our experiments on real-world images with various types of reflection demonstrate the effectiveness of our model with reflection-free flash-only cues: our model outperforms state-of-the-art reflection removal approaches by more than 5.23dB in PSNR, 0.04 in SSIM, and 0.068 in LPIPS. Our source code and dataset are publicly available at {github.com/ChenyangLEI/flash-reflection-removal}. | ['Qifeng Chen', 'Chenyang Lei'] | 2021-03-07 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Lei_Robust_Reflection_Removal_With_Reflection-Free_Flash-Only_Cues_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Lei_Robust_Reflection_Removal_With_Reflection-Free_Flash-Only_Cues_CVPR_2021_paper.pdf | cvpr-2021-1 | ['reflection-removal'] | ['computer-vision'] | [ 7.04195023e-01 -3.88882279e-01 4.71937358e-01 5.09569719e-02
-9.81723785e-01 -4.75570917e-01 2.46204272e-01 -4.27696943e-01
-1.84639677e-01 4.74839836e-01 2.40769699e-01 -3.74972761e-01
2.60476857e-01 -3.91947120e-01 -7.29355454e-01 -1.11565053e+00
2.24419340e-01 -8.05712044e-01 5.17613351e-01 -1.00799263e-01
2.81272680e-01 1.16054527e-01 -1.52385437e+00 2.96395659e-01
8.85777891e-01 1.16475976e+00 6.73843861e-01 8.63565981e-01
4.34894353e-01 7.94876039e-01 -8.12252760e-01 1.05603419e-01
6.15349412e-01 -4.14410055e-01 8.81498232e-02 1.83602665e-02
7.90233970e-01 -7.94765592e-01 -8.38751137e-01 1.09195590e+00
6.09668374e-01 -6.30725687e-03 2.07310662e-01 -9.22357142e-01
-7.37123907e-01 -1.95822090e-01 -9.84052420e-01 2.89186925e-01
7.27767766e-01 3.43665957e-01 2.80208021e-01 -1.08250141e+00
2.25091487e-01 9.25016284e-01 5.85723698e-01 1.77524492e-01
-9.89497006e-01 -8.38858068e-01 -2.87962258e-01 2.31651619e-01
-1.49797523e+00 -7.49278784e-01 7.69638777e-01 6.93346560e-02
6.35917306e-01 5.01099586e-01 4.01439786e-01 9.45541501e-01
4.01227504e-01 5.62785029e-01 1.45424604e+00 -5.81428111e-01
-1.12074055e-01 -5.13758361e-02 1.42777503e-01 6.12591803e-01
4.90134686e-01 3.29700828e-01 -6.60439134e-01 -4.69645374e-02
7.17846870e-01 2.93579459e-01 -1.07270873e+00 2.26192117e-01
-1.25405586e+00 -1.26083121e-01 3.27576011e-01 -2.49660779e-02
-1.36502907e-01 1.61268160e-01 -2.51409918e-01 4.85862792e-01
3.99307281e-01 1.24806628e-01 5.91473058e-02 1.45314038e-01
-8.39980662e-01 -1.85283095e-01 5.60702085e-01 1.11781883e+00
6.68646991e-01 2.13249639e-01 -2.90552348e-01 8.25157344e-01
1.38878554e-01 1.42057526e+00 -1.89319670e-01 -1.25973189e+00
5.62141120e-01 -7.18938112e-02 6.46603644e-01 -1.02097058e+00
-1.00530580e-01 -2.26811573e-01 -9.04566526e-01 5.14470339e-01
2.16390252e-01 -7.95986727e-02 -9.06095445e-01 1.47959805e+00
3.14038247e-02 5.83818734e-01 4.43840176e-02 1.24329102e+00
9.23089981e-01 7.85728872e-01 -6.80689335e-01 -4.92313594e-01
1.13087511e+00 -9.66955900e-01 -9.23196018e-01 -2.62708992e-01
-6.97402656e-03 -9.90476906e-01 1.20266080e+00 8.77392471e-01
-1.16250587e+00 -5.99852920e-01 -1.42689586e+00 -1.41662527e-02
3.55674714e-01 2.86275148e-01 1.24531582e-01 7.07036614e-01
-1.07451308e+00 9.28256884e-02 -4.95109230e-01 -1.82012156e-01
-5.01185879e-02 -1.75136134e-01 -2.42981806e-01 -7.30788469e-01
-9.74321187e-01 7.21240163e-01 -5.57921290e-01 2.21965313e-01
-6.40741765e-01 -7.44191647e-01 -4.86213863e-01 -1.20155543e-01
7.44837463e-01 -4.07620847e-01 9.42465007e-01 -9.81561303e-01
-1.37763941e+00 5.60836613e-01 -6.28664076e-01 4.22544517e-02
3.61454129e-01 -4.49250400e-01 -7.58156478e-01 5.67034721e-01
-2.32982010e-01 -2.16056705e-01 1.19184542e+00 -1.92683566e+00
-2.17548415e-01 -1.70895964e-01 -5.36423586e-02 9.97623727e-02
-1.42849922e-01 5.09923361e-02 -9.67771649e-01 -3.79544556e-01
1.54990941e-01 -8.76656294e-01 1.42194912e-01 1.89146876e-01
-5.63814044e-01 7.81189680e-01 9.71921802e-01 -7.05019057e-01
1.54859078e+00 -2.35002327e+00 -7.54616618e-01 -8.60106125e-02
3.29707354e-01 5.47268987e-01 -4.27025288e-01 5.65715134e-01
5.36823757e-02 -2.37140179e-01 -2.67616898e-01 -3.31186384e-01
-4.83140260e-01 -2.58375704e-01 -4.81000721e-01 7.42529631e-01
-2.24532887e-01 3.98466617e-01 -9.30582404e-01 -1.55750260e-01
3.81962746e-01 8.16161454e-01 -2.32056994e-02 3.15322727e-01
5.08165479e-01 5.18071175e-01 -2.81424761e-01 6.97719753e-01
1.39992714e+00 -5.89228906e-02 9.40928310e-02 -4.98448223e-01
-4.75782841e-01 -1.78614885e-01 -1.12883461e+00 1.28960323e+00
-6.90744102e-01 1.00531352e+00 9.31757465e-02 -3.61258313e-02
1.01856875e+00 -2.49157473e-02 1.44795299e-01 -1.32360744e+00
-6.35141805e-02 3.00158858e-01 -2.83600301e-01 -6.14304602e-01
6.68755233e-01 3.81310731e-01 1.85599476e-01 6.65128946e-01
-6.40652478e-01 8.51630121e-02 -4.80311997e-02 2.78313667e-01
1.58749712e+00 1.88940042e-03 -3.77138630e-02 -4.67338711e-02
4.43958610e-01 -7.77168632e-01 5.18236995e-01 1.06172943e+00
-1.79627120e-01 1.34117603e+00 6.54668584e-02 2.98505207e-03
-7.84611762e-01 -1.43657255e+00 1.12900160e-01 7.28469133e-01
8.07219028e-01 -5.16526878e-01 -6.33031547e-01 -9.82659683e-02
-2.88925737e-01 7.60265470e-01 -3.19404691e-01 -2.44578943e-01
-5.64047277e-01 -3.59091014e-01 3.31663877e-01 1.58465371e-01
8.80473018e-01 -6.36342645e-01 -8.02698791e-01 -2.92917877e-01
-6.86243892e-01 -1.24411190e+00 -6.63323879e-01 -1.35620087e-01
-4.96245742e-01 -1.24735594e+00 -9.40491736e-01 -2.10738778e-01
7.59983420e-01 1.55988085e+00 1.09697282e+00 4.77625519e-01
-5.90092540e-01 6.47513092e-01 -2.42394477e-01 -3.22753459e-01
-1.80627145e-02 -1.00340307e+00 -2.24074155e-01 3.03136587e-01
1.29094319e-02 -5.72412491e-01 -1.37520552e+00 5.61561167e-01
-1.05354071e+00 1.58789456e-01 5.68609476e-01 4.72669095e-01
4.08884645e-01 -6.63931370e-02 4.20162082e-02 -4.82907444e-01
3.96232754e-01 -1.68710679e-01 -5.60755014e-01 1.10006280e-01
-3.64392281e-01 -3.50495279e-01 5.68594933e-01 -2.95405358e-01
-1.53410196e+00 -2.13578582e-01 3.95057559e-01 -6.75422490e-01
3.63353714e-02 -2.64768839e-01 -2.52256274e-01 -6.45888820e-02
6.08925343e-01 2.45774895e-01 -1.76691711e-01 -4.71506447e-01
1.32836357e-01 8.56236458e-01 7.85903215e-01 -1.30907848e-01
8.20353806e-01 9.76874650e-01 1.56293780e-01 -1.39402986e+00
-8.21135283e-01 -6.32761359e-01 -2.40347892e-01 -4.14935410e-01
3.92709017e-01 -1.07967770e+00 -8.08817148e-01 8.98094893e-01
-1.02188933e+00 -7.18586922e-01 1.85095891e-01 3.22064817e-01
-3.78472179e-01 6.86625838e-01 -5.01840353e-01 -9.95827198e-01
-2.73053825e-01 -8.00052106e-01 1.17732215e+00 3.49836051e-01
2.66266227e-01 -5.95598042e-01 6.42932504e-02 2.65350461e-01
4.99444157e-01 -1.61352798e-01 4.69745874e-01 4.58329737e-01
-9.73304451e-01 -1.21309966e-01 -6.27273142e-01 5.44979930e-01
2.64726400e-01 -1.85340717e-01 -1.31928933e+00 -4.48362619e-01
1.50061488e-01 -4.88530882e-02 1.11389899e+00 3.90859485e-01
9.44332123e-01 7.44251767e-04 -2.00548589e-01 8.96152973e-01
1.81101251e+00 2.01922819e-01 1.43534482e+00 2.72022873e-01
6.26113892e-01 2.84664929e-01 9.64067578e-01 5.18433750e-01
2.30201371e-02 8.10692132e-01 4.75357175e-01 -5.61287165e-01
-7.33922124e-01 2.96022836e-03 5.24793983e-01 5.14912844e-01
-1.71431765e-01 -8.27565908e-01 -4.60561693e-01 1.82156920e-01
-1.62405312e+00 -1.00112629e+00 -6.53894961e-01 2.77750897e+00
5.04285395e-01 -3.04211825e-01 -4.78809744e-01 8.54211524e-02
5.29497087e-01 5.27499676e-01 -4.87820745e-01 -1.91857871e-02
-6.16323233e-01 1.49103001e-01 7.62491822e-01 7.97585070e-01
-5.96280992e-01 5.13639927e-01 6.04530430e+00 4.19020951e-01
-1.23230159e+00 -9.99608263e-03 2.62958497e-01 -2.90180564e-01
-4.68876690e-01 -1.08729422e-01 -4.28921431e-01 7.51313925e-01
7.74800479e-01 7.64186010e-02 3.22148174e-01 -8.96007493e-02
8.73606801e-01 -1.04029417e+00 -7.13780999e-01 1.24727952e+00
5.35652518e-01 -6.82256699e-01 -2.92712778e-01 1.02050742e-02
5.98914564e-01 -2.05189198e-01 4.25651103e-01 -3.42283398e-01
-1.19255826e-01 -7.57292926e-01 4.93885905e-01 1.04480278e+00
1.18434036e+00 -5.11363804e-01 3.82457882e-01 -5.94830699e-02
-1.01932395e+00 2.17410363e-02 -3.43408763e-01 2.35395148e-01
1.86567962e-01 9.27615345e-01 -2.09132344e-01 5.73365092e-01
9.97703373e-01 6.64070070e-01 -4.11154509e-01 1.05993974e+00
-3.97223502e-01 6.94398999e-01 -2.29660451e-01 5.24695516e-01
-3.83279949e-01 -5.07773459e-01 7.60562181e-01 1.25317419e+00
6.85499012e-01 3.95189703e-01 -6.95229769e-02 5.87715507e-01
2.75590837e-01 -5.12066185e-01 -5.63686013e-01 5.93687296e-01
5.59864998e-01 1.26652026e+00 -2.93622017e-01 -3.89727205e-02
-6.61424160e-01 1.46548498e+00 -4.57805097e-01 1.07716346e+00
-8.21060121e-01 -7.34382570e-01 6.54224753e-01 3.58029127e-01
3.12878102e-01 -3.40441853e-01 9.03394744e-02 -9.64182138e-01
3.89631659e-01 -7.36160457e-01 -1.44375056e-01 -1.60747528e+00
-8.32835197e-01 4.76363838e-01 -2.00964302e-01 -1.53165364e+00
3.64990979e-01 -5.31121135e-01 -7.23225832e-01 1.02533948e+00
-1.84020710e+00 -8.41077924e-01 -9.53181863e-01 8.21807981e-01
2.46820658e-01 3.52828532e-01 4.33516413e-01 4.12652820e-01
-4.07626510e-01 4.62406129e-01 6.11694217e-01 -1.49595931e-01
1.29160285e+00 -8.21701169e-01 3.96200977e-02 1.28948104e+00
-2.31572390e-01 6.17833614e-01 8.98026466e-01 -5.02589881e-01
-1.92670357e+00 -8.36710274e-01 2.85398215e-01 -1.88527271e-01
3.05946320e-01 -3.15467685e-01 -9.30513620e-01 3.67691875e-01
3.72592270e-01 2.15635702e-01 3.88636470e-01 -4.93129373e-01
-7.32007802e-01 -5.39822400e-01 -8.90313148e-01 8.63398373e-01
1.03429568e+00 -4.84487087e-01 -1.06683865e-01 -1.36917476e-02
6.15905762e-01 -1.68574974e-01 -2.23866194e-01 9.15968791e-02
6.91684067e-01 -1.71834958e+00 1.20831823e+00 5.21809638e-01
2.96825230e-01 -5.86580157e-01 -3.60444188e-01 -1.13363528e+00
-6.83743358e-02 -1.12931168e+00 -1.68305680e-01 9.61995363e-01
7.12295175e-02 -9.45641041e-01 2.74473041e-01 2.04739809e-01
-2.50933260e-01 -2.77695060e-01 -3.46031219e-01 -8.76021028e-01
-6.31267726e-01 -5.11879385e-01 1.21592589e-01 3.95593286e-01
-4.24207360e-01 -1.25866802e-03 -8.26121807e-01 3.46468210e-01
1.19274676e+00 2.88846642e-01 1.01641023e+00 -6.49987042e-01
-3.37322026e-01 3.69649738e-01 9.29672644e-02 -1.41528857e+00
-2.08769456e-01 -8.52071270e-02 4.01254833e-01 -1.55185127e+00
5.26483834e-01 -2.00247437e-01 -3.80958617e-01 1.84929490e-01
-4.29576904e-01 6.73261285e-01 3.92816067e-01 3.48291874e-01
-7.12295055e-01 3.63345385e-01 1.13394415e+00 1.79016933e-01
-2.68996179e-01 8.57375562e-03 -8.45885575e-01 6.88522935e-01
5.22011518e-01 -1.92761526e-01 -5.01835346e-01 -5.77581584e-01
2.11835265e-01 2.23737940e-01 5.92371225e-01 -9.90531087e-01
2.24615544e-01 3.55525687e-02 3.34869146e-01 -6.34882748e-01
7.49189138e-01 -7.73263633e-01 1.04587689e-01 1.38710439e-01
2.97246873e-02 -4.10015374e-01 1.89157575e-01 9.14168000e-01
-4.23215777e-02 5.73795363e-02 7.49689281e-01 -1.60384476e-02
-6.16881549e-01 1.01724319e-01 -3.42557043e-01 2.90811975e-02
6.71011627e-01 -5.07545114e-01 -1.04269636e+00 -9.26713526e-01
-3.16812731e-02 -2.45321289e-01 7.26939976e-01 1.00498565e-01
9.74402726e-01 -1.00871396e+00 -6.24580860e-01 2.60682851e-01
1.46598592e-01 -5.84125936e-01 6.20035231e-01 1.24302685e+00
-3.79512101e-01 1.54417127e-01 7.46621415e-02 -6.22351825e-01
-1.82123101e+00 5.50748825e-01 1.08125426e-01 1.33445829e-01
-8.30070913e-01 4.30740297e-01 6.82845831e-01 4.80004519e-01
6.95850104e-02 -1.37662709e-01 3.25796306e-01 -5.84443212e-01
1.12682843e+00 8.21240187e-01 1.54801439e-02 -5.12674093e-01
-1.54807091e-01 7.04054952e-01 3.00822407e-01 -2.93700993e-01
1.08355081e+00 -8.54716063e-01 1.28558114e-01 5.39178789e-01
1.28961134e+00 6.68026149e-01 -1.43928099e+00 -4.65509087e-01
-7.25530148e-01 -1.21416545e+00 2.64947057e-01 -8.77504170e-01
-1.09281445e+00 9.20479476e-01 6.42893195e-01 -6.65575638e-02
1.78755033e+00 -4.31393892e-01 7.60958850e-01 3.25898349e-01
4.40073848e-01 -6.71098650e-01 5.23375988e-01 3.71517122e-01
9.98040438e-01 -1.02730310e+00 3.14031899e-01 -6.29243195e-01
-4.33244944e-01 1.04798698e+00 3.25929642e-01 5.80557212e-02
4.57522124e-01 7.62044489e-01 3.96336168e-01 1.53938606e-01
-6.72056079e-01 -3.33930314e-01 1.16106346e-01 8.65925312e-01
2.07780674e-01 -3.99399042e-01 2.22793669e-01 1.91779867e-01
3.85526329e-01 -1.23005345e-01 8.50073338e-01 9.66909587e-01
-6.17767036e-01 -6.29725039e-01 -8.57573092e-01 3.36343572e-02
-3.42450768e-01 -4.17226970e-01 -3.23380083e-01 5.47951698e-01
-3.91160041e-01 1.52262652e+00 1.56151764e-02 -3.49621624e-01
3.89786482e-01 -6.51781380e-01 7.14968979e-01 -1.15720995e-01
-3.11049186e-02 4.50732797e-01 -3.29893231e-02 -1.03685772e+00
-4.70925421e-01 -3.53493392e-01 -1.00472152e+00 -3.14955324e-01
-2.50250310e-01 -2.51007140e-01 5.10384917e-01 2.43071914e-01
5.11731684e-01 4.22468007e-01 1.00880086e+00 -8.80028725e-01
-1.86150596e-02 -7.10277617e-01 -9.62070286e-01 2.98548967e-01
8.10380340e-01 -3.65651101e-01 -9.66585398e-01 9.76481363e-02] | [10.509469985961914, -2.6640737056732178] |
746a4c75-95d5-49be-98b8-ddf0590b4982 | the-power-of-many-a-physarum-swarm-steiner | 2110.08233 | null | https://arxiv.org/abs/2110.08233v1 | https://arxiv.org/pdf/2110.08233v1.pdf | The Power of Many: A Physarum Swarm Steiner Tree Algorithm | We create a novel Physarum Steiner algorithm designed to solve the Euclidean Steiner tree problem. Physarum is a unicellular slime mold with the ability to form networks and fuse with other Physarum organisms. We use the simplicity and fusion of Physarum to create large swarms which independently operate to solve the Steiner problem. The Physarum Steiner tree algorithm then utilizes a swarm of Physarum organisms which gradually find terminals and fuse with each other, sharing intelligence. The algorithm is also highly capable of solving the obstacle avoidance Steiner tree problem and is a strong alternative to the current leading algorithm. The algorithm is of particular interest due to its novel approach, rectilinear properties, and ability to run on varying shapes and topological surfaces. | ['Laura P. Schaposnik', 'Fidel I. Schaposnik Massolo', 'Sheryl Hsu'] | 2021-10-15 | null | null | null | null | ['steiner-tree-problem'] | ['graphs'] | [ 6.10003807e-02 1.03529036e-01 4.62660104e-01 3.02854002e-01
5.39520800e-01 -1.26032937e+00 5.11996627e-01 2.05900222e-01
-3.49782497e-01 8.99548829e-01 -6.91982031e-01 -3.15115243e-01
-5.55735826e-01 -9.60721552e-01 -2.60766029e-01 -9.69355583e-01
-6.99485838e-01 9.38789129e-01 7.79356658e-01 -5.44962645e-01
4.85078394e-01 8.53713512e-01 -1.11229873e+00 -5.06344557e-01
1.11915159e+00 4.02909786e-01 8.65358114e-01 1.06227422e+00
-1.46281406e-01 3.40042007e-03 -9.35390711e-01 4.90247644e-02
4.14715677e-01 -2.71512300e-01 -6.95329666e-01 8.70262086e-02
-4.41769004e-01 5.15702128e-01 7.23037198e-02 5.13616145e-01
1.09114200e-01 -5.34521155e-02 3.87898982e-01 -1.28693032e+00
-3.45639855e-01 1.75193146e-01 -3.04527164e-01 3.13039929e-01
5.64658642e-01 2.26549078e-02 3.97063822e-01 -3.20006579e-01
1.06653893e+00 1.12753654e+00 4.38541353e-01 5.42329788e-01
-5.93871236e-01 -3.01496148e-01 -2.30249003e-01 -3.81293833e-01
-1.25298262e+00 -1.94776449e-02 3.36369455e-01 5.04521169e-02
9.78814781e-01 7.15714395e-01 1.16976440e+00 7.41929770e-01
1.05247545e+00 1.23109065e-01 1.29176056e+00 -1.63770139e-01
8.19181323e-01 -2.68970907e-01 -2.76193768e-01 8.50182712e-01
6.24669909e-01 -2.45684043e-01 -7.55432025e-02 -3.61399293e-01
1.28326225e+00 -7.78245255e-02 -2.40030795e-01 -4.47808415e-01
-1.01625025e+00 4.12579864e-01 6.70406342e-01 7.86918521e-01
-1.58285007e-01 1.88757941e-01 -2.50145227e-01 3.37856174e-01
-2.11227804e-01 1.29247034e+00 -5.08106530e-01 1.75080057e-02
-2.62467116e-01 4.76195216e-01 1.47512710e+00 8.13738346e-01
2.89613217e-01 1.43290281e-01 6.53034866e-01 2.96055645e-01
3.09829831e-01 5.67262292e-01 7.40307719e-02 -1.34897149e+00
-1.21611301e-02 9.89540398e-01 1.97994247e-01 -1.01490068e+00
-7.27777600e-01 -5.39497256e-01 -5.97794116e-01 5.58491170e-01
1.33092508e-01 -1.01889081e-01 -8.81283939e-01 1.13927221e+00
6.31404638e-01 2.52148032e-01 3.32867652e-01 6.06920481e-01
6.00025833e-01 7.35692680e-01 -2.64471859e-01 -3.96746755e-01
9.88720775e-01 -7.63338149e-01 -3.03717405e-01 4.00840163e-01
4.76360649e-01 -7.02975810e-01 3.39124739e-01 3.37974459e-01
-1.32937455e+00 1.96579881e-02 -1.04265976e+00 7.84777582e-01
-7.51625001e-01 -8.92880738e-01 7.57428885e-01 6.26129866e-01
-1.70324314e+00 6.91898644e-01 -1.05105734e+00 -9.08090353e-01
-1.03695735e-01 8.58894229e-01 -1.42285690e-01 3.64401698e-01
-5.20202756e-01 8.27644289e-01 2.36496747e-01 -5.44877112e-01
-3.22192878e-01 -1.62492782e-01 -3.28542322e-01 -5.25055081e-02
2.74061024e-01 -1.03651273e+00 6.62868738e-01 -1.63979664e-01
-1.51406479e+00 6.85271919e-01 -2.08804950e-01 -2.44191006e-01
2.34136268e-01 7.14371085e-01 -1.82404637e-01 3.59252006e-01
9.17165074e-03 6.64570928e-01 2.91516110e-02 -1.72591627e+00
-5.20438910e-01 -2.94173509e-01 1.48844168e-01 -7.54467770e-02
-1.44769013e-01 -1.16542261e-02 -1.53039694e-01 -5.46720624e-01
5.62038124e-01 -1.17541265e+00 -7.84407258e-01 -3.24990124e-01
-2.01700017e-01 -4.03648704e-01 1.14764118e+00 5.35424240e-02
7.37004519e-01 -1.59797430e+00 4.80437815e-01 5.69887280e-01
5.29872656e-01 3.48220944e-01 -3.50068599e-01 8.88671756e-01
5.76152205e-01 4.39586490e-01 -3.13920230e-01 1.16925851e-01
-5.08671463e-01 8.31406891e-01 2.47687712e-01 3.23057175e-01
-2.91382037e-02 6.83935404e-01 -1.20702028e+00 -3.96730185e-01
-1.07927121e-01 1.62832469e-01 -3.42104971e-01 -2.60534585e-01
-1.38624772e-01 7.78977633e-01 -9.67049181e-01 9.42672968e-01
7.27523744e-01 -3.71599197e-01 2.61197299e-01 6.13233149e-01
-8.34809005e-01 -2.93690920e-01 -9.24714088e-01 1.50980043e+00
-6.69516027e-02 1.81427985e-01 7.34757125e-01 -5.55941939e-01
1.33695543e+00 1.68841660e-01 1.16945148e+00 -1.39374793e-01
4.14213985e-01 5.54546356e-01 1.00126691e-01 -3.44651043e-01
3.13345373e-01 3.59473974e-01 1.48520932e-01 3.69307160e-01
-2.50608087e-01 -8.68105352e-01 6.20257378e-01 2.27517337e-01
1.85114014e+00 -1.99835211e-01 2.82452144e-02 -9.71126318e-01
4.33399469e-01 2.95035064e-01 7.30095208e-01 3.79816025e-01
6.46097760e-05 8.83124173e-02 6.35974528e-03 -6.41692996e-01
-1.00205529e+00 -1.80821252e+00 -1.84252158e-01 6.72717988e-02
9.98702586e-01 -3.49974632e-01 -6.38097167e-01 -1.63788587e-01
1.02369435e-01 3.57906491e-01 -4.33272034e-01 3.75719219e-01
-7.84102142e-01 -4.68717456e-01 1.88413858e-01 -2.14964017e-01
5.08307099e-01 -8.05606425e-01 -9.76737738e-01 5.51970541e-01
3.29824299e-01 -7.73591638e-01 3.84374261e-01 3.39876503e-01
-7.73469150e-01 -1.42372370e+00 -3.70882392e-01 -1.08522081e+00
9.88456726e-01 6.46535575e-01 8.37076426e-01 6.85841322e-01
-5.91678500e-01 5.05253077e-01 -7.59825945e-01 -5.54182529e-01
-6.53450370e-01 -7.97745809e-02 4.01434273e-01 -8.00381780e-01
-2.15934873e-01 -1.12392056e+00 -4.07989681e-01 6.52746737e-01
-7.65809715e-01 -3.40620041e-01 5.40658459e-02 2.30420485e-01
3.89626503e-01 5.67019165e-01 4.07107770e-01 -3.51365238e-01
6.01511121e-01 -7.07766652e-01 -7.31469810e-01 2.76366085e-01
-1.81755990e-01 -4.91700441e-01 8.57898951e-01 -8.15639040e-04
-4.65738773e-01 1.17127232e-01 3.86556357e-01 9.13382247e-02
1.24630876e-01 -2.86648989e-01 2.75079727e-01 -1.08377659e+00
3.34720403e-01 -1.23075038e-01 -1.03652395e-01 -4.40992415e-01
-2.82087713e-01 2.90959388e-01 5.13025522e-01 -3.86616111e-01
9.77963686e-01 7.66544580e-01 7.38539398e-01 -1.03423429e+00
2.36461848e-01 -2.90032864e-01 -5.70501924e-01 -3.67212832e-01
8.55962992e-01 -1.41416311e-01 -1.19415069e+00 1.56706735e-01
-1.36372125e+00 -1.47784770e-01 -3.45944315e-01 1.21452548e-02
-5.89377582e-01 3.58261049e-01 -5.71529806e-01 -7.72594452e-01
-1.15505405e-01 -8.19454849e-01 7.01378465e-01 6.83475912e-01
-3.98536891e-01 -1.29289818e+00 4.08204824e-01 -8.92380849e-02
5.66209078e-01 9.95541036e-01 7.32371867e-01 -5.18798709e-01
-8.27295363e-01 -8.06133077e-03 1.47584260e-01 -6.05954051e-01
5.53301036e-01 4.90734458e-01 -1.10064610e-03 -6.51026011e-01
3.72221351e-01 2.39584550e-01 4.18816596e-01 3.17647129e-01
4.30251539e-01 -1.14802442e-01 -8.61115575e-01 5.89648962e-01
1.49010623e+00 9.44328547e-01 1.11161388e-01 6.00721598e-01
1.47358999e-01 6.31398857e-01 2.75780737e-01 5.20363152e-01
2.87320375e-01 3.20785999e-01 7.84174085e-01 -6.60887435e-02
2.91423529e-01 5.09408891e-01 -3.84106189e-02 1.30368447e+00
-5.35061598e-01 -8.07408988e-01 -1.23772514e+00 1.75982282e-01
-1.50870681e+00 -5.19891918e-01 -4.64053780e-01 1.58412004e+00
2.09050864e-01 -1.05690770e-01 -3.35920081e-02 1.11562997e-01
5.05727649e-01 -3.10368687e-01 -3.50221694e-01 -7.37848997e-01
-6.23679876e-01 3.95094812e-01 6.65585399e-01 3.56281936e-01
-8.29021692e-01 1.04706240e+00 7.75412846e+00 2.49856710e-01
-8.80234420e-01 -1.85410380e-01 1.52597874e-02 1.09428786e-01
-2.59456068e-01 2.13178217e-01 -5.38934588e-01 4.22836900e-01
4.46288526e-01 -3.68010849e-01 6.03145480e-01 1.09337568e-01
2.49547482e-01 -4.67757583e-01 -4.99693125e-01 4.64985102e-01
-1.97819218e-01 -1.68164611e+00 5.01108132e-02 3.40108454e-01
1.29291975e+00 5.79026602e-02 -2.28971481e-01 -6.73207700e-01
1.25081885e+00 -9.25766408e-01 -3.45237143e-02 2.78129816e-01
4.08184342e-02 -5.24943709e-01 3.39471668e-01 4.87043858e-01
-1.40948474e+00 -4.84140515e-01 -5.66928148e-01 -4.04248774e-01
5.69945455e-01 3.55716497e-01 -7.88279712e-01 7.67629206e-01
7.65094101e-01 2.12049186e-01 -3.30354571e-01 1.55484271e+00
1.11874700e-01 -1.58995152e-01 -7.86033988e-01 -5.75603724e-01
6.15499854e-01 -8.24869752e-01 1.34030282e+00 6.46627963e-01
6.49992168e-01 2.73142874e-01 5.95040739e-01 5.94484985e-01
4.86231923e-01 9.93628800e-02 -9.07735586e-01 1.59061193e-01
8.30334306e-01 9.82131243e-01 -1.54969513e+00 -2.10029528e-01
2.33351976e-01 7.02185512e-01 -7.90490806e-02 3.38101327e-01
-3.97816718e-01 -3.57554913e-01 7.10871458e-01 8.59837234e-02
2.60545671e-01 -9.93159175e-01 -4.95375961e-01 -5.51793694e-01
-4.08202976e-01 -4.56967354e-01 8.45145136e-02 -6.32439375e-01
-7.88277984e-01 7.62016892e-01 -3.56179625e-01 -7.37123489e-01
-3.17933634e-02 -4.76202756e-01 -9.75238562e-01 3.31498742e-01
-7.32629955e-01 -9.68721390e-01 -2.35006705e-01 4.57281113e-01
5.93829870e-01 -3.44829559e-01 1.07666504e+00 -6.79583967e-01
-2.67316878e-01 -3.20451409e-01 4.83050466e-01 -5.74562073e-01
-1.00305229e-01 -1.02977133e+00 5.46132565e-01 6.38870299e-01
-3.36761624e-01 7.41450548e-01 8.18427205e-01 -1.14672065e+00
-1.92851591e+00 -8.30674887e-01 6.06186628e-01 -5.65101266e-01
8.38921666e-01 -2.65260100e-01 -3.93137902e-01 1.18825912e-01
4.04221267e-01 -6.33252144e-01 4.16510493e-01 -4.83386666e-01
4.43932384e-01 2.09421694e-01 -1.50463915e+00 8.72968078e-01
1.40531683e+00 6.43012404e-01 -3.83790940e-01 6.55874789e-01
8.69221032e-01 -2.33692512e-01 -9.99183416e-01 5.60145259e-01
2.24283859e-01 -7.44336069e-01 9.04290736e-01 -3.00155431e-01
-1.92449559e-02 -5.10293782e-01 3.62913348e-02 -1.44905937e+00
-5.92887998e-01 -1.24319208e+00 4.21669900e-01 1.00091982e+00
1.78306088e-01 -1.09487009e+00 8.82149398e-01 5.87332249e-02
-8.51947889e-02 -8.21434677e-01 -1.32293749e+00 -1.38795960e+00
-9.49375704e-02 7.80461550e-01 6.63062394e-01 1.21080649e+00
2.12349832e-01 -1.52019458e-02 2.73543924e-01 3.06132510e-02
8.35877657e-01 2.30167180e-01 7.86292851e-01 -1.72742605e+00
-2.26825297e-01 -4.40555632e-01 -5.28795362e-01 -6.65727913e-01
-3.97190265e-02 -8.44189405e-01 -2.14543447e-01 -1.72480071e+00
-5.75448871e-01 -1.12174606e+00 -4.16362360e-02 8.96473378e-02
3.92550856e-01 2.33293045e-02 3.56527865e-01 2.78065413e-01
-4.59187716e-01 -8.29939172e-02 1.83436036e+00 3.34094435e-01
-5.30340433e-01 6.66064024e-02 -1.89114049e-01 7.39565372e-01
1.15254104e+00 -4.76252288e-01 -4.29430723e-01 -4.33217585e-01
9.39658657e-02 1.73193529e-01 -1.01606742e-01 -1.42604125e+00
5.27237058e-01 -4.25475061e-01 1.30312070e-01 -2.74495751e-01
5.33027828e-01 -1.01509833e+00 6.62353754e-01 1.30169642e+00
4.73353058e-01 6.97573900e-01 -1.27149880e-01 5.19530296e-01
1.81211188e-01 -8.10496956e-02 8.00036371e-01 -6.39976323e-01
-3.04174006e-01 9.07768384e-02 -9.07440305e-01 -4.39363152e-01
1.81220675e+00 -9.14177775e-01 -4.33557212e-01 1.24401316e-01
-6.03988111e-01 5.80555379e-01 1.46230090e+00 2.17664078e-01
4.56704497e-01 -6.49274826e-01 -7.12816358e-01 8.16848204e-02
-5.22513151e-01 -1.01434857e-01 -3.97919536e-01 5.11514425e-01
-1.23715472e+00 4.02430564e-01 -7.27962375e-01 -7.41326690e-01
-1.27643800e+00 6.53297842e-01 -1.17251843e-01 -1.31784767e-01
-4.79914457e-01 8.78881633e-01 6.66720942e-02 -3.29209328e-01
-2.03146487e-01 5.91799878e-02 -2.11253539e-01 -6.40604377e-01
2.90615018e-03 6.66902304e-01 -3.78406644e-01 -3.31657499e-01
-5.50116718e-01 9.63944972e-01 6.91782534e-01 2.30984688e-01
1.55387068e+00 -8.05560052e-02 -1.02105761e+00 2.97194030e-02
5.19959152e-01 2.55380608e-02 -5.89313328e-01 4.74480748e-01
1.08495103e-02 -6.69162393e-01 -7.29119539e-01 -5.02033234e-01
-6.77362204e-01 -5.57580777e-02 -1.27837405e-01 8.20606530e-01
9.93186951e-01 -3.13088926e-03 7.72940695e-01 5.18831968e-01
1.01718664e+00 -9.18109477e-01 2.32288420e-01 8.41680288e-01
6.06102347e-01 -3.45904768e-01 -6.51614666e-02 -1.25661111e+00
2.14432776e-01 1.08606267e+00 6.58297420e-01 -6.31026387e-01
4.79117930e-01 1.13968420e+00 -5.57698272e-02 -4.10502285e-01
-9.89104033e-01 1.92186773e-01 -8.25722992e-01 1.11347461e+00
-1.00437202e-01 1.01156443e-01 -8.33748698e-01 -1.65369004e-01
-6.26920283e-01 -3.56998652e-01 7.68381834e-01 1.43351030e+00
-1.10850096e+00 -1.52636611e+00 -7.28504181e-01 1.75749958e-01
3.40228863e-02 4.61718768e-01 -9.30978537e-01 7.82544076e-01
2.30457246e-01 1.23721886e+00 1.84088290e-01 -3.36267531e-01
-3.66131105e-02 -5.73878944e-01 5.58740079e-01 -2.18566731e-01
-8.05122614e-01 -1.83601916e-01 2.42242087e-02 -3.52061421e-01
-3.65220398e-01 -4.09125209e-01 -1.72713828e+00 -7.53231287e-01
-6.95321709e-02 7.70622671e-01 1.07269609e+00 6.47824228e-01
2.11366743e-01 3.87088537e-01 1.11729002e+00 -7.87533104e-01
1.01785906e-01 -8.81271735e-02 -7.82536209e-01 -3.15351337e-02
-2.92905539e-01 -7.16690004e-01 -7.07370341e-02 -2.98552990e-01] | [5.635213375091553, 4.127933025360107] |
93f321f8-7296-4b94-aa15-22999974d936 | unsupervised-generation-of-long-form | null | null | https://aclanthology.org/2022.pandl-1.3 | https://aclanthology.org/2022.pandl-1.3.pdf | Unsupervised Generation of Long-form Technical Questions from Textbook Metadata using Structured Templates | We explore the task of generating long-form technical questions from textbooks. Semi-structured metadata of a textbook — the table of contents and the index — provide rich cues for technical question generation. Existing literature for long-form question generation focuses mostly on reading comprehension assessment, and does not use semi-structured metadata for question generation. We design unsupervised template based algorithms for generating questions based on structural and contextual patterns in the index and ToC. We evaluate our approach on textbooks on diverse subjects and show that our approach generates high quality questions of diverse types. We show that, in comparison, zero-shot question generation using pre-trained LLMs on the same meta-data has much poorer quality. | ['Tapas Nayak', 'Pratik Saini', 'Arpita Kundu', 'Subhasish Ghosh', 'Indrajit Bhattacharya'] | null | null | null | null | pandl-coling-2022-10 | ['question-generation'] | ['natural-language-processing'] | [ 3.33892018e-01 7.23586619e-01 1.77572295e-01 4.80048470e-02
-1.90875590e+00 -1.11097503e+00 8.10261130e-01 4.55663383e-01
-4.57563549e-02 9.12142515e-01 7.46666491e-01 -7.77389228e-01
-4.33864981e-01 -9.24661219e-01 -7.14903891e-01 2.93088049e-01
7.80094504e-01 4.77965236e-01 4.52595711e-01 -5.50939381e-01
7.53064334e-01 -3.92494917e-01 -1.75401688e+00 4.63910371e-01
1.63711178e+00 4.76465255e-01 5.15911043e-01 1.11190236e+00
-1.21579909e+00 1.19546425e+00 -1.29496825e+00 -5.30647993e-01
-1.63805023e-01 -1.00258398e+00 -1.57442212e+00 4.46247160e-02
1.21606803e+00 -1.80033639e-01 -4.09821630e-01 7.65583992e-01
4.17371660e-01 1.66936293e-01 8.09177160e-01 -7.65481830e-01
-9.70915794e-01 1.07854176e+00 3.76039296e-01 4.82833743e-01
1.25794673e+00 -1.72800117e-03 1.30796516e+00 -5.81595361e-01
1.12604940e+00 7.00671613e-01 3.89265954e-01 7.85834789e-01
-1.11766815e+00 7.07728639e-02 -4.02222484e-01 3.74814481e-01
-7.17206895e-01 -3.89321953e-01 6.40047550e-01 -5.39400101e-01
6.61430001e-01 5.18400371e-01 5.83077848e-01 1.25052154e+00
3.79113615e-01 9.17261004e-01 1.23150098e+00 -8.43594372e-01
2.80060947e-01 1.69492170e-01 4.53228354e-01 6.73573375e-01
3.22937310e-01 -5.07002652e-01 -6.00045264e-01 -1.34988740e-01
4.02413219e-01 -7.25677609e-01 -2.06985697e-01 1.13857634e-01
-1.23668396e+00 7.24267304e-01 -3.09067369e-01 3.75994533e-01
-9.69617963e-02 -1.08367942e-01 5.84911034e-02 5.18468976e-01
-3.58949527e-02 1.46845758e+00 -4.00170684e-01 -5.45146883e-01
-1.44943571e+00 8.03741157e-01 1.58702123e+00 1.64374948e+00
3.34851146e-01 -1.85253307e-01 -1.15172851e+00 6.76840901e-01
-1.34042641e-02 2.50334680e-01 6.60909057e-01 -1.41536474e+00
8.34515512e-01 6.38729870e-01 2.70834446e-01 -6.89147174e-01
7.05947578e-02 -4.97574717e-01 9.70081463e-02 -2.28249565e-01
7.30440795e-01 -2.60905266e-01 -9.21392143e-01 1.27966845e+00
7.62293115e-02 -4.96246129e-01 1.13129243e-01 1.83380157e-01
1.79201829e+00 7.00077534e-01 -3.92198153e-02 1.47507519e-01
1.70748508e+00 -9.74801242e-01 -1.20648825e+00 -7.67403562e-03
6.08760953e-01 -1.21917832e+00 1.54667556e+00 2.25161806e-01
-1.71239996e+00 -6.17838860e-01 -9.38914001e-01 -5.55703223e-01
-4.39916342e-01 1.83474898e-01 -1.68008227e-02 5.73973954e-01
-1.14516592e+00 3.76561701e-01 3.07316147e-02 -3.25309575e-01
1.23793773e-01 -4.04547662e-01 3.02935064e-01 -8.99569094e-02
-1.13561559e+00 7.60527968e-01 -2.15657413e-01 -8.40559125e-01
-7.02187300e-01 -1.28089023e+00 -9.89352822e-01 2.41475642e-01
6.46784544e-01 -9.61257398e-01 2.03892231e+00 2.01666262e-02
-1.53965127e+00 8.21896076e-01 -2.57930636e-01 -2.84865834e-02
-5.95832244e-02 -1.57815129e-01 1.58484140e-03 3.81256253e-01
3.92438829e-01 5.91152906e-01 6.01232886e-01 -1.04456735e+00
-3.94754201e-01 2.44713157e-01 5.55079520e-01 2.64005959e-01
-1.95975855e-01 -8.23258907e-02 -1.09366611e-01 -8.35864663e-01
6.85889721e-02 -3.01055580e-01 -1.57594830e-01 -3.45053732e-01
-6.68751836e-01 -6.61036253e-01 3.43670577e-01 -1.11578524e+00
1.56931806e+00 -9.97913659e-01 -1.42818809e-01 -3.95299345e-02
4.41551328e-01 -2.93762147e-01 -3.20697397e-01 7.41838813e-01
3.30417812e-01 3.90265763e-01 4.56288345e-02 1.70332313e-01
6.00806057e-01 -3.71963650e-01 -6.61807001e-01 -6.89654231e-01
-7.67242908e-02 1.36957717e+00 -1.30631554e+00 -9.06421304e-01
-9.28378850e-02 -4.18465436e-01 -5.22055686e-01 7.34681368e-01
-8.23643804e-01 -1.87874928e-01 -4.50278699e-01 5.62897563e-01
-1.28385618e-01 -5.08657217e-01 -3.92984033e-01 2.24273413e-01
4.84993830e-02 1.23855746e+00 -7.29536951e-01 2.10152578e+00
-7.31904507e-01 9.39645231e-01 -4.54390615e-01 -8.66524726e-02
9.00112450e-01 5.98054945e-01 1.11345656e-01 -7.54888415e-01
-7.57856369e-02 1.48654565e-01 -1.94295138e-01 -9.35102403e-01
1.15569758e+00 2.42992043e-01 -4.21623737e-01 1.00061250e+00
6.67138040e-01 -1.08638835e+00 9.81003106e-01 1.00372398e+00
1.66402602e+00 6.06010377e-01 -1.19061135e-02 -3.37422043e-01
3.36202919e-01 4.70617890e-01 -4.04513806e-01 1.34967840e+00
3.69602054e-01 6.57450438e-01 4.56051052e-01 4.34093088e-01
-1.15584862e+00 -1.45879483e+00 3.45867164e-02 1.11814499e+00
-3.66067916e-01 -9.38588798e-01 -1.30234194e+00 -6.63796484e-01
-4.07392591e-01 1.41955936e+00 -4.11126316e-01 5.07320724e-02
-2.75025010e-01 9.62722600e-02 4.67184246e-01 2.08240882e-01
-5.43832593e-02 -1.21881127e+00 -6.40204966e-01 5.37265003e-01
-7.66776502e-01 -9.53359604e-01 -7.29023039e-01 -2.33108625e-01
-7.48603523e-01 -9.39577162e-01 -8.29907775e-01 -9.15521741e-01
7.05452263e-01 2.55715430e-01 1.93628013e+00 2.20280886e-01
-4.37882870e-01 1.14118934e+00 -5.55052221e-01 -4.45136905e-01
-6.51718676e-01 5.14888704e-01 -8.06172490e-01 -9.89683986e-01
4.15240556e-01 -4.31267262e-01 -2.43093506e-01 -2.38717914e-01
-1.23961771e+00 6.29274994e-02 4.06755149e-01 6.13963485e-01
2.75210053e-01 -5.43341756e-01 5.53245604e-01 -1.08771098e+00
1.41137052e+00 -6.50519967e-01 -4.81174439e-01 6.46628797e-01
-4.16081250e-01 3.80857587e-01 3.65175009e-01 -2.50791848e-01
-1.37871754e+00 -6.48062408e-01 -1.71314612e-01 3.53980809e-01
-3.17818433e-01 4.21215117e-01 -4.39218525e-03 1.49740189e-01
9.83642817e-01 4.30128813e-01 -4.89255279e-01 -5.14083207e-01
7.78276026e-01 4.86440033e-01 6.16244972e-01 -1.07715333e+00
1.25879347e+00 -2.65667617e-01 -2.83177108e-01 -6.74404263e-01
-1.34167683e+00 -4.19910938e-01 -4.58277166e-01 -6.61398828e-01
7.29838729e-01 -4.79899585e-01 -1.40920296e-01 -1.43662900e-01
-1.39325249e+00 -3.36416811e-01 -1.01346540e+00 -1.78989079e-02
-7.10048437e-01 3.82354110e-01 -8.57878923e-01 -5.23129702e-01
-4.27183688e-01 -8.32288027e-01 9.10264909e-01 7.44935691e-01
-9.88245547e-01 -1.16578841e+00 3.00148368e-01 9.77368057e-01
4.77150440e-01 7.20610842e-02 1.39000845e+00 -6.99842334e-01
-9.81930554e-01 4.39096875e-02 1.87148705e-01 -2.18164891e-01
6.04130365e-02 -1.21961609e-01 -6.74348354e-01 4.35126781e-01
7.16641620e-02 -8.10032785e-01 4.19164568e-01 5.18444069e-02
1.24311185e+00 -7.42007196e-01 1.51975945e-01 1.21663570e-01
1.12059557e+00 -1.34823009e-01 7.94485450e-01 4.31001037e-01
5.19979179e-01 1.25574124e+00 3.92736346e-01 2.60087520e-01
9.45746124e-01 3.16702485e-01 -1.23758167e-01 4.88032997e-01
-6.80843174e-01 -6.44276738e-01 3.15091848e-01 1.56783581e+00
5.40601552e-01 -3.33441108e-01 -8.13413918e-01 1.06856382e+00
-1.09244776e+00 -1.23278248e+00 -6.07388675e-01 1.67005956e+00
1.47621620e+00 1.00977845e-01 -2.13756397e-01 -7.36216828e-02
7.23295957e-02 4.33685370e-02 4.50683816e-04 -4.39919949e-01
-1.37435913e-01 1.06003773e+00 -9.22464952e-02 5.73556006e-01
-2.27890313e-01 8.30753565e-01 7.41106176e+00 1.00095522e+00
7.42012560e-02 1.26501590e-01 2.35433266e-01 -1.31792188e-01
-1.31791747e+00 1.60861775e-01 -9.58688736e-01 2.96298891e-01
1.22852361e+00 -8.57284307e-01 1.49500072e-01 6.67592585e-01
-2.73076929e-02 -2.66865849e-01 -1.00337946e+00 4.57994729e-01
5.20160556e-01 -1.77301049e+00 4.44184989e-01 -7.26487488e-02
1.24872088e+00 -8.43051910e-01 -4.40938063e-02 5.90725482e-01
8.22706342e-01 -1.00644064e+00 7.91344762e-01 8.49109650e-01
5.97959876e-01 -3.84284228e-01 4.04454112e-01 2.29475737e-01
-8.47632527e-01 2.37929836e-01 -3.91631186e-01 1.01058803e-01
3.13335568e-01 4.80136186e-01 -7.18870938e-01 5.03016293e-01
3.26371938e-01 -5.57447225e-02 -1.11282849e+00 1.23106802e+00
-5.36994815e-01 9.31115687e-01 1.10910252e-01 -6.71417892e-01
2.97029585e-01 -9.26928595e-02 6.26300931e-01 9.99917090e-01
6.38043463e-01 3.82571816e-01 -4.08789068e-02 1.25559342e+00
-3.10243011e-01 3.69775534e-01 -4.81683493e-01 -2.50402391e-01
6.78143680e-01 1.37728453e+00 -4.20326859e-01 -6.00150943e-01
-2.84797162e-01 5.80338061e-01 1.39163002e-01 2.37773597e-01
-2.96621501e-01 -1.13566017e+00 4.31347080e-03 4.11670506e-01
1.31841227e-01 -2.21775740e-01 -6.64317071e-01 -1.22471035e+00
2.17240915e-01 -1.42829144e+00 4.84814525e-01 -1.35341001e+00
-1.27632785e+00 5.39363503e-01 4.07598428e-02 -1.06579649e+00
-7.43710816e-01 -3.32649291e-01 -1.03389049e+00 9.28120732e-01
-1.03577507e+00 -5.51568866e-01 -6.41467214e-01 2.58847505e-01
1.06094551e+00 -1.72502786e-01 7.48217940e-01 -2.29899630e-01
8.25367346e-02 7.46216893e-01 -7.86210597e-02 -2.90417876e-02
6.36309147e-01 -1.97249734e+00 5.53284585e-01 9.71707404e-01
4.16456729e-01 8.65105271e-01 9.14615512e-01 -9.17396903e-01
-1.44469154e+00 -8.00731659e-01 1.41235673e+00 -1.19068503e+00
8.52462173e-01 -2.89827824e-01 -1.13447487e+00 2.03227609e-01
1.09232843e+00 -1.14797401e+00 1.19146335e+00 -7.89205059e-02
-1.87113941e-01 3.14056993e-01 -9.49929774e-01 1.06314194e+00
9.10577416e-01 -7.36528993e-01 -1.67593670e+00 6.75908625e-01
1.19909012e+00 -6.46935523e-01 -1.06866777e+00 -3.13129842e-01
4.63043898e-02 -4.03671443e-01 7.36315131e-01 -8.03838611e-01
1.10096514e+00 -9.59837586e-02 1.68033063e-01 -1.19966984e+00
-1.66197166e-01 -1.11050498e+00 -5.89908659e-01 1.47319973e+00
5.22248745e-01 1.77606806e-01 8.92367303e-01 6.65796816e-01
-4.56204563e-01 -6.11907899e-01 -6.10775054e-01 -5.88411689e-01
2.97761977e-01 -2.39599839e-01 5.39258063e-01 5.99446535e-01
4.17516619e-01 7.65799344e-01 2.43053436e-01 -6.39580667e-01
8.70887816e-01 1.67093366e-01 8.53713572e-01 -1.05328691e+00
-2.42123440e-01 -5.40623546e-01 3.41750532e-01 -1.45998347e+00
-7.18283132e-02 -9.94300008e-01 4.79672790e-01 -2.19192147e+00
2.03937128e-01 -6.03657728e-03 4.58291560e-01 -7.74179921e-02
-6.71287119e-01 -1.36495918e-01 2.92705651e-02 -1.70024276e-01
-9.30566549e-01 4.81637657e-01 1.56481183e+00 -6.20078892e-02
-5.09319045e-02 -2.47982860e-01 -1.14725447e+00 3.74067903e-01
6.02399528e-01 -4.54056799e-01 -5.20839572e-01 -3.06510538e-01
4.62745070e-01 3.97701085e-01 -4.78315428e-02 -1.19196165e+00
5.30521393e-01 -2.96394080e-01 2.88228840e-01 -8.73241544e-01
-1.00917354e-01 -2.44960878e-02 -5.15497684e-01 9.99754667e-02
-1.04932094e+00 3.57491255e-01 1.13941625e-01 1.31191939e-01
1.21569848e-02 -1.22063684e+00 1.00327305e-01 -5.90125024e-01
-1.91368565e-01 -1.08515890e-02 -1.10668886e+00 1.02500772e+00
4.07656312e-01 -1.92256242e-01 -6.33427501e-01 -9.21329916e-01
-2.77667254e-01 3.19992930e-01 2.13232294e-01 6.82641208e-01
5.82436621e-01 -1.37863028e+00 -8.70788276e-01 -5.07914424e-01
2.49336615e-01 -2.26591289e-01 8.81014392e-02 8.51280168e-02
-4.28247869e-01 9.49556828e-01 -1.02188744e-01 -1.71690077e-01
-9.55434620e-01 2.06928216e-02 -1.43093213e-01 -7.68251240e-01
-1.39489189e-01 7.28263378e-01 -3.91268790e-01 -7.27606535e-01
2.10819900e-01 -4.97032315e-01 -5.98957062e-01 3.01803201e-01
7.44781673e-01 4.57234323e-01 5.92405573e-02 1.43962234e-01
6.18072510e-01 1.55711621e-01 -1.07397512e-01 -7.55331278e-01
9.60268915e-01 -2.92384863e-01 8.42416286e-03 5.73530614e-01
6.25605404e-01 4.05466139e-01 -6.24879360e-01 -2.86294878e-01
6.66203499e-01 -1.61738664e-01 -2.05011383e-01 -1.11795199e+00
-1.26829790e-02 7.80316710e-01 -2.20160142e-01 3.73626113e-01
6.94587350e-01 2.37454921e-01 1.21845663e+00 8.31401408e-01
3.15513194e-01 -1.25895679e+00 9.07680571e-01 9.26371157e-01
9.28042173e-01 -6.96167886e-01 -1.75429597e-01 -3.34403992e-01
-2.58362055e-01 1.28892052e+00 1.10610843e+00 3.21273863e-01
-1.04010858e-01 -7.19838217e-02 1.80920348e-01 -4.04106647e-01
-1.06362152e+00 -2.91630477e-01 7.98023403e-01 5.99390864e-01
7.55544245e-01 -5.38804471e-01 -4.11996126e-01 8.18780124e-01
-1.07610631e+00 -1.01101920e-01 1.36793292e+00 1.30645812e+00
-7.71555543e-01 -1.24072385e+00 -5.51708877e-01 1.14918220e+00
-6.57678366e-01 -5.52596569e-01 -6.64117575e-01 3.20367515e-01
-3.77474576e-01 1.26780283e+00 -1.28912270e-01 -6.22058548e-02
3.44329774e-01 5.65729856e-01 9.45172548e-01 -1.35871184e+00
-1.01406658e+00 -5.27375698e-01 3.46705019e-01 -2.61557788e-01
1.43949881e-01 -7.24079192e-01 -8.52493286e-01 -1.30090073e-01
-1.51367992e-01 8.44471574e-01 5.55406511e-01 1.15260196e+00
2.79998243e-01 9.30696070e-01 2.70183414e-01 -7.93865547e-02
-1.00067401e+00 -1.25650907e+00 8.28335956e-02 4.21241581e-01
-2.31608991e-02 -1.58477619e-01 -2.50161052e-01 4.81666178e-01] | [11.45769214630127, 8.066893577575684] |
d7270265-52c4-4c30-bc46-a85df29225a7 | local-and-global-topics-in-text-modeling-of | 2104.01115 | null | https://arxiv.org/abs/2104.01115v1 | https://arxiv.org/pdf/2104.01115v1.pdf | Local and Global Topics in Text Modeling of Web Pages Nested in Web Sites | Topic models are popular models for analyzing a collection of text documents. The models assert that documents are distributions over latent topics and latent topics are distributions over words. A nested document collection is where documents are nested inside a higher order structure such as stories in a book, articles in a journal, or web pages in a web site. In a single collection of documents, topics are global, or shared across all documents. For web pages nested in web sites, topic frequencies likely vary between web sites. Within a web site, topic frequencies almost certainly vary between web pages. A hierarchical prior for topic frequencies models this hierarchical structure and specifies a global topic distribution. Web site topic distributions vary around the global topic distribution and web page topic distributions vary around the web site topic distribution. In a nested collection of web pages, some topics are likely unique to a single web site. Local topics in a nested collection of web pages are topics unique to one web site. For US local health department web sites, brief inspection of the text shows local geographic and news topics specific to each department that are not present in others. Topic models that ignore the nesting may identify local topics, but do not label topics as local nor do they explicitly identify the web site owner of the local topic. For web pages nested inside web sites, local topic models explicitly label local topics and identifies the owning web site. This identification can be used to adjust inferences about global topics. In the US public health web site data, topic coverage is defined at the web site level after removing local topic words from pages. Hierarchical local topic models can be used to identify local topics, adjust inferences about if web sites cover particular health topics, and study how well health topics are covered. | ['Robert E. Weiss', 'Jason Wang'] | 2021-03-30 | null | null | null | null | ['topic-coverage'] | ['natural-language-processing'] | [-2.38160402e-01 3.39470208e-01 -6.91125333e-01 -1.67281613e-01
-1.10866547e+00 -8.80861521e-01 8.34877372e-01 7.58205354e-01
-9.93506610e-02 5.31980693e-01 9.78796840e-01 -5.31275153e-01
-4.07669276e-01 -1.27658141e+00 -6.14660025e-01 -6.11598849e-01
-2.45965913e-01 7.50522971e-01 5.92582881e-01 6.69341013e-02
1.70850366e-01 -2.89136708e-01 -8.77716482e-01 3.70678186e-01
3.40970874e-01 4.89337258e-02 4.32724386e-01 3.77645046e-01
-7.46744633e-01 1.16783217e-01 -8.73427391e-01 1.00135736e-01
-4.78938818e-02 -1.92093134e-01 -7.51235545e-01 2.19114155e-01
2.48076528e-01 -3.18067610e-01 4.56887968e-02 9.22338188e-01
-1.13554634e-01 -2.03112483e-01 8.61640394e-01 -1.30626047e+00
-2.63915472e-02 9.54890072e-01 -8.25998485e-01 4.18686301e-01
3.85413438e-01 -4.60478395e-01 1.39380145e+00 -5.00607729e-01
1.01446426e+00 1.45519054e+00 6.37415767e-01 -3.32244068e-01
-1.51484025e+00 -6.61024690e-01 7.20726848e-01 -6.48612380e-01
-1.21026778e+00 3.13397974e-01 4.91218150e-01 -9.01230812e-01
5.85712016e-01 3.78332525e-01 7.33950436e-01 1.18490636e+00
1.05677736e+00 3.94268662e-01 8.09024692e-01 -1.66194767e-01
4.34372514e-01 2.23194286e-01 9.75758076e-01 1.69881299e-01
6.73593163e-01 -6.34964705e-01 -6.82439446e-01 -1.18320954e+00
5.20099938e-01 5.46805322e-01 -1.71393305e-01 -3.44581753e-01
-1.35647738e+00 1.40720367e+00 -1.19789101e-01 4.27071869e-01
-5.28541028e-01 1.66806374e-02 4.12114024e-01 6.40545711e-02
8.39111686e-01 1.18852898e-01 -5.94295144e-01 2.21175514e-02
-9.12865758e-01 6.66682184e-01 1.02956188e+00 8.66567433e-01
1.09888613e+00 -8.01661015e-01 -3.70584168e-02 9.55260336e-01
4.54752564e-01 4.69778448e-01 4.13171917e-01 -4.52358961e-01
5.56032598e-01 5.02104998e-01 2.27074459e-01 -1.13550782e+00
-5.21335721e-01 -9.23590735e-04 -1.90739378e-01 -4.00795788e-01
1.76860034e-01 -1.73101470e-01 -8.33077013e-01 1.61991358e+00
4.13478762e-01 -5.20547092e-01 -2.08923534e-01 2.24148124e-01
6.45178735e-01 1.21336877e+00 5.46737134e-01 -3.55815828e-01
2.13653588e+00 -3.29564929e-01 -1.03825414e+00 -1.77189931e-01
4.35757220e-01 -8.84582460e-01 9.93707538e-01 1.66695610e-01
-8.42944324e-01 4.06759866e-02 -6.17846072e-01 8.48740414e-02
-7.88939714e-01 -6.13641620e-01 5.18934190e-01 5.11309445e-01
-1.00479555e+00 -2.57759124e-01 -8.67900610e-01 -1.09837961e+00
-2.32260615e-01 -4.43197973e-03 -1.59438085e-02 3.16599756e-02
-1.51816833e+00 3.98664027e-01 3.77268612e-01 -9.03022468e-01
-9.98349071e-01 -9.32880938e-01 -6.79307044e-01 2.40134105e-01
5.39033353e-01 -4.88709033e-01 1.17853510e+00 -2.04280466e-01
-3.35674286e-01 4.69349772e-01 -5.15007138e-01 -2.41135865e-01
-9.58976671e-02 -6.33597448e-02 -5.64758003e-01 2.08513007e-01
1.05599201e+00 4.52532291e-01 3.72253984e-01 -1.11093235e+00
-1.22108793e+00 -3.74763161e-01 5.96525222e-02 1.63333505e-01
-5.42791367e-01 1.83337733e-01 -5.44155836e-01 -6.62019074e-01
5.09179711e-01 -6.98705077e-01 -9.51420143e-02 -6.45475686e-01
-6.38166010e-01 -6.92388535e-01 1.11549234e+00 -6.98413610e-01
1.34659684e+00 -1.92573822e+00 -5.76962650e-01 3.60475898e-01
3.47957104e-01 -1.19978571e+00 3.71904284e-01 7.95487761e-01
1.09570362e-02 8.15168858e-01 6.89116269e-02 2.59021878e-01
5.14335856e-02 6.47768080e-02 -6.08346343e-01 5.33044815e-01
-3.65219682e-01 1.70668855e-01 -6.25145853e-01 -4.98611838e-01
-2.23477527e-01 3.36254179e-01 -5.97175956e-01 -4.00553018e-01
-3.83029670e-01 -4.27666783e-01 -7.43233740e-01 2.28996173e-01
4.76627111e-01 -6.90675139e-01 3.48006994e-01 2.51316130e-01
-4.91719246e-01 1.12867224e+00 -1.13679445e+00 1.11267090e+00
-3.26352119e-01 7.79540658e-01 -1.11290000e-01 -4.11287218e-01
7.18702853e-01 6.64402008e-01 9.44041908e-01 9.26609114e-02
-5.56912482e-01 -2.49926671e-01 -3.20491254e-01 -9.69391093e-02
6.28328323e-01 -6.60529509e-02 -3.60158563e-01 1.05349851e+00
-2.73323447e-01 -2.06162836e-02 2.75428087e-01 6.50243700e-01
1.30556715e+00 -5.21271527e-01 4.02276605e-01 -8.93919826e-01
-3.48510534e-01 4.91278082e-01 5.82351267e-01 9.90463853e-01
2.42523983e-01 6.33121729e-01 1.05890799e+00 -4.72069949e-01
-1.16780877e+00 -1.41948068e+00 -9.03356850e-01 1.50519419e+00
3.74672227e-02 -8.59909296e-01 -5.86763620e-01 -6.22070789e-01
-1.22776747e-01 6.53086007e-01 -8.33202243e-01 5.21102428e-01
-5.98744228e-02 -1.19449246e+00 -1.67384773e-01 -3.19593772e-02
1.65887356e-01 -5.79867005e-01 -5.04502416e-01 4.80943412e-01
-7.12745368e-01 -7.44259536e-01 -6.64546907e-01 3.57600659e-01
-8.36775005e-01 -9.55900490e-01 -1.07167876e+00 -6.65556788e-01
7.15595305e-01 3.33021045e-01 1.12046766e+00 -4.15288597e-01
5.44135608e-02 6.76543236e-01 -1.86236590e-01 -5.66536784e-01
-3.33192140e-01 5.05311191e-01 -9.67531931e-03 -5.26341200e-01
6.30214334e-01 -7.80196339e-02 -5.74631929e-01 4.44607586e-01
-1.04914021e+00 -3.54252458e-01 -4.81368229e-02 5.96184492e-01
2.59842455e-01 6.86242342e-01 3.64178747e-01 -1.44624436e+00
6.58137679e-01 -1.38194704e+00 -4.46160048e-01 1.76194847e-01
-4.92930442e-01 -2.93733090e-01 -3.42009693e-01 -4.52250570e-01
-1.04561090e+00 -6.20351017e-01 5.33411562e-01 5.95465899e-01
-3.77170116e-01 1.05042362e+00 -5.00555225e-02 1.13971663e+00
6.60685897e-01 5.20119257e-02 -4.10001546e-01 -4.15053874e-01
-1.83605984e-01 6.55358732e-01 -3.72935325e-01 -4.89869624e-01
4.73796368e-01 7.62100160e-01 -6.26514256e-01 -9.59586859e-01
-6.78642094e-01 -1.02273953e+00 -2.46123821e-01 -7.04316273e-02
1.06449139e+00 -1.21946406e+00 -2.95309097e-01 -2.28717566e-01
-1.02450645e+00 -2.64253199e-01 -8.03026650e-03 6.26919806e-01
-1.30832016e-01 9.68770357e-04 -5.38221121e-01 -4.65502501e-01
2.17138588e-01 -9.94913995e-01 1.15031850e+00 5.82061987e-03
-1.01197636e+00 -1.72364128e+00 3.56451035e-01 5.82580678e-02
1.33336633e-01 2.31751408e-02 1.40060580e+00 -1.17945397e+00
-4.90120173e-01 -2.03283355e-01 1.03244744e-02 -6.13481522e-01
7.77325153e-01 8.88748690e-02 -3.97346467e-01 -4.54157859e-01
-6.02302663e-02 2.73314774e-01 4.85829800e-01 9.43305075e-01
4.53306079e-01 -7.92430043e-01 -1.10975718e+00 -1.55369788e-01
1.38432729e+00 6.02782369e-01 1.28872409e-01 7.65779972e-01
-5.95839415e-03 1.14291465e+00 2.17508867e-01 5.72433412e-01
4.23578620e-01 3.69344920e-01 -2.72658050e-01 -2.22467482e-01
4.29953575e-01 -2.42074445e-01 4.86393571e-01 6.36230409e-01
6.99684381e-01 -5.29472470e-01 -1.39834034e+00 1.15647054e+00
-1.45886302e+00 -8.72139335e-01 -4.48957264e-01 2.15564275e+00
9.86499906e-01 4.36891586e-01 1.54200137e-01 -5.82044780e-01
1.09959185e+00 3.00712794e-01 -1.95794791e-01 1.84455048e-02
9.98499319e-02 -5.68959653e-01 5.61676562e-01 8.43210042e-01
-9.91125822e-01 5.27147114e-01 7.49674606e+00 4.97077793e-01
-6.04569137e-01 2.16941729e-01 9.33779180e-01 -5.95075861e-02
-8.79071057e-01 3.90236020e-01 -1.47282875e+00 6.03671014e-01
8.63781273e-01 -5.71043670e-01 -5.61349154e-01 1.00158429e+00
5.92619002e-01 -5.47775030e-01 -9.05820012e-01 8.73868763e-02
-9.71843451e-02 -1.16732979e+00 3.38725120e-01 9.23668742e-01
1.12286329e+00 -1.99977428e-01 1.83021322e-01 -2.89532766e-02
1.07559013e+00 -4.21569169e-01 3.21575433e-01 6.49700612e-02
2.18810588e-01 -6.39590919e-01 7.59540439e-01 2.31581718e-01
-8.73742521e-01 3.83867979e-01 -5.30257761e-01 2.95934796e-01
2.93381393e-01 1.09853959e+00 -1.08087242e+00 3.04899961e-02
1.31895339e+00 2.00761855e-01 -1.43537506e-01 1.02349663e+00
1.82728618e-01 1.15645421e+00 -6.23589635e-01 -7.37966001e-02
3.87926757e-01 -5.95575981e-02 9.78350580e-01 1.21642816e+00
2.82061845e-01 -6.90227002e-02 3.86698931e-01 6.75256252e-01
2.51552433e-01 2.86650121e-01 -7.90218472e-01 -8.46871138e-02
4.82277155e-01 8.89181733e-01 -1.41525805e+00 -6.18639886e-01
-6.50713325e-01 2.16326550e-01 -5.08830369e-01 9.10084784e-01
-5.09739816e-01 -3.99331480e-01 7.53556848e-01 8.94275427e-01
1.35582030e-01 -1.21081673e-01 -3.94654572e-01 -7.62070298e-01
-3.89123172e-01 -3.75576347e-01 7.43400216e-01 -6.66373849e-01
-1.23118794e+00 2.42565125e-01 8.71299326e-01 -9.51659024e-01
-3.36419553e-01 -1.12295046e-01 -5.05823612e-01 8.00445497e-01
-7.79159963e-01 -5.67801714e-01 8.79238099e-02 5.15520871e-01
8.79086792e-01 -1.84291944e-01 6.16185904e-01 -3.40525359e-01
9.29544643e-02 -1.90408409e-01 4.44879860e-01 1.41378984e-01
8.15449297e-01 -1.46342540e+00 3.68762970e-01 4.21858490e-01
-1.37805790e-01 1.48341334e+00 1.04718232e+00 -1.51150823e+00
-4.22774702e-01 -7.78850317e-01 1.45534372e+00 -4.31599766e-01
1.08296788e+00 -6.52823269e-01 -9.96524513e-01 1.19029343e+00
5.69607496e-01 -1.13644874e+00 1.28836966e+00 1.04032111e+00
-1.93532526e-01 1.69262648e-01 -8.22231114e-01 7.49266267e-01
1.95187286e-01 -3.71413171e-01 -8.61075580e-01 1.17374897e+00
1.10559201e+00 1.39900118e-01 -8.76005292e-01 -2.09222987e-01
4.12562639e-01 -3.79355252e-01 6.63258135e-01 -3.68938267e-01
4.18019712e-01 -6.15137555e-02 1.40084943e-03 -1.18977928e+00
-4.86551404e-01 -3.86536360e-01 7.36724079e-01 1.53741896e+00
7.32377589e-01 -8.57562721e-01 1.00633144e+00 9.65620399e-01
1.48481503e-01 1.22335963e-02 -6.52523518e-01 -5.06163299e-01
2.99050182e-01 -2.68964738e-01 4.07866776e-01 1.20268440e+00
4.80348259e-01 1.25506565e-01 2.56606400e-01 3.47364664e-01
6.57103896e-01 2.24517900e-02 5.54606318e-01 -1.48014307e+00
5.75198457e-02 -2.01857597e-01 1.24257118e-01 -8.36801708e-01
-4.40510273e-01 -4.10026670e-01 5.01996018e-02 -2.05319428e+00
8.37071240e-01 -3.47163200e-01 2.88232118e-01 4.87493932e-01
-3.12122181e-02 -2.80052036e-01 -3.50249350e-01 7.15923250e-01
-2.80792445e-01 -5.43144457e-02 6.79011703e-01 -3.00554186e-01
-5.15726209e-01 -1.42930262e-02 -8.00471842e-01 9.04352784e-01
6.90942764e-01 -1.04542124e+00 -4.82088298e-01 5.81470653e-02
5.57105303e-01 1.61145136e-01 6.42144829e-02 -5.01958966e-01
2.41711870e-01 -3.35791528e-01 2.41990268e-01 -1.09911847e+00
-2.70160045e-02 -8.08880448e-01 4.11236316e-01 4.38926011e-01
-5.91865420e-01 3.46502550e-02 1.99608982e-01 7.11273968e-01
-3.41875285e-01 -2.86705315e-01 2.94948339e-01 -3.36544275e-01
-2.14698642e-01 -5.89574408e-03 -1.49699545e+00 -1.15437284e-01
9.92235601e-01 -7.07376227e-02 -4.15529013e-01 -8.22689712e-01
-1.06946862e+00 2.91087359e-01 2.66187102e-01 3.84919018e-01
1.16874449e-01 -1.18713474e+00 -5.21391153e-01 -2.54449785e-01
1.35212690e-01 1.17728347e-02 2.42136568e-01 3.57767522e-01
-2.62817353e-01 7.83676744e-01 2.03803346e-01 -7.28370667e-01
-1.23868430e+00 2.78138399e-01 -1.02109663e-01 -4.54367727e-01
-6.21095598e-01 5.13942301e-01 1.44479740e+00 -3.73204201e-01
1.63903117e-01 -4.50273961e-01 -2.58548826e-01 5.87676406e-01
5.95811248e-01 1.65042296e-01 -3.42886984e-01 -1.94538355e-01
-1.83958083e-01 3.56567979e-01 -5.78165412e-01 -6.76737666e-01
1.25493228e+00 -4.84623641e-01 -4.40405130e-01 1.19392586e+00
1.26877904e+00 1.14686012e-01 -8.81641507e-01 -8.59054923e-02
2.08522141e-01 -8.73622149e-02 -9.99446362e-02 -7.01044261e-01
-4.37710434e-01 2.57539093e-01 1.75935149e-01 9.64577436e-01
5.05431533e-01 7.81180203e-01 2.29004219e-01 -4.62986119e-02
2.77967632e-01 -1.00174451e+00 2.87670642e-02 3.51533234e-01
8.15747261e-01 -9.67298627e-01 2.98097283e-01 -7.67625511e-01
-4.79828507e-01 8.97345841e-01 2.64243633e-01 1.71850607e-01
1.31858039e+00 1.93184778e-01 -7.99040310e-03 -7.37923205e-01
-8.37413132e-01 2.92943090e-01 2.41143748e-01 2.72537678e-01
7.19767153e-01 -2.11885502e-03 -7.09198833e-01 4.78069454e-01
-5.43061078e-01 -6.06632113e-01 8.28116179e-01 7.08882987e-01
-7.03976810e-01 -9.02348340e-01 -9.48222578e-01 6.86229348e-01
-9.37720954e-01 -3.41554195e-01 -1.33561924e-01 1.18048394e+00
-9.81232151e-02 1.05229831e+00 9.06150043e-01 1.81609809e-01
-2.98050582e-01 5.55684209e-01 -6.38906837e-01 -1.06437194e+00
-3.89726281e-01 9.65793610e-01 3.27468328e-02 -1.60316825e-02
-1.56857669e-01 -1.21501088e+00 -1.19924462e+00 -1.26608565e-01
-3.47554088e-01 7.79918373e-01 8.10787141e-01 7.15034604e-01
-8.52231085e-02 3.11729461e-01 2.91381568e-01 1.71913192e-01
4.14224952e-01 -1.07359684e+00 -1.01900566e+00 1.11085013e-01
3.60523343e-01 -3.99931103e-01 -5.43311000e-01 5.90108573e-01] | [10.34074878692627, 7.014677047729492] |
85b5babd-12c7-4c88-8cd3-9d084d1bcac5 | one-stage-shape-instantiation-from-a-single | 1907.10763 | null | https://arxiv.org/abs/1907.10763v1 | https://arxiv.org/pdf/1907.10763v1.pdf | One-stage Shape Instantiation from a Single 2D Image to 3D Point Cloud | Shape instantiation which predicts the 3D shape of a dynamic target from one or more 2D images is important for real-time intra-operative navigation. Previously, a general shape instantiation framework was proposed with manual image segmentation to generate a 2D Statistical Shape Model (SSM) and with Kernel Partial Least Square Regression (KPLSR) to learn the relationship between the 2D and 3D SSM for 3D shape prediction. In this paper, the two-stage shape instantiation is improved to be one-stage. PointOutNet with 19 convolutional layers and three fully-connected layers is used as the network structure and Chamfer distance is used as the loss function to predict the 3D target point cloud from a single 2D image. With the proposed one-stage shape instantiation algorithm, a spontaneous image-to-point cloud training and inference can be achieved. A dataset from 27 Right Ventricle (RV) subjects, indicating 609 experiments, were used to validate the proposed one-stage shape instantiation algorithm. An average point cloud-to-point cloud (PC-to-PC) error of 1.72mm has been achieved, which is comparable to the PLSR-based (1.42mm) and KPLSR-based (1.31mm) two-stage shape instantiation algorithm. | ['Guang-Zhong Yang', 'Jian-Qing Zheng', 'Peichao Li', 'Zhao-Yang Wang', 'Xiao-Yun Zhou'] | 2019-07-24 | null | null | null | null | ['image-to-3d'] | ['computer-vision'] | [ 1.22799441e-01 5.80506444e-01 7.16744736e-02 -5.97530007e-01
-7.64780343e-01 -2.03639045e-01 5.43152809e-01 3.80088061e-01
-6.11331463e-01 3.44132453e-01 -4.02731389e-01 -4.33044046e-01
-2.51645088e-01 -5.40645719e-01 -6.86955214e-01 -5.67347288e-01
-1.13149360e-01 9.37036455e-01 4.03623074e-01 -9.06767771e-02
3.49250317e-01 1.26212633e+00 -9.95411396e-01 2.25821778e-01
5.45795977e-01 1.28402841e+00 4.95261788e-01 6.92760348e-01
-1.60429582e-01 7.10914582e-02 -2.34092832e-01 -3.14996660e-01
5.96888363e-01 9.73872244e-02 -4.84202415e-01 -1.87358737e-01
4.49695408e-01 -2.59090036e-01 5.88170486e-03 7.20533371e-01
8.05363417e-01 -2.20167875e-01 6.70764267e-01 -8.30528259e-01
-2.56440371e-01 3.82040329e-02 -4.45171237e-01 -2.03477964e-01
-2.38979489e-01 6.16851300e-02 3.93207133e-01 -1.01974714e+00
6.51831985e-01 1.02584910e+00 9.76628184e-01 4.81574446e-01
-1.15527987e+00 -7.11826921e-01 -3.95805180e-01 -2.96654671e-01
-1.35151577e+00 9.48535129e-02 1.02205408e+00 -6.93680286e-01
8.22492898e-01 2.51497895e-01 9.78289425e-01 5.37739098e-01
8.44777882e-01 5.43864012e-01 1.03943062e+00 -3.02641600e-01
-5.64036593e-02 5.40227350e-03 -3.87194574e-01 7.52668321e-01
2.48916939e-01 4.56250608e-01 6.20386228e-02 -1.72463343e-01
1.57813311e+00 1.14167750e-01 6.74366951e-02 -8.52865160e-01
-1.28520322e+00 7.35387981e-01 1.06310797e+00 3.56792957e-01
-6.12504840e-01 1.66763589e-01 1.69744834e-01 -1.16703250e-01
2.28056803e-01 5.86804509e-01 -4.90528524e-01 4.32885252e-02
-1.03515995e+00 2.65653342e-01 5.77023029e-01 9.77516413e-01
3.61018568e-01 -5.16518299e-03 -1.00844808e-01 5.69584727e-01
6.90570831e-01 6.35334373e-01 6.98187172e-01 -7.19934642e-01
2.27409512e-01 8.44270647e-01 -1.50667831e-01 -1.02678800e+00
-1.00742757e+00 -7.79570580e-01 -8.96610916e-01 6.17532909e-01
3.52799445e-01 9.19052679e-03 -1.33366930e+00 1.03221631e+00
2.94742107e-01 1.23837873e-01 -1.29060030e-01 1.12761414e+00
1.16132581e+00 -9.50874388e-02 -2.88749933e-01 -1.61556974e-01
1.11281192e+00 -6.29160702e-01 -9.93272811e-02 -1.78062811e-01
6.60424471e-01 -5.18008053e-01 7.20118880e-01 2.00434074e-01
-1.19537842e+00 -4.59706783e-01 -9.55219150e-01 -2.01496854e-03
6.36472926e-02 3.53049785e-01 3.33545804e-01 2.46978387e-01
-8.92101645e-01 6.87069952e-01 -1.17009282e+00 6.40992969e-02
8.36818337e-01 7.11649060e-01 -5.50134778e-01 3.12453002e-01
-5.96733689e-01 1.09024763e+00 3.38285953e-01 4.45081890e-01
-6.57004356e-01 -1.11416996e+00 -6.34396851e-01 -1.85932964e-01
-1.03208579e-01 -8.52629900e-01 9.46898580e-01 -5.35885334e-01
-1.62695217e+00 1.38340008e+00 1.57316446e-01 -4.90962058e-01
8.70116234e-01 2.52342001e-02 1.02531314e-01 2.44252786e-01
-1.52134702e-01 7.18603373e-01 6.18017018e-01 -1.57002890e+00
-6.58771619e-02 -7.75079548e-01 -3.38794649e-01 4.72271174e-01
4.89120245e-01 -3.08764905e-01 -3.29405040e-01 -4.68995601e-01
9.56713915e-01 -1.01936877e+00 -6.08553469e-01 3.61633301e-01
-3.70596647e-01 2.51879215e-01 3.75766665e-01 -6.58825397e-01
6.85061038e-01 -1.91133368e+00 -1.52939558e-01 5.18664479e-01
4.54738438e-01 2.24181205e-01 8.24060813e-02 -1.06774807e-01
-2.89982319e-01 -1.07096262e-01 -5.25594175e-01 -5.64884901e-01
-4.64593917e-01 6.67665824e-02 2.93922663e-01 6.84289873e-01
-5.99862672e-02 1.23446536e+00 -7.71072567e-01 -5.24674416e-01
4.17439193e-01 6.31473064e-01 -4.24362361e-01 1.62054952e-02
1.45961761e-01 8.80386710e-01 -3.98491144e-01 4.19625729e-01
8.54924381e-01 -9.86480191e-02 -1.38131768e-01 -3.52576107e-01
-7.35634044e-02 -4.21812922e-01 -6.69040084e-01 2.03149366e+00
-5.79264700e-01 4.83997643e-01 -2.28130240e-02 -5.57904601e-01
1.54757214e+00 2.55083382e-01 1.02087474e+00 -6.86350822e-01
3.81247461e-01 7.33347476e-01 2.56386876e-01 -3.68901163e-01
-2.97069885e-02 -6.49095774e-01 2.01876298e-01 -2.93915033e-01
-8.24186429e-02 -5.67849934e-01 -6.30015373e-01 -4.05283898e-01
5.40123940e-01 2.79495716e-01 3.68830740e-01 -3.18599403e-01
6.50846064e-01 1.28582254e-01 4.24709678e-01 5.57803035e-01
-1.26371592e-01 1.16059041e+00 4.07748550e-01 -5.59804797e-01
-1.26693714e+00 -1.09412789e+00 -6.36393249e-01 -1.20033950e-01
1.93872392e-01 1.36779964e-01 -5.33407688e-01 -5.84880829e-01
1.62682816e-01 6.12959206e-01 -4.89309937e-01 4.76552919e-02
-8.60093594e-01 -3.74808967e-01 3.35979640e-01 5.45203328e-01
2.13176787e-01 -9.92543757e-01 -8.23726535e-01 8.86887759e-02
1.50082991e-01 -8.79305661e-01 -1.47763401e-01 1.74434811e-01
-1.45650077e+00 -1.21055114e+00 -8.94940794e-01 -6.63746357e-01
1.08845890e+00 -3.56454939e-01 9.11167920e-01 3.80181000e-02
-3.76848727e-01 1.27468914e-01 -1.14504769e-01 -6.73671663e-01
-4.35229629e-01 -1.46370023e-01 -1.80436879e-01 -1.45968691e-01
-4.51515103e-03 -6.19028330e-01 -9.56588745e-01 2.54785806e-01
-6.60489082e-01 2.41562009e-01 1.12295997e+00 9.53151345e-01
1.15682745e+00 -3.11188102e-01 2.23443776e-01 -7.46273279e-01
3.01526725e-01 -1.07905455e-01 -7.93771803e-01 2.76013929e-03
-8.17875564e-01 -2.27913573e-01 3.75454307e-01 -3.61714751e-01
-6.46669567e-01 6.25350595e-01 -3.62251222e-01 -1.05702436e+00
-2.99280286e-01 4.04645324e-01 1.91102669e-01 -3.09882551e-01
8.15234125e-01 3.47710341e-01 7.13986278e-01 -3.69728416e-01
1.93528563e-01 4.79163408e-01 5.62064171e-01 -1.92050442e-01
7.67551541e-01 6.22848868e-01 6.01752222e-01 -7.59947896e-01
-5.36187351e-01 -6.43705010e-01 -1.29628158e+00 -2.21209824e-01
1.08311260e+00 -7.67452478e-01 -7.40075707e-01 4.13928151e-01
-1.15754843e+00 -4.65784460e-01 -3.38574082e-01 9.09082234e-01
-1.08617485e+00 1.16598867e-01 -2.54108477e-02 -6.68667734e-01
-7.82861173e-01 -1.51019156e+00 1.27086854e+00 -6.32420108e-02
2.15895716e-02 -9.98573780e-01 -1.88047141e-01 2.76467472e-01
3.99144828e-01 7.32123375e-01 6.22689247e-01 -8.71656835e-01
-5.01043916e-01 -6.25288725e-01 -2.07584456e-01 1.98308706e-01
-1.26970172e-01 -4.31802183e-01 -7.51169264e-01 -3.87602895e-01
2.94850677e-01 1.82850316e-01 1.58763856e-01 9.45507884e-01
1.19022298e+00 -4.01518010e-02 -4.07866389e-01 1.00780737e+00
1.69667089e+00 2.38255784e-01 5.61303794e-01 5.00733376e-01
8.16421270e-01 6.38422295e-02 7.00982928e-01 3.15742642e-01
2.00619847e-01 7.56098628e-01 8.14699948e-01 -3.25002164e-01
-2.21131369e-01 -2.48305932e-01 -3.55766654e-01 5.27991831e-01
-2.21076488e-01 6.20590150e-01 -1.25414145e+00 1.98612198e-01
-1.60112238e+00 -4.90803182e-01 -5.18567085e-01 2.45372987e+00
5.75643241e-01 3.38306457e-01 -1.74692482e-01 -3.20852846e-01
3.73553008e-01 -2.94619769e-01 -6.10797524e-01 -8.62310007e-02
2.78874129e-01 2.76758641e-01 7.22028375e-01 4.62925822e-01
-9.44101512e-01 6.65692747e-01 5.48515892e+00 6.37402177e-01
-1.54751003e+00 -8.32916424e-02 4.57705349e-01 1.04789928e-01
2.16662511e-03 -1.52822658e-02 -5.87988794e-01 2.38645434e-01
5.95104873e-01 -1.50124446e-01 9.54649597e-02 9.30772007e-01
4.88048136e-01 1.72610823e-02 -9.32986200e-01 1.37167668e+00
2.65630335e-02 -1.39116144e+00 -5.96633628e-02 1.97452694e-01
5.65167069e-01 2.98560470e-01 -2.06711203e-01 -1.23600811e-02
-4.39611048e-01 -1.17597187e+00 6.81553185e-01 1.15100145e+00
1.19357252e+00 -5.61809003e-01 9.78583336e-01 7.47357488e-01
-9.63360071e-01 -6.53032959e-02 -3.16974252e-01 4.58400905e-01
1.98316723e-01 3.73940408e-01 -1.57106721e+00 6.42596543e-01
5.98428726e-01 4.12080765e-01 -4.85055804e-01 1.44229937e+00
9.31891575e-02 -1.29472753e-02 -4.42472905e-01 1.68988228e-01
1.08146071e-01 -2.20502049e-01 9.05832112e-01 8.53183389e-01
3.72414142e-01 1.16050452e-01 2.38102451e-02 1.05512857e+00
2.46244833e-01 2.44176194e-01 -4.02352333e-01 4.98724610e-01
1.51719451e-01 1.41036403e+00 -9.93225694e-01 4.33828160e-02
5.34188794e-03 5.91222525e-01 -3.83921303e-02 2.14697663e-02
-5.64681590e-01 -2.81332552e-01 2.56250024e-01 6.83529377e-01
1.92858145e-01 -5.48693240e-01 -8.29199672e-01 -6.02252126e-01
-1.04595749e-02 -2.41418406e-01 2.70479750e-02 -9.25080419e-01
-1.01795292e+00 6.11611605e-01 5.00331447e-02 -1.39805436e+00
-4.39184941e-02 -7.61827707e-01 -5.36330938e-01 1.30093133e+00
-1.39449811e+00 -1.34958673e+00 -5.21962404e-01 2.41709292e-01
3.27198952e-01 -1.90160885e-01 8.32955956e-01 -1.42908543e-01
1.14543438e-01 5.05893826e-01 -1.04851723e-01 3.52804869e-01
3.57549757e-01 -1.39674354e+00 1.66670963e-01 3.36470902e-01
-1.48403913e-01 5.97430229e-01 2.92860270e-01 -8.99467945e-01
-1.19684875e+00 -1.04587400e+00 5.60692608e-01 -6.77430451e-01
2.65742719e-01 -1.31816091e-02 -7.12144911e-01 4.54488993e-01
-6.06517911e-01 4.63699341e-01 3.69547009e-01 -4.56624806e-01
3.61853629e-01 5.10014966e-02 -1.47265339e+00 3.51227045e-01
8.32641840e-01 -1.31492019e-01 -4.86374408e-01 1.43503472e-01
5.32184780e-01 -1.27148736e+00 -1.40593576e+00 8.73949766e-01
6.58822000e-01 -9.10821497e-01 1.12754452e+00 -4.67286021e-01
3.27494711e-01 -2.97195315e-01 5.30343689e-02 -1.07240224e+00
-2.59295017e-01 -3.63972485e-01 -3.41309346e-02 1.47451833e-01
4.91429359e-01 -4.72475797e-01 1.16735721e+00 7.98791587e-01
-7.66364098e-01 -1.38282585e+00 -1.26855373e+00 -7.67090917e-01
5.24589658e-01 -5.39553046e-01 4.36194688e-01 5.31628251e-01
-5.02829194e-01 -9.99118090e-02 3.84287357e-01 4.21360195e-01
8.14096570e-01 1.18723296e-01 7.47731745e-01 -1.54293561e+00
-1.67832207e-02 -7.05501676e-01 -8.44155788e-01 -9.70403016e-01
-1.67014584e-01 -1.42916012e+00 2.44549364e-02 -1.86779511e+00
-1.56909302e-01 -8.85818303e-01 -2.23759755e-01 3.27875137e-01
2.52411664e-01 4.76606973e-02 1.30146727e-01 5.41072011e-01
1.34956241e-01 3.47486079e-01 1.98198974e+00 2.16678217e-01
-4.86851960e-01 5.36886573e-01 -2.80124605e-01 6.08749151e-01
6.97639823e-01 -3.28223586e-01 -4.41458225e-01 -1.21212497e-01
-6.91164983e-03 5.12941301e-01 5.62526822e-01 -9.61827695e-01
3.51033241e-01 -5.68476226e-03 6.38352036e-01 -1.07525098e+00
5.07744551e-01 -1.27345181e+00 2.41694793e-01 6.62729919e-01
-2.69315213e-01 -2.85143465e-01 2.69688934e-01 3.47613186e-01
1.62584502e-02 -1.94961727e-01 1.02221322e+00 -3.25826585e-01
-1.86266765e-01 9.56383765e-01 5.12319624e-01 -1.87006697e-01
1.17060101e+00 -8.29092920e-01 4.09397840e-01 1.30421847e-01
-1.31730509e+00 -2.46048458e-02 3.18949670e-01 2.34050736e-01
1.22703040e+00 -1.50528157e+00 -7.20631301e-01 4.19860601e-01
8.35145786e-02 8.62053156e-01 4.20345128e-01 1.44937992e+00
-9.68285739e-01 5.66834927e-01 -4.08055596e-02 -1.19792342e+00
-1.17685819e+00 2.77968138e-01 9.19454396e-01 -9.06885639e-02
-1.18688226e+00 7.41067827e-01 9.23684984e-02 -7.18256295e-01
-1.18655907e-02 -4.74467427e-01 -2.81298727e-01 -3.79260778e-01
-7.12040290e-02 8.50889925e-03 2.87573904e-01 -8.79651666e-01
-2.75543153e-01 7.82252491e-01 2.07847163e-01 1.49266183e-01
1.47235644e+00 8.34491327e-02 2.22910829e-02 4.15840179e-01
1.45668101e+00 -3.14297646e-01 -1.38226306e+00 -4.39575128e-02
-2.02385992e-01 -4.40466613e-01 2.91448325e-01 -8.44254673e-01
-1.15233982e+00 7.55229831e-01 9.41116095e-01 -3.25368375e-01
7.04420209e-01 5.65374969e-03 6.54079914e-01 4.78607081e-02
3.36575031e-01 -6.68833077e-01 -1.83039606e-01 3.64872843e-01
1.43999994e+00 -1.33603394e+00 1.02233075e-01 -4.12136495e-01
-5.79589367e-01 1.52862370e+00 5.74997604e-01 -2.00154901e-01
1.04495382e+00 1.57873884e-01 2.71554768e-01 -5.74455678e-01
-9.58692804e-02 1.25386313e-01 8.65285933e-01 6.23357952e-01
2.55543321e-01 2.26673007e-01 -1.87736988e-01 4.97117579e-01
-4.71492589e-01 2.23343089e-01 9.45704281e-02 6.67633772e-01
-3.37247968e-01 -7.20726788e-01 -2.17981353e-01 6.86375856e-01
-1.86193660e-01 1.12657867e-01 5.95653802e-02 7.90790558e-01
1.28673121e-01 1.11835971e-01 2.51972377e-01 -1.87804967e-01
6.31233037e-01 -1.87779739e-01 5.61755776e-01 -5.78972340e-01
-5.93812108e-01 1.43978924e-01 -3.57683748e-01 -8.29243004e-01
-1.45684361e-01 -5.86784363e-01 -1.76413512e+00 1.78307071e-01
-2.52186626e-01 -2.60364056e-01 1.22295499e+00 6.69772565e-01
3.15271497e-01 3.98840934e-01 5.85602045e-01 -1.12560248e+00
-4.99919206e-01 -7.82055557e-01 -4.97958511e-01 2.54680812e-01
2.58635134e-01 -7.87448168e-01 -1.06197529e-01 -2.75885314e-01] | [14.064542770385742, -2.5804617404937744] |
ea671d61-4b3b-498b-8b4a-92a6ad161a6f | task-adaptive-pre-training-and-self-training | 2109.06466 | null | https://arxiv.org/abs/2109.06466v2 | https://arxiv.org/pdf/2109.06466v2.pdf | Task-adaptive Pre-training and Self-training are Complementary for Natural Language Understanding | Task-adaptive pre-training (TAPT) and Self-training (ST) have emerged as the major semi-supervised approaches to improve natural language understanding (NLU) tasks with massive amount of unlabeled data. However, it's unclear whether they learn similar representations or they can be effectively combined. In this paper, we show that TAPT and ST can be complementary with simple TFS protocol by following TAPT -> Finetuning -> Self-training (TFS) process. Experimental results show that TFS protocol can effectively utilize unlabeled data to achieve strong combined gains consistently across six datasets covering sentiment classification, paraphrase identification, natural language inference, named entity recognition and dialogue slot classification. We investigate various semi-supervised settings and consistently show that gains from TAPT and ST can be strongly additive by following TFS procedure. We hope that TFS could serve as an important semi-supervised baseline for future NLP studies. | ['Xifeng Yan', 'Wenhu Chen', 'Semih Yavuz', 'Shiyang Li'] | 2021-09-14 | null | https://aclanthology.org/2021.findings-emnlp.86 | https://aclanthology.org/2021.findings-emnlp.86.pdf | findings-emnlp-2021-11 | ['paraphrase-identification'] | ['natural-language-processing'] | [ 5.6523401e-01 4.2579386e-01 -7.8283262e-01 -8.7820506e-01
-1.0385566e+00 -7.7721006e-01 8.4415203e-01 1.5129682e-01
-4.9528354e-01 9.8763359e-01 4.3466070e-01 -5.4988605e-01
1.6067359e-01 -3.2749540e-01 -8.0355787e-01 -1.7495482e-01
3.2702246e-01 7.3420691e-01 -5.5218134e-02 -2.7196905e-01
1.1294833e-01 2.0165490e-02 -1.1439041e+00 6.3496995e-01
1.1514717e+00 7.2568446e-01 -3.9977971e-02 4.6930185e-01
-6.6407943e-01 1.0402418e+00 -2.4434091e-01 -3.1343919e-01
1.6877663e-01 -4.7900718e-01 -1.2948110e+00 2.1774633e-02
3.5495815e-01 -7.1659878e-02 2.4417201e-02 5.8388060e-01
5.0326222e-01 2.3861243e-01 6.8284732e-01 -1.0930198e+00
-5.2791077e-01 9.7259086e-01 -6.3328415e-01 1.6434203e-01
4.4982180e-01 -4.5629587e-02 1.2832643e+00 -9.0660882e-01
6.9285125e-01 1.2918229e+00 7.1654320e-01 7.7477390e-01
-1.3217341e+00 -5.5674112e-01 1.0401671e-02 -7.4663974e-02
-7.5338316e-01 -7.0182317e-01 5.7725763e-01 -9.0001412e-02
1.1548184e+00 1.5426025e-01 9.4122089e-02 1.2406342e+00
-2.7380896e-01 1.4617126e+00 1.4041218e+00 -8.4282488e-01
1.2410804e-01 5.0532693e-01 7.6432341e-01 6.5730882e-01
-6.8145432e-02 -7.2686546e-02 -6.9388425e-01 -2.7530554e-01
3.6153248e-01 -2.4698257e-01 -1.1234055e-02 -3.9772338e-01
-1.2748207e+00 1.0216322e+00 1.2165674e-01 2.9895428e-01
-1.4311239e-01 -1.3871452e-01 7.4494702e-01 7.2958541e-01
7.4730599e-01 8.2493198e-01 -1.0871228e+00 -3.0923921e-01
-7.4346864e-01 -1.1228968e-01 9.7564822e-01 1.1029152e+00
8.4796584e-01 1.2024211e-01 -2.8687447e-01 1.3733201e+00
-4.3283394e-03 3.6395469e-01 1.0130970e+00 -9.2194211e-01
5.6880623e-01 7.4204195e-01 2.0554036e-02 -7.3411278e-02
-4.9888521e-01 -7.0286378e-02 -5.9028161e-01 -4.8120108e-01
5.5824089e-01 -6.5945745e-01 -8.6276376e-01 1.9700118e+00
2.9298207e-01 1.9097762e-02 3.7061453e-01 4.3210468e-01
8.6649156e-01 6.0496926e-01 3.0926567e-01 -2.9681379e-01
1.3684551e+00 -1.2546862e+00 -6.5467000e-01 -5.4890209e-01
1.1613140e+00 -6.8126374e-01 1.2784022e+00 1.1356506e-01
-8.6950982e-01 -4.9982128e-01 -7.1644741e-01 -2.1371922e-01
-4.1750577e-01 2.5682077e-01 9.8861212e-01 8.2486492e-01
-8.3232641e-01 5.4884171e-01 -5.8522171e-01 -6.4112747e-01
4.5562744e-01 3.6698034e-01 -5.1028466e-01 -2.2480917e-01
-1.3155098e+00 8.5626346e-01 3.7141475e-01 -2.4049595e-01
-6.7859316e-01 -8.4036040e-01 -1.0424997e+00 2.6840644e-02
5.3952503e-01 -6.8778998e-01 1.5018528e+00 -1.0260345e+00
-1.7035553e+00 1.0601956e+00 -6.2515223e-01 -8.4226859e-01
2.8929040e-01 -5.5168027e-01 -1.2579205e-02 -4.3340698e-02
2.5875673e-01 8.4794879e-01 6.9116473e-01 -1.0156375e+00
-4.8752651e-01 -4.5346406e-01 -1.3062228e-02 4.5884234e-01
-5.9778458e-01 5.8361422e-02 -5.3917002e-02 -4.0163478e-01
-3.3711623e-02 -1.0427332e+00 -2.8974012e-01 -3.9832237e-01
-4.9932337e-01 -6.0016626e-01 6.8645298e-01 -4.1322717e-01
9.5680976e-01 -1.9125783e+00 -6.6984467e-02 -2.5366807e-01
-1.7412266e-02 6.0496366e-01 -3.4836099e-01 5.2550739e-01
-1.5606233e-01 1.7876065e-01 -2.5564393e-01 -4.3406752e-01
-5.6721345e-03 4.2561108e-01 -4.7578374e-01 1.5800556e-02
3.2812342e-01 1.2829419e+00 -9.8238903e-01 -4.4648358e-01
6.4721584e-02 -1.0899543e-01 -3.4370431e-01 2.3032381e-01
-5.2942479e-01 5.1376969e-01 -6.1607778e-01 5.0248110e-01
4.0688375e-01 -5.5222392e-01 2.7208456e-01 -9.2368402e-02
1.6251606e-01 5.8401489e-01 -7.1976477e-01 1.7273262e+00
-6.1857438e-01 5.8210862e-01 -8.4311798e-02 -1.4314936e+00
9.1203159e-01 3.6569920e-01 1.8947475e-01 -7.7353686e-01
-4.2644879e-05 1.2920451e-01 -2.3056835e-01 -4.9726805e-01
4.8015735e-01 -2.9477569e-01 -2.3367386e-01 8.7811315e-01
4.6443480e-01 -1.8684130e-02 1.5551730e-01 4.8589531e-01
9.9697107e-01 1.2408970e-01 5.8650243e-01 -3.8967186e-01
5.2101952e-01 3.2022065e-01 4.3893564e-01 9.8962319e-01
-3.3241394e-01 2.1244779e-01 5.2695358e-01 -2.7337149e-01
-9.9419576e-01 -8.4691530e-01 -1.9855630e-01 1.7976220e+00
-2.7229005e-01 -1.0120588e-01 -5.4942828e-01 -1.2231820e+00
-1.5361439e-02 7.1667647e-01 -5.7830548e-01 -8.6344928e-02
-3.8933212e-01 -6.9029301e-01 8.2075089e-01 6.8542820e-01
7.0531142e-01 -1.0889562e+00 4.5320433e-02 9.3937382e-02
-1.2225544e-01 -1.4079794e+00 -4.8638478e-01 7.2647005e-01
-1.0807170e+00 -8.2399768e-01 -7.3611605e-01 -1.0174460e+00
7.4866611e-01 5.4375625e-01 1.1119152e+00 -2.9219598e-01
8.4850930e-02 4.3311852e-01 -6.3754803e-01 -3.0939332e-01
-6.2866062e-01 4.9610490e-01 2.9166889e-01 -1.8768933e-01
5.4586244e-01 -4.5494941e-01 -2.7878416e-01 4.1134557e-01
-7.0366979e-01 1.2953568e-01 5.8465093e-01 1.1743242e+00
2.3718435e-01 -4.4815055e-01 1.0018456e+00 -1.8384895e+00
8.2017982e-01 -5.9931612e-01 -7.4902378e-02 4.3441337e-01
-7.4757731e-01 4.2768487e-01 6.7875022e-01 -5.5186492e-01
-1.5358814e+00 2.0740451e-01 -1.2878311e-01 -1.8586834e-01
-4.3409148e-01 6.2952065e-01 6.0977239e-02 2.5168173e-02
8.8877481e-01 2.4560846e-01 -9.5188312e-02 -5.0648761e-01
5.9633011e-01 1.0252390e+00 2.3427692e-01 -8.4506917e-01
5.7272542e-01 3.5856894e-01 -5.9440094e-01 -7.9906291e-01
-1.5732061e+00 -8.4092015e-01 -7.0497215e-01 3.6887941e-01
5.9109622e-01 -1.0928910e+00 -2.6040781e-01 1.5033688e-01
-7.9999840e-01 -6.0514909e-01 -2.8801733e-01 3.1287456e-01
-4.0450123e-01 5.5757165e-01 -6.5218282e-01 -8.4799606e-01
-5.4441851e-01 -6.9293493e-01 1.0258127e+00 4.9719304e-01
-4.9169225e-01 -1.1458172e+00 1.6453600e-01 9.8067206e-01
3.1908044e-01 -3.5407835e-01 8.9159089e-01 -1.4015670e+00
7.2431348e-02 -5.3532097e-02 -2.8701293e-01 5.2944773e-01
1.7669131e-01 -5.6073785e-01 -1.2673965e+00 -2.6305845e-01
-1.8589988e-01 -1.1781735e+00 9.0097982e-01 2.6842093e-01
8.4966272e-01 -4.2071530e-01 -2.4547535e-01 3.0509761e-01
1.0192512e+00 3.1374756e-02 3.5033479e-01 2.4214825e-01
5.8790028e-01 7.8252167e-01 7.1320850e-01 1.8454537e-01
2.9033557e-01 3.2193542e-01 -3.9317733e-01 -1.3195053e-01
7.8554526e-02 -4.9276221e-01 4.6349981e-01 7.8708625e-01
3.5094562e-01 -1.8563407e-02 -8.9862895e-01 6.2783813e-01
-1.8397335e+00 -7.4239802e-01 3.5523439e-03 2.0036077e+00
1.2007689e+00 1.5529610e-01 1.3926658e-01 8.7795794e-02
7.2279161e-01 5.3386785e-02 -7.9252362e-01 -5.3297865e-01
-1.6172615e-01 3.0286744e-01 3.2263476e-01 2.6887289e-01
-1.0992681e+00 1.3529676e+00 6.6966171e+00 9.5463449e-01
-9.1167074e-01 2.7946219e-01 7.8690755e-01 3.5131890e-01
-3.0571038e-01 6.7882299e-02 -9.4377005e-01 8.0045514e-02
1.1347682e+00 -3.5389984e-01 1.4940663e-01 9.7675675e-01
-4.6266474e-02 -1.1907067e-01 -1.1188450e+00 6.0089231e-01
-7.6975606e-02 -1.1606765e+00 4.8823006e-02 -3.4426793e-01
1.0445322e+00 3.3816388e-01 -9.0154476e-02 9.8893327e-01
7.5401717e-01 -8.0236441e-01 -2.0662803e-02 -1.0822734e-01
7.4626291e-01 -3.9078006e-01 7.0307744e-01 6.2344748e-01
-9.7384232e-01 -4.3559507e-02 -3.7861100e-01 -3.9794724e-02
1.0062277e-01 3.5536414e-01 -1.2177140e+00 4.6844319e-01
3.5473749e-01 7.3624438e-01 -5.1153004e-01 5.4304183e-01
-3.1416818e-01 1.1271955e+00 -3.1534752e-01 -2.9379076e-01
3.3751127e-01 -2.1409543e-02 3.1046060e-01 1.3010708e+00
-2.9569310e-01 1.3108021e-01 2.0980547e-01 7.0557147e-01
-3.8893530e-01 3.8865849e-01 -6.3567770e-01 -5.7246447e-01
5.1315904e-01 1.2021632e+00 -6.1560994e-01 -6.2379175e-01
-4.9572626e-01 9.7828609e-01 6.6947877e-01 2.9170257e-01
-4.6320564e-01 -2.1051255e-01 3.9413747e-01 -2.2631666e-01
2.5807917e-01 1.0767482e-01 -5.2378350e-01 -1.5898020e+00
-3.6774534e-01 -8.5554701e-01 5.6587988e-01 -6.4786482e-01
-1.5441808e+00 4.2156163e-01 -6.0918968e-02 -1.0720025e+00
-3.0544823e-01 -5.0200605e-01 -5.6348181e-01 3.6999732e-01
-1.5986682e+00 -1.1230773e+00 1.9467556e-01 6.4683521e-01
9.9036515e-01 -2.8452900e-01 1.0493727e+00 -6.8526268e-02
-7.7614462e-01 7.5963730e-01 3.6054429e-01 2.3552278e-01
8.8680011e-01 -1.4164432e+00 3.3206141e-01 5.3243214e-01
4.6165103e-01 8.8243759e-01 4.3306053e-01 -5.0237072e-01
-1.2165134e+00 -9.4045448e-01 9.5618594e-01 -5.2404606e-01
8.0001742e-01 -3.4993124e-01 -9.5366794e-01 8.7948585e-01
3.7679392e-01 -1.4457084e-01 1.0203122e+00 6.4470327e-01
-6.1270702e-01 2.0280087e-01 -1.0282975e+00 2.9599470e-01
9.4635046e-01 -7.8028131e-01 -9.8008662e-01 6.6144615e-01
9.4115865e-01 -3.1942043e-01 -8.2098091e-01 4.0611264e-01
3.2195491e-01 -7.2795761e-01 8.4915245e-01 -1.0060395e+00
5.6428087e-01 3.8383928e-01 7.2502345e-02 -1.2837905e+00
7.9003975e-02 -7.3177844e-01 1.1644169e-02 1.3800255e+00
8.1610429e-01 -6.9162440e-01 9.6202004e-01 6.5626007e-01
-7.7663079e-02 -6.2874979e-01 -7.5887346e-01 -6.9209766e-01
3.4170333e-01 -3.1091818e-01 1.3222955e-01 1.3239495e+00
5.2687681e-01 1.0833334e+00 -5.6372064e-01 -3.5770211e-01
4.6900997e-01 1.9894218e-01 8.4336019e-01 -1.2131112e+00
-3.1645793e-01 -1.6472827e-01 2.6221922e-01 -1.3821683e+00
5.6334448e-01 -9.2444932e-01 9.2994384e-02 -1.3003320e+00
3.8817060e-01 -4.5265016e-01 -1.6832139e-01 8.2186908e-01
-4.4461265e-01 -3.2611806e-02 -8.6184032e-02 2.9382855e-01
-8.9421284e-01 6.1837935e-01 1.1182067e+00 -2.4161277e-02
-3.3283922e-01 7.0468359e-02 -8.6263204e-01 5.5527341e-01
9.1587454e-01 -4.2017528e-01 -5.4180038e-01 -2.7841467e-01
6.9866307e-02 1.0672491e-01 -3.0394056e-01 -4.4125211e-01
7.7994347e-02 -4.6224728e-02 1.0619120e-01 -2.4023038e-01
1.6625375e-01 -3.4270000e-01 -7.4775350e-01 2.7117819e-01
-8.7902838e-01 -4.2189887e-01 4.3474397e-01 6.8448019e-01
-1.7474066e-01 -3.8417044e-01 7.1420091e-01 -2.4434355e-01
-7.3675478e-01 2.3660727e-02 -3.6686850e-01 5.1225179e-01
7.3572385e-01 5.1323674e-03 -4.3411559e-01 -5.0687152e-01
-7.0120585e-01 5.7174683e-01 8.0622941e-02 4.3097737e-01
2.3894049e-01 -9.1077989e-01 -6.3623893e-01 1.4275531e-01
1.7205076e-01 -4.6572588e-02 2.4960446e-01 7.5551641e-01
-4.1182302e-02 8.1180596e-01 4.4914979e-02 -5.3558183e-01
-1.2339878e+00 3.8271374e-01 -2.7204985e-02 -7.3031551e-01
-1.9415566e-01 1.0449196e+00 2.8633982e-01 -1.2046490e+00
1.7911209e-01 -7.6819710e-02 -2.8694531e-01 2.5331348e-02
4.1716644e-01 1.7394381e-02 -1.5546840e-01 -3.5511580e-01
-1.2098254e-02 1.9227131e-01 -6.1051911e-01 -1.0899873e-01
1.3335611e+00 -3.2819676e-01 2.2768202e-01 6.1763269e-01
1.2428267e+00 -1.9438116e-01 -9.0660250e-01 -8.1199974e-01
3.2976606e-01 6.7305104e-03 -1.7967770e-01 -8.2543242e-01
-6.8288755e-01 7.8814483e-01 1.3600595e-01 1.6634293e-01
7.2653270e-01 1.9527932e-01 9.7749919e-01 9.5011944e-01
3.6457792e-01 -1.0313604e+00 1.8831843e-01 6.4956987e-01
4.2777100e-01 -1.6305894e+00 -1.2947348e-01 -4.9436039e-01
-1.1624430e+00 7.9071242e-01 8.1061190e-01 1.9891713e-02
3.9610019e-01 9.0920523e-02 1.5925261e-01 2.3379480e-02
-1.0089118e+00 -4.1625196e-01 1.9878149e-01 5.2807635e-01
8.0368441e-01 -2.0956351e-01 -3.4492517e-01 6.9505250e-01
-1.7895247e-01 8.4963599e-03 2.6363581e-01 1.0373566e+00
-4.3046519e-01 -1.3381174e+00 -9.9840961e-02 7.9615474e-01
-3.2616526e-01 -2.8773430e-01 -6.2869734e-01 6.1601603e-01
-4.1845745e-01 9.7526515e-01 -3.2672751e-01 -3.5914221e-01
8.2918145e-02 6.5267354e-01 3.3931831e-01 -9.7586167e-01
-6.6768074e-01 -3.8181104e-02 5.5547404e-01 -3.7602496e-01
-6.5225857e-01 -5.9912127e-01 -1.2059890e+00 3.2220382e-02
-6.4457595e-01 5.0154650e-01 2.4367158e-01 1.3120354e+00
4.4114470e-01 1.2667739e-01 7.1522605e-01 -4.4168165e-01
-8.0255276e-01 -1.3996463e+00 -3.7384456e-01 5.2227604e-01
9.1216899e-02 -5.0077009e-01 -5.4550546e-01 1.7826429e-01] | [10.812566757202148, 8.158119201660156] |
eb54244d-4df7-43e8-96c6-b42d3217251a | causality-driven-hierarchical-structure | 2210.06964 | null | https://arxiv.org/abs/2210.06964v1 | https://arxiv.org/pdf/2210.06964v1.pdf | Causality-driven Hierarchical Structure Discovery for Reinforcement Learning | Hierarchical reinforcement learning (HRL) effectively improves agents' exploration efficiency on tasks with sparse reward, with the guide of high-quality hierarchical structures (e.g., subgoals or options). However, how to automatically discover high-quality hierarchical structures is still a great challenge. Previous HRL methods can hardly discover the hierarchical structures in complex environments due to the low exploration efficiency by exploiting the randomness-driven exploration paradigm. To address this issue, we propose CDHRL, a causality-driven hierarchical reinforcement learning framework, leveraging a causality-driven discovery instead of a randomness-driven exploration to effectively build high-quality hierarchical structures in complicated environments. The key insight is that the causalities among environment variables are naturally fit for modeling reachable subgoals and their dependencies and can perfectly guide to build high-quality hierarchical structures. The results in two complex environments, 2D-Minecraft and Eden, show that CDHRL significantly boosts exploration efficiency with the causality-driven paradigm. | ['Yunji Chen', 'Qi Guo', 'Ling Li', 'Zidong Du', 'Xishan Zhang', 'Ruizhi Chen', 'Qi Yi', 'Jiaming Guo', 'Ke Tang', 'Rui Zhang', 'Xing Hu', 'Shaohui Peng'] | 2022-10-13 | null | null | null | null | ['hierarchical-reinforcement-learning'] | ['methodology'] | [-4.39007461e-01 -1.59497689e-02 -2.35914662e-01 3.26676704e-02
-4.03911263e-01 -4.63742018e-01 2.97010750e-01 1.21527780e-02
-2.06485227e-01 1.25275660e+00 5.67336857e-01 -2.42503345e-01
-6.02744758e-01 -9.99343693e-01 -6.62177563e-01 -7.03725159e-01
-9.22507465e-01 6.35031521e-01 3.31770629e-01 -6.48824990e-01
3.47177178e-01 2.50545055e-01 -1.90416276e+00 -2.12849885e-01
1.36845899e+00 4.93531376e-01 6.04934454e-01 4.12263095e-01
4.13960814e-02 8.62563789e-01 -3.65581095e-01 4.05633003e-01
4.84579265e-01 -4.60585803e-01 -8.93785238e-01 -2.94272810e-01
-3.17152888e-01 -5.56842983e-01 -3.18476021e-01 7.80786514e-01
3.34087551e-01 3.94381225e-01 1.40954182e-01 -1.22731745e+00
-2.15093553e-01 1.15816760e+00 -4.26067531e-01 5.95822670e-02
3.10893953e-01 5.75764358e-01 1.53124511e+00 -2.78567612e-01
6.10311925e-01 1.56780255e+00 2.85767496e-01 5.11560559e-01
-1.24266088e+00 -7.47580528e-01 5.39510012e-01 4.18314725e-01
-1.08281577e+00 -5.60782338e-03 6.36071205e-01 -4.27182704e-01
9.86120701e-01 1.13231793e-01 1.11104965e+00 1.09624577e+00
2.63045996e-01 9.26968873e-01 1.47541797e+00 -8.11185092e-02
8.12575758e-01 -4.70292866e-01 -7.56852329e-02 8.98714423e-01
3.41100097e-01 1.07520223e+00 -9.74801898e-01 -8.56531635e-02
1.10600471e+00 -2.26766422e-01 8.31005424e-02 -5.99198580e-01
-1.18094945e+00 1.03956592e+00 7.97385633e-01 1.46104485e-01
-4.56557751e-01 2.53197581e-01 2.28435516e-01 2.20895857e-01
-5.53617597e-01 1.29636419e+00 -3.63438874e-01 -4.96451378e-01
-4.80069160e-01 6.20844066e-01 5.83950222e-01 1.00166106e+00
1.08115256e+00 3.49293560e-01 -2.90812433e-01 3.01943153e-01
3.25327814e-01 9.84084383e-02 5.03182590e-01 -1.12588620e+00
2.84133583e-01 8.84068251e-01 3.07647109e-01 -7.65508950e-01
-7.61061132e-01 -7.61954904e-01 -6.76228762e-01 6.84825420e-01
6.37158006e-02 -3.04353178e-01 -7.28015661e-01 2.03626752e+00
4.97852653e-01 -3.18510652e-01 2.36028120e-01 8.84367108e-01
4.88258511e-01 4.75699782e-01 -8.50438625e-02 -1.31533667e-01
1.04600537e+00 -1.19715786e+00 -7.19959497e-01 -2.21067861e-01
5.86014867e-01 2.25242957e-01 1.46221173e+00 4.07348394e-01
-1.01424658e+00 -4.58333284e-01 -1.11966097e+00 2.94601858e-01
-6.72665313e-02 -3.81402224e-01 1.08510244e+00 2.78355926e-01
-9.59919631e-01 6.87687278e-01 -8.81421924e-01 5.49125448e-02
3.63121271e-01 3.31773937e-01 -1.54945143e-02 3.98464166e-02
-1.42500603e+00 7.31427312e-01 8.81525218e-01 -1.54967859e-01
-1.75194252e+00 -2.59444445e-01 -9.03875828e-01 1.62082404e-01
9.53358769e-01 -4.76388395e-01 1.18514276e+00 -3.04381132e-01
-1.66748738e+00 -1.06163971e-01 5.97128868e-02 -5.24168134e-01
4.53521252e-01 -3.47384453e-01 1.93378970e-01 -6.84215799e-02
3.80310595e-01 8.30563784e-01 6.45683110e-01 -1.30815792e+00
-8.85293007e-01 -1.89223066e-01 4.10832345e-01 6.63113415e-01
-2.19963357e-01 -7.13975608e-01 -9.56200361e-02 -3.20191890e-01
5.59065607e-04 -8.63294184e-01 -9.81490195e-01 -6.07425749e-01
-5.26965320e-01 -5.05041778e-01 2.59698331e-01 -2.40550220e-01
1.45947504e+00 -1.68405259e+00 6.11911833e-01 2.22969890e-01
4.35764909e-01 -3.77857089e-01 -2.05535203e-01 6.03850305e-01
4.69801575e-01 8.58145580e-02 -1.24304831e-01 1.35306478e-01
3.07447940e-01 7.60803998e-01 -1.85707018e-01 -1.86567008e-01
-1.26969367e-01 1.09608769e+00 -1.46389818e+00 -3.50441128e-01
4.12604176e-02 -2.23227024e-01 -8.63408625e-01 6.03094697e-01
-5.28456211e-01 7.01478422e-01 -8.88505936e-01 5.96202374e-01
1.73367456e-01 -3.03859264e-01 5.46703696e-01 2.78409034e-01
-3.85777831e-01 4.00696754e-01 -1.18415272e+00 1.71845090e+00
-4.21293914e-01 2.00967789e-01 -9.47570335e-03 -5.65675855e-01
1.05088770e+00 -1.82451382e-01 3.85788649e-01 -1.20030665e+00
-2.84503460e-01 1.71051025e-01 3.11896026e-01 -3.22573364e-01
6.38414443e-01 -5.57102151e-02 -3.73513550e-01 2.77794838e-01
-2.18281269e-01 -1.23187058e-01 5.05187571e-01 1.11893773e-01
1.34868836e+00 4.39341128e-01 5.74872613e-01 -4.62785214e-01
-1.90828219e-02 2.39043564e-01 1.01778495e+00 1.19145250e+00
-7.66407326e-02 -1.80326328e-01 7.52575696e-01 -5.46262205e-01
-9.27873492e-01 -1.01263404e+00 2.00596839e-01 1.28550398e+00
4.74311650e-01 -8.25378358e-01 -4.95894641e-01 -6.87504828e-01
-9.82394964e-02 6.85757756e-01 -1.00383544e+00 -1.55057192e-01
-6.81154847e-01 -5.29725850e-01 2.47954920e-01 5.18073797e-01
6.14450991e-01 -1.46442389e+00 -1.46282852e+00 4.42130715e-01
-5.72868213e-02 -6.82808697e-01 -4.29283939e-02 7.90887356e-01
-9.71544802e-01 -1.11701620e+00 -1.62907287e-01 -5.22669673e-01
4.20527458e-01 -3.26467939e-02 1.30707812e+00 1.09500751e-01
-2.91540235e-01 -4.56477851e-02 -5.32846630e-01 9.98452455e-02
-1.23606689e-01 1.18362039e-01 1.96934819e-01 -7.34387100e-01
-1.61246523e-01 -8.69122028e-01 -5.42046607e-01 4.63968724e-01
-6.33442819e-01 2.88481861e-01 8.71858716e-01 1.22985137e+00
5.62405884e-01 8.50017607e-01 6.28828883e-01 -4.58093882e-01
7.99622416e-01 -4.50118572e-01 -9.56872761e-01 -3.13295498e-02
-8.33891094e-01 7.17389047e-01 8.03986073e-01 -2.15406746e-01
-8.31590116e-01 2.46519446e-02 -5.70319332e-02 -2.45572314e-01
-1.56632125e-01 6.37064159e-01 -2.67785907e-01 2.42665812e-01
7.71110356e-01 3.39494824e-01 -4.08086330e-01 -3.72549385e-01
4.21661079e-01 -1.73344314e-01 2.89113492e-01 -1.12972462e+00
6.95868134e-01 9.57754701e-02 1.75559700e-01 -3.83286893e-01
-7.74407744e-01 -1.38377815e-01 -3.55679840e-01 -2.40792241e-02
6.10715270e-01 -7.95655429e-01 -1.07785273e+00 5.57538308e-02
-5.95329881e-01 -9.36738431e-01 -6.41715765e-01 2.38290265e-01
-8.95645618e-01 9.34185088e-02 -4.39783543e-01 -9.53730941e-01
-1.71326455e-02 -1.54315245e+00 7.54036903e-01 3.66169214e-01
-2.03870624e-01 -6.51678920e-01 2.80481815e-01 -6.09828122e-02
3.01397443e-01 3.49938273e-01 1.15928805e+00 -1.46913901e-01
-1.07615256e+00 5.91800094e-01 9.38097313e-02 -5.00607789e-01
1.41317695e-01 -3.89638841e-01 -2.49779388e-01 -4.36948895e-01
-2.84635723e-01 -8.46723020e-01 6.05292916e-01 3.05850983e-01
1.15596950e+00 -7.86885917e-01 -1.92759216e-01 4.80211496e-01
1.23008156e+00 3.24973613e-01 5.57919383e-01 7.96748698e-01
4.30958629e-01 6.89376593e-01 9.50478435e-01 1.03293526e+00
6.78773761e-01 7.62155056e-01 1.02056122e+00 2.02378795e-01
1.23632565e-01 -9.36727107e-01 2.76587695e-01 5.14427662e-01
4.00886312e-02 8.67895558e-02 -8.95846844e-01 6.20662212e-01
-2.16063070e+00 -7.13400841e-01 8.91793594e-02 1.91347349e+00
1.24073219e+00 3.49675387e-01 3.89057249e-01 -1.39343813e-01
2.17706159e-01 9.87414196e-02 -1.08341515e+00 -1.20798878e-01
-2.21752599e-02 -1.19908974e-01 1.39957249e-01 6.46304309e-01
-6.88439786e-01 1.23286891e+00 6.79368973e+00 8.84403706e-01
-5.28157353e-01 -2.91736513e-01 2.18555406e-01 -1.58296451e-02
-4.93222564e-01 1.63949266e-01 -1.08034110e+00 3.11344296e-01
4.07273471e-01 -4.10500467e-02 8.72271121e-01 1.08150125e+00
2.33649597e-01 -1.77081913e-01 -1.02036130e+00 7.46943116e-01
-7.59041607e-01 -1.30424786e+00 6.69792816e-02 3.19066107e-01
8.36387873e-01 -1.80732116e-01 -1.13138231e-02 6.47506237e-01
1.39248621e+00 -1.46588707e+00 6.71138525e-01 2.26518333e-01
5.08587718e-01 -8.94111276e-01 2.91260064e-01 6.30799234e-01
-1.31955278e+00 -6.20570660e-01 -3.58292878e-01 -3.71499389e-01
3.75030972e-02 2.70055890e-01 -9.26959634e-01 5.27130604e-01
1.04508018e+00 8.54698718e-01 -4.10410970e-01 9.40584779e-01
-7.44293988e-01 4.52924430e-01 -1.46520242e-01 -4.58336771e-01
7.09712923e-01 -1.51438892e-01 7.06831992e-01 6.95245743e-01
1.50999293e-01 1.22912340e-01 6.02081120e-01 1.07068527e+00
4.24746454e-01 -3.13748807e-01 -5.47874510e-01 -2.34246533e-02
6.94129765e-01 1.18842459e+00 -5.54861903e-01 -9.02558118e-02
3.29586983e-01 3.75410467e-01 7.49473512e-01 2.87811279e-01
-6.52909517e-01 -9.27203521e-02 6.28955722e-01 -1.48798048e-01
2.94122875e-01 -5.76488733e-01 -4.66230929e-01 -9.13424134e-01
-3.17048341e-01 -1.34930301e+00 4.15926933e-01 -4.75771189e-01
-1.04277492e+00 7.17683792e-01 -1.03676960e-01 -1.12852383e+00
-3.63510668e-01 -9.32871178e-02 -4.91094112e-01 4.35273379e-01
-1.63544309e+00 -6.78273857e-01 -4.74520475e-01 6.56339586e-01
6.94824994e-01 -2.84538716e-01 7.71416843e-01 -5.11626422e-01
-5.57980239e-01 4.05138224e-01 -1.84630677e-02 -3.85634929e-01
3.47078294e-02 -1.67729402e+00 2.34397098e-01 6.14105284e-01
-1.23724982e-01 6.63905859e-01 7.78148770e-01 -9.19224203e-01
-1.46669912e+00 -7.86567032e-01 4.55527119e-02 -1.49894312e-01
6.14801288e-01 -3.28784704e-01 -8.82271588e-01 2.33651668e-01
1.89500824e-01 -5.96731782e-01 4.35037553e-01 5.56536555e-01
-2.50309467e-01 1.34265035e-01 -7.26345420e-01 9.91836071e-01
1.50200534e+00 3.37232724e-02 -8.10898662e-01 4.69565615e-02
9.80508924e-01 -3.63680691e-01 -8.01533699e-01 5.67964017e-01
4.33267802e-01 -1.16609037e+00 9.17549372e-01 -7.06522167e-01
5.63011110e-01 -5.39629638e-01 -1.24147244e-01 -1.60518920e+00
-7.15489924e-01 -1.09242451e+00 -4.17445332e-01 6.01109862e-01
1.54874220e-01 -4.33901489e-01 8.11839342e-01 1.68429807e-01
-1.69295371e-01 -9.55102205e-01 -8.48027468e-01 -1.01119542e+00
-4.04034331e-02 7.39052966e-02 9.64968562e-01 7.60258138e-01
2.86883205e-01 5.23936868e-01 -6.67279541e-01 1.10771179e-01
8.75171900e-01 5.34285009e-01 8.97992134e-01 -1.12285316e+00
-7.43288994e-01 -5.58867276e-01 2.45496020e-01 -1.30037093e+00
6.49331957e-02 -4.32947129e-01 5.09315312e-01 -1.67983592e+00
-1.20070076e-03 -1.11734498e+00 -2.22725689e-01 7.87107646e-01
-2.46371284e-01 -6.64317369e-01 1.23007894e-01 2.98582464e-01
-1.05987406e+00 1.22250557e+00 1.76297784e+00 1.41993001e-01
-6.05779529e-01 -2.07871705e-01 -8.13314736e-01 6.78168178e-01
9.49475586e-01 -3.22938591e-01 -6.71211064e-01 -1.31797180e-01
5.84871650e-01 5.13268709e-01 1.57983050e-01 -1.00554156e+00
2.17190966e-01 -7.47497261e-01 2.59865969e-01 -7.30415285e-01
1.42210916e-01 -6.67655289e-01 5.69671020e-02 8.96002293e-01
-6.89769030e-01 2.77708381e-01 1.63476437e-01 7.53841817e-01
-1.91514771e-02 -8.32580179e-02 4.87453938e-01 -5.03429294e-01
-1.04906762e+00 2.28139117e-01 -5.61606586e-01 2.09914267e-01
9.61455345e-01 -2.48672619e-01 -2.30043501e-01 -3.40821207e-01
-6.60586715e-01 9.06798482e-01 3.92849624e-01 3.78008246e-01
7.73450017e-01 -1.25165927e+00 -4.05383319e-01 1.89738959e-01
-4.81943563e-02 4.94664848e-01 1.72744051e-01 3.71963888e-01
-6.97043464e-02 2.65559673e-01 -7.89432466e-01 -4.03782368e-01
-8.10076416e-01 5.56933284e-01 2.57762462e-01 -9.93786514e-01
-7.21632242e-01 7.38040447e-01 5.93232989e-01 -4.86709207e-01
4.12793517e-01 -3.63699079e-01 -4.69683558e-01 -1.81865811e-01
3.57937604e-01 4.94055122e-01 -5.22742033e-01 1.86781630e-01
-2.12637350e-01 3.88473064e-01 5.53903468e-02 -1.73685998e-01
1.36873996e+00 -1.18201889e-01 3.59951071e-02 1.47537574e-01
3.62914801e-01 -1.32397875e-01 -1.90671468e+00 -6.86557516e-02
2.70035893e-01 -5.74640393e-01 3.73931788e-02 -9.99274909e-01
-6.77146673e-01 8.44653130e-01 1.88146427e-01 2.57679164e-01
1.02372479e+00 -7.38745555e-02 5.61354637e-01 7.94299662e-01
9.62811470e-01 -1.19849265e+00 8.31187189e-01 9.04339790e-01
1.05599964e+00 -8.53358984e-01 -1.68009877e-01 7.20617622e-02
-8.60958040e-01 9.77015734e-01 1.21811652e+00 8.00831337e-03
3.49528939e-02 2.62179852e-01 -5.13553917e-01 -2.78401315e-01
-1.09417617e+00 -6.61702394e-01 -7.69944489e-02 8.26676488e-01
-1.16986945e-01 2.94139594e-01 3.58921103e-02 4.97490793e-01
-5.28510809e-01 -3.35186660e-01 4.39215243e-01 9.33039725e-01
-8.94585490e-01 -1.19031084e+00 -2.12974429e-01 2.87015140e-01
2.84257025e-01 -2.19128840e-02 -1.32250458e-01 6.92726195e-01
-3.59948948e-02 8.52423489e-01 -3.77445638e-01 -5.69920659e-01
2.19451979e-01 -6.27945840e-01 4.02389228e-01 -6.81554198e-01
-5.64588845e-01 4.47697788e-02 1.65223181e-01 -1.09579480e+00
7.03514740e-02 -5.49975216e-01 -1.61120880e+00 -2.54376363e-02
4.19508778e-02 5.70605457e-01 1.76443458e-01 7.16900229e-01
4.15635824e-01 8.97034347e-01 6.99687719e-01 -6.38920903e-01
-7.15335965e-01 -5.16547263e-01 -5.84841847e-01 1.11913376e-01
4.46918458e-01 -1.24331629e+00 -8.00092444e-02 -6.10765338e-01] | [3.99023175239563, 1.675085186958313] |
01b41ac5-8ef9-4ff9-82fc-cf09bc2200d1 | dechorate-a-calibrated-room-impulse-response | 2104.13168 | null | https://arxiv.org/abs/2104.13168v1 | https://arxiv.org/pdf/2104.13168v1.pdf | dEchorate: a Calibrated Room Impulse Response Database for Echo-aware Signal Processing | This paper presents dEchorate: a new database of measured multichannel Room Impulse Responses (RIRs) including annotations of early echo timings and 3D positions of microphones, real sources and image sources under different wall configurations in a cuboid room. These data provide a tool for benchmarking recent methods in echo-aware speech enhancement, room geometry estimation, RIR estimation, acoustic echo retrieval, microphone calibration, echo labeling and reflectors estimation. The database is accompanied with software utilities to easily access, manipulate and visualize the data as well as baseline methods for echo-related tasks. | ['Sharon Gannot', 'Nancy Bertin', 'Antoine Deleforge', 'Cédric Foy', 'Pinchas Tandeitnik', 'Diego Di Carlo'] | 2021-04-27 | null | null | null | null | ['room-impulse-response'] | ['audio'] | [ 8.46731663e-02 -6.50098801e-01 1.11791289e+00 -3.72066915e-01
-1.43368161e+00 -9.66623008e-01 3.82342637e-01 6.94345459e-02
-2.73534179e-01 2.86809444e-01 7.51298189e-01 -3.85954410e-01
-6.09371886e-02 -1.56950448e-02 -1.73933744e-01 -8.82584751e-01
-4.97023672e-01 9.95917171e-02 1.35311961e-01 -2.03564450e-01
3.46775293e-01 7.70112216e-01 -1.32968760e+00 2.74572998e-01
-1.91631168e-02 1.03664243e+00 3.43988985e-01 1.83314824e+00
3.19261640e-01 5.38342714e-01 -1.00083864e+00 1.01778559e-01
1.69798806e-01 1.72030345e-01 -3.38194482e-02 -3.75548422e-01
7.87188053e-01 -1.99622437e-01 -5.54819822e-01 5.83187401e-01
1.59641933e+00 6.77296221e-01 7.32025623e-01 -5.99154174e-01
2.44587362e-02 5.48408687e-01 -7.44085908e-02 5.27170002e-01
8.09396744e-01 -1.21384956e-01 2.86908925e-01 -1.61010337e+00
2.75341690e-01 1.07866275e+00 1.09129274e+00 1.82080343e-01
-8.48194897e-01 -5.93817174e-01 -5.17085433e-01 -2.02198252e-01
-1.21401572e+00 -1.32370484e+00 6.24151170e-01 -1.46982074e-01
1.09950089e+00 9.54901874e-01 1.04415640e-01 1.29527223e+00
-3.37982744e-01 1.53088287e-01 1.34261775e+00 -7.61952579e-01
2.08213031e-01 3.61543328e-01 6.24856055e-02 6.93287432e-01
-6.68689132e-01 5.11227489e-01 -3.63920063e-01 -4.28340197e-01
7.43762672e-01 -5.86430669e-01 -7.48845518e-01 1.20690987e-02
-1.65702474e+00 9.75099951e-03 8.06156024e-02 3.92147928e-01
-8.16580653e-02 1.72485083e-01 4.04135942e-01 1.16059117e-01
1.58056661e-01 3.40490431e-01 -2.23665923e-01 -3.72072518e-01
-7.03252971e-01 4.81932759e-02 9.04028118e-01 1.27117312e+00
2.61912137e-01 4.94536310e-01 -2.50657529e-01 1.40920353e+00
3.34676355e-01 1.15757382e+00 -4.60010991e-02 -1.23461294e+00
6.77909076e-01 -6.87606812e-01 4.18769568e-01 -1.04941678e+00
-5.58936119e-01 -2.56125659e-01 -8.12823236e-01 -4.38397862e-02
3.27201247e-01 -3.13246101e-01 -7.80043066e-01 9.22528386e-01
3.07011455e-01 6.11425936e-01 5.65078110e-02 8.80691469e-01
1.22386837e+00 8.09493005e-01 -5.76290727e-01 -9.39051360e-02
1.42206430e+00 -9.27505851e-01 -8.18048477e-01 -2.41128057e-01
1.65200979e-01 -1.45666993e+00 6.86134636e-01 6.27043188e-01
-1.13279903e+00 -6.84151471e-01 -8.28150332e-01 2.67840683e-01
-3.41829389e-01 3.58466774e-01 2.55146831e-01 1.21907651e+00
-1.17860579e+00 1.78937733e-01 -4.56927776e-01 2.09753856e-01
-5.13818681e-01 1.11389197e-01 -2.74234295e-01 -7.76745752e-02
-7.55885184e-01 6.36486471e-01 -4.60209817e-01 6.20467305e-01
-1.13411474e+00 -8.95751417e-01 -8.80349696e-01 -9.47787538e-02
1.71299115e-01 -9.81139690e-02 1.49153459e+00 -7.23271444e-02
-1.62838411e+00 6.62159085e-01 -4.93979752e-02 4.47074138e-02
2.65729934e-01 5.43538034e-02 -1.04138923e+00 -4.79334034e-02
-4.76904541e-01 -2.03476250e-01 1.14521289e+00 -1.79325092e+00
-1.84382841e-01 -8.03163052e-02 -4.89134431e-01 1.29783705e-01
2.34638914e-01 4.42854494e-01 -1.98967606e-01 -5.35452187e-01
3.65399450e-01 -4.74956661e-01 -2.98905879e-01 -6.61041558e-01
-6.79563105e-01 6.24820173e-01 3.09827983e-01 -1.18405759e+00
1.18418944e+00 -2.20217323e+00 -1.52399272e-01 6.12136722e-01
4.23277430e-02 1.17455475e-01 2.25339178e-02 4.02045250e-01
-3.61607850e-01 -3.12861264e-01 -1.04087442e-01 -6.59826696e-01
1.42955706e-01 -4.67545003e-01 -3.98978174e-01 6.07868493e-01
-7.67799497e-01 1.89815015e-01 -6.14106357e-01 -4.83966410e-01
7.52125204e-01 7.46437907e-01 -1.38239279e-01 5.74729264e-01
8.13692093e-01 6.41731799e-01 -6.39892444e-02 9.10504937e-01
1.16186261e+00 6.11294031e-01 -1.81246534e-01 -4.19534504e-01
-5.89796185e-01 1.67026907e-01 -1.96309912e+00 1.34999561e+00
-1.30273128e+00 9.89717484e-01 1.04320300e+00 -5.10449588e-01
1.17469513e+00 7.22912371e-01 1.33627489e-01 -6.89747036e-01
9.31529626e-02 5.97476661e-01 -5.87076962e-01 -6.33541048e-01
8.15474808e-01 2.16983944e-01 1.75093785e-01 3.36537391e-01
-8.58817920e-02 -3.47782016e-01 -3.81843328e-01 -2.25549303e-02
1.39806712e+00 -4.23243910e-01 -3.05876005e-02 -1.41020462e-01
8.60537231e-01 -5.30681074e-01 -7.49653503e-02 8.37126732e-01
5.75302467e-02 1.04999316e+00 -5.18981218e-01 1.96582396e-02
-1.12828410e+00 -1.38324285e+00 -2.01349691e-01 1.24892735e+00
-3.09060782e-01 -2.35595405e-01 -6.57640100e-01 1.37990251e-01
-5.13160467e-01 6.57271981e-01 -2.32029006e-01 5.19532144e-01
-1.16905653e+00 -2.20096573e-01 7.61299431e-01 5.83368301e-01
1.03863306e-01 -8.75678420e-01 -4.27342415e-01 2.13110164e-01
-6.08503044e-01 -1.37458897e+00 -7.77937531e-01 5.96000493e-01
-5.67156672e-01 -6.46928430e-01 -1.03030217e+00 -8.39550495e-01
6.38675392e-01 5.27361512e-01 1.37272072e+00 -7.86938071e-02
-7.61873722e-01 1.32115781e+00 -3.60672444e-01 -3.90767902e-01
-5.63408971e-01 -6.64130509e-01 5.52375466e-02 -2.46042937e-01
-5.22563219e-01 -4.87169057e-01 -8.03609252e-01 9.21972215e-01
-2.81342715e-01 -6.06436789e-01 -8.53945836e-02 3.38249385e-01
3.03423077e-01 -3.51059794e-01 -5.21059856e-02 -1.32975206e-01
8.33584309e-01 2.70482264e-02 -8.97345006e-01 4.47693676e-01
4.55375090e-02 -2.86415607e-01 5.98906934e-01 -3.83081555e-01
-1.52791250e+00 -8.15218240e-02 -5.54786503e-01 -7.40571246e-02
-4.99138772e-01 -1.45484984e-01 -5.74924164e-02 -2.05427095e-01
1.01266611e+00 1.41916797e-01 -6.46244347e-01 -7.60785401e-01
1.41101331e-01 1.01174355e+00 8.60298753e-01 -5.53012609e-01
8.47445965e-01 2.93392092e-01 -8.81351009e-02 -1.40442359e+00
-1.65081024e-01 -1.10468280e+00 -5.46768546e-01 -5.00940919e-01
5.68526149e-01 -8.53623331e-01 -9.06236529e-01 3.82014751e-01
-1.26647723e+00 -4.88351762e-01 -4.18713316e-03 1.01067615e+00
-4.94567543e-01 2.04357073e-01 -4.92039889e-01 -1.54912424e+00
-2.32158870e-01 -9.43992794e-01 1.09256339e+00 -3.02647743e-02
1.80623457e-01 -1.10810137e+00 4.77819055e-01 4.28214073e-01
8.31469297e-01 -1.40542641e-01 5.54768026e-01 -2.23870471e-01
-5.03716707e-01 -2.42642477e-01 1.05879121e-01 3.76957059e-01
-2.73167700e-01 -1.02878265e-01 -1.65852928e+00 -1.71191275e-01
3.05875111e-02 -9.86915901e-02 6.73490405e-01 1.04938996e+00
9.31642056e-01 -1.44897789e-01 -3.31774980e-01 8.31087887e-01
1.28579450e+00 2.58978367e-01 8.08891475e-01 -1.89044148e-01
3.45266849e-01 6.73082769e-01 5.18613338e-01 8.54908347e-01
-1.95711479e-01 7.68484354e-01 4.54817772e-01 -2.93522030e-01
-2.17739567e-01 1.73543528e-01 4.04777974e-01 1.17963910e+00
-4.64585036e-01 -6.14049315e-01 -8.72887313e-01 1.38156503e-01
-8.40529859e-01 -1.02106023e+00 -3.16593260e-01 2.60146666e+00
7.12154284e-02 -6.80977464e-01 -6.93918264e-05 2.76631206e-01
8.02172303e-01 3.26812893e-01 1.59624591e-01 -6.74325943e-01
4.47653420e-02 3.18369299e-01 8.32988262e-01 9.75397527e-01
-8.90741467e-01 2.27691367e-01 7.73121071e+00 5.71519494e-01
-7.47301757e-01 1.98698103e-01 4.14681643e-01 2.74058551e-01
-2.81811625e-01 -5.17040074e-01 -9.53675449e-01 -3.50834191e-01
1.26957738e+00 6.04147494e-01 1.06466818e+00 5.41892767e-01
2.92981446e-01 -7.07249045e-02 -1.04527318e+00 1.26813638e+00
2.89423525e-01 -1.19216037e+00 -1.18807113e+00 -1.45032942e-01
3.10705870e-01 1.49147287e-01 2.30928078e-01 6.87363222e-02
-4.99957763e-02 -1.18024480e+00 7.74634719e-01 5.84671378e-01
9.43928242e-01 -4.02682662e-01 2.81115443e-01 4.24086452e-02
-1.61763167e+00 -1.16724543e-01 -4.79326069e-01 8.18765700e-01
3.17794472e-01 3.00002337e-01 -1.27807105e+00 4.40283328e-01
7.91271567e-01 -9.50640365e-02 -3.88048381e-01 1.34182847e+00
4.03589606e-01 6.48177207e-01 -6.19632244e-01 8.79144073e-02
-3.38038318e-02 1.38915284e-03 8.69938552e-01 1.98091114e+00
9.46292162e-01 3.00059319e-01 -2.76251584e-01 5.51151514e-01
2.51913309e-01 1.32949904e-01 -6.14449918e-01 3.19224387e-01
6.26820803e-01 1.64248550e+00 -8.41337562e-01 1.44369647e-01
-6.28462154e-03 6.06355965e-01 -6.94281161e-01 7.62239516e-01
-9.73706722e-01 -7.83752203e-01 1.97184756e-01 2.00234145e-01
2.43101358e-01 -8.25239360e-01 2.88629442e-01 -5.17835855e-01
-1.02280155e-01 -8.93745363e-01 -4.49233986e-02 -1.20541191e+00
-6.94429934e-01 7.98065484e-01 2.08981112e-01 -1.19482052e+00
-1.98788613e-01 -1.10652137e+00 -7.44885623e-01 8.90289605e-01
-1.11854053e+00 -4.91836965e-01 -6.53007984e-01 8.78361642e-01
4.87695158e-01 -2.35108003e-01 1.19340277e+00 7.68259048e-01
-7.50209987e-02 4.98142958e-01 5.37183642e-01 2.35429764e-01
8.34971428e-01 -1.41989803e+00 4.43070471e-01 4.45899814e-01
1.49852335e-01 6.08581424e-01 1.06940913e+00 7.22900033e-02
-1.50471866e+00 -3.92652452e-01 3.55416507e-01 -4.97559011e-01
2.21817076e-01 -8.83579671e-01 -6.72382176e-01 4.04575467e-01
2.07985178e-01 3.81536514e-01 6.13917530e-01 6.50003925e-02
-2.76171267e-01 -3.58641863e-01 -1.01472068e+00 2.71100730e-01
9.25596416e-01 -4.29069340e-01 1.54502511e-01 5.67303658e-01
2.91482598e-01 -9.02765691e-01 -7.95327067e-01 -8.09673592e-02
6.61314607e-01 -1.22439361e+00 1.76894677e+00 3.01530778e-01
-2.81693399e-01 4.40536588e-02 -6.40337467e-01 -1.33691990e+00
9.02358145e-02 -1.12109840e+00 6.97853416e-02 1.26154745e+00
5.44556022e-01 -4.96815264e-01 3.13125193e-01 1.88516915e-01
-7.28252769e-01 4.54757921e-02 -1.08893359e+00 -5.41907430e-01
-5.91908813e-01 -9.21082079e-01 4.60184306e-01 3.97609234e-01
-4.91842717e-01 1.19816579e-01 -5.29664814e-01 6.36681557e-01
8.76527309e-01 -3.07758152e-01 6.65702879e-01 -8.47578168e-01
-4.77561384e-01 -8.33328664e-02 1.69707030e-01 -1.25407279e+00
-1.17399544e-01 -4.72924799e-01 6.69222414e-01 -1.48566723e+00
-6.13955498e-01 -1.08605492e+00 -5.66563532e-02 -3.91935647e-01
4.98909593e-01 1.92946538e-01 -8.93337652e-02 -3.65821719e-01
-3.94720644e-01 4.24889624e-02 1.03486669e+00 -1.06319338e-02
-1.84487492e-01 6.85590208e-01 1.94035828e-01 7.81379521e-01
4.11769211e-01 -5.62320411e-01 -8.41236115e-02 -3.76643270e-01
2.69608378e-01 7.36074388e-01 6.85027778e-01 -1.20640194e+00
4.94416565e-01 2.74679929e-01 5.45754671e-01 -9.72998023e-01
1.08678067e+00 -1.03163052e+00 6.14438988e-02 -8.11023340e-02
-5.40416420e-01 4.15542871e-01 3.45927864e-01 5.58899045e-01
3.91130671e-02 -4.06191885e-01 5.43465316e-01 -3.53749573e-01
-3.29756588e-01 -9.56241116e-02 -6.83036983e-01 2.61207849e-01
2.81957150e-01 -3.87432575e-01 -2.34980345e-01 -7.28455961e-01
-8.39664757e-01 -4.58657712e-01 -2.93719053e-01 1.17063031e-01
1.10555851e+00 -1.09543312e+00 -6.60878301e-01 3.26289982e-01
-7.32695833e-02 -3.53320986e-01 6.38548374e-01 9.97953892e-01
-5.54817259e-01 4.31440920e-01 2.24214181e-01 -5.31415820e-01
-1.94696367e+00 3.27469260e-01 9.61811244e-01 1.62253231e-02
-6.31017983e-01 1.23064744e+00 -7.57178292e-02 -7.82065153e-01
8.00338924e-01 -4.61481571e-01 -2.96677291e-01 -2.00576052e-01
6.59860015e-01 9.51222539e-01 3.57297033e-01 -6.10758185e-01
-5.16008198e-01 8.12369823e-01 7.24620759e-01 -7.91204393e-01
1.00149441e+00 -5.23587227e-01 1.33677468e-01 6.94005907e-01
1.28701961e+00 9.35598254e-01 -1.00785828e+00 -1.99991494e-01
-3.25989097e-01 -7.94315755e-01 2.66619533e-01 -1.20314705e+00
-7.27061927e-01 1.21593010e+00 1.26591015e+00 4.53459881e-02
1.13312662e+00 -1.56613797e-01 4.44339156e-01 5.66341817e-01
3.38995039e-01 -1.16528022e+00 4.08192426e-01 4.70570415e-01
1.14223409e+00 -1.03221488e+00 -1.58901438e-01 -3.59426469e-01
-5.51333964e-01 1.23011720e+00 -1.00836523e-01 3.90268028e-01
9.98135746e-01 9.67843294e-01 5.06695926e-01 1.53123029e-02
-1.95893392e-01 2.52618939e-01 3.08834851e-01 1.17694020e+00
6.80199742e-01 2.78491024e-02 9.17324781e-01 1.68293029e-01
-1.35756135e-01 -9.12988245e-01 3.92557710e-01 5.79026699e-01
-4.91280586e-01 -7.87172318e-01 -1.29272413e+00 -2.87776709e-01
-4.20835614e-01 -4.81860310e-01 5.87752052e-02 3.35940897e-01
-2.52575517e-01 1.23602283e+00 -2.62215793e-01 -2.79530913e-01
7.64410853e-01 -1.26559496e-01 7.05417812e-01 -2.05309942e-01
-1.14705443e+00 5.76609313e-01 4.41706419e-01 -2.21246868e-01
-1.71778113e-01 -5.16785562e-01 -1.10108733e+00 -2.38744617e-01
-6.27306759e-01 3.30784529e-01 1.46618044e+00 2.23382965e-01
4.26131822e-02 9.15941000e-01 9.68677700e-01 -1.31946588e+00
-7.70132601e-01 -9.89263952e-01 -8.76764655e-01 -2.93609589e-01
6.91994250e-01 -2.21149296e-01 -9.03748989e-01 -8.57037231e-02] | [15.084566116333008, 5.811712741851807] |
67b72918-8041-405c-9513-5887c1b6a5bf | roft-real-time-optical-flow-aided-6d-object | 2111.03821 | null | https://arxiv.org/abs/2111.03821v1 | https://arxiv.org/pdf/2111.03821v1.pdf | ROFT: Real-Time Optical Flow-Aided 6D Object Pose and Velocity Tracking | 6D object pose tracking has been extensively studied in the robotics and computer vision communities. The most promising solutions, leveraging on deep neural networks and/or filtering and optimization, exhibit notable performance on standard benchmarks. However, to our best knowledge, these have not been tested thoroughly against fast object motions. Tracking performance in this scenario degrades significantly, especially for methods that do not achieve real-time performance and introduce non negligible delays. In this work, we introduce ROFT, a Kalman filtering approach for 6D object pose and velocity tracking from a stream of RGB-D images. By leveraging real-time optical flow, ROFT synchronizes delayed outputs of low frame rate Convolutional Neural Networks for instance segmentation and 6D object pose estimation with the RGB-D input stream to achieve fast and precise 6D object pose and velocity tracking. We test our method on a newly introduced photorealistic dataset, Fast-YCB, which comprises fast moving objects from the YCB model set, and on the dataset for object and hand pose estimation HO-3D. Results demonstrate that our approach outperforms state-of-the-art methods for 6D object pose tracking, while also providing 6D object velocity tracking. A video showing the experiments is provided as supplementary material. | ['Lorenzo Natale', 'Ugo Pattacini', 'Giulia Pasquale', 'Yuriy Onyshchuk', 'Nicola A. Piga'] | 2021-11-06 | null | null | null | null | ['6d-pose-estimation'] | ['computer-vision'] | [-1.50736764e-01 -5.92705548e-01 -1.42753720e-01 3.98092344e-02
-2.95397848e-01 -7.71443665e-01 4.33239222e-01 -1.50631607e-01
-8.14910412e-01 1.95870027e-01 -4.05839801e-01 -1.85318142e-02
-6.60075247e-02 -2.11005956e-01 -9.15664852e-01 -6.26734018e-01
-3.47047448e-01 6.60049558e-01 5.98229110e-01 9.69216675e-02
2.26099208e-01 1.14568925e+00 -1.57690334e+00 -1.56400427e-01
2.87194669e-01 1.24110770e+00 1.35115415e-01 1.20104110e+00
1.75354183e-01 5.36656439e-01 -5.94692409e-01 -1.43343121e-01
5.09164989e-01 1.89348266e-01 -6.72496557e-01 2.10331202e-01
1.05827582e+00 -8.05114388e-01 -6.64249063e-01 8.91500413e-01
6.51332438e-01 2.13338330e-01 3.70447695e-01 -1.52888215e+00
-3.09507668e-01 2.09019423e-01 -5.77225685e-01 2.60962307e-01
4.65766549e-01 6.98668361e-01 4.90514219e-01 -7.86493599e-01
9.51225936e-01 1.53676236e+00 7.65693963e-01 8.61535132e-01
-1.08138800e+00 -6.54853106e-01 2.82986850e-01 1.33175641e-01
-1.02441549e+00 -2.66370565e-01 6.24199212e-01 -6.35056973e-01
1.10630858e+00 2.59937122e-02 9.73858654e-01 1.16655600e+00
1.32281572e-01 1.06128085e+00 6.22632325e-01 -8.48122686e-02
2.70003011e-03 -3.68833244e-01 7.53161823e-03 6.17091358e-01
3.36802453e-01 4.81742084e-01 -6.11869276e-01 2.10790887e-01
9.89804983e-01 1.37299448e-01 -3.68542969e-01 -8.02835166e-01
-1.59054053e+00 2.29302168e-01 6.95191264e-01 -2.86231581e-02
-3.53803307e-01 8.28687787e-01 3.62806767e-01 7.28882849e-02
2.25859389e-01 9.41505283e-02 -5.31744301e-01 -3.73290956e-01
-6.05643213e-01 5.18928647e-01 6.95129931e-01 1.30799139e+00
1.92362398e-01 7.78982267e-02 -3.87845457e-01 6.37554973e-02
7.00368345e-01 9.78233874e-01 5.96332997e-02 -1.25662136e+00
4.10186648e-01 3.75025541e-01 5.74214041e-01 -9.70604479e-01
-5.93445122e-01 -4.33266908e-01 -3.43777448e-01 5.27737260e-01
9.94769096e-01 1.36416312e-02 -9.66805756e-01 1.42204952e+00
7.75372624e-01 5.71066916e-01 -9.06062499e-02 1.52663231e+00
8.04771841e-01 4.00030166e-01 2.64389962e-02 -5.95623106e-02
1.20885861e+00 -8.16890538e-01 -5.32736480e-01 5.01624420e-02
4.42854732e-01 -8.79548192e-01 6.33536637e-01 4.87580657e-01
-1.16323018e+00 -7.70953834e-01 -8.15674782e-01 4.36509401e-03
-2.07120836e-01 1.25783727e-01 6.09941483e-01 6.57219708e-01
-8.57426047e-01 7.06113577e-01 -1.39110291e+00 -4.04988080e-01
5.81736326e-01 5.74528158e-01 -2.50848740e-01 -2.75948457e-02
-7.40439534e-01 9.29529011e-01 2.24160746e-01 6.35886610e-01
-1.15563810e+00 -9.74982202e-01 -6.15525126e-01 -2.76408613e-01
6.37473583e-01 -6.69677079e-01 1.57172656e+00 -3.20898950e-01
-1.72357655e+00 7.45858967e-01 1.80731133e-01 -6.50688052e-01
1.01172960e+00 -7.79884160e-01 5.54995872e-02 1.75223172e-01
-3.20321739e-01 7.84576714e-01 1.10573494e+00 -1.17360926e+00
-7.41019428e-01 -7.15862691e-01 6.61299378e-02 1.05249658e-02
-1.01225384e-01 -6.28925562e-02 -8.20353925e-01 -3.44964594e-01
1.75631061e-01 -1.18984199e+00 -1.53508157e-01 7.32990801e-01
-2.53367215e-01 -3.16832095e-01 1.37990022e+00 -2.92046607e-01
5.92156112e-01 -1.97828233e+00 2.36498505e-01 -2.26454794e-01
2.57275611e-01 5.84247231e-01 -1.27260998e-01 -3.08028683e-02
2.03267500e-01 -4.41933632e-01 2.58114576e-01 -5.51404774e-01
1.84227645e-01 -1.83997862e-02 -1.46476418e-01 1.02818573e+00
1.86971903e-01 1.17064369e+00 -1.01865757e+00 -3.36283624e-01
5.93243301e-01 9.34970558e-01 -5.10751247e-01 1.61553785e-01
-5.14763474e-01 6.29720867e-01 -3.42299521e-01 7.92865157e-01
8.56877446e-01 -1.01703338e-01 -2.22643390e-01 -4.64705437e-01
-2.93786913e-01 -1.22450501e-01 -1.40716076e+00 1.98820078e+00
-1.75646067e-01 8.10631633e-01 5.61003899e-03 -5.62407136e-01
6.09493732e-01 2.07244545e-01 8.21194172e-01 -3.31984490e-01
5.46449244e-01 1.89838976e-01 2.52567008e-02 -5.14418125e-01
4.72412705e-01 4.57235515e-01 2.10079283e-01 2.82592416e-01
9.18110982e-02 -3.22525322e-01 1.29337460e-01 5.49495667e-02
1.01173913e+00 6.27971888e-01 -1.89149126e-01 2.24749222e-01
4.85098511e-01 2.60253847e-01 2.43927807e-01 8.28130245e-01
-6.91398799e-01 5.01102388e-01 1.11594260e-01 -3.75841886e-01
-1.12166905e+00 -1.00942075e+00 -3.68632078e-02 6.25490546e-01
5.50049603e-01 -7.18495846e-02 -5.25586188e-01 -5.98983645e-01
6.00417972e-01 8.97619203e-02 -4.96732622e-01 9.05239955e-02
-9.96663153e-01 -2.20386893e-01 3.91885459e-01 7.99651980e-01
3.22317183e-01 -9.69323099e-01 -1.37104428e+00 5.04370451e-01
3.03636998e-01 -1.60455370e+00 -3.84798437e-01 9.42546651e-02
-9.66800869e-01 -1.16943645e+00 -9.86495972e-01 -5.15425682e-01
2.25896463e-01 4.83720541e-01 7.27762938e-01 1.65076740e-02
-6.27580285e-01 8.19701791e-01 -2.45949432e-01 -3.35065573e-01
-2.16963395e-01 -9.81319249e-02 2.11658299e-01 -2.48376817e-01
2.02706218e-01 -1.28096789e-01 -8.57075989e-01 4.34135497e-01
-5.28059125e-01 -2.85801351e-01 2.27889702e-01 5.23493469e-01
5.05225837e-01 -3.26517314e-01 -1.08249016e-01 -1.83456466e-01
-7.20143020e-02 2.09609210e-01 -1.25336981e+00 -1.75795220e-02
-2.47898340e-01 -1.81436971e-01 1.58418864e-01 -8.76699150e-01
-7.85844207e-01 6.23349667e-01 -4.57168184e-02 -1.13805914e+00
-3.03815991e-01 -2.18147665e-01 1.52764007e-01 -4.19628650e-01
4.99245107e-01 -8.68795961e-02 8.17352682e-02 -4.80774879e-01
5.67433536e-01 5.09168029e-01 9.12813485e-01 -3.24376643e-01
1.01959383e+00 7.20842540e-01 1.41207263e-01 -6.23321712e-01
-6.30164564e-01 -7.21691966e-01 -8.57428491e-01 -8.26600432e-01
1.01057792e+00 -8.21594715e-01 -1.59880340e+00 9.16702151e-01
-1.53357458e+00 -3.81537735e-01 -2.38330901e-01 8.53243589e-01
-8.87541652e-01 1.23170666e-01 -6.95891321e-01 -9.58648860e-01
-2.71234214e-01 -1.36669135e+00 1.52917826e+00 2.89534599e-01
1.51700169e-01 -5.85564077e-01 -2.56525367e-01 -2.22867914e-02
3.05502534e-01 4.20867860e-01 9.03166011e-02 -2.46861149e-02
-1.11976850e+00 -2.31461629e-01 -2.79320121e-01 -6.94560036e-02
-6.30608052e-02 2.25472867e-01 -9.53673482e-01 -5.56190908e-01
-2.07182929e-01 -1.76451042e-01 7.32630193e-01 6.54571652e-01
8.92957985e-01 2.29767650e-01 -5.40546834e-01 8.60685766e-01
1.35469770e+00 1.82530716e-01 6.70613721e-02 4.26136643e-01
1.02072048e+00 3.52470160e-01 9.25846457e-01 5.37212193e-01
1.07256509e-01 1.02472520e+00 9.35315728e-01 2.51603335e-01
-4.57037896e-01 1.24893628e-01 2.48944819e-01 3.22190642e-01
-2.93862283e-01 -2.92894155e-01 -8.55595350e-01 2.21391395e-01
-1.89884090e+00 -7.57776201e-01 -4.80817169e-01 1.95303965e+00
3.60911250e-01 2.78308332e-01 3.45425516e-01 2.59660184e-01
6.29942417e-01 1.46317985e-02 -9.03641045e-01 2.92987227e-01
1.79519862e-01 -6.46937862e-02 7.28357613e-01 4.02649224e-01
-1.23252642e+00 1.19378054e+00 5.05100965e+00 1.70667320e-01
-1.22773659e+00 9.51326713e-02 -2.14473933e-01 -3.11315656e-01
4.11291718e-01 -2.45613024e-01 -1.15946543e+00 1.55324355e-01
5.25341272e-01 2.30393291e-01 2.55852222e-01 9.00000751e-01
1.46256849e-01 6.16653226e-02 -1.19632208e+00 1.14317667e+00
-2.54555531e-02 -1.23837364e+00 -4.55669850e-01 -7.70427361e-02
4.52263772e-01 1.34254187e-01 1.66900635e-01 8.42626672e-03
1.47869438e-01 -6.18018806e-01 1.17404556e+00 4.49519187e-01
4.13171649e-01 -6.14910066e-01 4.38549072e-01 2.39717707e-01
-1.36698055e+00 -5.52258007e-02 -2.66725242e-01 2.09746584e-01
4.28470880e-01 1.43251985e-01 -7.47957051e-01 3.47749203e-01
1.07306504e+00 1.03418982e+00 -4.81869310e-01 1.51088476e+00
-9.38309431e-02 2.21301973e-01 -5.10502875e-01 -2.33894587e-01
3.74063969e-01 3.66602898e-01 8.93899798e-01 1.01608169e+00
2.44916514e-01 3.30375619e-02 1.72671705e-01 7.82172799e-01
2.22907990e-01 -4.47524220e-01 -3.76658976e-01 3.73166166e-02
2.66073704e-01 1.07345390e+00 -9.81853545e-01 -1.89488679e-01
-2.12450340e-01 9.60522354e-01 8.15674756e-03 3.07566315e-01
-1.02517951e+00 -1.40296876e-01 8.22759867e-01 -1.13648348e-01
6.77290916e-01 -8.15882862e-01 1.08694799e-01 -1.10898042e+00
-9.12428126e-02 -5.17660201e-01 1.05458595e-01 -9.28688943e-01
-9.88947868e-01 4.96682703e-01 -5.61993271e-02 -1.39065766e+00
-1.28468826e-01 -1.13376248e+00 1.81334421e-01 5.23935676e-01
-1.45990980e+00 -9.92185056e-01 -7.26183534e-01 6.96665108e-01
5.54868281e-01 1.36657506e-01 3.39875937e-01 3.84679615e-01
-3.11003864e-01 3.95712495e-01 -1.41368747e-01 1.59120694e-01
6.69654310e-01 -1.14986968e+00 7.14641750e-01 8.06902170e-01
1.93921313e-01 2.85568416e-01 8.37399721e-01 -6.38827443e-01
-2.27984476e+00 -9.78735685e-01 2.43318379e-01 -8.45215082e-01
5.97704291e-01 -5.49705207e-01 -7.85862982e-01 6.83912694e-01
-1.97509170e-01 6.18555307e-01 -1.77677631e-01 -6.13618731e-01
-2.17890248e-01 -2.81445384e-02 -9.55360293e-01 4.64280903e-01
1.53737187e+00 -1.42009646e-01 -3.57616514e-01 3.93582255e-01
1.06716776e+00 -1.24908113e+00 -9.32491899e-01 4.63740915e-01
9.66204822e-01 -7.79454291e-01 1.31331909e+00 -5.24884582e-01
-9.75935459e-02 -6.76096082e-01 -9.39320847e-02 -8.49123240e-01
9.39424615e-03 -7.07291484e-01 -6.98295772e-01 7.97108293e-01
-2.90293545e-01 -2.48249471e-01 1.18747413e+00 3.14877152e-01
3.04491874e-02 -3.68550926e-01 -9.98833835e-01 -1.02643836e+00
-3.54295313e-01 -8.78802776e-01 4.63948280e-01 3.65361154e-01
-7.96936214e-01 -1.72171161e-01 -2.68198431e-01 4.66907948e-01
1.04654777e+00 3.85416687e-01 1.21079636e+00 -1.15129364e+00
-7.86996782e-02 -4.80942637e-01 -8.21778238e-01 -1.47698355e+00
6.60905987e-02 -6.25147402e-01 3.12899381e-01 -1.18523300e+00
-2.48751238e-01 -4.48113024e-01 1.51552245e-01 1.63099304e-01
9.33436677e-03 5.11678517e-01 7.27604210e-01 2.11621583e-01
-5.67120671e-01 3.62282127e-01 1.68252301e+00 -2.81751782e-01
-3.33248585e-01 1.98868632e-01 2.79017001e-01 5.13775229e-01
3.54900956e-01 -5.27301669e-01 -6.85589155e-03 -5.91506541e-01
-1.88716978e-01 2.41852626e-01 9.35542405e-01 -1.22137797e+00
5.34059048e-01 2.23338865e-02 6.55901611e-01 -1.20086837e+00
6.03862226e-01 -1.23828876e+00 1.98260367e-01 9.29860592e-01
-1.86712369e-01 1.57876804e-01 3.32552046e-01 6.67761385e-01
1.15016714e-01 4.59659547e-02 7.74992287e-01 3.82553227e-02
-1.05903590e+00 8.53951514e-01 -3.26458924e-02 -1.95984691e-01
1.19090581e+00 -3.49878371e-01 -2.99494177e-01 -5.02392044e-03
-7.75560737e-01 1.91018835e-01 3.06737959e-01 7.25190878e-01
6.14005864e-01 -1.25564253e+00 -5.56573927e-01 7.66925588e-02
-4.22814786e-02 3.79196942e-01 1.61298603e-01 1.00791705e+00
-7.80932546e-01 4.70510274e-01 -2.06138343e-01 -1.51147974e+00
-1.24835300e+00 6.93310678e-01 5.12902617e-01 2.23706439e-01
-9.22545075e-01 9.55922306e-01 -3.16475570e-01 -2.14786589e-01
7.70271122e-01 -7.59422719e-01 1.14728302e-01 3.50106470e-02
3.77207249e-01 5.31963587e-01 5.15934266e-02 -7.37654507e-01
-6.88129067e-01 1.07734859e+00 1.98310971e-01 -1.31755611e-02
1.31093252e+00 -7.52260536e-02 2.89895415e-01 2.74794966e-01
1.15173495e+00 -5.56155920e-01 -1.96627676e+00 -8.56197923e-02
1.26417084e-02 -8.51878285e-01 -2.35070940e-02 -5.33844411e-01
-1.37667799e+00 1.04395711e+00 1.06912410e+00 -1.87231395e-02
7.58868694e-01 7.60572851e-02 7.88436651e-01 4.99273777e-01
5.13055623e-01 -6.50986433e-01 4.00708497e-01 6.50082231e-01
7.25724459e-01 -1.23303103e+00 -8.09117705e-02 -4.26900595e-01
-1.31366387e-01 1.37836623e+00 8.95234108e-01 -3.56708020e-01
4.79959100e-01 5.95217466e-01 1.97882950e-01 4.72067203e-03
-5.41020811e-01 -3.90697807e-01 2.65333325e-01 6.87116861e-01
-1.47106498e-02 -3.01701039e-01 2.20021680e-01 -1.61140680e-01
8.78325254e-02 1.30069137e-01 1.46996707e-01 1.01899803e+00
-1.58471733e-01 -7.61364639e-01 -6.59304380e-01 -1.42515779e-01
-2.61946380e-01 5.65305829e-01 -2.02377904e-02 1.15176678e+00
4.68254089e-02 5.44189334e-01 2.62196273e-01 -1.87473506e-01
5.74956834e-01 -3.82958353e-01 1.07455587e+00 -1.93823427e-01
-6.89449787e-01 2.55828500e-01 -3.02230865e-01 -1.03503561e+00
-8.42477918e-01 -8.36515188e-01 -1.40124416e+00 -2.77156174e-01
-5.70553064e-01 -4.60478425e-01 9.46305752e-01 8.36357594e-01
5.21512106e-02 5.74529350e-01 1.24106191e-01 -1.59959328e+00
-6.60344899e-01 -6.05453134e-01 -3.68994594e-01 2.49490395e-01
7.62560725e-01 -9.93397772e-01 -2.03208998e-01 4.10861103e-03] | [6.8664774894714355, -2.2472951412200928] |
f26765ed-0e06-477e-983a-2df6758ce4e2 | unsupervised-doppler-radar-based-activity | 2103.10478 | null | https://arxiv.org/abs/2103.10478v2 | https://arxiv.org/pdf/2103.10478v2.pdf | Unsupervised Doppler Radar-Based Activity Recognition for e-Healthcare | Passive radio frequency (RF) sensing and monitoring of human daily activities in elderly care homes is an emerging topic. Micro-Doppler radars are an appealing solution considering their non-intrusiveness, deep penetration, and high-distance range. Unsupervised activity recognition using Doppler radar data has not received attention, in spite of its importance in case of unlabelled or poorly labelled activities in real scenarios. This study proposes two unsupervised feature extraction methods for the purpose of human activity monitoring using Doppler-streams. These include a local Discrete Cosine Transform (DCT)-based feature extraction method and a local entropy-based feature extraction method. In addition, a novel application of Convolutional Variational Autoencoder (CVAE) feature extraction is employed for the first time for Doppler radar data. The three feature extraction architectures are compared with the previously used Convolutional Autoencoder (CAE) and linear feature extraction based on Principal Component Analysis (PCA) and 2DPCA. Unsupervised clustering is performed using K-Means and K-Medoids. The results show the superiority of DCT-based method, entropy-based method, and CVAE features compared to CAE, PCA, and 2DPCA, with more than 5\%-20\% average accuracy. In regards to computation time, the two proposed methods are noticeably much faster than the existing CVAE. Furthermore, for high-dimensional data visualisation, three manifold learning techniques are considered. The methods are compared for the projection of raw data as well as the encoded CVAE features. All three methods show an improved visualisation ability when applied to the encoded CVAE features. | ['Bo Tan', 'Yanguo Jing', 'Wenda Li', 'Sara Sharifzadeh', 'Yordanka Karayaneva'] | 2021-03-18 | null | null | null | null | ['texture-classification'] | ['computer-vision'] | [ 2.58925170e-01 -8.83253291e-03 4.08052295e-01 -5.40083162e-02
-4.59377408e-01 1.12224254e-04 7.33098865e-01 -1.30520388e-02
-8.16275835e-01 9.38117921e-01 4.96018797e-01 -1.09411605e-01
-8.71218920e-01 -8.34035158e-01 2.28650421e-01 -1.13029969e+00
-5.46355724e-01 3.30474675e-01 -3.92553955e-02 -1.11416623e-01
-1.11627258e-01 7.73554742e-01 -1.69766200e+00 -1.17988989e-01
6.40399933e-01 9.86945510e-01 -4.78292145e-02 8.26884747e-01
2.66841114e-01 6.41274691e-01 -6.43798947e-01 2.12721258e-01
1.26826763e-01 -1.66094691e-01 -3.77279669e-01 -2.47507125e-01
-1.71281159e-01 -3.60390365e-01 -3.91923785e-01 4.66801673e-01
7.04583287e-01 4.61508811e-01 1.39999640e+00 -9.64174807e-01
-1.93403736e-01 -2.38138616e-01 -3.72091830e-01 5.49466491e-01
5.58606803e-01 -4.27532256e-01 4.85577047e-01 -5.45365036e-01
1.99687704e-01 7.44306862e-01 6.41508758e-01 2.28133112e-01
-8.72397304e-01 -3.42806548e-01 -5.64865112e-01 5.89179158e-01
-1.53337109e+00 -2.79911011e-01 7.88586378e-01 -6.09081864e-01
1.29707360e+00 3.39931041e-01 8.85192037e-01 1.02540052e+00
6.16910815e-01 2.70915538e-01 1.16070354e+00 -5.27365983e-01
6.38096273e-01 3.25010419e-02 3.46139073e-02 4.82482165e-01
5.42887270e-01 4.80121404e-01 -1.54455960e-01 -4.66692537e-01
5.46153784e-01 2.20665991e-01 -3.39616895e-01 -3.84290665e-01
-1.26810408e+00 9.27379847e-01 6.08319715e-02 8.71515095e-01
-8.89912128e-01 -1.60487667e-02 2.38900036e-01 5.05720712e-02
2.71220654e-01 2.26783723e-01 -4.30357546e-01 -6.28523529e-01
-1.18932652e+00 2.22905934e-01 8.80496562e-01 3.86302978e-01
4.05269653e-01 2.34808236e-01 1.31392017e-01 4.12569642e-01
6.51940823e-01 8.94213557e-01 7.45236039e-01 -6.72290146e-01
2.11951472e-02 3.99230957e-01 -1.36340931e-01 -1.14119911e+00
-8.34798872e-01 -4.00887251e-01 -1.03960466e+00 4.28406894e-01
1.37340056e-03 -4.39970016e-01 -6.56784296e-01 1.07275832e+00
3.12796414e-01 4.41419296e-02 4.47260082e-01 5.88939965e-01
6.24689102e-01 4.88457531e-01 -1.51311625e-02 -6.01371169e-01
1.41148627e+00 -1.94275960e-01 -1.04508245e+00 1.36633068e-01
5.83972692e-01 -4.19256777e-01 -1.15563069e-02 5.11323512e-01
-2.54130721e-01 -5.13672292e-01 -1.21144116e+00 5.84215939e-01
-5.17923057e-01 9.82019082e-02 4.33645070e-01 1.29651177e+00
-8.40301812e-01 4.06045377e-01 -1.06857884e+00 -4.61531639e-01
4.94918823e-01 4.22145218e-01 -5.55381060e-01 1.81782201e-01
-1.09732985e+00 1.01961029e+00 2.77725786e-01 9.04292986e-02
-1.46570504e-01 -3.82309139e-01 -8.80800605e-01 -6.67518377e-02
-4.32940096e-01 -4.66499627e-01 6.56278789e-01 -3.62313181e-01
-1.41442740e+00 1.40858203e-01 7.65488073e-02 -6.76032901e-01
3.14403057e-01 -3.69486541e-01 -9.70112681e-01 5.28496444e-01
-1.27277225e-01 -8.65384191e-03 1.08523905e+00 -6.55571163e-01
-5.19423187e-01 -5.51044703e-01 -4.80456561e-01 8.31101909e-02
-4.04304922e-01 -3.59337419e-01 6.37610376e-01 -5.84056199e-01
1.82445347e-01 -5.80238938e-01 -5.50843365e-02 -4.26020741e-01
1.91296339e-01 1.42497672e-02 1.11507666e+00 -9.25548494e-01
1.29765773e+00 -2.07450914e+00 -1.25754938e-01 5.81539929e-01
3.02335590e-01 2.71280527e-01 6.33607686e-01 5.30138195e-01
-4.63926010e-02 -5.31057179e-01 -6.02889895e-01 2.56595552e-01
-4.57324497e-02 2.35891774e-01 2.72610933e-01 7.78854430e-01
7.69668370e-02 4.96765852e-01 -6.18882000e-01 -6.48410738e-01
6.33409381e-01 1.03450561e+00 -2.80022889e-01 -2.30449393e-01
6.13324940e-01 3.39375615e-01 -4.67025578e-01 4.83606130e-01
6.51546299e-01 3.39620113e-01 1.46176904e-01 -4.09407228e-01
-7.52956644e-02 -5.04578471e-01 -1.38392687e+00 1.11162651e+00
-2.43791401e-01 8.04149449e-01 -3.67046684e-01 -1.40284503e+00
1.22466457e+00 6.76432610e-01 1.07400167e+00 -7.12512314e-01
3.93310875e-01 8.76131374e-03 5.39968163e-02 -8.25940251e-01
2.37351999e-01 -3.06224465e-01 1.39695769e-02 4.84531432e-01
1.73532709e-01 5.57416856e-01 -5.15184477e-02 -1.00006886e-01
1.33665633e+00 8.52295086e-02 8.24494898e-01 -2.92562723e-01
7.32661128e-01 -3.51860940e-01 1.33506864e-01 2.45183572e-01
-4.47319895e-01 3.41745406e-01 -1.89184338e-01 -2.56934911e-01
-5.71931422e-01 -1.04965246e+00 -3.92261952e-01 4.77055818e-01
-2.96293944e-01 -1.65866524e-01 -5.57868242e-01 -4.46786255e-01
-3.06598824e-02 5.37164807e-01 -6.49508178e-01 -8.18581283e-02
-5.61052084e-01 -7.75498509e-01 6.39062464e-01 6.95310295e-01
6.68100476e-01 -1.01859188e+00 -1.47816551e+00 3.70971560e-01
-4.11157832e-02 -9.09787953e-01 4.75827426e-01 2.78571635e-01
-1.19325745e+00 -1.04174101e+00 -1.09064174e+00 -5.71674883e-01
2.90844560e-01 5.62514625e-02 5.40911496e-01 -4.35259789e-01
-5.99268734e-01 1.00776160e+00 -6.42690122e-01 -4.88456547e-01
-1.25783850e-02 -8.07592049e-02 3.70457083e-01 7.13183805e-02
1.03294504e+00 -9.85273838e-01 -7.46911108e-01 -5.14746793e-02
-5.76858699e-01 -9.06588376e-01 9.56316054e-01 5.88788986e-01
2.69302547e-01 3.65728766e-01 5.87608933e-01 -3.24296594e-01
7.92539775e-01 -4.54033047e-01 -3.86630520e-02 -4.55181748e-02
-6.75579309e-01 1.30816519e-01 3.41547430e-01 -1.64540455e-01
-1.02263260e+00 9.44348499e-02 -1.01552717e-01 -1.71479627e-01
-6.42939806e-01 5.21060705e-01 1.01396263e-01 6.51515322e-03
1.12060797e+00 4.47519064e-01 2.00617909e-01 -3.84115547e-01
8.17249492e-02 9.48917747e-01 4.09152001e-01 3.93052734e-02
8.55791569e-01 7.14635909e-01 -1.19164428e-02 -1.54168546e+00
1.95378885e-01 -8.04725766e-01 -8.69097054e-01 -3.71329755e-01
1.21532154e+00 -7.54862130e-01 -6.48594141e-01 2.43784785e-01
-6.89708352e-01 2.99404770e-01 -3.29584092e-01 1.16445184e+00
-7.32936263e-01 7.57398307e-01 1.37723386e-01 -1.37658906e+00
-6.34540498e-01 -6.20100379e-01 7.25440383e-01 1.84682101e-01
-5.04555762e-01 -1.07292414e+00 4.43011075e-01 2.03724310e-01
5.19992292e-01 7.45683074e-01 6.19721472e-01 -6.67707145e-01
1.57809988e-01 -6.66005969e-01 2.48105973e-01 2.99443841e-01
3.78067762e-01 -3.80691081e-01 -9.90731537e-01 -3.55175108e-01
2.74286449e-01 3.20991933e-01 6.07211113e-01 7.89842427e-01
3.67709786e-01 -9.34374481e-02 -5.04645467e-01 2.92101711e-01
1.34462702e+00 6.35251164e-01 7.58481145e-01 5.60660481e-01
2.19908431e-01 3.30364853e-01 5.60460806e-01 8.48814189e-01
8.20704624e-02 4.33411121e-01 2.71281540e-01 2.05674261e-01
2.77197182e-01 4.50022817e-01 4.52932090e-01 7.88996041e-01
-8.30455124e-01 2.54353911e-01 -6.95753872e-01 4.50867921e-01
-1.61095202e+00 -1.40896857e+00 -1.14454091e-01 2.06740713e+00
-1.00116551e-01 5.46931438e-02 4.35544789e-01 1.05598760e+00
4.25273120e-01 2.39462808e-01 -1.80680621e-02 -4.47058737e-01
1.25693440e-01 8.59053552e-01 5.40737867e-01 2.48427391e-01
-1.35199034e+00 -8.93594101e-02 5.27851963e+00 4.25125450e-01
-7.12449729e-01 1.51093528e-01 -2.84734368e-01 1.04185209e-01
1.30985677e-01 -4.25189674e-01 -2.66486883e-01 5.23048222e-01
1.15676236e+00 1.22257307e-01 6.08255863e-02 9.36710835e-01
1.88581958e-01 -4.08733964e-01 -4.71066922e-01 1.24004519e+00
1.56732097e-01 -8.28053713e-01 -3.57614309e-01 5.17164111e-01
3.04687172e-01 -1.99450970e-01 -3.14613849e-01 2.97344029e-01
-3.16579044e-01 -8.85332584e-01 2.36147195e-01 8.78290415e-01
5.37609518e-01 -1.11125398e+00 1.04359496e+00 1.81159422e-01
-1.44070518e+00 -3.37143630e-01 -2.46639848e-01 -1.98020920e-01
4.08644766e-01 6.50079310e-01 -6.22182548e-01 7.13378608e-01
8.70371938e-01 3.08895886e-01 -2.83882529e-01 1.00799716e+00
7.69854337e-02 5.52518845e-01 -5.74476719e-01 -3.36467892e-01
7.51672462e-02 -2.67937809e-01 5.24810433e-01 1.29557121e+00
6.50920868e-01 2.68821299e-01 -4.44712192e-01 2.64453381e-01
8.46988857e-01 3.27503681e-02 -7.62511373e-01 -7.36930221e-02
2.85541534e-01 1.16926205e+00 -6.12210751e-01 -8.10504705e-02
-4.20896262e-01 8.50058615e-01 -5.43218195e-01 2.59590447e-01
-6.23822629e-01 -9.33035433e-01 5.31767368e-01 2.76275486e-01
5.49813986e-01 -6.06204748e-01 2.69272495e-02 -6.80745006e-01
-5.21476381e-03 -4.23195034e-01 5.20987213e-01 -3.71557862e-01
-9.17689145e-01 6.25390112e-01 2.54324317e-01 -1.60536051e+00
-7.26704657e-01 -7.71928370e-01 -4.45863962e-01 6.57403409e-01
-1.22259152e+00 -7.12417424e-01 -5.90819299e-01 1.01979506e+00
1.19641297e-01 -6.59452260e-01 1.37886071e+00 3.74339998e-01
2.31340248e-02 3.12648833e-01 5.12124658e-01 9.97225121e-02
2.21936569e-01 -1.16950905e+00 -1.16341718e-01 5.80385745e-01
4.29212786e-02 3.02054375e-01 6.06089354e-01 -6.38481140e-01
-1.18568158e+00 -7.52157211e-01 8.67219925e-01 -1.90970182e-01
3.45950395e-01 5.28317504e-02 -5.33653021e-01 2.85848767e-01
-8.96019936e-02 -4.23508175e-02 1.23346364e+00 -1.91662341e-01
2.61564732e-01 -8.54081586e-02 -1.42415869e+00 5.71521968e-02
7.35544682e-01 -4.30845261e-01 -1.01764965e+00 2.54653133e-02
-7.01989979e-02 2.32446328e-01 -1.29303575e+00 3.73173743e-01
1.07854104e+00 -1.05938244e+00 1.09737813e+00 -1.27905324e-01
-1.12125292e-01 -3.64590377e-01 -3.68740261e-01 -9.67587233e-01
-6.00210786e-01 -2.09704757e-01 -5.58083415e-01 6.13301456e-01
8.36714134e-02 -8.88525724e-01 8.92110765e-01 4.29159515e-02
2.94178694e-01 -4.87340510e-01 -1.28750253e+00 -7.70488501e-01
-4.28448021e-01 -5.70551813e-01 2.86523730e-01 7.30529428e-01
2.13661715e-01 6.46120682e-02 -2.81777054e-01 6.09561503e-02
8.05579305e-01 -4.71698165e-01 5.17267048e-01 -1.74870467e+00
-3.16962570e-01 -1.83051229e-02 -1.32488036e+00 -3.27105403e-01
-3.79812807e-01 -6.29230678e-01 -2.63832897e-01 -1.79986155e+00
-4.18997049e-01 2.56991200e-02 -3.93991739e-01 -5.36392629e-02
4.19600010e-01 4.69331518e-02 -1.78473726e-01 1.75242573e-01
1.72336459e-01 6.37923181e-01 6.24744296e-01 -1.17160335e-01
-4.91288573e-01 3.80755395e-01 -2.69054234e-01 6.32802427e-01
8.91713500e-01 -3.32537830e-01 -5.27017176e-01 1.24567017e-01
-1.31282568e-01 2.31651720e-02 4.42372710e-01 -1.79172266e+00
2.11524591e-01 3.15920621e-01 8.87066007e-01 -7.47409403e-01
5.42834163e-01 -1.11122119e+00 2.36721799e-01 7.08570957e-01
3.89576942e-01 2.19068080e-02 -6.04215525e-02 8.83425236e-01
-1.86414540e-01 -8.81083980e-02 5.61352909e-01 -3.16133574e-02
-7.99000621e-01 -1.05577171e-01 -1.15385783e+00 -4.90197420e-01
1.06328654e+00 -1.13626671e+00 -9.77175310e-03 -5.36493897e-01
-1.02976584e+00 -4.67963547e-01 -3.41747224e-01 1.13802552e-01
9.15477097e-01 -1.42115390e+00 -6.05209172e-01 4.11490023e-01
1.13563798e-01 -2.74582744e-01 6.18409932e-01 1.18128502e+00
-6.95739925e-01 5.28914809e-01 -6.18163049e-01 -4.94364828e-01
-1.11911917e+00 3.09041709e-01 1.52400851e-01 -7.63560385e-02
-8.23778510e-01 3.89852554e-01 -4.75676060e-01 -4.29391190e-02
4.33710925e-02 -8.71232376e-02 -9.47600365e-01 4.48548496e-01
5.39818048e-01 9.63289261e-01 2.79650301e-01 -9.17685211e-01
-7.95812547e-01 8.03906262e-01 1.34997517e-01 -3.35018575e-01
1.49974656e+00 2.03452915e-01 3.11890423e-01 8.19638073e-02
1.26502347e+00 -2.60925949e-01 -7.33861625e-01 -2.22073961e-02
2.56733596e-01 -6.10521510e-02 3.05812865e-01 -5.35427392e-01
-6.78417325e-01 8.61230671e-01 1.47775924e+00 3.67852896e-01
1.28907144e+00 -2.67672360e-01 5.50538123e-01 6.95209622e-01
2.82153368e-01 -1.26591694e+00 -2.67310828e-01 1.69343680e-01
7.06910253e-01 -6.52579963e-01 3.99573773e-01 3.27172354e-02
-5.17536938e-01 1.42948163e+00 -8.74989331e-02 -2.45779291e-01
1.28448284e+00 2.71737278e-01 1.18584037e-01 -3.25518608e-01
-4.85903881e-02 -4.19416904e-01 2.00540155e-01 1.44900382e+00
2.61528313e-01 2.29663178e-01 -5.03850639e-01 5.54313064e-01
-4.18250889e-01 1.33455008e-01 1.71395421e-01 1.39462030e+00
-7.88034379e-01 -5.79983711e-01 -6.38773024e-01 7.87515461e-01
-2.80473292e-01 3.17566186e-01 -1.50827086e-02 9.63583231e-01
1.82082042e-01 9.54056621e-01 3.54622491e-02 -6.26903474e-01
3.77138704e-01 2.45975256e-01 5.04296184e-01 3.12449113e-02
-1.88942298e-01 3.18341888e-02 -4.40322198e-02 -3.54091942e-01
-1.02137125e+00 -9.03094649e-01 -1.07764554e+00 1.80614844e-03
-3.73957664e-01 1.99037060e-01 9.05600429e-01 1.17595470e+00
1.39869541e-01 6.40689433e-01 5.26576757e-01 -8.30206454e-01
-2.79291809e-01 -1.11273265e+00 -9.21991765e-01 -9.34096873e-02
2.84407198e-01 -1.06629467e+00 -5.09155631e-01 7.20146373e-02] | [13.949509620666504, 1.547965168952942] |
afea1aab-ebcc-4ef9-8499-03a24c2b4148 | probabilistic-knowledge-graph-embeddings | null | null | https://openreview.net/forum?id=rJ4qXnCqFX | https://openreview.net/pdf?id=rJ4qXnCqFX | Probabilistic Knowledge Graph Embeddings | We develop a probabilistic extension of state-of-the-art embedding models for link prediction in relational knowledge graphs. Knowledge graphs are collections of relational facts, where each fact states that a certain relation holds between two entities, such as people, places, or objects. We argue that knowledge graphs should be treated within a Bayesian framework because even large knowledge graphs typically contain only few facts per entity, leading effectively to a small data problem where parameter uncertainty matters. We introduce a probabilistic reinterpretation of the DistMult (Yang et al., 2015) and ComplEx (Trouillon et al., 2016) models and employ variational inference to estimate a lower bound on the marginal likelihood of the data. We find that the main benefit of the Bayesian approach is that it allows for efficient, gradient based optimization over hyperparameters, which would lead to divergences in a non-Bayesian treatment. Models with such learned hyperparameters improve over the state-of-the-art by a significant margin, as we demonstrate on several benchmarks. | ['Stephan Mandt', 'Robert Bamler', 'Farnood Salehi'] | 2018-09-27 | null | null | null | null | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [-1.88910440e-01 6.71007514e-01 -4.96790260e-01 -2.25119755e-01
-5.56851327e-01 -5.98408282e-01 9.86974239e-01 5.68623900e-01
-3.72299135e-01 7.23913968e-01 4.02446330e-01 -2.88497567e-01
-5.63478649e-01 -1.12409413e+00 -9.00741696e-01 -4.64800239e-01
-2.09235236e-01 8.76645029e-01 4.58375126e-01 -4.20602225e-02
-1.87966511e-01 1.56457350e-01 -1.17551947e+00 -2.44153917e-01
4.30316627e-01 7.42468536e-01 -2.02847496e-01 4.50702101e-01
-5.25480472e-02 8.30228448e-01 -2.29879037e-01 -1.17864585e+00
-1.73155785e-01 2.69098908e-01 -9.98592377e-01 -4.18367594e-01
3.66032720e-01 -1.69833437e-01 -6.90295279e-01 1.05413365e+00
6.68522045e-02 2.34213397e-01 1.14152014e+00 -1.21996295e+00
-7.92286932e-01 1.07269824e+00 -3.85079116e-01 2.42122665e-01
3.36474597e-01 -3.50200862e-01 1.68347502e+00 -6.56031787e-01
6.99486315e-01 1.41952455e+00 6.69443429e-01 7.20395222e-02
-1.66012061e+00 -2.08963826e-01 3.59784395e-01 5.70509136e-01
-1.53408360e+00 -1.47420943e-01 5.51041663e-01 -6.42167032e-01
9.70259547e-01 5.28523251e-02 6.82316542e-01 1.14229977e+00
1.02120593e-01 7.45745182e-01 6.56735063e-01 -3.85690928e-01
2.35475570e-01 1.63672924e-01 4.04311597e-01 8.94001961e-01
8.70907784e-01 -8.72077420e-02 -7.92071462e-01 -5.10413885e-01
5.63252807e-01 -1.45170256e-01 -1.32716626e-01 -8.28477502e-01
-1.18408060e+00 9.67366278e-01 4.67467397e-01 5.17745763e-02
-2.07070202e-01 5.24558961e-01 2.12346777e-01 -8.21098685e-02
5.08301139e-01 1.90178499e-01 -5.84095418e-01 -1.68000728e-01
-6.11853719e-01 4.89755899e-01 1.20677471e+00 9.15491700e-01
7.75404215e-01 -4.63567406e-01 -5.29536828e-02 5.80796480e-01
6.88554883e-01 3.61678362e-01 -2.68883258e-01 -9.25145268e-01
4.17612135e-01 4.12800789e-01 2.42549032e-01 -1.01924467e+00
-2.35295475e-01 -2.25886717e-01 -4.65128779e-01 -1.99131474e-01
5.72176039e-01 9.15569887e-02 -7.56254554e-01 1.94073510e+00
4.13491726e-01 3.84736925e-01 1.21688925e-01 4.94039983e-01
7.24748433e-01 4.28418636e-01 3.10312957e-01 1.80183165e-02
1.57785094e+00 -3.92002612e-01 -6.74694598e-01 -1.78133950e-01
5.78863740e-01 -3.22309941e-01 8.12134206e-01 3.24464291e-01
-7.97866166e-01 2.10486408e-02 -8.86508942e-01 -2.84813434e-01
-6.14072859e-01 -2.21749812e-01 1.07813931e+00 7.51827836e-01
-8.23932946e-01 6.17677212e-01 -1.09590971e+00 -4.51863587e-01
4.86915052e-01 1.81313142e-01 -4.09018397e-01 -6.63837418e-02
-1.30429852e+00 1.24468994e+00 6.16578460e-01 6.70282766e-02
-7.10975230e-01 -8.30859661e-01 -1.05640292e+00 3.24010938e-01
9.51991737e-01 -9.48880255e-01 8.42270434e-01 -1.36298344e-01
-1.27780986e+00 8.33629310e-01 -9.93110165e-02 -4.48020875e-01
2.49469668e-01 -3.41328114e-01 -2.95230627e-01 -1.19518861e-02
-1.25941679e-01 3.28255326e-01 5.96011639e-01 -1.22103155e+00
-3.25675130e-01 -3.31999272e-01 4.13679421e-01 8.02356228e-02
-6.81565329e-02 -1.77289978e-01 -6.57045305e-01 -2.80650556e-01
1.19611859e-01 -8.79199326e-01 1.48054332e-01 -5.43781817e-02
-6.58690989e-01 -5.34853995e-01 3.41651231e-01 -5.35074532e-01
1.10179889e+00 -1.86053514e+00 5.18101871e-01 3.78954291e-01
3.63181293e-01 -8.67596120e-02 2.25652918e-01 4.10266846e-01
1.86949000e-01 2.68910527e-01 -1.46758750e-01 -3.68954659e-01
4.92986143e-01 4.89002645e-01 -4.55592155e-01 5.36211133e-01
1.62121639e-01 1.06664610e+00 -1.01303232e+00 -6.10978484e-01
5.33430651e-02 6.19014740e-01 -5.72008193e-01 -1.49037898e-01
-4.74467665e-01 -1.44826978e-01 -5.99186599e-01 4.55658972e-01
2.35016629e-01 -5.60368836e-01 6.00565672e-01 -4.16880488e-01
3.55632603e-01 6.84661746e-01 -1.46960831e+00 1.57206225e+00
-2.20207572e-01 4.12265182e-01 -1.68054447e-01 -8.78193378e-01
4.04786766e-01 2.23550379e-01 2.69741327e-01 2.26602450e-01
2.58317981e-02 -1.67078197e-01 -3.55933756e-02 -3.95719856e-01
6.27816379e-01 -1.98789895e-01 -9.80881527e-02 3.04264128e-01
3.80092919e-01 -2.75374472e-01 1.97630808e-01 6.85619891e-01
1.09210587e+00 4.47904050e-01 2.64928430e-01 -2.56006628e-01
-9.33186268e-04 -2.78702289e-01 6.35301888e-01 9.20110524e-01
1.08285751e-02 4.50470895e-02 9.37502503e-01 -9.32556018e-02
-8.10158312e-01 -1.54065323e+00 -4.23560947e-01 8.73748839e-01
3.08507271e-02 -8.93175900e-01 -1.99003577e-01 -7.99995840e-01
4.62350100e-01 8.73241603e-01 -8.68129909e-01 -1.95244923e-01
-5.24199521e-03 -8.40465367e-01 4.30566967e-01 5.37148595e-01
-1.65405855e-01 -4.64958221e-01 -2.51806110e-01 1.17403999e-01
8.59881341e-02 -1.04198492e+00 -1.57658681e-02 3.60968262e-01
-7.25593626e-01 -1.15973473e+00 -3.23864877e-01 -2.80493975e-01
2.92808086e-01 -3.98659825e-01 1.44553256e+00 -1.70730561e-01
-1.51801601e-01 6.79741800e-01 -1.37328222e-01 -3.45324934e-01
-1.64148301e-01 2.13715881e-02 1.60526037e-01 -2.63881475e-01
5.52129447e-01 -7.25715041e-01 -2.52878428e-01 -1.85924590e-01
-7.90854871e-01 -2.39442319e-01 3.63868237e-01 8.37759793e-01
6.13550186e-01 1.68003410e-01 2.63127953e-01 -9.95277941e-01
3.44543010e-01 -7.24567056e-01 -7.36294270e-01 6.89128578e-01
-7.72897899e-01 5.16694903e-01 1.13267072e-01 -3.48898351e-01
-9.04505253e-01 -3.69415879e-01 4.35684234e-01 -1.76046416e-01
1.88127428e-01 1.07257426e+00 -2.38509655e-01 1.83793545e-01
4.15006161e-01 -1.30493507e-01 -3.48023057e-01 -4.00658727e-01
9.65065837e-01 2.59378165e-01 3.04771602e-01 -1.11426795e+00
7.85860419e-01 5.00245452e-01 3.82103175e-01 -5.80911100e-01
-1.12424147e+00 -3.06444526e-01 -7.13911057e-01 9.14122090e-02
7.87004828e-01 -9.72318828e-01 -1.08068788e+00 -1.80229262e-01
-9.68401611e-01 -2.99990177e-01 -2.73003578e-01 6.91742659e-01
-5.21341443e-01 4.38607901e-01 -5.85152447e-01 -8.67603958e-01
2.50718325e-01 -7.16592550e-01 9.10685241e-01 2.23069936e-02
-3.89698744e-01 -1.33593988e+00 2.37279609e-01 2.36649066e-01
9.30138677e-03 6.95206374e-02 1.35936797e+00 -7.20095396e-01
-8.27835321e-01 -5.42085506e-02 -2.28822663e-01 7.86938705e-03
-7.91883618e-02 3.10464382e-01 -7.24652469e-01 -1.37442295e-02
-4.35237974e-01 -4.07745719e-01 1.19290316e+00 3.17122966e-01
7.86123037e-01 -3.41070652e-01 -6.39576018e-01 3.38047534e-01
1.41990066e+00 -4.57545966e-01 4.88454074e-01 5.44358194e-02
8.15546334e-01 5.18717468e-01 3.44711512e-01 3.06463063e-01
8.65890622e-01 7.92972386e-01 2.70583481e-01 6.46156728e-01
1.22778751e-01 -5.25208414e-01 9.18442905e-02 6.64555252e-01
-1.80525467e-01 -3.29125583e-01 -1.06735909e+00 8.38303864e-01
-2.12004757e+00 -1.12069571e+00 -1.11983076e-01 2.31418872e+00
1.29700029e+00 2.21363336e-01 6.44422024e-02 -2.04032913e-01
4.26815271e-01 2.95828819e-01 -4.09861922e-01 1.20108388e-01
-8.78834873e-02 2.46325165e-01 7.09597766e-01 9.08740222e-01
-1.03681636e+00 1.01950932e+00 6.61994600e+00 7.68212736e-01
-3.17678928e-01 1.31881073e-01 1.66644230e-01 -9.98919457e-02
-7.53767073e-01 4.64520544e-01 -8.96937490e-01 3.88337046e-01
9.15631890e-01 -2.24368975e-01 5.10209739e-01 6.33034110e-01
-4.74742472e-01 -4.06432092e-01 -1.41180980e+00 7.50973225e-01
-6.30523190e-02 -1.32033837e+00 3.54002090e-03 3.09255749e-01
7.40681529e-01 1.18370637e-01 -1.16635613e-01 3.60672057e-01
9.92511451e-01 -1.03918397e+00 6.48537040e-01 8.49183142e-01
2.60625690e-01 -7.39859998e-01 5.00014663e-01 8.11213478e-02
-9.80072320e-01 2.44313732e-01 -4.61845785e-01 2.08839387e-01
3.13665867e-01 9.89001274e-01 -5.59233665e-01 7.82044172e-01
6.38241291e-01 7.06372976e-01 -4.59278733e-01 6.56411946e-01
-6.66499674e-01 7.29383409e-01 -6.53327048e-01 -2.11777821e-01
-1.92468703e-01 -3.58387709e-01 5.66027939e-01 8.81141186e-01
3.69295999e-02 3.84043008e-02 -1.93533674e-02 1.11644351e+00
-2.59317845e-01 -2.07101971e-01 -5.20037234e-01 -1.70867935e-01
6.42123580e-01 1.06200051e+00 -5.47241628e-01 -3.60784709e-01
-5.36507905e-01 4.28345889e-01 7.64046907e-01 4.98375922e-01
-6.86986923e-01 -1.70706347e-01 5.77252388e-01 -9.32881013e-02
7.04075336e-01 -2.47471288e-01 -1.64105222e-01 -1.29543626e+00
1.58892721e-01 -4.37104255e-01 5.47397673e-01 -6.18533909e-01
-1.56309474e+00 -8.26256871e-02 4.82313395e-01 -3.98634255e-01
-3.37051928e-01 -8.26789916e-01 -2.10201591e-01 8.74671221e-01
-1.38266683e+00 -1.16186881e+00 2.26584151e-01 4.00160074e-01
-4.05200869e-01 5.92889376e-02 9.64110136e-01 -9.06112865e-02
-4.92665917e-01 3.71229023e-01 1.89430460e-01 1.04863383e-01
4.33208644e-01 -1.63963807e+00 2.16262266e-01 7.10537732e-01
6.10966682e-01 9.27076101e-01 9.15352225e-01 -8.75426352e-01
-1.54312289e+00 -7.21016049e-01 1.03555799e+00 -1.07283759e+00
1.12175655e+00 -4.79916006e-01 -8.37362289e-01 1.22310793e+00
-3.06652244e-02 1.46330029e-01 7.40902126e-01 1.14169312e+00
-8.39570284e-01 9.08569098e-02 -9.15797889e-01 7.28202760e-01
8.69927585e-01 -7.79387712e-01 -9.05291080e-01 3.53200078e-01
6.37441397e-01 -2.08737135e-01 -1.33730125e+00 3.19696128e-01
6.03105783e-01 -7.87595809e-01 1.04782164e+00 -8.28881383e-01
3.17038715e-01 -3.25792998e-01 -5.20541489e-01 -1.27370167e+00
-2.99884051e-01 -5.32248557e-01 -8.92121553e-01 1.22009170e+00
6.31694615e-01 -6.03145242e-01 6.75805211e-01 9.25253093e-01
3.32460463e-01 -7.20498264e-01 -1.05799496e+00 -9.54338431e-01
1.21130913e-01 -5.36554217e-01 4.89235282e-01 9.48128641e-01
3.41377616e-01 5.55883944e-01 -2.02710271e-01 4.31553006e-01
7.99318612e-01 -6.85760528e-02 5.59565544e-01 -1.67081702e+00
-6.19792879e-01 -3.42731506e-01 -7.44998693e-01 -9.54791427e-01
5.56697845e-01 -9.59216416e-01 -1.92182735e-01 -1.79703534e+00
5.38220465e-01 -3.88440281e-01 -3.30984205e-01 5.69598675e-01
-4.12910938e-01 -1.70031995e-01 -5.26063442e-02 -3.05787008e-02
-6.57660663e-01 7.82725632e-01 6.25637531e-01 -1.75897345e-01
1.63242370e-01 -1.76570743e-01 -6.86736643e-01 6.72992051e-01
3.56600761e-01 -7.24508345e-01 -5.51307619e-01 -2.81979382e-01
9.46474671e-01 -5.72015867e-02 8.27465057e-01 -2.98599780e-01
2.40494862e-01 -8.95732269e-02 1.49475589e-01 -4.19955999e-01
6.26498520e-01 -6.46158040e-01 2.02947736e-01 1.08972870e-01
-3.74903858e-01 -3.31821620e-01 8.32600147e-02 1.21258044e+00
-1.55575797e-02 -2.40466803e-01 2.60985672e-01 9.03729796e-02
-3.57869565e-01 1.99002787e-01 6.34925440e-02 2.55373687e-01
6.77559912e-01 1.13377333e-01 -4.74759996e-01 -4.54272270e-01
-8.26398253e-01 1.85090870e-01 3.33295643e-01 2.04019681e-01
2.51073956e-01 -1.36263049e+00 -5.48954010e-01 -4.45706367e-01
2.25863919e-01 9.26727578e-02 9.17233378e-02 8.56883883e-01
3.01514659e-02 4.48094517e-01 4.13268328e-01 -4.00419354e-01
-8.56648445e-01 5.87633789e-01 1.64055169e-01 -3.97480637e-01
-5.52666366e-01 9.49795246e-01 -3.30023430e-02 -4.50519651e-01
1.29744291e-01 -3.26237440e-01 7.89634660e-02 8.74975845e-02
1.15429558e-01 4.89185840e-01 -1.22650184e-01 -3.98065329e-01
-6.83804035e-01 2.86800504e-01 -2.09712118e-01 -3.40930700e-01
1.35977709e+00 -3.12785655e-02 -1.68844670e-01 7.64035940e-01
8.95849645e-01 2.92383224e-01 -1.25567245e+00 -5.65923691e-01
5.73957674e-02 -3.36555034e-01 2.75433689e-01 -8.27566624e-01
-6.72343254e-01 5.39664865e-01 1.63263068e-01 3.91562581e-01
3.12257946e-01 6.34613752e-01 1.72871858e-01 5.58105052e-01
5.25256932e-01 -9.65705752e-01 -3.84296805e-01 4.80808645e-01
5.94329417e-01 -1.21172523e+00 5.27356744e-01 -5.32383800e-01
-4.67169732e-01 8.47169399e-01 1.24622144e-01 -2.30478738e-02
1.03715336e+00 -3.29331640e-04 -4.74578559e-01 -5.13565183e-01
-9.93066549e-01 -3.79595309e-01 6.21513247e-01 6.91899955e-01
3.85315806e-01 3.56487960e-01 -8.25739503e-02 6.52178586e-01
-2.89809376e-01 -2.79417038e-01 4.15559173e-01 7.02609360e-01
-1.84894890e-01 -1.08270967e+00 -3.29750925e-02 5.91443539e-01
-4.83067781e-01 -3.96272719e-01 -2.54653007e-01 9.96223629e-01
-1.03118017e-01 8.59859169e-01 -4.34548333e-02 -8.88522491e-02
7.08706900e-02 3.11697155e-01 9.21534061e-01 -7.64258564e-01
5.87565415e-02 -2.66191661e-01 2.82011002e-01 -4.38287973e-01
-5.03317177e-01 -8.55251610e-01 -9.05903280e-01 -4.28797752e-01
-3.94676030e-01 1.15914010e-01 5.30182958e-01 1.13897729e+00
3.40505272e-01 3.88780773e-01 -6.55001849e-02 -4.13429588e-01
-6.31916702e-01 -9.99062657e-01 -8.54954064e-01 1.99131146e-01
8.01310763e-02 -1.01169109e+00 -3.74948084e-01 1.70452036e-02] | [8.753222465515137, 7.604669094085693] |
2365a4fa-968e-4698-b64e-5c42d68e01b5 | convolutional-neural-networks-for-fast | 1809.04440 | null | https://arxiv.org/abs/1809.04440v2 | https://arxiv.org/pdf/1809.04440v2.pdf | Learning-based Efficient Graph Similarity Computation via Multi-Scale Convolutional Set Matching | Graph similarity computation is one of the core operations in many graph-based applications, such as graph similarity search, graph database analysis, graph clustering, etc. Since computing the exact distance/similarity between two graphs is typically NP-hard, a series of approximate methods have been proposed with a trade-off between accuracy and speed. Recently, several data-driven approaches based on neural networks have been proposed, most of which model the graph-graph similarity as the inner product of their graph-level representations, with different techniques proposed for generating one embedding per graph. However, using one fixed-dimensional embedding per graph may fail to fully capture graphs in varying sizes and link structures, a limitation that is especially problematic for the task of graph similarity computation, where the goal is to find the fine-grained difference between two graphs. In this paper, we address the problem of graph similarity computation from another perspective, by directly matching two sets of node embeddings without the need to use fixed-dimensional vectors to represent whole graphs for their similarity computation. The model, GraphSim, achieves the state-of-the-art performance on four real-world graph datasets under six out of eight settings (here we count a specific dataset and metric combination as one setting), compared to existing popular methods for approximate Graph Edit Distance (GED) and Maximum Common Subgraph (MCS) computation. | ['Yunsheng Bai', 'Hao Ding', 'Yizhou Sun', 'Wei Wang'] | 2018-09-10 | null | null | null | null | ['set-matching', 'graph-similarity'] | ['computer-vision', 'graphs'] | [ 1.91835426e-02 -1.20420471e-01 -2.78421324e-02 -3.10713947e-01
-3.44704300e-01 -6.07023299e-01 4.62584943e-01 1.14567089e+00
-3.87809992e-01 2.33189374e-01 -1.52314156e-01 -3.02341759e-01
-4.52714413e-01 -1.27090549e+00 -3.52574706e-01 -4.39740509e-01
-3.90232176e-01 5.75912774e-01 1.18865713e-01 -2.12122679e-01
3.63318622e-01 5.62662899e-01 -1.44289565e+00 -4.02462631e-01
8.39803457e-01 9.40993249e-01 -4.01037559e-03 6.14675701e-01
-4.33496565e-01 1.80559307e-01 -3.26554358e-01 -8.80767822e-01
3.18910748e-01 -2.89980143e-01 -7.23083019e-01 -1.47664636e-01
6.73322678e-01 7.95271099e-02 -6.92707181e-01 1.45366335e+00
5.25643945e-01 3.37094277e-01 4.95998412e-01 -1.62704229e+00
-7.56771326e-01 5.89053214e-01 -4.67533708e-01 1.94752544e-01
6.79303885e-01 -1.86639711e-01 1.36687613e+00 -5.04276931e-01
5.50739706e-01 1.12480664e+00 7.50610232e-01 5.13618514e-02
-1.24851859e+00 -3.83167952e-01 -1.19931884e-01 5.19025505e-01
-1.73039258e+00 1.91705436e-01 8.44645500e-01 -2.53822446e-01
1.03265846e+00 2.45395496e-01 7.89688230e-01 3.32273275e-01
1.99612468e-01 1.23558268e-01 6.80241942e-01 -2.96343684e-01
2.38048241e-01 -1.86807662e-01 2.49972627e-01 8.66713703e-01
5.71798503e-01 -1.66457593e-01 -2.54761204e-02 -3.70173991e-01
3.32352936e-01 2.70612568e-01 -3.72377396e-01 -7.63519645e-01
-1.13500190e+00 1.03239489e+00 7.42590308e-01 6.09199584e-01
-1.05687253e-01 2.02630430e-01 7.72078037e-01 6.65247202e-01
3.27748716e-01 6.41845465e-01 4.68817763e-02 1.01226486e-01
-7.23714769e-01 2.25413769e-01 1.05594265e+00 9.67246830e-01
1.06082177e+00 -2.79146791e-01 4.31281440e-02 6.73390627e-01
4.42110077e-02 5.71587868e-02 5.19513667e-01 -4.11216259e-01
4.59842950e-01 1.16057014e+00 -5.21533430e-01 -1.79343426e+00
-2.66429484e-01 -1.16511546e-01 -1.25659609e+00 -2.76184022e-01
3.80593508e-01 3.73881251e-01 -5.72188079e-01 1.66293085e+00
5.65114677e-01 3.77398193e-01 -3.09644163e-01 7.00112820e-01
8.01049531e-01 4.88765299e-01 -2.15560645e-01 -5.35665900e-02
1.11495662e+00 -8.17331016e-01 -3.05411011e-01 1.04772216e-02
1.03892231e+00 -5.23777366e-01 1.01582587e+00 -3.17925774e-03
-8.56333375e-01 -2.75767893e-01 -1.19999218e+00 -1.97003931e-01
-9.80837882e-01 -4.66083616e-01 6.20163500e-01 5.34736216e-01
-1.20578432e+00 1.04786193e+00 -3.82891238e-01 -6.54570043e-01
3.66152599e-02 2.55722225e-01 -7.72354782e-01 -4.34092790e-01
-1.18000758e+00 8.35403204e-01 6.21729374e-01 -3.66548710e-02
-1.31248280e-01 -5.11278391e-01 -1.16481268e+00 3.65621537e-01
3.60631496e-01 -5.74723840e-01 5.13391495e-01 -6.15744174e-01
-9.09917235e-01 9.48525310e-01 1.95417345e-01 -5.24581969e-01
3.92940417e-02 3.73257667e-01 -5.30081987e-01 -1.07798092e-02
-4.36401442e-02 2.37340137e-01 6.33953452e-01 -5.21443903e-01
-2.78297216e-01 -4.83158976e-01 3.88229042e-01 9.88396257e-02
-6.85986698e-01 -1.13857538e-01 -4.61652040e-01 -5.94537675e-01
1.32902026e-01 -8.89626086e-01 -2.40258262e-01 2.34976500e-01
-3.02575320e-01 -5.59478343e-01 5.24597824e-01 -4.56051916e-01
1.55968797e+00 -2.01955676e+00 4.29444909e-01 6.06235564e-01
7.04751790e-01 3.13989401e-01 -3.40256721e-01 8.68999541e-01
-2.62333483e-01 1.39645457e-01 -5.88575602e-01 -7.99366981e-02
1.80153564e-01 1.82759538e-01 8.79527330e-02 6.69605255e-01
2.01510102e-01 7.44570673e-01 -1.06514525e+00 -5.37156820e-01
1.92775622e-01 3.24819624e-01 -3.79467934e-01 3.54657441e-01
2.05014437e-01 -3.66337121e-01 -1.63895294e-01 3.26043487e-01
7.85940409e-01 -2.59587497e-01 5.56314707e-01 -3.09207559e-01
3.38027328e-01 3.01761553e-02 -1.43431687e+00 1.73915648e+00
-5.10768890e-01 5.45607209e-01 -1.97822943e-01 -1.34043312e+00
1.06163073e+00 -4.82094325e-02 5.46043754e-01 -5.79963684e-01
8.08395594e-02 2.27167830e-01 1.97269082e-01 -7.74463043e-02
5.59889078e-01 2.02809945e-01 -1.75851002e-01 7.55412400e-01
-1.06840655e-01 -1.09157898e-01 6.77967489e-01 4.56118792e-01
1.50780869e+00 -5.44997811e-01 6.04430497e-01 -2.35092670e-01
7.10799813e-01 -2.62757093e-01 1.22202605e-01 2.85131902e-01
-7.54118487e-02 5.03182530e-01 6.04646027e-01 -7.39611030e-01
-9.99969900e-01 -1.03096354e+00 3.25013727e-01 8.37856293e-01
4.19039547e-01 -7.94671416e-01 -7.93288827e-01 -6.82652533e-01
2.92216063e-01 3.16157997e-01 -5.02113104e-01 -6.08650982e-01
-5.53291738e-01 -3.90326530e-01 4.19094831e-01 2.72144407e-01
1.29464641e-01 -7.35975504e-01 -1.87505841e-01 2.14894131e-01
8.37250501e-02 -1.10052061e+00 -1.05024517e+00 -1.46559939e-01
-8.14016819e-01 -1.40704727e+00 -4.57380533e-01 -9.33632731e-01
8.69822085e-01 5.45867801e-01 1.32448423e+00 6.93410099e-01
-4.03743118e-01 3.22763115e-01 -3.51314932e-01 3.01996380e-01
-4.42472637e-01 1.66639343e-01 6.22820184e-02 3.81427296e-02
3.51247787e-01 -8.97743523e-01 -3.82340968e-01 1.57975718e-01
-1.13234115e+00 -2.21097231e-01 5.00745475e-01 7.57243037e-01
6.93493843e-01 1.72289327e-01 2.55127400e-01 -7.52504230e-01
1.18745506e+00 -5.26998281e-01 -6.02984130e-01 4.99367714e-01
-9.10783947e-01 3.99396867e-01 9.94071245e-01 -3.43877375e-01
-5.26746400e-02 -2.24602297e-01 1.04019471e-01 -6.02773190e-01
2.43696958e-01 7.79514611e-01 -1.21360362e-01 -3.92198205e-01
4.17890549e-01 2.96612084e-01 9.08715799e-02 -2.11834848e-01
5.83909750e-01 4.24357742e-01 4.24335927e-01 -3.26280832e-01
9.19646025e-01 6.32830709e-02 3.86061907e-01 -6.34709239e-01
-2.45720163e-01 -6.35355949e-01 -6.96594656e-01 1.12223908e-01
4.22363013e-01 -3.02401364e-01 -6.74777389e-01 1.39186159e-01
-1.15039015e+00 1.64729640e-01 -1.28084034e-01 3.77003849e-01
-4.21915442e-01 9.76290286e-01 -4.27189291e-01 -1.73688963e-01
-7.29386508e-01 -1.08529592e+00 8.26502979e-01 -5.50330319e-02
-1.29305050e-01 -1.15102816e+00 3.55126828e-01 -2.70305693e-01
4.00684386e-01 5.35665691e-01 1.31050050e+00 -8.49638462e-01
-4.05300498e-01 -6.19660795e-01 -5.44337451e-01 1.16028041e-01
2.54125297e-01 -4.67343293e-02 -2.46214762e-01 -6.92601919e-01
-4.70405310e-01 -3.04582510e-02 4.69989866e-01 2.53156256e-02
1.18667042e+00 -2.90487140e-01 -3.07329416e-01 7.21712828e-01
1.66388857e+00 -3.10492575e-01 3.26709121e-01 -3.92759293e-02
9.12372649e-01 4.06923741e-01 4.50093389e-01 2.77782559e-01
4.94745344e-01 6.79383218e-01 5.16192317e-01 2.50392288e-01
-5.69527075e-02 -3.82097632e-01 -9.96539649e-03 1.16729987e+00
2.77072489e-01 -4.26778108e-01 -7.95860291e-01 4.80612099e-01
-1.69646871e+00 -7.52092600e-01 -4.26522046e-02 2.52170992e+00
2.95087427e-01 -1.04998387e-02 1.95238486e-01 2.76761591e-01
1.11693561e+00 3.07775468e-01 -4.77526814e-01 -8.63083899e-01
3.24501634e-01 2.63578743e-01 4.44307864e-01 2.87481099e-01
-7.59134531e-01 5.55739641e-01 5.44347334e+00 7.87087202e-01
-1.02889609e+00 -1.34461612e-01 2.92217493e-01 3.02034467e-01
-3.82959455e-01 1.23939849e-01 -8.57164264e-02 5.75577796e-01
9.74664867e-01 -6.99447393e-01 8.15667927e-01 8.07172835e-01
-3.62370580e-01 1.61937118e-01 -1.44469750e+00 1.26679051e+00
1.45938173e-01 -1.25080621e+00 1.83966473e-01 1.23483814e-01
4.09389824e-01 -7.27492645e-02 -3.81518513e-01 2.13212863e-01
-1.42343296e-02 -1.08169675e+00 1.12741031e-01 3.68227214e-01
9.44258630e-01 -9.70703125e-01 7.26234257e-01 2.77067721e-02
-1.93027627e+00 1.08400539e-01 -5.16720057e-01 4.55772877e-02
4.46260050e-02 7.31025755e-01 -7.70715892e-01 8.36915791e-01
5.87656379e-01 6.95941925e-01 -6.11399531e-01 1.12852359e+00
1.36389583e-01 -5.21710813e-02 -3.43278080e-01 -2.74200499e-01
2.53898799e-01 -5.51133037e-01 5.40975928e-01 8.65514100e-01
5.19489527e-01 -2.83984423e-01 2.08278686e-01 8.12176943e-01
-4.44950789e-01 3.39765847e-01 -9.82191086e-01 -3.86266291e-01
7.60896027e-01 1.44044077e+00 -8.04849207e-01 -2.58960366e-01
-3.92941207e-01 1.00975800e+00 8.24484289e-01 -1.74378380e-01
-8.11994314e-01 -1.03873467e+00 7.45050013e-01 1.17059760e-01
1.01138167e-01 -5.76390982e-01 1.83837637e-01 -8.53592038e-01
2.38320783e-01 -7.67755210e-01 7.72580028e-01 -4.00340408e-01
-1.46661401e+00 6.67977750e-01 2.03429293e-02 -1.19788516e+00
-3.15765649e-01 -5.54153860e-01 -8.73088837e-01 6.93476021e-01
-1.11921477e+00 -8.35179806e-01 -6.35067105e-01 6.02246344e-01
1.66396312e-02 1.26709461e-01 7.94105828e-01 6.17707193e-01
-4.09087420e-01 9.11052704e-01 1.98835805e-01 1.30106375e-01
4.52219874e-01 -1.26636195e+00 9.15646732e-01 7.09503174e-01
4.55483466e-01 4.87866968e-01 5.48298836e-01 -3.71406019e-01
-2.06007218e+00 -1.05086470e+00 9.58759964e-01 -5.56889363e-02
8.08744550e-01 -3.14207703e-01 -1.05872893e+00 3.28327268e-01
-1.44192129e-01 5.38530111e-01 3.36879045e-01 -6.95879161e-02
-4.01201665e-01 -2.77269632e-01 -1.33122826e+00 5.60588002e-01
1.42164445e+00 -7.37425864e-01 -1.38648108e-01 3.24557722e-01
8.32302988e-01 -2.14152157e-01 -1.35216331e+00 1.24641337e-01
4.85409260e-01 -8.66160870e-01 9.61581826e-01 -5.96042931e-01
-1.60112176e-02 -3.64059567e-01 -1.75163731e-01 -1.56373811e+00
-3.30050945e-01 -5.03425956e-01 -2.10422829e-01 1.06077766e+00
6.25592694e-02 -8.80104661e-01 7.16164947e-01 3.06919456e-01
2.25761518e-01 -9.48080480e-01 -9.27406907e-01 -8.99929464e-01
-2.99754947e-01 -9.44761038e-02 1.16107249e+00 1.13351870e+00
6.42844811e-02 2.76786298e-01 2.06301287e-02 8.95062834e-02
4.84997094e-01 5.13406992e-01 8.86505246e-01 -1.37063265e+00
-6.45684451e-02 -7.02564299e-01 -1.27240062e+00 -5.49992561e-01
2.43259117e-01 -1.36352766e+00 -2.78313607e-01 -1.67138767e+00
5.15884068e-03 -3.05690169e-01 -1.41756549e-01 9.03471708e-02
1.31208040e-02 2.89917383e-02 8.95546973e-02 -1.99663207e-01
-3.57365489e-01 5.53983927e-01 8.93473744e-01 -4.09365803e-01
1.11980271e-02 -2.33253032e-01 -6.27831340e-01 3.79996628e-01
6.60399735e-01 -5.25969565e-01 -6.07622385e-01 -4.31186050e-01
3.02358299e-01 -1.67753384e-01 1.98980600e-01 -1.02987158e+00
5.03223240e-01 -1.20808773e-01 -3.39852571e-01 -2.84963071e-01
1.21974833e-01 -8.42953920e-01 4.18403476e-01 6.37133420e-01
-1.20706514e-01 8.45360577e-01 -2.39642933e-01 8.75259817e-01
-4.25846040e-01 -2.32557118e-01 6.65597916e-01 4.20928784e-02
-7.12651372e-01 8.54768872e-01 2.87419379e-01 2.76857167e-01
1.07021654e+00 -4.33604807e-01 -1.71864837e-01 -2.94511795e-01
-3.30240577e-01 6.81254193e-02 6.02678657e-01 4.04375553e-01
8.08032274e-01 -1.59158576e+00 -6.11758411e-01 -2.53369696e-02
3.47373515e-01 -9.40484554e-02 6.41118512e-02 7.14545906e-01
-6.16100252e-01 1.57044366e-01 -1.45229593e-01 -3.31082076e-01
-1.26798403e+00 8.44046712e-01 2.86143541e-01 -6.55335784e-01
-5.87194800e-01 5.11481524e-01 -1.43513396e-01 -6.83551311e-01
7.00308532e-02 -4.33780193e-01 8.21269155e-02 -5.08187562e-02
2.65517622e-01 6.59815013e-01 2.62177438e-01 -6.61310971e-01
-3.99679601e-01 7.38080323e-01 3.04778554e-02 6.39815748e-01
1.07116210e+00 1.76403672e-02 -5.17722368e-01 7.73856640e-02
1.71889746e+00 -3.72384340e-01 -3.09443176e-01 -4.15994734e-01
2.07137167e-01 -6.76065624e-01 -1.27133086e-01 1.14006720e-01
-1.18597698e+00 8.25585663e-01 5.59823573e-01 6.04271710e-01
1.05099535e+00 -9.10442546e-02 1.14910114e+00 6.99586213e-01
3.73530984e-01 -9.91766393e-01 -3.04864030e-02 2.88322002e-01
8.22432876e-01 -1.15130770e+00 2.78041333e-01 -3.76572728e-01
-1.41245380e-01 1.24642360e+00 5.21030843e-01 -4.41260904e-01
7.77408361e-01 -2.06149831e-01 -7.82858074e-01 -3.64881217e-01
-4.03236270e-01 -2.42444202e-01 5.56645215e-01 5.17827511e-01
3.28853816e-01 1.97437540e-01 -5.70839345e-01 1.07448772e-01
-1.78063110e-01 -4.40895140e-01 3.20738405e-01 7.23950505e-01
-2.70932972e-01 -1.18716514e+00 1.52747780e-01 8.86309743e-01
1.29389599e-01 -1.36220783e-01 -5.68738937e-01 8.23279679e-01
-2.18185768e-01 6.38223231e-01 2.04040661e-01 -5.66928148e-01
3.00456703e-01 -3.38535964e-01 4.24939990e-01 -5.03580272e-01
-4.40699339e-01 -9.74014282e-01 -1.67811438e-02 -7.07712770e-01
-1.52532861e-01 -2.11497620e-01 -1.14176214e+00 -8.90834332e-01
-5.12578368e-01 1.27098992e-01 5.86149096e-01 7.36247838e-01
6.22554958e-01 1.14179976e-01 8.06958914e-01 -5.96981883e-01
-5.80310225e-01 -6.66630983e-01 -8.21810782e-01 8.07609797e-01
-7.18688741e-02 -6.48475349e-01 -3.72719020e-01 -6.31598234e-01] | [7.1628899574279785, 6.029603004455566] |
8da5544d-0e29-4ac1-8f27-7e6a59921993 | online-simulator-based-experimental-design | 2303.02227 | null | https://arxiv.org/abs/2303.02227v1 | https://arxiv.org/pdf/2303.02227v1.pdf | Online simulator-based experimental design for cognitive model selection | The problem of model selection with a limited number of experimental trials has received considerable attention in cognitive science, where the role of experiments is to discriminate between theories expressed as computational models. Research on this subject has mostly been restricted to optimal experiment design with analytically tractable models. However, cognitive models of increasing complexity, with intractable likelihoods, are becoming more commonplace. In this paper, we propose BOSMOS: an approach to experimental design that can select between computational models without tractable likelihoods. It does so in a data-efficient manner, by sequentially and adaptively generating informative experiments. In contrast to previous approaches, we introduce a novel simulator-based utility objective for design selection, and a new approximation of the model likelihood for model selection. In simulated experiments, we demonstrate that the proposed BOSMOS technique can accurately select models in up to 2 orders of magnitude less time than existing LFI alternatives for three cognitive science tasks: memory retention, sequential signal detection and risky choice. | ['Andrew Howes', 'Samuel Kaski', 'Luigi Acerbi', 'Suyog Chandramouli', 'Gregoire Clarte', 'Aini Putkonen', 'Alexander Aushev'] | 2023-03-03 | null | null | null | null | ['experimental-design'] | ['methodology'] | [ 4.64789212e-01 -3.56694400e-01 -2.00401306e-01 -4.23401207e-01
-6.85501993e-01 -6.62431538e-01 4.76534724e-01 4.20051217e-01
-9.04545188e-01 1.06847203e+00 -3.20799083e-01 -6.37986422e-01
-6.09201670e-01 -4.87254947e-01 -4.37850267e-01 -2.81537473e-01
-2.01181218e-01 5.32337368e-01 1.52879521e-01 -2.28201076e-02
9.41993535e-01 4.83601868e-01 -1.99159074e+00 -3.10468912e-01
1.04645157e+00 8.61250341e-01 4.56848383e-01 6.74301863e-01
3.27868104e-01 4.29084778e-01 -5.13309896e-01 7.07376227e-02
1.39782995e-01 -6.46389425e-01 -6.39007270e-01 -2.71105289e-01
-1.26913875e-01 -1.13945261e-01 9.07598529e-03 8.83489788e-01
9.29137051e-01 4.07122254e-01 5.65722346e-01 -1.12564802e+00
-2.31190220e-01 6.53154671e-01 -2.53610998e-01 6.27842665e-01
5.03179729e-01 2.03162313e-01 6.43572867e-01 -6.80436313e-01
3.57090771e-01 1.38938427e+00 1.04789920e-01 3.16781163e-01
-1.93836021e+00 -9.07404840e-01 1.57846898e-01 1.60821170e-01
-1.87338901e+00 -5.89932442e-01 5.03138304e-01 -3.34259599e-01
7.79182315e-01 5.01951456e-01 7.10037768e-01 9.09170210e-01
4.82025146e-01 6.43871605e-01 1.64401913e+00 -8.10975730e-01
9.77793813e-01 1.57591268e-01 3.91056806e-01 2.40962744e-01
6.49937928e-01 7.54478335e-01 -7.60603130e-01 -4.73333210e-01
6.92948759e-01 -4.27112877e-01 -1.36381865e-01 -2.86428422e-01
-8.72287631e-01 9.30330276e-01 -3.04417968e-01 2.57871449e-01
-4.01656270e-01 9.18712541e-02 -6.81027249e-02 5.07918417e-01
1.76006272e-01 1.01804233e+00 -2.84286708e-01 -2.24281311e-01
-1.06467772e+00 8.20915401e-01 7.81345606e-01 9.60221469e-01
3.05574268e-01 2.25856364e-01 -5.10588229e-01 5.24237156e-01
4.07992423e-01 4.42306429e-01 4.57946092e-01 -8.42878878e-01
5.65249436e-02 2.00916827e-01 6.35196149e-01 -6.12375498e-01
-6.53189063e-01 -7.50750124e-01 -2.76952535e-01 3.24626535e-01
3.49911481e-01 -1.86218694e-01 -6.23571396e-01 1.95852566e+00
1.65718749e-01 -1.48576945e-01 -2.45535120e-01 7.70628273e-01
-5.84174357e-02 4.01961088e-01 3.21633875e-01 -7.43636668e-01
1.30018771e+00 -4.18212190e-02 -8.96011889e-01 -3.42981994e-01
3.76891375e-01 -7.52124071e-01 1.27486277e+00 6.34748638e-01
-1.40965652e+00 -3.50981086e-01 -1.15833449e+00 3.01041901e-01
-4.05453779e-02 -1.14704676e-01 8.13823938e-01 1.23623264e+00
-1.24637663e+00 5.08507907e-01 -5.02841055e-01 -2.30777919e-01
1.45527065e-01 5.12748480e-01 3.39973330e-01 9.10386667e-02
-1.28189611e+00 1.05758774e+00 6.14317238e-01 -6.95621148e-02
-9.74239707e-01 -7.42895246e-01 -7.31611013e-01 3.17215532e-01
6.43966138e-01 -5.28079510e-01 1.41325068e+00 -6.96549654e-01
-1.73120248e+00 3.93526316e-01 -2.47453481e-01 -5.67171633e-01
3.71835738e-01 -1.26045749e-01 -3.64482969e-01 -1.35024533e-01
-1.99652120e-01 5.37199199e-01 5.78974664e-01 -1.17727697e+00
-3.66015375e-01 -1.91271588e-01 -8.14096779e-02 2.44744256e-01
7.05055892e-02 3.76648426e-01 1.43039197e-01 -6.46620452e-01
-6.96204510e-03 -1.00747359e+00 -5.58018565e-01 -9.52251405e-02
-2.06000999e-01 4.96710390e-02 -8.80643725e-02 -2.18253985e-01
1.73481381e+00 -1.81709599e+00 -1.23997860e-01 5.15422523e-01
-9.54199061e-02 2.99315881e-02 -9.02377218e-02 3.14271629e-01
1.34239987e-01 2.55538881e-01 -3.11174572e-01 -2.91108768e-02
3.00166160e-01 -3.62581879e-01 -7.04663396e-02 3.09103072e-01
9.06131696e-03 6.30262673e-01 -6.41636074e-01 -4.34146494e-01
-4.18333188e-02 1.33972093e-01 -8.95028174e-01 3.30155015e-01
-1.11872353e-01 2.25557163e-01 -5.01593232e-01 4.82243985e-01
4.94066954e-01 -1.46249488e-01 5.88410914e-01 2.30170533e-01
-4.73128945e-01 3.69960099e-01 -1.42456686e+00 1.07577693e+00
-3.08511913e-01 4.36847240e-01 -1.65295139e-01 -9.90056157e-01
8.18171680e-01 5.77590540e-02 3.60464752e-02 -1.01285243e+00
4.97153312e-01 2.63189763e-01 5.09044111e-01 -7.89385438e-02
4.38203484e-01 -1.61696672e-01 -2.42269680e-01 6.14124060e-01
7.96100590e-03 -1.87712193e-01 3.97801518e-01 -2.65233010e-01
9.06455696e-01 4.97547165e-02 8.22743177e-01 -8.32688391e-01
2.80483305e-01 -1.58527508e-01 4.43028182e-01 1.36833036e+00
-2.40920290e-01 3.92252393e-02 2.57089019e-01 -5.15483208e-02
-7.91423321e-01 -1.09084249e+00 -3.12072724e-01 9.79997993e-01
2.06429541e-01 -1.91524759e-01 -8.56457412e-01 1.49026617e-01
-1.73598453e-01 1.41570508e+00 -4.44526881e-01 -5.04435778e-01
-4.62403774e-01 -8.20634782e-01 2.34088555e-01 1.98569477e-01
2.47319579e-01 -1.06168950e+00 -1.22435260e+00 3.67626160e-01
1.17237113e-01 -6.08895957e-01 -3.85322958e-01 5.76736748e-01
-1.10942984e+00 -8.33886981e-01 -3.64123136e-01 -3.34937245e-01
5.44769883e-01 2.42541984e-01 1.19413686e+00 6.72006980e-02
-4.57782239e-01 4.00509387e-01 -6.71872720e-02 -5.67961216e-01
-1.52169734e-01 -1.14771456e-01 2.35272527e-01 -2.79566169e-01
5.41468859e-01 -3.87188464e-01 -6.61952615e-01 3.97193313e-01
-8.93024743e-01 9.09273699e-03 6.86439633e-01 9.66638446e-01
5.53639710e-01 1.56243324e-01 8.48360419e-01 -7.67254531e-01
1.14179039e+00 -4.75685388e-01 -1.04882336e+00 3.69331390e-01
-1.12133932e+00 3.90777111e-01 3.50763023e-01 -7.51443803e-01
-1.25165021e+00 -1.99955404e-01 3.34560305e-01 1.76655620e-01
-2.19278678e-01 8.18881035e-01 -2.29324028e-01 -3.10276508e-01
7.84224749e-01 2.66771793e-01 -1.96887657e-01 -2.77765155e-01
5.61640300e-02 2.75484711e-01 -2.28577435e-01 -9.30561304e-01
4.69267547e-01 -3.61654222e-01 1.17704421e-01 -8.26310694e-01
-3.12132508e-01 1.84481189e-01 -2.45995715e-01 -2.19661295e-01
3.81482899e-01 -5.88819146e-01 -9.77303803e-01 8.73044282e-02
-6.18477166e-01 -4.34586048e-01 -1.32227167e-01 8.85548592e-01
-8.47820640e-01 -5.71127236e-02 -2.17493594e-01 -1.38954437e+00
-7.10786320e-04 -1.16824830e+00 5.96382976e-01 3.82884353e-01
-7.14641809e-01 -6.04612052e-01 -5.05885519e-02 -1.08803011e-01
4.98084038e-01 -8.94551724e-02 1.03147900e+00 -6.68181181e-01
-8.28060627e-01 -2.40081355e-01 2.74416864e-01 -3.79290730e-01
-3.24014962e-01 -4.37607676e-01 -7.77486682e-01 -3.81847948e-01
3.18603635e-01 -2.73286521e-01 3.96294773e-01 8.75396967e-01
1.15118849e+00 -1.51585340e-01 -3.52317184e-01 1.84305340e-01
1.22535884e+00 1.07883918e+00 5.04568398e-01 3.12339395e-01
-2.74191022e-01 5.96448898e-01 8.66581261e-01 7.60770023e-01
-1.00157700e-01 7.84565270e-01 -2.17498064e-01 3.84103090e-01
5.65483272e-01 -1.35034814e-01 1.39922470e-01 3.82998377e-01
8.63814205e-02 -5.40689290e-01 -7.48014629e-01 2.14936927e-01
-1.61814117e+00 -9.30119991e-01 4.33803380e-01 2.82186866e+00
9.25472379e-01 5.42607546e-01 1.49748415e-01 2.98880398e-01
8.06583941e-01 -3.28878164e-01 -6.44054294e-01 -3.00737470e-01
1.09338341e-02 4.06921953e-01 4.22356397e-01 5.33025742e-01
-7.56881475e-01 5.25113046e-01 8.05661011e+00 8.59279811e-01
-6.12843394e-01 1.81466863e-02 7.70492017e-01 -3.77077162e-01
-4.34161246e-01 2.74370044e-01 -8.98264170e-01 4.82656121e-01
1.41858232e+00 -7.92946637e-01 5.79177082e-01 5.77649474e-01
6.64334059e-01 -6.79607749e-01 -1.32891524e+00 9.30925906e-01
-1.59697264e-01 -1.06143153e+00 -3.47574204e-02 1.88916700e-03
3.18120927e-01 -9.94904995e-01 2.76270390e-01 3.15926224e-01
1.32260770e-01 -1.17680323e+00 1.11487401e+00 5.02306223e-01
4.49128926e-01 -8.54948103e-01 3.88456196e-01 4.60799456e-01
-9.47079659e-01 -3.74976486e-01 -2.86963344e-01 -4.86997813e-01
1.63882643e-01 4.92451847e-01 -5.91839075e-01 8.90486911e-02
3.37192684e-01 -3.03944379e-01 -5.15029848e-01 1.58343911e+00
5.74554466e-02 8.58938217e-01 -3.78590822e-01 -7.04117239e-01
-1.07432291e-01 -8.37569684e-02 3.84350777e-01 9.88077760e-01
6.15176916e-01 4.31122154e-01 1.38103470e-01 1.25791228e+00
3.83469343e-01 1.22950323e-01 -5.24536133e-01 -9.43652540e-02
1.20230794e+00 4.67579424e-01 -9.96339023e-01 -1.25056982e-01
1.20899767e-01 4.21521187e-01 1.22575447e-01 5.05479336e-01
-7.37638533e-01 -3.77902091e-01 2.23568648e-01 1.37609079e-01
-1.20813258e-01 -4.49230611e-01 -3.94813031e-01 -7.61888444e-01
-2.33441770e-01 -9.93144155e-01 3.12661707e-01 -4.92362231e-01
-8.87269497e-01 3.22028667e-01 7.27132678e-01 -9.75848913e-01
-4.19012308e-01 -4.95307893e-01 -2.04479322e-01 1.22739363e+00
-1.03438139e+00 -2.84464240e-01 1.81017861e-01 2.12989062e-01
6.03649974e-01 -1.82118982e-01 7.08578169e-01 2.02331305e-01
-4.53342646e-01 6.82069540e-01 4.62191291e-02 -8.40606630e-01
4.57886010e-01 -1.01229203e+00 1.16317712e-01 7.10871339e-01
-3.67039144e-01 9.84993100e-01 1.18427157e+00 -6.72957838e-01
-1.39355564e+00 -3.77660394e-01 8.77058029e-01 3.25876810e-02
3.30871552e-01 -5.91681302e-01 -6.98232710e-01 2.70340979e-01
-1.14698485e-02 -5.82892239e-01 9.39064920e-01 1.14222907e-01
2.27486819e-01 1.79221824e-01 -1.16655970e+00 1.16999137e+00
1.21030521e+00 -2.50867665e-01 -4.41731036e-01 4.35528718e-02
4.25335199e-01 -2.04364315e-01 -5.41736722e-01 1.42496973e-01
6.56015754e-01 -7.39746094e-01 9.34894919e-01 -4.72606331e-01
-3.92182887e-01 -2.26060897e-01 -1.07295632e-01 -1.37958777e+00
-6.09005868e-01 -9.97335315e-01 3.78796160e-02 8.76148283e-01
4.94509608e-01 -7.24754393e-01 4.20340627e-01 8.81630838e-01
2.18701452e-01 -3.71823639e-01 -9.72780526e-01 -9.02657330e-01
-7.81270787e-02 -4.22591597e-01 5.12644827e-01 5.42379558e-01
-6.79731593e-02 3.84674072e-01 -3.49642277e-01 -1.01629086e-01
9.63118374e-01 3.47305052e-02 1.99147671e-01 -1.53468966e+00
-4.43976134e-01 -7.38499701e-01 -1.97854042e-01 -9.26555395e-01
8.98440629e-02 -4.54892486e-01 2.43333131e-01 -8.32668304e-01
1.74331307e-01 -4.43206936e-01 -3.32094431e-01 -1.71705693e-01
-2.50712186e-01 -3.74313504e-01 8.41095895e-02 -2.13061333e-01
-2.51685709e-01 6.62649989e-01 7.96113193e-01 2.31342483e-02
-4.44913626e-01 6.21929877e-02 -9.58428741e-01 4.17969137e-01
8.95554304e-01 -5.95595598e-01 -1.03400731e+00 1.63199544e-01
3.23172539e-01 4.23135817e-01 2.14537889e-01 -1.15472984e+00
1.98242664e-01 -5.00251710e-01 5.02292454e-01 -4.41733867e-01
1.14933521e-01 -6.58594728e-01 5.58190048e-01 6.12704873e-01
-8.93775284e-01 2.96007156e-01 4.68373895e-01 6.17739916e-01
2.34196857e-01 -5.81348538e-01 7.94040918e-01 9.53192711e-02
-5.20955920e-01 -1.24770917e-01 -1.08000886e+00 8.68693143e-02
1.03254330e+00 -2.20890388e-01 -1.15985475e-01 -3.01578015e-01
-8.00729334e-01 1.13996319e-01 1.30986571e-01 2.36879498e-01
7.43222177e-01 -1.19820535e+00 -3.52558434e-01 1.00729220e-01
-5.95633462e-02 -7.28106260e-01 3.07210028e-01 7.00426817e-01
-2.03881472e-01 7.87483990e-01 -2.75420934e-01 -2.18344450e-01
-9.29637671e-01 6.98801637e-01 2.41776302e-01 -2.17701465e-01
-4.38844860e-02 4.55120265e-01 2.19115585e-01 3.38697769e-02
1.69564322e-01 -1.67022571e-01 -2.25330606e-01 -6.63893893e-02
8.14324498e-01 4.54288512e-01 -4.66769971e-02 -1.49689242e-02
-3.53429407e-01 9.43558961e-02 -8.08499102e-03 -6.25419497e-01
8.83325160e-01 -2.73960203e-01 3.45609367e-01 7.20272541e-01
4.77113515e-01 -4.12470758e-01 -1.00398445e+00 1.67768542e-02
3.95289987e-01 -7.19306409e-01 4.10793155e-01 -9.95674074e-01
-2.04894111e-01 5.50826728e-01 9.77425396e-01 2.09510371e-01
1.20428157e+00 -5.52442014e-01 -1.20974198e-01 6.28127337e-01
8.73421967e-01 -1.32083619e+00 5.20476066e-02 2.79063761e-01
8.28844845e-01 -8.87503505e-01 6.66952655e-02 -2.37818390e-01
-3.97810400e-01 7.72015810e-01 7.13849545e-01 4.30605784e-02
8.03566754e-01 3.57946098e-01 -4.75994736e-01 4.16661169e-06
-8.96477818e-01 -1.23512603e-01 1.40418723e-01 6.54088616e-01
6.75694108e-01 8.34904611e-02 -9.91606057e-01 9.12530005e-01
-4.20911238e-02 4.05142516e-01 5.54447055e-01 1.44096410e+00
-7.87297070e-01 -1.13326144e+00 -7.57677436e-01 6.40716791e-01
-2.77965277e-01 -3.09408486e-01 -6.67626783e-02 8.43097270e-01
-4.82395232e-01 1.22338009e+00 6.40144423e-02 -1.28490686e-01
3.88428926e-01 1.58993602e-01 6.23761654e-01 -4.06257629e-01
-4.10554409e-01 2.29719028e-01 2.52521753e-01 -4.98810828e-01
-2.17841640e-01 -8.09953749e-01 -7.12120712e-01 -2.40670994e-01
-4.72774714e-01 4.47321981e-01 2.96528995e-01 8.94452333e-01
5.09272158e-01 5.83179474e-01 5.08053303e-01 -9.48671937e-01
-8.88830781e-01 -8.20715725e-01 -9.15753722e-01 -2.92129964e-01
-1.35155037e-01 -1.39055705e+00 -2.91706830e-01 -3.12581182e-01] | [4.934492588043213, 3.081486463546753] |
2f916580-7be0-40fd-9b44-9be9a6b57eb3 | interpretable-learning-for-self-driving-cars | 1703.10631 | null | http://arxiv.org/abs/1703.10631v1 | http://arxiv.org/pdf/1703.10631v1.pdf | Interpretable Learning for Self-Driving Cars by Visualizing Causal Attention | Deep neural perception and control networks are likely to be a key component
of self-driving vehicles. These models need to be explainable - they should
provide easy-to-interpret rationales for their behavior - so that passengers,
insurance companies, law enforcement, developers etc., can understand what
triggered a particular behavior. Here we explore the use of visual
explanations. These explanations take the form of real-time highlighted regions
of an image that causally influence the network's output (steering control).
Our approach is two-stage. In the first stage, we use a visual attention model
to train a convolution network end-to-end from images to steering angle. The
attention model highlights image regions that potentially influence the
network's output. Some of these are true influences, but some are spurious. We
then apply a causal filtering step to determine which input regions actually
influence the output. This produces more succinct visual explanations and more
accurately exposes the network's behavior. We demonstrate the effectiveness of
our model on three datasets totaling 16 hours of driving. We first show that
training with attention does not degrade the performance of the end-to-end
network. Then we show that the network causally cues on a variety of features
that are used by humans while driving. | ['John Canny', 'Jinkyu Kim'] | 2017-03-30 | interpretable-learning-for-self-driving-cars-1 | http://openaccess.thecvf.com/content_iccv_2017/html/Kim_Interpretable_Learning_for_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Kim_Interpretable_Learning_for_ICCV_2017_paper.pdf | iccv-2017-10 | ['steering-control'] | ['computer-vision'] | [ 1.98898658e-01 5.31120479e-01 -1.39764488e-01 -7.70108402e-01
-3.13589373e-03 -5.14115870e-01 8.72801542e-01 -8.99853632e-02
-1.61943808e-02 5.82075536e-01 4.32082117e-01 -8.14324439e-01
1.57203376e-01 -5.77360153e-01 -1.28145647e+00 -3.61987025e-01
9.00356937e-03 -7.00532552e-03 4.60416824e-01 -3.58987153e-01
3.62221420e-01 5.94710469e-01 -1.97688580e+00 7.18612850e-01
3.38940978e-01 7.11135864e-01 9.09350738e-02 9.21119452e-01
1.46927357e-01 1.20601702e+00 -5.87689996e-01 -2.33340874e-01
-5.18181780e-03 -1.82053655e-01 -4.53885794e-01 -9.32884961e-03
7.61741042e-01 -3.52839172e-01 -3.72117370e-01 8.18243682e-01
-1.27879575e-01 -9.40854847e-02 6.70976520e-01 -1.77059221e+00
-7.96833277e-01 3.44889820e-01 -4.55923170e-01 5.60084581e-01
9.19134840e-02 8.12115192e-01 8.68673682e-01 -9.32840645e-01
5.94812691e-01 1.58104658e+00 2.57868111e-01 5.91374457e-01
-1.15904701e+00 -7.56376982e-01 5.45775235e-01 4.63260353e-01
-7.61180997e-01 -6.08300865e-01 6.34045184e-01 -6.68264389e-01
1.02150702e+00 3.78979325e-01 5.49234986e-01 1.09323347e+00
6.15441978e-01 6.90006733e-01 8.41473877e-01 -6.45022094e-02
7.07801580e-02 3.14801961e-01 1.41596273e-01 6.21948898e-01
2.56199807e-01 6.00772321e-01 -7.10163236e-01 2.77629137e-01
5.16116679e-01 1.58484932e-02 -1.58577606e-01 -3.07883978e-01
-1.00966644e+00 8.26322377e-01 8.42520833e-01 -1.13032378e-01
-5.66925287e-01 9.24576938e-01 1.72424838e-02 -9.95880142e-02
1.10310860e-01 2.61196136e-01 -2.51539052e-01 7.49490559e-02
-5.15062749e-01 4.05752987e-01 3.53306949e-01 7.47682571e-01
6.90527439e-01 1.35448292e-01 -2.97392994e-01 2.93309778e-01
5.69063962e-01 3.40670198e-01 -1.59942493e-01 -9.41972792e-01
3.62760097e-01 4.34750468e-01 3.35660875e-01 -1.07746613e+00
-2.77213305e-01 -3.94924164e-01 -3.85210127e-01 9.16697800e-01
3.75883430e-01 -2.34816417e-01 -1.17900383e+00 1.67046440e+00
1.01102240e-01 2.55901486e-01 -1.33954793e-01 1.24998939e+00
6.59778953e-01 5.92043519e-01 3.56985152e-01 3.69691581e-01
1.47734296e+00 -9.01965141e-01 -7.78974235e-01 -7.38417387e-01
1.74836084e-01 -5.25096118e-01 1.08067060e+00 1.94216341e-01
-9.50109661e-01 -8.77761543e-01 -1.32529593e+00 -8.31380710e-02
-4.86422807e-01 -2.48818239e-03 5.00912666e-01 3.34356189e-01
-9.98476028e-01 5.43712258e-01 -5.34161687e-01 -1.71929851e-01
6.86676264e-01 2.95078516e-01 -1.82800993e-01 1.01575918e-01
-1.34901690e+00 1.12980556e+00 8.13900083e-02 3.16703916e-02
-1.31273901e+00 -8.23018909e-01 -6.64372921e-01 2.19346300e-01
2.93547809e-01 -8.52648497e-01 1.51123190e+00 -1.25242388e+00
-8.44367087e-01 5.62174082e-01 -5.76334894e-01 -5.75156450e-01
5.05696774e-01 -3.27743709e-01 -4.65238959e-01 6.15567109e-03
2.52526879e-01 1.05150938e+00 1.02483118e+00 -1.66812778e+00
-1.12125504e+00 -8.74406099e-02 3.31328630e-01 -7.99324960e-02
2.83123225e-01 -1.66163936e-01 -4.39462662e-01 -2.18219355e-01
-2.65144378e-01 -9.74832714e-01 -2.67145395e-01 2.18837619e-01
-6.77428007e-01 -1.39160916e-01 1.20511043e+00 -2.53851563e-01
8.14418733e-01 -2.23381186e+00 -4.32223648e-01 2.65247941e-01
5.56904197e-01 3.49620543e-02 5.42380288e-02 2.01457426e-01
-6.12034857e-01 3.70288819e-01 1.45083666e-01 -2.27188349e-01
2.64244862e-02 2.88123161e-01 -8.27694714e-01 3.78330052e-01
8.01824868e-01 9.63305175e-01 -8.25224519e-01 2.06695800e-03
4.88643944e-01 5.51196992e-01 -3.92959386e-01 3.67653638e-01
-3.76561373e-01 2.85620660e-01 -2.38192052e-01 7.40144327e-02
4.70551997e-01 -1.07764974e-01 -3.33824337e-01 -1.15181074e-01
-4.07381207e-01 4.81538743e-01 -8.32675099e-01 9.05916095e-01
-3.56582582e-01 1.35819280e+00 -4.04161662e-01 -4.60508496e-01
5.73315024e-01 9.11200419e-02 -1.70361832e-01 -6.65560126e-01
9.24504325e-02 -1.39519498e-01 2.87938118e-01 -7.72681653e-01
4.19040918e-01 -9.55535844e-02 1.94905251e-01 3.62884641e-01
-6.07718468e-01 2.87026614e-01 5.75071685e-02 5.30902982e-01
9.56538141e-01 -1.42037659e-03 9.64073762e-02 -2.18762085e-01
2.51241654e-01 2.72903383e-01 3.26047331e-01 8.22216630e-01
-2.04303443e-01 5.32107115e-01 1.02075243e+00 -7.15792835e-01
-1.03828132e+00 -9.86891866e-01 3.09308261e-01 1.15398514e+00
3.52457941e-01 -1.23542659e-01 -6.06144965e-01 -7.34518528e-01
4.11108136e-01 1.58630323e+00 -1.24276221e+00 -3.75392318e-01
-3.65957409e-01 -5.51358156e-04 1.82545140e-01 7.73156583e-01
1.50523350e-01 -1.20187140e+00 -1.17221320e+00 -8.37799236e-02
1.17844410e-01 -9.16968107e-01 -4.90786821e-01 2.54359782e-01
-3.42135996e-01 -1.18436670e+00 -2.84291245e-03 -2.26553410e-01
8.94589245e-01 5.42354405e-01 1.09263372e+00 5.01106381e-01
-2.27545783e-01 1.24705590e-01 1.93075523e-01 -9.16471183e-01
-6.48435175e-01 -4.91584569e-01 -9.98569280e-02 1.40051827e-01
4.59023029e-01 -3.70234877e-01 -7.89460778e-01 3.23273659e-01
-5.92000723e-01 4.87956941e-01 6.60089672e-01 4.12957460e-01
2.37118870e-01 -1.37241021e-01 3.46197426e-01 -1.01102197e+00
6.24816716e-01 -5.76321542e-01 -6.00815952e-01 -7.94696733e-02
-6.69305444e-01 3.74108672e-01 5.91840088e-01 -2.81484723e-01
-1.00749528e+00 3.39374036e-01 1.52695417e-01 -6.91822052e-01
-5.83954096e-01 1.05246365e-01 -4.50639538e-02 4.99335676e-01
9.03782666e-01 -1.42932072e-01 -1.12448893e-01 5.40781543e-02
6.08250678e-01 3.54795963e-01 7.45526135e-01 1.73171479e-02
8.54133725e-01 6.63565338e-01 1.32158457e-04 -4.48296398e-01
-7.39217341e-01 -1.96443483e-01 -3.84314805e-01 -7.02063143e-01
9.97071981e-01 -7.01888025e-01 -1.03045332e+00 -2.69331366e-01
-1.51779830e+00 -2.97205895e-01 -1.16435274e-01 5.87705374e-02
-3.20224434e-01 -4.38765943e-01 1.10014742e-02 -1.01991498e+00
2.64826298e-01 -1.30038083e+00 9.30640638e-01 3.98180753e-01
-4.56145465e-01 -8.14634800e-01 -4.13634926e-01 8.68384466e-02
3.72741997e-01 2.90913373e-01 8.49047363e-01 -3.56868148e-01
-7.59037793e-01 4.98326942e-02 -5.56746840e-01 -4.96053211e-02
-2.90526878e-02 3.79306167e-01 -1.49650729e+00 1.57025769e-01
-5.31325042e-01 2.34117284e-01 1.00045514e+00 5.51024199e-01
1.33992362e+00 -3.44506979e-01 -6.32744849e-01 2.00907677e-01
1.17734706e+00 3.62237453e-01 8.45989168e-01 1.66858181e-01
7.16568410e-01 1.00459194e+00 5.55761814e-01 -2.71919947e-02
4.29101706e-01 5.54460645e-01 1.19672489e+00 -8.18105638e-01
-3.37554276e-01 -4.64939624e-01 5.24758875e-01 -4.66510028e-01
9.48915407e-02 -2.93035746e-01 -7.98777878e-01 7.01250792e-01
-2.17339039e+00 -1.23587966e+00 -8.52963150e-01 1.96830225e+00
2.27821767e-01 6.65409148e-01 6.86790496e-02 -1.04879856e-01
6.29162669e-01 -1.39118144e-02 -9.31878388e-01 -9.54460144e-01
1.97097912e-01 -2.05328494e-01 7.46033907e-01 7.91666865e-01
-8.81187856e-01 8.14877808e-01 6.45759201e+00 1.38110250e-01
-1.18737125e+00 -6.91922456e-02 9.06265795e-01 -2.57634223e-01
-6.22065604e-01 7.59019852e-02 -6.65006220e-01 2.73138553e-01
9.35612559e-01 2.98513211e-02 3.08152825e-01 9.53258276e-01
9.67203856e-01 -3.38585913e-01 -1.34256816e+00 5.47169387e-01
-2.08287001e-01 -1.66194475e+00 -1.05554484e-01 5.14033511e-02
4.66353953e-01 -2.96829361e-02 1.65828407e-01 2.79171895e-02
4.54019547e-01 -1.42384076e+00 1.43422914e+00 5.18915057e-01
6.80277109e-01 -8.16752195e-01 3.72939616e-01 1.95063055e-01
-8.38278234e-01 -1.79303497e-01 -1.52729318e-01 -3.55975419e-01
2.14721680e-01 5.01017570e-01 -1.27316833e+00 -1.35117427e-01
6.99823380e-01 5.02660215e-01 -5.97640336e-01 6.20379627e-01
-9.27200496e-01 6.79918706e-01 9.42847133e-02 -3.27737540e-01
3.87547016e-01 4.71125245e-01 5.00113845e-01 1.35613585e+00
-8.63334909e-02 4.97687869e-02 -3.40666294e-01 1.43411076e+00
1.69817954e-01 -5.45262218e-01 -1.05031133e+00 2.60412037e-01
2.03330845e-01 1.14699602e+00 -4.71449405e-01 -5.35746098e-01
-2.82099068e-01 7.86915421e-01 1.26892596e-01 5.76472282e-01
-1.22599006e+00 -2.71208018e-01 1.24156928e+00 6.06894255e-01
3.95638227e-01 6.46023974e-02 -5.31993806e-01 -4.62982714e-01
-1.63352609e-01 -6.73010945e-01 2.24862751e-02 -1.55375361e+00
-8.75180840e-01 4.49338645e-01 -1.71900272e-01 -9.18582201e-01
-2.97081679e-01 -7.33021557e-01 -8.78392875e-01 1.19817078e+00
-1.70763779e+00 -9.73618686e-01 -4.22973037e-01 4.00899738e-01
7.64970720e-01 8.88522640e-02 2.60144353e-01 -1.30524248e-01
-3.94374669e-01 2.57915258e-01 -7.33538091e-01 -1.86918557e-01
4.40608978e-01 -1.34357548e+00 8.12091112e-01 1.18520141e+00
1.16866566e-01 7.23777413e-01 1.35388315e+00 -6.37674153e-01
-1.17037272e+00 -1.23104537e+00 9.06107545e-01 -7.69904912e-01
5.03283441e-01 -3.12205166e-01 -7.78832138e-01 8.04554701e-01
7.09613919e-01 -8.84563923e-02 2.42345542e-01 6.53454214e-02
-3.77937883e-01 -1.11973852e-01 -7.41944134e-01 9.58196104e-01
8.24326456e-01 -3.61680239e-01 -6.15496755e-01 1.79117844e-01
6.27779484e-01 -3.99539560e-01 1.74545169e-01 1.99364033e-02
6.41498923e-01 -1.38451588e+00 7.50807583e-01 -1.05201018e+00
9.58969533e-01 -7.84624815e-01 3.03703368e-01 -1.41196680e+00
-5.02937555e-01 -4.69576210e-01 1.99648112e-01 6.77102685e-01
9.65883493e-01 -5.90936780e-01 4.74986434e-01 8.61099780e-01
-3.33506763e-01 -5.83250523e-01 -5.19098997e-01 -2.80658722e-01
-3.10221553e-01 -9.42968786e-01 6.55881166e-01 7.40001440e-01
-1.06763644e-02 5.45911372e-01 -2.29696497e-01 6.64475858e-01
4.35343832e-01 -1.26319557e-01 9.81364965e-01 -8.17669511e-01
-5.83399944e-02 -5.12208164e-01 -1.89754307e-01 -1.00669241e+00
-1.69424534e-01 -4.61595565e-01 3.13418657e-01 -1.65456116e+00
1.09895833e-01 -8.47557411e-02 -8.98941979e-02 8.33761156e-01
-4.02407020e-01 5.62095791e-02 1.86137468e-01 8.89364444e-03
-2.20495865e-01 -2.40225159e-02 1.23879623e+00 -6.73003942e-02
5.87322563e-02 -1.60880372e-01 -1.10165930e+00 6.57445967e-01
8.46630692e-01 -5.50489008e-01 -6.08821452e-01 -4.17642474e-01
2.66136676e-01 -1.50390774e-01 1.00641489e+00 -6.64776504e-01
1.92331120e-01 -4.62258875e-01 6.47328615e-01 -6.27951741e-01
1.54141903e-01 -7.97834694e-01 1.47393063e-01 6.96300149e-01
-5.81922650e-01 3.26425582e-01 5.35925686e-01 5.96204937e-01
6.61439523e-02 1.86232105e-01 6.53863251e-01 1.72228694e-01
-8.95685017e-01 4.24976125e-02 -7.10174561e-01 -4.76506621e-01
1.12070942e+00 -2.39351898e-01 -5.21216691e-01 -7.65347779e-01
-6.22108102e-01 4.31684583e-01 2.05457449e-01 9.33245718e-01
7.09710896e-01 -1.25316310e+00 -6.20125175e-01 2.63553947e-01
2.04029009e-01 -2.96283841e-01 1.78167298e-02 6.55708313e-01
-3.13887089e-01 6.01203740e-01 -5.21328971e-02 -7.83835053e-01
-1.24210763e+00 6.66295111e-01 5.01922786e-01 3.31965864e-01
-4.94349778e-01 8.35341871e-01 7.68492639e-01 2.84552425e-01
-3.77341872e-03 -5.74730456e-01 -4.09739047e-01 -2.64514446e-01
7.22861409e-01 -1.26542315e-01 -2.07727790e-01 -5.64802289e-01
-4.99695241e-01 8.71958658e-02 1.17783241e-01 -2.72178859e-01
1.14174080e+00 -2.05762442e-02 3.76530737e-01 4.94385958e-01
8.80992234e-01 -4.60850187e-02 -1.87636709e+00 3.81788909e-01
-1.98537409e-01 -6.30855501e-01 2.02758461e-01 -1.15924919e+00
-1.23984516e+00 1.25739682e+00 6.02965176e-01 4.69350100e-01
7.81894267e-01 1.59882493e-02 3.11390132e-01 -1.92576170e-01
-1.96971104e-01 -8.61495018e-01 -2.93689710e-03 1.35286897e-01
1.42738771e+00 -1.23928499e+00 -1.49015129e-01 -4.16242272e-01
-8.94437969e-01 1.24718952e+00 8.14716280e-01 -8.45108330e-02
3.28997105e-01 2.95321405e-01 4.08161402e-01 -6.36216581e-01
-1.38008964e+00 -3.57031107e-01 5.87052822e-01 4.91408825e-01
2.73780018e-01 1.25781521e-01 2.19129249e-01 3.71089935e-01
-1.30836368e-01 -6.55590147e-02 6.30240083e-01 4.03451443e-01
-5.50068140e-01 -4.63428438e-01 -5.09432435e-01 2.87639648e-01
3.40573303e-03 1.13311717e-02 -6.58573151e-01 9.14688051e-01
3.78682137e-01 1.30601501e+00 2.85143167e-01 -4.36863452e-01
7.04037368e-01 -3.56920846e-02 -3.48486602e-02 -4.71050590e-01
-4.57451969e-01 -2.37362161e-01 3.52927566e-01 -8.69602621e-01
-3.35876122e-02 -7.22120166e-01 -1.55374014e+00 -3.66188139e-01
5.22223413e-02 -3.35698009e-01 8.80215347e-01 1.03540957e+00
5.00544846e-01 1.10907137e+00 6.20336354e-01 -7.12096751e-01
3.29865143e-02 -5.83594084e-01 6.69153556e-02 5.24856508e-01
8.15485060e-01 -7.45783269e-01 -4.53471601e-01 4.63639915e-01] | [8.924098014831543, 5.266473770141602] |
24d44c1c-f5eb-498a-9a0e-e8e0cf7b6b60 | enrich4all-a-first-luxembourgish-bert-model | null | null | https://aclanthology.org/2022.sigul-1.27 | https://aclanthology.org/2022.sigul-1.27.pdf | ENRICH4ALL: A First Luxembourgish BERT Model for a Multilingual Chatbot | Machine Translation (MT)-empowered chatbots are not established yet, however, we see an amazing future breaking language barriers and enabling conversation in multiple languages without time-consuming language model building and training, particularly for under-resourced languages. In this paper we focus on the under-resourced Luxembourgish language. This article describes the experiments we have done with a dataset containing administrative questions that we have manually created to offer BERT QA capabilities to a multilingual chatbot. The chatbot supports visual dialog flow diagram creation (through an interface called BotStudio) in which a dialog node manages the user question at a specific step. Dialog nodes can be matched to the user’s question by using a BERT classification model which labels the question with a dialog node label. | ['Dimitra Anastasiou'] | null | null | null | null | sigul-lrec-2022-6 | ['visual-dialogue', 'visual-dialogue'] | ['computer-vision', 'natural-language-processing'] | [-4.36706066e-01 5.67144394e-01 5.92518561e-02 -4.54240024e-01
-8.62152100e-01 -9.17719662e-01 8.71216118e-01 -1.75433636e-01
-4.72983420e-01 9.50433850e-01 2.59505987e-01 -9.18800235e-01
2.06355885e-01 -4.08778727e-01 1.43207327e-01 -1.15592621e-01
3.84524792e-01 1.36271691e+00 1.82880118e-01 -7.53388703e-01
-3.08168717e-02 2.05621019e-01 -5.20488858e-01 6.18446946e-01
8.40136409e-01 8.28792229e-02 6.59581363e-01 9.30931032e-01
-7.22339511e-01 1.37403214e+00 -9.65029776e-01 -6.05308533e-01
1.80617064e-01 -5.73434353e-01 -1.70194793e+00 -2.01090232e-01
6.65968806e-02 -3.88446242e-01 -3.56950164e-02 5.04429042e-01
1.92528814e-01 2.10630931e-02 3.18571657e-01 -1.35722268e+00
-2.70170033e-01 7.19496965e-01 3.57918113e-01 -1.40385553e-01
7.15591311e-01 3.91332626e-01 1.01651251e+00 -5.00507593e-01
1.26726532e+00 1.55933702e+00 5.16569138e-01 8.49775016e-01
-1.12357354e+00 -3.70370954e-01 -4.34695780e-01 1.60510510e-01
-9.57818270e-01 -2.97876775e-01 3.68944734e-01 -6.75126553e-01
1.17221797e+00 2.55157053e-01 2.15910539e-01 8.98186147e-01
1.91380709e-01 1.99971125e-01 1.21858084e+00 -7.03775227e-01
-2.26492509e-01 9.91458952e-01 5.08427992e-02 6.79214656e-01
-7.17661440e-01 -4.07299161e-01 -2.87252873e-01 -2.26596147e-01
6.65603280e-01 -4.92930204e-01 3.46919447e-01 1.20649457e-01
-9.48531091e-01 1.12590754e+00 2.30901748e-01 8.56069028e-01
-1.21470787e-01 -1.70073047e-01 4.11073864e-01 9.35598791e-01
3.62463683e-01 7.12422013e-01 -6.15680099e-01 -6.94932759e-01
-3.39816153e-01 3.46816182e-02 1.64760613e+00 1.12990987e+00
9.25572395e-01 -4.05145437e-01 -1.55947998e-01 9.68404472e-01
3.71498615e-01 1.02442637e-01 1.43720761e-01 -1.28253067e+00
7.12469518e-01 1.01432335e+00 4.71494079e-01 -3.26782107e-01
-3.77095431e-01 4.03189719e-01 -2.49453053e-01 3.92961025e-01
1.08558547e+00 -6.50478661e-01 -1.97926074e-01 1.27786314e+00
3.93142045e-01 -8.23718905e-01 3.02731663e-01 5.29563546e-01
1.04683316e+00 5.89825749e-01 2.79079556e-01 -8.95206407e-02
1.64838362e+00 -1.05580115e+00 -9.44794655e-01 3.12142279e-02
1.10328841e+00 -1.20649838e+00 1.33600569e+00 -7.87553843e-03
-9.12037611e-01 -3.59300733e-01 -3.17320287e-01 -3.90507013e-01
-5.28456807e-01 1.02061085e-01 5.62217593e-01 6.80747926e-01
-1.24357057e+00 1.00841217e-01 -4.86692995e-01 -9.48746204e-01
-2.57035404e-01 2.88471460e-01 -5.21367073e-01 1.35414213e-01
-1.29401600e+00 1.68622863e+00 -2.56330031e-03 -2.23800614e-01
-7.13050008e-01 -4.36997503e-01 -6.87512279e-01 -7.27422461e-02
2.67655790e-01 -4.32069659e-01 1.82937717e+00 -7.44661033e-01
-1.93107820e+00 1.14023888e+00 -1.61086425e-01 -4.55742508e-01
8.31095457e-01 1.33381367e-01 -2.29210615e-01 -4.84475447e-03
3.17955643e-01 7.43356764e-01 8.37615505e-02 -1.15858233e+00
-7.07566202e-01 -8.19041580e-02 8.04386854e-01 2.05911249e-01
3.46800545e-03 9.16181982e-01 -1.21202869e-02 1.50501400e-01
-4.30375367e-01 -7.06242800e-01 -2.11520538e-01 -2.63092369e-01
-1.82160184e-01 -4.46242183e-01 8.41268003e-01 -1.13428223e+00
1.08507490e+00 -1.56966352e+00 -1.55617833e-01 -3.79066795e-01
1.60230041e-01 3.28073740e-01 -8.32068175e-03 1.39454722e+00
3.63144130e-01 1.06894121e-01 1.77491516e-01 -3.83639276e-01
3.28431696e-01 2.79822737e-01 4.10898887e-02 -1.59227401e-02
-2.51788683e-02 7.15911269e-01 -8.02185714e-01 -9.42074597e-01
3.21883172e-01 1.80289105e-01 -4.35926795e-01 7.39330649e-01
-5.17413318e-01 8.78390789e-01 -3.32655758e-01 2.75817871e-01
2.84978896e-01 -1.66155566e-02 6.65972948e-01 4.21645164e-01
-7.56991088e-01 8.29590142e-01 -6.36345565e-01 1.77654350e+00
-1.21754920e+00 8.03992808e-01 6.88945651e-01 -4.87060726e-01
1.28604674e+00 8.12540472e-01 2.10083783e-01 -2.31721163e-01
1.88426554e-01 2.47339800e-01 2.79715359e-01 -8.30603540e-01
4.84216869e-01 -2.22567230e-01 -1.62363261e-01 7.39860654e-01
2.98557311e-01 -2.24124238e-01 2.40765497e-01 4.41064596e-01
9.64403689e-01 1.56480595e-01 3.06965142e-01 -2.40563124e-01
8.02837312e-01 5.98512292e-01 8.85398462e-02 5.39078891e-01
-4.17496681e-01 -2.58953065e-01 8.04236174e-01 -3.09730381e-01
-9.83343005e-01 -6.38614237e-01 1.59254372e-02 1.39397848e+00
-4.77659374e-01 -4.59019244e-01 -9.25438821e-01 -8.31330478e-01
-6.41354084e-01 8.15373957e-01 -3.62933762e-02 5.05009353e-01
-6.56089425e-01 1.23571120e-01 7.43039370e-01 -1.35844797e-01
6.24994576e-01 -1.28236079e+00 -4.25181240e-01 4.44839329e-01
-5.61999619e-01 -1.28228211e+00 -3.53118479e-01 -2.03933381e-02
-4.82531995e-01 -8.71994317e-01 -5.61116874e-01 -1.20437372e+00
2.31183231e-01 -1.48664176e-01 1.10163879e+00 5.04879793e-03
-1.42939553e-01 3.30503374e-01 -5.27011514e-01 -3.60226125e-01
-1.36950052e+00 6.02463365e-01 -4.32297379e-01 -5.20309329e-01
5.32297194e-01 -3.51660550e-01 -1.13219127e-01 5.80454826e-01
-3.10692072e-01 2.22387522e-01 1.96616650e-01 7.52678812e-01
-5.70371151e-01 -6.61754549e-01 8.63645732e-01 -1.00877237e+00
8.63254905e-01 -3.80139172e-01 -5.84900022e-01 5.35072267e-01
-1.89634383e-01 -1.42741084e-01 6.35524631e-01 -1.67989105e-01
-1.58213544e+00 6.85555711e-02 -4.53348398e-01 3.06016564e-01
-5.22591352e-01 3.54000330e-01 -2.62273371e-01 4.54735197e-02
6.45428002e-01 -3.21807176e-01 3.20501536e-01 -8.31328988e-01
6.10315144e-01 1.36480069e+00 2.01967180e-01 -6.32318437e-01
5.84570348e-01 6.92023858e-02 -5.45866311e-01 -8.82932842e-01
-1.98835880e-01 -7.16758132e-01 -1.00635779e+00 -7.18117952e-01
1.10841393e+00 -5.42354703e-01 -1.28592360e+00 9.77302492e-02
-1.67019355e+00 -8.35892081e-01 -2.34709978e-02 3.24322760e-01
-3.48232478e-01 1.52522713e-01 -9.53619421e-01 -1.21675014e+00
-3.20159107e-01 -9.94427800e-01 5.96400380e-01 2.13034108e-01
-7.14200974e-01 -1.34370768e+00 3.11718851e-01 9.39349413e-01
5.34259856e-01 -2.31381759e-01 1.17467391e+00 -1.03164744e+00
-4.72347468e-01 -1.38341472e-01 -1.97830603e-01 1.93494052e-01
1.55605868e-01 -9.77778733e-02 -8.67718518e-01 1.22859679e-01
8.25602040e-02 -5.64547777e-01 -2.99777538e-01 -3.64280283e-01
-1.83264792e-01 -6.67235136e-01 -6.29226565e-02 -2.93361515e-01
9.89061534e-01 5.29561222e-01 5.02709448e-01 4.65226442e-01
3.00248891e-01 1.26359999e+00 6.30225897e-01 1.92566495e-02
7.31139183e-01 9.34713602e-01 -1.57673750e-02 -1.56721607e-01
1.58610884e-02 -2.70129740e-01 4.43432540e-01 7.72385657e-01
1.10648848e-01 -1.88506767e-02 -1.20165598e+00 4.08004671e-01
-1.78041327e+00 -8.31721902e-01 -6.07828379e-01 1.82274175e+00
1.10982955e+00 -2.51628131e-01 5.28496683e-01 -4.84742075e-01
6.63032115e-01 -1.58851162e-01 1.14161596e-01 -9.15759921e-01
3.53753150e-01 1.96266010e-01 3.91688906e-02 1.30248725e+00
-6.11175835e-01 1.48605096e+00 5.64718962e+00 5.11523068e-01
-6.68421984e-01 6.36376798e-01 2.82034308e-01 4.31900501e-01
7.37642050e-02 5.98775804e-01 -7.37793148e-01 -4.87832539e-02
1.14774776e+00 -1.06732875e-01 6.12652898e-01 8.13593328e-01
7.20398545e-01 -2.38903880e-01 -9.42095637e-01 3.89677107e-01
-3.05717289e-01 -1.33968127e+00 -4.54357177e-01 1.08761579e-01
2.11202681e-01 -5.16490228e-02 -6.07810736e-01 6.20062947e-01
9.21035171e-01 -9.01857257e-01 1.53123915e-01 2.94251800e-01
6.91081047e-01 -2.75243253e-01 7.24862933e-01 8.02319944e-01
-9.25351143e-01 9.52190980e-02 -2.54026085e-01 -3.88328522e-01
3.91808987e-01 -2.59190410e-01 -1.66350329e+00 5.51293314e-01
3.75965774e-01 5.51979467e-02 -3.46585631e-01 5.95099568e-01
-2.19722435e-01 6.09026730e-01 -4.85221017e-03 -4.56568182e-01
4.28814352e-01 -7.42623329e-01 5.36539972e-01 1.30740643e+00
-8.31667185e-02 -6.23485595e-02 3.29979181e-01 8.35729003e-01
3.96219455e-02 4.91079301e-01 -6.82064772e-01 2.29608044e-02
4.28037047e-01 1.60983181e+00 -4.76980269e-01 -3.20754260e-01
-3.43180120e-01 8.97449374e-01 2.51868725e-01 9.11862776e-02
-4.44091499e-01 -4.06136274e-01 5.21576047e-01 -1.63348336e-02
-4.32153404e-01 -2.61774451e-01 -1.93297103e-01 -8.32784712e-01
-3.72300535e-01 -1.20981646e+00 4.67922419e-01 -9.20554101e-01
-1.13214111e+00 7.89799690e-01 -6.63063899e-02 -7.49970317e-01
-6.82600141e-01 -5.51740587e-01 -8.07229161e-01 1.27161157e+00
-8.95956099e-01 -1.43703735e+00 -4.07105945e-02 4.91974592e-01
7.24965155e-01 -2.19159052e-01 1.35821557e+00 5.18056512e-01
-2.50472158e-01 3.50169539e-01 -1.24729536e-01 4.09796685e-01
1.02571273e+00 -1.29924583e+00 2.24661082e-01 1.14472076e-01
-2.38621626e-02 6.46495938e-01 8.83382022e-01 -5.17943323e-01
-1.01069224e+00 -3.97598416e-01 1.87109697e+00 -8.50172162e-01
1.02553105e+00 -8.11521173e-01 -7.65771627e-01 8.01900625e-01
1.05200756e+00 -1.22355866e+00 5.99119365e-01 1.47549793e-01
-2.05918968e-01 2.02122822e-01 -1.31344092e+00 6.20234728e-01
5.70127130e-01 -7.62192249e-01 -8.32073331e-01 8.77338111e-01
6.72694325e-01 -3.22966009e-01 -9.62826550e-01 -4.33232099e-01
3.68442297e-01 -6.87931776e-01 5.42082787e-01 -9.94781971e-01
1.71631068e-01 -5.06425044e-03 1.71847522e-01 -1.09882236e+00
3.05976540e-01 -1.27823198e+00 7.77266443e-01 1.73440385e+00
7.56174028e-01 -8.43139291e-01 5.88889062e-01 9.51576829e-01
-4.99162190e-02 9.79907215e-02 -1.19478881e+00 -7.04935491e-01
3.34798813e-01 -2.45060459e-01 4.95015740e-01 1.02660108e+00
9.20264602e-01 1.09192932e+00 -4.28744912e-01 -2.72415340e-01
-1.01680800e-01 -1.57958150e-01 1.19709611e+00 -1.23715460e+00
-1.67024031e-01 -3.03625584e-01 2.10691020e-01 -9.25762653e-01
1.61163211e-01 -8.86342049e-01 2.05705792e-01 -1.91377592e+00
-1.57865256e-01 -4.32634681e-01 8.10829520e-01 4.90076631e-01
1.54681593e-01 -3.21038634e-01 4.26292360e-01 3.07159841e-01
-3.48548353e-01 9.17152986e-02 1.24120653e+00 5.61670177e-02
-4.07083839e-01 2.80273199e-01 -3.23477983e-01 5.78749120e-01
1.00359154e+00 -5.78202248e-01 -1.84333995e-01 -1.92371801e-01
-2.69982275e-02 6.18757606e-01 1.17907256e-01 -5.30059516e-01
2.68986970e-01 -2.25349516e-01 -4.74985003e-01 -1.37871906e-01
4.32769954e-01 -8.72264802e-01 3.67531702e-02 6.08215690e-01
-6.54187381e-01 2.10600242e-01 2.15810686e-02 -2.25173786e-01
-1.21166393e-01 -7.01260507e-01 7.33308136e-01 -4.48909491e-01
-2.20008850e-01 -2.53234118e-01 -1.20793128e+00 -4.45655873e-03
9.06471014e-01 1.24707788e-01 -6.13496065e-01 -1.11051667e+00
-6.81038141e-01 4.23818797e-01 3.55234712e-01 3.56726944e-01
-1.74374014e-01 -7.12849855e-01 -7.30343819e-01 -3.58336061e-01
-1.76997092e-02 -5.95699370e-01 2.01258257e-01 6.56688273e-01
-8.94715309e-01 7.19560444e-01 -3.33140016e-01 -2.16656461e-01
-1.57869041e+00 1.65797830e-01 4.85982388e-01 -8.25983942e-01
-3.47124666e-01 5.21677554e-01 -2.64353722e-01 -1.19065607e+00
1.47539631e-01 1.53098013e-02 -5.79910457e-01 2.34236106e-01
2.87325323e-01 5.17913997e-01 -1.23797365e-01 -6.68306649e-01
-2.09006369e-01 -1.42377362e-01 8.82031843e-02 -8.47480178e-01
1.01783442e+00 -2.09805861e-01 -3.82762581e-01 6.21823430e-01
8.53337824e-01 -4.64842357e-02 -7.50229180e-01 -7.61950612e-02
3.73734355e-01 -2.34910712e-01 -7.05644071e-01 -1.20162868e+00
-1.50368690e-01 1.06317949e+00 3.75839531e-01 6.89535379e-01
2.84045756e-01 1.54805854e-01 6.37811482e-01 7.95782983e-01
4.76828873e-01 -1.12293184e+00 4.21851277e-02 1.03828728e+00
9.13015425e-01 -1.26904011e+00 -6.15970552e-01 -3.79294723e-01
-1.15394187e+00 1.11958110e+00 7.68545091e-01 3.88458669e-01
3.40343982e-01 2.61576682e-01 1.08368504e+00 -2.81487852e-01
-9.10788417e-01 -3.16345602e-01 -3.73194903e-01 9.06579137e-01
7.91162312e-01 7.75245056e-02 -6.42910182e-01 2.60867298e-01
-6.01567686e-01 -1.24934614e-01 6.89087093e-01 8.66645932e-01
-4.46797967e-01 -1.77726150e+00 -4.00538385e-01 6.13523610e-02
-4.53866959e-01 -1.50113106e-01 -1.16366386e+00 1.28873026e+00
-1.43195957e-01 1.65082753e+00 -4.81215604e-02 -3.15812230e-01
3.53238821e-01 6.75858140e-01 1.89101733e-02 -1.14934313e+00
-1.56683326e+00 -5.63234240e-02 9.29084361e-01 -3.98377888e-03
-2.11438701e-01 -5.28970182e-01 -1.14666498e+00 -5.33153415e-01
-3.10802996e-01 7.47526467e-01 9.17598724e-01 1.01760590e+00
7.38811791e-02 3.72563526e-02 5.55274427e-01 -3.77249837e-01
-4.76661831e-01 -1.61607707e+00 -7.68125430e-02 7.39239082e-02
-1.95111245e-01 -1.14192225e-01 -1.41386569e-01 1.33163944e-01] | [12.690877914428711, 7.96993350982666] |
3ca6ddc7-fc81-4e4b-a439-22d44c3a1bb4 | unitrec-a-unified-text-to-text-transformer | 2305.15756 | null | https://arxiv.org/abs/2305.15756v1 | https://arxiv.org/pdf/2305.15756v1.pdf | UniTRec: A Unified Text-to-Text Transformer and Joint Contrastive Learning Framework for Text-based Recommendation | Prior study has shown that pretrained language models (PLM) can boost the performance of text-based recommendation. In contrast to previous works that either use PLM to encode user history as a whole input text, or impose an additional aggregation network to fuse multi-turn history representations, we propose a unified local- and global-attention Transformer encoder to better model two-level contexts of user history. Moreover, conditioned on user history encoded by Transformer encoders, our framework leverages Transformer decoders to estimate the language perplexity of candidate text items, which can serve as a straightforward yet significant contrastive signal for user-item text matching. Based on this, our framework, UniTRec, unifies the contrastive objectives of discriminative matching scores and candidate text perplexity to jointly enhance text-based recommendation. Extensive evaluation shows that UniTRec delivers SOTA performance on three text-based recommendation tasks. Code is available at https://github.com/Veason-silverbullet/UniTRec. | ['Kam-Fai Wong', 'Yiming Du', 'Huimin Wang', 'Zhiming Mao'] | 2023-05-25 | null | null | null | null | ['text-matching'] | ['natural-language-processing'] | [ 1.12946056e-01 -3.39801520e-01 -6.20894551e-01 -4.51801479e-01
-9.33578849e-01 -5.14311433e-01 9.04050946e-01 -1.96344536e-02
-4.63192642e-01 2.20694020e-01 9.28861260e-01 -5.49795389e-01
1.33878797e-01 -5.63533485e-01 -6.18328273e-01 -1.92493454e-01
2.62375861e-01 4.33745921e-01 -5.72570413e-02 -2.11898118e-01
8.15474391e-02 -5.24467975e-02 -1.25029969e+00 8.09252739e-01
8.42523694e-01 9.38776791e-01 5.40692627e-01 7.42123783e-01
-1.46976635e-01 9.07761514e-01 -1.79344311e-01 -5.67340136e-01
1.79598212e-01 -5.65734208e-01 -5.66450119e-01 -7.06850663e-02
4.98612702e-01 -8.27505648e-01 -9.26121712e-01 6.78788602e-01
5.74108720e-01 4.56080854e-01 6.66370571e-01 -6.31439805e-01
-1.04458272e+00 1.35161674e+00 -3.15244526e-01 2.76328117e-01
3.92638564e-01 -1.29295960e-01 1.83462715e+00 -1.10995126e+00
4.32478815e-01 1.18488061e+00 4.44886088e-01 3.15721512e-01
-1.02790487e+00 -5.89751601e-01 5.95360458e-01 7.14258775e-02
-1.07216787e+00 -4.93330151e-01 4.80701983e-01 -1.87493101e-01
1.17862892e+00 2.42247984e-01 4.79410738e-01 1.58407640e+00
4.81420755e-01 1.47152483e+00 3.58122975e-01 -2.78953701e-01
-2.09238395e-01 -7.84721449e-02 3.88456017e-01 6.16394401e-01
1.59155101e-01 -5.04986942e-02 -6.18107200e-01 -4.60034236e-02
6.52544081e-01 5.22290051e-01 -1.71393499e-01 2.24605948e-01
-1.03122258e+00 8.77694249e-01 1.90402985e-01 2.96168685e-01
-3.51475775e-01 2.01478586e-01 4.21293020e-01 3.65441263e-01
6.53450489e-01 3.15396369e-01 -5.62773407e-01 -9.76341441e-02
-1.09135389e+00 3.31344344e-02 7.37486780e-01 9.12371039e-01
5.19208968e-01 3.71225365e-02 -7.23545969e-01 1.05167985e+00
6.54410124e-01 6.33785367e-01 7.08406031e-01 -4.32666659e-01
6.20789766e-01 4.07734245e-01 -1.32131681e-01 -6.23013198e-01
-8.43159705e-02 -8.27063620e-01 -6.39584720e-01 -7.26620615e-01
8.13584626e-02 -1.67874023e-01 -9.50939059e-01 1.51808524e+00
-1.61453471e-01 1.48985326e-01 -3.33036810e-01 7.98257291e-01
7.91375399e-01 9.64676321e-01 -5.59699833e-02 1.02071045e-02
1.31048441e+00 -1.37304556e+00 -6.18954718e-01 -4.36607808e-01
7.06508994e-01 -9.31761205e-01 1.30878997e+00 3.59640628e-01
-1.11818850e+00 -4.84805048e-01 -9.35866714e-01 -5.33047199e-01
-4.51291800e-02 6.58437014e-01 6.00813329e-01 4.56542999e-01
-8.73309910e-01 5.64045787e-01 -9.77132916e-01 -2.89344370e-01
8.93318504e-02 1.24289565e-01 2.07441106e-01 -1.70381531e-01
-1.23343396e+00 4.96374607e-01 2.12946445e-01 2.18365211e-02
-7.94246972e-01 -7.34739542e-01 -7.04363823e-01 3.98787349e-01
3.25456172e-01 -7.93062687e-01 1.62009299e+00 -8.93778920e-01
-1.62793350e+00 2.59118676e-01 -4.76510316e-01 -4.67171043e-01
2.17613593e-01 -5.48142195e-01 -4.54996049e-01 -2.02325538e-01
-1.50667563e-01 3.52020770e-01 6.35121882e-01 -7.04797447e-01
-6.14519179e-01 -1.43267348e-01 7.11226463e-02 3.11827004e-01
-7.35925436e-01 1.92718998e-01 -1.07782245e+00 -1.15876234e+00
-2.09298134e-01 -7.19263196e-01 -1.43596545e-01 -3.50721568e-01
-2.53321707e-01 -4.20374960e-01 3.73181462e-01 -9.78479505e-01
1.96968412e+00 -2.03006864e+00 1.46852553e-01 1.20374724e-01
9.44015384e-02 3.20893452e-02 -4.91850704e-01 8.04445446e-01
3.22339058e-01 1.29863575e-01 2.45890066e-01 -7.88776338e-01
3.77392322e-01 1.24830164e-01 -6.69446588e-01 1.21368125e-01
-1.55944720e-01 1.20941269e+00 -7.46364713e-01 -1.96166515e-01
6.09976165e-02 6.96617961e-01 -9.25167739e-01 1.88450575e-01
-5.49442470e-01 7.90396035e-02 -5.60620606e-01 3.07635993e-01
2.36623645e-01 -7.21562922e-01 4.91580456e-01 -1.04284637e-01
2.01407537e-01 1.17439282e+00 -7.59675920e-01 1.86365604e+00
-6.68707192e-01 5.97708166e-01 -1.96025535e-01 -4.85855550e-01
4.63691443e-01 3.68966281e-01 5.39503872e-01 -1.06246567e+00
2.90670693e-01 -1.49579234e-02 -2.49594018e-01 -7.10501820e-02
9.92668390e-01 2.96046674e-01 -1.15305915e-01 1.03421581e+00
3.11975092e-01 4.99011576e-01 2.20304474e-01 5.68920732e-01
1.00227463e+00 2.43152842e-01 4.52558361e-02 4.82036024e-02
3.05173546e-01 -6.43550277e-01 3.26643229e-01 9.80720103e-01
3.34684730e-01 5.26463449e-01 9.54031795e-02 1.83008350e-02
-1.10106683e+00 -8.08445752e-01 4.09474671e-02 1.81495702e+00
-5.59129752e-02 -1.03732061e+00 -1.58920631e-01 -8.42105806e-01
-2.33449461e-03 8.14029694e-01 -4.49057579e-01 -1.84563309e-01
-7.29336917e-01 -5.99863708e-01 3.31028521e-01 5.89867175e-01
-1.41813224e-02 -6.67431533e-01 1.47003889e-01 4.38633949e-01
-4.04080957e-01 -9.08612370e-01 -1.33671463e+00 4.35519516e-02
-8.68129909e-01 -5.05710959e-01 -7.71303236e-01 -5.26569366e-01
3.36791188e-01 6.60141289e-01 1.22533226e+00 1.85606033e-01
1.41749918e-01 6.00616336e-01 -8.40996802e-01 -5.01839444e-02
-3.34129155e-01 3.74839097e-01 -1.30993411e-01 1.11202441e-01
3.26448590e-01 -6.28127337e-01 -8.84760022e-01 2.45187551e-01
-8.80377412e-01 2.72592396e-01 6.92481875e-01 7.82928586e-01
3.49420935e-01 -3.64584446e-01 3.84043068e-01 -9.49172735e-01
6.77831113e-01 -7.73396075e-01 -3.33914310e-01 5.12816668e-01
-9.21950459e-01 2.26367012e-01 7.29249597e-01 -5.74099898e-01
-1.18808222e+00 -3.54344845e-01 -5.09478152e-01 -1.94957450e-01
3.33376765e-01 8.75092089e-01 1.25888094e-01 6.07751071e-01
4.79912788e-01 4.06599730e-01 -5.77614725e-01 -9.41073000e-01
5.57595670e-01 8.40914249e-01 1.79909602e-01 -6.40843034e-01
3.65415156e-01 1.54681951e-01 -7.54407048e-01 -3.78611624e-01
-1.10133255e+00 -8.51870060e-01 -3.73492539e-01 -2.09767222e-02
5.59644759e-01 -1.22546136e+00 -3.96645427e-01 2.31614187e-01
-9.56193864e-01 -5.65055490e-01 1.20957710e-01 6.31887913e-01
-9.80556607e-02 5.99275231e-01 -1.01276863e+00 -7.21075237e-01
-7.97221005e-01 -9.76320446e-01 1.16263402e+00 -1.41756773e-01
-2.16306504e-02 -1.18922961e+00 -4.31810915e-02 3.50711405e-01
3.48238856e-01 -7.21486926e-01 8.37514341e-01 -8.89952123e-01
-4.08427179e-01 -3.18426967e-01 -1.58042803e-01 2.11485863e-01
-7.11879134e-02 -3.06615114e-01 -8.39737654e-01 -6.31330729e-01
-3.78534853e-01 -1.95415586e-01 1.24586594e+00 2.84023821e-01
1.14551747e+00 -5.82402587e-01 -2.48030767e-01 5.20689189e-01
1.14004433e+00 3.62494262e-03 4.82036501e-01 -1.48539841e-01
7.95938373e-01 1.23056835e-02 3.42907637e-01 8.32218766e-01
6.51170194e-01 8.37740541e-01 -1.09832719e-01 1.65880188e-01
-4.61515814e-01 -6.76486254e-01 1.00119328e+00 1.42344141e+00
1.09651558e-01 -9.77314353e-01 -4.43436027e-01 2.09521383e-01
-1.91523075e+00 -1.01796508e+00 1.31423429e-01 2.07423854e+00
8.47446442e-01 4.32931520e-02 8.37013349e-02 -3.03191990e-01
4.08152789e-01 3.22792411e-01 -5.74884415e-01 -7.41541311e-02
5.47385551e-02 2.13547602e-01 4.84637082e-01 7.04665482e-01
-8.43927503e-01 9.22512174e-01 6.16629028e+00 9.87280786e-01
-1.24201024e+00 4.20471191e-01 2.40069509e-01 -5.67563832e-01
-8.77148032e-01 -9.52613205e-02 -1.08639729e+00 6.49076164e-01
1.21203196e+00 -2.68464863e-01 5.96009433e-01 5.68101048e-01
3.21929932e-01 3.16728920e-01 -1.20759916e+00 7.20118940e-01
2.41811156e-01 -1.26717913e+00 5.27193427e-01 1.83920264e-01
7.00574040e-01 4.95768309e-01 2.88119614e-01 6.68238699e-01
6.99748158e-01 -5.93648255e-01 9.95977998e-01 4.83526558e-01
6.89922214e-01 -3.73199642e-01 4.62501407e-01 2.86767095e-01
-1.37409401e+00 -2.82017648e-01 -3.15896273e-01 5.68269193e-02
3.41480047e-01 5.97788751e-01 -8.46047342e-01 7.93113351e-01
3.68884951e-01 1.24438000e+00 -5.08339345e-01 8.07064533e-01
-1.95069298e-01 1.14831448e+00 -1.80535927e-01 2.18430571e-02
2.42073238e-01 -2.02421293e-01 4.06041801e-01 1.51517212e+00
3.94359529e-01 -7.68809468e-02 3.32381129e-01 6.68855906e-01
-4.22743708e-01 3.64075363e-01 -1.05231062e-01 -4.63808030e-01
4.67023700e-01 9.94605780e-01 -3.68886083e-01 -4.92644042e-01
-6.88297451e-01 1.32637298e+00 3.22015226e-01 5.69174826e-01
-8.09501886e-01 -7.01157227e-02 6.03103280e-01 2.69726850e-02
6.42266035e-01 -4.11838263e-01 -1.80876285e-01 -1.83826244e+00
-6.05509169e-02 -9.16966379e-01 6.05972946e-01 -6.26929402e-01
-1.27574325e+00 4.36914861e-01 -4.04416353e-01 -1.12259781e+00
-3.58964771e-01 -5.76225162e-01 -5.20272017e-01 8.32379639e-01
-1.54272044e+00 -1.49690032e+00 2.00283885e-01 5.55990040e-01
9.73943889e-01 -1.39923736e-01 6.49424255e-01 8.59716296e-01
-7.66495764e-01 1.19012523e+00 5.28732419e-01 1.65376261e-01
8.68359983e-01 -1.05453873e+00 6.55288160e-01 1.14176357e+00
5.39330065e-01 9.46002483e-01 2.88641155e-01 -7.26332724e-01
-1.77304912e+00 -1.26468003e+00 1.14522374e+00 -6.88113689e-01
6.20528579e-01 -6.55848682e-01 -7.69892573e-01 1.10849512e+00
1.99270546e-01 -5.26760399e-01 8.33765447e-01 5.14866829e-01
-5.98885715e-01 1.36850793e-02 -2.14907795e-01 7.64650941e-01
1.09776688e+00 -9.12281573e-01 -4.91508693e-01 2.63252079e-01
9.89701211e-01 -2.05477133e-01 -8.98493528e-01 1.34252489e-01
8.77427518e-01 -6.41382098e-01 9.04176772e-01 -5.82982659e-01
5.72964013e-01 1.20869085e-01 -2.62581021e-01 -1.00834095e+00
-6.97398782e-01 -6.64180577e-01 -4.74472225e-01 1.14254773e+00
7.29358196e-01 -4.56567526e-01 4.33112264e-01 2.04184383e-01
-4.52519506e-01 -6.50556922e-01 -4.21150535e-01 -5.86774170e-01
1.01432458e-01 -7.37385094e-01 5.17413914e-01 7.17563987e-01
1.81436956e-01 6.84992969e-01 -8.00549448e-01 -1.43892430e-02
1.51731536e-01 1.31432116e-01 4.67187881e-01 -1.05986953e+00
-7.86040962e-01 -7.44480968e-01 4.98679608e-01 -2.10379195e+00
1.58810288e-01 -1.52908278e+00 -3.40219773e-02 -1.72550488e+00
4.29549873e-01 -2.74369985e-01 -5.55941463e-01 6.17998421e-01
-2.60310739e-01 1.81804180e-01 1.81868434e-01 3.55737954e-01
-9.08573270e-01 7.02567399e-01 1.29279006e+00 -5.59503958e-02
-1.69287518e-01 7.11221546e-02 -8.78064930e-01 2.16490000e-01
5.56521654e-01 -2.90111214e-01 -5.74836969e-01 -9.06809986e-01
6.56262159e-01 1.16299137e-01 -1.18379600e-01 -4.72503245e-01
4.17973995e-01 -1.05525352e-01 2.72019386e-01 -6.29842877e-01
2.41685033e-01 -5.42721927e-01 -1.70716241e-01 8.68814439e-02
-7.46353447e-01 -4.76722345e-02 -5.48385084e-03 6.42718494e-01
-5.65921962e-02 -7.64703602e-02 2.06102937e-01 4.30788286e-02
-5.38065076e-01 7.02703118e-01 -6.68832362e-01 -1.19680636e-01
4.01014797e-02 1.81315660e-01 -1.69550031e-01 -6.49872124e-01
-4.09166187e-01 2.52651870e-01 2.04805493e-01 7.29501784e-01
4.77068335e-01 -1.37484360e+00 -7.74298787e-01 1.79151192e-01
7.93489441e-02 -6.47638202e-01 2.52077520e-01 7.48055875e-01
1.77957624e-01 6.80095673e-01 2.75908649e-01 -1.69081181e-01
-1.09884441e+00 4.23723608e-01 1.82335243e-01 -5.97657919e-01
-7.32630610e-01 7.92655706e-01 2.39853382e-01 -3.13327253e-01
4.44384217e-01 -5.06656170e-01 -7.82552883e-02 5.82865551e-02
7.73432851e-01 2.02130437e-01 8.27726200e-02 -1.79409355e-01
9.52335820e-02 1.84051484e-01 -5.13122320e-01 -2.59368777e-01
1.22025824e+00 -5.04958212e-01 4.86369841e-02 4.79195982e-01
1.19829428e+00 3.01225752e-01 -1.05296493e+00 -8.27851355e-01
-1.07982224e-02 -2.81224489e-01 4.44132596e-01 -9.71275270e-01
-8.58110666e-01 8.99282634e-01 7.04438314e-02 -1.63022980e-01
1.03926599e+00 -1.03745282e-01 1.31975150e+00 4.71582741e-01
9.47098136e-02 -8.64232004e-01 2.75115132e-01 1.04161346e+00
8.24393332e-01 -1.11899221e+00 -1.21973149e-01 2.32338794e-02
-5.60511947e-01 1.01959026e+00 4.14129257e-01 9.57572237e-02
7.59716451e-01 2.11851537e-01 -2.32029304e-01 2.01286167e-01
-1.26456380e+00 -2.65125155e-01 1.02323771e+00 -2.54834056e-01
1.04183900e+00 -9.92405266e-02 -3.85301739e-01 9.94667470e-01
-9.90061387e-02 2.84346249e-02 2.15467006e-01 6.40100479e-01
-6.30797684e-01 -1.43316197e+00 1.48736611e-01 7.42694497e-01
-6.15898371e-01 -6.78035259e-01 -2.01935455e-01 2.83505302e-02
-3.00240308e-01 9.51052487e-01 1.25179470e-01 -7.39971817e-01
9.82143078e-03 2.05806896e-01 3.85677904e-01 -7.74068415e-01
-7.72916496e-01 4.62549120e-01 2.42406622e-01 -4.42137331e-01
1.38385072e-01 -4.75112259e-01 -9.88057315e-01 -4.69339907e-01
-3.89371902e-01 1.32478178e-01 3.77432197e-01 8.32088292e-01
6.38115346e-01 4.53333676e-01 6.62518620e-01 -7.02623129e-01
-6.07919455e-01 -1.24631965e+00 -4.50501144e-01 2.84769326e-01
2.94966847e-01 -1.74178243e-01 -7.88064301e-02 1.75799150e-02] | [10.230107307434082, 5.757397174835205] |
28458a56-f035-43e4-b10b-fdad1e871fb5 | entropy-dissipation-informed-neural-network | 2303.11205 | null | https://arxiv.org/abs/2303.11205v1 | https://arxiv.org/pdf/2303.11205v1.pdf | Entropy-dissipation Informed Neural Network for McKean-Vlasov Type PDEs | We extend the concept of self-consistency for the Fokker-Planck equation (FPE) to the more general McKean-Vlasov equation (MVE). While FPE describes the macroscopic behavior of particles under drift and diffusion, MVE accounts for the additional inter-particle interactions, which are often highly singular in physical systems. Two important examples considered in this paper are the MVE with Coulomb interactions and the vorticity formulation of the 2D Navier-Stokes equation. We show that a generalized self-consistency potential controls the KL-divergence between a hypothesis solution to the ground truth, through entropy dissipation. Built on this result, we propose to solve the MVEs by minimizing this potential function, while utilizing the neural networks for function approximation. We validate the empirical performance of our approach by comparing with state-of-the-art NN-based PDE solvers on several example problems. | ['Zhenfu Wang', 'Zebang Shen'] | 2023-02-11 | null | null | null | null | ['type'] | ['speech'] | [-3.92359704e-01 -2.63267070e-01 5.16502023e-01 2.44554356e-02
-1.30563574e-02 -3.38095665e-01 5.53538978e-01 3.13856870e-01
-5.92948020e-01 1.31641746e+00 -4.14560348e-01 6.34227023e-02
-2.16005281e-01 -9.12813246e-01 -7.51596689e-01 -1.11142576e+00
-4.07110184e-01 6.34767830e-01 6.33227304e-02 -3.95862490e-01
2.98041940e-01 7.36038029e-01 -1.52189040e+00 -3.33593518e-01
1.14556766e+00 1.29170132e+00 -1.63023755e-01 7.75621474e-01
1.88187599e-01 5.22678435e-01 -4.03145254e-01 1.34862950e-02
2.61222959e-01 -3.62379253e-01 -6.15588725e-01 -3.44573349e-01
4.24816132e-01 3.97861749e-02 -3.16438824e-01 1.21763611e+00
4.05075073e-01 8.71024251e-01 7.94521570e-01 -1.17418218e+00
-9.11679804e-01 5.74213415e-02 -3.69275510e-02 2.89376825e-01
1.04651406e-01 3.62110883e-01 6.97066247e-01 -7.40651190e-01
7.26154387e-01 1.12286961e+00 1.13305390e+00 6.12473845e-01
-1.27334988e+00 -1.78626418e-01 -2.64605969e-01 -4.37490158e-02
-1.44880903e+00 1.73081070e-01 5.61926901e-01 -6.20155752e-01
9.42406774e-01 4.14053589e-01 1.00941539e+00 8.38819444e-01
1.04354823e+00 3.20455581e-01 1.43996334e+00 -1.17085114e-01
8.96230102e-01 2.06844792e-01 3.35996330e-01 4.87465471e-01
8.47922266e-01 2.40701720e-01 -3.43868852e-01 -6.57471478e-01
7.63073802e-01 -2.45344371e-01 -4.35349852e-01 -4.76411074e-01
-6.38530910e-01 9.22537684e-01 2.16895416e-01 2.90538967e-01
-5.28427184e-01 1.49095312e-01 2.95007706e-01 -1.32191092e-01
7.76247919e-01 8.23126614e-01 -3.30132008e-01 4.67655947e-03
-8.33582819e-01 1.02633190e+00 1.48040736e+00 5.25119960e-01
6.74688280e-01 1.38005093e-01 -3.52593698e-02 3.16436470e-01
3.93275678e-01 8.62504184e-01 1.80246294e-01 -1.14746189e+00
-1.15371808e-01 2.76653409e-01 7.06450284e-01 -8.55316818e-01
-3.56829196e-01 -6.87279105e-01 -1.17345285e+00 4.77786362e-01
3.48191559e-01 -2.93070436e-01 -4.73739713e-01 1.41319776e+00
6.45825922e-01 3.25368226e-01 3.99007797e-01 1.22127855e+00
5.44369340e-01 7.75627315e-01 -3.46450269e-01 -5.99543929e-01
7.03507125e-01 -8.26531351e-01 -8.94131839e-01 1.33129045e-01
5.07776439e-01 -3.02348107e-01 4.73948300e-01 3.08221579e-01
-1.52310777e+00 -2.02912807e-01 -9.90638256e-01 1.81654721e-01
-4.16620642e-01 -4.84472841e-01 2.82575577e-01 2.27311879e-01
-9.81187880e-01 1.51269662e+00 -1.03459370e+00 -2.16191933e-01
6.57964349e-02 2.72182286e-01 -1.67853892e-01 6.08130991e-01
-1.29170299e+00 1.23986292e+00 1.00488532e-02 4.53992277e-01
-5.77977657e-01 -1.03632748e+00 -5.52192211e-01 -1.74717251e-02
-3.65229286e-02 -7.37716436e-01 8.82900417e-01 -5.32308638e-01
-1.58612955e+00 3.73454303e-01 -1.01619855e-01 -6.37478054e-01
8.25046897e-01 -1.61289036e-01 -1.12819010e-02 -2.73622517e-02
-7.74620548e-02 2.92388201e-02 3.43333006e-01 -1.38245761e+00
9.42575186e-02 4.93935496e-02 -1.88628688e-01 -2.64293160e-02
-2.85393596e-02 -4.25006628e-01 3.54099065e-01 -4.53346670e-01
-1.39666542e-01 -1.22743058e+00 -2.61052430e-01 1.47097940e-02
-2.56962895e-01 -1.34846374e-01 8.28692913e-01 -6.82407141e-01
8.77382755e-01 -1.70069456e+00 4.77307707e-01 3.28833342e-01
3.35992165e-02 3.36176991e-01 3.27009708e-01 5.16548812e-01
2.37082675e-01 1.34645542e-02 -8.30618501e-01 -4.88857031e-01
3.98855656e-02 4.78607088e-01 -2.36681819e-01 9.50311899e-01
9.82315466e-02 7.61602342e-01 -8.43420386e-01 -2.27084994e-01
9.75377113e-02 8.56615484e-01 -4.70058143e-01 9.18859392e-02
-2.74663746e-01 7.74465144e-01 -3.97317380e-01 2.13659823e-01
1.22908568e+00 -3.85433614e-01 -1.59190074e-01 -4.53689098e-02
-5.60228527e-01 -9.78613924e-03 -1.38522661e+00 1.34239137e+00
-2.70305991e-01 4.56630886e-01 6.89043462e-01 -1.15044749e+00
7.16715932e-01 4.73219194e-02 6.30120456e-01 -5.17479181e-01
2.69753367e-01 3.78226250e-01 -2.11172830e-02 -4.12126452e-01
6.42046034e-01 -4.81973737e-01 2.92371839e-01 1.64976686e-01
-2.06454113e-01 -3.24766666e-01 3.38039279e-01 2.90863197e-02
1.08148074e+00 1.98519543e-01 7.96693116e-02 -1.06290185e+00
7.99137175e-01 1.77508190e-01 5.47688365e-01 8.16554844e-01
-3.83378230e-02 3.34889561e-01 2.71077186e-01 -5.22703230e-01
-1.17068028e+00 -8.49631846e-01 -7.03876317e-01 5.45200258e-02
4.35645729e-01 -1.68576017e-01 -1.05846441e+00 9.62476432e-02
5.60196936e-01 6.75514698e-01 -8.85491908e-01 -3.58150452e-02
-7.23685861e-01 -1.01187909e+00 3.30316991e-01 1.55607626e-01
5.02985597e-01 -9.52343225e-01 -7.64250040e-01 4.01656777e-01
2.65088707e-01 -9.41727042e-01 1.78446937e-02 3.11448164e-02
-8.50424230e-01 -9.08555567e-01 -6.32610023e-01 -1.17453948e-01
4.42389071e-01 -4.97455090e-01 1.07150781e+00 1.52183041e-01
-5.10683894e-01 5.41676641e-01 6.66460907e-03 4.95895445e-02
-7.56118417e-01 -4.19103801e-01 5.34324825e-01 -1.98265359e-01
-1.70883283e-01 -5.67940891e-01 -6.19425714e-01 2.57696450e-01
-9.61315155e-01 -3.71746689e-01 -1.02732740e-02 6.47838831e-01
7.35399187e-01 7.96412155e-02 8.28547478e-02 -4.63419586e-01
7.29418218e-01 -5.01107693e-01 -9.14957464e-01 -4.88039963e-02
-6.15437150e-01 2.76909590e-01 9.43823755e-01 -6.84523523e-01
-9.78150725e-01 -2.20074639e-01 1.43398698e-02 -5.47181308e-01
1.37588561e-01 3.12059909e-01 4.74390417e-01 -7.48479068e-01
4.62965250e-01 2.24199355e-01 6.98204711e-02 -3.69651705e-01
-1.94849506e-01 2.85714328e-01 3.99316281e-01 -8.10527980e-01
6.20960712e-01 8.18984985e-01 8.94114614e-01 -1.02293682e+00
-6.28804684e-01 -2.64949918e-01 -3.53675842e-01 -2.08644167e-01
9.27973866e-01 -2.83266932e-01 -1.44959950e+00 6.68618917e-01
-1.43881297e+00 -3.96835238e-01 -7.17707455e-01 5.34306586e-01
-5.48883080e-01 4.73954201e-01 -8.93858850e-01 -1.28766584e+00
-3.51154089e-01 -1.05374515e+00 6.46227300e-01 3.03946674e-01
-3.50665450e-02 -1.63638973e+00 8.31718266e-01 -2.64000893e-01
7.36931205e-01 6.48100734e-01 4.10897374e-01 -4.84597117e-01
-3.02896887e-01 -1.34124979e-01 1.47097439e-01 6.27443075e-01
-4.14531946e-01 1.23275153e-01 -5.98636627e-01 -3.40471029e-01
6.91507220e-01 -8.71481560e-03 7.26015270e-01 7.22441196e-01
5.66805065e-01 -4.33975309e-01 -3.01277786e-01 6.41648412e-01
1.74330974e+00 -6.30763220e-03 2.68384844e-01 2.88211644e-01
5.67435205e-01 4.59347904e-01 2.51196563e-01 6.18457615e-01
9.05926526e-02 5.57003915e-01 4.38770801e-01 1.74423963e-01
3.92954111e-01 4.74580437e-01 2.74041831e-01 7.35227108e-01
-4.30968910e-01 -2.85690308e-01 -8.05466115e-01 1.38987809e-01
-2.07494760e+00 -7.30999768e-01 -7.66926289e-01 2.11685920e+00
7.33167708e-01 -1.55020103e-01 -3.04284453e-01 -3.03011566e-01
6.26049399e-01 -9.48520228e-02 -6.64101422e-01 -5.77545345e-01
-3.69137973e-01 1.91277221e-01 7.77394772e-01 8.78398657e-01
-1.08195210e+00 4.07458693e-01 6.49828243e+00 4.24871445e-01
-1.21755612e+00 3.52017373e-01 1.92164958e-01 -1.35695487e-01
7.50601590e-02 -9.17901248e-02 -8.43868196e-01 7.37319887e-01
1.04775751e+00 -6.76873699e-02 5.37706554e-01 6.34230912e-01
3.35229784e-01 -3.85411978e-01 -8.93546700e-01 7.04010844e-01
9.01785865e-02 -1.51014805e+00 -3.54761660e-01 3.60836118e-01
1.14938271e+00 1.82454109e-01 -2.58920014e-01 -1.20221265e-01
1.09340668e-01 -7.93497324e-01 6.03415549e-01 1.06293690e+00
8.58551338e-02 -3.70045781e-01 8.74568701e-01 2.59338737e-01
-1.05199063e+00 2.31084019e-01 -5.17595768e-01 -2.70691246e-01
5.60821712e-01 1.01752269e+00 2.20768064e-01 4.16935623e-01
5.69496989e-01 6.15508258e-01 -7.56236026e-03 1.08241570e+00
3.11748385e-01 2.57965118e-01 -6.59957409e-01 -2.95706451e-01
2.74343461e-01 -8.67276728e-01 1.15585303e+00 1.09048510e+00
5.17904699e-01 2.05487147e-01 -8.51621032e-02 1.44350767e+00
1.60436362e-01 -1.36169484e-02 -2.90690184e-01 1.41218379e-01
5.47263734e-02 1.33402216e+00 -7.12886631e-01 -4.85772014e-01
8.16471279e-02 8.00507903e-01 9.77859423e-02 4.60270494e-01
-7.64716327e-01 -4.81540933e-02 9.92295802e-01 8.48596394e-02
3.54262978e-01 -3.26860040e-01 -9.81184989e-02 -1.33159733e+00
2.82278687e-01 -1.83319360e-01 -1.32340953e-01 -6.40595973e-01
-1.81323969e+00 3.63866895e-01 1.00795038e-01 -9.04980183e-01
6.93446174e-02 -1.03718817e+00 -7.92697906e-01 1.06575310e+00
-1.40456080e+00 -4.50599760e-01 -5.55637181e-02 1.62300035e-01
-2.37648696e-01 2.55358756e-01 6.52714491e-01 1.65343266e-02
-5.09418964e-01 -3.29410583e-01 8.18863392e-01 -3.65663856e-01
2.55350113e-01 -1.40531254e+00 4.11998838e-01 4.71689880e-01
-8.24491024e-01 7.99545527e-01 1.30992329e+00 -1.09832120e+00
-1.81044507e+00 -8.24791849e-01 6.97375417e-01 -1.86354637e-01
9.44842637e-01 -3.90667200e-01 -1.37007785e+00 2.36536130e-01
1.02802753e-01 5.56271911e-01 1.44670442e-01 -4.25575346e-01
2.19224393e-01 1.94692820e-01 -1.43914640e+00 7.97815919e-02
8.02080989e-01 -1.63427234e-01 -3.69107634e-01 5.67154408e-01
2.02005878e-01 -6.60025954e-01 -1.13856339e+00 3.33642542e-01
4.78728086e-01 -9.92111683e-01 9.33068395e-01 -5.27712286e-01
4.54366356e-01 -2.31012553e-01 2.33217068e-02 -1.30726659e+00
-8.47555026e-02 -7.83636570e-01 -3.57195139e-01 8.14673901e-01
3.63843329e-02 -9.34860408e-01 5.19617558e-01 9.50570643e-01
-1.14056900e-01 -1.09876466e+00 -1.34293139e+00 -1.11161160e+00
6.18216693e-01 -1.50805041e-01 2.25132570e-01 1.06671953e+00
-8.66395459e-02 -3.36792082e-01 -2.92418331e-01 3.80314499e-01
9.95293677e-01 -1.80924699e-01 3.35990727e-01 -1.57993579e+00
-4.79736745e-01 -3.86031151e-01 -3.12023789e-01 -6.07254624e-01
5.55032253e-01 -7.60390460e-01 3.41844141e-01 -1.59647334e+00
-7.87005946e-02 -3.20747703e-01 1.84507184e-02 -2.63138175e-01
1.13946974e-01 8.33344012e-02 5.16155660e-02 2.95951545e-01
-3.86696637e-01 6.98006153e-01 1.29884982e+00 1.41424611e-01
-3.53832543e-01 -2.46298254e-01 1.03424527e-01 7.21347392e-01
7.11696208e-01 -6.12925053e-01 2.13661313e-01 1.46191791e-01
5.21694005e-01 -1.01571247e-01 7.17049897e-01 -1.25980854e+00
4.66879964e-01 -2.44300947e-01 2.39100620e-01 -2.24211022e-01
5.84654033e-01 -5.43151915e-01 2.65759856e-01 7.34999299e-01
-2.67604850e-02 5.93093671e-02 3.83860886e-01 5.33307850e-01
-4.30225395e-02 -4.79908496e-01 9.89769399e-01 -2.34962180e-01
1.09229468e-01 1.27008185e-01 -5.87686300e-01 3.61355335e-01
9.14302170e-01 1.02436431e-02 -4.92113292e-01 -1.07505158e-01
-7.74475276e-01 2.92381980e-02 8.69524240e-01 -3.72328311e-01
3.48281622e-01 -1.20303881e+00 -4.66301471e-01 2.13154599e-01
-6.53694332e-01 4.40162681e-02 3.94553065e-01 1.02222180e+00
-9.04841900e-01 2.04420656e-01 -1.85213253e-01 -5.66267669e-01
-7.97669709e-01 2.94111818e-01 8.85331333e-01 -4.14322019e-01
-6.77216411e-01 7.19939351e-01 -1.37474081e-02 -3.97046208e-01
-1.93523467e-01 -5.49544811e-01 2.04516456e-01 -2.89793223e-01
1.88781247e-01 1.05005455e+00 -1.32975087e-01 -7.77251840e-01
-5.53095460e-01 6.91477954e-01 4.66486067e-01 -2.54232585e-01
1.35438526e+00 1.24531671e-01 -6.71643317e-01 5.88285923e-01
1.20718861e+00 -1.09923817e-01 -1.45717442e+00 2.95013607e-01
-4.98690486e-01 3.50829922e-02 1.55416876e-01 -4.60914791e-01
-7.57530868e-01 7.32333124e-01 3.68165135e-01 7.25043356e-01
3.42411429e-01 -2.91072518e-01 9.73595798e-01 5.34752369e-01
4.88851480e-02 -1.33362186e+00 -4.42164809e-01 8.37183058e-01
8.26763093e-01 -8.10850143e-01 3.57266188e-01 -3.21729362e-01
-3.04507226e-01 1.16331232e+00 3.67458016e-01 -8.24056029e-01
8.64284515e-01 5.62116563e-01 -4.05852407e-01 -1.41972369e-02
-4.87168700e-01 2.92688549e-01 2.19605386e-01 9.47060809e-02
2.80231953e-01 -1.53394833e-01 -6.04194999e-01 1.71827078e-01
1.12434933e-02 3.51811051e-02 6.95011854e-01 1.13245642e+00
-2.15323076e-01 -9.36235666e-01 -5.05014598e-01 2.57160813e-01
-3.30877066e-01 1.00422174e-01 -3.14707816e-01 8.82676423e-01
2.52139658e-01 5.87741733e-01 3.42237294e-01 2.79516548e-01
2.21217364e-01 1.11672133e-01 4.93888766e-01 -1.35523245e-01
-7.95322478e-01 -1.36413664e-01 -3.38369548e-01 -4.73349571e-01
-7.91369498e-01 -7.99042106e-01 -1.48087263e+00 -4.50168222e-01
-2.22568125e-01 6.63852394e-01 8.95957112e-01 1.14062202e+00
2.11436331e-01 4.67498779e-01 6.97349459e-02 -1.42882538e+00
-6.53213203e-01 -9.17053461e-01 -9.95031178e-01 4.31764156e-01
5.43531775e-01 -9.30842876e-01 -1.09327531e+00 -3.05308163e-01] | [6.476781845092773, 3.542633056640625] |
79278216-39d2-4fa2-b0f5-b40f8f22c4c7 | testing-human-ability-to-detect-deepfake | 2212.05056 | null | https://arxiv.org/abs/2212.05056v3 | https://arxiv.org/pdf/2212.05056v3.pdf | Testing Human Ability To Detect Deepfake Images of Human Faces | Deepfakes are computationally-created entities that falsely represent reality. They can take image, video, and audio modalities, and pose a threat to many areas of systems and societies, comprising a topic of interest to various aspects of cybersecurity and cybersafety. In 2020 a workshop consulting AI experts from academia, policing, government, the private sector, and state security agencies ranked deepfakes as the most serious AI threat. These experts noted that since fake material can propagate through many uncontrolled routes, changes in citizen behaviour may be the only effective defence. This study aims to assess human ability to identify image deepfakes of human faces (StyleGAN2:FFHQ) from nondeepfake images (FFHQ), and to assess the effectiveness of simple interventions intended to improve detection accuracy. Using an online survey, 280 participants were randomly allocated to one of four groups: a control group, and 3 assistance interventions. Each participant was shown a sequence of 20 images randomly selected from a pool of 50 deepfake and 50 real images of human faces. Participants were asked if each image was AI-generated or not, to report their confidence, and to describe the reasoning behind each response. Overall detection accuracy was only just above chance and none of the interventions significantly improved this. Participants' confidence in their answers was high and unrelated to accuracy. Assessing the results on a per-image basis reveals participants consistently found certain images harder to label correctly, but reported similarly high confidence regardless of the image. Thus, although participant accuracy was 62% overall, this accuracy across images ranged quite evenly between 85% and 30%, with an accuracy of below 50% for one in every five images. We interpret the findings as suggesting that there is a need for an urgent call to action to address this threat. | ['Bennett Kleinberg', 'Shane D. Johnson', 'Sergi D. Bray'] | 2022-12-07 | null | null | null | null | ['human-detection-of-deepfakes'] | ['miscellaneous'] | [ 2.49392986e-01 4.52201873e-01 5.09032048e-02 -1.49398714e-01
-6.59466028e-01 -8.96144390e-01 6.19171560e-01 8.39512572e-02
-6.99261904e-01 4.74941969e-01 2.50854045e-01 -5.73780537e-01
3.00844759e-01 -7.12580204e-01 -5.88815153e-01 -3.41570169e-01
3.95518810e-01 -2.69083232e-02 7.64585659e-02 3.28156748e-03
6.67236388e-01 5.57116449e-01 -1.22998393e+00 5.66458106e-01
6.27434075e-01 7.14970648e-01 -5.17637312e-01 6.10870302e-01
1.57580495e-01 9.77848589e-01 -1.02922821e+00 -9.15452003e-01
5.47994018e-01 -4.20982122e-01 -5.68297267e-01 -5.02154827e-02
1.07489955e+00 -9.20888662e-01 -2.80762315e-01 1.12819123e+00
4.53377306e-01 -1.03887260e-01 4.25141454e-01 -1.37902999e+00
-7.33996332e-01 6.27006292e-02 -5.98327339e-01 3.67887914e-01
6.90679193e-01 8.76522183e-01 3.89137119e-01 -5.72710931e-01
6.06766939e-01 1.43873942e+00 8.42251956e-01 4.50342596e-01
-1.29631889e+00 -1.44299889e+00 -5.83393984e-02 4.41752225e-02
-1.35270321e+00 -7.65243053e-01 1.96326837e-01 -7.95506656e-01
7.66395569e-01 2.99060971e-01 9.58457887e-01 1.09209323e+00
1.85111493e-01 -2.85164341e-02 1.53061604e+00 -1.55961066e-01
4.48394343e-02 6.53989553e-01 -6.74049109e-02 6.30334139e-01
7.36079812e-01 4.28666294e-01 -4.20247704e-01 -6.98596120e-01
5.96674860e-01 -3.56433868e-01 -2.51028657e-01 2.02964306e-01
-1.04215336e+00 9.29327965e-01 4.98165846e-01 2.51272112e-01
-3.69670630e-01 -1.93187483e-02 3.06922615e-01 2.29016945e-01
2.59531230e-01 8.04796875e-01 2.28264257e-01 -3.39449078e-01
-4.97242033e-01 2.25232020e-01 6.20915890e-01 2.14076173e-02
4.19809788e-01 1.04144886e-01 1.80326328e-02 4.26739752e-01
-1.57308325e-01 6.97999656e-01 4.35954221e-02 -1.12954557e+00
2.62113512e-01 4.28744793e-01 6.98543429e-01 -2.08979344e+00
-2.92418033e-01 -4.58489917e-02 -4.83054638e-01 5.27091026e-01
6.80447459e-01 -3.16171736e-01 -7.24641144e-01 1.55224693e+00
1.99451730e-01 -3.77656743e-02 -5.45814097e-01 1.21875632e+00
2.81131327e-01 6.38501585e-01 4.67811406e-01 -5.92407659e-02
1.44438565e+00 6.44112229e-02 -5.61625063e-01 -4.66409832e-01
5.06942868e-01 -7.42259145e-01 9.95601296e-01 6.69333279e-01
-7.60800481e-01 -4.20232475e-01 -9.61329818e-01 3.78117591e-01
-3.34347606e-01 -1.00509435e-01 3.52219671e-01 1.39228153e+00
-9.46786046e-01 2.41128534e-01 -1.78089857e-01 -4.94864672e-01
6.35111749e-01 2.07625311e-02 -5.32512069e-01 -1.30944073e-01
-1.18914306e+00 1.41633201e+00 2.18424380e-01 1.06279828e-01
-6.57976687e-01 -3.03342134e-01 -6.55284047e-01 -2.53937989e-01
1.21978477e-01 -2.59880662e-01 9.87052858e-01 -1.49866295e+00
-8.19904089e-01 1.26568604e+00 7.10225571e-03 -5.89910030e-01
6.27026260e-01 -1.10078998e-01 -6.75286353e-01 4.96402264e-01
4.82350796e-01 8.18869233e-01 1.20565665e+00 -1.14482820e+00
-4.61603522e-01 -5.10647833e-01 1.84571445e-01 8.99437666e-02
-1.78639993e-01 4.60951716e-01 5.78853965e-01 -3.98354977e-01
-2.56167591e-01 -9.77567136e-01 1.40617326e-01 1.45921394e-01
-2.02403322e-01 5.30176498e-02 7.90299416e-01 -9.13253844e-01
7.98527300e-01 -2.23894334e+00 -8.71835291e-01 2.48645812e-01
3.76933277e-01 4.59120125e-01 -3.73554304e-02 2.25349620e-01
5.19758984e-02 6.21015251e-01 1.88279450e-01 4.50429916e-01
-2.11599529e-01 -3.19060534e-01 -3.40488702e-01 9.03777242e-01
3.03014934e-01 6.46262467e-01 -1.05819404e+00 -1.45151049e-01
2.69054353e-01 3.60835254e-01 -1.03589602e-01 -2.44657204e-01
4.44369853e-01 2.36371964e-01 -4.59445901e-02 4.11412567e-01
8.05115581e-01 1.27477318e-01 1.86682120e-02 -1.79221883e-01
-2.36903116e-01 7.97063261e-02 -7.88133085e-01 4.80006754e-01
-2.65987013e-02 1.15965378e+00 1.61448941e-01 -6.38060451e-01
6.63569093e-01 3.05119187e-01 -1.16259910e-01 -8.91241729e-01
2.01542690e-01 2.37277433e-01 2.78828800e-01 -5.30305922e-01
5.66686153e-01 -2.99395561e-01 -1.27566501e-01 6.67751253e-01
-6.54074371e-01 -2.50188529e-01 -4.08777565e-01 3.21258157e-01
1.20763254e+00 -5.64170301e-01 5.33504263e-02 3.23298201e-02
1.05909519e-01 4.27552313e-01 3.38691175e-01 1.13355088e+00
-7.47261047e-01 2.79303521e-01 4.94132131e-01 -7.09431350e-01
-9.70938563e-01 -9.38062131e-01 1.30276173e-01 8.10424149e-01
3.32081467e-01 -7.26245418e-02 -8.83387923e-01 -7.04117894e-01
4.36426178e-02 1.06668890e+00 -5.50016940e-01 -5.62403619e-01
-1.10214956e-01 -3.60357493e-01 8.88634324e-01 9.34497118e-02
8.36462677e-01 -9.74800348e-01 -1.08903825e+00 -9.36866775e-02
-4.47020352e-01 -1.09458506e+00 -4.93130207e-01 -8.40609848e-01
-1.82605714e-01 -1.23631108e+00 -4.71826911e-01 -9.60039943e-02
7.93688118e-01 7.51301706e-01 7.44709671e-01 4.15016711e-01
-4.70482975e-01 6.23002708e-01 -1.80497676e-01 -5.28613448e-01
-7.33589888e-01 -6.44277096e-01 1.41055450e-01 9.65598002e-02
6.52074277e-01 1.76734194e-01 -6.50269628e-01 4.62192565e-01
-6.64062738e-01 -1.58310771e-01 5.25473237e-01 4.22668964e-01
-3.21278691e-01 1.07070088e-01 4.81552631e-01 -6.63212836e-01
8.34183037e-01 -4.05375808e-01 -4.31436986e-01 2.73598582e-02
-3.27837467e-01 -7.47638881e-01 2.19939187e-01 -6.62455559e-01
-7.76854396e-01 -1.32450312e-01 4.72685516e-01 -1.36520073e-01
-4.02839094e-01 9.91867781e-02 2.22840279e-01 -4.77023453e-01
1.15803027e+00 -9.53542218e-02 2.76178896e-01 2.60058731e-01
2.83730496e-02 7.14273214e-01 4.29664999e-01 -1.30580738e-01
8.61936867e-01 5.57349682e-01 -5.31571686e-01 -1.09103477e+00
-7.20459759e-01 -8.80491808e-02 3.57501674e-03 -7.46734619e-01
8.80746305e-01 -8.07918370e-01 -9.95910466e-01 6.28034949e-01
-1.25214338e+00 -3.22207212e-01 1.88958988e-01 5.18604517e-01
1.08621359e-01 3.33828062e-01 -3.58752996e-01 -1.13894331e+00
6.58910125e-02 -8.88873935e-01 6.62701905e-01 4.36083749e-02
-6.31285310e-01 -3.39999110e-01 -2.93032587e-01 7.97667801e-01
5.26100934e-01 4.12829638e-01 5.06825626e-01 -4.57777917e-01
-3.47874522e-01 -7.39658237e-01 -7.16176152e-01 2.77705491e-01
3.08390617e-01 -1.21313319e-01 -1.03759813e+00 -2.98749208e-01
-1.19140400e-02 -5.87336183e-01 4.79370952e-01 1.87485456e-01
7.16445148e-01 -7.18627572e-01 -3.64288986e-01 -2.61456549e-01
1.02253437e+00 4.06975567e-01 8.38749647e-01 3.84714365e-01
4.39656496e-01 8.18096161e-01 4.66619909e-01 1.66786179e-01
2.36164883e-01 6.04342163e-01 2.75858581e-01 1.22838460e-01
1.86713636e-01 -3.25957298e-01 8.03094089e-01 -3.14821184e-01
6.84449449e-02 -1.68547153e-01 -9.98184144e-01 4.43758249e-01
-1.21721661e+00 -1.51808465e+00 -2.06750691e-01 2.38668370e+00
6.24332055e-02 3.54805380e-01 4.70419854e-01 9.51843262e-02
1.14932120e+00 8.62926152e-03 -5.31368971e-01 -6.48962259e-01
-9.68302134e-03 -1.44010171e-01 7.92250872e-01 4.01154488e-01
-9.28058863e-01 7.68556595e-01 6.83468390e+00 4.64461356e-01
-1.20927250e+00 -1.47931166e-02 1.06017435e+00 1.63436010e-02
-1.27144635e-01 8.93293396e-02 -2.87805349e-01 5.78638434e-01
1.05400920e+00 -2.66926378e-01 3.92095625e-01 3.00542146e-01
5.77522099e-01 -5.04922092e-01 -3.72295320e-01 7.92787373e-01
2.45317012e-01 -1.16968036e+00 -5.11109456e-02 3.02144647e-01
3.75039577e-01 -2.87301600e-01 2.01953501e-01 8.50450061e-03
3.91706705e-01 -1.42896843e+00 9.22162533e-01 2.41468996e-01
9.32257175e-01 -6.13649011e-01 6.48601949e-01 4.97597635e-01
-4.88622040e-01 -2.52648473e-01 -8.78740996e-02 -6.97635412e-01
9.41114649e-02 2.44358435e-01 -1.00034654e+00 -2.63398916e-01
7.67782450e-01 -4.36802991e-02 -4.86864239e-01 6.58984542e-01
-1.53585225e-01 6.59169912e-01 -3.68999630e-01 2.68703438e-02
2.68856227e-01 1.95121884e-01 5.64696074e-01 1.09020746e+00
1.59131408e-01 5.64468622e-01 -2.65198916e-01 7.02955604e-01
1.84279785e-01 -3.45170945e-01 -1.07648981e+00 -4.94708210e-01
7.96865463e-01 1.04835796e+00 -9.44587767e-01 -2.74169058e-01
-3.97132605e-01 9.69387531e-01 -6.62250817e-02 1.54062271e-01
-6.26245499e-01 -3.56757760e-01 5.27949512e-01 6.43137336e-01
-2.16199443e-01 -4.78966236e-02 -1.71133220e-01 -8.73194695e-01
-1.46252513e-01 -1.22070122e+00 2.83892572e-01 -9.29047406e-01
-1.20091081e+00 4.15852547e-01 -2.09610108e-02 -1.03593063e+00
-1.26838267e-01 -3.79401565e-01 -3.55616182e-01 8.96585226e-01
-7.13795841e-01 -8.49922895e-01 -6.67941153e-01 2.31309995e-01
-9.33061261e-03 7.16023818e-02 6.55647516e-01 -7.90904686e-02
-2.12165892e-01 5.96833229e-01 -6.48024619e-01 3.57640386e-01
7.84940362e-01 -4.71910357e-01 5.84608495e-01 7.75883734e-01
-1.97811007e-01 7.08278358e-01 6.69650555e-01 -9.40270185e-01
-9.26785886e-01 -7.17815578e-01 7.11830795e-01 -6.02733552e-01
4.47591275e-01 -1.85852006e-01 -7.52650738e-01 5.05526423e-01
1.72000483e-01 -2.10876837e-01 5.68010747e-01 -3.75446051e-01
-5.13922453e-01 3.49754244e-01 -1.81009293e+00 6.34605110e-01
9.00450885e-01 -8.17828774e-01 -4.73575234e-01 1.03772141e-01
2.23053083e-01 -1.18167132e-01 -3.33260149e-01 4.29153033e-02
9.90078330e-01 -1.44236302e+00 8.19279492e-01 -4.19309676e-01
1.83797553e-01 -2.52142787e-01 9.90506932e-02 -1.20536160e+00
-5.36438346e-01 -5.57954013e-01 5.41904211e-01 8.90746057e-01
9.82083380e-02 -1.06039882e+00 7.72794187e-01 1.13783991e+00
3.28739345e-01 -3.29624973e-02 -9.20489073e-01 -5.78618705e-01
-8.31966773e-02 -5.74165285e-01 3.07741255e-01 1.24326420e+00
1.10814841e-02 1.69371832e-02 -4.44417387e-01 4.17943656e-01
6.41497314e-01 -5.16218245e-01 8.92985642e-01 -9.20435905e-01
1.00367777e-01 -5.58590829e-01 -7.82247484e-01 -2.73847729e-01
-1.58849582e-01 -2.46342495e-01 -2.87371755e-01 -1.25934970e+00
2.39401534e-01 -2.08180264e-01 2.38453254e-01 4.41644967e-01
-1.21856503e-01 6.32961810e-01 5.30608058e-01 2.28066638e-01
3.27894017e-02 -1.82937890e-01 9.61112440e-01 -1.37812033e-01
6.93666264e-02 -2.56454706e-01 -1.21081161e+00 7.04493761e-01
7.68157065e-01 -4.00115132e-01 -2.45178431e-01 -3.56178254e-01
3.63811761e-01 1.64410900e-02 1.20241964e+00 -9.89281952e-01
5.39760105e-02 -3.97189617e-01 6.69976294e-01 -1.35261342e-01
5.32955110e-01 -7.58755863e-01 3.65636706e-01 8.54401410e-01
-9.26322863e-02 7.90060833e-02 5.49495757e-01 3.19745928e-01
1.96529821e-01 -1.42548412e-01 8.45739305e-01 -3.34478855e-01
-4.69366699e-01 -3.26100945e-01 -8.78485858e-01 6.06330670e-02
1.07145274e+00 -6.55099213e-01 -7.09520876e-01 -9.89159465e-01
-2.65199602e-01 -1.91436037e-01 7.14965463e-01 3.46104801e-01
7.60000765e-01 -1.08860135e+00 -7.03150570e-01 1.67469800e-01
1.02826487e-02 -7.38260031e-01 3.05532962e-01 7.43491054e-01
-5.35306692e-01 2.47518718e-01 -3.71860415e-01 -1.35383621e-01
-1.07926464e+00 2.42680624e-01 3.82914454e-01 4.73500252e-01
-3.34244519e-01 8.19060504e-01 2.87955284e-01 -5.42512760e-02
-8.96409974e-02 2.37586454e-01 1.06773712e-01 1.10083774e-01
9.94656444e-01 5.27230859e-01 -3.47263813e-01 -1.15861094e+00
-5.22673070e-01 2.79732168e-01 -1.71134323e-01 -2.11904705e-01
6.40617132e-01 -3.54632619e-04 1.04850180e-01 8.98663551e-02
8.85618806e-01 1.25250518e-01 -9.92639720e-01 4.16756153e-01
-2.29067594e-01 -1.24277413e+00 -3.12835127e-02 -9.73563433e-01
-6.07247710e-01 5.94888747e-01 6.92458570e-01 6.08263016e-01
8.47375095e-01 -2.73978919e-01 6.69870377e-01 7.32690319e-02
6.15863979e-01 -7.54052877e-01 2.56136775e-01 -1.01415433e-01
8.78845394e-01 -1.12738049e+00 3.96618582e-02 -3.01373541e-01
-8.57613027e-01 8.41578543e-01 6.33169830e-01 -1.50989801e-01
6.89848661e-02 -2.80272365e-01 1.09450080e-01 -4.16090071e-01
-3.30718189e-01 1.50664315e-01 -3.45355794e-02 8.83136570e-01
1.47670761e-01 2.00311273e-01 -2.91857034e-01 1.76231787e-01
-2.60983378e-01 1.76031426e-01 7.67465770e-01 5.79834104e-01
-5.87943971e-01 -4.40867215e-01 -9.54984665e-01 7.61130452e-01
-4.27603453e-01 2.56494612e-01 -1.07177222e+00 7.56105840e-01
3.22610348e-01 1.22248328e+00 2.61569500e-01 -5.68125010e-01
3.72687221e-01 -1.88860208e-01 3.45748931e-01 -3.27105284e-01
-6.53749764e-01 -6.64024532e-01 4.76443559e-01 -5.18249929e-01
-2.14571565e-01 -8.76229227e-01 -6.46189332e-01 -8.20239782e-01
-3.80736291e-01 -1.24098584e-02 5.05474269e-01 7.79083490e-01
4.40454781e-01 -4.78242397e-01 5.62552512e-01 -8.10662091e-01
-3.34713906e-01 -7.50585914e-01 -1.72201335e-01 4.35560077e-01
3.88668865e-01 -6.47604048e-01 -6.80128098e-01 -3.09247106e-01] | [12.454084396362305, 1.1748028993606567] |
3d2c9d50-4aba-4b46-8c9b-4be563b4c321 | edpn-enhanced-deep-pyramid-network-for-blurry | 2105.04872 | null | https://arxiv.org/abs/2105.04872v1 | https://arxiv.org/pdf/2105.04872v1.pdf | EDPN: Enhanced Deep Pyramid Network for Blurry Image Restoration | Image deblurring has seen a great improvement with the development of deep neural networks. In practice, however, blurry images often suffer from additional degradations such as downscaling and compression. To address these challenges, we propose an Enhanced Deep Pyramid Network (EDPN) for blurry image restoration from multiple degradations, by fully exploiting the self- and cross-scale similarities in the degraded image.Specifically, we design two pyramid-based modules, i.e., the pyramid progressive transfer (PPT) module and the pyramid self-attention (PSA) module, as the main components of the proposed network. By taking several replicated blurry images as inputs, the PPT module transfers both self- and cross-scale similarity information from the same degraded image in a progressive manner. Then, the PSA module fuses the above transferred features for subsequent restoration using self- and spatial-attention mechanisms. Experimental results demonstrate that our method significantly outperforms existing solutions for blurry image super-resolution and blurry image deblocking. In the NTIRE 2021 Image Deblurring Challenge, EDPN achieves the best PSNR/SSIM/LPIPS scores in Track 1 (Low Resolution) and the best SSIM/LPIPS scores in Track 2 (JPEG Artifacts). | ['Zhiwei Xiong', 'Yueyi Zhang', 'Jie Huang', 'Zeyu Xiao', 'Ruikang Xu'] | 2021-05-11 | null | null | null | null | ['image-deblocking'] | ['computer-vision'] | [ 2.37848774e-01 -7.31774747e-01 1.49600551e-01 -7.29491338e-02
-6.26447320e-01 -2.66815543e-01 3.72914970e-01 -5.67775309e-01
-3.46226618e-02 7.51735508e-01 8.81487906e-01 -8.06336664e-03
-1.72943518e-01 -2.95576811e-01 -7.42861390e-01 -7.56218612e-01
-2.29102615e-02 -5.72646856e-01 2.36698449e-01 -2.71459073e-01
4.86465663e-01 2.88411409e-01 -1.41911209e+00 4.70737785e-01
1.30032086e+00 1.07205915e+00 6.50907695e-01 8.60244572e-01
3.60116720e-01 1.10676134e+00 -4.56172258e-01 -1.07446902e-01
2.77521193e-01 -3.77474010e-01 -7.30238080e-01 1.48802027e-01
6.85391009e-01 -1.08776844e+00 -1.01666915e+00 1.55944335e+00
6.74486995e-01 1.08089820e-02 1.63799882e-01 -7.37796605e-01
-1.39822364e+00 3.82580668e-01 -9.40896213e-01 8.88189137e-01
2.24957496e-01 5.62990427e-01 5.15305519e-01 -1.05084586e+00
2.76535869e-01 1.53288054e+00 6.67389750e-01 2.11207747e-01
-9.91753161e-01 -5.43699861e-01 -3.05544257e-01 9.61312354e-01
-1.14980102e+00 -6.02974951e-01 5.24725854e-01 -8.35779905e-02
6.89798176e-01 1.00099705e-01 1.87753007e-01 8.22355628e-01
4.35658842e-01 6.47902608e-01 1.25288498e+00 5.18344715e-02
-2.85558328e-02 -3.73326033e-01 -4.80954041e-04 9.17182714e-02
2.03263551e-01 3.14750493e-01 -4.20769364e-01 2.54491806e-01
1.08139956e+00 1.54003665e-01 -9.78963971e-01 -2.03664247e-02
-1.16931558e+00 2.25423768e-01 8.49927425e-01 2.32125759e-01
-6.54207706e-01 2.46125713e-01 3.81367117e-01 4.42436248e-01
4.37004626e-01 1.73717737e-01 -3.90928686e-01 -1.13607027e-01
-1.06414199e+00 5.83472326e-02 2.13491738e-01 7.71742404e-01
5.79812944e-01 -2.22208668e-02 -5.06668448e-01 1.12773740e+00
9.34542269e-02 4.79210228e-01 8.54269862e-01 -1.15867436e+00
5.69552660e-01 -7.51673896e-03 6.07775390e-01 -1.16225874e+00
1.22154996e-01 -5.55333197e-01 -1.43366730e+00 8.61172229e-02
-7.47481734e-02 2.13248432e-02 -1.00886047e+00 1.60563171e+00
-6.80327415e-02 5.31034946e-01 1.46418288e-01 1.44337654e+00
6.04440153e-01 6.85700417e-01 -1.19459853e-01 -2.90992439e-01
1.52849996e+00 -1.35291302e+00 -9.35490310e-01 -3.64371002e-01
-3.28554839e-01 -1.00187337e+00 9.93874013e-01 1.73240229e-01
-1.35600221e+00 -1.05929220e+00 -1.27043068e+00 -4.38890010e-01
2.04348832e-01 -4.99066571e-03 -5.55065386e-02 3.30884963e-01
-1.35627687e+00 9.40798402e-01 -6.42257273e-01 -8.24449658e-02
5.50278425e-01 2.98949350e-02 -1.89421162e-01 -6.74719632e-01
-1.50343215e+00 9.69625235e-01 2.11721942e-01 3.36340308e-01
-1.13694012e+00 -7.63761759e-01 -7.52402127e-01 5.28755248e-01
1.01961061e-01 -8.44162583e-01 1.14686084e+00 -9.30135012e-01
-1.23869419e+00 3.11059862e-01 -1.46526173e-01 -6.13500357e-01
6.80290878e-01 -7.30841339e-01 -6.90410435e-01 3.51792693e-01
1.31487653e-01 5.06925941e-01 1.39909661e+00 -1.30033123e+00
-5.85513949e-01 -2.27033153e-01 -7.44734183e-02 4.55172598e-01
-1.62574977e-01 1.24009438e-01 -5.44632673e-01 -9.06120241e-01
-2.05111410e-02 -4.78430808e-01 2.88082846e-02 -1.07436687e-01
-2.13912636e-01 4.15051550e-01 1.17945004e+00 -1.39897501e+00
1.30620956e+00 -2.49632335e+00 1.80865526e-01 -5.85134923e-01
4.00503218e-01 5.98417997e-01 -3.81385982e-01 1.23484656e-02
-4.35621172e-01 -9.53961760e-02 -3.60712379e-01 -2.25380063e-01
-3.03333253e-01 5.04157804e-02 -5.45720518e-01 4.63053435e-01
1.60318807e-01 8.81806970e-01 -1.07902622e+00 -7.89495260e-02
3.52091521e-01 7.33172536e-01 -3.36223394e-01 3.77269536e-01
1.30944908e-01 4.64461952e-01 6.27905950e-02 4.07575846e-01
1.38758802e+00 -4.22772706e-01 -2.14750215e-01 -7.70206571e-01
-2.39508227e-01 -2.48398143e-03 -8.56635749e-01 1.67564476e+00
-3.98918629e-01 8.05862427e-01 3.09110194e-01 -4.14597869e-01
5.29897392e-01 3.11443985e-01 6.59896433e-02 -8.16978574e-01
-6.01105653e-02 3.35980058e-01 -1.57805890e-01 -7.43300915e-01
7.37454772e-01 6.49505407e-02 4.24083531e-01 1.68076098e-01
-1.06968783e-01 1.47652291e-02 -2.71439180e-02 1.38267010e-01
1.16148472e+00 -8.16973075e-02 1.24933138e-01 -1.34553805e-01
7.14395285e-01 -5.24255514e-01 3.59221220e-01 6.98561907e-01
-5.68794310e-01 1.18932021e+00 1.22077532e-01 -3.31903577e-01
-1.42558014e+00 -1.11566710e+00 1.11079216e-01 6.92764521e-01
5.93347609e-01 1.76686589e-02 -8.18177998e-01 -1.62276596e-01
-1.80706233e-01 4.26399887e-01 -4.42924470e-01 -3.64527106e-01
-5.98446727e-01 -8.81842792e-01 2.95897573e-01 4.52295572e-01
1.41482782e+00 -9.75049615e-01 -3.07890892e-01 1.08894780e-01
-6.60437107e-01 -1.20013022e+00 -1.15731192e+00 -2.99161702e-01
-8.42843473e-01 -9.13096249e-01 -1.41279376e+00 -7.97287166e-01
4.11908537e-01 1.05385303e+00 7.17371345e-01 -4.18541431e-02
-7.51737878e-02 -9.47384909e-02 -2.45278299e-01 4.22901601e-01
-3.89711916e-01 -3.57019931e-01 1.78150963e-02 3.12955052e-01
-1.55286089e-01 -8.22956502e-01 -1.22472501e+00 3.49046737e-01
-1.15016139e+00 6.61622733e-02 9.52711642e-01 1.11092877e+00
1.01044392e-02 4.38745290e-01 4.10863549e-01 -9.09695178e-02
9.80178118e-01 -5.14423788e-01 -4.22250062e-01 8.64490420e-02
-6.47985876e-01 -4.69512194e-02 6.88488901e-01 -3.60950232e-01
-1.36680138e+00 -5.79281032e-01 8.52611586e-02 -9.19672966e-01
-1.54998660e-01 3.77921224e-01 -1.68661952e-01 -6.41531274e-02
5.96724391e-01 7.34425008e-01 3.50444391e-02 -8.92643034e-01
4.33357507e-01 9.92170453e-01 1.15616882e+00 8.46509114e-02
9.06659544e-01 3.09652627e-01 -3.83144081e-01 -5.70868075e-01
-7.58986115e-01 -2.43832633e-01 -1.94751322e-01 3.21977101e-02
6.70321941e-01 -1.40211570e+00 -5.64417779e-01 1.20762384e+00
-1.23225021e+00 -8.76963735e-02 -3.22687514e-02 3.47838104e-01
-3.28103364e-01 1.16322541e+00 -1.19222951e+00 -2.17467710e-01
-5.96091568e-01 -1.25764787e+00 7.85176218e-01 6.16825759e-01
2.81298995e-01 -5.81043303e-01 -1.92319050e-01 6.13243759e-01
1.13397384e+00 -2.90362358e-01 6.92400455e-01 1.92452669e-01
-7.46400595e-01 2.42414087e-01 -1.08469141e+00 9.38458562e-01
5.98704398e-01 -6.48134828e-01 -8.87807786e-01 -7.01747119e-01
4.24194694e-01 -3.31651084e-02 1.06464565e+00 4.34639633e-01
1.15524077e+00 -5.49939632e-01 1.69016376e-01 8.60578895e-01
1.48513532e+00 1.01626376e-02 1.24042082e+00 5.09937167e-01
7.54807472e-01 1.83421299e-02 2.68136948e-01 3.03397089e-01
3.53174329e-01 6.21270597e-01 4.79648560e-01 -1.16145715e-01
-8.67686391e-01 -5.68829626e-02 6.04970574e-01 7.51283765e-01
1.27902664e-02 5.66577278e-02 -4.84801739e-01 7.09995151e-01
-1.64212632e+00 -1.03962409e+00 6.13845959e-02 1.99311459e+00
1.17094886e+00 -1.72071591e-01 -4.68510985e-01 -1.49580851e-01
1.09415519e+00 4.77356285e-01 -8.28024745e-01 8.96741897e-02
-5.16373396e-01 -7.55122527e-02 5.49856126e-01 6.56799555e-01
-1.04576564e+00 7.01192796e-01 5.62652159e+00 1.01393163e+00
-1.08982217e+00 1.99764282e-01 8.78983974e-01 1.63172439e-01
1.03628278e-01 -3.68901566e-02 -1.72441989e-01 1.03671205e+00
8.83526802e-01 -2.88764298e-01 1.06221890e+00 4.65640813e-01
5.86170018e-01 -2.15063602e-01 -5.87317050e-01 1.18971646e+00
1.29644036e-01 -1.25918448e+00 -1.92640591e-02 -2.97210723e-01
9.38275218e-01 2.47539416e-01 4.30075377e-01 -4.58172485e-02
1.52638882e-01 -9.41481531e-01 5.63376248e-01 4.91797179e-01
1.06664491e+00 -4.58917856e-01 9.17757750e-01 9.37492549e-02
-9.87181008e-01 -3.08369517e-01 -3.77329350e-01 1.41765475e-01
2.75790423e-01 7.06287801e-01 -2.74037987e-01 8.00393403e-01
1.12569368e+00 1.06352305e+00 -5.56314230e-01 1.34519088e+00
-2.81249613e-01 3.73083532e-01 2.54888326e-01 8.48343492e-01
1.03585429e-01 -1.34195551e-01 7.35992074e-01 1.08013546e+00
5.52417517e-01 2.16210168e-02 -5.57853639e-01 8.34629059e-01
-2.20176011e-01 -5.93898952e-01 1.32009521e-01 3.57627183e-01
3.05340618e-01 1.02758741e+00 8.56367126e-03 -5.92990041e-01
-3.42247546e-01 1.64016533e+00 -7.66091347e-02 5.94553769e-01
-8.45414042e-01 -4.92250264e-01 1.03853202e+00 -2.65873581e-01
7.30306208e-01 -6.61003962e-02 -1.24552779e-01 -1.53301549e+00
3.08536086e-02 -1.20289934e+00 -9.61047318e-03 -1.48095751e+00
-1.37337482e+00 6.34349704e-01 -2.75350571e-01 -1.33862662e+00
4.20558870e-01 -2.18311012e-01 -4.28781599e-01 1.33367324e+00
-2.03675222e+00 -8.12925696e-01 -8.44631791e-01 5.87595761e-01
8.15580249e-01 1.80664748e-01 1.55836076e-01 6.65432155e-01
-5.88864565e-01 1.55875608e-01 4.81084555e-01 -4.31875773e-02
1.05421340e+00 -1.07274354e+00 6.51633978e-01 1.33341467e+00
-6.39344990e-01 5.77473700e-01 9.15599942e-01 -7.78599560e-01
-1.22116947e+00 -1.24589121e+00 5.12942672e-01 2.41408110e-01
5.95969141e-01 3.71036232e-01 -1.31402481e+00 4.30742711e-01
6.62054121e-01 2.37588286e-01 -2.36675888e-01 -6.77217126e-01
-4.73570466e-01 -2.47616947e-01 -1.07168269e+00 6.09747171e-01
7.62903512e-01 -6.68205023e-01 -6.96555912e-01 1.43094212e-01
1.01062059e+00 -6.20158494e-01 -7.39087999e-01 3.98402035e-01
2.74976283e-01 -1.27451324e+00 1.26175296e+00 -7.54056051e-02
8.84743035e-01 -7.31637180e-01 -1.49592906e-01 -1.68385947e+00
-8.56781721e-01 -7.64265299e-01 -2.85671324e-01 1.04964709e+00
-2.49352857e-01 -5.49975395e-01 2.88913637e-01 1.65834352e-01
-3.09708834e-01 -2.72323370e-01 -8.60218406e-01 -6.86490178e-01
-3.04118723e-01 1.68767288e-01 6.63820028e-01 8.39338601e-01
-3.93409371e-01 1.73107430e-01 -1.09061873e+00 5.77773154e-01
9.14491892e-01 6.27705157e-02 3.16470146e-01 -3.12713563e-01
-2.83550531e-01 -2.98993766e-01 -1.52612880e-01 -1.65937090e+00
-5.07999718e-01 -2.40236506e-01 1.78195223e-01 -1.47863317e+00
4.66297776e-01 4.62178588e-02 -4.94070053e-01 2.75833774e-02
-6.58571422e-01 3.45305622e-01 2.28646100e-01 6.56763315e-01
-4.69128221e-01 7.53400922e-01 1.51551354e+00 -2.40221873e-01
1.42278448e-02 -3.45434815e-01 -9.01991248e-01 4.54597056e-01
5.91686666e-01 -1.16531886e-01 -1.01988025e-01 -9.55225408e-01
-3.12851250e-01 2.56348580e-01 6.50451660e-01 -1.20053232e+00
4.46345359e-01 1.52014568e-02 6.56386137e-01 -4.97880459e-01
1.31240994e-01 -7.07162201e-01 3.09577107e-01 6.10548258e-01
-2.30563372e-01 6.84179887e-02 2.20254004e-01 7.46696413e-01
-5.57766914e-01 1.51927352e-01 1.15296805e+00 -1.32054180e-01
-7.18282282e-01 1.92316398e-01 -6.26693368e-02 -2.57194012e-01
5.23340940e-01 -5.18918745e-02 -8.71818244e-01 -5.75135589e-01
-5.02654195e-01 1.79775566e-01 6.15253210e-01 5.19676268e-01
9.45670009e-01 -1.16597700e+00 -9.71536100e-01 2.06739247e-01
-3.02312642e-01 -1.90531448e-01 1.07086408e+00 1.07774067e+00
-6.73106492e-01 2.84208864e-01 -6.67613149e-01 -3.80293906e-01
-1.13691437e+00 7.65695512e-01 4.30768877e-01 -2.37270474e-01
-5.74236691e-01 7.03894317e-01 3.46982270e-01 4.32309628e-01
-4.78195660e-02 -3.16348642e-01 -1.45883396e-01 -3.01703036e-01
1.00413144e+00 6.59912646e-01 -3.59179676e-02 -6.99065804e-01
-6.85099512e-02 4.64548856e-01 -4.17442977e-01 1.84779808e-01
1.38018978e+00 -9.12940919e-01 -4.07425463e-01 -2.35908434e-01
1.15032649e+00 -1.05115779e-01 -1.82488453e+00 -5.40739417e-01
-3.31740439e-01 -9.66252506e-01 3.59824240e-01 -1.05721474e+00
-1.29881501e+00 7.78851688e-01 9.26242948e-01 5.58734946e-02
1.76560938e+00 -2.82322019e-01 1.25479650e+00 -3.12686026e-01
-3.08889411e-02 -5.87957084e-01 3.26238364e-01 4.63290066e-01
1.27275872e+00 -1.19689524e+00 3.83563861e-02 -2.33851731e-01
-5.92906237e-01 9.86163914e-01 4.65452045e-01 -2.26669341e-01
2.93207616e-01 1.19407605e-02 1.42014148e-02 2.46875867e-01
-6.51323974e-01 1.89585149e-01 2.23009869e-01 3.80685061e-01
-4.30907831e-02 -2.68048733e-01 -1.88073590e-02 4.01089162e-01
2.44894966e-01 3.70988488e-01 7.21049130e-01 5.26361346e-01
-5.48135221e-01 -5.78633606e-01 -7.56967604e-01 2.15483189e-01
-4.90406781e-01 -6.06197357e-01 2.55223006e-01 8.38070512e-02
1.06047593e-01 1.06396115e+00 -8.87480155e-02 -5.30775487e-01
2.61712372e-01 -5.88126242e-01 2.78800249e-01 2.86317989e-02
-3.43180448e-01 9.26848873e-02 -3.76482815e-01 -6.98563576e-01
-3.97054672e-01 -3.51859391e-01 -5.78622460e-01 -6.32451177e-01
-1.39949679e-01 -1.81597635e-01 2.61137098e-01 6.86568975e-01
7.66712725e-01 6.98306084e-01 7.88142383e-01 -1.06905282e+00
-8.43414962e-01 -1.33434355e+00 -5.39833486e-01 5.80598712e-01
1.12033856e+00 -2.65902638e-01 -7.34242618e-01 3.66522044e-01] | [11.463809967041016, -2.5396111011505127] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.